PROCEEDINGS OF THE California Academy of Sciences Volume 44 SAN FRANCISCO PUBLISHED BY THE ACADEMY 1985-1986 PUBLICATIONS COMMITTEE Daphne G. Fautin, Scientific Editor Sheridan Warrick, Managing Editor Frank Almeda Luis F. Baptista Wojciech J. Pulawski Frank Talbot (US ISSN 0068-547X) The California Academy of Sciences Golden Gate Park San Francisco, California 94 1 1 8 PRINTED IN THE UNITED STATES OF AMERICA BY ALLEN PRESS, INC., LAWRENCE, KANSAS CONTENTS OF VOLUME 44 Pages No. 1 . XiN-Luo CHU, AND TYSON R. ROBERTS. Cosmochilus cardinalis, a new cyprinid fish from the Lancang-Jiang or Mekong River in Yunnan Province, China. Published August 29, 1 985 1-7 No. 2. McCosKER, JOHN E. Two new genera and two new species of deepwater western Atlantic worm eels (Pisces: Ophichthidae). Published August 29, 1985 9-15 No. 3. LEIPERTZ, STEVEN L. A review of the fishes of the agonid genus Xeneretmus Gilbert. Published August 29, 1 985 1 7-40 No. 4. BROWN, WALTER C., AND JOHN R. H. GIBBONS. Species of the Emoia samoensis group of lizards (Scincidae) in the Fiji Islands, with descriptions of two new species. Published January 3, 1986 41-53 No. 5 . ZULLO, VICTOR A. Quaternary barnacles from the Galapagos Islands. Published February 7, 1986 55-66 No. 6. KAVANAUGH, DAVID H. A systematic review of amphizoid beetles (Amphizoi- dae: Coleoptera) and their phylogenetic relationships to other Adephaga. Pub- lished February 7, 1 986 67-109 No. 7. DAY, ALVA G., AND REID MORAN. Acanthogilia, a new genus of Polemoniaceae from Baja California, Mexico. Published February 7, 1986 111-126 No. 8. MAHOOD, A. D., G. A. FRYXELL, AND M. MCMILLAN. The diatom genus Thalassiosira: Species from the San Francisco Bay system. Published May 6, 1986 127-156 No. 9. WASHINGTON, BETSY B. Systematic relationships and ontogeny of the sculpins Artedius, Clinocottus, and Oligocottus (Cottidae: Scorpaeniformes). Published May 6, 1986 157-223 No. 10. PARENTI, LYNNE R. Bilateral asymmetry in phallostethid fishes (Atherino- morpha) with description of a new species from Sarawak. Published May 6, 1986 225-236 No. 1 1 . ROTH, BARRY. Land mollusks (Gastropoda: Pulmonata) from early Tertiary Bozeman Group, Montana. Published May 6, 1986 237-267 No. 12. PITT, WILLIAM D., MATTHEW J. JAMES, CAROLE S. HICKMAN, JERE H. LIPPS, AND Lois ]. PITT. Late Cenozoic marine mollusks from Tuff Cones in the Galapagos Islands. Published May 6, 1986 269-282 Index to Volume 44 285-292 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 1, pp. 1-7, 5 figs. August 29, 1985 COSMOCHILUS CARDINALIS, A NEW CYPRINID FISH FROM THE LANCANG-JIANG OR MEKONG RIVER IN YUNNAN PROVINCE, CHINA By Xin-Luo Chu Kunming Institute of Zoology ofAcademia Sinica, Kunming, Yunnan Province, The People's Republic of China and Tyson R. Roberts California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: Cosmochilus cardinalis is a large, deep-bodied cyprinid fish recently discovered in southern Yunnan Province in the mainstream of the Lancang-jiang or Mekong River. It is distinguished from all other cyprinids in China and Southeast Asia in having all of its fins bright red; it is further distinguished from the two previously known species of Cosmochilus by having longer barbels, reduced labial papillae, a concave nape, more numerous scales (also more numerous vertebrae?) and nine instead of eight branched dorsal fin rays. INTRODUCTION In the present paper we describe a distinctive new species of large cyprinid fish from the Me- kong River or Lancang-jiang of China. The type- specimens were collected during an ongoing, long- term ichthyological survey of Yunnan Province undertaken by the Kunming Institute of Zoology of Academia Sinica. The Lancang-jiang or Mekong is the largest river in Southeast Asia and probably has the rich- est ichthyofauna of any river in Asia. It is some 4300 km long and arises at an elevation of 5090 m below an enormous glacier on the northern slopes of the Dza-Nag-Lung-Mung or Tanglha Range of the Tibetan highlands of China's Tsing- hai Province. In Tibet it is known as the Lan- cang-jiang or Dza Chu. It leaves Tibet and enters Yunnan Province at an elevation of about 2800 m, assuming a generally southerly course of near- ly 900 km through mountainous and hilly coun- try of Yunnan. In its upper reaches in Tibet and Yunnan it flows through canyonlike gorges be- tween and parallel to the Salween and Yangtze. In lower Yunnan, where the new species of cyp- rinid was collected, the elevation is only about 500 m. Here the river has a relatively gentle gra- dient and moderate current and is about 1 50 m wide during the dry season. The bottom is rocky in places but predominantly muddy. The local fishermen know the new species as bia Hang or hong chi ("red-finned fish"). We have identified it as an undescribed Cosmochilus. i PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 90C 100° \ 110° \ 120° \ 130° \ -20( * cardinalis A falcifer • harmandi FIGURE 1. Geographical distribution of the species of Cosmochilus. Cosmochilus Sauvage, 1878 Cosmochilus SAUVAGE, 1878:240 (type-species Cosmochilus harmandi SAUVAGE, by monotypy). DIAGNOSIS.— Large, deep-bodied and laterally compressed labeoine cyprinids; dorsal fin large and falcate, with 4 simple and 8-9 branched rays; last simple dorsal fin ray greatly enlarged, its pos- terior border more or less strongly serrated for its entire length; anal fin relatively small, with 3 simple and 5-6 branched rays; head relatively small, compressed; snout truncate, without en- larged tubercles or pores; mouth small and in- ferior, its opening transverse; rostral and max- illary barbels large and relatively elongate; lips moderately thick, entirely covered with large, contiguous papillae; horny jaw sheaths trans- verse, moderately thick but with relatively weak cutting edge; gill rakers fleshy, relatively unspe- cialized, 1 5-1 8 on first gill arch; pharyngeal teeth triserial, morphologically generalized for Cy- prinidae, usually 1,3,5/5,3,1 or 2,3,5/5,3,2; lat- eral line almost perfectly straight; each lateral line tubule with a short ventroposterior branch terminating in a small pore on exposed portion of posterior shield; scales in lateral line series 35- 48; circumpeduncular scales 16-18; scales ob- long, with relatively huge posterior shields; radii of posterior shield strongly convergent; radii of anterior shield frequently conjoined or bifurcate; vertebrate 35-43. In addition to C. harmandi from the Chao Phrya and Mekong rivers, the genus includes C. falcifer Regan, 1 906 from the Baram and Rejang rivers in Sarawak and Kapuas in Kalimantan Barat (western Borneo). The distribution of the species is shown in Figure 1 . Cosmochilus har- mandi is known to undertake lengthy spawning CHU AND ROBERTS: COSMOCHILUS CARDINALIS, A NEW CYPRINID FISH FIGURE 2. Cosmochilus cardinalis, holotype (KIZ 7351 13, 177 mm). migrations. This is suspected in C. falcifer, and thus may also be characteristic of C. cardinalis. Mekong localities for C. harmandi are from Rainboth et al. 1976. Cosmochilus cardinalis new species (Figures 2-5) HOLOTYPE. — KIZ (Kunming Institute of Zoology) 735113, 177 mm (standard length), mainstream of Lancang-jiang near Jinghong, southern Yunnan Province, lat. 21°50'N, long. 100°55'E, gill net, May 1973. PARATYPES.-KIZ 734079-80, 734082, 735107, 735109-1 12, 735025, 735030, 735075, 735160, 12: 165-326 mm, same locality and collecting method as holotype, April-May 1973.' DIAGNOSIS.— A large, deep-bodied, and later- ally compressed Cosmochilus, attaining at least 400 mm standard length. Rostral and maxillary barbels relatively long and thick, dorsal fin high with a falcate margin and elevated base; dorsal fin rays iv9-l/2, last simple ray an elongate stiff spine strongly serrate posteriorly; anal fin rays iii6-l/2. It differs from all or almost all other cyprinids in China and Southeast Asia in having all of the fins including the pectoral and pelvic 1 KIZ 735 111,251 mm, has been transferred to the Ichthy- ology Department collection of the California Academy of Sci- ences and is now CAS 55592. and both caudal lobes entirely bright cardinal red in life. Body dusky dorsally, silvery or whitish on sides and abdomen. Opercle golden. Scales in lateral series 46-48, between dorsal fin and lat- eral line 9-10, between lateral line and pelvic fin origin 5, circumpeduncular 16-18. Vertebrae 24+19 = 43 (holotype). Head relatively small and laterally com- pressed, its length 4.0-4.5; dorsal profile of head to nape moderately sloped, then abruptly steeper at nape until dorsal fin origin. Snout 2.7-3.2 in head, eye 4. 1-5.3 in head, interorbital width 2.4- 2.7 in head. Eye with a narrow gelatinous rim or hyaline eyelid. Barbels thick and relatively long; anterior or rostral barbel extending posteriorly almost to below posterior border of eye, its length 1.8-2.2 in head; posterior or maxillary barbel somewhat longer, extending posteriorly almost to pectoral fin, its length 1.4-1.7 in head. Mouth subinferior and moderately wide, extending pos- teriorly to directly below anterior margin of eye or slightly farther. Rostral cap well developed with deeply incised rostral groove complete be- tween rostral barbels of either side. Sublacrimal groove also well developed and deeply incised, extending from rostral barbel to posterior end of jaws. Lips moderately well developed; upper lip PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 mm FIGURE 3. Cosmochilus cardinalis, holotype. First gill arch (lateral view; bony supports of gill rakers in black). well defined, entirely separate from rostral cap and upper horny jaw sheath; lower lip with oral margin separated from horny jaw sheath only by a shallow groove, and with posterior margin well defined laterally but entirely interrupted for transverse portion of lower jaw; upper and lower horny jaw sheaths well developed, with a broadly rounded surface and very weakly developed transverse grooves or sulci. Anterior margin of lower jaw truncate. Rostral cap and horny jaw sheaths relatively smooth; lips and oral epithelium (including gular flap) covered with large, close-set or contiguous but low-lying papillae (probably unculiferous) comparable in distribution and basic morphol- ogy to the greatly enlarged and elevated contig- uous papillae characteristic of the other two species of Cosmochilus. Gill openings relatively broad, isthmus nar- row; upper portion of gill cover with posterior margin very slightly concave. First gill arch (Fig. 3) with 5 + 1 1 or 1 2 = 1 6 or 1 7 gill rakers on anterolateral margin and 0+11 gill rakers on posterodorsal margin in holotype; paratypes with 17-19 anterolateral gill rakers on first gill arch. Gill rakers all relatively short and fleshy, those on lower limb of gill arches more or less trian- gular in shape with broad bases; posterior gill rakers similar in shape to anterior rakers but relatively somewhat larger. Pharyngeal bones (Fig. 4) strongly arched; dorsal edentulous limb and dorsal half of toothbearing portion thick and evenly curved; external ala moderately broad; edentulous lower limb and ventral half of tooth- bearing portion below external ala with concave lateral margin. Entire medial surface of lower limbs of either pharyngeal bone forming a broad symphysis. Pharyngeal teeth basically uncinate, 2,3,5/4,3,2 in holotype, those of principal row compressed. Scales (Fig. 5) large, slightly longer than high, and relatively numerous. Anterior margin rela- tively straight but with a well defined convex median projection; posterior margin broadly CHU AND ROBERTS: COSMOCHILUS CARDINALIS, A NEW CYPRINID FISH 5 mm FIGURE 4. Cosmochilus cardinalis, holotype. Pharyngeal jaws (dorsal view). rounded; radii most numerous and longest on posterior field, least numerous and shortest on lateral fields; radii near central portion of pos- terior field strongly convergent. Lateral line com- plete and nearly straight. Scales of lateral line with straight primary tubules and a single small pore arising from a short ventroposteriorly di- rected secondary tubule originating from ante- rior half of primary tubule. Scales in lateral line series 46-48, 9-10 above and 5 below lateral line; predorsal scales 24-26, circumpeduncular 16-18. Two or three scale rows extend on caudal fin base posterior to last pored scale of lateral line series. Scales between vent and anal fin or- igin 2. Circumferential scales 35-37 (18-20 dor- sal + 2 lateral line + 1 5 ventral). Dorsal fin origin near middle of body, some- what closer to snout tip than to caudal fin base. Anal fin origin distinctly posterior to vertical through base of last dorsal fin ray. Pelvic fin or- igin well in advance of vertical through dorsal fin origin. Base of dorsal fin with well developed, strongly convex scale sheath comprising three somewhat irregular rows of large scales (of which the middle row is somewhat smaller); anal fin scale sheath only slightly convex, with one or two rows of scales. Pelvic fin with two moder- ately elongate axillary scales. Dorsal fin spine length 3.0-3.7, with 20 serrae in holotype. Anal fin much smaller than dorsal fin, with last simple and first two branched rays slightly prolonged to form a lobe. Pectoral and pelvic fins almost equal in shape and size, pectoral fin extending poste- riorly almost to pelvic fin origin and pelvic fin reaching or almost reaching vent; pectoral fin length 3.8-4.3, pelvic 4.0-4.3. Pectoral fin rays usually 15, pelvic 9. Caudal fin deeply forked, upper and lower lobes nearly equal in length and shape, with slightly pointed tips. Abdomen anterior to and between pelvic fins PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 5 mm FIGURE 5. Cosmochilus cardinalis, holotype. Scales. Below, 30th scale in lateral line scale series. Above, scale two scale rows about 30th scale in lateral line scale series. Dashed lines indicate areas of overlap by scales neighboring exposed portion of posterior shield. somewhat flattened from side to side; abdomen posterior to pelvic fins rounded or very slightly (almost imperceptibly) carinate. Body depth 2.4. Caudal peduncle laterally compressed and mod- erately deep, its length 6.1-7.9 and depth 6.4- 7.2. Dorsal surface of head and lateral surface of lacrimal region more or less uniformly covered with fine, widely spaced granular tubercles; stronger tuberculation or tubercles on fins not observed in specimens of either sex. Swim blad- der with two chambers. Intestine about three times as long as body. COMPARISON WITH OTHER SPECIES OF COSMO- cH/Lus.—The basically similar morphology of the scales, scaly fin sheaths, fleshy gill rakers, pharyngeal teeth, and papillose lips of C. car- dinalis, C. harmandi, and, so far as known, C. falcifer, leads us to conclude that the three are correctly placed together within Cosmochilus. Our new species differs from C. harmandi and C. falcifer in having more numerous scales (44- 48 vs. 36-39 in lateral series), three instead of two scale rows in scaly sheath on dorsal fin base, 9 instead of 8 branched dorsal fin rays, dorsal profile strongly concave at nape (vs. relatively evenly sloped), more elongate barbels, labial pa- pillae much reduced in size, and all fins bright red. It additionally differs from C. harmandi in having 6 instead of 5 branched anal fin rays, a less pointed snout viewed from above or from the side, and 43 instead of only 35 vertebrae (number of vertebrae unknown in C. falcifer). Cosmochilus cardinalis also differs from C. fal- cifer in having the fourth simple dorsal fin ray relatively less elongate, very straight, and with large, strong serrae on its posterior margin. In C. falcifer this ray is exceptionally elongate (some- times extending posteriorly to caudal fin base when adpressed), very strongly curved, and with very weak serrae on its posterior margin. Life color of C. harmandi is recorded as back rich pale blue, dorsal and caudal fins black-edged; in some specimens anal fin with black tip (Smith 1945:132). We note that the dorsal fin may also be black-tipped. Life color of C. falcifer has not been reported previously. A fresh 3 1 6-mm spec- imen caught in the Kapuas and photographed in the market at Sintang by the junior author had overall color white or milk white, especially ven- trolaterally and ventrally; dorsolaterally and dor- sally distinctly brownish or violaceous brown; posterior margins of scales, especially on upper parts of body, with broad dark margins; dorsal surface of head faintly yellowish and entire gill cover distinctly yellow; iris and ventral portion of head milk white; entire dorsal fin rosy pink or faintly orangish except black at tip; pectoral fin CHU AND ROBERTS: COSMOCHILUS CARDINALIS, A NEW CYPRINID FISH white, pelvic white or pinkish; anal and caudal fins dusky, caudal very dark, its posterior margin almost black. In conclusion, C. harmandi and C.falciferseem to be much more similar to each other than either of them is to C. cardinalis. RELATIONSHIPS OF COSMOCHILUS Sauvage ( 1 878:240) stated that Cosmochilus is related to Labeo but did not discuss this assess- ment. The pattern of strongly convergent radii on the posterior shield and conjoined or bifur- cating radii on the anterior shield of the scales of Cosmochilus is seen in relatively few Chinese cyprinids (Chu 1935). Those in which this pat- tern is most clearly present are species of Osteo- chilus, Sinilabeo, and Garra (Chu 1935, pis. 15- 16). These genera are referable to the cyprinid subfamily Labeoinae. We have observed scales with a similar pattern of radii in various non- Chinese Labeoinae including Morulius chryso- phekadion and various African and Asian species currently assigned to Labeo, but not in genera belonging to other subfamilies. This suggests that Cosmochilus might indeed belong in Labeoinae. On the other hand, the general morphology of Cosmochilus, including the relatively simple mouth parts; deep, laterally compressed body; and elongate, serrate dorsal fin spine suggest re- lationship to a group of Southeast Asian barbels including Cydocheilichthys. A serrate dorsal fin spine is unknown in the Labeoinae. ACKNOWLEDGMENTS We wish to express our gratitude to Walter J. Rainboth, who first noted the similarity of the new species to Cosmochilus, Susan Middleton for photographing the holotype, George Zorzi and David Catania for preparation of radio- graphs, and Reeve M. Bailey and Douglas W. Nelson for loaning specimens of C. harmandi. The manuscript was typed by Francis Bertetta. LITERATURE CITED CHU, Y. T. 1935. Comparative studies on the scales and on the pharyngeals and their teeth in Chinese cyprinids, with particular reference to taxonomy and evolution. Biol. Bull. St. John's Univ. 2, x + 226 pp., 30 pis. RAINBOTH, W. J., K. F. LAGLER, AND S. SONTIRAT. 1976. Maps of freshwater fish distribution in the lower Mekong basin. Mekong Secretariat Working Document 3 1 , xv + 406 pp. SAUVAGE, H. E. 1878. Note sur quelques poissons d'especes nouvelles provenant des eaux douces de 1'Indo-Chine. Bull. Soc. philomath. Paris 7(2):233-242. SMITH, H. M. 1945. The freshwater fishes of Siam or Thai- land. Bull. U.S. Nat. Mus. 188, xii + 622 pp., 9 pis. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 2, pp. 9-15, 6 figs. August 29, 1985 TWO NEW GENERA AND TWO NEW SPECIES OF DEEPWATER WESTERN ATLANTIC WORM EELS (PISCES: OPHICHTHIDAE) By John E. McCosker California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: Two new genera and two new species of Atlantic worm eels, family Ophichthidae, subfamily M yrophinae, tribe M yrophini, are described and illustrated. Mixomyrophis pusillipinna, gen. no v., sp. no v ., trawled from deepwater off the Lesser Antilles, is an elongate species with uniserial conical teeth, a labial posterior nostril, a minute pectoral fin, and 178 vertebrae. Asarcenchelys longimanus, gen. nov., sp. nov., collected off Belem Brazil, is an elongate species with biserial conical teeth, a labial posterior nostril, well- developed pectoral fins, and 149 vertebrae. Osteological characteristics of the new genera are described from radiographs and compared with those of related myrophines. INTRODUCTION The ophichthid worm eels of the subfamily Myrophinae (sensu McCosker 1977) occupy a variety of sand and mud habitats as well as the midwater environment, ranging from the shal- low intertidal to depths of 400 fathoms or more. The shallow- water species of the genera Myro- phis, Muraenichthys, and Ahlia are common in collections and are probably abundant within their milieu. The deeper water species are rare, being difficult to trawl or dredge, and are often known from but a single specimen. While preparing the ophichthid eel section of The Fishes of the Western North Atlantic (FWNA), the late James E. BOhlke discovered a single specimen of a new species of myrophine collected by trawl off Anguilla, Lesser Antilles. In taking over the completion of the FWNA proj- ect, I discovered another new species of deep- water myrophine, from Brazil, which is also ge- nerically distinct. I intend to make these taxonomic names available for the FWNA vol- ume and to describe the significant osteological characters that are visible by radiographic ex- amination. It is my hope that subsequent spec- imens will be discovered, which will allow a more thorough osteological examination and compar- ison with other myrophine genera. MATERIALS AND METHODS Measurements are straight-line, made either with a 300-mm ruler with 0.5-mm gradations (for total length, trunk length, and tail length) and recorded to the nearest 0.5 mm, or with dial calipers (all other measurements) and recorded to the nearest 0.1 mm. Body length comprises head and trunk lengths. Head length was mea- sured from the snout tip to the posterodorsal margin of the gill opening; trunk length was taken from the end of the head to mid-anus; maximum body depth did not include the median fins. Ver- tebral counts (which included the hypural) were [9] 10 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 2 FIGURE 1. Holotype of Mixomyrophis pusillipinna McCosker, sp. nov., ANSP 152305, 407 mm TL. Inset: Head of holotype of Mixomyrophis pusillipinna McCosker, sp. nov. taken from radiographs. Stained and cleared gill arches were prepared using the Taylor (1967) trypsin technique. Institutional abbreviations of material examined are explained in the acknowl- edgments section of this paper. Mixomyrophis McCosker, gen. nov. TYPE SPECIES.— Mixomyrophis pusillipinna McCosker, sp. nov. DIAGNOSIS.— An elongate myrophine, tribe Myrophini, with tail longer than head and trunk, laterally compressed, particularly posteriorly; snout subconical, broad from above, not grooved ventrally; anterior nostril tubular, posterior nos- tril on outer edge of lip and covered by a flap; dorsal fin origin in mid-trunk; pectoral fin a mi- nute flap in posterodorsal corner of upper gill opening; eye large, behind middle of jaw; third preopercular pore present; head and lips smooth, without cirri or lappets; teeth of jaws and vomer small, conical, uniserial, and close set, with slightly retrorse tips; gill arches well developed for a myrophine, first basibranchial ossified, up- per pharyngeal toothplates fused; neurocranium stout, slightly sloping posteriorly; suspensorium anteriorly inclined; pterygoid stout, not bracing maxilla; maxillae elongate, tapering posteriorly; opercular series apparently moderately devel- oped; pectoral girdle reduced to a slender clei- thrum and supracleithrum; epipleural ribs on all precaudal vertebrae; caudal transverse processes apparently absent; caudal vertebrae more nu- merous than precaudal. Other characteristics those of single species. ETYMOLOGY.— From the Greek /iu|ts, mixis, a mixing, and Myrophis (masculine), a genus of ophichthid eel. Named in reference to the com- bination of myrophine characters that this eel possesses. Mixomyrophis pusillipinna McCosker, sp. nov. (Figures 1, 2, 6b) HOLOTYPE. -ANSP 152305 (originally UMML 30290), 407 mm, a female with ripening ovaries, captured off Anguilla, Lesser Antilles ( 1 8°26.4'N, 63°1 2.6' W to 1 8°28'N, 63°1 1 . 1' W), by 10-m otter trawl, between 393-451 m depth, by the RV Pillsbury, sta. 984, on 22 July 1969. COUNTS AND MEASUREMENTS (IN MM).— Total length 407; head length 36.5; trunk length 115.5; tail length 255; body depth at gill openings 9.0; body width at gill openings 7.1; body depth at anus 8.8; body width at anus 6.0; gill opening 1.5; snout tip to origin of dorsal fin 94; left pectoral fin length 1.0; snout length 7.9; upper jaw length 12.1; eye diameter 2.9; fleshy interorbital distance 4.2 Total vertebrae 178; predorsal ver- tebrae 33; preanal vertebrae 57. DESCRIPTION.— Body elongate, its depth 45 in total length (TL), laterally compressed in tail re- gion. Head and trunk 2.7 and head 1 1.2 in TL. Snout subconical, broad as seen from above; lower jaw included, its tip reaches the anterior McCOSKER: TWO NEW SPECIES OF ATLANTIC WORM EELS 11 nostril bases. Anterior nostrils tubular, directed ventrally, their anterior edge looped upward; posterior nostril at outer edge of lip, covered by a flap. Eye large, its anterior edge behind midpoint of upper jaw. Gill opening mid-lateral, a constricted open- ing. Median fins low, lying partially within a groove, but elevated above last 20 vertebrae, meeting each other and extending beyond caudal tip. Dorsal fin arises above posterior trunk region. Pectoral fin minute. Head pores developed. Single temporal and interorbital pores. Six pores along left mandible; 5 along right. Two pores between anterior and posterior nostrils. Four supraorbital pores. Three preopercular pores. Left lateral line pores ca. 1 49; 9 above branchial basket; 57 before anus. Teeth small, conical, nearly uniform in size. An intermaxillary chevron of 8 teeth, followed by 2 on each side, closely adjoining ca. 20 uni- serial vomerine teeth and 25 maxillary teeth. Ap- proximately 30 uniserial mandibular teeth, with a secondary pair at symphysis. Gill arches removed, stained and cleared. Ba- sibranchial 1 ossified, basibranchials 2-4 absent. Hypobranchials 1 and 2 ossified; hypobranchial 3 cartilaginous. Ceratobranchials 1-4 ossified; ceratobranchial 5 absent. Infrapharyngobran- chials 2 and 3 ossified. Lower tooth plate small, with 2 rows of conical teeth, medial row largest. Upper pharyngeal tooth plate fused, subrectangular, with 4-5 rows of conical teeth, medial row largest. Body color in isopropyl alcohol yellow on head, chin, tail, and dorsal surface of trunk. Throat and belly whitish. Finely peppered throughout body and tail with small brown specks. Peritoneum black. ETYMOLOGY.— From the Latin pusillus, puny or insignificant, and pinna, fin, to be treated as a noun in apposition. REMARKS.— Mixomyrophis is separable from all other myrophines by the combination of its minute pectoral fin, elongate body, and posterior nostril located within the outer lip. Mixomyro- phis appears most similar to the elongate species of Pseudomyrophis, which differ by having the posterior nostril before the eye, more extreme body elongation, a reduced and rounded neu- rocranium (cf. Fig. 6b and 6d), and reduced gill FIGURE 2. Dentition of holotype of Mixomyrophis pusil- lipinna McCosker, sp. nov., ANSP 152305. arch components (Bohlke 1960; McCosker 1977). The nostril condition of Pseudomyrophis appears to be a slight posterodorsal translocation of the opening from its location along the lip (although it is difficult to interpret which state might be the primitive condition), and the other characters seem to be advanced specializations. The nearly bulbous snout, large eye, and body elongation of Mixomyrophis are conditions shared by other deepwater myrophines such as the species of Pseudomyrophis, Neenchelys, Asarcenchelys lon- gimanus, and Muraenichthys puhioilo McCosker (1979). 12 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 2 FIGURE 3. Reconstructed appearance of holotype of Asarcenchelys longimanus McCosker, sp. nov., MNHN 1968-215, 277 mm TL. The actual specimen is intact, but badly torn in the anterior trunk region. Asarcenchelys McCosker, gen. nov. TYPE SPECIES.— Asarcenchelys longimanus McCosker, sp. nov. DIAGNOSIS. —A very elongate myrophine, tribe Myrophini, with tail longer than head and trunk, laterally compressed throughout trunk and tail; snout subconical, tumid, not grooved ventrally; anterior nostril tubular; posterior nostril on outer edge of lip and covered by a flap that is incised posteriorly; dorsal fin origin in anterior trunk region; anal fin elevated; pectoral fin lanceolate, well developed, slightly longer than snout; eye large, behind middle of jaw; third preopercular pore present; head and lips smooth, without cirri or lappets; teeth of jaws and vomer large, not close set, conical and slightly recurved; teeth bi- serial anteriorly in jaws and vomer, outer row smaller; gill arches appear to be well developed for a myrophin; neurocranium stout, truncate posteriorly; supraoccipital crest developed; sus- pensorium posteriorly inclined; maxillae taper posteriorly; pectoral girdle reduced to stout cleithrum and thin supracleithrum; epipleural ribs present only on anterior trunk vertebrae; caudal temporal processes apparently absent; caudal vertebrae more numerous than precaudal. Other characteristics those of single species. ETYMOLOGY.— From the Greek aaapnos, asar- kos, lean, and evx&v$, enchelys, eel (treated as feminine according to Opinion 9 1 5 of the Bul- letin of Zoological Nomenclature, 1970), in ref- erence to its emaciated appearance. Asarcenchelys longimanus McCosker, sp. nov. (Figures 3-5, 6c) HOLOTYPE. -MNHN 1968-215, 277 mm, sex undeter- mined, captured near Bel6m, Brazil, at 55 m depth by P. Four- manoir, September 1966. PARATYPE.— MNHN B. 2994, 147 mm, sex undetermined, collected with the holotype. COUNTS AND MEASUREMENTS (IN MM). — Data for the paratype parenthetically follow those of the holotype. Total length 277 (147); head length 27 (18); trunk length 78 (46); tail length 172 (83); body depth behind gill openings 3.8 (~2); body width behind gill openings 3.6 (1.4); body depth at anus ~ 1.5 (~2); FIGURE 4. Head of holotype of Asarcenchelys longimanus McCosker, sp. nov., MNHN 1968-215. McCOSKER: TWO NEW SPECIES OF ATLANTIC WORM EELS 13 body width at anus 1.5 (1.2); gill opening 1.4 (0.9); snout tip to dorsal fin origin 55 (34); left pectoral fin length 4.0 (3.3); snout length 3.8 (3.7); upper jaw length 6.6 (5.8); eye diameter 1.4 (1.3); fleshy interorbital distance 1.6 (1.3). Total vertebrae 148 (1 3 1 , tail incomplete); predorsal vertebrae 27 (27); preanal vertebrae 53 (55). DESCRIPTION.— Body very elongate, its depth 72.9-73.5 in TL, laterally compressed behind head. Head and trunk 2.3-2.6, and head 8.2- 10.3 in TL. Snout subconical, bulbous; lower jaw includ- ed, its tip reaches to front of anterior nostril bas- es, leaving several intermaxillary teeth exposed. Anterior nostrils tubular, directed ventrally; pos- terior nostril at outer edge of lip, covered by a flap whose posterior edge is incised. Eye large, anterior edge of orbit above middle of upper jaw. Gill openings mid-lateral, not as constricted as those of most myrophines, about equal in length to isthmus. Dorsal fin low, arising in anterior trunk region. Anal fin elevated. Median fins expanded in pos- terior tail region, extended beyond caudal tip. Pectoral fin lanceolate, broad based, well devel- oped for a myrophine. Head pores developed, much more apparent than those of lateral line. Single temporal and interorbital pores. Five pores along mandible, widely spaced posteriorly. Two pores between anterior and posterior nostrils. Four supraorbital pores. Three preopercular pores. Lateral line pores difficult to discern; 1 4 above branchial basket. Teeth conical, fairly large for a myrophine, not close set, nearly uniform in size, recurved. An intermaxillary chevron of 6 teeth, visible when mouth is closed, followed by closely abutting vo- merine dentition consisting of 3-4 pairs of teeth and a uniserial row of 1 3 teeth. Maxillary teeth biserial anteriorly, with an inner row of 6 teeth and an outer row of 22-24 smaller teeth. Lower jaw biserial anteriorly, with an inner row of 5 teeth and an outer row of 23-25 smaller teeth. Gill arches, as viewed from radiograph, appear to be myrophin-like and not reduced. First ba- sibranchial is ossified, others appear to be car- tilaginous or absent. Hypobranchials 1 and 2 os- sified; hypobranchial 3 appears cartilaginous or absent. Ceratobranchials 1-4 ossified; cerato- branchial 5 not apparent. Upper and lower tooth plates appear to have 2-3 rows of conical teeth; upper plate appears to be fused. FIGURE 5. Dentition of holotype of Asarcenchelys longi- manus McCosker, sp. nov., MNHN 1968-215. Body coloration in isopropyl alcohol cream to white, numerous fine, chocolate-brown spots overlying snout, dorsal surface, and area behind eye. All fins transparent. Peritoneum light col- ored. ETYMOLOGY.— From the Latin longus, long, and manus, hand, to be treated as a noun in apposition. Named with reference to the elongate pectoral fins. REMARKS.— This new myrophine is separable from all related ophichthids by a combination of internal and external morphological charac- 14 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 2 B FIGURE 6. Enlarged radiographs of neurocrania of selected myrophins: A) Myrophis vafer, CAS 17823, 220 mm TL. B) Mixomyrophis pusillipinna, ANSP 1 52305, 407 mm TL. Gill arches have been removed. C) Asarcenchelys longimanus, MNHN 1968-215, 277 mm TL. D) Pseudomyrophis micropinna, CAS 50978, 109 mm TL. ters. Particularly significant are its well-devel- oped pectoral fins, posterior nostril located on the edge of the lip, elongate body and tail, and elongate dentition. The body elongation, anal fin development, nostril location, and snout shape are not unlike those of certain species of Neenchelys and Pseudomyrophis (cf. McCosker 1982; Smith and BOhlke 1983), and are probably McCOSKER: TWO NEW SPECIES OF ATLANTIC WORM EELS 15 adaptive for living in deepwater soft benthic hab- itats. Its close affinities, however, lie with the species ofMyrophis, which share with it the de- rived character state of having lost epipleural ribs beyond the fifteenth vertebra, a condition shared as well with Ahlia egmontis. Asarcenche- lys longimanus also shares with the species of Myrophis the primitive states of neurocranial shape (Fig. 6a, 6c), gill arch condition, and pec- toral fin development. Species of Pseudomyro- phis and Neenchelys are further specialized and separable from the "Myrophis group" in having a much-reduced neurocranium, and posterior nostrils before the eye and lacking a flap (Mc- Cosker 1977, 1982). It should be noted that both specimens of A. longimanus are damaged and thereby the total length measurement of each specimen may be in error by a few percent. The paratype is intact, but the radiograph indicates that the tail has probably been severed and regrown. The speci- men has 17 fewer vertebrae than the holotype. During capture the holotype was broken behind vertebra 15 and is twisted in preservative. The head remains attached by the skin to the trunk region and is sufficiently intact to allow precise measurements to be taken and characters to be analyzed. ACKNOWLEDGMENTS The curators of several institutions generously supplied information and specimens used in this study. I thank the ichthyological staffs of: Acad- emy of Natural Sciences of Philadelphia (ANSP), British Museum (Natural History) (BMNH), California Academy of Sciences (CAS), Paris Museum of Natural History (MNHN), Scripps Institution of Oceanography (SIO), Uni- versity of Miami Marine Laboratory (UMML), and the United States National Museum of Nat- ural History (USNM). Special assistance was supplied by Eugenia B. B6hlke (ANSP), Marie- Louise Bauchot (MNHN), and C. Richard Ro- bins (UMML). I also wish to thank: Mary Fuges and Amy Pertschuk for the preparation of illus- trations, Susan Middleton for photographic as- sistance, Lillian J. Dempster and W. I. Follett for nomenclatural assistance, and Eric Anderson for critically reading a draft of this manuscript. LITERATURE CITED BOHLKE, J. E. 1 960. A new ophichthid eel of the genus Pseu- domyrophis from the Gulf of Mexico. Notul. Nat., No. 329: 1-8. McCosKER, J. E. 1977. The osteology, classification and re- lationships of the eel family Ophichthidae. Proc. Calif. Acad. Sci., Ser. 4, 41:1-123. . 1 979. The snake eels (Pisces, Ophichthidae) of the Hawaiian Islands, with the description of two new species. Proc. Calif. Acad. Sci. 42(2):57-67. 1982. A new genus and two new species of remark- able Pacific worm eels (Ophichthidae, subfamily Myrophi- nae). Proc. Calif. Acad. Sci. 43(5):59-66. SMITH, D. G., AND J. E. BOHLKE. 1983. Neenchelys retropin- na: a new worm eel (Pisces: Ophichthidae) from the Indian Ocean. Proc. Acad. Nat. Sci. Philadelphia 135:80-84. TAYLOR, W. R. 1967. An enzyme method of clearing and staining small vertebrates. Proc. U.S. Natl. Mus. 122(3596): 1-17. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 3, pp. 17-40, 18 figs., 6 tables. August 29, 1985 A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT By Steven L. Leipertz School of Fisheries WH-10, University of Washington, Seattle, Washington 98195 ABSTRACT: The agonid genus Xeneretmus is reviewed and found to be composed of two subgenera: Xeno- pyxis, containing A . latifrons, X. leiops, and X. ritteri; and Xeneretmus, containing only A. triacanthus. The osteology of the type species of the genus, A. triacanthus, is described, illustrated, and compared with the other members of the genus, as well as to members of the agonid genera Agonus, Hypsagonus, and Podothecus. On the basis of a comparison with IS other agonid taxa, the subgenera Xeneretmus and Xenopyxis are demonstrated to be monophyletic. A key is provided, along with synonymies, diagnoses, and descriptions for the genus, the subgenera, and species. Lectotypes are designated for .V. triacanthus and V. latifrons. INTRODUCTION The family Agonidae is composed of typically small, benthic, scorpaeniform fishes that are al- most totally encased in rows of overlapping der- mal plates; the centers of these plates often bear spines or protuberances. The majority of the species are found in the North Pacific Ocean and Bering Sea; only 3 of the approximately 50 rec- ognized species (distributed among some 20 cur- rently recognized genera) are restricted to other regions: two in the North Atlantic Ocean (As- pidophoroides monopterygius and Agonus cata- phractus), and one off southern South America (Agonopsis chiloensis). Only two major reviews of the family have been written. The first is that of Jordan and Ev- ermann (1898). While this work was primarily concerned with American species, all known agonids were considered; little osteology was dis- cussed, and only a very few specimens of each species were examined. The second review of the family is that of Freeman ( 1 95 1 ), a widely known, but unpublished doctoral dissertation written at Stanford University. Once again, little osteology was examined, and few specimens were used. Only two notable osteological investigations of agonids have been published: Rendahl's (1934) work on Hypsagonus quadricornis and Ilina's (1978) work on the genera Podothecus and Agon- us. Both were limited to aspects of cranial os- teology. The taxonomic history of the genus Xeneret- mus began with Gilbert's (1890) erection ofXen- ochirus, established to contain three species: X. triacanthus, X. pentacanthus, and X. latifrons. Five years later, Gilbert (1895) described a fourth species, Xenochirus alascanus. In 1903 Gilbert (in Jordan 1 903) became aware of the prior use of the name Xenochirus by Gloger (1842) for a genus of marsupial mammals, and therefore of- fered Xeneretmus as a replacement name. In the following year, a fifth member of the genus, X. infraspinatus, was described by Gilbert (1904). In his final paper on this genus, Gilbert (1915) described two additional species, X. ritteri and X. leiops, moved X. alascanus, X. infraspinatus, [17] 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 and X. pentacanthus into a new genus, Astero- theca, and created two subgenera within Xeneret- mus: Xenopyxis containing X. latifrons, X. leiops, and X. ritteri; and Xeneretmus containing only X. triacanthus. Jordan et al. (1930), without ex- planation, raised the subgenus Xenopyxis to ge- neric status. Although a few later workers (Barn- hart 1936; Clemens and Wilby 1961) followed Jordan et al. (1 930), the majority of recent work- ers (Freeman 1951; Peden and Gruchy 1971; Miller and Lea 1972; Hart 1973; Barraclough and Peden 1976; Robins 1980; Eschmeyer et al. 1983) retained Gilbert's (1915) classification. Since Gilbert (1915), only minor publications on the genus have appeared. Bolin (1937), while noting that Gilbert's (1915) labels on the illus- trations of A', ritteri and X. leiops were switched, extended the geographic range of A', leiops north from Santa Catalina Island to Monterey Bay. Pe- den and Gruchy (1971) expanded the range of X. triacanthus into British Columbian waters. Ginn and Bond (1973) extended the range of A'. leiops north to the Columbia River. Three years later, Barraclough and Peden (1976) extended the range of X. leiops further north to southern British Columbia and noted, as did Bolin (1937), the switching of Gilbert's (1915) labels. The purposes of this study are to provide a complete osteological description of the genus Xeneretmus, to describe variation in a number of systematically important characters, to des- ignate type material for the species where it is in question, and to investigate the phyletic nature of the genus, subgenera, and closely related taxa. MATERIALS AND METHODS All measurements were taken from the right side of the fish. Standard length (SL), used throughout, was measured from the tip of the snout to the postero ventral corner of the last su- pralateral plate. Other measurements were made as follows: Anal, first dorsal, and second dorsal lengths.— from the tip of the snout to the insertion of the first ray of the respective fin Caudal peduncle length.— from the insertion of the posteriormost anal ray to the posteroven- tral corner of the last supralateral plate Vent length.— from the tip of the snout to the anterior margin of the anal opening Ventral head length.— from the tip of the snout to the posterior margin of the isthmus Depth at first and second dorsal.— the shortest distance from the first ray of that fin to the ventral contour of the body Head length.— from the tip of the snout to the posteriormost margin of the opercular mem- brane Supraoccipital pore to snout.— from the tip of the snout to the anterior edge of the acoustico- lateralis pore located dorsal to the supraoccipi- tal bone Snout length.— from the tip of the snout to the anterior margin of the orbit Upper jaw length. — from the anteriormost extent of the premaxilla to the posteriormost margin of the maxilla Length of orbit.— greatest distance between the rims of the orbit Interorbital width.— least distance between the lateral margins of the frontals Length of pectoral and pelvic fins.— from the base of the longest ray to its tip Caudal depth.— least depth of the caudal pedun- cle Pectoral width.— the greatest width measured between the pectoral bases Ural centra are included in vertebral counts ob- tained from radiographs. Nomenclature for, and the method of enu- meration of, dermal plates follow the system out- lined by Gruchy (1969) with the following ad- ditions: the number of plates anterior or posterior to a fin is counted to or from, but not including, the plate on which the first or last ray of the fin is inserted; the ventrolateral series of plates is considered to start at the pelvic fin base. Cladistic analysis was performed using Joseph Felsenstein's (Department of Genetics, Univer- sity of Washington) Package for Inferring Phy- logenies. Summary statistics were calculated us- ing SPSS and SCSS (Nie et al. 1975, 1980). All programs were run on a Digital Electronics Cor- poration VAX, under the VMS operating system. For the phylogenetic analysis, characters from the following in-group were recorded: Aspido- phoroides bartoni, A. olriki, Bathyagonus alas- canus, B. infraspinatus, B. nigripinnis, B. pen- tacanthus, Bothragonus swani, Odontopyxis trispinosa, Xeneretmus latifrons, X. leiops, X. rit- teri, and X. triacanthus. A close relationship of these taxa was hypothesized by Freeman (195 1); all are members of his subfamily Xeneretminae. A set of morphological characters (Table 1) was LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT 19 i°* £ e .2 3i~ gill U o o 2 J5 CJ U 0 i u "6 u J3 J3 JC U U U 33 3 o o j ^ - > a a a a E G E s * E S e E E E •a •* e <_ 3 « fll e e c £ e u C C m E ° -g H i> § !> i 1 1 i 1 1 01 1 01 1 S & 0 11 1 1 3 U Q sn KM OI | o *j <8 «- 00 "*l 09 99 99 (t> 4J 4> i >. 1 i I § C e >> >. C V S o >. e si. >< Ifff C C S « 5 3s 1 J 1 u i E 1 e e e c II 1 e S g s s e c a a " oi oi a a a a 03 a a a a a a a a "ot a a a 1 s g U (U U u > > > o 0 i 2 0 So o •S e c 0 II § § e e 8" S 1 e e e a a a c a e a e a e a e a a e a c a O *J *J 4J e C a oi oi § x> ot qi 5> M 35 01 01 II 01 01 II u X 1-3 u J3 C o e •o2 „ o o S ^ 0 0 0 0 *"* 5 * 0 ** e o Al <« «§ e i i i £ S J I 1 | 1 1 i i 1 G 5 c e c c S | £ •- a a « a « 01 03 01 a a a ma a a a a cO — . 3 S e e ~ c u u h e >* ~ Of -S •^ a o. 0 M U U S 111 III 1 1 1 1 1 1 1 1 1 1 § 1 e e S SJ X) X) C C 8 S^ S 8 W o 01 01 01 03 at 01 03 01 01 at « oi a oi m 01 01 a a 1 1 4) 4} U u u V o o II e e e o o o § 0 § O o o O O a i c 0 S 0 g § § B I 3 (A VI a a 1 1 it e e g 5 § 5S » en 41 X> X) C oi oi a a a a 1 03 1 01 a a a c c e QJ 1> O SA SA ft O CJ 1> a a a i X) 01 fli qi X) X 01 01 e e a a I 2.1 o a & cu S S g S S g X> X) C oi oi a 1 1 a j§ 03 Xi a a a U (G CA a a "§ 1 01 e e S^ S x> x> 01 01 present present o •s S o e e e e e a e "e a e e a eg S e e e g e <-' -5 III 1 1 1 1 1 1 1 1 II | S CTJ C5 BJ 03 a 03 01 01 at a 03 O. 01 01 01 01 a "S III o ,Q Cu u ° <^ e e g 5 S y? w « D X) Xl ft M M a e a c a e a 1 03 i i a C a c a Eg C a a a 1 01 s g O w) w y •§ a a a S"c3 « w ^ n ^ ._ — «- «- 3 ™ S cog c g G 2 e 2 g1 eg e e e g e e £ '•£ ^ U ° - III oi oi a a 01 I 03 a at c. « "§ a it 01 S S •§ a a a 1 1 S3 3 « C § §1 C/3 S S § 3 § &* & o a « ^ ^. ^. eq cq infraspinatl S> -S S> 6« Ci, gp ||| ll swani |i s a 5. ? S -3 S •§ s « latifrons Xeneretmus leiops Xeneretmus ritteri Xeneretmus JlH* *$i»$ dodecaedro Pallasina barbata Percis japonic Podothecus acipenserin Sarritor frena 20 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 ouids piomqin -Bjdns pasodxg souids IBSEN SOUldS [BJUOJJ OUldS IB13UBJ ouids DIIOJOJ,.) souids £ souids i OO— • — O—iOO — — — — — — OOOOO oo — — o — o — — — — — — — oo — — — — _o — — oo — — ooooooooooo o sojB(d 5(ooip JO JUOUIOSUBJJV ooo — — oo — oo — oooooooo — — — oo — oo — oo o sojBjd Apoq ' uo souids *IJ l ooooooo — — cu — oooooooo uo auids 3UQ '.' OOOOOOOOOO.O — OOOOOOO uo souids aaiqx 'J oo — — — — ooooooooooooo uo souids OAIJ oo — — — — o — — — — — ooooooo JEJISOJ ^^--H^M^^-^^H^H^^^H^H^^^HOOOOOOO JBJJSOJ pasodxg oooooooooooo — oooo — — oo — — — — oooo — — — o — oo — — PIOJ-33JJ snuiqisi ^.^o-oooo oo -o- oo — — — — ooooo — — — — — — o — oooooooooooo — oooo — — oooooooooooo — oooo — — oo — — — — o — — — — — — — oo — — — — oo — — — — oo — — — — — — ooo — — — oo — — — — oo — — — — — oooo — — oooo — — oooo — o — oooo — o oo — — — — oo — — — — — — oo — — — — oo — — o — o — oooo — ooo — — — — •a 3 3 3 *S .55 •« 1 '! I I -a i 1 I ill Hi a -S e ex * i; ^ 3 3 3 3 3 -a ^ iiili 11 Illlitl &> &> SN >> b c fc i £ « ill 341-S - Is II A u •o H v o * 8, e ? LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMVS GILBERT 21 collected and converted to binary characters (Ta- ble 2) by additive binary coding (Sokal and Sneath 1963; Kluge and Farris 1 969). The ancestral state of each binary character was established by ex- amining an out-group, that is, a group of taxa considered not to be members of the smallest monophyletic unit that contains all members of the in-group. The most frequent character state found among the species of the out-group was considered to be the primitive character state for the taxa of the in-group. The out-group was com- posed of Agonopsis vulsa, Hypsagonus quadri- cornis, Ocella dodecadron, Pallasina barbata, Percis japonicus, Podothecus acipenserinus, and Sarritor frenatus, seven agonid species consid- ered to be related to but not members of the in- group (Jordan and Evermann 1898; Freeman 1951). Several methods of estimating phylogenies from binary data have been proposed (Felsen- stein 1 982, and references cited therein). Of these methods, Wagner analysis (Farris 1 970; Farris et al. 1 970a, 1 9706) has been the most widely em- ployed (Baird and Eckhardt 1972; Simon 1979; Presch 1980; Jensen and Barbour 1981; Miya- moto 1983) and has been examined in detail (Colless 1981; Felsenstein, 1973, 1978, 1979; Mickevich 1978, 1980, Mickevich and Farris 1981; Schuh and Farris 1981; Schuh and Pol- hemus 1980; Sokal and Rohlf 1981). Felsenstein (1973, 1978, 1979) has shown some assumptions of the method: 1 . Characters evolved independently. 2. Changes of character states through time are a priori improbable. 3 . Polymorphisms of character states for a species are exceedingly unlikely. 4. Inequality of lengths of segments of the tree is not so extreme that two changes of states along a long segment is more probable than one change along a short segment. 5. Different lineages evolved independently. While these assumptions do not exactly express how the world is believed to work, without as- sumptions no explicit model could be advanced; with them at least we know the assumptions upon which the hypothesis rests. All systematists make assumptions when trying to work out a phylog- eny, but their assumptions are not as open to examination as are those of a model. The Wagner method searches for the cladogram that requires the fewest number of steps for all the characters. Some hypothesized monophyletic sets may not be supported by uniquely derived characters. Osteological material was cleared and stained with alizarin red S following the method of Tay- lor (1967). Osteological drawings were prepared with the aid of a Wild M5 stereomicroscope and camera lucida. Osteological terminology follows Weitzman(1974). The following cleared and stained specimens were examined: Xeneretmus latifrons: UW 18216, 3 (148-171 mm); X. leiops: OSU 7309, 2 (186, 191 mm); X. ritteri: SIO 59-92, 1 (141 mm); X. triacanthus: UW 20948, 3 (146-167 mm). The following abbreviations are used in the Osteological illustrations: ANG angular ARP articular process ASP ascending process BPT basipterygium BR branchiostegal ray BSB basibranchial BSO basioccipital CBR ceratobranchial CHY ceratohyal CL cleithrum CO circumorbital COR coracoid DH dorsal hypohyal DLP dorsolateral plate DN dentary ECT ectopterygoid EPB epibranchial EPH epihyal EPO epiotic EPU epural ER epiplural rib EXO exoccipital F frontal FHA first haemal arch HPP hypural plate HYB hypobranchial HYM hyomandibular IHY interhyal ILP infralateral plate INT intercalar IOP interopercle IPS infrapharyngo bran- chial LC lacrimal LE lateral ethmoid LLS lateral-line scale M maxilla MDP mid-dorsal plate MIS medial interopercular socket MSP mesopterygoid MTP metapterygoid MVP mid- ventral plate N nasal NZ neural zygapophysis OP opercle PC postcleithrum PHY parahypural PLT palatine PM premaxilla POP preopercle PPH parapophysis PRO prootic PRT parietal PSP parasphenoid PTG pterygiophore PTO pterotic PTS pterosphenoid PTT posttemporal Q quadrate R radial RAT retroarticular RP rostral plate SBO subopercle SCL supracleithrum SCP scapula SET supraethmoid SLP supralateral plate SOC supraoccipital SPH sphenotic SPN spine SYM symplectic T tabular UC ural centrum URN uroneural V vomer VH ventral hypohyal VLP ventrolateral plate Material examined is deposited at the follow- ing institutions: California Academy of Sciences, 22 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 San Francisco (CAS); Natural History Museum of Los Angeles County (LACM); Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts (MCZ); National Museum of Canada, Ottawa (NMC); Oregon State University, Corvallis (OSU); Scripps Institution of Oceanography, La Jolla, California (SIO); Stanford University (SU), material now housed at CAS; University of Alberta, Museum of Zo- ology, Edmonton, Alberta (UAMZ); United States National Museum of Natural History, Smithsonian Institution, Washington, D.C. (USNM); and School of Fisheries, University of Washington, Seattle (UW). COMPARATIVE MATERIAL EXAMINED Agonopsis vulsa: UW 4798 (27), UW 5359 (40). Aspidophoroides bartoni: CAS 10842 (2), CAS 15508 (1), CAS 15509 (1), CAS 22355 (2), CAS 26764 (3), CAS 26773 (1), MCZ 28323 (1), MCZ 32463 (1), SU 20421 (1), SU 26136 (5), SU 31699 (3), USNM 125584 (5), USNM 149047 (7), UW 20940 (1), UW 20941 (1), UW 20942 (1), UW 20943 (1), UW 20944 (4), UW 20945 (2), UW 20946 (2), UW 20947 (3). Aspidophoroides olriki: NMC 77-1537 (26), USNM 177610 (1), UW 20935 (3), UW 20936 (2), UW 20937 (1), UW 20938 (2). Bathyagonus alascanus: NMC 6502 1 9 (3), NMC 65-319 (3), NMC 66-16 (1), SIO 69-138 (3), SU 3088 (13), UAMZ 1985 (5), UAMZ 2774 (4), USNM 48741 (1), USNM 53582 (1), USNM 53583 (3), USNM 53586 (2), USNM 53589 (5), USNM 53592 (1), USNM 60484 (1), USNM 208391 (2), UW 1422 (5), UW 14392(11). Bathyagonus infraspinatus: CAS 14911 (2), NMC 65-259 (1), SIO 63-203 (2), SIO 69-1 10 (2), SIO 72-230 (1), SIO 72- 239 (1), SU 24967 (3), USNM 53593 (1), USNM 53595 (4), USNM 53596 (1), USNM 53597 (2), USNM 53598 (1), USNM 60416 (1), USNM 104676 (15), USNM 207968 (1), USNM 207990 (1), USNM 208117 (1), UW 1660 (1), UW 2886 (5), UW 5006 (1), UW 7583 (1). Bathyagonus nigripinnis: CAS 45524 (1), CAS 45525 (1), CAS 45526 (1), SIO 63-205 (8), SIO 69-140 (3), USNM 46613 (3), UW 7333 (4), UW 18147 (8), UW 20931 (1), UW 20932 (7), UW 20933 (1). Bathyagonus pentacanthus: CAS 15130 (4), NMC 65-397 (1), NMC 65-423 (6), NMC 71-693 (1), SIO 75-355 (3), SIO 80-9 (1), USNM 46612 (3), USNM 63444 (1), UW 18145 (15), UW 18472 (6), UW 19140 (3), UW 20934 (1). Bothragonus swani: UW 14155 (2), UW 17971 (1), UW 20929 (1), UW 20930 (2). Hypsagonus quadricornis: UW 1 1721 (32). Ocella dodecaedron: UW 20999 (5). Odontopyxis trispinosa: UW 1752 (4), UW 4375 (5). Pallasina barbata: UW 4206 (3). Percis japonicus: UW 21000 (1), UW 21001 (1). Podothecus acipenserinus: UW 3977 (125), UW 7340 (12). Sarritor frenatus: UW 20998 (5). OSTEOLOGY OF XENERETMUS TRIACANTHUS CRANIUM (Figs. 1, 9).— Rostral plate unpaired, situated anterodorsal to nasals, most anterior os- teological element and bears a single, dorsally directed spine; on either side, a laterally directed spine. Nasals in contact anteriorly, but separated posteromedially by anterior third of supraeth- moid. Each, bordered laterally by respective lac- rimal (Fig. 9), bears a strong, posterodorsally di- rected nasal spine. Lateral ethmoids lie posterior to nasals. Each comes into contact with frontal medially, lacri- mal laterally (Fig. 9), and parasphenoid ventro- laterally. Frontals are in contact with each other on mid- line for most of their length, separated by su- praethmoid anteriorly. Each frontal bordered posteriorly by sphenotic, pterotic, and parietal; posteroventrally by pterosphenoid. A sharp spine is present on the dorsal surface of each frontal, just posterodorsal to the orbit. Parietals meet on dorsal midline. Bordered along lateral margin by pterotic, tabular, and posttemporal. Each bears two posterodorsally di- rected spines: anteriormost spine knoblike, pos- teriormost strong and sharp. Most of the anterodorsal surface of supraoc- cipital, covered by parietals, comes into contact with exoccipitals along posterior margin. Pterot- ic meets sphenotic anteriorly, tabular posterior- ly, and prootic, exoccipital, intercalar, and post- temporal ventrally. Tabular dorsal to the epiotic; posttemporal bears a spine at posterior margin. Supracleith- rum articulates with posteroventral surface of posttemporal. Epiotic attaches at posterolateral corner of cranium, situated ventral to tabular and post- temporal. Exoccipital forms lateral and dorsal borders of foramen magnum; anteriorly, it forms posterolateral portion of otic capsule. A condyle at its posteroventral corner contacts the lateral process of the anteriormost vertebral centrum (preural centrum 41). Basioccipital, broad anteriorly, narrowing pos- teriorly, forms ventral margin of foramen mag- num. A single large condyle situated posteriorly, abuts against the anteriormost vertebral cen- trum. A posterior projection of parasphenoid overlaps anterior midline of basioccipital ven- trally; its anterolateral corner forms postero- medial portion of otic capsule. Parasphenoid runs from vomer anteriorly to basioccipital posteriorly, forms ventral margin of cranium; anteriorly receives shaft of vomer. SPH PTO PRT T PTT RP SET N LE EPO SOC EXO N PTT EPO INT SCL SPH PTS PTO FIGURE 1. Dorsal, left lateral, and ventral views of cranium of Xeneretmus triacanthus, UW 20948, 158 mm SL. Dotted lines portray canals of the acoustico-lateralis system. 23 24 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 ASP PM M FIGURE 2. Left lateral view of upper jaw of Xeneretmus triacanthus, UW 20948, 158 mm SL. Lateral ethmoids border on its anterolateral sur- along its anteroventral surface arranged in semi- face and its dorsolateral projections abut on pter- circular pattern. osphenoids dorsally and form posterior margin Prootic forms anterior portion of otic capsule; of orbits. Vomer is "tear"-shaped; teeth borne does not reach posterior margin of orbit. DN ANG RAT FIGURE 3. Left lateral view of lower jaw of Xeneretmus triacanthus, UW 20948, 1 58 mm SL. Dotted lines portray acoustico- lateralis canals. LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT 25 HYM SBO PLT ECT SYM POP I OP FIGURE 4. Lateral view of suspensorium and opercular apparatus ofXeneretmus triacanthus, UW 20948, 158 mm SL, right side reversed. Dotted lines portray acoustico-lateralis canals. Intercalar approximately circular; dorsally bordered by exoccipital and pterotic. Antero- medially directed projection of posttemporal overlies its posteroventral face. UPPER JAW (Fig. 2).— Premaxilla toothed along entire ventral surface in a broad band. The max- illa forked anteriorly to receive ascending process of premaxilla and widens abruptly posteriorly. LOWER JAW (Figs. 3, 4).— Anterodorsal three- fourths of dentary toothed. Angular bears socket POP EPH SBO MIS IOP FIGURE 5 . Left medial view of opercular apparatus of Xene- retmus triacanthus, UW 20948, 158 mm SL. FIGURE 6. Left lateral view of hyoid apparatus of Xene- retmus triacanthus, UW 20948, 158 mm SL. 26 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 BSB 1-3 CBR 5 HYB I- EPB 1-4 CBR I- FIGURE 7. Dorsal view of branchial basket of Xeneretmus triacanthus, UW 20948, 158 mm SL. on posteromedial surface to receive a process of quadrate (Fig. 4). Retroarticular attaches to the posteroventral corner of angular. SUSPENSORIUM (Figs. 1, 4-6, 9).— Palatine toothed for a third of its length, teeth centered about midpoint. It articulates posteriorly with mesopterygoid and ectopterygoid. Anterodorsal surface of palatine articulates with ventral sur- face of lateral ethmoid (Fig. 1). A lateral process of palatine articulates with medial surface of lac- rimal. Mesopterygoid borders ectopterygoid ventrally, quadrate posteriorly; does not contact metapter- ygoid. The ectopterygoid lies between palatine, mesopterygoid, and quadrate. Posteroventral surface of quadrate contacts preopercle. Metap- terygoid thin and flat; borders symplectic an- teroventrally and hyomandibular posteroven- trally. Hyomandibular has three dorsal articulating facets: anteriormost, articulating with sphenotic and prootic; medial articulating with pterotic; and posteriormost articulating with anterodorsal corner of the opercle (Fig. 1). Preopercle crescent-shaped with two spines along posterior margin. It is dorsally overlain by circumorbital 3 (Fig. 9). Elongate interopercle bears medial socket that fits onto posterior cor- ner of epihyal (Figs. 5, 6). Opercle triangular and slightly striated; a socket on anterodorsomedial face receives posterior facet of hyomandibular. Subopercle V-shaped, with crotch of the V lying dorsal to opercle (Fig. 5); posterior arm long and thin, lying on medial face of opercle for majority of its length; anterior arm short, its most dorsal point reaching only half the height of opercle. HYOID ARCH (Figs. 5, 6).— Dorsal hypohyal anterodorsal to ceratohyal. Ventral hypohyal forms an anterior cap over ceratohyal. Cerato- hyal has four branchiostegal rays connected to it: anterior two ventrally attached, posterior two ventrolaterally. Epihyal connected to the re- maining two basibranchials ventrolaterally. Pos- terodorsally, interhyal connects epihyal to hyo- mandibular. Posterior corner of epihyal fits into medial socket of interopercle (Fig. 5). BRANCHIAL ARCHES (Fig. 7).— Hypobranchials 1-3 broad and flat. Hypobranchial 2 two-thirds length of hypobranchial 1; hypobranchial 3 tear- shaped and two-thirds length of hypobranchial 2. All four ceratobranchials have anterior rows of gillrakers that bear toothlike structures; cer- atobranchials 1-3 also possess posterior rows of similar, "tooth"-bearing gillrakers. Ceratobran- chials 3 and 4 articulate with hypobranchial 3. Ceratobranchial 5 oval, completely toothed. Epi- LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT 27 FIGURE 8. Dorsal view of urohyal of Xeneretmus triacan- thus, UW 20948, 158 mm SL. FIGURE 9. Left lateral view of circumorbital bones of Xene- retmus triacanthus, UW 20948, 1 58 mm SL. Dotted lines por- tray acoustico-lateralis canals. branchials 1-4 all articulate dorsally with a sin- gle, large, well-toothed infrapharyngobranchial. Epibranchial 1 forked dorsally in some speci- mens. UROHYAL (Fig. 8).— Urohyal triangular, with a dorsomedial ridge rising posteriorly. CIRCUMORBITAL SERIES (Figs. 1, 9).— Lacrimal forms majority of dorsal surface area of snout and connects with lateral ethmoid and nasal (Fig. 1). Circumorbital 2 is tubelike. Circumorbital 3 has a single centrally located and posteriorly di- rected spine on lateral surface. Circumorbital 4 also tubelike, forming posterior margin of orbit. PECTORAL GIRDLE (Fig. 10).— Three rectan- gular radials and two postcleithra present. Scap- ula crescent-shaped and attached to posterior margin of cleithrum by two arms. Coracoid L-shaped, its anterior arm in contact medially with ventrolateral face of cleithrum; the dorsal arm with ventral margin of scapula and anterior borders of two ventralmost radials. Cleithrum the largest element of pectoral girdle; dorsally attaches to supracleithrum. PELVIC GIRDLE (Fig. 1 1).— Basipterygia paired and connected medially to each other along pos- terior tenth of their length. A ventral ridge runs anteroposteriorly. One spine and two rays pres- ent; lateralmost ray tightly bound to medial sur- face of spine. VERTEBRAL COLUMN (Fig. 12).— There are 41 preural centra in all the specimens of Xeneretmus triacanthus dissected. The 41st preural centrum has three anterior concave facets (a large central facet and two smaller lateral ones) that articulate with posterior surface of cranium; neurapophy- ses not dorsally ankylosed; 41st through 31st preural centra bear epiplural ribs. Neural zyg- apophyses become more pronounced posterior- ly. Haemal spines on 29th through first preural centra. Haemal spines posterior to posteriormost anal fin pterygiophore lie ventral to adjacent pos- terior centrum (this tendency increases poste- riorly). Posterior two rays of anal and second dorsal fin articulate with last pterygiophore of ventral and dorsal series, respectively. Four pterygio- phores lie between first and second dorsal fins. No ray articulates with these pterygiophores. An- terior two pterygiophores of ventral series do not articulate with any rays. Neural spines on 40th through first preural centra; centra posterior to last dorsal pterygio- phore have broad neural spines that lie between neural zygapophyses of adjacent posterior ural centrum. One parahypural fused to ventral margin of hypural plate. One uroneural tightly bound to dorsal margin of hypural plate. Epurals absent. DERMAL PLATES (Fig. 13).— All dermal plates of supralateral, dorsolateral, and mid-dorsal se- ries bear posteriorly directed spines. All infra- lateral plates bear similar spines as well, except those medial to pectoral fin. Ventrolateral plates spineless, except for those behind pelvic fin in- sertion to approximately four plates anterior to insertion of first anal fin ray. Mid- ventral plates spineless. Anterior supralateral plates overlain poste- riorly by next supralateral plate for approxi- mately half their length; those more posterior in position overlain by as little as 20% of their length. Supralateral plates overlain by dorsolateral and mid-dorsal plates dorsally, and lateral-line scales ventrally. Anterior infralateral plates have half their length overlain by adjacent posterior infralateral plate; length covered reduced to 20% posteriorly. Infralaterals overlain by lateral-line scales dor- 28 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 SCP PC I PC 2 COR FIGURE 10. Left lateral view of pectoral girdle of Xeneretmus triacanthus, UW 20948, 158 mm SL. sally and ventrolateral and mid-ventral plates ventrally. Ventrolateral plates bordering anal fin have medial projections that meet on midline such that fin rays surrounded by plates. A third of the length of each ventrolateral plate overlain pos- teriorly by next ventrolateral plate. Medially, ventrolateral plates slightly overlap each other on midline. Each mid- ventral plate posteriorly overlain by immediately posterior mid- ventral plate; length covered approximately 1 5%. RAYS FIGURE 11. Ventral view of pelvic girdle of Xeneretmus triacanthus, UW 20948, 158 mm SL. LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT 29 PTG FHA URN HPP FIGURE 12. Left lateral view of vertebral centra of Xeneretmus triacanthus, UW 20948, 158 mm SL. 30 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 SLP DLP MOP MVP VLP ILP VLP FIGURE 1 3. Dermal body plates of Xeneretmus triacanthus, UW 20948, 1 58 mm SL: (A) Dorsal view of plates immediately posterior to insertion of posteriormost ray of second dorsal fin. (B) Left lateral view of plates immediately posterior to pectoral fin. (C) Ventral view of plates immediately posterior to inser- tion of posteriormost anal fin ray. Each dorsolateral plate overlain posteriorly by next dorsolateral plate for approximately 35% of length. Like ventrolaterals, dorsolateral plates bordering first and second dorsal fins have me- dial projections that meet on dorsal midline such that only a small break in plates exists where dorsal rays insert. Each mid-dorsal plate overlain posteriorly for 20% of length by adjacent posterior mid-dorsal plate. ACOUSTICO-LATERALIS SYSTEM (Figs. 1, 3, 4, 9).— Acoustico-lateralis system of cranium passes posteriorly through nasals. Canal enters frontals through an anterior pore and extends along me- dial border to a medial pore where it turns lat- erally, continuing to posterolateral pore of fron- tal where it branches (Fig. 1). Anteriorly directed branch of acoustico-lateralis system passes through entire circumorbital series (Fig. 9). Pos- teriorly directed branch passes through pterotic and branches again beneath tabular. Medially directed branch passes beneath posterior spine of parietal to a medial pore dorsal to supraoc- cipital where it meets its counterpart from the other side. Posteriorly directed branch passes through posttemporal and supracleithrum, con- tinues posterolaterally along entire length offish (Fig. 1). A second acoustico-lateralis canal passes posteriorly through dentary, angular, and pre- opercle (Figs. 3, 4). COMPARATIVE OSTEOLOGY The osteology of the other species of the genus is almost identical with that of X. triacanthus described above; very few differences were found between species that exceeded variation within species. Xeneretmus triacanthus has two spines on the posterior margin of the preopercle where- as its congeners possess only one. The rostral plate of the other members of the genus do not have lateral spines (except in some individuals of X. leiops). Xeneretmus ritteri has two spines on circumorbital 3, whereas the other species of Xeneretmus have only one. Finally, the pterotic of X. ritteri bears two spines while in the other members of the genus only a ridge may be dis- cerned. The following comparisons can be made with Rendahl's (1934) work on the agonid species Hypsagonus quadricornis. The rostral plate, di- agnostic for Xeneretmus (Fig. 1), does not occur in Hypsagonus (Rendahl 1934, figs. 1-3). The vomer of Hypsagonus is toothless (Rendahl 1934, fig. 2). The frontal spine of Hypsagonus is much larger than that of Xeneretmus (Fig. 1 ; Rendahl 1934, figs. 1, 3). In Hypsagonus, Rendahl (1934, fig. 1) depicted the supraoccipital as lying along the entire medial border of the parietal, in such a configuration that the parietals are not in con- tact along their medial edges; the parietals meet along their entire medial edges in Xeneretmus (Fig. 1). Rendahl (1934, figs. 28 A, 28B) depicted the retroarticular as lying only on the medial face of the angular whereas in Xeneretmus the re- troarticular is visible from both a medial and lateral view (Fig. 3). Rendahl (1934, fig. 24 A) showed the mesopterygoid and metapterygoid of Hypsagonus to be in contact with each other; in Xeneretmus the quadrate is between these two bones such that they do not meet (Fig. 4). Finally, the posterior arm of the subopercle of H. quad- ricornis is considerably shorter than is the case for Xeneretmus (Figs. 4, 5; Rendahl 1934, figs. 24A, 24B). LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT 31 The following comparisons can be made with the work of Ilina (1978). Neither of the genera Podothecus and Agonus possesses a rostral plate. Ilina (1978, figs. 2-4, 6) portrayed the supraoc- cipitals of Podothecus acipenserinus, P. veternus, P. gilberti, and P. thompsoni as lying between the parietals, such that they do not meet along their medial edges, as they do in Xeneretmus (Fig. 1). The posttemporal of Podothecus acipenserinus, P. veternus, and Agonus cataphractus apparently makes no contact with the intercalar (Ilina 1978, figs. 2, 3, 7); in Xeneretmus the anteriorly di- rected projection of the posttemporal touches the intercalar (Fig. 1). SYSTEMATICS Genus Xeneretmus Gilbert Xenochirus GILBERT, 1890:90 (type-species Xenochirus tria- canthus GILBERT, 1 890, by original designation; preoccupied by Xenochirus GLOGER, 1842, a genus of marsupial mam- mal). Xeneretmus GILBERT, in Jordan 1 903:360 (substitute name for Xenochirus GILBERT, 1 890 [preoccupied, therefore taking the same type-species Xeneretmus triacanthus}). DIAGNOSIS.— The genus Xeneretmus is distin- guished from all other agonid genera by the ab- sence of a supraoccipital pit and by the presence of an exposed rostral plate bearing a single dor- sally directed spine. DESCRIPTION.— Body tapering uniformly from pectoral girdle to caudal fin; anterior cross sec- tions octagonal, cross section through caudal peduncle hexagonal; completely encased in over- lapping dermal plates. All dorsolateral, mid-dor- sal, and supralateral plates bearing posteriorly directed spines; all but those plates medial to pectoral fin of infralateral series bearing poste- riorly directed spines; ventrolateral plates be- tween pelvic fin insertion and insertion of first anal fin ray bearing posteriorly directed spines; no spines present on mid- ventral plates; dorso- lateral plates 22-24; mid-dorsal plates 12-19; supralateral plates 39-45; infralateral plates 35- 42; ventrolateral plates 21-23; mid- ventral plates 14-20. In comparison with the other genera of Agonidae, Xeneretmus has long spines on slightly flexible angular dermal plates. Cephalic spines. One nasal spine; one frontal spine dorsal to posterior edge of orbit; two pa- rietal spines; one posttemporal spine; one or two spines on circumorbital 3; one or two spines on posterior margin of preopercle. Fin rays. All simple; four ventralmost pectoral B FIGURE 14. Right lateral view of head of two species of Xeneretmus: (A) X. triacanthus; (B) X. latifrons. Arrows in- dicate dermal plates of cheek region. rays thickened (in comparison to the dorsalmost pectoral rays), and projecting fingerlike from the fin membrane; dorsal two thickened rays longest rays of pectoral fin. First dorsal, 5-8; second dor- sal, 6-8; anal, 5-8; pectoral, 12-16; pelvic I, 2; branchiostegal rays, 6. Mouth. Both jaws of equal length, mouth ter- minal; teeth present on premaxilla, dentary, pal- atine, and vomer. Barbels present along ventral margin of dentary at edges of acoustico-lateralis pores and at posterior corner of maxilla. Measurements. The following ranges for pro- portions of all species of the genus are expressed in thousandths of standard length (number of specimens measured in parentheses): anal length, 305-510 (201); vent length, 210-335 (197); cau- dal peduncle length, 370-479 (199); second dor- sal length, 443-592 (20 1 ); depth at second dorsal, 47-72 (197); first dorsal length, 258-386 (203); depth at first dorsal, 65-132(171); pectoral length, 123-217 (189); pelvic length, 51-105 (197); pec- toral width, 92-157 (189); head length, 166-241 (203); ventral head length, 80-155 (200); length from supraoccipital pore to snout, 147-1 89 (203). The following proportions, associated with characteristics of the head, are expressed in thou- sandths of head length (number of specimens measured in parentheses): orbit length, 249-476 (207); upper jaw length, 229-363 (141); snout 32 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 TABLE 3. RANGE, MEAN, STANDARD DEVIATION, AND SAMPLE SIZE FOR MERISTIC CHARACTERS OF SPECIES OF Xeneretmus. Character X. latifrons X. leiops X. ritteri X. triacanthus First dorsal spines 6-8 X = 6.9 6-7 X = 6.7 6-7 X = 6.4 5-7 X = o.O SD = 0.50 n = 110 SD = 0.45 n = 30 SD = 0.54 n = 7 SD = 0.28 n = 63 Second dorsal rays 6-8 Y = 7.0 7-8 X = 7.4 6-7 x = 6.9 6-7 X = O.O SD = 0.43 n = 110 SD = 0.49 n = 30 SD = 0.38 n = 7 SD = 0.50 tl ~ o3 Anal fin rays 6-8 x = 7.2 6-8 X = 6.9 6-7 X = 6.9 5-7 * = 6.1 SD = 0.15 n = 110 SD = 0.57 n = 30 SD = 0.38 n = 7 SD = 0.49 n = 63 Pectoral fin rays 13-15 X = 14.1 13-15 x = 14.0 16 x = 16 12-14 X= 13.0 SD = 0.40 n = 109 SD = 0.33 n = 29 SD = 0.00 n = 7 SD = 0.22 /z = o 1 Eyeball plates 3-6 X = 4.2 0 x = 0.0 3-6 x = 5.0 2-6 0. *l 0 SD = 0.62 n = 110 SD = 0.00 n = 30 SD= 1.0 n = 7 SD = 0.76 n = 63 Supralateral plates 39-42 X = 40.9 43-45 X = 43.9 40-41 x = 40.7 41-43 X = 42.0 SD = 0.65 n = 109 SD = 0.80 n = 28 SD = 0.49 n = 7 SD = 0.46 n = 58 Infralateral plates 35-40 X = 37.6 39^t2 x = 40.3 36-38 x = 36.9 38-40 X = 39.0 SD = 0.92 n = 108 SD = 0.98 n = 28 SD = 0.69 n = 7 SD = 0.49 n = 58 Mid-dorsal plates 12-16 X = 14.7 16-19 x = 17.7 14-15 x = 14.7 15-17 X= 16.2 SD = 0.63 n = 108 SD = 0.72 n = 29 SD = 0.49 n = 7 SD = 0.54 n = 58 Dorsolateral plates be- 3-5 x = 4.1 4-5 x = 4.4 4-5 x = 4.6 4-6 * = 5.1 tween first and sec- SD = 0.53 n = 110 SD = 0.50 n = 30 SD = 0.54 n = 7 SD = 0.33 n = 64 ond dorsal fins Mid-ventral plates 14-17 * = 15.4 16-20 Jf = 18.2 14-16 JC = 14.9 16-18 X= 16.7 SD = 0.66 n = 109 SD = 0.83 n = 28 SD = 0.69 n = 7 SD = 0.50 n = 59 Cheek plates 0-1 X = 0.0 0 X = 0.0 0 X = 0.0 1-4 * = 2.75 SD = 0.10 n = 110 SD = 0.00 n = 30 SD = 0.00 n = 7 SD = 0.69 n = 64 Vertebrae 40-42 x = 40.6 43-45 X = 43.7 40-41 X = 40.7 42 X = 42.0 SD = 0.79 n = 7 SD = 0.82 n = 6 SD = 0.52 n = 6 SD = 0.00 « = 6 length, 157-330 (206); interorbital length, 58- 123 (107). KEY TO THE SPECIES OF THE GENUS XENERETMUS la. Cheek plates present, filling area between circumorbitals and elements of the lower jaw (Fig. 14) Xeneretmus (Xeneretmus} triacanthus, p. 36 Ib. Cheek plates absent, or rarely represented by a single plate that does not fill the area between the circumorbitals and elements of the lower jaw (Fig. 14) Subgenus Xenopyxis 2 2a. Two or more dermal plates on eyeball 3 2b. No dermal plates on eyeball Xeneretmus (Xenopyxis) leiops, p. 35 3a. Two or more barbels present at posterior corner of maxilla, 1 6 pectoral rays Xeneretmus (Xenopyxis) ritteri, p. 36 3b. One barbel present at posterior corner of maxilla, 13-15 pectoral rays Xeneretmus (Xenopyxis) latifrons, p. 32 Subgenus Xenopyxis Gilbert Xenopyxis GILBERT, 1 9 1 5:345 (type-species Xeneretmus (Xen- opyxis) latifrons Gilbert 1890, by original designation). Jor- dan et al. 1930:396 (elevated to generic level). DIAGNOSIS.— Distinguished from the subgenus Xeneretmus by the absence of dermal plates in the cheek region (Fig. 14B), and the failure of the breast plates to abut against each other. It further differs in having higher average counts for rays in the first and second dorsal, anal, and pectoral fins (Table 3); having larger eyes (Table 4); and in general being more robust. Xeneretmus (Xenopyxis) latifrons (Gilbert) [Blacktip Poacher] (Figure 15) Xenochirus latifrons GILBERT, 1890:92 (original description, lectotype USNM 43091). Xeneretmus latifrons GILBERT, in Jordan 1903:360 (new com- bination, Xenochirus preoccupied). Xeneretmus (Xenopyxis) latifrons GILBERT, 1 9 1 5:345 (descrip- tion, key). Xenopyxis latifrons JORDAN ET AL., 1930:396 (checklist). MATERIAL EXAMINED.— One hundred and eighty- two spec- imens, 62 to 173 mm. LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT 33 TABLE 4. RANGE, MEAN, AND SAMPLE SIZE FOR BODY PROPORTIONS OF SPECIES OF Xeneretmus. Character X. latifrons X. leiops X. ritteri X. triacanthus Anal length/SL 305^86 X = 462 n = 109 X 434-493 = 457 n = 28 X 484-508 = 498 /i = 1 Jt 448-510 = 477 n = 57 Vent length/SL 220-335 je=253 n = 109 X 224-288 = 244 n = 24 X 253-280 = 265 n = 7 Jf 210-255 = 230 n = 57 Caudal length/SL 370-447 Jf =414 /! = 106 X 384-479 = 434 n = 30 X 380-409 = 393 n = 7 * 384-439 = 410 n = 56 Second dorsal length/SL 443-592 X = 487 n = 108 X 446-498 = 464 n = 29 X 490-516 = 501 n = 7 je 459-500 = 483 n = 57 Depth at second dorsal/SL 48-72 X=59 n = 106 X 48-64 = 57 n = 29 X 58-64 = 61 /! = 7 X 47-67 = 58 n = 55 First dorsal length/SL 258-386 X = 307 n = 108 X 269-319 = 291 n = 30 X 315-335 = 323 n = 7 X 287-317 = 302 n = 58 Depth at first dorsal/SL 73-132 X= 91 n = 88 X 72-98 = 85 /i = 22 X 86-117 = 95 /! = 7 X 65-109 = 82 n = 54 Pectoral length/SL 123-217 X= 167 n = 104 X 139-205 = 171 n = 29 X 163-188 = 173« = 7 X 151-202 = 175 n = 49 Pelvic length/SL 57-105 X = 84 n = 108 X 56-96 = 77 n = 25 X 72-99 = 83 n = 7 X 51-94 = 77 n = 57 Pectoral width/SL 97-157 X= 113 n = 102 X 92-151 = 106 n = 23 X 107-127 7 X 95-125 - 109 « - 57 Head length/SL 166-241 X= 212 /i = 109 X 191-227 = 204 n = 25 X 231-248 = 238 n = 7 X 190-215 = 201 « = 57 Ventral head length/SL 80-146 X= 121 n = 105 X 99-155 = 113 n = 30 X 133-151 = 142 n = 7 X 110-143 = 125 n = 58 Supraoccipital pore to snout length/SL Orbit length/head length 147-186 Jt= 166 n = 249^t76 X = 374 « = 109 109 X X 150-181 = 163« = 335^*19 = 374 n = 28 30 X X 167-189 = 180/i = 334-372 = 358 n = 7 7 X X 149-171 = 160 /i = 299-358 = 326 n = 57 61 Upper jaw length/head length 255-363 X = 293 n = 53 X 234-313 = 280 n = 28 X 278-300 = 286 n = 7 X 229-283 = 251 n = 53 Snout length/head length 157-330 *=258 n = 109 X 230-316 = 284 n = 30 X 249-271 = 258 n = 7 X 258-301 = 282 n = 60 Interorbital length/head length 58-123 JP= 100 « = 107 X 66-96 = 81 n = 30 X 84-125 = 95 « = 7 X 71-112 = 89 n = 60 Caudal depth/caudal length 37-68 X = 47 n = 95 X 35-48 = 40 « = 28 X 49-60 = 57 it- 7 X 39-58 = 45 n = 53 LECTOTYPE.— USNM 4309 1,131 mm, Albatross station 2935, San Diego, California, 32°45'N, 1 17°23'W, 227 m. PARALECTOTYPES.— CAS 5072, 3 (108-110 mm), Albatross station 2973, Point Conception, California, 34°20'N, 1 19°44'W, 124 m; USNM 46602, 8 (72-1 36 mm), Albatross station 2935, San Diego, California, 32°45'N, 117°23'W, 227 m; USNM 46605, 2 (1 10-1 12 mm), Albatross station 3059, Lincoln City, Oregon, 44°56'N, 124°13'W, 141m; USNM 46608, 120 mm, Albatross station 2972, Santa Barbara, California, 34°19'N, 1 19°41'W, 1 12 m; USNM 4661 1, 1 1 1 mm, Albatross station 2948, Santa Cruz Island, California, 33°56'N, 1 19°42'W; UW 1416, 2 (109-1 10 mm), Albatross station 2973, Santa Barbara, California, 34°20'N, 119°44'W, 124 m. ADDITIONAL NON-TYPE MATERIAL.— CAS 12572, 141 mm, Farallones, California, 37°43'N, 123°03'W; CAS 14282, 2 (128, 131 mm), San Pedro, California, 33°45'N, 118°11'W; CAS 26404, 2 (1 14, 134 mm), Port Heuneme, California, 34°09'N, 119°12'W; CAS 26447, 116 mm, Gaviota, California; CAS 26554, 3 (112-113 mm), Goleta Point, California, 34°27'N, 1 19°50'W; CAS 26560, 4 (97-1 12 mm), Santa Barbara Point, California, 34°30'N, 120°00'W; CAS 26563, 4 (1 18-130 mm), Santa Barbara Channel, California, 34°15'N, 119°55'W; CAS 26596, 120 mm, Santa Monica, California, 33°50'N, 1 18°38'W; CAS 26630, 19 (62-103 mm), Point Dume, California, 34°00'N, 1 18°50'W; CAS 37497, 149 mm, Half Moon Bay, California, 37°10'N, 122°42'W; CAS 40310, 136 mm, Santa Cruz, Cali- fornia, 34°07'N, 1 19°42'W; CAS 47107, 4 (99-1 10 mm), Go- leta, California, 34°27'N, 119°50'W; CAS 47110, 8 (91-128 mm), Santa Monica Bay, California, 33°58'N, 1 18°38'W; CAS 47111, 6 (66-108 mm), Morro Bay, California, 35°30'N, 121°15'W; CAS 471 12, 156 mm, Marin County, California. NMC 65-0259, 5 (121-153 mm), Kwatna Inlet, British Co- lumbia, 52°07'N, 127°38'W. SU 16903, 2 (77, 95 mm), Santa Barbara Channel, Califor- 34 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 V, FIGURE 15. Xeneretmus latifrons, 142 mm SL. Courtesy of R. H. Gibbs, Jr., and the Fish Division, National Museum of Natural History. nia, 30°26'N, 120°14'W; SU 39779, 118 mm, Santa Barbara Channel, California, 34°25'N, 120°18'W. USNM 6 1 1 76, 1 52 mm, Albatross station 3671, Santa Cruz, California, 37°00'N, 122°20'W; USNM 63435, 3 (70-1 14 mm), Point Soma, California, 32°41'N, 1 17°14'W; USNM 63437, 3 (82-123 mm), Point Soma, California, 32'41'N, 1 17°14'W. UW 1415, 131 mm, Albatross station 3174, Bodega Bay, California, 38°16'N, 123°14'W; UW 2943, 10 (1 16-150 mm), Camano Island, Washington, 47°59'N, 122°13'W; UW 3151, 123 mm, Burrard Inlet, British Columbia, 49°10'N, 123°00'W; UW 3168, 4 (93-138 mm), Elliot Bay, Washington, 47°36'N, 122°22'W; UW 3907, 43 (66-142 mm), Hoodsport, Washing- ton, 47°30'N, 123°10'W; UW 4224, 12 (76-130 mm), Hood Canal, Washington, 47°17'N, 122°42'W; UW 4308, 129 mm, Hood Canal, Washington, 47°30'N, 123°10'W; UW 5780, 2 (137, 144 mm), Hoodsport, Washington, 47°30'N, 123°10'W; UW 5861, 120 mm, Golden Gardens, Washington, 47°40'N, 122°24'W; UW 5872, 124 mm, Tulalip Bay, Washington; UW 5960, 132 mm, Meadow Point, Washington, 47°36'N, 122°22'W; UW 7347, 4 (104-133 mm), Puget Sound, Wash- ington; UW 8016, 142 mm, Ballard, Washington, 47°40'N, 122°25'W; UW 18216, 5 (158-163 mm), Columbia River, 46°N, 124°W; UW 18297, 162 mm, Columbia River, 46°N, 124°W; UW 18507, 3 (162-173 mm), 46°N, 124°W; UW 20939, 146 mm, Bainbridge Island, Washington, 47°37'N, 122°33'W. DIAGNOSIS.— Distinguished from other mem- bers of subgenus by following combination of characters: three to six spine-bearing dermal plates on each eyeball; one barbel at posterior corner of maxilla; 13-15 pectoral rays (Table 5). DESCRIPTION.— Posterior free-fold of bran- TABLE 5. CHARACTERS USED IN DISCRIMINATING AMONG THE SPECIES OF Xeneretmus. Characters Taxa Eye- ball plates Maxil- lary bar- bels Cheek plates Pectoral rays Xeneretmus latifrons Xeneretmus leiops Xeneretmus ritteri Xeneretmus triacanthus 3-6 0 3-6 2-6 1 1 2 2-3 0 or small 0 0 1-4 13-15 13-15 16 12-14 chiostegal membrane narrow; two barbels on ventral surface of dentary, one at each posterior margin of two anteriormost acoustico-lateralis pores; breast plates surrounded by skin, and hav- ing slightly raised centers; first dorsal fin with black distal margin; second dorsal fin membrane lightly pigmented along rays, clear between rays; counts and proportions are given in Tables 3 and 4. DISTRIBUTIONS.— Gilbert (1890) described X. latifrons from specimens obtained from Alba- tross stations situated off the coasts of California and Oregon, ranging between approximately 33° and 45°N latitude. Material examined in this study ranged from Ensenada, California to Kwat- na Inlet, British Columbia (Fig. 1 6). Gilbert ( 1 890, 1915) reported X. latifrons occurred in depths from 35 to 399 m. All the lots examined for this study fell within that depth range. COMMENTS.— In Gilbert's (1890) original de- scription of the species, no type-specimen was designated. When the single specimen in a lot now registered as USNM 4309 1 was transferred to the United States National Museum by Gil- bert and his associates, it was indicated in an accompanying letter that this specimen was soon to be described as the type of the species (Susan Jewett, USNM, personal communication, 7 June 1982). After examination of this specimen and the other members of the syntypic series avail- able to me (CAS 5072, USNM 46602, USNM 46605, USNM 46608, USNM 46611 and UW 1416), USNM 43091 is hereby designated as the lectotype. This decision was reached for the fol- lowing reasons: It appears to have been Gilbert's intention to designate this specimen as the type for the species, it is very close to the average for the species in the majority of characters, and its condition is as good as, if not better than, that of any other member of the syntypic series. In comparison to its congeners, X. latifrons has LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT 35 130 125 120 115 110 105 130 125 120 115 110 105 FIGURE 16. Distribution of Xeneretmus latifrons (circles) and X. ritteri (stars). low counts for numbers of mid- ventral plates, mid-dorsal plates, infralateral plates, supralater- al plates, and vertebrae; and high counts for numbers of unpaired fin rays (Table 3). Its snout is shorter, orbits longer, and interorbital distance greater than those of the other Xeneretmus (Table 4). Xeneretmus (Xenopyxis) leiops (Gilbert) [Smootheye Poacher] Xeneretmus (Xenopyxis) leiops GILBERT, 1915:348 (original description, key, illustration, holotype USNM 75813). Xenopyxis leiops JORDAN ET AL., 1930:396 (checklist). MATERIAL EXAMINED.— Forty-four specimens, 67-21 1 mm. HOLOTYPE.— USNM 75813, 163 mm, Albatross station 44 10, Catalina Island, California, 323-357 m. PARATYPE.— SU 22988, 2 (108, 136 mm) Albatross station 4410, Catalina Island, California. ADDITIONAL NON-TYPE MATERIAL. — LACM 93744, 5 (83- 160 mm), Catalina Basin, California, 32°N, 1 18°W. NMC 67-0348, 2 (147, 163 mm), Rennell Sound, British Columbia, 53°21'N, 133°04'W; NMC 72-061 3(192-211 mm), Barkley Sound, British Columbia, 48°30'N, 126°10'W. FIGURE 17. Distribution of Xeneretmus triacanthus (cir- cles) and X. leiops (stars). OSU 7305, 67 mm, Newport, Oregon, 44°40'N, 124°10'W; OSU 7309, 15 (134-191 mm), Columbia River, 46°10'N, 124°05'W. SIO 72-81, 206 mm, Neah Bay, Washington, 48°22'N, 126°10'W. SU 3623, 7 (132-172 mm), Central California Coast, 34°45'N, 121°29'W; SU 16711, 2 (153, 165 mm), Monterey Bay, Cal- ifornia, 36°49'N, 122°30'W; SU 26420, 3 (161-182 mm), Mon- terey Bay, California, 36°49'N, 122°30'W. UW 18123, 198 mm, Columbia River, 46°N, 124°W; UW 18473, 2 (171, 179 mm), Columbia River, 46°N, 124°W. DIAGNOSIS. — Distinguished from the other members of the subgenus by the following com- bination of characters: absence of dermal plates on eyeball; one barbel at posterior corner of max- illa; 13-15 pectoral rays (Table 5). DESCRIPTION.— Posterior free-fold of bran- chiostegal membrane wide; breast plates thin and completely surrounded by skin; one to three bar- bels at posterior margin of anteriormost acous- tico-lateralis pore of dentary, and one or none at posterior margin of middle acoustico-lateralis 36 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 pore of dentary; first dorsal whitish at base, black at distal margin, black pigmentation nearly reaching origin of fin, retreating distally poste- riorly; second dorsal black at distal margin. DISTRIBUTION.— Gilbert (1915) described X. leiops from specimens captured off Santa Cata- lina Island, southern California (Albatross sta- tion 4410). Xeneretmus leiops has a geographic distribution that ranges from Santa Catalina Is- land north to the Queen Charlotte Islands (Fig. 1 7). Specimens examined during this study were captured from depths between 183 and 357 m. COMMENTS.— Relative to its congeners, X. leiops has high counts for second dorsal rays, supralateral plates, infralateral plates, mid-dor- sal plates, mid- ventral plates, and vertebrae (Ta- ble 3). It also has a shorter precaudal region, longer caudal peduncle and orbits, and a smaller interorbital distance than the other members of the genus (Table 4). Xeneretmus (Xenopyxis) ritteri (Gilbert) [Stripefin Poacher] Xeneretmus (Xenopyxis) ritteri GILBERT, 1915:350 (original de- scription, key, illustration, holotype USNM 75814). Xenopyxis ritteri JORDAN ET AL., 1930:396 (listed). MATERIAL EXAMINED.— Nine specimens, 106-141 mm. HOLOTYPE.— USNM 758 14, 1 23 mm, Albatross station 4366, Point Loma, California, 320-331 m. PARATYPE.— SU 22980, 106 mm, Albatross station 4322, San Diego, California, 353-415 m. ADDITIONAL NON-TYPE MATERIAL. -LACM 88182 2 (1 1 1, 137 mm) Gulf of California, Mexico, 29°N, 1 12°W. SIO 59-92, 4 (121-141 mm), Cedros Island, Mexico, 28°23'N, 1 15°21'W; SIO H50-245B, 126 mm, Torrey Pines, California, 32°10'N, DIAGNOSIS.— Distinguished from the other members of the subgenus by the following com- bination of characters: three to six spine-bearing plates on eyeball, two barbels at posterior corner of maxilla, 1 6 pectoral rays (Table 5). DESCRIPTION.— Posterior free-fold of bran- chiostegal membrane narrow; two barbels on ventral margin of dentary, one at posterior mar- gin of each of two anteriormost acoustico-la- teralis pores; breast plates with bony prickles at centers, each surrounded by skin (such that they do not contact each other at their edges); dorsal fins with black bars along base and distal margin. DISTRIBUTION.— Gilbert (1915) described X. ritteri from specimens captured near San Diego (Albatross stations 4366 and 4322). Since that time X. ritteri has been obtained from Cedros Island, Baja California, north to Malibu, Cali- fornia, and in the northern section of the Gulf of California (Fig. 1 6). Specimens examined for this study were captured at depths from 274 to 415m. COMMENTS.— In comparison to its congeners, X. ritteri has low counts for supralateral plates, infralateral plates, mid-dorsal plates, mid-ven- tral plates, and vertebrae (Table 3). It also has a larger head, a longer precaudal region, and a shorter caudal peduncle than other species of Xeneretmus (Table 4). Its spines and ridges are more strongly developed than those of its con- geners. Subgenus Xeneretmus Gilbert Xeneretmus (Gilbert) 1915:345 [Type species Xeneretmus (Xeneretmus) triacanthus Gilbert, 1890, by original designation.] DIAGNOSIS.— Distinguished from the subgenus Xenopyxis by the presence of one to four dermal plates in the cheek region leaving little or no skin exposed in the cheek region (Fig. 14), and the tight arrangement of the breast plates. It further differs in having lower average counts for fin rays in the first and second dorsal, anal, and pectoral fins (Table 3); having larger eyes (Table 4); and in general being slender in comparison. Xeneretmus (Xeneretmus) triacanthus (Gilbert) [Bluespotted Poacher] Xenochirus triacanthus GILBERT, 1890:91 (original description, lectotype USNM 43089). Xeneretmus triacanthus GILBERT, in Jordan, 1903:360 (New combination, Xenochirus preoccupied). Xeneretmus (Xeneretmus) triacanthus GILBERT, 1915:345 (de- scription, key). MATERIAL EXAMINED.— Seventy-six specimens, 74-167 mm. LECTOTYPE.— USNM 43089, 1 5 1 mm, Albatross station 2893, Santa Barbara Channel, California, 34°13'N, 120°33'W, 265 m. PARALECTOTYPES.-USNM 46601, 2 (124, 154 mm), Al- batross station 2893, Santa Barbara Channel, California, 34°13'N, 120°33'W, 265 m; USNM 46606, 1 17 mm, Albatross station 3059, Lincoln City, Oregon, 44°56'N, 124°13'W, 141 m; USNM 125577, 5 (137-152 mm), Albatross station 2973, Point Conception, California, 34°20'N, 1 19°44'W, 124 m. ADDITIONAL NON-TYPE MATERIAL. -CAS 13100, 2 (78, 90 mm), San Pedro, California, 33°43'N, 1 18°23'W; CAS 14270, 4 (1 32-144 mm), Monterey Bay, California, 36°48'N, 122°07'W; CAS 26405, 134 mm, Port Hueneme, California, 34°10'N, 1 19°10'W; CAS 26441, 138 mm, Gaviota, California, 34°05'N, 1 19°02'W; CAS 471 13, 2 (142, 147 mm), Point Baja, Califor- nia, 30°05'N, 115°58'W. LACM 320303, 141 mm, Bahia San Quintin, Mexico, LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT 37 FIGURE 18. Cladogram for selected agonid taxa based on a Wagner analysis of data presented in Tables 1 and 2. 30°18'N, 115°53'W; LACM 322463, 5 (120-155 mm), Santa Monica Bay, California, 33°54'N, 1 18°25'W. NMC 65-258, 147 mm, Kwatna Inlet, British Columbia, 52°25'N, 127°34'W. SIO 5 1 -255-56, 7 (88-1 5 1 mm), Channel Islands, California, 34°01'N, 119°24'W; SIO 6047156, 154 mm, Baja California Norte, Mexico, 31°18'N, 1 16°38'W; SIO 63104256, 8 (75-154 mm), Point Arguello, California, 34°10'N, 120°00'W. SU 16712, 3 (141-142 mm), Monterey Bay, California, 36°44'N, 121°58'W; SU 19172, 104 mm, Santa Barbara Chan- nel, California, 34°25'N, 120°06'W; SU 21363, 4 (127-148 mm), Point Pinos, California; 36°37'N, 121°55'W; SU 39780, 1 3 1 mm, Santa Barbara Island, California, 33°37'N, 1 1 9°05'W; SU 39781, 6 (105-134 mm), Santa Barbara Island, California, 34°25'N, 120°18'W. USNM 59370, 74 mm, Albatross station 3171, Russian Riv- er, California, 38°21'N, 123°20'W; USNM 63422, 129 mm, Point Soma, California, 32°41'N, 117°16'W; USNM 63423, 122 mm, Point Pinos, California, 36°38'N, 121°56'W; USNM 63427, damaged, Santa Cruz, California, 36°58'N, 122°01'W; USNM 103719, 136 mm, Mukilteo, Washington, 47°57'N, 122°18'W. UW 4175,117 mm, Saratoga Passage, Washington, 47°50'N, 122°30'W; UW 4725, 4 (106-159 mm), Richmond Beach, Washington, 47°50'N, 122'30'W; UW 20948, 9 (129-167 mm), Ballard, Washington, 47°30'N, 122°30'W. DESCRIPTION.— Two to six spine-bearing plates on eyeball; two, rarely three, barbels at posterior corner of maxilla; 12-14 pectoral rays (Table 5); branchiostegal membrane without a posterior free-fold; three barbels on ventral surface of den- tary, one at each posterior margin of three acous- tico-lateralis pores; dorsal fins unpigmented, blue spots present on head. DISTRIBUTION.— Gilbert (1890) described X. triacanthus from specimens captured at Alba- tross stations located off the coasts of California and Oregon, between approximately 34° and 45° N latitude. Material examined in this study ranged from Point Baja, Baja California, north to Kwat- na Inlet, British Columbia (Fig. 17). Gilbert (1915) reported X. triacanthus occurred in depths from 73 to 364 m; all lots examined for this study fell within that depth range. COMMENTS.— As was the case for X. latifrons, Gilbert (1890) did not designate a type-species for X. triacanthus. When the single specimen of 38 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 TABLE 6. CHARACTER STATES, NUMBER OF EVOLUTIONARY STEPS, AND THE LOCATION OF THE EVOLUTIONARY STEPS FOR THE DATA OF TABLES 1 AND 2 ON THE CLADOGRAM ILLUSTRATED IN FIG. 18. Num- ber of Branches where Character states steps the steps take place Spination Circumorbital 1 spines Present, absent Circumorbital 3 spines Present, absent Pterotic spines Present, absent 3 6, 9, Bathyagonus nigripinnis 1 10 2 2, Xeneretmus ritteri Posttemporal spines Present, absent 1 7 Parietal spines Present, absent 1 10 Frontal spines Present, absent 1 9 Nasal spines One, two 0 — Exposed mesethmoid spines Present, absent 0 — Preopercular spines Present, absent 1 10 2<, >2 2 5, 10 Spines on rostral plate Present, absent 2 9, 11 5, *5 1 3 3, *3 1 Xeneretmus triacanthus 1, *1 2 5, Odontopyxis trispinosa Spines on dermal body plates Present, absent 1 Free-fold of isthmus Present, absent 4 Maxillary barbels Exposed rostral plate Present, absent Breast plates abutting True, false Presence and arrangement of cheek plates Present, absent Abutting, not abutting Development of thickened fingerlike projection of ventral pectoral fin rays Present, absent Slightly developed, greatly developed Greatly developed First dorsal fin Present, absent 5,8, 11, Bathyagonus infraspinatus 4, 10 5, 11 1, 10 Aspidophor- oides olriki 9, Bathyagonus nigripinnis 7, Bathyagonus nigripinnis a lot now registered as USNM 43089 was trans- ferred to the United States National Museum by Gilbert and his associates, it was indicated in an accompanying letter that this specimen was soon to be described as the type of the species. That lot is hereby designated the lectotype ofX. tria- canthus for the same reasons cited above (see p. 34). PHYLETIC RELATIONSHIPS A Wagner analysis of the binary data set (Table 2) yields Figure 18 as the most parsimonious hypothesis for the cladistic history of the agonid taxa under consideration. This cladogram re- quires 37 steps for the 25 characters examined. Xeneretmus, Bathyagonus, and Aspidophoroides are all hypothesized to be monophyletic, as is the subgenus Xenopyxis. Bathyagonus is the sis- ter group of Xeneretmus; together they form the sister group of Odontopyxis, Bothragonus, and Aspidophoroides. The character states, the num- ber of steps each character takes on the tree, and where the steps take place are given in Table 6. The least derived genus Bathyagonus, requires only three steps from the base of the cladogram to the node that unites its members. The most derived group consists of Bothragonus and As- pidophoroides; there are eleven steps required from the root to the node that unites this group. Xenopyxis is intermediate between these two groups. There are eight evolutionary steps on the lineage leading from the base of the cladogram to the node that connects its three species. Monophyly for Xeneretmus is evidenced by the loss of spines on Circumorbital 1 . This char- acter state, however, is not unique to Xeneret- mus; four other species also lack spines on cir- cumorbital 1, and this loss is hypothesized to have occurred on three separate lineages (Table 6). Wagner analysis attempts to find the clado- gram that best fits the entire data set. This may result in hypothesized monophyletic sets that are not supported by unique unreversed characters, as it has for Xeneretmus. Their existence is hy- pothesized because any other arrangement of the taxa would be less parsimonious. Monophyly for the subgenus Xenopyxis is sup- ported by five synapomorphies. The arrange- ment of the breast plates is the one unique char- acter that unites the subgenus. These plates are separated in all members of Xenopyxis, whereas LEIPERTZ: A REVIEW OF THE FISHES OF THE AGONID GENUS XENERETMUS GILBERT 39 they abut in all other members of the in-group and the majority of the out-group. The other four characters are not unique to Xenopyxis: (1) Pos- session of one preopercular spine is hypothesized to be a reduction from the ancestral state of two preopercular spines; a reduction in the number of preopercular spines has also occurred on the lineage leading to Odontopyxis, Bothragonus, and Aspidophoroides. (2) The presence of a free-fold of the branchiostegal membrane across the isth- mus is hypothesized to be a reversal back to the primitive state, and one that has occurred in two other lineages, Aspidophoroides and Bathyagon- us infraspinatus. (3) The possession of only one spine on the rostral plate is a derived character state shared with Odontopyxis. (4) The absence of cheek plates is considered to be a reversal back to the primitive condition. ACKNOWLEDGMENTS Many people have contributed their ideas and energy to aid my work. Chief among them is Theodore W. Pietsch, my major advisor, whose support, advice, patience, and prodding have greatly improved this work and its author. Jo- seph Felsenstein generously spent many hours of his time discussing numerical phyletic tech- niques. J. Ebaugh, D. Futuyma, K. Howe, R. Nawojchik, and D. Nelson aided my study in many tangible, and many more intangible, ways. I thank the following people and their insti- tutions for providing material and information: W. Eschmeyer, M. Hearne, and P. Sonoda (CAS); T. Adamson and J. Seigel (LACM); W. Fink and K. Hartel (MCZ); J. Aniskowicz, J. Frank, and D. McAllister (NMC); C. Bond, J. Long, and S. Sauer (OSU); J. Pulsifer and R. Rosenblatt (SIO); J. Nelson and W. Roberts (UAMZ); and K. Bru- welheide, R. H. Gibbs, Jr., S. Jewett, and S. Weitzman (USNM). Susan Jewett's excellent de- tective work concerning the proper recognition of type material is greatly appreciated. Support from a Grant-in- Aid of Research from Sigma Xi, The Scientific Research Society, was very helpful. Finally I thank my wife and best friend Stella Chao for an enormous amount of help in data analysis as well as for supplying consideration and love. This is contribution No. 677 from the School of Fisheries, University of Washington. LITERATURE CITED BAIRD, R. C., AND M. J. ECKHARDT. 