Memoirs of Museum Victoria Volume 62 Numbers 1 and 2 31 December 2005 Front cover Top left: a new species of seastar from Mauritius, Ailsastra eleaumei, described by P. Mark O’Loughlin and Francis W.E. Rowe. Top centre: an earbone of a fossil dolphin of the family Delphinidae from the Pliocene Whalers Bluff Formation, Portland, Victoria, described by Erich M. G. Fitzgerald. Bottom: a new species of fossil sea urchin, Spatagobrissus dermodyorum, described by Francis Holmes from the early Middle Miocene Glenforslan Formation near Blanchetown, South Australia. Back cover: male reproductive organs of a new species of millipede, Lissodesmus dignomontis , from Tasmania, described by Robert Mesibov. Memoirs of Museum Victoria 62(1): 1-66 (2005) ISSN 1447-2546 (Print) 1447-2554 (On-line) http://www.museum.vic.gov.au/memoirs/index.asp Homalonotid trilobites from the Silurian and Lower Devonian of south-eastern Australia and New Zealand (Arthropoda: Trilobita: Homalonotidae) Andrew C. Sandford Department of Earth Sciences, University of Melbourne, Victoria 3010, Australia (andrewsandford@hotmail.com) Abstract Sandford, A.C. 2005. Homalonotid trilobites from the Silurian and Lower Devonian of south-eastern Australia and New Zealand (Arthropoda: Trilobita: Homalonotidae). Memoirs of Museum Victoria 62(1): 1-66. Trilobites belonging to the Homalonotidae are well represented in the Silurian and Early Devonian of south-eastern Australia and New Zealand, and are a significant component of the family world- wide. Their description provides an opportunity to review relationships between species and higher order taxa. A new genus Wenndorfia and two new sub- genera Trimerus ( Ramiotis ) and T. ( Edgillia ) are described, and revised diagnoses are given for Trimerus, Homalonotus, Dipleura , Digonus and Parahomalonotus. Species described or redescribed from central Victoria include Homalonotus williamsi sp. nov., H. talenti sp. nov., Dipleura garratti sp. nov., Digonus wenndorfi sp. nov., Trimerus ( Trimerus ) vomer (Chapman, 1912), T. (T.) harrisoni (McCoy, 1876), T. {Edgillia) kinglakensis (Gill, 1949), T. (E.) jelli sp. nov., T. { Ramiotis ) rickardsi sp. nov., T. (R.) tomczykowa sp. nov., T. (R.) otisi sp. nov., T. (R.) thomasi sp. nov. and Wenndorfia lilydalensis (Gill, 1949). Tasmanian species described include T. (R.) iani sp. nov., Brongniartellal sp. and D. zeehan- ensis (Gill, 1949). Wenndorfia expansa (Hector, 1876) (= H. {Burmeisteria) huttoni Allan, 1935, = D. margaritifer Wenndorf, 1990) from New Zealand is redescribed. Complex relationships between trilobite faunal composition and taphonomy demonstrate that homalonotid assemblages are inadequately described by the biofacies concept. A recurrent relationship can be recognised between homalonotid-dominated low diversity assemblages and high diversity assemblages in relatively shallower-water facies in which homalonotids are minor faunal elements. These paired assemblages occur variously along a bathymetric gradient that reflects specific environmental tolerances, and precludes the definition of discrete assemblage-facies associations. Keywords Trilobita, Homalonotidae, Silurian, Devonian, Australia, New Zealand, systematics, biofacies Table of Contents Introduction 1 Geological Setting 2 Taphonomy and biofacies 6 Systematic palaeontology 12 Homalonotus 14 Homalonotus talenti sp. nov 16 Homalonotus williamsi sp. nov 19 Brongniartella sp 21 Digonus 21 Digonus wenndorfi sp. nov 23 Digonus zeehanensis (Gill, 1949) 28 Dipleura 28 Dipleura garratti sp. nov 30 Trimerus 33 Trimerus (Trimerus) harrisoni (McCoy, 1876) 34 Trimerus ( Trimerus ) vomer (Chapman, 1912) 35 Trimerus (Edgillia) subgen. nov 38 Trimerus (Edgillia) kinglakensis (Gill, 1949) 39 Trimerus (Edgillia) jelli sp. nov 41 Trimerus (Ramiotis) subgen. nov 43 Trimerus (Ramiotis) rickardsi sp. nov 44 Trimerus (Ramiotis) iani sp. nov 47 Trimerus (Ramiotis) otisi sp. nov 49 Trimerus (Ramiotis) thomasi sp. nov 53 Trimerus (Ramiotis) tomczykowae sp. nov 54 Wenndorfia gen. nov 55 Wenndorfia expansa (Hector, 1876) 56 Wenndorfia lilydalensis (Gill, 1949) 60 Acknowledgements 62 References 63 Introduction Homalonotid trilobites were first recorded from south-eastern Australia and New Zealand as early as the 1860s, and eight species have been described subsequently including Homalonotus expansus Hector, 1876, H. harrisoni McCoy, 1876, H. (Burmeisteria) huttoni Allan, 1935, Trimerus lily- dalensis Gill, 1949, H. vomer Chapman, 1912, T. kinglakensis Gill, 1949, T. zeehanensis Gill, 1949 and Digonus margaritifer 2 Andrew C. Sandford Wenndorf, 1990. The relationships of these taxa have been poorly understood, as most of them were inadequately described from limited populations and/or poorly preserved or incomplete individuals. Museum Victoria houses a large col- lection of homalonotid specimens from Australasia which not only permits the redescription of these species, but also the description of many new species. Both in taxonomic diversity and abundance the homalonotid fauna of south-eastern Australia is of world- wide significance. It is undoubtedly the best preserved representation of the family for the Llandovery to Lochkovian interval, representing about one-third of species so far described (Table 1). Three Llandovery species bring to eight the total number of homa- lonotids recorded from the interval, with two of them being among the three best known taxa. Seven Victorian Ludlow species bring to 15 the total number of homalonotids recorded from that interval. Similarly, four Lochkovian species make up one-quarter of the total number of homalonotids recorded from that interval. The species assigned below to Homalonotus are two of the total of five assigned to that genus globally, whilst the nine species assigned below to Trimerus represent almost half of the species assigned to that genus. A species assigned to Dipleura is the earliest known representative of the genus, and two species assigned to Digonus are among the earliest representatives of that genus. Historically, the systematics of the Homalonotidae has been heavily weighted towards the classification of European and North American species that represent about two-thirds the total number of species so far described. This trend persists in the two most recent taxonomic revisions accompanying the description of north European faunas (Tomczykowa, 1975; Wenndorf, 1990). In contrast and with the exception of Burmeisteria, species from Asia, North Africa and the southern hemisphere have been accommodated in taxa established for the European and North American faunas. However, problems arise with the uncritical application of generic concepts estab- lished for European and North American faunas to the classifi- cation of southern hemisphere homalonotids, and are manifest in the nomenclatural changes suffered by some of the few well documented species. Homalonotus clarkei Kozlowski, 1923 from the Devonian of Brazil has been variably assigned to Digonus (Wolfart et al., 1968), Trimerus! (Tomczykowa, 1975), Burmeisteria (Cooper, 1982), and Dipleura (Wenndorf, 1990). Similarly, B. ( Digonus ) accraensis Saul, 1967 from the Devonian of Ghana has been variably assigned to Trimerus (Tomczykowa, 1975), Burmeisteria (Cooper, 1982) and Dipleural (Wenndorf, 1990). The comprehensive redescription of the Australian homalonotid fauna in this work documents a morphological diversity that demands a more critical usage of established generic concepts. Revised diagnoses are given in this work for Homalonotus, Digonus, Parahomalonotus, Dipleura and Trimerus to accommodate the morphological diversity of the Australasian fauna. The Australasian fauna provides new perspectives on the relationships between species. Trimerus is well represented in the Australian fauna with nine species assigned to the genus. The relationships between these and other species of the genus are formalised in the description of the new subgenera Ramiotis and Edgillia. T. lilydalensis from Victoria shows closest relationships to European species previously assigned to Parahomalonotus. The new genus Wenndorfia is erected to accommodate T. lilydalensis and these closely related taxa, and the concept of Parahomalonotus is revised and restricted to those species close to the type. Reassignment of species to these groups (Table 1) reflects these revised concepts. The Silurian and to a lesser extent the Early Devonian are generally considered periods of cosmopolitanism, and this is reflected in the widespread distribution of species groups and higher taxa. Although restricted to the late Wenlock-early Ludlow, Trimerus ( Trimerus ) is represented in North America, South America, Europe and Australia, with T. (T.) vomer from Victoria and T. (T.) johannis Salter, 1865 from England defining a distinct trans-global species group. Wenndorfia lily- dalensis from the early Pragian of Victoria is closest to poorly known species from the late Lochkovian of Europe. The mor- phological diversity present in the Australian fauna is more a reflection of the age of the fauna rather than its palaeogeo- graphic setting. Most of the Australian species are from the Llandovery, Ludlow and Lochkovian intervals in which many of the species documented from elsewhere are poorly under- stood and have had little influence on systematics. Trimerus ( Edgillia ) is erected for a distinct group of globally distributed Pndolf-early Lochkovian species typified by T. (E.) king- lakensis from central Victoria but otherwise very poorly repre- sented in the fossil record. Similarly, T. ( Ramiotis ) is erected for a widespread group of Llandovery-Ludlow species poorly represented in the fossil record outside Australia. Siluro- Devonian homalonotid groups not represented in Australia and New Zealand include Parahomalonotus, Burmeisteria, Burmeisterella, Scabrella and Arduenella . Geological Setting In central Victoria, trilobite-bearing Silurian to Lower Devonian marine sedimentary rocks of the Murrundindi Supergroup crop out in a structurally bound area known as the Melbourne Zone (VandenBerg et al., 2000). Lithologies are predominantly interbedded mudstones, siltstones and fine to medium-grained sandstones, although coarse sandstones and conglomerates occur at various horizons. Carbonates, known only from the Lower Devonian in the central and eastern areas of the Melbourne Zone, are poorly represented and occur as isolated, discrete limestone bodies. Homalonotids are restricted in distribution to non-carbonate sequences between Melbourne, Heathcote and Lilydale in the western and central areas of the Melbourne Zone (Fig. 1). The lithostratigraphy of these homalonotid-bearing sequences has been variously described. The definition, correlation and age of the various trilobite-bearing lithostratigraphic units described from these sequences has been reviewed by Rickards and Sandford (1998) and Sandford (2000, 2002, 2003, 2004). Stratigraphic nomenclature for the Llandovery to early Ludlow sequence used in this work (Fig. 2) follows Rickards and Sandford (1998: fig. 2) and Sandford (2000: fig. 2), and that for the overlying Ludlow to Pragian sequences follows Sandford (2002, fig. 2). Homalonotids first appear in the latest Table 1. Distribution of Silurian and Devonian homalonotid trilobites. Australian species are underlined. Homalonotid trilobites from the Silurian and Lower Devonian 3 4 Andrew C. Sandford Figure 1. Distribution of fossil marine faunas with homalonotids in south eastern Australia and New Zealand. Localities for the Heathcote, Springfield- Kinglake West and Lily dale areas are detailed in Figures 8, 11 and 23. LATE LLANDOVERY WENLOCK LUDLOW PRfDOLI Homalonotid trilobites from the Silurian and Lower Devonian 5 EMStAN PRAGIAN LOCHKOVIAN LUDFORDIAN GORSTIAN HOMERIAN SHEINWOODIAN TELYCHIAN graptolite,conodont and brachiopod biozones recognised in Victoria ucATum-rc SPRINGFIELD- MELBOURNE- HtAIHLUIt kinglakewest ULYDALE pirenae sulcatus nilssoni iudensis nassa/parvus testis/ lundgteni griestonieftsis Cfispus turriculatus Boucotia loyolensis- Boucotia australis Boucotia jartaea Hotopormello pientlensis Aegiria thomasi TOP NOT EXPOSED PUCKAPUNVAL FORMATION III MCIVOR FORMATION MELBOURNE FORMATION VAN YEAN FORMATION WAPENTAKE FORMATION COSTERFIELD SLT BASE NOT EXPOSED- TOP NOT EXPOSED NORTON GULLY SANDSTONE WILSON CREEK SHALE PUCKAPUNVAL FORMATION Cullin' Qu.vry 55 M™ HUMEVALE SILTSTONE MELBOURNE FORMATION YANYEAN FORMATION BYLANDS SILTSTONE BASE NOT EXPOSED AAAAAAAA LILYDALE LIMESTONE Verirtgbeig 5 S Mjfm HUMEVALE SILTSTONE CHRISTMAS HILLS FORMATION MELBOURNE FORMATION YANYEAN FORMATION Mullum Mullum Sandstone Member ANDERSON CREEK FORMATION WarrandyiiCcjl Mem BASE NOT EXPOSED -J REEF TON GROUP BATON RIVER GROUP W, lilydolertsis | T(E.)jeili |d. wenndorfi T. (E.) kingiakensis BELL SHALE FLORENCE QUARTZITE ZEEHAN ci r.mjodsi y I | T(R.) thomasi jHiYi/tonosc D. garratti | Z(T.) harrisoni I 7 ™.! I T?spp. | T.(R-) tomczykowae C H I NTI N F M N | T.fRJ rickordsi TZRJfanl B.sp. TIGER RANGE RICHEA SLT SPRINGFIELD SANDSTONE CENTRAL VICTORIA DARRAWEIT GUI M PROVINCE CELL QUARTZITE TASMANIA NEW ZEALAND Figure 2. Stratigraphic distribution of homalonotids in the Heathcote, Springfield-Kinglake West and Lilydale sequences in central Victoria, and in Tasmania and New Zealand. Stratigraphic scheme for the Llandovery-Ludlow of central Victoria follows Rickards and Sandford (1998), and for the Ludlow-Lochkovian follows Sandford (2002). D.zeehanensis W.expansa 6 Andrew C. Sandford Llandovery Chintin Formation at Springfield and occur at various horizons in the overlying sequences, ranging high in the Mt Ida Formation at Heathcote in strata considered to be late Lochkovian in age, and ranging up to the top of the Humevale Formation at Lily dale in strata considered to be early Pragian in age. Homalonotids do not occur in the various Lochkovian-Pragian trilobite faunas from the western areas of the Melbourne Zone, east of Lilydale. Silurian-Early Devonian sequences occur widely in Tasmania, but homalonotids are known only from the south-western areas, in the Llandovery Richea Siltstone northwest of Maydena and in the Lochkovian Bell Shale at Zeehan. On the south island of New Zealand, homalonotid trilobites are known from the Lochkovian- ?Pragian Baton Formation, Baton River and from the ?Pragian- Emsian Lankey Limestone at Reefton. Silurian and Devonian trilobite faunas occur widely across the eastern states of Australia, and the restricted distribution of homalonotids to the south-western areas of the Melbourne Zone and to south-western Tasmania is noteworthy. Acastid trilobites are similarly distributed, whereas the distribution of harpetid trilobites complements that of the homalonotids and acastids, ranging from eastern Victoria to northern Queensland. The limited distribution of homalonotids in eastern Australia parallels their distribution in Europe across the Rheinischen- Bohemischen facies boundary. The barrier to dispersal in east- ern Australia is not apparent for other trilobites, particularly the phacopids and the calymenids. Echidnops, the Ananaspis typhlagogus (Opik, 1953) species group and individual species such as Sthenarocalymene sp. A of Chatterton, Johnson and Campbell, 1979 are represented both in central Victoria and New South Wales. It is currently held that the sequences of the Melbourne Zone were deposited on a low gradient shelf on the eastern margin of the Australian continental plate. In contrast, the sequences in eastern Victoria, New South Wales and Queensland were deposited on off-shore island-arc platforms many hundreds of kilometres from the mainland (Foster et al., 1999). It seems that although many trilobite groups were able to establish themselves in these settings, homalonotids (and acastids) were unsuccessful in dispersing from the continental margin. Isolation of the homalonotids can be attributed to pref- erences for continental margin environments. Taphonomy and biofacies The homalonotid-bearing trilobite faunas from the Silurian and Lower Devonian of central Victoria are quite variable in taxo- nomic composition and relative abundance of species, taphon- omy, associated lithofacies and associated non-trilobite faunal elements. Several faunas can be assigned to homalonotid- dominated biofacies including the Homalonotid Association (Mikulic, 1999), the Trimerus delphinocephalus-Dalmanites limulurus Association (Mikulic, 1999) and the Digonus Biofacies (Fortey and Owens, 1997). Other faunas resemble the Acaste-Trimerus Association (Thomas, 1979). Mikulic (1999) emphasised trilobite elements of Boucot’s (1975) shallow water ‘Homalonotid-P/ectorcotns BA-1 Community’ in defining the Homalonotid Association. Mikulic nominated the low diversity/monospecific Ludlow assemblages dominated by Homalonotus dawsoni Hall, 1860 of the Stonehouse Formation, Arasaig, Canada as typical. The bio- facies is associated with coarse- to fine-grained siliciclastics of nearshore environments and a taphonomy dominated by dis- articulated exoskeletal elements. Mikulic assigned faunas ranging worldwide from the Caradoc through to the Givetian to the Homalonotid Association. Taxa commonly associated with this assemblage include Acaste, Acastella, Encrinurus, Scotiella, Dalmanites and calymenids. The Homalonotid Association can be recognised in the Upper Silurian-Lower Devonian sequences at Heathcote in central Victoria, repre- sented by populations of Homalonotus talenti sp. nov., Trimerus ( Ramiotis ) thomasi sp. nov. and Digonus wenndorfi sp. nov. at successive horizons. Most typical is the mono- specific assemblage at the type locality of H. talenti (PL6650, Heathcote). The population is preserved in thick-bedded coarse sandstones of the Mclvor Formation and is represented pre- dominantly (94%) by isolated sclerites. At its type locality (PL2203, Heathcote), the population of D. wenndorfi is repre- sented entirely by isolated sclerites. The species occurs in medium to coarse sandstones of the Mt Ida Formation in asso- ciation with rare Acastella sp. and an abundance of very large bivalves and brachiopods (Fig. 3). The latter indicate alignment of the fauna to the ‘big-shell community’ of Boucot and Johnson (1967), later formally designated benthic assemblage BA-1 (Boucot, 1975) and interpreted to be a very shallow- water community. The D. wenndorfi fauna at PL2203 can also be assigned to the Digonus Biofacies of Fortey and Owens (1997). The Digonus Biofacies was recognised as the shallow- est in a succession of depth-related trilobite associations of Early Devonian shelf environments, with deeper associations named after their characteristic key taxa Scutellum, Lepidoproetus, Odontochile and Otarion. Although Fortey and Owens did not provide any description of these associations, their scheme follows Chlupac’s (1983, 1987) depth-related succession of lithofacies and trilobite associations described from the Czech sequences as Major Assemblage Groups I-V. The association of Digonus, Acastella and an abundance of bivalves at PL2203 also resembles Thomas’ (1979) Trimerus- Acaste Association, described from shallow water facies of the British Wenlock. The Trimerus -Acaste Association can be regarded as the Silurian equivalent of the Devonian Digonus Biofacies, both considered here as synonyms of the Homalonotid Association. Low diversity trilobite faunas with species of Trimerus and Dalmanites (or Odontochile) as the two dominant taxa occur at successive horizons in deeper-water sequences south of Heathcote. An earliest Ludlow fauna from PL386, Wandong is dominated by an association of D. wandongensis Gill, 1948 and T. (T.) vomer. The fauna is preserved in siltstones as isolated sclerites and partly disarticulated exoskeletons, some of which represent moult assemblages and indicate a depositional envi- ronment below storm wave base (Sandford, in press). A late Ludlow fauna from PL 1898, Eden Park is similarly preserved, dominated by T. ( Ramiotis ) otisi sp. nov. (relative abundance 96%) with rare Odontochile formosa Gill, 1948 (relative abun- dance 4%). A late Lochkovian fauna from the upper massive siltstone beds at Middendorps Quarry (PL252, Kinglake West) Homalonotid trilobites from the Silurian and Lower Devonian 7 Figure 3. Sandstone slab from PL2204 (Thomas’ F4, Parish of Dargile), Heathcote with bedding plane showing numerous isolated pygidia and cranidia of Digonus wenndorfi in various orientations and with large bivalves and brachiopods indicative of Boucot and Johnson’s (1967) ‘big shell’ community. is dominated by T. ( Edgillia ) kinglakensis (relative abundance 99%) with rare O. formosa. These assemblages can be assigned to the T. delphinocephalus-D. limulurus Association, typified by a trilobite fauna dominated by these taxa from the Wenlock Rochester Shale, USA. Mikulic (1999) interpreted the commu- nity as a deeper-water facies association, preserved as disartic- ulated exoskeletons in fine-grained, deep water/low energy sili- ciclastics and aligned with BA 4 and BA 5 communities, Brett’s (1983) Amphistrophia-Dalmanites Community in particular. Mikulic (1999) tentatively assigned assemblages occurring in shallower facies (BA3) to the T. delphinocephalus-D. limulurus Association. Equivalent shallower assemblages in central Victoria include the O. formosa- dominated fauna from the lower coquinal siltstones and sandstones at Middendorps Quarry, in which T. (E.) kinglakensis is rarer. At a number of localities in the uppermost Humevale Siltstone at Lilydale Wenndorfia lilydalensis occurs with Acastella frontosa Shergold, 1968, often as dominant elements of the fauna. These faunas resemble Thomas’ (1979) Trimerus- Acaste Association, although the facies at Lilydale suggest a deeper-water setting. Another acastid, Acaste lokii Edgecombe, 1993, occurs lower in the Humevale Siltstone in the Lily- dale area, its stratigraphic range overlapping with that of W. lilydalensis, but the two species never occur together. Despite the recognition of homalonotid biofacies at several localities, the homalonotid faunas of south-eastern Australia are otherwise inadequately described by the biofacies concept. Patterns can be recognised demonstrating more complex relationships between homalonotid relative abundance, faunal diversity, taphonomy and lithofacies. A skewed inverse relationship between relative abundance and faunal diversity can be demonstrated, with lower diversity trilobite associations having homalonotids proportionally over-represented, and higher diversity associations with homalonotids under- represented. In the higher diversity faunas, dalmanitids or proetids are frequently the dominant taxon, although acastids, phacopids and rarely odontopleurids may also dominate. The relationships between taxonomic diversity and homalonotid relative abundance is best expressed graphically (Fig. 4). Several species from central Victoria including Digonus wenndorfi, Trimerus ( Edgillia ) kinglakensis, T. ( Ramiotis ) otisi and Wenndorfia lilydalensis are known from both low Andrew C. Sandford AB2, AB1, AA34 r AA14, G62,G58 100 O i t Q PL 252 {upper slltstone) ijD 4 .PL1 898 1 LOWER DIVERSITY HOMALONOTIDS \ FAUNAS OVER-REPRESENTED LO Q 80 \ H 45 \ Q Oh46 0 S5R 5 Z C' C3 %\ ° 7 Qg 59 O _i < O i 1\ O a, ° o 60 q i X % u_ o 7 -\ O gm u z < 40 Q <v. z 3 00 < LU 0 Q2 J7 > O- ^ O G 1 Qa B 1 0 < *- _l LU CC 20 P L 3 0 0 Q 0-1 __ HIGHER DIVERSITY FAUNAS BWSO QPL1990 "O' Qg 69 G 1 3 O QR25 Q PL2S2 P G87 n n G5 ^ R 2 1 S R25E 0 HOMALONOTIDS UNDER-REPRESENTED (coquinas) 1 2 3 4 5 6 7 8 9 1 0 TAXONOMIC DIVERSITY Figure 4. Relationship between trilobite faunal diversity and homalonotid relative abundance. Only homalonotid-bearing faunas represented by more than 10 specimens are plotted. The curve connects plot-points of hypothetical faunas where species are represented in equal proportional relative abundance. Homalonotid faunas cluster in two groups, lower diversity faunas with homalonotids over-represented (above the line) and high diversity faunas with homalonotids generally under-represented (below the line). diversity, homalonotid-dominated assemblages and high diversity assemblages in which homalonotids are under-repre- sented. From these occurrences, a pattern emerges revealing a relationship between trilobite assemblage composition, taphon- omy and lithofacies (Table 2). This relationship is best illus- trated in the distribution of T. (E.) kinglakensis in the upper massive siltstones and the lower bedded coquinal siltstones and sandstones at Middendorps Quarry, where a marked contrast in facies accompanies differences in faunal composition. In the upper beds T. ( E .) kinglakensis is extremely dominant (relative abundance 99%) in a low-diversity assemblage accompanied by rare Echidnops hollowayi Sandford, 2002 and Odontochile formosa. Individuals are preserved predominantly as partly articulated and frequently incomplete exoskeletons or other- wise as isolated pygidia. Despite the exceptionally high pro- portion of partly articulated exoskeletons, complete fully artic- ulated exoskeletons are absent. Many of the partly articulated exoskeletons can be interpreted as moult assemblages, pre- served in typical phacopid fashion with the cephalon displaced, often inverted and/or lying underneath the thoracopygon (Figs 5 A, 5D). Partly-articulated thoraces can be interpreted as ele- ments of moult assemblages scattered by the movements of the animal during the exuvial process. Indeed, it is probable that the whole population of T. (E.) kinglakensis in the massive silt- stones (including the isolated cephala and pygidia) are derived from exuviae rather than dead individuals. Well bedded, siltstones and graded sandstones with poorly sorted, bioclastic basal coquinal lags underlie the siltstones at Middendorps Quarry and yield a much more abundant and diverse benthic fauna. By world standards ten taxa is low diver- sity for an Early Devonian trilobite assemblage, but for central Victoria this is one of the most diverse. The assemblage is dom- inated by Lepidoproetus sp. and Odontochile formosa with a combined relative abundance of 64%. Trimerus (Edgillia) Homalonotid trilobites from the Silurian and Lower Devonian 9 Table 2. Relationships between lithology, relative abundance and taphonomy for Trimerus (Edgillia) kinglakensis, Trimerus ( Ramiotis ) otisi and Wenndorfia lilydalensis. Populations of the latter species in the Eden Park and Humevale siltstones are taken as a whole. Trimerus (Edgillia) kinglakensis Humevale Formation PL252 (type locality) Trimerus (Ramiotis) otisi Eden Park Formation Wenndorfia lilydalensis Humevale Formation LITHOLOGY siltstone coquina siltstone coquina siltstone (Clonbinane Sandstone Member: PL300, PL6642) coquina (PL 1804) TOTAL specimens 286 15 132 8 58 1 RELATIVE ABUNDANCE 99% 15% 96% 24% 26% 2% TAPHONOMY complete dorsal exoskeleton 0 40% 1% 36% 0 100% cephalothorax <1% 0 1.5% 13% 2% 0 cephalon with displaced thoracopygon 12% 0 2% 0 2% 0 cephalon with displaced thorax 10% 0 1.5% 0 0 0 cephalon with displaced thorax and displaced pygidium 1% 0 0 0 0 0 cephalon 11% 0 11% 0 17% 0 isolated cranidium <1% 0 17% 0 12% 0 isolated librigena 0 0 2% 0 5% 0 isolated hypostome 0 0 1% 0 0 0 articulated/partly articulated thorax 9% 0 6% 0 0 0 loosely disarticulated thorax <1% 0 0 0 0 0 isolated thoracic segment 0 0 2% 13% 21% 0 partly disarticulated thorax, displaced pygidium 3% 0 0 0 0 0 complete/partly complete thoracopygon 36% 0 7% 13% 0 0 isolated pygidium 15% 60% 48% 25% 41% 0 articulated/partly articulated specimens 84% 40% 70% 75% 21% 100% TAPHOFACIES TIV Till TIV Till TIV Till kinglakensis is a minor faunal element (relative abundance <3%). In further contrast to its representation in the overlying siltstones, the population of T. (E.) kinglakensis in the coquinas has a high proportion (13%) of complete, fully articulated exoskeletons. Populations of other species in these beds also have a high proportion of complete exoskeletons, including Lepidoproetus sp. (42%), Maurotarion sp. (33%), Echidnops hollowayi (21%), Leonaspis sp. (11%) and Dicranurus king- lakensis Gill, 1947 (6%). Similar relationships between faunal composition and taphonomy can be demonstrated for Trimerus ( Ramiotis ) otisi and Wenndorfia lilydalensis. The single known complete exoskeleton of W. lilydalensis from PL 1804, Coldstream is exceptional in several respects. It occurs in a bioclastic coquinal lag within a fine-grained sandstone (Fig. 5C), ident- ical to the style of preservation of T. ( Trimerus ) harrisoni at PL 1820, West Brunswick (Fig. 5B) where most specimens (58%) are complete exoskeletons. The relative abundance of W. lilydalensis at PL 1804 is very low (2%), exceptionally under- represented in an assemblage of relatively high diversity (8 species). By contrast, all other specimens of the species including isolated tergites and partly disarticulated exo- skeletons occur in thick-bedded siltstones at various localities in the Lilydale area. The relative abundance of the species in the siltstones ranges between 6% and 100%, with assemblages comprising one to six species in which W. lilydalensis is moderately under-represented or over-represented (Fig. 4). In the siltstones of the Eden Park Formation in the Eden Park area T. (R.) otisi occurs in monospecific assemblages or dominates a low diversity assemblage also containing rare Odontochile formosa. In these beds T. (R.) otisi is preserved predominantly as isolated sclerites or partly articulated exoskeletons, many of which can be recognised as moult assemblages. The holotype is a complete exoskeletion with the cephalon displaced and lying underneath the thoracopygon. The ratio of articulated speci- mens is 25%, 5% being typical phacopid moult assemblages and less than 1% being fully articulated individuals. T. (R.) otisi also occurs in the overlying Clonbinane Sandstone Member, in medium grained sandstone and coquinal bioclastic lags. At PL300, Clonbinane it is under-represented (relative abundance 17%) in a higher diversity assemblage also containing Echidnops hollowayi, O. formosa, Dicranurus sp. and an 10 Andrew C. Sandford Figure 5 A, D. Moult assemblages of Trimerus ( Edgillia ) kinglakensis in the upper siltstones at PL252, Middendorps Quarry, Kinglake West, Humevale Siltstone. A, upper right, displaced and inverted cephalon lying underneath thoracopygon (only anterior margin visible); left, thora- copygon with displaced and rotated cephalon. D, thoracopygon with displaced and rotated cephalon. B-C. Complete dorsal exoskeletons in sand- stone bioclastic coquinas. B, Trimerus ( Trimerus ) harrisoni from PL 1820, Brunswick, Melbourne Formation. C, Wenndorfia lilydalensis from PL 1805, Coldstream. Humevale Siltstone. odontopleurid. More than one third of all trilobites in these beds are complete, fully articulated exoskeletons. As in the bedded silt- stones and coquinal lags at Middendorps Quarry, the assemblage at PL300 is associated with a rich echinoderm fauna. The taphonomy of the trilobite exoskeleton is an important palaeoenvironmental indicator. The multitude of sclerites of which the exoskeleton was composed responded in a complex way to the depositional environment in which they were ultimately preserved. Detailing the proportions of fragmented tergites, convex-up tergites, articulated tergites, enrolled individuals and moult assemblages in trilobite populations from the Middle Devonian Hamilton Group, New York, Speyer and Brett (1986) defined a sequence of trilobite taphofacies (taphofacies 1A-B, 2A-B, 3A-B and 4A-C) interpreted to reflect depth and rates of sedimentation. Sandford (2002) described a similar depth-related sequence of trilobite taphofa- cies for the phacopid-dominated assemblages (TpI-TpIV) and the homalonotid-dominated assemblages (ThI-ThIV) of the Late Silurian-Early Devonian of south-eastern Australia. Whilst there is some credit in distinguishing between taphonomies for thick-shelled forms and thin-shelled forms in shallower facies, these forms occur in both phacopids and homalonotids. The separate Tp-Th classification of tapho- nomies is not justified, and abandoned here in place of a broader TI-IV classification, which can be more widely applied to trilobite taphofacies. Although this study is based almost entirely on museum specimens and the proportion of convex-up specimens cannot be accurately determined, Speyer and Brett’s taphofacies can be recognised for the homalonotid faunas of south-eastern Australia. Assemblages preserved in fine-grained Homalonotid trilobites from the Silurian and Lower Devonian 11 articulated dement* in well wrtcd and bedded sjndilonrl in poorly sorted b*ocla*tic siltstone bt S coquina* (moult assemblage*] in siltstone; few complete cuoiketiorij Wenndorfia lityddensh Trimerus (EdgiHio)jelli Trimerus (Btgillfa) klngfakensis Trimenis ( Ramiotis ) otisi Homalonotus talenti Homalonotus willion Dipleura garratti Trimerus (Trimenis) horrisonl Trimerus (Trimerus) vomer Trimenis (Ramioiis) tomczykowae; Trimerus spp. Trimenis (Ramiotis) rlckardsl Trimerus (Ramiotis) iani Figure 6. Facies and environmental distribution of Australian homalonotids. lithologies with a high relative abundance of moult assem- blages (or partly disarticulated exoskeletons), a low abundance or absence of complete, fully articulated exoskeletons, a low degree of isolated sclerites and low or negligible breakage can be assigned to Sandford’s taphofacies TIV (-Speyer and Brett’s taphofacies 4A). The preservation of moult assemblages indicates deposition at depths below maximum storm wave base, where current activity was minimal and reworking of exoskeletal elements was negligible. Assemblages preserved in poorly sorted bioclastic coquinal lags with a high relative abundance of complete, outstretched exoskeletons, an absence of moult assemblages and a moderate degree of breakage can be assigned to Sandford’s taphofacies Till. This taphofacies is interpreted to reflect depths around storm wave base and is analogous to Speyer and Brett’s taphofacies 3A, although dif- fering from it in the lower representation of moult assemblages and enrolled specimens relative to outstretched specimens. The basal lags represent proximal tempestites generated from the peak of storm activity, during which highest energy currents rapidly concentrated and buried large bioclasts (both living and dead), with complete exoskeletons representing smothered individuals. Taphofacies associations for Wenndorfia lilydalensis, Trimerus ( Edgillia ) kinglakensis and Tn'merus ( Ramiotis ) otisi populations demonstrate that the low diversity, homalonotid- dominated assemblages inhabited deeper bathymetries than the higher diversity assemblages in which homalonotids are under- represented. A similar relationship can be demonstrated for populations of Digonus wenndorfi from the Heathcote sequence, but in a shallower water context. At several localities including the type locality PL2204, Heathcote, D. wenndorfi occurs in abundance in monospecific or very low diversity assemblages. The taphonomy is characterised by a high relative abundance of isolated sclerites (80-95%), minimal breakage and a high concentration of elements, with the articulated spec- imens mostly represented by cephala. Thick-bedded, medium to coarse-grained, well-sorted sandstone lithologies are associ- ated with this taphofacies. Sandford (2002) designated this preservation taphofacies Til, equivalent to Speyer and Brett’s (1986) taphofacies IB, and it is interpreted to reflect deposition at depths around normal wave base, where moderate energy current action gently winnowed finer sediments, and disartic- ulated and concentrated skeletal elements in coarser lithologies. At PL2327, Heathcote D. wenndorfi occurs in a high diversity 12 Andrew C. Sandford fauna of ten species as a proportionally represented species (relative abundance 11%). The assemblage is dominated by Echidnops hollowayi and Odontochile formosa. In contrast to the preservation of the D. wenndorfi- dominated assemblages, that at PL2327 is characterised by a high proportion of iso- lated and broken sclerites (broken specimens 20%, isolated sclerites 98%). This preservation is generally associated with coarser-grained lithologies, typically thick, coquinal sand- stone with poorly sorted, randomly oriented bioclasts and can be assigned to Sandford’s taphofacies TI, equivalent to Speyer and Brett’s taphofacies 1A. A high proportion of iso- lated and broken sclerites suggests an environment of deposi- tion in very shallow environments (<20 m) above normal wave base, with medium to high energy oscillatory wave- generated currents reworking, fragmenting and winnowing skeletal elements. The paired assemblages occur variously in very shallow, moderate or deep water facies that reflect specific environ- mental tolerances and preclude the definition of discrete assemblage-facies associations. The facies distribution of south-eastern Australian homalonotids is shown in Fig. 6. The relationship between homalonotid-dominated low diversity assemblages in deeper water facies and high diversity assemblages in which homalonotids are minor faunal elements in shallower-water facies are not generally consistent with Mikulic’s (1999) and Thomas’ (1979) observations that homa- lonotid species dominating shallow- water assemblages occur as rare elements of assemblages in deeper facies. For Digonus wenndorfi, Trimerus ( Edgillia ) kinglakensis and Trimerus ( Ramiotis ) otisi the opposite is true, being under-represented or rare in shallower environments and dominant in deeper environments. Rather than declining in abundance in even deeper environments, species are replaced by a different species. T. (E.) kinglakensis inhabits the deeper environments of the Lochkovian, whereas T (E.) jelli and D. wenndorfi inhabit the contemporary shallow and very shallow environ- ments. Similarly, T. ( R .) otisi inhabits the deeper environments of the late Ludlow, whereas T (R.) thomasi inhabits the con- temporary shallow environments. An exception is the distribu- tion of W. lilydalensis, which is consistent with both the pattern described above and with the observations of Mikulic and Thomas. As noted above, W. wenndorfi is rare in the shallower coquinal sandstones at PL 1804 and dominates many of the trilobite faunas in the deeper siltstones of the upper Humevale Siltstone at Lilydale. In the siltstones, the association of Acastella frontosa and W. lilydalensis alternates with another association of calymenids and/or phacopids, variously com- prising Sthenarocalymene angustior (Chapman, 1915), Nephranomma debrae, N. lynnae and Lochkovella longisulcata (Shergold, 1968). The latter association is dominant in the deepest water facies, with taphonomies including enrolled indi- viduals and often occurring with Harpidella sp. in abundance. At the few localities where the association of A. frontosa and W. lilydalensis occurs in these deepest water facies it is only in low abundance with the calymenid-phacopid associations dominant. These mixed faunas resemble the deeper water equivalents of Thomas’ Acaste-Trimerus association, where Calymene becomes common. Systematic palaeontology The description of the homalonotid trilobites from central Victoria is based on collections in Museum Victoria (NMV), registered with pre- fix P. This collection includes specimens previously held by the Geological Survey of Victoria (previously registered with prefix GSV) and the University of Melbourne, Geology Department. Other speci- mens cited include those at the Geological Survey of New Zealand (NZGS) (registered with prefix AR) and at the Canterbury Museum, Christchurch, New Zealand (registered with prefix ZFc). Trilobite specimens are preserved in mudstones as internal and external moulds. Internal moulds were coated with colloidial graphite, latex peels have been made from external moulds, and all have been whitened with a mm onium chloride for photography. Trilobite localities are registered with prefix PL in Museum Victoria. These include previously pub- lished fossil localites of Jutson (1908: pi. 3), Thomas (1940a, 1940b, 1941, 1956, 1960), Gill and Banks (1950), Gill (1940: fig.l; 1945: fig. 2), Talent (1964: fig. 1), Williams (1964: fig. 2), Moore (1965: fig.l), VandenBerg (1970), Garratt (1972, 1977), Jell and Holloway (1983: fig. 1), Holloway and Sandford (1993: fig. 1), Wall et al. (1995: fig. 1) Rickards and Sandford (1998: fig. 6) and Sandford (2000: fig. 7; 2002: fig. 1; 2003: text-fig. 1; 2004: fig. 1; 2005: figs 2-4). Trilobite local- ities in the Heathcote, Kilmore-Kinglake West, Springfield and Lilydale areas are shown in Figs 8, 11 and 23. Terminology used in this study for the description of the trilobite exoskeleton (Fig. 7) generally follows current standard nomenclature reviewed by Whittington (1997) and Whittington and Kelly (1997). Terminology specific to the description of homalonotid trilobites has been introduced by Tomczykowa (1975), Wenndorf (1990) and Whittington (1993). Wenndorf established several new terms including ‘AuBendom’, referred to as the posterolateral pleural spine in this work, and the ‘Gelenkleiste’ in reference to ridges that overhang and partly enclose the trench-like furrow traversing each thoracic segment. Wenndorf also established a set of measurements and derived ratios specifically for the description of homalonotids. In accord with Wenndorf, the glabellar width is measured across LI -LI rather than the across occipital ring (where the axial furrows are very poorly defined) and the pygidial axial width is measured across the second axial ring rather than the first ring, for the same reason. To avoid ambiguity this width is referred to as the preoccipital glabellar width. However in this work, following the description of other trilobites, the glabellar length includes the length of the occipital ring and the pygidial axial length includes the length of the terminal axial piece. Whittington (1993) introduced the term ‘articulating and pleural furrow’ to describe the furrow traversing thoracic segments, considered to be homologous to the independent pleural and articulating furrows of other trilobites. ‘Rib-ring offset’ is a term introduced in this work to describe differ- ences in segmentation of the pygidial axis and pleural lobe in homa- lonotids. Rib-ring offset is a modification of R-P ratios often described for encrinurids, but impractical in describing homalonotid pygidia where segmentation is often poorly defined or effaced posteriorly. It is expressed by the n-th pleural rib when the transverse midline of that rib intersects with the axial furrow opposite a pygidial ring furrow. Order PHACOPID A Salter, 1864 Suborder CALYMENINA Swinnerton, 1915 Superfamily CALYMENOIDEA Burmeister, 1843 Family Homalonotidae Chapman, 1890 Remarks. Organisation of the Homalonotidae here follows Thomas (1977) who discussed previous divisions of the family. Subfamily Homalonotinae Chapman, 1890 Homalonotid trilobites from the Silurian and Lower Devonian 13 CEPHALON (DORSAL VIEW) connective suture (dorsal section) rostral suture preglabellar field preglabellar furrow axial furrow r iK median indentation frontal lobe glabellar lobes (L1-L3) lateral glabellar furrows (SI -S3) apodemal pit muscle scar CEPHALON (VENTRAL VIEW) rostral plate (ventral section) librigenal doublure posterior branch of facial suture (ventral section) THORACIC SEGMENT (OF 13) PYGIDIUM (DORSAL VIEW) articulating half ring articulating furrow articulating facet inter-ring furrow axial ring axial furrow rostral plate (dorsal section) anterior branch of facial suture preocular fixigenal field . eye ridge lateral border furrow palprebral area palpebral lobe visual surface subocular groove librigenal field sutural furrow posterior branch of facial suture (dorsal section) postocular fixigenal area paraglabellararea (ala) posterior border furrow posterior border anterior wing anterior lobe of middle body lateral border furrow macula middle furrow posterior lobe of middle body posterior border furrow median notch posterior border lobe articulating half ring articulating and pleural furrow anterior band of pleura articulating facet posterior band of pleura axial furrow axial articulating process axial ring pleural furrow pleural rib lateral border lateral border furrow V hypostome pleural lobe posterior axial swelling terminal axial piece postaxial ridge/ field terminal pygidial spine Figure 7. Morphological terms for the homalonotid exo skeleton. 14 Andrew C. Sandford Genera and subgenera included. Homalonotus Konig, 1825, Trimerus ( Trimerus ) Green, 1832, Trimerus ( Ramiotis ) subgen. nov., Trimerus ( Edgillia ) subgen. nov., Digonus Gtirich, 1909, Dipleura Green, 1832, Brongniartella Reed, 1918, Burmeisteria Salter, 1865, Burmeisterella Reed, 1918, Scabrella Wenndorf, 1990, Arduenella Wenndorf, 1990, Parahomalonotus Reed, 1918, Platycoryphe Foerste, 1919, Wenndorfia gen. nov. Discussion. Classifications of the Homalonotinae by Reed (1918), Sdzuy (1959), Tomczykowa (1975), Thomas (1977) and Wenndorf (1990) have variously emphasised particular features of the exoskeleton including glabellar lobation, glabel- lar outline, the course of the cephalic suture, the length of the preglabellar field, the morphology of the anterior cephalic mar- gin, the shape of the rostral plate, the expression of a rostral process, eye position, pygidial outline, the outline of the pygidial axis, and the expression and degree of pygidial seg- mentation as characters of generic significance. Differences between these classifications reflect different emphases on diagnostic characters. The diversity of morphologies expressed within genera is greatly enhanced by the Australian fauna and necessitates a review of current generic concepts. A new genus Wenndorfia and two new subgenera Trimerus {Ramiotis) and T. ( Edgillia ) are described, and revised diagnoses are given for Trimerus, Homalonotus, Dipleura, Digonus and Parahomalonotus. Homalonotus Konig, 1825 Homalonotus Konig, 1825:104. Koenigia Salter, 1865: 119. Type species. Homalonotus knightii Konig, 1825 from the Ludlow of England, by monotypy. The type species has been recorded from Britain, Germany, Poland and Canada, although its the relationship to the Canadian H. dawsoni is not clearly established. Hall (1860) described dawsoni solely from pygidia, and his diagnosis clearly allies the species with knightii and the Swedish H. rhinotropis. McLeam (1924) suggested that differences in pygidial morphology easily dis- tinguish dawsoni from knightii, citing the posterior projection of the pygidial axis, narrower proportions and less convex profile of the lat- eral margin as distinguishing the type species. On these criteria McLeam identified specimens from the McAdam and Moydart Formations at Arasaig, Nova Scotia as knightii, and those from the slightly younger Stonehouse Formation in the same area as dawsoni. The cephalic morphology of dawsoni was described by Dawson (1868, 1877) and McLeam. McLearn noted that the two cephala known from the McAdam and Moydart Formations were very similar to those of dawsoni from the Stonehouse Formation. Indeed, a relatively unde- formed cephalon from the Moydart Formation examined by the author is clearly attributable to dawsoni rather than knightii. The specimen is comparable to the Stonehouse Formation cephala in the anterior place- ment of the eyes and more elongate and more-weakly tapering glabel- lar shape, distinguishing the Canadian species from the type species. The specimen also shows the acutely pointed lateral cusps of the ante- rior margin, not previously documented. Whether the differences in pygidial morphology between the McAdam and Moydart Formation populations are significant is a question that cannot be resolved with- out detailed re-examination of the faunas. Other species included. Homalonotus dawsoni Hall, 1860, H. rhinotropis Angelin, 1852, H. williamsi sp. nov., H. talenti sp. nov. Range. Ludlow, possibly early Pndoli . England, Germany, Poland, Sweden, eastern North America and south-eastern Australia. Revised diagnosis. Cephalon much wider than long. Glabella long (length 1.1- 1.3 times width), tapering forward weakly to moderately (15-30°), sides straight, lobation weak to indistinct. Paraglabellar area distinct. Preglabellar furrow present, of vari- able depth (shallow to very deeply impressed). Preglabellar field very short (<0.08 times cranidial length). Anterior margin of cephalon tricuspid with a strongly folded (M- shaped) profile in dorsal view. Central cusp triangular, and strongly convex downwards in anterior view. Facial suture and rostral suture meeting at invagination of anterior margin such that connective suture is absent dorsally. Eyes forwardly placed, opposite 0.55-0.7 glabellar length. Pygidial lateral border furrow distinct, anteriorly lateral border swollen, lip-like, fused ven- trally with rolled doublure and continuous with long (exsag.) articulating facet. Discussion. Sdzuy’s (1959) diagnosis of Homalonotus reiter- ates the characters listed by Reed (1918), omitting only the ‘angular course of the anterior branch of the facial suture’. Additional characters listed by Sdzuy include features of the glabella (indistinct lobation, trapezoidal shape), the rostral plate (lacking a process), the folding of the anterior margin, and the presence of a cephalic border (i.e. that defined by the preg- labellar furrow). The preglabellar furrow is variable in depth, being very deeply impressed in H. williamsi and H. talenti, moderately impressed in H. dawsoni (see McLearn, 1924: pi. 27, fig. 14) and H. rhinotropis (see Angelin, 1878: pi. 20, fig. 1). In the type species, the depth of the preglabellar furrow varies from shallow (e.g. Tomczykowa, 1975: pi. 1, figs 1, 3) to moderately impressed (e.g. Salter, 1865: pi. 12, fig. 2) and sug- gests the presence/absence rather than the depth of the furrow is of significance. The assignment of two new Australian species to the genus indicates a somewhat broader range of morphologies for the genus than represented by the northern hemisphere species, particularly in features of the pygidium. Homalonotus knightii, H. rhinotropis and H. dawsoni share a distinct pygidial mor- phology characterised by raised, long pleural ribs that are con- tinuous with the axial rings (separated by a deep, continuous, pleural and ring furrows), effaced axial furrows, by a fused postaxial ridge and posterior area of pleural field, and by a strongly acuminate tip. In contrast, the pygidial axial furrows of H. williamsi and H. talenti are only moderately impressed (Figs 10.3-10.11). In williamsi the axis is raised, defining distinct pygidial trilobation, the postaxial ridge is not fused with the posterior area of the pleural field, whilst in talenti the pleural and ring furrows are shallow and the posterior tip is rounded. The revised diagnosis given above accommodates these mor- phologies by excluding all pygidial characters listed in Sdzuy’s (1959) and Tomczykowa’s (1975) diagnoses, including the ‘acuminate triangular and acutely tipped pygidial outline’, the ‘weakly defined trilobation’, the ‘posterior fusion of axis and pleural field’ and the ‘distinct segmentation’. Nevertheless, the pygidial morphology characterising the contemporary northern- hemisphere species defines a species group distinct from the Australian taxa and reflects a degree of provincialism not otherwise seen in homalonotid distribution. The distinctive morphology of the pygidial lateral border is Homalonotid trilobites from the Silurian and Lower Devonian listed as an additional diagnostic character. The complex anter- ior cephalic margin of Homalonotus corresponds to a complex coaptive morphology. Laterally, the swollen, lip-like section of the pygidial border-doublure defines the overlap of the libri- gena, the pygidial border-doublure fitting against the cephalic doublure. The pygidial doublure is narrow posterior to the lip- like section, and at this point the doublure of the enrolled pygidium emerged from underneath the anterior cephalic border at the invagination of the anterior margin. The anterior margin of the rostral process fitted against the posterior margin of the pygidial doublure, as indicated by the matching convex- ity of these structures (in anterior/posterior view) in H. talenti (Figs 9.5b, 10.7c). Several of the cephalic features listed as generic characters by Sdzuy (1959) are excluded from the revised diagnosis, including the ‘indistinct glabellar lobation’ and the ‘median point on the rostral suture’. These characters are excluded to accommodate the Australian species, that lack a median point on the rostral suture, whilst Homalonotus williamsi exhibits weak glabellar lobation. Sdzuy noted the difficulty in interpret- ing the anterior border of Homalonotus. However, the junction of the facial suture and rostral suture at the invagination of the anterior margin (and the absence of the connective suture dorsally) is clearly evident on williamsi, and is comparable to the morphology of the anterior margin of H. rhinotropis described by Moberg and Gronwall (1909). Sdzuy (1957), Tomczykowa (1975) and Thomas (1977) considered Homalonotus to have been derived from Trimerus and interpreted H. ( T .) johannis Salter, 1865 as transitional between the genera. Wenndorf (1990) expressed uncertainty on this derivation of Homalonotus. Reed (1918) tentatively placed johannis in Homalonotus, primarily on account of its tricuspid, folded anterior margin and supposed pygidial similarities, but noted that long preglabellar field of johannis differed markedly from that of H. knightii. Although Prantl and Pribyl (1948) suggested a new genus or subgenus be erected for johannis, Sdzuy (1957) retained the species in Homalonotus. The length of the preglabellar field is only one of many differences between knightii and johannis indicating that these species are not related. Whilst johannis shares with the northern hemi- sphere species of Homalonotus ( knightii , H. dawsoni and H. rhinotropis ) deeply impressed pygidial pleural and ring fur- rows and an elongate, acutely tipped triangular outline, it lacks the effaced trilobation and fused postaxial area characterising those species. More importantly, johannis lacks the swollen lip-like pygidial border shared also by the Australian species and considered diagnostic of the genus. The species also differs markedly from Homalonotus in having a highly derived cranidial morphology. The raised glabella which is markedly expanded across LI -LI (tr.) and has deep SI apodemes and a distinct sagittal ridge, the strong expression of the medial indentation of the anterior glabellar margin and paraglabellar areas, and the long and weakly convex (tr.) preglabellar field indicates assignment of Salter’s species to Trimerus (Trimerus), in agreement with Tomczykowa’s (1975) and Morris’ (1988) assignment. Of the species assigned in this work to Homalonotus, H. williamsi appears to be the least derived. H. williamsi retains 15 Figure 8. Geological sketch map of the Heathcote area showing Wenlock-Lochkovian fossil localities yielding homalonotids. For other fossil localities see also Thomas (1940a, 1940b, 1941, 1956), Talent (1964, fig. 1), Sandford (2002, fig. 1A; 2005, fig. 4). 