Transactions of rhe Royal Society of Edinburgh, 76, 339-358, J985
CyrtoctefJUS wittebergensis sp. nov. (Chelicerata: Eurypterida), a large sweep-feeder from the Carboniferous of South Africa C. D. Waterston, B. W. Oelofsen, R. D. F. Oosthuizen
ABSTRACT: Cyrtoctenus winebergensis sp. nov. is described from a unique holotype from the Witteberg Group of the Cape Supergroup. It is a giant hibbertopteroid eurypterid having combs and specialised movable spines of crytoctenid type (St0rmer & Waterston 1968) on the more distal podomeres of the second to fourth prosoroal appendages. The function of the combs and their associated movable spines is discussed and it is suggested that together they formed a unique adaptation of eurypterid structures to sweep filter~feeding, the combs forming the filters and the spines the cleaners. The digestive tract is remarkably preserved and shows a spiral valve, posterior to tbe stomach, which is interpreted as an adaptive feature in this large arthropod to increase tbe absorptive area of the gut. The new evidence provided by the South African specimen bas required the re~interpretation of the disarticulated Cyrloclenus specimens previously described from Europe. Disjecla membra recently obtained from the Tournaisian of Foulden, Berwickshire, whicb may belong to CynoClenus, are described and show characters previously unknown in Scottish material but similar to certain features in the South African specimen. The taxonomic relationships within the Hibbertopteroidea are discussed in the light of tbe new combination of characters found in C. wittebergensis. Two families are recognised in the superfamily, the Hibbertopteridae, including Hibber/opterus and Campylocephalus, and the new family Cyrtoctenidae which is here erected to include Cyrtoclenus, Dunsoprerus and possibly also Haslimima.
KEY WORDS: Berwickshire, Gondwanaland, Hibbertopteroidea, intestine, sweep~feeding, taxonomy, Witte berg Group.
Our knowledge of the Eurypterida is derived, for the roost 1.2. Holotype part, from specimens found in North America, Europe and The species is known from a singie specimen (University of Asia. Discoveries of eurypterids in tbe southern hemispbere Stellenbosch, Zoology Department USS.IT.01) found at have been rare and usually fragmentary, often occurring as longitude 22"36' latitude 33°10', 6 km NE of Zwartskraal isolated examples (Kjellesvig-Waering 1961, p. 108-11). The which is 18 km N of Klaarstroom in tbe District of Prince finding of a large and almost complete' specimen in the Cape Albert, Cape Province, South Africa. The locality lies in the Province of South Africa is therefore notable. That it should southern Karoo some 390 km ENE of Cape Town and 60 k..m also provide a key to our understanding of the puzzling NE of Oudtshoorn. The specimen came from the lower part. European genus Cyrloctenus (St0Tmer & Waterston 1968) of the Waaipoort Formation, tbe uppermost subdivision of adds greatly to the interest of the find. the Witteberg Group of the Cape Supergroup (see 3.2). It is preserved in a nodule which has been broken open and now exists as a number of pieces. The specimen is displayed as 1. Description internal and external moulds. Because much of tbe gut is Cynoc/enus St~rmer & Waterston 1968 preserved (see 2.3) it is known that the specimen is a Revised diagnosis. Cyrtocterud ewypterid (see 5.1) in which cadaver and, from its position in tbe beds when found, it is the anterior movable spines of the more distal podomeres of known to have beeD preserved lying on its back. Because of the secood, third and fourth prosomal appendages are decay or predation, some of the ventral structures have been modified to form 'fingers' each associated with a fixed comb lost or have been slightly displaced. Nothing of the specimen probably developed as a modification of the posterior spine. was preserved beyond the nodule.
CynOClenus wittebergensis sp. nov. 1.3. Prosoma 1.1. Diagnosis The proportions of the prosoma are seen in the scale Cynoctenus having an oval oceUar node, and large-scale drawing Figure 1. The sbape is generally semi-circular with a crenulations 00 the distal margins of the podomeres. broad antero~median protrusion. The maximum wldth Ornament generally coarse V-shaped or rounded lunules. (485 mm) lies near the posterior margin. The postero-Iateral The axial portions of the tergites and post-abdominal angles are produced as rounded ears. The prosomallength is segments ornamented with coarse longitudinal ridges and 355 rom . In life the carapace must have been strongly furrows . elevated since in its present state the surface of the cardiac L
340 C. D . WATERSTON. B. W . OELOFSEN AND R . D. F. OOSTHUIZEN lobe lies 138 mm above the level of the postero-Iateral ears. The compound eyes are raised on palpebral lobes and lie just posterior to the mid-line with their axes converging anteriorly. The paired ocelli lie in the posterior half of (he oval ocellar node which occurs betlVeen the posterior portion of the lateral eyes. Anterior to each lateral eye (here is an arcuate furrow in the prosomal surface. The slope of the antero-median protrusion increases anteriorly until, at (he anterior margin, it is vertical. The surface of the protrusion bears irregularly spaced longitudin al grooves and ridges which radiate anteriorly. Three pairs of curved emarginations occur (Fig. 1) which increase in size posteriorly. The posterior pair occur where the prosomal margin swings anteriorly at the base of the antero-mc:dian protrusion. There is a marginal brim seen as an external mould some 10 mm wide at the anterior of the prosoma. The brim has a shallow embayment at the anterior mid-line. On either side of this, it turns outwards and extends laterally on the protrusion and is continued onto the semi-circular body of the prosoma. Here it becomes narrower posteriorly and fades out before reaching (he post-latera) angles. This brim is twisted into a more vertical position at the emarginations of (he antero-median protrusion. At the anterior margin the brim bears ~trongly developed circular bosses packed closely together each of which shows the base of a movable hair or spine (Fig. I1h). Posteriorly these give place to broad folliculated lunules. closing laterally, and then 10 rather narrow lunules some of which are folliculated. The cardiac region of the prosoma appears to be smoolh but fields of large V-shaped lunules closing post-laterally occur in the antero-lateral region extending post-laterally towards the post-lateral angles from a line joining the base of the antero-median protrusion and the lateral eyes. Mesially lunules in these fields become rounded and broader while laterally they become narrow and asymmetric. Numerous strong. narrow concentric tcrrace-lines are developed on the anterior slope of the eye-furrows. The posterior margin of the prosoma carries strong and large narrow lunules and ridges. The ventral structures of the prosoma have been shifted anteriorly and IOwards the left in relation to the carapace. The metaslOma i.. not present and was probably displaced beyond the nodule. The ventral marginal plates are largely obscured by the appendages and the sutures are not seen.
