Norntattes PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

AMERICAN MUSEUM Norntattes PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 3076, 15 pp., 8 figures November 1 , 993

A New Clupeomorph Fish from the Santana Formation (Albian) of NE Brazil

JOHN G. MAISEY'

ABSTRACT A new clupeomorph fish is reported from the cent of the engrauloid condition. Several plesio- Lower Cretaceous (Albian) Santana Formation of morphic character states conflict with that NE Brazil. It is ofsmall size; the largest individual hypothesis, such as the autogenous parhypural, lack is approximately 120 mm standard length, and of fusion between the first uroneural and first some are considerably smaller. The new fish dis- preural centrum, unreduced first ural centrum, plays several clupeomorph and clupeiform char- presence of a complete series of lateral line scales, acters, but its placement within the Clupeiformes and lack of dorsal scutes; therefore the fish is con- is problematic. It possesses obliquely inclined sus- sidered as a clupeiform incertae sedis. pensorium and elongate jaws, which are reminis- INTRODUCTION The Santana Formation ofthe Chapada do The clupeomorph described here is a rare Araripe is a well-known source offossil fishes constituent ofthe assemblage, and is the first representing a tremendously diverse ichthyo- clupeomorph species to be described from fauna. Many of these fossils (including those the Santana Formation. Other undescribed studied here) occur in carbonate concretions and somewhat different clupeomorphs occur of late diagenetic origin, within shales and lower in the Romualdo Member, from py- marls of the Romualdo Member (the upper- robituminous shales overlying gypsum de- most member of the Santana Formation). posits of the Ipubi Member. Recent palynological studies have deter- This new clupeomorph was first reported mined the age of these fossils to be Middle by Maisey (199lb) from two small individ- or Late Albian (Pons et al., 1990). uals in a single concretion belonging to the

' Curator, Department of Vertebrate Paleontology, American Museum of Natural History.

American Museum, associated with disartic- Smxa anterior supramaxilla ulated bones of the primitive gonorynchi- Smxp posterior supramaxilla form Tharrhias. Subsequently, another spec- Soc supraoccipital Sop subopercle imen was identified in the AMNH collection, Spo sphenotic and some others were found among material Steg stegural in the collections of the Departamento Na- tf temporal fenestra cional de Producao Mineral (DNPM), Rio de U ural centrum Janeiro. One of these DNPM specimens has Un uroneural been selected to serve as the holotype (fig. 1). VII, X, XII foramina for cranial nerves Although the total sample is small, these clupeomorph specimens seem to represent a sin- SYSTEMATICS gle taxon, which is treated here as a new genus and species. COHORT CLUPEOCEPHALA SUBCOHORT CLUPEOMORPHA ABBREVIATIONS ORDER CLUPEIFORMES Institutional Clupeiformes incertae sedis AMNH American Museum ofNatural His- tory Santanaclupea, new genus DNPM Departamento Nacional de Produ- ETYMOLOGY: After Santana do Cariri, the cao Mineral, Rio de Janeiro town giving its name to the famous Santana Anatomical Formation, and Clupea (Latin), small river A angular fish. af auditory fenestra DIAGNOSIS: Clupeomorph fish with pari- Bo basioccipital etals separated by supraoccipital, supratem- D dentary poral sensory canal passing through parietals Epo epioccipital but external to supraoccipital; dilatator fossa Exo exoccipital shallow, without roof; preepiotic fossa large; Fr frontal temporal foramen large; no frontal fonta- fp preepiotic fossa nelle; hyomandibula with slight posterior in- H hypural clination; complete series oflateral line scales Hm hyomandibula extending back to caudal region; continuous hmfa anterior hyomandibular facet hmfp posterior hyomandibular facet series of ventral scutes present from behind Ic intercalar pectoral fins as far as anal fin; 19 or 20 ab- lop interopercular dominal vertebrae plus 19 caudal and 2 ural Le lateral ethmoid centra. Proportions (as percent of standard Me mesethmoid length): head 22, maximum trunk depth 33, Mpt metapterygoid predorsal 45-47, prepelvic 51-53, preanal 75. Mx maxilla Pelvic fins slightly posterior to dorsal; fin ray Npu preural neural arch counts: D14; Al 1; P12; V8-9; C19 principal Op opercle rays (10 upper, 9 lower), plus 5 dorsal and 4 or foramen for orbital artery ventral procurrent rays. Os orbitosphenoid Pa parietal TYPE SPECIES: Santanaclupea silvasantosi. Ph parhypural Pmx premaxilla Santanaclupea silvasantosi, new species Pop preopercle ETYMOLOGY: silvasantosi after Prof. Ru- Pro prootic bens de Silva Santos, preeminent Brazilian Psp parasphenoid Pto pterotic authority on paleoichthyology, who de- Pu preural centrum scribed many species offossil fishes from the Q quadrate Santana Formation. rec openings for recessus lateralis HOLOTYPE: DNPM 51 5-P, fish lacking tail, S symplectic in part and counterpart, preserved length ap- 1993 MAISEY: NEW CLUPEOMORPH FISH 3