16 Andrew C. Sandford distinct glabellar lobation, the axial furrows are moderately impressed, the pygidial axis is raised posteriorly and the postaxial area is not fused with the pleural field. In these respects williamsi shares features of Trimerus possibly reflect- ing the relationship suggested by Sdzuy, Tomczykowa and Thomas. The interpretation of williamsi as morphologically primitive supports the derivation of Homalonotus from T. ( Ramiotis ) rather than from the more derived T. (Trimerus). Of species assigned to T. (Ramiotis), williamsi bears closest resemblance to T. (R.) otisi (Fig. 20), which bears a well-devel- oped tricuspid cephalic margin comparable to that of T. (T.) johannis, though more weakly folded. As with several other Upper Silurian species of T. (Ramiotis), otisi also a exhibits a well defined pygidial border furrow and swollen lip-like border comparable to that of Homalonotus, a morphology otherwise only poorly developed or absent in other groups including T. (Trimerus) and T. (Edgillia). T. (R.) otisi lacks the strongly derived glabellar features of johannis, sharing with Homa- lonotus the elongated, weakly tapering glabella with weak lobation. In these respects otisi is a more likely candidate than johannis in representing an intermediate morphology between Trimerus and Homalonotus, although its interpretation as an ancestral species is not in accord with its stratigraphic position, otisi appearing in strata immediately overlying those yielding williamsi. Homalonotus talenti sp. nov. Figures 9, 10.5-10.11 Homalonotinae gen. et sp. indet. 2. — Holloway and Neil, 1982: 146, fig. 4I-L. — Morzadec, 1986: 186. Type material. Holotype NMY P304936 (cephalon) from PL6650, Heathcote, Victoria (Fig. 9.1). Paratypes NMV P304927, NMV P304937-P304939 (cephala), NMV P304923-P304925, NMV P304928, P304929 (cranidia), NMV P304941, NMV P304943 (libri- gena), NMV P304946-P304948 (hypostomes), NMV P304945, NMV P304949 (thoracic segments), NMV P304917, P304918, NMV P304920-P304922 (pygidia) from PL6650. Paratypes NMV P304952, P304953 (pygidia) from PL6651, Heathcote. For localities see Fig. 8. Previously figured material. NMV P78296 (ex GSV35724, cephalon, figured Holloway and Neil, 1982: figs 41, 4J), NMV P78297 (ex GSV35725, pygidium, figured Holloway and Neil, 1982: figs 4K, 4L) from PL2239, Thomas locality F25, Parish of Heathcote, Heathcote. For locality see Fig. 8. Registered material. 127 specimens: 10 cephala, 41 cranidia, 9 librige- nae, 3 hypostomes, 8 thoracic segments, 56 pygidia. NMV P3049 1 7-P30495 1 , NMV P304959-305026 from PL6650. NMV P304952-P304957, NMV P305027-P305029 from PL6651. NMV P304958, NMV P305041 from PL2265, Thomas locality F46, Parish of Heathcote, Heathcote. NMV P78296, P78297 from PL2239. NMV P305030-P305040 from PL2207, Thomas locality F7, Parish of Dargile, Heathcote. For localities see Fig. 8. Stratigraphic distribution. Upper beds of the Mclvor Sandstone and lowermost beds of the Cornelia Member of the Mt Ida Formation, Notoparmella plentiensis Assemblage Zone, late Ludlow. The age of the Mclvor Sandstone is constrained by early Ludlow (upper nilssoni Biozone) graptolite assemblages from the underlying Melbourne Formation (Rickards and Sandford, 1998). In Europe, Homalonotus occurs in abundance in strata of Ludfordian age, occurring rarely in the Gorstian (Thomas et al., 1989) and possibly ranging into the Pr idoli. The first appearance of H. talenti about 380 m above the base of the Mclvor Sandstone (at PL2265) is considered here to approximate the Gorstian-Ludfordian boundary. The Ludlow-Pridoli boundary can be placed within the interval between the last appearance of talenti (at PL2239) and the first appearance of the post-Ludlow brachiopod Cyrtina at a slightly higher horizon, 50-100 m above the base of Cornelia Member. Derivation of name. For John A. Talent (Macquarie University), for his contribution to Victorian palaeontology. Diagnosis. Glabella trapezoid, sides straight and converging at 20°, anterior margin broadly rounded. Glabellar length about 1.1 times preoccipital glabellar width and 0.94 times cranidial length. Preglabellar furrow very deeply impressed. Eye placed with midline of palpebral lobe opposite 0.56 glabellar length/0.52 cranidial length. Dorsal surface of rostral plate 0.14 times cephalic length. Hypostome with length 0.85 width, lobes on the posterior border with length 0.17 times hypostomal length. Pygidium with length about 0.9 times width, postaxial ridge projecting posteriorly, tip rounded. Pygidial axis with width 0.5 times pygidial width, 11 axial rings. Axial furrows straight and tapering at about 30°. 7 pleural ribs, rib-ring medially offset at fourth rib. Ring, pleural and axial furrows shallow to moderately impressed. Description. Exoskeleton of moderate size, maximum length estimat- ed 12 cm (from NMV P305923), occipital convexity (tr.) moderate, pygidial convexity (tr.) strong. Cephalon wide, length 0.6 times width, with trapezoid outline, sides moderately convex and converging forwards at about 75°, anterior margin tricuspate, median cusp triangular in outline with rounded tip, length about 0.6 times width, lateral cusps with obtusely angled tips reaching forward to midlength of median cusp. Cranidial width 1.9 Figure 9. Homalonotus talenti sp. nov. la, holotype NMV P304936, cephalon, dorsal view X 3.0 (latex cast) from PL6650. lb, same, x 2.6 (inter- nal mould), lc, same, lateral view x 3.0 (latex cast). 2, paratype NMV P304938, cephalon, dorsal view x 2.6 (latex cast) from PL6650. 3, paratype NMV P304929, cranidium, dorsal view x 2.1 (internal mould) from PL6650. 4, paratype NMV P304923, cranidium, dorsal view x 1.65 (internal mould) from PL6650. 5a, paratype NMV P304927, cephalon, dorsal view x 5.5 (internal mould) from PL6650. 5b, same, anterior view x 5.0. 6, paratype NMV P304937, cephalon, dorsal view x 1 .9 (latex cast) from PL6650. 7, paratype NMV P304924, cranidium, dorsal view x 3.6 (inter- nal mould) from PL6650. 8a, paratype NMV P304939, cephalon, dorsal view x 2.0 (internal mould) from PL6650. 8b, same, x 2.1 (latex cast). 9, paratype NMV P304943, librigena, dorsolateral view x 1.22 (latex cast) from PL6650. 10, paratype NMV P304925, cranidium, dorsal view x 1.75 (internal mould) lfom PL6650. 11, paratype NMV P304928, cranidium, dorsal view x 1.6 (internal mould) from PL6650. 12, paratype NMV P304949, thoracic segment, lateral view x 1.75 (internal mould) from PL6650. 13, paratype NMV P304945, thoracic segment, dorsal view x 1.4 (internal mould) from PL6650. 14, paratype NMV P304941, librigenal doublure, ventral view x 2.4 (latex cast) from PL6650. 15, paratype NMV P304947, hypostome, ventral view x 2.2 (internal mould) from PL6650. 16a, paratype NMV P304946, hypostome, ventral view x 3.4 (latex cast) from PL6650. 16b, same (internal mould). 17a, paratype NMV P304948, hypostome, ventral view x 3.5 (internal mould) from PL6650. 17b, same (latex cast). Homalonotid trilobites from the Silurian and Lower Devonian 17 mm 18 Andrew C. Sandford Homalonotid trilobites from the Silurian and Lower Devonian 19 times length. Glabella with anterior margin strongly defined, arc of curvature centred at about glabellar midlength. Occipital ring about 0.1 times cranidial length, slightly wider medially. Occipital furrow mod- erately impressed to shallow, with weak forward flexure medially. Glabellar lobation extremely weak (best seen on NMV P304928) to indistinct. LI 0.32 times glabellar length, L2 0.12 times glabellar length, L3 0.10 times glabellar length and frontal lobe 0.20 times glabellar length. SI weakly convex, directed diagonally abaxially. S2 and S3 transverse. Axial furrows shallow to moderately impressed. Paraglabellar area very weakly defined. Length (sag.) of preglabellar furrow and ridge-like preglabellar field 0.03 times cranidial length. Length (exsag.) of posterior border equal to occipital length adaxially, lengthening slightly abaxially. Posterior border furrow transverse, very wide and shallow, meeting lateral border furrow distally. Postocular fixigenal area long, length (exsag.) 0.25 times cranidial length. Palpebral lobes placed anteriorly and remotely (6-6 1.5 times preoc- cipital glabellar width). Palpebral lobe length (exsag.) 0.15 times cranidial length, palpebral furrow indistinct. Preocular fixigenal area of moderate width, width 0.17 5-6 adjacent to palpebral area, eye ridges not distinct, narrowing markedly anteriorly. Anterior branches of facial suture converging at about 80°, curving inwards opposite the antero-lateral corner of the glabella and converging at about 140° anteriorly. Librigena with wide, moderately impressed border furrow, not continuing forward of antero-lateral comer of glabella, librigenal field weakly convex, steeply inclined, lateral border wide, convex. Anter- ior margin of librigena projecting far forwards from juncture of rostral, connective and facial sutures at cephalic margin as triangular cusp. Course of rostral suture broadly rounded, concentric to anterior margin of glabella. Dorsal surface of rostral plate rounded triangular, length (sag.) 0.4 times width, strongly concave (tr. sect.). On cephalic doublure connective sutures straight and weakly converging posterior- ly. Librigenal doublure without distinct vincular furrow, strongly con- vex (exsag. sect.) adjacent to connective suture (i.e. accommodating preglabellar furrow). Hypostomal suture very broadly rounded. Hypostome with middle furrow shallow to moderately impressed, anterior wing process small (width 0. 1 times hypostomal width), lobes on posterior border parabolic in outline, deep medial notch with sides converging at about 75°. Thorax with axial furrows poorly defined. Pleural furrows narrow (exsag.) and deep, pleural tips rounded. Pygidium with triangular outline, sides weakly convex and con- verging at about 80°. In lateral view pygidium high, height equal to length. Pygidial axis reaching to about 0.8 times pygidial length, con- tinuous posteriorly with wide postaxial ridge. Postaxial ridge raised, parallel-sided, wide, width (tr.) 0.45 times axial width, posterior mar- gin with parabolic outline. Axial furrows tapering, curving to the exsagittal posteriorly, shallow anteriorly, moderately impressed poster- iorly. Pleural furrows shallowing markedly adjacent to border furrow. Border furrow moderately impressed opposite pleural field, shallow opposite postaxial ridge. In dorsal view border narrow (tr.) but pro- tuberant from margin of pleural field. In lateral view border very wide anteriorly, merging with wide articulating facet, narrowing markedly posteriorly. Lateral border continuous with narrow pygidial doublure. In posterior view anterior margin of pygidium strongly convex, poster- ior margin with distinct medial and lateral arches. In ventral view, inner margin of doublure parabolic in outline with deep medial inden- tation. Dorsal exoskeleton finely granulose. Pygidial border with fine ridging. Discussion. Holloway and Neil (1982) suggested the affinities of an incomplete cephalon examined by them were with Homalonotus. Their suggestion is confirmed with the additional specimens listed. They noted similarities between the associated pygidium and that of the South American Lower Devonian species H. clarkei, but these similar- ities are superficial. Following Cooper (1982), and as discussed below, clarkei is assigned to Burmeisteria. Hypostomes are also known for Homalonotus knightii and H. rhinotropis (Salter, 1865: pi. 12 fig. 10; Angelin, 1878: pi. 20 fig. lc). The hypostome of H. talenti most closely resembles that of the type species. The hypostome of rhinotropis is relatively elongate (length times 1.1 width) compared to that of knightii (length 0.81 times width) and talenti (length 0.85 times width). The projections of the posterior margin are broadly based and long in talenti and knightii, but short in rhinotropis. Although the ventral surface of the rostral plate of Homalonotus tal- enti is not known, the course of the connective sutures (indicated by the shape of the librigenae) is more or less straight and slightly con- vergent posteriorly, indicating the rostral plate to be approximately pentagonal, with subequal sides. Apparent curvature of the connective sutures on Fig. 9.14 is due to the oblique orientation of the specimen. The broad, rounded medial indentation of the inner margin of the pygidial doublure of Homalonotus talenti (Fig. 10.7e) differs from that of H. knightii, in which the margin is more or less straight laterally, and defines an acute angle medially (see Salter, 1865: pi. 12 fig. 9a, Tomczykowa, 1975: pi. 1 fig. 5c). Environmental notes. Throughout its range Homalonotus talenti occurs in thick-bedded fine- to medium-grained sandstones, in low diversity or monospecific assemblages. At the type locality where the species occurs in greatest abundance the proportion of articulated specimens is low (6%) and the proportion of broken specimens low (3%).The fauna can be assigned to taphofacies Til and indicates shallow environments around normal wave base. Homalonotus williamsi sp. nov. Figures 10.1-10.4 Type material. Holotype NMV P308674 (cephalon) from PL6615, Jutson locality VII, Eden Park, Victoria (Fig. 10.1). Paratypes NMV P308675 (pygidium), NMV P308676 (cephalon) from PL6615. Paratype NMV P304511 (pygidium) from PL6614, Jutson locality VI, Eden Park. For localities see Fig. 1 1 . Figure 10.1-10.4 Homalonotus williamsi sp. nov. la, holotype NMV P308674, cephalon, dorsal view x 2.1 (internal mould) from PL6615. lb, same, anterior view, lc, same, oblique view x 2.3 (latex cast). Id, same, dorsal view. 2a, paratype NMV P308676, cranidium, dorsal view x 1.3 (internal mould) from PL6615. 2b, same, lateral view x 1.2. 2c, same, anterior view x 1.3. 3, paratype NMV P308675, pygidium, oblique view x 3.0 (latex cast) from PL6615. 4a, paratype NMV P304511, pygidium, oblique view x 1.35 (internal mould) from PL6614. 4b, same, dorsal view x 1.45. Figure 10.5-10.11, Homalonotus talenti sp. nov. 5a, paratype NMV P304952, pygidium, dorsal view x 1.1 (latex cast) from PL6651. 5b, same, posterodorsal view. 6, paratype NMV P304953, pygidium, oblique view x 1.2 (latex cast) from PL6651. 7a, paratype NMV P304921, pygidium, lateral view X 1.6 (internal mould) from PL6650. 7b, same, dorsal view. 7c, same, posteroventral view. 7d, same, posterior view. 7e, same, ventral view (doublure). 8a, paratype NMV P304917, pygidium, posteroventral view x 1.2 (internal mould) from PL6650. 8b, same, dorsal view x 1.6. 8c, same, lateral view x 1.4. 9, paratype NMV P304918, pygidium, dorsal view x 2.6 (latex cast) from PL6650. 10, paratype NMV P304920, pygidium, lateral view x 2.7 (internal mould) from PL6650. 11, paratype P304922, pygidium, dorsal view x 1.5 (internal mould) from PL6650. 20 Andrew C. Sandford Figure 11. Geological sketch map of the Springfield- Kinglake West area showing Llandovery -Lochkovian fossil localities yielding homalonotids. For other fossil localities see also Jutson (1908, pi. 3), Thomas (1960), Talent (1964, fig. 1), Williams (1964, fig. 2), Garratt (1972, 1977), Rickards and Sandford (1998, fig. 6), Sandford (2002, fig. 1C; 2005, figs 2-3). Other material. NMV P304512 from PL6614. Stratigraphic distribution. Macropleura band, Eden Park Formation (130-140 metres above base of unit), lowermost Notoparmella plentiensis Assemblage Zone, early-mid Ludlow. Derivation of name. For George E. Williams, for his contri- bution to Victorian stratigraphy. Diagnosis. Cephalon wide, length 0.6 times width. Glabella long, length 1.15 times width, sides more or less straight, taper- ing at about 15°, anterior margin broadly rounded, arc centred at 0.55 glabellar length. Glabellar length 0.92 times cranidial length. Glabellar lobation distinct, SI and S2 expressed as deep notches adjacent to the axial furrows but very shallow abax- ially, placed opposite 0.43 and 0.63 glabellar length respec- tively. Axial furrows moderately impressed on internal moulds (shallow on external moulds), preglabellar furrow very deeply impressed. Rostral plate with length of dorsal surface 0.11 times cephalic length, width (across rostral suture) 1.9 times Homalonotid trilobites from the Silurian and Lower Devonian length, triangular. Genae moderately swollen, lateral border furrow moderately impressed. Eye placed with midline of palpebral lobe opposite 0.67 glabellar length. Anterior branch of facial suture with posterior section straight and strongly con- vergent (at about 90°), anterior section evenly curving to the transverse. Rostral suture weakly convex. Pygidium triangular in outline, sides straight, converging posteriorly at about 100°. Axial furrows moderately impressed. Pygidial axis strongly convex, raised, sides converging at about 25°, continuous pos- teriorly with wide postaxial ridge. Pleural field with 7 ribs. Pleural and ring furrows moderately impressed, of subequal depth. Discussion. Homalonotus williamsi closely resembles H. talenti from the lower Ludlow-lower Pndolf strata of the Heathcote area. Differences in cephalic morphology are subtle, but talenti can be distinguished in that the glabella is less elon- gate (glabellar length 1.1 times width) SI -S3 are extremely weakly impressed to indistinct, the axial furrows are much shallower, the preglabellar furrow and preglabellar field are notably shorter (exsag. and sag., glabellar length 0.95 times cranidial length), the eyes are more posteriorly placed (midline of palpebral lobe opposite 0.55 glabellar length), and the rostral process is wider posteriorly and only weakly convex down- wards. The pygidium of williamsi is known only from frag- ments, but differs from that of talenti in having deeper axial, ring and pleural furrows. In addition to the pygidial characters and the very deep, trench-like preglabellar furrows distinguishing the two Australian species from the northern-hemisphere species, Homalonotus talenti and H. williamsi also share straight and strongly convergent anterior branches of the facial sutures, in contrast to the convex outwards course of the sutures in H. knightii, H. dawsoni and H. rhinotropis. The Australian species and rhinotropis share a rounded anterior glabellar margin, differing from the transverse margin of knightii and dawsoni. Environmental notes. See Dipleura garratti sp. nov. Brongniartella Reed, 1918 Type species. Homalonotus bisulcatus Salter, 1851 from the Upper Ordovician of England, by original designation. Brongniartella sp. Trimerus sp. — Holloway and Sandford, 1993: 93, fig. 4L, non figs 4C-D, 4F-G, 41- J, 4N-P. Material. NMV PI 37228 (pygidium, figured Holloway and Sandford, 1993: fig. 4L) from PL359, Tiger Range, northwest of Maydena, Tasmania. For locality see Holloway and Sandford: fig. 1 . Stratigraphic distribution. Richea Siltstone, Monograptus griestoniensis-M. crenulata biozones, upper Llandovery. Notes. The elongate, rounded parabolic outline of the pygid- ium, the strongly raised, narrow, convex pygidial axis and the indication of a distinct border suggest assignment to Brongniartella. The concave sides of the axis is shared with other species of Brongniartella. The species is a late appear- ance for Brongniartella, a predominantly Upper Ordovician 21 genus but known to extend into the Lower Llandovery atavus- cyphusl biozones in Britain (Temple, 1975) and also recorded from the Lower Llandovery of Jordan (Wolfart et al., 1968). The Upper Llandovery Tasmanian record narrows the gap between Lower Llandovery Brongniartella and the youngest known species, B. pamiricus (Balashova, 1966) from the Wenlock of Pamir, central Asia. Digonus Giirich, 1909 Type species. Homalonotus gigas Roemer, 1843, from the Lower Devonian (Emsian) Kahleberg Sandstone, Germany, by original designation. Other species and subspecies included. Burmeisteria ( Digonus ) antarcticus Saul, 1965, D. armoricanus Pillet, 1961a, D. asturco Kegel, 1927, Homalonotus crassicauda Sandberger and Sandberger, 1849, D. gigas posterior Wenndorf, 1990, H. goniopygaeus Woodward, 1882, H. (D.) harpyius Richter and Richter, 1932, H. inter- medius Vietor, 1919, H. laticaudatus Williams and Breger, 1916, H. (D.) ornatus disornatus Richter and Richter, 1932, D. ornatus linguatus Wenndorf, 1990, H. ornatus Koch, 1883a, D. comes Chlupac, 1981, H. roemeri de Koninck, 1876, D. wenndorfi sp. nov., Trimerus zeehanensis Gill, 1949, D. zemmourensis Pillet, 1961b, Digonus sp. A in Wenndorf (1990). Other species tentatively included. Digonus collini Renaud, 1942, Homalonotus derbyi Clarke, 1890. Range. Lochkovian-Emsian. Diagnosis. Cranidium trapezoidal in outline, anterior margin with median cusp. Dorsal section of connective sutures very short. Glabella trapezoidal in outline, sides straight or very weakly concave. Glabellar lobation very weak or absent. Paraglabellar area distinct. Anterior outline of cranidium quadrate, anterior branches of facial suture more or less straight between palpebral lobe and midlength of preglabellar field and converging at an acute angle (20-60°), curving abruptly to the transverse anteriorly. Rostral suture transverse to weakly con- cave. Ventral surface of rostral plate with thorn, spine or keel. Pygidium triangular to elongate triangular, with wide axis (0.5-0.7 times pygidial width). Axis with weak independent convexity (tr. sect.), postaxial ridge not raised. Pleural and ring furrows deep to very deep, more or less equal in depth. Axial furrow shallow or moderately impressed posteriorly, very shal- low or effaced anteriorly, anteriormost ribs and rings fused, with corresponding section of axial furrow effaced or reduced to shallow invagination in posterior margin of segment. Pleural furrows maintaining uniform depth to a point close to border. Border furrow and border poorly defined. Rib-ring offset high (fifth-seventh rib). Discussion. Previous diagnoses for Digonus are brief. Giirich (1909) listed the truncate and indented outline of the cephalic anterior margin, the angular margin of the antero-lateral corners of the cranidium, the pointed pygidial tip and the weakly tapered, trapezoid glabellar outline. Reed (1918) emphasised the strong expression of the pygidial segmentation as a further diagnostic character. Sdzuy (1959) included the absence of glabellar lobation and angular pleural tips as diagnostic charac- ters although, as he regarded Digonus as a subgenus of Burmeisteria, other characters were incorporated into the 22 Andrew C. Sandford generic concept. These included the distinct paraglabellar area, the presence of a ventral process on the rostral plate, a triangu- lar pygidial outline, distinct pygidial trilobation, and a poorly expressed postaxial ridge. Tomczykowa (1975) omitted many of Sdzuy’s characters from her revised diagnosis, closely fol- lowing Reed. In emending Tomczykowa’s diagnosis, Wenndorf (1990) noted the presence of weak glabellar lobation in some species, and that early representatives of the genus have less elongate pygidial outlines. The revised diagnosis given here adds substantially to the list of diagnostic characters and qualifies other characters pre- viously used. A substantially restricted concept of the genus is proposed by placing new emphasis on pygidial characters including the relative depth of pygidial ring and pleural furrows, the expression of the axial furrows, the expression of the postaxial ridge and the width and convexity of the pygidial axis. The revised diagnosis emphasises differences between Digonus, Trimerus and Burmeisteria, and results in sub- stantially different assignments of the species listed by Tomczykowa (1975) and Wenndorf (1990). Tomczykowa (1975) and Wenndorf (1990) considered Digonus to have been derived from Trimerus. The genera are certainly close, many of the characters listed in the diagnosis of Digonus occurring variously in species of Trimerus. Although the presence of a rostral keel or process is considered of pri- mary importance in differentiating the genus from Trimerus, it is the combination of characters that defines Digonus. Other cephalic characters, particularly the degree of glabellar loba- tion, the quadrate course of the facial and rostral sutures, and the trapezoid glabellar outline, distinguish Digonus from most species of Trimerus. Pygidia of Digonus can generally be dis- tinguished from those of Trimerus by the deeper pleural and ring furrows that are more or less equal in depth, by the shallower axial furrows, and by the less convex axis and postaxial ridge. For the morphologically convergent species, it is the continuity of the axial furrow anteriorly in Trimerus (and the fusion of the anteriormost axial ring and pleural rib in Digonus) that distinguishes the genera. Hence, Homalonotus crassicauda from the Emsian of Germany, known only from pygidia and previously interpreted as a temporally disjunct representative of Trimerus (Wenndorf, 1990, Schraut, 2000) clearly belongs in Digonus. The Lochkovian North American H. major Whitfield, 1881 is best assigned to Trimerus rather than to Digonus, because the pygidial axial furrow is deep and distinct anteriorly. Digonus wenndorfi and D. zeehanensis from the lower Lochkovian of south-eastern Australia are the best known of the early representatives of Digonus and support the derivation of this genus from Trimerus. D. wenndorfi and the French D. roemeri are the earliest Digonus, occurring immediately above the Silurian-Devonian boundary and suggesting a latest Silurian origin for the genus. In cranidial features, wenndorfi and zeehanensis show close affinities to the D. omatus group from the middle Pragian-Emsian of Europe rather than to the D. gigas group (see Wenndorf, 1990) whose members share longer preglabellar fields. Weak glabellar lobation and short pygidia with markedly obtuse (120-130°) angular tips set wenndorfi and zeehanensis apart from later members of the genus, with the exception of several species (D. antarcticus and D. omatus disornatus ) that retain weak lobation, and several others that have similarly short pygidia (e.g. D. armoricanus and D. crassicauda ). In addition to weak glabellar lobation and short pygidial proportions, wenndorfi also exhibits a moder- ately raised pygidial postaxial ridge (Figs 12.6, 12.12-12.18). In these features wenndorfi can be considered a ‘primitive’ Digonus, with relict features of Trimerus retained. D. zeehan- ensis bears a short (sag.) but wide (tr.) median cephalic cusp (Fig. 13.1) intermediate in morphology between the large semi- circular cusps of Trimerus on species such as T. ( Ramiotis ) otisi and T. ( Trimerus ) johannis, and the small, acute cusps of Emsian Digonus. The anteriorly distinct pygidial axial furrows of roemeri (see Morzadec, 1986: pi. 32 figs 1, 4, 6-7) and (to a lesser extent) wenndorfi, can also be interpreted as a Trimerus- like feature, although the wide pygidial axis and equally deep pleural and ring furrows indicates their affinities with other Digonus. Emphasising the significance of the transverse anterior margin of the cranidium and the depth of the pygidial pleural and ring furrows, Tomczykowa (1975) derived Digonus from Trimerus, suggesting T. ( Trimerus ) johannis as a transitional morphology between the genera. Thomas (1977) doubted this suggestion, emphasising the difference in age between T. (T.) johannis (late Wenlock) and the first appearance of Digonus (earliest Lochkovian). Wenndorf (1990) interpreted European Lochkovian species such as Homalonotus vialai Gosselet in Gosselet et al., 1912 and H. roemeri as representing early Digonus morphologies, noting similarities to the Polish Ludlow T. permutus Tomczykowa, 1978 (=T. lobatus Tomcykowa, 1975 non Prouty, 1923 in Swartz and Prouty). A more plausible transitional morphology is represented by the upper Ludlow T. ( Ramiotis ) otisi. The similarities of this species to contem- porary Homalonotus have been discussed above, but its similarities to species of Digonus are more marked. In addition to having a tricuspid cephalic margin, otisi exhibits a sub- quadrate course of the facial and rostral sutures, weak glabellar lobation, straight- sided glabellar outline, deep pleural and equally deep ring furrows, and a pygidial axial furrow that is indistinct anteriorly (Fig. 20). In the latter feature, otisi is an exception amongst congeners. However, the flat rostral plate, the long glabellar proportions, the narrow proportions of the pygidial axis and the raised postaxial ridge demonstrate its close affinities to Trimerus. Digonus wenndorfi and D. zeehanensis demonstrate the con- servative Digonus cephalic morphology to have been already established in the earliest Lochkovian. European species con- sidered by Tomcykowa (1975) as representing early Digonus morphologies, including Homalonotus vialai, D. bostoviensis and D. elegans, are variably assigned below to Paraho- malonotus and Wenndorfia, and exhibit few of the diagnostic cephalic or pygidial features of Digonus. In particular, the anterior branches of the facial sutures of these species are broadly curved and converge at a strongly obtuse angle opposite the preglabellar field, giving the anterior margin of the cranidium a more rounded/parabolic outline. Burmeisteria ( Digonusl ) delattrei Pillet and Waterlot, 1982 from the Emsian of France exhibits few features typical of Digonus. It has a Homalonotid trilobites from the Silurian and Lower Devonian 23 raised pygidial axis, a raised postaxial ridge with a posteriorly projecting spine, anteriorly distinct axial furrows and a moderately convex (tr.), elongate glabella. The pygidial morphology, including the posteriorly projecting axis, indicates assignment to Burmeisterella. Wenndorf (1990) questionably assigned the Brazilian Homalonotus oiara Hartt and Rathbun, 1875 to Digonus, but their emphasis on the concavity of the sides of the glabella is not compatible with the quadrate glabellar outline considered here as diagnostic of Digonus. The forward eye position exhibited by the single cranidium known suggests affinities with Burmeisteria rather than with Dipleura. Relationships between Digonus and Burmeisteria have been variously interpreted. Sdzuy (1959) considered Digonus to be a subgenus of Burmeisteria. Tomczykowa (1975) listed several differences between the taxa, emphasising the absence of spines and glabellar lobation in Digonus to support its independent status. Cooper (1982) argued that the range of variation in the course of the rostral suture, the degree of glabellar lobation and the glabellar outline in the highly polymorphic type species B. herschelii (Murchison, 1839) overlapped with morphologies typical of Digonus. Following the emphasis placed by Sdzuy on these characters, Cooper regarded Digonus as junior synonym of Burmeisteria. Wenndorf (1990) questioned many of the differences listed by Tomczykowa, noting a convergence in morphologies between Burmeisteria and Digonus, but nevertheless main- taining their independent status. Kennedy (1994) noted that the pygidial doublure of Digonus, unlike that of Burmeis- teria, is often narrower and bears a narrow groove in the margin of the pygidial doublure. This distinction between the genera also applies to Burmeisterella, which Kennedy consid- ered as a junior synomym of Burmeisteria. The independent status of Burmeisterella is supported here, with high generic significance accorded to the very short (sag.) pre- glabellar field and the morphology of the pygidial axis, which is raised high rather than fused with the pleural field posteriorly, and extends posteriorly over the margin as a narrow-based to spinose tip. In addition, the pygidium of Burmeisterella has a more convex lateral margin, and always bears a pair of prominent, large nodes or spines on the anteriormost pleural rib. The independent status of Bur- meisteria and Digonus is supported in this work. The taxa can be readily distinguished morphologically, despite the extreme polymorphism exhibited by herschelii. In Burmeisteria, the pygidial axial furrow is impressed anteriorly, and the pygidial pleural and ring furrows are typically shallower than those of Digonus and shallow posteriorly. Burmeisteria is further characterised by a high number of pygidial segments (>13 axial rings). Cranidia of Burmeisteria and Digonus may exhibit morphological convergence, but a glabellar outline markedly expanded across LI -LI (tr.) or with concave sides, and well defined lateral glabellar furrows immediately distinguish Burmeisteria when present. Following Hennig (1965), Wenndorf suggested that the mutually exclusive palaeogeographic ranges of Burmeisteria and Digonus represented a significant taxonomic feature supporting their independent status. Digonus wenndorfi. D. zeehanensis and D. antarctica extend the range of the genus into temperate palaeolatitudes of north-eastern Gondwana. However, the palaeogeographic ranges of the genera remain mutually exclusive. B. herschelii is restricted to lower palaeo- latitudes of southern Gondwana, occurring in southern Africa, the Falkland Islands and South America. The reassignment of several South American taxa to Burmeisteria emphasises the provincialism of the genus (see Table 1), perhaps as early as the Late Silurian. Homalonotus linares Salter, 1861, assigned below to Burmeisteria (see Trimerus) occurs in the Late Silurian of Bolivia and Uruguay. Affinities of the Lochkovian Homalonotus noticus Clarke, 1913 (?= H. caorsi Mendez-Alzola, 1938) from South America and possibly South Africa (see Cooper, 1982) are with Burmeisteria. Reed (1918) initially assigned the species to Burmeisteria but later reassigned it to Digonus (Reed, 1925). Sdzuy (1957) suggested that the species represents a transi- tional form between Burmeisteria and Digonus. The distinct S1-S3, the narrow pygidial axis and the anteriorly continuous pygidial axial furrow are incompatible with assignment to Digonus. Tomczykowa (1975) and Wenndorf (1990) assigned H. noticus to Trimerus, but the presence of a rostral projection, a sinusoidal rostral suture and high pygidial segmentation (up to 13 axial rings) excludes the species from the genus as defined here. Homalonotus clarkei from the Lower Devonian South America (?= H. buqueti Mendez-Alzola, 1938, ?= H. spat- ulirostris Mendez-Alzola, 1938), has been variably assigned to Digonus (Wolfart, 1968), to Trimerus (with question, Tomczykowa, 1975), to Burmeisteria (Cooper, 1982) and to Dipleura (Wenndorf, 1990). The strongly folded anterior cephalic margin excludes assignment to Trimerus, as this taxon is characterised in this work by having a flat rostral plate. The very shallow pygidial pleural furrows and somewhat deeper ring furrows of clarkei are typical of Dipleura, but the tricuspid anterior cephalic margin, the quadrate and medially concave anterior cranidial margin, and the strong forward tapering of the glabella do not conform to this assignment. The weakly expressed pygidial segmentation excludes the species from Digonus as defined here. The generic affinities of the species are uncertain. Burmeisteria ( Digonus ) accraensis from the Devonian of Ghana exhibits weak pygidial furrows, excluding the species from Digonus as defined in this work. The taxon has been vari- ably assigned to Trimerus (Tomczykowa, 1975) and to Dipleura (with question, Wenndorf, 1990), but the presence of a rostral process (see Saul, 1967: pi. 143 fig. 6) excludes the species from these genera.. The species is assigned here to Burmeisteria. Digonus wenndorfi sp. nov. Figs 3, 12, 22.1-22.6 Trimerus ( Dipleural ) sp. — Talent, 1964: 49 (pars), pi. 26 figs 1, 2 non text-fig. 6. Homalonotidae gen. et sp. indet. 1. — Holloway and Neil, 1982: 145, figs 4A-H. Type material. Holotype NMV P304717 (pygidium) from PL2203, Thomas locality F3, Parish of Dargile, Heathcote, Victoria (Fig. 24 Andrew C. Sandford Homalonotid trilobites from the Silurian and Lower Devonian 25 12.17). Paratypes NMV P304709-P304716 (cranidia), NMV P304874 (librigena), NMV P304872, P304872 (thoracic segments), NMV P304717, NMV P304860-P304866, NMV P304868 (pygidia) from PL2203. Paratype NMV P304913 (librigena) from PL2327, Heathcote. For localities see Fig. 8. Previously figured material. NMV P78295 (ex GSV 39474, pygidium, figured Holloway and Neil, 1982: figs 4E, 4H), NMV P59656 (ex GSV 39200 pygidium, figured Talent, 1964: pi. 26 fig. 2, Holloway and Neil, 1982: figs 4F, 4G), NMV P59655 (ex GSV 39476, pygidium, fig- ured Talent, 1964: pi. 26 fig. 1) from PL2203. NMV P78293 (thoracic segment, figured Holloway and Neil, 1982: fig. 4C), NMV P78294 (thoracic segment, figured Holloway and Neil, 1982: fig. 4D) from PL2202, Thomas locality F2, Parish of Dargile, Heathcote. NMV P78292 (cephalon, figured Holloway and Neil, 1982: figs 4A, 4B) from PL2204, Thomas locality F4, Parish of Dargile, Heathcote. For localities see Fig. 8. Registered material. 88 specimens: 2 cephala, 12 cranidia, 4 librigenae, 17 thoracic segments, 53 pygidia. NMV P59655, P59656, NMV P78295, NMV P82882, NMV P82886, NMV P304709-P304717, NMV P304859-P304903 from PL2203. NMV P78293, P78294, NMV P82880, NMV P82883-P82885, NMV P304904 from PL2202. NMV P82941, NMV P304905-P3049 14 from PL2327. NMV P78292 from PL2204. Stratigraphic distribution. Mt Ida Formation, from about 400 m above the base of the Dealba Member up to about the 100 m above the base of the Stoddart Member, Boucotia janae Assemblage Zone, early Lochkovian. The first appearance of Digonus wenndorfi at PL2204 is close stratigraphically to the first appearance of Boucotia (see Garratt, 1983: fig. 7) and together support a basal Lochkovian age for these horizons. Derivation of name. For Dr Klaus-Werner Wenndorf (Germany), for his contribution to the study of homalonotids. Diagnosis. Glabella trapezoid, length 1.0- 1.2 times width, sides straight and converging at about 20° opposite genae, anterior margin of glabella well defined with medial indentation moderately impressed to indistinct. Glabellar lobation weak to indistinct. Axial furrows moderately impressed opposite genae. Palpebral lobe short, 0.1 times cranidial length, placed with midline opposite 0.5 glabellar length/0.39 cranidial length. Preglabellar field long, length (sag.) 0.22 times cranidial length. Rostral suture transverse. Pygidium triangular, sides moderately convex and converging posteriorly at 85-100°, tip angular. Pygidial axis with width 0.5-0.6 times pygidial width, 11-12 axial rings, ring furrows deep, continuous with raised postaxial ridge. Axial furrows poorly defined opposite first-second axial rings, moderately impressed posteriorly, straight and tapering at 30-40°. Pleural furrows deep. 8-9 pleural ribs, rib-ring medially offset at fifth rib. Description. Exoskeleton large (maximum length estimated 25 cm from NMV P304860), occipital convexity (tr.) moderate, pygidial convexity (tr.) strong. Dorsal exoskeleton finely granulose. Cranidia with elongate and short forms. Elongate form with cranidial width about 1.6 times length, short form with cranidial width about 1.9 times length. Glabella with length 0.78 times cran- idial length, anterior margin very broadly rounded to transverse. Elongate form with glabellar length 1.2 times width, short form with glabellar length equal to width. Occipital ring with length 0.1 times glabellar length. Occipital furrow deeply impressed, with weak for- ward flexure medially. Glabellar lobation (best seen on NMV P304715 and NMV P304710) with LI 0.28 and L2 0.15 times glabellar length respectively. SI straight and directed posteromedially at about 17° from the transverse, S2 weakly convex forwards and transverse, S3 indistinct. In some specimens (e.g. NMV P304710, P304711) SI mod- erately impressed for short distance adjacent to axial furrow. Axial furrows very shallow and directed diagonally opposite occipital ring. Paraglabellar area weakly defined (best seen on NMV P304711). Length (exsag.) of posterior border equal to occipital length adaxially, lengthening slightly abax-ially. Posterior border furrow transverse, very wide, moderately impressed, terminating distally. Postocular fix- igenal area very short, length (exsag.) 0.16 times cranidial length. Palpebral lobes placed remotely (5-6 1.67 times preoccipital glabellar width/0.7 times cranidial width). Palpebral furrow weak to moderately impressed. Preocular fixigenal area of moderate width, 0.17 times 5-6 narrowing slightly anteriorly. Preglabellar field weakly concave (tr. sect.). Anterior branches of facial suture angular, directed exsagittaly for a short distance adjacent to the eye (to a point opposite 0.7 glabel- lar length), anteriorly converging at about 40° in elongate forms and 60° in short forms. Librigena without distinct border furrow or lateral border. In anterior view anterior margin of librigenae moderately convex. Thorax with axial furrows indistinct. Pleural furrows wide and deep. Pygidia with long form and narrow form, sides converging poster- iorly at about 85° in narrow form and 100° in wide form. Pygidial axis with width about 0.5 times pygidial width in wide form, 0.6 in narrow form. Pygidial axis reaching to about 0.85 times pygidial length, raised, continuous posteriorly with postaxial ridge. Postaxial ridge wide, width 0.3 times axial width. Axial furrows straight and tapering at about 33° in narrow form, 38° in wide form, curving to the exsagit- tal posteriorly. Pleural furrows not reaching margin. Border furrow and border not defined. In posterior view posterior margin of pygidium Figure 12. Digonus wenndorfi sp. nov. 1, paratype NMV P304715, cranidium, dorsal view x 1.25 (internal mould) from PL2203. 2a, paratype NMV P304714, cranidium, dorsal view x 1.6 (internal mould) from PL2203. 2b, same (latex cast). 3a, paratype NMV P304709, cranidium, dorsal view x 0.75 (internal mould) from PL2203. 3b, same, lateral view. 3c, same, dorsal view (latex cast). 4, paratype NMV P304716, cran- idium, dorsal view x 1.2 (internal mould) from PL2203. 5a, paratype NMV P30471 1, cranidium, dorsal view x 1.75 (internal mould) from PL2203. 5b, same (latex cast). 5c, same, oblique view (internal mould). 6, paratype NMV P304868, pygidium, dorsal view x 1.0 (internal mould) from PL2203. 7a, paratype NMV P304872, thoracic segment, lateral view x 2.3 (internal mould) from PL2203. 7b, same, dorsal view. 8, paratype NMV P304874, librigena, ventral view (doublure) x 1.75 (latex cast) from PL2203. 9, paratype NMV P304913, librigena, lateral view x 1.7 (internal mould) from PL2327. 10, paratype NMV P304873, thoracic segment, dorsal view x 2.0 (internal mould) from PL2203. 11, paratype NMV P3047 12, cranidium, dorsal view x 1 .25 (latex cast) from PL2203. 12, paratype NMV P304864, pygidium, dorsal view x 1 .4 (internal mould) from PL2203. 13, paratype NMV P304861, pygidium, dorsal view x 1.8 (latex cast) from PL2203. 14, paratype NMV P304860, pygidium, ventral view (doublure) x 1.25 (internal mould) from PL2203. 15, paratype NMV P304863, pygidium, dorsal view x 1.5 (internal mould) from PL2203. 16, paratype NMV P304865, pygidium, dorsal view x 1.7 (latex cast) from PL2203. 17a, holotype NMV P304717, pygidium, lateral view x 1.1 (internal mould) from PL2203. 17b, same, dorsal view x 1.3. 17c, same, x 1.1 (latex cast). 18, paratype NMV P304862, pygidium, lateral view x 1.5 (internal mould) from PL2203. 26 Andrew C. Sandford Homalonotid trilobites from the Silurian and Lower Devonian horizontal. In lateral view dorsal profile of axis inclined at about 30° to the horizontal, postaxial ridge very steeply inclined. Discussion. Digonus wenndorfi is abundant and dominates the trilobite fauna at several localities. Except for a cephalon (Holloway and Neil, 1982: figs 4A, 4B) the species is known only from isolated tergites. Relative proportions of many features (eg. glabellar and pygidial length/width ratios) are dif- ficult to ascertain due to deformation. Nevertheless, two distinct forms of wenndorfi can be recognised, elongate and short morphs differing in the relative proportions of the cephalon, glabella and pygidium. These morphs are not related to size, nor are they products of deformation as their ranges in size overlap, and they occur together on the same bedding plane (Fig. 3). The differing morphological proportions within the population are strongly bimodal rather than continuous. Pygidial length-width ratios range between 0.76-0.93 (average 0.86) in the short morphs, in the long morphs between 1.11-1.12. Similarly, glabellar length-width ratios range between 0.96-1.00 (average 0.99) in the short morphs, in the long morphs between 1.15-1.23 (average 1.19). Similar observations have been made for populations of D. gigas and attributed to sexual dimorphism (Dahmer, 1914), although Wenndorf (1990: table 31) demonstrated, with a much larger sample size of 67 pygidia, that the variation in gigas is con- tinuous rather than bimodal, with intermediate forms being the most numerous, and so not reflecting sexual dimorphism. Although the sample size for wenndorfi is much smaller (11 measurable specimens) the complete absence of intermediate forms (e.g. pygidia with length-width ratios range between 0.93-1.1) is statistically significant. Whether the bimodality reflects sexual dimorphism is uncertain, although the narrower morphs strongly outnumber the wider morphs and so suggest that it does not. The subquadrate outline of the anterior part of the cranidium, the weakly tapered trapezoid glabellar outline, the wide pygidial axis, and the depth of the pygidial ring and pleural furrows clearly indicate assignment to Digonus. D. wenndorfi differs markedly from the D. gigas group and more closely resembles the D. ornatus group in having a relatively longer preglabellar field, wide fixigenae anteriorly, and a transverse rather than weakly concave anterior cranidial margin. Digonus wenndorfi differs from most Pragian-Emsian species of Digonus in lacking the elongate, straight-sided tri- 27 angular outline and a long produced tip characteristic of the type and other taxa such as D. ornatus and D. intermedius. In having shorter pygidial proportions, wenndorfi more closely resembles other Lochkovian species, including D. laticaudatus from North America, D. armoricanus from France, and D. roe- meri from Europe. Compared to Pragian-Emsian species, Lochkovian species of Digonus lack some of the features characterising Digonus. These species can be interpreted as reflecting ancestral morphologies, as might be expected in the earliest representatives of a genus. ‘Trimerus- like’ features including the raised postaxial ridge and weak glabellar lobation of wenndorfi can be interpreted in this context. Of Lochkovian species of Digonus, only D. zeehanensis from Tasmania (redescribed below) is known sufficiently for detailed comparison with D. wenndorfi. These taxa are very close, but the latter differs in having slightly concave glabellar sides, a longer glabella, more distinctly expressed glabellar lobation, a shorter preglabellar field, a much shorter pygidium with a relatively narrower pygidial axis, a more obtusely- angled pygidial tip, and a lower postaxial ridge (Fig. 13). With the exception of the latter, these features suggest that zeehanensis represents a less derived morphology. Digonus roemeri is a near-contemporary of D. wenndorfi, recorded from the base of the Lochkovian ( [hesperius zone) in Europe (Chlupac et al., 2000). The species is poorly docu- mented and, as noted by Tomczykowa (1975), a number of taxa have been assigned to it. Topotypes of D. roemeri (Morzadec, 1986: pi. 32 figs 1, 4, 6-10) share with wenndorfi short pygidial proportions and the equally deep pygidial pleural and ring furrows typical of Digonus. D. roemeri differs from wenndorfi in that the pygidial axis is wider and the postaxial ridge is not raised. In contrast to other species assigned to the genus, pygidia of roemeri and the comparable D. laticaudatus exhibit anteriorly continuous pygidial axial furrows, interpreted here to reflect the ancestral morphology. Many species assigned to Digonus are poorly documented and comparison with D. wenndorfi is difficult. From Europe, D. harpyius is known only from a rostral plate. The Emsian D. goniopygaeus from England and D. crassicauda from Germany are kn own only from pygidia. Despite the compar- able short proportions, the pygidium of crassicauda differs from that of wenndorfi in having a weakly defined postaxial ridge and a narrow, recurved flange defining the pygidial tip (Wenndorf, 1990: pi. 9 figs 1-3), as in other Emsian Digonus. The Antarctic D. antarcticus differs markedly from wenndorfi Figure 13. Digonus zeehanensis (Gill, 1949). la, NMV P304656, cephalon, dorsal view x 1.2 (internal mould) from PL1726. lb, same, enlarge- ment of anterior margin x 2.5. 1c, same, lateral view of cephalon x 1.0. Id, same, dorsal view x 1.25 (latex cast). 2a, NMV P304657, cephalon, dorsal view x 2.6 (internal mould) from “Little Henty River”. 2b, same, lateral view x 2.25. 3, NMV P304650, hypostome, ventral view x 2.2 (internal mould) from PL1726. 4a, NMV P304653, librigena, dorsolateral view x 1.3 (latex cast) from PL1726. 4b, same, lateral view. 5, NMV P304652, thoracic segment, lateral view x 3.0 (latex cast) from PL1726. 6, NMV P14789, cranidium, dorsal view x 0.9 (latex cast) from PL1726. 7a, holotype NMV P14590, cranidium, dorsal view x 1.3 (latex cast) from PL1726. 7b, same, internal mould x 1.1. 8, NMV P304683, hypostome, ventral view x 2.1 (latex cast) from PL1726. 9, NMV P304658, cranidium, dorsal view x 1.6 (latex cast) from PL6653. 10, NMV P304649, pygidium, dorsal view x 1.1 (internal mould) from PL1726. 11, NMV P304655, cranidium, dorsal view x 1.0 (latex cast) from PL1726. 12, NMV P7651, pygidium, dorsal view x 1.65 (internal mould) from “MtZeehan”. 13a, NMV P304648, pygidium, lateral view x 1.1 (internal mould) from PL1726. 