jAMS HARVARD UNIVERSITY Library of the Museum of Comparative Zoology OF THE Museum of Comparative Zoology The Systematics and Evolution of the Subsaharan Africa, Seychelles, and Mauritius Scincine Scincid Lizards ALLEN E. GREER \x HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS, U.S.A. VOLUME 140, NUMBER 1 JUNE 26, 1970 PUBLICATIONS ISSUED OR DISTRIBUTED BY THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Beeviora 1952- Memoirs 1864-1938 Johnsonia, Department of Mollusks, 1941- Occasional Papers on Mollusks, 1945- Other Publications. Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine. Reprint, $6.50 cloth. Braes, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of In- sects. $9.00 cloth. Creighton, W. S., 1950. The Ants of North America. Reprint, $10.00 cloth. Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 12, 14, 15. (Price list on request.) Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Teredinidae (Mollusca: Bivalvia). $8.00 cloth. Whittington, H. B., and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948. ( Complete sets only. ) Publications of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138, U. S. A. © The President and Fellows of Harvard College 1970. THE SYSTEMATICS AND EVOLUTION OF THE SUBSAHARAN AFRICA, SEYCHELLES, AND MAURITIUS SCININE SCINID LIZARDS ALLEN E. GREER ABSTRACT Skull osteology and external morphology form the basis for a review of the relation- ships of the seineine scincid lizards of sub- saharan Africa, the Seychelles and Mau- ritius. There appear to be three natural groups in this area. ProsceJotes and Sepsina constitute the most primitive group in sub- saharan Africa, while Scelotes, Melanoseps, Scolecoseps and Typhlacontias form a second, perhaps more advanced group. The scincines of the Seychelles comprise two taxa worthy of generic rank (Pamelaesc in- cus and Janetaescincus, new genera) and together with the monotypic Mauritius genus Gongylomorphus form a third natural group. Evolutionary and zoogeographic relationships within each of the three groups are discussed in some detail, but only a passing attempt is made to relate them with each other or with the large, but virtually unknown complex of scincines on Madagas- car. In general, this complex seems to have more in common with the mainland Prosce- Jotes and Sepsina and the three genera of the Seychelles and Mauritius than with the mainland Scelotes and its relatives. In addi- tion to the systematic, evolutionary, and zoogeographic discussions there is also a key to the genera of scincines inhabiting mainland Africa south of the Sahara. INTRODUCTION Evidence has been presented elsewhere (Greer, 1970) demonstrating that at the subfamily level the Scincines are imme- diately ancestral to the other three sub- families of skinks. With the exception of the largest genus (Eumeces, 46 species) in the subfamily and the monotypic Neoseps of Florida, the Scincinae are entirely Old World in distribution and, again with the exception of the widespread Eumeces, show a relict distribution in southcentral and eastern Asia (Fig. 1). For example, the only seineine, with the exception of Eu- meces, in eastern Asia is Bracliymeles ( 13 species) in the Philippines. As one moves west through Asia, no other scincines are encountered until one reaches India, where the monotypic Barkudia is known from the regions around Chilka Lake and Calcutta. Further south in India there is a single species of Sepsophis in the central and southern part of the subcontinent and two genera, Nessia (8 species) and Chalcido- seps ( 1 species ) , on Ceylon. Moving still further west, it is not until one reaches Southwest Asia and the Medi- terranean area that one encounters widely distributed genera with many species: e.g., Ophiomorus (9 species); Scincus (12 spe- cies ) ; and Chalcides ( 14 species ) . And it is only south of the Sahara Desert in Africa, Madagascar, and the islands of the western Indian Ocean that the scincines become an important part of the skink fauna ( 76 of the 136 species of non-Eumeces scincines occur in this area). Two of the other three subfamilies are also found in subsaharan Africa. The Acon- tinae with approximately 15 species and the Bull. Mus. Comp. Zool., 140(1): 1-24, June, 1970 2 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 Figure Distribution of the Scincinae, exclusive of Eumeces. Feylininae with 4 species are undoubtedly derived from scincines in Africa, which indicates, along with the present number of species and their distribution in subsaharan Africa, Madagascar, and the islands of the western Indian Ocean, that the scincines have been in subsaharan Africa for much, if not most, of their evolutionary history. The reasons for the relict distribution of the scincines in south and east Asia and their prevalence in Southwest Asia, Africa, and Madagascar are undoubtedly complex, but may be due in part to the evolution and radiation of the Lygosominae in Southeast Asia and the Australian Region. The lygo- somines are undoubtedly derived from the Scincine Lizards • Gia i scincines (Greer, 1970), and are morpholog- ically the most advanced skinks. This group is most numerous and diverse in Southeast Asia and the Australian Region, and its expansion from this area of origin may account in part for the relict distribution of the scincines in south and east Asia. In Southwest Asia, Africa, and Madagascar, the area of the Old World furthest from their area of origin, the lygosomines are fairly well represented by species, but they are not morphologically diverse, i.e., there are not many genera. Presumably the lygo- somines are only recent arrivals in this area, and have not yet swamped their ancestral scincine relatives. Perhaps if we could return in several million years, the scincines would show a relict distribution in Africa, Madagascar, and the west Indian Ocean islands as they do in southern and eastern Asia today. Among scincines, relationships have re- mained most obscure in that area where extinction (due to competition from the lygosomines?) has done less to sharpen the differences between taxa, i.e., subsaharan Africa, Madagascar, and the islands of the western Indian Ocean.1 The purpose of this paper is to delimit some of the scincine taxa in this area more clearly than has been the case in the past and to discuss their relation- ships. As much of the data from this study are derived from comparative skull osteol- ogy, the Malagasy scincines, most of which are as yet too poorly known in collections to allow a skull to be prepared, will be largely excluded from the formal taxonomic section of the paper, and their relationships with other scincines will be discussed only in a general way. This is unfortunate, as Madagascar, with its possibilities as a refuge for groups facing extinction from new com- petitors on the mainland, undoubtedly holds many answers to important questions of scincine evolution. It is doubly unfortunate that Madagascar should be the repository of this information, as it is unlikely to be much more accessible' to collectors in the near future than it has been in the past. Madagascar, then, will probably be the "black box" of our analysis of scincine evolu- tion for a long time to come. The paper is divided into two major parts. In the first part taxonomic groups are defined, discussed, and defended, and in the second part the evolution and zoo- geography of these groups are discussed. The reason for this format is simply that the evolution of groups cannot be discussed without knowing what the groups are that are evolutionary significant. In addition to this, I have provided a key to the scincine genera of subsaharan Africa. SYSTEMATICS OF THE SUBSAHARAN AFRICA, SEYCHELLES, AND MAURITIUS SCINCINAE Since the appearance of Boulcnger's ( 1887 ) third volume of the Catalogue of the Lizards in the British Museum, several authors have discussed the relationships of the scincines of subsaharan Africa, Mada- gascar, and the islands of the western In- dian Ocean (Hewitt, 1921, 1927, and 1929; Barbour and Loveridge, 1928; Smith, 1935; de Witte and Laurent, 1943; Loveridge, 1957). All have relied heavily or exclu- sively on external characters and one inter- nal character, namely, whether or not the palatine and pterygoid bones meet along the midline of the secondary palate.1' Thus each successive discussant has had essen- tially the same characters available to him that were available to his predecessors. The present study adds new data from the comparison of whole skulls of most of the African, Seychelles, and Mauritius species8 (see Specimens Examined section of paper). 1 The other side of this coin, however, is the preservation of intermediate forms, which makes it easier to understand the evolutionary relationships. 2 Unfortunately, the relationships of the palatine and pterygoid bones have often been evaluated without removing the overlying buccal mucosa. This has led to misinterpretations. 3 Species of which whole skulls have been exam- ined are marked with an asterisk ( * ) in the account of the genera. 4 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 Scincine Lizards • Gret The relevant work of previous authors is discussed under the appropriate taxonomic grouping. GENERA OF SUBSAHARAN AFRICA Proscelotes de Witte and Laurent Proscelotes de Witte and Laurent, 1943, Mem. Mus. Roy. d'Hist. Nat., 2me ser., fasc. 26, p. 13 (Type species, Scelotes eggeli Tornier, 1902, by original designation). Diagnosis. Skull characters: Palatine bones closely apposed or meeting along midline; palatal rami of pterygoids sep- arated and diverging posteriorly; pterygoid teeth absent (Fig. 2). Postorbital bone present and relatively well developed; su- pratemporal arch strong and fenestra well developed; 17 to 22 maxillary teeth. External characters: Interparietal small (except in eggeli), not touching supraocular scales; a pair of supranasals meeting behind rostral; external ear opening present; 5 fin- gers and toes (i.e., digital formula, 5-5). Distribution. Lowlands of Mozambique (aenea); from Inyanga south to Melsetter in Rhodesia and Mlanje Mt. in Malawi (arnoldi) and Usambara Mts., Tanzania {eggeli). Species. Aenea* Barbour and Loveridge, 1928; arnoldi* Hewitt, 1932; eggeli" Tor- nier, 1902. Mode of reproduction. Both arnoldi and eggeli are known to be live bearing, arnoldi producing 5 young ( 1 female observed — FitzSimons, 1943: 205) and eggeli, 3-4 young (3 females observed — Barbour and Loveridge, 192S: 166; and Greer, personal observation ) . There is no information avail- able on reproduction in aenea. Discussion. With the exception of eggeli and uluguruensis, de Witte and Laurent ( 1943 ) divided all the subsaharan mainland African scineines, and the genera of what are now considered two other subfam- ilies, into two "phyla." One "phylum" was characterized by an interparietal narrower than the frontal and included the genera (as they conceived them) Sepsina, Dume- rilia, Acontias, Acontophiops, and Typhlo- saurus. The last three of these five genera are, however, now placed in the subfamily Aeontinae (Greer, 1970), and need not con- cern us here. The other "phylum" of de Witte and Laurent was characterized by the inter- parietals being wider than the frontal. This group included the genera Herpetosaura, Scelotes, Scolecoseps, Fitzsimonsia, Typhla- contias, Feylinia, and Chabanaudia. The last two genera are now considered mem- bers of the subfamily Feylininae (Greer, 1970). Two mainland African scineines, eggeli and uluguruensis, however, have frontals that are constricted anteriorly by the supra- occulars as in certain Madagascar scineines but not as in any other mainland subsaharan African scineines. This character, along with the fact that the two species are primi- tive in retaining 5 fingers and toes and seem to have interparietals intermediate in size between the sizes of the interparietal of the two "phyla," suggested to these authors that eggeli and uluguruensis formed a relict group ancestral to the two mainland "phyla." The status of these two species was empha- sized by their being placed in a new genus of their own — Proscelotes. The interparietal scale is indeed an im- portant taxonomic character in the Scin- cinae. Unfortunately, however, de Witte and Laurent chose a poor description of the character in the relative size of the scale. A characteristic of the interparietal that is more important, and is reflected less clearly in the size of the scale, is the position of the interparietal relative to the supraoculars; Figure 2. Ventral view of the secondary palate in the three species of the genus Proscelotes. Upper left: P. arnoldi MCZ 55145; upper right: P. eggeli MCZ 24217; bottom: P. aenea MCZ 18709. Abbreviations for this and other figures: p, pala- tine; e, ectopterygoid; pt, pterygoid. Not drawn to scale. 6 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 Figure 3. Ventral view of the secondary palate in three species of Sepsina. Upper left: S. bayoni BM 1967.80; upper right: S. tetradactyla MCZ 42885; bottom: S. angolensis FMNH 142793. Abbreviations as in Fig. 2. Drawn to scale. that is, whether the interparietal is in eon- taet with the supraoculars (wider than the frontal) or not in contact with the supra- oculars (narrower than the frontal). This character state does not cut across group- ings based on skull characters (see diag- noses ) and serves to divide the scincines of subsaharan Africa along essentially the same lines as those proposed by de Witte and Laurent ( 1943 ) . With this interpretation of the interpa- rietal, de Witte and Laurent's two species of Proscelotes can be easily assigned to one of the two "phyla," and on skull characters Scincine Lizards • C can be shown to be members of species groups within these "phyla." Thus eggeli, with a large interparietal, but one not touching the supraoculars, belongs to one "phylum," whereas uluguruensis, with the interparietal touching the supraoculars, be- longs to the other "phylum." Within the "phylum" characterized by the interparietals not touching the supra- oculars, eggeli is, on the basis of its skull morphology, similar to two other species (aenea and arnoldi) and forms with these a natural group. The diagnostic characters of this taxon warrant generic recognition. The available name is Proscelotes, of which eggeli is the type species. In showing a distinct postorbital bone, i.e., unfused to the postfrontal, and a well- developed supratemporal arch, Proscelotes does seem to be rather primitive, as sug- gested by de Witte and Laurent ( 1943 ) . The retention of 5 fingers and toes in the three species also supports this supposition. Sepsina Bocage Sepsina Bocage, 1866, J. Acad. Sci. Lisboa, vol. 1, p. 62 (Type species, Sepsina angolensis Bocage, 1866, by monotypy). Dumerilia Bocage, 1866, J. Acad. Sci. Lisboa, vol. 1, p. 63 (Type species, Dumerilia bayonii Bocage, 1866, by monotypy). Sepsina {Rliinoscincus) W. Peters, 1874, Monats- ber. Ak. Wiss. Berlin, p. 373 (Type species, Sepsina (Rliinoscincus) teiradactyla W. Peters, 1874, by monotypy). Scincodipus W. Peters, 1875, Monatsber. Ak. Wiss. Berlin, p. 551 (Type species, Scincodipus congi- cus W. Peters, 1875 = Sepsina bayonii Bocage, 1866, by monotypy). Diagnosis. Skull diameters: Palatine bones widely separated along the midline; palatal rami of pterygoids expanded medi- ally with a tendency toward emargination posteriorly; pterygoid teeth present (Fig. 3). Postorbital bone present and relatively well developed; supratemporal arch strong and fenestra well developed; 12 to IS maxillary teeth. External characters: Interparietal small, not touching supraocular scales; a pair of supranasals meeting behind rostral; external ear opening present; digits 4-4 or fewer. Distribution. Northern Southwest Africa northward through Angola (angolensis) and Cabinda (bayoni only) into the southeast part of the Democratic Republic of the Congo (angolensis) through the southeast- ern and eastern sections of the Democratic Republic of the Congo to eastern Tanzania and Malawi (tetradactyla). From comparison of the distribution of Sepsina with the distribution of the main vegetation types of Africa (Moreau, 1966: 17, fig. 3), it is clear that Sepsina is confined primarily to woodlands (moist and dry), savanna, and steppe, but is excluded from the lowland evergreen forest of west equa- torial Africa. Species. Angolensis* Bocage, 1866; bay- oni* Bocage, 1S66; tetraclacti/la* Peters, 1874. Incertae sedis. Alberti Hewitt, 1929; copei Bocage, 1873. On external characters Proscelotes can be distinguished from Sepsina only on the basis of 5 fingers and toes in the former and no more than 4 fingers and toes in the latter. Digital formulas, however, are so notori- ously variable in scincid genera that I hesi- tate to use such a character in assigning species like alberti and copei, for which no skull has been available, to genera that are diagnosed primarily on skull characters. For the moment, however, the digital formula is our only clue to the relationships of alberti and copei, and with only 4 fingers and toes, these two species are considered as most likely belonging in Sepsina. Mode of reproduction. Nothing is known about the mode of reproduction in Sepsina. Discussion. De Witte and Laurent (1943) resurrected the generic name Dumerilia Bocage, 1866 for bayoni on the grounds that the frontal in this species is wider than it is long, unlike the other species of scincines of the "phylum" consisting of those species with the interparietal narrower than the frontal (essentially the genera Proscelotes and Sepsina as construed here). 8 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 Table 1. Interspecific variation in several characters of the genus Scklotes. See text for discussion. Species Digital formula Ear opening: present (+) or absent ( — ) Herpetosaura (H) or Scelotes ( S), fide de Witte and Laurent Palatines meet ( + ) or not ( — ) Postorbital present ( + ) or absent ( — ) Supratemporal fenestra ca pens is 5-5 + minute H mira 5-5 + small H — + small + minute uluguruensis 5-5 + minute H1 + — ■ + small coffer 3-3 — H + + minute — ■ limpopoensis 3-3 + small H — + small + small bidigittata 0-2 + minute H — — + minute bi))cs 0-2 + minute S — + minute kasncri 0-2 + small S brevipes 0-1 + minute H + + small + minute gronovi 0-1 + minute S + + thin — guentheri 0-1 — H anguina 0-0 — H — i T minute — arenicolor 0-0 — S — + minute bicolor 0-0 — S inornatus 0-0 + minute H 1 Placed in Proscelotes by de Witte and Laurent (1943), but on the criterion of "frontal longer than wide" for distin- guishing Herpetosaura from Scelotes the species is similar to the former genus. This action seems undesirable since the broad frontal of bayoni is probably an adap- tation to burrowing habits, which, on the basis of other features of the external anat- omy (snout more wedge-shaped and digits reduced to the greatest extent: no forelimbs and a single styliform hindlimb1 ) , seem to be more well developed in bayoni than in the other species of scineines with the small interparietal south of the Sahara. In the genus Scelotes there is a similar trend to- ward the development of a relatively wider frontal in those species that, on the basis of other morphological characters, seem to be more highly adapted to a burrowing life ( see page 10 and Table 1 ) . Although the two taxa of scineines with 1 Although de Witte and Laurent (1943) de- scribe the monotypic genus Dumerilia as "pas de membres posterieurs," the two specimens I have seen ( BM RK 1907.80 and MCZ 27098) have styli- form hindlimbs and even Bocage (1866) in his description of the species states, "membres pos- terieurs mediocres, en forme de stylets simples, deprimes." the small interparietal ( which fails to touch the frontal) south of the Sahara Desert in mainland Africa, i.e., Proscelotes and Sep- sina, can be recognized on the basis of external characters (Proscelotes with 5 fin- gers and toes and Sepsina with 4 fingers and toes or fewer), it should be emphasized that the more important characters distinguish- ing the two genera are in the skull. The discrete, well-developed postorbital bone and the pterygoid teeth of Sepsina are definitely primitive characters. Sepsina, in fact, is the only mainland genus of scineines south of the Sahara to retain pterygoid teeth. Scelotes Fitzinger Scelotes Fitzinger, 1826, Neue Class. Rept., pp. 23, 53 (Type species, Bipes anguineus Menem, 1820 = Scelotes bipes Linnaeus, 1766, by monotypy). Zygnis (not Oken, 1816 or Fitzinger, 1826) Wagler, 1830, Nat. Syst. Amphib., p. 160 (Type species, Anguis bipes Linnaeus, 1766, by monotypy). ? Herinia Gray, 1838, Ann. Mag. Nat. Hist., vol. 2, p. 332 (Type species, //. capensis Gray, 1838 = ?, by monotypy). Figure 4. Ventral view of the secondary palate in three species of Scelotes. Upper left: S. brevipes MCZ 21237; upper right: S. arenicolor MCZ 14205; bottom: S. mira MCZ untagged specimen. Abbreviations as in Fig. 2. Drawn to scale. Scincine Lizards • Gre Scelotes brevipes 10 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 Lithophilus (not Frohlich, 1799) A. Smith, 1849, Illus. Zool. S. Africa, Rept, App., p. 12 (Type species Lithophilus inornatus A. Smith, 1849, by subsequent designation of de Witte and Laurent, 1943). Herpetosaura W. Peters, 1854, Monatsber. Ak. Wiss. Berlin, p. 619 (Type species, Herpetosaura arenicolor Peters, 1854, by monotypy). Sepomorphus W. Peters, 1861, Monatsber. Ak. Wiss. Berlin, p. 422 (Type species, Sepomorphus coffer Peters, 1861, by monotypy). Herpetoseps Boulenger, 1887, Cat. Lizards Brit. Mus., vol. 3, p. 416 (Type species, Herpetoseps anguinus Boulenger, 1887, by monotypy). Diagnosis. Skull characters: Palatine bones meeting or closely apposed on mid- line; palatal rami of pterygoids separated medially and diverging posteriorly; ptery- goid teeth absent ( Fig. 4 ) . Postorbital bone present, but small to minute, or absent; supratemporal arch weak, fenestra oblite- rated by apposition of bones of arch with parietal bone; 11 to 23 maxillary teeth, but only uluguruensis with 22 to 23 maxillary teeth, all other species with 11 to 19 teeth. External characters: Interparietal large, touching supraocular scales; external ear opening present or absent; a pair of supra- nasals meeting behind rostral (fused only in some bipes); digital formula 5-5 to 0-0 (see Table 1). Distribution. Africa south of an area from southern South-west Africa (cap- ensis), east through extreme southern Rhodesia {limpopoensis from Beitbridge), and southern Mozambique. Most of the species are found in southern and eastern Africa. Species. Anguina* Boulenger, 1887; are- nicolor* W. Peters, 1854; hidigittata* Fitz- Simons, 1930; bicolor A. Smith, 1849; bipes* Linnaeus, 1766; brevipes* Hewitt, 1925; coffer* W. Peters, 1861; capensis A. Smith, 1849; gronovi* Daudin, 1802; guentheri Boulenger, 1887; inornata A. Smith, 1849; kasneri FitzSimons, 1839; limpopoensis* FitzSimons, 1930; mira* Roux, 1907; ulugu- ruensis* Barbour and Loveridge, 1928. Mode of reproduction. The mode of re- production is known for six of the1 fifteen species of Scelotes and all are live bearing (available clutch sizes follow the species name): arenicolor — 4 (FitzSimons, 1943: 197); bidigittatus—1 to 2 (Pienaar and Fitz- Simons, 1966: 57); bipes — 2 (FitzSimons, 1943: 191); brevipes— 2 (FitzSimons, 1943: 195); mira— 4 (FitzSimons, 1943: 183); uluguruensis — ? (Barbour and Loveridge, 1928: 167). Discussion. Two characters that show a good deal of interspecific variation in Sce- lotes ( as conceived here ) have been empha- sized in generic diagnoses. One of these characters, the complete absence of hind- limbs (used to distinguish the genera Lithophilus, Herpetosaura, and Herpeto- seps), is most probably correlated with an increasing tendency toward a burrowing way of life, although actual ecological obser- vations are lacking to support this. The reduction of limbs is also correlated with a reduction in the size of the external ear opening, presumably also an adaptation to burrowing habits (Table 1). Whether or not the palatine bones touch along the midline of the palate has also been used as a generic character in this group, but it has long been known that this feature is an interspecifically variable character in Scelotcs (Hewitt, 1921 and 1927; Barbour and Loveridge, 1928; Table 1, this paper). De Witte and Laurent (1943) revived the name Herpetosaura for those species of Scelotes with the frontal longer than wide, leaving Scelotes to include those species with the frontal wider than long. In Table 1 the generic allocation of the species con- sidered here to be Scelotes is given accord- ing to de Witte and Laurent. It can be seen from the Table that a frontal wider than long (the "Scelotes" condition) is, in gen- eral, correlated with other morphological features indicative of a burrowing way of life. In my opinion this use of characters that show evolutionary trends within a taxon to subdivide that taxon into formal tax- onomic groupings, unnecessarily obscures the evolutionary significance of those char- acters for the sake of taxonomic conve- nience. Scincine Lizards • Greet 11 Scolecoseps boulenger Melanoseps occidentals Figure 5. Ventral view of the secondary palate in Sco/ecoseps and Melanoseps. Upper left: Sco/ecoseps boulengeri MCZ 18357; upper right: Melanoseps occldentalis BM 1907. 5. 22. 6A; bottom: M. ater rondoensis MCZ 52487. Drawn to scale. Ab- breviations as in Fig. 2. 12 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 It should be noted here that uluguruensis, a species with a large interparietal touching the supraoculars, was considered a species of Proscelotes by de Witte and Laurent ( see pages 6-7 above). In the diagnostic skull characters, however, uluguruensis is unlike the type species of Proscelotes ( eggeli ) and very much like the other species of Scelotes. Melanoseps Boulenger Melanoseps Boulenger, 1887, Cat. Lizards Brit. Mus., vol. 3, p. 422 (Type species Herpetosaura atra Giinttier, 1873, by monotypy). Diagnosis. Skull characters: Palatine bones separated medially; palatal rami of pterygoids expanded medially and emar- ginated posteriorly; pterygoid teeth absent ( Fig. 5 ) . Postorbital bone absent; supra- temporal arch weak and fenestra obliterated by apposition of bones of the arch with the parietal bone; 10 to 13 maxillary teeth. External characters: Interparietal large, touching supraoculars; a pair of supranasals meeting behind rostral; no external ear opening; limbless. Distribution. Extreme southern Kenya south through Tanzania and Malawi to the Zambesi (ater) and west through the Democratic Republic of the Congo, and extreme northeastern Angola to Cameroon and Fernando Poo (occidental is). In addi- tion, Sternfield (1911) reports a single speci- men of occidentalis from Windhuk, South- west Africa. This locality is, however, some distance from the range indicated by other specimens of occidentalis and is also from a considerably drier area than any other speci- mens. For these reasons the record should perhaps be regarded with suspicion (Mer- tens, 1955). Species. Ater* Giinther, 1873 (six sub- species); occidentalis* W. Peters, 1877. Mode of reproduction. Personal observa- tions on M. ater misukuensis indicate that the species is live bearing. Six gravid fe- males had 2 to 4 (average = 3.0) developing eggs in a clutch. Discussion. The generic distinction of Melanoseps from Scelotes, which has been argued for by most authors ( Boulenger, 1887; Tornier, 1901; Loveridge, 1957; Lau- rent, 1964), is confirmed on the basis of skull morphology: no Scelotes has the medi- ally expanded and posteriorly emarginated palatal rami of the pterygoids seen in Melanoseps (Figs. 4, 5). De Witte and Laurent (1943) regarded Melanoseps as a synonym of Scelotes. How- ever, when Loveridge ( 1957 ) pointed out that Melanoseps is "associated with primary forest, either montane or gallery," a habitat from which Scelotes is largely excluded, Laurent (1964) considered these ecological differences, along with the morphological differences, sufficient evidence for the generic separation of Melanoseps from Scelotes. Melanoseps appears to be a derivative of a Scelotes-\iko ancestor that has evolved primarily in the evergreen montane and lowland forests of central Africa. Scolecoseps Loveridge Scolecoseps Loveridge, 1920, Proc. Zool. Soc. London, p. 159 (Type species, Scolecoseps bou- lengcri Loveridge, 1920, by monotypy). Diagnosis. Skull characters: Palatines widely separated; palatal rami of pterygoids expanded slightly toward midline and deeply emarginated posteriorly, somewhat as in Melanoseps; pterygoid teeth absent ( Fig. 5 ) . Postorbital bone absent; supra- temporal arch weak and fenestra obliterated by apposition of bones in arch with parietal bone; twelve maxillary teeth. External characters: External naris in a large rostral and connected with the posterior edge of the rostral through a horizontal suture somewhat as in Typhla- eontias (Scincinae) or the subfamily Acon- tinae. Interparietal large, touching supra- oculars; a pair of supranasals meeting behind rostral; no external ear opening; limbless. Distribution. Lumbo, Mozambique (bou- lengeri ) and Dar es Salaam to Kilwa, Tanzania (acontias). Scincine Lizards • Grei , 13 Typhlacontias ngamiensis T. brevipes Figure 6. Ventral view of the secondary palate in three species of Typhlacontias. Upper left: T. ngamiensis FMNH 142787; upper right: T. gracilis USNM 159338; bottom: 7". brevipes MCZ 96702. Abbreviations as in Fig. 2. Drawn to scale. 14 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 Species. Acontias Werner, 1912; boulen- geri* Loveridge, 1920. Mode of reproduction. Absolutely nothing is known about the mode of reproduction in the two species of this rare genus. Discussion. I have been able to examine the secondary palate only in the type species of this genus, boulengeri. The palate is similar to that of Melanoseps, but the large rostral seems distinct enough to warrant generic separation from Melanoseps. I have not been able to see the palate in acontias. the other species in the genus. The enlarged rostral of this species is similar to that of Scolecoseps boulengeri, but it is also similar to the rostral of some Tijphla- contias. Since palatal characters are the only basis we would have for changing the generic status of acontias, it seems best to follow current taxonomic opinion ( de Witte and Laurent, 1943; Loveridge, 1957) and retain acontias in Scolecoseps. These two species of skinks are extremely poorly represented in herpetologieal collec- tions, and virtually nothing is known of their ecology. The absence of both an external ear opening and limbs would indicate that the genus is burrowing or at least cryptic in habits. Typhlacontias Bocage Typhlacontias Bocage, 1873, J. Acad. Sci. Lisboa, vol. 15, p. 5 (Type species, Typhlacontias punc- tatissimus Bocage, 1873, by monotypy). Fitzsimonsia de Witte and Laurent, 1943, Mem. Mus. Roy. d'Hist. Nat., 2me ser., fasc. 26, p. 33 (Type species Typhlacontias brevipes FitzSimons, 1939, by monotypy). Diagnosis. Skull characters: Palatine bones only slightly longer than wide, sep- arated medially; palatal rami of pterygoids expanded laterally, but not meeting medi- ally, and emarginated posteriorly (gracilis) or not (brevipes and ngamiensis); pterygoid teeth absent (Fig. 6). Postorbital and jugal bones lacking; supratcmporal arch weak and fenestra obliterated by the apposition of the bones in the supratcmporal arch with the parietal; fixe to six maxillary teeth. External characters: Interparietal large, touching supraoculars; three median, trans- versely enlarged head scales between the rostral and interparietal instead of a pair of supranasals and two median, transversely enlarged head scales; no external ear open- ing; limbless except for T. l)revipcs, which has a rudimentary hind leg. Distribution. Collecting records for the genus are few and widely separated but stretch from southern Angola and South- west Africa east through Botswana to the Upper Zambesi region and extreme south- eastern Rhodesia. This area is covered by the desert, subdesert, and dry woodlands, steppe vegetation tvpes of Moreau ( 1966, fig. 3). Species. Bogcrti Laurent, 1966; l>revipes* FitzSimons, 1938; gracilis* Roux, 1907; ngamiensis* FitzSimons, 1932; punctatissi- mus Bocage, 1873; rohani Angel, 1923. Mode of reproduction. The only informa- tion on the mode of reproduction in this genus is the statement by Laurent (1964: 84) in regard to T. bogerti, which would indicate that the species is live bearing: "Le grand Paratype contient des embryons avanccs dont les yeux sont fort developpes." The italics are mine. Discussion. The large interparietal indi- cates that Typhlacontias is a relative of Scelotes, and like Melanoseps the genus is probably a Scelotes derivative. But unlike Melanoseps, which has evolved in the moist evergreen lowland and montane forest of central Africa, Typhlacontias seems to have evolved in the arid areas of southern Africa. The adaptations of the genus, e.g., absence of an external ear opening, the almost com- plete reduction (brevipes) or absence of the limbs, and the large rostral, seem to be indicative of highly developed burrowing habits, but ecological data by which to judge this supposition are lacking. That the group is not just a composite of indepen- dently evolved Scelotes, well adapted to a burrowing life, is indicated by the large single postrostral scale, the unique propor- tion and relationships of the palatal bones, Scincine Lizards • Gret r the absence of a jugal bone, and the low number of maxillary teeth — features not found in those Scelotes (brevipes, gronovi, guentheri, anguina, arenicolor, bicolor, and inornatus ) similarly adapted in other fea- tures of their morphology to a burrowing life. Typhlacontias brevipes has the external naris situated elose to the posterior edge of the rostral and bordered posteriorly by a small postnasal scale; this is a Scelotes-\ike feature. It also retains minute posterior limbs unlike other Typhlacontias. Because of these features de Witte and Laurent (1943) created a new genus, Fitzsimonsia, for the species. Such special taxonomic recognition seems unjustifiable to me as it obscures evolutionary relationships. Typh- Jacontias brevipes is best viewed as the most primitive, i.e., the most Scelotes-likc, of its genus and relates Typhlacontias directly with Scelotes. T. brevipes possesses the large single postrostral scale and palatal characters of Typhlacontias and lacks the jugal bone as do other Typhlacontias, but it resembles Scelotes in the position of the external naris and the minute hindlimbs. It is interesting to note that among the seincines of subsaharan Africa only Scelotes bipes has postrostral scales (supranasals) that are sometimes fused to form a single median transverse scale. This condition is characteristic of Typhlacontias and is fur- ther evidence that a Scelotes was probably ancestral to Typhlacontias. MAURITIUS AND SEYCHELLES GENERA With the exception of Bermuda, which harbors the endemic Eu nieces longirostris, Mauritius and the Seychelles are the most remote oceanic islands on which seincines are known to occur now or in the recent past. On the Seychelles there are three species (gardinieri, braueri, and veseyfitz- geraldi ) that have always been called either Scelotes or Amphiglossus, generally by authors using these two generic names in their widest, catchall sense. The single scincine on Mauritius (bojeri) and the neighboring islands1 on the bank just north of the' main island has generally been placed in the genus Scelotes or referred to a mono- typie genus, Thyrus. Loveridge (1957) has noted that the generic name Thyrus for the endemic Mauritius scincine is antedated by the name Gongylomorphus. The seincines of the Seychelles and Mau- ritius show certain peculiar features in their skull osteology that have not been encoun- tered in any other scincine species for which skulls are available. The most striking of these features is an extensive secondary palate (Fig. 7) formed by the medial ap- position of the palatine bones (seen in a less well-developed state among seincines only in Proscelotes and some Scelotes) and the palatal rami of the pterygoids ( unknown in any other scincine to date ) . In addition, there are 11 teeth on the premaxillae in the Mauritius and Seychelles seincines, whereas all other seincines have only 10 ( only a few Brachymeles) or fewer (generally fewer than 9 in most other seincines). Slightly less significant, but of value in recognizing groupings of seincines in Africa, is the presence of a large, distinct post- orbital bone (partially fused to the post- frontal in the single skull of braueri exam- ined) and a well-developed supratemporal arch. These similarities, especially for seincines, would be indicative of close relationship in other species groups. But the fact that these species occur on two widely separated oceanic island banks poses a serious zoo- geographic problem and raises the possi- bility of convergence as an alternative ex- planation for the unique similarities among the four species. There is, however, no other evidence suggesting that the similar- ities among these species are convergent. Indeed, the independent evolution of a re- markably advanced secondary palate (for seincines) in these lizards seems less ered- 1 This species is actually known only from sub- fossil material on Mauritius itself ( Hoffstetter, 1945 and 1949), although it still occurs on several small islands just north of Mauritius (Vinson, 1965). 16 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 Gongylomorphus bojeri Figure 7. Ventral view of the secondary palate in Gongylomorphus, Pamelaesc/ncus and Janefaescincus. Upper left: Gon- gy/omorphus bojeri MCZ 46677; upper right: Pame/aesc/ncus gardinieri BM 1910.3.18.91; bottom: Janefaescincus braueri BM 1910.3.18.33. Abbreviations as in Fig. 2. Not drawn to scale. Lble to me than the notion of their close re- lationship in spite of the zoogeographic problem raised by that relationship. For the present, therefore, I am regarding the scincines of Mauritius and the Seychelles as a closely related group with a remark- able distribution. In spite of the remarkable similarities which serve to distinguish them as a group apart from all other scincines, the Seychelles Scincine Lizards • Gn and Mauritius scincines (4 species) them- selves show certain peculiar morphological features that warrant their separation into 6 genera as loll ows: Gongylomorphus Fitzinger Gongylomorphus Fitzinger, 1843, Syst. Rept., p. 22 ( Type species, Gongylus bojerii Dumeril and Bibron, 1839 = Scincus bojerii Desjardin, 1831, by monotypy ). Thyrus Gray, 1845, Cat. Specimens Lizards Coll. British Mus., p. 124 (Type species, Scincus bojerii Desjardin, 1831, by monotypy). Diagnosis. Skull characters: Palatine bones and palatal rami of pterygoids meet- ing medially; palatal rami of pterygoids "squared-off," not emarginated posteriorly as in the following two genera; pterygoid teeth absent (Fig. 7). Postorbital bone distinct, well developed; supratemporal arch well developed; 16 teeth on maxilla and 11 teeth on premaxillae. External characters: Interparietal small, not touching supraoculars; frontoparietals present (absent in all other subsaharan Africa, Madagascar, or west Indian Ocean island scincines); ear opening a horizontal slit; 38 longitudinal scale rows at midbody; digits 5-5. Distribution. At present G. bojeri occurs only on several small islands ( Round, Plate, Coin de Mire, and Forquets) lying on the shallow bank just north of Mauritius. The species is, however, known from subfossil remains on Mauritius (Mare aux Songes), and Vinson ( 1965 ) conjectures that in the past G. ]?ojeri also occured on Reunion. Species. Bojeri* Desjardin, 1831. Discussion. One of the most intriguing questions centering around bojeri is its re- cent extinction on the main island of Mau- ritius while it persists on the small bank- islands just north of the main island. Inter- estingly enough, however, bojeri is not uni- que in being subfossil on Mauritius while persisting on a bank island. Casarea, a monotypic genus of boiid snakes, is also subfossil on Mauritius in the same locality as bojeri (Mare aux Songes), but persists on Round Island in the bank islands ( Hoff- stetter, I960). Similarly the giant skink Didosaurus mauritianus is subfossil on Mau- ritius (Marc aux Songes) while its very close relative (Greer, personal observation), Leiolopisma telfairi, exists only on Round Island. In addition, there are endemic species on the bank islands (Round Island) which are not known today or in the past on Mauritius, but which have probably either evolved in situ or are relict popula- tions of species no longer extant on Mauri- tius, i.e., the gecko Phelsuma guentheri and the monotypic boiid genus Bolyeria. These data strongly suggest that extinc- tion among the endemic reptiles of the Mauritius bank has been greater on the main island than on the bank islands. The reasons for this differential extinction are unknown, but it is perhaps significant that a dramatic differential alteration of the ecology of these islands took place when man and his animals arrived and colonized the main island of Mauritius but, until quite recently, left the small bank islands rela- tively undisturbed. The case is unproved as yet, but man and his animals have already been deeply implicated in the extinction of the Mauritius birds, and their guilt may be equally great when the ease of the reptiles comes to trial. Pamelaescincus new genus1 Type species. Scelotes garclinieri Boulen- ger, 1909, Trans. Linn. Soc, ser. 2, Zool., vol. 12, p. 298. Diagnosis. Similar to Gongylomorphus and the following genus and differing from all other scincine skinks in having the pala- tines and palatal rami of the pterygoids meeting medially to form a broad secondary palate (Fig. 7), and in having 11, instead of 10 or fewer, premaxillary teeth. Among the Seychelles and Mauritius scincines, which appear to be the genus' closest relatives, Tamelaescincus is most similar in palatal characters to Gongylo- 1 The genus is named after Pamela, the older of my two sisters. 18 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 morphus, but differs from this monotypic Mauritius genus in lacking both the fronto- parietal scales and the clear spectacle in the lower eyelid. Pamelaescincus is similar to other Sey- chelles seincines in lacking frontoparietals (which all other subsaharan Africa and Madagascar seincines have) and in having scaly eyelids; but differs from these species in lacking the posterior emargination of the palatal rami of the pterygoids (Fig. 7), in having 5 instead of only 4 fingers, and in having a high midbody scale count (30 to 34 instead of 22 to 24)'. Distribution. The single species in the genus is reported from the following islands in the Seychelles archipelago: Mahe, Pras- lin. Silhouette, and Frigate. Species. Gardinieri* Boulenger, 1909. Janetaescincus^ new genus Type species. Scelotes 1)raueii Boettger, 1896, Zool. Anz., vol. 19, p. 349. Diagnosis. Similar to Gon Mel- anoseps — > Scolecoseps would be a reason- able morphological sequence by which to explain the evolution of these taxa. Typhla- contias, on the other hand, with its one species (brevipes) with a styliform hind- limb, may have evolved independently from a Sce/ores-like ancestor. (3) Finally, a third group, whose genera have been discussed in a formal systematic manner above, consists of the monotypic Mauritius Gongylomorphus, and the two genera (3 species), Pamelaescincus and Janetaescincus, on the Seychelles. The osteological similarities which align these skinks (11 premaxillary teeth, complete secondary palate involving both the pala- tines and pterygoids, large postorbital bone, and supratemporal fenestra) are striking and to me indicate close relationship instead of convergence. Gon(j,ylomorphus appears to be more primitive than Pamelaescincus and Janetaescincus in retaining a pair of frontoparietals scales, and Pamelaescincus, in turn, appears to be more primitive than janetaescincus in lacking the posterior emarginations on the palatal rami of the pterygoids and in having 5 instead of 4 fingers. Zoogeography I now would like to make two very gen- eral comments about the zoogeography of the groups discussed above. First, if the size of the interparietal scale, i.e., whether it touches the supraocular scales or not, is truly indicative of relationship within the seincines of subsaharan Africa, Madagascar, and the west Indian Ocean islands, then we have two major groups with highly sugges- tive distributions. On the one hand, the seincines with the small interparietal are widely distributed throughout the area, whereas the seincines with the large1 inter- parietal are confined to continental Africa. Simply on the basis of the distribution of these two groups, it would be reasonable to argue that the more widespread group, i.e., the group that occurs on Africa, Madagas- Scincine Lizards • Greer 21 car, and the islands of the west Indian Ocean, probably appeared in the area before the group that occurs only on the African mainland. This interpretation has the support of the morphological data. The' large interparietal of Scelotes, Melanoseps, Scolecoseps, and Typhlacontias is unusual in skinks and is probably derived from the small interpa- rietal condition. The reduction and loss of the postorbital bone and the predominate trend toward burrowing habits in this group also imply its specialized evolutionary role. This evidence suggests that perhaps the large interparietal line evolved on continen- tal Africa from the more widespread and less specialized small interparietal line. The second zoogeographic point I wish to make cornerns the Seychelles and Mau- ritius scincines. The single Mauritius genus (Gongylomorplius) and the two Seychelles genera (Pamelaescincus and Janetaescincus) appear to be each others closest relatives. There is a problem, however, in explaining how such closely related taxa can occur on two remote and widely separated oceanic island banks. The easiest explanation for this distribution might be that the three genera are simply relicts from a group of scincines that were once much more widely distributed over Africa and/or Madagascar. The flaw in this argument, however, is the notion that the complete secondary palate which unites the Mauritius and Seychelles scincines seems to be an extremely advanced character and is in no way primitive. That is, the complete secondary palate of these scincines appears to be an innovation uni- que to the scincines of the islands, and as the island banks were probably never con- nected in the past, skinks with this palatal character must have first appeared on one island bank and then dispersed over water to the second island bank. The most likely method of over water dispersal for these skinks is by rafting. The probability of successful rafting is greatly increased by the availability of large rivers to carry detritus out to sea, but at present neither the Seychelles nor Mauritius seems to offer a favorable source of rafts. In the recent, and perhaps more distant past, how- ever, this was not the case. Whereas the volcanic Mauritius bank is relatively small (only slightly larger than the main island itself) and rises from very deep water, the largely granitic Seychelles lie on a very shallow, but relatively extensive (approxi- mately 12,000 square miles), bank and would probably form "an extensive, low archipelago" with the standard Pleistocene lowering of sea levels. Under these circum- stances large tropical rivers may have ex- isted in the Seychelles and would perhaps have greatly increased the probability of dispersal from the Seychelles to Mauritius. If this were the direction of dispersal, we might visualize the primitive Gongylomor- phus or its ancestor rafting from the Sey- chelles to Mauritius and surviving there as a relict while the parental stock became extinct on the Seychelles but was repre- sented by the derived Pamelaescincus and Janetaescincus. A KEY TO THE SCINCINE GENERA OF SUBSAHARAN AFRICA Interparietal scale not touching supraocular scales; postorbital bone present and well de- veloped. Digital formula 5-5; pterygoid teeth lacking _ Proscelotcs Digital formula 4-4 or lower; pterygoid teeth present Sepsina Interparietal scale touching supraocular scales; postorbital bone present but only weakly de- veloped, or lacking entirely. A single postrostral scale (fused supranasals); jugal bone lacking; 5-6 maxillary teeth _ Typhlacontias A pair of postrostral scales (supranasals); jugal bone present; 10-23 maxillary teeth. External naris well within an enlarged postrostral Scolecoseps External naris near or bordering posterior edge of rostral scale. Palatal ramus of pterygoid bone deeply emarginated posteriorly (Fig. 5) Melanoseps Palatal ramus of pterygoid bone smoothly diverging along its postero- medial edge (Fig. 4) Scelotes 22 Bulletin Museum of Comparative Zoology, Vol. 140, No. 1 ACKNOWLEDGMENTS The brunt of the skeletonizing that was necessary for this paper has been borne by the herpetologieal collections of the Mu- seum of Comparative Zoology (MCZ). Dr. Ernest E. Williams, as curator of that collec- tion, has been extremely generous in making the collection available to me and has gone far in helping me obtain specimens not in the museum collections. The following people have provided the "crucial" speci- mens needed for this paper: Mr. Charles M. Bogert and Dr. Richard G. Zweifel of the American Museum of Natural History (AMNH); Miss A. G. C. Grandison of the British Museum of Natural History (BM); Dr. Robert Inger and Mr. Hymen Marx of the Field Museum of Natural History (FMNH); Monsieur Jean Gnibe of the Museum National d'Histoire Naturelle ( MNHN ) ; Dr. James A. Peters and the late Dr. Doris Cochran of the United States National Museum ( USNM ) . Dr. E. E. Williams has read the paper and offered several helpful criticisms. In addi- tion, part of the work for this paper was supported by funds from NSF grant GB 6944 to Dr. Williams. SPECIMENS EXAMINED I have seen the following complete skulls : GONGYLOMORPHUS: bojeri MCZ 46677. GRANDIDIERINA: lineata MNHN 3378. JANETAESCINCUS: braueri BM 1910.3. 18.33. MELANOSEPS: ater MCZ 50955, 52487; occidental is BM 1907.5.22.6A. PROSCELOTES: aenea MCZ 18709; ar- noldi MCZ 55145; eggeli MCZ 24217, 24218, 24220. PAMELAESC INCUS: gardinieri BM 1910. 3.18.91. PYGOMELES: braconnieri MNHN 1715. SCELOTES: anguina MCZ 96791; areni- color MCZ 14205; bidigittata MCZ 96789; bipes BM XVII.2.F; brevipes MCZ 21237; coffer MCZ 96792; gronovi BM 97.5.15.8; limpopoensis MCZ 96906; mira MCZ 96790; uluzuruensis MCZ 24206. SCOLECOSEPS: boulengeri MCZ 18357. SEPSINA: angolensis FMNH 142793, AMNH 40734; bm/oni BM RR 1967.80; tctradactyla MCZ 42885, 47770 (3 speci- mens), 47775, 56963, 56965, 56967, 85536. TYPHLACONTIAS: brevipes MCZ 96702; gracilis USNM 159338; ngamiensis FMNH 142787, 142791 (cleared and stained speci- men ) . VOELTZKOVIA: mira MCZ untagged specimen. Malagasy incertae sedis "Scelotes": astrolabi MCZ 20953; melanura MCZ 11733; splen- did™ FMNH 72086. LITERATURE CITED Barbour, T., and A. Loveridge. 1928. A com- parative study of the herpetologieal faunae of the Uluguru and Usambara Mountains, Tan- ganyika Territory with descriptions of new species. Mem. Mus. Comp. Zool., 50(2): 87- 265. Bocage, J. V. Barboza du. 1866. Beptiles nou- veaux ou peu connus recueillis dans les posses- sions portugaises de l'Afrique occidentale, qui se trouvent au Museum de Lisbonne. J. Sci. Math. Phys. Nat., 1: 57-78. Boulenger, G. A. 1887. Catalogue of the lizards in the British Museum (Natural History). Vol. III. London: Taylor and Francis, xii -f- 575 pp. FitzSimons, V. F. M. 1943. The lizards of South Africa. Transvaal Museum Memoir No. 1, Pretoria, xv + 528 pp. Greer, A. E. 1970. A subfamilial classification of scincid lizards. Bull. Mus. Comp. Zool., 139(3): 151-183. Hewitt, J. 1921. On some lizards and arachnids of Natal. Ann. Durban Mus., 3(1): 3-11. . 1927. Further descriptions of reptiles and batrachians from South Africa. Bee. Al- bany Mus., 3: 371-415. . 1929. On some Scincidae from South Africa, Madagascar and Ceylon. Ann. Trans- vaal Mus., 13(1): 1-8. Hoffstetter, B. 1945. Sur les Scincidae fos- siles. II. — Formes subfossiles de l'lle Maurice. Bull. Mus. Nat'l. Hist. Nat. (Paris), 17(1): 80-86. . 1949. Les reptiles subfossiles de l'lle Maurice. I. Les Scincidae. Ann. Paleont. (Paris), 35: 3-72. . I960. Sur la classification des boi'des de Madagascar et des Mascareignes. Bull. Mus. Nat'l. Hist. Nat. (Paris), 32(2): 131-138. S< [ncine Lizards • Greer Laurent, R. F. 1964. Reptiles et amphibiens de 1' Angola (troisieme contribution). Publ. Cult. Diamang, Mus. Dundo, 67: 1-165. Lovkhidce, A. 1957. Check list of the reptiles and amphibians of East Africa ( Uganda; Kenya; Tanganyika; Zanzibar). Bull. Mus. Comp. Zool., 117(2): 153-362 + xxxvi. MertenSj R. 1955. Die Amphibien und Rep- tilien Siidwestafrikas. Abh. senckenb. naturf. Ges., 490: 1-172. Moreau, R. E. 1966. The bird faunas of Africa and its islands. New York: Academic Press, viii + 424 pp. PlENAAR, U. DE V., AND V. F. M. FlTzSlMONS. 1966. The reptile fauna of the Kruger Na- tional Park. Koedoe, Monograph No. 1 : 1- 223. Smith, M. A. 1935. The fauna of British India, including Ceylon and Burma. Reptilia and Amphibia. Vol. II. — Sauria. London: Taylor and Francis, xiii + 440 pp. Sternfeld, R. 1911. Zur Herpetologie Siidwest- afrikas. Mitt. Zool. Mus. Berlin, 5: 395-411. Tornier, G. 1901. Die Reptilien und Amphibien der deutschen Tiefseeexpedition 1898/99. Zool. Anz., 24(635): 61-66. Vinson, J. 1965. Sur la disparition progressive de la flore et de la fauna de 1 ile Ronde. Proc. Roy. Soc. Arts Sci. Mauritius, 2(3): 247-261. Witte, G. F. de, and R. Laurent. 1943. Con- tribution a la systematique des formes de- gradees de la famille des Scincidae apparentees an genre Scelotes Fitzinger. Mem. Mus. Roy. d'Hist. Nat. Belgique, 2me ser., fasc. 26: 1-44. (Received 7 November 1968.) EuLLetln of the Museum of Comparative Zoology Studies on the Fossil Snake Dinilysia patagonica Woodward: Part I. Cranial Morphology RICHARD ESTES, T. H. FRAZZETTA, and ERNEST E. WILLIAMS HARVARD UNIVERSITY VOLUME 140, NUMBER 2 CAMBRIDGE, MASSACHUSETTS, U.S.A. AUGUST 6, 1970 PUBLICATIONS ISSUED OR DISTRIBUTED BY THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Breviora 1952- MEMoms 1864-1938 Johnsonia, Department of Mollusks, 1941- Occasional Papers on Mollusks, 1945- Other Publications. Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine. Reprint, $6.50 cloth. Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of In- sects. $9.00 cloth. Creighton, W. S., 1950. The Ants of North America. Reprint, $10.00 cloth. Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 12, 15. (Price list on request. ) Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Teredinidae (Mollusca: Bivalvia). $8.00 cloth. Whittington, H. B., and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948. ( Complete sets only. ) Publications of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138, U. S. A. © The President and Fellows of Harvard College 1970. STUDIES ON THE FOSSIL SNAKE DINILYSIA PATAGONICA WOODWARD: PART I. CRANIAL MORPHOLOGY RICHARD ESTES,1 T. H. FRAZZETTA,2 AND ERNEST E. WILLIAMS ABSTRACT A snake fossil described in 1901 by Sir Arthur Smith Woodward and named Dini- lysia patagonica is significant not only be- cause of late Cretaceous age but because it is essentially a complete specimen. It was recovered from sandstone deposits near the Rio Neuquen, Chubut Territory, Argentina. Relatively little notice has been accorded this significant fossil, primarily because it was incompletely prepared. Since many significant features of the specimen were thereby concealed, Woodward described Dinilysia as "a typical member of the Order Ophidia" and related it to the modern bur- rowing snake Anilius (=Ilysia). The pres- ent study indicates that he was mistaken in the first of these statements but at least partially correct in the second. The following snakelike characters are present: (1) prokinetic skull, (2) loss of both temporal arches, ( 3 ) ventral enclosure of brain by frontal and parietal bones, ( 4 ) loose articulation of dentary with posterior mandibular bones. Other snake resem- blances also occur, and the skull has a distinctly snakelike appearance. The lizardlike (or primitive) characters are more numerous than those present in modern snakes. The most striking are (1) palatine bones that are deeply channeled for the internal choanae and that lack the an- 1 Department of Biology, Boston University, Boston, Massachusetts 2 Department of Zoology, University of Illinois, Urbana, Illinois terior toothed projections of modern snakes, ( 2 ) presence of a jugal bone, ( 3 ) presence of both postfrontal and postorbital bones, the latter contacting the jugal as shown on the original cast of the specimen, but now broken on the fossil, (4) single trigeminal foramen, (5) pterygoid bones vertical pos- teriorly, ( 6 ) a single opening in the front of the braincase for olfactory tracts, (7) sta- pedial footplate not enclosed by a crista circumfenestralis, (8) large, laterally-pro- jecting basipterygoid processes. Unusual or unique features of Dinilysia are the robust nasal septum, the vomers underlain by palatines, the low position of the fenestra rotunda on the occiput, the peculiar shape of the jugals, the deeply- grooved transverse suture between para- sphenoid and basioccipital, and both the size of the paroccipital process and the relatively large supratemporal bone. The peculiar oval jaw joint is unique and unlike both the strongly saddleshaped articulation of snakes and the similar but less extreme joint of lizards. Characters in Dinilysia that show resem- blance to the modern aniliid snakes are: ( 1 ) the large paroccipital processes, ( 2 ) the dorsal exposure of the prootic as a tiny sliver surrounded by parietal, supratem- poral, supraoccipital, and exoccipital bones, ( 3 ) the shape of the supraoccipital, ( 4 ) the large stapes, (5) the broad attachment of the quadrate bone to the paroccipital proc- ess, (6) the wide, Cy1indrophis-\ike cultri- form process of the parasphenoid. Bull. Mus. Comp. Zool., 140(2) : 25-74, August, 1970 25 26 Bulletin Museum of Comparative Zoology, Vol. 140, Xo. 2 The lizardlike characters do not relate Dinilysia to any particular group of lizards; indeed most of the characters are prob- ably generalized lepidosaurian features. Some aspects of the occiput and the palate ( especially the posterior palatine and ptery- goid region) resemble those of the Recent platynotan Lanthanotus, but there are also non-platynotan resemblances, such as the depth of the choanal impressions on the palatines. Dinilysia is a mosaic, but in spite of its lizardlike features it is clearly a snake and shows detailed resemblances to the modern Aniliidae that cannot be dismissed as con- vergent. These modern forms are probably the closest relatives of Dinilysia among living snakes, and the unique features do not seem fundamental enough to weaken this relationship. Dinilysia should be placed in a family separate from the Aniliidae, and probably belongs at the base of the super- family Booidea. INTRODUCTION Fossil snake remains tend to be both infrequent and fragmentary. The majority of snake fossils consist of vertebrae and ribs, parts that, although furnishing certain characters of interest to systematists, are quite strikingly similar from group to group. Few snake skulls have been preserved as fossils and, thus, this greatest potential source of information pertinent to the phy- logeny of the Serpentes has been all but wanting. It is thus remarkable that one of the very oldest known fossil snakes, Dinilysia pata- gonica Woodward ( 1901 ) from the Upper Cretaceous of South America, is excellently preserved and includes a nearly complete skull, mandibles, and many vertebrae. Such a specimen might be the focal point of all speculations on the origin and early evolu- tion of the snakes, but, despite considerable interest in snake ancestry (see especially Bellairs and Underwood, 1951; McDowell and Bogcrt, 1954; Underwood, 1957), Dini- lysia has received little more than mention. This is not altogether surprising; Wood- ward's original description of Dinilysia is brief and draws attention to few characters that are not present in many living snakes. Woodward underlined this rather unexcit- ing portrayal of almost the oldest fossil snake with the statement that, on the basis of cranial structure, Dinilysia is "a typical member of the order Ophidia" (1901: 178). We have had the opportunity to re- examine the skull of Dinilysia, and further preparation by one of us ( RE ) has exposed the palate for study. We are unanimous in our opinion that Dinilysia is not a "typical" snake, and that it possesses both special- ized and primitive features so far unknown in other snakes. It is our aim, here, to re- deseribe what seems to us the most impor- tant fossil snake yet discovered and to note some of the ways in which Dinilysia re- sembles other members of the Squamata. We hope in this way to place Dinilysia in its proper context and to enable it to be discussed intelligently in relation to the origin and early evolution of snakes. We do not, however, propose such a discussion here. It is evident to us that further ad- vances in the study of snake origins will require the study of primitive living snakes in more meticulous detail than is found in any work now available, and definitive conclusions may require the discovery of critical fossils still unknown. ACKNOWLEDGMENTS We are most grateful to Dr. Rosendo Pascual (Museo de La Plata, La Plata, Argentina) for the loan of the type speci- men of Dinilysia patagonica, and also to Dr. Mario E. Teruggi of the same institu- tion, who brought the specimen to the United States for our use. We also thank Professor Bryan Patterson of the Museum of Comparative Zoology, whose constant harassment contributed materially to the speed of completion of this manuscript. The manuscript has been read critically Dinilysia Cranium • Estes, Frazzetta, and Williams by Dr. Samuel B. McDowell (Rutgers, The State University, Newark, New Jersey), Dr. Garth Underwood (Sir John Cass Col- lege, London) and Dr. Thomas Parsons ( University of Toronto ) . We have found their suggestions very helpful although we alone accept responsibility for the accuracy of the interpretations offered herein. Mr. Anthony Laska made excellent casts of the skull; these are the more remarkable when the soft and delicate condition of the fossil is considered. Figures 5-12 were pre- pared by Mr. Laszlo Meszoly, and the photographs were made by Mr. Fred Maynard of Boston University. We are grateful to a number of sources for financial aid during the course of this project. An American Philosophical Society Grant No. 3665 to Estes made possible preparation and some of the illustrations, and his research time was supported in part by National Science Foundation grants GB- 1683, GB-4303 and GB-7176. Travel and research time for Frazzetta was provided in part by his National Institutes of Health Postdoctoral Fellowship during 1964-65 and his N. S. F. grant GB-5S31. The Milton Fund of Harvard University supported the preparation of some of the figures. THE GEOLOGICAL OCCURRENCE AND PRESERVATION OF THE SPECIMEN The unique specimen of Dinilysia pata- gonica consists of an articulated skull, jaws, and most of a vertebral column. It is pre- served in a reddish brown sandstone that contains a high percentage of clay-sized particles cementing the medium sand grains. It was recovered from the sandstone deposits near the Rio Neuquen, Chubut Territory, Argentina (Feruglio, 1949). These widespread late Cretaceous sedi- ments were laid down following a great marine transgression in southern South America and are bounded above and below by strata bearing plant and animal fossils of characteristic late Cretaceous aspect; the age of the specimen is thus unquestionable. The deposits seem to have been formed on a flood plain; in this and their relationship to marine transgressions they resemble many North American late Cretaceous de- posits formed under a similar regime. The associated fauna includes giant horned turtles and large carnivorous dino- saurs. The preservation of the bone is excellent; it is white when newly prepared and re- sembles Recent bone. Unfortunately, it is also chalky and brittle and does not absorb protective substances such as Alvar or Glyptal readily. The skull and jaws are little eroded ex- cept at the terminal snout region. Some of the materials used in the past to protect the specimen have resulted in damage over the years: shellac has dried, shrunk and ex- foliated some of the surfaces, and mastic, a soft waxy substance (in part beeswax) used to fill some crevices, has permeated the sur- rounding bone, making it cheesy in texture and difficult to preserve. There is little, if any, distortion or crush- ing evident. The chief agent that has dis- turbed the specimen is weathering, most of which seems to have been post-depositional and subaerial. A large pre-depositional crack extends across the right temporal process and is the only instance of crushing significant enough to warrant comment. As a result, the process itself is depressed below the dorsal skull level and separated from the dorsal skull table by a broad, matrix-filled area. A fragment of bone, probably a part of the prootic in the region of the above-mentioned crack, lies on the right pterygoid. There is apparently no distortion of the individual bones; only a simple dislocation seems to have occurred. As originally preserved (Plates 1 and 2) the jaws were agape, and the elements of the right mandible were almost separated and bent at an unnatural angle. This indi- cates that some maceration had probably occurred before burial, but not sufficient to disarticulate the specimen. 28 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 Unfortunately, the most serious damage to the specimen has been suffered since the entrance of Dinilysia into the literature. The most important such changes are the breakage and loss of parts of the maxillae, the postorbital bar, and the loss of the left dentary region. These are discussed in more detail under the sections on the respective bones. We regret to say that there has been additional damage to the specimen during this study as a result of the very brittle nature of the bone. The posterior end of the right dentary with the splenial is now missing. The right postdentary bones have lost some bone dorsally. The right maxilla has been badly cracked posteriorly and some bone lost on the dorsal surface. Prep- aration of the skull necessary to study important features has resulted in many minor breakages. The snout region was broken and repaired with minor loss of bone and slight dislocation of elements after the casting of the specimen. Fortu- nately almost all of the above damage has been suffered subsequent to photographing ( except Plate 5 of the occiput ) , illustration and casting. Thus the figures represent the specimen at its best while available to us. A cast of the specimen as it was during Woodward's description is in the British Museum (Natural History). Dr. Alan Charig has very kindly had this copied for our study and states that the copy faith- fully reproduces the original cast. Two views of the copy are figured here as Plate 2. Unfortunately, the original cast does not give more than the gross shape of the speci- men, but it does allow some questions about the original specimen to be discussed: the number of maxillary and dentary teeth, the original extent of the maxillae, and the ex- tent of the postorbital bar. These will be discussed in the appropriate sections. We repeat below Woodward's original description; his figures ( la, lb, lc of his plate XX) are reprinted here as Plate 1. We do this ( 1 ) because the views expressed by Woodward are of interest; (2) because this is a report on the specimen before it suffered the damage that has resulted in its present diminished condition; and (3) be- cause reproduction of Woodward's descrip- tion in itself provides the reader with all previous substantive literature dealing with this fossil: other than the diagnosis of the family Dinilysiidae given by Romer (1956: 570), there have been no citations that are not mere mentions (or abstracts of this study; Estes, 1966; Estes et al., 1966). "II. An Extinct Ophidian, Dinilysia patagonica, gen. et sp. nov. ( Plate XX. ) "Mr. Roth's discovery of a fossil Ophidian in the red sandstone of Nequen, associated with typical Mesosuchian Crocodiles, has already been recorded,2 but the unique specimen re- ferred to has not hitherto been studied. It comprises the greater part of the skull and mandible, and fragmentary remains of the anterior half of the vertebral column; the cranium being in an especially good state of preservation. "The skull (Plate XX, figs. 1, la) is long, narrow, and depressed, with the cranial region as long as the facial region. It seems to have been widest at the occiput, where the otic region is very massive; and the maximum com- pression is immediately in front of this, where the parietals rise into a prominent sagittal crest. The constitution of the hinder part of the skull is best seen on the left side of the fossil, where there is only one slight antero- posterior crack (x) in the bones. The right postero-lateral angle, on the other hand, is fractured and displaced downwards. The fora- men magnum, which is filled with matrix, is completed above by the exoccipitals (ex.occ), which meet in the middle line. They are directly continuous on each side with a great, expanded piece of bone (op.), which curves backwards as well as outwards and abuts upon the bone at the upper end of the quadrate. This expansion of the exoccipital is probably the opisthotic, which is similarly fused with the exoccipital in Lacertilia and the extinct Mosa- saurs. Above the exoccipitals is the short but laterally-extended supraoccipital (s.occ), which completes the sagittal crest behind. Its lateral extremity on the left is in contact with a small trace of bone (o. ), which is exposed between the opisthotic, parietal, and the bone 2 A. Smith Woodward, Anales Mus. La Plata — Paleont. Argent, no. iv (1896), p. 1. Dinilysia Cranium • Estes, Frazzctta, and Williams 29 at the upper end of the quadrate. The frag- ment is doubtless the highest point of the anterior otic bone, of which the upper part is otherwise completely buried by the surrounding elements. The parietals (pa.) occupy nearly half the entire length of the cranium, curve downwards to form the side walls, are much compressed in their anterior two-thirds, and rise into a conspicuous sagittal crest. They are flattened in the middle line at their anterior end, and are not pierced by a pineal foramen. Each of the frontals ( //•. ) is nearly three times as long as broad, and slightly widest at its truncated anterior end. There is no supra- orbital bone; but posteriorly and anteriorly the outer border of the frontal is slightly notched for the accommodation of the postfrontal and prefrontal respectively. The postfrontal ( pt.f. ) seems to have partly bounded the compara- tively small orbit (orb.) behind; but this bar is broken away on both sides. The prefrontal (pr.f.), best preserved on the left (fig. la), is flattened and triangular in shape, almost equilateral; it is only slightly in contact with the postero-lateral angle of the nasal bone. The nasals ( na. ) are also flattened and tri- angular in shape, but antero-posteriorly elon- gated and with a somewhat concave outer side which bounds the relatively large narial open- ing (nar. ). They are widest at their articula- tion with the frontals. They are incomplete in front, and the premaxillae are unfortunately not shown. The greater part of the palate is obscured by matrix or broken away, but some features at the postero-lateral angles of the cranium and in the facial region are well shown. As observed especially on the left side (fig. 1), a long and narrow plate of bone (s.t.) forms the postero-superior boundary of the parietal and otic region, and seems to constitute the articulation for the quadrate. This is doubtless the element commonly named supra- temporal in Snakes, Lizards, and Mosasaurs. The quadrate (qu.) is evidently short and broad, but is only imperfectly shown in section on the left side. Its remains (PI. XX, fig. lc) are not readily interpreted; but the upper end of the bone seems to be displaced outwards and incomplete in the fossil, while the more expanded lower end shows the large notch which usually forms a loose articulation for the pterygoid in Snakes. At first sight, it might be supposed that the quadrate was of the same form as that of the Mosasaurs, with a deep posterior notch for the auditory meatus; but closer study seems to make this interpreta- tion impossible. At the side of the cranium, below the supratemporal and parietal, the upper border of a large prootic (pr.o.) is exposed; while between this bone and the orbit the downwardly curved portion of the parietal forms a sharp longitudinal lateral ridge (;•.). There are no traces of temporal arcades. The short pterygoids (pt.) are partly exposed, and a portion of the palatine below the orbit on the left side bears traces of two comparatively minute teeth. There are distinct remains of an ectopterygoid or transverse bone (cc.) on each side between the pterygoid and maxilla; and a fragment on the left side seems to show that this element overlapped the maxilla to a con- siderable extent. The maxilla itself ( mx. ) is relatively large, and best preserved on the right side. It is stout and curves inwards in front. It articulates not only with the ptery- goid behind by the intervention of the trans- verse bone, but also with the palatine by a broad articular palatal process which extends inwards from its middle. It likewise articu- lates directly with the prefrontal in an extensive suture. It shows 14 or 15 large shallow sockets for the implantation of teeth (fig. lb); and one dental crown preserved at the hinder end of the left maxilla is very slender and recurved. The fragmentary remains of the mandible show it to have been of the usual slender ophidian type, widi a very loose articulation between the dentary ( d. ) and articulo-angular region ( ag. ) ; and the dentary exhibits a series of large shallow tooth-sockets like those of the maxilla. "Behind the skull there are remains of a long series of typical ophidian vertebrae, which do not present any features worthy of special note. The neural arches are shown to have borne delicate low spines, though nearly all of these have been broken away and are only represented by their bases in the fossil (Plate XX, fig. 2, u. ) . The ribs ( r. ) are very stout. "From this description it is evident that the Patagonian fossil in question represents a typi- cal member of the order Ophidia. As shown, however, by the conformation of die occiput and the relatively small size of the quadrate, it belongs to one of the more generalized types. Its closest allies may therefore be sought among the Boidae and Ilysiidae, which still constitute so large and characteristic a part of the Ophidian fauna of South America. The skull bears much general resemblance to that of a Boa constrictor, but is readily distinguished from the latter by its non-projecting supra- temporal and relatively small quadrate. It is similarly distinguished from the skull of all the other Boidae.1 In precisely this character, 1 G. A. Boulenger, Catalogue of the Snakes in the British Museum (Natural History), vol. i (1893). 30 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 on the other hand, the fossil skull agrees with that of the existing Ilysiidae; and its occipital region is almost identical with that of the South American genus lltjsia." The resemblance to the latter, indeed, is so close that, although the coronoid region of the mandible is not observable in the fossil, there need be little hesitation in referring the extinct type now described to the family Ilysiidae. It differs from the existing genera of the family in its more numerous marginal teeth and relatively smaller palatine teeth; in its elevated sagittal crest; and in the presence of well-developed neural spines on the vertebrae. It also differs from the South American Ilysia, though agree- ing with the Javan Cylindrophis, in the posses- sion of a small postfrontal bone. It may, in fact, be regarded as a comparatively gigantic forerunner of the Ilysiidae, analogous to Glyp- todon among the Armadillos and Phororhachos among the Cariamas. Whereas the modern representatives of the family are small and degenerate burrowing snakes, the largest less than a metre in length, the extinct Patagonian snake, judging by the size of its vertebrae, must have attained a length of at least two metres. It had a relatively large head, and probably resembled the modern Boas in habit. "This fossil evidently represents a hitherto unknown genus, which may be named Dinilysia and defined thus:- Marginal teeth of moderate size, about 14 or 15 in the maxillary series; palatine teeth relatively minute. Head rather large, the occipito-parietal region constituting half of the skull, with elevated sagittal crest; f rentals longer than broad; small postf rentals present; prefrontals triangular, almost equi- lateral, only slightly in contact with nasals, which are long and narrow, tapering forwards. Vertebrae with low, delicate neural spines. "The type species, of which remains are now described, may be named D. patagonica, and defined by the minor characters of the head-bones already noted." Woodward's description of Dinilysia pata- gonica is not only short but unfortunately inadequate. Our major disappointment with Woodward's description is simple, though not trivial — Woodward stopped short of treating those characters that make Dini- lysia both a unique animal and a significant discovery. This complaint applies to those parts that were exposed when Woodward 2G. A. Boulenger, term. cit. (1893, p. 132, fig. 8.) examined the specimen as well as to those portions that were hidden and for which, of course, Woodward could not be held accountable. Woodward's failure to recog- nize the distinctive nature of Dinilysia might well have stemmed from the paucity of comparative material available to him. He mentions specific comparisons of Dini- lysia with Ilysia ( = Anilius ) and also with Boa constrictor, but lists no other boiform genera by name. Perhaps with a larger collection of cranial material to hand Wood- ward might have been more able — or more confident — to identify certain features that set Dinilysia sharply apart from all other snakes. We attempt below to correct the deficiencies of Woodward's description. A REDESCRIPTION OF THE CRANIAL REGION OF DINILYSIA PATAGONICA General Features of the Skull In general aspect, the skull is flat and broad anteriorly with the orbits opening dorsally as well as laterally (Figs, la-b and 3). Behind the level of the frontals the braincase becomes narrower and grace- fully lengthened. Caudally the posterior braincase swells smoothly transversely to produce a pair of strong, posterolateral projections. A more detailed account now follows in which each bone is treated indi- vidually. SNOUT COMPLEX Premaxilla Examination of the jagged cross section through the anterior ends of the nasals in Dinilysia reveals no indication of premaxil- lary structures. In addition, the copy of the original British Museum (NH) cast indicates that the maxillae extended close to the mid- line (see Plate 2). These facts might indi- cate that the premaxilla was of limited lat- eral and vertical extent were it not that, in many modern boids, the maxillae possess considerable freedom for transverse dis- placements. In some of these, the anterior Dinilysia Cranium • Estes, Frazzetta, and Williams i-H* Figure 1. Dinilysia patagonica; dorsal view of skull. Abbreviations on p. 62. X 1-5. maxillary tips can be brought fairly close to the cranial midline following removal of the premaxilla. Moreover, in certain boids with wide premaxillae (e.g. Boa, Eunectes), the premaxillary transverse process lies just in front of the tips of the maxillae. Nasal Figs. 1, 2a-b, 3, 7; na. The paired nasals together form a nearly equilateral triangle in dorsal view whose anterior apex is displaced slightly to the right. The anteriormost portions of the nasals are badly eroded and incomplete. Originally, the posterolateral nasal corners extended to the lateralmost limits of the anterodorsal edges of the frontals to pro- duce a wide zone of contact between these elements at the nasofrontal articulation ( Fig. 1 ) . The nasofrontal articulation is described below in the section on the frontal. In lateral view, the nasals are relatively thick. Anteriorly, their inferior surfaces slope gently ventromediad and form a thickened internasal partition. Septomaxilla Figs. 2a-b, 3, 7; sm. Although the snout region is broken, there is no real difficulty in interpretation 32 Bulletin Museum of Comparative Zoologij, Vol. 140, No. 2 Figure 2. Dinilysia pafogonico; ventral and occipital views of skull. Abbreviations on p. 62. X 1.5. Dinilysia Cranium • Estes, Frazzetta, and Williams of the bone remnants. On the right side, a thin bony shelf extends along the antero- ventral border of the nasal, beginning about 4 mm anterior to the nasofrontal contact. This shelf makes a gentle curve ventrally as it extends anteriorly, and its dorsal sur- face is inclined lateroventrally throughout its length. Anteriorly, the cross section through the snout (Fig. 7a) shows this plate of bone to be applied closely to bony fragments that appear to be continuous with the vomers (see below). On the left side, only a tiny bar of bone in the same position represents the septomaxilla. Pos- teriorly this fragment slopes dorsally away from the vomer, leaving space between them that must have housed the vomero- nasal organ. Vomer Figs. 2a-b, 3, 7; vo. Ventrally and posteriorly, the vomer meets the palatine at a somewhat indistinct junction. The junction is more easily dis- cerned on the right where the palatine appears to extend forward as a flat, pointed process ventral to the vomer. On the left, however, a line of demarcation between the corresponding palatine and vomer is not clear. Just anterior to the vomer-palatine junc- tion on the right, the vomer is slightly widened laterally; medially and more pos- teriorly, the vomers are visible between the palatines as a pair of slender, rodlike proc- esses (Figs. 2a-b), each of which is the ventral edge of a vertically oriented plate. Anteriorly the vomers are separated from their posterior portions by a matrix-filled crack. Their anteriormost portions directly underlie the septomaxillae (Fig. 7b). BRAINCASE Frontal Figs. 1, 2a-b, 3; fr. In dorsal view both frontals together are longer than wide and are separated by a longitudinal, very slightly irregular median suture. The upper surface of each is very gently convex anteromedially and bears a shallow concavity that is bounded medially by the convexity just noted, and laterally by a slightly raised frontal rim above the orbit. About six or seven tiny foramina extend in a crooked line along the medial slope of this rim. Posteriorly the frontals meet the parietal in a jagged transverse suture, while posterolaterally the frontals are indented by portions of the postfrontals. Anterolaterally the upper surfaces of the frontals are joined by dorsal tonguelike processes of the prefrontal bones. Dorsally, these elements each appear to extend into a frontal notch that accommodates them so that the prefrontal surfaces are at the same level as those of the frontals. The adjacent edges of the dorsal prefrontal and frontal surfaces lie against one another and appear to be tightly connected. Each prefrontal is very slightly notched to receive the laterally projecting frontal edge. The notch horizon- tally divides the prefrontal into the dorsal tonguelike process just described, and a smaller ventral tonguelike process that is pressed against the lower surface of the frontal and is visible ventrally and postero- laterally (Figs. 2a-b). Anterior to the prefrontal the dorsal frontal surface widens suddenly, trans- versely increasing the anterior frontal bor- der as it meets the nasals. A comparison between dorsal and ventral lines of demarcation separating the nasals and frontals reveals that, although the up- per line is relatively straight and transverse, the lower line is shallowly V-shaped and lies from one (medially, at the apex of the V) to nearly five millimeters (about 30% of the total nasal length ) in advance of the dorsal juncture (compare Figs, la with 2a, lb with 2b). Examination of the lateral edge of the right nasal shows that posteri- orly the nasal suddenly becomes vertically thinner where it lies upon an anteriorly projecting frontal shelf that underlaps it; the rostral borders of the paired frontal 34 Bulletin Museum of Comparative Zoology, Vol. 140, Xo. 2 left eo left qu Pa/ bS pt ju P$ pf po pa ps / pt ju rnx tooth Figure 3. Dinilysia patagonica; lateral views of skull. Abbreviations on p. 62. X 1-5. Dinilysia Cranium • Estes, Frazzetta, and Williams 35 shelves produce the V-shaped demarcation from the nasals that is seen in ventral view. The area of contact between nasals and frontals is thus extensive, both longitudinally and transversely. The frontals possess lateral descending walls that meet anteriorly to enclose a single, median opening for the olfactory tracts. This opening is triangular in form with the base formed by the flattened in- ferior surfaces of the dorsal frontal table, and with the apex lying in the midline at a point that is posterior and ventral to the base. The base itself lies about five milli- meters posterior to the level of the dorsal nasofrontal line of juncture. A suture be- tween the two descending walls of the frontal continues caudad from the ventro- median apex of the opening and is shortly hidden from view as it passes dorsal to the vomers. In the orbital region the descending walls of the frontals slope ventromedially at roughly 45 degrees, and are gently concave. The degree of concavity slightly increases posteriorly in the region of the optic fora- men. The frontal forms the anterior border of the optic foramen. The anteroventral border is provided by a small, caudally- pointed frontal process that projects be- neath the foramen and does not meet the parietal; instead, the posteroventral rim of the optic foramen is formed by the basi- parasphenoid, which is closely united to the frontals along the ventral midline. The dor- sal and posterior walls of the foramen are formed by the parietal. The frontal meets the parietal above the foramen and bulges laterally just beneath the postfrontal. Postfrontal Figs. 1, 3; pf. There are, on each side, two distinct elements in the postorbital region. The anteriormost pair of elements contact both frontal and parietal and seem to be homolo- gous to the postfrontals of lizards. On the right side, the postfrontal bears a lateral fingerlike portion (lacking on the left) and appears to be complete. Each postfrontal has a horizontally flat- tened and tapered anterior process that lies snugly in a dorsal frontal recess along the posterolateral edge of the frontal. The lateral border of the anterior process blends smoothly with that of the frontal, but caudally it expands laterally and ventrally to lie upon the corresponding postorbital bone and adjacent parietal as the fingerlike process noted above. Neither the anterior nor the lateral process extends onto the orbital surface of die frontal. Dorsomedially, and just posterior to the frontals, each postfrontal terminates in a short, longitudinal parietal-postfrontal su- ture. Postorbital Figs. 1, 3; po. The second pair of elements in the post- orbital region lies just beneath and slightly behind the postfrontals, and appears to represent the postorbital bones of lizards. No portion of the postorbital extends along the inferior surface of the frontal or parie- tal. Instead, each bone lies on the post- orbital process of the parietal so that the bone is wedged between the postfrontal above and die parietal below. Posteriorly the bone is exposed beneath and behind the postfrontal, the transverse caudal border of its lateral extension paralleling diat of the postfrontal. At the medial end of the short postfrontal-parietal suture, the parietal sends a short bony spur laterally that over- laps the postorbital medially (see Fig. 3). There is a small but distinct fossa ventral to each postorbital, floored and terminated medially by the parietal. On the left side the lateral projection of the postorbital is missing, while on the right the postorbital curves ventrally as it extends laterally beyond the tip of the postfrontal. The lateral postorbital process is wider and less horizontally flattened than is the post- frontal. 36 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 The free lateral end of the right postor- bital has a somewhat rough and uneven texture indicating that a still more latero- ventral extension of the bone has been broken away. The situation is confused by an inconsistency in Woodward's paper. His figure of the dorsal view of the skull ( Fig. 1 in Plate 1) shows the left postorbital region much as it still appears in the speci- men today. On the right, Woodward figures a postorbital bar that continues ventro- laterally from the fingerlike process of the postorbital bone as figured here to connect with a peculiar element (interpreted in the present paper as a jugal) lying on the sur- face of the maxilla. It is unfortunate that Woodward did not figure the right lateral view of the skull, for his written description seems to contradict his pictorial presenta- tion. He surmised that the postorbitals ( — postfrontals of Woodward ) "partly bounded" the orbit posteriorly, but goes on to say that the postorbital "bar is broken away on both sides." Woodward's choice of the word "broken" to describe the lack of a postorbital connection to the jugal makes it unclear whether he intended to convey that the specimen was damaged after it was illustrated, or that such a con- nection was never present in his specimen (in which case his figure is in error). Al- though his figures are somewhat unclear in several respects, they appear to be quite accurately and carefully rendered, and, if the postorbital bar is shown incorrectly, it is the only major mistake we can find in Woodward's illustrations. Furthermore, it is significant that Woodward raised the issue at all, since he regarded Dinilysia as a relatively unmodified snake, readily com- parable to Boa constrictor. It is easily possible that the postorbital bar could have been damaged during Woodward's examination of the specimen. The fossil is very brittle and, even with careful handling, several areas have crum- bled away and have required restoration so that the specimen as it appears at this writing does not match our own figures as well as it first did. Unfortunately the pres- ence or absence of a postorbital connection with the presumed jugal element influences our functional interpretation (to be pre- sented in a future paper). Evidence provided by the cast of the specimen in the British Museum (Natural History), and mentioned above in the sec- tion on preservation, indicates that the figure is correct, and that the postorbital bar was originally complete on the right side. The cast shows bone continuous from the skull roof to the maxilla. Wood- ward's statement that the postorbital bar was broken thus may have referred to breakage after the cast was made. We thus conclude, on the basis of available evidence, that when the specimen first came into Woodward's hands the postor- bital met the jugal. Parietal Figs. 1, 3; pa. Dorsally and anteriorly, the parietal be- gins between the postfrontals and postorbit- als at the jagged frontoparietal suture. In this region the parietal is flattened with a slight medial concavity; this essentially flat portion quickly narrows posteriorly and continues caudally as the dorsal edge of the sagittal crest. On either side of the anterior extremity of the crest, the parietal is hol- lowed to form a concavity that becomes shallower posteriorly and extends antero- laterally to the level of the postfrontal- parietal suture. More laterally, the parietal contour drops sharply downward onto a prominent shelflike modification of the bone. This shelf juts laterally and some- what ventrally so that its upper surface slopes at least 30 degrees below the hori- zontal; the inferior surface of the shelf is smoothly concave. Anteriorly the shelf curls upward to meet the postorbital process of the parietal. Posteriorly it extends ventrally to the anterior border of the trigeminal foramen. Dinilysia Cranium • Estes, Frazzctta, and Williams 37 &£ tf* -pq — o o _Q c o E 5 c c D V. a| to O - sO vT~ -* &"!< 38 Bulletin Museum of Comparative Zoology, Vol 140, No. 2 left pt Figure 5. Dinilysia pofogon/co; A, oblique dorsolateral view of left basiparasphenoid and surrounding elements; note anterior separation of basisphenoid from parasphenoid. B, oblique ventrolateral view of area between basipterygoid process and fe- nestra rotunda, showing partial excavation of vidian canal and canal dorsal to it opening at f on Fig. 5a, and area between sphenopalatine tuber and stapes. Foramina in dark stipple, matrix in coarse stipple, excavated or broken areas in hatched line; abbreviations on p. 62; both X 3. Diagrammatic. The elongated major portion of the parie- tal shows a faint cerebellar swelling that widens gradually until, posteriorly, it ex- pands smoothly but markedly laterad. The parietal contacts the prootic both laterally and dorsoposteriorly, extending narrowly between the supratemporal and a dorsally exposed sliver of the prootic (see below). The parietal joins the supraoccipital pos- teriorly in a long, transverse suture. Anteroventrally the parietal forms the posterior and dorsal rim of the optic fora- men. In this region the ventral parietal surface joins the parasphenoid with which it is firmly sutured as far posteriorly as the anterior border of the trigeminal foramen. At the level of the basipterygoid processes of the basiparasphenoid, the parietal is de- veloped ventrally and laterally to contrib- ute a small, lateral portion to the base of each basipterygoid process. Immediately posterior to the level of the basipterygoid Dinilysia Cranium • Estes, Frazzetta, and Williams 39 process, the parietal is narrowly included in the anteroventral border of the single trigeminal foramen. Prootic Figs. 1,2, 3,5; pr. The prootics are relatively large. On the left, the bone is complete, while on the right it is divided by a wide vertical frac- ture, which has dislocated the entire supra- temporal process ventrally. Much of the ventrolateral part of the right prootic is covered by matrix and by a small piece of bone that may be a dislocated piece of the right prootic. The anterior border of the prootic begins dorsal to the basipterygoid processes. The large, single trigeminal foramen lies in a notch on the anterior border and is almost entirely enclosed by the prootic except for a small contribution from the parietal an- teroventrally. The tiny facial foramen lies on the anterior margin of the fenestra ovalis; the latter is partly excavated into the posterior border of the prootic. Ventrally the prootic comes into broad contact anteriorly with the basiparasphe- noid and extends posteroventrally to the basioccipital, contributing to the basioccipi- tal tubera as noted below. At the level of the basiparasphenoid-basioccipital suture, the prootic is shallowly though sharply ex- cavated to form a small, bony "step" (see Fig. 2a). This step anteriorly bounds a depression that is medially and posteriorly delimited by tuberous processes of the basisphenoid and basioccipital. The prootic bulges laterally as it con- tinues posteriorly onto the supratemporal process. This portion of the bone, just before reaching its caudal sutural limit with the supratemporal, is abruptly flattened in a parasagittal plane as is the adjacent supra- temporal. At its posterior end, the prootic interdigitates strongly with the supratem- poral and the prootic-supratemporal suture courses medially at this point on the ventral surface of the supratemporal process. Dorsal to the supratemporal, and bounded by it, the parietal, and supra- and exoccipi- tal elements, there is a separate, sliverlike bone, which represents an internal portion of the prootic that "breaks through" the braincase to be exposed dorsally, in isola- tion from the rest of the prootic. Examina- tion of the large posterior fracture of the prootic on the right reveals that the break has cut this bony sliver, and that it is con- tinuous with the prootic but separated externally by the long diagonal suture of supratemporal and parietal. Supratemporal Figs. 1-3; st. The supratemporals begin on the trans- versely-expanded posterior portion of the braincase as rather narrow wedges that each lie between the parietal above and the prootic below. At the level of the dorsal prootic sliver each strip becomes vertically deepened and continues caudal to and beyond the posterior tip of the lateral part of the prootic. Here the supratemporal expands ventrally to form a parasagittally- flattened plane that is continuous witii the flattened posterior surface of the prootic. Caudally, the supratemporals extend well posterior to the level of the occipital condyle and cover the lateral, and part of the ventral, portions of the enlarged paroccipital processes of the exoccipital bones. Supraoccipital Figs. 1, 2c-d, 3; so. This element is a very wide, short, median bone sutured anteriorly to the pari- etal, laterally to the sliver-like process of the prootic, and caudally to the exoccipitals. The bone bears a median crest that is con- tinuous with the sagittal crest of the pari- etal. On either side of the crest there is a ridge that begins at the parieto-supra- occipital border and extends posterolaterally to the posterior supraoccipital border, thickening progressively toward its postero- lateral end. At the supraoccipital-exoccipi- 40 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 tal border, posterior and medial to the ridge, a prominent depression is present on each side. Exoccipital and Opisthotic Figs. 1-3; eo. Each exoccipital ( including a fused opis- thotic) dorsally makes contact with the supraoccipital, prootic, and supratemporal. Posterolaterally each is expanded laterally to form a large paroccipital process that continues caudally posterior to the occipital condyle and that possesses a wide, flat superior surface facing dorsomedially. Pos- teromedially, the exoccipitals overhang the foramen magnum to form an awninglike rim. Much of the occiput is formed by the exoccipitals. Here they descend to form the lateral borders of the foramen magnum and, since they join and provide the dorsal and lateral portions of the condyle, they also form most of the ventral border. Lateral to the condyle, the exoccipitals ex- tend transversely as a pair of wide, tongue- like processes (broken off on the left) that lie beneath the paroccipital processes and floor the posterior lacerate foramina ven- trally, concealing them from view. On the right, the bony tongue is well preserved, and is sutured to the basioccipital ventrally. The fenestra ovalis and the fenestra rotunda are separated by another, more ventral proc- ess of the exoccipital; this process continues ventrally and projects to meet the prootic and the basioccipital below the fenestra ovalis, where it contributes to the posterior portion of the rather complex spheno- occipital tubercle (Fig. 5b). Basioccipital Figs. 1-3; bo. This unpaired ventral bone begins an- teriorly at a broad, transverse suture with the basiparasphenoid. At the ends of this suture the basioccipital develops ventrally a pair of tuberous processes that lie adja- cent to similar developments of the basi- parasphenoid. Caudad to the level of these processes, the medial surface of the basi- occipital bends dorsally, while the lateral surfaces continue posterolaterally from the tuberous processes as ventrally-developed walls enclosing a wide, short concavity. The left lateral surface is partially broken away to reveal the opening of the fenestra ro- tunda; on the right, the lateral wall ex- tends toward the pterygoid. Anterolaterally the basioccipital joins the prootics, while posterolaterally it meets the exoccipitals. Posteromedially the bone narrows and produces a trans verse ridge at the base of the occipital condyle. The basioccipital forms the ventromedian part of the condyle. The basioccipital can be seen dorsally as a thin sliver separating the exoccipitals, and it contributes about one- half of the body of the condyle itself. Basiparasphenoid Figs. 2, 3, 5; bps. This ventral element begins posteriorly at its transverse suture with the basioccipital, where it is bounded laterally by the pro- otics. The posterolateral corners are strongly developed ventrally to form a pair of tuber- ous processes that lie in contact with similar developments of the basioccipital. Between these processes the ventral surface of the basiparasphenoid forms a wide trough that narrows anteriorly, where it is bounded by a pair of low but sharp ventral ridges that tend to converge toward one another anteriorly. These ridges originally bore laterally-developed crests, setting off groovelike pockets dorsally. Unfortunately these delicate crests have not survived this present study, but a small portion of the right one may be seen in Figure 2a and b. Lateral to these ridges and directly posterior to the basipterygoid processes are the paired posterior openings of the vidian canals. The laterally-placed basipterygoid proc- esses have convex articulating surfaces that face more laterally than ventrally; their axes are anteroposteriorly elongated, and Dinilysia Cranium • Estes, Frazzetta, and Williams 41 their ventromedial surfaces are gently con- cave. The ventral margin of each process is thickened and is separated slightly from the pterygoids. The joint between the processes and the pterygoids is straight and has a slight anteromedial-posterolateral ori- entation. The dorsal sides of the processes are partially covered by moderate ventral downgrowths of the parietals. The anterior vidian canals occur far laterally on the basi- ptt rygoid processes. The parasphenoid and basisphenoid are separate anterior to a foramen that opens dorsomedial to the anterior opening of the vidian canal (Fig. 5). The lateral and medial borders of this foramen are formed by the basisphenoid, its visible dorsal bor- der formed by the parietal, and its ventral border by the parasphenoid. The basi- sphenoid extends anteriorly from this fora- men about one-third of the distance to the optic foramen. At this point on both sides it stops abruptly at prominent, matrix-filled trabecular grooves ( Fig. 5a ) , and the para- sphenoid expands dorsally to meet the parietal, filling the gap in front of the basisphenoid. The trabecular grooves lie wholly in the cultriform process of the parasphenoid, although they approach the parietals closely. The cultriform process of the para- sphenoid is wide posteriorly near the basi- pterygoid processess, but tapers anteriorly and is closely pressed to parietal, basisphe- noid, and, presumably, frontals. The inferior surface of the bone, between the basi- pterygoid processes and its meeting with the palatines, bears a distinct, elongated concavity. The tapered anterior end of the para- sphenoid extends between and slightly dorsal to the palatines; the tip stops short of the posterior extremities of the vomers. STAPES Figs. 2, 3; s. On the left side, posteroventral to the prootic, there is a stapes consisting of a relatively large footplate and, as preserved, a delicate columellar process. Medially the bone is inserted into a fenestra ovalis, which is in turn set into a deep cleft in the pos- terior surface of the deeply emarginated prootie and the exoccipital. The fenestra ovalis is separated from the fenestra rotunda by an anteroventrally-projecting process of the exoccipital. The columellar process is broken, but extends toward the quadrate and approaches the posterior portion of the pterygoid. Whether or not it abutted against the quadrate is conjectural because of columellar breakage and quadrate dis- location. SUSPENSORIA OF THE PALATO-MAXILLARY ARCH; THE ARCH ITSELF Prefrontal Figs. 1-3; prf. The prefrontals are horizontally flattened so that their entire lateral surfaces could be equally well designated as dorsal. Since the prefrontal forms the anterior orbital wall, this flattening extends the orbit later- ally so that it opens as much superiorly as laterally. The external surface of each prefrontal (better preserved on the left than on the right) is roughly an equilateral triangle, with the lateral side lying upon a dorsal expansion of the maxilla, and the medial corner forming a dorsal tonguelike process that is inserted into the upper frontal sur- face. The anterior prefrontal margin ex- tends from this dorsal corner and crosses the lateral corner of the nasal to terminate against the maxilla. The ventral tonguelike process that lies against the lower frontal surface bears a pronounced transverse elevation that cau- dally delimits a deep prefrontal concavity that opens medioanteriorly. The ventral edge of this elevation is expanded anteriorly as a flattened, horizontal surface that comes into contact with the palatine and maxilla. Posteriorly, in the anterior orbital region, 42 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 Figure 6. Dinilysia pafagomca; reconsfruction of occiput. Abbreviations on p. 62. X 3.5. Dotted line = conjectural; struc- tures missing on one sick restored from the other; broken exoccipital-basioccipital hatched and not restored, in order to show fenestra rotunda. Diagrammatic. each prefrontal descends and then flattens horizontally where it rests upon the upper surface of the lateral palatine process (see below). Here the prefrontal is perforated by the combined orbitonasal canal and lac- rimal foramen, which on both sides is nearly closed ventrally by the prefrontal. Palatine Figs. 2a-b, 6; pa!. The palatines lack toothed anterior pro- jections. Instead, the dentigerous portion does not extend anteriorly beyond the level of the lateral process of each palatine. A distinct, deeply impressed choanal channel is present, which is partially closed ven- trally by sheetlike expansions from both lateral and medial edges of the palatine. Medial to the choanal channel, the sheetlike expansion is pressed against the lateral and inferior surfaces of the vomer as described above. The posterolateral dentigerous por- tion of each palatine is a thick, ribbonlike process facing ventromedially, and is rather smoothly continuous with the medial proc- ess. No teeth are preserved but small sockets for the subpleurodont teeth are per- ceptible. If we may judge by the diameters of their sockets, the palatine teeth were rela- tively small. They extended forward in a single row from the palatopterygoid joint to the posterior level of the lateral palatine process. It is impossible to determine pre- cisely their numoer; we estimate that each palatine bore about five or six teeth. The lateral process of the palatine is formed by a sudden horizontal expansion and flattening of the lateral palatine edge, just anterior to the level of the first palatine tooth. This expansion provides a partial floor beneath the choanal passage in that region, but its major portion stretches later- ally over the superior surface of the medial maxillary process and beneath the prefron- tal. The lateral palatine process is very extensive in that it intervenes between pre- frontal and maxilla in such a way as to all but prevent those elements from coming into contact, and it reaches transversely nearly to the ventrolateral edge of the pre- frontal. Below the lacrimal foramen, the lateral palatine process is perforated by a transversely-widened maxillo-palatine fora- men. Dinilysia Cranium • Estes, Frazzetta, and Williams 43 In dorsolateral view, just in advance of the dorsal lappet of the pterygoid, a tiny (venous?) foramen pierces the dorsal sur- face of the longitudinal portion of the palatine (Figs. 3, 8c). Posteriorly the palatines join the ptery- goids in a rather complex manner. Dorso- laterally the pterygoid sends forward a pointed projection that seems to lie in a correspondingly shaped shallow depression in the palatine surface. Ventrally (partially preserved on the left side) similar pointed processes are mostly broken away but presumably extended onto the undersurface of the palatine, as indicated by depressions in the bones. In effect, the palatine was clasped by these two projections (cf. Figs. 5a, 8c, 10a). Pterygoid Figs. 1-3, 5-6; pt. The anterior attachments of the ptery- goids with the palatines have just been described. The pterygoid tooth row is con- tinuous with that of the palatine. It ap- pears, upon examination of the tooth sock- ets, that the approximately five pterygoid teeth were subequal to those on the pala- tine and that they diminished in size posteri- orly. The last tooth socket lies just ahead of the level of the basipterygoid joint. The internal process of the pterygoid is moderately developed, arises at the level of the basipterygoid, and meets the ventro- laterally facing basipterygoid surface. At the level of the internal process, the lateral side of each pterygoid joins with an ectopterygoid. Again at the same trans- verse level, but dorsally, each pterygoid bears a sharp longitudinal ridge. Slightly posterior to this point the pterygoid be- comes sharply compressed, the upper edge of the ridge continuing caudally as the dorsal margin of the compressed pterygoid portion. The posterior, compressed portion of the pterygoid is convex laterally and concave medially. Posteriorly the vertical dimension increases to the level of the basi- cranial tubera and then decreases posteri- orly; the dorsal border remains longitudi- nally straight. In ventral view, the posterior halves of the pterygoids are essentially straight. The posterior tips of the ptery- goids (as seen on the left) are applied to the medial surface of the quadrate. A small, pebblelike element (labelled "?" in the figures) lies upon the dorsal surface of the right pterygoid. This is probably a dislodged fragment of the nearby broken area of the prootic as noted above. Maxilla Figs. 1, 2a-b, 3, 7; mx. Woodward's figure shows the right max- illa as complete (Fig. 1 in Plate 1). It is sad to note that today neither maxilla is complete anteriorly. It is also regrettable that Woodward figured only the dorsal view of the right maxilla. On the original cast (Plate 2) the maxilla curves conspicu- ously downward anteriorly, is strongly ex- cavated for the naris, and reaches almost to the midline. At the present time, the maxilla on the left is somewhat better preserved than that on the right. In the region of the prefrontal, the lateral maxillary surface is flattened and turns upward to meet the lower margin of that element. The lateral maxillary sur- face is smooth and convex anteriorly; pos- teriorly it is somewhat dorsoventrally ex- panded and concave. Dorsally, the superior maxillary surface receives the lateral palatine process that intervenes between prefrontal and maxilla. This process is very loosely articulated with the maxilla and is separated from it by a deep, matrix-filled groove. In this region the maxilla sends forth a flattened medial proc- ess that is seen in ventral view to extend mediad along the ventral surface of the palatine. More posteriorly the maxilla bears a marked ridge running lengthwise on the dorsal aspect of the bone. This ridge sharply 44 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 Figure 7. Dinilysia patagonica; A, anterior, and B, right lateral view of snout region; hatched line = broken surfaces. Ab- breviations on p. 62. X 3. Diagrammatic. separates the lateral maxillary surface just described from the medial side of the maxilla. The surface of the medial side faces dorsomedially; caudally it becomes increasingly concave to form a rather deep, longitudinal channel that extends to the posterior limit of the bone. Lying in this channel, on each side, is a small fragmentary element interpreted here as the jugal (see below ) . Ventrally each maxilla has a row of tooth sockets that follows the ventrolateral edge of the bone. Much of the ventral area is obscured on the right; on the left, there is evidence of seven maxillary teeth on the specimen today, but the cast indicates that at least thirteen maxillary teeth were origi- nally present. There is no clear indication that the teeth decrease markedly in size posteriorly. The posterior two sockets on the left contain remains of teeth. The anterior tooth is very fragmentary. The caudal surface of the more posterior tooth is hidden by ma- trix, and its tip is broken away. However, its anterior and ventral surfaces are intact and visible; it arises from the maxilla as a thick, basal portion from which it curves very markedly caudad to give the appear- ance of a sharply bent cone. The backward sweep of this tooth does not quite lie in a parasagittal plane, but slants slightly in- ward at an angle of approximately 20 de- grees to the longitudinal axis. Posteriorly the maxillae are joined by the ectopterygoids (mostly absent on the right). Here the left maxilla bears, along its medial margin, a longitudinal groove ( for the ecto- pterygoid ) that extends forward to the level of the middle of the jugal. Ectopterygoid Figs. 1, 2a-b, 3c-d; ec. Unfortunately neither ectopterygoid is well preserved, and the right element is al- Dinilysia Cranium • Estcs, Frazzctta, and Williams 45 most entirely lacking ( though it was nearly complete in 1901). On the left, the ecto- pterygoid arises from the lateral edge of the pterygoid and, in the form of a small cylinder, passes forward and outward to the maxilla. It appears to abut against the caudal maxillary tip and to cover a very small portion of the upper maxillary sur- face, and a groove on the medial surface of the maxilla indicates that it once extended anteriorly to the level of the middle of the jugal. Jugal Figs. 1, 2a-b, 3; ju. A small bone lies in the longitudinal channel formed on each side of the dorso- medial maxillary surface. These elements are essentially symmetrical in position and in shape and are unlikely to be adhering fragments broken from some other portion of the skull. Each of these peculiar bones has the form of a flattened ellipsoid whose anterior end is turned upward toward the tip of the postorbital. They appear not to be portions of the ectopterygoids that have been accidentally isolated, since grooves for the anterior extensions of the latter occur on the medial sides of the maxillae. These bones are best interpreted as jugals. In the natural condition, the dorsally turned anterior portion of the jugals must have made contact with the postorbitals as indicated on the original cast ( see above, Postorbital ) . Quadrate Figs. 1-3; qu. Much of the left quadrate remains and is still attached to the skuii. Its lateral surface is peculiarly flat, suggesting that a portion of this surface has been sheared off. The lower portion of the right quadrate is articu- lated with the right mandible; there is no remaining trace of the upper portion. The quadrate was apparently a some- what triradiate structure. The posterior arm of the left element extends caudally behind the enlarged paroccipital process and the anterior portion is deflected laterally, indi- cating that the bone has slipped posteriorly. The normal position of the quadrate must have been more anterior, and the anterior quadrate arm must have come close to the caudally flattened part of the prootic. The ventral arm of the essentially tri- radiate quadrate is somewhat flattened anteroposteriorly and expanded trans- versely. Its lower end curls posteriorly, presenting a smoothly convex hemispherical articulation surface ventrally. The lateral side of the posterior end of the pterygoid articulates with the medial quadrate surface fust dorsal to the level of the hemispherical expansion. The articula- tion surface was originally further anterior on the pterygoid before distortion of the quadrate. MANDIBLES Fig. 4 Three mandibular fragments have been preserved. These consist of the posterior half of the left mandible, whose broken an- terior end is neatly severed, and a nearly complete right dentary and a right "com- pound bone." The right elements can be fitted together, as shown in the figures, although they were originally distorted as shown in the original cast (PI. 2) and in Woodward's figure (PI. 1). The left den- tary, present at the time of Woodward's study, has been lost. Compound Bone Fig. 4; cp. Different portions of the two compound bones possess badly eroded surfaces and parts. Examined together, however, the two bones provide a rather complete inter- pretation of the natural condition. The articular area forms a shallow cup that opens dorsomedially to receive the hemispherical lower end of the quadrate 46 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 in a ball-and-socket joint. No retroarticular process of any sort is present, although the ventral surface of the mandible is broken posteriorly, and this region cannot be recon- structed with certainty. The surangular process is represented only by a low, smooth, longitudinal ridge on the lateral surface of the compound bone. The prearticular ridge is well devel- oped with a bluntly pointed dorsal contour. Between the surangular and prearticular processes the small mandibular fossa opens. Immediately anterior to this, the surangular bears a shallow longitudinal channel in its dorsal surface that extends alongside a nar- rower, rather inconspicuous channel on the anterolateral border of the prearticular la- mella. The two channels together form an elongated concave area for coronoid artic- ulation. The right compound bone has lost the elements (coronoid, angular, splenial) that naturally adhere to it, and its anterior structure can be studied. The anterior end bears a short medial, and a larger lateral, projection. At the point where the two projections begin to diverge, a shallow groove runs posteriorly in which lay the coronoid. The ventral surface of the larger projection itself bears a groove (for the upper posterior end of the dentary) that ends abruptly near the base of the projec- tion. When articulated with the dentary, the larger lateral surangular projection fits in a notch between the dorsal and ventral caudal ends of the dentary; the shorter, medial projection makes contact with the splenial bar that connects the medial edges of the dorsal and ventral dentary ends. Dentary Fig. 4; de. Only the right dentary is preserved, although fragments of the posterior end of the left still remain attached to the pos- terior bones. The dentary curves mediae! anteriorly, is slightly pointed and seems not to have been firmly united with its fellow on the left. Posteriorly the dentary forks into dorsal and ventral processes that are free laterally but are connected medially by the splen- ial. By comparison with the left mandible, the dorsal, tooth-bearing process on the right is seen to be incomplete at its lateral border, but the ventral process has only minor posterior breakage. The lateral notch between them communicates with Meckel's groove on the medial side. The lower, me- dial surface of the splenial bears a splenio- dentary suture. A large, elongated alveolar foramen is formed between the splenial and the ventral surface of the tooth-bearing bor- der of the dentary. Anteriorly, on the medial side, Meckel's groove becomes narrower and approaches the ventral dentary margin where it termi- nates shortly behind the dentary tip. Ten large tooth sockets are borne by the dentary as preserved, and also on the origi- nal cast of the left dentary. In addition to these, at least one smaller tooth socket may be present close to the anterior tip of the bone, but this is uncertain because of breakage. The size of the teeth appears to have diminished slightly caudad. Splenial Figs. 4c-h; sp. A portion of the splenial is preserved on the left in articulation with the posterior bones, and its articulation with the dentary is shown on the right. On the left, a somewhat rectangular piece of bone is identifiable as the splenial by its position. It is bordered dorsally by the coronoid and upper dentary process, posteriorly by the angular, and ventrally by the lower dentary process. Coronoid Figs. 4e— h; co. The right coronoid is missing. A portion of the left coronoid lies in a shallow groove Dixilysia Cranium • Estes, Frazzetta, and Williams- medial to the dorsal tooth-bearing process of the dentary. Anteriorly it lies between the dentary on the dorsolateral side, and the splenial on the medial side. Posteriorly this fragment stops short of the prearticular lamella. The coronoid must have extended pos- teriorly into the groove on the dorsal surface of the compound bone noted above; the latter bears striated attachment surfaces for its reception. The coronoid must have forked posteriorly, one part ascending the prearticular ridge medially as indicated by a roughened articulation surface, the other forming a coronoid process of unknown ex- tent laterally on the surangular; the two processes thus complete the relatively nar- row mandibular fossa anteriorly. Angular Figs. 4g-h; an. An angular appears only on the left mandible. It is short and thick as preserved, but can be seen to have extended posteriorly to the level of the mandibular fossa. It is traversed by a high ridge that extends the length of the preserved portion of the bone and probably did so throughout its original length, since a ridge is still present at the posterior end of the suture area. The angular is positioned somewhat ven- tromedially and is slightly slanted upward anteriorly. Its anterolateral corner ap- proaches, but no longer reaches, the coro- noid; a groove presently exists between breakage surfaces of the two bones and it can only be surmised that the two bones met. The angular also makes extensive contact with the splenial anteriorly and touches the posterolateral edge of the lower caudal dentary process. Foramina of the Skull Foramina are described briefly under the sections dealing with bones in which they occur; for convenience, a description of skull foramina in the context of their surrounding bones is given here. Termi- nology follows that of Bahl (1937), although we do not necessarily agree to the appro- priateness of the names applied. The tiny lacrimal foramen (If, Figs. 1, 8c) is completed dorsally by prefrontal and ventrally by palatine. Medially it seems to be confluent with the relatively large orbito- nasal canal (oc, Figs. 1, 8c); the latter is floored ventrolaterally by the palatine. The maxillo-palatine foramen (mp, Figs. 1, 8c) lies wholly within the lateral (maxil- lary) process of the palatine bone. Dorsally, the anterior border of the optic fenestra (of, Figs. 3c-d, 5) is formed by the frontal, while posteriorly it is sur- rounded by the parietal. Its ventral border is completed by the parasphenoid. Both anterior and posterior openings of the vidian canal (pvc, Figs. 2a-b, avc, Fig. 5) perforate the basipterygoid process of the basiparasphenoid. Both openings are about the same size. Mechanical excavation of the basipterygoid process shows that the vidian canal has a tiny, short anastomosis dorsally with a canal that appears to come from within the cranial cavity and leads anteriorly to an opening dorsal to the an- terior opening of the vidian canal (Fig. 5). The dorsal canal is formed dorsally by the parietal, laterally by the basisphenoid, and ventrally by the parasphenoid. After leav- ing the anterior opening, the contents of the canal have left a channel on the basi- sphenoid. The trigeminal foramen is single, and relatively large (V, Figs. 3c-d, 11a). It lies almost wholly within the prootic, but an- teroventrally it receives a small contribution from the parietal. The facial foramen (VII, Figs. 2c-d, 3c- d, 11a) opens through the prootic at the anterior margin of the fenestra ovalis. The fenestra ovalis (Figs. 2c-d, 12a) is quite large, but it cannot be seen directly because of the massive, superimposed sta- pedial footplate. The foramen is formed by deep emarginations in the posterior part 48 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 X o o D D a. ft .§ o Q E u § ltT to ■ o T3 Xe 0) oo" °° M U Q. 2 c * o 5 c -2 9 1! ° -5 .. CN o x-g lo X oo % S <*> So o M o u — u a o o X O LO -7^ Q O a> c o -Q ca Dinilysia Cranium • Estes, Frazzetta, and Williar, of the prootic and the anterior border of Without making every comparison or pro- the exoccipital. The latter sends a lappet viding final judgment on its position in the ventrally to the spheno-oeeipital tubercle; phylogenetic sequence, we do wish to call presumably this lappet represents a fused attention to its most conspicuous characters opisthotic. indicate its position relative to other snakes The fenestra rotunda (fro, Figs. 2c-d, and most obvious resemblances, and so 12a) is large. Dorsally it is separated from and to lizards, as we now see it. the fenestra ovalis by the lappet of exoccipi- To do this as nearly as possible without tal described above, and ventrally it is com- bias, we have chosen as the specimens to pleted by the basioccipital. be used for our first comparisons with The foramen lacerum posterius (plf, Figs. Dinilysia a few primitive snakes, Python 2c-d, 12a) opens through the posterior sur- sebae and Epicrates cenchris (described by face of the exoccipital, lateral to the fora- Frazzetta, 1959), and certain lizards, men magnum. In occipital view, it is Lantlianotus borneensis (discussed by Mc- concealed behind a small tongue of the Dowell and Bogert, 1954), Varanus monitor exoccipital. No separate hypoglossal fora- (elaborately described by Bahl, 1937), mina are associated with it on the occipital Tupinambis nigropunctatus (taken as an surface, nor does excavation show that such example by Jollie, 1960), Mabuya carinata foramina appear within the mouth of the (described by Rao and Ramaswami, 1952) foramen. and Ctenosaura pectinata (discussed by The foramen magnum (Figs. 2d, 12a) is Oelrich, 1956). We have added a few ob- surrounded by the exoccipitals, except ven- servations from skulls of Cylindrophis and trally, where the latter are separated on the Anilius at hand. By these admittedly very midline by a tiny wedge of basioccipital. specific, but we hope judiciously distrib- Comparisons dealing with the foramina uted comparisons, we have tried to avoid appear within the context of the discussion the unjustified generality of statement that below, afflicts so many discussions of the charac- ters and relationships of higher groups. We DISCUSSION admit out of hand that comparisons so se_ Comparisons of Dinilysia with modern cured will not necessarily be the most rele- snakes indicate that it shows greatest simi- vant. However, until thoroughgoing investi- larity to the primitive snake families Boidae gation provides greater assurance, we can and Aniliidae (sensu lato), especially to the present only a very preliminary assessment latter. However, the aniliid complex is itself of the position of Dinilysia. We do not in great need of study: no detailed osteo- pretend to attempt more. When, therefore, logical descriptions of Loxocemus, Xenopel- in the comparison below we contrast lizards tis, Cylindrophis, or Anilius exist. In addi- and snakes, ice mean no more than that tion, Dinilysia has some characters that within our sample all the lizards and all the appear unique and others that invite com- snakes were each congruent tvith the other parison with lizards rather than with snakes, examined members of their suborder. A full scale comparison of Dinilysia with all We shall score below (in italicized the forms that should be utilized will be a phrases) the ascertainable characters of lengthy task and one requiring an abun- Dinilysia as "snakelike," "lizardlike," "inter- dance of illustration. It is, in fact, still in mediate" or "unique." Here again the state- the future. ments should be taken only in the context We feel, however, the need to put Dini- of the comparisons expressly made. Certain lysia in its approximate context at this time, comparisons are omitted at this time, e.g. 50 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 comparisons with the as yet poorly under- stood typhlopids and leptotyphlopids.1 The skull of Dinilysia is incomplete, but its preserved parts will be discussed seriatim, following the sequence of the description. No indication of premaxillae exists at present; however, the original cast of the specimen (Plate 2) shows the right maxilla extending close to the midline, appearing to imply that the premaxilla was originally of relatively small lateral extent. The smooth tip of the maxilla on the cast probably indi- cates a loose articulation with the premax- illa, as in extant snakes, instead of the sutural connection characteristic of lizards. Tliis feature therefore would be scored as a strong resemblance to snakes. The descending lamellae of the two nasals in Dinilysia are extremely thick and short and together form an extraordinarily robust septum almost one-third as wide as the widest part of the two horizontal nasal lamellae above ( Fig. 7 ) . In snakes there are thinner, deep, sharply defined descending nasal lamellae ( Frazzetta, 1959 ) . In lizards there is instead a cartilaginous nasal sep- tum. In having a bony rather than a carti- laginous nasal septum Dinilysia is snake- like. In the thickness of its septum it is unique. The vomers of Dinilysia are visible be- tween the palatines ventrally as a pair of slender, rodlike processes with limited verti- cal or lateral extent (Figs. 2, 10), rather than the prominent vertical lamellae of the vomers in extant Boidae (Frazzetta, 1959: 457). In lizards the vomers may have con- siderable lateral expansion (e.g. Lanthano- 1 We have deliberately not included compari- sons with the Scolecophidia (Typhlopidae and Leptotyphlopidae ) . We have felt ( 1 ) that the most evident resemblances were as much lizard- like (or primitive) as scolecophidianlike, (2) that the striking differences were specializations of the Scolecophidia irrelevant to the phyletic position of Dinilysia, and (3) that the Scolecophidia, like the Aniliidae, still need much careful study before much worthwhile can be said about them, although we have included statements about the latter group because of its clear relationship to Dinilysia. tus but not Varanus). In neither snakes nor lizards are the vomers underlain by a portion of the palatines (see below ) as they appear to be in Dinilysia. The vomers of Dinilysia are thus neither snakelike nor spe- cifically lizardlike. In Dinilysia the frontals have lateral de- scending walls that meet to enclose a single median anterior opening. In boid snakes the frontals form a pair of openings by sending down not only lateral but median walls to surround the olfactory tracts. Lat- eral walls descend from the frontals in Varanus and Lanthanotus but not in Cteno- saura, Tupinambis or Mabuya. The pres- ence of lateral descending walls of the frontal in Dinilysia represents a step toward the snake condition and one that would be expected in an intermediate form. The nasofrontal articulation is extensive in Dinilysia and the frontals send elongated lappets anteriorly under the nasals. The overall extensiveness of the contact con- trasts with the type of articulation between frontal and snout complex described in snakes (Frazzetta, 1959, 1966). Evidently snout mobility about the nasofrontal joint was less than, or at least differed from, that of boids. Discussion of the type of kinesis possible in Dinilysia, and the comparison of this with lizard and snake kinesis is, how- ever, deferred to a future paper ( Frazzetta, in preparation ) . The difference from snakes in the extent or kind of mobility between nasals and frontals, and the relationship of nasals and frontals is lizardlike and primi- tive. This difference is surprising only if it is accepted that one of the primary and defining adaptations of snakes is a pro- kinetic joint that permits rotational move- ments (see Albright and Nelson, 1959; Frazzetta, 1966). The anteriormost element in the post- orbital region is probably homologous to the postfrontal of lizards; the posterior probably is the postorbital. There has been some uncertainty regarding homology of the saurian postfrontals and postorbitals Dinilysia Cranium • Estes, Frazzetta, and Williams with the ophidian elements in this region. The parietal downgrowths in Dinilysia are The relationships of the two bones present unequivocally snake characters. in this region of Dinilysia have the same In Dinilysia the descending lamina of relationships to the frontal, parietal, and each parietal forms the anterior border of jugal as do the postfrontal and postorbital the trigeminal foramen as the similar lamina of Tupinambis. Neither bone extends medi- does for the anterior trigeminal foramen in all}7 along the descending frontal wall or Epicrates, Anilius and Cylindrophis, but not contacts the prefrontal as does the pytho- in Python. In the compared lizards, lack of nine supraorbital (Frazzetta, 1959: 461). the descending parietal lamina results in We have thus no hesitation in homologizing the trigeminal nerve taking its exit from these elements with the comparable ones of a notch in the prootic rather than from a lizards. The supraorbital of pythons may foramen. The Dinilysia condition is again be of heterotopic origin; a specimen of a snake character. Python molurus examined by Frazzetta The prootic is large in Dinilysia, larger ( FMNH 100419 ) has several such bones in than in the compared boids and lizards. In the supraorbital area. On the other hand, sharp contrast with the boids, there is only it is equally possible that the pythonine one trigeminal foramen bounded anteriorly supraorbital is the postfrontal of lizards, by the parietal and hence anteriorly placed as suggested by McDowell and Bogert in a position, as before mentioned, com- (1954). The presence in Dinilysia of twin parable to the incisura prooticum of lizards bones having the saurian relationships to and the anterior trigeminal foramen of other skull bones gives the virtue of parsi- Epicrates, Anilius, and Cylindrophis. In mony to the latter interpretation. For rea- this feature Dinilysia is lizardlike and more sons noted in the description, we believe primitive than bold snakes but this feature that a complete postorbital bar was present is plausible in a very primitive snake. in Dinilysia, with the postorbital joining the In Dinilysia the prootic is broadly jugal. A postorbital arch is present in boids notched posteriorly for the relatively enor- but is joined to the palatomaxillary arch mous stapedial footplate, larger relatively only by a ligament; the Dinilysia condition than the very large footplates of Anilius differs importantly in retention of the jugal and Cylindrophis. Unlike these bones in — although a very peculiar jugal (see be- snakes, the otic and occipital bones do not low). The retention of both postfrontal and surround the footplate with a pericapsular postorbital is a primitive squamate feature chamber and the ivhole of the footplate and might plausibly be expected in any is thus exposed as in lizards. The exposed very priynitive snake. stapedial footplate is primitive and lizard- Dorsally the parietals of Dinilysia are like, and, while not found in boids or ad- applied to the occipital and otic bones, as vanced snakes, again might plausibly be in snakes; there is thus no posttemporal expected to occur in a very primitive snake, foramen like that of lizards. In this fea- The proportions of the stapes, with its ture Dinilysia is snakelike. large footplate and small, posterodorsally- The parietals in Dinilysia descend to directed columellar process, are as in provide a complete lateral covering to the Anilius and Cylindrophis, although even in brain, as in snakes. In none of the com- the latter two forms the footplate is not rel- pared lizards is there any trace of a similar atively so large. The relative size of the enclosure of the brain by the parietals. stapes is unexpected and may be special to (In lizards there is also an epipterygoid in Dinilysia. The stapes of Dinilysia has a this region, which is absent in snakes and generally lizardlike aspect, but in exposure of which there is no evidence in Dinilysia.) of footplate, proportions of footplate to 52 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 xl U 2 X m o n oo M a •* o -5 -s c c -= X a) m •> U - s o o ? Q 3 o -c LL. Q_ Dinilysia Cranium • Estes, Frazzetta, and Williams columella)' process, and connection of the latter to the footplate by a strong posterior crest (Fig. 6, etc.), it is specifically Anilius- like. The presence of a small dorsal exposure of the prootic, and its relation to surround- ing bones, is exactly like the situation in Anilius and Cylindrophis (Fig. 9). The massive supratemporals of Dinilysia are completely applied laterally to the par- occipital processes of the exoccipitals. They are like those of Anilius or Cylindrophis in being incorporated into the skull. They differ strikingly from the supratemporals of Anilius or Cylindrophis and resemble those of boids in being drawn out latero- posteriorly to a level well behind that of the occipital condyle, but are unlike those of boids in lacking a long free posterior pro- jection: the paroccipital processes of Dini- lysia are applied to the supratemporals all the way to their distal ends (Fig. 9). The supratemporals of lizards are also very different: small, almost vestigial in Ctenosaura and Mabuya; relatively large in Lanthanotus and Tupinambis. In the latter two, however, most of the slender supra- temporal lies alongside the posttemporal wing of the parietal; only a smaller pos- terior portion overlies the paroccipital proc- ess to make contact with the quadrate. In all compared lizards, posterior termination of the supratemporal lies only slightly be- hind the posterior level of the occipital condyle. The supratemporal of Dinilysia combines lizard and snake (in fact, boid) features in an unexpected way. The strong associa- tion of supratemporal and paroccipital proc- ess is lacertilian, as we have stated, but the way in which the supratemporal is carried well behind the occipital condyle is like the situation in boids and suggests a way in which the slender exposed supratemporal of boids may have evolved: to obtain the snake condition from that characteristic of eosuchians and lizards we need only to have supratemporal and paroccipital proc- ess extend well posteriorly with a post- temporal foramen present; next, as in Dini- lysia, to close the foramen by secondary growth of supraoccipital and opisthotic; and then to achieve the more usual snake condition by secondarily reducing the par- occipital process to the anterior level from which it began. The incoq^oration of such an intermediate stage would not be obvious were it not in front of us. Interpretation of the sequence of functional adaptations that may have been involved in such an evolu- tionary series will necessarily require analy- sis of the mechanics of the entire skull and, if this evolutionary sequence is verified as occurring in the main line of snake origins, will have great importance for our under- standing of the way of life of the first snakes. The supraoccipital in Dinilysia, in its shape and relationships to surrounding bones, is closely comparable to this element in Cylindrophis and Anilius and is relatively much lower than the wedge-shaped element of Python or the vertical blade of Epicrates. In Dinilysia, as in the boids, Anilius, and Cylindrophis, a median knob on the supra- occipital continues the strong sagittal crest on the parietal. In contrast, lizards have either a median crest on the supraoccipital (Ctenosaura, Tupinambis; "processus as- cendens" of Oelrich, 1956) rising to meet the parietal, the whole surface of which is at a distinctly higher level than that of the body of the supraoccipital, or (Lanthano- tus, Mabuya) parietal and supraoccipital are at the same level at the point of contact but a small unossified prong ("processus ascendens tecti synotici" of most authors; "cartilaginous portion of the processus an- terior" of Rao and Ramaswami, 1952) fits into a notch of the parietal. The difference in appearance of the two conditions in liz- ards inter se and as compared with condi- tions in Dinilysia and snakes is considerable. In this regard, Dinilysia is clearly snake- like and specifically Cylindrophis-Anilius- like. 54 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 Figure 10. Palatal region of a, Dinilysia pafogonica, X 1.5, right maxilla and ectopterygoid removed; b, Cylindrophis macu- latus, MCZ 34885, X 4.5; c, Python molurus, MCZ 4278, X 1.5; d, Anilius scytale, MCZ 17645, X 4.5; e, Lanlhanotus borne- Dinilysia Cranium • Estes, Frazzetta, and Williams The cxoccipitals of Dinilysia are in one regard strikingly snakelike: they meet on the midline and exclude the supraoccipital from the edge of the foramen magnum. The flat broad shelf that the exoccipitals form above the occipital condyle is, however, continued smoothly into the strong par- occipital process, which is carried back- ward with the supratemporal behind the occipital condyle, as noted above. This is in contrast to the condition in boids, in which the median exoccipital shelf ends abruptly in a notch separating the shelf from a rudimentary paroccipital process that at its extreme lateral projection does not extend further posteriorly than the ex- treme posterior level of the shelf and the condyle below it (Python), or does not ex- tend further posteriorly than the level of the lateral margins of the foramen magnum (Epicrates) . In lizards the exoccipitals are always well lateral. They do not exclude the supra- occipital from the foramen magnum and may provide a very rudimentary lateral shelf over the foramen magnum (Cteno- saura, Mabuya, Lanthanotus) or none at all (Tupinambis). In lizards the paroccipi- tal process is more robust than in snakes but in no case does it project strongly backward carrying the supratemporal with it behind the level of the occipital condyle, as in Dinilysia. As in both lizards and snakes, each ex- occipital in Dinilysia forms a lateral third of the occipital condyle. As in Python and Epicrates, a lappet of exoccipital extends posteriorly underneath the moderate-sized posterior lacerate foramen, which faces di- rectly posteriorly. In Anilius and Cylin- drophis the posterior lacerate foramen faces more laterally than posteriorly and the bony process underneath is likewise posterolat- eral. Of the lizards compared, only Lan- thanotus has a ledge beneath the posterior lacerate foramen and this forms a shallow trough leading posterolaterally (Fig. 12). The exoccipital in Dinilysia participates in the dorsal margin of a large, widely-open fenestra rotunda that is directed ventro- ensis, MCZ 8305, X 4.5; f, Tupinambis nigropunctatus, MCZ 109890, X about 2.5. Abbreviations on p. 62. Semidiagram- matic. 