,,'',,t%>;,r- i p V .j

C*1 r 57? < 4 2t

- r.

Fig. 1. Santanaclupea silvasantosi. (A, B) The holotype DNPM DGM 515-P; (A) part and (B) counterpart after acid preparation; (C) DNPM DGM 1338-P (B and C reversed). prox. 85 mm, estimated standard length ap- mated standard length of approx. 120 mm prox. 100 mm, Santana Formation, Chapada (figs. 2C, 3A). AMNH 12789 and 12790, two do Araripe, now acid-prepared (figs. IA, B, small associated individuals, both with a 4, 8). standard length of approx. 45 mm (figs. 2A, REFERRED MATERIAL: AMNH 18968, the B, 3B, C, 6A). DNPM 1338-P, poorly pre- largest known individual, having an esti- served individual with good tail, part of dor- 4 AMERICAN MUSEUM NOVITATES NO. 3076

1.m A

-.vI.

Fig. 2. Santanaclupea silvasantosi. (A) AMNH 12790; (B) AMNH 12789; (C) AMNH 18968 (A and B reversed). sal fin and opercular series, estimated standard length approx. 60 mm (fig. 6B). DNPM DESCRIPTION 683-P, complete fish approximately 60 mm The ethmoid region and anterior part of standard length. the orbit are incomplete in all the specimens Specimens have been acid-prepared apart examined except for the holotype, DNPM from DNPM 1338-P and 683-P. A part and 515-P. Most of the postorbital region of the a counterpart exist only in the holotype. braincase is well preserved in this specimen DIAGNOSIS: As for genus. and in AMNH 18968, which together pro- 1993 MAISEY: NEW CLUPEOMORPH FISH 5

.a -B

Fig. 3. Details of the head in (A) AMNH 12790; (B) AMNH 12789; (C) AMNH 18968 (B and C reversed). 6 AMERICAN MUSEUM NOVITATES NO. 3076

Fig. 4. Braincase in DNPM DGM 515-P. (A) Oblique dorsal view showing supraoccipital separating parietals, commissural canal not entering supraoccipital; (B) lateral view ofsame, with pitted "strawberry appearance" of the prootic; (C) lateral view of the counterpart, with displaced hyomandibula (A and B reversed). vide most ofthe information given here (figs. are juvenile. The head and jaws are well pre- 4, 5). In the two small associated individuals, served in DNPM 683-P. Nothing ofthe head AMNH 12789 and 12790, the neurocranium is preserved in DNPM 1338-P. is incompletely ossified and individual bones The parietals, supraoccipital, and posterior are disarticulated. Their size and degree of part of the frontals are ornamented by sin- ossification suggest that these two specimens uous ridges and rounded processes. Farther 1993 MAISEY: NEW CLUPEOMORPH FISH 7

ix!

Fig. 5. Reconstruction of (A) braincase and (B) parts of head in Santanaclupea. anteriorly the frontals are smoother. The su- the supraoccipital (fig. 4A, B). The parietal praoccipital completely separates the pari- has a broad contact with the frontal anteriorly etals. The supratemporal commissural canal and meets the epioccipital posteriorly. There is enclosed by the parietals, but is external to is a lateral flange on the parietal above the 8 AMERICAN MUSEUM NOVITATES NO. 3076

t t:};< - >-,A

N B 41A

C-~~~~~~~~~~5.n m . v.