13b, same, dorsal view. 14, NMV P14787, pygidium, dorsal view x 0.75 (latex cast) from PL1726. 15a, paratype NMV P14592, pygidium, lateral view x 1.0 (internal mould) from PL1726. 15b, same, dorsal view (latex cast). 15c, same, dorsoposterior view (internal mould). 15d, same, dorsal view. 16, NMV P304659, pygidium, dorsal view x 1.0 (latex cast) from “Mt Zeehan”. 28 Andrew C. Sandford in having weakly expressed pygidial axial furrows, exception- ally deep pleural furrows and a poorly expressed postaxial ridge (Saul, 1965: pi. 17 figs 1-11). Holloway and Neil (1982) considered specimens of Digonus wenndorfi to be close to Trimerus lilydalensis Gill, 1949, also from the Lochkovian of central Victoria. The availability of many new specimens of wenndorfi and a revision of Gill’s species does not support this comparison, with the latter assigned to the new genus Wenndorfia. Digonus zeehanensis (Gill, 1949) Figure 13 Trimerus zeehanensis Gill, 1949: 70, pi. 9 figs 1, 2, 4, text-fig. ID. — Tomczykowa, 1975: 11. — Wenndorf, 1990: 16. — Holloway and Sandford, 1993: 93. — Schraut, 2000: 382. Type material. Holotype NMV PI 4590 (cranidium, counterpart pre- viously registered NMV P14591, Fig. 13.7) and paratype NMV P14592 (pygidium, counterpart previously registered NMV P 1459 13) from PL1726, Gill and Banks locality 16, Zeehan, Tasmania. Registered material. 67 specimens: 2 cephala, 13 cranidia, 12 libri- genae, 2 hypostomes, 11 thoracic segments, 27 pygidia. NMV P14787, P14788 (counterparts), NMV P14789, NMV P14791, NMV P304648-P304656, NMV P304662, NMV P304664-P304708 from PL1726. NMV P304657 from “Little Henty River”, Zeehan. NMV P304661 from “near Little Henty River”, Zeehan. NMV P304663 from PL1725, Gill and Banks’ locality 15, Zeehan. NMV P304660, NMV P304658 from PL6653, Zeehan. NMV P7561, NMV P304659 from “Mt Zeehan”, Zeehan. Stratigraphic distribution. Bell Shale, Boucotia australis Assemblage Zone, mid-late Lochkovian. Diagnosis. Cephalon trapezoid, sides straight, anterior margin tricuspate, with short (0.05 times cephalic length) but wide (length 0.25 times width) median cusp, lateral cusps about half? the length of the median cusp. Glabella trapezoid, sides weakly concave and converging at about 20°, length about 1.2 times width, anterior margin well defined, lobation weakly defined. Axial furrows moderately impressed. Palpebral lobe placed with midline opposite 0.52 glabellar length/0.42 crani- dial length. Preglabellar field with length (sag.) 0.17 times cranidial length. Anterior branches of facial suture straight and converging at about 50°, curving abruptly to the exsagittal anteriorly. Rostral suture transverse. Dorsal surface of rostral plate triangular, short, length 0.15 times width. Hypostome with middle furrow deep adaxially, shallow abaxially. Posterior border of hypostome with long lateral lobes of length 0.17 times hypostomal length. Pygidium short, length 0.68 times width, sides moderately convex and converging posteriorly at about 110°. Pygidial axis 0.45 times pygidial width, 12 axial rings, continuous with low postaxial ridge. Axial furrows taper- ing at about 30°, moderately impressed posteriorly, not contin- uous anteriorly opposite first ring. Pleural and ring furrows deep. 8 ribs, rib-ring medially offset at fifth rib. Dorsal exoskeleton with coarse pitting. Discussion. The Bell Shale is strongly sheared, resulting in wide variations in shapes and proportions of specimens. Because Gill (1949) described the species from a limited number of specimens, examination of a larger population reveals a number of inaccuracies in his description. The concavity of the preglabellar field and the upturning of the anterior cranidial margin on the holotype (Fig. 13.7), consid- ered diagnostic of the species by Gill, are features attributable to tectonic distortion and are absent from other topotype cra- nidia. Gill recorded lateral glabellar furrows as absent, but they are weakly defined on most specimens. The pygidial axial fur- rows are moderately impressed rather than poorly defined. Despite the abundance of cephalic and pygidial tergites, thoracic pleurae are unusually rare in the Bell Shale. From the fragments available, there is no indication of spines on the thorax as suggested by Gill. Tomczykowa (1975) and Wenndorf (1990) assigned Gill’s species to Trimerus, but the quadrate course of the anterior branches of the facial and rostral sutures, the equally deep pygidial ring and pleural furrows and the effacement of the axial furrow anteriorly are more consistent with an assignment to Digonus. Pygidia have a much wider axis (average width 0.46 pygidial width) than recorded by Gill, who cited the axial width to be one third the pygidial width. However, as noted above, the longer glabellar proportions, more distinctly expressed glabellar lobation, shorter pygidial proportions with a relatively narrower axis and a more obtusely-angled pygidial tip suggest that D. zeehanensis represents a less derived morphology compared to other Digonus. In these features, zeehanensis is most closely comparable to D. roemeri from the basal Lochkovian of France. Environmental notes. Digonus zeehanensis dominates the tr- ilobite fauna of the Bell Shale. At the best-sampled locality (PL1726), the relative abundance of D. zeehanensis is >90%. The fauna is typical of homalonotid-dominated assemblages, being low diversity and otherwise comprising a proetid, a cheirurid and a dalmanitid (Gill, 1950). The latter suggests alignment with the deeper water Trimerus delphinocephalus- Dalmanites limulurus Community of Mikulic, 1999. The pro- portion of articulated specimens in populations of zeehanensis is low. Of the 67 specimens, all are isolated tergites with the exception of three cephala. Breakage is high in the sampled populations (Br=27%). Specimens occur on bedding planes in association with other fossil concentrations or in more sparsely fossiliferous layers between the lag bands. The taphon- omy is indicative of taphofacies Till and together with the lithofacies and biofacies suggests a setting around maximum storm wave base. Dipleura Green, 1832 Type species: Dipleura dekayi Green, 1832. From the Middle Devonian Hamilton Group, New York State, USA, by monotypy. Other species included. Dipleura boliviensis Wolfart, 1968, D. gar- ratti sp. nov., D. iberica Wenndorf, 1990, Homalonotus kayseri Thomas, 1905, H. laevicauda Quenstedt, 1852, = H. simplex Richter and Richter, 1926, D. lanvoiensis Morzadec, 1969, D. praecox Tomczykowa, 1975, D. sp. (in Richter and Richter, 1943), D. sp. (in Morzadec, 1981), D. sp. nov. (in Wenndorf, 1990). Range. Upper Silurian (mid Ludlow)-Middle Devonian (Givetian). Homalonotid trilobites from the Silurian and Lower Devonian 29 Revised diagnosis. Cephalon short, length 0.5 to 0.6 times cephalic width, rounded triangular to semicircular in outline. Glabella with sides straight or weakly concave, subparallel to moderately convergent (20°), length 1.1 to 1.25 times width, without lobation in maturity. Preglabellar field of moderate length, 0.15-0.18 times cranidial length. Paraglabellar areas indistinct. Palpebral lobes posteriorly placed, opposite 0.3 to 0.45 cranidial length. Eyes remote, 6-6 generally 1.7-1. 9 times preoccipital glabellar width. Anterior branches of facial sutures describing a broad curve and strongly convergent anteriorly, between 80-100° opposite midlength of preglabellar field. Rostral suture transverse to moderately convex forwards. Ventral surface of rostral plate flat, without projection. Pygidium with weakly defined trilobation, axial furrows very shallow to effaced. Axis with ring furrows moderately impressed to effaced, with weak to distinct posterior swelling, without well defined postaxial ridge. Pleural furrows shallow to effaced, invariably as shallow or shallower than ring furrows. Finely papillate ornament. Discussion. As with Digonus, Parahomalonotus and Burmeisteria, the type species of Dipleura is one of the youngest members of the genus and differs significantly from older forms. General morphological changes in the genus over time include a trend toward a less tapered glabellar shape with increasingly concave sides, and a decrease in the definition of pygidial segmentation and trilobation. In Middle Devonian species a stronger cephalic convexity is developed, and the contrast in depth between ring furrows and pleural furrows is not as marked. The revised diagnosis differs from previous ones in recognising the limited significance that pygidial out- line holds as a generic character. In the case of Dipleura , the oldest (early Ludlow) species, Dipleura garratti from Victoria, has a long, acuminate triangular-shaped pygidium with an acutely pointed tip, in contrast to the short parabolic outline and rounded pygidial tip exhibited by Devonian taxa and consid- ered characteristic of the genus. The upper Ludlow D. praecox from Poland has an intermediate morphology, having a short triangular outline with an angular tip. Emphasis is placed on several characters listed by Salter (1865), Reed (1918) and Sdzuy (1959) but omitted by Tomcykowa (1975) and Wenndorf (1990). These include the course of the facial and rostral sutures, eye position and weak expression of paraglabellar areas. Several new diag- nostic characters are added, including the absence of a projection or keel on the ventral surface of the rostral plate, the presence of a weak to distinct posterior swelling of the pygidial axis and the absence of a well defined postaxial ridge. The revised concept of the genus excludes a number of species that have been previously assigned to Dipleura (see Wenndorf, 1990: table 1, Tomczykowa, 1975: table 3). These include Burmeisteria ( Digonus ) accraensis from West Africa, Homalonotus clarkei from Bolivia, Dipleura salteri Morris, 1988 from England and Trimerus (Dipleura) fornix Haas, 1968 from Turkey. Generic assignment of the two former species is discussed under Digonus. D. salteri is tentatively assigned to Trimerus ( Ramiotis ), discussed below. T. (D.) fornix is assigned below to Wenndorfia, erected for many of the species previous- ly assigned to Parahomalonotus. Cephalic outline, the stronger segmentation of the pygidial axis compared to the pleural field, and the definition of the pygidial axial furrow posteriorly were considered by Wenndorf (1990) to be important generic char- acters in distinguishing Dipleura from morphologically con- vergent species of Parahomalonotus. In restricting Parahomalonotus to a group close to the type species, the dis- tinction between the two genera is marked. In particular, the strong glabellar and pygidial axial convexity, strong glabellar lobation and strong pygidial segmentation easily separate members of Parahomalonotus s.s. from those of Dipleura. The problem of convergence addressed by Wenndorf re-emerges however, between Dipleura and species assigned below to Wenndorfia. Dipleura and Wenndorfia share many characters including a weakly tapering glabellar outline, effaced glabellar lobation, a posterior eye position, a preglabellar field of mod- erate length, an effaced pygidial postaxial ridge and pygidial outlines that are for the most part rounded parabolic. For the latter character triangular outlines are exceptional, expressed in early representatives of each genus such as D. garratti and W. lilydalensis (Gill, 1949). Many species assigned to Wenndorfia can be immediately distinguished from Dipleura in having dis- tinct pygidial segmentation, with pleural furrows being moder- ately to deeply impressed. Such species include W. multicosta- tus (Koch, 1883a), W. miloni (Renaud, 1942), W. elegans (Tomcykowa, 1975), W. bostoviensis (Tomczykowa, 1975), W. plana junior (Wenndorf, 1990) and W. sp. (=Digonus vialai in Tomczykowa, 1975). However, in a number of taxa the pygidi- al axial and pleural furrows are weakly impressed to indistinct, including W. plana plana (Koch, 1883a), W. forbesi (Rouault, 1855) and W. fornix. Assignment of these convergent species to Wenndorfia is indicated by consistent differences in the course of the facial sutures between Dipleura and Wenndorfia. In Dipleura , the facial sutures tend to be more strongly convergent and the rostral suture more forwardly convex, giving a more parabolic or V-shaped outline to the anterior margin of the cranidium. By contrast, the rostral suture of Wenndorfia is transverse or weakly convex forwards. The anterior branches of the facial sutures tend to be less strongly convergent posterior- ly, abruptly curved medially and more strongly convergent anteriorly, giving the anterior margin of the cranidium a more rounded, U-shaped outline. Reed (1918) suggested that Dipleura was derived from Digonus. More recent phylogenies are in accord in deriving Dipleura directly from Trimerus, although there is disagree- ment in the chronology of this transition. Sdzuy (1957: fig. 3) considered Dipleura to have originated from Trimerus at the Siluro-Devonian boundary. Tomczykowa (1975) established an earlier origin for Dipleura, and suggested two species from the upper Ludlow of Poland, Dipleura praecox Tomczykowa, 1975 and Trimerus permutus Tomczykowa, 1978 from the immedi- ately underlying beds, to be closely related and to represent a transitional li nk between the genera. An earlier origin for Dipleura is indicated by its occurrence in lower Ludlow strata of Victoria, supporting Wenndorf’s (1990: fig. 6) suggestion of a Wenlock origin for the genus. Dipleura is presumably derived from Lower Silurian 30 Andrew C. Sandford Trimerus. Lower Silurian Trimerus are variably assigned in this work to T. ( Trimerus ) or to T. ( Ramiotis ). T. ( Trimerus ) is characterised by a suite of distinctive cephalic and pygidial features, notably a glabellar outline markedly expanded across LI -LI (tr.), raised glabellar profile, the strongly-defined glabellar lobation, L1-L2 muscle scars, sagittal ridge, exsa- gittal furrows and paraglabellar areas, the long and weakly con- cave (tr.) preglabellar field, and the produced pygidial tip. In the absence of most of these features species assigned to T. ( Ramiotis ) more closely resemble Dipleura. Llandovery species of T. ( Ramiotis ) exhibit features that show a strong resemblance to those typical of Dipleura, sharing trapezoid glabellar outlines, shorter preglabellar fields and marked contrast between the depth of the ring furrows and pleural furrows, with the latter very shallow to almost effaced. Sdzuy (1957) interpreted Homalonotus ( Trimerus ) mongo- licus Tchemycheva, 1937 from the Upper Silurian of Mongolia as an intermediate form between Dipleura and Trimerus. Sdzuy emphasised the posterior eye position, the rounded pygidial outline and the marked contrast between the depth of the ring furrows and almost effaced pleural furrows to support his inter- pretation. The significance of these characters is overstated by Sduzy. The markedly weak pleural furrows and rounded pygidial outline of mongolicus emphasised by Sdzuy are not features unique to Dipleura and occur amongst species of T. ( Ramiotis ) and T. ( Edgillia ). The eyes of mongolicus are clearly placed directly opposite the glabellar midlength (Tchemycheva, 1937: pi. 1 fig. 8). The rounded pygidial outline emphasised by Sdzuy is contradicted by Tchemycheva’ s description of the pygidium as “triangular, usually elongated, terminates in a thick spine ” (Tchemycheva, 1937: 25). In strong contrast to Dipleura, the species exhibits a very long preglabellar field (0.3 cranidial length) and the anterior branch- es of the facial sutures are relatively weakly convergent (60°). Tchemycheva’ s description of other features including a sub- quadrate, weakly tapering glabellar outline, an evenly vaulted glabellar profile, weak SI and indistinct S2 and S3, 13 pygidial axial rings, shallow pygidial pleural furrows, the absence of a postaxial ridge ( “rhachis falling short of poster- ior margin”) and the pygidial proportions (length 0.92 width) indicate assignment to T. (Edgillia). Dipleura garratti sp. nov. Figure 14 Homalonotus sp. — Chapman, 1908: 220. Type material. Holotype NMV P308644 (pygidium) from PL6615, Eden Park, Victoria (Fig. 14.13). Paratypes NMV P308638, NMV P308641 (cranidia), NMV P308639, NMV P308642, NMV P308645 (cephala), NMV P308643 (librigena), NMV P308637, NMV P308640, NMV P308646 (pygidia) from PL6615. Paratypes NMV P308635-6 (cephalothoraxes) , NMV P308649, NMV P308651, NMV P308663, (cephala), NMV P308657 (cranidium) from PL6614, Eden Park. NMV P308648 from “Whittlesea”, Victoria. For localities see Fig. 11. Registered material. 47 specimens: 4 cephalothoraxes, 8 cephala, 17 cranidia, 2 librigenae, 2 thoracic segments, 14 pygidia. NMV P3045 1 3-P3 045 1 5 , NMV P304572, P304573, NMV P308633-P308636, NMV P308649-P308657, NMV P308673 from PL6614. NMV P308637-P308647, NMV P308659-P308672 from PL6615. NMV P308658 from PL6625, Wandong, Victoria. NMV P308648 from “Whittlesea”. Unregistered specimens from PL1793, Clonbinane and Kenley locality 14c, Upper Plenty. For localities see Fig. 11. Stratigraphic distribution. As for Homalonotus williamsi. Derivation of name. For Michael J. Garratt, for his contribution to Victorian palaeontology and stratigraphy. Diagnosis. Glabella trapezoid, length 1.05 times preoccipital glabellar width, sides straight and converging at about 25°, anterior margin well defined, very broadly rounded to trans- verse. Length (sag.) of preglabellar field 0.15-0.18 times cran- idial length. Palpebral lobe placed with midline opposite 0.47 times glabellar length/0.38 times cranidial length. Anterior branches of facial suture angular, subparallel to axial furrows to a point opposite 0.75 glabellar length, anteriorly converging at about 80°. Rostral suture broadly curved. Ventral surface of rostral plate with length about equal to width, flat, connective sutures angular, anterior section converging at 40 degrees, pos- terior section at 80°. Pygidium triangular, length equal to width, sides straight and converging at 65°, tip acutely angular. Pygidial axis with width 0.5 times pygidial width, 12 axial rings, ring furrows moderately impressed, axis moderately swollen posteriorly, continuous with wide but poorly defined postaxial ridge. Axial furrows straight and tapering at about 33°, weakly impressed to indistinct. Pleural furrows weakly impressed to indistinct. 6 pleural ribs, rib-ring medially offset at second rib. Description. Exoskeleton small (maximum length estimated 8 cm from NMV P308644), occipital and pygidial convexity (tr.) moderate. Dorsal exoskeleton finely granulose. Cephalon with width about 1.6 times length, with semielliptic out- line, sides moderately convex. Cranidial width about 1.64 times length. Figure 14. Dipleura garratti sp. nov. 1, paratype NMV P308638, cranidium, dorsal view x 3.0 (internal mould) from PL6615. 2, paratype NMV P308645, cephalon, dorsal view x 3.0 (latex cast) from PL6615. 3, paratype NMV P308646, pygidium, dorsal view x 3.8 (internal mould) from PL6615. 4a, paratype NMV P308663, cephalon, ventral view (doublure) x 4.1 (latex cast) from PL6614. 4b, same, enlargement of rostral plate x 15. 5, paratype NMV P308641, cranidium, dorsal view x 3.0 (internal mould) from PL6615. 6, paratype NMV P308642, cephalon, dorsal view x 2.2 (internal mould) from PL6615. 7, paratype NMV P308643, librigena, ventral view (doublure) x 2.6 (latex mould) from PL6615. 8, paratype NMV P308657, cranidium, dorsal view x 1.9 (internal mould) from PL6614. 9, paratype NMV P308649, cephalon, dorsal view x 2.0 (internal mould) from PL6614. 10a, paratype NMV P308635, cephalothorax, dorsal view of cephalon x 2.2 (internal mould) from PL6614. 10b, same, dor- sal view of thorax x 2.0. 11, paratype NMV P308639, cephalon, dorsal view x 1.8 (internal mould) from PL6615. 12, paratype NMV P308640, pygidium, dorsal view x 3.5 (internal mould) from PL6615. 13a, holotype NMV P308644, pygidium, dorsal view x 2.1 (internal mould) from PL6615. 13b, same, posterior view. 13c, same, lateral view x 2.8. 14, paratype NMV P308648, cephalon, oblique view x 3.0 (latex cast) from PL6614. 15, paratype NMV P308637, pygidium, dorsal view x 3.5 (internal mould) from PL6615. 16, paratype NMV P308636, cephalothorax, lateral view x 3.0 (internal mould) from PL6614. 17, paratype NMV P308651, cephalon, dorsal view x 3.3 (internal mould) from PL6614. Homalonotid trilobites from the Silurian and Lower Devonian 31 32 Andrew C. Sandford Glabellar length about 0.8 times cranidial length. Occipital ring 0.12 times glabellar length, slightly wider medially. Occipital furrow deeply impressed, with weak forward flexure medially. Glabellar lobation extremely weak to indistinct, best seen on NMV P308638 as very shallow depression at adaxial end of SI placed opposite 0.3 times glabellar length. Paraglabellar area very weakly defined. Length (exsag.) of posterior border equal to occipital length adaxially, length- ening slightly abaxially. Posterior border furrow transverse adaxially, curving gently forwards abaxially, very wide, moderately impressed, terminating distally. Postocular fixigenal area very short, length (exsag.) 0.15 times cranidial length. Palpebral lobes placed remotely (6-6 1.65 times preoccipital glabellar width). Palpebral lobe length (exsag.) 0.15 times cranidial length, palpebral furrow indistinct. Preocular fixigenal area of moderate width, 0.18 5-6, narrowing slightly anteriorly, eye ridges very weakly defined. Librigena without distinct border furrow or lateral border. Dorsal surface of rostral plate very short (sag.), crescentic in outline. Ventral surface of rostral plate kite-shaped, posterior width 0.2 times maximum width. Hypostomal suture extremely weakly curved. Thorax with thirteen segments. Axial furrows extremely shallow, expressed as diagonally directed furrows on each segment, each meeting posterior margin at axial articulating process on internal mould. Pleural furrows wide and deep across axis, narrower (exsag.) and deeper across pleural field. Pygidial border furrow and border poorly defined. In posterior view posterior margin of pygidium horizontal. In lateral view dorsal profile evenly inclined, interrap ted by swelling of terminal piece. Discussion. The assignment of this species to Dipleura is indi- cated by the course of the facial sutures (anterior branches con- verging at 90°), the posterior position of the eye and the almost effaced pygidial axial furrows and pleural furrows. The short preglabellar field and indistinct glabellar lobation are in accord with this assignment. D. garratti is the earliest representative of the genus. A number of differences with later Dipleura can be interpreted as reflecting the closer affinities of garratti with Trimerus, particularly its triangular pygidial outline and its pos- teriorly raised axis that is continuous with a wide postaxial ridge (Fig. 14.12). In many features Dipleura garratti is most closely com- parable to the German upper Pragian-lower Emsian D. laevicauda, sharing a moderately tapered glabellar outline, more elongate pygidial proportions and a relatively wide pygidial axis. The course of the cephalic sutures is most like that of the Bolivian Givetian D. boliviensis. Originally described as a subspecies of D. dekayi, the relatively longer cephalic proportions, more tapering glabella, more anteriorly and less remotely placed eyes are significant in regarding boliviensis as an independent species. Environmental notes. Dipleura garratti occurs in a trilobite fauna dominated by proetids and dalmanitids, with Homalon- otus williamsi, aulacopleurids, odontopleurids and phacopids in very low abundance. Although known from various localities between Clonbinane and Eden Park, the fauna is well repre- sented at two richly fossiliferous localities in the Macropleura band at Eden Park, a horizon characterised by an abundance of the large, thick-shelled brachiopod M. densilineata. At PL6615, dalmanitids (relative abundance 50%) are more abundant than proetids (relative abundance 21%), whereas at PL6614 the proetids are dominant (relative abundance 55%) over dalman- itids (relative abundance 26%). The relative abundances of garratti and williamsi are similar at both localities (17-19% and 2% respectively). The taphonomy and facies of Dipleura garratti differs markedly between PL6614 and PL6615. At PL6615, 87% are isolated tergites and the remainder are cephala. With the excep- tion of the cephalon of Homalonotus williamsi all other trilobites associated with garratti are represented by isolated tergites, with the proportion of broken specimens 16%. The lithology is best described as a bioclastic coquina in a siltstone matrix. Assignment to shallow taphofacies TII is in accord with Garratt’s (1983) assignment of the brachiopod fauna to the shallow-water Notoconchidium Community (BA2). A moderate energy environment at depths around normal wave base is indicated. Slightly deeper conditions are indicated by the taphonomy and lithofacies at PL6614, where 21% of specimens of garratti are cephala and 21% are cephalothoraxes. The pro- portion of broken specimens (9%) is lower than at PL6615. The lithology at PL6614 is a poorly bedded micaceous siltstone, and although bioclasts are extremely abundant they have not been winnowed to form a coquina. The preservation suggests an environment at depths below normal wave base. Parahomalonotus Reed, 1918 Type species. Homalonotus gervillei De Verneuil, 1850. From the Pragian of France, Turkey, Spain, Morocco, Romania and Turkey, by original designation. Other species included. Parahomalonotus diablintianus Morzadec, 1976, Homalonotus vialai Gosselet, 1912 (in Gosselet et al.), P. sp. A (=P. gervillei in Haas, 1968), P. sp. B (=P. aff. gervillei in Wenndorf, 1990). Other species tentatively included. Digonus sp. cf. E in Morzadec, 1986. Range. Lower Devonian. Revised diagnosis. Glabella moderately to strongly convex (tr. sect.) and narrow, with length 1.2- 1.6 times width. Axial furrows moderately to deeply impressed. Lobation distinct, S1-S3 moderately to deeply impressed. Pygidial axis narrow (0.35-0.4 times pygidial width anteriorly), weakly tapering (-25-30°), with moderate to strong convexity independent of the pleural convexity, raised posteriorly, continuous with wide postaxial ridge reaching posterior margin, pygidial tip with small point or short slender spine. Discussion. The restricted diagnosis of Parahomalonotus defines a tight group distinct from Wenndorfia, to which many species previously assigned to Parahomalonotus have been assigned. Species assigned to Parahomalonotus also share weak glabellar tapering (0-18°), a preglabellar field of mod- erate (0.2) to short (0.05) length, eyes placed posteriorly (opposite 0.35-0.5 cranidial length), subparallel to moderately convergent anterior branches of facial suture, a transverse rostral suture, an obtusely angled (130-150°) to rounded pygidial tip and equally well defined axial rings and pleural ribs. Pygidial morphology of Parahomalonotus is distinct from that of Wenndorfia and is conservative, with the exception of Homalonotid trilobites from the Silurian and Lower Devonian 33 P. diablintianus in which pygidial ring and pleural furrows are relatively shallow and the pygidial outline distinctly rounded, hi cephalic features, the younger late Pragian-Emsian forms (P. sp. A, P. gervillei ) differ from earlier species in their stronger glabellar convexity and lobation, and greatly reduced preglabellar fields. Of the earlier taxa, P. vialai shows the closest affinities with the type species and represents an inter- mediate morphology in which glabellar furrows and lobation are more moderately expressed. To the species Parahomalonotus sp. A illustrated by Haas (1968: pi. 30 figs 1-4 as P. gervillei ) from the Emsian of Turkey can be added at least one of the cranidia (pi. 29 fig. 26) described as Dipleura fornix, differing from other specimens of fornix in having a transverse glabellar anterior margin and straight rather than concave glabellar sides. Trimerus Green, 1832 Type species. Trimerus delphinocephalus Green, 1832 from the Wenlock Rochester Shale, New York State, by monotypy. Subgenera included. Trimerus ( Trimerus ) Green, 1832, T. (Ramiotis) subgen. nov., T. ( Edgillia ) subgen. nov. Range. Llandovery-Lochkovian. Revised diagnosis. Cephalon triangular to rounded pentagonal in outline, length 0.6-0.7 times width, anterior margin entire, trilobed or tricusped. Preglabellar field of moderate length to very long, 0.16-0.35 times cranidial length, flat or weakly con- cave (tr.) medially. Glabella generally with distinct lobation and median indentation of anterior margin. Eyes placed forwardly, with midline of palpebral lobes opposite 0.5-0.75 glabellar length (0.38-0.5 cranidial length) and medially on genae (6-5 1.3- 1.7 times preoccipital glabellar width). Anterior branch of facial suture more or less straight and moderately convergent (45-60°) between eye and a point opposite midlength of preglabellar field. Rostral suture on dorsal surface, transverse or arching forwards evenly or with a median angle. Paraglabellar areas large. Ventral surface of rostral plate flat or weakly convex, without process. Thoracic axial furrows poorly defined. Pygidium with axis of moderate width (0.3-0.6 times pygidial width), not reaching posterior margin, axial furrows moderately impressed, axial ring and pleural furrows distinct but of variable depth, 6-12 rings, 5-10 ribs. Discussion. The revised diagnosis incorporates and quantifies most features listed in more recent diagnoses of Trimerus. New characters include the relatively forward position of the eye, the straight course of the anterior branch of the facial suture, and the moderate axial width. Early diagnoses, such as Salter’s (1865), predominantly list characters of a very general nature. Three species groups are recognised in this work and defined as the subgenera Trimerus (Trimerus), T. (Ramiotis) and T. (Edgillia). Previous diagnoses of Trimerus emphasise the well-defined glabellar lobation, triangular pygidial outline and acuminate pygidial tip (i.e. a mucro) exhibited by the type species and shared with closely related species including T. (T.) cylindricus, T. (T.) vomer (Chapman, 1912) and T. (T.) johan- nis. These features and a suite of other distinctive glabellar features are considered here to be diagnostic only of the typical subgenus T. (Trimerus). The forwardly arcuate course of the rostral suture and the short length (sag.) of the dorsal section of the rostral plate noted in previous diagnoses of Trimerus (Tomczykowa, 1975, Thomas in Curtis and Lane, 1998) is considered here to be only of specific significance for T. (T.) delphinocephalus. A second group comprising other Silurian species of Trimerus can be recognised, and is described in this work as T. (Ramiotis). Members of this group lack the produced pygidi- al tip and distinctive glabellar features of T. (Trimerus). The group is typified by a new upper Llandovery species from Victoria, T. (R.) rickardsi. The species assigned to this group share a suite of characters including less acute to obtusely angled pygidial tips lacking a mucro, and an elongate, weakly tapering, more or less straight-sided glabella with weakly defined lobation. Tomczykowa (1975) included a slight degree of tapering of the glabella in her diagnosis of Trimerus, but this character is peculiar to T. (Ramiotis). The degree of glabellar tapering in Trimerus varies between the strongly tapered (-55°) and posteriorly expanded morphologies of T. (Trimerus), moderately tapered trapezoid morphologies of T. (Ramiotis), and very weakly tapered (<15°) subquadrate morphologies of T. (Edgillia), the latter erected in this work for a group of Upper Silurian-Lower Devonian species. T. (Ramiotis) is further distinguished from T. (Trimerus) by a tendency towards a shorter preglabellar field, shorter pygidial proportions and more weakly defined pygidial segmentation. Trimerus (Edgillia) can be distinguished from the Silurian sub- genera by a distinctive morphology typified by the Victorian T. (E.) kinglakensis (Gill, 1949). The species assigned share a very weakly tapered, subquadrate glabellar outline, weakly defined glabellar lobation, and lack the raised postaxial ridge present on most Silurian representatives of Trimerus. Sdzuy (1959) considered Trimerus to comprise also the sub- genus Dipleura. The revised diagnosis of Dipleura is in accord with Tomczykowa (1975) and Wenndorf (1990) in recognising Dipleura and Trimerus as independent genera. Sdzuy’s diag- nosis of T. (Trimerus) emphasises differences from Dipleura, listing the longer cephalic proportions, a more tapering gla- bella, distinct glabellar lobation and paraglabellar areas, a for- wardly convex rostral suture, an acuminate pygidial tip, and more distinct pygidial segmentation. Diagnoses of Trimerus by Thomas (in Curtis and Lane, 1998) and Tomczykowa (1975) closely follow Sdzuy’s brief subgeneric diagnosis, but incor- porate only some of the features listed by him. Importantly, both Thomas and Tomczykowa omitted the morphology of the rostral plate as a diagnostic character. Holloway and Neil (1982) suggested the significance of homalonotid coaptative morphologies, and in accord with this view the absence of a rostral process is reinstated in the diagnosis. In emphasising this feature and in the presence of other differences, a number of Lower Devonian species previously assigned to Trimerus are excluded, including Homalonotus noticus from South America and South Africa), Burmeisteria (Digonus) accraensis from Ghana and Trimerus novus Tomczykowa, 1975 from Poland. Edgecombe and Fortey (2000) reassigned Homalonotus linares Salter, 1861 from the Pffdolf of Bolivia to Trimerus. The species exhibits a number of features that are not com- 34 Andrew C. Sandford patible with this assignment. The axial ring count (14-16) is higher than any species assigned to Trimerus (ranging 6-12), the width of the axis (up to 0.7 width) is greater than in any species assigned to Trimerus (ranging 0.3-0.6 width), the pygidium is longer (length/width ratio 1 . 1-1 .2) than all species of Trimerus (length/width ranging 0. 6-1.0) except for T. ( Trimerus ) cylindricus Salter (1865), and the cephalic pro- portions are longer (“ transverse width slightly exceeding length Edgecombe and Fortey, 2000: 332) than any species of Trimerus (length/width ranging 0.6-0.7), complementing its elongate pygidial proportions. These features are more com- patible with assignment to Burmeisteria. The type species, B. herschelii exhibits comparably high pygidial segmentation (16-17 axial rings, 9-11 pleural ribs), elongate pygidial pro- portions and a wide pygidial axis. Although the course of the rostral suture and the morphology of the rostral plate (including the presence of a ventral process) for B. linares remains uncer- tain, there is a suggestion of moderate convexity of the rostral plate in one of the cranidia illustrated by Edgecombe and Fortey (2000: pi. 2 fig. 16). In addition to the several species reassigned to Burmeisteria and Dig onus (discussed above), a considerable number of species assigned to Trimerus by Tomczykowa (1975) and Wenndorf (1990) are assigned to other genera. T. lilydalensis is assigned below to Wenndorfia. Pygidia from the upper Pragian of Morocco, identified by Schraut (2000) as T. sp. cf. crassi- cauda are too fragmentary for confident assignment to Trimerus , although the shallower axial ring and pleural furrows preclude assignment to the German species. The relationships of a number of poorly documented taxa previously assigned to Trimerus are uncertain, including Homalonotus acuminatus Tromelin and Lebesconte, 1876 and the French H. leheri Barrois, 1886. Trimerus ( Trimerus ) Green, 1832 Type species. Trimerus delphinocephalus Green, 1832, from the Wenlock Rochester Shale, New York State, by monotypy. Other species included. Homalonotus ( Trimerus ) cylindricus Salter, 1865, T. (T.) flexuosus Benedetto and Martel (in Baldis et al., 1976), H. harrisoni McCoy, 1876, H. (T.) johannis Salter, 1865, H. vomer Chapman, 1912, T. (T.) sp. A in Alberti (1970), T. sp. in Owens (1994). Range. Wenlock-Ludlow, Pfldoli? Diagnosis. Glabella moderately to strongly expanded across LI -LI (tr.), strongly raised, markedly narrower (tr.) anteriorly than posteriorly (sides tapering at 33-55°, width opposite S3 less than 0.65 times LI -LI). SI moderately to strongly expressed, S2 and S3 distinct or fused with SI to form a shal- low exsagittal depression. LI and L2 with low ovoid swellings. Sagittal ridge of glabella well defined. Preglabellar field long to very long (0.22-0.35 times cranidial length), distinctly concave (tr. sect.). Pygidial length equal to or greater than width. Pygidial ring furrows and pleural furrows of similar depth, 10-12 axial rings, 7-9 ribs. Pygidium with acutely produced tip or with short, spine-like process. Discussion. The type species Trimerus delphinocephalus belongs to a distinct species group within Trimerus, restricted in distribution to the mid Wenlock-lower Ludlow interval. The group is easily recognised in having distinctive glabellar features, particularly the glabella with an outline markedly expanded across Ll-Ll (tr.), raised profile, distinct lobation, swellings on LI and L2, and the sagittal ridge. In comparison to other members of Trimerus, these species also have more elongate preglabellar fields, more elongate and more highly segmented pygidia, and bear a short spine or acute process on the pygidial terminus. The species assigned are easily distin- guished from other Trimerus, in which the characters listed are only occasionally or weakly expressed. The characteristic cephalic features of Trimerus ( Trimerus ) are expressed most strongly in T. (T.) johannis from the upper Wenlock of England and T. (T.) vomer from the basal Ludlow in Victoria. These species are interpreted as the most highly derived of those assigned. Populations of T. (T.) delphinocephalus show wide morphological variability and the cephalic features characteristic of the subgenus are variably expressed, although they are clearly present in specimens fig- ured by Whittington (1993: figs 1, 2). Pygidia of delphino- cephalus differ from those of johannis and vomer in having shallower ring furrows and pleural furrows, and a short spine- like process posteriorly. In outline, the depth of the pygidial furrowing and presence of a posterior spine a single pygidium documented from the Upper Silurian of Morocco (Alberti, 1970, Schraut, 2000) closely resembles pygidia of delphinocephalus and is also assigned to the subgenus. Trimerus ( Trimerus ) flexuosus from the Wenlock-Ludlow? of Argentina is poorly known but is clearly attributable to T. (Trimerus). Benedetto and Martel (in Baldis et al., 1976) describe the species as having a raised glabella with distinct SI -S3, subcircular-ovate swellings on LI, and strongly defined paraglabellar areas, features characteristic of the subgenus. The pygidium closely resembles that of T. (T.) sp. (in Alberti, 1970) from Morocco in outline, in the depth of the pygidial furrowing and in the presence of a posterior spine. The rostral suture is convex, as in the type species. An incomplete cranidium figured as Trimerus sp. from the upper Wenlock Brinkmarsh Beds of England (Owens, 1994: pi. 2 fig. J) can be assigned to T. (Trimerus). In outline and height the glabella is not unlike that of T. (T.) johannis from the nearby area, to which it may belong. The anterior margin of the cephalon is variably expressed in Trimerus (Trimerus). In T. (T.) vomer, the angle of convergence of the sides of the cephalon increases abruptly opposite the antero-lateral comers of the cranidium, giving the cephalon an irregular pentagonal outline. In T. (T.) johannis (and possibly T. (T.) flexuosus) the anterior margin of the cephalon is tri- cuspate, with the sides of the cephalon invaginated opposite the connective suture. The weakly trilobate anterior margin of the cephalon of T. (T.) delphinocephalus is intermediate between these morphologies. Trimerus (Trimerus) harrisoni (McCoy, 1876) Figures 5B, 19.7, 19.9, 19.11 Homalonotus harrisoni McCoy, 1876: 19, pi. 23 fig. 11. Trimerus harrisoni. — Gill, 1949: 65, text-fig. 1A. — Sandford, 2000: 189, figs. 17A-E. Homalonotid trilobites from the Silurian and Lower Devonian Remarks. I recently provided a comprehensive synonymy, revised diagnosis and description for Trimerus ( Trimerus ) harrisoni (see Sandford, 2000). No new material has been obtained. Poor preservation of all known cranidia obscures many features of the cephalon pertinent to its subgeneric assignment. The glabella is not strongly raised, although this is true also of the type species Trimerus {Trimerus) del- phinocephalus. The glabella tapers very strongly forwards, although whether bell shaped or trapezoid in outline is not clear as the best-preserved specimen is somewhat crushed. The poor- ly known pygidia of harrisoni resemble those of the type species in the depth and relative depth of the pleural and ring furrows, although differ in having a prominent postaxial ridge and possibly in lacking a posterior spine. In respect to these latter differences, the pygidial morphology of harrisoni is not incompatible with T. ( Ramiotis ). However, the strong tapering of the glabella and exceptional length of the preglabellar field are emphasised here in assigning the species to T. {Trimerus). Trimerus {Trimerus) vomer (Chapman, 1912) Figure 15 Homalonotus vomer Chapman, 1912: 298, pi. 62 figs 2, 3, pi. 63 fig. 2, non pi. 63 fig. 1 (= Trimerus {Ramiotis) otisi sp. nov.) Homalonotus {Trimerus) vomer . — Reed, 1918: 323. Trimerus vomer . — Gill, 1949: 65, text-fig. 1C non IB {=Trimerus {Ramiotis) otisi sp. nov.). — Tomczykowa, 1975: 11. — Wenndorf, 1990: 16.— Schraut, 2000: 383. Type material. Holotype NMV PI 2301 (cephalon and displaced thorax, figured Chapman, 1912, pi. 62 fig. 1, 2, Gill, 1949a, text-fig. 1C, Fig. 15.4 herein) from “Wandong”, Victoria. Paratype NMV P12302 (thoracopygon, figured Chapman, 1912, pi. 63 fig. 2 from “Wandong”. Paratype NMV P12303 (cephalon, figured Chapman, 1912, pi. 63 fig. 1, Gill, 1949a, text-fig. IB) from “Wandong”. For localities see Fig. 11. Registered material. 55 specimens. NMV P455, NMV P529, NMV P16415, NMV P136623-P136627, NMV P136629-P136633, NMV P137158, NMV P137291, NMV P137910 from “Wandong”. NMV P 1 366 1 2-P 1 36622 , NMV P136628, NMV P136680, NMV P137120, P137121, NMV P137126-P137137 from PL286, Williams locality F22, Kilmore East, Victoria. NMV P136634-P 136636 from “Kilmore East-Sunday Creek Road”, Kilmore East. NMV P136637 from “Langford Park”, Kilmore East. NMV PI 36846, NMV PI 37887 from “Kilmore”, Victoria. NMV P137122-P137124, NMV P137144, NMV PI 38649 from “Kilmore- Wandong”. NMV P137144 from PL868, Kilmore East. Yan Yean Formation. NMV P136638 from “Heathcote”, Victoria. For localities see Fig. 11. Revised diagnosis. Cranidial width 1.5 times length. Glabella strongly raised, length equal to width, strongly bell shaped in outline, much narrower anteriorly than posteriorly, anterior margin well defined, with deep medial indentation. SI -S3 dis- tinct, SI with deeply impressed apodeme at adaxial end. Paraglabellar area very strongly defined. Preglabellar field very long, 0.3 times cranidial length, weakly concave (tr. sect.). Palpebral lobes placed opposite 0.5 cranidial length (0.7 glabel- lar length). Weak but distinct eye ridges. Anterior branches of the facial suture straight, converging at about 55° to a point opposite midlength of preglabellar field. Rostral suture trans- verse. Dorsal section of rostral plate triangular, long, 0.1 35 cephalic length. Ventral section of rostral plate elongate, width 0.85 times length, connective sutures straight and converging posteriorly at 35°. Pygidium triangular, length equalling width, with broad based, acutely pointed tip. Axial furrows moderate- ly to deeply impressed. Axial width (measured across second ring) 0.37 times pygidial width. Axis raised, markedly swollen posteriorly. Wide postaxial ridge. 10-12 axial rings. 8 pleural ribs with nineth rib reduced or absent. Pygidial pleural and ring furrows deep, of equal depth. Rib-ring medial offset at fifth rib. Description. Exoskeleton maximum length estimated 12 cm long. Exoskeleton granulate, more densely in the alae, finely tuberculate tho- racic segments. Cephalon triangular, length 0.8 width, in transverse and sagittal sec- tion posteriorly convex, anteriorly concave, lateral margins converging at 70-80°, genal angles and anterior tip rounded. Glabella 0.6 cephalic length, strongly bell shaped, narrowing markedly opposite palpebral area, posterior width equal to length and twice anterior width, in section (tr., sag.) weakly convex posteriorly, less so anteriorly. Occipital ring transverse, lower than glabella, 0.1 glabellar length, slightly wider at medial apex, in transverse section arched at medial apex, sloping at a uniform gradient abaxially, not interrupted by axial furrow and thus continuous with posterior border. Occipital furrow deeply impressed, strongly deflected forwards medially and less so abaxially, wider at medial apex. SI a wide shallow furrow directed inwards backwards between 0.2 and 0.4 preoccipital glabellar width, deeper at extremities. LI weakly defined as an independently convex, triangular lobe, widest abaxially, maximum exsagittal length 0.3 glabellar length. S2 and S3 weakly defined as wide shallow depres- sions not connected with axial furrows. L2 and L3 undefined. Frontal lobe impressed medially by a short (sag.) wide notch. Axial furrow developed posteriorly as large semicircular alae as long as and placed opposite LI. Anteriorly axial furrow very wide, deepening opposite antero-lateral comer of glabella. Preglabellar furrow indistinct, defined as a change in height from the frontal lobe to the frontal area. Frontal area long (sag.), 0.4 glabellar length. Medial area increasingly concave anteriorly, margins increasingly convex posteriorly in transverse section. Genae inflated. Fixigenae wide (tr.), approximately 0.5 pre- occipital glabellar width at midline. Palpebral area wide and convex (tr.), palpebral lobe inclined at 45°, semicircular, 0.25 glabellar length (sag.), placed anterior of and adjacent to glabellar transverse midline, weak eye-ridge directed inwards forwards to a point opposite S3. Anterior to palpebral area fixigenae sloping down to frontal area, not extending anteriorly as far as glabella. Posterior to palpebral lobe fixigenae wide (exsag.), 0.33 glabellar length, narrowing abaxially, strongly convex. Posterior border furrow wide, shallowing and curving forwards abaxially. Posterior border lower than genal area, increasing- ly convex (exsag.) and widening to twice occipital ring width (sag.) abaxially. Genal angle broadly rounded. Anterior facial suture con- verging forwards at about 60°, meeting connective suture very close to margin at a point opposite 2/3 sagittal length of frontal area (measured from glabella). Connective sutures short on dorsal surface, diverging forwards at 90°. Rostral suture transverse. Posterior branch of facial suture transverse adaxially, curving backwards abaxially, crossing margin just anterior of genal angle. Librigenae triangular, convex, steep. Eye small, raised above genal surface. Lateral border furrow defined as an angular change in slope, less distinct toward genal angle. Lateral border narrowing anteriorly by a factor of four from genal angle, flat and horizontal. Lateral border and furrow indistinct anterior to genae. Anterior facial sutures converging at 60°, close to margin and curving inwards anteriorly, meeting connective suture on a transverse line 0.3 sagittal length of frontal area from anterior tip. Connective suture short on dorsal surface, diverging at 90°. Ventral portion of 36 Andrew C. Sandford Homalonotid trilobites from the Silurian and Lower Devonian rostral plate flat, elongate pentangular, long (sag.), 0.4 cranidial length (extending to anterior margin of glabella), maximum width 0.8 length, placed at 0.75 rostral length. Connective sutures straight, converging posteriorly at 35°, posterior margin transverse, 0.4 maximum width. Dorsal portion of rostral plate triangular to weakly crescentic, flat, very short (sag.) length 0.2 width between posterolateral corners and 0.2 ventral length. Anterior margin straight abaxially, each side converging at 110°, medially rounded. Rostral suture convergent at 170° laterally (to one third maximum width), transverse medially. Connective sutures short (exsag.) apparently marginal between posterolateral comers of dorsal portion and anterolateral corners of ventral portion. Doublure posterior margin transverse, as wide as rostral plate adjacent to connective suture but narrowing adaxially. Adjacent to lateral margin doublure concave in section such that it closely parallels the dorsal exoskeleton. Thorax 1 1 segments. Axis cigar shaped, weakly convex (tr.), wide, 0.7 (max.) thoracic width in dorsal view. Pleural field narrow in dorsal view, 0.15 thoracic width, convex (tr.), flexed to subvertical, twice as long abaxial of flexure than adaxial. Axial rings convex (sag.), trans- verse, wider medially than abaxially. Pleural furrow continues onto rings, deep slot-like, defining anterior band wider abaxially. Axis dis- tinguished from pleural field by independent convexity and by a depressed area in the posterior edge of the pleurae. Pleural furrow divides narrow anterior band from posterior band twice as wide, shallows towards tip. Abaxially of flexure pleurae widen slightly, almost wholly occupied by articulating facet, which has a rounded tip. Pygidial proportions as for cephalon. Trilobation distinct. Anteriorly axis as wide as pleurae, tapering uniformly, 0.8 pygidial length, in transverse section axis weakly convex anteriorly to strongly convex posteriorly. In sagittal section axis concave anteriorly, posteriorly with long prominent swelling and strongly concave. 9-10 distinct rings, swelling 0.25 axial length and very feintly ringed. Axial rings very weakly convex (sag.) and weakly convex-forwards. Anterior ring furrow higher than others, second ring with pseudo-articulating half- ring. Rings 1-7 of u ni form width, posterior rings (8+) as narrow ridges Posterior of axis border strongly arched (tr. sect.) and high. Axial furrow shallow and wide. Pleural field uniformly convex. Pleural fur- rows distinct. 9 ribs, wider than rings, and widening abaxially, flat (exsag.). Interpleural furrows not distinct. Border narrow, concave, widening posteriorly. Remarks. Many of the characters listed as diagnostic of Trimerus ( Trimerus ) are strongly expressed in T. (T.) vomer, including the definition of SI and LI and L2 swellings, the paraglabellar area, the length of the preglabellar field and the bell-shape of the glabella. T. ( T .) vomer also displays a number of unusual characters including distinct eye ridges and deep S 1 apodemes, and further differs from other members of Trimerus 37 {Trimerus) in having a particularly long rostral plate (dorsally), and very forwardly placed eyes. The deep SI apodemes are one of the most distinctive features of vomer. The orientation and position of these apodemes appears to be somewhat variable, however, being transverse on some specimens, diagonal on others, and sagittal on others. These differences may well be attributable to deformation, as is suggested by the correlation of glabellar proportions with apodeme orientation. Specimens with tectonically shortened glabellae (L:W~0.8) have trans- verse apodemes, those with tectonically elongated glabellae (L:W~1.1) have sagittally directed apodemes. The prede- formational orientation was probably diagonal, as in the holotype. Although with many unique characters, Trimerus ( Trimerus ) vomer is most closely comparable to T. (T.) johannis from the Wenlock of Wales. These species share a transverse rostral suture, a markedly raised and very strongly expanded (tr.) glabella (more exaggerated in vomer), very distinct and large paraglabellar areas and a distinctly concave preglabellar field. The deep ring furrows and pleural furrows, the strong swellings on the pygidial axis posteriorly (more prominent on vomer) and the acute process defining the pygidial tip distinguish vomer and johannis from other species of T. {Trimerus). These two species differ in that vomer has more forwardly placed eyes, distinct eye ridges, a longer preglabellar field, an entire cephalic anterior margin, deep apodemes on SI and distinct S2-S3 (fused to define a shallow exsagittal furrow in johannis). As noted by Chapman (1912), Trimerus {Trimerus) vomer is also close to the type species, T. {T.) delphinocephalus. Feint eye ridges and very weak LI apodemes have been observed (Whittington, 1993, fig. 2a) on delphinocephalus, and are comparable to those on vomer (Figs 15.3-15.4) The most con- spicuous differences include the much wider pygidial axis, shallower pygidial furrows, the convex rostral suture, the angular (rather than straight) connective sutures, the low glabella and and the less elongate pygidial outline of delphinocephalus. The cephalon designated as paratype by Chapman and con- sidered by him to be a juvenile form is not conspecific with the holotype. The specimen (NMV P12303, see Fig. 20.1) lacks the bell shaped glabellar outline and deep lateral glabellar furrows of Trimerus {Trimerus) vomer, further differing from it in hav- ing distinct subocular and sutural furrows. The specimen is assigned in this work to T. {Ramiotis) otisi. Aspects of Gill’s Figure 15. Trimerus {Trimerus) vomer (Chapman, 1912). la, NMV P136613, cephalon, dorsal view x 1.2 (latex cast) from PL286. lb, same, ven- tral view (doublure) x 2.45. 