1.4. Prosomal appendages The proximal podomeres are shown in the ventral view of Ihe prosoma (Fig. 2a) and our interpretation is given in the scale drawing Figure 2b. The coxa of 11mb VI. lacking the gnathobase, is preserved on the right side but on the left side it is missing so that the 300mm ventral view of the whole of the coxa of left V is seen. The coxa of right VI can be removed with the matrix 10 reveal the ventral aspect of the coxa of right IV. This is again F"Igure 1 Cyrlocrenus willebergensis sp . nov. Diagrammatic recon complete except for the gnathobase and its shape is shown in struction (rorn scale drawings of the dorsal surface. Figure 2b, the covered margins being indicated by a broken
Fil!:ure 2 CYrlocrenus willebergensis sp. nov. (a) Photograph of the ventral structures of the prosoma. (b) Interpretative drawing oC the above. Stipple = matrix. black = large cracks. dark shade = prosomal doublure or fragments 01 margioal platcs, ligllt shade = anhrodial membrancs, horizontal lines = matrix infilling appendages. broken lines = margins of coxa of right rv revealed when overlying coxa of righ( VI is removed. r =- right, I = left. II-VI = second to sixth prosomal appendages, 1--4 '" first to fourth podomeres of prosomal appendages, a = spiral coprolite. b = venlral surface of prosoma, c = displaced spine or mould of edge of left coxa of VI. A NEW CYRTOCTENID EURYPTERID FROM SOUTH AFRICA 341
b L---..J 30mm 342 C. D. WATERSTON. B. w. OELOFSEN AND R. D. F. OOSTHUIZEN line. The coxal series on the left side is complete, each coxa widens to the crenulated distal margin. The groove is flanked overlapping that anterior to it and the character of the by acute shoulders ornamented by strong tile-like lunules. gnathobases of all these limbs is evident. The shape of the Podomere 5 carries a strong aotero-ventral spine only the coxa of limb II of both left and right is seen but that of limb base of which has been preserved. The edge of the nodule is III is obscured. It is evident from the leh side that the teeth reached at the proximal end of podomere 6. It is evident, on the gnathobase of limb III are 'normal' in that they are however, that ;t closely resembled 5, the posterior groove similar to those found on the same limb in other eurypterids being present. of comparable size. Those On the gnathobase of IV, As in limb VI, the more distal podomeres of limb V which however, are thinner and on V they are spine-like. are present on the left side occur in a separate piece (Fig. Unfortunately the gnathobases of VI are not known but it IOc, d), and closely resemble those of limb VI. Podomere 4 may be supposed from the apparent posterior reduction of is 80 mm long with a crenulated distal margin. The truncated the tooth size that those on VI would have been spine-like or base of a strong antero-ventral movable spine is present. may even have been absent. The gnathobases are marked by Podomere 5 is 70 mm long with a longitudinal posterior a series of short ridges and grooves normal to the toothed groove bounded by acute shoulders. Large distally closing margin. and overlapping lunules are well seell on the anterior margin Podomeres 2 and 3 of left VI are disarticulated and, on of this podomere but the antero-ventral spine has not been the right side, their form is obscured by the displaced spiral preserved. The edge of the nodule is reached at the distal valve of the gut. However, the arthrodial membrane margin of 5 and only the truncated base of podomere 6 is between the coxa and 2 can be seen passing beneath the seen. From this it is clear that the sixth podomere also had a endospire. Although podomere 2 is articulated with the coxa posterior longitudinal groove. The mould of an anomalous of left V, with the arthrodial membrane indicated, there is spine-like structure projects posteriorly from the proximal disarticulation between it and podomere 3 distal to the end of podomere 4 of limb V (marked c on Fig. 2b). We second arthrodial membrane. These membranes have a interpret this as a displaced anterior spine from limb V or characteristic appearance wherever they are seen, sometimes perhaps the mould of a remaining fragment of the coxa of showing stretch marks and always exhibiting circular limb VI of the leh side. pustules or hairbases. Podomeres 4 and 5 of limbs VI and V From Figures 2a and 2b it is clear, especially from the are preserved on the left side. While the proximal part of 4 right side of the specimen, that the second podomeres are or limb VI is preserved ventral to the carapace, the distal curved, the dorsal lengths being as much as twice the lengthS part of the podomere, together with podomere 5 is seen as a on the ventral side. separate internal mould (Fig. lOa, b). The fourth podomere The distal part of podomere ) of limb rv to the proximal is 86 mm in length, the anterior margin is slightly convex and part of podomere 6 is preserved on a separate piece as an the posterior slightly concave. Coarse rounded crenulations internal mould (Fig. 3). Podomere 3 is 125 mm long on the occur on the distal margin. There is a strong antero-ventral convex anterior margin and 62 mm along the posterior. It movable spine, 35 rom long. which overlaps podomere 5. widens distally to some 60 mm but is here flallened against The lauer is 85 mm in length with a straight anterior margin the carapace. A small aotero-ventral spine is present and the and a strong posterior longitudinal groove of rounded profile crenulations o( the distal margin are well seen on the which narrows distally for two thirds of the length and then external mould of the dorsal side of the podomele. The
IICnlral dorsal
! t
+---'LV.S. of 5---\1
"r---a':s. of 6---1/
a b c d
Figure 3 Cynocrenus winebergensis sp. nov . (a, b) Photograph and drawiag of limb IV of left side in ventral view. (c. d) Drawing and photograph of the SlIme in dorsal view. J-6 = third (0 sixth podomeres, a.v.s. = antero venlral spine. A NEW CYRTOCTENJD EURYPTERID FROM SOUTH AFRICA 343 internal mould shows that the anterior margin was 1.5. Melastoma ornamented by large and rather long scales and lunules while This plate has not been preserved but the overlap indicated the posterior surface also bore large scales but these were on the coxa oC the sixth appendage would suggest that the shorter and more rectangular. The fourth podomere is metastoma must have been triangular in form narrowing similar to 3, measuring 35 mm in length. The antero-ventral anteriorly with the length considerably greater than the spine, however, is longer (26 mm) and is flaltened and width. blade-shaped, rounded distally_ It has a thickened ridge forming the anterior margin from whicb one third curves 1.6. Abdomen posteriorly on tbe dorsal side and two thirds on (he ventral The first tergite of the pre-abdomen is fully exposed but the side . Podomere 5 is 40 mm long at the mid-line and the distal posterior ones have been telescoped (Fig. 11a), the exposed margin has tbe same massive and rounded crenulations as lengths at the mid-line being I, 57 nun; n, 27mm; III, seen in 4. From the proximal part of the ventral surface a 16mm; IV, 14mm; V, 15mm; VI, 16mm; V]J, 45mm. comb of Type A of St0rmer and Waterston (1968) is given Except for the first, all the pre-abdominal tergites are oCf with tbe convex margin of the rachis facing posteriorly. divided into a wide axial portion and pleural portions. The The rows of filaments are borne on the concave anterior axial portion is separated from the pleural parts by sharp margin of the rachis en echelon (op . Cil., text fig. 3e). Some longitudinal upfolds which must have acted as articulations 60 ml11 of the free rachis of the comb is preserved but the limiting lateral movement of succeeding tergites since the distal end is missing. The rachis measures 17 mm at its pair of upfolds on each tergite fit into the upfolds of the widest and the length of the longest filaments is 28 mm. tergite anterior to it. The two lines of upfolds begin on each Proximally the free rachis of the comb flattens and merges side posterior to the ears of the prosoma and diverge into the surface of the podomere and was thus a fixed posteriorly throughout the pre-abdomen. The distance structure. The antero-ventral