Fig. 6. Tails of(A) AMNH 12790 and (B) DNPM DGM 1338-P (both reversed); (C) reconstruction ofthe caudal skeleton in Santanaclupea; (D) caudal skeleton in Santanichthys diasii, from AMNH 12780 (from Maisey, 1991a). pterotic, ornamented in the largest specimen The frontals have a lower temporal flange. (AMNH 18968) by several small serrations. In AMNH 18968 a small notch is present in The preepiotic fossa (fp, fig. 5) is large, as in the posterior margin ofthe orbit, perhaps for most living clupeoids (see Patterson, 1970, innervation of part of the supraorbital sen- for references). Its limits are largely defined sory canal (fig. 3C). The parietals and frontals by the epioccipital, and it is floored by the contact each other dorsally, but farther lat- pterotic. erally these bones are separated by a gap. A 1993 MAISEY: NEW CLUPEOMORPH FISH 9

C recessus lateralis, although it is not clear which branches of the sensory canal network oc- cupied each opening. A dermosphenotic has not been observed; it was either absent or else very poorly developed (as in other clu- Npu, peomorphs with a well-developed recessus ul lateralis). Prootic and pterotic bullae are well developed. The external surface of the prootic is pitted, with a characteristic "strawberry appearance" (fig. 4) seen in many clupeomorphs (Forey, 1975). There is a large preepiotic fossa in the epioccipital. A membranous inter- \ \ calar is present (seen only in the holotype). p Pu3 A canal-bearing extrascapular bone over- lies part ofthe pterotic in AMNH 18968, but is too fragmentary for adequate description. A posttemporal bone is also present, more or less in its presumed original position behind the epioccipital, where it carries the posttemporal canal from the extrascapular. The orbitosphenoid (Os, fig. 5A) is small U1+ NpU2 but well ossified. There is a slender, almost straight edentulous parasphenoid. It is pierced by paired carotid foramina posteriorly. With- in the orbit of the holotype there is a ptero- sphenoid, displaced from its original position. This specimen is the only one examined where the mesethmoid is preserved (fig. 4). There are traces of what may be a poorly Ph- ossified lateral ethmoid within the anterior part of the orbit in AMNH 12789. The pos- Fig. 6. Continued. terior part of the vomer is visible in AMNH 12789 and 12790, but I have been unable to parietal branch of the supraorbita1 sensory determine whether the mesethmoid over- canal presumably lay below the lateral mar- hung the vomer, as in engrauloids. gin ofthe parietal, before entering tihe frontal The jaws are long and slender, and sharp, via the temporal foramen. The temporal pointed teeth are present on the premaxilla, flange of the frontal is sutured to tI he sphen- maxilla, and dentary (fig. 5B). The preda- otic, and forms a small 'depressed area that ceous habits of Santanaclupea are evidenced is interpreted as the dilatator fossa. If this in the holotype by the presence ofa complete interpretation is correct, however, the dila- skeleton ofa small Tharrhias within the body tator fossa is weakly developed aind is un- cavity (fig. 8). The premaxilla is considerably usual in lacking any pterotic contri.bution. shorter than the maxilla (approximately 20% The ventrolateral face ofthe sphen[otic bears of maxillary length), as in many engrauloids. a circular articulatory facet for the anterior A small flange of bone on the premaxilla hyomandibular condyle, and meetvs the pte- overlaps the maxilla anteriorly. There is a low rotic posteriorly (fig. SA). The pterotic has palatine articulation on the dorsal surface of an elongate facet (extending onto the sphe- the maxilla anteriorly, in front of which is a notic) for the posterior hyomandiboular con- slender ethmoid articulation mesial to the dyle, behind which is a short, posteroven- premaxilla. These