2, NMV P136619, cranidium, dorsal view x 1.0 (internal mould) from PL286. 3a, NMV P136622, cephalon, dorsal view x 1.15 (internal mould) from PL286. 3b, same, oblique view x 1.35. 3c, same, lateral view x 1.35. 4a, holotype NMV P12301, cephalon and displaced, incomplete and partly enrolled thorax, anterior view of cephalon x 0.95 (internal mould) from “Wandong”. 4b, same, dorsal view of cephalon and thorax x 0.88. 5, NMV P138649, cranidium, dorsal view x 0.92 (internal mould) from “Kilmore- Wandong”. 6, NMV P136638, cephalon, dorsal view x 0.85 (internal mould) from “Heathcote”. 7, NMV P16415, cephalon and incomplete thorax, dorsal view x 1.0 (internal mould) from “Wandong”. 8, paratype NMV P12302, thoracopygon, dorsal view x 0.92 (internal mould) from “Wandong”. 9a, NMV P136625, pygidium, dorsal view x 0.85 (internal mould) from “Wandong”. 9b, same, lateral view x 1.05. 9c, same, posterior view x 0.85. 10, NMV P137291, pygidium, dorsal view x 1.1 (internal mould) from “Wandong”. 11, NMV P137158, pygidium, dorsal view x 1.5 (internal mould) from “Wandong”. 12, NMV P137124, pygidium, dorsal view x 1.3 (internal mould) from “Kilmore- Wandong”. 13a, NMV P455, thoracopygon, Odorsal view x 1.1 (internal mould) from “Wandong”. 13b, same, lateral view x 1.2. 14, NMV P136637, pygidium, dorsal view x 1.05 (internal mould) from “Langford Park homestead, Kilmore East”. 15, NMV P136633, pygidium, dorsal view x 1.25 (internal mould) from “Wandong”. 16, NMV PI 379 10, pygidium, dorsal view x 1.0 (internal mould) from “Wandong”. 38 Andrew C. Sandford (1949) redescription of vomer are clearly based on this paratype, notably the description of the axial and glabellar furrows as shallow and faint (respectively). Documenting an increasingly distal eye position in a suc- cession of Victorian homalonotids, Gill suggested a migration of the Trimerus eye through time. In addition to the problem that that the width of the librigenae and eye position for T. ( Trimerus ) vomer is measured from the paratype cephalon of T. (. Ramiotis ) otisi, Gill’s model is flawed firstly, in that only two of the species in Gill’s model can be assigned to Trimerus, secondly, in that the relative stratigraphic positions of the species in Gill’s model are incorrect. Environmental notes. Although no complete exoskeletons are known from the population of Trimerus ( Trimerus ) vomer, the articulation ratio of the population as a whole is high (Ar=28%). T. (T.) vomer occurs in a fine siltstone lithology at all of the few localities from which it is known. The holotype is a cephalon displaced from an articulated thorax (Fig. 15.4), interpreted as a moult assemblage. As for T. ( Edgillia ) king- lakensis, articulated thoracopygons, cephalothoraxes and partly disarticulated thoraxes are interpreted as exuvial configurations (taphofacies TIV). An outer shelf deep-water habitat below maximum storm wave base is suggested for vomer. Trimerus ( Edgillia ) subgen. nov. Type species. Trimerus kinglakensis Gill, 1949 from the Lochkovian Hume vale Siltstone, central Victoria. Other species included. Trimerus ( Trimerus ) grandis Benedetto and Martel (in Baldis et al., 1976), T. ( Edgillia ) jelli sp. nov., Homalonotus (Trimerus) mongolicus Tchemycheva, 1937, H. vanuxemi Hall, 1859. Other species tentatively included. Homalonotus major Whitfield, 1881. Range. Upper Silurian-Lower Devonian. Derivation of name. For the late E. D. Gill, for his contribution to Victorian palaeontology. Diagnosis. Glabella subquadrate, length about 1.0 times width, sides straight and weakly tapered (converging at around 15-20°), anterior margin transverse, well defined. S2 and S3 indistinct, SI indistinct or weakly defined, not extending to axial furrows. Eye placed more or less opposite glabellar midlength. Hypostomal middle furrow with deep impressions abaxially. 10-14 pygidial axial rings. Pygidial postaxial ridge poorly defined. Ring furrows deep, pleural furrows shallow. Pygidial length 0.85-1.0 times width. Discussion. Represented by a population sample of over 300 partly articulated exoskeletons, Trimerus (Edgillia) king- lakensis is undoubtedly the best represented species of Trimerus from the Prfdolf-Lower Devonian interval. T. (E.) kinglakensis differs markedly from many earlier represen- tatives of Trimerus in its glabellar morphology, principally in having a subquadrate outline and a raised but evenly-vaulted surface. Species assigned here to T. (Ramiotis) differs in having a more strongly tapered and generally more elongate glabellar outline, with many exhibiting weak glabellar lobation and rounded anterior glabellar margins and pygidia with raised postaxial ridges. The glabellar outline strongly expanded across Ll-Ll (tr.), distinct glabellar lobation, paraglabellar areas, and the presence of a glabellar longitudinal ridge and glabellar antero-median indentation in T. (Trimerus) distinguishes kinglakensis from that group. The morphology exhibited by kinglakensis is shared with several poorly known Upper Silurian-Lower Devonian taxa and defines a species group for which T. (Edgillia) is established. Homalonotus vanuxemi from North America is roughly con- temporary with Trimerus (Edgillia) kinglakensis, occurring in Lower Devonian strata of New York State. The species was originally described from its thorax and pygidium (Hall, 1859). The later documentation of the cranidium (Hall and Clarke, 1888: p. 11, plate 5b, Williams and Breger, 1916: plate 22, figs 12, 21, Shimer and Shrock, 1944: plate 272, fig. 32) demon- strates a cephalic morphology closely comparable with king- lakensis. In particular, the straight sides, transverse anterior margin and weakly tapered outline and proportions of the glabella of vanuxemi are almost identical to the Victorian species. With respect to the morphology of the preglabellar field, the illustrations of the cranidium by Hall and Clarke and Shimer and Shrock are not in close agreement. Hall and Clarke noted an exceptionally long preglabellar field and a transverse rostral suture, although their observations were based on a single fragment. Shimer and Shrock illustrated a complete cranidium with a somewhat shorter preglabellar field, and although the rostral suture is not clearly depicted it appears forwardly convex in outline. In these respects the cephalon illustrated by Shimer and Shrock is very close to kinglakensis and supports assignment of vanuxemi to T. (Edgillia). The pygidia of these two species also share a relatively high degree of segmentation and both lack a raised postaxial ridge. The rel- ative depth of the pygidial pleural furrows of vanuxemi is diffi- cult to ascertain from the illustrations available. However, Hall and Clarke’s illustration of a large specimen in lateral view (Hall and Clarke: pi. 5b fig. 2) suggests shallower pleural furrows than the ring furrows, as in kinglakensis. Trimerus (Trimerus) grandis from the Upper Silurian-Lower Devonian of Argentina was described by Benedetto and Martel in Baldis et al. (1976) from a single specimen. Assignment of this species to T. (Edgillia) is supported by the raised, evenly vaulted glabellar profile, the subquadrate glabellar outline, the poorly defined glabellar lobation, and the relative shallowness of the pygidial pleural furrows relative to the ring furrows. The higher degree of pygidial segmentation (11-12 axial rings) is in accord with assignment to T. (Edgillia). A new species represented almost entirely by fragmented tergites from the Lochkovian of Victoria can also be assigned to Trimerus (Edgillia). The material available is closely compar- able to T. (E.) kinglakensis and T. (E.) vanuxemi and is described in this work as T. (E.) jelli. There are few other homalonotids recorded from the Lower Devonian that can be confidently assigned to T. (Edgillia). Homalonotus major was described from a single, incomplete but very large thoracopy- gon. Whitfield (1881) estimated the specimen to represent an individual 15.5 inches (31 cm) and in its very large size the species is comparable to kinglakensis, vanuxemi and T. (E.) grandis. The depth of the axial ring and pleural furrows of Homalonotid trilobites from the Silurian and Lower Devonian 39 major and kinglakensis are comparable, although those of major do not appear to shallow abaxially. The specimen has 10 axial rings, although the posterior part of the pygidium is broken off and whether there were further rings or a postaxial ridge cannot be determined. On these few features the species is only doubtfully assigned to Edgillia. Gill (1949) considered Trimerus C Edgillia ) kinglakensis to be most closely related to, and directly descended from, the earliest Ludlow T. (T.) vomer. Gill constructed his phylogeny erroneously considering these species to be much closer in age than they actually are. Redescription of both vomer and kinglakensis shows that these species are not closely related. T. (T.) vomer is a highly derived species and kinglakensis exhibits none of its characteristic features. With respect to the presumably ancestral T. ( Ramiotis ), morphoclines of increasing glabellar complexity defining T. ( Trimerus ) are contrary to morphoclines of decreasing glabellar complexity characterising T. ( Edgillia ) and suggest independent derivation of these subgenera. Trimerus {Edgillia) kinglakensis (Gill, 1949) Figures 5A, 5D, 16, 17.1-17.7 Homalonotus sp. — Gill, 1947: 13. Trimerus kinglakensis Gill, 1949: 67, pi. 8 figs 1-3, pi. 9 figs 3, 5-6, text-fig. IE. — Williams, 1964: table 1. — Wenndorf, 1990: 16. — Jell, 1992: 11, text-fig. IB .— Schraut, 2000: 382. Type material. Holotype NMV P14580 (cephalon, figured Gill, 1949: pi. 8 figs 1, 2, pi. 9 fig. 3, text-fig. IE, Fig. 16.12 herein, counterpart previously registered NMV P14581) from PL252, Williams locality W3, Middendorps Quarry, Kinglake West, Victoria. Paratype NMV P14582 (thoracopygon, figured Gill, 1949: pi. 8 fig. 3, counterpart previously registered NMV P14583) from PL252. For locality see Fig. 11. Previously jigured material. NMV P14584 (thoracopygon, figured Gill, 1949: pi. 9 fig. 5), NMV P14585 (pygidium, figured Gill, 1949: pi. 9 fig.6, counterpart previously registered NMV P14586) from PL252. Registered material. 307 specimens: 2 dorsal exoskeletons, 22 articu- lated thorax-pygidia with displaced cephala, 29 articulated thoraxes with displaced cephala, 12 articulated thorax-pygidia with inverted cephala, 3 articulated thoraxes with inverted cephala, 3 articulated thoraxes with displaced cephala and pygidia, 110 articulated or partly disarticulated thoraxes with attached pygidia, 10 articulated or partly disarticulated thoraxes with displaced pygidia, 32 cephala, 4 cranidia, 1 librigena, 30 articulated or partly disarticulated thoraxes, 50 pygidia. NMV P149634, P149635, NMV P305248-P305545 from PL252. NMV P305546 from “Kinglake”, Victoria. NMV P305547 from “King Parrot Creek”, Kinglake West. NMV P304849 from PL6640, Garratt locality M8, Heathcote Junction. NMV P304850 from PL6641, Williams locality G96, Clonbinane. NMV P304853, P304854? from PL6645, Williams locality W47, Humevale. NMV P304852? from PL6648, Humevale. NMV P304857? from PL6659, Williams locality W8, Kinglake West. NMV P304856 from PL6660, Humevale. Unregistered specimen from PL229, Williams locality W7, Collins Quarry, Kinglake West. For localities see Fig. 11. NMV P305548-P305550 from “Christmas Hills”, Victoria (probably in the vicinity of PL261). For locality see Jell and Holloway, 1983: fig. 1. NMV P305551, P305552 from old Scoresby Brick Pit, Wantirna South, Victoria. Stratigraphic distribution. Humevale Siltstone (from basal horizons at the old Scoresby Brick Pit to the Collins Quarry Sandstone Member at the top of the unit), upper Notoparmella plentiensis-Boucotia janae- Boucotia australis assemblage zones, Pfidoli -Lochkovian. Revised diagnosis. Cranidial width 1.5 times length. Glabella trapezoid, length more or less equal to width, sides straight, converging at about 15°, anterior margin well defined. Proximal section of SI moderately impressed, directed diag- onally, glabellar lobation otherwise poorly defined. Preglabellar field 0.2 times cranidial length. Palpebral lobes placed opposite glabellar midlength/0.4 cranidial length and remote (5-6 1.55 times preoccipital glabellar width). Anterior branches of facial suture angular, parallel to axial furrows for a short distance posteriorly, converging at about 80° anteriorly. Librigenal border furrow weakly defined. Course of rostral suture angular, defining an obtuse angle of about 140°, dorsal section of rostral plate boomerang-shaped. Hypostome with middle furrow very deep, posterior border with wide, short triangular lobes (length 0. 1 times hypostomal length) and deep, angular medial notch. Pygidium triangular, length 0.95 times width, with long pointed tip, sides straight, converging at about 75°. Axial furrows moderately impressed opposite first ring, deep posteriorly. Axial width 0.45 times pygidial width. 13-14 axial rings. 7 pleural ribs, rib-ring medially offset at third rib, pleural furrows markedly shallowing abaxially. Dorsal exoskeleton very strongly pitted. Description. Exoskeleton large, maximum size estimated 25 cm (from NMV P305258), occipital/pygidial convexity (tr.) moderate. Pygidial and librigenal doublure and hypostome finely ornamented with short ridges. Cephalon with rounded triangular outline with weakly convex sides, converging forwards at about 70° opposite genal field and at 115° opposite rostral plate, length 0.66 times width. Glabellar length about equal to width (although difficult to quantify due to the poor definition of the axial furrows posteriorly), width about 0.5 times cranidial width, sides straight to weakly convex, length 0.8 times cra- nidial length, anterior margin transverse, medial indentation moder- ately impressed to absent. Glabellar lobation very weakly defined, generally only the proximal part of SI distinct, S2 distinct only on smallest specimens (e.g. NMV P305272, see Fig. 16.10). S3 indistinct. Occipital ring 0.1 times cranidial length, slightly wider medially. Occipital furrow deep with weak forward flexure medially. LI 0.35 times glabellar length. SI strongly convex, weakly impressed to indis- tinct adjacent to axial furrow, moderately impressed and directed diagonally (posteromedially) abaxially. S2 a very weak depression adjacent to axial furrow or indistinct, rarely extending adaxially, trans- verse. Axial furrows shallow to indistinct and directed obliquely oppo- site paraglabellar area and across occipital furrow, deep anteriorly. Paraglabellar area weakly defined. Preglabellar field flat (tr. sect.). Length (exsag.) of posterior border equal to occipital length adaxially, lengthening to 1.5 times occipital length abaxially. Posterior border furrow transverse, very wide, moderately impressed, terminating dis- tally. Postocular fixigenal area long, length (exsag.) 0.21 times cran- idial length. Palpebral lobe 0.14 times cranidial length, 6-6 1.64 times glabellar width, palpebral furrow indistinct. Preocular fixigenal area of moderate width, 0.17 times 6-5, narrowing markedly anteriorly, eye ridges not distinct. Anterior branches of facial suture meeting connec- tive sutures opposite 0.9 cephalic length. Librigena with wide, moder- ately impressed border furrow anterior of eye, not defined opposite preglabellar field, librigenal field weakly convex, steeply inclined. Homalonotid trilobites from the Silurian and Lower Devonian Dorsal section of connective sutures diverging at about 90°. Dorsal surface of rostral plate length (sag.) 0.5 times cranidial length, very weakly concave (tr. sect.). Rostral suture moderately convex. Ventral surface of rostral plate kite-shaped, with length 0.9 times width, posterior width 0. 1 maximum times width. Connective sutures angular, anterior section converging posteriorly at about 30°, posterior section converging posteriorly at about 70°. Librigenal doublure with wide, shallow vincular furrow, indistinct anteriorly. Hypostomal suture broadly concave. Hypostome with length 0.75 times width, anterior wings large, width 0.35 times hypostomal width. Thorax with thirteen segments. Axial furrows defined by shallow to moderately impressed, diagonally directed furrows on each segment, each meeting posterior margin at indentation that correspond to deep pits (axial articulating process) on internal mould. Pleural furrows wide and deep, pleural tips rounded. Pygidium triangular, with sides weakly concave. Pygidial axis with sides straight and tapering at about 30°. Axis reaching 0.85 times pygidial length. Terminal piece parabolic in outline, length 0.08 times axial length, posterior margin moderately to weakly defined. Postaxial ridge not defined. Ring furrows short (sag.) and deep. Pleural furrows wide and deep adjacent to axial furrows, shallowing markedly abaxial- ly, reaching wide and very shallow border furrow. Interpleural furrows very weakly defined. Border weakly defined. Pygidial doublure flat and weakly inclined posteriorly, laterally with angular profile (tr. section), distal surface horizontal, proximal surface increasingly inclined anteriorly. In posterior view anterior margin of pygidium moderately convex, posterior margin without distinct medial arch. Remarks. Gill’s (1949) original description is not detailed but documents most of the characters listed in the revised diag- nosis. Gill described the glabellar furrows as absent, but the proximal portion of SI is manifest as a short, moderately impressed diagonal furrow. Trimerus ( Edgillia ) kinglakensis can be closely compared to the poorly known T. (E) vanuxemi from the Helderbergian of North America. In addition to the shared cephalic and pygidial features listed above as defining T. {Edgillia), the species share a forwardly convex rostral suture, a similar pygidial rib-ring offset (third-fourth? rib). The species differ in that vanuxemi exhibits longer glabellar proportions, shorter cephalic and pygidial proportions, a greater number of pleural ribs, a wider pygidial axis, and lacks a produced pygidial tip. In the depth of the pygidial axial furrows and ring and pleural furrows, kinglakensis shows closer resemblance to T. (E?) major , also from the Helderbergian of North America. However, major differs from kinglakensis in that the pleural furrows do not 41 shallow abaxially. Pygidial morphology of the Late Silurian T. (E.) mongolicus is closely comparable to that of kinglaken- sis. The Mongolian species exhibits similar proportions, 13 pygidial axial rings, deep ring furrows and shallow pygidial pleural furrows and an acuminate tip. T. (E.) kinglakensis differs in having deeper pygidial axial furrows, a shorter preglabellar field, and in the strong convexity of the rostral suture. Although the course of the rostral suture has been emphasised in the diagnoses of higher taxa (e.g. medial flexure in Burmeisteria, concavity in Digonus: see discussion above), variability in the forward curvature of the rostral suture amongst Trimerus is considered only of specific significance. Apart from kinglakensis and vanuxemi, few species of Trimerus exhibit strongly forwardly convex rostral sutures, the most notable exceptions being the type species T. (T.) del- phinocephalus and T. (T) flexuosus. Other features of these latter species are considered characteristic of T. ( Trimerus ) and T. {Ramiotis) respectively, and indicate that they are not closely related to kinglakensis. Environmental notes. See discussion on taphonomy and bio- facies. Trimerus {Edgillia) kinglakensis is considered to have inhabited mid- to outer-shelf environments. Trimerus ( Edgillia ) jelli sp. nov. Figure 17.8-17.19 Homalonotidae gen. et sp. indet. 3 — Holloway and Neil, 1982: 146, fig. 4M-R. Type material. Holotype NMV P82873 (pygidium) from PL2288, Thomas locality F25, Parish of Redcastle, Heathcote, Victoria (Fig. 17.9). Paratypes NMV P78300 (ex GSV48930, figured Holloway and Neil, 1982: fig. 4R), NMV P304610 (cranidia), NMV P304612 (thoracic segment), NMV P78298 (pygidium, ex GSV48942, figured Holloway and Neil, 1982: figs 4M, 4N, 4Q) from PL2319, Thomas locality F55, Parish of Redcastle, Heathcote. Paratypes NMV P304638 (cranidium), NMV P304633-4 (librigenae), NMV P304635 (hypos- tome), NMV P304636, P304637 (thoracic segments), NMV P304639 (pygidium) from PL6652, Heathcote. Paratype NMV P78299 (libri- gena, figured Holloway and Neil, 1982: figs 40, 4P) from PL2288. For localities see Fig. 8. Registered material. 45 specimens: 6 cranidia, 3 librigenae, 1 hypos- tome, 25 thoracic segments, 8 pygidia. NMV P82925 (ex GSV48928), NMV P3 046 1 0-P3 0462 5 , NMV P304626 (ex GSV48933), GSV48944, GSV48938 from PL2319. NMV P82873, NMV P304628-P304632 from PL2288. NMV P304633-P304647 from PL6652. Unregistered specimen from PL2328, Heathcote. Figure 16. Trimerus {Edgillia) kinglakensis (Gill, 1949). 1, NMV P305261, cranidium, dorsal view x 0.6 (internal mould) from PL252. 2a, NMV P305257, cephalon and hypostome, dorsal view of cephalon x 0.75 (latex cast) from PL252. 2b, same, dorsal view of hypostome x 1.6. 2c, same, ventral view of cephalon (doublure) and hypostome x 1.4. 3, NMV P305256, cephalothorax, dorsal view of cephalon x 0.85 (latex cast) from PL252. 4, NMV P305262, cephalon, dorsal view x 0.85 (latex cast) from PL252. 5, NMV P305265, cephalon, hypostome and displaced thora- copygon, dorsal view of hypostome x 1.55 (latex cast) from PL252. 6a, NMV P305250, cephalon, dorsal view x 0.95 (latex cast) from PL252. 6b, same, lateral view x 1.0. 7a, NMV P305264, pygidium, posterodorsal view x 0.85 (latex cast) from PL252. 7b, same, lateral view x 1.4. 8, NMV P305248, cephalon, ventral view (doublure) x 1.0 (latex cast) from PL252. 9, NMV P305252, cephalon, oblique view x 0.7 (internal mould) from PL252. 10, NMV P305272, cephalon, dorsal view x 3.1 (internal mould) from PL252. 11, NMV P305260, thoracopygon, dorsal view of pygidium x 0.7 (latex cast) from PL252. 12, holotype NMV P14580, cephalon, dorsal view x 0.75 (internal mould) from PL252. 13, NMV P305258, thoracopygon, dorsal view x 0.6 (latex cast) from PL252. 14a, NMV P305263, pygidium, dorsal view x 1.1 (latex cast) from PL252. 14b, same, x 1.0 (internal mould). 15, NMV P305253, cephalon, ventral view (doublure) x 1.6 (latex cast) from PL252. 16, NMV P305254, cephalon, dorsal view, enlargement of anterior margin x 2.9 (latex cast) from PL252. 42 Andrew C. Sandford Homalonotid trilobites from the Silurian and Lower Devonian 43 Stratigraphic distribution. Stoddart Member, Mt Ida Formation, 200 m to 500 m above the base of the unit, Boucotia australis Assemblage Zone, mid-late Lochkovian. Derivation of name. For Peter Jell (Queensland Museum), for his contribution to Victorian palaeontology. Diagnosis. Glabella with very weakly defined lobation, sides straight and converging at about 15°, anterior margin well defined, transverse, with broad medial depression. Axial fur- rows poorly defined posteriorly. Preglabellar field long, length 0.45 times anterior width of glabella (estimated 0.25 times cranidial length). Anterior branches of facial suture converging at 40°. Anterior margin of cranidium (including rostral suture) transverse. Palpebral lobe placed with midline opposite (estimated) 0.35 cranidial length. Hypostome with large, broad- based rounded lobes on the posterior margin. Thoracic pleurae with bilobed tips. Pygidium triangular, length equalling width, sides straight, converging at 60°. Axis with 10 rings, width 0.46 times pygidial width, length 0.8 (estimated) times pygidial length. Axial furrows very shallow. 6 pleural ribs. Pleural off- set at third rib. Ring furrows moderately impressed, pleural furrows very shallow to indistinct, increasingly so posteriorly. Discussion. Although most specimens assigned to this species are fragmentary, the morphology is distinctive enough to warrant description. The weakly convergent glabellar sides, the transverse and well-defined anterior margin and the simple glabellar morphology support assignment to Edgillia. In cephalic features Trimerus { Edgillia ) jelli differs from T. (E.) kinglakensis and T. (E.) vanuxemi in the transverse course of the rostral suture. Pygidial axial, ring and pleural furrows of vanuxemi are moderate to shallow in depth. In kinglakensis , the axial furrows and ring furrows are deep, with pleural furrows shallower, and markedly shallowing distally. The pygidium of jelli also exhibits pleural furrows shallower than the ring furrows, but is distinctive in that the pleural furrows are very shallow. The weakly defined segmentation is comparable to that of the Llandovery-lower Wenlock species of T. ( Ramiotis ), although jelli differs from these in having a more elongate pygidial outline and in lacking the raised postaxial ridge typical of T. {Ramiotis). In these respects the pygidia of jelli more closely resemble pygidia of species assigned in this work to Burmeisteria, in particular B. clarkei and B. linares. However, the strong segmentation shared by clarkei and linares suggest that similarities in the pygidial outline and depth of pygidial axial ring and pleural furrows are not of high taxo- nomic significance. The distinctive (straight-sided, weakly furrowed) pygidial morphology shared by these three taxa and others such as as Dipleura garratti suggests the morphology is homologous. Hypostomes are known for few species of Trimerus. The hypostome of T. {Edgillia) jelli (Fig. 17.12) closely resembles that of T. {E.) kinglakensis (Fig. 16.2). These both differ from the hypostome of T. (T.) delphinocephalus (Whittington, 1993: fig. IB) and T. {Ramiotis) rickardsi (Fig. 18.6) in having larger posterior border lobes and deeper middle furrows. Environmental notes. Trimerus {Edgillia) jelli is considered to have inhabited high-energy, very shallow environments. Trimerus ( Ramiotis ) subgen. nov. Type species. Trimerus ( Ramiotis ) rickardsi sp. nov., from the upper Llandovery ( crenulata Biozone) Chintin Formation, central Victoria, Australia. Other species included. Platycoryphe dyaulax Thomas, 1977, Trimerus ( Ramiotis ) iani sp. nov., T. permutus Tomczykowa, 1978 (nom. nov. for T. lobatus Tomczykowa, 1975 non Prouty, 1923), T. (R.) otisi sp. nov., T. {R.) tomczykowae sp. nov., T. (R.) thomasi sp. nov., T. sp. {in Edgecombe and Fortey, 2000), T. sp. {in Curtis and Lane, 1998), T. sp. A (in Wolfart, 1961), Other species tentatively included. Dipleura salteri Morris, 1988 (nom. nov. for Homalonotus {Koenigia) ludensis Salter (1865) non H. ludensis Murchison, 1839). Range. Silurian. Derivation of name. For my son Otis Rami. Diagnosis. Glabella trapezoid or weakly expanded across Ll- L1 (tr.), low, with sides more or less straight and weakly to moderately convergent. SI -S3 weakly impressed. Preglabellar field of short to moderate length (~0. 15-0.25 times cephalic length). Hypostomal middle furrow shallow. Pygidium short, length 0.5-1.0 times width. Postaxial ridge raised. Pygidium with moderately acute to obtusely convergent sides (>70°), tip without process or spine. Pygidial border furrow and ridge-like border present in late species. Figure 17.1-17.7 Trimerus {Edgillia) kinglakensis (Gill, 1949). 1, NMV P305264, thoracopygon, dorsal view x 0.65 (latex cast) from PL252. 2, NMV P14584, inverted cephalon and displaced thoracopygon, dorsal view of cephalon x 0.9 (internal mould) from PL252. 3, NMV P305276, cephalon and displaced, incomplete thorax, ventral view of cephalon (doublure) x 1.6 (latex cast) from PL252. 4, NMV P305269, incomplete thoracopygon, ventral view of pygidium (doublure) x 0.85 (latex cast) from PL252. 5, NMV P305252, cephalon, lateral view x 0.6 (internal mould) from PL252. 6, NMV P305255, cephalon and displaced thorax, dorsal view of cephalon x 0.85 (latex cast) from PL252. 7, NMV P305258, thoracopygon, dorsal view of pygidium x 0.95 (latex cast) from PL252. Figure 17.8-17.19 Trimerus {Edgillia) jelli sp. nov. 8, paratype NMV P304612, thoracic segment, dorsal view x 1.55 (internal mould) from PL2319. 9, holotype NMV P82873, pygidium, dorsal view x 1.4 (internal mould) from PL2288. 10, paratype NMV P304636, thoracic segment, lateral view x 2.0 (internal mould) from PL6652. 11, paratype NMV P304637, thoracic segment, lateral view x 1.9 (internal mould) from PL6652. 12, paratype NMV P304635, hypostome, ventral view x 2.0 (internal mould) from PL6652. 13a, paratype NMV P304610, cranidium, dorsal view x 1.3 (latex cast) from PL2319. 13b, same, x 1.35 (internal mould). 14a, paratype NMV P304634, librigena, dorsal view x 1.45 (latex cast) from PL6652. 14b, same, lateral view x 1.4. 15, paratype NMV P78300, cranidium, dorsal view x 1.1 (internal mould) from PL2319. 16a, paratype NMV P78298, pygidium, dorsal view x 1.25 (internal mould) from PL2319. 16b, same (latex cast). 16c, same, ventral view (doublure) x 1.0 (internal mould). 16d, same, lateral view x 1.0. 17, paratype NMV P304638, cranidium, dorsal view x 1.25 (internal mould) from PL6652. 18, paratype NMV P304633, librigena, ventral view (doublure) x 1.25 (latex cast) from PL6652. 19, paratype NMV P304639, pygidium, dorsal view x 5.5 (internal mould) from PL6652. 44 Andrew C. Sandford Discussion. Thomas (1977) noted that the rarity of Ashgill and Llandovery homalonotids emphasised the morphological gap between Ordovician and post-Ordovician genera. Reed (1918) suggested that links between Wenlock Trimerus and Ordovician homalonotid groups might be found in Llandovery strata. In this context Thomas interpreted his new Saudi Arabian Llandovery species Platycoryphe dyaulax as showing features of both Platycoryphe/Brongniartella and Trimerus. Thomas suggested that post-Ordovician homalonotines were derived from Platycoryphe/Brongniartella stock, rather than from Eohomalonotus as suggested by Sdzuy (1957). In assign- ing dyaulax to Platycoryphe , Thomas emphasised its wider and rounded pygidial outline, contrasting with the longer and posteriorly pointed outlines of ‘typical’ Trimerus, i.e. T. (Trimerus). Pygidia of dyaulax also exhibit shallow ring furrows and very shallow to effaced axial and pleural furrows, whereas pygidia of T. ( Trimerus ) exhibit pleural and ring furrows of moderate depth, and of equal depth. New species from the Llandovery and lower Wenlock of south-eastern Australia show strong resemblances to Platycoryphe dyaulax and define a tight species group. Similarities in cephalic morphology between dyaulax and Trimerus noted by Thomas are emphasised by the new Llandovery- Wenlock taxa, and all share weakly tapering, straight- sided, trapezoid glabellar outlines. The new south- eastern Australia taxa exhibit pygidial proportions and segmen- tation intermediate between dyaulax and T. ( Trimerus ) and support Thomas’s suggestion of close affinities between dyaulax and Trimerus. These similarities, and the weak defini- tion of thoracic trilobation, the width of the thoracic axis and the lower degree pygidial segmentation are emphasised here to justify assignment of dyaulax to Trimerus rather than to Platycoryphe or Brongniartella. T. (Ramiotis) is erected to embrace this group of species, which includes late Llandovery T. (R.) rickardsi from Victoria and T. (R.) iani from Tasmania, and the early Wenlock T. (R.) tomczykowae from Victoria. In pygidial outline and segmentation there is a morpho- logical gradient through time from Trimerus (Ramiotis) dyaulax (mid Llandovery) to T. (R.) rickardsi and T. (R.) iani (late Llandovery) with moderately impressed ring and axial furrows and shallow pleural furrows, to T. (R.) tomczykowae (early Wenlock) with pygidial furrows of subequal and moder- ate depth, and to T. (Trimerus) (Wenlock-early Ludlow). Although T. (Trimerus) appears to be derived from this group, there are a number of new or poorly known Ludlow-Pndolf Trimerus that share the distinctive cranidial morphology of the Llandovery-Wenlock T. (Ramiotis). In exhibiting elongate glabellar proportions, weakly tapered, straight-sided trapezoid glabellar outlines, low glabellar profiles and preglabellar fields of moderate length, these species are compatible with assignment to T. (Ramiotis) rather than to T. (Trimerus). These species include T. (R.) otisi and T. (R.) thomasi from Victoria, T. (R.) permutus from Poland, and T. (R.) sp. (in Wolfart, 1961) from Bolivia. Compared with the Llandovery-Wenlock T. (Ramiotis), these Ludlow-Pndolf species encompass a wider range of pygidial morphologies, particularly in length:width proportions and depth of pygidial furrows. Trimerus (Ramiotis) is easily distinguished from T. (Trimerus) in that it lacks most of the characteristic cephalic and pygidial features of the latter. Of the suite of features listed as diagnostic for T. (Trimerus), only T. (R.) permutus exhibits strongly expressed glabellar lobation. Several species exhibit sagittal glabellar ridges (albeit weak), several species exhibit distinct medial indentations of the anterior glabellar margin, and some specimens of T. (R.) otisi show a tendency towards outlines expanded across LI -LI (tr.) and strongly tapered glabellar outlines. However, there are no species of T. (Ramiotis) that exhibit a more than a few of the characters of T. (Trimerus), justifying the grouping of the latter. However, with the exception of the deep pygidial border furrow and raised border, the characters distinguishing T. (Ramiotis) from T. (Trimerus) can be considered as primitive, and hence only subgeneric status is justifiable. Dipleura salteri Morris, 1988 from the upper Ludlow (Ludfordian) of England is a poorly known species represented by a single cranidium (see Salter, 1865: p. 121, pi. 12 fig. 1). Following Salter’s (1865) suggestion, Tomczykowa (1975) and Morris (1988) assigned the species to Dipleura, but the forward eye position, the long preglabellar field and the quadrate course of the preocular facial sutures preclude assignment to that genus and are in accord with assignment to Trimerus. Subgeneric assignment of salteri is difficult as the species exhibits a very strongly tapered glabellar outline comparable to that of T. (Trimerus), but lacks the other diagnostic glabellar features of the subgenus such as the outline strongly expanded across Ll-Ll (tr.) and the distinct lobation. However, the strong tapering of the glabella may be partly attributable to sagittally oriented tectonic compression evident in the apparent con- cavity (sag.) of the preglabellar field. In all other respects the specimen is compatible with assignment to T. (Ramiotis), to which it is tentatively assigned. Trimerus (Ramiotis) dyaulax is from the Aeronian (mid- Llandovery) convolutus Biozone. A possibly earlier origin for Trimerus is suggested by a poorly known species from the lower Llandovery of Paraguay. Wolfart (1961) assigned the several cranidia representing this species to Trimerus. In having elongate glabellar proportions, weakly tapered, straight-sided trapezoid glabellar outlines, weak glabellar lobation and a preglabellar field of moderate length, these can be assigned to T. (Ramiotis). The only other Llandovery T. (Ramiotis) is rep- resented by fragmentary pygidia from the Aeronian of Wales, documented as T. sp. by Curtis and Lane (1998). Although closer in age to dyaulax, the Welsh species differs markedly from it in having moderately impressed pleural, ring and axial furrows, in this respect closely resembling T. (R.) rickardsi. Trimerus ( Ramiotis ) rickardsi sp. nov. Figure 18 Type material. Holotype NMV P305556 (cephalon) from PL6361, Springfield, Victoria (Fig. 18.1). Paratypes NMV P305553, P305554, NMV P305557, NMV P305563 (cephala), NMV P305559-P305562, NMV P305564 (cranidia), NMV P305582 (thoracic segment), NMV P305572-P305574 (pygidia) from PL6361. Paratype NMV P138204 (pygidium) from PL667, Springfield. Paratype NMV PI 38207 (pygidium) from PL599, Springfield. Paratypes NMV P147779 Homalonotid trilobites from the Silurian and Lower Devonian 45 (hypostome), NMV P147772 (pygidium) from PL256, Wallan, Victoria. For localities see Fig. 11. Registered material. 66 specimens. 5 cephala, 15 cranidia, 6 librigenae, 1 rostral plate, 3 hypostomes, 12 thoracic segments, 24 pygidia. NMV P305553-P305593 from PL6361. NMV P305626-P305628 from PL6390, Springfield. NMV P138213, NMV P147771, P147772, NMV P147774-P147777, NMV P147779-P147780, NMV P1477787 from PL256. NMV P 1 39442-P 1 39446, NMV P305595 from “Lancefield”, Victoria. NMV PI 38205-PI 38207, NMV P138270 from PL599. NMV P138204, NMV P305594 from PL667. NMV P138267, NMV P147778 from PL1964, GSV locality B25, Springfield. Lithological similarities suggests that the specimens labelled as coming from “Lancefield” almost certainly come from the Parish of Goldie, and possibly in the vicinity of allotments B19 and B23 (see Thomas, 1960: 1:31, 680 Lancefield Geological Sheet). Stratigraphic horizon. Chintin Formation, upper Llandovery. Rickards and Sandford (1998) suggested equivalence to the upper crenulata Biozone, based on the presence of Acemaspis sp. within the unit, and on the occurrence of crenulata Biozone graptolites in the uppermost beds of the underlying Springfield Sandstone. Derivation of name. For R. B. Rickards (University of Cambridge), for his contribution to Victorian palaeontology. Diagnosis. Cephalon rounded pentagonal, width about 1.6 times length, sides converging at 60° posteriorly, at 125° anter- iorly. Glabella trapezoid, length 1.2 times width, sides sub- parallel opposite paraglabellar areas, converging anteriorly at about 22°. S1-S3 moderately impressed adjacent to the axial furrow, shallow adaxially. Preglabellar field long, 0.23 times cephalic length. Palpebral lobes placed opposite 0.42 cranidial length/0.52 glabellar length. Anterior branches of facial suture straight, converging at about 45°, anterior-most section curving gently to the diagonal, continuous in curvature with broadly rounded rostral suture. Dorsal section of rostral plate cres- centic, length 0.1 times cephalic length. Ventral section of rostral plate with length 0.92 times width, connective sutures converging posteriorly at 50°. Posterior border of hypostome with short lobes of length 0.1 times hypostomal length. Pygidium wide, length 0.78 times width, with sides weakly convex, converging posteriorly at about 100°. Pygidial axis 0.42 times pygidial width, with 10 axial rings. Axial furrows straight, tapering at about 30°, shallow anteriorly, moderately impressed posteriorly. Ring furrows deep to moderately impressed, pleural furrows shallow. 7-8 ribs, rib-ring medially offset at fourth rib. Description. Exoskeleton small, maximum length 12 cm (estimated from NMV P138204), occipital and pygidial convexity (tr.) moderate. Dorsal exoskeleton and doublure finely granulose. Cranidium with length (sag.) 0.9 times cephalic length. Glabella trapezoid, length about 0.77 times cranidial length, width about 0.43 times cranidial width, anterior margin moderately well defined, broadly rounded, medial indentation variably expressed (indistinct to strong, e.g. NMV P305559). Glabellar lobation well defined. Occipital ring 0.09 times glabellar length, slightly wider medially. Occipital furrow deeply impressed with broad forward flexure medially. LI ~0.3 times glabellar length, L2 and L3 -0.1 times glabellar length and frontal lobe -0.25 times glabellar length. S 1 weakly convex forwards and directed posteromedially at about 25° from the transverse. S2 transverse. S3 directed antero-medially at about 10° from the trans- verse. Axial furrows moderately impressed opposite genae, shallow and directed diagonally opposite occipital ring. Paraglabellar area very weakly defined. Length (exsag.) of posterior border equal to occipital length adaxially, lengthening (exsag.) markedly abaxially to about twice occipital length. Posterior border furrow transverse, very wide, moderately impressed, terminating distally. Postocular fixigenal area with length (exsag.) 0.22 times cranidial length. Palpebral lobe short, 0. 1 times cranidial length, placed with 6-8 about 1 .77 times pre- occipital glabellar width/0.74 times cephalic width. Palpebral furrow wide and shallow. Preocular fixigenal area narrow (tr.), 0.17 times 8-5, uniform in width anteriorly. Preglabellar field flat (tr. sect.). Anterior branches of facial suture straight from eye to a point opposite midlength (sag.) of preglabellar field, anterior section curved. Librigenal border furrow wide and shallow opposite genal field, not defined opposite preglabellar field, border furrow indistinct on fixi- genae. Lateral border narrow and convex, narrowing posteriorly, indis- tinct opposite preglabellar field and on fixigenae. Dorsal surface of rostral plate with length 0.13 times width, anterior margin rounded. Dorsal section of connective sutures diverging forwards at about 30°. Ventral surface of rostral plate flat (tr, sag. sect.), kite-shaped, sides very weakly sinusoidal. Librigenal doublure without distinct vincular furrow. Hypostomal suture transverse. Hypostome with moderately impressed middle furrow. Thoracic segments with wide (sag.), deep articulating furrows, deep and narow (exsag.) pleural furrows. Axis very weakly defined by very shallow, wide, diagonally directed furrow. Pleural tips weakly bilobed with larger lobe opposite posterior band and smaller lobe opposite anterior band. Pygidium triangular. Pygidial axis extending to about 0.9 pygidial length, with semicircular terminal piece, length 0.1 times axial length, continuous posteriorly with raised postaxial ridge. Ring furrows deep to moderately impressed, pleural furrows shallow. Furrows markedly shallower on external casts. Pleural furrows not reaching margin. Border furrow and border poorly defined. In posterior view posterior margin of pygidium horizontal. In lateral view dorsal profile of axis moderately convex, postaxial ridge steeply inclined. Discussion. Trimerus ( Ramiotis ) rickardsi is the most com- pletely known homalonotid from the Llandovery. The species is of primary significance in the understanding of relationships between poorly known Llandovery homalonotids and better- known Wenlock and Ludlow taxa. In pygidial morphology Trimerus ( Ramiotis ) rickardsi is typical of other Llandovery- Wenlock species, particularly in having short pygidial proportions and much shallower ring furrows relative to the pleural furrows. In the latter feature rickardsi most closely resembles the very poorly known Welsh T. (R.) sp. (see Curtis and Lane, 1998), but can be distinguished by its higher segmentation (7 rings, 6 ribs in T. (R.) sp., cf. 10 rings, 8 ribs in rickardsi ). T. (R.) dyaulax differs from rick- ardsi in having even shorter proportions, a rounded pygidial tip, a low postaxial ridge, even shallower axial, ring and pleural furrows and lower segmentation (6 rings, 5 ribs) than counted for rickardsi. Pygidial morphology of T. (R.) iani from the upper Llandovery of Tasmania is poorly known. The single pygidium known exhibits segmentation comparable (9 rings, 6 ribs) to that of rickardsi, but differs in having shorter propor- tions and a rounded tip. The relative depth of the ring furrows compared to the pleural furrows distinguishes rickardsi from the Wenlock T. (R.) tomczykowae from Victoria, and from other Ludlow species of T. {Ramiotis). T. (R.) tomczykowae further differs in having a more rounded tip. Of the Upper Silurian 46 Andrew C. Sandford Homalonotid trilobites from the Silurian and Lower Devonian 47 species, T. (R.) thomasi from Victoria differs markedly in having elongate proportions (L:W~1.0), whilst pygidia of T. (R.) otisi from Victoria differ markedly in the depth of the pygidial pleural and ring furrows. Despite the pygidial differ- ences, in the more elongate glabellar outline (LAV- 1.2) the longer preglabellar field (0.25 cranidial length), and the expres- sion of glabellar lobation, rickardsi shows closest resemblance to otisi. T. (R.) rickardsi differs from most other T. ( Ramiotis ) in having a forwardly convex rostral suture, comparable to that of the Llandovery Bolivian T. (R.) sp. In having a relatively longer preglabellar field, lateral glabellar furrows of moderate length, a weak sagittal glabellar ridge and, in some specimens, a distinct medial indentation of the anterior glabellar margin, and higher pygidial segmentation Trimerus ( Ramiotis ) rickardsi shows closer affinities to T. ( Trimerus ) than other Llandovery T. {Ramiotis). However, in the absence of so many of the other derived features distin- guishing T. (Trimerus), it would be over stated to regard rickardsi as a transitional species between the subgenera. Environmental notes. Articulated specimens are only known from the type locality, and represented only by cephala. The proportion of articulated and broken specimens is 13% respec- tively, indicative of a taphofacies TII. Further east at PL256, Wallan, breakage of T. (R.) rickardsi and other trilobites is extremely high and indicates significant reworking. Associated tempestite lithologies comprising thick coquinal bioclastic lags and abundant mud rip-up clasts indicate deposition between normal wave base and storm wave base. The trilobite fauna may have been transport, possibly in a tempestite-generated mass-flow, although the profound compositional differences between the Springfield, Wallan and Lancefield faunas do not support this and suggest distinct, depth-related assemblages. Trimerus (Ramiotis) rickardsi is considered to inhabit inner shelf environments. Trimerus ( Ramiotis ) iani sp. nov. Figures 19.1-19.4, 19.6 Trimerus sp. — Holloway and Sandford, 1993: 93, figs 4C-D, 4F-G, 4I-J, 4N-P, non fig. 4L. Type material. Holotype NMV PI 37221 (cephalon lacking rostral plate, figured Holloway and Sandford, 1993: 4F-G, 4J, Fig. 19.1 here- in) from PL359, Tiger Range, northwest of Maydena, Tasmania. Paratypes NMV PI 37222 (cranidium, figured Holloway and Sandford: 41), NMV PI 37227 (pygidium, figured Holloway and Sandford: 4N-P), NMV PI 37224 (cephalon, figured Holloway and Sandford: 4C), NMV P137223 (cephalon and two displaced thoracic segments) from PL359. For locality see Holloway and Sandford: fig. 1 Previously figured material. NMV PI 37225 (librigena, figured Holloway and Sandford, 1993: 4D) from PL359. Other material. NMV P137226 (cephalon) from PL359. Stratigraphic distribution. As for Brongniartella sp. Derivation of name. For my brother Ian. Diagnosis. Cephalon rounded triangular, length -0.7 times width, sides (opposite midlength) converging at -60°. Glabella trapezoid in outline, length 1.2 times width, sides very weakly concave and tapering at 20° anteriorly, anterior glabellar margin transverse, without distinct medial indentation. SI -S3 and sagittal glabellar ridge very weakly defined. Preglabellar field of moderate length (0.2 times cranidial length), rostral suture transverse, dorsal section of rostral plate short, 0.05 times cephalic length. Eye placed opposite -0.5 cranidial length (-0.6 glabellar length). Anterior branches of facial sutures converging at 60°, curving abruptly adjacent to con- nective sutures. Ventral surface of rostral plate elongate pentangular in outline, with connective sutures converging pos- teriorly at 45°. Pygidium short. Axis 0.45 times pygidial width, raised posteriorly, 10 rings, ring furrows moderately impressed. 6 ribs, pleural furrows shallow. Axial furrows moderately impressed. Discussion. Trimerus (Ramiotis) iani is abundant at the type locality. Holloway and Sandford (1993) were uncertain whether the two dissimilar homalonotid pygidia from PL359 represented different species, or whether the differences were attributable to size differences or deformation. In reviewing this conclusion, I cannot attribute the contrast between the narrow, concave- sided and strongly raised axis of the smaller specimen (see Holloway and Sandford, 1993: fig. 4L) and the wide, convex-sided and low axis of the larger specimen entirely to deformation (see Fig. 19.6). The more elongate (length=width) and parabolic outline and the indication of a distinct border suggest that the smaller pygidium represents a separate species, referable to Brongniartella (see above). The diagnosis includes a number of features not listed in the brief description of the species by Holloway and Sandford (1993). Of the two cephala not figured by Holloway and Figure 18. Trimerus (Ramiotis) rickardsi sp. nov. la, holotype NMV P305556, cephalon, oblique view x 1.7 (internal mould) from PL6361. lb, same, dorsal view, lc, same (latex cast). 2, paratype NMV P305560, cranidium, dorsal view x 2.4 (latex cast) from PL6361. 3, paratype NMV P305561, cranidium, dorsal view x 1.4 (internal mould) from PL6361. 4, paratype NMV P305557, cephalon, dorsal view x 2.1 (internal mould) from PL6361. 5, paratype NMV P305564, cranidium, dorsal view x 2.1 (latex cast) from PL6361. 6, paratype NMV P147779, hypostome, ventral view x 2.7 (latex cast) from PL256. 7, paratype NMV P305562, cranidium, dorsal view x 1.6 (internal mould) from PL6361. 8a, paratype NMV P305559, cranidium, dorsal view x 1.5 (internal mould) from PL6361. 8b, same, x 1.6 (latex cast). 9, paratype NMV P305554, cephalon, ventral view (doublure) x 1.75 (internal mould) from PL6361. 10a, paratype NMV P305582, thoracic segment, lateral view x 2.5 (internal mould) fromPL6361. 10b, same, dorsal view x 2.2. 11, paratype NMV P305 5 63, cephalon, dorsal view x 2.4 (internal mould) fromPL6361. 12a, paratype NMV P305553, cephalon, dorsal view x 1.5 (internal mould) from PL6361. 12b, same, lateral view x 2.0. 13a, paratype NMV P305573, pygidium, dorsal view x 1.9 (latex cast) from PL6361. 13b, same, x 2.2 (internal mould). 14, paratype NMV P147772, pygidium, dorsal view x 2.0 (latex cast) from PL256. 15, paratype NMV P138204, pygidium, dorsal view x 1.4 (internal mould) from PL667. 16, paratype NMV P305574, pygidium, lateral view x 2.5 (latex cast) from PL6361. 17a, paratype NMV PI 38207, pygidium, dorsal view x 1.75 (latex cast) from PL599. 17b, same, x 1.5 (internal mould). 17c, same, posterior view x 1.75. 17d, same, lateral view (latex cast). 18a, paratype NMV P305572, pygidium, dorsal view x 1.6 (latex cast) from PL6361. 18b, same (internal mould). 