56 Bulletin Museum of Comparative Zoology, Vol 140, No. 2 VII Figure 11. Lateral view of posterior skull region of a, Dinilysia patagonica, X 1.5; b, Anilius scytale, MCZ 17645, X 4.5; c, Cylindrophis maculatus, MCZ 34885, X 4.5; d, Python molurus, MCZ 4278, X 1.5; e, Lanthanotus borneensis, MCZ 8305, X 4.5. Abbreviations on p. 62. Semidiagrammatic. laterally. In snakes (cf. Baird, 1960), the fenestra ovalis (partly) and fenestra ro- tunda (wholly) are enclosed and almost walled off from view by a crista circum- fencstralis, except in Anilius and Cylindro- phis, in which the crista is not prominent and the stapedial footplate relatively large. The crista circumfenestralis of snakes appears to correspond to the crista tuber- alis of lizards (Save-Soderbergh, 1947: 512; see also Oelrich, 1956: 1-17). The latter crista begins at the root of the paroccipital process and extends ventrally to the spheno- occipital tubercle (Oelrich), forming a pos- terior wall for the fenestra rotunda. The snake condition can be derived from that of lizards by the growth forward and up- ward of the crista tuberalis until the fenes- tra rotunda is deeply concealed at the base of the crista, which then partly occludes the fenestra ovalis and encloses the stape- dial footplate. The crista tuberalis in Dini- lysia does not fully enclose the stapedial footplate and the fenestra rotunda is thus widely open as in Anilius, Cylindrophis, and lizards, but the fenestra rotunda is strikingly low on the occiput (below, in- stead of on the level of the condyle), a condition not found in the lizards and a fortiori not in snakes (Fig. 12). Dinilysia Cranium • Estes, Frazzetta, and Willia Figure 12. Oblique posterior view of occiput of a, Dinilysia patagonica, X 1-5, hatched lines = breakage; b, Anilius scytale, MCZ 17645, X 4.5; c, /.anfhanofus borneensis, MCZ 8305, X 4.5; d, Python molurus, MCZ 4278, X 1.5. Right quadrates miss- ing in a and c. Abbreviations on p. 62. Semidiagrammatic. The absence of a crista circumfenestralis in Dinilysia is presumably a primitive con- dition and thus again is very plausible in a primitive snake. The low position of the fenestra rotunda, on the other hand, is clearly associated with the relatively enor- mous size of the stapedial footplate and may, like the latter, be special to Dinilysia. The basioccipital in Dinilysia, as is usual in lizards and snakes, provides a median third of the occipital condyle. As in lizards, there are large, projecting spheno-occipital tubera forming between them a concave channel on the ventral surface of the basi- occipital. The very large extent to which the basioccipital participates in the poste- rior as well as ventral wall of the fenestra rotunda is unusual. In snakes the crista circumfenestralis, which is wholly of ex- occipital origin, widely separates the basi- occipital from the fenestra rotunda. In Lanthanotus the basioccipital forms only a small part of the lower edge of the fenestra, and this narrow participation seems the usual lizard condition. The extensive par- ticipation of the basioccipital in the wall of the fenestra rotunda in Dinilysia is closer to the lizard than the snake condition. The laterally-placed basipterygoid proc- esses of the basiparasphenoid have convex articulating surfaces that face more laterally than ventrally, and the axis of the processes is anteroposteriorly elongated. This situa- tion is in contrast to the closely spaced, ventrally-facing, lateromedially-elongated structures seen in boids, and more closely resembles the lizard configuration. On the other hand, the processes in lizards have a narrower stem and an expanded distal artic- ulation in contrast to the broad, even width of the processes in Dinilysia. In Anilius and Cylindrophis the basipterygoid articulations 58 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 are lateral and anteroposterior^ elongated, lizards and seems more primitive than as in Dinilysia, but the articulations in boids, but shows both some special resem- these two snakes are scarcely raised from blances to Cylindrophis and Anilius and the body of the basiparasphenoid. Here the some unique features as well, resemblances are partly to lizards and The dorsal surface of the palatine en- partly to the Anilius-Cylindrophis complex, closes the posterior opening of the maxillo- The wide bladelike cultriform process of palatine foramen in Dinilysia, as in Python, Dinilysia, gently concave ventrally, bears Lanthanotus, Cylindrophis, and Varanus but a striking resemblance to that of Cylindro- not as in Mabuya, Tupinambis or Anilius. phis. Even Anilius has a narrower, if stout, In Ctenosaura, according to Oelrich, this cultriform process, while those of lizards foramen may be wholly within the palatine are feeble and thin as well as narrow, fre- or may be completed laterally by the max- quently distorted and curved upward in the ilia or the jugal or both. This character dry skull. The strength of the cultriform appears to have little value for the place- process in Dinilysia is snakelike but the ment of Dinilysia or any other form (Fig. specific resemblance is, as stated, to that 8). of Cylindrophis (Fig. 10). The tooth row on the posterior prong The prefrontals in Dinilysia have a wide, of the palatine of Dinilysia is gently curved firm, interlocking contact with both frontals and is continued by a row likewise gently and maxillae, much as in Anilius and Cylin- curved on the pterygoid. This is unlike the drophis. In contrast Epicrates and Python straight rows (and much larger teeth) of have only a loose contact between prefron- snakes. The curvature is like that of Lan- tals and maxillae. In lizards there are, of thanotus, but the pterygoid teeth are very course, firm sutures with both bones but much smaller in the latter. Other lizards the prefrontal has little lateral exposure, are even more dissimilar. It is important being covered by the ascending process of that on the pterygoid and palatine in Dini- the maxillae. In Dinilysia, as in Lanthano- lysia teeth appear to be enlarging and ap- tus, Varanus and snakes, the prefrontals proaching the snake condition, but clearly intervene between the maxillae and nasals, in both size and in the curvature of the row The resemblance again is snakelike but they are some steps behind the snake condi- specifically to Cylindrophis and Anilius tion. The implication would appear to be, rather than boids (Fig. 8). as with the absence of the anterior toothed The palatines of Dinilysia are strikingly prong of the palatine, that in Dinilysia the unsnakelike in the absence of anterior characteristic method of snake feeding was toothed projections. In this respect they are not yet perfected. like the palatines of lizards. Like lizards The posterior quadrate process of the they have strong, wide medial processes pterygoid of Dinilysia is a vertical plate to the vomers. Unlike Lanthanotus, Vara- with a sharp ventral edge as in lizards and nus, or Ctenosaura but like Tupinambis, Cylindrophis, not as in boids and Anilius, these medial processes are deeply exca- which have the ventral edge rotated medi- vated for the choanal passages. Rather ally so that the process is no longer a veri- similar medial processes of the palatines cal plate. It is interesting here that there occur in Anilius and Cylindrophis, but, as is a resemblance to Cylindrophis as well as in Tupinambis, these processes at their to lizards. medial terminations lie between or dorsal The maxilla of Dinilysia is snakelike in to the posterior ends of the vomers; they its long, gently tapering anterior end and do not broadly underlie them as in Dinilysia slight ascending process meeting the pre (Fig. 10). Here Dinilysia resembles some frontal; its general shape is specifically like Dinilysia Cranium • Estes, Frazzetta, and Williams that of Cylindrophis. The two teeth that are preserved are comparable to those of snakes in their sigmoid conical form and in their implantation. The snake resem- blance is clear, and within snakes the re- semblance is to Cylindrophis. The ectopterygoid in Dinilysia is a short and simple bone joining the lateral process of the pterygoid and the maxilla. If any slender process extended along the medial surface of the maxilla to provide much of the dorsolateral rim of the suborbital fe- nestra, as in boids, especially Python, it has not been preserved. There may have been a short anteromedial process, as in Cylin- drophis and Anilius. In Tupinambis such a process excludes the maxilla from the margin of the suborbital fenestra. In Vara- nus and Lanthanotus the maxilla is simi- larly excluded from the suborbital fenestra, but this seems to be a result of the shorten- ing of the maxilla (see McDowell and Bogert, 1954) rather than of any long anterior prong of the ectopterygoid. In other lizards a short anteromedial process is present: this is presumably a primitive character. The Dinilysia ectopterygoid seems unspecialized and, if taxonomically important, is so only in its differences from certain forms, not in its resemblances. The element we interpret as a jugal is absent in snakes and is quite unlike any lizard jugal in form and position. Its rounded shape, and its occurrence in a trough on the dorsal surface of the maxilla, are unique. In lizards the jugal is charac- teristically larger and has more extensive contacts, not only with the maxilla as in Dinilysia, but with the ectopterygoid (all examined cases) and with the postfrontal or postorbital and sometimes (Ctenosaura, Mabuya) the squamosal. According to Jollie (1960), reduction or loss of the jugal accompanies loss of orbital and temporal arches as in burrowing forms or geckos. In snakes, both arches are gone but, if our interpretation of Dinilysia (as influenced by the original British Museum cast) is correct, the jugal did in this form make contact with the postorbital and completed the orbital arch. It is thus peculiar, primar- ily in its shortened and compact form, lack of contact with the ectopterygoid, and in- sertion in a groove of the maxilla. The pres- ence of a jugal in Dinilysia is a primitive character and is thus expected in a very primitive snake. Its special features are, however, wholly surprising and require special functional study. At the moment, it would seem more probable that these special conditions are peculiar to a special- ized side branch and are not part of the main line transition in the shift from lizard to snake adaptive zones. The extensive attachment of quadrate to paroccipital process is unlike that of any boid. Anilius and Cylindrophis approach Dinilysia in degree of expansion of the head of the quadrate, and in the short and dorso- ventrally-compressed shape of the bone. The strongly curved tympanic process of the quadrate is unlike that of any other squamate, but is approached in Anilius and Cylindrophis. The oval distal quadrate articidation of Dinilysia is a unique feature and does not resemble the irregular articu- lation siwface of lizards, or tJie related, more saddle-shaped snake condition (see below, articular, and Frazzetta, 1959, figs. 1, 2). Again functional considerations and inteqoretations are important, but there now seems to be no reason to postulate the transformation of the lizard joint into the Dinilysia condition, and then to return to the saddle-shaped joint of snakes as part of the lizard-snake transition; this condition seems again to emphasize that Dinilysia is a specialized side branch of the early booid radiation. The dentary is like that of snakes in hav- ing a strong reentrant notch posteriorly for the surangular. No specific resemblances to lizards occur, and the general aspect is more as in boids than as in Anilius or Cylindrophis. In this feature Dinilysia re- sembles the snakes. 60 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 The splenial is small and meets the angu- lar in a vertical suture. The resemblance here is to snakes, although the vertical splenial-angular suture occurs also in Lan- thanotus. The angular, in its shape, position, and articulation with the splenial is like that of snakes. The compound bone is complex and its resemblances less clear-cut than those of the dentary. It resembles that of snakes in having a narrow surangular projection fit- ting loosely into the dentary notch, and in lacking ( apparently ) a retroarticular proc- ess. It is unique in having a small mandibu- lar fossa that is dorsal and dorsolateral in orientation, cut off from medial exposure by a strong ridge on the prearticular area, and exposed laterally by the absence of such a ridge on the surangular region. The pre- articular ridge occurs in snakes, but it is relatively larger in Dinilysia and more acutely angled. The lack of a strong sur- angular ridge for the coronoid differs from the condition in any snake or lizard, In Dinilysia, while the relations with the den- tary and the apparent absence of a retro- articular process constitute a general ap- proach to snake conditions, the mandibular fossa and articular joint form one of the most peculiar and unique features of this animal. The coronoid is in great part missing, but it has left articulation surfaces on the com- pound bone that allow some interpretation of its extent. In its relationship to the man- dibular fossa, it shares some of the unique features of the latter. Two distinct articu- lar surfaces occur at the anterior end of the fossa, one on the medial (prearticular) ridge, the other a deep notch immediately lateral to the fossa. This suggests that the coronoid forked narrowly around the an- terior border of the fossa. While the fork- ing is a lizard feature, the close apposition of the two parts of the fork is unique to Dinilysia. The main body of the coronoid projected vertically, forming a coronoid process of unknown extent, and then cul- minated anteriorly in a strong, flat ventro- medial strap of bone that is all that remains today of the coronoid bone itself. This preserved part of the coronoid is as in snakes and lizards generally, except in Anilius, in which the coronoid is almost vestigial. SUMMARY Dinilysia is conspicuously a mosiac of primitive and specialized characters. It has lizardlike features and also some star- tling and unique peculiarities, but it is also clearly a snake and it has some detailed resemblances to the modem primitive snakes Anilius and Cylindrophis that seem difficult to dismiss as convergent. The snakelike features— in fact booid snakelike — are as follows: (1) absence of temporal arches; (2) probable loose con- nection of premaxilla and maxilla; (3) pres- ence of a bony nasal septum; (4) ventral enclosure of the brain by both frontal and parietal downgrowths; (5) absence of a posttemporal foramen; (6) parietals level with oto-occipital complex; (7) descending process of parietal completing trigeminal foramen anteriorly; (8) supraoccipital par- ticipation in sagittal crest; (9) exoccipitals excluding supraoccipital from foramen mag- num; (10) lappet of bone present behind posterior lacerate foramen; (11) prefrontals intervening between maxillae and nasals; (12) reduced ascending process of maxilla; (13) strong, movable reentrant articula- tion of dentary and surangular on lateral side of mandible. Characters more specifically similar to the Anilius-Cylindrojmis complex are: (1) relatively large size of stapes, especially the footplate; (2) the peculiar dorsal exposure of the prootic; (3) wide-bladed Cylindro- phis-like cultriform process of the para- sphenoid; (4) prefrontal having a firm union with maxilla and frontals; (5) pos- terior (quadrate) process of pterygoid a vertical plate as in Cylindrophis; (6) quad- Dinilysia Cranium • Estes, Frazzetta, and Williams rate a dorsoventrally compressed bone with an expanded head. These resemblances to booid snakes are numerous, clear, and detailed; we believe that they require Dinilysia to be placed in the booid complex close to Anilius and Cylindrophis. Nevertheless, there are also a number of very primitive features, quite in keeping with the Cretaceous age of the fossil, that make Dinilysia more lizardlike than any other known snake. Generalized lizard fea- tures are: (1) frontal lappets under nasals, limiting nasofrontal mobility; (2) postfron- tal and postorbital both present; (3) tri- geminal foramen single; (4) exposure of the stapedial footplate ( = absence of a crista circumf enestralis ) ; (5) large, lat- erally-projecting convex basipterygoid proc- esses; (6) palatines with deep choanal grooves and without anterior toothed pro- jections; (7) posterior processes of ptery- goids vertical plates, concave medially; (8) jugal present ( but see below ) ; ( 9 ) a single opening in the anterior braincase for ol- factory tracts. These primitive lizardlike features are again numerous enough and impressive enough that, if Dinilysia is related to boids and aniliids, it must belong rather far down in the ancestry of the booid complex. In its lizardlike aspects, it must be em- phasized, Dinilysia does not show special affinity to any group of lizards. The lizard characters cited above are generalized and are primitive for squamates; the more de- tailed resemblances seem casual and ran- dom, inviting the suspicion that they are merely convergent. There are some specific resemblances to Lanthanohis or Varanus: (1) a somewhat L a ntha not us -like crista tuberalis behind the posterior lacerate fora- men; (2) a Lanthanotus-\ike curvature of the anterior (palatine) processes of the pterygoids; (3) prefrontals *that intervene between maxillae and nasals. But these are balanced by the ways in which Dinilysia appears to be more primitive than Lantha- notus or Varanus- e.g. the deep choanal im- pressions on the palatines and participation of the maxilla in the suborbital fenestra (ways in which primitive snakes generally seem more primitive than Lanthanotus) . There is neither special continuation nor denial in our study for an anguimorph origin of snakes ( cf. McDowell and Bogert, 1954 ) . However, it is very possible that the expectation that Dinilysia will shed much light on snake origins is erroneous. It is perhaps already too close to extant groups of snakes to be helpful. Yet, snake though it is and close though it seems to Anilius and Cylindrophis, there are aspects in which Dinilysia has provided wholly astonishing features. The strongly posterolaterally-produced paroccipital and supratemporal processes have been inter- preted as an intermediate between lizard and snake conditions. If it is really an intermediate, it is one quite outside expec- tation. Beyond this, however, there are other aspects of Dinilysia that are special and unique: (1) the vomers underlain by anterior portions of the palatines; (2) the extraordinarily robust nasal septum; (3) the low position of the fenestra rotunda on the occiput; (4) the peculiar shape of the quadrate, as well as the oval quadrate- articular joint; (5) the almost incredible element that we call the jugal, and ( 6 ) the large supratemporal applied closely to the skull. These are so special, and are so far from being in any simple sense intermediate between lizard and snake, that Dinilysia seems likely to be off on its own side branch and away from the main line of the ancestry of modernized snakes. Related to booids it seems to be, and among these is closest to Anilius and Cylindrophis, yet it appears also to have been a very early and eccen- trically divergent offshoot of this stock. Taxonomic expression of these conclusions need involve no change from that already given by Romer (1956: 570), who raised Dinilysia to family status and placed it at the base of the superfamily Booidea. 62 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 REFERENCES CITED Albright, R. G., and E. M. Nelson. 1959. Cranial kinetics of the generalized colubrid snake Elaplie obsoleta quadrivittata. I. De- scriptive morphology. J. Morph. 105: 193- 240. II. Functional morphology. J. Morph. 105: 241-292. Bahl, K. N. 1937. Skull of Varanus monitor (Linn.). Rec. Indian Mus. 39: 133-174. Baird, I. L. 1960. A survey of the periotic labyrinth in some representative Recent rep- tiles. Univ. Kansas Sci. Bull. 41: 891-981. Bellairs, A., and G. Underwood. 1951. The origin of snakes. Biol. Reviews 26: 193-237. Estes, R. 1966. Anatomy and relationships of the primitive fossil snake Dinilysia. Year Book of the American Philosophical Society 1966: 334-336. Estes, R., T. Frazzetta, E. E. Williams, and M. Hecht. 1966. Abst. paper given at 46th An- nual Meeting, Am. Soc. Ichth. Herp. Feruglio, E. 1949. Descripcion geologica de la Patagonia 1 : xv + 334 pp. Buenos Aires. Frazzetta, T. H. 1959. Studies on the mor- phology and function of the skull in the Boidae (Serpentes). Part I. Cranial differ- ences between Python sebae and Epicrates cenchris. Bull. Mus. Comp. Zool. 119: 453- 472. . 1966. Studies on the morphology and function of the skull in the Boidae (Serpentes). Part II. Morphology and func- tion of the jaw apparatus in Python sebae and Python molurus. J. Morph. 118: 217- 296. Jollie, M. I960. The head skeleton of the lizard. Acta Zool. 41: 1-64. McDowell, S. B., Jr., and C. M. Bogert. 1954. The systematic position of LantJmnotus and the affinities of the anguinomorphan lizards. Bull. Amer. Mus. Nat. Hist. 105: 1-142. Oelrich, T. M. 1956. The anatomy of the head of Ctenosaura pectinata (Iguanidae). Misc. Publ. Mus. Zool., Univ. Michigan No. 94: 1-122. Rao, M. K. M., and L. S. Ramaswami. 1952. The fully formed chondrocranium of Mabnya with an account of the adidt osteocranium. Acta Zool. 33: 209-275. Save-Soderbergh, G. 1947. Notes on the brain- case in Sphenodon and certain Lacertilia. Zool. Bidrag 25: 489-516. Underwood, G. 1957. On the lizards of the family Pygopodidae. A contribution to the morphology and phylogeny of the Squamata. J. Morph. 100: 207-268. Woodward, A. S. 1901. On some extinct rep- tiles from Patagonia of the genera Miolania, Dinilysia and Genyodectes. Proc. Zool. Soc. Lond. 1901 1: 169-184. (Received 18 November 1968.) ABBREVIATIONS an r= angular art = articular avc = anterior opening of vidian canal bo = basioccipital bp = basipterygoid process bs = basisphenoid ca = coronoid articulation surface co = coronoid cp = compound bone de = dentary ec = ectopterygoid eo = exoccipital f = unknown foramen fo = fenestra ovalis fp r= foramen for palatine artery fr = frontal fro = fenestra rotunda ju = jugal If = lacrimal foramen nip = maxillopalatine foramen mx = maxilla na = nasal oc = orbitonasal canal of = optic fenestra pa = parietal pal = palatine pf = postfrontal plf = posterior lacerate foramen po = postorbital pr = prootic pra = prearticular prf = prefrontal ps = parasphenoid pt = pterygoid pvc = posterior opening of vidian canal qu = quadrate s = stapes sm = septomaxilla so = supraoccipital sot = spheno-occipital tuber sp = splenial st = supratemporal sur = surangular t = trabecular pit and groove vo = vomer V = trigeminal foramen VII = facial foramen Dinilysia Cranium • Estes, Frazzetta, and Williams 64 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 Plate 1. Dinilysia paragom'ca; original plate from Woodward (1901); 1, dorsal and left lateral views of skull; lb, ventral view of left maxilla showing tooth sockets; lc, lateral view of left quadrate; 2, a portion of the vertebral column, dorsal view; all X 1. Abbreviations used on this original plate may not coincide with those on p. 62. Dinilysia Cranium • Estes, Frazzetta, and Williams orb ul ' I ■ nr ■in 11)11 111 .r 66 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 Plate 1. Dinilysia patagonica; right lateral and ventral views of copy of British Museum (Natural History) cast of original specimen of Dinilysia. Note complete postorbital arch and complete maxilla. X 1-5. Dinilysia Cranium • Estes, Frazzetta, and Williams 6S Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 Plate 3. Dinilysia pafagonico; dorsal view of skull. Dinilysia Cranium • Estes, Frazzetta, and Williams 70 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 Plate 4. Dinilysia patagonica; ventral view of skull Dinilysia Cranium • Estes, Frazzetta, and William : 72 Bulletin Museum of Comparative Zoology, Vol. 140, No. 2 Dinilysia Cranium • Estes, Frazzetta, and Williams c o O) D D a Q uLietin of the seum The Lynx Spider Genus Hamataliwa in Mexico and Central America (Araneae: Oxyopidae) ALLEN R. BRADY HARVARD UNIVERSITY VOLUME 140, NUMBER 3 CAMBRIDGE, MASSACHUSETTS, U.S.A. AUGUST 6, 1970 PUBLICATIONS ISSUED OR DISTRIBUTED BY THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Breviora 1952- Memoibs 1864-1938 Johnsonia, Department of Mollusks, 1941- Occasional Papers on Mollusks, 1945- Other Publications. Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine. Reprint, $6.50 cloth. Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of In- sects. $9.00 cloth. Creighton, W. S., 1950. The Ants of North America. Reprint, $10.00 cloth. Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 12, 14, 15. (Price list on request.) Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Teredinidae (Mollusca: Bivalvia). $8.00 cloth. Whittington, H. B., and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948. ( Complete sets only. ) Publications of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138, U. S. A. © The President and Fellows of Harvard College 1970. THE LYNX SPIDER GENUS HAMATALIWA IN MEXICO AND CENTRAL AMERICA (ARANEAE: OXYOPIDAE) ALLEN R. BRADY1 ABSTRACT Spiders of the genus Hamataliua in Mexico and Central America are described and illustrated. A new definition of Hama- taliwa includes eight species originally assigned to Oxyopeidon and seven species originally assigned to Oxyopes. Four spe- cies of Hamataliua from Mexico and Central America are retained in the genus and five species are described as new. In short, 24 species of Hamataliwa are re- corded from Mexico and Central America, where previously only four were reported. The genus Hamataliwa, undoubtedly, will prove to be as widespread as Oxyopes when additional studies in tropical regions are completed. Three species groups of Hamataliwa have been established on the basis of positive correlation between eye arrange- ment and the structure of the palpus and/or epigynum. The banksi group con- j sists of seven species, the puta group consists of eight species, and the grisea : group consists of three species. On the basis of present information no clearly de- fined relationship could be established for J the remaining six species of Hamataliwa. Although distribution data are scarce, records and maps are given for those speci- mens examined. Much more work needs to be done in Mexico and Central America to provide a clear picture of species ranges. 1 Department of Biology, Hope College, Hol- land, Michigan. In general, intraspecific populations of oxyopids in this region tend to be much more variable than comparable groups from North America, north of Mexico. INTRODUCTION This paper is an outgrowth of an earlier work on the Oxyopidae of North America, north of Mexico (Brady, 1964). In that investigation 17 species of oxyopids rep- resenting three genera were recorded and described from North America. For that study, I examined numerous specimens of Neotropical oxyopids to determine the geographic range of the North American species. I uncovered problems of inade- quate descriptions and figures, as well as numerous errors in systematic placement. The present study is primarily an effort to correct this situation and to establish a foundation on which future investigations might be based. The collections examined from Mexico and Central America were not extensive, and although collecting has been concentrated only in certain areas of this region (for example, Barro Colorado Island), I judged the amount of material adequate. Because the number of oxyopid species increases considerably as one moves southward into Mexico and Central America, it seemed advisable to report on the genera in this area individually rather than to treat the entire family in one monograph. This paper is the first in a series I plan on the Neotropical oxyopids. In addition to shedding some light on the Bull. Mus. Comp. Zool., 140(3) : 75-128, August, 1970 75 76 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 evolutionary relationships of the species involved, I hope that this study will pro- vide pertinent information about the dis- tributional patterns of spiders in the Neo- tropical Region. In the present investigation, I cover eight species found in Mexico and Central America that were originally described in the literature as Oxyopeidon. In addition, I have placed in Hamataliica seven species originally assigned to Oxyopes and re- corded from that region. These changes are based on an intensive study that neces- sitated a redefinition of Hamataliica (Brady, 1964). This new diagnosis indi- cated both that Oxyopeidon was a synonym of Hamataliica and that certain species placed in Oxyopes were much closer to Hamataliica than investigators previously thought. Four species of Hamataliica from Mexico and Central America remain in the genus, and five species are newly described in this paper. It is likely that numerous species of Oxyopes and the remaining species described in the literature under Oxyopeidon belong in Hamataliica as characterized here. After additional studies are completed, the genus Hamataliica will undoubtedly prove to be as widespread as Oxyopes. For example, in a recent work on spiders from south New Guinea, Father Chrysanthus ( 1967 ) figures Oxyopes tap- poniformis Strand. The figures, as well as measurements provided by the author, indicate that this species belongs in Hamataliica. The revision of Hamataliica as it is treated here is based primarily on morpho- logical evidence. Although my approach to delimiting species is based on morpho- logical distinctness, I have considered carefully other factors, such as ecological amplitude and individual variation demon- strated in field investigations of the North American species. I have used Hamataliica grisea and //. helia, two of the more closely studied American representatives of this group, as "standards" for testing assump- tions and for drawing conclusions about reproductive isolation. My preliminary studies of the behavior of these two species in the field and their natural history have been reported elsewhere (Brady, 1964). I have considered other factors which I do not yet fully understand. For example, it appears that members of a single popu- lation of the same species in Mexico and Central America tend to varv more than their North American counterparts. Per- haps this is a result of the increased interspecific competition among tropical populations, or it may simply be a conse- quence of local diversity in the physical environment. This intraspecific variation must, however, be considered in judging the significance of differences in allopatric populations. In this day of molecular analysis and comparative behavioral studies, the value of a strict morphological approach to system- atic problems may be questioned, but one must lay a foundation at some point. The time necessary to gather information to establish this base is a primary question. How long would it take to acquire enough ecological, behavioral, or molecular data so that one would have sufficient evidence to modify the conclusions drawn from morphology? For the Neotropical Oxyopi- dae, it would take months and perhaps years. I wish to make clear that I am not arguing against the acquisition of addi- tional information from ecological, be- havioral, and molecular studies, nor am I questioning the value of data from these areas. I hope that this paper might stimu- late further investigations in ecology, be- havior, and molecular analysis. I am emphasizing the need for presenting basic morphological revisions where adequate numbers of specimens are available and qualified systematists are present. I feel that morphological studies cognizant of the factors mentioned above provide an ade- quate basis for establishing genetic relation- ships and that such studies provide a firm foundation on which to build future inter- pretations of phylogeny. Because this Hamataliwa in Mexico and Central America • Brady 71 morphological study modifies considerably the findings of earlier authors, and because it clears up some difficult nomenclatural problems and consolidates scattered bits of information, I felt that it should be pre- sented without further delay. ACKNOWLEDGMENTS I initiated this study in 1964 while I was a Research Fellow in Arachnology at the Museum of Comparative Zoology, Harvard University. This appointment was under the auspices of a grant from the Evolution- ary Biology Committee. I am especially grateful to Dr. H. W. Levi of the Museum of Comparative Zoology, who has offered advice and aided in many ways the prep- aration of this paper. Collections from the Museum of Comparative Zoology were available throughout this investigation. Much of this material was collected by Dr. A. M. Chiekering, and it is through his efforts in the field that the study became practicable. I am also indebted to Dr. W. J. Gertsch for collections from the Ameri- can Museum of Natural History. I wish to thank Dr. G. Owen Evans and Mr. Douglas Clark, whose hospitality I enjoyed for three weeks in the summer of 1963 at the British Museum (Natural History). A grant from the Evolutionary Biology Committee, Biological Labora- tories, Harvard University, made this visit possible. Type specimens of O.P.- and F.O.P.-Cambridge were drawn and ex- amined at that time. As my investigation progressed, I realized that critical measure- ments and additional drawings would be necessary to diagnose properly certain of the Cambridges' types. I appreciate the further courtesy of Dr. Evans and Mr. Clark for making this possible. I would also like to thank Dr. O. Kraus of the Senckenberg Museum, Frankfurt, and Dr. E. Kritscher of the Natural History Museum, Vienna, for making available critical type specimens. Father Chrysan- thus aided by checking the proper Latin endings for many of the specific names. A 1967 Summer Faculty Grant from Hope College allowed much needed time for the preparation and writing of this paper. National Science Foundation Grant GB- 13925 helped to defray expenses con- nected with this study. METHODS The methods for measuring specimens during this study were essentially the same as those I employed in my earlier paper on the family Oxyopidae (Brady, 1964). The color descriptions and illustrations are based on alcoholic specimens that were in reasonably good condition (except where noted to the contrary). Locality records are listed geographically in a sequence from north to south and from east to west. The number of specimens collected at each localitv is indicated; the lower case "o" represents immature speci- mens. For most species, the face view as well as the dorsal view of a male and female were drawn (when both were available). A ventral external view of the epigynum (after all of the hair had been removed) was drawn. This drawing often reveals some internal structures through the integu- ment. In addition, a dorsal internal view with the genitalia separated from the spider and submerged in clove oil was figured. The female genitalia of all species are drawn to the same scale. The scales are indicated on the plates. Two views, a ventral and a lateral, were drawn for each species. These were drawn after the palpus had been gently scraped free of hair to reveal as clearly as possible the palpal sclerites and patellar or tibial apophyses. No attempt was made to indicate spination or hairiness in the drawings. All palpi are drawn to the same scale. SCIENTIFIC NAMES OF UNCERTAIN STATUS R. V. Chamberlin (1924) described one new species and one new subspecies of Oxyopeidon from the shores and islands 78 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 of the Gulf of California. Immature specimens of Oxyopeidon absolutum were collected from San Esteban Island, Con- cepcion Bay, Puerto Escondido, Angel de la Guarda Island, and San Josef Island. There are no distinguishing characteristics that differentiate these specimens from im- mature Hamataliwa grisea. A geographic race, Oxyopeidon absolutum obliquum was described from Coronados Island because of different coloration than other speci- mens of O. absolutum. The holotype is an early instar of Hamataliwa. Hamataliwa grisea varies considerably in coloration, as do other species of Hamataliwa, and color- ation alone does not warrant subspecific recognition. Until mature specimens are collected from the above localities along the shore and on islands of the Gulf of California, it seems best to consider Oxyopeidon absolutum as a synonym of Hamataliwa grisea. Reimoser (1939) described two new species of Hamataliwa from San Jose, Costa Rica. One of these, H. schmidti, is newly described and figured in this study. The other species, H. tristani, is based on two female specimens supposedly de- posited in the Natural History Museum, Vienna. These two specimens were un- available for study, and the original de- scription and Reimoser's sketch are not sufficient to provide an accurate determin- ation of //. tristani. Hamataliwa Keyserling Hamataliwa Keyserling, 1887, Verh. Zool.-Bot. Ges. Wien, 6:458, fig. 24, 9 . Type species by monotypy: Hamataliica grisea Keyserling, op. cit., 6:458, fig. 24, 9, from North America in British Museum (Natural History), examined. Oxyopeidon O. P. -Cambridge, 1894, Biologia Centrali-Americana, Arachnida, Araneidea, 1: 139. Type species designated by F.O. P. -Cam- bridge, 1902, Biologia Centrali-Americana, Arachnida, Araneidea, 2:346; Oxyopeidon putum O.P.-Cambridge, 1894, op. cit., 1:140, in British Museum (Natural History), examined. Characteristics. For general character- istics of the genus refer to Brady (1964), p. 496. Diagnosis. In Hamataliwa the eye rows differ in position and/or proportional width from those of Oxyopes. The face is often not vertical as it is in other oxyopids, but slopes more gradually toward the clypeus. In many species of Hamataliwa, the carapace is clothed with long hair, often with tufts in the eye region. In addition, there may be long hairs on the lateral surfaces of the legs and along the sides of the abdomen. These features add to the cryptic effect offered by their color- ation and provide concealment against bark of trees and twigs or against woody shrubs. Many species are undoubtedly arboreal in habit. Hamataliwa seems to be as well defined ecologically as it is morphologically. Leg development and structure appear to be correlated with arboreal habits. In all species studied, except H. tricuspidata, the relative leg length is I-II-III-IV. The first two pairs of legs are long and robust, the third and fourth pairs weakly de- veloped. In Oxyopes the fourth pair of legs is strongly developed concomitant with their jumping habits. Observed species of Hamataliwa are more sedentary than Oxyopes. The general form of the epigyna in Hamataliwa is a semi-circular or U-shaped, heavily sclerotized rim surrounding a shal- low median depression with a character- istic shape in each species. Male palpi are also similar in basic construction, with the embolus following a definite route and forming a characteristic twist or loop near the base at the mesal edge of the cymbium. The above combination of character- istics distinguish members of Hamataliwa from Oxyopes. SPECIES GROUPS OF HAMATALIWA Mexican and Central American species of Hamataliica can be separated into several groups based on the comparative width of the eye rows and the position of certain eyes relative to others. I made comparisons of the structure of the geni- Hamataliwa in Mexico and Central America • Brady 79 talia, bodily proportions, and coloration of those species that have similarities in eye arrangement. Most species of Hamo- taliica can be placed in species groups based on a positive correlation between eye arrangement and the structure of the palpus or epigynum. Color patterns and bodily proportions are also useful, but they are not as reliable in preserved specimens. Although the species groups thus estab- lished may not be strictly natural assem- blages, they do include species that have certain common characteristics and are apparently related. A few species are arbitrarily included in a particular species group because of a similarity in eye ar- rangement. In these cases we know only one sex; the discovery of the other sex will determine whether or not the placement is valid. Banksi group. In the banksi group (banksi, helxa, brunnea, triangularis, bar- roana, globosa and cheta) the ALE row is wider than or subequal to the PME row. Hamataliwa banksi, II. helia, H. brunnea, and H. triangularis have the ALE row wider than the PME row. Of these four species, H. banksi and H. helia are very closely related (see discussion under H. banksi). Hamataliwa brunnea agrees closely with II. banksi and II. helia in eye arrangement (compare Fig. 39 with Fig. 3), but the epigynum of H. brunnea is different (compare Fig. 59 with Figs. 54- 58). Although the epigynum of //. tri- angularis is quite distinct from that of other members of this group, the eye ar- rangement resembles that of H. banksi and the palpus of the male bears a close re- semblance to that of II. helia (compare Figs. 120, 121 of this paper with figs. 130- 133 of Brady, 1964). Hamataliwa barroana, H. globosa, and H. cheta have the ALE row subequal to the PME, i.e. the PME row is less than .05 mm wider than the ALE. This eye ar- rangement is much nearer to that of mem- bers of the banksi group than to that of other species of Hamataliwa. The epigynum and internal genitalia in H. barroana bear a strong resemblance to those of II. banksi (compare Figs. 60-62 with Figs. 54-58). Hamataliwa cheta has an epigynum similar to that of H. barroana. Hamataliwa globosa is included in this group because of the eye arrangement. The palpus of H. globosa (Figs. 122-123) dis- tinguishes it from all other species of Hamataliwa. Puta group. In the puta group (puta, ursa, cavata, hista, flebilis, difficilis, laeta, crocata ) , the PME row is much wider than the ALE row. These species have the PME much closer to the PLE than do the mem- bers of the banksi species group. The male palpi also strongly resemble one another (see Figs. 107-118). Hama- taliwa puta, H. ursa, and H. cavata have very similar epigyna (compare Figs. 65- 67, with Figs. 68, 69 and Figs. 73, 74). These three species may prove to be ge- ographic races after more data on their biology and distribution is collected. On the basis of present materials and infor- mation, however, they appear to be mor- phologically distinct species. In H. puta and H. ursa, the male palpi easily separate the two species (compare Figs. 113-119 with Figs. 111-112). The seminal recep- tacles of H. cavata are considerably more elongate than those in H. puta or H. ursa (compare Fig. 73 with 65, 68). Hamataliwa flebilis and H. laeta have epigyna resembling those of H. banksi, but these may also be associated with II. puta. The male palpus of H. flebilis is similar to that of other males in the puta group ( com- pare Figs. 124, 125 with Figs. 107-119). Because of this similarity and because of the correspondence in eye arrangement, H. flebilis and H. laeta are included in the puta group. Hamataliwa hista has an epigynum readily distinguished from that in all other species of Hamataliwa; however, the male palpus bears a strong resemblance to that in other members of the puta species 80 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 group, (compare Figs. 107-108 with Figs. 109-119). Hamataliwa diffidlis and H. crocata are placed in the puta group primarily because of eye arrangement. Although their epigyna are distinct, they may arbitrarily be con- sidered as similar to those of the puta complex. When the male of H. difficilis is known, it will be easier to place this species. Hamataliwa crocata has a distinct male palpus (Figs. 126, 127) and is the most divergent members of this species group. Grisea group. The epigyna and internal genitalia of H. grisea, H. facilis, and H. schmidti bear a strong resemblance to one another. The eye relationships in this species group are not as uniform as are those in the previous two species groups, but the structure of the cephalothorax and the general arrangement of the eyes, to- gether with the genitalic similarities, sup- port their amalgamation into a separate complex. Discovery of the males in H. facilis and H. schmidti will clarify the situation. Of the remaining six species, H. positiva and H. unca have eye arrangements and epigyna that closely resemble one another, thus indicating kinship, but they could not be linked with other species. Hamataliwa positiva has an epigynum resembling that in certain specimens of //. facilis (compare Fig. 93 with Figs. 97, 98); however, the eye arrangements in the two species are completely different ( compare Fig. 43 with Fig. 44). Hamataliwa circularis and //. subfacilis have eye dispositions remi- niscent of those in members of the puta species group, but they do not agree in proportion. The bodily structure and epigynum of H. circularis ( Figs. 37, 38, 99, 100) and the eye arrangement and epigy- num of //. subfacilis (Figs. 45, 92) make it difficult to relate them to any group. The general body structure, profuse hair, and structure of the epigynum of //. bufo, as well as the absence of the male, exclude it from any of the above groups. Hamataliwa tricuspidata is distinct from all species of Hamataliwa thus far studied. The order of leg length is I-II-IV-III, and coloration and eye arrangement resemble those in certain species of Oxyopes, but the epigynum and palpi, together with the proportions of the legs, are akin to those in Hamataliwa (see discussion under H. tricuspidata) . KEY TO SPECIES OF HAMATALIWA MALES la. ALE row wider than or subequal0 to PME row 2 lb. PME row distinctly wider than ALE row 5 2a. Cymbium of palpus almost as wide as long, ALE row subequal to PME. Color pattern and eye arrangement as in Figs. 11, 12. Palpus as in Figs. 122, 123 globosa 2b. Cymbium of palpus much longer than wide, ALE row slightly wider than PME 3 3a. Distinct color pattern and eye arrange- ment as in Figs. 52, 53. Palpus as in Figs. 128, 129 . tricuspidata 3b. Color pattern not as in 3a. Palpus not resembling those in Figs. 128, 129 _ 4 4a. Palpus illustrated in Figs. 120, 121. Color pattern and eye arrangement as in Figs. 9, 10 -__ triangularis 4b. Palpus illustrated in Figs. 130-133. Color pattern and eye arrangement as in figs. 124, 125 (Brady, 1964) . helia 5a. Palpus with two large tibial apophyses as in Figs. 126, 127. Color pattern and eye arrangement as in Figs. 35, 36 crocata 5b. Palpus with only a single tibial apophysis or a lateral apophysis with a large tooth or spur at its base _ 6 6a. Palpus with a single lateral apophysis without a tooth or spur at its base 7 6b. Palpus with a single lateral apophysis with a conspicuous tooth or spur at its base as in Figs. 107-119 _ 8 7a. Palpus illustrated in Figs. 124, 125. Color pattern and eye arrangement as in Figs. 25, 26 flebilis 7b. Palpus illustrated in Figs. 128, 129. Color pattern and eye arrangement as in figs. 122, 123 (Brady, 1964) . . unca 7c. Palpus illustrated in Figs. 134, 135. * PME row less than .05 mm wider than ALE row. Hamatauwa in Mexico and Central America • Brady Color pattern as in figs. 126, 127 (Brady, 1964) __ grisea 8a. Palpus illustrated in Figs. 107, 108. Color pattern and eye arrangement as in Figs. 27, 28 hista 8b. Palpus illustrated in Figs. 109, 110. Color pattern and eye arrangement as in Figs. 29, 30 cavata 8c. Palpus illustrated in Figs. Ill, 112. Color pattern and eye arrangement as in Figs. 31, 32 ursa 8d. Palpus illustrated in Figs. 113-119. Color pattern and eye arrangement as in Figs. 33, 34 puta KEY TO SPECIES OF HAMATAUWA FEMALES la. ALE row wider than or subequal* to PME row 2 lb. PME row distinctly wider than ALE row 6 2a. Patella-tibia IV slightly longer than patella-tibia III. Distinctive contrasting color pattern as illustrated in Figs. 50, 51. Epigynum and internal genitalia as in Figs. 104-106. tricuspidata 2b. Patella-tibia III slightly longer than patella-tibia IV. Color pattern not as in 2a 3 3a. ALE row subequal to PME row, AME touching a line drawn tangent to lower edge of ALE, and AME less than own diameter from ALE on same side — - 4 3b. ALE row slightly but distinctly wider than PME row, AME well below a line drawn tangent to lower edge of ALE, and AME at least its own diameter from ALE on same side 5 4a. Epigynum and internal genitalia as in Figs. 60-62. Color pattern and eye arrangement as in Figs. 5, 6 barwana 4b. Epigynum and internal genitalia as in Figs. 63, 64. Color pattern and eye ar- rangement as in Figs. 1, 2 cheta 5a. Epigynum as in Fig. 59. Face view as in Fig. 39 brunnea 5b. Epigynum and internal genitalia as in Figs. 54-58. Color pattern and eye ar- rangement as in Figs. 3, 4 banksi 5c. Epigynum and internal genitalia as in Figs. 119, 120. Color pattern and eye arrangement as in figs. 112—114 (Brady, 1964 ) helia 5d. Epigynum and internal genitalia as in Figs. 81-84. Color pattern and eye ar- rangement as in Figs. 7, 8 triangularis PME row less than .05 mm larger than ALE. 6a. Line drawn tangent to lower edge of ALE running above AME row. Epigy- num and internal genitalia as in Fij_rs. 101-103. Color pattern as in Figs. 48, 49 . hufo 6b. Line drawn tangent to lower edge of ALE running below AME, running through AME, or tangent to the upper edge 7 7a. Line drawn tangent to lower edge of ALE running below center of AME 8 7b. Line drawn tangent to lower edge of ALE, tangent to upper edge of AME, or running above center of AME 9 8a. Face view as in Fig. 44. Epigynum and internal genitalia as in Figs. 94-98 __ facilis 8b. Epigynum as in Figs. 89-91. Color pat- tern and eye arrangement as in Figs. 46, 47 schmidti 8c. Epigynum as in Figs. 115, 116. Color pattern as in figs. 108, 109 (Brady, 1964) grisea 9a. AME more than own diameter from ALE 10 9b. AME own diameter or less from ALE ... 11 10a. Face view as in Fig. 43. Epigynum as in Fig. 93 positiva 10b. Epigynum and internal genitalia as in Figs. 77, 78. Color pattern and eye ar- rangement as in Figs. 23, 24 crocata 10c. Epigynum and internal genitalia as in Figs. 73, 74. Color pattern and eye ar- rangement as in Figs. 17, 18 cavata 11a. Posterior sclerotized rim of epigynum more or less U-shaped 15 lib. Posterior sclerotized rim of epigynum V-shaped, scalloped, or straight 12 12a. Posterior rim of epigynum scalloped or V-shaped 13 12b. Posterior rim of epigynum straight, median depression rectangular 14 13a. Posterior rim of epigynum scalloped as in Fig. 92. Face view as in Fig. 45 _... subfacilis 13b. Posterior rim of epigynum V-shaped as in Figs. 70-72. Color pattern and eye arrangement as in Figs. 13, 14. flebilis 14a. Epigynum and internal genitalia as in Figs. 85-88. Face view as in Figs. 40, 41 difficilis 14b. Epigynum and internal genitalia as in Figs. 117, 118. Color pattern and eye arrangement as in figs. 110, 111 (Brady, 1964 ) unca 15a. Posterior rim of epigynum broadly U- shaped, almost circular; seminal recep- tacles widely separated as in Figs. 99- 100. Color pattern and eye arrange- ment as in Figs. 37, 38 circularis 82 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 MAP 1 15b. Posterior rim U-shaped, seminal recep- tacles close together as in Figs. 65-72 .... 16 16a. Epigynum and internal genitalia as in Figs. 65-67. Color pattern and eye ar- rangement as in Figs. 21, 22. puta 16b. Epigynum and internal genitalia as in Figs. 68, 69. Color pattern and eye ar- rangement as in Figs. 19, 20 . ursa 16c. Epigynum and internal genitalia as in Figs. 73, 74. Color pattern and eye ar- rangement as in Figs. 17, 18 ... . cavata 16d. Epigynum and internal genitalia as in Figs. 75, 76. Color pattern and eye arrangement as in Figs. 15, 16 hista SPECIES DESCRIPTIONS Hamataliwa banksi (Mello-Leitao) Figures 3, 4, 54—58. Map 1. Oxyopes brevis Banks, 1898, Proc. California Acad. Sci., 1(7):278, pi. 17, fig. 26, $. Female lectotype, here designated, from Cerro del Taste, Territorio Sur, Baja California, in Museum of Comparative Zoology, examined. Name preoccupied, not Oxyopes brevis Thorell, 1881. Oxyopes annulipes F.O.P-Cambridge, 1902, Bio- logia Centrali-Americana, Arachnida, Araneidea 2:345, pi. 32, fig. 27, 9. Female holotype from Amula, 9.5 km NW of Chilapa, Guerrero, Mex- ico, in the British Museum (Natural History), examined. Name preoccupied, not Oxyopes an- nulipes Thorell, 1892. NEW SYNONYMY. Oxyopes banksi Mello-Leitao, 1928, Bol. Mus. Bio de Janeiro, 4(3):50. New name for Oxyopes brevis Banks. Oxyopes cambridgei Mello-Leitao, 1928, Bol. Mus. Bio de Janeiro 4(3):50. New name for Oxyopes annulipes F.O. P. -Cambridge. NEW SYNONYMY. Discussion. Coincidentally, the names Oxyopes brevis and Oxyopes annulipes, applied to this species by Banks and F.O.P.- Cambridge respectively, were both pre- occupied. Mello-Leitao (1928) noticed this and gave new names to the species. The two names are considered synonymous in this paper because only one species is involved ( compare Figs. 54-57 with Fig. 58). Measurements. Length of eight females 4.1-5.1 mm, mean 4.84 mm; carapace width 1.4-2.0 mm, mean 1.76 mm; carapace length 1.7-2.5 mm, mean 2.16 mm. Width of eye rows: AME .25-30 mm, mean .272 mm; ALE .50-.68 mm, mean .631 mm; PLE .88-1.13 mm, mean 1.025 mm; PME .45-.63 mm, mean .547 mm. Ham iTALiwA in Mexico and Central America • Brady 83 Segments of leg I ( five females ) : femur 1.9-2.4 mm, mean 2. IS mm; patella-tibia 2.0-2.7 mm, mean 2.43 mm; metatarsus 1.4- 1.6 mm, mean 1.51 mm; tarsus .8-9 mm, mean .82 mm; total length I 5.9-7.5 mm, mean 6.94 mm. Length of patella-tibiae: II 1.9-2.5 mm, mean 2.23 mm; III 1.4-1.9 mm, mean 1.72 mm; IV 1.3-1.9 mm, mean 1.59 mm. Color. Female. Pattern illustrated in Figures 3 and 4. Face pale yellow to light orange, chelicerae with slightly darker orange tint. Lighter, inversely T-shaped mark from AME row to lower edge of clypeus. Flattened white hairs, heaviest in eye region and along sides of face. Carapace pale orange to orange, with scattered spatulate-shaped white hairs, most abundant along sides and at posterior declivity. Dorsum of abdomen cream. Venter pale yellow to cream without distinct median stripe. Legs pale yellow to light orange, some- what darker distally. Labium, endites, and sternum pale yel- low to light orange. Diagnosis. Hamataliwa banksi is very close to H. helia in body dimensions, eye arrangement, and the shape of the epigy- num. These two species apparently overlap in distribution. Hamatalhca banksi is larger than H. helia, and the females can be distinguished by epigynal structure. In H. helia the posterior rim of the epigynum is not as heavily sclerotized, and the central de- pression of the epigynum is larger and more oval than it is in //. banksi (compare figs. 119-120 of Brady, 1964, with Figs. 54- 58 of this paper). Separation of these two species may be- come impossible after larger series of speci- mens are collected. Until males of H. banksi are found and are compared with H. helia males, it seems best to maintain them as separate species. Distribution. Mexico and Central America (Map 1). Records. MEXICO. Baja California. Territorio Sur, Cerro del Taste, 9 9 . Guer- rero. Amula, 9.5 km NW of Chilapa, (H. II. Smith). GUATEMALA. Coban, July 1947, 9 (C, P. Vaurie). COSTA RICA. San Jose, 9 9 (E. Schmidt). Hamataliwa helia (Chamberlin) Oxyopes helius Chamberlin, 1929, Ent. News, 40:19, fig. 4, 9. Female holotype from Mix- son's Hammock, Okefenokee Swamp, Georgia, in the American Museum of Natural History, examined. Hamataliwa helia is closely related to H. banksi and may be synonymous with that species (see discussion under II. banksi). For illustrations of the color patterns and genitalia and locality records, refer to Brady (1964, p. 497). Distribution. Florida to Texas and south to Yucatan. Hamataliwa brunnea (F.O. P. -Cambridge) Figures 39, 59. Map 1. Oxyopes brunneus F.O. P. -Cambridge, 1902, Bio- logia Centrali-Americana, Arachnida, Araneidea, 2:346, pi. 32, fig. 29, 9. Female holotype from Atoyac, Veracruz, Mexico, in the British Museum (Natural History), examined. Discussion. This species is represented by the unique female above. Specimens designated as Oxyopes brunneus F.O. P.- Cambridge in other collections did not agree specifically with this one. The holo- type was in such poor condition that the original color description is used below and only partial measurements were possible. Drawings of the epigynum and face were made. Measurements. Length of female holo- type 6.2 mm, carapace width 2.0 mm, carapace length 2.5 mm. Width of eye rows: AME .28 mm, ALE .69 mm, PLE 1.22 mm, PME .62 mm. Segments of leg I: femur 2.7 mm, patella- tibia 3.2 mm, metatarsus 1.9 mm, tarsus 1.0 mm, total length 8.8 mm. 84 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Length of patella-tibiae: II 3.0 mm, III 2.0 mm, IV not present. Color. Following is the original de- scription of the holotype by F.O.P.-Cam- bridge: "The scales have been almost en- tirely rubbed off from the single specimen received of this species, and with these the colour and pattern have vanished; but the form of the vulva is quite distinct from that of any other Oxyopes in the collection be- fore me. The general ground-colour is deep brown, whereas that of all the other members of the genus here described is yellow or orange." Diagnosis. The structure of the epigy- num in H. brunnea is similar to that of H. crocota (compare Fig. 59 with Fig. 78), but the eye arrangement in these two spe- cies is quite different (compare measure- ments). Body size and eye arrangement of H. brunnea ally it with H. banksi. Record. MEXICO. Veracruz. Atoyac, 9 (H. H. Smith). Hamataliwa triangularis (Kraus) Figures 7-10, 81-84. Map 1. Oxyopes globosus F.O. P. -Cambridge, 1902, Bio- logia Centrali-Americana, Arachnida, Araneidea, 2:343 (in part), pi. 32, figs. 19, 19a $ only. Female allotype, from "Bugaba, Panama, in the British Museum (Natural History), examined. Not Oxyopes globosus F.O.P.-Cambridge $ holotype. Oxyopeidon triangularis Kraus, 1955, Abh. Sen- ekenb. Naturf. Ges., 493:38, pi. 5, fig. 97. Female holotype from San Salvador, El Sal- vador, in Senekenberg Museum, examined. Discussion. The female described by F.O.P.-Cambridge as Oxyopes globosus docs not agree in size or eye arrangement with the male holotype. In all species of Hamataliwa investigated the males are smaller than the females. The male holo- type of O. globosus is larger than 20 fe- males of O. globosus F.O.P.-Cambridge that were measured. In addition the ALE row of the male holotype is not wider than the PME row as in the female. Therefore, //. triangularis (Kraus) becomes the valid 0 Chiriqui, 22 km NW of David. name for the female described as Oxyopes globosus F.O.P.-Cambridge. Measurements. Length of two males 3.8, 4.2 mm, carapace width 1.5, 1.6 mm, carapace length 1.8, 1.9 mm. Width of eye rows: AME .25, .27 mm, ALE .53, .55 mm, PLE .87, .93 mm, PME .50, .53 mm. Segments of leg I: femur 2.0, 2.1 mm, patella-tibia 2.5, 2.8 mm, metatarsus 1.7, 1.9 mm, tarsus .9. 1.0 mm, total length 7.1, 7.8 mm. Length of patella-tibiae: II 2.4, 2.5 mm, III 1.8, 1.9 mm, IV --*, 1.6 mm. Length of 10 females 4.2-5.3 mm, mean 4.90 mm; carapace width 1.6-1.9 mm, mean 1.78 mm; carapace length 2.0-2.3 mm, mean 2.4 mm. Width of eye rows: AME .27-.30 mm, mean .281 mm; ALE .57-.63 mm, mean .606 mm; PLE .97-1.03 mm, mean 1.005 mm; PME .53-.60 mm, mean .569 mm. Segments of leg I: femur 2.2-2.5 mm, mean 2.40 mm; patella-tibia 2.7-3.0 mm, mean 2.85 mm; metatarsus 1.8-2.0 mm, mean 1.92 mm; tarsus .8-1.0 mm, mean .91; total length 7.7-8.4 mm, mean 8.08 mm. Length of patella-tibiae: II 2.5-2.8 mm, mean 2.68 mm; III 1.9-2.3 mm, mean 2.07; IV 1.6-1.9 mm, mean 1.78 mm. Color. Male. Pattern illustrated in Figures 9 and 10. Face and chelicerae yellow- orange. Distal ends of chelicerae lighter, yellowish. Cymbia of palpi brown. Carapace yellow-orange to orange. Dorsum of abdomen cream colored. Sides darker, brownish. Venter of abdo- men cream colored, slightly darker medi- ally. Legs yellow-orange without dusky mark- ings. Labium and endites ivory to pale cream. Sternum ivory. Color. Female. Pattern illustrated in Figures 7 and 8. Face and chelicerae yel- low to yellow-orange, with relatively thick clothing of white appressed hairs, thickest at lateral and ventral margins of face. * Two dashes indicate a missing leg segment. Hamataliwa in Mexico and Central America • Br 85 Carapace- pale- yellow to yellow-orange. Vertical sides with white or mixture of white and dark brown spatulate hairs. Dorsum of abdomen cream colored, without darker markings or, in a few specimens, an irregular spotted pattern formed from an intermixture of white and dark brown hairs. Large dark brown spots along sides of abdomen about one-third of the distance from the spinnerets to the base in these hirsute specimens. Venter of abdomen cream colored. Legs pale yellow to cream with dusky brown bands at distal ends of femora and tibiae, tibiae dusky at proximal ends as well. Dusky bands formed by spatulate- shaped hair. Labium and endites pale yellow to yel- low-orange. Sternum cream to pale yellow. Diagnosis. Hamataliwa triangularis is similar to II. banksi and II. brunnca in eye arrangement. The palpus of the male re- sembles that of H. helia (compare Figs. 120-121 of this paper with figs. 130-133 of Brady, 1964). Ha?nataJhca triangularis can be readily differentiated from other mem- bers of this group of species by the struc- ture of the epigynum (Figs. 81-84). Distribution. El Salvador to Panama (Map 1). Records. EL SALVADOR. San Salvador, 21 June 1951, 39 9 (A. Zilch). PANAMA. Canal Zone. Barro Colorado Island, numerous $ i 9 9 (A. M. Chicke- ring); Madden Dam, 8 Aug. 1939, 9 (A. M. Chickering), 28 May 1956, 9 (W. E. Lundy ) . Hamataliwa barroana (Chamberlin and Ivie) Figures 5, 6, 60-62. Map 1. Oxijopcs barroanus Chamberlin and Ivie, 1936, Bull. Univ. Utah, Biol. Ser., 3(5):18, pi. 4, fig. 27, 9 • Female holotype from Barro Colo- rado Island, Panama Canal Zone, in the Ameri- can Museum of Natural History, examined. Measurements. Length of 10 females 4.7-5.7 mm, mean 5.16 mm; carapace width 1.7-1.9 mm, mean 1.81 mm; carapace length 2.1-2.3 mm, mean 2.19 mm. Width of eye rows: AME .25-.2S mm, mean .274 mm; ALE .68-.73 mm, mean .70S mm; PLE 1.17-1.25 mm, mean 1.211 mm; PME .68-77 mm, mean .730 mm. Segments of leg 1: femur 2.0-2.4 mm, mean 2.17 mm; patella-tibia 2.5-3.0 mm, mean 2.72 mm; metatarsus 1.5-1.9 mm, mean 1.74 mm; tarsus .8-.9 mm, mean .82 mm; total length I 6.8-8.0 mm, mean 7.44 mm. Length of patella-tibiae: II 2.4-2.6 mm, mean 2.48 mm; III 1.8-2.0 mm, mean 1.88 mm; IV 1.7-1.9 mm, mean 1.81 mm. Color. Female. Pattern illustrated in Figures 5 and 6. Face yellow-orange with broad light brown vertical stripes from ALE to lower edge of clypeus. Chelicerae yellow-orange, overlaid with brown. Hexag- onal area bounded by eyes, reddish. In- terior distal ends lighter in color. Carapace yellow-orange. Dorsum of abdomen cream colored. Cardiac area translucent gray. Lateral areas with scat- tered spots of reddish hair. Venter of abdomen cream colored with broad light brown stripe from epigastric furrow to base of spinnerets. Legs pale yellow to cream. Labium pale orange-yellow. Endites pale orange-yellow, distal ends tipped with cream. Sternum cream. Diagnosis. Hamataliwa barroana is readily distinguished from other species of Hamataliwa by the shape of the epigynum. (Figs. 60-62). In both II. barroana and 77. globosa, the ALE row is subequal to the PME row. They may be related to the H. banksi group in which the ALE row is wider than the PME rows. In all other species of Hamataliwa, the PME row is much wider than the ALE row, with the exception of H. tricuspidata. Distribution. Mexico and Central America. Records. MEXICO. Veracruz. La Buena Ventura, July 1909, 9. Yucatan, Colonia, 13-19 Aug. 1952, 9 (J., D. Pallister). PANAMA. Canal Zone. Barro Colorado Island, numerous 9 9, various collectors; 86 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Gamboa, 11 Aug. 1939, 49 9; Madden Dam, 18 Aug. 1936, 49 9 (A. M. Chicke- ring). Hamataliwa globoso (F.O. P. -Cambridge) Figures 122, 123. Map 1. Oxyopes globosus F.O. P. -Cambridge, 1902, Bio- logia Centrali-Americana, Arachnida, Araneidea, 2:343 (in part), pi. 32, figs. 18, 18a-c, $ only. Male holotype from Bugaba, Panama, in British Museum (Natural History), examined. Discussion. The female described by F.O.P.-Cambridge as Oxyopes g,lobosus be- longs to the species described by Kraus (1955) as Oxyopes triangularis. For further comments see the discussion under Hama- taliwa triangularis. Measurements. Length of five males 4.7-5.5 mm, mean 4.92 mm; carapace width 1.8-1.9 mm, mean 1.85 mm; carapace length 2.1-2.4 mm, mean 2.18 mm. Width of eye rows: AME .25-28 mm, mean .264 mm; ALE .63-72 mm, mean .66S mm; PLE 1.10-1.25 mm, mean 1.149 mm; PME .65-.75 mm, mean .685 mm. Segments of leg I: femur 2.1-2.4 mm, mean 2.19 mm; patella-tibia 2.7-3.0 mm, mean 2.84 mm; metatarsus 1.9-2.0 mm, mean 1.91 mm; tarsus .9-1.0 mm, mean .92 mm; total length I 7.6-8.3 mm, mean 7.86 mm. Length of patella-tibiae: II 2.4-2.7 mm, mean 2.57 mm; III 1.9-2.2 mm, mean 1.97 mm; IV 1.7-2.0 mm, mean 1.80 mm. Color. Male. Pattern illustrated in Figures 11 and 12. Face yellow with broad vertical stripes of brown from ALE row to lower edge of clypeus. Chelicerae darker, brownish. Cymbia of pedipalpi dark brown. Carapace light orange-yellow without darker markings or sometimes dusky along sides. Dorsum of abdomen white to cream colored with brownish margins and sides. Often with a few scattered darker spatulate hairs about mid-point of abdomen. Venter of abdomen white to cream colored with or without median dusky band. Legs yellow without darker markings. Labium and endites yellow with distal ends lighter, ivory. Sternum ivory. Diagnosis. Hamataliwa globosa is dis- tinct from all other species of Hamataliwa on the basis of palpal structure ( Figs. 122, 123). The fact that the cymbium of the palpus is almost as wide as it is long makes for easy recognition. The eye arrangement is nearest to that of H. triangularis and H. barroana. Distribution. Mexico and Central America (Map 1). Records. MEXICO. San Luis Potosi. Tamazunchale, 20 May 1952, 6 (M. Cazier, W. Gertsch, R. Schrammel). PANAMA. Canal Zone. Barro Colorado Island, 7-8 May 1946, 3i6(T. C. Sch- neirla); Bugaba (Chiriqui, 22 km NW of David), i (G. C. Champion). Hamataliwa cheta sp. n. Figures 1, 2, 63, 64. Map 1. Holotype. Female from Coban, Guate- mala, July 1947 (C, P. Vaurie), in the American Museum of Natural History. The specific name is an arbitrary combination of letters. Measurements. Length of three females 6.4, 6.9, 7.5 mm; carapace width 2.3, 2.3, 2.4 mm; carapace length 2.8, 2.8, 2.9 mm. Width of eye rows: AME .28, .28, .30 mm; ALE .69, .72, .72 mm; PLE 1.20, 1.22, 1.27 mm; PME .70, .74. .75 mm. Segments of leg I: femur 2.5, 2.7, 2.7 mm; patella-tibia 3.3, 3.5, 3.5 mm; meta- tarsus 2.0, 2.0, 2.1 mm; tarsus 1.0, 1.0, 1.0 mm; total length I 8.8, 9.2, 9.3 mm. Patella-tibiae: II 2.9, 3.1, 3.1 mm; III 2.4, 2.5, - - mm; IV 2.3, 2.3, 2.4 mm. Color. Female. Pattern illustrated in Figures 1 and 2. Face yellow-orange to orange-brown, lower edge of clypeus yel- lowish. Chelicerae yellow-orange to orange- brown with tuft of white hairs along inner margins of basal segments. Carapace yellow-orange to orange-brown with darker brown hairs along vertical sides. Hamataliwa in Mexico and Central America • Bradu 87 MAP 2 Dorsum of abdomen white to cream with cardiac area translucent white. Lateral areas darker brown. Venter of abdomen cream to pale yellow with broad median dusky stripe from epigastric furrow to base of spinnerets, margined by thin pale yel- low stripe laterally. Legs cream to pale yellow with metatarsi and tarsi darker, brownish. Labium and endites pale yellow-orange, distal ends ivory. Sternum yellow. Diagnosis. In H. cheta the PME row is subequal to the ALE row in width and the AME are well below the ALE ( see Fig. 1 ) as in the banksi species group. The shape of the epigynum also conforms generally to that of H. barroana. For these reasons H. cheta is considered a member of the banksi species group. This species is easily dis- tinguished from H. barroana and other species of the banksi group by its larger size and the structure of the epigynum (Figs. 63, 64). Distribution. Guatemala (Map 1). Records. GUATEMALA. Coban, July 1947,39 9 (C., P. Vaurie). Hamataliwa puta (O.P.-Cambridge) Figures 21, 22, 33, 34, 65-67, 113-119. Map 2. Oxyopeidon putum O.P.-Cambridge, 1894, Bio- logia Centrali-Americana, Arachnida, Araneidea, 1:140, pi. 16, figs. 7, 7a-e, 8, 8a-c, & $. Male lectotype, here designated, from Bugaba, Panama, in British Museum (Natural History) examined. F.O.P.-Cambridge, 1902, op. cit., 2:347. Measurements. Length of six males 4.7-5.3 mm, mean 5.04 mm; carapace width 1.8-2.2 mm, mean 1.93 mm; carapace length 2.1-2.5 mm, mean 2.23 mm. Width of eye rows: AME .25-.27 mm, mean .256 mm; ALE .70-.75 mm, mean .737 mm; PLE 1.30-1.42 mm, mean 1.350 mm; PME .92-.99 mm, mean .960 mm. Segments of leg I: femur 2.4-2.8 mm (5), patella-tibia 3.3-3.5 mm (3), meta- tarsus 2.1-2.5 mm (3), tarsus 1.0-1.2 mm (2), total length I 8.9-9.9 mm (2). Length of patella-tibiae: II 2.6-3.4 mm (3), III 1.9-2.6 mm (3), IV 1.9-2.3 mm (3). In most cases a leg or leg segment was missing, so for these particular measure- ments only the range is given, with the number of specimens in parentheses. Length of ten females 4.7-6.1 mm, mean 5.36 mm; carapace width 1.6-2.0 mm, mean 1.79 mm; carapace length 2.0- 2.4 mm, mean 2.18 mm. Width of eye rows: AME .22-.28 mm, mean .254 mm; ALE .70-80 mm, mean .746 mm; PLE 1.30-1.50 mm, mean 1.363 mm; PME .95-1.07 mm, mean .985 mm. 88 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Segments of leg I: femur 2.0-2.4 mm, mean 2.18 mm; patella-tibia 2.5-2.9 mm, mean 2.68 mm; metatarsus 1.5-1.8 mm, mean 1.64 mm; tarsus .7-.9 mm, mean .80 mm; total length I 6.7-7.8 mm, mean 7.29 mm. Length of patella-tibiae: II 2.2-2.8 mm, mean 2.45 mm; III 1.6-2.0 mm, mean 1.79 mm; IV 1.5-1.9 mm, mean 1.69 mm. Color. Male. Pattern illustrated in Figures 33 and 34. In one male in good condition, the face and chelicerae are yellow-orange and clothed with white spatulate hairs. The sides of the face above the eheliceral artic- ulation are darker brown. In rubbed speci- mens the face and chelicerae are orange- brown, with the lower edge of the clypeus and distal ends of chelicerae lighter. Palpi dark brown. Carapace orange to orange-brown with darker brown hairs along sides. Dorsum of abdomen pale yellow to cream. Venter pale yellow to cream with- out darker median stripe. Legs pale yellow to yellow-orange. Sternum cream to yellow-orange or amber. Labium and endites pale yellow to orange, cream distally. Female. Pattern illustrated in Figures 21 and 22. Face and chelicerae orange-brown, usually with lighter cream color along lower edge of clypeus and sometimes at distal ends of chelicerae. Carapace orange-brown. Dorsum of abdomen cream colored to tan with scat- tered patches of brown spatulate hairs. Venter cream to pale yellow. No darker markings. Legs yellow to yellow-orange. Labium and endites yellow to light orange, distal ends paler. Sternum cream to yellow. Diagnosis. Hamataliwa puta is closely related to //. ursa. The structure of the epigyna and the female genitalia is very similar (compare Figs. 65-67 with Figs. 68-69), but the tibial apophysis of the palpus in //. ursa is much larger than it is in II. puta (compare Figs. 111-112 with Figs 113-119). For further discussion see the diagnosis of H. ursa. Distribution. Mexico and Central America (Map 2). Records. MEXICO. Veracruz. Vera- cruz, $ (N. Banks). COSTARICA. Antonios, 6 (N.Banks). PANAMA. Bugaba (Chiriqui, 22 km NW of David), $ S :109 9 (C. G. Cham- pion). Canal Zone. Barro Colorado Island, 31 July 1954, & (A. M. Chickering). Hamataliwa ursa sp. n. Figures 1 9, 20, 31 , 32, 68, 69, 1 1 1 , 1 1 2. Map 2. Holotypc. Male from Barro Colorado Island, Panama Canal Zone. June 1950 (A. M. Chickering), in the Museum of Com- parative Zoology. The specific name is a noun in apposition meaning bear. Measurements. Length of nine males 4.4-5.0 mm, mean 4.74 mm; carapace width 1.7-2.0 mm, mean 2.19 mm; car- apace length 2.0-2.4 mm, mean 2.19 mm. Width of eye rows: AME .22-.25 mm, mean .239 mm; ALE .65-.75 mm, mean .705 mm; PLE 1.19-1.39 mm, mean 1.306 mm; PME .S7-.97 mm, mean .926 mm. Segments of leg I: femur 2.1-2.6 mm, mean 2.39 mm; patella-tibia 2.7-3.3 mm, mean 2.99 mm; metatarsus 1.8-2.2 mm, mean 2.01 mm; tarsus .9-1.0 mm, mean .96 mm; total length 7.5-8.9 mm, mean 8.36 mm. Length of patella-tibiae: II 2.4-3.0 mm, mean 2.73 mm; III 1.9-2.2 mm, mean 2.06 mm; IV 1.8-2.0 mm, mean 1.87 mm. Length of ten females 5.3-6.7 mm, mean 6.04 mm; carapace width 1.9-2.2 mm, mean 2.02 mm; carapace length 2.3-2.6 mm, mean 2.43 mm Width of eye rows: AME .27-30 mm, mean .277 mm; ALE .78-85 mm, mean .815 mm; PLE 1.45-1.64 mm, mean 1.536 mm; PME 1.02-1.19 mm, mean 1.102 mm. Segments of leg I: femur 2.4-2.7 mm, mean 2.48 mm; patella-tibia 2.9-3.5 mm, Hamataliwa in Mexico and Central America • Brady 89 mean 3.13 mm; metatarsus 1.8-2.0 mm, mean 1.84 mm; tarsus .8-1.0 mm, mean .88 mm; total length I 7.9-9.2 mm, mean 8.33 mm. Length of patella-tibiae: II 2.7-3.0 mm, mean 2.82 mm; III 1.9-2.2 mm, mean 2.02 mm; IV 1.8-2.1 mm, mean 1.91 mm. Color. Male. Pattern illustrated in Figures 31 and 32. Coloration very similar to that of the female. In the male illustrated there are more brown spatulate hairs along the sides of the abdomen than in the female. Cymbia of palpi brown. Color. Female. Pattern illustrated in Figures 19 and 20. Face and chelicerae yellow-orange to brownish orange, overlaid with darker hairs. White spatulate hairs at sides of face, thinning anteriorly. Lower edge of clypeus with yellowish tinge. Carapace yellow-orange overlaid with brown spatulate hairs imparting a brownish orange tint. Darker along vertical sides. Dorsum of abdomen cream to yellow with scattered patches of brown hair. Brown indentations about one-third of the distance from spinnerets to base of abdo- men. Cardiac region and sometimes muscle depressions well marked. Venter cream to yellow with only a faint broad dusky stripe from epigastric furrow to base of spinner- ets. Legs yellow with darker brown spatulate hairs. Labium yellow to light brown. Endites cream to yellow. Sternum ivory to cream. Diagnosis. Hamataliwa ursa is very similar to H. pitta in body size, leg length, and eye arrangement (compare measure- ments of these components). However, the epigynum of H. ursa is more broadly rounded than that of //. pitta (compare Fig. 69 with Fig. 66) and the tibial apophysis of the male palpus in H. ursa is considerably larger than that in H. pitta (compare Figs. 111-112 with Figs. 113- 119). The internal genitalia of the females are very much alike (compare Fig. 68 with Fig. 65); this similarity may be conside as good evidence for conspecificity. Males of //. ursa, however, are easily distin- guished from H. puta males by the tibial apophysis. Because of the differences in the males of the two groups, they are considered as separate species here. Further collections and field studies should elucidate the relationships of the populations concerned. Distribution. Panama (Map 2). Records. PANAMA. Canal Zone. Barro Colorado Island, numerous 6 6 9 9 (A. M. Chickering); Madden Dam, 18 Aug. 1936, 9, 25-31 July 1950, 9 (A. M. Chickering); Summit, 7-10 July 1950, 4 9 9, 21-29 July 1950, 59 9, 16-17 Aug. 1950, 79 9:oo, 2.3-28 Aug. 1950, 49 9 (A. M. Chickering). Hamataliwa cavata (Kraus) Figures 17, 18, 29, 30, 73, 74, 109, 110. Map 2. Oxyopeidon cavatum Kraus, 1955, Ahb. Senckenb. Naturf. Ges., no. 493, p. 39, figs 99-101, CM oo lO ro to 106 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Figs. 13-14. Hamataliwa flebiin (O. P. -Cambridge), female from Gamboa, Panama, 24 July 1954. 13. Face view. 1 4. Dorsal view. Figs. 15-16. Hamataliwa hista sp. n., female from Boquete, Panama, 4-11 Aug. 1954. 15. Face view. 16. Dorsal view. Figs. 17-18. Hamataliwa cavata (Kraus), female from San Salvador, El Salvador, 30 Apr. 1951. 17. Face view. 18. Dorsal view. Figs. 19-20. Hamataliwa una sp. n., female allotype from Barro Colorado Island, Panama Canal Zone, 1-4 July 1950. 19. Face view. 20. Dorsal view. Figs. 21-22. Hamataliwa puta (O. P. -Cambridge), female allotype from Bugaba, Panama. 21. Face view. 22. Dorsal view. Figs. 23-24. Hamataliwa zrocata sp. n., female from Barro Colorado Island, Panama Canal Zone, 21-29 July 1950. 23. Face view. 24. Dorsal view. Lower scale is for dorsal views. Upper scale is for face views. Hamataliwa in Mexico and Central America • Brady c\j ro ro ro E E 00 yr .: £ ' ■■•- ■.:......-■ - 112 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Figs. 46-47. Hamataliwa schmidti Reimoser, female syntype from San Jose, Costa Rica. 46. Face view. 47. Dorsal view. Figs. 48-49. Hamataliwa bufo sp. n., female holotype from Barro Colorado Island, Panama Canal Zone, 1-4 July 1950. 48. Face view. 49. Dorsal view. Figs. 50-53. Hamataliwa tricuspidata (F.O. P. -Cambridge), from Arraijan, Panama, 6-9 July 1950. 50. Face view of female. 51. Dorsal view of female. 52. Face view of male. 53. Dorsal view of male. Hamataliwa in Mexico and Central America • Brady 46 48 50 52 Imm 53 * w 114 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Figs. 54-57. Hamataliwa banksi (Mello-Leitao), females from Cerro del Taste, Territorio Sur, Baja California. 54. Epigynum of paralectotype. 55. Internal genitalia of paralectotype. 56-57. Epigynum of lectotype drawn from different angles. Fig. 58. Oxyopes annulipes F.O. P. -Cambridge = Hamafa//wa banksi (Mello-Leitao), epigynum of holotype from Amula, Guerrero. Fig. 59. Hamataliwa brunnea (F.O. P. -Cambridge), epigynum of female holotype from Atoyac, Veracruz. Fig. 60. Hamataliwa barroana (Chamberlin), internal genitalia of female from Barro Colorado Island, Panama Canal Zone, 4 July 1954. Fig. 61. Hamataliwa barroana (Chamberlin and Ivie), epigynum of female holotype from Barro Colorado Island, Panama Canal Zone, Aug. 1928. Fig. 62. Hamataliwa barroana (Chamberlin), epigynum of female from Barro Colorado, Panama Canal Zone, 4 July 1954. Figs. 63-64. Hamataliwa cheta sp. n., female from Coban, Guatemala, July 1947. 63. Internal genitalia. 64. Epigynum. Scale is for all figures of epigyna. Hamataliwa in Mexico and Central America • Brady 54 55 ■ 58 60 61 62 -" - 0.1 mm 116 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Figs. 65-67. Homataliwa puta (O. P. -Cambridge), female paratypes from Bugaba, Panama. 65. Internal genitalia. 66. Epi- gynum. 67. Epigynum of "allotype." Figs. 68-69. Homataliwa ursa sp. n., female from Barro Colorado Island, Panama Canal Zone, 1-4 July 1950. 68. Internal genitalia. 69. Epigynum. Figs. 70-72. Homataliwa flebilis (O. P. -Cambridge), female paratypes from Bugaba, Panama. 70. Epigynum of "allotype." 71. Epigynum. 72. Internal genitalia. Figs. 73-74. Homataliwa cavata (Kraus), female allotype from San Salvador, El Salvador, 30 Apr. 1951, epigynum mounted in glycerine gel. 73. Dorsal view. 74. Ventral view. Figs. 75-76. Homataliwa h/'sfa sp. n., female from Boquete, Panama, 4-11 Aug. 1954. 75. Internal genitalia. 76. Epigynum. Hamataliwa in Mexico and Central America • Brady ■ '.■:..' 68 70 ... ■-■. ■ 73 74 71 72 118 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Figs. 77-78. Hamataliwa crocata sp. n., female from Summit, Panama Canal Zone, 21-29 July 1950. 77. Internal genitalia. 78. Epigynum. Figs. 79-80. Hamataliwa laeta (O. P. -Cambridge), female holotype from Dos Caminos, Guerrero. 79. Epigynum, drawn Mar. 1968. 80. Epigynum, drawn June 1963. Figs. 81-82. Hamataliwa triangularis (Kraus), female from Barro Colorado Island, Panama Canal Zone, 23-30 June 1939. 81. Internal genitalia. 82. Epigynum. Fig. 83. Hamataliwa triangularis (Kraus), epigynum of female holotype from San Salvador, El Salvador, 21 June 1961, mounted in glycerine gel. Fig. 84. Oxyopes globosus F.O. P. -Cambridge — Hamataliwa triangularis (Kraus), epigynum of female paratype from Bugaba, Panama. Figs. 85-88. Hamataliwa difficilis (O. P. -Cambridge), females from Amula, Guerrero. 85. Epigynum of lectotype. 86. Internal genitalia of paralectotype. 87. Epigynum of paralectotype. 88. Epigynum of holotype of Oxyopeidon molestum O. P. -Cam- bridge ~ Hamataliwa difficilis. Hamataliwa in Mexico and Central America ■ Brady ■. 77 78 ,""'. % ^ . ■.-•; •'- ■■ 83 tf0&$$§!t%l#;- " £& 84 86 87 88 120 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Fig. 89. Hamataliwa schmidti Reimoser, epigynum of female from Guatemala. Figs. 90-91. Hamataliwa schmidti Reimoser, female syntype from San Jose, Costa Rica. 90. Internal genitalia. 91. Epi- gynum. Fig. 92. Hamataliwa subfacilis (O. P. -Cambridge), epigynum of female holotype from Amula, Guerrero. Fig. 93. Hamataliwa positiva Chamberlin, epigynum of female holotype from San Carlos Bay, Sonora, 7 July 1921. Figs. 94-95. Hamataliwa facilis (O. P. -Cambridge), female holotype from Chilpancingo, Guerrero. 94. Epigynum drawn April 1968. 95. Epigynum drawn June 1963. Figs. 96-98. Hamataliwa facilis (O. P. -Cambridge), female paratypes from Chilpancingo, Guerrero. 96. Internal genitalia. 97. Epigynum. 98. Epigynum. Hamataliwa in Mexico and Central America • Brady i 122 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Figs. 99-100. Hamataliwa circularis (Kraus), female holotype from 6 km N of Los Blancos, El Salvador, genitalia mounted in glycerine gel. 99. Ventral view. 100. Dorsal view. Figs. 101-102. Hamataliwa bulo sp. n., female from Barro Colorado Island, Panama Canal Zone, 3-5 July 1936. 101. Inter- nal genitalia. 102. Epigynum. Fig. 103. Hamataliwa bufo sp. n., epigynum of female holotype from Barro Colorado Island, Panama Canal Zone, 1-4 July 1950. Fig. 104. Hamataliwa tricuspidata (F.O. P. -Cambridge), internal genitalia of female from Arraijan, Panama, 6-9 July 1950. Fig. 105. Oxyopes clypeatus F.O. P. -Cambridge ~ Hamataliwa tricuspidatus, epigynum of female holotype from Bugaba, Panama. Fig. 106. Hamataliwa tricuspidatus (F.O. P. -Cambridge), epigynum of female from Arraijan, Panama, 6-9 July 1950. Hamataliwa in Mexico and Central America • Brady 100 102 105 103 104 10 6 124 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Figs. 107-108. Hamataliwa hista sp. n., male holotype from Boquete, Panama, 4-11 Aug. 1954. 107. Left palpus, ventral view. 108. Left palpus, retrolateral view. Figs. 109-110. Hamataliwa cavata (Kraus), male holotype from San Salvador, El Salvador, Nov. 1951. 109. Palpus, ventral view. 110. Palpus, retrolateral view. Figs. 111-112. Hamataliwa una sp. n., male holotype from Barro Colorado Island, Panama Canal Zone, June 1950. 111. Palpus, ventral view. 112. Palpus, retrolateral view. Figs. 113-115. Hamataliwa puta (O. P. -Cambridge), male lectofype from Bugaba, Panama. 113. Palpus, ventral view. 114. Tibia of palpus. 115. Palpus, retrolateral view. Figs. 116-119. Hamataliwa puta (O. P. -Cambridge), paralectotypes from Bugaba, Panama. 116. Palpus, ventral view. 117. Tibia of palpus. 118. Palpus, retrolateral view. 119. Tibia of palpus, second paralectofype. Figures 114, 117, 119 demonstrate variability in tibial apophyses of palpi. Scale is for all palpi. Hamataliwa in Mexico and Central America • Brady CO rr> o& e E 126 Bulletin Museum of Comparative Zoology, Vol. 140, No. 3 Figs. 120-121. Hamataliwa triangularis (Kraus), male from Barro Colorado Island, 23-30 June 1939. 120. Left palpus, ventral view. 121. Left palpus, retrolateral view. Figs. 122-123. Hamataliwa g/obosa (F.O. P. -Cambridge), male holotype from Bugaba, Panama. 122. Palpus, ventral view. 123. Palpus, retrolateral view. Figs. 124-125. Hamataliwa flebilis (O. P. -Cambridge), male holotype from Bugaba, Panama. 124. Palpus, ventral view. 125. Palpus, retrolateral view. Figs. 126-127. Hamataliwa crocata sp. n., male holotype from Summit, Panama Canal Zone, 23-28 Aug. 1950. 126. Palpus, ventral view. 127. Palpus, retrolateral view. Figs. 128-129. Hamataliwa tricuspidata (F.O. P. -Cambridge), male holotype from Bugaba, Panama. 128. Palpus, ventral view. 129. Palpus, retrolateral view. Hamataliwa in Mexico and Central America • Brady 111 CM CO cvi O cvj mlk etin Museum of Comparative Zoology The Spider Family Oecobiidae in North America, Mexico, and the West Indies WILLIAM A. SHEAR HARVARD UNIVERSITY VOLUME 140, NUMBER 4 CAMBRIDGE, MASSACHUSETTS, U.S.A. AUGUST 6, 1970 PUBLICATIONS ISSUED OR DISTRIBUTED BY THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Breviora 1952- Memoies 1864-1938 Johnsonia, Department of Mollusks, 1941- Occasional Papers on Mollusks, 1945- Other Publications. Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine. Reprint, $6.50 cloth. Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of In- sects. $9.00 cloth. Creighton, W. S., 1950. The Ants of North America. Reprint, $10.00 cloth. Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 12, 14, 15. (Price list on request.) Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Teredinidae (Mollusca: Bivalvia). $8.00 cloth. Whittington, H. B., and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948. ( Complete sets only. ) Publications of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138, U. S. A. © The President and Fellows of Harvard College 1970. THE SPIDER FAMILY OECOBIIDAE IN NORTH AMERICA, MEXICO, AND THE WEST INDIES WILLIAM A. SHEAR* ABSTRACT Two genera of the spider family Oecobi- idae are recognized as occurring in the area studied: Oecobius Lucas, 1846, and Platoe- \cobius Chamberlin and Ivie, 1935. The | genera Omamis Thorell, 1870, Thalomia Hentz, 1850, Ambika Lehtinen, 1967, and Tarapaca Lehtinen, 1967, are synonyms of Oecobius. The female of O. beatus, the male of P. floridanus, and mature speci- mens of O. isolatus are described for the first time. Oecobius interpellator, O. bracae, O. juangarcia, O. civitas, O. piaxtla, O. rivulo, O. culiacanensis, and O. iso- htoides are described as new. The social behavior of O. civitas is briefly described. INTRODUCTION Members of the family Oecobiidae are I listed in nearly every faunal survey of North American spiders, primarily on the basis of one or two widespread, synan- ithropic species. An examination of the tropical and subtropical Mexican fauna for this study has increased the number of known species in the region from six to I Fifteen, most of which are well established •apart from human dwellings. Except for the ubiquitous Oecobius hnnulipes Lucas, little is known of the ! habitats of these minute cribellate spiders. *Debski (1922) published a few ecological observations on Oecobius templi O. P.- * Department of Biology, Concord College, Ath- ens, W. Va. 24712. Cambridge, in Egypt. He found webs stretched over small projections on walls and on the bark of Eucalyptus trees; the major item of prey seemed to be ants. Millot (1931, 1938) studied the anatomy of the spinning apparatus of Oecobius cellario- rum (Duges) in great detail, but did not publish any observations on how these or- gans were used by the spider. Most valuable is a recent study by Glatz (1967) on the web construction, feeding behavior, and mating of O. annulipes. Glatz found that in southern France, O. an- nulipes constructs two kinds of webs, one of an irregular starlike shape, and another, more tubular type. I have observed the first type constructed by specimens of O. an- nulipes in Torreya State Park, Liberty County, Florida. The main food of the French specimens was found to be two spe- cies of ants, Plagiolepis pygmaea and Lasius flavus (Glatz, 1967); in Florida I observed the spiders feeding on minute Diptera. Of great interest are Glatz's observations on the use of Oecobius extraordinary anal tubercle. The stout, curved bristles with which the apical joint of the anal tubercle is ringed are used to comb the silk from the large pos- terior spinnerets, much as other cribellate spiders card silk from the cribellum with the calamistrum on the metatarsus of the fourth legs. Only rarely did Glatz observe O. annulipes using cribellar silk to swath prey. Copulation takes place in a special mat- ing web constructed by the male. The Bull. Mus. Comp. Zoo!., 140(4) : 129-164, August, 1970 129 130 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 male takes a position facing the female and slightly beneath her, and inserts the right palpus into the left side of the female epigynum, and vice versa, five or more times. Each egg sac contains three to ten eggs, which receive no protection or care from the female after they are laid (Glatz, 1967). Several partly synanthropic species, such as O. putus, O. concinnus, and O. beatus, presumably exhibit similar behavior under stones and in leaf litter or on plants. Of considerably greater interest is the large number of new species ranging from northern Arizona to Oaxaca. While many of these species are to be found in situ- ations similar to that described for O. annulipes (when living apart from man), at least one, O. civitas, is known to be communal. The great numbers of individu- als comprising single collections of O. culi- acanensis and O. rivula point to a possible communal existence for these species also. The synanthropic or partly synanthropic species of Oecobius are widely distributed. Oecobius annulipes is known from North and South America, Europe, North Africa, Southeast Asia, and numerous oceanic islands. Oecobius concinnus occurs in the West Indies to the exclusion of other oecobiids, and has also been found in peninsular Florida, Venezeuela (Simon, 1892) and coastal Mexico. Records of this species in the literature (as O. nieborow- skii) include Costa Rica ( Kulczyinski, 1909, type of O. nieljorowskii, illustrated) and the Galapagos Islands (Banks, 1931). In contrast, the "native" American species are often quite local, although more in- tensive collecting may show them to be widelv distributed. Oecobius isolatus is found only in Baja California and the lower Colorado River basin; O. culiacanensis is limited to central Sinaloa and western Durango. I wish to gratefully acknowledge the aid of Dr. Willis J. Certsch, American Museum of Natural History; Dr. Herbert W. Levi, Museum of Comparative Zoology; Mr. Charles L. Bailey, Oklahoma State University; Miss Lois B. O'Brien, Cali- fornia Academy of Sciences; and, Miss Patricia Rodgers, Philadelphia Academy of Sciences, all of whom kindly loaned ma- terial from the collections under their care. Dr. Gertsch made many valuable sug- gestions during the course of the study, and Dr. Levi read and edited the manu- script. M. Jean-Claude Ledoux sent me drawings of type material in the London and Paris Museums, without which I would have made numerous errors in assigning species names. Mr. Vincent D. Roth, Dr. Joseph A. Beatty, Dr. Andrew A. Weaver, Dr. Martin H. Muma, Dr. B. J. Kaston, and M. J.-C. Ledoux loaned or donated material from their private collections. The faculty and administration of Concord College, Athens, West Virginia, where most of the work for this study was completed, made time and work space available. This investigation was supported in part by Public Health Service Research Grant AI-01944 from the National Institute of Allergy and Infectious Diseases to H. W. Levi. TAXONOMY OECOBIIDAE Oecobiidae Blackwall, 1862, Ann. Mag. Nat. Hist., 3(9) 382. Family Oecobiidae including only the type genus Oecobius Lucas, 1846. Omanoidae Thorell, 1869, Nova Acta Reg. Soc. Sci. Upsaliensis, 3: 44. Type genus Oniamts ( = Oecobius, objective synonym ) . Oecobiinae Chamberlin and Ivie, 1935, Ann. Entomol. Soc. Amer., 28: 267. Subfamily under Uroteidae; Lehtinen, 1967, Ann. Zool. Fenn., 4: 303. Subfamily under Oecobiidae. Note. Regardless of whether one con- siders Uroctea confamilial with Oecobius or not, the family name of the family in- cluding Oecobius must be Oecobiidae, a family name which antedates Urocteidae Thorell, 1869, by seven years. Until more detailed investigations are carried out on the structure and development of the cribellate Oecobius and the ecribellate Spider Family Oecobiidae • Shear 131 Uroctea, I feel it is best to maintain the two separate families. Diagnosis. Oecobiidae are separated from all other families of spiders except Urocteidae by the large, two-jointed anal tubercle with a double fringe of curved hairs. They may be separated from the Urocteidae by the possession of a cribellum and calamistrum. Description. Cribellate spiders of the suborder Labidognatha. Respiratory sys- tem-pair of book lungs at base of abdomen and single tracheal spiracle opening im- mediately in front of cribellum. Spinnerets six, set close together, median pair smallest, posterior pair largest, with apical joint long, pointed, turned dorsally. Cribellum partially divided distally. Anal tubercle large, movable, two-jointed, with fringe of long, simple, sinuate hairs around base of apical joint; apical joint tipped with long sensory hairs. Carapace suboval to reni- form, usually wider than long. Chelicerae small, without boss, convergent and touch- ing throughout their length; fang minute, fang groove and cheliceral teeth lacking. Labium wider than long, free. Endites convergent over labium, almost touching. Sternum heart-shaped, wider than long, pointed posteriorly, separating posterior coxae. Eight eyes, in two rows, AME dark, PLE dark or light, others degenerate. Three claws, paired claws with single row of teeth, median claw small, accessory claws present; female pedipalp with single toothed claw. Calamistrum somewhat sup- pressed, in double series, absent in males. Female copulatory organ of the entelegyne type, one or two pairs of spermathecae present, opening directly into vaginal area or by means of copulatory tubes. Male palpus complex, pre-embolic sclerites sup- pressed, males with fringe of specialized spatulate hairs on sternal margin. As regards non-genitalic characters, the members of the family are remarkably uniform, though several general trends of variability may be noted. Carapace shape varies from nearlv round to reniform. and seems to be correlated with body size-larger species having comparative K broader and flatter carapaces. There are considerable age and sexual differences in carapace shape as well. Juveniles and males have much rounder, higher carapaces than adult females of a given species, and the carapace of the males of some species is modified for muscle attachments of the massive palpi. This modification usually takes the form of two lunate depressions on either side of the carapace. Macrosetae associated with the eyes have an arrange- ment characteristic of each genus studied, and will be described below. The legs lack macrosetae in most of the species, but some rather strong setae appear on the legs of larger species, and this character also appears to be correlated with size. Additional secondarv sexual characters in the males are the lack of a calamistrum and great reduction of the cribellum, and the presence on the margins of the sternum of strong, spatulate hairs of unknown function. The most important non-genitalic char- acter of the oecobiids is the remarkable anal tubercle. This large, two-jointed, movable structure has been described in detail by Millot (1931, 1938), and by Glatz ( 1967 ) . The possibility of the anal tubercle and the elongated posterior spin- nerets replacing the calamistrum and cribellum respectively has already been mentioned. The loss of the calamistrum in males and the poorly differentiated nature of this structure in females provides ad- ditional evidence for this view. To my knowledge, the only species er- roneously placed in the Oecobiidae is "Oecobius" sapporensis (Saito, 1934). Al- though specimens were not available for study, it is clear from the excellent color figures that the type specimen is probably a theridiid. Yaginuma (1962) was the first to point out the erroneous placement of this species, but his paper was overlooked by Kritscher (1966), who included sap- porensis in Oecobius. 132 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 Key to Oecobiid Genera of the World la. Tibia I about six to seven times longer than wide; calamistrum running two- thirds the length of metatarsus IV Oecobius. p. 135 lb. Tibia I about four times longer than wide; calamistrum running entire length of metatarsus IV . Platoecobius. p. 161 THE GENERIC PROBLEM IN OECOBIIDAE As is usual in spiders, different oecobiid genera and species are separated primarily by distinctions in the genitalia and eye patterns. The eyes in species of Oecobius assume two distinct patterns. In the majority of species in the genus, the eyes are arranged in two straight to slightly procurved rows, with the posterior lateral eyes the largest, followed, in order of size, by the anterior medians, anterior laterals, and posterior medians. The anterior laterals and pos- terior medians are opalescent and irregular in shape (Fig. 14). In O. putus and Platoecobius floridanus, the rows are more distinctly procurved and the anterior medians are largest, followed, in order of size, by the posterior laterals, the anterior laterals, and the posterior medians. The anterior laterals and posterior medians are irregular and opalescent, and the posterior laterals are round, but light in color ( Figs. 17, 27). Oecobius cellariorum (Fig. 13) has a pattern that is variable, but usually intermediate between these two. The palpal organs of male Oecobius ap- pear simple and easily described in the contracted state, but treatment of the palpus with an expanding agent reveals considerable complexity. The most obvious unusual feature of Oecobius male palpi is the extensive elaboration of the embolic division. The apparent tegulum in an un- expanded palpus is in reality the much enlarged radix, or basal part of the em- bolus. In the more complex representatives of the genus, the embolus is further sub- divided, forming three distinct sclerites. Homologies in such an unusual palpal type cannot clearly be established, but it is useful to apply the same terms to struc- tures that are analogous in other spider groups. The terms used here are based on those originated by Comstock (1940), and further described by Shear (1967). The order in which the following descriptions are arranged is not intended to suggest a possible phylogeny, rather, simpler palpi are described first. Oecobius putus shows most of the features common to the palpi of a majority of the species of Oecobius. In a lateral view of an extended palpus (Fig. 1), the tegulum (t) can be seen to articulate with the enlarged basal part of the embolus, which may be termed the radix (r), by means of a strong sclerotized rod fitting into an internal socket (possibly this rod could be called a median apophysis). The conductor (c) is large and complex. In a mesal view (Fig. 2), the incomplete, spiral nature of the radix is revealed, and the articulation between the basal part and the intromittent portion of the embolus (e) is clear. Oecobius cellariorum, the type species of the genus, differs from O. putus only in that the intromittent portion of the embolus (e) bears a large curved hook near its articulation with the radix (r), The actual intromittent portion of the embolus cannot be seen in either of the figures (Figs. 3, 4), but is a short, slightly curved spine. The palpus of O. concinnus (Figs. 5, 6) is more complex than the preceding two examples. The large, curved hook of the embolus is present as a separate sclerite, which can be designated the stipes (s). The distal portion of the1 embolus (e) articulates with the stipes and not with the radix (r). The radix bears a thin sclerotized rod on its lateral surface. The conductor (c) is reduced in size. Oecobius rivula (Figs. 7, 8) is repre- sentative of a group of Mexican species whose palpi differ in a few particulars from O. concinnus. The process of the radix ( r) is present in this group as a thick, Spider Family Oecobiidae • Shear 133 heavy apophysis (radical apophysis, a), and the radix is reduced in size. The stipes (.s) is enlarged mesally. The distal portion of the embolus is a small, curved spine ( e ) . Of some interest is the perforation in the stipes of these species, opposite the median apophysis of the tegulum (f). In O. annulipes (Figs. 9, 10), a similarity to the O. putus-O. cellariorum plan is obvious. The embolus (e) and conductor (c) are subequal in size, and the radix (r) is a spiral making slightly more than one turn. In Platoecobius floridamts, illustrated in Figures 11 and 12, the palpus is simpler than in O. annulipes. The tegulum (?) may be represented only by a ringlike patch of slightly heavier sclerotization on the basal hematodocha ( /;/; ) . The spiral radix (r) is much enlarged and makes one and one-half full turns. The conductor (c) is small and partly membranous. The stipes (s) is also small, but bears a distal hook opposite its articulation with the intromittent part of the embolus (e). The track of the sperm tube inside the palpus is much convoluted, and is valuable in ascertaining the relationships of the scle- rites. The female genitalia may be discussed from two aspects, external and internal. Externally, features common to most species of Oecobius are a scape of ex- tremely variable size and shape, and a large, common fossa from which the open- ings of the copulatory ducts lead to the I spermathecae. In O. annulipes (Fig. 29), I the scape is long and narrow, while in O. I civitas (Fig. 43), O. isolatus (Fig. 35), O. I isolatoides (Fig. 37), and related Mexican I species, the scape is broad. In each of j these, however, the fertilization tubes are I plainly visible, and they open at the distal J end of the scape, or into notches that run to the distal end of the scape (O. culia- ucanensis, Fig. 45). In O. cellariorum (Fig. j 28), O. beatus (Fig. 33), and O. bracae I (Fig. 41), the scape is very broad and \ short, but otherwise comparable to the above species. A slightly different pattern occurs in O. concinnus (Fig. 31), O. juan- garcia (Fig. 40), and O. rivula (Fig. 39). The scape is folded anteriorly, then some- what posteriorly, and the fertilization tubes open into or near a common fossa that may be set back from the distal margin of the scape. Oecobius interpolator (Fig. 36) is apparently not very closely related to the other species studied here; both a scape and common copulatory fossa are absent. Platoecobius floridanus (Fig. 46) departs radically from this basic external pattern by being practically devoid of any external epigynal modifications, except for an an- terior sclerotized notch of unknown func- tion. Internally, the basic pattern in Oecobius shows little variation. The copulatory ducts vary in length; in O. annulipes (Fig. 30) and O. civitas (Fig. 44) they are moder- ately long, but in O. bracae (Fig. 42) and O. concinnus (Fig. 32) they are practically nonexistent. The spermathecae are heavy- walled, and vary in shape from rather rounded (O. civitas, Fig. 44) to more elongate (O. concinnus, Fig. 32). In species with long copulatory ducts, the spermathecae are always lateral to the copulatory openings, except in O. annu- lipes (Fig. 30), where they are anterior and mesal. The fertilization tubes are long in all species, and traverse the scape for at least the distal third of their length. Once again, Platoecobius floridanus (Fig. 47) differs considerably from the other oecobiids studied. The heavy-walled primary spermathecae are sessile and pos- terior to the copulatory openings. Con- necting ducts run laterad and anterior to open into large, weakly sclerotized second- ary spermathecae, from which they con- tinue mesoposteriad as fertilization tubes to open, without external modification, on the ventral surface of the abdomen an- terior to the copulatory openings. Recently, Lehtinen ( 1967 ) has divided the family Oecobiidae into six genera, of which three are new. Two of the three 134 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 genera are monotypic. Thalamia is resur- rected from synonymy, and Platoecobius Chamberlin and Ivie 1935, is retained. An examination of the chart (Lehtinen, 1967: 304) in which the characters of the pro- posed genera are given reveals that the new genera are erected primarily on the grounds of differences in the eyes and genitalia. Examples of characters in the chart are such items as abdominal color pattern, which I believe can vary with feeding and the reproductive state of the spider, the presence of strong leg macrosetae and trichobothrial distribution, [which Lehti- nen (1967: 305) indicates "...are easily explained by the general correlation of these characters to the average size of the spiders."], and a series of other characters that are uniform throughout the family. In addition, errors of observation and inter- pretation have crept in; the "fovea" (thoracic groove) of all oecobiids ex- amined by me was very shallow and transverse, yet Lehtinen ( 1967 ) lists five different forms, ranging from "totally ab- sent" to "long and rather deep." The comments on genitalia are filled with terms not explained nor illustrated. The genus Oecobius sensu Lehtinen is said to be Mediterranean in distribution, but includes O. formosensis Kishida, of the Orient, and O. cellariorum, synanthropically wide- spread in Europe and North America; Platoecobius is said to be distributed through the "N. Neotropical- S. Nearctic" regions, but the only species, P. jloridanus (Banks), is known only from Florida, Georgia, and South Carolina in the United States. In addition, three of the genera are based on a knowledge of only the female sex of the included species. The discussion of eye characters and genitalia above delimits the ranges of vari- ability in the fifteen species and nearly 1500 specimens studied by me, including both sexes in all eases. Certainly, the eye arrangement of Oecobius putus, for ex- ample, selected by Lehtinen as the type species of Ambika, is distinct from that of most other members of Oecobius sensu latu; but in many of the males of Oecobius cellariorum, type species of Oecobius, a similar pattern with enlarged anterior median eyes, is found. Clearly, O. putus is not in any other way except eye arrange- ment related to Platoecobius jloridanus, but its genitalia are quite within the range of variation of Oecobius. Thus it would seem that eye pattern is subject to con- vergence in otherwise unrelated species. Perhaps similar habits, as yet undescribed, provide the selection pressures producing this convergence. The genitalia of O. nieborowskii ( = O. concinnus Simon), type species of Lehtinen's genus Tarapaca, also fall within the variation found in Oecobius. Most modern taxonomists feel that the genus represents a group of closely related species that occupy a similar ecological niche and may be separated from other such groups by a distinct morphological discontinuity, although this latter criterion is often soft-pedalled (Mayr, Linsley, and Usinger, 1953: 57-59; Mayr, 1963:' 588- 592). It has also been emphasized that the criteria by which genera are delimited are essentially subjective, and that little is gained by splitting long-recognized, seem- ingly natural higher taxa (Mayr, Linsley, and Usinger, 1953: 59). Taking this into account, along with the evidence of the similar genital pattern of Oecobius species presented above, it is my opinion that three of Lehtinen's genera, Aml)ika, Tarapaca, and Tlialamia, must be considered syno- nyms of Oecobius. I have refrained from commenting on the genus Maitreja Lehti- nen 1967, type species Maitreja marathaus (Tikader), from India, since I have been unable to examine any specimens of this species, but the published figures (Tikader, 1962) do not seem to warrant the retention of Maitreja as a generic name. Indeed, M. marathaus may be a synonym of O. cella- riorum. Spider Family Oecohiidae • Shear 135 However, after studying in detail the general structure and genitalia of Phitoe- cobius floridanus including, for the first time, male specimens, I concur with Chamberlin and Ivie (1935) and Lehtinen (1967) that it deserves generic recognition. In addition to its unique genitalia, it is the only oecobiid known not to weave a web for capture of prey (Chamberlin and Ivie, 1935), thus occupying a distinct ecological niche. Therefore, two genera, Oecohius Lucas 1846 and Plotoecobius Chamberlin and Ivie 1935, are recognized from the area of this study. Genus Oecobius Lucas 1846 Oecohius Lucas, 1846, Expl. Sci. Algerie, Zool. I, Arach., p. 100, type species designated by Thorell, 1869, (On European Spiders, p. 112), Oecohius domesticus Lucas ( = O. ccllariorum (Duges); Simon, 1875, Arach. France, 2: 6; 1892, Ann. Entomol. Soc. France, p. 435; Comstock, "1912" (1913), The Spider Book, p. 288; Chamberlin and Ivie, 1935, Ann. Entomol. Soc. Amer., 28: 267; Kaston, 1948, Bull. Connecticut Ceol. Nat. Hist. Surv., no. 70, p. 499; Kaston, 1953, How to Know the Spiders, p. 36; Lehtinen, 1967, Ann. Zool. Fenn., 4: 253, p. 304. Omanus Thorell, 1870, Nova Acta Beg. Soc. Sci. Upsaliensis, ser. 3, p. 114, type species by original designation Oecohius navus Blackwall ( = O. annulipes Lucas); Keyserling, 1891, Die Spinnen Amerikas, Vol. 3 ( Brasilianische Spinnen), p. 160. Thalamia Hentz, 1850, J. Boston Soc. Nat. Hist., 6: 35, type species by monotypy Tlialamia parietalis Hentz ( = O. annulipes Lucas); Banks, 1890, Proc. Entomol. Soc. Washington, 2: 125; Lehtinen, 1967, Ann. Zool. Fenn., 4: 269, 304. Ambika Lehtinen, 1967, Ann. Zool. Fenn., 4: 212, 304, type species by monotypy Oecohius putus O. P.-Cambridge. NEW SYNONYMY. Tarapaca Lehtinen, 1967, Ann. Zool. Fenn., 4: 267, 304, type species by monotypy Oecohius nieborowskii Kulczynski ( = O. concinnus Simon). NEW SYNONYMY. Diagnosis. Oecobius is separated from Plotoecobius, the only other genus in the family, by having longer, thinner legs, the calamistrum of the females extending two- thirds the length of metatarsus IV, and the more complex genitalia. Description. Small oecobiid spider (1.5-4.5 mm total length) with characters of the family. Carapace with sides rounded, clypeus prolonged into subtriangular pro- jection. Carapace flat to moderately high, head region moderately elevated. Eyes on low tubercle, eye area wider than long. Anterior median eyes ( AME) and posterior lateral eyes (PLE) round, dark; anterior lateral eyes (ALE) and posterior median eyes (PME) irregular, light. Maerosetae of eye area as follows: three between AME, one behind each PME, two nearly contig- uous on midline behind eye area; position of maerosetae varying somewhat with eye arrangement of species. Thoracic groove transverse, absent to indistinct. Carapace depressed and indented above pedicel. Abdomen flattened, elongate to suboval, somewhat pointed behind. Leg formula usually 4213 or 421=3. Legs long, thin, spines weak to lacking, without definite arrangement. Tarsi without trichobothria, metatarsal trichobothria one or two, with- out definite arrangement. Hind coxae separated by sternum. Calamistrum lack- ing in mature males, extending in immature specimens and females along proximal two- thirds of metatarsus IV. Palpi of mature males with tegulum and subtegulum greatly suppressed, visible only in treated palpi, conductor large, well sclerotized, embolus consisting of two or three selerites, intromittent portion a short, curved spine. Female with more or less elaborate epig- ynal modification, one pair of heavily sclerotized spermathecae. Fourteen species are known from the region of this study; most of them are very similar, and best separated by reference to the figures. Only the most significant references are given under each species; for more detailed references, see Roewer (1954), and Bonnet (1958). Key to Species of Oecobius: la. AME larger than or equal to PLE (Figs. 13, 17) 2. lb. AME five-sixths the diameter of PLE, or smaller (Fig. 16) 3. 136 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 2a. Eye area nearly twice as wide as long, or wider (Fig. 17); PLE pale but not opalescent putus. p. 144 2b. Eye area nearly square (Fig. 13); PLE dark cellariorum. p. 136 3a. Carapace yellowish white, unmarked, or with dark markings on a yellowish white ground (Figs. 13, 16) 4. 3b. Carapace medium brown to black, lighter individuals with a dark pattern (Figs. 18, 22) 9. 4a. Carapace unmarked except for marginal dark line and often a dusky central band (Fig. 13) cellariorum. p. 136 4b. Carapace with additional markings, usually taking the form of three dark spots on each side (Figs. 14, 15), or of submarginal bands (Figs. 20, 24) ._ 5. 5a. Submarginal spots well separated from each other and from marginal line (Figs. 14, 15) 6. 5b. Submarginal spots fused to each other (Fig. 24), or to marginal line (Fig. 20), or both 8. 6a. Epigynum without a scape (Fig. 36); male palpus in lateral view with a large, serrate, distally hooked conductor (Fig. 58); Cambridge, Mass., and New Cale- donia (?) interpellator. p. 146 6b. Not as in 6a 7. 7a. Submarginal spots usually commalike, nearly touching (Fig. 15); epigynal scape broader than long (Fig. 31 ); male palpus with a large stipal process in mesal view (Fig. 52); West Indies, Florida, coastal Mexico, Central America, Venezuela, Columbia concinnus. p. 141 7b. Submarginal spots oval, well separated (Fig. 14); epigynal scape three to four times longer than broad ( Fig. 29 ) ; stipal process of male palpus small (Fig. 51); pantropical, north to Massachu- setts and Oregon on the North American coasts annulipes. p. 138 8a. Submarginal spots usually connected to form a submarginal band which is then connected to the marginal line ( Fig. 24); epigynal scape squared-off poste- riorly and notched (Fig. 41); male palpus as in Figure 70; Oaxaca hracae. p. 156 8b. Submarginal spots usually separated from each other, but connected to marginal line (Fig. 20); epigynum with a very short, broad scape bearing two posterior digital processes (Fig. 33); male palpus as in Figure 54; Cuerrero, Tamaulipas hcatus. p. 143 9a. Carapace of adults evenly dark brown to black (Fig. 23), immatures more lightly colored, often marked darker; epigynal scape with a central depression ( Fig. 40 ) ; male palpus lacking radical apophysis (Fig. 67; compare with Fig. 69); Oaxaca juangarcia. p. 154 9b. Carapace of adults dark to light brown, but generally with a much darker pat- tern or shading (Figs. 18, 19, 21); epigynal scape without a central depres- sion; male palpus with radical apophysis (except for O. civitas, Figs. 72, 73) ._.. 10. 10a. Epigynal scape as wide or wider than long, not distinctly narrowed distally (Fig. 43); male palpus without a radical apophysis (Figs. 72, 73); west-central Mexico civitas. p. 157 10b. Epigynal scape longer than wide, or if not, then distinctly narrowed distally; male palpus with radical apophysis 11. 11a. Epigynal scape as wide or wider than long, distinctly narrowed distally ( Fig. 38); male palpus as in Figures 64, 65; Sinaloa piaxtla. p. 150 lib. Epigynal scape longer than wide, nar- rowed distally or not (Figs. 35, 37); male palpus not as in Figures 64, 65; .... 12. 12a. Fertilization tubes opening into distal notches on epigynal scape (Fig. 45), scape not narrowed distally; male palpus as in Figures 74, 75; Sinaloa culiacanensis. p. 159 12b. Fertilization tubes not opening into dis- tal notches; male palpus not as in Figures 74, 75 13. 13a. Scape with a small perforation near tip, narrowed distally ( Fig. 39 ) ; male palpus massive, as in Figures 68, 69; Sinaloa rivula. p. 152 13b. Scape without such a perforation; male palpus otherwise 14. 14a. Scape narrowed distally ( Fig. 37 ) ; male palpus as in Figures 62, 63; Arizona, Sonora, N. Sinaloa .... isolatoides. p. 150 14b. Scape not narrowed distally (Fig. 35); male palpus as in Figures 60, 61; lower Colorado River valley, Baja California isolatus. p. 148 Oecobius cellariorum (Duges) Figures 3, 4, 13, 28, 48, 49; Map 1 Clotho cellariorum Duges, 1836, Ann. Sci. Nat., 2: 161. Types presumed lost. Type locality not designated. Oecobius cellariorum, Simon, 1875, Arach. France, 2: 7. Roewer, 1954, Katalog der Araneae, 2: 1288. Bonnet, 1958, Bibliographia Araneorum, 2: 3132. Kritscher, 1966, Ann. Naturhist. Mus. Wien., 69: 287, pi. 1, fig. 4, 5, 6, $9. Spider Family Oecobiidae • Shear 137 • cellari o annu Map 1. North America, showing distribution of O. cellariorum and O. annuiipes Oecobius domesticus Lucas, 1846, Expl. Sci. Algerie, Zool. I, Arach., 100-101, pi. 2, figs, la-lg, 2 . Types presumably in the Paris museum, not examined; type locality, Algiers, Algeria. Oecobius texanus Bryant, 1936, Psyche, 43: 87, figs. 8a-8e, c5 $ . Male holotype, female para- types from Dallas, Texas, in Museum of Com- parative Zoology, examined. Gertsch and Mulaik, 1940, Bull. Amer. Mus. Nat. Hist., 97 335. Muma, 1944, Amer. Mus. Novitates, 1257 1. Boewer, 1954, Katalog der Araneae, 2 1290. Bonnet, 1958, Bibliographia Araneorum, 2: 3135. NEW SYNONYMY. Diagnosis. This species is distinguished from all other Nearctic species except O. putus by the virtually unmarked carapace, and from O. putus by the eye relations ( cf. Figs. 13, 17) and the smaller size of O. cellariorum. Description. Female from Las Cruces, New Mexico: carapace (Fig. 13) suboval, wider than long (L/W = 1/1.3), almost glabrous, pale hairs on margins and lateral to eye area. Clypeus slightly prolonged, steeply sloping, about ten degrees from vertical. Eye area highest point of cara- pace, sloping steeply to clypeus and later- ally, prolonged behind almost to thoracic groove, then declining sharply to posterior margin of carapace. Eye area wider than long (L/W =1/1.5), eyes in two rows, posterior row recurved, anterior row nearly straight. PLE the largest, dark, separated 138 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 by about two and one-half times their diameter; PME irregular, opalescent, sepa- rated by scarcely one-fourth diameter of PLE, and from each other by their diameter, nearly contiguous with PLE; ALE irregular, opalescent, nearly con- tiguous with AME. Labium subtriangular, wider at the base than long. Sternum longer than wide ( L/W = 1/.83 ) , anterior margin slightly excavated to receive labium, bluntly pointed behind, separating pos- terior coxae by slightly more than their width, clothed sparsely with long hairs, more dense around margins. Abdomen suboval, pointed behind, widest slightly anterior to middle of length, densely set with colorless hairs. Epigynum (Fig. 28) with two strongly chitinized lateral lobes overhanging a transversely wrinkled area. Carapace pale yellowish white, except for dark marginal line, dusky area sur- rounding eyes; eyes ringed with black. Abdomen pale gray with chalky white blotches, except in cardiac area, three pairs of distinct dark spots marking ab- dominal apodemes. Venter pale gray blotched chalky white, spinnerets slightly darker. Endites, labium, sternum, and coxae of legs uniform pale yellowish white. Legs slightly darker distally, femoral bases off-white, tarsi somewhat dusky; vaguely indicated annuli median on femora and tibiae of anterior legs and distally on meta- tarsi. Total length, 2.90 mm. Carapace .90 mm long, 1.02 mm wide. Tibiae I-IV .83, .87, .84, .90 mm long, respectively. Metatarsi I-IV .78, .87, .84, .98 mm long, respectively. Male from Las Cruces, New Mexico, with structure essentiallv as in female, except as follows: carapace proportionally wider than in the female; eye area higher; and clypeus more steeply sloping, nearly vertical. PLE reduced in size, subequal to AME. Legs longer and thinner than in female. Sternum margined with special hairs which are spatulate and darkened dis- tally. Coloration as in female. Palpus as in Figures 48, 49. Total length, 2.18 mm. Carapace .67 mm long, .95 mm wide. Tibiae I-IV .82, .87, .92, .96 mm long, respectively. Metatarsi I-IV .82, .90, .92, 1.00 mm long, respectively. Ecological Notes. In the southwest, this species is fairly common both in and out of buildings; in the more northerly parts of its range it is restricted to indoor habi- tats. Males in Texas and New Mexico mature in June, and disappear in August. Immature males with swollen palpi appear in October and December. Females mature outdoors in July, but persist indoors the year round. The mode of life is essentially that described by Glatz (1967) for 6. annulipes. Distribution. (Map 1) From extreme western New Mexico across Texas and Oklahoma to Arkansas, Missouri, and Kan- sas; in Tennessee, Maryland, and North Carolina in the east. Absent from most of the range of O. anulipcs. Common in Europe, but probably more closely related to North American forms, and may have originated here. Oecob/us annulipes Lucas Figures 9, 1 0, 1 4, 29, 30, 50, 51 ; Map 1 Oecobius annulipes Lucas, 1849, Expl. Sci. Algerie, Zool. I, Arach., 102, figs. 2a-2g, 9 ; types probably in Paris museum, not examined; type locality, Algiers. Simon, 1875, Arach. France, 2: 9. Kritscher, 1966, Ann. Naturhist. Mus. Wien., 68: 285, pi. 1, figs. 1-3, $ 9. Thalamia parietalis Hentz, 1850, J. Boston Soc. Nat. Hist., 6: 35, pi. 15, fig. 16, 9 ; types de- stroyed; type locality, "S. Alabama." Oecobius parietalis, Simon, 1892, Histoire Xatur- elles des Araignees, 1: 247, figs. 192, 197, 9. Emerton, 1909, Trans. Connecticut Acad. Arts Sci., 14: 212, pi. 7, figs, la-le, i 9 . Comstock, Expanded right palpi of males. Figs. 1-12. 1. Oecob/us pufus, lateral view. 2. O. putus, mesal view. 3. O. cellariorum, lateral view. 4. O. cellariorum, mesal view. 5. O. concinnus, lateral view. 6. O. concinnus, mesal view. 7. O. rivula, lateral view. 8. O. rivula, mesal view. 9. O. annulipes, lateral view. 10. O. annulipes, mesal view. 11. P/a/oecobius lloridanus, lateral view. 12. P. llondanus, mesal view. Spider Family Oecobiidae • Shear 139 140 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 "1912" (1913), The Spider Book, p. 288, fig. 275, 9. Chamberlin and Ivie, 1935, Ann. Entomol. Soe. Amer., 28: 267, figs. 1, 3-17, 9 $ . Kaston, 1948, Connecticut State Geol. Nat. Hist. Bull., 70: 599, pi. 101, figs. 1878-1881, 9 $ ; Kaston, 1953, How to Know the Spiders, p. 36, figs. 69-70, 9. Roewer, 1954, Katalog der Araneae, 2: 1288. Bonnet, 1958, Biblio- graphia Araneorum, 2: 3131-3132. Diagnosis. The pattern of the carapace (Fig. 14) is characteristic, but similar to O. inter])e11ator and O. concinnus. Oe- cobius annulipes is intermediate between these two in size. The palpus of O. con- cinnus (Fig. 52) has two large, projecting apophyses, and the epigynal scape ( Fig. 31 ) is wider than long. An epigynal scape is lacking in O. interpellator, and the palpus is distinctive (Fig. 58). The palpus of O. annulipes (Fig. 50) is compact, with- out large projections, and the epigynal scape (Fig. 29) is long and pointed. Description. Female from Sunland, Cali- fornia: carapace (Fig. 14) suboval, sparsely clothed with small hairs, wider than long (L/W = 1/1.3). Clypeus pro- longed beyond eyes by one-tenth length of carapace and evenly rounded. Eye area wider than long (L/W =1/1.3). PLE much the largest, separated by little less than two diameters. PME irregular, opales- cent, nearly contiguous with PLE, sepa- rated by their greatest width. AME dark, about two-thirds diameter of PLE, sepa- rated from each other by a radius and from the PLE by a radius. ALE light, irregular, nearly contiguous with AME. Anterior eye row and posterior eye row of nearly the same width, both rows procurved so that anterior margin of PME is on line drawn through centers of PLE, and pos- terior edge of ALE is on line drawn through centers of AME. Epigynum as in Figures 29, 30, with prominent scape bear- ing two small openings at its distal end and traversed by pair of tubes, openings of epigynal ducts concealed at base of scape, running short distance internally to two seminal receptacles. Legs clothed with fine hairs, some plumose, weak spines ar- ranged irregularly, strongest situated dor- sally at distal ends of femora and patellae, and ventrally at distal ends of tibiae. Number of bristles found in calamistrum varies from specimen to specimen, lowest number found being 19, greatest 27, but in all cases inner row has greater number of bristles. Remainder of structure in close agreement with O. cellariorum. Carapace yellowish white, marginal band gray. Eyes surrounded by black rings blending off into gray area extending backwards and tapering to thoracic groove, then continu- ing as band as wide as groove to posterior margin of carapace. Anteriorly, gray area constricted between AME, widening abruptly, extending down clypeus, tip of clypeal projection light. Three dusky spots on carapace about midway between the marginal band and median dark area, opposite coxae of legs I, II, and III. Legs paler yellow, annulated as follows. Femur with two annuli, incomplete dorsally, one about middle of its length, other at distal end; patella dark ventrally; tibia with two annuli, incomplete dorsally; metatarsus with two annuli paler dorsally but com- plete; tarsi without annuli but slightly darker distally. Sternum, labium, endites, and coxae of legs uniform pale yellow; palpus dusky towards distal end. Abdo- men grayish, mottled white, with darker, unmottled cardiac mark. Anterior margin of abdomen with dark band extending laterad about one-half length of abdomen, then replaced by three pairs of slanting dark marks that extend mesally almost to cardiac mark. Spinnerets and anal tubercle whitish, gray dorsally, surrounded by dusky area. Venter pale. Total length 2.91 mm. Carapace .70 mm long, .98 mm wide. Tibiae I-IV .83, .86, .82, .83 mm long, respectively. Metatarsi I-IV .74, .82, .83, .90 mm long, respectively. Male from Sunland, California, with structure essentially as in female, except as follows. Carapace wider in proportion to its length (L/W = 1/1.4), clypeal pro- jection slightly more acute. Eye area more Spider Family Oecobiidae • Shear 141 compressed, but only slightly wider in proportion to its length. Carapace more nearly glabrous than in female, but hairs of legs and sternum much longer. Abdo- men much smaller in proportion to cara- pace, with longer hairs. Fringe on anal tubercle less well developed. Cribellum small. Calamistrum absent from fourth metatarsus. Color pattern less well de- veloped than in female. Marginal band of carapace broken between coxae of legs, paired spots of carapace only faintly indi- cated, median dusky area limited to eye region and thoracic groove. Markings on abdomen fainter, dark coloration on dorsal side of apical segment of posterior spin- nerets more pronounced. Palpus as in Figures 50, 51. Total length, 2.58 mm. Carapace .70 mm long, 1.05 mm wide. Tibiae I-IV .74, .81, .78, .81 mm long, respectively. Metatarsi I-IV .75, .78, .81, .85 mm long, respectively. Ecological Notes. Discussed in detail by Glatz (1967) and in section INTRO- DUCTION above. Mature individuals are found throughout the year in houses. It has become established out-of-doors only in the southwestern United States and south into Mexico where it matures in early summer. Distribution. (Map 1) Pantropical. I have examined specimens from Europe, India, South America, Africa, Hong Kong, South Viet Nam, and various oceanic islands, and found little variation. There is little evidence as to the point of origin of this highly synanthropic species. Oecobius concinnus Simon Figures 4, 5, 15, 31, 32, 52, 53; Map 2 Oecobius concinnus Simon, 1892, Ann. Entomol. Soc. France, p. 435, pi. 9, fig. 2, $ ; female holotype and male paratype from Puerto Ca- bello, Venezuela, in the Paris Museum, ex- amined and photographed by J.-C. Ledoux. Roewer, 1954, Katalog der Araneae, 2: 1289. Bonnet, 1958, Bibliographia Araneorum, 2: 3133. Oecobius nieborowskii Kulezynski, 1909, Bull. Internat. Acad. Sci. Cracovie, ann. 1909, sem. II, p. 454, pi. 22, fig. 30, $. Type locality Turrialba, C. R.; types not examined, disposi- tion unknown. Roewer, 1954, Katalog der Araneae, 2: 1290. Bonnet, 1958, Bibliographia Araneorum, 2: 3134. NEW SYNONYMY. Thalamic! nieborowskii, Banks, 1931, Nyt Mag. Naturvidensk (Oslo), 118, p. 272, pi. 2, figs. 7, 8, S. Oecobius benneri Petrunkevitch, 1929, Trans. Connecticut Acad. Arts Sci., 30: 75, figs. 64- 66, $ ; female holotype from Rio Piedras, P. R., in American Musuem of Natural History, ex- amined. Roewer, 1954, Katalog der Araneae, 2: 1289. Bonnet, 1958, Bibliographia Araneo- rum, 2: 3132. NEW SYNONYMY. Oecobius vokesi Gertsch and Davis, 1942, Amer. Mus. Novitates, 1158: 19, Fig. 40, $; female holotype from 30 miles south of Jesus Carranza, Veracruz, in American Museum of Natural History, not examined, presumed lost. Roewer, 1954, Katalog der Araneae, 2: 1290. NEW SYNONYMY. Oecobius audanti Bryant, 1948a, Bull. Mus. Comp. Zool., 100: 334, pi. 1, fig. 4, $, pi. 2, figs. 10, 12, $ ; male holotype, female paratype from Port-au-Prince, Haiti, in Museum of Com- parative Zoology, examined. NEW SYNONYMY. Tarapaca nieborowskii, Lehtinen, 1967, Ann. Zool. Fenn., 4: 433, fig. 30, $. Oecobius beat us, Bryant, 1948, Psyche, 55: 57, (in part) pi. 10, figs. 1, 2, $ ; specimen designated "allotype male" in Museum of Com- parative Zoology, examined. Not O. beatus Gertsch and Davis, 1937, Amer. Mus. Novitates, 961, p. 2, figs. 1,2, 9 ; female holotype of O. beatus in American Museum of Natural History, examined. See Oecobius beatus. Note. This species has accumulated the most impressive synonymy of any Nearctic- Neotropical species. The specimens desig- nated as "allotypes" by Bryant (1948) be- long to O. concinnus, while females found supposedly at the same place and time are typical O. beatus. A vial from the Ameri- can Museum of Natural History from Tamaulipas contained O. beatus males and females. The synonymy of O. vokesi was established by the examination of speci- mens in the American Museum, verified by Dr. Gertsch. The placement of Banks' specimens is based on the close agreement between his figures and male specimens of O. concinnus. Diagnosis. This species overlaps with others only in eastern Mexico and in Penin- sular Florida, where it may be dis- 142 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 Map 2. Caribbean region, showing distribution of O. conc/nnus and P. tloridanus. tinguished from O. annulipes and O. beatus by reference to the figures, and the dis- cussions under those species. No other species is known to occur in the West Indies, as Petrunkevitch (1929) observed in describing O. benneri. Description. Female from St. Augustine, Trinidad: carapace (Fig. 15) suboval, wider than long (L/W = 1/1.2), clypeal projection evenly rounded; carapace high- est in eye area and sloping evenly in all directions, more sharply posteriorly, gla- brous, with a few plumose hairs. Eye area wider than long (L/W =1/1.9), anterior eye row slightly procurved. PLE the larg- est, separated by one and one-third their diameter, PME irregular, opalescent, twice as long as wide, separated by one and one- Spider Family Oecobiidae • Shear 143 third their width, contiguous with inner margins of PLE. AME slightly more than one-half diameter of PLE, separated from each other by one and one-third diameters and from PLE by one-half diameter. ALE irregular, opalescent, contiguous with pro- lateral margins of AME. Epigynum as in Figures 31, 32. Remainder of structure typical. Carapace off-white, eyes ringed with black and surrounded by extensive dusky area. Clypeus entirely dusky, slightly lighter in center. Marginal line dark gray, three comma-shaped dusky marks con- tiguous end-to-end longitudinally between marginal line and central dusky area, faintly indicated gray spots extending to- wards these markings from marginal line. Abdomen grayish white, with a brown cardiac mark and brownish gray lines on sides. Venter mostly chalk-white, spin- nerets light brown. Legs yellow-white with lateral maculae on coxae and trochanters of first two pairs, at midlength and distally on all leg segments except patellae and tarsi, patellae entirely dark laterally, tarsi somewhat darker distally. Maculae form- ing complete annuli on metatarsi. Total length, 1.37 mm. Carapace .68 mm long, .84 mm wide. Tibiae I-IV .60, .65, .60, .67 mm long, respectively. Metatarsi I-IV .48, .62, .62, .71 mm long, respectively. Male from St. Augustine, Trinidad, with structure essentially as in female, except as follows: Abdomen smaller in comparison to carapace, carapace higher in eye area and sloping more sharply at posterior de- clivity. Eye area wider in proportion to length, AME slightly larger. Calamistrum lacking on metatarsus IV. Coloration some- what paler than in female. Palpus as in Figures 52, 53. Total length, .92 mm. Cara- pace .65 mm wide, .77 mm long. Tibiae I-IV .53, .60, .55, .57 mm long, respectively. Metatarsi I-IV .45, .57, .54, .64 mm long, respectively. Ecological Notes. This species is very common both inside and outside of build- ings in the West Indies, and is the only species occurring there. The types of O. o beat us • putus a ci vi t as Map 3. Mexico and Southern United States, showing dis- tribution of O. beatus, O. putus, and O. civitas. audanti and O. benneri were both taken in buildings. Bryant (1940) found this spider on garden foliage in Cuba; Dr. A. M. Chickering has collected literally hundreds of individuals from buildings. Mature males and females occur in the West Indies at all times of the year. The Florida specimens are all from areas of good tree cover, and only one was found on the outside of a building; the others were collected from under boards on the ground and under flakes of bark. Distribution. Islands of the Caribbean, Peninsular Florida, coastal Mexico, Central America, Venezuela, and Columbia (Map 2). Oecob/'us beatus Gertsch and Davis Figures 20, 33, 54, 55; Map 3 Oecobius beatus Gertsch and Davis, 1937, Amer. Mus. Novitates, 961: 2, figs. 1, 2, $; female holotype from Acapulco, Guerrero, in American Museum of Natural History, examined. Bryant, 1948, Psyche, 55: 57 (in part). Roewer, 1954, Katalog der Araneae, 2: 1289. Bonnet, 1958, Bibliographia Araneorum, 2: 1289. Note. The "allotype male" so designated by Bryant ( 1948 ) in the Museum of Com- parative Zoology is a male of O. concinnus. 144 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 The Acapulco females from a vial with the same data as the "allotype" are O. beatus. Specimens from Tamaulipas in the Ameri- can Museum included true O. beatus, both males and females, and descriptions given here are based on those specimens. Diagnosis. Distinguished by its small size and color from all other sympatric species except O. bracae, from which it may be separated by reference to the figures of the genitalia. Description. Female from Rio Frio, Tamaulipas: carapace (Fig. 20) suboval, wider than long (L/W = 1/1.1), clypeal projection evenly rounded, with distinct shoulders at base of clypeus; carapace highest in eye area and sloping evenly in all directions, more sharply posteriorly, glabrous, with few scattered plumose hairs. Eye area wider than long (L/W= 1/1.6), eyes in two rows, both rows slightly pro- curved. PLE much the largest, separated by two and one-fourth times their diameter. PME irregular, opalescent, less than twice as long as wide, separated from each other by one and one-half times their greatest width, from ALE by their greatest width. AME round, dark, one-half diameter of PLE, separated from each other by two diameters, from PLE by one diameter. ALE irregular, opalescent, separated from AME by less than radius of AME. Epigy- num as in Figure 33. Remainder of struc- ture typical for the genus. Carapace yellow-white, PLE and AME ringed with black, dusky area surrounding eyes faint. Posterior part of carapace with irregular dusky band extending from just behind eye area to posterior margin, nearly discon- tinuous just anterior to thoracic groove, where it shows broad lateral extensions, "spectacle-like" light markings midway be- tween thoracic groove and posterior margin of carapace. Clypeus light. Marginal band black, with dusky extensions at clypeal shoulder, three dusky extensions contact- ing three dark spots midway between cen- tral band and marginal line. Abdomen grayish white, with chalky blotches except in cardiac area. Dark brown markings nearly covering abdomen laterally and extending ventrally, leaving central light band on venter. Spinnerets and anal tubercle light, surrounded by dark area. Legs yellowish white, with annuli as in O. concinnus, but lighter and less distinct. Total length, 1.95 mm. Carapace .67 mm long, .74 mm wide. Tibiae I-IV .51, .50, .50, .52 mm long, respectively. Metatarsi I-IV .50, .51, .50, .56 mm long, respectively. Male from Rio Frio, Tamaulipas, with structure and coloration essentially as in female, except as follows: carapace much wider in proportion to length ( L/W = 1/1.14), larger in proportion to abdomen, flatter than in female, sloping more gradu- ally from eye area. Eye area wider in proportion to length than in female (L/W = 1/2.6), eyes smaller, more dispersed. Legs proportionally somewhat longer than in female. Calamistrum absent from meta- tarsus IV. Palpus as in Figures 54, 55. Total length, 1.80 mm. Carapace .71 mm long, .(SI mm wide. Tibiae I-IV .60, .56, .55, .60 mm long, respectively. Metatarsi I-IV .60, .56, .60, .64 mm long/respectively. Ecological Notes. This species is found both on buildings and under stones. Distribution. Tamaulipas, San Luis Potosi, Guerrero (Map 3). Oecobius putus O. P.- Cambridge Figures 1,2, 17, 34, 56, 57; Map 3 Oecobius putus O. P. -Cambridge, 1876, Proc. Zool. Soc. London, p. 544-545, pi. 58, figs. la-Id, $ ; type specimens from ruined temple between Denderah and Assouan, additional material from Temple of Pbilae, Egypt. Roewer, 1954, Katalog der Araneae, 2: 1289. Bonnet, 1958, Bibliographia Araneorum, 2: 3134. Tika- der, 1962, J. Bombay Nat. Hist. Soc., 59: 683, Dorsal views of carapaces. Figs. 13-20. 13. Oecobius cellariorum. 14. O. annu//pes. 15. O. concinnus. 16. O. inferpe/ lafor. 17. O. putus. 18. O. iso/afus. 19. O. isolatoides. 20. O. beafus. Spider Family Oecobhdae • Slt< at 145 zo 15 16 8 1 mm 146 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 figs la-Id, $ 9 . Kritscher, 1966, Ann. Natur- hist. Mus. Wien., 69: 290-291, pi. 1, fig. 11, pi. 2, fig. 12, $. Ambika putus, Lehtinen, 1967, Ann. Zool. Fenn., 4: 433, fig. 32, 2. Diagnosis. Distinguished from all other Nearctic and Neotropical species of Oeco- bius by its large size, having the AME much the largest, and its lack of pattern. Description. Female from Yuma, Arizona: carapace (Fig. 17) oval, wider than long (L/W= 1/1.2), with distinct shoulders on margin opposite eye region, clypeal pro- jection evenly rounded. Carapace highest in eye area, sloping gradually and evenly in all directions. Thoracic groove trans- verse, indistinct. Eye area wider than long (L/W = 1/1.6), eyes in two transverse, slightly procurved lines. AME much the largest, separated by one diameter, ALE irregular, opalescent, separated from AME by their width. PLE slightly oval, one- half diameter of AME, separated from each other by nearly five times their diameter and from AME by their radius. PLE light, but not opalescent. PME irregular, sub- triangular, opalescent, separated by nearly four times their width, almost contiguous with PLE. Spinnerets, cribellum, anal tubercle surrounded by a lightly sclerotized rim which bears a circle of long, incurved plumose hairs similar to those on anal tubercle. Legs with scattered hairs and few spines, spines irregularly arranged, but always present on ventrodistal portion of tibiae and metatarsi as pair of prolateral articular spines. Epigynum as in Figure 34. Remainder of structure as described for genus. Carapace off-white, unmarked except for black area around AME and PLE. Dusky marginal line faintly indi- cated, vague gray area around thoracic groove. Remainder of prosoma and legs off-white, unmarked. Abdomen pale gray, blotched chalky white. Venter pale gray, unmarked. Spinnerets and anal tubercle off-white. Total length, 3.15 mm. Cara- pace .94 mm long, 1.11 mm wide. Tibiae I-IV .91, 1.01, 1.01, 1.06 mm long, re- spectively. Metatarsi I-IV .86, 1.03, 1.12, 1.20 mm long, respectively. Male from Yuma, Arizona, with structure essentially as in female, except as follows: carapace more nearly circular ( L/W = 1/1.08), legs longer, thinner in proportion to body size. Calamistrum absent, cri- bellum suppressed. Body much more heavily clothed with plumose hairs, those on ridge around spinnerets much less obvious. Palpus as in Figures 56, 57. Color- ation similar to that of female. Total length, 2.63 mm. Carapace 1.05 mm long, 1.13 mm wide. Tibiae I-IV .85, .86, .89, .97 mm long, respectively. Metatarsi I-IV .98, 1.00, .99, 1.10 mm long, respectively. Ecological Notes. Mature males and fe- males of this species were taken in January and February in Baja California, in April in Arizona and in May in Big Bend National Park, Texas. Mature specimens taken from June to August are represented from all these localities. At Tucson, Ari- zona, the species is common on the outsides of University buildings, with small webs stretched over "superficial inequalities," as Cambridge (1876) elegantly stated in the original description. Debski (1922) pub- lished more detailed reports on the ecology of this species and of O. templi. Only one collection places this spider away from buildings: a female was taken in a packrat nest near Sufford, Arizona. Distribution. See Map 3. Southern Cali- fornia, southern Arizona, extreme westen New Mexico, southwestern Texas, northern Sonora and Chihuahua, and Baja Cali- fornia. Because of its synanthropic habits, this species is spread by man, and occurs throughout Africa ( Paris Museum speci- mens fom Tanganyika examined by Le- doux) and India (Tikader, 1962). Oecobius interpellator new species Figures 16, 36, 58, 59; PI. 1, Figures 1, 2 Holotijpe. Female from Cambridge, Massachusetts, Biological Laboratories, Harvard University, 20 February 1963, collected by J. A. Beatty, in the Museum Spider Family Oecobhdae • S. 147 »• • »■ • ' •'.-•■ Plate I. Fig female later 1. Oecobius interpellator female, dorsal view. Fig. 2. O. interpellator female and eggs in old web. The left the old web and built a new one in another part of the container. of Comparative Zoology. Paratype male from the same locality in the Museum of Comparative Zoology. The specific epithet is a Latin noun in apposition, meaning in- vader, foreigner, or stranger. Diagnosis. As a result of its limited dis- tribution in the Nearctic region, this species is sympatric only with O. annulipes, to which it bears superficial resemblance. However, the epigynum of O. interpellator lacks a scape, and the palpus (Fig. 58) is distinct from that of O. annulipes in having a much larger conductor. Description. Female from Cambridge, Massachusetts: carapace (Fig. 16) wider than long (L/W = 1/1.4), suboval, clyp- eal projection evenly rounded, almost squared off when seen from above. Eye area highest portion of carapace, carapace sloping evenly in all directions, slightly more abruptly to clypeus. Eyes in two nearly straight rows, PLE the largest, separated by slightly more than one diam- eter; PME irregular, opalescent, nearly contiguous with PLE, separated from each other by slightly less than their greatest width; PME round, dark, three-fourths as large as PLE and separated from PLE by a radius, from each other by one and one- fourth diameters; PLE oval, opalescent, nearly as large as PME, almost contiguous with PME. Epigynum (Fig. 36) of the O. concinnus-O. beatus pattern; copula- tory openings anterior, separated by raised, wrinkled ridge resembling scape; ridge traversed medially by fertilization tubes, which turn abruptly laterad near posterior end of ridge, then curve back mesad and ventrad to open on either side of an oval depressed region limited anteriorly by blunt end of wrinkled ridge and posteriorly by transverse sclerotized rim. Internal geni- talia fitting the O. concinnus pattern. Re- mainder of structure as descibed for genus. Coloration very similar to O. annulipes; carapace pale off-white with dusky mar- ginal line; eyes ringed with black; vaguely indicated dusky area behind eyes extending in darker specimens to posterior margin o\ carapace, most specimens marked as shown in Figure 16; three submarginal dusky spots on each side. Abdomen unmarked with dusky areas, yellowish gray blotched chalky white. Sternum, coxae of legs and venter of abdomen off-white. Legs un- marked. The general impression is one of an intermediate between O. annulipes and O. eella riorum. Total length, 2.13 mm. Carapace .77 mm long, .88 mm wide. Tibiae I-IV .75, .71, .68, .79 mm long, 148 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 respectively. Metatarsi I-IV .68, .64, .65, .73 mm long, respectively. Male from Cambridge, Massachusetts, with structure essentially as in female, except as follows: carapace somewhat rounder (L/W = 1/1.08); eye area wider in proportion to length than in female (L/W =1/1.8); eyes smaller. Abdomen much smaller in proportion to carapace than in female. Calamistrum lacking from metatarsus IV. Palpal organ (Figs. 58, 59) of moderate size; radical apophysis absent; radix slightly produced at posterior lateral edge. Remainder of palpal structure close to O. concinnus. Coloration similar to fe- male, but paler in most cases. Total length, 1.29 mm. Carapace .76 mm long, .82 mm wide. Tibiae I-IV .65, .71, .67, .73 mm long, respectively. Metatarsi I-IV .60, .65, .64, .72 mm long, respectively. Ecological Notes. The mode of life of this species is very similar to that of O. annutipes. Despite extensive remodelling of the rooms in which it was originally collected, O. interpellator is still (Septem- ber, 1968) remarkably common in the two rooms composing the insectary, and has spread to several adjacent rooms. All sizes except mature males were observed 20 September 1968, including six egg masses ( PI. 1, Fig. 2) left in situ, all containing four eggs. The collection from which the types were taken contained nearly forty mature males and females ( February, 1963 ) . Webs ( PI. 1, Fig. 1 ) cover virtually every corner and depression between cinder blocks, and are about a centimeter square, of the tubu- lar type described by Glatz ( 1967). Debris in the webs consisted of tiny cockroach nymphs and small ants, all of which had been sucked dry without obvious damage to the cuticle. When prodded from their webs, the spiders run with great agility over the walls, but seem unable to cling to smooth, clean surfaces. Distribution. Known only from the type locality. Presumably this species was im- ported with tropical cockroaches or ants. A sketch by Ledoux of a female in the • isolatoides o isolatus Map 4. Southwestern North America, showing distribution of O. isolatus and O. isolatoides. British Museum (Natural History) from Noumea, New Caledonia, closely resembles this species. Oecobius isolatus Chamberlin Figures 1 8, 35, 60, 61 ; Map 4 Oecobius isolatus Chamberlin, 1924, Proc. Cali- fornia Acad. Sci., 12: 584; immature holotype from Carmen Island, Gulf of California, in American Museum of Natural History, ex- amined. Chamberlin and Ivie, 1935, Ann. Entomol. Soc. Amer, 28: 270, pi. 1, fig. 2, 9. Roewer, 1954, Katalog der Araneae, 2: 1290. Bonnet, 1958, Bibliographia Araneorum, 2: 3133. Oecobius parvus Chamberlin and Ivie, 1942, Bull. Univ. Utah, Biol. Ser., 32; 13; immature holo- type from Virgin Narrows, near Littlefield, Arizona, in American Museum of Natural His- tory, examined. Roewer, 1954, Katalog der Araneae, 2: 1290. NEW SYNONYMY. Note. Both O. isolatus and O. parvus were described from immature specimens, a single example in the case of O. parvus Spider Family Olx.obiidae • Shear 149 and two specimens of O. isolatus. Para- types and holotypes of both were studied, and no structural differences in eyes and other features could be discerned. Only color and pattern differed, O. parvus re- sembling a very light O. isolatus. The northernmost record of O. isolatus is in southwestern Yuma Countv, Arizona, and the southernmost for O. parvus is just over the Arizona-Utah border. But even in the absence of intermediate records, I am inclined to synonymize O. parvus under O. isolatus, since the only mature female available from the range of O. parvus has an epigynum identical to O. isolatus. Also, there is a distinct tendency for specimens of O. isolatus from the Baja peninsula to be somewhat darker than those from Ari- zona and California. If this trend is extrap- olated, a light, O. parvus-\ike O. isolatus would be expected to occur in southern Utah. Additional collecting throughout extreme western Arizona is much to be desired. Diagnosis. This species belongs to what might be termed the "Mexican group," a series of similar species (dark carapaces, broad epigynal scapes, palpi with large stipal and/or radical apophyses) obviously native to the region and not synanthropic. These species are best separated by refer- ence to the figures, though O. isolatus is not sympatric with any of the others ( Map 4). It is most closely related to O. isolatoi- des, but the palpus differs (Fig. 60), and the epigynal scape is not narrowed dis- tally. Description. Female from Fortuna Mine, Yuma County, Arizona: carapace (Fig. IS) subcircular, clypeal projection evenly rounded, more than one-fourth length of carapace, distinct shoulders at base of clypeal projection. Carapace nearly gla- brous, with scattered plumose hairs. Eyes on distinct tubercle, sloping evenly to clypeus, but abruptly behind, then sloping gradually to posterior declivity. Posterior declivity sharp, limiting distinctly lower, lunate posterior area. Eyes in two rows, both moderately procurved. PLE the largest, separated from each other by two and one-half diameters; PME irregular, opalescent, separated from PLE by their width and from each other by their length. AME three-fourths diameter of PLE, separated from each other by one diameter and from PLE by one-half diameter of PLE. ALE irregular, opalescent, as large as AME, nearly contiguous with AME. Remainder of structure essentially as de- scribed for the genus. Epigynum ( Fig. 35 ) large, with a broad scape, tracks of fertili- zation duets visible in scape. Carapace entirely dark brown, eyes ringed with black, marginal line black, three lighter spots on each side just above marginal line. Coxae of legs, endites, labium pale yellow, sternum light brown. Abdomen blotched chalky dorsally, dark brown pattern near that described for O. annulipes. Venter pale, spinnerets and anal tubercle dark brown, surrounded by dark brown area. Legs with lateral maculae as follows: two maculae mesal and distal on all segments except patella and tarsi, patella wholly dark laterally, tarsi dark distally. Distal lateral maculae on tibiae and metatarsi forming complete or nearly complete rings. Total length, 2.40 mm. Carapace .87 mm long, .85 mm wide. Tibiae I-IV .68, .72, .65, .67 mm long, respectively. Metatarsi I-IV .65, .67, .68, .78 mm long, respectively. Male from Roosevelt Dam, Yuma County, Arizona, with structure essentially as in female, except as follows: carapace flatter, broader in relation to width, much larger in comparison with abdomen. Eye area nearly twice as wide as long, eyes more dispersed. Coloration paler, but of the same pattern. Palpus (Figs. 60, 61) with a falcate apophysis on stipes and large, weakly sclerotized radical apophysis. Total length, 2.17 mm. Carapace .86 mm long, .98 mm wide. Tibiae I-IV .67, .68, .68, .76 mm long, respectively. Metatarsi I-IV .68, .74, .75, .82 mm long, respectively. Ecological Notes. Nothing is known of this species' ecology. 150 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 Distribution. (Map 4) Lower Colorado Basin, Baja California. Oecobius isolatoides new species Figures 19, 37, 62, 63; Map 4 Holotype. Female from El Coyote, Sonora, 17.7 miles (28.3 kilometers) east of Rio Bavisbe and 48 miles (76.8 kilo- meters) east of Moctezuma, elev. 3200', 12 July 1960, J. A. Beatty; paratype male from the same locality, in the American Museum of Natural History. Diagnosis. The scape of the epigynum is nearly twice as long as wide, narrowed distally, and not folded; this combination of characters occurs in no other species of the "Mexican group." The male palpus resembles that of O. isolatus, but the radical apophysis of O. isolatoides is wider in proportion to its length, and not knobbed at the tip. Description. Female from El Coyote, Sonora: carapace (Fig. 19) oval, wider than long (L/W = 1/1.1), clypeal pro- jection evenly tapering from lateral mar- gins, bluntly pointed. Carapace highest in eye area, sloping gently in all directions. Thoracic groove indistinct. Eye area wider than long (L/W = 1/1.7), eyes in two transverse, distinctly procurved rows. PLE the largest, separated by slightly more than two and one-half diameters. AME little more than half the diameter of PLE, separated from each other by one diameter and from PLE by a radius. PME opales- cent, irregular-triangular, separated from each other by one and one-fourth their greatest width and from PLE by half their greatest width. ALE oval, opalescent, nearly as large as AME, separated from AME by their width. Epigynum (Fig. 37) with prominent scape, tapering to less than half its basal width, edges sigmoid. Fertilization tubes diverging at midlength, converging slightly near their openings. Remainder of structure as described for genus. Carapace nearly even medium tan, margins darker, submarginal dark shadings obsolete in some specimens. Eyes ringed with black. Remainder of coloration as described for O. isolatus. Total length, 2.04 mm. Carapace .77 mm long, .86 mm wide. Tibiae I-IV .56, .64, .59, .67 mm long, respectively. Metatarsi I-IV .56, .60, .60, .71 mm long, respectively. Male from El Coyote, Sonora, with struc- ture essentially as in female, except as follows: carapace slightly flatter, wider in proportion to length (L/W= 1/1.12), eyes slightly smaller, more dispersed. Abdomen smaller in proportion to carapace, body generally more densely hairy. Sternum margined with specialized hairs. Coloration as described for female, but lighter in some specimens. Palpus (Figs. 62, 63) large, bulky. Radical apophysis blunt, gnarled, almost meeting process of stipes, which is smoothly curved and truncate. Conductor complex, intromittent portion of embolus concealed, visible only in direct ventral view as a short, heavily sclerotized curved spine. Total length, 1.82 mm. Carapace .75 mm long, .83 mm wide. Tibiae I-IV .56, .60, .57, .64 mm long, respectively. Metatarsi I-IV .53, .60, .59, .63 long, re- spectively. Ecological Remarks. The Sonoran and Sinaloan localities are in desert thorn scrub. The type species were found under stones on a dry hillside. In Arizona, the species occurs in mountains, probably in the transition between Sonoran desert and woodland. Mature individuals have been collected in April, July, and August. Distribution. (Map 4) Mountain foot- hills in Sonora, northern Sinaloa, and southern Arizona. Oecobius p/axf/a new species Figures 21 , 38, 64, 65 Holotype. Female from 20 miles (32 kilometers) west of Piaxtla, Sinaloa, 2 August 1964, W. J. Gertseh and J. Woods, collectors, in the American Museum of Natural History; male paratype from the same locality in the same museum. The specific name is a noun in apposition, after the type locality. Spider Family Okcobiidae • Shear 151 Dorsal views of carapaces. Figs. 21-27. 