spine is modified to (orm a between the upfolds being as follows: II, 401 mm; III. nattened movable 'finger', lacking filaments, which is almost 404mm; IV, 420mm; V, 426mm; VI, 430mm; VII, complete and measures 80 mm in Jengtb. It is widest at Ihe 438 mOl. Strong lunules and longitudinal ridges and furro~ proximal end (23 mm). The anterior margin is straight and are developed on the posterior parts of the axial portion of the posterior slightly concave in the proximal part but the tergites and the posterior margin is finely but irregularly convex distaUy. Toward the distal tip there are radiating crenulated (Fig. 11e) . The pleural portions are smooth. ridges. The posterior comb and tbe anterior 'finger' of the From the longitudinal upfolds the pleural portions of the fifth podomere both extend well beyond the distal margin of tergites swing ventro-Iaterally, curving steeply ventrally to the succeeding podomere. Podomere 6 is similar to 5, a small join the ventral body wall at their doublures (Fig. 4) . The part only of the comb being preserved but the filaments can deep body of the holotype allows these marginal features of be clearly seen. The proximal 50 mOl of the antero-ventral tbe pre-abdomen to be seen exceptionally clearly. The first 'finger' are preserved and it was of a much more massive tergite terminates laterally within the post-lateral ears of the nature than that of the fifth podomere. In tbe preserved prosoma while the pleural portion of the second tergite portion the margins are almost parallel and the width is curves ventrally, i.n the manner described, to meet the 25 mm. The veotral surface is ornamented by distally closing ventral body wall at the position of the operculum although, narrow lunules and a row of large overlapping scales occurs in the holotype, the opercular plates bave been displaced. on the anterior margin . The pleural pl'lrtion of the second tergite overlaps that of the Figure 2b shows the proximal podomeres of limbs II and third and so on in tum until the last pre-abdominal tergite is III. The more distal podomeres of these limbs are reflexed reached. and may be traced in separate pieces of the specimen (Fig. The left opercular plate has been displaced anteriorly, lOe) postero-veutral in position to the anterior podomeres having been tom off with part of the skin of the opercular shown in the drawing. In tbe holotype the distal podomeres gill cbamber and DOW lies to the left of the prosoma, at right o( limbs II and III of the right side are continued angles to the life position. The left opercular plate is 120 mm post-ventraUy on the left side of the specimen with tbose of long and 200 mm wide, the ventral surface being heavily the second appendage lying in a more mesial position than ornamented lateraUy and post-lateraUy with rather wide, those of the third. TIle distal parts of limbs II and III of the rounded asymmetrical lunules which are largest in the left side are also deflected to the left and their distal post-lateral area and become smaller anteriorly and mesiaUy extremities are found anterior 10 limb IV of the left side. (Fig. nC). Although the shaft of the median abdominal The distal anatomy of these limbs is hard to interpret since appendage has been lost, tbere are indications that the the structures are bunched together and, without destroying • the specimen by serial sectioning, it is not possible to disentangle tbem. It is apparent, however, Ihal the distal podomeres were constructed on the same pattern as those of limb IV, each having a movable 'finger' associated with a fixed comb. The most proximal comb and 'finger' are associated with podoroere 5 of the third appendage and possibly podomere 4 of limb II. It is also apparent tbat the most distal combs of limbs II and ru are the longest; the longest one seen being associated with a distal podomere of limb III measuring more than 135 mm. The associated movable 'fingers' are of matching lengtb and ornamented by folliculated pustules. Although it is not possible to see the wbole of the distal comb of limb II it is clear that it too must have been a very long oDe. Tbe combs overlapped the podomeres distal to them so that a complex of up to three Figure 4 Cyrloctenus wUlebergerois sp. nov. Diagrammatic skelch pairs of overlapping combs and 'fingers' results. of pre-abdominal lergites. 344 C. D. WATERSTON, B. W. OELOFSEN AND R. D . F. OOSTHUIZEN proximal hastation is still attached to the mesial margin of ante-lateral shoulders and posferiorly as large narrow the left opercular plate. The skin of the ventral body wall lunules at the telson margins. Otherwise the ventral surface which is still attached to the anterior margin of the left is unomamented. The sandy preservation of the holotype in opercular plate formed the anterior par! of thc opercular gill the telson region does not permit preservation of follicles. chamber and is what has been described as the anterior granular skin (Waterston 1975). Although the abdominal 1.8. Digestive tracl appendages posterior 10 the opercular plates are not well This is remarkably displayed in the hOlotype. It has been preserved, skin surfaces of the ventral body wall are partially displaced to the right side and emerges post-laterally near exposed and can be identified with the various elements of the fourth prosomal appendage and fonns a loop posteriorly the gill chamber. (0 be ltidden again below the first tergite (Fig. lOf). The The post-abdominal segmenb decrease in width and anterior 80 mm lacks form but can be recognised by the increase in length posteriorly. They are not epimerate but characteristic black colouration and flinty appearance of the have a rather wide lateral flattened portion separated gut content. This represents probably the stomach region from the axial part of the segment on the dorsal side by a since it is followed postcnorly by a spiral structure. The furrow. The lengths and anterior widths of the post stomach content shows plates of small arthropods, probably abdominal segments are VIII, 44 mm and 550 mm; IX, crustaceans and finely comminuted plant material. 53 mm and 530 mm; X, 71 mm and 504 mm; XI, 105 mm and The spiral structure (Fig. JOi) is 80 mm long. well 444 mm; XII, 129 mm and 354 mm . The posterior parts of preserved and consists of five dextral whorls, the first seen ilS each post-abdominal segment are ornamented by longitudin an impression only. The whorls increase in radius to the al furrows and ridges which are most strongly developed on fourth which measures 50 mm in diameter. Each whorl is (he posterior segments, the posterior margins being finely 18 mm in length. The whorls are crossed sinistrally by well but irregularly crenulated. Post-laterally, large narrow marked regular transverse fine ridges and one or tIVO similar lunules are developed. fine ridges foJlow the whorb irregularly longitudinally. Posterior to the spiral structure the tract turns sharply 1.7. Telson inwards in a mesial direction and a further 100 mm is A striking feature of the telson is the presence of a pair of exposed before it is lost below the /irst tergite. The anterior strong median longitudinal keels on the ventral surface part of this section IS 45 mm wide and it narrows to 25 mm separated by a deep groove on the mid-line (Fig. Ilb). At before being covered. rn the anterior 50 mm closely spaced the anterior the ventral crest~ of these keels are 18 mm transverse ridges and constrictions cross the tract, there below the level o{ the ventral plane of the lateral blades of being eight of these ridges and constrictions in 10 mm. Tn the the telson and their outer edges some 62 mm apart while posterior region the ridges and furrows (fig. 109) are their inner edges are 45 mm apart. The transverse section of covered by an enclosing membrane showing irregular spiral the groove separating the keels has the form of an inverted 'stretching' into little folds and furrows. The membrane itself 'Y·. The telson is thickest at the keel bases and thins carries II finely pustular ornament. The intestine can be laterally. The keels converge and lessen in depth posteriorly detected in cross-section at the posterior of the pre and C 2.2. Method of feeding The balance of evidence suggests that CyrlOclenus was a bottom-living filter- or sweep-feeder. Pelagic filter-feeders a~ong crustacea are characterised by lightness of body and small size and an exaggeration of appendages used in swimming (Fretter & Graham 1967, p. 307). Cyrtocrenw:, on the other hand, is massively constructed and of gigantic size and was therefore certainly not a pelagic feeder. Its massive .