articulations were proba- trally directed spine. Within the pt!erotic are bly simple sliding surfaces. Two supramax- two foramina, interpreted as openinigs for the illae are present (AMNH 18968). In the lower 10 AMERICAN MUSEUM NOVITATES NO. 3076 jaw there is a long, oblique suture between tends anteriorly as far as the articulation be- the dentary and the angular. The mandibular tween ural centra 1 and 2. sensory canal lies in an open groove on the There are 10 upper and 9 lower principal dentary for almost three-quarters ofits length, caudal fin-rays. The last dorsal principal ray but is partly enclosed in a tube anteriorly. lies on the diastema between the second and The canal then passes through the angular to third hypural. In addition to the principal emerge on its posterior face below and behind rays, there are five dorsal and four ventral the mandibular joint. Unfortunately the in- procurrent rays. ner aspect ofthe lower jaw cannot be studied The description ofthe caudal fin given here in this material, and it is consequently un- essentially agrees with that presented by Silva known whether the angular, articular, and Santos (1991: 30, fig 10) of DNPM 683-P, retroarticular are fused or separate. which is regarded here as another specimen The anterior ceratohyal contains a large of Santanaclupea. The identity of that spec- "beryciform" foramen; this primitive tele- imen will be discussed in a note at the end ostean feature is absent in clupeiform fishes of this paper. (i.e., clupeiformes sensu Grande, 1985). The There are approximately 19-20 abdominal hyomandibula is slender, has separate ante- and 19 caudal vertebrae, although a precise rior and posterior articular condyles, and is count cannot be made on the basis of avail- inclined obliquely backward from the brain- able specimens. Each rib is borne by a short case (fig. 5B). In this last respect Santanaclu- parapophysis, and the posterior ribs do not pea resembles engrauloids. diminish in length appreciably. The neural Dermal bones of the cheek are poorly pre- spines of the anterior abdominal vertebrae served, and nothing can be said about the are unfused, as in most clupeomorphs; the infraorbital series. The opercular series is condition of the neural spines on more pos- poorly known, mainly from AMNH 12789, teriorly positioned abdominal vertebrae is 12790, and 18968. uncertain. The caudal vertebrae bear single The pectoral and pelvic girdles are pre- median spines. served in these examples and also in the ho- The body is entirely covered by thin scales, lotype. There are at least 15 pectoral fin-rays although these are clearly visible only in the and at least nine pelvic fin-rays. largest specimen, AMNH 18968 (fig. 2). Con- In AMNH 18968 the tail has been folded trary to the earlier report (Maisey, 1991 b) the over itself, but the caudal endoskeleton is well lateral line series is completely developed, a preserved in DNPM 1338-P, 683-P, and primitive condition among clupeomorphs. A AMNH 12790 (fig. 6). Five hypurals are pres- continuous series ofventral scutes is present. ent; among other clupeomorphs only Denti- There is no sign of any dorsal scutes. ceps is known to have this many. Preural centrum 1 is fused to a short, expansive neural arch dorsally and to the parhypural ventrally DISCUSSION (fig. 6). A. Santanaclupea as a clupeomorph. Clu- Hypural 1 is autogenous, and articulates peomorph fishes have been characterized with ural centrum 1, which lacks a neural (Grande, 1985) by several synapomorphies, arch. Hypural 2 is fused to ural centrum 1. including: There is a distinct diastema between hypurals 1. abdominal scutes at the ventral midline; 2 and 3. Hypurals 3-5 are all autogenous, and 2. otophysic connection that penetrates the all meet ural centrum 2 (which also lacks a exoccipital and forms ossified bullae in prootic neural