1972. Divergence and relationships in deep-sea hatchetfishes (Stemoptychidae). Sys. Zool. 21:80-89. BARNHART, P. S. 1 936. Marine fishes of Southern California. Univ. Calif. Press, Berkeley. 209 p. BARRACLOUGH, W. E., AND A. E. PEDEN. 1976. First records of the pricklebreast poacher (Stellerina xyosterna), and the cutfin poacher (Xeneretmus leiops) from British Columbia, with keys to the poachers (Agonidae) of the Province. Syesis 9:19-23. BOLIN, R. L. 1937. Notes on the California fishes. Copeia 1937:63-64. CLEMENS, W. A., AND G.V.WILBY. 1961. Fishes of the Pacific Coast of Canada. Bull. Fish. Res. Bd. Can. 68:1-443. COLLESS, D. H. 1981. Predictivity and stability in classifi- cations: Some comments on recent studies. Syst. Zool. 30: 325-331. ESCHMEYER, W. N., E. S. HERALD, AND H. HAMMANN. 1983. A field guide to Pacific Coast fishes of North America. Houghton Mifflin Co., Boston. 336 p. FARRIS, J. S. 1970. Methods for computing Wagner trees. Syst. Zool. 19:83-92. FARRIS, J. S., A. G. KLUGE, AND M. J. ECKHARDT. 1970a. On predictivity and efficiency. Syst. Zool. 19:363-372. . 19706. A numerical approach to phylogenetic sys- tematics. Syst. Zool. 19:172-189. FELSENSTEIN, J. 1973. Maximum likelihood and minimum- steps methods for estimating evolutionary trees from data on discrete characters. Syst. Zool. 22:240-249. . 1978. Cases in which parsimony and compatibility methods will be positively misleading. Syst. Zool. 27:401- 410. . 1 979. Alternative methods of phylogenetic infer- ences and their interrelationships. Syst. Zool. 28:49-62. . 1982. Numerical methods for inferring evolutionary trees. Quar. Rev. Bio. 57:379-404. FREEMAN, H. W. 1951. Contribution to the evolution and classification of the fishes of the family Agonidae. Ph.D. Dissertation, Stanford University. GILBERT, C. H. 1890. A preliminary report on the fishes collected by the Steamer Albatross on the Pacific coast of North America during the year 1889, with descriptions of twelve new genera and ninety-two new species. Proc. U.S. Nat. Mus. 13:49-126. . 1 904. Notes on fishes from the Pacific coast of North America. Proc. Cal. Acad. Sci. Ser. 3, 3(9):255-271. . 1915. Fishes collected by the United States Fisheries Steamer Albatross in southern California in 1 904. Proc. U.S. Nat. Mus. 48:305-380. . 1895. The ichthyological collections of the steamer Albatross during the years 1 890 and 1 89 1 . Rept. U.S. Comm. Fish. 19:393-476. GINN, T. C., AND C. E. BOND. 1973. Occurrence of the cutfin poacher, Xeneretmus leiops, on the continental shelf off the Columbia River mouth. Copeia 1973:814-815. GLOGER, C. W. L. 1842. Gemeinnutziges Hand- und Hilfs- buch der Naturgeschichte. 1:1-85. GRUCHY, C. G. 1969. Canadian records of the warty poacher, Occa verrucosa, with notes on the standardization of plate terminology in Agonidae. J. Fish. Res. Bd. Can. 26:1467- 1472. 40 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 3 HART, J. L. 1973. Pacific fishes of Canada. Bull. Fish. Res. Bd. Can. 180:1-740. ILINA, M. B. 1978. On the systematic status of the genus Podothecus Gill in the family Agonidae. Pages 1 3-24 in L. V. Shorisikova (ed.), Morphology and systematics of fishes (Collected scientific works). Utverzh. k pech. Zool. Inst. Akad. Nauk SSSR (in Russian), 90 p. JENSEN, R. J., AND C. D. BARBOUR. 1981. A phylogenetic reconstruction of the Mexican cyprinid fish genus Algansea. Syst. Zool. 30:41-57. JORDAN, D. S. 1903. Correspondence to the editor of the American Naturalist. Amer. Nat. 37:360. JORDAN, D. S., AND B. W. EVERMANN. 1898. The fishes of North and Middle America. Bull. U.S. Nat. Mus. 47(2): 1241-2183. JORDAN, D. S., B. W. EVERMANN, AND H. W. CLARK. 1930. Checklist of the fishes and fish-like vertebrates of North and Middle America north of the northern boundary of Vene- zuela and Columbia. Rep. U.S. Fish. Comm. Fisc. Yr. 1928, 670 p. KLUGE, A. G., AND J. S. FARRIS. 1 969. Quantitative phyletics and the evolution of anurans. Syst. Zool. 18:1-32. MICKEVICH, M. F. 1978. Taxonomic congruence. Syst. Zool. 27:112-128. . 1980. Taxonomic congruence: Rohlf and Sokal's misunderstanding. Syst. Zool. 29:162-176. MICKEVICH, M. F., AND J. S. FARRIS. 1981. The implications of congruence in Menidia. Syst. Zool. 30:351-369. MICKEVICH, M. F., AND M. S. JOHNSON. 1976. Congruence between morphological and allozyme data in evolutionary inference and character evolution. Syst. Zool. 25:260-270. MILLER, D. J., AND R. N. LEA. 1972. Guide to the Coastal Marine Fishes of California. Fish. Bull. 157:1-249. MIYAMOTO, M. M. 1983. Biochemical variation in the frog Eleutherodactylus bransfordii: geographic patterns and cryp- tic species. Syst. Zool. 32:43-51. NIE, N. H., G. HULL, M. FRANKLIN, J. JENKINS, K. SOURS, N. NORUSIS, AND V. BEACLE. 1980. SCSS: a user's guide to the SCSS conversational system. McGraw-Hill, New York. 595 p. NIE, N. H., G. HULL, J. JENKTNS, K. STEINBRENNER, AND D. BENT. 1975. SPSS: statistical package for the social sci- ences. McGraw-Hill, New York. 675 p. PEDEN, A. E., AND C. G. GRUCHY. 1971. First record of the blue spotted poacher, Xeneretmus triacanthus in British Co- lumbia. J. Fish. Res. Bd. Can. 28:1347-1348. PRESCH, W. 1980. Evolutionary history of the South Amer- ican microteiid lizards (Teiidae: Gymnophthalminae). Co- peia 1980:36-56. RENDAHL, H. 1934. Studien iiber die Scleroparei. I. Zur Kenntis der kranialen Anatomic der Agoniden. Ark. Zool. 26(3) pare 13:1-106. ROBINS, R. C. 1980. A list of common and scientific names of fishes from the United States and Canada. Am. Fish. Soc. Spec. Pub. 12:1-174. SCHUH, R. T., AND J. S. FARRIS. 1981. Methods for inves- tigating taxonomic congruence and their application to the Leptopodomorpha. Syst. Zool. 30:331-351. SCHUH, R. T., AND J. T. POLHEMUS. 1980. Analysis of taxo- nomic congruence among morphological, ecological, and biogeographic data sets for the Leptopodomorpha (Hemip- tera). Syst. Zool. 29:1-26. SIMON, C. M. 1979. Evolution of periodical cicadas: phylo- genetic inferences based on allozyme data. Syst. Zool. 28: 22-39. SOKAL, R. R., AND F.J. ROHLF. 1981. Taxonomic congruence in the Leptopodomorpha re-examined. Syst. Zool. 30:309- 325. SOKAL, R. R., AND P. H. A. SNEATH. 1963. Principles of numerical taxonomy. W. H. Freeman and Co., San Fran- cisco. 359 p. TAYLOR, W. R. 1967. An enzyme method of clearing and staining small vertebrates. Proc. U.S. Nat. Mus. 122:1-17. WEITZMAN, S. H. 1974. Osteology and evolutionary rela- tionships of the Sternoptychidae, with a new classification of stomiatoid families. Bull. Am. Mus. Nat. Hist. 153:327- 478. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 4, pp. 41-53, 4 figs., 1 table. January 3, 1986 SPECIES OF THE EMOIA SAMOENSIS GROUP OF LIZARDS (SCINCIDAE) IN THE FIJI ISLANDS, WITH DESCRIPTIONS OF TWO NEW SPECIES By Walter C. Brown Department of Herpetology, California Academy of Sciences, San Francisco, California 94118 and John R. H. Gibbons School of Natural Resources, University of the South Pacific, P.O. Box 1168, Suva, Fiji ABSTRACT: Emoia lizards, generally referred to as E. samoensis in the Fiji, Samoa, and Tonga islands, actually represent four distinct species, two of which are newly described. Emoia samoensis and Emoia murphy i are limited to the Samoa and Tonga islands; three species, Emoia concolor, E. trossula n. sp., and E. cambelli n. sp. occur in the Fiji Island group. A key to the species of the Emoia samoensis group in the Samoan, Tonga, and Fiji islands is provided. thought to be endemic to the Fijis. The recog- INTRODUCTION . ,, ., nized species were the apparently wide ranging The Fiji Island group in the South Pacific Ba- ones, E. caeruleocauda, E. cyanura, E. nigra, E. sin, about midway between Vanuatu (formerly samoensis, and possibly E. cyanogaster. New Hebrides) to the west and the Samoa and The present study is primarily concerned with Tonga islands to the east, is comprised of about those populations of the E. samoensis evolu- 320 islands. Several small, limestone and coral tionary line that occur in the Samoa, Tonga, and islands surround a number of larger, ancient, Fiji islands and that have generally been referred volcanic islands. The principal large islands are to the species E. samoensis. Other species of this Viti Levu, Vanua Levu, Taveuni, Kadavu, Ova- evolutionary line that have previously been rec- lau, Koro, Gau, Rabi, and Moala. ognized as distinct from E. samoensis are not The Fijis are the Pacific outpost for amphib- considered in detail, although some of them are ians, with two endemic species of the ranid genus included in Table 1 and the key at the close of Platymantis. Other terrestrial vertebrates in- this paper. elude a number of endemic species, and some The '(evolutionary) line of the genus Emoia endemic genera. Until recently, however, no includes species which range from relatively species of the scincid lizard genus Emoia were small, E. parkeri, to the largest in the genus, E. [41] 42 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 4 TABLE 1 . SCALE COUNTS AND OTHER PERTINENT CHARACTERS FOR SPECIES OF THE Emoia samoensis GROUP OF SPECIES IN THE SAMOA AND Fui ISLANDS.* Snout-vent Number Scale rows between length at in sample parietals maturity for scale Midbody and base Fourth toe (mm) counts scale rows of tail lamellae E. campbelli n. sp. (Fiji Islands) 68.9-97.8 13 30-36 56-65 44-56 E. concolor (Viti Levu Island) 61.3-78.9 16 28-32 56-60 44-52 E. concolor (other Fiji Islands) 52.8-88.9** 66 28-34 54-66 44-64 E. murphyi (Samoa and Tonga islands) 52.2-74.9 14 26-32 52-58 59-82 E. nigra 85.0-121.0 20 33-40 59-72 30-39 E. parkeri 44.5-53.8 18 28-32 52-59 34-43 E. samoensis (Samoa Islands) 78.0-118.0 21 30-34 58-68 45-54 E. trossula n. sp. (Fiji Islands) 66.0-103.0 37 32-38 62-76 52-54 * Four other species of Emoia occurring in these islands belong to other groups (evolutionary lines). ** This size range is based on 50 adults. One specimen measuring 100 mm from Vanisea, Kandavu Islands is referred to this species with some reservations. nigra. The genus may be characterized as follows: habitus varies from slender to fairly stout; snout moderately tapered and slightly to moderately depressed; subdigital lamellae usually broadly rounded (moderately thinned for two species), and number 30-82 under the fourth toe; number of midbody scale rows 26-42; number of para- vertebral rows between parietals and base of tail 52-84; frontoparietals fused; interparietal nearly always distinct, ranging from long and narrow to small; nasal bones separate; parietal eye present; alpha-type palate. This group of species ranges through the south Pacific Islands from Samoa in the east to Vanuatu and Bismarcks in the west. Emoia samoensis and E. concolor were the first species of this group to be described (Du- meril 1851). The type locality for samoensis was given as Samoa and that for concolor as Ambon, an island in the Moluccas to the west of New Guinea. The latter locality was not given in the original description, but subsequently by Jac- quinot and Guichenot (1853). Peters (1877) de- scribed another species of this complex, E. re- splendens, from the Fijis, stating that the type was in the Godeffroy Museum. Boulenger (1887) recognized the close rela- tionship of these three species and placed E. con- color and E. resplendens in the synonymy of E. samoensis. Most later authors (e.g., Werner 1899; Schmidt 1923; Burt and Burt 1932; Smith 1937; Brown 1956), followed this synonymy, assigning specimens from various islands between Samoa and Vanuatu to E. samoensis. Exceptions were Roux (1913) who described a race of samoensis from the Loyalty Islands and E. speiseri and E. nigromarginata from Vanuatu, Burt (1930) who described E. murphyi from Samoa, and Schmidt and Burt (1930) who described E. sanfordi from the Vanuatu and Solomons. Medway (1974) de- scribed yet another species, E. aneityumensis, from Vanuatu. At the same time, these authors as well as Medway and Marshall (1975) contin- ued to refer all Fiji specimens as well as examples from some populations in Vanuatu to E. sa- moensis. Brown (1953:20) recognized that the species E. concolor was distinct from E. samoensis, but without examining the types and assuming the type locality to be Ambon, erroneously suggested that the species might belong to the E. physicae group. Thus it was not until the 1970s that field work in the Fijis by several zoologists, J. C. Pernetta, D. Watling and W. Beckon among them, began to raise serious questions about the taxonomic status of the Fijian Emoia populations. Both Per- netta and Beckon pointed out the coexistence of uniformly colored populations and more typical dark-spotted or banded samoensis-like popula- tions on various islands in the Fiji group. Several of the uniformly colored specimens were com- pared with the type specimens of E. concolor by W. C. Brown and were judged to belong to the same species. This evidence, supported by the fact that no examples of the species other than the types have ever been recorded from Ambon, was interpreted as an indication that the types ofE. concolor were doubtless from the Fijis, and BROWN AND GIBBONS: SPECIES OF THE EMOIA SAMOENSIS GROUP 43 the locality Ambon was in error. In a subsequent paper, Pernetta and Watling (1979) listed both E. samoensis and E. concolor (the latter appar- ently endemic) as occurring in the Fijis, and not- ed differences in habit as well as color pattern but not other characters for the two groups of populations represented in their samples. Beck- on (personal communication) also recognized sa- moensis and concolor, suggesting that both could probably be divided into several island races or subspecies. Brown et al. (1980) described a sec- ond endemic Fijian species, E. parked, which is possibly related to E. nigromarginata from Van- uatu. The purpose of this study is to determine the status of those populations of Emoia (previously referred to E. samoensis) in the Samoa, Tonga, and Fiji islands. ACKNOWLEDGMENTS We are indebted to the following persons for their assistance in this study. For the loan of critical material, Alain Dubois, Museum Na- tional d'Histoire Naturelle (MNHN); Allen E. Greer and Harold Cogger, Australian Museum (AM); Pere Alberch, Museum of Comparative Zoology (MCZ); Richard G. Zweifel, American Museum of Natural History (AMNH); Robert F. Inger, Field Museum of Natural History (FMNH); George Zug and W. R. Heyer, United States National Museum (USNM); Edwin Ar- nold, British Museum of Natural History (BMNH); Fergus Clunie, Fiji Museum (FM); Hans Wilhel Koepcke, Zoologisches Museum, Hamburg (ZMH); Robert Drewes and Alan Lev- iton, California Academy of Sciences (CAS and CAS-SU); and the University of the South Pacific (USPM). John Campbell, consultant geologist with the Monasavu Hydro Electric Scheme pro- vided extensive assistance in the field work on the Rairaimatuku Plateau. We especially wish to thank Harold Cogger, John Pernetta, and Wil- liam Beckon for notes from their field observa- tions as well as for specimens. Photographs of Emoia concolor and E. trossula are by Harold Cogger. Allen Greer, Alain Dubois, Robert Drewes, and Richard Zweifel were most helpful with their observations and/or critique of the manuscript. The senior author was assisted by travel grants from the Australian Museum, the Science and Industry Endowment Fund of the Common- wealth Scientific and Industrial Research Orga- nization, Australia, and The Penrose Fund of the American Philosophical Society. Field work of the junior author was funded by the University of the South Pacific Research Committee. MATERIALS AND METHODS We have examined the types of Emoia sa- moensis, E. concolor, and E. murphyi as well as other examples from the Samoan and Tonga is- lands. In addition, relatively large samples of E. sanfordi from Vanuatu and populations on some of the Fiji Islands, as well as small samples (one to a few specimens) from other islands of the Fijis have also been studied. Data on size at ma- turity, size of eye, length of snout, length of limbs, variation in color patterns, and scale characters such as number of midbody scale rows, para ver- tebral scale rows between the parietals and the base of the tail, the number of lamellae beneath the fourth toe of the hind foot, and the pattern of the squamation of the head were determined. RESULTS Our analysis of populations in the Samoa, Tonga, and Fiji islands, which were generally referred to Emoia samoensis, have shown that these populations represent five distinct taxa. We treat them as separate species and provide de- tailed descriptions of E. samoensis and the Fijian species. Emoia samoensis (A. Dum6ril) Eumeces samoensis (part) Dumdril, 1851:157 (type loc.: Sa- moa; type in Mus6um National d'Histoire Naturelle, Paris); Jacquinot and Guichenot, 1853:10, in Hombron and Jac- quinot 1853. Emoia samoensis (part) Girard, 1858:265. Lygosoma samoense (part) Boulenger, 1887:293; Sternfeld, 1920:407; Boettger 1893:106; Boulenger, 1897:307. Emoia samoensis (part) Burt and Burt, 1932:531; Mertens, 1934:160; Smith, 1937:227; (part) Brown, 1956:1487; Mit- tleman, 1952:30; Greer, 1970:171. Emoia samoense, Schwaner, 1980:8. Dum6ril (1851:157) described Eumeces sa- moensis on the basis of two specimens in the Museum National d'Histoire Naturelle in Paris which were said to be from Samoa. He also at- tributed the name to Hombron and Jacquinot based on an illustration published earlier (some- time between 1845 and 1851). This illustration 44 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 4 is part of Plate V in the Atlas (1845-1853) which accompanied Jacquinot and Guichenot's text (1853). However, the plate bears only the French name Eumeces de Samoa and therefore does not establish publication of the valid scientific name. Scale counts and color (even in the faded con- dition) of the syntypes of Emoia samoensis, however, indicate that only one of them (MNHN 7070) is in close agreement with other examples from the populations in the Samoa Islands. And even though the head is damaged on one side, this specimen must be chosen as the lectotype. The second syntype (MNHN 7070a), the undam- aged one and the one apparently used for the illustration in Hombron and Jacquinot, even in its present faded condition, exhibits some dis- tinct white longitudinal dashes on the dorsum (prominent in the illustration), which are typical of most examples of the larger, previously un- described species from the Fijis. MATERIAL EXAMINED.— Samoa Is. (without definite locality): MNHN 7070 (syntype), 81; RMHN 3103; ZMC 373-374; BMNH 66.8.25.2; MCZ 3951, 8963. Upola Is.: CAS 157233; MCZ 69487; AMNH 29241, 29245; BMNH 1969.631-633; AM R1473. Tutiula Is.: CAS 50236-38; CAS-SU 13600-603, 18071; AMNH 27206, 27695, 27702, 27706. Savaii Is.: AMNH 41737-38, 41744. Tau Is.: AMNH 27668-73, 27675-76. Western Samoa: BMNH 1969.622-24, 1969.628-30. LECTOTYPE (NEW DESIGNATION).— MNHN 7070, collected on Samoa during the voyage of the Astrolabe and the Zelee, 1 837- 1840. DESCRIPTION OF LECTOTYPE.— Adult male, snout- vent length 105+ mm; head damaged but most head-shield characters recognizable; rostral broader than high, forming slightly curved suture with frontonasal; prefrontals in contact; frontal damaged but was in contact with first and second supraoculars; four large supraoculars; interpari- etal moderate size; anterior loreal about as long as posterior, in contact with first and second and probably third upper labials; sixth upper labial on right side enlarged and beneath eye; dorsal scales smooth, vertebral rows not distinctly en- larged; 32 midbody scale rows; 59 para vertebral rows between parietals and base of tail; 52 round- ed lamellae under fourth toe; 1 6 under first toe. COLOR (IN PRESERVATIVE).— Surface layer lost from most scales, but remaining undamaged scales indicate a dorsal pattern of olive green with blackish blotches. DEFINITION (BASED ON SAMPLE OF ABOUT 20 SPECIMENS).— A relatively large Emoia, snout- vent length 78-118 mm for 120 mature males and 84-114 mm for 80 mature females (data from Schwaner 1980:8); habitus moderately stout with well-developed limbs; snout moderately ta- pered, rounded at tip, its length 36-40% of head length and 56-68% of head breadth; head breadth 57-60% of head length and 13-16% of snout- vent length; eye moderate, its diameter 52-73% of snout length and 30-42% of head breadth; ear diameter % -% of eye diameter with three or four small lobules anteriorly; rostral broader than high, forming moderate, nearly straight suture with frontonasal; prefrontals in moderate contact (oc- casionally narrowly separated); frontal longer than broad, about as long as or slightly longer than fused frontoparietals, broadly rounded poste- riorly, in contact with first and second supra- oculars; four large supraoculars; six to seven su- praciliaries; interparietal moderately long and narrow to moderately wide; parietals in contact posteriorly; one pair of nuchals; anterior loreal shorter than to about as long as and higher than posterior loreal, in contact with first and second, second and third, or first, second, and third upper labials; six to eight upper labials, sixth (rarely fifth or seventh) largest and beneath eye; usually seven lower labials; scales smooth, paravertebral rows not enlarged or only slightly enlarged; 30- 35 midbody scale rows (very rarely greater than 34); 58-68 paravertebral rows between parietals and base of tail; limbs well developed, length of extended hind limb 86-1 10% of axilla-groin dis- tance and 46-54% of snout-vent length; 45-54 rounded lamellae beneath fourth toe; 13-16 la- mellae beneath first toe; rank of adpressed toes from longest to shortest four, three, two through five, one; tail longer than body. MEASUREMENTS (IN MM) OF LARGE FEMALE (CAS 50238).— Snout- vent length 107.0; axilla-groin distance 54.4; hind limb length 55.2; head length 25.6; head breadth 16.0; snout length 9.6; eye diameter 6.2; ear diameter 2.1. COLOR (IN PRESERVATIVE). — Dorsal ground color is greenish-tan to tan, marked with few to numerous dark brown spots varying from less than scale-size to vague, irregular transverse bands or partial bands involving several scales in transverse rows. Occasionally these show short whitish bars as in E. trossula. The venter is yel- lowish ivory to dusky tan. The top of head is not distinctly darker than the body. NOTE ON REPRODUCTION.— Schwaner (1980:8) states that the clutch size for 30 specimens ranges BROWN AND GIBBONS: SPECIES OF THE EMOIA SAMOENSIS GROUP 45 FIGURE 1 . Emoia concolor. from 4-7 eggs. He gives a snout-vent length of 3 1 mm for one hatchling. HABITAT NOTE.— Schwaner (1980:8) states that examples of this species were found primarily on tree trunks and low vegetation at heights from near ground level to several meters. This skink is diurnal. RANGE.— This species occurs in Samoa. Emoia concolor Dumeril (Figure 1) Euprepes concolor Dumeril, 1851:62 (type loc.: Fijis (?); type in Museum d'Histoire Naturelle, Paris). Emoa samoensis (part) Girard, 1858:264. Euprepes samoensis (?) Steindachner, 1867:44. Euprepes resplendens Peters, 1877:416. Lygosoma samoense (part) Boulenger, 1887:293. Lygosoma cyanogaster (part?) Boettger, 1893:106. Lygosoma cyanogaster tongana (part, Fiji Islands) Werner, 1899:375. Emoia samoense (part) Schmidt, 1923:52. Emoia samoensis (part) Burt and Burt, 1932:531; (part) Smith 1937:227; (part) Brown, 1956:1487; Greer, 1 970: 1 7 1 ; (part) Pernetta and Watling, 1979:236. Emoia concolor, Pernetta and Watling, 1979:236. As noted in the introduction, this species has long been regarded as a synonym of E. samoen- sis, the latter having been thought to range from Samoa to Vanuatu and to be the only species of the complex represented in the Fijis. Only re- cently has E. concolor been recognized as a valid species and name correctly applied to some Fi- jian populations. Peters (1877) based his brief description of E. resplendens on a single example from Ovalau Island. Several early specimens labeled L. sa- moense of the Godeffroy Museum collections are now in the Zoologisches Museum, Universitat Hamburg, so it is reasonable to assume that the type specimen of E. resplendens was also trans- ferred to that museum. However, Professor Hans Wilhelm Koepcke (personal communication) states that many of the herpetological types (in- cluding E. resplendens) and type catalogues were lost or destroyed during World War II. Thus the status of E. resplendens (Peters) must be based on the original description. Peters' s count for midbody scale rows (30) and note on color, "metalic gold luster with numer- ous, dark brown dots arranged in transverse lines," would indeed seem to identify this spec- imen as an example of E. concolor, since this color pattern is exhibited by some examples of E. concolor from various islands. Thus, basing our conclusion on Peters's data and color de- scription, we regard E. resplendens as a synonym of E. concolor. Werner's (1899:375) specimen 46 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 4 from the Fijis, which was referred to E. cyano- gaster tongana, is almost certainly an example of E. concolor although he states that there are only 26 midbody scale rows. Emoia concolor, though similar to E. sa- moensis in midbody scale rows, is a smaller species than either E. samoensis or E. trossula and typically has fewer scale rows between the parietals and the base of the tail, and more fre- quently exhibits a nearly uniform color on the dorsum. Emoia concolor has a wide range in the Fijis, and there is evidence that some island popula- tions differ sufficiently from one another to war- rant recognition as subspecies. However, pend- ing the availability of samples from as yet unexplored islands and larger samples from some already explored we hold such a decision in abey- ance. MATERIAL EXAMINED.— Fiji Is. (without definite locality): MNHN 7084, 7084a (syntypes); AM R6448a-b, 6449a-b, 6450a-c, 6451a-c; MCZ 9133, 9144; BMNH 55.11.7.24, 63.5.11.14-15, 75.12.31.10-13. Eastern Lau Group: AMNH 41750, 40195, 48058. Moala Is.: AMNH 41708. Kadavu Is.: BMNH 82.8.29.169-70; AM R30442, 30445; MCZ 15014, 1 6943-44; FMNH 3497; CAS 155974-85. Cikobia Is.: AMNH 29007. Vanua Levu Is.: BMNH 87.8.25.41; Pernetta coll. 252, 282. Viwa Is.: Pernetta coll. 117-119, 122-123. Yaduataba Is.: CAS 156002-04. Dravuni Is.: MCZ 16930-40; FMNH 3498a-d. Nagasan Is. (=Yagasa Is.?): MCZ 16947-48. Lami Is.: MCZ 48958. Viti Levu Is.: MCZ 16459; FMNH 62796, 69241,69639, 7 1764, 7 17.72; BMNH 1940.1.17.7,1945.11.5.9, 1947.3.1.86; Pernetta coll. 181, 167, 203, 205, 207-90; FM RA1,4, 12, 43; CAS 102361, 156136-37. Maluku Is.: USNM 230221-26. Kori Is.: USNM 230019-21. Ovalau Is.: FMNH 170716, 13641; USNM 230104-05. Taveuni Is.: BMNH 1959.1.2.32. Rotuma Is.: BMNH 97.7.29.8. Gau Is.: Watling coll. 501, 526, 543. Bird Is. (small island near Viti Levu): AM R 109939-43. LECTOTYPE.— MNHN 7084, collected during the voyage of the Astrolabe and the Zelee, 1837-1840. DESCRIPTION OF LECTOTYPE.— Male, 87 mm snout-vent length; rostral broader than high, forming long, nearly straight suture with front- onasal; prefrontals in moderate contact; frontal longer than broad, about as long as frontopari- etals and interparietal, in contact with first and second supraoculars; four large supraoculars; in- terparietal moderate in size; frontoparietals fused; one pair of nuchals; anterior loreal nearly as long as and not much higher than posterior, in contact with second upper labial and supranasal; sixth upper labial enlarged and under eye; some dorsal scales with three or four faint keels; paravertebral rows not enlarged or only slightly enlarged; 34 midbody scale rows; 60 paravertebral rows be- tween parietals and base of tail; 60 moderately rounded lamellae beneath fourth toe; 16-17 be- neath first toe; tail longer than body. COLOR (IN PRESERVATIVE).— Relatively uni- form yellowish olive-green on dorsum and upper lateral surfaces; lower lateral surfaces have bluish tinge and venter dirty ivory. The following definition, based on a series of more than 50 specimens, provides data on the variation exhibited by this species. DEFINITION.— An Emoia of moderate size, snout- vent length 52.3-88.7 mm for 22 males and 52.8-73.5 mm for 6 females (the sex of the syntype at 87 mm has not been determined, also a unique specimen, FMNH 3497, labeled from Vunisea on Kadavu Island and measuring about 100 mm, is referred to this species with some reservations); habitus moderately slender; snout moderately tapered, rounded at tip, its length 35- 42% of head length and 55-66% of head breadth; head breadth 52-68% of head length and 13- 17% of snout- vent length; eye diameter 52-70% of snout length and 30-45% of head breadth; ear diameter !/5-'/2 of eye diameter; ear usually with three small, rather pointed lobules anteriorly; rostral broader than high, forming a long, straight or slightly concave suture with frontonasal; su- pranasals long and narrow, in contact with an- terior loreal; prefrontals in moderate contact (rarely very narrowly separated); frontal longer than broad, usually longer than fused frontopari- etals, broadly rounded posteriorly, in contact with first and second supraoculars; four large supra- oculars; interparietal of moderate length and breadth; parietals in contact posteriorly; one pair of nuchals; anterior loreal somewhat shorter than to nearly as long as posterior and slightly higher, in contact with first and second, second, or first, second, and third upper labials; seven or eight upper labials, sixth (very rarely fifth) enlarged and beneath eye; usually seven lower labials; dor- sal scales smooth or with two or three weak stria- tions; paravertebral rows not enlarged or scarcely enlarged; 28-34 (see Table 1) midbody scale rows; 54-63 paravertebral rows between parietals and base of tail; limbs well developed; length of ex- tended hind limb 89-104% of axilla-groin dis- tance and 44-54% of snout-vent length; 46-68 rounded lamellae under fourth toe and 14-18 beneath first toe; rank of adpressed toes from longest to shortest four, three, two through five, one; tail longer than body. BROWN AND GIBBONS: SPECIES OF THE EMOIA SAMOENSIS GROUP 47 FIGURE 2. Emoia trossula. MEASUREMENTS (IN MM) OF MALE (AM 6448 A).— Snout- vent length 73.5; axilla-groin distance 34.8; hind limb length 36.0; head length 19.3; head breadth 11.7; snout length 7.4; eye diameter 4.8; ear diameter 2.0; tail broken. COLOR (IN PRESERVATIVE). — Dorsal ground color is nearly uniform greenish tan or with a few scattered brown spots or short bars to nearly solid, brownish longitudinal stripes (see Fig. 1); ground color more greenish yellow on lower lat- eral surfaces; venter more yellowish white or nearly turquoise (pale yellow to lime green in life); top of head and lower limbs often somewhat darker tan than rest of dorsum; sometimes yel- low spots on posterior surface of thighs; digits sometimes with dark brown, transverse bands. NOTE ON REPRODUCTION.— Three hatchlings measure 26.3-28.1 mm in snout- vent length. Gravid females have two eggs. HABITAT NOTE.— Emoia concolor is a lowland species from sea level to 500 m. It occurs both in the relatively open, intermediate-zone wood- lands and lowland forests, and in agricultural and suburban areas such as coconut and mango groves. RANGE.— This species is widely distributed in the Fiji Islands. Steindachner's 1869 reference to Stuart Island is presumably Unganga Island in the Fijis which some nineteenth century charts call Stuart. Emoia trossula n. sp. (Figure 2) Eumeces samoensis (part) Dumeril, 1851:157. Lygosomasamoense(part)Bo\dene,er, 1887:293; Werner, 1899: 375. Emoia samoense (part) Schmidt, 1923:52. Emoia samoensis (part) Burt and Burt, 1 932:53 1 ; (part) Brown 1956:1487; (part) Pernetta and Watling, 1979:236. This distinctively colored species from the Fi- jis has long been confused with E. samoensis. Even one of the two syntypes of the latter in the National Museum in Paris (MNHN 7070a), al- though indicated as being from the Samoa Is- lands in the catalogue, is an example of this species. It exhibits the color pattern and scale counts of the Fijian populations referred to this species (see also p. 44). HOLOTYPE.— AM R30433, an adult male, collected on Ova- lau Island, Fijis, 6 May 1970, by Harold G. Cogger. PARATYPES.— Fiji Is. (without specific locality): MNHN 7070a (one of syntypes of E. samoensis), 5573, 5573a-b; USNM 58155, 58166; AMNH 20927; AM R6446, R8566, A9463; BMNH 75.12.31.6, 62.10.23.4. Ovalau Is.: FMNH 13642, 48 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 4 13644-45; AM R30586, R30616-18; AMNH 40491. Yad- uataba Is.: USNM 230301; CAS 156128-29, 155960-62. Ka- davu Is.: AM R30446; AMNH 40489; BMNH 82.8.29.185. Aiwa Is.: AMNH 29917-22. Koro Is.: AMNH 40506. Thithia Is.: AMNH 40196. Moala Is.: AMNH 40223. Vatu Vara Is.: CAS 156130; AMNH 29010-1 1. Gau Is.: AMNH 40503. Buki Levu Is. (possibly part of Kadavu): MCZ 16945. Lakeba Lau Is.: MCZ 16965. Doi Lau Is.: MCZ 16941-42. Tuvuca Is.: AMNH 40539; BMNH 8 1 . 1 0. 1 2. 10. Namena Is.: BPBM 1 504; AMNH 40441-43, 40445. Taveuni Is.: CAS 155958; BMNH 1 938.8.2.9. Viti Levu Is.: ZMH RO 1 976-77. Rotuma Is.: BMNH 97.7.29.9-10. Gau Is.: Watling coll. 524-25, 541-42. Kia Is. (a small island near Viti Levu): R 1 1 6 1 60. BMNH 1860.3.18.8 and 1860.3.18.11, which were pur- chased from Mr. Cuming, are stated to be from Eumonga and Vanuatu. However, since they agree in most characters with examples from populations of E. trossula from the Fijis and not with samples from any of the species known from Vanuatu, we assume that the locality data are probably in error and have referred these two specimens to E. trossula, but have not in- cluded them in the paratypes. DIAGNOSIS.— This species differs from other species of the Emoia samoensis group in the fol- lowing combination of characters: (1) 32-40 (rarely less than 34) midbody scale rows; (2) 62- 76 paravertebral scale rows between the parietals and the base of the tail; (3) 42-54 lamellae under the fourth toe; (4) snout-vent length for adults 66.5-103.0 mm; (5) some features of color pat- tern that are generally present, such as short, narrow, greenish white longitudinal dashes on dorsum and upper lateral surfaces. These dashes are more or less in rows occupying the middle region of the concerned scales and are most prominent on the lighter, dorsal transverse bands and above the spaces between the dorsolateral dark blotches. DESCRIPTION. — A relatively large Emoia, snout- vent length 67.0-101.6 mm for 7 mature females and 66.5-103.0 mm for 14 mature males (2 specimens measuring 58.7 and 69.0 mm ap- pear immature); habitus moderately stout with well-developed limbs; snout moderately tapered, rounded at tip, its length 33-39% of head length and 51-65% of head breadth; head breadth 56- 60% of head length and 13-17% of snout- vent length; eye moderate, its diameter 60-81% of snout length and 34-48% of head breadth; ear diameter about Va-'/z of eye diameter, with three or four small, white lobules anteriorly; rostral broader than high, forming a moderate, slightly concave suture with frontonasal; supranasals slightly broader anteriorly than posteriorly, in contact with anterior loreal; prefrontals narrowly separated to moderately in contact; frontal longer than broad, about as long as fused frontopari- etals, rounded posteriorly, in contact with first and second supraoculars; four large supraoculars; interparietal relatively long and moderately nar- row; parietals in contact posteriorly; one pair of nuchals; anterior loreal slightly shorter than to nearly as long as posterior and slightly higher, usually in contact with first and second, second only, or first, second, and third upper labials; six to eight upper labials, sixth (rarely fifth or sev- enth) largest and beneath eye; usually seven low- er labials; scales smooth for adults, a hatchling with faint keels; middorsals only slightly en- larged; 32-38 (very rarely less than 34) midbody scale rows; 61-76 paravertebral rows between parietals and base of tail; limbs well developed, length of extended hind limb 96-109% of axilla- groin distance and 47-53% of snout- vent length; 43-54 rounded lamellae beneath fourth toe and 13-16 beneath first toe; rank of adpressed toes from longest to shortest four, three, two through five, one; tail longer than body. MEASUREMENTS (IN MM) OF HOLOTYPE, AN ADULT MALE.— Snout-vent length 94.3; axilla-groin distance 48.2; hind limb length 49.1; head length 22.4; head breadth 14.3; snout length 8.6; eye diameter 6.45; ear diameter 2.6; tail length 140. COLOR (IN PRESERVATIVE).— Ground color of the dorsum and upper lateral surfaces nearly uni- form medium brown to greenish olive-brown, or most often marked by irregular, lighter and dark- er transverse bands and by a series of dark blotches on the dorsolateral surface. Usually there are few to numerous, narrow, greenish white, longitudinal dashes occupying the median part of each affected scale. These are pri- marily on the lighter bands and above the dark, dorsolateral blotches, usually in longitudinal rows. The lower lateral surfaces are bluish gray fading into the bluish white of the venter which is lightly to densely spotted or flecked with small, black marks, at least posteriorly. ETYMOLOGY.— The name trossula is from the Latin meaning dandy, and refers to the colorful spotting like a brightly colored coat. COMPARISONS.— The number of paravertebral rows (62-76) is greater than that of other species of the E. samoensis group except for E. aneityu- mensis from Vanuatu which it barely overlaps. Also, the white dashes in the dorsal color pattern (nearly always present) are most prominent in this species. Boulenger (1 887:294), Werner (1 899: 375), and Burt and Hurt (1932:531) noted these white markings for some specimens but did not BROWN AND GIBBONS: SPECIES OF THE EMOIA SAMOENSIS GROUP 49 , FIGURE 3. Emoia campbelli. observe that they were primarily limited to Fiji specimens. Nor did they have a sufficiently large series to recognize that in addition to being di- visible on the basis of two color patterns, the Fiji samples also show bimodal curves for the num- ber of midbody scale rows and scale rows be- tween parietals and base of the tail. Emoia trossula is most closely related to E. samoensis. They are similar in size, but E. tros- sula differs in color pattern and has a greater number of midbody scale rows and a slightly higher number of para vertebral rows between the parietals and the base of the tail (see Table 1). Emoia sanfordi from Vanuatu is also a large species but differs in color pattern and has a much greater number of lamellae. Emoia concolor is smaller than E. trossula and has a lower number of midbody scale rows. NOTE ON REPRODUCTION.— Gravid females have two to five large eggs in the oviducts. One hatchling measures 32.3 mm from snout to vent. HABITAT NOTE.— This species is primarily a semi-arboreal forest form. On Yaduataba Island it was found both in the trees and on the forest floor. Some specimens were also seen asleep on open tree branches. Such habits as well as its relative boldness may have led to its extinction on many islands, possibly due to predation by mongooses and feral cats. Beckon's notes indi- cate that in inhabited areas on Taveuni Island this species was found in trees but on Kadavu Island this species was found in the forest, on or near the ground. On Gau Island E. trossula was found from the coastal areas up to an elevation of about 650 m in the rain forest. RANGE.— This species now has a patchy dis- tribution in the Fiji Islands and is almost cer- tainly extinct on the main islands of Viti Levu and Vanua Levu. The two specimens in the Hamburg Museum (ZMH RO 1976-77) that are recorded from Viti Levu were collected early in the nineteenth century. It is suggested that the introduction of the mongoose in 1887 may have led to this extinction. This theory would appear to be indirectly supported by the fact that E. nigra, a primarily terrestrial skink of about the same size as E. trossula, is also absent from Viti Levu but common on some other islands on which E. trossula still occurs. Thus far E. tros- sula has been recorded from the Fiji Islands. Emoia campbelli n. sp. (Figure 3) HOLOTYPE. — CAS 1 56256, an adult female collected by John Gibbons in the upper canopy of the cloud forest at Monsasavu 50 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 4 FIGURE 4. Section of "Ant plant" showing chambers in which eggs of Emoia campbelli were found. on the Rairaimatuku Plateau at an elevation of about 750 meters on Viti Levu Island in December 1983. PARATYPES.-CAS 155967-73, CAS 156257-58, CAS 156710-12 and USPM 46-48 from the same locality as the holotype; Viti Levu Island: ZMH R01978. DIAGNOSIS.— The species can be distinguished from other species of the E. samoensis group on the basis of the following combination of char- acters: (1) midbody scale rows 30-36 (rarely be- low 32); (2) scale rows between parietals and base of tail 56-64; (3) fourth toe lamellae 48-54; (4) snout- vent length at maturity 69-98 mm; (5) pre- frontals in relatively broad contact; and (6) color pattern: usually large, pale (yellowish in life) blotches along dorsolateral margin, separated by short, irregularly margined, blackish bars. (These spots are small or very faint in two specimens.) DESCRIPTION.— A moderate sized to relatively large Emoia, snout-vent length 70.4-97.8 mm for four males and 68.9-96.0 mm for five females (a female measuring 64.2 mm is apparently im- mature); habitus moderately stout with well-de- veloped limbs; snout rather strongly tapered, rounded at tip, its length 56-68% of head breadth and 35-42% of head length; head breadth 56- 70% of head length and 13-17% of snout- vent length; eye moderate, its diameter 56-72% of snout length and 33-43% head breadth; ear di- ameter 33-40% of eye diameter; rostral broader than high, forming long, nearly straight suture with frontonasal; supranasals narrowly triangu- lar, in contact with anterior loreal; prefrontals in moderately broad contact; frontal longer than broad, about as long as fused frontoparietals, in contact with first and second supraoculars; four large supraoculars; interparietal moderate; pa- rietals in contact posteriorly; one pair of nuchals; anterior loreal nearly as long as posterior, in con- tact with first and second, second and third, or first, second and third upper labials; seven or eight upper labials, sixth (rarely seventh) en- larged and beneath eye; six or seven lower labials; scales smooth; middorsal scales slightly enlarged; 30-36 (rarely less than 32) midbody scale rows; 56-64 para vertebral rows between parietals and base of tail; limbs well developed, length of ex- tended hind limb 88-107% of axilla-groin dis- tance and 45-52% of snout- vent length; 44-56 rounded lamellae under 5th toe; 14-17 under 1 st toe; rank of adpressed toes from longest to shortest four, three, two through five, one; tail longer than body. MEASUREMENTS (IN MM) OF HOLOTYPE.— Snout- vent length 97.8; axilla-groin distance 48.0; length of hind limb 45.6; head length 21.6; head breadth 13.8; snout length 8.9; eye diameter 5.0; ear diameter 1.7; tail length 136. COLOR (OF FRESHLY PRESERVED SPECIMENS).— Middorsal three or four rows of scales grayish to grayish olive green or light grayish brown, marked by black spots or dashes (in some specimens the black scales form either broken or very irregular transverse bands). Top of head is usually darker (slate brown), occasionally the same as the body; upper lateral surfaces usually marked by small to large (two to eight scales), yellow blotches al- ternating with black blotches between the nape and the groin (they are evident for some exam- ples of both sexes); lower lateral surfaces mottled grayish tan and bluish green marked by blackish flecks and dashes; venter bright sulfur yellow to greenish yellow, often with a blood red diffusion posteriorly, and on the base of the tail in life (fading in preservative); with small black dashes posteriorly, along midline, on preanals and sometimes base of tail. ETYMOLOGY.— This species is named for Mr. John Campbell, who collected the first example of this species in the Monasavu area. NOTE ON REPRODUCTION. —Eggs of this species have been found in the chambers of "ant plants." Figure 4 shows the chambers in a section of one of these plants. One gravid female has two large eggs. HABITAT NOTE.— Field observations by the ju- nior author indicate that this species uses as shel- ter primarily, if not exclusively, "ant plants" of the genus Hydnophytum, epiphytic in trees in the montane forests. It forages on the branches of BROWN AND GIBBONS: SPECIES OF THE EMOIA SAMOENSIS GROUP 51 the trees. Only one juvenile has been found on the ground. RANGE.— This species has thus far been found only on the Nadrau Plateau in the mountains of Viti Levu Island. COMPARISONS.— Emoia campbelli is probably most closely related to E. concolor and E. ni- gramarginata. It differs from both in color pat- tern, a slightly higher number of midbody scale rows, and a somewhat larger size. It also differs from E. nigromarginata in the somewhat greater number of lamellae. The morphological differences separating E. campbelli from E. concolor are not as great as those separating E. trossula from E. concolor, and we were at first inclined to regard this pop- ulation as a montane subspecies of E. concolor. However, E. campbelli, based on our available sample from the Monasavu area on the Rairai- matuku Plateau (part of the Nadrau Plateau), 750 to 1 ,200 m elevation, in the montane rain forest, is apparently strictly arboreal, at least in the adult stage. Also it seems to prefer the arboreal "ant plant" as a resting place and even deposits its eggs in cavities of that plant. Furthermore the montane forests of the plateau are effectively iso- lated from the lowland forest on three sides by high vertical cliffs and by a partially grassland corridor on the fourth side. Emoia concolor is a less specialized, primarily lowland species, which is at home in a variety of habitats, many of them much drier than the montane forest (see note on habitat for E. concolor). Because of its special- ized habitat preference and isolation, we treat E. campbelli as a distinct species. DISCUSSION Much larger samples were available to us than to earlier authors, and we were able to assess more accurately the limits of variation for many populations and therefore more clearly define species and determine their ranges. As stated in the introduction and the diagnostic key, the pop- ulations from the Fijis represent taxa distinct from E. samoensis in the Samoa Islands. Emoia samoensis generally attains a larger size than does E. concolor from the Fijis although these two overlap in scale counts and color pat- tern. E. trossula n. sp. is similar in size to E. samoensis and E. sanfordi from Vanuatu, but differs from these as well as from E. concolor in some scale counts (Table 1) and usually in some features of the color pattern. Emoia campbelli, the other new Fiji species, is thus far known only from a population on the Nadrau Plateau in the mountains on Viti Levu Island. For most species of the E. samoensis complex the dorsal color tends to vary, but the basic patterns are different. The least variation is characteristic of E. camp- belli and E. trossula. The latter only infrequently exhibits a nearly uniform brown or greenish olive brown color on the dorsum. Emoia concolor, as presently diagnosed, exhibits a uniform greenish color, or various patterns of brownish markings. Also, for E. concolor at least on Viti Levu, in- dividuals exhibiting a uniform pattern are found almost exclusively in the coastal and open wood- lands of the lowlands while those exhibiting varying density of dark spots, often bands or lines, on the dorsum occur primarily in the more dense lowlands and montane forests (see Per- netta and Watling 1979). These authors assigned these apparent color morphs to E. concolor and E. samoensis respectively. Emoia murphyi is closely related to E. con- color and E. trossula is closely related to E. sa- moensis. The following key and table set forth the di- agnostic characters and known ranges for the sev- en species included in this study. Populations in Vanuatu heretofore identified with E. samoensis must be reexamined to determine their true taxo- nomic status and relationship to other species of Emoia now recognized as occurring in Vanuatu and the Loyalty Islands. KEY TO THE SPECIES OF THE EMOIA SAMOENSIS GROUP IN THE SAMOA AND FIJI ISLANDS la. Midbody scale rows 26-36 (rarely greater than 34) 2 1 b. Midbody scale rows 32-40 (rarely less than 34) _ 3 2a. Fourth toe lamellae 30-39; snout-vent length at maturity 85-121 mm; interpa- rietal very small; dorsal and upper lateral surfaces dark brown to almost black, near- ly uniform, with scattered pale spots, or sometimes with vague, irregular trans- verse bands E. nigra 2b. Fourth toe lamellae 42-54; snout-vent length at maturity 66-103 mm for 20 specimens; interparietal long; dorsal color pattern greenish olive brown to medium brown usually with darker blotches or 52 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 4 bands and with few to numerous whitish dashes (Fiji Islands) E. trossula n. sp. 3a. Fourth toe lamellae 34-43; snout-vent length at maturity 46-54 mm; color pat- tern marked by a golden bronze head, a greenish bronze to greenish blue vertebral stripe about two scale rows in breadth; bordered by a blackish or dark brown band with scattered pale scales; a similar band on the upper labial surface (Fiji Islands) E. parked 3b. Fourth toe lamellae 44-83 (very rarely less than 45); snout- vent length at maturity 52-118 mm; color variable but not as above _._ 4 4a. Fourth toe lamellae 59-82 (rarely less than 63); color of dorsum dull grayish to gray- ish tan, darker posteriorly with a few vague dark and light spots, especially dorsolat- erally (known from Samoa and Tonga is- lands) _ E. m urphyi 4b. Fourth toe lamellae 44-66 (very rarely greater than 62); color of dorsum usually greenish tan, with or without darker markings 5 5a. Midbody scale rows 30-36 (very rarely fewer than 32, mean 33.1); color pattern of middorsal area greenish olive with scat- tered, dark and light spots or sometimes nearly complete, dark and light bands; dorsolateral area usually marked by a se- ries of large pale to yellowish blotches sep- arated by narrow, blackish bars (moun- tains of Viti Levu Island, Fijis) E. campbelli 5b. Midbody scale rows 28-35 (rarely greater than 32); color variable but not as above _ _ _ 6 6a. Scale rows between parietals and base of tail 54-63 (rarely greater than 60); snout- vent length at maturity 53-88 mm; color pattern of dorsum, four phases: (1) rela- tively uniform greenish to greenish tan, (2) greenish, marked by few to numerous brown to blackish spots, (3) dark spots in narrow longitudinal lines, or (4) occasion- ally marked by a series of pale and dark (in preservative usually brownish) more or less complete, transverse bands (Fiji Islands) E. concolor 6b. Scale rows between parietals and base of tail 56-68 (rarely less than 58); snout- vent length at maturity 78-1 1 8 mm; color pat- tern of dorsum: (1) nearly uniform green- ish to greenish tan; or (2) with scattered brownish to blackish spots, sometimes forming transverse bands, very rarely marked by whitish dashes (Samoa) E. samoensis LITERATURE CITED BOETTGER, O. 1893. {Catalog der Reptilien Sammlung im Museum der Senken— bergischer Naturforschender Gesell- schaft in Franfurt am Main, 1:X+ 140 pp. BOULENGER, G. A. 1887. Catalogue of lizards in the British Museum (Natural History), Vol. 3. Brit. Mus. Nat. Hist., London, i-xii+575 pp., 40 pis. . 1897. On the reptiles of Rotuma Island, Polynesia. Ann. Mag. Nat. Hist. (6)20:306-307. BROWN, W. C. 1953. Results of the Archibold Expeditions. No. 69. A review of New Guinea lizards allied to Emoia baudini and Emoia physicae. Amer. Mus. Novitates No. 1627:1-25. . 1956. The distribution of terrestrial reptiles in the islands of the Pacific Basin. Proc. 8th Pac. Sci. Congress 3: 1479-1491. BROWN, W. C., J. C. PERNETTA, AND D. WATLING. 1980. A new lizard of the genus Emoia (Scincidae) from the Fiji Islands. Proc. Biol. Soc. Wash. 93(2):350-356. BURT, C. E. 1 930. Herpetological results of the Whitney South Sea Expedition. IV. Descriptions of new species of lizards from the Pacific Islands (Scincidae). Amer. Nat. Hist. 43: 461-597. BURT, C. E. AND M. D. BURT. 1932. Herpetological results of the Whitney South Sea Expedition. VI. Pacific Island amphibians and reptiles in the collection of the American Museum of Natural History. Bull. Amer. Mus. Nat. Hist. 53:461-597. DUMERIL, A. M. C. 1851. Catalogue methodique de la Col- lection des Reptiles (Paris Mus.). 224 pp. GIRARD, C. 1858. Herpetology of the U.S. Exploring Expe- dition (1838-1842) under the command of Capt. Charles Wilkes, U.S.N. 20:1-496. GREER, A. E. 1970. A subfamilial classification of scincid lizards. Bull. Mus. Comp. Zool. Harvard 139:151-183. JACQUINOT, H. AND A. GUICHENOT. 1853. "Reptiles," in Hombron and Jacquinot, Voyage au Pole Sud et dans 1'Oceanie sur les Corvettes 1' Astrolabe et la Zelee. Zoologie 3:1-28. MEDWAY, LORD. 1974. A new skink (Reptilia: Scincidae: genus Emoia) from the New Hebrides. Bull. Brit. Mus. Nat. Hist. (Zool.) 27:53-57. MEDWAY, LORD AND A. G. MARSHALL. 1975. Terrestrial ver- tebrates of the New Hebrides: origin and distribution. Phil. Trans. R. Soc. London 272:423-465. MERTENS, R. 1934. Die Insel Retilien, ihre Ausbreitung, vari- ation und Artbildumg. Zoologica, Stuttgart 32:1-209. MITTLEMAN, M. B. 1952. A generic synopsis of the lizards of the subfamily Lygosominae. Smithsonian Misc. Coll. 1 1 7 (no. 17): 1-35. BROWN AND GIBBONS: SPECIES OF THE EMOIA SAMOENSIS GROUP 53 PERNETTA, J. C. AND D. WATLINO. 1 979. The introduced and native vertebrates of Fiji. Pac. Sci. (978)32:223-243. PETERS, W. 1877. Herpetologische Notizen. Monatsb. Akad. Wiss. Berlin 1877:407-423. Roux, J. 1913. Les reptiles de la Nouvelle-Caledonie et des lies Loyalty, in Sarasin and Roux, Nova Caledonia. Zoologie 1:79-152. SCHMIDT, P. 1923. A list of Fijian lizards. Copeia 1923:50- 52. SCHMIDT, P. AND E. BURT. 1930. Herpetological results of the Whitney South Sea Expedition. V. Description ofEmoia sanfordi, a new skink from islands in the western Pacific (Scincidae). Amer. Mus. Novitates, No. 436:1-3. SCHWANER, T. D. 1 980. Reproductive biology of lizards on the American Samoan Islands. Occas. Papers Mus. Nat. Hist. Univ. Kansas, No. 86:1-53. SMITH, M. A. 1937. A review of the genus Lygosoma (Scin- cidae: Reptilia) and its allies. Rec. Ind. Mus. 39:213-234. STEINDACHNER, F. 1867. Reptilien. In Reise der Novara um die Erde (1857, 1858, 1859), Zool. Thiel, Bd:l-98. STERNFELD, R. 1 920. Zur Tiergeographie Papuasiens und der pazifischen Inselevelt. Abh. Senkenb. naturf. Ges. 36:375- 436. WERNER, F. 1 899. Beitrage zur Herpetologie der Pazifischen Inselwelt und von Kleinasien. Zool. Anz. 22:371-378. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 5, pp. 55-66, 37 figs., 1 table. February 7, 1986 QUATERNARY BARNACLES FROM THE GALAPAGOS ISLANDS By Victor A. Zullo Department of Earth Sciences, University of North Carolina at Wilmington, North Carolina 28403 ABSTRACT: Quaternary marine deposits on six islands in the Galapagos Archipelago have yielded at least nine species of balanomorph barnacles, seven of which are present in the extant Galapagan fauna. The micl- intertidal species Tetraclita milleporosa Pilsbry and the lower intertidal to subtidal species Megabalanus galapaganus (Pilsbry) and Balanus trigonus Darwin are most common. The whale barnacle Coronula diadema (Linnaeus) and the turtle barnacle Chelonibia testudinaria (Linnaeus) are represented by unique specimens at separate localities. Balanus poecilus Darwin and a shell tentatively identified with K. calidus Pilsbry were found at one locality. A species of Concavus that might represent ('. (Arossia) panamensis eyerdami (Henry), and a Tetraclita shell bearing marked similarity to that of T. rubescens rubescens Darwin were present at single localities. Neither Concavus nor Tetraclita rubescens is known from the extant Galapagan fauna. INTRODUCTION As a participant in the 1964 Galapagos Inter- national Scientific Project, I had an opportunity to study the extant cirriped fauna of the Gald- pagos Archipelago, and to collect a few fossil barnacles from Cerro Colorado on Isla Santa Cruz. Some aspects of the extant fauna were pub- lished (Zullo 1966; Zullo and Beach 1973). The lack of adequate fossil material has prevented any serious speculation on the antiquity of the extant fauna and, indirectly, on the antiquity of intertidal and shallow-water habitats in the Ga- lapagos Archipelago. During February 1982, Carole Hickman, Mat- thew James, Jere Lipps, and Lois and William Pitt made an extensive survey of fossiliferous marine deposits in the Galapagos. Of the 84 sam- ples taken on seven islands, 1 2 localities on six islands contained barnacle remains (Fig. 36, 37). Lipps and Hickman (1982) argued that none of the Galapagan fossil localities is older than two million years, and that some types of deposit are only a few hundred years old. This conclusion is contrary to previous Miocene or Pliocene age estimates for several of these localities (e.g., Dall and Ochsner 1928; Durham 1964), but the com- pletely modern aspect of the fossil barnacle fauna would appear to support a Quaternary age as- signment. PALEONTOLOGY All of the species represented by fossils are either found today in the intertidal zone or at depths less than 20 m. The mid-intertidal species Tetraclita milleporosa Pilsbry, and the low in- tertidal zone and shelf species Megabalanus galapaganus (Pilsbry) and Balanus trigonus Dar- win are the most abundant fossils. The remaining species, including Balanus sp., cf. B. calidus Pils- bry, B. poecilus Darwin, Concavus (Arossia) sp., cf. C. (A.) panamensis eyerdami (Henry), Tetra- clita sp. indet., Chelonibia testudinaria (Lin- naeus), and Coronula diadema (Linnaeus) are represented by one or a few specimens from single localities. Quaternary distribution of barnacles mirrors modern distribution patterns. Tetraclita [55] 56 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 5 milleporosa and, particularly, Megabalanus gal- apaganus prefer high energy environments and are most abundant and grow to greatest size on windward sides of islands in areas of consider- able wave action. Balanus trigonus, on the other hand, prefers low energy environments on the leeward sides of islands or in protected areas below wave base. Balanus calidus and B. poe- cilus are most common in current-swept areas below wave base. Table 1 indicates the relationship between fos- sil barnacle occurrences and orientation of lo- calities with respect to prevailing wind direction. The majority of localities yielding specimens of Tetraclita milleporosa and Megabalanus gala- paganus are on southeast-facing shores that pres- ently bear the brunt of wave energies generated by the southeast trade winds. Balanus trigonus, on the other hand, is found predominantly at localities on west-facing, or present leeward shores. Only at CASG locality 61392 does B. trigonus occur with Megabalanus galapaganus. This apparent contradiction can be explained by the presence of Balanus sp., cf. B. calidus and B. poecilus, both subtidal species, and the small size of the Megabalanus galapaganus specimens, typical of subtidal populations. Locality 61392 probably represents a depositional environment below wave base with substantial current action. The two leeward Tetraclita localities are notable in that neither has yielded B. trigonus, suggesting that local wave energies were sufficient to main- tain Tetraclita populations, but too high to per- mit establishment of B. trigonus. ORIGIN OF THE GALAPAGAN BARNACLE FAUNA The major objective sought in this study, but not completely attained, was a clue to the time of origin of the Galapagan barnacle fauna. Clear- ly, the modern Galapagan fauna was already es- tablished in the Pleistocene, and its origins must be looked for in Neogene deposits, if such de- posits exist. This conclusion is supported by studies of north temperate and tropical eastern Pacific Cenozoic barnacle faunas. The major fau- nal break occurs at the Tertiary-Quaternary boundary, with the barnacles of the Pleistocene being essentially of modern aspect, whereas those of the Pliocene are primarily of extinct species- groups that evolved at the end of the Oligocene. The presence of Concavus cannot be adequate- ly explained. The two subspecies of Concavus (Arossia) panamensis (Rogers) range throughout much of the Panamic faunal province (Newman 1982). It is possible that the species has been overlooked in the extant fauna, or was elimi- nated from the fauna in the recent past. SYSTEMATICS Superfamily CORONULOIDEA Newman and Ross Family CORONULIDAE Leach Subfamily CHELONIBIINAE Pilsbry Genus Chelonibia Leach Chelonibia testudinaria (Linnaeus, 1767) Figures 3, 4 MATERIAL. — One lateral compartment, CASG locality 61281. DISCUSSION.— The single lateral plate in the collection is 3 1 mm high and has a basal width of 28 mm. The presence of deep cavities between basal septa and well-developed oblique grooves and ridges on the radial and alar edges of the paries readily identify this specimen with C. tes- tudinaria. This common and widely distributed turtle barnacle has been reported from the Pacific loggerhead, green, hawksbill, and ridley turtles. Fossils are known from Miocene and younger TABLE 1. DISTRIBUTION OF ENVIRONMENTALLY SENSITIVE GALAPAGAN FOSSIL BARNACLES WITH RESPECT TO LOCALITY ORI- ENTATION. Species Orientation of localities* E-facing SE-facing (windward) W-facing (leeward) 392 387 281 282 285 286 386 388 389 390 391 Tetraclita milleporosa Megabalanus galapaganus Balanus trigonus Balanus sp., cf. B. calidus Balanus poecilus X X X X XXX XX X X X ? X XXX * Locality numbers in table are last three digits of CASG numbers (e.g., 67392). ZULLO: GALAPAGOS QUATERNARY BARNACLES 57 FIGURES 1-1 1. Fig. 1, 2. Coronula diadema (Linnaeus, 1758), basal and side views of shell, hypotype CASG 61364, CASG locality 61229; xl.3. Fig. 3, 4. Chelonibia testudinaria (Linnaeus, 1767), lateral and interior views of lateral plate, hypotype CASG 61365, CASG locality 61281; x 1.6. Fig. 5-11. Tetraclita milleporosa Pilsbry, 1916. Fig. 5, 6. Exterior and basal views of shell, hypotype CASG 6 1 366, CASG locality 6 1 387; x 1 .6. Fig. 7, 8. Basal and exterior views of shell, hypotype CASG 61367, CASG locality 61386; x 1.6. Fig. 9. Interior of scutum, hypotype CASG 61368, CASG locality 61286; x2.7. Fig. 10. Interior of scutum, hypotype CASG 6 1 369, CASG locality 6 1 286; x 2.7. Fig. 1 1 . Exterior of shell rasped by fish, hypotype CASG 6 1 370, CASG locality 61282; xl.6. 58 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 5 deposits in Mediterranean, Atlantic, and Carib- bean regions (Zullo 1982). To my knowledge, this is the first reported fossil occurrence of Che- lonibia from the Pacific basin. Subfamily CORONULINAE Leach Genus Coronula Lamarck Coronula diadema (Linnaeus, 1767) Figures 1 , 2 MATERIAL. —One complete specimen, CASG locality 6 1 229. DISCUSSION.— The single specimen is 27 mm in height and its greatest diameter is 3 1 mm. The well-developed transverse corrugations on the external surfaces of the transverse flanges suggest the ornamentation seen in the Pliocene species C. barbara Darwin, but the absence of similar corrugations on the inner surfaces of the flanges and the lack of infilling between radii and the alar plates indicate that the Galapagos Coronula is merely a highly corrugated specimen of C. diadema. Coronula diadema, with a modern cosmopol- itan distribution on humpback, fin, blue, and sperm whales (Newman and Ross 1 976), has been reported from numerous Pliocene and Pleisto- cene localities in the Pacific basin region. Family TETRACLITIDAE Gruvel Subfamily TETRACLITINAE Gruvel Genus Tetraclita Schumacher Tetraclita milleporosa Pilsbry, 1916 Figures 5-1 1 Tetraclita porosa var. communis Darwin, 1854:329 (in part). Tetraclita squamosa milleporosa Pilsbry, 1916:257, pi. 60, fig. 1-ld; Newman and Ross 1976:48. MATERIAL.— One shell, CASG locality 61386; one shell, CASG locality 61387; one shell, CASG locality 61391; one shell, CASG locality 6 1 28 1 ; 2 1 shells, two partial shells, CASG locality 61282; 28 shells, four compartmental plates, six scuta, and one partial tergum, CASG locality 61286. DISCUSSION.— The extant tropical American taxa T. milleporosa, T. panamensis Pilsbry, T. stalactifera stalactifera (Lamarck), T. stalactifera confinis Pilsbry, and T. stalactifera floridana Pilsbry form a group within the genus Tetraclita characterized by similarities in shell coloration and opercular plate morphology that readily dis- tinguish them from other Tetraclita species and suggest close phylogenetic relationships. It is as- sumed that T. milleporosa, known only from the Galapagos Archipelago, was derived from a mainland T. stalactifera stock. In the eastern Pa- cific, subspecies of T. stalactifera are restricted to Panamic faunal province mainland localities. Tetraclita panamensis occurs along the Central American Pacific coast, but is also found on Bay of Panama islands. The intertidal Tetraclita of Cocos Island off the coast of Costa Rica appears to be conspecific with T. panamensis, but may represent a distinct subspecies. The opercular plates of T. milleporosa are sim- ilar to those of T. stalactifera, but the shell of the Galapagos species differs in being thicker and having much smaller and more numerous pari- etal tubes. Tetraclita milleporosa approaches T. panamensis in thickness and density of small pores, but differs particularly in opercular plate morphology. None of the fossils in the present collections shows any deviation from morphologies exhib- ited by extant T. milleporosa populations. The shells (Fig. 5-8) are typical of T. milleporosa in being peltate, and in having tiny orifices, obscure sutures, eroded external surfaces exposing in- filled parietal tubes, and thickened walls with very small and numerous parietal tubes. A few shells show evidence of rasping by fish (Fig. 1 1). The well-preserved scuta from CASG locality 61286 (Fig. 9,10) are typical as well, being about as high as wide, with small, closely set denticulae on the inflected occludent margin, and a rela- tively short adductor ridge that nearly merges with the lower part of the articular ridge, being separated by only a shallow groove. The tergum is too worn to be of aid in identification. Tetraclita sp. indet. Figures 12-14 MATERIAL.— One shell without opercular plates, CASG lo- cality 61286. DISCUSSION.— A single shell associated with numerous specimens of T. milleporosa, from CASG locality 61286, represents a second species of Tetraclita. The shell is high conic, with a rel- atively thin shell wall and correspondingly fewer rows of parietal tubes formed of larger individual tubes. The radii are narrow, but well developed and conspicuous, and the exposed filling of the upper parts of the parietal tubes is red. This shell is remarkably similar to that of T. rubescens ru- bescens which presently ranges between San Francisco, California and Cabo San Lucas, Baja California. The combination of shell features, ZULLO: GALAPAGOS QUATERNARY BARNACLES 59 FIGURES 12-21. Fig. 12-14. Tetradita sp. indet., lateral and basal views of shell, hypotype CASG 61371, CASG locality 61286; x2.5.Fig. 15, 16. Balanussp., cf. B. calidusPilsbry, 1916, top and lateral views of shell, hypotype CASG 61 372, CASG locality 61392; x2.5. Fig. 17-19. Balanus poecilus Darwin, 1854, CASG locality 61392; x2.5. Fig. 17, 18. Top and side views of shells, hypotype lot CASG 61373. Fig. 1 9. Lateral view of shell, hypotype CASG 61374. Fig. 20, 2 1 . Balanus trigonus Darwin, 1854, CASG locality 61388. Fig. 20. Top view of shells, hypotype lot CASG 61375; x2.5. Fig. 21. Shells on Anomia peruviana Orbigny, hypotype lot CASG 61376; x 1.6. particularly the color of the internal filling of the parietal tubes, is unlike that of T. milleporosa or any of the known Panamic faunal province species. The Panamic species, related to or con- specific with T. stalactifera, range in color from gray to purple-black, usually lack well-defined 60 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 5 radii, and usually have a conic shell with a small orifice. If this unique specimen is indeed repre- sentative of T. rubescens, I am at a loss to explain its presence in the Pleistocene Galapagan fauna. Superfamily BALANOIDEA (Darwin) Newman and Ross Family BALANIDAE Darwin Subfamily BALANINAE Darwin Genus Balanus Da Costa Balanus trigonus Darwin, 1854 Figures 20, 21 MATERIAL.— Ten shells, CASG locality 61388; two shells, CASG locality 61389; one shell, CASG locality 61390; six shells, CASG locality 61392. DISCUSSION.— Although shells characteristic of B. trigonus were collected at the four localities listed above, no opercular plates were present in the collections. Balanus trigonus is found in most of the warm- water regions of the world from the lower intertidal zone to the edge of the shelf, but is most common in those parts of the immediate subtidal and inner-shelf zones that are protected from wave shock. In the Galapagos Archipelago, extant B. trigonus is abundant on the leeward sides of islands in the lower intertidal and im- mediate subtidal. Considering the widespread distribution and abundance of this species in modern shallow seas, few verifiable reports of fossil B. trigonus exist. The species is fairly common in Pleistocene de- posits of the Gulf of California region (Ross 1962) and has been identified by William A. Newman (personal communication 1982) from the Pleis- tocene of Hawaii. To my knowledge, fossil B. trigonus has not been reported from the western margins of the Pacific basin. Western Atlantic reports of B. trigonus include those of Withers (1953) from the (?)Miocene of Cuba, and of Ross (1965) from the Pliocene Tamiami Formation of Florida. Ross's (1964) report of this species from the Pliocene Yorktown Formation of Vir- ginia was later stated to be in error (Ross 1965). Western Tethyan reports include those of Ko- losvary (1957) from the Tortonian (Miocene) of Hungary, and Davadie (1963) from the Pliocene of Italy, the Red Sea, and the Coralline Crag of England. These fossil occurrences, coupled with the modern distribution of the species, would suggest that B: trigonus is an old Tethyan element that has managed to survive to the present. There are two problems, however, that cause me to ques- tion this conclusion. First, many of the afore- mentioned reports are based on species lists without substantiating descriptions or illustra- tions. Their validity is placed in question par- ticularly in the knowledge that other students of barnacles who have monographed the faunas of the same regions (e.g., Darwin 1854; Alessandri 1906; Menesini 1966) did not uncover B. tri- gonus. Secondly, although both extinct and ex- tant representatives of the B. trigonus complex are common in Neogene and Pleistocene depos- its of southern California and the southeastern United States that I have examined (e.g., Zullo 1979), B. trigonus is absent. This is particularly odd, because the faunas of these units indicate that hydroclimates were substantially the same or warmer than those in the same regions today, and that depositional environments were fully within the present bathymetric range of B. tri- gonus. The origin and historical biogeography of B. trigonus remain in doubt, and their resolution will, in part, be dependent on a thorough eval- uation of previously reported occurrences. Balanus sp., cf. B. calidus Pilsbry, 1916 Figures 15, 16 MATERIAL.— One complete shell, CASG locality 61392. DISCUSSION.— A single shell, lacking opercular plates, is tentatively identified with B. calidus based on its coarsely ribbed, volcaniform shell and small orifice. Only a few extant specimens of B. calidus were obtained during the 1964 ex- pedition, and all came from shallow, subtidal depths. Off the East Coast of the United States, B. calidus is found on the shelf at depths below significant wave action. Balanus poecilus Darwin, 1854 Figures 17-19 Balanus poecilus Darwin, 1854:246, pi. 5, fig. 3a, b; Henry 1960:142, pi. 2, fig. a, c, d, pi. 5, fig. b-d. MATERIAL.— Eight shells without opercular plates, CASG locality 61392. DISCUSSION.— The "west coast of South Amer- ica, Mus. Cuming; attached to an Avicula" was cited by Darwin (1854) as the type and only lo- cality in his original description of B. poecilus. The species went unreported until Henry (1960) obtained some individuals of Pteria sterna (Gould) from the vicinity of Guaymas in the Gulf ZULLO: GALAPAGOS QUATERNARY BARNACLES 61 of California. Because of the unusually broad distribution indicated by the recorded occur- rences, and because of the ambiguity of the type locality, I requested the aid of J. P. Harding, British Museum (Natural History) in attempting to refine these data through identification of the "Avicula" to which the types are attached. Dr. Harding kindly located the type-lot and for- warded the following information provided by S. P. Dance (personal communication August 5, 1965): The shell to which the type specimens of Balanus poe- cilus Darwin are attached closely resembles a recently described species, Pteria beilana Olsson. The type lo- cality for this species is Venado Beach, Canal Zone, Pan- ama. Pteria peruviana Reeve may be an earlier name for this taxon but there is not enough material in the British Museum (Natural History) collections to decide this. Whichever name is used for it there can be little doubt that the shell to which the Darwinian barnacles are at- tached is a member of the Panamic-Pacific faunal prov- ince. Dr. Harding also reported that the specimens bear the label "West coast of America," rather than South America, and as it is known that Hugh Cuming made extensive collections on the west coast of Central America, and especially in Panama during the period 1832-1856 (Keen 1958:2), it seems likely that the types of B. poe- cilus are from the same region. Based on collections made during the 1964 Galdpagos expedition, and previously unre- ported specimens in the collections of the Allan Hancock Foundation and the California Acad- emy of Sciences, Balanus poecilus is found to range throughout the Panamic faunal province. The Allan Hancock Foundation collection in- cludes specimens from off San Pedro Nolasco Island in the Gulf of California, off Jicarita Island and Bahia Honda, Panama, off Gorgona Island, Colombia, and off La Libertad, Ecuador, as well as from the Galapagos off Gardiner Island, near Espanola. Genus Concavus Newman Subgenus Arossia Newman Concavus (Arossia) sp., cf. C. (A.) panamensis eyerdami (Henry, 1960) Figures 22-24 MATERIAL.— Two shells without opercular plates, CASG locality 61390. DISCUSSION. — The genus Concavus is not known to be represented in the extant Galapagan fauna. According to Newman (1982), modern representatives of this Tethyan Tertiary genus are restricted to the eastern Pacific, ranging from San Francisco, California to Valparaiso, Chile. Newman (1982) established two subgenera for extant species: Menesiniella for C. aquila (Pils- bry) and C. regalis (Pilsbry); and Arossia for C. henryae Newman, C. panamensis panamensis (Rogers), and C. panamensis eyerdami (Henry). The Galapagan fossils, with their plicate, but not regularly or strongly ribbed parietes, appear to be assignable to Arossia in the absence of the more definitive features of the opercular plates. Within Arossia, these fossil shells most closely approach those of C. panamensis eyerdami in having a high conic shell with the rostrum higher than wide, a straight carina, and no evidence of beaded growth lines. The preserved reddish-pur- ple coloration of the shell and the closely spaced transverse septa appear to distinguish the fossils from C. henryae, the sole representative of Con- cavus in the Peruvian faunal province. Subfamily MEGABALANINAE Newman Genus Megabalanus Hoek Megabalanus galapaganus Pilsbry, 1916 Figures 28-35 Balanus tintinnabulum galapaganus Pilsbry, 1916:70, pi. 12, fig. 1-lb. MATERIAL.— Eight shells, 5 shell fragments, CASG locality 61281; 2 partial shells, CASG locality 61282; 24 shells, 26 compartmental plates, two bases, four scuta, and one tergum, CASG locality 61286; 2 shells, CASG locality 61392. DISCUSSION.— After Tetraclita milleporosa, shells of Megabalanus galapaganus are the most abundant barnacle fossils obtained during the 1982 expedition. The specimens range from re- cently settled juveniles to mature individuals over 5 cm in height and 4 cm in greatest diameter. Many individuals retain the parietal color or col- or striping, and the parietal spines characteristic of extant populations. The opercular plates, al- though somewhat worn, are typical for M. gala- paganus. The scutum is flat, bears a well-defined adductor ridge, and lacks a definite lateral de- pressor muscle pit. The tergum has a longer and narrower spur than the closely related species M. clippertonensis (Zullo) from Clipperton Island and M. tanagrae (Pilsbry) from the Hawaiian Islands (Zullo 1969). Extant M. galapaganus is relatively abundant in low intertidal rocky areas subject to heavy 62 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 5 FIGURES 22-27. Fig. 22-25. Concavus sp., cf. C. (Arossid) panamensis eyerdami (Henry, 1960), CASG locality 6 1 390; x 1.6. Fig. 22, 23. Lateral views of shell, hypotype CASG 61377. Fig. 24. Basal view of shell, hypotype CASG 61378. Fig. 25. Broken radial sutural edge showing tubes, hypotype CASG 61378. Fig. 26, 27. Megabalanus sp. indet, lateral views of shell, hypotype CASG 61379, CASG locality 61285; x2.5. wave shock. It is in this region that the species reaches its maximum size. At subtidal depths specimens are locally abundant on lobster car- apaces, gastropod shells, and coral heads, but rarely attain more than 2 cm basal diameter and are usually less than 1 cm high. Pilsbry (1916) based this species on specimens from the inter- tidal of Espanola Island. Collections made during the 1964 expedition and augmented by collec- tions from the Allan Hancock Foundation and the California Academy of Sciences extend the range of M. galapaganus not only through most of the Galapagos Archipelago but to Cocos Is- land (Costa Rica) to the north and Port Utria, Colombia on the South American mainland. Megabalanus sp. indet. Figures 26, 27 MATERIAL.— One shell without opercular plates, CASG lo- cality 6 1285. ZULLO: GALAPAGOS QUATERNARY BARNACLES 63 FIGURES 28-35. Megabalanus galapaganus (Pilsbry, 1916). Fig. 28, 29. Exterior and interior of tergum, hypotype CASG 61380, CASG locality 61286; x 1.3. Fig. 30, 31. Exterior and interior of scutum, hypotype CASG 61381, CASG locality 61286; x 1.3. Fig. 32, 33. Interior and exterior of scutum, hypotype CASG 61382, CASG locality 61286; x 1.3. Fig. 34. Lateral view of shells, hypotype lot CASG 61383, CASG locality 61286; x 1.0. Fig. 35. Top view of shell clump, hypotype CASG 61384, CASG locality 6 1282; xl.O. DISCUSSION.— The single barnacle specimen from CASG locality 61285 differs sufficiently from the typical growth form of M. galapaganus to question its identification. The shell is 1 5 mm high, 22 mm in carinorostral diameter, and is low conic, rather than cylindric to subglobose in 64 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 5 91° GALAPAGOS ISLANDS 90° I. Pinta & I. Marchena I. Genovesa 61386 San Salvador I. Fernandina 61387^y> 61388- I. Rabida 61391 / I. Pinzoh 61229 l. Baltra ^^.61392 I. Santa Cruz see Fig.37 San Cristobal, Santa Fe 'jX/ Santa Maria r?N Espanola FIGURE 36. Generalized map of CASG Galapagos localities containing fossil barnacles (map provided by J. H. Lipps). shape. The sub-diamond-shaped orifice is, re- sultingly, rather small, and the radii are corre- spondingly narrow. The parietes show no evi- dence of color or striping, and bear low, rounded, irregular ribs. In the absence of opercular plates, however, there is no way to determine its iden- tity. LOCALITY DESCRIPTIONS All barnacle specimens are in the collection of the Department of Geology, California Academy of Sciences, San Francisco (CASG). Locations of collection sites are shown in Figures 36 and 37. 6 1 229 Isla Isabella. White to tan, loose, silty sand containing abundant shells at site of airport at Villamil. Collected 3 February 1982. Coronula diadema. 61281 Isla Santa Fe. Beach deposit about 8 m above sea level on southeast shore north of Punta Miedo. Calcareous, sometimes stratified sand up to 2.5 m thick and inter- mixed with basalt boulders and cobbles. Collected 14 February 1982. Chelonibia testudinaria, Tetraclita milleporosa, Megabalanus galapaganus. 6 1 282 Isla Santa Fe. Fossils from top of sedimentary sequence overlain by basalt. Red, tuffaceous, crossbedded sand- stone with stratified fossils at top, about 8 m above sea level. Same horizon as CASG locality 61281, but 30 m farther seaward. Collected 14 February 1982. Tet- raclita milleporosa, Megabalanus galapaganus. 61285 Isla Santa Fe. Terrace deposits of boulders, cobbles, pebbles, and sand containing molluscs and barnacles at top of cliff in small cove at landing site; 3.5-4 m above sea level. Locality is north of CASG locality 61282. Collected 15 February 1982. Megabalanus sp. indet. 61286 Isla Santa Fe. Terrace deposit about 100 m from shore near eastern end of south coast. Loose, white to tan, medium- to coarse-grained sand containing many bar- nacles. Collected 15 February 1982. Tetraclita mille- porosa, Tetraclita sp. indet., Megabalanus galapaga- nus. 61386 Isla San Salvad6r, James Bay. Shelly, basaltic sand in line of trees north of mining camp. Collected 8 Feb- ruary 1982. Tetraclita milleporosa. 61387 Isla Rabida. Storm-tossed shell and bone in small, cliff- ZULLO: GALAPAGOS QUATERNARY BARNACLES 65 0°50'S + 9005'W CAS 61285* CAS 61283 CAS 61284 300' ncj IJ^CdS 67282* ^~ & °AS 61281* CAS 61286* ISLA SANTA FE km contour interval in feet from U.S. Naval Hydrographic chart 5939 FIGURE 37. Collecting sites on Isla Santa Fe. CASG localities with asterisks yielded barnacles (map provided by J. H. Lipps). backed cove on south side of island. Collected 9 Feb- ruary 1982. Tetraclita milleporosa. 6 1 388 Isla Baltra. Bedded to crossbedded, reddish-brown, silty sandstone with abundant Codakia shells at basal con- tact (Unit 4). South shore of Caleta Aeolian, directly south of Punta Noboa. Collected 10 February 1982. Balanus trigonus. 61389 Isla Baltra. Crossbedded, white sandstone containing shell debris and abundant pectinids (Unit 1). Locality about 30 m east of CASG locality 61388 along a 1 00-m stretch of exposure. Collected 10 February 1982. Bal- anus trigonus. 61390 Isla Baltra. Basal 0.5-1. 5-m-thick boulder and cobble bed containing abundant coralline algae and casts and molds of molluscs. Same area as CASG locality 6 1 389. Collected 10 February 1982. Balanus trigonus, Con- cavus (Arossia) sp., cf. C. (A.) panamensis eyerdami. 61391 Isla Baltra. Bulldozed pit (?old anti-aircraft gun em- placement) about 170 m back of sea cliff. Collected 10 February 1 982. Tetraclita milleporosa. 61392 Isla Santa Cruz, Cerro Colorado. Fossils from top of limestone shelf on north side of Cerro Colorado. Col- lected 17 February 1982. Balanus trigonus, Balanus sp., cf. B. calidus, B. poecilus, Megabalanus galapaga- ACKNOWLEDGMENTS I thank Jere H. Lipps, Department of Geology, University of California, Davis, and William D. Pitt, Sacramento, California for providing the specimens, maps, and locality data used in this study, and for their advice during manuscript preparation. Funding for this study was provided by the Marine Sciences Research Program of the University of North Carolina at Wilmington. This paper is contribution number 372 of the Charles Darwin Foundation. LITERATURE CITED ALESSANDRI, G. DE. 1906. Studi monografici sui Cirripedi fossili d'ltalia. Palaeontogr. Italica 12:207-324. DALL, W. H. AND W. H. OCHSNER. 1928. Tertiary and Pleis- tocene Mollusca from the Galapagos Islands. Proc. Calif. Acad. Sci., ser. 4, 17:89-139. DARWIN, C. 1854. A monograph on the sub-class Cirripedia, the Balanidae, the Verrucidae. Ray Society, London. 684 pp. DAVADIE, C. 1963. Syst6matique et structure des Balanes d'Europe et d'Afrique. Editions Centre Natl. Recherche Scient. 146 pp. DURHAM, J. W. 1964. The Galapagos Islands expedition of 1964. Am. Malacol. Union, Inc. Annu. Rep. 1964:53. HENRY, D. P. 1960. Thoracic Cirripedia of the Gulf of Cal- ifornia. Univ. Washington Publ. Oceanogr. 4(4): 135-1 58. KEEN, M. A. 1958. Seashells of tropical west America. Stan- ford Univ. Press, Calif. 624 pp. 66 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 5 KOLOSVARY, G. 1957. Enumeration des Balanides fossiles de la Hongrie. Bull. Soc. Linn. Lyon 26(2):30-32. LINNAEUS, C. 1758. Systema naturae. Holmiae, Editio De- cima, Reformata, Vol. 1 . 824 pp. . 1767. Systema naturae per regnatria naturae— editio duodecima, reformata. Holmiae 1(2):533-1327. LIPPS, J. H. AND C. S. HICKMAN. 1982. Paleontology and geologic history of the Galapagos Islands. Abs. with Pro- grams, Geol. Soc. Am. 14(7):548. MENESINI, E. 1966. I balani Miocenici delle "Arenarie di Posano" (Volterra, Provincia di Pisa). Palaeontogr. Italica 60:99-129. NEWMAN, W. A. 1 982. A review of extant taxa of the "Group ofBalanus concavus" (Cirripedia, Thoracica) and a proposal for genus-group ranks. Crustaceana 43:25-36. NEWMAN, W. A. AND A. Ross. 1976. Revision of the ba- lanomorph barnacles; including a catalog of the species. San Diego Soc. Nat. Hist. Mem. 9:1-108. PILSBRY, H. A. 1916. The sessile barnacles (Cirripedia) in the collections of the U.S. National Museum; including a mono- graph of the American species. U.S. Natl. Mus. Bull. 93:1- 366. Ross, A. 1962. Results of the Puritan-American Museum of Natural History expedition to western Mexico, 15. The lit- toral balanomorph Cirripedia. Am. Mus. Novitates 2084: 1-44. . 1964. Cirripedia from the Yorktown Formation (Miocene) of Virginia. J. Paleontol. 38:483-491. . 1965. A new barnacle from the Tamiami Miocene. J. Fla. Acad. Sci. 27(4):273-277. WITHERS, T. H. 1953. Catalogue of fossil Cirripedia in the Department of Geology, Vol. III. Tertiary. British Mus. (Nat. Hist.), London. 396 pp. ZULLO, V. A. 1966. Zoogeographic affinities of the Balano- morpha (Cirripedia: Thoracica) of the eastern Pacific. Pp. 139-144 in The Galapagos, R. I. Bowman, ed. Univ. Cal- ifornia Press, Berkeley. . 1969. A new subspecies of Balanus tintinnabulum (Linnaeus, 1758) (Cirripedia, Thoracica) from Clipperton Island, eastern Pacific. Proc. Calif. Acad. Sci., ser. 4, 36(16): 501-510. . 1979. Thoracican Cirripedia of the lower Pliocene Pancho Rico Formation, Salinas Valley, Monterey County, California. Contrib. Sci., Nat. Hist. Mus. Los Angeles Co. 303:1-13. 1982. A new species of the turtle barnacle Chelonibia Leach, 1817, (Cirripedia, Thoracica) from the Oligocene Mint Spring and Byram formations of Mississippi. Mississippi Geol. 2(3): 1-6. ZULLO, V. A. AND D. B. BEACH. 1973. New species of Mem- branobalanus Hoek and Hexacreusia Zullo (Cirripedia, Ba- lanidae) from the Galapagos Archipelago. Contrib. Sci., Nat. Hist. Mus. Los Angeles Co. 249:1-16. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94 1 1 8 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 6, pp. 67-109, 28 figs., 2 tables. February 7, 1986 A SYSTEMATIC REVIEW OF AMPHIZOID BEETLES (AMPHIZOIDAE: COLEOPTERA) AND THEIR PHYLOGENETIC RELATIONSHIPS TO OTHER ADEPHAGA By David H. Kavanaugh Department of Entomology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: Rediscovery of type material for Amphizoa davidi Lucas, 1882, the only known Palaearctic amphizoid, is reported, with a lectotype designated, and the type area emended (from Tibet to Szechwan Province, China). A key is provided for identification of adults of the four known amphizoid species. Form and structure, geographical and habitat distributions, and geographical relations with other taxa are described and illustrated for each species. Amphizoa carinata Edwards, 1951, is recognized as a junior synonym of A. lecontei Matthews, 1872. Through cladistic analysis, using out-group and character correlation criteria, and a review of known Mesozoic fossil material, a hypothesis of phylogenetic relationships among extinct and extant Adephaga is developed, discussed, and related to geologic time. A semi-aquatic, rather than terrestrial, common ancestor is proposed for Adephaga. Amphizoids diverged from their sister-group, which includes all Hydradephaga except haliplids, in Triassic time. AH extant amphizoid species had differentiated by late Pliocene time, in response to a series of vicariant events. Quaternary climatic and geologic events resulted in changes in geographical distributions of these species and structural, physiological, and behavioral adapta- tions of their members. TABLE OF CONTENTS Evidence for relationship between amphizoids and trachypachids 85 A hypothesis of adephagan phylogeny 8 9 Introduction 68 Phylogenetic relationships Materials and Methods 69 of amphizoid species 98 Systematics of Amphizoidae 69 Zoogeography and Evolution 100 Introduction 69 Present pattern of amphizoid A Key for Identification of distribution 100 Amphizoa Adults 70 Mesozoic events and the Amphizoa davidi Lucas 70 origin of amphizoids 101 Amphizoa insolens LeConte 72 Tertiary events and Amphizoa striata Van Dyke 75 amphizoid radiation 102 Amphizoa lecontei Matthews 75 Quaternary history and development Phylogeny 78 of the present amphizoid fauna 105 Phylogenetic relationships of Prospectus for Future Research 107 amphizoids 8 1 Acknowledgments 1 07 Evidence from the fossil record 84 Literature Cited 107 [67] 68 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 INTRODUCTION For several decades, the location of type-ma- terial for Amphizoa davidi Lucas, 1882, de- scribed from Tibet, remained a mystery (Ed- wards 1951; Kavanaugh 1980; Kavanaugh and Roughley 1981). Although material Lucas stud- ied was known to have been deposited in the Museum National d'Histoire Naturelle in Paris, several independent efforts to locate specimens of A. davidi in appropriate parts of that collection had failed. Equally perplexing was the fact that no additional specimens representing this taxon had been found since its original description. Amphizoa davidi is an especially important taxon for two reasons. First, it is a member of the small family Amphizoidae, which is consid- ered by many workers to represent an interme- diate evolutionary grade between the so-called Geadephaga, or terrestrial Adephaga (i.e., Ca- rabidae, in the broadest sense), and the remain- ing Hydradephaga, or aquatic Adephaga (i.e., Dytiscidae, Hygrobiidae, Gyrinidae, etc.). + , FIGURE 1. Amphizoa davidi Lucas: lectotype male, dorsal aspect, total length = 1 1.4 mm. Knowledge of amphizoids is seen as a major key to understanding adephagan evolution and phy- logeny; and knowledge of A. davidi in particular is critical for understanding Amphizoidae. Sec- ond, Amphizoa kashmirensis Vazirani, 1 964: 1 45, described from the Himalaya of India, has re- cently been shown to be a dytiscid, referable to genus Hydronebrius Jakovlev, rather than an am- phizoid (Kavanaugh and Roughley 1981). As a result, A. davidi is the only known Palaearctic amphizoid; and because no specimens of this species had ever been seen by current workers, doubts had arisen with regard to its familial af- finities (Kavanaugh 1980; Kavanaugh and Roughley 1981). What are the phylogenetic re- lationships between A. davidi and the Nearctic species, and what are the zoogeographic impli- cations of this phylogeny and the disjunct dis- tribution of genus Amphizoal Answers to these questions might shed new light on the origins and history of the Holarctic fauna in general and of certain relict, taxonomically isolated taxa in particular. In early 1983, Terry L. Erwin (U.S. National Museum, Washington, D.C.) discovered several amphizoids pinned in one corner of a Schmidt box labelled "Australian Carabidae" at the Mu- seum National d'Histoire Naturelle in Paris. In- cluded were a few specimens of Amphizoa in- solens LeConte from western North America and one specimen (Fig. 1) from Mou-pin, Tibet, the type-locality for A. davidi. Suspecting that he had found the long-sought type of A. davidi, Erwin arranged for shipment of the specimen to me on loan. Jean Menier, curator at the museum in Paris, provided photocopies of relevant entries in the museum's catalog, specifically for the accession of material from Mou-pin, Tibet, re- ceived from Armand David and upon which Lu- cas's description was based. Subsequently, I have determined that the specimen is the type-speci- men of Amphizoa davidi Lucas through a study of the specimen itself and the labels it bears (in- cluding one with the proper catalog number). The purposes of this paper are: ( 1 ) to report on the rediscovery of type-material for A. davidi Lucas; (2) to designate a lectotype for same; (3) to redescribe this material in comparison with Nearctic forms, and illustrate certain character- istics of form and structure for the first time; (4) to update distributional records that have ac- cumulated since Edward's (195 1) revision of the family; (5) to propose one new synonymy; (6) to KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 69 provide a revised key to species that reflects new findings; and (7) to initiate consideration of the phylogenetic relationships among extant Am- phizoa species and the zoogeographic implica- tions of these relationships. A cladistic analysis of relationships among major extant and known extinct adephagan groups is presented as a basis for the intrageneric analysis of Amphizoa. MATERIALS AND METHODS Descriptions of form and structure, taxonomic conclusions, geographical distributions, and oth- er findings reported here are based on exami- nation of more than 1,250 adult specimens of Amphizoa and more than 300 specimens rep- resenting other extant adephagan taxa. The fol- lowing acronyms are used in the text to refer to collections from which specimens were received for study and/or in which specimens are depos- ited. Curators and collecuon managers respon- sible for these collections are also listed, and I thank them sincerely for their help in providing specimens on loan for study. BMNH British Museum (Natural History), London SW7 5BD, England; M. E. Bacchus. BYUM Brigham Young University Museum of Natural His- tory, Provo, Utah 84602; R. W. Baumann. CAS California Academy of Sciences, San Francisco, Cal- ifornia 941 18; D. H. Kavanaugh. CNC Canadian National Collection of Insects, Biosys- tematics Research Institute, Ottawa, Ontario Kl A OC6; A. Smetana. DMad D. Maddison, University of Alberta, Edmonton, Al- berta T6G 2E3. GCha G. Challet, Orange County Vector Control District, Garden Grove, California 92643. GLPa G. L. Parsons, Oregon State University, Corvallis, Oregon 97331. GLPe G. L. Peters, Oregon State University, Corvallis, Or- egon 97331. LGBe L. G. Bezark, California State Department of Food and Agriculture, Sacramento, California 95814. MCZ Museum of Comparative Zoology, Harvard Univer- sity, Cambridge, Massachusetts 02 1 38; A. F. Newton, Jr. MNHP MusSum National d'Histoire Naturelle, Paris, 75005 France; J. Menier. NSDA Nevada State Department of Agriculture, Reno, Ne- vada 89504; R. C. Bechtel. OSUO Oregon State University, Corvallis, Oregon 9733 1 ; J. Lattin, G. L. Peters. PUCA Pacific Union College, Angwin, California 94508; L. E. Eighme. RERo R. E. Roughley, University of Manitoba, Winnipeg R3T 2N2. SJSU San Jose State University, San Jose, California 95 1 14; J. G. Edwards. UASM University of Alberta, Strickland Museum, Edmon- ton, Alberta T6G 2E3; G. E. Ball. UCB University of California, Essig Museum of Entomol- ogy, Berkeley, California 94720; J. A. Chemsak and G. Ullrich. UCD University of California, Davis, California 956 1 6; R. O. Schuster. USNM United States National Museum, Smithsonian Insti- tution, Washington, D.C. 20560; P. J. Spangler, T. L. Erwin. UZMH Universitetets Zoologiska Museum, Entomologiska Avdelningen, SF-00100 Helsingfors 10, Finland; H. Silfverberg. Methods, including techniques for dissection of male and female genitalia and criteria for rank- ing taxa, are discussed by Kavanaugh (1979). The only measurement used in this paper, stan- dardized body length (SBL), is the sum of three measurements: length of head along midline from apical margin of labrum to a point opposite pos- terior margin of left eye; length of pronotum along midline from anterior to posterior margin; and length of elytron along midline from apex of scu- tellum to a point opposite apex of longer elytron. Line drawings were made with the aid of a camera lucida attached to a Wild Model M-5 stereoscopic dissecting microscope. The scan- ning electron micrograph (Fig. 2) was obtained using a Hitachi model S-520 SEM (with accel- erating voltage = 5 kV and specimen uncoated). Cladistic analyses were carried out using man- ual methods (see Phylogeny below for further discussion); but results were compared with those generated using "WAGNER" and "SOKAL" programs from the "Phylogenetic Inference Package" (PHYLIP) for microcomputers created by J. Felsenstein (University of Washington, Se- attle), as modified by T. K. Wilson (Miami Uni- versity, Oxford, Ohio). In general, cladograms obtained using manual and computer-assisted methods were similar. However, placement of individual taxa in cladograms generated by the PHYLIP programs varied markedly, subject to changes in the order in which taxa were listed in the database (and, therefore, compared by com- puter). SYSTEMATICS OF AMPHIZOIDAE Introduction Edwards's (1951) monograph of Amphizoidae stands as the definitive systematic treatment of this group. His extensive review of the literature and detailed descriptions, comparative studies, and discussions of form and structure serve as a sound basis for all subsequent work on amphi- 70 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 zoids, as well as for comparisons of members of this group with those of other adephagan taxa. Because of more liberal institutional lending policies, I was able to borrow type-material that was unavailable to Edwards and designate lec- to types for Amphizoa insolens LeConte, A. jo- sephi Matthews, and A. lecontei Matthews (Ka- vanaugh 1980). Moreover, a great deal of new material has been collected during the past 30 years. Through loans and my own fieldwork, I have had access to almost five times as many specimens as Edwards studied, many of these from areas in which amphizoids were previously unknown. These new distributional records have important taxonomic and zoogeographic impli- cations. In order to make this report minimally redundant with respect to Edwards's paper, I have limited my descriptive presentations to brief list- ings and discussions of distinguishing character- istics, except where my findings depart from Ed- wards's. The reader should consult Edwards (1951) for more detailed descriptive and com- parative information on amphizoids, as well as comprehensive coverage of the literature prior to that date. The format used for presentation of Amphizoa species below is as follows: (1) a synonymy (in- cluding for each name the author, date, and page citation for original description; status, sex, and depository for holotype or lectotype; type-local- ity; and literature citations that were not listed by Edwards [195 1]); (2) additional comments on nomenclature, type-specimens, and/or type- locality as needed; (3) a brief listing of distinguish- ing characteristics of adults, with additional dis- cussion of form and structure as needed; (4) habitat distribution; (5) geographical distribu- tion, including distributional summary state- ment, map illustrating known localities, and for- mal listing of localities for specimens studied (with area and month[s] of collection, number of spec- imens studied, and depositories] for same); and discussions of (6) geographical variation; and (7) geographical relationships with other Amphizoa species. A Key for Identification of Amphizoa Adults 1 . Elytron (Fig. 6b) with blunt but distinct carina on fifth interval, area medial to carina elevated, flat, area lateral to ca- rina slightly concave Amphizoa lecontei Matthews 1'. Elytron (Fig. 3b, 4b, 5b) evenly convex or slightly concave paralaterally, with- out carina 2 2(1'). Prosternal intercoxal process (Fig. 11) short, round; body form narrower (Fig. 3a); specimen from southwestern China (Fig. 17) Amphizoa davidi Lucas 2'. Prosternal intercoxal process (Fig. 12) long, spatulate; body form (Fig. 4a, 5a) relatively broader; specimen from west- ern North America 3 3(2'). Elytral silhouette (Fig. 5a) broad basally and distinctly narrowed subapically, elytral surface only faintly rugose in lat- eral one-half; pronotum (Fig. 9) broad- est at base, with lateral margins not or only slightly crenulate Amphizoa striata Van Dyke 3'. Elytral silhouette (Fig. 4a) subovoid, slightly narrowed basally, slightly broader subapically, elytral surface moderately or coarsely rugose in lateral one-half; pronotum (Fig. 8) at least as broad at middle as at base, with lateral margins markedly crenulate _ _ Amphizoa insolens LeConte Amphizoa davidi Lucas (Figures 1-3,7, 11, 13, 17) Amphizoa davidis Lucas, 1882:157 [incorrect spelling]. Lec- totype (here designated), a male, in MNHP, labelled: "Mu- seum Paris, Mou-pin, A. David 1870'V "398"/ "774 70" [yellow-backed disk]/ "Amphizoa davidis, Lucas" [label double-pierced by pin, hence vertical on pin]; "Type" [red label]/"Mus6um Paris"/ "Lectotype Amphizoa davidi Lu- cas designated by D. H. Kavanaugh 1983" [red label]. Type- Locality.