48 Andrew C. Sandford Homalonotid trilobites from the Silurian and Lower Devonian Sandford, one specimen shows the anterior margin of the cephalon, with the impression of the rostral suture evident on the surface of the doublure (Fig. 19.4). The other speci- men shows the posterior portion of the ventral surface of the rostral plate. The pygidial pleural furrows are shallow, although the greater depth of the first furrow appears to be partly exaggerated by longitudinal shearing. Holloway and Sandford (1993) distinguished Trimerus ( Ramiotis ) iani from previously documented homalonotids from the Ludlow-Pragian strata of south-eastern Australia. Of the species discussed by Holloway and Sandford, only speci- mens described by Talent (1964) as “ Dipleura sp.” can be referred to T. {Ramiotis). In cephalic features Trimerus ( Ramiotis ) iani is similar to T. (R.) otisi from the Ludlow of Victoria. These species share a glabella with elongate proportions (L:W~1 .2) and moderately tapering (-25°) and weakly concave sides, a preglabellar field of similar length (0.2 cranidial length), similar cephalic propor- tions (L:W ~0.7), a transverse rostral suture, very short pygidial proportions and comparable pygidial segmentation. T. (R.) otisi differs in having well defined glabellar lobation and a distinctive tricuspid anterior cephalic margin. In cephalic features T. (R.) permutus from the Ludlow of Poland is also closely comparable, although it differs in having strongly expressed glabellar lobation. In pygidial features iani demon- strates its closer affinities to other Lower Silurian T. (Ramiotis), particularly in the weak expression of the pleural furrows, the deeper ring furrows, and the rounded tip. In pygidial morph- ology iani is closest to T. (R.) dyaulax, that differs in having weaker segmentation. T. (R.) iani is closest in age to the type species T. (R.) rickardsi, but differs in having a shorter preglabellar field, weaker glabellar lobation, a transverse rostral suture, and a rounded pygidial tip. Trimerus ( Ramiotis ) otisi sp. nov. Figure 20 Homalonotus vomer . — Chapman, 1912: 298, pi. 63 fig. 1, non pi. 62 figs 2, 3, pi. 63 fig. 2. Trimerus vomer . — Gill, 1949: 65, text-fig. IB, non text-fig. 1C. Trimerus cf. lilydalensis. — Williams, 1964: 285, table 1. Trimerus sp. — Garratt, 1968: 164, chart 1. Trimerus cf. vomer . — Garratt, 1968: 164, chart 1. Type material. Holotype NMV P304741 (cephalon and displaced tho- racopygon) from PL1898, Eden Park, Victoria (Fig. 20.5). Paratypes 49 NMV P304727 (cephalon and displaced thoracopygon), NMV P304722-P304724 (cephala), NMV P304726 (cranidium), NMV P304718-P304720 (pygidia) from PL6633, Eden Park. Paratypes NMV P304739, NMV P304797 (incomplete cephalothoraxes), NMV P304738 (thoracopygon), NMV P304742 (pygidium) from PL1898. Paratypes NMV P304730, NMV P304733, P304734 (cephala), NMV P304729 (pygidium) from PL6632, Eden Park. Paratype NMV P304737 (pygidium) from PL300, Clonbinane, Victoria. For localities see Fig. 11. Registered material. 141 specimens: 4 dorsal exoskeletons, 3 dorsal exoskeletons with displaced cephala, 3 articulated cephalothoraxes, 2 thoraxes with displaced cephala, 15 cephala, 22 cranidia, 3 librigenae, 1 hypostome, 8 thoracic segments, 4 articulated thoraxes, 10 articulat- ed thoracopygons, 66 pygidia. Eden Park: NMV P304766-P304770 from PL6627, Jutson locality VIII. NMV P304743, NMV P304775- P304778 probably from PL6627. NMV P304771-P304773 from PL6628. NMV P304832-P304836 from PL6629, Williams locality F130. NMV P304745, NMV P304837-P304840 from PL6630. NMV P304738-P304742, NMV P304779-P304799 from PL1898. NMV P304832-P304838 from PL6631. NMV P304728-P304736, NMV P304807-P304822 from PL6632. NMV P304718-P304727, NMV P304747-P304765 from PL6633. NMV P304800-P304806 from PL6635. NMV P304744 from “Eden Park”. Unregistered specimens from PL6634 and PL6616, Williams locality XI 2. Upper Plenty-Clonbinane area, Victoria: NMV P304774 from PL6636, Jutson locality 111= Williams locality FI 2, Upper Plenty. NMV P304746, NMV P304830, P304831 fromPL6638, Kenley locality 32g, Upper Plenty. NMV P304844-P304846 from PL6639, Upper Plenty. NMV P304841-P304843 from PL6642, Williams locality G23, Wandong. NMV P304737, NMV P304823-P304826 from PL300. NMV P304827-P304829 from PL1782, Clonbinane. In addition, homalonotids recorded by Garratt (1968) from PL6618, Jutson locality IX=Williams locality F90, Eden Park and by Williams (1964) from Williams locality F41, Clonbinane undoubtedly refer to Trimerus ( Ramiotis ) otisi, although specimens from these localities are not rep- resented in the MOV collections. Humevale: NMV P304851 from PL6646, Williams locality W17. Unregistered specimen from PL6648. For localities see Fig. 11. Stratigraphic distribution. Eden Park Formation, from 350 m above the base to the top of the unit, Notoparmella plentiensis Assemblage Zone, late Ludlow. Diagnosis. Cephalon with length 0.7 times width, anterior margin tricuspid. Glabella trapezoid, length about 1.17 times width, sides converging at about 25°, anterior margin strongly defined, transverse. Glabellar lobation moderately well defined. Palpebral lobe placed opposite 0.6 glabellar length (0.5 cranidial length). Moderately impressed sutural and Figure 19.1-19.4, 19.6. Trimerus ( Ramiotis ) iani sp. nov. la, holotype NMV P137221, cephalon, dorsal view (internal mould) from PL359. lb, same, oblique view, lc, same, lateral view. Id, same, dorsal view (latex cast). 2, paratype NMV P137224, cephalon, dorsal view (internal mould) from PL359. 3, paratype NMV P137222, cranidium, dorsal view (internal mould) from PL359. 4, paratype NMV P137223, cephalon, dorsal view (internal mould) from PL359. 6a, paratype NMV P137227, pygidium (internal mould) lateral view, from PL359. 6b, same, posterior view. 6c, same, dorsal view. Figure 19.5, 19.8, 19.10, 19.12. Trimerus ( Ramiotis ) tomczykowae sp. nov. 5, paratype NMV P305191, cranidium, dorsal view x 1.1 (internal mould) from PL2263. 8, holotype NMV P138203, pygidium (internal mould) x 1.5, from the “ Illaenus band”, Costerfield South. 8a, posterior view. 8b, dorsal view. 8c, lateral view. 10, NMV P138291, cephalon (fragment), dorsal view (latex cast) from PL1460. 11, paratype NMV P138293, rostral plate, dorsal view x 6.0 (latex cast) from PL1460. Figure 19.7, 19.9, 19.11. Trimerus ( Trimerus ) harrisoni (McCoy, 1876). 7, holotype NMV P7503, complete exoskeleton, dorsal view x 1.8 (inter- nal mould) from PL1620. 9, NMV P136645, complete exoskeleton, dorsal view of cephalon with impression of hypostome x 2.1 (internal mould) from PL1620. 11, NMV P2500, complete exoskeleton, lateral view x 2.6 (internal mould) from PL1620. Homalonotid trilobites from the Silurian and Lower Devonian sub-ocular furrows, weak eye ridges. Preglabellar field 0.2 times cranidial length. Anterior branch of facial suture straight poster-iorly, converging at about 55°, then anteriorly curving strongly to the transverse, rostral suture transverse. Dorsal sur- face of rostral plate semicircular, without connective sutures, equal in length to preglabellar field. Ventral surface of rostral plate with length equalling width, connective sutures moder- ately convex and converging posteriorly at about 90°. Pygidium triangular, short, length 0.6 times width, with sides weakly con- vex and converging posteriorly at about 110° to an obtusely pointed tip. Pygidial axis narrow, width (measured across sec- ond ring) 0.3 times pygidial width, 10 axial rings, ring furrows deep. 7-8 pleural ribs, pleural furrows as deep as ring furrows and uniform in depth to lateral border furrow, rib-ring medial- ly offset at fifth rib. Axial furrows moderately impressed poste- riorly, poorly defined anteriorly, postaxial ridge posteriorly raised. Lateral border furrow well defined, border a raised ridge. Discussion. The distinctive tricuspid cephalic margin immed- iately distinguishes Trimerus ( Romiotis ) otisi from all other members of the genus. The pygidium of otisi is also distin- guishes the species from many congeners, with deep pleural and ring furrows and the anteriorly indistinct axial furrow shared only with the Bolivian T. (R.) sp. A. The Bolivian species also displays a well defined border furrow and border comparable to that of otisi and T. (R.) thomasi, but the species differs in having longer pygidial proportions (L:W~0.9) and a wider pygidial axis (width 0.5 pygidial width). The very short pygidial proportions of otisi are closest to those of T. (R.) iani, that also shares comparable cephalic features including moder- ate length of the preglabellar field, longer glabellar proportions, a glabellar outline weakly expanded across LI -LI (tr.), and a forward eye position (opposite 0.5 cranidial length). In addition to the tricuspid margin, otisi differs from iani in having distinct subocular furrows and eye ridges. The tricuspid cephalic margin of Trimerus ( Ramiotis ) otisi is comparable to that of Digonus, as is the quadrate course of the preocular facial sutures and the very weak expression of the pygidial axial furrows anteriorly, with up to the third rib fused with the axial rings. The species also resembles Digonus in the very deep and equally deep pleural and ring furrows. The mod- 51 erately well defined glabellar lobation and glabellar outline weakly expanded across LI -LI (tr.) are comparable to early Digonus morphologies best represented by D. zeehanensis, and in these respects otisi is the best candidate as an intermediate species between Trimerus and Digonus. Significant features precluding the assignment of otisi to Digonus include the lack of a ventral process on the rostral plate, the very narrow pygidi- al axis, and the strongly raised postaxial ridge. T. (R.) otisi and Digonus further differ in the expression of the dorsal sections of the connective sutures (short in Digonus but absent in otisi), and in the size of the medial cephalic process (small in Digonus but very large in otisi). In the large size of the medial cephalic cusp, otisi is most closely comparable to T. (T.) johannis. However, marked differences in cephalic morphology between tricuspid T. {Trimerus) (represented by johannis), T. ( Ramiotis ) (represented by otisi) and Homalonotus suggest that the tri- cuspid cephalic margin was independently derived in these groups. The paratype cephalon of Homalonotus vomer (NMV P12303, Fig. 20.1 and figured Chapman, 1912, pi. 63 fig. 1, Gill, 1949, text-fig. IB) belongs to Trimerus {Ramiotis) otisi, and is designated a paratype of the new species. Chapman (1912) considered the specimen to be a juvenile form of vomer, but did not otherwise compare it with the holotype cephalon. Although Chapman’s description of vomer seems to be based entirely on the holotype, Gill (1949) included features of the paratype in his redescription of the species. Chapman and Gill recorded the localities of the paratype and the holotype of vomer only as “Wandong”. The holotype was almost certainly collected from PL286 (Williams locality F22, Wandong), where the species occurs in relative abundance and in an identical lithology to that of the holotype. As noted by Gill, the paratype occurs in a yellow-brown, fine grained sandstone. T. {R.) otisi has been collected from similar lithologies in the Merriang Syncline to the east of Wandong, and it is presumed the specimen was collected in that area, and probably in the vicinity of PL300, Comet Creek Mine, Clonbinane. It is remarkable that although Trimerus {Ramiotis) otisi is very common in the Eden Park Form a tion in the Merriang Syncline between Merriang and Clonbinane, trilobites (pre- dominantly homalonotid) are exceedingly rare in the equivalent horizons of the nearby Kinglake Basin sequence. Figure 20. Trimerus {Ramiotis) otisi sp. nov. la, paratype NMV P12303 (and paratype of Homalonotus vomer Chapman, 1912), cephalon, dorsal view x 1.9 (latex cast) from Wandong (exact locality unknown), lb, same, x 1.7 (internal mould). 2, paratype NMV P304726, cranidium, dorsal view x 2.7 (latex cast) from PL6633. 3, paratype NMV P304739, incomplete cephalothorax, dorsal view x 3.2 (latex cast) from PL1898. 4, paratype NMV P304797, incomplete cephalothorax, dorsal view of cephalon x 2.7 (internal mould) from PL 1898. 5a, holotype NMV P304741, cephalon and displaced thoracopygon, dorsal view of cephalon x 1.15 (internal mould) from PL 1898. 5b, same, enlargement of anterior margin x 2.3. 5c, same, anterior view of cephalon x 1.15. 6, paratype NMV P304727, cephalon and displaced thoracopygon, dorsal view of cephalon x 2.75 (internal mould) from PL6633. 7, paratype NMV P304723, cephalon, enlargement of anterior margin x 1.6 (internal mould) from PL6633. 8, paratype NMV P304722, cephalon, enlargement of anterior margin x 2.3 (latex cast) from PL6633. 9a, paratype NMV P304730, cephalon, anterior view x 2.1 (latex cast) from PL6632. 9b, same, oblique view. 9c, same, dorsal view. 10, paratype NMV P304738, thoracopygon, dorsal view x 3.0 (latex cast) from PL1 898. 11, paratype NMV P304734, cephalon, dorsal view x 4.3 (internal mould) from PL6632. 12, paratype NMV P304733, cephalon, ventral view (doublure) x 2.0 (latex cast) from PL6632. 13, paratype NMV P304724, cephalon, ventral view (doublure) x 3.4 (latex cast) from PL6633. 14a, paratype NMV P304720, pygidium, lateral view x 1.9 (latex cast) from PL6633. 14b, same, dorsal view x 1.8. 15a, paratype NMV P304719, pygidium, dorsal view x 1.15 (latex cast) from PL6633. 15b, same, lateral view. 16a, paratype NMV P304729, pygidium, dorsal view x 1.7 (latex cast) from PL6632. 16b, same, posterior view x 1.6. 17, paratype NMV P304742, pygidium, dorsal view x 2.4 (internal mould) from PL 189 8. 18, paratype NMV P304737, pygidium, ventral view (doublure) x 1.45 (latex cast) from PL300. 19, paratype NMV P304718, pygidium, dorsal view x 1.4 (internal mould) from PL6633. Homalonotid trilobites from the Silurian and Lower Devonian Environmental notes. See taphonomy and biofacies. Trimerus ( Ramiotis ) otisi is considered to have inhabited mid shelf environments. Trimerus ( Ramiotis ) thomasi sp. nov. Figure 21 Trimerus ( Dipleura ?) sp.. — Talent, 1964: 49 (pars), fig. 6 non pi. 26 figs 1-2 ( =Digonus wenndorfi sp. nov.) Type material. Holotype NMV P305067 (pygidium) from PL6653, Heathcote, Victoria (Fig. 21.15). Paratypes NMV P305046-P305047 (cephala), NMV P305048, NMV P305053-P305056, NMV P305058, P305059, (cranidia), NMV P305072-P305073 (librigena), NMV P305061, NMV P3 04063, NMV P305068-P305070, (pygidia) from PL2264, Thomas locality F45, Parish of Heathcote, Heathcote. Paratype NMV P63385 (cranidium), NMV P305043 (pygidium) from PL2265, Thomas locality F46, Parish of Heathcote, Heathcote. For localities see Fig. 8. Previously figured material. NMV P63385 (ex GSV38132, cranidium, figured Talent, 1964: text-fig. 6) from PL2265. Registered material. 94 specimens. 3 cephala, 40 cranidia, 5 librigenae, 1 thoracic segment, 45 pygidia. NMV P305042-P305045, NMV P305 127-P305 130 from PL2265. NMV P305046-P305053, NMV P305055-P305066, NMV P305068-P305123, NMV P305126 from PL2264. NMV P305124-P305125, NMV P305054, NMV P305067 from PL6653. NMV P305131 from PL6654, Heathcote. NMV P305132 from PL6655, Heathcote. NMV P305133 from PL6656, Heathcote. Unregistered specimens from PL2234, Thomas locality F21, Parish of Heathcote, Heathcote, PL6727, Thomas locality F28, Parish of Redcastle, Heathcote. For localities see Fig. 8. Stratigraphic distribution. Mclvor Formation, -500 m above the base of the unit, lower Notoparmella plentiensis Assemblage Zone, mid Ludlow. Derivation of name. For Alan T. Thomas (University of Birmingham), for his contribution to the study of homalonotids. Diagnosis. Cephalon with length about 0.7 times width. Glabella with length equalling width, trapezoid in outline, sides more or less straight, tapering weakly forwards at about 25°. Glabellar lobation distinct, SI -S3 very shallow, LI, L2, L3 and frontal lobe subequal in length (exsag.). Anterior margin of glabella well defined, transverse or with broad medial inden- tation. Preglabellar field long, 0.25 times cranidial length, slightly concave (tr. sect.). Palpebral lobe placed with midline opposite 0.45 cranidial length (0.6 glabellar length) and with 53 8-6 1.6 times preoccipital glabellar width. Anterior margin of cranidium transverse to very broadly convex. Pygidium with length equal to width, sides converging at about 70°, angular tip. Pygidial axis wide, 0.43 times pygidial width, continuous with wide, raised postaxial area. 9 distinct axial rings, 7 distinct pleural ribs, axial furrows shallow, pleural furrows moderately impressed, rib-ring offset at fifth-sixth rib. Remarks. Although specimens are abundant, the quality of the material of Trimerus ( Ramiotis ) thomasi and its similarity to T. (R.) otisi does not warrant a full description. The trapezoid, straight-sided and moderately tapered glabellar outline indicate assignment to Trimerus {Ramiotis). The moderate length of the preglabellar field and the weak glabellar lobation are in accord with this assignment. The cranidium figured by Talent (1964: fig. 6) as T. (Dipleural) sp. is significantly larger than cranidia from the type locality of T. (R.) thomasi which lies about 500m south-east of PL2265 and more or less on strike. The larger specimen differs conspicuously in that the anterior branches of the facial sutures are more strongly convergent (-70°) and the rostral suture correspondingly shorter (tr.), the preglabellar field is slightly shorter (0.22 cranidial length), and the medial flexure of the occipital furrow is strongly expressed (see Fig. 21.12). Whether these differences are attributable to the size or to variability within the species cannot be resolved without further specimens of large individuals. Within the range of intraspecific variability, there are specimens of thomasi exhibiting very weak lobation, and short and more strongly tapered glabellar outlines (e.g. Figs 21.4, 21.5). These morphs are closely comparable to the contemporary T. (R.) salteri from England. Original illustrations of the holo- type [Salter, 1865: pi. 12 fig. 1, as Homalonotus ( Koenegia ) ludensis ] indicates that the English species differs from thomasi in having a wider (tr.) preocular fixigenal field and deeper axial furrows. Trimerus {Ramiotis) permutus from the upper Ludlow of Poland is close in age to T. {R.) thomasi. In cephalic morph- ology permutus differs markedly from thomasi in exhibiting strong glabellar lobation, an elongate glabellar outline, more strongly convergent facial sutures, and a short preglabellar field. The species are more closely comparable on pygidial features. Although pygidia of permutus are poorly known, they share with thomasi the elongate pygidial proportions and mod- erately and equally impressed axial, pleural and ring furrows. Figure 21. Trimerus {Ramiotis) thomasi sp. nov. 1, paratype NMV P305046, cephalon, dorsal view x 2.5 (internal mould) from PL2264. 2, paratype NMV P305048, cranidium, dorsal view x 2.5 (latex cast) from PL2264. 3, paratype NMV P305055, cranidium, dorsal view x 2.6 (inter- nal mould) from PL2264. 4, paratype NMV P305059, cranidium, dorsal view x 2.4 (latex cast) from PL2264. 5, paratype NMV P305053, cranidium, dorsal view x 2.7 (internal mould) from PL2264. 6, paratype NMV P305056, cranidium, dorsal view x 2.9 (latex cast) from PL2264. 7, paratype NMV P305054, cranidium, dorsal view x 2.6 (latex cast) from PL6653. 8, paratype NMV P305058, cranidium, dorsal view x 2.6 (latex cast) from PL2264. 9a, paratype NMV P305047, cephalon, anterior view x 2.3 (latex cast) from PL2264. 9b, same, oblique view x 1.8. 9c, same, dorsal view x 2.4. 10a, paratype NMV P305072, librigena, lateral view x 2.0 (internal mould) from PL2264. 10b, same, dorsal view (latex cast). 11, paratype NMV P305073, librigena, ventral view (doublure) x 1.7 (latex cast) from PL2264. 12, NMV P63385, cranidium, dorsal view x 1.95 (internal mould) from PL2265. 13a, paratype NMV P305068, pygidium, lateral view x 1.9 (internal mould) from PL2264. 13b, same, dorsal view. 14, paratype NMV P304063, pygidium, dorsal view x 3.0 (internal mould) from PL2264. 15a, holotype NMV P305067, pygidium, dorsal view x 3.0 (internal mould) from PL6653. 15b, same, posterior view. 15c, same, oblique view. 16, paratype NMV P305070, pygidium, ventral view x 4.0 (internal mould) from PL2264. 17, paratype NMV P305061, pygidium, dorsal view x 2.5 (latex cast) from PL2264. 18a, paratype NMV P305043, pygidium, lateral view x 2.4 (latex cast) from PL2265. 18b, same, dorsal view x 2.5. 19, paratype NMV P305069, pygidium, lateral view x 3.2 (internal mould) from PL2264. 54 Andrew C. Sandford The pygidia differ in that thomasi exhibits a markedly raised postaxial ridge, whilst that of permutus is low. Trimerus ( Ramiotis ) otisi from southern central Victoria is another contemporary of T. (R.) thomasi. Despite their geo- graphic and temporal proximity, otisi and thomasi can be easily distinguished. The pygidia of thomasi are much more elongate, have a wider axis, and exhibit shallower pleural and ring furrows and deeper axial furrows. The cranidia of thomasi generally exhibit a more straight-sided and shorter glabellar outline, more weakly defined glabellar lobation, and a longer preglabellar field. Despite these differences, otisi and thomasi are considered to be closely related. Together with the poorly known T. ( R .) sp. A from Bolivia, the species share a distinct pygidial morphology including pleural furrows that continue to a deep border furrow, and a raised ridge-like border. This pygidial morphology distinguishes this species group from Llandovery-Wenlock T. {Ramiotis). In other respects T. (R.) sp. A is morphologically intermediate between otisi and thomasi, exhibiting longer pygidial proportions of the latter and deeper pygidial pleural and ring furrows of the former. Of the Llandovery-Wenlock species of Trimerus (Ramiotis), T. (R.) thomasi shows greatest resemblance to T. (R.) tom- czykowae from the Wenlock of Victoria (described below), that also exhibits very weak lobation, a transverse rostral suture, shorter glabellar proportions and a long preglabellar field. The species also share moderate and subequal depth of the pygidial pleural and ring furrows, distinguishing them from other Llandovery-Wenlock T. (Ramiotis). T. (R.) tomcykowae differs in having a short, rounded pygidial outline, slightly longer glabellar proportion, and a deep medial indentation of the glabellar anterior margin. Environmental notes. Trimerus (Ramiotis) thomasi is the dom- inant element of monospecific or low diversity assemblages. Three cephala from PL2264 are the only known articulated specimens, isolated tergites representing 97% of the entire population, typical of taphofacies Til. The species occurs in medium-grained sandstones, associated with a shelly fauna dominated by gastropods and brachiopods. T. (R.) thomasi is considered to have inhabited inner shelf environments. Trimerus ( Ramiotis ) tomczykowae sp. nov. Figures 19.5, 19.8, 19.10, 19.12 Type material. Holotype NMV PI 38203 (pygidium) from the “ Illaenus band”, Costerfield, Victoria (Fig. 19.8). Paratype NMV P305191 (cranidium) from PL2263, Thomas locality F44, Parish of Dargile, Costerfield. Paratypes NMV P138292, P138293 (rostral plates) from PL1460, Thomas locality F43A, Parish of Dargile, Costerfield. For localities see Fig. 10. Registered material. 8 specimens: 1 cephalon, 3 cranidia, 2 rostral plates, 2 pygidia. NMV P138291-P138294, NMV P138296, NMV P139358 from PL1460. NMV P305191 from PL2263. NMV P138203 from the “ Illaenus band”. Stratigraphic distribution. Illaenus band, Wapentake Formation. Rickards and Sandford (1998) interpreted Ananaspis typhlagogus to indicate a Wenlock age for these beds. Derivation of name. For Ewa Tomczykowa (Instytut Geologiczny, Poland) for her contribution to the study of homalonotids. Diagnosis. Glabella trapezoid, elongate, length ~1.1 times width, sides parallel opposite paraglabellar area, straight and converging at about 20° anteriorly, with extremely weak lobation, anterior margin strongly indented. Paraglabellar area distinct. Palpebral lobe placed at about 0.55 glabellar length. Rostral suture more or less transverse. Dorsal surface of rostral plate long, one third length of ventral surface, anterior margin of rostral plate with sides converging at 115°, rounded medi- ally. Connective sutures straight, diverging forwards at 55°. Ventral surface of rostral plate flat, length equalling width. Pygidial axial furrows moderately impressed. Axis about 0.4 times pygidial width, sides straight and tapering at 25°, poster- ior third of axis moderately convex and raised, terminal piece continuous with and not distinguishable from wide, raised, subparallel- sided postaxial ridge. 10 axial rings, ring furrows moderately impressed, 7 pleural ribs, pleural furrows moder- ately impressed and equal in depth to ring furrows, shallowing distally. Pleural offset at fourth rib. Discussion. Trimerus (Ramiotis) tomczykowae is a very rare element of the Illaenus band fauna. Despite the very few and fragmentary specimens available, a fairly complete reconstruc- tion of the cephalon can be made, and together with the single, well-preserved pygidium clearly represents a new species of homalonotid and a full description of the species is possible with the material at hand. Significant cephalic characters include the elongate, trapezoid, straight- sided, moderately tapering glabella, the weakly expressed glabellar lobation, the moderate length of the preglabellar field (0.25 cranidial length) and the short pygidial proportions (LAV-0.9) indicate assign- ment to T. (Ramiotis). In cephalic features tomczykowae differs from the type species T. (R.) rickardsi in having a transverse rostral suture and strong medial indentation of the anterior glabellar margin. In these respects tomczykowae is closest to T. (R.) dyaulax. The equal depth of pleural and ring furrows attained by Trimerus ( Ramiotis ) tomczykowae is considered here to repre- sent a significant stage of homalonotine development. This feature distinguishes Wenlock-Ludlow T. (Ramiotis) from Llandovery members of the group. Although in its short pygidial proportions and rounded outline the affinities of tomczykowae to Llandovery T. (Ramiotis) are clear, it repre- sents a transitional morphology between these and Upper Silurian species of T. (Ramiotis) and T. (Trimerus). Indeterminate homalonotids from the Wenlock of central Victoria include a partly disarticulated exoskeleton from the old Tooronga Brick Pit, high in the Anderson Creek Formation, only briefly examined by the author. The abundant and diverse trilobite fauna occurring in the Bylands Siltstone at PL206, Wallan has yielded only an single homalonotid thoracic segment. Environmental notes. Trimerus (Ramiotis) tomczykowae is a very rare element of a moderately diverse fauna occurring in the nodular mudstones of the Illaenus band. The fauna is dom- inated by the blind illaenid Thomastus thomastus Opik, 1953. Homalonotid trilobites from the Silurian and Lower Devonian 55 Together with Ananaspis typhlagogus the composition of the fauna closely resembles a contemporary fauna from Argentina described by Waisfeld and Sanchez (1993). The taphonomy of the Illaenus band is characterised by an abundance of complete exoskeletons of Thomastus in the ‘bumastoid stance’, pre- sumed to be individuals preserved in their infaunal life position (Sandford and Holloway, 1998). The taphonomy and the aute- cology of Thomastus in the mudstones suggest a deeper water environment for tomczykowae. Several specimens of tom- czykowae occur in a richly fossiliferous sandstone presumably interbedded within the mudstone, although the bed is not known in outcrop. Wenndorfia gen. nov. Type species. Homalonotus mutabilis Koch, 1880 from the Lower Devonian (upper Emsian) of Germany. Other species included. Parahomalonotus angustico status Tomczykowa, 1975, P. bostoviensis Tomczykowa, 1975, Digonus ele- gans Tomczykowa, 1975, Homalonotus expansus Hector, 1876 (=H. ( Burmeisteria ) huttoni Allan, 1935, -D. margaritifer Wenndorf, 1990), H. forbesi Rouault, 1855, Dipleura fornix Haas, 1968, Trimerus lily- dalensis Gill, 1949, H. miloni Renaud, 1942, H. multicostatus Koch, 1883a, H. mutabilis Koch, 1880, T. novus Tomczykowa, 1975, H. obtusus Sandberger and Sandberger, 1849, P. planus junior Wenndorf, 1990, H. planus Sandberger, 1849, U P. sp. nov. A” in Wenndorf, 1990, “D. sp.” in Le Menn et al., 1976. Range. Lower Devonian (Lochkovian-Emsian). Diagnosis. Glabella of moderate height to very low, weakly tapering. Lobation indistinct. Anterior margin of cranidium U-shaped and more or less concentric with rounded anterior margin of glabella, anterior branches of facial suture broadly curved and strongly convergent anteriorly, rostral suture short (tr.), transverse or continuous in curvature with anterior section of facial suture. Eyes placed posteriorly, opposite 0.3-0.45 cephalic length. Sides of glabella more or less straight and parallel, varying to strongly concave. Ventral surface of rostral plate with low or prominent anterior node. Pygidial outline parabolic with rounded tip, but triangular with pointed tip in early species. Pygidium with axial furrows very shallow to indistinct, axis without independent convexity, wide (0.4-0.65 time pygidial width anteriorly), strongly tapering (furrows con- verging at 30-45°), not reaching margin but terminating at about 0.85-0.95 times pygidial length, semicircular terminal piece short (sag.), slightly raised, but often indistinct from postaxial field. Discussion. Wenndorfia has been erected to accommodate many species previously assigned to Parahomalonotus (see above). In establishing Parahomalonotus, Reed (1918) listed a number of characters in the diagnosis that are not apparent in the type species Homalonotus gervillei DeVerneuil, 1850 from the Emsian of France. Reed described the “facial sutures uniting in a regular wide arched commissure close to (the) anterior margin” as diagnostic of the genus (Reed, 1918: 326). However, in gervillei the anterior cranidial margin is quadrate, the rostral suture being transverse and the anterior branches of the facial sutures subparallel (see Pillet, 1973: fig. 120). Reed’s original diagnosis also states a pygidial entire margin and more or less indistinct trilobation as diagnostic. However, in con- tradiction to these definitive parameters, trilobation is well defined by the raised pygidial axis of gervillei, and the pos- taxial ridge extends posteriorly as a short posteromedial spine (see Pillet, 1973, pi. 37 fig. 5). The revised diagnosis for Parahomalonotus s.s. (see above) emphasises the distinctive features of P. gervillei, and hence differs markedly from that given by Reed (1918). Of species assigned to Wenndorfia, Reed assigned Homalonotus planus, H. obtusus and H. multicostatus to Parahomalonotus with question. In the course of their facial sutures and in pygidial features, these species are in closer agreement with Reed’s diagnosis of Parahomalonotus than with the type gervillei. These species further differ from gervillei in having pygidial axes that are much wider anteriorly, taper more strongly, are not independently convex, lack postaxial ridges, and terminate between 0.85 and 0.95 pygidial length, and glabellae that are very low and lack distinct lobation. Species included in Wenndorfia were previously assigned variously to Digonus, Dipleura, Parahomalonotus and Trimerus. Wenndorfia is most reliably distinguished from these genera by the rounded outline of the anterior cranidial margin, concentric to the strongly rounded glabellar anterior margin. This morphology contrasts most strongly with the angular, quadrate anterior cranidial margin of Digonus. In pygidial morphology early representatives of Wenndorfia and Digonus are alike, although marked differences between these groups are manifest in later representatives in trends toward short, rounded outlines, and elongate, triangular outlines with acute processes respectively. Later representatives of Wenndorfia further differ from Digonus in exhibiting moderately to strongly concave glabellar sides. Rostral morphology is recognised in this work as a reliable character distinguishing Wenndorfia from Trimerus and Dipleura. The rostral plate of Wenndorfia exhibits a diamond- shaped outline and exhibits a variably developed antero- medial node, whereas rostral plates of Trimerus and Dipleura are elongate pentangular and pentangular in outline respectively, and lack rostral processes. Developmental trends in the pygidial morphology of Wenndorfia, manifest in the replacement of triangular outlines by rounded outlines in the mid-Pragian, are comparable to developmental trends in Dipleura in the Late Silurian. However, pygidia of species assigned to Wenndorfia vary markedly in the depth of the pleural and ring furrows. In a number of species the pleural and ring furrows are very shallow to indistinct, and approach morphologies exhibited by species of Dipleura. The distinction of these morphologically conver- gent species relies on the course of the cephalic sutures, as dis- cussed above (see Dipleura ). Pygidia of Trimerus ( Ramiotis ) and T. ( Trimerus ) are distinguished from the Wenndorfia pygidial morphology in having a distinct postaxial ridge. Wenndorfia mutabilis, unlike most other species of the genus, is abundant and its dorsal and ventral morphology is both completely known and recently redescribed (see Wenndorf, 1990). The species exhibits the rounded anterior cranidial margin, the weakly expressed pygidial trilobation and 56 Andrew C. Sandford the entire pygidial margin included in Reed’s original diag- nosis of Parahomalonotus and in the diagnosis of Wenndorfia. Other features exhibited by mutabilis and considered typical of Wenndorfia include the strongly rounded anterior margin of glabella, the posteriorly placed eyes, the small node on the ventral surface of rostral plate, and the wide, strongly tapering axis terminating at about 0.85 pygidial length. W. mutabilis is distinguished from its congeners in having a glabella of moder- ate height with strongly concave sides, a wide palebral area, and a shorter pygidial outline with shallow pygidial ring and pleural furrows. Lochkovian-lower Pragian Wenndorfia differ from upper Pragian-Eifelian species in variously exhibiting features that can be regarded as underived. These features include a trapezoid glabellar outline, the expression of glabellar lobation, moderately impressed pygidial axial furrows, deep pleural and ring furrows, and a triangular pygidial outline with an obtusely angled tip. In these respects these early species represent a transitional morphologies with early Digonus. Early represen- tatives of Wenndorfia include several Polish homalonotids orig- inally described as species of Digonus by Tomczykowa (1975) and regarded by Wenndorf (1990) as representing an early Digonus morphology. Lacking in these Polish species is the quadrate anterior cranidial outline exhibited by the earliest Digonus (see above, D. wenndorfi). The strongly rounded anterior glabellar margin, the concentric cranidial margin and posteriorly placed eyes indicate assignment of these species to Wenndorfia. The upper Lochkovian W. bostoviensis from Poland is the earliest species assigned to Wenndorfia. W. bostoviensis differs from W. mutabilis and other Wenndorfia in having and a short triangular pygidial outline. This pygidial morphology suggests affinities of the earliest Wenndorfia to the earliest Digonus. Tomczkowa (1975) assigned homalonotids from upper Lochkovian-lower Pragian strata immediately overlying Wenndorfia bostoviensis to Digonus vialai. The number of specimens representing these two morphologies is limited, with four cranidia representing vialai and two representing bostoviensis. Tomczkowa distinguished vialai from bostov- iensis by its trapezoid rather than rectangular glabellar outline, but the significance of these supposed differences cannot be determined without reference to larger populations. In all other respects it is difficult to distinguish vialai from bostoviensis, and the specimens are best considered conspecific. Similarly, Tomczykowa described Parahomalonotus angustico status from the lower Pragian of Poland using a limited collection of pygidia, only one complete. Tomczykowa (pi. 6 figs 1-4) assigned pygidia from a slightly lower horizon to Parahomalonotus forbesi (Rouault, 1855), distinguished from those of angusticostatus in having fewer axial rings and ribs and shallower pleural and ring furrows. However, the pygidia figured do not differ markedly from angusticostatus and are considered to be conspecific. Wenndorfia expansa (Hector, 1876) Figure 22.7-22.14 Homalonotus expansus Hector, 1876: 602, pi. 27 fig. 2. Homalonotus sp. — Hutton, 1888: 257. Homalonotus (Burmeisteria) huttoni. — Allan, 1935: 28, pi. 1 figs 4-5. Homalonotus (Digonus) expansus. — Allan, 1935: 29, pi. 1 fig. 1. Homalonotus ( Digonus ) cf. expansus. — Allan, 1935: 29, pi. 1 fig. 3. Burmeisteria ( Digonus ) expansus. — Saul, 1965: 271. Burmeisteria huttoni. — Saul, 1965: 271. — Tomczykowa, 1975: 11. Digonus expansus. — Tomczykowa, 1975: 11. — Wenndorf, 1990: 16. Burmeisteria ( Digonus ) cf. expansus. — Speden and Keyes, 1981: pi. 4 fig. I. Burmeisteria expansa. — Cooper, 1982: 27. Burmeisteria huttoni. — Cooper, 1982: 27. — Wenndorf, 1990: 16. Digonus margaritifer — Wenndorf, 1990: 77, pi. 11 figs 8-9. Type material. Homalonotus expansus. Lectotype (pygidium, figured Hector, 1876, Allan, 1935: pi. 1 fig. 1, Fig. 22.10 herein, in the Geological Survey of New Zealand, cast in Museum Victoria regis- tered NMV PI 6844) from McKay locality 129, Rainy Creek, Reefton, south island, New Zealand. Paralectotype (pygidium, also the holotype of Digonus margaritifer, figured Allan, 1935: pi. 1 fig. 3, Wenndorf, 1990: pi. 11 fig. 8, in the Geological Survey of New Zealand, cast in Museum Victoria registered NMV PI 6846) from McKay locality 130, LankeyGully, Reefton. Paralectotype (pygidium, figured Hector, 1876, in the Geological Survey of New Zealand, cast in Museum Victoria registered NMV P16847) from locality 129. Homalonotus (Burmeisteria) huttoni. Holotype ZFc33 (dorsal exoskeleton, figured Allan, 1935: pi. 1 figs 4, 5) from “near Reefton”. Allan (1935) suggested the specimen is from locality 129. Figure 22. 1-22.6. Digonus wenndorfi sp. nov. 1, paratype NMV P304713, cranidium, dorsal view x 1.4 (internal mould) from PL2203. 2, paratype NMV P304862, pygidium, dorsal view x 1.5 (internal mould) from PL2203. 3a, paratype NMV P304864, pygidium, posterior view x 1.35 (internal mould) from PL2203. 3b, same, lateral view x 1.25. 4, paratype NMV P304866, pygidium, dorsal view x 1.25 (internal mould) from PL2203. 5, paratype NMV P304710, cranidium, dorsal view x 1.0 (internal mould) from PL2203. 6, paratype NMV P304717, pygidium, posterior view x 1.15 (internal mould) from PL2203. Figure 22.7-22.14. Wenndorfia expansa (Hector, 1876). 7a, holotype of Homalonotus ( Burmeisteria ) huttoni Allan, 1935, ZFc 33, enrolled exoskeleton, dorsal view of pygidium x 0.7 (internal mould) from near Reefton (exact locality unknown). 7b, same, lateral view of enrolled exoskeleton. 7c, same, dorsal view of cephalon. 7d, same, posterior view of enrolled exoskeleton. 8a, NMV P64122, pygidium, dorsal view x 0.9 (internal mould) from “Reefton”. 8b, same, posterior view. 9, paralectotype of Homalonotus expansus Hector, 1876 and holotype of Digonus margaritifer, pygidium, dorsal view x 0.9 (internal mould) from locality 130, Reefton (NMV P16846, plastercast of original). 10a, lectotype of Homalonotus expansus Hector, 1876, pygidium, dorsal view x 1.15 (internal mould) from locality 129, Reefton (NMV P16844, plastercast of original). 10b, same, external mould. 11, paralectotype of Homalonotus expansus Hector, 1876, pygidium, dorsal view x 0.7 (internal mould) from locality 129, Reefton (NMV P16847, plastercast of original). 12, NMV P64121, pygidium, dorsal view x 1.23 (internal mould) from “Reefton”. 13, NMV P64120, pygidium, dorsal view x 0.95 (internal mould) from “Reefton”. 14, ZFc 58, incomplete thoracopygon, dorsal view x 0.75 (internal mould) from “Reefton”. 58 Andrew C. Sandford Digonus margaritifer. Holotype (pygidium, and paralectotype of Homalonotus expansus , see above). Paratype (pygidium, figured Wenndorf, 1990: pi. 11 fig. 9, in the Palaontologisches Museum der Humboldt-Universitat Berlin, Germany), documented as coming from Kilmore, Victoria, acquired through the fossil dealer Kranz in 1885. However as the paratype is undoubtedly conspecific with the New Zealand material, as no homalonotids other than Trimerus ( Trimerus ) vomer , Trimerus ( Ramiotis ) otisi and Dipleura garratti are yet known in the well-collected Wenlock-Ludlow trilobite faunas from the Kilmore area, and as Wenndorfia would be most unexpected in strata of this age, I conclude that the Berlin specimen is from the Reefton area and its documentation confused, possibly by Kranz, with a specimen of T. (T.) vomer. Kranz may have acquired the Berlin specimen from Theodore Ranft who was the first to collect fossils in Lankey Gully at Reefton in 1872. Part of Ranft’s collection was purchased by the Melbourne Museum, who also had dealings with Kranz. Previously figured material. NZGS AR 676 (thoracopygon, figured Speden and Keyes, 1981: pi. 4 fig. I) from locality GS3737, S38/f523, Rainy Creek, Reefton. Registered material. 12 specimens: 1 complete exoskeleton, 2 thora- copygons, 1 incomplete thorax, 8 pygidia,. NMV PI 6844, NMV P16847 from locality 129. P16846 from locality 130. NZGS AR676 from locality GS3737. NMV P64120-P64122 from Lankey Gully (Ranft collection). ZFc33, ZFc58, ZFc309, ZFc350 from “Reefton”. One pygidium in the Palaontologisches Museum der Humboldt- Universitat Berlin. Stratigraphic distribution. Lankey Limestone, IBoucotia loyolensis Assemblage Zone, Emsian. Diagnosis. Glabella extremely low, sides parallel, without dis- tinct lobation, anterior margin broadly rounded, length 1.1 times width. Palpebral lobes placed opposite -0.55 glabellar length. Axial furrows very poorly defined. Pygidium short, with length 0.7 times width, weakly concave sides, converging posteriorly at about 100° to an obtusely angular tip. Pygidial axis wide, about 0.4 times pygidial width and very poorly defined anteriorly, tapering moderately with sides straight and converging at about 30°, length about 0.85 times pygidial length, posterior margin parabolic, 12-13 rings, semicircular terminal piece with distinct posterior margin. Axial furrows poorly defined, indicated by flexure of pleurae, almost indis- tinct anteriorly. 9 pleural ribs. Pleural furrows not reaching margin. Pleural and ring furrows deep and of equal depth. Anteriorly ribs and rings continuous, pleural offset at seventh rib. Row of equidimensional, regularly spaced tubercles on axial area of thoracic and pygidial segments, with two or three supplementary rows of smaller tubercles. Tubercles large on larger specimens, small to fine on smaller specimens. Similar, but less regular tuberculation on posterior border of fixigenae and thoracic and pygidial pleurae. Genal field with densely spaced, coarse granules. Discussion. Three homalonotids have been named from the Lower Devonian at Reefton. The best known of these is Homalonotus ( Burmeisteria ) huttoni, based on a single, near- complete, enrolled exoskeleton (Fig. 22.7). The specimen was described in detail in open nomenclature by Hutton (1888), and later named by Allan (1935). Several points in Hutton’s original description of the specimen cannot be substantiated, in particular those concerning the cephalic shape and proportions, and the presence of distinct glabellar lobation. Although only the first six segments of the pygidium are preserved in the holo- type, the ornament of tubercles on the pleural ribs and axial rings are distinctive and shared by several more completely preserved pygidia in the Museum of Victoria (NMV P64 120-2, Figs 22.8, 22.12-22.13) and in the New Zealand Geological Survey (NZGS AR 676, Fig. 22.14). These pygidia also share very deep ring and pleural furrows that are continuous to the seventh segment. The axial furrow is not impressed on the anterior half of the pygidium, but its position is indicated by a posteriorly increasing flexure at the junction of the ring and pleural furrows. These specimens are considered to be conspecific with the holotype of huttoni. The other two species, Homalonotus expansus and Digonus margaritifer, were based only on pygidia. Wenndorf ’s (1990) description of margaritifer includes features identical with those of H. (. Burmeisteria ) huttoni, such as the ornament, the deep axial ring and pleural furrows continuous to the seventh rib, the very shallow axial furrows, and the wide rounded- triangular outline. Wenndorf noted expansus and margaritifer shared pleural offsetting at the seventh-eighth rib, but distin- guished the former by the smooth surface texture (absence of tubercules) and the weak definition of the axis of the former species. Close inspection of the external mould of the lectotype of expansus reveals, however, a row of small tubercles on the first four axial rings and pleurae (see Fig. 22.10b). Corresponding ornament on the anteriormost rings on the internal mould were not recorded either by Hector (1876) or by Allan (1935) as this area of the pygidium is damaged. The spacing of these tubercles on the lectotype is similar to that of huttoni. The differences in the coarseness of tuberculation between the type specimens of expansus and specimens of huttoni are attributed here to the size of the pygidia. The lectotype of expansus, with the finest tubercles, is the smallest specimen. Moderately sized specimens including the holotype of margaritifer have slightly larger tubercles (see Fig. 22.9), whereas the largest of specimens including the holotype of huttoni have much larger tubercles. This interpreta- tion is in accord with Cooper’s (1982) suggestion that expansus and huttoni might be synonyms. The differences in definition of the axis between margaritifer and expansus noted by Wenndorf (1990) can be attributed to the flattening of the lectotype of expansus, axial convexity obscured by crushing and longti-tudinal folding of the tergite. Taking this deformation into account, whatever differences may have existed between the specimens do not appear to be significant. In the absence of significant differences in ornament or other features between the types of expansus and pygidia of margaritifer and huttoni, the species are regarded as conspecific. There is little to support the assignment of Homalonotus expansus to Burmeisteria. The most reliable character of Burmeisteria recognised in this work is the exceptionally high number of pygidial segments, with B. herschelii having up to 17 axial rings and 11 pleural furrows. H. expansus has only 12 axial rings and 9 pleural ribs. Reed (1918) emphasised the biconcave course of the rostral suture in herschelii as a generic Homalonotid trilobites from the Silurian and Lower Devonian 59 character. Cooper (1982) suggested the course of the rostral suture of herschelii to be variable and probably ontogenetic- ally related, supporting earlier doubts on its significance (Sdzuy, 1957, Saul, 1965). Notwithstanding this debate, the anterior portion of the cephalon (and the rostral suture) of the holotype of huttoni is not preserved. B. herschelii is strongly polymorphic, with variable glabellar morphology, but the glabella of expansus differs in having no trace of lobation or spines, and in that the axial furrows are subparallel and much shallower. In the absence of rostral morphology, the assignment of huttoni to Burmeisteria by various workers (Allan, 1935, Wenndorf, 1990) was presumably based on the similarity of the tuberculate ornament to that of herschelii. The arrange- ment of these tubercles suggests, however, they are not homo- logous, but independently derived. As noted by Cooper (1982), the tubercules of herschelii are arranged longitudinally. This arrangement differs from the transversely arranged tubercles on expansus. Tuberculation similar to that exhibited by Homalonotus expansus is known amongst members of Digonus. Wenndorf (1990) noted a close similarity in pygidial ornament between the paratypes of margaritifer and the ornatus group. Wenndorf added that margaritifer shared few other ornatus- group characters. The narrow pygidial axis is incompatible with assignment of expans a to Digonus. Generic assignment of Homalonotus expansus to Wenndorfia emphasises its low glabellar height, rounded anter- ior glabellar margin, posteriorly placed eyes, indistinct pygidial axial furrows and narrow axial width, and the contin- uity of the ring and pleural furrows, the termination of the axis at about 0.9 pygidial length, and the absence of a postaxial ridge. Its relationships to other species of Wenndorfia are diffi- cult to assess, but it differs from all other species in exhibiting a coarsely tuberculate thoracic and pygidial ornament. In over- all appearance, pygidia assigned to W. expansa most closely resemble the deeply furrowed pygidia of W. multicostata. The triangular pygidial outline, however, distinguishes it from multicostata and most other Wenndorfia, as discussed above. As the cephalic morphology of multicostata is poorly known, it is difficult to make a detailed comparison with expansa. The very low height and poor definition of the glabella of expansa is approached only by W. plana plana, although the latter differs in that the glabellar axial furrows are quite distinct posteriorly. Shirley (1938) recorded Homalonotus sp. from the Baton Formation at Baton River, east of Nelson, North Island, New Zealand. He considered ten crushed and deformed pygidia to be identical with pygidia of Wenndorfia expansa. Shirley’s description of the number of axial rings and pleural ribs and the continuity of the rings and pleurae are in accord with expansa, but as I have not examined these specimens I can make no judgement on their affinities. The Baton Formation is gener- ally considered younger than the Lankey Limestone at Reefton, although its age is variably cited as Pragian (Boucot et al., 1969) or Lochkovian (Wright, 1990). The newly documented material warrants a revised diagno- sis of the species, but otherwise is too limited to warrant a com- plete description to replace those of Hutton (1888), Allan (1935) and Wenndorf (1990). Figure 23. Geological sketch map of the Lily dale area showing Lochkovian-basal Pragian fossil localities yielding homalonotids. For other fossil localities see also Gill (1940, fig. 1, 1945, fig. 2), Moore (1965, fig.l), VandenBerg (1970), Garratt (1972), Wall et al. (1995, fig. 1), Sandford (2003, text-fig. 1A, 2004, fig. 1). 60 Andrew C. Sandford Environmental notes. The taphonomy of the population of Wenndorfia expansa (as a whole) are attributable to Speyer and Brett’s taphofacies 4B, with Ar=33% and the relative abundance of completely enrolled specimens (8%). W. expansa is interpreted to inhabit deep, outer-shelf environments. This suggested habitat is comparable to the deepest water facies associated with populations of many other species of Wenndorfia (see Wenndorf, 1990) Wenndorfia lilydalensis (Gill, 1949) Figures 5C, 24 Homalonotus harrisoni. — Cresswell, 1894: 156. Homalonotus sp. — Chapman, 1907: 239. — Gill, 1938: 170. Trimerus lilydalensis Gill, 1949: 69, pi. 8 figs 4-5, pi. 9 fig. 7, text-fig. IF — Holloway and Neil, 1982: 145. Type material. Holotype NMV P14587 (cephalon, figured Gill, 1949: pi. 8 figs 4-5, text-fig. IF, Fig. 24.1 herein, counterpart previously registered NMV P14588) from PL1801, Gill locality 1, Mooroolbark, Victoria. Paratype NMV PI 45 89 (pygidium, figured Gill, 1949: pi. 9 fig. 7) from PL1801. For locality see Fig. 23. Registered material. 57 specimens: 1 articulated dorsal exoskeleton, 1 disarticulated dorsal exoskeleton, 1 enrolled cephalothorax, 9 cephala, 7 cranidia, 3 librigenae, 12 thoracic segments, 23 pygidia. NMV P304527 from PL1805, Gill locality 5, Coldstream, Victoria. NMV P304531, P304532 from PL1990, Coldstream. NMV P304528- P304530 from PL1813, Gill locality 13, Mooroolbark. NMV P304524-P304526 from PL 1887, Gill locality 87, Mooroolbark. NMV P3204533, NMV P304570 from PL1869, Gill locality 69, Mooroolbark. NMV P304549-P304558 from PL1801. NMV P304567, P304569 from PL6661, Lilydale, Victoria. NMV P457, NMV P304559-P304562 from PL1802, Gill locality 2, Lilydale. NMV P3045 1 6-P3045 1 8 from PL1810, Gill locality 10, Lilydale. NMV P304522 from PL1858, Gill locality 58, Lilydale. NMV P304534- P304548, NMV P304566 from PL1859, Gill locality 59, Lilydale. NMV P304523 from PL 1862, Gill locality 62, Lilydale. NMV P304520, P304521 from PL 1829, Gill locality 29, Kilsyth, Victoria. NMV P304563-P304565 from “Lilydale”. NMV P304519 from “with- in 1 km of Mooroolbark Railway Station”. NMV P304571 probably from the Lilydale area. For localities see Fig. 23. Stratigraphic distribution. Humevale Siltstone, from 1750 m above the base of the unit up to a horizon 3700 m above the base of the unit, upper Boucotia australis Assemblage Zone to lower Boucotia loy-olensis Assemblage Zone, late Lochkovian-earliest Pragian. Diagnosis. Cephalon weakly pentangular in outline, with length 0.56 times width. Glabellar length equalling width, sides straight, weakly tapering (15°), strongly rounded anteriorly. Glabellar lobation weakly defined. Palpebral lobe placed with midline opposite 0.5 cranidial length. Preglabellar field (excluding rostral plate) 0.15 times cranidial length. Anterior cranidial margin evenly rounded, concentric with glabellar anterior margin. Dorsal surface of rostral plate crescentic in shape, length (sag.) 0.05 times cephalic length, anterior margin obtusely angular (130°). Ventral surface of rostral plate wide, kite-shaped, width 1.15 times length. Pygidium triangular, length 0.8 times width, tip obtusely pointed (-100°). Axis very wide anteriorly, width 0.66 times pygidial width, length 0.92 times pygidial length, sides straight. 