arch). and usually also in the pterotic; 3. supratem- Two uroneurals are present in AMNH poral sensory canal penetrates parietals (or 12790 and DNPM 683-P; the first extends parietals and supraoccipital); 4. second hy- anteriorly above the first preural centrum but pural fused with first ural centrum; first hy- is not fused to it. There is no evidence ofany pural autogenous; 5. well-defined preepiotic dorsal outgrowth of membrane bone on the fossa; 6. dorsal scutes present. first uroneural (this is commonly found in Santanaclupea is known to possess char- clupeomorphs). The second uroneural ex- acters 1, 3, 4, and 5, but lacks character 6. I 1 993 MAISEY: NEW CLUPEOMORPH FISH I1I have not seen the otophysic connection (character 2), but assume it was present because of the presence of well-developed prootic and pterotic bullae, which are typi- cally connected to a diverticulum ofthe swim bladder via the otophysic connection. The supratemporal canal passes through the parietals (character 3), but as in Recent clupeomorphs, the sensory canal is superficial to the supraoccipital and not enclosed in bone. The articulation between hypural 1 and ural centrum 1 is well developed, a primitive condition that is lost in more advanced clupeomorphs (clupeoids; Cavender, 1966). The first ural centrum is not reduced in size, another primitive feature (also found in Denticeps and Diplomystus). This centrum is much smaller than the first preural centrum in clupeoid clupeomorphs; Grande, 1985) It is concluded that Santanaclupea is a clupeomorph on the basis ofcharacters 1-5, but that it also retains some plesiomorphic characters in its caudal skeleton. Grande (1985) subdivided clupeomorphs into two "divisions," but only a single fossil taxon (Erichalcis arcta) was included in his "Division 1" because it lacks derived characters 4-6; in fact, it has not been shown to possess an otophysic connection or prootic and pterotic bullae, and the grounds for including it in clupeomorphs at all are extremely weak. B. Comparison with Spratticeps. Another Lower Cretaceous (Albian) clupeomorph, Spratticeps gaultinoides Patterson, is known only from some isolated braincases (Patter- son, 1970). Spratticeps and Santanaclupea provide essentially all the information available about cranial anatomy in Early Creta- ceous clupeomorphs, and it is instructive to compare the braincase in these two taxa. There is a large prootic bulla in Spratticeps, but no pterotic bulla. The pterotic bulla is well developed in Santanaclupea, and there is an auditory fenestra ventrally. The supraoccipital separates the parietals in both taxa; in Spratticeps all three bones contain the supratemporal canal, whereas in Santan- aclupea the canal lies external to the supraoccipital. The preepiotic fossa is extremely small Fig. 7. Reconstruction of the skeleton in San- in Spratticeps, and is much better developed tanaclupea silvasantosi. in Santanaclupea. The skull roof of Spratti- ceps is perforated by small openings for var- 12 AMERICAN MUSEUM NOVITATES NO. 3076 ious branches of the supraorbital sensory ca- thyiformes (his putative successive sister taxa nal, but the skull roof is more complete than to the clupeiformes) is speculative. Spratti- in living clupeoids. In Santanaclupea the re- ceps may be the most primitive taxon that cessus lateralis is apparently better developed can be reliably considered a clupeomorph (it and there are large openings for sensory ca- has a swim-bladder diverticulum entering the nals in the skull roof. The temporal foramen prootic, which houses a large bulla; Patter- is exposed, as in living clupeoids, and is not son, 1970: fig. 6). More parts of the skeleton covered over by the frontals and parietals as ofSantanaclupea are known, and it certainly in Spratticeps. The overall pattern of open- represents a clupeomorph even in