- .. '. " ,I . " • build, its powerful walking legs, and the presence of the .. lateral eyes in a central position OD the dorsal surface of the . -.. .. prosoma all suggest that CyrloClenus was a bottom-feeder...... This environment is the second major habitat occupied by tilter-feeding arthropods today where they utilise organisms :'. .. ., . living on the substrate and the rain of particles which falls .. ' . from above. 'Those which inhabit soft, muddy bottoms are hardly to be separated from detritus feeders since the Fi&ure 7 CyrloCienus peach; Sl0nner and Walerslon. Drawing of mechanisms for filtering and coJlecting Dtinute particles will comb, movable linger and associaled podomers of GSE 2127 (rom serve for both' (Marshall & Orr 1960, p. 229). It would lhe CemenlSlonc Group, Dinantian. of Lenoel Braes, Berwickshire. appear, therefore, that any feeding mechanism proposed for Stipple =; matrix, dark shading = cuticle, lighl shading = mould where cuticle has been los I. Cyrloclenus should be consistent with its life as a A NEW CYRTOcrENID EURYPTERID FROM SOUTH AFRICA 347 bouom-liVlng animal feeding either directly upon the rich current passes through the mechanism the setules spread out source of plant and animal food available in that babitat, or to form a filter. When the current passes the other way, the on Ihe smaller animals, such as Crustacea, which themselves paired setules come together and the filter ccases to operate. may have been feeding on the same food source. Among fossil arthropods, Storch, Eriksson, St~rmer and Before presenting support for the view that CyrloclelllLt Bergstrom (summarised by Bergstrom 1969) have all pointed combs could be used for sweep- or filter-feeding, it is out the possible use in filter-feeding of the filaments of the worthwhile considering what olher possible uses the combs outer branch of the biramous appendages of trilobites. On could have been put to and why these uses are considered the other hand, Whittington (1975) believed that these unlikely in this case. Their use as respiratory organs, as for filaments were used primarily in respiration and swimming example in the trilobite exite (Whittington 1975) is rejected and that food catching and transport to the mouth, at least in because the combs were probably the most heavily chitinised such forms as Olenoides serralllS, was accomplished by the structures of Cyrtoclenlls and, as a consequence, were ventral spines of the leg branches and the gnathobases. preserved in the Scottish and Belgian material almost to the CYrJoclenlls combs have a superficial resemblance to exclusion of all else except for the associated 'fingers'. It is modem arthropod filters and to the filaments of the tnlobite known also that, like other eurypterids, C. willebergensis exite but their structure differs from both of these in had gills on the ventral body wall contained in gill chambers fundamental ways. The [act that when a comb is present on closed by plate-like abdominal appendages. Thus the a Cyrloclenus podomere it is invariably associated with primary respiratory function was performed by the gills and. specialisation of the antero-ventral movable spine of that while it remains possible that ancillary respiration could podomere to a finger-like process of similar length to the have laken place through thin cuticle at other sites, it is comb, strongly suggests that these structures were unlikely that the primary function of the heavily chitinised functionally linked and appear to provide a unique comb filaments could have been respiratory. While the adaptalion of eurypterid structures to sweep- or filter combs were still thought of as possible abdominal feeding requirements. appendages (Sl0rmer & Waterston 1968) it was reasonable That the specialised movable 'finger' ;s derived from lhe to compare them with the combs of scorpions which might antero-ventral spine of the podomere is clear from limb IV imply a sensory function. Wills (1959, p. 261; 1960, p. 331) of the SOllth African specimen in which the progression from had noted, however, tbat the Cyrtoclenus comb lacked the a small unspecialised movable spine on the third podomere sensory peg-organs of true scorpions. They also lack sensory to lhe speCialised 'finger' o[ the fifth is obvious. The combs hairs on the comb rachis and in this they contrast with the were fixed structures but their position suggests that tlley movable spines in which setae are present. developed from the postero-ventral spine of the podomere. The central position of the lateral eyes in the posterior The general similarity of the cyrtoctenid comb lo the half of the carapace of CYrloClenus is consistent with a food-gathering fi.lter mechanisms of modem anhropods sweep-feeding, bottom-living animaL In an active predator suggests a simllar function since the flexible filaments of the or striker. such as a plerygotoid, the lateral eyes are brought combs could have acted as strainers and the row of denticles forward to a posilion where Ihe prey and the position of the along the distal edge of each filament as barbs upon which chelae are in view and so can be co-ordinated. In the case of food particles were held. The movable 'fingers' would probable trap-feeders, such as Megalograp Ius , Mixoprerus appear to be ideally adapted to scrape food from the fixed and Lanark op lerus, the lateral eyes also have a forward combs in the course of lransferring it to the mouth. position to see the prey and the array of spines on the Firstly, in considering the (unction of lhe combs it is second and third limbs in which it was trapped. Sweep- or helpful to visualise the probable reaction of their known tilter-feeding, however, being a continuing activity could structures to a waler current meeting the comb at right proceed without the need to observe water-borne particles. angles to the rachis. The filaments facing the current would Here it would be more important to have a wide field of close one upon the other as each filament overlaps its distal vision to see the approach of a threat and this accords with neighbour. At the same time the leeward filaments would the cyrtoctenid condition. become spread out. The opposite would bappen if the Direcl evidence that large pieces of food could not have direction of the current were reversed since the comb is been masticated by Cyrtocrenus is provided by the reduction bilaterally symmetrical. In neither case would the barbs on of the teeth on the posterior gnathobases to spine-like the dislal edges of the filaments be exposed to the oncoming processes. This is again consistent \vitb sweep-feeding. water currenl, but would be closed up against the next distal Sweep-feeding may be regarded as one fonn of filament. We therefore believe that the combs, if moved at filter-feeding in that the passive filter-feeder relies on water rikht angles to the rachis, could not have acted as filters but currents to bring food to the straining mechanism while tbe only as sweeps. The complexity of the comb slrUcture, sweep-feeder actively sweeps the filtering mechanism however, suggests tbat it has an adaptive function and sucb through the food-laden medium. Both require a filter to complexity would not be required for a simple sweeping strain food from the aqueous environment and a means of action. seraping it from the filter and transferring it to the mouth Secondly, consider the reaction of the structures if the (Manton 1977, p. 170-1). Filter-feeding appears to have water current moved parallel to the comb rachis. Let us evolved independently many times in different animal suppose that the current moved distally escaping between groups and, among arthropods, it is widespread (Manton the filaments by way of the rows of denticles near their distal 1977, p. 156-71). In Manton's opinion none of the many margins. Such a stream migbt be maintained by