9-11 axial rings, 8 pleural ribs, ring and pleural furrows deep. Surface of cephalon with fine tubercles, more distinct and densely distributed adjacent to the cephalic margin. Description. Exoskeleton with maximum size moderate (estimated 20 cm from NMV P304562). Cephalon with pentangular outline, sides converging forwards at 70-75° opposite glabella and at about 135° opposite preglabellar field, length about 0.57 times width. Glabella with length 0.75 times cephalic length, width 0.45 times cephalic width, sides converging at 13-17°, length ranging between 1.06-0.91 times width. Occipital ring length 0.1 times cranidial length, slightly wider medially. Occipital furrow deep on internal moulds, shallow on external moulds, weak forward flexure medially. Glabellar anterior margin well defined, with arc of curvature centred at 0.55 times glabellar length. Glabella weak- ly convex (tr., sag.). Glabellar lobation very weakly defined (best seen on NMV P304562 and NMV P304516, Figs 24.2, 24.4), LI 0.22 times glabellar length, L2 and L3 0.11 times glabellar length. S3 transverse, S2 directed inwards-backwards at about 7° from the transverse and S 1 at 15° from the transverse. Axial furrows wide, shallow posteriorly, moderately impressed anteriorly, paraglabellar areas poorly defined. Preglabellar field of uniform width, flat. Length (exsag.) of posterior border equal to occipital length abaxially, lengthening adaxially to 2.0 times occipital length. Posterior border furrow transverse, shallow and very wide, meeting lateral border furrow distally. Postocular fixigenal area short, length (exsag.) 0. 14 times cranidial length. Palpebral lobe 0.13 times cranidial 1.75 times preoccipital glabellar width, very shallow 6-6 length, palpebral furrow. Preocular fixigena area wide, 0.22 narrowing forwards anteriorly. Anterior branches of facial suture broadly curved, meeting connective sutures almost at margin opposite 0.9 cephalic length. Librigena with wide, moderately impressed border furrow, librigenal area weakly convex, steeply inclined, lateral border wide and convex, shallow, wide subocular furrow. Dorsal surface of rostral plate slightly inclined (in lateral view, see Figs 24.2c, 24.3c), length (exsag.) 0.18 times width (tr.). Ventral surface of rostral plate kite-shaped, with slightly medial flexure (tr.) giving anterior Figure 24. Wenndorfia lilydalensis (Gill, 1949). la, holotype NMV P14587, cephalon, dorsal view x 2.3 (latex cast) from PL1801. lb, same, oblique view x 2.5. lc, same, dorsal view x 1.8 (internal mould). 2a, NMV P304516, cephalon and displaced thoracopygon, dorsal view, enlarge- ment of anterior margin of cephalon x 5.0 (latex cast) from PL1810. 2b, same, dorsal view of cephalon x 1.5 (internal mould). 2c, same, oblique view of cephalon x 1.4 (latex cast). 2d, same, dorsal view of thoracopygon x 1.3 (latex cast). 3a, NMV P304536, cephalon, ventral view (dou- blure) x 2.2 (latex cast) from PL1859. 3b, same, dorsal view x 2.25. 3c, same, lateral view x 2.55. 4, NMV P304562, cephalon, dorsal view x 2.2 (internal mould) from PL1802. 5, NMV P304531, cephalon, ventral view (doublure) x 3.3 (latex cast) from PL1990. 6, NMV P304528, cephalon, dorsal view x 1.3 (latex cast) from PL1813. 7, NMV P304550, pygidium, dorsal view x 1.4 (internal mould) from PL1801. 8a, NMV P304565, pygidium, posterior view x 2.9 (latex cast) from “Lilydale” 8b, same, dorsal view. 8c, same, lateral view. 9, NMV P304534, pygidium, dorsal view x 2.2 (internal mould) from PL1859. 10, paratype NMV P14589, pygidium, dorsal view x 2.1 (internal mould) from PL1801. 11a, NMV P304549, pygidium, lateral view x 1.6 (internal mould) from PL1801. lib, same, dorsal view (latex cast). 12, NMV P304566, pygidium, ventral view (doublure) x 2.8 (internal mould) from PL1859. 13, NMV P457, pygidium, dorsal view x 1.5 (internal mould) from PL1802. 14, NMV P304550, pygidium, dorsoposterior view x 1.7 (internal mould) from PL 1801. Homalonotid trilobites from the Silurian and Lower Devonian 61 62 Andrew C. Sandford margin of cephalon a very weak M-shaped outline, flat posteriorly. Connective sutures weakly sinusoidal, diverging at 90° posteriorly and at 50° anteriorly. Posteriorly, connective sutures are either confluent immediately anterior to the meeting the hypostomal suture (NMV P304536, Fig. 24.3a), or meet the hypostomal suture separately (NMV P304531, Fig. 24.5). Thorax with 13 segments. On external moulds axial furrows and the pleural flexure are not defined, although the axis is marked on internal moulds by exsagittal line of deep depressions (axial articulating processes) on the posterior margin of each segment. Pleural furrows wide and deep, pleural tips rounded. Pygidium with sides weakly and uniformly convex. Axial furrows converging at about 45°, indistinct opposite first and second rings, shallow posteriorly. Axis poorly defined posteriorly without postaxial ridge. Pleural furrows not reaching margin, border poorly defined and lacking independent convexity from pleural field, without border furrow. In posterior view anterior margin strongly convex, posterior margin weakly arched to accommodate rostral flexure. In lateral view dorsal margin steep (-45°) and gently curved, ventral margin more or less horizontal. Discussion. Trimerus lilydalensis exhibits several features indicative of assignment to Wenndorfia, including the strongly rounded anterior glabellar margin and the concentric anterior cranidial margin, the antero-medial rostral node, the weak pygidial trilobation, and the termination of the pygidial axis at about 0.9 pygidial length. The species was originally described from a limited number of poorly preserved specimens, but is redescribed from a much larger population. Gill’s (1949) description of the species includes several of the diagnostic characters listed above, although the revised diagnosis differs in a number of points. Gill described the cephalon as being markedly inflated (it is only moderately convex, see Fig. 24.2c), and the exoskeleton as being smooth (it bears fine tubercules and pits, see Fig. 24.2a). Holloway and Neil (1982) considered specimens of Digonus wenndorfi from the Heathcote area to be possibly conspecific with Wenndorfia lilydalensis. W. lilydalensis can be easily distinguished from wenndorfi in having pygidia lacking a raised postaxial ridge, and cranidia with a U-shaped rather than quadrate anterior margin, a more weakly tapered glabella with a more rounded anterior margin, and a forwardly convex rather than transverse rostral suture. Wenndorfia lilydalensis is most closely comparable to the contemporary W. bostoviensis from the upper Lochkovian of Poland. The species share a similar trapezoid, moderately tapering, straight-sided glabellar outline and a very wide (tr.) palpebral area. Glabellar lobation is variably expressed in lilydalensis, and although Tomczykowa (1975) noted an absence of glabellar lobation for bostoviensis, shallow SI -S3 can be seen on specimens considered here to be con- specific with bostoviensis (see discussion above). More significantly, bostoviensis also exhibits a relatively underived pygidium, with a triangular pygidial outline with an acutely angled tip. Pygidial axial furrows are moderately impressed on bostoviensis, distinguishing it from lilydalensis and younger species of Wenndorfia, that share very shallow to effaced pygidial axial furrows. The only other Wenndorfia to exhibit a triangular pygidial outline is the lower Pragian Polish W. nova, with less elongate pygidial proportions and a very narrow pygidial axis. In cranidial morphology nova shows some resemblance to lilydalensis, although the former differs in having narrow palpebral areas. Wenndorfia lilydalensis is confined to the Lilydale sequence. One specimen labelled as coming from the Seville area, collected by F. Chapman, was found to have a counterpart labelled as coming from the Lilydale area, collected by J. Jutson. The lithology of the specimen matches lithologies from the Lilydale sequence rather than those from Seville. Gill’s (1938) record of the species from PL 1835, Gill locality 35, Seville is also probably erroneous, as the lithology of the specimen is unlike that at PL1835, but again closely matches lithologies of the Lilydale sequence. Environmental notes. The faunal associations of Wenndorfia lilydalensis are complex. W. lilydalensis first appears as a rare element in the trilobite fauna (relative abundance 2%), but through its range steadily increases in relative abundance, reaching maximum relative abundance in the upper horizons of its range. In the lower part of its range lilydalensis frequently occurs in phacopid-dominated assemblages, but in the upper parts of its range lilydalensis is frequently part of a distinct acastid/homalonotid dominated association With the exception of a complete exoskeleton from a bio- clastic coquinal sandstone at PL 1805 (taphofacies Till), all specimens of Wenndorfia lilydalensis are found in a siltstone lithology preserved as isolated sclerites or partly disarticulated specimens including cephala, a cephalon with 7 thoracic segments, and a moult assemblage, being a cephalon lying obliquely over the thorax. A cranidium and nearby pygidium on the same bedding plane might also be interpreted as a moult assemblage (taphofacies TIV). An outer shelf setting is sug- gested as the preferred environment of W. lilydalensis, which ranges only as a rare faunal element to mid shelf settings and to even deeper settings where it occurs with a calymenid- phacopid association. The taphonomy of the latter, with a high degree of articulation and containing sparsely distributed moult assemblages and rare enrolled specimens represents a new taphofacies for central Victoria, designated taphofacies TV, equivalent to Speyer and Brett’s (1986) taphofacies 4B (see Fig. 6). Acknowledgements I extend my sincere thanks to Dr Alan Thomas (University of Birmingham) and to an anonymous referee for their most detailed and helpful reviews of the manuscript, and for alerting me to a number of publications on British homalonotids which I had overlooked. Research for this work was funded from an Australian Postgraduate Award. I gratefully acknowledge Dr David Holloway (Museum Victoria), Dr Malcolm Wallace (University of Melbourne) and Dr Stephen Gallagher (University of Melbourne) for their supervision of this project. Mr Peter Duncan (Eden Park) kindly donated specimens of Trimerus ( Ramiotis ) otisi collected from his property. I also thank Dr Norton Hiller, Canterbury Museum, Christchurch, New Zealand for the loan of specimens of Wenndorfia expansa including the holotype of Homalonotus ( Burmeisteria ) huttonv. Homalonotid trilobites from the Silurian and Lower Devonian 63 and Mr Robert Jones at the Australian Museum for the loan of a well-preserved cranidium of Trimerus ( Trimerus ) harrisoni , regretfully lost. References Alberti, G.K.B. 1970. Trilobiten des jiingeren Siluriums sowie des Unter- und Mitteldevons. 2. Abhandlungen der Senckenbergischen Naturforschenden Gesellschaft 525: 1-233. Allan, R.S. 1935. The fauna of the Reefton beds (Devonian) New Zealand, with notes on Lower Devonian animal communities in relation to the base of the Devonian System. New Zealand Geological Survey Palaeontological Bulletin 14: 1-72. Angelin, N.R 1852. Paleontologica Scandinavica. Lundae. Angelin, N.R 1878. Paleontologica Scandinavica. P. 1. Crustacea Formationis Transitions Fascicula I and II. Samson and Wallin: Holmiae. Balashova, E.A. 1966. [Trilobites from the Ordovician and Silurian beds of Pamir], Trudy Upravlenia Geologii. Sovieta Ministrov Tadzhiskoi 2: 191-262 (in Russian). Baldis, B.A., Benedetto, L., Blasco, G. and Martel, M. E. 1976. Trilobites Silurico-Devonicos de la Sierra de Zapla (Nordeste de Argentina). Ameghiniana 13: 185-225. Barrois, C. 1886. Aper§u de la constitution geologique de la rade de Brest. Bulletin de la Societe Geologique de France (3) 14: 678-707. Boucot, AJ. 1975. Evolution and Extinction Rate Controls. Elsevier: Amsterdam. 427 pp. Boucot, A.J. and Johnson, J.G. 1967. Paleogeography and correlation of Appalachian Province Lower Devonian sedimentary rocks. Tulsa Geological Society Digest 3:1-53. Boucot, A.J., Johnson, J.G. and Talent, J.A. 1969. Early Devonian brachiopod zoogeography. Geological Society of America Special Paper 119: 1-113. Brett, C.E. 1983. Sedimentology, facies and depositional environments of the Rochester Shale (Silurian, Wenlockian) in western New York and Ontario. Journal of Sedimentary Petrology 53: 947-971. Burmeister, H. 1843. Die Organisation die Trilobiten aus ihren leb- enden Verwandten entwickelt; nebst einer systematischen ubersicht aller seither beschriebenen Arten, Reimer: Berlin. 147 pp. Chapman, E.J. 1890. Some remarks on the classification of trilobites as influenced by stratigraphical relations: with outline of a new group- ing of these forms. Transactions of the Royal Society of Canada 1: 113-120. Chapman, F. 1907. On the occurrence of Yeringian fossiliferous mud- stones at Croydon. Victorian Naturalist 23: 237-239. Chapman, F. 1908. Notes on a collection of Silurian fossils from the Whittlesea district made by Mr J. T. Jutson. Proceedings of the Royal Society of Victoria 21: 218-225. Chapman, F. 1912. New or little known Victorian fossils in the National Museum. Part 14. On some Silurian trilobites. Proceedings of the Royal Society of Victoria 24: 293-300. Chapman, F. 1915. New or little known Victorian fossils in the National Museum. Part 18. Some Yeringian trilobites. Proceedings of the Royal Society of Victoria 28: 157-171. Chatterton, B.D.E., Johnston, B.D. and Campbell, K.S.W. 1979. Silicified Lower Devonian trilobites from New South Wales. Palaeontology 22: 799-837. Chlupac, I. 1981. Homalonotid trilobites from the metamorphic Devonian of the Hraby jesernk Mts, Czechoslovakia. Casopsis pro Mineralogii a Geologii 26: 361-370. Chlupac, I. 1983. Trilobite assemblages in the Devonian of the Barrandian area and their relationships to palaeoenvironments. Geologica et Palaeontologica 17: 45-73. Chlupac, I. 1987. Ecostratigraphy of Silurian trilobite assemblages of the Barrandian area, Czechoslovakia. Newsletter Stratigraphy 17(3): 169-186. Chlupac, I., Feist, R. and Morzadec, P. 2000. Trilobites and standard Devonian stage boundaries. Pp. 87-98 in: Bultynck, P. (ed), Subcommision on Devonian Stratigraphy. Fossil groups important for boundary definition. Courier Forschungsinstitut Senckenberg 220. Clarke, J.M. 1890. As trilobitas do Grez de Erere e Maecuru, Estado do Para, Brazil. Archivos do Museu Nacional do Rio de Janeiro 9 (Revista 1): 1-58. Clarke, J.M. 1913. Fosseis Devonianos do Parana. Monographias do Servigo Geologico e Mineralogico do Brasil 1: 1-353. Cooper, M.R. 1982. A revision of the Devonian (Emsian-Eifelian) Trilobita from the Bokkeveld Group of South Africa. Annals of the South African Museum 89: 1-174. Cresswell, A.W. 1894. Additional notes on the Lilydale Limestone. Proceedings of the Royal Society of Victoria 6: 156-160. Curtis, N.J. and Lane, P.D. 1998. The Llandovery trilobites of England and Wales. Part 2. Monograph of the Palaeontographical Society, London 152: 51-101. Dahmer, G. 1914. Ein Hautungsplatz von Homalonotus gigas A. Rom. im linksrheinischen Unterdevon. Jahrbiicher des Nassauischen Vereins fur Naturkunde 67: 16-21. Dawson, J.W. 1868. Acadian Geology. The geological structure, organic remains and mineral resources of Nova Scotia, New Brunswick and Prince Edward Island, (second Edition). MacMillan and Co.: London. 694 pp. Dawson, J.W. 1877. Note on two Palaeozoic crustaceans from Nova Scotia. Geological Magazine 2(4): 56-58. Edgecombe, G.D. 1993. Silurian acastacean trilobites of the Americas. Journal of Paleontology 67: 535-548. Edgecombe, G.D. andFortey, R.A. 2000. Silurian trilobites from the El Carmen Formation, Bolivia. Senckenbergiana Lethaea 79: 329-355. Foerste, A.F. 1919. Silurian fossils from Ohio, with notes on related species from other horizons. Ohio Journal of Science 19: 367-404. Fortey, R.A. and Owens, R.M. 1997. Evolutionary history. Pp. 249-287 in: Whittington et al. (eds), Treatise on Invertebrate Palaeontology, Part O. Arthropoda 1. Trilobita (revised). Volume I: Introduction, Order Agnostida, Order Redlichiida. Geological Society of America and University of Kansas Press: Lawrence. Foster, D.A., Gray, D.R. and Bucher, M. 1999. Chronology of de formation within the turbidite-dominated, Lachlan orogen: implica- tions for the tectonic evolution of eastern Australia and Gondwana. Tectonics 18: 452^485. Garratt, M.J. 1968. Studies in Devonian Faunas. Part 1, Devonian spiriferids from the Fitzroy Basin, Western Australia. Part 2, Biostratigraphy of the Merriang Syncline, Eden Park. MSc thesis. University of Melbourne. Garratt, M.J. 1972. Yan Yean 1:63,000 geological map. Department of Mines: Melbourne. Garratt, M.J. 1977. Kinglake 1:63,000 geological map. Department of Mines: Melbourne. Garratt, M.J. 1983. Silurian and Devonian biostratigraphy of the Melbourne Trough, Victoria. Proceedings of the Royal Society of Victoria 95: 77-98. Gill, E.D.1938. Yeringian trilobites. Victorian Naturalist 54: 167-171. Gill, E.D. 1940. The Silurian Rocks of Melbourne and Lilydale: a discussion of the Melbournian- Yeringian boundary and associated problems. Proceedings of the Royal Society of Victoria 52: 249-261. Gill, E.D. 1945. Trilobita of the Family Calymenidae from the Palaeozoic rocks of Victoria. Proceedings of the Royal Society of Victoria 56: 171-186. 64 Andrew C. Sandford Gill, E.D. 1947. Anew trilobite from the Yeringian (Lower Devonian) Rocks of Kinglake, Victoria. Proceedings of the Royal Society of Victoria 59: 8-19. Gill, E.D. 1948. A genus of dalmanitid trilobites. Journal and Proceedings of the Royal Society of New South Wales 82: 16-24. Gill, E.D. 1949. Some Australian homalonotid trilobites. Proceedings of the Royal Society of Victoria 61: 61-74. Gill, E.D. 1950. Preliminary account of the formation of the Zeehan area, Tasmania: Palaeontology and Palaeoecology of the Eldon Group. Papers and Proceedings of the Royal Society of Tasmania 1949:231-258. Gill, E.D. and Banks, M.R. 1950. Silurian and Devonian stratigraphy of the Zeehan area, Tasmania. Papers and Proceedings of the Royal Society of Tasmania 1949: 259-277. Gosselet, J., Barrois, C., Leriche, M. and Crepin, A. 1912. Description de la faune siluro-devonienne de Lievin. Memoires de la Societe de Geologie Nord 6: 1-67. Green, J. 1832. A monograph of the trilobites of North America, with coloured models of the species. Joseph Brano: Philadelphia. 94 pp. Giirich, G. 1909. Leitfossilien, 2nd. Lieferung. Leitfossilien des Devonian Gebriider Bomtrager: Berlin, p. 97-199. Haas, W. 1968. Trilobiten aus dem Silur und Devon des Bithynien (NW-Turkei). Palaeontographica, AbteilungA 130: 60-207. Hall, J. 1859. Palaeontology of New York: Volume 3, containing descriptions and figures of the organic remains of the Lower Helderberg group and the Oriskany sandstone. New York State Geological Survey: New York. 532 pp., plates published separately in 1861. Hall, J. 1860. Descriptions of new species of fossils from the Silurian rocks of Nova Scotia. Canadian Naturalist and Geologist 5: 144-159. Hall, J. and Clarke, J.M.. 1888. Palaeontology of New York, Volume 7. Descriptions of the trilobites and other Crustacea of the Oriskany, Upper Helderberg, Hamilton, Portage and Chemung, and Catskill groups. New York State Geological Survey: New York. 236 pp. Hartt, C.F. and Rathbun, R. 1875. Morgan expeditions: 1870-1871. On the Devonian trilobites and mollusks of Erere, province of Para, Brazil. Annals of the Lyceum of Natural History, New York 11: 110-127. Hector, J. 1876. On a new trilobite ( Homalonotus expansus). Transactions of the New Zealand Institute 9: 602; pi. 27; fig.2, p. 474. Hennig, W. 1965. Phylogenetic Systematics. Annual Review of Entomology 10: 97-116. Holloway, DJ. and Neil, J.V. 1982. Trilobites from the Mount Ida Formation (Late Silurian-Early Devonian), Victoria. Proceedings of the Royal Society of Victoria 94: 133-154. Holloway, D.J and Sandford, A.C. 1993. An Early Silurian trilobite fauna from Tasmania. Memoirs of the Australian Association of Palaeontologists 15: 85-102. Hutton, F.W. 1888. On a trilobite from Reefton, New Zealand, new to Australasia. Proceedings of the Linnaean Society of New South Wales 2(2): 257-258. Jell, P.A. 1992. Middendorps quarry: a classic Australian Early Devonian fossil site. The Fossil Collector 38: 7-16 Jell, P.A. and Holloway, D.J. 1983. Devonian and Late? Silurian palaeontology of the Winneke Reservoir site, Christmas Hills, Victoria. Proceedings of the Royal Society of Victoria 95: 1-21. Jutson, J.T. 1908. The Silurian rocks of the Whittlesea district. Proceedings of the Royal Society of Victoria 21: 211-217. Kegel, W. 1927. Homalonotus asturco n. sp. aus dem Unterdevon Asturiens. Centralblatt fiir Mineralogie, Geologie und Palaeontologie. Abteilung B 1927. 8: 334-340. Kennedy, R. J. 1994. British Devonian Trilobites. Part 1. Monograph of the Palaeontographical Society, London 148: 1- 33. Koch, C. 1880. Uber das Vorkommen von Homalonotus - Arten im rheinischen Unterdevon. S itzungsberichte der Naturhistorie Vereins Rheinland und We stphalen. [fur 1880]: Bonn. Koch, C. 1883a. Monographic der Homalonotus - Arten des rheinischen Unterdevon. Koniglich Preussischen geologischen Landesanstalt, . H. Neumann: Berlin, xxxi + 85 pp (plates published separately: Koch, 1883b). Koch, C. 1883b. Atlas von acht lithographischen Tafeln zur Monographic der Homalonotus- Arten des rheinischen Unterdevon. Abhandlungen zur geologischen Specialkarte von Preussen und den Thuringischen Staaten, Berlin 4 (2): pis 1-8. Konig, C.E. 1825. leones fossilium sectiles. Centuria Prima: London. Pp. 1-4, pis 1-8. Koninck, L.G. de. 1876. Notice sur les quelques fossiles, recueillis par G. Dewalque dans le systeme Gedinnien de A. Dumont. Annales de la Societe Geologique Belgique 3: 25-52. Kozlowski, R. 1923. Faune devonienne de Bolivie. Annales de Paleontologie 12: 1-112. Le Menn, J., Plusquellec, Y., Morzadec, P. and Lardeux, H. 1976. Incursion Hercynienne dans les faunes Rhenanes du Devonien inferieur de la rade de Brest (Massif Armoricain). Palaeonto- graphica Abteilung A 1 53: 1-61. McCoy, F. 1876. Prodromus of the Palaeontology of Victoria; or, figures and descriptions of Victorian organic remains. Decade 3. Geological Survey of Victoria: Melbourne. 40 pp. McLeam, F.H. 1924. Palaeontology of the Silurian rocks of Arisaig, Nova Scotia. Geological Survey of Canada, Memoir 137. Mendez- Alzola, R. 1938. Fossiles devonianos del Umguay. Boletin del Instituto Geologico del Uruguay 24: 1-115. Mikulic, D.G. 1999. Silurian trilobite associations in North America. Pp. 793-798 in: Boucot, A.J. and Lawson, J.D. (eds), Palaeo- communities: a case study from the Silurian and Lower Devonian. Cambridge University Press: Cambridge. Moberg, J.C. and Gronwall, K.A. 1909. Om Fyledalens Gotlandium. Lunds Universitets Arsskrift, Neue Folge 2: 1-86. Moore, B.R. 1965. The structure and stratigraphy of the Siluro-Devonian sediments of the middle Yarra basin, central Victoria. Proceedings of the Royal Society of Victoria 79: 205-213. Morris, S.F. 1988. A review of British trilobites, including a synoptic revision of Salter’s Monograph. Monograph of the Palaeonto- graphical Society, London 140: 316 pp. Morzadec, P. 1969. Le Devonien de la Rive Nord de la Riviere du Faou (Finistere), etude stratigraphique, etude des trilobites. Bulletin de la Societe Geologique et Mineralogique de Bretagne 1968: 1-52. Morzadec, P. 1976. Les trilobites. Pp. 285-294 in: Babin, C. et al. (eds), Les schistes et calcaires eodevoniens de Saint-Cenere ( Massif armoricain, France). Memoire Societe Geologique et Mineralogique de Bretagne 19. Morzadec, P. 1981. Les trilobites. Pp. 389-294 in: Babin, C. et al. (eds), Les schistes et calcaires de TArmorique ( Devonien inferieur. Massif Armoricain. Memoire Societe geologique et mineralogique de Bretagne 23. Morzadec, P. 1986. Trilobites. Pp. 185-200 in: Racheboeuf, PR. (ed), Le Groupe de Lievin. Pndoli-Lochkovian de T Artois (N. France ). Biostratigraphie du Paleozoique 3. Murchison, R.I. 1839. The Silurian System, founded on geological researches in the counties of Salop, Hereford, Radnor, Montgomery, Caermarthen, Brecon, Pembroke, Monmouth, Worcester, Gloucester and Stafford; with descriptions of the coalfields and overlying formations. John Murray: London, xxxii + 768 pp. Homalonotid trilobites from the Silurian and Lower Devonian 65 Opik, A. A. 1953. Lower Silurian fossils from the ‘ Illaenus Band’ Heathcote, Victoria. Memoirs of the Geological Survey of Victoria 19: 1-42. Owens, R.M. 1994. The trilobites of the Bristol district. Proceedings of the Bristol Naturalists’ Society 54: 67-84. Pillet, J. 1961a. Contribution a T etude des Homalonotidae (Trilobites) des Gres a Dalmanella monnieri (Siegenian inferieur) du Massif armoricain. Bulletin de la Societe Geologique de France 7: 457-462. Pillet, J. 1961b. Contribution a l’etude des faunes de trilobites du Zemmour (Mauri tanie septentrionale). Annales de la Faculte des Sciences, Universite de Dakar 6: 93. Pillet, J. 1973. Les trilobites du Devonien inferieur et du Devonien moyen du Sud-Est du Massif armoricain. Memoires de la Societe desEtudes Scientifiques d’ Anjou 1: 1-307. Pillet, J. and Waterlot, M. 1982. Decouverte d’une nouvelle espece de trilobite (Homalonotidae) dans les Gres de Vireux (Ardennes, France). Annales de la Societe Geologique du Nord 52: 183-185. Prantl, F. and Pribyl, A. 1948. Roztradena ceskych Homalonotidu (Trilobitae). Rozpravy II Trady Cesku Akademie 58: 1-23. Quenstedt, F. 1852. Handbuch der Petrefaktenkunde. Second Edition. H. Laupp: Tubingen. Reed, F.R.C. 1918. Notes on the genus Homalonotus. Geological Magazine (6)5: 263-276, 314-327. Reed, F.R.C. 1920. Notes on the fauna of the Lower Devonian beds of Torquay. Geological Magazine 299-306, 341-347. Reed, F.R.C. 1925. Revision of the fauna of the Bokkeveld beds. Annals of the South African Museum 22: 27-225. Renaud, A. 1942. Le Devonien du synclinorium median Brest-Laval. 2e fasc: Paleontologie. Memoire de la Societe Geologique et Mineralogique de Bretagne 7: 1-436. Richter, R. and Richter, E. 1926. Die Trilobiten des Oberdevon. Beitrage zur Kenntnis devonischer Trilobiten. Abhandlungen der Preuszischen Geologischen Landesanstalt. Neue Folge 99: 1-314. Richter, R. and Richter, E. 1932. Unterlagen zum Fossilium Catalogs VI. Senckenbergiana 14: 359-37 1 . Richter, R. and Richter, E. 1943. Studien im Palaozoikum der Mittelmeer-Lander. 4a. Trilobiten aus dem Devon von Morokko; mit einem Anhang fiber Arten des Rheinlands. Senckenbergiana 26: 116-199. Rickards, R.B. and Sandford, A. C. 1998. Llandovery-Ludlow grapto- lites from central Victoria: new correlation perspectives of the maior formations. Australian Journal of Earth Sciences 45: 743-763. Roemer, F.A. 1843. Die Versteinerungen des Harzgebirges. Hahn: Hannover. 40 pp. Rouault, M. 1855. Notice sur quelques especes de fossiles du terrain devonien du Nord du department de la Manche. Bulletin de la Societe Geologique de France 12: 1040-1045. Salter, J.W. 1851. Appendix, in: Sedgwick, A. and McCoy, F. A Synopsis of the Classification of the British Palaeozoic Rocks With a Systematic Description of the British Palaeozoic Fossils in the Geological Museum of the University of Cambridge. John W. Parker and Son: London. 550 pp. Salter, J.W. 1861. On the fossils from the high Andes, collected by David Forbes esq., FRS FGS. Quarterly Journal of the Geological Society of London 17: 62-73. Salter, J.W. 1864. A monograph of the British trilobites from the Cambrian, Silurian and Devonian formations. Part 1 . Monograph of the Palaeontographical Society, London for 1862: 1-80. Salter, J.W. 1865. A monograph of the British trilobites from the Cambrian, Silurian and Devonian formations. Part 2. Monograph of the Palaeontographical Society, London for 1863: 81-128. Sandberger, G. and Sandberger, F. 1849. Die Versteinerungen des Rheinischen Schichtensystems in Nassau. Mit einer kurzgefassten Geognosie dieses Gebietes und mit steter Beriicksichtigung analoger Schichten anderer Lander. Kreidel, Niedner: Wiesbaden, i-xv, 1-564 pp. Sandford, A.C. and Holloway, D.J. 1998. The effaced styginid trilobite Thomastus from the Silurian of Victoria, Australia. Palaeontology 41: 913-928. Sandford, A.C. 2000. Trilobite faunas and palaeoenvironmental setting of the Silurian (early Ludlow) Melbourne Formation, central Victoria. Alcheringa 24: 153-206. Sandford, A.C. 2002. Systematics, biostratigraphy and palaeoenviron- ments of Echidnops, a new genus of trilobite from the Late Silurian- Early Devonian of south-eastern Australia: Phacopidae of Victoria, Part 1. Memoirs of the Association of Australasian Palaeontologists 27: 1-31. Sandford, A.C. 2003. A revision of Nephranomma Erben, 1952 (Trilobita, Phacopidae), with new species from the Early Devonian of Victoria, Australia. Phacopidae of Victoria, Part 2. Special Papers in Palaeontology 70: 309-329. Sandford, A.C. 2004. A revision of Lochkovella Chlupac, 1972 (Arthropoda: Trilobita) and a species from Lilydale and Tyers. Phacopidae of Victoria, Part 5. Proceedings of the Royal Society of Victoria 116: 209-222. Sandford, A.C. in press. Systematics, palaeoenvironments and strati- graphy of the Silurian trilobite Dalmanites wandongensis Gill, 1948 and its bearing on the structural geology of the Kilmore area, Victoria. Alcheringa. Saul, J.M. 1965. Trilobita. Pp. 269-271 in: Doumani, G.A., Boardman, R.S., Rowell, A.J., Boucot, A.J., Johnson, J.G., Lee-McAlester, A., Saul, J., Fisher, D.W. and Miles, R.S. Lower Devonian Fauna of the Horlick Formation, Ohio Range, Antarctica Pp. 241-280 in: Hadley, J.B. (ed), Geology and Paleontology of the Antarctic. Antarctic Research Series 6, American Geophysical Union, Publication 1299. Saul, J.M. 1967. Burmeisteria (Digonus) accraensis, a new homa- lonotid trilobite from the Devonian of Ghana. Journal of Palaeontology 41: 1126-1136. Schraut, G. 2000. Trilobiten aus dem Unter-Devon des siido- stlichen Anti-Atlas, Slid Marokko. Senckenbergiana Lethaea 79: 361^133. Sdzuy, K. 1957. Bemerkungen zur Familie Homalonotidae, mit der Bescreibung einer neuen Ait von C alymenella. Senckenbergiana Lethaea 38: 275-290. Sdzuy, K. 1959. Family Homalonotidae E.J. Chapman, 1890. Pp. 454-0461 in: R. C. Moore (ed.), Treatise on Invertebrate Paleontology, Part O, Arthropoda 1. Geological Society of America and University of Kansas Press: Lawrence. Shergold, J. H. 1968. On the occurrence of the trilobite genera Acaste and Acastella in Victoria. Proceedings of the Royal Society of Victoria 81: 19-30. Shimer, H.W. and Shrock, R.R. 1944. Index Fossils of North America. John Wiley and Sons: New York, London. 837 pp. Shirley, J. 1938. The fauna of the Baton River Beds (Devonian), New Zealand. Quarterly Journal of the Geological Society of London 94: 459-506. Speden, I.G. and Keyes, I.W. 1981. Illustrations of New Zealand fossils. New Zealand Department of Scientific and Industrial Research, Information Series, Wellington 150. Speyer, S.E. and Brett, C.E. 1986. Trilobite taphonomy and Middle Devonian taphofacies. Palaios 1: 312-327. Swartz, C.K. andProuty, W.F. 1923. Trilobita. Pp. 712-714 in: Silurian. Maryland Geological Survey, John Hopkins Press: Maryland. 66 Andrew C. Sandford Swinnerton, H.H. 1915. Suggestions for a revised classification of trilobites. Geological Magazine 6(2): 487-496, 538-545. Talent, J.A. 1964. The Silurian and Early Devonian Faunas of the Heathcote district, Victoria. Memoirs of the Geological Survey of Victoria 26: 1-55. Tchernycheva, N. 1937. [Silurian and Devonian trilobites of Mongolia], Akademiya Nauk SSSR, Trudy Mongol’ skoi Komiteta 28: 1-31 (in Russian). Temple, J.T. 1975. Early Llandovery trilobites from Wales with notes on British Llandovery calymenids. Palaeontology 18: 137-159. Thomas, A.T. 1977. Classification and phylogeny of the homalonotid trilobites. Paleontology 20: 159-178. Thomas, A.T. 1979. Trilobite associations in the British Wenlock. Pp. 447-451 in: Harris, A.L., Holland, C.H. and Leake, B.E. (eds), The Caledonides of the British Isles-reviewed. Special Publications of the Geological Society of London 8. Thomas, A.T, Owens, R.M. and Snajdr, M. 1989. Trilobites. Pp. 248-251 in: Holland, C.H. and Bassett, M.G. (eds), A global standard for the Silurian System. National Museum of Wales, Geological Series 9. Thomas, D.E. 1940a. Heathcote, County of Dalhousie, 40 chains to 1 inch geological map. Mines Department of Victoria: Melbourne. Thomas, D.E. 1940b. Redcastle, Counties of Dalhousie and Rodney, 40 chains to 1 inch geological map. Mines Department of Victoria: Melbourne. Thomas, D.E. 1941. Dargile, County of Rodney, 40 chains to 1 inch geological map. Mines Department of Victoria: Melbourne. Thomas, D.E. 1956. Heathcote: geological and topographic map. Mines Department of Victoria: Melbourne. Thomas, D.E. 1960. Lancefield 1:31,680 geological map. Geological Survey of Victoria: Melbourne. Thomas, I. 1905. Neue Beitrage zur Kenntnis der devonischen Fauna Argentiniens. Zeitschrift der Deutschen Geologischen Gesellschaft 57: 233-290. Tomczykowa, E. 1975. The trilobite subfamily Homalonotinae from the Upper Silurian and Lower Devonian of Poland. Acta Palaeontologica Polonica 20: 3-46. Tomczykowa, E. 1978. Anew specific name for Tomczykowa, 1975, and the taxonomic value of Homalonotus lobatus Prouty, 1923. Acta Palaeontologica Polonica 23: 195-198. Tromelin, G. and Lebesconte, P. 1876. Observations sur les terrains primaires du Nord de department d’llle-et-Vilaine et de quelques parties du Massif Breton. Bulletin de la Societe Geologique de France (3) 4: 583-623. VandenBerg, A.H.M. 1970. Ringwood 1:63000 Geological Map. Department of Mines: Melbourne. VandenBerg, A.H.M. , Willman, C.E., Maher, S., Simons, B.A., Cayley, R.A., Taylor, D.H., Morand, V.J., Moore, D.H. and Radojkovic, A. 2000. The Tasman Fold Belt System in Victoria. Geological Survey of Victoria Special Publication: Melbourne. Vemeuil, E. de. 1850. Note sur les fossiles devoniens du distrikt de Sabero (Leon). Bulletin de la Societe Geologique de France 2(7): 155-186. Vietor, W. 1919. Der Koblenzquartzit, seine Fauna, Stellung und linksrheinishe Verbreitung. Jahrbuch Preussischen Geologischen Landesanstalt. Neue Folge: 317-346. Waisfeld, B.G. and Sanchez, T.M. 1993. Trilobites Siluricos de la Formation Lipeon en el Noroeste Argentino (Sierra de Zapla, Provincia de Jujuy). Ameghiniana 30: 77-90. Wall, R., Mawson, R., Talent, J.A. and Cooper, B.J. 1995. Conodonts from an environmentally hostile context, the Lilydale Limestone (Early Devonian: Pragian) of central Victoria. Courier Forschungsinstitut Senckenberg 182: 371-387. Wenndorf, K. 1990. Homalonotinae (Trilobita) aus dem Rheinischen Unter-Devon. Palaeontographica Abteilung A 211: 1-184. Whitfield, R.P 1881. Notice of a very large species of Homalonotus from the Oriskany Sandstone Formation. Bulletin of the American Museum of Natural History 6: 193-5. Whittington, H.B. 1993. Morphology, anatomy and habits of the Silurian homalonotid trilobite Trimerus. Memoirs of the Association of Australasian Palaeontologists 15: 69-83. Whittington, H.B. 1997. Morphology of the exoskeleton. Pp. 1-85 in: Kaesler, R. L. (ed), Treatise on Invertebrate Palaeontology. Part O. Arthropoda 1. Trilobita, revised. Volume 1: Introduction, Order Agnostida, Order Redlichiida. University of Kansas and the Geological Society of America: Boulder, Lawrence. Whittington, H.B. and Kelly, S.R.A. 1997. Morphological terms applied to Trilobita. Pp. 313-329 in: Kaesler, R.L. (ed). Treatise on Invertebrate Palaeontology. Part O. Arthropoda 1. Trilobita, revised. Volume 1: Introduction, Order Agnostida, Order Redlichiida. University of Kansas and the Geological Society of America: Boulder, Lawrence. Williams, G.E. 1964. The geology of the Kinglake district, central Victoria. Proceedings of the Royal Society of Victoria 77: 213—321. Williams, H.S. and Breger, C.L. 1916. The fauna of the Chapman Sandstone of Maine including descriptions of some related species from the Moose River Sandstone. Professional Papers of the United States Geological Survey 89: 1-347. Wolf art, R. 1961. Stratigraphie und Fauna des alteren Palaozoikums (Silur/Devon) in Paraguay. Geologisches Jahrbuch 78: 29-102. Wolfart, R. 1968. Die Trilobiten aus dem Boliviens und ihre Bedeutung fiir Stratigraphie und Tiergeographie. Beihefte zum Geologischen Jahrbuch 74: 5-201. Wolfart, R, Bender, F. and Stein, V. 1968. Stratigraphie und Fauna des Ober-Ordoviciums (Caradoc-Ashgill) und Unter-Silurs (Unter Landovery) von Siidjordanien. Geologisches Jahrbuch 85: 517-564. Woodward, H. 1882. Additional notes on Homalonotus from the Devonian. Geological Magazine 9: 157-158. Wright, A.J. 1990. Acastid trilobites from the Baton Formation (Early Devonian), New Zealand. New Zealand Journal of Geology and Geophysics 33: 49-53. Memoirs of Museum Victoria 62(1): 67-89 (2005) ISSN 1447-2546 (Print) 1447-2554 (On-line) http://www.museum.vic.gov.au/memoirs/index.asp Pliocene marine mammals from the Whalers Bluff Formation of Portland, Victoria, Australia Erich M.G. Fitzgerald School of Geosciences, Monash University, Vic. 3800, Australia and Museum Victoria, G.P.O. Box 666, Melbourne, Vic. 3001, Australia (efitzger@museum.vic.gov.au) Abstract Fitzgerald, E.M.G. Pliocene marine m a mmals from the Whalers Bluff Formation of Portland, Victoria, Australia. Memoirs of Museum Victoria 62(1): 67-89. The most diverse and locally abundant Australian fossil marine mammal assemblages are those from late Neogene (Late Miocene through Late Pliocene) sediments in Victoria and Flinders Island, Tasmania. However, none of these assemblages have hitherto been described. The Pliocene (>2.5^1. 8 Ma) Whalers Bluff Formation, exposed in beach cliff sections and offshore reefs, at Portland, western Victoria (38°19'S, 141°38'E) has yielded a small but moderately diverse assemblage of marine mammals represented by fragmentary material. Taxa present include: right whales (Balaenidae); rorqual whales (Balaenopteridae); a physeterid similar to the extant sperm whale (cf. Physeter sp.); the first Australian fossil record of pygmy sperm whales (Kogiidae); at least three genera of dolphins (Delphinidae: cf. Tursiops sp., Delphinus sp. or Stenella sp., and an undetermined genus and species); and probable earless or tme seals (Phocidae). This small assemblage represents the first Australian fossil marine mammal assemblage to be described in detail. The taxonomic composition of this Pliocene marine mammal assemblage is generally similar to the present day marine mammal assemblage in north-west Bass Strait. The occurrence of extant cetacean genera in the Portland Pliocene and Blinders Island Cameron Inlet Formation assemblages indicates that the marine m a mmal fauna off south-east Australia had acquired an essentially modern aspect by the Late Pliocene. Several of the cetacean genera recorded in the Portland Pliocene assemblage also occur in similar-aged assemblages in other ocean basins. This corroborates the hypothesis that many cetacean taxa that are widely distributed in the world’s oceans today were equally widespread during the Pliocene. Keywords Cetacea, Carnivora, Pinnipedia, Phocidae, Mysticeti, Odontoceti, Australia, Victoria, Portland, Whalers Bluff Formation, Pliocene Introduction The Pliocene epoch (1.8-5. 3 Ma) is generally considered to be the time during which the modem marine mammal fauna evolved, with the extinction of archaic taxa (as well as some taxa with novel adaptations), and widespread geographic distribution of extant families and genera (Bames, 1977; Fordyce, 1989; Fordyce and Barnes, 1994; Fordyce and Muizon, 2001; Fordyce et al., 2002; Demere et al., 2003). Although the Pliocene marine mammals of the North Pacific (e.g. Barnes, 1973a, 1977, 1998) and eastern tropical Pacific (e.g. Muizon, 1981, 1984; Muizon and DeVries, 1985; Muizon and Domning, 1985, 2002) have been described and discussed in some detail, the Pliocene marine mammals of the Southwest Pacific (Australia and New Zealand) remain poorly known (Fordyce et al., 2002: 53). This is despite the fact that Australian late Neogene marine mammals (mostly cetaceans) are relatively abundant in museum collections, especially the Palaeontology Collections of Museum Victoria, Melbourne. The majority of Pliocene marine mammal fossils in these collections are rather fragmentary with one partially complete skull and associated skeleton known (NMV PI 79005, cf. Megaptera sp.). Despite this general lack of diagnostic skull material, some details of the SW Pacific Pliocene marine mammal fauna may be filled in by the study of certain isolated skeletal elements such as periotics. As noted by Bames (1977: 322), study of these isolated elements may provide data on the taxonomic diversity within an assemblage and wider fauna. Pliocene marine mammal assemblages have hitherto not been described from Australia, so descriptions of even fragmentary material provide an initial basis for understanding marine mammal evolution off southern Australia during the Pliocene. This description of marine mammals from the Pliocene Whalers Bluff Formation assemblage comprises a preliminary basis for the late Neogene fossil record of marine mammals in Australia. Erich M.G. Fitzgerald Figure 1. Locality of Portland in Victoria, south-east Australia, and the Portland fossil marine vertebrate localities. Fossils have been collected as float along the beach and from adjacent cliffs between Dutton Way and Portland Harbour. Black shading indicates areas of cliff outcrop of the Whalers Bluff Formation. Pliocene marine mammals All fossils were collected from coastal exposures of the Whalers Bluff Formation lining Portland Bay in western coastal Victoria, southeast Australia (38°19'S, 141°38'E) (Fig. 1). Unfortunately, details of the geological context of vir- tually all fossils are unknown apart from whether the fossils were collected from the Whalers Bluff Formation or underlying limestone. Fitzgerald (2004a, 2004b) mentioned the Portland fossil marine mammals in previous publications. Bearlin (1987: 177) briefly noted the occurrence of cf. Balaena sp., and cf. Balaenoptera sp. in private collections in an unpublished Ph.D. thesis. The vertebrate faunal list for Portland given by Fitzgerald (2004b: 186) was completed prior to the recognition of two distinct, stratigraphically/temporally disjunct marine vertebrate assemblages at Portland. The majority of the verte- brates listed by Fitzgerald (2004b) were derived from the Pliocene Whalers Bluff Formation although some vertebrates from the Portland Late Miocene assemblage were listed under the Whalers Bluff Formation assemblage. The Portland Late Miocene marine vertebrate assemblage, from the Port Campbell Limestone, is considerably more diverse than the Pliocene Whalers Bluff Formation assemblage, and is to be described in a subsequent publication. Below is an emended list of the vertebrates recorded in the assemblage from the Whalers Bluff Formation at Portland. Chondrichthyes Isurus sp. Carcharodon carcharias Linnaeus, 1758 Carcharodon megalodon Agassiz, 1835 Myliobatis sp. Ischyodus dolloi Leriche, 1902 Mammalia Marsupialia ?Dasyuromorphia incertae sedis Diprotodontoidea gen. et sp. undet. Diprotodontoidea gen. et sp. undet. C Diprotodontidae gen. et sp. undet. A Zygomaturinae gen. et sp. undet. T Palorchestes sp. Vombatidae gen. et sp. undet. Sthenurus sp. Protemnodon sp. Macropus sp. Macropodidae gen. et sp. undet. C Ektopodontidae gen. et sp. undet. Rodentia Rodentia incertae sedis Carnivora ?Phocidae gen. et sp. indet. Cetacea Balaenidae gen. et sp. indet. Balaenopteridae gen. et sp. indet. cf. Physeter sp. Kogiidae gen. et sp. indet. Delphinoidea incertae sedis cf. Tursiops sp. Delphinus sp. or Stenella sp. Delphinidae gen. et sp. undet. A Materials and methods All fossil specimens were collected by Mr Sean Wright of Portland and are in the Palaeontology Collections of Museum Victoria. Anatomical terminology for periotic s and tympanies follows Evans (1993), Fordyce (1994), Fordyce and others (2002) and Kasuya (1973) with mi nor modifications. All periotic and tympanic measurements follow the methods and dimensions outlined by Kasuya (1973) and were made using vernier callipers. Photographs were taken using a 35 mm Nikon Nikkormat EL SLR with a 105 mm macro-lens, and a Nikon D70 digital SLR with a 60 mm macro-lens. Where indicated, specimens were coated with a sublimate of ammonium chloride to enhance contrast in black and white (denoted by AC in figure captions). Where necessary, fragile specimens were consolidated with a hardener con- sisting of 3% solution of Paraloid B72 (ethyl methacrylate/methyl acrylate copolymer) in acetone. Institutional abbreviations. NMV C, Museum Victoria Comparative Anatomy Collection, Melbourne; NMV P, Museum Victoria Palaeontology Collection, Melbourne; CD, Phylum Chordata cata- logue, New Zealand Geological Survey, Lower Hutt; USNM, National Museum of Natural History (formerly United States National Museum), Smithsonian Institution, Washington, DC. For a complete list of specimens referred to in this study, see table 1. Geology and age of the Whalers Bluff Formation For about 4 km along cliffs north of Portland Harbour the Pliocene Whalers Bluff Formation is exposed (Singleton et al., 1976). This formation is about 7.6 m thick, being comprised of horizontally bedded fossiliferous clay, oyster-rich beds and sandy limestones (Abele et al., 1988). These sediments uncon- formably overlie the Upper Miocene Port Campbell Limestone and infill a karst topography developed in the top of the Miocene limestone (Boutakoff and Sprigg, 1953; Dickinson et al., 2002). The Whalers Bluff Formation is unconformably capped by basalts. The age of the Whalers Bluff Formation is well constrained relative to some other Neogene marine mammal-bearing units in the SW Pacific (Fig. 2). However, the determination of the younger age limit of the Whalers Bluff Formation has proved problematic. The Port Campbell Limestone which underlies the Whalers Bluff Formation is Late Miocene (indicated by pres- ence of Globorotalia miotumida\ planktonic foraminiferal zones N16-basal N17; Tortonian; 8-10.8 Ma) (Dickinson et al., 2002). Deposition of the Whalers Bluff Formation began during planktonic foraminiferal zone N19 (indicated by the presence of Globorotalia puncticulata at the base of the for- mation) and ensued into the base of planktonic foraminiferal zone N21 (Fig. 2) (Singleton et al., 1976; Dickinson et al., 2002). K-Ar dates of 2.5 1 Ma from basalts that cap the Whalers Bluff Formation have been reported (Singleton et al., 1976). Beu and Darragh (2001) have suggested that the Whalers Bluff Formation is latest Pliocene to earliest Pleistocene (>1.5-2.0 Ma) based on the presence of the pectinid bivalve Pecten fumatus Reeve, 1852 in a section along the Glenelg River. However, Pecten fumatus does not occur in the Whalers Bluff Formation at Portland (Darragh in Singleton et al., 1976; T.A. Darragh, pers. comm.). Zenatiopsis ultima Darragh and Kendrick, 1971 occurs in the Whalers Bluff Formation at Portland but is a Pliocene species and never occurs with 70 Erich M.G. Fitzgerald Table 1. Specimens referred to in this study. For more detailed locality/stratigraphic data see Fitzgerald (2004b). Abbreviations: e=Early; m=Middle; l=Late; M=Miocene; P=Pliocene; Pt=Pleistocene; R=Recent; Vic. = Victoria; Tas. = Tasmania; NZ = New Zealand. Specimen Taxon Locality Lormation Age NMV P221242 hums sp. Portland, Vic. Whalers Bluff P NMV P218415 Carcharodon megalodon Portland, Vic. Whalers Bluff P NMV P218418 Carcharodon carcharias Portland, Vic. Whalers Bluff P NMV P218470 Myliobatis sp. Portland, Vic. Whalers Bluff P NMV P2 18296 Ischyodus dolloi Portland, Vic. Whalers Bluff P NMV P221241 ?Dasyuromorphia gen. et sp. undet. Portland, Vic. Whalers Bluff P NMV P2 18500 Diprotodontoidea gen. et sp. undet. Portland, Vic. Whalers Bluff P NMV P2 18498 Diprotodontoidea gen. et sp. undet. C Portland, Vic. Whalers Bluff P NMV P21 8499 Diprotodontidae gen. et sp. undet. A Portland, Vic. Whalers Bluff P NMV P221230 Zygomaturinae gen. et sp. undet. T Portland, Vic. Whalers Bluff P NMV P22 1231 Palorchestes sp. Portland, Vic. Whalers Bluff P NMV P221238 Vombatidae gen. et sp. undet. Portland, Vic. Whalers Bluff P NMV P221232 Protemnodon sp. Portland, Vic. Whalers Bluff P NMV P221235 Macropus sp. Portland, Vic. Whalers Bluff P NMV P221227 IMacropus sp. Portland, Vic. Whalers Bluff P NMV P221237 Sthenurus sp. Portland, Vic. Whalers Bluff P NMV P221229 Macropodidae gen. et sp. undet. C Portland, Vic. Whalers Bluff P NMV P197795 Ektopodontidae gen. et sp. undet. Portland, Vic. Whalers Bluff P NMV P221240 Rodentia indet. Portland, Vic. Whalers Bluff P NMV P2 18273 ?Phocidae gen. et sp. indet. Portland, Vic. Whalers Bluff P NMV P21 8465 ?Phocidae gen. et sp. indet. Portland, Vic. Whalers Bluff P NMV P2 18269 Balaenidae gen. et sp. indet. Portland, Vic. Whalers Bluff P NMV P2 18268 Balaenopteridae gen. et sp. indet. Portland, Vic. Whalers Bluff P NMV P2 18298 cf. Physeter sp. Portland, Vic. Whalers Bluff P NMV P21 8407 Kogiidae gen. et sp. indet. Portland, Vic. Whalers Bluff P NMV P21 8283 Delphinoidea incertae sedis Portland, Vic. Whalers Bluff P NMV P2 18284 Delphinoidea incertae sedis Portland, Vic. Whalers Bluff P NMV P2 18286 Delphinoidea incertae sedis Portland, Vic. Whalers Bluff P NMV P2 18266 cf. Tursiops sp. Portland, Vic. Whalers Bluff P NMV P2 18265 Delphinus or Stenella sp. Portland, Vic. Whalers Bluff P NMV P218264 Delphinidae gen. et sp. undet. A Portland, Vic. Whalers Bluff P NMV P16198 ?Phocidae gen. et sp. indet. Beaumaris, Vic. Black Rock Sandstone lM-eP NMV P41759 Phocidae gen. et sp. indet. Beaumaris, Vic. Black Rock Sandstone lM-eP NMV P42523 ?Phocidae gen. et sp. indet. Beaumaris, Vic. Black Rock Sandstone eP NMV PI 60399 Phocidae gen. et sp. undet. Beaumaris, Vic. Black Rock Sandstone eP NMV PI 60433 Phocidae gen. et sp. undet. Beaumaris, Vic. Black Rock Sandstone eP NMV PI 60441 Phocidae gen. et sp. undet. Grange Burn, Vic. Grange Bum eP NMV P215759 ?Phocidae gen. et sp. indet. Beaumaris, Vic. Black Rock Sandstone eP NMV P16195 Balaenidae gen. et sp. indet. Beaumaris, Vic. Black Rock Sandstone eP NMV P48865 Balaenidae gen. et sp. indet. Grange Burn, Vic. Grange Bum eP NMV PI 6043 8 Balaenidae gen. et sp. indet. Beaumaris, Vic. Black Rock Sandstone lM-eP NMV PI 97824 Balaenidae gen. et sp. indet. Beaumaris, Vic. Black Rock Sandstone lM-eP NMV P171503 Balaenoptera edeni or B. brydei Tidal River, Vic. N/A R NMV P179005 Megaptera sp. Lakes Entrance, Vic. Jemmys Point eP NMV P23961 IMesoplodon sp. Cameron Inlet, Tas. Cameron Inlet P NMV P13033 “Stem” cudmorei Beaumaris, Vic. Black Rock Sandstone lM-eP NMV P204352 Delphinidae gen. et sp. undet. A Henley, England Red Crag P-Pt NMV P2 18481 Delphinidae gen. et sp. undet. A Henley, England Red Crag P-Pt NMV C24972 Kogia breviceps Cape Conran, Vic. N/A R NMV C24976 Kogia breviceps Shelley Beach, Vic. N/A R NMV C27879 Eubalaena australis Altona Bay, Vic. N/A R NMV C28892 Megaptera novaeangliae Venus Bay, Vic. N/A R CD 53 Delphinidae gen. et sp. undet. A Chatham Island, NZ Unnamed TP USNM 22953 Orycterocetus crocodilinus Calvert County, USA Calvert mM USNM 183007 Physeteridae gen. et sp. undet. Lee Creek Mine, USA Yorktown eP USNM 452993 Scaphokogia cochlearis Aguada de Lomas, Peru Pisco 1M USNM 183008 Kogiidae gen. et sp. undet. Lee Creek Mine, USA Yorktown eP USNM 251118 Kogiidae gen. et sp. undet. Lee Creek Mine, USA Yorktown eP Pliocene marine mammals 71 Figure 2. Stratigraphic correlation of the Portland fossil marine mammal-bearing formations with selected major late Neogene marine mammal- bearing units. Stratigraphy and geochronology are from Bames (1973, 1977, 1984, 1998), Muizon and DeVries (1985), Muizon and Bellon (1986), Gottfried et al. (1994), Whitmore (1994), Prothero (1998), Fordyce (2002a), Fordyce et al. (2002), Fitzgerald (2004b), Muizon et al. (2004), Bames et al. (2005) and Gradstein et al. (2004). Abbreviations: AGL, Pisco Formation, Aguada de Lomas level; BL, Batesford Limestone; BRS, Black Rock Sandstone; CLB, Pisco Formation, Cerro la Bruja; ELJ, Pisco Formation, El Jahuay level; GBF, Grange Bum Formation; LAF, Lower Member, Almejas Formation; MTM, Pisco Formation, Montemar level; SAO, Pisco Formation, Sacaco level; SAS, Pisco Formation, Sud- Sacaco level; SDF, San Diego Formation; UAF, Upper Member, Almejas Formation; WBF, Whalers Bluff Formation. P. fumatus, the latter first appearing at the base of the Pleistocene (T.A. Darragh, pers. comm.). These data indicate that the Whalers Bluff Formation at Portland is Early to Late Pliocene (Zanclean to Piacenzian; >2.5-4. 