— Pao-hsing, Szechwan Province, People's Repub- lic of China. Edwards 1951:322. Kavanaugh 1980:289. Amphizoa davidi Lucas [justified emendation]. Edwards 1951: 322. Kavanaugh 1980:289. Kavanaugh and Roughley 1981: 269. Leech and Chandler 1956:301. NOTES ON NOMENCLATURE AND TYPE- SPECIMEN.— Mou-pin, Tibet, the area originally cited as type-locality, is now called "Pao-hsing" (30°22'N, 1 02°50'E). This region is no longer part of Tibet, but rather the western part of Szechwan Province, People's Republic of China. DISTINGUISHING CHARACTERISTICS. — Size small, SBL male = 1 1 .4 mm; body form (Fig. 1 , 3a) narrow; body color piceous, with antennae, maxillary and labial palpi, and tarsi rufopiceous; head (Fig. 2) finely and densely punctate; prono- tum (Fig. 2) coarsely and densely punctate, with KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 71 areas between punctures convex, granulate in ap- pearance; elytra finely and densely punctate, slightly rugose at base and in lateral one-fourth; pronotum (Fig. 7) broadest at base, with lateral margins arcuate at middle, markedly sinuate an- terior to basal angles, not crenulate, median lon- gitudinal impression present but faintly im- pressed; prosternal intercoxal process (Fig. 1 1) short, round; posterolateral angle of proepister- num and posteromedial angle of proepipleuron abut evenly to form smooth prothoracic margin (see Edwards 1951:321, "Plate 4"); elytral sil- houette (Fig. 3a) moderate in width basally and distinctly narrowed subapically, elytra (Fig. 3b) evenly convex, without carinae; elytral striae complete but faintly impressed and finely punc- tate; front tibia with posterodorsal groove pres- ent on apical three-fourths, with fringe setae in groove very short and restricted to apical one- half; male median lobe (Fig. 1 3) with shaft slen- der at middle, evenly arcuate ventrally, apex slightly deflected ventrally, left paramere narrow basally, with vestiture restricted to apical one- fourth; female unknown; specimen from south- western China (Fig. 1 7). Edwards's description of A. davidi (1951:322) was an English translation of the original de- scription in French (Lucas 1882). Based on my examination of the type-specimen, additional comments and certain amendments to the orig- inal description seem appropriate. Lucas de- scribed the type of A. davidi as "noir mat . . . avec les palpes . . . d'un brun teinte de ferrugi- neux. Les antennes . . . d'un brun ferrugineux brilliant" (i.e., dull black, with reddish-brown antennae and palpi). In my view, the specimen is as dull as adults of A. insolens and A. striata but less dull than adults of A. lecontei. Its body color is piceous, not black as in A. insolens adults; and its antennae and palpi are rufopiceous, not reddish-brown. The median longitudinal impression (median furrow), which was de- scribed as "ne presente pas" (i.e., absent), is pres- ent and as deeply impressed as in A. striata adults, less so than in A. lecontei and A. insolens mem- bers. Lucas described the scutellum as "tres fine- ment chagrine" (very finely granulate); but be- cause this character state is shared with adults of the other Amphizoa species, it is of no taxo- nomic use. According to the original description, the elytral striae are "les parcourent obsolete- ment accusees et non ponctuees" (obsolete and FIGURE 2. Amphizoa davidi Lucas: scanning electron mi- crograph of head and pronotum, dorsal aspect, magnification = 25 x (specimen uncoated). impunctate); but they appear to be complete and clearly (although very shallowly) impressed. Due to dense punctation of the entire elytral surface, it is difficult to distinguish the fine punctures which are found in the striae. Close examination of the elytra of the lectotype of A. davidi has revealed a previously unrecorded feature. Due to the relatively faint development of macrosculpture on the elytral surface of this specimen, I found small but distinctly foveate punctures on the third, fifth, seventh, and ninth intervals. No setae appear to be associated with these punctures. Identical punctures were sub- sequently found in adults of all three Nearctic Amphizoa species, although they are much less obvious in Nearctic specimens, least so in A. striata adults. Similar, but seta-bearing, punc- tures are found among adults of a broad spectrum of tribes and genera of Carabidae. The presence of such setiferous punctures on odd-numbered elytral intervals (except the first) in amphizoids suggests that this may be an ancestral (plesiotyp- ic) adephagan trait. Absence of setae from the punctures may be an apotypic trait associated with development of an aquatic lifestyle. Ab- sence of the punctures themselves (such as is seen in adults of the other hydradephagan groups) may represent a more highly evolved trait. However, a majority of carabid groups also lack some or all of these punctures; so the evolution of this character has been complex and homoplastic, no matter how the polarity of its states is interpret- ed. HABITAT DISTRIBUTION.— Unknown. 72 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 GEOGRAPHICAL DISTRIBUTION.— This species is known only from the type-locality in south- western China (Fig. 1 7), in the Min River drain- age, an upper tributary of the Yangtze River. This watershed flows first south, then east more than 2,000 km to the Pacific Ocean, at 30°N latitude, and has no Himalayan drainage com- ponent. Past mislocation of the type-locality (i.e., Tibet rather than Szechwan, China) has apparently led collectors astray. Several workers, all with knowledge of the habits and habitat preferences of Nearctic amphizoids, have collected in var- ious parts of the Himalayas (e.g., India, Nepal, Tibet, Sikkim) in recent years without finding representatives of this species. This suggests that the range of Amphizoa in the Palaearctic Region may not extend west to include the main Hi- malayan ranges. Furthermore, the People's Re- public of China has been closed to most western collectors for decades (and until very recently); and this may account for the lack of additional specimens in European or North American col- lections during this century. GEOGRAPHICAL RELATIONSHIPS WITH OTHER SPECIES.— The known range of this species is al- lopatric with respect to ranges of all other known species ofAmphizoa. Amphizoa insolens LeConte (Figures 4, 8, 14, 18,28) Amphizoa insolens LeConte, 1853:227. Lectotype (designated by Kavanaugh 1980) male in MCZ. Type-Locality.— Sac- ramento, California. Edwards 1951:323, 1954:19. Hatch 1953:194. Kavanaugh 1980:290. Leech and Chandler 1956: 301. Dysmathes sahlbergii Mannerheim, 1 853:265. Location of type- specimen unknown. Type-Locality.— Si tka, Alaska. Ed- wards 1951 :323. Kavanaugh 1 980:29 1 . Synonymized by Salle 1874:222. Amphizoa josephi Matthews, 1872:1 19. Lectotype (designated by Kavanaugh 1 980) male in BMNH. Type-Locality.— Van- couver Island, British Columbia. Edwards 1951:323. Hatch 1953: 194. Kavanaugh 1980:290. Synonymized by Horn 1873: 717. NOTES ON NOMENCLATURE AND TYPE- SPECIMENS.— The problem with location of the holotype of Dysmathes sahlbergii Mannerheim was discussed by Kavanaugh (1980). DISTINGUISHING CHARACTERISTICS.— Size var- ied (small, medium, or large), SBL male = 10.9- 13.6 mm, female 11.1-15.0 mm; body form moderately broad (Fig. 4a); body black, with an- tennae, maxillary and labial palpi, and tarsi black or rufopiceous; head finely and densely punctate; pronotum medially with coarse, sparse punc- tures with areas between punctures flat, laterally with punctures coarser, denser, confluent, sur- face markedly gnarled; elytra finely and densely punctate, markedly rugose at base and in lateral one-half; pronotum (Fig. 8) as broad (or broader) at middle as (than) at base, with lateral margins arcuate at middle, moderately or markedly sin- uate anterior to basal angles, markedly crenulate, median longitudinal impression deeply im- pressed; prosternal intercoxal process (Fig. 12) slightly elongate, spatulate; posterolateral angle of proepisternum and posteromedial angle of proepipleuron abut evenly to form smooth pro- thoracic margin (see Edwards 1951:321, "Plate 4"); elytral silhouette (Fig. 4a) subovoid, slightly narrowed basally, less narrowed subapically, ely- tra (Fig. 4b) evenly convex, without carinae; ely- tral striae complete but faintly impressed (diffi- cult to define laterally because of macrosculpture) and finely punctate; front tibia with posterodor- sal groove restricted to apical one-half or two- thirds, with fringe setae in groove restricted to apical one-third or one-half; male median lobe (Fig. 14) with shaft slightly thickened at middle, evenly arcuate ventrally, apex slightly deflected ventrally and extended apicodorsally, left par- amere narrow basally, with vestiture restricted to apical one-third; female coxostylus ("coxite" of Edwards 1951:321, see his "Plate 4") with stylar region short, with vestiture of only a few scattered, minute setae; specimen from western North America (Fig. 1 8). Among specimens studied, I observed greater variation in pronotal shape than that reported by Edwards (1951). Although many adults of A. insolens have pronota broadest at the middle, in most they are equally broad at middle and base. There is also notable variation among individ- uals with respect to tibial grooves and associated fringe setae ("hairs" of Edwards 1951). These structures are discussed more fully below in my treatment of A. lecontei. In A. insolens adults, the posterodorsal groove on the front tibiae is varied in length, either restricted to the apical one-half of the tibia or extended basally to oc- cupy the apical two-thirds of the tibia. Fringe setae in this groove are restricted to the apical one-third of the tibia in most individuals, but several adults were seen with fringe setae also at the middle of the tibia or even on the apical part of the proximal one-half. As noted by Edwards KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 73 (1951:324), A. insolens adults have the least completely developed complement of fringe se- tae among extant amphizoids (see further dis- cussion in section on Zoogeography and Evolu- tion). HABITAT DISTRIBUTION. — Members of this species are most often found at the edges of cold, swift-flowing streams, under rocks or in coarse gravel at shoreline, clinging to exposed roots be- neath undercut banks, or in floating debris that has collected in backwater eddies. Adults are often found in greatest numbers at the bases of water- falls, which represent a first stretch of quiet water after a steep drop downstream. The occasional occurrence of these beetles in ponds or lakes, where they are almost always found near the in- lets of torrential streams, probably results from their being washed downstream and does not represent permanent residence in such standing bodies of water. GEOGRAPHICAL DISTRIBUTION.— The known range of this species (Fig. 1 8) extends from south- ern Yukon Territory and southeastern Alaska south to the San Bernardino, San Gabriel, and San Jacinto mountains of southern California, and from the Pacific Coast, including the Queen Charlotte Islands and Vancouver Island, east across the Columbia Plateau and Great Basin to western Alberta, central Montana, western Wy- oming, central Idaho, and eastern Nevada. I have examined 398 males and 360 females from the following localities: CANADA Alberta: Banff National Park, Banff [May] (1; CAS). British Columbia: Yoho National Park, Kicking Horse River (20.9 km W of Field) [June] (2; USNM); other localities, Ainsworth Hot Springs [July] ( 1 ; USNM), Fernie [Aug.] ( 1 ; CAS), Haines High- way (km 92.1) [July] (1; UASM), Inverness [July-Aug.] (2; USNM), Kaslo [June] (1; USNM), Kay Falls [July] (1; CAS), Kuskanook (Kootenay Lake [530 m]) [Oct.] (1; RERo), Kwi- nitza (Telegraph Point) [June] (2; UASM), Nicomen Ridge [July] (1; CAS), North Bend [June] (1; USNM), Prince Rupert (north slope of Mount Hayes near base [ 1 20 m]) [July] ( 1 ; CAS), Revelstoke (25.7 km W) [July] (1; CAS), Seltat Creek (Haines Highway km 78.5) [June] (1 ; UASM), Skagit (40 km E of Hope) [July] (1; RERo), Stanley [June] (1; CAS), Stawamus River (2 km S of Squamish on Highway 99) [July] (1; CAS), Wynndel [Aug.] (10; CAS, OSUO); Queen Charlotte Islands, Graham Island (Ghost Creek drainage 7.3 km NW of Rennell Sound Road [240 m], Juskatla area, Nebria Peak at Lower Nebria Lake [620 m]) [July] (11; CAS), Moresby Island (3 km NE of Jedway [6-50 m], Mount Moresby at High Goose Lake [640 m]) [July-Aug.] (2; CAS); Vancouver Island, Tyee (4.9 km NW) [June] (2; UASM). Yukon Territory: Upper Frances River (at Route 10) [June] (3; DMad). UNITED STATES OF AMERICA Alaska: Juneau [June] (1; CAS), Lituya Bay (9.7 km N [240- 590 m]) [Aug.] (21; CAS). California: El Dorado County, Pino Grande ([1,370 m]) [July] (3; UCD), Pollock Pines [July] (1; UCD), Riverton [July] (2; CAS, UCD), Whitehall area [June] (1; CAS); Fresno County, Barton Flat ([1,580 m]) [May] (1; UCB), Huckleberry Meadow [May] (1; CAS); Inyo County, Lone Pine ( 1 2.9 km N) [June] ( 1 ; CAS); Kings Canyon National Park, Bubbs Creek Canyon ([3,200 m]) [July] (1; CAS); Lake County, Bartlett Springs [June] ( 1 ; CAS); Lassen County, Susan River (12.9 km N of Susanville on Highway 36) [Aug.] (1; CAS); Los Angeles County, Coldwater Canyon [Aug.] (2; CAS), Little Jimmy Creek [June] (1; GCha), Los Angeles area (1; USNM), San Antonio Creek [June] (6; GLPe); Madera County, Boggy Meadows ([1,830 m]) [July] (10; CAS, NSDA, SJSU); Mariposa County, Sweetwater Creek ([ 1 ,220 m]) [July] (2; CAS); Mono County, Twin Lakes [Aug.] (4; USNM); Nevada County [Aug.] (1; CAS), Sagehen Creek (near Hobart Mills) [Aug.] (9; UCD), Truckee [Aug.] ( 1 ; CAS); Placer County, Emigrant Gap ([1,620 m]) [June] (1; UCD), Shirttail Creek (below Yellow Pine Reservoir) [May] (1; BYUM); Plumas County, North Fork Feather River ([910 m]) [Apr.] ( 1 ; CAS); Riverside Coun- ty, San Jacinto Mountains (Idlewild) [July] (6; CAS); San Ber- nardino County, Camp Baldy [July, Sep.] (9; CAS, UCD), Cie- nega Seco (6.4 km E of Barton Flats on Highway 38) [Aug.] (1; GCha), Mill Creek (0.16 km E of Forest Falls [1,650 m]) [May] (4; CAS), San Gorgonio Mountain ([2,130 m]) [Sep.] (22; CAS); San Mateo County, Tunitas Creek [Aug.] (1; UASM); Santa Clara County, Corte Madera Creek [Apr.] (1; CAS), Los Gatos [June] (1; CAS), San Francisquito Creek (Stanford Uni- versity Campus) [July] (1; USNM), San Jose [Sep.] (1; CAS), Uvas Creek (Sveadal, Uvas County Park, Uvas Meadows) [Mar.-May, July-Aug.] (1 1; LGBe, SJSU); Santa Cruz County, Boulder Creek [Apr.] (1; SJSU), Castle Rock State Park [May] (1; LGBe); Sequoia National Park ([610-910 m]) [May-June] (7; CAS, UCD), Ash Mountain (Kaweah Powerhouse #3) [June- July, Sep.] (22; UCB, UCD), Cahoon Meadow ([2,290 m]) [Aug.] (1; CAS), Giant Forest [Aug.-Sep.] (1; CAS), Paradise Valley ([910-2,130 m]) [May, July] (2; CAS), Potwisha ([610- 1,520 m]) [May, July] (5; CAS, UCD, USNM), Wolverton ([2,130-2,740 m]) [June] (1; CAS); Shasta County [July] (1; USNM), Castle Crags [July] (4; CAS), Lost Creek (at Twin Bridges Road [1,450 m]) [Aug.] (3; CAS), McArthur-Burney Falls State Park ([910m]) [June-Sep.] (36; CAS, OSUO, SJSU), Old Station ([1,220 m]) [Sep.] (2; CAS), Viola ([1,370 m] and 6.4 km W) [June] (3; CAS, NSDA); Sierra County, Sierraville (8 km S [1,830 m]) [Aug.] (1; CAS); Siskiyou County [July] (7; CAS), Big Flat Campground [Aug.] (6; CAS), Cement Creek (S of Callahan [1,220 m]) [Aug.] (1; CAS), East Fork of South Fork Salmon River (headwaters at Cecilville/Callahan Road [1,830 m]) [July] (1; CAS), McCloud [June] (4; CAS, USNM), Mount Shasta (Panther Creek [2,440 m]) [July] ( 1 ; CAS), Shasta Springs (Shasta Retreat [730 m]) [July] (14; CAS), Taylor Lake Road ([1,750 m]) [Aug.] (1; CAS), Yreka area (1; USNM); Tehama County, Soap Creek ([6 1 0 m]) [July] ( 1 ; CAS); Trinity County, Boulder Creek (at Goldfield Campground [1,070 m]) [July] (5; CAS), Doe Gulch (1.6 km W of Altoona Mine on Ramshorn/Castella Road [1,230 m]) [Aug.] (1; CAS), Emerald Lake ([1,680 m]) [Aug.] (1; CAS), Rarick Gulch Creek (8 km S of Dedrick [640 m]) [Aug.] (1; CAS), Swift Creek ([1,520 m]) [May] (12; PUCA); Tulare County, California Hot Springs [July] (7; LGBe), Franklin Creek ([2,500-2,990 m]) [July] (1; CAS), Kaweah [June-Aug.] (10; CAS), Mineral King [Aug.] (2; CAS), South Fork Kaweah River [July] (2; USNM); Tuolumne 74 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 County, Herring Creek ([1,980 m]) [Aug.] (1; CAS); Yosemite National Park, Yosemite Valley (Lower Merced River) [June] (3; CAS); county unknown, Alpine Lake ( 1 ; CAS). Idaho: Blaine County, Petit Lake Creek (4.8 km WSW of Highway 93 on Twin Lakes Trail [2,130-2,440 m]) [Aug.] (1; CAS); Boise County, Rocky Bar ([1,830 m]) [June] (1; CAS); Elmore Coun- ty, South Fork Boise River (4.8 km N of Pine at Dog Creek [1,460 m]) [Aug.] (1; CAS); Nez Perce County, Waha [June] (1; CAS); Shoshone County, Wardner [July] (6; CAS, OSUO); county unknown, Twin Creek Forest Camp ([1,520 m]) [July] (2; OSUO). Montana: Cascade County, Belt Creek (27.4 km S of Monarch on Highway 89 [2,100 m]) [July] (4; CAS); Glacier National Park [July, Sep.] (5; CAS, SJSU, UCD), Howe Creek [July] (4; SJSU), St. Mary Lake [July] (3; CAS), Swiftcurrent Creek (at Many Glacier Ranger Station) [Aug.] (3; SJSU), Two Medicine Lake [July] (7; CAS); Sweetgrass County, Big Timber Creek (at Half Moon Campground [2,230-2,290 m]) [July] (4; CAS). Nevada: Elko County, Lamoille Creek (near headwaters) [June] (1; BYUM), Thomas Creek (12.9 km SE of Lamoille at Thomas Creek Campground [2,320-2,380 m]) [Aug.] (1 ; CAS); Lander County, Hilltop [Aug.] (1; NSDA), Skull Creek [Sep.] (3; NSDA); Washoe County, Galena Creek (17.7 km W of Highway 395 on Highway 27 [2,290 m]) [July] (3; CAS), Third Creek (at Highway 28 [2,210 m]) [July] (22; CAS), Whites Creek (near Reno) [Oct.] (1; NSDA); White Pine County, Taft Creek ([2, 1 30-2,440 m]) [July] (3; CAS). Oregon: Baker Coun- ty, Pine Creek (16.1 km W of Baker [1,220 m]) [June-July, Sep.] (17; CAS, OSUO, USNM); Benton County, Marys Peak (Parker Creek at Road 1245 and Road 1296) [June] (9; GLPe, OSUO, SJSU), Yew Creek (14.5 km E of Alsea) [May] (1; OSUO); Clackamas County, Brightwood [May] (1; OSUO); Deschutes County, Indian Ford Creek (8 and 9.7 km NW of Sisters) [May-July, Sep.-Oct.] (81; GLPe, OSUO, SJSU), Squaw Creek (Highway 20 at Sisters [980 m]) [Aug.] (1; CAS); Hood River County, Mount Hood (Sand Creek) [July] (4; CAS); Jef- ferson County, Camp Sherman [Aug.] (2; UCD), Metolius (9; OSUO), Metolius River [June] (2; OSUO); Klamath County, Deming Creek (17.7 km NE of Ely) [June] (4; GLPa); Lane County, McKenzie River (8.4 km W of McKenzie Bridge [350 m]) [May] (2; CAS), South Fork McKenzie River [Sep.] (1; OSUO); Linn County, H. J. Andrews Forest (Mack Creek at Road 1553 [810 m]) [May] (1; CAS), North Santiam River (near Idanha) [May] (1; OSUO); Multnomah County, Bonne- ville [July] (1; BYUM), Horsetail Falls ([120 m]) [May, July] (8; CAS, GLPe, OSUO), Multnomah Falls [July] (5; CAS, OSUO); Wallowa County, Lostine River ([1,310m]) [Aug.] ( 1 ; OSUO), Wallowa Lake ([1,830 m]) [June-July] (9; CAS, OSUO, USNM), West Fork Wallowa River (at Sixmile Meadow [1,830 m]) [July] (1; CAS); Wasco County, Bear Springs (40 km W of Maupin [980 m]) [May] (1; OSUO). Washington (1; OSUO): Chelan County, Buck Creek [Aug.] (1; SJSU); Clallam County, Soleduck River [Sep.] (3; CAS, SJSU); King County, Fall City [July] (1; OSUO), Greenwater River (1; OSUO), North Bend [July] ( 1 ; CAS), Seattle [July] ( 1 ; OSUO), Snoqualmie ( 1 ; OSUO), Snoqualmie Pass [Sep.] (2; OSUO), Tokul Creek (at Tokul) [July] (4; CAS, GLPe, OSUO, UCD), Wellington [July] (4; CAS, USNM), White River (8 km W of Greenwater on High- way 410 [490 m]) [Aug.] (35; CAS); Kitsap County, Seabeck [Aug.] (1; OSUO); Kittitas County, Iron Creek Pass ([1,520 m]) [Aug.] (2; OSUO); Lewis County, Horse Creek (near Long- mire) [July] (1; CAS); Mason County [June] (2; OSUO), Pebble Ford Creek [June] (1; OSUO), Skokomish River [May] (1; OSUO); Mount Rainier National Park, Longmire [July] (2; CAS), Narada Falls ([1,370 m]) [July] (1; CAS), Paradise River ([1,490 m]) [June] (1; USNM); Olympic National Park, De- ception Creek (at Dosewallips Trail [960 m]) [July] (1; CAS), Olympic Hot Springs ([760 m]) [June-July] (9; CAS, OSUO), Pass Creek (at Dosewallips Trail [560 m]) [July] (1; CAS), Sol Due Hot Springs [Aug.] (2; USNM), Upper Twin Creek (at Dosewallips Trail [670 m]) [July] (2; CAS); Pierce County, Goat Creek (6.4 km E of Ashford on Highway 706 at Nisqually River [900-910 m]) [July] (4; CAS), Poch Creek (Carbon River Canyon) [Aug.] (1; OSUO); Whatcom County, Mount Baker (3.3 km E of Picture Lake on Highway 542 at Bagley Creek [670 m]) [Aug.] (1; CAS); Yakima County, Glenwood (2.7 km N [700 m]) [May] (2; CAS), Mount Adams (Bird Creek [1 ,370- 2, 1 30 m]) [July] (44; CAS, OSUO, USNM), Naches River (6.9 km SE of Cliffdell on Highway 410 [740 m]) [July] (1; CAS), White Pass [June] (2; SJSU); county unknown, Mount Adams ([1,830-2,440 m]) [July] (3; CAS). Wyoming: Yellowstone Na- tional Park, Gardiner River (at Mammoth Hot Springs) [Aug.] (1; OSUO). Locality unknown: (2; CAS, USNM). GEOGRAPHICAL VARIATION.— Although con- siderable intrapopulational variation is evident for several characters, the only character in which I observed variation associated with distribution is body size. Adults from southern California, at the southern range limit for A. insolens, are the largest specimens I have seen. The smallest adults are from coastal Alaska, at the extreme northern range limit of the species. Adults from inter- mediate areas are intermediate in size, but the pattern is not strictly clinal. For example, adults from the area around Portland, Oregon, are larg- er than those from the Mount Rainier, Wash- ington, area; and adults from interior localities (e.g., Alberta, Montana, and Wyoming) are al- most as small as Alaskan specimens and clearly smaller than specimens from west coast localities at equivalent latitudes. Hence, the pattern is one of decreasing size from south to north and from west to east, with minor exceptions to the pattern in a few areas (such as Portland). GEOGRAPHICAL RELATIONSHIPS WITH OTHER SPECIES.— The known range of A. striata (Fig. 19) is completely within the range of A. insolens (Fig. 1 8). Nevertheless, the two taxa may not be mi- crosympatric. The only record of their co- occurrence (perhaps in different streams) is at North Bend, King County, Washington. The geographical ranges of A. insolens and A. lecontei (Fig. 20) overlap broadly: from south- central and southeastern British Columbia, east to southwestern Alberta (Banff National Park) and northwestern Montana (Glacier National Park), and south to northeastern Oregon (Wal- lowa Mountains) and central Idaho (Sawtooth Mountain system). Ranges of these species ap- pear to overlap also in northwestern British Co- KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 75 lumbia/southern Yukon Territory. Adults of both species have been found together in several lo- calities (see respective locality lists). Amphizoa striata Van Dyke (Figures 5, 9, 15, 19,28) Amphizoa striata Van Dyke, 19276:197. Holotype male in CAS. Type-Locality.— North Bend, King County, Washing- ton. Edwards 1951:324. Hatch 1953:194. Kavanaugh 1980: 291. Leech and Chandler 1956:301. DISTINGUISHING CHARACTERISTICS.— Size large, SBL male = 13.1-14.2 mm, female 13.2-14.9 mm; body form very broad (Fig. 5a); body dark brown or piceous, with antennae, maxillary and labial palpi, and tarsi piceous or rufopiceous; head very finely and densely punctate; pronotum coarsely, moderately densely punctate over en- tire surface, with areas between punctures flat; elytra finely and densely punctate, slightly rugose in lateral one-third; pronotum (Fig. 9) broadest at base in most individuals (as broad at middle as at base in a few individuals), with lateral mar- gins slightly or moderately arcuate at middle, not sinuate or slightly sinuate anterior to basal an- gles, slightly or moderately crenulate, median longitudinal impression present but faintly im- pressed; prosternal intercoxal process moderate- ly elongate, spatulate; posterolateral angle of proepisternum and posteromedial angle of pro- epipleuron abut evenly to form smooth protho- racic margin (see Edwards 1951:321, "Plate 4"); elytral silhouette (Fig. 5a) very broad basally, markedly narrowed subapically, elytra (Fig. 5b) convex, except slightly concave in lateral one- half posterior to humeral area, without carinae; elytral striae complete but faintly impressed, coarsely punctate; front tibia with posterodorsal groove extended along entire length, with fringe setae in groove restricted to apical two-thirds or four-fifths of tibia; male median lobe (Fig. 1 5) with shaft distinctly thickened at middle, slightly bulged ventrally, apex slightly deflected ventral- ly, not extended apicodorsally, left paramere broad basally, with vestiture restricted to apical one-fourth; female coxostylus ("coxite" of Ed- wards 1951:321, see his "Plate 4") with stylar region medium in length, with dense vestiture of minute setae; specimen from western North America (Fig. 1 9). HABITAT DISTRIBUTION. — Members of this species have been found in cool (but not cold), slow-flowing streams (Edwards, pers. comm.) and in roadside ditches. Their distribution in such streams is similar to that of members of A. in- solens. GEOGRAPHICAL DISTRIBUTION.— The known range of this species (Fig. 1 9) extends from south- ern Vancouver Island and the Olympic Penin- sula and Cascade Range of northern Washington, south to southwestern Oregon, and east to Yak- ima County, Washington, and Wasco County, Oregon (both east of the Cascade Range). I have examined 73 males and 63 females from the following localities: CANADA British Columbia: Vancouver Island, Duncan (Koksilah Creek) [Aug.] (5; MCZ, OSUO, USNM), Little Qualicum Falls Provincial Park (Little Qualicum River) [Aug.] (4; CAS, OSUO). UNITED STATES OF AMERICA Oregon: Benton County, Sulphur Springs (9.7 km NW of Corvallis) [May] (2; GLPe); Clackamas County, Colton [Aug.] (1; CAS); Jackson County, Little Applegate River (7.2 km S of Ruch [520 m]) [May] (1 ; CAS); Lincoln County, Deer Creek (12.9 km S of Toledo) [June] (1; OSUO); Wasco County, Tygh Valley [June] (1; OSUO). Washington: Clallam County, La Push [July] (1; OSUO); King County, Bothell (North Creek, Swamp Creek) [May-July] (12; CAS, GLPe, OSUO, SJSU, UCD), North Bend [July] (3; CAS), Seattle (Swamp Creek) [July, Sep.] (10; BYUM, CAS, OSUO, SJSU, UCD), Swamp Creek [May-Aug.] (71; GLPe, OSUO); Kitsap County, Brem- erton [Apr.] (1; NDSA); Kittitas County, Parke Creek (near Kittitas) [Aug.] (1; LGBe); Mason County, South Fork Sno- homish River [July] (2; OSUO); Snohomish County, Hazel (Stillaguamish Club) [May] (1; OSUO), Swamp Creek [Sep.] (5; SJSU); Yakima County, Satus Creek (near Toppenish [610 m]) [Aug.] (7; CAS, UCD, USNM); county unknown, Pack Forest [Aug.] (1; OSUO), "Pisht R." [July] (1; CAS). GEOGRAPHICAL VARIATION.— Although intra- populational variation is evident in body size, pronotal shape, and several other characters, I found no characters in which variation is asso- ciated with distribution. GEOGRAPHICAL RELATIONSHIPS WITH OTHER SPECIES. — Refer to discussions under this head- ing for A. insolens and A. lecontei. Amphizoa lecontei Matthews (Figures 6, 10, 12, 16,20,28) Amphizoa lecontei Matthews, 1872:121. Lectotype (designated by Kavanaugh 1980) male in BMNH. Type-Locality.- Van- couver Island, British Columbia [doubtful record, see com- ments below]. Ed wards 1951:327, 1954:19.Hatch 1953:195. Kavanaugh 1980:290. Leech and Chandler 1956:301. Amphizoa planata Van Dyke, 1927a:98. Holotype female in CAS. Type-Locality.- Beaver Creek, Alberta. Edwards 1951: 327. Hatch 1953:195. Kavanaugh 1980:291. Synonymized by Van Dyke 19276:197. Amphizoa carinata Edwards, 195 1 :326. Holotype male in CAS. 76 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 Type-Locality.— Conejos River near Menkhaven, Conejos County, Colorado. Kavanaugh 1980:289. Leech and Chan- dler 1956:301. NEW SYNONYMY. NOTES ON NOMENCLATURE AND TYPES.— The lectotype of A. lecontei is supposed to have been collected on Vancouver Island, British Colum- bia, as noted both in Matthews's original de- scription and on labels affixed to the specimen. However, I have not seen any other specimens from the island nor from the adjacent coastal mainland. Because the present type-locality ap- pears to be well outside the known geographical range of A. lecontei (see Fig. 20 and text below), it is probable that the lectotype is mislabeled and that the type-locality should be emended. How- ever, I choose not to do so at this time, pending further field efforts on Vancouver Island. DISTINGUISHING CHARACTERISTICS.— Size me- dium, SBL male = 1 1.7-12.7 mm, female 12.2- 14.0 mm; body form moderately broad (Fig. 6a); body dark brown or piceous (specimens from Arizona almost black), with antennae, maxillary and labial palpi, and tarsi piceous or rufopiceous; head very finely and densely punctate; pronotum medially with coarse, sparse punctures, with areas between punctures flat, laterally with punctures coarser, denser, more or less confluent, surface unevenly rugose in appearance; elytra finely and densely punctate, with punctures confluent over large areas, moderately rugose in lateral one-half; pronotum (Fig. 10) broadest at base in most in- dividuals (as broad at middle as at base in a few individuals), with lateral margins slightly or moderately arcuate at middle, not or slightly sin- uate anterior to basal angles, slightly or moder- ately crenulate, median longitudinal impression faintly or deeply impressed; prosternal intercoxal process moderately elongate, spatulate or sub- lanceolate; posterolateral angle of proepisternum and posteromedial angle of proepipleuron either abut evenly to form continuous posterior pro- thoracic margin or proepipleuron is distinctly shorter than proepisternum and the two do not abut evenly, posterior prothoracic margin there- fore with distinct jog (see Edwards 1951:321, "Plate 4"); elytral silhouette (Fig. 6a) broad ba- sally, markedly narrowed subapically, elytron (Fig. 6b) with blunt but distinct carina on fifth interval, area medial to carina elevated, flat, area lateral to carina slightly concave; elytral striae complete but faintly impressed, coarsely punc- tate; front tibia with posterodorsal groove ex- tended along entire length or restricted to apical four-fifths, with fringe setae in groove restricted to apical one-half or two-thirds of tibia; male median lobe (Fig. 1 6) with shaft distinctly thick- ened at middle, slightly bulged ventrally, apex slightly deflected ventrally, not extended apico- dorsally, left paramere broad basally, with ves- titure restricted to apical one-fourth; female cox- ostylus ("coxite" of Edwards 1951:321, see his "Plate 4") with stylar region long and slender, with dense vestiture of minute setae; specimen from western North America (Fig. 20). There is considerable intrapopulational vari- ation in the development of tibial grooves and associated fringe setae among adults of all Am- phizoa species; for this reason, I have experi- enced considerable difficulty in trying to interpret tibial characters that Edwards used to distinguish A. lecontei and A. carinata adults. I have found no differences between specimens Edwards iden- tified as A. carinata and specimens of A. lecontei from various localities throughout its range in development of tibial grooves or in length or distribution of fringe setae, except such as can be attributed to intrapopulational variation. Edwards also described and illustrated differ- ences in shape of valvifers and paraprocts be- tween A. lecontei and A. carinata females. Among my own dissections of females from within the range of A. carinata and from other localities for A. lecontei, I found only the lecontei form illus- trated by Edwards (1951:321, "Plate 4"). I have also examined material dissected by Edwards, including the specimen that he illustrated for A. carinata. Although his drawing is a true repre- sentation of form of the valvifers and paraprocts of the latter specimen, other specimens from the same series differ from it in form and are, in fact, similar to other females of A. lecontei. It appears, therefore, that A lecontei and A. carinata females are similar in form of valvifers and paraprocts, and that the A. carinata specimen illustrated by Edwards is atypical in this regard. HABITAT DISTRIBUTION. — Members of this species are found in cool or cold, slow- or fast- flowing streams, in the same microhabitats as those described above for A. insolens members. However, they are more common in stretches of slow-moving water and streams that drop less steeply than are members of the latter species. GEOGRAPHICAL DISTRIBUTION.— The known range of this species (Fig. 20) extends from south- KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 77 ern Yukon Territory south along the Rocky Mountain system to the Chuska Mountains of northeastern Arizona and Sangre de Cristo Range of northern New Mexico, and from the Wallowa Mountains of northeastern Oregon and Indepen- dence Mountains of northeastern Nevada east to the Bighorn Mountains of northcentral Wyo- ming and Front Range of central Colorado. I have examined 1 90 males and 202 females from the following localities: CANADA Alberta: Banff National Park, Banff (and at Cascade River) [May, July-Aug.] (5; CAS, USNM); other localities, Beaver Creek [May] (4; CAS, USNM), Blairmore [Aug.] (1; CAS), Edmonton (1; OSUO), Happy Valley [Aug.] (2; CAS, USNM), Lundbreck [Aug.] (1; CAS), Mill Creek (72.4 km W of Fort Macleod) [Aug.] (2; USNM), Whitecourt (21 km SE on High- way 43) [May-June] (2; RERo). British Columbia: Creston (Goat River) [July-Aug.] (16; CAS, GLPe, OSUO), Fernie (Liz- ard Creek) [July] (1; CAS), Golden [Apr.] (1; CAS), Midday Valley (near Merritt) [July-Aug.] (4; CAS), Stanley [June] (1; CAS), Vernon (1; USNM). Yukon Territory: Haunka Creek (Highway 8 N of Atlin, British Columbia) [July] (1; UASM). UNITED STATES OF AMERICA Arizona: Apache County, Lukachukai Creek (8 km NE of Lukachukai at Wagon Wheel Campground [2,250-2,260 m]) [May, July-Aug.] (21; BMNH, CAS, MCZ, UASM, USNM). Colorado (2; MCZ): Archuleta County, Pagosa Springs area ([2,440-2,740 m]) [Aug.] (2; MCZ), Upper San Juan Valley ([2,130-3,200 m]) [Aug.] (6; MCZ, USNM); Boulder County, Coal Creek (3.2 km E of Wondervu) [May] (1; CAS), Lefthand Creek (9.7 km WSW of Highway 36 [2,010 m]) [Aug.] (12; CAS); Conejos County, Menkhaven (Conejos River) [June] (2; CAS); Jackson County, Cameron Pass ([2,740-2,930 m]) [Aug.] (5; CAS, SJSU), Gould (Michigan River near Cameron Pass) [Aug.] (2; BYUM); Larimer County, Virginia Dale [June] (1; USNM); Pueblo County, Beulah [Aug.] (4; MCZ); San Miguel County, South Fork San Miguel River ([2,590 m]) [July] (12; MCZ). Idaho: Adams County, New Meadows [June] (2; CAS, OSUO); Bear Lake County, Bloomington Creek (1 1.1 km SW of Bloomington [2,130 m]) [Aug.] (2; CAS); Camas County, Carrie Creek (57.9 km ESE of Ketchum [2,100 m]) [Aug.] (13; CAS), Little Snake Creek [Sep.] (1; GLPe), South Fork Boise River (22.5 km E of Featherville at Skeleton Creek [1,550 m]) [Aug.] (2; CAS); Cassia County, Goose Creek [July] (2; GLPe), Magic Mountain (Rock Creek at Ranger Station [1,890 m]) [July] (5; OSUO); Clark County, Birch Creek [July] (1; GLPe); Elmore County, South Fork Boise River (4.8 km N of Pine at Dog Creek [1,460 m]) [Aug.] (3; CAS), Wood Creek (1.6 km S of Pine [1,370 m]) [Aug.] (16; CAS); Valley County, Bear Valley [July] (1; GLPa). Montana: Cascade County, Dry Fork Belt Creek (at Henn Gulch [ 1 ,620 m]) [July] (9; CAS); Chouteau County [Aug.] (1; OSUO); Glacier National Park [July-Sep.] (10; CAS, SJSU), Kintla Lake [June] (1; CAS), Swiftcurrent Creek (at Many Glacier Ranger Station [ 1 ,460 m]) [June-Aug.] (32; CAS, SJSU). Nevada: Elko County, North Fork Humboldt River [Oct.] (1; BYUM). New Mexico: Taos County, Red Riv- er (6.6 km W of Red River [2,580 m]) [June] (1 ; CAS). Oregon: Baker County, Cornucopia (14.5 km NW of Halfway) [July] (1; GLPe), Richland area ([1,220 m]) [June] (1; CAS); Grant County, Clear Creek (3.2 km W of Granite) [Aug.] (1; GLPe); Wallowa County, Bear Creek (at Boundary Camp) [Sep.] (1; USNM), Lostine River (16.1 km S of Lostine [1,310 m]) [July- Aug.] (7; CAS, OSUO, UCD, USNM). Utah: Box Elder County Clear Creek (at Clear Creek Campground) [Mar.] (1; BYUM), George Creek Campground [Apr.] (1; BYUM); Emery County, Huntington Creek (at Stuart Ranger Station) [July] ( 1 ; BYUM); Garfield County, Steep Creek [Aug.] ( 1 ; BYUM); Kane County, East Fork Virgin River (7.9 km NE of Glendale [1,860 m]) [June] (2; CAS); Salt Lake County, City Creek [June-July] (14; USNM); Piute County, Beaver Creek (below national forest boundary) [May] (1; BYUM); Sevier County, Mount Marvine (0.2 km N of Johnson Valley Reservoir at Sevenmile Creek [2,590 m]) [Aug.] (15; CAS); Summit County, Tryol Lake (1; BYUM); Utah County, Hobble Creek ([1,830 m]) [July-Aug.] (29; BYUM, CAS, NSDA, SJSU), Prove ([1,490 m]) (1; CAS); Wasatch County, Little South Fork Provo River [July] (1; BYUM), Lost Lake Campground ([2,990 m]) [Aug.] (1; CAS), Upper Provo River (5.5 km E of Hailstone Junction on High- way 89A/150 [1,890 m]) [Aug.] (35; CAS), West Fork Du- schesne River [Aug.] (1; BYUM); Weber County, Ogden [July] (2; USNM), Weber River (Highway 30 at Mountain Green [1,510 m]) [Aug.] (3; CAS); county unknown, Uinta Moun- tains [June] (2; BYUM), "Wasatch" [June] (1; USNM). Wash- ington: Pend Oreille County, Sullivan Lake [Aug.] (2; CAS, OSUO); Stevens County, Crystal Falls [Aug.] (1; CAS). Wy- oming: Big Horn County, Granite Creek ( 1 2.9 km SW of Gran- ite Pass on Highway 14 [2,380 m]) [July] (2; CAS); Converse County, LaPrele Creek (61.2 km SW of Douglas on Highway 91 at Camel Creek Campground [2,530 m]) [July] (7; CAS); Grand Teton National Park, Colter Bay [Aug.] ( 1 ; SJSU), Delta Lake ([2,730 m]) [July] (1; SJSU); Johnson County, South Fork Clear Creek (25.7 km W of Buffalo on Highway 16 [2,350 m]) [July] (7; CAS), Tie Hack Camp [Aug.] (2; SJSU); Sheridan County, Little Tongue River (20.9 km WSW of Dayton on Highway 14 [2,380 m]) [July] (5; CAS); Sublette County, Ho- back River (3.2 km NW of Bondurant [2,100 m]) [Aug.] (28; CAS); Teton County, Jackson (1; USNM); Washakie County, Tensleep Creek (1 7.7 km NE of Tensleep on Highway 1 6 [ 1 ,890 m]) [July] ( 1 ; CAS); Yellowstone National Park, Grand Canyon of the Yellowstone (above Tower Falls) [Aug.] ( 1 ; MCZ), Indian River Campground [Aug.] (1; USNM), Spirea Creek [Aug.] (2; SJSU). GEOGRAPHICAL VARIATION.— In his original description of A. carinata, Edwards (1951:327) suggested that this form might represent "merely a geographical subspecies" of A. lecontei, but added that "it seems probable that no intergra- dation occurs between these populations." How- ever, subsequent collections from geographically intermediate areas demonstrate intergradation, and the incongruence found among geographical variation patterns of different characters has led me to treat A. lecontei and A. carinata as con- specific. Nonetheless, the pattern of variation in A. lecontei merits description. Mature (i.e., non-teneral) adult specimens from northeastern Arizona are black whereas mature specimens from other parts of the species range are piceous or dark brown. 78 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 Characters of pronotal shape, including shape of lateral margins, of apical and basal angles, and relative width at base versus at middle, are all highly varied among adults of A. lecontei. All character states cited by Edwards as unique for A. carinata adults fall within the range of vari- ation seen among A. lecontei adults from other geographical areas. Edwards described the me- dian longitudinal impression as deep in A. car- inata adults but shallow and indistinct in A. lecontei adults. Specimens with pronotal char- acteristics of the carinata form predominate in the region from southcentral Wyoming, south through Colorado and northern New Mexico, east through northeastern Arizona, and north through south and central Utah. Adults with the typical lecontei form predominate in all other areas. Specimens with prominent excavations of the prosternum anterior to the front coxal cavities, described by Edwards as a feature unique to A. carinata adults, are found in localities through- out the range of A. lecontei, although always in lower numbers than specimens from which these excavations are lacking. Furthermore, not all specimens exhibiting other features characteris- tic of A. carinata have these excavations (e.g., most specimens from Arizona). Similarly, the relationship between the proepisternum and proepipleuron described and illustrated by Ed- wards (1951:321, "Plate 4") does not hold up as a distinguishing feature of A. carinata adults. Samples from localities in southcentral Wyo- ming, northern Colorado, northern New Mexico, northeastern Arizona, and southern and east- central Utah include specimens exhibiting both states of this character, as well as intermediates between these extremes. In most adults from northeastern Arizona, northern New Mexico, and Colorado, the pro- sternal intercoxal process is slender, elongate, and sublanceolate, whereas it is slightly broader, shorter, and spatulate in adults from other areas. Several of the characters noted above are use- ful for describing the carinata form. Its geograph- ical range is centered at the southern extreme of the range of A. lecontei, in northeastern Arizona, and extends northwestward (through Utah) and northeastward (through New Mexico, Colorado, and southcentral Wyoming). In successively more northern populations within this range, the car- inata form is represented by a lower percentage of individuals. However, adults demonstrating one or more A. carinata traits are found in low numbers throughout the range of A. lecontei; adults that are intermediate between the A. car- inata and typical lecontei forms (for one or more characters) are abundant in northern parts of the range of the A. carinata form and present in low numbers throughout that range. Given this pat- tern, there appears to be insufficient reason for retaining the name A. carinata even at subspe- cific rank. GEOGRAPHICAL RELATIONSHIPS WITH OTHER SPECIES.— The geographical ranges of A. lecontei and A. insolens overlap extensively over a broad north/south area (see above under this heading for A. insolens); and adults have been collected together in several localities (e.g., at Swiftcurrent Creek, Glacier National Park, Montana; see also respective locality lists). Based on material I have examined, the ranges of A. lecontei and A. striata are allopatric. If, however, A. lecontei is represented on Vancou- ver Island, the original type locality for the species, then these two species are at least macrosym- patric in that area. Phylogeny Prerequisite to understanding the evolutionary and distributional histories of the species ofAm- phizoa is formulation of a hypothesis of phylo- genetic relationships among them. Cladistic analysis is the best available technique for elu- cidation of these relationships (Hennig 1 966; Ka- vanaugh 1972, 1978a). Briefly, the analytical procedure is as follows. (1) For each character, the direction of its evolution (i.e., the so-called "polarity" of the transformation of its different character states) is determined, from most prim- itive (plesiotypic) to most derived (apotypic) state or states. (2) Taxa are then grouped together, solely on the basis of shared derived (synapo- typic) character states, into successively more in- clusive groups. (3) Because synapotypy is ac- cepted as evidence for common ancestry, and because degree of phylogenetic (cladistic) rela- tionship is equivalent to relative recency of com- mon ancestry, the hypothetical branching pat- tern of phylogenetic relationships inferred is simply the grouping sequence read in reversed order (i.e., from most to least inclusive). The crucial step in cladistic analysis is deter- mination of the polarity of transformations of character states for each character. Several cri- teria have been proposed and/or used (Ball 1975; KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 79 FIGURES 3-6. Body form (a = dorsal aspect, right elytron omitted; b = cross-sectional dorsal silhouette of left elytron at point one-third of elytral length from base); scale line = 1.0 mm. Figure 3. Amphizoa davidi Lucas (Pao-hsing, China). Figure 4. Amphizoa insolens LeConte (Indian Ford Creek, Oregon). Figure 5. Amphizoa striata Van Dyke (Swamp Creek, Washington). Figure 6. Amphizoa lecontei Matthews (Upper Provo Canyon, Utah). Ross 1974; Ekis 1977; Kavanaugh 19786; Crisci and Stuessy 1980; Watrous and Wheeler 1981; and references therein) to determine which states are relatively plesiotypic and which are relatively apotypic. Of these, only two have intrinsic merit. First, and most important, is the so-called "out- group" criterion, which can be stated as follows: for a given character with two or more states within a group, the state occurring in related groups is assumed to be the plesiomorphic [=ple- siotypic] state (Watrous and Wheeler 1981). This criterion is relatively straightforward and easy to apply, except when an appropriate out-group is difficult to recognize or when more than one character state is represented in the out-group. Recently, Maddison, Donoghue, and Maddison (1 984) proposed a practical method for out-group analysis using parsimony criteria which should prove useful even when phylogenetic relation- ships among out-group components are inade- quately known. The second criterion, "character correlation" (Ekis 1977; Hennig 1966; Kavanaugh 19786), can be stated as follows: characters for which the polarities of transformation series have been de- termined with confidence can be used to infer polarities in transformations of other characters in which evolutionary sequence is less easily in- ferred. This is the criterion of choice only when the out-group criterion cannot be applied on its 80 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 FIGURES 7-10. Pronotum, dorsal aspect; scale line = 1.0 mm. Figure 7. Amphizoa davidi Lucas (Pao-hsing, China). Figure 8. Amphizoa insolens LeConte (Indian Ford Creek, Oregon). Figure 9. Amphizoa striata Van Dyke (Swamp Creek, Washington). Figure 10. Amphizoa lecontei Matthews (Upper Provo Canyon, Utah). own. Without formally recognizing it, Watrous and Wheeler (1981) invoked the character cor- relation criterion in order to recognize functional in-groups and functional out-groups where con- ventional (i.e., their so-called "taxonomic") in- groups and out-groups proved useless. In prac- tice, a tentative phylogenetic tree (cladogram) is constructed based on one or more characters for which character-state polarities are well estab- lished. Depending on the structure of the clado- gram derived, it may be possible to recognize a functional out-group (e.g., the most basal diver- gent lineage in the cladogram), which can be used in analysis of other characters. The distribution of states of another character, polarity of which cannot be determined by reference to the out- ^s Q£ *\J FIGURES 11, 12. Prosternal intercoxal process, ventral as- pect; scale line = 1.0 mm. Figure 11. Amphizoa davidi Lucas (Pao-hsing, China). Figure 12. Amphizoa lecontei Matthews (Lukachukai Creek, Arizona). FIGURES 13-16. Median lobe and left paramere of male genitalia, left lateral aspect; scale line = 1.0 mm. Figure 13. Amphizoa davidi Lucas (Pao-hsing, China). Figure 14. Am- phizoa insolens LeConte (Indian Ford Creek, Oregon). Figure 15. Amphizoa striata Van Dyke (Swamp Creek, Washington). Figure 16. Amphizoa lecontei Matthews (Swiftcurrent Creek, Montana). KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 81 FIGURE 17. Map of geographical distribution ofAmphizoa davidi Lucas. group criterion using the conventional (taxo- nomic) out-group, is then determined over the tentative cladogram. A derivative out-group cor- relation may then be possible, making use of the functional out-group recognized using other characters. The choice of a suitable out-group for Am- phizoa is not a simple one. Although amphizoids have long been considered to represent an evo- lutionary grade between the Geadephaga and the more specialized groups of Hydradephaga (LeConte 1853; Edwards 1951), their phyloge- netic relationships with other adephagan groups are not clearly understood. Among the Adephaga are both terrestrial and aquatic groups, each with generalists, specialists, and hyperspecialists (Er- win 1979) in their ranks. Structural, functional, and behavioral diversity within the suborder is great, and independent evolutionary trends, some in opposite directions, are numerous. Character state distributions of many important characters are so complex within the suborder, at least in our present understanding of them, that the log- ical out-group for amphizoids, the Adephaga- minus-Amphizoidae, is not a particularly useful group for cladistic analysis. I have, therefore, tried to limit the scope of the out-group to some subgroup of Adephaga to maximize the useful- ness of the out-group criterion as a tool in anal- ysis. Phylogenetic relationships of amphizoids To many workers (Edwards 1951, and refer- ences therein), amphizoids appear to represent a primitive grade of dytiscoid evolution. Amphi- zoids lack structural adaptations of highly spe- cialized swimmers (Hlavac 1975; Evans 1982), such as dytiscids. In fact, they are much more efficient as runners on land than as swimmers in water. Characteristics of thoracic and male and female genitalic structure and numerous other features approximate what could be expected in a suitable common dytiscoid ancestor. Among 82 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 20 FIGURES 18-20. Map of geographical distribution. Figure 18. Amphizoa insolens LeConte. Figure 19. Amphizoa striata Van Dyke. Figure 20. Amphizoa lecontei Matthews. extant forms, no other hydradephagan adults ap- pear to represent the early Mesozoic grade of dytiscoid evolution (Ponomarenko 1977) as well as amphizoids. But are amphizoids related only to the dytiscoids, and if so, to which most closely? Several more particular affinities have been proposed for amphizoids. Horn (1881) and a few later authors have suggested close relationship with hygrobiids. Might the latter alone serve as a suitable out-group? Probably not. Evidence linking hygrobiids with amphizoids is minimal and likely based on symplesiotypic traits (Ham- mond 1979). Bell (1966) suggested that amphizoids, living in habitats where swimming is hazardous, may have evolved from more advanced dytiscids, with their apparent plesiotypic characteristics repre- senting secondary loss or reduction of swimming adaptations. However, characteristics of pro- thoracic (Hlavac 1975) and pterothoracic struc- ture (Ponomarenko 1977; Evans 1982, also in press) support a phylogenetically more remote (basal) relationship between amphizoids and dy- tiscids. Although a suitable out-group for am- phizoids must include dytiscids, extant forms of the latter represent a more highly specialized grade of adephagan evolution and are probably not suf- ficient as an out-group. Several lines of evidence suggest that Trachy- pachidae form a monophyletic group with the dytiscoid families, including all the hydradepha- gan groups except, perhaps, haliplids and gy- rinids (Bell 1966, 1967, 1982; Crowson 1981; Evans 1977, 1982; Hammond 1979; Forsythe 1981; Roughley 1981). This relationship is es- tablished on the basis of numerous supposed syn- apotypic features (Hammond 1979; Roughley 1981; Evans 1982) involving characters of an- tennal pubescence, locomotory function and structure (of legs, wings, and associated struc- tures), male and female genitalic structure, and female reproductive system. If trachypachids are KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 83 closely related to amphizoids and other dytis- coids, then they should be included in any suit- able out-group for analysis. Determination of apotypic states of at least several characters linking trachypachids with dy- tiscoids is based on the assumption, either stated or implied, that the common adephagan ancestor was terrestrial rather than aquatic in habits (Crowson 1955, 1981; Bell 1966, 1967, 1982; Evans 1977, 1982, also in press; Hammond 1979; Forsythe 1981; Roughley 1981). However, this contention is not universally accepted. Erwin (1979) suggested that the Adephaga arose from an aquatic neuropteroid ancestor similar to ex- tant amphizoids, at least in habits. Based on re- view of both Palaeozoic and Mesozoic fossil bee- tles, Ponomarenko (1977) proposed an aquatic origin of Adephaga, probably in late Permian time, from aquatic schizophoroid Archostemata. In fact, his separation of fossil specimens of Adephaga from those of Schizophoridae of that age was admittedly somewhat arbitrary (Po- nomarenko 1977). Perhaps this distinction is one of grade rather than clade. Crowson (1981) and Ponomarenko (1977) agree on both the time (Late Permian) and source group (Archostemata: Schizophoridae) for the probable origin of Adephaga. They differ, how- ever, in their views on the ancestral adephagan habitat (whether terrestrial or aquatic), which by extension, could have been inherited from either terrestrial or aquatic schizophoroid ancestors, both of which are known from Permo-Triassic time. Amphizoids are only semiaquatic in habits. Adults are able to carry on most if not all life functions (e.g., feeding, locomotion, oviposition) at least as well out of water as in it and based on my fieldwork, do so routinely in nature. At least under laboratory conditions, eggs and larvae also thrive out of water, and pupation occurs on land. These habits are reflected by structure. Adults lack special adaptations for fast swimming and are barely able to move freely underwater except by clinging to substrata. Their most effective mode of locomotion in water is passively drifting with stream currents. Here I use the term "semi- aquatic" to refer to the combination of amphib- ious habits, structure that is relatively unspe- cialized for aquatic life, and ineffective swimming capability that is characteristic of extant amphi- zoids. If the ancestral adephagan was a terrestrial or- ganism, then amphizoids may represent, at least structurally, a first stage in adephagan adaptation to aquatic life. Primitive (plesiotypic) character states for Hydradephaga, including amphizoids, should be represented among their near terres- trial (i.e., geadephagan) relatives, including the Mesozoic Eodromeinae and Protorabinae (Po- nomarenko 1977), both living and extinct Trachypachinae (Bell 1966, 1982; Evans 1977, 1982; Ponomarenko 1977; Roughley 1981), and living basal-grade Carabinae, such as Notioka- siini, Nebriini, and Opisthiini. It might, there- fore, be a waste of effort to include other extant hydradephagan groups in the out-group for cla- distic analysis because their members may dem- onstrate only relatively apotypic character states associated with more advanced stages of spe- cialization to aquatic life. Alternatively, if the common adephagan an- cestor were aquatic, then plesiotypic character states should be associated with aquatic rather than terrestrial organisms. Any suitable out-group for cladistic analysis of amphizoids would have to include extinct aquatic groups, such as the Mesozoic Parahygrobiidae, Coptoclavidae, and Liadytidae (Ponomarenko 1 977), as well as other living dytiscoids (Hygrobiidae, Dytiscidae, and Noteridae). If, however, the common adephagan ancestor were only semiaquatic, similar in both habits (Erwin 1979) and structure to living am- phizoids, then extant dytiscoids might again be too specialized to be useful in out-group com- parisons. Composition of a suitable out-group for anal- ysis of extant Amphizoa species depends, at least in part, on the evolutionary hypothesis proposed to account for the origin and initial radiation of Adephaga— whether from a terrestrial, aquatic, or semiaquatic common ancestor. Faced with a choice from among five conflicting hypotheses (none of which he could reject with available evidence) to explain the relationships of trachy- pachids with other Adephaga, Bell (1982) called for additional efforts to discover new evidence bearing on the question. Perhaps a similar call for additional data is most appropriate here as well. However, even a preliminary cladistic anal- ysis of Amphizoa species at this time requires selection of an out-group for comparative pur- poses; such a selection requires a choice among alternative hypotheses for adephagan origin. In 84 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 my view, evidence favors the origin of Adephaga from a semiaquatic common ancestor for reasons outlined below. Evidence from the fossil record Thanks to Ponomarenko's (1977, and other papers cited therein) outstanding work on late Palaeozoic and Mesozoic beetle fossils, infor- mation about the early stages of adephagan evo- lution is now available. It is evident, for example, that a significant aquatic radiation of schizoph- orid Archostemata, presumptive ancestors of Adephaga, had occurred by Permo-Triassic time (Ponomarenko 1977). By early Mesozoic time, the adephagan radiation was already diverse. Forms that, structurally, could have given rise to all major extant adephagan groups— gy- rinoids, haliploids, dytiscoids, and caraboids— were represented in the Jurassic fauna of Asia. However, the aquatic adephagan component was clearly more diverse and more advanced (i.e., more similar to extant forms) than the terrestrial component of that time. The carabid fauna, for example, did not take on a modern aspect (i.e., one in which middle- and higher-grade carabids are evident) until mid- to late-Cretaceous time (Ponomarenko 1977). This suggests that the aquatic radiation of Adephaga preceded that of terrestrial groups. Much can also be learned about plesiotypic (primitive) versus apotypic (derived) character states for Adephaga from study of the diverse and beautifully preserved Mesozoic fossil ma- terial illustrated by Ponomarenko (1977). For example, it is clear, from review of these fossil specimens and out-group comparisons with schi- zophorid fossil material, that contribution to the lateral wall of the mesocoxal cavity by the met- episternum is plesiotypic in Adephaga. This trait was widespread among extinct (and extant) Ar- chostemata as well as the extinct eodromeine trachypachids, protorabine carabids, triaplids, and some (but not all) Mesozoic dytiscoid groups. Among extant forms it is restricted to Amphi- zoidae, some Dytiscidae, and members of genus Spanglerogyrus among Gyrinidae. Similarly, the form of hind coxae seen among extant trachy- pachids, dytiscids, amphizoids, hygrobiids, gy- rinids, and haliplids— in which the lateral coxal wing extends laterally to the elytral epipleuron, completely separating thoracic from abdominal sclerites (i.e., the "incomplete" form of Bell 1967, or "interrupted" form of Roughley 1981)— ap- pears to be plesiotypic, based on out-group com- parisons with schizophorids and Mesozoic fossil adephagans. In form and structure of hind coxae, relation- ships of mesepimera and metepisterna to me- socoxal cavities, and every- other structural detail that can be observed in the fossil material, am- phizoids appear to demonstrate the character state that can be interpreted as plesiotypic in relation to a semiaquatic ancestor and divergent lines of more specialized forms. Liadytids (Ponoma- renko 1977), which probably represent a basal grade of Mesozoic dytiscoids, have hind coxae more specialized (hence, apotypic) for rapid swimming than amphizoids, and coptoclavids (Ponomarenko 1977) have metepisterna exclud- ed from lateral walls of mesocoxal cavities by anterolateral extensions of the metasternum. Amphizoids are very similar in appearance and structure to Mesozoic eodromeine trachy- pachids, except that their metacoxae are slightly larger and more closely contiguous medially than the latter. Presumably, eodromeines were ter- restrial beetles, not aquatic or semiaquatic. Perhaps the only known form more similar to eodromeines than amphizoids is Necronectulus (Ponomarenko 1977), described from a single, legless specimen of Early Jurassic age from Asia. Its metacoxae were typical of those in eodro- meines, but nothing is known of its distal leg structure. Based on body structure and form of antennae, Ponomarenko suggested that it could have been either terrestrial or aquatic in habits, but he favored the latter view. Possibly, it rep- resents the first stage of adaptation to purely ter- restrial life among Adephaga, although the ear- liest known eodromeines predate the only known occurrence of Necronectulus in the fossil record. In summary, I suggest that a review of Me- sozoic fossil material provides two insights. First, character states demonstrated by extant amphi- zoid adults can, in almost every instance, be in- terpreted as plesiotypic in relation to respective character states in known Mesozoic and extant aquatic Adephaga, as well as extant trachy- pachids and carabids. Second, there is little with which to distinguish amphizoids and eodro- meine trachypachids, except their habitats. If this similarity is based on synapotypic features, then adephagan relationships could be as illustrated in Figure 2 la or 21b. If it is based on symple- KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 85 TRACHYPACHIDAE HYDRADEPHAGA / CARABIDAE HYDRADEPHAGA TRACHYPACHIDAE III CARABIDAE TRACHYPACHIDAE HYDRADEPHAGA \ CARABIDAE FIGURE 2 1 . Illustrations of alternative hypotheses of phylogenetic relationships among Hydradephaga, trachypachids, and carabids. siotypic features, then adephagan groups could be related as in Figure 21c. Evidence for relationship between amphizoids and trachypachids As noted above, several workers (Bell 1966, 1982; Evans 1977, 1982; Hammond 1979; and Roughley 1981) have provided evidence in sup- port of close relationship between trachypachids and the dytiscoids, including amphizoids, hy- grobiids, noterids, and dytiscids. All of these au- thors presumed a terrestrial origin for Adephaga. Hammond (1979) listed 7 and Roughley (1981) 10 (for a total of 14 different) proposed syn- apotypies uniting these groups. Each should be considered separately, in light of all available data about extant and fossil forms. 1. Antennal pubescence. Both Hammond and Roughley considered the glabrous antennae of adult hydradephagans to be apotypic, with the plesiotypic state— antennae pubescent— associ- ated with carabids (i.e., terrestrial forms). The condition in trachypachids— glabrous, except for an apical whorl of setae on each antennomere and fine pubescence on antennomere 1 1 only— was considered synapotypic with the condition found among Hydradephaga. Most Coleoptera have flagellar antennomeres covered with a dense coat of short sensory setae, as do adults of most other insect orders; and presence of such pubes- cence would appear to be plesiotypic for Adeph- aga. If we assume an aquatic or subaquatic an- cestry for the suborder, however, presence of pubescence could be interpreted as apotypic in 86 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 carabids; pubescence on antennae of trachy- pachids, albeit greatly restricted, could be syn- apotypic for trachypachids and carabids. 2. Open procoxal cavities with postcoxal bridge. This combination of two characters (i.e., (a) procoxal cavities open or closed and (b) post- coxal bridge absent or present) is difficult to in- terpret. Most workers agree that open procoxal cavities represent the plesiotypic state of the first character (a). However, presence or absence of a postcoxal bridge (b) is more difficult to interpret. A bridge has been reported from trachypachids and dytiscoids and cited as a synapotypy for these groups. However, Hlavac (1975) and Hammond (1979) noted the presence of a bridge in adults ofCarabus, Hiletus, and the nebriine genus Leis- tus, whereas no bridge is evident in members of related carabid groups, including other nebriine genera (i.e., Nebria and Pelophild). Presence or absence of a postcoxal bridge does not appear to be a reliable character for demonstrating phy- logenetic relationships among Adephaga. 3. Prosternal process. Roughley (1981) pro- posed a similarity (synapotypy) among trachy- pachids and Hydradephaga in length and shape of the prosternal intercoxal process. I disagree with this contention. The process in both trachy- pachids and amphizoids (Fig. 11, 12) is very similar to that in Nebriini, Notiokasiini, and oth- er basal-grade carabids in size and shape and unlike more specialized aquatic adephagans such as dytiscids and hygrobiids. I regard this char- acter as symplesiotypic in trachypachids, am- phizoids, and carabids, apotypic in higher dytis- coids. 4. Prosternal-metasternal contact. Roughley (1981) suggested that contact between the pro- sternal intercoxal process and the anteriormost portion of the metasternum was possible in trachypachids as in most dytiscoids and, further, that this represented a synapotypic feature. As with character 3 (prosternal process) above I dis- agree with this interpretation. Contact between prosternum and metasternum is no greater in either trachypachids or amphizoids than in ne- briines or other basal-grade carabids. The con- dition in amphizoids, trachypachids, and basal- grade carabids is surely symplesiotypic. 5. Coadaptation of posterior border of prono- tum and anteriorly truncate elytra. Hammond (1979) proposed that coadaptation of the pos- terior pronotal margin and elytral base among trachypachids and hydradephagans represents a synapotypic feature. It is true that pronota and elytra are very closely juxtaposed in trachy- pachids and most dytiscoids, but no more so than in a number of carabid groups (e.g., omophron- ines, amarines, and some migadopines, bembi- diines, pterostichines, and harpalines). Close, relatively inflexible association of prothoraces and pterothoraces and a continuous, evenly arcuate, uninterrupted lateral silhouette (also including the head in many instances) is broadly charac- teristic of aquatic beetles. Assuming an aquatic origin for Adephaga, this form may represent the plesiotypic condition. Apparent support for this interpretation is provided by Ponomarenko's (1977) numerous illustrations of Mesozoic Adephaga. Among beetles illustrated, including eodromeine trachypachids and protorabine ca- rabids, a form typical of extant trachypachids predominates. This evidence suggests that early carabids were more similar in form to extant trachypachids than to a majority of extant cara- bids. Perhaps the relatively narrow-waisted, flexibly-joined carabids are apotypic rather than plesiotypic in this regard, with members of groups such as omophronines and amarines having ac- quired a trachypachid-like form secondarily, as an adaptation to particular, specialized biotopes. 6. Metacoxal cavities interrupted ("incom- plete," Bell 1966). As discussed above, meta- coxae of trachypachids, amphizoids, and other Hydradephaga (both extinct and extant) are sim- ilar in form and lateral extent to those seen in Archostemata, including presumptive schizo- phoroid ancestors of Adephaga. The main dif- ference between these adephagan metacoxae and archostematan metacoxae is that the former, unlike the latter, are countersunk into the body wall (i.e., into the basal abdominal sterna), so that they appear to divide the first (basal) visible sternum into two lateral parts. Continuity of this sternum internal to the metacoxa (i.e., dorsally) can be confirmed by dissection. I agree with Po- nomarenko (1977) that this form of metacoxae is plesiotypic in Adephaga, rather than apotypic as suggested by most recent workers (Bell 1966, 1967, 1982; Evans 1977, 1982; Hammond 1979; Forsythe 1981). Evans (1977, 1982), Forsythe (1981), Hammond (1979), and others have con- structed and/or reviewed various hypotheses to explain why trachypachids should have meta- coxae preadapted for aquatic life, and why car- KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 87 abids should have even partially immobilized coxae for rapid running. Again, these workers assumed a terrestrial origin of Adephaga. In light of both out-group comparisons with Archoste- mata and form and structure of known Mesozoic Adephaga, it seems simplest to suggest that trachypachids have legs adapted for aquatic life because their ancestors were aquatic, and cara- bids have immobilized coxae because, like trachypachids, their ancestors lived in the water and such coxae are advantageous there. Carabid leg structure is adapted to terrestrial life, but it still reflects the constraints of ancestry. 7. Metacoxal fusion. Roughley (1981) stated that "in trachypachids and Hydradephaga the metacoxae are fused medially, the fusion being marked by a single internal intercoxal septum continuous with the metafurca and the median sternal ridge." My own dissections do not sub- stantiate the extent effusion Roughley reported. In both trachypachids and amphizoids, the me- dial walls of the metacoxae are not fully fused to form a single septum, but are merely very closely approximated, slightly more so in amphizoids than in trachypachids. Roughley is correct in not- ing the close association between the metacoxae and the metafurca and median sternal ridge in adults of these groups. In the carabids examined, the metafurca is positioned far forward in rela- tion to the metacoxal base medially. It is unclear, however, which of these conditions (states) is apotypic. Presumably the state seen in trachy- pachids and Hydradephaga is an adaptation to "aquatic existence" (Roughley 1 98 1 :276). Again, assuming an aquatic ancestry, the state seen in carabids could be apotypic, rather than plesio- typic as Roughley and others have suggested. 8. Similarities in wing venation and folding. According to Hammond (1979), hindwings of trachypachids and dytiscoids share numerous features (e.g., wing folding pattern, position of oblongum cell in relation to apical and posterior wing margins). Adephaga are characterized by having an exceptionally strong spring mecha- nism for wing folding, which is aided, in a ma- jority of groups, by one or another kind of ab- dominal movement. Almost complete reliance on the spring mechanism alone is seen among the related, basal-grade carabid tribes Nebriini, Opisthiini, Notiophilini, Carabini, and Cicin- delini. Hammond (1979) interpreted this latter condition as (probably) plesiotypic for Adepha- ga; but he noted that this hypothesis requires that increased reliance on abdominal movements, and development of special structures associated with same, occurred independently in several adeph- agan lineages. Again, without information from direct out-group comparisons with other Co- leoptera, especially Archostemata, it is difficult to recognize the most plesiotypic condition with any confidence. Although they may represent only a basal grade of carabid evolution, nebriines, opisthiines, and the other groups listed above may also form a monophyletic assemblage that diverged from other carabids at an early evolu- tionary stage, members of which are character- ized by sole reliance on the spring mechanism for wing folding. 9. Subcubital binding patch of hindwing. Both Hammond (1979) and Roughley (1981) cited presence of this binding patch, posteriorly near the apex of the hindwing, as a synapotypic fea- ture uniting trachypachids with dytiscoids. Ab- sence of such a binding patch from hindwings of carabids, haliplids, and gyrinids was seen as a plesiotypic condition. Not all dytiscoids, how- ever, have the binding patch; in all Systolosoma (Trachypachidae) adults that I examined, the patch was nonpigmented and very poorly de- fined, if it could be claimed as present at all. A subcubital binding patch is absent from hygro- biid wings and from wings of members of some bidessine, hydrovatine, and hyphydrine dytiscid genera. Hammond (1979) noted a marked as- sociation between small body size and absence of the binding patch in the dytiscid groups cited above. He suggested functional reasons why the subcubital binding patch might not be necessary in small beetles and proposed that its absence represented a secondary loss in the above dytis- cid groups. Size considerations do not, however, account for the absence of the patch from hy- grobiid hindwings nor its reduction or absence from Systolosoma adults. Although I see no rea- son to doubt that presence of the subcubital bind- ing patch is an apotypic feature in Adephaga, I suggest that its absence from hygrobiid and car- abid hindwings may also represent secondary losses. If this is correct, presence of the binding patch may, in fact, be synapotypic for Adephaga, with its secondary loss having evolved indepen- dently in some or all members of the dytiscoid, trachypachid, carabid, and haliplid lineages. 10. Male genitalia with long, apically nar- 88 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 rowed parameres. Parameres of trachypachid males and of at least some dytiscoid group males are very similar in length, shape, and vestiture. Hammond (1979) suggested that the long, api- cally narrowed form seen in males of these groups represented a synapotypic feature. However, males of certain carabid groups, including car- abines, cychrines, pamborines, and cicindelines also have parameres resembling those of trachy- pachids in form. It seems simpler to suggest that this condition represents a plesiotypic rather than apotypic condition, with the great diversity of forms seen among extant carabids having evolved through several independent, apotypic trends di- verging from the basic form. 11. Size and armature of internal sac. Rough- ley (1981) and other workers have assumed that presence of a well-developed internal sac found in the median lobe of the aedeagus, such as in carabid males, represents the plesiotypic condi- tion among Adephaga; he further suggested that presence of an armature of setae and spines, on or in the sac, is also plesiotypic. Without knowl- edge of these characteristics in proposed schi- zophoroid common ancestors of Adephaga, nor even in extant archostematan males (such as in Omma species), it is difficult to interpret differ- ences in size and development of the internal sac among extant Adephaga in a cladistic sense. As Roughley suggested, it is also possible that the small, slightly developed internal sac of trachy- pachids and dytiscoids represents the plesiotypic adephagan condition. Male gyrinids, which ap- pear to be only distantly related to other Adeph- aga, based on many other characters (Evans 1 982; Ponomarenko 1977), also have a slightly devel- oped internal sac. This suggests that the large, well-developed internal sac of carabids repre- sents an apotypic, rather than plesiotypic, con- dition. Some basal-grade carabid males (e.g., ne- briines, notiokasiines, and opisthiines) lack ev- ident armature on the internal sac. Although as- sociated spines and/or setae are found in males of some basal-grade carabids and are widespread among those of middle- and high-grade carabid groups, I see no reason to suggest that their oc- currence represents a plesiotypic condition among Adephaga, and I do not consider their absence to be synapotypic for trachypachids and dytis- coids. 12. Dilator muscle of vagina. The presence of this muscle in a majority of dytiscoids examined (Burmeister 1 976) led Roughley ( 1 98 1 ) to suggest that its occurrence represents a synapotypic fea- ture among dytiscoids (including amphizoids) and trachypachids. Its absence from carabid fe- males was considered plesiotypic. The source of Roughley's data for trachypachids and amphi- zoids (Roughley 1981, table 1) is unclear; but I assume that these data are from his own dissec- tions because Burmeister (1976) did not present data for these groups (see his table 1, p. 216). Assuming that Roughley is correct, and this mus- cle is present in trachypachids and amphizoids as well as in haliplids, gyrinids, hygrobiids, and most dytiscids, but not in carabids (Burmeister 1976), it would seem simpler to suggest that its presence is plesiotypic, and its absence (in car- abids and a few dytiscids) apotypic among Adephaga. As with the previous character, it will be useful to examine extant archostematans as a possible out-group test of alternative hypotheses. 13. Giardina bodies. Roughley (1981) suggest- ed that the nature of so-called "Giardina bod- ies," which contain extrachromosomal DNA and appear in oogonia at the preoocyte stage of oo- genesis, might represent a synapotypic feature for dytiscoids and trachypachids. He noted that these bodies "appear to be of a different type in Dy- tiscoidea than in other insects." They have been found in female representatives of Gyrinidae, Hygrobiidae, and some Dytiscidae studied. For example, they occur in Colymbetinae, Lacco- philinae, and some (e.g., Hydaticus, Dytiscus), but not all (e.g., Eretes, Cybister), dytiscines, and are absent from the few hydroporines studied. More significantly, however, their presence (or absence) remains unknown for noterids, hali- plids, amphizoids, trachypachids, and carabids. Roughley's primary intent was to initiate a sur- vey of the occurrence of Giardina bodies among Adephaga— to introduce a new character into adephagan systematics. Available data cannot support the hypothesis that presence of a partic- ular type of Giardina body is synapotypic for dytiscoids and trachypachids. 14. Ligula absent from labium of larva. Ham- mond (1979) cited this character state as a pos- sible synapotypic feature uniting trachypachids and dytiscoids; but he noted that a ligula is absent from larvae of various carabid groups (e.g., Bra- chinus, Gehringia, and lebiines) as well. Distri- bution of this characteristic among extant Adephaga is not yet fully known, nor have de- tailed out-group comparisons with archostema- KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 89 Paleogene Neogene SCHIZOPHORIDAE ( A RCHOSTEMATA ) (8a) ,,,, .llc.iaa", (1a'.3a,l?a'.16o.l7a.18a,19a') 8a,11b,13a,25a 12a*'l6a* / ?8a,29a (8a , 1 1b, 13a,22a) A GYRINIDAE B SpanglerogyruE E HYCROBIIDAE F NOTER1DAE G DYTISCIDAE I AMPHIZOIDAE L TRACHYPACHINAE N CARABINAE 0 HALIPLIDAE P TRIAPLIDAE FIGURE 22. Reconstructed phylogeny of Adephaga, including both extinct and extant groups. Time is represented by the horizontal axis; but neither position nor gap width on the vertical axis is intended to reflect divergence considerations. Thickened portions of tree branches indicate known temporal occurrence in the fossil record. Number and letter symbols placed adjacent to solid dots refer to synapotypic features presented in Table 1 and discussed in the text. Symbols in parentheses refer to apotypic features found in some, but not all, members of the lineage directly below them. tan larvae been made. Therefore, significance of the co-occurrence of this feature among dytis- coids and trachypachids cannot be properly eval- uated. It may represent another symplesiotypy for Adephaga. In summary, there is little, if any, unequivocal evidence to support strict monophyly of a group including dytiscids, hygrobiids, amphizoids, and trachypachids but excluding carabids. This view is based on a re-evaluation of character polarities proposed and/or supported by Hammond (1979), Roughley (1981), and several other workers (e.g., Bell, Crowson, Evans, and Forsythe, as previ- ously cited). These workers may be correct in their interpretations. Nonetheless, I offer an al- ternative interpretation of data and relationships as perceived from my studies of nebriines and other basal-grade carabids over the past few years; I hope that these interpretations and conclusions will be rigorously tested by current and future colleagues. A hypothesis ofadephagan phylogeny The hypothesis ofadephagan relationships that I have used below as a basis for out-group com- parison in cladistic analysis of amphizoids is il- lustrated in Figure 22. Some relationships pro- posed are highly speculative in relation to available data, and relatively few characters have been adequately studied and applied to a cladis- tic analysis of Adephaga. Consequently, the monophyly of certain groups proposed is not substantiated, or is inadequately substantiated, 90 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 by synapotypic features at present. Nevertheless, I hope that others will be encouraged to challenge proposed relationships through a search for ad- ditional apotypic features that support or refute the phylogenetic hypothesis. In this regard, a comprehensive comparative study of larval structure, including what can be gleaned from review of fossil material, will undoubtedly pro- vide extremely valuable data. In both comparisons made and conclusions drawn, I have accepted adephagan family limits as presently denned. Some of the family-group taxa so delimited may not represent strictly monophyletic groups, and better understanding the phylogenetic relationships among some of these so-called "families" (e.g., dytiscids and no- terids) will require further cladistic analyses of member subgroups and relationships among them. Familial status of certain taxa known only from fossils (e.g., liadytids, parahygrobiids, and coptoclavids) is unclear, but I have accepted pro- posed familial ranking for each herein to facili- tate comparisons with extant taxa of familial rank. Based mainly on Ponomarenko's (1977) re- view of Mesozoic fossil material, I have also tried to relate the branching sequence of the proposed cladogram to geologic time (but not specifically to events in Earth history). Among possible sources of error in establishing timing of diver- gent events in adephagan phyletic history are: (1) that fossil occurrence of a group provides only a minimum estimate of the time of its origin, and disappearance or absence of a group from the known fossil record does not rule out its existence at a particular time or place; and (2) that the geographical distribution of currently available material which represents the Mesozoic (and ear- ly Cenozoic) adephagan fauna is highly biased. Almost all useful specimens are from Asia, and some groups, such as amphizoids, hygrobiids, and haliplids, may well have evolved in other areas and much earlier than the known fossil record suggests. Of the five hypotheses of adephagan relation- ships reviewed by Bell (1982), that of Ponoma- renko (1977) is most similar to the one proposed here. Ponomarenko suggested that the common adephagan ancestor gave rise to three major, in- dependent lineages, which Bell (1982) termed the "haliplomorph," "dytiscomorph," and "cara- bomorph" ancestral lineages, respectively. Ac- cording to Ponomarenko, extant haliplids may be descendants of the Triassic haliplomorph group, Triaplidae; gyrinids diverged, probably in Lower Triassic time, from the common ancestor of other dytiscoids (including amphizoids, hy- grobiids, dytiscids, and a number of extinct Me- sozoic forms); and extant carabids and trachy- pachids are descendants of a common, terrestrial carabomorph ancestor, which also evolved in, or just before, the Triassic. The only major difference between Ponoma- renko's hypothesis and that illustrated in Figure 22 involves the relationship of haliplids to other Adephaga. I doubt that any close phylogenetic relationship exists between haliplids and tria- plids. Evidence cited by Ponomarenko as linking these two groups more likely represents conver- gence. Instead, several synapotypic features link haliplids with caraboids, and triaplids probably have no extant descendants or near relatives. Several other relationships proposed here are noteworthy. The recently discovered gyrinid ge- nus Spanglerogyrus (Folkerts 1979) appears to be a relict form less closely related to other extant gyrinids than is the Upper Triassic form, Tri- adogyrus (Ponomarenko 1977) (see details be- low). Nothing is known about external structure of parahygrobiid adults, and placement of this group in the cladogram is problematic at present. Some evidence exists to link hygrobiids with the extinct coptoclavids rather than with other ex- tant dytiscoids. Both coptoclavids and hygro- biids appear to be more closely related to gyrinids than to dytiscids and amphizoids. The Lower Jurassic form, Necronectulus (Ponomarenko 1977), known from only a single specimen with- out legs, shares apotypic features with no known adephagan lineage. I have, therefore, indicated its derivation from an unresolved trichotomy with the dytiscomorph and carabomorph lineages. It may be related to either of these lines, but evi- dence for one or the other affinity is currently lacking. Evidence in support of relationships proposed in Figure 22 is presented in Table 1 . Code letters used for taxa in the table are the same as those used in Figure 22. Coding of character states, both in Figure 22 and Table 1, is as follows: (1) each character is represented by a unique, Arabic number; (2) the plesiotypic state of each char- acter is represented by the letter o; (3) indepen- dently derived apotypic states are represented by different letters (a, b, etc.), where states a and b KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 91 evolved independently from state o; (4) sequen- tially derived apotypic states are represented by a letter (a) or a letter plus asterisk (a*), where state a evolved from state o and state a* evolved from a; and (5) apotypic states that include a combination of independently and sequentially derived conditions are represented by letters (a, b, etc.) and letters with different symbols (a*, a#, etc.), where states a and b evolved independently from state o, and both a* and a# evolved inde- pendently from state a. Polarities of transfor- mation for 25 of the 29 characters used for cla- distic analysis were determined by means of the out-group criterion. The character correlation criterion was used to determine polarities for characters 4 (antennal pubescence), 27 (gonostyli of female ovipositor), 28 (thoracic defense glands), and 29 (pygidial defense gland cells). Implica- tions of the distributions of states of the char- acters presented in Table 1 in relation to the cladogram in Figure 22 are as follows. Character 1. General habitat. If a semiaquatic lifestyle, similar to that of extant amphizoids, is accepted as plesiotypic for Adephaga, then a fully aquatic lifestyle may have evolved only twice: in a lineage including all Hydradephaga except amphizoids and haliplids, and in haliplids. A more highly evolved lifestyle, one specializing in water surface activity apparently evolved twice— once in gyrinids, and again in some coptoclavids (see Ponomarenko 1977). Haliplids and amphi- zoids swim with an alternating (walking) leg mo- tion. In the former group, this trait may reflect a semiaquatic (or even terrestrial) ancestry and independent adaptation to fully aquatic life. Ad- aptation to passive drifting in streams shown by amphizoids is no doubt an apotypic feature. Character 2. Food habits/feeding. Ponoma- renko (1977) suggested that triaplids and hali- plids shared herbivorous feeding habits, but he noted also that this trait could have been plesi- otypic in triaplids. If the relationship of haliplids to caraboids proposed here is correct, then algal feeding must be apotypic in haliplids. Character 3. Compound eyes. Both gyrinids and a majority of known coptoclavids have com- pound eyes divided into dorsal and ventral por- tions. Based on other characteristics, this co- occurrence appears to be convergent in the two groups. In all extant gyrinids, except Spanglero- gyrus adults, the dorsal and ventral eye portions are moderately or broadly separated by an an- terior extension of the gena. In Spanglerogyrus adults, the eye portions are broadly contiguous, with their division marked by only a thin sep- tum. This feature, in combination with others listed below, suggests a very ancient divergence of this monobasic group from the main line of gyrinid evolution. Character 4. Antennal pubescence. As noted above, this character is problematic. Other au- thors (e.g., Roughley 1981) have suggested that absence of antennal pubescence is apotypic, a trait evolved in association with the change to an aquatic lifestyle. Yet terrestrial trachypachids lack antennal pubescence (except on antenno- mere 1 1) and aquatic gyrinids have pubescence (but of a peculiar form and distribution). If, as I suggest here, presence of antennal pubescence is an apotypic feature where it occurs in Adephaga, then this trait may have evolved only twice: once in the lineage including trachypachids and car- abids, and again in gyrinids. The minimal pu- bescence seen in extant trachypachids can be in- terpreted as a first step in a transformation series leading to the condition found in a majority of carabids. Character 5. Orientation of mouthparts. Po- nomarenko (1977) suggested that apparent opis- thognathy seen in triaplid fossil specimens may reflect a grazing style of feeding, characteristic of a variety of herbivorous beetle groups. The known occurrence of opisthognathy among schizopho- roid Archostematan an Adephagan is such that it must be apotypic for triaplids. Character 6. Prosternal intercoxal process. Most extant and extinct Archostemata and Adephaga have a well-developed prosternal in- tercoxal process. Known triaplids appear to have lacked such a process, at least externally. This probably represents an apotypic feature. In re- lation to those of other groups, haliplids, omo- phronine carabids, and some noterids have intercoxal processes markedly expanded and strikingly similar in form and degree of contact with the mesothorax. However, adult haliplids and noterids have open procoxal cavities, where- as omophronines have them closed. While this difference may be significant, Bell (1967) pointed out that the type of procoxal closure found in omophronines was apparently unique to them. Hence, it is likely that the immediate ancestor of omophronines had open procoxal cavities. Shape of the prosternal intercoxal process is just 92 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 TABLE 1 . DISTRIBUTION OF STATES OF SELECTED CHARACTERS AMONG MEMBERS OF CERTAIN SUPRA-SPECIFIC TAXA OF ADEPH- AGA (COLEOPTERA) (See Fig. 22 for Code Letters for Taxa and Text for Discussion of Characters). Taxa and character state distributions Character Character state ABCDEF GHIJK LMNOP 1. General habitat a* a* a, a* a a a a ?c?b b bb a? Semiaquatic, o Aquatic, a Aquatic, surface, a* Terrestrial, b Semiaquatic, passive drifter, c 2. Food habits/feeding o oo ooo o oooo o oo a? Predaceous, o Herbivorous (on algae), a 3. Compound eyes a a o, a ? o o o oooo o oo oo Undivided, o Dorsoventrally divided, a 4. Antennal pubescence a* a*? ?oo o ?o?? a ?a*o? Without pubescence, o Only antennomere 1 1 pubescent, a Pubescence widespread, a* 5. Orientation of mouthparts o oo ?oo o oooo o oo oa Prognathous, o Opisthognathous, a 6. Prosternal intercoxal process o oo ? o o, bo oooo o oo ba Narrow, o Absent, a Broad, b 7. Protibial antenna cleaner o o o ? o o o oo?a a aa o? Absent, o Present, a 8. Scutellum o, a o o ?oa o, a o o o o o oo ao Visible externally, o Concealed, a 9. Mesothoracic length a ao ?oo o oooo o oo oo Short, o Long, a 10. Mesocoxal shape a ao ?oo o oooo o oo oo Round, o Laterally expanded, a 1 1 . Ventral mesocoxal articulation c c? ?ob o, b ? o ? ? a ? a, a* a* ? Absent, o Coxal lobe, sternal stop, a Coxal peg, sternal socket, a* Coxal groove, sternal ridge, b Coxae otherwise immobilized, c 12. Metasternal transverse ridge a* a* a, a* ? a a* a* a a o o a o o o o Present, laterally extended, o Present, laterally reduced, a Absent, a* 13. Relationship of metepisternum to mesocoxal cavity a oa ?aa o, a o o o o a oa ao Forms part of lateral wall, o Excluded from lateral wall, a 14. Metacoxal position a aa ? a a a aaaa a aa aa Free of abdomen, o Countersunk into abdomen, a 15. Metacoxal width o oo ?oo o oooo o aa oo Wide, o Narrow, a KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 93 TABLE 1. CONTINUED. Taxa and character state distributions Character Character state A BCDEF GHIJK LMN OP 16. Metacoxal length o o o, a ? a a* a* aaoo o oo oo Short, o Medium, slightly expanded anteriorly, a Long, markedly expanded anteriorly, a* 17. Metacoxal fusion a* a* a, a* ? a* a* a* a* a o o o o o o o Not fused medially, o Partially fused medially, a Extensively fused medially, a* 18. Metacoxal femoral plates o o o, a ? o o, a o o o o o, a o oo a* a Absent or small, o Present, moderately large, a Present, very large, a* 19. Legs, distal modifications for swimming a* a* a, a* ? a a a oo?o o oo b? Absent, o Slight modifications, a Extensive modifications, a* Femoral modifications only, b 20. Legs, fringe setae a aa ?aa a ao?b b bb a? Present, slightly developed, o Present, well developed, a Absent, b 21. Hindwing apex in repose a a? ?aa a ?a?? a ?a a? Spirally rolled, o Folded, a 22. Hindwing, subcubital binding patch a* a*? ?oa o, a ? a ? ? o, a ? o o? Present, o Absent, a Absent, suboblongum patch present, a* 23. Oblongum cell position a a? ?ao o ?o?? o ?a a? Posteroapical, o Near center of wing, a 24. Male median lobe, internal sac o o? ?oo o ?o?? o ?a o? Short, slightly developed, o Large, better developed, a 25. Male genitalia, parameres o o? ?oa o ?o?? o ? o, a a ? Symmetrical in length and shape, o Asymmetrical in length and shape, a 26. Male genitalia, ring sclerite o o? ?oo o ?o?? a ?a a? Split posterodorsally, o Complete posterodorsally, a 27. Female ovipositor, gonostylus a a? ?aa a ?a?? a ? a, a* a ? Distinct, o Fused with gonocoxite, a Apparently distinct, a* 28. Thoracic defense glands o o? ?ao a ?o?? o ?o o? Absent, o Present, a 29. Pygidial defense gland cells o o? ?oo a ?o?? o ?o o? Type I cells absent, o Type I cells present, a 94 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 one of several similarities (see below) among no- terids, haliplids, and omophronines that appear to represent convergences, based on data from other characters. Character 7. Protibial antennal cleaner. Sev- eral authors (e.g., Hammond 1979) have sug- gested that absence of a protibial cleaning organ may be an apotypic feature among hydradepha- gans. Also