positioning methods employed today can be regarded as primitive or the limb in relation to existing water currents or perhaps by ancentral to another. creating a current by flapping the broad tel son. Movement of Despite the independent development of lilter-feeding in the telson, however, would create an anteriorly moving many modem arthropods, the filter mechanism is usually of current which would meet the combs of tbe reftexed anterior one general type (Dennell 1934), consisting of filtering setae limbs at their distal rather than their proximal ends. It is each with two rows of setules carried at an angle to the plane unlikely also that the combs of limb IV would be effective as of the setal array and facing the food current. When the filters in an anteriorly moving stream since the current would 348 C. D.. WATERSTON. B. W. OELOFSEN AND R. D. f OOSTHUJZEN be met by the curved posterior margin of the rachis. The eTl whence by mechanical action of Ihe spine-like coxal teeth or echelon orientation of the filaments would again tend to possibly hydraulic action generated by the sucking action of close one upon the other under these conditions and. in any the mouth, the food was ingested. event, the space between the two rows of filaments on the The proposed co-ordination of the combs and 'fingers' in comb is too little to allow a sufficient volume of water to CyrtOClenlLs presenls a remarkable adaptation to sweep pass through the system and does not therefore satisfy the feeding of the basic food gathering function of the antero prerequ.isite of a food-coJ!ecting filter to present the maximum and postero-ventral spines of the hughmilleriid lype straining surface to the water current. Let us suppose, on the eurypterid leg. other hand, that the water current ft.owed parallel to the It has been suggested thaI eurypterids swam on Iheir backs rachis proximaHy. Because this would bring into effect all (Patten 1912, p. 371; St0rmer 1934 , p. 62-4) and, if the known structures of the comb in a number of swimming could be so accomplished, why not feeding? TIlis advantageous ways, we consider this the most probable appears to us most improbable since in such a posture option. The en echelon orientation of the filaments along the Cyrtoctenus 1V0uid have been most vulnerable since its rachis (St0rmer & Waterston 1968, text-fig. 3e) would cause ventral anatomy is much less impenetrable than the dorsal the two rows to spread apart. The twist along the length of carapace and terg1tes and the lateral eyes would be rendered the filaments (op. cit., text-fig. 3f) would assist in this useless and unable to warn of Ihe approach of danger. We separation. The flaring of the two rows would expose the therefore conclude that food gathering was carried on while maximum filtration surface to the oncoming current and the standing or walking. If this were so, the antero-median row of barbs On the distal edges of the filaments would be extension of the carapace will be seen to be adapted 10 directed towards the oncoming current and so would arrest accommodate the anterior appendages allowing them free suspended food particles. Water currents would be movement to sweep food into the region of the mouth whjle generated proximally along the comb rachis on movement of affording them dorsal protection. It is likely that the the reftexed second and third limbs towards the ventral body radiating grooves and ridges in the dorsal su·rface of the wall. This movement would also bring the structures into antero-median extension strengthened that part of Ihe close proximity to the mourh; the natural position for the carapace. transfer of food from the combs to the mouth. On these The use of the second to fourth prosomal appendages as limbs being withdrawn from the mouth, the en echelon food gatherers is common among eurypterids and the trend orientation of the comb filaments would cause them to be towards specialisation of these limbs and their spines {or forced together and the filtering action would cease to different forms of food gathering is seen in miiny lineages. function . The elasticity of the cuticle of the filaments would, The movement of the second and Ihird appendages we believe, be such as to facilitate the spreading and closing envisaged for CyrlocreJIus is of a similar kind to that which of the filament rows by the water currents generated through must have taken place in other eurypterids, having their the movement of the limbs in the successive sweep and appendages specialised for food gathering. in transferring retraction strokes. The directional nature of the filter and its food to the mouth. For example. a morphological successive opening and closing (with the resultant reduction progression may be envisaged from the trap-feeder, with of drag on the return slroke) closely parallels the filtering ensnaring spines and well-spaced forward-looking laleral action in modern models. Wbi1e the second and third eyes, laking single prey such as MegaJograplus, MixoplertLS appendages would therefore appear to have been used as and LallQrkoplerUS, to trap-feeders taking smaller prey such active foo" collectors, the anteriorly directed combs of the as CtenoprertLS and Pagea which have eyes in an intermediate fourth appendage would have completed the food trap position, to sweep-feeders, such as Cyrtoclenus, with spines between the ventral body surface and the substrate into adapted to comb filters and eyes close together at the centre which the fiiter-sweeps of the anterior limbs carried the or posterior half of the prosoma. food. As noted above, we believe that one of the functions of 2.3. Digestive tract the movable 'fingers' was to remove food particles from the A remarkable feature of the digestive Iract is the spiral combs during transfer to the moulh. The coincidence of structure which resembles an amphipolar spiral coprolite as shape of the curved posterior margin of Ihe 'fingers' of C. described in fishes by Neumayer (1904). Price (1927) and willebergensis with the cu(Ve of the rachis of the associated Williams (1972) believed that amphipolar coprolites could comb suppons this suggestion. The fine structures of the have been produced by the action of a simple spiral valve 'fingers' moreover prm'ide additional evidence to support and that such a valve may have been present in the this view. The presence of the unique toothed fulcral scales palaeoniscoids. We are not aware however that such a on the 'fingers' (St0rmer & Waterston 1968, text-figs 4, 6, 7) coprolile has yet been found within a palaeoniscoid fossil. would provide a cleaning mechanism when the 'finger' was Since palaeoniscoids, some of conSIderable size, were drawn over the filaments of the comb. The co-ordinal ion of present in the Wilteberg fauna, we must ask whether the this function would be assisted by the sensory structures spiral structure associated with Cyrlocrenus could have which are present on the 'fingers' (op. cit., text-fig. 5) . The originated from a fish, or whether it forms part of the fact that the barbs on the comb filaments face away from the digestive tract (or of its contents) of the eurypterid itself. To rachis indiC IWAOPE CApE ANO .tARoe UA.$INS PARANA. BA.!IH (8t.1IJH~ A.hll:l.~ ( Dtulll W •• '.,h Ceoe Alo Dnn,S. 60 $ul p.,.". $60 Paello ."6 Sant .. Ct.tr. .. ,ln& COfumt..dl Ira II c ..· 1 · - ----~"1----- ~ .... i W."p".Uetl, ! e Narnl6lM 1 ...... -- ---"".~- ..... 3 TC)IUft&I.I&fI o --- ... --'1----.. ~ A F...... ,.... ~ rN.nI.,. , , ,.. f . d .... 19"•• 8 ~. p ...,c:tt.! (; <,::. O'III,UltUI. Figure 8 Correlation of South African and Brazilian eurypttrid-beariog strata. 