8 Ma). Further study of terrestrial mammals from the Whalers Bluff Formation may help refine the geological age of this unit. Unfortunately, the exact bed from which fossil verte- brates were collected within the Whalers Bluff Formation is unknown. Thus, a finer age resolution than that given above for the assemblage is currently unavailable. The age range of the Whalers Bluff Formation presented herein is considered to be the best estimate based on available data and the reader must be cautioned that future work may yield a younger limit on the minimum age. The Whalers Bluff Formation (as well as, in part, laterally equivalent Victorian marine mammal-bearing units such as the Grange Burn Formation, Black Rock Sandstone and Jemmys Point Formation) has been interpreted as representing a prograding quartz-carbonate barrier system with the elastic- dominated units listed above representing an initial marine incursion (Dickinson et al., 2002: 290). Systematics of marine mammals Order Cetacea Brisson, 1762 Suborder Mysticeti Flower, 1864 Family Balaenidae Gray, 1821 Genus and species indeterminate Referred specimen. NMV P2 18269, incomplete right periotic; anterior and superior processes virtually complete, but pars cochlearis almost entirely worn off, and only anteriormost base of posterior process preserved (Fig. 3A). Description. P2 18269 is highly polished and abraded. The anterior process is blunt and globose, being indistinct from the superior process. There is marked lateral exostosis of the super- ior process lateral to the epitympanic recess. The lateral aspect of the anterior process is rugose and pitted. Posteriorly, this pit- ting decreases in density. Only the lateralmost region of the pars cochlearis is preserved. In medial view, the most notable feature is the sulcus for the facial nerve (cr. VII), the course of cr. VII being preserved from its entry into the body of the periotic at the aperure of the internal facial foramen, to its ventral exit into the epitympanic cavity via the ventral facial foramen. All other features of the pars cochlearis and epitym- panic recess have been obliterated. Posterior to the broad and shallow hiatus epitympanicus is a remnant of the base of the posterior process (which is directed posterolaterally and somewhat ventrally). Discussion. Miller (1924: 8-9) listed the following features that distinguish the periotics of Balaenidae from those of Balaenopteridae (and other baleen-bearing Mysticeti): (1) axis of anterior process of periotic parallel with axis of internal acoustic meatus; (2) [longitudinal] axes of anterior and posteri- or processes converge at an acute angle; and (3) pars cochlearis small relative to rest of periotic. In addition to the preceding features, the possession of massive lateral exostosis of the anterior process and anterolateral superior process, such that the anterior process appears swollen (as noted by Fordyce, 1982: 48), seems to be a feature shared by all extant and late Neogene balaenid periotics. It is largely on the basis of the latter character and the phenetic similarity of P2 18269 to a periotic (PI 6 195) from the Lower Pliocene Black Rock Sandstone of Beaumaris identified as belonging to cf. “ Balaena ” (Gill, 1957) that P2 18269 is referred to Balaenidae, genus and species indeterminate. 72 Erich M.G. Fitzgerald The fossil record of Balaenidae begins in the Late Oligocene (c. 28 Ma: Fordyce, 2002b), although the record only becomes reasonably well known from the Mio-Pliocene boundary onwards (McLeod et al., 1993; Bisconti, 2003). Morenocetus parvus Cabrera, 1926 is the geologically oldest named balaenid, from the early Early Miocene (Aquitanian) of Patagonia. From the end Aquitanian to early Tortonian of the Miocene the evolutionary history of Balaenidae is virtually unknown. The extant balaenids include Balaena mysticetus Linnaeus, 1758, Eubalaena australis Desmoulins, 1822, E. glacialis Muller, 1776, and E. japonica Lacepede, 1818 (e.g., Cummings, 1985; Reeves and Leatherwood, 1985; Bannister, 2002). Note that Rice (1998) included all extant balaenids in the genus Balaena and recognised only two species, B. mysticetus and B. glacialis. The taxonomic scheme of Bannister (2002) is used herein. Balaena is known from the Early Pliocene of the North Atlantic (McLeod et al., 1993; Westgate and Whitmore, 2002). There are very few confirmed pre-Quatemary fossil records of Eubalaena. Bisconti (2003, 2005) referred the Pliocene Balaena belgica Abel, 1941 to Eubalaena belgica. McLeod and others (1993: 63) suggested that a balaenid periotic from the Early Pliocene of South Australia (originally recorded by Howchin: 1919) could repre- sent Eubalaena (as opposed to its original referral to Balaena). Dixon (1990) described an incomplete Recent Eubalaena australis skeleton from Altona Bay, near Melbourne, Victoria. The latter specimen (C27879) includes tympanies and perio- tics. The extinct genera Balaenula and Balaenotus have been recorded from the Late Miocene through Pliocene of the N Pacific (Barnes, 1977; McLeod et al., 1993) and N Atlantic (McLeod et al., 1993; Bisconti, 2003 and references therein). Recently, Bisconti (2005) described a new genus and species of relatively small balaenid, Balaenella brachyrhynus, from the Early Pliocene of Belgium. The incompleteness of P2 18269 and lack of information on the extent of intraspecific and ontogenetic variation in balaenid periotics, hampers comparisons with described extant and fossil balaenid taxa. Furthermore, there are as yet no published criteria for discriminating between the periotics of Balaena and Eubalaena. Despite these problems, it may be noted that P2 18269 is similar in overall size to several isolated balaenid periotics from the uppermost Miocene to Lower Pliocene Black Rock Sandstone and Grange Burn Formation of Victoria (e.g. P16195, P48865, P160438, and P197824). The discovery of a more complete periotic (including the pars cochlearis) is neces- sary before any further comparisons between the Portland Pliocene balaenid and the other Victorian specimens listed above can be made. Family Balaenopteridae Gray, 1864 Genus and species indeterminate Referred specimen. NMV P2 18268, incomplete right periotic; lacking medial three-quarters of pars cochlearis and posterior process (Fig. 3B). Description. P2 18268 is polished, rolled and may be secondar- ily phosphatised. The anterior process is elongated and some- what attenuated anteriorly. The dorsal surface of the anterior process is smooth, with only slight rugosity, as seen in the periotics of extant Balaenopteridae. An oblique groove on the dorsolateral surface of the anterior process near its preserved apex is interpreted as a trace of a vascular sulcus. The latter feature has previously been considered a sulcus for the cap- suloparietal emissary vein (Geisler and Luo, 1998; Geisler and Sanders, 2003) or as a sulcus marking the path of an artery, specifically part of the middle meningeal artery (Fordyce, 1994). Fordyce (1994) and Watson and Fordyce (1994) described this feature as the anteroextemal sulcus whereas Geisler and Sanders (2003) treated the anteroextemal sulcus and sulcus for the capsuloparietal emissary vein as separate fea- tures. Further work is required to better establish the venous/arterial correlate of this osteological feature which in this study is referred to as the sulcus for the capsuloparietal emissary vein. In extant balaenopterids, this sulcus usually courses posteriorly to a point level with the position of the mallear fossa. However, in P2 18268 any more posterior con- tinuation of the sulcus for the capsuloparietal emissary vein, if formerly present, no longer occurs due to abrasion. The ventral presentation of the periotic exhibits several features. As occurs in extant Balaenoptera and Megaptera, the lateralmost eminence of the ventrolateral ridge of the superior process (sensu Geisler and Luo, 1996) is situated at the same level as the anterior margin of the pars cochlearis. The mallear fossa is poorly differentiated from the rest of the epitympanic recess. The preserved posterolateral margin of the periotic is formed by the hiatus epitympanicus. The course of the facial nerve on the ventral surface of the periotic is marked by the facial sulcus which is bounded anteriorly by the aperture of the ventral facial foramen. In ventral view, a distinct bridge of bone at the anterolateral comer of the preserved pars cochlearis represents the ventral roof of the ventral facial foramen. The endocranial aspect of the pars cochlearis preserves the aperture of the internal facial foramen. Anterior to this aperture is a deep excavation in the medial surface of the periotic at the base of the anterior process. This region (composed of cancellous bone in extant balaenopterids) marks the site of ankylosis between the anterior process and the body of the periotic. As in other Balaenopteridae, the pars cochlearis appears to have been elongated towards the cranial cavity. Discussion. That P2 18268 is a balaenopterid periotic is evident by the possession of: (1) elongated, triangular and anteriorly attenuated anterior process; (2) a triangular lateral eminence of the ventrolateral ridge; and (3) a relatively large pars cochlearis elongated towards the cranial cavity. Because P2 18268 is rep- resented only by an incomplete periotic, it is not possible for it to be identified below family level. The size of P2 18268 is comparable to that of periotics of subadult Balaenoptera edeni Anderson, 1879 or B. brydei Olsen, 1913 (P171503) and juve- nile Megaptera novaeangliae Borowski, 1781 (C28892). However, the periotics of extant Megaptera novaeangliae and an undescribed species of Megaptera from the Early Pliocene of Victoria (PI 79005) possess the following features which differentiate them from P2 18268: (1) the anterior process is Pliocene marine mammals 73 Figure 3. Mysticeti and Physeteridae from the Pliocene Whalers Bluff Formation, Portland. A, Balaenidae gen. et sp. indet., incomplete right peri- otic, NMV P2 18269, in ventrolateral view (AC); B, Balaenopteridae gen. et sp. indet., incomplete right periotic, NMV P2 18268, in ventral view (AC); C, cf. Physeter sp., apical crown of tooth, NMV P218298, in side view (AC). Scale bars equal 10 mm. 74 Erich M.G. Fitzgerald relatively shorter; (2) anterior process is more dorsoventrally compressed; (3) in endocranial view, there is an anteroposter- iorly thickened region of cancellous bone in the pars cochlearis anterior to the internal facial foramen; and (4) no deep excava- tion in the endocranial surface of the periotic anterior to the pars cochlearis. Many of these features may be related to onto- genetic variation. At least, the lack of these four features in P2 18268 may indicate that this periotic is not referrable to Megaptera and may belong to an indeterminate species in the genus Balaenoptera. Suborder Odontoceti Flower, 1867 Family Physeteridae Gray, 1821 cf. Physeter sp. Linnaeus, 1758 Referred specimen. NMY P2 18298, isolated, worn-down apical region of tooth crown (Fig. 3C). Description. P2 18298 has an ovoid cross-section at its base which becomes more ellipsoid towards the preserved apex of the crown. In occlusal view, the apex of the tooth crown is incised by a deep anteroposterior cleft which is attributed to tooth wear. Shallow longitudinal grooves in the surface occur on all faces of the crown. The basal end of the tooth is broken to reveal a cross-section through the crown. The only notable feature of this aspect of the tooth is the presence of thick layers of cementum. Discussion. The large size of this tooth and the thick layers of cementum suggest that P2 18298 represents a physeterid. Little more can be said about the systematic s of this speci- men although its size and similarity to teeth of large adult Physeter macrocephalus suggest affinities with the extant sperm whale. Family Kogiidae Gill, 1871 Genus and species indeterminate Referred specimen. NMV P218407, incomplete left tympanic bulla (Fig. 4A-B). Description. The most striking aspect of the morphology of P2 18407 is its small size. The tympanic is polished with rolled edges being rounded off. The preserved portion includes only the medial half of the bulla with very little of the outer lip pres- ent. The base of the posterior process of the tympanic has been worn off. In dorsal and ventral view, there is a distinct furrow in the medial edge of the involucrum between the inner poster- ior prominence and the inner anterior prominence. In ventral view, the furrow between the prominences of the involucrum forms an obtuse angle. There is no preserved ventral keel and the median furrow is very shallow such that it is poorly differ- entiated from the surrounding ventral surface of the tympanic. The interprominential notch is relatively broad. The anterior edge of the involucrum and outer lip is squared-off. In dor- sal view, the involucrum has a consistent width along its length, but is expanded at the level of the inner posterior prominence. Discussion. P2 18407 is referred to the Kogiidae on the basis of the following features shared with extant and fossil kogiids: (1) small overall size [as Kasuya (1973: 25) noted among extant Odontoceti only Pontoporia blainvillei Gervais and d’Orbigny, 1844 has a smaller tympanic, and P2 18407 does not resemble the tympanic of that taxon]; (2) distinct embayment in the medial edge of the involucrum between the inner anterior and posterior prominences; and (3) squared-off anterior edge of the involucrum and outer lip. It should be noted that the second feature is also seen in the tympanic bullae of Physeter macrocephalus Linnaeus, 1758, Orycterocetus croco- dilinus Cope, 1868 (USNM 22953) (Kellogg, 1965) (Fig. 5C-D), and an Early Pliocene physeterid (USNM 183007) (Fig. 5A, B). Despite this similarity between P2 18407 and the tympanies of Physeteridae, the relatively large size of physeterid tympanies precludes P2 18407 from being con- sidered as a physeterid. Furthermore, the inner posterior prominence in physeterid tympanies is generally more pointed in outline than the rounded outer posterior prominences in kogiid tympanies. Among fossil Kogiidae, the tympanies of Praekogia cedrosensis (Barnes, 1973b) have not been described. A skull of Scaphokogia cochlearis Muizon, 1988 (USNM 452993) includes an associated incomplete left tympanic (Fig. 4E, F). The tympanic of S. cochlearis is similar to the tympanies of extant Kogia in its relatively small size and possession of a dis- tinct embayment in the medial side of the involucrum between the inner anterior and posterior prominences. The most notable difference between the tympanic of S. cochlearis and the Portland kogiid tympanic lies in the more marked inflation of the inner posterior prominence of P2 18407. S. cochlearis pos- sesses a less expanded inner posterior prominence, such that the embayment in the medial face of the involucrum is not as deep as in P2 18407. In this respect, Scaphokogia cochlearis is similar to two undescribed Early Pliocene kogiids from the Lee Creek Mine, North Carolina (USNM 183008, Fig. 4G-H; USNM 251118, Fig. 4C, D) and the extant Kogia breviceps Blainville, 1838 (Fig. 41, J). The features which Kasuya (1973) used to differentiate between Recent Kogia breviceps and K. sima Owen, 1866 are not preserved in P2 18407. However, comparison between fig- ures of the tympanic of K. sima (Kasuya, 1973: plate VIII), and actual specimens of K. breviceps (C24972, C24976), indicate that P2 18407 differs from both extant Kogia species in: (1) embayment in medial edge of involucrum between the inner prominences is markedly deeper; and (2) the involucrum is less dorsoventrally and mediolaterally inflated. Despite these dif- ferences, P2 18407 is almost identical in size to the tympanies of Kogia breviceps. Given that the currently available evidence is meagre, P2 18407 is not identified below family level. Table 2 compares some dimensions of the tympanies of kogiids and physeterids discussed above. P2 18407 is the first fossil record of Kogiidae from Australia. Fossil kogiids have previously been reported in the SW Pacific region, from the ?Late Miocene of the Chatham Rise, east of New Zealand (Fordyce, 1984a) but that record has a poorly constrained age (Fordyce, 1989, 1991b). Pliocene marine mammals 75 Table 2. Measurements of tympanies of Physeteridae and Kogiidae. Measurements follow methodology of Kasuya (1973). All measurements are in mm. NMV C24972 NMV P2 18407 USNM 22953 USNM 183007 USNM 452993 USNM 183008 USNM 251118 Standard length of tympanic bulla, distance from anterior tip to posterior end of outer posterior prominence 26.0 31.0 42 29.0 31 Distance from anterior tip to posterior end of inner posterior prominence 19.9 21.6 27.0 38 28 31.0 28 Distance from posteroventral tip of outer posterior prominence to tip of sigmoid process 21.28 _ _ _ _ _ _ Distance from posteroventral tip of outer posterior prominence to tip of conical process 15.3 _ _ _ _ _ _ Width of tympanic bulla at the level of the sigmoid process 26.0 _ _ _ _ _ _ Height of tympanic bulla, from tip of sigmoid process to ventral keel 26.5 Width across inner and outer posterior prominences 19.0 - 20.2 24 22.5 20+ Superfamily Delphinoidea Gray, 1821 Incertae sedis Referred specimens. NMV P218283, P218284 and P218286, all isolat- ed teeth (not figured). Description. P21283, P218284 and P218286 all represent small odontocete teeth possessing conical enamel-covered crowns that bear fine wrinkling ornamentation, and curve lingually towards the crown apex. As in kentriodontids, there is a lingual bulge at the base of the crown but this feature is not as promi- nent in the teeth from Portland. None possesses an open pulp cavity suggesting that all were derived from adult individuals. P2 18283 is an incomplete tooth, its preserved maximum length and maximum width of the crown being 16 mm and 6 mm respectively. Due to the incomplete nature of this tooth it does not warrant further description. P2 18284 is the most highly polished and the most complete. It differs from the others in having a mediolaterally compressed root with a more prominent mesial-distal bulge at its midpoint. The preserved apex of the root curves posteriorly. The most notable feature of P2 18286 distinguishing it from the other teeth is its bulbous root, which contrasts with the transversely flattened morphology of P2 18284. Discussion. Only one delphinoid odontocete has previously been described from the Tertiary of Australia, the latest Miocene-earliest Pliocene “ Steno ” cudmorei Chapman (1917) from the Black Rock Sandstone of Beaumaris, Victoria (Fitzgerald, 2004b). Fordyce (1982) questioned the taxonomic identity of “5?’ cudmorei (the holotype, P13033, being an iso- lated tooth) and Fitzgerald (2004b) referred “S.” cudmorei to Delphinidae, genus and species indeterminate. Chapman (1917) assigned P13033 to Steno on the basis of the resem- blance of its crown ornamentation to that seen in the teeth of the extant Steno bredanensis Cuvier in Lesson, 1828 (e.g., Miyazaki and Perrin, 1994). Given that Steno is probably in a basal position in the phylogeny of Delphinidae (Miyazaki and Perrin, 1994; LeDuc et al., 1999) and that some of the pre- sumed ancestors of Delphinidae, the Kentriodontidae (Barnes, 1978, 2002; LeDuc, 2002), possessed Steno-like crown orna- mentation (e.g., Kellogg, 1966), the anastomosing striae on the crown of P13033 (and P218283, P218284, P218286) are of dubious use in assessing the phylogenetic affinities of isolated teeth. Furthermore, non-delphinid small odontocetes such as Lipotes vexillifer Miller, 1918 possess anastomosing wrinkling on tooth crown enamel (Miller, 1918; Brownell and Herald, 1972; Barnes, 1985) which casts doubt on any perceived taxonomic or phylogenetic signal present in these teeth. The isolated teeth from Portland are assigned to Delphinoidea incertae sedis. None of the Portland Pliocene teeth share demonstrably close affinities with the holotype tooth of “ Steno ” cudmorei. Family Delphinidae Gray, 1821 cf. Tursiops sp. Gervais, 1855 sp. Referred specimen. NMV P2 18266, virtually complete right periotic (Fig. 6). Description. P2 18266 is typically delphinid in possessing: (1) posterior process of the periotic projects laterally and postero- laterally; (2) longitudinal grooves on the articular facet of the posterior process of the tympanic (this feature also occurs in Monodontidae); (3) relatively low crista transversa; (4) internal facial foramen opens at the same level as the tractus spiralis foraminosus in the internal acoustic meatus; (5) a short, blunt, rectangular anterior process of the periotic which in anterior, dorsal and ventral views appears laterally compressed; (6) a large fovea epitubaria for the accessory ossicle eliminates the anterior bullar facet on the anterior process; (7) prominent parabullary ridge; (8) inflated pars cochlearis; and (9) rela- tively shallow internal acoustic meatus (Kasuya, 1973; Fordyce et al., 2002). The first character is usually only seen in Delphinidae but several taxa in the extinct delphinoid grade- 76 Erich M.G. Fitzgerald A B Figure 4. Miocene to Recent Kogiidae tympanies. A-B, Kogiidae gen. et sp. indet. (Pliocene Whalers Bluff Formation, Portland, Victoria, Australia), incomplete left tympanic, NMV P2 18407 (AC). C-D, Kogiidae gen. et sp. undet. (Lower Pliocene Yorktown Formation, Lee Creek Mine, North Carolina, U.S.A.), incomplete left tympanic, USNM 251118. E-F, Scaphokogia cochlearis (Upper Miocene Pisco Formation, Aguada de Lomas level, Arequipa Department, Peru), incomplete left tympanic, USNM 452993. G-H, Kogiidae gen. et sp. undet. (Lower Pliocene Yorktown Formation, Lee Creek Mine, North Carolina, U.S.A.), incomplete right tympanic, USNM 183008. 1-J, Kogia breviceps (Recent, Shelley Beach, Victoria, Australia), incomplete left tympanic, NMV C24976. A, C, E, G, I, all in dorsal view. B, D, F, H, J, all in ventral view. Scale bars equal 10 mm. Pliocene marine mammals 77 C D Figure 5. Miocene to Pliocene Physeteridae tympanies. A-B, Physeteridae gen. et sp. undet. (Lower Pliocene Yorktown Formation, Lee Creek Mine, North Carolina, U.S.A.), right tympanic, USNM 183007. C-D, Orycterocetus crocodilinus (Middle Miocene Calvert Formation, Zone 14, south of Randle Cliff Beach, Calvert County, Maryland, U.S.A.), right tympanic, USNM 22953. A and C in dorsal view. B and D in ventral view. Scale bars equal 10 mm. taxon Kentriodontidae possess a posterolaterally projecting posterior process of the periotic (eg., Barnes and Mitchell, 1984; Dawson, 1996) as does Albireo whistleri (Albireonidae) (Barnes, 1984). P218266 and the other odontocete periotics described below possess all of the characters listed above, justifying their assignment to Delphinidae. Table 3 presents comparative measurements of selected dimensions for all delphinid periotics described herein. Discussion. Within Delphinidae, P218266 is most similar to Tursiops and Lissodelphis. P2 18266 is similar to the periotics of Tursiops in: (1) its overall size; (2) having the aperture of the internal acoustic meatus opening at the same level as the endocranial surface of the body of the periotic; (3) having a deep internal facial foramen; and (4) having an aquaeductus vestibuli located at the same level as the foramen for the vestibular branch of the vestibulocochlear nerve. P2 18266 78 Erich M.G. Fitzgerald Table 3. Measurements of delphinid periotics from the Whalers Bluff Formation and delphinid periotics from the Red Crag, England. Measurements follow methodology of Kasuya (1973). All measurements are in mm. NMV NMV NMV NMV NMV P218264 P2 18265 P2 18266 P204352 P218481 Standard length of periotic, from tip of anterior process to posterior end of posterior process, measured on a straight line parallel with cerebral border 28.00 26.66 32.50 29.94 29.40 Thickness of superior process at the level of upper tympanic aperture Width of periotic across pars cochlearis and superior process, at the level of 11.00 10.00 11.90 12.14 11.58 upper tympanic aperture Least distance between the margins of fundus of internal acoustic meatus 20.00 17.76 23.00 20.00 19.58 and of aperture of aquaeductus vestibuli Least distance between the margins of fundus of internal acoustic meatus 2.56 2.16 1.58 3.00 2.50 and of aperture of aquaeductus cochleae 3.00 3.00 3.14 1.50 2.20 Length of the posterior bullar facet 11.00 11.00 11.00 12.46 12.62 Antero-posterior diameter of pars cochlearis 17.00 16.00 17.80 14.66 16.16 differs from Tursiops in: (1) the possession of a more acute angle between the anterior process and pars cochlearis; (2) having an aquaeductus cochleae positioned further from the fenestra cochleae, and further dorsally on the endocranial side of the pars cochlearis; (3) having a lower crista transversa; (4) having a poorly developed septum between the tractus spiralis foraminosus and the foramen for the vestibular branch of the vestibulocochlear nerve; and (5) having a less rounded pars cochlearis. P2 18266 shares with the periotic of Lissodelphis : (1) endocranial surface of the periotic bordering the posterior margin of the internal acoustic meatus, aperture of aquaeductus vestibuli, and aquaeductus cochleae, is uniformly flat; and (2) aperture of the aquaeductus cochleae opens on the same plane as the aperture of the internal acoustic meatus (note that this condition is also seen in Steno). P2 18266 differs from the periotic of Lissodelphis by: (1) being markedly larger in over- all size; and (2) having a mediolaterally broader aperture of the internal acoustic meatus. The balance of features noted above indicates that P2 18266 is a delphinid with close affinities to Tursiops. This periotic is not identified as undoubtedly as such due to the subtle differ- ences between P2 18266 and the periotic of extant Tursiops spp. Barnes (1990) provided a thorough review of the fossil record of Tursiops. He recognised four fossil species of Tursiops-. T. cortesii Sacco, 1891 (Italy; Late Pliocene, c. 1.75-3.5 Ma, and possibly Early Pliocene, c. 5 Ma); T. astensis Sacco, 1891(Italy; early Late Pliocene, c. 3-3.5 Ma); T. capellinii del Prato, 1898 (Italy; middle Pliocene, c. 3.5 Ma); T. ossenae Simonelli, 1911 (Italy; middle Pleistocene or late Pleistocene, c. 0.5-0. 8 Ma). Other unnamed pre-Pleistocene records of Tursiops include: aff. Tursiops from the Purisima Formation, California (Late Pliocene, 1.8-3 Ma) (Barnes, 1977); and Tursiops sp. from the Yorktown Formation, North Carolina (Early Pliocene, 3. 5-4.5 Ma) (Whitmore, 1994). The record from the Pliocene of Victoria is the first prob- able pre-Pleistocene fossil record of Tursiops from the Southwest Pacific. Furthermore, this fossil occurrence provides corroboration for Barnes’ (1990: 18) hypothesis that Tursiops has been as geographically widespread during the last six million years as it is at the present time. Delphinus sp. Linnaeus, 1758 or Stenella sp. Gray, 1866 Referred specimen. NMV P2 18265, virtually complete left periotic (Fig. 7). Description. P2 18265 is smaller than P2 18266 (cf. Tursiops sp.) (Fig. 6), closely matching the periotics of both Delphinus and Stenella in overall proportions. Morphological similarities between P2 18265 and Delphinus include: (1) relatively large diameter of the aperture of the fenestra cochleae; and (2) aper- ture of the aquaeductus vestibuli is ellipsoid to slit-like in out- line and opens posteriorly. However, P2 18265 differs from Delphinus in: (1) the suprameatal area of the periotic, lateral to the aperture of the internal acoustic meatus, is less planar and more convex; (2) the lateral wall of the internal acoustic meatus lacks an elevated platform which obscures the vesti- bular foramen in endocranial view; (3) the crista transversa is lower; and (4) the aperture of the internal facial foramen is narrower. Discussion. Whitmore (1994) noted that it is very difficult to distinguish between Delphinus and Stenella on the basis of periotic morphology. Confounding generic identification of P21265 is the fact that the periotics of Stenella show wide indi- vidual variation (Kasuya, 1973). In both Delphinus and Stenella, the aquaeductus cochleae opens at a similar position on the posterior wall of the pars cochlearis. Most preserved fea- tures indicate closer affinities with Stenella than Delphinus. However, all of the identified differences between P2 18265 and Delphinus are subtle, representing differences in degree rather than kind. I do not advocate the definitive placement of P2 18265 within either Delphinus or Stenella. In any event, the somewhat intermediate morphology of P2 18265 between the periotics of Delphinus and Stenella may indicate that P2 18265 is from an extinct genus closely related to both of the extant genera in question. In either case, P2 18265 is the first record of a periotic of the Delphinus-Stenella type recorded from the Neogene of Australia. Fossils representing Delphinus or Stenella have been recorded from: Upper Member Almejas Formation, Baja California Sur (latest Miocene to Early Pliocene, 3.5-6 Ma) (Barnes, 1998); San Mateo Formation, California (Late Pliocene marine mammals 79 Figure 6. cf. Tursiops sp. (Pliocene Whalers Bluff Formation, Portland, Victoria, Australia), right periotic, NMV P218266 (AC). A, ventral view. B, cranial view. C, medial view. D, lateral view. Scale bar equals 10 mm. 80 Erich M.G. Fitzgerald Figure 7. Delphinus sp. or Stenella sp. (Pliocene Whalers Bluff Formation, Portland, Victoria, Australia), left periotic, NMV P2 18265 (AC). A, ventral view. B, cranial view. C, medial view. D, lateral view. Scale bar equals 10 mm. Pliocene marine mammals 81 Miocene to Early Pliocene, 5-9 Ma) (Barnes et al., 1981); unnamed blue clays at Waihi Beach, Hawera, New Zealand (Early Pliocene, 3-3.6 Ma; Beu, 1995) (McKee and Fordyce, 1987; Fordyce, 1991a: 1262); Yorktown Formation, North Carolina (Early Pliocene, 4.5 Ma) (Whitmore, 1994); Salada Formation, Baja California Sur (Early to Late Pliocene, 3-5 Ma) (Barnes, 1998); Tirabuzon Formation, Baja California Sur (middle Pliocene, 3-4 Ma) (Barnes, 1998); SAO Level of the Pisco Formation, Peru (late Early to early Late Pliocene, 3. 0-4.0 Ma) (Muizon and DeVries, 1985; Muizon and Domning, 2002); and San Diego Formation, California (Late Pliocene, 1.8-3.4 Ma) (Barnes, 1973, 1977, 1998). Genus and species undetermined A Referred specimen. NMV P2 18264, virtually complete right periotic (Fig. 8). Description. P2 18264 differs from all previously described delphinid periotics from the Pliocene of Portland in its relative- ly good state of preservation, indicating less post-mortem trans- port and abrasion, with fine surface details intact. P2 18264 is distinct from P2 18265 and P2 18266 but intermediate in overall size. Notable differences between P2 18264 and P2 18265 and P2 18266 include: (1) a mediolaterally thickened periotic body lateral to the pars cochlearis; (2) apex of anterior process with distinct tubercle; (3) possession of two deep creases in the anteromedial face of the anterior process; (4) aperture of the internal acoustic meatus is mediolaterally compressed, such that in endocranial view the meatus is more ellipsoidal and slit- like in outline than that of either or the others; (5) the aperture of the internal facial foramen is narrow; and (6) the aperture of the aquaeductus vestibuli is so narrow that it appears closed. Discussion. P2 18264 does not closely resemble the periotics of any Recent delphinids in the Museum Victoria collections, nor any of the Recent delphinid periotics figured by Kasuya (1973). However, P2 18264 (Fig. 8) closely resembles two fossil del- phinid periotics from the Plio-Pleistocene Red Crag of England (P204352, P2 18481: Fig. 9). Both periotics were originally identified as Globicephalus uncidens Lankester, 1864 (sensu Lydekker, 1887) by persons unknown. The Globicephalus uncidens periotic figured by Lydekker (1887: plate II fig. 11) probably represents a Globicephala periotic, Globicephalus being a junior synonym of Globicephala. However, P204352 and P2 18481 (Fig. 9) do not represent Globicephala periotics. An incomplete delphinid periotic (CD 53) has been described from late Neogene sediments of uncertain age on Chatham Island, east of New Zealand (Fordyce and Campbell, 1990). The following features of CD 53 are shared with P2 18264: well-developed, laterally expanded, parabullary ridge with a globose, tubercle-like anterior apex; anteroventral angle is rounded off; creases in ventrolateral parabullary ridge; and short, broad, fan-like posterior process with heavily fis- sured posterior bullar facet. Based on these similarities, it is hypothesised that CD 53 is in the same genus, at least, as P2 18264. Given the Pliocene age of P2 18264, and Plio- Pleistocene age of P204352 and P2 18481, this taxonomic assignment of CD 53 suggests a lowest Opoitian (Early Pliocene) or younger age for the Chatham Island delphinid periotic and its host sediments. Fordyce and Campbell (1990: 62-63) suggested a Late Miocene or younger age for CD 53. Pending the discovery of more complete cranial material including periotics of this type, P2 18264 is referred to an undetermined genus and species of Delphinidae (as are P204352, P218481 and provisionally CD 53). Order Carnivora Bowdich, 1821 Suborder Pinnipedia Illiger, 1811 Family ?Phocidae Gray, 1821 Genus and species indeterminate Referred specimens. NMV P2 18465, incomplete left horizontal ramus of mandible (Fig. 10). NMV P2 18273, isolated upper incisor or canine tooth (not figured). Description. P2 18465 is an incomplete left horizontal ramus with a preserved length of 69 mm, depth at level of dorsal con- cavity of 21 mm, and mediolateral thickness at level of poster- ior alveolus for ml of 10 mm. The surface detail of P218465 is well preserved relative to most of the other marine vertebrate fossils recovered from the Whalers Bluff Formation. P2 18465 lacks all of the horizontal ramus anterior to the anterior alveo- lus for p4 and all of the ascending and horizontal rami poster- ior to the anteriormost corner of the fossa for m. masseter pars profundus. No teeth are preserved in situ in the mandible. In overall proportions, the ramus is lightly built. In posterior and dorsal view, slight medial inflection of the ramus is visible. In lateral aspect, there is a small mental foramen located ventral to the position of the interalveolar septum between the anterior and posterior alveoli of p4 (Fig. 10B). This position is almost identical to that of the posteriormost mental foramen on a mandible referred to Piscophoca sp. by Walsh and Naish (2002). Such small mental foramina are also present on the mandibles of Zalophus californianus Lesson, 1828 (Howell, 1929; pers. obs.). Immediately posterior to the posterior alveolus of ml is a prominent dorsal concavity 25 mm long (measured from the ml alveolus to the anterior margin of the fossa for m. masseter pars profundus). This dorsal concavity resembles that of Piscophoca. The preserved anterior region of the fossa for m. masseter pars profundus differs from that of Piscophoca in being dorsoventrally broader with a more rounded outline. P2 18273 represents an isolated incisor or canine tooth, 28 mm long; its mesiodistal width at a level just below the base of the crown is 8 mm while its buccolingual width at the same level is 6 mm. The crown has smooth enamel and is recurved towards the apex. The root has a consistent thickness and is buccolingually compressed. Discussion. The pinniped fossils recovered from the Whalers Bluff Formation do not include elements preserving unequivo- cal synapomorphies of suprageneric pinniped taxa. Berta (1991) and Berta and Wyss (1994) listed the possession of a bony flange below the angular process of the mandible as an unequivocal synapomorphy of Superfamily Phocoidea (sensu Wyss and Flynn, 1993) but the posterior region of the mandible is not preserved in P2 18465. However, P2 18465 is tentatively 82 Erich M.G. Fitzgerald Figure 8. Delphinidae gen. et sp. undet. A (Pliocene Whalers Bluff Formation, Portland, Victoria, Australia), right periotic, NMV P2 18264 (AC). A, ventral view. B, cranial view. C, medial view. D, lateral view. Scale bar equals 10 mm. Pliocene marine mammals 83 Figure 9. Delphinidae gen. et sp. undet. A (Pleistocene-Pliocene Red Crag, Henley, England), right periotic, NMV P2 18481 (AC). A, ventral view. B, cranial view. C, medial view. D, lateral view. Scale bar equals 10 mm. 84 Erich M.G. Fitzgerald Figure 10. ?Phocidae gen. et sp. indet. (Pliocene Whalers Bluff Formation, Portland, Victoria, Australia), incomplete left mandible, NMV P2 18465 (AC). A, dorsal view. B, lateral view. C, medial view. Black arrow in B points to mental foramen. Scale bar equals 10 mm referred to the Phocidae on the basis of its morphology and pro- portions being most similar to the mandibles of phocid seals, as opposed to otariids and odobenids. It remains possible that P2 18465 represents an otariid mandible. If this were the case, then a major rethinking of otariid evolutionary biogeography would be necessary, as current estimates place the otariid dispersal into the Southern Hemisphere at around the Pliocene/Pleistocene boundary (Demere et al., 2003; contra Repenning and Tedford, 1977, who indicated a latest Miocene dispersal event at the earliest), and the age of P2 18465 (and P2 18273) probably predates that horizon. Given that P2 18465 is not similar in morphology to Otariidae but shares certain features with some phocids (see description and discussion below), there is no firm evidence to suggest that the Portland mandible represents an otariid. The fact that phocid seal fossils have previously been reported from Pliocene-aged sediments in Victoria (Fordyce and Flannery, 1983) whereas otariids have not lends further support to the assignment of P2 18465 to the Phocidae. However, P2 18465 and P2 18273 are not referred unquestionably to Phocidae because the late Neogene fossil record of marine mammals in the SW Pacific remains too poorly documented to provide any absolute idea of the composition of the marine mammal fauna during the Pliocene. Among extant and fossil phocid mandibles, P2 18465 is most similar to those of the Pliocene taxa Acrophoca lon- girostris Muizon, 1981, Homiphoca capensis Hendey and Repenning, 1972, and Piscophoca pacifica Muizon, 1981. P2 18465 may be clearly distinguished from Acrophoca, as it lacks the wide diastema between cheek teeth characteristic of Acrophoca (Muizon, 1981). The Portland mandible can be fur- ther distinguished from Homiphoca (Hendey and Repenning, 1972; Muizon and Hendey, 1980) by the possession of a well-developed dorsal concavity posterior to ml. P2 18465 is generally very similar to the mandible of Piscophoca (Muizon, 1981) in its overall proportions, relative length of the dorsal concavity posterior to ml, subequal diameters of the alveoli and relatively closely spaced alveoli along the tooth row. However, it is not possible at this stage to determine whether the Portland mandible belongs to a species of Piscophoca or not. The Australian fossil record of pinnipeds is currently poor. Fordyce (1991b) summarised the state of knowledge at the beginning of the 1990s, and virtually nothing has been added since that time. The oldest fossil pinnipeds from the SW Pacific are latest Miocene (c. 6 Ma) at the earliest and are from Australia (Fitzgerald, 2004b). Fordyce and Flannery (1983) provided a preliminary assessment of these fragmentary fossils suggesting that they represented monachine phocids. The fos- sils represent one ?incisor tooth (P16198), two right temporals Pliocene marine mammals 85 (P160399 and P160441), two fused sacral vertebrae (P41759) and an articulated series of eight thoracic vertebrae with five ribs (P160433). None of these specimens has yet been described and only one of the temporals (PI 60399) has been figured (Fordyce and Flannery, 1983: 99). Recently, two other pre-Pleistocene ?phocid fossils have been discovered: P42523, isolated right metatarsal V; and P2 15759, isolated left metatarsal V. Both P42523 and P215759 were derived from beds immediately overlying a phosphatic nodule horizon at the base of the Black Rock Sandstone (Beaumaris, Victoria), and are thus early Early Pliocene in age. The exact relationships of the phocids represented by temporals to extant Monachinae and their fossil sister-taxa ( Acrophoca , Homiphoca and Pisco - phoca) have yet to be determined. The report herein of two probable phocid pinniped fossils from Portland brings the number of known Australian pre-Pleistocene pinniped specimens to nine. Conclusions The fossil marine mammal assemblage from the Pliocene-aged Whalers Bluff Formation is the first to be described in detail from Australia. All marine mammal taxa represent members of extant families and for the most part extant genera: Physeter, Tursiops , and Delphinus or Stenella. Other taxa, of uncertain affinities below family level, include a balaenid, balaenopterid, a third undetermined genus of delphinid and a phocid pinniped. The diversity of marine mammals in the Portland Pliocene assemblage is impoverished relative to extant faunas, with at least 18 marine mammal species regularly occurring in north- west Bass Strait (Wameke in Menkhorst, 1995; pers. obs.). Published details of other SW Pacific Pliocene marine mam- mal assemblages are scanty (Fordyce, 1991a, 1991b; Fitzgerald, 2004b). Only three Australian assemblages provide a reasonable basis for comparison with the Portland Pliocene assemblage; the Beaumaris Local Fauna (Victoria), Grange Burn assemblage (Victoria) and Cameron Inlet assemblage (Flinders Island, Bass Strait). However, only the Cameron Inlet assemblage is approximately contemporaneous with the Portland Pliocene assemblage (about 2. 5-4. 8 Ma), the Cameron Inlet Formation being late Early to Late Pliocene (about 2.0-4.0 Ma; Fitzgerald, 2004b and references therein). The Beaumaris Local Fauna spans the Miocene-Pliocene boundary, with an age range for the Black Rock Sandstone of about 4.5-5. 8 Ma (Dickinson et al., 2002; Wallace et al., 2005). The Grange Burn assemblage is perhaps slightly younger than the Beaumaris Local Fauna, with the Grange Burn Formation being Early Pliocene (>4.35-5.3 Ma) (Dickinson et ah, 2002; Fitzgerald, 2004; Wallace et ah, 2005). Both the Beaumaris Local Fauna and the Grange Burn assemblage are (at least in part) phosphatic nodule bed deposits. A significant problem in making meaningful comparisons between the Portland Pliocene assemblage and the other assem- blages listed above lies in the great disparity in numbers of specimens collected. Whereas the Portland Pliocene and Cameron Inlet assemblages are known from 30-40 specimens, the Beaumaris Local Fauna and Grange Bum assemblage are known from hundreds of fossils. Nevertheless, with this bias in mind some preliminary comparisons can be made. hi all four assemblages almost all marine mammal families present are extant. The sole exception to this is the occurrence of the paraphyletic family ‘Cetotheriidae’ (Fordyce and Barnes, 1994; Fordyce, 2003; Geisler and Luo, 1996; Geisler and Sanders, 2003; Kimura and Ozawa, 2002) in the Beaumaris Local Fauna and Grange Bum assemblage (Fitzgerald, 2004b). At the generic level, there appear to be some differences between the known diversity of marine mammals in the Portland Pliocene assemblage/Cameron Inlet assemblage and the Beaumaris Local Fauna/Grange Bum assemblage. In the latter two, extant genera include cf. Eubalaena, Balaenoptera, Megaptera, Physeter and cf. Mesoplodon. The delphinids from the phosphatic nodule bed assemblages do not appear to be closely related to extant genera, contrary to earlier assessments (e.g., Chapman, 1917). Rather, delphinid periotics and middle ear ossicles from the Beaumaris Local Fauna possess a rela- tively high number of primitive characters with respect to extant Delphinidae. This is markedly different from the Portland Pliocene assemblage in which most described del- phinid periotics seem to represent extant genera. Other appar- ently archaic aspects of the Beaumaris Local Fauna/Grange Bum assemblage include the high diversity of physeterid taxa (three genera, including a small form similar to the form genus Scaldicetus and another similar to Physeterula dubusii Van Beneden, 1877) relative to the present (one genus). The Cameron Inlet assemblage appears to be essentially modem in aspect based on fossils recovered to date. Fordyce (1991b: 1183) listed ziphiids (including Mesoplodon sp., based on isolated periotics), physeterids (cf. Physeter macrocephalus) and balaenopterids in this assemblage. Fitzgerald (2004: 198) included cf. Balaenoptera, cf. Megaptera and a possible del- phinid. Sutherland and Kershaw (1971) reported an incomplete skull (NMV P23961) as Ziphius sp. but Fordyce (1984b: 939) later questioned this assignment and re-identified NMV P23961 as ? Mesoplodon sp. The Cameron Inlet assemblage is generally similar to the Portland Pliocene assemblage which is perhaps not unexpected given their similar geological age and geographic proximity. It has previously been noted that latest Miocene through Pliocene marine mammal assemblages across the globe include extinct, often aberrant, genera and families and formerly wider (in some cases unexpected) geographic ranges for extant taxa that today have restricted geographic distributions (Fordyce et al., 2002 and references therein). Indeed, the apparently recent evolution of marine mammal faunas of modem aspect led Fordyce and colleagues (2002) to suggest that among cetaceans at least a major ecological change occurred about 3-4 million years ago. This change is marked by the last appearance in the fossil record of genus and family-level taxa displaying rela- tively primitive grades of evolution (e.g., Albireonidae, Herpetocetus : Barnes, 1984; Whitmore, 1994; Oishi and Hasegawa, 1995; Barnes and Furusawa, 2001; Fordyce and Muizon, 2001) as well as novel morphological adaptations and inferred palaeoecology (e.g., Odobenocetopsidae, Austral- odelphis : Fordyce et al., 2002; Muizon and Domning, 2002). The early Pliocene is marked by a global warming trend beginning at c. 4.5-5. 5 Ma with the rapid development of full- scale Northern Hemisphere and Antarctic glaciation occurring 86 Erich M.G. Fitzgerald in the late Pliocene at approximately 2.75-3.2 Ma (Zachos et al., 2001; Ravelo et al., 2004; Wara et al., 2005). Gallagher and colleagues (2003) presented data indicating that the earlier Pliocene (3. 1-5.3 Ma) was a time of generally stable marine temperatures in Bass Strait with surface temperatures per- haps as much as 3°C warmer than today (Ravelo et al., 2004). The Late Pliocene in Bass Strait was characterised by a fluctuating, overall cooler climate than the preceding Early Pliocene (Gallagher et al., 2003). As Fordyce and colleagues (2002) have indicated, climatic and oceanic changes associ- ated with rapid global cooling at c. 3.2 Ma were likely major influences on the evolution of the modem marine mammal fauna. A larger sample of marine mammal fossils from Pliocene-aged assemblages in south-east Australia with finer resolution of stratigraphic distribution of marine mammal taxa during the Pliocene is required for more detailed correlations between Pliocene marine mammal and climatic evolution. The description of the Pliocene marine mammal assemblage from Portland alleviates the dearth of information on SW Pacific marine mammal assemblages during the late Neogene. The occurrence of extant cetacean taxa ( Balaenoptera , Physeter, Delphinus/Stenella and Tursiops ) in the Whalers Bluff Formation (>2. 5-4.8 Ma) (and Cameron Inlet Formation) indicates that the marine mammal fauna off south-east Australia had begun to take on a modem aspect by at the earliest 4.8 Ma and latest around 2.5 Ma. Indeed, most of the cetacean species still exist in the seas off Portland (Wameke in Menkhorst, 1995; pers. obs.). That the latter cetacean taxa occur in Early-Late Pliocene deposits in the SW, SE and NE Pacific, and NW Atlantic, suggests that several extant cetacean genera were widespread in the world’s ocean basins prior to about 2.5 Ma. Acknowledgements Our rapidly improving knowledge of late Neogene fossil marine vertebrates off southeast Australia would not be possible without the collecting efforts of Mr Sean Wright and his donation of hundreds of fossils to Museum Victoria, for which he is thanked. The Vertebrate Palaeontology and Mammalogy departments of Museum Victoria are thanked for providing research facilities and access to specimens. D.J. Bohaska is thanked for his assistance with access to specimens during a visit to the National Museum of Natural History (Smithsonian Institution) and R. Purdy assisted with photo- graphic facilities there. The Museum Victoria Library, with the National Museum of Natural History (Smithsonian Institution) Remington Kellogg Library for Marine Mammalogy, were crit- ical in providing access to some otherwise intractable literature. Dr T.A. Darragh is thanked for discussions on late Neogene biostratigraphy. This paper was greatly improved by the helpful review by Ewan Fordyce and editorial comments from Gary Poore. This work forms part of a Ph.D. thesis undertaken in the School of Geosciences, Monash University and Museum Victoria, which was financially supported by an Australian Postgraduate Award. References Abele, C., Gloe, C.S., Hocking, J.B., Holdgate, G., Kenley, P.R., Lawrence, C.R., Ripper, D., Threlfall, W.F., and Bolger, P.F. 1988. Tertiary. Pp. 251-350 in: Douglas, J.G. and Ferguson, J.A. (eds), Geology of Victoria. Victorian Division, Geological Society of Australia Incorporated: Melbourne. Bannister, J.L. 2002. Baleen whales: mysticetes. Pp. 62-72 in: Perrin, W.F., Wiirsig, B., and Thewissen, J.G.M. (eds), Encyclopedia of Marine Mammals. Academic Press: San Diego. Bames, L.G. 1973a. Pliocene cetaceans of the San Diego Formation, San Diego, California. Pp. 37 — 42 in: Ross, A. and Dowlen, R.J. (eds), Studies on the Geology and Geologic Hazards of the Greater San Diego Area, California. San Diego Association of Geologists: San Diego. Bames, L.G. 1973b. Praekogia cedrosensis, a new genus and species of fossil pygmy sperm whale from Isla Cedros, Baja California, Mexico. Contributions in Science, Natural History Museum of Los Angeles County 247: 1-20. Bames, L.G. 1977. Outline of eastern North Pacific fossil cetacean assemblages. Systematic Zoology 25: 321-343. Bames, L.G. 1978. A review of Lophocetus and Liolithax and their relationships to the delphinoid family Kentriodontidae (Cetacea: Odontoceti). Bulletin of the Natural History Museum of Los Angeles County 28: 1-35. Bames, L.G. 1984. Fossil odontocetes (Mammalia: Cetacea) from the Almejas Formation, Isla Cedros, Mexico. PaleoBios 42: 1-46. Bames, L.G. 1985. Fossil pontoporiid dolphins (Mammalia: Cetacea) from the Pacific coast of North America. Contributions in Science, Natural History Museum of Los Angeles County 363: 1-34. Bames, L.G. 1990. The fossil record and evolutionary relationships of the genus Tursiops. Pp. 3-26 in: Leatherwood, S. and Reeves, R.R. (eds), The Bottlenose Dolphin. Academic Press: San Diego. Bames, L.G. 1998. The sequence of fossil marine mammal assemblages in Mexico. Avances En Investigacion. Publicacion Especial 1 (Paleontologia de Vertebrados): 26-79. Bames, L.G. 2002. Delphinoids, evolution of the modem families. Pp. 314-316 in: Perrin, W.F., Wiirsig, B., and Thewissen, J.G.M. (eds), Encyclopedia of Marine Mammals. Academic Press: San Diego. Barnes, L.G., and Furusawa, H. 2001. A primitive albireonid dolphin (Cetacea, Odontoceti, Delphinoidea) from Hokkaido, Japan. Journal of Vertebrate Paleontology 21 (Supplement to No. 3): 31 A. Bames, L.G., and Mitchell, E.D. 1984. Kentriodon obscurus (Kellogg, 1931), a fossil dolphin (Mammalia: Kentriodontidae) from the Miocene Sharktooth Hill Bonebed in California. Contributions in Science, Natural History Museum of Los Angeles County 353: 1-23. Bames, L.G., Howard, H., Hutchison, J.H., and Welton, B.J. 1981. The vertebrate fossils of the marine Cenozoic San Mateo Formation at Oceanside, California. Pp. 53-70 in: Abbott, PL. and O’Dunn, S. (eds), Geologic Investigations of the San Diego Coastal Plain. San Diego Association of Geologists: San Diego. Bames, L.G., Pyenson, N.D., and McLeod, S.A. 2005. Middle Miocene marine ecosystems and cetacean diversity as exemplified by the Sharktooth Hill Bonebed, California, USA. Pp. 6-7 in: Uhen, M.D. (ed), Evolution of Aquatic Tetrapods: Fourth Triannual Convention Abstracts. Cranbrook Institute of Science Miscellaneous Publications 1 . Bearlin, R.K. 1987. The morphology and systematics of Neogene Mysticeti from Australia and New Zealand. Ph.D. Thesis, University of Otago: Dunedin. 