a more re- ings for sensory canals thus is more advanced stricted context than that advocated by Gran- in Santanaclupea than in Spratticeps. de (1985). The hyomandibular facet in Spratticeps is Santanaclupea possesses two of Grande's single, and is framed by the sphenotic pos- (1985) three clupeiform characters (presence teromedially, the prootic laterally, and the of a recessus lateralis; parietals completely pterotic anteromedially (Patterson, 1970: fig. separated by supraoccipital), but it primi- 4B). The facet is double in Santanaclupea tively retains the "Beryciform foramen." and many extant clupeomorphs. In Santan- "Higher" clupeiforms appear to have lost this aclupea the prootic does not help to frame foramen independently of other teleosts. the hyomandibular facets. A comparison of Spratticeps lacks the first ofthese characters, the hyomandibular articulations in modern but shares the second and is unknown in the clupeomorphs has not been made, but it is third. On this admittedly slender basis, both noted in passing that there are also two ar- taxa are clupeiforms, but Santanaclupea is a ticular condyles on the hyomandibula of more derived clupeiform than Spratticeps. Denticeps (Greenwood, 1968: 240; the sister Within this group (= Clupeiformes of group of clupeiforms according to Grande, Grande) Santanaclupea cannot be convinc- 1985). Here, the anterior facet is formed ingly placed within either the Denticepitoidei mainly in the sphenotic, but partly in the or the Clupeoidei without special pleading for prootic, while posterior facet is formed in the homoplasy (for example, if Santanaclupea is pterotic. Comparison with primitive fossil an engrauloid lacking clupeiform characters). and living teleosts (e.g., Elops, Pholidophorus Santanaclupea fails to meet Grande's four bechei, Leptolepis coryphaenoides; Patterson, criteria (all unknown in Spratticeps) for in- 1973: fig. 10; 1975: figs. 62, 89) suggests that clusion in the Clupeoidei (first uroneural fused the teleostean hyomandibular facet is prim- to first preural; first ural reduced; parhypural itively single and formed by three bones, the separate from first preural centrum; lateral pterotic, sphenotic, and prootic. Comparison line scales lost). Santanaclupea neverthless with modern Amia and gars is relatively un- agrees with one higher category of clupeoid informative, because the facet (although sin- fishes (engrauloids) in its suspensorial ar- gle) is largely formed in cartilage. A double rangement and possibly its mesethmoid ar- articulation may have developed among clu- rangement. In engrauloids, the gape of the peomorphs as a corollary to the pterotic bul- mouth is elongate and the snout region (via la, although it also occurs in other fishes (no- the mesethmoid) projects anterior to the low- tably among ostariophysans) where bullae are er jaw (when the mouth is closed). In San- never developed (Arratia, 1992: 24). tanaclupea, the jaws are elongate, although It is concluded that Santanaclupea dis- as preserved they seem to project slightly be- plays more clupeomorph synapomorphies yond the snout. This state of affairs invites than Spratticeps, and of the two is phyloge- further consideration of clupeomorph rela- netically closer to Recent clupeomorphs. Both tionships and an appraisal of the characters taxa are more convincing clupeomorphs than used to defend Grande's hypothesis of clu- Erichalcis (cf. Forey, 1975; Grande, 1985). peomorph phylogeny. C. Placement within clupeomorphs. As Fusion ofthe first preural centrum with the Grande (1985: 253) admitted, the presence first uroneural occurs not only in clupeoids, of an otophysic connection of clupeomorph but is common in ostariophysans and also type in Erichalcis, Armigatus, and Ellimich- occurs in some teleosts (Fink and Fink, 1981; 1 993 MAISEY: NEW CLUPEOMORPH FISH 13