352 C. D . WATERSTON. B. W. OELOFSEN AND R. D. F. OOSTHUIZEN & Rolfe \985) fragments of large eurypterids were found eye , the furrows being lined with shallow rounded lunules together with scorpions and xiphosuroids (Waterston 1985) . closing outwards and. except for their postero·lateral Of panicular interest among these are a fragment of extremities, the ridges are smooth. Rounded lunnles occur carapace and of walking legs which were found in association "Iso bet ween the lateral eyes posterior to the ocellar node with combs of Cyrtoctenlls peac/li S{0nner and Waterston, and postero-[aterally between the lateral eyes and the and which may prove to belong to that species. Since one of posterior termination of the concentric ridges which they the limbs has the distal podomeres preserved. a feature not cross. yet known in any Cyrtocrenus, and since there IS. as yet, no These characters, and particularly the form and relative record of the carapace of a European cyrtoctenid, these position of the latera! eyes. are very similar to those of C. specimens. now in the Royal Scoltish Museum. Edinburgh, willebergensis although the elongation of the ocellar node are described here. distinguished the Fou/den specimen from that species. It may well be that this fragment is pan of the carapace of C. 4.1. Prosomal rragmen. peachi since other Scottish genera with which it might be The central part of a large prosomaJ carapace (RSMGY compared, such as HibberlOpterus and perhaps DW1sopterus 1984.69.3, lig. [OJ) was obtained from Plant Bed A and (Waterston 1957, text-fig. 8) have the lateral eyes funher shows the lateral eyes and ocellar node but no natural apan with a marked interocular inflation which is lacking in margins have been preserved. Each lateral eye has an outer both Cyrtoctenus lVil/ebergensis and the Foulden fragment. annular thickemng 18 mm in length and 14 mm wide and they are some 5 mm apart. Within the annular thickening there is a flattened palpebral lobe. The visual area of the 4.2. Limb fragments right eye is well preserved being kidney-shaped and Fra~ment L RSMGY 1984.67.6 (Fig. 9c) shows the four extending from a mid-anterior position it wraps round the most distal podomeres of a walking limb in which the tarsus lateral side of the lobe and continues to a postero-median (median length 62 mm) and post-tarsus (23 mm) remain position. The oce!li\r node is long and narrow, its narrow articulated but the basi-tarsus (60 mm) lies alongside the anterior end originating anterior to the lateral eyes. [t extends tarsus at 180" to it and the tibia (39 mm), also disarticulated. to a point about two thirds of the length of the eyes from lies almost at right angles to the basi-tarsus. The podomeres their anterior where it terminates in a rounded posterior are ornamented by narrow V-shaped [unu\es. many of which margin. The ocelli are situated near the posterior margin of are folliculated. which. towards the tip of the post-tarsus. the node which becomes increasingly inflated posteriorly. A become very elongated and asymmetrical giving a striated concentric furrow and ridge occur anterior to each lateral appearance, (he striae converging toward the point of the .. IDmm Figure 9 Limb fragments associated with Crr/Dc/enus peachi St0nner and Waterston, from the Cemenlstone Group, Dinantian, of Foulden, Berwickshirc. (a) RSMGY \984.67.7, (b) RSMGY 1984.67.5, (c) RSMGY 1984.67.6. A NEW CYRTOCTENID EURYPTERlD FROM SOUTH AFRICA 353 podornere. Fragment 2, RSMGY 1984.67.5 (Fig. 9b), by provided with basal extensions, a character known only from comparison with fragment 1, may be interpreted as an the Carboniferous genus Hibberloplems. The Hibbertopter isolated but complete basi-tarsus which is 60 mm in length. It oidea was the only superfamily which St0rmer included in is slightly smaller than the incomplete basi-tarsus of the Hibbertopterina. The establishment of this new fragment 1 and may belong to a more anterior walking limb suborder, on the basis of criteria different from those or to a smaller individual. Fragment 3, RSMGY 1984.67.7 distinguishing the other suborders, was unfortunate. In no (Fig. 9a), by comparison with fragment I, may be hibbertopteroid genus is the form of the chelicerae known. interpreted as a tibia (41 mm) articulated with the femur There is no reason to believe, however, that the first (34 mm). The proximal end of the disarticulated basi-tarsus prosomal appendages are other than small and chelicerate. is also seen. Were they large and specialised it is likely that they would The nature of the ornament and the compatibility of have been observed. The criteria used by St0rmer to dimensions of the podomeres shown in the limb fragments distinguish the new suborder are of the same kind as those strongly suggest that they all belong to tbe same species and which he used successfully to distingusih taxa at the probably represent walking limbs VI and V. Such a limb is superfamily and family level. We use similar criteria in our quite different from the walking limbs of HibbenopteTlis and revised definition of the Hibbertopteroidea and therefore the ornament is quite distinct from that of Dunsopterus. The prefer to regard it as a superfamily of the suborder ornament is closer to that of Woodwardopterus scabrosus but Eurypterina until any evidence is available to the contrary. the relative dimensions of the podomeres are quite different Our work also suggests that members of Ihe Hibbertopter from that form. The ornament is not inconsistent with that oidea, although specialised, may be derived morphologically of C. willebergensis although in general much finer, from more primitive members of the Eurypterina. especially the marginal crenulations. Except on the tarsus St0rmer and Waterston (1968, p. 92) proposed a new there is no evidence in the Foulden material of movable order Cyrtoctenida for their new genus Cyrtoctenus in the spines on tbe podomeres; neither does there appear to be a belief thaI the unique combs and finger-like appendages of posterior groove on the more proximal podomeres that genus formed a succession of paired appendages on the comparable to that seen in Dunsoplems or the bolotype of anterior part oC the abdomen (op. cif., p. &3, text-fig. 10). C. willebergensis although in the flattened condition this may The new evidence provided by the South African specimen be more difficult to detect. As noted, however, the presence shows that Cyrtoctenus can be classified within the of the groove may be a response to directional stress in very Eurypterina and that it must now be included in the large eurypterids. Since the podomeres of the Foulden superfamily Hibbertopteroidea. fragments are less than half the size of the comparable ones in Dunsoplerus and C. winebergensis 00 which grooves arc 5.1. Classification present it may be that grooves were developed only in later Suborder Eurypterina Burmeister l843 instars and their apparent absence from the Foulden Superfamily Hibbertopteroidea Kjellesvig-Waering 1959 material may not be significant. The Foulden fragments Revised diagnosis. Large deep-bodied eurypterids having therefore appear to belong to a walking limb which is semi-circular to sub-elliptical prosomae with central or distinct from other large Scottish Dinantian eurypterids and sub-central ocelli and lateral eyes. Inter-ocular inflations which has some similarities to the ornament of Cyrloctenus may be present on the prosomal surface. The walking legs peachi and C. winebergensis. More evidence would be are of specialised 'Hughmilleria' type with coxal teetb required before these fragments could be identified as reduced and anterior movable spines on the more distal CYlroclenus. podomeres. Second and third prosomal appendages specialised, spinose or comb-bearing. Metastoma triangular. 5. Taxonomic relationships of the Opisthosoma broad, without constriction between the pre Hibbertopteroidea and post-abdomen. Family Hibbertopteridae Kjellesvig-Waeriog 1959 Caster and KjeUesvig-Waering (1964, p. 306) recognised two Revised diagnosis. Hibbertopteroid eurypterids io which suborders within the Eurypterida which they distinguished the walking legs have basal extensions ('Laden') but do not on the basis of the chelicerae. The Eurypterina with small have a longitudinal posterior groove in any of the chelicerae devoid of teetb and the Pterygotina wjth large podomeres. The post-abdomen narrows and it is supposed denticulate chelicerae. St0rmer (1974, p. 377) adopted this that the telson must therefore be styliform or narrowly major division but added a third and new suborder, the hasta·te. Ornament includes semi-circular and V-shaped Hibbertopterina, for large forms which deviate considerably lunules, the latter may become elongate to form narrow from the others and in which the posterior prosomal legs are pointed spines. Figure 10 All specimens of CynoClenus winebergensis sp. nov. unJess otherwise stated. (a) Dorsal view of VIth prosomal appendage x 33%. (b) Veotral view of same showing posterior groove x33%. (e) Ventrnl view of Vth prosonal appendoge sbowing posterior groove X33% . (d) Dorsal view o[ same x 33%. (e) Combs and movable spines associated with limbs n and III x 50% . (f) View o[ gut showing at left the spiral valve and, at right, the intestine xSO%. (g) Detail of tbe anterior portion of the intestine xl. (h) Luoules ornamenting the right post-lateral portioo of the prosoma x 2. (i) Detail of transverse and longitudinal striae on one whorl oC the spiral valve x 2. m?Cyrloclenus pea chi St0rmer and Waterston, lateral eyes and ocellar node of prosomal fragment RSMGY 1984.69 .3 from the Cemeotstone Group, Dinantian, of Foulden, 8erwick.shire x J .5. (k) Detail of comb of C. IVirrebergensis from the 5th podomere of prosomal appendage IV x2. ~ II 354 C. D. WATERSTON, B. w. OELOFSEN AND R. D. F. OOSTHUIZEN r ~ b.