212 pp. Pliocene marine mammals 87 Berta, A. 1991. New Enaliarctos* (Pinnipedimorpha) from the Oligocene and Miocene of Oregon and the role of “enaliarctids” in pinniped phylogeny. Smithsonian Contributions to Paleobiology 69: 1-33. Berta, A., and Wyss, A.R. 1994. Pinniped phylogeny. Pp. 33-56 in: Berta, A. and Demere, T.A. (eds), Contributions in Marine Mammal Paleontology Honoring Frank C. Whitmore, Jr. Proceedings of the San Diego Society of Natural History 29. Beu, A.G. 1995. Pliocene limestones and their scallops: lithostrati- graphy, pectinid biostratigraphy and paleogeography of eastern North Island late Neogene limestone. Institute of Geological and Nuclear Sciences Monograph 10: 1-243. Beu, A.G., and Darragh, T.A. 2001. Revision of southern Australian Cenozoic fossil Pectinidae (Mollusca: Bivalvia). Proceedings of the Royal Society of Victoria 113: 1-205. Bisconti, M. 2003. Evolutionary history of Balaenidae. Cranium 20: 9-50. Bisconti, M. 2005. Skull morphology and phylogenetic relationships of a new diminutive balaenid from the Lower Pliocene of Belgium. Palaeontology 48: 793-816. Boutakoff, N., and Sprigg, R.C. 1953. Summary report on the petrol- eum possibilities of the Mount Gambier sunklands. Mining and Geological Journal, Department of Mines, Victoria 5: 28-42. Brownell, R.L., Jr., and Herald, E.S. 1972. Lipotes vexillifer. Mammalian Species 10: 1-4. Chapman, F. 1917. New or little-known Victorian fossils in the National Museum. Part XXI. Some Tertiary cetacean remains. Proceedings of the Royal Society of Victoria 30: 32-43. Cummings, W.C. 1985. Right Whales, Eubalaena glacialis (Muller, 1776) and Eubalaena australis (Desmoulins, 1822). Pp. 275-304 in: Ridgway, S.H. and Harrison, R. (eds), Handbook of Marine Mammals. Volume 3: The Sirenians and Baleen Whales. Academic Press: London. Dawson, S.D. 1996. A description of the skull and postcrania of Hadrodelphis calvertense Kellogg 1966, and its position within the Kentriodontidae (Cetacea; Delphinoidea). Journal of Vertebrate Paleontology 16: 125-134. Demere, T.A., Berta, A., and Adam, PJ. 2003. Pinnipedimorph evolu- tionary biogeography. Bulletin of the American Museum of Natural History 279: 32-76. Dickinson, J.A., Wallace, M.W., Holdgate, G.R., Gallagher, S J., and Thomas, L. 2002. Origin and timing of the Miocene-Pliocene unconformity in southeast Australia. Journal of Sedimentary Research 72: 288-303. Dixon, J.M. 1990. Record of a southern right whale ( Eubalaena australis ) skeleton from Altona Bay, Victoria, Australia. The Victorian Naturalist 107: 159-162. Evans, H.E. 1993. Miller’s Anatomy of the Dog. Third Edition. W. B. Saunders Company: Philadelphia. 1113 pp. Fitzgerald, E.M.G. 2004a. The fossil record of cetaceans (Mammalia) on the Australian continent. Fossil Collector Bulletin 72: 5-32. Fitzgerald, E.M.G. 2004b. A review of the Tertiary fossil Cetacea (Mammalia) localities in Australia. Memoirs of Museum Victoria 61: 183-208. Fordyce, R.E. 1982. A review of Australian fossil Cetacea. Memoirs of the National Museum of Victoria 43: 43-58. Fordyce, R.E. 1984a. Preliminary report on cetacean bones from Chatham Rise (New Zealand). Geologisches Jahrbuch, Reihe D Heft 65: 117-120. Fordyce, R.E. 1984b. Evolution and zoogeography of cetaceans in Australia. Pp. 929-948 in: Archer, M. and Clayton, G. (eds), Vertebrate Zoogeography and Evolution in Australasia. Hesperian Press: Perth. Fordyce, R.E. 1989. Origin and evolution of Antarctic marine mammals. Pp. 269-281 in: Crame, J.A. (ed), Origins and Evolution of the Antarctic Biota. Geological Society Special Publication No. 47. Fordyce, R.E. 1991a. Anew look at the fossil vertebrate record of New Zealand. Pp. 1191-1316 in: Vickers-Rich, P, Monaghan, J.M., Baird, R.F., and Rich, T.H. (eds), Vertebrate Palaeontology of Australasia. Pioneer Design Studio in cooperation with the Monash University Publications Committee: Melbourne. Fordyce, R.E. 1991b. The Australasian marine vertebrate record and its climatic and geographic implications. Pp. 1165-1190 in: Vickers- Rich, P, Monaghan, J.M., Baird, R.F., and Rich, T.H. (eds), Vertebrate Palaeontology of Australasia. Pioneer Design Studio in cooperation with the Monash University Publications Committee: Melbourne. Fordyce, R.E. 1994. Waipatia maerewhenua, new genus and new species (Waipatiidae, new family), an archaic Late Oligocene dolphin (Cetacea: Odontoceti: Platanistoidea) from New Zealand. Pp. 147-176 in: Berta, A., and Demere, T.A. (eds), Contributions in Marine Mammal Paleontology Honoring Frank C. Whitmore, Jr. Proceedings of the San Diego Society of Natural History 29. Fordyce, R.E. 2002a. Fossil sites. Pp. 471-482 in: Perrin, W.F., Wiirsig, B., and Thewissen, J.G.M. (eds), Encyclopedia of Marine Mammals. Academic Press: San Diego. Fordyce, R.E. 2002b. Oligocene origins of skim-feeding right whales: a small archaic balaenid from New Zealand. Journal of Vertebrate Paleontology 22 (Supplement to No. 3): 54A. Fordyce, R.E. 2003. Cetacean evolution and Eocene-Oligocene oceans revisited. Pp. 154-170 in: Prothero, D.R., Ivany, L.C., and Nesbitt, E.A. (eds), From Greenhouse to Icehouse: the Marine Eocene-Oligocene Transition. Columbia University Press: New York. Fordyce, R.E., and Bames, L.G. 1994. The evolutionary history of whales and dolphins. Annual Review of Earth and Planetary Sciences 22: 419^455. Fordyce, R.E., and Campbell, H.J. 1990. Fossil dolphin bones from the Chatham Islands, New Zealand. New Zealand Natural Sciences 17: 61-65. Fordyce, R.E., and Flannery, T.F. 1983. Fossil phocid seals from the late Tertiary of Victoria. Proceedings of the Royal Society of Victoria 95: 99-100. Fordyce, R.E., and Muizon, C. de. 2001. Evolutionary history of cetaceans: a review. Pp. 169-233 in: Mazin, J.-M. and Buffrenil, V. de (eds). Secondary Adaptation ofTetrapods to Life in Water. Verlag Dr. Friedrich Pfeil: Munchen. Fordyce, R.E., Quilty, P.G., and Daniels, J. 2002. Australodelphis mirus, a bizarre new toothless ziphiid-like fossil dolphin (Cetacea: Delphinidae) from the Pliocene of Vestfold Hills, East Antarctica. Antarctic Science 14: 37-54. Gallagher, S.J., Greenwood, D.R., Taylor, D., Smith, A.J., Wallace, M.W., and Holdgate, G.R. 2003. The Pliocene climatic and envi- ronmental evolution of southeastern Australia: evidence from the marine and terrestrial realm. Palaeogeography, Palaeoclimatology, Palaeoecology 193: 349-382. Geisler, J.H., and Luo, Z. 1996. The petrosal and inner ear of Herpetocetus sp. (Mammalia: Cetacea) and their implications for the phylogeny and hearing of archaic mysticetes. Journal of Paleontology 70: 1045-1066. Geisler, J.H., and Luo, Z. 1998. Relationships of Cetacea to terrestrial ungulates and the evolution of cranial vasculature in Cete. Pp. 163-212 in: Thewissen, J.G.M (ed), The Emergence of Whales: Evolutionary Patterns in the Origin of Cetacea. Plenum Press: New York. Erich M.G. Fitzgerald Geisler, J.H., and Sanders, A.E. 2003. Morphological evidence for the phylogeny of Cetacea. Journal of Mammalian Evolution 10: 23-129. Gill, E.D. 1957. The stratigraphic al occurrence and palaeoecology of some Australian Tertiary marsupials. Memoirs of the National Museum of Victoria 21: 135-203. Gottfried, M.D., Bohaska, D.J., and Whitmore, F.C., Jr. 1994. Miocene cetaceans of the Chesapeake Group. Pp. 229-238 in: Berta, A. and Demere, T.A. (eds), Contributions in Marine Mammal Paleontology Honoring Frank C. Whitmore, Jr. Proceedings of the San Diego Society of Natural History 29. Gradstein, F.M., Ogg, J.G., Smith, A.G., Agterberg, F.P., Bleeker, W., Cooper, R.A., Davydov, V., Gibbard, P, Hinnov, L., House, M.R., Lourens, L., Luterbacher, H.-P., McArthur, J., Melchin, M.J., Robb, L.J., Shergold, J., Villeneuve, M., Wardlaw, B.R., Ali, J., Brinkhuis, H., Hilgen, F.J., Hooker, J., Howarth, R.J., Knoll, A.H., Laskar, J., Monechi, S„ Powell, J., Plumb, K.A., Raffi, I., Rohl, U., Sadler, P, Sanfilippo, A., Schmitz, B., Shackleton, N.J., Shields, G.A., Strauss, H., Van Dam, J., Veizer, J., van Kolfschoten, T., and Wilson, D. 2004. A Geologic Time Scale 2004. Cambridge University Press: Cambridge. 500 pp. Hendey, Q.B., and Repenning, C.A. 1972. A Pliocene phocid from South Africa. Annals of the South African Museum 59: 71-98. Howchin, W. 1919. [Tympanic bone of Balaena from the Pliocene of South Australia]. Transactions and Proceedings of the Royal Society of South Australia 43: 430. Howell, A.B. 1929. Contribution to the comparative anatomy of the eared and earless seals (genera Zalophus and Phoca ). Proceedings of the United States National Museum 73: 1-142. Kasuya, T. 1973. Systematic consideration of Recent toothed whales based on the morphology of tympano-periotic bone. The Scientific Reports of the Whales Research Institute 25: 1-103. Kellogg, R. 1965. Fossil marine mammals from the Miocene Calvert Formation of Maryland and Virginia. Part 2. The Miocene Calvert sperm whale Orycterocetus. United States National Museum Bulletin 247: 47-63. Kellogg, R. 1966. Fossil marine mammals from the Miocene Calvert Formation of Maryland and Virginia. Part 4. Anew odontocete from the Calvert Miocene of Maryland. United States National Museum Bulletin 247: 99-101. Kimura, T., and Ozawa, T. 2002. Anew cetothere (Cetacea: Mysticeti) from the Early Miocene of Japan. Journal of Vertebrate Paleontology 22: 684-702. LeDuc, R.G., Perrin, W.F., and Dizon, A.E. 1999. Phylogenetic relationships among the delphinoid cetaceans based on full cytochrome b sequences. Marine Mammal Science 15: 619-648. LeDuc, R.G. 2002. Delphinids, overview. Pp. 310-314 in: Perrin, W.F., Wiirsig, B., and Thewissen, J.G.M. (eds), Encyclopedia of Marine Mammals. Academic Press: San Diego. Lydekker, R. 1887. The Cetacea of the Suffolk Crag. Quarterly Journal of the Geological Society 43: 7-18. McKee, J.W.A., and Fordyce, R.E. 1987. Dolphin mandible (Delphinidae) from the Waipipian Stage (Pliocene), Waihi Beach, Taranaki, New Zealand. New Zealand Journal of Geology and Geophysics 30: 321-323. McLeod, S.A., Whitmore, F.C., Jr., and Barnes, L.G. 1993. Evolutionary relationships and classification. Pp. 45-70 in: Burns, J.J., Montague, J.J., and Cowles, C.J. (eds), The Bowhead Whale. The Society for Marine Mammalogy Special Publication Number 2. Menkhorst, P.W. 1995. Mammals of Victoria: Distribution, ecology and conservation. Oxford University Press: Melbourne. 359 pp. Miller, G.S., Jr. 1918. A new river-dolphin from China. Smithsonian Miscellaneous Collections 68(2486): 1-12. Miller, G.S., Jr. 1924. A Pollack whale from Florida presented to the National Museum by the Miami Aquarium Association. Proceedings of the United States National Museum 66 (Article 9): 1-15. Miyazaki, N., and Perrin, W.F. 1994. Rough- toothed dolphin Steno bredanensis (Lesson, 1828). Pp. 1-21 in: Ridgway, S.H., and Harrison, R. (eds), Handbook of Marine Mammals, Volume 5: The First Book of Dolphins. Academic Press: London. Muizon, C. de. 1981. Les vertebres fossiles de la formation Pisco (Perou). Premiere partie: Deux nouveaux Monachinae (Phocidae, Mammalia) du Pliocene de Sud-Sacaco. Travaux de I’lnstitut Frangais d’ Etudes Andines 22: 1-161. Muizon, C. de. 1984. Les vertebres fossiles de la Formation Pisco (Perou). Deuxieme partie: Les odontocetes (Cetacea, Mammalia) du Pliocene inferieur de Sud-Sacaco. Travaux de l ’Institut Frangais d’ Etudes Andines 27: 1-188. Muizon, C. de. 1988. Les vertebres fossiles de la Formation Pisco (Perou). Troisieme partie: Les odontocetes (Cetacea, Mammalia) miocenes. Travaux de T Institut Frangais d’ Etudes Andines 42: 1-244. Muizon, C. de, and Bellon, H. 1986. Nouvelles donnees sur V age de la Formation Pisco (Perou). Comptes-Rendus de TAcademie des Sciences Paris. Serie II. Mecanique-Physique-Chimie, Sciences de I’Univers, Sciences de la Terre 303: 1401-1404. Muizon, C. de, and DeVries, T.J. 1985. Geology and paleontology of late Cenozoic marine deposits in the Sacaco area (Peru). Geologische Rundschau 74: 547-563. Muizon, C. de. and Domning, D.P. 1985. The first records of fossil sirenians in the southeastern Pacific Ocean. Bulletin du Museum Nationale d’Histoire Naturelle (Paris) (4)7, Sect. C, 3: 189-213. Muizon, C. de, and Domning, D.P. 2002. The anatomy of Odobenocetops (Delphinoidea, Mammalia), the walrus-like dolphin from the Pliocene of Peru and its palaeobiological implications. Zoological Journal of the Linnean Society 134: 423-452. Muizon, C. de, and Hendey, Q.B. 1980. Late Tertiary seals of the South Atlantic Ocean. Annals of the South African Museum 82: 91-128. Muizon, C. de, McDonald, H.G., Salas, R., and Urbina, M. 2004. The youngest species of the aquatic sloth Thalassocnus and a reassess- ment of the relationships of the nothrothere sloths (Mammalia: Xenarthra). Journal of Vertebrate Paleontology 24: 387-397. Oishi, M., and Hasegawa, Y. 1994 (1995). Diversity of Pliocene mysticetes from eastern Japan. The Island Arc 3: 436-452. Prothero, D.R. 1998. The chronological, climatic, and paleogeo- graphic background to North American mammalian evolution. Pp. 9-36 in: Janis, C.M., Scott, K.M., and Jacobs, L.J. (eds), Evolution of Tertiary Mammals of North America. Volume 1: Terrestrial Carnivores, Ungulates, and Ungulatelike Mammals. Cambridge University Press: Cambridge. Ravelo, A.C., Andreasen, D.H., Lyle, M., Lyle, A.O., and Wara, M.W. 2004. Regional climate shifts caused by gradual cooling in the Pliocene epoch. Nature 429: 263-267. Reeves, R.R., and Leatherwood, S. 1985. Bowhead Whale, Balaena mysticetus Linnaeus, 1758. Pp. 305-344 in: Ridgway, S.H. and Harrison, R. (eds), Handbook of Marine Mammals. Volume 3: The Sirenians and Baleen Whales. Academic Press: London. Repenning, C.A., and Tedford, R.H. 1977. Otarioid seals of the Neogene. United States Geological Survey Professional Paper 992: 1-93. Rice, D.W. 1998. Marine Mammals of the World: Systematics and Distribution. The Society for Marine Mammalogy Special Publication Number 4. 231 pp. Pliocene marine mammals 89 Singleton, O.P., McDougall, I., and Malle tt, C.W. 1976. The Pliocene- Pleistocene boundary in southeastern Australia. Journal of the Geological Society of Australia 23: 299-311. Sutherland, F.L., and Kershaw, R.C. 1971. The Cainozoic geology of Flinders Island, Bass Strait. Papers and Proceedings of the Royal Society of Tasmania 105: 151-175. Wallace, M.W., Dickinson, J.A., Moore, D.H., and Sandiford, M. 2005. Late Neogene strandlines of southern Victoria: a unique record of eustasy and tectonics in southeast Australia. Australian Journal of Earth Sciences 52: 219-291. Walsh, S., and Naish, D. 2002. Fossil seals from late Neogene deposits in South America: a new pinniped (Carnivora, Mammalia) assemblage from Chile. Palaeontology 45: 821-842. Wara, M.W., Ravelo, A.C., and Delaney, M.L. 2005. Permanent El Nino-like conditions during the Pliocene warm period. Science 309: 758-761. Watson, A.G., and Fordyce, R.E. 1994. Skeleton of two minke whales, Balaenoptera acutorostrata, stranded on the south-east coast of New Zealand. New Zealand Natural Sciences 20: 1-14. Westgate, J.W., and Whitmore, F.C., Jr. 2002. Balaena ricei, a new species of bowhead whale from the Yorktown Formation (Pliocene) of Hampton, Virginia. Pp. 295-312 in: Emry, R.J. (ed), Cenozoic Mammals of Land and Sea: Tributes to the Career of Clayton E. Ray. Smithsonian Contributions to Paleobiology 93. Whitmore, F.C., Jr. 1994. Neogene climatic change and the emergence of the modem whale fauna of the North Atlantic Ocean. Pp. 223-227 in: Berta, A. and Demere, T.A. (eds). Contributions in Marine Mammal Paleontology Honoring Frank C. Whitmore, Jr. Proceedings of the San Diego Society of Natural History 29. Wyss, A.R., and Flynn, J.J. 1993. A phylogenetic analysis and defini- tion of the Carnivora. Pp. 32-52 in: Szalay, F.S., Novacek, M.J. and McKenna, M.C. (eds), Mammal Phylogeny. Springer- Verlag: Berlin. Zachos, J., Pagani, M., Sloan, L., Thomas, E., and Billups, K. 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292: 686-693. Memoirs of Museum Victoria 62(1): 91-99 (2005) ISSN 1447-2546 (Print) 1447-2554 (On-line) http://www.museum.vic.gov.au/memoirs/index.asp Two new Middle Miocene spatangoids (Echinoidea) from the Murray Basin, South Australia Francis C. Holmes 1 , Christopher Ah Yee and Janice Krause 2 ‘15 Kenbry Road, Heathmont, Victoria 3135, Australia, and Invertebrate Palaeontology, Museum Victoria, PO Box 666, Melbourne, Victoria 3001, Australia (fholmes@bigpond.net.au) 2 P.O. Box 581, Hamilton, Victoria 3300, Australia Abstract Holmes, F.C., Ah Yee, C., and Krause, J. 2005. Two new Middle Miocene spatangoids (Echinoidea) from the Murray Basin, South Australia. Memoirs of Museum Victoria 62(1): 91-99. Two new spatangoid taxa are described from the Glenforslan Formation cropping out in the Murray River cliffs near Blanchetown, South Australia. One taxon, Murraypneustes biannulatus gen. et sp. nov., a large species of spatangoid with two ‘peripetalous’ fascioles (one circling the margin and the other close to the distal end of the relatively short petals), two distinct sizes of aboral primary tubercles, and a depressed apical system. The other spatangoid described, Spatagobrissus dermodyorum sp. nov. differs from the only other fossil species of this genus recorded from Australia, S. laubei (Duncan, 1877), in having a much shorter labrum, markedly larger peristome and periproct and larger primary tubercles within the peripetalous fasciole. Keywords Echinoidea, Spatangoida, Murrraypneustes, Spatagobrissus, new taxa, Middle Miocene, South Australia Introduction Of all the Australian Tertiary sedimentary sequences that con- tain extensive echinoid faunas, the 150 km of Murray River cliffs, from Overland Comer, east of Waikerie, to Murray Bridge, South Australia, have over the last 120 years been more comprehensively studied by palaeontologists, students and amateur collectors than any other similar area. In the Miocene stratigraphic sequences along the Murray River and elsewhere in Australia, species belonging to the Spatangoida constitute approximately 50 percent of the recorded taxa of irregular echinoids. In view of the number of spatangoids occurring in this section of the Murray River cliffs, the discovery in 2003 of three specimens of a new genus belonging to this order, seem- ingly unrelated to any other genus known from the continent’s fossil record, was unexpected. The specimens were found within 300 m of each other in a single bed of the Glenforslan Formation, cropping out on the left bank of the Murray River, 7 km NNE of Blanchetown, South Australia [Museum Victoria locality PL3203]. Further investigation of the surface exposure of the formation in the general vicinity of this discovery failed to produce any additional specimens of the new genus. Materials and methods. Specimen numbers prefixed P, on which the studies are based, are housed in the Invertebrate Palaeontology collection, Museum Victoria (NMV). In addition to type material, the new species of Spatagobrissus is repre- sented by several specimens in private collections. Measure- ments were made with a dial calliper to an accuracy of 0. 1 mm. Parameters are expressed as a percentage of test length (%TL) or test width (%TW). Age and stratigraphy The Glenforslan Formation, synonymous with the Lower Morgan limestone, conformably overlies the Finniss Formation and is of early Middle Miocene (Batesfordian, Langian) age. The thickness of the unit is relatively consistent at 13-15 m although this is reduced in southern exposure due to post- Middle Miocene uplift and subsequent erosion. The formation is sublithified to lithified, weathering to whitish-cream colour in outcrop (Lukasik and James, 1998). It is composed of cycles of mollusc-bryozoan floatstone, with a microbioclastic pack- stone matrix, grading upward into Celleporaria rudstone tops (infaunal bivalve and gastropod-rich microbioclastic matrix with large branching and sheeted Celleporaria) which also contain pectens and oysters. Echinoids tend to be found at or above the rudstone-floatstone contact at the base of the cycles, the latter containing delicate branching and uni-laminar sheet bryzoans of 1-3 cm length. Sediments are pervasively mottled, obscuring all physical sedimentary textures. The middle 92 Francis C. Holmes, Christopher Ah Yee and Janice Krause Glenforslan Formation is interpreted as being deposited in relatively shallow waters, possibly less than 10 m, based on the presence of calcareous algae and mixotrophic foraminifers (Lakasik, pers. com. 2005). It form s part of the richest warm- water biotic record from southern Australia at a time of maxi- mum transgression of the sea across the continental shelf (McGowran and Li, 1994, and papers cited therein). The three specimens of the new genus were found about 7.4 m above the base of the formation, approximately 2 m above the Lepidocyclina Zone. Associated fauna Nineteen species of echinoids have been recorded from the Glenforslan Formation within 500 m upstream and downstream of PL3203 (Table 1), compared with 28 confirmed species known to occur within the Morgan Group (Glenforslan, Cadell, and Bryant Creek Formations). Table 1. Echinoids recorded within the vicinity of locality PL3203. Letters in brackets indicate the frequency of occurrence of each species. [A], abundant; [C], common; [F], fairly common; [U], uncom- mon; [R], rare. References to authors and supporting literature cited below, but not listed in the main text references, can be found in Holmes (1993). Cidaroida Goniocidaris murrayensis Chapman and Cudmore, 1934 [C] Arbacoida Murray echinus paucituberculatus (Gregory, 1890) [F] Temnopleuroida Cryptechinus humilior ( Bittner, 1892) [C] Ortholophus morganensis Philip, 1969 [C] O. pulchellus (Bittner, 1892) [C] Clypeasteroida Monostychia australis Laube, 1869 [A] M. sp. ‘C’ in Holmes, 1999 [F] Scutellinoides patella (Tate, 1891) [C] Spatangoida Brissopsis tatei Hall, 1907 [U] Brissus sp. nov? [R] Cyclaster archer i (Tenison Woods, 1867) [C] Eupatagus rotundus Duncan, 1877 [U] Eupatagus sp. indet. [U] Hemiaster (Bolbaster) planedeclivis Gregory, 1890 [C] Lovenia cf. forbesi (Tenison Woods, 1862) [C] Murraypneustes biannulatus gen. et sp. nov. [R] Pericosmus compressus (Duncan, 1877) [R] Protenaster antiaustralis (Tate, 1885) [U] Spatagobrissus dermodyorum sp. nov [F] Systematic Palaeontology Order Spatangoida Claus, 1876 Suborder Micrasterina Fisher, 1966 Family Incertae sedis Remarks. The combination of generic features, particularly the presence of a marginal ‘peripetalous’ fasciole, intermittent hor- izontal fasciole bands and a rudimentary non re-entrant peripetalous fasciole clear of the distal ends of paired petals, clearly distinguish the new genus Murraypneustes from virtu- ally all other taxa assigned to the Micrasterina, in particular the 22 genera included in the Asterostomatidae Fisher, 1966. Only one of these genera, Asterostoma Agassiz, 1847, has tenta- tively been retained in the Asterostomatidae (together with Stomaporus Cotteau, 1888, a genus originally assigned to the Brissidae Gray, 1855) by Smith et al. (2003). Of the remaining 21 genera, Smith et al. have conditionally reassigned 13 to other families, only six of which are included in the Micrasterina. The remaining eight are listed as incertae sedis or as belonging to an unnamed taxon. Until a detailed revision of these latter genera is published, it is considered imprudent to assign the new genus to a specific family. Murraypneustes gen. nov. Type and only known species. Murraypneustes biannulatus sp. nov. Diagnosis. Large ovoid spatangoid with centrally depressed adapical surface, apex well posterior of centre. Apical system ethmolytic with 4 gonopores. Aboral primary tubercles of 2 dis- tinct sizes, small and randomly spaced, the larger proximal to the ambitus. Labrum long, narrow, partially tuberculate, extending to third ambulacral plate. Two ‘peripetalous’ fascioles present; one marginal, passing above the periproct (pseudolateral); the second, rudimentary, non re-entrant and clear of the distal ends of paired petals. Intermittent horizontal fasciole bands also occur between the 2 ‘peripetalous’ fascioles. Subanal fasciole in contact with marginal periproct. Etymology. For the Murray River cliffs, the origin of the fossils, and “pneustes”, a common suffix used for spatangoid echinoids. Gender masculine. Remarks. The following analysis of morphological features of genera similar to Murraypneustes gen. nov. is based primarily on Mortensen (1950a) and Smith et al. (2003). Although approximately 30 genera within the Micrasterina have been investigated, only eight warranted further scrutiny; four cur- rently unassigned to a family by Smith et al. (2003), Elipneustes Koehler, 1914, Eurypatagus Mortensen, 1948, Linopneustes A. Agassiz, 1881, and Platybrissus Grube, 1865; three belonging to the Maretiidae (Lambert, 1905), Eupatagus L. Agassiz, 1847, Mazzettia Lambert and Thiery, 1915 and Spatagobrissus H. L. Clark, 1923; and Macropneustidae genus Lajanaster Lambert and Sanchez Roig, 1924. Using a broad comparison of 38 features, Linopneustes stands out from the other seven genera as having closest affin- ity with Murraypneustes ; four of its recorded six species, L. longispinus (A. Agassiz, 1878), L. fragilis (de Meijere, 1903), L. spectabilis (de Meijere, 1903), and L. brachipetalus Mortensen, 1950b having distinct marginal peripetalous fascioles passing above the periproct. In the other two, L. murrayi (A. Agassiz, 1879) and L. excentricus de Meijere, 1903, the peripetalous fasciole is not marginal but somewhat higher up on the test (Mortensen, 1950a: 221). Although L. fragilis has multiple fasciole bands around the ambitus, Two new middle Miocene spatangoids 93 L. murrayi a rounded margin (Smith, pers. com. 2005) and L. longispinus relatively short closing petals; Murmypneustes is distinguished by its centrally depressed aboral surface form- ing four apices (domed on Linopneustes ) with the highest point of the test posterior to the apical disk (anterior in Linopneustes ), the presence of two ‘peripetalous’ fascioles and intermittent horizontal fasciole bands, two distinct sizes of small randomly spaced aboral primary tubercles, the larger restricted to the area outside the upper ‘peripetalous’ fasciole, and a prominent angular subanal fasciole. Based on the same criteria, three other genera show a mod- erate degree of morphological similarity to Murraypneustes, namely Elipneustes, Mazzettia, and Eupatagus. Elipneustes, however, can be distinguished by its very long, parallel- sided, open-ended, flush petals with conjugate pore pairs; single, narrow, marginally situated peripetalous fasciole; peristome situated immediately below the apical disk; small thickset plastron with minimal posterior swelling; and much larger primary tubercles scattered over all the aboral inter- ambulacra. Mazzettia, a poorly known fossil genus, has an elongated heart-shaped test with a low sharp margin, very long weakly- bowed petals closing distally, no recorded peripetalous fasciole and only occasionally a weak shield-shaped subanal fasciole, an elongated labrum just reaching the relatively small plastron, and random but closely spaced coarse tubercles aborally. Eupatagus, although possessing well developed peri- petalous and subanal fascioles is easily categorised by its lack of an anterior sulcus, predominately short lanceolate closed petals, primary tubercles of varying density restricted to the area within the peripetalous fasciole in aboral interlambracra 1-4, and very small evenly spaced tubercles distally over the remainder of the aboral surface. Comparison of other features is difficult because of variability among the very large number of described species. In particular, the eight Australian fossil species vary considerably in profile, are generally more round- ed, and have an extremely wide range of primary tubercle densities when compared with the type species on which the generic description is based. Two other genera, Platybrissus and Eurypatagus, are distin- guished by their complete lack of fascioles, although the former may have a subanal fasciole present in juvenile specimens. Lajanaster, although depressed aborally, has no anterior sulcus, a relatively sharp ambitus, a markedly narrow plaston, and primary tubercles confined to the posterior column of anterior and lateral interambulacra within the peripetalous fasciole. Spatagobrissus was included in the comparative analysis, primarily as a new species of the genus occurs at the same locality (PL3203) as Murraypneustes. In most respects the for- mer is very similar to Eupatagus but has shorter petals and only small tubercles over the whole of the aboral surface. Murraypneustes biannulatus. sp. nov. Figures 2A-F, 3A-C, 4A, B, 5 Type material. Holotype, NMY P3 12370 from early Middle Miocene Glenforslan Formation (Batesfordian), Morgan Group, 7 km NNE of Murray River Lock 1, Blanchetown, South Australia [NMV locality PL3203]. Paratypes, NMY P3 12371 and P3 12372 from the same location. Diagnosis. As for genus. Description. Test moderately large, ovoid in outline with shal- low anterior sulcus and pointed, slightly truncated posterior. Specimens range from 72 to 81 mm in length, with maximum width 81-86%TL occurring at 45%TL from anterior ambitus. Test 44.1%TL high (uncompressed specimen) with apex of all specimens well posterior of centre, 63-67%TL from anterior ambitus. Centre of test on adapical surface depressed around apical disk and proximal end of paired petals to form minor apices in interambulacra 1, 4 and 5, and conjointly, a raised area across ambulacrum III and interambulcra 2 and 3. Adoral surface mildly concave in vicinity of peristome with ambulacrum III slightly recessed anteriorly and the plastron forming a fairly pronounced keel posteriorly, terminating at the anterior edge of the subanal fasciole. Primary tubercles on adapical surface of 2 distinct sizes, both widely and randomly spaced. Larger of the two restricted to the distal 35% of the radius on interambulacrum 5, and 25% elsewhere. Adorally, larger primary tubercles closely and evenly spaced throughout, with exception of naked areas in ambulacra I and V and phyllode plates of ambulacra II, III and IV. Overlapping scrobicules on adoral surface form distinct diagonal ridges. Small tubercles closely spaced immediately below ambitus increase in size to that of larger primary tubercles below curvature of margin. Large primary tubercles, perforate with undercut mamelon and crenulated platform, and maximum scrobicular diameter of approximately 2.5 mm, twice size of smaller counterparts. Ring of scrobicular tubercles not always present. ‘Peripetalous’ and subanal fascioles present, the former, though not continuously visible on any specimen clearly passes above periproct, while latter is distinctly angular, form- ing hexagonal outline where in contact with lower edge of periproct. Intermittent horizontal fasciole bands also present laterally above and parallel to marginal ‘peripetalous’ fasciole. Uppermost very narrow and more continuous fasciole, although clear of distal ends of petals, appears to be a rudi- mentary peripetalus fasciole not indented interradially (Fig. 3). Apical system anterior of centre, 39.5-42.2%TL from anter- ior ambitus to centre of disk, level or slightly below proximal end of paired petals. Ethmolytic, with 4 small closely spaced gonopores approximately 0.3 mm in diameter, anterior pair closer together than posterior pair. Detail of ocular plates indeterminate. Hydropores numerous, approximately 70 visible in one specimen, centrally located but extending between posterior pair of gonopores and possibly ocular plates I and V (Fig. 4A, B). Petals lanceolate, moderately wide at midlength, closed dis- tally, situated in gentle concave depressions incorporating adradial edges of adjacent interambulacra and continuing prox- imally across apical disk. Anterior paired petals shorter than posterior pair, extending on average 60% of the radius meas- ured along the surface of the perradial suture from centre of 94 Francis C. Holmes, Christopher Ah Yee and Janice Krause apical disk to ambitus; posterior pair about 56% radius. Inner pores of petals oval, outer pores slot-like, slightly curved and 50% wider. Pore pairs not conjugate but linked by a fine ridge which extends along both sides of each pore (Fig. 5). Maximum width of interporiferous zone slightly more than twice width of poriferous zone. Anterior paired petals diverge at approxim- ately 135° and contain on average 23 pore pairs, posterior petals 297° and 26 pairs. Secondary tubercles extend randomly across interporiferous and poriferous zones and for a distance outside petals without primary tubercles. Ambulacrum III not petaloid, basically flush with adjoining interambulacra for about 50% radius, then gradually becoming concave towards anterior sulcus. Other details unknown, no sign of pores or regularly spaced tubercles being visible on specimens. Peristome reniform, centre situated 27-30%TL from anterior ambitus, longitudinal dimension approximately 6.5%TL, transverse dimension 12%TL. Phyllodes short and not particularly well developed. Labrum long and narrow, averaging 20.5%TL, slightly curved at junction with peristome and abutting the plastron at centre of third pair of adjacent ambulacral plates. Numerous small tubercles adjacent to peristome with a few larger ones towards the posterior end (Fig. 2C). Plastron closely tuberculate, width approximately 75% length measured from posterior edge of labrum to anterior edge of subanal fasciole. Strong posterior taper suggests sixth and subsequent plates of ambulacra I and V indent behind paired episternal plates. Periproct opening marginal, not visible from above, tear shaped, slightly wider than high, set in a slight truncation approximately 65° to the horizontal. Etymology, biannulatus (L) - two-ringed, referring to the pres- ence of two ‘peripetalous’ fasciole rings. Family Maretiidae Spatagobrissus H. L. Clark, 1923 Type species. Spatagobrissus mirabilis H. L. Clark, 1923, by original designation. Diagnosis. See H. L. Clark (1923: 402) Spatagobrissus dermodyorum sp. nov Figures 6A-E, 7A-C, 8A-E Type Material. Holotype. NMV P3 12570 from early Middle Miocene Glenforslan Formation (Batesfordian), Morgan Group, in the vicinity of NMV locality PL3203, 7 km NNE of Murray River Lock 1, Blanchetown, South Australia. Paratypes, NMV P312571-P312373 from the same general area. Other material used for statistical purposes is held in Museum Victoria and private collections. Description. Test small, subcircular in outline with minimal posterior truncation and no anterior sulcus. Specimens range 30.0-45.5 mm in length with maximum width 82-90%TL occurring 51-58 %TL from anterior ambitus. Maximum height 49.5-60%TL at 52.6-65.5%TL from anterior ambitus. Adapical surface moderately inflated, evenly curved transversely above well-rounded margin with ambitus situated at about 30%TH. Adoral surface very mildly convex but with prominent posterior keel caused by sharp rise of ambulacra I and V posterior to centre. In lateral view, posterior truncation covers about 35%TH. Apical system ethmolytic with 4 genital pores, centre 31.3-38.6%TL from anterior ambitus (Fig. 7A). Paired petals short (26.5-32.5%TL measured along surface of perradial suture from centre of apical disk), narrow (7.0-9.0%TL), lance- olate, closing to closed, posterior pair marginally longer than anterior pair. Anterior pair diverge at about 135°, posterior pair 310°. Pore pairs conjugate, inner pores oval, outer-tear shaped. Anterior row of pore pairs in anterior paired petals distinctly narrower than posterior row and atrophied adapically. Interporiferous zone up to 1.5 times width of poriferous zone at widest point. Ambulacrum III flush adapially, with 2 rows of indistinct longitudinally orientated pore pairs and interporifer- ous zone containing few secondary tubercles and numerous miliaries. Peripetalous fasciole subcircular, not indented. Numerous small, randomly spaced perforate, crenulate, primary tubercles occur in interambulacra within fasciole. Outside fasciole, tubercles in posterior half of test very small, but anterior to centre, gradually increasing in size towards interambulacra 2 and 3. Peristome reniform, mildly sunken, width 16.0-19. 3%TL, length 8.9-10.3%TL with anterior border 23.4-26. 8%TL from anterior ambitus. Phyllodes moderately developed with slot- shaped pores in circular depressions. Labrum short and wedge- shaped extending only to centre of second pair of adjacent ambulacral plates. Anterior edge raised above surrounding ambulacra and slightly projecting over peristome. Miliary tubercles present with few secondary tubercles in anterior half (Fig. 7B). Plastron long, width about 55-60% length, with ambulacral plates indenting posteriorly. Subanal fasciole circular to trans- versely oval, enclosing 3 pore pairs each side of interradial suture, and with slight anterior projection at posterior end of prominent plastronal keel. Posterior edge of fasciole margin- ally clear of periproct opening. Adorally, ambulacra I and V relatively wide, covered with very fine, randomly spaced miliary tubercles up to sixth plate, then by tubercles of similar size to rest of adoral surface. Periproct, tear-drop shaped, generally positioned vertically on truncated posterior margin but slightly visible from above on some specimens; height 17.1-19.1%TL, width 13.2-16.0%TL. Etymology. Named for Michael and Marie Dermody, owners of Glenforslan Station. Remarks. Spatagobrissus dermodyorum sp. nov. differs pri- marily from the Middle Miocene Port Campbell Limestone species, Spatagobrissus laubei (Duncan, 1877), in having narrower test with more posterior maximum width (Fig. 8A), markedly larger peristome and periproct (Fig. 8D, E), and very much shorter labrum (Fig. 8B). In addition, anterior paired petals are shorter and posterior petals longer (Fig. 8C), with divergent angle of latter greater than S. laubei. Aboral primary Two new middle Miocene spatangoids 95 tubercles larger within peripetalous fasciole and interambulacra 2 and 3, while fine tubercles outside fasciole on posterior half of test are much smaller in diameter. Adorally, plastron wider and longer, and ambulacra I and V narrower. The extant type species Spatagobrissus mirabilis is charac- terised by a larger (up to 110 mm long) and less inflated test, more posteriorly located apical system in line with maximum width, greater area enclosed by peripetalous fasciole, shorter peristome, and periproct situated on an obliquely truncated surface below the ambitus. Primary tubercles within the peripetalous fasciole of S. mirabilis are also larger and more closely spaced than in S. dermodyorum. Spatagobrissus incus Baker and Rowe, 1990, an extant species endemic to southeast Australian waters, particularly between Flinders Island (Tasmania) and western Spencer Gulf, South Australia, has a larger and more rounded test (up to 80 mm long) and, similar to S. mirabilis, more posteriorly located apical system and greater area enclosed by peripetalous fasciole. Compared with S. dermodyorum, it also has a much wider and longer plastron and narrower adoral ambulacra I and V. Miskelly (2002: 156) noted two pairs of pore pairs occur in each side of the subanal fasciole, a feature also recorded for S. laubei (McNamara et al., 1986: 80). This contrasts with the three pairs found on S. dermodyorum (Fig. 1C). Discussion The major diagnostic features of Murraypneustes, particularly the position of the two ‘peripetalous’ fascioles, the intermittent horizontal fasciole bands, and the random pattern and separ- ation of two distinct sizes of primary tubercles across the aboral surface, give little indication of its lineage. Certainly, within the Australian Cenozoic sequences that predate the early Middle Miocene Glenforslan Formation, all the 13 recorded genera of Micrasterina have their peripetalous fasciole close or in contact with the paired petals. Only the brissid Cyclaster is recorded as having developed multiple fascioles in some specimens (McNamara et al., 1986: 68). Similarly, there are no extant genera in Australian waters that have any specific combination of characters that link them to Murraypneustes. Even the extant species Linopneustes brachipetalus, found off the east coast of Australia, has long petals in contact with its marginal peripetalous fasciole. The lack of any juvenile specimens of the new genus, or indeed any marked variation in the size of the three known specimens, precludes any useful comment on the disposition and specific function of its somewhat unusual arrangement of fascioles, as such development takes place at a very early stage of ontogeny. Excluding Murraypneustes, only Hemiaster, of the nine spatangoid genera known to occur in the Glenforslan Formation sequence which embraces locality PL3203 (Table 1), has no extant record in Australian waters. The eight remain- ing genera are today almost exclusively benthic filter feeders, living inshore or on the continental shelf. Five, Brissopsis, Brissus, Cyclaster, Lovenia and Pericosmus, occur in tropical waters, and three, Eupatagus, Protenaster and Spatagobrissus in temperate waters (Rowe and Gates, 1995). Species of these eight extant genera, with the possible exception of Cyclaster, are known to occur at depths of less than 20 m in Australian or New Zealand waters. This depth range is consistent with the sedimentary deposition of the Glenforslan Formation. On the other hand, Indo-Pacific species of Linopneustes, including L. brachypetalus, are found at depths exceeding 270 m (range 272-1788 m), with only the West Indies species, L. longispinus, extending up into sublitoral waters (70-570 m) (Mortensen, 1950a). Based on the available evidence, it is reasonable to assume Murraypneustes dermodyi was a benthic filter feeder inhabiting relatively shallow, warm (?sub-tropical) waters. Acknowledgements We are indebted to David Holloway (Invertebrate Palaeontol- ogy, Museum Victoria) for valuable advice and support during the preparation of this manuscript, and to Kenneth McNamara (Western Australia Museum), Rich Mooi (California Academy of Sciences) and Andrew Smith (Natural History Museum, London), for suggesting improvements to the manuscript. Jeff Lukasik (Petro-Canada) is thanked for providing specific strati- graphic information and Ashley Miskelly (Blackheath, NSW) for details of extant echinoids from Australian waters. Michael and Marie Dermody and Donald and Miriam Griffen (Blanchetown District, South Australia) are also thanked for permission to collect on their properties. As always, Val Hogan and Sandra Winchester (Library, Museum Victoria) were helpful with access to references. References Agassiz, A. 1878. Report on the Echini in No. 9(2) — Reports on the results of dredging, under the supervision of Alexander Agassiz, in the Gulf of Mexico, by the United States Coast Survey Steamer “Blake”, Lieutenant-Commander C. D. Sigsbee, U. S. N., com- manding. Bulletin of the Museum of Comparative Zoology, Harvard University 5: 185-196, pis 1-5. Agassiz, A. 1881. Report on the Echinoidea. Reports on the scientific results of the voyage of H.M.S. Challenger during the years 1873-76, Zoology 3(9): 1-321, 45 pis. Agassiz, L. in: Agassiz, L., and Desor, E. 1847. Catalogue raisonne des families, des genres et des especes de la classe des echinoderms. Annales des Sciences Naturelles 7: 129-168. Baker, A.L., and Rowe, F.W.E. 1990. Atelostomatid sea urchins from Australia and New Zealand waters (Echinoidea: Cassiduloida, Holasteroida, Spatangoida, Neolampadoida). Invertebrate Taxonomy 4: 281-316. Clark, H.L. 1923. The echinoderm fauna of South Africa. Annals of the South African Museum 113: 221^435. Claus, C.F.W. 1876. Grundzuge der Zoologie. Third edition. Marburg and Leipzig, xii + 1254 pp. Cotteau, G. 1875. Description des echinides tertiaries des lies St Barthelemy et Anguilla. Kungliga Svenska Vetenskapsakademiens Handlingar 13(6): 1-48. Cotteau, G. 1888. Echinides eocenes de la province d’ Alicante (Espagne). Compte Rendu de VAcademie des Sciences Paris 107(25): 976-978. 96 Francis C. Holmes, Christopher Ah Yee and Janice Krause Duncan, P.M., 1877. On the Echinodermata of the Australian Cainozoic (Tertiary) deposits. Quarterly Journal of the Geological Society of London 33(1): 42-73, pis 3, 4. Fell, H.B. 1963. New genera of Tertiary echinoids from Victoria, Australia. Memoirs of the National Museum of Victoria 26: 211-217. Fisher, A.G. 1996. Spatangoids. Pp. U543-U628 in: Moore, R.C. (ed.), Treatise on Invertebrate Paleontology Part U, Echinodermata 3. Geological Society of America and University of Kansas Press. Gray, J.E. 1855. Catalogue of the Recent Echinida, or sea eggs, in the collection of the British Museum 1: Echinida Irregularia. Woodfall and Kinder: Eondon. 69 pp., 6 pis. Grube, A.E. 1865. Einige neue Seesterne des hiesigen zoologischen Museums. Jahresbericht der Schlesischen Gesellschaft fiir vaterlandische Cultur 1865: 35-37. Holmes, F.C. 1993. Australian fossil echinoids: annotated bibliography and list of species. Occasional Papers of the Museum of Victoria 6: 27-54. Koehler, R. 1914. Rectification. Zoologischer Anzeiger 44: 191. Lambert, J. 1905. Echinides eoceniques de l’Aude et de l’Herault. Pp. 129-164 in: Doncieux, L. (ed.), Catalogue descriptif des fossils nummuliques de T Aude et de l’Herault. Premiere partie: Montagne Noire et Minervous. Annales de TUniversite de Lyon 17: 1-184,5 pis. Lambert, J., and Sanchez Roig, M. 1924. P. 449 in: Lambert J., and Thiery, P. Essai de nomenclature raisonnee des echinides 6 and 7. Libraire L. Ferriere: Chaumont. Lambert, J., and Thiery, P. 1915. Descriptions des echinides des Terrains Neogenes du Basin du Rhone. Memoires de la Societe Paleontologique Suisse ( Abhandlungen der Schweizerischen Palaontologischen Gesellschaft ) 41: 155-240, pis 13-17. Lukasik, J.J., and James, N. P. 1998. Lithostratigraphic revision and correlation of the Oligo-Miocene Murray Supergroup, western Murray Basin, South Australia. Australian Journal of Earth Sciences 45: 889-902. McGowran, B., and Li, Q. 1994. Miocene oscillation in southern Australia. Records of the South Australian Museum 27: 197-212. McNamara, K.J., Philip, G.M., and Kruse, P.D. 1986. Tertiary brissid echinoids of southern Australia. Alcheringa 10: 55-84. Meijere, J.C.H. de 1903. Vorlaufige Beschreibung der neuen, durch die Siboga-Expedition gesamelten Echiniden. Tijdschrift der Nederlandsche Dierkundige Vereeniging (ser. 2) 8: 1-16. Miskelly, A. 2002. Sea urchins of Australia and the Indo-Pacific. Capricornia Publications: Sydney. 179 pp. Mortensen, T. 1948. Report on the Echinoidea collected by the United States Fisheries Steamer “Albatross” during the Philippine Expedition, 1907-1910. Part 3. The Echinoneidae, Echinolampadidae, Clypeasteridae, Arachnoididae, Laganidae, Fibulariidae, Urchinidae, Echinocorythidae, Palaeostomatidae, Micrasteridae, Palaeopneustidae, Hemiasteridae and Spatanidae. Bulletin of the United States National Museum 100(14): 93-140. Mortensen, T. 1950a. A monograph of the Echinoidea 5(1). Spatangoida 1. Proto sternata, Meriodosternata, Amphisternata 1. Palaeopneustidae, Palaeostomaridae, Aeropsidae, Toxasteridae, Micrasteridae, Hemiasteridae. C.A. Reitzel: Copenhagen. 432 pp., 25 pis. Mortensen, T. 1950b. New Echinoidea (Spatangoida). Videnskabelige Meddelelser fra Dansk Naturhistorisk Forening I Kj0benhavn 112: 157-163. Rowe, F.W.E., and Gates, J. 1995. Echinodermata. In Wells, A. (ed.). Zoological Catalogue of Australia , Vol. 33. CSIRO Australia: Melbourne, xiii + 510 pp. Smith, A.B., Stockley, B., and Godfrey, D. 2003. Spatangoida. In: Smith, A.B. (ed.), The echinoid directory. http:Avww.nhm.ac.uk/ palaeontology/echinoids [accessed 19 Jun 2005]. Figure 1. A, B, general location maps; C, map of the Murray River between Waikerie and Swan Reach, South Australia, showing locality of NMV locality PL3203, north of Blanchetown. Two new middle Miocene spatangoids 97 Figure 2. Murraypneustes biannulatus gen et sp. nov. A-D, adapical, posterior, adoral and left lateral views of holotype NMV P3 12370; E, adapical view of paratype NMV P3 12371; F, left lateral view of paratype, NMV P3 12372. All from the early Middle Miocene Glenforslan Formation in the vicinity of NMV locality PL3203, north of Blanchetown, South Australia. Figure 3. Fasciole details of Murraypneustes biannulatus gen. et sp. nov. A, B, oblique lateral view and ambulacrum IV detail above margin of paratype, NMV P3 12371; C, oblique posterior view of paratype, NMV P3 12372. Scale bar 10 mm. Francis C. Holmes, Christopher Ah Yee and Janice Krause Figure 4. Apical disk details of Murraypneustes biannulatus gen. et sp. nov. A, holotype NMV P3 12370; B, drawing of paratype, NMV P3 12372, showing extent of hydropores (black circles), tubercles (white circles), and gonopores and oculars (black or stippled). Location of posterior oculars (marked *) is assumed. Scale bars 1 mm. Figure 5. Murraypneustes biannulatus gen. et sp. nov. Detail of pore pairs and tubercles in and adjacent to ambulacrum V of paratype, NMV P3 12372. Figure 6. Spatagobrissus dermodyorum sp. nov. A, B, D, adapical, adoral, and left lateral views of holotype, NMV P3 12570; C, adoral view of paratype, NMV P3 12571; E, posterior view of paratype, NMV P3 12572 All specimens from the early Middle Miocene Glenforslan Formation in the vicinity of locality NMV PL3203, north of Blanchetown, South Australia. Scale bars 10 mm. test length {mm) Two new middle Miocene spatangoids 99 Figure 7. Spatagobrissus dermodyorum sp. nov. A, detail of petals and apical disk of holotype, NMV P3 12570; B, peristome and labrum tuberculation of paratype, NMV P3 12571; C, subanal fasciole and pore pairs of paratype, NMV P3 12573. Scale bars 1mm. Figure 8. Comparative biometric data on specimens of Spatagobrissus laubei (Duncan, 1877) from the Middle Miocene Port Campbell Limestone, Port Campbell, Victoria (•), and S. dermodyorum sp. nov. from the early Middle Miocene Glenforslan Formation, Blanchetown district, South Australia (o). Contents Memoirs of Museum Victoria Volume 62 Number 1 2005 1 > Homalonotid trilobites from the Silurian and Lower Devonian of south-eastern Australia and New Zealand (Arthropoda: Trilobita: Homalonotidae) Andrew C. Sandford 67 > Pliocene marine mammals from the Whalers Bluff Formation of Portland, Victoria, Australia Erich M.G. Fitzgerald 91 > Two new Middle Miocene spatangoids (Echinoidea) from the Murray Basin, South Australia Francis C. Holmes, Christopher Ah Yee and Janice Krause Volume 62 Number 2 2005 103 > The millipede genus Lissodesmus Chamberlin, 1920 (Diplopoda: Polydesmida: Dalodesmidae) from Tasmania and Victoria, with descriptions of a new genus and 24 new species Robert Mesibov 147 > A review of pygal-furrowed Synallactidae (Echinodermata: Holothuroidea), with new species from the Antarctic, Alantic and Pacific oceans P. Mark O’Loughlin and Cynthia Ahearn 181 > A new asterinid genus from the Indo-West Pacific region, including five new species (Echinodermata: Asteroidea: Asterinidae) P. Mark O’Loughlin and Francis W.E. Rowe 191 > A review of soles of the genus Aseraggodes from the South Pacific, with descriptions of seven new species and a redescription of Synclidopus macleayanus John E. Randall 213 > The species of Dasycercus Peters, 1875 (Marsupialla: Dasyuridae) P.A. Woolley