21. Oecobius piaxtla. 22. O. rivula. 23. O. juangarcia. civitas. 26. O. culiacanensn. 27. Platoecobius lloridanus. 24. O. bracae. 25. O. 152 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 Diagnosis. Distinct from others of the "Mexican group" in details of the genitalia. The epigynal scape (Fig. 38) is as wide as long, and narrowed distally; the only species with a similar pattern is O. civitas, in which the scape is not narrowed distally. The male palpus (Figs. 64, 65) is some- what smaller than the others bearing both stipal and radical apophyses. Description. Female from 20 miles (32 kilometers) west of Piaxtla, Sinaloa: carapace ( Fig. 21 ) subcircular ( L/W = 1 1.12), clypeal projection low and rounded, shoulders at base of clypeal projection distinct. Carapace highest in eye area, sloping evenly in all directions; thoracic groove transverse, poorly indicated. Eyes in two slightly procurved rows; eye area nearly half again as wide as long ( L/W = 1/1.35). PLE the largest, separated by nearly two diameters; PME irregular, opalescent, separated from each other by their greatest width, nearly contiguous with PLE. AME two-thirds diameter of PLE, separated from each other by slightly more than one radius, nearly contiguous with PLE; ALE irregular, opalescent, nearly contiguous with AME. Epigynum (Fig. 38) with scape as wide or slightly wider than long, traversed by fertilization ducts. In- ternal genitalia of the O. civitas pattern, but copulatory ducts open almost immediately into spermathecae. Remainder of structure as described for genus. Carapace with light brown ground color, overlaid with dark brown pattern as in Figure 21. Eyes ringed with black, eye tubercle dark brown to black; clypeus light mesally, with two parallel dark bands laterally; marginal line black, with dark brown areas extending into submarginal region as figured; dark brown band extending from eye tubercle to posterior margin of carapace. Labium light brown, sternum light brown, mar- gined dark brown. Leg annulations of the typical pattern; all incomplete. Total length, 2.82 mm. Carapace .77 mm long, .83 mm wide. Tibiae I-IV .48, .49, .48, .56 mm long, respectively. Metatarsi I-IV .45, .49, .47, .54 mm long, respectively. Male from 20 miles (32 kilometers) west of Piaxtla, Sinaloa, with structure essentially as in female, except as follows: carapace wider in proportion to length, eyes smaller, more dispersed than in female. Carapace with two lunate submarginal depressions, presumably serving for muscle attachment for palpus. Palpus (Figs. 64, 65) large, tegular apophysis complex, nearly trifid; conductor broadly spatulate, embolus with all the usual processes well developed and large. Coloration as in female, somewhat lighter. Total length, 2.00 mm. Carapace .80 mm long, .90 mm wide. Tibiae I-IV .53, .59, .56, .62 mm long, respectively. Metatarsi I-IV .51, .56, .58, .64 mm long, respectively. Ecological Notes. Found under rocks in desert thorn scrub. Distribution. Known only from the type locality. Oecobius rivula new species Figures 7, 8, 22, 39, 68, 69 Holotype. Female from 32 miles (51.2 kilometers) east of Villa Union, Sinaloa, 26 August 1965, collected by W. J. Gertsch and J. Hastings; paratype male from the same locality, in the American Museum of Natural History. The specific epithet is a noun in apposition, Latin for crevice. Diagnosis. Easily distinguished from all others of the "Mexican group" by the folded epigynal scape with a perforation at the tip (Fig. 39), and by the extraor- dinarily massive male palpus with an elongate cluster of small knobs on the inner side of the radical apophvsis (Fig. 69). Description. Female from 32 miles (51.2 kilometers) east of Villa Union, Sinaloa: carapace (Fig. 22) suboval, wider than long (L/W = 1/1.16), clypeal projection evenly rounded, shoulders lacking at base of clypeus. Carapace highest in eye region, sloping gradually in all directions, some- what more steeply to clypeus. Eye area Spider Family Oecobiidae • Shear 153 28 31 29 33 36 emale epigyna. Figs. 28-38. 28. Oecobius cellanorum, ventral view. 29. O. annulipes, ventral view. 30. O. annu/ipes, lorsal view. 31. O. concinnus, ventral view. 32. O. concinnus, dorsal view. 33. O. beafus, ventral view. 34. O. putus, entral view. 35. O. isolafus, ventral view. 36. O. interpellator, ventral view. 37. O. isolatoides, ventral view. 38. O. iiaxlla, ventral view. 154 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 wider than long (L/W= 1/1.88); eyes in as described for female, somewhat lighter. two slightly procurved rows. PLE much Total length, 2.34 mm. Carapace 1.05 mm the largest,' separated from each other by long, 1.14 mm wide. Tibiae I-IV .70, .75, two and one-quarter diameters; PME ir- .71, .79 mm long, respectively. Metatarsi regular, opalescent, separated from each I-IV .70, .71, .70, .80 mm long, respectively. other by one and one-quarter their greatest Ecological Notes. The specimens were width, from PLE by one-half their greatest collected at the type locality from a deep, width- AME two-thirds diameter of PLE, wet crevice. separated from each other by slightly less Distribution. This species is known only than one diameter, from PLE by less than from the type locality. a radius; ALE irregular, opalescent, nearly . ••-i A\fT? T7™n-w«,,™ /T^ct Oecobius luanqarcia new species contiguous with AME. Epigynum (rig- i » r 39) large, well-defined, lying in depression Figures 23, 40, 66, 67 that is longer than wide. Scape long, Holotype. Female from Juan Garcia, folded anteriorly, then posteriorly, most Oaxaca, 1 September 1964, collected by J. anterior portion when seen in ventral view anc] \y. Ivie; male paratypc from same broadly rounded, tapering to less than half locality, both in American Museum of its width. Fertilization ducts clearly visible Natural History. The specific name is a in scape; tip of scape with small, round noun in apposition, after the type locality, perforation. Remainder of structure typical Diagnosis. Distinct from others of the for genus. Ground color of carapace light "Mexican group" in the centrally depressed brown, eyes ringed with black; central dark epigynal scape (Fig. 40), and the less brown band forked anterior to eyes, central complex palpus (Figs. 66, 67) which lack part of clypeus light. Marginal band dark a radical apophysis. A radical apophysis brown, widened opposite coxae of legs. js lacking in O. civitas also, but the general Submarginal spots in some cases connected appearance of the palpus is quite different, to marginal band. Sternum, labium, and particularly in the conductor, venter light, as described for O. isolatus. Description. Female from Juan Garcia, Leg annulations incomplete, as described Oaxaca: carapace (Fig. 23) suboval (L/W for O. isolatus. Total length, 3.05 mm. =1/1.3), clypeal projection indistinct, Carapace .92 mm long, 1.05 mm wide. r0unded, shoulders at base of clypeal pro- Tibiae I-IV .76, .76, .75, .80 mm long, jection poorly developed. Carapace with respectively. Metatarsi I-IV .75, .75, .75, scattered plumose white hairs. Eyes on .83 mm long, respectively. low tubercle, sloping evenly in all di- Male from 32 miles (51.2 kilometers) rections to margin of carapace; thoracic east of Villa Union, Sinaloa, with structure furrow indistinct. Eyes in two slightly essentially as in female except as follows: procurved rows; eye area wider than long carapace slightly longer in proportion to (L/W = 1/1.28). PLE the largest, sepal width (L/W = 1/1.08); eyes smaller more rate(J from each other by slightly more than dispersed than in female. Carapace with one diameter; PME irregular, opalescent, two deep submarginal lunate depressions, nearly contiguous with PLE, separated possibly associated with muscle attach- from eacn other by their greatest width, ments for massive palpi. Remainder of AME two-thirds diameter of PLE, nar- structure as described for genus. Palpus rowly separated from PLE; separated from (Figs. 68, 69) unusually large, massive; each other by slightly more than one diam- radical apophysis forked apically, coarsely eter; ALE irregular, opalescent, nearly rugose on mesa] edge; hooks of conductor contiguous with AME. Epigynum (Fig. close together; stipal hook large, as long 40) as wide as long, somewhat reniform; as cymbium, spatulate distally. Coloration scape margined, perforated anteriorly with Spider Family Oecoihidai: • Shear 155 41 45 42 44 Female epigyna. Figs. 39-47. 39. Oecobius rivula, ventral view. 40. O. juangarc/a, ventral view. 41. O. bracae, ventral view. 42. O. bracae, dorsal view. 43. O. civitas, ventral view. 44. O. civitas, dorsal view. 45. O. culiacanensis, ventral view. 46. P/afoecobius floridanus, ventral view. 47. P. floridanus, dorsal view. a large fossa. Internal genitalia not ex- amined on only available female specimen, but presumably similar to O. rivula. Re- mainder of structure as described for genus. Carapace entirely medium brown, eyes ringed with black, eye area dark brown. Coxae of legs, endites, pale yellow, labium and sternum medium brown. Abdo- men blotched chalky dorsally, with typical pattern of dark brown areas; venter mostly pale. Legs annulated and blotched as described for O. isolatus. Total length, 2.56 mm. Carapace .80 mm long, .91 mm wide. Tibiae I-IV .52, .53, .51, .57 mm long, respectively. Metatarsi I-IV .50, .50, .49, .59 mm long, respectively. Male from Juan Garcia, Oaxaca, with structure essentially as in female except as follows: carapace wider in proportion to length (L/W= 1/1.22); eyes smaller, more dispersed, eye area wider in proportion to length than female (L/W = 1/1.4). Cara- 156 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 pace with two submarginal lunate de- pressed areas, presumably serving as points of muscle attachment for palpi. Palpus (Figs. 66, 67) strongly modified from basic form described for O. isolatus, distal portion of stipes greatly lengthened and curved around face of radix; intromittent portion emerging from behind conductor ventro- mesally; conductor reduced, radical apoph- ysis absent. Total length, 2.01 mm. Carapace .76 mm long, .93 mm wide. Tibiae I-IV .48, .49, .46, .52 mm long, respectively. Metatarsi I-IV .49, .51, .46, .60 mm long, respectively. Ecological Notes. Little is known of the ecology of this species. A male from near Tehuantepec, Oaxaca, was taken from a small web between the ribs of a cereoid cactus. Distribution. Central Oaxaca. Oecobius bracae new species Figures 24, 41, 42, 70, 71 Holotype. Female from three miles (4.8 kilometers) west of Tehuantepec, Oaxaca, 28 April 1963, collected by W. J. Gertsch and W. Ivie; paratype male from the same locality, both in the American Museum of Natural History. The specific epithet is a Latin noun in apposition, "a short, pants- like garment," and is derived from the unusual form of the epigynum. Diagnosis. The color pattern distinguishes this species from all others in the "Mexican group," but could lead to confusion with O. heatus. The epigynal scape is highly distinctive in both species; in O. bracae (Fig. 41) the fertilization tubes open flush with the surface, in O. beatus (Fig. 33) they open on the tips of short projections. The palpus of O. bracae (Fig. 71) has a much smaller stipes than that of O. beatus (Fig. 55). Description. Female from three miles (4.8 kilometers) west of Tehuantepec, Oaxaca: carapace (Fig. 23) nearly rounded (L/W = : 1/1.03), clypeal projection blunt, clypeal shoulders fairly distinct. Carapace evenly domed at eye area, sloping almost at an equal angle in all directions. Eyes in two rows, first row nearly straight, second perceptibly procurved. Eye area wider than long (L/W= 1/1.5). PLE the largest, separated by twice their diameter; PME irregular, opalescent, nearly con- tiguous with PLE, separated from each other by slightly less than their greatest width. AME nearly three-fourths diameter of PLE, separated from each other by a little less than their diameter and from PLE by their radius; ALE opalescent, irregular, nearly contiguous with AME. Female epigynum very distinctive, consist- ing of a deep, semi-rectangular depression in which prominent, flared openings of vulvae are situated (Fig. 41). Anteriorly, fertilization tubes open into a shallow, rimmed depression. Internally (Fig. 42), a single pair of spermathecae open almost directly to outside, fertilization tubes curve far anteriorly just under cuticle until they loop back to open as described above. Remainder of structure typical for genus. Carapace off-white, eyes ringed black which separates them into two lateral groups, eye tubercle covered by extensive dusky area interrupted by two symmetrical light areas immediately behind eyes, and extending as irregular band to posterior margin of carapace. Submarginal band dark brown, connected with brownish black marginal line just behind clypeal shoulders and at two other points evenly spaced posteriorly. Clypeus dark brown, with central white area. Abdomen pale gray, blotched chalky, cardiac mark brown. Legs banded as follows: incomplete median and distal bands on femur, tibia and metatarsus; patella with a single, distal band, incomplete dorsally. Tarsus some- what darker distally. Remainder of color- ation as described for O. beatus. Total length, 1.96 mm. Carapace .63 mm long, .65 mm wide. Tibiae I-IV .38, .38, .40, .48 mm long, respectively. Metatarsi I-IV .39, .42, .41, .50 mm long, respectively. Male from three miles (4.8 kilometers) west of Tehuantepec, Oaxaca, with struc- Spider Family Oecobiidae • Shear 157 hire essentially as in female except as fol- lows: carapace somewhat wider in relation to length (L/W = 1/1.1), eye group more compact (L/W = 1/.96). Legs longer in proportion to body, body more densely hairy. Coloration as in female, carapace of some specimens more lightly marked. Papal organ (Figs. 66, 67) of the moderately large type, as figured. Total length not known, abdomen missing on all mature male specimens. Carapace .55 mm long, .61 mm wide. Tibiae I-IV .34, .35, .33, .43 mm long, respectively. Metatarsi I-IV .36, 37, .37, .45 mm long, respectively. Ecological Notes. Nothing is known of the ecology of this species. Distribution. Central Oaxaca. Oecobius civitas new species Figures 25, 43, 44, 72, 73; Map 3 Holotype. Female from west side of Lake Sayula, Jalisco, 30 July 1964, collected by W. J. Gertsch and J. Woods; paratype male from the same locality, both in the American Museum of Natural History. The specific epithet is a Latin noun in appo- sition, "community," and is derived from the evident communal existence of the species. Diagnosis. Resembles O. culiacanensis in having the epigynal scape as wide as long, but the scape of O. civitas is not narrowed distally (Fig. 43). The palpus is distinct from all others of the "Mexican group," lacking both a radical apophysis and a large stipal process (Figs. 72, 73). Description. Female from Lake Sayula, Jalisco: carapace (Fig. 28) suboval, wider than long (L/W =1/1.2), clypeal pro- jection evenly rounded, shoulders at base of clypeus indistinct. Carapace highest in eye region, sloping gradually behind and abruptly to clypeus, fairly heavily clothed with white plumose hairs, otherwise gla- brous. Eye area wider than long, ( L/W = 1/1.7), eyes in two slightly procurved rows. PLE much the largest, separated by two and one-half times their diameter, round, dark. PME irregular, opalescent, twice as long as wide, separated from each other by their greatest width and from PLE by two- thirds their greatest width. AME round, dark, nearly one-half diameter of PLE, separated from each other by twice their diameter, from PLE by one diameter. ALE irregular, opalescent, separated from AME by slightly more than one radius of AME. Epigynum as in Figures 43, 44. Remainder of structure as described for genus. Cara- pace with ground color of dark grayish tan, irregularly mottled dark brown, eyes ringed with black. Clypeus dark brown, margins of carapace dark brown with three lighter areas opposite legs I, II, and III. Central band of carapace grayish tan, procurved darker arc at thoracic groove, two recurved darker arcs behind, the last surrounding lighter "spectacle-like" area. Sternum, labium, coxae of legs tan ventrally. Abdo- men dark brown ventrally, laterally, and in cardiac area with two light tan stripes paralleling cardiac area, large light tan macula over anal tubercle. Spinnerets and anal tubercle brown. Legs tan, all with a similar pattern of maculae and bands: coxae banded dark brown distally, femora with median band incomplete dorsoposte- riorly, distal annulus complete; patellae dark brown proximally and dorsally; median annulus on tibiae complete, distal annulus complete except for narrow pos- terior band; median and distal annuli of metatarsi complete; tarsi shaded dark brown distally. Total length, 1.97 mm. Carapace .82 mm long, .98 mm wide. Tibiae I-IV .62, .69, .66, .75 mm long, respectively. Metatarsi I-IV .60, .66, .72, .80 mm long, respectively. Male from Lake Sayula, Jalisco, with structure essentially as in female, except as follows: carapace wider in proportion to length (L/W =1/1.3), eyes somewhat more dispersed, carapace more heavily clothed with procumbent white plumose hairs, abdomen smaller in relation to cara- pace. Calamistrum absent from leg IV. Coloration similar to female, somewhat lighter. Palpus as in Figures 72, 73. Total 158 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 55 H Male right palpi. Figs. 48-57. 48. Oecob/us cellariorum, lateral view. 49. O. cellariorum, mesal view. 50. O. annu/ipes, lateral view. 51. O. annulipes, mesal view. 52. O. concinnus, lateral view. 53. O. conc/nnus, mesal view. 54. O. beafus, lateral view. 55. O. beatus, mesal view. 56. O. putus, lateral view. 57. O. putus, mesal view. Spider Family Oecobhdae • Shear 159 length, 1.72 mm. Carapace .72 mm long, .95 mm wide. Tibiae I-IV .60, .66, .66, .75 mm long, respectively. Metatarsi I-IY .60, .68, .68, .76 mm long, respectively. Ecological Notes. Communities of these medium-sized oecobiids weave large webs on the undersides of rocks on the shores of Lake Sayula, Jalisco. Intact whole webs were collected in July, 1964, by Willis J. Gertsch and J. Woods and deposited in the American Museum of Natural History. Each web contained between 60 and 140 .individuals of approximately the same age, in most eases in the penultimate instar, with a few7 adults. Each individual seemed to have its own pocket or retreat in the communal web. Large egg sacs were found near the centers of two of the webs, each containing nearly 200 deutova clearly identifiable as oecobiids; in both these cases, most of the previous (?) generation were in the penultimate instar. Numerous insect parts were found tangled in the webs, mostly from ants, small beetles, and from thysanurans. The vials containing [the webs also contained numerous spiders of the families Pholcidae, Filistatidae, Ganphosidae, Oonopidae, and Scytotidae. The relationship of these spiders to the oecobiids is not known; perhaps they were merely collected at the same time. Distribution. (Map 3) Jalisco, Morelos, Guerrero, Oaxaca. Decobius culiacanensis new species Figures 26, 45, 74, 75 Holotype. Female from six miles (9.6 kilometers) south of Culiacan, Sinaloa, 22 July 1954, collected by W. J. Gertsch; male paratype from the same locality, both in the American Museum of Natural History. The specific epithet is formed from the name of the type locality: "of Culiacan." Diagnosis. Females may be distinguished from other members of the "Mexican group" by the epigynal scape, which is longer than wide, not folded, and bears two distal notches (Fig. 45). The stipal process of the male palpus is more spatu- late than in any other species (Figs. 74, 75). Description. Female from six miles (9.6 kilometers) south of Culiacan, Sinaloa: carapace (Fig. 26) wider than long (L/W = 1/1.12), suboval, clypeal projection slightly pointed. Eye area highest region of carapace, sloping evenly in all directions, slightly more gradually posteriorly. Eye area wider than long (L/W= -74/1), eyes in two slightly procurved rows. PLE dark, the largest, separated from each other by one and one-fourth their diameter; PME irregular, opalescent, nearly contiguous with PLE, separated from each other by their greatest width; AME dark, slightly more than one-half diameter of PLE, separated from PLE by a radius, from each other by one and one-half diameters; ALE opalescent, irregular, nearly contiguous with AME. Epigynum (Fig. 45) of the O. civitas type, with a broad oval scape traversed mesally by fertilization ducts, posterior margin deeply notched; fertili- zation ducts opening into notches. Internal genitalia typical. Remainder of structure as described for genus. Carapace evenly dusky brown-black, eyes ringed with black, submarginal markings in three series, slightly darker. Clypeus with two pale spots. Abdomen marked with dark blotches and chalky patches as described for O. isolatus. Sternum pale, margined heavily posteriorly with brown. Remainder of color as described for O. civitas, but leg bands on tibiae and metatarsi in the form of complete annuli. Total length, 2.20 mm. Carapace .66 mm long, .76 mm wide. Tibiae I-IV .55, .60, .55, .64 mm long, respectively. Metatarsi I-IV .54, .58, .55, .67 mm long, respectively. Male from six miles (9.6 kilometers) south of Culiacan, Sinaloa, with structure essentially as in female, except as follows: carapace with two submarginal crescentic depressions running from eye area to thoracic furrow, presumably acting to pro- vide muscle attachments for the large palpi. Palpus (Figs. 74, 75) massive, radical 160 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 Male right palpi. Figs. 58-67. 58. Oecobius interpellator, lateral view. 59. O. interpellator, mesal view. 60. O. isolatus, lateral view. 61. O. isolatus, mesal view. 62. O. holafoides, lateral view. 63. O. isolatoides, mesal view. 64. O. piaxtla, lateral view. 65. O. piaxtla, mesal view. 66. O. juangarcia, lateral view. 67. O. juangarcia, mesal view. Spider Family Oecobiidae • Shear 161 apophysis blunt, stipes large, cupped, terminating in spatulate point. Conductor complex. Coloration as in female, some- what lighter. Total length 1.86 mm. Cara- pace .82 mm wide, .72 mm long. Tibiae I-IV .47, .50, .50, .53 mm long, respectively. Metatarsi I-IV .45, .50, .46, .59 mm long, respectively. Ecological Notes. All the known locali- ties for this species are in subtropical thorn scrub. Distribution. Southern and western Sinaloa. Genus Platoecobius Chamberlin and Ivie, 1935 Platoecobius Chamberlin and Ivie, 1935, Ann. Entomol. Soc. Amer., 28: 270; type species by original designation and monotypy Platoe- cobius floridanus (Banks); Lehtinen, 1967, Ann. Zool. Fenn., 4: 259, 304. Diagnosis. Separated from Oecobius, the only other genus in the family, by the shorter, stouter legs, the calamistrum ex- tending the entire length of metatarsus IV, and the simpler genitalia. Description. Small oecobiid spiders (2-3 mm total length), with the characters of the family. Carapace with sides slightly sinuate, clypeus broad, evenly rounded. Carapace flat, head region slightly ele- vated. Eyes on low tubercle, eye area much wider than long. AME round, dark; PLE round, dark, but lighter in color than AME; PLE and ALE light and irregular in shape. Macrosetae of eye area as follows: one immediately behind each AME and PLE; two behind eye area at posterior declivity, separated by more than the diameter of AME. Thoracic groove very indistinct, transverse. Carapace slightly depressed above pedicel. Abdomen flat- tened, almost elongate-pentagonal. Legs stout, spines entirely lacking. Tarsi and metatarsi without trichobothria. Hind coxae separated by sternum. Calamistrum lack- ing in mature males, extending in immature specimens and females along entire length of metatarsus IV. Palpi of mature males with tegulum lacking, radix spiral, making more than one turn, conductor small, mem- branous. Females without elaborate epigy- nal modification, two pair of spermathecae, one heavily sclerotized, the other saclike. Platoecobius floridanus (Banks) Figures 27, 46, 47, 76, 77; Map 2 Thalamia floridanus Banks, 1893, Trans. Amer. Entomol. Soc., 23: 58; holotype ? from Lake Worth, Florida, in American Museum of Natural History, examined. Platoecobius floridanus, Chamberlin and Ivie, 1935, Ann. Entomol. Soc. Amer., 28: 270, pi. 4, figs. 22-32, 9. Roewer, 1954, Katalog der Araneae, 2: 1290; Bonnet, 1958, Bibliographia Araneorum, 2: 3699. Description. Female from Highlands Hammock State Park, Florida: carapace (Fig. 27) suboval, slightly wider than long (L/W = 1/1.08), widest just posterior to eyes, edge slightly undulate, clypeal projection broad, evenly rounded. Cara- pace highest in eye area, elevation pro- longed somewhat behind eyes, sloping evenly in all directions. Thoracic groove very indistinct. Eye area wider than long (L/W = 1/2.25), eyes in two rows, the first slightly procurved, the second straight. AME the largest, separated by one diam- eter, dark. ALE opalescent, subequal in size to AME, slightly irregular, nearly con- tiguous with AME, their posterior edges lying on a line through centers of AME. PLE light, but not opalescent, separated from each other by nearly six times their diameter, nearly contiguous with AME. PME nearly one-half the diameter of PLE, slightly irregular, opalescent, nearly con- tiguous with PLE. The general impression is one of two well-separated tetrads of eyes. Legs proportionally stouter than in any Oecobius species. Epigynum as in Figures 46, 47. Remainder of structure as described for the genus. Ground color medium tan. Carapace with darker mar- ginal bands, eyes surrounded by black, clypeal projection with two parallel light stripes set off by darker lines on each side. 162 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 Male right palpi. Figs. 68-77. 68. Oecobius rivula, lateral view. 69. O. rivu/a, mesal view. 70. O. bracae, lateral view. 71. O. bracae, mesal view. 72. O. civ/fas, lateral view. 73. O. c/Vlfas, mesal view. 74. O. cu/iacanens/s, lateral view. 75. O. cu/iacanensis, mesal view. 76. P/atoecobius lloridanus, lateral view. 77. Platoecobws Horidanus, mesal view. Spider Family Oecobhdae • Shear 163 Thoracic groove slightly darker. Abdomen pale dorsally, dark brown on sides. Venter nearly white, spinnerets and anal tubercle dark brown dorsally. Sternum, coxae of legs, labium nearly white. Legs irregularly blotched darker, blotching very indistinct on most specimens. Total length, 2.40 mm. Carapace .82 mm long, .88 mm wide. Tibiae I-IV .52, .52, .48, .54 mm long, respectively. Metatarsi I-IV .38, .44, .38, .49 mm long, respectively. Male from Highlands Hammock, Florida, with structure essentially as in female, ex- cept as follows: carapace somewhat wider in proportion to length (L/W = 1/1.1), eye area not so wide in proportion to its length (L/W= 1/1.9). Proportions of eyes as in female. Calamistrum absent, cri- bellum suppressed. Body much more heavily clothed with hairs. Palpus as in Figures 76, 77. Coloration similar to fe- male, somewhat lighter. Total length, 2.02 mm. Carapace .75 mm long, .82 mm wide. Tibiae I-IV .49, .49, .45, .47 mm long, respectively. Metatarsi I-IV .36, .38, .38, .44 mm long, respectively. Ecological Notes. Chamberlin and Ivie (1935) record this species as being found beneath the bark of trees, and note that no prey-catching web has been observed. They did not remark on retreats that may have been built by the spider. It is possible that the enlarged, forward-directed median eyes are an adaptation to a vagrant, hunt- ing existence. Mature representatives of both sexes have been taken in December, February, and August. Distribution. (Map 2) Central Florida, inland Georgia, and South Carolina. LITERATURE CITED I Banks, X. 1931. Expedition to L conducted by Nyt Mag. N 270-273. on net, P. 1957 Toulouse, 2 ( 4 Bryant, E. 1940. of Comparativ Zool., Harvard. The Norwegian Zoological the Galapagos Islands, 1925, Alf Wollebaek. I. Arachnida. aturvidensk (Oslo), 68(22): 1959. Biliographia Araneorum. ): 3130-3135, 3699. Cuban spiders in the Museum e Zoology. Bull. Mus. Comp. 86(7): 247-533. . 1948. Some spiders from Acap Mexico. Psyche, 55(2): 5.5-77. Comstock, J. 1940. The Spider Book. Rex, and edited by W. J. Gertsch. Comstock Publ. Co., New York: 729 pp. Chamberlin-, R., and W. Ivie. 1935. Nearctic spiders of the family Urocteidae. Ann. Ento- mol. Soc. Amer., 28: 265-270. Debski, B. 1922. Quelques observations sur les moeurs de YOecobius templi Cambridge 1876, retrouve a Helouan (Arachnida). Bull. Soc. Entomol. d'Egypte, seance du 13 Decembre 1922: 121-126. Glatz, L. 1967. Zur Biologic und Morphologie von Oecobius annulipes Lucas (Araneae, Oecobiidae). Z. Morphol. Tiere, 61: 185- 214. Kritscher, E. 1966. Die paliiarktischen Arten der Gattung Oecobius (Aran., Oecobiidae). Ann. Naturhist. Hofmus. Wien., 69: 285- 295. Kulczynski, W. 1909. Fragmenta arachnologica XIII: Araneorum et opilonum species aliquot novae. Bull. Internat. Acad. Sci. Cracovie, 1909(2): 447-472. Lehtinen, P. 1967. Classification of the cribellate spiders and some allied families, with notes on the evolution of the suborder Araneo- morpha. Ann. Zool. Fenn. 4: 199-468. Mayr, E., E. G. Linsley, and R. L. Usinger. 1953. Methods and Principles of Systematic Zoology. New York: McGraw-Hill, 231 pp. Mayr, E. 1963. Animal Species and Evolution. Cambridge, Mass.: Harvard University Press, 729 pp. Millot, J. 1931. Le tubercule anal des Uroc- teides et des Oecobiides (Araneidae). Bull. Soc. Zool. France, 56: 199-205. . 1938. L'appareil serieigene d'Oecobius cellariortim Duges suivi de quelques conside- rations generates sur les glandes secretrices de soie des Araneides. Trav. Stat. Zool. Wimeroux, 13: 479-487. Roewer, C. 1954. Katalog der Araneae. Brux- elles, vol. 2: 1288-1290. Saito, S. 1934. Spiders from Hokkaido. J. Fac. Agr. Hokkaido Imp. Univ. 33: 267-362. Shear, W. 1967. Expanding the palpi of male spiders. Mus. Comp. Zool., Harvard, Breviora 259: 1-28. Simon, E. 1892. Arachnides du Venezuela. Ann. Entomol. Soc. France, 61 : 423-462. Tikader, B. 1962. Studies on some spiders of the genus Oecobius (Family Oecobiidae) from India. J. Bombay Nat. Hist. Soc, 59 (2): 682-685. Yaginuma, T. 1962. The Spider Fauna of Japan. Arach. Soc. E. Asia. Osaka, Japan. (Received 28 November 1968.) 164 Bulletin Museum of Comparative Zoology, Vol. 140, No. 4 INDEX Valid names are in italics; only major discussions cited. Ambika, 134, 135 annulipes, Oecobius, 138 andanti, Oecobius, 141 beatus, Oecobius, 143 beatus, Oecobius, status of "allotype male," 141 benneri, Oecobius, 141 bracae, Oecobius, 156 cellariorum, Clotho, 136 cellariorum, Oecobius, 136 civitas, Oecobius, 157 concinnus, Oecobius, 141 culiacanensis, Oecobius, 159 domesticus, Oecobius, 137 floridanus, Platoecobius, 161 floridanus, Thalamia, 161 interpellator, Oecobius, 146 isolatoides, Oecobius, 150 isolatus, Oecobius, 148 juangarcia, Oecobius, 154 Maitreja, 134 marothaus, Maitreja, 134 nieborowskii, Oecobius, 141 nieborowskii, Tarapaca, 141 nieborowskii, Thalamia, 141 Oecobius, 135 Omanus, 135 parietalis, Oecobius, 138 parietalis, Thalamia, 138 parvus, Oecobius, 148 piaxtla, Oecobius, 150 Platoecobius, 161 putus, Ambika, 146 putus, Oecobius, 144 rivula, Oecobius, 152 sapporensis, "Oecobius," 131 Tarapaca, 134, 135 templi, Oecobius, observed by Debski, 129, 146 texanus, Oecobius, 137 Thalamia, 134, 135 Uroctea, 130 vokesi, Oecobius, 141 x^ 133 SZ Museum of Comparative Zoology ■mnM Systematics and Biology of The Stromateid Fishes of the Genus Peprilus MICHAEL H. HORN HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS, U.S.A. VOLUME 140, NUMBER 5 21 SEPTEMBER 1970 PUBLICATIONS ISSUED OR DISTRIBUTED BY THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Breviora 1952- Memoirs 1864-1938 Johnsonia, Department of Mollusks, 1941- Occasional Papers on Mollusks, 1945- Other Publications. Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine. Reprint, $6.50 cloth. Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of Insects. $9.00 cloth. Creighton, W. S., 1950. The Ants of North America. Reprint, $10.00 cloth. Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 12-15. (Price list on request. ) Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Teredinidae ( Mollusca: Bivalvia ) . $8.00 cloth. Whittington, H. B., and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948. ( Complete sets only. ) Publications of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138, U. S. A. © The President and Fellows of Harvard College 1970. SYSTEMATICS AND BIOLOGY OF THE STROMATEID FISHES OF THE GENUS PEPRILUS1 MICHAEL H. HORN2 TABLE OF CONTENTS Abstract 1 65 Introduction 166 Materials and Methods 169 Systematics 170 Genus Peprilus 170 Key to the Species of Peprilus 173 Peprilus meclius 175 Peprilus ovatus 182 Peprilus simillimus 185 Peprilus snyderi 189 Peprilus burti 192 Peprilus triacanthus 197 Peprilus paru .. 202 Aspects of Functional Morphology . 210 Aspects of Life History and Ecology .. 220 Evolutionary Relationships and Zoogeography 239 Acknowledgments 252 Summary 254 Literature Cited 255 ABSTRACT A complete revision is presented of the genus Peprilus, one of the three genera of the family Stromateidae. The nominal genera Poronotus and Palometa are placed in the synonymy of Peprilus. Seven species 1 This paper is based on a thesis presented to Harvard University in partial fulfillment of the requirements for the Ph.D. in Biology. 2 Museum of Comparative Zoology, Harvard University. Present Address: Department of Bi- ology, California State College, Fullerton, Cali- fornia 92631. are recognized, and one, Peprilus ovatus from the northern Gulf of California, is newly described. Several aspects of functional morphology are considered. The vertebral column, skull, and pectoral fins appear to ossify earlier than the caudal skeleton and median fins, a sequence interpreted as being correlated with an early planktonic life followed by an independent nektonic existence. Vertebral number is relatively constant within a species and is considered to be of possible selective value in main- taining a certain body form. The absence of pelvic fins, the long pectoral fins which are used extensively for propulsion in adult fishes, and the compressed body, may all be correlated with the continuous swim- ming habit of these fishes, especially those larger than 100 mm SL. An hypothesis is proposed that the swimbladder is of hydro- static advantage to juvenile fishes which hover under jellyfish medusae and that it becomes nonfunctional in larger fishes which swim continuously. The scales are highly deciduous, and the skin is underlain by an extensive canal system, the function of which is unknown. The alimentary canal includes a small mouth with nipping teeth, a toothed, muscular pharyngeal sac which shreds food material, and numerous pyloric caeca and a long intestine which provide a great absorptive area. Considerations of life history and ecology generally are of four species: P. triacanthus, P. buiii, P. paru, and P. simillimus. Spawn- Bull. Mus. Comp. Zool., 140(5): 165-262, September 165 166 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 ing occurs in the pelagic surface waters at be the most primitive type, and the deep- varying distances from shore. The eggs bodied P. paru the most highly derived and larvae are planktonic, the latter be- form in the genus. The Camin-Sokal coming capable of independent locomotion method for deducing relationships of con- at a size of about 10 to 15 mm SL. The temporaneous species is used to reconstruct species occur in a wide range of salinities a dendrogram of species relationships. Two and variously inhabit all depths over the somewhat subtle species groups are recog- continental shelf and generally over a sand nized in the genus, and each group is or mud bottom. The genus is essentially represented on both sides of the Central a tropical and warm temperate one with American isthmus. Character displacement only two species, P. triacanthus and P. is invoked as a possible mechanism to ex- simiUimus, reaching cooler waters. Sea- plain the existence of two apparently dis- sonal movements appear to be most pro- tinct populations of P. triacanthus in the nounced in P. triacanthus, the species oc- Atlantic off the southeastern coast of the curring most abundantly in temperate United States. regions. Fishes smaller than 100 mm SL The distributions of the species of Pepri- associate with jellyfish medusae of several lus appear to correspond generally to the genera. This association is apparently im- major faunal provinces of the Atlantic Coast portant during the early critical growth and the Pacific Coast of the Americas. The phases of the fishes. Peprilas is a low-level species generally traverse the zoogeo- carnivore; jellyfish medusae seem to be graphic subdivisions established from the an important element in the diet, especially study of small fishes inhabiting rocky of juveniles. Other food items include a shores. Sympatry involves the more diverse variety of small crustaceans, polychaete species, and the similar or closely related worms, and small fishes. Fishes of the species tend to parallel one another in dif- genus are evidently significant forage fishes ferent oceans or displace one another lati- for a number of larger fishes, some of tudinally along a continuous coastline, which are of great commercial importance. Niche separation seems to be produced The economically important species of largely by spatial arrangement and eco- Peprilus are generally taken commercially logical displacement, in a region much smaller than the total range of the species, and this seems to INTRODUCTION reflect the pattern of migration and center of abundance of the particular species. . PurPOse. One of the main purposes of Disruption of the Tethys Sea in the *1S 1stud>; has been to examine specimens Miocene apparentlv facilitated the segre- °f *«. st">mateid genus Peprilus from gation of the early'members of the familv throuShout its geographic range, and to Stromateidae and led to the evolution of determlne the limits of the species and the the three extant and essentially allopatric sPecies relationships within the genus. The stromateid genera. The formation of the systematics of the species are considered Central American land bridge in the Plio- in the first section of this paper and include cene, the emergence and submergence of a generic description, a key to the species, land areas associated with the Pleistocene species descriptions and distributions, and glacial and interglaeial periods, and the an account of the geographic variation and prevailing current systems all appear to ontogenetic change. The functional aspects have been important in producing the of certain skeletal features, of body form, current level of differentiation and speci- of the shape, position, and loss of fins, of ation in the genus. the integument, and of the alimentary The elongate P. snyderi is considered to canal are considered in the second section. Systematics and Biology of Pe prills • Horn 167 Aspects of life history and ecology are considered in the third section and include a general account of the distribution in relation to certain physical factors, of sea- sonal distribution and abundance, of faunal associations, of larval and juvenile ecology, and of food habits. In this section, the dis- cussion is concentrated on the better known and more accessible species. The final section concerns the evolutionary relation- ships and zoogeography of the genus and species. The proposed primitive and de- rived conditions in the genus are discussed, and the Camin-Sokal method is used to reconstruct a dendrogram of probable species relationships. The zoogeographical regions of the North and South American coasts are discussed in relation to the dis- tribution of the species of Peprilus. A brief analysis is made of the degree of divergence in allopatric versus sympatric sets of species. The nature and level of speciation in the genus are considered and briefly compared to that of other types of coastal marine fishes. Background. The fishes of the family Stromateidae comprise a small, well-de- fined marine group of some 12 to 15 species. Members of this perciform family occur on the continental shelf of oceans of both the Old and New Worlds; none have reached Australia or New Zealand. Haed- rich (1967) recognized three genera in the family, Stromateus, Pampus, and Pepri- lus. These genera are allopatric except for a small area of overlap between Stromateus and Peprilus on the east coast of South America, and possibly a similar area on the west coast. Stromateus occurs along both coasts of South America, along the West African coast, and in the Mediterranean Sea; Pampus occurs from the Iranian Gulf to Japan; and, Peprilus is a New World genus of the east and west coasts of North, Central, and northern South America. Stromateids are small (shorter than 18 inches), essentially pelagic, schooling fishes of coastal regions. Several of the species are commercially important. All possess <*> Figure 1. Pelvic bones of the three stromateid genera: A zz Stromateus; B rr Pampus; C = Peprilus. toothed pharyngeal sacs, a characteristic of the entire suborder Stromateoidei. According to Haedrich ( 1967 ) , the Stroma- teidae are the apex in stromateoid evo- lution. The genus Peprilus is the most speciose of the family, seven species being recog- nized in the current study. Peprilus is distinguished from the other stromateid genera by the combination of: deep body; large eye; long pectoral fin; two to four small spines ahead of the dorsal and anal fins; a ventral spine on the pelvic bone (Fig. 1); and, no pelvic fins. Differences also exist among the three genera in meristic values; in the number of epural elements in the caudal skeleton (three in one species of Stromateus, and two in all other species); and, in the number of branchiostegal rays (five in Pampus, six in the other two genera). At least three species of Peprilus form important fisheries. The flesh is considered to be of rich and delicate flavor. Although the stromateid fishes are famil- iar coastal fishes often collected in large numbers, work on the group has been sporadic; and the relationships within the family have remained enigmatic. Much remains to be known of the life history and distribution of even the most common species. Early revisions of the family include those of Fordice (1884), Jordan and Evermann (1896; 1898), and Regan 168 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 ( 1902 ) . Each of these studies, however, had only limited material available from restricted geographic areas. Haedrich ( 1967 ) has provided the most compre- hensive and recent revision of the sub- order; in his revision he arrived at a classification consisting of five families and fourteen genera. Caldwell ( 1961 ) has dis- cussed the populations of one species, Peprilus triacanthus. Studies of the early life historv of the Stromateidae, particularly of Peprilus, have been largely confined to P. triacanthus and, to a lesser extent, P. paru. Such works in- clude those of Kuntz and Radcliffe (1918), Hildebrand and Schroeder (1927), Pearson (1941), and Colton and Honey (1963). Bigelow and Schroeder (1953) reviewed the life history and bionomics of Peprilus triacanthus (the butterfish) in the Gulf of Maine. T. J. Hart (1946) discussed at length the bionomics and seasonal distri- bution of Stromateus maculatus ( = S. hrasilicnsis) off the Atlantic Coast of South America, and he compared these aspects of the species with those of Peprilus tria- canthus off the Atlantic Coast of North America. Rege (1958), and Rege and Bal ( 1964 ) , among others, have contributed to the knowledge of the life history of the species of Pampas off the coast of India. A well-known aspect of the biology of stromateid fishes is their association with coelenterates, either medusae or siphono- phores, in the pelagic environment. Man- sueti (1963) has reviewed the literature on this association and has provided additional information on the interaction between Peprilus paru and the scyphomedusa, Chrysaora quinquecirrha. Stromateoids are also known to hover beneath flotsam and Sargassum weed ( Besednov, I960; Gooding and Magnuson, 1967). The hovering of stromateids beneath floating objects has not been well documented but probably does occur. Kennedy and Milkman (1956) demon- strated that the lens of the butterfish, Peprilus triacanthus, transmits a consider- able band of ultraviolet (UV) light. They indicated that a plankton-feeding fish, for example, might use UV sensitivity to locate areas of suspended matter, including organisms, by responding to the light of shortwave length scattered from such particles. Such a mechanism may facilitate the search for food or the location of jellyfish medusae. Morphological and anatomical studies of stromateoid fishes are few. Haedrich's ( 1967 ) comparative study of the osteology of the group, and Biihler's ( 1930 ) exten- sive account of the digestive tract of the Centrolophidae and Stromateidae, are among the most notable contributions in this area. Gilchrist (1922) examined the teeth of the pharyngeal sac of stromateoids of South Africa, and Barnard (1948) re- vised and extended the former's work. Isokawa et al. ( 1965) have made additional contributions to the study of pharyngeal sacs and teeth. Fossil stromateoids are few. Two new fossil genera have been found by Bonde ( 1966 ) in the lower Eocene of Denmark. However, no discoveries to date can be reliably referred to the Stromateidae. Since the three stromateid genera are essentially allopatric, it is reasonable to treat each genus separately. This study is primarily concerned with the genus Pepri- lus, although reference to, and comparison with, other stromateids and other members of the suborder are made. Three species of Peprilus, P. paru, P. triacanthus, and P. simillimus, are familiar fishes of North American coasts, but other members of the genus are less well known. Currently in fishery and fauna! works, three genera are frequently recognized within what is here considered the genus Peprilus. The characters used for generic separation seem more reasonably to be criteria for delineating species; the species form a reasonably compact group. There are close species pairs in the genus, and there has been doubt as to whether certain populations constitute valid species. Active Systematics and Biology of Peprilus • Horn 169 speciation is in the process and creates uncertainty as to rank. MATERIALS AND METHODS The list of specimens examined for each species is included in the systematic section under the particular species in alphabetical order of museum and institutional names. Museum and institutional names are ab- breviated as follows: ANSP — Academy of Natural Sciences of Phila- delphia BM — Institut fiir Spezielle Zoologie und Zoologisches Museum, Berlin BC — University of British Columbia, Van- couver CAS — California Academy of Sciences, San Francisco FMNH — Field Museum of Natural History, Chicago FSBC — Florida Board of Conservation Marine Laboratory, St. Petersburg IMS — University of Texas Marine Science Institute at Port Aransas LACM — Los Angeles County Museum MCZ — Museum of Comparative Zoology, Har- vard University SIO — Scripps Institution of Oceanography, La Jolla SU — Division of Systematic Biology, Stanford University TABL — U.S. Bureau of Commercial Fisheries Tropical Atlantic Biological Laboratory, Miami TU — Tulane University, New Orleans USNM — United States National Museum, Washington, D.C. UCLA — University of California, Los Angeles UMML — Institute of Marine Sciences, Uni- versity of Miami U W — University of Washington, Seattle WHOI1 — Woods Hole Oceanographic Institu- tion Distributional and ecological data were obtained from three Bureau of Commercial Fisheries laboratories: Biological Labora- tory, Woods Hole, Massachusetts; Explora- tory Fishing and Gear Research Base, Pascagoula, Mississippi; and, Fishery- Oceanography Center, La Jolla, California. Certain data on larval fishes were obtained 1 These specimens will ultimately be deposited in the MCZ. from the Bureau of Commercial Fish* Biological Laboratory, Brunswick, ia. Two lots of specimens collected Cruise 16 of the R/V ANTON BRUUN were received from the Smithsonian Ocean- ographic Sorting Center, Washington, and are listed as uncatalogued specimens of Peprilus medius in the MCZ. Three lots of specimens, two of Peprilus burti and one of P. para, were collected personally at Port Aransas, Texas, and are listed as uncatalogued specimens in the MCZ. Measurements were made point-to-point to the nearest tenth millimeter with a pair of fine point dial calipers. A dissecting microscope was used in measuring very small specimens. Measurements made were: total length (TL); standard length (SL); head length; snout length; eye diameter; length of upper jaw; interorbital width; length of pectoral fin; predorsal distance I; predorsal distance II; preanal distance; maximum depth of body; and, least depth of caudal peduncle. Counts in all but the largest specimens were made using a fine needle and under low magnification. Median fin-ray counts of very small specimens were made using transmitted, polarized light. Counts made were: total dorsal fin ele- ments (D); total anal fin elements (A); total pectoral fin elements (P); total gill rakers; precaudal and caudal vertebrae; and, lateral line scales. Except for predorsal distances I and II, all measurements and counts conform to the methods of Haedrich (1967). Pre- dorsal distance I is the distance from the tip of the snout to the base of the first element of the dorsal fin. Predorsal distance II is the distance from the tip of the snout to the anterior edge of the first free inter- neural. Skeletal features were studied from radiographs and cleared-and-stained speci- mens. Radiographs suitable for counting vertebrae and for general osteological examination were made with three differ- 170 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Table 1. Number* of Dorsal Fin-rays in Species of Peprilus. Species 38 39 40 41 42 43 44 45 46 47 48 49 50 51 medius 1 1 7 33 40 34 33 ovatus 11 10 33 52 22 12 5 simillimus 1 7 13 15 31 15 13 3 snyderi 1 4 7 6 4 burti 1 2 2 5 14 55 30 31 12 2 1 triacanthus 4 5 7 25 30 53 15 7 4 pant 1 2 8 23 29 59 23 25 5 3 * Excluding spines. ent types of industrial X-ray film at set- tings of 30 kilovolts and 200 milliampere- seconds or of 50-54 kilovolts and 100-150 milliampere-seconds, depending upon the type of film and size of specimens. The enzyme method of clearing-and-staining small vertebrates developed by W. R. Taylor ( 1967 ) was used exclusively in this study. Specimens ranging in size from six to 115 mm SL were successfully prepared using this method. Drawings of each species were made from actual specimens with the aid of radiographs. Other drawings were made similarly, but with the aid of a camera lucida attached to a Wild dissecting micro- scope. The drawing of the digestive tract is semi-diagrammatic although made from a specific preparation. Size-on-size and ratio-on-size diagrams are used for each species to show changes with growth and to make species compari- sons. Five characters, head length, eye diam- eter, body depth, length of pectoral fin, and depth of caudal peduncle, each ex- pressed as a percentage of standard length, were plotted on standard length. The re- sulting scatter diagrams show trends and indicate change in size of body parts relative to standard length, but because of the wide scatter and curvilinearity, straight lines were not fitted to the data. Parr ( 1956 ) has advocated the use of ratio-on- size diagrams in preference to size-on-size diagrams. Ratios, however, have a greater inherent error than the measurements of which they are composed and are to be used cautiouslv ( Simpson, Roe, and Lewontin, 1960: 18; Sokal, 1965: 346). Five characters, eye diameter, length of upper jaw, length of pectoral fin, body depth, and depth of caudal peduncle, each expressed as a direct value, were plotted on standard length. Straight lines (Y = A + RX) were fitted to these size-on-size diagrams without transformation of the values by the method described by Rartlett (1949) and outlined by Simpson et al. (1960: 232). The regression lines for each character are presented in a single diagram for all the species, and the statistics of the lines are tabulated. Measurements of either 24, 90, or 100 specimens were used for the size-on-size diagrams and of either 24, 90, or 109 speci- mens for the ratio-on-size diagrams. The specimens were randomly chosen from the broadest size and geographical ranges pos- sible. Except for the few additional speci- mens used in certain of the ratio-on-size plots, measurements of the same specimens were used for both types of diagrams. Computation of statistics describing the size-on-size regressions was performed at the Harvard Computation Laboratory on an IBM 7094 digital computer. Unless indicated otherwise, other sta- tistical procedures follow methods de- scribed by Simpson et al. ( 1960 ) . SYSTEMATICS Genus Peprilus Cuvier, 1829 Rhombus Lacepede, 1800: 60. (Type species: Chaetodon alepidotus Linnaeus, 1766: 460, by monotypy. Charleston, South Carolina. Pre- Systematics and Biology of Peprilvs • Horn 171 Table 2. Number* of Anal Fix-rays en Species of Peprilvs. Species 35 36 37 38 39 40 41 42 43 44 45 46 47 mcdius 1 2 17 44 41 32 16 4 ovatus 7 14 29 50 20 20 6 simil limits 1 6 15 30 28 12 5 2 1 snyderi 3 8 1 7 4 burti 1 1 4 15 51 43 26 11 3 triacanthus 2 10 33 40 34 23 9 1 paru 1 2 7 20 32 29 31 30 18 4 2 * Excluding spines. occupied by Rhombus Humphrey, 1797, Mol- lusca. ) Peprilus Cuvier, 1829: 213. (Type species: Sternoptyx gardenii Bloch and Schneider, 1801: 494; by subsequent designation of Gill, 1862: 126. Charleston, South Carolina. A synonym of Chaetodon alcpidotus Linnaeus, 1766: 460.) Poronotus Gill, 1861: 35. (Type species: Strom- ateus triacanthus Peck, 1804: 48, fig. 2, pi. II, by monotypy. Piscataqua Biver, New Hamp- shire. ) Palometa Jordan and Evermann, 1896: 966. (Sub- genus. Type species: Stromateus palometa Jordan and Bollman, 1889: 156, by original designation. Elevated to generic rank, Jordan and Evermann, 1898: 2849. Panama Bay, Pacific Ocean. ) Simobrama Fowler, 1944: 2, fig. 2. (Type species: Seserinus xanthurus Quoy and Gaimard, 1824: 384, by original designation. Bio de Janeiro, Brazil. A synonym of Stromateus paru Lin- naeus, 1758:' 248.) Diagnosis. The genus Peprilus is dis- tinguished from other stromateid genera by the combination of deep body, large eye, long pectoral fin, two to four bladelike spines anterior to the rays of the dorsal and anal fins, a ventral spine on the pelvic bone, and no pelvic fins. Also, Peprilus is distinguished specifically from the stroma- teid genus Pampus in having six rather than five branchiostegal rays and in having a movable rather than a fixed maxillary bone. Peprilus is further distinguished from Stromateus in having 29 to 36 verte- brae rather than approximately 42 to 48. Description. Body deep, ovate to elon- gate, depth 35 to 87 per cent of standard length, highly compressed. Caudal pe- duncle short, slender, and compressed. Dor- sal and ventral profiles convex and similar; anterior dorsal profile slightly to strongly convex. Snout short and blunt; jaws equal; maxilla just reaching to line of anterior margin of eye; eye large with surrounding adipose tissue reaching to nostrils; eye diameter either as great or greater than length of snout. Nostrils double, anterior round, posterior a slit, directly anterior to eye, near end of snout. Dorsal and anal fins with long base, similar to each other, either slightly, moderately, or extremely falcate, as produced by elongation of the first 10 to 13 rays; dorsal and anal co- terminal just anterior to caudal peduncle. Dorsal fin with two to four small, bladelike spines anterior to and continuous with 40 to 51 fin-rays; spines often irregular in shape, free surface jagged or crenulated, anterior spine pointed on both ends; third or fourth spine often difficult to distinguish from anteriormost rays. Anal fin equal to or shorter than dorsal, with two or three (rarely four) small spines preceding and continuous with the 35 to 47 fin-rays; spines usually easier to distinguish than in dorsal since rays of anal are more dis- tinctly segmented. Pectoral fin lateral, Table 3. Number of Pectoral Fin-rays in Species of Peprilus. Species 17 18 19 20 21 22 23 24 medius 2 10 52 75 19 ovatus 1 9 74 54 12 simillimus 1 15 65 25 1 snyderi 3 10 9 burti 2 16 60 65 5 triacanthus 1 1 4 51 71 21 paru 1 1 17 39 63 40 11 172 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Table 4. Total Number of Gill Rakers in Species of Peprilus. Species 2d 21 22 23 24 25 26 27 medius 15 47 67 24 ] ovatus 2 7 56 63 17 1 simillimus 19 29 19 8 snyderi 6 6 4 2 hurti 1 14 26 26 3 1 triacanthus 4 8 10 11 paru 9 39 44 6 just below level of eye, long, winglike, lower rays much shorter than upper ones, base of fin slightly inclined, becoming relatively longer with age, often short and fanlike in larvae and small juveniles; 17 to 24 rays. No pelvic fins, a small spine on the posterior end of the fused pelvic bones projects posteroventrally through the skin at mid-ventral line below base of pectoral fin; pelvic bones closely adjacent ante- riorly, fused posteriorly. Edge of coracoid may project underneath head at level of preopercular margin. Caudal fin long to very long, deeply forked, about 18 to 40 per cent of total length; lobes equal. Scales small, cycloid, irregular in shape, decidu- ous; occasionally with small, blunt, knob- like structures on posterior surface; scales often crowded, usually absent in preserved material; cheek, suborbital area, preopercle, and sometimes opercle scaled; top of head and nape generally scaleless; scales extend- ing onto caudal, dorsal, and anal fins, less so on pectorals, usually only to base of latter. Lateral line of trunk of simple, tubed scales which are often less deciduous than those of rest of body; lateral line above intermuscular septum, either follow- ing dorsal profile or more highly arched anteriorly, extending from cleithrum onto caudal peduncle to base of caudal fin. A branch of lateral line extends from immedi- ately above eye posteriorly to head of hyomandibula where it joins a similar branch then turns upward toward mid- dorsal line in a wide, bony tract; cephalic- lateral line of pores and branching canals on cheek, opercular area, snout, lower jaw, and top of head. Subdermal canal system on body, top of head and nape; main canal along intermuscular septum with numerous, parallel side-branches; often only partially visible or completely invisible in preserved material. Skin of top of head underlain by a series of small, dendritic canals extend- ing backward over nape. In two species there is an irregular row of about 17 to 25 relatively large pores just below the an- terior half of the dorsal fin. A third species often has a series of slightly smaller, ir- regularly-spaced pores along entire dorsal surface. All species have very small pores scattered over body, apparently in associ- ation with the subdermal canal system. Head about 25 to 40 per cent of standard length, depending upon species and age. Premaxilla not protractile; maxilla movable. Lacrimal bone small, end of maxilla exposed when mouth closed. No supra- maxilla. Jaw teeth small, uniserial, close- set; premaxillary teeth slightly recurved, either pointed and simple, or, in two species, with three small cusps; in at least Table 5. Number of Vertebrae in Species of Peprilus. Precaudal Caudal Species 12 13 14 15 16 17 18 19 20 21 22 medius 173 8 20 153 8 ovatus 3 154 5 143 9 simillimus 158 2 138 22 snyderi 3 20 20 3 hurti 273 5 262 6 triacanthus 279 7 62 208 2 paru 1 181 3 176 3 Systematics and Biology of Peprilus • Horn 173 two other species both types of premaxil- lary teeth may be present; teeth of lower jaw with three to five small cusps; slightly more closely spaced than premaxillary teeth. Vomer, palatines, and basibranchi- als toothless. Gill membranes united across isthmus, divided from about level of an- terior part of eye. Opercle and preopercle thin, flat; preopercle finely striated; oper- cular margins entirely with two flat, blunt, ill-defined spines. Gill rakers similar in all species, slightly more than one-half the length of filaments; with fine teeth on inner edge; closely set; six to nine on epi- branchial, one at junction of elements, and 13 to 18 on lower elements of first arch (cerato- and hypobranchials). Pseudo- branch small. Six branchiostegal rays, four on ceratohyal, two on epihyal. Vertebrae, 12 to 15 precaudal, 16 to 22 caudal, 29 to 36 total. Three free interneural bones anterior to dorsal fin. Sclerotic bones well ossified. Toothed pharyngeal sac behind last gillarch as in all other stromateoid fishes, joining tubular esophagus; stomach a simple sac; intestine long and in loops, about two and one-half to three times the standard length of the body; pyloric caeca numerous, short, in a dendritic mass adjacent to stomach. Swimbladder thin- walled, physoelistous, continuous with dor- sal peritoneum; easily deflated; possibly nonfunctional or absent in adults; found only in fish smaller than 80 to 100 mm SL; not yet seen in all species. Buccal and pharyngeal cavities light in color; peri- toneum silvery, with peppering of black pigment that is more dense dorsally. Gonads paired, in posteroventral region of body cavity; ovary in mature and maturing specimens elongate, yellow, and granular; eggs in ripe ovary spherical, yellowish, less than 1 mm in diameter; testis in mature and maturing individuals elongate, thin, yellowish white, and smooth in texture. No external sexual dimorphism apparent. Color in life iridescent bluish or greenish silver dorsally to silver ventrally. At least one species often has large, dark spots on the dorsal and upper ventral surface, in preservative varying from bluish br to brown dorsally to silver ventrally or body completely brown or bluish brown. Individual melanophores conspicuous in specimens up to about 80 to 100 mm SL, dense and small dorsally, larger and more dispersed ventrally. Anal and dorsal fins with distinct melanophores on membranes between fin-rays, slightly less evident on the paler caudal and pectoral fins. Eye usually darker than rest of body, lens opaque surrounded by iris which is either completely black or partly black and partly yellowish white. Key to the Species of Peprilus Ranges of proportional measurements in thou- sandths of standard length and ranges of meristic values are each followed by the mean or two different means (when for two species) in paren- theses. la. Row of about 17 to 25 relatively large pores immediately below anterior half of dorsal fin; premaxillary teeth usually with three small cusps 2 lb. No row of pores below anterior half of dorsal fin; premaxillary teeth pointed, simple 3 2a. Body elongate, shallow to moderately deep 364 to 600 (x 458); eye moderately large, 061 to 133 (086); caudal vertebrae 17 to 20, usually 19, rarely 17 or 20; dorsal and upper ventral surface in adults often mot- tled with dark spots P. triacanthus (Peck, 1804) p. 197, Figs. 21, 22. 2b. Body moderately elongate, moderately deep to deep, 460 to 640 ( x 551 ) ; eye large, 065 to 144 ( 100); caudal vertebrae 16 to 18, usually 17; dorsal or upper ventral surface rarely if ever mottled P. burti Fowler, 1944 p. 192, Fig. 19. 3a. Dorsal and anal fins except in larvae and juveniles smaller than 50 to 75 mm SL moderately to extremely falcate, the longest anal ray six or more times the length of the shortest anal ray; dorsal often slightly less falcate 4 3b. Dorsal and anal fins only slightly falcate, the longest dorsal and anal rays less than six times the length of the shortest of each 5 4a. Body ovate, very deep, 565 to 877 (x 710); dorsal rays 38 to 47, usually 41 to 45 (42.9); gill rakers 20 to 23, usually 21 or 22 (21.5); caudal vertebrae 16 to 18, usually 17 — 174 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 125 185 STANDARD LENGTH mm Figure 2. Regression of body depth on standard length for each species of Peprilus: A = P. medius; B = P. ovatus: C = P. simillimus) D = P. snyderi; E = P. borti; F = P. rriacanfhus; G = P. paru. Statistics in Table 6. 125 185 STANDARD LENGTH mm Figure 3. Regression of eye diameter on standard length for each species of Peprilus. Symbols as in Figure 2. Statis- tics in Table 6. Systematics and Biology of Peprilus • Horn 175 P. paru (Linnaeus, 1758) p. 202, Fig. 24. 4b. Body moderately elongate, moderately deep to deep, 459 to 618 (x 529); dorsal rays 42 to 51, usually 45 to 48 (46.7); gill rakers 23 to 27, usually 24 to 26 (24.7); caudal vertebrae 20 to 22, usually 21 _ P. medius (Peters, 1869) p. 175, Fig. 8. 5a. Body ovate, deep, 542 to 676 (x 619); eye moderately large, 078 to 115 (093); snout length considerably less than eye diameter, 052 to 074 ( 063 ) ; dorsal spines 3 or 4, most frequently 4; often a series of irregularly-spaced, medium-sized pores vis- ible along dorsal surface; total vertebrae 31 to 33, usually 32 (32.0) . P. ovatus sp. n., p. 182, Fig. 10. 5b. Body elongate, shallow to moderately deep, 371 'to 517 (x 427, 461); eye small, 050 to 119 (070, 073); snout length about equal to eye diameter, 058 to 083 (068, 070); dorsal spines 2 to 4, usually 3; no series of medium-sized pores usually visible along dorsal surface; total vertebrae 30, 31, or 36 6 6a. Dorsal rays 43 to 49, usually 45 to 48 (x 46.5); anal rays 40 to 44 (42.0); caudal vertebrae 21 or 22, usually 21; total vertebrae 36 P. snyderi Gilbert and Starks, 1904 p. 189, Fig. 16. 6b. Dorsal rays 41 to 48, usually 43 to 47 (x 44.8); anal rays 35 to 44, usually 38 to 41 (39.5); caudal vertebrae 17 or 18, usually 17; total vertebrae 30 or 31 P. simillimus (Ayres, 1860) p. 185, Fig. 13. Peprilus medius (Peters, 1869) Figures 8, 18a Stromateus medius Peters, 1869: 707 (original description, Mazatlan, Mexico, holotype seen, 138.4 mm SL, BM 7073); Jordan, 1883: 284; Fordice, 1884: 314. Stromateus palometa Jordan and Bollman, 1889: 156 (original description, Pacific Ocean, off coast of Panama, 8°16'30"N, 79°37'45"W, ALBATBOSS 2804, holotype not seen; four paratypes seen, 50.0 to 58.3 mm SL, SU 434). Rhombus (Palometa) palometa, Jordan and Ever- mann, 1896: 966, Palometa erected as subgenus. Rhombus (Palometa) medius, Tordan and Ever- mann, 1896: 967. Palometa palometa, Jordan and Evermann, 1898: 2849, subgenus Palometa elevated to generic level; Jordan, Evermann, and Clark, 1930: 267. Palometa media, Jordan and Evermann, 1898: 2849; Jordan, Evermann, and Clark, 1930: 267. Peprilus palometa, Gilbert and Starks, 1904: 85; Meek and S. F. Hildebrand, 1925: 412; Haed- rich, 1967: 107. Peprilus medius, Gilbert and Starks, 19l Meek and S. F. Hildebrand, 1925: 413; rich, 1967: 107. Material examined. Those specimens marked with an asterisk (*) have been radiographed. The number radiographed equals the number measured unless other- wise indicated. CS indicates specimens cleared and stained. Size ranges, in mm, are standard lengths (SL). *BM 7073 (1:138.4 mm, holotype of Stromateus medius Peters, Mazatlan, Mexico); BC 56- 164 (1:113.0, Paita, Peru, CS); *BC 56-233 (5:143.7-186.0, Paita, Peru); *BC 59-66S (28:108.6-144.0, Pacific Ocean, Panama Bay near Panama City); *BC 59-685 (6:119.3-139.5, Pacific Ocean, Chiman area SE of Panama City); *BC 59-686 (3:125.7- 139.6, Pacific Ocean, near Puerto Obaldia, Panama); *BC 59-687 (4:124.8-174.0, Panama City market); *BC 60-25 (1:143.5, Mexico, Golfo de Tehuantepec, Puerto Arista to Salina Cruz); *BC 60-115 (2: 37.4, 144.5, Pacific Ocean, Taboga I., off Panama); BC 61-126 (1:189.0, Mex., Acapulca market); *BC 61-151 (1:131.0, Mazatlan, Mex.); *FMNH 20500 to 20503 (4:123.5-158.5, Panama City market, Feb- ruary-March, 1911); *FMNH 73840 (8: 109.6-185.0, Mex., 1-5 miles off Chiapas coast at Guatemala border to above San Benito, 18-27 m, 14-18 December 1954, 5 radiographed); *LACM 6545-2 (1:112.9, Sinaloa, Mex., 30 August 1962); *LACM 6917-8 (1:163.2, Pacific Ocean, El Savador, Fondo del Mar, August, 1965); *MCZ uncat. (15:111.2-140.6, R/V ANTON BRUUN Cr. 16, Sta. 624E, 04°53'S, 81°23' W, 75-91 m, 2 June 1966, 14 radio- graphed); *MCZ uncat. (11:75.3-90.8, R/V ANTON BRUUN Cr. 16, Sta. 624b, 04°48'S, 81°17'W, 16 m, 2 June 1966, 1 CS); *SIO H51-324 (1:141.3, Golfo de Fonseca, Honduras, about 13°00'N, 87°40/ W, 1 August 1951, M/V RENOWN); *SIO 55-58 (1:130.5, Pacific Ocean, between San Diego, California, and Ecuador, dipnet from tuna clipper, July, 1955); *SIO 59- 263 (5:76.5-80.5, Sinaloa, Mex., off Rio San 176 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 125 5TANDARD LENGTH Figure 4. Regression of length of upper jaw on standard length for each species of Peprilus. Symbols as in Figure 2. Statistics in Table 6. STANDARD LENGTH mir Figure 5. Regression of length of pectoral fin on standard length for each species of Peprilus. Symbols as in Figure 2. Statistics in Table 6. Systematics and Biology of Peprilus • Horn 177 125 185 STANDARD LENGTH mm 245 Figure 6. Regression of depth of caudal peduncle on standard length for each species of Peprilus. Symbols as in Figure 2. Statistics in Table 6. Lorenzo, 23°50.5'N, 107°18.2'W, 77-88 m, 16 May 1959, SIO Vermilion Sea Expe- dition, 4 radiographed; *SIO 64-749 (1:112.4, Panama, off San Carlos, 17 August 1961); *SU 434 (4:50.0-58.3, paratypes of Stromateus palometa Jordan and Bollman, 1889, off coast of Panama, "N, 79o8o16'3037'45"W, ALBATROSS Sta. 2804, 86 m); *SU 1867 (1:68.0, same as SU 434 above, part of type material of Stromateus palometa, incorrectly labeled as From Galapagos Is., according to S. F. Hil- debrand (MS)); *SU 6998 (5:136.0-180.4, Panama); *USNM 41272 (1:65.0, Pacific Ocean, ALBATROSS Sta. ?); *USNM 200351 (1:150.0, Peru, Caleta Cruz, Tumbes, 37-55 m, 25 March 1965); *UCLA W51-246 (2:140.0, 146.5, Ecuador, vicinity of Guayaquil, 16 November 1951); *UCLA W53-275 (19:110.6-178.0, Panama Bay be- tween Panama city and Punta Gorda, 4-27 m, 1-5 July 1953); UCLA W53-311 (2: 109.7, 111.8, Panama, between Islas Perlas and Isla Otogne, 8°32'10"N, 79°21'11"W, 64 m, 19 March 1953); *UCLA W54-55 (1:121.3, Costa Rica, S of Golfo de Nicoya, 9-27 m, 29 November 1952); *UCLA W54-167 (1:141.2, Costa Rica, Golfo de Nicoya, 5-8 m, 16-18 April 1952); *UCLA W54-334 (6:121.8-163.5, Panama Bay, mouth of Anton River, 15 m, 21 May 1954 ) ; *UCLA W56-112 (2:111.1, 112.5, Gulf of California, Sinaloa, Mex., Bahia Topolo- bampo, from mouth to 6-9 miles N, 2-9 m, 22 May 1956); *UCLA W56-113 (15:84.5- 115.4, Gulf of Calif., Sinaloa, Mex., Bahia Topolobampo, 6-11 m, 24 May 1956); *UCLA W56-115 (1:96.0, Gulf of Calif., Sinaloa, Mex., Bahia Topolobampo, in main channel from islands E of town to Punta Copas, 13 m, 31 May 1956); *UCLA W56-117 (11:84.7-108.2, Gulf of Calif., Sinaloa, Mex., Bahia Topolobampo, N of entrance, 9-13 m, June, 1956); *UCLA W56-123 (2:109.6, 112.2, Gulf of Calif., Sinaloa, Mex., vicinity of I. de Altamura, 25°15'N, 108°30'W, 9 m, 10 March 1956); *UCLA W58-46 (1:163.8, Gulf of Calif., 178 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Table 6. Statistics Describing Regressions (y = a + bx)oF Body Measurements on Standard Length, Figs. 2-6, for the Species of Peprilus. x = mean size (SL); y = mean of body proportion: N = number of specimens; a = y intercept; b = slope; CL-b = 95% confidence limits for b; r = cor- relation coefficient. Body Proportion X y N a b CL-b r Body Depth medius 121.6 63.1 100 -8.4 0.59 0.49-0.69 0.96 ovatus 85.6 52.1 90 4.6 0.56 0.50-0.62 0.99 simillimus 123.8 57.3 100 -4.6 0.50 0.43-0.57 0.97 snyderi 170.0 70.8 24 5.4 0.38 0.30-0.46 0.99 burti 82.0 44.9 90 1.7 0.53 0.46-0.60 0.99 triacanthus 108.4 48.3 100 3.1 0.42 0.35-0.49 0.98 paru 91.6 63.5 100 9.7 0.58 0.48-0.68 0.95 Eye Diameter medius 121.6 9.9 100 3.5 0.052 0.042-0.062 0.93 ovatus 85.6 7.9 90 1.5 0.074 0.067-0.081 0.97 simillimus 123.8 8.6 100 3.9 0.038 0.027-0.049 0.81 snyderi 170.0 11.2 24 3.3 0.046 0.037-0.055 0.97 burti 82.0 7.8 90 1.0 0.083 0.07.3-0.093 0.90 triacanthus 108.4 8.4 100 2.4 0.056 0.046-0.066 0.84 paru 91.6 9.0 100 1.8 0.078 0.070-0.086 0.90 Length of Upper Jaw medius 121.6 10.0 100 2.60 0.061 0.054-0.068 0.96 ovatus 85.6 6.8 90 0.59 0.072 0.065-0.079 0.98 simillimus 123.8 9.3 100 1.00 0.067 0.060-0.074 0.96 snyderi 170.0 13.2 24 2.20 0.065 0.060-0.070 0.99 burti 82.0 7.1 90 0.43 0.081 0.077-0.085 0.99 triacanthus 108.4 8.9 100 0.81 0.075 0.069-0.081 0.98 paru 91.6 7.5 100 1.30 0.068 0.062-0.074 0.97 Length of Pectoral F 'in medius 122.2 47.0 100 -6.10 0.43 0.39-0.47 0.97 ovatus 85.8 32.4 90 -6.60 0.45 0.40-0.50 0.97 simillimus 123.8 41.6 100 1.70 0.32 0.25-0.39 0.93 snyderi 170.0 57.4 24 -0.57 0.33 0.27-0.39 0.99 burti 81.5 28.4 90 -3.00 0.38 0.32-0.44 0.98 triacanthus 108.3 34.9 100 -1.20 0.33 0.26-0.40 0.97 paru 92.1 35.9 100 -2.50 0.42 0.38-0.46 0.99 Depth of Caudal Peduncle medius 121.6 9.4 100 -3.30 0.104 0.088-0.120 0.95 ovatus 85.6 6.9 90 -0.67 0.090 0.077-0.103 0.97 simillimus 123.8 8.1 100 -1.10 0.075 0.062-0.088 0.95 snyderi 170.0 12.4 24 -2.10 0.085 0.079-0.091 0.99 burti 82.0 6.3 90 -0.37 0.081 0.074-0.088 0.99 triacanthus 108.4 7.6 100 -0.31 0.073 0.064-0.082 0.97 paru 91.6 8.4 100 -1.04 0.103 0.095-0.111 0.97 Sinaloa, Mex., S of Bahia Topolobampo, off I. San Ignacio and I. Macapule, 7-13 m, 10-14 February 1958); UCLA W58-218 (1:126.2, Gulf of Panama, 10 miles WNW I. Pacheca, 08°44'N, 79°13,W, 48 m, 8 156.0, Panama Bay, Punta Bay, Punta Chame and Punta Anton, 6-27 m, 6-9 September 1958); *UCLA W58-304 (8: 100.4-131.6, Panama Bay, between Punta de Hicacal and Bio Pasiga, 3-4 m, 7-9 March 1957); *UCLA W58-303 (14:124.3- September 1958); UCLA W58-305 (1: Systematics and Biology of Peprilus • Horn 179 R pgru Figure 7. Distribution of the species of Peprilus. 136.0, Panama Bay between Rio Chico and Punta de la Plata, 3-6 m, 7-11 September 1958). Diagnosis. P. medius is moderately elon- gate and is the only Pacific Coast species in the genus with moderately to extremely falcate dorsal and anal fins. Specifically and aside from the length of the lobes of the dorsal and anal fins, it is distinguished from the similar P. ovatus by having usually 21 rather than 19 caudal vertebrae; a more elongate body -mean body depth of 529 compared to 619 for P. ovatus, each in thousandths of SL; a mean number of dorsal fin-rays of 46.7 compared to 42.8 for P. ovatus; and usually three dorsal spines rather than frequently four as in P. ovatus. It is distinguished from P. sm/deri by having a deeper body -mean body depth of 529 compared to 427 for P. snyderi, each in thousandths of SL; an eye diameter greater than the length of the snout; and 33 to 35 rather than 36 total vertebrae; and from P. simittimus by having a deeper body -mean body depth of 529 compared to 461 for P. simillimus, each in thousandths of SL; a mean number of anal fin-rays of 43.9 compared to 39.5 for P. simillimus; and 33 to 35 rather than 30 or 31 total vertebrae. The species differs from the related P. paru in the Atlantic by having a more elongate body - mean body depth of 529 compared to 710 for P. paru, each in thousandths of SL; a mean number of dorsal fin-rays of 46.9 compared to 42.9 for P. paru; and 33 to 35 rather than 29 to 31 total vertebrae. Description. Proportional measurements are given in Table 7 and meristic values in Table 8. Body moderately elongate, moderately deep to deep, compressed; anterior dorsal profile moderately convex. 180 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Figure 8 Peprilus medius, 141.3 mm SL, Golfo de Fonseca, Honduras, SIO H51-324. Eye diameter greater than length of snout. Dorsal and anal fins moderately to ex- tremely falcate, longest rays of anal six or more times the length of shortest anal ray; dorsal often slightly less falcate; both usually becoming more falcate with age. Dorsal and anal with three or four ( usually three) small spines preceding the rays. Base of anal fin shorter than base of dorsal (fewer rays). Caudal fin long to very long, deeply forked, about 22 to 40 per cent of total length. Subdermal canal system in- distinct in preserved specimens, seen as vertical, parallel lines on body and dendritic canals on top of head and nape. Premaxillary teeth slightly recurved, pointed, simple. Swimbladder not seen in this species. Coloration as described for the genus. Maximum length probably 250 mm SL. Variation. Coefficients of variation (V) for proportional measurements range from four to about 11 (Table 7) and from 1.2 to about 3.6 for meristic characters (Table 8 ) . According to Simpson et al. ( 1960 ) , V is a good measure of relative dispersion and usually a suitable measure of vari- ability; it is useful in comparing the variability of samples or species. The sample size of P. medius (Table 7) used is considered to be reasonably ade- quate, and specimens from all parts of the known geographic range were examined. Except for vertebral number, ranges of meristic values appear, by inspection, to approach a normal distribution (Tables 1 to 4). Vertebral number, as in all the species of Peprilus, is relatively constant (Table 5). Geographic variation. No decided geo- graphic variation was detected in P. medius. A slight tendency, however, for the median fins to be longer in the lower latitudes of the distribution was noted, but was too small to quantify. Ontogenetic change. Few and moderate changes accompany growth in P. medius in a size range of about 50 to 189 mm SL. Systematica and Biology of Peprilus • Horn 181 Table 7. Proportional Measurements of Pep- rilus MED1US. N = number of specimens; R = range of values; x = mean; SE = standard error of mean; V = coefficient of variation. Size range, 50.0-189.0 mm SL; mean size, 124.4 mm SL. Table 8. Meristic Values of Peprilus medius. Symbols as in Table 7. N R SE In thousands of SL: Head length 162 Snout length 162 Eye diameter 162 Length of upper jaw 162 Interorbital width 162 Length of pectoral fin 160 Predorsal distance I 162 Predorsal distance II 162 Preanal distance 162 Maximum depth of body 182 Least depth of caudal peduncle 162 In thousandths of HL: Snout length 162 Eye diameter 162 Length of upper jaw 162 Interorbital width 162 260-384 295 1.61 6.96 046-080 060 0.37 7.95 061-113 082 0.64 9.87 069-111 082 0.62 9.59 077-109 091 0.41 5.78 290-442 386 2.13 6.99 351-436 390 1.27 4.14 239-312 273 1.18 5.52 369-496 434 1.50 4.39 459-618 529 2.34 5.98 058-094 078 0.68 11.04 152-284 203 1.32 8.31 235-333 279 1.29 5.86 225-333 278 1.10 5.01 240-352 307 1.49 6.18 With growth, head length and eye diam- eter decreases in size relative to SL (Fig. 9). The correlation coefficient in the size- on-size regression of eye diameter is, how- ever, relatively high (Fig. 3; Table 6). Body depth, length of the pectoral fin, and depth of the caudal peduncle increase in size relative to SL with growth up to about 125 to 150 mm SL, beyond which the values become nearly constant (Figs. 9 and 12). The anterior lobes of the dorsal and anal fins tend to increase in length with growth (compare Figs. 8 and 18a). Distribution (Fig. 7). P. medius N R SE Dorsal fin- rays* Anal fin-rays* Pectoral fin-rays Total gill rakers Total vertebrae Lateral-line scales 158 42-51 46.7 0.12 3.19 157 40-47 43.9 0.11 3.03 158 20-24 22.6 0.07 3.63 154 23-27 24.7 0.07 3.56 181 33-35 34.0 0.03 1.21 23 105-117 - is * Excluding spines. known from off the Pacific Coast of north- ern South America, Central America, and Mexico northward to the southern region of the Gulf of California. The southern extent of the range is near Paita, Peru, about 5°S, and just south of Golfo de Guayaquil. The northern known limit of the range is near Bahia Topolobampo, Sinaloa, Mexico, about 25°30'N. One speci- men is known from the Galapagos Islands, but was not examined in this study ( 128 mm SL, TABL uncat. ). Taxonomic comments. The holotype of Stromateus medius Peters is very similar to specimens identified as Peprilus or Palo- meta palometa. The supposed anal ray count of 32 for the holotype of S. medius is incorrect; the correct value is 42, within the range of values for P. palometa. The count can be made accurately only by examining the radiograph of the holotype since the anterior portion of the anal fin is externally mutilated. I have thus placed the two names, S. medius and P. palometa, in synonymy of Peprilus medius. It has been a common practice to identify questionable stromateids from the Gulf of California with either Palometa media or Peprilus medius. A distinct popu- lation does exist in the northern Gulf of California and is described below as new. 182 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Peprilus ovatus sp. n. Figures 10, 18b Holotype. A 128.7 mm SL specimen, USNM 203304; 3-5 February 1949, Gulf of California, Sonora, Mexico, about 20 miles SW of Golfo de Santa Clara near the mouth of the Colorado River, shrimp trawl, 13 to 27 m depth. Paratypes. Two, 56.6, 118.4 mm SL, LACM 30175-1 and -2; data as above for holotype. Two, 56.9, 123.9 mm SL, MCZ 46202; data as above for holotype. Thirty-six, 38.3-126.0 mm SL, UCLA W49-55; data as above for holotype. Additional material examined. Those specimens marked with an asterisk ( * ) have been radiographed. The number radiographed equals the number measured unless otherwise indicated. CS indicates specimens cleared and stained. Size ranges, in mm, are standard lengths (SL). All specimens are from the northern Gulf of California. *CAS 24158 (15:41.4-127.5 mm, Sonora, Mexico, 10 miles off Santa Clara, 5 February 1952, 2 CS); *SIO H47-53 (2:62.5, 67.7, Baja California, between San Felipe and mouth of Colorado River, 31°18'-22'N, 114°47'-50,W, 4-9 m, 6-9 April 1947, shrimp trawler FELIPE ANGELES); *SIO 58-164 (3:51.5-75.6, Baja Calif., 10 miles N of San Felipe, 4 April 1956); *SIO 63-484 (6:55.9-84.3, mouth of Colorado R.); *UCLA W49-91 (36:37.5-124.8, Baja Calif., about 10 miles SE of San Felipe, February, 1950); *UCLA W49-422 (4:70.6-78.0, Baja Calif., be- tween San Felipe and mouth of Colo. R., 6-9 April 1947); *UCLA W49-423 (2:80.8, 84.1, Baja Calif., San Felipe Bay, 10 April 1947); UCLA W49-429 (1:83.5, Baja Calif., off Punta Diggs S of San Felipe, 10 April 1947); *UCLA W53-196 (29:63.2- 88.0, Baja Calif., between San Felipe and mouth of Colo. R., 6-9 April 1947, same collection as UCLA W49-422); *UCLA W53-19S (3:69.0-87.9, Baja Calif., San Felipe Bay, 10 April 1947); *UCLA W54- . • • Eye Diameter ■ • • • ■ " • • • •# • • • • ••• • • • • • ■ • . - - • • • - - 1 ' 1 p i i - ? 40 36 32 Length of Pectoral Fin J L 100 125 150 Standard Length mm Figure 9. Ratio-on-size scatter diagrams for four morpho- metry characters of Peprilus medius. 366 (1:145.2, Baja Calif., 2 to 3 miles N of Punta Diggs, about 7 miles S of San Felipe, 11 m, 23 September 1954, shrimp trawler SAN LUIS); *UCLA 55-2 (9:45.4- 132.5, Baja Calif., Punta Diggs, 9 miles E of San Felipe, 24 m, 1-2 February 1955, shrimpboat YUKY); UCLA W55-23 (2: 60.3, 111.6, Baja Calif., 1 mile S of San Felipe, 0-7 m, 9-11 March 1955, 600 ft Systematics and Biology of Peprilus • 183 (Qmurt. Figure 10. Peprilus ovatus, holotype, 128.7 mm SL, 20 miles SW of Golfo de Santa Clara, northern Gulf of California, USNM 203304. beach seine); *UCLA W55-27 (6:56.2- 110.6, Baja Calif., Punta Diggs, 5-10 miles S of San Felipe, 26 m, 11 March 1955, shrimpboat YUKY); *UCLA W55-28 (3: 112.7-132.0, Baja Calif., 4 to 5 miles N of San Felipe, 27 m, 11 March 1955, shrimp- boat YUKY). Diagnosis. P. ovatus is a short, deep- bodied species with slightly falcate dorsal and anal fins. It is distinguished from P. snyderi by having a much deeper body- mean body depth of 619 compared to 427 for P. snyderi, each in thousandths of SL; an eye diameter greater than the length of the snout; a mean number of dorsal fin- rays of 42.8 compared to 46.5 for P. snyderi; and 31 to 33 rather than 36 total vertebrae. It is distinguished from P. simillimus by having a much deeper body - mean body depth of 619 compared to 461 for P. simillimus; an eye diameter greater than the length of the snout; a mean num- ber of anal fin-rays of 43.0 compared to 39.5 for P. simillimus; and 31 to 33 rather than 30 or 31 total vertebrae. Characters that distinguish P. ovatus from P. medius are listed in the diagnosis of the latter species. The specific epithet, ovatus, is from the Latin and is descriptive of the short, deep body of the species. Description. Proportional measurements are given in Table 9 and meristic values in Table 10. Body ovate, deep, compressed; anterior dorsal profile moderately convex. Eye diameter greater than length of snout. Dorsal and anal fins slightly falcate, the longest dorsal ray three to six times the length of the shortest dorsal ray, the longest anal ray two to five times the length of the shortest anal ray. Dorsal fin with three or four (usually four) small, bladelike spines preceding the rays; anal fin with three or four (usually three) spines preceding the rays. Dorsal and anal fins originating at about the same level anteriorly, and the bases about equal in length. Caudal fin long to very long, deeply forked, about 24 to 35 per cent of total length. Lateral line of trunk highly arched 184 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Table 9. Proportional Measurements of Pep- RILUS ovatus. Symbols as in Table 7. Size range, 37.5-145.2 mm SL; mean size, 79.0 mm SL. Table 10. Meristic Values of Peprilus ovatus. Symbols as in Table 7. N R SE V In thousandths of SL: Head length Snout length Eye diameter Length of upper jaw Interorbital width Length of pectoral fin Predorsal distance I Predorsal distance II Preanal distance Maximum depth of body Least depth of caudal peduncle In thousandths of HL: Snout length Eye diameter Length of upper jaw Interorbital width 158 158 158 158 158 154 158 158 158 163 259-325 296 052-074 063 078-115 093 065-094 080 080-106 092 267-441 360 369-465 408 259-347 298 427-552 481 542-676 619 1.08 4.60 0.39 7.75 0.62 8.35 0.43 6.75 0.41 5.62 2.68 9.26 1.52 4.67 1.33 5.62 2.13 5.57 2.36 4.86 158 060-094 079 0.54 8.59 158 176-244 213 158 267-377 315 158 234-313 271 158 26.3-357 311 1.23 7.23 1.69 6.76 1.20 5.55 1.72 6.97 anteriorly, following dorsal profile poste- riorly. Subdermal canal system visible as indistinct lines on the body; dendritic canals on top of head and nape often ill-defined. A series of medium-sized, irregularly-spaced, slitlike pores frequently visible along dorsal and upper ventral surfaces. Premaxillary teeth slightly recurved, pointed, and simple. Swimbladder not seen in this species. Col- oration as described for the genus. Maxi- mum length probably 200 mm SL. Variation. Coefficients of variation (V) for proportional measurements range from about 4.6 to 9.3 (Table 9) and from 0.8 to 3.5 for meristic characters (Table 10). N R SE V Dorsal fin-rays* Anal fin-rays* Pectoral fin-rays Total gill rakers Total vertebrae Lateral-line scales 145 40-46 42.8 0.12 3.27 147 40-46 43.0 0.12 3.40 150 19-23 21.4 0.06 3.50 146 21-26 23.6 0.07 3.35 157 31-33 32.0 0.02 0.81 * Excluding spines. These values indicate that the relative dispersion of character values is nearly as great as that of its presumed closest relative, P. medius. It might be expected that such a species of restricted distribution would be less variable than that of the wide-ranging species from which it is de- rived. The sample size is considered to be reasonably adequate. Except for vertebral number, ranges of meristic values appear, by inspection, to approach a normal dis- tribution (Tables 1 to 4). Vertebral num- ber, as in all the species of Peprilus, is relatively constant (Table 5). Geographic variation. No geographic variation was found in P. ovatus because of its apparently restricted distribution. Ontogenetic change. Few and moderate changes accompany growth in P. ovatus in a size range of about 40 to 145 mm SL. With growth, eye diameter and head length decrease in size relative to SL, while the length of the pectoral fin and the depth of the caudal peduncle increase in size rela- tive to SL (Fig. 11). The length of the pectoral fin, however, becomes nearly con- stant beyond about 100 mm SL. In con- trast to the situation in the related P. medius, body depth in P. ovatus decreases in size relative to SL with growth (Fig. 12). The anterior lobes of the dorsal and SYSTEMATICA AND BlOLOGY OF PePRILVS • Hom 185 33 29 25 11 .. . • '•. Head Len gth - • - " • • • • • 1 1 i • 1 - • • •• • • • • • • Eye D • ■ ame ter - • • * • • • « • . • •• • • • * •. •• I ' i • -. 44 "S 38 32 26 Length of Pectoral Fin • • • 9 • . .* • - 8 • • •• • • ••• • • • • ft • •• • • ••'. ^ •• • • • / •• • • - • • Dep th of Cauda 1 Peduncle 6 • ' i i - 25 50 75 100 125 Standard Length mm 150 Figure 11. Ratio-on-size scatter diagrams for four morpho- metry characters of Peprilus ovatus. anal fins tend to slightly increase in length with growth (compare Figs. 10 and 18b). Distribution (Fig. 7). P. ovatus is known only from the northern Gulf of California from Punta Diggs south of San Felipe, Baja California, northward to the mouth of the Colorado River and from Golfo de Santa Clara, Sonora, to the mouth of the Colorado River. i /\ 1 -V *-".J ■■ ■ i ■: * STANDARD LENGTH Figure 12. Ratio-on-size scatter diagram of body depth for Peprilus medius and Peprilus ovatus. Taxonomic comments. See taxonomic comments under P. medius. Peprilus simillimus (Ayres, 1860) Figures 13, 18c Poronotus simillimus Ayres, 1860: 84, fig. 1 (original description, San Francisco, California, holotype, a CAS specimen apparently destroyed by fire in 1906). Stromateus simillimus, Jordan and Gilbert, 1881: 46; Jordan and Gilbert, 1882a: 451; Fordice, 1884: 314. Rhombus (Palometa) simillimus, Jordan and Ever- mann, 1896: 967, Palometa erected as a sub- genus. Palometa simillima, Jordan and Evermann, 1898: 2849, subgenus Palometa elevated to generic level; Jordan, Evermann, and Clark, 1930: 266; Roedel, 1953: 81, fig. 78; Batts, 1960: 146, occurrence in Washington; Berry and Perkins, 1966: 673, distribution in California Current area; High, 1966: 53, fig. 1, occur- rence in Puget Sound. Peprilus simillimus, Clemens and Wilbv, 1946: 201, fig. 38; J. L. Hart, 1949: 101; Haedrich, 1967: 107. Material examined. Those specimens marked with an asterisk (*) have been radiographed. The number radiographed equals the number measured unless other- wise indicated. CS indicates specimens cleared and stained. Size ranges, in mm, are standard lengths (SL). *BC 62-242 (6:151.5-169.5 mm, British Columbia, Strait of Georgia, Pender Harbor and 186 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Bargain Harbor, 4 radiographed); *BC 63-1447 (1:143.0, British Columbia, San Juan de Fuca Strait); CAS 1546 (1:168.5, San Francisco, Calif.); *CAS 1552 (1:157.0, San Francisco, Calif.); CAS 21929 (1:92.3, San Francisco Bay, Calif., 4 May 1955); *CAS 21943 (1:105.9, San Francisco Bay, Calif., 22 May 1955); *LACM 401 (11 meas.: 98.2-131.8, Calif., Los Angeles County, Long Beach, Belmont Shores, 25 October 1957, 58 radiographed); *LACM 3000 (1:76.2, Calif., Los Angeles Co., Long Beach, Belmont Shores, 19 October 1962); *LACM 21044 (1:68.2, Calif., Los Angeles Co., Venice Pier, 15 August 1914); MCZ 16885 (2:118.0, 121.5, San Francisco, Calif.); MCZ 17337 (1:192.5, San Fran- cisco, Calif.); MCZ 23394 (1:202.0, San Francisco, Calif.); MCZ 23395 (1:168.0, San Francisco, Calif.); MCZ 23399 (1: 174.0, San Francisco, Calif.); *SIO H52- 221 (4:28.4-49.8, Pacific Ocean, Baja Calif., San Martin I., Hassler Cove, found in bell of jellyfish, Pelagia sp., 14 Septem- ber 1952, 1 radiographed, 1 CS); *SIO 59- 70 (1:71.0, Pacific Ocean, Baja Calif., 26°4S'N, 113°25'W, 3 July 1958); *SIO 60- 394 (1:86.4, Pacific Ocean, Baja Calif, Se- bastian Vizcaino Bay, 28°34.5'N, 114°24'W, 19 August I960); *SIO 60-409 (1:48.0, Calif, San Diego Co, La Jolla Bay, found in bell of jellyfish, Pelagia sp, 6 Julv 1960); *SIO 62-236 (1:137.5, Gulf of Calif, Baja Calif, Bahia de Los Angeles, 28°55'-56'N, 113°31'-33'W, 20 m, 26 April 1962); *SIO 62-580 (1:68.7, Calif, Imperial Co, on beach at Salton Sea, apparently a live-bait transfer, 3 September 1962); *SIO 62-681 (20 meas.: 78.7-144.9, Calif, off San Diego, 32°46.5'N, 117°18.5'-19.7'W 0-82 m, 1 September 1962, B/V JOHN N. COBB, Cobb Mark II pelagic trawl, 23 radio- graphed); *SIO 64-641 (5:107.4-129.0, Pacific Ocean, Baja Calif, E of Cedros I, 2 miles from shore, 15 m, 17 August 1961); *SIO 64-747 (2:78.2, 86.0, Pacific Ocean, Baja Calif, Bahia Magdalena, from bait sample, 7 September 1954, M/V MARY C. CANAS); *SU 4382 (6:96.3-158.5, Monterey Bay, Calif.); SU 48000 (1:85.0, off Oceanside, Calif, 17 November 1943); *USNM 4472 (1:172.0, San Francisco, Calif.); *USNM 26800 (4:101.0-143.0, San Diego, Calif.); *USNM 26910 (4:113.0- 144.0, Santa Barbara, Calif.); *USNM 38015 (1:108.0, San Diego, Calif.); *USNM 42045 (2:135.4, 150.0, Wash, Seattle market); *USNM 52998 (1:168.0, Calif, San Pedro market); *UCLA W56-360 (4: 79.0-96.4, Pacific Ocean, Baja Calif, Turtle Bay, 27°41'N, 114°53'W, 0-7 m, 22 July 1956, bait net); *UW 13405 (4:140.3-153.3, Wash, Strait of Georgia, Bellingham Bav, 8 February 1946); *UW 14656 (9:164.7- 198.0, Wash, Puget Sound, Fletcher Bay, 11 December 1959); *UW 18210 (6:119.3- 152.4, Wash, Orcas I, East Sound, 0-27 m, 10 September 1963). Diagnosis. P. simillimus is an elongate, shallow to moderately deep species with only slightly falcate dorsal and anal fins. The species is distinguished from P. snyderi by having a mean number of dorsal fin-rays of 44.8 compared to 46.5 for P. snyderi; a mean number of anal fin-rays of 39.5 com- pared to 42.0 for P. snyderi; and 30 or 31 rather than 36 total vertebrae. Characters that distinguish P. simillimus from P. medius and P. ovatus are listed in the diagnosis of each of these species, respec- tively. P. simillimus is distinguished from the similar P. triacanthus by having no row of relatively large pores beneath the an- terior half of the dorsal fin. The specific epithet, simillimus, is from the Latin, "similis," meaning like or resembling; the species was so named because of its similarity to P. triacanthus. Description. Proportional measurements are given in Table 11 and meristic values in Table 12. Body elongate, shallow to moderately deep, compressed; anterior dor- sal profile slightly convex. Eye diameter about equal to the length of the snout; eye relatively small. Dorsal and anal fins slightly falcate, the longest dorsal ray three to five times the length of shortest dorsal ray, longest anal ray two to four times the Systematics and Biology of Peprilvs • Horn 187 Figure 13. Peprilus simillimus, 164.7 mm SL, Puget Sound, Washington, UW 14656. length of shortest anal ray. Dorsal with two to four (usually three) small spines preceding the rays; anal with two or three (usually three) spines anterior to the rays. Base of anal fin shorter than base of dorsal (fewer rays). Caudal fin long, moderately to deeply forked, about 18 to 32 per cent of total length. Subdermal canal system visible as indistinct parallel lines on body in preserved material; dendritic canals on top of head and nape distinct. Premaxil- lary teeth slightly recurved, mostly pointed, simple, but a few on a particular specimen may have two or three small cusps. Swim- bladder not seen in this species. Coloration as described for the genus. Maximum length probably 280 mm SL. Variation. Except for eye diameter, coefficients of variation (V) for propor- tional measurements range from 3.8 to about 9 (Table 11) and from 1.2 to 4.2 for meristic characters (Table 12). The V value for eye diameter as a percentage of SL is 17.44, a high value resulting from allo- metric growth. The sample size (Table 11) of P. similli- mus used is considered to be reasonably adequate, and specimens from almost all parts of the known range have been ex- amined. Except for vertebral number, ranges of meristic values appear, by in- spection, to approach a normal distribution (Tables 1 to 4). Vertebral number, as in all species of Peprilus, is relatively constant (Table 5). Geographic variation. There is slight but inconclusive evidence that northern mem- bers of P. simillimus are larger and deeper- bodied. The dorsal, anal, and caudal fins of northern specimens seem to be shorter and less conspicuous. The differences are slight in the dorsal and anal fins, but more prominent in the caudal fin. Length of the caudal fin expressed as a percentage of the total length in specimens from San Francisco to British Columbia is less on the average than in specimens taken from Monterey to Baja California (Fig. 14). Although the ranges of the two groups overlap, the difference in the means is highly significant with a probability of much less than 0.001 using a two-tailed Student's Mest. The functional significance of shorter or lower fin lobes in northern populations of fish species is apparently unknown. Per- 188 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Table 11. Proportional Measurements of Peprilus simillimus. Symbols as in Table 7. Size range, 28.4-202.0 mm SL; mean size, 123.8 mm SL. Table 12. Meristic Values of Peprilus simil- limus. Symbols as in Table 7. N R SE V In thousandths of SL: Head length 109 252-408 Snout length 109 058-081 Eye diameter 109 050-119 Length of upper jaw 109 064-100 Interorbital width 109 076-110 Length of pectoral fin 109 272-379 Predorsal distance I 103 338-407 Predorsal distance II 109 222-339 Preanal distance 109 402-518 Maximum depth of body 109 394-517 Least depth of caudal peduncle In thousandths of HL: Snout length 109 192-287 Eye diameter 109 188-347 Length of upper jaw 109 218-323 Interorbital width 109 259-369 285 2.28 8.35 070 0.45 6.74 073 1.22 17.44 076 0.64 8.82 089 0.53 6.26 336 2.28 7.41 369 1.37 3.76 263 1.82 7.23 440 2.01 4.78 461 2.01 4.57 109 055-074 065 0.43 6.83 246 2.11 8.95 255 2.62 10.71 267 1.63 6.39 314 2.07 6.89 haps the difference is allied to differential growth rates as influenced by regional conditions of temperature. The shorter fin lobes may possibly be associated with sur- face-volume ratios and with adaptations to cooler temperatures. Ontogenetic change. Few and moderate changes accompany growth in P. similli- mus in a size range of about 28 to 202 mm SL. With growth, head length and eye diameter decrease in size relative to SL while depth of body and of caudal peduncle remain nearly constant (Fig. 15). The correlation coefficient of the size-on- size regression of eye diameter is the lowest among members of the genus ( Fig. N R SE V Dorsal fin-rays* Anal fin-rays* Pectoral fin-rays Total gill rakers Total vertebrae Lateral-line scales 98 41-48 44.8 0.16 3.50 100 35-44 39.5 0.15 3.67 107 19-23 21.1 0.06 3.13 75 23-26 23.9 0.12 4.18 160 30-31 30.2 0.03 1.19 9 95-110 - * Excluding spines. 3; Table 6). The length of the pectoral fin increases slightly in size relative to SL with growth up to about 125 mm SL, bevond which there is a slight decrease (Fig. 15). Distribution (Fig. 7). P. simillimus is known along the Pacific Coast of North America from southern Baja California to southern British Columbia. The southern limit of the range is in the vicinity of Bahia Magdalena, Baja California, at about 24°30'N. One specimen here referred to as P. simillimus is known from the Gulf of California: a female, 137.5 mm SL, from Bahia de Los Angeles, Baja California, 28°55'-56'N, 113°31'-33'W (SIO 62-236). The northern known extents of the range are near Clayoquot Sound on the west Coast of Vancouver Island, near Nanaimo on the east coast of the Island, and near the mouth of the Fraser River at Van- couver, British Columbia, all at about 49°N. Taxonomic comments. Ayres (1860) in his original description of this species placed it "with doubt" in the genus Poro- notus, noting that it lacked the relatively large dorsal pores characterizing Poronotus triacanthus. Until he received Gill's (1861) "Catalogue of the fishes of the eastern coast of North America from Greenland to Georgia," Ayres had recorded the new Systematics and Biology of Peprilus • Horn 189 i i I I i i A(53) 10 _r 5 x = 21.3 c V 3 u_ ^S x = 25.1 B(47) ■ i i i i i 5 10 1 Caudal Fin Length {7. of Total Length) Figure 14. Geographic variation of caudal fin length in Peprilus simillimus: A =z specimens collected between British Columbia and San Francisco; B = specimens collected be- tween Monterey, California, and southern Baja California; numbers in parentheses are the sample sizes; x =z mean. The means are highly significantly different at the 2 per cent level (two-tailed Student's f-test) . species in his manuscript notes as Peprilus simillimus. Since Gill reserved the genus Peprilus for the deep-bodied species with falcate median fins, and the genus Poron- otus for the elongate species with relatively large dorsal pores, Ayres with reluctance placed the new species in the genus Poronotus because of the overall similarity to Poronotus triacanthus. This species is often referred to in faunal and fishery works as Palometa simillimo, after Jordan and Evermann's ( 1898 ) desig- nation. The common names of P. simillimus on the Pacific Coast, "California pompano," "Pacific pompano," or "pompano," are somewhat misleading since "pompano" is most frequently used in reference to cer- tain fishes of the family Carangidae. A more appropriate name might be "Pacific butterfish" after the familiar butterfish, P. triacanthus, of the Atlantic Coast. Peprilus snyderi Gilbert and Starks, 1904 Figures 16, 18d Peprilus snyderi Gilbert and Starks, 1904: 87, fig. 23, pi. XII (original description, Panama Head Length H .• • • • - - - .• Eye Diameter - - • • 4 • • - - * •• • * • *« - - 1 i i * • •• " ' ' ;- ::'V;. . Body Depth _l l_ 100 125 150 Standard Length mm Figure 15. Ratio-on-size scatter diagrams for five morpho- metric characters of Peprilus simillimus. Bay, holotvpe not seen; three paratypes seen, 2:188.0, 193.0 mm SL. SU 7009; 1:182.3, USNM 50448); Meek and S. F. Hildebrand, 1925: 413; Haedrieh, 1967: 107. Palometa snyderi, Jordan, Evermann, and Clark, 1930: 267. Material examined. Those specimens marked with an asterisk (*) have been radiographed. The number radiographed equals the number measured unless other- 190 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Figure 16. Peprilus snyderi, 227.0 mm SL, Acapulco market, Mexico, BC 61-126. wise indicated. CS indicates specimens cleared and stained. Size ranges, in mm, are standard lengths (SL). *BC 59-247 (1:205.0 mm, Mazatlan, Mexico); *BC 59- 687 (3:190.0-213.0, Panama City market); *BC 60-25 (1:193.0, Mex., Golfo de Tehuantepec, Puerto Arista to Salina Cruz); *BC 61-126 (1:227.0, Mex., Aca- pulco market); *FMNH 73840 (1:227.0, Mex., 1-5 miles off Chiapas coast at Guatemala border to above San Benito, 18-27 m, 14-18 December 1954); IMS 1310 (1:143.6, Mex., Salina Cruz market, 16 August 1963); *LACM 8859-2 (3: 152.5-159.2, Gulf of California, 24°29'N, 110°26'W, 0-15 m captured in 82-110 m water, 6 April 1964, R/V ALASKA); *SIO H50-181 (3:238.0-245.0, Pacific Ocean, Baja Calif., Bahia San Juanico, San Juanico Anchorage, 26°15'N, 112°28'W, 12 April 1950, snag hooks); *SU 7009 (2:188.0, 193.0, paratypes, Panama Bay); *USNM 50448 (1:182.3, paratype, Panama Bay); *UCLA W53-267 (1:232.0, Panama City market, 23 June 1953); *UCLA W55-23 (1:179.0, Gulf of Calif., Baja Calif., 1 mile S of San Felipe, 0-7 m, 9-11 March 1955, 600 ft beach seine); *UCLA W56-113 (1: 121.2, Gulf of Calif., Sinaloa, Mex., just S of mouth of Bahia Topolobampo, 6-11 m, 24 May 1956); *UCLA W56-129 (4:42.1- 65.7, Jalisco, Mex., Bahia de Banderas, 20°32'N, 105°17'W, 0-7 m, 18 March 1956, bait net, 1 CS). Diagnosis. P. snyderi is an elongate, wide-ranging, rarely-collected species with slightly falcate dorsal and anal fins. Char- acters that distinguish P. snyderi from P. medius, P. ovatus, and P. simillimus, are listed in the diagnosis of each of these species, respectively. Description. Proportional measurements are given in Table 13 and meristic values in Table 14. Body elongate, shallow to moderately deep, compressed; anterior dor- sal profile slightly convex. Eye diameter about equal to the length of the snout; eye relatively small. Dorsal and anal fins slightly falcate, the longest dorsal ray three to five times the length of the shortest dorsal ray, the longest anal ray two to five times the length of the shortest anal ray. Dorsal and anal with two or three (usually three) small spines preceding the rays; spines appear to be somewhat smaller and less conspicuous than in other members of the genus. Base of anal fin shorter than base of dorsal (fewer rays). Caudal fin long, moderately to deeply forked, about 20 to 30 per cent of total length. Scales Systematics and Biology of Peprilus • Horn 191 Table 13. Proportional Measurements of Pep- rill s SNYDERI. Symbols as in Table 7. Size range, 42.1-245.0 mm SL; mean size, 170.0 mm SL. Table 14. Meristic Values of Peprilus sny- deri. Symbols as in Table 7. N R SE V In thousandths of SL: Head length Snout length Eye diameter Length of upper jaw Interorbital width Length of pectoral fin Predorsal distance I Predorsal distance II Preanal distance Maximum depth of body Least depth of caudal peduncle In thousandths of HL: Snout length Eye diameter Length of upper jaw Interorbital width 24 262-380 24 059-083 24 057-103 24 066-107 24 079-104 22 289-361 24 332-430 24 224-315 24 367-505 24 371-503 297 6.97 11.50 068 1.23 8.85 070 2.83 19.80 081 2.17 13.14 087 1.14 6.43 329 3.34 4.78 364 6.18 8.32 253 5.79 11.22 424 7.38 8.53 427 7.35 8.43 24 050-082 071 1.75 12.10 24 191-250 24 209-279 24 250-290 24 217-322 2.30 2.84 6.05 234 4.40 9.21 272 1.96 3.53 295 5.19 8.62 apparently somewhat less deciduous than in other members of the genus. Subdermal canal system visible as indistinct parallel lines on body in preserved material; dendritic canals on top of head and nape distinct. Premaxillary teeth slightly re- curved, mostly pointed, simple, but a few on a particular specimen may have two or three small cusps. Swimbladder not seen in this species. Coloration as described for the genus. Maximum length probably 300 mm SL. Variation. Except for eye diameter, co- efficients of variation (V) for proportional measurements range from 3.5 to 13.1 N R SE Dorsal fin-rays* Anal fin-rays* Pectoral fin-rays Total gill rakers Total vertebrae Lateral-line scales 24 43-49 46.5 0.28 2.88 23 40-44 42.0 0.30 3.35 22 21-23 22.3 0.15 3.14 18 23-26 24.1 0.24 4.23 23 36 36.0 0.00 0.00 12 110-130 - * Excluding spines. (Table 13). The V values for meristic characters range from 2.9 to 4.2, excluding vertebral number which showed no vari- ation (Table 14). The V value for eye diameter as a percentage of SL is 19.8. This range of V values is wider than for other species and may be at least partially explained by the small sample size (24), which consisted of several large fish and a few small fish. Geographic variation. No geographic variation was noted in this species. Ontogenetic change. Few and moderate changes accompany growth in P. snyderi although the sample, 24 specimens be- tween 42 and 245 mm SL, was small and lacked individuals between 60 and 120 mm. With growth, head length, eye diameter, and body depth decrease in size relative to SL while the relative depth of the caudal peduncle increases slightly and the relative length of the pectoral fin re- mains nearly constant (Fig. 17). Distribution (Fig. 7). P. snyderi is known along the Pacific Coast from Panama Bay northward to the upper part of the Gulf of California near San Felipe (about 31°N) and to Bahia San Juanico (about 26°15'N) on the outer coast of Baja Cali- fornia. Collections from intermediate lo- calities are few, but the species probably occurs along this entire coastal region. 192 Bulletin Museum of Comparative Zoology, Vol. 140, No. Stondofd Length Figure 17. Ratio-on-size scatter diagrams for five morpho- metric characters of Peprilus snyderi. Peprilus burti Fowler, 1944 Figures 19, 31 Poronotus triacanthus (not of Peck, 1804), Fowler, 1933: 61, Breton Island, Louisiana ( misidentification ) ; Briggs, 1958: 292, in part, Gulf of Mexico populations; Caldwell, 1961: 19, pi. 10, in part, bottom fish in photograph, Gulf of Mexico Peprilus burti synonymized with Atlantic Poronotus triacanthus. Peprilus burti Fowler, 1944: 1, fig. 1 (original description, Breton Island, Louisiana, holotype seen, 113.3 mm SL, ANSP 55841; five para- types seen, 2:73.3, 82.0, ANSP 55842-55843; 3:66.5-71.5, Galveston, Texas, ANSP 70943- 70945); Haedrich, 1967: 108. Poronotus burti, Collette, 1963: 582, valid species based on vertebral number. Material examined. Those specimens marked with an asterisk (*) have been radiographed. The number radiographed equals the number measured unless other- wise indicated. CS indicates specimens cleared and stained. Size ranges, in mm, are standard lengths (SL). All localities are in the Gulf of Mexico. *ANSP 55841 (1:113.3 mm, holotype of Peprilus burti Fowler, 1944, Breton I., Louisiana, Novem- ber, 1930); *ANSP 55842, 55843 (2:73.3, 82.0, para types of Peprilus burti Fowler, 1944, Breton I., La., November, 1930); *ANSP 70943 to 70945 (3:66.5-71.5, para- tvpes of Peprilus burti Fowler, 1944, Gal- veston, Texas, 1943); FSBC 372 (10:34.4- 47.3, Florida, Escambia County, E of Pensacola Bav Bridge, 1 February 1958); FSBC 1640 (1:44.5, Fla., Pinellas Co., Madeira Beach, February, 1960); FSBC 2283 (6:96.2-106.1, Mississippi, Horn I., 27 October 1962); FSBC 3211 (1:63.6, Fla., Tampa Bay, grass flats, 9 October 1964); FSBC 3832 (1:71.2, Fla., Tampa Bay, 23 February 1966); FSBC 3954 ( l:86.0,>la., Pinellas Co., 27 December 1966, R/V CORTEZ); IMS 324 (1:80.5, Bayou, 11-15 m, 23 October 327 (1:24.8, Port Aransas, Tex. lophus medusa, 27 November 1950; 1:28.2, Port Aransas, Tex., tide trap, 16 June 1950; 2:29.5, 30.0, Port Aransas, Tex., Humble Docks, dipnet, 1 Mav 1948); IMS 328 (5:79.0-88.8, 60-70 miles S of Port Aransas, Tex., 33-42 m, 13-16 Mav 1951); IMS 329 (1:120.8, 100 miles S of Port Isabel, Tex., 57-68 m, 6-11 March 1951; 2:82.5, 109.0, 40-60 miles S of Port Aransas, Tex., 22-33 m, 6-11 July 1951); IMS 330 (1:122.6, Mexico, Punta Frontera, 24-37 m, 29 July to 7 August 1951; 1:128.8, Tex., Corpus Christi Pass, 33-42 m, 22-23 January 1951); IMS 981 (3:100.8-107.9, Port Aransas, Tex., 37 m, 24 July 1962); IMS 1179 (3: 98.7-105.3, Port Aransas, Tex., 10 July 1962); IMS uncat. (8:28.4-88.4, Port Aransas, Tex., spring, 1964); *MCZ 35118 (4:68.0-76.0, Pensacola, Fla.); MCZ 41924 (4:6.7-9.0, 28°51'N, 88°37'W, at surface over 695 m, M/V OREGON Sta. 852, 25 October 1953, 2 CS); MCZ 45015 (10: Tex., Green 1950); IMS from Stoma- Systematics and Biology of Peprilus • Horn 193 Figure 18. Juveniles of the four Pacific species of Peprilus-. a c = P. simillimus, 71.0 mm SL; d =: P. snyderi, 66.4 mm SL. = P. medius, 77.9 mm SL; b = P. ovofus, 60.0 mm SL; 113.5-136.0, 29°45'N, S8°20rw, 37 m, M/V SILVER BAY Sta. 296, 12 March 1958); *MCZ uncat. (22:55.6-102.2, Port Aransas, Tex., 7-20 m, R/V LORENE, 30 June 1967, 16 radiographed); TABL uncat. (1:136.4, 27°04'N, 96°43'W, 82 m, M/V SILVER BAY Sta. 266, 28 January 1958); TABL uncat. (1:133.5, 20°01'N, 91°47'W, 55 m, M/V SILVER BAY Sta. 844, 17 No- vember 1958); TABL uncat. (4:44.0-52.0, Fla., lower Pensacola Bay, between ship channel and south shore from Big Lagoon entrance to U. S. Coast Guard Station, 20 February 1964); TABL uncat. (1:133.7, Fla., 45 miles SW of Pensacola); TU 3979 (3:86.0-104.5, La., Terrebonne Co., Oyster Bayou, 2-4 m, 31 May 1952); TU 4084 (10:38.5-56.2, La., Terrebonne Co., Oyster Bavou 4 m, 12 March 1952); TU 4461 (10:12.5-17.5, La., Jefferson Co., NE end of Grand I., 2-3 May 1952); TU 8945 (7:78.7-122.9, La., Miss. Sound near Grand I. channel, 30°09'N, 89°18'W, M/V ANN INEZ, 26 August 1954); TU 23966 (10: 16.3-30.6, 29°50.5'N, 85°32.5'W, 22 m, M/V OREGON Sta. 2431, 4 March 1959); *USNM 118620 (4:72.0-79.0, Aransas Pass, Tex., 12 May 1940); USNM 147777 (1: 109.0, Miss., 1948); USNM 156116 (1: 100.0, La., Grand I., Barataria Bay, 5 July 1930); *USNM 156118 (2:88.0, 91.0, La., Barataria Bav, 27 November 1951 ) ; USNM 156121 (1:97.0, La., Grand I., 5-8 m, 2 July 1930); USNM 156122 (1:97.0, La., Grand I., 21 Julv 1930); *USNM 156123 (5:94.0-107.0, La., Grand I., 12 miles off Grand Terre, 22 July 1930); *USNM 156124 (1:86.0, Fla., Appalachicola Bay, June, 1932); *USNM 157709 (2:99.0, 100.0, Alabama, 9 m, 29 September 1951); *USNM 159714 (3:56.0-73.0, 30°12.5'N, 88°15'W, Ala., 11 m, M/V OREGON Sta. 194 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 /O t»i 7*1 Figure 19. Pepn'lus burti, 113.7 mm SL, Mississippi Sound, 30°09'N, 89°18'W, TU 8945. 2142, 12 March 1958); *USNM 159808 (1:145.0, 28°58'N, 84°44'W, 144 m, M/V SILVER BAY Sta. 154, 22 August 1957); USNM 159825 (1:123.1, 28°30'N, 94°97'W, 35 m, M/V SILVER BAY Sta. 187, 25 September 1957); USNM 163517 (1:143.0, 28°09'N, 84°54'W, 146 m, M/V OREGON Sta. 920, 11 March 1954); *USNM 187922 (4:142.0-167.0, 29°04'N, 85°49'W, 183-186 m, M/V SILVER BAY Sta. 156, 22 August 1957); *USNM 187924 (1:54.2, 29°42'N, 88°40/W, 15-18 m, M/V OREGON Sta. 2193, 25 May 1958); *USNM 187987 (6 meas.: 69.0-93.0, 29°07'N, 88°34'W, 229 m, M/V OREGON Sta. 3646, 14 June 1962, 11 radiographed); USNM 188649 (2:103.0, 104.0, 28°47'N, 91°49.5'W, 26m); USNM uncat. (4:32.0-63.0, 30o09yN, SS^W, 13 m, M/V OREGON Sta. 2115, 24 February 1958); *USNM uncat. (3:58.0-76.0, 30° 12.5'N, 88°15'W, 11 m, M/V OREGON Sta. 2142, 12 March 1958); *USNM uncat. (31 meas. =53.9-102.1, 30°02'N, 88°42.5'W, 15-17 m, M/V OREGON Sta. 2389, 29 January 1959, 103 radiographed, 2 CS); *USNM uncat. (10:132.0-155.0, Yucatan Peninsula, W of Campeche, 19°42'N, 91° 47'W, 49-55 m, M/V SILVER BAY Sta. 1131, 25 April 1959); UMML 750 (4:53.5- 68.4, Fla., Appalachicola Bay, 1951); UMML 1804 (1:61.3, Yucatan Peninsula, Campeche, SW of Monas Pt.); UMML 11002 (1:130.8, 28°13.5'N, 92C56'W, 68 m, M/V OREGON Sta. 3804, 14 September 1962); UMML 13268 (1:83.3, 30°07.5'N, 88°42.5'W, 16 m, M/V SILVER BAY Sta. 5006, 22 June 1963); UMML 15813 (1: 65.9, 30°07.5'N, 88°43'W, 15 m, M/V SILVER BAY Sta. 5007, 22 June 1963). Diagnosis. P. burti is a moderately elon- gate species with slightly falcate dorsal and anal fins. It is distinguished from the closely related P. triacanthus by having a deeper body -a mean body depth of 551 compared to 458 for P. triacanthus, each in thousandths of SL; 16 or 17 rather than 18 to 20 caudal vertebrae; and, no dark spots along dorsal surface as are frequently present in P. triacanthus. The species is distinguished from P. paru by having a Systematics and Biology of Peprilus • Horn 195 Table 15. Proportional Measurements of Pep- rilus BVRTi. Symbols as in Table 7. Size range, 7.8—167.0 mm SL; mean size, 79.5 mm SL. Table 16. Meristic Values of Peprilus burti. Symbols as in Table 7. N R SE In thousandths of SL: Head length 216 Snout length 215 Eye diameter 229 Length of upper jaw 215 Interorbital width 215 Length of pectoral fin 219 Predorsal distance I 205 Predorsal distance II 215 Preanal distance 215 Maximum depth of body 232 Least depth of caudal peduncle In thousandths of HL: Snout length 215 Eye diameter 216 Length of upper jaw 215 Interorbital width 214 255-392 303 048-096 069 065-144 100 074-128 089 059-131 097 228-391 332 331-529 395 239-448 292 404-582 465 460-640 551 1.78 8.66 0.58 12.26 0.97 14.68 0.63 10.39 0.65 9.88 3.03 13.49 2.59 9.39 2.72 13.66 2.38 7.52 2.45 6.76 215 050-096 075 0.52 10.08 167-278 226 251-424 328 200-371 295 200-411 319 1.28 8.31 2.32 10.40 1.47 7.33 2.02 9.29 row of relatively large pores below the anterior half of the dorsal fin, and by hav- ing slightly falcate dorsal and anal fins rather than moderately or extremely falcate ones. Description, Proportional measurements are given in Table 15 and meristic values in Table 16. Body moderately elongate, moderately deep to deep, compressed; an- terior dorsal profile moderately convex. Eye diameter greater than length of snout; eye relatively large. Dorsal and anal fins slightly falcate, longest dorsal ray three to five times the length of shortest dorsal ray, longest anal ray two to four times the N R SE Dorsal fin-rays* Anal fin-rays* Pectoral fin-rays Total gill rakers Total vertebrae Lateral-line scales 155 38-48 43.6 0.12 3.53 155 35-43 39.7 0.11 3.35 148 19-23 21.4 0.06 3.59 71 21-26 23.3 0.11 3.91 273 29-31 30.0 0.01 0.67 4 90-100 - * Excluding spines. length of shortest anal ray. Dorsal with two to four (usually three) small spines preceding the rays; anal with two or three (usually three) small spines preceding the rays. Base of anal fin shorter than base of dorsal fin (fewer rays). Caudal fin long to very long, deeply forked, about 25 to 35 per cent of total length. Subdermal canal system usually indistinct in preserved specimens, seen as vertical, parallel lines on body and dendritic canals on top of head and nape. The system is often very conspicuous in living or freshly collected specimens. There is an irregular row of about 17 to 25 relatively large pores just below the anterior half of the dorsal fin. Premaxillary teeth slightly recurved, usu- ally with three small cusps. Swimbladder delicate, thin-walled, elongate; found only in specimens smaller than about 100 mm SL. Coloration as described for the genus. Maximum length probably 200 mm SL. Variation. Coefficients of variation (V) for proportional measurements range from 6.8 to 14.7 (Table 15) and from 0.7 to 3.9 for meristic characters (Table 16). Com- paratively high V values for proportional measurements result partly from some de- gree of allometry, especially in eye size, and from the inclusion in the sample of a wide size-range of individuals. The sample size of P. burti used (Table 196 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Head Length 65 •J • \ * Body Depth ■n »' • - • • . • • O • • • • M 57 " • • **"".. m a c 53 - • • • i • • • • • • - a. • • 49 - • • - 45 i i i i i i 80 105 Standard Length mrt Figure metric 20. Ratio characters ■size scatter diagrams for five morpho- Peprilus burti. 15) is considered to be reasonably ade- quate, and specimens from all parts of the known geographic range were examined. Ranges of meristic values appear, by in- spection, to approach a normal distribution (Tables 1 to 4). The main exception is vertebral number which, as in all the species of Peprilus, is relatively constant (Table 5). Geographic variation. No geographic variation was noted in this species. Ontogenetic change. Few and moderate changes accompany growth in P. burti in a size range of about 8 to 167 mm SL. As shown by ratio-on-size diagrams (Fig. 20), head length, eye diameter, and body depth decrease in size relative to SL with growth, while the relative length of the pectoral fin increases up to about 100 mm SL then becomes nearly constant; the relative depth of the caudal peduncle remains nearly constant. The correlation coefficient of the size-on-size regression of eye diameter is relatively low (Fig. 3; Table 6). The size- on-size regressions of body depth for P. burti and the closely related P. triacanthus (Fig. 2) show the two to be similar in this character up to about 30 mm SL, beyond which size P. burti becomes deeper-bodied. Distribution (Fig. 7). P. burti is known only from the Gulf of Mexico. The range extends from the vicinity of Tampa Bay, Florida (about 27°45'N), along the conti- nental shelf to the western side of Yucatan, Mexico (about 21 °45'N). Caldwell (1961) lists two records for the species outside of the above range: 1) Big Sarasota, Florida, about 27°30'N, CAS 17237; and 2) Cape Romano, Florida, about 25°54'N, M/V SILVER BAY Sta. 524, the latter based on an unsupported field report. Specimens agreeing closely to the descrip- tion of P. burti are known from shallow water in the Atlantic off Florida. These specimens may represent disjunct members of the species. All such specimens are, how- ever, listed under P. triacanthus, the closely related Atlantic Coast species. Systematics and Biology of Peprilus • Horn 197 Taxonomic comments. P. burti was originally described as a new species by Fowler ( 1944 ) from specimens which he had previously (1933) designated as Gulf of Mexico members of Poronotus tria- canthus. Fowler (1944) stated that the species was apparently related to Peprilus palomcta of the Pacific Coast even though he recognized that the former species was different in several respects, including the possession of a row of relatively large pores below the dorsal fin. P. burti is very closely related to P. triaconthus of the Atlantic Coast, a relation- ship resulting in different interpretations of taxonomic status. Populations in the Gulf of Mexico are frequently referred to as Peprilus or Poronotus triacanthus. Cald- well ( 1961 ) proposed the synonymy of Peprilus burti with Poronotus triacanthus. Collette (1963), on the basis of differing vertebral counts, considered P. burti to be valid. I consider it distinct but close to Peprilus triacanthus. The relationships and possible interactions of the two species are discussed on p. 247. Peprilus triacanthus (Peck, 1804) Figures 21, 22 Stromateus triacanthus Peck, 1804: 48, fig. 2, pi. 2 (original description, Piscataqua River, New Hampshire, holotype not seen); Giinther, 1860: 398; Jordan and Gilbert, 1882a: 451; Jordan and Gilbert, 1882b: 597. Stromateus cryptosus Mitchill, 1814: 3 (original description, New York Bay, holotype not seen); Mitchill, 1815: 365, fig. 2, pi. I; Cuvier and Valenciennes, 1833: 408. Rhombus cryptosus, Cuvier and Valenciennes, 1833: 408." Peprilus triacanthus, Storer, 1839: 60; Haedrich. 1967: 106, fig. 40. Rhombus triacanthus, DeKay, 1842: 137, fig. 80, pi. 26. Poronotus triacanthus, Gill, 1861: 35; Jordan and Gilbert, 1878: 377; Bean, 1880: 91; Jordan and Evermann, 1898: 2849, fig. 405, pi. CL (in part IV); Jordan, Everniann, and Clark, 1930: 267; Bigelow and Schroeder, 1953: 363, fig. 192; Briggs, 1958: 292, in part. Rhombus (Poronotus) triacanthus, Jordan and Evermann, 1896: 967, Poronotus designated as a subgenus. Material examined. Those spei marked with an asterisk (*) have beer radiographed. The number radiographei equals the number measured unless other- wise indicated. CS indicates specimens cleared and stained. Size ranges, in mm, are standard lengths (SL). ANSP 11350 (1: 95.6 mm, New Hampshire); ANSP 11353 (1:147.8, Maine); FSBC 1024 (2:81.6, 98.2, Atlantic Ocean, Florida, Duval County, jettv at Atlantic Beach, 29 November 1958); MCZ 2375 (1:185.0, Nahant, Mas- sachusetts); MCZ 16722 (1:197.5, Waquoit, Mass.); MCZ 16745 (1:152.5, Nahant, Mass.); MCZ 16815 (2:118.0, 119.0, Mai- den, Mass.); MCZ 16822 (1:168.5, Prov- incetown, Mass.); MCZ 16906 (1:67.0, Eastport, Me.); MCZ 16911 (2:116.5, 120.0, Tarry town, New York); MCZ 16939 (2: 50.0. 72.0. Penikese I., Elizabeth Is., Mass.); *MCZ 17017 (5:94.5-108.0, Hampton Roads, Virginia); MCZ 17173 (1:195.5, Martha's Vineyard, Mass.); MCZ 17175 (1:164.5, Trenton, New Jersey); MCZ 17207, 17208 (2:138.0, 194.5, Waquoit, Mass.); MCZ 17304 (1:166.5, Cape Cod, Mass.); MCZ 23406 (1:179.0, Nahant, Mass.); MCZ 23928 (1:186.0, New York Harbor, N. Y.); MCZ 25768 (16:10.6-24.6, Newport, Rhode Island, 6 CS); MCZ 34601 (3:57.0-68.0, ALBATROSS II 20852, Mass., 58 m, 14 July 1930); MCZ 34602 (2:26.5, 28.3, Nantucket I., Mass., 24 August 1925); MCZ 42122 (4:7.7-29.7, 39°28'N, 71°10'W, to 39°27'N, 71°26'W, 57 m, R/V AT- LANTIS RHB 604, IKMT, 22-23 July 1954, CS); MCZ 42123 (1:113.0, Port Royal, South Carolina); *MCZ 43227 (3:144.0- 156.0, Assateague I., Maryland, 10 July 1957); *MCZ 44974 (9:60.0-115.0, Georgia, off Jekyll Island, 25 March 1959, 1 CS); MCZ 44977 (1:121.0, 31°06'N, 79°56.5'W, 46-55 m, M/V SILVER BAY Sta. 5709, 8 May 1964); MCZ 45038 (4:76.5-175.0, Ga., commercial trawling area, 19 August 1959); MCZ 45117 (10:94.5-138.6, Point Judith, R. I., June, 1958); MCZ 45126 (2: 110.2, 121.9, North Carolina, 34°45'N, 76°21,W, 16 m, M/V SILVER BAY Sta. 198 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 lo min Figure 21. Peprilus triacanthus, 157.8 mm SL, fish market, Cambridge, Massachusetts. 1641, 24 February I960); MCZ uncat. (11:115.0-183.5, Waquoit, Mass.); MCZ uncat. (15:20.5-40.8, E of Long Island, N. Y., 40°46'N, 70°55'W, 5 m, EUGENIE VIII, Cr. 8, Sta. 1, 19-20 August 1961, plankton net on end of longline, 2 CS); *MCZ un- cat. (4:122.5-153.0, E of Long I., N. Y., 40°01'N, 71°23'W, bottom to 128 m, CAP'N BILL III, Cr. 65-1, Bait sta. 1, 5 May 1965); *MCZ uncat. (14:106.0-198.0, N. C, 36°30,N, 74°45'W, 110 m, M/V SIL- VER BAY Sta. 4104, 31 May 1962, 1 radio- graphed); *TABL 101984 (13 meas.:55.2- 138.8, Atlantic Ocean, 30°31'N, 81°22'W, 13 m, M/V SILVER BAY Sta. 3464, 5 October 1961, 17 radiographed); TABL 101990 (1:23.0, Atlantic Ocean, 30°15'N, 80°59'W, 29 m, M/V SILVER BAY Sta. 2795, 7 February 1961, larval trawl); TABL 101991 (1:21.0, Atlantic Ocean, 27°38'N, 80°02'W, 48-57 m, M/V SILVER BAY Sta. 5559, 8 March 1964, nekton net); TABL 101992 (3:9.6-19.8, Atlantic Ocean, 30°46'N, 80°48'W, 25 m, M/V SILVER BAY Sta. 2804, 9 February 1961, larval trawl); TABL 101993 (1:19.0, Atlantic Ocean, 27°43'N, 80°00.5'W, 57 m, M/V SILVER BAY Sta. 5560, 8 March 1964, 1 m, nekton net); TABL 102050 (1:15.0, Atlantic Ocean, 27°21.5'N, 79°57.5'W, 101 m, M/V SILVER BAY Sta. 5551, 7 March 1964); TABL uncat. (1:132.0, Atlantic Ocean, 29°48'N, 80°12'W, 348-366 m, M/V SILVER BAY Sta. 470, 17 June 1958); TABL uncat. (1:119.5, N. C, 34°17'N, 75°55'W, 183-201 m, M/V OREGON Sta. 4973, 29 July 1964); *USNM 311? (1:122.0, Atlantic Ocean, 29°38'N, 80°09'W, 320-329 m, M/V SILVER BAY Sta. 471, 17 June 1958); USNM 156113 (1:72.0, Cape Ken- nedy, Florida); *USNM 156149 (6 meas.: 54.0^79.0, Fernandina, Fla., March, 1920, 13 radiographed); *USNM 156153 (5:69.0- 76.0, Fernandina, Fla., March, 1918); *USNM 158133 (3:116.0-127.0, Atlantic Ocean, 28°03'N, 79°52'W, 274-320 m, M/V PELICAN Sta. 25, 8 April 1956); *USNM 159807 (1:131.0, Atlantic Ocean, 29°15'N, 80°05'W, 384 m, M/V COMBAT Sta. 329, 31 Mav 1957); USNM 159813 (2:73.0, 79.0, Atlantic Ocean, 30°24'N, 81°22'W, 11-15 m, M/V COMBAT Sta. 504, 2 October 1957); *USNM 159818 (1: 61.0, Atlantic Ocean, 29°30'N, 80°11'W, 137 m, M/V COMBAT Sta. 489, 19 August 1957); USNM 159823 (1:75.0, Ga., 31° 29'N, 79°33'W, 110 m, M/V COMBAT Sta. 512, 3 October 1957); USNM 188468 (4: System atics and Biology of Pepriias • Horn 199 Figure 22. Peprilus triacanthus, 122.0 mm SL, off eastern Florida at 320 to 330 m, USNM, SILVER BAY Station 471. 101.0-110.0, Atlantic Ocean 30°39'N, 76° 39'W, 229 m, M/V COMBAT Sta. 180, 16 November 1956); *USNM 190367 (10: 61.0-141.0, N. C, 35°01.5'N, 76°02.5'W, 9-15 m, M/V SILVER BAY Sta. 1487, 7 December 1959, 8 radiographed); USNM uncat. (5:95.0-111.0, Delaware, about 15 miles NE of Indian River Inlet, 16 m, 4-5 August 1958); *USNM uncat. (10:133.0- 180.0, Atlantic Ocean, 29°45'N, 80°10'W, 329-348 m, M/V SILVER BAY Sta. 489, 21 June 1958, 8 radiographed); *USNM uncat. (4:108.0-132.0, Atlantic Ocean, 29° 45'N, 80°20/W, 64-68 m, M/V SILVER BAY Sta. 5687, 1 May 1964); UMML 2479 (1:110.5, Atlantic Ocean, 30°02'N, 80° 06'W, 320 m, M/V PELICAN Sta. 49, 12 May 1956); UMML 2924 (1:169.0, Jack- sonville, Fla., to Brunswick, Ga., 37-91 m, January, 1956); UMML 4370 (2:105.1, 106.2, N. C, 34°07'N, 76°06'W, 229 m, M/V COMBAT Sta. 175, 15 November 1956); UMML 4497 (1:110.4, Fla., ESE off St. John's River, 1 May 1952); UMML 6937 (2:77.3, 81.7, Atlantic Ocean, St. Augustine, Fla., 7-13 m, 4 August 1960); UMML 7053 (1:111.8, N. C, 34°36'N, 76°13'W, 33-37 m, M/V SILVER BAY Sta. 1643, 24 February I960); UMML 7340 (10:76.7-98.8, N. C, 34°57'N, 76°03'W, 16-18 m, M/V SILVER BAY Sta. 1631, 23 February I960); UMML 7614 (1:144.6, S. C, 32°59.5'N, 79°14'W, 9-11 m, M/V SILVER BAY Sta. 1365, 21 October 1959); UMML 8102 (1:64.4, Atlantic Ocean, St. Augustine, Fla., 15-16 m, 29 November I960); UMML 8527 (1:138.3, Atlantic Ocean, St. Augustine, Fla., 14-15 Septem- ber 1960); UMML 15049 (1:108.2, Ga., 31°26'N, 79°44'W, 84-102 m, M/V SIL- VER BAY Sta. 5385, 4 December 1963); UMML 17171 (1:103.0, Ga., 32°01'N, 79° 10'W, 137-146 m, M/V SILVER BAY Sta. 5397, 7 December 1963); *UW 13407 (1: 136.6, Gulf of Mexico, NE of Tortugas Light, 12 February 1954, shrimp trawl). Diagnosis. P. triacanthus is an elongate species with slightly falcate dorsal and anal fins. It is distinguished from P. paru by having a more elongate body -mean body depth of 458 compared to 710 for P. paru, each in thousandths of SL; a row of relatively large pores below the anterior half of the dorsal fin; and slightly falcate dorsal and anal fins rather than moder- ately or extremely falcate ones. Characters that distinguish P. triacanthus from P. burti and P. simillimus are listed in the diagnosis 200 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Table 17. Proportional Measurements of Pep- kills triacanthus. Symbols as in Table 7. Size range, 10.6-198.0 mm SL; mean size, 102.4 mm SL. Table 18. Meristic Values of Peprilus tria- canthus. Symbols as in Table 7. N R x SE V SE V In thousandths of SL: Head length 182 Snout length 182 Eye diameter 194 Length of upper jaw Interorbital width 182 182 Length of pectoral fin 177 Predorsal distance I 138 Predorsal distance II 182 Preanal distance 182 Maximum depth of body- 203 Least depth of caudal peduncle 182 In thousandths of HL: Snout length 182 Eye diameter 182 Length of upper jaw Interorbital width 182 182 251-358 289 1.87 8.72 055-090 071 0.57 10.86 061-133 086 1.12 18.15 064-113 085 0.69 10.93 072-111 092 0.62 9.09 179-364 312 2.60 11.06 332-492 381 3.23 9.98 228-407 278 2.82 13.67 375-544 445 2.23 6.76 364-600 458 3.28 10.22 055-089 070 0.54 10.36 175-309 252 1.96 10.49 232-377 292 2.41 11.14 239-348 295 1.50 6.87 245-389 320 2.26 9.51 of each of these species, respectively. The specific epithet, triacanthus, is from the Latin and means "three spines"; it refers to the first dorsal spine, the first anal spine, and the spine on the ventral surface of the pelvic bone, not to the usual presence of three dorsal and three anal spines. Description. Proportional measurements are given in Table 17 and meristic values in Table 18. Body elongate, shallow to moderately deep, compressed; anterior dor- sal profile slightly to moderately convex. Eye diameter greater than length of snout, eye moderately large. Dorsal and anal fins slightly falcate, longest dorsal ray three to Dorsal fin-rays* Anal fin-rays* Pectoral fin-rays Total gill rakers Total vertebrae Lateral-line scales 150 40-48 44.3 0.13 3.63 152 37-44 40.3 0.11 3.47 149 17-22 20.7 0.07 3.96 33 22-25 23.8 0.18 4.33 279 30-33 31.7 0.03 1.61 9 96-105 - * Excluding spines. five times the length of shortest dorsal ray, longest anal ray two to four times the length of shortest anal ray. Dorsal with two to four (usually three) small spines preceding the rays; anal with two or three (usually three) small spines preceding the rays. Base of anal fin shorter than base of dorsal (fewer rays). Caudal fin long to very long, deeply forked, about 25 to 35 per cent of total length. Subdermal canal system usually indistinct in preserved speci- mens, seen as vertical, parallel lines on body and dendritic canals on top of head and nape. The system is often very con- spicuous in living or freshly collected speci- mens. An irregular row of about 17 to 25 relatively large pores just below the an- terior half of the dorsal fin. Premaxillary teeth slightly recurved, usually with three small cusps. Swimbladder delicate, thin- walled, elongate, found only in specimens smaller than about 100 mm SL. Dorsal and upper ventral surfaces of the body quite often mottled with large, dark spots; other- wise, coloration as described for the genus. Maximum length probably 280 mm SL. Variation. Coefficients of variation (V) for proportional measurements range from 6.9 to 18.2 (Table 17) and from 1.6 to 4.3 for meristic characters (Table 18). High V values for proportional measurements indicate considerable variability. Partly it System atics and Biology of Peprilus - Horn 201 is due to allometry, especially in eye size, and partly to the inclusion in the sample of members of both the shallow and deep- bodied populations from the southern part of the range. Another source of variability, compared to some of the other species, is the inclusion in the sample of individuals of a broad size-range. The sample size of P. triacanthus used (Table 17) is considered to be reasonably adequate, and specimens from all parts of the known geographic range were ex- amined. Ranges of meristic values appear, by inspection, to approach a normal distri- bution ( Tables 1 to 4 ) . The main exception is vertebral number which, as in all the species of Peprilus, is relatively constant (Table 5). Variation in vertebral number I is slightly higher in this species than in the others and partly due to differences in the two populations in the southern part of the range. Geographic variation. Two distinct popu- lations apparently exist off the southeastern coast of the United States, one in deep water generally over a mud bottom, and lone in shallower water generally over a sand bottom (Table 22). The deep-water fish (Fig. 22) tend to be more elongate with numerous spots on the body and with 18 or 19, usually 19, caudal vertebrae. Those from shallow water tend to be {deeper-bodied with no spots and with 17 to 19, rarely 19, caudal vertebrae (similar to P. burti, Fig. 19). The triacanthus - fburti situation is discussed on p. 247. No other geographic variation was noted. Ontogenetic change. Few and moderate changes accompany growth in P. tria- canthus in a size range of about 10 to 198 mm SL. As shown by ratio-on-size dia- grams (Fig. 23), head length, eye diam- eter, and body depth decrease in size relative to SL with growth while the relative length of the pectoral fin increases up to about 100 mm SL, then becomes nearly constant; the relative depth of the caudal peduncle remains nearly constant. The correlation coefficient of the size-on- Eye Diometer • Body Depth • * " •. • * . • .". . * • • • • _ • , . . *. . . . • ' * , * • 1 ••. "• . _ • " * " • - • * •' • .. • - ■ I < i " Length of Pectoral Fn Figure 23. Ratio-on-size scatter diagrams for five morpho- metry characters of Peprilus triacanthus. size regression of eye diameter is relatively low (Fig. 3; Table 6). Distribution (Fig. 7). P. triacanthus is known from off the Atlantic Coast of North America from about 48 °N in the Gulf of St. Lawrence to about 27°30/N off southern Florida. The species is infrequently col- lected as far north as the south and east 202 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 coasts of Newfoundland (Bigelow and Schroeder, 1953), Cape Breton Island ( Hoar, 1937 ) , and the outer coast of Nova Scotia (McKenzie, 1939). It occurs in the Gulf of Maine in the summer and autumn. Southward, the species ranges to southern Florida in shallow water (for example, 27°28'N, 80°02'W, 35 m, M/V OREGON Sta. 5319) and to a similar region in deeper water (for example, 28°03'N, 80°29'W, 274-320 m, M/V PELICAN Sta. 25). I have identified one specimen of P. tria- canthus from the Gulf of Mexico, a male, 136.6 mm SL, from "NE of Tortugas Light" (UW 13407). Taxonomic comments. Until Haedrich's ( 1967 ) revision of the suborder Stroma- teoidei, the species had been known al- most exclusively as Poronotus triaeanthus. The chief basis for the generic distinction was the presence of a row of relatively large pores below the dorsal fin, a char- acter more reasonably considered as a specific one. For other relevant taxonomic comments, see under P. burti. Peprilus paru (Linnaeus, 1758) Figures 24, 29 Stromatcus paru Linnaeus, 1758: 248 (based on Sloane's (1725) description, Old Harbour, Ja- maiea, holotype not seen); Jordan and Gilbert, 1882a: 914; Jordan and Gilbert, 1882b: 596; Fordice, 1884: 312. Chaetodon alepidotus Linnaeus, 1766: 460 ( original description, Charleston, South Caro- lina, holotype not seen ) . Rhombus alepidotus, Lacepede, 1800: 321. Sternoptyx gardenii Bloch and Schneider, 1801: 494 ( original description, Carolina, holotype not seen ( after Linnaeus ) ) . Stromateus longipinnis Mitchill, 1815: 366 ( original description, New York Bay, holotype not seen). Seserinus xanthurus Quoy and Gaimard, 1824 ( original description, Rio de Janeiro, Brazil, holotype not seen); Fowler, 1906: 119. Peprilus crenulatus Cuvier, 1829: 214 (original description, locality ?, holotype not seen); Cuvier, 1836-1849: 141, fig. 3, pi. 64. Rhombus longipinnis, Cuvier and Valenciennes, 1833: 401, pi. 274; DeKay, 1842: 136, fig. 239, pi. 74. Rhombus argentipinnis Cuvier and Valenciennes, 1833: 405 (original description, Montevideo, Uruguay, holotype not seen ) . Rhombus xanthurus, Cuvier and Valenciennes, 1833: 405; Jordan and Evermann, 1898: 2849. Rhombus crenulatus, Cuvier and Valenciennes, 1833: 410, pi. 275. Peprilus longipinnis, Cuvier, 1836-1849, fig. 2, pi. 63 (in Atlas by Valenciennes); Gill, 1861: 35. Stromateus gardenii, Giinther, 1860: 399. Rhombus orbicularis Guichenot, 1866: 245 (original description, Cayenne, French Guiana, holotype not seen ) . Peprilus alepidotus, Goode, 1879: 112; Goode and Bean, 1879: 130; Bean, 1880: 92; Bigelow and Schroeder, 1953: 368; Briggs, 1958: 292; Springer, 1961: 482; Haedrich, 1967: 106, ^ fig. 41. Stromateus alepidotus, Jordan and Gilbert, 1882a: 451; Jordan, 1884: 149; Bean and Dresel, 1884: 156. Rhombus (Rhombus) paru, Jordan and Evermann, 1896: 965, Rhombus a genus and a subgenus. Rhombus (Rhombus) xanthurus, Jordan and Ever- mann, 1896: 966, Rhombus a genus and a sub- genus. Rhombus paru, Jordan and Evermann, 1898: 2849, fig. 404, pi. CL (in Part IV). Peprilus paru, Evermann and Marsh, 1900: 141, fig. 39; Meek and S. F. Hildebrand, 1925: 411; Jordan, Evermann, and Clark, 1930: 266; Briggs, 1958: 292; Haedrich, 1967: 106. Seserinus paru, Fowler, 1916: 402; Fowler, 1942: 152. Sesserinus xanthurus. Fowler, 1942: 152 (error in spelling of generic name; referable to Seseri- nus xanthurus Quoy and Gaimard, 1824). Simobrama xanthura, Fowler, 1944: 3, fig. 2. Material examined. Those srjecimens marked with an asterisk (*) have been radiograjohed. The number radiographed equals the number measured unless other- wise indicated. CS indicates specimens cleared and stained. Size ranges, in mm, are standard lengths (SL). FSBC 306 (1:63.6 mm, Gulf of Mexico, Florida Keys, 24°41'-52'N, 82°00'-35'W, 14-15 January 1958); FSBC 1745 (2:90.9, 95.5, Gulf of Mex., Fla., Pinellas Co., Johns Pass, Madeira Pass, 25 April I960); FSBC 2282 (4:75.1-83.8, Mississippi, Horn I., 27 Oc- tober 1962); FSBC 2572 (1:71.0, Gulf of Mex., Fla., 27°43'N, 82°45'W, 24 February 1963); FSBC 2562 (6:65.6-95.9, Gulf of Mex., Fla., 27°43'N, 82°45'W, 19 Decern- Systematics and Biology of Peprilus • Horn 203 ber 1962); FSBC 3831 (2:92.0, 138.0, Gulf of Mex., Fla., Tampa Bav, 1 mile S of MacDill AFB, 23 January 1966); FSBC 3956 (3:74.0-149.0, Gulf of Mex., Fla., [Pinellas Co., 27 December 1966); IMS 325 (1:63.0, near Pass Cavallo, Texas, 6-11 m, 22 October I960); IMS 326 (1:106.7, Tex., 18 miles W of Sabine, 12 m, 15 June 1951); IMS 333 (2:42.5, 54.0, Tex., Mustang I., 6-7 m, 21 October 1950); IMS 1019 (1: 68.3, near Port Aransas, Tex., 10 Julv 1962); IMS 1178 (1:137.0, 50 miles E 'of Port Aransas, Tex., 57-58 m, 6 December 1961); IMS uncat. (9:21.2-61.2, near Port Aransas, Tex., spring, 1964); LACM 4489 (2:61.5, 69.3, Mavagiiez, Puerto Rico, October, 1963, 1 CS); *LACM 4988 (1:175.0, Mavagiiez, P. R., March, 1964, beach seine); *LACM 4989 (1:172.0, Mavagiiez, P. R., March, 1964, beach seine); *LACM 5725 (1:103.7, Jamaica, Kingston market); *LACM 6741-21 (1:70.0, Mavagiiez, P. R., 3 December 1966); *LACM 7884 (1:103.3, Mavagiiez, P. R., 9 February 1963); MCZ 1687 (1:75.0, Beaufort, North Carolina); MCZ 4600 (1:118.0, Brazil); MCZ 16778 (1:106.0, Brazil); MCZ 16999 (1:73.0, Fla.); MCZ 17104 (2:123.0, 141.0, Rio de [aneiro, Brazil); MCZ 17129 (2:132.0, '137.0, Pensacola, Fla.); MCZ 17.336 (1: 123.0, Fla. Keys); MCZ 17364 (1:131.0, Penikese I., Elizabeth Is., Mass.); MCZ 26291 (1:106.0, Mobile, Alabama); MCZ 41064 (2:47.4, 56.3, Port au Prince, Haiti, August, 1950); *MCZ uncat. (5:112.0- 148.5, Port Aransas, Tex., 6-7 m, 3 July 1967, tide trap, 2 radiographed); MCZ Lmcat. (1.-60.0, N. C, Bogue Sound, August, 1966, CS); *SU 2774 (3:124.3-140.0, Wash- ington market); *SU 4900 (2:136.9, 168.6, Jamaica); *SU 14025 (1:222.0, Argentina, (Buenos Aires market, 1934); *SU 38704 (3:76.5-80.6, Aransas Pass, Tex., 12 Mav 1940); *SU 51811 (2:98.7, 105.5, Recife, (Brazil, 23 November 1944); *TABL 101976 (1:137.8, Honduras, off Caratasca Lagoon, 15°49.5'N, 83°44'W, 31 m, R/V UN- DAUNTED Cr. 6703, 7 April 1967); HTABL 101977 (7:115.7-130.1, Honduras, off Caratasca Lagoon, 15°54'N, 83°4( 37 m, R/V UNDAUNTED Cr. 6703, 8 April 1967); *TABL 101978 (2:85.4, 103.6, Hon- duras, off Caratasca Lagoon, 15°45'N, 83° 32'W, 33-37 m, R/V UNDAUNTED Cr. 6703, 9 April 1967); *TABL 101979 (1: 132.4, Honduras, off Caratasca Lagoon, 15°56'N, 83°41'W, 37-40 m, R/V UN- DAUNTED Cr. 6703, 12 April 1967); *TABL 101980 (2:102.9, 107.5, Honduras, off Caratasca Lagoon, ^"l^N, 83°26'W, 9 m, R/V UNDAUNTED Cr. 6703, 10 April 1967); *TABL 101982 (3:84.8-96.5, Honduras, off Caratasca Lagoon, 15°2rN, 83°34'W, 9 m, R/V UNDAUNTED Cr. 6703, 10 April 1967); *TABL 101983 (2: 114.8, 123.0, British Honduras, Belize, 17° 12'N, 88°11.2'W, 18-20 m, R/V UN- DAUNTED Cr. 6703, Sta. 72, 18 May 1967); *TABL uncat. (2:106.2, 134.9, Guy- ana, about 8°45'N, 59°15'W, 29 m, M/V CALAMAR Sta. 72, 18 June 1967); *TABL uncat. (7:82.2-97.7, Guyana, about 7°N, 58°30/W, near mouth of Essequibo River, M/V CALAMAR Sta. 82, 20 June 1967); *TABL uncat. (6:56.4-121.2, Surinam, about 6°N, 55°30'W, 20 m, M/V CALA- MAR Sta. 102, 15 July 1967); *TABL uncat. (1:88.1, Surinam, about 6°N, 54°30'W, 20 m, M/V CALAMAR Sta. 100, 14 July 1967); *TABL uncat. (1:104.2, Surinam, about 6°N, 54°30'W, 27-29 m, M/V CALAMAR Sta. 121, 11 August 1967); *TABL uncat. (2:113.9, 154.2 Surinam, about 6°N, 54° 30'W, 22-24 m, M/V CALAMAR Sta. 84, 21 June 1957); *TABL uncat. (3:78.5-90.4, Surinam, about 6°N, 54°30'W, 16-18 m, M/V CALAMAR Sta. 133, 15 August 1967); TABL uncat. (1:45.6, Georgia, Jekyll I. beach, 19 July 1956); TABL uncat. (1:62.6, Ga., commercial trawling area, 16 January 1957); TABL uncat. (1:100.2, Ga., commer- cial trawling area, 6 July 1959 ) ; TABL un- cat. (2:80.6, 115.4, Ga., commercial trawling area, 12-18 July 1959); TABL uncat. (4: 50.7-108.0, Ga., commercial trawling area, 26-30 July 1959); TABL uncat. (3:61.3- 85.7, Ga., commercial trawling area, 6 August 1959); TABL uncat. (1:177.0, 204 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Figure 24. Peprilus paru, 126.8 mm SL, Port Aransas, Texas, MCZ uncatalogued. Trinidad, Port of Spain market, M/V OREGON, 26 September 1964); TABL uncat. (2:146.0, 172.0, Venezuela, 9°32'N, 60°24'W, 60 m, M/V OREGON Sta. 2348, 20 September 1958); TABL uncat. (2: 140.0, 146.0, off Campeche, Mexico, 20° 01'N, 91°47'W, 55 m, M/V SILVER BAY Sta. 844, 17 November 1958); TABL un- cat. (1:68.4, N. C., 33°57.5'N, 77°50.5'W, 11 m, M/V SILVER BAY Sta. 1666, 27 February 1960); TABL uncat. (5:67.4-76.3, South Carolina, 32°55.5'N, 79°29/W, 9 m, M/V SILVER BAY Sta. 5423, 11 January 1964); TABL uncat. (1:135.8, Atlantic Ocean, Fla., 28°04'N, 80°11.5'W, 27-29 m, M/V SILVER BAY Sta. 5513, 29 February 1964); *TU 4396 (10:57.6-101.5, Louisiana, Terrebonne Co., mouth of Oyster Bayou, 31 October 1952); TU 9224 (10:25.5-33.1, La., Lake Pontchartrain, 2 miles SE of south draw, 30°10'N, 89°55'W, 11 August 1954); *USNM 3083 (3:50.0-63.0, Plum Point, Maryland, 22 August 1949); *USNM 30064 (1:153.0, Jamaica); *USNM 63660 (1:147.0, P. R., San Juan market); *USNM 63655 (1:114.0, Tex., Matagorda Bay); *USNM 83362 (2:93.0, 94.0, Rio de Janeiro, Brazil ) ; *USNM 86716 ( 1 : 131.0, Uruguay ) ; *USNM 117650 (1:152.0, Paraguay ?) *USNM 118637 (6:118.0-140.0, Galveston, Tex., 2 June 1940, 4 radiographed); *USNM 123044 (3:66.0-68.0, Golfo de Venezuela, Piedras Bay, 14 March 1925, USS NIAGARA); *USNM 156134 (8:55.0- 93.0, Aransas Pass, Tex., GRAMPUS Sta. 10476); *USNM 158505 (10:115.0-155.0, Systematics and Biology of Peprilvs • Horn 205 Tex., E of Padre I., 27°32.5'N, 96°28'W, pressed; anterior dorsal profile mc 60 m, M/V OREGON Sta. 4, 25 May to strongly convex. Eye diameter 1950); UMML 187 (1:155.0, Belize, British than length of snout; eye relatively Honduras, 17°10'N, 88°18'W, 27 September Dorsal and anal fins moderately to ex 1956); UMML 2085 (1:87.7, Dry Tortu- tremely falcate, longest dorsal and anal gas, 24°45'-50/N, 82°10'-30'W, 1955-1956); rays six or more times the length of shortest UMML 8271 (1:57.3, Atlantic Ocean, St. dorsal and anal rays; anterior anal rays Augustine, Fla., 21 December 1960); often extremely long, as much as 20 to 25 UMML 12422 (1:93.0, French Guiana, times the length of shortest ray. Dorsal 5°48'N, 52°53/W, 44 m, M/V OREGON fin with two to four (usually three) small Sta. 4190, 22 February 1963); UMML spines preceding the rays; anal with two or 13462 (1:60.8, Atlantic Ocean, St. Augus- three (usually three) small spines preced- tine, Fla., 1 February 1961); UMML 13986 ing the rays. Base of anal fin slightly (1:86.5, Surinam, 6°16'N, 55°56'W, 27 m, shorter than base of dorsal (fewer rays). M/V OREGON Sta. 4171, 19 February Caudal fin long to very long, deeply forked, 1963); UMML 16374 (1:49.2, Gulf of Mex., about 25 to 35 per cent of total length. Fla., Okaloosa Co., 9-10 September 1963); Subdermal canal system usually indistinct UMML 16817 (2:18.2, 25.7, Fla., Monroe in preserved specimens, seen as vertical, Co., Everglades Natl. Park, Shark River parallel lines on body and dendritic canals Delta, 5 November 1964); UMML 17068 on top of head and nape; pores very small (1:142.5, Colombia, 9°35'N, 76°04.5'W, on body and usually not visible except 40-55 m, M/V OREGON Sta. 4895, 26 in small individuals. Premaxillary teeth Mav 1964); UMML 21799 (2:99.5, 102.4, slightly recurved, pointed, simple. Swim- Gol'fo de Venezuela, 11°46'N, 71°16'W, bladder delicate, thin-walled, elongate; 24 m, M/V OREGON Sta. 5670, 6 October found only in specimens smaller than about 1965); UMML 21812 (2:30.3, 50.1, Golfo 100 mm SL. Coloration as described for the de Venezuela, 11°27'N, 71°39/W, 27 m, genus. Maximum length probably 280 mm M/V OREGON Sta. 5673, 6 October 1965 ) ; SL. UMML 22232 (1:110.0, Colombia, 8°48'- Variation. Coefficients of variation (V) 46.8'N, 76°39.7'-42.8'W, R/V PILLSBURY for proportional measurements range from Sta. 360, 12 July 1966); WHOI 66264B 5.8 to 14.5 (Table 19) and from 0.6 to 5.1 (1:83.0, Carmen, Mex., about 18°N, 92°W, for meristic characters (Table 20). Rel- 15 March 1950). atively high V values indicate considerable Diagnosis. P. pant is a very deep-bodied variability and are partly due to: allometry, species with moderately to extremely fal- especially in eye size; the inclusion of a cate dorsal and anal fins. The longest rays wide size-range of individuals in the of each fin are six or more times the length sample; and, to irregular variation over the of the shortest rays of each fin. Char- broad geographic range of the species, acters that distinguish P. paru from P. Meristic characters other than vertebral burti, P. triacanthus, and the related P. number and gill raker number vary more medius are listed in the diagnosis of each than in the other species of the genus of these species, respectively. The specific (Tables 1 to 5). Dorsal, anal, and pectoral epithet, paru, is a Brazilian name used by fin-ray numbers vary geographically and Linnaeus ( 1758 ) following Sloane's ( 1725 ) are discussed below. designation of the species as "Paru pisci Geographic variation. P. paru occurs Brasiliensi Congener." along the Atlantic Coast of the United Description. P oportional measurements States from Chesapeake Bay or just north- are given in Table 19 and meristic values ward to Florida, in the Gulf of Mexico, in in Table 20. Body ovate, very deep, com- the West Indies, along the coast of Central 206 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Table 19. Proportional Measurements of Pep- rilus PARU. Symbols as in Table 7. Size range, 18.2-222.0 mm SL; mean size, 94.3 mm SL. Table 20. Meristic Values of Peprilus parv. Symbols as in Table 7. N R SE V In thousandths of SL: Head length 182 Snout length 195 Eye diameter 202 Length of upper jaw Interorbital width Length of pectoral fin 196 Predorsal distance I 182 Predorsal distance II Preanal distance 182 Maximum depth of body 205 Least depth of caudal peduncle In thousandths of HL: Snout length 182 Eye diameter 182 Length of upper jaw 182 Interorbital width 182 251-396 305 1.93 8.54 044-084 062 0.56 12.63 058-147 103 1.05 14.46 182 065-112 084 0.68 10.90 182 085-144 112 0.84 10.14 274-454 386 2.51 9.11 369-503 435 2.12 6.57 179 258-399 319 2.02 8.48 369-636 484 3.27 9.13 565-877 710 4.25 8.57 191 066-112 092 0.58 8.66 148-247 204 1.41 9.35 228-442 334 2.35 9.51 233-318 275 1.19 5.82 245-479 368 1.74 6.37 America, and along the coast of South America, rather abundantly to Rio de Janeiro, and less commonly to Argentina. After the publication of "The Marine Fishes of Panama" by Meek and S. F. Hildebrand (1925), Hildebrand (MS) stated that there were two distinct species among the populations of P. paru: 1) P. alepidotus from the Atlantic Coast of the United States and the Gulf of Mexico; and 2) P. paru from the West Indies, Central America, and South America. He noted that P. paru differed from P. alepidotus in having a more robust body, larger scales, and slightly shorter (fewer rays) dorsal and anal fins, particularly the anal. Hilde- N R SE V Dorsal fin-rays* Anal fin-rays* Pectoral fin-rays Total gill rakers Total vertebrae Lateral-line scale 178 38-47 42.9 0.12 3.75 176 35-45 40.4 0.14 4.73 172 18-24 21.9 0.09 5.11 98 20-23 21.5 0.07 3.12 182 29-31 30.0 0.01 0.57 22 80-95 * Excluding spines. brand also stated that while each has an equal number of pectoral rays, P. paru has a narrower pectoral and a less falcate dor- sal fin than P. alepidotus. His findings were based on numerous specimens from the United States, two from the West Indies, two from Panama, five from Vene- zuela, four from Brazil, and five from Uruguay. Following Hildebrand's analysis, P. paru and P. alepidotus have been considered distinct in faunal and fishery works. Based upon the examination of 207 speci- mens from throughout the species range, I find only partial agreement with Hildebrand's conclusions. Although body proportions do vary, I could not detect geo- graphic variation of characters reflecting the "robustness" of the body, nor could I find decided differences in the size of the scales or in the shape of the pectoral fin. Some slight differences may exist in the above three characters but do not seem of sufficient magnitude to warrant distinction. Contrary to Hildebrand's statement that the dorsal fin of P. alepidotus is more falcate than that of P. paru, I find that the length of the anterior lobes of both dorsal and anal fins tends to increase in a north-to-south direction. This follows a general trend existent in a number of fish groups -that the fins are longer in more Systematics and Biology of Pephua s • Horn 207 i:::.:::::::i I V///A -wk. P<< .001 - I K0&fr'"r-| •.:•.: kaa&gfl —1 DORSAl PATS Figure 25. Geographic variation in the number of dorsal fin-rays of Peprilus paru-. A z= United States, Atlantic Ocean; B zz Gulf of Mexico; C zz Central America; D — West Indies; E zz South America; numbers in parentheses are sample sizes. The horizontal line represents the range of values; the vertical line represents the mean; the large, hol- low rectangle represents one standard deviation on either side of the mean; the small, barred rectangle represents the 95 per cent confidence interval. Each P value is for comparison of the two sample means immediately above and below the value; only P values of 2 per cent or less are listed. tropical waters (Frederick H. Berry, per- sonal communication). In P. paru it ap- pears to be partly an individual variation since certain fish from a single sample may have extremely falcate dorsal and anal fins. The counts of the dorsal, anal, and pec- toral rays vary geographically, but in each case the variation is clinal rather than being a basis for dividing the populations into two distinct species. The clines are in an irregular north-to-south pattern and are diagrammed in Figures 25 to 27 in a manner generally following the method of Hubbs and Hubbs (1953). The number of dorsal rays shows a regular gradient in the manner in which I have arranged the regions ( Fig. 25 ) al- though, in a strictly latitudinal sense, the positions of the West Indies and Central America should probably be reversed. The ranges of values overlap widely, but in proceeding from one coastal region to an- other in a north-to-south direction, the sample from the United States Atlantic shows a significant difference from the z:::. ,:::::-. "::::::z ANAL RA.i Figure 26. Geographic variation in the number of anal fin-rays of Peprilus paru. Explanation of symbols in Figure 25. Gulf of Mexico sample, and the Central American population has a significantly dif- ferent number from the West Indian and the South American populations. The levels of significance were set at the 2 per cent level using a two-tailed Student's f-test. The greatest difference among popu- lations of the five regions occurs in the number of anal fin-rays, although the ranges of values overlap considerably (Fig. 26). In this character the United States Atlantic and Gulf of Mexico populations are very similar. The latter sample differs signifi- cantly from the Central American sample and in turn the Central American one is A (28) d V/////A B I72L £ A WZK C (12) D (13) E (48) .01 > P >.001 Y//////A I Y/////////A d VZ7A I PECTORAL RAYS Figure 27. Geographic variation in the number of pectoral fin-rays of Peprilus paru. Explanation of symbols in Figure 25. 208 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 significantly different at the 2 per cent level from the West Indian sample. The Central American sample is different from the South American one, but onlv significant at the 2 to 5 per cent level. The West Indian members are the most aberrant among the five regions; however, the sample size (13) is the smallest. The differences in this character among the regional populations do not seem strong enough to suggest a subspecific designation, especially since other characters vary in a different manner. In the pectoral ray number the gradient is slight, the overlap is considerable, and there is less difference among the popu- lations than in the above two meristic characters (Fig. 27). The Central Ameri- can population is significantly different from the Gulf of Mexico member, but the southern populations are completely con- tained in the ranges of values of the United States Atlantic and Gulf populations. Further evidence for divergence among the populations of P. paru has come from an examination of the parasitic isopods which inhabit the gill chambers of this species. Thomas E. Bowman (personal communication) has found that the cymo- thoid isopods taken from United States Atlantic and Gulf populations are different from those taken from P. pant collected off Surinam. The species in the former regions is Lironeca ovalis Say (possibly a synonym of L. redmanni Leach), while those from the Surinam specimens are very similar in some ways but in other characters are different enough to possibly be considered as members of a distinct species. Speci- mens from the intermediate areas, West Indies and Central America, are needed to clarify the status of the populations. It may be that the isopods are geographically varying in a manner parallel to their hosts. If there are two species of isopods involved, this would lend support to the recognition of two separate species among the popu- lations of P. paru. It would not be con- clusive evidence since it is to be expected that rates of differentiation and speciation would be different in the two animal groups. In summary, there is apparently semi- isolation and some differentiation among regional populations of P. paru. Although there is probably no complete break in gene flow, it does appear to be reduced, the reduction effected by barriers of deep water and unfavorable coastlines. P. paru is essentially a shallow-water species and would cross deep water only infrequently, probably most commonly by passive trans- port of eggs and larvae. The species is most abundant in areas where the con- tinental shelf is broad and with large ex- panses of shallow water. Zones where the shelf is narrow or where the shore is rocky may serve as partial barriers to P. paru. Although the distribution of P. paru is continuous around Florida, it is uncommon in southern Florida; and there is evidence of slight differentiation between Atlantic and Gulf populations (Fig. 25). Atlantic and Gulf populations differ more signifi- cantly in the three meristic characters discussed above from West Indian forms than from the Central American ones. This disparity may be partly due to the gap of deep water separating the United States from the West Indies as compared to con- tinuous coastline from the Gulf to Central America. Some differentiation is apparent between Gulf and Central American popu- lations and mav be effected somewhat bv the lack of favorable coastlines between the two regions. The West Indian form ap- pears to be most similar to South American members of the species. Continuity and genetic interchange are probably being maintained between these two regions via the shallow coastal areas associated with the chain of West Indian islands. Deep water and unfavorable coastlines are ap- parently serving to partially isolate Central American populations from the West Indian and South American populations, respectively. Mayr (1963: 361) states that clines are the product of two conflicting forces: Systematics and Biology of Peprilvs • Horn 209 Head Length J I 1_ * 85 '.: -J L_ 170 195 Figure 28. Ratio-on-size scatter diagrams for five morpho- metric characters of Pepn'/us paru. selection, which would make every popu- lation uniquely adapted to its local environ- ment; and, gene flow, which would tend to make all populations of a species identical. Mayr (1963: 365) further states that in many cases gene flow seems responsible for the maintenance of clines to a greater degree than environmental gradients. The cohesive effect of gene flow seems to be exceedingly potent. Although a tempera- ture gradient is probably partially respon- sible for the clines in characters of P. paru, it seems that gene flow between contiguous populations is of equal or greater impor- tance. If temperature were of prime im- portance, an increase in meristic values in the extreme southern portion of the species range might be expected. I have not found such a pattern. Ontogenetic change. Few and moderate changes accompany growth in P. paru in a size range of about 25 to 177 mm SL. As shown by ratio-on-size diagrams (Fig. 28), head length, eye diameter, and body depth decrease in size relative to SL with growth, while the relative length of the pectoral fin increases up to about 100 mm SL then becomes nearly constant; the relative depth of the caudal peduncle remains nearly con- stant. The correlation coefficient in a size- on-size regression for eye diameter is relatively low (Fig. 3; Table 6). Distribution (Fig. 7). P. paru is known from Chesapeake Bay or just north on the Atlantic Coast of the United States south- ward around peninsular Florida to through- out the Gulf of Mexico, Central America, and at least part of the West Indies to South America where it commonly occurs to about 23° S in the vicinity of Rio de Janeiro, Brazil, and less abundantly as far south as Argentina. The species infre- quently ranges as far north as Long Island and rarely as far as the Gulf of Maine, the most northerly capture being at Cape Elizabeth, Maine, about 43°N (Bigelow and Schroeder, 1953). P. paru has been collected near a number of West Indian Islands including Jamaica, Puerto Rico, Haiti, and Trinidad. Taxonomic comments. Since the descrip- tions by Linnaeus of Stromateus paru in 1758 and Chaetodon alepidotus in 1766, 210 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 the taxonomic status of the species has Skeletal features. Ossification of various been in question. Various names have been parts of the skeleton of Teprilus appears applied to this widely-distributed and vari- to occur differentially. Study of cleared- able species; hence, the synonymic list is and-stained material of two species, P. an extensive one. trkicanthus and P. burti, shows that the It has become customary in faunal and vertebral column is ossified very early, fishery works to list P. alepidotus as the Specimens of 6 or 7 mm SL have the an- species along the Atlantic and Gulf coasts terior two-thirds of the column almost of the United States and P. paru as the completely ossified, while the posterior species along the West Indies, Central and one-third and the hypural plate are in the South America. P. paru is the name some- early stages of ossification. In these same times given to the Gulf of Mexico popu- specimens the bones of the head are in- lations, and P. alepidotus to the population completely ossified; the pectoral girdle is off the Atlantic Coast of the United States, apparently relatively well ossified and the Briggs (1958) lists both as occurring off pectoral fin is short and fanlike; the jaw the Atlantic and Gulf coasts of the United teeth and the teeth of the pharyngeal sac States; this seems to be the least likely situ- have formed; and the dorsal and anal fins ation. In this study, all populations are are almost completely absent at this stage, included within P. paru. In specimens of 8 to 10 mm SL the epural Lutken (1880) listed a fish he called elements remain unossified while the Stromateus paru as occurring at the south- hypurals are more completely ossified; the ern end of the Strait of Malacca between median fins, particularly anteriorly, are Malaya and Sumatra in southeast Asia, beginning to ossify. In specimens of 12 His drawing was of a small fish which had to 18 mm SL the rays of the median fins pelvic fins. P. paru is restricted to the New are visible and all parts except the epurals World and has no pelvic fins. Therefore, are near complete ossification (stromateids this species is excluded from being the do not have strongly ossified skeletons), fish of Liitken's description. Liitken's draw- The caudal skeleton has by this size range ing may be referable to Parastromateus achieved a higher degree of symmetry ( = Apolectus) niger, a fish representing with the hypural plate occupying a more a monotypic family of uncertain affinity. equal position dorsally and ventrally, and P. paru is commonly known as the with the urostyle becoming more slender; "harvestfish." Names less frequently used the epurals have reached almost complete are "starfish" and "poppyfish." ossification and the caudal fin has begun to fork. In the smallest specimens examined ASPECTS OF FUNCTIONAL (5 mm SL) tne hypural plate had the MORPHOLOGY same degree of fusion as in the adult skeleton - two epural and four hypural Information on the morphology and elements> At a size range of 25 to 30 mm ecology of the members of the genus SL the epurak are Qssified ,md the pectoral Peprilus is sufficient to allow some inter- fin has become more elongate. No differ] pretations of functional moqohology, and ences in pattern 0f ossification between the differences among the species are great two Species were found, enough to allow consideration of some The pattern of ossification may correlate trends and specializations which appear to with the behavior of the young fish, be developing. It is instructive to see Spawning occurs offshore generally, and whether certain trends correspond to the young are in the pelagic surface layers those thought to be occurring generally in of the ocean. After hatching, and until perciform fishes. several days after the yolk sac is absorbed, Systematics and Biology of Peprilus • Horn 211 the larvae are essentially planktonic with probably little locomotor effort of their own. Early vertebral ossification results in a supporting element of some rigidity. Also, the pectoral fins appear early at : about four or five days in P. triacanthus (Colton and Honey, 1963). As the caudal apparatus and the median fins strengthen I through ossification, and as the caudal fin I becomes forked, the young fish becomes I more independent in locomotion and at a I size of 10 to 15 mm SL probably becomes \ more nektonic than planktonic. The juve- |i nile fish frequently become associated with \ coelenterate medusae or siphonophores at a size of 10 to 30 mm SL. Fishes of the genus Peprilus and the family Stromateidae in general have deep, compressed bodies, which are deepest in the region of and just posterior to the body cavity. All seven species of Peprilus are similar osteologieally, and one of the most conspicuous skeletal regions is that sur- rounding the body cavity and supporting the viscera. This "visceral basket" gives obvious support and protection to the viscera and abdomen. It is formed ante- riorly and ventrally by the eleithrum and the large, posteroventrally directed pelvic bones which unite posteriorly to form a single structure. The vertebrae and pleural ribs house the body cavity dorsally. Pos- teriorly, the enlarged first interhaemal swings posteriorly to meet the haemal spine of the first caudal vertebrae. The pair of elongate posteleithra provide lateral support. An increased number of anal fin- . rays in the stromateids has apparently }! resulted in the angular position of the first interhaemal. While most meristic characters of spe- ; cies of Peprilus vary considerably, the I vertebral number varies only slightly Table 5). No geographic variation in vertebral number was detected in any of pdie species, with the possible exception of the complex burti-triacanihus situation off the Atlantic coast of the southeastern United States. Variation is so slight that . the vertebral count becomes an important specific character in this genus. For the seven species, 1255 vertebral counts were made. Coefficients of variation (Tables 8, 10, 12, 14, 16, 18, and 20) are very low for this, character, indicating great homo- geneity within a species. Several authors (Hubbs, 1926; Vladykov, 1934; Taning, 1952; Bailey and Gosline, 1955; and others) have shown that meristic elements, including vertebrae, are influ- enced by temperature and that the number of such elements is progressively greater to the north, or at higher altitudes, or in locally colder areas of spawning. Variation in vertebral number in certain fish groups is often so great that the character must be used only with caution, as has been shown by Bailey and Gosline ( 1955 ) for darters of the family Percidae. But in Peprilus variation is slight and vertebral number is an important distinguishing char- acter. The caudal vertebrae of Peprilus vary more than the precaudal ones, a situation found to be a somewhat general one by C. L. Hubbs (1922). It may be that precaudal vertebrae are fixed earlier in development than most other meristic characters, as is true in the paradise fish, Macropodus opercularis (Lindsey, 1954). Constancy of vertebral number may be correlated with certain other characters involved in evolutionary trends. Bailey and Gosline ( 1955 ) have stated that in the family Percidae decreased size is ap- parently, and probably casually, correlated with reduction in vertebral number. In- crease in attenuation within the Percidae appears to be primarily or entirely associ- ated with the elongation of the vertebral centra, rather than with an increase in vertebral number; deepening of the body is marked by some foreshortening of the vertebrae. Barlow (1961) reacted skepti- cally to the idea that the addition or subtraction of a few elements would be of selective value. However, as in the Per- cidae, vertebral number in the species of Peprilus seems to be significantly associated 212 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 with size and body shape (Table 21). Size of individual vertebrae seems to be about the same among the species. P. snyderi, the most elongate and prob- ably the largest species in the genus, has the greatest number of vertebrae, 36. P. paru, the deepest-bodied of the species, and also considered to be the most highly derived member of the genus, has the fewest vertebrae, 29 to 31. The close relative of P. paru in the Pacific, P. medius, is a more elongate species and has 33 to 35 vertebrae. P. ovatus, the deep-bodied, small derivative of P. medius, has 31 to 33 vertebrae. The relatively elongate P. tria- canthus has 31 to 33 vertebrae while its close but slightly deeper, smaller relative, P. burti, has 29 to 31. The only exception to the trend occurs in P. simillimus, a relatively elongate species, which has 30 or 31 vertebrae. Vertebral number in as- sociation with other character trends ap- pears to have selective value in the genus, and this selection apparently reduces the variability of the character. In stromateoid evolution the general tendency has been toward an increase in the number of vertebrae ( Haedrich, 1967 ) . This trend seems to have culminated in the rather high number among the Stroma- teidae. However, from the basal stromateid stock the trend in Peprilus seems to be toward a reduction in number of vertebrae in correlation with the trend toward a deep, less elongate body. Body form: shape, position, and loss of fins. Body form is associated with the shape and position of the fins and both in turn are associated with the locomotion and mode of life of the fish. Myers ( 1958), Liem (1963), and Patterson (1964) have stated and presented evidence that within the teleostean fishes and particularly the acanthopterygians there has been a trend toward a shorter, deeper, more highly compressed body. Associated with this trend have been several changes in skeletal features and in the shape and location of the fins. The Stromateidae represent the zenith of stromateoid evolution (Haedrich, 1967) and are among the smallest fishes of the suborder. Within the Stromateidae, par- ticularly the genus Peprilus, a number of specific conditions and trends exemplify some of the general tendencies occurring in acanthopterygian fishes. All species of Peprilus are relatively deep, varying from elongate to very deep, and are highly compressed. A trend in spiny-rayed fishes has been for the pec- toral fins to move to a lateral position on the body and for the pelvic fins to migrate forward from an abdominal position. The pectoral fin in a lateral position has as- sumed the primary functions of braking and turning (Harris, 1937; Patterson, 1964). The braking movements of the pectoral produces a lift as well as a drag force. Harris ( 1937 ) has shown that the neutrali- zation of this lift force is effected by an equal downward force produced by the anteriorly-placed pelvic fins. If migration of the pelvics had not occurred in evolution along with the migration upward of the pectorals, the fish would either tilt up- wards or rise bodily when it comes to a stop. Extension of a pectoral facilitates turn- ing; and turning is most rapid if the mass is concentrated at the level of the pectorals, the fulcra in turning (Patterson, 1964). Such a concentration of mass can be achieved by deepening the body, and for retention of streamlined form, greatest depth is required about at the level of the pectoral. The dorsal and anal fins function to stabilize against rolling motion, i. e., they serve primarily as keels (Breder, 1926; Harris, 1937). A fish inclined to roll will do so about the long axis of the body; thus, for the dorsal and anal fins to be most effective they need to be as far as possible from the long axis. If the body is deepened, the dorsal and anal fins are farther from the long axis, and rolling stability is in- creased. Systematics and Biology of Peprilus • Horn 213 That the pelvic fins act as bilge keels vices and often following the form of the has been advocated by Breder (1926). pectorals. The pelvics are the fins most Harris (1937) differs and states that the commonly modified so as to lose all loco- pelvics may be used rather to produce motor function. The following discussion rolling movements for establishing equilib- considers the loss of the pelvics only in the rium or for swimming between rocks or stromateid fishes, but may also apply to into crevices in other than an upright certain other perciform groups, position. The position and function of the pec- The swimming behavior of adults of torals, the shape of the body, and the mode three species, P. triacanthus, P. burti, and of swimming all should be relevant to the P. para, was observed in this study. In loss of the pelvics. In Peprilus the pectorals each instance the fishes were accustomed appear to perform at least three functions: to the tank and locomotion was apparently braking, turning, and propelling. Propul- normal. All species were observed in sion is of the carangiform type, supple- public aquaria or in tanks in which I had mented by the synchronous flapping of the placed freshly collected fish. Swimming pectoral fins. The latter is more important appears to be a continuous process with no at lower speeds. This use of the pectorals hovering in midwater. Locomotion is par- is common in short, deep-bodied forms, tially produced by the usual serial con- according to Breder (1926). tractions of myomeres, but propulsion by Breder stated that with the development the flapping of the pectoral fins is also of a short, deep, compressed body, and a well developed, especially at less than centrally-located pectoral, may come a maximum speeds. Fishes of this genus reduction or loss of pelvic fins - as in the appear to be moderately rapid swimmers genus Vomer (Carangidae) or in Peprilus with the ability for quick turning and for triacanthus. Vomer is a deep, extremely continuous swimming for long periods of compressed pelagic fish with very small time. pelvics which seem, certainly, to be of Pelvic fins are present in only one species little functional importance. P. triacanthus of the Stromateidae, in specimens smaller is also a deep, highly compressed pelagic than 100 mm SL of Stromateus fiatola fish, the adults of which swim continuously ( Haedrich, 1967 ) . Other stromateoids have without hovering. If frequent braking does pelvic fins that are generally situated under not occur, as seems to be the case in Pepri- the pectorals; the pelvics are very large in lus, less of a downward force is needed or Nomeus. The loss of pelvic fins might less frequently needed to counteract the seem to be a major evolutionary change, lift force created by the pectorals in brak- A rather specific geometric and hydro- ing. Also, the deep portion of the body dynamic condition of the body in relation beneath the pectorals should provide some to swimming and mode of life of the fish downward force which would further re- should exist or develop if a species or group duce the need for pelvic fins. Breder of species can dispense with a set of paired reasoned that the pelvics may be large, as appendages. In S. fiatola the pelvics are in the somewhat primitive perciform genus, apparently in the process of being lost, Lepomis ( Centrarchidae ) , which fre- and in the other stromateids these fins have quently stops and hovers, already been lost. In several groups of Fishes of the genus Peprilus appear to perciform fishes pelvic fins are quite small possess several attributes for moderately and in some they are absent. The actual rapid swimming at least for short periods, form of the pelvic fins appears to be of The caudal fin is deeply forked. Gero little significance in rapid swimming, these (1952) has demonstrated that the tips of fins being no more than maneuvering de- the caudal fin produce less turbulence than 214 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 the center portion. Thus, the more deeply forked the fin, the less should its oscillations disturb the water. As the aspect ratio ( = (span of the fin)2/area 0f the fin) is increased the drag diminishes. Aspect ratios of Pcprilus range from about 3.5 in P. simillimus to about 5.0 or more in P. medius, which puts the group into the upper range or above of Nursall's ( 1958 ) category of "typical fish." This category includes fish which have moderate ampli- tude and fairly high frequency waves to form swimming oscillations, a flexible vertebral column of 24 to 100 vertebrae, a definitely narrowed but muscular caudal peduncle, and a flexible, forked tail of intermediate aspect ratio (two to four). Breder ( 1926 ) stated that pectorals of spatulate or fanlike form accompany fishes of slow or moderate speeds, while long and falcate ones accompany rapid swimmers. The rapid swimmers were said to use such fins largely for quick-turning and seldom, if ever, for either propulsion or for main- taining a stationary position. The above does not agree with the condition in Pepri- lus in which the pectorals of large juve- niles and adults are long and winglike, but are used in propulsion. One function of the dorsal and anal fins is to produce stability against rolling (Breder, 1926; Harris, 1937). The cross section of many fish, including Peprilus, is elliptical at the level of the pectoral fins (Fig. 29). The corresponding position of maximum tangential velocity during rolling movements are not at the sides, but at the top and the bottom of the body, the positions of the dorsal and anal fins. The dorsal and anal fins of rapidly swimming fish are often somewhat similar in size and shape, especially in the latter. Frequently in less active forms, one fin may be larger than the other. A trend or at least a morphological arrangement mentioned by Breder ( 1926 ) which seems to be oper- ative in Peprilus is that in reasonably elongate forms that have the pelvic fins greatly reduced or wanting, the anal fin is Figure 29. Frontal view of Peprilus paru, 117.0 mm SL specimen. usually quite long with the vent displaced forward and the first interhaemal directed backward. Another observation of Breder which seems to have some relevance in the genus Peprilus is that of the relation of the development of the dorsal and anal fins to the area of concentration of the bulk of the body. Breder (1926) observed that in fishes with their greater bulk above a line from the tip of the snout to the middle of the peduncle, the dorsal fin is generally larger than the anal. And conversely, when the bulk is below that line, the anal fin is generally the larger. In fishes nearly sym- metrical about such a line, the dorsal and anal fins are nearly equal in size. Five species of Peprilus, P. snyderi, P. similli- Systematics and Biology of Peprilus • Horn 215 mus, P. triacanthus, P. burti, and P. medius, have the greater bulk above the line and all have dorsal fins which are longer (more rays) than the anal fins (Table 21). P. paru has slightly greater bulk above the line and has only a slightly longer dorsal than anal fin ( Table 21 ) , while the anal is usually more falcate. P. ovatus is nearly symmetrical about this longitudinal line and, in turn, the dorsal and anal fins are very similar in size and shape (Table 21); the mean ray counts are about equal (D, 42.8; A, 43.0). The above pattern suggests that selection favors the size and position- ing of these steadying keels which provide I the optimum range of stability. The ex- tremely long anterior lobes of the dorsal and anal fins of P. paru and P. medius may serve to increase rolling stability in these deep-bodied fishes, and the anal lobes may somewhat replace the pelvic fin in produc- ing a downward force in counteraction to the upward force produced by the pectoral fin. Patterson ( 1964 ) has stated that during teleost evolution the number of precaudal vertebrae has been reduced more drastically than the number of caudal vertebrae. This feature is characteristic of stromateoid Fishes. The results of this change in pro- portion, according to Patterson, is that in icanthopterygian fishes the less muscular ibdominal region is reduced more than the highly muscular and propulsive caudal ; "egion, and the resistance to lateral oscil- ation of the anterior part of the fish is •educed. These trends can be compensated or by increasing the depth of the trunk, or the streamlined form of the fish de- mands that the greatest depth be anterior :o the point of contraction in oscillating rhe tail. The most elongate species, P. hnyderi, has 15 precaudal and 21 caudal i vertebrae; the shorter, deeper species of Weprilus have 13 precaudal and as few as 1 16 caudal vertebrae. Swimbladder and mode of life. Changes n the mode of life during development ire probably closely correlated with the size and shape of the pectoral fins and with the state of the swimbladder. The larvae and small juveniles of Pepri- lus are surface dwellers after hatching in offshore waters. They then generally mi- grate to coastal waters and bays. At a length of about 10 to 30 mm SL, the young frequently become associated with jellyfish medusae and may maintain such a relation- ship up to a size of 80 to 100 mm SL. This association is well documented (Pearson, 1941; Bigelow and Schroeder, 1953; Man- sueti, 1963; and others), and appears to be an important aspect of the life history. The young fish apparently seek pro- tection among the coelenterate tentacles and also feed on the tentacles themselves. Mansueti ( 1963 ) has shown in laboratory experiments that the association requires of the fishes an ability to hover, a loco- motor attribute not evident in adult Pepri- lus. The adults that I have observed (three species) swim continuously with little or no hovering in midwater. Two characteristic changes seem to cor- relate with the above apparent shift in habit and locomotor pattern. The first is the change of the pectoral fin from a short, fanlike structure to a long, winglike one. The change begins early and continues until the fish reaches a size of 80 to 100 mm SL (Figs. 9, 11, 15, 17, 20, 23, and 28). The second change is the apparent regres- sion of the swimbladder in individuals larger than 80 to 100 mm SL. Although the information is from only three species, P. triacanthus, P. burti, and P. paru, a swim- bladder is present in these fishes up to at least a size of about 100 mm SL. The organ is elongate, thin-walled, and slightly in- clined anteriorly. In adult specimens of all species no swimbladder has been found, although fresh or well-preserved material of every species has not been examined. It seems, then, that the swimbladder becomes regressed and nonfunctional in fishes larger than 100 mm SL. It may be that the swim- bladder is functional in juveniles and impor- tant in hovering and maintaining position 216 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Table 21. Associations and Trends of Certain Character Values Among the Species of Pep- rilvs. The deepest-bodied species is listed at left followed in order of decreasing body depth by the other species. Mean morphometric values are in thousandths of SL. para ovatus burti medius simillimus triacanthus smjderi Body depth 710 619 551 529 461 458 427 Preanal distance 484 481 465 434 440 445 424 Predorsal distance I 435 408 395 390 369 381 364 Caudal peduncle depth 092 079 075 078 065 070 071 Eye diameter 103 093 100 082 073 086 070 Pectoral fin length 386 360 332 386 336 312 329 Mean total vertebrae 30.0 32.0 30.0 34.0 30.2 31.7 36.0 Ratio of dorsal rays to anal rays 1.04 1.00 1.10 1.06 1.13 1.10 1.11 Relative length of anterior lobes of median fins* 1 3 4 2 6 4 5 * Longest lobes given rank of 1. under floating objects such as jellyfish me- dusae, and that the organ becomes non- functional when swimming becomes con- tinuous. Jordan and Evermann (1896) indicated that the swimbladder is "usually" absent in stromateid fishes. Breder (1926), in discussing the hydrostatic situation for fishes which lack a swimbladder, stated that some fishes such as P. triacanthus (indicating that this fish has no swim- bladder) are so close to the specific gravity of water due to a sufficient amount of fat, that only the slightest movement suffices to keep them from sinking or rising. Fishes of this genus are quite oily and perhaps this condition does facilitate the mainte- nance of position and the ease of ascent or descent. I have observed the movements of adult P. triacanthus in aquaria, and it appears that slow sinking follows cessation of active swimming, but that only the slightest pectoral movements return the fish to a horizontal or an upward direction. H. F. Taylor (1922) calculated that ab- sence of a swimbladder in other than bot- tom forms should occur only in marine fishes since the quantity of fat necessary to float a fish in the less dense fresh water is practically prohibitive. He estimated that a marine fish with a fat content of 29.34 per cent and with no swimbladder would be in equilibrium with sea water. Many other factors, however, including the relative amount and density of bone need to be considered in this situation. N. B. Marshall (personal communication) states that a number of shallow-water marine fishes have regressed, age-de- pendent swimbladders. He includes in the list certain gobies, blennies, flatfishes, certain Solenichthys spp., most muraenid eels, gadids such as Gaidropsarus spp., and some percomorphs such as Thalassoma bifasciatum, Holacanthus tricolor, Upeneus prayensis and an oceanic species, Nomens gronovii. The latter species is a stromateoid and lives closely associated with the siphon! ophore, Plujsalia. Whether Nomeus stays with Pln/salia all its life is uncertain. The familiar small, blotched Nomeus found under Plujsalia may only be the young form of a bigger fish which lives in deep water. In Nomeus, as in Pcprilus, regression of the swimbladder may correlate with a change in mode of life. The above discussion of swimbladder loss is speculative at this stage. A number of questions remain to be resolved. Fur- ther examination of specimens, observations of living fish, and a series of experiments are being planned in an attempt to answer some of the questions. A comparative study of stromateoids and other fishes with Systematics and Biology of Pbprili • Hum 217 Figure 30. Scales of Peprilus snyderi: region. From 227.0 mm SL specimen. = scale from trunk below lateral line; b = lateral line scale from mid-trunk regressed, age-dependent swimbladders is being initiated. Character associations. Listed in Table 21 are the mean values of some morpho- metric and meristic characters which illus- trate trends in sets of characters among the species of Peprilus. P. para, the short- est, deepest-bodied species, is listed first in the table, followed in order bv the more elongate species. From this ranking the table shows how other characters relate to that of body depth. Greatest body depth is associated with the greatest preanal and predorsal distances, the longest pectoral, the largest eye, and the fewest vertebrae. In the most elongate species, P. snyderi, these character values are largely reversed. P. snyderi probably does not have the quick-turning ability attained by P. paru or P. ovatus, for example. I have not analyzed unit characters or their roles by the techniques of Olson (1964), although this approach might be instructive, especially if applied to the entire suborder. Skin and scales. The skin of Peprilus is thin and, as in most other stromateoid fishes, the scales (Fig. 30) are thin, cycloid, irregular in shape, highly deciduous, and in shallow pockets in the skin. The lateral line scales (Fig. 30b) have a simple tube passing through the middle, and are less deciduous than scales of other parts of the body. The scale condition in Peprilus may be important in achieving greater maneu- verability and may represent a stage in a trend toward complete loss of scales. The cephalic lateral line consists of pores and branching canals on the cheek, opercular area, snout, lower jaw, and top of head. In the stromateoid fishes there is usually a well-developed subdermal canal system which communicates to the surface through small pores scattered over the head and body (Haedrich, 1967). The system is particularly well developed in Peprilus (Fig. 31), and is conspicuous in living or freshly collected fish. The canals often are only slightly visible after preservation. There are some differences in the system among the species of the genus. P. burti and P. triacanthus (Fig. 19 and 21-22) have a row of large pores beneath the anterior half of the dorsal fin. A series of 218 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 lormrm Figure 31. Subdermal canal system and portions of the cephalic lateral line of Peprilus burti, 124.6 mm SL specimen. small, irregularly-spaced, but conspicuous pores are sometimes evident along the dorsal surface of P. ovotus (Fig. 10). The function of this extensive canal system is unknown. Whether the system is independent of the lateral line and whether sensory structures are present are unknown. If the canals contain a sensory mechanism, they may function as receptors important in the formation and maintenance of schools. These fishes produce large amounts of mucous, and the system might be used in the dispersal of the mucous which may contain an agent that counter- acts coelenterate toxins. Walters (1963), in a study of the integument and subdermal canal system of the trachipterid fishes, has proposed that the canal system functions in boundary-layer control through dis- tributed dynamic damping. Such control could apparently decrease drag by 60 to 75 per cent. According to the hypothesis, as the skin pressure mounts beneath a growing boundary-layer perturbation, the layer will sink through the surface pores to enter the subdermal canals; the canal fluid will simultaneously flow toward regions of lower pressure and re-enter the boundary layer through the surface pores. This simultaneous series of events would dynamically damp the disturbance. Walters also found similar integuments in two species of Gempylidae. LaMonte ( 1958 ) reported that the skin of marlin, genus Makaira, has many minute, round open- ings. Whether such a system operates in Peprilus or in gempylids and marlin is un- known. Metabolic economy in Peprilus does not seem critical, and selection pres- sure for such a mechanism would not appear to be great. However, all the above possibilities as to function remain in question, and further investigation is re- quired. Alimentary canal. The alimentary canal with its several elaborations (Fig. 32) sug- gests that feeding and food conversion are rather specialized. Each part of the canal is considered below, and an attempt is made to correlate the part with its probable function. The jaw teeth (Fig. 33) are small, uni- Systematics and Biology of Peprilus • Horn 219 Figure 32. Alimentary canal of Peprilus paru: A = pha- ryngeal sac; B = esophagus; C = stomach; D = pyloric caeca; E :^ intestine. From 124.6 mm SL specimen (semi- diagrammatic). serial, and laterally compressed. The teeth of the upper jaw are slightly recurved, either simple and pointed (Fig. 33a), or have three small cusps (Fig. 33b). The teeth of the lower jaw are similar to those of the upper, but have either three cusps (Fig. 33a), or four or five cusps (Fig. 33b), and are not recurved. The jaw apparatus seems suited for nip- ping parts from larger objects or for taking small objects whole. The significance of the differences in cusp pattern of the pre- maxillary teeth among the species of Pepri- lus is unknown. No teeth are present on the vomer, palatines, or basibranchials. The gill rakers are long, slender, closely spaced, and beset with small teeth. They appear to serve as an effective screen in preventing small food particles from enter- ing the gill area or from escaping through the opercular opening. However, they do not appear to serve as a straining device. Small differences in the number of gill rakers among the species of Peprilus do not seem sufficient to indicate differences in food habits. The suborder Stromateoidei is character- ized by toothed pharyngeal sacs located immediately behind the last gill arch. Biihler (1930) showed the sacs to be of Figure 33. Jaws and teeth of two species of Peprilus-. A = P. paru; B = P. burti; 1 = tooth of upper jaw; 2 = tooth of lower jaw. From cleared-and-stained specimens. pharyngeal origin and worked out the morphology of them in two stromateoid families, the Centrolophidae and the Stro- mateidae. Haedrich ( 1967 ) has studied in detail the comparative structure of the sacs in the five stromateoid families. The 220 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 sac has muscular walls and appears to function as a shredding or grinding organ; and stomach contents are often shredded beyond recognition. Immediately posterior to the pharyngeal sac is the short and thick-walled esopha- gus. The stomach is a distinct, thick-walled, muscular, and roughly U-shaped organ which is capable of considerable distension. The intestine is long, about two and one-half to three times the length of the body in Peprilus, and arranged into long loops in the body cavity. The intestine of fishes is generally considered to be the site of absorption. At the anterior end of the intestine is a dendritic mass of pyloric caeca. These structures, which develop as blind outgrowths of the anterior end of the intestine, are present in a number of fami- lies of teleost fishes. The number of caeca may vary greatly, from one to more than 1000 (Suyehiro, 1942). Suyehiro estimates there to be about 600 in the stromateid fish, Pampus orgenteus. There are several hundred long, narrow caeca in Peprilus. It has been suggested (Al- Hussaini, 1947; Barrington, 1957) that the caeca are merely an adaptation which increases the surface area of the intestinal epithelium of fish in which the size of the body cavity limits the length of the intestine itself, or whose habits are car- nivorous. Both of these criteria apply to Peprilus. The body cavity is small, prob- ably because of other selection pressures for a short, compressed body. The fishes are also carnivorous, but there is no close correlation between gut length and food habits. Food of Peprilus appears to consist of not only jellyfish but also small fishes, a variety of small crustaceans, and other small invertebrates. The pharyngeal sac is formed very early in life, at least by a size of 10 mm SL, and the young appar- ently feed in a manner similar to the adults. Juvenile Peprilus may eat proportionately more jellyfish than do the adults. ASPECTS OF LIFE HISTORY AND ECOLOGY Emphasis here is placed upon the four better known species, P. triaconthus. P. burti, P. paru, and P. simillimus. Little is known of the life history of the other three species. It seems reasonably certain that the overall biology of the latter species will prove not to be substantially different from that of the other species. The fishes of this genus are similar and wherever they occur fill a similar niche in the ecosystem. Early life history. The testes, ovaries, and eggs are similar among all the species. Dioecy seems to be the rule in the genus. Eggs are buoyant, transparent, spherical, and contain one or more oil globules. Egg diameters range from 0.75 to 0.79 mm in P. triacanthus, and 70 per cent of the eggs have a single oil globule (Colton and Honey, 1963). They reported that other eggs had two or three smaller oil globules. Eggs of P. paru in Chesapeake Bay are ap- proximately 1 mm in diameter ( Hildebrand and Schroeder, 1927). Ovaries in a ripe fish occupy more than one-half of the body cavity. Maturity is reached in P. triacanthus at about two years of age (Bigelow and Schroeder, 1953), and at a length of about 140 to 180 mm SL; how- ever, I have examined nearly ripe individu- als which were as small as 120 mm SL. The size at maturity of the smaller species such as P. burti and P. ovatus is probably less than that for P. triacanthus. Hatching time for P. triacanthus is around 48 to 72 hours, and size of larvae at hatching is slightly less than 2 mm (Pearson, 1941; Colton and Honey, 1963). The identification of larval P. triacanthus and Urophycis chuss, the squirrel hake, has been confused. Pearson (1941) and Miller and Marak (1959) indicate that the early larval stages considered by Kuntz and Radcliffe (1918) to be P. triacanthus are those of Urophycis chuss. Colton and Honey (1963) have shown that eggs of P. triacanthus differ from those of U. chuss Systematics and Biology of Peprilvs * Horn 221 described by Hildebrand and Cable (1934). Spawning generally occurs in the spring and early summer, although the period of spawning may be longer in some of the warm-water species. Bigelow and Schroeder (1953) indicate that spawning of P. tria- canthus in the Gulf of Maine begins in June, is at a maximum in July, and con- tinues into August. Pearson (1941), on the basis of appearance of larvae, postulated a late spring and early summer spawning for P. triacanthus and P. paru in lower Chesapeake Bay. P. paru in the Gulf of Mexico apparently spawns in the spring (J. M. Miller, 1965). On the basis of the occurrence of young in winter and spring, Miller suggested a winter and possibly an autumn spawning for P. burti. The type of eggs, the seasonal distri- bution of mature individuals, and the larval distribution should give some indi- cation of the area of spawning. The eggs are buoyant and transparent which indi- cates that they are pelagic and near the surface. Kuntz and Radcliffe (1918), and Bigelow and Schroeder (1953), stated that spawning in P. triacanthus occurs a few miles offshore. Mature individuals are known to leave inshore waters during the spawning season, and those returning later are spent. Schaefer (1967) reported that P. triacantlius of spawning size does not appear in the surf zone until June, which indicates that before this time the fish were offshore. According to Miller ( 1965 ) , P. burti moves offshore with increasing water temperatures in the spring, which indicates that spawning is offshore. Data from M/V GILL cruises along the coast of the southeastern United States show that Peprilus sp. larvae are several miles off- shore. Larvae of P. simillimus occur both near shore and several miles offshore, but generally in large protected areas (Fig. 39). Offshore concentrations of both larvae and mature fish of this species suggest that spawning is generally a few miles offshore in surface waters. Almost nothing is known of the spawn- ing act. Aggregations of mature fish prob- ably develop in offshore waters followed by mass extrusion of reproductive products. There is probably little courtship behavior before the spawning act. No external sexual dimorphism has been detected ex- cept the distension of the abdomen of ripe females. Segregation of males and females is not known to occur; large collections usually contain both sexes in reasonably equal numbers. The eggs and early larval stages are in the surface layers and constitute part of the plankton. Ahlstrom (1959) stated that larvae of P. simillimus occurs between the surface and a depth of 48 m. Plankton volume data of the California Current Survey (Staff, South Pacific Fishery In- vestigations, 1956; Thrailkill, 1957, 'l959, 1961, and 1963), in combination with in- formation supplied by Dr. Ahlstrom, show that the larvae of P. simillimus are in the surface layers usually at depths of 50 m or less. Larvae which I identified as Peprilus sp. and which were taken on M/V GILL cruises off Georgia and South Carolina were in the surface layers to a depth of 72 m. Distribution of larvae of P. simillimus on the Pacific coast (Fig. 39) is mostly near shore and in bays or other protected areas, although many were collected at consider- able distances from shore. The greatest numbers of larvae are taken mainly around Cedros Island and Bahia Sebastian Viz- caino off upper central Baja California, and smaller concentrations are taken in the Channel Island area near Los Angeles and near Bahia Magdalena off lower central Baja California (Figs. 39 and 40). Bige- low and Schroeder (1953) stated that al- though a considerable number of eggs of P. triacanthus are produced in the Gulf of Maine, very few larvae are taken there. They reported that no young P. triacanthus have been taken in the Bay of Fundy, but that they are very plentiful along the shores of southern New England. That juveniles of 222 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Peprilus occur in bays and other inshore areas during the summer has also been shown by Reid ( 1955) for P. poru in Texas, Mansueti (1963) for P. poru and P. tria- canthus in Chesapeake Bay, and by Schaefer (1967) for P. triaconthus in the surf zone of Long Island, New York. In summary, the young of Peprilus ap- pear to move inshore after hatching in pelagic surface waters to bays and other protected areas which serve as favorable nursery grounds. This general inshore movement coincides with the presence and abundance of certain coelenterates with which the young fishes of this genus associ- ate in the summer and early fall ( Mansueti, 1963). Distribution in relation to certain physi- cal factors. Temperature. Temperature is probably the most significant factor affecting the distribution of members of the genus Peprilus. Stromateids are basically warm-water fishes and secondarily have invaded tem- perate regions. Stromateus off southern South America and off southern Africa, and Peprilus in the western North Atlantic off Newfoundland and in the eastern North Pacific off British Columbia, represent the farthest penetrations into temperate waters. Peprilus in its most northerly range is not abundant and generally is regarded as a summer visitor. Even though limited by temperature, some members of the genus do occur over a wide range of temperature. P. triaconthus off the northeastern United States is found where bottom temperatures range in autumn from 4.4° to 20.6°C (Fritz, 1965). Data from groundfish sur- veys (Figs. 34-36) off the northeastern United States show that P. triaconthus has been taken over areas where the mean bot- tom temperatures ranged from 5.0° to 14.2°C, these figures representing the total range for summer, autumn, and winter for the years 1963 to 1966. Schaefer (1967) has collected P. triaconthus in the surf zone of Long Island in August in water tempera- tures as high as 21.6°C. P. burti, a close relative of P. triaconthus, has been taken in waters of temperatures as high as 28°C and as low as 12.6°C (Gunter, 1945). P. para has been collected in waters ranging in temperature from 13.7° to 30.0°C (Gunter, 1945) and from 14.6° to 27.5°C (J. M. Miller, 1965). P. paru is a shallow- water species, usually in waters of less than 50 m depth, and probably does not en- counter as wide a range of temperatures as does P. triaconthus. P. simillimus on the Pacific coast probably encounters a range of temperatures similar to that of P. tria- conthus since their latitudinal distributions are complementary. The largely tropical and subtropical spe- cies of the Pacific, P. medius, P. ovatus, and P. snyderi, are almost certainly limited to a narrower, warmer range of tempera- tures than the above species. Salinity. The habit of entering bays and even river mouths indicates that members of this genus are euryhaline. Reid ( 1955 ) collected a juvenile P. paru at a bayou mouth in a salinity of 10.3 ppm and two others at salinities of 17.3 and 17.6 ppm. Schultz (1962) collected P. paru in Texas bays in salinities ranging from 18.9 to 19.5 ppm. Gunter (1945) caught P. paru in the Gulf of Mexico in high salinities, 33.0 to 36.7 ppm. P. burti has also been collected in waters of widely varying salinity. Gun- ter (1945) reported that P. burti occurs in a salinity range of 15.6 to 35.2 ppm, but mainly in salinities over 30 ppm. Schultz (1962) gave a range of 21.5 to 25.9 ppm for P. burti in Texas bays. Schaefer (1967) recorded a salinity range of about 29.0 to 33.0 ppm for the surf waters of Long Island in which P. triaconthus is an abun- dant species. P. burti and P. poru have both been collected in Lake Pontchartrain, Louisiana (Tulane University Collections), a lake of varying salinity. P. ovatus has been taken at the mouth of the Colorado River (SIO 63-484) in the northern Gulf of California. Currents. It is possible to correlate ex- Systematics and Biology of Pepriu - Horn 223 isting current patterns with the distribution P. ovatus in the northern Gulf of of almost every species of Peprilus. The has been collected over mud ar influences of currents in the distribution bottoms. P. simillimus is found ove of the species of Peprilus is discussed be- sand and mud bottoms. low in the section on zoogeography. Depth. Fishes of this genus may Sediment or bottom type. In general, found in waters as deep as 420 m on the members of the genus Peprilus are distrib- continental slope (Table 22), but more uted either over a sand or a mud bottom frequently are found in much shallower or a combination of both. None of the spe- waters. cies are residents of rocky shores or coral Fritz ( 1965 ) reported that P. triacanthus zones, though they may occur in such an occurs in a depth range of 30 to 270 m off area temporarily. The preference for either the northeastern United States, but stated a sand or a mud bottom is not usually that the species is most abundant in 50 to species specific. 90 m of water. Data for other seasons P. triacanthus off the coast of the north- (Figs. 34, 36, and 37) do not expand the eastern United States is usually distributed above depth range. over a sandy bottom or a combination of Table 22 gives the depth ranges of P. silt and sand. Fritz (1965) showed that triacanthus, P. paru, and P. burti off the the autumn distribution of this species is southeastern United States and in the Gulf over the latter type. Bigelow and Schroeder of Mexico. As mentioned earlier in the ( 1953 ) stated that this species has such a discussion of sediment type, P. triacanthus decided preference for sandy bottoms in its southern distribution has a wide rather than for rocky or muddy bottoms depth distribution correlated with the pos- that few are taken in traps on muddy sible existence of two separate populations ground, while other traps along a sandy or species in that region. This species has beach may yield considerable numbers, been taken in depths ranging from 20 to Leim and Scott (1966) report that small 420 m off the southeastern coast of the schools of P. triacanthus occur over sandy United States. The closely related P. burti bottoms along the Canadian coast during in the Gulf of Mexico has a less extensive the wanner months. In the southern part depth range with greatest abundance at a of the range of P. triacanthus there appear range of about 9 to 135 m. P. paru is a to be two distinct populations: one pri- relatively shallow-water form which occurs marily in deep water and distributed over in inshore areas throughout the year, a mud bottom; and the other in shallow, The depth distribution of P. simillimus inshore water over a sand bottom ( Table is probably similar to that of P. triacanthus. 22). Most of the shelf area off the south- There is little information on the vertical eastern United States is sandy with some distribution of the tropical Pacific species, small areas of mud and gravel ( Moore and P. snyderi, P. medius, and P. ovatus have Gorsline, 1960). While there seems to be been taken in inshore areas in depths of a preference of each apparent population less than about 50 m. P. ovatus probably for either a sand or a mud bottom, there does not occur beyond this depth while P. is overlap and no strict segregation. The snyderi may descend to greater depths. P. apparent existence of two distinct popu- medius seems to compliment in the Pacific lations in the Atlantic off the southeastern the ecology of P. paru in the Atlantic; thus, i states is discussed on p. 247. its depth range may be similar to that of P. burti of the Gulf of Mexico is most the latter species, frequently distributed over a mud or a Seasonal distribution and abundance. sand-silt bottom (Table 22). Seasonal movements are generally char- There are few data on the other species, acteristic of the more temperate species of 224 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Table 22. Summary of the Depth of Capture and Associated Major Bottom Type of Collec- tions of Three Species of Peprilis Made by the M V OREGON and M V SILVER BAY in the Southwestern North Atlantic (Cape Hatteras to Southern Florida), and in the Gulf of Mexico for the Years 1950 through 1966. Species Region Depth Range ( meters ) Bottom Type P. pant Atlantic 6-88 sand and/or mud P. pant Gulf of Mexico 2-115 sand and/or mud P. burti Gulf of Mexico 2-274 sand and/or mud; infrequently coral P. triacanthus (deep-bodied) Atlantic (generally inshore) 7-145 usually sand; less often gravel, mud, or mud and sand P. triacanthus (shallow-bodied: often spotted) Atlantic (generally further offshore) 20-420 usually mud; less often sand, coral, or mud and sand the genus. Seasonal migrations apparently serve to maintain the species within a favorable range of environmental con- ditions. The seasonal movements of P. triacanthus (the butterfish) have been generally recognized for some time. Bigelow and Schroeder (1953) indicated that while definite evidence is lacking, butterfish in the Gulf of Maine and southward seldom appeared to descend deeper than 25 to 55 m during the summer, and to spend the winter and early spring near bottom in depths down to about 180 to 210 m. Fritz (1965) showed that in the same area but- terfish are in shallow, warm waters during autumn and are abundant at depths of 50 to 90 in and at a water temperature of 11 °C. He suggested that the fish migrates north- ward into the Gulf of Maine during the summer months. Pearson (1932) reported that small quantities of butterfish have been trawled in offshore waters in winter in Virginia and North Carolina. Leim and Scott (1966) stated that small schools of butterfish occur over sandy bottoms along the Canadian coast during the period of warm water from May to November, but are offshore in depths to 100 fms (183 m) in the winter. Schaefer (1967) found that butterfish of a wide size-range occur in the surf zone of Long Island from June through October. The distribution of P. triacanthus in sum- mer, autumn, and winter over a three-year period in the Gulf of Maine and in coastal waters south to just beyond 40°N is pre- sented in Figures 34 to 36. Length-fre- quency values of the total catch calculated from a sample length-frequency are given for each seasonal catch. Figure 37 shows the depth zones for the strata of this geographic region. The strata maps of Figures 34 through 37 were designed, de- scribed, and presented in extended form by Grosslein (1969). P. triacanthus is widespread both in in- shore waters and over deeper waters near the edge of the continental shelf in the summer months (Fig. 34). The fish occurs to a limited extent in the Gulf of Maine, especially in the southern portion. The greatest number of individuals were col- lected in strata nine and ten (Fig. 37), and in waters of less than 55 m in depth. The species was collected over an area in which the mean bottom temperatures ranged from 5.8° to 13.3°C and in waters ranging from 40 to 247 m in depth. The autumn distribution (Fig. 35) is similar to that of the summer distribution. Systematics am) Biology of Peprili s • Horn 225 FREQUENCY , ^^ N= 8825 J OVER 1000 Figure 34. Summer (July and August) distribution of the butterfish, Peprilus triacanthus, off the northeastern coast of the United States, 1963-o5. Data compiled from R/V ALBATROSS IV groundfish survey cruises of the U. S. Bureau of Com- mercial Fisheries, Woods Hole, Massachusetts. Number of fish per stratum is the cumulative total catch per stratum for the three-year period. Length-frequency is of the total catch calculated from the sample length-frequency; N = total catch of all strata for the three-year period. Strata map courtesy of Marvin D. Grosslein. The greatest numbers of individuals were collected in stratum nine (Fig. 37), in waters of less than 110 m in depth. The species was collected over an area in which the mean bottom temperatures ranged from 5.8° to 14.2 °C and in waters ranging from 33 to 250 m in depth. The winter distribution (Fig. 36) is different from that of either the summer or the autumn. The winter collections made in the same region and with reason- ably the same fishing effort, indicate that the species has a more restricted horizontal distribution during this season. The dis- tribution is entirely south of the Gulf of Maine and in deep water. No fish were collected in the shallowest strata. The greatest numbers of individuals were cap- tured in stratum seven (Fig. 37), in which the depth range is 112 to 183 m. Con- siderable numbers were collected in deeper water near the edge of the continental shelf. Collections were made over an area in which the mean bottom temperatures ranged from 5.0° to 10.2°C. The spring distribution (no map, data for 1968 only ) is similar to the winter distri- bution. The greatest numbers of individuals were collected in deep water along the edge of the continental shelf. The fish was 226 Bulletin Museum of Comparative Zoology, Vol 140, No. 5 i FREQUENCY 10 N = 6101 101-1000 OVER 1000 -Z. Figure 35. Autumn (October, November, and early December) distribution of the butferfish, Pepnlus triacanthus, off the northeastern coast of the United States, 1963-65. Explanation of symbols and other information in Figure 34. captured in waters ranging in depth from 44 to 274 m. In the summer months along the coasts of the middle and northeastern Atlantic states, P. tricanthus has a wide, shallow- water distribution with movements both inshore and northward. The autumn dis- tribution is similar. During the winter the species becomes restricted horizontally and concentrated in deep, offshore waters near the edge of the continental shelf. This winter concentration appears to be in- fluenced strongly by water temperatures, and even though the seasonal distributions are different, the ranges of mean bottom temperatures strongly overlap for all sea- sons. The early spring distribution is similar to the winter distribution. In late spring, May to June, the species moves inshore and northward as upper level water temperatures begin to rise. The fish then assumes the summer distribution described above. As spawning continues or is completed in late spring and summer, subadult and mature individuals of P. triacanthus range widely in shallow waters and occur in the surf zone from June through October on Long Island (Schaefer, 1967). Schaefer has shown that in June surf zone catches are mainly of large individuals, many of which are large enough to have already spawned offshore. In July, fish of a wider size-range are captured in the surf zone. The length-frequency diagram presented by Schaefer for August is bimodal with all Systematics and Biology of Peprilvs • Horn 227 ". FREQUENCY Figure 36. Winter (January and February) distribution of the butterfish, Peprilus triacanthus, off the northeastern coast of the United States, 1964-66. Explanation of symbols and other information in Figure 34. fish of 100 mm FL (fork length in P. tria- canthus approximately 10 per cent greater than SL) or less. This indicates that juve- niles spawned earlier in the year are now entering the surf waters along with larger juveniles. Still larger fish were captured in August, September, or October. These large individuals probably move long dis- tances during the summer and autumn even into waters as far north as Newfound- land and the Gulf of St. Lawrence. The seasonal distributions and length- frequency diagrams (Figs. 34-36) do not include surf zone catches but only captures from trawlable waters. The length-fre- quency data seem to correlate well with Schaefer's (1967) data from the surf zone. In summer (Fig. 34), the diagram is essentially unimodal at around 140 mm FL, but it also indicates the presence of juveniles of 50 mm or less. All the fish were collected with a standard 36' trawl having a one-half inch mesh codend liner; the use of a finer mesh probably would have resulted in the capture of more small fish. The data for autumn (Fig. 35) are weakly bimodal and consist mostly of large individuals of around 160 mm FL. This corresponds with the surf zone data which indicated that such large fish were absent from the surf in autumn. This further suggests that different size groups move independently, at least during the summer and autumn. The length-frequency dia- gram for winter (Fig. 36) is strongly bimodal with one peak around 100 mm, 228 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Figure 37. Strata map with depth zones corresponding to Figures 34 to 36. another around 180 mm, and has a total size-range of 70 to 230 mm FL. The length- frequency diagram for the spring (Fig. 38) is similar to the one for winter. The curve is bimodal with one peak at around 100 mm and another at around 140 mm, but with many larger individuals. The total size-range is from 70 to 210 mm FL. These data for the winter and spring months indicate that two- and possibly three-year classes are concentrated in deep water dur- ing these seasons. Thus, there seem to be wide dispersal and independent movement of size-classes in the warmer months, and a concentration of a large part of the popu- lation in deep water in the colder months. The seasonal distribution of Stromateus maculatus (=Stromateus brasiliensis) off the southern Atlantic coast of South America parallels in some respects that just described for P. triacanthus in the western North Atlantic. T. J. Hart (1946) described the main features of the sea- sonal movements of S. maculatus on the Patagonian shelf as follows: a double trend of seasonal movement - inshore in spring and summer with maximum con- centration, offshore in autumn and winter with maximum dispersal; and, a meridional movement southward in summer and north- ward in winter. The latter aspect is similar to that of P. triacanthus, but, unlike the pattern of S. maculatus, maximum dispersal of P. triacanthus seems to occur in summer and autumn with maximum concentration in a restricted region occurring in the winter and early spring. A number of fishes with warm-water af- Systematics and Biology of Peprilus • Horn 229 130 150 Fork Lenglh mm Figure 38. Length-frequency diagram for the spring (March, April, and early May) collection of butterfish, Peprilus triacanthus, made by R/V ALBATROSS IV groundfish survey cruises of the U. S. Bureau of Commercial Fisheries, Woods Hole, Massachusetts, off the northeastern coast of the United States, 1968. Length-frequency is of the total catch calculated from the sample length-frequency. See Figures 34 to 37. finities migrate seasonally in a manner simi- lar to P. triacanthus. Atlantic mackerel, Scomber scombrus, scup, Stenotomus versi- color, and weakfish, Cynoscion regalis, among others, migrate coastwards to their spawning grounds in spring and summer; before winter advances, when shallow, in- shore waters become colder than those off- shore, these species move offshore to warmer, deeper waters (Marshall, 1966). Whereas, cold-adapted fishes such as had- dock. Melanogrammus aeglefinus, and pol- lock, Pollachius virens, may retreat in sum- mer from warm, inshore waters to cooler, deeper reaches. Cape Cod is often recog- nized as the southern boundary of most boreal species and the northern boundary of many temperate species (Fritz, 1965). Some fishes are restricted to either north or south of Cape Cod while others range seasonally into the adjacent region. The distribution of P. triacanthus largely is south of Cape Cod except in the summer and autumn when it migrates northward into the Gulf of Maine. According to Fritz, Atlantic cod, Gadus morhua, a typical boreal species, may migrate as far south as Virginia during the winter months when temperature conditions are suitable. Silver hake. Merluccius bilinearis, and spiny dog- fish, Squalus acanthias, for example, range widely both north and south of Cape Cod. The other species of Peprilus which might be expected to migrate seasonally is P. simillimus, since it ranges into temperate waters as far north as the southern coast of British Columbia. Ulrey and Greeley (1928) reported that the fish is abundant in summer especially around Santa Cruz, California. Recent publications (J. L. Hart, 1949; Batts, I960; High, 1966) have cited either the increased abundance or local oc- currences of P. simillimus in Washington and British Columbia waters. Hart also pointed out that an increase in poundage of landings had occurred in the Monterey Bay area of California, whereas, previously, the major portion of the landings had oc- curred in southern California. Hart's list of occurrences of the species in British Columbia waters mostly includes records for the winter months of December, Janu- ary, and February. The distribution of P. simillimus larvae off the coasts of California and Baja Cali- fornia (Fig. 39) strongly suggests the warm-water affinities of the species. From 1955 to 1959, larvae of this fish have not been collected north of Point Conception, California, about 34°N. The greatest abundance of larvae was taken in the upper central Baja California area (Figs. 39 and 40), mainly around Bahia Sebas- tian Vizcaino. Larvae were collected in greatest numbers in the months of May, June, and July (Fig. 41), although they were taken from January through August. There seems to be no significant time difference between the appearance of larvae in the northern (Point Conception) and in the southern (southern Baja Cali- fornia) parts of the distribution. Berry and Perkins (1966), in a survey of the pelagic fishes of the California Current area from central California to central Baja California, reported that P. simillimus ranging in size from 12 to 165 mm SL occurred from Point Conception to the vicinity of Cedros Island near Bahia Sebastian Vizcaino, Baja California. This distribution agrees closely with that of the larval distribution just described. 230 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 130° 40' 35' 30 2C. NcP 115° 110° i r CALCOFI BASIC STATION PLAN SINCE 1950 POINT CONCEPTION 40" 35° P ^ 3 *° " ". V • ' * . • * • * n I .•* * *S * V^«- XX #V^TtX3 Lf** \ 4 .-v>.--*.,%i _ X Captures of R simi 1 1 imus Larvae _ • Stations Occupied / 20' / ++ ' „ '°. * .P • * * • * J I L I I I 30° 20° I30" 125" 120° 115° Figures 39. Distribution of larvae of Peprilus simillimus off the coasts of California and Baja California, 1955-59: 1 r= northern California; 2 = central California; 3 = southern California; 4 — northern Baja California; 5 = upper central Baja California; 6 = lower central Baja California; 7 = southern Ba|a California. Data compiled from collections made by California Cooperative Oceanic Fisheries Investigations (CALCOFI). Map courtesy of Elbert H. Ahlstrom. Part of the adult population may move of an increase in the northern abundance, northward long distances along the Pacific yet the occurrence is still of sufficient coast during favorable periods. As has novelty that reports continue to be pub- been mentioned, there have been reports lished of the fish's capture in such northern Systematics and Biology of Peprilus • Horn 231 Latitude Figure 40. Relative regional abundance of larvae of Peprilus simillimus according to CALCOFI collections, 1955-59: 3 zz southern California; 4 zz northern Baja California; 5 ~ upper central Baja California; 6 zz lower central Baja Cali- fornia; 7 zz southern Baja California. The number of larvae per 10 square meters of sea surface is an estimate calculated from standard haul totals. See Figures 39 and 41. areas as Washington and British Columbia. The species remains to be of commercial importance only in California and mainly from Monterey to San Pedro. Surface currents probably influence the extent of the northern distribution of P. simillimus. The data on the occurrence of this species in Washington and British Columbia are all similar in that the ma- jority of captures are in the winter months, November through February, and that the size of the fish captured has been 150 mm FL or more. During the season of upwell- ing on the California coast (approximately March to July) a countercurrent (counter to the southward-flowing California Cur- rent) that contains considerable quantities of equatorial water flows close to the coast at depths below 200 m (Sverdrup, John- son, and Fleming, 1942). In the fall up- welling ceases, and in the surface layers S 500 ° 300 M J J A Months, 1955-59 Figure 41. Cumulative monthly catch of larvae of Peprilus simillimus off the coasts of California and Baja California made by CALCOFI survey cruises, 1955-59. See Figures 39 and 40. a countercurrent develops, the Davidson Current, which in November, December, and January flows northward along the coast to at least latitude 48°N. This latter current may particularly influence the movements of P. simillimus. Individuals may follow this current northward, and this may at least partially explain the captures of this species in northern areas in the winter months. In addition, the greatest numbers are landed commercially in Cali- fornia in the summer and fall with a de- cline in the catch in the winter and early spring (Fig. 43). However, at least two factors cast some doubt on the proposal of a northward mi- gration in winter and rather suggest the possibility of a resident, breeding popu- lation in the northern part of the range. 232 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 One factor is the gonadal condition of northern individuals. I have examined both male and female fish from Washington and British Columbia which had maturing and nearly mature gonads. These were fishes of 150 mm SL or larger. This is evidence for a northern spawning unless a southern migration occurs first. Eggs and larvae of this species have apparently not been found in northern waters. A second factor which suggests a resi- dent northern population is the apparent geographic variation which occurs in the length of the caudal fin. Individuals oc- curring from San Francisco to British Columbia have significantly shorter caudal fins than do those fishes taken from Monte- rey to Baja California (Fig. 14). Northern individuals also seem to have lower median fins than do southern forms. The im- portance of these variations is difficult to assess but it is suggestive of a distinct, resident population in the north, or at least of a clinal effect due to changing tempera- ture conditions. The differences may onlv be between northern and southern Cali- fornia populations with those having the shorter caudal fin being resident off north- ern California and periodically migrating into more northerly waters. While most northern occurrences have been in the winter months, P. simillimus has been collected at other times of the year. J. L. Hart (1949) reported that on 29 July 1946, about 150 pounds of this fish were taken in a bait seine in Bargain Harbor, British Columbia. Hart also quoted a crewman as saying that the species occurs throughout the year in British Columbia waters. More information is needed in order to resolve the problems of breeding range versus geographic range and of seasonal migrations in P. .simillimus. P. paru, widely distributed in tropical, subtropical, and warm temperate waters in the Atlantic, is probably the most coastally-oriented fish of the genus. This species apparently does not migrate sea- sonally to offshore areas, but rather re- mains in relatively shallow water through- out the year (Table 22). H. H. Hildebrand (1954) caught P. paru during all months of the year on the shrimp grounds in the western Gulf of Mexico. Compton and Bradley (1963) reported that the fish was abundant in March at depths of 6 to 15 fms (11 to 27 m) in Texas coastal bays. J. M. Miller (1965) collected subadults in depths of 6 fms or less in the Gulf of Mexico near Port Aransas in February, March, and June. Copeland (1965) classed P. para as a rare emigrant at Aransas Pass Inlet, Texas, and caught it in a tide trap in all seasons. I have collected mature and nearly mature individuals in the same tide trap during late June and early July. P. paru also appears to occur throughout the year in Chesapeake Bay on the Atlantic- Coast. P. burti occupies in the Gulf of Mexico a less extensive depth range than does P. triacanthus on the Atlantic Coast (Table 22). The species does, however, season- ally migrate to a limited extent. Gunter (1945) collected P. burti in Aransas Bay, Texas, in March, August, and December; and in the Gulf of Mexico in March, April, May, June, and November. In Louisiana, the fish was taken most abundantly by Gunter in shallow areas in March. H. H. Hildebrand (1954) noted the presence of this fish on the shrimp grounds in the Gulf of Mexico throughout the year, but he stated that it was in greatest abundance in May and July, while during the winter it appeared sporadically and never in great abundance. Compton and Bradley (1963) found that it was one of the five most abundant fishes in Texas bay areas in March. They also stated that the fish breeds at a time of rising or high temperatures. According to J. M. Miller (1965), P. burti moves offshore with in- creasing water temperatures in the spring. Copeland ( 1965 ) listed this species as an occasional emigrant at Port Aransas Inlet, Texas, taking it in tide trap collections Systematics and Biology of Peprilus • Horn 233 from November through May and most abundantly in April. The pattern seems to be that individuals are in shallow inlets and bays during the winter and earlv spring; then as the water temperature increases, the fish move offshore into the Gulf, yet not into much deeper water. Spawning probably occurs offshore in the surface layers of the Gulf during late spring, summer, and perhaps early fall. The seasonal movements of P. burti seem to correlate with the temperate affinities of the species. The movements are the reverse of that of its close relative, P. triacanthus, but similar in that both remain in temperate waters. P. burti avoids subtropical temperatures, and P. tria- canthus migrates to avoid boreal tempera- tures in the northern part of its range. Little information is available on the other three species. P. ovatus is a shallow- water species of restricted distribution in the northern Gulf of California and prob- ably does not migrate seasonally. The spe- cies has been collected at depths of 45 m or less. P. medius is often collected in shal- low water, including bays and inlets and in depths of 90 m or less. This species prob- ably migrates seasonally only to a limited extent. P. snijderi has been collected in inshore areas; however, its rareness in collections suggests that it may periodically inhabit deep water. Faunal associations. At a size of about 10 to 30 mm SL, a portion at least of the populations of the species of Peprilus be- comes associated with scyphomedusae and siphonophores ( hereafter, both referred to as jellyfish medusae ) . This association may continue until the fish reach a size of 100 mm SL. The relationship is a widely rec- ognized and well-documented one. Man- sueti ( 1963 ) has reviewed the literature of the subject and has provided new data on the symbiotic behavior between P. pant and the jellyfish, Chrysaora quinque- cirrlia. The association is widespread among the stromateoid fishes according to Haedrich (1967). The relationship also oc- curs in the Carangidae, a pelagic family, in the Gadidae, Girellidae, and Centriscidae, all demersal families, and in the Zaproridae, an abyssal family ( Mansueti, 1963 ) . The importance of the association in the life of these fishes is not well understood. The young fish may gain several advan- tages from the relationship including pro- tection from predators, a source of food (the jellyfish itself), a concentration of food resulting from presence of small in- vertebrates among the tentacles and under the bell of the jellyfish, a means of dis- persal, and possibly a means of reducing interspecific competition. Walford (1958) stated that jellyfish are among the most valuable animals in the sea in that they give shelter to the young of a number of species of fishes. He added that the small fish travel with their host in the plankton, feeding around it within a radius of a few feet, darting to safety beneath the um- brella when threatened by enemies. The fish continue this mode of life as long as it is advantageous to them, e.g., until they are ready to become independent. Walford suggests that this association may be an essential stage in the life cycle of some fishes. At least in the stromateid fishes, and particularly in Peprilus, the association may not be obligatory. I have collected two species of Peprilus, P. burti and P. paru, which were not with jellyfish, as have others for P. triacanthus and P. paru (Bige- low and Sehroeder, 1953; Mansueti, 1963). Mansueti classifies the association as a temporary ecological phenomenon result- ing from a series of extrinsic chance fac- tors in which jellyfishes are the passive hosts and fishes the active opportunists. P. paru begins a commensal association by consuming plankton and other organisms found on or near the hosts and then gradu- ally becomes parasitic as it feeds on parts of the medusa. Finally, in autumn, as a large non-symbiont, it becomes predatory upon the medusae. Cargo and Schultz (1966) stated that schools of P. paru larger 234 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 than 50 mm in length may destroy large numbers of scyphozoan medusae in minutes. The fish itself is apparently rarely killed or eaten by the medusae. According to Mansueti ( 1963 ) , the nature of the sym- bioses suggests that most associated fishes have some immunity to jellyfish toxin, and that either the immunity is gradually ac- quired through gradual adaptation, or that some or all fish symbionts may possess natural antibodies. A protective agent may be present in the mucous secreted by symbiotic fish which may either inhibit nematocyst discharge or counteract the toxin after stinging occurs. Species of Peprilus do secrete considerable amounts of mucous, and the extensive canal system may be involved in achieving immunity as suggested by Miner (1936). The temporary association between Peprilus and jellyfish may be discontinued because of several factors: rapid growth of the fish and feeding independently from the jellyfish; the physical size attained by the fish; seasonal changes in hydrographic conditions - cold water temperatures in late autumn resulting in the death of certain jellyfish such as Chnjsaora in Chesapeake Bay (Mansueti, 1963); or, similar to the first factor, a change in food habits of the fish, such as feeding in deeper waters. The association does not involve the entire population of the fish species, which is to be expected since it is considered to be a fortuitous relationship. Stromateid fishes are known to be nonsymbiotic at the same size and in the same months as symbiotic stromateid fishes. The contacts between fish and jellyfish may be explained by the random search of the fishes for concentrations of food with such move- ment bringing the two organisms into close association. The coincidence of a period of pelagic existence among the young fish in the same area of a high incidence of pelagic jellyfishes seems important in de- veloping the association. The association is not species-specific, at least in P. paru, P. triacanihus, and P. burti. P. paru is known to consort with five jelly- fish genera, Chnjsaora, Chiropsalmus, Cyanea, Aurelia, and Physalia (a siphono- phore). P. triacanthus has been found with Chnjsaora, Cyanea, Stomolophus, and Physalia. P. burti in the Gulf of Mexico is known to associate with Stomolophus meleagris and Aurelia sp. Specimens of P. simillimus of about 30 to 50 mm in length have been found with Pelagia (SIO H52-221 and 60-409). This information suggests that the association is a loose one and gives additional support to the theory that the relationship is largely a fortuitous one and not geared to a species-specific, obligatory condition. The association is unique in the sense that the symbiont is not physically attached to the host (jelly- fish) but is free in the immediate vicinity. This may account for the looseness of the relationship and increases the difficulty of attempting to categorize the association. The young of many pelagic fishes gather under a variety of animate and inanimate floating objects at sea. The objects may be seaweed, large living animals, driftwood, or rafts of various origins. Gooding and Magnuson (1967) refer to the floating object in the pelagic environment as being a relatively rare "superstrate" in an en- vironment notable for its horizontal homo- geneity. This superstrate, they suggest, has some of the same ecological signifi- cance to certain pelagic fishes that a substrate has to inshore fishes. They cite shelter from predation as probably the most significant factor in the evolution of fish communities that gather beneath in- animate drifting material in the open ocean. The fish-jellyfish association seems to be different in that it persists where the in- animate relationship may terminate (Man- sueti, 1963). The fish-jellyfish relationship, even though temporary itself, seems to be of a more sustaining nature in that the host provides shelter, protection, and food. If precise categorization and a more Summer Autumn Winter (Jul-Aug) (Oct-Nov-Dec) (Jan-Feb) X X X X X X X X X X X X X X X X X X X X Systematics and Biology of Peprilus • Horn 235 Table 23. List of the Tex Fish Species Seasonally Collected Most Regularly and Abundantly with the Butterfish, Peprilus triacanthus, by R/V ALBATROSS IV Groundfish Survey Cruises (U. S. Bureau of Commercial Fisheries, Woods Hole, Mass.) off the Northeastern Coast of the United States for the Years 1963 through 1966. Species Spiny dogfish (Squahis acanthias)* Little skate ( Raja erinacea ) * Haddock {Melanogrammus acglefinus) Silver hake (Merluccius bilinearis)* Squirrel hake (Urophycis chuss)* Alewife (Alosa pscudoharengus) Yellowtail flounder ( Limanda jerruginea ) Winter flounder (Pseudoplcuronectes americanus) Fourspot flounder (Paralichlhys oblongus) Longhorn sculpin (Myoxocephalus octodecimspinosus) * Butterfish known to be a part of diet. complete understanding of the fish-jellyfish particularly north of Cape Cod (41°N) association are to be achieved, underwater (Fritz, 1965). P. triacanthus occurs north observations in the natural environment of Cape Cod only in the summer and early and further careful laboratory studies are autumn when temperatures are warmer; necessary. hence, the two species are collected to- Table 23 is a list of the species of fishes gether only in summer and autumn. which were collected in greatest abundance Yellowtail ( Limanda jerruginea ) are also with P. triacanthus over a three-year period collected with P. triacanthus mainly in the during three seasons, summer, autumn, summer and autumn. The former species winter, off the northeastern coast of the is apparently confined to shoal waters United States (Fig. 37) by groundfish (Bigelow and Schroeder, 1953) which P. surveys of the Bureau of Commercial triacanthus abandons during the winter Fisheries, Woods Hole, Massachusetts. A for deeper water. The longhorn sculpin number of the species are large carnivores (Myoxocephalus octodecimspinosus) is a and feed on small fishes. Some of them winter spawner, probably in shallow, are known to prey on P. triacanthus, and coastal waters (Bigelow and Schroeder, probably all the species listed in Table 23 1953 )? wnich would explain its absence occasionally feed on this species. The in areas of concentrations of P. triacanthus importance of P. triacanthus as a forage ^ the winter The Httle skate (Rfl/fl species becomes especially evident since it . v , ., . . „ n„ „„jar / T>c.„1,jr> , ,, r T./-1 • r erinacea) and the winter flounder ( tseuao- probablv forms a substantial portion ot 7 ' . >. , .. , jwu,^ £i j. i r , ■ 11 1 ii pleuronectes americanus) are bottom fishes the diet ot such commercially valuable r , ,. , species as haddock, silver hake, and squir- u/ 1 /„. , -, >. 1 i rgj hake about 136 m (Bigelow and Schroeder, The season of the year that certain of 1953)> ^vhich wou1^ tend to explain why the species listed in Table 23 occur or are both sPecies are taken Wlth P- trtacatUhus collected most abundantly with P. tria- only in the summer and autumn. Three canthus reflects the habits and temperature species, silver hake (Merluccius bilinearis), affinities of these species as compared with squirrel hake ( Urophycis chuss), and spiny the same characteristics of P. triacanthus. dogfish (Squalus acanthias), range widely The haddock (Melanogrammus aegjefinus) both north and south of Cape Cod at is a boreal species inhabiting cool waters, various depths (Fritz, 1965), and, accord- 236 Bulletin Museum of Comparative Zoology, Vol 140, No. 5 ingly, the three are captured in abundance in all three seasons with P. triacanthus. General food habits. Stomach contents of the fishes of this genus are often very difficult to identify. The type of material making up a portion of the diet and the shredding action of the teeth of the pharyn- geal sac frequently result in stomachs containing chunks of whitish or trans- parent, amorphous tissue. No attempt has been made to present a quantitative ac- count of food habits of any species. Rather, a general summary is given of the results of examination of food items which I have made in three species of Peprilus and of reports in the literature of food studies in this genus and in the related genus Pampas. The teeth, mouth size, and digestive tract are very similar among all the species of Peprilus. Food habits and feeding be- havior are probably similar in all species of this genus. Differences which do exist may be quite subtle and might result from shifts in diet with age, from an increase in size associated with a shift in mode of life, from seasonal availability of certain food items, or from niche separation and char- acter displacement in sympatric situations. The latter possibility deserves investigation. The species of Peprilus consume a variety of food items. Bigelow and Schroeder ( 1953) stated that P. triacanthus in the Gulf of Maine feeds on small fish, squid, crusta- ceans such as shrimp and amphipods, an- nelid worms, and to a lesser extent, cteno- phores. Haedrich (1967) reported that the shredded transparent tissues which predom- inated in the stomachs he examined were probably jellyfish, ctenophores, and salps. I have examined the stomach contents of freshly captured specimens of three species, P. triacanthus, P. burti, and P. paru. In one group of specimens of P. triacanthus ranging in size from 24 to 41 mm SL, the stomachs were generally full of chunks of whitish material which was identified as parts of jellyfish, especially the tentacles. Jellyfish remains constituted the bulk of the stomach contents of 20 specimens of P. burti ranging in size from 55 to 100 mm SL. Crustaceans including small amphipods, shrimp, and zoea larvae, and postlarval fish were also present in the stomachs of these fish. The stomachs of a few large (up to 150 mm SL ) specimens of P. paru contained jellyfish remains and small crustaceans. The association of young Peprilus with jellyfish medusae seems to be correlated with the feeding habits of the fish. That the young fish feed on their jellyfish hosts has been reported on several occasions (Miner, 1936; Mansueti, 1963; Cargo and Schultz, 1966; and others). Mansueti made monthly collections of symbiotic P. paru and found that the stomachs of these fish invariably contained masses of whitish, coagulated material which he identified as parts of scyphomedusae. A small amount of other food items, mainly small inverte- brate remains, were also found in the stomachs of these fish which ranged in size from 21 to 80 mm TL. Stomachs of small, nonsymbiotic P. paru examined by Mansueti also contained masses of co- agulated medusan tissue, some small in- vertebrates, and other material, largely unidentifiable. The young fish considered to be nonsymbiotic may actually be sym- biotic except that they were away from the host when captured. The association is admittedly a loose one. Although medusae are usually considered to be low in nutrients and high in water content per unit volume, young Peprilus seems certain to derive some nourishment from the protein structure of jellyfish. Mansueti (1963) found that in aquaria, P. paru, satiated and markedly distended from voracious feeding on Chrysaora, rapidly lost the distension suggesting the efficient utilization of jellyfish tissue and absorption of the liquid contents. Jellyfishes may be an important source of protein for young fishes during the critical early stages of growth in the summer and autumn. Changes in the feeding and other be- havioral patterns appear to occur in young Systematics and Biology ok Peprilvs • Horn 237 Peprilus by late autumn. Autumn seems to be the period when the symbiotic re- lationship dissolves, at least in Atlantic coastal waters, particularly around Chesa- peake Bay. This separation may be fol- lowed by a short period of voracious feed- ing upon scyphomedusae and ctenophores (Dunnington and Mansueti, 1955; Cargo and Schultz, 1966). The end of the associ- ation often correlates with a loss in hosts, especially of Chnjsaora which undergoes mass mortalities with a drop in water temperatures in late October and early November ( Mansueti, 1963 ) . The pattern seems to consist of feeding largely upon jellyfish until a size of about 100 mm SL is reached. The association ends as the fish increase in size and hydro- graphic conditions change. This is fol- lowed by continued feeding upon jellyfish but to a lesser extent. The diet then con- sists of proportionately more crustaceans, worms, squid, and small fish. Suyehiro ( 1942 ) reported that another stromateid fish, Fampus argenteus, feeds principally on jellyfish in Japanese waters, although his conclusion was based on limited data. Nath (1966) in a study of the fishes off the Travancore coast of India found that Pampus argenteus was a macro- plankton feeder and that the variation in diet of the species depended largely upon the seasonal changes in abundance of crustaceans and polychaetes. From No- vember through March amphipods and copepods are abundant, and during this period these crustaceans predominated in the stomachs of Pampus argenteus. In April, the diet consisted almost entirely of crustaceans. By June, the crustaceans de- clined and were replaced in the diet by polychaete worms until September. Nath did not find medusae in the stomachs of this species. Kuthalingam (1963) found results similar to those of Nath with the same species in the Bay of Bengal. He reported that young Pampus argenteus ate primarily copepods and other small crusta- ceans, while the adult fish are principally a variety of small crustaceans, as well as larger crustaceans, polychaetes, and small fish. Menon (1942) stated that larger specimens of Pampus chinensis off the Trivandrum coast of India feed on fishes, mainly anchovies, and younger ones on crustaceans and polychaetes. The above data on Pampus indicates that the food and feeding habits of these fishes depend upon such factors as size and growth rate of the fish, seasonal and local abundance of food organisms, and the type and length of symbiotic relationships. Commercial importance. Some of the species of all three genera of stromateid fishes are of commercial value. The fishes are oily and all species which are used commercially are said to be of delicate texture and fine flavor. Pampus is the most important genus commercially of the sub- order Stromateoidei. Species of Pamj)us are important food fishes along the coasts of India, China, and Japan, being much in demand and bringing a high price in most areas. Species of Stromateus support only local fisheries along the coasts of west Africa, Chile, Argentina, and Uruguay. The potential fishery value of the species along the coasts of the latter two countries is probably much greater than generally suspected (T. J. Hart, 1946). Three species of Peprilus, P. triacanthus, P. simillimus, and P. paru are of considerable commer- cial value, although the catch of each species is subject to large yearly fluctu- ations, and none of the species forms a major part of the fishery in their respective regions. P. triacanthus forms the largest fishery among the fishes of the genus. P. triacanthus, commonly known as the "butterfish" in fisheries literature, is mainly of importance along the coasts of the northeastern and middle Atlantic states. Its fishery has been discussed by Hilde- brand and Schroeder ( 1927 ) for the Chesa- peake Bay region and by Bigelow and Schroeder (1953) for the Gulf of Maine. This species is landed in recordable amounts from Maine to North Carolina, 238 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 no 1959 1960 Year Figure 42. Yearly commercial catch of Peprilus simillimus off the coast of California, 1955-64. Data compiled from Staff, Marine Resources Operations (1958), Biostatistical Sec- tion, Marine Resources Operations (1960a, 1960b, 1961, 1963, 1964, 1965), and Greenhood and Mackett (1965). and the largest landings are in Rhode Island, New Jersey, New York, Massa- chusetts, and Virginia (Power, 1962; Lyles, 1966). The species is not of importance south of North Carolina even though its range extends to southern Florida. In Massachusetts, it is available in fresh fish markets throughout most of the year and is usually sold at a weight of about one- fifth to one-half pound and a length of six to nine inches. It is usually captured by otter trawl, haul seines, or pound nets. In the Chesapeake Bay region, about equal quantities are taken in the Bay proper and in oceanic areas. The characteristics of small body cavity, relatively soft bones, and high muscle content combined with its habit of forming large concentrations in winter in deep water may indicate that this species is of considerable potential value for use in the manufacture of protein supplements and fishmeal. P. pant, the "harvestfish" of fisheries literature, is of important commercial value only on the Atlantic Coast of the United 5. 0 M J J A Months, 1955 -64 Figure 43. Cumulative monthly commercial catch of Peprilus simillimus off the coast of California, 1955-64. Source of data same as for Figure 42. States despite its much wider distribution in the Atlantic. This species is landed in largest amounts in the Chesapeake Bay region mainly off Virginia and North Caro- lina. It was not recorded in the commer- cial catch either north of Virginia or south of North Carolina in 1960 and 1964 ( Power, 1962; Lyles, 1966). P. paru is not sold in the Gulf of Mexico in any recordable num- bers although it occurs throughout the region in considerable abundance. Baugh- man (1941) has stated that this fish does not reach sufficient size in Texas to be of economic value. The size difference seems to be slight, however. Small local fisheries may exist for this fish southward to Brazil. In the Chesapeake Bay region almost twice the poundage is landed in the Bay proper compared to the adjacent oceanic areas (Lyles, 1966). This demonstrates some- what the inshore nature of this species compared to P. triacanthus which in the same region is caught in about equal amounts in the Bay proper and in oceanic- areas. P. simillimus, the "Pacific pompano" or just "pompano" of Pacific fisheries litera- Systematics and Biology of Peprilus • Horn 239 hire, has a wide distribution (British Columbia to southern Baja California), but is of commercial value only in the southern half of California from Monterey ■ to San Diego. The biggest landings by far are in San Pedro and Monterey, Cali- fornia (Power, 1962; Lyles, 1966). The annual and monthly catches of this species fluctuate widely (Figs. 42 and 43) accord- ing to data compiled from Staff, Marine Resources Operations (1958), Biostatistical Section, Marine Resources Operations (1960a, 1960b, 1961, 1963, 1964, and 1965), and Greenhood and Mackett ( 1965 ) . Largest monthly landings occur in the summer and early autumn. P. simittimus is a minor species in terms of the total catch; but with its rich flavor and flaky texture, it is considered a delicacy and brings a high price. It is sold entirely in fresh fish markets. The major portion of the catch is taken in lampara nets and purse seines, and a fair quantity is taken on hook-and-line by pier fishermen. P. burti of the Gulf of Mexico appar- ently does not attain the size of its close relative, P. triacanthus, on the Atlantic Coast. As already mentioned, P. paru in the Gulf of Mexico is reported to not reach the size of members of populations of the same species on the Atlantic Coast. These two situations may illustrate the trend observed by Lindsey (1966) that fishes tend to be larger in higher latitudes. P. burti is of little or no commercial im- portance in the Gulf of Mexico. According to Gunter (1945), few members of this species reach large size in Texas waters and they are never sold for food there. H. H. Hildebrand (1954), who caught large numbers of this species on the Texas shrimp grounds, reported that the largest specimen caught measured 173 mm (SL or FL?). Gunter (1945) caught one speci- men which measured 200 mm ( SL or FL? ) which may be nearly maximum size. P. triacanthus mav reach a length of 300 mm FL. No information is available on the economic status of P. ooatus in the northern Gulf of California, but, because of its ap- parently small size and restricted distri- bution, it is probably of little or no com- mercial importance. The extent of the commercial value of P. medius and P. snyderi is uncertain although both species do appear in fish markets along the coasts of Mexico and Central America. Each species probably constitutes a minor element of the fishery in these regions. EVOLUTIONARY RELATIONSHIPS AND ZOOGEOGRAPHY A number of attributes of the stromateid fishes and of the genus Peprilus in par- ticular invite study into the relationships and zoogeography of the group. The genus Peprilus is a member of an advanced family and has a widespread, continuous distri- bution along the coasts of the New World. Some of the species are widely distributed, two are more restricted, and all are con- fined to the continental shelf. None of the species cross oceanic barriers. Peprilus is largely allopatric with the other stroma- teid genera, and it is possible to correlate tectonic activity of recent epochs with the present distributional pattern of the mem- bers of the genus. Origin of the genus Peprilus and its position among stromateoid fishes. Certain evidence suggests a Tethyan distribution for early members of the family Stroma- teidae or its ancestors. These fishes, though low in the number of species, were prob- ably widely distributed in shallow coastal areas of the large Tethys Sea which in the early Tertiary combined the Indo-West Pacific, the Mediterranean, the tropical Atlantic, and the East Pacific faunas into one major unit (Ekman, 1953). The first tectonic activity in the western Tethys which led to eventual isolation of the Mediterranean probably occurred in the early Tertiary (Fell, 1967). The dis- ruption of the Tethys by emergence of 240 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 land areas began in the Oligocene and was well developed in the Miocene (Kummel, 1961). From that time on the Indo-West Pacific realm was isolated from the western end of the Tethys Sea, and the Mediter- ranean Sea was formed. The events described above may have divided and isolated segments of the an- cestral stromateid fishes. Pompus probably evolved in isolation from other stromateids in the eastern Tethyan region. The an- cestral stromateid stock in the western Tethyan region probably gave rise to Stromateus and Peprilus. Stromateus prob- ably emerged in the Mediterranean and West African region, and in at least partial isolation, Peprilus probably evolved in the American region. The divergence of the evolutionary lines leading to Stromateus and Peprilus had possibly begun before the division of the Tethys Sea since Fell (1967) has indicated that passage and ex- change of faunal elements through the western Tethyan region had become in- creasingly difficult before the Miocene. Stromateus occurs along the southern coasts of South America and is the only stromateid to occur on both sides of an ocean. However, its transoceanic dispersal is across the shortest possible route, the South Atlantic. Stromateus of southern South America is similar to Peprilus snyderi of the eastern tropical Pacific. The two genera may share a common ancestor in the western end of the Tethyan element. Stromateus of South America has retained its generic affinities with Stromateus of Africa, but approaches Peprilus in several characteristics including: a reduced num- ber of precaudal vertebrae; two epural elements in the caudal skeleton (three in the African Stromateus); and no pelvic- fins in the young (see Haedrich, 1967: 102). Its systematic position seems to be somewhat intermediate between Stroma- teus fuitola of Africa and Peprilus snyderi. It may be more closely related to Peprilus, but further study is needed. There was a direct connection between the Atlantic and Pacific Oceans before the formation of the Central American land bridge in the Pliocene, and there were probably few or no serious obstacles to widespread distributions of shallow, warm- water, marine organisms. The cooling of the climates which began in the middle and late Cenozoic and culminated in the Pleistocene glaciations (Kummel, 1961) apparently had not progressed by this time (lower Pliocene) so as to be a strongly limiting force to the distribution of tropical faunas. C. L. Hubbs (1960) has attributed in considerable part the low generic endemism in the Panamic fauna to the fact that this fauna is largely a part of the common Atlantic-Pacific tropical fauna of the New World. Hence, the stock of Peprilus was probably rather wide- ranging in an Atlantic-Pacific province of warm water. The distribution probably extended into what would be today temper- ate latitudes on the Atlantic and Pacific coasts. Few barriers existed to segregate populations, and speciation was probably at a low level. North and South America became con- nected by an isthmus in the Pliocene (Schuchert, 1935; Lloyd, 1963; Simpson, 1950), and since that time, the Atlantic and Pacific faunas have evolved separately. The genus Peprilus has been divided geographically for several million years, depending upon the time of complete emergence of the isthmus and the duration of the Pliocene and the Pleistocene. After this separation, the cooling of the climates, the alternating effects of the Pleistocene glacial and interglacial periods, the result- ant emergence and submergence of land barriers, and the establishment of persist- ent tropical-temperate conditions have pro- vided the barriers for isolation and the resulting speciation in the genus. Despite the basic morphological simi- larity among members of the genus, differ- ences are apparent in body proportions, meristic values, premaxillary teeth, in the length of the anterior lobes of the median Systematica and Biology of Peprilus • Horn 2 1 1 fins, and in the presence and size of Pampus and Peprilus, the latter especially, body pores. Two somewhat subtle species have lower meristic values than Stromateus, groups, or at least two separate trends, and Haedrich (1967) considers Pampus exist within the genus. One group consists to be the most advanced stromateid genus, of relatively elongate fishes with low The ancestral or early stromateid forms median fins; the species in this group are probably had an increased number of P. snyderi, P. simillimus, P. triacanthus, and vertebrae and median fin-rays relative to P. burti. The second group exemplifies the the centrolophid form from which they trend toward a deep body and falcate evolved. I therefore consider the high median fins and is composed of P. paru, number of vertebrae and other meristics of P. medius, and P. ovatus. Both trends or P. snyderi to represent the primitive con- species groups apparently were established dition in the genus, and the reduced before the closing of the Central American number to be the derived condition. P. seaway since each group is represented in para ( Fig. 24 ) is considered to be the most both the Atlantic and Pacific oceans, highly derived species, although no sug- Rosenblatt ( 1963 ) likewise has found that gestion is made that this species is the the two species groups of the tripterygiid most recently evolved. The divergence genus Enneanectes are each represented began early in the history of the group, in both oceans. before separation by the Central American Primitive and derived conditions. With- land bridge, and has continued to the out fossil material, it is difficult and some- present. P. paru, with its very deep body what conjectural to attempt to determine and extremely falcate dorsal and anal fins, which taxa within a group are primitive has diverged further from the presumed and which are derived. However, the situ- ancestral condition than any other member ation becomes more approachable when a of the genus. Pampus has in several re- larger group containing the one in question spects paralleled the evolutionary direction can be placed in a reasonable evolutionary taken by P. paru. perspective. Haedrich ( 1967 ) , largely on A fairly strong continuum of types exists the basis of osteological evidence, has between the primitive type, P. snyderi, and provided a concept of the evolutionary the derived type, P. paru. This range of trends in the suborder Stromateoidei. forms strongly supports the concept that Haedrich places the family Stromateidae this group of species constitutes a single at the zenith of stromateoid evolution and genus rather than the three (including considers the ancestor of the family to be Palometa and Poronotus) in common among the hard-spined members of the usage. Although P. paru differs in several family Centrolophidae, probably a fish sim- respects from P. snyderi and the other ilar to the genus Psenopsis. Members of elongate species, its affinities with these this genus are small (up to 200 mm SL), species are strong. The group fits well elongate, with small mouths, deciduous into the category genus as defined by scales, the dorsal fin with only slightly Mayr, Linsley, and Usinger ( 1953 ) . more rays than the anal fin, and a verte- There does exist in P. triacanthus and bral number of 10+15. P. burti the possible beginnings of another I consider P. snyderi (Fig. 16) to be the divergence or evolutionary direction. These most primitive species of Peprilus and most two species have a row of relatively large nearly like the ancestral form which gave pores of varying number and regularity rise to Stromateus and Peprilus, and prob- below the dorsal fin (Figs. 19 and 21-22), ably Pampus as well. Compared to Pse- and also have cusped rather than pointed nopsls, all stromateids have an increased or simple premaxillary teeth (Fig. 33b). number of vertebrae and median fin-rays. The function of either attribute is unknown 242 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 _P .JUL Figure 44. Dendrogram of probable species relationships in the genus Peprilus. The evolutionary steps which increase character-state codes for each character are shown as lines across the branches. The number of the character repre- sented is next to the horizontal line and corresponds to one of the list of characters in Tables 24 and 25. Explana- tion in text. and neither seems to represent as yet a major evolutionary shift. The presence of the row of pores has been used as a generic character; however, it does not appear to be specially innervated, but rather to be a slight elaboration of the subdermal canal system which is present in all the members of the genus. P. ovatus (Fig. 10) frequently has a series of medium-sized, irregularly-spaced pores along the dorsal surface of the body, a somewhat intermediate condition. Regard- ing the premaxillary teeth, two other members of the elongate species group, P. snyderi and P. simillimus, generally have pointed premaxillary teeth, but a particular specimen of either species may have a few teeth with two or three small cusps, an intermediate condition. In other aspects of morphology and ecology P. ttiacanthus and P. burti have strong affinities with the other members of the genus. As Cain ( 1956 ) has stated, the genus is not necessarily definable by one peculiar attribute, nor are its constituents mono- typic, equivalent, essentially merely sub- divisions of it, or themselves wholly dis- crete. Inclusion of the seven stromateid species within a single genus appears to be a natural and instructive classification consistent with the evolutionary, morpho- logical, and distributional characteristics of the members. Intra gene ric relationships. A dendro- gram of species relationships in the genus Peprilus is presented in Figure 44. The dendrogram fits the definition given by Mayr, Linsley, and Usinger (1953) and the definition of a cladogram given by Camin and Sokal ( 1965 ) . A time scale or a phylogeny is not inferred. Several methods have recently been published (Edwards and Cavalli-Sforza, 1964; Camin and Sokal, 1965; Wilson, 1965) concerning the reconstruction of cladistic relationships based on the char- acters of contemporaneous species. In re- constructing the dendrogram of Figure 44, I have used the quantitative method of Camin and Sokal in association with con- Systematics and Biology of Peprilus • Horn 243 Table 24. Characters and Character States Applied to Dendrogram of Species Relation- ships in Peprilus. Mean morphometries values in thousands of SL. See Table 25 and Figure 44. Character Character-State Codes 0 1 2 1. Eye diameter X 070-086 093-103 2, Interorbital width X 087-097 112 3. Pectoral fin length X 312-336 360-386 4. Preanal distance X 424^45 465-484 5. Body depth X 427-461 529-619 710 6. Caudal peduncle depth X 065-079 092 7. Mean dorsal ray number 46.5-46.7 42.8-44.8 8. Mean anal ray number 42.0-43.9 39.5-40.4 9. Precaudal vertebrae 14-15 13 10. Caudal vertebrae 20-22 16-19 11. Anterior lobes of median fins < 6 times length of shortest ray ^ 6 times length of shortest ray 12. Ratio of dorsal rays to anal ray s >1.0 = 1.0 13. Row of dorsal pores absent present 14. Premaxillary teeth simple cusped ventional methods. The distribution of the species and the probable position of the genus in the suborder have been important in establishing the species relationships. The reasons for using the Camin-Sokal method are mainly for systematization and quantification and for increased objectivity. The method, although not greatly different from conventional methods, allows the procedure of reconstruction to be more easily followed. The Camin-Sokal method requires several assumptions to be made: that characters be expressed in discrete states differing among at least some of the species; that characters can be arrayed in some logical order; that there is a knowledge of evo- lutionary trends so the characters can be arrayed in a presumed evolutionary se- quence from primitive to derived; that the ancestral state arose only once in the taxa, although derived character states may have arisen repeatedly in different branches; and, that evolution is irreversible for each character state. The most parsimonious cladogram is considered to be the most plausible one. Primitive character states are coded as zero, derived states positively or negatively as required (Tables 24 and 25). Table 25. Data Matrix of Character-State Codes for the Species of Peprilus. See Table 24 and Figure 44. Explanation in text. Character snyderi simillimus triacanthus burti medius ovatus paru 1. Eye diameter 0 0 0 1 0 1 1 2. Interorbital width 1 1 1 1 1 1 0 3. Pectoral fin length 0 0 0 0 1 1 1 4. Preanal distance 0 0 0 1 0 1 1 5. Body depth 0 0 0 1 1 1 2 6. Caudal peduncle depth 0 0 0 0 0 0 1 7. Mean dorsal ray number 0 1 1 1 0 1 1 8. Mean anal ray number 0 1 1 1 0 0 1 9. Precaudal vertebrae 0 1 1 1 1 1 1 10. Caudal vertebrae 0 1 1 1 0 1 1 11. Anterior lobes of median fins 0 0 0 0 1 0 1 12. Ratio of dorsal rays to anal rays 0 0 0 0 0 1 0 13. Row of dorsal pores 0 0 1 1 0 0 0 14. Premaxillary teeth 0 0 1 1 0 0 0 244 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 I began reconstruction of the cladogram using a total of 31 characters. Some of these could not be coded with any certainty or showed no differences among the spe- cies and thus provided no additional in- formation. As a result, 14 characters were applied to the cladogram. These are pre- sented in a data matrix with the appropri- ate character-state codes for each of the 14 characters (Table 25). The character states with the assigned character-state codes are listed in Table 24. The primitive and derived character states were determined by examining the members of the genus in the perspective of the entire suborder. Certain character states correspond to apparent trends exist- ing in the genus and represent conditions different from, or advanced relative to, the basal or primitive character states of some members of the genus or the pre- sumed ancestral types of the suborder. The character states considered to be de- rived include: an increased eye diameter; a greater interorbital width; a longer pec- toral fin; a greater preanal distance; in- creased body and caudal peduncle depths; and, lower numbers of dorsal and anal rays and of caudal and precaudal vertebrae (see Table 24 and the preceding discussion of primitive and derived conditions). The elongate lobes of the median fins, the ap- proximate one-to-one ratio of dorsal and anal rays, the presence of a row of dorsal pores, and the cusped premaxillary teeth (characters 11, 12, 13, and 14; Table 24), are considered to be derived character states and were easier to code than the above states because they are discrete and unique to certain species in the sense of being present or absent. The cladogram was reconstructed by the monothetic method described by Camin and Sokal. Evolutionary steps in- creasing character-state codes are shown graphically as short lines across the branches (Fig. 44). The number of the character represented by a short line is listed next to that line. It was necessary to test for parsimony and to make adjust- ments. The final cladogram requires 28 steps and is parsimonious in the context of information available. The Camin-Sokal method was of use in determining the relationships of a newly described species, P. ovatus. The clado- gram supports the assumption of a close relationship between P. ovatus and P. medius. P. snyderi has the assumed primitive condition of all but one of the characters used in the reconstruction of the clado- gram. The line leading to P. simillimus, P. triacanthus, and P. burti developed be- fore the formation of the Central American land bridge. P. burti and P. triacanthus are seen as recently diverged species; P. burti has attained some advanced character states including a deeper body, a larger eye, and fewer vertebrae. The paru-medius line was probably divided by formation of the Central American land bridge. The high number of changes in character states for P. paru is illustrated in the cladogram and indicates the derived condition. P. ovatus secondarily approaches P. paru in a number of aspects including the short, deep body, the large eye, and the reduced number of vertebrae. Parallelism and species pairs. As men- tioned above, P. triacanthus of the Atlantic and P. simillimus of the Pacific closely resemble one another. Both are moder- ately elongate fishes with low dorsal and anal fins (Figs. 13 and 21). These two species show the closest similarity of any two in the genus in the summed differences of certain morphometric and meristic characters (Table 26). Their respective distributions complement one another on opposite coasts. Both reach their northern distributional limits at about 48 °N, P. triacanthus as an occasional visitor on the southern coast of Newfoundland, and P. simillimus as an element in low abundance in coastal waters of southern British Co- lumbia. The southern limits of their dis- tributions are also similar. P. triacanthus Systematics and Biology oe Pi • Horn 245 ranges to about 27°N off the coast of south- subtropical distribution chiefly within the ern Florida, with likely infrequent occur- limits of the 20C surface isotherms. The fences in the Gulf of Mexico; one record opposite pattern of surface currents m the exists for the southeastern Gulf of Mexico Pacific relative to that in the Atlantic (UW 13407). P. simillimus occurs as far accounts for the disparity in the extent of south as about 24 30'N in Bahia Magda- the tropical-subtropical regions of the two lena, Baja California; one specimen is coasts and apparently in the distributions known from the Gulf of California, in of the two species. Whereas P. triacanthus Bahia Los Angeles (SIO 62-236). Each and P. simillimus probably acquired tem- speeies migrates seasonally in at least part perate adaptations early in the Pleistocene, of its geographic range. P. paru and P. medius seem to have had a P. triacanthus is of greatest abundance continuous warm-water existence. P. paru in the northern part of its range, North has probably extended its range northward Carolina to the Gulf of Maine, as is at- and southward into more temperate waters tested by its relative abundance in the only since the last glacial period. The commercial landings in this region. P. species has a continuous distribution simillimus is of most common occurrence around Florida which is indicative of in the southern part of its distribution, warm-water affinities. chiefly from San Francisco to Baja Cali- It is possible to postulate two possible fornia. It is caught commercially only mechanisms for the derivation of P. ovatus, from Monterey to San Diego in California, apparently restricted to the northern Gulf Catches of the species in Washington and of California. One mechanism is to con- British Columbia are small and sporadic, sider it to have differentiated from P. The major surface currents, the Gulf simillimus- stock off the outer coast of Cali- Stream in the Atlantic, and the California fornia and Baja California. C. L. Hubbs Current in the Pacific, probably influence (1960) stated that during glacial periods the regional distributions of P. triacanthus the outer coast fauna was displaced south- and P. simillimus, respectively. ward to yield Pleistocene relicts in the P. medius in the Pacific and P. paru in upper Gulf of California. Walker (1960) the Atlantic seem to parallel one another counted 22 fish species endemic to the morphologically and ecologically. Both upper Gulf and added that a striking ele- species are deep-bodied with falcate dorsal ment of the upper Gulf fauna is composed and anal fins, but P. paru has diverged of fishes common to the outer Baja Cali- more sharply to become a very deep- fornia and southern California coasts, but bodied fish with extremely falcate dorsal absent in the extreme southern Gulf. In and anal fins. Although little is known of most cases the degree of differentation is the habits of P. medius, the available evi- slight. Walker is hesitant to suggest how dence suggests that both are essentially the northern faunal element reached its shallow-water species that do not migrate present position. He offered the possibility seasonally. The geographic ranges are that there may have been a movement complementary. P. medius is distributed through an oceanic connection which ex- from the southern Gulf of California to isted in the region of the La Paz plain about 5°S near Paita, Peru, in the eastern during the late Pleistocene. If movement tropical-subtropical Pacific and is largely did occur through such a passage, only a confined within the limits of the 20° C slight general cooling of the waters of the surface isotherms. P. paru ranges from just Gulf would have been necessary to allow north of Chesapeake Bay to throughout the the entrance of cold-water forms from the Gulf of Mexico and the Caribbean region outer coast. Because of a temperature to southern Brazil and also has a tropical- front which exists in the cape region, a 246 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 much more extensive cooling would seem to be required to allow movement around the cape of Baja California according to Walker. However, P. ovottts is closer to P. medius morphologically, and I consider the former species to have been derived from P. medius stock even though the mechanism may be less obvious. Derivation from P. simillimus would involve an increase in the number of anal rays, in the number of caudal vertebrae, and an increase in body depth, eye size, and depth of caudal peduncle. Derivation from P. medius re- quires a decrease in the number of dorsal rays, in the number of caudal vertebrae, along with an increase in body depth and a decrease in the length of the anterior lobes of the dorsal and anal fins. The latter seems to be a more reasonable and co- ordinated change of character states and consistent with trends occurring in the genus. Much of the upper Gulf is quite shallow with wide areas of shelf; the sediment is mostly sand with transition to mud in the delta area of the Colorado River (Walker, 1960). This type of habitat seems suitable for members of this genus, particularly ones closely related to P. medius, which seems to be a shallow-water species. Several factors may interact to prevent the southward movement of P. ovatus or the northward dispersal of P. medius into the upper Gulf: almost the entire Gulf coast of Baja California south of the area men- tioned above is rocky, and few stromateid fishes have been collected along this coast; the shelf is narrow in the central Gulf region; temperature and salinity fronts exist in the southern Gulf; two islands with rocky shores are present in the central Gulf with a considerable extent of deep water south of them; increased wind convection and the more extreme ranges of sea sur- face-temperatures occur in the upper Gulf (Roden, 1958); and, a region of cold water exists around Isla Angel de la Guarda and adjacent islands in the upper central Gulf (Walker, 1960). P. triacanthus and P. burti are the most closely related species in the genus. Their similarity has resulted in different taxo- nomic interpretations (Caldwell, 1961; Collette, 1963). P. burti occupies the now zoogeographically familiar "temperate pocket" of the Gulf of Mexico which in past epochs was continuous with the At- lantic Ocean of the United States because of the submergence of peninsular Florida. Both species are absent as residents from the southern part of Florida, the popu- lations of both forming a disjunct distri- bution. C. L. Hubbs (1960) considers the Gulf of Mexico to be geographically and zoo- geographically analogous to the Gulf of California in that both contain Pleistocene relicts. Peninsular Florida has been var- iously submergent and emergent since the Miocene (Cooke, 1945). During the glacial periods of the Pleistocene when the sea level receded, Florida was a peninsula, and conversely, during the high stand of the sea of interglacial periods Florida was in large parts submerged (MacNeil, 1950). Thus, there have been opportunities for junction and disjunction of closely related populations. According to Hedgpeth (1953), the occurrence of nontropical spe- cies on both sides of Florida, but not in the southern part of the peninsula, appar- ently was as characteristic of Pleistocene periods when Florida was a peninsula as it is today. Hedgpeth characterized the Gulf of Mexico as a transitional zone between tropical and temperate regions with en- vironmental conditions favoring invasions from one region or the other at some time during the year. Isothermal configurations presented by Hedgpeth, and studies on the coastal faunas of Florida, show that the southern part of Florida is more tropical than the upper regions. According to Briggs (1958), the shorefish fauna of Florida, particularly the southern element, Systematics and Biology of Pi , rili s • Horn 2 I i has much in common with the fauna of the West Indies, South America, and Bermuda. The degree of differentiation varies among species which have disjunct popu- lations in the Atlantic and in the Gulf of Mexico. Some have diverged to the species level, others only slightly, and in most cases the decision as to rank is largely arbitrary. The situation has been described for cer- tain species or has been discussed by several authors including Ginsburg ( 1952 ) , Briggs (1958), R. J. Miller (1959), Cald- well (1961), Collette (1963), and Gutherz (1966) for fishes, and by Hedgpeth (1953) for crabs and molluscs. The burti-triacanthus situation is com- plex and requires special mention. Caldwell ( 1961 ) considered the Atlantic and Gulf populations to be a single polymorphic species, P. triacanthus. He recognized three distinct populations of the species: 1) a deep-bodied form in the Gulf of Mexico (equivalent to P. burti); 2) a shallow- bodied form in coastal waters of the Atlantic north of Cape Hatteras and in waters usually deeper than 22 m to the south of Cape Hatteras; and 3) a some- what intermediate form in shallow water of usually less than 22 m in depth south of Cape Hatteras. The Gulf population and the widely-distributed Atlantic population (2 above) are very closely related, but I have found the two forms to be distinct and, in combination with evidence given below, have reinstated P. burti as a valid species in the Gulf of Mexico. The mor- phological differences are in: caudal verte- brae number (17 in P. burti, 18-19 in P. triacanthus); body depth (Fig. 2); eye diameter (Fig. 3); and, the apparent lack of spots on the body in P. burti (Fig. 19) which are frequently present in P. tria- canthus (Figs. 21 and 22). P. burti does not reach the size of P. triacanthus, occu- pies a less extensive depth range than P. triacanthus (Table 22), and appears to be less migratory between inshore and off- shore waters than P. triacanthus. A main problem is to explain the exis- tence of the shaUow-watei population in the Atlantic (3 above), which presumably results in two distinct po] as of the same species in the Atlantic south of Cape Hatteras. Caldwell (1961) gave two rea- sons for recognizing two such populations: 1) Mature individuals have been taken from both populations, and there is thus apparently no significant spawning mi- gration; and 2) collections during each season of the year and in almost every month have produced specimens larger than 100 mm SL from most depths. In addition, I have found that most of the deep-water specimens in the Atlantic have been taken largely over a mud or silt bottom in waters up to 420 m depth, while1 the shallow- water form has been mainly collected over a sand bottom and nearer the coast (Table 22). The shallow-water form is deeper-bodied, usually has 17 or 18 (sometimes 19) caudal vertebrae, ap- parently has no spots on the body, and agrees closely with the description of P. burti (Fig. 19). The deep-water form (Fig. 22) is even more shallow-bodied than more northern members of P. tria- canthus (Fig. 21), usually has 19 (some- times 18) caudal vertebrae, and frequently has numerous spots on the body. There is likely some interbreeding between the two populations as evidenced by the over- lap in vertebral number; however, the interbreeding is apparently not of the level that would result in the merging of the populations. No external sexual dimor- phism or unusual sex ratios have been found in either population. While it may be possible to explain the development and maintenance of two dis- tinct populations in the same region by considering some type of hybridization, by presuming a strict segregation or diver- gence as to bottom type, or perhaps by assuming differential growth rates, I tend to favor as an explanatory device the speciation phenomenon of character dis- placement with the shallow- water popu- lation in the Atlantic being a disjunct 248 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 element of the Gulf of Mexico species, P. breeding population in the Atlantic; and, burti. in a first post-isolation contact have Character displacement may be defined morphologically and ecologically displaced (Brown and Wilson, 1956) as the situation the southern members of P. triacanthns to in which, when two species of animals the extent that the latter are further off- overlap geographically, the differences be- shore, in deeper water, and largely over a tween them are accentuated in the zone of mud-silt substrate. Kohn and Orians ( 1962 ) sympatry and weakened outside this zone, have stated that all cases of character The characters involved may be morpho- displacement should involve displacement logical, ecological, behavioral, or physio- of ecological characters since the two logical, the situation probably resulting species tend to exploit different niches most commonly from the first post-isolation where they occur together, contact of two newly evolved cognate That these fishes have migrated or can species. Upon meeting, the two popu- migrate around peninsular Florida or at lations interact through genetic reinforce- least are present on both sides of Florida ment of species barriers, or ecological is indicated by at least two pieces of evi- displacement, or both, in such a way as to dence: 1) one record of a fish agreeing diverge further from one another where in all characters with P. triacanthus (espe- they occur together. Plasticity of the par- cially the southern deep-water form) is ticular species may result in short-term known from the southeastern Gulf of ecological differences, say in food habits, in Mexico, northeast of the Dry Tortugas situations where species may infrequently (UW 13407); and 2) as has been de- come into contact. Whereas, character dis- scribed, individuals which agree quite placement should be considered as de- closely to the description of P. burti have veloping over a relatively longer period of been taken in the Atlantic off Florida time and involving genetic segregation as (USNM 156149, for example), to morphology, ecology, and other char- It is indeed difficult to advocate char- acteristics, acter displacement as occurring in the Invoking character displacement requires marine environment, especially since so that the probable history and vagility of many factors of the interaction cannot be the two species be considered. From a observed or monitored. Although the above knowledge of the history of the genus explanation is as yet only an hypothesis, discussed previously, it seems probable the phenomenon of character displacement that the Atlantic and Gulf populations seems with present evidence to be a rea- have existed since the beginning of the sonable mechanism for explaining the exis- Pleistocene. P. burti and P. triacanthus or tence of two distinct populations in a given their precursor populations were probably region. continuously distributed and panmictic Faunal regions. Much has been written during an early interglacial period. A sue- on the extent and delineation of faunal ceeding glacial period would have divided provinces of coastal marine organisms, the populations and allowed differentiation oft(,n thp ions ftre ded or con. and presumably isolating mechanisms to , . °, ,. . , .. , -ii rr-.i i . < ii • • tracted to suit the discussion or a particular develop, hither during a following inter- r . , „ , glacial period when contact would have SrouP oi organisms and to allow the been facilitated, or because of an increased §reatest number of generalizations. Cer- ability to migrate around Florida at some- tain of the faunal boundaries, including time in the recent past, members of P. ichthyological ones, remain speculative. In burti have become distributed in shallow the discussion below, no attempt is made water in sufficient numbers to establish a to alter the size of the provinces since the Systematics and Biology of Peprilus ' Horn 249 imprecise distributional knowledge of the species does not allow it. Atlantic Coast. P. paru is distributed throughout the tropical-subtropical region of the western Atlantic and occupies a number of faunal subregions. In ranging from just north of Chesapeake Bay south- ward throughout the Gulf of Mexico and the Caribbean to southern Brazil, this spe- cies is largely confined within the limits of the 20°C surface isotherms (Sverdrup, Johnson, and Fleming, 1942), a boundary often cited for warm-water species. Warm currents flowing away from the equatorial regions result in a relatively wide expanse of tropical and subtropical conditions in the western Atlantic. The southern limit of P. paru seems to concur with the faunal boundary and transition zone which begins at about 23° S in the region of Cabo Frio, Brazil. The partial boundary and transitional area are apparently created by the mixing of the waters of the cold, northward-flowing Falkland Current with those of the warm, southward-flowing Brazil Current, and have been discussed by several South American workers (Balech, 1954; Lopez, 1963; Stuardo, 1964; Vannucci, 1964). P. paru ranges southward as far as Buenos Aires, but is less common south of Rio de Janeiro in an area which is largely transi- tional between the tropical and south temperate regions (Balech, 1954). P. paru, in the northern part of its range, occupies a somewhat transitional region be- tween subtropical and warm temperate zones. P. burti and P. triacanthus are fishes of more temperate waters than P. paru. P. triacanthus ranges from the Gulf of Maine to just south of Cape Kennedy, Florida, and thus occupies the so-called Trans- atlantic Province (Hedgpeth, 1953) which some authors divide into the Virginian, Cape Cod to Cape Hatteras, and the Caro- linian, Cape Hatteras to southern Florida. The latter subregion is often considered to be a transitional or mixed zone. Cape Cod is often listed as the southern boundary of the Boreal Province but as Ekman (1953) has indicated, the fluctuating seasonal conditions of temperature in this area make it difficult to define a sharp boundary at Cape Cod. P. triacanthus is found season- ally north or south of Cape Cod. This illustrates that water conditions must be of primary importance in zoogeography rather than mere coastal configurations. P. burti, found throughout the Gulf of Mexico, is a disjunct warm temperate ele- ment closely related to P. triacanthus. P. burti occupies the so-called "temperate pocket" of the Gulf of Mexico. Pacific Coast. Until recently, two spe- cies of Peprilus, P. snyderi and P. medius, were known only from the Gulf of Panama. I have found that both species occupy al- most the whole of the tropical-subtropical region of the eastern Pacific. This region extends from the cape region of Baja Cali- fornia to about 6°S off northern Peru near Point Aguja (Ekman, 1953). Ekman in- cludes in this warm-water region the Gulf of California and the Galapagos Islands and regards the area from 3° to 6°S as possibly a transitional zone containing a subtropical fauna. Rosenblatt (1967) con- siders the southern limit of the tropical region to be at Cabo Blanco, Peru, about 4°S. P. snyderi occurs as far north as the upper Gulf of California, and Punta San Juanico (about 26° N) on the lower outer coast of Baja California, conforming to C. L. Hubbs' (1960) statement that tropical elements persist much farther north in the Gulf than on the outer coast. This species is not known southward beyond the Gulf of Panama, but it is a rarely-collected species and probably will be found as far south as the Gulf of Guayaquil, Ecuador (about 3°S). P. medius ranges from the vicinity of Paita, Peru (about 5°S), northward to Bahia Topolobampo (about 25°N) in the southern Gulf of California, thus occupy- ing almost all of the warm-water region. 250 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 One uncertain record exists for the Gala- pagos Islands. P. simillimus occupies essentially the warm temperate region of the eastern Pacific, but ranges into so-called tran- sitional areas northward and southward. This species occurs from Bahia Magda- lena (about 25CN) on the outer coast of Baja California to the southern coast of British Columbia (about 48°N). This species occurs mainly in the Californian Province which extends to about 48 °N (Schenck and Keen, 1936). P. simillimus is most abundant in an area corresponding closely to the warm temperate San Diegan division of C. L. Hubbs (1960). It is less abundant in the cool temperate division (Montereyan of Hubbs), and even less abundant in the somewhat transitional region between 42°N and 4"8°N of Schenck and Keen (1936). In general, the species of Peprilus oc- cupy broadly the major faunal regions of the continental shelf. Especially in the warm-water region of the eastern Pacific, the species cross the faunal subregions which have been established mainly from the study of generally small, more re- stricted fishes of rocky shores such as clinids, gobiesocids, tripterygiids, and chaenopsids (C. Hubbs, 1952; Briggs, 1955; Springer, 1958; Rosenblatt, 1959; and Stephens, 1963). Degree of divergence in allopatric and sympatric sets of species. Sympatry over an extensive area occurs only in three situations among the seven species of Peprilus. The group is somewhat reminis- cent of the superspecies concept, super- species being defined (Mayr, 1963) as a monophyletic group of entirely or essen- tially allopatric species that are morpho- logically too distinct to be included in a single species and, thus, that reproductive isolation between them can be assumed. The genus has a rather narrow ecological diversity. Therefore, it is important to know which species are sympatric and how Table 26. Values of Coefficient of Differ- ence (CD.) for Three Pairs of Species of Pep- rilus. Character =0 to ( Morphometries in thousandths of SL ) ti H O -5 £ ~ JgJt ■•— 5. to Head length 0.27 0.08 0.31 Snout length 0.12 0.08 0.58 Eye diameter 0.46 0.46 0.56 Upper jaw length 0.22 0.56 0.05 Interorbital width 0.28 0.22 1.01 Pectoral fin length 0.25 0.40 1.06 Predorsal distance I 0.21 0.23 0.81 Predorsal distance II 0.18 0.26 0.63 Preanal distance 0.31 0.10 0.53 Body depth 1.11 0.04 2.34 Caudal peduncle depth 0.34 0.43 1.45 Dorsal rays 0.22 0.15 0.43 Anal rays 0.22 0.28 0.03 Pectoral rays 0.44 0.20 0.62 Gill rakers 0.26 0.05 1.35 Vertebrae 2.39 1.72 2.50 Sum of CD. 7.28 5.26 14.26 Sum of CD.* 4.89 3.54 11.76 * Excluding vertebrae. niche separation is effected in sympatric situations. Coefficients of difference (CD.) for morphometric and meristic characters were calculated for three pairs of species. The individual and summed CD. values for each of the either allopatric or sympatric pairs are given in Table 26. The value is obtained by dividing the difference in the means of a character of two populations by the sum of their standard deviations (Mayr, Linsley, and Usinger, 1953: Chap- ter 7). A value of CD. which these authors consider to be the conventional level for designating subspecies is 1.28 or above. The method is used here not for subspecific distinctions, but to show that generally in the genus Peprilus the diverse sets of spe- cies are sympatric and the similar ones allopatric. P. snyderi and P. medius are sympatric Systematica and Biology of Peprilus • Horn 251 throughout much of the warm-water region within the genus toward a she p body of the eastern Pacific. They have been (Table 21; Figs. 2-6). It is a shallow- taken in the same hauls and are members water, inshore species while P. triacanthus of different species groups. P. snyderi is and P. burti seasonally migrate between the more rarely collected of the two spe- inshore and offshore waters and are more cies, and although the data are meager, it regularly in deeper water, seems to be a more vagile fish and one The premaxillary teeth of P. triacanthus more frequently in deep water than P. and P. burti each have three small cusps medius. Collections of the latter species (Fig. 33b) while those of P. paru and the are mostly from shallow coastal areas and rest of the species are generally simple and bays. pointed ( Fig. 33a ) . The teeth of the lower P. snyderi and P. simillimus, two quite jaw of all the species of Peprilus are multi- similar species (Table 21), are sympatric cusped but the cusp pattern in P. burti in a small region on the outer coast of Baja and P. triacanthus is slightly different. The California; however, there is only one significance of these differences in the record of P. snyderi from this region. I teeth is as yet unknown; however, they believe that the two species very rarely may indicate a slight difference or segre- come in contact, especially during the gation in food habits between say, P. paru breeding season. P. snyderi probably and P. triacanthus. Mouth size (Fig. 4) breeds only farther south along the Mexi- and, apparently, food habits are similar can coast. P. simillimus replaces the tropi- among the species. cal P. snyderi northward in temperate The above comparisons show that sym- waters. patry involves the more diverse sets of One record exists for P. snyderi in the species, whereas the sets or pairs of species upper Gulf of California where it would which are very similar tend to parallel one come in contact with the considerably dis- another in different oceans or to replace similar P. ovatus. P. snyderi is apparently one another in adjacent faunal regions with quite rare in this region, also. perhaps a small area of sympatry or none P. snyderi differs noticeably from P. at all. Niche separation is to a considerable medius and P. ovatus, but less so from P. degree effected spatially, i.e., one species simillimus (Figs. 2-6; Table 21). The in shallow water and a second ranging over differences seem to reflect the acquisitions a more extensive area of the shelf and associated with the trend toward a deep, generally in deeper water. There may be short body in P. ovatus and P. medius, and some shifts in food habits among sympatric include a larger eye, deeper body, slightly forms or some type of resource subdivision, longer pectoral fin, and deeper caudal although no differences in diet among peduncle. These differences may be cor- species have been discovered. This ap- related with an existence in a shallow, parent similarity in food habits further essentially inshore habitat. The longer suggests that spatial arrangement and pectoral fin, shorter body, and deeper ecological displacement are important in caudal peduncle seem suited for short- niche separation. distance, nonmigratory swimming, and the The nature of speciation in the genus larger eye an adaptation for living con- Peprilus. Most of the differentiation and tinuously in highly productive, but less speciation in the genus seems to have transparent inshore waters. taken place since the emergence of the In the Atlantic P. paru is sympatric with Central American land bridge in the Plio- P. burti and with P. triacanthus, except in cene. Before the emergence, the tropical the northern portion of the range of the faunas of the eastern Pacific-western latter. P. paru best exemplifies the trend Atlantic were apparently continuous and 252 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 homogeneous. Rates of evolution in the group are difficult to estimate since there is substantial difference of opinion as to the time in the Pliocene when the land bridge was completely formed. Briggs (1967), in following Simpson's (1950) statement that the Isthmus of Panama made its last emergence in the latest Plio- cene or earliest Pleistocene, estimates an age of one and one-half million years for the Central American barrier. Schuchert (1935) and Durham and Allison (1960) consider the time of emergence as the early Pliocene (or perhaps the late Miocene according to a reference of the latter authors). With the length of Pliocene as approximately ten million years (Kummel, 1961), a discrepancy of perhaps six to eight million years exists for the time of isolation of the Atlantic and Pacific faunas, depending upon whether one considers the early Pliocene or the late Pliocene as the date of closing of the seaway. Ira Rubinoff (personal communication) favors a period of three to four million years as the length of time since the effective separation of the Atlantic and Pacific shore faunas. In general, the isolation of the two faunas has resulted in divergence to the species level, which may indicate the later closure of the seaway since the average age of a species might cautiously be con- sidered to range from 100,000 to a few million years (Rensch, 1959). Rosenblatt ( 1967 ) finds that the shorefish faunas of the two sides of the Americas exhibit pro- found similarities on the familial, sub- familial, and generic levels, but that few species are common to both coasts. Ac- cording to Briggs (1967), only about 12 of the approximately 1000 fish species along the tropical American coasts are still identi- cal, making the Central American barrier 99 per cent effective. Members of the genus Peprilus conform to this pattern in that the genus is found on both coasts, but none of the species occur on both sides. The members of the genus have con- siderably higher vagility and a greater ability for genetic interchange over a broad geographic region than do a number of other coastal fishes, particularly those of rocky shores. The fishes of this genus possess considerable ability not only for active dispersal but for passive disperal as well via the planktonic stages of eggs and larvae. They are of medium size (smaller than 300 mm in length) and have a life span of several years (at least three or four years), characteristics which should contribute to the maintenance of genetic interchange and a certain homogeneity within a particular species. In contrast, the controlling factors in the distribution of fishes of rocky littoral areas are more restrictive and localizing in nature. Rosen- blatt (1963) states that fishes of rocky shores and coral areas are usually small, short-lived (perhaps one or two years), often territorial, and either have a short pelagic larval period or are even viviparous as is characteristic of a number of families of rocky littoral zones. He reasons that these factors curtail gene flow, frequently change the composition of the gene pool, and lead to differences in gene frequency either randomly or in response to local conditions. Such a population structure is favorable for the fragmentation of a species and the evolution of new species. The genus Peprilus and similar genera with wide-ranging coastal species are usually much less speciose than genera of rocky shores, and usually traverse the zoogeo- graphic subregions occupied by fishes of rocky littoral areas. ACKNOWLEDGMENTS I am deeply grateful to my adviser at Harvard University, Giles W. Mead, who throughout this work has provided both patient and enthusiastic support to my efforts. He has read and carefully criti- cized the manuscript in all its stages. Richard L. Haedrich of the Woods Hole Oceanographic Institution has been gen- Systematics and Biology of Peprilus • Horn 253 erous with information on stromateoid fishes, has given invaluable assistance of various types, and has examined the manu- script. I am very appreciative of his help and encouragement. H. B. Fell and E. O. Wilson have read and criticized the manuscript and I extend my thanks to them. A number of people have given their full cooperation either in sending specimens from their respective institutions or in directly providing working space, facilities, and access to collections. I wish to sin- cerely thank the following: Christine Karrer, Institut fur Spezielle Zoologie und Zoologisches Museum, Berlin; J. D. Mc- Phail, University of British Columbia, Vancouver; James E. Bohlke, Academy of Natural Sciences of Philadelphia; W. I. Follett, California Academy of Sciences, San Francisco; Loren P. Woods and Pearl Sonoda, Field Museum of Natural History, Chicago; David K. Caldwell and Robert J. Lavenberg, Los Angeles County Mu- seum; Richard H. Rosenblatt, Scripps Institution of Oceanography, La Jolla; Warren C. Freihofer, Division of System- atic Biology, Stanford University; Leslie W. Knapp, Smithsonian Oceanographic Sorting Center, Washington, D. C; Boyd W. Walker and John E. Bleck, University of California, Los Angeles; Gerald W. Wadley, University of Washington, Seattle; Frederick H. Berry and George C. Miller, Bureau of Commercial Fisheries Tropical Atlantic Biological Laboratory, Miami; Jack W. Gehringer and Elmer J. Gutherz, Bureau of Commercial Fisheries Biological Laboratory, Brunswick, Georgia; Gordon Gunter and C. E. Dawson, Gulf Coast Re- search Laboratory, Ocean Springs, Missis- sippi; Frank Hoff and Martin A. Moe, Jr., Florida Board of Conservation Marine Laboratory, St. Petersburg; C. Richard Robins and Eva-Maria Kiene, Institute of Marine Sciences, University of Miami; Royal D. Suttkus, Tulane University, New Orleans; Bruce B. Collette, Edgar N. Gramblin, and Ernest A. Lachner, United States National Museum, Washington, D. C; Donald E. Wohlschlag, Jniversity of Texas Marine Science Institute at Port Aransas. Personnel at three government labo- ratories have supplied or allowed access to distributional and ecological data on four species of stromateid fishes. I wish to express my appreciation to Elbert II. Ahlstrom, Bureau of Commercial Fisheries Fishery-Oceanography Center, La Jolla, California; to Harvey R. Bullis, Jr., Bureau of Commercial Fisheries Exploratory Fish- ing and Gear Research Base, Pascagoula, Mississippi; and to Robert L. Edwards and Marvin D. Grosslein, Bureau of Com- mercial Fisheries Biological Laboratory, Woods Hole, Massachusetts. Dr. Ahlstrom and Dr. Grosslein have also provided and allowed the use of base maps and station plans of their respective regions of study. This service is gratefully acknowledged. I have benefited from discussions and correspondence with Frederick H. Berry, Tropical Atlantic Biological Laboratory, Miami. John E. Fitch, California State Fisheries Laboratory, Terminal Island, and Heater Heyamoto, Bureau of Commercial Fisheries Exploratory Fishing and Gear Research Base, Seattle, have given valu- able information during the course of this work. Thomas E. Bowman, United States National Museum, kindly identified and provided information on the parasitic iso- pods from the gill chamber of Peprilus parti. Andrew Konnerth, Woods Hole Ocean- ographic Institution, has provided assist- ance with the X-ray equipment at that institution. The staff of the Museum of Comparative Zoology, especially Myvanwy M. Dick, has been most helpful in various ways. Ronald C. Baird, a fellow graduate stu- dent, has assisted with the computer programming. Sharon L. Horn has made the drawings, drafted most of the figures, and typed the final manuscript. She has contributed very 254 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 substantially to the completion of this study, and I am most grateful for her ideas, assistance, patience, and encouragement. Financial support during the course of this study has come largely from a National Defense Education Act Title IV Fellowship in evolutionary and population biology awarded through Harvard University. Ad- ditional funds for summer research and travel to various institutions have been gratefully received from the Committee on Evolutionary Biology, Department of Biology, Harvard University, NSF Grant No. GB3167. A National Science Founda- tion Summer Predoctoral Award provided support for study at the Duke University Marine Laboratory during the summer of 1966. SUMMARY A revision is presented of the genus Peprilus, one of the three genera of the family Stromateidae. The nominal genera Poronotus and Palometa are placed in the synonymy of Peprilus. Seven species are recognized in the genus. P. ovatus is de- scribed as a new species and is apparently restricted to the northern Gulf of Cali- fornia. P. merlins and P. palometa are synonyms, and the former is the valid name. P. alepidotus is treated as a synonym of P. paru. Accounts of each species con- sist of a synonymy, diagnosis, description, distribution, the geographic variation, and the ontogenetic change. The genus Peprilus occurs in tropical and temperate waters along the coasts of North, Central, and northern South America. Four species, P. medius, P. ovatus, P. simillimus, and P. snyderi, are distributed along the Pacific Coast, and three, P. triacanthus, P. burti, and P. paru, along the Atlantic Coast. Several aspects of functional morphology are considered. The vertebral column, skull, and pectoral fins appear to ossify earlier than the caudal skeleton and median fins, a sequence inteqoreted as being correlated with an early planktonic life followed by an independent nektonic ex- istence. Vertebral number is relatively constant within a species and is considered to be of possible selective value in main- taining a certain body form. The absence of pelvic fins, the long pectoral fins which are used extensively for propulsion in adult fishes, and the compressed body may all be correlated with the continuous swim- ming habit of these fishes, especially those larger than 100 mm SL. An hypothesis is presented that the swimbladder is of hydrostatic advantage to juvenile fishes which hover under jellyfish medusae and that it becomes nonfunctional in larger fishes which swim continuously. The scales are highly deciduous, and the skin is underlain by an extensive canal system, the function of which is unknown. The ali- mentary canal is composed of a small mouth with nipping teeth, a toothed, muscular pharyngeal sac, a U-shaped stomach, numerous pyloric caeca, and a long intestine. The food is shredded in the pharyngeal sac, and the great absorptive area of the caeca and intestine probably allows for maximum utilization of jellyfish and other food items. Considerations of life history and ecology are generally of four species - P. tria- canthus, P. burti, P. paru, and P. simillimus. Spawning occurs in the pelagic surface layers at varying distances from shore. The eggs and larvae are planktonic, the latter becoming capable of independent loco- motion at a size of about 10 to 15 mm SL. The species occur in a wide range of salinity and variously inhabit all depths over the continental shelf and generally over a sand or mud bottom. The genus is essentially tropical and warm temperate, with only two species, P. triacanthus and P. simillimus, reaching cooler waters. Seasonal movements appear to be most pronounced in P. triacanthus, the species occurring most abundantly in temperate regions. Fishes smaller than 100 mm SL associate with jellyfish medusae of several Systematics and Biology of Peprilus • Horn 255 genera. This association is apparently im- portant during the early critical growth phases of the fishes. Peprilus is a low- level carnivore; Jellyfish medusae seem to be an important element in the diet, especially of juveniles. Other food items include a variety of small crustaceans, polyehaete worms, and small fishes. Mem- bers of the genus are evidently significant forage fishes for a number of larger fishes, some of which are of great commercial importance. The economically important species of Peprilus are generally taken commercially in a region much smaller than the total range of the species, and this seems to reflect the pattern of migra- tion and center of abundance of the par- ticular species. Disruption of the Tethys Sea in the Miocene apparently facilitated the segre- gation of the early members of the family Stromateidae and led to the evolution of the three extant and essentially allopatric genera. The formation of the Central American land bridge in the Pliocene, the emergence and submergence of land areas associated with the Pleistocene glacial and interglacial periods, and the prevailing current systems all appear to have been important in producing the current level of differentiation and speeiation in the genus. The elongate P. snyderi is considered to be the most primitive type and the deep- bodied P. paru the most highly derived form in the genus. The Camin-Sokal method for deducing relationships of con- temporaneous species is used to reconstruct a dendrogram of species relationships. Two somewhat subtle species groups are recog- nized in the genus, and each group is represented on both sides of the Central American isthmus. Character displacement is invoked as a possible mechanism to explain the existence of two apparently distinct populations of P. triacanthus in the Atlantic off the southeastern coast of the United States. The distribution of the species of Pepri- lus appears to correspond generally to the major faunal provinces of the Atlantic Coast and the Pacific Coast of the Americas. The species generally traverse the zoogeo- graphic subdivisions established from the study of small fishes inhabiting shores. Sympatry involves the more di- verse species, and the similar or closely related species tend to parallel one another in different oceans or displace one another latitudinally along a continuous coastline. Niche separation seems to be produced largely by spatial arrangement and eco- logical displacement. LITERATURE CITED Ahlstrom, E. H. 1959. Vertical distribution of pelagic fish eggs and larvae off California and Baja California. U. S. Fish Wildl. Serv. Fish. Bull., 60(161): 107-146. Al-Hussaini, A. H. 1947. The feeding habits and the morphology of the alimentary tract of some teleosts living in the neighborhood of the Marine Biological Station, Ghardaqa, Bed Sea. Publ. Mar. Biol. Sta., Ghardaqa, No. 5: 4-61. Ayres, W. O. 1860. On new fishes of the Cali- fornia coast. Proc. Calif. Acad. Natur. Sci., ser. 1., 2: 81-86. Bailey, B. M., and W. A. Gosline. 1955. Vari- ation of systematic significance of vertebral counts in the American fishes of the family Percidae. Misc. Publ. Mus. Zool. Univ. Mich., 93: 1-44. Balech, E. 1954. Division zoogeografica del litoral sudamericano. Bev. Biol. Mar. Val- paraiso, 4: 184-195. Barlow, G. W. 1961. Causes and significance of morphological variation in fishes. Syst. Zool., 10(3): 105-117. Barnard, K. H. 1948. Further notes on South African marine fishes. Ann. S. Afr. Mus., 36: 341-406. Barrington, E. J. W. 1957. The alimentary canal and digestion. In Brown, Margaret E. (ed.), The Physiology of Fishes. Vol. I. New York: Academic Press, pp. 109-162. Bartlett, M. S. 1949. Fitting a straight line when both variables are subject to error. Biometrics, 5: 207-212. Batts, B. S. 1960. Further occurrence of the California pompano, Palometa simillima (Ayres), in Puget Sound, Washington. Copeia, 1960(2): 146-147. Baughman, J. L. 1941. Scombriformes, new, rare or little known in Texas waters, with 256 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 notes on their natural history or distribution. Trans. Tex. Acad. Sci., 24: 14-26. Bean, T. H. 1880. Check-list of duplicates of North American fishes distributed by the Smithsonian Institution in behalf of the United States National Museum, 1877-1880. Proc. U. S. Nat. Mus., 3: 75-116. Bean, T. H., and H. G. Dresel. 1884. A cata- logue of fishes received from the Public Museum of die Institute of Jamaica, with descriptions of Pristipoma approximans and Tylosurus euryops, two new species. Proc. U. S. Nat. Mus., 7(10): 151-170. Berry, F. H., and H. C. Perkins. 1966. Survey of pelagic fishes of the California Current area. U. S. Fish Wildl. Serv. Fish. Bull., 65 (3): 625-682. Besednov, L. N. 1960. Some data on the ichthyo- fauna of the Pacific Ocean driftwood. Tr. Inst. Okeanol. Akad. Nauk. SSSB, 41: 192- 197. (In Bussian). BlGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U. S. Fish Wildl. Serv. Fish. Bull., 53(74): 1-557. BlOS TATISTICAL SECTION, MARINE BESOURCES Operations. 1960a. The marine fish catch of California for the years 1957 and 1958. Calif. Dept. Fish Game Fish Bull., No. 108: 1-74. . 1960b. The marine fish catch of Cali- fornia for the year 1959. Calif. Dept. Fish Game Fish Bulk, No. Ill: 1-44. . 1961. The marine fish catch of Cali- fornia for the year 1960. Calif. Dept. Fish Game Fish Bulk, No. 117: 1-45. . 1963. The marine fish catch of Cali- fornia for the year 1961. Calif. Dept. Fish Game Fish Bulk, No. 121: 1-47. . 1964. The marine fish catch for 1962. Calif. Dept. Fish Game Fish Bulk, No. 125: 1-45. -. 1965. The California marine fish catch distribution. Bull. Fla. State Mus. Biol. Sci., 2(8): 223-318. 1967. Belationship of the tropical shelf for 1963. Calif. Dept. Fish Game Fish Bulk, No. 129: 1-45. Bloch, M. E., and J. G. Schneider. 1801. Systema ichthyologiae iconibus ex illustratum. Post obitum auctoris opus inchoatum absolvit, correxit, interpolavit Jo. Gottlob Schneider, Saxo. Berolini, 584 pp. Bonde, N. 1966. The fishes of the Mo-Clay Formation (Lower Eocene). Medd. Dansk Geol. Foren., 16: 198-202. Breder, C. M., Jr. 1926. The locomotion of fishes. Zoologica (N.Y.), 4(5): 159-297. Bricgs, J. C. 1955. A monograph of the cling- fishes (Order Xcnopterygii ) . Stanford Ich- thyol. Bull., 6: 1-244. . 1958. A list of Florida fishes and their regions. Studies Tropical Oceanogr. Miami, 5: 569-578. Brown, W. L., Jr., and E. O. Wilson. 1956. Character displacement. Syst. Zook, 5(2): 49-64. Buhler, H. 1930. Die Verdauungsorgane der Stromateidae ( Teleost. ) . Zeitschr. Morphok Okolog. Tiere, 19: 59-115. Cain, A. J. 1956. The genus in evolutionary taxonomy. Syst. Zook, 5(3): 97-109. Caldwell, D. K. 1961. Populations of the butterfish, Pownotus triacanthus (Peck), with systematic comments. Bull. S. Calif. Acad. Sci., 60(1): 19-31. Camin, J. H., and B. B. Sokal. 1965. A method for deducing branching sequences in phylog- eny. Evolution, 19: 311-326. Cargo, D. G., and L. P. Schultz. 1966. Notes on the biology of the sea nettle, Chrysaora quinquecirrha, in Chesapeake Bay. Chesa- peake Sci., 7(2): 95-100. Clemens, W. A., and G. V. Wilry. 1946. Fishes of the Pacific Coast of Canada. Bull. Fish. Bes. Bd. Canada, 68: 1-368. Collette, B. B. 1963. The systematic status of the Gulf of Mexico butterfish, Poronotus burti (Fowler). Copeia, 1963(3): 582-583. Colton, J. B., Jr., and K. A. Honey. 1963. The eggs and larval stages of the butterfish, Poronotus triacanthus. Copeia, 1963(2): 447-450. Compton, H., and E. Bradley. 1963. Analysis of populations of sport and commercial fin- fish and of factors which affect these popu- lations in the coastal bays of Texas; survey of the fish found in Gulf Area 20 from 0-15 fathoms. Job No. 14, Project Beports Coastal Fisheries 1961-62. Tex. Game Fish Comm., 14 pp. Cooke, C. W. 1945. Geology of Florida. Bull. Fla. Geol. Surv., 29: 1-339. Copeland, B. J. 1965. Fauna of the Aransas Pass Inlet, Texas. I. Emigration as shown by tide trap collections. Publ. Inst. Mar. Sci. Univ. Tex., 10: 9-21. Cuvier, G. 1829. Le regne animal, distribue d'apres son organisation, pour servir de base a l'histoire naturelle des animaux et d'intro- duction a l'anatomie comparee. Edition II. Tome II, Beptiles, Batraciens, et Poissons. Paris, 406 pp. . 1836-1849. Le regne animal, distribue d'apres son organisation, pour servir de base a l'histoire naturelle des animaux, et d'intro- duction a l'anatomie comparee. Edition III. Poissons. Paris, 392 pp. Systematics and Biology of Peprilvs • Horn 257 Cuvier, G., and A. Valenciennes. 1833. Histoire naturelle des poissons. Tome IX. Paris, 512 pp. DeKay, J. E. 1842. Zoology of New York, or the New York fauna. Comprising detailed descriptions of all the animals hitherto ob- served within the state borders. Class V. Fishes. Natur. Hist. N. Y. Ceol. Surv., Part I, Zoology. Albany, 415 pp. Dunnington, E., and R. Mansueti. 1955. School of harvestfish feeds on sea walnuts. Md. Tidewater News, 12(5): 1, 4. Durham, J. W., and E. C. Allison. 1960. The geologic history of Baja California and its marine faunas. In Symposium: The Bio- geography of Baja California and Adjacent Seas. Part I, Geologic History. Syst. Zool., 9(2): 47-91. Edwards, A. W. F., and L. L. Cavalli-Sforza. 1964. Reconstruction of evolutionary trees. In Heywood, V. H., and J. McNeill (eds.), Phenetic and Phylogenetic Classification. Syst. Assoc. Publ. No. 6: 67-76. Ekman, S. 1953. Zoogeography of the Sea. London: Sidgwick and Jackson, 417 pp. Evermann, B. W., and M. C. Marsh. 1900. The fishes of Puerto Rico. Bull. U. S. Fish Comm, pt. 1, 20: 49-350. Fell, H. B. 1967. Cretaceous and Tertiary sur- face currents of the oceans. Oceanogr. Mar. Biol, Ann. Rev, 5: 317-341. Fordice, M. W. 1884. A review of the Ameri- can species of Stromateidae. Proc. Acad. Natur. Sci. Philadelphia, 1884: 311-317. Fowler, H. W. 1906. New, rare or little known scombroids, no. 3. Proc. Acad. Natur. Sci. Philadelphia, 58: 114-122. . 1916. Cold-blooded vertebrates from Costa Rica and the Canal Zone. Proc. Acad. Natur. Sci. Philadelphia, 68: 389-414. . 1933. Notes on Louisiana fishes. Proc. Biol. Soc. Washington, 46: 57-64. . 1942. A list of the fishes known from the coast of Brazil. Arquiv. Zool. Estad. Sao Paulo, 3(6): 115-184. -. 1944. Description of a new genus and a new species of American stromateid fishes. Notulae Naturae, No. 142: 1-4. Fritz, R. L. 1965. Autumn distribution of groundfish species in the Gulf of Maine and adjacent waters, 1955-1961. Serial Atlas Marine Environment, Amer. Geog. Soc, Folio No. 10: 1-3. Gero, D. R. 1952. The hydrodynamic aspects of fish propulsion. Amer. Mus. Novitates, No. 1601: 1-32. Gilrert, C. H, and E. C. Starks. 1904. The fishes of Panama Bay. Mem. Calif. Acad. Sci, 4: 1-304. Gilchrist, J. D. F. 1922. Note on the oesoph- ageal teeth of the Stromateidae. Ann. Mag. Natur. Hist, ser. 9, 9: 249-255. Gill, T. N. 1861. Catalogue of the fishes of the eastern coast of North America, from (Green- land to Georgia. Proc. Acad. Natur. Sci. Philadelphia, Suppl, 13: 1-63. . 1862. On the limits and arrangements of the family of scombrids. Proc. Acad. Natur. Sci. Philadelphia, 14: 125-127. Ginsrurg, I. 1952. Eight new fishes from the Gulf Coast of the United States with two new genera and notes on geographic distribution. J. Washington Acad. Sci, 42(3): 84-101. Goode, G. B. 1879. A preliminary catalogue of the fishes of the St. John's River and the east coast of Florida, with descriptions of a new genus and three species. Proc. U. S. Nat. Mus, 2: 108-121. Goode, G. B, and T. H. Bean. 1879. Catalogue of a collection of fishes sent from Pensacola, Florida, and vicinity, by Mr. Silas Stearns, with descriptions of six new species. Proc. U. S. Nat. Mus, 2: 121-156. Gooding, R. M, and J. J. Magnuson. 1967. Significance of a drifting object to pelagic fishes. Pacific Sci, 21(4): 486-497. Greenhood, E. C, and D. J. Mackett. 1965. The California marine fish catch for 1964. Calif. Dept. Fish Game Fish Bull, Xo. 132: 1-45. Grosslein, M. D. 1969. Groundfish survey pro- gram of BCF, Woods Hole. Commercial Fish. Rev, 31(8-9): 22-30. Guichenot, A. 1866. Notice sur un nouvelle espece de poissons appartenant au genre des rhombes du Museum de Paris. Mem. Soc. Sci. Natur. Cherbourg, 12: 243-247. Gunter, G. 1945. Studies on the marine fishes of Texas. Publ. Inst. Mar. Sci. Univ. Tex, 1(1): 1-190. Gunther, A. 1860. Catalogue of the Acanthopte- rygian Fishes in the Collection of the British Museum. Vol. II. London: Taylor and Francis, 548 pp. Gutherz, E. J. 1966. Revision of the flounder genus Ancylopsetta ( Heterosomata: Bothi- dae) with descriptions of two new species from the Antilles and the Caribbean Sea. Bull. Mar. Sci, 16(3): 445-479. Haedrich, R. L. 1967. The stromateoid fishes: systematics and a classification. Bull. Mus. Comp. Zool, 135(2): 31-139. Harris, J. E. 1937. The role of the fins in the equilibrium of the swimming fish: 2. The role of the pelvic fins. J. Exp. Biol, 15: 32-47. Hart, J. L. 1949. Increased abundance of an unusual British Columbia fish, the California 258 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 pompano. Can. Field-Natur., 63(3): 101— 102. Hart, T. J. 1946. Report on trawling surveys on the Patagonian continental shell. Dis- covery Rept., 23: 223-408. Hedgpeth, J. W. 1953. An introduction to the zoogeography of the northwestern Gulf of Mexico with reference to the invertebrate fauna. Publ. Inst. Mar. Sci. Univ. Tex., 3(1): 107-224. High, W. L. 1966. Recent captures of the Cali- fornia pompano (Palometa simillima) and the sandfish ( Trichodon trichodon ) in Puget Sound. Wash. Dept. Fish, Fish. Res. Pap 2(4): 53-54. Hildebrand, H. H. 1954. A study of the fauna of the brown shrimp (Penaeus aztecus Ives) grounds in the western Gulf of Mexico. Publ. Inst. Mar. Sci. Univ. Tex., 3(2): 234-366. Hildebrand, S. F. Manuscript: The marine fishes of Panama. (Partially completed revision of the earlier work by Meek and Hildebrand, 1925. Stromateidae section obtained from FMNH through I. Rubinoff and R. L. Haed- rich. ) Hildebrand, S. F., and Louella E. Cable. 1934. Reproduction and development of whitings or kingfishes, drums, spot, croaker, and weak- fishes or sea trouts, family Sciaenidae, of the Atlantic Coast of the United States. Bull. U. S. Bur. Fish., 48(16): 41-117. Hildebrand, S. F„ and W. C. Schroeder. 1927. Fishes of Chesapeake Bay. Bull. U. S. Bur. Fish., pt. 1, 43: 1-388. Hoar, W. S. 1937. The occurrence of Poronotus triacanthus in the Gulf of St. Lawrence. Copeia, 1937(4): 238. Hubbs, Carl L. 1922. Variation in the number of vertebrae and other meristic characters of fishes correlated with the temperature of the water during development. Amer. Natur., 56: 360-372. . 1926. The structural consequences of modifications of the developmental rate in fishes, considered in reference to certain problems of evolution. Amer. Natur., 60: 57-81. 1960. The marine vertebrates of the outer coast. In Symposium: The Biogeography of Baja California and Adjacent Seas. Part II, Marine Biotas. Syst. Zool., 9(3&4): 134- 147. Hubbs, Carl L., and Clark Hubbs. 1953. An improved graphical analysis and comparison of series of samples. Syst. Zool., 2(2): 49-56. Hubbs, Clark. 1952. A contribution to the classification of the blennioid fishes of the family Clinidae, with a partial revision of the Eastern Pacific forms. Stanford Ichthyol. Bull., 4(2): 41-165. Humphrey, G. 1797. Museum Calonnianum. Specification of the various articles which comprise the .... Museum of Natural History collected by M. de Calonne in France, etc. Part I. 84 pp. Isokawa, S., K. Kubota, T. Kosakai, I. Sato- mura, M. Tsubouchi, and A. Sera. 1965. Some contributions to study of esophageal sacs and teeth of fishes. J. Nilion Univ. Sch. Dent, 7(3): 103-111. Jordan, D. S. 1883. Notes on American fishes preserved in the museums at Berlin, London, Paris, and Copenhagen. Proc. Acad. Natur. Sci. Philadelphia, 1883: 281-293. . 1884. List of fishes collected at Key West, Florida, with notes and descriptions. Proc. U. S. Nat. Mus., 7(8): 103-150. Jordan, D. S., and C. H. Bollman. 1889. De- scriptions of new species of fishes collected at die Galapagos Islands and along the coast of the United States of Colombia, 1887- 1888. Proc. U. S. Nat. Mus., 12: 149-183. Jordan, D. S., and B. W. Evermann. 1896. The fishes of North and Middle America: a de- scriptive catalogue of the species of fish-like vertebrates found in the waters of Nordi America, north of the Isthmus of Panama. Part I. Bull. U. S. Nat. Mus., No. 47: 1- 1240. Reprint 1963, Smitiisonian Institution. . 1898. The fishes of North and Middle America: a descriptive catalogue of the spe- cies of fish-like vertebrates found in the waters of North America, north of the Isthmus of Panama. Part III. Bull. U. S. Nat. Mus., No. 47: 2183-3136. Reprint 1963, Smithsonian Institution. Jordan, D. S., B. W. Evermann, and H. W. Clark. 1930. Check list of the fishes and fish-like vertebrates of North and Middle America north of the northern boundary of Venezuela and Colombia. Appendix X, Rep. U. S. Coram. Fish., 1928. Reprint 1955, Washington, U. S. Govt. Printing Office, 670 pp. Jordan, D. S., and C. H. Gilbert. 1878. Notes on the fishes of Beaufort Harbor, North Caro- lina. Proc. U. S. Nat. Mus., 1: 365-388. . 1881. Notes on the fishes of the Pacific Coast of the United States. Proc. U. S. Nat. Mus., 4: 29-70. . 1882a. Synopsis of the fishes of North America. Bull. U. S. Nat. Mus., No. 16: 1-1018. . 1882b. Notes on a collection of fishes from Charleston, South Carolina, with de- scriptions of three new species. Proc. U. S. Nat. Mus., 5: 580-620. Systematics and Biology of Peprilus • Horn 259 Kennedy, D., and R. D. Milkman. 1956. Selec- tive light absorption by the lenses of lower vertebrates, and its influence on spectral sensitivity. Biol. Bull., 111(3): 375-386. Kohx, A. J., and G. H. Orians. 1962. Eco- logical data in the classification of closely related species. Syst. Zool., 11: 119-127. Kummel, B. 1961. History of die Earth: An Introduction to Historical Geology. San Francisco: W. H. Freeman and Co., 610 pp. Kuntz, A., and L. Radcliffe. 1918. Notes on die embryology and larval development of twelve teleostean fishes. Bull. U. S. Bur. Fish., 35: 87-134. Kuthalingam, M. D. K. 1963. Observations on the fishery and biology of the silver pomfret, Pampas argenteus (Euphrasen), from the Bav of Bengal. Indian J. Fish., 10(1): 59-74. Lacepede, B. G. 1800. Histoire Naturelle des Poissons. Volume II. 372 pp. LaMonte, F. R. 1958. Scales of the Atlantic species of Makaira. Bull. Amer. Mus. Natur. Hist., 114(5): 381-395. Leim, A. H., and W. B. Scott. 1966. Fishes of the Atlantic Coast of Canada. Fish. Res. Bd. Canada, Bull. No. 155: 1-485. Liem, Karel F. 1963. The comparative oste- ology and phylogeny of the Anabantoidei (Teleostei, Pisces). 111. Biol. Monogr., No. 30: 1-149. Lindsey, C. C. 1954. Temperature controlled meristic variation in the paradise fish, Macropodus opercularis (L.). Can. J. Zool., 30: 87-98. . 1966. Body sizes of poikilotherm verte- brates at different latitudes. Evolution, 20 (4): 456-465. Linnaeus, C. 1758. Systema Naturae. 10th Edition. Vol. I. 824 pp. . 1766. Systema Naturae. 12th Edition. Vol. I (1). 532 pp. Lloyd, J. J. 1963. Tectonic history of the south Central-American orogen. In Backbone of the Americas. Publ. Amer. Assoc. Petrol. Geol., Tulsa, Okla., pp. 88-100. Lopez, R. B. 1963. Peces marinos de la Republica Argentina. Evaluacion Recursos Natur. Ar- gentina, 2(3): 105-219. Lutkex, C. F. 1880. Spolia Atlantica. Bidrag til Kundskab om Formforandringer hos Fiske under deres vaext og Udvikling, sa=rligt hos nogle af Atlanterhavets H0js0fiske. Danske Vidensk. Selsk. Skr., (5) Nat. Math. Aid., 12(6): 409-613. Lyles, C. H. 1966. Fisherv statistics of the United States 1964. U. S. Fish Wildl. Serv. Statist. Dig., No. 58: 1-541. MacNeil, F. S. 1950. Pleistocene shorelines in Florida and Georgia. Prof. Papers U. S. Geol. Surv., 221-F: 91-107. Mansueti, R. 1963. Symbiotic behavior between small fishes and jellyfishes, with new data on that between die stromateid, Peprilus alepi- dotus, and the scyphomedusa, Chrysaora quinquecirrha. Copeia, 1963(1): 40-80. Marshall, N. B. 1966. The Life of Fishes. Cleveland: World Publishing Co., 402 pp. Mayr, E. 1963. Animal Species and Evolution. Cambridge, Mass.: Harvard Univ. Press, 797 pp. Mayr, E., E. G. Linsley, and R. L. Usinger. 1953. Methods and Principles of Systematic Zoology. New York: McGraw-Hill, 336 pp. McKenzie, R. A. 1939. Some marine fish and salp records. Proc. Nova Scotian Inst. Sci., 20: 13-20. Meek, S. E., and S. F. Hildebrand. 1925. The marine fishes of Panama. Part II. Publ. Field Mus. Natur. Hist. (Chicago), 15(226): 331- 707. Menon, M. A. S. 1942. The food and feeding habits of the shoal fishes of the Trivandrum coast and their general correlations with the plankton distribution. Thesis, Univ. Travan- core. Part II. (Unpublished). Miller, D., and R. R. Marak. 1959. The early larval stages of the red hake, Urophycis chuss. Copeia. 1959: 248-250. Miller, J. M. 1965. A trawl survey of die shallow gulf fishes near Port Aransas, Texas. Publ. Inst. Mar. Sci. Univ. Tex., 10: 80-107. Miller, R. J. 1959. A review of the seabasses of the genus Centropristes (Serranidae). Tulane Stud. Zool., 7(2): 35-68. Miner, R. W. 1936. Sea creatures of our Atlantic- shores. Nat. Geog. Mag., 70(2): 209-231. Mitchill, S. L. 1814. Report, in part, of the fishes of New York. No. 301. D. Carlisle, Broadway, New York, 28 pp, (Ed. by Theo- dore N. Gill, 1898). New York, 30 pp. . 1815. The fishes of New York, described and arranged. Trans. Lit. Phil. Soc. N. Y., 1: 355-492. Moore, J. E., and D. S. Gorsline. 1960. Physical and chemical data for bottom sediments South Atlantic Coast of the United States, M/V THEODORE N. GILL cruises 1-9. U. S. Fish Wildl. Serv. Spec. Sci. Rep. -Fish., No. 366: 1-84. Myers, G. S. 1958. Trends in the evolution of teleostean fishes. Stanford Ichthyol. Bull., 7(3): 27-30. Nath, P. R. 1966. Biology and seasonal distri- bution of the pelagic food fishes of Travan- core coast. Kerala Univ. Publ., India, 1-140. Nursall, J. R. 1958. The caudal fin as a hydro- foil. Evolution, 12: 116-120. 260 Bulletin Museum of Comparative Zoology, Vol. 140, No. 5 Olson, E. C. 1964. Morphological integration and the meaning of characters in classifi- cation systems. In Heywood, V. H., and J. McNeill (eds. ), Phenetic and Phylogenetic Classification. Syst. Assoc. Publ. No. 6: 123- 156. Parr, A. E. 1956. On the original variates of taxonomy and their regressions upon size in fishes. Bull. Amer. Mus. Natur. Hist., 110 (5): 369-398. Patterson, C. 1964. A review of Mesozoic acanthopterygian fishes, with special refer- ence to those of the English Chalk. Phil. Trans. Roy. Soc. London, ser. B, Biol. Sci., 247(739): 213-482. Pearson, J. C. 1932. Winter trawl fishery off the Virginia and North Carolina coasts. U. S. Bur. Fish. Invest. Rep. No. 10: 1-31. . 1941. The young of some marine fishes taken in lower Chesapeake Bay, Virginia, with special reference to the gray sea trout, Cijnoscion regalis (Bloch). U. S. Fish Wildl. Serv. Fish. Bull., 50(36): 79-102. Peck, W. D. 1804. Description of four remark- able fishes, taken near the Piscataqua River in New Hampshire. Mem. Amer. Acad. Arts Sci., 2(2): 46-57. Peters, W. 1869. Neue oder weniger bekannte Fische des Berliner zoologischen Museums. Monatsber. Konigl. Akad. Wiss. Berlin, 1869: 70.3-711. Power, E. A. 1962. Fishery statistics of the United States 1960. U. S. Fish Wildl. Serv. Statist. Dig., No. 53: 1-529. Quoy, J. R. C, and J. P. Gaimard. 1824. Voyage autour du monde . . . les corvettes . . . l'Uranie et la Physicienne pendant les annees 1817- 1820. 3. Zoologie. Poissons, Parts I and II. Paris, pp. 192-401. Regan, C. T. 1902. A revision of the fishes of the family Stromateidae. Ann. Mag. Natur. Hist, 10(56): 115-131. Rege, M. S. 1958. A study of the stromateid fishes of Bombay. Ph.D. Thesis, Univ. Bom- bay. Rege, M. S., and D. V. Bal. 1964. Some ob- servations on the food and feeding habits of the silver pomfret, Pampus argenteus (Eu- phrasen), in relation to the anatomy of its digestive system. J. Univ. Bombay, N. S., 31(5): 75-79. Reid, G. K. 1955. A summer study of the biology and ecology of East Bav, Texas. Part II. Tex. J. Sci., 7(4): 430^53. Rensch, B. 1959. Evolution Above the Species Level. New York: Columbia Univ. Press, 419 pp. Roden, G. I. 1958. Oceanographic and meteoro- logical aspects of the Gulf of California. Pacific Sci., 12: 21-45. Roedel, P. M. 1953. Common ocean fishes of the California coast. Calif. Dept. Fish Game Fish Bull., No. 91: 1-184. Rosenblatt, R. H. 1959. A revisionary study of the blennioid fish family Tripterygiidae. Ph.D. Thesis, Univ. Calif., Los Angeles, 376 pp. . 1963. Some aspects of speciation in marine shorefishes. In Harding, J. P., and N. Tebble (eds.), Speciation in the Sea. Syst. Assoc. Publ. No. 5: 171-180. 1967. The zoogeographic relationships of the marine shorefishes of tropical America. Stud. Trop. Oceanogr. Miami, 5: 579-587. Schaefer, R. H. 1967. Species composition, size, and seasonal abundance of fish in the surf waters of Long Island. N. Y. Fish Game J., 14(1): 1-46. Schenck, H. G., and M. A. Keen. 1936. Marine molluscan provinces of western North America. Proc. Amer. Phil. Soc. 76(6): 921-938. Schuchert, C. 1935. Historical Geology of the Antillean-Caribbean Region, or Lands Border- ing the Gulf of Mexico and the Caribbean Sea. New York: John Wiley and Sons, Inc., 811 pp. Schultz, R. L. 1962. Fisheries investigations in the Aransas-Copano Bay system; a survey and inventory of the vertebrate species present in Mesquite Bay and Cedar Bayou. Job No. A-2, Project Reports Coastal Fish- eries. Tex. Game Fish Coram., 15 pp. Simpson, G. G. 1950. History of the fauna of Latin America. Amer. Sci., 38(3): 361-389. Simpson, G. G., A. Roe, and R. C. Lewontin. 1960. Quantitative Zoology, rev. ed. New York: Harcourt, Brace and World, Inc., 440 pp. Sloane, Sir Hans. 1725. . . . Natural History of Jamaica . . . Vol. II, 499 pp. Sokal, R. R. 1965. Statistical methods in sys- tematics. Biol. Rev., 40(3): 337-391. Springer, V. G. 1958. Systematica and zoo- geography of the clinid fishes of the subtribe Labrisomini Hubbs. Publ. Inst. Mar. Sci. Univ. Tex., 5: 418^92. . 1961. Notes on and additions to the fish fauna of the Tampa Bay area in Florida. Copeia, 1961: 480-482. Staff, Marine Resources Operations. 1958. The marine fish catch of California for the years 1955 and 1956. Calif. Dept. Fish Game Fish Bull., No. 105: 1-104. Staff, South Pacific Fishery Investigations. 1956. Zooplankton volumes off the Pacific Coast, 1955. U. S. Fish Wildl. Serv. Spec. Sci. Rep. -Fish., No. 177: 1-31. Systematics and Biology of Pi-wrilus • Horn 261 Stephens, J. S., Jr. 1963. A revised classification of the blennioid fishes of the American family Chaenopsidae. Univ. Calif. Publ. Zool., 68: 1-133. Storer, D. H. 1839. Reports on the ichthyology and herpetology of Massachusetts. Mass. Zool. Bot. Surv. Fishes, 206 pp. Stuardo, J. 1964. Distribution de los moluscos marinos litorales en Latinoamerica. Bol. Inst. Biol. Mar., 7: 79-91. Suyehiro, Y. 1942. A study on the digestive system and feeding habits of fish. Japan. J. Zool., 10(1): 1-303. Sverdrup, H. U., M. W. Johnson, and R. H. Fleming. 1942. The Oceans, their Physics, Chemistry, and General Biology. Englewood Cliffs, N. J.: Prentice-Hall, 1087 pp. Taxing, A. V. 1952. Experimental study of meristic characters in fishes. Biol. Rev., 27: 169-193. Taylor, H. F. 1922. Deductions concerning the air bladder raid the specific gravity of fishes. Bull. U. S. Bur. Fish., 38(921): 121-126. Taylor, W. R. 1967. An enzyme method of clearing and staining small vertebrates. Proc. U. S. Nat. Mus., 122(3596): 1-17. Thrailkill, J. R. 1957. Zooplankton volumes off the Pacific Coast, 1956. U. S. Fish Wildl. Serv. Spec. Sci. Rep. - Fish., No. 232: 1-50. . 1959. Zooplankton volumes off the Pacific Coast, 1957. U. S. Fish Wildl. Serv. Spec. Sci. Rep. -Fish., No. 326: 1-57. . 1961. Zooplankton volumes off the Pacific Coast, 1958. U. S. Fish Wildl. Serv. Spec. Sci. Rep. -Fish., No. 374: 1-70. -. 1963. Zooplankton volumes off the Pacific- Coast, 1959. U. S. Fish Wildl. Serv. Spec. Sci. Rep. -Fish., No. 414: 1-77. Ulrey, A. B., and P. O. Greeley. 1928. A list of the marine fishes (Teleostei) of southern California with their distribution. Bull. S. Calif. Acad. Sci., part I, 27: 1-53. Vannucci, M. 1964. Zoogeografia marinha do Brazil. Bol. Inst. Biol. Mar., 7: 113-121. Vladykov, V. D. 1934. Environmental and taxonomic characters of fishes. Trans. Roy. Can. Inst., 20: 99-140. Walford, L. A. 1958. Living Resources of the Sea. New York: Ronald Press, 321 pp. Walker, B. W. 1960. The distribution and affinities of the marine fish fauna of the Gulf of California. In Symposium: The Bio- geography of Baja California and Adjacent Seas. Part II, Marine Biotas. Syst. Zool., 9 (3&4): 123-133. Walters, V. 1963. The trachipterid integument and an hypothesis on its hydrodynamic func- tion. Copeia, 1963(2): 260-270. Wilson, E. O. 1965. A consistency test for phylogenies based on contemporaneous spe- cies. Syst. Zool., 14(3): 214-220. (Received 4 December 1968.) ill latin o Museum of oology The Systematics and Zoogeography of the Unionidae (Mollusca: Bivalvia) of the Southern Atlantic Slope Region RICHARD I. JOHNSON HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS, U.S.A. VOLUME 140, NUMBER 6 27 NOVEMBER 1970 PUBLICATIONS ISSUED OR DISTRIBUTED BY THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Breviora 1952- Memoirs 1864-1938 Johnsonia, Department of Mollusks, 1941- Occasional Papers on Mollusks, 1945- Other Publications. Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine. Reprint, $6.50 cloth. Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of Insects. $9.00 cloth. Creighton, W. S., 1950. The Ants of North America. Reprint, S10.00 cloth. Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 12-15. (Price list on request. ) Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Teredinidae (Mollusca: Bivalvia). $8.00 cloth. Whittington, H. B., and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948. ( Complete sets only. ) Publications of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138, U. S. A. © The President and Fellows of Harvard College 1970. THE SYSTEMATICS AND ZOOGEOGRAPHY OF THE UNIONIDAE (MOLLUSCA: BIVALVIA) OF THE SOUTHERN ATLANTIC SLOPE REGION RICHARD I. JOHNSON CONTENTS Introduction 264 Acknowledgments 266 Part I. The zoogeography of the Unionacea of the Apalachieolan and Atlantic Slope regions 267 Chapter 1. The distribution of the Union- acea in the Apalachieolan region as evidence of a former confluence of the headwaters of the Alabama- Coosa, Apalachicola, and Savannah river systems 267 The Apalachieolan region defined 267 Analysis of the distribution of the species 268 Chapter 2. The distribution of the Union- idae in the Southern Atlantic Slope region as evidence of a former stream confluence of the headwaters of the Apalachicola and Savannah river sys- tems 273 The Atlantic Slope region defined and subdivided 273 Analysis of the distribution of the species 275 Chapter 3. The Unionacea of the Northern Atlantic Slope region, with a discus- sion of the probable origin of some of the Atlantic slope species . 278 Analysis of the distribution of the species 278 Chapter 4. The generic affinities of the Unionacea of the Apalachieolan region, Peninsular Florida, and Southern and Northern Atlantic Slope regions 281 Chapter 5. The relationship of the geo- morphology and topography of the Bull. Mus. Comp. Apalachieolan and Atlantic Slope re- gions to their unionid faunas .... 283 Evidence of stream capture .. . 284 The Piedmont Plateau . 285 The Coastal Plain _____ . 286 Chapter 6. Summary and Conclusions 291 Part II. A revision of the Unionidae from the St. Marys River, Florida, to the Po- tomac River, Maryland 293 Introduction 293 History of the taxonomists and the collectors of the Atlantic Slope Unionacea 293 The taxonomists 293 The collectors . _ 294 The primary systematic studies of Apala- chieolan and Atlantic Slope Union- acea 294 Classification of the Unionacea _ 295 Systematic Section 296 Key to the Unionidae found between the St. Marys River, Florida, and the Po- tomac River, Maryland 297 PLEUROBEMA - 299 Lexingtonia 300 Pleurobema (Lexingtonia) collina (Con- rad) - 300 Pleurobema (Lexingtonia) masoni (Con- rad) - 301 ELLIPTIC- 303 Canthyria 303 Elliptio (Canthyria) spinosa (Lea) . - 303 Elliptio s. s. 304 Elliptio (Elliptio) crassidens crassidens (Lamarck) 305 Zool., 140(6) : 263-450, November, 1970 263 264 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Elliptio (Elliptic) crassidens downiei (Lea) - 307 Elliptio (Elliptio) congaraea (Lea) ... 308 Elliptio (Elliptio) dariensis (Lea) 310 Elliptio (Elliptio) fraterna (Lea) 312 Elliptio (Elliptio) waccamawensis (Lea) 313 Elliptio (Elliptio) complanata (Light- foot) . 314 Elliptio (Elliptio) hopetonensis (Lea) 324 Elliptio ( Elliptio ) icterina ( Conrad ) .. 325 Elliptio (Elliptio) arctata (Conrad) .. 331 Elliptio (Elliptio) lanceolata (Lea) .. 333 Elliptio ( Elliptio ) shepardiana ( Lea ) 338 UNIOMERUS 339 Uniomerus tetralasmus (Say) 339 LASMIGONA 343 Platynaias 343 Lasmigona (Platynaias) subviridis (Con- rad) ..... 343 ALASMIDONTA 346 Prolasmidonta 346 Alasmidonta (Prolasmidonta) heterodon (Lea) . 347 Alasmidonta s. s. 348 Alasmidonta (Alasmidonta) undulata (Say) . 349 Alasmidonta (Alasmidonta) triangulata (Lea) . 351 Alasmidonta (Alasmidonta) arcula (Lea) 352 Decuramhis 353 Alasmidonta (Decuramhis) varicosa ( Lamarck ) 354 ANODONTA 356 Pyganodon 356 Anodonta (Pyganodon) cataracta cata- racta Say 356 Anodonta ( Pyganodon ) gibbosa Say .. 359 Anodonta (Pyganodon) implicata Say 360 Utterbackia 362 Anodonta (Utterbackia) imbecilis Say 362 Anodonta (Utterbackia) couperiana Lea 365 STROPHITUS 366 Strophitus undulatus (Say) 367 CARUNCULINA 369 Caruncidina pulla ( Conrad ) 370 VILLOSA 371 Villosa villosa (Wright) . 372 Villosa vibex (Conrad) 373 Villosa delumbis (Conrad) 375 Villosa constricta (Conrad) . 378 LIGUMIA 380 Ligumia nasuta (Say) 380 LAMPSILIS . 382 Lampsilis s. s. 382 Lampsilis (Lampsilis) cariosa (Say) .. 382 Lampsilis (Lampsilis) dolabracformis (Lea) 384 Lampsilis (Lampsilis) ovata (Say) .... 386 Lampsilis (Lampsilis) ochracea (Say) 388 Lampsilis (Lampsilis) radiata radiata (Gmelin) . 390 Lampsilis (Lampsilis) splendida (Lea) 393 Bibliography 395 Index to relevant taxa 399 ABSTRACT The Atlantic Slope region consists of those streams flowing into the Atlantic- Ocean from the Altamaha River system, Georgia, to the lower St. Lawrence River system, Canada, including rivers in New- foundland and Labrador. Forty species comprise the unionid fauna. Most of the species that are clearly of Interior Rasin origin are a northern group that migrated around the northern end of the Appala- chian Mountains before the Pleistocene; there is a southern group that entered the Atlantic Slope region through a confluence of the Apalachicola and Savannah river systems, also in pre-Pleistocene time; there is also an additional fauna originally of Interior Rasin origin, but whose ancestry is more remote. INTRODUCTION The Unionacea and Mutelacea (Parodiz and Ronetto, 1963), or freshwater mussels, are found throughout the world, but it is in the Mississippi River system that the Unionidae have especially radiated and achieved their greatest diversity. This, the Interior Basin, embraces 1,200,000 square miles. The rivers are very old, and flow over vast limestone beds. Here is found a variety of shell forms which is rivaled only in a few species found in the rivers of China. Within the Interior Basin, H. and A. van der Schalie (1950: 450) recognized the Ozark and Cumberland regions, each of which has an indigenous unionid fauna of its own, as well as the Interior Basin one. Continental North America is made up of the following additional regions. They are given essentially as defined by H. and A. van der Schalie (1950). The Pacific region is the area west of the Rocky Mountains. It has a very limited unionid fauna, which is clearly of Asiatic origin. Found there are several species of Ano- donta, Margaritifera margaritifera (Lin- naeus), and "Gonidea" angulata (Lea), which may belong to the Asiatic genus Solenaia Conrad. The West Gidf Coastal region consists of the streams flowing into the Gulf of Mexico from the eastern slope of Mexico north to, but not including, the Alabama- Coosa River system. Found there are several endemic genera of Unionidae. It has been tentatively suggested by both Simpson (1892: 406) and H. and A. van der Schalie (1950: 452) that this might be regarded as a subregion of the Interior Basin. The Apalachicolan (or West Florid ian) region was mentioned by H. and A. van der Schalie (1950: 450) and is now defined as consisting of those river systems flowing into the Gulf of Mexico from the Escambia to the Suwannee, and also including the St. Marys and the Satilla, which flow directly into the Atlantic Ocean. Peninsular Florida has representatives of only six genera. These consist of species derived mostly from the Apalachicolan and Atlantic Slope regions. This area will be discussed as a separate region in a sub- sequent report. The Atlantic Slope region includes the river systems flowing into the Atlantic- Ocean from the Altamaha, Georgia, to the lower St. Lawrence, Canada, as well as those of Newfoundland and Labrador. The Unionidae of this region are of Interior Basin origin, but on the Atlantic slope the Atlantic Slope Unionidae • Johnson 265 species are smaller and are neutral in color. Yet, the only two species of Unionidae with true spines. Elliptic- spinosa (Lea) and Pleurobema collina (Conrad), occur here. Because the Atlantic Slope region has distinct northern and southern assemblages of species, it has been divided here into the Southern Atlantic Slope region, which ex- tends from the Altamaha River system, Georgia, to the James River system, Vir- ginia, and the Northern Atlantic Slope region, which extends from the York River system, Virginia, to the lower St. Lawrence River system, Canada. Between the York River and the glacial drift border ( roughly along the Pennsylvania-New York bound- ary) are found all of the Unionacea that repopulated the Northern Atlantic Slope region north of the limit of glacial drift. The Unionacea, or freshwater mussels, offer two advantages for zoogeographic study: 1) As shown in the systematic por- tion of this paper, there are a reasonable number of species, most of which are clearly distinguishable, and whose generic affinities have been revealed rather clearlv by Ortmann (1911; 1912a). 2) They have a limited mode of distribution, being un- able to pass over land from one drainage system to another. Their ability to move between drainage systems is dependant on the mobility of fishes to which the glochidia attach themselves. Of the primary fresh- water fishes, Myers (1938: 343) stated, "They are inescapably confined to their own particular drainage systems and can migrate from one isolated stream basin to the next only through the slow physiographical change of the land itself (stream capture, etc. [base- leveling] )." For this reason the distribution of the Unionacea gives evidence of former stream confluences and of flooding in baseleveled coastal regions. Van der Schalie (1945) convincingly illustrated the use of Unionidae as a means of tracing major stream confluences. He also reviewed the old controversy over mechanical distribution, which implies that 266 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 unionid distribution is haphazard and for- tuitous, caused by aquatic birds to whose feet gravid females may occasionally be- come attached. This theory postulates that the shell will be carried to a different river system, and that there the glochidia will be released and will find a suitable host fish. No one has actually established that Unionidae have been successfully spread by birds in this manner, but even granting this possibility, there is no evidence that it is, or was, an important method of dis- tribution, since zoogeographic data fail to substantiate unionid distribution by any agent other than fish. In their discussion of the freshwater mol- lusks of the Apalachicolan region, Clench and Turner (1956: 103) support the idea that that fauna was distributed by me- chanical means, though they are vague as to what those means might have been. Exception is taken here to their theory of distribution, but only as far as the Unionacea are concerned. The Apalachicolan Unionacea were re- studied, and it is demonstrated here that their general distribution is not fortuitous. References are made to some of the Unionidae of the Alabama-Coosa River system to illustrate the origin of some of the Apalachicolan species and to illuminate the sequence of stream captures which affected the distribution of both Apalachi- colan and Southern Atlantic Slope Unioni- dae. In his classic study of the influence of the Alleghenian Divide on the distribution of mollusks and crayfish, Ortmann ( 1913a ) assumed that the Unionidae were dis- tributed by natural means. Roth by choice and because of the confusion in which he found the systematics of the southern species, his work was limited to the Atlantic side of the divide, primarily to the species of Unionacea found on the Northern Atlantic slope. In the present study, the distribution of the Apalachicolan and Atlantic Slope Unionacea is analyzed, and, with the ex- ception of Margaritifera mar^aritifera ( Linnaeus ) , Alasmidonta marginata ( Say ) , and Anodonta catorocta fragilis (Lamarck), which are beyond the scope of this paper, all of the Atlantic Slope Unionacea are included and their nomenclature is revised. ACKNOWLEDGMENTS In 1961, Dr. William J. Clench, my long- time mentor, received a grant from the National Science Foundation (G18922) to make a collection of the freshwater mollusks of the Atlantic Slope of Georgia. Upon completion of the expedition, he generously gave me the Unionidae for study. As the study progressed, it gradu- ally became apparent that the unionids of the entire Southern Atlantic slope had to be included if the work was to be of any significance. The opportunity for examining relevant types was provided by: Dr. Norman Tebble, Rritish Museum ( Natural History ) ; Dr. Gilbert Ranson, Paris Museum; Dr. G. Mermod, Geneva Museum; Dr. Arthur H. Clarke, Jr., National Museum of Can- ada; Dr. David H. Stansbery, Ohio State Museum; Dr. Juan Parodiz, Carnegie Mu- seum, Pittsburgh; Dr. R. Tucker Abbott, Academy of Natural Sciences of Philadel- phia; Dr. Henry van der Schalie, Museum of Zoology, University of Michigan. Drs. Harold Render, Joseph Rosewater, and Joseph Morrison, United States National Museum, kindly permitted me to study the collections under their care and to borrow all relevant types. The cost of photo- graphing the types was generously borne by the William F. Milton Fund, Harvard Uni- versity. Dr. Ruth Patrick, Academy of Natural Sciences of Philadelphia, made available the Unionidae collected for her by John N. Rates and Samuel L. H. Fuller; Mr. Rates allowed me to study the portion of this material under his care at the Museum of Zoology, University of Michigan. Dr. Patrick kindly deposited most of the Atlantic Slope Unionidae • Johnson 267 material collected by Mr. Fuller in the Museum of Comparative Zoology. Herbert D. Athearn, Cleveland, Ten- nessee, is gratefully acknowledged for the material which he has presented to the Museum of Comparative Zoology over the years, and which has proved useful in this study. Thanks are also extended to Dr. William II. Heard, University of Florida. Tallahassee, for specimens. Finally, thanks are offered to Drs. Ken- neth J. Boss, Raymond A. Paynter, Jr., and Ruth D. Turner, who critically read the manuscript, and to Drs. Arthur II. Clarke. Jr. and Joseph P. E. Morrison, who offered opinions on some matters of taxonomy. PART I. THE ZOOGEOGRAPHY OF THE UNIONACEA OF THE APALACHICOLAN AND ATLANTIC SLOPE REGIONS CHAPTER 1 The distribution of the Unionacea in the Apalachicolan region as evidence of a former confluence of the headwaters of the Alabama-Coosa, Apalachicola, and Savan- nah river systems. Figure 1 illustrates the propinquity of the headwaters of the Alabama-Coosa, Apalachicola, and Savannah river systems. Van der Schalie (1945) made it clear that there was once a connection between the Tennessee and Alabama river systems; this connection is of interest here since certain unionid species of the Alabama-Coosa River system are discussed in this paper. Matteson (1948a: 131) following, in part, Hayes and Campbell (1894), suggested that in the late Tertiary, the Chattahoochee River of the Apalachicola River system captured a tributary of the Etowah River, of the Alabama-Coosa River system, and that the Savannah River then captured one of the tributaries of the Chattahoochee River. This present study supports these geomorphological assertions, but in a dif- ferent order. The distribution of Apalachi- colan species in the Savannah River system seems to indicate that this system may first have been connected to, and later separated from, the Chattahoochee River before the latter was connected to the Alabama-Coosa River. These conditions would explain why certain species found in both the Alabama-Coosa and Apalachi- cola river systems are absent in the Savannah River system. To understand the origins of the Unioni- dae of the Southern Atlantic Slope, it was necessary to study the distribution of the superfamily Unionacea in the Alabama- Coosa and Apalachicola river systems. Further, the whole Apalachicolan region had to be considered, to determine if it could be established that unionid distri- bution there is not fortuitous. Tfie Apalachicolan region defined. This region is regarded here as consisting of the river systems flowing into the Gulf of Mexico, from the Escambia to the Su- wannee. Also included are the St. Marys and Satilla, although they flow into the Atlantic Ocean ( Plate 1 ) , because, as Table 1 shows, their modest unionid faunas con- sist entirely of species found in the Apa- lachicolan region, and further, because EUiptio crassidens crassidens and E. c. downiei occur in them, respectively, and are dominant. According to Cooke (1945: 273), in the early Pleistocene, during the Aftonian interglacial stage when the Brandywine terrace ( Citronelle Formation in the South- east) was formed, the sea level was 270 feet above the present level. If this were so, most of the area occupied by the present St. Marys and Satilla river systems, and virtually all of Peninsular Florida, were inundated, with the exception of 268 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Text-figure 1. Northern Georgia, illustrating the propinquity of the headwaters of the Etowah River (ER), of the Alabama-Coosa; of the Chattahoochee River (CR), of the Apalachicola; and of the Savannah (HSR) river systems. The topographic areas are the Appalachian Mountains (AM), Appalachian Valley (AV), and the Piedmont Plateau (PP). The Blue Ridge divide (BR) is indicated by large dots. The minor divides are indicated by smaller dots. (After Keith, 1925, pl. 29.) several small islands in the vicinity of present Polk County (Cooke, 1945: 273). x It is now generally agreed that this flood- ing took place in the Upper Miocene (Alt and Brooks, 1965: 408). MacNeil (1950: 98, 99) casts doubt on Cooke's evidence on the extent of Aftonian flooding, and on the basis of his identifi- cation of marine terraces concludes that the Citronelle Formation is of subaerial origin, and that there is no evidence to indicate that the sea ever transgressed it to an alti- tude of more than 150 feet during the Yarmouth interglacial stage. This flooding is now thought to have occurred during the Pliocene. MacNeil's detailed map (pl. 19) shows that even at 150 feet Peninsular Flor- 1 Orange (Ocala) Island, referred to by Clench and Turner (1950: 104), was a land mass that was separated from the continent by the Suwannee Strait dining the late Oligocene (Vanghan, 1910: 156), and its existence appears to have had no bearing on the present nnionid fauna. ida was reduced to a number of small is- lands where a unionid fauna might have had refugia. In any case, except for a small portion of the Satilla River which is above the area of maximum flooding, the remain- der of it, and all of the St. Marys River, is of more recent origin (see p. 289). It is assumed that these two rivers were mostly repopulated from the west. The ap- pearance of Elliptio c. cra.ssiden.s in the Pliocene of Florida (see note on p. 271) indicates that it has been present in this general area over a long period of time. Analysis of the distribution of the species. The Apalachicolan and Atlantic Slope regions consist of a number of independent river systems, some with quite different faunas. Table 1 shows all of the Unionacea known from the individual Apalachicolan river systems, systematically arranged. Some differences in species concepts and Atlantic Slope Unionidae • Johnson 269 Table 1. Distribution ok the Umonacea of the Apalachk :olan hegion and RELEVANT SPECIES OF THE AlABAMA-CooSA RlYKR SYMHM. 0) ua s g u g M 1) .S« to -gS 2 >> >> £ >• > "> « » £ <- a qj 3 V > 2 > t).S ja.s- WPS >H« U« X X X X X X c o .Sts ja >. o <« a h 73 g «5 H .St Offi c/jPG 1> c>- a°> . > c/3« (/J 1. 1/ o. 3. 4. 5. X 6. 7. X 8. 2/ 9. 10. X 11. 12. 13. 3/ 14. 15. X 16. 17. 18. 19. 20. 21. X 22. 4/ 23. 24. X 25. X 26. X 27. 28. 29. 30. X 31. X 32. 5/ 33. X 34. X 35. X 36. 37. 38. X 39. X 40. 41. 42. X 43. 44. 45. Margaritifera hembeli ( Conrad ) Fusconaia succissa ( Lea ) Fusconaia escambia Clench & Turner Quincuncina infurcata (Conrad) Quincucina burkei Walker Amblema boykiniana (Lea) Amblema neisleri (Lea) Amblema perplicata* (Conrad) Plcurobema strodcanum (Wright) Plcurobema pyriforme (Lea) Elliptic) crassidens crassidens (Lamarck) EUiptio crassidens downiei (Lea) Elliptic fraterna (Lea) EUiptio complanata (Lightfoot) EUiptio icterina (Conrad) EUiptio arctata (Conrad) EUiptio lanceolata ( Lea ) EUiptio jayensis (Lea) EUiptio nigclla (Lea)* EUiptio chipolaensis (Walker) EUiptio sloatiana (Lea) Uniomerus tetralasmus (Say) Alasmidonta triangulata (Lea) Alasm idonta wrightiana ( Walker ) * Anodonta grandis Say Anodonta cataracta cataracta Say Anodonta imbecilis Say Anodonta peggyae Johnson* Anodonta couperiana Lea Anodonta suborbiculata Say* Anodontoides radiatus ( Conrad ) * Strophitus snbvexus (Conrad)* Obovaria rotulata (Wright)* Carunculina parva (Barnes) Villosa vibex (Conrad) Villosa lienosa (Conrad) Villosa choctawensis Athearn* Villosa villosa (Wright) Lampsilis excavatus (Lea) Lampsilis claibornensis (Lea) Lampsilis haddletoni Athearn* Lampsilis binominatus Simpson* Lampsilis anodontoides (Lea) (6) Lampsilis australis Simpson Lampsilis jonesi van der Schalie* X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X x X X X X XXX X X X X X 270 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Table 1. (Continued) c - c S O a, — ii -S.S; 0) rt t/i -" 1) w« >-« ■8 0 is u5 0 tb o IS a > C w 5 * O > Ok ^ -^ W5 « ■i.a\ region, or are widely distributed throughout and beyond it. wps ;>*ps o ^ to -c >> 5 > ^ >• O 1H "S.S5 OK a §& S "> a >- II :/3pS i> fe- ns ■» . > :/)PS C/2PS .=: P-.W •I s E •5 i U. tetralasmus C. pa rv a L. claibornensis V. vibcx V. licnosa L. anodontoides A. perplicata L. excavatus M. hembeli A. suborbicaJata X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X Nine species of the Alabama-Coosa River system (Table 1, B) appear to have reached the Apalachicola River through a former stream confluence (Text-fig. 1, see p. 285), since they are mostly missing from intervening systems. With the probable exception of Anodonta grandis, A. imbe- cilis, and Glebula rotundata, it is possible that the remaining six species were once endemic to the Alabama-Coosa River system. Matteson (1948a: 131) suggests that EUiptio c. crassidens and E. dilatata Rafinesquc (figured by Ortmann, 1919: 95, pi. 8, fig. 2 ) reached the Interior Basin from the Alabama-Coosa River system. If this can be shown, then EUiptio s. s. is a genus originally from the coastal regions. Four species of Unionidae appear to have reached the Apalachicolan region from a former confluence of the Apalach- icola and Savannah river systems (Table 1, C). They will be discussed with the Southern Atlantic Slope fauna. Three ad- ditional species entered the Savannah Table 1, B. Unionidae, derived from the west, found in both the Alabama- Coosa and Apalachicola river systems, supporting the assumption of a former confluence between them. E. crassidens A. boijkiniana A. grandis A. imbecilis E. arctata A. radiatus S. subvextis G. rotundata M. penicillatus U u 5 ^ •a a X X X X X X X X X .3- a M S.fe WPS X X X X X X >HpS X o §. J3.2 UPS X .a ►. X X X X X X X X X n> a a v a ■*-* O u Ops a >> a "> a in c/3PS p £ ID 53 u s.« ■j; a 13 & a3 ■a >- . > ■M > F c tops c/5pS PhW X X X X X1 i E. crassidens has been found as a fossil in the Pliocene deposits of St. Petersburg, Hillsboro Co., Florida. It renamed EUiptio pachyodon Pilsbry (see p. 306). 272 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Table 1, C. Unionidae, derived from the east or west, found in the Alabama- COOSA AND APALACHICOLA RIVER SYSTEMS, AND WHICH GIVE EVIDENCE OF A FORMER CONFLUENCE WITH THE SAVANNAH RlVER SYSTEM, EXCLUDING THE POSTULATION OF TWO MINOR STREAM CONFLUENCES. a o g -a U E t> 0> S JJ o 5 « «> V *-. u s r* x wa "— Vi C vi it X v. S" X e >■ >t £ >. •~ *~. p* >» e >■ £■>• •«•-> H j — ^ C i-i 1" rt wi * 1- O WD c * 2 £ ■2" ;■« M 1m — . a a .2 aj ? o "3 ^ 3 a» > ^ — n *— c s? ^ a > o > ? > ■A > £ > . > a > r c > > X ^PS A S s s - 2 n o E tn & 5 E lH x. - oa > >. •— cfl 2 >> >> A >. ^c >. C '^- C>- >. 5 '- g" a x C t_ "S J- « hi £ t £ j> ~ it "P 13 l> 2 j< ~ i< •" t- o > — > 8 > 3 > . > — > ij. u «•- oj^; WPS ><« urt 0) 4} c3 ■" ~ — o — v 4-> •ii U5 VI 5 x •— -y: o *& a> w 3 tn Cfl x> >• >* A >, ^ ►, — s. £•>• >. ^Z v. .2 s « w O W5 u en - W5 5 » m w B i- o > o u ji .5 0<.S > *> 5.5 was >H(C urt £ § C 5 -c £ C 1* > > £ to to > 0 CD C C5 tn >> > > £ est 0) > £ est U > ca IH cd > £ s > £ EQ £ CO >. VI U > CD u 0) c 3 C3 £ o I h CD PC CD g 5 cj 9) fa 0 PC en c cd M PC 0J > •II r. -73 pi p ^« est O p O CZ) o u 01 £ C3 U cd 2 £ PC J3 H^ PC PC PC c X X X X X X X X X X X 1. Pleurobema collina (Conrad) 2. Pleurobema masoni (Conrad) 3. Elliptio spinosa ( Lea ) 4. Elliptic congaraea ( Lea ) 5. Elliptio waccamawensis ( Lea ) 6. Elliptio dariensis (Lea) 7. Elliptio fraterna ( Lea ) 8. Elliptio complanata (Lightfoot) 9. Elliptio hopetonensis (Lea) 10. Elliptio icterina (Conrad) 11. Elliptio aretata (Conrad ) 12. Elliptio lanceolata (Lea) 13. Elliptio shepardiana (Lea) 14. Uniomerus tetralasmus (Say) 15. Lasmigona subviridis (Conrad) 16. Alasmidonta heterodon (Lea) 17. Alasmidonta undulata (Say) 18. Alasmidonta triangulata (Lea) 19. Alasmidonta arcula (Lea) 20. Alasmidonta varicosa (Lamarck) 21. Anodonta cataracta eataraeta Say 22. Anodonta implicata Say 23. Anodonta gibbosa Say 24. Anodonta imbecilis Say 25. Anodonta couperiana Lea 26. Strophitus undulatus (Say) 27. Caruneulina pulla (Conrad) 28. Villosa vibex ( Conrad ) 29. Villosa delumbis ( Conrad ) 30. Villosa constricta ( Conrad ) 31. higumia nasuta (Say) 32. LampsUis eariosa (Say) 33. LampsUis dolabraeformis (Lea) 34. Lampsilis ovata (Say) 35. Lampsilis ochracea (Say) 36. Lampsilis rudiata radiata (Cmelin) 37. Lampsilis splendida (Lea) X X X X X X X X X X X X X X X X X X X X X X X X xxxxxxxxxxxxxxxx X xxxxxxxx XX X xxxxxx xxxxxxxxx X xxxxxxxxxxxx X XX XXX xxxx X XX XX XXX xxxxx XX X XXX X X XXXXX XXX X X X X X X X XXXXXX X X XXX X X X xxxxxxxxx XXX XX xxxxxxxxx xxxxxxxxx X XXXX X X X XX X XXXX X XXX X X XXXX X X X XXX X 11 TOTAL NUMBER OF SPECIES 18 12 21 8 20 13 9 19 14 12 8 9 12 7 8 14 Atlantic Slope Unionidae • Johnson 275 Table 2, A. Unionidae, most of which are widely dim huh n i> in the Northern Atlantic Slope region, all but one of which terminate in the Southern Atlantic Slope region, but are of ddverse origins. £ -5 >> £ CD I- Ok £ 0) •■n d p. X J- s u Vj >. en S * 7 > -- > s 6 >. a w O I. O CD >'■ fct, XB e S* >2 :« Ck «n s© O en — - 5 > ©I. i- A. implicate L. nasuta A. heterodon L. radiate radiata A. undulata S. undulatus L. suhviridis A. varicosa L. cariosa L. ochracea X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X J .5 X X X S > AS x x X X X X e (A ><3S X X X X X o o s |£ .3 *• =■> sfix X X X X •n o 1> (22 X X X X X X X X X X Ana/ys/s of fhe distribution of the species ( Table 2 ) . All of the species of Unionacea found in the Atlantic Slope region are covered in this paper except Margaritifera margaritifera (Linnaeus), whose southern terminus is in Pennsylvania, below the Wisconsin drift border;1 Alasmidonta mar- ginata (Say), which is restricted to the upper Susquehanna River, Pennsylvania; and Anodonta cataracta fragilis Lamarck from Nova Scotia and northward (figured 1 Flint (1957: 361) mentioned that along the Atlantic coast, the limit of the Nebraskan drift border extends somewhat farther south of the limit of the Wisconsin drift border ( sometimes earlier referred to as the terminal moraine), but that it is ill defined. by Athearn and Clarke, 1962: 28, pi. 2, figs. .3-4, and Clarke and Rick, 1963: 15). Forty species comprise the unionid fauna of the Atlantic Slope region, but one, Lampsilis ouata (Say), was artificially introduced from the west at the turn of the century (p. 387). Of the remaining thirty-nine species, sixteen occur in the Northern Atlantic Slope region. Twelve of these are also found in the Southern Atlantic Slope region. Of the twenty- three additional species found on the Southern Atlantic Slope, seven appear to have been derived from the west, nine are endemic to the region, and seven are endemic to individual river systems. The Atlantic Slope unionid fauna is Table 2, B. Unionidae thought to have obiginated on the Atlantic Slope, FOUND IN THE APALACHICOLAN REGION, AND WHICH AFFORD EVIDENCE OF A FORMER CONFLUENCE OF THE APALACHICOLA AND SAVANNAH RIVER SYSTEMS. ? E 0) £ 01 0> >§ i- E 9) E E a> = E a> o £ i) - X en i >< Si w V) VI C3 " rE^. GO >> en tn 0 >> O tn «2 a >. V) C3 w •3 tn S S4 £ '■" 3 1- ££ £ > <2 1 E Crt 0 t u2 7 > 05 53 v. — - a i> ^E g > E. camphmata X X X X X X X X X X X X X X X X X E. lanceolata X X X X X X X X X X X X X X X X A. cataracta cataracta X X X X X X X X X X X X X A. couperiana X X X X X 276 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Table 2, C. Unionidae derived from the Apalachicolan region, found in the Savannah River system, which afford evidence of a former confluence with the apalachicola rlver system, and which terminate in the Southern Atlantic Slope region. £ * II h 2 '■ El 4) S«i X > u£ £s (3 y Zj & "3 > le- Sutf O j- O OJ -1 > JB R, 2*> > ££ C;£ s:£ ££ (D V o» 0) 53 5 ^ 0) v. g& C >. >. >-, r- ^, « "= v> ■y. ""^ iy; £ 'Si 5 ^ r* )x — — s > % 1> ^4 1/ ^ 0j C — ' B > ft > -.c *■* •> UPS Art >tf sec &,as C7. tetralasmus X X X X X X X X X X X X A. imbecilis X X X X X X X V. vibex X X X X X X X X E. icterina X X X X X X X X E. arctata X X X A. triangulate! X X X E. fraterna X distinct, consisting of species clearly de- rived from the fauna of the Interior Basin, and a secondary, or Atlantic (Simpson, 1900: 505), fauna also originally of Interior Basin origin, but whose ancestry is more remote. The Atlantic fauna consists of Elliptio spinosa (Lea), Alasmidonta hetero- don (Lea), and A. undulata (Say). It may also include A. triangulate! (Lea) and Pleurobema collina (Conrad). All of these ten species (Table 2, A) are found in the Northern Atlantic Slope region. One of them, Anodonta implicata, is not found in the Southern Atlantic Slope region, whereas the remaining nine species terminate there. A. implicata, Lampsilis Table 2, D. Summary of the relationships of the Unionidae of the Savannah RlVER SYSTEM. THE ARROWS IN THE FOLLOWING ILLUSTRATION INDICATE THE POSSIBLE DIRECTION OF MIGRATION. Savannah River system Apalachicola River system 1. Pleurobema masoni ( Conrad ) <- 2. Elliptio congaraea (Lea) <- 3. Elliptio fraterna (Lea) <- 4. Elliptio complanata (Lightfoot)— 5. Elliptio icterina ( Conrad ) <- 6. Elli]>tio arctata ( Conrad ) <- 7. Elliptio lanceolata (Lea) — 8. Uniomerus tetralasmus (Say) <- 9. Lasmigona subviridis (Conrad) 10. Alasmidonta triangulata (Lea) <- 11. Alasmidonta varicosa (Lamarck) 12. Anodonta c. cataracta Say — 13. Anodonta imbecilis Say <- 14. Anodonta couperiana Lea — 15. Strophitus undulatus (Say) 16. Carunculina pulla ( Conrad ) <- 17. Villosa vibex (Conrad) <- 18. Villosa delumbis (Conrad) <- 19. Lampsilis cariosa (Say) 20. Lampsilis ochracea (Say) 21. Lampsilis splendida (Lea) — ( 1 ) Pleurobema pyriforme ( Lea ) —(2) Elliptio fraterna (Lea) — ( 3 ) Elliptio fraterna ( Lea ) ->(4) Elliptio complanata (Lightfoot) —(5) Elliptio icterina (Conrad) — ( 6 ) Elliptio arctata ( Conrad ) ->(7) Ellij>tio lanceolata (Lea) — ( 8 ) Uniomerus tetralasmus ( Say ) -(9) Alasmidonta triangulata (Lea) -»(10) Anodonta c. cataracta Say — ( 11 ) Anodonta imbecilis Say -»(12) Anodonta couperiana Say -(13) Carunculina parva (Barnes) -( 14) Villosa vibex ( Conrad ) ■(15) Villosa lienosa (Conrad) Atlantic Slope I'momdak • Johnson 271 Table 2, E. Unionidae ENDEMIC, Willi AN EXCEPTION, TO THE SOI Mil RN Atlas i i< Si OPI REGION. £ a 0) g Is a 3 4) © E p P c = c > - = H PJ c £ £ ?\ M -r- -./: Cfl V3 '75 Cfl — en r w Cfl W5 <^ ~ Cfl Cfl Cfl — V. fi • r >» 3 >» >. i >< >t E **. r" >> >. c S* Ss = >, >» >-. g >. ^ >. - 0 _ Cfl vi © Cfl 5- tfl O i- y> i& 3 .a 8> a > §£ = > §.s i.s = > j- ^ ~.> %.z <5 cs vjrt we u« art ?rt urt £s ph2 ass urt A« >H x~ ShK P. masoni X X X X X X X X X V. const ri eta X X X X X X P. collina X X V. delumbis X X X X X X X X X c. pulla X X X X X E. congaraea X X X X X X L. splendid a X X X X E. dariensis1 X i Also found in the St. Johns River system, Florida. radiata, Alasmidonta varicosa, L. ochracea, and L. cariosa originated on the Northern Atlantic Slope. Strophitus undulatus and Lasmigona subviridis are thought to have crossed the divide of the Appalachian Mountains. Lignmia nasuta probably origi- nated in the lake drainage of Ohio. Alasmi- donta hetewdon and A. undulata are prob- ably relicts. The origin of these species are discussed in Chapter 3. Elliptio complanata and Anodonta cata- racta are found throughout the Northern Atlantic Slope Region, and probably origi- nated there. Elliptio lanceolata extends only to the Susquehanna River system, Pennsylvania, and is thought to be of southern origin. All three of these species (Table 2, B) are discussed (in detail) in Chapter 3. Anodonta couperiana is abun- dant in Peninsular Florida, but since it does not appear to be present in the Suwanee and Withlacoochee river systems, it prob- ably spread into the Apalachicolan region by stream confluence. In the Apalachicolan region, E. com- planata is found only within the Apalaeh- icola River system, where it does not reach the size it does on the Atlantic Slope. E. lanceolata and A. cataracta are more widely distributed than E. complanata; but all three species are much less abun- dant there than they are on the Atlantic Slope. Seven species of Unionidae (Table 2, C) entered the Southern Atlantic Slope region directly from a former confluence of the headwaters of the Apalachicola and Savan- nah river systems, and they all reach the limits of their distribution within this region. Of the twenty-one species of Unionidae found in the Savannah River system, eleven species (Table 2, D) also occur in the Apalachicola River system, and another five were probably derived from Apalach- icolan species. The remaining five spe- cies are clearly of Atlantic Slope origin. Seven species of Unionidae are endemic to the Southern Atlantic Slope (Table 2, E ) and one other extends into the Atlantic- drainage of Peninsular Florida. The fol- lowing possible species pairs, or analogs, are suggested: Savannah River system P. masoni E. congaraea C. pulla V. delumbis Southern Atlantic Slope region P. collina E. dariensis L. splendida Southern Atlantic Slope region V". constricta1 Apalachicola River system P. pyriforme E. fratema or c. crassidens C. parva V. lienosa Southern Atlantic Slope region P. masoni K. congaraea L. radiata radiata Tennessee River system V. vanuxemensis 1 Discussed in Chapter 3, p. 280. 278 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Table 2, F. Unionidae endemic to individual river systems of the southern Atlantic Slope region. 8.1 4) r! -t-1 Pj ^ i a> S-S &£ g-S -5-S <& Otfi c/s« Wrt a a j. c £ o> 5 ^ 'O > IS £. waccamawensis X E. spinosa X E. hopetonensis X E. shepardiana X A. arcula X A. gibhosa X L . dolah raefo rm is X Seven species of Unionidae are endemic to individual Atlantic Slope river systems (Table 2, F), six of them to the Altamaha. These are discussed below. The species pair Elliptio waccamawensis — E. congaraea is recognized. It is suggested that the species in the left hand list evolved in some part of the Alta- maha River system that must have been iso- lated for considerable time. Text-figure 3 shows that the Hazlehurst Terrace caused the Ocmulgee and Oconee rivers to be sep- arated. This terrace was formerly thought to have been formed during the early Pleis- tocene, but has recently been dated as Up- per Miocene (Alt and Brooks, 1965: 407). If this latter date is correct, this isolation probably bears little or no relevance to the present fauna. (See Table 2, G.) CHAPTER 3 The Unionacea of the Northern Atlantic Slope region, with a discussion of the prob- able origin of some of the Atlantic Slope species. The Northern Atlantic Slope region has already been defined as extending from the York River system, Virginia, to the lower St. Lawrence River system, Canada, and beyond to Newfoundland and Labra- dor. All of the thirteen species that re- populated the Northern Atlantic Slope at the end of the Pleistocene above the limit of Wisconsin drift (near the Pennsylvania- New York boundary) are found between the York River system and the drift border. Analysis of the distribution of the species. Sixteen species comprise the Unionacea of the Northern Atlantic Slope region (Table 3). Eight are considered to be of northern origin (Table 3, A); three are of southern origin; three crossed the divide of the Ap- palachian Mountains; one is of ancient origin from palearctic Europe; one is of rather recent western origin. Clearly west- ern species found at certain areas of post- glacial contact, such as upper New York State and Lake Champlain, but which have Table 2, G. Relationships of some of the Unionidae of the Altamaha River system. With the one exception noted above, all of the Unionidae that are endemic to individual river sys- tems in the Southern Atlantic Slope occur in the Altamaha. The following relationships are suggested for some of the species that occur there. Elliptio spinosa (Lea) EUiptio dariensis ( Lea) EUiptio hopetonensis (Lea)1 Elliptio shepardiana ( Lea ) Alasmidonla areata (Lea) Anodonta gibbosa Say Lampsilis dolabraeformis (Lea) Lampsilis splendida ( Lea)2 [Not closely related to any other recent spe- cies, probably a relict.] EUiptio congaraea (Lea) (Allopatric) EUiptio eomplanata ( Lightfoot) (Sympatrie) Elliptio laneeolata (Lea) (Sympatrie) Alasmidonta triangulata (Lea) (Allopatric) Anodonta cataracta cataracta Say (Sympatrie) Lampsilis eariosa (Say) (Allopatric) Lampsilis radiata radiata Gmelin (Allopatric) i Also found in the St. Johns River system, Florida. 2 Endemic to the Southern Atlantic Slope. Atlantic Slope Uniomdak • Johnson 279 Table 3. The Unionacea of the Northern Atlantic Slope region. 1. Margaritifera margaritifera (Linnaeus) 2. Ettiptio complanata (Lightfoot) 3. Ettiptio lanceolata (Lea) 4. Lasmigona subviridis (Conrad) 5. Alasmidonta undulata (Say) 6. Alasmidonta marginata (Say) 7. Alasmidonta varicosa (Lamarck) 8. Alasmidonta heterodon (Lea) 9. Anodonta cataracta cataracta Say 10. Anodonta cataracta fragilis Lamarck 11. Anodonta implicataSay 12. Strophitus undulotus (Say) 13. Ligumia nasuta (Say) 14. LampsUis cariosa (Say) 15. LampsUis ochracea (Say) 16. LampsUis radiata radiata (Gmelin) not spread significantly into the region, are not discussed. Ortmann (1906, and 1913a: 364) sug- gested that the ancestors of the Atlantic Slope Unionidae listed above, with the ex- ception of Elliptio complanata (Table 3, A), migrated around the northern end of the Appalachian Mountains, in preglacial times, by way of the Eirigan River, which flowed in the direction of the present St. Lawrence River. There was no barrier to their dispersal southward because the base- leveled coastal plain then extended much farther seaward. During glacial episodes of the Pleisto- cene, the species found on the eastern side of the mountains were effectively separated from those of the Interior Basin, and speci- ation took place among them, probably south of the limit of glacial drift, early in the period. Later some of these species spread into the Southern Atlantic Slope region where the coastal region was base- leveled. After the Pleistocene, this assem- blage of species occupied most of the northern territory lost by its predecessors. Ortmann (1913a: 361) was unaware of the complete distribution of E. complanata and suggested that it had its center of radiation somewhere on the Southern At- lantic Slope. Matteson (1948a: 131) indi- Taule 3, A. I'mionidae eoi \n i\ the Noiitukkn Atlantic Slope region with close affinities to species of the interior b\m\, believed to have come around the northern end of mil Appalachian Mountains. Interior Basin Northern Atlantic Slope E. dilatata A. marginata A. grandis A. grandis A. grandis L. ovata L. ovata L. radiata sUiquoidea E. complanata A. varicosa A. cataracta cataracta A. cataracta fragilis A. implicata L. cariosa L. ochracea L. radiata radiata cated that the predecessor of complanata reached the Savannah River system from the Alabama-Coosa River system by stream capture. In Chapter 1 of this paper, a dif- ferent order of stream capture is postu- lated. It appears that the Apalachicola and Savannah river systems were connected and separated before the Apalachicola and Alabama-Coosa were joined, judging from the distributional evidence afforded by the unionid fauna. The present distribution of complanata in the Apalachicolan region, as it is restricted to the Apalachicola River system, also indicates this order of stream capture. It was probably already a species before its ancestor, unidentified by Matte- son, was supposed to have arrived on the Atlantic Slope. E. complanata is related to dilatata (figured by Ortmann, 1919: 95, pi. 8, fig. 2) of the Interior Basin. Both spe- cies enjoy a certain ubiquitousness within their areas of distribution. The history of complanata is, in short, probably similar to that of the other species discussed above.1 1 Walker (1910: 131) suggested the same origin of E. complanata as that presented here. He further thought that L. sabrostrata and L. nasuta, and A. calceolus (figured by Clarke and Berg, 1959: 29, fig. 30) and A. heterodon had similar relationships and the same origins as the species pairs discussed above. However, Ortmann's argu- ments as to their origins are thought to be more trenchant and are followed here. 280 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Ortmann (1919: 109) called attention to the presence of complanata in the small tributaries of the West Branch of the Sus- quehanna River, Pennsylvania, close to the divide, and to dilatata in the Allegheny River drainage, but he did not suggest that dilatata entered the Atlantic Slope here and speciated. He did, however, suggest that Strophitus undulatus and Alasmidonta marginata entered the Atlantic Slope here. These species are discussed later. The Margaritiferidae is represented on the Northern Atlantic Slope region by a single species, Margaritifera margaritifera. It is a special case because of its origin, but it is included here since its post-glacial history is similar to that of the other species discussed above. There is good evidence that M. mar- garitifera (figured by Ortmann, 1919: 2, pi. 1, fig. 1 ) reached the Pacific region from Asia during the Miocene or early Pliocene when the two continents were connected by a North Pacific (Bering) bridge (Walker, 1910). Walker further speculated that this species separately reached the Atlantic Slope by an Atlantic bridge over Iceland and Greenland, prob- ably at roughly the same time.1 In any event, M. margaritifera appears to have reached eastern North America before the Pleistocene, during which time its range was restricted. Ortmann (1913a: 378) found it in Pennsylvania just south of the limit of glacial drift and suggested that Pennsylvania was at least one re- fugium where M. margaritifera survived the Pleistocene Epoch. Its present distri- bution, from Pennsylvania to Newfound- land and Labrador, is the result of a Table 3, B. Unionidae found in the Atlantic Slope region which appear to have crossed the divide of the Allegheny Mountains, or are represented by allopatric species. Interior Basin Atlantic Slope region L. subviridis V. oanuxemensis Allegheny River drainage L. subviridis V. const rid a Upper Susquehanna River drainage A. marginata S. undulatus A. marginata S. undulatus 1 The North Pacific bridge is well established. Kurten (1966: 4) believes there were three sepa- rate intermigrations by this route, one in the late Paleocene, another in the early Eocene, and a third at some time in the late Middle or beginning of the late Upper Eocene. He further suggests a North Atlantic connection including the British Isles, Iceland, and Greenland, in the late Paleo- cene and early Eocene. reoccupation of territory. Besides differing anatomically, this species contrasts with the Unionidae in its ecological preferences, since it lives in cold water of high alka- linity. Ortmann (1913a: 371) found Lasmigona subviridis to be the dominant species in the Greenbrier and New rivers of the Kanawha River system, which extends across the divide but drains on the western side of it. With the exception of this one western drainage, the species is limited to, and rather widelv distributed on, the Atlantic Slope. He suggested that this species developed in the western mountain streams from Lasmigona compressa (Lea) (figured by Clarke and Berg, 1959: 31, fig. 29), with which it appears to be allopatric; though, at present, the range of compressa is not close to that of subviridis. It lives in small creeks where the best opportunities for fishes to cross from one river system to another develop. Ortmann thought that it spread into the Atlantic Slope region by way of stream capture with the upper James River, but it has since been reported from the Roanoke River system, permitting the assumption that it might have reached the Atlantic Slope by stream capture be- tween the New and Roanoke rivers. (See p. 285.) Villosa eonstrieta is restricted to the Southern Atlantic Slope, between the upper Catawba River of the Cooper-Santee River system and the James River system. It is Atlantic Slope Unionidae • Johnson 2NI verv close to Villosa vanuxemensis (Lea), (for reference to a figure see Simpson 1914, 1: 165) which is found in the Holston River and elsewhere in the Tennessee River system. It is not obvious where the an- cestor of constricta came from; it may have entered the Roanoke from the New River along with L. suhviridis, if it can be as- sumed that vanuxemensis was more widely distributed in pre-glacial time. Strophitus undulabus and Alasmidonta marginata crossed the divide in central Pennsylvania, in the vicinity of Cambria, Westmoreland, and Indiana counties, where the Susquehanna drainage of the east has largely encroached on the Alle- gheny River of the west, and where stream capture has taken place (Ortmann, 1913a: 368). A. marginata remained restricted to the upper Susquehanna River system, while S. undulatus spread throughout the Atlantic Slope region. Both of these species go into very small streams and, in general, avoid large rivers. Ligumia nasuta is allopatric with L. subrostrata Say (for reference to a figure see Simpson, 1914, 1: 99) of the central and western part of the Interior Basin. It probably developed in the lake drainage of Ohio in postglacial time and spread eastward into the Hudson River, following a course along the baseleveled coastal plain. It reached the James River system in the south and the St. Lawrence River system, Canada, in the north (Ortmann 1913a: 379). This species has a preference for quiet water. Elliptic) lanceolata occurs in the Apa- lachicolan region and on the Atlantic Slope but there it extends only to the Susque- hanna River system, Pennsylvania. The present distribution argues that this is a southern species, but its immediate an- cestors or the area of its speciation is obscure. Alasmidonta undulata and Alasmidonta heterodon appear to be relicts of a much older, westerly derived fauna, since they no longer closelv resemble anv of the mem- bers of the genus found in the Interior Basin, nor do they resemble one another. These two Alasmidonta appear to have long existed on the Southern Atlantic Slope, but are now widely distributed above the drift border. CHAPTER 4 The generic affinities of the Unionacea of the Apalachicolan region, Peninsular Florida, and Southern and Northern At- lantic Slope regions. Summary Over forty genera of Unionidae occur in the Interior Basin, many of which are also found in the Alabama-Coosa River system. In the Apalachicolan region are fourteen genera that occur in the Interior Basin as well as in the Alabama-Coosa River system. An additional genus, Glebula, is limited to the Coastal regions. In the Apalachicolan region also occur Margaritifera, Alasmi- donta s. s., and Quinctincina. The latter is endemic. The six genera found in Penin- sular Florida occur in both the Apalach- icolan and Atlantic Slope regions but are probably derived mostly from the former. Nine of the eleven genera of Unionidae found in the Southern Atlantic Slope region are found in the Apalachicolan region. The exceptions are Lasmigona and Ligumia. Four of the genera appear to be of south- ern origin, and the remainder have affin- ities with both the north and south. Only seven of these genera occur in the North Atlantic Slope region. This region is also occupied by Margaritifera, which does not occur on the Southern Atlantic Slope. Ortmann (1913a: 323), speaking of the Unionidae that reached the Atlantic Slope around the northern end of the Appala- chian Mountains, noticed the diminution of genera from the Interior Basin, and con- cluded that "the Allegheny Mountains formed an important barrier to the east- ward distribution of the bulk of the west- ern fauna." It is noted here that along the 282 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Table 4, A. The genera of Unionacea in the Apalachicolan region. Margaritifera? Fusconaia Ambelma2 Quincuncina Pleurobema Ettiptio Uniomerus Alasmidonta3 Anodonta Anodontoides Strophitus Obovaria4 Carunculina Villosa Lampsilis Medionidus Glebula 1 See note on p. 280. - See note on p. 270. 3 Alasmidonta is represented in the Alabama-Coosa River system by A. mccordi Atheam (1964: 134, pi. 9, figs, a, b). 4 Obovaria is found in the Alabama-Coosa River system and is represented in the Apalachicolan region only in the Escambia River, the next system immediately east of the Alabama-Coosa River system. southern coastal plain, where no such great barrier existed, there is a similar diminution of genera from west to east. It can be seen in Table 1 that fifteen of the seventeen genera found in the Apa- lachicolan region (Table 4, A) occur in the Alabama-Coosa River system. The exceptions are Margaritifera (represented by M. hembeli, a relict), which is also found in one other drainage to the west, the Bayou Teche, Louisiana, and Quincun- cina, which is endemic to the Apalachicolan region. Although Alasmidonta occurs in the Alabama-Coosa River system, it does not appear close to Alasmidonta s. s., which is found only in the Apalachicolan and Atlantic Slope regions. Peninsular Florida is considered to be the area south and east of the Suwannee River system and south of the St. Marys River system. (Both of these systems are Table 4, B. The genera and subgenera of Umonidae in Peninsular Florida. Elliptio s.s. Uniomerus Anodonta s.s [Not represented in this area] Utterbackia Carunculina Villosa Lampsilis s.s Table 4, C. The genera and subgenera of UNIONroAE IN THE SOUTHEBN Atlantic Slope region. Pleurobema s.s. [Not represented in this area] Lexingtonia. Endemic to the Southern Atlantic Slope Elliptio s.s. Canthyria. Endemic to the Altamaha River sys- tem Uniomerus Lasmigona s.s [Not represented in this area] Platynaias Alasmidonta s.s. Prolasmidonta. Endemic to the entire Atlantic Slope Decurambis Anodonta s.s. [Not represented in this area] Pyganodon Utterbackia Strophitus Carunculina Villosa Ligumia Lampsilis s.s. included here in the Apalachicolan region.) In this area occur only six genera (Table 4, B ) derived from the west and north-east. Of the eleven genera ( Table 4, C ) found in the Southern Atlantic Slope region, nine are represented in the Apalachicolan region where there is a total of seventeen genera. Strophitus, Lasmigona, and Ligumia are clearly from the North. Of these, Stro- phitus alone is represented in the Apalach- icolan region by S. subvexus (Conrad), which is endemic to the region. Of clearly southern origin are Uniomerus and Carunculina. Pleurobema and Villosa are probably of southern origin. The genera Elliptio, Alasmidonta, Anodonta, and Lampsilis have affinities with both the southern and northern elements of the fauna. Of the eleven genera ( Table 4, D ) found in the Southern Atlantic Slope region, seven occur in the Northern Atlantic Slope region, where there also occurs an additional genus of Unionacea, Margaritifera. These eight genera repopulated the entire Northern Atlantic Slope above the limit of glacial Atlantic Slope Unionidae • Johnson 283 Table 4, D. The genera and subgenera of Unionacea in the Northern Atlantic Slope region. Margaritifera Elliptio s.s Lasmigona s.s Platynaias Alasmidonta s.s. Prolasmidonta. Endemic to the Atlantic Slope region Decurambis Anodonta s.s. [Not represented in this area] Pyganodon Strophitus Ligumia Lampsilis s.s. drift, with the exception of Lasmigona, which only reaches the Hudson River system, New York. One of these, Ligumia, appears to have entered from the west at the close of the Pleistocene. With this exception, the other genera were estab- lished below the limit of glacial drift be- fore that period. CHAPTER 5 The relationship of the geomorphology and topography of the Apalachicolan and Atlantic Slope regions to their unionid faunas. The Appalachian Mountains have been an effective barrier to the spread of Union- acea from the Interior Basin to the coastal area since postcretaceous time (Ortmann, 1913a: 383), although the original Appala- chian Mountains were formed much earlier in the Permian Period as the result of lateral pressure. They consisted of a num- ber of parallel anticlines and synclines that ran in a northeast-southwestern direction. These folds were pressed against an old block of Archaic rocks, the present Pied- mont Plateau, that lay to the east of them. The highest elevation of these folds was in the south, and the divide was well to the east, close to the old Archaic land. The rivers followed the structure of the mountains, running first between the parallel ridges (anticlines) in the synclinal valleys, and finding their outlets in a north- westerly direction toward the Interior Basin. On the Atlantic side of the divide, shorter streams originated in the highest elevations and flowed east and southeasl across the Archaic rocks. Rivers also formed on the tops of the anticlines, and anticlinal valleys developed, parallel to the synclinal ones. The uppermost rocks of the anticlines were largely hard sandstones of Carboniferous age, below which were softer shales and limestones of Devonian age or older. The synclinal rivers ran uni- formly over the same hard sandstone, but the anticlinal rivers, primarily those on the highest elevations, were the first to cut through the sandstone, allowing the rivers access to the softer rocks below. In time, the anticlinal valleys became more exca- vated than the synclinal valleys. This process advanced farthest in the eastern section of the mountains, so that what was once the highest elevation became a deep valley. By the Cretaceous Period, nearly all of the Appalachian area was peneplaned, with the exception of a chain of monadnocks between eastern Tennessee and North Carolina, and scattered hills in northern New England. (These unreduced areas form the modern Great Smoky Mountains and the summits of the White Mountains in New Hampshire. Elsewhere there are widespread remnants of the peneplane even on the crests of the highest present ridges.) The coastal plain was covered with a shallow sea that deposited a veneer of beds on the Archaic rocks. Because of certain characteristics of the patterns of some of the present streams, it is thought that this sea extended further inland than is now indicated by existing beds. The relief of the present Appalachian Mountains is due almost wholly to a series of broad regional upwarps that occurred in the early Cenozoic Era, giving the pres- ent elevation both to the already extant 284 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 mountain structures and to the subsequent sculpture of this complex mass by erosion. The Appalachian Mountain system of east- ern North America extends in a wide belt, more or less parallel to the coast line for more than 1600 miles from the Gulf Coastal Plain in central Alabama to the St. Law- rence River, Quebec. The mountains are divided from west to east into four prov- inces. 1 ) The Cumberland, or Appala- chian Plateau, which extends from Ala- bama to New York and merges with the Allegheny Plateau. This western area has been peneplaned, but the soft rocks of pre-Carboniferous age have not been reached. 2) The Appalachian Valley and Ridges that reach from Alabama to New York. In the south, this province has a valley-like appearance, but in the north, it becomes the Allegheny Mountains. The eastern part of these were originally much higher. They are cut into by anticlinal streams, which have carved out broad limestone valleys, with high ridges of harder rock between them. Within the Allegheny Mountains, and parallel to them on the east, where there was once the highest elevation, there is now the ex- ceptionally broad Great Allegheny Valley. 3) The Blue Ridge is the highest part of the Southern Appalachian Mountains and extends from northeastern Georgia across western South and North Carolina into southwestern Virginia. The Blue Ridge of the Central Appalachians extends from Central Virginia into Pennsylvania where it continues as South Mountain. It is the flank of an anticline, consisting largely of Archaic rocks. It is less elevated here than in the south. 4) The Piedmont Plateau (or Upland) goes from central Alabama to the Hudson River, New York. To the east of these mountain provinces lies the Coastal Plain; this is a continuation of the Archaic granites of the Piedmont Plateau, over which are much younger Cretaceous and Tertiary deposits of sea- lain limestone and poorly consolidated sand and clay. Waterfalls are so numerous at the boundary between the Piedmont Plateau and the Coastal Plain that it is known as the Fall Line. During the Pleisto- cene, a number of changes in the level of the ocean left terraces on the Coastal Plain that will be discussed below. The whole Appalachian Mountain system has also been divided from south to north into three regions. 1) Southern. North- eastern Georgia to the New or Kanawha River, Virginia, and West Virginia. 2) Central. New River to the Hudson River valley (covered here to the Potomac River system). 3) Northern. Hudson River val- ley to the lower St. Lawrence River. This adumbration of the complex history of the Appalachian Mountains, primarily of the area under study, is offered as background. The relevant details follow. The main backbone of the present Appa- lachian Mountains extends continuously from the divide between the Toccoa and Coosawattee rivers in northeastern Georgia to northern Virginia, and is known as the Blue Ridge. From it the streams flow northwest, southeast, or east. It has been mentioned that the original divide was well to the east, as is the present divide, and that the eastern streams did, and do, have a steeper grade and a shorter, more direct route to the sea. Since they had, and have, more eroding power than the streams of the western drainage, they have encroached on it. There has also been a tendency of the rivers on the Atlantic Slope to capture streams between systems. Evidence of stream capture. The southern extremity of the Blue Ridge is the water- shed of a number of rivers. The Toccoa, the Little Tennessee, and the French Broad all flow into the Interior Basin. Of concern here are the two systems that originate on the southeast side of the mountains and that flow into the Gulf of Mexico; the Etowah River of the Alabama-Coosa River system, which flows westward around Pine Mountain, Georgia, into the Appalachian Atlantic Slope Unioniuai. • Johnson 285 Valley and then into the Gulf; the Chatta- hoochee River of the Apalachicola River system, which flows sonthwestward into the Gulf; and the Tugaloo River of the Savannah River system, which also heads in the mountains and flows southeast into the Atlantic Ocean. On the Piedmont Plateau, as can be seen in Text-figure 1, (see p. 268) all three of these rivers have tributaries that presently are separated by minor divides. Matteson (1948a: 131) sug- gested that confluence occurred among these streams during the late Tertiary. It may have been later, but in any event, the distribution of the Unionidae in the several river systems indicates that confluence did take place. A number of other Atlantic Slope river systems, in addition to the Savannah, have their sources near or in the Blue Ridge. The Cooper-Santee River system, which drains some 15,000 square miles of central South Carolina, has three main tributaries, the Saluda, Broad, and Catawba, all of which originate close to the divide. The Pedee River system's main tributary, the Yadkin, begins in Watauga and Caldwell counties, North Carolina, near the divide. The Roanoke River system, which has the Dan River as a major tributary, forms in Patrick County, Virginia. The Roanoke it- self commences in the eastern foothills of the Blue Ridge. It is only in the southern Appalachians that streams that flow westward still have their sources near, or in, the Blue Ridge. The best example is the New River, which preserves its ancient course and cuts across the divide. According to Campbell (1896: 674) there is a good example of stream capture in the uppermost Roanoke River, Virginia (Plate 1). The headwaters of the Roanoke's North Fork run first in a southwesterly direction in a valley that clearly continued to New River; but just north of Christiansburg, Montgomery County, Virginia, the North Fork makes a sharp bend, cuts through Paris Mountain, then flows eastward and northeastward. It is clear that the Roanoke River has captured here a tributary of the New River. The presence of La.smigona subviridis and Villosa constricta in the Atlantic Slope region adds credence to the geomorphological evidence. Beginning with the Roanoke, the suc- ceeding river systems, the James, Potomac, and Susquehanna, all flow more directly east and each one cuts farther back into the mountains (Campbell, 1896: 675). Campbell also assumes that stream capture has taken place between the headwaters of the Roanoke and James river systems, an assumption which might explain the presence of Villosa constricta in both systems. The Potomac River cuts across the mountains and has, in northeastern West Virginia, reached the Allegheny Plateau. It now drains a longitudinal synclinal val- ley. Campbell (1896) further states that in the mountains, the Potomac River robbed the James River of tributaries, and that the Susquehanna River of Pennsyl- vania encroached on the Potomac. How- ever, these sequences of stream captures seem too ancient to be relevant to the present unionid fauna north of the James. Many of the Unionidae appear to have spread along the baseleveled coastal plain; this spreading is discussed below. The Piedmont Plateau. In Alabama, the Piedmont Plateau is divided into two regions. 1. The Ashland Plateau, the high- est division, is an area of crystalline rocks, diversified by ridges. It is the mountainous portion of the Piedmont and includes Cheaha Mountain, the highest point in the state. It is drained by portions of the Alabama-Coosa River system. 2. The Ope- lika Plateau is lower and is underlain by Archean rocks; it has no striking topo- graphic features. It is drained by the tributaries of the Tallapoosa River of the Alabama-Coosa system, which lies adjacent to the northwest border of the Archaic 286 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 rocks, and by the Chattahoochee River, which forms the eastern boundary of the state. On the Opelika Plateau the rivers have cut valleys some 200 feet below the general surface. The region continues into Georgia as the Greenville Plateau, and is drained by longitudinal streams that flow southwestward and westward to the Gulf. The remaining portion of the Piedmont of Georgia to the north is drained by dendritic streams. Those rivers that originate in the Piedmont of Georgia include: the Flint of the Apalachicola system, which flows into the Gulf, and the Ocmulgee and Oconnee of the Altamaha system and the Ogeechee system, which both flow into the Atlantic. In South Carolina the rivers have their sources in the mountains or on the Coastal Plain. In North Carolina the Piedmont rivers include: the Cape Fear, Neuse, Pamlico, and Chowan systems; and in Vir- ginia, the York and Rappahannock systems. Ortmann (1913a: 340) observed, in the area he studied, that the great majority of Atlantic Slope Unionidae go up into the mountains (where the rivers may have reached Cretaceous limestone) and ap- proach the headwaters without significant depauperation. In the Apalachicola River system, pres- ent pollution prevents verification, but older records indicate that some of the unionid fauna of the Chattahoochee River extended on the Piedmont at least to the vicinity of Atlanta, Georgia. In the Flint River, which forms south of Atlanta, Unionidae are found in Line Creek, close to the headwaters. On the Atlantic Slope of Georgia the unionid faunas of the major river systems — the Altamaha, Ogeechee, and Savannah — end abruptly at the Fall Line. The ex- ceptions in the Altamaha River system are EAliptio complanata and Anodonta c. cata- rocta, which have been found near Stone Mountain, DeKalb County, the headwaters of the Ocmulgee River. In the Savannah River system the exceptions are Elliptio complanata, E. congaraea, and Villosa delumbis, which are found in the Broad River, opposite to which, in Abbeville County, South Carolina, have also been found Pleurobema masoni, Elliptio fra- tema, E. arctata, Lasmigona subviridis, and Lampsilis cariosa. The siliceous character of the Piedmont water, as illustrated in Text-figure 4, might offer an adequate explanation for the dim- inution of the fauna above the Fall Line. However, elsewhere on the Atlantic Slope siliceous water does not appear to be such an inhibiting factor to the distribution of the Piedmont Unionidae. The Coastal Plain. The Coastal Plain is a continuation of the Archean rocks of the Piedmont over which occur Cretaceous and Tertiary deposits of sea-lain limestone and poorly consolidated sand and clay (Text- fig. 2 ) . In spite of the apparent propinquity of the Apalachicola and Altamaha river systems in the Piedmont where the Chatta- hoochee passes close by the source of the Ocmulgee, the unionid fauna does not indicate any direct exchange of species. The Flint River has no large tributaries to the east, and in some places it is as much as fifteen miles from the divide that separates it from the drainage to the east, though Big Indian Creek, near Fort Valley, Houston County, a tributary of the Ocmul- gee River, is only three miles distant. It would appear that the height of the Tifton Upland and Fort Valley Plateau, and the depth of the river channel on the Dough- erty Plain, have prevented stream capture between these river systems for a long time. During the Pleistocene the Coastal Plain was subjected to a number of invasions by the sea. According to Cooke (1945: 248), the maximum and earliest flooding left a record, the Brandywine Terrace [Hazlehurst in Georgia], which correlates with the Aftonian interglacial stage. The sea was thought to be 270 feet above the present level. As mentioned on p. 268, MacNeil cast doubt on the extent of Afonian flood- Atlantic Slope Unionidae • Johnson 287 Text-figure 2. The topographic divisions of the Coastal Plain of Georgia. (From La Forge, 1925, fig. 2.) ing, and concluded that there is no evidence to indicate that the sea ever reached a level more than 150 feet above the present one during the Yarmouth interglacial stage. He regarded the Brandywine Terrace to be of subaerial origin. Russell (1957: 427-428) suggested that the shore lines might have been subjected to differential uplifting. Oaks and Coch ( 1963 ) , on the basis of cores made in Virginia, postulated six cy- cles of the Pleistocene seas, with maximum heights of 45 feet above present levels. They disagreed with the theory of "terrace- stratigraphy" and disapproved of the no- menclature of the terraces, or shore lines. Alt and Brooks (1965) also objected to the nomenclature of the shore line scarps and their associated terraces, since the im- plication is that they refer to independent physiographic units rather than to com- ponent parts of a single physiographic unit. In spite of these valid objections, some of the terrace names are well established in the literature and are retained here, but are used in the broader sense implied by Alt and Brooks. On the basis of the geological evidence, Alt and Brooks (1965) concluded that the 215-250 [270] foot shore line was occupied during the Upper Miocene. This dating was confirmed by Laesse (1968) with bo- tanical evidence. Alt and Brooks also con- cluded that the 90-100 foot shore line was occupied during the Pliocene. Alt ( 1968 ) further concluded that the 70-80 foot shore line also dated from the Pliocene. He rec- ognized three Pleistocene shore lines: one at 40-45 feet, which he assigned to the Af- tonian interglacial; a second at 2.5-30 feet, which he assigned to the longer Yarmouth interglacial; and a third at 5-10 feet, which he assigned to the Sangamon interglacial. 288 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 ''Georgetown oVienna. tericds \ «Cordelp if «Cuthberto[)aw30n S *Lec»burp °Ashburn [ATbnx.y OcFlla. ••Sylvester Tifton A; < — J-)°No_5hvifle ^■'•C ami 1 1 a. Wi wfe— IE— it— Jfe—JTw 'lib Text-figure 3. The Coastal Terraces on the Coastal Plain of Georgia. (From Cooke, 1925, pi. 10, fig. A. The coastal terraces of Georgia are shown in detail in Text-figure 3 because some are especially relevant. These ter- races were formerly defined as sea bottom uncovered and converted into land by a drop in sea level, but Hoyt and Hails ( 1967 ) now regard these Georgia terraces as former lagoon salt marshes. The terraces are not entirely confined to the coastal re- gion, since all of the larger rivers have ter- races of their own. A river terrace may be defined as the old floor of a valley within which an inner valley has been cut by a stream when deepening its channel. Thus, river terraces are fossil flood plains. The various coastal terraces (or shore lines) are correlated below from oldest (top) to youngest: Murray (1961: 508) Cooke (1925: 35) MacNeil (1950: 99) Alt and Brooks (1965) and Alt (1968) General maximum height Georgia Range Florida Time or Stage Brandywine Coharie Sunderland Wicomico IVnholoway Talbot Pamlico 272 feet Hazlehurst 215 170 100 70 42 25 Claxton Okefenokee Penholoway Satilla 270-215 feet Upper Miocene 250-215 feet Okefenokee Pliocene 100- 90 215-160) 160-100 J 100-60 1 Wicomico Late Pliocene 80- 70 60-0 J Aftonian interglacial 50- 45 Pamlico Yarmouth interglacial 30- 25 Sangamon interglacial 10- 5 Atlantic Slope Unionidal • Johnson 289 Text-figure 4. Water types of Georgia. (From Cherry, 1961, fig. 6.) A number of the Coastal Plain rivers appear to have their sources above the level of maximum flooding by the sea; among these are the Escambia, Yellow, Choctawhatchee, Ochlockonee and Su- wanee river systems in the Apalachicolan region and the Satilla and Edisto river sys- tems on the Atlantic Slope. The Yellow, Satilla, and Edisto river systems appear to have been more effectively inundated by the sea, because their small unionid faunas suggest subsequent repopulation. Among the river systems that are com- pletely on that part of the coastal plain that was flooded during the Pliocene is the St. Marys. Text-figure 3 shows that it is entirely on the Olcefenokee and Penholo- way (Wicomico) formations. The river drains the Okefenokee Swamp, and its water, with a low pH, is high in chlorides or silicates (Text-figs. 4 and 5). Although the Satilla River is similar, it differs from the St. Marys in that a small portion of its headwaters is on the Tifton Upland, be- 290 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Text-figure 5. Ranges in pH of Georgia water. (From Cherry, 1961, fig. 8. yond the area of maximum Pliocene and Pleistocene flooding. Both rivers have a small unionid fauna (Table 1). The domi- nant species in the St. Marys is E. c. crassi- dens, and in the Satilla, it is E. crassidem downiei. If the Satilla was a refugium, it is possible that downiei had a chance to evolve there and that E. c. cra.ssidens came back from the west by a route not explained here. The Waecamaw River system in South and North Carolina is of the same origin as the St. Marys. It is completely on the Wicomico and Chowan formations and has a typical coastal assemblage of Unioni- dae (Table 2), but it is remarkable in having an endemic species, E. uaccomo- ivensis. Among the small river systems of more recent origin on the Coastal Plain in the Apalachicolan region in which Unionidae have been found are the Econfina and the Wasassasa. In the Econfina occur Elliptic icterina, ViUosa lienosa, V. vibex, and Atlantic Slope Unionidae • Johnson 291 Lampsilis anodontoides. E. icterina occurs in the Wasassasa. The systems of more recent origin on the Coastal Plain in the Atlantic Slope region are the Combahee, Ashley, Black, and White Oak rivers. The Combahee has five species: E. complanata, E. icterina, Unio- merus tetralasmus, V. delumbis, and V. vibex. In the other river systems occur either, or both, E. complanata and icterina, with the exception of the Black, in which U. tetralasmus also occurs. It is generally regarded that the Pleisto- cene lasted for a million years, although Flint (1957: 301) states that, "all we can say with confidence is that the last major glaciation occurred within 30,000 years, and on the basis of sea floor stratigraphy, the Pleistocene as a whole, embraces at least 300,000 years, perhaps much more." In any event, during the Pleistocene at least five ice sheets pushed southward into the United States from centers of accumulation in Canada. The interglacial times are thought to have been of long duration with sea levels high. During the shorter glacial times the sea must have been relatively low, and the coastal plain often extended much farther to the east than it does at present. Emery (1967) has convincingly shown that as early as 19,000 years ago (lowest level of the sea during the Wisconsin glacial stage) that this was so then. Ortmann (1913a: 383) pointed out that there is a dispersal line directed both north and south on the Coastal Plain, because geophysical barriers have been largely re- moved by baseleveling. Lowland flooding allows fishes to migrate from one system to another. The present paucity of Unionidae in most of the coastal streams of recent origin indicates that dispersal was probably easier during one of the inter-glacial periods. The importance of baseleveling and its faunal significance was well illustrated by Adams (1901). CHAPTER 6 Summary and Conclusions The Apalachicolan region as defined herein includes not only those river systems flowing into the Gulf of Mexico, from I lie Escambia to the Suwannee, but also the St. Marys and Satilla, which flow into the Atlantic Ocean. Forty-nino species com- prise the unionid fauna of this region. Nineteen have affinities with species to the west or Interior Basin, twelve are en- demic to the region, and eleven are re- stricted to individual river systems. Three species that were probably originally en- demic to the region have spread into the Atlantic Slope region, and four others probably originated on the Atlantic slope. The Atlantic Slope region is defined as those streams flowing into the Atlantic Ocean from the Altamaha Biver system, Georgia, to the lower St. Lawrence Biver system, Canada, including rivers in New- foundland and Labrador. Since the area is occupied by an assemblage of species of both southern and northern origin, it can be divided into Southern and Northern regions. The former extends north to the James Biver system, Virginia, the northern- most system with a clearly southern assem- blage of species. Forty species comprise the unionid fauna of the Atlantic Slope region, but one, Lampsilis ovata (Say), was artificially introduced from the west at the turn of the century. Of the remaining thirty-nine species, sixteen occur in the Northern Atlantic Slope region. Of the twenty-three additional species found on the Southern Atlantic Slope, seven appear to have been derived from the west, seven are endemic to the region, and seven are endemic to individual river systems. Two other species, thought to have originated on the Southern Atlantic Slope, are found in Peninsular Florida, and one of them has spread into the Apalachicolan region. 1. Because of their unique mode of dis- tribution, the Unionidae are able to furnish 292 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 important evidence for past conditions of drainage by their present distribution, as was demonstrated by Ortmann (1913a: 381). The discontinuous distribution of the Unionidae of the Apalachicolan region indicates that there was once a confluence of the headwaters of the Alabama-Coosa and Apalachicola river systems. The pres- ence of the same species and analogs in the Savannah River system on the Atlantic Slope as in the Apalachicola River also indicates confluence. The absence of some of the species found in both the Alabama- Coosa and Apalachicola river systems indi- cates that the confluence of the Apalach- icola and Savannah river systems took place first. Two minor confluences are postulated: One between Uphauppee Creek of the Alabama-Coosa River system and Uchee Creek of the Apalachicola River system, the other between the headwaters of the Choctawhatchee River system and a tributary of the Chattahoochee River of the Apalachicola River system. 2. Contained in the fauna of the Apa- lachicolan region are species that are found in the Alabama-Coosa River system, having entered this system through a former con- fluence with it and the Tennessee River system. The region contains mainly en- demic species of Interior Basin origin. A few species have reached it from the At- lantic Slope Region. 3. The presence of endemic species, primarily in the larger Apalachicolan river systems, indicates that these species had refugia above the maximum Pliocene and Pleistocene flooding, and originated before then. 4. Ortmann (1913a: 385) pointed out that the Allegheny Mountains formed an old and well-marked boundary between aquatic animals of the Interior Basin and those of the Atlantic Slope region since post-Cretaceous time. For the Unionidae this boundary includes the entire Appa- lachian Mountain system. 5. The Atlantic Slope unionid fauna is distinct, consisting of species clearly de- rived from the fauna of the Interior Basin, and a secondary, or Atlantic fauna (Simp- son, 1900: 505), originally of Interior Basin origin, but whose ancestry is more remote. 6. In the Atlantic Slope region, most of the species that are clearly of Interior Basin origin are a northern group that migrated around the northern end of the Appa- lachian Mountains before the Pleistocene; there is also a southern group that entered the Atlantic Slope region through a con- fluence of the Apalachicola and Savannah river systems, also in pre-Pleistocene time. 7. The large number of endemic unionids in the Altamaha River system, Georgia, of both old Atlantic and Northern Atlantic Slope origin suggests that this system is particularly old and that it was sufficiently isolated during some of the Pleistocene to have facilitated speciation. 8. Dispersal in the Apalachicolan region appears to have been largely by stream capture. On the Atlantic Slope there is a dispersal line, directed both north and south, where barriers are rendered ineffec- tive by baseleveling ( Adams, 1901 ) . The rate of dispersal was undoubtedly influ- enced by changes in sea level during the Pleistocene. Although it cannot be specifi- cally demonstrated, stream confluence in the uplands also may have played a part in the dispersal of the Atlantic Slope species. 9. In the Appalachian Mountains, there are a few examples of species crossing the divide by stream capture. In addition to the confluence of the Apalachicola and Sa- vannah River systems, there were conflu- ences of the headwaters of the Roanoke and New rivers and of the Allegheny and Sus- quehanna river systems. Evidence of the former existence of these confluences is borne out by the present distribution of some unionid species in the adjoining sys- tems and by the physiography. 10. Margaritifera margaritifera (Lin- naeus), an ancient species from Europe, Atlantic Slope Unionidae • Johnson 293 had a refugium in Pennsylvania below from the St. Lawrence Basin to the Atlantic maximum Pleistocene glaciation, and Ligu- Slope in postglacial time. These are two mia nasuta (Say) appears to have migrated special cases of distribution. PART II. A REVISION OF THE UNIONIDAE FROM THE ST. RIVER, FLORIDA, TO THE POTOMAC RIVER, MARYLAND. MARYS INTRODUCTION In spite of the parochial title of this section, I hold the belief, as do many evo- lutionary biologists, that most faunal studies are an anathema. Their authors seldom accurately determine subspecies (Mayr, et al., 1953: 181) and often even their determination of species is suspect. My excuse for this faunal approach is thought to be justified by the findings presented in Part I. I have studied the superfamily Unionacea for almost thirty years and each of the forty species covered in this section has been completely mono- graphed, including those whose range ex- tends beyond the area studied. Although the synonymy of each species is believed to be complete, and the modern species concept has been assiduously applied, infallibility of judgment is not claimed. If I am guilty of creating composite species, they will probably be found among the ubiquitous Elliptic). Nevertheless, if this is so, the postulations in Part I should not require serious modification since most of the species on which they are based are old and sufficiently discrete to satisfy a typologist. History of the taxonomists and the col- lectors of the Atlantic Slope Unionacea. The Taxonomists Three species of Unionacea found in the Atlantic Slope region were figured in the first printed book devoted entirely to the study of mollusks. This was the great folio volume by Martin Lister, Synopsis Method- icae Conchyliorum, which appeared in London in 1685, at the close of the English Renaissance. This work was cited by Lin- naeus in 1758 for Margaritifera margariti- fera; by Lightfoot1 in 1786 for Elliptio complanata; and by Gmelin in 1791 for Lampsilis radiata radiata. It was not until January of 1817, when volume two of Nicholson's First American Edition of the British Encyclopedia appeared, which in- cluded Thomas Say's article on Conchol- ogy, that many of the common Atlantic Slope Unionidae were described and figured. Eleven of Say's species are recog- nized here. Many of these were redescribed two years later, in 1819, by Lamarck in vol- ume VI of his Histoire Naturelle des Ani- maux sans Vertebres. Only three of his names are recognized in this paper. In 1830 and 1831 Isaac Lea of Philadelphia de- scribe several new species from the Southern Atlantic Slope region, and in 1834 Timothy A. Conrad published a little volume, New Fresh Water Shells of the United States, which included many of the species he had found in the southeastern states. A bitter battle arose between Lea and Conrad over the priority of some of the species, but their polemics, discussed by Wheeler (1935), are only of historical in- terest. Lea continued to describe Unioni- dae from the Southern Atlantic Slope until 1874, by which time he had burdened the literature concerning the few species found there with over one hundred and fifty- names. In this paper, nine of Conrad's taxa and fifteen of Lea's are considered valid. With the demise of Isaac Lea, the de- scribing of the Unionidae of the south- eastern states, exclusive of Florida, was 1 See p. 314. 294 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 essentially over. Between the years 1883- 1934, the Wrights, father and son, de- scribed fifty-two species of mollusks, most of which were from Florida (Johnson, 1967a). The Wrights were a pair of un- sophisticated naturalists who redescribed many species, and aside from Villosa villosa, which is described in this paper, their other relevant taxa are included in the various synonymies. The Collectors The first serious student of fresh-water mollusks to make a collecting trip to South Carolina and Georgia was Timothy A. Conrad. His route of 1833 is shown in de- tail by Wheeler (1935: 26, 27). Before and after this time, most collections were made by interested local residents who sent their material to Isaac Lea, and oc- casionally to Conrad, for description. In the 1830's Lea received Unionidae from Lewis Leconte, J. H. Couper, Professor Shepard, and Edmund Ravenel. These men made it possible for Lea to describe most of the endemic species in the Altamaha River, Georgia. Later, Dr. J. P. Barratt sent him shells from the Abbeville District of South Carolina. From the Caro- lina's Lea received, in his later years, shells from E. Emmons and especially from C. M. Wheatley. No extensive collecting took place on the Southern Atlantic Slope after Lea's time, until W. J. Clench and P. Okkelberg made a trip to Georgia in 1929. This trip was followed by another by Clench and H. van der Schalie in 1932; and then by Clench, C. V. MaCoy, and II. D. Russell in 1934. The collections made in 1929 and 1932 were shared by the Museum of Compara- tive Zoology and the University Museum, University of Michigan. The remaining collections are in the Museum of Compara- tive Zoology. In 1937 van der Schalie made a survey of the mollusks of the Ogeechee River, Georgia, for the Museum of Zoology, University of Michigan. The last major expedition to the Southern Atlantic Slope of Georgia was made by W. J. Clench, K. J. Boss, and S. L. H. Fuller in 1962, for the Museum of Comparative Zoology. Since then, Ruth Patrick ( 1967 ) of the Academy of Natural Sciences of Philadelphia has had J. Bates and S. L. H. Fuller make collec- tions of Unionidae in the Savannah and other Atlantic Slope river systems, in con- nection with her limnological studies. In 1950-51 W. M. Walter made an ex- tensive survey of the Neuse River system, North Carolina. This collection was di- vided among the Museum of Comparative Zoology, United States National Museum, and Museum of Zoology, University of Michigan. In the Carolinas and northward, collect- ing has been adequate, but not organized in a way that can be discussed. Many stations have been made by casual collec- tors, and Ortmann collected systematically in Virginia. THE PRIMARY SYSTEMATIC STUDIES OF APALACHICOLAN AND ATLANTIC SLOPE UNIONACEA. Charles T. Simpson's Notes on the Unionidae of Florida and the Southeastern States (1892) was the first attempt at a revision of any of the species of the South- ern Atlantic Slope region. This work was followed (1900) by his Synopsis of the Naiades, which was expanded into the Descriptive Catalogue of the Naiades (1914). These two works include a re- vision of the Naiades on a worldwide basis. Simpson's species concepts, in so far as they concern the Apalachicolan and At- lantic Slope regions, have been generally accepted, with but few changes, until recently. His were scholarly works. Simp- son was among the more competent mala- cologists of his time. He included in his works almost every taxon relevant to the North American Unionacea that had ap- peared before 1914. After Simpson's work, aside from some Atlantic Slope Unionidae • Johnson 295 studies by Ortmann, little was done on the Unionaeea of the Apalachieolan region until van der Schalie wrote the Naiad Fauna of the Chipola River, in North- western Florida (1940). This was followed by Clench and Turner's Freshwater Mol- lusks of Alabama, Georgia, and Florida from the Escambia to the Suwannee River (1956). Further additions were made by Athearn in Three new unionids from Ala- bama and Florida and a note on Lampsilis jonesi ( 1964 ) and by Johnson in Additions to the unionid fauna of the Gulf drainage of Alabama, Georgia and Florida (1967b). The Unionaeea of the Northern Atlantic Slope region were elegantly monographed by Ortmann, who wrote the Naiades of Pennsylvania, pt. 3 (1919). Recently, Clarke and Berg (1959) made a modern study of the Freshwater Mussels of Central New York which complemented Ortmann's work to the north. This work was followed by the Freshwater Mussels of Nova Scotia by Athearn and Clarke (1962). This brief review of the primary litera- ture on the Unionaeea of the Apalachieolan and Atlantic Slope regions itself explains the reason for carrying the present study to the Potomac River. To have carried it further would have been redundant, in view of the quality of the work of Ort- mann, Clarke and Berg, and Athearn and Clarke. Classification of the Unionaeea The classification based on the anatomy of the genera of Unionaeea proposed by Ortmann (1911, 1912a) is used here. Since Ortmann's time a few necessary nomen- clatural changes have been made, but no one has substantially modified his concepts of the evolution of the North American genera. In 1927 Frierson, an ardent naturalist and an oldfashioned typologist, published A Classified and Annotated Check List of the North American Naiades. He stated his method and philosophy: "Many generic divisions have Keen proposed, based upon characters of diverse kinds, some being those of the hard parts, others those of the soft parts. These characters resemble each other in the respect that they all vary very much, and that none are constant. The general plan adopted herein is to select groups which may he typed by one member, the rest being evident kinfolk. .Such new genera as are named herein are unaccom- panied by diagnoses, a type shell being named, whose chief characters will forever give such diagnoses."' The paper is sprinkled with numerous generic and subgeneric names. It is not unfair to say that most of these names would have fallen into obscurity except for two events. In 1934, Thiele, in his Handbuch der Si/stematischen Weichtier- kunde (2, pt. 3: 815-844), fell under both the systematic and taxonomic influence of Frierson. This is understandable, as he was a compiler, not a specialist in the Unionaeea, and Frierson's work was the latest synthesis, and should have been the last word. In 1942, 1949, and 1964, Modell, in Das natiirliche System der Najaden, pro- posed that the relationships within the Unionaeea can be found in the beak sculpture. He proposed forty-three new names for family groupings. He expected his system to be accepted on the basis of a few diagrammatic sketches of beak sculp- ture, and faith in a single character, which I have not found constant. Regarding the North American Unionaeea he says: "... after Simpson's divisions were not sufficient for these purposes [i.e. questions concerning nomenclature]. I had to go along completely with the uniform system of Frier- son in order to give a better picture. With this however I also had to use Frierson's concepts of Rafinesque's names completely."1 Of course, as mentioned before, the objection is to Frierson's systematics. His nomenclatural eccentricities, in the sense implied by Modell, are not of concern here. 1 Translation by Stansbery and Soehngen, 1964, Sterkiana, no. 14, p. 18. 296 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 In short, it is not implied that Ortmanns anatomical system is perfect, only that it has not been improved on. Haas (1969a), in SuperfarrriUa Unionacea, does not use Modell's system of classification for the higher categories. His species concepts of the North American Unionacea are those of Frierson ( 1927 ) , with a few additions. In Superfamily Unionacea Haas (1969b) only discusses the genera. SYSTEMATIC SECTION The following abbreviations have been used in the text and on the plate captions. ANSP — Academy of Natural Sciences of Philadelphia, Pennsylvania MCZ — Museum of Comparative Zoology, Cambridge, Massachusetts MZUM — Museum of Zoology, University of Michigan, Ann Arbor, Michigan USNM— United States National Mu- seum, Washington, D. C. Synonymy: For ease of reference, full citations are included for each taxon. Else- where in the text, references are abbrevi- ated and require the use of the bibliog- raphy. Isaac Lea often rushed brief Latin de- scriptions of his new species into print, a practice common in his time. These were subsequently followed by adequate de- scriptions and figures, which were then reprinted as "Observations on the Genus Unto." Only page references to this work are included here since the plates and figures were never renumbered, but always kept the same numbers as in the preceed- ing reference. Lea generally gave several localities where each of his species had been found and did not select types, though he always figured a single speci- men, for which he gave measurements. In lieu of the use of the word type,' under Ar- ticle 73 (b) Int. Code Zool. Nomen. (1964), this is an "equivalent expression" and these specimens are regarded as holotypes. Dur- ing the early part of the century, W. B. Mar- shall located most of these figured speci- mens in the United States National Mu- seum. These have been rechecked, and their localities restricted when possible. With but a few mentioned exceptions, all type localities were relocated, and are often rendered more specific from data on original labels and by references to standard atlases, modern county maps, or United States Geological Survey 1: 250,000 maps. These additional data are placed in square brackets. Unless specifically mentioned to the con- trary, all extant types have been examined and usually photographed. Almost none of Say's primary types have survived, and many of Conrad's are missing, whereas most of Lea's have been located. Only pertinent references are included. Simpson (1914) is referred to only when our species concepts are similar or when the differences are easily reconcilable. The papers of Ortmann (1919), Clench and Turner (1956), Clarke and Berg (1959). and Atheam and Clarke ( 1962 ) are all re- ferred to when relevant. Descriptions: The measurements arc only intended to convey the general size of specimens from a given station, or to illustrate sexual differences when they are relevant. Anatomy and Breeding Season: These entries are included when such information is available. For widely distributed species, it should be borne in mind that the dates of the breeding season are not necessarily applicable to their entire range. Habitat: Given when known: Remarks: Statements made in this sec- tion, though applicable to each species over its entire range, are designed primarily to aid in differentiating one species from an- other within the Apalachicolan and At- lantic Slope regions. Range: The ranges are intended to be all-inclusive for each species. Records: The records, limited to the area of the study, are based mostly on speci- Atlantic Slope Unionidae • Johnson 297 mens in the major collections mentioned above under Abbreviations. All of these specimens have been examined, with a lew noted exceptions. Sometimes similar records arc found in several museums. If so, those in the Museum of Comparative Zoology are generally given preference. It is to be assumed that all records are in this museum, unless specifically mentioned to the contrary. A few records from Virginia and Mary- land, of common species, listed by Ort- mann ( 1919 ) have not been reexamined, but are accepted with confidence. In each of these instances, references are given. Walter (1956) did not list his data, but gave numbers on a figure of the Neuse River system. Because the first series of his collection is in the Museum of Com- parative Zoology, his station numbers follow those of his localities which are included here. (See note on p. 323). The records of ubiquitous Elliptio complanata have been selected. They are more copious than is necessary to illustrate the distri- bution of the species, but they are an accurate index to the scope of collecting that has been done in the area covered in : this portion of the paper. In so far as possible, the records are arranged from headwaters to the mouth of the rivers, and from south to north, or west to east, as the case may be. Abbeville District, South Carolina, which occurs many times over in the records, includes the present counties of Abbeville and Greenwood, and most of McCormick. Figures: When available, the holotypes are generally used to illustrate the various species. Often more than one illustration is included to show variability. Frequently, the data on the plate captions is not re- peated elsewhere. Key to the Unionidae Found Between the St. Marys River, Florida, and the Potomac River, Maryland. The following key includes all of the species of Unioniclae that are found between the St. Marys and the Potomac rivers, and it is designed to be used with specimens primarily from this area. The key in Clarke and Berg ( L959: 13) may be used for the identification of die northeastern species of Unionacea. The key is artificial and is based on shell characters only. Some species of Unioni- dae are quite variable and some specimens may not key out correctly. Therefore it is advisable to confirm identifications made with the key by reference to the text and figures. 1. Hinge teeth completely absent. 4 Articulating or vestigial hinge teeth pres- ent. 2 2. Hinge teeth vestigial. 3 Articulating pseudocardinal teeth present. 8 3. Radial wrinkles on the posterior slope, each valve with one thin, depressed tri- angular pseudocardinal tooth; no lateral teeth ...Alasmidonta varicosa p. 354 Radial wrinkles absent on the posterior slope, each valve with a slight depression and then a slightly raised thickening just before the umbos Strophitus undulatus p. 367 4. Umbos inflated, raised above the hinge line. 5 Umbos flat, not raised above the hinge line. 7 5. Shell greatly inflated, especially toward the posterior ventral margin; Altamaha River system only „ Anodonta gibbosa p. 359 Shell somewhat inflated, but uniformly so. 6 6. Shell uniformly thin, nacre bluish white Anodonto cataracta cataracta p. 356 Shell prominently thickened along die an- terior ventral margin below the pallial line; nacre salmon or copper colored Anodonta implicata p. 360 7. Dorsal and ventral margins almost parallel; periostracum with broad green rays. Shell very thin and not much inflated Anodonta imbecilis p. 362 Ventral margin very broadly curved; periostracum with fine green rays. Shell very thin and not much inflated Anodonta couperiana p. 365 8. Lateral teeth absent or vestigial. 9 Articulating lateral teeth present. 11 9. Posterior ridge very rounded; shell ellipti- cal or subrhomboid _ Alasmidonta nndidata p. 349 Posterior ridge sharp; shell subtriangu- lar. 10 10. Umbos anterior to the center, incurved; posterior slope acute Alasmidonta triangulata p. 351 Umbos centrally located, acutely incurved; posterior slope at almost 90° angle with 298 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 posterior ridge; Altamaha River system only Alasmidonta areola p. 352 11. Lateral teeth double in the right valve, single in the left one. Shell quite small Alasmidonta heterodon p. 347 Lateral teeth single in the right valve, double in the left one. 12 12. Pseudoeardinal teeth chunky and serrated, the more prominent one in the right valve not parallel to the hinge line. 13 Pseudoeardinal teeth long and lamellate, with a tendency of the larger one in the right valve to be parallel to the hinge line. 29 13. Shell less than, or about, twice as long as high. 14 Shell distinctly more than twice as long as high. 37 14. Posterior ridge with faint ridges crossing lines of growth. 15 Posterior ridge without faint ridges cross- ing line of growth. 20 15. Shell subrhomboidal, subtriangular or quadrate, often solid, sometimes quite ponderous. 16 Shell subelliptical or elongated rhomboid, generally quite thin. 19 16. Posterior ridge of variable acuteness. 17 Posterior ridge sharp; shell rather thin, widi green rays; Altamaha River system only . ..Elliptio dariensis p. 310 17. Shell ponderous, not rayed, or with thin dark green rays when young, generally blackish; St. Marys River system only Elliptio c. crassidens p. 305 Shell not ponderous, with green rays of varying width, yellowish or chestnut ..Elliptio congaraea p. 308 18. Posterior ridge sharp near the umbos, be- coming faint and biangulate toward the posterior margin. Shell compressed, brown- ish. Young specimens occasionally with faint green rays Elliptio fratema p. 312 Posterior ridge consistently acute. Shell subinflated, brownish, often with narrow green rays over entire surface - ..Elliptio ivaccammuensis p. 313 19. Shell ponderous, inflated; posterior ridge double and of variable acuteness, brown- ish, rhomboid to obovate; Satilla River system only . Elliptio c. downiei p. 307 Shell not ponderous, disks flattened; pos- terior ridge double, but uniformly rounded. 20 20. Dorsal margin long, forming an acute angle with the obliquely descending pos- terior margin. 21 Dorsal margin very long, forming an ex- tremely acute angle with the obliquely descending posterior margin, rendering the posterior end rather wedge shaped. 22 21. Shell rather uniformly trapezoidal, disk flattened; periostracum not usually shiny, often rayed, yellowish green to black Elliptio complanata p. 314 Shell subrhomboidal, often somewhat pointed, very variable as to shape and degree of inflation; periostracum usually subshiny to shiny, often rayed, yellowish to brownish . ...Elliptio icterina p. 325 22. Valves with greatest degree of inflation at midpoint of posterior ridge; shell often rayed when young, becoming blackish widi age. Periostracum smooth on the disk, rough toward the margins; Altamaha River system only Elliptio hopctonensis p. 324 Valves rather uniformily inflated; shell not rayed, occasionally greenish when young, becoming blackish with age. Periostracum generally distinctly satiny Uniomerus tetralasmus p. 339 23. Shell small, not exceeding 65 mm in length, seldom rayed; periostracum yel- lowish, greenish, or brownish. Sexual dimorphism generally present. 24 Shell large, generally rayed, periostracum greenish yellow. Sexual dimorphism pres- ent. 28 24. Shell without spines on disk. 25 Shell with spines on disk. Outline sub- quadrate or subrhomboid; periostracum yellowish or brownish. Sexual differences not well marked Pleurobema collina p. 300 25. Shell with a constriction of the ventral margin before a double, but faint, pos- terior ridge; periostracum smooth. 26 Shell with a regularly rounded ventral margin before a sharp double posterior ridge; periostracum with numerous heavy ridges. Outline of female long obovate; outline of male elliptical Carunculina pulla p. 370 26. Outline subquadrate or rhomboid; perios- tracum often clothlike, blackish brown, occasionally with rays. Sexual difference not well marked ...Pleurobema masoni p. 301 Outline of female short obovate; of male subelliptical. Periostracum yellowish green, widi fine green rays over the entire sur- face — . —Vittosa constricta p. 378 27. Outline a long ellipse; greenish rays over entire surface, periostracum satiny and clothlike - Villosa villosa p. 372 Shell not as above and much larger. 28 28. Posterior ridge well developed, especially toward the umbos. Shell subovate with narrow or wide rays generally distributed over a rather shiny yellowish background. Atlantic Si oil Unionidae • John-son 299 Sexual differences well marked in the shell - Lampsilis ovata p. 386 Posterior ridge low and rounded. Shell suhelliptical, with wide rays generally distributed over a dull yellowish green background. Sexual differences not al- ways well marked in the shell Lampsilis radiata radiata p. 390 29. Left valve without a rudimentary inter- dental projection. 