among carabid groups (e.g., paussines) in which specialized antennal structure precludes grooming by means of a protibial cleaning organ, such an organ is absent. I see no evidence, how- ever, to suggest that presence of a protibial clean- ing organ is plesiotypic among Adephaga, and I view its occurrence as an apotypic feature linking trachypachids and carabids. Character 8. Scutellum. A scutellum is visible externally in extant and extinct Archostemata and in Adephaga, except noterids, haliplids, omophronine carabids, some gyrinids, and some dytiscids. Because the distributions of apotypic states of several other characters are incongruous with the distribution of a concealed scutellum among Adephaga (i.e., character correlation cri- terion), there is little evidence to suggest that this trait is synapotypic for any two or more of the exceptional taxa. It probably evolved indepen- dently in each, although its co-occurrence among noterids and certain dytiscids may reflect close phylogenetic relationship. Character 9. Mesothoracic length. A signifi- cant increase in length of the mesothorax is seen in extant and fossil gyrinids, including adults of Spanglerogyrus and the inadequately known Triassic fossil form, Triadogyrus. This feature appears to be autapotypic (i.e., uniquely derived) in gyrinids. Character 10. Mesocoxal shape. Distribution of states of this character is identical with that in character 9. No doubt, the two characters are closely correlated. Among known Adephaga, only gyrinids have laterally expanded mesocoxae. Even coptoclavids, which share several other features with gyrinids, had round mesocoxae typical of the remainder of the suborder. Character 11. Ventral mesocoxal articulation. Evans (1977) reviewed various structural means found among Adephaga for ventral articulation of mesocoxae with the metasternum. He noted that amphizoids, hygrobiids, and some dytiscids evidently lack special structural means of ventral articulation. Carabids and extant trachypachids (i.e., adults of both Trachypachus and Systolo- soma species) have a coxal lobe/sternal stop mechanism, but noterids and those dytiscids with evident ventral articulations have a sternal ridge/ coxal groove arrangement. Gyrinids have me- socoxae that are practically immobilized by a unique structural arrangement which is probably independently derived. It seems that the absence of ventral articular structure is plesiotypic in Adephaga, and therefore, that articular struc- tures evolved independently in (1) noterids and some dytiscids, and (2) the lineage including trachypachids and carabids. A special coxal peg/ sternal socket arrangement is found in haliplids and omophronines (Evans 1977), although po- sition of the socket is different in members of the two groups. This is yet another similarity be- tween these groups, but it probably evolved in- dependently in each from the coxal lobe/sternal stop arrangement seen in other caraboids. Character 12. Met asternal transverse ridge. Evans (1977) discussed this structure (also known as the "metasternal suture"), its functional sig- nificance, and its distribution among Adephaga. Its presence appears to be plesiotypic and its loss or lateral reduction apotypic within the suborder. The single known Necronectulus specimen has a well-developed transverse ridge. Presence of the laterally reduced ridge in amphizoids and hygro- biids suggests that, if the cladogram is correct, its loss has occurred three times independently: in dytiscids and noterids, in gyrinids, and in some (but not all) coptoclavids. Eodromeines appar- ently had well-developed, laterally extended transverse ridges, like extant carabids, and so the laterally reduced ridge found in extant trachy- pachids probably represents reduction conver- gent with that in the dytiscomorph lineage. Stei- ner and Anderson (1981) reported presence of a metasternal ridge in adults of Spanglerogyrus. In my own examination of representatives of this genus, I found the structure in question to be wholly part of the metacoxa rather than the metasternum. I suggest that the suture (or ridge) at the base of the metacoxa in Spanglerogyrus adults is autapotypic among them and not ho- mologous with the metasternal ridge found among Adephaga as listed in Table 1 . Character 13. Relationship of metepisternum to mesocoxal cavity. As noted earlier, the plesio- typic condition among Adephaga is that in which the metepisternum contributes to the lateral wall of the mesocoxal cavity. Among extant forms, this condition is found only in amphizoids, some KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 95 dytiscids, and adults of the gyrinid genus, Span- glerogyrus. Although this condition may have been achieved secondarily in members ofSpan- glerogyrus, it is more likely that it represents yet another feature suggesting ancient ancestry for this unique genus. Exclusion of the metepister- num from the mesocoxal cavity appears to have evolved at least seven times: in (1) gyrinids (after divergence of Spanglerogyrus from the main lin- eage), (2) the lineage including hygrobiids and coptoclavids, (3) noterids, (4) some dytiscids, (5) trachypachines [this is the feature that distin- guishes them from eodromeines], (6) carabids [again, this feature distinguishes extant carabids from protorabines], and (7) haliplids. This struc- tural change must be highly advantageous me- chanically for it to have become fixed in so many different lineages making use of both terrestrial and aquatic habitats. Character 14. Metacoxal position. In all known Adephaga, the metacoxae are countersunk into the base of the abdomen so that they divide the first visible abdominal sternum externally into two triangular lateral portions. This feature dis- tinguishes Adephaga from other Coleoptera, in- cluding Archostemata. It is, no doubt, a syn- apotypic feature. Character 15. Metacoxal width. The narrowed metacoxae found in protorabines and all extant carabids (except gehringiines) are clearly apotyp- ic. The condition found in rhysodines is not equivalent to the plesiotypic state, because the metacoxae extend laterally only to the postero- lateral corner of the metasternum, just as in car- abids. The metepisterna are hidden posteriorly under the elytral epipleura, but they are com- pletely laterad of the lateral margins of the meta- coxae. Evans (1977) noted that, unlike those in dy- tiscids, gyrinids, and other Hydradephaga, meta- coxae of haliplids have a lateral coxal condyle, as do carabid metacoxae. I agree this feature in- dicates close affinity with a presumed terrestrial ancestor, namely carabids. However, a coxal condyle is also present, although not as well de- veloped, in extant trachypachines and amphi- zoids, but not in hygrobiids and other dytiscoids. Presence of a coxal condyle may represent the plesiotypic condition among Adephaga. The lat- erally extended metacoxae of haliplids may rep- resent either the plesiotypic adephagan condition or secondary acquisition of a similar condition as part of an adaptation of metacoxae for a new function (see further discussion under character 18). Character 16. Metacoxal length. Slight to moderate expansion of metacoxae anteriorly, and attendant reduction in size of the metasternum, is seen in amphizoids, liadytids, hygrobiids, and some coptoclavids. In dytiscids and noterids, metacoxae are greatly expanded anteriorly. Based on distributions of states of other characters, it is likely that the trend for anterior expansion, which was initiated in the common dytiscoid ancestor, has been reversed at least twice inde- pendently: in gyrinids and in some coptoclavids. Character 17. Metacoxal fusion. This char- acter was discussed above in consideration of the relationship between amphizoids and trachy- pachids. Based on character states represented in extinct and extant Archostemata and Mesozoic Adephaga, it is clear that the unfused metacoxae are plesiotypic. As noted by Evans (1977) and others, the metacoxae of haliplids are not fused medially. It is therefore likely that at least partial medial fusion of metacoxae represents a syn- apotypic feature for Hydradephaga exclusive of haliplids (and Necronectulus, if its members were aquatic). A trend for more extensive medial fu- sion of metacoxae may have evolved only once— in the common ancestor of all dytiscomorphs except amphizoids. If so, then this trend was also reversed at least once, in the ancestor of some (but not all) coptoclavids. Eodromeine trachy- pachids appear to have had more widely sepa- rated metacoxae than extant trachypachids. Hence, a trend toward increased medial conti- guity, if not fusion as suggested by Evans (1977) and Roughley (1981) for extant trachypachids, probably represents a development independent of that in Hydradephaga. Character 18. Metacoxal femoral plates. Po- nomarenko (1977) described posteroventral ex- tensions of metacoxae, which he termed femoral plates, in triaplids, some coptoclavids, and some eodromeines among Mesozoic fossil forms. He noted that such plates could be rather easily bro- ken off and, therefore, that their distribution may have been taxonomically more extensive than present fossil material illustrates. He also sug- gested that presence of metacoxal plates may be plesiotypic for Adephaga. Such structures are ap- parently unknown among schizophoroid Ar- chostemata, however, and among extant forms, femoral plates are found only in noterids and haliplids. In my view, it is simplest to consider 96 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 presence of femoral plates as apotypic where they occur among Adephaga. There is little or no evi- dence to suggest that this feature is synapotypic for any two or more of the extinct and/or extant groups whose members are known to possess them. Haliplid metacoxal femoral plates are much larger, both posteriorly and laterally expanded, than those of noterids and the extinct groups listed above, including triaplids. The role of these plates in haliplid respiration has been well doc- umented (Hickman 1931). It is unlikely that they served this highly specialized function in triap- lids, coptoclavids, or eodromeines, and no respi- ratory role has been suggested for them among noterids. Character 19. Legs, distal modifications for swimming. Based on comparisons with Archo- stemata and Mesozoic Adephaga, it appears that distal leg structure in amphizoids represents the plesiotypic state among Adephaga. Special struc- tural modifications of femora, tibiae, and/or tarsi as adaptations for rapid swimming are consid- ered apotypic. Relatively slight modifications of this kind have apparently evolved twice inde- pendently: in the common ancestors of (1) no- terids and dytiscids and (2) hygrobiids, copto- clavids, and gyrinids. In each of these groups, structure of distal leg parts is quite distinctive in detail. Gyrinids and members of the extinct ge- nus Coptoclava (Ponomarenko 1977) are similar in that their middle and hind legs are (or were) markedly flattened and expanded. This feature is probably apotypic relative to more conserva- tive leg modification, but distributions of states of other characters suggest that it is not syn- apotypic for these two groups. Middle and hind legs of haliplids show no spe- cial structural adaptations for swimming. Hali- plid hind femora are unique in that they are markedly narrowed basally— a feature probably evolved to facilitate leg movement within the narrow space between abdominal venter and metacoxal femoral plates. Character 20. Legs, fringe setae. Unfortu- nately, some of the most important extinct Me- sozoic groups are known only from specimens without distal leg parts. Among these are Tri- aplidae and genus Necronectulus. Hence, it is dif- ficult to know whether or not ancestral Adephaga had legs bearing fringe setae (or so-called "swim- ming hairs"). Assuming a semiaquatic ancestry, the condition found in extant amphizoids, in which fringe setae are present but short and lim- ited in distribution, could be considered the ple- siotypic condition. Absence of fringe setae would then be synapotypic for trachypachids and car- abids. If the cladogram in Figure 22 is correct, then more extensive development of fringe setae would also be apotypic. But this feature would have had to have evolved at least twice inde- pendently: in (1) the common ancestor of all Hy- dradephaga except amphizoids, and (2) haliplids. Fringe setae are longer and more extensively dis- tributed in haliplids than in amphizoids— this is perhaps associated with a slightly better devel- oped aquatic lifestyle. Character 21. Hindwing apex in repose. Mem- bers of all Adephaga groups examined have the hindwing apex folded, rather than spirally rolled as in Archostemata. This feature is probably syn- apotypic for the suborder Adephaga. Character 22. Hindwing, subcubital binding patch. If presence of the subcubital binding patch is synapotypic for suborder Adephaga (hence, plesiotypic within Adephaga, see above), then loss of the patch has evolved at least five times: in (1) the common ancestor of carabines and (probably) protorabines, (2) haliplids, (3) trachy- pachids of genus Systolosoma, (4) some dytis- cids, and (5) the common ancestor of gyrinids, hygrobiids, and (probably) coptoclavids. Gyrinid specimens examined have a narrow patch of short setae or long microtrichia along the posterior margin of the oblongum cell that may aid in wing folding as an alternative to or replacement for the subcubital patch. Character 23. Oblongum cell position. Ham- mond (1979) noted that the oblongum cell is positioned closer to the posterior margin of the wing apex in trachypachids, amphizoids, noter- ids, and dytiscids than in other Adephaga and considered this to represent a synapotypic fea- ture for the groups noted. Position of the oblon- gum cell in archostematan hindwings, however, is also close to the posterior margin of the apex, just as in amphizoids and other taxa noted by Hammond. I conclude that this feature is ple- siotypic, and further, that a more anterior and basal placement of the cell is apotypic. If this view is correct, then the apotypic state could have evolved as few as three times: in (1) cara- KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 97 bids, (2) haliplids, and (3) the common ancestor of hygrobiids, coptoclavids, and gyrinids. Character 24. Male median lobe, internal sac. As noted above, it is likely that a large, well- developed internal sac, such as is found in most carabid males, is apotypic among Adephaga. Males of basal-grade rhysodid lineages have larg- er internal sacs than those of more highly evolved lineages, but this trend appears to reverse that seen among carabids in general. Character 25. Male genitalia, parameres. Based on comparisons with genitalia of extant Archostemata, it appears that the plesiotypic form of parameres among Adephaga demonstrates symmetry in both length and shape. Asymmet- rical parameres are found in noterids, haliplids, and most, but not all, carabids. Based on the distribution of this feature in relation to char- acter-state distributions of other characters, it is likely that asymmetry of parameres evolved in each of these groups independently. Character 26. Male, ring sclerite. The ring sclerite (Kavanaugh 1 9786) and associated struc- tures probably represents the sclerotized remains of the ninth abdominal segment (the genital seg- ment, or urite X of Jeannel 1941), and it serves as a rim for attachment of muscles from the base of the median lobe. In all Hydradephaga ex- amined, except haliplids, the ring is split postero- dorsally in the midline, into what might be termed "hemitergites," but is continuous anteroventral- ly (see Edwards 1951, "Plate 2"). This condition is shared with Archostemata males examined. In trachypachids, carabids, and haliplids, however, the ring is complete posterodorsally as well as anteroventrally— a feature that is probably syn- apotypic for these three groups. Character 27. Female ovipositor, gonostylus. Bell (1982) and others have suggested that the apparent absence of a gonostylus (or stylomere two) from ovipositors of female trachypachids, isochaetous carabids, and hydradephagans may represent a synapotypic feature uniting these groups. In fact, a majority of basal-grade carabid groups (e.g., opisthiines, notiokasiines, nebri- inies, and notiophilines) also have females in which a gonostylus is either absent from the ovi- positor or fused with the gonocoxite (stylomere one) so as to appear absent. I agree with Bell that this feature is apotypic, but suggest that it is syn- apotypic for the suborder Adephaga rather than just for a subgroup of that taxon. The structures that have been called gonostyli (or second sty- lomeres) in female carabines, cychrines, cicinde- lines, and a majority of intermediate- and advanced-grade carabids are probably not ho- mologous with the gonostyli of female Archo- stemata and Polyphaga. Character 28. Thoracic defense glands. For- syth (1968, 1970) noted that, among Adephaga, only hygrobiids and dytiscids possess thoracic defense glands in addition to the pygidial defense glands common to all Adephaga. Presence of such thoracic glands is no doubt apotypic in hygro- biids and dytiscids, but based on the character correlation criterion, I agree with Forsyth that this similarity represents convergence rather than common ancestry. Character 29. Pygidial gland cells. In a series of papers describing the structure of pygidial and other defense glands among Adephaga, Forsyth (1968,1970,1972) provided numerous excellent characters, while making detailed comparisons among members of included taxa, but he did not consider states of these characters from a cladis- tic perspective. The relationships he suggested were based on simple similarity, rather than on synapotypy, and I have been unable to recognize patterns of synapotypy among the mass of data he provided for included adephagan taxa. Forsyth (1968) recognized two types of secre- tory cells (Type I and Type II cells) in the pygidial glands of dytiscids. Apparently, only Type II cells are found in these glands in other Adephaga, and presence of Type I cells in dytiscid pygidial glands must be autapotypic. Summary of phylogenetic reconstruction. Sev- eral final points should be made in reference to the proposed cladogram and data provided in Table 1 . First, the monophyly of a lineage in- cluding all Adephaga except triaplids is unsup- ported at present by evidence in the form of syn- apotypic features. We know too little about triaplid structure and lifestyle to recognize fea- tures in which their proposed sister-group may be considered specialized (i.e., apotypic). I also failed to discover any synapotypic feature uniting Trachypachinae with Eodromeinae. However, eodromeines probably represent the ancestral stock from which both trachypachines and car- abids evolved. A group including trachypachines and eodromeines but excluding carabids would 98 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 therefore be paraphyletic, which may explain why synapotypic features for such a group are lacking. Monophyly of a group including both extinct and extant trachypachids, carabids, and haliplids is supported by fewer and less compelling syn- apotypic features than might be desirable. The only proposed synapotypies for this group are the following: (1) mesocoxal ventral articulation by means of a coxal lobe and sternal stop or derivative of this arrangement, and (2) male gen- italia with ring sclerite complete posterodorsally. Nonetheless, available evidence supports a clos- er phylogenetic relationship between haliplids and carabids than between the former and other Hy- dradephaga. As can be seen in Table 1 , noterids share apo- typic features (e.g., see characters 8, 1 1, and 13) with some, but not all, dytiscids. This suggests a close relationship between noterids and only some dytiscids. It is therefore possible that if Dytis- cidae (in the broad sense) is a monophyletic tax- on (and there is considerable doubt in this regard; Roughley, pers. comm.) then it would be a para- phyletic taxon if noterids were excluded and/or recognized as a separate family. On the other hand, dytiscids possess thoracic defense glands and Type I secretory cells in their pygidial de- fense glands, whereas noterids studied to date have neither of these features. Available evi- dence is therefore equivocal with regard to the question of relationship between noterids and dytiscids. However, I suggest that noterids and dytiscids should be taken together as a mono- phyletic unit of greater inclusiveness, whether at the familial or some higher taxonomic level, to assure that appropriate comparisons are made in future studies. The proposed relationship between hygrobiids and coptoclavids also requires further comment. Among characters used in this study, I found no apotypic states that distinguish all members of either group from all members of the other. Some coptoclavids have apotypic features not shared with hygrobiids, but the reverse does not apply, except perhaps for the presence of thoracic de- fense glands in hygrobiids (but coptoclavids may also have had such glands). Hygrobiids are most similar to certain members of Necronectinae (Ponomarenko 1 977). Together, these groups ap- pear to represent a basal grade of coptoclavid evolution, and I predict that future studies will indicate that hygrobiids and coptoclavids should be included in a single family. Phylogenetic relationships of amphizoid species Based on assumed adephagan phylogenetic re- lationships as illustrated in Figure 22, a cladistic analysis was conducted to ascertain relationships among extant amphizoid species. A total of 1 4 selected characters was used. For each, the out- group criterion was used to establish polarity (from plesiotypic to most apotypic) of character- state transformation. Characters and character- state distributions among amphizoid species are presented in Table 2, and the cladogram that results from analysis of these data is illustrated in Figure 23. Format and coding for characters and character states used in Table 2 and Figure 23 are as explained above for Table 1 and Figure 22. If the hypothesis of phylogenetic relationship proposed— namely that Amphizoa davidi is the sister-group of the other three species, and that A. insolens is the sister of the group including A. striata and A. lecontei— is correct, then the fol- lowing comments are appropriate. Character-state distributions of all characters analyzed are compatible with each other over the cladogram, except for character 3 (sinuation of the lateral margin of the pronotum). Develop- ment of a deep sinuation basolaterally is evident in adults of A. davidi and A. insolens. Although nothing is presently known about habitat re- quirements and/or tolerances of A. davidi mem- bers, those of A. insolens are often found in swift- er-flowing, more precipitous streams than are members of A striata or A. lecontei. A deep sub- basal sinuation of the lateral pronotal margin is also found in certain dytiscids (e.g., members of genus Hydronebrius Jakovlev and of the cordatus group ofAgabus), which also live in fast-flowing streams. This suggests that the apotypic state of this character (i.e., lateral margin deeply sinuate sub-basally) may be associated with adaptation to life in swift-flowing streams, and distributions of states of other characters suggest that this fea- ture evolved independently in A. davidi and A. insolens. An alternative, equally parsimonious interpretation of the distribution of character states is that a deep, sub-basal sinuation evolved among members of the common ancestor of ex- tant Amphizoa species and is therefore synapo- typic for the genus. An evolutionary reversal then occurred in members of the common ancestor of A striata and A. lecontei. If this interpretation KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 99 TABLE 2. DISTRIBUTIONS OF STATES OF SELECTED CHARACTERS AMONG MEMBERS orAmphizoa SPECIES (See Text for Discussion of Character Coding). Taxa and character state distributions Character Character state davidi insolens striata lecontei 1 . Macrosculpture, elytra o a o b Not rugose or slightly rugose basally, punctures distinct, o Markedly rugose basally, punctures distinct, a Slightly rugose basally, punctures confluent, b 2. Pronotum, shape o a o o Widest at base, o Width at middle and base equal, a 3. Pronotum, sinuation of lateral margin a a o o Absent or shallow, o Deep, a 4. Pronotum, lateral margin o a* a a Not crenulate, o Slightly crenulate, a Markedly crenulate, a* 5. Prosternal intercoxal process, shape a o o o Elongate, spatulate, o Short, circular, a 6. Elytra, silhouette (dorsal aspect) o a b b Moderately broad basally, narrowed subapically, o Subovoid, slightly narrowed basally, less narrowed subapically, a Very broad basally, narrowed subapically, b 7. Elytra, silhouette (cross-sectional aspect) o o a a* Evenly convex, o Convex medially, slightly concave laterally, a Carinate, flat medially, concave laterally, a* 8. Male median lobe, shaft thickness o a a* a* Slender at middle, o Slightly thickened at middle, a Markedly thickened at middle, a* 9. Male median lobe, ventral margin o o a a Evenly arcuate, o Slightly bulged, a 10. Male median lobe, shape apex o a o o Slightly deflected ventrally, o Extended apicodorsally, a 1 1 . Male left paramere, shape o o a a Narrow basally, o Broad basally, a 12. Male parameres, vestiture o a o o Restricted to apical one-fourth, o Restricted to apical one-third, a 13. Female ovipositor, length of coxostylus ? o a a* Short, o Medium, a Long, a* 14. Female ovipositor, vestiture of coxostylus a o o Dense, evenly distributed setae, o Sparse, scattered setae, a is correct, then absence of a deep sinuation from others used for descriptive purposes only), no adults of the last two species mentioned repre- apotypic feature was found to unite all members sents yet another synapotypy for these taxa. of A. striata, although six (or seven, see above) Among the characters used in this analysis (and synapotypic features support a sister-species re- 100 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 davidi striata lecontei insolens 2a,3a,'»a«,6aI10a,12af1l»a FIGURE 23. Reconstructed phylogeny of species ofAmphizoa. Number and letter symbols placed adjacent to solid dots refer to synapotypic features presented in Table 2 and discussed in the text. lationship for A. striata and A. lecontei. Adults of A. striata are distinctly larger than most mem- bers of other amphizoid species; but it was not possible, using the out-group criterion, to affirm that this represents an apotypic feature. Zoogeography and Evolution In this section, I briefly review the present geo- graphical and habitat distributions of amphi- zoids and then discuss what can now be inferred about the zoogeographic and evolutionary his- tory of this group. Present pattern of amphizoid distribution The present pattern of geographical distribu- tion of Amphizoidae is disjunct across the north- ern Pacific Basin, with three species (Fig. 18- 20) restricted to western North America and one (Fig. 1 7) to central China. This pattern reflects a vicariance relationship, with the Palaearctic species recognized as the sister-group of the three Nearctic forms. Among North American species, the distri- bution of A. insolens (Fig. 18) is mainly coastal (i.e., east to the Sierra Nevada and Cascade Range), with range extensions east into mountain ranges of the Great Basin in Nevada, Oregon, and Idaho, and to the Northern Rocky Moun- tains of Idaho, Montana, Alberta, British Colum- bia, and Yukon Territory. The sister-group of this species includes A. lecontei, restricted to the Rocky Mountain region (Fig. 20), and A. striata, restricted to western Oregon, western and central Washington, and Vancouver Island, British Co- lumbia (Fig. 1 9). A vicariance relationship is ap- parent between A. lecontei and A. striata across the northern Great Basin and Columbia Plateau. However, because the ranges of A. striata and A. insolens overlap extensively, A. insolens and its sister-group are not strictly vicariant at present. The habitat distribution of extant amphizoids is apparently quite limited. Members of all three Nearctic species are confined to cool or cold streams. Members of A. striata are found in slow- flowing, relatively warm streams, those of A. le- contei in cooler or cold, moderate- to fast-flowing streams, and those of A. insolens in cold, fast- flowing or cascading streams. Habitat is un- known for A. davidi members; however, the type- locality of this species is in a region occupied by vegetation types that Wolfe (1979) called "no- KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 101 tophyllous broad-leaved evergreen forest" and "mixed broad-leaved evergreen and coniferous forest." Western North American vegetation types with apparently equivalent temperature re- quirements include "mixed coniferous forest," "mixed evergreen forest," and "California wood- lands" (in part) (Griffin and Critchfield 1972). These vegetation types are almost completely re- stricted to areas in California at present, and members of A. insolens are found in streams as- sociated with such forests. I suggest that the hab- itat of A. davidi members will be found to be similar to that for A. insolens members, although the former may prefer slightly warmer and slow- er-flowing streams than the latter. Mesozoic events and the origin ofamphizoids According to the hypothesis of adephagan phylogenetic relationships proposed above and illustrated in Figure 22, amphizoids are the sis- ter-group of all other Hydradephaga, except hal- iplids. If this is correct, then divergence of these sister-groups probably occurred at about the Per- mo-Triassic boundary, and certainly no later than Upper Triassic time. Although there are no am- phizoid fossil specimens known from that time, fossils representing a diverse array of other hy- dradephagan taxa document a relatively exten- sive radiation of the structurally more advanced sister-group ofamphizoids by Upper Triassic and Jurassic time. In the Triassic, the supercontinent of Pangaea was still intact (Smith, Briden, and Drewry 1977), and climate was apparently warm and equable over the entire landmass (Hallam 1981). Local climatic anomalies, associated with physiogra- phy and/or relative proximity to the ocean, may have provided some diversity of habitats, but there is no evidence for broad, latitudinally lim- ited climatic zones such as occur on continents at present. Both early and late Palaeozoic glacia- tions have been recognized (Tarling 1978), with most of these associated with high latitude po- sitions of the continents affected. No major gla- ciations appear to have occurred during all of the Mesozoic, however, probably because the con- tinents were all positioned at relatively low lat- itudes. At present, we have no information from which to infer the geographical and habitat distribu- tions ofamphizoids during early Mesozoic time. Again, these beetles are not known from the fossil record; their sister-group includes both extinct groups, presently known only from Mesozoic Asia, and extant groups with widely disjunct (e.g., Hygrobiidae) or worldwide (e.g., Dytiscidae and Gyrinidae) distributions. Because (1) there is ex- tensive sympatry at the familial level and (2) comprehensive hypotheses of phylogenetic re- lationships within families have not yet been for- mulated, it is currently impossible to recognize vicariance relationships between amphizoids and their sister-group. Hence, amphizoids could have been either widely distributed in Pangaea or geo- graphically restricted to some unknown part of that supercontinent. Structurally, extant amphizoids appear to have diverged little, if at all, from the hypothetical common ancestor of all Hydradephaga (exclud- ing haliplids). It is their sister-group, whose ex- tant descendants include hygrobiids, dytiscoids, and gyrinoids, that evolved rapidly away from the presumed ancestral form and lifestyle in adapting to a more fully aquatic existence. How then were amphizoids able to survive presumed early competition with members of their ad- vanced sister-group, whereas other lineages with- in their sister-group (e.g., liadytids and most cop- toclavids) appear to have been replaced by more highly evolved forms? Amphizoids may have persisted in geographical isolation from their sis- ter-group for an extended period. Eventually, a shift of habitat— namely to faster-flowing water- may have reduced the potential for competition with other, more rapidly diversifying Mesozoic hydradephagan groups. Even to the present, dy- tiscoids and their allies have exploited lotic hab- itats in only a limited manner, especially in geo- graphical areas where amphizoids now occur. There is no reason to suggest that amphizoids also became adapted to cool- or cold-water hab- itats so early in their history. Such habitats may have been available locally, but as noted above, climate was generally warm and equable throughout Pangaea (Hallam 198 1) at that time. Cool- or cold-water specialization would seem to have been a risky adaptive strategy at that time— one that could well have led to extinction during or before early Cenozoic time (see below). At present, there is no way to infer what (if any) effect Mesozoic plate-tectonic processes, re- sulting in fragmentation of Pangaea, may have had on the Mesozoic amphizoid fauna. Of po- tentially greater impact, however, were eustatic changes in Jurassic and Cretaceous time that re- 102 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 suited in formation of epicontinental seas in Eur- asia (the so-called "Turgai Sea," late Middle Ju- rassic through Oligocene) and North America (mid-to-late through latest Cretaceous) (Hallam 1981). Because there appear to have been con- tinental connections between eastern North America and western Europe on one hand and western North America and Asia on the other, two new land masses were formed, which Cox (1974) called "Euramerica" and "Asiamerica," respectively. Fossil evidence suggests that biotas subsequently evolved independently on each landmass (Cox 1974; Hallam 1981), resulting in increased endemism in each area by the end of the Mesozoic. The geographical range of extant amphizoids is confined to land area derivatives of Asiamerica, and it is tempting to suggest that amphizoids were at least present on that land- mass, if not also restricted to it, during Creta- ceous time. Tertiary events and amphizoid radiation As just noted, there is no evidence to suggest that the late Mesozoic distribution of amphi- zoids extended outside an area including eastern Asia and western North America (Fig. 24), al- though a more extensive distribution was cer- tainly possible. The first direct land connection between these areas occurred well before the end of the Cretaceous, as a consequence of spreading of the North Atlantic (Hallam 1981), and per- sisted continuously until late Miocene time (Hopkins 1967). Then, between 10 and 12 mil- lion years before present (mybp), a trans- Beringian seaway developed, which linked the North Pacific and Arctic basins but interrupted the exchange of terrestrial and freshwater (aquat- ic) biota between North America and Asia. A land connection was re-formed in Pliocene time and permitted renewed biotic exchange until about 3.5-4.0 mybp, at which time the trans- Beringian seaway opened again (Hopkins 1967). In Quaternary time, the Beringian land connec- tion was re-established during several, if not each, of the major glaciations, and further biotic ex- change is known to have occurred during this period (Repenning 1967). Finally, the seaway opened for the last time more than 1 1 ,000 years ago, and it has remained a substantial barrier to east-west biotic movement since that time. Palaeobotanical and other evidence indicates that early Cenozoic climates were as warm and equable as those of the Mesozoic. Then, in late Eocene time, an abrupt cooling occurred in the northern hemisphere. This cooling trend leveled off in Oligocene time; but cool conditions have persisted, with both major and minor fluctua- tions (e.g., the various Pleistocene glaciations), to the present. Another set of events that had a profound effect on climate, especially in western North America, were the episodes of erogenic and volcanic activity in Miocene and Pliocene times. This activity produced topographic relief that resulted in local and regional rain-shadow effects, increased diversity of microclimates, and increased seasonality. Geographical regions of Asia and North Amer- ica that are now occupied by extant Amphizoa species appear to have shared closely related flo- ras in early Tertiary time. These floras were of the evergreen sclerophyllous broad-leaved and mixed mesophytic forest types (Leopold and MacGinitie 1972). Floral affinities between Asia and western North America were very close in Paleocene and early Eocene time. However, by middle-to-late Eocene time, floras of the Rocky Mountain region were quite distinctive. Leopold and MacGinitie (1972) suggested that edaphic conditions associated with local volcanic activity may have stimulated selection for xeric-adapted vegetation. Although affinities between floras of southeastern Asia and the Pacific coast of North America decreased more gradually, they were nonetheless very slight indeed by late Miocene time (Wolfe and Leopold 1967). Differentiation of the North American floras appears to have been closely related to general cooling begun in late Eocene time and to middle through late Ter- tiary erogenic activity in the Pacific Northwest region. Two features that seem to characterize devel- opment of the North American floras more than contemporary floras of southeastern Asia include wholesale selective elimination of broad-leaved evergreen elements, and recruitment of subtrop- ical and temperate elements from Neotropical floras (Wolfe 1978). The first feature is no doubt related to decreasing temperatures and/or in- creased seasonality in the region; the second may simply indicate that derivative Neotropical ele- ments were already well suited to life in arid regions and could readily move into such habi- tats as they appeared and expanded. The historical factors that resulted in the vi- cariance relationship observed between A. davidi and the three Nearctic species may be the same KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 103 24 - ofc? r - FIGURE 24. Hypothetical distribution of ancestral amphizoid stock, late Cretaceous to middle Eocene time. factors that led to initial isolation of ancestral stocks and thereby permitted differentiation to proceed. These factors may include: (1) the gen- eral cooling trend that began abruptly in late Eocene time, which resulted in the elimination of subtropical and warm temperate vegetation types and their biotic associates from the Be- ringian region by late Miocene time; (2) Miocene erogenic and volcanic activity, particularly in western North America, which resulted in lati- tudinal and altitudinal climatic zonation, in- creased climatic and habitat diversity, and de- velopment of physiographic and local climatic barriers to north-south and east-west continuity of biotic distribution; and (3) opening of the trans- Beringian seaway in late Miocene time, which effectively severed faunal continuity between Asia and North America for about two million years. Any of these factors, either singly or in combi- nation, could have effected a division of the geo- graphical range of the common ancestor of extant amphizoids into Asian and North American iso- lates, and all three point to a middle-to-late Mio- cene age for the vicariance event in question. Because extant Nearctic and Palaearctic am- phizoids all appear to be cool-adapted, it is likely that their common ancestor was also cool-adapt- ed rather than that such an adaptation was ac- quired independently in the two lines. If this is correct, then it is another indication that isola- tion of respective ancestral stocks occurred after initiation of the late Eocene cooling trend, hence in Miocene time. Because there do not appear to have been any extensive areas of cool-temperate climate in the northern Pacific region prior to late Eocene or Miocene time, it is unlikely that amphizoids had specialized at an earlier time for life in a cool climate. As noted above, a vicariance relationship is not readily apparent between A. insolens and its sister-group, including A. lecontei and A. striata, due to rather extensive sympatry. However, present distribution patterns of these species are at least suggestive of an initial split of the an- cestral Nearctic stock into eastern and western vicars, the latter represented at present by A. insolens, the former by its sister-group (Fig. 25). 104 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 26 FIGURES 25, 26. Hypothetical distributions of amphizoid ancestral stocks. Figure 25. Late Miocene to early Pliocene time; ancestral stocks of A. insolens (stippled areas) and A. striata and A. lecontei (cross-hatched area). Figure 26. Middle Pliocene to end of Tertiary; ancestral stocks of A. insolens (stippled area), A. lecontei (cross-hatched area), and A. striata (obliquely hatched area). The present range of A. insolens is primarily cen- tered in and west of the Sierra Nevada and Cas- cade Range. Present populations in mountain ranges of the Great Basin, in the northern Rocky Mountains, and in Yukon Territory, could be viewed as representing more recent dispersal eastward from areas along the Pacific Coast. Development of the Cascade Range and Sierra Nevada was a gradual process (King 1977) that apparently had little effect on Pacific Northwest biota before late Miocene time. At that time, differences between floras east and west of the divide first became apparent (Wolfe 1969). Flo- ras east of the divide began to include elements adapted to drier summers and increased season- ality, while composition of the western flora con- tinued to reflect a more humid, somewhat less seasonal climate. From late Miocene time to the present, topographic relief has continued to in- crease, resulting in greater seasonality and aridity in the east, and increasingly greater differences between trans-montane climates and associated biotas. Based on proposed phylogenetic relationships among extant Nearctic amphizoid taxa and re- spective habits of their members at present, it seems likely that Nearctic amphizoids were adapted for life in cool (but not cold), slow- to only moderately fast-flowing, lowland or lower- montane streams during late Miocene time. Con- sequently, development of the extensive north- south trending Sierra-Cascade mountain system served as a barrier that effectively isolated the ancestors of A. insolens west of the divide and the common ancestors of A. lecontei and A. stria- ta east of it (Fig. 25). Based on inferred associations of amphizoids with particular early and mid-Tertiary vegeta- tion types and the known distributions of the latter and/or their descendant vegetation types during mid-Tertiary time (Leopold and Mac- Ginitie 1972; Wolfe 1969, 1978), I suggest that the common ancestor of A striata and A lecontei occupied a broad geographical range— one that extended from the eastern flank of the Sierra- Cascade divide eastward to include at least parts KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 105 of the Rocky Mountain region— during late Mio- cene and/or early Pliocene time (Fig. 25). The northern extent of this range was probably lim- ited by development of a much cooler, conti- nental climate east of the Coast Mountain system in British Columbia. Pliocene fossil assemblages from areas east of the Cascades reflect increasing aridity, probably due to the enhanced rain-shadow effect of the rising Cascade Range, and increased seasonality in the region (Wolfe 1 969). Eventually, this trend resulted in isolation of the last (relict) broad- leaved deciduous remnants of early Tertiary flo- ras on opposite sides of the Columbia Plateau and northern Great Basin (i.e., just east of the Cascades in central Oregon [Wolfe 1969] and on the western fringe of the Rocky Mountain system in central Idaho [Leopold and MacGinitie 1972]). This climatic change may have been the histor- ical event that isolated respective ancestral stocks of A. striata (in the west) and A. lecontei (in the east) (Fig. 26) and led to their divergence and, ultimately, speciation. A vicariance relationship between these taxa is still apparent at present. Quaternary history and development of the present amphizoid fauna If the sequence and timing of vicariance and speciation events suggested above is correct, then extant amphizoid diversity was achieved prior to Quaternary time (Fig. 26). Pleistocene and Recent events appear to have played a relatively minor role in the evolution of the present am- phizoid fauna. Nonetheless, available evidence suggests that important changes in geographical (Fig. 27) and habitat ranges of the Nearctic species and in structural, physiological, and behavioral characteristics of their members occurred during Quaternary time. Geologic, climatic, and biotic events of the Quaternary are relatively well known, and the reader is referred to Black, Goldthwait, and William (1973), Heusser (1960), Wright and Frey (1965), and references therein for pertinent information on the period. Amphizoa insolens LeConte. The ancestral stock of this species appears to have been isolated in the area west of the Cascade-Sierra divide in late Miocene time (Fig. 25). Subsequently, and probably in response to profound cooling (as- sociated with local and regional glaciation) and the continued rise of the Cascade-Sierra and Coastal mountain systems during early Pleisto- cene time, members of this species acquired sev- eral adaptations for life in cold, fast-flowing montane streams. Adult structural changes apparently associated with adaptation to such streams included (1) modification in pronotal and elytral shape, which actually appears to have reduced streamlining, and (2) reduction in the size and extent of fringe setae on legs. Both of these changes may have accompanied a shift in locomotory behavior among members of this species from limited use of both swimming movements and passive trans- port with stream current to almost complete reliance on the latter locomotory mode. This lo- comotory strategy, common to all extant am- phizoids, is most highly developed in A. insolens adults. Reliance on passive transport with stream cur- rent in montane areas presents amphizoids with a high risk of drifting downstream into lowland areas of warmer climate where they cannot sur- vive. To counteract downstream displacement, they may resort to either crawling back upstream on the substratum (in the water against the cur- rent, or out of water along stream banks) or flight. I have observed the former activity repeatedly, but this must result in very slow progress. Am- phizoids have very large, thick-veined hind- wings, and they appear to be capable of strong flight. The only record for amphizoid flight to date, however, is that of Darlington (1929). Finally, increased cold-tolerance is also evi- dent among A. insolens members, and this trait probably accounts for the success of this species in extending its geographical range so remarkably (Fig. 18). Eastward range expansion across the Great Basin and more northern Columbia and Central plateaus probably occurred during a ma- jor glacial (or pluvial in this area) period (Fig. 27). A general lack of evident differentiation among members of widely isolated populations over a large part of the Great Basin and western Rocky Mountain flank suggests that the present extent of range was achieved relatively recently, perhaps during the Wisconsinan. Similarly, members of populations in coastal Alaska and British Columbia are undifferentiated from those in populations to the south. Hence, occurrence of these populations in formerly glaciated areas probably represents postglacial range extension through dispersal from the south (Fig. 28). Amphizoa striata Van Dyke. Although they share several apotypic features with members of 106 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 FIGURES 27, 28. Figure 27. Hypothetical distribution of ancestral amphizoid stocks, mid-Pleistocene glacial period. Figure 28. Present distributions of Nearctic Amphizoa species. Limits of geographical distribution: A. insolens = solid line; A. striata = dotted line; A. lecontei = dashed line. A. lecontei and no doubt represent the sister- group of that species, A. striata adults are sur- prisingly similar to the hypothetical common ancestor of Nearctic amphizoids in general form, structure, and habits. Their occurrence in rela- tively warm, slow-flowing streams is unique among extant amphizoids, but such streams probably represent the ancestral (plesiotypic) habitat. The present geographical distribution of this species (Fig. 19) suggests that its members are only marginally adapted to a continental cli- mate. I have proposed that the ancestral stock of this species first became isolated and differentiated on the eastern flank of the Cascade-Sierra divide, at the western limit of the Columbia Plateau and northern Great Basin, during Pliocene time (Fig. 26). This hypothesis requires that the present distribution pattern (Fig. 1 9) resulted from sub- sequent westward range extension over, around, or through the divide in Pliocene or Quaternary time. Several present lowland routes through or around the divide (e.g., through the lower Fraser and Columbia River valleys or across the low area north and east of the Pit River in northern- most California) probably also existed through at least part of Pliocene and Pleistocene time. Populations of A. striata, members of which were marginally adapted to the regional climate of the Great Basin and Columbia Plateau, were appar- ently able to disperse westward along lowland routes and subsequently expand their range through the Willamette and Puget lowlands and into adjacent low mountains. Because potential dispersal routes were probably either filled with, or greatly restricted by, montane glaciers during major glaciations (Fig. 27), it is more likely that westward range extension coincided with some interglacial period. Nevertheless, an early post- glacial origin for the present pattern cannot be ruled out (Fig. 28). Amphizoa lecontei Matthews. Adaptation to a continental climate was probably well under way among western Rocky Mountain populations of the common ancestor of A. lecontei and A striata (Fig. 25) even before the complete isolation of eastern and western descendant stocks (Fig. 26). The present geographical distribution of A. le- KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 107 contei (Fig. 20) suggests that members of this species now require such a climatic regime for survival. At present, Amphizoa lecontei is widely dis- tributed in the Rocky Mountain region. Many extant populations, especially at the southern limits of distribution, occupy mountain ranges that are now widely separated by warm, arid low- lands. There is considerable geographical vari- ation in characters of form and structure among members of these disjunct populations, but the pattern of variation is highly discordant (see above). This suggests that the ancestral stock of this species became widely distributed through- out the central and southern Rocky Mountain regions during a major glacial period (probably the Illinoian) (Fig. 27). During a subsequent in- terglacial (e.g., the Sangamon), the formerly con- tinuous geographical range became fragmented, and isolated populations differentiated to a lim- ited degree. During one or more subsequent gla- cial periods (probably the Wisconsinan glacia- tions), ranges of previously isolated populations came in contact, and secondary intergradation occurred among several differentiated forms. Ex- tant populations achieved their present geo- graphical relationships (Fig. 20, 28), as disjunct isolates, in response to postglacial warming; the present pattern of discordance in geographical variation reflects a history of repeated episodes of isolation and dispersal among several evolving populations or groups of same. Adults of A. lecontei are similar to those of A. insolens in their physiological and behavioral ad- aptations for life in cooler, relatively faster-flow- ing streams. Perhaps the most striking features of A. lecontei adults are the broad elytral carinae. The functional significance of these carinae is yet unknown, but their dorsal position suggests that they may somehow contribute to stability during passive transport in stream currents. PROSPECTUS FOR FUTURE RESEARCH Clearly, much remains to be learned about ex- tant amphizoids and their evolutionary history. More information is needed about Amphizoa davidi— its geographical and habitat ranges, adult locomotory habits, and the form and structure of females. Because amphizoids are often diffi- cult to find, even in areas where they are known to occur, it is yet uncertain whether or not other species occur in eastern Asia. Concerted field- work in this region, carried out by individuals familiar with the habits of Nearctic amphizoids, is required to resolve this question. Comparative morphological study of amphi- zoid larvae and those of other adephagan groups should provide valuable new data that can be used in tests of hypotheses of phylogenetic re- lationship among both amphizoid species and adephagan families. This potential source of data has gone largely untapped and much basic de- scriptive work on larvae is still lacking. In order to learn more about the historical development of amphizoids in space and time, search must continue among fossil materials of Mesozoic as well as Cenozoic age. To the best of my knowledge, amphizoids are not represented anywhere in the known fossil record, even during Quaternary time. Organisms living in lotic en- vironments are much less likely to be preserved as fossils than are their lentic equivalents, and this punctuates the notion that absence from the fossil record at any particular time does not pre- clude occurrence at that time. Clearly the search for additional fossil assemblages of appropriate age must be continued. ACKNOWLEDGMENTS On different occasions, George E. Ball, Terry L. Erwin, Jean Menier, and Rob E. Roughley each tried to locate type-material for A. davidi Lucas on my behalf in MNHP and/or in other European collections, and I thank them for their efforts in this regard. Both J. Gordon Edwards and Rob Roughley provided helpful comments and insights, based on their extensive personal experience with amphizoid beetles, during var- ious phases of my study. The photograph of Am- phizoa davidi was taken by Susan Middleton, and Mary Anne Tenoria took the scanning electron micrograph. Access to the PHYLIP program package was made possible through J. Russo (Of- fice of Information Management, Smithsonian Institution). Alan Leviton provided access to the computer facility in the Department of Herpe- tology (CAS) and assisted me with use of the PHYLIP programs. LITERATURE CITED BALL, G. E. 1975. Pericaline Lebiini: notes on classification, a synopsis of the New World genera, and a revision of the genus Phloeoxena Chaudoir (Coleoptera: Carabidae). Quaest. Entomol. 11:143-242. 108 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 6 BELL, R. T. 1966. Trachypachus and the origin of the Hy- dradephaga. Coleopt. Bull. 20:107-1 12. . 1967. Coxal cavities and the classification of the Adephaga (Coleoptera). Ann. Entomol. Soc. Am. 60:101- 107. . 1982. What is Trachypachus (Coleoptera: Trachy- pachidae)? Coleopt. Bull. 36:590-596. BLACK, R. F., R. P. GOLDTHWAIT, AND H. B. WILLIAM, EDS. 1 973. The Wisconsinan Stage. Geological Society of Amer- ica, Memoir 136. ix + 334 pp. BURMEISTER, E. -G. 1976. Der Ovipositor der Hydradephaga (Coleoptera) und seine phylogenetische Bedeutung unter be- sonderer Beriicksichtigung der Dytiscidae. Zoomorphologie 85:165-257. Cox, C. B. 1974. Vertebrate palaeodistributional patterns and continental drift. J. Biogeogr. 1:75-94. CRISCI, J. V. AND T. F. STUESSY. 1980. Determining primitive character states for phylogenetic reconstruction. Syst. Bot. 5:112-135. CROWSON, R. A. 1955. The natural classification of the fam- ilies of Coleoptera. N. Lloyd Company, Ltd., London. 187 pp. . 1 98 1 . The biology of the Coleoptera. Academic Press, London, xii + 802 pp. DARLINGTON, P. J., JR. 1929. Notes on the habits of Am- phizoa. Psyche 36:383-385. EDWARDS, J. G. 1951. Amphizoidae (Coleoptera) of the World. Wasmann J. Biol. 8:303-332. . 1 954. Observations on the biology of Amphizoidae. Coleopt. Bull. 8:19-24. EKIS, G. 1977. Classification, phylogeny, and zoogeography of the genus Perilypus (Coleoptera: Cleridae). Smithson. Contr. Zool. No. 227. 138pp. ERWIN, T. L. 1979. Thoughts on the evolutionary history of ground beetles: hypotheses generated from comparative fau- nal analyses of lowland forest sites in temperate and tropical regions. Pp. 539-592 in Carabid beetles: their evolution, natural history, and classification, T. L. Erwin, G. E. Ball, D. R. Whitehead, and A. L. Halpern, eds. (Proceedings of the First International Symposium of Carabidology, Smith- sonian Institution, Washington, D.C., August 21, 23, and 25, 1976). W. Junk b.v., Publishers, The Hague, x + 635 pp. EVANS, M. E. G. 1 977. Locomotion in the Coleoptera Adeph- aga, especially Carabidae. J. Zool. (Lond.) 181:189-226. . 1982. Early evolution of the Adephaga— some lo- comotor speculations. Coleopt. Bull. 36:597-607. . In press. Hydradephagan comparative morphology and evolution: some locomotor features and their possible phylogenetic implications. Proc. Acad. Nat. Sci. Phila. 137. FOLKERTS, G. W. 1979. Spanglerogyrus albiventris, a prim- itive new genus and species of Gyrinidae (Coleoptera) from Alabama. Coleopt. Bull. 33: 1-8. FORSYTH, D. J. 1968. The structure of the defense glands in the Dytiscidae, Noteridae, Haliplidae and Gyrinidae (Co- leoptera). Trans. R. Entomol. Soc. Lond. 120:159-182. . 1970. The structure of the defense glands of the Cicindelidae, Amphizoidae, and Hygrobiidae (Insecta: Co- leoptera). J. Zool. (Lond.) 160:51-69. . 1972. The structure of pygidial defense glands of Carabidae (Coleoptera). Trans. Zool. Soc. Lond. 32:249- 309. FORSYTHE, T. C. 1981. Running and pushing in relation to hind leg structure in some Carabidae (Coleoptera). Coleopt. Bull. 35:353-378. GRIFFIN, J. R. AND W. B. CRITCHFIELD. 1972. The distri- bution of forest trees in California. U.S. For. Serv. Res. Pap. PSW-82. 114pp. HALLAM, A. 1981. Relative importance of plate movements, eustasy, and climate in controlling major biogeographical changes since the early Mesozoic. Pp. 303-330 in Vicariance biogeography: a critique, G. Nelson and D. E. Rosen, eds. Columbia University Press, New York, xvi + 593 pp. HAMMOND, P. M. 1 979. Wing-folding mechanisms of beetles, with special reference to investigations of adephagan phy- logeny (Coleoptera). Pp. 113-180 in Carabid beetles: their evolution, natural history, and classification, T. L. Erwin, G. E. Ball, D. R. Whitehead, and A. L. Halpern, eds. (Pro- ceedings of the First International Symposium of Carabi- dology, Smithsonian Institution, Washington, D.C., August 21, 23, and 25, 1976). W. Junk b.v., Publishers, The Hague, x + 635 pp. HATCH, M. H. 1953. The beetles of the Pacific Northwest. Part 1: introduction and Adephaga. University of Washing- ton Press, Seattle, vii + 340 pp. HENNIG, W. 1966. Phylogenetic systematics. University of Illinois Press, Urbana, Illinois. 263 pp. HEUSSER, C. J. 1960. Late-Pleistocene environments of North Pacific North America. American Geographical Society, New York, xxiii + 308 pp. HICKMAN, J. R. 1931. Respiration of the Haliplidae (Co- leoptera). Pap. Mich. Acad. Sci. Arts Lett. 13:277-289. HLAVAC, T. F. 1975. The prothorax of Coleoptera: (except Bostrichiformia— Cucujiformia). Bull. Mus. Comp. Zool. 147: 137-183. HOPKINS, D. M. 1967. The Cenozoic history of Beringia— a synthesis. Pp. 451-484 in The Bering land bridge, D. M. Hopkins, ed. Stanford University Press, Stanford, Califor- nia, xiii + 495 pp. HORN, G. H. 1873. Coleoptera. P. 7 1 7 in Sixth Annual Re- port of the United States Geological Survey of the Territo- ries, embracing portions of Montana, Idaho, Wyoming, and Utah; being a report of progress of the explorations for the year 1872, F. V. Hayden, ed. U.S. Government Printing Office, Washington, xi + 844 pp. . 1881. On the genera of Carabidae with special ref- erence to the fauna of Boreal America. Trans. Am. Entomol. Soc. (Phila.) 10:91-196. JEANNEL, R. 1941. Faune de France, 39. Coleopteres Cara- biques. Premiere partie. P. Lechevalier et Fils, Paris. 571 pp. KAVANAUGH, D. H. 1972. Hennig's principles and methods of phylogenetic systematics. Biologist 54:115-127. . 1978a. Hennigian phylogenetics in contemporary systematics: principles, methods, and uses. Pp. 139-150 in Beltsville symposia in agricultural research, 2. Biosystema- tics in agriculture. Allenheld, Osmun and Company, Mont- clair, New Jersey, xii + 340 pp. . 1978&. The Nearctic species of Nebria Latreille (Co- leoptera: Carabidae: Nebriini): classification, phylogeny, zoogeography, and natural history. Ph.D. Dissertation, Uni- versity of Alberta, xlvii + 1041 pp. . 1979. Studies on the Nebriini (Coleoptera: Carabi- dae), III. New Nearctic Nebria species and subspecies, no- menclatural notes, and lectotype designations. Proc. Calif. Acad. Sci. 42:87-133. KAVANAUGH: SYSTEMATICS OF AMPHIZOID BEETLES 109 . 1980. On type specimens of /I mp/»'zoaLeConte(Co- leoptera: Amphizoidae). Pan-Pac. Entomol. 56:289-292. KAVANAUGH, D. H. AND R. E. ROUGHLEY. 1981. On the identity of Amphizoa kashmirensis Vazirani (Coleoptera: Amphizoidae). Pan-Pac. Entomol. 57:269-272. KING, P. B. 1977. The evolution of North America [revised edition]. Princeton University Press, Princeton, xvi + 197 pp. LECONTE, J. L. 1853. Description of twenty new species of Coleoptera inhabiting the United States. Proc. Acad. Nat. Sci. Phila. 6:226-235. LEECH, H. B. AND H. P. CHANDLER. 1956. Aquatic Coleop- tera. Pp. 293-371 in Aquatic insects of California, R. L. Usinger, ed. University of California Press, Berkeley and Los Angeles, ix + 508 pp. LEOPOLD, E. B. AND H. D. MAcGiNiriE. 1972. Development and affinities of Tertiary floras in the Rocky Mountains. Pp. 147-200 in Floristics and paleofloristics of Asia and eastern North America, A. Graham, ed. Elsevier Publishing Com- pany, Amsterdam, xii + 272 pp. LUCAS, H. 1882. Description d'une espece nouvelle du genre Amphizoa. Bull. Soc. Entomol. Fr. (Ser. 2) 21:157-158. MADDISON, W. P., M. J. DONOGHUE, AND D. R. MADDISON. 1 984. Outgroup analysis and parsimony. Syst. Zool. 33:83- 103. MANNERHEIM, C. G. 1853. Dritter Nachtrag zur Kaefer-Fau- na der Nord-Amerikanischen Laender des Russischen Reiches. Bull. Soc. Imp. Nat. Mosc. 26:95-273. MATTHEWS, A. 1872. Descriptions of two new species of Amphizoa discovered in Vancouver's Island by Mr. Joseph Beauchamp Matthews. Cist. Entomol. 1:119-122. PONOMARENKO, A. G. 1977. Suborder Adephaga, etc. Pp. 3- 104 in Mesozoic Coleoptera [in Russian], L. V. Arnoldy, V. V. Jerikin, L. M. Nikritin, and A. G. Ponomarenko, eds. Tr. Paleontol. Inst. Akad. Nauk SSSR 161:1-204. REPENNING, C. A. 1967. Palearctic-Nearctic mammalian dis- persal in Late Cenozoic. Pp. 288-311 in The Bering land bridge, D. M. Hopkins, ed. Stanford University Press, Stan- ford, California, xiii + 495 pp. Ross, H. H. 1974. Biological systematics. Addison- Wesley Publishing Company, Inc., Reading, Massachusetts. 345 pp. ROUGHLEY, R. E. 1981. Trachypachidae and Hydradephaga (Coleoptera): a monophyletic unit? Pan-Pac. Entomol. 57: 273-285. SALLE, M. A. 1874. [Remarques synonymiques sur une es- pece de Coldopteres]. Bull. Soc. Entomol. Fr. (Ser. 4) 14: 222. SMITH, A. G., J. C. BRIDEN, AND G. E. DREWRY. 1977. Pha- nerozoic world maps. Pp. 1-39 in Organisms and continents through time, N. F. Hughes, ed. Palaeontological Society, London, vi + 334 pp. STEINER, W. E., JR. AND J. J. ANDERSON. 1981. Notes on the natural history of Spanglerogyrus albiventris Folkerts, with a new distributional record (Coleoptera: Gyrinidae). Pan- Pac. Entomol. 57:124-132. TARLING, D. H. 1978. The geological-geophysical framework of ice ages. Pp. 3-24 in Climatic change, J. Gribben, ed. Cambridge University Press, Cambridge, xi + 280 pp. VAN DYKE, E. C. 1927a. A new species of Amphizoa (Co- leoptera). Pan-Pac. Entomol. 3:97-98. . 1927 b. The species of A mphizoa (Coleoptera). Pan- Pac. Entomol. 3:197-198. VAZIRANI, T. G. 1964. On a new species of aquatic beetle of the genus Amphizoa LeConte, 1853 [Insecta: Coleoptera: Amphizoidae] from Kashmir, India. Proc. Zool. Soc. (Cal- cutta) 17:145-147. WATROUS, L. E. AND Q. D. WHEELER. 1981. The out-group comparison method of character analysis. Syst. Zool. 30:1- 11. WOLFE, J. A. 1969. Neogene floristic and vegetational history of the Pacific Northwest. Madrono 20:83-1 10. . 1978. A paleobotanical interpretation of Tertiary climates in the Northern Hemisphere. Am. Sci. 66:694-703. 1979. Temperature parameters of humid to mesic forests of eastern Asia and relation to forests of other regions of the Northern Hemisphere and Australasia. U.S. Geol. Surv. Prof. Pap. 1 106. iii + 37 pp. WOLFE, J. A. AND E. B. LEOPOLD. 1967. Neogene and Early Quaternary vegetation of northwestern North America and northeastern Asia. Pp. 193-206 in The Bering land bridge, D. M. Hopkins, ed. Stanford University Press, Stanford, xiii + 495 pp. WRIGHT, H. E. AND D. G. FREY, EDS. 1965. The Quaternary of the United States. Princeton University Press, Princeton, x + 922 pp. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 7, pp. 111-126, 14 figs., 1 table. February 7, 1986 ACANTHOGILIA, A NEW GENUS OF POLEMONIACEAE FROM BAJA CALIFORNIA, MEXICO By Alva G. Day and Reid Moran Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 Published by California Academy of Sciences San Francisco PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 7, pp. 111-126, 14 figs., 1 table. February 7, 1986 ACANTHOGILIA, A NEW GENUS OF POLEMONIACEAE FROM BAJA CALIFORNIA, MEXICO By Alva G. Day and Reid Moran Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: We propose the genus Acanthogilia for the spiny desert shrub first named Gilia gloriosa Bran- degee. The gametic chromosome number is nine, also the basic number for nine other genera and for the family. Acanthogilia is unique among Polemoniaceae in its extreme leaf dimorphism, its persistent woody-spinose primary leaves, and its coarsely verrucate zonocolporate pollen grains. Though formerly placed in Gilia, Ipomopsis, Leptodactylon, and Loeselia, it differs further from all these genera in its persistent secondary leaf bases with deciduous blades, its numerous closely spaced corolla veins connected at several levels, and its winged seeds. It does share several unusual characters, such as the superficially adnate filaments, with species of Gilia sect. Giliastrum. Acanthogilia seems closest to the Andean genus Cantua. Cantua, like Acanthogilia, is shrubby, with leaves dimorphic, on long shoots and axillary short shoots, with persistent leaf bases, with corolla veins connected at several levels, with winged seeds, with superficially adnate filaments (in some species), and with coarsely verrucate pollen grains (in one other species). Cantua differs in having the leaves broad and herbaceous, only weakly dimorphic, and neither woody-spinose nor with deciduous linear blades, the calyx entirely herbaceous, the pollen pantoporate, and the chromosome number hexaploid. INTRODUCTION scribed genera of the family and has some unique „.,. , r-^c.™ /,00™ • characters. We therefore propose for it the fol- Gilia gloriosa, of T. S. Brandegee (1889), is a \. I ft. lowing new monotypic genus, spiny but truly glonose desert shrub of rather local occurrence on the Pacific drainage of north- 0 _ ,„.„,.- • M-.- ^ ^ TI.- 1 * • SYSTEMATIC TREATMENT central Baja California (Fig. 1-3). This plant is seldom seen and little known, and its best generic Acanthogilia Day et Moran, genus novum position has remained uncertain. Brand (1907) mexicanum Polemoniacearum, ob folia valdedi- placed it in Gilia sect. Leptodactylon, and Wher- morpha, primariis rigide spinosis persistentibus, ry (1945) called it Leptodactylon gloriosum. granaque pollinis zonocolporata supraverrucata Johnston (1924) informally listed it as Loeselia bene distinction; Cantuae Juss. fortasse proxi- gloriosa. Current floras (Wiggins 1964, 1980) treat mum, quae autem calyce toto herbaceo, aetate it as Ipomopsis gloriosa, following Alva Grant non rumpenti, pollinis grants pantoporatis, chro- (in V. Grant 1956). mosomatumque numero polyploideo differt. Si New information on the chromosome num- vis descriptionem latine recipere, involucrum ber, pollen grain type, and some other aspects praeinscriptum praesolutumque mitte. shows that Gilia gloriosa differs from all de- Stiff spiny shrub with dimorphic leaves, the [ill] 112 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 7 DAY AND MORAN: ACANTHOGILIA. NEW GENUS OF POLEMONIACEAE 113 II 11 FIGURE 2. Acanthogilia gloriosa in flower and fruit. A. Flowering and fruiting branches on an older branch, with dehisced capsules from previous season; B. Green, submature seeds, contents of a single locule; C. Dry seeds from dehisced capsule; D. Calyx with mature, undehisced capsule; E. Calyx with dehisced capsule from previous season; F. Segment of branch with spinose primary leaves and fascicled herbaceous secondary leaves. FIGURE 1. 1976. Inflorescence of Acanthogilia gloriosa (Brandg.) Day and Moran, El Colosal, Baja California, Mexico, 1 3 June 114 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 7 El Rosario San Fernando Luis Gonzaga Santa Catarina Punta Canoas Punta Prieta •San Andres •Rosarito iffMillers 3$ Landing FIGURE 3. North-central Baja California, Mexico, showing distribution of Acanthogilia gloriosa. + of collection yielding chromosome count, A = stated type locality. : collection site, * = site primary alternate, woody-persistent, pinnate, with terete spinose divisions, the secondary fas- cicled on axillary short shoots, with persistent bases and deciduous, flat, linear, herbaceous blades. Calyx tubular, with equal spinose lobes and narrower scarious intervals that rupture in fruit. Corolla regular, salverform. Stamens sub- equally attached near middle of tube, superfi- cially adnate below, subequal, well exserted. Pol- len yellow, the grains zonocolporate, perreticulate, supraverrucate. Seeds elongate, flat, winged, mu- cilaginous when wet. Chromosomes: x = 9. DAY AND MORAN: ACANTHOGILIA, NEW GENUS OF POLEMONIACEAE 115 TYPE-SPECIES.— Gilia gloriosa Brandegee. Acanthogilia gloriosa (Brandegee) Day and Moran, comb. nov. Gilia gloriosa Brandegee, Proc. Calif. Acad. Sci. Ser. 2, 2:184, pi. 9. 1889. Loeselia gloriosa I. M. Johnston, Proc. Calif. Acad. Sci. Ser. 4, 12:962. 1924. Leptodactylon gloriosum Wherry, Am. Midi. Naturalist 34:383. 1945. Ipomopsis glo- riosa A. Grant in V. Grant, Aliso 3:357. 1956. TYPE. -Mexico, Baja California, Ubi, 8 May 1 889, T. S. Bran- degee s.n. (holotype, UC 101896!; isotypes, DS!, GH). The type locality, Ubi, is the tinaja, or waterhole, of Yubay, near 29°1 1'N, 1 13°59'W, elevation ca. 650 m, ca. 9 km NE of the abandoned mine of Desengano and ca. 52 km from the Pacific coast. Brandegee remarked that the plant appeared to be very local, having been observed only during an hour's journey and not again met with. He customarily gave bare locality names without direction or distance, and presumably he may have meant within half a day or so north or south of Yubay. His itinerary (Moran 1952) suggests south. Moran failed to find the plant about Yubay or along the old trail just to the south. Stiff spiny shrub 1-3 m high and 1-5 m wide, much branched at base, the young parts glan- dular-pubescent and glutinous with two-many- celled trichomes mostly less than 0.5 mm long, each tipped with yellowish globule. Trunks to 6 cm thick, the bark light to dark gray, flaking in small plates; lower branches arching, sometimes rooting. Branching sympodial, the branches mostly flowering terminally the first year and so not elongating further, 1-125 cm long, 1-4 mm thick the first year, tan becoming gray, subterete, persistently spiny with old leaves; internodes av- eraging 5-8 mm, exceeded by leaves. Primary leaves subopposite to mostly alternate, 1-3 cm long, woody-spinose, rigidly divaricate and straight except terminal segment usually de- clined, green becoming tan and finally dark gray, persisting two to three (sometimes to six) years but weak after first or second year, the lowermost sometimes simple but most pinnate with nar- rowly linear rachis and one to two (sometimes to three) pairs of spreading spinose lobes to 9 mm long, the base 1.5-3.0 mm wide. Secondary leaves fascicled in axils, few and short the first season (on flowering branches), later to 25 per season, the bases whitish to tan, persistent, the blades herbaceous, mostly simple, rarely with one to two short lobes, linear-oblanceolate, spi- nose-tipped, flattened, 5-20 mm long, to 1 mm wide, rather sparsely glandular, deciduous throughout plant all at about one time; first leaves of new shoot with enlarged semiglobular bases to 1.5 mm wide and blades sometimes less than 1 mm long. Short shoots producing leaves for three to four (sometimes to six) years, to 8 mm long or some becoming long shoots. Inflores- cence densely glandular-puberulent and gluti- nous, a thyrse to 2 dm long on a new shoot, with terminal flower mostly opening first and with up to 30 short one-few-flowered branches below; or inflorescence reduced to short one-few-flowered shoot, though sometimes several such shoots borne on one older branch to form two-genera- tional inflorescence. Flowers January to July, Oc- tober, protandrous, open ca. four to five days and nights, odorless, visited by hummingbirds. Pedicels erect to spreading, 1-6 mm long, 0.5- 1.0 mm thick. Calyx 10-16 mm long, 3.0-4.5 mm wide, cylindric, tapering to rounded at base, densely glandular and glutinous without, more sparsely so within, tubular in lower %; segments equal, erect or slightly outcurved, 3-8 mm long, triangular-lanceolate, pungent-acuminate, car- tilaginous, with many crowded veins within, scarious-margined except near apex; sinuses V-shaped, the scarious intervals much narrower than ribs, distended at anthesis, folding inward as segments later converge, mostly rupturing in fruit. Corolla salverform, 3.0-4.5 cm long, glan- dular without, in bud pale yellow becoming or- angish; tube stout, slightly upcurved, 2.0-3.2 cm long, 2.5-3.5 mm wide below, gradually flaring to orifice 5-7 mm wide, dull orange-red to or- ange-brown, becoming paler and more purplish; throat yellow; limb 2-4 cm wide, white and mur- iculopapillose inside, rose-veined outside, the lobes in bud convolute, in anthesis widespread- ing or somewhat reflexed, in age strongly re- flexed, 8-20 mm long, 4-12 mm wide, obliquely oval to strap-shaped, obtuse to slightly emargin- ate, sometimes apiculate, with 30-50 close-spaced parallel veins per lobe. Filaments glabrous, 1 5- 27 mm long, subequally attached at middle of tube or slightly above, superficially adnate below, with margins free throughout, subequally exsert- ed 4-13 mm from throat; anthers oblong, sag- ittate, 4-5 mm long before anthesis, dehiscing as corolla begins to open. Pollen grains suboblate to spheroidal (P 55-64 ^m, E 6 1-7 1 /xm). Nectary disk green, ca. 2 mm wide, shallowly cupped, the margin regularly undulate to form erect lobules opposite calyx segments and spreading ones be- tween. Ovary three-celled, 4-5 mm long, ca. 1.5 mm thick; style 20-40 mm long, slightly shorter to slightly longer than filaments (consistent in each plant); stigma lobes acute, 1 .5-2.0 mm long, 116 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 7 outcurved by third day. Ovules 14-24 per cell, ovoid, ca. 0.6 mm long, many aborting. Capsules 7-15 mm long, 4-5 mm thick, cylindric, beaked, crustaceous, yellowish brown, loculicidally de- hiscent, the valves often recurving. Seeds over- lapping, one to six per cell, narrowly oblong, flat, 6-7 mm long, 1.5 mm wide, the brownish body surrounded by elongated membranous wing, the body and wing swelling and mucilaginous when wet. Chromosomes: n = 9. DISTRIBUTION. — Mexico, Baja California Norte: locally common on desert flats and hill- sides and in arroyo beds from 40 km SSE of El Rosario to Miller's Landing, a span of 200 km, and from the coast inland at least 45 km and to 525 m elevation (Fig. 3). ADDITIONAL SPECIMENS EXAMINED.— San Fernando, 24 May 1894, Anthony s.n. (UC); Rosalia Bay, Jul.-Oct. 1896, Anthony 92 (DS, UC); 2 km NW of Cajiloa, 29°41'N, 1 15°34'/2'W, Aug. 1980, Binneys.n. (SD); 2.9 m E of mouth of Arroyo San Jos6, 29°12'N, 1 14°44'W, 28 Jun. 1969, Bostic s.n. (SD); 6.2 m S of Santa Catarina, 29°38'N, 115°10'W, 26 Aug. 1969, Bostic s.n. (SD); San Andreas Canyon above Santa Rosalillita, 20 Mar. 1 984, Breedlove 60808 (CAS); 2-3 km NE of Santa Rosalillita, 20 Mar. 1984, Breedlove 60834 (CAS); 5-15 m N of Puerto Santa Catarina, road to San Agustin, 1 Mar. 1985, Breedlove 62269(CAS); 1 mNE of Rancho Santa Catarina, 20 Jun. 1979, Clark 3167 (CAS); 10 m S of Punta Prieta, 9 Feb. 1947, Con- stance 3 1 25 (DS); 2 km W of La Ramona, 29°49'N, 1 1 5°07'W, 10 Jul. 1976, Day and Moran 76-126 (CAS, SD); same data, Day and Moran 76-129 (CAS, SD); 4.3 m S of El Colosal, 29°47'N, 115°06'W, 10 Jul. 1976, Day and Moran 76-133 (CAS); Sierra Lino, 25 m S of Punta Prieta, 6 Mar. 1 947, Gentry 7345 (DS, SD, UC); S of Arroyo San Borja, 26 Mar. 1947, Gentry 7617 (DS, UC); San Andreas, 26 Jul. 1941, Harbison s.n. (SD); Arroyo San Jose, 29°10'N, 1 14°45'W, 18 Oct. 1966, Hastings and Turner 66-154 (DS, SD); Rancho La Ramona, Santa Catarina, 21 Jun. 1947, Hueys.n. (SD); 3 m S of Miller's Landing, 9 Jul. 1937, Lindsay s.n. (DS); Arroyo Santo Do- minguito, 6.7 m S of San Andres, 28°42'N, 1 14°15'W, 28 May 1959, Moran 7498 (DS, SD, UC); 1 '/2 m N of Rancho Ramona, 29°50'N, 115°05'W, 25 Mar. 1970, Moran 16896 (SD); 11 m N of Puerto Santa Catarina, 29°39'N, 1 1 5°1 2' W, 28 Mar. 1 970, Moran 17030 (SD); 3 m SE of Santa Rosalillita, 28°40'N, 1 14°13'W, 2 Jan. 1976 and ex hort. San Diego, 16 Jul. 1976, Moran 22779, (CAS, SD); 10 km S of El Aguila, 290523/VN, 1 15°04'/4'W, 12 Jun. 1976, Moran 23518 (SD); 2 km W of La Ramona, 29°49'N, 115°07'W, 12 Jun. 1976, Moran 23519 (SD); 2 km W of La Luciana Mine, 29°42'N, 115°02'W, 13 Jun. 1976, Moran 23521 (CAS, SD); 3 km NW of Santa Ca- tarina, 29°44'/2'N, HSW/z'W, 13 Jun. 1976, Moran 23522 (CAS, SD); coastal region near Rosarito, 28°38'N, 1 14°05'W, 5 Oct. 1970, Rauh 25416 (HEID, SD); 1 m NW of Santa Catarina, 29°44'N, 115°06'W, Robinson s.n. (SD); 23 m S of Punta Prieta, 1 Jun. 1931, Wiggins 5731 (DS UC); 17 m S of Punta Prieta, 9 Apr. 1961, Wiggins 16193, (DS). At Rancho Santa Catarina this plant was called "mala mujer" (bad woman). That name is used in Baja California and elsewhere in Mexico for some other prickly plants, as well as for several stinging and poisonous plants (Martinez 1937). FLORAL BIOLOGY.— Floral characters ofAcan- thogilia predominantly suggest outcrossing. The anthers dehisce as the flower opens, but stigmas do not open out until the third day. Styles usually exceed stamens, as in many Polemoniaceae that are insect- or hummingbird-pollinated. In some individuals, however, styles are consistently shorter, with stigmas opening just beneath the anthers, as in various autogamous flowers. This heteromorphism in the population may help en- sure some seed production even if outcrossing fails. The floral characters strongly suggest ad- aptation to hummingbird pollination, and Mor- an has observed hummingbirds visiting the flow- ers. The flowers are open by day, odorless, with stamens and style well exserted. The corolla is robust, with a long and ample tube. Its color pattern is well marked, with glistening white lobes around a yellow orifice, and with an orange-red to orange-brown tube. Other hummingbird flow- ers in the family have similar characters. In Can- tua, Gilia, Ipomopsis, Loeselia, and Polemon- ium, hummingbird flowers are diurnal and odorless, with long red or yellow corolla tubes and usually with exserted stamens and style (Grant and Grant 1965). Superficially, some hummingbird flowers of different genera look more like each other than like bee- or fly-polli- nated flowers of their own genus. COMPARISONS WITH OTHER POLEMONIACEAE CHROMOSOME NUMBER.— The basic chromo- some number of Acanthogilia is nine. We base it on counts from propionic-carmine squashes of anthers from three collections of A. gloriosa (mapped in Fig. 3): (1) Moran 17030 from 18 km N of Puerto Santa Catarina; (2) Moran 22779, ex hort. San Diego, from 4.8 km SE of Santa Rosalillita; and (3) Moran 23519, from 2 km W of La Ramona. Meiotic counts from collection 1 showed n = 9; chromosome behavior was reg- ular, with 9n at Mj (Fig. 4). From premeiotic sporogenous cells of collection 2 (Fig. 4) and from tapetal cells of collection 3, mitotic counts showed 2n= 18. Nine is the basic number for more than half the genera of the family (Table 1) and is regarded as the primitive basic number in the Polemo- niaceae. The other genera have lower basic num- bers apparently derived independently in differ- ent tribes by aneuploid reduction (Grant 1959). DAY AND MORAN: ACANTHOGILIA, NEW GENUS OF POLEMONIACEAE 117 Thus, Acanthogilia stands among the ten genera that still have the primitive number. These are a mixed lot, from all five tribes in the classifi- cation of Grant (1959). Hence chromosome number gives no clue to the tribal placement of Acanthogilia. The basic number is mostly constant within genera of the family (Table 1); where it is not (Allophyllum, Gilia), the number varies by only one chromosome pair (x = 9, 8). Gilia gloriosa was placed in Ipomopsis (A. Grant in V. Grant 1956) before its chromosome number was known. Since Ipomopsis has x = 7, a count of n = 9 for /. gloriosa showed us it was in the wrong genus. That was the starting point for this study. POLLEN.— Pollen studies in the Polemoniaceae (Erdtman 1952; Stuchlik 1 967; Taylor and Levin 1975) have not included Acanthogilia. We sent a pollen sample to Dr. Leon Stuchlik, who kindly prepared the following diagnosis, in 1980, with permission to include it here (see Fig. 5-8). 5 um 10 pm FIGURE 4. Chromosomes of Acanthogilia gloriosa. Left, mitosis; right, meiosis. Traced from micrographs. Pollen grains 5-6 colporate (zonocolporate), suboblate to spheroidal; diameter 55-64 /^m x 61-71 ^m. Colpi short, only slightly longer than pores are broad. Pores lalongate to circular; diameter 5-7 ^m x 7-10 ^m. Exine 2.4-2.9 nm thick; nexine 0.8-1.2 nm thick, thickened up to 1 .7 jim in pore area, finely perreticulate. Lumina vari- able in shape and size; diameter less than 0.5 urn to 1 Mm; muri supported by simple bacula densely spaced, TABLE 1 . COMPARISON OF THE GENERA OF POLEMONIACEAE. 1.9,8 = intrageneric aneuploidy; 9/8 arid 8/7 = dibasic polyploidy. II. Pollen groups 1-4 are the alliances of Taylor and Levin (1975), with Acanthogilia added. III. Pin = leaves pinnately veined, dissected, or lobed; PinC = leaves pinnately compound; Palm = leaves palmately lobed; * = true foliage leaves lacking. IV. N = seeds not winged; NW = seeds very narrowly winged; W = seeds broadly winged. V. A = filaments superficially adnate; M = filaments merged with corolla; I = filaments intermediate: merged below. VI. M = calyx membranous below sinuses; H = calyx herbaceous throughout. VII. A = veins connected at base of lobe and in upper lobe; B = veins connected only at base; C = veins connected only well above base; D = veins free; * = venation too simplified to classify. I Basic chromosome number II Pollen group III Leaf form IV Seed type V Filaments VI Calyx type VII Venation of corolla lobe Acanthogilia 9 1 Pin W A M A Cantua 9 1 Pin W A, M H A, B (In = 27n) Huthia 7 1 Pin W I H B Cobaea 9/8 2 PinC W A H A (In = 26n) Phlox 7 2 Pin N M M A,B Microsteris 7 2 Pin N M M * Gymnosteris 6 2 * N M H * Polemonium 9 3 PinC N M,I H A, B Bonplandia 8/7? 3 Pin N, NW M H A (2n= 15n)t Gilia 9,8 4 Pin N A, M, I M A, B, D Collomia 8 4 Pin N M H B,C Eriastrum 7 4 Pin N M, I M B Navarretia 9 4 Pin N M M B,* Ipomopsis 7 4 Pin N M, I M B, D Langloisia 7 4 Pin N M, I M B,C Allophyllum 9,8 4 Pin N M M C Loeselia 9 4 Pin N, NW M M C,D Leptodactylon 9 4 Palm N M M D Linanthus 9 4 Palm N M M D t Bonplandia geminiflora chromosome number, 2n = 1 5 bivalents, determined from the following collections: Sinaloa, Mexico, D. E. Breedlove 44637; Chiapas, Mexico, D. E. Breedlove 56256. Vouchers deposited at CAS. Counted by A. Day. 118 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 7 sometimes merged 2-3 together. Diameter of bacula ca. 0.5 nm. Reticulum supra verrucate. Verrucae on surface of exine very variable in shape and size, from very small and flat with diameter ca. 1 tim to circular or oval with diameter to 1 0 nm, in young pollen grains very densely spaced. Surface of verrucae ultra-finely striate or rugu- late, as seen with scanning electron microscope at 7000 x magnification. This diagnosis is based on a single collection (DayandMoran 76133, CAS). Other collections show that the grains may be 7-8-colporate (Mo- ran 7498, CAS) or the colpi may be longer (Fig. 6; Wiggins 5731, DS). Variation in the number and distribution of verrucae is seen by comparing Figures 5 and 6. The most distinctive feature of Acanthogilia pollen grains is the coarsely verrucate exine (Fig. 5-8). Among other Polemoniaceae with zono- colporate grains, only Eriastrum and Gilia sect. Giliastrum have the exine verrucate, but there the verrucae are minute. Dr. Stuchlik (pers. comm. 1980) remarked that Acanthogilia has probably a new pollen type for the family. Pollen grains with large verrucae do occur, however, in Cantua. In C. buxifolia Juss. ex Lam. (Fig. 9, 10) the exine appears much as in Acan- thogilia. In both C. buxifolia and Acanthogilia the verrucae are diverse in size and shape, the larger ones supported by groups of bacula. Viewed with SEM (Fig. 9, 10), the verrucae of C. bux- ifolia differ from those of Acanthogilia only in being somewhat broader and flatter. In the Cantueae (Cantua and Huthid) the exine is semitectate and, as illustrated (SEM) by Taylor and Levin (1975), generally consists of large, closely spaced areoles (Taylor and Levin's term) or insulae (Stuchlik 1967). However, Cantua buxifolia is exceptional in having areoles of such small diameter that they have been described as large verrucae (Erdtman 1952). This exine pat- tern may have evolved through reduction of larg- er areoles. Despite the similarity in exine, the pollen grains of Cantua buxifolia differ from those of Acan- thogilia in aperture type; for, as in other Can- tueae, they are pantoporate, not zonocolporate. Since, however, both zonocolporate and panto- porate grains can occur within a single genus else- where (Collomia, Loeblich 1964, Chuang et al. 1978; Gilia, Stuchlik 1967), this difference be- tween Acanthogilia and Cantua is not necessarily fundamental. In view of other notable shared characters (Table 1), we interpret the similarity in exine as a mark of relationship. On the basis of pollen morphology, Taylor and Levin (1975: fig. 1) grouped the genera of Pole- moniaceae into four unnamed alliances (pollen groups 1-4 of our Table 1). One alliance included only Cantua and Huthia, but we would add Acanthogilia. LEAVES.— In Acanthogilia the leaves of long shoots and axillary short shoots are markedly different, with no intermediates (Fig. 2F). The leaves of long shoots are woody-spinose and per- sistent, as in no other Polemoniaceae. Base and blade are scarcely delimited, and the blade is pinnately divided, with terete rachis and lobes. On the contrary, the fascicled axillary leaves are each clearly divided by a constriction into a per- sistent base and a deciduous blade (Fig. 11 A). The broadened bases remain indefinitely in a compact spiral on the short shoot, but the blades fall at one time throughout the plant with drying of the season. These are smaller blades than those of the primary leaves, mostly simple, linear but flattened, herbaceous, and greener. Cantua, Huthia, Leptodactylon, Loeselia, and some species of Ipomopsis also have leaves on long shoots and in axillary fascicles; but although the fascicled leaves may be smaller, all leaves are nearly alike. Cantua is somewhat exceptional: the primary leaves are large and more or less lobed and fall early, whereas the secondary, ax- illary leaves are more persistent and in most species are smaller and have entire margins (In- fantes Vera 1962; Gibson 1967). In C. buxifolia grown in San Francisco, we note that, except for young shoots, leafy stems bear only the smaller secondary leaves. In various other Polemonia- ceae, especially annuals, leaves are gradually dif- ferent from base to apex, grading into bracts above. Only Acanthogilia, however, has mark- edly dimorphic leaves. In most perennial Polemoniaceae the leaves wither persistent, though they may finally erode away. In several evergreen shrubs (Cantua, Hu- thia, Loeselia mexicana (Lam.) Brand, L. pur- pusii Brandegee), however, leaf blades finally fall, leaving the persistent bases conspicuous (Fig. 1 1 B). In Cantua buxifolia and C. pyrifolia Juss. ex Lam., both of which produce fascicled leaves, the short shoots and crowded leaf bases, with blades gone, somewhat resemble those of Acan- thogilia (Fig. 1 1A, B). DAY AND MORAN: ACANTHOGILIA, NEW GENUS OF POLEMONIACEAE 119 5|xm FIGURES 5-10. Pollen grains of Acanthogilia gloriosa and Cantua buxifolia Juss. ex Lam.; Fig. 5, 6. (light microscope) Acanthogilia; Fig. 5. Day and Moran 76-133 (CAS); Fig. 6. Wiggins 5731 (DS); Fig. 7, 8. (SEM) Acanthogilia, Wiggins 5731 (DS); Fig. 9, 10. (SEM) Cantua buxifolia, cultivated, McClintock s.n., 15 Mar 1976 (CAS). 120 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 7 4 mm i m 111 mi I mm B 2 mm 4 mm FIGURE 1 1 . Short shoots and calyces. A. Short shoot of Acanthogilia gloriosa with persistent leaf bases after most blades have fallen, the leaf bases from previous seasons compacted below, primary leaf mostly eroded away; B. Short shoot of Cantua buxifolia with crowded leaf bases, growing out into long shoot above; C. Calyx of Acanthogilia gloriosa at anthesis; D. Part of calyx, ventral side, showing venation in rib; E. Calyx of Cantua buxifolia at anthesis. SEEDS.— The seed of Acanthogilia is flat and is bordered by a membranous wing 1-3 mm wide (Fig. 2C). In most Polemoniaceae, seeds are wingless, though in Bonplandia and Loeselia they are sometimes very narrowly winged. Only in Cantua, Cobaea, and Huthia are the seeds like- wise flat and broadly winged. In these genera, however, both seeds and wings are considerably DAY AND MORAN: ACANTHOGILIA, NEW GENUS OF POLEMONIACEAE 121 wider; and the wings, though thin, are opaque rather than membranous. Acanthogilia has only 1-6 seeds per locule, whereas Cantua, Cobaea, and Huthia have many, in platelike layers. However, the ovary of Acan- thogilia has 1 4-24 ovules per locule, suggesting that the ancestral capsule may have had many more seeds. A hint of layering in the capsules suggests that if more seeds were present they might form layers as in the other genera. STAMENS.— The filaments of Acanthogilia are attached to the corolla tube about midway but are only superficially adnate below; they are well formed, with free margins, and are distinguish- able to the base. Likewise in Cobaea, Cantua (C. candelilla Brand and C. quercifolia Juss. but not C. buxifolia), and Gilia sect. Giliastrum (G. ri- gidula Benth. and G. ripleyi Barneby but not G. insignis (Brand) Cory and Parks or G. incisa Benth.) the filaments are superficially adnate. This appears to be a rare and primitive condition in the family. In our sampling of other genera, the filaments are so merged with the corolla, at least basally and commonly to the point of insertion, that they are not distinguishable. CALYX.— The calyx of Acanthogilia is narrow- ly membranous and veinless below the sinuses and has many veins crowded in the herbaceous ribs (Fig. 1 1C, D); it ruptures between ribs as the capsule grows. In eleven other genera (Table 1), including Gilia and most allies as well as Phlox and Microsteris, similarly, the calyx is narrowly to broadly membranous below the sinuses, with veins again confined to the ribs; it may or may not rupture in fruit. In all examples seen, veins are fewer and less crowded than in Acanthogilia. In all these genera, including Acanthogilia, lat- eral veins of adjacent ribs are connected only near the base of the calyx. On the other hand, in Bonplandia, Cantua (Fig. HE), Cobaea, Collomia, Gymnosteris, Huthia, and Polemonium the calyx is not alternately ribbed and membranous but is herbaceous or somewhat chartaceous throughout, and it en- larges without rupturing as the capsule grows. Venation is various but generally is spread out more than in the genera with membranous calyx. Lateral veins of adjacent sectors may be con- nected just below the sinuses (Fig. 1 IE) or much lower. They are connected in Cantua quercifolia near the base of the calyx but in C. buxifolia at various levels, even in the same calyx (Fig. 1 1 E). The herbaceous calyx type, found also in re- lated families, presumably is primitive in the Polemoniaceae, the membranous calyx perhaps arising independently in more than one line in arid habitats. A division of the family by calyx types then would separate some related genera. Thus Collomia (herbaceous calyx) belongs with the Gilia group (otherwise membranous), and Phlox and Microsteris (membranous calyx) seem related to genera with herbaceous calyx (Table 1). Similarly, Acanthogilia appears related to Cantua despite the difference in calyx (Table 1). COROLLA VENATION.— Surveying corolla ve- nation in the family, Day has found patterns to link Acanthogilia with some genera and to sep- arate it from others. Generally in the family, each sector of the corolla has a median vein and two laterals more or less parallel in the tube, with branches in the lobes and commonly with con- nections; positions of vein connections are char- acteristic for many taxa. Figures 12-14 show ex- amples from 1 2 out of 1 9 genera— all traced from photographs of dissected corollas stained with safranin. The staminal veins, alternating with the corolla veins in the tube, are omitted. Venation patterns fall mainly into four types (identified in Table 1 by letters A-D): A. Cantua type— veins connected near corolla orifice, curv- ing and connected once or twice above in the lobe (Fig. 12G-H, 13K-O);B. Gilia type -veins connected near orifice but straighter above and without other connections (Fig. 1 2B-D); C. Loe- selia type— veins often connected near middle or apex of lobe but not near orifice (Fig. 1 4P-R); and D. Leptodactylon type— veins free, not con- nected in orifice or lobes, even in large corollas with many veins; often each sector with only a single vein in basal half of tube (Fig. 1 2E-F, 14S-U). Some genera show only one venation type, some two, and one three (Table 1). In very small corollas (Microsteris, Fig. 1 3J; many Na- varretia species, Fig. 1 2A; Gymnosteris; and oc- casional species in other genera) venation may be so simplified that it tells little of relationship. In Acanthogilia the corolla veins are connected at several levels in the lobes (Fig. 12H), much as in Cantua, Cobaea, and Phlox (Fig. 1 3). Acan- thogilia differs from them in having more closely spaced veins that are nearly parallel and less curved. Its pattern is perhaps most closely ap- 122 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 7 B bar = I mm FIGURE 1 2. Corolla venation patterns in Polemoniaceae, each showing a sector from base of tube to apex of lobe. Dashed lines show where corolla tube cut. Stamens and staminal veins not shown. A. Navarretia fossalis Moran; B. N. mitracarpa Greene; C. Gilia tricolor Benth.; D. G. leptomeria Gray; E. G. incisa Benth.; F. G. rigidula Benth.; G. G. ripleyi Barneby; H', H". Acanthogilia gloriosa Day and Moran. preached in Cantua candelilla (Fig. 13M). On the other hand, despite more numerous veins with connections at several levels, the pattern of Acanthogilia resembles that of Gilia and allies (Fig. 1 2B-D) in its straighter and closer-spaced veins. And although most species of Gilia sect. Giliastrum have free veins (Fig. 12E-F), the anomalous G. ripleyi (Pig. 12G) has connections DAY AND MORAN: ACANTHOGILIA, NEW GENUS OF POLEMONIACEAE 123 bar = i mm FIGURE 13. Corolla venation patterns, cont. J. Microsteris gracilis (Hook.) Greene; K. Cobaea biaurita Standl.; L. Cantua pyrifolia Juss. ex Lam.; M', M". C. candelilla Brand; N. Phlox andicola Nutt. ex Gray; O. Bonplandia geminiflora Cav. at several levels, thus somewhat approaching Acanthogilia. Although the species gloriosa has been placed in Gilia, Ipomopsis, Leptodactylon, and Loese- lia, each of these genera has a venation pattern different from that of Acanthogilia. The distinc- tive patterns of Leptodactylon and Loeselia es- pecially seem to make close relationship with Acanthogilia unlikely. WOOD ANATOMY.