> ., + wP * 1cm t. i -- _ .X!4vv.. ofi tAd

Fig. 8. Abdominal region in the holotype of Santanaclupea silvasantosi showing remains of prey, a small Tharrhias (arrowed); (A) caudal skeleton; (B) head and vertebral column (reversed).

perhaps independently ofclupeoids, Schultze osteoglossomorphs, elopomorphs, and some and Arratia, 1989). Modem clupeoids have euteleosts, although these bones are fused in lost all but the first one or two lateral line cladistically primitive members ofthese taxa, scales, but more primitive clupeomorphs all e.g., Lycoptera, Hiodon, Elops, Chanos retain a complete series (Grande 1985: 259); (Schultze and Arratia, 1988; Arratia, 1991). the clupeoid condition could have arisen in- Moreover, these bones are not separate in dependently in different lineages, although it two genera of dussumieriine clupeoids seems unlikely. Separation ofthe parhypural (Grande, 1985: 260). and first preural is also widespread among Grande's (1985) characters and hypothesis 14 AMERICAN MUSEUM NOVITATES NO. 3076 of clupeomorph relationships are founded I agree with Silva Santos (1991) that "L." largely on extant forms, and information on diasii differs from "typical" Leptolepis, and fossils is scarce. Santanaclupea displays a its placement within another genus is justi- combination of characters that conflict with fied. From well-preserved material at my dis- the interpretation of some of his characters posal, I suspect that Santanichthys is related and therefore with his hypothesis, but con- to Clupavichthys Gayet, 1989, perhaps rep- siderably more data from other fossil clu- resenting a distinct species. peomorphs are required before an alternative phylogenetic hypothesis can be presented. CONCLUSIONS D. Note on "Leptolepis" diasii. This small fish also occurs in concretions from the San- It is concluded that Santanaclupea is a clu- tana Formation (Silva Santos, 1958). Its re- peomorph, and its cranial morphology is more lationships were discussed briefly elsewhere advanced toward the modern clupeomorph (Maisey, 1991a: 273); it differs from Lepto- condition than in Spratticeps. Within clu- lepis sensu Nybelin (1974) and Patterson and peomorphs, Santanaclupea can be referred Rosen (1977) in its caudal fin morphology, to the Clupeiformes on the basis of two of particularly in having a stegural of euteleos- the three clupeiform characters recognized by tean type formed by fusion between the first Grande (1985). Its placement within pres- preural neural arch and first uroneural (fig. ently recognized groups of clupeiform fishes 6D). is more problematic. One ambiguous apo- Recently, Silva Santos (1991: 30) erected morphic character (orientation of the sus- a new genus (Santanichthys) for "L." diasii, pensorium, especially the hyomandibula) and suggested it has "clupeiform" affinity. suggests engrauloid affinity within the Clu- Our respective descriptions and illustrations peoidei. At this time it is impossible to de- of the caudal skeleton of this fish are pro- termine whether its mesethmoid/vomer/pre- foundly different (cf. Maisey, 199 la: 273; Sil- maxilla arrangement is also engraulidlike. If va Santos, 1991: fig. 10). Santanaclupea is an engrauloid, it would be The specimen on which Silva Santos based by far the oldest one known (the next being his interpretation is DNPM 683-P, which I the Miocene Engraulis tethensis; Grande and examined in September 1992 after complet- Nelson, 1985). ing most of this study. This specimen agrees in its cranial and caudal morphology with ACKNOWLEDGMENTS Santanaclupea, not with "L. " diasii. In par- ticular, it has an elongate, slender, toothed Thanks are given to Dea Campos for ar- maxilla and dentary, not the short edentulous ranging the loan of the DNPM material to jaws of "L. " diasii. Moreover, it possesses a me at the AMNH, and to Diogenes de Al- vertebral count consistent with that of San- meida Campos for kindly providing facilities tanaclupea and higher than that of "L. " diasii to examine the specimens in Rio de Janeiro (characteristically the latter has only about beforehand, and for giving permission to acid- 30 preural centra, of which 11 are caudal; prepare the DNPM material. The American both fishes have about 19 abdominal verte- Museum funded my visits to Brazil in 1991 brae). The caudal endoskeleton in DNPM and 1992. I acknowledge the kindness ofPro- 683-P agrees with that ofSantanaclupea, and fessor Rubens da Silva Santos in permitting lacks several features found in the caudal fin me to study specimens in his care; I recognize of "L. " diasii (e.g., fused PU I + U + NPU 1 his achievements by naming this new genus articulating with the parhypural and first two after him, and apologize if some remarks I hypurals; stegural present). Unfortunately it have made should appear critical; it was not was not possible to include a photograph of my intention. Preparation ofspecimens at the DNPM 683-P in the present work. Although American Museum was carried out by R. the holotype of "L." diasii (DNPM 647-P) Evander and illustrations were made by L. has not been prepared in acid, its morphology Meeker and C. Tarka. I thank G. Arratia, L. clearly differs from that ofSantanaclupea and Grande, and M. Wilson for their constructive it represents a different taxon. reviews of this paper. 1993 MAISEY: NEW CLUPEOMORPH FISH 15