30 Left valve with a rudimentary projection, which fits into an interdental groove in the right valve; shell generally small, 45 to 65 mm in length, though giant speci- mens occur; subrhomboid, often with a low posterior wing when young . ...Lasmigona suhoiridis p. 343 30. Shell without spines on surface. 31 Shell with hollow spines in front of a sharp posterior ridge. Outline subtri- angular; nacre purple. Altamaha River system only Elliptio spinosa p. 303 31. Posterior ridge present and well defined, at least toward the umbos. Shell rather large and solid, yellowish, with varying amounts of the surface covered with green rays. Sexual dimorphism present. 32 Posterior ridge very broadly rounded. Shell small and thin, with numerous green rays over the entire surface, more greenish than yellow. 35 32. Surface of the shell either without rays or with some on posterior slope only. Periostracum shiny. 33 Surface of the shell with rays over the entire surface. Periostracum more often subshiny than shiny. 34 33. Posterior ridge low and rounded except toward the umbos, interdentum present. Nacre bluish white ..Lampsilis cariosa p. 382 Posterior ridge very sharp, interdentum present; Altamaha River system only . Lampsilis dolabraeformis p. 384 34. Nacre bluish; rays narrow. Posterior ridge rounded except toward the umbos. Shell thin, no interdentum Lampsilis ochracea p. 388 35. Periostracum shiny. 36 Periostracum distinctly satiny. Villosa villosa p. 372 36. Narrow green rays, interrupted by growth rests Villosa delumhis p. 375 Wide green rays, not interrupted by growth rests Villosa vibcx p. 373 37. Shell with ventral margin incurved, often sufficiently as to render entire shell ar- cuate. 38 Shell with ventral margin straight or curved, sometimes with a slight dorsal- ventral sulcus. Dorsal and ventral margins roughly parallel. 39 38. Shell somewhat over twice as long as high, suhelliptical, posterior end pro- duced; seldom rayed Elliptio arctata p. 331 Shell over three times as long as high, trapezoidal, with the highest point at the anterior end tapering to a long point; rays visible when young. Altamaha River only Elliptio shepardiana p. 338 39. Nacre silvery white; posterior end dis- tinctly lance-head shaped; rays always present. Sexual dimorphism present. Fe- males with a distinct postbasal swelling. Ligumia nasuta p. 380 Nacre coppery or purple; dorsal and vent- ral margin almost parallel, often rayed, somewhat over twice as long as high Elliptio laneeolata p. 333 Superfamily Unionacea Thiele 19351 Family Unionidae (Fleming 1828) Ortmann 1911 Subfamily Unioninae (Swainson 1840) Ortmann 1910 Genus Pleurobema Rafinesque Subgenus Pleurobema s.s. Rafinesque Pleurobema Rafinesque 1819, Jour. Phys. Chim. Hist. Nat. (Paris), 88: 427. Species listed: P. mytiloides Rafinesque, P. conica ? Rafinesque [nomina nuda]. Rafinesque, 1820, Ann. Gen. des Sci. Phys. (Bruxelles), 5: 313. Species listed: P. mytiloides Rafinesque, P. cuneata Rafinesque. Type species, Unio mytiloides Rafinesque. Sub- sequent designation, Herrmannsen, 1847, Indi- cis Generum Malacozoorum, 1: 292. Ortmann, 1912, Ann. Carnegie Mus., 8: 261. Ortmann and Walker (1922, Occ. Pap. Mus. Zool., Univ. Mich. no. 112, p. 20) state that the description of P. mytiloides in 1820 ren- dered the genus monotypic. This is not so, as the first use of P. mytiloides has no bearing on the subsequent recognition of the genus. Under Pleurobema s. s., Frierson (1927: 40_44) includes thirty-seven species and subspecies. It is probable that some of these taxa will be subsequently reduced to synonymy, but it is clear that this genus 1 Or, Unionoidea ( oidea is the standard ending for superfamily names, though acea has been gen- erally used in molluscan taxonomy). 300 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 has achieved its greatest diversity in the Interior Basin. Subgenus Lexingtonia Ortmann Lexingtonia Ortmann 1914, Nautilus, 28: 28. Type species, Unio subplanus Conrad. Original designation. Ortmann instituted Lexingtonia as a genus because he believed that the six to eight subconcentric ridges of the beak sculpture and the red placentae differenti- ated it from Elliptic- and Pleurobema. He mentions that the shell resembles Fusconaia in which all four gills serve as marsupia, but since he also says that only the outer gills are so used by U. sul)planus Conrad = masoni Conrad, as in Pleurobema, the present author is placing Lexingtonia under Pleurobema as a subgenus, as did Frierson (1927: 44). Under Lexingtonia, Frierson (1927: 44, 45) includes three species. Two of these are included here under the synonymy of Pleurobema (Lexingtonia) masoni (Con- rad). The third, Pleurobema utriculum (Lea) belongs to the Tennessee River drainage. Its synonymy has not been re- investigated, but it probably does not be- long in this subgenus, which appears to be restricted to the Southern Atlantic Slope. Pleurobema (Lexingtonia) collina (Conrad) Plate 2: 1-2 Unio collinus Conrad 1836, Monography Unioni- dae, no. 7, pi. 36, fig. 2 [plate caption only]. 1837, op. cit., no. 8, p. 65 (North [= Maury] River, a branch of the James River [Rockbridge Co.], Virginia; type lot ANSP 20408 [lost]). 1840, op. cit., no. 12, p. 109, pi. 60, fig. 3; this subsequently figured metatype, here se- lected, lectotype ANSP 41007. Alasmidonta collina (Conrad). Simpson, 1914, Cat. Naiades, 1: 501. Pleurobema collina (Conrad). Boss and Clench, 1967, Occ. Papers on Moll., 3: 45-51. Description. Shell small, seldom reach- ing more than 55 mm in length. Outline subquadrate or subrhomboid. Valves sub- compressed, subsolid. Anterior end regu- larly rounded; posterior end angular and slightly cut away below. Ventral margin straight or slightly arcuate centrally. Dor- sal margin almost straight, terminating in a sharp angle with the obliquely truncated posterior margin. Hinge ligament short but prominent. Posterior ridge full, rounded, weakly biangulate toward the base of the shell. Umbos slightly elevated and located somewhat anterior to the middle of the shell, their sculpture consisting of three to four strong subconcentric ridges that form an indistinct rounded angle on the pos- terior ridge. Surface of the disk smooth, with undifferentiated concentric growth rings, surface sculptured with short differ- entiated bilateral spines especially in the region of the posterior dorsal slope and posterior quarter of the disk. Periostracum bright yellow or greenish yellow in young shells, becoming brownish in adults; rarely rayed with fine, narrow, broken, brownish lines. Left valve with two ragged, triangular pseudocardinal teeth and two short, obliquely descending parallel lateral teeth. Right valve with one ragged, triangular pseudocardinal with a vestigial tooth above it; one lateral tooth. Beak cavities com- pressed; anterior adductor muscle scars deep, posterior ones very faint. Pallial line distinct, especially anteriorly. Nacre white, occasionally slightly pinkish, often bluish posteriorly. Length Height Width mm mm mm 55.0 34.7 20.3 Rivanna River, 2 mi. W Columbia, Fluvanna Co., Virginia. 46.0 30.0 19.0 North River [Rockbridge Co.], Virginia. Lecto- type. Anatomtj. Discussed by Boss and Clench (1967: 47). Breeding season. Probably tachytictic, releasing the glochidia during the sum- mer. (Boss and Clench, 1967: 48.) Habitat. Lives on sandy bottoms, in rather swift water. Atlantic Sloi i Unionidae • Johnson 30] Remarks. Pleurobema collina (Conrad) is a distinct Southern Atlantic Slope species. It is remarkable for the spines on the posterior slope of the valves. Usually the maximum number of pairs is three, and, in general, they are nearly bilaterally symmetrical. The first set of spines is about 3-4 mm from the umbo, the second 6-8 mm, and the third from 10-17 mm. The largest spines approach 5 mm in length and have a thickness of more than 1.5 mm. The spines are originally hollow, being closed ventrally at the end of their for- mation. (Boss and Clench, 1967: 47.) Boss and Clench (1967) convincingly pointed out the similarity of P. masoni and collina. When the latter does not have spines, the two species can be easily con- fused, except that masoni is slightly more rhomboid and often has a brownish cloth- like periostracum, generally with at least a trace of green rays on the disk, whereas collina has a periostracum that is more yellowish and smooth, with only an oc- casional hint of brownish ravs. The range of this species was extended by Miss Carol Stein, who collected two small specimens in the Tar River in 1964. Range. Southern Atlantic Slope: Tar River, of the Pamlico River system. North Carolina, and James River system, Virginia. Specimens Examined Pamlico River System Tar River Drainage. North Carolina: Tar River, Old Sparta, 3.5 mi. W Pine Tops, Edgecombe Co. (Ohio State Mu- seum ) . James River System James River Drainage. Virginia: Calf- pasture River (Conrad, 1846: 407); North [= Maury] River, Lexington; James River, near Natural Bridge (USNM), all Rock- bridge Co. James River, Buchanan, Bote- tourt Co. Rivanna River, 2 mi. W Columbia, Fluvanna Co. James River, op- posite Maidens, Goochland Co. Pleurobema (Lexingtonia) masoni (Conrad) Plate 2: 3-10 Unio inaso)ii Conrad 1834, New Fresh Water Shells United States, p. 34, pi. 5, fig. 2 (Savannah River, Augusta [Richmond Co.], Georgia; figured holotype ANSP 41333). Con- rad, 1836, Monography Unionidae, no. 3, p. 28, pi. 12, fig. 2; figured paratype ANSP 41332. Unio subplanus Conrad I : 41 (Neuse River, 6 miles from [East] Raleigh [Wake Co.], North Carolina). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 61, pi. 4, fig. 209: figured holotype USNM 85003. Lea, 1860, Obs. Unio, 8: 65. Unio exactus Lea 1858, Proc. Acad. Nat. Sci. Phila., 10: 41 (Neuse River, 6 miles from [East] Raleigh [Wake Co.], North Carolina). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 62, pi. 4, fig. 210; figured holotype USNM 85002. Lea, 1860, Obs. Unio, 8: 66. Unio postellii Lea 1858, Proc. Acad. Nat. Sci. Phila., 10: 165 (Randall's Creek, near Colum- bus [Muscogee Co.]; Carter's Creek, Baldwin Co.; both Georgia). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 214, pi. 26, fig. 94; figured holotype USNM 85470 from Carter's Creek. Lea, 1859, Obs. Unio, 7: 32. Unio roswellensis Lea 1858, Proc. Acad. Nat. Sci. Phila., 10: 165 (Chattahoochee River, Roswell, Cobb Co., Georgia). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 205, pi. 24, fig. 87; figured holotype USNM 85467. Lea, 1859, Obs. Unio, 7: 23. Unio hurkensis Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 112 (Buckhead Creek, Burke Co., Georgia). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 215, pi. 27, fig. 95; figured holotype USNM 85998. Lea, 1859, Obs. Unio, 7: 33. Unio hallenbeckii Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 170 (Flat Rock Creek; Four Mile Creek; both near Columbus [Muscogee Co.], Georgia). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 328, pi. 51, fig. 154; figured holotype USNM 85537, from Black Dirt Creek, near Columbus [Muscogee Co.], Georgia. Lea, 1860, Obs. Unio, 8: 10. Unio haldwinensis Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 170 (Carter's Creek, Baldwin Co., Georgia). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 330, pi. 51, fig. 155; figured holotype USNM 85420. Lea, 1860, Obs. Unio, 8: 12. Unio salcbrosus Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 170 (Flat Rock Creek; Bull Creek; Chattahoochee River; all near Columbus [Mus- cogee Co.], Georgia). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 332, pi. 52, fig. 157; figured holotype USNM 85574, from Flat Rock Creek. Lea, 1860, Obs. Unio, 8: 14. Unio raeensis Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 171 (Chattahoochee [River], near Columbus [Muscogee Co.] and Rae's Creek [not located] , Georgia). Lea, 1869, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 331, pi. 52, fig. 156; figured holotype USNM 85571. Lea, 1860, Obs. Unio, 8: 13. Unio latus Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 171 (Savannah River, near Savannah [Chatham Co.], Georgia). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 334, pi. 53, fig. 159; figured holotype USNM 85898. Lea, 1860, Obs. Unio, 8: 16. Unio (iiiad ratus Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 172 (Carter's Creek, [Baldwin Co.]; Factory Creek, near Columbus [Muscogee Co.]; both Georgia). Lea, 1860, Jour. Acad. Nat. Atlantic Slope Untonidae • Johnson 317 338, pi. 54, fig. 163; Sci. Phila., ser. 2, I figured holotype USNM 85718 from Factor) Creek. Lea, 1860, Obs. Unio, 8: 20. Unio squamous Lea 1861, Proc. Acad. Nat. Sci. Phila., 13: 391 (North Carolina). Lea, 1862, lour. Acad. Nat. Sci. Phila., ser. 2, 5: 200, pi. 28, fig. 269; figured holotype USNM 85530. Lea, 1863, Obs. Unio, 9: 22. Unio rostrum Lea 1861, Proc. Acad. Nat. Sci. Phila., 13: 391 (Davidson Co., North Caro- lina). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 201, pi. 29, fig. 270; figured holotype USNM 85556. Lea, 1863, Obs. Unio, 9: 23. Unio northamptonensis Lea 1861, Proc. Acad. Nat. Sci. Phila., 13: 392 (Connecticut River, Northampton [Hampshire Co.]; [Connecticut River] Springfield [Hampden Co.]; both Massa- chusetts. [Connecticut River] below Hartford [Hartford Co.], Connecticut; Neuse River, North Carolina). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 190, pi. 25, fig. 260; figured holotype USNM 85527 from the Con- necticut River, Northampton [Hampshire Co.], Massachusetts. Lea, 1863, Obs. Unio, 9: 12. Unio decumbens Lea 1861, Proc. Acad. Nat. Sci. Phila., 14: 40 (Alabama). Lea, 1862, Tour. Acad. Nat. Sci. Phila., ser. 2, 5: 87, pi. 12, fig. 236; figured holotype USNM 86150 [not seen]. Lea, 1862, Obs. Unio, 8: 91. Is E. complanata, fide Frierson, 1922, Nautilus, 36: 44. Unio rcdeighensis Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 191 (Neuse River, 6 mi. East of Raleigh [Wake Co.], North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 6, pi. 1, fig. 2; figured holotype USNM 85592. Lea, 1867, Obs. Unio, 11: 10. Unio aberrans Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 191 (Neuse River, 6 mi. East of Raleigh [Wake Co.], North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 7, pi. 1, fig. 3, figured holotype USNM 85600. Lea, 1867, Obs. Unio, 11: 1. Unio weldonensis Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 191 (Roanoke River, Weldon [Hali- fax Co.], North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 11, pi. 3, fig. 8; figured holotype USNM 85407. Lea, 1867, Obs. Unio, 11: 15. Unio mecklenbergensi.s Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 191 (near Charlotte, Mecklenberg [sic] Co., North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 17, pi. 5, fig. 15, figured holotype USNM 85419. Lea, 1867, Obs. Unio, 11: 21. Unio chathamensis Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 191 (Rocky Run, Chatham Co., North Carolina; James River, near Richmond [Henrico Co.], Virginia). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 21, pi. 6, fig. 19; figured holotype USNM 85672 from Rocky Run. Lea, 1867, Obs. Unio, 11: 25. Unio gastonensis Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 191 (Mine Creek, Gaston Co., North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 20, pi. 6, fig. 18; figured holo- type USNM 85409. Lea, 1867, Obs. Unio, 11: 24. Unio quadrilatcrus Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 192 (Abbeville District [Savan- nah River drainage], South Carolina; Neuse River, near [6 miles East of] Raleigh [Wake Co.]; Catawba River; both North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 5, pi. 1, fig. 1; figured holotype USNM 85385 from the Neuse River. Lea, 1867, Obs. Unio, 11: 9. Unio indefinilus Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 192 (Long Creek, Mecklenburg Co.; Neuse River, near [6 miles East of] Raleigh [Wake Co.]; both North Carolina). Changed to: Unio indefinitus Lea 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 15, pi." 4, fig. 12; figured holotype USNM 85388, from Long Creek. Lea, 1867, Obs. Unio, 11: 19. Unio cistclliformis Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 192 (Neuse River, near [6 miles East of] Raleigh [Wake Co.], North Caro- lina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 19, pi. 6, fig. 17; figured holotype USNM 85533. Lea, 1867, Obs. Unio, 11: 23. Unio mediocris Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 192 (Neuse River, [6 miles East of] Raleigh [Wake Co.], North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 24, pi. 7, fig. 22; figured holotype USNM 85611. Lea, 1866, Obs. Unio, 11: 28. Unio perlucens Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 193 (Six Runs, Sampson Co., North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 18, pi. 5, fig. 16; figured holotype USNM 85076. Lea, 1867, Obs. Unio, 11: 22. Unio curatus Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 193 (Sugar Creek [Mecklenburg Co.], North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 23, pi. 7, fig. 21; type, C. M. Wheatley collection in ANSP; the only specimen located, ANSP 127200, is not figured type. Lea, 1867, Obs. Unio, 11: 27. Unio protensus Lea 1865, Proc. Acad. Nat. Sci. Phila., 17: 33 (North Carolina). Lea, 1868, Tour. Acad. Nat. Sci. Phila., ser. 2, 6: 256, pi. 31, fig. 71; figured holotype USNM 85994, from Eno River, near Hillsboro, Orange Co., North Carolina. Lea, 1868, Obs. Unio, 12: 16. Unio lazarus Sowerby 1868 in Reeve, Conch. Iconica 16, Unio, pi. 68, fig. 348 (Abbeville 318 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 District [Savannah River drainage], South Caro- lina) non Lea, 1852. Unio beaverensis Lea 1868, Proc. Acad. Nat. Sci. Phila., 20: 161 (Beaver Creek [Gaston Co.]; Long Creek [Gaston Co.]; both North Caro- lina. Carter's Creek [Baldwin Co.], Georgia). Lea, 1868, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 297, pi. 44, fig. 109; figured holotype USNM 85598 from Beaver Creek. Lea, 1869, Obs. Unio, 12: 57. Unio nubilus Lea 1868, Proc. Acad. Nat. Sci. Phila., 20: 161 (Paw Creek, Mecklenburg Co., North Carolina). Lea, 1868, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 298, pi. 44, fig. 110; figured holotvpe USNM 85417. Lea, 1869, Obs. Unio, 12: 58. Unio datus Lea 1868, Proc. Acad. Nat. Sci. Phila., 20: 161 (Paw Creek [Mecklenburg Co.]; Beaver Co. [ = Creek, Gaston Co.]; Long Creek [Gaston Co.]; all North Carolina). Lea, 1868, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 299, pi. 44, fig. Ill; figured holotype USNM 85097 labeled, "Paw Creek, Beaver Co., North Caro- lina." Lea, 1869, Obs. Unio, 12: 59. Unio humerosus Lea 1868, Proc. Acad. Nat. Sci. Phila., 20: 161 (Charlotte, Mecklenburg Co., North Carolina). Lea, 1868, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 301, pi. 45, fig. 113; figured holotype USNM 85414. Lea, 1869, Obs. Unio, 12: 61. Unio uhareensis Lea 1868, Proc. Acad. Nat. Sci. Phila., 20: 145 (Uharee [Uwharree] River, Montgomery Co., North Carolina). Lea, 1868, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 304, pi. 46, fig. 116; figured holotype USNM 85576. Lea, 1869, Obs. Unio, 12: 63. Unio tortuosus Sowerby 1868, Conch. Iconica, 16, Unio, pi. 65, fig. 330 (Maryland; figured holo- type BMNH 74.12.11.25). Unio santeensis Lea 1871, Proc. Acad. Nat. Sci. Phila., 23: 193 (Santee Canal, South Carolina; Oconee River, Georgia). Lea, 1874, Tour. Acad. Nat. Sci. Phila., ser. 2, 8: 20, pi. 6, fig. 17; figured holotype USNM 85635, from Santee Canal. Lea, 1874, Obs. Unio, 13: 24. Unio yadkinensis Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 156 (Yadkin River, near Salisbury [Rowan Co.], North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 32, pi. 10, fig. 29; figured holotype USNM 85387. Lea, 1874, Obs. Unio, 13: 36. Unio ampins Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 157 (Irwin's Creek, Mecklenburg Co., North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 39, pi. 13, fig. 36; figured holotype USNM 85591. Lea, 1874, Obs. Unio, 13: 43. Unio ligatus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 157 (Irwin's Creek, Mecklenburg Co.; Long Creek, Gaston Co.; both North Caro- lina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 49, pi. 17, fig. 47; figured holotype USNM 85590, from Irwin's Creek. Lea, 1874, Obs. Unio, 13: 53. Unio differtus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 158 ([Savannah River], Georgia ?). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 42, pi. 14, fig. 39; figured holotype USNM 85399. Lea, 1874, Obs. Unio, 13: 46. Unio sirfjparallclus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 158 (Irwin's Creek [Mecklen- burg Co.]; Fox River [not located]; both North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 46, pi. 16, fig. 44; figured holotype USNM 85418, localities not separated. Lea, 1874, Obs. Unio, 13: 50. Unio oblongus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 158 (Irwin's Creek, Mecklenburg Co., North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 52, pi. 18, fig. 50; figured holotype USNM 85572. Lea, 1874, Obs. Unio, 13: 56. Unio curvatus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 159 (Pfeiffers Pond [Charlotte], Mecklenburg Co., North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 47, pi. 13, fig. 35; figured holotype USNM 85550. Lea, 1874, Obs. Unio, 13: 42. Unio invinensis Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 159 (Irwin's Creek, [Mecklenburg Co.], North Carolina). Lea, 1874, Tour. Acad. Nat. Sci. Phila., ser. 2, 8: 44, pi. 15, fig. 42; figured holotype USNM 85680. Lea, 1874, Obs. Unio, 13: 48. Unio subsquamosus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 160 (Yadkin River, North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 47, pi. 16, fig. 45; figured holotype USNM 85557. Lea, 1874, Obs. Unio, 13: 51. Unio infuscus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 160 (Fox River [not located]; Irwin's Creek, Mecklenburg Co.; both North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 51, pi. 17, fig. 49; figured holotype USNM 85587 from Fox River. Lea, 1874, Obs. Unio, 13: 55. Unio ratus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 160 (Neuse River, North Carolina). Lea, 1874, Tour. Acad. Nat. Sci. Phila., ser. 2, 8: 52, pi. 18, fig. 51; figured holotype USNM 85382. Lea, 1874, Obs. Unio, 13: 56. Unio basalis Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 161 (Carter's Creek, near Colum- bus [Muscogee Co.], Georgia). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 48, pi. 16, fig. 46; figured holotype USNM 85703. Lea, 1874, Obs. Unio, 13: 52. Atlantic Slope Umomdae • Johnson 319 Unio dissimilis Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 161 (Long Creek, Gaston Co.; Pfeiffers Pond, [Charlotte], Mecklenburg Co.; both North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 53, pi. 18, fig. 52; figured holotype USNM 86019 from Long Creek. Lea, 1874, Obs. Unio, 13: 57. Unio cirratus Lea 1874, Proc. Acad. Nat. Sci. Phila. for 1873, 25: 422 (Abbeville District [Savannah River drainage], South Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 55, pi. 19, fig. 53; figured holotype USNM 85584. Lea, 1874, Obs. Unio, 13: 59. Unio subolivaceus Lea 1874, Proc. Acad. Nat. Sci. Phila. for 1873, 25: 422 (Catawba River; Fox River [not located]; Yadkin River; all North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 57, pi. 20, fig. 56; figured holotype USNM 85564. Lea, 1874, Obs. Unio, 13: 61. Unio inful gens Lea 1874, Proc. Acad. Nat. Sci. Phila. for 1873, 25: 422 (Stewart's Pond, Union Co., North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 60, pi. 21, fig. 59; figured holotype USNM 85403. Lea, 1874, Obs. Unio, 13: 64. Unio corneus Lea 1874, Proc. Acad. Nat. Sci. Phila. for 1873, 25: 423 ([Chattahoochee River] Columbus [Muscogee Co.]; Marietta [Cobb Co.]; both Georgia. Abbeville District [Savannah River drainage], South Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 59, pi. 20, fig. 58; figured holotype USNM 85580. Lea, 1874, Obs. Unio, 13: 63. Unio dooleijcnsis Lea 1874, Proc. Acad. Nat. Sci. Phila. for 1873, 25: 424 (Flint River, Dooley Co., Georgia; Abbeville District [Savannah River drainage] South Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 64, pi. 22, fig. 60; figured holotype USNM 85538 from Flint River. Lea, 1874, Obs. Unio, 13: 68. Unio gesnerii Lea 1874, Proc. Acad. Nat. Sci. Phila. for 1873, 25: 424 (Uchee River [Russell Co., Alabama], near Columbus [Muscogee Co.], Georgia). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 65, pi. 22, fig. 64; figured holotype USNM 85670. Lea, 1874, Obs. Unio, 13: 69. Unio invenustus Lea 1874, Proc. Acad. Nat. Sci. Phila. for 1873, 25: 424 ( [Chattahoochee River] Columbus [Muscogee Co.]; Russell Co.; both Georgia. Irwin's Creek [Mecklenburg Co.], North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 66, pi. 22, fig. 62; figured holotype USNM 85704 from Columbus, Georgia. Lea, 1874, Obs. Unio, 13: 70. Unio (Arconaia) provancheriana Pilsbry 1890, Nat. Canadienne, 20: 171 (locality unknown). Pilsbry, 1891, Nautilus, 4: 127. Pilsbry, 1892, Proc. Acad. Nat. Sci. Phila., 44: 132, pi. 7, figs. 4-6; figured holotype ANSP 63094. Unio palliatus 'RaveneV Simpson 1900, Proc. United States Natl. Mus., 22: 730. [nomen nudum]. Listed under the synonymy of Unio errans Lea. Unio pullatus majusculus De Gregorio 1914, II Naturalista Siciliano, 22: 57, pi. 8, figs, a-d (Connecticut; type Palermo Museum, Sicily [not seen] ). Unio complanatus mainensis Rich 1915, Science, n. s., 42: 580 ( Songo Pond about 3 miles South of Rethel [Oxford Co.], Maine; type, [location not known]). Elliptio violaceus (Spengler). Ortmann, 1919, Mem. Carnegie Mus., 8: 94, pi. 8, figs. 4, 5. Elliptio strigosus (Lea), partim. Clench and Turner, 1956, Bull. Florida State Mus., 1: 165. Elliptio complanatus (Solander). Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem. 367, p. 21, figs. 26-28. Athearn and Clarke, 1962, Natl. Mus. Canada, Bull., 183, p. 22, pi. 1, figs. 5, 6; pi. 4, figs. 7, 8. Description. Shell large, often reaching over 100 mm in length. Outline generally long, trajDezoidal, sometimes decidedly rhomboid, occasionally subelliptical. Valves generally flat or subinflated, though oc- casionally considerably inflated, thin to solid. Anterior end regularly rounded; posterior end usually broader and some- what biangulate. Ventral margin straight or slightly arcuate, roughly rjarallel to the long straight or slightly curved dorsal mar- gin, which forms an obtuse angle with the obliquely descending joosterior margin. Hinge ligament prominent. Posterior ridge usually broad, double and rounded, some- times rather angular, ending in a biangu- lation near the base. Posterior slope broad and unsculptured. Umbos low and unin- flated, their sculpture consisting of strong ridges that run nearly parallel to the growth lines and are carried back to the nucleus behind as delicate radial lirae. Surface with irregular growth lines, often nearly smooth in young shells, becoming rougher in old ones. Periostracum brown- ish, or yellowish green, greenish brown, to almost black, often with green rays over the entire surface. 320 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Left valve with two stumpy pseudo- cardinal teeth, one in front of the other, the more anterior one triangular, the hinder one generally not much elevated above the hinge line. Hinge line short and narrow; two long, straight, granular lateral teeth. Right valve with one chunky, ser- rated pseudocardinal, with a vestigial tooth in front of it; one lateral tooth. Beak cavities very shallow, with a few dorsal muscle scars. Anterior and posterior adduc- tor muscle scars and pallial line all distinct. Nacre generally purplish, though often white, or slightly orange; iridescent, espe- cially toward the margins. Length Height Width mm mm mm 154 99 77 82 50 43 44 28 20 Tar River, Bruce, 9 mi. NW Greenville, Pitt Co., North Carolina. Savannah River, 7 mi. NE Newington, Screven Co., Georgia. Turkey Creek, 4 mi. NE Allentown, Wilkinson Co., Georgia. Anatomy. See Ortmann (1912a: 269). Discussed and figured by Reardon (1929: 11, pi. 4, figs. 1-10). Breeding season. Ortmann (1919: 104) found gravid females from April 26 through July 16 during several years in Pennsyl- vania; the breeding season was restudied by Matteson (1948b: 704), who found that in the Ocqueoc River, lower Michigan, it extended from the middle of June to the middle of July. Habitat. Found in lakes, ponds, small streams, and large rivers on nearly every type of substrate, though it appears to prefer sand. Throughout its range it is sometimes the only unionid found at some stations, and at others it is generally found in greater numbers than any other species save where its distribution overlaps Ellip- tio icterina (Conrad). Remarks. Elliptio complanata (Light- foot), which is found at scattered localities in the Apalachicola River system, is the most widely distributed and abundant unionid in the Atlantic Slope region, ex- tending from the upper Altamaha River, Georgia, to northern Canada. It is a very variable species, but most nineteenth cen- tury authors agreed on uniting all of the ecophenotypes north of about the latitude of Washington under complanata. South of Washington, Isaac Lea, especially when in his seventies and eighties, applied spe- cific names to the ecophenotypes of com- planata and icterina as assiduously as Bourguignat and Locard did to the com- mon European unionids, with the same confusion resulting. Simpson (1914, 2: 651 ) did a great deal to rectify the synonymy of complanata. This species, al- though variable in outline, degree of in- flation, and color of both periostracum and nacre, is generally trapezoidal in outline, with the valves rather compressed, and when swollen, the greatest width is in the region of the posterior ridge. In large rivers such as the Savannah, Neuse, Tar, Roanoke, lower Potomac, and Connecti- cut, specimens reach maximum size, and tend to be subelliptical in outline with flattened valves, whereas in smaller rivers and tributaries they tend to be more rhomboid in outline and more inflated. Walter (1956: 266, 270) regarded the small creek form as E. c. complanata and called the large river form E. complanata roanokcnsi.s (Lea), even though both forms were found together at a number of the stations. In the Altamaha River system E. com- planata can be confused with E. hope- tonensis (Lea) (under which see: Remarks on p. 325) which replaces it in the Alta- maha River proper. It can also be confused with E. icterina (Conrad) (under which see: Remarks on p. 328) with whose several ecophenotypes it is often found, south of the White Oak River, North Caro- lina. In Georgia, complanata can not be confused with Uniomerus tetralasmus (Lea) with its acutely angled meeting of Atlantic Slope Unionidae • Johnson 321 the dorsal and posterior margins and its satiny periostracum. These characteristics I become somewhat less constant in the Carolinas and northward, and complanata can be confused with it. However, tetra- lasmus tends to be more quadrate and to exhibit more swelling of the valves, and is always unrayed. The specimens of E. complanata from Lake Creek, at Stone Mountain, De Kalb Co., Georgia, are among just a few records of Unionidae collected above the Fall Line in Georgia. Unfortunately, I misidentified them as hopetonensis for Raulerson and Burbank (1962, 9: 39). Range. Apalachicolan region: restricted to the Apalachicola River system. Atlantic Slope region: Altamaha River system, Georgia, north to the St. Lawrence River system, Canada. Interior Basin: westward to Lake Superior, also parts of the Hudson Bay drainage. Discussed in detail and illustrated by Matteson (1948a: 13, fig. 2). Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Lake Creek at Stone Mountain, De Kalb Co. Walnut Creek, E Macon; Tobesofkee Creek, 5 mi. S Macon (USNM); both Bibb Co. Flat Creek, 2 mi. S Perry; Big Creek, 3 mi. S Henderson; both Houston Co. Buck Creek, 5 mi. NW Hawkins ville; Limestone Creek, 4.3 mi. E Hawkinsville; Cedar Creek, 5 mi. WSW Hawkinsville; Tusca- whatchee Creek, 5 mi. SW Hawkinsville; all Pulaski Co. Bluff Creek, 8 mi. NW Abbeville, Wilcox Co. Dicksons Creek, 10 mi. NW Fitzgerald, Ben Hill Co. Oconee River Drainage. Georgia: Car- ters Creek, Baldwin Co. Turkey Creek, 4 mi. NE Allentown, Wilkinson Co. Ford Branch, 4 mi. W Dublin; Turkey Creek, 9 mi. W Dublin; both Laurens Co. Ohoopee River Drainage. Georgia: Ohoopee River, 4 mi. S Wrightsville, John- son Co. Ogeechee River System Ogeechee River Drainage. Georgia: Ogeechee River, Shoals, Warren Co. Ogeechee Creek, 4 mi. SW Mitchell, Glas- cock Co. Big Creek, 3 mi. SE Louisville; Williamson Swamp Creek, Bartow; Nails Creek, 2 mi. S Bartow; Rocky Creek, 2 mi. S Wadley; Big Creek, 3 mi. E Louisville; all Jefferson Co. Mill Creek, 5 mi. E Mid- ville; Buckhead Creek, 14 mi. W Waynes- boro; Buckhead Creek, 10 mi. SW Waynes- boro; all Burke Co. Ogeechee River, Scarboro (MZUM); Ogeechee River, bridge, 1 mi. S Dover (MZUM); both Screven Co. Ogeechee River, bridge, 1 mi. E Blitchton (MZUM); Ogeechee River, Jinks Bridge, Route 16 (MZUM); Ogeechee River, Morgan Bridge, 14 mi. SE Pembroke (MZUM); all Bryan Co. Savannah River System Broad River Drainage. Georgia: Cadya Creek [not located] (MZUM). Broad River, 1.5 mi. S Bell; Broad River, Hugue- not; both Elbert Co. Savannah River Drainage. South Caro- lina: Abbeville Co. (USNM). Turkey Creek, 7 mi. NW Edgefield, Edgefield Co. Georgia: Savannah River, Augusta, Richmond Co. (MZUM). South Carolina: Lower Three Runs Creek, 8 mi. SW Barn- well, Barnwell Co.; Savannah River, 6 mi. W. Martin; Savannah River, Johnsons Landing, 10 mi. W Allendale; Savannah River, King Jaw Point, 10 mi. WSW Allen- dale (MZUM); all Allendale Co. Georgia: Brier Creek, Keysville (MZUM); Mill Race, 2 mi. N Sardis; both Burke Co. Savannah River, 7 mi. NE Newington, Screven Co. South Carolina: Mouth of Vermezobre Creek (Savannah River Wild- life Res. ) , Jasper Co. ( MZUM ) . Savannah River, Savannah, Chatham Co. COMBAHEE Rr\T:R SYSTEM Salkehatchee River Drainage. South Carolina: Lemon Creek, 2.5 mi. S Bam- 322 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 berg; Little Salkehatchee River, 4 mi. N Ehrhardt; both Bamberg Co. Whippy Swamp Creek, 2.5 mi. NE Crocketville, Hampton Co. Edisto River System Edisto River Drainage. Soutli Carolina: Snake Swamp Creek, Orangeburg Co., 6 mi. NE Bamberg, Bamberg Co. N. Fork Edisto River, Orangeburg, Orangeburg Co. Edisto River, Givhans Ferry, 2 mi. NW Givhans, Dorchester Co. Ashley River System Ashley River Drainage. South Carolina: [Ashley River] near Summersville, Dor- chester Co. Cooper-Santee River System Saluda River Drainage. South Caro- lina: Saluda River, 2.3 mi. above Ware Shoals, Abbeville Co. Broad River Drainage. Tennessee: Green River, 5 mi. ENE Mill Spring, Polk Co. South Carolina: Headwaters, Broad River, SE Reidsville, Spartenburg Co. Congaree River Drainage. South Caro- lina: Congaree River, Columbia, Richland Co. Catawba River Drainage. North Caro- lina: Catawba River, Bridgewater, Burke Co. Leppers Creek, Lincolnton, Lincoln Co. Mine Creek; Beaver Creek; both Gas- ton Co. Paw Creek; Stewarts Creek; Irwins Creek; Sugar Creek; Ashleys Creek; Long Creek; Catawba River; Bissels Pond, Char- lotte; Pfeiffers Pond, Charlotte; Elias Pond, 10 mi. from Charlotte; all Mecklenburg Co. Soutli Carolina: Branch, Little Dutchmans Creek, Rockhill, York Co. Wateree River Drainage. South Caro- lina: Wateree River, 2.5 mi. W Camden, Kershaw Co. Santee River Drainage. South Caro- lina: Poplar Creek, Schulers Fish Pond, near Santee State Park, Orangeburg Co., Wilsons Landing, 5 mi. NW Pineville; Santee Canal; both Berkeley Co. Black River System Black River Drainage. South Carolina: Sammy Swamp Creek, 6 mi. S Paxville, Clarendon Co. Pedee River System Lynches River Drainage. South Caro- lina: Lynches River, 2 mi. NE Bishopville, Lee Co. Yadkin River Drainage. North Caro- lina: Yadkin River, Boomville, Yadkin Co. Yadkin River, Salisbury, Rowan Co. (MZUM). Ditch, Buffalo Creek, 6 mi. E Concord, Cabarrus Co. (MZUM). Uwhar- rie River, Montgomery Co. (USNM). Stewarts Pond, Union Co. (USNM). Waccamaw River System Waccamaw River Drainage. North Carolina: Lake Waccamaw, [town of] Lake Waccamaw, Columbus Co. Soutli Caro- lina: Ditch, Waccamaw River, 1.25 mi. SE Conway, Horry Co. Waccamaw River, Wachasaw Landing, 2 mi. W Murrells Inlet, Georgetown Co. Cape Fear Rwer System Deep River Drainage. North Carolina: Sandv Creek, Randolph Co. (MZUM). Rocky Run (USNM); Deep River, Gulf (Lea); both Chatham Co. Cape Fear River Drainage. North Caro- lina: Haw River, 1.25 mi. NE Benaja, Rockingham Co. (MZUM). Buffalo Creek, 1 mi. E Greensboro, Guilford Co. (MZUM). Morgan Creek, 1 mi. SE Chapel Hill, Orange Co. (MZUM). New Hope River, Burke Forest, Durham Co. (MZUM). Rocky River, 11 mi. N Sanford, Chatham Co. Cape Fear River, Carlos; Cape Fear River, Kinnon; both Cumber- land Co. (both MZUM). Ashe River, Pender Co. Greenfield Mill Pond, Wil- mington, New Hanover Co. (USNM). South River Drainage. North Carolina: Six Runs, Sampson Co. (USNM). Neuse River System Neuse River Drainage. North Carolina: North Flat River, 5.5 mi. S Roxboro Center, Atlantic Slope Unionidae • Johnson 323 Person Co. (SI)1 Eno River, 1.75 mi. ESE Ilillshoro Center ( 117); S Fork, Little River, 0.5 mi. N Schley ( 104 ); Lake Michie, 12 mi. NNE Durham Center (75); Neuse River, 6 mi. ENE Durham Center (64); all Dur- ham Co. Neuse River, 10.5 mi. NE Raleigh (49); Walnut Creek, 4.75 mi. WSW Raleigh (32); Swift Creek, 3 mi. WSW Garner (21); all Wake Co. Little River, 1.25 mi. NW Ragley (2); Neuse River, 3.5 mi. NNE Smithfield (26); both John- ston Co. Neuse River, Cliffs of Neuse State Park, 2 mi. NW Seven Springs, Wayne Co. Neuse River, 13.5 mi. WSW Kingston Cen- ter, Lenoir Co. Neuse River, Streets Ferry, 8.5 mi. NNW New Bern, Craven Co. Trent River, near Pollocksville, Jones Co. (MZUM). Pamlico River System Tar River Drainage. Nortli Carolina: Tar River, Providence, Granville Co. Sandy Creek, Franklin Co. Stony Creek, Nash- ville, Nash Co. [Beach Swamp Creek] Enfield, Halifax Co. Swift Creek, Whit- akers ( MZUM ) ; Tar River, Tarboro; Tar River, Old Sparta, 3.5 mi. W Pinetops; all Edgecomb Co. Tar River, Bruce, 9 mi. NW Greenville; Mill Run, Greenville; both Pitt Co. Roanoke River System Dan River Drainage. Virginia: Aarons Creek, 3 mi. W Buffalo Lithia Springs; Dan River; both Halifax Co. Bluestone Creek, near Clarksville, Mecklenburg Co. Roanoke River Drainage. Virginia: Roanoke River, 4 mi. SW Elliston; N. Fork 1 Walter (1956) marked on a map of the Upper Neuse River some of the 136 stations he made during his survey in 1950-51. He collected E. complanata at forty of them. A selection of these with the station numbers included are presented here, since the localities have never been published. According to the author, most of the collections were made with a 1/8 inch mesh scraper net, which accounts for the paucity of specimens in most of the lots and also why this and other species were not found at more of the stations. Roanoke River, Ironto (USNM); both Montgomery Co. Black Water River, Franklin Co. (MZUM). North Carolina: Roanoke River, Washington Co. Chowan River System Meherrin River Drainage. Virginia: Meherrin River, 2 mi. SW Grandy, Bruns- wick Co. Three Creek, near Emporia, Greensville Co. (MZUM). North Caro- lina: Meherrin River, Murfreesboro, Hert- ford Co. (USNM). Nottoway River Drainage. Virginia: Tommeheton Lake, near Camp Pickett, Nottoway Co. Waqua Creek, 2 mi. SE Rawlings; Nottoway River, 3 mi. E Rawl- ings; both Brunswick Co. Blaek water River Drainage. Virginia: Swift Creek, Rolling Mills, Chesterfield Co. Nansemond River, Nansemond Co. Chowan River Drainage. Virginia: Chowan River, Edenhouse, 13 mi. E Wind- sor, Bertie Co. James River System James River Drainage. Virginia: Calf- pasture River (Conrad, 1846: 407). North [ = Maury] River, Lexington, Rockbridge Co. (USNM). Rivanna River, 2 mi. W Columbia, Fluvanna Co. James River, opposite Maidens, Goochland Co. James River, near Cartersville, Cumberland Co. James River, Richmond, Henrico Co. (MZUM). Chiekahominy River Drainage. Vir- ginia: Chickahominy River, 4 mi. S Ash- land, Hanover Co. York River System North Anna River Drainage. Virginia: Mine Run; Church River; both Orange Co., {both MZUM). Rappahannock River System Rapidan River Drainage. Virginia: Mountain Run, Orange Co. (USNM). Rappahannoek River Drainage. Vir- ginia: Rappahannock River, 1 mi. S Rem- ington, Faquier Co. Mountain Run, 3 mi. N 324 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Lignum, Culpeper Co. Rappahannock River, Fredericksburg, Spotsylvania Co. Occoquan Creek System Occoquan Creek Drainage. Virginia: Rull Run, 3 mi. N Catharpin; Broad Run, [Fairfax Co.], 3 mi. W Manassas; Kettle Run, 1.5 mi. N Nokesville; all Prince William Co. Potomac River System Potomac River Drainage. West Vir- ginia: Cacapon River, 1 mi. below Inter- mont, Hampshire Co. (USNM). Potomac River, Brosius, Morgan Co. Canal at Harp- ers Ferry, Jefferson Co. (MZUM). Virginia: [North River of S Fork Shenandoah River], Weyers Cave, Augusta Co. Pass- age Creek, 2 mi. SW Seven Fountains (USNM); N Fork Shenandoah River, E Woodstock; both Shenandoah Co. Mary- land: Potomac River, 0.75 mi. W Point of Rocks, Frederick Co. Virginia: Potomac River, near Great Falls; Little Hunting Creek, near Mt. Vernon (MZUM); both Fairfax Co. Elliptio {Elliptio) hopetonensis (Lea) Plate 9: 1-2 Unio hopetonensis Lea 1838, Trans. Amer. Philos. Soc, 6: 29, pi. 9, fig. 24 ( [Altamaha River] Hopeton, near Darien [Mcintosh Co.], Georgia; figured holotype USNM 85391). Lea, 1838, Obs. Unio, 2: 29. Simpson, 1914, Cat. Naiades, 2: 668. Unio inusitatis Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 171 (Swift Creek, below Macon [Bibb Co.], Georgia). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 333, pi. 52, fig. 158; figured holotype USNM 85531. Lea, 1860, Obs. Unio, 8: 15. Simpson, 1914, Cat. Naiades, 2: 670. Description. Shell large, often exceed- ing 150 mm in length. Outline subrhom- boid, narrower anteriorly. Valves rather flat to subinflated, occasionally quite in- flated, the greatest inflation being about midway along the posterior ridge, they are thin to solid, inequilateral. Anterior end regularly rounded; posterior end broader, wedge-shaped or slightly biangulate. Ven- tral margin generally straight or slightly arcuate. Dorsal margin long and straight, tending to be a bit winglike where it joins the obliquely descending posterior margin in either a sharp or imperceptible angle. Hinge ligament prominent. Posterior ridge usually rather broad, single, and narrowly rounded above, becoming double below and ending in a wide biangulation at or above the posterior base. Posterior slope rather broad and unsculptured. Umbos low, slightly inflated, located in the an- terior third of the shell, their sculpture consisting of several bars. Disk rather flat, or slightly concave when a slight umbonal- ventral sulcus is present. Periostracum brownish or yellowish green to olive green, often with green rays in young individuals, becoming rough and brownish or blackish in matures. Left valve with two stumpy pseudo- cardinal teeth, one in front of the other, the more anterior one triangular, the hinder one generally not much elevated above the hinge line. Hinge line short and narrow, two long, straight, granular lateral teeth. Right valve with one chunky, serrated pseudocardinal, with a vestigial tooth in front of it; one lateral tooth. Beak cavities very shallow, with a few dorsal muscle scars. Anterior and posterior adductor muscle scars and pallia] line all distinct. Nacre generally white or pinkish, occasion- ally purplish and iridescent, especially toward the margins. Length Height Width mm mm mm 150 75 38 Altamaha River, 3 mi. NW Everett City, Glynn Co., Georgia. 103 55 37 Ohoopee River, 3 mi. N Leman, Emanuel Co., Georgia. 84 43 20 Altamaha River, 4 mi. NE Jesup, Wayne Co., Geor- gia. Anatomy. Discussed by Lea (1863: 415). Habitat. Lives in sand or sandy mud in Atla\ii< Slope Unionidai'. • Johnson 325 the Altamaha River and its major tribu- taries. Remarks. Elliptic) hopetonensis (Lea), which is restricted to the Altamaha River system, is sympatric with Elliptic com- phinata (Lightfoot), which replaces hope- tonensis in the smaller streams of the system. While E. complanata is very vari- able in both outline and degree of inflation, E. hopetonensis is quite constant in outline and in its lack of inflation. When some- what inflated, as it is at the limits of its range, the greatest degree is near the mid portion of the posterior ridge. The posterior ridge is better defined in hopetonensis and the long dorsal margin tends to render most individuals wedge-shaped. The large river form of E. complanata is admittedly very close to hopetonensis, but in river svstems other than the Altamaha, where the large flat-sided form is found, more typical intergrades are generally present. E. hopetonensis (Lea) has been confused with older specimens of E. dariensis (Lea), but the latter has a proportionally higher shell, with a sharper posterior ridge, often with sculpture on the posterior slope. Range. Southern Atlantic Slope: re- stricted to the lower Altamaha River system, Georgia. Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Swift Creek below Macon, Bibb Co. Ocmulgee River, Hawkinsville; Cedar Creek, Fountains Mill, 7 mi. SW Hawkins- ville; both Pulaski Co. House Creek, Bowens Mill, 10 mi. N Fitzgerald, Ben Hill Co. Ocmulgee River, 1.5 mi. S Jackson- ville, Telfair Co. Little Ocmulgee River Drainage. Geor- gia: Gum Swamp Creek, 1 mi. N McRae, Telfair Co. Oconee River Drainage. Georgia: Oco- nee River, 8 mi. SW Soperton. Treutlen Co. Oconee River, 2.5 mi. N Glenwood, Wheeler Co. Ohoopee River Drainage. Georgia: Ohoopee River, Norristown; Ohoopee River, 3 mi. N Leman; both Emanuel Co. Ohoopee River, above Reidsville, Tattnall Co. (ANSP). Altamaha River Drainage. Georgia: Altamaha River, 7 mi. N Ha/.lehurst, Jeff Davis Co. Altamaha River, 10 mi. N Baxley, Appling Co. Altamaha River, "Riverside Park," 4 mi. N Jesup, Wayne Co. Altamaha River, 3 mi. NE Everett City, Glynn Co. Altamaha River, near Fort Barrington; Altamaha River, Hopeton, near Darien; both Mcintosh Co. Elliptio (Elliptio) icterina (Conrad) Plate 9: 3-10 Plate 10: 1-3 Undo icterinus Conrad, [May] 1834, New Fresh Water Shells United States, p. 41, pi. 6, fig. 5 (muddy shore, Savannah River, opposite Augusta [Richmond Co.], Georgia; figured holotype ANSP 41381). Published in May, teste Conrad, 1853, Proc. Acad. Nat. Sci. Phila., 6: 244, and not disputed by Lea, 1854, Proc. Acad. Nat. Sci. Phila., 7: 336-349. Conrad, 1836, Monography Unionidae, no. 4, p. 39, pi. 18, fig. 2. Unio raveneli Conrad, [May] 1834, New Fresh Water Shells United States, p. 39, pi. 6, fig. 4 (Wateree Canal; since found in the small creeks near Cooper River; vicinity of Santee Canal; all South Carolina. Two syntypes ANSP 41370; the smaller one agrees with Conrad's description, but is not the figured specimen, which appears to be Elliptio complanata ( Light- foot). The second specimen is Elliptio lanceo- lata (Lea), non Unio ravenelianus Lea, 1834). Unio ivatereensis Lea 1836, Synopsis Unionidae, p. 31. New name for Unio raveneli Conrad, 1834, non Unio ravenelianus Lea, 1834. As pointed out by Simpson, 1900, Proc. United States Natl. Mus., 22: 748, this change was unnecessary. Unio confertus Lea, [August or September] 1834, Trans. Amer. Philos. Soc, 5: 103, pi. 16, fig. 47 (Santee Canal, South Carolina; type not in USNM [presumed lost]). Lea, 1834, Obs. Unio, 1: 215. Published in August or Septem- ber 1834, teste Lea, 1854, Proc. Acad. Nat. Sci. Phila., 7: 244. Unio lugubris Lea 1834, Trans. Amer. Philos. 326 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Soc, 6: 30, pi. 9, fig. 25 ([Altamaha River], Hopeton, near Darien [Mcintosh Co.], Georgia; figured holotype USNM 85638). Lea, 1838, Obs. Unio, 2: 30, non Say, 1832. Unio geddingsianus Lea 1840, Proc. Amer. Philos. Soc., 1: 285 (Congaree River, South Carolina). Lea, 1842, Trans. Amer. Philos. Soc, 8: 202, pi. 11, fig. 15; figured holotype USNM 85650. Lea, 1842, Obs. Unio, 3: 40. Unio fuscatus Lea 1843, Desc. Twelve Uniones (Black Creek, Florida). Lea, 1846, Trans. Amer. Philos. Soc, 9: 277, pi. 40, fig. 4; figured holotvpe USNM 85243. Lea, 1848, Obs. Unio, 4: 35. Unio occulta* Lea 1843, Desc. Twelve Uniones (Black Creek; Lake Monroe; both Florida). Lea, 1846, Trans. Amer. Philos. Soc, 9: 279, pi. 41, fig. 7; figured holotype USNM 85247, from Black Creek. Lea, 1848, Obs. Unio, 4: 37. Unio limatulus Conrad 1849, Proc. Acad. Nat. Sci. Phila., 4: 154 (Savannah River, [Georgia]; type not in ANSP, [presumed lost]). Conrad, 1850, Jour. Acad. Nat. Sci. Phila., ser. 2, 1: 276, pi. 37, fig. 9. Conrad, 1853, Proc. Acad. Nat. Sci. Phila., 6: 251. Unio tuomeyi Lea 1852, Trans. Amer. Philos. Soc, 10: 256, pi. 13, fig. 4 (Abbeville District [Savannah River drainage], South Carolina; figured holotype USNM 85669). Lea, 1852, Obs. Unio, 5: 12. Unio whiteianus Lea 1852, Trans. Amer. Philos. Soc, 10: 258, pi. 14, fig. 8 (near Savannah [Chatham Co.], Georgia; figured holotype USNM 85658). Lea, 1852, Obs. Unio, 5: 14. Unio barrattii Lea 1852, Trans. Amer. Philos. Soc, 10: 256, pi. 13, fig. 5 (Abbeville District [Savannah River drainage], South Carolina; figured holotype USNM 86010). Lea, 1852, Obs. Unio, 5: 12. Unio pullatis Lea 1856, Proc. Acad. Nat. Sci. Phila., 8: 262 (creeks near Columbus [Mus- cogee Co.], Georgia). Changed to: Unio puUatus Lea 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4:57, pi. 8, fig. 39; figured holo- type USNM 86020. Lea, 1858, Obs. Unio, 6: 57. Unio coruscus Could 1856, Proc Boston Soc. Nat. Hist., 6: 15 (River Saint John's, near Lake Beresford, Florida; measured holotype MCZ 169097, figured by Frierson, 1911, Nautilus, 25, pi. 1, figs. 1-3, and by Johnson, 1964, United States Natl. Mus., Bull. no. 239, p. 60, pi. 32, fig. 3). Unio micans Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 85 (Catawba River, Gaston Co.; Deep River, Gulf [Chatham Co.]; both North Carolina). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 59, pi. 3, fig. 207; figured holotype USNM 85077 from the Catawba River. Lea, 1862, Obs. Unio, 8: 63. Unio obnubilus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 169 (Buckhead Creek, Burke Co., Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 84, pi. 17, fig. 64; figured holotype USNM 85646. Lea, 1858, Obs. Unio, 6: 84. Unio opacus Lea 1857, Proc Acad. Nat. Sci. Phila., 9: 169 (Buckhead Creek, Burke Co., Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 86, pi. 18, fig. 66; figured holotype USNM 85546. Lea, 1858, Obs. Unio, 6: 86. Unio similis Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 169 (Buckhead Creek, Burke Co., Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 91, pi. 19, fig. 71; figured holotype USNM 85653. Lea, 1858, Obs. Unio, 6: 91. Unio sublatus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 169 ([Chattahoochee River], Uchee Bar, below Columbus [Muscogee Co.], Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 82, pi. 16, fig. 62; figured holotype USNM 85897. Lea, 1858, Obs. Unio, 6: 82. Unio viridicatus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 170 (Buckhead Creek, Burke Co., Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 87, pi. 18, fig. 67; figured holotype USNM 85551. Lea, 1858, Obs. Unio, 6: 87. Unio tetricus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 170 (Flint River, near Albany [Dougherty Co.], Georgia). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 195, pi. 22, fig. 78; figured holotype USNM 85655. Lea, 1859, Obs. Unio, 7: 13. Unio aequatus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 170 (Buckhead Creek, Burke Co., Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 89, pi. 19, fig. 69; figured holotype USNM 85561. Lea, 1858, Obs. Unio, 6: 89. Unio aquilus Lea 1857, Proc Acad. Nat. Sci. Phila., 9: 172 (Flint River, Macon [County], Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 92, pi. 20, fig. 72; figured holotype USNM 85993. Lea, 1858, Obs. Unio, 6: 92. Unio viridiradiatus Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 154 (Big Uchee River [Creek, Russell Co., Alabama] near Columbus, Georgia). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 336, pi. 53, fig. 161; figured holotype USNM 86018). Lea, 1860, Obs. Unio, 8: 18. Atlantic Slope Unionidae • Johnson 327 Unto hepaticus Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 151 ( Salkehatchie River, South Carolina). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 348, pi. 57, fig. 173; figured holotype USNM 85559. Lea, 1860, Obs. Unio, 8: 30. Unio viridans Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 170 (near Columbus [Muskogee Co.], Georgia). Lea, 1860, lour. Acad. Nat. Sci. Phila., ser. 2, 4: 337, pi. 54, fig. 162; figured holotype USNM 85579. Lea, 1860, Obs. Unio, 8: 19. Unio verutus Lea 1859, Proc. Acad. Nat. Sci. Phila., 11: 171 (Flat Rock Creek, near Co- lumbus [Muscogee Co.], Ceorgia). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 335, pi. 53, fig. 160; figured holotype USNM 85899. Lea, 1860, Obs. Unio, 8: 17. Unio ocmulgeensis Lea 1861, Proc. Acad. Nat. Sci. Phila., 13: 38 (Little Ocmulgee River, Lumber City [Telfair Co.], Georgia). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 95, pi. 14, fig. 243; figured holotype USNM 85901. Lea, 1862, Obs. Unio, 8: 99. Unio merceri Lea 1862, Proc. Acad. Nat. Sci. Phila., 14: 169 (Lee Co., [Flint River drain- age], Georgia). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 209, pi. 31, fig. 278, figured holotype USNM 86057. Lea, 1863, Obs. Unio, 9: 31. Unio lucidus Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 192 (Livingston's Creek, Bruns- wick Co., [NE corner of Columbus Co., Cape Fear River drainage], North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 9, pi. 2, fig. 6; figured holotype USNM 85242. Lea, 1867, Obs. Unio, 11: 13. Unio livingstonensis Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 192 (Livingston's Creek, Bruns- wick Co., [NE corner of Columbus Co., Cape Fear River drainage] North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila, ser. 2, 6: 14, pi. 4, fig. 11; figured holotype USNM 85536. Lea, 1867, Obs. Unio, 11: 18. Unto ablatus [sic] Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 193 (Long Creek, Gaston Co., North Carolina). Changed to: Unio oblatus Lea 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 13, pi. 4, fig. 10; figured holotype USNM 86001. Lea, 1867, Obs. Unio, 11: 17. Unio radiolus Lea 1871, Proc. Acad. Nat. Sci. Phila., 23: 192 (Ogeechee River, Liberty Co., Georgia). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 21, pi. 6, fig. 18; figured holotype USNM 85621. Lea, 1874, Obs. Unio, 13: 25. Unto cuspitatus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 159 (Buckhead Creek, Burke Co., Ceorgia; Abbeville District I Savannah River drainage], South Carolina). I. '-a, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 43, pi. 14, fig. 50; figured holotype USNM 86014, from Buck- head Creek. Lea, 1874, Obs. Unio, 13: 47. Unto hastatus Lea 1873, Proc. Acad. Nat. Sci. Phila., 25: 423 (New Market, Abbeville District [Savannah River drainage], South Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 56, pi. 19, fig. 54; figured holotype USNM 86013. Lea, 1874, Obs. Unio, 13: 60. Unio fryanus B. H. Wright 1888, Proc. Acad. Nat. Sci. Phila., p. 113, pi. 2, fig. 1 (Lake Ashby, Volusia Co., Florida; figured holotype USNM 151032, refigured by Johnson, 1967, Occ. Papers on Moll., 3: 6, pi. 8, fig. 5). Unto nolani B. H. Wright 1888, Proc. Acad. Nat. Sci. Phila., p. 116, pi. 4, fig. 11 (a creek flow- ing into St. Johns River, near Palatka [Putnam Co.], Florida; holotype USNM 151030, re- figured by Johnson, 1967, Occ. Papers on Moll., 3: 7, pi. 10, fig. 4). Unio simpsoni B. H. Wright 1888, Proc. Acad. Nat. Sci. Phila., p. 117, pi. 5, fig. 1 (Lake Woodruff, Volusia Co., Florida; holotype USNM 151038, refigured by Johnson, 1967, Occ. Papers on Moll., 3: 8, pi. 8, fig. 2). Unio hurtchianus S. H. Wright 1897, Nautilus, 10: 137 (St. Marys River, Nassau Co., Florida; lectotype USNM 149653, selected by Johnson, 1967, Occ. Papers on Moll., 3: 5, pi. 8, fig. 4, possibly the specimen figured by Simpson, 1900, Proc. Acad. Nat. Sci. Phila., p. 80, pi. 4, fig. 8). Unio diazensis S. H. Wright 1897, Nautilus, 11: 5 (Lake Diaz, Volusia Co., Florida; lectotype USNM 149652, selected by Johnson, 1967, Occ. Papers on Moll., 3: 6, pi. 8, fig. 6). Unio dispalans B. H. Wright 1899, Nautilus, 13: 50 (Suwannee River, Florida; holotype USNM 159986, figured by Simpson, 1900, Proc. Acad. Nat. Sci. Phila., p. 80, pi. 1, fig. 9, refigured by Johnson, 1967, Occ. Papers on Moll., 3: 6, pi. 8, fig. 3). Unto singuUris B. H. Wright 1899, Nautilus, 13: 75 (Spring Creek, [a branch of the Flint River], Decatur Co., Georgia; measured holotype USNM 159988, figured by Johnson, 1967, Occ. Papers on Moll., 3: 8, pi. 5, fig. 7). EUiptio cylindraccus Frierson 1927, Check List North American Naiades, p. 29, new name for Unio lugubris Lea, 1838, non Say, 1832. EUiptio maywebbae B. H. Wright 1934, Nautilus, 48: 28; ibid. 47, pi. 13, figs. 5-8 (near Seminole Springs [3.4 miles NE Sorrento], 15 miles SE of Eustis [Lake Co.], Florida, refigured by 328 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Johnson, 1967, Occ. Papers on Moll., 3: 7, pi 10, fig. 3). Elliptio strigosus (Lea), partim. Clench and Tur- ner, 1956, Bull. Florida State Mus., 1: 165. Description. Shell generally small to medium, seldom reaching over 100 mm in length. Outline variable, subquadrate to subelliptical, sometimes rather pointed. Valves subinflated, subsolid to very solid, inequilateral. Anterior end regularly rounded; posterior end generally biangu- late near the base, though sometimes rather produced and pointed. Ventral mar- gin straight or slightly arcuate, sometimes obliquely descending. Dorsal margin slightly curved or almost straight, meeting the obliquely descending posterior margin in a more or less distinct angle. Hinge ligament long and low7. Posterior ridge broadly rounded, generally faintly double. Posterior slope flat to slightly concave, sometimes with very faint radial sculpture present. Umbos broad and full but very low, located in the anterior quarter of the shell, their sculpture consisting of several double-looped ridges. Disk surface gen- erally flat, or slightly concave when a slight umbonal-ventral sulcus is present. Perios- tracum generally fine and shiny, though sometimes heavy and rough, black, brown- ish black, or yellowish brown or bright yellow, chestnut, often with numerous very fine green rays. Left valve with two stumpy pseudo- cardinal teeth, one in front of the other, often of about equal height. Hinge line short and narrow; two long straight lateral teeth. Right valve with two roughly paral- lel pseudocardinals, the posterior one apt to be serrated and chunky, the more an- terior one low and vestigial; one lateral tooth. Beak cavities very shallow, with a few dorsal muscle scars. Anterior and pos- terior adductor muscle scars and pallia! line all distinct. Nacre generally purplish, though sometimes salmon, bluish white, or pinkish, posteriorly iridescent. Length Height Width mm mm mm 98 81 88 49 42 50 29 28 30 Napiers Creek, 6 mi. N Melntyre, Wilkinson Co., Georgia. Mill Race, 2 mi. N Sardis, Burke Co., Georgia. Canoochee River, 2 mi. N Claxton, Evans Co., Geor- gia. Habitat. Found in lakes, ponds, small streams, and large rivers, in nearly every type of substrate. E. icterina (Conrad) is sometimes found with E. complanata (Lightfoot) and other Unionidae at a given station, but like complanata, it is often found alone. In any case, one or the other of these species is generally more abundant than any other unionid at a given station. Remarks. Elliptio icterina (Conrad), which ranges from the Escambia River system in the Apalachicolan region, east through northern Florida, and north to the White Oak River of North Carolina, is a very variable species and a number of populations have been named, some of them several times over. While some popu- lations are more or less identifiable, there is usually a gradual transition of them be- tween one river system and the next, such that while specimens from extremes of the range bear little resemblance to one an- other, there appears to be no point at which subspecies can be established. There is often a great deal of ecophenotypic vari- ation, even at what appears to be a single station, the extremes usually connected by intergrades. In the Apalachicolan region, E. icterina can be confused with E. complanata (Lightfoot) and arctata (Conrad). The latter is a rare species outside of the Ala- bama River system. It is distinctly and con- sistently arcuate with compressed valves, whereas icterina has a generally straight or curved ventral margin, is bluntly or acutely pointed posteriorly, and when oc- casionally produced post-basally, the valves are somewhat inflated. Elliptio icterina Atlantic Slope Unionedae • Johnson 329 occurs with E. complanata in the Apalach- icola River system, and it can be dis- tinguished from complanata by its less rhomboid, more elongate, often pointed shape. The most common shape E. icterina takes in the Apalaehicolan region is sub- rhomboidal to subelliptical, sometimes appearing quite pointed posteriorly if the triangulated posterior ridge ends near the medial line. The tendency to be pointed is more prevalent in specimens from the Apalaehicolan region than in those from the Southern Atlantic Slope. The similarity of the populations of this species from the Chattahoochee River and the upper Savan- nah River, first noted in the localities of some of the taxa described by Isaac Lea, further confirms the commingling of the headwaters of these two systems. The shell form just described includes most of the taxa grouped by Simpson (1914, 2: 661) under U. tuomei/i Lea and by Clench and Turner (1956: " 165-169) under Elliptic strigosus (Lea). Noteworthy are the populations in Moc- casin Creek of Econfina Creek, Bay Co., Florida, on the Gulf Coast: Black Creek, Florida, St. Marys and Canoochee rivers, Georgia, all on the Atlantic Slope; they resemble one another more than they do those from the several intervening river systems. The shells from these rivers tend to be more solid than those from elsewhere, to be more uniformly biangulate poste- riorly, and to have a rather characteristic yellowish brown to shiny chestnut perios- tracum, often with fine dark green rays. In northern Florida, there is an eco- phenotype that lives in lakes and attains much less size than the river one. It has a heavy shell, but tends to be generally oval. In the Wekiva River and Oklawaha rivers of the St. Johns River system, Florida, there occurs toward the headwaters of the streams and in springs a very thin, compressed, subrhomboidal ecophenotype that ends in a broad biangulation below the medial line, the shells of which tend to be heavier, more inflated, and produced post- basally. This shell form occurs again in abundance in Buckhead Creek of the Ogeechee River system; Brier Creek, of the Savannah River system; both Georgia; and the Salkehatchie River of South Car- olina; all on the Atlantic Slope, which, similar to the Floridian rivers mentioned, are known to be rich in carbonates, Text- figure 4. That the shape of the shell is apparently environmentally controlled is also illustrated by the close resemblance of shells from Magnesia Springs, 3 miles W Hawthorn, Alachua Co., Florida; the head- waters of the Ogeechee River; and Cedar Spring, 2 mi. SE Bamberg, Bamberg Co., South Carolina. Though from widely separated drainage systems, shells from these stations bear a closer resemblance to one another than they do to specimens from other stations in their respective drainage systems. The shell form just de- scribed includes most of the taxa grouped by Simpson (1914, 2: 641) under Unio obnubilis Lea and some of those (1914, 2: 639) under Unio confertus Lea. Especially on the Southern Atlantic Slope, often in the lower parts of rivers, specimens of E. icterina tend to be less subrhomboidal than elsewhere, but more elongate and inflated, with a tendency for the ventral margin to be slightly arcuate. The periostracum is often rough, black, and rayless. The shell form just described includes most of the taxa grouped by Simpson (1914, 2: 639) under Unio con- fertus Lea. On the Atlantic Slope of Georgia, E. icterina (Conrad) is most easily confused with Elliptio complanata (Lightfoot) with which it is associated at many of the same stations, but complanata is quite consist- ently rhomboidal, and the valves are less inclined to be inflated. The periostracum of icterina is sometimes bright yellow or chestnut and is generally more shiny and smooth than that of complanata. Range. Apalaehicolan region: Escambia 330 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 River system, east to the St. Marys River system, Georgia. Peninsular Florida. South- ern Atlantic Slope: Altamaha River system, Georgia, north to the White Oak River, North Carolina. Specimens Examined St. Marys River System St. Marys River Drainage. Georgia: St. Marys River, 2 mi. E St. George; St. Marys River, Traders Hill; Spanish Creek, W Folkston; St. Marys River, 4 mi. SSE Folkston; all Charleton Co. Altamaha River System Ocmulgee River Drainage. Georgia: Swift Creek, near Macon; Tobesofkee Creek, 5 mi. S Macon; both Bibb Co. Mossy Creek, 2 mi. S W Kathleen, Houston Co. Limestone Creek, Brumbrys Mill, 4 mi. NE Hawkinsville; Cedar Creek, Foun- tains Mill, 7 mi. SW Hawkinsville; both Pulaski Co. House Creek, Bowens Mill, 9 mi. N Fitzgerald; Dicksons Creek, 10 mi. NE Fitzgerald; both Ben Hill Co. Little Ocmulgee River Drainage. Geor- gia: Little Ocmulgee River, 6 mi. NE Cochran, Beckley Co. Alligator Creek, 2 mi. W Alamo, Wheeler Co. Gum Swamp Creek, 1 mi. N McRae; Little Ocmulgee River, Lumber City (USNM); both Tel- fair Co. Oconee River Drainage. Georgia: Na- piers Creek, 6 mi. N Mclntyre, Wilkinson Co. Ford Branch, 4 mi. W Dublin; Turkey Creek, 9 mi. W Dublin, both Laurens Co. Ohoopee River Drainage. Georgia: Ohoopee River, 4 mi. S Wrightsville; Ohoopee River, 2 mi. NE Adrian; both Johnson Co. Ohoopee River, 1 mi. E Adrian; Mulepen Creek, 2 mi. E Adrian; Ohoopee River, 1 mi. S Norristown; all Emanuel Co. Little Ohoopee River Drainage. Geor- gia: Little Ohoopee River, 1 mi. E Kite, Johnson Co. Little Ohoopee River, 11 mi. W Swainsboro, Emanuel Co. Altamaha River Drainage. Georgia: [Altamaha River], Hopeton near Darien, Mcintosh Co. Frederica, Glynn Co. Ogeechee River System Canoochee River Drainage. Georgia: Canoochee River, Bulloch Co. Canoochee River, 2 mi. N Claxton, Evans Co. Canoo- chee River, near mouth, Bryan Co. (MZUM). Ogeechee River Drainage. Georgia: Rocky Comfort Creek, 1 mi. N Louisville; Ogeechee River, 2 mi. S Louisville; Wil- liamson Swamp Creek, Bartow; Nails Creek, 2 mi. S Bartow; all Jefferson Co. Barkcamp Creek, 7 mi. E Midville; Buck- head Creek, 14 mi. W Wavnesboro; Buck- head Creek, 4 mi. E Vidotte; Buckhead Creek, 9 mi. SW Waynesboro; Mill Creek, 5 mi. NE Midville; Rocky Creek, 5 mi. SW Waynesboro; all Burke Co. Ogeechee River, Scarboro (MZUM); Chew Mill Creek, 8 mi. W Millen; Little Buckhead Creek, 1 mi. N Millen; all Jenkins Co. Savannah River System Savannah River Drainage. South Caro- lina: Abbeville District. (USNM). Turkey Creek, 7 mi. NW Edgefield, Edgefield Co. Georgia: Savannah River, 0.5 mi. S Augusta; Little Spirit Creek, DeBruce; both Rich- mond Co. South Carolina: Lower Three Runs Creek, 8 mi. SW Barnwell, Barnwell Co. Savannah River, 6 mi. W Martin; Savannah River, Johnsons Landing, 10 mi. W Allendale; both Allendale Co. Georgia: Savannah River, 7.5 mi. NE Shell Bluff; Newberry Creek, 2 mi. SE Shell Bluff; Beaverdam Creek, 7 mi. NW Girard; Sweetwater Creek, 2 mi. N Girard; all Burke Co. Brier Creek, 7.5 mi. NE Wrens, Jefferson Co. Brier Creek, Chalker Bridge, 6 mi. N Wavnesboro, Burke Co. Brier Creek, 6 mi. N Sylvania, Screven Co. Rocky Creek, 6 mi. NW Sardis; Miami Branch, 3 mi. NW Sardis; Mill Race, 2 mi. N Sardis; all Burke Co. Near [city of] Savannah, Chatham Co. (USNM). Atlantic Slope Unionidae • Johnson 331 COMBAHEE RlVER SYSTEM Salkehatchee River Drainage. South Carolina: Salkehatchee River, Broxton Bridge; Whippy Swamp Creek, 2.5 mi. NE Crocketville; both Hampton Co. Edisto River System Edisto River Drainage. South Caro- lina: Snake Swamp Creek [Orangeburg Co.], 6 mi. NE Bamberg, Bamberg Co. N. Fork, Edisto River, Orangeburg, Orange- burg Co. Edisto River, Canadys Landing, 8 mi. SW St. George; Edisto River, Givhans Ferry, 2 mi. NW Givhans; both Dorchester Co. ' Cooper-Santee River System Broad River Drainage. North Carolina: Green River, 5 mi. ENE Mill Spring, Polk Co. Catawba River Drainage. North Caro- lina: Catawba River, Bridgewater, Burke Co. Catawba River; Long Creek; Beaver Creek; all Gaston Co. Ashleys Creek; Bissels Pond, Charlotte; both Mecklenburg Co. Wateree River Drainage. North Caro- lina: Wateree River, 2.5 mi. W Camden; Big Pine Tree Creek, 1.5 mi. S Camden; both Kershaw Co. Santee River Drainage. South Carolina: Poplar Creek, Schulers Fish Pond, near Santee State Park, Orangeburg Co. Santee Canal; Cooper River (Conrad); near Charleston; all Berkeley Co. Black River System Black River Drainage. South Carolina: Cane Savannah Creek, 4 mi. SW Sumter; Pocotaligo Creek, 3 mi. S Sumter; [right branch] Pocotaligo River, 9 mi. SSE Sum- ter; Green Swamp Creek, 2.5 mi. W Sumter; all Sumter Co. Sammy Swamp Creek, 3 mi. S Paxville, Clarendon Co. Pedee River System Yadkin River Drainage. North Caro- lina: Yadkin River. Stewarts Pond, Union Co. Pedee River Drainage. South Carolina: Lake Swamp Creek, 7.5 mi. ENE Oates, Darlington Co. Waccamaw River System Waccamaw River Drainage. North Carolina: Lake Waccamaw, [town of] Lake Waccamaw, Columbus Co. South Carolina: Ditch, Waccamaw River, 1.25 mi. SE Con- way, Horry Co. Cape Fear River System Deep River Drainage. North Carolina: Deep River, Gulf (Lea), Chatham Co. Cape Fear River Drainage. North Caro- lina: Rocky River, 11 mi. N Sanford, Chat- ham Co. Livingston Creek, Columbus Co. Ashe River, Pender Co. Cape Fear River. White Oak Ri\t:r System White Oak River Drainage. North Carolina: White Oak River, Maysville, Onslow Co. Elliptio (Elliptio) arctata (Conrad) Plate 10:4-9 Unio arctatus Conrad 1834, Amer. Jour. Sci., 25: 340, pi. 1, fig. 9 (Black Warrior and Alabama Rivers [Alabama]; figured type not in ANSP. Lectotype, here selected, ANSP 41356, pi. 10, fig. 4, Alabama River, from T. A. Conrad). Conrad, 1834, New Fresh Water Shells United States, p. 36, pi. 5, fig. 4. Unio strigosus Lea 1840, Proc. Amer. Philos. Soc, 1: 287 (Chattahoochee River, Columbus [Mus- cogee Co.], Georgia). Lea, 1842, Trans. Amer. Philos. Soc, 8: 198, pi. 9, fig. 9; figured holo- type USNM 85890. Lea, 1842, Obs. Unio, 3: 36. Unio tortivus Lea 1840, Proc. Amer. Philos. Soc, 1: 287 (Chattahoochee River, Columbus [Mus- cogee Co.], Georgia). Lea, 1842, Trans. Amer. Philos. Soc, 8: 204, pi. 12, fig. 17; USNM 85674 labeled type, does not agree exactly with the figure and is from Lee Co., Georgia, [Flint River drainage]. Lea, 1842, Obs. Unio, 3: 42. Unio lazarus Lea 1852, Proc. Amer. Philos. Soc, 5: 251 (Abbeville District [Savannah River drainage], South Carolina). Lea, 1852, Trans. Amer. Philos. Soc, 11: 259, pi. 14, fig. 9; figured holotype USNM 86155. Lea, 1852, Obs. Unio, 5: 15. 332 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Unio perstriatus Lea 1852, Proc. Amer. Philos. Soc, 5: 252 (Abbeville District [Savannah River drainage], South Carolina). Lea, 1852, Trans. Amer. Philos. Soc., 10: 255, pi. 12, fig. 3; figured holotype USNM 85892. Lea, 1852, Obs. Unio, 5: 11. Unio gracilentus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 85 (Catawba River, Gaston Co., North Carolina). Lea, 1862, Tour. Acad. Nat. Sci. Phila., ser. 2, 5: 58, pi. 3, fig. 205; figured holotype USNM 85976. Lea, 1862, Obs. Unio, 8: 62. Unio perlatus Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 193 (Cape Fear River, Black Rock Landing [ = 2 mi. S Kings Bluff, Bladen Co.], North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 15, pi. 4, fig. 13; figured holotype USNM 86006. Lea, 1867, Obs. Unio, 11: 19. EUiptio strigosus (Lea), partim. Clench and Turner, 1956, Bull. Florida State Mus., 1: 165. Description. Shell small, seldom exceed- ing 60 mm in length. Outline elongated, subelliptical, distinctly arcuate. Valves compressed, subsolid, inequilateral. An- terior end regularly rounded; posterior end more broadly rounded, somewhat biangu- late, produced and extending below the base line. Ventral margin arcuate. Dorsal margin slightly curved, usually indistinctly joining the obliquely descending posterior margin. Hinge ligament prominent, lo- cated near the middle of the shell. Posterior ridge low and rounded, usually with a faint second ridge above. Posterior slope slightly concave and smooth. Umbos not swollen, nor raised above the hinge line, located in the anterior sixth of the shell, their sculp- ture consisting of strong ridges. Disk generally slightly concave, caused by a slight umbonal-ventral sulcus. Surface with well-marked but delicate growth lines, often smooth and shiny, becoming darker and duller toward the borders. Periostra- cum usually greenish or yellowish with green rays over the entire surface becom- ing wider and more prominent posteriorly, sometimes dark burnt-brown especially on older shells. Left valve with two low, rather vestigial, somewhat triangular pseudocardinal teeth. Hinge line short and very narrow; two low, short, straight lateral teeth. Right valve with one low but rather chunky pseudo- cardinal; one lateral tooth. Beak cavities very shallow, with a few dorsal muscle scars under the hinge plate. Anterior ad- ductor muscle scars well impressed, pos- terior ones less so. Pallial line distinct, especially anteriorly. Nacre dirty white or purplish with yellowish splotches, slightly iridescent posteriorly. Length Height Width mm mm mm 55 22 14 56 24 13 52 23 12 49 25 15 Alabama River, Alabama. Lectotype of Unio arctatus Conrad. Chattahoochee River, Co- lumbus, Muscogee Co., Georgia. Holotype of Unio strigosus Lea. Abbeville District, South Carolina. Holotype of Unio lazarus Lea. Chattahoochee River, Co- lumbus, Muscogee Co., Georgia. Figured type of Unio tortivus Lea, after Lea. Habitat. Lives in rivers near the shore, among and under rocks. Remarks. On the Atlantic Slope, EUip- tio arctata (Conrad) is known only from the several type lots, and from a rather large unlocalized series from the Cape Fear River, North Carolina. In spite of the very erratic distributional pattern, there appears to be no morphological dif- ference between the specimens from the Cape Fear River and those from the Ala- bama River system, where it is most abundant. It becomes distinctly rare east- ward. The thinness and compression of the valves, subelliptical outline, somewhat pro- duced posterior end, and quite consistently arcuate ventral margin distinguish this species from any other EUiptio in the Atlantic Slope region. Range. Alabama-Coosa River system. Apalachicolan region: Escambia River system, east to the Apalachicola River system, but discontinuous. Southern At- Atlantic: Slope Unionidae • Johnson 333 lantic Slope: upper Savannah River system, South Carolina; Catawba River, North Carolina of the Cooper-Santee River system, Lower Cape Fear River system. North Carolina. Specimens Examined Savannah River System Savannah River Drainage. Soutli Caro- lina: Abbeville District (USNM). Cooper-Santee River System Catawba River Drainage. North Caro- lina: Catawba River, Gaston Co. (USNM). Cape Fear River System Cape Fear River Drainage. North Caro- lina: Cape Fear River, Rlack Rock Land- ing, 2 mi. S Kings Bluff, Bladen Co. (USNM). Cape Fear River. Elliptio {Elliptio) lanceolata (Lea) Plate 10: 10 Plate 11: 1-6 Obliquaria cuprca Rafinesque 1820, Ann. Gen. Sci. Phys. (Bruxelles), 5: 304, pi. 81, figs. 8, 9 (Le Monongahela et le Potowmak; type not in ANSP [lost]). Ortmann and Walker (1922: 31) have pointed out that this species might be either E. dilatata Rafinesque, of the Monon- gahela River or producta Conrad [ = lanceo- lata Lea] of the Potomac River. Conrad (1834: 68), as first reviser of this species, referred it to dilatata. The type locality is here restricted to the Monongahela River, Pennsylvania, and O. cuprea remains a synonym of E. dilatata Rafinesque. Unio lanceolatus Lea 1828, Trans. Amer. Philos. Soc, 3: 266, pi. 3, fig. 2 (Tar River, Tar- borough [Edgecombe Co.], North Carolina; figured holotype USNM 85905). Lea, 1834, Obs. Unio, 1: 8. Unio angustatus Lea 1831, Trans. Amer. Philos. Soc, 4: 114, pi. 17, fig. 43 (Cooper River, South Carolina; figured holotype USNM 85896; Congaree River, South Carolina). Lea, 1834, Obs. Unio, 1: 124. Unio productus Conrad 1836, Monography Union- idae, no. 3, p. 31, pi. 14, fig. 1 (Savannah River, Augusta [Richmond Co., Georgia]; figured holotype ANSP 41397). Unio fislicrianus Lea 1838, Trans. Amer. Philos. Soc, 6: 8, pi. 4, fig. 8 (Head of Chester River [Kent Co.], Maryland; figured holotype USNM 86022). Lea, 1838, Obs. Unio, 2: 8. Unio foliculatus Lea 1838, Trans. Amer. Philos. Soc, 6: 38, pi. 11, fig. 33 (Savannah River, [Georgia]; figured holotype USNM 85861). Lea, 1838, Obs. Unio, 2: 38. Unio duttonianns Lea 1841, Proc Amer. Philos. Soc, 2: 31 (Ogeechee Canal, Savannah [Chat- ham Co.], Georgia). Lea, 1842, Trans. Amer. Philos. Soc, 8: 236, pi. 22, fig. 50; type not in USNM [lost]. Lea, 1842, Obs. Unio, 3: 74. Unio sagittiformis Lea 1852, Trans. Amer. Philos. Soc, 10: 277, pi. 22, fig. 35 (Oconee River, near Athens [Clarke Co.], Georgia; figured holotype USNM 85970). Lea, 1852, Obs. Unio, 5: 33. Unio rostraeformis Lea 1856, Proc. Acad. Nat. Sci. Phila., 8: 262 (Swift Creek, near Macon [Bibb Co.], Georgia). Changed to: Unio rostriformis Lea 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 64, pi. 10, fig. 46; probable figured holotype USNM 85973. Lea, 1858, Obs. Unio, 6: 64. Unio extensus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 31 (Dry Creek, near Columbus [Muscogee Co.], Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 68, pi. 12, fig. 49; figured holotype USNM 85995. Lea, 1858, Obs. Unio, 6: 67. Unio emmonsii Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 86 (Roanoke River, Weldon [Hali- fax Co.], North Carolina). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 56, pi. 2, fig. 203; figured holotype USNM 86028. Lea, 1862, Obs. Unio, 8: 60. Unio naviculoid.es Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 170 (Buckhead Creek, Burke Co.; Macon [Bibb Co.]; both Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 94, pi. 20, fig. 74; figured holotype USNM 86005, from Buckhead Creek, Lea, 1858, Obs. Unio, 6: 94. Unio maconensis Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 172 (Flint River, near Macon [Co.], Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 93, pi. 20, fig. 73; figured holotype USNM 86004. Lea, 1858, Obs. Unio, 6: 93. Unio hazelhurstianus Lea 1858, Proc. Acad. Nat. Sci. Phila., 9: 166 (Satilla River, Camden Co., Georgia). Lea, 1859, Tour. Acad. Nat. Sci. Phila., ser. 2, 4: 211, pi. 26, fig. 92; figured holotype USNM 86009. Lea, 1859, Obs. Unio, 7: 29. Unio viridulus Lea 1863, Proc Acad. Nat. Sci. Phila., 15: 193 (Neuse River, near [6 miles E of] Raleigh [Wake Co.], North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 334 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 10, pi. 3, fig. 7; figured holotype USNM 85908. Lea, 1867, Obs. Unio, 11: 14. Unio nasutilus Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 192 (Livingston's Creek, Brunswick- Co. [NE corner of Columbus Co., Cape Fear River drainage], North Carolina). Changed to: Unio nasutulus Lea 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 12, pi. 3, fig. 9; figured holo- tvpe USNM 86027. Lea, 1867, Obs. Unio, 11: 16. Unio attenuatus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 157 (Savannah River; Beaver Creek [few miles E of Bonaire, Ocmulgee River drainage], Houston Co.; both Georgia). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 41, pi. 14, fig. 38; figured holotype USNM 85974, from Beaver Creek. Lea, 1874, Obs. Unio, 13: 45. Unio rostellum Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 160 (Macon [Bibb Co.], Georgia). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 44, pi. 15, fig. 41; figured holotype USNM 85972. Lea, 1874, Obs. Unio, 13: 48. Unio exactus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 159 (Savannah River at Governor Hamilton's [Mansion, 8 mi. S of Hardeeville, Jasper Co., South Carolina], Georgia). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 45, pi. 15, fig. 43; figured holotype USNM 85872. Lea, 1874, Obs. Unio, 13: 49. Unio subcylindraceus Lea 1873, Proc. Acad. Nat. Sci. Phila., 25: 422 (Rocky Creek, near Macon [Bibb Co.]; Carter's Creek [Baldwin Co.]; both Georgia). Lea, 1874, lour. Acad. Nat. Sci. Phila., ser. 2, 8: 58, pi. 20, fig. 57; figured holotype USNM 85863, from Rocky Creek. Lea, 1874, Obs. Unio, 13: 62. Unio arctior fisheropsis De Gregorio 1914, II Naturalista Siciliano, 22: 45, pi. 5, figs. 3a-c (Ogeecho [Ogeechee] Canal [Chatham Co.], Georgia; type Palermo Mus., Sicily [not seen]). Elliptio cupreus (Rafinesque). Ortmann, 1919, Mem. Carnegie Mus., 8: 110, pi. 8, fig. 6. Elliptio fisherianus (Lea). Ortmann, 1919, Mem. Carnegie Mus., 8: 113, pi. 8, fig. 7. Description. Shell generally medium throughout most of its range, usually not exceeding 100 mm in length, though reach- ing over 130 mm in a few localities. Out- line elongated trapezoid, or elongated ovate, lance-head shaped, a little over twice as long as high. Valves usually subcom- pressed, thin to subsolid, occasionally quite solid; inequilateral. Anterior end regularly rounded; posterior end elongated and either sharply or bluntly pointed. The posterior point may occasionally be sharp and turned up above the medial line; sometimes it drops below the medial line, but more generally it is evenly and broadly pointed. Ventral margin gently curved, sometimes almost straight, or a bit arcuate. Dorsal margin straight, forming an indis- tinct angle with the obliquely descending posterior slope. Hinge ligament rather long and low. Posterior ridge rather broadly rounded or subangular, generally with a secondary ridge above it; the ridges sometimes cause the point to be slightly biangulate. Umbos very low, located in the anterior fourth of the shell, their sculpture consisting of several corrugated, longi- tudinal ridges. Disk rather flat, or just slightly concave caused by a slight um- bonal-ventral sulcus. Surface of the shell with numerous concentric striae. Perios- tracum fine, subshiny, greenish yellow, olive, often with distinct greenish rays, esrjecially when young, but often becoming rough and brownish or dirty black with age. Left valve with two stumpy pseudo- cardinal teeth, one in front of the other, both somewhat triangular; the hinder one may be vestigial. Hinge line rather short and narrow; two long, straight lateral teeth. Right valve with two roughly parallel pseudocardinals, the posterior one inclined to be triangular, serrated and chunky, the more anterior one, low and vestigial; one lateral tooth. Beak cavities very shallow, with a few dorsal muscle scars. Anterior and posterior adductor muscle scars and pallial line all distinct. Nacre white, bluish white, pinkish, purplish, posteriorly iridescent. Length Height Width mm mm mm 34 54 29 Ochee Creek, about 5 mi. SE Toomsboro, Wilkinson Co., Georgia. 99 42 20 Mill Race, 2 mi. N Sardis, Burke Co., Georgia. 74 28 14 As above. 68 22 12 As above. Atlantic Sloik Umonidak • Johnson 335 Anatomy. Discussed and figured by Reardon (1929: 11, pi 5, figs. 1-10). The glochidia are hookless. Breeding season. Ortmann (1919: 111) found gravid females on May 6, June 3-8, 1912, in the Potomac and James Rivers, Virginia. Habitat. Lives on sandy bottoms, often found crawling about with much of the shell out of the sand, also found among rocks and in mud, where the current is not too swift. Remarks. Elliptio lanceolata (Lea), which ranges from the Escambia River system of Florida to the Juniata River of the Susquehanna River system of Penn- sylvania, is allied to Elliptio icterina (Con- rad) and complanata (Lightfoot) of the Apalachicolan and Atlantic Slope regions. While complanata has a subtrapezoidal shape, lanceolata is generally easily dis- tinguished from it by its more elongate, lanceolate shape. Even though the pos- terior point may be somewhat rounded, it can usually be separated from complanata or icterina because it is over twice as long as high and because the dorsal and ventral margins are roughly parallel. On the Atlantic Slope, E. lanceolata is most easily confused with Ligumia nasuta (Say), which ranges from the James River. Virginia, to the St. Lawrence River, Canada. Anatomically they are quite dif- ferent, and the shells of nasuta show sexual dimorphism and a pattern of green rays that is not found in lanceolata. The male shells can be confused, but nasuta has a silvery white or cream colored nacre, whereas lanceolata may be coppery or purple, with more stumpy hinge teeth. The females of nasuta have a peculiar expansion of the lower margin, rendering them easy to distinguish. Ortmann (1919: 113) separated fisheri- anus from lanceolata with some reluctance, on the basis that fisherianus was slightly more elongated, with a greater taper to the posterior part of the shell. A few popu- lations of this ecophenotype are found in the northern range of the species and again in the ponds near Wilmington, New Han- over Co., North Carolina. Reardon (1929: 9) notes that there is no constant distinction either in the anat- omy or in the shell morphology. Except for their large size, specimens from Ochee Creek, about 5 mi. SE Toomsboro, Wilkin- son Co., Georgia, show the same range of variation as do specimens from the Potomac River, Virginia. There is much local variation in this species throughout its range, and it is easy to see how Issac Lea fell to naming all of the shell forms. The specimens from the type locality and from several other of the rivers in North Carolina are usually small, shiny, and yellowish, without a trace of rays, whereas over the rest of its range, lanceolata generally has greenish rays. The species varies considerably as to the ratio of height to length, the sharpness of the posterior point, whether the point is above or below the medial line, and even to the degree of inflation. It can generally be easily separated from any other species of Elliptio because it is more than twice as long as high, with the dorsal and ventral margins roughly parallel. Also found on the Southern Atlantic Slope is Elliptio shepardiana (Lea), which is restricted to the Altamaha River system, Georgia, and although it is obviously closely related to lanceolata, it differs from lanceolata by being over three times as long as high. In the Suwannee River system of the Apalachicolan region and in Peninsular Florida, E. lanceolata is replaced by a closely related species, Elliptio jayensis (Lea). While jayensis exhibits a range of variation similar to that of lanceolata, it is generally smaller, with a tendency to be higher behind. The dorsal and ventral margins are not parallel, and it has more numerous darker green rays. In the Apalachicolan region, it is often difficult to separate lanceolata from Ellip- tio icterina (Conrad), which also some- 336 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 times tends to have a lanceolate shell, but even so, these strikingly long shells can generally be separated from any other Elliptic) by their height to length ratio. Elliptic) lanceolata (Lea) seems to be most abundant on the Atlantic Slope of Georgia. Here and in the Apalachicolan region, individuals reach their greatest size. Specimens from the Carolinas and the north are generally smaller and more deli- cate. Range. Apalachicolan region: Escambia River system, east to the Apalachicola River system, but discontinuous; Satilla River system, Georgia. Atlantic Slope: Altamaha River system, Georgia, north to the Juniata River of the Susquehanna River system, Pennsylvania. Specimens Examined Satilla River System Satilla River Drainage. Georgia: Sa- tilla River, Camden Co. Altamaha River System Ocmulgee River Drainage. Georgia: Walnut Creek, 2 mi. E Macon; Rockv Creek (USNM); Swift Creek, 4 mi. E Macon; Tobesofkee Creek, 5 mi. S Macon; all Ribb Co. Reaver Creek, E Ronaire (USNM); Mossy Creek, 2 mi. SW Kath- leen; Flat Creek, 2 mi. S Perry; all Houston Co. Ruck Creek, 5 mi. NW Hawkins ville; Limestone Creek, 4.3 mi. E Hawkinsville; Rig Tuscawhatchee Creek, 5 mi. SW Hawkinsville; Cedar Creek, Fountains Mill, 7 mi. SW Hawkinsville; Mosquito Creek, 8 mi. SE Hawkinsville; all Pulaski Co. House Creek, Rowens Mill, 9 mi. N Fitz- gerald; Dicksons Creek, 10 mi. NE Fitz- gerald; both Ren Hill Co. Little Ocmulgee River Drainage. Geor- gia: Little Ocmulgee River, 7 mi. NE Cochran, Reckley Co. Gum Swamp Creek, 1 mi. N McRae, Telfair Co. Oconee River Drainage. Georgia: Oco- nee River, near Athens, Clarke Co. (USNM). Carters Creek, Raldwin Co. (USNM). Napiers Creek, 6 mi. N Mc- Intyre; Ochee Creek, about 5 mi. SE Toomsboro; Turkey Creek, 4 mi. NE Allentown; all Wilkinson Co. Ford Branch, 4 mi. W Dublin; Turkey Creek, 9 mi. W Dublin; Rocky Creek, 8 mi. W Dudley; all Laurens Co. Ochwalkee Creek, 2 mi. E Glenwood, Wheeler Co. Ohoopce River Drainage. Georgia: Ohoopee River, 4 mi. S Wrightsville, John- son Co. Mulepen Creek, 2 mi. E Adrian; Ohoopee River, Norristown; both Emanuel Co. Little Ohoopee River Drainage. Geor- gia: Battleground Creek, 1 mi. N Kite, Johnson Co. Ogeechee River System Ogeeehee River Drainage. Georgia: [Williamson Swamp Creek] Rartow; Nails Creek, 2 mi. S Bartow; both Jefferson Co. Bark Camp Creek, 7 mi. E Midville; Mill Creek, 5 mi. E Midville; Buckhead Creek, 9 mi. S Waynesboro; all Burke Co. Ogee- chee River, bridge 1.5 mi. SW Oliver, Screven Co. (MZUM). Ogeechee River, bridge, 1 mi. E Blitchton; Ogeechee River, Jinks Bridge [Rte. 16]; Dolly Lake, about 1.5 mi. below Kiterlighter Camp; all Bryan Co. (all MZUM). Savannah River System Savannah River Drainage. Georgia: Savannah River, Augusta, Richmond Co. (ANSP). Brier Creek, 7.5 mi. NE Wrens, Jefferson Co. Savannah River, 7.5 mi. NE Shell Bluff; Brier Creek, Chalker Bridge, 6 mi. N Waynesboro; Mill Race, 2 mi. N Sardis; all Burke Co. South Carolina: Savannah River, 2 mi. SW Millettville ( MZUM ) ; Savannah River, Johnsons Land- ing, 10 mi. W Allendale; both Allendale Co. Georgia: Savannah River, 8 mi. S Hardeville, Jasper Co. (USNM). Rice fields, near Savannah; Ogeechee Canal, Savannah (Lea); both Chatham Co. Edisto River System Edisto River Drainage. South Carolina: Edisto River [not localized]. Atlantic Slope Unionidae • Johnson 337 Cooper-Santee River System Cooper River Drainage. Sontli Caro- lina: Cooper River [not localized]. Saluda River Drainage. South Caro- lina: Saluda River, 2.3 mi. above Ware Shoals, Abbeville Co. Congaree River Drainage. South Caro- lina: Congaree River, Richland Co. Wateree River Drainage. South Caro- lina: Wateree River, 2.5 mi. W Camden, Kershaw Co. Black River System Blaek River Drainage. South Carolina: Green Swamp Creek, 2.5 mi. W Sumter; Poeotaligo River, 3 mi. S Sumter; both Sumter Co. Waccamaw River System Waccamaw River Drainage. South Carolina: Lake Waccamaw, [town of] Lake Waccamaw, Columbus Co. Cape Fear River System Cape Fear River Drainage. North Caro- lina: Livingstons Creek, Brunswick Co. (USNM). Wilmington, New Hanover Co. Neuse River System Neuse River Drainage. North Carolina: Little River, Tarpleys Mill, 2 mi. NE Wendell (5); Little River, 2 mi. WSW Zebulon (6); both Wake Co. Neuse River, 6 mi. E Raleigh (USNM). Pamlico River System Tar River Drainage. North Carolina: Tar River, Tarboro (USNM); Tar River, Old Sparta, 3.5 mi. W Pinetops; both Edgecombe Co. Roanoke Rpver System Dan River Drainage. Virginia: Dan River [Halifax Co.]. Roanoke River Drainage. Virginia: Roanoke River, Weldon, Halifax Co. Chowan River System Nottoway River Drainage. Virginia: Nottoway River, 3 mi. E Rawlings, Bruns- wick Co. James River System James River Drainage. Virginia: Calf- pasture River (Conrad, 1846: 407); North [ = Maury] River, Lexington; [James River], near Natural Bridge; all Rockbridge Co. James River, Buchanan, Botetourt Co. James River, Lynchburg, Campbell Co. Rivanna River, 2 mi. W Columbia, Flu- vanna Co. James River, opposite Maidens, Goochland Co. York River System South Anna River Drainage. Virginia: South Anna River, 1 mi. N. Dabneys, Louisa Co. South Anna River, Gum Tree, Hanover Co. Rappahannock River System Rapidan River Drainage. Virginia: Rapidan River, N Unionville, Orange Co. Branch, N Lignum, Culpeper Co. Rappahannoek River Drainage. Vir- ginia: Rappahannock River, Remington, Faquier Co. Rappahannock River, Water- loo, Culpeper Co. Occoquan Creek System Oeeoquan Creek Drainage. Virginia: Broad Run [Fairfax Co.] 3 mi. W Manas- sas, Prince William Co. Potomac River System Potomac River Drainage. West Vir- ginia: Back Creek, 0.5 mi. E Glengary, Berkeley Co. Shenandoah River, Jefferson Co. Virginia: N Fork Shenandoah River, E Woodstock, Shenandoah Co. Pennsyl- vania: Rock Creek, 0.75 mi. N Gettysburg, Adams Co. Maryland: Aqueduct Lake Washington Co. Flat Run, Emmitsburg; Monocacv River, 2 mi. W Pleasant; both Frederick Co. Potomac River, Great Falls; 338 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Potomac River, below Alexandria; both Fairfax Co. Elliptio (Elliptio) shepardiana (Lea) Plate 11:7 Unio shepardianus Lea 1834, Trans. Amer. Philos. Soc, 5: 95, pi. 13, fig. 38 [Altamaha River] Hopeton, near Darien [Mcintosh Co.], Georgia; type, not in USNM [lost]; three presumed para- types MCZ 155568 ex Prof. Shepard. Lea, 1834, Obs. Unio, 1: 207. Simpson, 1914, Cat. Naiades, 2: 698. Description. Shell often large, reaching over 185 mm in length. Outline extremely elongate, trapezoidal, lance-head shaped, over three times as long as high. Valves subeompressed to subinflated, subsolid, quite inequilateral. Anterior end regularly rounded, usually higher than any other part of the shell; posterior end very elongated and pointed. The point may be at the medial line or below it. Ventral margin almost always arcuate. Dorsal mar- gin roughly parallel to the ventral margin, long and straight, forming an indistinct angle with the obliquely descending pos- terior margin. Hinge ligament long and low. Posterior ridge rather broadly rounded or subangular, generally with a secondary ridge above it; the ridges some- times cause the point to be slightly bi- angulate. Umbos very low, located in the anterior fifth of the shell, their sculpture consisting of several corrugated, longi- tudinal ridges. Disk surface flat or just slightly concave caused by a slight um- bonal- ventral sulcus. Surface of the shell with numerous concentric striae. Perios- tracum fine, greenish or yellowish, with faint rays when young, becoming rough and brownish or blackish with age. Left valve with two stumpy pseudo- cardinal teeth, one in front of the other, both somewhat triangular, the hinder one less high. Hinge line rather short and narrow; two long straight lateral teeth. Right valve with two roughly parallel pseudocardinals, the posterior one inclined to be triangular, serrated, and chunky, the more anterior one low and vestigial; one lateral tooth. Beak cavities very shallow, with a few dorsal muscle scars. Anterior and posterior adductor muscle scars and pallial line all distinct. Nacre sometimes bluish white or pinkish but usually pur- plish or coppery, iridescent posteriorly. Length Height Width mm mm mm 185 127 42 29 26 15 Altamaha River, 3 mi. NW Everett City, Glynn Co., Georgia. Ocmulgee River, 1 mi. S Lumber City, Telfair Co., Georgia. Anatomy. Discussed by Lea (1863: 415). Habitat. Lives in muddy banks, or buried in sand in flowing water. Remarks. Elliptio shepardiana (Lea) is a singular species and it can be confused with no other one. It is obviously related to E. lanceolata (Lea), but differs from it by being over three times as long as high, whereas lanceolata is just over twice as long as high. In shepardiana the an- terior end is the widest part of the shell; it is almost alwavs arcuate, and is con- sistently pointed. It replaces E. lanceolata in the main tributaries of the Altamaha River. Range. Southern Atlantic Slope: re- stricted to the Altamaha River system, Georgia. Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Ocmulgee River above Jackson, Butts Co. Big Tucsawhatchee Creek, 6 mi. S Hawk- insville, Pulaski Co. Ocmulgee River, Jacksonville; Ocmulgee River, 1 mi. S Lumber City; both Telfair Co. Oconee River Milledgeville, Baldwin Co. Oconee River, 2.5 mi. N Glenwood; Ochwalkee Creek, 2 mi. E Glenwood; both Wheeler Co. Altamaha River Drainage. Georgia: Altamaha River, 7 mi. N Hazlehurst, Mont- gomery Co. Altamaha River, 10 mi. NE Atlantic Slope Umonidae • Johnson 339 Surrency, Appling Co. Altamaha River, 11 mi. N Odum; Altamaha River, "River- side Park," 4 mi. N Jesup; both Wayne Co. Altamaha River, 3 mi. NE Everett City, Glynn Co. Altamaha River, near Fort Rar- rington; Penhollaway Creek, 5 mi. S Fort Rarrington; Altamaha River, Hopeton, near Darien; all Mcintosh Co. Genus Uniomerus Conrad Uniomcrus Conrad 1853, Proc. Acad. Nat. Sci. Phila., 6: 268. Species listed: U. dcclivis Say, camptodon Say, suhcroccus Conrad, sayii Ward, rivularis Conrad, porrectus Conrad, sym- metricus Lea, excultus Conrad. Conrad, 1854, Jour. Acad. Nat. Sci. Phila., ser. 2, 2: 296. Type species, Unio tetralasmus Say. Subsequent designation, Simpson, 1900, Proc. United States Natl. Mus., 22: 739. Since U. excultus Conrad is included in Simpson's synonymy of U. tetra- lasmus, the subsequent selection of the former by Clench and Turner, 1956, Bull. Florida State Mus., 1 : 176, is invalid under Article 69 ( a ) ( iv) of the Int. Code Zool. Nomen. ( 1964 ). Ort- mann, 1912a, Ann. Carnegie Mus., 8: 272. Frierson (1927: 34-35) lists a number of species and subspecies under Uniomerus. Like Elliptio, Uniomerus has a wide range of environmental tolerance, and while there are a number of ecophenotypes, the genus appears to be monotypic. Uniomerus fefralasmus (Say) Plate 12: 1-6 Unio tetralasmus Say, [September] 1831, American Conchology, no. 3 [no pagination] pi. 23 (Bayou St. John [not located] near New Or- leans, Louisiana; type not in ANSP [lost]). Unio obesus Lea 1831, Trans. Amer. Philos. Soc, 4: 96, 108, pi. 13, fig. 26 (York River, Virginia [corrected to Georgia; Maj. Leconte on p. 108]; figured holotype USNM 85366, labeled, "Little Ogeechee River [Hancock Co.], Georgia; Maj. Leconte"). Lea, 1834, Obs. Unio, 1: 106, 118. Clench and Turner, 1956, Bull. Florida State Mus., 1: 178, did not see the type, which has the corrected locality written on it, and un- fortunately restricted the type locality to Co- lumbus, Georgia. Lea, 1854 (Proc. Acad. Nat. Sci. Phila., 7: 243), claimed that this description appeared during the latter part of 1831. It was reported on in 1832, (Jan. -March number of: Amer. Jour. Sci., 22: 169 [probably appeared in April]). There is no way to be sure which name has priority, l)i it Say's name is certainly better known, and it is given primary status here without equi- vocation. Unio dcclivis Say 1831 [1832], Transylvania Jour. Med., 4: 527 (Bayou Teche, Louisiana). Say, 1832, American Conchology, no. 4 [no pagi- nation], pi. 35; three syntypes ANSP 41698 from Mr. Barabino, all smaller than figured type. Unio camptodon Say 1832, American Conchology, no. 5 [no pagination], pi. 42 (opposite to New Orleans [Jefferson Parish, Louisiana], in ponds; type not in ANSP [lost]). Unio geometricus Lea 1832, Trans. Amer. Philos. Soc, 5: 38, pi. 4, fig. 10 (Bayou Teche, Louisiana; figured holotype USNM 85712) Lea, 1834, Obs. Unio, 1: 150. Unio bland in gianns Lea 1834, Trans. Amer. Philos. Soc, 5: 101, pi. 15, fig. 44 (St. Johns River, Florida; figured holotype USNM 85715). Lea, 1834, Obs. Unio, 1: 213. Unio declivis Conrad 1836 non Say. See under: Unio rivularis Conrad, 1853. Unio excultus Conrad, 1838, Monography Unioni- dae, no. 11, p. 99, pi. 55, fig. 1 (New Orleans [Orleans Parish], Louisiana; type ANSP 20427 [lost]). Unio sayii Ward 1839, [in Tappan], Amer. Jour. Sci., 35: 268, pi. 3, fig. 1 (Walnut Creek" and Ohio Canal, near Circleville [Pickaway Co.], Ohio; type [location unknown]). Unio paralellus [sic] Conrad 1841, Proc. Acad. Nat. Sci. Phila., 1: 20 non Sowerby 1840. Changed to: Unio porrectus Conrad 1845, Jour. Acad. Nat. Sci. Phila., ser. 2, 2: 296, pi. 26, fig. 7 ([Pearl River] Jackson [Hinds Co.], Mississippi; figured holotype ANSP 42847). Unio buddianus Lea 1843, Desc. Twelve Uniones (Lake George and Lake Monroe, Florida). Lea, 1845, Trans. Amer. Philos. Soc, 9: 277, pi. 40, fig. 5; figured holotype USNM 85606, from Lake George. Lea, 1848, Obs. Unio, 4: 35. Unio symmetricus Lea 1845, Trans. Amer. Philos. Soc, 10: 73, pi. 4, fig. 11 (Red River, Alex- andria [Rapides Parish] Louisiana; figured holotype USNM 85604). Lea, 1848, Obs. Unio, 4: 47. Unio paludicolus Gould 1845, Proc. Boston Soc. Nat. Hist., 2: 53 (Florida Everglades; lecto- type MCZ 169278, selected by Johnson, 1964, Bull. United States Natl. Mus., 239: 121, pi. 31, fig. 3). Unio ineptus Lea 1852, Trans. Amer. Philos. Soc, 10: 261, pi. 15, fig. 12 (Abbeville District [Savannah River drainage], South Carolina; 340 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 figured holotype USNM 85326). Lea, 1852, Obs. Unio, 5: 17. Unio hebes Lea 1852, Trans. Amer. Philos. Soc, 10: 267, pi. 18, fig. 21 (Oconee River, near Athens [Clarke Co.], Ceorgia; figured holotype USNM 85383). Lea, 1852, Obs. Unio, 5: 23. Unio rivularis Conrad 1853, Proc. Acad. Nat. Sci. Phila., 6: 257. New name for Unio declicis Conrad 1836 non Say 1831, in Monography Unionidae, no. 5, p. 45, pi. 23, fig. 1 (small creek in Green Co., Alabama; figured holotype ANSP 42852). Unio paludicolor Conrad 1853, Proc. Acad. Nat. Sci. Phila., 6: 254. Error for U. paludicolus Gould. Unio subcroceus Conrad 1854, Jour. Acad. Nat. Sci. Phila., ser. 2, 2: 297, pi." 27, fig. 1 (one of the tributaries to Canadian River, Arkansas; type not in ANSP [lost]). Unio manubius Could 1855, Proc. Boston Soc. Nat. Hist., 5: 229 (Chihuahua, 60 mi. from Camp Ringgold = Rio Agualeguas, 3 mi. NE General Trevino, Nuevo Leon [State, Mexico], teste Taylor, 1967, Veliger, 10: 154; holotype MCZ 169447, figured by Johnson, 1964, Bull. United States Natl. Mus., 239: 108, pi. 32, fig. 5). Unio columbensis Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 31 (Creeks near Columbus [Mus- cogee Co.], Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 75, pi. 14, fig. 55; figured holotype USNM 85360. Lea, 1858, Obs. Unio, 6: 75. Unio jamesianus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 84 ([Pearl River] Jackson [Hinds Co.], Mississippi). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 53, pi. 6, fig. 35; figured holo- type USNM 85365). Lea, 1858, Obs. Unio, 6: 52. Unio plantii Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 171 (Flint River, near Macon [Co.], Georgia). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 192, pi. 21, fig. 76; figured holotype USNM 85005. Lea, 1859, Obs. Unio, 7: 10. [Known only from the holotype, which is a pathological specimen]. Unio cicur Lea 1861, Proc. Acad. Nat. Sci. Phila., 13: 39 (Little Ocmulgee River, Georgia). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 93, pi. 13, fig. 241; figured holotype USNM 85532. Lea, 1862, Obs. Unio, 8: 97. Unio squalidus Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 192 ( Neuse River, near Raleigh (Wake Co.]; Roanoke River, near Weldon [Halifax Co.]; Deep River; all North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 22, pi. 7, fig. 20; figured holotype USNM 85376, from Roanoke River. Lea, 1867, Obs. Unio, 11: 26. Unio electrinus Reeve 1865, Conch. Iconica, 16, Unio, pi. 25, fig. 121 (Hab. ? Cuming colln; type, British Museum (Nat. Hist.) [lost]). Unio bisselianus Lea 1867, Proc. Acad. Nat. Sci. Phila., 15: 81 (Bissels Pond, Charlotte [Meck- lenburg Co.], North Carolina). Lea, 1868, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 277, pi. 37, fig. 90; figured holotype USNM 85373. Lea, 1869, Obs. Unio, 12: 37. Unio jewettii Lea 1867, Proc. Acad. Nat. Sci. Phila., 11: 81 (sink of Noonan's [Newnans] Lake [Alachua Co.], Florida). Lea, 1868, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 276, pi. 37, fig. 89; figured holotype USNM 85374. Lea, 1869, Obs. Unio, 12: 36. Unio rivicolus Conrad 1868, Amer. Jour. Conch., 4: 280, pi. 18, fig. 4 (brook near Tampa [Hillsborough Co.], Florida; figured holotype ANSP 41411). Unio pawensis Lea 1868, Proc. Acad. Nat. Sci. Phila., 20: 161 (Paw Creek [Mecklenburg Co.], Beaver Co. [ = Creek, Gaston Co.], Catawba Run [Gaston and Mecklenburg Cos.]). Lea, 1868, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 302, pi. 45, fig. 114; figured holotype USNM 85380, labeled, "Beaver Creek, [into ?] Catawba Run, North Carolina." Lea, 1869, Obs. Unio, 12: 62. Uniomerus obesus (Lea). Clench and Turner, 1956, Bull. Florida State Mus, 1: 177, pi. 5, fig. 2. Description, Shell medium to large, reaching 114 mm in length. Outline rhomboid or long rhomboid. Valves sub- inflated or inflated, subsolid. Anterior end regularly rounded or slightly truncated; posterior end usually somewhat produced. Ventral margin slightly incurved. Dorsal margin slightly curved, generally forming a sharp angle with the almost straight posterior margin. Hinge ligament long and narrow, located posteriorly of the umbos. Posterior ridge rounded, ending in a point or feeble biangulation at the base of the shell, sometimes rendering older specimens a bit arcuate. Posterior slope often with two radial sulci. Umbos low to slightly elevated, located in the an- terior quarter of the shell, their sculpture consisting of five or six heavy ridges that form a rounded angle on the posterior ridge, in front of which they tend to be corrugated. Periostracum generally black and slightly roughened, but with a satiny Atlantic Slope Unionxdae • Johnson 341 sheen over most of the surface. Sometimes the surface is smooth and shiny, especially in the umbonal area, and may then be brownish yellow or yellowish mixed with green, not rayed. Left valve with two ragged, subequal pseudoeardinal teeth and two straight lateral teeth. Right valve with one tri- angular pseudoeardinal often with a vesti- gial tooth above it; one lateral tooth. Beak cavities compressed, but with several muscle scars; anterior adductor muscle scars deep, posterior ones faint. Pallial line distinct. Nacre white, bluish white or pinkish to lurid purple. Length Height Width 114 102 70 61 59 40 39 36 29 Ogeeehee River, Bartow, Jefferson Co., Georgia. Little Cedar Creek, 1 mi. E Wrightsville, Jefferson Co., Georgia. Richardson Creek, 2 mi. NE Thrift, Jenkins Co., Georgia. Anatomy. Discussed by Lea (1863: 404). Habitat. Generally lives in smaller streams and ponds on muddy bottoms, where it may be locally abundant. Some- times found in sand with Elliptic com- planata (Lightfoot) in larger Southern Atlantic Slope rivers, but in this habitat it is relatively scarce. Remarks. In the Apalachicolan and Southern Atlantic Slope regions, Uniomerus tetralasmus (Say) can be confused with both Elliptio complanata (Lightfoot) and E. icterina (Conrad). In general, tetralas- mits is more inflated than either of them, is proportionately higher, more acutely angu- lar where the dorsal margin meets the posterior one, and very often has a satiny periostracum. The yellowish brown, un- rayed periostracum and the bluish white or pinkish nacre of tetralasmus are distinct from the brown, often rayed periostracum and the dark purplish bronze nacre of complanata. The periostracum of icterina is sometimes bright yellow or chestnut, but like complanata, is generally rayed, though often but faintly, whereas tetra- lasmits scarcely ever appears rayed. Uniomerus tetralasmus (Say) is generally common and abundant throughout the Apalachicolan region, Peninsular Florida, and in the Southern Atlantic Slope rivers of Georgia. It becomes noticeably scarce in the Carolinas. The periostracum is more inclined to be smooth on individuals from this northern end of the range. Range. Interior Basin: Mississippi drain- age generally, north to about latitude 40 degrees, Ohio River. West Gulf Coastal region, Alabama-Coosa River system, and Apalachicolan region: Rio Grande River system, Texas, east to the Suwannee River system, Florida. Peninsular Florida. South- ern Atlantic Slope: Altamaha River system, north to the Nottoway River of the Chowan River svstem, North Carolina. Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Swift Creek, SE Macon, Bibb Co. Mossy Creek, 3.5 mi. SW Kathleen, Houston Co. Tucsawhatchee [ = Big] Creek, 5 mi. SW Hawkinsville; Cedar Creek, Fountains Mill, 7 mi. WSW Hawkinsville; Mosquito Creek, 8 mi. SE Hawkinsville; all Pulaski Co. Brushy Creek, 8 mi. NW Abbeville, Wil- cox Co. House Creek, Bowens Mill, 9 mi. N Fitzgerald, Ben Hill Co. Little Ocmulgee River Drainage. Geor- gia: Little Ocmulgee River, 6 mi. NE Cochran, Bleckley Co. Alligator Creek, 2.5 mi. SW Alamo, Wheeler Co. Gum Swamp Creek, 1 mi. N McRae, Telfair Co. Oconee River Drainage. Georgia,: Oco- nee River, near Athens, Clarke Co. (USNM). Black Creek, 7 mi. N Mclntyre; Napiers Creek, 6 mi. N Mclntyre; Ochee Creek, about 5 mi. SE Toomsboro; all Wilkinson Co. Rocky Creek, 6 mi. E Dudley; Turkey Creek, 6 mi. W Dublin; both Laurens Co. Ochwalkee Creek, 2 mi. E Glenwood, Wheeler Co. 342 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Ohoopee River Drainage. Georgia: Ohoopee River, 4 mi. S Tennible, Wash- ington Co. Little Ohoopee River, 10 mi. NE Wrightsville; Little Cedar Creek, 1 mi. E Wrightsville; Big Cedar Creek, 1 mi. SE Wrightsville; outlet of Rollins Pond, 2 mi. SE Wrightsville; Battleground Creek, 1 mi. N Kite; Dry Creek, 9 mi. NW Adrian; all Johnson Co. Pendleton Creek, 8 mi. N Soperton, Treutlen Co. Mulepen Creek, 2 mi. E Adrian; Yam Grande Creek, 3 mi. W Swainsboro; Little Ohoopee River, 6 mi. SW Swainsboro; all Emanuel Co. Altamaha River Drainage. Georgia: Hopeton, near Darien, Mcintosh Co. Brunswick, Glynn Co. Ogeechee River System Canoochee River Drainage. Georgia: Rocky Creek, 1.5 mi. E Swainsboro; Canoo- chee River, 4 mi. E Swainsboro; Little Canoochee Creek, 6 mi. E Swainsboro; all Emanuel Co. Little Lotts Creek, 2 mi. SE Statesboro; Lotts Creek, 7 mi. SW Statesboro; both Bulloch Co. Ogeeehee River Drainage. Georgia: Little Ogeechee River [Hancock Co.] (USNM). Ogeechee River, Bartow; Nails Creek, 2 mi. S Bartow; Rockv Creek, 2 mi. S Wadley; all Jefferson Co. Mill Creek, 5 mi. E Midville; Barkcamp Creek, 7 mi. E Midville; Barkcamp Creek, 9 mi. SW Waynesboro; all Burke Co. Spring, Ogeechee River, 1 mi. S Millen; Richardson Creek, 2 mi. NE Thrift; both Jenkins Co. Ogeechee River, Flat Ford Landing (MZUM); Dolly Lake, about 1.5 mi. below Kiterlighter Camp (MZUM); both Bryan Co. Savannah River System Savannah River Drainage. Soatl} Caro- lina: Abbeville District (USNM), Turkey Creek, 7 mi. NW Edgefield, Edgefield Co. Georgia: Beaverdam Creek, 7 mi. NW Girard; outlet of Waynesboro Lake, Waynesboro; Mill Race, 2 mi. N Sardis; all Burke Co. Brier Creek, 6 mi. N Sylvania, Screven Co. Savannah River, Atomic En- ergy Plant, 25 mi. S Augusta (ANSP). South Carolina: Savannah River, 2 mi. SW Millettville (ANSP); Savannah River, Kingjaw Point, 10 mi. WSW Allendale (ANSP); both Allendale Co. Mouth of Vermezobre Creek, Savannah Wildlife Res- ervation, Jasper Co. (MZUM). Georgia: Pond near Savannah, Chatham Co. Combahee River System Salkehatchie River Drainage. South Carolina: Lemon Creek, 5 mi. W Smoaks, Colleton Co. Edisto River System Edisto River Drainage. South Caro- lina: 20 mi. W Charleston, Charleston Co. Edisto River [not localized]. Cooper-Santee River System Cooper River Drainage. South Caro- lina: Cooper River (USNM). Catawba River Drainage. North Caro- lina: Beaver Creek, Gaston Co. Paw Creek; Irwins Creek; Bissels Pond, Charlotte; Elias Pond, 10 mi. from Charlotte; all Mecklenburg Co. Wateree River Drainage. South Caro- lina: Wateree River, 2.5 mi. W Camden; Big Pine Tree Creek, 1.5 mi. S Camden; both Kershaw Co. Santee River Drainage. South Caro- lina: Green Swamp Creek, 2.5 mi. W Sumter, Sumter Co. Poplar Creek, Schu- lers Fish Pond, near Santee State Park, Orangeburg Co. Santee River, Cross, Berkeley Co. Black River System Blaek River Drainage. South Carolina: Cowpen Swamp, 11 mi. NE Sumter, Sum- ter Co. Pedee River System Lynches River Drainage. South Caro- lina: Lynches River, 2 mi. NE Bishopville, Lee Co. Yadkin River Drainage. North Caro- lina: Yadkin River. Atlantic Slope Unionidae • Johnson 343 Waccamaw River System Waccamaw River Drainage. North Carolina: Lake Waccamaw (USNM); drainage canal beside Lake Waccamaw, 1 mi. NNW Dupree Landing; both Colum- bus Co. Sontli Carolina: Waccamaw River, Wachesaw Landing, 2 mi. W Murrells Inlet, Georgetown Co. Cape Fear River System Deep River Drainage. North Carolina: Deep River (USNM). Cape Fear River Drainage. North Caro- lina: Rocky River, Chatham Co. 11 mi. N Sanford, Lee Co. Stream below Greenfield Mill Pond, Wilmington, New Hanover Co. (ANSP). Neuse River System Neuse River Drainage. North Caro- lina: Neuse River, near Raleigh, Wake Co. Pamlico River System Tar River Drainage. North Carolina: Sandy Creek, Franklin Co. Fishing Creek, 4 mi. E Leggett; Tar River, Tarboro; both Edgecombe Co. Roanoke River System Roanoke River Drainage. North Caro- lina: Roanoke River, near Weldon, Hali- fax Co. (USNM). Chowan RrvER System Nottoway River Drainage. Virginia: Nottoway River. Subfamily Anodontinae (Swainson 1840) Ortmann 1910 Genus Lasmigona Rafinesque Subgenus Lasmigona s.s. Rafinesque Lasmigona Rafinesque 1831, Continuation of Monog. Bivalve Shells of River Ohio (Phila.), p. 4. Species listed: Alasmidonta marginata Say, A. costata Rafinesque. Type species, Alasmidonta costata Rafinesque. Subsequent designation, Simpson 1900, Proc. United States Natl. Mus., 22: 664. Symphynota Simpson 1900, Proc. United States Natl. Mus., 22: 662. Type species, Symphynota compressa Lea, origi- nal designation, non Symphynota Lea 1829, the type species of which is Unio alata Say. Original designation. There is only one species in Lasmigona s. s., L. costata (Rafinesque), which is found in the Interior Basin and the upper St. Lawrence River system. In addition to the subgenera discussed here are Alas- minota Ortmann (Sidcularia Rafinesque is not identifiable, teste Ortmann and Walker, 1922: 36) and Pterosyna Rafines- que. They do not occur on the Atlantic Slope, but belong to the Interior Basin. Subgenus Platynaias Walker Symphynota Simpson 1900, Proc. United States Natl. Mus., 22: 662. [non] Lea. (See under Lasmigona, above.) Platynaias Walker 1918, Occ. Pap. Mus. Zool., Univ. Mich., 49: 2. Type species, Symphynota compressa Lea, origi- nal designation. Platynaias probably contains only two, probably allopatric species, L. compressa (Lea) (Ortmann 1919: 116, pi. 9, figs. 1, 2 as L. viridis Rafinesque), which is found in the Interior Basin, Hudson Bay; Upper Mississippi, Ohio, and St. Lawrence River systems from Saskatchewan and Nebraska to Vermont; extending into the Hudson River on the Atlantic Slope; and L. sub- viridis (Conrad), which is primarily a spe- cies of the Atlantic Slope, but which ex- tends into the Lake Ontario drainage. Most of the other taxa listed under this subgenus by Frierson (1927: 20, 21) are included elsewhere in this paper in synon- ymy, save for L. neglecta (Lea) and L. quadrata ( Lea ) which are out of the scope of this paper. Lasmigona (Platynaias) subviridis (Conrad) Plate 12:7-9 Plate 13: 1 Unio subviridis Conrad 1835, New Fresh Water Shells United States; appendix, p. 4, pi. 9, fig. 1 (Schuylkill River; Juniata River [Rlair Co.]; 344 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 figured holotype ANSP 2105 [lost]; creeks in Lancaster Co.; [all Pennsylvania]). Unio viridis Conrad 1836, Monography Unioni- dae, no. 4, p. 35, pi. 17, fig. 1, non Rafinesque 1820. Unio tappanianus Lea 1838, Trans. Amer. Philos. Soc, 6: 62, pi. 17, fig. 55 ( [Frankstown Branch], Juniata River, near Hollidaysburg [Blair Co.]; figured holotype USNM 85240; Schuylkill [River]; and a small stream near Lancaster; [all Pennsylvania] ). Lea, 1838, Obs. Unio, 2: 62. Unio hyalinus Lea 1845, Proc. Amer. Philos. Soc, 4: 164 ([James River drainage], Richmond [Henrico Co.], Virginia). Lea, 1848, Trans. Amer. Philos. Soc, 10: 69, pi. 2, fig. 4; figured holotype USNM 86131. Lea, 1848, Obs. Unio, 4: 43. Unio decoratus Lea 1852, Trans. Amer. Philos. Soc, 10: 257, pi. 13, fig. 6 (Abbeville District [Savannah River drainage] South Carolina; figured holotype USNM 83972). Lea, 1852, Obs. Unio, 4: 43. Unio pijgmacus Lea 1852, Proc. Amer. Philos. Soc, 5: 52 (Abbeville District [Savannah River drainage] South Carolina). Lea, 1852, Trans. Amer. Philos. Soc, 10: 262, pi. 15, fig. 14; figured holotype USNM 85240. Lea, 1852, Obs. Unio, 5: 18. Unio charlottensis Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 191 ( [Sugar Creek], near Char- lotte, Mecklenburg Co., North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 8, pi. 2, fig. 5; figured holotype USNM 85402. Lea, 1867, Obs. Unio, 11: 12. Unio pertenuis Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 193 (Neuse River, near [6 mi. E of] Raleigh [Wake Co.], North Carolina). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 8, pi. 2, fig. 4; figured holotype USNM 86139. Lea, 1867, Obs. Unio, 11: 12. Unio insolidus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 159 (Abbeville District [Savannah River drainage] South Carolina; Fredericksburg [Spotsylvania Co.], Virginia; Irwin's Creek, Mecklenburg Co., North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 40, pi. 13, fig. 37; figured holotype USNM 83974 from Irwin's Creek, Lea, 1874, Obs. Unio, 13: 44. Symphynota viridis (Conrad). Simpson, 1914, Cat. Naiades, 1: 484. Symphynota charlottensis (Lea). Simpson, 1914, Cat. Naiades, 1: 484. Symphynota decorata (Lea). Simpson, 1914, Cat. Naiades, 1 : 486. Unio pygmaeus (Lea). Simpson, 1914, Cat. Naiades, 2: 649. Lasmigona (Platynaias) sidiviridis (Conrad). Ortmann, 1919, Mem. Carnegie Mus., 8: 121, pi. 9, figs. 3, 4. Ortmann and Walker, 1922, Occ Pap. Mus. Zool. Univ. Michigan, no. 112, p. 35. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem. no. 367, p. 32, fig. 31. Description. Shell generally small, sel- dom reaching over 55 mm in length. Out- line subrhomboid or subovate. Valves subeompressed to slightly inflated, subsolid. Anterior end regularly rounded; posterior end more broadly rounded and biangulate near the base. Ventral margin almost straight or slightly curved. Dorsal margin straight, sometimes forming a winglike angle with the obliquely descending pos- terior margin. Hinge ligament prominent. Posterior ridge low, but double, ending near the base of the shell in a slight bi- angulation. Posterior slope sometimes slightly concave above the upper ridge. Umbos elevated above the hinge line, located anterior to the middle of the shell, their sculpture consisting of about four subnodulous bars, the first two subcon- centric, the others deeply double-looped, anterior loops rounded, posterior ones angular with a few radial threads behind. The disk is smooth or slightly roughened by growth rests; periostraeum yellowish or brownish green, occasionally olivaceous, subshining to dull, usually faintly rayed over the entire surface, darker on the pos- terior slope. Hinge teeth well developed but rather delicate. Left valve with two lamellate, pseudocardinals that are directed forward and almost parallel to the hinge with a rudimentary interdental projection which fits into an interdental groove in the right valve; two long, straight, thin, lateral teeth. Right valve with one long pseudoeardinal, directed forward, almost parallel to the hinge line; interdentum wide; one thin lateral tooth. Beak cavities shallow, con- taining dorsal muscle scars. Anterior ad- ductor muscle scars distinct, posterior ones less so. Pallia] line distinct anteriorly, where the shell may be slightly thickened. Atlantic Slope Unionidae • Johnson 345 Nacre bluish white and iridescent, some- times pinkish, especially toward the urn- bonal cavities. Length Height Width mm mm 60 37 19 30 18 11 116 56 38 117 59 29 Abbeville District [Savan- nah River drainage], South Carolina. Holotype of Unio decoratus Lea. As above. Holotype of Unio pygmaeus Lea. Pfeiffers Pond, Charlotte, Mecklenburg Co., North Carolina (ANSP). Lasmigona compressa (Lea). Small brook, Mid- dlebury, Addison Co., Ver- mont. Anatomy. Discussed by Ortmann (1911: 283). Glochidia are subtriangular, almost semicircular, with hooks. They measure 0.36 mm in length, 0.30 mm in height. The host fish is not known. Breeding season. This hermaphroditic- species is bradytictic, the breeding season beginning in August and lasting until June (Ortmann, 1919: 122). Habitat. Usually found in gravel or sandy bottoms in medium or small streams, as well as canals and the pond areas of streams; seems to avoid large rivers. Remarks. With its lamellate pseudo- cardinal teeth and characteristic rudiment- ary interdental projection in the left valve, Lasmigona subviridis (Conrad) is not con- fused with any other Atlantic Slope species, except in New York, where its distribution overlaps that of Lasmigona compressa (Lea) in the lower St. Lawrence and Hudson river systems. L. compressa, which is found abundantly in the Ohio- Mississippi drainage of the Interior Basin, has a rather heavy, compressed, trapezoidal shell, and the left valve has a well-de- veloped interdental projection. L. sub- viridis has almost invariably a smaller, thinner, more delicate shell which is more inflated (see comparative measurements, above), and the interdental projection is rudimentary. Ortmann (1919: 123) calls attention to some specimens of L. subviridis from Co- nococheague Creek, Greencastle, Franklin Co., Pennsylvania, remarkable for their large size. The largest one measured 63 mm in length, 38 mm in height, and 26 mm in width. Lea described a single giant specimen of this species under the name Unio charlottensis, from Sugar Creek, near Charlotte, Mecklenburg Co., North Caro- lina, part of the Catawba River drainage of the upper Cooper-Santee River system, which measured 112 mm in length. During the last century, a few additional giant specimens, as well as a number of smaller ones, were collected in Pfeiffers Pond, Charlotte, Mecklenburg Co., North Caro- lina, by C. M. Wheatley (ANSP and MCZ). The original label on the lot in the Academy of Natural Sciences of Phila- delphia (ca. 1900) states that, "the pond from which they were taken has long since been filled in." Ortmann (1913a: 371; 1919: 124) found L. subviridis to be especially abundant on the west side of the Alleghenian divide in the Greenbrier and New rivers of the up- per Kanawha drainage in Virginia and West Virginia, where he suggested this species originated. Range. Interior Basin: New and Green- brier rivers of the upper Kanawha River drainage, Virginia and West Virginia. At- lantic Slope: upper Savannah River system, South Carolina; north to the Hudson River system, westward through the Mohawk River and Erie [now New York State Barge] Canal to the Genesee River; all New York. Specimens Examined Savannah River System Savannah River Drainage. South Caro- lina: Abbeville District (USNM). Cooper-Santee River System Catawba River Drainage. North Caro- lina: Irwins Creek (MZUM); Sugar Creek, 346 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 near Charlotte (USNM); Pfeiffers Pond, Charlotte; all Mecklenburg Co. [Ten Mile Creek], Wolfsville [7 mi. W Monroe], Union Co. (MZUM). Pedee River System Yadkin River Drainage. North Caro- lina: Crosiers Branch, Cabarrus Co. (ANSP). Cape Fear River System Cape Fear River Drainage. North Caro- lina: Cape Fear River, Kinnon, Cumber- land Co. (MZUM). Neuse River System Neuse River Drainage. North Carolina: Flat River, 2.5 mi. S Mt. Tirzah, Person Co. (80). Flat River, 2.3 mi. ENE Rouge- mont, Durham Co. (78). Eno River, 1.75 mi. ESE Hillsboro Center, Orange Co. (117). Neuse River, 9.5 mi. ENE Durham, Durham Co. (62). Neuse River, 5 mi. NNW Bayleaf (56); Neuse River, Falls (50); Neuse River, 5.75 mi. E Raleigh (45); Neuse River, 3 mi. S Milburnie (33); Stirrup Iron Creek, 2 mi. NE Morrisville (43); Swift Creek, 3 mi. SSW Garner (20); Little River, Tarpleys Mill, 2 mi. NE Wendell (5); all Wake Co. Neuse River, 3.5 mi. NNE Smithfield, Johnston Co. (25). Pamlico River System Tar River Drainage. Nortli Carolina: Tar River, Providence, Granville Co. ( MZUM ) . Tar River, 2 mi. W Springhope, Nash Co. Roanoke River System Roanoke River Drainage. Virginia: Stone River, Clarksville, Mecklenburg Co. (MZUM). James River System Appomattox River Drainage. Virginia: Appomattox River, Petersburg, Dinwiddie Co. (MZUM). James River Drainage. Virginia: North River, Buena Vista, Rockbridge Co. (Ort- mann, 1919: 124). Rivanna River, 2 mi. W Columbia, Fluvanna Co. James River, opposite Maidens, Goochland Co. [James River], Richmond, Henrico Co. (Lea). York River System North Anna River Drainage. Virginia: Church River, Orange Co. (MZUM). Rappahannock River System Rapidan River Drainage. Virginia: Rapidan River, 3 mi. N Orange, Orange Co. Rapidan River, Rapidan, Culpeper Co. (Ortmann, 1919: 124). Rappahannock River Drainage. Vir- ginia: Rappahannock River, Remington, Fauquier Co. (Ortmann, 1919: 124). [Rap- pahannock River], Fredericksburg, Spot- sylvania Co. (Lea). Potomac River System Potomac River Drainage. West Vir- ginia: Sleepy Creek, 8 mi. E Berkeley Springs, Morgan Co. (MZUM). Virginia: S. Fork, Shenandoah River, Elkton, Rock- ingham Co. (Ortmann, 1919: 124). West Virginia: S. Fork, Shenandoah River, Harp- ers Ferry, Jefferson Co. (Ortmann, 1919: 124). Virginia: Potomac River, 5 mi. S Lucketts, Loudoun Co. Potomac River, Great Falls County Park, Fairfax Co. (USNM). Genus Alasmidonfa Say Subgenus Prolasmidonta Ortmann Prolasmidonta Ortmann 1914, Nautilus, 28: 44. Type species, Unio heterodon Lea, original designation. This subgenus is monotypic. Alasmidonfa heterodon (Lea) is restricted to the Atlantic Slope region. The general shape and the distinct posterior ridge give the shell the unmistakable look of an Alasmidonfa, but it is not close to the other members of the genus. There are two lateral teeth in the right valve and one in the left which is the reverse arrangement Atlantic Slope Unionidae • Johnson 347 of these teeth in all other Atlantic Slope Unionidae. Alasmidonta (Prolasmidonta) heterodon (Lea) Plate 13:2 Unio heterodon Lea 1830, Trans. Amer. Philos. Soc, 3: 428, pi. 8, fig. 11 (Schuylkill [River] and Derby Creek [mostly in Delaware Co.], Pennsylvania; figured type, belonged to Mr. Mason, not in USNM or ANSP [presumed lost]. Lea mentions specimens in possession of T. A. Conrad, now ANSP 41004, 41005). Lea, 1834, Obs. Unio, 1: 42. Alasmidonta heterodon (Lea) 1914, Simpson, Cat. Naiades, 1: 499. . \lasmidonta ( Prolasmidonta ) heterodon ( Lea ) . Ortmann, 1919, Mem. Carnegie Mus., 8: 173, pi. 12, figs. 1, 2. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem. no. 367, p. 24, fig. 35. Description. Shell small, generally from 30 to 45 mm, seldom reaching more than 50 mm in length. Outline subrhomboidal or subtrapezoidal, sometimes more or less elongate. Valves rather flat, though fe- males are somewhat swollen in the region of the posterior ridge, thin but strong. Anterior end regularly rounded; posterior end somewhat produced and angular. Ventral margin straight or slightly arcuate. Dorsal margin short and straight, almost always forming a sharp angle with the obliquely descending posterior margin. Hinge ligament occupying almost all of the posterior dorsal margin. Posterior ridge distinct though broadly rounded, some- times almost angular in females, usually ending in a broad point near the base. Posterior slope flat, or slightly concave, with one or two fine radiating lines. Umbos somewhat inflated but not very prominent, located slightly anterior to the middle of the shell, their sculpture consisting of three or four, sometimes five, bars, the first two concentric, the others with a distinct angle on the posterior ridge in front of which is a shallow sinus. The bars are rather heavy; the sinus does not assume the shape of a re-entering angle, and the sculpture can not be regarded as double-looped. The disk is smooth; the periostracum is green- ish olive to brownish. In addition to in- distinct concentric bands of lighter and darker green, there are often obscure green rays which are occasionally sharp and well defined. Hinge teeth delicate. Left valve with two lamellate pseudocardinal teeth which are parallel to the hinge, the anterior one sometimes rudimentary; an interdental tooth that may be well developed and iso- lated or connected with the posterior pseudocardinal; one lateral tooth. Right valve with two compressed crenulated pseudocardinals, the more anterior one often rudimentary; interdentum narrow and short; two delicate lateral teeth, the upper one often partially rudimentary. Beak cavities shallow, containing dorsal muscle scars. Anterior adductor muscle scars distinct, posterior ones less so. Pallial line distinct anteriorly only. Nacre bluish or silvery white, often cream colored or yellowish toward the umbonal cavities. Sexual differences are visible in the shell. In general the male shell is compressed, ovate, and elongate, with the ventral mar- gin curved, the posterior ridge not sharp, and the posterior slope not truncate, whereas the female shell is swollen in the region of the posterior ridge, rendering the posterior slope truncate, the outline of the shell more trapezoidal, and the ventral margin slightly concave. Length Height Width mm mm mm 52 40 38 26 23 22 17 16 18 [Ten Mile River], ML\- ville, Cheshire, New Haven Co., Connecticut. Male. Neuse River, Pooles Bridge, Wake Co., North Carolina. Male. Little River, 2 mi. NE Wendell, Wake Co., North Carolina. Female. Anatomy. Discussed by Lea (1863: 442) and Ortmann (1911: 295). Glochidia are subtriangular with strong hooks. They measure 0.30 mm in length, 0.25 mm in height. The host fish is unknown. 34S Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Breeding season. Ortmann (1919: 174) reports gravid females having been taken in February and April. It is bradytictic. Habitat. Lives in water that does not flow very fast, in gravel, sand, or in muddy sand. Remarks. Alasmidonta heterodon (Lea) is an Atlantic Slope species, not closely related to the other members of the genus. It can be confused with no other unionid in the region because of its distinct den- tition. A. heterodon has two lateral teeth in the right valve, though the upper one may be rather rudimentary, and one in the left valve. This tooth arrangement is just the opposite of all other Atlantic Slope Unionidae. This species is usually rare, and its er- ratic distribution may be merely a reflec- tion of inadequate collecting. Judging from the available records, it appears to be most widely distributed and abundant in the Connecticut River system of Vermont, New Hampshire, Massachusetts, and Connecti- cut. Range. Atlantic Slope: Neuse River system, North Carolina, north to the Petitcodiac River system, New Brunswick, Canada, but discontinuous. Known only from the following additional intervening drainage systems; Rappahannock River system, Virginia; Potomac River system, Maryland; Delaware River system, Penn- sylvania; Housatonic River system, Con- necticut (Linsley, 1845); Quinnipiac River system, Connecticut; Connecticut River sys- tem, Vermont, New Hampshire, Massa- chusetts, Connecticut; Merrimac River sys- tem, Massachusetts. Specimens Examined Neuse River System Neuse River Drainage. Nortli Carolina: Neuse River, Pooles Bridge [not located]; Neuse River [6 mi. E Raleigh]; Little River, Tarplcys Mill, 2 mi. NE Wendell (5); all Wake Co. Chicod Creek, Pitt Co. (MZUM). Rappahannock River System Rapidan River Drainage. Virginia: Blue Run, Orange Co. (MZUM). Rappahannock River Drainage. Vir- ginia: Mountain Run, Culpeper Co. (Ort- mann, 1919: 175). Potomac River System Potomac River Drainage. Maryland: Mcintosh Run, 4 mi. N Leonardtown, St. Marys Co. (H. D. Athearn). Subgenus Alasmidonta s.s. Say Monodonta Say 1817, Nicholson's Encyclopedia, 2 [no pagination] non Lamarck 1799. Alasmidonta Say 1818, lour. Acad. Nat. Sci. Phila., ser. 1, 1: 459. Type species, Monodonta undulata Say. Mono- typic. Alasmodonta [sic] Say 1819, Nicholson's Encyclo- pedia, ed. 3, 4 [no pagination]. Hemiodon Swainson 1849, Treatise on Malacology, pp. 288, 381. Species listed: H. undulatus [Say], An. rugosus, purpurascens, arcolata [all Swainson], non Hemiodon Swainson, op. cit., p. 191. Type species, A. undulata Say. Subsequent designation, Hermannsen, 1847, Indicis Gene- rum Malacozoorum, 1: 527. Unio))sis Swainson 1840, Treatise on Malacology, pp. 289, 382. Type species, U. mytiloides Swainson and U. radiata Swainson. Both names are based on the same fig., no. 64 = Alasmidonta undulata (Say). BuUella Simpson 1900, Proc. United States Natl. Mus., 22: 672. Type species, Margaritana arcida Lea. Original designation. Jugosus Simpson 1914, Cat. Naiades, 1: 357. Type species, Strophitus icrightianus Walker. Original designation, Simpson, 1900, Proc. United States Natl. Mus., 22: 666. Ortmann, 1914, Nautilus, 28: 45. Alasmidonta s. s. is restricted to the Apalachicolan and Atlantic Slope regions. There is one other species in this subgenus that is not covered in this paper, A. icright- iana (Walker) (Johnson, 1967: 8, pi. 2, fig. 5), from the Ochlockonee River system, Florida. In addition to the subgenera discussed here, or placed in synonymy above, Frier- Atlantic Slopk Umonidae • Johnson 349 son (1927: IS, 19) includes in his list two other subgenera of Alasmidonta, Piesso- donta Simpson, which appears to be repre- sented in the Interior Basin by a single species, A. calccola (Lea); and Sulcularia Rafinesque, which Ortmann and Walker (1922: 33) show is not identifiable. The species listed under it belong in the genus Lasmigona, subgenus Alasminota Ortmann. Alasmidonta (Alasmidonta) undulata (Say) Plate 13:3-4 Unto undulata Say 1817, Nicholson's Encyclo- pedia, 2 [no pagination], pi. 3, tig. 3 (Dela- ware and Schuylkill rivers [near Philadelphia, Philadelphia Co., Pennsylvania]; type, not in ANSP [lost]). Unto glabratus Sowerby 1823, Genera of Recent and Fossil Shells, no. 16, fig. 3 (no locality, type not in British Museum (Nat. Hist.) [lost]). non Lamarck 1819. Alasmidonta sculptilis Say 1829, New Harmony [Indiana] Disseminator, 2, no. 22, p. 339 (Vir- ginia; type not in ANSP [lost]). Unto hians Valenciennes 1827, in Humboldt and Bonpland, Voyage aux Regions Equinoxiales du Nouveau Continent, Pt. 2, 2: 235, pi. 54, figs. 2a, b (environs de Philadelphie [Philadelphia Co., Pennsylvania]; type [location unknown]). Uniopsis radiata Swainson 1840, Treatise on Malacology, p. 289, fig. 62 ([North America] type [location unknown]). Uniopsis mytiloides Swainson 1840, Treatise on Malacology, p. 382, fig. 62 ( [North America] type [location unknown]. [Both this and the preceding were found on the same figure.] Unio swainsoni Sowerby 1868, Conch. Iconica, 16, Unto, pi. 76, fig. 396 ( Hab. ? figured holo- type British Museum [Nat. Hist.] 1900.3.19.21.). Alasmidonta undulata (Say). Simpson, 1914, Cat. Naiades, 1: 494. Alasmidonta (Alasmidonta) undulata (Say) Ort- mann 1919, Mem. Carnegie Mus., 8: 117, pi. 11, fig. 7. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem. no. 367, p. 25, fig. 33. Athearn and Clarke, 1962, Natl. Mus. Canada, Bull. 183, p. 24, pi. 3, figs. 3, 4. Description. Shell generally small, from 50 to 65 mm, though sometimes reaching 80 mm in length. Outline elliptical or sub- rhomboid. Valves somewhat inflated, thin, becoming quite solid anteriorly. Anterior end regularly rounded; posterior end rather broadly pointed. Ventral margin almost always slightly rounded. Dorsal margin short and straight, forming an angle with the obliquely descending posterior margin. Hinge ligament located posteriorly of the umbos, occupying most of the dorsal mar- gin. Posterior ridge rounded, sometimes with a faint second ridge above it. Posterior slope slightly concave, sometimes with irregular oblique corrugations. Umbos rather inflated and somewhat raised above the hinge line, located anterior to the middle of the shell, their sculpture consisting of four or five strong ridges running nearly parallel to the growth lines. These ridges extend some distance on the disk, and posteriorly are angular, thickened, and raised. They are crossed by fine, wrinkled, radiating threads. In front of and behind the regular umbonal sculpture there are radiating lirae. Generally the en- tire surface of the shell is smooth, though occasionally slightly roughened, especially posteriorly; periostracum greenish, yellow- ish, reddish brown, with more or less distinct green to blackish rays of varying width over the entire surface. The pos- terior slope is often somewhat lighter than the rest of the shell, with finer and sharper rays. Older shells become blackish, but the rays are still visible through trans- mitted light. Left valve with a compressed, sometimes stumpy pseudocardinal tooth, often with a rudimentary one in front; interdentum rather broad, occasionally with a visible but connected interdental tooth; lateral tooth barely distinct. Right valve with one rather high, chunky, subtriangular tooth; often with a broad interdentum; with just a trace of a lateral tooth. Beak cavities moderately deep with muscle scars under the hinge plate. Anterior adductor muscle scars deep, posterior ones faint. Pallial line distinct anteriorly where the shell is thickened. Nacre white, salmon pink, or reddish, iridescent posteriorly. 350 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Sexual differences are not visible in the shell. Length Height Width mm mm mm 80 51 72 45 54 32 39 31 23 Tar River, Bruce, 9 mi. NW Greenville, Pitt Co., North Carolina. Neuse River, Falls, Wake Co., North Carolina. Nottoway River, 3 mi. E Rawlings, Brunswick Co., Virginia. Anatomy. Discussed by Ortmann (1911: 296). Glochidia are moderately large with strong hooks. They measure 0.34 mm in length and 0.36 mm in height. The host fish is unknown. Breeding season. Ortmann (1919: 178) states that this species is clearly bradytietic, with a short interim in June and July. Habitat. Lives in water that does not flow very fast, in gravel or sand, appears to avoid mud. Remarks. Alasmidonta undulata (Say), of the Atlantic Slope, is not closely related to any Alasmidonta of the Interior Basin. Its nearest relatives are A. arcula of the Altamaha River system, Georgia, and A. triangulata of the Apalachicolan and South- ern Atlantic Slope regions; but both of these species are more triangular in outline, with shaqoer posterior ridges. A. undulata also resembles A. wiightiana (Walker) of the Ochlockonee River system, Florida, in shape, and it has a similar tooth structure and rounded posterior ridge, but wrighb- iuna has strong ridges which cover the posterior slope, whereas undulata is un- sculptured. A. undulata is a common species, espe- cially on the Northern Atlantic Slope, and is often locally abundant. It seems to be somewhat less common on the Southern Atlantic Slope. It avoids the larger rivers, preferring smaller streams, in which it often goes far up toward the headwaters. Range. Atlantic Slope: Catawba River, North Carolina, of the upper Cooper- Santee River system, north to the lower St. Lawrence River system, Canada. Specimens Examined Cooper-Santee River System Catawba River Drainage. Nortli Caro- lina: Catawba River (USNM). Pedee River System Yadkin River Drainage. North Caro- lina: Uwhairie River (USNM). Cape Fear River System Cape Fear River Drainage. North Caro- lina: Morgan Creek, 1 mi. SE Chapel Hill, Orange Co. (MZUM). Rocky River, 11 mi. N Sanford, Chatham Co. (MZUM). Cape Fear River ( MZUM ) . Neuse River System Neuse River Drainage. North Carolina: Neuse River, Falls (50); Neuse River, 5.75 mi. E Raleigh Center (45); Swift Creek, 3 mi. SSW Garner (20); Middle Creek, 6.25 mi. SE Apex (18); Little River, Tar- pleys Mill, 2 mi. NE Wendell (5); all Wake Co. Pamlico River System Tar River Drainage. North Carolina: Tar River, Old Sparta, 3.5 mi. W Pinetops, Edgecomb Co. Tar River, Bruce, 9 mi. NW Greenville, Pitt Co. Chowan River System Nottoway River Drainage. Virginia: Nottoway River, 3 mi. E Rawlings, Bruns- wick Co. James River System James River Drainage. Virginia: Calf- pasture River (Conrad, 1846: 407); North [ = Maury] River, 2 mi. WNW Lexington (MZUM); North River, Buena Vista (USNM); all Rockbridge Co. Rivanna River, 2 mi. W Columbia, Fluvanna Co. James River, opposite Maidens, Goochland Co. York River System North Anna River Drainage. Virginia: Mine Run, Orange Co. (MZUM). Atlantic Slope Unionidm: • Johnson 351 Rappahannock River System Rapidan River Drainage. Virginia: Blue Run (MZUM); Mountain Run (MZUM); both Orange Co. Rappahannock River Drainage. Vir- ginia: Rappahannock River, Remington; Marsh Run, Remington (both Ortmann, 1919: ISO); both Fauquier Co. Occoquan Creek System Occoquan Creek Drainage. Virginia: Broad Run [Fairfax Co.], 3 mi. W Ma- nassas, Prince William Co. Potomac River System Potomac River Drainage. Maryland: Wills Creek, Ellerslie, Allegany Co. (Ort- mann, 1919: 179). Potomac River, Han- cock, Washington Co. (Ortmann, 1919: 179). Pennsylvania: Great Tonoloway Creek, Thompson Township, Fulton Co. (Ortmann, 1919: 179). Conococheague Creek, Greencastle and Scotland, Franklin Co. (Ortmann, 1919: 179). Virginia: N Fork, Shenandoah River, Broadway, Rockingham Co. (Ortmann, 1919: 179). S River of S Fork, Shenandoah River, Waynesboro, Augusta Co. (Ortmann, 1919: 179). S Fork, Shenandoah River, Elkton, Rockingham Co. Ortmann, 1919: 179). West Virginia: Shenandoah River, Harpers Ferry, Jefferson Co. (USNM). Virginia: Potomac River, Great Falls, Fairfax Co. Alasmidonta (Alasmidonta) triangulata (Lea) 91016 13:5,6 Margaritana triangulata Lea 1858, Proc. Acad. Nat. Sci. Phila., 10: 138 (Upper Chattahoochee [River], Georgia; Columbus, Georgia; Polato Co. [sic] [Potato Creek, Upson Co.], Georgia; Sawney's Creek [about 8 mi. NW Camden, Kershaw Co.], South Carolina). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 228, pi. 32, fig. Ill; figured holotype USNM 86249, from Upper Chattahoochee [River], Lea, 1859, Obs. Unio, 7: 46. Alasmidonta triangulata (Lea). Simpson, 1914, Cat. Naiades, 1: 509. Clench and Turner, 1956, Bull. Florida State Mus., 1: 180, pi. 5, fig. 4. Strophitus wrightianus Clench and Turner 1956, Bull. Florida State Mus., 1: 180 non Walker 1901, Nautilus, 15: 65, pi. 3. Description. Shell generally small, though reaching 70 mm in length in the Apa- lachieola River system. Outline subtri- angular. Valves inflated, thin to rather solid. Anterior end regularly rounded; posterior end rather broadly pointed. Ven- tral margin almost always slightly rounded. Dorsal margin short and straight, forming an angle with the obliquely descending posterior margin. Hinge ligament occupy- ing almost all of the posterior dorsal mar- gin. Posterior ridge high and angular with a second or third faint ridge above it, the second one ending in a slight biangulation near the base of the shell. Umbos full and high, located anterior to the center of the shell, their sculpture consisting of a number of strong ridges running nearly parallel to the growth lines. These ridges extend some distance on the disk and are crossed by fine, wrinkled, radiating threads. In front of and behind the regular umbonal sculp- ture there are radiating lirae. The disk is smooth; the periostraeum is usually fine and limited to the posterior slope. Young shells may be yellowish or greenish yellow, often with dark green rays of varying width. Older shells become dark brownish black, but the rays and yellowish ground color are still visible when seen through transmitted light. Left valve with a compressed, sometimes stumpy pseudocardinal tooth, often with a rudimentary one in front; interdentum broad, occasionally with a visible but con- nected interdental tooth; lateral tooth barely visible. Right valve with one rather high subtriangular tooth; broad interden- tum; lateral tooth barely visible. Beak cavities deep, with dorsal muscle scars under the hinge plate. Anterior adductor muscle scars deep, posterior ones faint. 352 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Pallial line distinct anteriorly where the shell is thickened. Nacre bluish white to salmon pink. Sexual differences are not visible in the shell. Length Height Width 68 48 47 32 38 27 Savannah River [Johnsons Landing], 10 mi. W Allen- dale, Allendale Co., South Carolina (MZUM). Mill Raee, 2 mi. N Sardis, Burke Co., Georgia. Habitat. Prefers sandy mud in rocky pools and eddies of large creeks and rivers. Remarks. Alasmidonta triangulata (Lea) of the Apalachicolan and Southern Atlantic Slope regions cannot be confused with A. arcula ( Lea ) of the Altamaha River system because it is proportionally not as high. The umbos are anterior to the center, and while incurved, do not render the hinge line concave, and the posterior ridge is not as acutely sharp. A. triangulata is closest to A. undulata (Say) of the Atlantic Slope region, but the latter has a shape which is usually an elongated oval, a very broadly rounded posterior ridge, more ponderous pseudo- cardinals, and more of an interdentum. In the Apalachicolan region, A. triangulata can be confused only with A. wrightiana (Walker); the latter has the shape, rounded posterior ridge, and tooth struc- ture of A. undulata, differing from un- dulata and triangulata in the strong ridges which cover the posterior slope. A. triangulata is a rare species through- out its range, but is especially so on the Southern Atlantic Slope where each of the lots examined consists of only one or two specimens. Range. Apalachicolan region: restricted to the Apalachicola River system. Southern Atlantic Slope: Ogeechec and Savannah river systems, Georgia; and Cooper-Santee River system, South Carolina. Specimens Examined Ogeechee River System Ogeechee River Drainage. Georgia: Ogeechee River, Scarboro (MZUM); Ogeechee River, bridge, 1 mi. S Dover; both Screven Co. (MZUM). Savannah River System Savannah River Drainage. South Caro- lina: Savannah River [Johnsons Landing], 10 mi. W Allendale, Allendale Co. (MZUM). Georgia: Mill Race, 2 mi. N Sardis, Burke Co. Cooper-Santee River System Wateree River Drainage. South Caro- lina: Sawney's Creek [about 8 mi. NW Camden, Kershaw Co.] (USNM). Alasmidonta (Alasmidonta) arcula (Lea) Plate 13: 7, 8 Margaritana arcula Lea 1836, Synopsis Naiades, p. 43 [nomen nudum]. Lea, 1838, Trans. Amer. Philos. Soc., 6: 71, pi. 22, fig. 69 (Altamaha [River], Liberty [now Long] Co., Georgia; figured holotype USNM 86170). Lea, 1838, Obs. Unio, 2: 71. Alasmidonta arcula (Lea). Simpson, 1914, Cat. Naiades, 1: 508. Description. Shell medium, reaching 75 mm in length. Outline triangular. Valves much inflated, thin but strong. Anterior end regularly rounded; posterior end sharply truncated. Ventral margin straight or slightly rounded. Dorsal margin short and curved, forming a sharp angle with the obliquely truncated posterior margin. Hinge ligament occupying almost all of the posterior dorsal margin. Posterior ridge high and sharply angular, with a sec- ond faint ridge above, ending in a slight biangulation near the base of the shell. Posterior slope very truncated. Umbos extremely full and high, located at the center of the shell, their sculpture consist- ing of a number of strong ridges running nearly parallel to the growth lines. These ridges extend well out on the disk and Atlantic Slope Unionidae • Johnson 353 posteriorly are angular, considerably thick- ened, and raised. They are crossed by fine, wrinkled, radiating threads. In front of and behind the regular umbonal sculp- ture there are radiating lirae. The disk is smooth; the periostracum is usually fine and limited to the posterior slope. Shell dull greenish yellow with distinct green rays of varying length over the entire sur- face. Mature shells are often blackish, but the rays are still visible through trans- mitted light. Left valve with a long compressed pseudocardinal tooth, often with a rudi- mentary one in front; interdental tooth distinct but connected; lateral teeth barely visible. Right valve with one compressed, triangular tooth; scarcely any interdentum; lateral tooth barely visible. Beak cavities very deep, with dorsal muscle scars under the hinge plate. Anterior adductor muscle scars deep, posterior ones faint. Pallial line distinct anteriorly where the shell is thickened. Nacre bluish white or white. Sexual differences are not visible in the shell. Length Height Width mm mm mm 80 67 60 50 51 44 50 43 38 Altamaha River [Long Co.], Georgia. Ocmulgee River, Hawkins- ville, Pulaski Co., Georgia. Altamaha River, 4 mi. NE Jesup, Wayne Co., Geor- gia. Anatomy. Discussed by Lea (1863: 447). Habitat. Lives in sandy mud below sand bars in sluggish water and eddies. Remarks. Alasmidonta arcula (Lea) of the Altamaha River system cannot be con- fused with any other unionid in the At- lantic Slope region, though it is close to A. triangulata (Lea), which is found in both the Apalachicolan and Atlantic Slope regions. Alasmidonta arcula differs from triangulata by being proportionally higher with centrally located umbos which are so elevated as to be incurved, by its concave hinge line, and by its extremely sharp posterior ridge which forms almost a 90° angle with the posterior slope. This is a rare species; not more than a few specimens have ever been collected at any station in this century save for a series of twenty-two specimens collected by H. D. Athearn in 1962 in the Ocmulgee River, below Lumber City, Telfair Co., Georgia. The old unlocalized specimens of areata in the major museums are presumed to have been collected near the type locality in the tidewater region of the Altamaha River, where it may be, or was, more abundant. Range. Southern Atlantic Slope: re- stricted to the Altamaha River system, Georgia. Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Ocmulgee River, Hawkinsville; Limestone Creek, near Hartford (H. D. Athearn); both Pulaski Co. Ocmulgee River, below Lumber City, Telfair Co. (H. D. Athearn). Ohoopee River Drainage. Georgia: Ohoopee River, 3 mi. N Leman, Emanuel Co. Altamaha River Drainage. Georgia: Altamaha [River], Liberty [now Long] Co. (USNM). Altamaha River, Fort Barring- ton, Mcintosh Co. (ANSP). Subgenus Decurambis Rafinesque Decurambis Rafinesque 1831, Continuation of Monograph Rivalve Shells of River Ohio (Phila. ), pp. 4, 5. Species listed: Alasmodon scriptum Rafinesque and A. atropurpureum Rafinesque. Type species, Alasmodon scriptum Rafinesque = Alasmidonta marginata (Say). Subsequent designation, Ortmann and Walker, 1922, Occ. Pap. Mus. Zool., Univ. Mich., no. 112, p. 38. Rugifera Simpson 1900, Proc. United States Natl. Mus., 22: 670. Type species, Alasmidonta marginata Say. Original designation. The present author follows Ortmann and Walker (1922: 38) and regards Decuram- 354 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 bis as a subgenus of Alasmidonta, though Fricrson (1927: 21) raised Decurambis to a genus. Decurambis includes but two allopatric species: A. marginata Say (Ortmann, 1919: LSI, pi. 12, fig. 3), primarily a species of the Interior Basin, but which is also found on the Northern Atlantic Slope in the Sus- quehanna and upper St. Lawrence drainage systems; and A. varicosa (Lamarck), which is limited in distribution to the Atlantic- Slope. Alasmidonfa {Decurambis) varicosa (Lamarck) Plate 14: 1,2 Unio varicosa Lamarck 1819, Hist. Nat. des Animaux sans Vertebres, 6: 78 (la riviere de Schuglkill [Schuylkill] pies de Philadelphie [Philadelphia Co., Pennsylvania]; holotype, Geneva Museum, teste Johnson, 1953, Nautilus, 66: 95; aussi dans le lac Champlain, [Vermont]). Alasmodon corrugata De Kay 1843, Zool. New York, Moll., pt. 5: 198, pi. 24, fig. 259 (Passaic River, New York; type, New York Lyceum of Nat. Hist, [destroyed by fire]). Mya rugulosa Wood 1856, in Hanley, Index Test, p. 199, pi. 1 supp., fig. 7 ( North America; type [probably lost]). Alasmidonta varicosa (Lamarck). Simpson, 1914, Cat. Naiades, 1: 506. Alasmidonta (Decurambis) varicosa (Lamarck). Ortmann, 1919, Mem. Carnegie Mus., 8: 190, pi. 12, fig. 5. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem. no. 367, p. 28, fig. 34. Athearn and Clarke, 1962, Natl. Mus. Canada, Bull. 183, p. 25, pi. 3, figs. 5, 6. Description, Shell usually small, from 40 to 65 mm, though sometimes reaching 80 mm in length. Outline rhomboid or sub- elliptical. Valves not much inflated, thin but strong. Anterior end regularly rounded; posterior end slightly elongated and bi- angulate. Ventral margin straight or slightly arcuate. Dorsal margin straight, forming an indistinct angle with the obliquely descending posterior margin. Hinge ligament located posteriorly of the umbos, short but prominent. Posterior ridge broadly rounded, with a second very faint ridge above it, ending in a biangu- lation near the base. Posterior slope broadly rounded, with a system of fine radial wrinkles running to the upper posterior margin. Umbos large and somewhat in- flated, but not high, located in the anterior third of the shell, their sculpture consisting of three or four thick bars. The disk is smooth; the periostraeum is yellowish, greenish, brownish, or blackish, generally with distinct narrow or wide, straight, greenish or blackish rays. The rays are usually broken by growth lines. Each valve has one thin, distinct, de- pressed, triangular pseudocardinal; no lateral teeth. Beak cavities moderately deep with muscle scars under the hinge plate; anterior adductor muscle scars deep, posterior ones less so. Pallia] line distinct anteriorly where the shell is slightly thick- ened. Nacre bluish white, often with salmon, pinkish, or purplish shades. Sexual differences are visible in the shell; females are slightly more swollen in the region of the posterior ridge. Occasionally the swelling is so great that the lateral faces of the shell in front of the ridge ap- pear flat or even concave. The ventral margin may be slightly emarginate, but male shells are also sometimes emarginate so that the sex can not alwavs be ascer- tained from the shell alone. Length Height Width mm mm mm 73.0 40.0 31.0 44.0 26.0 18.0 33.0 28.0 12.0 32.0 29.5 13.0 N Fork, Shenandoah Biver, E Woodstock, Shenandoah Co., Vir- ginia. Female. Turkey Creek, 8 mi. NW Edgefield, Edgefield Co., South Carolina. Female. Bocky Biver, 11 mi. N Sanford, Chatham Co., North Carolina. Male. As above. Female. Anatomy. According to Ortmann ( 1919: 191 ) , the anatomy of this species is very similar to that of Alasmidonta marginata Say. Breeding season. Ortmann (1919: 191) found that in Pennsylvania breeding begins Atlantic Slope Unionidae • Johnson 355 lin August, when eggs are present; glochidia are found in September, and are carried over the winter and discharged the follow- ing May. Habitat. Lives among rocks on gravel substrates; also on sandy shoals, especially in rapids and riffles of small rivers and creeks. Remarks. Alasmidonta varicosa (La- marck), having rudimentary pseudocar- dinals, a lack of lateral teeth, and radial wrinkles on the posterior slope, is dis- tinguishable from the other Alasmidonta of the Atlantic Slope region. It is replaced in the upper Tennessee and Cumberland river systems by A. ravenelina (Lea) which differs from varicosa by being more elongate, by having a more broadly bian- gulate posterior ridge, and by usually not having more than a trace of radial wrinkles on the posterior slope. Alasmidonta varicosa is the Atlantic re- placement of A. marginata Say which is found in the Interior Basin throughout the upper Mississippi drainage, including the Ohio, Tennessee and Cumberland river systems. In the Northern Atlantic Slope region, marginata is found in the Susque- hanna River, Pennsylvania, and the upper St. Lawrence River, Canada. A. marginata has a very sharp posterior ridge and a truncated posterior slope, which renders the shell subtrapezoidal; also, the rays tend to be broken into spots. A. varicosa has a gradually rounded, broadly biangu- late posterior ridge, a gently rounded pos- terior slope, and a rhomboidal shape. The rays are never broken into spots; the shell is not as heavy, nor does it grow as large as does marginata. Though marginata pre- fers larger rivers while varicosa prefers smaller streams, they are nevertheless sometimes found together where their ranges overlap, but they are easily sepa- rated. Simpson (1914, 1: 506) correctly re- corded this species from as far south as South Carolina, but I was unable to find the material on which his record was based. Ortmann (1919: 193) personally collected A. varicosa in the upper Catawba River, North Carolina, and Clench and Okkelberg collected a single specimen in 1929 in a tributary of the Savannah River, South Carolina. Seemingly rare below the Potomac River, Maryland, A. varicosa is relatively abundant on the Northern At- lantic Slope. Range. Atlantic Slope: tributary of the upper Savannah River system, South Caro- lina, north to the lower St. Lawrence River system, Canada. Specimens Examined Savannah River System Savannah River Drainage. South Caro- lina: Turkey Creek, 8 mi. NW Edgefield, Edgefield Co. (MZUM). Cooper-Santee River System Catawba River Drainage. North Caro- lina: Catawba River, Bridgewater, Burke Co. (Ortmann, 1919: 193). Cape Fear River System Cape Fear River Drainage. North Caro- lina: Rocky River, 11 mi. N Sanford, Chat- ham Co. James River System James River Drainage. Virginia: Calf- pasture River (Conrad, 1846: 407 as A. marginata Say). Occoquan Creek System Oeeoquan Creek Drainage. Virginia: Broad Run [Fairfax Co.], 3 mi. W Manassas, Prince William Co. Potomac River System Potomac River Drainage. Virginia: N Fork, Shenandoah River, E Woodstock, Shenandoah Co. S Fork, Shenandoah River, Riverton, Warren Co. Maryland: Potomac River, Hancock, Washington Co. (Ortmann, 1919: 193). 356 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Genus Anodonta Lamarck Subgenus Anodonta s.s. Lamarck Anodonta Lamarck 1799, Memoires tie la Soc. d'Hist. Nat. de Paris, p. 87. Type species, Mytilus cygneus Linnaeus. Mono- typic. Placed on the Official List of Generic Names in Zoology, in 1926, Opinion 94. Re- confirmed, 1959, Opinions and Declarations rendered by Int. Comm. Zool. Nomen., 20 (28): 303-10, Opinion 561. Ortmann, 1912, Ann. Carnegie Mus., 8: 286, partim. Anodonta s. s. is found primarily in the Palearctic region of Europe and Asia and like the type species may be monoecious and have flat umbones. The several western American species of the Pacific region clearly belong to Anodonta s. s. Subgenus Pygonodon Crosse and Fischer Pyganodon Crosse and Fischer 1894, in Fischer and Crosse, Mission Sci. an Mexique, pt. 7, 2: 518. Species listed: A. glohosa Lea; A. tabascens Morelet; A. nopalatensis Sowerby; A. grijalvae Morelet; A. glauca Valenciennes. Introduced as a section. Type species, Anodonta globosa Lea. Subsequent designation, Frierson, 1927, Check List North American Naiades, p. 9, teste Errata et Cor- rigenda. Raised to a subgenus. Frierson (1927: 14-16) used Pyganodon as a subgenus to include most of the Anodonta of middle and eastern North America. It is suspected that when the Interior Basin taxa included in this subgenus are revised, it will be found that there are about an equal number of species in the Interior Basin and Atlantic Slope region. Anodonta (Pyganodon) cataracta cataracta Say Plate 14: 3,4 Plate 15: 1 Anodonta fluviatilis Cmclin, of authors. See note under Lampsilis ochracca Say, p. 388. Anodonta cataracta Say 1817, Nicholson's Encyclo- pedia, 2 [no pagination], pi. 3, fig. 2 (deep part of a mill dam [presumably near Philadel- phia, Philadelphia Co., Pennsylvania, and here so restricted]; type, ANSP [lost]). Simpson, 1914, Cat. Naiades, 1: 386. Ortmann, 1919, Mem. Carnegie Mus., 8: 152, pi. 10, fig. 5, pi. 11, fig. 1. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem. 367, p. 39, fig. 46. Athearn and Clarke, 1962, Natl. Mus. Canada Bull. no. 183, p. 27, pi. 2, figs. 5, 6. Anodonta marginata Say 1817, Nicholson's En- cyclopedia, 2 [no pagination], pi. 3, fig. 5 ( very common in our rivers [presumably near Philadelphia, Philadelphia Co., Pennsylvania, and here so restricted]; type, ANSP [lost]). Simpson, 1914, Cat. Naiades, 1: 388 [partim]. The measurements in the description, as tran- scribed into millimeters, indicate the type to have been 51 mm in length and 38 mm in height. The figured shell was 69 mm in length and 36 mm in height. The first set of di- mensions are those of a cataracta from a creek environment, while the second are of the elongated form which is sometimes found in ponds, and which becomes more common in the northern range of cataracta. Anodonta teres Conrad 1834, New Fresh Water Shells United States, p. 47, pi. 7, fig. 2 (Santee Canal [Berkeley Co.], South Carolina [type, ANSP lost]). Simpson, 1914, Cat. Naiades, 1: 390. Anodon excnrvata De Kay 1843, Zool. of New York, Moll., pt. 5, p. 202, pi. 17, fig. 233 (Shaker Pond, Niskayuna, Albany Co., New York; Passaic River [Paterson, Passaic Co., New Jersey]; type, New York State Colin., now in USNM [not yet located]). Anodonta virgulata Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 86 (Roanoke River, Weldon [Halifax Co.], North Carolina; Washington Co., Georgia). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 213, pi". 33, fig. 282; figured holotype USNM 86593 from Roanoke River. Lea, 1863, Obs. Unio, 9: 35. Anodonta lacnstris Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 84 (Crooked Lake and Little Lakes [Herkimer Co.], New York). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 363, pi. 62, fig. 188; figured type USNM [not located]; though a slightly smaller specimen, USNM 86597 marked Mohawk, New York. ex Lewis, is labeled holotype. Lea, 1860, Obs. Unio, 8: 45. Anodonta liallenheckii Lea 1858, Proc. Acad. Nat. Sci. Phila., 10: 138 (Uphaupee [Uphapee] Creek, Macon Co., Georgia [Alabama]). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 50, pi. 32, fig. 112; figured holotype USNM 86428. Lea, 1859, Obs. Unio, 7: 50. Simpson, 1914, Cat. Naiades, 1: 392. Clench and Turner, 1956, Bull. Florida State Mus., 1: 186, pi. 9, fig. I- Anodonta gcsnerii Lea, 1858, Proc. Acad. Nat. Sci. Phila., 10: 139 (Uphaupee [Uphapee] Creek, Macon Co., Georgia [Alabama]). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: Atlantic Slope Unionidae • Johnson 357 231, pi. 31, fig. 109; figured holotype USNM S6427. Lea, 1859, Obs. Unio, 7: 49. Anodonta dariensis Lea 1S58, Proc. Acad. Nat. Sci. Phila., 10: 139 (Hopeton, near Darien [Mcintosh Co.]: Swift Creek, near Macon [Bibb Co.]; Flint River; all Georgia). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 230, pi. 28, fig. 99; figured holotype USNM 86600 from Swift Creek. Lea, 1859, Obs. Unio, 7: 48. Simpson, 1914, Cat. Naiades, 1: 394. Anodonta williamsii Lea 1862, Proc. Acad. Nat. Sci. Phila., 14: 169 (Potomac River at the Wbite House [District of Columbia]; below Mount Vernon [Fairfax Co., Virginia]; Oth- calooga Creek [Gordon Co.], Georgia. Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 27, pi. 10, fig. 26; type, USNM [not located]. Lea, 1867, Obs. Unio, 11: 31. Anodonta tryoni Lea 1862, Proc. Acad. Nat. Sci. Phila., 14: 169 (Schuylkill River above Phila- delphia; Delaware River at League Island [both Philadelphia Co., Pennsylvania]; Flem- ington [ = Farmington, Hartford Co.], Con- necticut; Westfield [Hampden Co.], Massa- chusetts; Potomac River near Chain Bridge, above Washington, D. C). Lea, 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 28, pi. 10, fig. 27; type, USNM [not located], paratype, USNM 86572 from Potomac River. Lea, 1867, Obs. Unio, 11: 32. Anodonta dolcaris Lea 1863, Proc. Acad. Nat. Sci. Phila., 15: 193 (Stewart's Mill Dam, Union Co., North Carolina). Changed to: Anodonta doliaris Lea 1866, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 24, pi. 8, fig. 23; figured holotype ANSP 126522a. Lea, 1867, Obs. Unio, 11: 28. Simpson, 1914, Cat. Naiades, 1: 393. Description. Shell medium to large, reaching over 165 mm in length. Outline subelliptical, more or less elongate. Valves of young specimens not much inflated, those of older ones quite inflated, thin, sometimes very thin. Anterior end regularly rounded; posterior end rather pointed. Ventral margin straight, or slightly curved, very occasionally slightly incurved. Dorsal margin slightly curved, often forming a winglike angle where it meets the obliquely descending posterior margin. The posterior margin joins the ventral margin in a broad point, somewhat below the medial line. Hinge ligament long and low. Posterior ridge indistinct. Posterior slope slightly concave, sometimes with two faint ridges. Umbos slightly swollen and moderately convex, located in the anterior third or fourth of the shell, their sculpture consist- ing of five to seven double-looped bars, not appreciably lower in front or behind the sinus, but of uniform elevation. Perios- tracum generally smooth and shiny, slightly roughened toward the margins and on the posterior slope. Surface of the shell sometimes uniformly straw yellow, but usually lighter to darker green with con- centric lighter and darker bands, with darker rays on the disk. The rays on the posterior slope are darker, at least those rays on the faint ridges, though generally the whole posterior slope is blackish green. No hinge plate or teeth; muscle scars inconspicuous and poorly defined. Nacre bluish white and iridescent. Females are sometimes indicated by a swelling of the valves posterior to the middle, but this is not a uniform character. Length Height Width mm mm mm 165 86 62 Lake Michie, 12 mi. NNE Durham, Durham Co., North Carolina. Savannah River, Johnsons Landing, 10 mi. W Allen- dale, Allendale Co., South Carolina ( MZUM ) . Swift Creek, near Macon [Bibb Co.], Georgia. Holo- type of A. dariensis- Lea. Anatomy. Thoroughly discussed by Simpson (G. B., 1884) and commented upon by Reardon (1929: 8, pi. 3, figs. 1-10). The glochidia are hooked. Breeding season. From August to May (Ortmann, 1919: 154). Habitat. Lives in rivers and creeks in soft mud and sand; also in sand bars, generally in slowly moving water; prefers ponds. Remarks. In the Atlantic Slope region Anodonta cataracta Say can be confused in the lower St. Lawrence drainage with Anodonta cataracta fragilis Lamarck (see Athearn and Clarke, 1962: 28; 1963: 22); Anodonta implicata Say under which see: 152 79 65 107 58 49 358 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Remarks on p. 361; and in central New York with Anodonta grandis Say, of the Interior Basin. A. cataracta Say differs from grandis primarily in the umbonal sculpture, that of the latter consists of four to five bars of which the first two are concentric, while the others are distinctly double-looped, with a sharp re-entering sinus between the loops. The anterior loop is broadly rounded, the posterior loop is angular, narrow, and characteristically ele- vated into a tubercle. In cataracta the bars are of uniform height and are not tubercular. These distinctions were prop- erly pointed out by Marshall (1890: 188, 189). Except for this minor difference the two species are very close, though cataracta tends to be a brighter green, and while short high examples of cataracta are found in rivers, these forms are more common in grandis. In general, cataracta is more elongated, especially individuals from ponds, and the posterior point is more elevated above the base line, rendering the ventral margin more convex. Ortmann (1919: 154) has clearly pointed out the different reaction of the two species to environment. Specimens from the Alabama-Coosa River system show no morphological dif- ferences separating hallcnbeckii (Lea) (Clench and Turner, 1956: 186) from cataracta. It is possible that cataracta ar- rived there through the commingling of the upper Alabama and Coosa rivers di- rectly, though it is suggested elsewhere in this paper that cataracta may more likely have spread into the Apalachicola River system through a confluence of it with the Savannah, and then subsequently reached the Alabama system through a mingling of the waters of its Uphapee Creek and Uchee Creek of the Apalachicola River system. Most of the shells from Uphapee Creek are the long pond forms. I did not find any Anodonta in this broad, swift flowing creek in 1964, suggesting that the original lot was from a quiet back- water which I did not locate. In the southern part of the Atlantic Slope A. cataracta is not very common. In the Altamaha River, at the stations listed below, only a few specimens were taken. Lea's specimen of dariensis from Hopeton, near Darien, Mcintosh Co., Georgia, was not located, hence it is not included in the records since it might be gibbosa. In the Museum of Comparative Zoology there is a single specimen of cataracta from the C. B. Adams collection made before 1850, labeled Washington Co., Georgia. If this record is valid, then cataracta is found either in the upper Oconee River of the Altamaha River system or in the upper Ogeechee River system. This species be- comes abundant and reaches large size in impounded waters, such as Lake Moultrie of the Cooper-Santee River system, South Carolina, Stewarts Mill Dam, Union Co., North Carolina, and Lake Michie on Flat River, North Carolina. Range. Alabama-Coosa River system. Apalaehicolan region: Choctawhatchee and upper Apalachicola River systems. Atlantic Slope: Altamaha River system, Georgia, north to the lower St. Lawrence River system, Canada. Interior Basin: probably extending from the St. Lawrence River system westward to Michigan. Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Possum Lake, NE Stone Mountain, De Kalb Co. Swift Creek, near Macon, Bibb Co. Limestone Creek, near Hartford, Pulaski Co. (H. D. Athearn). Dicksons Creek, 10 mi. NE Fitzgerald, Ben Hill Co. Ocmulgee River, below Lumber City, Tel- fair Co. (H. D. Athearn). Savannah River System Savannah River Drainage. Sotitli Caro- lina: Savannah River, 6 mi. S Martin; Savannah River, Johnsons Landing, 10 mi. W Allendale; both (MZUM); both Allen- dale Co. Atlantic Slope Unionidae • Johnson 359 Cooper-Santee River System Congaree River Drainage. South Caro- lina: Congaree River. Catawba River Drainage. North Caro- lina: Bissels Pond, Charlotte, Mecklenburg Co. (USNM). Wateree River Drainage. Soutli Caro- lina: Wateree River, 2.5 mi. W Camden, Kershaw Co. Santee River Drainage. South Caro- lina: Poplar Creek, Schulers Fish Pond, near Santee State Park; Eutavv Springs, [town of] Eutaw Springs; both Orange- burg Co. Lake Moultrie, Cross, Berkeley Co. Pedee River System Yadkin River Drainage. North Caro- lina: Salem, Forsyth Co. (USNM). Yadkin River, Rowan Co. Stewarts Mill Dam, Union Co. (USNM). Waccamaw River System Waccamaw River Drainage. North Carolina: Lake Waccamaw, [town of] Lake Waccamaw, Columbus Co. Cape Fear River System Cape Fear River Drainage. North Caro- lina: Travis Creek, 1.5 mi. N Gibsonville, Guilford Co. (MZUM). Rocky River, 11 mi. N Sanford, Chatham Co. Livingstons Creek [Brunswick Co.] (USNM). Neuse River System Neuse River Drainage. North Carolina: Lake Michie, 12 mi. NNE Durham Center, Durham Co. (75). Roanoke River System Dan River Drainage. Virginia: Aarons Creek, 3 mi. W Buffalo, Lithia Springs; Dan River; both Halifax Co. Roanoke River Drainage. North Caro- lina: Roanoke River, Weldon, Halifax Co. (USNM). James River System James River Drainage. Virginia: Calf- pasture River (Conrad, 1846: 407). Rappahannock River System Rapidan River Drainage. Virginia: Mountain Run, Orange Co. (MZUM). Potomac River System Potomac River Drainage. Pennsyl- vania: Great Tonoloway Creek, Thompson Township, Fulton Co. (Ortmann, 1919: 156). Conococheague Creek, Greencastle, Franklin Co. (Ortmann, 1919: 156). West Virginia: South Branch, Potomac River, Romney, Hampshire Co. (Ortmann, 1919: 157). Virginia: South River (upper Shen- andoah River), Waynesboro, Augusta Co. (Ortmann, 1919: 157). Potomac River, Great Falls Co. Park, Fairfax Co. District of Columbia: Aqueduct Lake. Anodonta (Pyganodon) gibbosa Say Plate 15:2,3 Anodonta gibbosa Say 1824 in W. H. Keating, Narrative of an Expedition to the Source of the St. Peters River, 2: 265, pi. 14, figs. 3-4 (South Carolina; type, ANSP [lost]. Type locality is here restricted to Altamaha River, Hopeton, near Darien, Mcintosh Co., Georgia. See: Remarks helow). Simpson, 1914, Cat. Naiades, 1: 397. Anodonta gibbosa Say, partim. Clench and Turner, 1956, Bull. Florida State Mus., 1: 184. Description. Shell medium to large, reaching 120 mm in length. Outline sub- elliptical. Valves very much inflated, thin and smooth. Anterior end regularly rounded; posterior end broadly rounded or somewhat pointed. Ventral margin usually broadly curved, though centrally some- times straight, or slightly incurved. Dorsal margin slightly curved, usually forming a distinct winglike angle where it meets the obliquely descending posterior margin. The posterior margin joins the ventral margin in a broad point near the medial line. Hinge ligament long and low. Posterior ridge broadly rounded, though occasionally with a trace of an angle. Posterior slope slightly concave. Umbos very high and broad, located in the anterior third of the shell, their sculpture consisting of a num- ber of low double-looped ridges. Perios- 360 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 tracum generally smooth and shiny, slightly roughened toward the margins and on the posterior slope. Surface of the shell straw yellow to greenish, older specimens becoming chestnut to blackish. Usually with very fine greenish rays; the several on the posterior slope usually darker and broader than those on the disk. No hinge plate or teeth; muscle scars inconspicuous and poorly defined. Nacre bluish white and iridescent, sometimes pinkish toward the umbonal cavities. Length Height Width mm mm mm 123 72 69 97 59 48 74 49 40 71 47 37 Altamaha River, 3 mi. NW Everett City, Glynn Co., Georgia. Cedar Creek, Fountains Mill, 7 mi. SW Hawkins- ville, Pulaski Co., Georgia. [after Say]. Altamaha River, Hopeton, near Darien, Mcintosh Co., Georgia. Habitat. Lives in soft mud and sand, also in sand bars; generally in slowly mov- ing water. Remarks. In the Atlantic Slope region, Anodonta gibbosa Say can only be con- fused with A. cataracta Say, but gibbosa has a very swollen shell, the umbos are greatly protruded, and the swelling con- tinues with the growth of the shell, the greatest swelling extending toward the posterior ventral margin. Frierson (1912: 129) correctly pointed out the relationship of this species to A. cataracta Say based on the beak sculpture, but thought gibbosa might be merely a subspecies. They ap- pear to be separate and both species occur in the Altamaha River system. It is obvious that Major Le Conte who sent T. Say the type of gibbosa made a mistake as to its locality, since there are no specimens of this species in any of the collections studied from South Carolina. Le Conte was to later send I. Lea numerous species he collected in the lower Altamaha River, so it is presumed that this species was also from this region, especially since a specimen in the Museum of Comparative Zoology from B. Walker, collected at Hope- ton on the lower Altamaha, has almost the same measurements as the type. To avoid confusion the restriction of the type lo- cality to this region seems warranted. Clench and Turner (1956: 185) included under this name the Anodonta of the Ochlockonee, Apalachicola and Chocta- whatchee River systems, but these Ano- donta are more elliptical and elongate, much less globose, and with the more tuberculous beak sculpture of Anodonta grandis Say, which they appear to be. Range. Southern Atlantic Slope: re- stricted to the Altamaha River system, Georgia. Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Cedar Creek, Fountains Mill, 7 mi. SW Hawkinsville, Pulaski Co. Ocmulgee River, 1 mi. S Lumber City, Telfair Co. Altamaha River Drainage. Georgia: Altamaha River, 10 mi. NE Surrency, Ap- pling Co. Altamaha River, "Riverside Park," 4 mi. N Jesup, Wayne Co. Altamaha River, 3 mi. NW Everett City, Glynn Co. Button- wood Swamp, Liberty [now Long] Co. Altamaha River, Hopeton, near Darien, Mcintosh Co. Anodonta {Pyganodon) implicata Say Plate 15: 4 Plate 16: 1,2 Anodonta implicata Say 1829, New Harmony [Indiana] Disseminator 2, no. 22, p. 340 (pond in Danvers [Essex Co.], Massachusetts; type, ANSP [lost]). The type locality was changed to Agawani River (outlet of Halfway Pond), Plymouth [Plymouth Co.], Massachusetts, and a neotype MCZ 176769 was selected by Johnson, 1946, Occ. Papers on Moll., 1: 112, pi. 16, figs. 1, 2). Simpson, 1914, Cat. Naiades, 1: 391. Ortmann, 1919, Mem. Carnegie Mus., 8: 159, pi. 11, figs. 2, 3. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem., no. 367, p. 40, fig. 42. Athearn and Clarke, 1962, Natl. Mus. Canada, Bull. no. 183, p. 26, pi. 2, figs. 1, 2. Atlantic Slope Unionidae • Johnson 361 Anodonta newtonensis Lea 1838, Trans. Amer. Philos. Soc, 6: 79, pi. 21, fig. 66 (Newtown [Newton] Creek, [Camden Co., New Jersey], opposite Philadelphia; figured type not in USNM or ANSP [lost]; Schuylkill [River], Fairmonnt [Philadelphia, Philadelphia Co., Pennsylvania], lectotype USNM 86561 selected by Johnson, 1946, Oec. Papers on Moll., 1: 112. [The selection of lectotypes, without figuring them, is a dubious contribution to knowledge, if, indeed, such selections are valid at all.] Figured here. Pi. 15: 4. Lea, 1838, Obs. Unio, 2: 79. Anodonta housatonica Linsley 1845, Amer. Jour. Sci., ser. 1, 48: 277 (Housatonic [River], Corum [near Huntington, Fairfield Co.], Con- necticut), [nomen nudum]. Gould, 1848, Amer. Jour. Sci., ser. 2, 6: 234, figs. 4, 5; measured and figured holotype USNM 678302. Description. Shell medium to large, reaching 165 mm in length. Outline elon- gate-elliptical to elongate-ovate. Valves quite inflated, somewhat subcylindrical, rather solid and heavy for an Anodonta. Anterior end regularly rounded; posterior end more acutely rounded, older speci- mens often becoming subtruncated. Ven- tral margin slightly rounded, becoming straight or slightly arcuate in old individu- als. Dorsal margin straight, occasionally forming a slight winglike angle where it meets the obliquely descending posterior margin. Hinge ligament long and low. Posterior ridge not very acute, but gener- ally distinctly biangulate. Posterior slope slightly concave. Umbos slightly swollen and moderately convex, located in the an- terior third of the shell, their sculpture consisting of five to seven double-looped bars, not appreciably lower in front or behind the sinus but of uniform elevation. Periostracum generally rather smooth, save for growth lines and occasional plaiting, yellowish brown or greenish brown, some- times becoming reddish brown or almost black in old specimens. Immatures are sometimes greenish and obscurely rayed. No hinge plate or teeth. Beak cavities rather shallow. Muscle scars and pallial line distinct especially in mature speci- mens. Shell distinctly thickened along the anterior margin below the pallial line. Nacre dull opalescent, generally pale cop- per, pinkish, or, more rarely, white or bluish white, sometimes with a bluish cast toward the margins. Females are usually more swollen than males in the middle portion of the disk, which tends to make the lower margin more curved and the shell somewhat shorter and higher than that of the male, but this is not a uniform character. Length Height Width mm mm mm 162 77 66 120 60 48 97 52 39 94 51 34 Agawam River, Plymouth, Plymouth Co., Massachu- setts. Male. As above. Neotype of A. implicata Say. Male. Schuylkill River, Fair- mount, Philadelphia Co., Pennsylvania. Lectotype of A. newtonensis Lea. Fe- male. Housatonic River, Corum, Fairfield Co., Connecticut. Holotype of A. housatonica Gould. Male. Anatomy. The anatomy of this species has not been investigated. The glochidia are rather large, subtriangular, with a spine at the tip of each valve. The host fish is the alewife, Alosa pseudoha renins (Wilson) (Johnson, 1946: 112). Breeding season. According to Ortmann (1911: 303) all Anodonta are bradytictic, or winter breeders, the marsupia filling sometime in summer between July and September. The glochidia are carried through the winter and set free when warmer weather begins. I collected gravid females at Plymouth, Massachusetts, on May 8, 1943, and June 9, 1946. Habitat. Lives in sand or gravel, rarely in mud. Found abundantly, close to the coast in ponds that have an unobstructed outlet to the ocean. Seems to prefer a stream environment; the largest examples have been taken in a relatively swift, sandy stream. Remarks. Anodonta implicata Say of the 362 Bulletin Museum of Comparative Zoology, Vol 140, No. 6 Northern Atlantic Slope can generally be easily separated from all other species of Anodonta. The most reliable and constant feature in identifying adults of this species is the pronounced thickening of the an- terior-ventral portion of the shell below the pallial line. This thickening is not al- ways pronounced in immatures, but may still be detected. Other characteristics of this species are the dark and usually yel- lowish or brownish rayless periostracum and the pale copper or salmon color of the nacre. Anodonta implicata is especially abun- dant in the region of Cape Cod, Massachu- setts, where it reaches its greatest size. The tide waters of the Potomac River, Virginia, seem to be its southern terminus. All the specimens under this name I have seen from rivers below the Potomac were heavy cataracta, which have an especially distinct pallial line anteriorly, but which lack the special anterior-ventral thickening of the shell and the characteristic nacre color. It is on the basis of such specimens that Atheam and Clarke (1962: 26) suggest that this species extends to South Carolina. Range. Northern Atlantic Slope: Potomac River, Maryland, north to New Brunswick and northern Nova Scotia, both Canada. Specimens Examined Potomac River System Potomac River Drainage. District of Columbia: Potomac River (USNM). Subgenus Utterbackia F. C. Baker Utterbackia F. C. Baker 1927, American Midland Nat., 10: 221, 222 (misspelled as Utterbackia on p. 221). Type species, Anodonta imbecillis [sic] Say. Orig- inal designation. Utterbackiana Frierson 1927, Check List North American Naiades, p. 17. Type species, Anodonta suborbiculata Say. Mono- typic. Under the subgenus Lastena Rafinesque (not available for use here since the type is Anodonta lata Rafinesque, teste Ort- mann and Walker, 1922: 32), Frierson included all of the taxa mentioned in the present paper, except A. suborbiculata Say of the Interior Basin for which he intro- duced the subgeneric name Utterbackiana, on the basis that that species is dioecious. But, as mentioned below, this as an un- reliable basis for classification and the shell morphology is clearly that of Utter- backia. Morrison (in Walter, 1956: 265) states that Anodonta imbecilis, like A. cygnea of Europe and Asia, is monoecious and has flat umbos, and that therefore imbecilis belongs to Anodonta s. s. Heard (1966: 31) has clearly shown that sexuality is an unreliable means of classi- fying Anodonta. Neither cygnaea nor im- becilis are uniformly monoecious. The flat umbos of Utterbackia and Anodonta s. s. appear to be a convergent character. Utter- backia is quite isolated from Anodonta s. s. in North America, the latter being restricted to the Pacific region. All four species of Ut- terbackia are more delicate than cygnea, and some individuals of each of the species exhibit fine rays toward the umbos. These are lacking in cygnea or any other Ano- donta. There are two species of Utterbackia in the Interior Basin, suborbicidata and im- becilis. The latter is found almost every- where to the east except Peninsular Florida and the Northern Atlantic Slope region. Speciation has taken place in the South- eastern states, where two species, in addition to imbecilis, occur. They are couperiana Lea, covered here, and A. peggyae Johnson ( 1965 ) , which occurs in the Apalachicolan region and Gulf drain- age of Peninsular Florida. Anodonta (Utterbackia) imbecilis Say Plate 16: 3 Anodonta (Lastena) ohiensis Rafinesque 1S20, Ann. Gen. Sci. Physiques (Bruxelles), 5: 316 (l'Ohio et toutes les rivieres adjacentes; sup- posed type, ANSP 216908, is Proptera laevissima (Lea)). Ortmann and Walker (1922, Occ. Papers, Mus. Zool. Univ. Michigan, no. 112, p. Atlantic Slope Unionidae • Johnson 363 37) clearly showed that A. ohiensis is not identi- fiable, nevertheless Frierson ( 1927, Check List North American Naiades, p. 18) persisted in using this name For A. imbecilis Say. Anodonta imbecilis Say 1829, New Harmony [Indiana] Disseminator, 2(23): 355. (Wabash River, Indiana; type, ANSP [lost]). Neotype Senckenberg Mus. 4301 [not seen] selected, but not figured, by Haas, 1930, Senckenberg- iana, 12: 326. Type locality restricted to Wabash River, New Harmony, Indiana, by Clench and Turner, 1956, Bull. Florida State Mus., 1: 188. Anodonta incerta Lea 1834, Trans. Amer. Philos. Soc, 5: 56, pi. 6, fig. 16 (Ohio River, near Cincinnati [Hamilton Co., Ohio]; type, not in USNM or ANSP [lost]. Lea, 1834, Obs. Unio, 1: 158. Anodon horda Could 1855, Proc. Boston Soc. Nat. Hist., 5: 229 (Comanche Creek [Tributary of The Llano River, close to the present site of Mason, Mason Co.], Texas [teste, Taylor, 1967, Veliger, 10: 153]; measured holotype USNM 678301). Anodonta henryana Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 102 (Matamoras, Tamaulipas [State], Mexico). Lea, 1860, Tour. Acad. Nat. Sci. Phila., ser. 2, 4: 373, pi. 66, fig. 198; figured holotype USNM 86692. Lea, 1860, Obs. Unio, 8: 55. Utterbackia imbeciUis [sic] fusca Baker 1927, American Midland Nat., 10: 222 (Sturgeon Bay, Door Co., Wisconsin; holotype MZUM 209141, figured by Baker, 1928, Bull. Wisconsin Geol. Nat. Hist. Survey, no. 70(2): 175, pi. 103, fig. 14). Anodonta imbeoillis [sic] Say. Simpson, 1914, Cat. Naiades, 1: 396. Anodonta henryana Lea. Simpson, 1914, Cat. Naiades, 1: 396. Anodonta ohiensis Rafinesque. Ortmann, 1919, Mem. Carnegie Mus., 8: 162. Anodonta imbecilis Say, partim. Clench and Turner, 1956, Bull. Florida State Mus., 1: p. 187. See: lohnson, 1965, Breviora, Mus. Comp. Zool., no. 213, pi. 2, fig. 5, 6. Anodonta (Utterbackia) imbecilis Say. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem. no. 367, p. 40, fig. 42. Description. Shell small to medium, usually not reaching over 90 mm in length. Outline rather elongated, subelliptical. Valves compressed, or only slightly swollen in young shells, but often becoming quite inflated with age, very thin, fragile, and smooth. Anterior end regularly rounded; posterior end somewhat pointed. Ventral margin slightly convex, often almost straight in the middle. Dorsal margin straight and long, usually forming a distinct angle where it meets the obliquely descending posterior margin. The posterior margin joins the curved ventral margin in a point near the medial line. Hinge ligament short and inconspicuous. Posterior ridge broadly rounded, though occasionally rather an- gular. Posterior slope slightly concave. Umbos low and broad, seldom extending above the dorsal margin, located in the anterior third of the shell, their sculp- ture consisting of a number of delicate subconcentric undulations. Periostracum smooth and shiny, except the posterior slope which may be slightly roughened. Surface of the shell straw yellow to green- ish yellow, but more generally darker green, sometimes grass green. Toward the umbos the color may be greyish, brownish or yellowish, with concentric darker growth-rests. The disk may be obscurely rayed; the posterior slope usually has two or three dark green to blackish rays. No hinge plate or teeth; muscle scars inconspicuous and poorly defined. Nacre bluish white and iridescent, sometimes pinkish toward the umbonal cavities. This species is hermaphroditic and does not show anv sexual differences. Length Height Width mm mm mm 90 84 71 42 44 39 34 31 24 Magnolia Springs, Mag- nolia State Park, Perkins, Jenkins Co., Georgia. Outlet of Waynesboro Lake, Waynesboro, Burke Co., Georgia. Buck Creek, 5 mi. NW Hawkinsville, Pulaski Co., Georgia. Anatomy. Clarke (1959: 42) discusses the work that has been done on the anat- omy of this species. He suggests that the host fish is Semotilus atromaculatus atro- maculatus (Mitchill). Breeding season. Ortmann (1919: 163) says this species is undoubtedly brady- tictic. 364 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Habitat. Lives in soft mud or sand in ponds, creeks, and near the banks of larger rivers. Remarks. In the Atlantic Slope region, Anodonta imbedlis Say can only be con- fused with A. coupcriana Lea and in the Apalachicolan region it can only be mis- taken for Anodonta peggyae Johnson. These three species of Anodonta are the only ones in these drainage systems whose umbos do not extend above the dorsal margin. In imbedlis the ventral margin is almost parallel to the dorsal one, while in couperiana the ventral margin is broadly curved, which renders the shell much higher in proportion to its length than imbecilis. See: Remarks under Anodonta couperiana Lea on p. 365. The general distribution of A. imbecilis has been discussed by Ortmann (1919: 165). It is found in the Apalachicolan region east to the Ochlockonee River system, Florida. In the Atlantic Slope region it does not extend below the Alta- maha River system, Georgia. It apparently does not occur in Peninsular Florida, since the author made an extensive collecting trip to this area in 1962 and did not find A. imbecilis. It is replaced on the Gulf side of Florida by Anodonta. peggyae John- son (1965: 1) and by A. couperiana Lea on the Atlantic side. (See: Johnson, 1965, pi. 1.) On the Atlantic slope, A. imbecilis extends to the Nottoway River, Virginia, of the Chowan River system. Ortmann (1919: 163) records this species from Pennsylvania in the region of Lake Erie, but not from the Atlantic Slope. It has since been found at Gunpowder River, Loch Raven Dam, Loch Raven, Baltimore Co., Maryland (ANSP). Range. Interior Basin: Mississippi drain- age generally, Ohio River, West Gulf Coastal region, Alabama-Coosa River system; and Apalachicolan region: Rio Grande River system, Texas, east to the Ochlockonee River system, Florida. South- ern Atlantic Slope: Altamaha River system, Georgia, north to the Chowan River system. Northern Atlantic Slope: Gunpowder River system, Maryland. Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Buck Creek, 5 mi. NW Hawkinsville; Cedar Creek, Fountains Mill, 7 mi. WSW Hawkinsville; both Pulaski Co. Ohoopee River Drainage. Georgia: Yam Grande Creek, 3 mi. W Swainsboro, Emanuel Co. Altamaha River Drainage. Georgia: Altamaha River, 4 mi. NE Jesup, Wayne Co. Ogeechee River System Ogeeehee River Drainage. Georgia: Magnolia Springs, Perkins, Jenkins Co. Mill Creek, 5 mi. E Midville; Barkcamp Creek, 7 mi. E Midville; both Burke Co. Savannah River System Savannah River Drainage. Georgia: Savannah River, 7.5 mi. NE Shellbluff; outlet of Waynesboro Lake, Waynesboro; Savannah River, 6 mi. NE Girard; all Burke Co. South Carolina: Savannah River, 6 mi. S Martin; Savannah River, Johnsons Landing, 10 mi. W Allendale; Savannah River, Kingjaw Point, 10 mi. WSW Allen- dale (ANSP); all Allendale Co. Cooper-Santee River System Catawba River Drainage. North Caro- lina: Bissels Pond, Charlotte, Mecklenburg Co. (ANSP; USNM). Wateree River Drainage. South Caro- lina: Wateree River, 2.5 mi. W Camden, Kershaw Co. Santee River Drainage. South Carolina: Santee River, above dam, near Cross; Charleston (USNM); both Berkeley Co. Pedee River System Yadkin River Drainage. North Caro- lina: Salem, Forsyth Co. (USNM). Atlantic Slope Unioniime • Johnson 365 Cape Fear River System Cape Fear River Drainage. North Caro- lina: Greenfield Mill Pond, Wilmington, New Hanover Co. (Ortmann, 1919: 164). Chowan River System Nottoway River Drainage. Virginia: Birchen Lake, on Birchen Creek, Camp Pickett, Nottoway Co. Gunpowder River System Gunpowder River Drainage. Maryland: Gunpowder River, Loch Raven Dam, Loch Raven, Baltimore Co. (ANSP). Anodonta {Utterbackia) couperiana Lea Plate 16:4 Plate 17: 1 Anodonta cowperiana [sic] Lea 1840, Proc. Amer. Philos. Soc, 1: 289 (Hopeton, near Darien [Mcintosh Co.], Georgia). Changed to: Anodonta couperiana Lea 1842, Trans. Amer. Philos. Soc., 8: 227, pi. 20, fig. 46; figured type, not in USNM [lost]. Lectotype, USNM 86673, selected by Johnson, 1965, Breviora, Mus. Comp. Zool., no. 213, p. 3, pi. 2, fig. 4. Lea, 1842, Obs. Unio, 3: 65. Anodonta dunlapiana Lea 1842, Proc. Amer. Philos. Soc, 2: 225 (South Carolina). Lea, 1842, Trans. Amer. Philos. Soc., 8: 248, pi. 27, fig. 65; figured type, not in USNM [lost]. Lectotype, here selected, USNM 86564, Pi. 17, fig. 1, [Charleston, Chatham Co.] South Caro- lina. Lea, 1842, Obs. Unio, 3: 86. Anodonta cowperiana [sic] Lea. Clench and Turner, 1956, Bull. Florida State Mus., 1: 183, pi. 6, fig. 3. Description. Shell medium to large, reaching 110 mm in length. Outline sub- elliptieal to subcircular. Valves somewhat inflated, thin, fragile, and smooth. Anterior end regularly rounded; posterior end somewhat pointed. Ventral margin broadly- curved. Dorsal margin straight and long, usually forming a distinct winglike angle where it meets the obliquely descending posterior margin. Hinge ligament short but prominent. The posterior margin joins the curved ventral margin in a point near the medial line. Posterior ridge broadly rounded. Posterior slope slightly concave. Umbos low and broad, seldom extending above the dorsal margin, located in the anterior third of the shell, their sculpture consisting of a number of delicate subcon- centric undulations. Periostraeum smooth and shiny, except the posterior slope which may be roughened. Surface of the shell straw yellow to yellowish green, usually with numerous and generally fine green rays, sometimes with distinctly darker rays on the posterior slope. No hinge plate or teeth; muscle scars inconspicuous and poorly defined. Nacre bluish white and iridescent. Length Height Width mm mm mm 110 67 49 Sandpit, near Edisto Biver, Givhans Ferry State Park, 2 mi. NW Givhans, Dor- chester Co., South Caro- lina. 87 50 35 Pond, near Savannah, Chatham Co., Georgia. 83 46 35 Charleston, South Carolina. Lectotype of A. dunlapiana Lea. 67 38 29 Hopeton, near Darien, Mc- intosh Co., Georgia. Lecto- type of A. couperiana Lea. Anatomy. Discussed by Lea (1863: 451). Habitat. Prefers sandy or muddy bot- toms of ponds and sluggish streams. Remarks: In the Southern Atlantic Slope region, Anodonta couperiana Lea can be confused only with A. imbecilis Say. Both species have umbos which do not extend above the dorsal margin, a character which distinguishes them from the other Ano- donta of the Atlantic Slope. A. couperiana is differentiated from imbecilis by its green rays, which are especially fine on the disk, and by its broadly curved ventral margin, which renders the shell much higher in proportion to its length than imbecilis, whose ventral margin is almost straight and parallel to the dorsal one. The height/length ratio of couperiana is about 2 to that of 1.5 in imbecilis. A. couperiana is found in the Apalach- 366 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 icola River system with A. peggyae John- son, which differs from couperiana by being subrhomboidal rather than elliptical in outline, and by being more biangulate pos- teriorly, with the biangulation ending near the base. A. couperiana is more pointed posteriorly with the point ending near the medial line. In A. peggyae, when the dorsal margin is held straight, the ventral margin is often obliquely descending toward the broad posterior basal biangulation. A. couperiana is a rare species in Geor- gia, but the specimens from the northern part of the range are remarkable for their size and beauty. It is common and abun- dant in Central Florida, where individuals tend to be small. Range. Apalachicolan region: Apalach- icola, Ochlockonee, and St. Marys river systems. Peninsular Florida. Southern At- lantic Slope: Altamaha River system, Geor- gia, north to the Cape Fear River system, North Carolina. Specimens Examined Saint Marys River System Saint Marvs River Drainage. Florida: St. Marys River (ANSP; MZUM). Altamaha River System Altamaha River Drainage. Georgia: Rice field ditches on the Altamaha River, Hopeton, near Darien, Mcintosh Co. Savannah River System Savannah River Drainage. South Caro- lina: Brickyard Pond, Hamburg, Aiken Co. (USNM). Savannah River [Johnsons Landing], 10 mi. W Allendale; Savannah River, Kingjaw Point, 10 mi. WSW Allen- dale (ANSP); both Allendale Co. Georgia: Pond near Savannah, Chatham Co. Combahee River System Salkehatchie River Drainage. South Carolina: Pauline Cave, near Kline, Barn- well Co. (ANSP). Edisto River System Edisto River Drainage. South Caro- lina: Sandpit, near Edisto River, Givhans Ferry State Park, 2 mi. NW Givhans, Dorchester Co. Charleston, Chatham Co. (USNM). Cooper-Santee River System Saluda River Drainage. South Carolina: Milton, Laurens Co. Cape Fear River System Cape Fear River Drainage. North Caro- lina: Greenfield Mill Pond, Wilmington, New Hanover Co. (USNM). Genus Strophitus Rafinesque Strophitus Rafinesque 1820, Ann. Gen. ties Sei. Physiques (Bruxelles), 5: 316. Type species, Anodonta undulata Say. Monotypic. Ortmann, 1912, Ann. Carnegie Mus., 8: 299. Pscudodontoideus Frierson 1927, Cheek List North American Naiades, p. 9, 23. Instituted as a subgenus. Type species, Margaritana alabamensis Lea. Original designation. Strophitus contains two species, undu- latus (Say), which is widely distributed throughout the Interior Rasin and Atlantic Slope region, and suhvexus (Conrad) (Johnson, 1967: 3, pi. 1, figs. 1-3), which extends from the eastern part of the West Gulf Coastal region through the Apalach- icolan region. All of the taxa included by Frierson under Strophitus s. s. and Vseudodontoideus Frierson are included in the synonymy of the two species mentioned above, except for: Strophitus undulatus tennesscensis Frierson, which may be a geographic subspecies in the modern sense; and Strophitus radiatus (Conrad), which belongs under Anodontoidcs radiatus (Conrad) (Johnson, 1967: 6, pi. 2, figs. 1-4). Generally, Frierson gave neither de- scriptions nor reasons for instituting new subgenera, but it is presumably on the basis of the vestigial pseudocardinal teeth found in alabamensis ( = subvexus) that Atlantic Slope Uniontoae • Johnson 367 he instigated Pseudodontoideus as a sub- genus. If the present synonymy of the species is correct, Strophitus does not need subdivision. Strophitus undulatus (Say) Plate 17: 2, 3 Anodonta undulata Say 1817, Nicholson's Encyclo- pedia, 2 [no pagination], pi. 3, fig. 6 (no type locality, here restricted: Schuylkill River, near Philadelphia, Philadelphia Co., Pennsylvania; type, AXSP [lost]). Anodonta pensylvanica [sic] Lamarck 1819, Hist. Nat. des Animaux sans Vertebres, 6: 86 (la riviere de Schuglkill [Schuylkill], pres de Phila- delphie [Philadelphia Co., Pennsylvania]; holo- type, Geneva Museum, figured by Delessert, 1841, Rec. Coquilles de Lamarck, pi. 13, figs. 4a, 4b). Anodon wgosus Swainson 1822, Zool. Illustrations, ser. 1, 2, pi. 96 (United States; type [location unknown] ) . Anodon areolatus Swainson 1829, Zool. Illustra- tions, ser. 2, 1, pi. 18 (North America; type [location unknown]). Alasmondonta edcntula Say 1829, New Harmony [Indiana] Disseminator 2, no. 22, p. 340 (Wabash River [near New Harmony, Posey Co., Indiana]; type, ANSP [lost]). Anodonta virgata Conrad 1836, Monography Unionidae, no. 5, back cover ( Buck Creek, Clarke Co., Ohio; type ANSP [lost]). Anodonta pavonia Lea 1836, Trans. Amer. Philos. Soc, 6: 78, pi. 21, fig. 65 (headwaters of the Little Beaver [River, Lawrence Co.], Ohio; figured holotype USNM 86514). Lea, 1838, Obs. Unio, 2: 78. Anodonta wardiana Lea 1838, Trans. Amer. Philos. Soc, 6: 46, pi. 14, fig. 42 ([Scioto River], near Chillieothe [Ross Co.], Ohio; figured holo- type USNM 86488). Lea, 1838, Obs. Unio, 2: 46. Anodon unadilla De Kay 1843, Zool. New York, Moll., pt. 5: 199, pi. 15, fig. 228 (Unadilla River, a tributary of the Susquehanna [River system], Otsego Co., New York; type, New York Lyceum of Nat. Hist, [destroyed by fire]). Anodonta tetragona Lea 1845, Proc. Amer. Philos. Soc, 4: 165 (Alexandria [Rapides Parish], Louisiana). Lea, 1848, Trans. Amer. Philos. Soc, 10: 82, pi. 8, fig. 25; figured holotype USNM 86682. Lea, 1848, Obs. Unio, 4: 56. Anodonta arkansensis Lea 1852, Trans. Amer. Philos. Soc, 10: 293, pi. 29, fig. 56 (Little Arkansas River, where the road to Santa Fe crosses it [Sedgwick Co., Kansas]; figured holo- type USNM 86603). Lea, 1852, Obs. Unio, 5: 49. Anodonta sliacffcriana Lea 1852, Trans. Amer. Philos. Soc, 10: 288, pi. 26, fig. 50 (Horn Lake Creek [Shelby Co.], Tennessee; figured holotype USNM 86685). Lea, 1852, Obs. Unio, 5: 44. Alasmodon rhombica Anthony 1865, Amer. Jour. Conch., 1: 158, pi. 12, fig. 5 (Michigan; figured holotype MCZ 50296). Anodon pa))ijracca Anthony 1865, Amer. Jour. Conch., 1: 161, pi. 15, fig. 2 (locality unknown [Potomac River, Virginia]; figured holotype MCZ 150656). Anodon annulatus Sowerby 1867, Conch. Iconica, 17, Anodon, pi. 18, fig. 67 (Hab. ?; Cuming collection in BMNH [lost]). Anodon quadriplicatus Sowerby 1867, Conch. Iconica 17, Anodon, pi. 28, fig. 110 (Potomac River, [Virginia]; holotype in Walpole coll., not in BMNH, [probably lost]). Anodonta sahnonia Clessin 1873, Conch. Cab., ser. 2, 9, pt. 1: 91, pi. 24, figs. 1, 2 (Norda- merika, im Ohio-und Wabash-Fluss ) ; Jion Lea 1838. Strophitus undulatus ovatus Frierson 1927, Check List North American Naiades, p. 22 (Middle West [Lyon Creek, Edwards Co., Illinois]; syn- type MZUM 87584). Strophitus rugosus pepinensis Baker 1928, Bull. Wisconsin Geol. Nat. Hist. Survey, no. 70(2), p. 204, pi. 74, fig. 8 (Lake Pepin near Lake City [Wabasha Co.], Minnesota; holotype MZUM 209137). Strophitus rugosus ivinnchagoensis Baker 1928, Bull. Wisconsin Geol. Nat. Hist. Survey, no. 70(2), p. 205, pi. 74, figs. 1-6 (Lake Winne- bago, Long Point Island, Wisconsin; holotype MZUM [original no. Univ. of Wisconsin 943a]). Stro))hitus rugosus lacustris Baker 1928, Bull. Wisconsin Geol. Nat. Hist. Survey, no. 70(2), p. 207, pi. 75, figs. 6-8 (Oconomowoc Lake, Waukesha Co., Wisconsin; holotype Univ. Illinois, Urbana Z-22073 [not seen]). Strophitus edentulus (Say). Simpson, 1914, Cat. Naiades, 1: 345. Ortmann, 1919, Mem. Car- negie Mus., 8: 197, pi. 12, figs. 7, 8. Strophitus undulatus (Say). Simpson, 1914, Cat. Naiades, 1: 349. Ortmann, 1919, Mem. Car- negie Mus., 8: 195, pi. 12, fig. 6. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem., no. 367, p. 43, fig. 41. Athearn and Clarke, 1962, Natl. Mus. Canada, Bull. no. 183, p. 29, pi. 3, figs. 7, 8. Descri))tion. Shell medium, seldom ex- ceeding 90 mm in length in the Atlantic boid to subrhomboid. Valves subinflated, generally rather thin, inequilateral. An- terior end regularly rounded; posterior end more broadly rounded, somewhat pointed 368 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 or subtruncated. Ventral margin slightly curved, occasionally incurved. Dorsal mar- drainage. Outline subelliptical, long rhom- gin straight, ending in a slight angle with the obliquely descending posterior margin. Hinge ligament not very prominent. Pos- terior ridge very rounded, but occasionally slightly angular with a trace of a secondary ridge above, ending in a slight biangulation near the base. Umbos rather full, located in the anterior third of the shell, their sculpture consisting of a few very strong ridges which turn up posteriorly where there are a few radial lirae. Surface of the shell with fine irregular growth lines, gen- erally smooth and shiny, though sometimes covered with brownish periostracum. When not heavily covered with periostracum, the shell may be yellowish, greenish or tawny, often with greenish rays over the entire surface or on the posterior slope. Each valve has vestiges of pseudo- cardinal teeth which are merely slight swellings, anterior to the umbos. The hinge plate is narrow, and slightly incurved in front of the umbos. Lateral teeth are ab- sent. Beak cavities shallow, with muscle scars on the hinge plate; anterior adductor muscle scars distinct, posterior ones faint. Pallial line scarcely visible. Nacre bluish, straw-colored or pinkish toward the umbos. Li-imth Height Width mm mm mm 88 74 47 55 44 29 24 27 17 Swift Creek, 3 mi. SSW Gainer, Wake Co., North Carolina. Turkey Creek, 8 mi. NW Edgefield, Edgefield Co., South Carolina. As above. Anatomy. Figured by Lea ( 1838, pi. 15, fig. 47). Discussed by Ortmann (1911: 299). The glochidia are subtriangular, with hooks, and measure 0.36 mm in length and 0.30 mm in height. They may either complete their development in the mar- supia of the parent or metamorphose while parasitizing the largemouth bass, Micro- pterus salmoides (Lacepede), or north- ern creek chub, Semotilus atromaculatus (Mitchill) (Baker, 1928: 201); both extend over the entire range of S. undulaius (Say). Breeding, season. From July to April or May of the following year. Habitat. Lives in small rivers and creeks in mud, sand or gravel. Remarks. Strophitus undulatus (Say), is the only species of Strophitus in the Interior Basin and Atlantic Slope region. The shell has no striking characters and might be confused with species of Ano- donta, since members of both genera lack lateral teeth. Anodonta have no pseudo- cardinals, and do have a straight hinge line; whereas Strophitus have vestigial pseudo- cardinals which appear as slight swellings anterior to the umbos, and a hinge line which is incurved in front of the umbos. S. undulatus is replaced in the Alabama- Coosa River system and Apalachicolan region by S. subvexus (Conrad) (Johnson, 1967: 2) which differs from undulatus by being distinctly rhomboidal, having a sharper posterior ridge, and by having at least one pseudocardinal in each valve elevated slightly above the hinge line. Strophitus undulatus might be confused with Anodontoides radiatus (Conrad) (Johnson, 1967: 6) in these areas, except that radiatus is more elliptical, has broad green rays, and each valve has a long, narrow pseudocardinal tooth which is slightly elevated and roughly parallel to the hinge line. Clarke and Berg (1959: 43) have al- ready discussed the necessity of discarding the name rugosus Swainson for this species. Ortmann (1919: 195) stated that undulatus was only a poorly defined variety of cdentulus (= rugosus), "connected by intergrades," and that they were the same species. He retained the two names, "thus avoiding inconvenience." Since only one species of Strophitus occurs in the Interior Basin and Atlantic Slope region, it is clear that, by priority, undulatus must be used. Range. Interior Basin: Mississippi and Ohio drainages, from Central Texas to Atlantic Slope Unionidae • Johnson 369 Lake Winnipeg, Canada. Atlantic Slope: tributary of the upper Savannah River system, South Carolina, north to the St. Lawrence River system, Canada. Specimens Examined Savannah River System Savannah River Drainage. South Caro- lina: Turkey Creek, 8 mi. NW Edgefield, Edgefield Co. Cooper-Santee River System Congaree River Drainage. Soutli Caro- lina: Congaree River. Cape Fear River System Cape Fear River Drainage. North Caro- lina: Rocky River, 11 mi. N Sanford, Chat- ham Co. Neuse River System Neuse River Drainage. North Carolina: Beaverdam Creek, 4.75 mi. S Creedmoor (57); Crabtree Creek, 6 mi. NW Raleigh (38); Neuse River, Pooles Bridge; Swift Creek, 3 mi. SSW Garner (20); all Wake Co. Pamlico River System Tar River Drainage. North Carolina: Tar River, 2 mi. W Springhope, Nash Co. Roanoke River System Roanoke River Drainage. Virginia: Mason Creek, Salem; Roanoke River, Salem; both (Ortmann, 1919: 204); both Roanoke Co. Chowan River System Nottoway River Drainage. Virginia: Nottoway River, 3 mi. E Rawlings, Bruns- wick Co. James River System James River Drainage. Virginia: Calf- pasture River, Goshen (Ortmann, 1919: 204); North [ = Maury] River, 2 mi. WNW Lexington; North River, Buena Vista (Ort- mann, 1919: 204); all Rockbridge Co. James River, opposite Maidens, Goochland Co. James River, Cartersville, Cumberland Co. York River System South Anna River Drainage. Virginia: [North Fork, S Anna River] Orange, Orange Co. Rappahannock River System Rapiclan River Drainage. Virginia: Rapidan River, Rapidan, Culpeper Co. (Ortmann, 1919: 204). Rappahannock River Drainage. Vir- ginia: Rappahannock River, Remington, Fauquier Co. Mountain Run, Culpeper, Culpeper Co. both (Ortmann, 1919: 204). Potomac River System Potomac River Drainage. West Vir- ginia: South Branch Potomac River, Rom- ney, Hampshire Co. (Ortmann, 1919: 204). Pennsylvania: Conochocheague Creek, Greencastle and Scotland; East Branch, Little Antietam Creek, Waynesboro; both (Ortmann, 1919: 202); both Franklin Co. Virginia: North Fork, Shenandoah River, E Woodstock, Shenandoah Co. South River, Waynesboro, Augusta Co. (Ortmann, 1919: 204). South Fork, Shenandoah River, Elk- ton, Rockingham Co. (Ortmann, 1919: 204). Maryland: Potomac River, Hancock, Washington Co. (Ortmann, 1919: 204). Potomac River, Paton Id., .75 mi. W Point of Rocks, Frederick Co. District of Co- lumbia: Aqueduct Lake. Subfamily Lampsilinae (Ihering 1901) Ortmann 1910 Genus Carunculina Baker Toxolasma Rafinesque 1831, Continuation of Monog. Bivalve Shells River Ohio (Phila.), p. 2. Species listed: Unio cyclips; U. cinerescens; U. lividus; U. flexus; all Rafinesque. Type species, Unio lividus Rafinesque. By elimi- nation, Frierson, 1914, Nautilus, 28: 7. Ortmann and Walker, 1922, Occ. Papers, Mus. Zool. Univ. Mich. no. 112, pp. 54, 55, show that U. lividus is a women duhium and that therefore Toxolasma must be disregarded. 370 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Corunculina [sic] Simpson 1898, in F. C. Baker,1 Bull. Chicago Acad. Sci., 3(1): 109. Type species, Unto parvus Barnes. Monotypic. Corunculina, corrected in die index and on errata sheet, and reconfirmed by Simpson, 1900, Proc. United States Natl. Mus., 22: 563. Ortmann, 1912, Ann. Carnegie Mus., 8: 377, partim. (Under Eunjnia.) Call (1896) monographed Carunculina, and indicated that it included only a few very variable species, and that most of its radiation was in the Interior Basin. He probably correctly reduced to synonymy many of the taxa subsequently recognized by Simpson (1914, 1: 148-161) and Frier- son (1927: 87-89). Carunculina pulla (Conrad) Plate 17: 4-7 U)iio pullus Conrad 1838, Monography Unionidae, no. 11, p. 100, pi. 55, fig. 2 (Wateree River, South Carolina; figured type ANSP [lost]; Warm Springs, [= Hot Springs, Madison Co.], North Carolina [Tennessee River system]; prob- ably either Villosa vancuxemensis (Lea) or Carunculina glans (Lea) (Ortmann and Walker, 1922: 55). Conrad (1838: 101) states that this specimen was seen subsequent to the prepa- ration of the plate. Therefore, the type locality is here restricted to the former locality. ) Lampsilis pulla (Conrad). Simpson, 1914, Cat. Naiades, 1: 160. Carunculina patrickae Bates 1966, Occ. Papers, Mus. Zool. Univ. Michigan, no. 646, pp. 1-9, 3 text figs., 1 pi. (Savannah River at approxi- mately mile point 134.5 U.S. Army Corps of Engineers Map [ = Johnsons Landing, 10 mi. W Allendale, Allendale Co.], on the South Carolina bank; holotype MZUM 85274). Carunculina pulla (Conrad). Johnson, 1967, Nautilus, 80: 127, pi. 10, figs. 1-4. Description. Shell small, seldom reach- ing over 35 mm in length. Outline of female long obovatc, of male, elliptical. Valves subinflated, generally thin, somewhat thickened anteriorly. Anterior end regu- larly rounded; posterior end of females more broadly rounded and subtruncated 1H. B. Baker (1964, Nautilus, 78: 33) has pointed out that since Simpson contributed noth- ing in the original publication of this genus (under article 51 (c), 1964 edition of the In- ternational Code Zool. Nomen.), the authority for Carunculina must be F. C. Baker. below the medial line, somewhat pointed in males. Ventral margin straight in males. In females marsujoial swelling causes the margin to be somewhat convex a little jdos- terior of the center. Dorsal margin slightly curved, forming a distinct angle with the obliquely descending posterior margin. Posterior ridge double, sometimes broadly rounded, but more often both ridges very angular. Posterior slojoe slightly concave. Umbos jorominent, not much elevated above the hinge line, located in the anterior third of the shell, their sculjDture consisting of several concentric sharp ridges parallel to the growth lines, which curve upward to the posterior ridge. Perios- tracum generally with distinct growth lines, often satiny, generally blackish, though sometimes brownish, greenish, olivish, and with obscure very fine green rays. Left valve with two raised triangular rjseudocardinal teeth, one in front of the other, occasionally crenulate. Hinge line short and very narrow before two short, straight, lateral teeth. Right valve with one rather chunky, triangular pseudo- cardinal; one lateral tooth. Beak cavities shallow, with a few dorsal muscle scars under the hinge plate. Anterior adductor muscle scars well imjDressed, posterior ones faint. Pallial line distinct anteriorly. Nacre bluish white, pink to purplish, iridescent. Length Height Width mm mm mm 32.0 19.0 15.5 Savannah River, 10 mi. W Allendale, Allendale Co., South Carolina. Fe- male. 22.0 1 3.5 10.0 As above. Male. 33.0 19.0 13.5 Wateree River, South Carolina. Male, [after Conrad]. 25.0 17.0 11.0 University Lake, 1 mi. W Chapel Hill, Orange Co., North Carolina. Habitat. Lives in shallow still water near the edges of streams and ponds, generally in mud, sometimes in sand. Remarks. In the Atlantic Slope region Carunculina pulla (Conrad) cannot be confused with any other species. The species of Villosa exhibit similar sexual dimorphism, but otherwise they are quite different from pulla. Villosa delumbis (Conrad) and vibex (Conrad) attain much larger size, have thinner, generally yellow- ish, distinctly rayed shells, whereas pulla is heavier for its size, blackish, and if rayed, only obscurely so. Villosa villosa ( Lea ) , on the other hand, is broadly rayed, though the rays are sometimes visible only in transmitted light. Villosa constricta (Conrad) attains much larger size, has a similarly heavy shell, but it is shiny, has no distinct posterior ridge, and the females have a distinct "constriction" in the post- basal region. Carunculina pulla most closely resembles Carunculina parva (Barnes) ( —C. paula (Lea) Clench and Turner, 1956: 193) which is found throughout the Interior Basin and Apalachicolan region. The um- bonal sculpture is similar, but pulla differs from parva by having heavy growth lines and a generally rougher periostracum; but more especially it differs by having a sharp posterior ridge, with a second, less promi- nent ridge above it. The posterior ridge is usually not present in typical parva, and when it is, it is generally not as acute as in pulla. Occasional specimens of C. pulla have an indistinct posterior ridge, as does one of the specimens figured. Nevertheless, the sharp posterior ridge is so generally present that pulla, with its relatively iso- lated distribution from parva, appears to be quite distinct from it. I have pointed out elsewhere (Johnson, 1967c) why it was not necessary to create a new taxon for this species. Carunculina parva (Lea) is abundant in Black Creek, northern Florida, but no Carunculina have been reported from either the St. Marys or Satilla river systems, which are between Black Creek and the Altamaha River system. To the north, no Caruncu- lina have been found on the Atlantic Slope beyond the Neuse River system. Atlantic Slope Unionidae • Johnson 371 It is likely that the ancestors of C. pulla entered the Southern Atlantic Slope region through a commingling of the headwaters of the Apalachicola and Savannah river systems. Range. Southern Atlantic Slope: Alta- maha River system, Georgia, north to the Neuse River system, North Carolina. Specimens Examined Altamaha RrvER System Ocmulgee River Drainage. Georgia: Ocmulgee River, below Lumber City, Tel- fair Co. (H. D. Athearn). Altamaha River Drainage. Georgia: [Altamaha River] Darien, Mcintosh Co. (USNM). Savannah River System Savannah River Drainage. South Caro- lina: Savannah River, approximately mile point 134.5 (U.S. Army Corps of Engineers Map ) = Johnsons Landing, 10 mi. W Allen- dale, Allendale Co. Santee River System Wateree River Drainage. North Caro- lina: [Headwaters of] Catawba River; Pfeiffers Pond; Stewarts Pond; Bissels Pond; Beaver Creek; all Charlotte, Meck- lenburg Co. (all ANSP). South Carolina-. Wateree River (Conrad). Cape Fear River System New Hope River Drainage. North Caro- lina: University Lake, an impoundment on Morgan Creek, 1 mi. W Chapel Hill, Orange Co. Neuse River System Neuse River Drainage. North Carolina: Neuse River, Raleigh, Wake Co. (Lea). Genus Villosa Frierson Micromya Agassiz 1852, Archiv fur Naturge- schichte, 18 (1): 47. Species listed: Unio lapillus Say, Mar gar it ana fahula Lea, M. cur- reyana Lea, non Micromya Rondani 1840 (Insecta). 372 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Type species, Unio lapillus Say. Subsequent designation, Herrmannsen, 1852, Indicis Gene- rum Malacozoorum, Supp. et Corr., p. 83. Ortmann, 1912, Ann. Carnegie Mus., 8: 337. partim. Villosa Frierson, 1927, Check List North Ameri- can Naiades, pp. 11, 80. Type species, Unio villosus Wright. Original designation. At this writing, it is impossible to tell how many species there are in Villosa. Frierson (1927: 70-79) includes under Lampsilis, subgenus Ligumia, many taxa which have been included by other authors under Micromya ( —Villosa). Except for the type species, the taxa listed by Frier- son '(1927: 80, 81) under Villosa are species of Carunculina. In spite of the state of the synonymy of this genus, it is clear that the majority of its species occur in the Interior Basin. Villosa villosa (Wright) Plate 17: 8, 9 * Unio villosus B. H. Wright 1898, Nautilus, 12: 32 (Suwannee River [Luraville], Suwannee Co., Florida; syntype USNM 150503, figured by Simpson, 1900, Proc. Acad. Nat. Sci. Phila., p. 77, pi. 1, fig. 1, selected as lectotype by Johnson, 1967, Occ. Papers on Moll., 3: 9, pi. 8, fig. 1). Lampsilis villosus (Wright). Simpson, 1914, Cat. Naiades, 1: 143. Villosa villosa (Wright). Clench and Turner, 1956, Bull. Florida State Mus., 1: 213, pi. 4, fig. 2. Description. Shell usually small in size, seldom exceeding 60 mm in length. Outline long, elliptical. Valves subinflated, gener- ally thin and translucent. Anterior end regularly rounded; posterior end of females slightly more broadly rounded, that of males quite pointed. Ventral margin al- most always broadly curved except in females where a slight marsupial swelling, somewhat posterior of the center, renders it straight or slightly convex. Dorsal mar- gin straight, with a very slight, if noticeable, angle where it meets the obliquely de- scending posterior margin. Hinge ligament small. Posterior ridge broadly rounded, double in the male; obscured by a slight marsupial swelling in the female. Posterior slope slightly concave. Umbos moderately swollen, slightly elevated above the hinge line, located in the anterior quarter of the shell, their sculpture consisting of several fine, low, slightly double-looped ridges. Surface of the shell with irregular growth lines, occasionally smooth and shiny, but usually covered with either rough or dis- tinctly satiny periostracum, especially on the posterior slope. Periostracum some- times subshiny, greenish yellow, dark greenish, or more often brownish black, the entire surface of the shell with broad green rays interspersed with narrow ones, some- times only visible in transmitted light. Left valve with two delicate pseudo- cardinal teeth, one in front of the other, the anterior one somewhat triangular, the hinder one inclined to be vestigial. Hinge line short and narrow before two short, straight lateral teeth. Right valve with two triangular, narrow, parallel pseudo- cardinals separated by a narrow pit, the more anterior tooth quite vestigial, some- times absent; one low lateral tooth. Beak cavities shallow, a few dorsal muscle scars under the hinge plate. Anterior adductor muscle scars well impressed, posterior ones faint, if visible. Pallial line distinct an- teriorly. Nacre bluish white, occasionally yellowish white, iridescent, especially pos- teriorly. Length Height Width mm mm mm 57 28 56 29 51 27 18 20 20 Suwannee River, Luraville, Suwannee Co., Florida. Lectotype. Female. St. Marys River, Nassau Co., Florida. Female. Spring Creek, Decatur Co., Georgia. Male. Habitat. "Limited to spring-fed streams and clear rivers," Clench and Turner (1956: 214); but their report did not in- clude the rather acid and muddy St. Marys River. Remarks. Villosa villosa (Wright) bears a resemblance to Villosa oibex (Conrad), which is more widely distributed and ex- tends over the whole range of V. villosa, as they both show similar sexual di- morphism. In the female of villosa there is, in general, less tendency for the post- basal swelling to extend below the ventral margin, and if inclined to be somewhat posteriorly pointed, the point is higher. The male is pointed posteriorly as in vibex, but the shell is proportionately longer. V. villosa often has a distinctive roughened periostracum which produces a satiny luster. Range. Apalachicolan region: Apalach- icola River system, east to the St. Marys River system, Georgia. Peninsular Florida. Specimens Examined St. Marys River System St. Marys River Drainage. Georgia: St. Marys River, Traders Hill, Charleton Co. (MZUM). Florida: St. Marys River, Nas- sau Co. (USNM). Villosa vibex (Conrad) Plate 17: 10,11 Plate 18: 1-3 Unio vibex Conrad, [May] 1834, New Fresh Water Shells United States, p. 31, pi. 4, fig. 3 (Black Warrior River, South of Blount's Spring [Blount Co.], Alabama; figured holotype ANSP 56488a ) . Published in May, teste Conrad, 1853, Proc. Acad. Nat. Sci. Phila., 6: 243. Unio modioliformis Lea, [August or September] 1834, Trans. Amer. Pbilos. Soc, 5: 97, pi. 13, fig. 40 ( Santee Canal, South Carolina; probable figured holotype USNM 85029 [differs slightly from figure]). Lea, 1834, Obs. Unio., 1: 209. Published in August or September, teste Lea, 1854, Proc. Acad. Nat. Sci. Phila., 7: 244. Unio exiguns Lea 1840, Proc. Amer. Philos. Soc, 1: 287 ( Cbattaboochee River, near Columbus [Muscogee Co.], Georgia). Lea, 1842, Trans. Amer. Philos. Soc, 8: 191, pi. 7, fig. 1; figured holotype USNM 84974. Lea, 1842, Obs. Unio, 3: 29. Unio stagnalis Conrad 1849, Proc. Acad. Nat. Sci. Phila., 4: 152 (inhabits mill ponds, Ogeechee Atlantic Slope Unionidae • Johnson 373 River, Georgia, J. II. Couper [loaned]). Con- rad, 1850, Jour. Acad. Nat. Sci. Phila., ser. 2, 1: 275, pi. 37, fig. 2; figured holotype MCZ 178778, purchased from J. II. Couper. Unio prevostianus Lea 1852, Trans. Amer. Philos. Soc, 10: 269, pi. 19, fig. 24 (Eutowah [Etowah] River [North West], Georgia; figured holotype, C. M. Wheatley collection in ANSP [lost]). Lea, 1852, Obs. Unio, 5: 25. Unio nigrinus Lea 1852, Trans. Amer. Philos. Soc, 10: 284, pi. 24, fig. 44 (West Florida; figured holotype USNM 86132). Lea, 1852, Obs. Unio, 5: 40. Unio gracilior Lea 1856, Proc. Acad. Nat. Sci. Phila., 8: 262 (Buckhead Creek [Burke Co.]; Tobesaufke [Tobesofkee] Creek near Macon [Bibb Co.]; both Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 56, pi. 8, fig. 38; figured holotype USNM 85088 [localities not separated]. Lea, 1858, Obs. Unio, 6: 56. Unio rutilans Lea 1856, Proc. Acad. Nat. Sci. Phila., 8: 262 Othcalooga [Oothkalooga] Creek, Gordon Co.; Columbus [Muscogee Co.]; both Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 59, pi. 9, fig. 41; figured holotype USNM 85093 from [Oothkalooga] Creek. Lea, 1858, Obs. Unio, 6: 59. Unio subcUipsis Lea 1856, Proc. Acad. Nat. Sci. Phila., 8: 262 (creeks near Columbus [Mus- cogee Co.], Georgia). Lea, 1858, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 62, pi. 10, fig. 44; figured holotype USNM 85095. Lea, 1858, Obs. Unio, 6: 62. Unio sudus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 170 (Dry Creek, near Columbus [Muscogee Co.]; Macon [Bibb Co.]; both Georgia). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 194, pi. 21, fig. 77; figured holotype USNM 85155 from Dry Creek. Lea, 1859, Obs. Unio, 7: 12. Unio obfuscus Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 172 (Flint River, near Macon [Co.], Georgia). Lea, 1859, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 197, pi. 22, fig. 80; figured holotype USNM 85089. Lea, 1859, Obs. Unio, 7: 15. Unio dispar Lea 1860, Proc. Acad. Nat. Sci. Phila., 12: 305 (Columbus [Muscogee Co.], Georgia). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 327, pi. 51, fig. 153; figured holotype USNM 85101. Lea, 1860, Obs. Unio, 8: 9. Unio averillii Wright 1888, Proc. Acad. Nat. Sci. Phila., 40: 115, pi. 3, fig. 4 (Lake Ashby, Volusia Co., Florida; syntype figured by Simpson, 1892, Proc. United States Natl. Mus., 15: 414, pi. 56, fig. 6 [not located]). Lectotype USNM 91142, 374 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 selected by Johnson, 1967, Occ. Papers on Moll., 3: 5, pi. 7, fig. 4. Villosa vibex (Conrad). Clench and Turner, 1956, Bull. Florida State Mus., 1: 209, pi. 4, fig- 4. Description. Shell usually small in size, not exceeding 60 mm in length, though occasionally reaching 100 mm. Outline subelliptical. Valves subinflated, generally thin and translucent. Anterior end regu- larly rounded; posterior end of females more broadly rounded, somewhat pointed in males. Ventral margin straight or slightly curved in males, often slightly arcuate in females. Dorsal margin straight with a very slight, if noticeable, angle where it meets the obliquely descending posterior margin. Hinge ligament small. Posterior ridge broadly rounded. Posterior slope slightly concave, occasionally with faint wrinkles and ridges. Umbos moder- ately swollen, slightly elevated above the hinge line, located in the anterior quarter of the shell, their sculpture consisting of several fine, low, slightly double-looped ridges. Surface of the shell generally rather smooth, but roughened by periostracum posteriorly. Periostracum usually subshiny, greenish yellow, yellowish brown to almost black, the entire surface with numerous broad greenish rays, which in darker speci- mens can be seen in transmitted light. Left valve with two delicate pseudo- cardinal teeth, one in front of the other, the anterior one somewhat triangular, the hinder one inclined to be vestigial. Hinge line rather long and very narrow before two short, straight lateral teeth. Right valve with two triangular, narrow, parallel pseudoeardinals separated by a narrow pit, the more anterior tooth vestigial, sometimes absent; one lateral tooth. Beak cavities shallow, a few dorsal muscle scars under the hinge plate. Anterior adductor muscle scars well impressed, posterior ones faint, if visible. Pallial line distinct anteriorly. Nacre bluish white, sometimes pinkish or purple, iridescent posteriorly. Length Height Width mm mm mm 100 48 32 52 29 17 48 27 17 Ochee Creek, about 5 mi. SE Toomsboro, Wilkinson Co., Georgia. Male. Mill Race, 2 mi. N Sardis, Burke Co., Georgia. Female. As above. Male. Habitat. Lives in mud or soft sand, particularly where rich in vegetable de- tritus, in small rivers and creeks. Remarks. In the Southern Atlantic Slope region, Villosa vibex ( Conrad ) can be con- fused with Villosa delumbis (Conrad) with which it is often found living. V. vibex has broader, less distinct, green rays which are less inclined to be broken by sharp growth rests than those of delumbis. Sexual di- morphism is not as strongly developed in vibex. The males of both species tend to be somewhat pointed posteriorly. Females of vibex tend to be broadly rounded, rendering the shell slightly arcuate, while the females of delumbis are greatly in- flated, with the posterior margin subangu- late dorsally and truncate below. In the Apalachieolan region, vibex can be con- fused with Villosa lienosa (Conrad) which is allopatric with V . delumbis. See Remarks under Villosa delumbis (Conrad) on p. 376. Range. West Gulf Coastal region, Alabama-Coosa River system, and Apalach- ieolan region: Pearl River system, Mis- sissippi, east to the Suwanee River system, Florida. Peninsular Florida. Southern At- lantic Slope: Altamaha River system, Geor- gia, north to the coastal ponds of the Cape Fear River system, North Carolina. Specimens Examined Altamaha River System Ofimilgee River Drainage. Georgia: Tobesofkee Creek, near Macon, Bibb Co. (USNM). Flat Creek, 2 mi. S Perry, Houston Co. Ruck Creek, 5 mi. NW Hawk- insville; Limestone Creek, 4.3 mi. E Ilawkinsville; Tucsawatchee Creek, 5 mi. SW Hawkinsville; Cedar Creek, Fountains Mill, 7 mi. SW Hawkinsville; Mosquito Creek, 8 mi. SE Hawkinsville; all Pulaski Co. Brushy Creek, 8 mi. NW Abbeville, Wilcox Co. House Creek, Bowens Mill, 9 mi. N Fitzgerald; Dieksons Creek, 10 mi. NE Fitzgerald; both Ben Hill Co. Little Ocmulgee River Drainage. Geor- gia: Little Ocmulgee Biver, 7 mi. NE Cochran, Bleckley Co. Alligator Creek, 2 mi. W Alamo, Wheeler Co. Gum Swamp Creek, 1 mi. N McBae, Telfair Co. Oconee River Drainage. Georgia: Ochee Creek, about 5 mi. SE Toomsboro; Turkey Creek, 4 mi. NE Allentown; both Wilkinson Co. Ford Branch, 4 mi. W Dublin; Bocky Creek, 8 mi. W Dudley; both Laurens Co. Ohoopee River Drainage. Georgia: Little Ohoopee Biver, 12 mi. NE Wrights- ville; Little Ohoopee Biver, 2 mi. N Kite; Ohoopee Biver, 4 mi. S Wrights ville; all Johnson Co. Ohoopee Biver, Norristown, Emanuel Co. Ogeechee Biver System Ogeechee River Drainage. Georgia: Bocky Comfort Creek, 1 mi. N Louisville; Williamson Swamp Creek, Bartow; Nails Creek, 2 mi. S Bartow; Bocky Creek, 2 mi. S Wadley; all Jefferson Co. Barkcamp Creek, 7 mi. E Mid ville; Mill Creek, 5 mi. E Midville; Buckhead Creek, 4 mi. E Vidette; Bocky Creek, 5 mi. SW Waynes- boro; all Burke Co. Chew Mill Creek, 8 mi. W Millen, Jenkins Co. Savannah Biver System Savannah River Drainage. Georgia: Brier Creek, 7.5 mi. NE Wrens, Jefferson Co. Mill Bace, 2 mi. N Sardis; Beaverdam Creek, 7 mi. NW Girard; both Burke Co. COMBAHEE BlVER SYSTEM Salkehatchee River Drainage. South Carolina: Little Salkehatchee Biver, 4 mi. N Ehrhardt, Bamberg Co. Salkehatchee Biver, Broxton Bridge; Whippy Swamp Creek, 2.5 mi. NE Crocketville; both Hampton Co. Atlantic Slope Unioniuak • Joluison 375 Edisto Biver System Edisto River Drainage. SoutJi Caro- lina: Edisto Biver. Cooper-Santee Biver System Wateree River Drainage. South Caro- lina: Wateree Biver, 2.5 mi. W Camden, Kershaw Co. Santee River Drainage. South Caro- lina: Lake Moultrie, Cross, Berkeley Co. (USNM). Santee Canal (Lea). Pedee Biver System Lynches River Drainage. South Caro- lina: Lynches Biver, 2 mi. NE Bishopville, Lee Co. Waccamaw Biver System Waccamaw River Drainage. North Carolina: Lake Waccamaw, [town of] Lake Waccamaw, Columbus Co. Cape Fear Biver System Cape Fear River Drainage. North Caro- lina: Sprunts Pond [not located] ( USNM ) ; stream below Greenfield Mill Pond, Wil- mington (ANSP); both New Hanover Co. Villosa delumbis (Conrad) Plate 18:4-8 Unio delumbis Conrad 1834, New Fresh Water Shells United States, p. 35, pi. 5, fig. 3 (small streams near Cooper River, South Carolina; figured type not in ANSP [lost]). Unio tenerus Ravenel 1834, Cat. Recent Shells Cabinet Edmund Ravenel, p. 7 (Cooper River, South Carolina) [nomen nudum]. Simpson 1892, Proe. United States Natl. Mus., 15: 416, pi. 58, figs. 5, 8 (South Carolina; figured syn- type USNM 85030 [male]; [Savannah River], Georgia; figured syntype USNM 85032 [female]). Walker, 1919, Oec. Papers, Mus. Zool. Univ. Mich. no. 74, p. 1, pi. 1, figs. 1-6. Unio vaughanianus Lea 1838, Trans. Amer. Philos. Soc, 6: 5, pi. 3, fig. 5 (Sawney's Creek, near [about 8 mi. NW] Camden [Kershaw Co.], South Carolina; two syntypes USNM 86106, one a bit smaller than the figure, the other with the outline of the figure, but with less prominent rays. The figure appears to be a composite.) Lea, 1838, Obs. Unio, 2: 5. Unio ogeecheensis Conrad 1849, Proc. Acad. Nat. 376 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Sci. Phila., 4: 153 (Ogeechee River, Georgia, T. H. Couper [loaned]). Conrad, 1850, Jour. Acad. Nat. Sei. Phila., ser. 2, 1: 275, pi. 37, figs. 3-4; figured syntype, fig. 3, MCZ 146971 purchased from J. H. Couper, selected as lecto- type by Johnson, 1956, Bull. Mus. Comp. Zool., 115: 126. Unio concavus Lea 1S52, Trans. Amer. Philos. Soc, 10: 260, pi. 15, fig. 11 (Abbeville District [Savannah River drainage], South Carolina; figured holotype USNM 85154). Lea, 1852, Obs. Unio, 5: 16. Unio proximus Lea 1852, Proc. Amer. Philos. Soc, 5: 252 (Georgia). Lea, 1852, Trans. Amer. Philos. Soc, 10: 271, pi. 20, fig. 27; figured holotype USNM 85131. Lea, 1852, Obs. Unio, 5: 27. Unio contiguus Lea 1861, Proc. Acad. Nat. Sci. Phila., 13: 392 (Stewarts Mill Dam, Union Co., North Carolina). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 199, pi. 28, fig. 268; figured holotype USNM 85120. Lea, 1863, Obs. Unio, 9: 21. Unio genuinus Lea 1868, Proc. Acad. Nat. Sci. Phila., 20: 161 (Bissels Pond, Charlotte [Meck- lenburg Co.] North Carolina). Lea, 1868, Jour. Acad. Nat. Sci. Phila., ser. 2, 6: 305, pi. 46; fig. 117, figured holotype USNM 85123. Lea, 1869, Obs. Unio, 12: 64. Description. Shell usually small in size, seldom exceeding 60 mm in length. Outline of female long obovate; of male long ellip- tical. Valves subinflated, generally thin and translucent. Anterior end regularly rounded; posterior end of females more broadly rounded and subtruncated below the medial line, somewhat pointed in males. Ventral margin straight in males, in females marsupial swelling causes the margin to be somewhat convex a little posterior of the center. Dorsal margin straight forming a sharp angle with the obliquely descending posterior margin. Posterior ridge broadly rounded. Posterior slope slightly concave, occasionally with faint ridges and wrinkles. Umbos moderately swollen, slightly ele- vated above the hinge line, located in the anterior third of the shell, their sculpture consisting of several fine, low, slightly double-looped ridges. Surface of the shell generally rather smooth and shiny, but roughened by periostraeum posteriorly especially on the posterior slope. Perios- traeum often subshiny, greenish yellow, brownish, or greenish, the entire surface with narrow, sometimes very narrow, greenish rays which are broken by growth rests to form concentric bands. Left valve with two delicate pseudo- cardinal teeth, one in front of the other, the anterior one somewhat triangular, the hinder one inclined to be vestigial. Hinge line rather long and very narrow before two short, straight lateral teeth. Right valve with two triangular, narrow, parallel pseudo- cardinals separated by a narrow pit; the more anterior tooth vestigial, sometimes absent; one lateral tooth. Beak cavities shallow, a few dorsal muscle scars under the hinge plate. Anterior adductor muscle scars well impressed, posterior ones faint, if visible. Pallial line distinct anteriorly. Nacre sometimes dirty white or pinkish, but usually bluish white and iridescent. Length Height Width mm mm mm 64 57 53 39 35 30 30 19 19 Ogeechee River, Georgia. Lectotype of U. ogeeche- ensis Conrad. Female. Mill Race, 2 mi. N Sardis, Burke Co., Georgia. Female. As above. Male. Habitat. Lives in mud or soft sand, particularly where rich in vegetable detritus, in small rivers and creeks. Remarks. In the Southern Atlantic Slope region, Villosa delumbis (Conrad) can be confused only with Villosa vibex (Con- rad) under which see: Remarks on p. 374. Villosa delumbis (Conrad) is very close to Villosa lienosa ( Conrad ) of the Apalach- icolan region. Both species are similarly shaped and the females show the same sexual dimorphism. V. lienosa has a much heavier shell; the posterior ridge, while rounded, tends to be faintly biangulate. The periostraeum is usually black or very dark yellowish green, and when visible the rays are solid green. The nacre is often pinkish or coppery. V. delumbis has a thinner shell, there is scarcely ever a hint of the posterior ridge1 being faintly bi- angulate, the periostracum is usually a light yellow or light greenish yellow, and the rays are almost always distinct and char- acteristically broken by growth rests. The nacre is very seldom pinkish. Although the figured type of V. delumbis has been lost, the life size figure, with its distinct, narrow rays, clearly distinguishes it from vibex, and it corresponds almost exactly in outline with the male shell figured in this report. Range. Southern Atlantic Slope: Alta- maha River system, Georgia, north to the Neuse River system, North Carolina. Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Macon, Bibb Co. Echeconnee Creek, Eche- connee, Houston Co. Buck Creek, 5 mi. NW Hawkinsville; Limestone Creek, 4.3 mi. E Hawkinsville; Tuesawhatchee Creek, 5 mi. SW Hawkinsville; Cedar Creek, 8 mi. SW Hawkinsville; all Pulaski Co. Oconee River Drainage. Georgia: Ochee Creek, about 5 mi. SE Toomsboro; Turkey Creek, 4 mi. NE Allentown; both Wilkinson Co. Ford Branch, 4 mi. W Dub- lin; Palmetto Creek, 7 mi. S Dublin; both Laurens Co. Ohoopee River Drainage. Georgia: Ohoopee River, 4 mi. S Wrightsville; Little Ohoopee River, 1 mi. E Kite; both Johnson Co. Ohoopee River, 1 mi. E Adrian; Ohoopee River, Norristown; Yam Grande Creek, 3 mi. W Swainsboro; all Emanuel Co. Altamaha River Drainage. Georgia: Darien, Mcintosh Co. (USNM). Ogeechee River System Ogeechee River Drainage. Georgia: Ogeechee River, Shoals, Warren Co. Ogeechee River, 4 mi. SW Mitchell, Glas- cock Co. Williamson Swamp Creek, Bar- tow, Jefferson Co. Barkcamp Creek, 7 mi. E Midville; Mill Creek, 5 mi. E Midville; Atlantic Slope Uniontoae • Johnson 377 both Burke Co. Chew Mill Creek, 8 mi. W Millen, Jenkins Co. Ogeechee River, Scarboro; Ogeechee River, bridge 1 mi. 5 Dover; Ogeechee River, 1.5 mi. SW Oliver; all Screven Co. (all MZUM). Ogeechee River, bridge 1.5 mi. E Blitch- ton; Ogeechee River, Jinks Bridge [Rte. 16]; Ogeechee River, Morgan Bridge, 14 mi. SE Pembroke; Ogeechee River, mouth of Arnold Lake; Jones Lake; all Brvan Co. (all MZUM). Savannah River System Broad River Drainage. Georgia: Broad River, Elbert, Huguenot Co. (USNM) Broad River, Lincoln Co. Savannah River Drainage. Georgia: Savannah River, Augusta, Richmond Co. South Carolina: Savannah River, 2 mi. SW Millettville (MZUM); Savannah River, Johnsons Landing, 10 mi. W Allendale; both Allendale Co. Georgia: Savannah River, 7.5 mi. NE Shell Bluff; Brier Creek, 6 mi. N Waynesboro; outlet of Waynesboro Lake, Waynesboro; Mill Race, 2 mi. N Sar- dis; all Burke Co. Combahee River System Salkehatchee River Drainage. South Carolina: Pauline Cave, near Kline, Barn- well Co. (MZUM). Edisto River System Edisto River Drainage. South Caro- lina: Snake Swamp Creek, Orangeburg Co. 6 mi. NE Bamberg, Bamberg Co. Edisto River, Canadys Landing, 8 mi. SW St. George, Dorchester Co. Edisto River, 1 mi. E Jacksonboro, Colleton Co. (MZUM). Cooper-Santee River System Congaree River Drainage. South Caro- lina: Congaree River. Catawba River Drainage. North Caro- lina: Long Creek; Catawba River; Bissels Pond, Charlotte; Pfeiffers Pond, Charlotte (ANSP); Flannigans Pond, Charlotte (ANSP); all Mecklenburg Co. [Ten Mile 378 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Creek] of Catawba River, Wolfsville [7 mi. W Monroe], Union Co. (MZUM). Wateree River Drainage. South Caro- lina: Sawneys Creek, about 8 mi. NW Camden, Kershaw Co. (USNM). Cooper River Drainage. South Caro- lina: near Cooper River (Conrad). Pedee River System Yadkin River Drainage. North Caro- lina: Stewarts Mill Dam, Union Co. (USNM). Waccamaw River System Waccamaw River Drainage. North Carolina: drainage canal beside Lake Waccamaw, 1 mi. NNW Dupree Landing, Columbus Co. Cape Fear River System Cape Fear River Drainage. North Caro- lina: University Lake, an impoundment on Morgan Creek, 1 mi. W Chapel Hill, Orange Co. Rocky River, 11 mi. N Sanford, Chatham Co. (MZUM). North East Cape Fear River, Dublin Co. (MZUM). Green- field Mill Pond, Wilmington, New Hanover Co. (USNM). Neuse River System Neuse River Drainage. North Carolina: North Flat River, 5.5 mi. S Roxboro Center, Person Co. (81). Neuse River (USNM). Villosa constricta (Conrad) Plate 18: 9-10 Unio lienosus constrictus Conrad 1838, Mono- graphy Unionidae North America, no. 10, p. 91, pi. 49, fig. 4 ( North [ = Maury] River, Rock- bridge Co., Virginia; figured holotype ANSP 20423, now ANSP 56465a.) Unio genthii Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 85 (Catawba River, Caston Co.; Deep River, Gulf [Chatham Co.]; both North Caro- lina). Lea, 1862, Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 57, pi. 2, fig. 204; figured holotype USNM 84834, labeled, Roanoke River, Raleigh, North Carolina, Prof. Emmons. (This locality is nonexistent. There is no way to tell which locality the type was from; so it is assumed here that since it was probably Prof. Emmons' specimen that was figured, it was from Deep River, Gulf, Chatham Co., North Carolina, and the type locality is, here, so restricted. ) Lea, 1862, Obs. Unio, 8: 61. Lampsilis constricta (Conrad). Simpson, 1914, Cat. Naiades, 1: 111. Eunjnia (Micromya) constricta (Conrad). Ort- mann, 1915, Nautilus, 29: 66. Description. Shell small in size, seldom exceeding 50 mm in length. Outline of female short obovate; of male subelliptical. Valves not much inflated, rather heavy and strong. Anterior end regularly rounded; posterior end of females more broadly rounded and pointed a little more than midway up from the base, truncated below the medial line. In mature specimens there is a rather distinct "constriction" in the middle of this truncation which marks the boundary between the branchial opening and the papillar part of the mantle edge in front of it; somewhat pointed in males, the point being generally less than midway up from the base. Ventral margin quite convex in females, slightly convex in males. Dorsal margin rather rounded, forming a more or less distinct angle with the obliquely descending posterior margin. Posterior ridge broadly rounded. Posterior slope slightly concave, occasionally with fine radiating wrinkles. Umbos not much swollen, slightly elevated above the hinge line, located in the anterior quarter of the shell, their sculpture consisting of evenly double-looped ridges. Surface of the shell generally rather smooth and shiny, some- times roughened with concentric growth lines. Periostracum often subshiny, yellow- ish green, or bottle green, becoming brownish with age, generally with distinct, but very fine, green rays over the entire surface. Left valve with two rather heavy pseudo- cardinal teeth, one in front of the other, both rather triangular. Hinge line very short and very narrow before two straight lateral teeth. Right valve with two pseudo- cardinals, the posterior one chunky and serrated, the more anterior tooth vestigial; one lateral tooth. Beak cavities shallow, 41.4 27.0 16.0 39.0 27.0 18.0 a few dorsal muscle scars under the hinge plate. Anterior adductor muscle scars well impressed, posterior ones faint, if visible. Pallial line generally distinct anteriorly. Nacre sometimes dirty white or yellowish, especially toward the beak cavities, but usually bluish white and iridescent. Length Height Width mm mm mm 49.0 28.0 17.0 Pedlar River, 1 mi. W Pleasantview, Amherst Co., Virginia. Male. North River, Rockbridge Co., Virginia. Female. Holotvpe of Unto con- strictus Conrad. [Deep River, Gulf, Chat- ham Co.], North Caro- lina. Male. Holotype of Unio genthii Lea. Breeding season. The breeding season begins in August and ends in June. Glo- chidia subspatulate, higher than long. Length 0.21 mm, height 0.27 mm (Ort- mann, 1915: 66, 67). Habitat. Lives in sand in rather swift, flowing rivers. Remarks. Villosa constricta (Conrad) of the Southern Atlantic Slope region does not much resemble any other unionid found there. There is often quite a bit of differ- ence in the degree of inflation and form of the shell, but it differs from all of the other Atlantic Slope Villosa in the shape of the postbasal swelling which generally does not project beyond the base line of the shell, and in the emargination behind the swelling. For its size constricta is much heavier and stronger than any other Villosa in the region under study, and has a dis- tinctive yellowish green or shiny dark chestnut-brown periostraeum, and the fe- male has a distinct constriction in the postbasal region. Females of Caninculina pulla (Conrad) show a similar sexual dimorphism, but they lack the constriction. In addition, pulla usually has a sharp double posterior ridge, a rough periostra- eum and if rayed at all, obscurely so, whereas constricta has a broadly rounded Atlantic Slope Unionidae • Johnson 379 posterior ridge and a generally smooth, shiny surface that is distinctly rayed. Villosa constricta (Conrad) is allopatric with Villosa vanuxemensis (Lea) (Simp- son, 1914, 1: 105) of the Cumberland and Tennessee rivers. The constriction found rather regularly in old females of con- stricta is not as well developed in vanuxe- mensis, and the latter has a distinct nacre color (usually an unreliable character) which varies from dirty purplish white through salmon-tinted to dark purple, usu- ally darker toward the beak cavity of the shell, whereas the nacre color of constricta is dirty white, bluish white or yellowish toward the beak cavities. Ortmann (1915: 66) collected a number of specimens of vanuxemensis on September 17, 1912 in the North Fork, Holston River, Saltville, Smyth Co., Virginia, and at once recognized its affinity with constricta. Range. Southern Atlantic Slope: Catawba River, North Carolina, of the upper Cooper-Santee River system, north to the James River system, Virginia. Specimens Examined Cooper-Santee River System Catawba River Drainage. North Caro- lina: Catawba River, Gaston Co. (Lea). Beaver Creek [Gaston Co.] (MZUM). Long Creek [Mecklenburg Co.] (MZUM). [Ten Mile Creek] of Catawba River, Wolfsville [7 mi. W Monroe], Union Co. (USNM). Cape Fear RrvER System Deep River Drainage. North Carolina: Deep River, Gulf, Chatham Co. (Lea). Cape Fear River Drainage. North Caro- lina: Rocky River, 11 mi. N Sanford, Chat- ham Co. (MZUM). Neuse RrvER System Neuse River Drainage. North Carolina: North Flat River, 5.5 mi. S Roxboro, Person Co. (81). Eno River, 1.8 mi. S Cedar Grove ( 122 ) ; Eno River, Hillsboro; both 380 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Orange Co. Eno River, 6 mi. NW Durham (113); Fork, Little River, 11 mi. NNW Durham ( 101 ) ; both Durham Co. Neuse River, 6 mi. E Raleigh; Hare-Snipe Creek, 6.25 mi. NW Raleigh (37); Little River, Tarpleys Mill, 2 mi. NE Wendell (5); all Wake Co. Pamlico River System Tar River Drainage. North Carolina: Tar River, 2 mi. ENE Bunn, Franklin Co. Chicod Creek, Pitt Co. (MZUM). Roanoke River System Staunton River Drainage. Virginia: Mason Creek, near Roanoke River, Salem; Tinker Creek, Roanoke (MZUM); both Roanoke Co. James River System James River Drainage. Virginia: Dun- lap Creek (MZUM); Jackson River, Cov- ington (MZUM); Cowpasture River, near Longdale (MZUM); all Alleghany Co. Calfpasture River, Goshen (MZUM); North [ = Maury] River, 2 mi. WNW Lexington; both Rockbridge Co. James River, Buchanan, Botetourt Co. Pedlar River, 1 mi. W Pleasantview, Amherst Co. James River, opposite Beaver Creek, 6 mi. E Lynchburg, Campbell Co. (MZUM). James River, Cartersville, Cumberland Co. James River, opposite Maidens, Goochland Co. James River, Richmond, Henrico Co. (ANSP). Genus Ligumia Swainson Ligumia Swainson 1840, Treatise on Malaeology, pp. 268, 274, 378. Type species, [Unio] recta Lamarck. Original designation, p. 274. Ortmann, 1912, Ann. Carnegie Mns., 8: 338, partim. Ortmann and Walker, 1922, Occ. Papers, Mns. Zool. Univ. Midi., no. 112, p. 59. Frierson (1927: 70-79) includes many taxa in this genus. Most of them have been placed by subsequent authors under Lam))- silis s. s. and Villosa. It appears that Ligumia contains two species, L. recta (Lamarck) (Ortmann, 1919: 276, pi. 16, figs. 12, 13) (the so-called subspecies latis- sima (Rafinesque) is an ecophenotype, oc- curring in rivers) and L. nasuta (Say). Ligumia nasuta (Say) Plate 19: 1,2 Unio nasutus Say 1817, Nicholson's Encyclopedia, 2 [no pagination], pi. 4, fig. 1 (Delaware and Schuylkill [rivers near Philadelphia, Philadel- phia Co., Pennsylvania]; type, ANSP [lost]). Ohiiquaria attentuata Rafinesque 1820, Ann. Gen. des Sci. Physiques (Bruxelles), 5: 304 (la fleuve Hudson [New York]; tvpe, not in ANSP [lost]). Unio rostrata Valenciennes 1827, in Humboldt and Bonpland, Voyage aux Regions Equinoxi- ales du Nouveau Continent, pt. 2, 2: 233, pi. 53, fig. 3 (Philadelphia vicinis [near] [Philadelphia Co., Pennsylvania]; type [location unknown] ). Unio vaughanianus Sowerby 1868, Conch. Iconica, 16, Unio, pi. 61, fig. 308, non Lea, 1838. Unio fisherianus Kuester 1860, Conch. Cabinet, 9, pt. 2, pi. 68, fig. 6, non Lea, 1838. Lampsilis nasuta (Say). Simpson, 1914, Cat. Naiades, 1: 97. Eurynia nasuta (Say). Ortmann, 1919, Mem. Carnegie Mus., 8: 271, pi. 16, figs. 10, 11. Ligumia nasuta (Say). Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem. 367, p. 51, fig. 54. Description. Shell generally medium in size throughout most of its range, usually not exceeding 80 mm in length, though reaching over 110 mm in a few localities. Outline elongated, subelliptical, or sub- lanceolate; distinctly over twice as long as high. Valves usually subcompressed, thin to subsolid, inequilateral. Anterior end regularly rounded; posterior end elongated and bluntly pointed. Ventral margin curved. Dorsal margin straight, slightly curved, forming a distinct angle, at least in immatures, with the obliquely descend- ing posterior slope. Hinge ligament rather long and low. Posterior ridge distinct and angular, becoming toward the umbos broader, more rounded, and indistinct. Posterior slope slightly concave, near the umbos becoming flatter or even somewhat convex posteriorly. Umbos very low, lo- cated in the anterior quarter of the shell, their sculpture consisting of fine close ridges which are looped in front and run parallel to the axis of the shell behind. Surface of the shell with irregular growth lines, sometimes with a few nearly vertical ridges below the posterior ridge. Perios- tracum usually fine, subshiny, greenish yellow, dark olive to brownish, often with distinct fine rays, especially posteriorly. Left valve with two compressed pseudo- cardinal teeth, one in front of the other, both subtriangular and crenulated; no interdentum; two long straight laterals. Right valve with two pseudocardinals, the posterior one triangular, the more anterior one low and vestigial; one lateral tooth. Beak cavities very shallow, with a few dorsal muscle scars. Anterior adductor muscle scars well impressed; posterior ones often scarcely visible. Pallial line distinct anteriorly. Nacre bluish white, often cream color or salmon, especially toward the beak cavity, posteriorly iridescent. Sexual differences well marked in the shell. The male shell tapers uniformly behind into a point, but the female is dis- tinctly produced in the postbasal region, behind which the ventral margin slopes up more acutely than in the male, and is straight or concave. Length Height Width mm mm mm 112 44 28 66 29 14 62 25 12 Connecticut River, Hart- ford, Hartford Co., Con- necticut. Male. Potomac River, Washing- ton, District of Columbia. Female. As above. Male. Anatomy. Discussed by Ortmann (1911: 343); discussed and figured by Reardon (1929: 7, pi. 2, figs. 1-10). The glochidia are subovate, with an undulate hinge line, and measure 0.25 mm in length and 0.29 mm in height. The host fish is unknown. Breeding, season. Ortmann (1919: 272) indicates from specimens he examined in Atlantic Slope Qniontdae • Johnson 3S1 Lake Erie that this species breeds from August until the following June. Habitat. Lives in sand and mud, mostly in protected areas in ponds, lakes, canals, and in slack water areas of streams. Remarks. Ligumia nasuta (Say) of the upper Ohio River drainage and Atlantic- Slope region is distinguishable from any other North American unionid by its usually distinctly lance-head shaped pos- terior end. On the Atlantic Slope the male shell can be confused with Elliptio lanceo- lata (Lea), but the latter does not have the same distinct pattern of rays, nor the dull olive green color. Also the nacre of nasnta is generally silvery white or cream colored, whereas that of lanceolata may be coppery or purple. The females of nasuta have a distinct expansion in the postbasal region which render them more easily distinguishable from lanceolata than the male shell. Clarke and Berg (1959: 51), apparently following Simpson ( 1914, 1 : 97 ) , give the southern range of this species as North Carolina. Simpson's record appears to be based on series of rather posteriorly pro- duced specimens of Elliptio lanceolata (Lea) from Wilmington, New Hanover Co., North Carolina. Ligumia nasuta is decidedly part of the northern element of the Atlantic Slope fauna, the James River, Virginia, being its southernmost record. (Conrad, 1836: 38, pi. 18, fig. 1, figures a female specimen from the James River, but it has not been collected there since his time.) As observed by Conrad, nasuta does not go very far up into the rivers in the southern part of its range, and we are led by Conrad to believe that his record from the James is from the tidewater region. It is abundant in the Potomac River in the tidewater region, but is not found very far up the river. Ortmann (1919: 275) found the distribution to be similar in Pennsylvania. Range. Atlantic Slope: James River system, Virginia, north to the St. Lawrence 382 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 River system, Canada, westward through the Mohawk River and Erie [now New York State Barge] Canal, both New York. Interior Basin: Lake Erie, Ohio, and Michigan. Specimens Examined James River System James River Drainage. Virgino: James River (Conrad). Potomac River System Potomac River Drainage. District of Columbia: Potomac River. Genus Lampsilis Rafinesque Lampsilis Rafinesque 1820, Ann. Gen. des Sci. Physiques (Bruxelles), 5: 298. Species listed: Lampsilis cardium Rafinesque, Lampsilis ovata (Say), Lampsilis fasciola Rafinesque. Type species, Unio ovatus Say. Subsequent designation, Herrmannsen, 1847, Indicis Gene- rum Malacozoorum, 1: 575. Ortmann, 1912, Ann. Carnegie Mus., 8: 345. Aeglia Swainson 1840, Treatise on Malacology, pp. 265, 378. Species listed: Ac. ovata (Say), occidens ( Lea ) . Type species, Unio ovatus Say. Subsequent designation, Herrmannsen, 1846, Indicis Gene- rum Malacozoorum, 1: 20. Subgenus Lampsilis s.s. All of the species described in this paper belong to Lampsilis s. s. Frierson (1927: 67-86) lists ten other subgenera, three of which are of his own creation. One of them, Villosa, is now used as a replace- ment name for Micromya Agassiz (see under Villosa on p. 371. Ligumia Swainson now replaces Eurynia Rafinesque. To comment on the other subgenera is not in the scope of this paper, but on cursory examination, the present author disagrees substantially with Frierson's classification, both on a generic and specific level. Lampsilis s. s., while clearly of Interior Basin origin, appears to have1 speciated about equally there and in the combined Apalachicolan and Atlantic Slope regions. lampsilis (Lampsilis) cariosa (Say) Plate 19:3-5 Unio cariosus Say 1817, Nicholson's Encyclopedia, 2 [no pagination], pi. 3, fig. 2 (Delaware and Schuylkill Rivers; [Susquehanna River], Wilkes Rarre; [Luzerne Co.; all Pennsylvania]; type, ANSP [lost]). The type locality was restricted to: Schuylkill River, near Philadelphia, Penn- sylvania, and a neotype MCZ 178839 was selected by Johnson ( 1947, Occ. Papers on Moll, 1: 148, pi. 19, fig. 1). Lampsilis pallida Rafinesque 1820, Ann. Gen. des Sci. Physiques (Bruxelles), 5: 299 (Hudson River, [New York]; type not in ANSP [lost]). Unio ovata Valenciennes 1827, in Humboldt and Bonpland, Voyage aux Regions Equinoxiales du Nouveau Continent, pt. 2, 2: 226, pi. 50, figs. 1 a-c, (pres de Washington [District of Co- lumbia]; type [location unknown]) non Say 1817. Unio crocatus Lea 1841, Proc. Amer. Philos. Soc, 2: 31 (Savannah River, Georgia). Lea, 1842, Trans. Amer. Philos. Soc, 8: 238, pi. 22, fig. 52; figured holotype USNM 84908. The lecto- type selected by Johnson ( 1947, Occ. Papers on Moll., 1: 156) but not figured, is invalid. The holotype which was catalogued under the same number has since been located. Lea, 1842, Obs. Unio, 3: 76. Unio oratus Conrad 1849, Proc. Acad. Nat. Sci. Phila., 4: 153 (Flint River, Georgia; here re- stricted to: Ogeechee River, Georgia1). Conrad 1850, Jour. Acad. Nat. Sci. Phila., ser. 2, 1: 267, pi. 37, fig. 6; type, coll. of J. H. Couper [lost]. Lampsilis cariosa (Say). Simpson, 1914, Cat. Naiades, 1: 43. Ortmann, 1919, Mem. Carnegie Mus., 8: 313, pi. 20, figs. 3-5. Johnson, 1947, Occ. Papers Moll., 1: 145, pi. 19, figs. 1-2. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem., no. 367, p. 54, figs. 59-60. Clarke and Rick, 1963, Natl. Mus. Canada, Bull. no. 189, p. 27, pi. 1, fig. 3. Description. Shell usually medium, though reaching 130 mm in length. Out- line of male elliptical and somewhat elon- gate; of female subovate, or obovate, rather short and high. Valves inequilateral, some- 1 Also described on the same page in both publications is Unio contrarius = Lampsilis clai- bornensis ( Lea ) with Ogeechee River, Georgia, as the type locality. The localities for the two species are obviously transposed, since oratus is not found in the Flint River, but in the Ogeechee River, and contrarius is found in the Flint River and not in the Ogeechee River. Atlantic Slofe Umontdae • Johnson 383 what inflated; moderately heavy. Anterior end regularly rounded; posterior end more broadly rounded and somewhat pointed in the male, truncated in the female. Ven- tral margin slightly curved. Dorsal margin straight, terminating in a broad angle with the obliquely descending posterior margin. Hinge ligament prominent, running under the umbos and appearing anteriorly. Pos- terior ridge rounded and poorly defined. Posterior slope slightly convex, usually with a few faint ridges and wrinkles. Umbos moderately swollen, slightly raised above the hinge line, located somewhat anterior to the middle of the shell, their sculpture consisting of four or five wavy recurved ridges. Surface of the shell generally smooth except for a few growth ridges. Periostracum usually shiny, bright wax or straw yellow7, infrequently greenish yellow, becoming a dirty brownish yellow or red- dish brown in matures. Green or blackish rays, when present, usually restricted to the posterior slope. Left valve with two compressed, ser- rated pseudocardinal teeth, one in front of the other, the anterior one somewhat tri- angular, the hinder one lower, but chunky, located directly under the umbo. Hinge line rounded, with a considerable inter- dentum under the umbo before two short lateral teeth. Right valve with two oppos- ing pseudocardinals, the anterior one low and laminate, the other higher, chunky and serrated; one lateral tooth, sharply truncated behind. Beak cavities moder- ately deep, with deep dorsal muscle scars. Anterior adductor muscle scars well im- pressed, posterior ones less so. Pallial line distinct. Nacre bluish white or tinged with salmon. Length Height Width mm mm mm 30 85 56 Savannah River, 7 mi. NE Newington, Screven Co., Georgia. Male. 98 62 42 As above. Male. 98 69 45 As above. Female. Anatomy. Figured by Lea ( 1838, pi. 15, fig. 45). Ortmann (1912: 353) says that the anatomy of L. cariosa is similar to that of Lampsilis ovata uentricosa (Barnes). Breeding season. According to Ortmann (1919: 315) this species is probably brady- tictic. Habitat. The largest examples are found in the swift waters of large rivers on sand bars or gravelly bottoms, though it also oc- curs in smaller creeks and, occasionally, in ponds. Remarks. In the Atlantic Slope region, Lampsilis cariosa (Say) can be confused with L. clolabraeformis (Lea) of the Alta- maha River system, but the latter species has much higher umbos and a very sharp posterior ridge, especially toward the umbos, whereas in cariosa the umbos are slightly elevated above the hinge line, and the posterior ridge is poorly defined. Lampsilis cariosa has been most often confused with L. ochracea (Say) with which it is sometimes found throughout its range. In general, L. ochracea has a thin- ner, smaller shell, and unlike cariosa, which is rarely rayed except posteriorly, ochracea is often rayed over the entire surface of the shell and it has a rougher, duller, more greenish periostracum. The hinges are quite different. The pseudo- cardinals of L. cariosa are pyramidal, and an interdentum is present; but in L. och- racea the pseudocardinals are lamellate, almost parallel to the hinge line, and there is no interdentum. Lampsilis cariosa is found with Lamp- silis ovata (Say) in the Potomac River system, where the latter was introduced, and in the St. Lawrence River system. While the two species are close, they are distinct. L. ovata has a rather sharp pos- terior ridge, is somewhat inflated, has a dull olivaceous tint to the shell, and is inclined to be rayed over the entire surface. L. cariosa has a poorly defined posterior ridge, is not especially inflated, usually has a distinctive bright yellow, glossy periostracum, and rays when present are usually confined to the posterior slope. 384 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Range. Atlantic Slope: Ogeechee River system, Georgia, north to the Sydney River, Nova Scotia (Clarke and Rick, 1963: 27), and in the St. Lawrence River system, Canada, westward to the Madawaska River, Ontario, of the lower Ottawa River drainage. Specimens Examined Ogeechee River System Ogeechee River Drainage. Georgia: Ogeechee River, Shoals, Warren Co. Ogeechee River, Scarboro, Screven Co. (MZUM). Ogeechee River, mouth of Arnold Lake [Rryan Co.] (MZUM). Savannah River System Savannah River Drainage. South Caro- lina: Abbeville District (USNM). Edge- field District. Savannah River, 2 mi. SW Millettville (ANSP); Savannah River, Johnsons Landing, 10 mi. W Allendale; Savannah River, Kingjaw Point, 10 mi. WSW Allendale (ANSP); all Allendale Co. Georgia: Savannah River, 7.5 mi. NE Shell Rluff; Savannah River, 6 mi. NE Girard; both Rurke Co. Relow Haga Slage Landing, E Sylvania (ANSP); Savannah River, 7 mi. NE Newington; both Screven Co. Savannah River, in sandbank, Savan- nah, Chatham Co. (MZUM). Cooper-Santee Rpver System Congaree River Drainage. South Caro- lina: Congaree River. Wateree River Drainage. South Caro- lina: Wateree River, 2.5 mi. W Camden, Kershaw Co. Santee River Drainage. South Caro- lina: Santee River, Wilsons Landing, 5 mi. NW Pineville (ANSP); below Santee River Dam, near St. Stephens; both Berkeley Co. Waccamaw River System Waccamaw River Carolina: Waccamaw River, Wachasaw Drainage. South Landing, 2 mi. W Murrells Inlet, George- town Co. Cape Fear River System Cape Fear River Drainage. North Caro- lina: Cape Fear River, Cumberland Co. (USNM). Neuse RrvER System Neuse River Drainage. North Carolina: Eno River, 6 mi. NNW Durham, Durham Co. (112). Neuse River [6 mi. E] Raleigh; tributary of Swift Creek, 5 mi. SW Raleigh (22); both Wake Co. Pamlico River System Tar River Drainage. North Carolina: Tar River, 2 mi. W Springhope, Nash Co. Chowan River System Nottoway River Drainage. Virginia: Nottoway River, 3 mi. E Rawlings, Bruns- wick Co. York River System Pamunkey River Drainage. Virginia: Pamunkey River, 2 mi. N Hanover, Han- over Co. Potomac River System Potomac River Drainage. Maryland: Monocacy River, 2 mi. W Mt. Pleasant, Frederick Co. Potomac River, Cabin John, Montgomery Co. (Ortmann, 1919: 316). Potomac River, Great Falls, Fairfax Co. District of Columbia: Potomac River. Lampsilis (Lampsilis) dolabraeformis (Lea) Plate 20: 1-4 Unto dolabraeformis Lea 1838, Trans. Araer. Philos. Soc, 6: 103, pi. 24, fig. 113 (Altamaha River, Liberty [now Long] Co.; Altamaha River, near Darien [Mcintosh Co]; both Georgia; figured holotype USNM 84888). Lea, 1838, Obs. Unio, 2: 103. Lampsilis dolabraeformis (Lea). Simpson, 1914, Cat. Naiades, 1: 46. Description. Shell large, often exceed- ing 130 mm in length. Outline of male Atlantic Slope Unioxidae • Johnson 385 elliptical and somewhat elongate; of fe- male obovate, rather short and high. Valves inflated, rather heavy. Anterior end regu- larly rounded; posterior end more broadly rounded and somewhat pointed in the male, truncated in the female. Ventral margin slightly curved. Dorsal margin straight terminating in a broad angle with the obliquely descending posterior margin. Hinge ligament prominent, running under the umbos and appearing anteriorly. Pos- terior ridge very sharp and angular near the umbos, gradually becoming less so toward the ventral margin, especially in older individuals. Posterior slope slightly convex, usually with a few faint ridges and wrinkles. Umbos swollen, much elevated above the hinge line, located somewhat anterior to the middle of the shell, their sculpture consisting of four or five rather evenly raised bars which terminate at the posterior ridge. Surface of the shell smooth except for a few concentric ridges ante- riorly. Periostracum usually shiny, yellowish green, sometimes becoming chestnut brown or black in matures. The posterior slope usually has a concentration of green or black rays which extend with less fre- quency over the posterior portion of the disk. Left valve with two compressed pseudo- cardinal teeth, one in front of the other, the anterior one higher. Sometimes these are broken into three imperfect teeth or are united in one narrow ridge. Hinge line rounded, with a considerable interdentum under the umbo before two short lateral teeth. Right valve with two opposing pseudocardinals, the anterior one low and laminate, the other higher, chunky, and serrated; one lateral tooth, shaqily trun- cated behind. Beak cavities deep and wide, with deep dorsal muscle scars under the pseudocardinal teeth. Anterior adductor muscle scars deep, posterior ones distinct, but not impressed. Pallial line distinct. Nacre white, bluish white, or pinkish, usuallv iridescent. Length Height Width mm nun mm 129 91 64 108 78 55 91 69 49 79 63 50 Altamaha River, 4 mi. N Jesup, Wayne Co., Georgia. Male. As above. Female. Altamaha River, 10 mi. N Raxley, Appling Co., Geor- gia. Female. As above. Male. Anatomy. Discussed by Lea (1863: 413). Habitat. Usually found in sand bars, where the river is wide and the water is rather swift, though occasionally found in mud. Remarks. Lampsilis dolabraeformis (Lea) is known only from the Altamaha River system. It is close to Lampsilis cariosa (Say) which is found throughout the Atlantic Slope region, but differs from cariosa by having umbos which are much higher and a posterior ridge which is con- siderably sharper, especially toward the umbos. In the Apalachicolan region, L. dolabrae- formis resembles L. excavatus (Lea). The latter does not extend east of the Escambia River system. It is a smaller species than dolabraeformis, the umbos are not as high, the posterior ridge is not quite as sharp, especially in females, and the entire sur- face has wide green or blackish rays which are often broken into concentric bands of green. In dolabraeformis the rays are restricted to the area toward the posterior ridge and posterior slope, and the rays are fine, sharp, and not broken. L. dolabrae- formis also resembles L. binominatus Simpson (Johnson, 1967b: 9) of the Upper Apalachicola River system. The rays are very similar, but in binominatus they cover the entire surface. Further, binominatus is a much smaller species, which does not have the high umbos or sharp posterior ridge of dolabraeformis. Simpson (1914, 1: 46) reports dolabrae- formis from the Ogeechee and Savannah River systems, but the specimens in the United States National Museum are all Lampsilis cariosa (Say). 386 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Range. Southern Atlantic Slope: re- stricted to the Altamaha River system, Georgia. Specimens Examined Altamaha River System Ocmulgee River Drainage. Georgia: Ocmulgee River, Hawkinsville, Pulaski Co. Ocmulgee River, 1.5 mi. S Jacksonville; Gum Swamp Creek, 1 mi. N McRae; Ocmulgee River, 1 mi. S Lumber City; all Telfair Co. Oeonee River Drainage. Georgia: Oconee River, 2.5 mi. N Glenwood; Och- walkee Creek, 2 mi. E Glenwood; Oconee River, 8 mi. SW Soperton; all Wheeler Co. Altamaha River Drainage. Georgia: Altamaha River, 7 mi. N Hazlehurst, Jeff Davis Co. Altamaha River, 10 mi. N Baxley; Altamaha River, 10 mi. NE Sur- rency; both Appling Co. Altamaha River, "Riverside Park," 4 mi. N Jesup, Wayne Co. Altamaha River, 3 mi. NW Everett City, Glynn Co. Altamaha River, Fort Barrington; Altamaha River, Hopeton, near Darien; both Mcintosh Co. Lampsilis (Lampsilis) ovata (Say) Plate 21: 1-2 Unio ovatus Say 1817, Nicholson's Encyclopedia, 2 [no pagination], pi. 2, Fig. 7 (Ohio River and its tributary streams; type, ANSP [lost]). Neotype Senckenberg Mus. 4338 [not seen] selected, but not figured, by Haas, 1930, Senckenbergiana, 12: 328. Lampsilis cardium Rafinesque 1820, Ann. Gen. des Sci. Physiques (Rruxelles), 5: 298, pi. 30, figs. 16-19; ([Ohio River]: supposed type ANSP 20210 teste Vanatta, 1915, Proc. Acad. Nat. Sci. Phila., 66: 5511). Unio ventricosus Barnes 1823, Ainer. Tour. Sci., 6: 267, pi. 13, fig. 14 (The Wisconsan [Wis- consin River] near Prairie du Chien [Crawford Co., Wisconsin] type, Lyceum of Natural History of New York [destroyed by fire]). Unio occidens Lea 1829, Trans. Amer. Philos. Soc, 3: 435, pi. 10, fig. 16 (Ohio; figured holotype USNM 84866). Lea, 1834, Obs. Unio, 1: 49. Unio suhovatus Lea 1831, Trans. Amer. Philos. Soc, 4: 118, pi. 18, fig. 46 (Ohio River; figured holotype USNM 84509). Lea, 1834, Obs. Unio, 1: 128. Unio lenis Conrad 1838, Monography Unionidae, no. 11, back cover. Conrad, 1840, op. tit., no. 12, p. 106, pi. 58, fig. 2 (upper part of White River, Illinois, possible type ANSP 42307, [smaller than figured specimen]). Unio canadensis Lea 1857, Proc. Acad. Nat. Sci. Phila., 9: 85 (St. Lawrence River, near Mon- treal [Quebec], Canada). Lea, 1860, Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 268, pi. 44, fig. 148; figured holotype USNM 84503. Lea, 1860, Obs. Unio, 7: 86. Unio latissimus Sowerby 1868, Conch. Iconica, 16, Unio, pi. 66 fig. 337 (United States, Sowerby coll. not in BMNH [lost]). Lampsilis ventricosa lurida Simpson 1914, Cat. Naiades, 1: 41 (throughout the St. Lawrence drainage; holotype USNM [not located]). Lampsilis ventricosa cohongownta Ortmann 1912, Nautilus, 26: 53 (no type selected. Lecto- type, selected by Parodiz (1967: 28) Carnegie Mus. 61.3999, from Potomac River, Hancock, Washington Co., Maryland), pi. 21: 1; allotype 61.4000, pi. 21:2. Lampsilis ventricosa winnchagoensis Baker 1928, Bull. Wisconsin Geol. Nat. Hist. Survey, no. 70(2), p. 291, pi. 94, figs. 1-4 (Winnebago Lake, near Oshkosh [Winnebago Co.], Wiscon- sin; holotype MZUM 209219). Lampsilis ventricosa perglohosa Baker 1928, Bull. Wisconsin Geol. Nat. Hist. Survey, no. 70(2), p. 285, pi. 93, figs. 1-4 (Lake Pepin, near Lake City [Wabasha Co.], Wisconsin; holotype MZUM 209195). Lampsilis ovata (Say). Ortmann, 1919, Mem. Carnegie Mus., 8: 297, pi. 17, figs. 8, 9; pi. 18, figs. 1-3. Lampsilis ovata ventricosa (Barnes). Ortmann, 1919, Mem. Carnegie Mus., 8: 301, pi. 18, fig. 4 pi. 19, figs. 1-3. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem., no. 367, p. 55, figs. 47, 48. Description. Shell usually medium, though reaching 140 mm in length. Outline of male elliptical or subelliptical, of female subovate, rather short and high. Valves somewhat to considerably inflated, thin when young, becoming thick with age, in- equilateral. Anterior end regularly rounded; posterior end more broadly rounded and somewhat pointed in the male, truncated in the female. Ventral margin slightly Atlantic Slope Unionidae • Johnson 387 curved. Dorsal margin short and straight, terminating in a broad angle or gentle curve with the obliquely descending pos- terior margin. Hinge ligament prominent, occupying most of the dorsal margin. Pos- terior ridge variable; it may be rounded and poorly defined, though often becoming distinct toward the umbos, and in some habitats the posterior ridge becomes very sharp toward the umbos. Posterior slope varies from gently convex to almost flat, or even concave toward the umbos. Umbos moderately swollen, slightly raised above the hinge line, located somewhat anterior to the middle of the shell, their sculpture consisting of four or five rather coarse bars, of which the second and third have a slight tendency to fall into two loops, with a small sinus in the middle, while the first, fourth, and fifth bars are indistinct. Surface of the shell generally smooth ex- cept for growth ridges. Periostracum usually light or dark yellowish green to olive brown. Dark green or blackish rays which are straight and continuous, finer or broader, often cover the entire surface or only part of it, but some rays are almost always present. On the posterior slope the periostracum is less smooth and usually darker. Left valve with two slightly compressed pseudocardinal teeth, one in front of the other, the anterior one somewhat triangu- lar, the hinder one lower, located directly under the umbos, both teeth quite chunky. Hinge line rounded with a narrow inter- dentum under the umbos; two short lateral teeth. Right valve with two opposing pseudocardinals, the anterior one low and laminate, the other higher, quite chunky, and serrated; one lateral tooth, sharply truncated behind. Beak cavities moder- ately deep, with deep dorsal muscle scars. Anterior adductor muscle scars well im- pressed, posterior ones less so. Pallial line distinct. Nacre silvery or bluish white, sometimes suffused with pink or purple. Length Ileiuht Width mm nun mm 115 78 56 111 73 47 95 60 41 87 52 57 Potomac- River, Hancock, Washington Co., Maryland. Female. South Branch, Potomac River, South Branch, Hamp- shire Co., West Virginia. Male. Carnegie Mus. Potomac River, Hancock, Washington Co., Maryland. Lectotype of L. v. cohon- goronta Ortmann. Male. As above. Allotype. Fe- male. Anatomy. Discussed by Ortmann (1911: 351). The glochidia are subelliptical, with a straight line hinge, without hooks, and measure 0.25 mm in length and 0.29 mm in height. Coker, et al. (1921: 153) re- port the following fish as hosts: bluegill, Lepomis macrochirus Rafinesque; white crappie, Pomoxis annularis Rafinesque; largemouth bass, Micropterus salmoides (Lacepede); smallmouth bass, M. dolo- mieui dolomieui Lacepede; yellow perch, Perca flave.scens (Mitchill); and yellow pikeperch, Stizostedion vitreum (Mitchill). Breeding season. According to Ortmann (1919: 298, 303, 308) this species is brady- tictic. Habitat. Lives in rivers and creeks on gravel, sand, or even mud bottoms; also occurs in lakes. Remarks. Lampsilis ovata (Say) is a species of the Interior Basin and St. Law- rence River system which was accidently introduced into the Shenandoah River in 1889 and in other parts of the Potomac River system in 1894, while in the larval stage, on species of bass and other fishes that were transplanted at those times from the west. On the Atlantic Slope ovata can be confused only with Lampsilis cariosa (Say), with which it is now found in the lower Potomac River system. L. ovata generally has a sharper posterior ridge, especially near the umbos, tends to be inflated, has a dull olivaceous tint to the shell, is inclined to be rayed over the entire surface, and has rather heavy chunky 388 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 pseudocardinals. L. cariosa always has a poorly defined posterior ridge, is not much inflated, usually has a distinctive bright waxy yellow glossy periostracum. Rays when present are usually confined to the posterior slope, and the pseudocardinals are less chunky than those of ovata. The present synonymy of ovata may not be the last word spoken on the subject ( the author is not yet prepared to comment on the validity of Unio satur Lea, which is found in the Western Gulf Coastal region ) , but it is an attempt to use modern system- atic concepts. Ortmann (1919: 298, 303) clearly states that L. ovata occurs in large rivers and gradually gives way to ovata ventricosa in smaller rivers and streams, that numerous intergrades occur in inter- mediate areas and, further, that this is a general phenomenon and occurs in widely separated streams. Van der Schalie (1938: 70) discusses a similar change from ovata ventricosa, the river form, to the ovata canadensis, the lake form, which also inter- grade. Goodrich and van der Schalie (1944: 315) again note that, "as one pro- gresses into the headwaters the sharp pos- terior ridge of the true ovata is seen to round off and we pass gradually to the more common form of the species in Indi- ana, known as LampsUis ovata ventricosa ( Barnes ) ." Cvancara's ( 1963 ) work ap- pears to indicate that ovata and ventricosa are ecophenotypes. Ortmann (1912: 53) proposed cohongo- ronta as a variety of LampsUis ventricosa, "on account of the small size, the shorter and higher outline, and somewhat more distinct posterior ridge." He found no difference in the anatomy or glochidia from L. ventricosa. When Ortmann wrote in 1912, LampsUis ovata was found on the Atlantic Slope only in the Great Allegheny Valley, an area from which cariosa has never been re- ported. Since that time ovata has spread throughout the Potomac River, having reached Great Falls, Maryland, by 1915 (Marshall, 1917: 40). Writing again, Marshall (1930: 19) indicated that cariosa was being replaced by ovata. This is con- firmed by collections made in the Potomac River in 1965 by S. L. H. Fuller. Range. Interior Basin: Mississippi and Ohio drainages. St. Lawrence drainage from Lake Superior to the Ottawa River and Lake Champlain. Hudson Bay drain- age. Northern Atlantic Slope: restricted to the Potomac River system, Maryland (in- troduced ) . Specimens Examined Potomac River System Potomac River Drainage. West Vir- ginia: South Branch, Potomac River, Rom- ney (Carnegie Mus.); South Branch, Potomac River, South Branch [Depot = French Station] (Carnegie Mus.); both Hampshire Co. Cacapon River, 1 mi. above confluence with Potomac River, Morgan Co. (USNM). Maryland: Potomac River, Hancock, Washington Co. (Carnegie Mus.). Virginia: Shenandoah River, Harp- ers Ferry, Jefferson Co. (Carnegie Mus.). Maryland: Potomac River, Paton Island, 0.75 mi. W Point of Rocks, Frederick Co. Potomac River, 2.5 mi. WSW Dickerson; Potomac River, Harrison Island, 5 mi. above Great Falls (USNM); Potomac River, Great Falls (USNM); Chesapeake and Ohio Canal, Great Falls (USNM); all Montgomery Co. LampsUis (LampsUis) ochracea (Say) Plate 21: 4, 5 Mytilus fluviatilis Gmelin 1791, Systema Naturae, ed. 13, 1, pi. 6, p. 3359 (Europae, aquis dulci- bus [erroneous], type [probably based on Lister's figure only]) refers to: Lister, 1685, Synopsis Methodicae Conehyliorum, pi. 157, fig. 12 (Virginia; figured specimen Oxford Univ. Mus. [probably lost] teste Dance, 1966, Shell Collecting, p. 292). Conrad (1836, Monog- raphy Unionidae, no. 4, p. 37) considered Lister's figure to represent LampsUis ochracea (Say), but the figure is sufficiently ambiguous to have subsequently allowed Isaac Lea to claim successfully, for over fifty years, that M. fluviatilis Gmelin was Anodonta cataracta Say. Since Lister makes no reference to dentition or lack of it, his figure cannot be positively identified. Gmelin's name is certainly a nomen dubium. Unio ochraceus Say 1817, Nicholson's Encyclo- pedia, 2 [no pagination], pi. 3 fig. 8 (Delaware and Schuylkill Rivers; type ANSP [lost]). The type locality was restricted to Schuylkill River, mar Philadelphia, Pennsylvania, and a neotype MCZ 178838 was selected by Johnson, 1947, Occ. Papers on Moll., 1: 153, pi. 20, fig. 2. Lampsilis rosea Rafinesque 1820, Ann. Gen. des Sci. Physiques (Bruxelles), 5: 229 (Hudson River [New York]; type not in ANSP [lost]). Unio wsaceus Conrad 1849, Proc. Acad. Nat. Sci. Phila., 4: 153 (Savannah River, Georgia, J. H. Couper [loaned]). Conrad, 1850, Jour. Acad. Nat. Sci. Phila., ser. 2, 1: 275, pi. 37, fig. 5; figured holotype MCZ 178779, purchased from J. H. Couper. Lampsilis ochracca (Say), Simpson, 1914, Cat. Naiades, 1: 49. Ortmann, 1919, Mem. Car- negie Mus., 8: 318, pi. 20, figs. 6, 7. Johnson, 1947, Occ. Papers on Moll., 1: 150, pi. 20, figs. 1-2. Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem., no. 367, p. 57, figs. 55, 56. Athearn and Clarke, 1962, Natl. Mus. Canada, Bull. no. 183, p. 30, pi. 4, figs. 3, 4. Description. Shell usually small, seldom exceeding 70 mm in length. Outline of male elliptical and somewhat elongate; of female subovate, rather short and high. Valves inequilateral, subinflated, rather thin, though strong, sometimes translucent. An- terior end regularly rounded; posterior end more broadly rounded and somewhat pointed in the male, truncated in the fe- male. Ventral margin usually straight or slightly curved, sometimes concave in females. Dorsal margin straight or slightly curved, terminating in a broad angle with the obliquely descending posterior margin. Hinge ligament prominent, running under the umbos and appearing anteriorly. Pos- terior ridge rounded and generally poorly defined. Posterior slope slightly convex, usually with a few faint ridges and wrinkles. Umbos moderately swollen, slightly raised above the hinge line, located somewhat anterior to the middle of the shell, their sculpture consisting of four or five wavy recurved ridges. Surface of the shell generally smooth in the region of the Atlantic Slope Umomdak • Johnson 389 disk, becoming roughened with perios- tracum posteriorly. Periostracum usually subshiny, brownish olive, brownish, green- ish yellow, reddish yellow, or yellow. Dull, rather fine, greenish rays sometimes found over the entire surface of the shell, espe- cially in immatures. Left valve with two compressed, ser- rated pseudocardinal teeth, one in front of the other, the anterior one somewhat triangular, the hinder one lower, inclined to be vestigial. Hinge line rather long and narrow before two short, curved lateral teeth. Right valve with two triangular, narrow, rjarallel pseudocardinals separated by a deep narrow pit, the more anterior tooth inclined to be vestigial; one lateral tooth. Beak cavities shallow, with deep dorsal muscle scars. Anterior adductor muscle scars well impressed, posterior ones less so. Pallial line not very distinct. Nacre white, bluish white, or pinkish, usually iri- descent. Length Height Width mm mm mm 73 50 54 44 31 38 27 19 24 Savannah River, Georgia. Holotype of U. rosaceus Conrad. Male. Schuylkill River, near Philadelphia, Pennsylvania. Neotype. Male. Halfway Pond, Plymouth, Plymouth Co., Massachu- setts. Female. Anatomi/. Figured by Lea (1838: pi. 15, fig. 44) and described (1863: 455) also by Reardon (1929: 1, pi. 1, figs. 1-10). The host fish is unknown, but since ochrocea is generally restricted to the lower regions of streams or bodies of water directly con- nected with the ocean, it may, like Ano- donta implicata Say, parasitize a migratory fish. Breeding season. Lea (1863: 455) found this species gravid in the autumn. I found gravid females on May 3, 1943 in Plym- outh, Massachusetts. It is probably brady- tictic. Habitat. Lives in sand or mud in ponds, canals, and the lower portions of rivers to- 390 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 ward the tidal region in areas where the current is not strong. Remarks. In the Atlantic Slope region, Lampsilis ochracea (Say) has been con- fused with L. cariosa (Say), but ochracea has a thinner, smaller shell and, unlike cariosa which is rarely rayed, ochracea is often rayed over the entire surface. It has a rougher, duller, more greenish perios- tracum. The hinges are quite different. The pseudocardinals of cariosa are pyra- midal and an interdentum is present, but in ochracea the pseudocardinals are lamel- late, almost parallel to the hinge line, and there is no interdentum. Lampsilis ochracea may also be mis- taken for immature Lampsilis splendida (Lea), but only in that the shells of both may be rayed over the entire surface. The rays of ochracea are finer, and the shell is more yellowish, whereas that of splendida is more brownish, less delicate, and more inflated. Simpson (1914: 49) gives the range of this species as from "New England to the Ogeechee River, Georgia." The specimen in the United States National Museum on which he based the Ogeechee record from Le Conte appears to be a young cariosa. Unfortunately, in none of the major mu- seums are there any precisely labeled spec- imens of ochracea from the Savannah River, the apparent southern limit of its distribu- tion. Conrad (1836: 37) reports ochracea as occurring in, "most of the tide waters north of Savannah River." Range. Atlantic Slope: Savannah River system, Georgia, north to River Herbert, Nova Scotia (Athearn and Clarke, 1962: 30). Specimens Examined Savannah River System Savannah River Drainage. Georgia: Savannah River (MCZ and USNM). Waccamaw River System Waeeamaw River Drainage. North Carolina: Lake Waccamaw, [town of] Lake Waccamaw; drainage canal, beside Lake Waccamaw, 1 mi. NNW Dupree Landing; both Columbus Co. Soutli Carolina: Waccamaw River, Wachasaw Landing, 2 mi. W Murrells Inlet, Georgetown Co. Pamlico River System Tar River Drainage. North Carolina: Tar River, Old Sparta, 3.5 mi. W Pinetops, Edgecombe Co. Tar River, Bruce, 9 mi. NW Greenville, Pitt Co. Roanoke River System Roanoke River Drainage. North Caro- lina: Roanoke River, Weldon, Halifax Co. (USNM). James River System Appomattox River Drainage. Virginia: Petersburg, Dinwiddie Co. (USMN). James River Drainage. Virginia: James River (MCZ and USNM). York River System Mattaponi River Drainage. Virginia: Mantapike Creek, Mantapike, King and Queen Co. (USNM). York River Drainage. Virginia: York River (MZUM). Potomac River System Potomac River Drainage. Virginia: Po- tomac River, Great Falls, Fairfax Co. District of Columbia: Potomac River. Ana- costia River. Virginia: Canal, Alexandria (MZUM); Little Hunting Creek, near Mt. Vernon ( MZUM ) ; both Fairfax Co. Lampsilis (Lampsilis) radiata radiata (Gmelin) Plate 22: 1-3 Mya radiata Gmelin 1791, Systema Naturae, ed. 13, 1, pt. 6, p. 3220 (Malabariae [erroneous]) refers to: Martini and Chemnitz, 1782, Conch. Cabinet, 6, pi. 2, fig. 7 ( [source of Cmelin's locality]; figured specimen Spengler colln. Copenhagen, [lost] teste Haas, 1913, Vidensk. Meddr. Dansk naturh. Foren. [Copenhagen], 65: 59); also Lister, 1685, Synopsis Methodicae Conchyliorum, pi. 152, fig. 7 (Virginia; figured specimen, Oxford Univ. Mus. [probably lost] Atlantic Slope Unionioae • Johnson 391 teste, Dance, 1906, Shell Collecting, p. 292). The figure in Martini and Chemnitz is here selected as the type figure, and the type lo- cality, following Lister and Simpson (1914, Cat. Naiades, 1: 64), is here restricted to Potomac River, District of Columbia ( approxi- mately opposite, Fairfax Co., Virginia). [//i/'o hiteola Lamarck 1819, Hist. Nat. des Ani- maux sans Vertebres, 6: 79 (la riviere Susque- hana [Susquehanna River, Pennsylvania] et celle Mohancks [Mohawk River, New York]; measured holotype Paris Museum, figured by Wheeler, 1963, Nautilus, 77: 58, pi. on p. 59, figs. 1-2. Refigured by Johnson, 1969, Nautilus, 83: 54, fig. 11, the type locality was restricted to the Susquehanna River, Columbia, York Co., Pennsylvania ) . [7/i/'o lincata 'Valenciennes' Bory de St. Vincent 1827, Encyclopedic Methodique, 2 of atlas, explanation of pis., p. 151, pi. 248, fig. 5 (no locality, type [presumed lost]). [Named from figure published in 1797.] [//i/'o tenebrosus Conrad 1834, New Fresh Water Shells United States, p. 42, pi. 7, fig. 1 (Rap- pahannock [sic] River, near Falmouth, Stafford Co., Virginia; tvpe presumably in Poulson colln., ANSP [lost]). ' [7/i/o melinus Conrad 1838, Monography Unioni- dae, no. 11, p. 101, pi. 56, fig. 1 (Salina Lake [Onondaga Co.], New York; type, not in ANSP, presumed lost). Unio boydianus Lea 1840, Proc. Amer. Philos. Soc, p. 286 (Oak Orchard Creek, Orleans Co., New York). Lea, 1843, Trans. Amer. Philos. Soc, 8: 216, pi. 16, fig. 32; figured holotype USNM 86126). Lea, 1942, Obs. Unio, 3: 54. Unio rosaceus De Kay 1843, Zool. New York, Moll., pt. 5, p. 192, pi. 39, figs. 355, 356; pi. 40, fig. 357 (Seneca Lake [North Western], New York; lectotype, here selected, USNM 678303, formerly New York State Cabinet 391, specimen figured on pi. 39, fig. 356). Mija oblongata Wood 1856, Index Testaceologicus, ed. Hanley, p. 199, Supplement, pi. 1, fig. 2 (North America, type [lost]). Unio elongata S. G. Goodrich 1859, Illustrated Natural History, 2: 523, text fig. (No locality, type [probably not saved].) Unio obliquiradiatiis Reeve 1856, Conch. Iconica, 16, Unio, pi. 29, figs. 151 (Hab?, Mus. Cuming, type [not in BMNH, lost]). Unio conspicuus Lea 1872, Proc. Acad. Nat. Sci. Phila., 24: 156 (Yadkin River, Salisbury, [Rowan Co.], North Carolina). Lea, 1874, Jour. Acad. Nat. Sci. Phila., ser. 2, 8: 34, pi. 11, fig. 31; figured holotype USNM 85056. Lea, 1874, Obs. Unio, 13: 38. Unio virginiana Simpson, 1900, Proc. United States Natl. Mus., 22: 536, non Lamarck 1819. 1 .amarck's type examined and identified as Elliptio complanatus Solander [sic] teste John- son, 1953, Nautilus, 66: 95. Lampsilis radiata (Gmelin). Simpson, 1914, Cat. Naiades, 1: 64. Ortmann, 1919, Mem. Car- negie Mus., 8: 292, pi. 17, figs. 6, 7. Lampsilis radiata oneidensis Baker 1916, Nautilus, 30: 74, pi. 2 (Oneida Lake, Central New York; syntypes, New York State College of Forestry, Syracuse, New York; ANSP 114854; MZUM). Unio virginca Frierson 1927, Check List North American Naiades, p. 72, error for Unio vir- giniana Lamarck 1819, non Lamarck. Lampsilis radiata radiata (Gmelin). Clarke and Berg, 1959, Cornell Univ. Exp. Sta. Mem. no. 367, p. 58, fig. 53. Athearn and Clarke, 1962, Natl. Mus. Canada, Bull. no. 183, p. 31, pi. 4, figs. 5, 6. Lampsilis radiata var. conspicua (Lea). Simpson, 1914, Cat. Naiades, 1: 66. Description. Shell large, often reaching over 100 mm in length. Outline subellipti- cal or subovate, moderately elongated, distinctly over one and a half times as long as high. Valves generally subinflated, though occasionally quite inflated, and solid. Anterior end regularly rounded; jdos- terior end more broadly rounded, and in females often somewhat expanded in the joostbasal region, though not uniformly so. Ventral margin generally regularly curved, sometimes straight, roughly parallel to the almost straight dorsal margin which forms a blunt angle with the posterior margin in immatures, and becomes almost impercep- tible with age. Hinge ligament prominent. Posterior ridge feeble or wanting. Posterior slojie broad and undistinguished. Umbos rather sharp, but not full or high, their sculpture consisting of six to ten double- looped bars, with a distinct re-entering angle in the middle. The posterior looj) is slightly angular and indistinct on the jdos- terior slope. Surface roughened by close concentric wrinkles of periostracum, gen- erally yellowish or brownish green, with dark greenish or blackish rays, both nar- row and wide, over the entire surface. Left valve with two stumrjy pseudo- cardinal teeth, both rather triangular and crenulate. No interdentum. Two long, slightly curved lateral teeth. Right valve 392 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 with one rather high, triangular pseudo- cardinal, and a vestigial one before it; one lateral tooth. Beak cavities very shallow, with a few dorsal muscle scars. Anterior and posterior adductor muscle scars and pallial line all distinct. Nacre generally white, bluish white, sometimes tinted with pink or salmon, sometimes quite pink or salmon. Length Height Width mm mm mm 138 86 46 112 64 39 99 52 29 84 49 27 62 39 20 Lake Michie, 12 mi. NNE Durham, Durham Co., North Carolina. Male. Yadkin River, near Salis- bury, Rowan Co., North Carolina. Holotype of U. conspicuus Lea. Male. Susquehanna River, Co- lumbia, York Co., Penn- sylvania. Male. Potomac River, Washing- ton, District of Columbia. Male. As above. Female. Anatomy. Figured by Lea ( 1838, pi. 15, figs. 48, 49). According to Ortmann ( 1911: 349) the anatomy of L. radiata radiata agrees "in all essential respects" with that of L. r. siliqaoidca (Barnes). The glochidia are suboval, without hooks, and measure 0.22 to 0.23 mm in length and 0.27 to 0.28 mm in height. The host fish is unknown. Breeding season. Ortmann (1919: 293) indicates that the breeding season begins in August and ends the following August. Habitat. Lives in rivers and lakes of all sizes, usually in gravel or sand, occasionally in mud. Prefers tidewaters, but ascends some of the larger rivers, such as the Susquehanna River, Pennsylvania, and the Hudson River, New York, for a consider- able distance. Remarks. In the Southern Atlantic Slope region Lampsilis radiata radiata (Gmelin) is replaced by Lampsilis splendida (Lea) below the Pedee River system. But the two species do not much resemble one another, as splendida has a sharp posterior ridge and r. radiata docs not. See: Remarks under splendida on p. 394. Lampsilis r. radiata, an Atlantic Slope species, most closely resembles Lampsilis radiata sili- quoidea (Barnes) which is found through- out the Interior Basin. Clarke and Berg (1959: 58-62, 68-70) have shown that in the Lower St. Lawrence drainage of New York where these two otherwise quite distinct "species" commingle, they inter- grade completely. Typically, L. r. radiata is easily dis- tinguished from L. r. siliquoidea. The former is more compressed; the difference between the male and female shell is much less marked; the periostracum is always rough and not shiny like that of siliquoidea. It has broader rays and often has a reddish or salmon nacre. In siliquoidea the nacre is always bluish white. Ortmann (1919: 296) suggested for the type locality of r. radiata Saratoga Lake in New York, "should there not be any earlier record," and was probably unaware that Simpson (1914, 1: 65) had already selected Virginia as the type locality. Ortmann and Walker (1922: 61) have shown that Unio luteolus Lamarck is not recognizable on the basis of the original description. Wheeler (1963: 58) has tried to resurrect this name on the basis of the identification of the type. Lamarck origi- nally gave two localities, the Susquehanna and the Mohawk rivers. The former has only L. r. radiata, and the latter, which is now in the region of hybridization, prob- ably did not have L. r. siliquoidea in it before the completion of the Erie Canal, some years after Lamarck's specimen was collected. Therefore, to avoid further con- fusion the type locality was restricted to the Susquehanna. See: synonymy above. Simpson (1914, 1: 66) treats Unio con- spicuus Lea as a variety of Lampsilis r. radiata, stating that "there are intermedi- ates, which seem to fully connect the two." I am in agreement with him. The main difference in the southern specimens seems to be a general tendency to grow large more consistently than elsewhere. There Atlantic Slope Unionidae • Johnson 393 does seem to be a hiatus for this species between the Neuse River, North Carolina, and the Potomac River, Virginia. Ortmann (1919: 296) calls attention to the fact that while this species is abundant in the tide- water region of the Potomac River, it does not go up the river into the mountains west of the Blue Ridge, and is rare or absent on the Piedmont Plateau. Range. Southern Atlantic Slope: dis- continuous, Pedee River system, South Carolina, Cape Fear and Neuse river systems, both North Carolina. Northern Atlantic Slope: Potomac River, Maryland, north to the St. Lawrence River system, westward to Lake Ontario. Specimens Examined Pedee River System Yadkin River Drainage. North Caro- lina: [Yadkin River], Salem, Forsyth Co. (MZUM). Yadkin River, near Salisbury, Rowan Co. (MZUM). Cape Fear River System Cape Fear River Drainage. North Caro- lina: Greenfield Pond, Wilmington, New Hanover Co. (USNM). Neuse River System Neuse River Drainage. North Carolina: Lake Michie [an impoundment on Flat River], 12 mi. NNE Durham Center, Dur- ham Co. (75). Potomac River System Potomae River Drainage. Virginia; Po- tomac River, Great Falls, Fairfax Co. District of Columbia: Potomac River. Vir- ginia: Mount Vernon, Fairfax Co. Lampsilis (Lampsilis) splendida (Lea) Plate 22: 4, 5 Unio splendidus Lea 1838, Trans. Amer. Philos. Soc, 6: 70, pi. 19, fig. 61 (Altamaha River, near Darien [Mcintosh Co.]; Altamaha [River], Liberty [now Long] Co.; both Georgia; figured holotvpe USNM 84893). Lea, 1838, Obs. Unio, 2: 70. Unio regularis Sowerby 1866, Conch. Iconica, 16, Unio, pi. 34, fig. 181, non Lea 1841. Lampsilis splendida Lea. Simpson, 1914, Cat. Naiades, 1 : 50. Description. Shell large, often exceeding 110 mm in length. Outline obovate and elongated, the female proportionally higher than the male. Valves considerably in- flated, subsolid. Anterior end rounded; posterior end of the male broader and very bluntly pointed; the female even more broadly rounded. Ventral margin usually slightly curved, or slightly arcuate in males when they are swollen behind the middle of the base. Dorsal margin straight, some- times forming a distinct angle with the obliquely descending posterior margin or merging gradually into it. Hinge ligament prominent, running under the umbos and appearing anteriorly. Posterior ridge rather sharp, generally with a second ridge above it. Posterior slope considerably wrinkled. Umbos very high and full, their sculpture consisting of a number of strong, nearly straight bars. Surface of the disk some- times smooth, but the periostracum is usually concentrically wrinkled, especially posteriorly and toward the margins. Perios- tracum yellowish green to brownish with the entire surface covered with numerous wide and narrow green rays. Old shells sometimes dark brown with the rays ob- scured. Left valve with two compressed pseudo- cardinal teeth slightly anterior of the umbos, one in front of the other, the an- terior one somewhat triangular; the hinder one considerably lower. Hinge line very narrow, before two short, curved, lateral teeth. Right valve with two triangular, narrow, parallel pseudocardinals separated by a deep narrow pit, the more anterior tooth inclined to be vestigial; the hinder tooth inclined to be chunky in old speci- mens. One rather high lateral tooth present. Beak cavities deep and wide, with dorsal muscle scars. Anterior adductor muscle scars well impressed, posterior ones less so. Pallial line distinct. Nacre white, orange, 394 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 violet, sometimes bluish, but more inclined to be pinkish and iridescent. Length Height Width 112 68 54 Ocmulgee River, 1 mi. S Lumber City, Telfair Co., Georgia. Female. 113 62 59 As above. Male. 73 46 36 As above. Female. 70 42 33 As above. Male. Anatomy. Discussed by Lea (1863: 414). Habitat. Usually lives in the swift waters of large rivers on sand bars and gravelly bottoms, but also lives in lakes. Remarks. Lampsilis splendida (Lea) does not closely resemble any of the Lamp- silis in either the Apalachicolan or Atlantic Slope regions. It replaces Lampsilis radiata radiata (Gmelin) in the Atlantic Slope region below the Pedee River system, South Carolina. L. splendida resembles L. r. radiata in its tendency to have rays which cover the entire shell, but r. radiata does not have a posterior ridge, while splendida often has quite an angular one. In splendida the shell of the female is proportionally higher than that of the male and more blunt behind, while in radiata the main difference in the shell is a tend- ency for the female to be expanded in the postbasal region. Young specimens of L. ochracea (Say) can be confused with splendida when the former are rayed over the entire surface, but the rays of splendida are narrow and wide, while those of ochracea are rather uniformly narrow. L. splendida has a much heavier shell and even in immatures the shell is heavier than the shells of ochracea of corresponding size. Range. Southern Atlantic Slope: Alta- maha River system, Georgia, north to the Cooper-Santee River system, South Caro- lina. Specimens Examined Altamaha Ringer System Ocmulgee River Drainage. Georgia: Limestone Creek, Brumbys Mill, 4 mi. NE Hawkinsville; Ocmulgee River, Hawkins- ville; Mosquito Creek, 8 mi. SE Hawkins- ville; all Pulaski Co. Dicksons Creek, 10 mi. NE Fitzgerald; House Creek, Bowens Mill, 9 mi. N Fitzgerald; both Ben Hill Co. Ocmulgee River, Jacksonville; Ocmulgee River, 1 mi. S Lumber City; both Telfair Co. Oconee River Drainage. Georgia: Oco- nee River, 2.5 mi. N Glenwood, Wheeler Co. Altamaha River Drainage. Georgia: Altamaha River, 7 mi. N Glenwood, Wheeler Co. Altamaha River, 10 mi. NE Surrency, Appling Co. Altamaha River, "Riverside Park," 4 mi. N Jesup, Wayne Co. Altamaha River, 3 mi. NW Everett City, Glynn Co. Altamaha River, Fort Barrington; Altamaha River, Hopeton, near Darien; both Mcintosh Co. Oceechee River System Ogeechee River Drainage. Georgia: Re- ported by Simpson (1914: 51) but the specimens were not located in the USNM. Savannah River System Savannah River Drainage. Georgia: Savannah River, 0.75 mi. SE Augusta, Rich- mond Co. South Carolina: Savannah River, 2 mi. SW Millettville; Savannah River, Johnsons Landing, 10 mi. W Allendale; Savannah River, Kingjaw Point, 10 mi. WSW Allendale; all Allendale Co. (all ANSP). Georgia: Savannah River, 7.5 mi. NE Shell Bluff; Savannah River, 6 mi. NE Girard; both Burke Co. Cooper-Santee River System Wateree River Drainage. Soutli Caro- lina: Wateree River, 2.5 mi. W Camden, Kershaw Co. Santee River Drainage. South Carolina: Lake Moultrie, Cross; Santee River, 1 mi. N Moncks Corner; Bunnon Lake [not lo- cated], near Alvin (ANSP and USNM); all Berkeley Co. SELECTED BIBLIOGRAPHY The references were selected to include those which are briefly cited in the text, as well as those which are especially relevant to the area under study. Adams, C. C. 1901. Baseleveling and its faunal significance, with illustrations from the Southeastern United States. Amer. Natur., 35: 839-851. Adams, G. I. 1926. In Adams, G. I., et al., Geology of Alabama. Geol. Survey Alabama, Spec. Rept. no. 14: 25-27. Alt, D. 1968. Pattern of Post-Miocene eustatic fluctuation of sea level. Paleogeog., Paleoclim., Paleoecol, 5: 87-94. Alt, D., and H. K. Brooks. 1965. Age of Florida marine terraces. Tour. Geol., 73: 406-411. Athearn, H. D. 1964. Three new unionids from Alabama and Florida and a note on Lamp- silis joncsi. Nautilus, 77: 134-139. Athearn, H. D., and A. H. Clarke, Jr. 1962. The freshwater mussels of Nova Scotia. Natl. Mus. Canada, Bull. 183, Cont. Zool., 1960- 61: 11-41. Bailey, J. L., Jr. 1940. Wilmington, N. G, Records. Nautilus, 54: 69. Elliptio fisheri- (11111.1 Lea [= Elliptio htnccolata (Lea)], Lake Greenfield, Wilmington, N. G, is the only unionid mentioned. Ball, G H. 1922. Variation in freshwater mus- sels. Ecology, 3: 93-121. Bates, J. M. 1966. A new species of Canmculina (Unionidae: Pelecypoda) from the Savannah River, South Carolina. Occ. Pap. Mus. Zool., Univ. Mich., no. 646: 1-9. Boss, K. J., and W. J. Clench. 1967. Notes on Pleurohema collina (Conrad) from the James River, Virginia. Occ. Pap. Moll., Mus. Comp. Zool., 3: 45-52. Call, R. E. 1896. A revision and synonymy of the parvus group of Unionidae. Proc. Indiana Acad. Sci., 1895: 109-119. Campbell, M. R. 1896. Drainage modifications and their interpretation. Jour. Geol., 4: 567- 581. Cherry, R. N. 1961. Chemical quality of water of Georgia streams, 1957-58. Georgia State Division of Conservation. Geol. Survey Bull, no. 69: 1-100. Clarke, A. H., Jr., and G O. Berg. 1959. The freshwater mussels of central New York. Cornell Univ. Agr. Expt. Sta., Mem. no. 367: 1-79. Clarke, A. H., Jr., and A. M. Rick. 1963. Sup- plementary records of Unionacea from Nova Scotia with a discussion of the identity of Atlantic Slope Unionidae • Johnson 395 Anodonta fragilis Lamarck. Natl. Mus. Canada, Bull. 199. Cont. Zool., 1963: 15-27. Clench, W. J. 1962. Collecting freshwater mol- lusks in south central Georgia. Shells and their neighbors, 12: 1, 7. [A popular publi- cation, discontinued in 1964.] Clench, W. J., and K. J. Boss. 1957. Freshwater Mollusca from James River, Virginia. Nauti- lus, 80: 99-101. Clench, W. J., and R. D. Turner. 1956. Fresh- water mollusks of Alabama, Georgia and Florida from the Escambia to the Suwannee River. Bull. Florida State Mus., 1: 97-239. Conrad, T. A. 1834 (May). New fresh water shells of the United States with colored illustrations, and a monograph of the genus Anculotus of Say; also a synopsis of the American naiades. Philadelphia, Penn., pp. 1-76. . 1835 (October), ibid. Appendix, pp. 1-8, pi. 9 (col.). . 1836-40. Monography of the family Unionidae, or naiades of Lamarck, ( Fresh water bivalve shells ) of North America. Phila- delphia, Penn., pp. i-v, 1-118 [pp. 13-16 never printed]. Published in thirteen parts. . 1846. Notices of fresh water shells, etc., of Rockbridge Co., Virginia. Amer. Jour. Sci., ser. 2, 1: 405-407. 1853. A synopsis of the family of naiades of North America, with notes, and a table of some of the genera and sub-genera of the family, according to their geographical dis- tribution, and descriptions of genera and sub- genera. Proc. Acad. Nat. Sci. Phila., 6: 243- 269. Cooke, C. W. 1925. Physical geography of Georgia: The coastal plain. Geol. Survey Georgia, Bull., no. 42: 19-54. . 1930. Pleistocene seashores. Jour. Wash. Acad. Sci., 20: 389-395. . 1936. Geology of the coastal plain of South Carolina. Geol. Survey, Washington, D. G, Bull., no. 867: 1-12. . 1943. Geology of the coastal plain of Georgia. Geol. Survey, Washington, D. G, Bull., no. 941: 103-112. 1945. Geology of Florida. Florida Geol. Survey, Geol. Bull., no. 29: 1-339. Cvancara, A. M. 1963. Clines in three species of Lampsilis (Pelecypoda: Unionidae). Mala- cologia, 1: 215-225. Dawley, G 1965. Checklist of freshwater mol- lusks of North Carolina. Sterkiana, no. 19: 35-39. An uncritical list; most of the identi- fications appear to be based on whatever appeared on the labels of specimens examined or are from the literature. 396 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 Emery, K. O, 1967. The Atlantic continental margin of the United States dining the past 70 million years. Geol. Assn. of Canada, Special Paper 4, Geol. of the Atlantic Region, pp. 53-70. Flint, R. F. 1957. Glacial and Pleistocene ge- ology. New York, pp. i-xiii, 1-553. Frierson, L. S. 1912. Notes on Anodonta couperiana and A. gibbosa. Nautilus, 25: 129-130. . 1915. Lasmigona subviridis Conrad, redi- vivus. Nautilus, 29: 57-59. -. 1927. A classified and annotated check list of the North American naiades. Waco, Texas: Baylor Univ. Press, pp. 1-111, errata sheet. Includes a reference to almost all of the taxa, within its scope, descrihed before 1927. Gibbes, L. R. 1848. Mollusea [of South Carolina]. In Tuomey, M., Report on the geology of South Carolina. Appendix, pp. xix-xxii. Goodrich, C. 1930. Unio spinosa Lea. Nautilus, 43: 140. . 1939. Certain mollusks of the Ogeechee River, Georgia. Nautilus, 52: 129-131. Nothing on Unionidae, but refers to van der Schalie's expedition of 1937. Goodrich, C., and H. van der Schalie. 1944. A revision of the Mollusea of Indiana. Amer. Midland Nat., 32: 257-326. Gould, A. A. 1848. Descriptions of shells found in Connecticut, collected and named by the late Rev. J. H. Linsley. Amer. Jour. Sci., ser. 2, 6: 233-236. Grier, N. M., and J. F. Mueller. 1926. Further studies in correlation of shape and station in fresh water mussels. Bull. Wagner Free Instit. Sci., Philadelphia, 1: 11-28. Haas, F. 1969a. Superfamilia Unionacea. In Das Tierreich. Berlin. Lief. 88: I-X, 1-663. . 1969b. Superfamily Unionacea. In Treatise on Invertebrate Paleontology, Part N. Mollusea 6 (Bivalvia), pp. N411-N471. Hayes, C. W., and M. R. Campbell. 1894. Geomorphology of the southern Appalachians. Natl. Geographic Mag., 6: 63-126. Heard, W. II. 1966. Population sexuality in Anodonta. Amer. Malacol. Union, Ann. Rept. for 1966: 31-33. Hoyt, J. H., and J. R. Hails. 1967. Pleistocene shoreline sediments in coastal Georgia: Depo- sition and modification. Science, 155: 1541- 1543. Johnson, R. I. 1946. Anodonta implicata Say. Occ. Papers Moll., Mus. Comp. Zool., I: 109- 116. . 1947a. Lampsilis cariosa Say and Lamp- silis ochracea Say. Occ. Pap. Moll., Mus. Comp. Zool., 1 : 145-156. — . 1947b. The audiorship of EHiptio com- planatus. Nautilus, 62: 36. — . 1956. The types of Naiades (Mollusea: Unionidae) in the Museum of Comparative Zoologv. Bull. Mus. Comp. Zool., 115: 101- 142. . 1965. A hitherto overlooked Anodonta (Mollusea: Unionidae) from the Gulf drain- age of Florida. Breviora, Mus. Comp. Zool., no. 213: 1-7. . 1967a. Illustrations of all the mollusks described by Berlin Hart and Samuel Hart Wright. Occ. Pap. Moll., 3: 1-35. . 1967b. Additions to the unionid fauna of the Gulf drainage of Alabama, Georgia and Florida (Mollusea: Bivalvia). Breviora, Mus. Comp. Zool., no. 270: 1-21. . 1967c. Carunculina pulla (Conrad), an overlooked Atlantic drainage unionid. Nauti- lus, 80: 127-131. . 1968. EHiptio nigella, overlooked unionid from Apalaehicola River system. Nautilus, 82: 22-24. . 1969a. Further additions to the unionid fauna of the Gulf drainage of Alabama, Georgia and Florida. Nautilus, 83: 34-35. 1969lx Illustrations of Lamarck's types of North American Unionidae mostly in the Paris Museum. Nautilus, 83: 52-61. Keith, A. 1925. Physical geography of Georgia: The highland. Geol. Survey Georgia, Bull., no. 42: 93-132. Kurten, B. 1966. Holarctic land connexions in the early Tertiary. Comm. Biol. Soc. Sci. Fennica, 29(5): 1-5. Laessle, A. M. 1968. Relationship of sand pine scrub to former shore lines. Quart. Jour. Florida Acad. Sci., 30: 270-286. LaForce, L. 1925. Physical geography of Geor- gia: The central upland. Geol. Survey Georgia, Bull., no. 42: 57-92. Lea, I. 1834-74. Observations on the genus Unio. Philadelphia, Penn., 1-13. A reprint of Lea's papers from various journals. They are repaged, but the plate and figure num- bers are the same as those in the original journals. . 1838. [Notes on the anatomical structure of Naiades]. Trans. Amer. Philos. Soc, 6: 48-57; also 1838, Obs. Unio, 2: 48-57. . 1854. Rectification of Mr. T. A. Conrad's "Synopsis of the family of naiades of North America," published in the "Proceedings of the Academy of Natural Sciences of Phila- delphia," February, 1853. Proc. Acad. Nat. Sci. Phila., 7: 236-249. . 1858. Descriptions of the embryonic forms of thirty-eight species of Unionidae. Atlantic Slope Unionidae • Johnson 39" Jour. Acad. Nat. Sci. Phila., ser. 2, 4: 43-50, pi. 5; also 1858, Obs. Unio, 6: 43-50. 1863. Descriptions of the soft parts of one hundred and forty-three species and some embryonic forms of Unionidae of the United States. Jour. Acad. Nat. Sci. Phila., ser. 2, 5: 401-456; also 1863, Obs. Unio, 10: 37-92. Linsley, J. H. 1845. Catalogue of the shells of Connecticut. Amer. Jour. Sci., 48: 271-286. MacNeil, F. S. 1950. Pleistocene shore lines in Florida and Georgia. Geol. Survey, Washing- ton, D. C, Prof. Paper 221-F, pp. 95-106. Marshall, W. B. 1890. Beaks of Unionidae in- habiting the vicinity of Albany, New York. Bull. New York State Mus., 2: 169-189. . 1917. Lampsilis oentricosa cohongoronta in the Potomac River. Nautilus, 31: 40-41. . 1918. Lampsilis ventricosa cohongoronta in the Potomac Valley. Nautilus, 32: 51-53. . 1930. Lampsilis ventricosus cohongoronta in the Potomac River. Nautilus, 44: 19-21. Matteson, M. P. 1948a. The taxonomic and distributional history of the freshwater mussel EUiptio complanatus ( Dillwyn, 1817). Nauti- lus, 61: 127-132; 62: 13-17. . 1948b. Life history of EUiptio com- planatns (Dillwyn, 1817). Amer. Midland Nat., 40: 690-723. . 1955. Studies on the natural history of the Unionidae. Amer. Midland Nat., 53: 126-145. Mayr, E., E. G. Lixsley, and R. L. Usinger. 1953. Methods and Principles of Systematic Zoology. New York, McGraw-Hill, pp. 1- 336. Mazyck, W. G. 1913. Catalog of Mollusca of South Carolina. Contributions from the Charleston Museum No. 2: 1-39. [Includes an uncritical list of Unionidae based on the literature and on previous identifications of specimens in the museum.] McCallie, S. W. 1925. Physical geography of Georgia. Introduction: The state as a whole. Geol. Survey Georgia, Bull., no. 42: 1-17. Modell, H. 1942. Das natiirliche system der Najaden. 1. Archiv fur Molluskenkunde, 74: 161-191. . 1949. Das natiirliche system der Najaden. 2. Archiv fur Molluskenkunde, 78: 29-48. -. 1964. Das natiirliche system der Najaden. 3. Archiv fur Molluskenkunde, 93: 71-126. [Authorized English translation of part 1 by D. H. Stansbery and U. Soehngen, 1964, Sterkiana, no. 14: 1-18.] Murray, G. E. 1961. Geology of the Atlantic and Gulf Coastal Province of North America. New York. pp. 1-523. Myers, G. S. 1938. Fresh-water fishes and West Indian zoogeography. Smithsonian Rept., [for] 1937: 339-364. Oaks, R. Q., and N. K. Coch. 1963. Pleistocene sea levels, southeastern Virginia. Science, 140: 979-983. Ortmann, A. E. 1906. The crawfishes of the state of Pennsylvania. Mem. Carnegie Mus., 2: 343-524. . 1909. The breeding season of Unionidae in Pennsylvania. Nautilus, 22: 91-95, 99- 103. . 1910. A new system of the Unionidae. Nautilus, 23: 114-120. . 1911. A monograph of the najades of Pennsylvania. Parts 1 and 2. Mem. Carnegie Mus., 4: 279-347. . 1912a. Notes upon the families and genera of the najades. Ann. Carnegie Mus., 8: 222-365. . 1912b. The geological origin of the fresh-water fauna of Pennsylvania. Penn. Topog. Geol. Survey, 1910-12: 130-149. . 1913a. The Alleghenian Divide and its influence upon the freshwater fauna. Proc. Amer. Philos. Soc, 52: 287-390. [This is a classic study, with which I have noted only a few minor disagreements.] . 1913b. Studies in najades. Nautilus, 27: 88-91. -. 1914. Studies in najades. Nautilus, 28: 41-47, 65-69, 129-131. . 1915. Studies in najades. Nautilus, 29: 63-67. . 1919. A monograph on the naiades of Pennsylvania. Part 3. Systematic account of the genera and species. Mem. Carnegie Mus., 8: i-xiv, 1-384. . 1920. Correlation of shape and station in freshwater mussels. Proc. Amer. Philos. Soc. 19: 269-312. . 1923. The anatomy and taxonomy of certain Unionidae and Anodontinae from the Gulf drainage. Nautilus, 36: 73-84. -. 1923-1924. Notes on the anatomy and taxonomy of certain Lampsilinae from the Gulf drainage. Nautilus, 37: 56-60, 99-104, 137-144. Ortmann, A. E., and B. Walker. 1922. On the nomenclature of certain North American naiades. Occ. Pap. Mus. Zool. Univ. Mich., no. 112: 1-75. Parodiz, J. J. 1967. Types of North American Unionidae in the collection of the Carnegie Mu- seum. Sterkiana, no. 28: 21-30. [A lectotype is selected for L. v. cohongoronta Ortmann male: C. M. 61.3999 and female: C. M. 61.4000 as allotype, p. 28. The author was unaware that I had published a list of the 398 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 types of B. H. and S. H. Wright (Johnson, 1967a) a few months previous to his paper.] Parodiz, J. J., and A. A. Bonetto. 1963. Taxon- omy and zoogeographic relationships of the South American naiades (Pelecypoda: Unio- aacea and Mutelacea). Malaeologia, 1: 179- 213. Patrick, R., J. Cairns, Jr., and S. Roback. 1967. An ecosystematic study of the fauna and flora of the Savannah River. Proc. Acad. Nat. Sci. Phila., 118: 109-407. [Unionidae, pp. 345- 346.] [The following species of Unionidae are listed from several localities: (1) Elliptio hopetonensis (Lea)=E. complanata (Light- foot), (2) E. incrassatus (Lea) = E. con- garaea (Lea), (3) E. lanceolata (Lea), (4) Uniomerus obesus (Lea) = U. tetralasmus (Say), (5) Anodonta couperiana Lea, (6) A. hallenbecki Lea = A. cataracta Say, (7) A. imbecilis Say, (8) Lampsilis eariosa (Say), (9) L. splendida (Lea), (10) Villosa ogee- cheensis (Conrad) = V. delumbis (Conrad).] Pilsbry, H. A. 1894. Critical list of mollusks collected in the Potomac Valley. Proc. Acad. Nat. Sci. Phila., 46: 11-30. Rafinesque, C. S. 1831. Continuation of a monograph of the hivalve shells of the river Ohio, and other rivers of the western states. Philadelphia, Penn. pp. 1-7. [This, and the papers from the various journals, were re- printed by: Binney, William C, and George W. Tryon. 1864. The complete writings of Constantine Smaltz Rafinesque on recent and fossil conchology. Philadelphia, Penn. pp. 1-96, 7, 3 pis.] Raulerson, L., and W. D. Burbanck. 1962. The life cycle and ecology of Elliptio hopetonen- sis Lea. Assoc. Southeast. Biol., 9: 39. [An abstract of a master's thesis, "The ecology of a small Georgia mountain stream, with special emphasis on the mussel, Elliptio hopetonensis Lea," by Claire Lynn Raulerson ( 1960, Emory University, Atlanta, Georgia ) . I misidentified these specimens; they are E. complanata ( Lightfoot ) .] Ravenel, E. 1874. Catalogue of the recent and fossil shells in the cabinet of Edmund Ravenel. [Edited by L. R. Gibbes] Charles- ton, South Carolina, pp. 1-67. [Not seen. Includes a nomen nudum, Alasmidonta coarctata Ravenel (rivers in the upper part of South Carolina), p. 58.] Reardon, L. 1929. A contribution to our knowl- edge of the anatomy of the fresh-water mussels of the District of Columbia. Proc. U. S. Natl. Mus., 75: 1-12. Rehder, H. A. 1949. Some land and freshwater mollusks from the coastal region of Virginia and North and South Carolina. Nautilus, 62: 121-126. [Only two Unionidae are noted: Lampsilis ochraceus (Say) and Elliptio com- planatus quadrilaterus (Lea)=E. complan- ata ( Lightfoot ) , both from Lake Waccamaw, North Carolina.] -. 1967. Valid zoological names of the Portland catalogue. Proc. U. S. Natl. Mus., 121, 1-15. Russell, R. J. 1957. Instability of sea-level. American Scientist 45: 414-430. Say, Thomas. 1830-34. American conchology, or descriptions of the shells of North America, il- lustrated by colored figures. New Harmony, Indiana. 68 col. pis. with letterpress. Pub- lished in 7 parts. [Reprinted by: Binney, William G. 1858. The complete writings of Thomas Say on the conchology of die United States. New York. 71 col. and 4 plain pis.] SCHNABLE, J. E., AND H. G. GOODELL. 1968. Pleistocene-Recent stratigraphy, evolution and development of the Apalachicola Coast, Florida. Geol. Soc. Amer., Special paper 112, pp. 1-72. Sellards, E. H., and H. Gunter. 1918. Geology between the Apalachicola and Ochlockonee Rivers in Florida. Geology between the Choctawhatchee and Apalachicola Rivers in Florida. lOth-llth Ann. Rept. Florida Geol. Surv., pp. 9-56, pp. 77-102. Sickel, J. B. 1969. A survey of the mussel popu- lations ( Unionidae ) and Protozoa of the Altamaha River with references to their use in monitoring environmental changes, pp. 2 [Mimeographed abstract of a master's thesis, Emory University, Atlanta, Georgia.] Simpson, C. T. 1892. Notes on the Unionidae of Florida and the southeastern states. Proc. U. S. Natl. Mus., 15: 405-436. [Many of the figures are line drawings of the types.] . 1900. Synopsis of the naiades, or pearly freshwater mussels. Proc. U. S. Natl. Mus., 22: 501-1044. -. 1914. A descriptive catalogue of the naiades or pearly freshwater mussels. Detroit, Michigan. Parts 1-3, pp. xi, 1-1540 Simpson, G. B. 1884. Anatomy and physiology of Anodonta fluviatilis. Thirty-fifth Ann. Rept. New York State Mus. Nat. Hist., pp. 169-191. Stephenson, L. W. 1912. In W. B. Clark, et al., The Coastal Plain of North Carolina. North Carolina Geol. Econ. Surv., 3, The Quater- nary Formations, pp. 266-290. Swain, G. F., et al. 1899. Papers on the water- power in North Carolina. North Carolina Geol. Survey: Bull., no. 8. Thomas, G. J., and D. C. Scott. 1965. Note on EUiptio spinosa in Georgia. Nautilus, 79: 66-67. Tomkins, I. R. 1955. Elliptic spinosus in the Altamaha River. Nautilus, 68: 132-133. Vanatta, E. G. 1915. Rafinesque's types on Unio. Proc. Acad. Nat. Sci. Phila., 67: 549-559. van der Schalie, H. 1938. The naiad fauna of the Huron River, in southeastern Michi- gan. Misc. Pub. Mus. Zool., Univ. Michigan, no. 40: 1-83. . 1940. The naiad fauna of the Chipola River, in northwestern Florida. Lvoydia, 3: 191-208. [Published by the Lyoyd Library, Cincinnati, Ohio.] . 1945. The value of mussel distribution in tracing stream confluence. Pap. Michigan Acad. Sci. Arts Letters, 20: 355-373. . 1966. Hermaphroditism among North American freshwater mussels. Malacologia, 5: 77-78. van der Schalie, H., and A. van der Schalie. 1950. The mussels of the Mississippi River. American Midland Nat., 44: 448—466. Vaughn, T. W. 1910. A contribution to the geologic history of the Floridian plateau. Carnegie Inst., Washington, Pub. no. 133: 99-185. Walker, R. 1905. List of shells from north- western Florida. Unionidae from Moccasin Creek, a tributary of the Econfine River. Nautilus, 18: 136. . 1910. The distribution of Margaritana margaritifcra ( Linn. ) in North America. Proc. Mai. Soc. London, 9: 126-145. -. 1918. A synopsis of the classification of the fresh-water Mollusca of North America. Misc. Pub. Mus. Zool. Univ. Michigan, no. 6: 1-213. . 1919. Notes on North American naiades. II. Unio tenerus Ravenel and Lampsilis modioUformis (Lea). Occ. Pap. Mus. Zool. Univ. Michigan, no. 74: 1-8. Walter, W. M. 1956. Mollusks of the upper Neuse River Rasin, North Carolina. Jour. Elisha Mitchell Sci. Soc, 72: 262-274. Walter, W. M., and R. A. Parker. 1957. EUip- tio complanatus roanokensis in the Neuse River. Nautilus, 71: 60-64. Wentworth, C. K. 1930. Sand and gravel re- sources of the coastal plain of Virginia. Vir- ginia Geol. Survey, Rull. no. 32: 1-146. Wheeler, H. E. 1935. Timothy Abbott Conrad, widi particular reference to his work in Alabama one hundred years ago. Rull. Amer. Paleont., 23(77): 1-157. Wheeler, M. J. 1963. Type of Unio luteolus Lamarck 1819. Nautilus, 77: 58-61. Atlantic Slope Unionidae • Johnson 399 INDEX TO RELEVANT TAXA Principal references in boldface type. abbevillensis Lea, Unio, 315 aberrans Lea, Unio, 317 ablatus Lea, Unio, 327 Acglia Swainson, 382 aequatus Lea, Unio, 326 alabumensis Lea, Margaritana, 366 Alasmidonta Say, 346, 348 Alasminota Ortmann, 343 Alasmodonta Say, 348 ampins Lea, Unio, 318 angulata (Lea), Gonidea, 265 angustatus Lea, Unio, 333 annulatus Sowerby, Anodon, 367 Anodonta Lamarck, 356 anodontoides (Lea), Lampsilis, 271 anthonyi Lea, Unio, 312 Anodontinae (Swainson) Ortmann, 343 aquilus Lea, Unio, 326 arctata (Conrad), EUiptio (EUiptio), 271, 274, 276, 331 arctatus Conrad, Unio, 331 arcula (Lea), Alasmidonta (Alasmidonta), 27 '4, 278, 352 arcula Lea, Margaritana, 352 areolatus Swainson, Anodon, 367 arkansasensis Lea, Anodonta, 367 atropurpureum Rafinesque, Alasmidon, 353 attentuata Rafinesque, Obliquaria, 380 attenuatus Lea, Unio, 380 aurata Rafinesque, Unio, 315 australis Simpson, Lampsilis, 272 averellii B. H. Wright, Unio, 373 baldwinensis Lea, Unio, 316 barrattii Lea, Unio, 326 basalis Lea, Unio, 318 beaverensis Lea, Unio, 318 binominatus Simpson, Lampsilis, 273, 385 bisselianus Lea, Unio, 340 blandingianus Lea, Unio, 339 boydianus Lea, Unio, 391 boijkiniana (Lea), Amblema, 271 brimleyi S. H. Wright, Unio, 301 buddianus Lea, Unio, 339 Bullella Simpson, 348 burkei Walker, Quincuncina, 273 hurkensis Lea, Unio, 316 burtchianus S. H. Wright, Unio, 327 buxcus Lea, Unio, 308 calceola (Lea), Alasmidonta, 279, 349 camptodon Say, Unio, 339 canadensis Lea, Unio, 386 Canthijria Swainson, 303 cardium Rafinesque, Lampsilis, 386 carinifera Lamarck, Unio, 314 400 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 cariosa (Say), Lampsilis (Lampsilis), 274, 275, 276, 278, 279, 382 cariosus Say, Undo, 382 CaruncuUna Baker, 369 castas Lea, Unio, 301 cataracta Say, Anodonta (Pyganodon) cataracta, 272, 274, 276, 278, 279, 286, 356 catawbensis Lea, Unio, 315 charlottensis Lea, Unio, 344 chathamensis Lea, Unio, 317 chipolancnsis (Walker), Elliptio, 273 choctawensis Athearn, Villosa, 273 dear Lea, Unio, 340 cirratus Lea, L/n/o, 319 cistelliformis Lea, l/n/o, 317 claibornensis (Lea), Lampsilis, 271, 382 coarctata Lamarck, [/mo, 314 cohongoronata Ortmann, Lampsilis ventricosa 386 collina Conrad, Pleurobcma (Lexingtonia), 274, 277, 300 collinus Conrad, Unio, 300 columbensis Lea, l/nio, 340 complanata ( Lightfoot ) , Elliptio (Elliptio), 272, 274, 278, 279, 286, 314 complanata (Lightfoot), Mija, 314 compressa (Lea), Lasmigona, 280, 343, 345 compressa Lea, Symphynota, 343 concavus Lea, l/n/o, 376 confertus Lea, l/rao, 325, 329 congaraea (Lea), Elliptio (Elliptio), 274, 276, 277, 278, 308 congaraeus Lea, l/n/o, 308 conica Rafinesque, Pleurobcma, 299 conspicuus Lea, l/n/o, 391, 394 constricta (Conrad), V/Z/o.sa, 274, 277, 278, 280, 285, 378 constrictus Conrad, Unio lienosus, 378 contiguus Lea, Unio, 376 contractus Lea, l/n/o, 316 contrarius Conrad, l/n/o, 382 corneas Lea, C/ra'o, 319 corrugata DeKay, Alasmidon, 354 Corunculina Simpson, 370 coruscus Gould, Unio, 326 corvus Lea, l/n/o, 308 costata Rafinesque, Ala-smidonta, 243 coupcriana Lea, Anodonta ( Utterbackia ) , 272, 274, 275, 276, 365 cowperiana Lea, Anodonta, 365 crassidens (Lamarck), Elliptio (Elliptio) cras- sidens, 267, 271, 277, 305 crassidens- var. b. Lamarck, C/mo, 305 crocatus Lea, l/n/o, 382 cuneata Rafinesque, Pleurobcma, 299 cuncalus Barnes, Unio, 305 Cunicula Swainson, 305 cuprea Rafinesque, Obliquaria, 333 curatus Lea, l/n/o, 317 curvatus Lea, L'mo, 318 cuspidatus Lea, L/n/o, 327 cuvierianus Lea, L/n/o, 315 cygneus Linnaeus, Mytilus, 356, 362 cylindraceus Frierson, Elliptio, 327 danielsii B. H. Wright, C/ra'o, 305 dariensis Lea, Anodonta, 357 dariensis (Lea), Etf/pf/o (Elliptio), 274, 277, 278, 310 dariensis Lea, l/n/o, 310 datus, Lea, L/n/o, 318 declivis Say, l/n/o, 339 decoratus Lea, l/n/o, 344 Decurambis Rafinesque, 353 delumbis Conrad, l/n/o, 375 delumbis (Conrad), V///osa, 274, 276, 277, 375 diazensis S. H. Wright, l/n/o, 327 differtus Lea, l/n/o, 318 dilatata (Rafinesque), Elliptio, 270, 279, 280, 305 dilatatus Rafinesque, Unio, 305 discus Sowerby, Unio, 305 dispalans B. H. Wright, l/n/o, 327 dispar Lea, L/n/o, 373 dissimilis Lea, l/n/o, 319 dolabraeformis (Lea), Lampsilis (Lampsilis), 274, 278, 384 dolabraeformis Lea, L/n/o, 384 dolearis Lea, Anodonta, 357 doliaris Lea, Anodonta, 357 dooleyensis Lea, l/n/o, 319 dorsatus Lea, L/n/o, 308 downiei (Lea), Elliptio (Elliptio) crassidens, 267, 273, 307 downiei Lea, L/n/o, 307 dunlapiana Lea, Anodonta, 365 duttonianus Lea, l/n/o, 333 edentula Say, Alasmidonta, 367 cdentulus Say, Strophitus, 368 elect rinus Reeve, l/n/o, 340 £///p//o Refinesque, 303, 304 clongata Goodrich, l/n/o, 391 emmonsii Lea, l/m'o, 333 errans Lea, l/n/o, 315 cscambia Clench and Turner, Fusconaia. 272 Eunjria Rafinesque, 305 exact us Lea, l/n/o, 316 exacutus Lea, l/n/o, 334 excavatus (Lea), Lampsilis, 271, 385 excultus Conrad, l/n/o, 339 excurvata DeKay, Anodon, 356 exiguus Lea, l/n/o, 373 extensus Lea, C/rwo, 333 fishcrianus Kuester, l/n/o, 280 fishcrianus Lea, C/rao, 333, 334 fisheropsis De Gregorio, l/n/o arctior, 334 floridensis (Lea), Lampsilis anodontoides, 270 fluviatilis Gmelin, Mytilus, 388 fluviatilis Green, l/n/o, 315 folliculatus Lea, Unto, 333 forbesianus Lea, Unto, 308 fragilis Lamarck, Anodonta cataracta, 266, 279, 357 fraterna (Lea), EUiptio (Elliptic), 272, 274, 276, 277, 312 fraternus Lea, Unto, 312 fryanus B. H. Wright, C/niO', 327 fuliginosus Lea, C/nio, 315 fulvus Lea, C/nio, 308 fumatus Lea, C/nio, 316 /(/.«•« Baker, Utterbackia imbecUlis, 363 fuscatus Lea, C/nio, 326 Fusconaia Bafinesque, 300 gastonensis Lea, C/nio, 317 geddingsianus Lea, C/nio, 326 geminus Lea, C/nio, 315 genthii Lea, C/nio, 378 genuinus Lea, C/nio, 376 geometricus Lea, C/nio, 339 georgina Lamarck, C/nio, 314 gesnerii Lea, Anodonta, 356 gesnerii Lea, C/nio, 319 gibbesianus Lea, C/nio, 308 gibbosa Say, Anodonta (Pyganodon), 27 '4, 278, 359 glabrata Lamarck, C/nio, 314 glabratus Sowerby, C/nio, 349 globosa Lea, Anodonta, 356 gracilenius Lea, C/nio, 332 gracilior Lea, C/nio, 373 grandis Say, Anodonta, 271, 279, 360 griffithianus Lea, C/nio, 315 haddletoni Athearn, Lampsilis, 373 hallenbeckii Lea, Anodonta, 356, 358 hallenbcckii Lea, C/nio, 316 hartwrightii B. H. Wright, C/nio, 310 hazelhurstianus Lea, C/nio, 333 hefe, 333 waccamaivensis (Lea), Elliptio (Elliptio), 274, 278, 313 waccamaivensis Lea, £7nio, 313 walkeri (Wright), Medionidus, 272 wardiana Lea, Anodonta, 367 watereensis Lea, t/ra'o, 325 websterii B. H. Wright, l/n/o, 310 iveldonensis Lea, Unio, 317 ivheatleyi Lea, l/m'o, 315 whiteianus Lea, C/n/o, 326 williamsii Lea, Anodonta, 357 winnebagoensis Baker, Lampsilis ventricosa, 386 winnebagoensis Baker, Strophitus rugosus, 367 wrightiana (Walker), Alasmidonta, 273, 350 wrightianus Clench and Turner, Strophitus, 351 wrightianus Walker, Strophitus, 348 yadkinensis Lea, L/mo, 318 406 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 1 The Southeastern United States A.M. The solid line indicates the limit of the Appalachian Mountains. The divide is somewhat to the west of it. P.P. Piedmont Plateau. The line thus indicated, , is the Fall Line. C.P. Coastal Plain. The line thus indicated, '-■-', is the limit of the Brandywine (maximum) interglacial (now thought to be Upper Miocene) flooding, and is, in part, conjectural. A. Area of stream capture of the New River and the North Fork of the Roanoke River. B. Area of confluence of the headwaters of the Alabama-Coosa, Apalachicola and Savannah River systems. C. Area of suspected stream confluence between Uphauppee Creek of the Alabama-Coosa River system and Uchee Creek of the Apalachicola River system. D. Area of suspected confluence of the headwaters of the Choctawhatchee River and the Chattahoochee River. The Principal Drainage Systems 1. ALABAMA-COOSA RIVER SYSTEM 2. Coosa River Drainage 3. Tallapoosa River Drainage Apalachicolan Region 4. ESCAMBIA RIVER SYSTEM 5. YELLOW RIVER SYSTEM 6. CHOCTAWHATCHEE RIVER SYSTEM 7. APALACHICOLA RIVER SYSTEM 8. Chattahoochee River Drainage 9. Flint River Drainage 10. OCHLOCKONEE RIVER SYSTEM 11. ECONFINA RIVER SYSTEM 12. SUWANNEE RIVER SYSTEM 13. ST. MARYS RIVER SYSTEM 14. SATILLA RIVER SYSTEM Southern Atlantic Slope Region 15. ALTAMAHA RIVER SYSTEM 16. Ocmulgee River Drainage 17. Oconee River Drainage 18. OGEECHEE RIVER SYSTEM 19. SAVANNAH RIVER SYSTEM 20. EDISTO RIVER SYSTEM 21. COOPER-SANTEE RIVER SYSTEM 22. Saluda River Drainage 23. Broad River Drainage 24. Catawba River Drainage 25. BLACK RIVER SYSTEM 26. PEDEE RIVER SYSTEM 27. Yadkin River Drainage 28. WACCAMAW RIVER SYSTEM 29. CAPE FEAR RIVER SYSTEM 30. NEUSE RIVER SYSTEM 31. PAMLICO RIVER SYSTEM 32. ROANOKE RIVER SYSTEM 33. Dan River Drainage 34. Roanoke River Drainage 35. CHOWAN RIVER SYSTEM 36. JAMES RIVER SYSTEM Northern Atlantic Slope Region 37. YORK RIVER SYSTEM 38. RAPPAHANNOCK RIVER SYSTEM 39. POTOMAC RIVER SYSTEM 40. Shenandoah River Drainage Atlantic Slope Unionidae • Johnson 407 SOUTHEASTERN UNITED STATES JO 408 Bulletin Museum of Comparative Zoology, Vol 140, No. 6 PLATE 2 Pleurobema (Lexingfonia) collina (Conrad) Fig. 1. Unio collinus Conrad. North [= Maury] River, a branch of the James River [Rockbridge County], Virginia. Lec- totype ANSP 41007. Length 46 mm, height 30 mm, width 19 mm (nat. size). Fig. 2. [James River] USNM 84376. Length 18.8 mm. Specimen sprayed with ammonium chloride vapor. After Boss and Clench (1967, pi. 15, fig. 2). Pleurobema [Lexingtonia] masoni (Conrad) Fig. 3. Unio masoni Conrad. Savannah River, Augusta [Richmond County], Georgia. Holotype ANSP 41333. Length 26.6 mm, height 19.3 mm, width 13.3 mm (nat. sire). Fig. 4. Mill Race, 2 mi. N of Sardis, Burke County, Georgic, MCZ 234387. Length 47 mm, height 28 mm, width 19 mm (nat. size). Fig. 5. Unio masoni Conrad. Savannah River, Augusta [Richmond County], Georgia. Paratype ANSP 41332. Length 37 mm, height 26 mm, width 16 mm (nat. size). Fig. 6. Unio casfus Lea. South Carolina. Holotype USNM 84782. Length 40 mm, height 27 mm, width 17 mm (nat. size). Fig. 7. Unio brimleyi Wright. [Walnut Creek of] Neuse River, Raleigh [Wake County], North Carolina. Lectotype USNM 149651. Length 38 mm, height 29 mm, width 15 mm (nat. siz.e). Fig. 8. Unio merus Lea. Abbeville District [Savannah River drainage], South Carolina. Holotype USNM 85698. Length 38 mm, height 29 mm, width 15 mm (nat. size). Fig. 9. Unio pumi'/us Lea. Black River [about 10 mi. W of Benson, Johnston County], North Carolina. Holotype USNM 84545. Length 28 mm, height 20 mm, width 13 mm (nat. size). Fig. 10. Unio striatums Lea. Roanoke River, Weldon [Halifax County], North Carolina. Holotype USNM 84548. Length 39 mm, height 28 mm, width 17 mm (nat. size). Elliptio (Canthyria) spinosa (Lea) Fig. 11. Altamaha River, 4 mi. NE of Jesup, Wayne County, Georgia. MCZ 234055. Length 59 mm, height 37 mm, width 22 mm (approximately 1.2 X)- Atlantic Slope Unionidae • Johnson 409 410 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 3 Elliptio (Elliptio) crassidens crass/dens [Lamarck) Fig. 1. Unio lehmanii Wright. St. Marys River [Nassau County], Florida. Lectotype USNM 149650. Length 60 mm, height 40 mm, width 24 mm (nat. size). Fig. 2. Unio polymorphus Wright. Spanish Creek [a tributary of the St. Marys River W of Folkston], Charleton County, Georgia. Lectotype USNM 152060. Length 75 mm, height 46 mm, width 29 mm (slightly reduced). Fig. 3. Spanish Creek, Charleton County, Georgia. MCZ 269228. Length 62 mm, height 40 mm, width 30 mm (slightly reduced). Fig. 4. As above. Length 94 mm, height 57 mm, width 38 mm (slightly reduced). Atlantic Slope Unionidae • Johnson 411 412 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 4 Elliptio (Elliptio) crassidens downiei (Lea) Fig. 1. Um'o downiei Lea. Buck Lake, a bayou of the Sati 1 1 a River, Wayne [Branfly] County, Georgia. Holotype USNM 84854. Length 84 mm, height 49 mm, width 36 mm (slightly reduced). Fig. 2. Unio satillaensis Lea. Sati! la River, Camden County, Georgia. Holotype USNM 84855. Length 75 mm, height 45 mm, width 30 mm (slightly reduced). Fig. 3. Sati II a River, 3 mi. S of Hortense, Brantly County, Georgia. MCZ 237460. Length 27 mm, height 16 mm, width 8 mm (nat. size). Fig. 4. Unio spissus Lea. Satilla River, Wayne County, Georgia. Holotype USNM 84853. Length 76 mm, height 57 mm, width 40 mm (approximately nat. size). Atlantic Slope Unionidae • Jolinsoii 413 414 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 5 EUiptio (Elliptio) congaraea (Lea) Fig. 1. Unio congaraeus Lea. Congaree River [Cooper-Santee River system], South Carolina. Holotype USNM 85693. Length 41 mm, height 25 mm, width 14 mm (slightly reduced). Fig. 2. Unio (ulvus Lea. South Carolina. Holotype USNM 85679. Length 40 mm, height 21 mm, width 15 mm (slightly reduced). Fig. 3. Unio vestitus Lea. Ogeechee River [Georgia]. Holotype USNM 85332. Length 39 mm, height 25 mm, width 15 mm (nat. size). ■ Fig. 4. Unio rorbesianus Lea. Savannah River, Georgia. Holotype USNM 84542. Length 45 mm, height 31 mm, width 29 mm (nat. size). Fig. 5. Unio leconlianus Lea. Canoochee River, Liberty County, Georgia. Holotype USNM 84852. Length 70 mm, height 45 mm, width 29 mm (slig-htly reduced). ■'•'.' ■ • • ' Fig. 6. Unio dorsatus Lea. Catawba River, North Carolina. Holotype USNM 84494. Length 57 mm, height 40 mm, width 23 mm (nat. size). Fig. 7. Unio moussonianus Lea. Georgia. Holotype USNM 85168. Length 72 mm, height 49 mm, width 28 mm (slightly reduced). Fig. 8. Unio corvus Lea. Buckhead Creek, Burke County, Georgia. Holotype USNM 84539. Length 62 mm, height 40 mm, width 25 mm (nat. size). Atlantic Slope Unionidae • Johnson 415 416 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 6 Elliptic [Elliptio] dariensis (Lea) Fig. 1. House Creek, Bowens Mills, 9 mi. N of Fitzgerald, Ben Hill County, Georgia. MCZ 234237. Length 41 mm, height 23 mm, width 14 mm (nat. size). Fig. 2. Unio dariensis Lea. [Altamaha River] near Darien [Mcintosh County], Georgia. Holotype USNM 85691. Length 78 mm, height 55 mm, width 28 mm (slightly reduced). Fig. 3. House Creek, Bowens Mill, 9 mi. N of Fitzgerald, Ben Hill County, Georgia. MCZ 234237. Length 76 mm, height 45 mm, width 26 mm (nat. size). Fig. 4. As above. Length 98 mm, height 58 mm, width 35 mm (nat. size). Atlantic Slope Unionidae • Johnson 417 '- ■*% 418 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 7 Elliptio [Elliptio) fraterna (Lea) Fig. 1. Un/o traternus Lea. Abbeville District [Savannah River drainage], South Carolina. Holotype USNM 85396. Length 60 mm, height 31 mm, width 16 mm (slightly reduced). Fig. 2. Unio Iraternus Lea. [Chattahoochee River] Columbus [Muscogee County], Georgia. Pararype USNM 85398. Length 69 mm, height 34 mm, width 17 mm (slightly reduced). Fig. 3. Unio anthonyi Lea. Florida. Holotype USNM 84986. Length 58 mm, height 35 mm, width 22 mm (slightly reduced). Fig. 4. Flat Creek, 8 mi. SW of Samson, Geneva County, Alabama. MCZ 186967. Length 75 mm, height 38 mm, width 20 mm. Fig. 5. Elliptio mcmichaeli Clench and Turner. Choctawhatchee River, 8 mi. W of Miller Cross Roads, Holmes County, Florida. Holotype MCZ 191922. Length 91 mm, height 50 mm, width 27 mm. Elliptio waccamawens/s (Lea) Fig. 6. Unio waccamawens/s Lea. Lake Waccamaw, North Carolina. Holotype USNM 84437. Length 36 mm, height 17 mm, width 16 mm (nat. size). Fig. 7. Canal beside Lake Waccamaw, 1 mi. NNW of Dupree Landing, Columbus County, North Carolina. MCZ 214235. Length 89 mm, height 45 mm, width 28 mm (approximately nat. size). Atlantic Slope Unionidae • Johnson 419 420 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 8 Elliptio (Elliptio) complanata (Lightfoot) Fig. 1. Savannah River, 7 mi. NE of Newington, Screven County, Georgia. MCZ 269257. Length 99 mm, height 50 mm, width 28 mm (slightly enlarged). Fig. 2. Turkey Creek, 4 mi. NE of Allentown [Altamaha River system], Wilkinson County, Georgia. MCZ 234355. Length 73 mm, height 36 mm, width 18 mm (nat. size). Fig. 3. Unio quadrilaterus Lea. Neuse River, near [6 mi. E of] Raleigh, [Wake County], North Carolina. Holotype USNM 85385. Length 60 mm, height 36 mm, width 20 mm (slightly reduced). Fig. 4. Unio roanokensis Lea. Roanoke River [road between Norfolk, Virginia, and Tarborough, North Carolina]. Holo- type USNM 85423. Length 119 mm, height 57 mm, width 29 mm (slightly reduced). Fig. 5. Unio mediocris Lea. Neuse River, [6 mi. E of] Raleigh [Wake County], North Carolina. Holotype USNM 85611. Length 57 mm, height 30 mm, width 19 mm (slightly reduced). Fig. 6. North shore, Lake Waccamaw, town of Lake Waccamaw, Columbus County, North Carolina. MCZ 258813. Length 60 mm, height 31 mm, width 18 mm (approximately nat. size). Atlantic Slope Unionidae • Johnson 421 422 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 9 Elliptio (Elliptic*) hopetonensis (Lea) Fig. 1. Unio inusitatis Lea. Swift Creek, below Macon [Bibb County], Georgia. Holotype USNM 85531. Length 57 mm, height 32 mm, width 19 mm (slightly reduced). Fig. 2. Unio hopetonensis Lea. [Altamaha River], Hopeton, near Darien [Mcintosh County], Georgia. Holotype USNM 85391. Length 84 mm, height 43 mm, width 20 mm (slightly reduced). Elliptio [Elliptio] icjerina (Conrad) Fig. 3. Unio icterinus Conrad. Savannah River, opposite Augusta [Richmond County], Georgia. Holotype ANSP 41381. Length 55 mm, height 28 mm, width 18 mm (approximately nat. size). Fig. 4. Unio lugubris Lea. [Altamaha River], Hopeton, near Darien [Mcintosh County], Georgia. Holotype USNM 85638. Length 53 mm, height 37 mm, width 22 mm (slightly enlarged). Fig. 5. Unio fuomeyi Lea. Abbeville District [Savannah River drainage]. South Carolina. Holotype USNM 85669. Length 65 mm, height 33 mm, width 19 mm (slightly reduced). Fig. 6. Unio opacus Lea. Buckhead Creek, Burke County, Georgia. Holotype USNM 85546. Length 66 mm, height 37 mm, width 24 mm (slightly reduced). Fig. 7. Unio hastatus Lea. New Market, Abbeville District [Savannah River drainage], South Carolina. Holotype USNM 86013. Length 66 mm, height 28 mm, width 17 mm (slightly reduced). Fig. 8. Unio obnubilus Lea. Buckhead Creek, Burke County, Georgia. Holotype USNM 85646. Length 65 mm, height 38 mm, width 23 mm (slightly reduced). Fig. 9. Unio similis Lea. Buckhead Creek, Burke County, Georgia. Holotype USNM 85653. Length 57 mm, height 31 mm, width 20 mm (slightly reduced). Fig. 10. Unio aequotus Lea. Buckhead Creek, Burke County, Georgia. Holotype USNM 85561. Length 66 mm, height 38 mm, width 21 mm (slightly reduced). Atlantic Slope Unionidae • Johnson 423 424 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 10 EUiptio [Elliptio] icterina (Conrad) Fig. 1. Un/'o micans Lea. Catawba River, Gaston County, North Carolina. Holotype USNM 85077. Length 44 mm, height 23 mm, width 14 mm (slightly reduced). Fig. 2. Un/'o ocmu/geens/'s Lea. Little Ocmulgee River, Lumber City, [Telfair County], Georgia. Holotype USNM 85901. Length 104 mm, height 50 mm, width 35 mm (slightly reduced). Fig. 3. Un/'o confertus Lea. Santee Canal, South Carolina, from J. Lewis. Metatype USNM 85634. Length 60 mm, height 33 mm, width 22 mm (slightly reduced). The holotype was not located. This specimen, identified by Lea, is almost the same width and height as the original specimen, though the latter was more inflated, measuring 28 mm in width. Elliptio [EUiptio] arctata (Conrad) Fig. 4. Un/'o arcfatus Conrad. Alabama River, Alabama. Lectotype ANSP 41356. Length 55 mm, height 22 mm, width 14 mm (nat. size). Fig. 5. Un/'o sfrigosus Lea. Chattahoochee River, Columbus [Muscogee County], Georgia. Holotype USNM 85890. Length 56 mm, width 24 mm, height 13 mm (slightly reduced). Fig. 6. Un/'o lazarus Lea. Abbeville District [Savannah River drainage], South Carolina. Holotype USNM 86155. Length 52 mm, height 23 mm, width 12 mm (slightly reduced). Fig. 7. Un/'o perstriatus Lea. Abbeville District [Savannah River drainage], South Carolina. Holotype USNM 85892. Length 57 mm, height 25 mm, width 12 mm (slightly reduced). Fig. 8. Un/'o graci/entus Lea. Catawba River, Gaston County, North Carolina. Holotype USNM 85976. Length 71 mm, height 30 mm, width 13 mm (slightly reduced). Fig. 9. Un/'o perlatus Lea. Cape Fear River, Black Rock Landing [= 2 mi. S of Kings Bluff, Bladen County], North Caro- lina. Holotype USNM 86006. Length 66 mm, height 22 mm, width 12 mm (slightly reduced). EUiptio [EUiptio] lanceolata (Lea) Fig. 10. Un/'o productus Conrad. Savannah River, Augusta [Richmond County, Georgia]. Holotype ANSP 41397. Length 66 mm, height 26 mm, width 16.5 mm (approximately nat. size). Atlantic Slope Unionidae • Johnson 425 426 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE n Elliptio [Elliptio] lanceolata (Lea) Fig. 1. Unio subcylindraceus Lea. Rocky Creek, near Macon [Bibb County], Georgia. Holotype USNM 85863. Length 95 mm, height 41 mm, width 36 mm (slightly reduced). Fig. 2. Unio lanceolatus Lea. Tar River, Tarborough [Edgecombe County], North Carolina. Holotype USNM 85905. Length 41 mm, height 18 mm, width 11 mm (slightly reduced). Fig. 3. Unio angustatus Lea. Cooper River, South Carolina. Holotype USNM 85896. Length 72 mm, height 29 mm, width 17 mm (slightly reduced). Fig. 4. Un/o emmonsii Lea. Roanoke River, Weldon [Halifax County], North Carolina. Holotype USNM 86028. Length 110 mm, height 46 mm, width 26 mm (slightly reduced). Fig. 5. Unio hazelhurstianus Lea. Satil I a River, Camden County, Georgia. Holotype USNM 86009. Length 78 mm, height 35 mm, width 22 mm (slightly reduced). Fig. 6. Unio fisherianus Lea. Head of Chester River [Kent County], Maryland. Holotype USNM 86002. Length 70 mm, height 29 mm, width 15 mm (slightly reduced). Elliptio [Elliptio] shepardiana (Lea) Fig. 7. Ocmulgee River, 1 mi. S of Lumber City, Telfair County, Georgia. MCZ 234042. Length 127 mm, height 29 mm, width 15 mm (slightly enlarged). Atlantic Slope Unionidae • Johnson 427 428 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 12 Uniomerus tetralasmus (Say) Fig. 1. Unio obesus Lea. Little Ogeechee River [Hancock County], Georgia. Holotype USNM 85366. Length 82 mm, height 50 mm, width 35 mm (slightly reduced). Fig. 2. Unio bisse/zanus Lea. Bissels Pond, Charlotte [Mecklenburg County], North Carolina. Holotype USNM 85373. Length 69 mm, height 40 mm, width 24 mm (slightly reduced). Fig. 3. Unio squalidus Lea. Roanoke River, near Weldon [Halifax County], North Carolina. Holotype USNM 85376. Length 52 mm, height 30 mm, width 18 mm (slightly reduced). Fig. 4. Unio cicur Lea. Little Ocmulgee River, Georgia. Holotype USNM 85532. Length 41 mm, height 25 mm, width 17 mm (slightly reduced). Fig. 5. Unio ineptus Lea. Abbeville District [Savannah River drainage], South Carolina. Holotype USNM 85326. Length 33 mm, height 18 mm, width 10 mm (slightly reduced). Fig. 6. Unio pawensis Lea. Beaver Creek [into ?], Catawba Run [Gaston and Mecklenburg Counties], North Carolina. Holotype USNM 85380. Length 74 mm, height 42 mm, width 29 mm (slightly reduced). Lasmigona (P/afyna/as) subviridis (Conrad) Fig. 7. Unio pertenuis Lea. Neuse River, near [6 mi. E of] Raleigh [Wake County], North Carolina. Holotype USNM 86139. Length 35 mm, height 22 mm, width 13 mm (slightly reduced). Fig. 8. Unio hyalinus Lea. [James River drainage], Richmond [Henrico County], Virginia. Holotype USNM 86131 Length 35 mm, height 21 mm, width 12 mm (slightly reduced). Fig. 9. Unio charlottensis Lea. [Sugar Creek], near Charlotte, Mecklenburg County, North Carolina. Holotype USNM 85402. Length 101 mm, height 58 mm, width 34 mm (slightly reduced). Atlantic Slope Unionidae • Johnson 429 430 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 13 Lasmigona (Platynaias) subviridis (Conrad) Fig. 1. Un/o decorafus Lea. Abbeville District [Savannah River drainage], South Carolina. Holotype USNM 83972. Length 60 mm, height 37 mm, width 19 mm (slightly reduced). Alasmidonta [Prolasmidonta] heterodon (Lea) Fig. 2. Neuse River [6 mi. E of] Raleigh [Wake County], North Carolina. MCZ 231191. Length 34 mm, height 19 mm, width 14 mm (1.5 X). Alasmidonta {Alasmidonta) undulata (Say) Fig. 3. Nottoway River, 3 mi. E of Rowlings, Brunswick County, Virginia. MCZ 237455. Length 54 mm, height 32 mm, width 23 mm (approximately nat. size). Fig. 4. James River, opposite Maidens, Goochland County, Virginia. MCZ 261313. Length 29 mm, height 19 mm, width 13 mm (1.5 X)- Alasmidonta (Alasmidonta) triangulata (Lea) Fig. 5. Margan'fana triangulata Lea. Upper Chattahoochee [River], Georgia. Holotype USNM 86249. Length 57 mm, height 38 mm, width 31 mm (nat. size). Fig. 6. Mill Race, 2 mi. N of Sardis, Burke County, Georgia. MCZ 237453. Length 48 mm, height 32 mm, width 27 mm (slightly enlarged). Alasmidonta (Alasmidonta) arcu/a (Lea) Fig. 7. Margaritana areola Lea. Altamaha [River], Liberty [now Long] County, Georgia. Holotype USNM 86170. Length 57 mm, height 44 mm, width 41 mm (nat. size). Fig. 8. Altamaha River, 4 mi. NE of Jesup, Wayne County, Georgia. MCZ 237551. Length 50 mm, height 43 mm, width 38 mm (slightly enlarged). Atlantic Slope Unionidae • Johnson 431 432 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 14 Alasmidonta (Decurambis) varicosa (Lamarck) Fig. 1. Turkey Creek, 8 mi. NW of Edgefield, Edgefield County, South Carolina. MZUM 58024. Length 44 mm, height 26 mm, width 18 mm (slightly reduced). Fig. 2. North Fork of Shenandoah River, E of Woodstock, Shenandoah County, Virginia. MCZ 216721. Length 45 mm, height 25 mm, width 18 mm (2X)- Anodonfa (Pyganodon) cataracta cataracta Say Fig. 3. Anodonta doliaris Lea. Stewarts Mill Dam, Union County, North Carolina. Topotype MZUM 103859, from the original lot collected by C. M. Wheatley. Lea saw only the holotype in the ANSP 126522a. Length 111 mm, height 67 mm, width 50 mm (slightly reduced). Fig. 4. Anodonfa dariensis Lea. Swift Creek, near Macon [Bibb County], Georgia. Holotype USNM 86600. Length 107 mm, height 58 mm, width 49 mm (slightly reduced). Atlantic Slope Uxionidae • Johnson 433 434 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 15 Anodonta [Pyganodon] cataracta cataracta Say Fig. 1. Anodonta virgulata Lea. Roanoke River, Weldon [Halifax County], North Carolina. Holotype USNM 86593. Length 78 mm, height 44 mm, width 33 mm. Anodonta [Pyganodon] gibbosa Say Fig. 2. Altamaha River, "Riverside Park," 4 mi. N of Jesup, Wayne County, Georgia. MCZ 234056. Length 40 mm, height 30 mm, width 22 mm (l3/s X)- Fig. 3. Cedar Creek, Fountains Mill, 7 mi. SW of Hawkinsville, Pulaski County, Georgia. MCZ 111438. Length 107 mm, height 61 mm, width 48 mm (slightly reduced). Anodonta (Pyganodon) implicata Say Fig. 4. Anodonta newtonensis Lea. Schuylkill [River], Fairmount [Philadelphia, Philadelphia County, Pennsylvania]. Lectotype USNM 86561. Length 97 mm, height 52 mm, width 39 mm (slightly reduced). Atlantic Slope Unionidae • Johnson 435 436 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 16 Anodonta (Pyganodon) implicata Say Fig. 1. Anodonta housalonica ' Linsley ' Gould. Housatonic [River], Corum [= near Huntington, Fairfield County], Con- necticut. Holotype USNM 678302. Length 94 mm, height 51 mm, width 34 mm (slightly reduced). Fig. 2. Anodonta implicata Say. Agawam River (outlet of Halfway Pond), Plymouth [Plymouth County], Massachusetts. Neotype MCZ 176769. Length 120 mm, height 60 mm, width 48 mm (somewhat reduced). Interior view, showing the especially distinct pallial line caused by the characteristic anterior-ventral thickening of the shell. Anodonta (Utterbackia) imbecilis Say Fig. 3. Magnolia Springs, Perkins, Jenkins County, Georgia. MCZ 234050. Length 90 mm, height 42 mm, width 34 mm (nat. size). Anodonta [Utterbackia) couperiana Lea Fig. 4. Pond near Savannah, Chatham County, Georgia. MCZ 119137. Length 87 mm, height 50 mm, width 35 mm (nat. size). Atlantic Slope Unionidae • Johnson 437 438 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 17 Anodonta [Utterbackia] couperiana Lea Fig. 1. Anodonta dunlapiana Lea. [Charleston, Chatham County], South Carolina. Lectotype USNM 86564. Length 83 mm, height 46 mm, width 35 mm (slightly reduced). Sfroph/fus undulatus (Say) Fig. 2. Turkey Creek, 8 mi. NW of Edgefield, Edgefield County, South Carolina. MCZ 82807. Length 47 mm, height 29 mm, width 17 mm (nat. size). Fig. 3. As above. Length 74 mm, height 44 mm, width 27 mm (nat. size). Carunculina pulla (Conrad) Fig. 4. Unio pullus Conrad. Wateree River, South Carolina. Type lost. (Figures after Conrad). Length 33 mm, height 19 mm, width 13.5 mm (nat. size). Fig. 5. Caruncu/ina patrickae Bates. Savannah River, Johnsons Landing, 10 mi. W of Allendale, Allendale County South Carolina. Topotype MCZ 255220. Length 32 mm, height 19 mm, width 15.5 mm. Female (1.25 X)- This topotype lacks the sharp posterior ridge of the holotype. Fig. 6. As above. Length 22 mm, height 13.5 mm, width 10 mm. Male (1.25 X)- This topotype closely resembles Bates' sketch of the allotype. Fig. 7. University Lake, an impoundment on Morgan Creek, 1 mi. W of Chapel Hill, Orange County, North Carolina. MCZ 261347. Length 25 mm, height 17 mm, width 11 mm. Male (1.25 X). Villosa villosa (Wright) Fig. 8. St. Marys River, Nassau County, Florida. USNM 152066. Length 56 mm, height 29 mm, width 20 mm. Male (slightly reduced). Fig. 9. Unio villosus Wright. Suwannee River [Luraville], Suwannee County, Florida. Lectotype USNM 150503. Length 57 mm, height 28 mm, width 18 mm. Female (slightly reduced). V///osa vibex (Conrad) Fig. 10. Unio vibex Conrad. Black Warrior River, S of Blount s Spring [Blount County], Alabama. Holotype ANSP 56488a. Length 49 mm, height 28 mm, width 17 mm. Male (approximately nat. size). Fig. 11. As above. Allotype ANSP 56488. Length 54 mm, height 31 mm, width 19 mm. Female (approximately nat. size). Atlantic Slope Unionidae • Johnson 439 - / 10 440 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 18 Villosa vibex (Conrad) Fig. 1. Um'o gracilior Lea. Buckhead Creek [Burke County]; or Tobesaufke [Tobesofkee] Creek, near Macon [Bibb County],- both Georgia. Holotype USNM 85088 [exact locality not indicated]. Length 50 mm, height 28 mm, width 18 mm. Male (slightly reduced). Fig. 2. Um'o modioliformis Lea. Santee Canal, South Carolina. Probable holotype USNM 85029. Length 68 mm, height 40 mm, width 25 mm. Female (slightly reduced). Fig. 3. Mill Race, 2 mi. N of Sardis, Burke County, Georgia. MCZ 234263. Length 48 mm, height 27 mm, width 17 mm. Male (nat. size). Villosa delumbis (Conrad) Fig. 4. Un/'o concavus Lea. Abbeville District [Savannah River drainage], South Carolina. Holotype USNM 85154. Length 55 mm, height 32 mm, width 21 mm. Male (slightly reduced). Fig. 5. Um'o genui'nus Lea. Bissels Pond, Charlotte [Mecklenburg County], North Carolina. Holotype USNM 85123. Length 52 mm, height 30 mm, width 20 mm. Male (slightly reduced). Fig. 6. Mill Race, 2 mi. N of Sardis, Burke County, Georgia. MCZ 234340. Length 53 mm, height 30 mm, width 19 mm. Male (nat. size). Fig. 7. As above. Length 57 mm, height 36 mm, width 19 mm. Female (nat. size). Fig. 8. Um'o ogeecheensi's Conrad. Ogeechee River, Georgia. Lectotype MCZ 146971. Length 64 mm, height 39 mm, width 30 mm. Female (approximately nat. size). Villosa constricta (Conrad) Fig. 9. Um'o lienosus constrictus Conrad. North [:= Maury] River, Rockbridge County, Virginia. Lectotype ANSP 56465a. Length 41.4 mm, height 27 mm, width 16 mm. Female (approximately nat. size). Fig. 10. Um'o genthii Lea. [Probably from] Deep River, Gulf [Chatham County], North Carolina. Holotype USNM 84834. Length 39 mm, height 27 mm, width 18 mm. Male (nat. size). Atlantic Slope Uxionidae • Johnson 441 442 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 19 Ligumia nasuta (Say) Fig. 1. Potomac River, Washington, D.C. MCZ 119087. Length 62 mm, height 25 mm, width 12 mm. Male. Fig. 2. As above. MCZ 5653. Length 66 mm, height 29 mm, width 14 mm. Female. Lampsilis [Lampsilis) can'osa (Say) Fig. 3. Savannah River, 7 mi. NE of Newington, Screven County, Georgia. MCZ 234241. Length 98 mm, height 62 mm, width 42 mm. Male (slightly reduced). Fig. 4. As above. Length 98 mm, height 69 mm, width 45 mm. Female (slightly reduced). Fig. 5. Un/'o crocatus Lea. Savannah River, Georgia. Holotype USNM 84908. Length 42 mm, height 27 mm, width 19 mm. Male (slightly reduced). Atlantic Slope Unionidae • Johnson 443 444 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 20 Lampsilis [Lampsilis) dolabraelormh (Lea) Fig. 1. Altamaha River, 11 mi. N of Odum, Wayne County, Georgia. MCZ 234197. Length 16 mm, height 12 mm, width 7.5 mm. Male (4X1- Fig. 2. Altamaha River, 10 mi. N of Baxley, Appling County, Georgia. MCZ 234010. Length 91 mm, height 69 mm, width 49 mm. Female (slightly reduced). Fig. 3. As above. Length 79 mm, height 63 mm, width 50 mm. Male (slightly reduced). Fig. 4. Unio dolabraeformh Lea. Altamaha River, Liberty [now Long] County, Georgia. Holotype USNM 84888. Length 116 mm, height 79 mm, width 51 mm. Male (reduced). Atlantic Slope Unioxidae • Johnson 445 446 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 21 Lampsilis- [Lampsilis] ovata (Say) Fig. 1. Lampsilis ventricosa cohongoronfa Ortmann. Potomac River, Hancock, Washington County, Maryland. Lecto- type Carnegie Museum 61.3999. Length 95 mm, height 60 mm, width 41 mm. Male. Fig. 2. As above. Allotype Carnegie Museum 61.4000. Length 87 mm, height 62 mm, width 42 mm. Female. Lampsilis [Lampsilis] ochracea (Say) Fig. 3. Un/'o rosaceus Conrad. Savannah River, Georgia. Holotype MCZ 178779. Length 73 mm, height 44 mm, width 27 mm. Male (approximately nat. size). Fig. 4. Lake Waccamaw, [town of] Lake Waccamaw, Columbus County, North Carolina. MCZ 234327. Length 40 mm, height 30 mm, width 17 mm. Female (approximately nat. size). Fig. 5. As above. Length 50 mm, height 35 mm, width 21 mm. Male (approximately nat. size). Atlantic Slope Unionidae • Johnson 447 448 Bulletin Museum of Comparative Zoology, Vol. 140, No. 6 PLATE 22 Lampsilis (Lampsilis) radiata radiata (Gmelin) Fig. 1. Unio conspicuus Lea. Yadkin River, Salisbury [Rowan County], North Carolina. Holotype USNM 85056. Length 112 mm, height 64 mm, width 39 mm. Male (slightly reduced). Fig. 2. Greenfield Pond, Wilmington [New Hanover County], North Carolina. USNM 523976. Length 70 mm, height 43 mm, width 28 mm. Male (slightly reduced). Fig. 3. As above. USNM 452041. Length 76 mm, height 40 mm, width 28 mm. Female (slightly reduced). Lampsilis (Lampsilis) splendida (Lea) Fig. 4. Ocmulgee River, 1 mi. S of Lumber City, Telfair County, Georgia. MCZ 234019. Length 71 mm, height 47 mm, width 35 mm. Female (approximately 1.2 X). Fig. 5. Unio splendidus Lea. Alfamaha River [Mcintosh or Long County], Georgia. Holotype USNM 84893. Length 70 mm, height 43 mm, width 36 mm. Male (slightly reduced). Atlantic Slope Unionidae • Johnson 449 Bulletin of the Museum of Comparative Zoology ■:'^Mm^mM Ecological-Behavioral Studies of the Wasps of Jackson Hole, Wyoming HOWARD E. EVANS HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS, U.S.A. VOLUME 140, NUMBER 7 27 NOVEMBER 1970 PUBLICATIONS ISSUED OR DISTRIBUTED BY THE MUSEUM OF COMPARATIVE ZOOLOGY HARVARD UNIVERSITY Bulletin 1863- Breviora 1952- Memoirs 1864-1938 Johnsonia, Department of Mollusks, 1941- Occasional Papers on Mollusks, 1945- Other Publications. Bigelow, H. B., and W. C. Schroeder, 1953. Fishes of the Gulf of Maine. ' Reprint, $6.50 cloth. Brues, C. T., A. L. Melander, and F. M. Carpenter, 1954. Classification of Insects. $9.00 cloth. Creighton, W. S., 1950. The Ants of North America. Reprint, $10.00 cloth. Lyman, C. P., and A. R. Dawe (eds.), 1960. Symposium on Natural Mam- malian Hibernation. $3.00 paper, $4.50 cloth. Peters' Check-list of Birds of the World, vols. 2-7, 9, 10, 12-15. ( Price list on request. ) Turner, R. D., 1966. A Survey and Illustrated Catalogue of the Teredinidae (Mollusca: Bivalvia). $8.00 cloth. Whittington, H. B., and W. D. I. Rolfe (eds.), 1963. Phylogeny and Evolution of Crustacea. $6.75 cloth. Proceedings of the New England Zoological Club 1899-1948. ( Complete sets only. ) Publications of the Boston Society of Natural History. Publications Office Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138, U. S. A. © The President and Fellows of Harvard College 1970. ECOLOGICAL-BEHAVIORAL STUDIES OF THE WASPS OF JACKSON HOLE, WYOMING HOWARD E. EVANS ABSTRACT During three summers of study, 190 species of wasps ( Hymenoptera, Aeuleata ) were collected in Jackson Hole, Wyoming, principally in six selected sites along the Snake River. The majority of these were fossorial species, about 50 of which were found nesting. For the most part, the species studied appeared to avoid complete competition for prey and for nesting sites, but some possible exceptions are noted. Many species were attacked by parasitic flies, some by cuckoo wasps and other insects. Many of these natural enemies are not host-specific, and maintain high populations at the expense of many different species of wasps. The behavior patterns of the wasps are discussed as mechanisms for ( 1 ) avoiding competition with other wasp species, and (2) reducing the success of parasites. Following a general discussion of these subjects, a list of the wasps of Jackson Hole, annotated with ecological and behavioral data, is presented. CONTENTS Introduction 451 Study Areas 453 Major Aspects of Ecology 455 Seasonal cycles 455 Nesting sites and nest type 455 Food relationships 457 Ecological displacement in the species of Ph ilan th us 46 1 Kinds and specificity of parasites and pred- ators - 463 Summary of interrelationships 466 Behavior Patterns as Adaptations to Life in the Community 468 Behavior related to competition 468 Behavior related to parasitism 469 Summary of behavioral adaptations 472 Bull. Mus. Comp. Annotated List of the Wasps of Jackson Hole 474 Family Dryinidae 475 Family Chrysididae 475 Family Sierolomorphidae 477 Family Tiphiidae 477 Family Mutillidae 477 Family Sapygidae 477 Family Eumenidae 477 Family Masaridae 479 Family Vespidae 479 Family Pompilidae 480 Family Sphecidae 483 Subfamily Sphecinae 483 Subfamily Pemphredoninae . 486 Subfamily Astatinae 487 Subfamily Larrinae 488 Subfamily Crabroninae 491 Subfamily Nyssoninae 493 Subfamily Philanthinae 496 References Cited 502 INTRODUCTION It was unfortunate that during the flowering of natural history in the 19th century it became fashionable to separate ecology (from oikos, home) from ethology (from ethos, habits). Surely the habitat and behavior of an animal are all of a piece. Behavior is merely the impingement of an animal upon its environment, while the environment is that portion of the bio- sphere in which an animal's behavior per- mits it to thrive. A biotic community is a collection of co-adapted behaviors, no one Zool, 140(7): 451-511, November, 1970 451 452 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 .■<& JACKSON LAKE RES. STA. MOR Bl B2 B3 MOR,AI AZ' SIGNAL MT. 1 y OUTWASH PLAIN SNAKE RIVER N A ELK ELK P.O. ALLUVIAL BOTTOMS 2 mi. Map 1. Sketch of portion of Jackson Hole, Wyoming, showing location of major study areas. Wasps of Jackson Hole • Evans 453 of which is fully understandable of and by of view of a specialist on digger wasps, the itself. That such communities are often area is especially attractive because much enormously complex is troublesome, but of it is unsuitable for these insects; but this should not divert us from assuming, at those areas which are suitable (i.e., where least occasionally, a broad view of the the soil is friable and more or less devoid interactions of its components. Only such of vegetation) contain concentrations of a view can provide insights into the species and individuals perhaps unrivalled adaptive value and biological significance anywhere. Another attractive feature is of observed behaviors; and only behavioral that much of Jackson Hole is part of Grand studies can supply insights into the signifi- Teton National Park and thus immune to cance of the structures with which system- development or to extensive disturbance, atists are so preoccupied. Still another is the presence of the Jackson Most areas of bare, friable soil in tern- Hole Research Station at Moran, which perate and tropical regions are inhabited provided a pleasant and effective base of by numerous species of digger wasps which operations during three summers of study often show marked behavioral differences (1961, 1964, and 1967). I am much in- with respect to nesting and hunting be- debted to the officers of the station, par- havior. These behavior patterns are stereo- ticularly to its director, Dr. L. Floyd typed and relatively easy to observe and Clarke, for facilitating this research in to describe. Students have often concluded many ways. I am also indebted to the that the behavioral peculiarities of the authorities of Grand Teton National Park various species represent adaptations for for permission to collect specimens for avoiding competition with other members identification, of the community, or for reducing the incidence of attacks by the great numbers ^'UUY AKbAb of parasites which occur in these restricted Most of the floor of Jackson Hole is habitats. There have, however, been few classified as glacial outwash plain ( Fryxell, efforts to consider an entire community in 1930). East of Jackson Lake much loess the attempt to document these impressions, has been deposited. This loess not only The present study, although grossly in- renders the glacial outwash relatively more complete, represents an effort in that di- friable than elsewhere, but also is the rection. source of much of the alluvial sand which Jackson Hole, Wyoming, is in many ways has been deposited here and there on the ideally suited for studies of this nature, immediate banks of the Snake River. These This hat valley surrounded by mountains studies were conducted partly in areas of is relatively high (6750 feet elevation at "lower outwash plain" and partly in "allu- Moran) and far enough north (about 44° vial bottoms," following Fryxell's termin- north latitude) so that its insect fauna is ology (see Map 1). The outwash plains more limited in species than that of many are of Pleistocene origin (Wisconsin glacial more southerly localities. Yet during the stage) and consist chiefly of quartzite brief summer season insects are exceed- gravel with a variable content of loess and ingly abundant and much more diverse a great many cobbles of various sizes; for than one might expect, including species the most part they are covered with sage- of eastern, Pacific coast, and northern dis- brush (Artemisia). The alluvial sand is of tribution, as well as a few Upper Sonoran recent origin; it is pale in color and of elements that apparently follow the Snake relatively uniform texture. It occurs in low River drainage, and a few characteristic terraces or in bars which often slope into Rocky Mountain elements. From the point the river, these bars and terraces often 454 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 being surrounded by willows and cotton- woods. Groves of lodgepole pine and aspen occur here and there in the outwash plain or adjacent to the alluvial deposits. Where the outwash gravels abut upon the river, the banks are usually high and are con- stantly being eroded, while the alluvial deposits are several feet lower and in areas of accretion during flooding. In some cases there is a bank separating the margin of the plain from the alluvial bottoms, but more often there is a gradual slope between the two with no sharply marked separation of soil type or cover. Although I collected wasps at several places in Jackson Hole in an effort to understand the distribution of various spe- cies, most of my studies were conducted in three sites near the Research Station. These localities were chosen because of the high concentration of nesting wasps they con- tained and because each was slightly dif- ferent as to soil type, vegetation, and composition of the fauna. The three sites are considered below, followed by a de- scription of three additional areas in which extensive collecting was done. In each case I have noted the abbreviation by which these sites are indicated in the text. The three primary localities are shown on Map 1. (1) Moran, area A (MOR-A). Alluvial sand close beside the Snake River, both near the Station (Al) and at the "Cattle Bridge," about 0.5 miles east of the Station (A2). These are areas of fine-grained, light sand, largely bare but in places partially overgrown with short grass or low herbs. The sand flats are relatively sharply separated from adjacent outwash plain and have a rich and distinctive fauna of digger wasps, including especially species of Bemhix, Oxybelus, AmmopJiila, and Tachy- sphex. A2 is shown in more detail in Map 2, and a photograph of a portion of the area is shown in Figure 1. (2) Moran, area B (MOR-B). Outwash plain between the Station and the Cattle Bridge, especially three areas with very sparse vegetation (Bl, 2, 3) which are inhabited by numerous species of wasps including Philanthus pulcher, P. crabroni- formis, Eucerceris spp., and others (Fig. 2). Certain wasps occupied both the more friable portions of these areas and the more compact soil in the alluvial bars and ter- races: such species as Hoplisoides spilo- graphus and Stenodynerus papogorum. Wasps not largely restricted to either areas A or B are simply labeled "MOR" in the text. Since wasps nesting in A or B gen- erally ranged widely for their prey, the two communities were by no means separate entities. Also, ground squirrels nesting in area B often produced mounds of loosened, friable soil which were oc- casionally occupied by typical inhabitants of alluvial sand such as Episyron quin- quenotatus. ( 3 ) Four miles S W of the Elk post office (ELK), roughly seven miles due south of MOR. This area is at a very slightly lower altitude (6650 feet) and is one in which the rather extensive alluvial sand grades almost imperceptibly into glacial outwash, the two soil types intergrading over a low terrace of several acres ( Fig. 3 ) . The rich wasp fauna here is dominated by Bembix americana spinolae and Philanthus zebrotus nit ens. (4) Five miles north of the city of Jack- son, near the junction of the Gros Ventre and Snake Rivers (JAC). This is at a still lower altitude (6300 feet) and represents an area of extensive alluvial deposits of cobblestones with patches of sand. (5) Pilgrim Creek (PCR), in Grand Teton National Park and Teton National Forest, some four to seven miles NE of Moran. This is a small stream cutting through forested country, but having nu- merous sandy deposits along its banks. The sand-inhabiting wasps are similar to those in MOR-A; in addition, there are many twig-nesting wasps here, as is generally true in more wooded areas. (6) Huckleberry Hot Springs (HHS), in Teton National Forest just north of Grand Wasps of Jackson Hole • Evans 455 Teton National Park and south of Yellow- stone. This area is of special interest be- cause of its hot springs, which favor cer- tain species of generally more southerly distribution (e.g., Bembix amoena). Studies were conducted in a sandy field adjacent to the major springs. MAJOR ASPECTS OF ECOLOGY During the period of study, 190 species of aculeate wasps were collected in these six areas (six other species, listed below in brackets, in peripheral localities). Some species were taken only once or twice and appear to maintain low populations in Jackson Hole, while others were exeeed- ! ingly abundant in suitable sites. The majority of these wasps (slightly over 100 species) are solitary ground-nesters or are parasites of such wasps. How do so many species of generally similar behavior man- ! age to survive side-by-side during the same brief active season, especially when the ' presence of so many fossorial wasps per- mits the build-up of great numbers of i parasites? To what extent are the observed behavioral differences among these species understandable in the contexts of inter- species competition and of rampant para- ' sitization? An attempt will be made to answer these questions in this and the next j section, leaving the actual documentation to an annotated list of wasps to follow. I was able to study the nesting and j predatory behavior of only about half of the ground-nesters (about 50 species in all), but these were by and large the com- moner fossorial species. The majority of the other species have either been studied elsewhere or are closely related to species of known behavior. While most of the conclusions I shall draw here are based on those species studied in detail in Jackson Hole, there is basis for extrapolating them to the remainder of the wasp fauna. Seasonal Cycles Since adult female solitary wasps gen- erallv live from three to six weeks, it is obvious that to thrive in Jackson Hole they must emerge some time in July in order to have time to complete their nesting cycles before cold temperatures restrict activity in late August or early September. Hence, most species nest more or less simultane- ously. Nevertheless some differences in time of emergence and completion of nest- ing can be noted. For example, Philanihus pulcher emerges in late June or early July and has largely disappeared by August first, while P. pacificus is a species char- acteristic of August. Among the spider wasps, Episyron quinquenotatus is already active in early July, while Ageniella blais- delVi does not appear until about August 1. However, differences in seasonal cycle apparently functioning to reduce competi- tion among species are minimal in this area. Nesting Sites and Nest Type Most of the fossorial wasps studied nested in one particular type of soil and no other. Bembix americana spinolae, for example, is a gregarious species requiring a fairly large expanse of fine-grained, friable sand; populations of this species occurred only at MOR-A2, ELK, and HHS. The related wasp Steniolia obliqua, in con- trast, occurs in smaller aggregations in patches of coarser soil, not necessarily en- tirely bare (although this species is highly gregarious at night and during unfavorable weather, forming massive "sleeping clus- ters" on vegetation ) . This species nested at MOR-B3 and on the peripheiy of A2, in no case forming mixed nesting aggregations with Bembix. The differences among the five species of Philantluts with respect to soil type are summarized in a subsequent section. While related wasps in many cases tend to occupy slightly different soil types, any one site often contains a diversity of species belonging to different genera (a wide- spread phenomenon studied by Elton, 1946). If the soil is of uniform texture, the nests of different wasps may be inter- spersed, but more often there is some 456 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 \ A^v_ \A. V • Bembix a. spinolae ^Steniolia obliqua ■ Hoplisoides sp i log raph us if A mmoph i la azteca O Oxybelus u. quadrinotatus A N itelopter us eva n si Q S tenodyner us papagorum tV Xy locel ia a rgent inae o \ • s __ o Q PARKING LOT * OLD DIRT ROAD o • □ D TREE ■___ A * ______ D |*D* o o ■ — TRAIL . o 10 meters o D. O O o • o VTmTTTTTTTTTTrT^TTTTZ^TTTTTTZWT^TTTm. ^^^^ ^/'/>>>/^/y//W///^^^^^ SNAKE RIVER Map 2. Detailed map of study area MOR-A2, showing location of marked nests of 8 species over a period of several days in late July, 1964. There were many more, unmarked nests of each of these species as well as nests of numerous other species. tendency for segregation of species in ac- cordance with minor differences in con- sistency of the soil. Location of nests of some of the species occurring in MOR-A2 are shown in Map 2 (see also Fig. 1). In this community, certain wasps (notably Bembix and Oxybelus) tended to occupy the very friable sand near the river, certain others occupied more compact sand along trails and roads (especially Ammophila and Stenodynerus) , while still others occu- pied places in partial shade or with more ground cover (Steniolia, Diodontus). In this site, the eight species listed on the map all nested in great numbers during late July and early August, along with many other species, some of them nearly or quite as abundant as those listed (e.g., Episyron quinquenotatus, Nitelopte- rus evansi, Podalonia communis) . Thus the scene in midsummer is one of constant activity involving many hundreds of ground-nesting wasps belonging to literally dozens of species. Moving to a neighboring site of quite different soil type (e.g., to MOR-B3; Fig. 2), one finds a rather different set of wasps, in this case dominated by species of Philanthus, Pisonopsis, and Belomicrus, but with several species in common with A2: Astata nubecula and Steniolia obliqua, for example. Each of the areas studied had its own particular complex of species, with some overlap in content with other areas; Wasps of Jackson Hole • Evans 457 yet much the same parasites (particularly miltogrammine flies) occurred in all areas in considerable numbers. Even bare places only a few square meters in extent (such as MOR-B2 and B3) are able to support many fossorial wasps. The nest burrows and entrances are, after all, only a few millimeters in diameter, and each nest is a temporary affair, being permanently closed and aban- doned after the completion of provisioning (this is less true of Philanthinae, many of which maintain a single, multicellular nest for their entire lives). Female wasps of the same or different species nesting in close proximity rarely show extensive aggression toward one another; thus, it is possible for many nest-eells to be packed into a limited space. To a certain extent the cells tend to be stratified, for some species make very shallow nests, others relatively deep ones. Nest depth is not necessarily correlated with size of the wasp, for some large species (for example, all species of Ammophila) make their nest- cells only a few centimeters beneath the surface. Stratification of nest-cells is well shown in the sandier parts of MOR-A2, where excavations during the inactive season reveal many cocoons of Bembix americana spinolae at a depth of about 8 cm (range 5-11 cm) and many of the smaller cocoons of OxybeJus uniglumis quadrinotatus at a depth of about 5 cm (range 3-7 cm). An assortment of typical nests from MOR-A2, drawn to the same scale, is shown in Plate V. While most attention was focused on fossorial species in this study, it should be pointed out that many of the 190 species recorded from Jackson Hole do not nest in the soil but above ground in trees, bushes, or herbs. Some of them nest in hollow twigs (e.g., Eumenidae such as Ancistrocerus and Symmorphus) , others bore in pith or rotten wood (e.g., Spheci- dae such as Ectemnius and Pemphredon), still others build mud nests on steins (Eumenes), or on rocks (Pseudomasaris) , or paper nests in trees, shrubs, or cavities in the soil (Vespidae). It is usually con- sidered that the primitive wasps were fossorial and that various groups have transferred to aerial nesting sites, thus free- ing themselves from an attachment to bare, friable soil and from competition with soil- nesters. Competition for hollow twigs is, however, often severe (though not studied here). In the areas of study, twig-nesters were especially prevalent at PCR, which is heavily wooded. Since nesters in the re- stricted areas of bare soil in Jackson Hole did most of their foraging for prey in sur- rounding vegetation, and since ground- nesters and aerial-nesters often fed on the nectar of certain flowers side by side, it cannot be said that ground-nesters and aerial-nesters occupied fundamentally dif- ferent food sites. From the point of view of adult and larval food, they were mem- bers of one community, and from the point of view of nesting sites, of two. It should be added that the parasites of aerial-nesters are almost totally different from those of ground-nesters; evidently very different behavioral adaptations are required for exploiting nests in the two situations. Food Relationships Collectively, the wasps of Jackson Hole prey upon virtually all kinds of arthropods available in quantities. Yet to a remarkable degree they "divide up" the prey, each species of wasp specializing on one or a few kinds. A general summary of the prey of the commoner fossorial species is pre- sented in List 1, where the wasps are ar- ranged systematically. Consideration of the prey systematically is also instructive. For example, spiders are exploited as prey by at least two species of Sphecidae and by numerous Pompilidae. Yet any one family of spiders is often utilized by only one predator: Dictynids ( Dictynidae ) by Nitelopterus evansi 458 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 Combfooted spiders (Theridiidae) by Pisonopsis clypeata Orbweavers (Araneidae) by Episyron quinquenotatus Crab spiders (Thomisidae) by Dipogon sayi Jumping spiders (Salticidae) by Pompi- lus angularis Wolf spiders (Lycosidae) Large Lycosa by Pompilus scelestus Medium-sized Lycosa by Cryptochei- lus terminatum Small lycosids by Pompilus occident- alis, Anoplius tenebrosus. LIST 1 Prey of commoner ground-nesting wasps of Jackson Hole (wasps arranged system- atically; see Annotated List, beginning p. 474, for details ) TIPHIIDAE: 3 spp. - Beetle larvae in soil ( Coleoptera ) EUMENIDAE Stenodynerus papagorum - Leaf -mining beetle larvae (Coleoptera) POMPILIDAE Cryptocheilus t. terminatum - Lycosa spi- ders (immature) (Araneae) Episyron q. quinquenotatus - Araneus spiders ( Araneae ) Pompilus angularis - Small errant spiders, mostly Salticidae (Araneae) Pompilus scelestus- Lycosa spiders ( adult ) ( Araneae ) SPHECIDAE SPHECINAE Palmodes carbo -Cyphoderris (Gryl- lacrididae ) ( Orthoptera ) Palmodes hesperus - Anabrus (Tettigoni- idae ) ( Orthoptera ) Podalonia communis - Soil-inhabiting lar- vae of Noctuidae (Lepidoptera) Ammophila azteca - Larvae of small moths or of sawflies (Lepidoptera, Hymenoptera) Ammophila dysmica - Larvae of leaf- feeding Noctuidae (Lepidoptera) Ammophila ma era -Larvae of Sphingi- dae ( Lepidoptera ) PEMPHREDONINAE Diodontus: 3 spp. - Aphids (Hemiptera) ASTATINAE Astata nubecula -Pentatomidae, im- mature ( Hemiptera ) Dryudella montana - Reduviidae, Scutel- leridae, Cydnidae (Hemiptera) LARRINAE Plenoculus d. davisi - Miridae, adult and immature ( Hemiptera ) Solierella affinis - Nabidae, immature (Hemiptera) Nitelopterus evansi - Dictyna spiders ( Araneae ) Pisonopsis clypeata - Theridiid spiders (Araneae) Tachysphex: 5 spp. -Acridid grasshop- pers, immature (Orthoptera) CRABRONINAE Lindenius columbianus - Small parasitic wasps, flies, and bugs (Hymenoptera, Diptera, Hemiptera) Crossocerus maculiclypeus - Very small flies (Diptera) Belomicrus f. forbesii - Miridae (He- miptera) Oxybelus uniglumis quadrinotatus -Di- verse small flies (Diptera) NYSSONINAE Gorytes canaliculatus asperatus - Idioce- rus leafhoppers (Hemiptera) Hoplisoides spilographus - Immature tree- hoppers, Membracidae (Hemiptera) Sticticlla emarginata - Adult moths, Noc- tuidae (Lepidoptera) Steniolia obliqua-BeetUes, Bombyliidae ( Diptera ) Beml)ix americana spinolae - Diverse, medium-sized flies (Diptera) PHILANTHINAE Aphilanthops subfrigidus- Queen For- mica ants (Hymenoptera) Wasps of Jackson Hole • Evans 459 Philanthus: 5 spp. -Bees and wasps ( Hymenoptera ) Eucerceris flavocincta - Adult, medium- sized weevils (Coleoptera) Eucerceris fulvipes - Adult, very small weevils ( Coleoptera ) In some cases this specialization in pre- dation is absolute; for example, there are now literally hundreds of records from many areas which indicate that Episyron quinquenotatus takes only orbweavers. In other cases there is overlap which is more apparent than real: Pisonopsis clypeata, for example, takes an occasional orbweaver, but only small species occurring close to the ground and not utilized by Episyron. Inclusion of other, less common Pompilidae in this listing would complicate the picture, but by no means smudge it completely. Two species of Aporinellus, for example, prey upon Salticidae, but these are very small wasps and undoubtedly take smaller spiders than Pompilus angularis. A number of small Pompilidae besides those listed are known to employ small lycosids, e.g., Anoplius itliaca, A. imbellis, Priocnemis notlia, and some of these use errant spiders of other families as well. It is probable that Cryptocheilus utilizes immature Ly- cosas which, when mature, might serve as prey for Pompilus scelestus. Thus, there would appear to be competition for prey among the numerous predators on errant spiders. Two points should be made here: (1) these are abundant spiders, readily available and not requiring highly special- ized hunting behavior; and (2) the wasps involved are parasitized by certain Pompil- idae (Ceropales, Evagetes) and by milto- grammine flies, and their populations may thus be suboptimal. I would assume that most pompilids, like many other insects, occur at relatively low densities as a result of parasite pressure and thus are able to occupy similar or even identical food niches. In any case, the pompilids in question have different nesting behaviors and often occupy different nesting sites; thus, they are by no means ecological homologues. For example, of the two predators on small lycosids listed, Pom- pilus occidentalis nests primarily in wooded areas, Anoplius tenebrosus in open country. Examination of the predators on major groups of insects reveals situations similar to that prevailing among the spider-hunt- ers. The fauna of Jackson Hole includes quite a number of predators on Orthoptera. Two species of Polmocles take large long- horned grasshoppers; present records indi- cate that they use different species, but this may not hold up. Larropsis capax is known to utilize camel crickets (Ceutho- philus). Five species of Tachysphex use immature short-horned grasshoppers (Acrid- idae), but these species are not all the same size. Furthermore, the larger species use one grasshopper per cell, hence, they take grasshoppers slightly larger than them- selves, while the smaller species use several per cell and generally use grasshoppers smaller than themselves. Hence, the spread in size of prey is greater than that in size of the wasp. The following is a list of the species of Tachysphex, giving the mean length of the females and the mean length of the grasshoppers utilized (in the case of aethiops and nigrior only one prey each was taken; see Annotated List for data). Species of Tachysphex aethiops tarsatus nigrior terminatus sp. nr. linsleyi Mean length of females 13.0 9.5 9.0 7.5 6.5 mm mm mm mm mm Mean length of prey 16.0 mm 10.0 mm 9.5 mm 6.5 mm 6.0 mm Since these species nest more or less simultaneously, and since each species of grasshopper tends to be at one growth stage at this season, it is unlikely that the species of Tachysphex at the ends of this spectrum often use the same species of prey. However, by presenting only the means I have concealed the fact that each species takes grasshoppers over a consider- able size range (see especially Kurczewski, 1966). Undoubtedly considerable overlap 460 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 in prey does occur, and much the same reasoning may apply as in the case of the spider-hunters. The predators on Coleoptera seem to show little overlap in prey preferences. Eucerceris flavocincto is a large wasp using rather large weevils, E. fulvipes a smaller wasp using very small weevils. Leaf-min- ing larvae of weevils and leaf beetles are employed by Stenodynerus papagorum, and it is probable that here as elsewhere twig-nesters of the genera Symmorphus and Leptochilus utilize beetle larvae feed- ing externally on leaves. The species of Tiphia presumably attack the subterranean larvae of scarab beetles here as elsewhere, and Methocha the larvae of tiger beetles. Quite a number of wasps in Jackson Hole employ Lepidoptera as prey. Of the two that are known to employ adult moths (though not actually studied here), Sticticlla emarginata uses Noctuidae, the twig-nester Lestica interrupta small Micro- lepidoptera. Species employing lepidop- terous larvae include especially several Podalonia (using subterranean larvae or "cutworms") and several Ammophila (using external leaf feeders). Among the Ammophila, macra uses very large sphingid larvae, dysmica medium-sized Noctuidae, azteca small larvae of Geometridae and other groups, including sawflies, and pilosa uses principally larvae of lycaenid butter- flies. Five other less common species of Ammophila, unstudied in this area, may well show some prey overlap with these species. Many aerial nesters also utilize lepidopterous larvae. Ancistrocerus catskilh for example, is common in Jackson Hole; this species uses small Microlepidoptera and thus may not compete seriously with any of the ground-nesters. Some of the other Eumcnidae may, however, do so. In the case of the several predators on Diptera nesting in the ground, some slight overlap in prey is apparent between Steniolia obliqua and Bembix americana spinolae and between the latter species and Oxybelus uniglumis quadrinotatus; Crosso- cerus maculiclypeus utilizes extremely small flies not otherwise employed by any ground-nesters studied. There are major gaps in knowledge here, as species of Crabro (ground-nesters) and Ectemnius (nesters in wood) were common in Jackson Hole and are predators on Diptera: how- ever, I obtained no prey records and thus cannot compare them directly with the spe- cies studied. In the case of predators on true bugs ( Hemiptera ) , it should be pointed out that Astata nubecula preys on fairly large stink- bugs, Dryudella montana on small, im- mature bugs of several related families, Plenoculus davisi and Belomicrus forbesi on plant bugs (Miridae) (but so far as known on different species). Other Hemiptera-predators attack treehoppers (Hoplisoides), leafhoppers (Gorytes and presumably Dienoplus, Mimesa, Crosso- cerus wickhamii, and the Dryinidae), and aphids (various Pemphredoninae, includ- ing three species of Diodontus). Leaf- hoppers and aphids are exceedingly abun- dant insects, and it is not surprising that they are attacked by a variety of wasps. None were studied in sufficient detail to determine how much prey overlap occur- red, but it can confidently be said that the wasps involved are diverse in habitat and nest type. For example, two of the three species of Diodontus were common, yet one was confined to friable, flat sand at MOR-A2, the other to roadside banks of coarse glacial outwash at MOR-B3. Rather than prolong this discussion, I should like to consider one other group in greater detail: the species of Pliilanthus, predators on wasps and bees, and among the most abundant wasps in Jackson Hole. These exhibit ecological displacement in an unusually striking manner and also serve to point up the great complexity of communities such as this, for several spe- cies utilize as prey fossorial wasps nesting close to them, while some utilize the para- sites of wasps or of their prey. Wasps of Jackson Hole • Evans 461 Ecological Displacement in the Species of Philanthus1 Despite the abundance of members of this genus in Jackson Hole, it is an unusual experience to encounter more than one speeies at one time and place. At MOR-B, P. pulcher is a dominant species through July, each female making a series of rather shallow nests ( Fig. 36 ) in bare places relatively free of stones; during three sum- mers, I have never collected or seen this species after August 3. P. crabroniformis is an equally abundant speeies, making its first appearance about July 22, but begin- ning to nest in numbers a week or so later. On a few occasions I have seen crabroni- formis females digging in sites occupied by pulcher, but for the most part they select places where the soil is notably harder and stonier. Since wasps of this species make relatively deep, complex nests which they usually occupy for life, and since all ex- pansion of the nest is at lower soil strata (Fig. 38), friability of the surface soil is evidently less critical. Mean cell depth of pulcher is 8.3 cm (range 6-10 cm), while that of crabroniformis is 13 cm (range 9-21 cm). P. pulcher preys upon bees and wasps in approximately equal numbers, including such diverse forms as cuckoo wasps, leaf-cutter bees, and several para- sitic bees (List 6). Most of these are small insects, but a few are about as large as P. pulcher; the mean length of females of this species is 10 mm, the mean length of the prey 6 mm (range 4-11 mm).2 1 1 use the word "displacement" to refer to different ways of exploiting the habitat, not in the sense of physical displacement of one species by another. That is, a species occurring in a particular area and habitat will have features which are displaced by other features in related species also occurring there. Darwin (1859) spoke of this as "divergence of character" and stressed that "more living beings can be sup- ported in the same area the more they diverge in structure, habits, and constitution" ( Mayr, 1963). 2 Length of the prey is a poor measurement of size, since bees and some wasps are notably P. crabroniformis, on the other hand, although averaging slightly larger than pulcher, uses a great many halictid bees much smaller than itself; only occasionally does this species employ wasps as prey (List 7). The mean length of female crabroniformis is 11.5 mm, that of the prey only 5.5 mm ( range 4—8 mm ) . One species of wasp and six species of bees appear on both lists of prey; 15 per cent of the species utilized by pulcher were also utilized by crabroniformis, but these make up about one third of the records for pulcher (spe- cies of Dialictus being the most commonly used prey of both species). Thus there is a fair amount of overlap in prey, the larger species curiously tending toward smaller prey than the smaller. However, the two species displace one another almost com- pletely with respect to soil type and stratum and with respect to nesting season. In the final analysis, they may be char- acterized as non-eompetitors. A third species, P. pacificus, appears in late July, at about the same time as crabroniformis. It is a considerably smaller species (mean length of females about 9 mm), yet a comparison of available prey records reveals that the two compete for much the same bees and wasps (mean length of prey of pacificus: 5.2 mm, range 4-7 mm). Both use only an occasional small wasp, but use ground-nesting bees in great numbers, small Halictidae making up 78 per cent of the prey of pacificus, and 96 per cent of the prey of crabroni- formis. The majority of species taken as prey by pacificus (List 9) also appear on the list for crabroniformis. Evidently these bees are sufficiently abundant in August to support aggregations of two species of Philanthus. However, these two wasps ex- hibit a total separation in nesting sites, pacificus being confined to fine-grained, broader and heavier than such wasps as Mimesa, Ammophila, etc. However, it was the only con- venient measurement I could make with the available time and facilities. 462 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 light sand in bars close beside the river genera (no spp.) of wasps and eight (at MOR-A2) or in the center of sandy genera (3 spp.) of bees in common. How- roads near the river (at ELK). In no case ever, I found the two species nesting in the did I find these two species nesting within same situation only once (at ELK) and 200 meters of one another. Like most then on a date (July 10) when only pulcher Philanthus, they probably do most of their was nesting. hunting on flowers (taking many more The fifth species of Philanthus occurring male bees than females), and it is probable in Jackson Hole, bicinctus, I shall discuss that they exploit different patches of Soli- only briefly, since I encountered it only dago and other flowers, at least for the occasionally, and since Armitage (1965) most part. has discussed a nesting aggregation at the A fourth species of Philanthus, zebratus South Gate of Yellowstone. This species nitens (referred to hereafter simply as makes very deep nests in thinly vegetated zebratus), appears in mid- July and is active slopes where the soil is veiy coarse and until mid-August. This is a large, colorful stony; the prey consists principally of species which nested in large numbers in bumblebees of several species, rarely of moderately friable sandy soil at ELK and other large bees. Thus it is isolated from also along a sandy road at HHS; I did not the other four species both with respect take the species at MOR at any time. At to habitat and with respect to prey. ELK, zebratus occupied more hard-packed Differences among the five species in sand than pacificus, and at HHS more seasonal cycle and size of the females are friable sand than crabroniformis. I would summarized in Text-figure 1; however, it characterize this species as preferring a should be remembered that size of the soil type intermediate between those two wasp is not a wholly dependable reflection species, and the absence of a suitable ex- of size of the prey, for cral)roniformis uses panse of such soil at MOR may explain prey averaging smaller than pulcher. Also, its absence there. In any case, zebratus the figure makes no attempt to indicate exhibits almost no prey overlap with either differences in soil type utilized for nesting, of those species, since it uses notably larger which ( although characterized here only prey. The mean length of female zebratus subjectively) is evidently at least slightly is about 14 mm, the mean length of the different in each of the five species, and recorded prey 11.3 mm (range 8-18 mm), quite conspicuously different in some cases This figure is somewhat deceptive, since (e.g., pacificus vs. crabroniformis). Some the longer prey consisted of Ichneumonidae of the several conspicuous behavioral dif- and such Sphecidae as Ammophila, which ferences between the five species evidently are very slender-bodied. Nevertheless, serve to sustain the almost complete eco- zelnatus does clearly occupy a different logical displacement among these species, food-size niche than pacificus and crab- Others probably serve to maintain reason- roniformis; no species and only one genus ably high population sizes in spite of the appear in common on the lists for pacificus presence of an abundance of parasites. It and zebrat us (List 8), and only one species will be more profitable to explore this and two genera in common on the lists for subject after surveying the parasites briefly. crabroniformis and zebratus. As a predator, zebratus has much more in common with KlNDS AND Specificity of Parasites pulcher, since it uses wasps and bees in about equal numbers and uses prey averag- The approximately 100 species of ing only slightly smaller than itself. The ground-nesting wasps occurring in the prey lists for these two species have five areas of study were attacked by a consider- Wasps of Jackson Hole • Evans 463 bicinctus zebratus 10 cm- pulcher pacificus crabroniformis 4-10 II- 17 18-24 25-31 1-7 JULY 8-14 15-21 AUGUST 22-28 Text-figure 1. Comparison of the five species of Philanthus occurring in Jackson Hole with respect to mean size of females (width of bars) and period during which females provision their nests. able array of parasites and predators1; 25 species of Diptera and Hymenoptera were either actually shown to attack these wasps, or can be assumed to attack them with a high degree of probability. Some of these are relatively host-specific, attacking one or a few related species of wasps, while others show little or no specificity, attack- ing ground-nesters of many diverse species. Specificity implies a measure of coevolution of host and parasite; that is, the two may acquire behavioral traits that permit them to coexist without marked population de- pressions. On the other hand, non-specific parasites may have varying success depend- 1 In fact, all of these are predators as that word is most commonly used. Like many entomologists, I use the word "parasite" loosely and here apply it to two kinds of associations: (1) parasitoids, which feed upon the host slowly and destroy it as they reach maturity; and (2) cleptoparasites, which feed principally upon the prey in die cell, but usually also kill the host larva. "Predators," as I use the word here, are also of two kinds: ( 1 ) those that capture and feed directly upon the prey, e.g., asilid flies; and (2) those that capture and paralyze prey and take it to the nest as food for the larvae, e.g., digger wasps. ing upon the behavior of their host, which cannot usually have evolved behavior pat- terns fitted to reducing the success of all of its various attackers. Thus, non-host- specific parasites and predators may be more significant in producing population crashes in certain hosts ill-adapted to them and in bringing about rapid but narrowly adaptive changes in nesting behavior. Good examples of host-specific parasites are provided by members of the cuckoo- wasp genus Pa mopes, several species of which have a wide distribution but are seldom abundant, perhaps because the sand wasps they attack have evolved nest closures sufficient to delay or prevent entry by the female parasites. In Jackson Hole, P. edwardsii is a parasitoid of Steniolia obliqua and probably of Bembix americana spinolae and B. amoena, the female digging through the nest closure to oviposit on the larva ( Evans, 1966a ) . Chrysidids of the genera Hedychrum, Hedychridium, and Ceratochrysis were also found associated with certain ground- nesters (see Annotated List below), but the degree of host-specificity is not known. 464 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 The species of Bembix are also attacked by two apparently host-specific dipterous parasites: Physocephalo texana, a conopid fly attacking the adults, and Exoprosopa dorcadion, a bee-fly that oviposits in open holes, its larva attacking the wasp larva. Hoplisoides spilographus and probably the species of Gorytes and Dienoplus are at- tacked by wasps of the genus Nysson, while various spider wasps are attacked by members of the eleptoparasitic pompilid genera Ceropales and Evogetes. (Records will be found in the Annotated List; for further details on these and related species see Evans and Yoshimoto, 1962, and Evans, 1966a.) This essentially completes the list of parasites that confine their attacks to a restricted group of hosts in this area. None of them are overly abundant in Jackson Hole, and there is no evidence from here or elsewhere that they cause profound de- pressions in the population sizes of their hosts. Some of the non-host-specific parasites, notably the miltogrammine flies (Sarco- phagidae), stand in strong contrast to these. These flies are abundant in virtually all areas inhabited by ground-nesting wasps in Jackson Hole as elsewhere. The most ubiquitous species is Senotainia tri- lineata (Fig. 10), often spoken of as a "satellite flv," since the females follow* closely behind prey-laden wasps and larvi- posit upon the prey, often just as it is being taken into the nest. The small maggots typically destroy the egg of the wasp within a few hours after it is laid, then develop at the expense of the prey in the cell. Generally several maggots are de- posited at once, and several (rarely up to 20) may develop successfully in one cell. In Jackson Hole, I have reared this fly from the cells of six species of Sphecidae, and altogether 17 species belonging to three families may be regarded as probable hosts (List 2). In fact, records from many localities suggest that these satellite flies follow wasps of many genera of several families and that their larvae develop suc- cessfully in the nests of most species. How- ever, in the case of maggots introduced by progressive-provisioners (such as Bembix) after the larva has begun to grow, the wasp larva and the maggots may both survive; I record one such case for B. americani spinolae and Senotainia trilineata below. LIST 2 Hosts of Senotainia trilineata at Jackson Hole1 EUMENIDAE Stenodynerus p. papagorum (Viereck)* POMPILIDAE Episyron q. quinquenotatus (Say)** SPHECIDAE Aphilanthops subfrigidus Dunning* Astata nubecula Cresson Belomicrus f. forbesii (Robertson) Bembix americana spinolae Lepeletier Gorytes canaliculatus asperatus Fox* Nitelopterus evansi Krombein* Oxybelus uniglumis quadrinotatus Say Philanthus crabroniformis Smith PJiilanthus pacificus Cresson* Philanthus pulcher Dalla Torre* Philanthus zebratus nitens (Ranks) Plenoculus davisi Fox** Tachysphex nigrior Fox* Tachysphex tarsatus (Say)** Tachysphex terminatus (Smith)** Two other satellite flies, Hilarella hilar- ella and Taxigramma heteroneura, were taken at MOR, the first associated with three unrelated species of Sphecidae, the second with two unrelated sphecids. I did not rear these flies from any nests and little is known of their biology, but I would assume that they are similar to Senotainia 1 Reared from nests unless marked with an asterisk: * Seen following females but not reared from cells; ** Known to be attacked on basis of studies elsewhere. Wasps of Jackson Hole • Evans 465 in their mode of attaek and development, high: about 20 per cent in Philanthus as well as in their lack of host-specificity. zebratus nitens, a species possessing several The miltogrammine fly Metopia argyro- behavioral mechanisms evidently function- cephala was reared from a nest of Gonjtes ing to reduce the success of parasites. canaliculatus asperatus and was found There is evidence that these flies sometimes elosely associated with the nests of five cause marked decline in local populations other spheeids. This fly also sometimes of certain digger wasps, follows female wasps to their nests, but Among the general predators at Jackson rather than larvipositing on the prey in Hole should be mentioned asilid flies, j transit, it enters the nest and larviposits in including at least two species of Laphria, the cell. "Hole-searchers" such as this prob- which tend to prey extensively upon ably also locate open nests by random Hymenoptera. These flies were often seen searching. perching on logs or bushes overlooking One of the most abundant Miltogram- nesting sites and dashing periodically at iminae in Jackson Hole is Phrosinella pilosi- flying insects; one was taken with a \frons, a relatively large fly with slightly sphecid wasp in its grasp, while others expanded front tarsi that is often seen were seen with bees and sawflies. Spiders digging at closed nest entrances (Fig. 14). of many kinds also occur on the ground I have never observed this fly trailing and in vegetation and probably take a con- female wasps or being attracted to open siderable toll among the small wasps, al- holes; rather it moves from one closed nest though I obtained only one specific record entrance to another where it now and then of predation by a spider, attempts to dig through the closure, when Finally, it should be remembered that successful presumably larvipositing in the four species of Philanthus prey upon wasps, burrow or cell. Whether it is attracted to two of them quite extensively. While the nests by the sight of fresh digging or by majority of wasps taken as prey are males, odor has not been determined. The mag- a good many females are also taken, in- gots of Phrosinella are large, and usually eluding species nesting in close proximity one develops per nest cell (in a rearing tin to Philantluis (Lists 4 and 7). It is inter- one maggot may move to other cells and esting to note that while Sphecidae, destroy them). In Jackson Hole, I found Eumenidae, Masaridae, Chrysididae, and this species associated only with the species even parasitic Hymenoptera appear on the of Philantluis, and reared numerous flies lists of Philantluis prey, Pompilidae do not. from the cells of three species: pulcher, Pompilids are certainly abundant at Jack- crabronifonnis, and zebratus. However, son Hole both on the ground and on other species of Phrosinella are known to various flowers visited by Philanthus. Pos- attack a diversity of ground-nesters that sibly their swift and erratic flight patterns maintain closures, and I would expect the render them difficult to capture. It should same to be true of pilosifrons. also be noted that in no case has a There is need for detailed field studies Philanthus been found to take another of these and other species of Miltogram- Philanthus (of any species) as prey, even minae, for these flies are unquestionably though zebratus has been found to use the major pests of digger wasps and may have very similar and closely related wasp played an important role in molding Aphilanthops subfrigidus. Although Rath- various attributes of their nesting behavior, mayer (1962) has shown that Philanthus as discussed below. Despite various de- is immune to its own venom, it seems more vious flight patterns, closures, accessory probable that these wasps are able to avoid burrows, and so forth, the percentage of attacking members of their own genus, cells successfully attacked may be quite perhaps by olfactory cues. This would be 466 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 an exciting problem to study experimen- tally. Summary of Interrelationships Digger wasp communities of Jackson Hole involve food webs that cannot ade- quately be summarized because of their complexity and because of shortcomings in our knowledge. The wasps are energetic predators that harvest great numbers of arthropods within and around their nesting sites. For example, if there were 50 nesting female Bembix americana spinolae at MOR-A2 and each prepared five nests containing 20 flies each (all highly con- servative estimates), these wasps would have collected 5000 flies from the sur- rounding countryside. The aggregation of Philanthus zebratus at ELK contained an estimated 200 females, each of which pre- pared an estimated 15 cells provisioned with an average of about six prey; thus the members of this population must have collected at least 18,000 bees and wasps, evidently chiefly from flowers in the gen- eral vicinity of the nesting area. Considering the abundance of ground- nesting wasps in certain areas and their efficiency as predators, it is not surprising that most species occupy different food niches. That some species do appear to prey upon the same arthropods during the same season may be attributable to: (1) the great abundance of such insects as aphids, leafhoppers, small halictid bees, etc.; and (2) the fact that virtually all wasps may maintain submaximal popu- lation sizes as a result of widespread parasitization, especially by miltogrammine flies. During seasons when, for climatic or other reasons, certain arthropods are unusually scarce, wasps which occupy similar food niches may be especially af- fected, and during seasons when one par- ticular non-specific parasite is especially abundant, those species relatively unpro- tected against it may suffer profound population declines. Various short-range and long-range climatic factors may affect predators, prey, and parasites differently, thus providing a continually changing environment which must be met by be- havioral adjustments of one kind or an- other, or decline and local extinction may follow. Some of these changes were obvious during the three summers I worked in Jackson Hole. In 1964, deerflies (Chrysops) were exceedingly abundant at MOR, and these flies were used as prey by Bembix almost to the exclusion of other flies. In 1967, doubtless as a result of a different rainfall or temperature pattern, deerflies were rarely in evidence, and the Bembix had shifted wholly to other flies (List 3). Some of these other flies provide the com- mon prey of Oxybehts and of Steniolia obUqua. The latter species underwent a steady decline in numbers at MOR from 1961 to 1967. To what extent this has been due to competition for prey with Bembix, to success of its several parasites, or to changing climatic factors, is unknown. Since Steniolia often nests in the middle of paths and dirt roads, its decline may be largely the result of increasing numbers of humans and horses. Human interference may, in fact, be having an important in- fluence on wasp populations even in this supposedly sacrosanct area. I suspect, for example, that construction of the Jackson Lake dam and resulting control of flooding along the Snake River may eventually diminish the alluvial deposits along the river. Sand bars have their own life cycles, and unless refreshened by wind or water eventually become overgrown and the soil greatly altered. Even in the absence of human interfer- ence, changes in soil conditions and in vegetative cover are continually occurring, and these changes result in the expansion and contraction of suitable nesting space for certain species. It is difficult to evaluate such factors, as little is known of the precise edaphie relationships of wasp spe- cies. As in the case of food resources, it is probable that the abundance of natural Wasps of Jackson Hole • Evans 467 SPIDERS MIDGES MOTHS CATERP ILLARS GRASSHOPPERS LEAFBUGS STINKBUGS LEAFHOPPERS TREEHOPPERS APH IDS BEETLE LARVAE WEEVILS OTHER FLIES AS ILID FLIES BEEFLIES CONOP ID FLIES MILTOGRAMMINE FLIES CHNEUMONS CUCKOO WASPS BEES ANTS Text-figure 2. Diagrammatic representation of some of food relationships of wasps and their prey and parasites at Jackson Hole. Direction of arrow indicates direction of predatory or parasitic relationship (for energy flow, read arrows backward). enemies ameliorates the intensity of com- petition for space; that is, many populations may be kept at levels sufficiently low that they do not fill all suitable nesting areas. At the same time, the need for bare, friable soil restricts many of these wasps to areas of loess or alluvium and thus pushes many species into close proximity. In these re- stricted areas, the build-up of non-specific parasites may play a major role in permit- ting such a diversity of species to co-exist (see, e.g., Paine, 1966). I have attempted to delineate a few of the food relationships in the areas of study in Text-figure 2. Obviously, the arthropods used as prey by the wasps have other natural enemies, some of them in them- selves serving as prey for wasps. The common orb-weaving spider Araneus pata- giatus, for example, is the major prey of Episyron quinquenotatus, but itself preys on the small flies that provide the major prey of Oxybelus and Crossocerus. Thus an increase in the population of Episyron might permit an increase in the species of the latter two genera. The Ichneumonidae preyed upon by Philanthus zebratus are parasites of cateq}illars, some of which may serve as prey of Ammophila and other wasps. Thus a high population of Philanthus might favor such a wasp as Ammophila azteca — except that, ironically, P. zebratus also preys directly upon Am- mophila. P. pulcher preys upon cuckoo wasps to some extent, including species known to attack other ground-nesting wasps. The species of Philanthus, indeed, provide the most interesting elements in this fauna, preying as they do on so many wasps, wasp parasites, and parasites of other insects — also, of course, on many bees, which to a certain extent compete with wasps for nesting sites and nectar sources. Insufficient information is available to explore these interrelationships further — in fact, I have already gone farther than the data warrant — and I should like to turn to the second question posed at the beginning of this section: to what extent 468 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 are species differences in behavior under- standable in the contexts of interspecific competition and of widespread parasiti- zation? BEHAVIOR PATTERNS AS ADAPTATIONS TO LIFE IN THE COMMUNITY There exists a very extensive literature describing the behavior patterns of solitary wasps, for it is true" that these patterns are subject to relatively little intraspecific variation and that species differences, even between closely related forms, are often pronounced. However, few attempts have been made to explain these different be- havior patterns in terms of their adaptive value. Why do related species often show pronounced differences in prey preferences, manner of approach to the nest, nest closure, mound-leveling, accessory bur- rows, and so forth? In my opinion the two factors emphasized here — competition and parasitism — are sufficient to explain the adaptive value of nearly all of the species- specific behavior patterns of solitary wasps (see also Evans, 1966a, Chapter XV). Behavior Related to Competition1 The relatively high degree of ecological displacement among the wasps of Jackson Hole is by no means a unique phenomenon; a very similar picture is presented by the solitary wasps of eastern Massachusetts, for example. Indeed, that ecological homo- logues cannot coexist is a widely recognized biological phenomenon (DeBach, 1966). 1 It may be argued that neither food or nesting sites have been shown to be in short supply, and thus competition cannot be demonstrated. But, as pointed out in this section, speed is important to these short-lived insects, and if two species prey on the same thing or nest in the same place, any interference or diminution in space or in the numbers of prey which lengthens their nesting cycles will result in a reduction in number of offspring. I use the word "competition" in this broad sense, to include any interspecies rivalry that is selectively disadvantageous (see also De- Bach, 1966). There are no well-documented exceptions among the wasps of Jackson Hole, for those few species appearing to have identical or broadly overlapping food niches occupy different nesting sites — and this in itself suggests that they may hunt in somewhat different areas. Obviously several sets of behavior patterns are involved here: selec- tion of a nesting site; construction of a nest compatible with that soil type; selection of a hunting site; and response to a particular kind of prey. No studies were made of hunting be- havior, but it can be assumed that even slight differences in prey selection involve important differences in responses to stimuli. A given species of wasp may be narrowly or fairly broadly adapted with respect to prey selection. Aphilanthops subfrigidus, which (like its congeners) preys only on queen ants of the genus Formica during their nuptial flights, must have highly specialized hunting behavior indeed. On the other hand, Bembix ameri- cana spinolae is known to forage widely and to respond to Diptera of various shapes and sizes — including such diverse flies as asilids, tabanids, and sciomyzids. Aphilan- tliops has the advantage of being the only wasp in this area using Formica ants as prey, but the success of these wasps is very much tied up with the size of nuptial flights and weather during such flights. Bembix uses large numbers of deerflies (Chrysops) when these are plentiful (a prey not known to be used by other wasps in this area), but when they are not in abundance it is readily able to utilize small muscoids (a common prey of Oxijbelus) or beeflies (the almost exclusive prey of Steniolia). Such a broadened response to potential prey may have evolved in an area where there were fewer species of digger wasps and no competitors for larger flies. Under conditions of severe competi- tion, one would expect a narrowing of the spectrum of response, perhaps at first non- genetic but later becoming fixed in a local population. Under conditions of prolonged Wasps of Jackson Hole • Evans 469 isolation, such a population might emerge diy sand tend to make a series of separate as a species with restricted prey prefer- nests, often with one or a few cells; pre- enees. sumably such nests are less amenable to There is evidence that various species wholesale destruction by parasites than of Bcmbi.x show differences in prey type nests in which many cells arise from a in various parts of their ranges, though to single burrow. But in hard-packed, stony what extent this is genetically determined soil this presumed advantage is outweighed is unknown. The same is true in Philan- by the problems of digging. That is, it is thus. P. pacificus, for example, has been advantageous for a species such as Philan- studied at Antioch, California, and found thus crabroniformis to make a single to prey upon wasps and bees in approxi- burrow and ultimately to prepare many mately equal numbers, also on parasitic cells from it, even though such a nest may Hymenoptera such as Braconidae and be more subject to parasitization by cuckoo Ichneumonidae (Powell and Chemsak, wasps or Phrosinclla flies than a more 1959). P. crabroniformis has been ob- temporary nest with only a few cells. The served at Delta, Utah, capturing honeybees female Philanthus crabroniformis requires at the hive entrances (under the name several days to dig her burrow, while P. flavifrons Cresson; G. E. Bohart, 1954). pacificus may complete a burrow in a sand Both of these species prey very largely on bar in a few hours. The first species small halictid bees in Jackson Hole. usually nests from a single burrow through- It should be pointed out that the success out its life, making fifteen or more cells of a predatory wasp is very much de- ( Fig. 38 ) , while the second prepares a pendent upon the speed with which it is series of shallow nests, each with about able to find and capture prey; if it is forced three cells ( Fig. 35 ) . The correlation be- to search for an uncommon prey or to tween soil texture and number of cells per subdue an unusual prey, the result will be nest is by no means absolute, but it is that during its limited lifetime each female true that most of the nesters at MOR-B will have time to provision fewer cells than made multicellular nests (e.g., Astata usual. This being true, there will be great nubecula, two species of Eucerceris, and selective advantage in finding and captur- BeJomicrus forbesii) although a few pre- ing a particular kind of prey quickly, that pared cells from pre-existing burrows is, toward more specialized hunting be- (Cryptocheilus terminatum, and Pisonopsis havior. My observations on Steniolia clypeata). Here again, speed of work is obliqua hunting beeflies (Evans and Gil- important to a wasp living only a few laspy, 1964) suggest that it is a much more weeks, and those species nesting in rela- effective predator on these swiftflying tively intractable soil have evolved mecha- Diptera than is Bembix. Thus Bembix may nisms for reducing the total amount of be at a disadvantage if forced to rely digging that must be done, heavilv on beeflies in an area where a specialist on beeflies, Steniolia, is abun- Behavior Related to Parasitism dant. I include here all behavior that appears As already pointed out, related wasps to function in reducing the success of seldom nest in identical sites. This results natural enemies. It is, of course, possible in a partitioning of available space and also to consider these behavioral devices also tends to disperse aggregations of different as a form of competition. Crombie (1947) species such that their hunting arenas are remarks that "agencies such as predators at least partially separate. Occupation of may produce a different kind of competi- different soil types also influences the type tion rather than preventing it altogether, of nest. Species nesting in highly friable, They may, for instance, prevent compe- 470 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 tition for food among their hosts, but cause them to compete to avoid the preda- tors, e.g., there may be selection for reflexes which assist this." I would agree with this statement, but nevertheless feel that it is useful to distinguish between direct interspecies competition and the competitive success of various species in reducing the attacks of natural enemies. For example, prey selection is evidently related to interspecies competition, since it results in a partitioning of the available resources, as already pointed out. How- ever, the manner in which the prey is carried to the nest has apparently evolved in response to parasite pressure, as I have argued elsewhere (Evans, 1963a). Jackson Hole has many wasps that cany the prey with their mandibles, but these wasps must either deposit the prey while they open the nest, thus rendering the prey tempo- rarily unprotected against satellite flies, tiger beetles, ants, and so forth; or they may simply leave the nest open, making it available to beeflies, hole-searching milto- grammine flies, and so forth. Jackson Hole also has many species that carry the prey with their middle legs, where it is well protected beneath the body and where it remains while the nest entrance is opened with the front legs. Only one species in this area carries the prey on the end of its abdomen, but that is one of the common- est species, Oxybelus uniglumis quadrino- tatus. This species is a very rapid pro- visioned and I assume speed has been selected for at the expense of exposing the prey to satellite flies. Two of the larger species of Philanthus, both heavily attacked by Senotainia flies, exhibit strikingly devious flight patterns upon approaching their nest with prey. In the case of P. crabroniformis, prey-laden females (whether or not followed by satel- lite flies) almost always fly not directly to the nest but to a low herb or grass blade 0.5-2.0 m from the entrance. Here they often cling for a few seconds before pro- ceeding to another perch or ( if no flies are present) to the nest. Once at the entrance, females almost always dig through the closure and enter, even though a satellite fly may have appeared in the meantime. Some females followed by flies move about from perch to perch (always within 20 cm of the ground) at varying distances from the nest for as long as five minutes before entering. On several occasions females were seen to lose their would-be parasites before reaching the nest. The corresponding flight patterns in P. zebratus differ in some details. These wasps, with their generally rather large prey, typically descend into the nesting area from a considerable height, landing with an audible sound on the ground and sitting motionless for several seconds be- fore continuing to the nest. If followed by satellite flies, they fly off close to the ground in a circuitous pattern, often dis- appearing from sight for several seconds to a minute or two. This may be repeated several times if the flies persist or reappear. Although prey-laden females may land on the ground occasionally, they seldom land on vegetation in the manner of crabroni- formis. Some females followed by flies rise high in the air and descend abruptly as in their initial return flight. These two species of Philanthus are also the only wasps in Jackson Hole known to maintain accessory burrows. These short burrows are dug beside the true burrow and are left open while the true burrow is closed. In crabroniformis, a nester in hard soil, newly completed nests lack an acces- sory burrow, but as females replace their closures again and again they often scrape much of the soil from small quarries beside the nest entrance, so that within a few days numerous nests have such quarries, later in the season still more. For example, on August 3, I counted 26 nests, of which four had one quarry each. On August 12, a count of 40 nests (including all of the original 26) showed that about half had quarries, several of them two (one on each side or both on one side). The quar- Wasps of Jackson Hole • Evans All ries varied in depth from 2 mm to 2.5 cm. Females were often seen entering the quarries and occasionally remaining within them for a few seconds. Their primary function was unquestionably as a source of fill in this hard soil. I did not observe any parasites being attracted to them, although it is probable that this sometimes occurs (for a general discussion of acces- sory burrows, see Evans, 1966b). P. zebratus nests in much more friable soil, yet some females appear to quarry much of the sand for closure at one or two points, creating short accessoiy burrows which may be redug from time to time. On August 2, I estimated that 20 per cent of the nests had accessory burrows, and there was no evidence of an increase as the sea- son progressed. Actually, about half of these nests had two such burrows, and several had three. They varied from barely measurable' to 3 cm in depth. On three occasions hole-searching flies, Metopia argyrocephala, were seen entering these accessory burrows. It seems probable that in zebratus the burrows serve a more im- portant role in diverting parasites, than as a source of soil. These two species of Philanthus do not level the mound of soil at the nest entrance, merely leaving it in a fan-shaped pile. In contrast, mound-leveling behavior follow- ing completion of a new nest is elaborate in P. pulcher and pacificus. Mound-level- ing behavior is also elaborate in Bembix americana spinolae, although absent in the related wasps B. amoena and Steniolia obliqiia. In Ammophila azteca and other species of this genus quite another method of removing soil from the edge of the bur- row is employed: the female carries small loads forward in flight and drops them some distance away. A variation on this "soil-carrying" theme is employed by Belo- micrus forbesi, the females of which fly swiftly backward from the nest, each time carrying a lump of earth and dropping it on the ground. It is assumed that removal of the soil from the nest entrance renders the site less evident to parasites (as it does to humans). Cuckoo wasps and Phrosinella flies are often seen flying from one nest entrance to another, landing on the mound of soil and searching about for covered holes. They will also alight, as least briefly, on artificial piles of sand placed in nesting areas, suggesting that their initial attrac- tion is to the sight of the mounds. Closure of the nest entrance between hunting flights is also regarded as a device for rendering the nest less visible and also for delaying or preventing the entiy of diggers such as Phrosinella or Pa mopes. This belief is supported by observations and simple experiments. Bombyliid flies oviposit into many kinds of holes, includ- ing artificial holes made by a pencil in the soil. Metopia flies also enter various open holes, including blind accessory burrows, as mentioned above. There is, however, a paradox here: Phrosinella evidently is not attracted to open holes, but only to closed nests (though presumably preferring thin closures to thick ones); and Senotainia is favored by closures, since they cause the wasps to pause at the entrance to dig it open and thus provide a better target for larviposition. The fact that the vast ma- jority of the fossorial wasps of Jackson Hole do maintain a closure suggests that these factors are generally outweighed by others, that is, when and where closing behavior evolved Senotainia was a less important parasite than some others. The apparent success of Phrosinella pilosifrons against Philanthus may reflect the fact that al- though these wasps do close the nest, the closure is usually rather weak. The species of Tachysphex ought to pro- vide unusually good material for a study of the influence of closure on success of parasites, especially since the nests are shallow and easy to excavate. One would assume that the species making no closure (tarsatus and nigrior) might be more heavily parasitized by Metopia and less heavily parasitized by Senotainia than the 472 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 species that make a closure (terminatus and sp. nr. linsleyi). On the other hand, the fact that the first two species use relatively larger grasshoppers, which protrude fur- ther behind their bodies in transit, may mean that they present a better target for satellite flies, thus offsetting the advantage of quick entry into the nest. Unfortunately, none of the species of Tachysphex are abundant in Jackson Hole. The digger wasps that omit the outer closure include several that make vertical burrows. Such burrows cannot be closed in the usual manner of simply scraping soil into the entrance; either they must be left open, or special mechanisms evolved. The species of Eucerceris leave the entrance open, but possibly compensate for this by keeping a certain amount of loose sand in the lower part of the burrow. Stenodynerus papagorum makes a vertical burrow, but in this case it is capped by a mud turret ( Fig. 23 ) . These turrets may have evolved as deterrents to parasites, but I have seen cuckoo wasps entering them, and Olberg (1959) has provided a photograph of a cuckoo wasp entering the turret of a dif- ferent species. Although the species of Ammophila make short, vertical burrows, they have evolved unique behavior en- abling them to close their burrows effectively. Briefly, the female finds an object that just fits the lumen of the burrow, fits it in place, then puts additional lumps and loose soil above it ( Fig. 31 ) . Final closure of the nest in Ammophila azteca is also enhanced by the "tool using" behavior of this species, which surely serves to provide a firm barrier against parasites and to form a plug which does not erode and expose the location of the shallow nest to predators. Many other wasps make a fairly lengthy final closure (Bembix americana spinolae and Philanthus pulcher, for example), but whether the species of Philanthus that nest in one bur- row for their entire lives usually fill up the burrow before they die is doubtful. Progressive feeding of the larva probably also arose in response to parasite pressure (Evans, 1966a). In Jackson Hole, Am- mophila azteca, Bemhix americana spin- olae, and Steniolia obliqua are the only known progressive provisioners. Since these wasps require five days or more to com- plete a single cell (as compared to usually about one day for mass provisioners), they would seem to be at a disadvantage when the seasons are as short as they are here. The three may thrive simply because al- most none of the larvae succumb to para- sites; as noted above, miltogrammine fly maggots and wasp larvae may both survive in such nests, at least at times. Ammophila azteca, furthermore, has evolved an un- usual behavioral feature peculiarly adaptive for these short summers: successive nests overlap, the female maintaining two to four nests at different stages of develop- ment and remembering the location of each (Evans, 1965). Admittedly the success of these various behavioral devices has yet to be docu- mented quantitatively. In fact, it is un- certain whether it can be, both because of the difficulty in obtaining such data foi ground-nesting species and because the data, if obtained, may not mean a great deal. The parasites and their hosts have undoubtedly undergone a measure of co- evolution, and if a particular behavioral device does not seem effective, it may mean that the parasites are, at this stage, "one step ahead." Summary of Behavioral Adaptations No two of the wasp species studied in Jackson Hole are behaviorally alike in every respect. This suggests that each has evolved to meet somewhat different prob- lems. There is no way of knowing where or when these distinctive patterns evolved. The present is the only time-transect we have, and the interplay among the various members of the fauna today can only sug- gest some of the problems these species have surmounted. At least in this context many otherwise puzzling behavior patterns Wasps of Jackson Hole • Evans 473 make sense." Our explanations may not be complete, or even correct, but at least they provide the basis for a fuller under- standing of the biological role of each spe- cies attribute than any amount of study of museum specimens is likely to provide. In these restricted areas of bare, friable- soil in Jackson Hole — and in similar areas elsewhere — a wasp must have certain be- havioral capabilities in order to be success- ful. Some of these are as follows: ( 1 ) It must hunt prey which is readily available and not overly exploited by other predators, and its hunting behavior must be closely attuned to the habitat and be- havior of that particular prey. (2) It must utilize available sources of nectar or honeydew for its own nourish- ment without undue interference or over- exploitation of the source by bees or other wasps. (3) It must nest where there is available space not densely filled with other wasps or with ground-nesting bees, and it must construct a nest appropriate to that soil type. (4) It must be able to provision its nest quickly and with a minimum of delay due to prey scarcity or to interference from other wasps or from parasites, especially in an area with a short summer season. (5) It must possess mechanisms for re- ducing the success of parasites to such an extent that each female is able to leave at least two progeny that survive to adult- hood. Examples of such mechanisms in- clude pedal prey carriage, devious flight patterns, nest closure, mound leveling, accessoiy burrows, progressive provision- ing, and so forth. Possession of these attributes in various forms — and, of course, such other obvious ones as the ability to survive the winters at this latitude and altitude — has enabled these species to coexist, some with greater success than others. Paradoxically, it may be the presence of so many natural enemies that permits so many species to occur to- gether, as discussed earlier; yet each spe- cies appears committed to evolutionary strategies for incurring a lessened incidence of successful attack by these enemies. It may be instructive to summarize these strategies for the species of one genus. Philanthus has five species in Jackson Hole, and, by coincidence, five common species in eastern Massachusetts (all dif- ferent). Eventually I hope to complete a detailed comparative study of these ten species, but for the present I offer a tabular comparison of some of the more salient attributes of each (Table 1). In each area there are certain species in which mound- leveling is well developed (all of these are members of the politus species-group). In each area there are species in which acces- sory burrows are sometimes constructed, and in the eastern states one (lepidus) in which such burrows form a fixed feature of behavior (although varying in number). All Jackson Hole species make a nest closure, but two eastern species omit the closure. Such differences presumably re- flect the relative importance of satellite flies, beeflies, and other natural enemies at the times and places when these be- havior patterns evolved; e.g., open nest entrances permit rapid entry and may re- duce the success of Senotainia, while per- mitting access to the nest by Metopia, bee- flies, etc. Others (prey type and soil type, also size, time of emergence, and other qualities not tabulated here) probably represent mechanisms for reducing direct competition with congeners. A similar analysis of other genera would provide a very similar picture. It is obvious that the functional signifi- cance of these various behavior patterns remains unproved. The conclusions I have drawn are based on extensive field work in Jackson Hole and elsewhere, but I would be the first to admit the need for many more observational data, for experiments designed to test critical points, and for mathematical approaches such as those of many contemporary population biologists. 474 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 Table I Comparison of ten species of Philanthus with respect to six behavioral features. Levels mound Temporary Accessory Species Nest type1 Soil type2 Prey type3 at entrance closure burrows Jackson Hole, Wyoming paeificus S 1 SB + + 0 pulcher s 2 BW + + 0 crabroniformis p 3 SB — + 0-2 zebratus p 2 BW — + 0-3 bicinctus p 3 LB — + 0 Eastern Massachusetts politus s 1 SB + + 0 gibbosus p 2 SB — + 0-2 lepidus s 1 SB — + 1-5 solivagus p la BW — — 0 sanbornii p 2 LB — — 0 1 Nest type: S — a succession of simple nests with few cells. P — a more or less permanent, multicelled nest. 2 Soil type: 1 — light, fine-grained sand; la — such sand, but only on slopes. 2 — coarse, moderately friable sand. 3 — very coarse-textured sandy gravel. 3 Prey type: SB — small bees; LB — large bees; BW — both bees and wasps used in numbers. The complexity of these communities is such that they must be examined from many points of view. A behavioral adjust- ment on the part of one species, or the extinction of a species or arrival of a new one — or, indeed, the varying population sizes of each species belonging to the com- munity, even peripherally — each of these factors may have subtle effects upon many members of the community. Hence no factor, no behavior can a priori be re- garded as unworthy of study. The twin sciences of ecology and ethology are certain to become increasingly important with our growing concern with the environ- ment and our growing awareness of its complexity. Despite welcome advances in theory and the promise of computerization, it appears that the need is greater than ever for what is sometimes called "old fashioned natural history." ANNOTATED LIST OF WASPS OF JACKSON HOLE The following is a list of wasps collected in the areas of study, with a brief summary of their nesting behavior, if known. Ar- rangement of families and genera is the conventional one, approximating that in the Second Supplement of the Synoptic Cata- log of North American Hymenoptera (Krombein and Burks, 1967). A few spe- cies taken outside the immediate area of study are included but are enclosed in brackets. The major localities were de- scribed earlier (see "Study Areas"), but the abbreviations used are repeated here for the sake of ready reference: ELK: 4 miles SW of the Elk Post Office (Map 1); HHS: Huckleberry Hot Springs, Teton National Forest; J AC: 5 miles north of Jackson; MOR: Moran, near Jackson Hole Re- search Station ( Map 1 ) ; MOR-A: Largelv restricted to alluvial sand; MOR-B: Chiefly in bare places in lower outwash plain; PCR: Pilgrim Creek. 4 to 7 miles NE of Moran. Records of adults feeding on the nectar of flowers are also included, and again several abbreviations have been employed: Wasps of Jackson Hole ♦ Evans 475 Eri: Eriogonum spp. (wild buckwheat); Per: Perideridia gairdneri (yampa); Sol: Solidago spp. ( goldenrod ) ; Several species were taken in Malaise traps set up in MOR-A1, but not elsewhere. These are identified simply by the word trap. The insect fauna of Jackson Hole is by no means thoroughly known, and in a num- ber of cases I have been able to identify wasps only to genus. Some species identi- fications may prove incorrect as the system- atics of Hymenoptera improves. However, all material collected has been placed in the collections of the Museum of Compar- ative Zoology, and specimens with asso- ciated behavioral data are cross-referenced by number to my field notes, also on per- manent file at the Museum. Thus, if prob- lems arise regarding the identity of any of the species studied, it should be possible for specialists to re-examine this material. I identified many of the wasps myself, but received valuable help from several persons, especially from R. M. Bohart ( Chrysididae, Eumenidae, and certain Sphecidae), A. S. Menke (Ammophila), F. D. Parker (Astatinae and Leptochilus) , F. E. Kurczewski (Tachysphex) , and K. V. Krombein (various Sphecidae). Most of the prey were identified by various special- ists located at the U. S. National Museum, but H. W. Levi identified the spiders, and H. J. Reinhard and F. C. Thompson as- sisted with the flies. P. H. Timberlake undertook the arduous task of identifying the many bees taken as prey by species of Philanthus. Family DRYINIDAE These small wasps are parasites of leaf- hoppers and plant hoppers; they make no nest but leave the prey on vegetation, where it recovers from paralysis and con- tinues feeding. The American species have been little studied, and species identifi- cation is nearly impossible. Chalcogonatopus sp. MOR, July 28-Aug. 15, 1 9,2 $ S. Anteon sp. MOR, July, 1 6 in trap. Aphelopus comesi Fenton. MOR, Aug., 1 $ in trap. Family CLEPTIDAE Cleptids have been reared from sawfly larvae. Only one species has been taken in Jackson Hole. Cleptes provancheri Aaron. MOR, July 11-Aug. 16, 5 9 9,1 6, 9 on Per, one 9 in trap; PCR, Aug. 2, 1 9,1 6. Family CHRYSIDIDAE (Cuckoo wasps) Cuckoo wasps are parasites of various wasps and bees, usually entering the nest and laying their egg while the host is absent. A good recent account of several species is provided by Krombein ( 1967 ) . Although cuckoo wasps are abundant in Jackson Hole both in individuals and in numbers of species (24), I obtained no important new information regarding their biology. Omalus aeneus (Fabricius). MOR-B, July 6-30, 1 9, 2 6 6; note no. 1981: 1 9 taken from a nest of Philanthus pulcher, as prey. This wasp is known to parasitize species of Passaloecus, Pemphredon, and Stigmus nesting in cavities in wood. Omalus plicatus (Aaron). MOR-B, Aug. 1, 1 9. Omalus purpuratus (Provancher). MOR- A, July, 2 9 9 , in trap. Omalus speculum (Say). MOR, July 28- Aug. 1,1 9,1 $. Omalus variatus ( Aaron ) . MOR, July 19- Aug. 13, 1 9,2 $ 6 ; ELK, July 10, 1 9 . Elampus viridicyaneus Norton. MOR, July 11-Aug. 3, 5 9 9; JAC, July 15, 2 9 9. Note no. 2110B; 1 6, MOR-B3, taken from nest of Philanthus pulcher as prey. 476 Bulletin Museum of Comparative Zoology, Vol. 140, No. 7 Holopyga hora Aaron. MOR, July 16-30, 2 5 2. Holopvga ventralis (Say). MOR-B, July 11-Aug. 11, 2 ?9,4'ii, 2<5 on Sol, Per; ELK, Aug. 4. 1 6 ■ PCR, Aug. 2, 1 2. Note no. 1981: 1 9, MOR-B3, taken from nest of Philanthus pulcher, as prey. Hedychridium fleteheri Bodenstein. MOR, July 11-Aug. 23, 8 5 9, 4 S S ; ELK, Aug. 4, 1 <5; HHS, July 17, 1 $. Note nos. 1981, 1984, 2118: 1 2,3 $$, MOR- B3, taken from nests of Philanthus pulclier, as prey. Note nos. 2096, 2109: 2 2 2, both taken on ground in P. pulclier nesting areas, MOR-B3, July 14, 21, 1967. These females were flying from one nest entrance to another, apparently attracted by the fresh earth at the entrances. They were not seen to enter any nests (other wasps and various bees were also nesting here ) . The species of this genus apparently attack ground-nesting wasps; H. fleteheri is reported to attack Tachysphex, while some European species are said to attack Dienoplus spp. Hedyehrum nigropilosum Mocsary. MOR, Aug. 1, 1 $. Hedyehrum parvum Aaron. MOR, July 16-Aug. 14, 1 2,3 S $ , $ on Per. Note no. 2048: 2 found in burrow of Eucer- ceris fulvipes Cresson while I was dig- ging out nest, Aug. 14, 1964, MOR-B 1. Since other species of this genus are known to attack Cerceris spp., it seems probable that this individual was in fact attacking E. fulvipes. Chrysura densa (Cresson). MOR, July 6, 1 $; Death Canyon, Aug. 15, 1 2. Chrysura pacifiea (Say). JAC, July 15, 1 2 ; MOR, July 4-16, 3 2 2, 2 S $, 1 S no. 1981: taken as prey by Philan- thus pulclier. Ceratoehrvsis eyanosoma ( Mocsary ) . MOR, Aug. 26, 1 2. Ceratoehrvsis perpulehra ( Cresson ) . MOR, July 13, 1 2; PCR, Aug. 2, 1 2. Ceratoehrysis traehypleura Bohart. MOR, July 20- Aug. 26, 11 2 2, 3 $ $ ; ELK, July 28, 1 2; HHS, Aug. 11, 1 2. Note nos. 1996, 2109, 2135, all MOR-B, July, Aug. Several females were seen landing at nest entrances and walking over soil at entrances, but not entering; nests were those of Philanthus pulcher, P. crahroni- formis, and Ammophila azteca (cuckoo wasp erroneously identified as Ceratoehry- sis perpulehra in my 1965 paper on A. azteca). This species is common in Jackson Hole and is surely a parasite of one or more species of ground-nesting wasps or bees, but I have no evidence as to which ones. C. perpulehra is reported as a para- site of Ammophila aberti, but the type spe- cies of the genus, enhuueki Cooper, is a parasite of the twig-nesting eumenid Lepto- chilus. There are two distinct color forms of female traehypleura in Jackson Hole, one blue-green, the other with a strong overlay of coppery and rose tints; the two are represented in about equal numbers. Triehrysis doriae (Gribodo). MOR, July, Aug., 3 2 2, in trap. Chrysis coerulans Fabricius. MOR, July 11-Aug. 15, 9 2 2, 1