— Carlquist et al. (1984) studied the wood anatomy of the Polemoniaceae, comparing the relatively few woody species. Be- 124 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 7 u bar= I mm FIGURE 14. Corolla venation patterns, cont. P. Loeselia greggii S. Wats.; Q. L. amplectens (Hook, and Am.) Benth.; R. Allophyllum glutinosum (Benth.) A. and V. Grant; S. Linanthus dianthiflorus (Benth.) Greene; T. L. grandiflorus (Benth.) Greene; U. Leptodactylon pungens (Torr.) Rydb. DAY AND MORAN: ACANTHOGILIA, NEW GENUS OF POLEMONIACEAE 125 sides Acanthogilia, these species fall in Cantua and Huthia (Cantueae), Cobaea (Cobaeeae), and Eriastrum, Ipomopsis, and Leptodactylon (Gi- lieae). In general, they thought the wood anato- my probably more closely correlated with growth form as related to habitat, than with systematic relationships. The authors noted that Acanthogilia and a species of Ipomopsis are alike in having banded axile parenchyma— unusual in the family but oc- curring also, in its most incipient form, in Can- tua. Likewise, in the imperforate tracheary ele- ments and in the vascular rays, Acanthogilia is similar to Ipomopsis on the one hand and to the Cantueae on the other. The coincidence in these characters among Acanthogilia, the Gilieae, and the Cantueae somewhat parallels other similar- ities we report here (Table 1) and would seem to be due to relationships rather than to environ- mental factors alone. RELATIONSHIPS Acanthogilia is unique among Polemoniaceae in its extreme leaf dimorphism, its persistent woody-spinose primary leaves, and its coarsely verrucate zonocolporate pollen grains. Although A. gloriosa, the sole species, has been placed in Gilia, Ipomopsis, Leptodactylon, and Loeselia, it differs further from all these genera in its persis- tent secondary leaf bases with deciduous blades, its numerous closely spaced corolla veins with interconnections at several levels, and its winged seeds, and from all these except for two species of Gilia in its superficially adnate filaments. It differs still further from Gilia and Ipomopsis in its large shrubby habit, from Ipomopsis in its basic chromosome number of nine, and from Leptodactylon in its pinnate leaves and its three corolla veins instead of one in each sector of the lower tube. Among North American Polemoniaceae, Acanthogilia seems to have most in common with Gilia and allies, and especially with species of Gilia sect. Giliastrum. As in Acanthogilia, all Gilia species have the calyx membranous below the sinuses, and most, including sect. Giliastrum, have zonocolporate pollen and have the primi- tive x = 9. In this polymorphic genus of five sec- tions, usually the pollen is blue and the exine reticulate to striate and not verrucate. In sect. Giliastrum, however, the pollen is yellow as in Acanthogilia, and the exine is somewhat similar, being pertectate and minutely verrucate whereas in Acanthogilia it is perreticulate and coarsely verrucate. Although in most species of Gilia the filaments merge with the corolla below, in G. ripleyi and G. rigidula, of sect. Giliastrum, the filaments are superficially adnate, as in Acantho- gilia. Most species of Gilia are annual and none are truly woody, but G. ripleyi is a suffrutescent perennial. Finally, although most species of sect. Giliastrum have free corolla veins, G. ripleyi is unique in Gilia and further resembles Acantho- gilia in having the veins connected at several levels. Acanthogilia is perhaps most closely related to the Andean genus Cantua. Cantua, like Acan- thogilia, is shrubby, with leaves dimorphic, borne on long shoots and axillary short shoots, with crowded leaf bases remaining on the short shoots after the blades have fallen, with corolla veins connected at more than one level, with seeds flattened and broadly winged, with a basic chro- mosome number of nine, with superficially ad- nate filaments in C. candelilla and C. quercifolia, and with coarsely verrucate pollen in C. buxi- folia. The lower branches of C. buxifolia (grown in San Francisco) take root, as do those of Acan- thogilia. Cantua differs in that the leaves are broadly herbaceous and only slightly dimorphic, with primary leaves deciduous, not at all woody- persistent, and secondary leaves more persistent; the calyx herbaceous, not membranous below the sinuses, and not rupturing in age; the pollen pan- toporate, not zonocolporate; the chromosome number hexaploid, not diploid. We suggest that Acanthogilia may be a specialized desert descen- dent of a diploid line also ancestral to Cantua. Since Cantua is hexaploid, however, and prob- ably amphiploid, such divergent characters as the herbaceous calyx may perhaps derive from some other line. Grant (1959) divided the Polemoniaceae into five tribes. Acanthogilia probably belongs to the Cantueae but apparently has some relationship also with the Gilieae. Much new evidence bear- ing on generic relationships has accumulated, es- pecially from pollen studies, since Grant's clas- sification, and the time seems ripe for a new tribal arrangement. ACKNOWLEDGMENTS We are grateful to all who have helped with this paper, and especially the following: Dr. Leon 126 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 7 Stuchlik of the Botanical Institute, Cracow, Po- land, for his pollen diagnosis of Acanthogilia and observations concerning it; Jerry Morgan of Uni- versity of California, San Francisco, for the SEM micrographs; the Gift Fund, Dudley Herbarium of Stanford University for support for the SEM work; Terry Bell for line drawings of Acantho- gilia; Colleen Sudekum for the corolla venation tracings; Barbara Stewart for drawing the distri- bution map; and Dave and Dru Binney for send- ing two points for the map, from far, out-of-the- way places. We appreciate suggestions given by Dr. Robert Patterson on the pollen morphology section, by John Thomas Howell on the Latin diagnosis, and by Dr. Dennis Breedlove and Dr. Leslie Landrum on various points along the way. An early preview of the wood-anatomy manu- script was kindly offered by the senior author, Dr. Sherwin Carlquist. We thank Dr. Frank Al- meda, Dr. Christopher Davidson, Dr. Verne Grant, and two unsigned reviewers for reading this manuscript and making many helpful com- ments. We appreciate the courtesies extended at the following herbaria: California Academy of Sciences (CAS), Dudley Herbarium of Stanford University (DS), San Diego Museum of Natural History (SD), and University of California, Berkeley (UC). Lastly, for making the color plate possible, we are most grateful to seventy-two col- leagues and friends enlisted by Annetta Carter and Lincoln Constance as Los Amigos de Acan- thogilia. BRAND, A. 1907. reich 4(250). LITERATURE CITED Polemoniaceae. A. Engler, Das Pflanzen- BRANDEGEE, T. S. 1889. A collection of plants from Baja California, 1889. Proc. Calif. Acad. Sci. Ser. 2, 2:1 17-216. CARLQUIST, S., V. M. ECKHART, AND D. C. MICHENER. 1984. Wood anatomy of Polemoniaceae. Aliso 10(4):547-572. CHUANG, T., W. C. HSIEH, AND D. H. WILKEN. 1978. Con- tribution of pollen morphology to systematics in Collomia (Polemoniaceae). Am. J. Bot. 65:450-458. ERDTMAN, G. 1952. Pollen morphology and plant taxonomy: angiosperms. The Chronica Botanica Co., Waltham, Mas- sachusetts. 539 pp. GIBSON, D. N. 1967. Polemoniaceae. In Flora of Peru, Mac- bride, ed. Field Mus. Nat. Hist., Bot. Ser. 13(5a:2):l 12-131. GRANT, V. 1 956. A synopsis of Ipomopsis. Aliso 3:35 1-362. . 1959. Natural history of the Phlox family: 1. Sys- tematic botany. Martinus Nijhoff, The Hague, Netherlands. GRANT, V. AND K. A. GRANT. 1965. Flower pollination in the Phlox family. Columbia Univ. Press, New York. INFANTES VERA, J. G. 1962. Revisi6n del genero Cantua (Polemoniaceae). Lilloa 31:75-107. JOHNSTON, I. M. 1 924. Expedition of the California Academy of Sciences to the Gulf of California in 1921: the Botany (the vascular plants). Proc. Calif. Acad. Sci. Ser. 4, 12:951- 1218. LOEBLICH, A. R. 1964. The pollen grain morphology of Col- lomia as a taxonomic tool. Madrono 17:205-216. MARTINEZ, M. 1937. Catalogo de nombres vulgares y cien- tificos de plantas Mexicanas. Ediciones Botas, Mexico. MORAN, R. 1952. The Mexican itineraries of T. S. Brande- gee. Madrono 11:253-262. STUCHLIK, L. 1967. Pollen morphology and taxonomy in the Polemoniaceae. Grana Palynol. 7:146-240. TAYLOR, T. N. AND D. A. LEVIN. 1975. Pollen morphology of Polemoniaceae in relation to systematics and pollination systems: scanning electron microscopy. Grana 15:91-1 12. WHERRY, E. T. 1945. Two Linanthoid genera. Am. Midi. Naturalist 34:38 1-387. WIGGINS, I. L. 1964. Flora of the Sonoran Desert. Pp. 189- 1740 in Vegetation and flora of the Sonoran Desert, Shreve and Wiggins, eds., 2 vols. Stanford Univ. Press, Stanford, California. . 1 980. Flora of Baja California. Stanford Univ. Press, Stanford, California. 1025 pp. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94118 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 8, pp. 127-156, 106 figs., 2 tables. May 6, 1986 THE DIATOM GENUS THALASSIOSIRA: SPECIES FROM THE SAN FRANCISCO BAY SYSTEM By A. D. Mahood Department of Geology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 G. A. Fryxell Department of Oceanography, Texas A & M University, College Station, Texas 77843 M. McMillan Department of Botany and Range Science, Brigham Young University, Prove, Utah 84602 ABSTRACT: Twenty species of the diatom genus Thalassiosira, including one previously undescribed species, were collected from diverse habitats of San Francisco Bay, California, extending from nearly freshwater in Suisun Bay to marine salinity near the Golden Gate. In this paper the morphology of these 20 species is elucidated by scanning electron microscopy (SEM) and light microscopy (LM). The species range from large taxa with linear areolae to small, lightly silicified forms with eccentrically arranged areolae. The basic form of the genus is seen as a moderately small diatom with rectangular outline in girdle view and a simple process pattern of one central and one marginal ring of strutted processes and a single labiate process. The distribution of species is influenced by salinity. Major species show limited distributions: T. visurgis Hustedt, fresh to brackish water; T. decipiens (Grunow) Jorgensen, brackish water; T, nodulolineata (Hendey) Hasle and G. Fryxell, tidal marine to brackish; T. hendeyi Hasle and G. Fryxell, and T. wongii Mahood sp. nov., tidal marine water; T. nordenskioeldii Cleve and T. pacifica Gran and Angst, coastal marine. These distributions demonstrate the value of species as indicators of salinity patterns within the bay ecosystem. INTRODUCTION . . . . . . their interactions begun at the University of Cal- The San Francisco Bay system has been stud- ifornia, Berkeley (Hedgepeth 1979). The phy- ied by many investigators since 1816, when the toplankton flora was not examined until 1939, Russian ship Rurik anchored in the bay (Hedg- when F. W. Whedon, using a Sedgewick-Rafter peth 1979). The bay estuarine system extends chamber, made a limited study of San Francisco from the mouth of the Guadalupe River in the Bay phytoplankton and presented a brief species south to the lower reaches of the Sacramento- list, including Thalassiosira rotula Meunier. San Joaquin delta near the city of Pittsburg (Fig. The flora of San Francisco Bay remained large- 1) in the north. Early studies of the bay concen- ly unstudied until 1958, when the Sanitary En- trated on hydrology, fisheries, and physical pa- gineering Research Laboratory (SERL) of the rameters of the system. Not until the early 1 920s University of California, Berkeley, began a mul- was a serious effort to study the bay's biota and tidisciplinary study of the bay (Harris et al. 1961; [127] 128 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 TABLE 1 . KEY TO COLLECTION SITES, DATES OF COLLECTION, AND COLLECTORS. Date Collector Location 3 September 1980 R. L. J. Wong 1 16 October 1980 R. L. J. Wong 2 2 May 1971 A. D. Mahood 3 6 July 1981 G. A. FryxelV 3 A. D. Mahood 6 June 1971 A. D. Mahood 4 20 February 1972 C. A. McNeil 4 6 July 1981 A. D. Mahood 5 24 April 1971 A. Hauser 6 June 1972 P. Wares 6 1 November 1979 L. L. Lack 7 3 September 1980 R. L. J. Wong 8 13 November 1980 R. L. J. Wong 9 12 September 1979 A. D. Mahood 10 7 July 1981 A. D. Mahood 10 Storrs et al. 1 966). Although the species list gen- erated by SERL was established from raw ma- terial (i.e., not cleaned of organic material so that the siliceous parts can be clearly seen) and was thus of limited systematic value, it is one of the few comprehensive references to the San Fran- cisco Bay diatoms available. Hasle (1978#, b) published the first careful taxonomic papers on Thalassiosira from the bay system. The most recent inventory of bay diatom species (Wong and Cloern 1981), though still in progress, offers a basis for future diatom floristics and phyto- plankton ecology studies of the bay. Thalassiosira Cleve, one of the dominant dia- tom genera in the bay's phytoplankton, has been frequently mentioned in species lists: T. decip- iens (Grunow) Jorgensen; T. eccentrica (Ehren- berg) Cleve; T. lacustris (Grunow) Hasle (=Cos- cinodiscus lacustris Hustedt); T. nordenskioeldii Cleve; T. punctigera (Castracane) Hasle (=T. angstii (Gran) Makarova); and T. rotula Meu- nier. The diversity of species of the genus in the bay system is not surprising because the Tha- lassiosira probably evolved in a similar estuarine and coastal environment (Round and Sims 1 98 1). The species diversity of diatoms is limited by salinity variations within the bay system (pers. comm. Wong, United States Geological Survey, 1983). The use of the genus Thalassiosira as an indicator of marine influence in the bay system has been hindered by several conditions: the con- fusion with other centric genera (e.g., Coscino- discus), difficulty in identification (especially in water mounts), problems connected with sample preparation (including the need for mounting in a medium with a high refractive index for light microscopy), and widely scattered pertinent taxonomic literature. In this paper we report on the distribution of Thalassiosira species in the bay and discuss mor- phology and its application to systematic prob- lems within the bay. METHODS AND MATERIALS Samples were collected with a Kemmerer water bottle and net over a broad range of the navigable section of the San Francisco Bay estuarine system (Fig. 1). Net hauls are particularly useful in con- centrating material for identification before quantitative work is attempted on water sam- ples. Fixed samples were examined in Utermohl settling chambers using an Olympus IMT in- verted microscope. Most often positive identi- fication must be made from cleaned material, mounted in medium of high refractive index such as Hyrax (Hanna 1930). Material was cleaned using the Van der Werff(1955) hydrogen per- oxide method. Strewn and arranged slides were prepared for light microscope (LM) from cleaned materials and mounted in Hyrax. Similar strewn and also arranged material (mounted by the se- nior author) was prepared on scanning electron microscope (SEM) stubs for examination and photography using Jeolco JSM-35, Texas A&M University Electron Microscopy Center. Texas A&M Department of Oceanography phyto- plankton cultures F190, F206, F209, and F226 were isolated by T. P. Watkins from collections made about a mile north of the Golden Gate Bridge, 6 July 1981. They were maintained in the Department of Oceanography phytoplankton culture collection: F/2 medium, 30%o salinity, 16 hours light and 8 hours dark cycle, in 1 6°C growth chambers. Terminology follows that of the work- ing Party on Diatom Terminology (Anonymous 1975; Ross et al. 1979). RESULTS AND DISCUSSION Twenty species of Thalassiosira were re- covered from 33 samples from San Francisco Bay: T. anguste-lineata (A. Schmidt) G. Fryxell and Hasle; T. decipiens (Grunow) Jergensen; T. eccentrica (Ehrenberg) Cleve; T. endoseriata Ha- sle and G. Fryxell; T. hendeyi Hasle and G. Fryxell; T. incerta Makarova; T. lacustris (Gru- now) Hasle; T. lundiana G. Fryxell; T. minuscula MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 129 Pacific Ocean FIGURE 1 . San Francisco Bay System. 130 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 Krasske; T. nodulolineata (Hendey) Hasle and G. Fryxell; T. nordenskioeldii Cleve; T. oestrupii var. venrickae G. Fryxell and Hasle; T. pacifica Gran and Angst; T. punctigera (Castracane) Ha- sle; T. rotula Meunier; T. simonsenii Hasle and G. FryxeU; T. stellaris Hasle and Guillard; T. tenera Proschkina-Lavrenko; T. visurgis Hus- tedt; and T. wongii Mahood sp. nov. An understanding of the individual species, their similarities and differences, the habitats in which they occur, and their distributions within the bay system is essential to understanding the phytoplankton dynamics of the bay and the role of the Thalassiosira species in the bay ecosystem. The genus is characterized by a number of morphological features: one to a few labiate pro- cesses (Fig. 16); many strutted processes (Fig. 1 1); none to many occluded processes (Fig. 38); eccentric, linear, or fasciculated patterns of the rows of areolae, in a basically radial pattern; and the placement of the areola cribrum on the in- ternal side of the loculate areolae and the fora- men on the external side (Fig. 14, 17). For this paper and to assist the light microscopist, we have stressed the characters particularly visible in LM: number and location of labiate processes; number and arrangement of the areolae; and lo- cation and number of strutted processes. Several species can be identified under LM, whereas oth- ers require SEM in order to resolve definitive characters. Because our primary purpose includes gaining a greater understanding of San Francisco Bay, an essential goal is to distinguish between marker species with similar morphological characteris- tics but differing distributions, species which have been confused in earlier literature. For such com- parisons, we have separated the Thalassiosira species in this paper into five morphological groups: 1) species with basic linear areola pat- terns, 2) species with eccentric areola patterns, 3) species with one central strutted process and one marginal ring of strutted processes, 4) species with no central strutted processes and modified rings of strutted processes, and 5) two otherwise dissimilar species that have radial areola patterns and multiple central strutted processes. These are form groupings only; they are not placed together to imply close taxonomic relationships. GROUP I.— Thalassiosira species with linear ar- eola patterns. Thalassiosira simonsenii Hasle and G. Fryxell, 1977 (Figures 2-5) DETAILED DESCRIPTION. Hasle and Fryxell (1977).— Cell di- ameter 30-57 ^m; areolae 4-5.5 in 10 ^m across the valve, 8- 1 0 in 1 0 urn at mantle edge (Fig. 2, 4); one small central strutted process (Fig. 5); two rows of alternating strutted processes on margin, five to six in 10 nm (Fig. 3); two opposing labiate processes (Fig. 2); distinctive marginal ribs, eight in 10 nm; large tubular occluded processes, one in 10 jim on margin above strutted processes (Fig. 4). Marginal ribs distinctive in SEM, but not always clear in LM. DISTRIBUTION.— Marine, found only in central San Francisco Bay in association with other ma- rine diatoms. Observed previously at 28°00'N, 112°17.5'W, Pacific (Hasle and Fryxell 1977). Thalassiosira hendeyi Hasle and G. Fryxell, 1977 (Figures 6-1 1,86) DETAILED DESCRIPTION. Hasle and Fryxell (1977).— Cell di- ameter 38-120 ion; areolae, regularly linear, five to six 10 tan; prominent central strutted process (Fig. 6, 8) set to one side of central areola; two closely adjacent rings of marginal strutted processes (Fig. 9) alternating in orientation (Fig. 1 1 , internal view), not easily resolved with LM (Fig. 86); wavy mantle ridge (Fig. 9, 86); two labiate processes (Fig. 6, external [arrow]; Fig. 10, internal); labiate process with two adjacent strutted pro- cesses (Fig. 7); valve slightly concentrically undulated (Fig. 6). DISTRIBUTION.— Common but never abun- dant from San Pablo Bay to south San Francisco Bay. Previously found in warm coastal waters of West Africa (Hasle and Fryxell 1977), Uruguay and Brazil (Muller-Melchers 1953). Thalassiosira nodulolineata (Hendey) Hasle and G. Fryxell, 1977 (Figures 12-17, 87, 90) DETAILED DESCRIPTION. Hasle and Fryxell (1977).— Cell di- ameter 26-58 Mm; areolae linear, regular, 3.5 (one with 9.0) in 10 nm (Fig. 12); four strutted processes in 10 nm with spines (Fig. 12, 15) along the margin; five to six strutted processes inside the central areola (Fig. 14, external; Fig. 17, internal; Fig. 90); single labiate process at valve margin (Fig. 12, 16). In our samples, central areolae surrounded by six symmetrical areolae (Fig. 90), although central areolae surrounded by five areolae have been reported (Hasle and Fryxell 1977). DISTRIBUTION.— Central San Francisco Bay, San Pablo Bay, and Suisun Slough; common but never in large numbers. Thalassiosira tenera Proschkina-Lavrenko, 1961 (Figures 18-23, 103-104) DETAILED DESCRIPTION. Hasle and Fryxell (1977).— Cell di- ameter 10-29 urn; areolae 9-16 in 10 pm; marginal strutted MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 131 FIGURES 2-5. Thalassiosira simonsenii Hasle and G. Fryxell. SEM. Figure 2. Scale = 5 urn, external view of valve, two labiate processes across valve from each other (arrow). Figure 3. Scale = 2 /im, two rows of alternating strutted processes on margin. Figure 4. Scale = 1 /*m, large tubular occluded process on margin above strutted processes. Figure 5. Scale = 1 pm, view of central strutted process. processes three to five in 10 jtm, one central strutted process (Fig. 18-20, 23); one labiate process (Fig. 20, 22, arrow); ar- rangement of areolae linear (Fig. 23, 103, 104) or fasciculated; marginal strutted processes often with siliceous overgrowths and flattened (Fig. 18, 21, 23). DISTRIBUTION.— Restricted to coastal waters (Hasle and Fryxell 1977). Material examined from San Francisco Bay entirely prepared from cultures from samples taken just north of the Golden Gate (6 July 1981). DISCUSSION.— T. hendeyi, T. simonsenii, and T. nodulolineata have been confused in bay stud- ies and erroneously reported as Coscinodiscus lineatus Ehrenberg (=T. leptopus [Grunow] Ha- sle and G. Fryxell). In our samples, T. nodulo- lineata specimens were most easily distinguished 132 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 from the other two by having a single labiate process and six symmetrical areolae surrounding the central areola (Fig. 14, 17). The five to six strutted processes inside the central areola fur- ther confirm the identification. Although T. te- nera and T. incerta Makarova (to be discussed in Group 5) have a similar appearance to T. nod- ulolineata, the coarser areolae of T. nodulolineata distinguish it from the two with finer areolae. In addition, the marginal strutted processes of T. tenera often are in heavily silicified areas of the margin and are markedly distinct in LM, in con- trast to other species. Distinguishing specimens of T. hendeyi and T. simonsenii is extremely difficult under LM. Ex- amination of SEM micrographs of the margin of T. hendeyi (Fig. 6) and the central area (Fig. 8) will assist the light microscopist in differentiating these two species. Both species have two labiate processes and similar central areas. For identi- fication under the LM (Fig. 86), careful exami- nation of the margin is necessary. Under the SEM (Fig. 6, 9), the margin of T. hendeyi is irregular and wavy, but with the light microscope a clean break can be noticed when focusing from the margin to the valve face. The valve face forms a rather sharp angle with the margin in T. hen- deyi, while in T. simonsenii (Fig. 4) the transition from valve face to margin is smooth. Although the marginal strutted processes of T. simonsenii are more prominent than those of T. hendeyi, this characteristic is not easily resolved under LM. In our samples the salinity ranges for T. hen- deyi and T. nodulolineata overlap, and both are found from south San Francisco Bay to Suisun Bay, indicating a more brackish environment. Thalassiosira simonsenii, much less common than the others, was confined to the central San Francisco Bay and was associated with other ma- rine diatoms. Thalassiosira tenera was collected near the Golden Gate Bridge in a marine habitat. Thalassiosira tenera is the only one of this group with strutted processes in the middle of the cen- tral areola. GROUP 2.— Thalassiosira species with eccentric areola patterns. Thalassiosira wongii Mahood, sp. nov. (Figures 24-29, 99-101) DETAILED DESCRIPTION. Mahood.— Diameter 27-51 tmi; ar- eolae radial, fasciculated, 9-11 in 10 |im with areolar rows parallel to central areolar row of fascicle; strutted processes at margin three to five in 10 ^m (Fig. 24-26, 29, 99-101); four to five strutted processes observed around central areola (Fig. 25, 26); strutted processes usually forming two irregular rings between central areola and margin (Fig. 26, 99, 100); one mar- ginal ring of strutted processes near margin, four to five in 10 ttm (Fig. 29); one labiate process set slightly inside marginal ring (Fig. 27, internal and external, Fig. 28, 99, 100), three spines in 10 jim around margin in same ring as labiate process (Fig. 24); small spines often above each strutted process on outside of valve (Fig. 27, internal and external; Fig. 29). Diameter 27-51 nm; areolae 9-11 in 10 nm; fultoportulae ad marginem, 3-5 in 10 nm; 4-5 fultoportulae circum areolam centralem visae; fultoportulae plus minusve annulos duos ir- regulares formantes inter areolam centralem et marginem; an- nulus marginalis unicus fultoportularum prope marginem 4-5 FIGURES 6-11. Thalassiosira hendeyi Hasle and G. Fryxell. SEM. Figure 6. Scale = 10 ^m; external view of valve; two labiate processes (arrow), valve slightly concentrically undulated, distinctive marginal ridge, linear areolae. Figure 7. Scale = 1 Mm, labiate process with two adjacent processes. Figure 8. Scale = 1 nm, prominent central process. Figure 9. Scale = 1 ion; wavy marginal ridge, two closely adjacent rows of strutted processes. Figure 10. Scale = 10 iaa, internal view of valve, alternating marginal strutted processes. Figure 1 1 . Scale = 1 /urn, internal view of valve, alternating marginal strutted processes. FIGURES 12-17. Thalassiosira nodulolineata (Hendey) Hasle and G. Fryxell. SEM. Figure 12. Scale = 5 pan; external view of valve; linear areolae, marginal strutted processes (sp), marginal spines (ms), single labiate process (Ip). Figure 13. Scale = 1 Aim, single labiate process, bands. Figure 14. Scale = 1 ton; strutted processes in the central areola, six symmetrical areolae surrounding central areola with radial threads. Figure 15. Scale = 0.5 /tm, external marginal spine, strutted processes. Figure 16. Scale = 1 nm, internal view of valve, internal labia with external labiate process (arrow). Figure 17. Scale = 1 pan, internal view of central strutted processes. FIGURES 18-23. Thalassiosira tenera Proschkina-Lavrenko. SEM. Figure 18. Scale = 2 pun; external view of valve; marginal strutted processes, one central strutted process. Figure 19. Scale = 1 pun, distinctive central strutted process. Figure 20. Scale = 2 /im, internal view of valve, one central strutted process. Figure 2 1 . Scale = 1 ^m, external view of marginal strutted process with siliceous overgrowths. Figure 22. Scale = 0.5 pan, internal view of labia (arrow). Figure 23. Scale = 2 ^m; external view of valve; single external labiate process (arrow), marginal strutted processes with siliceous overgrowths. MAHOOD, FRYXELL, AND McMILLAN: THALASS1OSIRA ' -, 133 134 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 135 136 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 FIGURES 24-29. Thalassiosira wongii Mahood, sp. nov. SEM. Figure 24. Scale = 10 nm; external view of valve; fasciculated areolae, small marginal spines, single labiate process (arrow). Figure 25. Scale = 1 nm, external view of irregularly arranged processes surrounding the central areola. Figure 26. Scale = 10 pm; internal view of valve; single labia, two irregular rings of strutted processes, marginal ring of strutted processes. Figure 27. Scale = 2 /zm; internal view of labia and external labiate MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 137 in 10 nm; rimportula unica annuli marginalis parum penitus posita, spinis 3 in 10 pm circum margineum in eodem annulo cum rimoportula; spinae parvae saepe supra fultoportulam omnem extus valvarum. DISTRIBUTION.— South and central San Fran- cisco Bay, associated with marine diatoms. HOLOTYPE. — Slide, deposited at California Academy of Sciences, CAS 61243. Thalassiosira oestrupii var. venrickae G. Fryxell and Hasle, 1980 (Not figured) DETAILED DESCRIPTION. Fryxell and Hasle (1980).— Cell di- ameter 5.5-39 Aim; areolae 6-9 in 10 Aim in central area, 7-1 1 in 10 Mm toward margin; one labiate process, usually located three areolae from center; one central strutted process, one to two in 10 Aim, on margin. May be confused with T. decipiens, but marginal strutted processes internal rather than external. Dominant characteristics: labiate process away from margin and marginal strutted processes with internal projection seen in same plane of focus (Fryxell and Hasle 1980, fig. 15 A, B, 17). DISTRIBUTION. — Coastal, temperate waters. Central San Francisco Bay, rare. Thalassiosira eccentrica (Ehrenberg) Cleve, 1904 (Figures 30-35, 102) DETAILED DESCRIPTION. Fryxell and Hasle (1972).— Cell di- ameter 12-101; areolae 5-8 in 10 ^m in central area, 7-10 in 10 Aim toward margin (Fig. 102); scattered strutted processes across valve face (Fig. 32); irregular spines around margin, three to four in 10 /im; central areola surrounded by seven areolae with single strutted process next to the central areola (Fig. 30, 33); two to three rings of strutted processes in 10 /tm near margin (Fig. 31; Hasle 1979); one prominent labiate pro- cess (Fig. 34, internal; Fig. 35). Valve face relatively flat. DISTRIBUTION.— Possibly brackish to definite- ly saline waters, common in central San Fran- cisco Bay. DISCUSSION. —The confusion within this group stems in part from the areola patterns of T. ec- centrica (Fryxell and Hasle 1972), which varies from the classic eccentric pattern to a fascicu- lated arrangement. For example Cupp (1943), a primary reference source for Bay studies, does not clearly distinguish T. decipiens (to be dis- cussed in Group 3) and T. eccentrica (=Cosci- nodiscus eccentricus). The overall eccentric pat- tern is representative of three species in this group. Other morphological characteristics may be used to facilitate identification. Thalassiosira oestru- pii var. venrickae lacks external labiate process tubes. Only T. wongii has multiple central pro- cesses, fasciculated valves, and spines above each strutted process. A single row of marginal strut- ted processes further distinguishes T. wongii from T. eccentrica. The spines above each strutted process of T. wongii are only seen under SEM. GROUP 3. — Thalassiosira species with one cen- tral and one marginal ring of strutted processes and one labiate process near the margin. Thalassiosira minuscula Krasske, 1941 (Figures 36, 37) DETAILED DESCRIPTION. Hasle (1976).— Cell diameter 10- 20 /im; areolae small, 30 in 10 ^m in rows parallel to radial row; one large radially oriented labiate process located in from margin (Fig. 36); strutted processes in ring on margin four to five in 10 tim near margin, one strutted process in center (Fig. 36), plus one adjacent to labiate process (Fig. 37). Distin- guished from other fasciculated species of Thalassiosira in San Francisco Bay by the unique arrangement of a strutted process adjacent to the labiate process and by the finer areolae. DISTRIBUTION. — Originally described from coastal plankton of Chile (Krasske 1 94 1). Central San Francisco Bay, rare. Thalassiosira lundiana G. Fryxell, 191 5a (Figures 36-41, 89) DETAILED DESCRIPTION. Fryxell (1975a).— Cell diameter 7- 43 Mm; areolae 24-30 in 10 jtm, fasciculated; marginal striae; strutted processes in ring inside margin (Fig. 38), approxi- mately 10 in 10 Aim; often irregularly arranged large occluded processes in ring farther from margin (Fig. 38, 39); one central strutted process (Fig. 40, arrow; Fig. 41, csp); one labiate pro- cess inside marginal ring (Fig. 41, arrow) in same ring as oc- cluded processes; weakly silicified. DISTRIBUTION.— Inshore euryhaline (Fryxell 1975a); found in our samples from central San Francisco Bay to Suisun Bay, indicating a broad- er freshwater range than previously proposed. Thalassiosira punctigera (Castracane) Hasle, 1983 (=T. angstii (Gran) Makarova) (Figures 42-48, 92) DETAILED DESCRIPTION. Gran and Angst (1931), G. Fryxell (1978).— Cell diameter 43-145 nm; areolae across valve face process, small spines above marginal strutted processes. Figure 28. Scale = 2 ^m; external view of margin at the single labiate process, valve strutted processes (arrow). Figure 29. Scale = 2 ^m, marginal strutted processes with short spines, large spines back from edge of margin. 138 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 1 5 in 1 0 Mm; areolae fasciculated with areolae arranged parallel to center of fascicle (Fig. 45); strutted processes four to five in 10 nm along margin (Fig. 44, 46, 92); single central process (Fig. 43, arrow); one labiate process (Fig. 46; Fig. 47, internal); large occluded processes irregularly arranged around margin (Fig. 42), although some specimens lack large occluded pro- cesses entirely (Fig. 48). DISTRIBUTION.— Coastal, large population of T. punctigera reported in Richardson Bay (A. Hauser, pers. comm., 1976), which is constantly flushed by waters from the Golden Gate. Thalassiosira nordenskioeldii Cleve, 1873 (Figure 106) DETAILED DESCRIPTION. Hasle (19786).— Cell diameter 10- 50 /tin; areolae 14-18 in 10 pm across valve face; marginal strutted processes three in 10 Mm; separated from margin by ca. 6-8 areolae. Single strutted process in center of valve (Fig. 106); one labiate process in same ring as marginal strutted processes but not in a constant position relative to a strutted process; marginal striae 18-20 in 10 Mm in mantle rim. DISTRIBUTION.— Marine, cold water (Hasle 1978&); only in samples near the Golden Gate Bridge. Thalassiosira pacific Gran and Angst, 1 93 1 (Figures 49-55, 105) DETAILED DESCRIPTION. Hasle (1978ft).— Cell diameter 7- 46 Mm; areolae 10-18 in 10 /tin in central area, 20 in 10 nm at margin (Fig. 49); one labiate process (Fig. 54); pronounced, regular marginal strutted processes, four to seven in 10 nm (Fig. 52); single central strutted process (Fig. 51, internal; Fig. 53, external; Fig. 105) adjacent to central areolae. Areolae usually in fasciculated rows with areolae parallel to radius. DISTRIBUTION.— Marine, from central San Francisco Bay through Golden Gate to Gulf of Farallons. Thalassiosira visurgis Hustedt, 1957 (Figures 56-6 1,95, 96) DETAILED DESCRIPTION. Hasle (1978a).— Cell diameter 9- 18 Mm; areolae in central area 13-14 in 10 /mi, 18 in 10 /tm at the margin (Fig. 58); one central strutted process (Fig. 58, external; Fig. 60, internal [arrow]), and four to five strutted processes in 10 /im in ring near margin (Fig. 56, external; Fig. 57); two labiate processes on opposing sides, each found be- tween marginal strutted processes and slightly inside the ring of strutted processes (Fig. 56, 58-60, 95, 96). In our prepa- ration, valves often convex or concave, suggesting heteroval- vate cells with one convex and one concave valve. Granules on valve often extending onto processes (Fig. 57). DISTRIBUTION.— Usually found surrounded with silt and clay, fresh to brackish water. Tha- lassiosira visurgis occasionally the dominant Thalassiosira in Suisun Slough, a brackish en- vironment. Thalassiosira decipiens (Grunow) Jorg., 1905 (Figures 62-67, 97, 98) DETAILED DESCRIPTION. Hasle (1979).— Cell diameter 9-29 Mm; areolae across valve face 8-12 in 10 i*m, much smaller on mantle (Fig. 62); single ring of marginal strutted processes four to six in 10 nm (Fig. 62, 64); one central strutted process (Fig. 63, arrow); labiate process located between two marginal strut- ted processes (Fig. 62, arrow; Fig. 97, arrow; Fig. 98), closer FIGURES 30-35. Thalassiosira eccentrica (Ehrenb.) Cleve. SEM. Figure 30. Scale = 20 Mm; external view of valve; eccentric areolar pattern, irregular spines. Figure 31. Scale = 2 Mm, large marginal spines, two rings of marginal strutted processes. Figure 32. Scale = 20 nm, internal view of valve, scattered processes across valve. Figure 33. Scale = 3 nm; internal view of central area, seven areolae with cribra surrounding central areola (ca), central strutted process (csp) just off center. Figure 34. Scale = 2 /on, internal view of labia, scattered strutted processes. Figure 35. Scale = 3 Mm; external view of valve; single labiate process (Ip), scattered strutted processes, marginal strutted processes. FIGURES 36, 37. Thalassiosira minuscula Krasske. SEM. Figure 36. Scale = 10 Mm; internal view of valve, single labia set back from margin, single row of marginal strutted processes, central strutted process. Figure 37. Scale = 2 fim; same valve, internal view of marginal strutted processes, labia set back from margin. FIGURES 38-41. Thalassiosira lundiana G. Fryxell. SEM. Figure 38. Scale = 10 Mm; external view of valve; irregular strutted processes ring valve, irregular occluded processes (arrow). Figure 39. Scale = 2 Mm; large occluded process, marginal strutted processes. Figure 40. Scale = 2 Mm; external view of central area (arrow), fasciculation. Figure 41. Scale =10 Mm; internal view of valve; scattered valve strutted processes, single labiate process (arrow), single central strutted process (csp). FIGURES 42-48. Thalassiosira punctigera (Castracane) Hasle. SEM. Figure 42. Scale = 20 Mm; external view of valve; marginal strutted processes, irregular large occluded processes. Figure 43. Scale = 2 Mm, single strutted process in central area. Figure 44. Scale = 2 Mm, marginal strutted processes. Figure 45. Scale = 20 Mm, internal view of fasciculated areolae, marginal strutted processes. Figure 46. Scale = 2 Mm, external view of labiate process, marginal strutted processes. Figure 47. Scale = 2 Mm, internal view of labia and external labiate process. Figure 48. Scale = 20 Mm, external view of valve lacking occluded processes. MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 139 140 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 141 142 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 to one than the other. Often found coated with sediment (Fig. 67), DISTRIBUTION.— San Joaquin estuary, a brack- ish environment (Hasle 1979). Dominant Tha- lassiosira in the Suisun Bay area (Arthur and Ball 1980; Wong and Cloera 1981). DISCUSSION.— Confusion of the species T. punctigera and T. lundiana may lead to misin- terpretation of environmental conditions. Tha- lassiosira punctigera is the larger of the two species, although there is a slight overlap between the smaller diameters of T. punctigera and the larger diameters of T. lundiana. The larger di- ameter of T. punctigera, the more heavily silic- ified valve, and the prominent marginal strutted processes are the most important distinguishing features (Fig. 44, 46, 48, 92), with only a single strutted process in the center of the valve. Tha- lassiosira lundiana is fragile and weakly silicified making the fine areolar pattern more difficult to resolve, but strutted processes are scattered over the valve face (Fig. 89). The differentiation of T. decipiens and T. vi- surgis has proven both difficult and interesting. In our samples these species commonly appear in large numbers in Suisun Slough, Suisun Bay, and the delta of the Sacramento and San Joaquin rivers. The presence of the two labiate processes clearly distinguishes T. visurgis (Fig. 56, 60, 95, 96) from T. decipiens. Confusion of the species is possible if one process is obscured by detritus (Fig. 67). The best distinguishing character in both cleaned and uncleaned specimens is the ar- rangement of the areolae. In T. decipiens the number of areolae in 10 too. is constant across the valve face (Fig. 97, 98). On these small valves, the cell diameter, number of areolae, and the presence of the second labiate process may not be clear enough under the light microscope to differentiate the two species. A peculiar charac- teristic of T. visurgis is seen as the light micro- scope focuses through the valve: the central area areolae display a winking effect, optically sepa- rating the central area from the margin (Fig. 95, 96). Unless the material is cleaned and mounted in Hyrax or other suitable medium, however, we cannot differentiate these two species. The overall eccentric pattern, overlap of cell diameter, and overlap of number of areolae in 10 nm make it difficult to differentiate T. eccen- trica and T. decipiens in the 30 fim diameter range without reference to more recent works. Distinctive characteristics that aid their differ- entiation include pronounced, regular strutted processes seen on T. decipiens versus the irreg- ular spines of T. eccentrica and the concave or convex valves of T. decipiens versus the rela- tively flat valve of T. eccentrica. Thalassiosira eccentrica is a coastal marine species, while vi- able T. decipiens cells in the bay system are usu- ally restricted to the Suisun Bay-Delta, a brack- ish environment; non-viable cells may be flushed from Suisun Bay by tidal action. It is possible that T. eccentrica reported in eco- FIGURES 49-55. Thalassiosira pacifica Gran and Angst. SEM. Figure 49. Scale = 10 jtm; external view of valve; fasciculated areolae, single central strutted process, marginal strutted processes. Figure 50. Scale = 2 /mi, single labiate process (arrow) on margin between strutted processes. Figure 51. Scale = 1 /mi, internal view of single strutted process. Figure 52. Scale =10 /mi, regular marginal strutted processes. Figure 53. Scale = 1 /mi, external details of central strutted process. Figure 54. Scale = 1 /mi, internal view of labia and external labiate process. Figure 55. Scale = 10 /mi; internal view of valve; regular marginal strutted processes, single central strutted process, internal labia. FIGURES 56-61. Thalassiosira visurgis Hustedt. SEM. Figure 56. Scale = 2 /mi; external view of valve; strutted processes (sp) at margin, two labiate processes Op), single central strutted process (csp). Figure 57. Scale = 2 /mi, labiate process between two strutted processes on margin. Figure 58. Scale = 2 /mi, central areolae distinct from those toward margin. Figure 59. Scale = 2 /mi, internal view of convex valve, regular marginal strutted processes. Figure 60. Scale = 2 /tin, internal view of concave valve, two opposing labia. Figure 61. Scale = 2 /mi, external view of concave valve. FIGURES 62-67. Thalassiosira decipiens (Grunow) Jorgensen. SEM. Figure 62. Scale = 10 /mi; external view of valve; single labiate process (arrow), regular marginal strutted processes. Figure 63. Scale = 1 /mi; single central process (arrow), fine siliceous granulations on external side of valve. Figure 64. Scale = 10 /mi; internal view of valve; single labia (arrow), regular marginal strutted processes. Figure 65. Scale = 10 pan; external view of valve; consistent areola pattern across valve, regular arrangement of marginal strutted processes. Figure 66. Scale = 1 /an, labiate process between two strutted processes. Figure 67. Scale = 10 nm, detritus accumulation surrounding cell. MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 143 144 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 145 146 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 logical studies of San Francisco Bay (Arthur and Ball 1980) is in reality a complex of T. decipiens and T. visurgis, two species usually found in the entrapment zone in Suisun Bay. Freshwater from the Sacramento and San Joaquin rivers flows westward to meet the tidal wedge moving east- ward in Suisun Bay. This mixing area produces vertical and horizontal circulation that tends to concentrate populations of phytoplankton (Ar- thur and Ball 1980). The average salinity is low (0. l-5%o, Arthur and Ball 1 980; Sitts and Knight 1979) and well within the range established for T. decipiens, T. visurgis, and T. incerta. GROUP 4.—Thalassiosira species with no central strutted process but a modified ring of processes on the face of the valve. Thalassiosira stellaris Hasle and Guillard, 1977 (Not figured) DETAILED DESCRIPTION. Fryxell and Hasle (1977).— Cell di- ameter 6-20 nm; areolae elongate, fasciculated, 30 in 10 ion; marginal strutted processes three to five (sometimes six) in 10 Mm; a ring of two to seven strutted processes Vz the distance between the center and the margin. DISTRIBUTION.— Marine (Fryxell and Hasle 1977), San Francisco Bay samples described from cultures (F226) developed at Texas A&M Uni- versity and maintained at ca. 30%o salinity. Thalassiosira lacustris (Grunow) Hasle and Fryxell, 1977 (Figures 68-73, 88) DETAILED DESCRIPTION. As Coscinodiscus lacustris, Hustedt (1930).— Cell diameter 20-75 /im; areolae 10-14 in 10 nm in central area; valve face tangentially undulated (Fig. 68, 88); areolae fine (Fig. 71), arranged in dichotomous branching ra- diating rows (Fig. 71); five to seven strutted processes in ring ca. '/3 radius from center of valve (Fig. 68, external; Fig. 69, internal); also ring on edge of mantle (Fig. 69). Labiate process large, with slit parallel to the margin (Fig. 72). This species marked by tangentially undulated structure of the valve face (Fig. 68, 69) and areolar pattern. Other tangentially undulated centric species usually in San Francisco Bay include members of the genus Cyclotella. DISTRIBUTION.— Specimens in our samples from Suisun Slough indicate a brackish prefer- ence. Thalassiosira endoseriata Hasle and G. Fryxell, 1977 (Figure 91) DETAILED DESCRIPTION. Hasle and Fryxell (1977).— Cell di- ameter 20-60 /tm; areolae ll-18in 10/tm; one labiate process, located V* distance from margin to center; marginal strutted process projecting internally 5-6 in 10 nm; central irregular ring of 4-14 strutted processes. Areolae fasciculated with rows of areolae parallel to the radius (Fig. 91). DISTRIBUTION.— The number of specimens from samples was too small to draw positive conclusion concerning the distribution of this species, but it apparently lives at ca. 20%o salin- ity. Thalassiosira anguste-lineata (A. Schmidt) G. Fryxell and Hasle, 1977 (Figures 74-79, 93) DETAILED DESCRIPTION. Fryxell and Hasle (1977).— Cell di- ameter 17-78 nm; areolae fasciculated 8-18 in 10 urn; one FIGURES 68-73. Thalassiosira lacustris (Grunow) Hasle. SEM. Figure 68. Scale = 10 nm; external view of valve; tangentially undulated, marginal strutted processes. Figure 69. Scale = 1 tarn; internal view of valve; ring of valve strutted processes, pronounced tangential undulation. Figure 70. Scale = 1 fim; external view of margin, labiate process (arrow) between marginal strutted processes. Figure 7 1 (arrow). Scale = 1 /im; internal view of valve strutted process, areolae in dichotomous branching rows. Figure 72. Scale = 1 jim, internal view of labia and marginal strutted process. Figure 73. Scale = 10 /im; internal view of marginal strutted processes, dichotomous branching rows of areolae. FIGURES 74-79. Thalassiosira anguste-lineata (A. Schmidt) G. Fryxell and Hasle. SEM. Figure 74. Scale = 20 /xm; external valve view; areolae fasciculated, regular marginal strutted processes, arc of valve strutted processes in each fascicle (arrow). Figure 75. Scale = 1 ^m, external view of valve strutted processes. Figure 76. Scale = 2 ^m, marginal strutted processes. Figure 77. Scale = 2 nm; single labiate process (arrow) between two marginal strutted processes, small processes above each strutted process. Figure 78. Scale = 2 nm, internal view of labia. Figure 79. Scale = 2 /xm, internal view of valve strutted processes. FIGURES 80-85. Thalassiosira rotula Meunier. SEM. Figure 80. Scale = 2 jim, external view of central area strutted processes. Figure 81. Scale = 2 nm, internal view of central area strutted processes and scattered strutted processes on valve. Figure 82. Scale = 10 /im; external view of center and margin; weakly silicified valve, many marginal and valve strutted processes. Figure 83. Scale = 10 ion, internal view of valve, central and valve strutted processes. Figure 84. Scale = 10 ion, external view of valve, single labiate process (arrow), radial arrangement of areolae. Figure 85. Scale = 2 /tm, external labiate process (arrow). MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 147 148 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 149 150 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 FIGURES 86-94. LM. Scale = 10 nm. Figure 86. Thalassiosira hendeyi Hasle and G. Fryxell. Linear arranged areolae, two labiate processes (arrow), wavy margin. Figure 87. T. nodulolineata (Hendey) Hasle and G. Fryxell. Linear arranged areolae, six symmetrical areolae surrounding central area, marginal strutted processes, single labiate process (arrow). Figure 88. T. MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 151 labiate process (Fig. 77, arrow; Fig. 78); ring of strutted pro- cesses in small arcs (Fig. 74, 75, 79) in each fascicle located ca. '/z the distance between the margin and the central areolae (Fig. 74). Characteristic arc-shaped grouping of strutted pro- cesses regularly arranged in each fascicle diagnostic for this species. DISTRIBUTION.— Coastal, prefers cold to tem- perate waters, usually associated with other ma- rine forms. Central San Francisco Bay to Golden Gate. DISCUSSION.— Thalassiosira stellaris, a lightly silicified species, was found in culture from a sample taken north of Golden Gate Bridge. Tha- lassiosira endoseriata was found only once. As previously mentioned, Thalassiosira lacustris displays a tangentially undulated valve face (Fig. 68), an aid in identification. The largest concen- tration of viable cells of T. lacustris was found in the brackish waters of Suisun Slough (Mahood 1981). Thalassiosira anguste-lineata, with its fasciculated areolar pattern and distinguishing arrangement of strutted processes in each fascicle (Fig. 74, 79, 93) is easily distinguished from other fasciculated species seen within the Bay in cleaned material in permanent mounts. However, the strutted processes on the valve face are not al- ways easily seen in an uncleaned sample, and lack of resolution may result in some confusion. When cells are united in chains, however, several threads can be seen to extend from one cell to the next, one from each cluster of strutted pro- cesses instead of the single central thread com- monly seen in the genus. GROUP 5.— Thalassiosira, otherwise dissimilar species that have radial areolae patterns and mul- tiple central processes. Thalassiosira rotula Meunier, 1910 (Figures 80-85, 94) DETAILED DESCRIPTION. Fryxell (19756); Syvertsen ( 1 977). — Cell diameter 8-61 urn (40-61 urn, Gran and Angst 1931); areolae very fine, only clearly seen in the central area (Fig. 84); cluster of strutted processes in center (Fig. 80, external; Fig. 81, internal), with scattered processes across entire valve face (Fig. 83, 84); single labiate process (Fig. 84, arrow; Fig. 85, arrow); valve weakly silicified (Fig. 82). Usually observed in chains (Fig. 94). DISTRIBUTION.— In our samples T. rotula re- stricted to the more saline portion of the central bay and Golden Gate with other marine forms. Thalassiosira incerta Makarova, 1961 (Not figured) DETAILED DESCRIPTION. Hasle (1978a).— Cell diameter 13- 38 tan; areolae 8-16 in 10 ^m; three to four strutted processes in 10 fim at margin; four to six strutted processes surrounding central areolae. DISTRIBUTION.— Fresh to brackish water, rare in our samples. DISCUSSION.— Both brackish and marine members of this group show distinctive char- acteristics that clearly differentiate them from other Thalassiosira species of the bay. Thalas- siosira rotula, first mentioned by Whedon (1939), is usually found in the central bay (Wong and Cloern 1981). This species has been reported from similar estuarine environments (Marshal 1980). Our experience has been to associate T. rotula with other coastal or marine forms. Its charac- teristic chain formation with coin-shaped cells connected by a thick central thread, narrow gir- dle bands, delicate structure, and scattered strut- ted processes on the valve face aid in the iden- tification. Thalassiosira incerta, T. decipiens, and T. vi- surgis are all characterized by small size, eccen- tric to linear areolar patterns, and an accumula- tion of detritus around the margin of the living cell. This group has proven to be the most dif- ficult to differentiate by light microscopy. All three species have similar diameters and general char- acteristics (Table 2) and can only be differen- tiated after the sample has been cleaned of or- ganic material. In cleaned material T. incerta possesses characteristics that distinguish it from lacustris (Grunow) Hasle. Tangential undulations, radial arrangement of areolae. Figure 89. T. lundiana G. Fryxell. Weakly silicified, large irregular occluded processes at margin, irregularly arranged strutted processes on valve. Figure 90. T. nodulolineata (Hendey) Hasle and G. Fryxell. Detail of central area, strutted processes in central areola. Figure 91. T. endoseriata Hasle and G. Fryxell. Fasciculated areolae, ring of strutted processes near center of valve. Figure 92. T. punctigera (Castracane) Hasle. Regular marginal strutted processes, weakly silicified. Figure 93. T. anguste-lineata (A. Schmidt) G. Fryxell and Hasle. Fasci- culated, arc of strutted processes in each fascicle. Figure 94. T. rotula Meunier. Central strutted process, random valve strutted processes, forming chains. 152 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 FIGURES 95-106. LM. Scale = 10 nm. Figure 95. Thalassiosira visurgis Hustedt. Focused on center, two labiate processes (arrows), radial pattern of areolae. Figure 96. T. visurgis Hustedt. Same valve as Figure 95, focused on margin. Figure 97. T. decipiens (Grunow) Jorgensen. Focused on center, regular marginal strutted processes, one labiate process (arrow). Figure 98. T. decipiens (Grunow) Jorgensen. Same valve as Figure 97, focused on margin, uniform areolae across valve. Figures 99-101. T. wongii Mahood sp. nov. Fasciculated, two circles of strutted processes on valve, one labiate process (arrow), raised central MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 153 TABLE 2. THALASSIOSIRA SPECIES FROM SAN FRANCISCO BAY AREA. Measurements from the literature and our observations. Marginal Labiate strutted Central Diameter Areolae pro- processes strutted Distinctive Species (Mm) in 10 /im cesses in 10 nm processes characteristics Group 1 —Species with a linear areolar pattern T. hendeyi 38-120 5-6 2 5-6 1 wavy margin T. nodulolineata 26-58 3.5-7.0(9.0) 1 3-5 5-6 strutted processes in central areola T. simonsenii 30-57 4.0-5.5 2 5-6 1 two marginal rings of strutted processes T. tenera 10-29 9-16 1 3-5 1 flattened marginal processes Group 2— Species with eccentric patterns T. eccentrica 12-101 5-10 T. oestrupii var. venrickae T. wongii 5.5-39.0 27-51 6-9 9-11 1 irregular 1 seven areolae around central 10-20 areola 1 1-2 1 strutted processes project in- ward 1 3-5 4-5 fasciculated central ring of strutted processes Group 3— Species with 1 central process and 1 marginal ring of strutted processes T. lundiana T. punctigera T. nordenskioeldii T. pacifica T. minuscula T. visurgis T. decipiens 7-43 43-145 10-50 7-46 10-20 9-18 9-29 13-14 8-12 1 5-10 1 fasciculated, weakly silicified 1 4-5 1 fasciculated, regular marginal strutted processes 1 3 1 radial, marginal strutted pro- cesses back from margin 1 4-7 1 fasciculated, raised central process 1 4-5 labiate process away from margin 2 4-5 1 irregular radial pattern 1 4-6 1 eccentric radial pattern Group 4— Species with no central strutted processes but a modified ring of strutted processes T. stellaris T. lacustris T. endoseriata T. anguste-lineata 6-20 20-75 20-60 17-78 30 10-14 11-18 8-18 Group 5— Dissimilar species with radial patterns T.rotula 8-61 20 T. incerta 13-38 8-16 1 3-5 (6) 2-7 processes in ring, fascicu- lated 1 5-7 tangentially undulated 1 5-6 4-14 processes in ring, fasci- culated 1 3-4 ring of arcs on face of valve, fasciculated 1 4-7 cluster cluster of central strutted pro- cesses 1 3-4 4-6 central processes around cen- tral areola radial pattern T. decipiens and T. visurgis. The central areola of T. incerta is surrounded by five to six strutted processes similar to the arrangement seen in T. nodulolineata (Fig. 14). Unlike those in T. nod- ulolineata, the central strutted processes in T. incerta are extremely small and are just visible under the light microscope. Thalassiosira incerta was rare in our collections and was only found strutted process, pronounced and regular marginal strutted processes. Figure 102. T. eccentrica (Ehrenb.) Cleve. Eccentric areolar pattern, irregular marginal spines. Figures 103, 104. T. tenera Proschkina-Lavrenko. Linear arrangement of areolae, robust flattened marginal strutted processes, central area raised around the central areola. Figure 105. T. pacifica Gran and Angst. Fasciculated pronounced regular marginal strutted processes, distinctive central process. Figure 106. T. nordenskioeldii Cleve. Large marginal strutted processes set back from margin, raised central process. 154 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 8 from the Suisun Bay material. Hasle (1978a) has also reported T. incerta from waters of the San Joaquin delta. CONCLUSIONS The 20 species presented require additional study to delineate more clearly their ecological variables. Other than salinity, little is known of their habitat requirements. Of the species pre- sented, T. decipiens, T. nodulolineata, T. eccen- trica, T. hendeyi, and T. wongii appear to be best suited as indicators of salinity. ACKNOWLEDGMENTS Samples were collected by R. L. J. Wong, C. A. McNeil, P. Wares, and L. L. Lack. T. P. Wat- kins isolated clonal cultures from the study area and provided technical assistance. P. A. Fryxell kindly provided the Latin description of Tha- lassiosira wongii. Expert assistance was provided by the Texas A&M University Electron Micros- copy Center. M. M. Hanna and the California Academy of Sciences were helpful with literature resources. We would like to express our thanks to M. A. Hoban of the California Academy of Sciences for his careful and constructive editing. G. R. Hasle and L. K. Medlin participated in helpful discussions, and the Texas A&M Uni- versity Department of Oceanography provided partial support for maintenance of phytoplank- ton cultures. Partial support for this work was provided by National Science Foundation Grants DEB 79-23159 and DEB 77-15908 and the Cal- ifornia Academy of Sciences, G Dallas Hanna Memorial Publication Fund, and is gratefully ac- knowledged. Figure 1 was prepared by J. F. DeMouthe. LITERATURE CITED ANONYMOUS. 1975. Proposal for a standardization of diatom terminology and diagnoses. Nova Hedwigia, Beih. 53:323- 354. ARTHUR, J. F. AND M. D. BALL. 1980. The significance of the entrapment zone location to the phytoplankton standing crop in the S.F. Bay-Delta Estuary. U.S. Department of Interior, Water and Power Resources Service, p. 89. CLEVE, P. T. 1873. On diatoms from the Arctic Sea. Bih. K. Svenska Vetensk.-Akad. Handl. 1(1 3): 1-28. . 1904. Plankton table for the North Sea. Bull. Cons. Explor. Mer. 1903-1904:216. CUPP, E. E. 1 943. Marine plankton diatoms of the west coast of North America. Bull. Scripps Inst. Oceanogr. Tech. Ser. 5(l):l-238. FRYXELL, G. A. 1975a. Three new species of Thalassiosira: with observations on the occluded process, a newly observed structure of diatom valves. Nova Hedwigia, Beih. 53:57-75. . 1975ft. Morphology, taxonomy, and distribution of selected species of the diatom genus Thalassiosira Cleve in the Gulf of Mexico and antarctic waters. Ph.D. Dissertation, Texas A&M University, College Station, Texas. . 1978. The diatom genus Thalassiosira: T. licea sp. nov. and T. angstii (Gran) Makarova, species with occluded processes. Bot. Mar. 21:131-141. FRYXELL, G. A. AND G. R. HASLE. 1972. Thalassiosira ec- centrica (Ehrenb.) Cleve, T. symmetrica sp. nov. J. Phycol. 8:297-317. . 1977. The genus Thalassiosira: some species with a modified ring of central strutted processes. Nova Hedwigia, Beih. 54:67-98. . 1980. The marine diatom Thalassiosira oestrupii: structure, taxonomy, and distribution. Am. J. Bot. 67(5): 804-814. GRAN, H. H. AND E. E. ANGST. 1931. Plankton diatoms of Puget Sound. Puget Sound Mar. (Biol.) Stn. 1 929-3 17:417- 519. HANNA, G D. 1930. Hyrax, a new mounting medium for diatoms. J. Microsc. (Oxf.), Series 3, 50(4):424-426. HARRIS, H. S., D. L. FEURSTEIN, AND E. H. PEARSON. 1961. A pilot study of physical, chemical, and biological charac- teristics of waters and sediments of San Francisco Bay, SERL. College of Engineering and Public Health, University of Cal- ifornia Berkeley, pp. 9-14. HASLE, G. R. 1976. Examination of diatom type material: Nitzschia delicatissima Cleve, Thalassiosira minuscula Krasske, and Cyclotella nana Hustedt. Br. Phycol. J. 11: 101-110. . 1978a. Some freshwater and brackish water species of the diatom genus Thalassiosira Cleve. Phycologia, 17(3): 263-292. . 1978ft. Some Thalassiosira species with one central process (Bacillariophyceae). Norw. J. Bot. 25:77-1 10. . 1979. Thalassiosira decipiens (Grun.) Jerg. (Bacil- lariophyceae). Bacillaria 2:85-108. . 1983. Thalassiosira punctigera (Castr.) comb, nov., a widely distributed marine planktonic diatom. Nord. J. Bot. 3:593-608. HASLE, G. R. AND G. A. FRYXELL. 1977. The genus Thalas- siosira: some species with a linear array. Nova Hedwigia, Beih. 54:15-66. HEDGPETH, J. W. 1979. San Francisco Bay: the unsuspected estuary. San Francisco Bay, the urbanized estuary. Pacific Division of the American Association for the Advancement of Science. California Academy of Sciences, Golden Gate Park, San Francisco, California. P. 493. HUSTEDT, F. 1 930. Die Kieselalgen Deutschlands, osterreichs und der Schweiz unter Berucksichtigung der ubrigen Lander Europas sowie der angrenzenden Meeresgebiete in Krypto- gamen-Flora von Deutschland, Osterreich und der Schweiz. (ed.) L. Rabenhorst. Pp. 609-920. . 1957. Die Diatomeen Flora des Fluss-systems der Weser im Gebiet Hansestadt Bremen. Abh. naturw. ver. Bremen, 34(3). J0RGENSEN, E. 1905. Protist plankton. Diatoms in bottom samples. Pp. 1-254 in Hydrographical and biological inves- tigations in Norwegian fiords, O. Nordgaard, ed. Bergens Museum Strift, 1905. John Grieg, Bergen. MAHOOD, FRYXELL, AND McMILLAN: THALASSIOSIRA 155 KRASSKE, G. 1941. Die Kieselalgen des chilenischen Kus- tenplankton. Arch. Hydrobiol. 38:260-286. MAHOOD, A. D. 1981. Phytoplankton analysis of Suisun Slough, Fai rfield-Suisu n Sewer District, Fairfield, California, Unpublished report to the Fairfield-Suisun Sewer District. MAKAROVA, I. V. 1961. Diatomaceae novae familiae Cos- d nodiscaceae e Mari Caspico borealis. Notulae Systematicae e Sectione Cryptogamica Instituti Botanic! Nomine V. L. Komarovii Academiae Scientiarum U.S.S.R. 14:49-52. MARSHAL, H. G. 1980. Seasonal phytoplankton composition in Lower Chesapeake Bay and Old Plantation Creek, Cape Charles, VA. Estuaries 3(3):207-216. MEUNIER, A. 1910. Microplancton des Mers de Barents et de Kava. Due d'Orleans, Campagne Archique 1907. MULLER-MELCHERS, F. C. 1953. New and little known dia- toms from Uruguay and South Atlantic coast. Comun. Bot. Mus. Hist. Nat. Montev. 3(30): 1-11. PROSCHKINA-LAVRENKO, A. I. 1961. Diatomeae novae e Mari Nigro (Ponto Exino) et Azoviano (Maeotico). Notulae Sys- tematicae e Sectione Cryptogamica Instituti Botanici Nom- ine V. L. Komarovii Academiae Scientiarum U.S.S.R. 14: 33-39. Ross, R., E. J. Cox, N. I. KARAYEVA, D. G. MANN, T. B. B. PADDOCK, R. SIMONSEN, AND P. A. SIMS. 1979. An am- mended terminology for the siliceous components of the diatom cell. Nova Hedwigia, Beih. 64:513-533. ROUND, F. E. AND P. A. SIMS. 1981. The distribution of diatom genera in marine and freshwater environments and some evolutionary considerations. Proceedings of the sixth symposium on recent and fossil diatoms. Budapest Sept. 1- 5, 1980. Otto Koeltz, Konenstein, Germany:30 1-320. Srrrs, R. M. AND A. W. KNIGHT. 1979. Plankton ecology in the Sacramento-San Joaquin estuary. Water Science and Engineering No. 4509. University of California, Davis. P. 145. STORRS, P. M., E. A. PEARSON, AND F. F. SELLECK. 1966. A comprehensive study of San Francisco Bay. Final Report, V. Summary of physical, chemical, and biological water and sediment data. University of California, Berkeley SERL, 67(2): 1-140. SYVERTSEN, E. E. 1977. Thalassiosira rotula and T. gravida: ecology and morphology. Nova Hedwigia, Beih. 54:99-1 12. VAN DER WERFF, A. 1955. A method of concentrating and cleaning diatoms and other organisms. Int. Assoc. Theor. App. Limno. Proc. 12:276-277. WHEDON, W. R. 1939. A three year survey of the phyto- plankton in the region of San Francisco, California. Int. Rev. Hydrobiol. Pp. 459-476. WONG, R. L. J. AND J. E. CLOERN. 1981. Plankton studies in San Francisco Bay. II Phytoplankton and Species com- position. July 1977-December 1979. U.S. Geological Sur- vey. Open File Report 81-214:103. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94 1 1 8 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 9, pp. 157-224, 33 figs., 7 tables. May 6, 1986 SYSTEMATIC RELATIONSHIPS AND ONTOGENY OF THE SCULPINS ARTEDIUS, CLINOCOTTUS, AND OLIGOCOTTUS (COTTIDAE: SCORPAENIFORMES) By Betsy B. Washington1 Gulf Coast Research Laboratory, East Beach Drive, Ocean Springs, Mississippi 39564 ABSTRACT: Using the methods of phylogenetic analysis proposed by Hennig (1966), characters of the larvae of 13 species of Artedius, Clinocottus, Oligocottus are examined in terms of synapomorphic states. Number and pattern of preopercular spines, gut di verticula, body shape, and a bubble of skin at the nape are identified as synapomorphic characters useful in systematic analysis of this group. The synapomorphic character, multiple preopercular spines, provides strong evidence that Clinocottus acuticeps, C. analis, C. embryum, C. globiceps, C. recalvus, Oligocottus maculosus, O. snyderi, Artedius fenestralis, A. harringtoni, A. lateralis, and A. Type 3 form a monophyletic group within the Cottidae. Within this group, the species of Clinocottus and Oligocottus are very closely related; each genus, however, appears to be monophyletic. Larvae of all species of Clinocottus possess the synapomorphy, auxiliary preopercular spines. Larval Oligocottus maculosus and O. snyderi share two derived characters, dorsal gut bumps and a bubble of skin at the nape. Artedius fenestralis, A. harringtoni, A. lateralis, and A. Type 3 also form a monophyletic group closely related to Clinocottus and Oligocottus on the basis of a unique multiple preoper- cular spine pattern. Synapomorphic characters of the larvae provide strong evidence that A. creaseri and A. meanyi are more closely related to Icelinus than to species of Clinocottus, Oligocottus maculosus, O. snyderi, Artedius fenes- tralis, A. harringtoni, A. lateralis, and A. Type 3. Characters of the larvae strongly indicate that the genus Artedius as defined by Bolin (1934, 1947) is not monophyletic and that A. creaseri and A. meanyi should be placed separately from the other species of Artedius. Clarification of the exact position of these two species in relation to the Artedius-Clinocottus-Oligocottus group and other cottids must await ^examination of characters of adults. Complete, identified, developmental series of larval cottids Artedius fenestralis, A. creaseri, A. meanyi, Oligocottus snyderi, Clinocottus embryum, and C. globiceps are described for the first time. Partial devel- opmental series of two species, Artedius Type 3 and Clinocottus analis, are also described and illustrated for the first time. In addition, four species, Artedius harringtoni, A. lateralis, Oligocottus maculosus, and Clino- cottus acuticeps are redescribed providing new and comparative information on larval development. INTRODUCTION era and 300 species (Nelson 1 976). Most of these TU r- **-j • j- species are marine and generally distributed in The Cottidae are a large, morphologically di- r „ T ,. c ., f~ , f , ,-J coastal waters of all oceans except the Indian verse family of fishes composed of nearly 67 gen- ~ /-, ^-j • XT _L Ocean. Cottids are most speciose in the North Pacific where 90 species distributed in 40 genera • Current address: National Marine Fisheries Service, Sys- are ™V°^ to OCCUr between Baja, California tematics Laboratory, National Museum of Natural History, and the Aleutian Islands, Alaska. Sixteen of these Washington, D.C. 20560. species belonging to the genera Artedius, Clino- [157] 158 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 coitus, and Oligocottus are common intertidal and subtidal inhabitants of the northeast Pacific coast (Table 1). Despite their abundance the sys- tematic status and early life history of members of Artedius, Clinocottus, and Oligocottus have received little study. The few systematic studies that have ad- dressed these genera have yielded contradictory results. Most workers have placed members of Artedius, Clinocottus, and Oligocottus in the family Cottidae (Jordan and Evermann 1898; Regan 1913;Bolin 1934, 1944, 1947; Berg 1940; Taranets 1941; Howe and Richardson 1978); Jordan (1923), however, placed the genus Arte- dius in the family Icelidae. Bolin (1934, 1947), in a review of marine cottids of California, pro- posed that Artedius, Clinocottus, and Oligocottus were closely related genera that evolved from an evolutionary line of cottids tending towards a reduction of gills, pelvic fin rays, preopercular spines, and squamation. In contrast, Taranets (1941) separated members of these genera into two subfamilies. He placed Clinocottus, Oligo- cottus, and five species of Artedius in the subfam- ily Oligocottinae, and placed Artedius creoseri and A. meanyi in the Icelinae. Howe and Rich- ardson (1978) suggested that Artedius, Clinocot- tus, and Oligocottus form a closely related group, Oligocottus and Clinocottus being most closely related. However, they did not discuss characters that led to their proposal. Much of the confusion in the systematic treat- ment of these genera is due to the use of primitive and reductive characters in classifications, and the failure of past workers to define the genera on the basis of unique, derived characters. Char- acters used in past studies have not been exclu- sive to this group and are present in other cottid genera. The usefulness of characters of larvae and ju- veniles in elucidating systematic relationships has been demonstrated in several groups (Bertelsen 1951; Moser and Ahlstrom 1970, 1972, 1974; Johnson 1974; Okiyama 1974; Kendall 1979; Richardson 1 98 1 ; and Moser et al. 1984). Results of these studies have indicated that ontogenetic characters provide an independent set of char- acters with which to evaluate phylogenetic re- lationships. Characters of larvae have been par- ticularly helpful in groups in which characters of adults have been reductive or generalized (Moser and Ahlstrom 1972, 1974). Although larvae of most species of Artedius, Clinocottus, and Oligocottus are frequently col- lected in nearshore plankton samples in the northeast Pacific, until recently larvae of few species have been described. Of the 1 6 nominal species, developmental series of identified larvae have only been described for A. harringtoni, A. lateralis, C. acuticeps, C. recalvus, and O. mac- ulosus. Other forms that belong to this group based on larval morphology, but not identified to species, were described by Richardson and Washington (1980) as Artedius Type 2, Cottidae Type 1, Type 2, and Type 3. Previous descrip- tions of A. lateralis and O. maculosus are inad- equate for specific identification. Descriptions by Richardson and Washington (1980) were based on incomplete developmental series and too few specimens for specific and/or generic identifi- cation. The objectives of this study were: 1) to eval- uate phylogenetic relationships of Artedius, Cli- nocottus, and Oligocottus within the Cottidae following the phylogenetic methodology of Hen- nig (1966) and 2) to describe the ontogeny of larvae and juveniles of as many species of Ar- tedius, Clinocottus, and Oligocottus as possible. METHODS AND MATERIALS Systematic Procedures The investigation of systematic relationships in this study follows the phylogenetic approach first put forth by Hennig (1966) and modified and debated by Brundin (1966, 1968), Cracraft (1 974), Sneath and Sokal (1 973), Ashlock (1 974), Mayr (1974), and others. This methodology is considered best suited to the objectives of this study because the meth- odology is a phylogenetic approach, and is well defined in regard to character evolution. Other approaches currently used are not as well suited to the purpose of this study. Numerical taxon- omy, best described by Sneath and Sokal (1973), is a phenetic approach in which taxa are clustered by overall similarity. Evolutionary systematics described by Mayr (1969) and Simpson (1961) combines both phyletic and phenetic informa- tion; however, a well-defined, repeatable meth- odology has not been incorporated into this ap- proach. The basic tenet of Hennig's approach is that the shared possession of derived character states is the only valid criterion for establishing phy- logenetic relationships. Hennig (1966) defines a WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 159 Branchio- ertebrae2 stegal rays PT m rT m ^ "^ f^ rr> ^ f> ff r^-i m f*> ^ fT f> m t?> rr> rs" rT f> <»1 S NO 1 — NO ^ ^ f> m r^1 fT f> r^ NO NO <*1 r^ rr> r^i NO NO NO" r^t f> f^ i/^ m V ro in Pelvic fin Principal rays caudal rays Total v m 1 VO NO NO NO NO NO NO S SI*" J, _!. NO NO NO NO NO NO 'T CN NO NO 4 NO 1 1 NO t- NO £ « NO t^ iXiX iXiX NT) ^ ^ «0 W^ i 4 «k4 «A4 44 A4 d. «J» A ri. cA J> r!) 18 R f ^ I- NO NO W> >r> NO •^ NO NO r^ 00 NO ON o Jo u2 S 0 NO rf oo oo r- r~ NO t- 00 r~ t- NO 00 ON P~ o « 1 Q •A r!) NO •A .A 4 4 A 4 4 A 4 A NO tA rl a M 1 tion of Howe illinus or A. n this study. *^ « X C- X X X X X % > > .-3 ? i fa 0 M 7 X*, 22 $x 2* 22 2 2 ^™^ x j, 3 X I ^j*| §s^ > | i creased ] fenestralis V §1 §> Si 1 " ^ 2 ™ £ ^ ^ r -T ^ ^ U 0 U 0 0 0 0 0 * 160 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 monophyletic group as a group in which all members are descended from a single stem. The common possession of one or more derived char- acters is the only conclusive evidence that a group is monophyletic. Shared, plesiomorphic char- acters are not used because primitive character states inherited from an ancestral taxon may re- main unchanged in various divergent lineages and may not be evidence of close relationship. Monophyletic groups that arose from a common stem by the same splitting process are called sis- ter groups. Every monophyletic group, together with its sister group, constitutes a monophyletic group of higher taxonomic rank. Since only derived character states are used in determining phylogenetic relationships, the po- larity of character states was determined through outgroup comparisons. The outgroup taxa ex- amined in this study included larvae of seven different cottid genera: Scorpaenichthys mar- moratus, Hemilepidotus spinosus, Leptocottus armatus, Enophrys bison, Myoxocephalus sp., Icelinus sp., and Radulinus asprellus. Members of these genera are quite varied and represent several divergent lineages within the Cottidae (Bolin 1934, 1947; Taranets 1941; Howe and Richardson 1978). Larvae of several other scor- paeniform families also were examined for out- group comparisons. These taxa included: Se- bastes flavidus (Scorpaenidae); Hexagrammos sp. (Hexagrammidae), Cyclopteridae Type 1 (Cy- clopteridae), and Stellerina xyosterna (Agoni- dae). SELECTION OF CHARACTERS FOR ANALYSES.— A variety of characters were examined in Arte- dius, Clinocottus, and Oligocottus larvae includ- ing meristics, morphology, pigmentation, spi- nation, and developmental osteology. However, of the 50 characters initially examined, many were deleted from final analysis. The criteria used in deleting characters are as follows: (1) Characters that exhibit a large amount of variability were deleted from analysis. High- ly variable characters are poor indicators of phylogenetic relationships (Bolin 1947; Simpson 1961; Mayr 1969). Examples of variable characters are head pigmentation and number of posttemporal-supracleithral spines. (2) Derived character states found in only one species were deleted. Again characters of this nature are of no value in determining intra- group relationships. Hindgut diverticula of Clinocottus acuticeps are an example of a specific character. (3) Characters in which the sequence of change or the primitive and derived states could not be identified were deleted from the analysis. Many of the morphometric and pigmenta- tion characters fell into this category. Taxonomic Procedures Larval descriptions are based on both labo- ratory-reared larvae and field-collected speci- mens. Egg masses were spawned from ripe Cli- nocottus globiceps, Oligocottus maculosus, and O. snyderi collected from tidepools along the central Oregon coast during winter-spring 1979 and 1980. Larvae were maintained at 12-13°C. In addition to reared specimens, developmental series were put together with larvae obtained with 70 cm bongo nets and neuston nets off the coast of Oregon between 1969 and 1978. Samples were taken in all months of the year from an area concentrated along an east-west transect off Newport, Oregon (lat. 44°39.1'N). Specimens were also obtained from estuarine and coastal collections of the Southwest Fisheries Center, La Jolla Laboratory, National Marine Fisheries Ser- vice; Scripps Institution of Oceanography; Los Angeles County Museum of Natural History; Marine Ecological Consultants; California Acad- emy of Sciences; Humboldt State University; Northwest Fisheries Center, Seattle Laboratory, National Marine Fisheries Service; and Univer- sity of Washington. Reared specimens of Arte- dius lateralis from College of Fisheries, Univer- sity of British Columbia and Oligocottus maculosus and Clinocottus acuticeps from Van- couver Public Aquarium were also utilized. Transforming and juvenile specimens were col- lected monthly from 1977 to 1980 from tide- pools along the central Oregon coast. All specimens were preserved in 5 or 1 0% buff- ered formalin and some material was subse- quently transferred to 36 or 40% isopropyl al- cohol. Developmental series of larvae and juveniles were assembled for 12 of the 16 species of Ar- tedius, Clinocottus, and Oligocottus. The num- ber of specimens examined in each series varies from 1 1-38 according to availability of material. Developmental series were formed utilizing field- caught larvae, except as noted below, because of WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 161 the large amounts of variation in morphology and pigmentation in laboratory-reared larvae. Newly hatched, reared larvae are included in se- ries of Artedins fenestralis, A. later alis, Clino- cottus acuticeps, C. globiceps, Oligocottus mac- ulosus, and O. snyderi. In addition, reared larvae were used to supplement incomplete develop- mental series of field-caught larvae oiClinocottus globiceps and Oligocottus maculosus. Marked differences between developmental series based on field specimens and laboratory-reared speci- mens are noted in the descriptions. Developmental stages follow the terminology of Ahlstrom et al. (1976), except that the tran- sitional period between the larval and juvenile stages is marked by an increase in pigmentation particularly over the head and in saddles along the dorsum, a reduction in the size and number of preopercular spines, an ossification of the pel- vic fin spine and rays, and the formation of scales. Specimens are referred to as juveniles when they settle from the plankton and assume a benthic existence. MORPHOMETRICS.— Measurements of selected body parts were made to the nearest 0. 1 or 0.0 1 mm using an ocular micrometer in a stereo- microscope. Measurements were made following the definitions of Richardson and Laroche (1979) except as follows: body depth at anus = vertical distance from the dorsal to ventral body margin at the anus, snout to pelvic fin origin = horizon- tal distance from the tip of the snout to a vertical through the origin of the pelvic fin, and origin of pelvic fin to anus = horizontal distance from a vertical through the origin of the pelvic fin to the anus. Head length is abbreviated as HL. Detailed tables documenting the development of meristic elements for larvae of Artedius, Clinocottus, and Oligocottus are presented by Washington (1 98 1). All body lengths given in this study refer to either notochord length (NL), which is defined as snout tip to notochord tip preceding devel- opment of the caudal fin; standard length (SL), which is defined as snout tip to the posterior margin of the hypural plates; or total length (TL), which is defined as snout tip to the posteriormost margin of the caudal fin. Unless otherwise in- dicated, all lengths given are standard length. MERISTICS.— Following the methods of Din- gerkus and Uhler (1977), several larvae were cleared and stained with Alcian Blue and Ali- zarin Red S for each species when specimens were available in sufficient numbers. Counts were made of dorsal fin spines and rays, anal fin rays, pelvic fin spines and rays, principal caudal rays, branchiostegal rays, preopercular spines, and vertebrae. Vertebral counts always included the urostyle. All meristic elements were counted if they absorbed Alizarin stain. Principal caudal rays are defined as the number of caudal fin rays that articulate with the upper and lower hypural plates. Counts of meristic elements were also made on unstained larvae from the developmental se- ries used in the morphometric examination. All fin rays and spines, branchiostegal rays, pre- opercular spines, and myomeres were counted when visible under magnification. In this study, all fin rays and spines were counted, regardless of whether they arose from the same pterygio- phore. (For detailed meristic and spination tables of developmental series of known Artedius, Cli- nocottus, and Oligocottus larvae see Washington 1981). SPINATION.— Spine terminology generally fol- lows Richardson and Washington ( 1 980) in which spines are named for the bones from which they originate. TAXONOMIC TERMINOLOGY.— Results of this study do not agree with previously recognized limits of the genus Artedius. In order to avoid confusion, species in this group are designated as Artedius Group A including A. fenestralis, A. harringtoni, A. lateralis, and A. Type 3 (either A. corallinus or A. notospilotus) or Artedius Group B including A. creaseri and A. meanyi. Larvae referred to as Artedius Type 3 are either A. cor- allinus or A. notospilotus. Positive identification is not possible at this time because of the lack of late-stage larval specimens. (See descriptions for discussion of identification.) RESULTS Description of Characters Considered PREOPERCULAR SPINATION.— The number of preopercular spines is a relatively stable, con- servative character in larval cottids. Most cottid larvae (22 of 28 known genera) possess four ap- proximately equal-sized spines situated along the posterior margin of the preopercle. Generally the dorsalmost spine increases in size with devel- opment while the lower three spines are reduced or lost. A modification of this basic preopercular pat- tern is found in larvae of several species of Ice- 162 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 H FIGURE 1. Multiple preopercular spines in larval Artedius, Clinocottus, and Oligocottus. A) Artedins harringtoni, B) A. fenestralis, Q A. lateralis, D) A. Type 3, E) Clinocottus acuticeps, F) C. embryum, G) C. globiceps, H) C. analis. I) Oligocottus snyderi, J) 0. maculosus. linus and Myoxocephalus. These larvae possess an additional small, auxiliary spine situated on the inner shelf of the preopercle anterior to the bases of the four principal preopercular spines. A third pattern of preopercular spination is found in larvae of Clinocottus, Oligocottus, and Artedius Group A (Fig. 1). These larvae possess 5-24 small spines situated along the posterior margin of the preopercle. Two basic patterns of multiple preopercular spines occur in larvae of this group. In four species of Artedius (Group A), the dorsalmost, middle, and ventralmost spines become enlarged relative to the other preoper- cular spines. During transformation, the dorsal- most spine continues to increase in size, while the middle spines (7-9), midventral spines (11- 14), and ventralmost spines each fuse together, forming three large bumps on the preopercular margin. The other preopercular spines gradually disappear. In larvae of Clinocottus and Oligocottus the dorsalmost spine increases in size relative to the other preopercular spines. During transforma- tion the lower spines regress and disappear while the dorsalmost spine remains prominent. Outgroup comparisons with larvae of closely related Scorpaeniformes indicate the presence of five or fewer approximately equal-sized pre- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 163 FIGURE 2. Preopercular spines of larval Artedius meanyi and A. creaseri. opercular spines. Sebastes and Stellerina larvae possess five and four spines, respectively; hexa- grammid larvae possess five to six, three, or no spines, and cyclopterid larvae have lost all pre- opercular spines. The cottid taxa that possess four equal-sized preopercular spines also tend to have many other primitive character states. The presence of four equal-sized spines is considered the plesiomorphic state for preopercular spines in cottid larvae. The modified pattern of spines found in larval Icelinus and Myoxocephalus could easily be de- rived from the basic pattern of four preopercular spines. In fact, larvae of several species of Ice- linus and Myoxocephalus possess only four pre- opercular spines. The presence of an auxiliary spine on the preopercle probably represents an intermediate character state leading toward mul- tiple preopercular spines. The multiple preopercular spines of larvae of Artedius (Group A), Clinocottus, and Oligocottus are unique to this group. Multiple preopercular spines are not present in any other known cottid or scorpaeniform larvae. Multiple preopercular spines are derived character states indicative of the monophyletic origin of this group. BASAL PREOPERCULAR SPINE.— Larvae of Ar- tedius meanyi and creaseri, larvae of at least two species of Icelinus, and larvae of Myoxocephalus possess small projections or spines on the base of each of the four main preopercular spines (Fig. 2). These basal spines project out at 90° angles to the axis of the main preopercular spines. The basal spines are most pronounced in early post- flexion larvae. With development, four bony ridges form on the inner shelf of the preopercle, parallel to each basal spine. These bony ridges grow toward the basal spines and gradually fuse with them, forming bony arches over the forming lateral line canal of the preopercle. These basal spines are not present in other cottid or scor- paeniform larvae examined and probably are a derived character state. INNER SHELF PREOPERCULAR SPINES.— Larval Clinocottus possess one or two tiny spines on the inner shelf margin of the preopercle. Clinocottus acuticeps larvae have only one inner shelf spine. All other Clinocottus larvae examined have two. These spines are transient features which form in postflexion larvae and are lost before trans- formation. They appear to be unique to this group and, as such, to be derived character states. NAPE BUBBLE.— Larvae of Oligocottus macu- losus and O. snyderi possess a distinctive bubble of skin in the nape region just anterior to the origin of the dorsal fin (Fig. 3). This bubble is present at hatching and persists for two or three weeks (to about the beginning of flexion of the notochord). No other known cottid larvae pos- sess a bubble of skin at the nape; accordingly, this bubble is probably a derived character unique to these two species. (Larvae of O. rimensis and O. rubellio are unidentified and it is not known if they also possess this character.) GUT DIVERTICULA.— Long protrusions or di- verticula extend dorsolaterally from either side of the abdominal cavity in larvae of Artedius fenestralis, A. lateralis, and A. Type 3 (Fig. 4). These di verticula are present at hatching and per- sist throughout larval development. Newly 164 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 FIGURE 3. Nape bubble of larval Oligocottus snyderi. hatched larvae of Oligocottus maculosus and O. snyderi possess similar but less pronounced bumps or protrusions on either side of the dorsal surface of the abdominal cavity. These bumps are present at hatching, but they disappear after two to three weeks at about the onset of noto- chord flexion. The diverticula of Artedius fenes- tralis, lateralis, and Type 3 and the smaller pro- trusions of Oligocottus appear to be homologous structures, with the smaller bumps constituting an intermediate form. These diverticula are unique, derived characters not known in other cottid larvae. Larvae of C. acuticeps also possess long di- verticula which extend posteriorly on either side of the anus. Larvae of C. globiceps, C. embryum, and C. analis have bulges on either side of the anus. Although these bulges appear to form an intermediate state in the evolution of hindgut diverticula, histological sections of the guts of larval Clinocottus yielded inconclusive results. Larval C. analis, C. embryum, and C. globiceps possess an enlarged coelom on either side of the hindgut, but no distinct diverticula. Hindgut di- verticula appear to be unique to C. acuticeps. PARIETAL AND NUCHAL SPINES.— Most larval cottids develop two spines, a large anterior pa- rietal and a smaller posterior nuchal spine at the posterior edge of the parietal bones. The anterior parietal spine develops first, followed by a small- er nuchal spine that forms just posterior to it. These spines are generally transient structures that form late in larval development and are re- duced or lost during transformation. In many species of cottids, the spines appear to fuse to- gether enclosing a small canal between the bases of the two spines. This canal eventually becomes part of the cranial lateral line system. In other species, the spines decrease in size without fusing together. Concurrently, sheets of bone extend an- teriorly and posteriorly from the spines and eventually fuse together, forming an incipient cranial arch. Similar parietal and nuchal spines occur in lar- vae of most other scorpaeniform families and appear to be homologous to those of cottid lar- vae. The presence of a parietal and nuchal spine is probably the primitive or ancestral condition in the Cottidae. Parietal and nuchal spines have undergone modification and elaboration in larvae of most of the species of Clinocottus and Oligocottus, and in Artedius fenestralis, A. harringtoni, A. later- alis, and A. Type 3. Two species (Artedius har- ringtoni and Clinocottus acuticeps} have lost these spines completely. Three species (Clinocottus analis, C. embryum, and C. recalvus) have re- tained the primitive condition of possessing a parietal and nuchal spine. The remaining species (A. fenestralis, A. lateralis, O. maculosus, O. sny- deri, and C. globiceps) have tended toward an elaboration and increase in number of parietal spines. Generally, these larvae develop a cluster of three to six spines, which are situated in two transverse rows at the posterior margin of the parietal region. During transformation, these clusters of spines decrease in size and disappear. At the same time, sheets of bone extend ante- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 165 FIGURE 4. Dorsal gut diverticula in larval Artedius fenestralis. riorly and posteriorly from the bases of the two rows of spines and eventually fuse together. This bony arch becomes a part of the cranial lateral line system in juveniles. The absence of parietal spines in A. harringtoni and C. acuticeps is probably a secondary loss and, as such, represents a derived condition. The elaboration of spines into clusters is apparently unique to larvae of Artedius, Clinocottus, and Oligocottus and is also a derived state. PIGMENTATION. — Melanistic pigmentation varies greatly among cottid larvae ranging from relatively unpigmented forms to heavily pig- mented ones. Larvae of Artedius, Clinocottus, and 166 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 Oligocottus are all lightly pigmented and possess numerous, intense melanophores over the dor- solateral surface of the gut and in a row posterior to the anus. The shape and number of midline melanophores varies between species. Larvae of all but one species, A. creaseri, possess several melanophores in the nape region. The presence of head pigment and anteroventral gut melano- phores varies among the species ofArtedius, Cli- nocottus, and Oligocottus. Although pigment patterns are diagnostic at the specific level, they are difficult to evaluate for use in systematic analysis. Many fish larvae in distantly related families, and even orders, are similarly pigmented. Several cottid genera other than Artedius, Clinocottus, and Oligocottus also possess similar pigment patterns. Hence, it is dif- ficult, if not impossible, to determine which pig- ment patterns are primitive and which are de- rived. Nevertheless, trends in certain areas of pig- mentation can be discerned. Among known cot- tid larvae, a discrete nape pigment patch is found only in members of Artedius, Clinocottus, Oli- gocottus, Enophrys, Myoxocephalus, and Gym- nocanthus. Nape pigment is probably derived in cottid larvae. The number of ventral midline me- lanophores situated posterior to the anus ranges from 2 to 33 in larvae ofArtedius, Clinocottus, and Oligocottus. In addition, the shape and spac- ing of these melanophores varies from small dots to long slashes extending onto the ventral finfold to large pigment blotches. Artedius creaseri and A. meanyi both possess irregularly shaped blotches of pigment along the ventral midline. Artedius fenestralis, A. harringtoni, and A. lat- eralis all possess distinctive pigment slashes. Oli- gocottus and Clinocottus larvae possess small, round melanophores. Although these pigment patterns bind certain species together, it is dif- ficult to determine ancestral versus derived states. Both pigment blotches and distinctive midline slashes are found in larval Icelinus. Pigment, as well as other characters, indicates that Icelinus shares close affinities with Artedius; however, the direction of the evolution of pigment patterns cannot be determined. MORPHOMETRICS.— Cottid larvae exhibit a di- versity of body forms. Body shape ranges from short and stubby (Artedius [Group A] and Eno- phrys) to long and slender (Radulinus and Icelus) to globose (Malacocottus). Larval Artedius (Group A), Clinocottus, and Oligocottus have short, stubby bodies with blunt, rounded snouts. Gut length is moderately long and the posterior por- tion of the hindgut trails well below the rest of the body. Measurements of body parts frequently over- lap in larval Artedius (Group A), Clinocottus, and Oligocottus because of their similarity in body shape. These similarities make it difficult to de- termine discrete character states or transforma- tion series. In addition, many body parts change markedly during larval development, frequently exhibiting allometry. Because of the extreme di- versity of forms found in larval cottids, it is dif- ficult to evaluate morphometric characters for systematic analysis. Trends can be observed in only a few body parts. Larvae ofArtedius creaseri and A. meanyi have long, pointed snouts. They develop relatively long ascending processes on the premaxillary. The un- derlying ethmoid cartilage is relatively large, causing a pointed, humped appearance of the snout. Larval Icelinus exhibit pointed snouts similar to that of A. meanyi and A. creaseri. In contrast, all other larval Artedius, Clino- cottus, and Oligocottus have blunt, rounded snouts. The ascending processes of their pre- maxillaries are relatively short, and the ethmoid cartilage forms late in development. Snout length is variable in the outgroup taxa. Sebastes larvae have a somewhat long, pointed snout, but the hexagrammid, cyclopterid, and agonid larvae examined have shorter, rounded snouts. Within the cottids, Enophrys, Leptocot- tus, Hemilepidotus, and Scorpaenichthys larvae all have blunt, rounded snouts. This condition is probably the primitive condition relative to larvae ofArtedius, Clinocottus, and Oligocottus because it is widespread in several divergent gen- era of cottids and scorpaeniforms. The pointed snout appears to be a derived condition. Although gut length varies greatly among cot- tid larvae (Richardson and Washington 1980), larval Artedius (Group A), Clinocottus, and Oli- gocottus possess distinctive guts with the hindgut coiled very loosely and extending posteriorly. The tip of the hindgut extends ventrally well below the rest of the body and posteriorly to the origin of the anal fin. This condition is most pro- nounced in Clinocottus larvae. Larval C. acuti- ceps and C. embryum have especially long, trail- ing hindguts which extend posteroventrally past WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 167 the anal fin origin. A trailing gut is unique to larvae of Artedius (excepting A. meanyi and A. creaseri), Clinocottus, and Oligocottus, and is as- sumed to be a derived condition. PELVIC FIN RAYS.— The number of pelvic fin rays in cottids ranges from one spine and five rays to no spines or rays. Pelvic fin rays are generally considered to be undergoing reduction in the cottids. The primi- tive state is 1,5 fin rays as in other Scorpaeni- formes. Reduction in number of rays is a derived state. Larvae of Artedius, Clinocottus, and Oligocot- tus all possess 1,3 pelvic fin rays, except for A. meanyi. Artedius meanyi usually possesses 1,2 pelvic fin rays. The outermost of these fin rays is markedly long and thickened, and the tips of this ray are separated. Icelinus also possesses 1,2 pelvic fin rays; however, both rays are relatively short and fine. The thickened outer ray of A. meanyi may have evolved through the fusion of two fin rays. If so, this condition may constitute an intermediate state between the three pelvic fin rays of Artedius, Clinocottus, and Oligocottus and the two pelvic fin rays of Icelinus. BRANCHIOSTEGAL RAYS.