REFERENCES Arratia, G. 1991. The caudal skeleton ofJurassic teleosts, tana Fossils; an illustrated atlas, pp. 272- based on both fossil and extant teleosts. 273. Neptune, NJ: TFH Publications. In M.-M. Chang, H. Liu, and G. Zhang 199 1b. Undetermined Santana clupeiform. In (eds.), Early vertebrates and related J. G. Maisey (ed.), Santana Fossils; an problems in evolutionary biology, pp. illustrated atlas, pp. 297-302. Neptune, 249-340. Beijing: Science Press. NJ: TFH Publications. 1992. Development and variation of the sus- Nybelin, 0. pensorium of primitive catfishes (Te- 1974. A revision ofthe leptolepid fishes. Acta leostei: Ostariophysi) and their phylo- Regiae Societatis Scientiarum et Litte- genetic relationships. Bonn. Zool. rarum Gothoburgensis, Zool. 9: 202 pp. Monogr. nr. 32: 1-149. Patterson, C. Cavender, T. 1970. A clupeomorph fish from the Gault 1966. The caudal skeleton of the Cretaceous (Lower Cretaceous). Zool. J. Linn. Soc. teleosts Xiphactinius, Ichthyodectes and 49(3): 161-182. Gillicus, and its bearing on their rela- 1973. Interrelationships of holosteans. In P. tionship with Chirocentrus. Occas. Pap. H. Greenwood, R. S. Miles, and C. Pat- Mus. Zool. Univ. Michigan 650: 1-15. terson (eds.), Interrelationships of fish- Fink, S. V., and W. L. Fink es. Zool. J. Linn. Soc. 53(suppl. 1): 233- 1981. Interrelationships of the ostariophysan 305. fishes (Teleostei). Zool. J. Linn. Soc. 1975. The braincase of pholidophorid and 72(4): 297-353. leptolepid fishes, with a review of the Forey, P. L. actinopterygian braincase. Philos. Trans. 1975. A fossil clupeomorph fish from the Al- R. Soc. B 269: 275-579. bian of the Northwest Territories of Patterson, C., and D. E. Rosen Canada, with notes on cladistic rela- 1977. Review of ichthyodectiform and other tionships of clupeomorphs. J. Zool., Mesozoic teleost fishes and the theory London 175: 151-177. and practice of classifying fossils. Bull. Gayet, M. Am. Mus. Nat. Hist. 158(2): 81-172. 1989. Note preliminaire sur le materiel pa- Pons, D., P. Y. Berthou, and D. A. Campos leoichthyologique eocr6tacique du Rio 1990. Quelques observations sur la palynolo- Benito (sud de Bata, Guinee Equatori- gie de l'Aptien superieur et de l'Albien ale). Bull. Mus. Natl. Hist. Nat., Paris, du Bassin d'Araripe. In I Symposio sobre 4 ser., 11: 21-31. a bacia do Araripe e bacias interiores Grande, L. do Nordeste DNPM: 241-252. 1985. Recent and fossil clupeomorph fishes Schultze, H.-P., and G. Arratia with materials for revisions of the sub- 1988. Reevaluation of the caudal skeleton of groups ofclupeoids. Bull. Am. Mus. Nat. actinopterygian fishes: Hiodon, Elops Hist. 181(2): 231-372. and Albula. J. Morphol. 195: 257-303. Grande, L., and G. Nelson 1989. The composition ofthe caudal skeleton 1985. Interrelationships of fossil and recent of teleosts (Actinopterygii: Osteichthy- anchovies (Teleostei: Engrauloidea) and es). Zool. J. Linn. Soc. 97: 215-241. description of a new species from the Silva Santos, R. da Miocene of Cyprus. Am. Mus. Novi- 1958. Leptolepis diasii, novo peixe fossil da tates 2826: 16 pp. Serra do Araripe, Brazil. Notas Prelim. Greenwood, P. H. Estud., Rio de Janeiro 108: 1-15. 1968. The osteology and relationships of the 1991. Fosseis do nordeste do Brasil; Paleoic- Denticepidae, a family of clupeomorph tiofaunula da Chapada do Araripe. fishes. Bull. B.M.(N.H.) 16(6): 215-273. Universidade do Estado do Rio de Ja- Maisey, J. G. neiro, 62 pp. 199 la. "Leptolepis"' In J. G. Maisey (ed.), San- Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates, Bulletin, and Anthropological Papers published during the last five years are available free of charge. Address orders to: American Museum of Natural History Library, Department D, Central Park West at 79th St., New York, N.Y. 10024.

THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER.