iT 1 JUNE 1982 | | ISSN 0303-2515 La fi PP H ; ANNALS OF THE SOUTH AFRICAN CAPE TOWN INSTRUCTIONS TO AUTHORS 1. MATERIAL should be original and not published elsewhere, in whole or in part. 2. LAYOUT should be as follows: (a) Centred masthead to consist of Title: informative but concise, without abbreviations and not including the names of new genera or species Author’s(s’) name(s) Address(es) of author(s) (institution where work was carried out) Number of illustrations (figures, enumerated maps and tables, in this order) (b) Abstract of not more than 200 words, intelligible to the reader without reference to the text {c) Table of contents giving hierarchy of headings and subheadings (d) Introduction 4 (e) Subject-matter of the paper, divided into sections to correspond with those given in table of contents (f) Summary, if paper is lengthy (g) Acknowledgements (h) References (i) Abbreviations, where these are numerous 3. MANUSCRIPT, to be submitted in triplicate, should be typewritten and neat, double spaced with 2,5 cm margins all round. First lines of paragraphs should be indented. Tables and a list of legends for illustrations should be typed separately, their positions indicated in the text. All pages should be numbered consecutively. Major headings of the paper are centred capitals; first subheadings are shouldered small capitals; second subheadings are shouldered italics; third subheadings are indented, shouldered italics. Further subdivisions should be avoided, as also enumeration (never roman numerals) of headings and abbreviations. Footnotes should be avoided unless they are short and essential. Only generic and specific names should be underlined to indicate italics; all other marking up should be left to editor and publisher. 4. ILLUSTRATIONS should be reducible to a size not exceeding 12 « 18 cm (19 cm including legend); the reduction or enlargement required should be indicated; originals larger than 35 x 47 cm should not be submitted; photographs should be rectangular in shape and final size. A metric scale should appear with all illustrations, otherwise magnification or reduction should be given in the legend; if the latter, then the final reduction or enlargement should be taken into consideration. All illustrations, whether line drawings or photographs, should be termed figures (plates are-not printed; half-tones will appear in their proper place in the text) and numbered in a single series. Items of composite figures should be designated by capital letters; lettering of figures is not set in type and should be in lower-case letters. The number of the figure should be lightly marked in pencil on the back of each illustration. 5. REFERENCES cited in text and synonymies should all be included in the list at the end of the paper, using the Harvard System (ibid., idem, loc. cit., op. cit. are not acceptable): (a) Author’s name and year of publication given in text, e.g.: ‘Smith (1969) describes .. .” ‘Smith (1969: 36, fig. 16) describes .. .’ “As described (Smith 1969a, 19696; Jones 1971)’ ‘As described (Haughton & Broom 1927)...’ ‘As described (Haughton et a/. 1927)...’ Note: no comma separating name and year Dagination indicated by colon, not p. names of joint authors connected by ampersand et al. in text for more than two joint authors, but names of all authors given in list of references. (b) Full references at the end of the paper, arranged alphabetically by names, chronologically within each name, with suffixes a, 5, etc. to the year for more than one paper by the same author in that year, e.g. Smith (1969a, 19694) and not Smith (1969, 1969a). For books give title in italics, edition, volume number, place of publication, publisher. For journal article give title of article, title of journal in italics (abbreviated according to the World list o, scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number (only if independently paged) in parentheses, pagination (first and last pages of article). Examples (note capitalization and punctuation) BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan. FIsCHER, P.—H. 1948. Données sur la résistance et de le vitalité des mollusques. J. Conch., Paris 88: 100-140. FIsCHER, P.-H., DuvaL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archs Zool. exp. gén. 74: 627-634. Konn, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320. Konn, A. J. 19606. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51. THEELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 89 Band June 1982 Junie Part 1 Deel eel CS AP Ip = =) n Uy, ~S Qy S oun Now NY A REVISION OF THE DEVONIAN (EMSIAN-EIFELIAN) TRILOBITA FROM THE BOKKEVELD GROUP OF SOUTH AFRICA By MICHAEL R. COOPER Cape Town Kaapstad The ANNALS OF THE SOUTH AFRICAN MUSEUM are issued in parts at irregular intervals as material becomes available Obtainable from the South African Museum, P.O. Box 61, Cape Town 8000 Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM word uitgegee in dele op ongereelde tye na gelang van die beskikbaarheid van stof Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad 8000 OUT OF PRINT/UIT DRUK 12123558) NS (12h 425, Shep) o(l=3e Se =-9)s 6(1, t.p.i.), 701-4); 8, 94-2, 7), 10(1-3), 11(1-2, 5, 7, t-p.i.), 15(4-5), 24(2), 27, 31(1-3), 32(5), 33, 36(2), 45(1) EDITOR/REDAKTRISE Ione Rudner Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 86813 035 4 Printed in South Africa by In ‘Suid-Afrika gedruk deur The Rustica Press, Pty., Ltd., Die Rustica-pers, Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap A REVISION OF THE DEVONIAN (EMSIAN-EIFELIAN) TRILOBITA FROM THE BOKKEVELD GROUP OF SOUTH AFRICA By MICHAEL R. COOPER National Museum of Zimbabwe, Bulawayo (With 120 figures) [MS accepted 20 January 1982) ABSTRACT The Bokkeveld trilobite fauna is revised and shown to comprise the following species: Ormistoniella malaca (Lake), Malimanaspis? dereimsi (Kozlowski), Burmeisteria herscheli (Murchison), B. notica (Clarke), B. fontinalis (Reed), Francovichia clarkei (Ulrich), Gamon- edaspis boehmi (Knod), Oosthuizenella ocellus (Lake), Pennaia pupillus (Lake), Metacryphaeus caffer (Salter), M. venustus Wolfart, M. cf. koukharskii (Baldis & Longobucco), Typhloniscus baini Salter, Phacopina hexensis (Reed), P. lakei (Reed), P. cf. padilla Eldredge & Branisa, Phacopina? sp. nov?, Renniella rossouwi gen. et sp. nov., Bainella africana (Salter), B. arbuteus (Lake), B. cristagalli (Woodward), B. gamkaensis Rennie, Bainella sp. nov., Deltacephalaspis pseudoconvexus (Reed), and Kozlowskiaspis ceres (Schwarz). The affinities of the fauna are clearly with Bolivia. Three new genera are proposed, the dechenellid Ormis- toniella, and the calmoniids Oosthuizenella and Renniella, as well as the new calmoniid subgenus Metacryphaeus (Wolfartaspis). The only new species is Renniella rossouwi. At this stratigraphic level Digonus and Burmeisteria are inseparable, and the former taxon is herein considered superfluous; it is treated as a junior subjective synonym of Burmeisteria s.s. Higher calmoniid taxonomy is unsatisfactory and in need of phylogenetic analysis. The trilobite assemblages from the Gydo, Voorstehoek and Waboomberg Formations are distinct, and form the basis for the recognition of the Bainella africana, Metacryphaeus caffer, and M. venustus Assemblage Zones respectively. Commencement of Bokkeveld sedimentation is related to late Emsian eustatic transgression, an age determination supported by the trilobite faunas, with marine deposition extending well into the Eifelian. CONTENTS PAGE PH EL OCUCHION er oA areata es ie eatin ernie ke te eens 2 SVSLOMIALICS terse ne etre ane iteens terete nine amet ded) tan eubuenie eh 5) BamilyDechencellidactere, armen tens re acme eee 5) Genus Crmistoniellacen, NOV «5. ne ae 5 Ramilva@tanomdaeerr. =e meer ee 14 GenlsHVMahianaspis® eerie ts ee ein 14 Bamilyaomalonotidactartwre acarics cr: pacer rs ee 17 GENUSES UIICISICLIG einen ee iain yt oe cen mee 17 HamilyeDalmaniti dace rites rete ee ea ee 45 Genustirancovichian ae re re 45 BamilyaSyvophoni dace steer an we ee 51 GenusiGamoedasp ism een yer ra 55 anulyi@almonnddes sce. ars Merny ite en foe eee 61 GenustOosthuizencila cen Ove sec. ee 63 1 Ann. S. Afr. Mus. 89 (1), 1982: 1-174, 120 figs. 2 ANNALS OF THE SOUTH AFRICAN MUSEUM PAGE GenusiPenndia s. a ocudc poeta Rie ees SEE 72 (CENWS MAME OOS coascsccveccosve0cac0nc0cccc 84 GENUS INHUDANELS coccocboocooc0psssccusEKdoKC 109 (GENUS JAVED DAG 000000000009 cd000000n0000000000 113 Genus (Rag wAlG BAMs WOT. bocccecccca0cc0K0000 006 124 GenuspBainclla ee. ay tise e soe eee eR oe N27 GenusvKOzlowsiiaspise seen eee a eee ace 155 Genus! Deltacephalaspiswerem nn eee ee 158 Biostrationaphyerye eerche sch soci ae Make Oo oe eee eet 161 Eaunaltattinities#) 2. Agi. oer. ce es ean a 163 Age ofthe BokKeveldifaunas) qers.o2 eee eee 167 SUMIM ALY eras ore errs acre eae eine aie eae 170 ACknOWledgements ane. cet eee pads case Ieee 170 IRNEfEEN CES Heresy hereon eae a eR oe SOR ee 171 INTRODUCTION On 25 May 1836 Sir Roderick Murchison announced to the Geological Society of London the discovery by Sir John Herschel of trilobites *. . . in rocks which occur to the north of the Cape of Good Hope’, associated with other fossils which, together, ‘. .. were found in abundance in the Cedar Bergen, according tc Captain Alexander 150 English miles north of Cape Town and 2 000 feet above the sea, in an ash-grey quartzose sandstone’ (Sandberger 1853: 581). The trilobites were subsequently identified by Murchison (1839) as Homalonotus herscheli sp. nov. and Calymene blumenbachi Brongniart, the latter a misidentification of what is now Oosthuizenella ocellus (Lake) (see p. 65). It was Salter (1856), however, who was the first to describe a major collection of invertebrate fossils from the Bokkeveld succession, including the trilobites Homalonotus herscheli Murchison, Phacops (Cryphaeus) africanus sp. nov., Phacops caffer sp. nov., and Typhloniscus baini gen. et sp. nov. Thus, this work is important in providing the first descriptions of the wholly Malvinokaffric trilobite genera Bainella, Metacryphaeus, and Typhloniscus. In his subsequent monograph of the British trilobites, Salter (1865) created the new subgenus Burmeisteria within Homalonotus, with H. herscheli as the type species. Woodward (1873) described a highly ornate trilobite with long axial spines from the Cockscomb Mountains in the eastern Cape as Encrinurus cristagalli sp. nov., whilst Schenk (1888) mentioned a ‘Praetus ricardi’ from the same locality, but without figure or description and hence the name is a nomen nudum. In 1897 Frech figured a specimen from the ‘Cape Colony’ as Homalo- notus perarmatus sp. nov. In the first work devoted entirely to a study of the Bokkeveld Trilobita, Philip Lake (1904) recorded the following species: Phacops pupillus sp. nov., P. arbuteus sp. nov., P. ocellus sp. nov., P. africanus Salter, P. cristagalli (Woodward), P. (Cryphaeus) caffer Salter, Typhloniscus baini Salter, Dalman- ites lunatus sp. nov., Proetus malacus sp. nov., Homalonotus herscheli Mur- chison, H. quernus sp. nov., and H. colossus sp. nov., and considered DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 3 H. perarmatus Frech a synonym of H. herscheli. Soon afterwards Schwarz (1906) further added to the list of Bokkeveld trolibites when describing the material housed in the Albany Museum, Grahamstown, including Phacops cristagalli (Woodward), P. acacia sp. nov., P. (Cryphaeus) callitris sp. nov., P. (C.) ceres sp. nov., P. (C.) gydowi sp. nov., Proetus ricardi Schenk, Homalono- tus herscheli Murchison, H. perarmatus Frech, H. lex sp. nov., H. agrestis sp. nov., H. horridus sp. nov. and H. hippocampus sp. nov. In a discussion of the Bokkeveld fauna, Reed (1907) pointed out that the subgenus Cryphaeus could not be applied to trilobites for reasons of priority, and noted that the Bokkeveld phacopids fell into two groups. For Phacops caffer, P. ocellus, P. impressus, P. callitris, P. gydowi, and P. ceres he proposed the name Metacryphaeus, while considering P. africanus, P. acacia, P. arbuteus, and P. cristagalli to ‘... fall into another group which in many respects resembles that containing Dalmanites anchiops Green’ (p. 168). For the latter group he proposed the name Anchiopella. In describing Pennaia africana sp. nov. from Osplaats in the Hex River valley, Shand (1914) was the first to report the presence of this genus in the Bokkeveld fauna although, as noted by Rennie (1930), the species was generi- cally misidentified. In a discussion of the genus Homalonotus, Reed (1918) referred all the South African species to the subgenus Burmeisteria. The most comprehensive discussion of the Bokkeveld Trilobita till now is Reed’s (1925a) monographic ‘Revision of the fauna of the Bokkeveld Beds’ in which he described or discussed the following taxa: Proetus malacus Lake, P. hexensis sp. nov., Cyphaspis dereimsi Kozlowski, Dalmanites (Anchiopella) africanus (Salter), D. (A) cristagalli (Woodward), D. (A.) arbuteus (Lake), D. (A.) baini sp. nov., D. (Corycephalus?) capensis sp. nov., D. (Hausmannia) lunatus Lake, D. (H.) dunni sp. nov., D. (Proboloides) ensifer sp. nov., D. (Acastella?) pseudoconvexus sp. nov., D. (Cryphaeus) caffer (Salter), D. (C.) caffer var. albana var. nov., D. (C.) cf. pentlandi Salter, D. (C?) cf. rostratus Kozlowski, D. (C?) ceres (Schwarz), Phacops (Calmonia) callitris Schwarz, P. (C.) impressus Lake, P. (C?) lakei sp. nov., P. (C?) pupillus Lake, Phacops (Pennaia) gydowi Schwarz, P. (P.) africanoides nom. nov. (pro Pennaia afri- cana Shand), P. (Phacopina) hiemalis sp nov., Homalonotus (Burmeisteria) herscheli Murchison with the varieties colossus Lake, rectisuturalis var. nov., fusiformis var. nov., grahami var. nov., bituberculata var. nov., perarmata Frech, and sodalis var. nov., H. (B.) quernus Lake, H. (B?) hippocampus Schwarz, H. (Digonus) fontinalis sp. nov. and H. (D.) noticus var. africana var. nov. Reed (19255) subsequently proposed the new genus Eocorycephalus to include Dalmanites (Corycephalus?) capensis Reed from the Bokkeveld sequence. In reviewing work on the Phacopidae, Reed (1927) contended that Pha- cops (Cryphaeus) caffer Salter, which he had earlier listed as a species of Metacryphaeus, was, in fact, a typical Asteropyge and, since the remaining species cited in his original proposal of the genus Metacryphaeus could all 4 ANNALS OF THE SOUTH AFRICAN MUSEUM satisfactorily be included in Calmonia, the name should be abandoned. Moreover, he stated that ‘. . . the type which was chosen for this special group (Anchiopella) and exemplified by Dalmanites anchiops Green was Phacops crista-galli (Woodw.)’ (p. 210), while now referring Phacops gydowi and Acaste verneuili d’Orbigny to the genus Pennaia. Rennie’s (1930) revision of the Bokkeveld ‘phacopids’ was important for its attempt to stabilize nomenclature. Thus he showed that Metacryphaeus was a valid taxon whose type species was Phacops caffer, that Eocorycephalus was a junior subjective synonym of Synphoria, and that Eocorycephalus capensis (Reed) was founded upon damaged cephala of Metacryphaeus caffer. Rennie also transferred all Reed’s (1925a) Bokkeveld species of Dalmanites (Anchiopella) to his new genus Bainella, because he erroneously believed (see Cooper 1982) Dalmanites anchiops Green to be the type species of Anchiopella. In this work Rennie (1930) described and/or figured Dalmanites (Metacry- phaeus) caffer (Salter), D. (M?) ceres (Schwarz), Acaste (Pennaia) impressus (Lake), A. (Calmonia) callitris (Schwarz), A. (C.) africana (Shand), Bainella bokkeveldensis sp. nov. B. gamkaensis sp. nov., B. acacia (Schwarz), B? africanus (Salter), and Typhloniscus baini Salter. In 1951 Brink figured and described a large trilobite eye from the Bokke- veld Beds that was identified only as Phacops sp. Pillet (1954) discussed the classification of the phacopid trilobites and created the new genus Paradalman- ites for Phacops acacia Schwarz. Paradalmanites was correctly included in the synonymy of Bainella by Struve (in Harrington et al. 1959) who also rightfully elevated Metacryphaeus, Calmonia, and Pennaia to generic status. Although not dealing directly with the Trilobita from the Bokkeveld succession, the works of Wolfart (1968), Branisa & Vanék (1973), Eldredge & Ormiston (1979), and Eldredge & Brani§a (1980) on the Bolivian faunas have a direct bearing on the classification of the Bokkeveld material. Wolfart’s (1968) important study showed that Metacryphaeus caffer in South America was a catch-all for a number of well-defined species of Metacryphaeus, thus pointing to the impressive speciation and adaptive radiation that occurred in the strongly transgressive deposits of the Malvinokaffric Realm. Branisa & Vanék (1973) created a number of new calmoniid and dalmanitid genera, some of which are now known also to occur in South Africa. Eldredge & Ormiston (1979) reviewed the Malvinokaffric faunas and provided a tentative modern supraspe- cific revision of the Bokkeveld fauna. Salient features of this review include reference of Dalmanites dunni to the genus Francovichia, Proetus malacus to Dechenella?, Phacopina hiemalis to Bainella?, Proboloides ensifer to Schizo- stylus (Curuyella), the eye figured by Brink (1951) to the genus Fenestraspis, and Cyphaspis dereimsi to Otarion (Maurotarion). The study of Eldredge & BraniSa (1980) is of the utmost importance to an understanding of calmoniid taxonomy, again highlighting the complex speciation that occurred in this closed evolutionary pool. These workers provided a revised diagnosis of Bainella and created the new subgenus B. (Belenops) to include Bainella DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 5 gamkaensis Rennie. They also figured a cephalon of Bainella bokkeveldensis Rennie, and provided a schematic representation of the auxiliary impression system of a Pennaia from Gamkapoort. As it is now 50 years since the last paper was published dealing specifically with Bokkeveld trilobites, the group is obviously in need of modern revision. This is emphasized by the study of Eldredge & BraniSa (1980), which has shown the Malvinokaffric trilobites to be largely endemic. Thus, such European taxa as Acaste, Phacops, Dalmanites, Homalonotus, and Otarion are not present in the Bokkeveld fauna, but are represented by closely allied or homoeomorphic taxa. The aim of this paper is, therefore, to provide a modern revision of the Bokkeveld fauna on the basis of type and additional material. In this way it is hoped to recognize any zonal succession that may exist and, if possible, to attempt to date the Bokkeveld succession on the basis of the trilobite faunas. The following abbreviations are used to indicate the repositories of material discussed in the text: AM _ Albany Museum, Grahamstown BPI Bernard Price Institute for Palaeontological Research, Johannes- burg BM _ British Museum (Natural History), London KM Alexander Macgregor Museum, Kimberley PEM Port Elizabeth Museum, Port Elizabeth PRV _ Geological Survey, Pretoria RO Roy Oosthuizen Collection, Zwartskraal, Prince Albert SAM_ South African Museum, Cape Town SMC Sedgewick Museum, Cambridge SU Geological Collections, Stellenbosch University SYSTEMATICS Order PROETIDA Fortey & Owens, 1975 Superfamily PROETACEA Salter, 1864 Family Dechenellidae Pribyl, 1946 Genus Ormistoniella gen. nov. Type species Proetus malacus Lake, 1904; by original designation herein Etymology For Allen R. Ormiston of Amoco Production Company, Tulsa, who assisted the writer with literature pertaining to this species and who has contributed greatly to an understanding of the dechenellid trilobites. 6 ANNALS OF THE SOUTH AFRICAN MUSEUM Diagnosis Cephalon semicircular with very prominent, long, genal spines. Glabella pear-shaped, tapering anteriorly. Glabellar furrows weak to obsolete; 1p deepest, with weak adaxial branches tending to isolate L1. Preglabellar field lacking. Eyes large, crescentic, close to axial furrows. Facial sutures strongly divergent anteriorly. Thorax with 9 segments. Pygidium macropygous, oval, elongate, with narrow but distinct border. Axis with 14-16 rings and pleural fields with 6-7 pairs of ribs. There is a short post-axial field. Discussion Ormistoniella gen. nov. differs from Dechenella (Dechenella) in having less deeply incised glabellar furrows, lacking a marked angulation of the anterior facial sutures at the border furrow, possessing only 9 thoracic segments, and with a pygidium with only 6—7 ribs in the pleural field and a very narrow post-axial border. It is perhaps closest to D. (Basidechenella) which it resem- bles in the shallowness of the glabellar furrows and very weak preoccipital lobes, but may be distinguished by fewer thoracic segments and pleural ribs to the pygidium and with a much narrower pygidial border. Ormistoniella may be distinguished from D. (Monodechenella) by its pear-shaped glabella and fewer pleural ribs to the pygidium. It is clear from the above that Ormistoniella gen. nov. is closely allied to Dechenella and that the differences may warrant only subgeneric separation, a procedure originally favoured by the writer. However, in view of the constancy of the count of 10 thoracic segments in all other known dechenellids, the presence of only 9 in Ormistoniella malacus seems to be meaningful taxonomi- cally. As such the differences are herein considered sufficient for generic separation. Ormistoniella malaca (Lake, 1904) Figs 1-4, 5C, 6B, 7C Praetus ricardi Schenk, 1888: 227 (nomen nudum). Proetus malacus Lake, 1904: 213, pl. 25 (fig. 10). Reed, 1925a: 119, pl. 11 (fig. 3). Proetus ricardi Schenk, Schwarz, 1906: 390, pl. 10 (fig. 5). Dechenella? malaca (Lake) Eldredge & Ormiston, 1979: 164. Material The holotype, by monotypy, SAM—K559, together with SAM-—PDB 4203, AM-35, SU-K460, RO-8, 412, 748, 759-761, 830, 832, P55—-56, C27, E4, 13, 63, 68, PRV-—26, 71, and an unnumbered KM specimen. Description Cephalon. The cephalon is semicircular, much wider than the sagittal length, and with long genal spines extending backward as far as the 5th or 6th thoracic segment. The lateral and anterior borders are broad, well developed, DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 7 Fig. 1 Ormistoniella malaca (Lake). A reconstruction based upon the available material. x 2. and striated along their lengths. The rather shallow lateral border furrows meet the moderately deep posterior border furrows in very acute angles at the bases of the genal spines and are produced backward almost to their distal tips. The occipital furrow is narrow and deep, and the occipital ring is of uniform width. The glabella is urceolate, waisted opposite the 3p furrows, very broad and short with a maximum transverse width slightly less than the sagittal length, and tapers rather strongly to the anterior. The axial furrows are shallow, especially opposite the eye lobes. Glabellar segmentation is poorly developed; 3p is faint, very short, and almost normal to the glabellar margin; 2p is slightly longer and weakly curved but equally indistinct; 1p is the deepest, but still weakly developed, and curves back almost to the occipital furrow, so as to separate a subtrigonal, incipient preoccipital glabellar lobe (L1). The lp furrows show a faint but distinct adaxial bifurcation. The glabella reaches to the anterior furrow and there is no preglabellar field. The eyes are fairly large, holochroal, and comprise hundreds of extremely small ocelli arranged in diagonal rows. 8 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 2. Ormistoniella malaca (Lake). The holotype, by monotypy, SAM-PDB559, preserved as an internal mould. Note the broad pygidial axis. x 1,7. Thorax. The thorax comprises 9 segments only (Fig. 7C) with a well- defined, strongly convex axis. The pleural fields are strongly down-turned distally, with well-defined fulcra. There are deep, prominent pleural grooves to the pleurae, which are separated by finer interpleural furrows. Pygidium. Large, macropygous, oval in outline, and somewhat wider than the sagittal length. As with the British dechenellids (cf. Selwood 1965) there are two morphotypes amongst the pygidia, the differences perhaps reflecting sexual dimorphism. In the Type I pygidia (Fig. 3D) the moderately convex axis is rather broad, tapering fairly rapidly to the posterior, and comprises 14-16 distinct axial rings. In the Type II pygidia (Fig. 7C), the axis is relatively much narrower, more convex and tapers less strongly to the posterior. In both types, the pleural fields are moderately convex, subtrigonal, down-turned distally and with 6-7 prominent ribs separated by equally broad pleural furrows. The pleural ribs become obsolete distally so as to leave a narrow, smooth border with an entire margin. There is a narrow post-axial field which is terminally acuminate. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 9 Fig. 3. Ormistoniella malaca (Lake). A. RO-760. B-—C. RO-748, an enrolled specimen. D. RO~-761. Note the broad pygidial axis. E-F. An enrolled specimen, unnumbered, in the Alexander MacGregor Museum. G. RO-P55. H. RO-P56. I. RO-90, an unusually large individual. All preserved as internal moulds and all x 1. 10 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 4. Ormistoniella malaca (Lake). Lateral view of RO-8, preserved as an internal mould. x 3. Discussion No other dechenellids are currently recorded from the Malvinokaffric Realm and hence O. malaca is easily distinguished. *“Proetus’ hexensis Reed (1925a: 121, pl. 7 (fig. 8)) is based upon pygidia which differ from O. malaca in being semicircular in outline, with fewer axial rings and pleural ribs, indistinct segmentation and a distinct post-axial field. ‘Proetus’ problematicus Swartz (1925: 56, pl. 1 (figs 3-5)) is based upon fragmentary material from Bolivia that shows this species to be micropygous and thus probably an otarionid. Dechenella (Basidechenella?) boteroi Caster & Richter (in Richter & Richter 1950: 161; Eldredge & Ormiston 1979: 164) from the Devonian Floresta Beds of Colombia has median tubercles to the axial rings of the pygidium, 9 pairs of pleural ribs that are also tuberculate, and a broad, flattened border and hence is not liable to be confused with the present species. ‘Proetus’ chengi Sun (1937: 350, pl. 1 (fig. 2)) from the Upper Devonian Hsikuangshan Formation of Hunan, China, resembles O. malaca in the outline and narrow border of the pygidium, and with ‘13 or more’ axial rings, but its axis tapers more rapidly to the posterior and there are said to be 12 pairs of ribs in the pleural fields. 11 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP ‘T x ‘souids Jeuds SUC] OY] NION ‘pynouw [euIdjUI Ue ‘//D-OY ‘(pleMpooM) 1JDSVISIAD DYJaUuIDg “Q ‘[ X ‘YNou pozIsApoo ue Jo pynoul jeUIOJUI Ue ‘QOPY-NS “(ey¥e]) vovjou je saseq ouids juduiwoid aj0N “‘pynow jeurojUI ue se paArasaid JUdWISIS DIOeIOY}] Paie[OsI Ue ‘Gg/-OY DYAUOISIUMO “DQ “ST X ‘(ayeq) snaing 1D Dau *¢ “SOxe pex9]n] -V ‘S$ 34 12 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 6. A. Deltacephalaspis pseudoconvexus (Reed). The lectotype, designated herein, SAM-7201, preserved as an internal mould. x 1,9. B. Ormistoniella malaca (Lake). SAM- PDB4203, preserved as an internal mould. x 1,2. C. Pennaia pupillus (Lake). SAM—9605, preserved as an internal mould. x2. D. Metacryphaeus caffer (Salter). SAM-—PDB4347, an internal mould. x 1,1. Boliviproetus branisai Eldredge & Ormiston (1979: 163, pl. 1 (figs 1-7)) has a strongly inflated cephalon with a narrow preglabellar field. The strongly inflated glabella is oval in outline, not urceolate as in O. malaca. The pygidium of B. branisai is readily distinguishable from that of O. malaca in its weak axial segmentation and effaced axial furrows, with all but the first interpleural furrows obsolete. Dechenella clara (Hall) (Hall & Clarke 1888: 104, pl. 20 (figs 12-14), pl. 22 (figs 28-30)) from the Onondaga Limestone bears a general resemblance to O. malaca, but possesses pronounced occipital lobes and a pygidium twice as wide as long, with 7-11 axial rings and 4-5 pleural ribs. Dechenella mackayi Allan (1935: 27, pl. 1 (fig. 2)) from the Reefton Beds of New Zealand resembles O. malaca in its multisegmented pygidial axis with 12-15 axial rings and pleural fields with 7-8 pairs of ribs. It differs, however, in having an elliptical outline to the pygidium, pleural ribs that are much wider DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 13 Fig. 7. A. Gamonedaspis boehmi (Knod). PRV-126a, an internal mould. x 2,6. B. Typhlo- niscus baini (Salter). PRV-625, preserved as an internal mould. x 2,4. C. Ormistoniella malaca (Lake). PRV-71, an internal mould. Note nine segments to thorax and narrow pygidial axis. X 3. 14 ANNALS OF THE SOUTH AFRICAN MUSEUM than the intervening furrows and with distinct pleural grooves, and in possess- ing granulose ornament. Occurrence Ormistoniella malaca (Lake) is common in the Gydo Formation, while a single specimen, PRV-26, in green shale preservation may be from the Voorstehoek Formation. Family Otarionidae Richter & Richter, 1926 Subfamily Otarioninae Richter & Richter, 1926 Genus Malimanaspis Baldis & Longobucco, 1977 Type species Malimanaspis sarudianskii Baldis & Longobucco; by original designation Malimanaspis ? dereimsi (Kozlowski, 1923) Fig. 8 Cyphaspis dereimsi Kozlowski, 1923: 61, pl. 4 (figs 15-17). Swartz, 1925: 34. Reed, 1925a: 12, pl. 10 (fig. 10). Otarion dereimsi (Kozlowski) Branisa, 1965: 96, pl. 16 (figs 29-30). Otarion (Maurotarion) dereimsi (Kozlowski) Eldredge & Ormiston, 1979: 165. Malimanaspis dereimsi (Kozlowski) Baldis & Longobucco, 1977a: 154. Material SMC-A3062a, obtained from a road cutting between De Doorns and Tunnel (Reed 1925a), is still the only example known from the Bokkeveld Group. Description Cephalon. The cephalon is semicircular, with the anterior border slightly pointed sagittally, about twice as wide as the sagittal length, and with pro- minent genal spines extending backward to about the fourth thoracic segment. There is a well-developed preglabellar field, with a broad, fairly deep anterior furrow and a convex anterior border. The lateral borders to the cephalon are also very broad. The glabella is strongly inflated, trapezoidal in outline, almost rectangular, with maximum width at the occipital furrow. The deep, well- developed Ip glabellar furrows curve strongly backward to separate pro- nounced pre-occipital lobes (L1). L1 is half the exsagittal length of the glabella, whilst the 2p and 3p furrows are obsolete. The glabella is surrounded by deep, narrow axial and preglabellar furrows. The occipital furrow is rather narrow, anteriorly convex, and separates off an occipital ring of constant width. Although the eyes are missing from the present specimen, their broken bases suggest they were moderately large. The anterior branches of the facial suture are strongly divergent and the posterior branches extremely so, the latter meeting the posterior margin somewhere in the genal spines. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP IS) Fig. 8. Malimanaspis? dereimsi (Kozlowski). SMC- A3062a, preserved as an internal mould. x 1,6. Thorax. The thorax comprises at least 12 segments and has a broad axis, about one-third of the thoracic width. The pleurae bear deep pleural grooves and display prominent fulcra. Pygidium. Unknown from the present material. Discussion The generic assignment of the Bokkeveld specimen poses some problems. It was regarded as conspecific with the South American Cyphaspis dereimsi Kozlowski both by Reed (1925a) and the writer. However, Baldis & Longo- bucco (1977a) have recently assigned Kozlowski’s species to their new genus Malimanaspis, distinguished by a pustulose ornament over the entire exoskele- ton, which is reflected on internal moulds, the presence of a median tubercle to the occipital ring, and median tubercles (spines?) to the thoracic axial rings, and a pygidium with less than 5 axial rings. Kozlowski (1923) makes no mention of pustulose ornament in his original description of Cyphaspis der- eimsi, nor of the presence of a median tubercle to the occipital ring or sagittal tubercles to the thoracic axis. Similarly, none of these features is evident in the 16 ANNALS OF THE SOUTH AFRICAN MUSEUM Bokkeveld specimen, although the writer does not have the original for close inspection. The cranidium figured by Wolfart (1968, pl. 1 (fig. 1)) and assigned to Otarion dereimsi does, however, show pustulose ornament but, since there are now known to be at least 2 species of otarionid in the Malvinokaffric Realm, is perhaps better referred to Malimanaspis sarudianskii Baldis & Longobucco. Clearly the problems surrounding Otarion dereimsi (Kozlowski) can be resolved only by a restudy of the type material. The Bokkeveld example is, therefore, only questionably assigned to the genus Malimanaspis, since it also resembles Otarion (Maurotarion). Amongst Bokkeveld species, the only possibility of confusion of Malima- naspis ? dereimsi lies with Ormistoniella malaca (Lake). The latter species has, however, an urceolate glabella that lacks the deep 1p furrows and well- developed pre-occipital lobes of Malimanaspis? dereimsi, as well as being macropygous. Otarion (Maurotarion) maurum (Alberti) (1967: 500, pl. 1 (fig. 28); 1969: 384, pl. 36 (figs 1-4), pl. 46 (fig. 19)) is a Silurian (Lower Ludlow) species from Morocco which differs from M? dereimsi in having relatively smaller pre-occipital lobes, convex axial furrows to the glabella, a more strongly convex preglabellar furrow, and a sagittal node to the occipital ring. Otarion (Mauro- tarion) neptis (Alberti) (1969: 386, pl. 36 (figs 9-10)) is easily separated from the present species by its elliptical glabella with relatively much smaller pre-occipital lobes and much broader occipital ring with a sagittal node. These are also much the same features that distinguish O. (M.) ambiguum (Alberti) (1969: 388, pl. 36 (fig. 11)), which also has a narrower anterior border not weakly pointed sagittally as in M? dereimsi. Otarion (Maurotarion) elegantulum (Loven) (Alberti 1969: 384, pl. 36 (figs 5-6)) is a Silurian species with a pronounced anterior border, which is distinctly pointed on the sagittal line, and with a strongly convex preglabellar furrow, relatively smaller pre-occipital lobes, and a much broader occipital ring than the present species. Other species referred to O. (Maurota- rion) by Alberti (1969) include O. (M.) novellum (Barrande), O. (M.) instita Whittington & Campbell, O. (M.) plautum Whittington & Campbell, O. (M?) cristyi (Hall), and O. (M?) girardeauensis (Shumard), but the writer does not have access to the literature pertaining to these species. Kozlowski (1923) and Eldredge & Ormiston (1979) have pointed to the similarity of this species with Otarion miniscula Hall (Hall & Clarke 1888: 140, pl. 24 (figs 7-12)) from the Bois Blanc, Schoharie, and Onondaga of Ontario and New York, from which it was distinguished by its larger size and stouter genal spines. Otarion tiaomachienensis (Sun) (1937: 349, pl. 1 (fig. 1)) has much smaller pre-occipital lobes than the present species and its glabella is more strongly inflated anteriorly, with a