— The scorpaenids, considered to be the most generalized scorpae- niform (Bolin 1947; Quast 1965), possess seven branchiostegals. The hexagrammids and the zan- iolepids, which occupy an intermediate position between the scorpaenids and cottids (Quast 1 965), both possess six branchiostegals. Most cottids possess six branchiostegal rays; however, the psychrolutids and some freshwater Cottus species have seven branchiostegals. The psychrolutids are a distinct group which possess many derived characters and have apparently diverged from other cottids. Similarly, members of Cottus also possess many derived characters that apparently reflect adaptation to a freshwater habitat. Cottids that are generally considered to be primitive be- cause they retain many primitive features all pos- sess six branchiostegal rays. Although the possession of seven branchioste- gal rays is probably the primitive condition in the scorpaeniforms, possession of six branchi- ostegals appears to be the primitive state within the cottids. Cottid genera such as Icelinus and Hemilepidotus, which Bolin (1947) considered to have evolved from the evolutionary line lead- ing to Artedius, Clinocottus, and Oligocottus, all possess six branchiostegal rays. Six branchioste- gals is probably also the primitive state relative to Artedius, Clinocottus, and Oligocottus. Be- cause those cottids generally considered to be from the same evolutionary lineage as Artedius (Bolin 1934, 1947) all have six branchiostegals, it is assumed that the seven branchiostegals found in Artedius harringtoni and in some Clinocottus globiceps are secondarily derived. POSTTEMPORAL-SUPRACLEITHRAL SPINES. — Larvae of most known scorpaeniforms (includ- ing most known cottids) develop three spines in the posttemporal-supracleithral region of the head. Generally, two spines form first on the ven- tral portion of the posttemporal bone, and one spine forms midway along the posterior margin of the supracleithrum. These spines persist dur- ing transformation at which time the surround- ing portions of the posttemporal and supracleith- ral bones undergo modification and canals form in these bones between the spines. This entire complex then develops into the junction point of the cephalic lateral line system and the lateral line system. This pattern of spines probably rep- resents the plesiomorphic condition in larval cot- tids. In larvae of Artedius, Clinocottus, and Oligo- cottus the posttemporal-supracleithral spines are frequently modified. The modifications appear to be correlated with those of the parietal and nuchal spines. Larvae that have lost parietal spines do not develop posttemporal-supracleith- ral spines, and larvae that have evolved complex clusters of parietal spines also develop clusters of posttemporal-supracleithral spines. Neither/!. harringtoni nor C. acuticeps larvae develop any spines in the posttemporal-supracleithral region. Larvae of A. creaseri, A. meanyi, C. embryum, and C. recalvus possess two posttemporal and one supracleithral spine. Artedius fenestralis, A. lateralis, O. maculosus, O. snyderi, and C. glo- biceps all develop more than three posttemporal- supracleithral spines. As described for parietal and nuchal spines, the posttemporal-supracleithral spines vary among species of Artedius, Clinocottus, and Oli- gocottus. This variability among closely related species suggests that these spines may be undergoing rapid modification in this group and loss of spines or possession of clusters of spines may represent convergent or parallel evolution. As with the parietal and nuchal spines, the absence of posttemporal-supracleithral spines in 168 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 TABLE 2. CHARACTER STATES USED IN SYSTEMATIC ANALYSIS AND THEIR DISTRIBUTION AMONG LARVAL ARTEDIUS, CUNOCOTTVS, AND OUGOCOTTVS AND THE OuTGROUP TAXA. Preopercular spine pattern Auxiliary f opercular sp Bubble of >rc- skin at nape No. of . ' preopercular preopercular Dorsal m,d, & tpinp vpntrfll spines Fqiial- larg- cpmp Pr<>«- unes Pres- <5 >5 sized est largest Absent ent Absent One Two Absent ent Artedius fenestralis X XX X X Artedius harringtoni X XX X X Artedius lateralis X XX X X Artedius Type 3 X XX X X Oligocottus maculosus XXX X X Oligocottus snyderi XXX X X Clinocottus acuticeps XXX X X Clinocottus analis XXX X X Clinocottus embryum XXX X X Clinocottus globiceps XXX X X Clinocottus recalvus XXX X X Artedius creaseri XX X X X Artedius meanyi XX X X X Scorpaenichthys marmoratus XX X X X Hemilepidotus hemilepidotus XX X X X Leptocottus armatus XX X X X Enophrys bison XX X X X Myoxocephalus sp. XX X X X Icelinus sp. XX X X X Radulinus asprellus X X X X Sebastes flavidus X X X X Hexagrammos sp. * * * X X Cyclopteridae Typel * * * X X Stellerina xyosterna XX X X X * Character absent. A. harringtoni and C. acuticeps is probably a sec- ondary loss and hence a derived state. The trend toward an elaboration of these spines is found only in members of Artedius, Clinocottus, and Oligocottus and is also considered a derived state. CHARACTERS SELECTED FOR SYSTEMATIC ANALYSIS.— Of the 50 characters examined, 10 characters were selected for use in the phyloge- netic analysis (Table 2). These 10 best fit the criteria for character selection listed in the meth- ods. Character 1 : Number of preopercular spines. A) 5 B) 0 Q 4 D) >5 Character 2: Preopercular spine pattern. A) preopercular spines equal-sized WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 169 TABLE 2. CONTINUED. Parietal Hindgut shape Dorsal gut diverticula spines No. pelvic fin rays Snout shape Trails Modi- Absent Small Large Two fied >3 3 2 Rounded Pointed Compact ly greatly 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 XX 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 XX X X X X X X X X X X X X X X X X X * X X X X XX X B) dorsalmost spine largest C) dorsal, middle, and ventral spines larg- est Character 3: Basal preopercular spines. A) absent B) present Character 4: Auxiliary preopercular spines. A) absent B) 1 C) 2 Character 5: Bubble of skin at nape. A) absent B) present Character 6: Diverticula on dorsal surface of gut. A) absent B) small bumps C) long diverticula 170 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 I0a. trailing gut «>b. greatly trailing gut FIGURE 5. Cladogram of systematic relationships between Artedius, Clinocottus, and Oligocottus. Characters numbered on cladogram indicate synapomorphies. Character 7: Parietal spines. A) 2 B) modified: clusters of spines C) modified: absent Character 8: Number of pelvic fin rays. A) >3 B) 3 C) 2 Character 9: Snout shape. A) rounded B) pointed Character 10: Hindgut shape. A) compact; not trailing below body; end- ing anterior to anal fin origin B) hindgut trailing slightly below body; ex- tending to origin of anal fin C) hindgut trailing well below body; ex- tending posterior to anal fin origin PHYLOGENETIC RELATIONSHIPS A hypothesis of evolutionary relationships among species in the cottid genera Artedius, Cli- nocottus, and Oligocottus is presented in Figure 5. Two main lineages or sister groups of Artedius, Clinocottus, and Oligocottus larvae are repre- sented in the resulting cladogram based on shared derived characters of the larvae. Species with larvae possessing the synapomorphic characters, multiple preopercular spines and trailing guts, form one major group. In addition to the pos- session of shared, derived characters, larvae of this group are extremely similar in pigmentation and body shape. The second major evolutionary line consists of species sharing two derived char- acters, basal preopercular spines and pointed snouts. This line includes Artedius creaseri, A. meanyi, and Icelinus. The two main evolutionary lines or groups of Artedius, Clinocottus, Oligocottus, and Icelinus correspond to Taranets's (1941) classification of these species. Taranets (1941) placed Artedius meanyi and A creaseri in the subfamily Icelinae, along with members of Icelinus and Chitonotus. He based this decision on the following char- acters: the upper preopercular spine larger than the lower spines; two rows of bony plates on the body— one along the lateral line, the other at the base of the dorsal fins; and scales usually present on other parts of the body. All of these characters are undergoing reduction and are present in sev- eral different genera of cottids. As such, they have WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 171 low systematic value. Taranets (1941) combined Clinocottus, Oligocottus, Artedius corallinus, A. fenestralis, A. harringtoni, A. lateralis, and A. notospilotus in the subfamily Oligocottinae. This subfamily was characterized by the absence of spines or ridges projecting through the skin on the head, weakly developed preopercular spines, body naked or with bony plates reduced, and ". . . other characters." Unfortunately, Taranets did not mention which other characters he ex- amined. In contrast, other investigators have placed Ar- tedius meanyi and A. creaseri in the genus Ar- tedius along with the five other species of Arte- dius discussed above. Howe and Richardson (1978) and Bolin (1947) proposed that although Artedius shared close affinities with Icelinus, it was more closely related to Clinocottus and Oli- gocottus. Jordan (1923), however, stated that^r- tedius was most closely related to Icelinus be- cause of the common possession of bony plates on either side of the dorsal fins. The evolutionary lineage (Fig. 5) containing larvae with multiple preopercular spines in- cludes three distinct groups of species. One of these groups includes all the species of the genus Clinocottus. Larvae of this group share one syn- apomorphic character, inner-shelf preopercular spines. Larvae of Clinocottus possess one or two auxiliary spines on the inner preopercular shelf. These spines appear to be a derived character unique to members of this genus and provide evidence that the genus is monophyletic. This genus has been previously recognized and de- fined on the basis of adult characters, e.g., loss of scales, an advanced anus, and possession of a heavy, blunt penis (Hubbs 1926; Bolin 1934, 1944, 1947; Taranets 1941; Howe and Richard- son 1978). None of these characters is unique to members of Clinocottus. Several different lin- eages of cottids exhibit reduction in squamation, advanced anuses, and penes. Within the genus Clinocottus, C. embryum and C. acuticeps are grouped together on the basis of a synapomorphic character, a trailing hindgut which extends posterior to the origin of the anal fin. Clinocottus acuticeps larvae are unique in their possession of the autapomorphy, distinct hindgut diverticula. Clinocottus embryum and C. acuticeps larvae also are similar in possessing moderately pointed snouts relative to the blunt, rounded snouts of other Clinocottus larvae, a loose bubble of skin in the head region, and light pigmentation. Synapomorphic characters for clarifying inter- specific relationships among C. analis, C. recal- vus, and C. globiceps were not identified. Never- theless, several pigmentation and morphological characters, for which direction of evolution is not known, do suggest possible relationships among these species. Both Clinocottus globiceps and C. recalvus larvae have intense pigmentation over the snout, head, and nape. Both have very blunt, globose heads, large bulging guts, and relatively deep bodies. Juveniles of the two species are nearly inseparable based on external morphol- ogy. These characters suggest that C. globiceps and C. recalvus may be a very closely related species pair. Postflexion larvae of C. analis differ from all other postflexion Clinocottus larvae in possessing an intense band of melanistic pigment over the lateral surface of the body. Unfortu- nately, the polarity (direction of evolution) of many of the transformation series of morpho- metric and pigment characters could not be de- termined; hence, these characters could not be used in the phylogenetic analysis. Pigmentation and morphometric characters have been useful in several systematic studies based on larvae and have frequently been correlated with other de- rived characters (Johnson 1 974; Moser and Ahl- strom 1974; Okiyama 1974; Richardson 1981). Clarification of relationships among the species of Clinocottus must await identification of de- rived characters or a better understanding of the evolution of pigmentation and body shape with- in cottid larvae. The relationships among species of Clinocottus postulated by Bolin ( 1 947) are in close agreement with relationships suggested by larval characters. Bolin placed C. acuticeps in its own subgenus because of its unique possession of a modified penis with a tri-lobed tip, and a membrane con- necting the innermost pelvic fin ray with the ab- domen. Bolin also placed C. analis in its own subgenus because of the retention of minute prickles covering the body. All other members of the genus are scaleless. Clinocottus embryum, C. globiceps, and C. recalvus were placed in the same subgenus because of their large, rounded heads and the retention of a pore behind the last gill. The latter character is plesiomorphic and, hence, of little value for evaluating relationships. Bolin (1947:163) described C. recalvus and C. 172 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 globiceps as "two extremely closely related species" because of their hemispherical head shape, an increased number of cirri on the head, and a pair of lateral knobs near the tip of the penis. Another group of species within the lineage having multiple preopercular spines consists of Oligocottus maculosus and O. snyderi. They share two synapomorphies— a bubble of skin at the nape, and dorsal gut bumps. Larval O. rimensis and O. rubellio are not yet identified and, there- fore, it is not known if these larvae also possess the synapomorphic characters binding O. snyderi and O. maculosus together. Bolin (1934, 1947) defined the genus Oligo- cottus on the basis of the following adult char- acters: absence of scales; presence of a long, slen- der, simple penis; and modification of the anterior anal fin rays in males. Only the last character appears to be unique to members of this genus. As mentioned above, evolution of penes and loss of scales have occurred in several diverse cottid genera. Bolin placed O. maculosus, O. snyderi, and O. rubellio in the same subgenus because of the greatly modified anal fin in males, a per- manently external penis, and loss of all but lateral line scales. He further speculated that O. mac- ulosus was the least specialized member of the subgenus and O. rubellio was the most special- ized. Larval O. snyderi possess the derived, auta- pomorphic characters of multiple prickles cov- ering the parietal and posttemporal regions of the head. In addition, larval O. snyderi possess an accessory spine at the anterior base of most of the main spines on the posterior margin of the preopercle. Both of these conditions are unique specializations of larval O. snyderi. Clarification of relationships of other Oligocottus species must await identification of larval O. rimensis and O. rubellio. Larval characters indicate that while Oligo- cottus and Clinocottus are each a monophyletic group, they are closely related. Larvae of both genera are linked together into a higher-level monophyletic unit by the possession of a dis- tinctive preopercular spine pattern. Taranets (1941) also concluded that Oligocot- tus and Clinocottus are closely related. He placed both genera in the supragenus Oligocottini be- cause of the presence of a penis and the absence of bony plates in both groups. Artedius fenestralis, A. harringtoni, A. later- alis, and A. Type 3 form the third group of larvae with multiple preopercular spines. These larvae share one synapomorphy, an Artedius-type pre- opercular spine pattern. In addition, larvae of this group possess distinctive pigment slashes on the ventral midline posterior to the anus and a strongly humped appearance in the nape region. Known larvae of Oligocottus and Clinocottus do not possess either of these characters; however, larvae of several species of Icelinus and Myox- ocephalus do possess similar pigment slashes on the ventral midline. Although these characters are not unique to Artedius species, they provide additional support for the cohesiveness of this group. Within the Artedius group, A. fenestralis, A. later alis, and A. Type 3 form a distinct subgroup. Larvae of these species share one synapomorphic character, dorsal gut diverticula. Characters identified in this study do not define relation- ships among these three species. Artedius har- ringtoni is probably less specialized than the three species possessing gut diverticula. Artedius har- ringtoni is further distinguished from other Ar- tedius larvae by the possession of seven bran- chiostegal rays. All other cottid larvae examined in this study possess six branchiostegal rays ex- cept several laboratory-reared Clinocottus glo- biceps. Although seven rays are probably a prim- itive character in scorpaeniforms, A. harringtoni appears to have secondarily derived this condi- tion, since none of the outgroup cottids have seven branchiostegals. This grouping of Artedius larvae with multiple preopercular spines corresponds to Taranets's (1941) classification. He placed Artedius coral- linus, A. fenestralis, A. harringtoni, A. later alis, and A. notospilotus together in the supragenus Artediini in the subfamily Oligocottinae. Other workers have placed all species of Ar- tedius in the same subfamily and genus (Hubbs 1926; Bolin 1934, 1947; Rosenblatt and Wilkie 1963; Howe and Richardson 1978). Bolin (1947: 161) included A. creaseri in Artedius because of the retention of hemilepidotid-like scales "in various degrees of reduction," large head, an un- advanced anus, and "normal structure of the pel- vic fins." Artedius meanyi was not reported to occur off California at the time of Bolin's work. Hence, he did not include this species in his classification. WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 173 Rosenblatt and Wilkie (1963) described A. meanyi as being extremely similar to A. creaseri and placed it in Bolin's subgenus Ruscariops along with A. creaseri. Reduction in squamation and number of pel- vic fin rays is found in many different cottid genera and appears to have evolved separately several times. Icelinus, Chitonotus, and Ortho- nopias also possess hemilepidotid-like scales in various degrees of reduction. Several species of Icelinus, Orthonopias, and Chitonotus possess large heads and unadvanced anuses. Characters of the larvae indicate that Artedius creaseri and A. meanyi form a distinct grouping separate from the other species of Artedius. Two synapomorphic characters, a pointed snout and basal preopercular spines provide strong evi- dence that A. creaseri, A. meanyi, and Icelinus form a monophyletic group. In addition, A. meanyi, A. creaseri, and Icelinus larvae are very similar in other pigmentation, morphometric, and spination characters, giving further support for the cohesiveness of this group. In a phenetic study of larval cottids, Richardson (1981) also placed A. meanyi in a group with Chitonotus, Parice- linus, Triglops, and Icelus. (A. meanyi was mis- identified as Icelinus spp. in Richardson's study. See literature section of A. meanyi description.) Although her study was based on similarities of the larvae and not synapomorphies, it supports the grouping of A. meanyi and A. creaseri with Icelinus. Both phenetic and synapomorphic characters of the larvae provide strong evidence that the genus Artedius (as defined by Bolin 1934, 1947) is not monophyletic and that A. creaseri and A. meanyi should be placed separately. Artedius meanyi and A. creaseri appear to be more closely related to species of Icelinus and other members of Richardson's Group 2 than to other species of Artedius. Clarification of relationships among A. meanyi and A. creaseri must await a reex- amination of characters of adult Artedius. Although larvae of the A. meanyi-creaseri group and the Artedius-Clinocottus-Oligocottus group are distinct from one another, they share certain similarities in comparison to other cottid larvae. Both groups have similar pigment pat- terns, morphology, and meristics, suggesting that species of these two groups share a common ancestor. Bolin (1947:159) also speculated that Icelinus, Chitonotus, Artedius, Clinocottus, and Oligocottus constitute a single evolutionary line within the Cottidae. He suggested that "certain details of the more primitive members, partic- ularly the scales, indicate that while these forms undoubtedly did not spring from the modern ge- nus Hemilepidotus, they shared a common and not particularly remote ancestor with the fishes of that genus." Although characters of the larvae do not exclude the possibility of a hemilepidotid- type ancestor, they do indicate that it would be a relatively distant ancestor. Larval Hemilepi- dotus differ markedly from larvae of Artedius, Clinocottus, Oligocottus, and Icelinus in many characters including meristics, morphometrics, osteology, spination, and pigmentation. It is much more likely that the ancestor of this group pos- sessed characteristics similar to both the Arte- dius-Icelinus group and the Artedius- Clinocot- tus-Oligocottus group. Larvae of at least one species of Icelinus and several species of Myox- ocephalus possess a fifth or sixth accessory pre- opercular spine. Larvae of Myoxocephalus also possess two distinct patterns of pigment: one type is lightly pigmented similar to the two Artedius groups, whereas the other has intense bands of lateral pigmentation. An ancestor similar to Ice- linus or Myoxocephalus may well have given rise to Artedius, Clinocottus, Oligocottus, and Iceli- nus. This hypothesis is supported by the presence of one or two accessory preopercular spines in Myoxocephalus larvae. This preopercular spine condition appears to be intermediate between the primitive pattern of four preopercular spines and the derived pattern of multiple preopercular spines. Hence, larvae of the ancestor of Artedius, Clinocottus, and Oligocottus were probably rel- atively lightly pigmented with melanophores present on the head, nape, dorsal surface of the gut, and along the ventral midline posterior to the anus. In addition, the ancestral larvae prob- ably possessed four large preopercular spines with one accessory spine on the inner preopercular shelf, two parietal spines, and three posttempor- al-supracleithral spines. In summary, the hypotheses of relationships between Artedius, Clinocottus, and Oligocottus based on larvae characters is in general agree- ment with previous classifications based on adult characters. Synapomorphic characters of the lar- vae provide strong evidence that Clinocottus, Oligocottus maculosus, O. snyderi, A. fenestralis, A. harringtoni, A. lateralis, and A. Type 3 form 174 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 a monophyletic group within the cottids. Within this group, the genera Clinocottus and Oligocot- tus are very closely related; however, each genus appears to be monophyletic. Artedius fenestralis, A. harringtoni, A. lateralis, and A. Type 3 also form a monophyletic species group closely re- lated to Clinocottus and Oligocottus. However, synapomorphic characters of the larvae provide strong evidence that A. creaseri and A. meanyi are more closely related to Icelinus than to species of Clinocottus, Oligocottus, and other Artedius. The genus Artedius as denned by Bolin (1934, 1947) does not appear to be monophyletic; A. meanyi and A. creaseri should be placed sepa- rately. Clarification of the exact position of these two species in relation to Icelinus and Myoxo- cephalus and the Artedius-Clinocottus-Oligocot- tus group must await identification and exami- nation of larvae of additional species of cottids and reexamination of adult characters. TAXONOMIC DESCRIPTIONS Larvae of Artedius, Clinocottus, and Oligocot- tus have been difficult to identify at both the specific and generic levels because of their strik- ing similarities. Previous descriptions of larvae of this group have been inadequate to separate larvae at both the specific and generic levels be- cause of inaccuracies or insufficient detail. Many of the diagnostic characters useful in separating these larvae are transient features which are pres- ent during only a part of larval development (e.g., head spines, nape bubble). Hence, frequently a combination of several characters is necessary for identification of the larvae. Therefore, to fa- cilitate identification, larval descriptions are ar- ranged in species groups formed by the shared presence of diagnostic characters (Table 3). This matrix table is based on a set of characters that will allow identification of the early life history stages of 1 3 species of Artedius, Clinocottus, and Oligocottus. Larvae of Artedius fenestralis, A. harringtoni, A. lateralis, A. Type 3, Oligocottus maculosus, O. snyderi, Clinocottus acuticeps, C. analis, C. embryum, C. globiceps, and C. recalvus (Groups A, B, and C) are very similar in morphology, pigmentation, and spination. They are all rela- tively lightly pigmented with melanophores pres- ent on the nape, dorsolateral surface of the gut, and in a series on the ventral midline of the tail. Presence and amount of head pigmentation var- ies within the group. All of these larvae possess blunt, rounded snouts, stubby bodies, and a bulg- ing gut which trails somewhat below the rest of the body. These larvae are readily distinguished from all other known cottid larvae by the pres- ence of multiple preopercular spines (>5). Larvae in Group A, Artedius fenestralis, A. harringtoni, A. lateralis, and A. Type 3, all have a distinctively stubby shape, a rounded snout, and a humped appearance in the nape region. They are further distinguished by a series of ven- tral midline melanophores posterior to the anus that extend onto the ventral finfold as charac- teristic pigment slashes in flexion and postflexion larvae. These Artedius larvae possess distinctive preopercular spination; postflexion larvae have a relatively high number (>14) of preopercular spines. The dorsalmost, middle, and ventralmost spines are larger than the other spines creating the "Artedius" spine pattern unique to larvae of this group. Characters such as number of pre- opercular spines, number of ventral midline me- lanophores, size at formation of head pigmen- tation, presence of gut diverticula, and number of branchiostegal rays distinguish larvae of each species of Artedius. Larvae in Group B are Oligocottus maculosus and O. snyderi. These larvae can be distin- guished by the presence of a distinctive bubble of skin situated just anterior to the origin of the dorsal finfold in preflexion and early flexion lar- vae. Larvae of both species are more slender than larvae in Groups A and C and have a relatively short, compact gut. In contrast to larvae of Group A, the dorsalmost preopercular spine becomes larger than other spines in flexion and postflexion larvae. Characters useful in distinguishing larvae of the two species of Oligocottus are number and position of ventral midline melanophores, num- ber of preopercular spines, number of parietal spines or prickles, and presence of melanophores on the nape bubble. Group C includes Clinocottus acuticeps, C. an- alis, C. embryum, and C. globiceps. Larvae of C. recalvus, described by Morris (1 95 1), also belong to Group C based on morphology. This is the least cohesive group in that larvae vary more in morphology and pigmentation than in the other groups. In general, larvae have a long gut, the posteriormost portion of which trails below the rest of the body. Larvae of all species except C. WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 175 III 1 1 1 1 II 1 + 1 1 1 II 'a lit! 1 1 1 1 II 1 1 1 1 1 1 + ? "S S ++.+++ + + + + + \ + ! 1 '§ 1 1 + 1 + ~- J i u ci p .§> a M 0 a 1 g z i Q "S 8 S E K 1 1 + 1 + ^ + + 1 + + + + o ! V 3 1 § G g Jr * 1 1 1 1 II + + 1 + + 1 1 VO 3 « 2 z — , 1 o •- 1 1 + + + 1 1 1 1 1 1 1 II n § £ o .2 <« s Q •fi "3. ^** s z S I t» vo VO vo VO VO *O *O "sO o ^O *O ^O 9 * 2 « ^o •5 I « 1 1 § 1 -s JI > 's s a § i ; 4) , +- + + + + 1 1 1 1 1 1 1 II I 1 O ^ ^ S1 a ^ '6 1 i b S> O 1 0 6 S% (N pO rt «Ai 3 Cu CX t 2 "— ^ /^ o cx ,2 c -^ p^ 13 in « £ <« 5 1 1 '5 3 ^>5 m 2 "S it! « 1 11 1 fc fc & 5 &o S 8 a 33333 g 1 « 111! •o o o. c°5 6 . ^ — « u « u O T3 b e "* «? iipi| illlil H Illl 1 1 .R .K .S .S .K £j 5j b B. S.S lllllt ^ ^ ^ T O O G G G G G ^ ^ 3 O, CX m « _ Z D D OS ffl £ « 4- 4+ =*; < S3 0 Q 176 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B FIGURE 6. Larvae of Artedius fenestralis: A) 3.0 mm NL, B) 3.0 mm NL, C) 4.7 mm NL, D) 6.0 mm NL (from Richardson and Washington 1980). WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 177 B FIGURE 7. Larvae of Artedius fenestralis: A) 7.2 mm SL, B) 9.9 mm SL, C) 1 1.8 mm SL (from Richardson and Washington 1980). embryum have melanistic pigmentation on the head and nape. The dorsalmost preopercular spine is larger than other preopercular spines in postflexion larvae. Characters such as number of preopercular spines, number and spacing of ven- tral midline melanophores posterior to the anus, and presence of hindgut diverticula or bulges are useful in separating larvae of each of these Cli- nocottus species. Group D consists of Artedius creaseri and A. meanyi. These larvae differ from all other larvae of Artedius, Clinocottus, and Oligocottus species listed above in morphology, pigmentation, and spination. They have pointed snouts and large heads, light pigmentation, and four preopercular spines. These characters bind them more closely with Icelinus larvae. In addition, A. creaseri and A. meanyi larvae are further distinguished by large blotch-like melanophores situated along the ventral midline posterior to the anus. Snout to anus length, meristics, finfold pigmentation, and nape pigmentation are useful characters in sep- arating larvae of the two species. Artedius fenestralis (Figures 6-8; Table 4) LITERATURE.— Blackburn (1973) illustrated an 8.5 mm SL larva similar to Artedius fenestralis, 178 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 FIGURES. Juvenile of Artedius fenestralis, 19.1 mm SL. which he described as Cottid 4. Eldridge (1970) and White (1977) briefly described and illus- trated 3.2 mm and 3.9 mm larvae, respectively, which are similar to A. fenestralis. These illus- trations also closely resemble A. later alls larvae. Richardson and Pearcy (1977) listed these larvae as Artedius sp. 2. Richardson and Washington (1980) described and illustrated specimens 3.0, 4.7, 6.0, 7.2, 9.9, and 1 1.8 mm long as Artedius Type 2. IDENTIFICATION.— Juveniles and adults were identified by the following combination of char- acters: high dorsal fin ray counts, absence of nasal and preorbital cirri, and the presence of scales on the head under the entire orbit and in a dense patch on the caudal peduncle. The develop- mental series was linked together primarily by pigmentation, body shape, gut diverticula, and preopercular and parietal spination. Identifica- tion of larvae was further confirmed through comparison with larvae reared from known eggs. Postflexion and transforming larvae were linked with juveniles using pigmentation, cirri patterns, spination, and meristics. DISTINGUISHING FEATURES.— A combination of characters is useful in distinguishing preflexion A. fenestralis larvae including prominent gut di- verticula protruding from the dorsal surface of the abdominal cavity, melanistic nape pigmen- tation, lack of head melanophores, and a series of 13-19 ventral midline melanophores poste- rior to the anus. Late flexion and postflexion larvae are further distinguished by the presence of 1 8-22 preoper- cular spines with the dorsalmost, middle, and ventralmost spines being larger than the others. Postflexion larvae also have a cluster of 5 or 6 spines situated on the posterior margin of each parietal bone. Juveniles of A. fenestralis are distinguished by meristics, dark pigmentation over the dorsolat- eral surface of the body, and 13-16 ventral mid- line melanophores posterior to the anus. Other useful characters include the absence of a nasal and preorbital cirrus, the presence of one or two small cirri on the eyeball, and two frontoparietal cirri. PIGMENTATION.— Newly hatched larval Arte- dius fenestralis reared in the laboratory have no melanistic pigmentation on the head or nape. Intense melanophores are scattered over the dor- solateral surface of the gut. These lateral gut melanophores are frequently faded and difficult to see in field-collected larvae. Posterior to the anus, a series of 13-19 melanophores originates under the third or fourth postanal myomere and extends posteriorly along the ventral body mid- line. An additional 1 or 2 melanophores extend onto the ventral finfold near the notochord tip. These ventral midline melanophores are evenly spaced approximately one every other myomere. During larval development, the head region remains unpigmented. Two to four melano- phores are added on the nape in larvae 3.4 mm long and become embedded in musculature over the notochord by ~7 mm. By that size the pos- terior half of the series of ventral midline me- lanophores appear as distinctive slashes that ex- tend onto the ventral finfold. During transformation (planktonic specimens ~ 12-14 mm long) juvenile pigmentation begins to develop. Melanophores are added on the dor- sal surface of the head, on the tip of the lower jaw, and on the pectoral fin base. Gradually, me- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 179 TABLE 4. BODY PROPORTIONS OF LARVAE AND JUVENILES OF ARTEDIUS FENESTRALIS, A. HARRINGTONI, A. LATERAUS, AND A. TYPE 3. Values given are percent standard length (SL) or head length (HL) including mean, standard deviation, and range in parentheses. Item Artedius fenestralis Artedius harringtoni Artedius lateralis Artedius Type 3 Head length/SL: Preflexion 22.2 ±1.41 (21.1-25.0) 20.9 ± 2.30(18.7-22.3) 20.2 ± 2.68(16.4-22.9) 21.8 ± 0.92(20.7-23.2) Flexion 22.1 ± 1.22(20.3-23.3) 23.1 ± 3.86(24.9-27.0) 23.6 ± 1.92(19.9-26.1) 24.6 ± 1.71 (23.3-27.1) Postflexion 26.1 ± 2.85(20.8-29.4) 29.0 ± 0.50(22.4-34.1) 26.9 ± 1.89(24.3-30.8) 29.0* Juvenile 33.9 ± 2.08 (32.0-34.9) 33.6 ± 3.46 (30.4-36.5) 22.7 ± 3.21(19.1-25.3) - Snout length/HL: Preflexion 15.1 ± 3.77(10.5-20.0) 18.9 ± 5.52(9.9-25.1) 28.0 ± 1.83(25.8-30.1) 19.8 ± 3.85(15.4-24.6) Flexion 20.4 ± 5.96(12.3-26.6) 23.7 ± 6.14(18.3-19.1) 25.2 ±1.91 (22.9-29.4) 23.2 ± 3.51(19.2-26.7) Postflexion 21.8 ± 3.57(17.5-24.5) 24.2 ± 4.19(15.8-31.4) 24.7 ± 3.77(19.4-30.6) 29.4* Juvenile 21.6 ± 2.08 (19.8-23.2) 23.1 ± 1.00(22.3-24.3) 23.0 ± 1.73(18.9-24.7) - Eye diameter /HL: Preflexion 39.7 ± 4.32 (33.7-46.9) 44.7 ± 4.57(39.7-50.1) 44.7 ± 2.87(40.8-47.1) 46.7 ± 6.53 (39.0-57.6) Flexion 41.2 ± 3.42 (37.4-46.5) 40.4 ± 7.33 (38.4-43.2) 40.4 ± 2.88 (37.4-45.5) 37.1 ± 2.16(35.3-40.2) Postflexion 36.9 ± 7.10(27.3-52.1) 32.6 ± 4.09 (23.2-39.8) 36.6 ± 4.06 (37.9-44.6) 37.9* Juvenile 29.1 ± 1.00(24.6-27.3) 31.2 ± 5.77(29.8-31.2) 35.3 ± 5.03(22.1-24.8) - Snout to anus length/SL: Preflexion 45.3 ± 3.59(41.2-50.8) 42.0 ± 3.21 (38.2-48.2) 40.4 ± 4.83 (33.3-45.6) 44.8 ± 2.76(41.4-49.8) Flexion 45.9 ± 2.49 (44.5-50.1) 47.3 ± 8.30 (36.5-55.9) 42.7 ± 3.65 (38.2-47.9) 45.1 ± 2.63(43.2-48.1) Postflexion 48.4 ± 2.26 (42.7-51.4) 49.9 ± 2.67 (45.9-55.4) 48.4 ± 2.18(44.4-50.9) 50.2* Juvenile 49.1 ± 2.00(47.0-51.5) 47.4 ± 3.21 (43.1-49.2) 48.7 ± 4.04(44.1-51.2) - Snout to pelvic fin origin/SL: Preflexion - - - Flexion 25.6 ± 3.21 (20.2-29.8) - - 25.3 ±4.24 (2 1.9-27.6) Postflexion 26.1 ± 3.21 (20.2-29.8) 27.8 ± 1.72(24.5-30.8) 28.1 ± 3.11(24.9-34.1) 29.3* Juvenile 30.4 ± 3.51 (26.5-30.1) 27.4 ± 2.65(24.1-29.4) 27.0 ± 3.61 (23.8-31.0) - Pelvic fin origin to anus/SL: Preflexion - - - Flexion 22.2 ± 5.35 (16.1-27.0) - - 23.1 ±4.95(19.1-26.1) Postflexion 26.1 ± 3.21 (20.2-29.8) 22.8 ± 4.68(17.2-34.8) 20.1 ± 4.20(14.4-26.7) 20.9* Juvenile 1 8.9 ± 1.15(1 8.2-20.2) 19.9 ± 1.15(18.9-21.3) 22.3 ± 2.08(20.3-24.1) - Body depth at pectoral fin base/SL: Preflexion 21.7 ± 1.81 (19.0-23.8) 23.7 ± 2.99 (20.9-29.7) 23.0 ± 3.16(17.9-26.3) 25.9 ± 1.77(23.4-28.2) Flexion 28.5 ± 2.70(24.9-30.1) 28.1 ± 2.45(28.0-31.2) 26.6 ± 2.33(24.2-27.1) 28.2 ± 1.76(26.3-30.1) Postflexion 28.2 ± 1.82(24.2-32.1) 30.5 ± 2.74(23.4-34.1) 28.1 ± 2.29(24.1-32.2) 30.0* Juvenile 25.8 ± 1.53 (24.4-26.8) 22.3 ± 4.15(24.6-25.2) 19.7 ± 3.06(21.9-25.0) - Body depth at anus/SL: Preflexion 19.0 ± 2.49 (14.8-22.2) 27.8 ± 2.45(17.2-27.6) 20.4 ± 2.51 (15.8-21.8) 22.9 ± 3.28(18.2-27.2) Flexion 24.1 ± 2.59 (23.6-27.1) 30.4 ± 3.32(28.1-34.7) 26.3 ± 2.29(24.5-31.3) 29.1 ± 2.06(26.1-31.3) Postflexion 27.9 ± 2.79 (21.4-32.8) 20.9 ± 5.77 (23.9-34.8) 26.3 ± 2.82 (22.4-33.0) 30.3* Juvenile 21.6 ± 2.65 (18.9-24.4) 28.0 ± 1.72(20.8-21.9) 30.3 ± 3.51 (17.4-22.8) - Pectoral fin length/SL: Preflexion 9.1 ± 0.98 (8.2-10.9) 7.4 ± 2.34(4.4-11.1) 11.0± 1.08(9.6-12.2) 10.4 ± 2.91(7.1-13.3) Flexion 9.4 ± 0.71 (8.4-10.4) 12.1 ± 3.54(9.9-15.4) 12.3 ± 1.98(9.6-15.9) 11.2 ± 3.50(7.4-15.1) Postflexion 22.6 ± 6.21 (9.7-29.7) 24.6 ± 7.59(10.3-34.6) 19.8 ± 4.66(13.3-27.6) 21.0* Juvenile 26.2 ± 1.53 (27.1-28.4) 34.1 ± 3.61(29.9-36.6) 30.3 ± 3.51(26.8-34.2) - - = Not present at this stage. * = Only one specimen available in this stage. 180 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 lanophores develop on the anteriormost portion of the spinous dorsal fin, then extend ventrally as a band of pigment stretching from the fourth or fifth dorsal spine to the pigmentation over the dorsal surface of the gut just posterior to the pectoral fin base. Juvenile pigmentation increases markedly in newly settled individuals 1 3 mm SL. Numerous melanophores are added over the dorsolateral surface of the head and become concentrated in the parietal-interorbital region. Additional me- lanophores extend down onto the snout and lips. Laterally, melanophores are added in the cheek region between the eye and the preopercle and the dorsal portion of the opercle. Several mela- nophores are clustered at the posterior edge of the lower jaw. The ventral surface of the head remains unpigmented. Pigmentation gradually extends from the head posteriorly across the dor- solateral surface of the body until it fuses with bands of pigment reaching from the middle of the spinous dorsal fin to the gut. Pigmentation increases on the anterior end of the dorsal fin creating a dark blotch of pigment across the first four dorsal spines. Melanistic pigmentation also increases on the pectoral fin base with melano- phores extending onto the pectoral fin rays and eventually forming several bands of pigmenta- tion. Several irregular clusters of melanophores appear along the lateral midline and gradually form a band of pigment reaching from the gut to the caudal peduncle. As juvenile pigmentation develops, saddles of pigment form along the dorsum in an anterior to posterior sequence. The first saddle or band of pigment forms under the 4th-7th dorsal fin rays. Gradually, melanophores extend ventrally from the pigment saddle and merge with the lat- eral midline melanophores. Concurrently, a sec- ond saddle of pigment forms under the 9th- 10th dorsal fin rays, while a third saddle of pigment begins to develop under the 1 3th-l 5th dorsal fin rays. Melanophores from these pigment saddles also extend ventrally and fuse with the lateral midline pigment. At the same time, melano- phores are added on the dorsal fin forming three to four bands. Melanophores extend ventrally from the lateral midline band and form a series of five to eight scallops which reach just below the lateral midline. The rest of the ventrolateral surface of the body remains characteristically un- pigmented until juveniles reach about 19-20 mm. As the dorsal pigment saddles are forming, the lateral midline melanophores extend posteriorly to the base of the caudal fin where they form a dark band. Gradually, melanophores extend onto the caudal fin rays forming three or five indistinct bands of pigment. Approximately 13-16 ventral midline melanophores remain visible in juve- niles up to ~20 mm long. MORPHOLOGY. — Larvae of Artedi us fenestralis hatch at ~ 3. 5-3. 8 mm NL. Rexion of the no- tochord occurs between 5.9 and 6.8 mm NL. The largest planktonic larva collected is 1 3.9 mm and is beginning to undergo transformation. The smallest ben thic juvenile examined is 13.1 mm. Thirty-four selected specimens, 3.2-21.2 mm, were examined for developmental morphology. Larval A. fenestralis have stubby bodies with a humped appearance in the nape region. Dis- tinctive diverticula extend dorsolaterally from the dorsal surface of the gut just posterior to the origin of the pectoral fin base. These diverticula are present in newly hatched larvae and remain prominent in the largest planktonic larvae. The diverticula completely disappear in benthic ju- veniles shortly after settling. The gut itself is moderately long and the posterior portion of the hindgut trails well below the rest of the body. Snout to anus length increases from 43% to 45% SL during larval development, then increases to 49% SL in benthic juveniles. Artedius fenestralis larvae have a short, rounded snout with snout length increasing from 1 5% HL in preflexion lar- vae to 23% HL in postflexion larvae and juve- niles. FIN DEVELOPMENT.— Caudal fin rays begin to form at ~6 mm. The adult complement of prin- cipal caudal rays is present in larvae ~7 mm long. The bases of the dorsal and anal fin rays appear in 7-7.5 mm larvae. The full complement of fin rays is formed by ~8.5-9 mm. Dorsal fin spines begin to form at ~8 mm, and the full complement of spines (VIII-IX) is present by ~9.5 mm. Although pectoral fin rays are visible by ~7 mm, the adult complement (14-16) is not formed until ~9 mm. Pelvic buds form between 6.5 and 7 mm and the adult complement of 1,3 pelvic fin rays is formed in larvae ~ 1 0 mm long. SPINATION.— Seven to 13 tiny spines begin to form along the posterior margin of the preopercle in larvae ~4.7 mm NL. The preopercle appears to develop in two arc-shaped sections, which overlap slightly at the angle of the preopercle. Three to 7 spines are present along the dorsal- most section and 6-8 spines occur on the lower WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 181 B FIGURE 9. Larvae ofArtedius harringtoni: A) 3.0 mm NL, B) 4.7 mm NL, C) 6.9 mm NL (from Richardson and Washington 1980). section. The two sections fuse together in post- flexion 7 mm long. Two spines located at the site of fusion begin to increase in length relative to the other preopercular spines. Concurrently, the preopercular spines increase in number during larval development and range between 18 and 21 in larvae >8 mm long. The dorsalmost and middle two spines continue to increase in size relative to the other spines, becoming nearly three times as long. The ventralmost 2 or 3 spines also increase in size, becoming 1.5 to 2 times as long as the other spines. The number of preopercular 182 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 spines decreases in transforming larvae > 1 3 mm long. The smaller spines (4-6, 8-10, and 13-15) begin to regress first. Newly settled juveniles pos- sess one large dorsal preopercular spine. The low- er spines are visible only as serration, or bumps, on the preopercular margin. The dorsalmost spine continues to increase in size while the lower bumps eventually disappear in juveniles ~ 1 9 mm long. Clusters of spines also develop in the parietal and supracleithral-posttemporal regions. One or two spines form at the posterior end of the pa- rietal in larvae ~6 or 7 mm long. A third spine is added in 7-8-mm larvae. Larvae >8 mm have four to six spines located in two rows on each side of the head. Usually, three spines occur in the anterior row while two or three spines are present in a second row posterior to and parallel to the first row. These spines begin to regress in size in transforming larvae >12-13 mm long. The anterior spines curve posteriorly, eventually fusing with spines from the posterior row form- ing a hollow arch and canal. This canal develops into part of the cranial lateral line system in ju- veniles ~14-15 mm long. Two small spines develop on the ventral por- tion of the posttemporal in larvae 6-7 mm long. A third spine is added on the posttemporal in larvae >8 mm long. Concurrently, another spine develops on the dorsal tip of the supracleithrum. These spines remain prominent in planktonic larvae <12 mm long; however, in transforming juveniles the spines gradually curve dorsally and ventrally and fuse together forming a bony tube or canal. This canal becomes the anteriormost juncture of the lateral line and cephalic lateral line systems in juveniles. Artedius harringtoni (Figures 9-11; Table 4) LITERATURE.— Blackburn (1973) described a 4.6 mm larva that he called Cottid 6 that is sim- ilar to A. harringtoni. Richardson and Washing- ton (1980) illustrated and described specimens 3.0, 4.7, 6.9, 7.3, 9.3, and 13.6 mm long. IDENTIFICATION.— Juveniles and adults were identified primarily on the basis of the following characters: high dorsal fin ray counts (16-18), low pectoral fin ray counts (usually 14), presence of seven branchiostegals, presence of a preorbital cirrus, scales extending onto the head under only the posterior portion of the orbit, and scales ab- sent on the snout. The developmental series of larvae was linked together by pigmentation, pre- opercular spination, absence of gut diverticula, body shape, and the possession of seven bran- chiostegals. Postflexion and transforming larvae were linked with juveniles primarily on the basis of pigmentation, meristics, and presence of a preorbital cirrus. DISTINGUISHING FEATURES.— Characters use- ful in distinguishing small larval A. harringtoni are a combination of presence of melanistic nape pigment, lack of head pigmentation, a series of 21-33 pigment slashes along the ventral midline of the tail, and a humped appearance in the nape region. Absence of dorsal gut diverticula distin- guishes larval A. harringtoni from similarly pig- mented larvae of A. later alis, A. fenestralis, and A. Type 3. Postflexion larvae 6.5 mm are distinguished by the presence of 1 8-22 spines along the pos- terior margin of the preopercle. The dorsalmost and middle preopercular spines are characteris- tically larger than the other spines. Larvae >7 mm have seven branchiostegal rays. Larvae of all other species of Artedius have only six bran- chiostegal rays. Juvenile A. harringtoni may be recognized by the dark pigmentation over the head and nape, possession of seven branchiostegals, retention of 1 8-22 ventral midline melanophores, possession of a preorbital cirrus, and dorsal and pectoral fin ray counts. PIGMENTATION.— Preflexion larvae have no melanistic pigmentation on the head; however, 3-5 small, external melanophores are concen- trated in a dense patch on the nape. The dor- solateral surface of the gut is covered with nu- merous large, intense melanophores. One to 8 tiny melanophores encircle the anus. Posterior to the anus, the only pigmentation consists of a series of 23-33 melanophores positioned along the ventral midline. This series originates under the first to third postanal myomere and extends posteriorly toward the tail tip with 1 or 2 me- lanophores positioned under each myomere. An additional 1 to 3 melanophores frequently occur on the caudal finfold near the tail tip. During larval development the head region re- mains unpigmented. The nape melanophores be- come embedded in the musculature over the no- tochord in larvae >7 mm. Concurrently, the number of ventral midline melanophores de- creases to between 21 and 30, and the posterior WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 183 FIGURE 10. Larvae ofArtedius harringtoni: A) 7.3 mm SL, B) 9.3 mm SL, C) 13.6 mm SL (from Richardson and Washington 1980). half of the series appears as characteristic pig- ment slashes that extend onto the ventral finfold. During transformation, planktonic larvae > 10 mm begin to develop juvenile pigmentation. Me- lanophores are added on the tip and base of the lower jaw, on the cheek between the eye and the dorsalmost preopercular spine, on the opercu- lum, and on the isthmus. Pigmentation increases markedly over the head in newly settled benthic juveniles. Melanophores develop on the snout and upper lip and on the dorsal surface of the head over the brain. Me- lanophores gradually extend posteriorly from the head and eventually join with the nape pigmen- tation. Concurrently, melanophores extend pos- teroventrally from the posttemporal region to- ward the dorsal gut pigment. Numerous large melanophores form over the base of the pectoral 184 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 FIGURE 11. Juvenile of Artedius harringtoni, 13.9mmSL. fin, and subsequently extend onto the pectoral fin rays forming four or five distinct pigment bands across the fin. A band of melanophores also extends ventrally from the pectoral fin base and covers the isthmus. In juveniles > 1 3 mm long, the entire head is heavily pigmented. Melanophores extend pos- teriorly from the head to a vertical line under the seventh dorsal fin spine. A dense patch of me- lanophores develops at the anterior end of the spinous dorsal fin forming a dark blotch across the fin membrane between the first four dorsal spines. Scattered melanophores are added along the rest of the dorsal fin eventually forming three or four bands of pigment. Posterior to the head, pigmentation is added in three saddles along the dorsum. The first sad- dle of pigment forms under the 2nd-4th dorsal fin rays, the second saddle forms under the 7th- 10th dorsal fin rays, and the third forms under the 1 3th- 1 5th fin rays. Concurrently, an irregular band of faint melanophores develops along the lateral midline. This lateral pigmentation grad- ually extends posteriorly from the abdominal re- gion to the caudal peduncle. As development proceeds, bands of melanophores extend ven- trally from each of the saddles on the dorsum and merge with the lateral midline pigment. As a result, the dorsolateral surface of the tail is covered by bands of pigmentation, which enclose small unpigmented saddles and circles creating a characteristic pattern. Subsequently, groups of melanophores extend ventrally from the lateral midline pigmentation creating a scalloped edge of pigment along the ventrolateral body surface. Eventually, melano- phores from the tips of each scallop extend lat- erally and join together enclosing four to six dis- tinctive unpigmented circles, characteristic of juvenile A. harringtoni. In late stages of juvenile pigmentation, the lat- eral band of melanophores extends posteriorly to the base of the caudal fin. Melanophores are added on the caudal fin rays forming five to seven bands of pigmentation. Between 18 and 21 ventral midline melano- phores remain visible in juveniles < 1 5 mm. MORPHOLOGY.— The smallest A. harringtoni larva from plankton collections is 3.0 mm NL and still retains remnants of its yolk. Larvae undergo flexion of the notochord between 5.2 and 6.4 mm NL. The largest planktonic larva examined is 13.6 mm long and beginnning to undergo transformation. The smallest benthic juvenile collected in tidepools is 12.9 mm and is just beginning to develop juvenile pigmenta- tion on the head and pectoral fin base. Thirty- five selected specimens, 3.0-13.7 mm, were ex- amined for morphometrics. Larvae of A. harringtoni are stubby with a dis- tinctive humped appearance in the nape region. Unlike larval A. fenestralis, A. lateralis, and A. Type 3, larval A. harringtoni have no dorsal gut diverticula. The gut is moderately long with snout to anus length ranging from 42% in preflexion larvae to 50% SL in postflexion larvae. Relative snout to anus length decreases slightly in benthic juveniles. The hindgut appears to trail below the rest of the body. Relative body depth at the pec- toral fin base increases from 23% in preflexion larvae to 30% in flexion and postflexion larvae. Artedius harringtoni have blunt heads and rounded snouts. Head length increases relative to body length during development, averaging 21% in preflexion larvae, and 34% SL in juve- niles. Snout length increases from 19% to 22% HL during larval development. FIN DEVELOPMENT.— A thickening in the hy- pural region of the developing caudal fin is first visible at 4.7 mm NL, just prior to the onset of WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 185 notochord flexion which occurs at ~ 5.2 mm NL. Caudal fin rays begin to form in larvae ~6 mm; however, the adult complement of principal cau- dal rays is not complete until larvae reach ~7 mm long. Bases of the dorsal and anal fin rays form in larvae ~6-7 mm long. Dorsal spines begin to form in larvae ~7-8 mm long. The adult com- plement of dorsal and anal fin rays is complete at 9.3 mm. Pectoral fin rays are first visible be- tween 6 and 7 mm, and the adult complement (13-15) is countable at ~7.5 mm. The pelvic fin bud begins to form at ~7.1 mm, and the adult complement of 1,3 is complete by ~ 10 mm. SPINATION.— Eight to ten tiny spines begin to form along the posterior margin of the preopercle in larvae ~4.5 mm NL. The number of spines increases to 1 8-22 in flexion and postflexion lar- vae. By the end of flexion, ~6.7 mm, the middle two spines (7-9) begin to increase in size relative to the other preopercular spines. In larvae >7.5 mm, the dorsalmost two or three spines also in- crease in size relative to other spines. As devel- opment proceeds, the dorsalmost and middle spines increase in length and diameter creating a characteristic pattern with small, inconspic- uous spines situated between the dorsalmost and middle spines, and ventral to the middle spines. In larvae >8.5 mm, the ventralmost four or five spines also become somewhat larger than the spines directly above them. When larvae reach ~10-11 mm SL, the preopercular spines begin to regress with the small, inconspicuous spines disappearing first. At the onset of transformation (~ 12-1 3 mm) only four spines remain in the approximate position of the original spines (1- 2, 4-9, 12-14, and 18-22). In newly settled ju- veniles, the dorsalmost preopercular spine be- comes quite long and stout while the lower three spines gradually become smaller and visible only as slight bumps on the margin of the preopercle. Spines never develop in the parietal and post- temporal region of the head. However, in cleared and stained larvae, bony thickenings are visible in the parietal region at the same position as parietal spines found in other cottid larvae. Artedius lateralis (Figures 12, 13; Table 4) LITERATURE.— Budd (1940) described and il- lustrated a newly hatched larva of A. lateralis 4. 1 mm TL. Marliave (1975) described larvae of A. lateralis and illustrated specimens 4 mm TL, 8 mm TL, 1 1 mm TL, and 14 mm TL long. IDENTIFICATION.— Small larval A. lateralis were reared from eggs spawned from known adults. Juveniles and adults were identified using the following characters: pigmentation, absence of scales on the head and caudal peduncle, absence of nasal and preorbital cirri, and the presence of 3-1 1 scales in the longest row in the dorsal scale band. The developmental series was linked to- gether primarily on the basis of pigmentation, preopercular spination, presence of gut divertic- ula, and body shape. Postflexion and transform- ing larvae were linked to juveniles by pigmen- tation, cirri patterns, spination, and meristics. DISTINGUISHING FEATURES.— Characters use- ful in distinguishing preflexion larvae of Artedius lateralis are prominent diverticula which extend dorsolaterally from the dorsal surface of the gut just posterior to the pectoral fin bases, the lack of head and nape pigment in larvae <6 mm NL, and a series of 22-32 melanophores that lie along the ventral midline posterior to the anus. The anterior half of the series is characterized by one large melanophore per myomere while the pos- terior half of the series consists of two or three smaller pigment slashes per myomere. Postflexion larvae of A. lateralis >6.2 mm can be distinguished from other Artedius larvae by melanistic pigmentation over the brain. Ju- venile A. lateralis are distinguished by two dark bars of melanophores extending ventrally from the dorsal fins across the lateral surface of the body trunk, the series of 1 1-21 ventral midline melanophores, and meristics. PIGMENTATION.— Newly hatched larvae of A. lateralis have no melanistic pigmentation on the head or nape. Dense, round melanophores are concentrated over the dorsolateral surface of the gut and extend dorsally onto the gut diverticula. A cluster of 4 to 6 small melanophores sur- rounds the anus. Posterior to the anus, a series of 22-32 melanophores lies along the ventral midline of the body. These melanophores orig- inate under the third or fourth postanal myomere and extend posteriorly toward the tail tip where several additional melanophores extend onto the caudal finfold. Melanophores in the anterior half of this series are relatively large and spaced one per myomere. The posteriormost melanophores appear as small pigment slashes, which extend onto the ventral finfold and are closely spaced two or three to every myomere. During larval development, melanophores form on the dorsal surface of the head in larvae >6.3 mm. Two to five melanophores also form 186 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B FIGURE 12. Larvae ofArtedius lateralis: A) 4.6 mm NL, B) 6.4 mm SL, C) 7.1 mm SL. at the base of the cleithrum and along the ventral midline of the gut in larvae >5.2 mm. These melanophores are arranged in a characteristic T shape with two melanophores positioned as hor- izontal slashes at the base of the cleithrum and one to three melanophores extending posteriorly along the ventral midline of the gut. During transformation, in planktonic larvae >8 mm, melanistic pigmentation increases markedly on the dorsal surface of the head with 33-44 dark melanophores covering the brain. Melanophores also form just posterior to the lower jaw, on the cheek between the eye and the preopercle and on the operculum. Ventral mid- line melanophores remain unchanged in number and spacing. Pigmentation increases markedly in newly set- tled juveniles > 10 mm long. Dark melanophores form on the dorsolateral surfaces of the head and extend anteriorly onto the snout and upper and lower lips. Several melanophores are added to the gular region beneath the lower jaw. Intense pigment forms on the bases of the pectoral fins and several large melanophores extend onto the pectoral fin rays. Gradually, melanophores from the base of the pectoral fin extend ventrally form- ing a band of pigment across the isthmus. With development, pigmentation increases on the head WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 187 FIGURE 13. Young ofArtedius lateralis: A) 9.1 mm SL, B) 13.3 mm SL. so that in larvae > 1 2 mm SL, the entire head is darkly pigmented. Shortly after settling, in larvae between 10 and 1 1 mm, a patch of melanophores is added to the dorsal fin between the fourth and sixth spines. These melanophores extend ventrally across the dorsum toward the pigmentation on the pectoral fin base. A second band of melanophores forms on the second dorsal fin membrane between the 2nd and 4th fin rays. Gradually this band extends anteroventrally below the lateral midline. With development, the two vertical bands of pigment become very dark and intense. Melanophores from these bands extend dorsally across the dor- sal fins. Concurrently, three saddles of faint me- lanophores are added posteriorly along the dor- sum. The first saddle of pigment forms under the 8th- 10th dorsal fin rays, the second saddle is added under the 1 2th-l 5th fin rays, and the third saddle forms on the dorsal surface of the caudal peduncle. Gradually, melanophores from these pigmented saddles extend ventrolaterally and join together forming a band along the lateral mid- line. Lateral pigmentation extends posteriorly and forms a band along the base of the caudal fin. Melanophores extend onto the caudal fin rays, gradually forming two or four distinct bands across the caudal fin. Ventral midline melano- phores decrease in number in juveniles from 1 1 to 21. These melanophores remain visible in ju- veniles < 1 5 mm long. MORPHOLOGY.— Artedius lateralis larvae are 3.9-4.5 mm long at hatching. Rexion of the no- tochord occurs between 5.0 and 6.3 mm NL. The largest planktonic specimen observed is 9.2 mm and beginning to develop juvenile pigmentation. A. lateralis settle at a relatively small size, ~9.5 to 10.5 mm. Thirty-three specimens (4.1-12.1 mm) were examined for developmental mor- phometrics. Larvae of A. lateralis are rather stubby with a moderately short gut. Snout to anus length av- erages 40% in preflexion larvae, then increases to 48% SL in postflexion larvae and juveniles. Pronounced diverticula extend dorsally from each side of the gut just posterior to the pectoral fin base. The diverticula are present in newly hatched larvae and remain prominent throughout larval development. Tiny remnants of the diverticula are present in newly settled juveniles between 9 and 10.5 mm long. Larvae of A. lateralis have a long rounded snout 188 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 relative to other Artedius Group A larvae. Snout length decreases from 28% to 24% HL during larval development. FIN DEVELOPMENT.— The notochord begins to flex in larvae ~5 mm long, and is fully flexed in larvae between 5.7 and 6.3 mm. Caudal rays begin to form during flexion in larvae ~5.5 mm; however, the adult complement of 6 + 6 prin- cipal caudal rays is not complete until ~6 mm. Bases of the dorsal and anal fins are first visible in larvae between 6 and 6.5 mm and the adult complement of dorsal (15-17) and anal (12-14) rays is complete by about 7.5-8 mm. Pectoral fin rays begin to form between 7 and 8 mm, but the full complement of rays (14-16) is not formed until larvae are ~8 mm long. Pel- vic fin buds are first visible in a 7.4 mm larva; however, the fin rays are not countable until 9 mm. SPINATION.— In preflexion larvae ~4.5 mm NL 8-9 tiny spines are visible on the posterior mar- gin of the preopercle. As larvae undergo flexion of the notochord, the number of preopercular spines increases to 9-14. By the end of flexion, larval A lateralishave 14-16 preopercular spines. In larvae >7 mm, the dorsalmost and middle (spines 6-9 from the top of the preopercle) be- come slightly longer than the other preopercular spines. These spines never become more than 1.5 times larger than the other preopercular spines, in contrast to the situation in larvae of A. harringtoni, A. fenestralis, and A. Type 3 in which the dorsalmost and middle preopercular spines may be nearly 2.5 times larger than the other spines. Preopercular spines begin to regress in transforming specimens >9 mm. The dorsal- most spine increases in size while the lower spines (4-6 and 9-1 2) decrease in size becoming visible only as small serrations or irregularities on the preopercular margin. Spines 7-8, 12-13, and 1 6- 1 8 fuse together to form blunt bumps along the preopercular margin. In juveniles > 13 mm, only the large, dorsalmost spine remains. Transform- ing larvae reared in the laboratory possess 4-5 small spines at the posterior margin of the pa- rietals. These spines are not present in planktonic larvae from field collections, nor are they visible in newly settled juveniles from tidepools. Artedius Type 3 (Figures 14, 15; Table 4) LITERATURE.— Larvae of Artedius Type 3 have not been previously described. IDENTIFICATION.— Only a partial size series (2.9-7.6 mm) of Artedius Type 3 larvae are avail- able, all from California collections. The pres- ence of prominent gut diverticula and the char- acteristic Artedius-type preopercular spine pattern (dorsalmost, middle, and ventralmost spines larger than the others) identifies this larval type as an Artedius. Larvae remain unknown for only two species of Artedius, A. corallinus and A. no- tospilotus. Meristics of the largest larva of Ar- tedius Type 3 coincide with those recorded for both A. corallinus and A. notospilotus. However, pectoral counts fit those of A. notospilotus most closely. The 7.6 mm larval A. Type 3 possesses 16 pectoral fin rays. Ninety % of the A. noto- spilotus examined by Howe and Richardson (1978) possessed 16 pectoral fin rays while only 10% of A corallinus specimens possessed 16 pec- toral fin rays. Pigmentation along the ventral midline pos- terior to the anus of A. Type 3 larvae (9-1 3 widely spaced melanophores) coincides most closely with that of juvenile A. corallinus. Several A. coral- linus 13.5-14 mm long, possess 3-6 widely spaced ventral midline melanophores. In con- trast, a 16-mm juvenile A notospilotus possesses 24 ventral midline melanophores spaced one every one or two myomeres. Adult A. corallinus are common in the inter- tidal areas of the southern California coast where Artedius Type 3 larvae were collected (Miller and Lea 1972). Artedius notospilotus adults are rare in the same area. Additional larger specimens are needed before larvae of Artedius Type 3 can be specifically iden- tified. DISTINGUISHING FEATURES.— Artedius Type 3 larvae are distinguished as an Artedius by the distinctive diverticula that extend dorsolaterally from the dorsal surface of the gut just posterior to the pectoral fin base. Artedius Type 3 larvae are distinguished from small larvae of A. fenes- tralis, which possess similar diverticula, by the low number (9-13) of ventral midline melano- phores posterior to the anus. Other characters useful in distinguishing small A. Type 3 larvae are absence of head pigmentation and presence of a cluster of 2-4 melanophores in the nape region. Preopercular spines begin to form in lar- vae <4.1 mm NL. Preopercular spines do not form in other Artedius larvae with multiple pre- opercular spines until ~4.5 mm NL. Flexion and postflexion larval Artedius Type 3 possess 21- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 189 FIGURE 14. Larvae ofArtedius Type 3: A) 3.2 mm NL, B) 4.1 mm NL, C) 4.9 mm NL. 24 preopercular spines, more than other larval Artedius (groups A and D), all of which have < 2 1 preopercular spines. PIGMENTATION.— Small preflexion larvae of Artedius Type 3 possess no melanistic head pig- mentation. Two to four small external melano- phores are clustered on the surface of the nape. Numerous dark, rounded melanophores are con- centrated over the dorsolateral surface of the gut and extend dorsally onto the gut diverticula. One to four small melanophores are clustered around the anus. Posterior to the abdominal cavity, the only pigmentation consists of a series of 9-13 mela- nophores located along the ventral midline. This series of melanophores originates under the third to fourth postanal myomere and extends poste- riorly toward the tail tip. Each melanophore is 190 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B FIGURE 15. Larvae of Artedius Type 3: A) 6.8 mm SL, B) 7.3 mm SL. spaced approximately two to three myomeres apart. An additional one to five pigment slashes extend onto the caudal finfold near the tail tip. Pigmentation changes little during larval de- velopment. Melanophores are added in the nape region and become embedded in the musculature over the notochord in larvae <5.5 mm. Mela- nophores are also added in the isthmus region and along the ventral midline of the gut. MORPHOLOGY.— The smallest larval Artedius Type 3 is 2.9 mm NL and possesses remnants of the yolksac. Larvae undergo flexion of the no- tochord between 5.6 and 6.9 mm NL. The largest specimen examined is 7.6 mm long and has re- cently completed notochord flexion. Size at transformation is unknown. Thirteen larvae (2.9- 7.6 mm long) were examined for developmental morphology. Larvae of Artedius Type 3 are rather stubby with a moderately long gut; the posteriormost portion of this gut trails somewhat below the rest of the body. A prominent diverticulum extends dorsally from each side of the gut just posterior to the base of the pectoral fin. Diverticula are present in the smallest larva examined (2.9 mm NL) and remain pronounced in the largest spec- imen. Snout to anus length averages 45% SL in both preflexion and flexion larvae. Body depth at the pectoral fin base increases during development from 26% in preflexion larvae, to 28% in flexion larvae, and 32% SL in the single postflexion larva. Relative body depth at the anus also in- creases with development, from 23 to 30% SL. The distances from the snout to the origin of the pelvic fins and from the origin of the pelvic fins to the anus averages 26 and 22% SL, respectively, in late flexion and early postflexion larvae. Artedius Type 3 larvae have a rather large head with a blunt, rounded snout. With development relative head length increases from an average of 22% in preflexion larvae to 25% in larvae undergoing flexion of the notochord, and 29% SL in the postflexion larva. Jaw length averages about 43% HL throughout early larval devel- opment. In contrast, eye diameter decreases dur- ing development from an average of 47% in pre- flexion larvae to 37% HL in flexion and early postflexion larvae. FIN DEVELOPMENT.— A 4.9-mm NL larva ex- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 191 hibits a slight thickening of the hypural region of the forming caudal fin. By 5.6 mm, the no- tochord of larval A. Type 3 is strongly flexed and caudal rays are beginning to form. Notochord flexion is nearly complete by ~7 mm and the adult complement (6 4- 6) of principal caudal rays is countable. The dorsal and anal fin bases are first visible in larvae between 6.8 and 6.9 mm long. In the largest specimen examined, 7.6 mm, the adult complement of dorsal spines (IX), dorsal rays (13-15), and anal fin rays (12) is complete. Pec- toral fin rays begin to form in a larva 6.8 mm NL, and 1 6 pectoral fin rays are countable in a larva 7.6 mm. Pelvic fin buds are first visible at ~6.8 mm NL; however, pelvic fin rays are not yet formed in the largest specimen. SPINATION.— Preopercular spines begin to form in small preflexion larvae of A. Type 3 at ~ 4.1 mm NL. A series of 1 5-1 7 tiny, equal-sized spines is visible along the posterior margin of the pre- opercle in preflexion larvae between 4.1 and 5 mm NL. During development, preopercular spines increase in number ranging from 2 1 to 24 in flexion and early postflexion larvae. In late flexion larvae (~6. 8-6.9 mm NL) the middle 2 or 3 preopercular spines (the 8th- 1 1th spine from the dorsal margin of the preopercle) begin to increase in size relative to other pre- opercular spines. In the 7.6-mm larva, the dor- salmost and ventralmost 1 or 2 spines are also larger than other preopercular spines. This forms the characteristic preopercular spine pattern found in Artedius larvae with multiple preoper- cular spines: the dorsalmost, middle, and ven- tralmost spines are markedly larger than the oth- er preopercular spines. No other spines develop on the head in larvae <7.6 mm. Head spination in larger larvae re- mains unknown. Oligocottus maculosus (Figures 16, 17; Table 5) LITERATURE. —Stein (1 972, 1 973) described O. maculosus larvae and illustrated specimens 4.6, 6.0, 6.6, and 9.2 mm TL. IDENTIFICATION.— Larvae in this series were reared from eggs spawned from known adults. Adults and juveniles were identified by the fol- lowing combination of characters: high vertebral (33-34) and dorsal fin ray (15-18) counts, small size at transformation (8-9 mm), absence of cirri on the nasal spines and along the base of the dorsal fins, and pigmentation. The develop- mental series was linked together primarily on the basis of pigmentation, preopercular and pa- rietal spination, and body shape. Postflexion and transforming larvae were linked to juveniles by the serial method utilizing pigmentation, spi- nation, and size at transformation. DISTINGUISHING FEATURES.— Newly hatched larvae of O. maculosus reared in the laboratory are distinguished by the following pigmentation characters: intense melanistic nape pigment, dark dendritic melanophores that extend onto a prominent bubble of skin in the nape region just anterior to the origin of the dorsal finfold, 1 or 2 melanophores situated anteriorly on the vis- ceral mass beneath the pectoral fins, and a series of 18-36 ventral midline melanophores poste- rior to the anus. In addition to distinctive pig- mentation, larvae possess two rounded humps or protrusions that extend dorsally on either side of the gut just posterior to the pectoral fin bases. These protrusions are similarly positioned and reminiscent of the gut diverticula found in larvae of Artedius; however, they never develop into distinct diverticula. These protrusions disappear at the completion of yolk absorption about five to ten days after hatching. Oligocottus maculosus larvae also possess a distinctive bubble of skin in the nape region just anterior to the origin of the dorsal finfold. This bubble persists in larvae up to 7.5 mm SL. Flexion and postflexion larvae >6.5 mm pos- sess a relatively low number of preopercular spines (9-13). Postflexion larvae and juveniles may be dis- tinguished by meristics, especially the high ver- tebral and dorsal fin ray counts, and the small size at transformation (8-9 mm SL). In addition, juveniles possess a slender postorbital cirrus and two frontoparietal cirri. PIGMENTATION.— Newly hatched larvae of O. maculosus possess no melanistic head pigmen- tation. Fourteen to 16 intense, stellate melano- phores are concentrated in the nape region. One to 3 dendritic melanophores extend anteriorly from the nape pigment patch onto a prominent elevation or bubble of skin located just anterior to the origin of the dorsal finfold. In live larvae, xanthophores cover the bubble of skin and the nape. Three to 4 dendritic, embedded melano- phores are positioned in the otic capsule. The dorsal surface of the gut is darkly pig- 192 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B FIGURE 16. Larvae of Oligocottus maculosus: A) 4.3 mm NL, B) 7.2 mm NL, Q 6.9 mm NL. mented with 100-150 dark melanophores. Two to 4 pale, dendritic melanophores are located along the anteroventral margin of the gut, just beneath the pectoral fins. These melanophores are frequently embedded in the gut musculature and are difficult to see. One to 5 small melano- phores are clustered around the anus. Posterior to the anus, larvae of O. maculosus possess a relatively high number of ventral mid- line melanophores. The actual number of me- lanophores appears to vary with the geographic location at which the larvae were collected. Stein (1973) recorded between 1 1 and 20 ventral mid- line melanophores in his reared larvae, while lar- vae reared in Oregon possessed between 1 6 and 20 melanophores along the ventral midline. J. B. Marliave (Vancouver Public Aquarium, Van- couver, B.C., Canada, pers. comm.) found be- tween 26 and 36 ventral midline melanophores in reared larvae from the Straits of Georgia in British Columbia. Regardless of the number of melanophores, this series begins under the third or fourth myomere posterior to the anus and extends toward the tail tip. The first four mela- nophores in the series are usually spaced one every two to three myomeres, while the remain- der of the melanophores are spaced one per myo- mere. Five or nine additional pigment slashes extend onto the ventral finfold near the tail tip. During larval development in larvae <6 mm, 1 5-20 melanophores form over the midbrain and interorbital region of the head. Two to 5 mela- nophores form on the snout and 1-3 melano- phores form on the cheek just anterior to the WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 193 ; - $j$j$00t& FIGURE 17. Young of Oligocottus maculosus: A) 7.8 mm SL, B) 10.4 mm SL. dorsalmost preopercular spine in larvae > 7 mm. At this size, melanophores are also added in the otic capsule; however, they become obscured by the developing musculature and are difficult to see. By ~6 mm, several melanophores are added just ventral to the nape pigment patch. Five to 7 of the centrally positioned melanophores be- come embedded while the other nape melano- phores form a prominent U-shape anterolater- ally around the central melanophores. During transformation, ~7-8 mm, melanistic pigmentation increases markedly over the dorsal surface of the head. Melanophores are added on the snout, on the cheek region anterior to the preopercle and on the dorsal portion of the oper- culum. Melanophores also form on the pectoral fin base and gradually extend ventrally onto the isthmus. Pigmentation over the dorsal surface of the head is intense in benthic juveniles. A band of melanophores forms on either side of the snout, extending from the upper lip to the ventral mar- gin of each eye. Each band continues posteriorly reaching from the eye to the dorsalmost pre- opercular spine. Melanophores are also added ventrally along the entire margin of the pre- operculum and on the anterior tip of the lower lip. In juveniles >8.5 mm, tiny melanophores cover the entire dorsolateral surface of the head; however, the bands of pigment extending through each eye remain prominent. An irregular band of tiny melanophores forms along the surface of the lateral midline in juveniles >9 mm. This band gradually extends posteriorly to the caudal fin base. Two additional bands of pigment form along the dorsum. A third band forms under the 8th-10th dorsal rays and a fourth band develops under the 14th- 16th dorsal fin rays. These pig- ment bands eventually extend ventrally and fuse with the lateral midline pigmentation. Tiny me- lanophores are added over the dorsolateral body surface in juveniles ~ 1 3 mm; however, the in- tense pigment bands along the dorsum remain distinct. Melanophores extend out onto the dor- sal and caudal fin rays, forming three or five bands of pigment. MORPHOLOGY.— Newly hatched Oligocottus maculosus larvae range in length from 4.2 to 4.5 mm NL. Larvae undergo flexion of the noto- chord at 7.2-7.6 mm NL. Transformation occurs at a relatively small size, ~7.5-8 mm. The small- est benthic juvenile examined was 8 mm long. 194 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 TABLE 5. BODY PROPORTIONS OF LARVAE AND JUVENILES OF OLIGOCOTTVS MACULOSUS AND O. SNYDERI. Values are percent standard length (SL) or head length (HL) including mean, standard deviation, and range in parentheses. Item Oligocottus maculosus Oligocottus snyderi Head length/SL: Preflexion 17.3 ± 0.61 (17.2-18.3) 21.4± 1.26(20.1-23.4) Flexion 19.7 ± 1.84(18.4-21.0) 20.8 ± 1.34(18.9-23.1) Postflexion 25.3 ± 1.89(22.3-26.6) 23.4 ±1.71 (21.4-28.2) Juvenile 30.1 ± 3.06(27.1-32.8) 26.9 ± 1.00(26.5-28.0) Snout length/HL: Preflexion 25.7 ± 2.08 (24.3-27.9) 23.7 ± 3.37(21.2-25.9) Flexion 26.5 ± 1.08(26.0-28.1) 21.4± 2.43(16.9-23.3) Postflexion 29.3 ± 2.83 (24.4-33.0) 16.1 ± 2.87(13.9-20.4) Juvenile 27.1 ± 3.22(24.9-31.4) 32.0 ± 0.00 (32.0-32.0) Eye diameter/HL: Preflexion 55.6 ± 2.08 (53.9-57.2) 47.8 ± 4.80(40.1-52.3) Flexion 44.9 ± 4.24 (42.3-48.2) 42.1 ± 2.95(37.6-46.1) Postflexion 41.2 ± 4.03(33.1-45.0) 33.3 ± 3.21 (31.2-37.9) Juvenile 31.4 ± 2.01 (28.9-34.1) 31.0± 2.52(29.0-34.3) Snout to anus length/SL: Preflexion 39.1 ± 1.53(37.2-40.4) 42.1 ± 1.47(40.1-44.4) Flexion 39.8 ± 2.83(38.1^2.3) 41.8 ± 2.30(39.2-44.9) Postflexion 43.9 ± 3.39(40.8-48.1) 44.5 ± 1.83(42.1-46.4) Juvenile 45.0 ± 2.65 (42.7^8.5) 42.2 ± 1.53(40.3-43.1) Snout to pelvic fin origin/SL: Preflexion _ _ Flexion 24.0* 21.0± 1.00(20.0-22.0) Postflexion 24.9 ± 2.15(23.1-26.9) 23.3 ±1.71 (21.0-25.5) Juvenile 28.1 ± 3.61(25.3-32.1) 24.6 ± 1.53(22.9-26.1) Pelvic fin origin to anus/SL: Preflexion _ _ Flexion 18.0* 20.0 ± 1.15(19.1-21.1) Postflexion 20.2 ± 1.89(18.1-22.8) 21.2 ± 1.50(20.1-23.3) Juvenile 16.9 ± 1.73(16.1-19.4) 18.1 ± 3.21 (16.0-22.1) Body depth at pectoral fin base/SL: Preflexion 19.4 ± 1.00(17.8-20.2) 23.2 ± 1.17(21.8-25.5) Flexion 24.0 ± 0.00 (24.0-24.0) 23.4 ± 2.00 (20.2-26.6) Postflexion 25.9 ± 2.17(23.7-29.1) 20.6 ± 1.50(19.4-21.8) Juvenile 23.9 ± 3.06 (20.6-27.2) 21.2 ± 2.00(19.0-23.4) Body depth at anus/SL: Preflexion 15.5 ± 0.61(14.9-16.2) 21.0± 2.37(18.1-24.9) Flexion 18.0 ± 0.00(18.0-18.0) 24.3 ± 2.91 (17.9-28.1) Postflexion 25.9 ± 2.17(23.7-29.1) 26.2 ± 1.50(21.9-27.3) Juvenile 20.3 ± 2.08(18.1-22.3) 22.1 ± 2.31 (19.0-23.0) Pectoral fin length/SL: Preflexion 10.1 ± 0.61(9.8-11.2) 8.2 ± 2.53(6.1-12.1) Flexion 18.0 ±0.00 (18.0-18.0) 10.1 ± 1.77(9.3-13.4) Postflexion 21.7 ± 2.96(17.3-26.3) 14.4 ± 1.41 (13.1-16.0) Juvenile 29.1 ± 2.65(26.8-31.9) 24.3 ± 1.53(22.9-26.3) - = Not present at this stage. * = Only one specimen available in this stage. Eighteen specimens of O. maculosus (4.3-10.8 mm) were examined for developmental mor- phology. In newly hatched larvae two prominent bumps or protrusions appear on the dorsal surface of the gut just posterior to the pectoral fin base. These bulges are similar to the dorsal gut diver- ticula of Artedius larvae; however, they never WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 195 develop into distinct diverticula. The gut pro- trusions disappear approximately 5-10 days af- ter hatching. Oligocottus maculosus larvae also possess a distinctive bubble of skin in the nape region just anterior to the origin of the dorsal finfold. This bubble persists in larvae <7.5 mm. Small O. maculosus are slender with a rela- tively short gut. Snout to anus length averages 39% in preflexion larvae and increases to 44% in postflexion larvae and 45% SL in juveniles. Body depth at both the pectoral fin base and the anus increases during larval development from 19% and 15%, respectively, to 25% SL. FIN DEVELOPMENT.— Larval Oligocottus mac- ulosus begin to undergo notochord flexion at ~6- 7 mm. The adult complement of 12 principal caudal rays is complete in larvae ~6.8-7 mm long at about the completion of notochord flex- ion. Dorsal and anal fin rays begin to form in larvae ~6.6 mm long; however, the full complement of fin rays (15-18, 12-14, respectively) is not com- plete until larvae are 8 mm long. Dorsal spines (VIII-IX) also form between 7 and 8 mm. Al- though pectoral fin rays are visible in larvae by 6.6 mm, the adult complement (12-15) is not fully formed until larvae reach about 7.6 mm. Pelvic buds are first visible in 7 mm larvae but the fin rays are not formed until ~8.5 mm. SPINATION.— Six to 7 tiny spines are first vis- ible on the posterior margin of the preopercle in larvae ~ 5.8 mm long. Spines increase in number to 9 or 10 in larvae undergoing notochord flex- ion. In postflexion larvae 6.9-7.8 mm long, pre- opercular spines number 10 or 1 1. Two or 3 of these spines appear as tiny accessory spines that form just anterior to the bases of the other spines. The dorsalmost spine becomes slightly larger than the lower spines. The 3rd, 4th, and 5th preoper- cular spines also increase slightly in size relative to the lower spines. In the largest planktonic lar- vae (~8 mm long) the preopercular spines begin to decrease in size and number and are covered with skin. In newly settled benthic juveniles ~8- 10 mm long, the dorsalmost spine is quite large and stout with a strong upward curvature. The lower spines persist only as three blunt, bony protrusions on the preopercular margin. Three tiny spines also form in the parietal re- gion in larvae ~6-7 mm long. Two spines de- velop anteriorly with a third nuchal spine form- ing just posterior to them. These parietal spines persist through the larval period, but they regress in benthic juveniles. During regression, the an- terior spines decrease in size and their tips bend posteriorly and fuse with the nuchal spine, form- ing an arch. This canal and arch become incor- porated into the cephalic lateral line system. Two spines also form on the posttemporal in larvae ~6-7 mm long. By ~7-8 mm, a third spine forms on the posttemporal and a fourth spine forms on the supracleithrum. These su- pracleithral-posttemporal spines persist through larval development and eventually form the junction of the cephalic and lateral line systems. Oligocottus snyderi (Figures 18, 19; Table 5) LITERATURE.— Stein (1973) described and il- lustrated 4.5- and 5. 5 -mm TL larvae of 0. sny- deri. Richardson and Washington (1980) called these larvae Cottidae Type 1 and illustrated spec- imens 4.2, 6.7, and 9 mm long. IDENTIFICATION.— Small larvae in this series were reared from eggs spawned from known adults. Adults and juveniles were identified by the following combination of characters: high vertebral (34-37) and dorsal fin ray (17-20) counts, light pigmentation, the presence of cirri on the nasal spines and along the bases of the dorsal fins, and the absence of scales (prickles). The developmental series was linked together primarily on the basis of pigmentation, body shape, and preopercular and parietal spination. Postflexion and transforming larvae were linked to juveniles by pigmentation, meristics, and pre- opercular and parietal spination. DISTINGUISHING FEATURES. — Distinguishing pigmentation of preflexion larval O. snyderi are melanistic nape pigmentation, relatively light pigmentation over the dorsolateral surfaces of the gut, and a low number of ventral midline melanophores (5-9) situated posterior to the anus. This series of ventral midline melanophores originates beneath the fifth to seventh postanal myomeres and extends posteriorly toward the tail tip. One melanophore is spaced approxi- mately every four or five myomeres. This char- acteristic pigmentation changes little during lar- val development. In newly hatched larvae, a hump or bubble of skin is present just anterior to the origin of the dorsal finfold. Although diffuse xanthophores are present over this bump in laboratory-reared lar- vae, no melanophores extend onto this bubble of skin. In contrast, O. maculosus larvae, which 196 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B FIGURE 18. Larvae ofOligocottus snyderi: A) 4.7 mm NL, B) 5. 1 mm NL, C) 6.7 mm NL (C from Richardson and Washington 1980). also possess this distinctive bubble of skin at the nape, have one to three large dendritic mela- nophores that extend up onto the bubble of skin from the nape pigment patch. Larvae of O. snyderi >4.2 mm NL may be further distinguished from all other known cottid larvae by the presence of a cluster of 10-20 mi- nute prickles situated in the parietal region of the head. Larvae undergoing notochord flexion, > 6 mm long, possess a distinctive pattern of multiple preopercular spination, in which approximately 15 equal-sized spines are positioned along the posterior margin of the preopercle. Ten to 1 1 small, accessory spines are situated at the ante- rior bases of the other spines and point antero- laterally. Postflexion larvae and juveniles may be dis- tinguished by their relatively light pigmentation, the prominent bands of pigment through the eye, and the low number of widely spaced ventral midline melanophores. In addition to pigmen- tation, juvenile O. snyderi are characterized by high vertebral and dorsal fin ray counts (34-37 and 1 7-20, respectively), and by the presence of very long, slender nasal, postorbital, and fron- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 197 FIGURE 19. Young of Oligocottus snyderi: A) 8.2 mm SL, B) 10.2 mm SL, C) 14.4 mm SL. toparietal cirri. In juveniles >\5 mm, a row of distinctive cirri develop along either side of the bases of the dorsal fins. PIGMENTATION.— Newly hatched, reared lar- val O. snyderi are lightly pigmented with no mel- anistic head pigmentation. Several xanthophores are situated over the midbrain in live larvae. Two to 8 external melanophores are clustered over the notochord in the nape region. In live specimens several diffuse xanthophores extend dorsally from the nape onto a distinctive bump or bubble of skin just anterior to the origin of the dorsal unfold. In contrast to larvae of O. maculosus, however, melanophores never ex- tend onto this bubble of skin in larvae of O. snyderi. The dorsolateral surface of the gut is lightly pigmented with 50-60 small melano- phores forming an elliptical patch over the body cavity. Intense xanthophores also cover the dor- solateral surfaces of the gut. The only pigmen- tation posterior to the anus consists of a series of 5-9 melanophores that originates under the fifth to seventh postanal myomere and extends posteriorly. Each melanophore is positioned four to six myomeres apart. Occasionally, one or two melanistic pigment slashes extend onto the cau- dal finfold just beneath the notochord tip. During larval development, several melano- phores form on the dorsal surface of the head. Size of larvae at formation of this melanistic pig- 198 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 mentation appears somewhat variable. One or two melanophores are present over the midbrain in 21 -day-old, laboratory-reared larvae (~6.5 mm long). Morris (n.d.) reported that four or five melanophores develop in the midbrain region of larvae at four weeks of age (~8.5 mm). Mela- nistic pigmentation does not appear over the brain in field-caught larvae until about 10 mm. During transformation, head pigmentation in- creases markedly in late postflexion larvae 10- 14 mm long. Thirteen to 18 large, stellate me- lanophores form over the midbrain and inter- orbital regions of the head. Concurrently, several small melanophores form anterior to the orbit and extend anteriorly across the snout onto the upper lip forming a distinct band. Several intense melanophores develop posterior to the orbit forming a dark band extending from the orbit to the dorsalmost preopercular spine. Melano- phores are also added at the posterior margin of the lower jaw just ventral to the preopercle, along the dorsal margin of the operculum, and along the pectoral fin rays. A row of intense, embedded melanophores forms just above the spinal cord and extends posteriorly from the nape region two- thirds of the way to the caudal fin. The ventral midline melanophores remain unchanged. In juveniles between 1 3 and 1 5 mm long, nu- merous tiny melanophores form over the dor- solateral surfaces of the head and extend ante- riorly onto the snout between the eyes, ventrally along the preopercle, and along the opercular margin. These diffuse melanophores extend pos- teriorly to the seventh dorsal spine. Numerous small melanophores also form posterior to the pectoral fin base in an irregular band of pigment along the lateral midline. With development, melanophores are added along the dorsum in an anterior to posterior sequence. Concurrently, melanophores extend dorsally onto the dorsal fins forming four to five distinct bands of pig- ment. Gradually, the melanophores along the dorsal midline extend ventrally and posteriorly and join the dorsal and lateral areas of pigmen- tation. This lateral pigmentation extends poste- riorly and forms a dark band at the base of the caudal fin. Melanophores extend onto the caudal fin rays forming four to five bands of pigment. Juvenile O. snyderi are characterized by uniform diffuse pigmentation over the head and dorso- lateral surfaces of the body with a distinct, dark band of pigment extending from the snout, through the orbit, to the dorsalmost preopercular spine. The characteristic low number of widely spaced ventral midline melanophores remains visible in juveniles < 1 8 mm. MORPHOLOGY.— Newly hatched O. snyderi larvae range in size from 4 to 4.5 mm NL. No- tochord flexion occurs between 6.2 and 8.4 mm NL. The largest planktonic specimen taken in the field is 10.2 mm and has not yet begun to undergo transformation. The smallest benthic juvenile examined is 12.4 mm. Twenty-four specimens, ranging in length from 4 to 1 5. 1 mm, were examined for development morphology. Newly hatched O. snyderi larvae are rather slender with a relatively short gut, the poste- riormost portion of which trails well below the body. Snout to anus length averages 42% in pre- flexion and flexion larvae, then increases slightly to 44% SL in postflexion larvae. Relative body depth at the pectoral fin base increases from 23% to 25% SL during larval development. A small, rounded protrusion extends dorsally from the dorsal surface of the gut just posterior to the pectoral fin base in newly hatched larvae. This protrusion is reminiscent of the gut diverticula found in larvae of several species ofArtedius but is much less pronounced and never develops into distinct diverticula. This protrusion decreases in size shortly after hatching and is no longer visible by yolk absorption five days after hatching. In addition, O. snyderi larvae possess a prominent bump or bubble of skin that protrudes dorsally in the nape region just anterior to the origin of the dorsal finfold. This bubble persists in larvae up to ~6.5-7 mm. FIN DEVELOPMENT.— Larvae of O. snyderi undergo notochord flexion between 6.2 and 8.4 mm. Caudal fin rays first appear at 7.8 mm; how- ever, the full adult complement of 6 + 6 prin- cipal caudal rays is not complete until larvae reach ~9-10 mm. Rays begin to form in the dorsal and anal fins of larvae between 7.5 and 8 mm long; however, these rays are not fully formed in larvae <9 mm. Adult complements are 17- 20 and 12-15, respectively. Dorsal fin spines be- gin forming in larvae 9-10 mm long, and the adult complement of spines (VIII-IX) is count- able in a 10.2 mm specimen. Pectoral fin rays (12-15) form at 9 mm and are complete by 10 mm. Pelvic buds are first visible in larvae be- tween 8.2 and 9 mm, but the fin rays are not fully formed until larvae reach 10-12 mm. SPINATION.— Five to nine tiny bumps form along the posterior margin of the preopercle WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 199 in larvae >4.2-5 mm. By ~ 5.1 mm, 10-15 tiny equal-sized spines are visible. During notochord flexion the preopercular spines increase in size and number, ranging between 17 and 22. The preopercular spines of O. snyderi larvae are unique in that 10-12 spines form along the pos- terior margin of the preopercle as in other cottids with preopercular spines, yet by ~7 mm NL, between 8 and 10 small accessory spines form anteriorly at the bases of the original spines. In larvae > 9 mm, the dorsalmost preopercular spine becomes stouter and longer than the other spines and is separated from the lower spines by a short gap on the preopercular margin. The 5-8 spines just ventral to the dorsalmost spine also become slightly larger relative to the lower preopercular spines. Between 12 and 14 preopercular spines are visible in newly settled benthic juveniles. The dorsalmost spine becomes much larger relative to the other spines. The smaller, accessory spines begin to atrophy and are represented only by small bumps or irregularities on the preopercle. By ~ 14 mm, only the dorsalmost spine persists. Distinctive spines also form in the parietal re- gion of the head of young O. snyderi. Larvae as small as 4.2 mm have 7-10 small bumps or prickles visible over the parietals. These prickles increase in number during development; 10-20 prickles are present in the parietal region in lar- vae >6.2 mm. Eight to 12 tiny prickles remain visible along the posterolateral margin of the pa- rietal bones in 12-13-mm cleared and stained benthic juveniles. A cluster of spines also develops in the supra- cleithral-posttemporal region in larvae > 8 mm. One spine forms on the supracleithrum and five spines, situated in two rows, form on the dorsal portion of the posttemporal. These persist throughout larval development but atrophy dur- ing transformation until only three bony projec- tions are present in benthic juveniles. These bony projections represent the rudiments of the incip- ient lateral line system. Clinocottus acuticeps (Figures 20-22; Table 6) LITERATURE. — Blackburn (1973) illustrated and described an 8.6-mm specimen, which he called Cottid 1 "Biramous anus." Richardson (1977) and Richardson and Pearcy (1977) listed larvae of C. acuticeps as Cottidae sp. 12. Larvae of this species were described by Richardson and Wash- ington (1980). They illustrated specimens 3.7, 3.9, 6.9, 7.6, 10.4, 13.8, and 16.5 mm long. IDENTIFICATION.— Small larval C. acuticeps were reared from eggs spawned from known adults. Adults and juveniles were identified by low dorsal fin ray (13-17) and anal fin ray (9- 13) counts, the presence of nasal cirri, and a membrane connecting the innermost pelvic fin ray with the abdomen. The developmental series was linked together primarily by pigmentation, body shape, and hindgut diverticula. Postflexion and transforming larvae were linked with juve- niles by pigmentation, meristics, and the mem- brane attaching the pelvic fin rays to the abdo- men. DISTINGUISHING FEATURES. — Clinocottus acu- ticeps larvae are distinguished from all other known cottid larvae by long protrusions (diver- ticula) that extend posteriorly from the gut on either side of the anus. These diverticula are pres- ent in yolk-sac larvae and persist in the largest pelagic specimens (14.5 mm). The gut itself is distinctively long and the posterior portion trails well below the body. Snout to anus length, av- eraging 62.5% SL, is greater than in other known larvae of Artedius, Clinocottus, or Oligocottus. In addition, these larvae have a flabby appear- ance with an outer bubble of skin, which is es- pecially pronounced in the head region. Other characters useful in distinguishing C. acuticeps larvae are melanistic pigmentation on the snout and head, and relatively few ventral midline melanophores (4-10). Transforming and juvenile C. acuticeps are distinguishable from all other known cottids by the presence of a membrane attaching the inner pelvic fin ray to the belly. Other characters useful in separating juveniles are the relatively light, uniform pigmentation over the body; a band of pigment extending from the snout posteriorly through the orbit toward the preopercle; a dark blotch of pigment at the anterior end of the spinous dorsal; and a low number of ventral mid- line melanophores. PIGMENTATION. —Newly hatched larvae reared in the laboratory exhibit 4 or 5 dendritic mela- nophores on the snout and 2 faint melanophores in each otic capsule. In field-collected larvae <3.7 mm NL, the presence of snout pigment varies; however, all larvae >3.7 mm NL possess at least 2 melanophores on the snout. Eight to 15 me- lanophores are clustered in the nape region of even the smallest larvae. Numerous melano- 200 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 FIGURE 20. Larvae of Clinocottus acuticeps: A) 3.7 mm NL, B) 3.7 mm NL, C) 3.9 mm NL, D) 6.9 mm NL (from Richardson and Washington 1980). WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 201 FIGURE 2 1 . Larvae ofClinocottus acuticeps: A) 7.6 mm SL, B) 1 0.4 mm SL, C) 1 3.8 mm SL (from Richardson and Washington 1980). phores are scattered over the dorsolateral surface of the gut extending posterolaterally over the sur- face of the gut diverticula. These melanophores are much fainter and more irregular in shape than in larvae of other species of Clinocottus. A series of 4-1 0 inconspicuous ventral midline melanophores originates beneath the 7th- 10th postanal myomeres and extends posteriorly to- ward the tail tip. Several additional melano- phores appear as streaks of pigment on the ven- tral finfold near the tail tip. Pigmentation increases on the head during lar- val development. Melanophores form first on the head over the midbrain in larvae 5.5 mm NL. Concurrently, several embedded melanophores appear on the nape and extend anteriorly onto the head. Four to five internal melanophores oc- cur in or near the otic capsule. In larvae >6.5 mm, scattered melanophores extend continu- ously from the snout to the nape region. Ventral midline melanophores persist in flexion and postflexion larvae, and the posteriormost mela- 202 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 FIGURE 22. Young of Clinocottus acuticeps: A) 13.8 mm SL, B) 16.5 mm SL (B from Richardson and Washington 1980). nophore, which is located near the notochord tip in small larvae, occurs near the middle of the caudal fin base between the forming hypural plates. The posteriormost melanophores which extended onto the ventral finfold now occur on the caudal fin. In the largest planktonic larvae, which are beginning to undergo transformation, melanophores are added in a patch just posterior to the orbit. Melanophores also are added along the bases of the pectoral fins and extend onto the pectoral fin rays. In newly settled benthic juveniles (13-14 mm) head pigmentation increases markedly. Mela- nophores extend anteriorly across the dorsal sur- face of the snout and onto the upper lip. The melanophores at the ventral margin of the snout are especially intense and closely spaced, forming a prominent band that extends from the upper lip to the ventral margin of the orbit. This band continues from the posterior margin of the orbit to the dorsal margin of the preopercle. Addi- tional melanophores are added along the ventral edge of the lower lip, at the base of the preopercle, and on the dorsal portion of the operculum. As development proceeds, a second band of pig- ment forms between the ventral margin of the orbit and the melanophores at the base of the preopercle. Simultaneously, pigmentation in- creases on the pectoral fin bases while two to three bands of pigment form across each fin. Between 14 and 15 mm, a dense patch of me- lanophores forms at the anterior end of the first dorsal fin between the first and third spines. As juvenile pigmentation progresses this patch ex- pands posteriorly to include the fourth dorsal spine, and a second patch of melanophores forms between the seventh and eighth dorsal spines. Melanophores extend ventrally from these two pigment patches forming two distinct bands across the dorsum. Pigmentation proceeds pos- teriorly along the dorsum. In juveniles between 15 and 16 mm long, a third band (or saddle) of pigment forms under the second to sixth dorsal fin rays; a fourth band forms under the 8th to 1 1th dorsal fin rays; and a fifth band forms under the last two dorsal fin rays. As these bands of pigment develop along the dorsum, they extend ventrally and eventually unite into a uniform band of pigment above the lateral midline. Con- currently, another band of pigment extends pos- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 203 TABLE 6. BODY PROPORTIONS OF LARVAE AND JUVENILES OF CLINOCOTTVS ACUTICEPS, C. EMBRYUM, C. GLOBICEPS, AND C. ANALIS. Values given are percent standard length (SL) or head length (HL) including mean, standard deviation, and range in parentheses. Item Clinocottus acuticeps Clinocottus embryum Clinocottus globiceps Clinocottus analis Head length/SL: Preflexion 27.1 ± 3.44(22.3-30.1) Flexion 29.2 ± 1.92(26.8-32.4) Postflexion 27.9 ± 2.71 (22.9-32.6) Juvenile 32.2 ± 0.65 (31.5-32.7) Snout length/HL: Preflexion 24.6 ± 4.80(21.3-25.9) Flexion 26.4 ± 2.97 (22.9-30.6) Postflexion 23.6 ± 4.38 (16.3-30.9) Juvenile 27.5 ± 4.36 (26.8-29.4) Eye diameter/HL: Preflexion 39.7 ± 6.63 (31.5-54.2) Flexion 34.6 ± 2.41 (31.9-36.4) Postflexion 32.7 ± 3.30(27.1-38.2) Juvenile 27.9 ± 2.06 (25.5-29.3) Snout to anus length/SL: Preflexion 60.7 ± 4.64 (54.4-67.2) Flexion 62.8 ± 2.74(60.3-67.1) Postflexion 62.5 ± 4.61 (57.5-70.3) Juvenile 50.2 ± 1.64(48.4-51.6) Snout to pelvic fin origin/SL: Preflexion Flexion Postflexion 33.4 ± 2.63 (29.4-39.5) Juvenile 33.3 ± 1.16 (32.5-34.6) Pelvic fin origin to anus/SL: Preflexion Flexion Postflexion 29.4 ± 4.56 (23.1-37.3) Juvenile 16.9 ± 1.70(15.9-18.9) Body depth at pectoral fin base/SL: Preflexion 24.3 ± 3.25 (17.8-29.1) Flexion 27.6 ± 2.30 (24.3-30.2) Postflexion 31.3 ± 2.29 (28.4-35.0) Juvenile 26.1 ± 1.99 (23.9-27.8) Body depth at anus/SL: Preflexion 21.4 ± 2.94(18.0-24.5) Flexion 25.4 ± 2.07 (23.3-28.1) Postflexion 28.4 ± 2.43 (26.4-35.2) Juvenile 24.0 ± 2.28(21.4-25.4) Pectoral fin length/SL: Preflexion 1 1.4 ± 1.27 (9.6-13.3) Flexion 1 1.0 ± 1.87 (9.9-14.5) Postflexion 26.4 ± 5.95 (18.2-35.0) Juvenile 32.1 ± 1.74(30.8-34.1) 26.1 ± 2.65(26.0-30.1) 23.9 ± 1.15(21.8-25.3) 24.0 ± 0.00 (24.0-24.0) 31.5 ± 1.13(30.2-32.2) 21.3 ± 1.53(19.9-23.7) 24.1 ± 3.83(18.4-30.8) 22.2 ±0.58(21.8-23.6) 27.5 ± 1.47(22.7-31.2) 39.3 ± 3.06(35.8-42.1) 35.7 ± 3.79 (34.8-39.5) 32.1 ± 3.06(29.0-35.3) 29.8 ± 0.74 (29.2-30.6) 51.6 ± 3.51(48.2-55.1) 49.9 ± 3.87 (43.9-54.4) 50.0 ± 3.06 (47.3-53.8) 49.0 ± 3.10(47.0-52.6) 28.0* 32.1 ± 3.54(29.0-34.2) 31.1 ± 1.31(29.6-32.1) 26.1* 18.2 ± 0.71 (17.1-18.4) 17.9 ± 2.50(15.4-20.4) 25.8 ± 4.36(21.2-29.4) 26.3 ± 2.29 (23.7-30.2) 26.1 ± 1.53(25.0-28.2) 21.6 ± 1.01 (23.5-25.5) 22.8 ± 4.36(18.2-26.5) 25.1 ± 1.51(22.9-27.3) 27.0 ± 0.00 (27.0-27.0) 23.3 ± 1.93(22.1-25.5) 9.9 ± 3.46(6.3-12.1) 11.2 ± 2.75(7.1-17.5) 32.0 ± 4.04(30.1-37.0) 33.4 ± 1.16(32.1-34.2) 20.6 ± 2.21 (17.0-25.0) 23.0 ± 1.79(21.0-27.2) 27.3 ± 4.09 (22.4-33.3) 30.1 ± 1.25 (27.8-32.1) 31.6 ± 0.46(31.2-32.1) 21.3 ± 4.40(14.5-26.7) 25.7 ± 3.39 (20.0-30.5) 25.3 ± 4.84 (20.0-33.0) 28.3 ± 2.70 (22.3-32.2) 27.1 ± 0.92(26.1-27.9) 50.4 ± 9.16(46.2-63.1) 43.7 ± 7.39(37.9-51.4) 38.2 ± 8.09 (24.4-46.0) 31.3 ± 1.41 (27.8-33.0) 27.1 ± 0.91(26.1-27.9) 44.0 ± 3.59 (39.6-52.9) 48.2 ± 3.27 (44.4-56.0) 50.0 ± 3.48 (44.7-56.8) 48.9 ± 2.40 (46.4-54.3) 24.6 ± 2.94(21.4-28.0) 26.7 ± 2.33 (23.4-30.9) 30.7 ± 1.04(29.5-31.4) 21.5 ± 1.77(18.7-23.4) 22.9 ± 1.42(21.4-25.2) 20.7 ± 0.64(20.0-21.2) 20.8 ± 2.87(15.8-25.5) 23.9 ± 2.02 (22.2-29.3) 26.8 ± 2.78(22.1-30.9) 27.3 ± 0.53 (26.7-27.7) 17.5 ± 2.46(13.6-21.6) 21.2 ± 2.07(17.7-24.0) 25.7 ± 3.04(21.1-30.4) 25.7 ± 2.08 (23.3-27.0) 12.1 ± 1.19(9.6-13.8) 12.3 ± 4.40(7.5-22.0) 22.7 ± 6.00(11.1-30.4) 29.0 ± 0.82(28.1-29.7) 29.3 ± 1.34(27.2-31.4) 19.9 ± 2.26(17.3-22.3) 28.1 ± 1.36(25.4-29.1) 25.7 ± 1.50(25.0-29.1) 29.3 ± 2.15(24.8-32.3) - = Not present at this stage. * = Only one specimen available in this stage. 