-: ~ d g h j k 355 A NEW CYRTOCfENID EURYPTERID FROM SOUTH AFRICA 1 ·1 356 C. D. WATERSTON. B. W. OELOFSEN AND R. D . F. OOSTHUrZEN Genera. HibberlopcfIIS Kjellesvig-Waering J959. Campy these characters may well be convergent and that the family (acephalus Eichwald 1860 may be polyphyletic. Such is almost certainly the case in Family Cyrtoctenidae fam. nov . other eurypterid groups sllch as the Pterygotina. whose Diagnosis. Hibbertopteroid eurypterids in which the members arc also specialised for a common mode of walking legs are large and. in the later instars. the food-gathering as free-swimming predators (Waterston 1964. podomeres in the femoral and tibial regions of che fifth and p. 26). Whether polyphyly exists in such a group can only be sixth appendages have a longitudinal poste(ior groove. Basal testc::d by reference to characters not directly affected by extensions (,Laden') are noC present. The post-abdomen is food-gathering and of these the most valuable taxonomically broad and the telson broadly hastate with. on the ventral would be the nature of the median abdominal appendages. side, a strong median keel composed of two laceral shoulders Unfortunately we are not, at present, able to do so wich the separated by a medIan indentation. The omament is often of HibbertOpleroidea since this evidence is lacking. two orders of size and includes drop-shaped lunules which St0rmer (1951. p. 4Jl-2) pointed out that the prosoma of become elongated to strong adpressed spines with rounded HibberloplcrUS has much in common with the large Lower terminations. Devonian genus Tarsoplerefla and Kjellesvig-Waering (1959. Genera. Cyrloclenus St0rmer and Waterston 1968, p. 253) believed that Tarsoplerel/a might have to be included . Dunsoplerus Walerston 1968, ?Haslimima White [908 . within the Hibbertopteridae. Willweralhta (St0rmer 1936, p. 59: 1969, p. 26) from the Lower Emsian and known only 5.2. Discussion from the carapace, is a form which St0rmer has compared Criteria used In the classification of eurypterids have been with Tarsoptere/lo and which was originally described as a reviewed by St0rmer (1974 . p. 369-72) and Waterston (1979. species of that genus. The relationship of these Devonian p. 292-5) but in no genus now inc/uded in the genera with the Hibbertopteroidea, if any. remains in doubt. Hibbenopteroidea are all these criteria known. [ndeeJ even in the beller known genera Hibberloplems and Cyrioclerllls 6. Acknowledgements we lack information essential for a confident classification. It is not yet possible to construct a meaningful dendrogram We thank P. J. Brand of the British Geological Survey, illustrating the relationships of the genera which we include Edinburgh and the Director of the South African Museum in the superfamily. for the opportunity to study specimens in their charge. For Dunsoplerus and Cyrloclenlls are very similar in their photographic services we are indebted to Clive Booth of the known characters of prosomal shape. eyes. posterior walking South African Museum and to Ken Smith of the Royal legs and tergite shape. The ornament is also similar although Scottish Museum, and we thank the Directors of both in DunsoplerliS it appears to be stronger than in Cyrloclenus. institutions for the provision of facilities. We are grateful to We have no evidence of the telson of Dunsoptcrus nor have Nancy Bryce for preparing Ihc drawings for publication and any combs or movable 'fingers'. such as those developed in Karin Malan for the stratigraphic table . C. D. Waterston CyrtoClenus. been found in association wich DUlIsoplerus. expresses thanks to the Director of the Royal Scottish We cannot prove, therefore. that these genera are Museum for making possible his visit to South Africa and 10 congeneric although further evidence may show this to be the University of Stellenbosch for assistance there. B. W. the case. t Oelofsen is indebted to the Universicy of Stellenbosch and We consider that the telson of Haslimima must have been the C.S.1. R . of South Africa for financial support. strikingly similar to that of Cyrloclenus in che uncompressed state. The inverted key-hole lunules of Hastimima also are not dissimilar to the drop-shaped lunules of CyrlaClcmts and 7. References Dunsoplerus although Haslimima does not appear to show two orders of ornament size as seen in the other two genera. Bergs1r6m, J. 1969. Remarks on Ihe Appendages of Trilobites. LEl1--IAIA 2, 395-414 . The coxa with aICached leg segments of Haslimima, Boucot. A. J ., Brunlon. C. H. C. & Theron. J. N. 1983. illustrated by Clarke and Ruedemann (19[2. text-fig. 93) Implications for Ihe age of Soulh African Devonian rocks in shows well developed masticatory teeth which differs from which Tropidoleptus (Brachiopoda) has been found. GEOL other genera of the Hibbertopteroidea. On the other hand MAG 120, 51-8. there is nothing to prove that che coxa is that of HasliJllima BriW. D. E. G., Rolfe. W. D. I. & Brannan, J. 1979. A giant mynapod trail from the Namurian of Arran. Scotland. and we know that other genera are found in association with ~ PALAEONTOLOGY 22, 273--91. it. We associate the Brazilian genus with the Cyrtoctenidae Caster, K. E. 1947. Expedicao a 0 Eslado do Piaui. MIN METAL with hesitation since no other family of Eurypterina accounts 12, 271-2. so well for such characters as are known in Haslimima. Caster. K. E. & Kjellesvig-Waering. E. N. 1964. Upper Ordovician Eurypterids of Ohio. PALAEONTOGR AM 4, 301-42. The characters which we have used to define the Clarke. 1. M. & Ruedemann, R . 1912 . The Euryplerida of New Hibbertopteroida are related to specialisations for a York. MEM NEW YORK STATE MUS NAT HIST 14, vol. I, common mode of food-gathering. We are thus aware that 1-439_ vol. 2. plates. Figure 11 A II specimens of C)'rlOClenus willebergensis sp. nov . (a) View of external mould of Ihe dorsal surface of prosoma, abdomen lind lelson x 15%. (b) Venlral view of inlernal mould of lelson x2S%. (c) Lateral parts of the pre-abdominal lergites x 33% . (d) Posterior cu(icle of gill chamber x2. (e) Ornament of posterior edge of pre-abdominal lergite x I. (f) Lunules ornamenting Ihe left opercular plate x 1.5. (g) Ventral view of prC-lelson x 25% . (h) Anterior view of Ihe right stde of the internal mould of the prosomal margin showing the re·enHant curve and circular follicles wilh seta-bases x I . A NEW CYRTOCTENID EURYPTERTD FROM SOUTH AFRICA 357 Clarke, K. U . 1979. Visceral anatomy and anhropod phylogeny. In Plumstead, E . P. 1967. A general review of the Devonian fossil Gupta, A . P . (ed.) ArthropOd phylogeny, 467-549. New York: plants found in the Cape System of South AfriC3 . Van Nostrand Reinhold. PALAEONTOL AFR 10, J-&3 . Cooper, M. R. 1982. A Revision of the Devonian (Emsian-Eifeliao) Plumslead, E. P. 1969. Three thousand million years of plant life in Trilobita from the Bokkeveld Group of South Africa. S AFR Africa. GEOL SOC S AFR TRANS 80, 267-77. MUS ANN 89, (I), 1-174. Price, P. J927. The cop(olite limestone horizon of the Conemaugb Copeland, M. 1. & Bolton, T. E . 