strongly convex preglabellar furrow. ‘Proetus’ problematicus Swartz (1925: 56, pl. 1 (figs 3—-5)) may be an otarionid; it shows a small, subtrigonal pygidium that differs from that of the present species in lacking trilobation and segmentation. The Cyphaspis sp. of DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 7) Kozlowski (1923: 62, pl. 4 (fig. 18)) was assigned to Andinacaste legrandi sp. nov. by Eldredge & Brani§a (1980). Eldredge & Ormiston (1979) refer to a second species of Otarion in the Belen Beds of Bolivia, which was said to show considerable resemblance to O. (M.) novellum (Barrande). It was distinguished from M. dereimsi by its semicircular cephalon and slender genal spines, relatively smaller pre-occipital lobes, convex preglabellar furrow, narrow anterior and lateral borders, and median node to the occipital ring. It must also, therefore, be very close to O. (M.) ambiguum (Alberti). Occurrence Malimanaspis? dereimsi (Kozlowski) ranges from the zone of Bainella insolita through to the Metacryphaeus giganteus Zone in Bolivia, but its level of occurrence in South Africa is uncertain. Order PHACOPIDA Salter, 1864 Suborder CALYMENINA Swinnerton, 1915 Family Homalonotidae Chapman, 1890 Subfamily Homalonotinae Chapman, 1890 Genus Burmeisteria Salter, 1865 Type species Homalonotus herscheli Murchison, 1839; by original designation Discussion Two subgenera are currently recognized in this genus; they are Burmeis- teria (Burmeisteria) with a supposedly urceolate glabella showing distinct loba- tion and a biconcave rostral suture, and B. (Digonus) in which the trapezoidal glabella lacks lobation and the rostral suture is transverse or slightly concave. However, as remarked by Sdzuy (1957) and Saul (1965), the form of the rostral suture is highly variable and thus of little taxonomic significance. This is confirmed by the present study, which suggests its form may in part be ontogenetically related and that the shape of the glabella and the strength of its lobation cannot be used to distinguish B. (Burmeisteria) and B. (Digonus). Indeed, available material of B. herscheli, the type of the genus, shows every transition between these two subgenera and at this level the criteria on which they were established do not appear to be of specific let alone subgeneric importance. Consequently, the subgenus Digonus is herein considered a junior subjective synonym of Burmeisteria s.s. In providing the most recent discussion of this subfamily, Thomas (1977) pointed to the similarity of the cephalon of Burmeisteria with that of juvenile Dipleura and suggested that Burmeisteria may be a neotenous descendant of Dipleura. 18 ANNALS OF THE SOUTH AFRICAN MUSEUM Saul (1965) has suggested that this genus may be dimorphic, the dimorphs distinguished by the characters of the pygidium. If present in B. herscheli (Murchison), any such dimorphism is certainly inconspicuous and masked by polymorphism. Burmeisteria herscheli (Murchison, 1839) Figs 9-13, 14A—-B, D-F, 15-28, 39E, 56H Homalonotus herscheli Murchison, 1839: 652, pl. 7 (fig. 2). Salter, 1856: 215, pl. 24 (figs 1-8). Lake, 1904: 214, pl. 26 (figs 1-3). Schwarz, 1906: 383, pl. 8 (fig. 8), pl. 9 (fig. 3). Clarke, 1913: 93, pl. 3 (figs 1-4). Homalonotus perarmatus Frech, 1897: 218. Homalonotus quernus Lake, 1904: 216, pl. 27 (fig. 1). Homalonotus colossus Lake, 1904: 216, pl. 28 (figs 1-3). Homalonotus sp., Lake, 1904: 217, pl. 27 (fig. 2). Homalonotus herscheli Salter var?, Schwarz, 1906: 383, pl. 8 (fig. 8). Homalonotus horridus Schwarz, 1906: 385, pl. 9 (fig. 1). Homalonotus agrestis Schwarz, 1906: 386, pl. 9 (fig. 2). Homalonotus hippocampus Schwarz, 1906: 388, pl. 9 (fig. 5). Homalonotus lex Schwarz, 1906: 389, pl. 9 (fig. 4). Homalonotus (Burmeisteria) herscheli Murchison, Reed, 1918: 314; 1925a: 163. Homalonotus (Burmeisteria) herscheli var. rectisuturalis Reed, 1925a: 167, pl. 10 (figs 3, 7). Homalonotus (Burmeisteria) herscheli var. colossus Lake, Reed, 1925a: 169. Homalonotus (Burmeisteria) herscheli var. fusiformis Reed, 1925a: 171. Homalonotus (Burmeisteria) herscheli var. grahami Reed, 1925a: 172. Homalonotus (Burmeisteria) herscheli var. bituberculata Reed, 1925a: 173, pl. 10 (fig. 2). Homalonotus (Burmeisteria) herscheli var. perarmata Frech, Reed, 1925a: 175. Homalonotus (Burmeisteria) quernus Lake, Reed, 1925a: 181, pl. 9 (fig. 13), pl. 10 (fig. 1). Homalonotus (Burmeisteria?) hippocampus Schwarz, Reed, 1925a: 183. Homalonotus (Burmeisteria) sp., Reed, 1925a: 184. Burmeisteria herscheli (Murchison) Saul, 1965: 271. Burmeisteria hippocampus (Schwarz) Sdzuy, 1957: 283, fig. 1. ? Burmeisteria herscheli (Murchison) BraniSa, 1965: 84, pl. 10 (figs 1, 4), pl. 11 (fig. 6). ? Burmeisteria (Burmeisteria) sp., Wolfart, 1968: 61. Material Salter’s (1856) original material in the British Museum, abundant material in the South African Museum including SAM-—7199, SAM-PDB598, 600, 619, 626-627, 1406, 4292-4293, 4343, 4345, 4352, 4356, 4359, 4360-4364, 7193, 7201, 7795-7796, SU-C6, D182, K810, AM-—1-2, 6, 8, 13, 21, 61-64, 1432, 1444, 1457, 1462, 1613, 1616, 2552-2554, together with RO-16, 175, 257, 271, 807-808, 810-812, 815-819, 825, E22, 50-52, 57-59, 72, 90-92, L20, 60, and 93. Description Cephalon. The cephalon is moderately to very convex (depending on the size and the nature of preservation), with downturned cheeks and a subtrigonal to trapezoidal outline. It is wider than the sagittal length, with a prominent occipital furrow and fairly deep posterior border furrows that shallow abaxially. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 19 Fig. 9. Burmeisteria herscheli (Murchison). The original of Salter (1856, pl. 24 (fig. 1)), preserved as an internal mould. x 1. The lateral border furrows are very shallow and indistinct and there is no anterior border furrow, although the whole of the preglabellar field is gently concave. The preglabellar field is moderately broad, generally between 17-20 per cent of the sagittal length of the cephalon, and may be flattish or distinctly upturned. The glabella is very slightly elevated and generally subtrapezoidal in outline, though tending to urceolate in some individuals. The axial furrows are moderately deep, diverging slightly to the posterior, whilst the anterior margin of the glabella is abruptly truncate but still with rounded corners and a gently convex preglabellar furrow. Glabellar segmentation is variably developed from individual to individual, tending to be most pronounced in juveniles and commonly very weak or obsolete in the middle and adult growth stages. The 3p furrows are weak to indistinct, short, directed posteromedially and subparallel to the slightly longer and more pronounced 2p furrows. The 1p furrows are strongly curved adaxially, almost reaching the occipital furrow and tending to separate weakly developed, subtrigonal, pre-occipital lobes. The occipital fur- row is transverse, sometimes arched forward sagittally, and commonly nar- rower than the occipital ring. The posterior borders broaden slightly abaxially and commonly bear from_1 to 3 small dorsal spines (Fig. 22) just before the rounded genal angles. The subtrigonal paraglabellar areas (Fig. 9) vary from indistinct to prominent, tending to be more pronounced in larger individuals. In some examples the paraglabellar areas tend to invade the L1 lobes, giving them slightly concave lateral margins. The strongly convex cheeks bear prominent rounded bosses on top of which are situated the small elevated eyes surrounded by shallow circumocular furrows. The eyes are situated just posterior to the 20 ANNALS OF THE SOUTH AFRICAN MUSEUM abaxial terminations of the 2p glabellar furrows. The rostral plate is triangular, varying from equilateral to distinctly longer (sag.) than wide, and bears a prominent, hooked, beak-like, median spine. The hypostoma (Fig. 39E) is similar to that of Homalonotus but with only a very weak posterior indentation. The facial suture is proparian, the posterior branches cutting the lateral margins just in front of the genal angles. The anterior facial suture is highly variable, with the rostral suture either transverse, concave or biconcave. This character may be ontogenetically related (Fig. 23) since small individuals generally have a concave rostral suture, while a biconcave suture is encountered only in mod- erately large to large specimens. The entire surface of the cephalon is granu- lose, with randomly scattered tubercles (Fig. 24) that vary in distribution from individual to individual, but tend to be concentrated along the posterior borders. Thorax. The thorax comprises 13 segments with a very broad axis, which is at least 60 per cent of the thoracic width. Ornament comprises scattered granules and randomly distributed tubercles (Fig. 25), the latter tending, however, to form a longitudinal row just inside the axial margins. As shown by SAM-K1406, the tubercles represent spine bases (Fig. 56H). The axial furrows are indistinct but marked by prominent pits at the adaxial ends of the interpleu- ral furrows. The pleural furrows are deep, diverging abaxially from the anterior borders of the pleurae (Fig. 17D), and terminate just before the truncate distal margins of the pleurae. Pygidium. The pygidium is strongly convex, triangular, longer than wide, and with fairly distinct trilobation. The axis is triangular, rather broad an- teriorly but tapering to the posterior and terminating before the posterior margin is reached to leave a short post-axial field. In large individuals the axis comprises 16-17 rings, with the segmentation tending to become indistinct for the last few rings. On internal moulds the terminal point of the pygidium is sharply rounded, acuminate. The axial rings are much wider than the ring furrows and may, or may not, bear a few scattered tubercles (Figs 25-26), while the entire cuticular surface of the pygidium is granulose. The pleural fields are triangular, convex, and curved almost vertically downwards at their margins. They are commonly ornamented with 9-11 pairs of low ribs that are wider than the interpleural furrows. The ribs do not reach the lateral margins, leaving a narrow entire border. The pleural grooves vary from very indistinct to obsolete. Discussion Since a species is defined on the basis of the characters of a population of individuals, it is clear that the thirteen or so taxa of Burmeisteria currently recognized in the Bokkeveld fauna represents a gross splitting of a small number of variable species. Thus, it is readily apparent that B. herscheli displays an extreme range of intraspecific variation and hence genetic poly- morphism. This is most apparent in the characters of the tuberculation and it is worth recalling Reed’s (1925a: 179) comment that *... the lack of bilateral DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 21 symmetry and the general irregularity in the distribution of the tubercles and spine bases on the thorax and pygidium is a noticeable feature in all the forms here regarded as varieties of H. Herscheli, and for this reason the tuberculation must be regarded as of small specific or varietal value. Individuals, otherwise identical, differ considerably in this respect’. The author agrees wholeheartedly with this statement although the varietal rank is here used only for recurring morphotypes that form a significant percentage of the individuals within a contemporaneous population. With regard to the rostral suture, it has already been noted that its form seems to be ontogenetically related and it is not herein considered of specific importance. Similarly, the strength of the glabellar segmentation is so variable, depending on ontogenetic stage and mode of preservation, that it, too, does not appear to be of taxonomic importance among the present material. Burmeisteria herscheli was originally based upon a thorax and pygidium (Murchison 1839, pl. 7 (fig. 2)), with the original of Salter’s (1856) plate 24, figure 1 generally being regarded a typical cephalon. However, as noted by Reed (1925a), Salter’s figure is a restoration based upon more than one individual since the original of his figure 1 (Fig. 9) lacks the occipital ring and posterior borders. Lake (1904) introduced Homalonotus colossus for a very large but frag- mentary cephalon that Clarke (1913: 95) considered to be merely a large B. herscheli. Reed (1925a: 170), however, felt that‘... it may be worthy of being a variety, for there are more numerous (5) tubercles or spine bases on the pleuro-occipital rings, and the thoracic segments bear 2 pairs of lateral spines or tubercles’. As can be seem from Figures 24-26, however, the distribution of tubercles in B. herscheli is highly variable and the use of this character, even at the varietal level, would result in a superfluous multiplicity of names. Conse- quently, the writer follows Clarke (1913) in considering H. colossus a junior subjective synonym of B. herscheli. Homalonotus quernus Lake (Fig. 27A) was based upon a very large fragmentary specimen showing coarse tuberculation on the glabella and thorax. Since, however, there are transitional forms to the sympatric B. herscheli (Fig. 24), the differences are not herein considered of specific importance, although Lake’s name may usefully be retained at the varietal level for the hypernodose individuals within B. herscheli populations. As suggested by Reed (1925a), the fragmentary pygidium figured by Lake (1904: 217, pl. 27 (fig. 2)) as a Homalonotus sp. (Fig. 27B) should be assigned to this variant. Homalonotus perarmatus Frech (Fig. 20) and H. horridus Schwarz are both multituberculate ‘species’ transitional from B. herscheli s.s. to the var. quernus. Consequently, the writer would agree with Lake (1904) and Reed (1925a) in regarding H. perarmatus as a junior subjective synonym of B. herscheli, as well as following Reed (1925a) in including H. horridus Schwarz in the synonymy of Murchison’s species. The retention of these names even at the varietal level serves little practical purpose and they are not used here. ANNALS OF THE SOUTH AFRICAN MUSEUM Up) jenjen) wnasnyy ysnig “ST X ‘pjnow jeusdjuUI ue se paarasaid ‘(A10]SIH] ay) ut ‘((Z ‘3Yy) pz Id ‘ggg]) JaIeSg Jo JeuIsiIo oy, ‘(UOSTYyoIN|,) ayossay pDlualsiawung “QT “3I4 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP M8; Fig. 11. Burmeisteria herscheli (Murchison). The original of Salter (1856, pl. 24 (fig. 4)), in the British Museum (Natural History), preserved as an internal mould. This specimen was referred to the var. perarmata Frech by Reed (1925a). x 1. Homalonotus agrestis Schwarz (1906: 386) was distinguished by ‘ . . . the very rapid tapering of the pygidium and the incurved margin’ but, as noted by Reed (1925a: 178), is so similar to H. horridus that it, too, is included in the synonymy of B. herscheli. Homalonotus hippocampus Schwarz was based upon a number of small cephala distinguished by their subtrigonal outlines and flat glabellae. According to Reed (1925a: 183), however, * ... the shape of the glabella, the practical absence or obsolescence of the lateral furrows on its ANNALS OF THE SOUTH AFRICAN MUSEUM 24 “Tl x IP pue sp[nou [Buto}UL se poarasaid ITV “SIJDANINSIIIA “IVA oq} Oj Ppoltojol a ‘3y) Ol ‘Id “DCZ6I S,Pp9I9y jo [BUIsIIO oy) “C6ILHdd-WNVS “Ste D) “QINJNS [PIISOI DAKIUOD dy) ION ‘ZIEPACd-NVS “A ‘Wowsesy yerpistd ve ‘76z7pPadd-NVS “V “(UOSTYoINI) Yayosiay YLaIsiaULng “7 “B14 25 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP Wek [Ihe pure spjnow jeusajur se paarasaid Ty “S6LLUAdd-WVS ‘O ‘09¢PAdd-WVS “A ‘C9EhAdd-WVS 'V ‘(UOSIYOINIA]) YaYyISday Dldajsiauing “¢] “3I4 26 ANNALS OF THE SOUTH AFRICAN MUSEUM surface, the straight, transverse, connecting suture in front, the presence of a pair of spine bases close to the genal angles, and other features suggest that it is only a young individual or narrow form of H. herscheli var. rectisuturalis. In the Ordovician trilobite Ogygia buchi Brongniart we have a narrow and a wide form, the differences being regarded as sexual. Possibly the same explanation is applicable here.’ Perhaps significantly, Saul (1965) has recorded size dimorph- ism in Burmeisteria antarctica and hence Reed’s opinion may be correct. In any case, the differences are not herein considered of specific importance and Homalonotus hippocampus is included in the synonymy of B. herscheli. Homalonotus lex Schwarz was founded upon several thoracic segments (Fig. 17D) which Schwarz (1906: 389) felt might have belonged to the cephala he had named H. hippocampus, while noting that ‘ . the rings show no particular features differing from the usual H. herscheli type, except in the ends, which in H. herscheli are rounded in the lower margin, in this species are pointed, and probably bore spines’. The writer would agree with Reed (1925a) that this is an unnecessary species that should be included in the synonymy of B. herscheli. Reed (1925a) created the var. rectisuturalis (Fig. 16E) within B. herscheli for a cephalon ‘ ... which can hardly be separated specifically from H. herscheli, but certainly constitutes a variety remarkable for the shortness and breadth of its glabella’. The writer finds little use for this varietal name and it is not used here. The var. fusiformis was created by Reed (1925a) for the pygidium figured by Lake (1904) (Fig. I1SA-B herein) and, as the varietal name implies, distinguished by its subfusiform outline. While this variant is distinctive, no further individuals have been forthcoming and the characters seem to be those of an aberrant individual. As such, the perpetuation of Reed’s name is tantamount to naming individuals and it is not used here. Reed (1925a: 172) introduced the var. grahami (Fig. 21) for forms in which the glabella was said to be more oblong in shape and relatively longer than in typical B. herscheli, as well as having the cephalon more pointed anteriorly, narrower, with steeper cheeks, and with‘... the meso-occipital furrow more sharply bent forwards in the middle, the paraglabellar areas invade the basal lobes more deeply, and the minute ornament is slightly different, the tubercles being more numerous and closer together; the transverse suture is typically straight and not angular’. Since, however, most of these characters are also found in the ‘var. rectisuturalis’ and in ‘H. hippocampus’, this variety seems to be based upon trivial differences of the individual and the name is not used here. Burmeisteria notica (Clarke) (1913: 89, pl. 1 (figs 1-2), pl. 2 (figs 1-13)) differs from B. herscheli in the complete lack of tubercles and spine bases to the exoskeleton, a broader preglabellar field (about 25% of the sagittal length of the cephalon) and a pygidium with fewer (13-15) axial rings. It is perhaps also worthy of note that the pygidium figured by Clarke (1913) as plate 2, figures DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP Off 11-12 shows a distinctly deeper first ring furrow than the rest. Burmeisteria clarkei (Kozlowski) and B. accraensis Saul are both very closely allied to B. notica and distinguished from B. herscheli by the complete lack of tubercula- ton. Burmeisteria antarctica Saul (1965: 269, pl. 17 (figs 1-11)) (Fig. 29 herein) differs from B. herscheli in its lack of tuberculation, shorter rostral suture, allegedly smaller eyes and weaker glabellar segmentation, its relatively broader cephalon, and particularly in features of the pygidium. Thus, in the “Type I’ pygidia (Fig. 29) of B. antarctica only the anterior 7-9 axial rings are distinct and there are only 7 pairs of well-developed ribs to the pleural fields with very deep interpleural furrows. The ‘Type II’ pygidia are closer to B. herscheli in showing 12 or more axial rings, but still with only 6—7 pairs of pleural ribs. Burmeisteria expansa (Hector) (Allan 1935: 29, pl. 1 (fig. 1)) from the Reefton Beds of New Zealand is based upon pygidia which differ from B. herscheli in being relatively much broader and with only 12-13 distinct axial rings and 9-12 pairs of pleural ribs. Burmeisteria huttoni Allan (1935: 28, pl. 1 (figs 4-5)) is from the same beds and, in view of the range of variation displayed by B. herscheli, probably does not warrant specific separation from B. expansa. So far as can be judged, B. huttoni differs from B. herscheli in that the coarse tuberculation of the thorax forms distinct transverse rows, whereas in the South African species the rows tend to be longitudinal. Occurrence Burmeisteria herscheli (Murchison) is presently recorded from the Gydo, Gamka, and Voorstehoek Formations of the Bokkeveld sequence, the Fox Bay Beds of the Falkland Islands and, perhaps, from Bolivia. Burmeisteria notica (Clarke, 1913) Figs 14C, 30 Homalonotus noticus Clarke, 1913: 89, pl. 1 (figs 1-2), pl. 2 (figs 1-13). Homalonotus cf. noticus Clarke, Kozlowski, 1923: 23, pl. 1 (fig. 11). Homalonotus sp., Kozlowski, 1923: 13, pl. 3 (figs 8-9). ? Homalonotus (Burmeisteria) herscheli var. rectisuturalis Reed, 1925a: 160 (SU-—C1 only). Homalonotus (Digonus) noticus var. africana Reed, 1925a: 184, pl. 10, (fig. 4). Digonus noticus (Clarke) Sdzuy, 1957: 279. Digonus cf. noticus (Clarke) Branisa, 1960: 60, pl. 6 (fig. 13); 1965: 84, pl. 10 (figs 5-7), pl. 11 (fig. 13), pl. 75 (fig. 2). Burmeisteria (Digonus) noticus (Clarke) Saul, 1965: 271. Burmeisteria (Digonus) cf. noticus (Clarke) Wolfart, 1968: 62. Material The holotype of Reed’s var. africana, SAM-7796, together with RO-17, 60, 88, 174, 297 and P57, and questionably SU-C1. 28 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 14. A-B, D-F. Burmeisteria herscheli (Murchison). A-B. SAM—PDB619, the original of Lake (1904, pl. 26 (fig. 1)). D. SAM-—7201, with the lectotype of Deltacephalaspis pseudocon- vexus (Lake) impressed in its terminal portion. E-F. SAM-PDB4345. C. Burmeisteria notica (Clarke). SAM-—7796, the holotype of the var. africana Reed (1925a, pl. 10 (fig. 4)). A-B, D-F preserved as internal moulds and all x 1. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 29 Fig. 15. Burmeisteria herscheli (Murchison). A-B. SAM-PDB627, the holotype of the var. subfusiformis Reed, figured by Lake (1904, pl. 26 (fig. 2)). C. SAM-PDB4296. D-E. SAM-PDB3075. F-G. SAM-PDB4353. All preserved as internal moulds and all x 1. 30 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 16. Burmeisteria herscheli (Murchison). A. SU-K810. B-C. A tectonically foreshortened pygidium in the South African Museum. D. The original of Salter (1856, pl. 24 (fig. 7)), in the British Museum. E. SU-C6, figured by Reed (1925a, pl. 10 (fig. 3)) as the var. rectisuturalis Reed. All preserved as internal moulds and all x 1. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP Sil Fig. 17. Burmeisteria herscheli (Murchison). A. SAM-7199, the holotype of the var. sodalis Reed, figured by Reed (1925a, pl. 10 (fig. 5)), and preserved as an internal mould. B-C. SAM-PDB600, the original of Lake (1904, pl. 26, (fig. 3)), which retains cuticle. D. The lectotype, designated herein, of Homalonotus lex Schwarz, AM-1462, preserved as an internal mould, showing the distal terminations of the thoracic pleurae. All x 1. 32 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 18. Burmeisteria herscheli (Murchison). A-B. SAM-PDB4356. C. SAM-PDB4293. D. SAM-PDB4361, with a concave rostral suture. E. SAM-PDB4343. F. SAM-PDB4352. G. An unnumbered cephalon in the South African Museum. All preserved as internal moulds and all x 1. 33 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP ‘Tx [[@ pue spjnow eussjur se poasasaid [TV ‘6serddd-WVS ‘O ‘p9EPAdd-WNVS “A ‘“wWhasny uROLyY YING oy) Ul UsUDads poroqunuUN uy “Vy ‘(UOSIYDIN,) Yayosiay visajsiauung ‘61 “S14 34 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 20. Burmeisteria herscheli (Murchison). The holotype. by monotypy, of Homalonotus perarmatus Frech, almost certainly an internal mould (after Frech 1897). x 1. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 35 Fig. 21. Burmeisteria herscheli (Murchison). AM-—2554, the original of Schwarz (1906, pl. 8 (fig. 8)), which was referred to the var. grahami Reed by Reed (1925a), preserved as an internal mould. x 1. Fig. 22. Burmeisteria herscheli (Murchison). A genal fragment in the collection of R. Oost- huizen, showing the tubercles of the posterior border to represent spine bases, with some of the spines bifurcating. Scale = 5 cm. 36 ANNALS OF THE SOUTH AFRICAN MUSEUM GLABELLAR LENGTH 80mm 0 0 80mm GLABELLAR WIDTH AT S1 Fig. 23. Burmeisteria herscheli (Murchison). Plot of glabellar length (sag.) against glabellar width at the S1 sutures for the various rostral suture types. Circles =concave rostral suture; triangles=straight rostral sutures; dots = biconcave rostral sutures. Discussion This species is so similar to B. herscheli that it does not warrant additional description on the basis of the available material. The small cephalon that formed the basis of Reed’s var. africana (Fig. 14C) shows the lack of tubercula- tion and relatively broad preglabellar field herein considered diagnostic of Clarke’s species. It was assigned to B. notica by Reed (1925a: 185) ©... on account of its proportions, the truncate and emarginate anterior margin, the characters of the glabella, the course of the facial sutures, the well-rounded genal angles, and especially in the absence of tubercles and spine bases. But our specimen has a coarsely-granulated instead of smooth surface’. This latter feature is surely an artefact of preservation, the Bokkeveld example retaining cuticle, and hence Reed’s varietal name is not used here. The small individual (Fig. 30) referred by Reed (1925a: 168) to B. her- scheli var. rectisuturalis shows the lack of tuberculation typical of B. notica, with the exception of a pair of spine bases on the posterior border just adaxial DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP oF Fig. 24. Burmeisteria herscheli (Murchison). Schematic representation of the cephalic tuberculation of various individuals. Hatched areas not preserved. 38 ANNALS OF THE SOUTH AFRICAN MUSEUM TS Fi nN Pewee acs IMAL eri LUT TTT ETT PTTL 1) Benes ant SONA ELUTE TT] = ALLER \ AL RRNNNSSSANAS SO S cca Ns 2S Jay) 1 Schematic representation of the thoracic and pygidial tuberculation of various individuals. Hatched areas not preserved. (ess is DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 39 Fig. 26. Burmeisteria herscheli (Murchison). Schematic representation of the pygidial tuberculation of various individuals. Hatched areas not preserved. of the genal angles. The importance of this feature is uncertain, since this individual also shows a pygidium with fewer axial rings (probably about 14) than B. herscheli, and hence also approaches B. notica in this respect. Burmeisteria clarkei (Kozlowski) (1923: 24, pl. 1 (figs 12-15)) is very close to B. notica from which it was distinguished by the weaker trilobation and segmentation of the pygidium, with a slightly different outline, and in apparently having only 8 axial rings and 8 pairs of pleural ribs discernible. The differences are slight and objective population analysis is required to confirm the distinction of this species from B. notica. Burmeisteria accraensis Saul (1967: 1129, pl. 143 (figs 1-8), pl. 144 (figs 1-16)) from the Devonian (Eifelian?) of Ghana was characterized by its lack of tuberculation and lobation, and especially by having the first ring furrow of the pygidium significantly deeper than the remainder. However, as noted by Saul (1967: 1134) this latter feature is also found in at least some individuals of B. clarkei while © ... the relative dimensions of H. clarkei, as measured from Kozlowski’s plates, usually fall on the extreme peripheries of the scatter diagrams which exhibit the range of variation of the Accraian species’. Kozlow- ski’s species, however, was said to have a more acuminate pygidium. Since the 40 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 27. Burmeisteria herscheli var. quernus (Lake). A. SAM-626, the holotype of this variety and original of Lake (1904, pl. 27 (fig. 1)). B. The pygidial fragment figured by Lake (1904, pl. 27 (fig. 2)), in the South African Museum. Both preserved as internal moulds and both x 1. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP Fig. 28. Burmeisteria herscheli var. quernus (Lake). A. SAM-2454, x 2. B. A speci- men in the South African Museum, 1,5. Both preserved as internal moulds. 41 42 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 29. Burmeisteria antarctica Saul. A. ‘Type I’ pygidium in lateral view (after Saul 1965). x 1. Fig. 30. Burmeisteria cf. notica (Clarke). SU-C1, a juvenile preserved as an internal mould that may belong here. This specimen was referred to Homalonotus herscheli var. rectisuturalis by Reed (1925a). x 1. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 43 population characteristics of B. clarkei are currently unknown, the differences between the two species are slight and questionably of specific importance. Saul (1967) failed to compare B. accraensis with B. notica. The lack of tuberculation and glabellar segmentation are all features of B. notica and in relative propor- tions the cephala of the two species are all but indistinguishable. Moreover, the pygidium of B. notica figured by Clarke (1913, pl. 2 (figs 11-12)) shows the same deep first ring furrow. Burmeisteria accraensis seems to differ from B. notica only in its slightly shorter (sag.) preglabellar field and slightly nar- rower pygidium, with all individuals apparently having the first ring furrow deepened. ‘Homalonotus derbyi Clarke (1890: 7, pl. 1 (figs 4, 7, 19)) was said to be . an excellent representative of the subdivision Trimerus’, while the lack of segmentation to the pleural areas of the pygidium of ‘H’. kayseri Thomas (1905: 245, pl. 11) suggests that it, too, may be a species of Trimerus. Burmeisteria antarctica Saul was said to differ from B. notica in having a straighter and shorter rostral suture, a more indented posterior margin to the glabella, and a pygidium with deeper interpleural furrows becoming smooth near the posterior termination, and with a sharper break at the connection of ring furrows and rib furrows. Certainly “Type I’ pygidia are easily distinguished from B. notica, but the “Type II’ pygidia, which Saul (1965) failed to compare with B. notica, are very similar to Clarke’s species. Clearly, however, the problem will be resolved only with the much-needed modern revision of Clarke’s material. Occurrence Burmeisteria notica (Clarke) is currently recorded from Argentina, Bolivia, and the Bokkeveld succession. The definitely localized specimen in the collec- tion of R. Oosthuizen comes from the Gydo Formation. Burmeisteria fontinalis (Reed, 1925) Fig. 31 Homalonotus (Digonus) fontinalis Reed, 1925a: 185, pl. 10 (fig. 6). Material Only the holotype, SMC—A3069, from Ezelfontein is currently known. Description Cephalon. The anterior portion of the cephalon is indistinct in the photo- graph of the holotype supplied by C. L. Forbes but, according to Reed (1925a: 185), it has a ‘ ... truncate anterior edge, excavated in the middle’. The glabella is weakly inflated, almost flat, and with indistinct lobation. It is subrectangular in outline, narrowing slightly to the anterior. The glabellar 44 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 31. Burmeisteria fontinalis (Reed). The holotype, by monotypy, SMC-A3069, pre- sumably an internal mould. Note the lack of tuberculation and numerous pleurae to the pygidium. Approx. X 5. furrows are all directed strongly posteromedially and all seem to be slightly convex anteromedially. The occipital furrow is straight, narrow, of uniform depth and distinctly narrower than the occipital ring. The posterior border furrows are distinct and the posterior borders broaden (exsag.) abaxially. The cheeks are poorly preserved but there appear to be weakly developed para- glabellar areas, and the small eyes are raised on prominent bosses. The surface of the cephalon is granulose but discrete tubercles are lacking. Thorax. The thorax is poorly preserved, but with a wide axis, 13 segments and a granulose surface. Pygidium. The pygidium is strongly convex, subtrigonal in outline and with well-developed trilobation. The axis is rather broad anteriorly and tapers posteriorly into a bluntly rounded tip to leave a short post-axial field. The axis comprises about 15 rings which are slightly wider than the ring furrows. The pleural fields are subtrigonal and ornamented with 14-15 sharply raised, distinct DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 45 ribs which reach the pygidial margin and are said to be ornamented with coarse granules. Discussion This species is most closely related to B. notica (Clarke) and its allies. It differs from B. notica by its shorter preglabellar area, its relatively much longer glabella, and especially in the characters of the pygidium, B. fontinalis having many more pleural ribs than Clarke’s species. The above differences also serve to distinguish B. fontinalis from B. clarkei (Kozlowski) and B. accraensis Saul. Occurrence Burmeisteria fontinalis (Reed) is known only by the holotype which comes from an unknown level in the Bokkeveld sequence. Superfamily DALMANITACEA Vogdes, 1890 Family Dalmanitidae Vogdes, 1890 Subfamily Dalmanitinae Vogdes, 1890 Genus Francovichia Brani§a & Vanék, 1973 Type species Odontochile branisi Wolfart, 1968; by original designation Francovichia clarkei (Ulrich, 1893) Figs 32A, 33D, 34A, 35-36, 37A-B Dalmanites clarkei Ulrich, 1893: 19, pl. 1 (fig. 13). Lake, 1906: 429. Knod, 1908: 500. Kozlow- ski, 1923: 106. Swartz, 1925: 34. Dalmanites lunatus Lake, 1904: 212, pl. 25 (fig. 6). Dalmanites sp., Lake, 1904: 212, pl. 25 (fig. 7). Non Dalmanites lunatus Lambert, 1904: 482, pl. 44 (figs 1-3, 5) (=D. limurulus (Green) fide Delo, 1940). Dalmanites (Hausmannia) dunni Reed, 1925a: 133, pl. 11 (fig. 5). Dalmanites (Hausmannia) lunatus Lake, Reed, 1925a: 135. Phacops sp., Brink, 1951: 162, figs 1-2. Odontochile clarkei (Ulrich) Wolfart, 1968: 75. Material The holotypes of Dalmanites lunatus Lake, SAM-—66, and D. dunni Reed, SAM-3949, together with RO-5, 51, 780, 783, PRV-—34-35, 623, and BPI-i113 and 1243. Description Cephalon. The cephalon is known, among South African material, largely from the holotype of D. lunatus Lake (Fig. 35). This shows it to have been semicircular, transversely much wider than long (sag.), and with prominent genal spines extending backwards as far as the fourth thoracic segment. The 46 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 32. A. Francovichia clarkei (Ulrich). The holotype, by monotypy. of Dalmanites (Hausmannia) dunni Reed (1925a, pl. 11 (fig. 5)). in the South African Museum, SAM-3949, preserved as an external mould. B. Gen. et sp. indet. SAM-—784, the original of Lake (1904, pl. 25 (fig. 5)), preserved as an internal mould. Both x 1,5. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 47 glabella expands anteriorly, with a much enlarged frontal lobe, but the anterior region of the cephalon is too poorly preserved to be certain of the characters in this region. The 3p furrows are fairly deep and directed strongly postero- medially. The 2p and 1p furrows are moderately deep, transverse, and seem to connect with the well-developed axial furrows. Both the L1 and L2 lobes are of more or less constant width (exsag.). The occipital furrow is transverse, and the smooth occipital ring seems to be broadest sagittally. The posterior border furrows are deep and transverse, whereas the lateral border furrows are shallow. The eye is very large, certainly 50 per cent of the exsaggital cephalic length. A well-preserved eye that seems to belong to this species (Fig. 36) comprises about 50 vertical files of lenses, with 15 ocelli in the central file and a total of about 577 ocelli to the visual surface. It also shows an unornamented palpebral lobe weakly separated from the unornamented palpebral area, as well as a Shallow subocular groove separating the eye from the socle. The posterior branch of the facial suture curves strongly forward before recurving to meet the lateral border well in front of the posterior margin to the eye. Thorax. The thorax comprises 11 segments, with a rather low, weakly convex axis about one-third of the thoracic width. The ring furrows are very much narrower than the axial rings themselves. The axis broadens (trans.) to the third or fourth ring whereafter it tapers slowly to the posterior. The pleural grooves are pronounced, with finer, narrower interpleural furrows. Pygidium. Although the pygidium of D. lunatus is poorly preserved, Dalmanites dunni is based upon adult topotype material (Fig. 32A) and, together with additional topotype material, allows for a relatively comprehen- sive description of the pygidium of this species. The pygidium is large, triangular, distinctly wider (trans.) than long (sag.), and with a long, narrow axis. The axis tapers regularly to the posterior, and comprises 18-21 rings. After the tenth axial ring, the ring furrows become obsolete sagittally. In none of the available material is the terminal portion of the pygidium preserved, but in BPI-i243 (Fig. 37A-B) it is at least distinctly mucronate; it is assumed to have had a terminal spine as in the South American material. The pleural fields are broad, subtrigonal, strongly downturned at the margins in well-preserved material, and with an entire border. There are 11 pairs of well-developed, flat-topped ribs, each bearing a very faint interpleural furrow and separated by a deep, but narrow (exsag.) pleural groove, with a very reduced and indistinct twelfth pair to the posterior. In the anterior pleurae, the pleural grooves meet the axis almost at right angles, but the angle becomes increasingly acute to the posterior, until the final few ribs are almost parallel to the axis. The pleural segmentation is effaced distally to leave a smooth narrow border. BPI-i243 retains granulose cuticle. Discussion Both Lake (1904) and Reed (1925a) remarked on the close similarity between D. dunni and D. clarkei without effectively distinguishing them. The 48 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 33. A, C. Metacryphaeus caffer (Salter). A. SU-H76, the holotype of Eocorycephalus capensis (Reed) (1925a, pl. 8 (fig. 5)). C. SU-C2, the holotype of Dalmanites (Proboloides) ensifer Reed (1925a, pl. 8 (fig. 4)), which probably belongs to Salter’s species. Compare with Figure 39F. x 1.5. B. Bainella africana (Salter). SU-K461. x1. D. Francovichia clarkei (Ulrich). RO-5. x 1. E-F. Typhloniscus baini Salter. E. RO-702:; note the ‘eye ridges’. x 1. F. RO-753, an individual retaining cuticle. X 2. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 49 Fig. 34. A. Francovichia clarkei (Ulrich). PRV—623, an external mould. x 1,3. B. Metacry- phaeus caffer (Salter). PRV—1028, an internal mould showing a prominent auxiliary impres- sion system. X 1,8. 50 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 35. Francovichia clarkei (Ulrich). Plasticine squeeze of the holotype of Dalmanites lunatus Lake (1904, pl. 25 (fig. 6)), an external mould. Note the very large eyes. x 5. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 51 writer cannot find any significant characters to separate these two species, and D. dunni is considered a junior subjective synonym of Francovichia clarket. Dalmanites lunatus Lake is based upon an incomplete juvenile (Fig. 35) that was collected from the same locality (Gamkapoort) as the holotype and topotype material of D. dunni. As such, the two species seem to be based upon different parts of different ontogenetic stages of the same animal and they are herein regarded as conspecific. Dalmanites lunatus Lake is, therefore, included in the synonymy of Ulrich’s species. In the same year that Lake (1904) created his species, Lambert (1904) also described a new species as Dalmanites lunatus from the Silurian of North America. Since, according to Delo (1940: 44), D. lunatus Lambert (non Lake) is a synonym of D. limurulus (Green), while Lake’s homonym is here included in the synonymy of F. clarkei, this is a case for letting sleeping dogs lie and the nomenclatural priorities of these two names are ignored. Francovichia branisi (Wolfart) (1968: 72, pl. 6 (fig. 3), pl. 7 (figs 1-2)) is very close to the present species. According to Wolfart (1968), there are 20-22 axial rings and 14 pairs of pleural ribs to the pygidium of F. branisi, which was also said to have a somewhat narrower pygidium (with margins that are weakly concave posteriorly) than F. clarkei. The differences in outline and width are slight and may reflect merely intraspecific variation. Moreover, it is difficult to determine more than 17 axial rings on the pygidium of F. branisi figured by Wolfart (1968, pl. 6 (fig. 3)), or more than 12 pairs of pleural ribs. Franco- vichia branisi does, however, seem to have smaller eyes and is for the present retained as distinct. ‘Dalmanites’ andii Kozlowski (1923: 38, pl. 2 (figs 3-4)) differs from the present species in having a much wider pygidium, and with prominent inter- pleural furrows that are almost as wide as the ribs. These are also much the same characters that distinguish ‘Dalmanites’ patacamayaensis Kozlowski (1923: 36, pl. 2 (fig. 2)), Odontochile dunbari Delo (1940: 56, pl. 5 (fig. 7)) and O. micrurus (Green) (Delo, 1940: 58, pl. 5 (figs 1-2)). The ‘Dalmanites’ maecurua Clarke figured by Knod (1908: 500, pl. 21 (fig. 3)) seems to have fewer axial rings and wider interpleural grooves, which extend to the pygidial margin, than the present species. Occurrence Francovichia clarkei (Ulrich) is currently known only from the Icla Forma- tion of Bolivia and the Gydo Formation of the Bokkeveld sequence. Family Synphoriidae Delo, 1935 Discussion The evolution and phylogeny of this largely North American family has been discussed at length by Lespérance & Bourque (1971, 1973) and Lespér- ance (1975), and two subfamilies are currently recognized, the Synphoriinae 52 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 36. Franchovichia clarkei (Ulrich). BPI-i113. The large eye figured by Brink (1951, figs 1-2), preserved as an internal mould. x 2,5. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 53 Fig. 37. A-B. Francovichia clarkei (Ulrich). BPI-i243, which seems to retain cuticle. x 2,4. C-D. Bainella sp. nov. BPI-i244, a pygidium retaining cuticle. Note the weak tubercles on the ribs at the pleural axis. x 2. 54 ANNALS OF THE SOUTH AFRICAN MUSEUM and Trypaulitinae, with the origins of the family lying in ‘*... a generalized Silurian Odontochile’ (Lespérance 1975: 99). Branigsa & Vanék (1973) intro- duced three new genera from the Devonian of Bolivia that they believed should be included here; these were Gamonedaspis, Fenestraspis, and Chacomurus. However, only Chacomurus was accepted as a true representative of the Synphoriidae by Lespérance (1975), who included it in the subfamily Trypauli- tinae. Both Gamonedaspis and Fenestraspis were excluded because *... they do not possess the characteristics of the family as now understood’ (Lespérance 1975: 101). It is perhaps significant, however, that Lespérance & Bourque (1971) provisionally included Dalmanites accola Clarke in the Synphoriinae, although Lespérance (1975: 103) later concluded that *... no compelling evidence exists for [its] inclusion in the Synphoriidae’. Contrary to the opinion of Lespérance (1975), however, Dalmanites boehmi Knod and the perhaps conspecific D. accola Clarke are very close to certain Synphoriidae, especially Roncellia and Anchiopsis. The Malvinokaffric species, on the face of it, seem to be very closely related to Roncellia perceensis (Clarke) (Delo 1940: 59, pl. 5 (figs 5-6), pl. 6 (fig. 7); Lespérance & Bourque 1971: 190, pl. 26 (figs 1-3)), the main points of difference being the relatively narrower pygidium with fewer pleural ribs of the Malvinokaffric species. Since these differences are no greater than those between many species of dalmanitid trilobites, one is left wondering whether the Malvinokaffric species should not be included in the genus Roncellia. However, Dalmanites boehmi and D. accola both differ from typical Roncellia in having a median tubercle to the occipital spine, in which respect they approach Anchiopsis. The pygidium of the latter genus is sufficiently distinct, however, to preclude reference of the Malvinokaffric species to this taxon. The type species of Gamonedaspis, G. scutata Branisa & Vanék (1973: 98, pl. 1 (fig. 6), pl. 4 (figs 1-2)), is known only from pygidia that show the following features: longer than wide, subtrigonal in outline, and terminat- ing in an uptilted postaxial spine. The axis tapers rather rapidly to the posterior and comprises 13-14 rings, whereafter the segmentation becomes indistinct. There are 9-10 pairs of pleural ribs, of which the anterior 3-4 show faint pleural grooves. There are two longitudinal rows of tubercles on the axis and randomly scattered tubercles on the pleural ribs. Thus, G. scutata differs from Dalmanites boehmi and D. accola mainly in its uptilted post-axial spine. On the basis of available evidence the differences are not herein considered of generic importance and the writer would also refer D. boehmi Knod, D. boehmi var. boliviensis Kozlowski, and D. accola to the genus Gamonedaspis. Whether the resemblances between Gamonedaspis and Roncellia are another example of the perturbing homoeomorphy between boreal and Malvinokaffric trilobites is at present uncertain. On the face of it, Roncellia, Anchiopsis, Dalmanitoides, and Gamonedaspis show sufficient similarities to Suggest they may form a third lineage within the Synphoriidae. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 5 Subfamily Synphoriinae Delo, 1935 Genus Gamonedaspis Brani§a & Vanék, 1973 Type species Gamonedaspis scutata Brani$a & Vanék, 1973; by original designation Gamonedaspis boehmi (Knod, 1908) Figs 7A, 38, 39D, 40, 41A, 42C, 43, 44A Dalmanites b6hmi Knod, 1908: 568, pl. 21 (fig. 1) Dalmanites boehmi var. boliviensis Kozlowski, 1923: 36, pl. 2 (fig. 1). Swartz, 1925: 34. Wolfart 1968: 70. Dalmanites boehmi (Knod) Reed, 1925a: 133. Francovichia? boehmi (Knod) Eldredge & Ormiston, 1979: 162. Material Thirteen specimens, SAM-—PDB3997, 4056, 4070, 4332, 4336, 4338-4339, 4348, 4358, RO-282, and PRV—34, 126, 1096, all preserved as internal moulds. Description Cephalon. The cephalon is semicircular in outline, about twice as wide as the sagittal length, and weakly convex. The genal angles are not preserved in the South African material but are assumed to be as in Knod’s (1908) figured type. The axial furrows are strongly impressed and diverge markedly to the anterior. There is a narrow preglabellar field that forms a bluntly rounded median rostrum (Fig. 41A). The 3p furrows are rather deep, directed strongly BR : RES / / i f Fig. 38. Gamonedaspis boehmi (Knod). \ Reconstruction of the cephalon based upon the available material. x 1. 56 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 39. A-B. Bainella africana (Salter). SAM-PDB3076. x 2. C. Typhloniscus baini Salter. RO-702a, a pygidium. x1.1. D. Gamonedaspis boehmi (Knod). SAM-PDB4338. X 2. E. Burmeisteria herscheli (Murchison). An hypostoma, RO-279. x 1. F. Metacryphaeus caffer (Salter). SAM-3960. x 2. All specimens preserved as internal moulds. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 57 posteromedially, and almost meet adaxially. They are weakly connected proximally by a shallow transglabellar furrow that serves to isolate the pro- minent, coarsely tuberculate, rhomboidal frontal lobe. The 2p and 1p fur- rows are moderately deep, subparallel, with prominent adaxial apodemal pits, and distinctly joined to the axial furrows. In large individuals (Fig. 39D), the L3 lobes protrude laterally almost as much as the frontal lobe, while the Ll lobes are very narrow (exsag.) and reduced. In PRV-126a (Fig. 7A), the 3p and 2p furrows are joined proximally, so as to isolate the L3 lobes, while the 2p and Ip furrows are also weakly joined, tending to isolate the L2 lobes. Moreover, the 2p and 1p furrows are weakly connected across the glabella (Fig. 43A) by variably developed transglabellar furrows. The occipital furrow is rather narrow, with apodemal pits distally, and separates the moderately broad occipital ring that bears a prominent median tubercle (spine base?). The cheeks are subtrigonal in outline, mod- erately downturned, with very deep posterior border furrows and posterior borders which broaden (exsag.) abaxially. There are distinct palpebral fur- rows that connect with the axial furrows anteriorly and, seemingly, with the furrow separating the eye from its socle posteriorly. The cheeks lack tuber- culation, although there are small tubercles on the palpebral lobes above the eyes. The eyes are large, arcuate, about half the exsagittal length of the cephalon, and separated from the L3 lobes anteriorly only by the axial furrows. The visual surface comprises about 30-32 subvertical rows of ocelli, with 8-9 lenses in the central files. The anterior ramus of the facial suture touches, but does not transect, the lateral margin of the frontal lobe. The posterior branch curves strongly forward before recurving sharply to meet the lateral border in front of the posterior margin of the eye. Thorax. Poorly known from South African material. So far as can be ascertained there are no median tubercles to the axis nor any tubercles on the pleurae. The axial furrows are moderately impressed, and there are deep pleural grooves to the pleurae, which are separated by distinct interpleural furrows. The distal terminations of the pleurae are unknown. Pygidium. The pygidium is triangular, slightly wider than the sagittal length (if the length of the post-axial spine is excluded), with gently convex borders and a short terminal spine. The axis tapers uniformly to the posterior and comprises between 15-18 rings, the segmentation becoming indistinct posteriorly, each ornamented with a pair of tubercles close to the sagittal line. The pleural fields are subtrigonal and separated from the axis by prominent axial furrows. There are 11 pairs of rather broad, flat-topped ribs, the anterior 6-7 with faint pleural grooves. The interpleural furrows are deep and clearly defined, and much narrower than the ribs. The pleural ribs meet the axis at an acute angle (almost at right angles for the first few pleurae), which increases to the posterior. A few randomly scattered tubercles of varying sizes ornament the pleural fields, some situated directly upon the faint pleural grooves, and show a tendency to mark the fulcral axis. 58 ANNALS OF THE SOUTH AFRICAN MUSEUM Fae inal .. ee Fig. 40. Gamonedaspis boehmi (Knod). A. SAM-—PDB4339, a fragmentary cephalon. B. SAM-PDB4332, a pygidium retaining the terminal spine. Both preserved as internal moulds and both x 2. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 59 Fig. 41. A. Gamonedaspis boehmi (Knod). SAM-PDB4358. Note the smooth anterior border and well-developed eye socle. x 1,5. B-C. Typhlo- niscus baini (Salter). B. SAM-7799. The holotype of Acidaspis capensis Reed (1925a, pl. 11 (fig. 6)). x 1,5. C. A specimen in the South African Museum. Note the strong posterior taper of the thorax. x 1. All specimens preserved as internal moulds Discussion Until the full range of intraspecific variation within G. boehmi is documented, on the basis of a population of individuals, the largely trivial characters on which Kozlowski (1923) separated his var. boliviensis are mean- ingless and this taxon is not used here. Gamondedaspis scutata Branisa & Vanék differs from the present species in its uptilted post-axial spine, fewer (9-10) and narrower pleural ribs, and ANNALS OF THE SOUTH AFRICAN MUSEUM 60 ‘7 X [Je pue ‘spynow jeusojUI se poarosoid suautoads IV “o49 oyi JO dovJINs [eNsIA oy} Surmoys ‘uoyeydao ay} JO MIIA [e1DIvT “(pouy) muya0q sidsppauowwyH ‘DJ ‘aqnuids yeuds ou Jo uorssaidtut ay} SuIMoYs */pOPAdd-WVS ‘G “Winasnyy uespy YNOS ay) ul UoWOods y “g "R] [9qe[s Ss} JO Pole URIpoW payeyul oy} Ul (YSH]) snxeauos “pw soyseosdde yey) yenpraipul ue “goopadd-WWS “WV ‘(4aues) 4affvo snavyd&uoviaw “Gq “q-V ‘Zp ‘Sly DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 61 seemingly fewer axial rings. It is also a somewhat older species. Gamonedaspis accola (Clarke) (1913: 101, pl. 4 (figs 9-20)) is very close to the present species, as noted by Kozlowski (1923). So far as can be judged from Clarke’s perhaps somewhat idealized drawings, it differs from G. boehmi only in the irregular development of the tubercles on the pygidial axis. Since this is perhaps an artefact of preservation, G. accola is likely to become a synonym of Knod’s species. Occurrence Gamonedaspis boehmi (Knod) is currently known with certainty only from Bolivia and the Voorstehoek Formation of South Africa, but may also prove to be present in Brazil. Superfamily ACASTACEA Discussion On the basis of Eldredge’s (1979) phylogenetic analysis of this group of trilobites, the writer is inclined to recognize the following subdivisions: Family Acastidae Subfamily Acastinae Subfamily Acastavinae Family Asteropygidae Family Calmontidae Subfamily Calmoniinae Subfamily Bouleiinae These taxa are united by the form of their auxiliary impression systems which is *.. . a highly organized and distinctly triangular series of two major radiating rows of impressions (with accessory impressions in subsidiary rows), with the midline devoid of impressions, and with the widest portion of the field situated adjacent to the pre-glabellar furrow or facial suture’ (Eldredge & Branisa 1980: 191). Family Calmoniidae Delo, 1935 Discussion The quite remarkable pattern of speciation among the calmoniid trilobites is undoubtedly due to adaptive radiation, causally related to marine transgres- sion (Cooper 1977) and the niché diversity and abundance brought about by the flooding of large areas of Gondwanaland by the late Emsian-early Eifelian eustatic transgression. As such, there is likely to be a very high degree, if not complete endemicity, of the taxa involved. Hence, the writer would agree with Eldredge & Branisa (1980) in regarding the family as wholly Malvinokaffric. ANNALS OF THE SOUTH AFRICAN MUSEUM 62 “Spynow [eusa}UI se paArosaid [[y ‘“¢‘] x ‘ogerpddd -NVS ‘Cd ‘S‘I x ‘sa]oroqny Jerxe posted oy) Surmoys */66€ AAd-WVS “A “91 x “Saqo] ET pur [eo 34} JO UONRIOSI puke SMOLIN Ie[JOqe|ssUeI] 9Y) NION ‘9SOPFAdd-WVS ‘O‘V ‘(pouy) muyaog sidsppauouwy ‘¢p “314 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 63 Such alleged calmoniids as Neocalmonia (Pillet 1969) are better assigned to the family Asteropygidae, while Dalmaniturus is also excluded from the current concept of the family. Hupé (1955) created the monotypic subfamily Bouleiinae, within the Phacopidae, for Bouleia but, as shown by Eldredge (1972), the latter genus is merely a phacopid homoeomorph whose affinities are with the Calmoniidae. As such, the Bouleiinae becomes a subfamily within the Calmoniidae. On hypo- stomal characters and the strong reduction of the L1 lobes, the writer is also inclined to place Parabouleia, Malvinella, and Vogesina in this subfamily. Provisionally, therefore, the following taxa may be included in the sub- family Calmoniinae: the Calmonia plexus comprising Calmonia, Paracalmonia, Metacryphaeus (Metacryphaeus), M. (Wolfartaspis), Tibagya and Punillaspis, the Phacopina plexus of Phacopina, Pennaia, Oosthuizenella and Andinacaste, the Probolops group comprising Probolops, Tarijactinoides, Renniella, Cry- phaeoides and perhaps Schizostylus, the unique Typhloniscus, and the group comprising Bainella, Deltacephalaspis (Deltacephalaspis), D. (Prestalia), Koz- lowskiaspis (Kozlowskiaspis) and K. (Romanops). Genus Oosthuizenella gen. nov. Type species Phacops ocellus Lake, 1904; by original designation herein Etymology For Mr Roy Oosthuizen of Zwartskraal, Prince Albert, who allowed the writer to study his magnificent collection of Bokkeveld trilobites, collected and meticulously prepared over many years, and freely shared his expert knowledge of this group. Diagnosis Cephalon subtrigonal in outline, wider than sagittal length. Genal angles with small spiniform processes. 2p and 3p glabellar furrows subparallel and characteristically sinuous and equally impressed; 2p furrows not reaching axial furrows. 1p furrows very deep, concave anteriorly, and weakly connected to axial furrows. Frontal lobe with pronounced ‘acastid’ auxiliary impression system. Occipital furrow with strongly convex anterior margin and broadest sagittally. Occipital ring smooth, of constant width. Eyes small (about one-third of exsagittal length), situated between abaxial terminations of lp and 3p furrows. Pygidium semicircular, broader than long, with 4 pairs of well- developed pleural spines and a short post-axial spine. Axis rather broad, comprising 8 rings of which the anterior 2 are most pronounced. Pleural fields with 4, perhaps 5, pleurae with deep pleural grooves. ANNALS OF THE SOUTH AFRICAN MUSEUM 64 “Sp[nouwl [euIO\UI se poArosoid l@ “Opp 10J 1dsoxq “€ x “S-AUd ‘(493[eS) saffv2 snanyddaonjaw “q “¢ x “Y99YO OY} JO JoRJANs IeynoyNd oy} SurMoys ‘LO-OU “WUUY sisuapyuins vyjauing “DQ °8°7 x “pooy sypwuaiy (vuidosvyg) sdosvyg se ((Z] BY) 6 “1d ‘vSzI) pooy Aq poinsy “¢6LL-WWS “(24"T) Snaingin vyauiwg “q “87x ‘PE-AUd ‘(pouy) Mya0q sidsypauowmy “Wy “pp “BIg DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 65 Discussion Although Phacops ocellus Lake was originally included by Clarke (1913) in his proposal of the genus Calmonia, the South African material shows certain features that suggest generic separation. In all the South American species of Calmonia the glabella shows transverse 2p furrows that deepen adaxially and are more pronounced than the 3p furrows; there is an anterior border visible dorsally and with a short median frontal process, and there are invariably 6 pairs of pleural spines to the pygidium. In Phacops ocellus, however, the 3p and 2p furrows are equally developed, sinuous and parallel, without adaxial deepening of the 2p furrows, the glabellar frontal lobe overhangs the anterior border, there is no sign of a frontal process, and there are only 4 pairs of pleural spines to the pygidium. Oosthuizenella ocellus (Lake, 1904) Figs 45-49, 50C, 85C Calymene blumenbachii Murchison (non Brongniart), 1839: 654. Phacops (Cryphaeus) africanus Salter, 1856: 218, pl. 25 (figs 6-7, 9 only). Phacops ocellus Lake, 1904: 207, pl. 24 (figs 9-10). Phacops (Cryphaeus) callitris Schwarz, 1906: 393, pl. 10 (fig. 2). Pennaia africana Shand, 1914: 24, pl. 6 (figs 1-14). Swartz 1925: 51. Phacops (Calmonia) ocellus Lake, Reed, 1925a: 153. Phacops (Pennaia) africanoides Reed, 1925a: 158. Phacops (Calmonia) callitris (Schwarz) Reed, 1925a: 149. Acaste (Calmonia) callitris (Schwarz) Rennie, 1930: 344, pl. 9 (fig. 19). Acaste (Calmonia) africana (Shand) Rennie, 1930: 344. Bainella? ocellus (Lake) Struve (in Harrington et al.), 1959: 0484, fig. 383, 2f. Calmonia ocellus (Lake) Eldredge & Ormiston, 1979: 158. Calmonia? callitris (Schwarz) Eldredge & Ormiston, 1979: 158. Material The holotype, by monotypy, of Phacops ocellus Lake in the British Museum, BM-In4957, together with the holotype of Pennaia africana Shand, SU-E417B, the holotype and paratype of Phacops callitris Schwarz, AM-29 and AM-—34 respectively, AM-18 and RO-42, 777-778, E69, E94~-95, E42-45, E61 and E66. All are preserved as internal moulds. Description Cephalon. The cephalon is subtrigonal in outline, convex, and about twice as wide as the sagittal length. The glabella is moderately convex, pentagonal, and with well-developed axial furrows diverging slightly to the anterior. There are weak axial furrow pits. The frontal lobe is rhomboidal, with a pronounced ‘acastid’ auxiliary impression system, and overhangs the anterior border slightly. The 3p and 2p furrows are equally impressed, parallel, distinctly flexuous, and directed posteromedially. The 3p furrow does not reach the axial furrow. Because the 2p furrows do not meet the axial furrows, the L3 and. L2 lobes are fused distally. The 1p furrows are anteriorly concave, with deep apodemal pits, and connect weakly with the axial furrows. The occipital furrow 66 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 45. Oosthuizenella ocellus (Lake). Reconstruction based upon the available material. x 2. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 67 Fig. 46. Oosthuizenella ocellus (Lake). The holotype, BM-—In4957, figured by Salter (1856, pl. 25 (fig. 6)) and Lake (1904, pl. 24 (fig. 9)). x 2. is well developed with a strongly convex anterior border and an almost straight posterior margin, and is thus broadest sagittally. There are deep apodemal pits to the occipital furrow, immediately behind the L1 lobes. The posterior border furrows are moderately deep, shallowing abaxially, while the lateral border furrows are weakly developed and become obsolete anteriorly. The genal angles are subrounded and bear short spiniform processes. The eyes are relatively small about one-third of the exsagittal cephalic length, and situated 68 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 47. Oosthuizenella ocellus (Lake). RO-42, a somewhat crushed but relatively complete internal mould. Note the prominent pygidial lappets and short terminal spine. X 3. 69 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP ‘TT x Yjoqg puke spjnow yeusojul se poalosoid ylog “((Z “3Y) OT ‘Id ‘9061) zIeMYDS stamyjvI Sdoovyg Jo adAjojoy 9y) *67-WV “G-d ‘Z1eMYoS siuyyjv9 sdoovyd Jo adAjered ay ‘po-WV ‘VW ‘(24eT) Snqjas0 vyauazmyisogQ ‘gp ‘S14 yp ay pL faa La a RCO a mamma aT reas tO 70 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 49. Oosthuizenella ocellus (Lake). RO-777. Posterior portion of thorax and anterior portion of pygidium, showing the distinct pleural spines. Preserved as an internal mould. x 2,8. between the distal terminations of the 1p and 3p furrows. They comprise 23-24 vertical files, with 7-8 ocelli in the central files. The anterior branches of the facial suture just transect the lateral extremities of the frontal lobe. Thorax. The thorax is moderately convex, with an axis about one-third the thoracic width. The axial rings are distally spatulate and separated by fairly deep ring furrows with apodemal pits developed abaxially. The axial furrows are moderately deep and distinct. The pleurae are convex, distally with short, posteriorly directed spines, and with pronounced, anteriorly concave pleural grooves that do not reach the distal extremities of the pleurae. The interpleural furrows are moderately developed. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP TAL Fig. 50. A-B. aff. Pennaia sp. AM-25, one of Schwarz’s (1906) syntypes of Phacops gydowi. C. Oosthuizenella ocellus (Lake). SU-E417B, the holotype of Pennaia africana Shand (1914, pl. 6 (figs 1-4)), renamed Phacops (Pennaia) africanoides Reed (1925a). D. Pennaia? sp. AM-—26, selected by Schwarz (1906) as the type for the body of Phacops gydowi Schwarz. E. Typhloniscus baini Salter. SU-K463. F. Bainella africana (Salter). RO-784. All preserved as internal moulds and all x 1. Pygidium. The pygidium is subtrigonal in outline, almost semicircular, convex, and with a pronounced axis. The latter comprises 8 rings, of which only the anterior 2 are separated by well-developed ring furrows. The axis tapers posteriorly for the first 3 rings, after which the axial furrows become subparal- lel, and is bluntly rounded posteriorly. The pleural fields are gently convex and comprise 4, perhaps 5, pleurae, each with deep pleural grooves and faint interpleural furrows. The pleurae terminate in well-developed, short, pos- teriorly directed spines of which there are 4 on either margin. There is a short, pointed post-axial spine. 2 ANNALS OF THE SOUTH AFRICAN MUSEUM Discussion This species differs from all described species of Calmonia in features of — the glabellar segmentation and in possessing only 4 pairs of pleural spines to the pygidium. It is also close to Pennaia pupillus (Lake) but the latter species lacks spiniform processes to the genal angles, has transverse 2p glabellar furrows which deepen adaxially, possesses fewer axial rings and pleurae to the pygidium, with only 3 pairs of minute pleural spines, and lacks a terminal spine. The material assigned by Clarke (1913, pl. 5 (figs 10-15)) to this species differs in possessing deep, transverse 2p furrows and at least 5 pairs of pleural spines to the pygidium. According to N. Eldredge (in litt. 1980), this material is assignable to Oosthuizenella. Occurrence Oosthuizenella ocellus (Lake) is currently known with certainty only from the Gydo Formation, but may be present in the Falkland Islands. Genus Pennaia Clarke, 1913 Type species Pennaia pauliana Clarke, 1913; by original designation Discussion Eldredge & BraniSa (1980) have recently discussed the Malvinokaffric species of Acastoides, noting that they differ from the Boreal species (which include the type of the genus) in having (i) shorter and narrower pygidia, (ii) a prominent axis, bluntly rounded posteriorly and with a distinct post-axial field, (111) a less well-differentiated palpebral complex, and (iv) the posterior borders of the cephalon strongly convex so that the genal angles are situated opposite the posterior margins of the eyes. Moreover, they noted, as had Kozlowski (1923), that in well-preserved material there were small lappets to the anterior three pleurae of the pygidium. In the latter character, ‘Acastoides’ verneuili (d’Orbigny) is obviously allied to Pennaia and was, in fact, assigned to this genus by Reed (1927). It is highly significant, therefore, that the Bolivian ‘Acastoides’ species differ from the type species of Acastoides in having well- developed L1 lobes to the glabella and a hypostoma with the maculae situated on the anterior margin of the middle body, not laterally as in A. henni (Richter). In consequence, the Bolivian “Acastoides’ species are here transfer- red to Pennaia although the strongly convex posterior border of P. verneuili (d’Orbigny), with the concomitant shift in position of the genal angles, the pronounced auxiliary impression system and the anteriorly convex form of the 2p glabellar furrows, may prove to warrant subgeneric separation. As such, Pennaia should be regarded as a synchronic Malvinokaffric homoeomorph of the Boreal Acastoides. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP Ws Pennaia pupillus (Lake, 1904) Figs 6C, 51-55, 56C—G, 58E-F, 59C-D, 99D Phacops pupillus Lake, 1904: 203, pl. 24 (fig. 1). Phacops impressus Lake, 1904: 209, pl. 25 (fig. 1). Phacops (Cryphaeus) gydowi Schwarz, 1906: 396, pl. 10 (fig. 3). Phacops (Calmonia) impressus Lake, Reed, 1925a: 150. Phacops (Calmonia) impressus var. vicina Reed, 1925a: 151, pl. 11 (fig. 8). Phacops (Calmonia?) pupillus Lake, Reed, 1925a: 156. Phacops (Pennaia) gydowi Schwarz, Reed, 1925a: 157, pl. 9 (fig. 7), pl. 11 (fig. 2). Acaste (Pennaia) impressus (Lake) Rennie, 1930: 338, pl. 9 (figs 5-13). Metacryphaeus ? pupillus (Lake) Eldredge & Ormiston, 1979: 160. ? Phacops (Cryphaeus) africanus Salter, 1856: 218, pl.25 (figs 2, 9c, 10, 13 only). Material The holotype, by monotypy, SAM-K60, together with SAM-1167, 2463, 7187, 8959, 8961, 8963, 8966-8967, 8969, 9605, PDB3078, PRV-37, 3200, AM-26, 28, 69, KM-418, BM-—52061, 1In28613-28614, 28616, 28618-28619, IM29236, and RO-4, 213, 273, 704, 758, 762, 774, 781, 791, 797, 831, C70, E26, 28, 78-79, L76, and P2. Description Cephalon. The cephalon is rather convex, subtrigonal in outline and about twice as wide (trans.) as the sagittal length. The glabella is moderately convex, subpentagonal, widening slightly to the anterior and typically with a thomboidal frontal lobe. The auxiliary impression system is poorly developed and only occasionally visible, e.g. PRV-3200. The axial furrows are well defined, subparallel, and with prominent axial furrow pits opposite the distal terminations of the 3p furrows. The latter are moderately impressed, straight to very slightly sinuous, and distinctly connected to the axial furrows. The 2p furrows are deep, transverse, and distally obsolete, so that L2 and L3 coalesce distally. The 1p furrows are very deep, directed slightly posteromedially, and distinctly connected to the axial furrows. There is a slight forward curvature of the adaxial ends of the 1p furrows as well as a weak, but distinct adaxial bifurcation well seen in the lectotype. The L1 lobes broaden abaxially. The cheeks are subtrigonal, moderately convex and down-turned and apparently lacking spiniform processes. The posterior border furrows are deep, trans- verse, becoming shallower distally, whereas the lateral border furrows are shallow and poorly defined. The anterior border is very narrow, and generally obscured dorsally by the frontal lobe. The occipital furrow is rather deep, anteriorly convex, widest sagittally, and with deep apodemal pits distally. The smooth occipital ring is of more or less constant width, and the posterior borders broaden (exsag.) slightly towards the genal angles which are sub- rounded. The eyes are rather large, somewhat more tha one-third of the exsagittal cephalic length. The visual surface comprises 28-29 vertical rows of ocelli, with 7-8 lenses in the middle row. The anterior branch of the facial suture only just transects the lateral extremities of the frontal lobe (fig. 54), 74 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 51. Pennaia pupillus (Lake). A reconstruction based upon the available material. x 2. while the posterior branches meet the lateral margins more or less in line with posterior ends of the eyes. Thorax. The thorax is strongly convex, with an axis about one-third of the thoracic width. The axial rings are distally spatulate and separated by rather deep intersegmental grooves that bear apodemal pits distally. The pleurae are convex, with deep, anteriorly concave pleural grooves that taper distally and become obsolete before reaching the distal extremities of the pleurae. The latter are ornamented with minute, posteriorly directed spines. Pygidium. The pygidium is semicircular to subtrigonal in outline, mod- erately convex, and with an axis composed of 8-f0 rings. The axis tapers uniformly to the posterior and there is no post-axial field. The pleural fields are DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 1D gently convex and comprise 5, possibly 6, pairs of pleurae, each with deep, well-defined pleural grooves that become obsolete before reaching the pygidial border. Distinct interpleural furrows are developed only between the anterior pleurae; posteriorly they are only faintly visible mesially. Like the pleural grooves, the interpleural furrows are distally obsolete, so the pygidium is left with a smooth border. The pygidial margin is ornamented with 3 (? occasionally 4, fide Rennie 1930) pairs of very short, broad, triangular spines that are directed posteriorly and correspond with the anterior 3 pleurae. Behind these spines the pygidial margin is entire. Discussion Lake (1904) considered the holotype of Phacops pupillus to have spiniform processes to the genal angles but, as can clearly be seen from Figure 52, this is an artefact of preservation. The posterior border of the left cheek of the holotype has been broken off for almost its entire length, save for a small, narrow remnant at the genal angle; this gives the erroneous impression of a spiniform process to the genal angle. In creating Phacops impressus, Lake (1904: 209) noted that ‘.. . this form is represented by several nearly complete specimens, but they are so much crushed and distorted that it is impossible to determine accurately the charac- ters of the head’. Since then, however, Rennie (1930) has provided an excellent description of this species, based upon well-preserved topotype material, and it is now clear that P. pupillus and P. impressus were originally based upon different parts of the same animal. Unfortunately, because of page priority, the better known Pennaia impressus (Lake) must become a junior subjective synonym of P. pupillus. A perturbing feature of the Pennaia material from the Bokkeveld succes- sion is that few examples have identical cephala (Fig. 59), but seemingly all have identical thoraxes and pygidia. On the basis of current taxonomic practice within the group, a number of these could perhaps be given new specific names. However, the writer is not yet certain that more than a single, highly variable, polytypic species is present. Thus, PRV-—3100 (Fig. 58E-F) differs from typical P. pupillus in having the posterior borders arched strongly forward so that the genal angles, which are acuminate, are situated opposite the posterior portions of the eyes. This specimen also has relatively somewhat smaller eyes than typical P. pupillus, and with a broader frontal lobe. The glabellar segmentation is also somewhat more deeply impressed and the 2p furrows are not quite transverse. It thus approaches an unnumbered specimen in the South African Museum (Fig. 99D) that differs in having transverse 2p furrows and a relatively broader (trans.), almost elliptical frontal lobe. The latter example approaches a specimen in the British Museum (Fig. 62B) that shows a similar frontal lobe, but now with a distinct auxiliary impression system and similar 2p furrows, but this specimen now has a transverse posterior border, thus resembling typical P. pupillus. From the latter it differs, however, in having a relatively broader, 76 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 52. Pennaia pupillus (Lake). The holotype, by monotypy, SAM—PDB60, which retains cuticle. Note the pseudo-spiniform process formed by the destruction of the posterior border, the distinct posterior branch to the lp furrow, and the strong homoeomorphic resemblance to Phacopidina in lateral view. X 4. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP Wa A Thi | } be = Fig. 53. Pennaia pupillus (Lake). SAM-1167, figured by Reed (1925a, pl. 9 (fig 7)), and preserved as an internal mould. x 2. ellipsoidal frontal lobe with a distinct auxiliary impression system and Ip furrows that do not show adaxial bifurcation. Phacopina lakei (Reed) differs from Pennaia pupillus in its very much broader (trans.), elliptical frontal lobe, with the anterior border distinctly visible in dorsal view. It also has less deeply impressed 2p and 3p furrows, 2p furrows that are anteriorly convex, narrower L1 lobes, and posterior borders that curve weakly forward (Fig. 59A). Until very much bigger collections are available for study, the writer prefers to subscribe most of the above differences to intraspecific and ontogenetic varia- tion. This tends to be supported by their sympatric association and the fact that only one type of thorax and pygidium can, at present, be identified. Pennaia pauliana Clarke (1913: 133, pl. 5 (figs 18—25)) is very similar to the present species but was said to have only 3-4 well-defined axial rings to the pygidium, whereas in P. pupillus there are 8-10. Furthermore, the pleural grooves of the thorax of P. pauliana are not anteriorly concave as in the South African species. Although the differences are slight, the author prefers to maintain P. pauliana as distinct until it is revised on the basis of type and topotype material. 78 Fig. the ANNALS OF THE SOUTH AFRICAN MUSEUM 54. Pennaia pupillus (Lake). SAM-—8969a, preserved as an internal mould. Note how anterior ramus of the facial suture transects the extreme lateral margin of the frontal lobe. X 4,1. Fig. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 55. Pennaia cf. pupillus (Lake). SAM-8961, preserved as an internal mould. Note the oval shape to the frontal lobe. x 4. 79 80 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 56. A-B. Metacryphaeus caffer (Salter). A. SAM-—PDB4345. x 1,5. Note the length and shape of the pleural lappets to the pygidium. B. SAM-—PDB7787. x1. C-G. Pennaia pupillus (Lake). C-D. SAM-—PDB3078. x 1. E-F. AM-28, the holotype of Phacops gydowi Schwarz (1906, pl. 10 (fig. 3)). x1. G. KM-418, the original of Reed (1925a, pl. 11 (fig. 8)). x 1. H. Burmeisteria herscheli (Murchison). A specimen in the South African Museum retaining spines to the thoracic axial ring. X 2. All specimens preserved as internal moulds. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 81 Fig. 57. A. Bainella africana (Salter). SAM-PDB3080, the internal mould of an hypostoma. x4. B. Pennaia sp. SAM-11915, the posterior portion of the thorax and terminally eroded pygidium, preserved as an internal mould. x 4. 82 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 58. A-D. Bainella africana (Salter). A. RO-30. x2. B-—C. SAM-—782, the original of Lake (1904, pl. 24 (fig. 6)). x 2. D. SAM-PDB3945. x 2. E-F. Pennaia aff. pupillus (Lake). PRV-3100. Note the strongly curved posterior border and acuminate genal angles. X 2,7. ‘€ x “xoidde ]je pue spnow jeurdjur WOT} PINASUOSAI [TW “OOLE-AUd uodn paseq ‘(axe T) snyidnd -jye vivuuag ‘q ‘(ayeq) snypidnd vwuuag ‘Dd *((OT “8Y) SZ ‘Id ‘oggq) saqes a JO [eursi1o ay ‘ds gvuidoopyg *g ‘(paay) 12yn] vundosvyg “Ww ‘eyeydao puidospyd pure vivuuag pojon14su0de1 uIOS JO uostiedulog ‘6S “314 oe) DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP RW 84 ANNALS OF THE SOUTH AFRICAN MUSEUM ‘Acastoides’ koukharskii Baldis & Longobucco (1977b: 171, figs 1-3) is neither an Acastoides nor a Pennaia. It differs from both these genera in having 2p furrows that connect with the axial furrows and L2 lobes that narrow (exsag.) abaxially. Pennaia gamonedensis (Eldredge & BraniSa) (1980: 237, fig. 19) differs from P. pupillus in having smaller eyes, lightly impressed 2p and 3p furrows, the latter almost transverse, and with a pygidium with only 2 distinct axial rings and 4 pleural grooves. Occurrence Pennaia pupillus (Lake) is known with certainty only from the Gydo Formation, although there is a single pygidium, PRV-39, in green shale preservation that may be from the Voorstehoek Formation. Genus Metacryphaeus Reed, 1907 Subgenus Metacryphaeus Reed, 1907 Type species Phacops caffer Salter, 1856; by the subsequent designation of Rennie, 1930 Discussion Wolfart (1968) recognized two species groups within Metacryphaeus; the group of M. caffer was said to comprise M. caffer (Salter), M. australis (Clarke), M. praecursor Wolfart, and M. venustus Wolfart, and characterized by a comparatively small exoskeleton, relatively large eyes and a limited capacity for spine formation. The group of M. giganteus was said to comprise M. giganteus (Ulrich), M. convexus (Ulrich), M. dereimsi (Groth), M. boulei (Kozlowski), M. cornutus Wolfart, and M. tuberculatus Kozlowski, and typified by the relatively larger size of the exoskeleton, relatively smaller eyes, and a marked capacity for spine formation. The writer finds this division of little practical use since M. tuberculatus is a probable synonym of M. caffer, while M. caffer is very closely related to, and probably descended from M. giganteus, as well as being of a comparable size. It is pertinent to note, however, that Eldredge & BraniSa (1980) report the presence of two distinct types of hypostomata among Metacryphaeus species. Among the large number of Metacryphaeus species now known, one species stands alone. This is M. cornutus Wolfart (1968: 95, pl. 12 (figs 2-5), pl. 13 (figs 1-4), pl. 14 (figs 1-2) (Fig. 60 herein), which can be distinguished from all other described species of Metacryphaeus in the possession of a cylindrical terminal spine to the pygidium and a thick, cylindrical median spine to the occipital ring. It is accordingly made the type species of the new subgenus M. (Wolfartaspis). DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 85 Fig. 60. Metacryphaeus (Wolfartaspis) cornutus Wolfatt. A reconstruction based upon the figured material. Approx. x 1,5. Metacryphaeus (Metacryphaeus) caffer (Salter, 1856) Figs 6D, 33A, C, 34B, 39F, 42A-B, D, 44D, 56A-B, 61, 62A, 63-64, 65C-D, 66A, 67A, 68B, 69-70, 78D, 98A, C, 99C Phacops caffer Salter, 1856: 219, pl. 25 (figs 11-12 only). Phacops (Cryphaeus) africanus Salter, 1856: 218, pl. 25 (fig. 4 only). Phacops (Cryphaeus) caffer Salter, Lake, 1904: 210, pl. 25 (figs 3-4). Schwarz, 1906: 395. Clarke, 1913: 72. Dalmanites (Cryphaeus) caffer (Salter) Reed, 1925a: 139, pl. 9 (figs 10-11), pl. 11 (fig. 4). Dalmanites (Cryphaeus) caffer var. albana Reed, 1925a: 144, pl. 11 (fig. 1). Dalmanites (Cryphaeus) cf. australis (Clarke) Reed, 1925a: 146, pl. 11 (fig. 9). 86 ANNALS OF THE SOUTH AFRICAN MUSEUM Dalmanites (Cryphaeus) cf. pentlandi (Salter) Reed, 1925a: 146, pl. 11 (fig. 10). Dalmanites (Cryphaeus) caffer (Salter) var., Reed, 1925a: 145. Dalmanites (Corycephalus ?) capensis Reed, 1925a: 131, pl. 8 (figs 5-6). Dalmanites (Eocorycephalus) capensis Reed, 1925b: 133. Asteropyge caffer (Salter) Reed, 1927: 310. Dalmanites (Metacryphaeus) caffer (Salter) Rennie, 1930: 336, pl. 9 (figs 3-4). Metacryphaeus caffer (Salter) Reed, 1907: 168. Struve (in Harrington et al.), 1959: 0484, fig. 383, 3c. Wolfart, 1968: 81. Eldredge & Ormiston, 1979: 159. ? Dalmanites (Cryphaeus ?) cf. rostratus Kozlowski, Reed, 1925a: 147, pl. 8 (fig. 2). ? Dalmanites (Proboloides) ensifer Reed, 1925a: 136, pl. 8 (fig. 4). ? Schizostylus ensifer (Reed) Wolfart, 1968: 113. ? Schizostylus (Curuyella) ensifer (Reed) Eldredge & Ormiston, 1979: 159. Eldredge & Brani§a, 1980: 228. Fig. 61. Metacryphaeus (Metacyphaeus) caffer (Salter). A reconstruction based upon the available material. Approx. X 1,5. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 87 Material Numerous specimens including the lectotype, BM—In28620 (Fig. 62A), and BM-28783, BM-Itl1619, ?SMC-A3063, SAM-PDB2463, 3088, 3850, 3943, 4000, 4041, 4047, 4068, 4075, 4078, 4329, 4335, 4340-4341, 4346-4347, 4350-4352, 4354, 4363, 7184-7185, 7788-7790, 7792, 7799, 11900, PRV-5, 922, 1028, 1070, AM-74, SU-C2, 13, SU-H73, 76, 172-3, SU-E414, 483, together with RO-68, 207, 241, 285, 703, 820-828, 833, E60, 77, 93, and 96. Description Cephalon. The cephalon is moderately convex, semicircular to semi- elliptical in outline, and more than twice as wide as the sagittal length. The glabella expands evenly towards the anterior with deep, straight axial furrows. The inflated frontal lobe is coarsely pitted by internal muscle scars, and there is a well-developed posterior median impression. The 3p furrows are strongly developed, directed posteromedially at about 60° to the axial furrows and distinctly connected to the latter. The 2p furrows are transverse, deepest adaxially, and only very weakly connected to the axial furrows. The 1p furrows are normal to the axial furrows and thus directed slightly anterome- dially, with the result that the L1 lobes tend to have a uniform width (exsag.). They have deep apodemal pits adaxially and are slightly more strongly con- nected to the axial furrows than the 2p furrows. The occipital furrow is rather deep, with abaxial apodemal pits, and a moderately convex anterior margin but an almost transverse posterior border, with the result that it is broadest sagittally. The occipital ring is of uniform width, straight, transverse, and wider (sag.) than the occipital furrow. The posterior border furrows are rather broad (exsag.) and deep, and meet the moderately developed lateral border furrows in acute angles at the genal angles which bear short spiniform processes. The anterior border furrow is very shallow and weakly developed but there is a short but distinct uptilted median frontal process. The cheeks are subtrigonal in outline and moderately convex. The eyes are moderately large, about one-third of the exsagittal length, and comprise 23-27 vertical files of lenses, with 8-9 lenses in the central file. The eyes are strongly arcuate, with distinct palpebral furrows, and touch the axial furrows anteriorly immediately behind the abaxial terminations of the 3p furrows. The posterior borders of the eyes lie in a plane slightly anterior to the distal terminations of the 1p furrows. Thorax. The thorax is moderately convex, with an axis about one-third the thoracic width. The axial rings are distally spatulate and separated by mod- erately wide, deep, ring furrows with abaxial apodemal pits. The axial furrows are deep and well defined. The pleural fields are moderately convex, with pronounced fulcra at about the middle of their transverse widths. The promi- nent deep pleural grooves are almost straight, tapering distally and becoming obsolete before the abaxial terminations of the pleurae are reached. The latter 88 ANNALS OF THE SOUTH AFRICAN MUSEUM bear small, posteriorly directed spines. The interpleural furrows are well developed, narrower than the pleural grooves, deepest adaxially, and show distinct geniculations at the fulcra. Pygidium. The pygidium is almost semicircular in outline, distinctly wider than the sagittal length, and moderately convex. The axis is broad anteriorly and tapers rather rapidly to the posterior for the anterior 5 rings, whereafter the moderately deep axial furrows become subparallel. There are about 10 axial rings, of which the last few are rather indistinct, separated by rather broad (sag.) ring furrows. The anterior 34 rings retain spatulate distal terminations, while the anterior 4-5 ring furrows have prominent apodemal pits distally. The axis terminates in a bluntly rounded tip inside the pygidial margin so as to leave a short but distinct post-axial field. The pleural fields are subtrigonal, gently convex, each comprising 5 distinct pleu- rae that extend beyond the pygidial border to form prominent posteriorly directed lappets. The pleural grooves are broad and rather deep, tapering abaxially to become obsolete before the distal terminations of the pleurae are reached. The interpleural grooves are fine, tending to broaden and become more pronounced distally. There is an incompletely developed sixth pair of pleurae which does not develop lappets. The posterior margin of the pygidium is evenly rounded to slightly pointed between the fifth pair of lappets and is arched sagittally. Discussion In the interpretation of this species, and all the material discussed herein, the writer has been strongly governed by the consideration that a species is typified by the statistics of a population of individuals, not solely the characters of the holotype. Consequently, many of the trivial differences of previous workers are not herein considered of specific importance. As noted by Rennie (1930), Dalmanites (Eocorycephalus) capensis (Fig. 33A) Reed was based upon damaged cephala of Metacryphaeus caffer of which it is, therefore, a synonym. Both Reed’s (1925a) var. albana, said to have a more subquadrate glabella and to lack a median frontal process, and his Dalmanites (Cryphaeus) caffer var., based upon a pygidium distinguished by the more rapidly tapering axis and broader (sag.) post-axial field, are based upon characters of the individual that are not herein considered of taxonomic importance. Dalmanites (Proboloides) ensifer Reed (Figs 33C, 64H) was based upon a unique cephalon that subsequently has generally been assigned to the genus Schizostylus. The writer has had the opportunity of studying this specimen; there seems little justification for interpreting the median frontal process as long as did Reed (1925a), and this is certainly a species of Metacryphaeus. It differs from M. caffer mainly in its state of preservation: it represents nodule preservation whereas most of the M. caffer material is preserved as dorsoven- DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 89 trally compressed internal moulds. It is only because of the different modes of preservation, reflecting perhaps different stratigraphic levels, that the writer is somewhat hesitant unequivocally to include Dalmanites ensifer in the synonymy of Salter’s species. The Dalmanites (Cryphaeus) cf. rostratus Kozlowski figured by Reed (1925a) (Fig. 70 herein) would appear to represent nothing more than a well-preserved frontal process such as might have been found on the holotype of D. (Proboloides) ensifer. Perhaps significantly, however, it is associated with Bainella and hence may represent a lower stratigraphic horizon than typical M. caffer. The true Dalmanites rostratus Kozlowski (1923: 44, pl. 5 (figs 1-9)) is the type species of the genus Cryphaeoides Delo (1935) and bears only a very superficial resemblance to M. caffer. Dalmanites (Cryphaeus) cf. pentlandi Salter (Reed 1925a, pl. 11 (fig. 10)) (Fig. 67A herein) was said to differ from M. caffer in being rather coarsely and irregularly tuberculate. It seems probable, on account of the large size of this individual (sagittal length 71 mm excluding the cephalon), that the slight differences are due to a comparison of different ontogenetic stages. Reed’s (1925a: 145, pl. 11 (fig. 9)) Dalmanites (Cryphaeus) cf. australis Clarke was distinguished from M. caffer by its smaller eyes placed further to the anterior, its less inflated glabella and its more pointed genal angles. The writer has studies this specimen and is of the opinion that it is an individual of M. caffer. However, M. australis (Clarke) (1913: 110, pl. 4 (figs 1-5)) is indeed very close to M. caffer and, to judge from Clarke’s perhaps idealized drawings, differs mainly in its relatively larger cheeks and the shape of the genal angles. It is noteworthy, however, that Clarke (1913: 114) wrote ‘ ... the writer finds figure 4 of Salter’s Cryphaeus africanus a true Cryphaeus tail and identical with C. australis; he believes that Salter’s figure 1 belonged to the same animal and is Cryphaeus australis... . While the Phacops africanus rests well enough on Mr. Lake’s type, let it be understood that Cryphaeus australis is present in the Bokkeveld fauna.’ Clearly the taxonomic status of M. australis must await the modern revision of Clarke’s material. Metacryphaeus boulei (Kozlowski) (1923: 46, pl. 2 (fig. 15); Wolfart 1968: 88, pl. 11 (figs 1-3)) is a coarsely tuberculate species bearing a homoeo- morphic resemblance to Bainella arbuteus (Lake), which is easily distinguished from M. caffer. M. cornutus Wolfart (1968: 95, pl. 12 (figs 2-5), pl. 13 (figs 1-4), pl. 14 (figs 1-2)) is readily separable from M. caffer in the possession of a long, cylindrical median spine to the occipital ring and a terminal spine to the pygidium. Metacryphaeus venustus Wolfart (1968: 99, pl. 15 (figs 2-4), pl. 16, pl. 17 (figs 1-6)) is very close to M. caffer and perhaps especially to M. australis. It seems to differ from the South African species in its relatively longer (sag.) cephalon, with more strongly diverging axial furrows to the glabella, and its somewhat larger eyes that are not quite as close to the axial furrows as in M. caffer. 90 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 62. A. Metacryphaeus caffer (Salter). The lectotype, BM—In28620, figured by Salter (1856, pl. 25 (fig. 11)). x 1,5. B. Phacopina? sp. BM-In28619, the orig- inal of Salter (1856, pl. 25 (fig. 10)), tentatively referred to Phacopina lakei (Reed) by Reed (1925a). Note the oval frontal lobe, distinct auxiliary impression system and relatively large eyes. X 3. Both preserved as internal moulds. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP wil Fig. 63. Metacryphaeus caffer (Salter). An internal mould in the collection of R. Dick, Cape Town. x 1. Metacryphaeus convexus (Ulrich) (1893: 16, pl. 1 (figs 9-11)) was based upon two isolated cephala and a pygidium drawn with an entire margin. As figured by Groth (1912, pl. 18 (fig. 4), pl. 19 (fig. 1)) and Wolfart (1968: 94, pl. 12 (fig. 1)), however, this species differs from M. caffer in having the sagittal area of the glabella, between the adaxial terminations of the glabellar furrows, strongly inflated and raised, with a tendency for the 2p furrows to be deepest abaxially and distinctly connected with the axial furrows. Metacryphaeus dereimsi (Groth) (1912: 606, pl. 18 (figs 1-3)) was based upon isolated pygidia association with M. convexus. The post-axial field seems rather more triangular than is typical for M. caffer and thus this species closely approaches M. venustus. However, since most Metacryphaeus species have closely comparable pygidia, this name is probably best regarded as a nomen dubium. Metacryphaeus praecursor Wolfart (1968: 82, pl. 8 (fig. 7)) was based upon an isolated cephalon which differs from M. caffer in having the 2p and 3p 92 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 64. Metacryphaeus caffer (Salter). A. SAM-—PDB4335. B. SAM-PDB4352. C. SAM-PDB4340. D. SAM-PDB4075. E. SAM-PDB4363. F. SAM-PDB4354, in lateral view. G. SAM—PDB4329. H. SU-C2, the holotype of Dalmanites (Probo- loides) ensifer Reed in lateral view. All preserved as internal moulds and all x 1,3. 93 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP ‘spynowr [euroqur se paarosord [PV “O°T x “(OL SY) 6 ‘Id “YSZHT S.P2e¥ JO [BUISIIO OY) ‘OGLL-NVWS “A ‘FI X “LEOPAdd-WVS ‘O ‘(1oyes) 4affv2 snapyddsovjay “G-D “61 X “aANyNS [RIOR] oy} JO snes JOLID]UR dy} WoIy payesedas [Jam st ey} 9qo] [RIVOLI a1R[NDIAqN} IY) AION “(p9o2y) Ming vjyjauing jo sadAyuds pansy ((¢ “3y) 6 ‘Id ‘vSZI) S,peeY JO 9UO ‘O8//-WYS ‘(ZIeMYOS) sasad sidsp1ysmojz0y ‘ ‘pod “TX “winasnyy uKoY YINog ay} ur ‘((¢ “B3Yy) pZ “Id ‘pO6L) BAeT JO [euIsiIO oY], “(preMpooM) YjV8vISIUD DYaUuIDg “YW ‘Sg “314 94 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 66. A. Metacryphaeus caffer (Salter). AM-—69, plaster cast of an external mould in nodule preservation. Note the weak ocular ridge. B. Bainella africana (Salter). Latex cast of the holotype of Phacops acacia Schwarz, AM-—2124, preserved as an external mould, and figured by Schwarz (1906, pl. 10 (fig. 4)) and Rennie (1930, pl. 10 (fig. 5)). Note how both these workers misinterpreted the nature of the genal spines, which are situated inside (adaxial) of the genal angles and were almost certainly recurved distally. Both X 1,7. 95 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP ‘ST X ‘Spfnoul jeusojut SP paalosaid (7 BY) 6 ‘Id ‘yCZH] sty Jo speuIsIUO ay) pu ‘(podxyz) luipq vyjauwg jo satias adAjudAs datsuajxo s.paay jo yaed BurmM0j “Sp[NOwW [BUIIIUL se poatosoid ‘vyeydao yo dnois ve ‘QOZL-WVS ‘(491[eS) UDO “Jo pyauwg °{ “ST x ‘layes ipuvjjuad “yo (snavyd&up) sajuvuyog se ((OT “3y) [1 ‘Id ‘ysz6t) peoy Aq painsy pyjnow jeusa}xo 241 “€9P7-NVS ‘(4aIeS) daffpa snanyddisvjaw “vy °L9 “314 96 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 68. A. Bainella africana (Salter). RO-708, showing the visual surface of the eye. Much enlarged. B. Metacryphaeus caffer (Salter). SAM-— PDB4341. X 1,8. Both preserved as internal moulds. DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP ay Fig. 70. Aff. Metacryphaeus caffer (Salter). SMC-A3063, the ros- tral fragment figured by Reed (1925a, pl. 8 (fig. 2)) as D. (Cry- phaeus) cf. rostratus Kozlowski, associated with a cephalon of Bainella ‘baini’ (Reed). Both preserved as internal moulds. x 2. 98 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 71. Metacryphaeus giganteus (Ulrich). A reconstruction based upon the figured material of Wolfart (1968). x 1,8. glabellar lobes rather inflated, with the L1 lobes narrow and reduced. M. curvioculatus (Wolfart) (1968: 76, pl. 6 (fig. 2)) is very close to M. venustus and differs from M. caffer in much the same respects as the latter species. Metacryphaeus giganteus (Ulrich) (1892: 14, pl. 1 (figs 6-8)) (Fig. 71 herein) is very close to M. caffer, at least some of the differences relating to the nodule preservation of the Bolivian material. However, it may be distinguished from the South African species by its more rounded frontal lobe and in the development of seemingly true genal spines. Its occipital furrow also appears to be relatively broader (sag.). DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 99 Fig. 72. Punillaspis argentina (Baldis). A reconstruction based upon the figured material. x 2. Metacryphaeus tuberculatus (Kozlowski) (1923: 43, pl. 3 (figs 20-21)) was originally described as a variety of M. australis (Clarke) but elevated to specific status by Wolfart (1968). As figured by Kozlowski (1923), the cephalon is indistinguishable from M. caffer, while the pygidium merely differs in its slightly more triangular outline (an artefact of nodule preservation?). The cephala figured by Wolfart (1968: 102, pl. 17 (fig. 7), pl. 18 (fig. 1)) cannot satisfactorily be separated from M. caffer and should probably be regarded as conspecific. However, since N. Eldredge of the American Museum of Natural History is currently revising the Bolivian calmoniids, the taxonomic position of this species should await his conclusions. Metacryphaeus argentina Baldis (1967: 792, pl. 1 (figs 7-11), pl. 2 (figs 1-6)) (Fig. 72 herein) is a curious species whose very reduced, minute eyes readily distinguish it from M. caffer and, indeed, all other species of Metacry- 100 ANNALS OF THE SOUTH AFRICAN MUSEUM phaeus. It was recently (Baldis & Longobucco 1977a) made the type of the new genus Punillaspis. Occurrence Metacryphaeus caffer (Salter) is abundant in the Voorstehoek Formation of the Bokkeveld Group, and the probably conspecific M. tuberculatus (Kozlow- ski) ranges from the upper part of the Belen Formation into the Sicasica Formation of Eifelian age in Bolivia. There is definitely a Metacryphaeus species present in the Gydo Formation but none of the material available to the writer has been sufficient to identify M. caffer positively; it could be M. gigan- teus (Ulrich) to judge by the stratigraphic relationships of the various species of Metacryphaeus in Bolivia. Metacryphaeus (Metacryphaeus) venustus Wolfart, 1968 Figs 73-75, 76A, C-D, 77C-D, F, 78B-—C Cryphaeus australis Clarke, Kozlowski, 1923: 41, pl. 3 (figs 3-6, 8, 14-17). Reed, 1925a: 139 (pars). Fricke et al., 1964: 15-16. Metacryphaeus australis (Clarke) Branisa, 1960: 68, 74. Metacryphaeus caffer (Reed) Braniga, 1960: 66, 71; 1965: 106, pl. 21 (figs 10-15). Fricke et al., 1964: 16. Metacryphaeus venustus Wolfart, 1968: 99, pl. 15 (figs 2-4), pl. 16, pl. 17 (figs 1-6). Eldredge & Ormiston, 1979: 160. Material Fourteen specimens, PRV-—716, 718-719, 721, 724, 729, 731-734, 777, 780, 791, and 796, all preserved as internal moulds, and all from the Waboomberg Formation at Klein Tafelberg. Description Cephalon. The cephalon is semicircular, much broader (trans.) than long (sag.), with the cephalon diverging strongly to the anterior. The frontal lobe is subrhomboidal, with a prominent posterior median impression and no distinct auxiliary impression system. The 3p furrows are deep, directed postero- medially, and distinctly connected to the well-developed axial furrows. The 2p furrows are deep, transverse, and connected to the axial furrows, while the 1p furrows are very deep, anteriorly concave, and also distinctly connected to the axial furrows. The occipital furrow is arched forward sagittally and with deep apodemal pits distally. The L1 lobes are very narrow (exsag.), anteriorly concave, and in some individuals give the impression of being weakly con- nected across the glabella (Fig. 76A, C). The L2 lobes are more or less of constant width, and the L3 lobes broaden (exsag.) distally. The median glabellar region, between the adaxial terminations of the L2 and L3 lobes, seems to be slightly raised in some individuals (Fig. 77C). The occipital furrow is arched forward sagittally, and with deep apodemal pits distally. The occipital ring is of constant width (exsag.) and lacks tuberculation. There is a narrow DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 101 Fig. 73. Metacryphaeus venustus Wolfart. A reconstruction based upon the material from the Waboomberg Formation. x 2. anterior border that, in some of the better preserved cephala (Fig. 76A), shows the hint of a very short, median rostrum. The cheeks are subtrigonal and curved strongly downward distally. The posterior border furrows are deep, transverse, whereas the lateral border furrows are shallow and indistinct. The posterior border broadens (exsag.) rather markedly towards the genal angles. In PRV-796 (Fig. 75), the genal angles are produced posteriorly, with a rather distinct spinule. The eyes are large, about 40 per cent of the exsagittal cephalic length, and comprise subvertical rows of ocelli, with 6-7 lenses in the middle row. In PVR-796 only about 18 rows of ocelli could be counted. The anterior branch of the facial suture is not clearly visible in any of the available material; presumably it was in close juxtaposition with the glabellar frontal lobe. The posterior suture curves strongly forward after leaving the eye, and then recurves to meet the lateral border approximately opposite the posterior margin of the eye. 102 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 74. Metacryphaeus venustus Wolfart. PRV-791, preserved as an internal mould. x 2. Thorax. There is little to separate the thorax of Metacryphaeus venustus from that of M. caffer, and hence it does not merit redescription. Pygidium. The pygidium is wider (trans.) than long (sag.), with a convex axis that tapers strongly to the posterior for the first 5 rings, whereafter the distinct axial furrows become subparallel. After the eighth axial ring the segmentation becomes indistinct but there may have been as many as 11 rings; certainly in some individuals 10 can be counted. The anterior 4 ring furrows have apodemal pits distally and the axis is bluntly terminated to the posterior, leaving a very pronounced post-axial field. The pleural fields are subtrigonal DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP 103 Fig. 75. Metacryphaeus venustus Wolfart. PRV-796, preserved as an internal mould. Note the very reduced L1 lobes, large eyes, and distinct spines to the genal angles. x 3. 104 ANNALS OF THE SOUTH AFRICAN MUSEUM and rather strongly convex. There are 6 distinct pleurae, each with a pro- nounced pleural groove and faint interpleural furrows, with a very reduced seventh to the posterior. The anterior 5 pairs of pleurae terminate in prominent lappets distally. In PRV—724 there are weak, oblique tubercles, on the anterior moiety of each pleura (Fig. 78B), marking the fulcrum. This feature is not, however, persistent and its taxonomic significance is uncertain. The post-axial field is relatively rather broad, about 22 per cent of the sagittal length of the pygidium, and appears to be abruptly truncate to the posterior (Fig. 78C). Discussion The Bokkeveld material shows a few minor points of differences when compared with the Bolivian material. These include apparently fewer lenses to the eyes and the presence in at least one individual of weak tubercles marking the fulcral axis of the pygidium. In view of the remaining marked similarities between the two forms, however, the differences are not at present considered of taxonomic importance. Differences from the other species of Metacryphaeus are discussed under M. caffer. Occurrence Metacryphaeus venustus Wolfart is currently known only from the upper part of the Belen Formation and the Sicasica Formation of Bolivia, and the Waboomberg Formation of South Africa. Metacryphaeus (Metacryphaeus ?) cf. koukharskii (Baldis & Longobucco, 1977) Figs 76B, 77A-B, 79 Compare Acastoides koukharskii Baldis & Longobucco, 1977b: 171, figs 1-3. Material Two specimens PRV-175 and 428, both preserved as internal moulds, the latter from the farm Bon Chretien, 15 km due east of Ceres, south-western Cape. Description Cephalon. The cephalon is subtrigonal in outline and rather convex, with strongly downturned cheeks. The axial furrows are pronounced and diverge anteriorly. The glabella is convex, with a somewhat inflated frontal lobe that is oval in shape and without a distinct auxiliary impression system. The 3p furrows are weakly impressed, straight, and directed anterolaterally to connect with the axial furrows. The 2p furrows are moderately impressed, transverse, anteriorly convex, and distally connected with the axial furrows. The 1p furrows are very deep, transverse, and strongly connected to the axial furrows. 105 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP “sp[now [eusd}UT se paarasaid [py °Z x “wmuprsdd peordAy & *76—-A Ud ‘(491[eG) PUotfo vYJaUIDY “| “ZX “SLI-AUd ‘(OsonqosuoT 7 sipjeg) IL{SADYYNOY “JO snavyddsovjaw “| *7X “LLL “Aud “7X ‘TeL-AUd ‘DO 'S‘€ X “MOLI IE[Poqe|sueN jounsIp ev Aq paysouuos Ayjeurxoid ose smosiny dy oy) MOY 210N “O8/-AUd “W “MWeOA snisnuaa snavyddiovjaew ‘G-D ‘V '9L ‘314 106 ANNALS OF THE SOUTH AFRICAN MUSEUM 5, oe Fig. 77. A-B. Metacryphaeus cf. koukharskii (Baldis & Longobucco). PRV-—428. Note the abaxial taper of the L2 lobes and the connection of the 2p furrows with the axial furrows. C-D, F. Metacryphaeus venustus Wolfart. C. PRV-796. Note the very reduced L1 lobes. D. PRV-729. F. PRV-733. E. Bainella africana (Salter). PRV-—57. All preserved as internal moulds and all x 2. 107 DEVONIAN TRILOBITA FROM THE BOKKEVELD GROUP ‘parejs dSIMIOYIO ssoyuN Z x [[B pue Sp_NOU [eUIa}UI se paArosoid ITV ‘wojsds uolssoidun Areyixne podojaasp-[[am eB SUIMOYS ‘7Z76-AUd ‘(4aIeS) /affva snavydduonjaw “q “¢ x ‘sapreqny [eropny Suryory sjduexs uy “gT/-AUd ‘DO ‘SEXe [esO[NJ ay) 3 saposoqmy anbiygo ‘a eBuoja oy) SION ‘pZL-AUd ‘GF “URIOAA —AUd ‘(191[8S) vuvoufo vyjauig “VW SL ‘S14 snisnuaa snavyddsovjay ‘Q-q ‘(peey) jung, ‘g Jo sadAquds ayl Jo auo ‘/6T¢ 108 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 79. Metacryphaeus cf. koukharskii (Baldis & Longobucco). A reconstruction based upon PRV—428.