204 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 teriorly along the lateral midline to the caudal fin where melanophores form a dark band at the base of the caudal fin. Melanophores extend onto the caudal fin rays where they form four or five bands across the fin. Clusters of melanophores extend first ventrally and then laterally from the lateral midline pigment band and gradually unite enclosing three to five small unpigmented circles below the lateral line. Three to six tiny mela- nophores remain visible along the ventral mid- line posterior to the anus in juveniles < 1 7 mm long. MORPHOLOGY. — Clinocottus acuticeps larvae hatch at the smallest length (3.1-3.3 mm NL) of any member of the genus. Notochord flexion oc- curs between 5.6 and 7.3 mm NL). The largest planktonic specimen collected is 14.5 mm and is beginning to undergo transformation. The smallest benthic juvenile examined is 12.6 mm. Thirty-seven selected specimens of C. acuticeps, ranging in length from 3.1-16.2 mm, were ex- amined for developmental morphology. Larvae of C. acuticeps have a distinctive, flab- by appearance as if a loose bubble of outer skin surrounds the anterior part of the body. Larvae are deep-bodied with a long, distinc- tive gut, the posterior portion of which trails well below the body. Snout to anus length remains relatively constant during larval development, averaging 63% SL. Prominent diverticula extend posteroventrally from the hindgut on either side of the anus. These diverticula are well developed throughout larval development but are not vis- ible in benthic juveniles. FIN DEVELOPMENT.— A 5.6 mm NL larva is just beginning notochord flexion and a concur- rent thickening of the hypural region of the form- ing caudal fin. The adult complement of 6 + 6 principal caudal rays is present in a 6.8 mm spec- imen prior to completion of notochord flexion at ~7.5 mm. Bases of the dorsal and anal fin rays begin forming on a 6.9 mm larva. The full adult complement of dorsal (13-17) and anal (9-13) fin rays is present by ~8 mm. The adult com- plement of dorsal fin spines (VII-IX), however, is not present until ~8.7 mm. Although pectoral fin rays are visible on a 6.9 mm larva, the full adult complement (13-15) is not complete until >7.6 mm. Pelvic fin buds appear just after completion of notochord flexion in a 7.6 mm larva; however, the fin rays are not fully formed until ~ 10 mm. SPINATION.— Preopercular spines first appear as small bumps at 5.2 mm NL. Nine to 1 1 small spines are present by the onset of notochord flex- ion at ~6 mm. During flexion, spines remain small and evenly spaced with the 2nd and 3rd spines becoming slightly longer than the others. By completion of flexion, at 7.6 mm, 1 1-1 2 spines are present along the margin of the preopercle. The dorsalmost 3 spines are beginning to elon- gate and point dorsally. In a 10-mm cleared and stained specimen, the dorsalmost 3 spines are nearly four times as long as the ventral spines. In the largest planktonic larvae (13-14 mm long) the ventralmost spines are beginning to atrophy. The 3 dorsalmost spines are still prominent in a 15.2 mm juvenile, but the 8 ventral spines are minuscule, with their tips twisted and bent an- teriorly. By ~ 1 9 mm, the lower spines have atro- phied completely, and only the single large dor- salmost spine persists. No spines develop in the parietal or supra- cleithral-posttemporal regions of the head in lar- vae or juveniles of this species. Clinocottus embryum (Figures 23-25; Table 6) LITERATURE.— Richardson (1977) and Rich- ardson and Pearcy (1977) listed larvae of this species as Cottidae sp. 20. Richardson and Washington (1980) described these larvae as Cottidae Type 2 and illustrated specimens 4.0, 6.4, and 7.4 mm long. IDENTIFICATION.— Juveniles and adults were identified using a combination of the following characters: an advanced anus, light pigmenta- tion, presence of a nasal cirrus, low anal fin ray counts (9-12), and absence of a membrane at- taching the pelvic fin rays to the abdomen. The developmental series was linked together pri- marily on the basis of pigmentation, body shape, and preopercular spination. Postflexion and transforming larvae were linked to juveniles by pigmentation, cirri patterns, meristics, and pre- opercular spination. DISTINGUISHING FEATURES.— Characters use- ful in distinguishing preflexion larvae of C. em- bryum are lack of head pigment, relatively light gut pigmentation, large number of ventral mid- line melanophores (15-21), and relatively long, trailing gut. Head and/or snout pigment is pres- ent in larvae of all other species of Clinocottus. C. embryum larvae are further distinguished from WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 205 yolk-sac C. acuticeps larvae (in which snout pig- ment is sometimes absent) by the absence of dis- tinct hindgut diverticula. In addition to the pigmentation characters mentioned above, flexion and postflexion larvae of C. embryum are distinguished by their head spination. Larvae 7.3 mm have 11-14 preoper- cular spines; the dorsalmost spine is the largest. In benthic juveniles of C. embryum, the anus is advanced midway between the origin of the pelvic fins and the anal fin, as in other members of the genus. Juvenile C. embryum are distin- guished from C. globiceps, C. analis, and C. re- calvus by relatively light, mottled body pigmen- tation and long and slender nasal, postorbital, and frontoparietal cirri. Clinocottus embryum juveniles are distinguished from C. acuticeps by the presence of a large number of ventral midline melanophores (15-21) and the absence of a membrane connecting the inner pelvic fin ray to the abdomen. PIGMENTATION.— Melanistic pigmentation is absent on the head in preflexion C. embryum larvae. One to 5 melanophores are scattered over the nape region. The dorsolateral surface of the gut is relatively lightly pigmented and occasion- ally several faint melanophores are present on the anterolateral surface of the gut below the pec- toral fins. Posterior to the anus, a series of 1 5- 19 melanophores extends along the ventral mid- line. This series begins on the fourth or fifth myo- mere posterior to a vertical line through the anus; the melanophores are spaced approximately 1 per myomere. Several specimens have 1 or 2 melanophores on the ventral finfold near the no- tochord tip. During larval development, the formation of head pigmentation varies in larvae between 6 and 9 mm long. Three out of eight larvae ob- served possess one to five tiny melanophores over the brain. One or two melanophores are consis- tently present beneath the pectoral fin on the anterolateral surface of the gut in larvae >6.5 mm. Otherwise, pigmentation remains un- changed. In transforming larvae >9.6 mm long, nu- merous melanophores appear over the brain. Several melanophores appear on the cheek re- gion between the orbit and the preopercle. Me- lanophores are also added on the pectoral fin base. Melanistic pigmentation increases over the head in newly settled juveniles. Large melano- phores cover the surfaces of the head over the midbrain and interorbital regions. Several large, intense melanophores are embedded at the pos- terior margin of the parietal region. A distinct, dense band of melanophores extends from the orbit anteriorly onto the upper lip and posteriorly from the orbit to the dorsal tip of the preopercle. Several melanophores form a dark patch on the cheek beneath the orbit. Melanophores are also added to the dorsal surface of the operculum and to the pectoral fin base with several melano- phores extending onto the pectoral fin rays. As development proceeds, pigmentation in- creases markedly over the head. In a 16-mm juvenile, the bands of pigment extending through the eye are prominent, but numerous small me- lanophores cover the entire dorsal surface of the head above these bands of pigment. Pigmenta- tion increases on the operculum and pectoral fin base. Three to four bands of melanophores are added across the pectoral fin rays. Five bands of pigment develop on the body along the dorsal midline in an anterior to pos- terior sequence. The first band of pigment forms under the third to fifth dorsal fin spines, and a second smaller band begins to form under the seventh to ninth dorsal spines in juveniles be- tween 13 and 14 mm long. By ~15 mm, three additional bands of pigment are present on the dorsum beneath the second dorsal fin. The third band forms under the 2nd-4th dorsal fin rays, the fourth band forms under the 7th-9th fin rays, and the fifth band forms under the 12th-15th rays. At the same time, a series of embedded melanophores develops in a row just above the notochord, extending from the nape region to- ward the caudal fin. A few diffuse patches of external melanophores also form along the lat- eral midline posterior to the gut. As juvenile pigmentation develops, the dorsal bands of pigment extend ventrally until they unite above the lateral line, forming four unpigmented saddles between the bands. The melanophores lying along the lateral midline increase in number and extend posteriorly and ventrally toward the caudal fin. As this lateral pigmentation extends posteriorly, it fuses dorsally with the pigment bands. As pigmentation expands and unites over the lateral surface of the body, numerous, irreg- ular, unpigmented circles remain above and be- low the lateral line, giving juvenile C. embryum 206 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B FIGURE 23. Larvae of Clinocottus embryum: A) 4.0 mm NL, B) 5.4 mm NL, Q 6.4 mm NL (A and B from Richardson and Washington 1980). a distinctively mottled appearance. Eighteen to 2 1 small melanophores remain visible along the ventral midline in juveniles up to ~ 1 9 mm long. MORPHOLOGY.— The smallest C. embryum examined is 4.0 mm NL and is recently hatched. Larvae undergo flexion of the notochord between 6.4 and 9.6 mm NL. The largest specimen taken in the plankton is 14.0 mm and is beginning to undergo transformation. The smallest benthic juvenile is 13.7 mm. Eighteen C. embryum, ranging from 4 to 14 mm long, were examined for developmental morphology. Larval C embryum have a distinctively shaped gut with the posterior portion trailing well below WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 207 B FIGURE 24. Larvae of Clinocottus embryum: A) 7.4 mm SL, B) 9.6 mm SL, Q 13.9 mm SL (A from Richardson and Washington 1980). the body. The walls of the hindgut protrude on either side of the anus, reminiscent of the hindgut diverticula of C acuticeps; however, these bulges never develop into pronounced diverticula. Snout to anus length is relatively long throughout larval development, averaging 50% SL. FIN DEVELOPMENT.— The onset of notochord flexion is first apparent in a 6.4 mm larva. Caudal rays are present by ~7.4 mm but the adult com- plement of 6 + 6 principal caudal rays is not complete until about 8.4 mm. Bases of the form- ing dorsal and anal fin rays are first visible at ~7.4 mm; however, the adult complement of dorsal (14-17) and anal (9-12) fin rays is not present until ~8.3 mm. Dorsal spines (VIII-X) are beginning to form at ~8.3 mm but are not fully formed until 9.6 mm. Pectoral fin rays begin to form at ~8 mm, and the adult complement of fin rays (12-15) is present by 9.6 mm. Pelvic fin buds are first apparent at ~9.6 mm, and the adult pelvic fin complement (1,3) is present in postflexion larvae > 12.4 mm long. SPINATION.— Eight to 10 tiny, evenly spaced spines increases, ranging in number from 1 1 to 1 4. opercle at ~5.2 mm NL. In larvae undergoing notochord flexion, the number of preopercular spines increases, ranging in number from 1 1 to 14. During the flexion stage, the dorsalmost pre- opercular spine increases in size relative to the rest of the preopercular spines so that by the end 208 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 FIGURE 25. Juveniles of Clinocottus embryum: A) 13.7 mm SL, B) 16.2 mm SL. of flexion, the dorsalmost spine is much longer and stouter than the other spines. In the largest planktonic larvae, 13-14 mm long, the upper spine is >2.5 times longer than the other spines and is separated from them by a slight gap. In newly settled benthic juveniles, the number and size of the lower preopercular spines are reduced. By > 1 5 mm, only a second, tiny spine persists just ventral to the large dorsal spine. The other spines appear as five to seven small bumps or irregularities along the preopercular margin. In completely transformed juveniles > 1 6 mm long, only the uppermost spine is visible. Two spines develop in the parietal region of C. embryum larvae. A single small spine is first present at the posterior margin of each parietal at ~6.7 mm. By 9.6 mm, this parietal spine has increased in size and a second smaller parietal spine is present just behind it. As larvae undergo transformation, between 12 and 14 mm, these spines undergo a reduction in size, and the first parietal spine eventually fuses with the second parietal spine, forming a hollow central canal between the spines. This canal is part of the in- cipient cephalic lateral line system. In newly set- tled juveniles, only a skin-covered bony protu- berance is visible in the parietal region. Three spines also form in the supracleithral- posttemporal region of the head at ~9.6 mm. These spines persist through transformation and eventually become associated with the lateral line system in juveniles > 1 5 mm long. Clinocottus globiceps (Figures 26-28; Table 6) LITERATURE.— Larvae of this species were list- ed by Richardson (1977) and Richardson and Pearcy (1977) as Oligocottus sp. 1. Richardson and Washington (1980) described larvae of this species as Cottidae Type 3. They illustrated spec- imens 6.3, 7.5, and 12.5 mm long. IDENTIFICATION.— Larvae were reared from eggs spawned from known adults. Field-collected larvae were identified through comparison with reared larvae. Identification of larvae and juve- niles was further confirmed by the following characters: pigmentation, body shape, an ad- vanced anus, and absence of a nasal cirrus. DISTINGUISHING FEATURES.— Preflexion and flexion larvae of C. globiceps may be distin- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 209 guished from other cottid larvae, except C. re- calvus and C. analis, by the presence of heavy pigmentation on the head, nape, and dorsolateral surface of the gut. Larval C. globiceps are distin- guished from C. recalvus and C. analis by the number (4-8) and spacing of ventral midline me- lanophores. Late flexion and postflexion larvae of C. globiceps differ from all other Clinocottus larvae in preopercular and parietal spination. Transforming and juvenile C. globiceps are distinguished from other cottid larvae by the combination of a blunt, rounded snout and head, heavy pigmentation over the anterior third of the body, and two or three inconspicuous ventral midline melanophores which persist on the cau- dal peduncle. PIGMENTATION.— Newly hatched and preflex- ion larvae of C. globiceps have intense melanistic pigmentation on the head and nape. Eight to 1 1 large stellate melanophores are present over the midbrain and 21-30 melanophores are concen- trated in the nape region. These nape melano- phores are arranged in a distinctive pattern in which 7-10 melanophores are embedded along the dorsal midline of the nape and are surround- ed anteriorly and laterally by 14-23 dark mela- nophores lying on the external surface of the nape. Eight to 10 dendritic melanophores occur on both the anterior and posterior walls of the otic cap- sules. The dorsolateral surface of the gut is heavi- ly pigmented with 100-150 large, round mela- nophores. The only pigmentation occurring posterior to the anus is a series of 4-8 discrete, ventral midline melanophores. These are situ- ated under the 10 posteriormost myomeres near the tail tip. Frequently, 2-5 additional small me- lanophores extend beyond the tail tip onto the caudal finfold. Pigmentation changes little during larval de- velopment. The midbrain melanophores in- crease in number ranging from 1 2 to 1 6 in larvae >6 mm. By about 8 mm, melanophores are densely concentrated over the nape and extend anteriorly onto the head. Melanophores are added in the midbrain region and several melanophores extend anteriorly over the forebrain onto the snout. As head musculature develops, melano- phores in the otic region become obscured so that only 5 or 6 melanophores are visible on the posterior wall of the otic capsule. During transformation, in planktonic larvae 12-14 mm long, head pigmentation increases markedly. Several melanophores are added on the upper lip and beneath the orbit. Melano- phores are also added in a row along the pre- opercle and on the dorsal portion of the oper- culum. Pigmentation over the brain intensifies and expands posteriorly, merging with the nape pigmentation. Concurrently, nape melanophores extend ventrally from the nape forming a con- tinuous band of pigment between the nape and gut. Pigmentation also increases over the body cavity as melanophores extend ventrally over the lateral surfaces of the gut. Several melanophores also are added on the pectoral fin base. Ventral midline melanophores decrease in size and num- ber, until only two to four inconspicuous mela- nophores persist beneath the caudal peduncle. Newly settled benthic juveniles of C. globiceps are distinctively pigmented with the anterior third of the body covered with dark melanophores ex- tending posteriorly to about a vertical through the seventh dorsal spine. Only the posterior two- thirds of the pelvic fin rays remain unpigmented. Posterior to the intense head pigment, the two to four small, ventral midline melanophores con- stitute the only pigment. Between 1 4 and 1 6 mm SL, juvenile pigmentation is added posteriorly along the dorsum. By about 14 mm, a dark ver- tical bar of melanophores forms under the sec- ond to fourth dorsal fin rays and extends ven- trally two-thirds of the way below the lateral midline. Between 15 and 16 mm, three addi- tional saddles of melanophores are added pos- teriorly along the dorsum. The first saddle forms under the 8th- 10th dorsal fin rays, the second forms under the 14th-l 5th fin rays, and the third saddle forms on the dorsal surface of the caudal peduncle. Concurrently, melanophores are added posteriorly along the lateral midline forming a dark band of pigment at the base of the caudal fin. Several melanophores appear on the pector- al, dorsal, and caudal fin rays. MORPHOLOGY.— Larval C. globiceps hatch at a relatively large size, 5.1-5.4 mm NL. Flexion of the notochord occurs between 6.2 and 8. 1 mm NL. The largest planktonic larva taken in field collection is 1 2.9 mm and is beginning to undergo transformation. The smallest benthic juvenile is 13.5 mm long. Thirty-eight specimens (5.1-14.6 mm) were examined for developmental mor- phometrics. Because only 10 larvae were avail- able from field collections, this morphometric series includes 25 laboratory-reared larvae. 210 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B '\ FIGURE 26. Larvae of Clinocottus globiceps: A) 5.0 mm NL, B) 6.3 mm SL, Q 7.5 mm NL (B and C from Richardson and Washington 1980). Larval C. globiceps are relatively deep-bodied, and the posterior portion of the gut trails below the rest of the body. When viewed ventrally, the hindgut bulges slightly on either side of the anus similar to, but less pronounced than, the bulges in C. embryum. Relative body depth at the pectoral fin base increases during larval development from 20.7% in preflexion larvae to 28.5% SL in transforming larvae and juveniles. Larval C. globiceps have a notably blunt, rounded head and snout with relative head length increasing from 17.0% in preflexion larvae to about 31% SL in transforming juveniles. FIN DEVELOPMENT.— The smallest larva begin- ning to undergo flexion of the notochord is 6.2 mm long. Notochord flexion is complete in lar- vae between 7.5 and 8.0 mm long. Although cau- dal rays are present in late-flexion stage larvae (7.0-7.5 mm NL), the adult complement of 6 + 6 principal caudal rays is not countable until the completion of flexion at 7.4 mm. WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 211 FIGURE 27. Larvae of Clinocottus globiceps: A) 8.5 mm SL, B) 12.5 mm SL, C) 12.9 mm SL (B from Richardson and Washington 1980). Dorsal and anal fin bases are just beginning to form at completion of notochord flexion. The full complement of dorsal (13-17) and anal (11- 12) fin rays is complete at ~9.5 mm. The dorsal spines (VIII-X) are completely formed at ~10 mm. Development of the pectoral fin corre- sponds to that of the dorsal and anal fins. Pectoral fin rays are visible on a 7.5 mm larva. The adult complement of rays (13-15) is fully formed by ~9-9.5 mm. Pelvic fin buds are first visible in larvae between 6.5 and 7 mm long. The adult complement (1,3) is present between 9.5 and 10 mm. SPIN AXIOM.— Preopercular spines first appear as seven to nine small bumps along the posterior margin of the preopercle in larvae 5.5-6 mm. Larvae undergoing notochord flexion have 9-14 small, evenly spaced spines along the preoper- cular margin. During postflexion, spines increase in number from 1 6 to 22, and the dorsalmost spine becomes separated from the rest of the preoper- cular spines by a short gap. Simultaneously, this dorsalmost spine becomes longer and stouter than the other preopercular spines. In the largest planktonic larvae (12.5 mm SL) this dorsalmost spine is about 2.5 times as long as the other spines. The lower preopercular spines decrease in size and number during transformation. The uppermost spine continues to become longer and stouter in benthic juveniles and is over four times 212 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B FIGURE 28. Juveniles of Clinocottus globiceps: A) 13.7 mm SL, B) 14.6 mm SL. as long as the lower spines in a 14.5 mm speci- men. The other preopercular spines are reduced to small bumps or serrations along the lower preopercular margin. In a 17-mm juvenile all remnants of the lower spines have disappeared and only the large dorsal spine persists. Clusters of spines develop on the head in the parietal region of C. globiceps larvae. One tiny spine is visible on each side of the head in larvae 6-7 mm and two spines are present on each side of the head in larvae 7-8 mm. By ~9-10 mm, five to six spines are present on each side of the head, arranged in a parallel pair of rows with two to three spines in the anterior row and three spines in the posterior row. These spines persist in the largest planktonic specimens examined (12.9 mm). In newly settled benthic juveniles, how- ever, these spines appear reduced and are present only as bony protuberances situated at the pos- terior margin of the parietals. Each protuberance has a hollow canal running through it which eventually forms the incipient cranial lateral line system in the parietal region of the head. Similar spine clusters also form in the supra- cleithral-posttemporal region. One or two small spines are first visible in larvae ~9 mm long. Five or six spines, arranged in two rows of three spines each, are present in both of the 1 2-mm specimens. These spines eventually become as- sociated with the lateral line system in benthic juveniles. Clinocottus analis (Figure 29; Table 6) LITERATURE.— Eigenmann (1892) and Budd (1940) briefly described and illustrated 4-mm specimens of C. analis. IDENTIFICATION.— Juveniles and adults were identified by the following combination of char- acters: an advanced anus, cirri, head shape, and pigmentation. Pigmentation, preopercular spi- nation, and body shape linked postflexion larvae of C. analis with juveniles and adults. DISTINGUISHING FEATURES.— Late postflexion and transforming specimens of C. analis were identified in collections from southern Califor- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 213 FIGURE 29. Young of Clinocottus analis: 10.5 mm SL. nia. Apart from the two descriptions of newly hatched larvae, intermediate larval stages of C. analis are unknown. A brief diagnosis of post- flexion larval C. analis is presented in the hope that this information may facilitate the identi- fication of a complete developmental series of C. analis. Eleven postflexion larval C. analis (9.9-11.4 mm SL) were examined for developmental mor- phology, pigmentation, and spination. Clinocot- tus analis postflexion larvae may be distin- guished from all other larvae belonging to the Artedius, Clinocottus, Oligocottus groups by the intense band of melanistic pigmentation on the lateral body surface between the bases of the sec- ond dorsal and anal fins. Intense melanophores are also present on the dorsolateral surface of the head, the snout, the tips of the lips, and on the operculum. A patch of melanophores is present on the pectoral fin base and in a band on the dorsum beneath the spinous dorsal fin. Sixteen to 22 small, round melanophores are situated on the ventral midline posterior to the anus. Postflexion larval C. analis have blunt, round- ed snouts and relatively large heads. Snout length and head length are 28% HL and 30% SL, re- spectively, longer than in other Clinocottus lar- vae. In addition, C. analis larvae have moder- ately long, bulging guts. Snout to anus length averages 49% SL in postflexion larvae. Body depth at the pectoral fin base is 28% SL, while body depth at the anus is 26% SL. Six to 1 1 spines are present on the posterior margin of the preopercle. The dorsalmost spine is longer and stouter than the other spines. In the smallest specimens (9.9-1 1 .0 mm) the spines are situated in two groups of three to five spines. The ventralmost spines begin to regress in larvae > 1 1 mm long; these spines decrease in size and num- ber and gradually become covered by skin. Two small spines are also present in the parietal region of the head in the 9.9 mm specimen. These spines decrease in size and remain only as bony bumps by ~ 1 1 mm. Artedius creaseri (Figures 30, 31; Table 7) LITERATURE.— Larval Artedius creaseri have not been previously described. IDENTIFICATION.— Juveniles and adults were identified by the following combination of char- acters: low dorsal fin ray (12-14) and anal fin ray (9-10) counts, low vertebral counts (30-3 1), scales extending onto head under the orbit and on the snout, and the presence of a preorbital cirrus. The developmental series was linked together primarily by preopercular and parietal spination, pigmentation, body shape, and meristics. Post- flexion and transforming larvae were linked to juveniles by the cirri pattern, pigmentation, body shape, and meristics. DISTINGUISHING FEATURES.— Preflexion lar- vae of A. creaseri are characterized by a pointed snout, large head, and relatively deep body. Dis- tinguishing pigmentation includes intense, large, round melanophores covering the dorsolateral surface of the gut, 1-3 large melanophores at the anteroventral margin of the gut, and a series of 7-1 1 large, evenly spaced melanophores along the ventral midline posterior to the anus. A large, distinctive, blotch-like melanophore is located on the ventral finfold near the tail tip, and another smaller melanophore occurs just beneath the tail tip. Larvae of A. creaseri >7 mm are further dis- tinguished by the presence of four large, evenly 214 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B FIGURE 30. Larvae of Artedius creaseri: A) 5.0 mm NL, B) 6.6 mm NL, C) 7.9 mm SL. spaced preopercular spines and a prominent pa- rietal and nuchal spine. Late postflexion larvae may be recognized by their pointed snout and long jaw, large head, and the characteristic ven- tral midline pigmentation. In addition, meristics, especially the low dorsal fin, anal fin, and ver- tebral counts are diagnostic of this species. Small juveniles possess a long, slender nasal cirrus, a broad, ribbonlike postorbital cirrus with a fringed tip, and two pairs of frontoparietal cirri. PIGMENTATION.— Small preflexion larvae of A creaseri are relatively lightly pigmented. They possess no melanistic pigmentation on either the head or the nape. Pigmentation over the dor- solateral surface of the gut is heavy and intense. Melanophores are large, round, and closely packed together. One or 2 melanophores are present on the ventral surface of the gut lying just posterior to the cleithrum. Posterior to the anus, the sole pigmentation consists of a series of 7-11 large, rounded melanophores evenly spaced along the ventral midline, positioned ap- proximately 1 to every three myomeres. This series originates under the third or fourth post- anal myomere and extends posteriorly toward the tail tip. The posteriormost 1 or 2 melano- phores in this series lie on the ventral finfold and are notable: large and blotchlike. Pigmentation increases markedly during larval development. Two melanophores form over the WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 215 B FIGURE 31. Larvae ofArtedius creaseri: A) 9.1 mm SL, B) 13.0 mm SL. midbrain in larvae >5.7 mm. During flexion, melanophores increase in number and extend anteriorly onto the forebrain. By ~8.0 mm, the dorsal surface of the head is entirely pigmented. Pigmentation extends dorsally along the anterior wall of the gut so that 3 or 4 large melanophores lie just posterior to the cleithrum. Several me- lanophores form at the posterior margin of the gut in larvae >6 mm, frequently forming a ring around the anus. The number of postanal ventral midline melanophores ranges from 6 to 12, and the 4th or 5th melanophore in the series increases markedly in size and extends below the body wall onto the ventral finfold. The posteriormost 2 me- lanophores in the series move up onto the base of the caudal finfold. One is positioned just pos- terior to the lower hypural plate and the other lies just below the tip of the notochord at the dorsal base of the upper hypural plate. Transforming larvae (>10 mm) have mela- nophores extending ventrally along the preoper- cle and opercle and two to four melanophores on the lower jaw. Pigment is also added on the pectoral fin. MORPHOLOGY.— The smallest larval A. crea- seri examined is 3.5 mm NL and recently hatched. Flexion of the notochord occurs between 5.7 and 7.9 mm NL. The largest planktonic specimen is 1 3 mm and beginning to develop juvenile pig- mentation. The smallest benthic juvenile is 13.5 mm and still undergoing transformation. Thirty- two specimens ranging from 3.5 to 13.6 mm were measured for developmental morphology. Artedius creaseri larvae are relatively deep- bodied with distinctively large heads and pointed snouts. Body depth at the pectoral fin base av- erages 26% in preflexion larvae and increases to 29% SL in postflexion larvae. Relative head length averages about 25% in preflexion and flexion lar- vae and increases to 33% SL in postflexion lar- vae. Snout length remains 30% HL during larval development. FIN DEVELOPMENT.— Initiation of a thickening in the hypural region of the developing caudal 216 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 TABLE 7. BODY PROPORTIONS OF LARVAE AND JUVENILES OF ARTEDIUS CREASERI AND A. MEANYI. Values are percent standard length (SL) or head length (HL) including mean, standard deviation, and range in parentheses. Item Artedius creaseri Artedius meanyi Head length/SL: Preflexion 24.4 ± 1.25(22.9-26.5) 18.6 ± 1.77(17.9-21.3) Flexion 26.0 ± 2.46 (22.8-28.8) 22.7 ± 3.53(19.4-38.0) Postflexion 32.6 ± 3.80 (26.5-38.5) 32.1 ± 4.68(25.4-38.0) Juvenile 40.3 ± 2.97 (38.2-42.4) - Snout length/HL: Preflexion 29.7 ± 2.74 (25.0-33.0) 27.5 ± 7.50(15.1-31.0) Flexion 31.9 ± 3.58(26.7-36.9) 29.5 ± 5.01 (22.3-35.4) Postflexion 30.4 ± 2.95(25.1-35.5) 29.4 ± 4.34(23.1-37.3) Juvenile 23.2 ± 1.20(22.4-24.1) - Eye diameter/HL: Preflexion 42.0 ± 4.42 (34.7-47.3) 44.0 ± 3.39 (35.0-47.7) Flexion 37.4 ± 3.11(32.5-43.1) 36.6 ± 4.93(30.4-43.1) Postflexion 33.2 ± 4.01 (27.5-38.6) 29.0 ± 3.31 (23.2-35.3) Juvenile 30.4 ± 1.20(29.6-31.3) - Snout to anus length/SL: Preflexion 44.6 ± 3.01 (41.5-50.0) 33.0 ± 2.86 (27.2-35.4) Flexion 43.5 ± 3.83 (39.7-50.0) 39.6 ± 2.97(36.1-44.3) Postflexion 51.5 ± 3.66(44.6-55.8) 48.3 ± 4.43 (38.0-55.2) Juvenile 54.4 ± 1.77(53.2-55.7) - Snout to pelvic fin origin/SL: Preflexion _ _ Flexion _ 23.0* Postflexion 29.5 ± 2.17(27.0-33.1) 29.1 ± 3.16(21.9-33.6) Juvenile 32.0 ± 2.12(30.5-33.5) - Pelvic fin origin to anus/SL: Preflexion _ - Flexion - 20.1* Postflexion 24.6 ± 4.96(17.5-33.9) 19.4 ± 1.60(15.7-22.1) Juvenile 22.4 ± 0.35 (22.2-22.7) - Body depth at pectoral fin base/SL: Preflexion 26.0 ± 2.54 (23.4-29.7) 18.0 ± 1.64(16.2-20.0) Flexion 26.0 ± 1.78(22.7-28.1) 19.9 ± 1.25(17.7-22.2) Postflexion 28.8 ± 2.67 (25.3-33.8) 24.6 ± 3.65(17.2-30.1) Juvenile 24.8 ± 1.98(23.4-26.2) - Body depth at anus/SL: Preflexion 21.8 ± 2.35(19.4-26.0) 15.2 ± 3.11(11.9-19.1) Flexion 24.0 ± 3.08(19.3-28.1) 18.9 ± 1.55(15.8-21.3) Postflexion 28.2 ± 3.07 (25.4-36.2) 24.6 ± 2.73(19.3-28.4) Juvenile 21.2 ± 2.19(19.6-22.7) - Pectoral fin length/SL: Preflexion - 8.4 ± 1.35(6.4-9.1) Flexion 12.2 ± 1.97(9.5-15.1) 9.2 ± 1.77(6.5-12.1) Postflexion 23.7 ± 6.11(14.5-35.2) 19.0 ± 6.21 (10.1-28.3) Juvenile 30.1 ± 1.34(29.1-31.0) - - = Not present at this stage. * = Only one specimen available at this stage. fin is first evident at 5.7 mm, coincident with the onset of notochord flexion. Caudal rays are pres- ent at 6.4 mm, but the adult complement of 6 + 6 principal caudal rays is not present until larvae reach ~8.0 mm. Bases of the second dorsal (12-14) and anal WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 217 (9-11) fins are countable in larvae ~7-8 mm long, and fin rays are formed between 9 and 10 mm. The adult complement of dorsal fin spines (IX-X) is first countable at 9.7 mm. Pectoral fin rays begin forming at ~7-8 mm, and the adult complement is present at 8.6 mm. Pelvic buds begin to form at >8 mm; however, the adult complement of rays is not present until larvae are > 1 1 mm. SPINATION.— Artedius creaseri larvae develop prominent head spines. In contrast to Artedius, Clinocottus, and Oligocottus larvae which have multiple preopercular spines, A. creaseri larvae develop four equal-sized preopercular spines. Two spines develop first on the posterior margin of the preopercle at >5.7 mm. At a state of growth between ~6.4 and 7 mm, two additional spines develop, one dorsal and one ventral to the original two spines. The middle two spines re- main slightly longer than the outer two through- out larval development. These spines persist through transformation and are present in ju- veniles. In larvae >10 mm, small basal spines or projections form on the base of each of the four main preopercular spines. With develop- ment, four bony ridges form on the inner shelf of the preopercle parallel to each basal spine. These ridges grow toward the basal spines and gradually fuse with them forming bony arches over the incipient lateral line canal of the pre- opercle. Prominent spines also form in the pa- rietal region of the head. A single parietal spine is first visible at 5.7 mm. By >9 mm, a second smaller parietal spine forms just posterior to the first. These spines are quite large and distinctive, and they are present in the largest planktonic larvae (> 13 mm long). When larvae reach ~8 mm, a spine forms in the supracleithral-posttemporal region. The su- pracleithral spine points dorsolaterally. A second supracleithral spine forms just ventral to the first and points dorsally. Larvae >9.5 mm form one posttemporal spine. These spines persist in the largest planktonic larvae (13.6 mm) but regress in young juveniles becoming incorporated into the developing lateral line canal system. Artedius meanyi (Figures 32, 33; Table 7) LITERATURE.— Blackburn (1973) described a 4.3-mm larva resembling A. meanyi, which he called Cottid 3. Richardson (1977) and Richard- son and Pearcy (1977) listed three larvae as Ice- linus sp. 1. Richardson and Washington (1980) illustrated and described specimens 3.3, 8.6, 10.9, 12.5, 13.5, 15.2, 16.5, and 16.6 mm long as Ice- linus spp. IDENTIFICATION.— Larval A. meanyi were mis- identified as Icelinus spp. by Richardson and Washington (1980) on the basis of meristics and the pelvic fin ray count of 1,2, which is charac- teristic of Icelinus. Meristics also match those of A. meanyi, which possess 1,3 (rarely 1,2) pelvic fin rays (Rosenblatt and Wilkie 1963; Lea 1974). Recently, Howe and Richardson (1978) reex- amined Lea's specimens of A. meanyi and re- ported that "... only one small specimen ap- peared to have two rays— all others had three rays." Lea's specimens were reexamined in this study. Cleared and stained specimens clearly have 1,2 pelvic fin rays. The outermost ray is greatly thickened and branched at the tip in all speci- mens examined. All of the misidentified "Iceli- nus" larvae possess this distinctive, thickened outer ray. In addition, during the present study, large transforming specimens of A. meanyi were ob- tained that possess scales on the dorsal surface of the head, the opercle, and in four or five rows on either side of the dorsal fins. Specimens also possessed preorbital cirri and distinctive post- ocular cirri with three tentacles arising from a single base. The combination of these morpho- logical and meristic characters conclusively iden- tifies these transforming larvae and juveniles as A. meanyi. The developmental series was linked together primarily on the basis of pigmentation and body shape. DISTINGUISHING FEATURES.— Small preflexion larval A. meanyi are distinguished by their short, compact guts (snout to anus length averages 33% SL) and pointed snouts. Characteristic pigmen- tation includes a low number of ventral midline melanophores posterior to the anus (< 13), sev- eral large melanophores situated anteriorly on the visceral mass at the base of the cleithrum, and two distinctive blotches of pigment on both the dorsal and anal finfolds. Notochord flexion begins at a relatively large size, ~6.2 mm in A. meanyi larvae, and is com- plete by ~9.4 mm. Four large, evenly spaced spines form along the margin of the preopercle in postflexion larvae >9 mm. Two parietal spines develop at the posterior margin of each parietal in larvae > 1 1 mm. Postflexion and juvenile A meanyi (13-18 mm 218 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 B FIGURE 32. Larvae ofArtedius meanyi: A) 3.3 mm NL, B) 8.6 mm NL, C) 10.9 mm NL (from Richardson and Washington 1980). long) are distinguished by a low number of blotchy ventral midline melanophores posterior to the anus, a relatively pointed snout and large head (33% SL), a pelvic fin ray count of 1,2 with the outermost ray thickened as if two rays are fused together, and other meristics. In addition, ju- venile A. meanyi possess a single, slender preor- bital cirrus, an eyeball cirrus, and a distinctive postorbital cirrus having three tentacles that arise from a single base. The largest specimens (16- 1 8 mm) possess rows of prickle-like scales on the parietal, cheek, and opercular regions of the head and on the dorsal surface of the body and caudal peduncle. PIGMENTATION.— Small A. meanyi larvae are relatively lightly pigmented. Melanistic pigmen- tation is absent on the head of preflexion larvae. Two to 5 round, external melanophores are clus- tered on the nape. The dorsolateral surface of the gut is lightly pigmented. Two or 3 large dendritic melanophores are embedded in the anterior musculature of the body cavity just posterior to the cleithrum. Posterior to the anus, a series of 7-13 large, blotch-like melanophores is posi- tioned along the ventral midline originating un- der the second to fourth postanal myomere and extending toward the tail tip. These melano- phores vary in size with the 3rd or 4th and the WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 219 B FIGURE 33. Young ofArtedius meanyi: A) 13.8 mm SL, B) 15.2 mm SL, Q 16.5 mm SL (from Richardson and Washington 1980). posteriormost melanophores of the series being markedly larger and frequently extending onto the ventral finfold. Two large distinct pigment blotches are present on both the dorsal and ven- tral finfolds in small larvae. One specimen out of 45 examined possessed three pigment spots on both the dorsal and anal finfolds. Melanistic pigmentation increases during lar- val development. Several melanophores are added over the brain in larvae between 7.4 and 8 mm. Melanophores in the nape region become embedded in larvae > 6 mm as body musculature develops. Pigmentation increases slightly over the lateral surfaces of the gut. With the onset of notochord flexion and development of the caudal fin, the posteriormost melanophore of the ven- tral midline series is characteristically positioned at the ventral margin of the forming caudal fin. A second, large melanophore is frequently added at the dorsal margin of the caudal fin base dorsal to the notochord tip. The blotches of pigment on the dorsal and ventral finfolds disappear in lar- vae >9 mm as fin rays begin to form. During transformation, between 13 and 19 mm, head pigmentation increases markedly. Me- lanophores extend anteriorly over the interor- 220 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 bital region and onto the snout. Several mela- nophores are added just ventral to the orbit, between the eye and preopercle, and along the dorsal margin to the opercle. With development, melanophores are also added along the pectoral fin base, in a band across the dorsum, and on the dorsal fin around the first four dorsal spines. Melanophores are also added to the upper and lower lips, across the cheek, along the ventral margin of the opercle, and on the dorsal surface of the head. MORPHOLOGY.— The smallest larval A. mean- yi collected from plankton samples are ~ 3 mm NL long and appear recently hatched. Larvae undergo notochord flexion between 6.2 and 9.4 mm NL. Specimens as large as 18-19 mm were collected in plankton and neuston tows. Plank- tonic larvae > 1 5 mm are beginning to undergo transformation indicated by the development of juvenile pigmentation and the formation of scales on the head and dorsum. The smallest benthic juveniles examined were 15-16 mm long and were fully transformed. Thirty specimens, rang- ing in size from 3.3 to 17.9 mm, were examined for morphometrics. Small larval A. meanyi are relatively slender with a characteristic body shape. Body depth is constricted just posterior to the anus; the body bulges slightly in the midtail region and narrows again near the tail tip or caudal peduncle. This distinctive body shape remains apparent throughout larval development. The gut of A. meanyi is short and tightly coiled. Snout to anus length averages 33% SL in preflexion larvae in- creasing markedly to 48% in postflexion larvae. Prior to flexion of the notochord, body depth averages 18% at the pectoral fin base and 15% SL at the anus and increases to 25% SL at both the pectoral fin base and the anus in postflexion larvae and juveniles. Artedius meanyi larvae have small heads with a distinctively pointed snout. Head length av- erages 19% in preflexion larvae, then increases dramatically to 33% SL in late postflexion larvae and juveniles. Snout length remains relatively constant throughout larval development, ranging from 28 to 30% HL. FIN DEVELOPMENT.— The fins develop rela- tively late in A. meanyi. Initiation of notochord flexion begins at ~6.2 mm NL. Although caudal rays are first visible in larvae > 7 mm long, prin- cipal caudal ray number (6 + 6) is not complete until after notochord flexion at ~ 1 1 mm. Dorsal and anal soft rays begin to form in larvae 9.5- 10 mm long. The full complement of fin rays is visible in larvae ~ 1 2 mm. Dorsal spines (IX-X) begin to form at ~ 1 1 mm and are all present by 12-13 mm. Pelvic fin buds form in larvae >9.5 mm; however, the adult complement of pelvic fin rays (1,2) is not complete until larvae reach ~ 12-1 3 mm. SPINATION.— Preopercular spines form rela- tively late in the development of A. meanyi lar- vae. Two tiny spines are first visible along the central portion of the preopercle in larvae >6.2 mm with a third spine forming dorsal to these spines between 8 and 8.5 mm. By 9.4 mm, a fourth spine is added at the ventral margin of the preopercle. These four spines remain prominent and approximately equal-sized throughout larval development. In larvae >13 mm, small basal spines or projections form on the base of each of the four main preopercular spines. With de- velopment, four bony ridges form on the inner preopercular shelf parallel to each basal spine. These ridges grow toward the basal spines and gradually fuse with them forming bony arches over the incipient lateral line canal of the pre- opercle. During transformation, between 1 5 and 17 mm, the dorsalmost preopercular spine be- comes longer and stouter than the other spines; however, all four preopercular spines remain clearly visible on the largest pelagic juveniles ex- amined (~ 19 mm long). Spines also develop in the parietal and supra- cleithral-posttemporal regions of the head. A sin- gle tiny spine first forms at the posterior margin of the parietal in larvae >7 mm long. This spine gradually becomes longer, and in larvae between 12 and 13 mm a second, smaller nuchal spine forms immediately posterior to it. A small spine forms on the dorsal margin of the posttemporal bone between 9 and 10 mm. A second, similarly sized spine is added ventrally on the posttemporal in larvae ~ 1 1 mm. At about the same time, a third spine forms posteroven- trally to the two posttemporal spines on the dor- sal portion of the supracleithrum. These three spines increase in size during transformation and eventually become associated with the junction of the cephalic and lateral line systems. ADDENDUM Since this study was accepted for publication, a subsequent review of cottid relationships has been published (Washington et al. 1984). Wash- WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS 221 ington et al. (1984) presented a hypothesis of phylogenetic relationships of cottoids and briefly summarized characters that supported eight pro- posed monophyletic groups of cottid genera. Much of this work was based on an unpublished manuscript (Washington and Richardson n.d.) in which a hypothesis of cottid relationships based on osteological characters of early life history stages was presented. Results of Washington et al. (1984) support the proposed monophyly of Artedius Group A, Clinocottus, and Oligocottus and the exclusion of Artedius creaseri and A. meanyi from this group. They listed five syn- apomorphic characters in addition to the multi- ple preopercular spines which support the mono- phyly of Artedius Group A, Clinocottus, and Oligocottus. These characters include anterior neural arches enlarged and elevated, arms of the anterior neural arches open in a broad u-shape un- til late in the juvenile period, a greatly expanded cleithrum base, posterior extensions or bony plates at the cleithrum base which enclose the pelvic bones, and loss of ventral postcleithrum and reduction or loss of dorsal postcleithrum. Washington et al. (1984) also placed Artedius creaseri and A. meanyi in a proposed monophy- letic group of genera including Icelinus and Myoxocephalus as well as 10 other cottid genera. This placement gives additional evidence for the exclusion of Artedius creaseri and A. meanyi from the genus Artedius (sensu Bolin 1944) and sup- ports their close relationship to Icelinus and Myoxocephalus. Further results of Washington et al. (1984) provide additional characters that strengthen the hypothesis of monophyly of Ar- tedius Group A, Clinocottus, and Oligocottus. ACKNOWLEDGMENTS I am indebted to many people for the loan of materials. Dr. Geoffrey Moser, National Marine Fisheries Service, La Jolla; Dr. Richard Rosen- blatt, Scripps Institution of Oceanography, La Jolla; Dr. Robert Lavenberg, Los Angeles Coun- ty Museum, Los Angeles; H. J. Walker, Ecolog- ical Marine Consultants, Solano Beach; Dr. Wil- liam Eschmeyer, California Academy of Sciences, San Francisco; David Rice, Lawrence Livermore Laboratory, Livermore; Dr. Arthur Kendall, Na- tional Marine Fisheries Service, Seattle; Dr. Theodore Pietsch, University of Washington, Se- attle; Dr. Norman Wilimovsky, University of British Columbia, Vancouver; and Dr. Jeffrey Marliave, Vancouver Public Aquarium, Van- couver; all were helpful in making materials in their collections available to me. Special thanks are due Dr. Robert Morris for generously allowing me access to an unpublished manuscript on laboratory-reared Oligocottus snyderi larvae. I would also like to thank Kevin Howe, Joanne Laroche, Wayne Laroche, James Long, Bruce Mundy, Dorinda Ostermann, and Waldo Wakefield for volunteering their help in tidepool sampling. I gratefully acknowledge Dr. William Pearcy for his encouragement and for providing laboratory space. Special thanks are due to Kevin Howe for stimulating conversa- tions about cottid relationships, for reviewing parts of this manuscript, and for first teaching me about sculpins. I wish to thank Dr. Carl Bond, Wayne Laroche, and Dr. Sally Richardson for helpful discussions about cottids and for review- ing initial drafts of this manuscript. I gratefully acknowledge Margaret Snider and Lucia O'Toole for carefully typing drafts of this paper. This study was funded in part by Oregon State University Sea Grant College Program support- ed by NOAA Office of Sea Grant, Department of Commerce and by a Grant-in-Aid of Research from Sigma Xi, The Scientific Research Society of North America. LITERATURE CITED AHLSTROM, E. H., J. L. BUTLER, AND B. Y. SUMIDA. 1976. Pelagic stromateoid fishes (Pisces, Perciformes) of the east- ern Pacific. Kinds, distributions, and early life histories and observations on five of these from the northwest Atlantic. Bull. Mar. Sci. 26:285-402. ASHLOCK, P. H. 1974. The uses of cladistics. Ann. Rev. Ecol. Syst. 5:81-89. BERG, L. S. 1940. Classifications of fishes, both recent and fossil. Travaux de L'Institut Zoologique de 1' Academic des Sciences de 1'URSS 5(2):87-517. BERTELSEN, E. 1951. The ceratioid fishes. Ontogeny, taxon- omy, distribution, and biology. Dana Rep. 39. 276 pp. BLACKBURN, J. E. 1973. A survey of the abundance, distri- bution, and factors affecting distribution of ichthyoplankton in Skagit Bay. M.S. Thesis, University of Washington, Se- attle. 1 36 pp. BOLIN, R. L. 1934. Studies on California Cottidae: an analysis of the principles of systematic ichthyology. Ph.D. Disser- tation, Stanford University, Stanford, California. 337 pp. . 1944. A review of the marine cottid fishes of Cali- fornia. Stanford Ichthyol. Bull. 3(1): 1-1 35. . 1 947. Evolution of the marine Cottidae of California with a discussion of the genus as a systematic category. Stan- ford Ichthyol. Bull. 3:153-168. BRUNDIN, L. 1966. Transantarctic relationships and their significance, as evidenced by chironomid midges with a monograph of the subfamilies Podonominae and Aphrote- 222 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 9 niinae and the austral Heptagyiae. K. Svenska Vetenskadad. Handl. (Ser. 4) ll(l):l-472. . 1968. Application of phylogenetic principles in sys- tematics and evolutionary theory. Pp. 473-495 in Current problems in lower vertebrate phylogeny, T. Orvig, ed. Nobel Symposium 4. Almquist and Wiksell, Stockholm. BUDD, P. 1940. Development of the eggs and early larvae of six California fishes. Calif. Fish Game Bull. 56. 53 pp. CRACRAFT, J. 1974. Phylogenetic models and classification. Syst.Zool. 23:71-90. DINGERKUS, G. AND L. D. UHLER. 1977. Enzyme clearing of alcian blue stained whole small vertebrates for demonstra- tion of cartilage. Stain Tech. 52(4):229-232. EIGENMANN, C. H. 1892. The fishes of San Diego, California. Proc. U.S. Natl. Mus. 15:123-178. ELDRIDGE, M. B. 1970. Larval fish survey of Humboldt Bay. M.S. Thesis. Humboldt State College, Arcata, California. 52 pp. HENNIG, W. 1966. Phylogenetic systematics. University of Illinois Press, Inc. Urbana, Illinois. 263 pp. HOWE, K. M. AND S. L. RICHARDSON. 1978. Taxonomic re- view and meristic variation in marine sculpins (Osteichthys: Cottidae) of the northeast Pacific Ocean. Fin. Rep., NOAA- NMFS Contr. No. 03-78-M02-120. 142 pp. HUBBS, C. L. 1 926. A revision of the subfamily Oligocottinae. Occ. Pap. Mus. Zool. Univ. Michigan 171. 1 8 pp. JOHNSON, R. K. 1 974. A revision of the alepisauroid family Scopelarchidae (Pisces:Myctophiformes). Fieldiana (Zool.) 66:1-249. JORDAN, D. S. 1 923. A classification of fishes including fam- ilies and genera as far as known. Stanford Univ. Publ., Univ. Sen, Biol. Sci. 3:77-243. JORDAN, D. S. AND B. W. EVERMANN. 1898. The fishes of North and Middle America. Bull. U.S. Natl. Mus. 47(pt. 2): 1241-2183. KENDALL, A. W. 1979. Morphological comparisons of North American sea bass larvae (Pisces:Serranidae). U.S. Dep. Comm., NOAA Tech. Rep. NMFS Circ. 428. 50 pp. LEA, R. N. 1974. First record of Puget Sound sculpin, Ar- tedius meanyi, from California. J. Fish. Res. Board Can. 3 1 : 1242-1243. MARLIAVE, J. B. 1975. The behavioral transformation from the planktonic larval stage of some marine fishes reared in the laboratory. Ph.D. Dissertation, University of British Co- lumbia, Vancouver. 231 pp. MAYR, E. 1969. Principles of systematic zoology. McGraw- Hill, Inc. New York, New York. 428 pp. . 1974. Cladistic analysis of cladistic classification. Z. Zool. Syst. Evol.-forsch. 12:94-128. MILLER, D. J. AND R. N. LEA. 1972. Guide to the coastal marine fishes of California. Calif. Dep. Fish Game Fish Bull. 157. 235 pp. MORRIS, R. W. 1951. Early development of the cottid fish Clinocottus recalvus (Greeley). Calif. Fish Game Fish Bull. 37(3):28 1-300. . n.d. Unpubl. MS, University of Oregon, Eugene, Oregon. MOSER, H. G. AND E. H. AHLSTROM. 1970. Development of lanternfishes (family Myctophidae) in the California Cur- rent. Pt. 1. Species with narrow eyed larvae. Los Angeles Co. Mus. Nat. Hist. Sci. Bull. 7:1-45. . 1972. Development of the lanternfish, Scopelopsis multipunctatus Brauer 1906, with a discussion of its phy- logenetic position in the family Myctophidae and its role in a proposed mechanism for evolution of photophore patterns in lanternfish. Fish. Bull., U.S. 70(3):54 1-564. -. 1974. Role of larval stages in systematic investiga- tions of marine teleosts: the Myctophidae, a case study. Fish. Bull., U.S. 70(2):391-413. MOSER, H. G., W. J. RICHARDS, D. M. COHEN, M. P. FAHAY, A. W. KENDALL, JR., AND S. L. RICHARDSON. 1984. On- togeny and systematics of fishes. Am. Soc. Ichthyol. Herp. Spec. Publ. No. 1. Allen Press, Inc., Lawrence, Kansas. 760 pp. NELSON, J. S. 1976. Fishes of the world. Wiley-Interscience, New York, New York. 416 pp. OKJYAMA, M. 1 974. The larval taxonomy of the primitive myctophiform fishes. Pp. 609-62 1 in The early life history offish, J. H. S. Blaxter, ed. Springer- Verlag, New York. QUAST, J. C. 1965. Osteological characteristics and affinities of the hexagrammid fishes with a synopsis. Proc. Calif. Acad. Sci. 31(21):563-600. REGAN, C. T. 1913. Osteology and classification of the te- leostean fishes of the order Scleroparei. Ann. Mag. Nat. Hist. 8(11): 169-1 84. RICHARDSON, S. L. 1977. Larval fishes in ocean waters off Yaquina Bay, Oregon: abundance distribution and season- ality, January 1971-August 1972. OSU Sea Grant Publ. ORESU-T-77-003. 72 pp. . 1981. Current knowledge of northeast Pacific sculpin larvae (Family Cottidae) with notes on relationships within the family. Fish. Bull., U.S. 79:(1). RICHARDSON, S. L. AND W. A. LAROCHE. 1979. Development and occurrence of larvae and juveniles of the rockfishes Se- bastes crameri, Sebastes pinniger, and Sebastes helvoma- culatus (Family Scorpaenidae) off Oregon. Fish. Bull., U.S. 77:1-46. RICHARDSON, S. L. AND W. G. PEARCY. 1977. Coastal and oceanic fish larvae in an area of upwelling off Yaquina Bay, Oregon. Fish. Bull., U.S. 75(1): 125-145. RICHARDSON, S. L. AND B. B. WASHINGTON. 1980. Guide to identification of some sculpin (family Cottidae) larvae from marine and brackish waters off Oregon and adjacent areas. NOAA Tech. Rep. NMFS Circ. 430. 56 pp. ROSENBLATT, R. H. AND D. WILKIE. 1963. A redescription of the rare cottid fish, Artedius meanyi, new to the fauna of British Columbia. J. Fish. Res. Board Can. 20:1505-1511. SIMPSON, G. G. 1961. Principles of animal taxonomy. Co- lumbia Univ. Press, New York, New York. 247 pp. SNEATH, P. H. H. AND R. R. SOKAL. 1973. Numerical tax- onomy. W. H. Freeman and Company, San Francisco, Cal- ifornia. STEIN, R. 1972. Identification of some larval Pacific cottids. M.S. Thesis, Humboldt State College, Arcata, California. 4 1 PP- . 1973. Description of laboratory-reared larvae ofOli- gocottus maculosus Girard (Pisces:Cottidae) Copeia 1973: 373-377. TARANETS, S. Y. 1 94 1 . On the classification and origin of the family Cottidae. Issues. Akad. Nauk. SSR, Old. Biol. 1943(3): 427-447. (Trans, from Russian, Univ. British Columbia Mus. Contr. 5:1-28.) WASHINGTON, B. B. 1981. Identification and systematics of larvae of Artedius, Clinocottus, and Oligocottus (Scorpae- niformes:Cottidae). M.S. Thesis, Oregon State University, Corvallis, Oregon. 192 pp. WASHINGTON: RELATIONSHIPS AND ONTOGENY OF THE SCULPINS WASHINGTON, B. B., W. N. ESCHMEYER, AND K. M. HOWE. Note: Since this paper went to press, a phylogenetic study of 1984. Scorpaeniformes: relationships. Pp. 438-447, in On- the Cottoidea has been published that addresses relation- togeny and systematics of fishes, Moser et al., eds. Am. Soc. ships of 42 cottid genera including Artedius, Clinocottus, and Ichthyol. Herp. Spec. Publ. No. 1. Oligocottus (Yabe 1985). A discussion of Yabe's work will WASHINGTON, B. B. AND S. L. RICHARDSON n.d. Systematic be included in a forthcoming paper (Washington and Rich- relationships among cottid fishes based on osteological char- ardson n.d.). acters of early life history stages with notes on their allies. YABE, M. 1985. Comparative osteology and myology of Unpubl. MS. NMFS Systematics Lab., Natl. Mus. Nat. Hist., the super-family Cottoidea (Pisces: Scorpaeniformes), and Washington, D.C. its phylogenetic classification. Mem. Fac. Fish. Hokkaido WHITE, W. A. 1977. Taxonomic composition, abundance, Univ., 32(1):1-130. distribution and seasonally of fish eggs and larvae in New- port Bay, California. M.S. Thesis, California State Univer- sity, Fullerton, California. 107 pp. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 10, pp. 225-236, 5 figs., 2 tables. May 6, 1986 BILATERAL ASYMMETRY IN PHALLOSTETHID FISHES (ATHERINOMORPHA) WITH DESCRIPTION OF A NEW SPECIES FROM SARAWAK By Lynne R. Parent! California- Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: Phenacostethus trewavasae, the first phallostethine from Borneo, is described from clay- and gravel-bottom freshwater streams of the Baram River, Sarawak. One of the distinguishing characteristics of P. trewavasae is a minute eye-lens. The subfamily Phallostethinae comprises Phallostethus Regan, known only from a single collection of one species; P. dunckeri Regan, from the mouth of the Muar River, in Johore on the Malay Peninsula; and Phenacostethus Myers, known previously from two species, P. smith! Myers and P. posthon Roberts, from coastal peninsular Malaysia and Thailand. Male phallostethids are bilaterally asymmetric. The subcephalic copulatory organ, the priapium, is oriented so that the aproctal side of the body is either the left or right; hence, males are termed sinistral or dextral, respectively. Both Phallostethus dunckeri and Phenacostethus smith! have, in about equal numbers, males that are either sinistral or dextral. In P. posthon, all males are dextral, whereas, in P. trewavasae, all males are sinistral. One species of neostethine, Mirophallus bikolanus Herre, is known in which all males are dextral. Bilateral asymmetry is compared among phallostethids to assess better the nature of this phenomenon, and its importance in determining the homology of priapial structures. INTRODUCTION and Neostethinae. The closest living relative of the freshwater, brackish, and occasionally salt- Phallostethids are a group of some 20 known water phallostethids is hypothesized to be the species of Indo- Australian atherinomorph fishes monotypic western Pacific marine silverside (or distinguished from all other teleosts by the pres- hardyhead) Dentatherina Patten and Ivantsoff ence in males of the priapium, a complex, sub- (Parenti 1984). cephalic copulatory organ (Regan 1913, 1916). The subfamily Phallostethinae includes two Phallostethids have been divided into two groups genera: Phallostethus Regan with one species, P. (and classified traditionally in two families, Phal- dunckeri Regan, 1 9 1 3, and Phenacostethus Myers, lostethidae and Neostethidae, as in Roberts with three species, P. smithi Myers, 1928, P. pos- 197 la, b) on the basis of gross differences in thon Roberts, 197 la, and P. trewavasae, de- priapial morphology. Rosen and Parenti (1981) scribed herein. and Parenti (1984) treated the entire group as Phallostethids are found throughout coastal one family, the Phallostethidae sensu lato, and peninsular Malaysia, Thailand, Borneo, the Phil- that convention is followed here; the two groups ippines, and Java. Phallostethus and Phenaco- are referred to as the subfamilies Phallostethinae stethus were known previously only from penin- [2251 226 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 10 sular Malaysia and Thailand (see Roberts 1 97 1 a, fig. 1). The third species of Phenacostethus, P. trewavasae, was collected from the Baram River in Sarawak, Malaysian Borneo. Fowler (1937) described Phenacostethus thai from a series of nine specimens, four males and five females (Academy of Natural Sciences, Philadelphia, ANSP 51352-51360). Roberts (1971a) followed Herre (1942) in treating Phenacostethus thai as a synonym of Phenacostethus smithi, and I ten- tatively concur. All Phallostethinae and Neostethinae males are bilaterally asymmetric with regard to orientation of the major supporting bones and other struc- tures associated with the priapium. Most females are bilaterally symmetric; the anus is just ante- rior to the urogenital opening along the ventral midline, under the throat (see Regan 1913, fig. 2). In just one species of neostethine are bilat- erally asymmetric females found (see Parenti, in press). Bilateral asymmetry is well documented in fishes (see Hubbs and Hubbs 1945, and refer- ences therein, for a comprehensive review). Most phallostethid species have both sinistral and dex- tral males in more or less equal numbers. Roberts (19710) described Phenacostethus posthon, the first known species in which males are exclu- sively dextral. Phenacostethus trewavasae males are exclusively sinistral. Both species are known from relatively large numbers of specimens, such that unique or fixed asymmetry can only be in- terpreted as a natural phenomenon in some phal- lostethids. Phenacostethus trewavasae is also distin- guished by a minute eye-lens. The structure and function of the retina and, hence, the quality of vision, is unknown. MATERIALS AND METHODS The material on which the description of Phe- nacostethus trewavasae is based was made avail- able for study by Dr. E. J. Grossman, of the Royal Ontario Museum (ROM), where the holotype and majority of paratypes and additional specimens are deposited. Remaining paratypes have been deposited in the California Academy of Sciences (CAS), American Museum of Natural History (AMNH), British Museum (Natural History) (BMNH), and the United States National Mu- seum of Natural History (USNM), through the courtesy of Dr. Grossman. Included in the comparative material are spec- imens of Phallostethus dunckeri Regan from the single known collection by G. Duncker from the Muar River, in Johore on the Malay Peninsula (Duncker 1904). Regan's (1913) description was based on seven specimens from this collection. Both Roberts (19716) and Parenti (1984) be- lieved that the only known specimens were the BMNH syntypes. However, additional speci- mens from the single collection by Duncker have been discovered in the Zoologisches Museum, Hamburg (ZMH) and have been made available for study through the courtesy of Prof. H. Wil- kens. Some of the ZMH material was given lec- totype and paralectotype status erroneously by Ladiges et al. (1958), who did not refer to the BMNH syntypes. Osteological structures were examined in, and counts made on, material counterstained with alcian blue and alizarin red S following the pro- cedure of Dingerkus and Uhler (1977), or solely alizarin stained. See text and Table 2 for catalog numbers of phallostethid material examined. Al- cohol-preserved (USNM 230367, USNM 230181 ), solely alizarin-stained (USNM 23037 1 , USNM 230366), and counterstained prepara- tions (USNM 230374) of the western Pacific Dentatherina merceri were used for outgroup comparison. Additional comparative material was obtained on loan from the University of Michigan, Museum of Zoology (UMMZ) and the Museum of Comparative Zoology, Harvard Uni- versity (MCZ). A Zeiss SV8 stereomicroscope with drawing tube and photomicrography apparatus was used for dissection of specimens and recording of data. Phenacostethus trewavasae, new species (Figures 1-3, 4b) HOLOTYPE.— ROM 41826, a mature, sinistral male, 14.1 mm standard length, collected 3 August 1 98 1 , by Dwight Wat- son, from Malaysia: Sarawak (Fourth Division), Baram River, Sungei Kejin Tugang, tributary of Sungei Kejin, depth to 1 m, clay- and gravel-bottom stream (03041'30"N, 1 14°27'15"E). PARATYPES.— ROM 44289 (8 sinistral males, 11 females); ROM CS 812 (2 sinistral males, 1 female, all cleared and stained with alizarin red S); ROM 41827(1 adult female), taken with the holotype. ROM 41829 (1 sinistral male); ROM 41830 (7 sinistral males, 6 females); CAS 55454 (3 sinistral males, 2 females); BMNH 1984.7.12:1-5 (3 sinistral males, 2 females); AMNH 55570 (3 sinistral males, 2 females); USNM 267266 (2 sinistral males, 3 females), all collected 1 1 February 1980, by Dwight Watson, from Malaysia: Sarawak (Fourth Division), Baram River, Sun- gei Kejin, station at confluence of Kejin Tugang and Kejin PARENTI: PHALLOSTETHID FISHES 227 FIGURE 1. Phenacostethus trewavasae, 1 4.1 -mm holotype (ROM 41826). River, depth to 1 m, clay- and gravel-bottom stream, no vege- tation (03°4r30"N, 114°27'15"E). ADDITIONAL MATERIAL EXAMINED (no type status).— ROM 41828(11 juveniles), collected 27-30 July 1981 from Malaysia: Sarawak (Fourth Division), Baram River, Loagan Titad. ROM 44290 (9 sinistral males, 17 females, 3 juveniles or of undetermined gender, of which 2 sinistral males and 2 females have been counterstained with alcian blue and alizarin red S); ROM 44291 (24 sinistral males, 12 females, 14 juveniles or of undetermined gender) taken with the holotype. DIAGNOSIS.— Phenacostethus trewavasae is distinguished from its sister species, P. posthon, by having only sinistral males, that is, with aproctal side of body on the left. The hooklike toxactinium arises on right side of head and curves very strongly under head towards left side of body. Males of P. posthon are exclusively dex- tral. Four characters distinguish the sister species from Phenacostethus smithi: distal portion of pe- nis smooth; penial bone absent; ctenactinium small or absent; and stout and distinctly curved, hooklike toxactinium (see Fig. 5). Males of P. smithi are either sinistral or dextral and occur in about equal numbers (see Introduction, Rela- tionships, Bilateral Asymmetry, and Table 2). A second diagnostic character is a minute eye- lens (Fig. 4b), as compared with the relatively large eye-lens of P. posthon and P. smithi (Fig. FIGURE 2. Phenacostethus trewavasae, left lateral view of head and anterior portion of body of 14.1 -mm holotype (ROM 41826). 228 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 10 FIGURE 3. Diagrammatic representation of dorsal mela- nophore pattern, Phenacostethus trewavasae, male paratype (ROM 44289). 4a), more typical of that found in diminutive teleost fishes (see below). DESCRIPTION.— Slender, laterally compressed diminutive phallostethid fish, distinguished by minute eye-lens and by males with sinistral pria- pium. Phenacostethus trewavasae like congeners in meristic data (Table 1). No vestigial pelvic fin rays or supports in females; males with pelvic and parts of pectoral fins modified into priapium that is invariably sinistral: prominent external- ized subcephalic bone a toxactinium arising on right side of body, articulating with right axial bone, and curving strongly under head towards left side of body (Fig. 4b). Cartilaginous pulvin- ular pad lateral to and covering articulation point of, toxactinium and axial bone; small antepleural bone just posterior to point of articulation. Mi- nute ctenactinium articulates with posterior base of right axial bone. Penis smooth not ruffled as in P. smithi (see Roberts 197 la). Pleural ribs of fifth? vertebra, each with a posterior flange, elongate dorsoventrally, meeting just dorsal to right axial bone (Fig. 4a). First pleural rib on fourth vertebra in females, fifth vertebra in males. Skull, gill arches, and jaws like those illustrated for Phallostethus dunckeri Regan by Parenti (1984), with following qualifications. Frontals project above dorsal head profile (Fig. 1 , 2, 4b). Three infraorbital bones: preorbital, second in- fraorbital, and dermosphenotic. Outer jaws with few unicuspid teeth; paradentary with cartilagi- nous core and slight perichondral ossification, no teeth. Lower jaw protrudes beyond anterior ex- tent of upper jaw. Submaxillary element cartilag- inous. Rostral cartilage pear-shaped, wider pos- teriorly. Two small accessory cartilages between medial ramus of maxilla and rostral cartilage (as in Ceratostethus bicornis, Roberts 19716, fig. 5). Gill arch skeleton highly cartilaginous. Unicus- pid teeth on fourth ceratobranchial and infra- pharyngobranchial toothplates. Three cartilagi- nous basibranchials posterior to cartilaginous basihyal. Caudal skeleton with two epurals and autoge- nous parhypural. Caudal fin forked, dorsal and ventral rays forming incipient lobes. Pectoral fin narrow and elongate. Two dorsal fins, the first with a single spine or ray supported by single pterygiophore. Ventral dermal keel extending from base of priapium in males or urogenital opening in fe- males, to anal fin origin. Scales on body small and deciduous, absent from dorsal surface of head. Color pattern in alcohol similar to that of congeners (as in Rob- erts, 197 la, confirmed by personal observation): melanophores scattered on dorsal surface of head and anterior portion of body (Fig. 3), along mid- lateral intermuscular septum, around orbit, on operculum and priapium, and along basal por- tion of anal fin, dorsal midline, and ventral mid- line. Ground coloration very pale yellow or light brownish in alcohol; coloration in life unknown, although P. trewavasae is probably nearly trans- lucent in life, as are its congeners. Largest spec- imens reported by Roberts (197 la: 13- 14) of P. posthon and P. smithi with a bright orange yellow bar on caudal peduncle and a smaller orange yellow bar on the body "next to the anal fin origin." PARENTI: PHALLOSTETHID FISHES 229 ETYMOLOGY.— trewavasae, in honor of Dr. Ethelwynn Trewavas, British Museum (Natural History), to express my deep appreciation of her continued contribution to the field of ichthyol- ogy. EYE-LENS SIZE The eye-lens is a nearly perfect sphere at the center of the eyeball (Fig. 4). In P. trewavasae (Fig. 4b), the eye-lens is minute compared with that of P. smithi (Fig. 4a). A minute eye-lens has been observed in all seven of the cleared and stained specimens of P. trewavasae. Four of these specimens were chosen at random from a lot of alcohol-preserved specimens for study: a mature male, an immature male, and two adult females (ROM 44290) that were counterstained for bone and cartilage. Three of the seven specimens— two mature males and one adult female (ROM CS 812)— stained solely with alizarin, were not prepared by me, but were probably chosen at random for preliminary identification at ROM. The presence of a minute eye-lens has been con- firmed, by dissection, in alcohol-preserved spec- imens. Size of the eye-lens varies from minute to barely detectable with a dissecting microscope, so that the character of a minute eye-lens may represent a stage in a transition series from a small eye-lens to eye-lens absent. The ratio of the distance between the center of the lens and the retina to the radius of the lens is nearly a constant in adult teleost fishes. This constant of 2.55, known as Matthiessen's ratio, has been demonstrated in numerous teleosts, and has been confirmed in the cichlid Haplochromis elegans Trewavas (Otten 1981). During growth of H. elegans, the ratio increases rapidly from about 2.2 to 2.8, then decreases slowly to about 2.5 before leveling off at about 2.55 in the adult (Otten 1981, fig. 15). Matthiessen's ratio in P. trewavasae could not be measured directly as part of this study. How- ever, a minute eye-lens at the center of the eyeball and a normal retina will not affect the distance from the center of the lens to the retina. But, obviously, Matthiessen's ratio will be greatly in- creased by a small eye-lens radius, and the dis- tance from the center of the lens to the retina will be greater than the focal-length of the eye- lens. Visual acuity at any given stage in ontogeny is a function of retinal structure as well as shape of the eyeball (Otten 1981; Levine and MacNichol 1982; Fernald 1985). Growth of the eye-lens is probably retarded very early in ontogeny. Struc- ture and function of the retina, as well as other accommodation made during ontogeny for a mi- nute eye-lens, is unknown. Without such infor- mation, the quality of natural vision in P. tre- wavasae will remain open to speculation, but several statements can be made. First, P. trewavasae may have poor visual acu- ity simply because of the optical properties of a minute eye-lens (Kirschfeld 1976). The short fo- cal length of the eye-lens can be correlated with low resolving power, decreased ability to distin- guish among wavelengths of light, and high chro- matic aberration. In very small lenses, absolute aperture limits resolving power (Otten 1981:681). Second, P. trewavasae lives in clay- and gravel- bottom, freshwater streams of the Baram River, Sarawak. The species is apparently omnivorous, with sample gut-contents including, for example, larval or juvenile P. trewavasae and adult dip- terans. Field notes state that there was no vege- tation at P. trewavasae collecting sites. If the mi- nute eye-lens limits the visual acuity of P. trewavasae, then we may assume that the species does not seek out prey visually. RELATIONSHIPS Myers (1928) distinguished Phenacostethus from Phallostethus by a shorter anal fin (Table 1), a protruding lower jaw, and the absence in the female of a groove on the abdomen. Regan ( 1 9 1 3, 1 9 1 6) did not state whether Phallostethus dunckeri has a spinous first dorsal fin, and Myers (1928) and Roberts (197 la, b) could only spec- ulate about its presence. The first dorsal fin is absent in the syntypes in the BMNH (Parenti 1984) and absent in the material in the ZMH. Further, Myers (1928) said that Phenacoste- thus resembled Phallostethus and differed from Neostethus (and, in fact, from all Neostethinae) in having a priapium that has a prominent hook- like anterior element (the toxactinium) and a shieldlike pulvinular pad. Nevertheless, these elements, or their homologs, are present in most phallostethids. They are well developed and hence are the prominent priapial elements in Phallo- stethus and Phenacostethus. Division of phallostethids into two groups em- phasized gross differences between the types of priapia. However, no assessment of whether one 230 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 10 tox lens B ax FIGURE 4. Left lateral view of head and anterior portion of body of cleared and stained preparations of male a. Phenacostethus smithi (MCZ 47299), and b. Phenacostethus trewavasae (ROM 44290), focused on eye-lens. Abbreviations: tox = toxactinium; pul = pulvinular pad; ant = antepleural; ax = axial bone. PARENTI: PHALLOSTETHID FISHES 231 or both represented a derived priapium has been incorporated into a classification. For example, Roberts's (19716:396) branching diagram of phallostethid genera clearly indicates a paraphy- letic Neostethinae. Furthermore, he interprets the neostethinae priapium as primitive (Roberts 19716:395): "The priapium of Neostethinae, in which the only externalized elements are derived from pelvic spines and rays, is evidently more primitive than [the priapia of all other phallo- stethids]." Phallostethus and Phenacostethus together can be defined as monophyletic by the following shared derived characters (some characters mod- ified from Roberts 1 97 1 a: 5-6; his numbering not followed): 1. Slender, elongate atherinomorph fishes, with deciduous scales; diminutive— maximum stan- dard length recorded 17.0 mm (Roberts 197 la). 2. Body translucent or transparent, melano- phores scattered on top of head (Fig. 3), middle of dorsum, midlateral intermuscular septum, priapium, and basal and distal portion of fin rays (Fig. 1, 2). 3. Dorsum of head with translucent, membra- nous dome (not as noticeable in alcohol-pre- served specimens owing to dehydration). 4. Teeth on premaxilla, paradentary, and den- tary small, fewer in number than in other phal- lostethids and atherinomorphs; no large outer teeth on lateral ramus of premaxilla. 5. Main externalized bone of priapium a tox- actinium (Regan 1916; Myers 1928). 6. Large, oval, concave, cartilaginous pulvin- ular pad covering point of articulation of tox- actinium with axial bone (Regan 1916; Myers 1928). 7. Coiled vas deferens terminates in fleshy gen- ital pore or penis that projects from posterior section of priapium. 8. Pelvic spines or rays reduced or absent. 9. Vas deferens highly coiled, forming what has been termed an "epididymus" (Regan 1913, 1916). Most of these characters represent reductions; that is, we might think of Phallostethus and Phenacostethus as diminutive phallostethids that, perhaps because of small size, have lost or re- duced characters such as pelvic spines and fin rays, complete squamation and heavy pigmen- tation, and more complete, fuller outer dentition. However, the priapium of Phallostethus and Phenacostethus cannot be regarded as a reduced T3 114 113 111 ho 11-9 FIGURE 5. Cladogram of relationships among the four species of Phallostethus and Phenacostethus. Black squares represent one or more synapomorphies; open squares represent one or more symplesiomorphies. Character 1) slender, elongate, di- minutive fishes with deciduous scales; 2) body translucent or transparent, melanophores scattered on top of head, middle of dorsum, midlateral intermuscular septum, priapium, and basal and distal portion of fin rays; 3) dorsum of head with translucent, membranous dome; 4) teeth on premaxilla, para- dentary and dentary small; 5) main externalized bone of pria- pium a toxactinium; 6) large, oval, concave, cartilaginous pul- vinular pad covering point of articulation of toxactinium with axial bone; 7) coiled vas deferens terminates in fleshy genital pore or penis; 8) pelvic spines or rays reduced or absent; 9) vas deferens highly coiled; 10) protruding lower jaw; 1 1) distal portion of penis smooth; 12) penial bone absent; 13) ctenac- tinium small or absent; 1 4) stout and distinctly curved hooklike toxactinium. See text for defining characters of each species. character complex. Several elements, including the pulvinular and the externalized toxactinium (homologous with the internalized secondary pulvinular; Roberts 19716), are more well de- veloped than in other phallostethids. The fact that pelvic fin rays at the base of the priapium are absent in Phallostethus and Phenacostethus, whereas rudimentary rays are present in neo- stethines, is interpreted as a derived condition representing a further modification of the pelvic fin supports and rays. Relationships among the four species of phal- lostethines are summarized in the cladogram in 232 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 10 TABLE 1 . MERISTIC CHARACTERS OF PHALLOSTETHUS AND PHENACOSTETHUS Branchi- First Second Pectoral ostegal dorsal rays dorsal rays Anal rays rays rays Vertebrae Phallostethus dunckeri 0 8-10 26-28 9-10 4 40 Phenacostethus trewavasae 1 6 14-15 9-10 5 34 Phenacostethus posthon 1 5-6 14-15 9-10 4 34-35 Phenacostethus smithi 1 6-7 14-15 9-11 4-5 33-35 Figure 5. Phallostethus is readily defined by its extremely long anal fin, relatively high number of vertebrae (Table 1), serrated ctenactinium (Regan 1913), and absence of the first dorsal fin (Parenti 1984). The last two characters are also found in some neostethines. Phenacostethus is defined as monophyletic by the protruding lower jaw (Myers 1 928). Myer's additional character, absence of a groove on the abdomen of females, seems to be correlated with quality of preservation, and therefore is not used here. Phallostethus and Phenacostethus are not syn- onymized here solely for reasons of tradition. Phenacostethus trewavasae resembles P. smithi in that the first and second dorsal fins are sep- arated by a relatively large distance, as opposed to being rather close together, as in P. posthon (see Roberts 197 la, fig. 2, 3). The distance be- tween the first and second dorsal fin is a primitive character in P. smithi and P. trewavasae, and serves as a defining character of P. posthon. Roberts (197 la) described P. posthon as re- duced in a number of character states relative to P. smithi, then its only congener. Three character states that may be described as shared reductions in P. posthon and P. trewavasae are: (1) distal portion of penis smooth, as opposed to being ruffled as in P. smithi; (2) penial bone absent; (3) ctenactinium, if present, small and barely de- tectable. These characters were illustrated in both P. posthon (Roberts 197 la, fig. 6) and P. smithi (Roberts 1 97 la, fig. 7). (Figures and captions were switched inadvertently when printed, as noted in a published erratum.) These three reductions are considered synapo- morphies of P. posthon and P. trewavasae. They are correlated with a stouter and more distinctly curved hooklike toxactinium (compare Fig. 4a and 4b). Thus, even though the two species have characters that can only be described as reduc- tions, these characters are treated as synapo- morphies in phylogenetic reconstruction because of their correlation with uniquely derived char- acters (Fig. 5). BILATERAL ASYMMETRY A. Anatomical Homology Bilateral asymmetry is well documented in fishes (Hubbs and Hubbs 1945). The phenom- enon usually concerns reproductive structures offset to, or more complex on, one side of the body; although, the well-known asymmetry of flatfishes (pleuronectiforms) is not necessarily as- sociated with reproduction. TABLE 2. BILATERAL ASYMMETRY IN FIVE SPECIES OF PHALLOSTETHIDAE (SENSU LATO) Total Dextral males Sinistral males Females Juveniles or undetermined Phallostethus dunckeri (a) 25 4 1 16 4 Phenacostethus trewavasae (b) 148 0 63 57 28 Phenacostethus posthon (c) 237 107 0 99 31 Phenacostethus smithi (d) 334 179 155 0 0 (e) 31 4 6 10 11 Mirophallus bikolanus (f) 300 173 0 108 19 (a) BMNH 1913.5.24:18-20,21,22, ZMH 193-195; (b) total of known specimens, catalog numbers in text; (c) MCZ 47300, 47301 A,B, USNM 229302; (d) Hubbs and Hubbs 1945:290, table XIX; (e) BMNH 1927.12.29:1-10; (f) AMNH 50592, CAS 50722, CAS 53165, UMMZ 21 1665. PARENTI: PHALLOSTETHID FISHES 233 Regan (1916:23) was the first to use the terms "sinistral" and "dextral" to refer to the orien- tation of the priapium: "In its asymmetry and in being either dextral or sinistral the priapium agrees with the copulatory organ of Anableps," a possibly bilaterally asymmetric killifish genus (see Hubbs and Hubbs 1945:289-291). Accord- ing to the convention established by Regan, in a sinistral male, the anal opening is on the right, what he termed the "proctal side." Hence, the left side of a sinistral male is termed the "aproctal side." The opposite is true for a dextral male, which may be thought of as the mirror image of a sinistral conspecific male. All female phallo- stethines examined exhibit no bilateral asym- metry; the anus is just anterior to the urogenital opening along the ventral midline under the throat. (See Parenti in press for report of bilateral asymmetric females in one neostethine species.) Regan (1916:21) also pointed out marked dif- ferences between priapia of phallostethines and neostethines: "The approximate symmetry of the priapial ribs and cleithra in Phallostethus, as compared with their marked asymmetry in Neo- stethus, is no doubt due to the symmetrical at- tachment of the priapium in the former . . . and its asymmetrical attachment ... in the latter." Proportions of sinistral and dextral males of Phenacostethus smithi were shown to be equal (Hubbs and Hubbs 1945; Table 2, herein). It was assumed that, with sufficient sample sizes, phal- lostethid species would be represented by more or less equal numbers of sinistral and dextral males, until Roberts (197 la) described Phenac- ostethus posthon, an exclusively dextral species. Roberts (197 la) termed P. posthon sinistral. However, it is more properly called dextral in keeping with homologies of priapia among all phallostethids. Confusion about whether to call a male sinistral (a left) or dextral (a right), may be traced to a casual statement by Myers (1928: 5): "The proctal side may be indifferently either the right or the left of the fish; in other words, the males are either 'rights' or 'lefts.' " In inter- preting this statement strictly, one would assume Myers meant that a male with the proctal side on the right should be called a dextral, or a right. However, this is contrary to the terminology es- tablished by Regan and followed by most phal- lostethid systematists (e.g., Herre 1942; and Myers 1 928 himself). Regan emphasized that the proctal side was away from the female during copulation, and that the aproctal side, the side with the fleshy genital pore or penis, was the obviously functional side of the male with regard to internal fertilization (see also Villadolid and Manacop 1934). Herre (1942:139) followed Regan's conven- tion but was ambiguous when describing asym- metry: "The coiled, enlarged vas deferens lies within the posterior end of the priapium, from which its penis-like tip projects. The proctal side may be either side, so that males of the same species may be either 'rights' or 'lefts.' " Hubbs and Hubbs (1945:290, table XIX) followed Re- gan strictly and documented bilateral asymmetry by tabulating the "location, left (L) or right (R), of aproctal side of males of Phallostethidae Phe- nacostethus smithi." Division of phallostethids into two families was based primarily on the type of prominent external priapial bones. Phallostethines have a hooklike toxactinium that articulates with the axial bone (Fig. 4), a homolog of the pelvic fin girdle (Bailey 1936; Aurich 1937), and curves underneath the head toward the aproctal side (Regan 1916; Herre 1 942; Fig. 2, 4 herein). Hence, a male phallostethine with a toxactinium arising on the right side of the head and curving toward the left, aproctal side, is called sinistral because the aproctal side is the left. Such sinistral phal- lostethine males also have a rudimentary cten- actinium that articulates with the left, aproctal axial bone at its posterior extent. The prominent externalized priapial bones of neostethines are the one or two ctenactinia that arise on the aproctal side of the body. Hence, a male neostethine with one or two ctenactinia arising on the left side of the body is termed sinistral not because of the position of these prominent priapial bones, but because the aproc- tal side is the left. This terminology should be adhered to strictly because of the consistency with the inferred homology of priapial structures. Fur- thermore, this convention for describing bilat- eral asymmetry of male phallostethids should be followed because the division between phallo- stethines and neostethines is not supported by unambiguous, derived characters. Roberts (197 la: 13), in describing Phenaco- stethus posthon, was explicit in describing bilat- eral asymmetry: ". . . the priapium is invariably sinistral (toxactinium arising on left side) in the material examined." It is clear, therefore, that 234 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 10 Roberts did not follow the convention estab- lished by Regan. Hence, I recommend that P. posthon be referred to as an exclusively dextral species, not sinistral as Roberts described it. This correction need not be made to Roberts's ( 1 97 1 b) discussion of the anatomy of Ceratostethus bi- cornis (Regan) because that species is a neo- stethine and the prominent external priapial bones are the ctenactinia, which arise on the aproctal side in every known example. B. Unique or Fixed Asymmetry Of the 148 known specimens of Phenacoste- thus trewavasae, 63 are sinistral males, 57 are females, and 28 are juveniles or of otherwise undetermined gender (see Diagnosis and Table 2). The collection of large samples of the exclu- sively sinistral Phenacostethus trewavasae allows us to conclude with certainty that unique or fixed asymmetry is a natural phenomenon in some phallostethids. In addition to the phallostethines P. trewavasae and P. posthon, the neostethine Mirophallus bikolanus Herre has males of fixed asymmetry (Tyson R. Roberts, personal com- munication). Three large lots of M. bikolanus, all collected from the Cabangan River, Albay Province in the Bicol (Bikol) region of Luzon, Philippine Islands, contain dextral males only. Of 300 specimens, 173 are dextral males, 108 are females, and 19 are juveniles or of undeter- mined gender (Table 2). The relatively high num- ber of male and low number of juvenile or un- determined M. bikolanus is probably related to the fact that immature males are readily iden- tifiable as such by the presence of a heavily pig- mented anal region. Unique or fixed asymmetry is a natural phe- nomenon in other atherinomorph fishes, as re- viewed by Hubbs and Hubbs (1945). Females of the ricefish Horaichthys setnai Kulkarni may be considered sinistral in that the urogenital open- ing is to the left of the midline in a majority of females, and the right pelvic fin girdle and rays are absent in females (Kulkarni 1940; Hubbs 1941). Males of//, setnai have an anal fin mod- ified into an elaborate gonopodium that is not bilaterally asymmetric as far as known. The vi- viparous poeciliids, Carlhubbsia kidderi (Hubbs) and Xenodexia ctenolepis Hubbs have a concav- ity on the right side of the gonopodium (Hubbs and Hubbs 1 945). Males of the latter species also have a right pelvic fin modified into a so-called "pectoral clasper" (Hubbs 1950), and a "... thickened fleshy ridge along the ventromesial edge of the proximal third of the outer ray of the right pelvic fin" (Rosen and Bailey 1963:143). Phenacostethus posthon and P. trewavasae are sister species, the males of which are nearly mir- ror images of each other. Apart from type of bilateral asymmetry, they differ in several char- acteristics of priapial structure, placement of fins, and relative size of eye-lens (see Diagnosis, De- scription, and Relationships). One might assume that the common ancestor of these sister species, like most other phallostethids, contained both sinistral and dextral males (Fig. 5). One might assume further that it was the separation of the ancestral species into a sinistral and dextral pop- ulation that precipitated (or, in fact, was) the speciation event. The problem with such a spe- ciation hypothesis is that it presents a series of untestable statements, the first concerning states of the ancestral species, the second concerning isolation of sinistral and dextral subgroups. Experimental data are needed to answer the questions: What is the genetic basis of bilateral asymmetry in phallostethids? Does a male phal- lostethid determine type of bilateral asymmetry of offspring? That is, does a sinistral male have only sinistral male offspring, and likewise, does a dextral male have only dextral male offspring? Breeding experiments to answer these questions, performed when live phallostethids are available for study, will further our understanding of the evolution of bilateral asymmetry, and the special case of fixed or unique asymmetry, in phallo- stethids. CONCLUSIONS Phenacostethus trewavasae new species, is de- scribed from the Baram River, Sarawak. It is the first phallostethine species known from Borneo. This subfamily had been reported previously from Thailand and peninsular Malaysia. Two characters distinguish P. trewavasae from all other phallostethines: a minute eye-lens and males that are exclusively sinistral with regard to orientation of priapial structures. We may hy- pothesize reduced visual acuity in P. trewavasae because of the size of the eye-lens; however, a clear statement on vision awaits knowledge of structure of the retina. All phallostethid males are bilaterally asym- metric, described by position of the anus: in si- PARENTI: PHALLOSTETHID FISHES 235 nistral males, the aproctal side is the left; in dex- tral males, the aproctal side is the right. Females exhibit no apparent asymmetry. In most species, sinistral and dextral males are represented in more or less equal numbers. Sam- ple sizes of the exclusively sinistral Phenac- ostethus trewavasae and the exclusively dextral P. posthon and Mirophallus bikolanus allow us to conclude with certainty that unique or fixed asymmetry is a natural phenomenon in some phallostethids. ACKNOWLEDGMENTS This study would not have been possible with- out the loan of recently collected phallostethid fishes from Sarawak generously provided by Dr. E. J. Grossman, ROM. For loans of additional comparative material, I thank Ms. M. Norma Feinberg and Dr. Gareth Nelson, AMNH; Dr. Barry Chernoff and Mr. William Saul, ANSP; Ms. Bernice Brewster, Dr. P. Humphry Green- wood, and Mr. Gordon Howes, BMNH; Mr. Karsten Hartel and Dr. Karel Liem, MCZ; Dr. Robert R. Miller and Mr. Douglas Nelson, UMMZ; Ms. Susan Jewett and Dr. Richard P. Vari, USNM; and Prof. H. Wilkens, ZMH. Dr. Tyson R. Roberts has continued to be a source of information on Southeast Asian fishes, and a willing discussant of all aspects of fish bi- ology and systematics. His comments and those of several reviewers greatly improved the manu- script. Members of the CAS Department of Ich- thyology staff provided assistance during the course of this study, including Dr. M. Eric An- derson, Dr. Stuart G. Poss, and Ms. Pearl M. Sonoda. Mr. Jim Patton, CAS Department of Photog- raphy, prepared the photographs in Figures 1 and 2. My studies of phallostethid fishes began during the tenure of a North Atlantic Treaty Organi- zation postdoctoral fellowship at the BMNH where I had the opportunity to benefit from the wisdom and experience of Dr. Ethelwynn Tre- wavas. Support of this project by the National Science Foundation through grant DEB 83-15258 is gratefully acknowledged. LITERATURE CITED AURICH, H. 1937. Die Phallostethiden (Unterordnung Phal- lostethoidea Myers). Int. Rev. Gesamten. Hydrobiol. Hy- drogr. 