1960. Canadian Fossil Anhropoda Series in and around Morgantown, West Virginia . CARNEGIE Eurypterida, Phyllocarida and Decapoda. BULL GEOL SURV MUS ANN 17, 211-54. CAN 60, 1-84. Rennie, 1. V. L. 1934 . Note on an Eslheria from the Willeberg Dennell, R. 1934. The feeding mechanism of the common cumacean Series. S AFR 1 SCT 31, 233-5. crustacean Diastylis bradyi. TRANS R SOC EDINBURGH 58, Ruedcmann, R. 1921. Preservation of Alimentary Canal in an 125-42. Eurypterid. BULL NEW YORK STATE MUS 2:l71'J.2B, 92-5. Du Toit, A. 1954. The Geology of SOUlII Africa, 3rd cdn. Salamuni, R. & Bigarella, J. J. 1967. The pre-Gondwana basement. Edinburgh: Oliver & Boyd. In Bigarella, 1. 1., Becker, R. D . & Pinto, I. D. (eds) Problems Fage. L 1949. C1asse des M~rostoma~ (Merostomata, Woodward in GOMwana Geology, 3-22, TSI . IntI. Gondwana Symposium. (1866)) . In Grasst, P .-P. (ed.) Traut de Zoologic VI, 219-62. Salter, 1. W. 1862. On some Fossil Crustacea from the Coal Paris: Masson. Measures and Devonian Rocks of British North America. Q J Forsythe, R . 1982. The lale Palaeozoic to early Mesozoic evolution GEOL SOC LONDON 19, 75-80. of southern America: a plate tectonic interpretation. 1 GEOL Seward, A . C. 1909. Notes 00 Fossil Plants from the Wiueberg SOC LONDON 139, 671-82. Series of Cape Colony. GEOL MAG (5) 6, 482-5. FrClter, V. & Graham, A . 1976. A Functional Anatomy of Simpson, S. 1970. Notes on Zoophycos and Spirophyton. Tn Crimes, IJlvertebrates. London: Academic Press. T . P. & Harper, 1. C. (cds) Tra.ce Fossils, 505-l4. Liverpool: Fritsch, A. 1895. Fauna der Gaskohle UM der Kalksreine der Seel House. Permfomwtion Bohmens, 3, no. 4, I-J32. Prague. Spencer, W . K. 1950. A new Briltlestar and an Eurypterid from the Gardiner, B. G . 1969 . New paJaeoniscoid fish from the Willeberg Bokkeveld straLa . S APR J SCI, June 1950, p. 300. Scries of South Africa. ZOOL 1 LINN SOC 48, 423--52. Stapleton, R. P. I 977a. Carboniferous unconformity in southern Heubusch, C. A. 1962. Preservation of the intestine in three Africa. NATURE LONDON 268, 222-3. specimens of £uryprerus. J PALEONTOL 36,222-4. Stapleton, R. P. 1977b. Carbonised Devonian spores {rom South Jubb. R. A . 1965 . A new paJaeoniscid fisb from the Willeberg Nrica. POLLEN SPORES 19, 427-40. Scries (Lower Carboniferous) of South Africa . S AFR MUS St0rmer. L 1934. Meroslomata from the Downtonian Sandstone of ANN 48, 267-72. Ringerike, Norway. SKR NOR VIDENSK-AKAD MAT Kaestner, A . 1968. [nvertebrale Zoology, vol. 2, New York: NATURVIDENSK KL 1933,10,1-125. Interscience. (Translated by H . W. and L. R . Levi.) St0rmer, L. 1936. Eurypteriden aus dem Rheinischen Unterdevon. Kjellcsvig-Waering, E. N. 1948. Thc Mazon Creek Eurypterid: a ASH PREUSS GEOL LANDESANST 175, 1-74. Revision of the Genus Lepidoderma. TLLrNOiS STATE MUS St0rmer. L. 1951. A New Eurypterid from the Ordovician of SCI PAP 3 (4), 1-46. Montgomeryshire, Wales. GEOL MAG 88, 409--22. Kjellcsvig-Waering, E . N. 1958. Some previously unknown St0rmer, L. 1969. Eurypterids from tbe Lower Devonian of morphological structures of Cardnoso1lUJ lIewli'li (Claypole). 1 Willwerath. Eifel. SENCKENBERGJANA LETIiAEA SO, PALEONTOL 32, 295-303. 21-35. Kjellcsvig· Wareing, E . N. 1959. A Taxonomic Review of some Late St0rmer, L. 1974 . Arthropods {rom the Lower Devonian (Lower Paleozoic Eurypterida. J PALEONTOL 33,251-6. Emsian) of AIken an der Mosel, Germany. Part 4: Eurypterida, Kjellcsvig-Waering, E. N. [961. Nota sobra la p(esencia. de un DrepaDopteridae and otber groups. SENCKENBERGlANA Euryp(eridao en el Devonico Inferior de la Argentina. REV LETHAEA 54, 359-451. ASSOC GEOL ARGENTINA 15, (102), 109-11 . St0rmer, L. & Waterston, C. D . 1968. CyrlOClenU9 gen. nov., a Kjcllesvig-Waering, E. N. 1963. PeDJlsylvanian Invenebnlles of the large late Palaeozoic Anhropod with pectinate Appendages. Mazon Creek Area; Illinois: Eurypterida. FlELDlANA TRANS R SOC EDINBURGH 68, 63-104. (GEOL) U, 85-106. Strand, E. 1928. Miscellanea nomenclatorica zoologic.a et Kjcllcsvig-Waering, E. N. 1966. A Revision of the Families and paleontologica HI. ARCH NA TURGESCH 92A, H B, 30-75. Genera of the Stylonuracea (Eurypterida). FTELDlANA Tankard, A . 1., Jackson, M. P. A ., Eriksson, K. A ., Hobday, P. (GEOL) 14, 169-97. K., Hunter, D. R. & Minter, W. E . L. 1982. Curs tal evolution LIIke . P. 1904. The trilobites of tbe Bokkeveld Beds. S AFR MUS in South Africa. New York: Springer. ANN 4, 201-20. Theron, J. N. 1962. An analysis of the Cape folding in the district of Wock. 1. C. 1967. The sfrluigraphy of the Wu/eberg·Dwy/w contact WiUowmore, Cape Province. UNIV STELLENBOSCH ANN beds. Unpublished M.Sc. thesis, University of SteUenboscb, A37, 247-419. South Africa. Vonk, H. J . 1960. Digestion and Metabolism. In Waterman, T. H . Manton, S. M. 1977 . The AnhropotUJ, Habits, Functional (ed.) The PhySiology of CrusUleea, vol. l: Metabolism and Morphology OM Evolution. Oxford: Clarendon. Growth. New York: Academic. Ma~hall, S. M. & Orr, A. P. 1960. In Waterman, T. H . (ed.) The Waterslon, C. D. 1957. The Scottish Carboniferous Eurypterida. Physiology of Crustaaa, vol. 1,227-58. New York: Academic. TRANS R SOC EDINBURGH 63, 265-8B. Mcualira, S. 1951. Ocorrencias do euripteridio Htwimima no Wa~ton, C. D. 1964. Observations on Pterygotid Eurypterids. estado de Sao Paulo. D G M NOT PREL E EST 52, l-14. TlUNS R SOC EDINBURGH 66,9-33. Neumayer, L. 1904. Die Koprolitheo des Perms von Texas. Walerston, C. D. 1968. Further Observations on tbe Scottish PALAEONTOGRAPHICA 51, 121-8. Carboniferous Eurypterids. TRANS R SOC EDrNBURGH 68, OclofseD, ~. W. 1981. An anatomical and syst~atic sTUdy of the 1-20. family Mesosauri.dJle (Reptilia, Progonc)3(uiTUJ) with special Waterston, C. D. 1975. Gill structures in the Lower Devonian refermce . to U.s as90ciaud fauna and palaco·ecological euryptcrid Ta.rsoplerella seotica. FOSSILS STRATA 4, 241-54. environment in the WhiIehili sea. Unpublished Ph.D. thesis, Watemon, C. D. 1979. Problems of functional morphology aod University of Stellenbosch, Soutb AIrica. classification in stylonurid eurypterids (Cbelicerata, Merostoma Parker. T. 1. 1885. On the intestinal spiral valve in the genus Raw. ta), with observations on the Scottish Silurian StyloDLtridae. TRANS ZOOL SOC LONDON 11, 4C}--61. TRANS R SOC EDINBURGH 70, 251-322. Patten, W. 1912. The evolwion of Ihe venebrQ/e3 and their kin. Waterston, C. D. 1985. Chelicerala from the Dinantian of Foulden, Philadelphia. Berwickshire. TRANS R SOC EDINBURGH EARTH SCI Pinto. J. D . 1972. Permian insects from Parana Basin, Soutb Brazil 76,25-33. J. Mecoptera. REV BRAS GEOC 2, 105-16. White, D. 1908. Report on the fossil flora of the coal measures of Pinto, I. D., Ornellas, L. P. & Purper, I. 1978. Fossil insects and Brazil. Tn White, 1. C. (ed.) Final repon on the coal measures the problem of Permo-Carboniferous boundaries. Act. II and associated roeks of Sowh Brazil, 377-(:iJ7. Rio de 1aneiro: Congr. Argentino Pateont. Biocstr. I CONGR LATINOAM National Press. PALEONTOL B A 1978 IV (1980) , 207-13. Whittington, H . B. 1975. Trilobites with appendages from the P!otnik, R . E. 1985. Lift based mechanisms for swimming in Middle Cambrian, Burgess Sbale, British Columbia. FOSSTLS eurypterids and ponunid crabs. TRANS R SOC EDIN STRATA 4, 97-136. BURGH EARTH SCI 76, 325-37. Williams, M. E . 1972. The Origin of 'Spiral Coprolites'. UNJV 358 C. D . WATERSTON, B. W . OELOFSEN AND R. D . F. OOSTI-fUrZEN KANSAS PALEONTOL CONTRIB Paper 59, 1-19. 1-6. Wills, L J. 1959. The ,External Anatomy of some Carboniferous Woodward, H. 1909. Note on the genus Haslimimo from Bra2il and 'Scorpions' Pan I. PALAEONTOLOGY 1,261-82 the Cape. GEOL MAG (5) 6, 485-8. Wills, L. 1. 1960. The External Anatomy of some Carboniferous Zidek, J. 1980. Acanrhodes /undi, new species (Acanthodii). and 'Scorpions' Pan 11 . PALAEONTOLOGY 3, 276-332. associated coprolites from Uppermost Mississippian Heath Wood, S. P. & Rolfe, W. O. I. 1985 . IntrodlJction to the Formation of Central Montana. ANN CARNEGIE MUS 49, palaeontology of the Dinantian of Foulden, Bcrwickshire, 69-78. Scotland. TRANS R SOC EDINBURGH EARTH SCI 76, C. D. WA TERSTON. Department of Geology, Royal Sco\lish Museum, Chambers Street, Edinburgh EHI JJF, Scotland. B. W. OELOFSEN, Department oC Zoology, Faculty of Science. University of Stellenbosch, Stcllenbosch 7(fYJ, South Africa. R. D . F. OOSTHUIZEN, Zwanskraal, Klaarstroom, District of Prince Alben, South Afrie