34:263-286. BAILEY, R. J. 1936. The osteology and relationships of the phallostethid fishes. J. Morphol. 59(3):453-483. DINGERKUS, G. AND L. D. UHLER. 1977. Enzyme clearing of alcian blue stained whole small vertebrates for demonstra- tion of cartilage. Stain. Tech. 52(4):229-232. DUNCKER, G. 1 904. Die Fische der Malayischen Halbinsel. Mitt. Naturh. Mus. Hamburg 21:135-207. FERNALD, R. D. 1985. Growth of the teleost eye: novel so- lutions to complex constraints. Environ. Biol. Fishes 13: 113-123. FOWLER, H. W. 1937. Zoological results of the third de Schauensee Siamese Expedition, Part VIII— fishes obtained in 1936. Proc. Acad. Nat. Sci. Philadelphia 89:125-264. HERRE, A. W. 1942. New and little known phallostethids, with keys to the genera and Philippine species. Stanford Ichthyol. Bull. 2(5): 137-1 56. HUBBS, C. L. 1941. A new family of fishes. J. Bombay Nat. Hist. Soc. 42:446-447. . 1950. Studies of cyprinodont fishes. XX. A new subfamily from Guatemala, with ctenoid scales and a uni- lateral pectoral clasper. Misc. Publ. Mus. Zool. Univ. Mich- igan, no. 78:1-18. HUBBS, C. L. AND L. C. HUBBS. 1945. Bilateral asymmetry and bilateral variation in fishes. Pap. Michigan Acad. Sci. Arts Letters 30:229-310. KJRSCHFELD, K. 1976. The resolution of lens and compound eyes. Pp. 354-360 in Neural principles in vision, F. Zettler and R. Weiler, eds. Springer Verlag, Berlin. KULKARNI, C. V. 1 940. On the systematic position, structural modifications, bionomics and development of a remarkable new family of cyprinodont fishes from the province of Bom- bay. Rec. Indian Mus. 42:379-423. LADIGES, W., G. VON WAHLERT, AND E. MOHR. 1958. Die Typen und Typoide der Fischsammlung des Hamburgischen Zoologischen Staatsinstituts und Zoologischen Museums. Mitt. Hamburg Zool. Inst. 56:155-167. LEVINE, J. AND E. F. MACNICHOL. 1982. Color vision in fishes. Sci. Am. 246(2): 140- 149. MYERS, G. S. 1 928. The systematic position of the phallo- stethid fishes, with diagnosis of a new genus from Siam. Am. Mus. Novitates, no. 295:1-12. OTTEN, E. 1981. Vision during growth of a generalized Hap- lochromis species: H. elegans Trewavas 1933 (Pisces, Cich- lidae). Neth. J. Zool. 31(4):650-700. PARENTI, L. R. 1984. On the relationships of phallostethid fishes (Atherinomorpha), with notes on the anatomy ofPhal- lostethus dunckeri Regan, 1913. Am. Mus. Novitates, no. 2779:1-12. . In press. Homology of pelvic fin structures in female Phallostethid fishes (Atherinomorpha, Phallostethidae). Co- peia. REGAN, C. T. 1913. Phallostethus dunckeri, a remarkable new cyprinodont fish from Johore. Ann. Mag. Nat. Hist. 12:548- 555. . 1916. The morphology of the cyprinodont fishes of the subfamily Phallostethinae, with descriptions of a new genus and two new species. Proc. London Zool. Soc. 1916: 1-26. ROBERTS, T. R. 197 la. The fishes of the Malaysian family Phallostethidae (Atheriniformes). Breviora, no. 374:1-27. . 1971 b. Osteology of the Malaysian phallostethoid fish Ceratostethus bicornis, with a discussion of the evolution 236 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 10 of remarkable structural novelties in its jaws and external Oryzias, and the groups of atherinomorph fishes. Am. Mus. genitalia. Bull. Mus. Comp. Zool. 142(4):393-418. Novitates, no. 2719:1-25. ROSEN, D. E. AND R. M. BAILEY. 1963. The poeciliid fishes VILLADOLID, D. V. AND P. R. MANACOP. 1934 (Issued 1935). (Cyprinodontiformes), their structure, zoogeography, and The Philippine Phallostethidae, a description of a new species, systematics. Bull. Am. Mus. Nat. Hist. 126(1):1-176. and a report on the biology of Gulaphallns mirabilis Herre. ROSEN, D. E. AND L. R. PARENTI. 1981. Relationships of Philippine J. Sci. 55(3): 193-220. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 11, pp. 237-267, 40 figs., 2 tables. May 6, 1986 LAND MOLLUSKS (GASTROPODA: PULMONATA) FROM EARLY TERTIARY BOZEMAN GROUP, MONTANA By Barry Roth Museum of Paleontology, University of California, Berkeley, California 94720 ABSTRACT: The Bozeman Group consists of fluvial, eolian, and lacustrine rocks deposited in intermontane basins of western Montana after the Laramide Orogeny. In the Three Forks Quadrangle, three early Tertiary formations have yielded land mollusk fossils. The Milligan Creek Formation (of probable Eocene age) con- tains snails of the genera Gastrocopta (two species), Radiocentrum, and Helminthoglypta; the Climbing Arrow Formation (middle or late Eocene) contains Gastrocopta and Polygyrella;tlie Dunbar Creek Formation (latest Eocene or early Oligocene) contains Gastrocopta (two species), Pupoides (two species), Radiocentrum, and Helminthoglypta. Three species of Gastrocopta, one of Pupoides (Ischnopupoides), two of Radiocentrum, and one of Helminthoglypta are described as new. Two others (Pupoides and Polygyrella) are scarcely distinguish- able from extant species. No interregional correlations are suggested because the land mollusk faunas of the western interior are too spottily known at present. In the Bozeman Group, genera and species groups that are now allopatric occur together. The land mollusks indicate a change in terrain through time: from sparsely vegetated to forested and back again. Climates were temperate and, at least toward the end of the interval, seasonally variable. Numerous land mollusk taxa in upper Cretaceous and Tertiary rocks of western North America occur outside the Holocene ranges of their families and genera. It is suggested that the evolutionary and biogeo- graphic history of North American land mollusks through the Tertiary has involved (1) the sorting out of component taxa into different geographic/adaptive zones; (2) restriction of many forms to lower latitudes, concurrent with climatic cooling; and (3) eastward and westward displacements, probably related to avail- ability of rainfall. For land mollusks, the late Eocene-early Oligocene was a time not so much of evolutionary innovation as of local extinction and biogeographic rearrangement. INTRODUCTION acterized in greater detail. Table 1 presents a summary of the fauna. Land and freshwater mollusks from early Ter- Robinson (1963) presented a detailed account tiary continental sediments of the Bozeman of the geology of the Three Forks Quadrangle. Group in the Three Forks Quadrangle, southwest He denned the Bozeman Group— which he an- Montana (Fig. 1), have been reported in check- ticipated would be recognizable on a regional lists by Taylor (in Robinson 1963; Taylor 1975). scale— as the Tertiary fluvial, eolian, and lacus- The terrestrial gastropods were not figured or trine rocks that accumulated in the basins of discussed taxonomically; many were not iden- western Montana after the Laramide Orogeny, tified beyond family. However, the assemblage In the Three Forks Quadrangle the group consists is an unusual one and bears strongly on the origins mainly of four formations (Fig. 2). The Sphinx of present-day American land mollusk faunas. Conglomerate, stratigraphically the lowest, is a Preservation of the fossils ranges from fair to limestone conglomerate probably originating as excellent. Seven species are represented by ma- an alluvial apron; it is not fossiliferous. terial good enough to permit description of them Next lowest is the Milligan Creek Formation, as new herein, and nearly all taxa can be char- consisting of light-colored, fine-grained, tuflfa- [237] 238 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 112* FIGURE 1 . Index map of southwestern Montana. Three Forks Basin diagonally hatched; other intermontane basins stippled. After Robinson (1961). ceous lake deposits that range from limestone to calcareous mudstone. Where it crops out, in the southwest part of the quadrangle, it produces whitish dissected benchlands. Its maximum thickness is probably not much greater than 90 m. Fossils from the Milligan Creek Formation include five taxa of land snails, all from one lo- cality (USGS 20007). Freshwater mollusks are evidently more widespread and are reported from five additional localities (Taylor 1975). The freshwater gastropod Lymnaea is also present at USGS 20007. Ostracods and charophyte re- mains occur in the formation. The age is prob- ably Eocene but none of the fossils is age-diag- nostic. The limestone and other fine-grained rocks of the Milligan Creek Formation were deposited in a perennial lake. Partly based on the snail here described as Radiocentrum laevidomus, Robin- son (1963) inferred that the lake basin lay in mountainous terrain not much different from that of the present. The Climbing Arrow Formation conformably overlies the Milligan Creek Formation in places but elsewhere may have formed contempora- neously with it. The Climbing Arrow is made up of olive, thick-bedded, sandy bentonitic clay and coarse sand, with subordinate light-colored silt- stone, sandstone, conglomerate, and limestone. TABLE 1 . OCCURRENCE OF LAND MOLLUSKS IN EARLY TERTIARY BOZEMAN GROUP IN THREE FORKS QUADRANGLE, SOUTHWEST MONTANA. Locality numbers are those of U.S. Geological Survey Cenozoic Series. Mil- ligan Creek Climbing Forma- Arrow tion Formation Dunbar Creek Formation Taxon 20007 20008 20009 20011 20012 20013 20014 20015 20016 20017 Class Gastropoda Subclass Pulmonata Family Pupillidae Gastrocopta (Albinula) montana n. sp. G. (A.) sp. a G. (A.) sagittaria n. sp. G. cordillerae n. sp. Pupoides (Ischnopupoides) tephrodes n. sp. P. (/.) sp., cf. P. (I.) hordaceus (Gabb, 1866) Family Oreohelicidae Radiocentrum taylori n. sp. R. laevidomus n. sp. Family Ammonitellidae Polygyrella sp., cf. P. polygyrella (Bland and Cooper, 1861) Family Helminthoglyptidae Helminthoglypta bozemanensis n. sp. cf. cf. X X X X - - - cf. XXX -XX- X X cf. cf. ROTH: EARLY TERTIARY LAND MOLLUSKS 239 It is extensively exposed in the Three Forks Quadrangle and tends to form subdued topog- raphy of low, rounded hills separated by broad, smooth valleys. The formation is not less than 220 m thick, and may be fully twice that. Two fossil localities (USGS 20008, 20009) have yield- ed three taxa of land snails. Freshwater gastro- pods are also present in the formation, including one, Physal, from USGS 20009 (Taylor 1975). The Climbing Arrow Formation ranges in age from middle or late Eocene to early Oligocene. A small assemblage of vertebrates from the lower part of the formation was assigned a probable Uintan age (middle to late, but not latest, Eocene) (G. E. Lewis in Robinson 1963). Pipestone Springs (late Eocene) and Chadronian (latest Eocene to early Oligocene) vertebrates are known from higher in the formation (Hough and Lewis in Robinson 1963). A single locality well down in the lowest stratigraphic unit of the Climbing Arrow has yielded the freshwater planorbid gas- tropod Biomphalaria pseudammonius (Schlot- heim), diagnostic of middle to late Eocene age and a tropical climate (McKenna et al. 1962; Taylor 1985). This locality is several hundred meters stratigraphically below the Uintan ver- tebrate locality. The terrestrial snail localities are probably in the Uintan rather than the Chad- ronian part of the formation. A diverse fossil microflora exists but has not been studied (Rob- inson 1963). The Climbing Arrow originated largely as the product of an aggrading stream system. The coarser sediments are stream-chan- nel deposits; the finer-grained ones, evidently overflow deposits that accumulated on the flood- plains in short-lived ponds and lakes. The Dunbar Creek Formation, stratigraph- ically the highest named formation of the group, consists of white to grayish yellow thick-bedded tuffaceous siltstone, partly lacustrine and partly eolian in origin, laced with fluvial sandstone and conglomerate; minor limestone and bentonitic clay are also present. The formation is 80-240 m thick in the Three Forks Quadrangle and forms a topography much like that of the Milligan Creek— white benchlands rising steplike from the floodplain, dissected by many steep-walled can- yons. The constituent sediments were evidently deposited in a more or less enclosed local basin (part of the larger Three Forks structural basin) which, at least part of the time, contained stand- ing water. Whether the ash of a particular stra- co * « » 0> ~" « ™ ,J _ Rock Units Ma Q\ 0) CO CO E E < E • CO Z 2 LU Z Z < LU <_X>-^-N^^. ? ^v> ^-~ O o 0 -34 CC Dunbar Creek _ 0 _l < Formation o X o a -36 3 O cc HP O ^-^_ - < Y////// c / i t- Y/y//// CO -38 ~- E LU O Climbing Arrow N •™ t— O Formation LU CD Z -40 LU z O < o H LU Z Z_9 ^ A 9 Miingan . *r £. Creek Fm Ak Sphinx Cgl — ^•**-*- ?" -44 FIGURE 2. Stratigraphic and inferred temporal relations of early Tertiary formations of the Bozeman Group, Three Forks Basin, Montana. Filled triangles indicate sources of land mol- lusks reported in this study. Ma = millions of years before present. turn fell in water or on dry land is not always easy to determine, but from the predominance of channeling by "sands and gravel derived from the neighboring highlands and containing little contemporaneous pyroclastic debris," Robinson (1963:80) concluded that if the bulk of the tuff- aceous sediments were deposited in a lake, it must have been shallow and often dry. Seven sampling sites yielded a fauna of six species of terrestrial snails, while two other localities in the formation contain freshwater gastropods (Taylor 240 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 1 975). Thin limestones in the Dunbar Creek For- mation may be evaporites or possibly fossil ca- liche horizons, suggestive of bolson rather than lake conditions. Based on the presence of brontothere and ore- odont bones, Hough and Lewis (in Robinson 1963) assigned a Chadronian (latest Eocene to early Oligocene) age to the lower 80 m of the Dunbar Creek Formation. Vertebrate fossils and land snail remains are not recorded from iden- tical localities, but the land snail localities are within this part of the formation. Other parts of the formation are of less certain age and regarded simply as Oligocene. Remnant patches of middle and late Tertiary sands and gravels, too small and discontinuous to be mapped, also occur in the Three Forks Quadrangle. These are correlated with the Mad- ison Valley beds of Douglass (1903), which else- where in the Three Forks Basin lie with angular unconformity on the Dunbar Creek Formation (Robinson 1 96 1). They have yielded a variety of mammal bones, indicating Miocene and Plio- cene ages, but as yet no molluscan fossils are recorded. The unconformity within the Bozeman Group that divides lower, predominantly fine-grained, Eocene and Oligocene strata from upper, pre- dominantly coarse-grained, Miocene and Plio- cene strata is probably of regional extent. Kuenzi and Fields (1971, fig. 4) correlated similar rock sequences in the Ruby, Jefferson, Three Forks, Townsend, and Clarkston basins. A biota from the Douglass Creek Basin (Konizeski 1961) sug- gests time correlation with the Climbing Arrow Formation. Lillegraven and Tabrum (1983, fig. 2) presented an interbasinal correlation diagram including the Dunbar Creek and Climbing Arrow formations, but the placement of these units with reference to the radiometric time scale remains highly inferential. LOCALITY DESCRIPTIONS Locality numbers given are those of the U.S. Geological Survey Cenozoic series. Altitudes are given in feet as in the original locality register (with metric equivalents supplied) and are cor- rect to ± 1 0 ft. All localities are in the Three Forks Quadrangle (USGS, Topographic, 1:62,500, edi- tion of 1950), southwest Montana. Localities are plotted on map by Robinson (1963, pi. 2). Milligan Creek Formation 20007. NWV4 SW/4 sec. 36, T 1 N, R 1 W; al- titude 4,260 ft (1,300 m). Climbing Arrow Formation 20008. SE'/4 NE'/» NE1/. sec. 1 1, T 2 N, R 1 W; altitude 4,600 ft (1,400 m). 20009. NEVi NE1/. NE1/. sec. 1 1, T 2 N, R 1 W; altitude 4, 5 80 ft (1,400m). Dunbar Creek Formation 2001 1. SE'/4 NW/. sec. 6, T 2 N, R 2 E; altitude 4,360 ft (1,330m). 200 1 2. NW/. NW/» sec. 5, T 2 N, R 2 E; altitude 4,350 ft (1,330m). 200 1 3. NW'/4 SW/4 sec. 3, T 2 N, R 1 E; altitude 4,230 ft (1,290m). 20014. Same location as USGS 20013 but 6 m stratigraphically higher. 2001 5. NW1/. SW1/. sec. 3, T 2 N, R 1 E; altitude 4,235 ft (1,290m). 20016. S'/2 NE1/. SE'/4 sec. 19, T 3 N, R 1 E; altitude 4,450 ft (1,360m). 20017. NW/. SW/4 sec. 3, T 2 N, R 1 E; altitude 4,235 ft (1,290m). SYSTEMATIC PALEONTOLOGY The following institutional abbreviations are used: CAS— Department of Invertebrate Zool- ogy, California Academy of Sciences; USGS— United States Geological Survey; USNM- Di- vision of Paleobiology, United States National Museum of Natural History, Smithsonian Insti- tution. Class GASTROPODA Subclass PULMONATA Order ORTHURETHRA Family PUPILLIDAE Gastrocopta Wollaston, 1878 TYPE-SPECIES: Pupa acarus Benson, 1856, by subsequent des- ignation (Pilsbry 1916-18). DIAGNOSIS.— Shell "rimate or perforate, cylin- dric or ovate-conic, having the angular and pa- rietal lamellae more or less completely united into one biramose, bifid, lobed or sinuous lamella (or rarely the angular is wanting). Columellar la- mella present; palatal folds present . . . ; lip well expanded" (Pilsbry 1948:871). ROTH: EARLY TERTIARY LAND MOLLUSKS 241 REMARKS. — Gastrocopta is the most widely distributed genus of the Pupillidae, with a re- corded stratigraphic range of Eocene to Recent (Pilsbry 1948; Zilch 1959; Preece 1982). Pilsbry distinguished two main geographic groups— a northern, and a tropical and southern continent series— each with several subgenera. The sub- genus Albinula Sterki, 1892, is part of the north- ern group. The "Gastrocopta^ sp." reported by La Rocque (1960) from the Flagstaff Formation, Paleocene of Utah, does not show sufficient detail for as- signment to a subgenus. If correctly allocated to Gastrocopta, it is the oldest known member of the genus. Taylor (1 975) reported an undescribed species of Gastrocopta (Gastrocopta) from the Tepee Trail (=Wagon Bed) Formation, upper Eocene of Wyoming, and other fossils question- ably referred to Gastrocopta from the Wagon Bed and White River formations (lower Oligocene), Wyoming. Pilsbry (1916-18) noted that the Oli- gocene and Miocene species of Europe seem "a little too specialized" to have been ancestral to the American species. It is reasonable therefore to expect ancestral forms in Eocene, Paleocene, and perhaps upper Cretaceous strata in America and elsewhere. (Albinula) Sterki, 1892 TYPE-SPECIES: Pupa contracta Say, 1 822, by original designa- tion. DIAGNOSIS. — "Whitish-translucent gastro- copts having the inner end of the parietal lamella curved towards the periphery; angular lamella well developed, concrescent in varying degree with the parietal; the palatal folds stand upon a white palatal callus, and a suprapalatal fold is usually developed. Except in G. armifera, the columellar lamella is horizontal in front and curves toward the base within. The lip is thin and expanded" (Pilsbry 1948:874). REMARKS.— Albinula occurs in the Eocene of England, the middle Oligocene through upper Miocene of Germany, the Miocene and Pliocene of France, and the Pliocene of Italy (Pilsbry 1916- 18; Preece 1982). It is widespread at present in North America, although absent from the Pacific slope, but there are no other American fossil rec- ords earlier than Pliocene. The Three Forks Quadrangle is just within the western edge of the Holocene range of the subgenus. Gastrocopta (Al- binula) holzingeri (Sterki, 1889) is reported to range west to Helena, Montana (Pilsbry 1 948). Several species of the subgenus, G. armifera (Say, 1821), G. contracta (Say, 1822), G. holzin- geri, G. falcis Leonard, 1946, G. proarmifera Leonard, 1946, and G. tridentata (Leonard, 1946), occur in Pliocene and early Pleistocene faunas in Kansas (Franzen and Leonard 1947; Leonard 1950; Taylor 1960). Gastrocopta ar- mifera also occurs in late Pleistocene deposits in Kansas and Arizona (Franzen and Leonard 1947; Bequaert and Miller 1973), and G. contracta oc- curs in Quaternary deposits in west Texas (Al- britton and Bryan 1939). Gastrocopta (Albinula) montana new species (Figure 3) Pupillidae C, D. W. Taylor in Robinson 1963:68, table 4 (in part). -Taylor 1975:206 (non p. 209). Pupillidae D, D. W. Taylor in Robinson 1963:68, table 4.— Taylor 1975:206. DIAGNOSIS.— A cylindric-ovate Gastrocopta (Albinula) with sinuous, unbranched angular la- mella, sulcate body whorl, strong crest, and low callus ridges inside inner and outer lips. DESCRIPTION.— Shell dextral, umbilicate, cy- lindric-ovate, of about 5.25 whorls; suture lightly to moderately impressed, early whorls strongly convex, later ones less so. Nuclear whorls smooth; neanic sculpture of oblique growth lines, strong- est just anterior to suture. Body whorl narrowly, roundly shouldered anterior to suture, com- pressed at periphery, attenuated toward base, sulcate behind aperture; last 0.3 whorl nearly straight in basal view, then angling toward um- bilicus just behind peristome. Strong crest pres- ent behind aperture. Aperture rounded-triangu- lar; peristome broadly reflected, thickened within by callus, limbs connected by thin callus film across face of body whorl. Outer lip sinuous, most produced medially, outer-posterior quad- rant retractive. With prominent, sinuous, an- gular lamella, projecting not quite as far as plane of aperture, inner end thickened along axial side and deflected toward periphery. Low callus bar- riers present inside inner and outer lips, inner one bearing two faint denticles. Dimensions of holotype: height 2.0 mm, diameter 1.1 mm, whorls 5.2. TYPE-MATERIAL. -Holotype: USNM 377373, from U.S. Geological Survey Cenozoic Locality 20007, Montana: Gal- latin County: NWy4 SW'/4 sec. 36, T 1 N, R 1 W, Three Forks 242 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 FIGURES 3-7: Figure 3. Gastrocopta (Albinula) montana, new species, holotype, USNM 377373; height 2.0 mm. Figure 4. Gastrocopta (Albinula) contracta (Say), CAS 046506; height 2.4 mm; Holocene, Lee County, Texas. Figure 5. Gastrocopta (Albinula) species a, USNM 377375; height 2.75 mm. Figure 6. Gastrocopta (Albinula) sagittaria, new species, holotype, USNM 377376; height 1.9 mm. Figure 7. Gastrocopta cordillerae, new species, holotype, USNM 377378; height 1.8 mm. Quadrangle (1950) 1:62,500; altitude 4,260 ft (1,300 m). Mil- ligan Creek Formation, Eocene (?). Two paratypes, USNM 377374, from same locality as holotype. REFERRED MATERIAL (all, Gastrocopta sp., cf. G. mon- tana).— Dunbar Creek Formation: USGS 20012, two speci- mens. USGS 20013, two specimens. USGS 20017, one spec- imen. REMARKS.— The type-lot consists of well-pre- served original shells filled with colorless calcite matrix. The holotype is an adult shell. One para- type is an adult shell with reflected peristome, height 2. 1 mm, diameter 1 . 1 mm, with 5.3 whorls. The other paratype lacks the adult peristome; dimensions: height 2.1 mm, diameter 1.1 mm, with 5.0 whorls. ROTH: EARLY TERTIARY LAND MOLLUSKS 243 Gastrocopta montana is distinguished from Gastrocopta (Albinula) species a, next described, of the Milligan Creek Formation by its smaller size, more cylindrical shape, and low callus ridges thickening the inner and outer lips internally. The base of the body whorl of Gastrocopta (Al- binuld) species a is more strongly compressed, almost forming a keel around the umbilicus. Gastrocopta sagittaria of the Climbing Arrow Formation is more conic in shape, having a broad rather than narrowed anterior end. Material from the Dunbar Creek Formation (USGS 20012, 20013, 20017) is provisionally referred to the species. Most of the shells are slightly more cylindrical than the type-lot of G. montana, but one of two specimens from USGS 20012 is as broadly ovate as the types. The spec- imens from USGS 20013 are internal molds of tuffaceous siltstone with very little shell remain- ing. They show the impressions of upper and lower palatal barriers a short distance behind the position of the crest. The lower barrier is larger and more deeply immersed than the upper. The better-preserved specimen from USGS 20012 consists of original shell, partly filled with recal- citrant siltstone matrix that obscures most of the apertural dentition, but a strong, unbranched an- gular lamella is present, projecting almost as far as the plane of the aperture. The peristome is everted; the outer lip is sinuous, most produced medially. The Pliocene to Holocene G. contracta (Say, 1822) (Fig. 4), type-species of the subgenus Al- binula, is the modern species most similar to G. montana, in its sinuous, unbranched angular la- mella, strong crest, and callus ridges inside the inner and outer lips. Gastrocopta contracta is conic rather than cylindrical, with the penulti- mate whorl substantially broader than the an- tepenult; however, the basal configuration of the last whorl is quite similar. Gastrocopta holzingeri (Sterki, 1889), Pliocene to Holocene, is similar in shape to G. montana but has a forked, lamb- da-shaped angulo-parietal lamella. Gastrocopta (Gastrocopta) cristata (Pilsbry and Vanatta, 1 900) is similar in shape and also has a distinct crest, but its angular lamella is smaller and does not turn to the right within; there is no toothed callus ridge paralleling the inner lip. Along with English Eocene species (Preece 1982), this and the following species probably constitute the oldest known occurrence of the subgenus Albinula, but as noted in the introduc- tion the exact age of the Milligan Creek For- mation is not well established. Gastrocopta (Albinula) species a (Figure 5) Pupillidae A, D. W. Taylor in Robinson 1963:68, table 4.- Taylor 1975:206. Pupillidae B, D. W. Taylor in Robinson 1963:68, table 4 (in part).-Taylor 1975:206 (non p. 209). DESCRIPTION.— Shell dextral, umbilicate, ovate to elongate-ovate, of 5.2-5.5 whorls, suture mod- erately to deeply impressed, spire profile convex, fourth and fifth whorls about equally broad. Sculpture of fine, oblique growth lines and a trace of puckering anterior to the suture. Body whorl compressed anteriorly, sulcate behind aperture, very much narrowed toward base; last 0.3 whorl nearly straight in basal view, cross section tri- angular, the anterior vertex slightly pinched off on either side. Umbilicus wide, excavated. Faint crest behind aperture. Aperture ovate-triangular; peristome everted, thin, continuous, appressed to body whorl. With prominent, strongly sin- uous, angular lamella, projecting as far as plane of aperture, inner end thickened and deflected toward periphery. Columellar lamella horizon- tal. Peglike upper and lower palatal folds some- times present inside aperture. REFERRED MATERIAL. — Milligan Creek Formation: USGS 20007, five specimens. REMARKS.— The specimens at hand, while not good enough for formal taxonomic description, demonstrate that a second species of Gastrocopta (Albinula), distinct from G. montana, occurs in the Milligan Creek Formation. The material con- sists of one adult shell, height 2.75 mm, diameter 1.6 mm, with 5.5 whorls (outer lip broken in manipulation); two other adults 2.3 mm in height and 1.4 mm in diameter with 5.2 whorls; an intact juvenile shell of four whorls, height 1.4 mm; and a fragment of spire (4+ whorls) about 2 mm in height. Original shell is present in all, showing fine surface incremental lines. The pal- atal folds present in the largest specimen are not borne on a palatal callus but arise separately within the aperture. No palatal folds are detect- able on the two smaller adult shells. Otherwise the dentition is basically similar to the modern G. contracta (Say). The first bend in the angular lamella points toward the middle of the outer lip 244 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 at about the level of the upper palatal fold. Vari- ation of 0.4-0.5 mm in height and 0.25-0.5 whorl in adult shells is not uncommon in modern species of Gastrocopta (Pilsbry 1948). Gastrocopta montana of the Milligan Creek Formation is smaller, more cylindrical, with cal- lus ridges thickening the inner and outer lips. Gastrocopta sagittaria of the Climbing Arrow Formation is also smaller and more conic in shape, having a broad rather than narrowed an- terior end. This species and G. montana both seem more closely related to G. contracta than to any other modern species; G. contracta may be the surviv- ing member of a once more diversified lineage. Gastrocopta ( Albinula) sagittaria new species (Figure 6) Pupillidae E, D. W. Taylor in Robinson 1963:68, table 4.— Taylor 1975:208. DIAGNOSIS.— A small, ovate-conic Gastrocop- ta with prominent, sinuous, unbranched angular lamella, broadly rounded, perforate base, and low, untoothed callus ridges inside inner and outer lips. DESCRIPTION.— Shell small for the genus, dex- tral, umbilicate, ovate-conic, of 4.5-4.8 whorls; suture moderately impressed, spire profile con- vex, body whorl not much broader than penult. Sculpture of oblique growth lines. Body whorl either narrowly shouldered or compressed below suture, well rounded toward base; last 0.5 whorl nearly straight in basal view, laterally com- pressed and narrowed anteriorly into a triangular cross section. Moderately strong crest behind ap- erture. Aperture subquadrate; peristome everted, solid, continuous, appressed to body whorl; with low callus barriers inside inner and outer lips. With prominent, sinuous, angular lamella, pro- jecting as far as plane of aperture, its inner end broadening to a flange extending toward periph- ery. Dimensions of holotype: height 1.9 mm, di- ameter 1.3 mm, whorls 4.8. TYPE-MATERIAL. -Holotype: USNM 377376, from U.S. Geological Survey Cenozoic Locality 20009, Montana: Jeffer- son County: NE'A NE'/4 NE'/< sec. 11, T 2 N, R 1 W, Three Forks Quadrangle (1950) 1:62,500; altitude 4,580 ft (1,400 m), Climbing Arrow Formation, Eocene. Fourteen paratypes, USNM 377377, from same locality as holotype. REMARKS.— The material consists of internal molds and thoroughly recrystallized shells, of pink to colorless calcite. Even in those specimens that preserve the shape of the aperture and the prom- inent angular lamella, calcite fills the aperture to an extent that conceals the presence or absence of other lamellae or deep-seated folds. When specimens are immersed in toluene (refractive index 1.49693), it can be seen that the angular lamella first curves toward the outer lip, then recurves toward the axis. Its inner end broadens and bears a rounded flange extending toward the periphery, much as in G. contracta (Say). The three most complete paratypes measure Locality Height Diameter No. of whorls USGS 20009 1.9 mm 1.3 mm 4.7 1.8 1.3 4.6 1.8 1.3 4.5 Measurements include the expanded portion of the peristome. Several younger species of Albinula have a similar, ovate-conic shape, including Gastrocop- ta (Albinula) dupuyi (Michaud, 1855), Pliocene of France, and the Pleistocene to Recent North American G. contracta. The broad, perforate base and the presence of fewer than five whorls dis- tinguish this from any other species of Gastro- copta in the Bozeman Group. The species is named for the Climbing Arrow (Latin, sagittd) Formation. Subgenus indeterminate Gastrocopta cordillerae new species (Figure 7) Pupillidae B, D. W. Taylor in Robinson 1963:68, table 4 (in part). -Taylor 1975:209 (non p. 206). Pupillidae F, D. W. Taylor in Robinson 1963:68, table 4 (in part).-Taylor 1975:209 (in part). Pupillidae I, D. W. Taylor in Robinson 1963:68, table 4 (in part).-Taylor 1975:209 (in part). Pupillidae indet., D. W. Taylor in Robinson 1963:68, table 4. -Taylor 1975:209. DIAGNOSIS.— An ovate-oblong Gastrocopta with short spire whorls but moderately tall body whorl, flattened or weakly sulcate behind aper- ture, narrowed base, small aperture, and strong to moderate crest. DESCRIPTION. — Shell dextral, umbilicate, ovate-oblong with obtusely conic summit, of about 5.0-5.5 whorls; suture moderately im- pressed, whorls of spire short, convex. Nuclear whorls 1 .4, smooth. First neanic whorl smooth; thereafter with sculpture of fine, markedly oblique growth lines, strongest just anterior to suture. Body whorl moderately tall, compressed, nar- ROTH: EARLY TERTIARY LAND MOLLUSKS 245 rowed toward base, flattened or weakly sulcate behind aperture, with a strong to moderate crest. Aperture small, rounded-triangular. Upper and lower palatal barriers present, discrete; rest of apertural dentition not known. Dimensions of holotype: height 1.8 mm, diameter 1.1 mm, whorls 5.1. TYPE-MATERIAL.— Holotype: USNM 377378, from U.S. Geological Survey Cenozoic Locality 200 1 5, Montana: Broad- water County: NW'/4 SW/4 sec. 3, T 2 N, R 1 E, Three Forks Quadrangle (1950) 1:62,500; altitude 4,235 ft (1,290 m). Dun- bar Creek Formation, Eocene or Oligocene. Thirteen para- types, USNM 377379, from same locality as holotype. REFERRED MATERIAL.— Dunbar Creek Formation: USGS 20012, one specimen, juvenile. USGS 20013, three speci- mens. USGS 20016, one specimen, a calcitic internal mold with little shell remaining. USGS 20017, one specimen. REMARKS.— The holotype and paratypes con- sist of moderately well preserved original shells filled with grayish yellow tuffaceous siltstone. No specimen in either the type-lot or the referred material shows fully the characteristics of the aperture, and for this reason it is not possible to assign the species to a subgenus. However, it is a characteristic species of the Dunbar Creek For- mation, occurring at five out of the seven local- ities that yielded land snails, and is readily rec- ognized by its ovate-oblong profile and the contrast between the short whorls of the spire and the relatively tall body whorl. On the juvenile specimen from USGS 20012, axial ribs extend across the flat base but are much weaker than those above the peripheral angula- tion. The largest paratype is 2.4 mm long and 1.3 mm in diameter without having the complete aperture present. A referred specimen from USGS 20013 is 2.2 mm long, 1.3 mm in diameter, and ovate in outline, the base distinctly compressed; it shows a narrow angular lamella. Pupoides Pfeiffer, 1854 TYPE-SPECIES: Bulimus nitidulus Pfeiffer, 1839, by subsequent designation (Kobelt 1880 [1876-81]). DIAGNOSIS.— Shell "about 3 to 6 mm long, ri- mate; long-ovate, turrited or rarely cylindric, with obtuse apex and few (generally 5-6) rather long whorls. Aperture ovate, toothless except for a small, tuberculiform angular lamella close to the insertion of the outer lip, or united with it, some- times wanting; peristome expanded, reflected and usually thickened within. Internal axis slender, perforate" (Pilsbry 1948:920). REMARKS.— Pupoides is mainly a tropical and subtropical genus, distributed on all continents except Europe; it is also absent from Southeast Asia and the East Indies. Pilsbry (1920-21) as- sociated it with arid regions and relatively dry stations in humid areas. Pilsbry (1922-26:249, 265) placed Pupa in- colata White, 1876, from the Eocene of south- western Wyoming, in Pupoides, but also included it in a list of Pupillidae of uncertain affinities. The figures by White (1883, pi. 29, fig. 15-17) show a conical shell with an externally thickened outer lip, doubtfully pupillid in my opinion. (Ischnopupoides) Pilsbry, 1926 TYPE-SPECIES: Pupa hordacea Gabb, 1866, by original desig- nation. DIAGNOSIS.— "Shell cylindric or subcylindric; diameter decidedly less than half the length" (Pilsbry 1948:921). REMARKS.— Ischnopupoides is a New World group with two closely related species living north of Mexico. Pupoides (Ischnopupoides) hordaceus (Gabb, 1866) occurs today in Wyoming, Colo- rado, Kansas, Utah, Arizona, and New Mexico (Pilsbry 1948; Bequaert and Miller 1973). It is recorded from late Pleistocene or early Holocene deposits in Kansas, Texas, New Mexico, and Ar- izona (Bequaert and Miller 1973). Pupoides (Ischnopupoides) inornatusVanatia, 1915, ranges from South Dakota to New Mexico, although as pointed out by Taylor (1960), it is documented to be living at only two localities; it also occurs in Blancan (late Pliocene and early Pleistocene) faunas in Nebraska, Kansas, and Texas. Other species referred to the subgenus occur in Mexico, Ecuador, Peru, Bolivia, Argentina, and Chile (Bequaert and Miller 1973). There are no pre- vious records of Ischnopupoides in Montana. Taylor (1975:430) reported an unnamed Pu- poides (Ischnopupoides) species from the Tepee Trail (=Wagon Bed) Formation, upper Eocene of Wyoming, as follows: "one well-preserved in- ternal mold shows the features of shape, size, apertural thickening, reflected peristome and lack of apertural lamellae. In diameter of shell and general proportions of aperture, the specimen agrees well with P. (/.) inornatus Vanatta, but its whorls are lower and hence total shell length is less in the fossil." In the following new species, P. tephrodes, di- ameter is approximately half the length (or height) 246 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 FIGURES 8-12: Figure 8. Pupoides (Ischnopupoides) tephrodes, new species, holotype, USNM 377380; height 3.0 mm. Figure 9. Pupoides (Ischnopupoides) sp., cf. P. (I.) hordaceus (Gabb), USNM 377382; height 3.0 mm. Figure 10. Pupoides (Ischno- pupoides) hordaceus (Gabb), CAS 046507; height 3.4 mm; Holocene, drift of Chaco River at Pueblo Bonito, Chaco Canyon National Monument, San Juan County, New Mexico. Figure 1 1 . Pupoides tephrodes detail of apex of holotype showing transition from smooth nuclear to ribbed neanic whorls. Figure 12. Pupoides hordaceus, detail of apex of CAS 046507. of the shell. This leaves little but the cylindric form, absence of apertural tubercles, and (in some species) the presence of fine axial riblets to dis- tinguish the subgenus. Miller (in Bequaert and Miller 1973) found that the anatomy of P. hor- daceus differed only in minor details from that of Pupoides (Pupoides) albilabris (C. B. Adams, 1841). Nevertheless, the presence in the Boze- man Group of P. tephrodes and another species hardly distinguishable from P. hordaceus points to the existence in the eastern Cordillera of a lineage distinct from Pupoides, sensu stricto, since at least Oligocene time. Pupoides (Ischnopupoides) tephrodes new species (Figures 8, 11) Pupillidae C, D. W. Taylor in Robinson 1963:68, table 4 (in part).— Taylor 1975:209 (in part) (non p. 206). Pupillidae F, D. W. Taylor in Robinson 1963:68, table 4 (in part).— Taylor 1975:209 (in part). ROTH: EARLY TERTIARY LAND MOLLUSKS 247 DIAGNOSIS. — A small, cylindric Pupoides (Ischnopupoides) with diameter equal to about half of height, sculpture of slender, well-spaced, retractive riblets, and peristome narrowly everted. DESCRIPTION.— Shell dextral, narrowly umbil- icate, cylindric with convexly low-conic summit, of about 5.2 tall whorls; suture moderately im- pressed, crenulated by axial riblets. Early whorls convex; fourth and fifth whorls roundly shoul- dered below suture, compressed at periphery and anteriorly. Nuclear whorls 1.5, smooth; neanic sculpture of slender, well-spaced, retractive rib- lets. Body whorl narrowly, slopingly shouldered, compressed at periphery, slightly attenuated to- ward base, rising gently on penult behind aper- ture. Aperture oblique, ovate; peristome narrow- ly everted, not thickened within; parietal margin oblique, covered by a thin callus. No angular tubercle present. Dimensions of holotype: height 3.0 mm, diameter 1.5 mm, whorls 5.3. TYPE-MATERIAL.— Holotype: USNM 377380, from U.S. Geological Survey Cenozoic Locality 20015, Montana: Broad- water County: NW'/4 SW/4 sec. 3, T 2 N, R 1 E, Three Forks Quadrangle (1950) 1:62,500; altitude 4,235 ft (1,290 m). Dun- bar Creek Formation, Eocene or Oligocene. Two paratypes, USNM 377381, from same locality as holotype. REFERRED MATERIAL.— Dunbar Creek Formation: USGS 20013, three specimens, internal molds. USGS 20016, three specimens, internal molds with some shell preserved, showing fine oblique riblets. REMARKS.— The type-lot consists of moder- ately well preserved original shells filled with grayish yellow tuffaceous siltstone. The axial rib- lets are worn on all specimens so that in places they are visible only near the suture; similar wear occurs in Recent species of Pupoides (Ischno- pupoides). One paratype measures: height 2.8 mm, diameter 1.3 mm, with 5.2 whorls. The other paratype measures: height 2.7 mm, di- ameter 1 .2 mm, with 5. 1 whorls. Diameter/height ratios for intact material range from 0.44-0.50, compared to a range of 0.40-0.45 for P. hor- daceus and 0.44-0.46 for P. inornatus (calculated from dimensions given by Pilsbry [1948]). The shells are smaller than in either P. hor- daceus or P. inornatus, both of which may exceed 3.5 mm in length. The peristome is less sharply turned out. The rather tall, loosely coiled whorls (parietal wall encroaching little upon the aper- ture) and the cylindrical outline are wholly typ- ical of the subgenus Ischnopupoides. The name tephrodes combines the Greek teph- ra, ash, with the suffix -odes, denoting fullness. Pupoides (Ischnopupoides) sp., cf. P. (/.) hordaceus (Gabb, 1866) (Figure 9) Pupillidae C, D. W. Taylor in Robinson 1963:68, table 4 (in part).-Taylor 1975:209 (in part) (non p. 206). DESCRIPTION.— Shell dextral, narrowly umbil- icate, cylindric with convexly conic summit, of 5.1-5.5 tall whorls; suture not deeply impressed. Early whorls convex; fourth and fifth whorls roundly shouldered below suture, compressed at periphery and anteriorly. Nuclear whorls 1.4, smooth; neanic sculpture of faint, slender, irreg- ularly spaced, retractive riblets. Body whorl elon- gate, smoothly rounding toward base, rising slightly on penult behind aperture. Aperture ovate; parietal margin sinuous, not strongly en- croaching on aperture; peristome broken on all specimens at hand. REFERRED MATERIAL.— Dunbar Creek Formation: USGS 200 1 2, five specimens. REMARKS.— The material consists of recrys- tallized shells with filling of cream-colored tuff- aceous siltstone. Although the peristome is not preserved in any of the specimens at hand, the shape of the shell is almost identical to Holocene P. hordaceus (Fig. 10, 12). The surface of the fossils is somewhat worn, so the original strength of the axial ribbing cannot be evaluated. The strongest riblets seem to have been more irreg- ularly spaced than those of P. hordaceus. Better- preserved material would also show whether the ribbing was stronger than that of P. inornatus Vanatta. The four most nearly intact specimens mea- sure Locality USGS 200 12 * Broken. Height 3.2 mm 3.1 3.0 3.0 Diameter No. of whorls 1.2 mm* 1.3 1.3 1.4 5.5 5.4 5.5 5.1 These specimens differ from P. tephrodes in being slimmer and less distinctly ribbed, and in having a more steeply conical summit. Order SIGMURETHRA Family OREOHELICIDAE Radiocentrum Pilsbry, 1905 TYPE-SPECIES: Oreohelix chiricahuana Pilsbry, 1905, by orig- inal designation. 248 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 FIGURES 13-15. Radiocentrum hendersoni (Russell), holotype, USNM 497659; diameter 7.8 mm. DIAGNOSIS.— Oreohelicids with "embryonic shell of about 1 1/2 radially ribbed whorls; penis club-shaped, the walls of its cavity plain in the anterior part, having oblique ridges irregularly en chevron in the middle and posterior parts, wide and truncate at the end, epiphallus about as long as the penis, slender anteriorly, the penial retractor inserted on it a short distance from its entrance in the penis. Reproduction oviparous" (Pilsbry 1939:540). Babrakzai et al. (1975) found a large number of submetacentric chromosomes and a haploid chromosome number of 32 to be characteristic of Radiocentrum. The few-whorled, radially costulate proto- conch is the only paleontologically useful diag- nostic character, but it is adequate for recogni- tion of this group. The shells are otherwise much like Oreohelix, depressed-helicoid to lenticular in shape, the periphery ranging from rounded, through obtusely subangular, to distinctly cari- nate. REMARKS.— The endemic North American pulmonate family Oreohelicidae consists of two genera: Oreohelix Pilsbry, 1905, and Radiocen- trum, the latter originally proposed as a subgenus of the former. Conchologically, Radiocentrum has a distinc- tive, radially ribbed 1 .5-whorled embryonic shell. Living Radiocentrum deposit eggs, whereas Re- cent Oreohelix are ovo viviparous. On reproduc- tive characters and chromosome data, Babrakzai et al. (1975) elevated Radiocentrum to generic status. The modern distribution of Radiocentrum in- cludes southern New Mexico, southern Arizona, trans-Pecos Texas (late Pleistocene; Metcalf and Johnson 1 97 1), northern Coahuila (probable late Pleistocene; Metcalf 1980), Chihuahua, Baja California Sur (Miller 1973a; Christensen and Miller 1976), and Santa Catalina Island off southern California (Hochberg et al., in press). Tozer (1956) found Oreohelix angulifera (Whiteaves, 1885) from the St. Mary River and Edmonton formations, upper Cretaceous of western Alberta, and Oreohelix thurstoni (Rus- sell, 1926) from the Paskapoo, Porcupine Hills, and Willow Creek formations, Paleocene of west- ern Alberta, to have regular costae on the em- bryonic whorls, strongly suggestive of Radiocen- trum. Additional fossil oreohelicids probably assignable to Radiocentrum include R. grangeri (Cockerell and Henderson, 1912) from the Eocene of Park County, Wyoming, and R. hendersoni (Russell, 1938) from the Oligocene of Colorado. The latter was originally described in the genus Gonyodiscus Fitzinger, 1833 (Discidae), but is here reassigned to Radiocentrum based upon ex- amination of photographs of the holotype (USNM 497659) (Fig. 13-15), supplied by D. W. Taylor. Sculpture of the protoconch is not preserved, but the whorl diameter increases suddenly after 1.5 whorls, as it does at the beginning of neanic growth in many Radiocentrum. Helix nacimientensis White, 1886, from the Paleocene of New Mexico, assigned to Radiocentrum by Cockerell (1914), is probably not an oreohelicid snail and may belong to the Helminthoglyptidae (Taylor 1 975). Cretaceous and Tertiary species of Radiocen- trum are all from north of the present range of the genus, along the eastern Cordillera. The genus has undergone a southward restriction or dis- placement of range since Paleogene time; the Pleistocene and Holocene range includes dis- junctions probably related to the late Cenozoic ROTH: EARLY TERTIARY LAND MOLLUSKS 249 FIGURES 16-23: Figures 16-20. Radiocentrum taylori, new species. Figures 16-18, holotype, USNM 377383; diameter 12.4 mm. Figure 19, paratype, USNM 377384; diameter 8.1 mm. Figure 20, referred specimen, USNM 377386; detail of apical sculpture, x 35. Figures 2 1-23. Radiocentrum chiricahuanum obsoletum (Pilsbry and Ferriss), CAS 046508; diameter 12.6 mm; Holocene, Whitetail Canyon, Chiricahua Mountains, Cochise County, Arizona. All specimens coated for photographing. emergence of the Sonoran and Chihuahuan des- erts as arid environments of regional extent (Hochberg et al., in press). The type localities of the two new species of Radiocentrum described here are approximately 1,600 km north of the northernmost Holocene occurrences of the genus. The present distribu- tion of Radiocentrum consists of scattered, high- ly local enclaves. Taken together, the fossil and Recent range data suggest restriction from a for- merly more widespread and continuous range that included the northeastern Cordillera. Ra- diocentrum has been distinct from Oreohelix since at least the late Cretaceous; species of Oreohelix are known from upper Cretaceous, Paleocene, and Eocene strata from Alberta to Utah (Pilsbry 1939; Tozer 1956; La Rocque 1960). No other generic groups are recognized in the family. In contrast to the Helminthoglyptidae of the arid Southwest, which have responded to fragmen- tation of range and isolation by a dramatic, ev- idently saltational, generic diversification (Miller 1973&, 198 la), the Oreohelicidae have been ev- olutionarily conservative. Radiocentrum taylori new species (Figures 16-20) Oreohelix n. sp., D. W. Taylor in Robinson 1963:68, table 4 (in part). -Taylor 1975:209. DIAGNOSIS.— A small, solid, low-trochoid Ra- diocentrum of about 5.25 whorls; periphery an- gulate to obtusely keeled; protoconch strongly radially ribbed, ribs overhanging abaperturally; neanic sculpture of irregular ribs, lightly decus- sated on base by incised spirals. DESCRIPTION. — Shell low-trochoid, solid, height about 0.65 times diameter, apical angle 124°. Whorls about 5.25; spire profile weakly convex; suture not deeply impressed. Body whorl 250 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 slightly shouldered, about equally convex above and below periphery, not markedly descending except immediately behind aperture. Periphery angulate to obtusely keeled (particularly on last 0.5 turn). Aperture circular, at about 40° angle to axis of coiling; limbs of peristome approaching closely, parietal wall with thin wash of callus; peristome not expanded but inner lip reflected toward umbilicus. Base well rounded, umbilicus contained about six times in diameter of shell. Protoconch consisting of 1.5 whorls, nuclear tip smooth, thereafter with prominent, elevated ra- dial ribs, slightly convex-forward, overhanging on abapertural side, and almost as wide at the interspaces between. End of protoconch some- times slightly thickened; onset of neanic growth marked by abrupt increase in whorl diameter and/or change in obliquity of radial ribs. Ribs of neanic shell at first upstanding and bladelike, quickly becoming lower, more solid, and less reg- ularly spaced. Ribbing on body whorl crude, slightly sinuous over periphery, and lightly de- cussated by fine spiral striae particularly evident on base and on shoulder of last 0.25 whorl. Di- mensions of holotype: diameter 12.4 mm, height 7.9 mm, whorls 5.3. TYPE-MATERIAL.— Holotype: USNM 377383, from U.S. Geological Survey Cenozoic Locality 200 1 5, Montana: Broad- water County: NWA SWA sec. 3, T 2 N, R 1 E, Three Forks Quadrangle (1950) 1:62,500; altitude 4,235 feet (1,290 m). Dunbar Creek Formation, Eocene or Oligocene. Figured para- type, USNM 377384; 39 additional paratypes, USNM 377385, from same locality as holotype. REFERRED MATERIAL.— Dunbar Creek Formation: USGS 20013, nine specimens— three external molds of spires (the largest 12.0 mm in diameter, with 5.0 whorls) with clear impressions of surface sculpture, one also having an internal mold counterpart; one internal mold with traces of exterior sculpture, diameter 12.1 mm, height 8.2 mm, with 5.5 whorls; four specimens with considerable shell remaining and very good preservation of protoconch sculpture (Fig. 20), diameter 4.6-9.7 mm, with 3.25-4 whorls. USGS 20014, two speci- mens—one a basal external mold 10.4 mm in diameter of an umbilicate shell with somewhat tumid base, low, forwardly concave, radial ribbing, and minor spiral rugosity; the other a shell 3.5 mm in diameter, in matrix, no sculpture preserved. USGS 20016, one external mold with internal mold counter- part, 9 mm in diameter, the matrix coarse but preserving radial riblets on the protoconch. USGS 20017, five specimens— one external and four internal molds, the largest 13.5 mm in di- ameter; three of the internal molds with some original shell remaining, showing angulate periphery, retractive axial rib- bing, and fine, incised spiral lines on the shoulder; one with strongly ribbed protoconch well preserved. REMARKS.— The type-lot consists of moder- ately well preserved to very well preserved orig- inal shells and internal molds with matrix of cream-colored to yellowish gray, limy, tufFaceous siltstone. The paratypes range from 2.2 mm in diameter with 2.2 whorls to 1 1 .4 mm in diameter with 5.3 whorls. Specimens of fewer than 4 whorls are acutely carinate, with the carina above the middle of the whorl, set off in some instances by faintly impressed grooves above and below. Some adult shells show minor spiral ribbing in addition to incised spiral striae on the last 0.25-0.5 whorl. This is the most ubiquitous species in the Dun- bar Creek Formation, occurring at five out of the seven localities that yielded land snails. Radiocentrum taylori resembles the Holocene R. chiricahuanum (Pilsbry, 1 905), another low- trochoid species with strong, irregular ribbing and an angulate to carinate periphery. The strongest resemblance is to the subspecies R. c. obsoletum (Pilsbry and Ferriss, 1910) (Fig. 2 1-23), in which the incised spiral sculpture is weak, the ribbing coarse and blunt, and the peripheral keel not especially prominent. The periphery of/?, taylori is less carinate, most often acutely angular in subadult to adult shells, less commonly with spi- ral grooves setting off a peripheral keel. Radio- centrum c. obsoletum attains larger size, up to 1 5 mm diameter; R. c. chiricahuanum, at about 1 1 mm, is similar in size to R. taylori. Radiocentrum taylori differs from R. laevido- mus, next described, from the Milligan Creek Formation, in its coarser sculpture and higher- spired, trochoid shape. Radiocentrum laevido- mus lacks the spiral sculpture of R. taylori; its protoconch has fine, simple ribs instead of the heavy, overhanging ribs of R. taylori. Radiocentrum (?) anguliferum (Whiteaves, 1885) from the upper Cretaceous of Alberta dif- fers in its very low spire and nearly involute mode of coiling. Radiocentrum thurstoni (Russell, 1926) from the Paleocene of Alberta has a subangular to rounded periphery and narrow umbilicus. It is also larger than any R. taylori specimens yet seen. Radiocentrum hendersoni (Russell, 1938) from the Oligocene of Colorado is similar to R. taylori in its low-trochoid shape, prominent radial rib- bing, and shouldered whorls. At 7.8 mm in di- ameter, with 4.5 whorls, the holotype of/?, hen- dersoni (Fig. 13-15) may be immature. However, the last whorl descends below the peripheral an- gle, the shouldering of the whorl is intensified, and the inner lip expands toward the umbilicus, ROTH: EARLY TERTIARY LAND MOLLUSKS 251 FIGURES 24-34: Figures 24-26, 30. Radiocentrum laevidomus, new species. Figures 24-26, holotype, USNM 377387; di- ameter 10.1 mm. Figure 30. Paratype (hatchling young), USNM 377388; diameter 1.6 mm. Figures 27-29. Radiocentrum hachetanum (Pilsbry), CAS 046509; diameter 14.2 mm; Holocene, summit of Hacheta Grande Mountain, Hidalgo County, New Mexico. Figures 31, 32. Polygyrella sp., cf. P. polygyrella (Bland and Cooper), USNM 377391; diameter 10.1 mm. Figures 33, 34. Polygyrella polygyrella (Bland and Cooper), CAS 046510; diameter 10.1 mm; Holocene, 2.4 km south of Selway Falls, Idaho County, Idaho. All specimens coated for photographing. as in mature specimens of other oreohelicids. A 4.5-whorled paratype of/?, taylori (Fig. 1 9) is 8.2 mm in diameter, lacks the incised spirals that decussate the ribs in R. hendersoni, and has a much less profoundly impressed suture. I take pleasure in naming this species for Dwight W. Taylor, expert on freshwater Mol- lusca of western North America and author of the first reports on its occurrence. Radiocentrum laevidomus new species (Figures 24-26, 30) Oreohelix n. sp., D. W. Taylor in Robinson 1963:68, table 4 (in part). -Taylor 1975:206. DIAGNOSIS.— A small, lenticular Radiocen- trum of about 5 whorls; periphery angulate above middle of whorl: orotoconch sculpture of fine. low-standing, smooth radial riblets separated by wider interspaces; neanic sculpture of low, irreg- ular radial striae; no spiral sculpture present. DESCRIPTION.— Shell lenticular, apical angle 133°. Whorls about 5, enlarging rapidly; whorls of spire moderately convex, suture impressed. Body whorl expanding at about same rate as spire whorls, not descending behind aperture; periph- ery angulate above middle of whorl, becoming gently rounded on last 0.25 turn of 5-whorled specimen. Base rather deep, moderately inflated; umbilicus conical, its diameter contained three to four times in diameter of shell, with a hint of carination at its rim. Aperture simple, without thickening. Protoconch of 1.75 whorls, nuclear tip smooth, followed by smooth, regular, for- wardly convex radial ribs separated by inter- spaces of greater width, increasing in strength 252 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 and distance to end of protoconch; ribs extending from suture to shoulder, weak to obsolete over periphery, then reappearing as a series of fine riblets spiraling inward around umbilicus of ju- venile shells. Neanic whorls sculptured with low, irregular, forwardly convex radial striae. Striae strongest below suture, sometimes interrupted or deflected backward at peripheral angulation. Di- mensions of holotype: diameter 10. 1 mm, height 5.6 mm, whorls 4.6; of largest well-preserved paratype (USNM 377389): diameter 10.3 mm, height 6.5 mm, whorls 5.3. TYPE-MATERIAL. -Holotype: USNM 377387, from U.S. Geological Survey Cenozoic Locality 20007, Montana: Gal- latin County: NW'/« SWA sec. 36, T 1 N, R 1 W, Three Forks Quadrangle (1950) 1:62,500; altitude 4,260 ft (1,300 m). Mil- ligan Creek Formation, Eocene (?). Measured paratype, USNM 377389; figured paratype, USNM 377388; and 24 additional paratypes, USNM 377390, all from same locality as holotype. REMARKS.— The type-lot consists of well-pre- served original shells with matrix of solid to fri- able tan to cream-colored limestone. Color pat- tern is preserved on some specimens. The paratypes range from 1.5-5+ whorls. Sixteen of these are little more than hatchling young of 1 .5- 2.0 whorls (Fig. 30). The presence of such small juveniles free in the matrix indicates that the species was probably oviparous like modern Ra- diocentrum. In Oreohelix, the embryos are re- tained in utero to a stage of 2 whorls or more. The spire of the holotype and of the largest well-preserved paratype are mottled with or- ange-brown blotches at intervals of 0.1-0.5 whorl. The blotches extend either inward or out- ward from the suture, in a few cases crossing an entire whorl, but do not pass from one whorl to the next. They are probably remnants of an orig- inal color pattern. Many specimens of the Ho- locene species R. hachetanum (Pilsbry, 1915) and R. chiricahuanum (Pilsbry, 1905) show similar mottling. The species most similar to R. laevidomus in general shape and sculpture is the Holocene R. hachetanum (Fig. 27-29). The fine spiral lines in the intervals between protoconch ribs on R. hachetanum are not visible on R. laevidomus. This species differs from R. taylori of the Dun- bar Creek Formation in its finer, more delicate sculpture on both protoconch and neanic whorls, its lower spire, and lenticular rather than tro- choid shape. The radial ribs on the protoconch ofR. taylori are elevated, prominent, somewhat overhanging on the abapertural side, and nearly as wide as their interspaces. Those of R. laevi- domus are low-standing, evenly rounded, and distinctly narrower than their interspaces. Ribs on the very young paratypes, which could not have been subjected to much wear, have the same character (Fig. 30). Radiocentrum laevidomus shows no spiral sculpture like that present on R. taylori. Approximately the same characters dif- ferentiate R. laevidomus from R. chiricahua- num, which is very similar to R. taylori as noted above. Radiocentrum^} anguliferum (Whiteaves, 1885) from the upper Cretaceous of Alberta dif- fers from R. laevidomus in its strongly angular periphery and nearly involute mode of coiling. Radiocentrum hendersoni (Russell, 1938) from the Oligocene of Colorado differs in its higher- spired, trochoid shell, strong radial ribbing, and very deeply impressed suture. Radiocentrum thurstoni (Russell, 1926) from the Paleocene of Alberta is another trochoid species, with a higher, more conical spire. According to Tozer (1956) the embryonic whorls of Oreohelix obtusata (Whiteaves, 1885), upper Cretaceous of Alberta, are apparently smooth, which would rule out as- signment to Radiocentrum. The name proposed combines the Latin laevis, smooth, with domus, house, in reference to the relatively faint sculpture of the shell. Family AMMONITELLIDAE Polygyrella Binney, 1863 TYPE-SPECIES.— Helix polygyrella Bland and Cooper, 1861, by monotypy. DIAGNOSIS.— "The shell is widely umbilicate, discoidal with convex to nearly flat spire of nar- row, closely coiled costulate whorls; base smooth, translucent. Aperture lunate-triangular, the unexpanded peristome somewhat thickened within, the ends connected by an erect, triangular parietal tooth. Within the last whorl there are one or two radial series of three teeth each. Jaw with flat plaits and fine vertical striae. Soft anat- omy. . .about as in A mmonitella" (Pilsbry 1939: 555-556). REMARKS.— Polygyrella is represented in the Holocene by one species, Polygyrella polygyrella (Bland and Cooper, 1 86 1), with a range of north- ern Idaho, adjacent Montana, southeastern Washington, and northeastern Oregon. Polygy- rella from the John Day Formation (late Oli- ROTH: EARLY TERTIARY LAND MOLLUSKS 253 gocene or early Miocene) of central Oregon are referred to the same species (Hanna 1920). Taylor (1975) assigned specimens from the Eocene Kingsbury Conglomerate Member of the "Wasatch" Formation in the Powder River Ba- sin, Wyoming, to Polygyrella. He further sug- gested that Planorbis amplexus Meek and Hay- den, 1857 (upper Cretaceous, Judith River Formation, Montana), and Anchistoma parvu- lum Whiteaves, 1885 (upper Cretaceous, St. Mary River Formation, Alberta), are both species of Polygyrella. Based on its pattern of coiling, basal configuration, and parietal barrier, Polygyra ve- nerabilis Russell, 1937, from the upper Creta- ceous Belly River Formation, Alberta, seems to be another. Indeterminate species of Polygyrella are reported from upper Eocene strata in Glacier National Park, Montana (D. W. Taylor in Ross 1959), and (questionably) from an unnamed con- glomeratic sequence of presumed early Tertiary age on Little Granite Creek, Hoback Basin, northwestern Wyoming (Taylor 1975). Polygyrella sp., cf. P. polygyrella (Bland and Cooper, 1861) (Figures 3 1,32) Polygyrella, D. W. Taylor in Robinson 1963:68, table 4.— Taylor 1975:208. DESCRIPTION.— Shell subdiscoidal with broad umbilicus contained approximately three times in diameter. Spire flat to low-convex, suture lightly impressed. Whorls tightly coiled, some- times with closely spaced, forwardly convex ra- dial grooves extending outward from suture, not reaching shoulder of whorl. Body whorl with up to four shallow transverse constrictions. Last whorl not markedly descending except imme- diately behind aperture. Aperture oblique, peri- stome thickened and slightly expanded outward. REFERRED MATERIAL.— Climbing Arrow Formation: USGS 20008, 5 specimens. USGS 20009, 14 specimens. REMARKS.— The material from the Climbing Arrow Formation consists of internal molds (and some thoroughly recrystallized shells) shaped much like the modern Polygyrella polygyrella (Fig. 33, 34), with the subdiscoidal shape, flat to low-domed spire, tightly coiled whorls, and broad, circular umbilicus of that species. The fossils are composed in part of translucent, colorless to hon- ey-colored, coarsely crystalline calcite, either re- placing the shell or conforming to its interior, and in part of pinkish tan, finely crystalline cal- cite probably representing a limy mud that par- tially filled the shells upon burial. The peristome is moderately thickened in sev- eral specimens from locality USGS 20009; these are probably mature individuals. In them, the aperture slants at a 45° angle to the axis of coiling, the same as in P. polygyrella. Juvenile specimens of 4 whorls or less are planorboid, with much of the protoconch visible in the umbilicus. Two distinctive shell features of P. polygyrella are not detectable: an erect parietal tooth and a radial series of barriers inside the last whorl. Careful preparation around the aperture of two specimens with mature, expanded peristomes re- vealed no parietal tooth. If completely recrys- tallized, such a tooth might not be distinguish- able from other calcite filling the whorl; the same may be true for the series of barriers that would be expected about one-half whorl back of the aperture. One specimen from locality USGS 20009 shows radial grooves outboard of the su- ture on what is apparently the fourth whorl, cor- responding to interspaces between radial costae on the spire of P. polygyrella. The six largest specimens measure Locality Diameter No. of whorls USGS 20008 USGS 20009 10.2 mm 9.1 8.6 9.0 9.0 8.0 7.1 6.8 6.7 6.2* 6.2* 5.9 Asterisks denote specimens with mature ex- panded peristome. Polygyrella parvula (Whiteaves, 1885) differs from P. polygyrella and the present specimens in its conoidal spire, sharply descending body whorl, and single slanting internal barrier. As illustrated (Meek 1876, pi. 42, fig. 16a-16e; Shi- mer and Shrock 1944, pi. 213, fig. 14, 15), P. amplexa (Meek and Hayden, 1 857) is flat-spired or nearly so, with an umbilicus that is more con- ical than the steep-sided, pitlike umbilicus of P. sp., cf. P. polygyrella. No information is avail- able on the presence or configuration of internal barriers in P. amplexa, and the aperture is un- known. Family HELMINTHOGLYPTIDAE Helminthoglypta Ancey, 1887 TYPE-SPECIES.— Helix tudiculata Binney, 1 843, by original des- ignation. 254 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 FIGURES 35-40: Figures 35-38. Helminthoglypta bozemanensis, new species, holotype, USNM 377392; diameter 16.6 mm. Figure 38. oblique view showing color banding. Figures 39, 40. Helminthoglypta californiensis (Lea), CAS 04651 1; diameter 19.3 mm; Holocene, Pacific Beach, north of Monterey, Monterey County, California. All specimens except Figure 38 coated for photographing. DIAGNOSIS.— "Helices of moderate or large size, the shell globose or depressed with conic or low spire and open or covered umbilicus; periphery rounded at all stages of growth. Embryonic shell of IVz to !3/4 whorls; after the smooth tip and a few radial wrinkles it has sculpture of close, mi- croscopic, waved, radial wrinkles, over which there are papillae in forwardly descending trends (often indistinct or practically absent). Adult sculpture of simple growth lines or with spiral engraved lines, malleation, papillae or granula- tion also. A dark band revolves above the pe- riphery (sometimes absent). Peristome narrow, expanded outwardly, usually reflected at base, dilated at columellar insertion" (Pilsbry 1939: 63). REMARKS.— The genus Helminthoglypta today is distributed from southeastern Oregon, through California west of the Cascade Range and the crest of the Sierra Nevada, into northern Baja California, Mexico. About 55 species are rec- ognized. At the northern end of the range of the genus, Helminthoglypta mailliardi Pilsbry, 1 927, and H. hertleini Hanna and Smith, 1937, inhabit low elevations in the Klamath Mountains. At the southern end, H. tudiculata (Binney, 1 843) and H. traskii (Newcomb, 1861) extend along the coast to the vicinity of San Antonio del Mar, Colonet, and H. reederi Miller, 1981, occurs in the Sierra San Pedro Martir (Miller 198 \b). One species, Helminthoglypta alfi Taylor, 1954, oc- curs in the Barstow Formation, upper Miocene of the Mojave Desert, California; it is similar to Recent Mojave Desert species. An undescribed species occurs in strata of probable Pliocene age in the Tehachapi Mountains, California (Roth, unpublished data). The Eocene species Hel- minthoglypta obtusa Anderson and Hanna, 1925, from the Tejon Formation, and H. (?) stocki Han- na, 1924, from the Sespe Formation, California, are probably incorrectly assigned to Helminth- oglypta, but more material will have to be stud- ied before a better allocation can be made. A land snail tentatively identified as Helmin- thoglypta is present in the Wiggins Formation (Oligocene), Wind River Basin, Wyoming (Tay- lor 1975). ROTH: EARLY TERTIARY LAND MOLLUSKS 255 Helminthoglypta bozemanensis new species (Figures 35-38) Helminthoglyptidae n. gen., D. W. Taylor in Robinson 1963: 68, table 4. Hemitrochus'! Taylor 1975:208, 209. DIAGNOSIS.— A small, globose-conic minthoglypta of about five whorls; body whorl weakly constricted behind reflected outer lip; umbilicus narrow, obliquely entering; sculpture of blunt incremental rugae and granulose ver- miculation sometimes resolving into radial rows of granules; narrow dark peripheral band present, bordered above and below by wider light zones; less distinct dark zones on shoulder and just be- low suture. DESCRIPTION.— Shell globose-conic, wider than tall, apical angle 107°. Whorls about 5, convex, rapidly expanding; suture impressed. Body whorl tumid, strongly descending for last 0.25 turn and more strongly for last 0. 1 turn, weakly constrict- ed behind outer lip. Aperture subcircular; pari- etal wall shallowly sigmoid, with thin wash of callus; peristome reflected and narrowly expand- ed, smooth at edge, more strongly reflected and thickened in umbilical region. Base well-round- ed; umbilicus very narrow, scarcely perforate, obliquely entering behind inner lip. Protoconch smooth, probably consisting of about 1 .8 whorls. Neanic whorls sculptured with (a) moderately strong, blunt, oblique incremental rugae, most irregularly spaced but some rhythmically spaced at intervals of about 1.0 mm, strongest on whorl shoulder but continuing over base, and (b) low granulose vermiculation, strongest between ru- gae and generally trending parallel to them. Ver- miculation weaker on base and behind outer lip, where it generally appears as radial rows of blunt, axially elongate granules. Body whorl with nar- row (0.8 mm- wide) brown supraperipheral band, bordered above and below by slightly wider (1 .2 mm) zones lighter than ground color of shell, and less distinct dark zones on midshoulder and just below suture. Dimensions of holotype: diameter (exclusive of expanded lip) 16.6 mm, height 13.3 mm, whorls 4.8; dimensions of paratype: di- ameter (slightly distorted) 15.4 mm, height 13.9 mm, whorls 5.2. TYPE-MATERIAL.— Holotype: USNM 377392, from U.S. Geological Survey Cenozoic locality 200 1 1 , Montana: Broad- water County: SE'/4 NW/4 sec. 6, T 2 N, R 2 E, Three Forks Quadrangle (1950) 1:62,500; altitude 4,360 ft (1,330 m). Dun- bar Creek Formation, Eocene or Oligocene. Paratype: USNM 377393, from same locality as holotype. REFERRED MATERIAL.— Four additional specimens from the USGS collection are referable, with differing degrees of con- fidence, to Helminthoglypta and H. bozemanensis. Climbing Arrow Formation: USGS 20009, one specimen, a fragment of smooth apical whorls and partial spire, 5 mm in greatest di- ameter, probably assignable to Helminthoglypta. Dunbar Creek Formation: USGS 20015, two specimens, internal molds, (1) diameter 7.4 mm, height 5.6 mm, whorls 3.4, and (2) diameter 9.8 mm, height 8.9 mm, whorls 3.8; small amount of shell material remaining on latter shows spaced incremental rugae oriented as on holotype, surface detail not preserved; raised line on internal mold in position of supraperipheral band; = Helminthoglypta sp., cf. H. bozemanensis. USGS 20016, one specimen, internal mold, diameter 13.0 mm, height 10.5 mm, whorls 4.1, =Helminthoglypta sp., cf. H. bozemanensis. REMARKS.— The holotype (Fig. 35-38) is a very well preserved specimen with almost all the orig- inal shell remaining and clear indication of the former position of color bands. The paratype is an internal mold with little shell remaining, the spire somewhat collapsed into the body whorl. This species is assigned to Helminthoglypta because of the distinct supraperipheral band, shape, sculpture, and reflected peristome. There are particularly strong similarities to the Recent Helminthoglypta californiensis (Lea, 1838) (Fig. 39, 40) of the central California coast. The weak constriction of the body whorl immediately be- hind the evenly expanded outer lip, and the ex- tent to which the last quarter- whorl descends are identical in both species. The fine sculpture of H. californiensis consists of rows of axially elon- gated granules paralleling the incremental lines. In many specimens, the granules correspond rather loosely from one row to the next. In other specimens, and in other members of the "Hel- minthoglypta nickliniana series" (Pilsbry 1939), to which H. californiensis belongs, the granules may line up quite precisely in diagonal rows, producing a distinctive clothlike pattern. In H. bozemanensis, where the granulose vermicula- tion most clearly resolves into rows of discrete granules (mainly on the base and behind the out- er lip), the sculpture strongly resembles that of H. californiensis. Taylor (1975) ruled out assignment to Hel- minthoglypta because of the multiple color bands, not otherwise known in the genus, and question- ably referred these specimens to Hemitrochus Swainson, 1 840 (type-species, Hemitrochus hoe- mastomus Swainson, 1840 [=Helix varians Menke, 1829], by monotypy). Hemitrochus con- 256 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 sists of about 25 Recent species distributed in southern Florida and the Antilles (Pilsbry 1889- 90, 1893-95; Turner 1958). The dim bands at the shoulder and suture of Helminthoglypta bozemanensis are located where the margins of the dark shoulder zone occur in modern Helminthoglypta. Several instances ex- ist, in Helminthoglyptidae and other families, of vacant-centered bands as a polymorphism of banded shells (Roth 198 la, Monadenia; Roth and Began 1984, Liguus). The upper and lower shell bands in the helminthoglyptid genus Hum- boldtiana Ihering, 1 892, for which Pilsbry (1939) could not suggest a homology, may have origi- nated in this way, as the emphasized edges of two now- vanished broad zones. The central band in Humboldtiana is probably homologous with the supraperipheral band of most other Hel- minthoglyptidae. The upper and lower bands bracket the central band about as evenly as the dark zones top and bottom on a Helminthoglypta shell. The strengths of the upper and the lower band are strongly correlated, but apparently in- dependent of the strength of the central band, suggesting a different derivation. Because of the apparent ease of the transfor- mation from solid to vacant-centered band, I do not think that the banding pattern of H. boze- manensis precludes assignment to Helmintho- glypta. Other similarities to Hemitrochus do ex- ist, however. They include the smooth protoconch, obliquely entering umbilicus, and spaced incremental rugae. In no Recent species of Hemitrochus do the whorls enlarge as rapidly; in globose-conic species, the spire makes up more of the shell. The base is typically shorter, tending to be flattish rather than tumid, and the strongest color band is usually at or below the periphery. The extent to which the lip turns out at maturity varies with the species, but often there is addi- tional thickening inside the edge of the aperture, which is not present in Helminthoglypta boze- manensis. Taylor (1975) also reported shells that he as- signed to Hemitrochusl from the White River Formation (Oligocene), Beaver Divide area, cen- tral Wyoming. La Rocque (1960) noted the sim- ilarity of "Helix" riparia White, 1876, from the Flagstaff Formation, Paleocene and Eocene of Utah, to Hemitrochus but declined to make a firm generic assignment for the species. Of the other helminthoglyptid genera that might be compared, Leptarionta Fischer and Crosse, 1872, has a glossy or silky shell with growth lines hardly evident in relief. Xerarionta Pilsbry, 19 13, and Plesarionta Pilsbry, 1939, both have incised spiral sculpture that is not present in Helminthoglypta bozemanensis. Humbold- tiana has compound sculpture consisting of in- cremental rugae and blunt granulation (some- times, as in H. palmeri Clench and Rehder, 1 930, partly fusing into an irregular vermiculation). The possible homology of Humboldtiana 's three bands with the pattern of Helminthoglypta boze- manensis has already been mentioned. However, no Recent Humboldtiana has the globose-conic shape of//, bozemanensis; in Humboldtiana the protoconch is finely pustulose, its juncture with the teleoconch well marked; the outer lip is bare- ly reflected, and there is no constriction behind the aperture. Lysinoe Adams and Adams, 1855, has, in addition to growth rugae, regularly spaced, discrete papillae (corresponding to the bases of periostracal bristles) in diagonal series. Nothing similar is present in H. bozemanensis. Because of the supraperipheral band (a form of disruptive coloration), it is assumed that the periostracum of H. bozemanensis was transpar- ent, although vestigial banding occurs under an opaque periostracum in Monadenia (Roth 1981ft). FAUNAL COMPOSITION AND PALEOECOLOGY MILLIGAN CREEK FORMATION.— The faunule from the Milligan Creek Formation consists of two species of Gastrocopta (Albinuld) and one species of Radiocentrum. The only species that may have a stratigraphic record outside the for- mation is G. montana, which is provisionally recognized in the Dunbar Creek Formation. The two Gastrocopta species seem to be most closely related to Gastrocopta (Albinuld) con- tracta, which I regard as a plausible ecological analog. Gastrocopta contracta ranges over much of eastern North America, as far west as South Dakota, Oklahoma, and western Texas (Pilsbry 1948; Cheatum and Fullington 1973). It extends farther west in Mexico, reaching southern Sonora (Arroyo San Rafael, San Bernardo) and northern Sinaloa (San Bias) (Pilsbry 1953). In the United States its western limit coincides approximately with the 16-inch [41 -cm] normal annual isohyet (U.S. Department of Commerce 1968). Gastro- copta contracta does not occur in western Mon- tana but Albinula in the broad sense is present in the form of G. holzingeri. ROTH: EARLY TERTIARY LAND MOLLUSKS 257 The modern habitat of Gastrocopta contracta is on shaded slopes along watercourses, under dead wood, leafmold, and grass (Franzen and Leonard 1 947, referring to Kansas). It is recorded from dense vegetation in mixed mesophytic for- est in Kentucky (Branson and Batch 1970). No ecological notes on its Mexican occurrences are available. Other species of Albinula occur on wooded slopes, either near or away from streams, under dead wood, bark, stones, and in moist grass around seepages (Franzen and Leonard 1947; Taylor 1960). Radiocentrum does not now range north of Santa Catalina Island, California, southern Ar- izona and New Mexico, and trans-Pecos Texas. Its habitat is generally in mountainous terrain, among cliffs and rockslides, usually where vege- tation is sparse. It is not an inhabitant of forests. At least two species are reported to be restricted to limestone (Pilsbry 1939), but others occur on shale or among lava rockslides (Pilsbry 1939; Miller 1973a). On Santa Catalina Island, R. ava- lonense (Pilsbry, 1 905) occurs on steep slopes of talus where the country rock is granitic, around the roots of black sage (Salvia melliferd) shrubs (Hochberg et al., in press). Modern disjunctions within the range of Radiocentrum are closely linked to zones of extreme aridity— the Sonoran and Chihuahuan deserts (Hochberg et al., in press). The southern Arizona and New Mexico occurrences are in isolated areas receiving at least 4 1 cm annual precipitation, at least half of it in the summer months (U.S. Dept. Commerce 1968). Radiocentrum and G. contracta are not known to be sympatric anywhere at present; the two groups come closest in western Texas— Radio- centrum in El Paso County (Metcalf and Johnson 1971) and G. contracta in Culberson, Jeff Davis, and Presidio counties (Cheatum and Fullington 1973). However, the range of Radiocentrum is apparently contracting, with several peripheral occurrences known only from empty shells, and there may have been limited sympatry within the recent past in either the southwestern United States or northern Mexico. The G. contracta-like species in the Milligan Creek Formation imply somewhat more mesic conditions (possibly at the microhabitat level) than does Radiocentrum. Robinson ( 1 96 1 , 1 963) concluded that the limestone and other fine- grained rocks of the Milligan Creek Formation were deposited in a perennial lake lying in moun- tainous terrain. The habitat of the Gastrocopta species may have been in leafmold along wooded watercourses feeding the lake. Radiocentrum laevidomus may have lived in more sparsely veg- etated habitats nearby. Its source area was prob- ably not remote, however, as it is the numerically dominant species in the USGS sample. The in- dicated climate is warm-temperate to subtropi- cal. A minimum of 40 cm annual precipitation is suggested, much of it in the summer months. The land snails do not indicate a frost-free cli- mate; within the southern Arizona-New Mexico range of Radiocentrum there is a period of gen- erally less than 200 days between the last freeze of spring and the first freeze of autumn. Over much of the eastern range of Gastrocopta con- tracta the frost-free period is even shorter (U.S. Dept. Commerce 1968). CLIMBING ARROW FORMATION.— The faunule from the Climbing Arrow Formation consists of Gastrocopta (Albinula) sagittaria, Polygyrella sp., cf. P. polygyrella, and a fragment of apical whorls and incomplete spire probably assignable to Hel- minthoglypta. The Gastrocopta is not known to occur outside the formation. The present mate- rial does not allow a determination whether the HelminthoglyptaC?) is the same as H. bozema- nensis of the Dunbar Creek Formation. The Polygyrella species, although not adequately pre- served, does not seem to differ in any significant way from Polygyrella polygyrella, which ranges from late Oligocene or early Miocene (Hanna 1920)toHolocene. As already noted, Gastrocopta (Albinula) is ba- sically an eastern group at present, with one species entering western Montana. The only modern species of Polygyrella, P. polygyrella, ranges through northern Idaho and adjacent parts of Montana, southeastern Washington, and northeastern Oregon (Pilsbry 1939; Smith 1943). There are no records from as far east as the Three Forks Quadrangle. The original locality for P. polygyrella was on the eastern slope of the Coeur d'Alene Mountains, in moss and decaying wood in damp spruce forest. At Cataldo, Idaho, it was found in schist rockslides near the base of east- facing slopes near the Coeur d'Alene River; on lava rock, as at Stites, Idaho, it buries itself in the black, coarsely granular soil beneath the rock- slides (Pilsbry 1939). Smith (1943) found it in lava rockslides and reported it to be common at lower elevations in the Clearwater Mountains of Idaho. The main part of its range is in a moun- 258 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 tainous region that receives ~ 50-7 5 cm annual precipitation; just eastward in the rain shadow of the Continental Divide the precipitation falls off sharply to half or a third of this total (U.S. Dept. Commerce 1968). A large proportion of this precipitation falls as rain in the months of April, May, and June. The indicated environment is temperate, cool- er and wetter than that of the Milligan Creek Formation. Both Polygyrella and the Gastrocop- ta are suggestive of wooded conditions, perhaps mature forest with plenty of wood on the ground. Shaded rockslides adjoining wooded stream- banks are another possibility. The fine-grained, bentonitic sediments of the Climbing Arrow For- mation suggest low relief in the immediate area, so perhaps humid forest is the more plausible environment. It is compatible with the flood plain deposition inferred by Robinson (1963). The Three Forks Basin may have drained eastward (Robinson 1961), but it was undoubtedly not in a Cordilleran rain shadow as at the present time. The ecological significance of Helminthoglypta is discussed more fully below under the Dunbar Creek Formation. It is consistent with an equa- ble, mesic climate, although Helminthoglypta and Polygyrella are strictly allopatric at present. The presence of the tropical freshwater snail Biomphalaria pseudammonius contradicts the temperate climatic inferences based on the land snails. According to Taylor (1985), B. pseudam- monius is doubtfully distinct from the living B. glabrata, which has an optimum reproductive temperature of 25°C and fails to reproduce at 20°C. The Biomphalaria is from a different lo- cality (USGS 20010) than the land snails, in an adjoining section, 50-55 m topographically low- er. It seems unlikely that there was enough local relief to throw temperate (i.e., as from higher- altitude) and tropical faunal elements into jux- taposition. Konizeski (1961) described the paleoecology and inferred climate of a biota (including Biom- phalaria cf. B. pseudammonius) from the Doug- lass Creek Basin, Montana, approximately 150 km northwest of the Three Forks Basin. The ver- tebrate fauna is similar to the Pipestone Springs fossil assemblage, correlative with the upper part of the Climbing Arrow Formation but probably somewhat younger than the beds yielding the land gastropods. Plant remains, determined by Axelrod, indicate that the climate was temperate "but whether it was warm temperate ... is not now known. Rainfall was distributed in summer and winter and was not less than 30 to 35 inches. Winters appear to have been comparatively mild, but the frequency of frost or snow cannot be determined from the material at hand" (D. I. Axelrod in Konizeski 1961:1639). Konizeski concluded that the climate was temperate and probably varied seasonally; winters were mild compared to the present. Plant associations were stratified by altitude. The vertebrate assemblage suggests montane woodland rather than a savan- nah or open plains environment. The sedimen- tology indicates a basin profile of low relief with erosion a function of chemical as well as me- chanical weathering. Robinson (1963) suggested that part of the Climbing Arrow Formation may have been de- posited contemporaneously with the Milligan Creek Formation, but the environments inferred from Polygyrella, on the one hand, and Radio- centrum on the other, are so distinct that facies difference seems an inadequate explanation. Taylor (1975:204) remarked that the mollus- can collections from the Climbing Arrow For- mation are of special interest "because they pro- vide a stratigraphic tie with late Eocene and early Oligocene fossil vertebrates in this area." It is unfortunate therefore that the Polygyrella may be a stratigraphically long-ranging species and that the helminthoglyptid is not represented by better material. Gastrocopta sagittaria, with its distinctive, ovate-conic shape, may prove to have biostratigraphic utility. DUNBAR CREEK FORMATION.— The faunule from the Dunbar Creek Formation consists of two species of Gastrocopta, two of Pupoides (Ischnopupoides}, Radiocentrum taylori, and Helminthoglypta bozemanensis. Gastrocopta cordillerae, Pupoides tephrodes, and Radiocen- trum taylori are known only from this formation. Gastrocopta sp. cf. G. montana may be the same species present in the Milligan Creek Formation. Helminthoglypta bozemanensis is questionably present in the Climbing Arrow Formation. The phylogenetic affinities of G. cordillerae axe not known, so it adds no ecological or geographic information to that derivable from G. sp. cf. G. montana, discussed above under the Milligan Creek Formation. The same comments made above for Radiocentrum laevidomus also apply to R. taylori. Pupoides (Ischnopupoides) is indicative of dry conditions. Today, P. hordaceus is "a species of ROTH: EARLY TERTIARY LAND MOLLUSKS 259 the arid plateaus and foothills [of Colorado, Utah, New Mexico, and Arizona], not found in the humid upper zone of the mountains" (Pilsbry 1948:925). Bequaert and Miller (1973) report it from northern Arizona, living in litter of an arid biotope, under shrubs and low trees, near the top of a steep rocky bluff at an elevation of 5,800 ft [1,800 m]. Both P. hordaceus and P. inornatus are found only as empty shells in many more localities than they are known living. This sug- gests that, like Radiocentrum, the subgenus is now undergoing local extinction in many parts of its range. The known distribution is not ob- viously linked to any thermal or precipitation gradient on a regional scale. Helminthoglypta is now basically a Califor- nian genus, with a few species extending into Oregon and Baja California. It occurs over a wide range of habitats from equable, maritime situa- tions along the coast to highly arid conditions in the Mojave Desert. There is some indication of correlation between shell form and climate. The globose-conic species and subspecies of Hel- minthoglypta— H. californiensis (Lea); H. mail- liardi Pilsbry; H.fieldi Pilsbry, 1930; H. nickli- niana awania (Bartsch, 1919)— are all coastal forms, living in temperate climates with few ex- tremes of temperature. Inland forms on the whole tend to be flatter and more tightly coiled, the extremes being the "Mojave Desert series" (Pils- bry 1939) consisting of small, almost planispiral, widely umbilicate shells with the whorls increas- ing slowly in size. An exception is Helmintho- glypta berryi Hanna, 1929, from the foothills of the Sierra Nevada and Tehachapi ranges, an aberrant form that is the most highly fossorial species of the genus. The presence of a strong color pattern in H. bozemanensis does not sug- gest a fossorial mode of life, so that H. berryi is probably not as good a modern analog as H. mailliardi or H. californiensis. To the extent that shell shape in Helminthoglypta is correlated with environment, H. bozemanensis suggests an equable climate with a mean annual range of temperature of less than 1 3°C (data from Wolfe 1979). For the coastal species mentioned above the total annual rainfall varies from over 200 cm in the range of H. mailliardi to less than 30 cm in the range of H.fieldi. The precipitation is con- centrated in the winter months, with practically none between May and October (U.S. Dept. Commerce 1968). Helminthoglypta today is nowhere sympatric with the other genera; its closest approach to any is on the mainland of Los Angeles County, op- posite Santa Catalina Island where Radiocen- trum avalonense occurs. Without postulating dif- ferent climatic tolerances for one or more of these genera in the Paleogene, it is hard to reconcile their joint occurrence in the Dunbar Creek For- mation. About the only inferences one can draw about the Dunbar Creek environment are that it was probably drier than that of the Climbing Arrow Formation, with sparser vegetation (pos- sibly scrub, savannah, or open woodland), and moderate seasonal variation in temperature and precipitation. Presumed caliche horizons in the Dunbar Creek Formation suggest deposition in a seasonally dry basin or bolson (Robinson 1 963) and are compatible with this interpretation. GENERAL TRENDS.— The land mollusk fauna of the Bozeman Group shows three salient char- acteristics: (1) the occurrence of several genera well outside their modern ranges; (2) the seem- ingly paradoxical co-occurrence of genera now widely separated geographically and environ- mentally; and (3) change through time from sparsely vegetated to forested terrain and back again. Tropical to subtropical climates extended to high latitudes during the early Tertiary (Durham 1950;Savinetal. 1975; Savin 1977; Wolfe 1978; Lillegraven 1979). Not all taxa in the Bozeman Group, however, show the simple southward shift of range that one would expect if temperature tolerance were the sole determining factor. Poly- gyrella still occupies the same general region. Helminthoglypta now lives along the Pacific Coast, extending from cool-mesic to warm and arid environments. Radiocentrum has under- gone a southward shift to the American south- west, where it lives in rocky habitats but is ap- parently excluded from regions of extreme aridity. Gastrocopta is absent from the west coast but widely distributed in the eastern states; the sub- genus Albinula approaches but does not overlap the range of Radiocentrum. The present range limitations of Radiocen- trum may involve interactions between genera as well as simple environmental tolerances. The ovoviviparous Oreohelix is now the dominant genus of large land snails throughout much of the Cretaceous-early Tertiary range of Radio- centrum (Bequaert and Miller 1973, fig. 4). Ex- cept for some work on the agonistic behavior of slugs (Rollo and Wellington 1977, 1979), little 260 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 is known about molluscan interactions that could lead to competitive exclusion. However, it is pos- sible that in a climatic context of increasing sum- mer drought and seasonal temperature variation the ovoviviparous mode might have permitted larger and more stable populations of Oreohelix to exist, which then outcompeted Radiocentrum for limiting resources such as shelter sites. A mixture of diverse present-day biogeograph- ical elements also existed in the British Isles dur- ing the Paleogene, among plants and insects as well as land mollusks (Preece 1982 and refer- ences cited therein). Daley (1972) argued that those apparently paradoxical associations rep- resent a climate that has no close modern analog: seasonal but frostless, with high rainfall, and temperatures elevated but not as high as those of tropical rain forest areas today. Table 2 lists land mollusks from western North America (the western Great Plains to the Pacific Coast) that have a Tertiary or late Cretaceous fossil record outside the Holocene ranges of their families or genera. Many fossil land snails are difficult to assign taxonomically (there is even debate over whether certain species are proso- branchs or pulmonates), hence a number of the generic assignments are provisional. Major dif- ferences of interpretation are annotated. It is be- yond the scope of this paper to resolve some of the more difficult taxonomic problems involving these species, but in most cases the biogeographic and paleoenvironmental conclusions are not rad- ically affected. Taxa that are wholly problemat- ical because of inadequate type material (e.g., "Eucalodium" eophilum Cockerell, 1915) or without any convincing modern analogs (Gran- gerellidae) are excluded. From this table it is apparent that families and genera now distributed in many other parts of the world were present in western North America in the late Cretaceous and early Tertiary. The largest block of taxa now exists in the American tropics, but almost as many groups now occur in tropical regions of Africa, Asia, and the Pacific Islands as well as in a Pacific coastal belt ex- tending from Alaska to northern Baja California. Another large group now lives in warm-temper- ate to subtropical parts of the southern United States and Mexico. Others show minor displace- ment within the western interior. None of the taxa are now restricted to far northern America or Eurasia; those are young molluscan faunas, derived as species from middle latitudes colo- nized deglaciated regions in Pleistocene to Ho- locene time. Lower-latitude Tertiary fossil lo- calities (such as those of the lower Miocene of Florida) contain no land snail genera now re- stricted to higher latitudes. The main bulk of local extinction seems to have taken place by the end of the Oligocene, but to some extent the evidence is negative: Mio- cene and Pliocene deposits yielding land mol- lusks are few. However, Miocene and Pliocene faunas are overwhelmingly composed of genera still extant in the region. The extinction of many genera of land mol- lusks over parts of their west North American range may represent the sorting out of formerly sympatric groups into different ecologic/geo- graphic zones. A similar scenario was proposed for the subgenera of Monadenia in the Pacific states (Roth 198 la), and a comparable pattern is evidently involved in the origin and devel- opment of coniferous forests in the west (Axelrod 1976). It is also possible that if the climatic pa- rameters in the microhabitats of snails were bet- ter known, the former association of genera now geographically separated would seem less para- doxical—and the lack of congruence in their modern ranges more attributable to the opera- tions of chance. The answer awaits a closer study of the ecology of living land mollusks. How much of the environmental change shown by Bozeman Group mollusks is the result of sec- ular climatic change, and how much due to local factors such as tectonism? Radiometric dates as- sociated with faunas of the Chadronian North American Land Mammal "age" range between 37.4 ± 1.2 Ma and 32.3 ± 0.7 Ma (Evernden et al. 1964; Prothero, Dunham, and Farmer 1982). The transition from Uintan to Chadronian fau- nas occurs within the Climbing Arrow Forma- tion (Robinson 1963). The Eocene-Oligocene boundary, placed at 36.6 Ma (Palmer 1983), probably also occurs within the Climbing Arrow or Dunbar Creek Formation (Fig. 2). (See also correlation by Lillegraven and Tabrum [1983, fig. 2], except that they place the Eocene-Oligo- cene boundary at 38 Ma.) Lillegraven summarized the evidence for a late Eocene climatic deterioration, beginning perhaps 5-7 million years before the advent of the Oli- gocene. He concluded that "the late Eocene and early Oligocene was represented by a world- wide pulse of increased continentality, oceanic cool- ing, and a significant compression of tropical zones with dilation of temperate conditions. The time was marked by increased rates of extinc- ROTH: EARLY TERTIARY LAND MOLLUSKS 261 TABLE 2. TERTIARY AND UPPER CRETACEOUS LAND MOLLUSK. TAXA FROM WESTERN NORTH AMERICA (WESTERN GREAT PLAINS TO THE PACIFIC COAST) THAT OCCUR OUTSIDE THE HOLOCENE RANGES OF THEIR FAMILIES OR GENERA Taxon Fossil occurrence Holocene range Subclass Prosobranchia Family Helicinidae Eohipptychia eohippina (Cockerell, 1915) Hendersonia evanstonensis (White, 1878) H. oregona (Hanna, 1920) Lucidella (?) buttsi (Russell, 1955) Tozerpina mokowanensis (Tozer, 1956) T. douglasi (Tozer, 1956) T. rutherfordi (Russell, 1929) "Helicina" cretacea Yen, 1954 "H." cokevillensis Yen, 1954 "H." vokesi Hanna, 1936 Family Cyclophoridae Paleocydotusl sp. (Yen 1952) Pseudarinia convexa Yen, 1952 P. pupilla Yen, 1952 P. uniplica Yen, 1954 Rhiostoma americana Hanna, 1920 Subclass Pulmonata Family Ellobiidae Carychium sp. (La Rocque 1960; Taylor 1975) Family Tornatellinidae Protornatellina isoclina (White, 1895) Family Pupillidae Gastrocopta (Gastrocopta) sp. (Taylor 1975) Pupoides (Ischnopupoides) sp. (Taylor 1975) P. (I.) tephrodes n. sp. P. (I.) sp., cf. P. (I.) hordaceus (Gabb, 1866) (this paper) Family Strobilopsidae Eocene, Wyoming Paleocene and Eocene, Wyoming Oligocene or Miocene, Oregon Eocene or Oligocene, British Colum- bia; Oligocene, Montana Upper Cretaceous, Alberta Paleocene, Alberta Paleocene, Alberta Upper Cretaceous, Wyoming Upper Cretaceous, Wyoming Eocene, California Upper Cretaceous, Wyoming Upper Cretaceous, Wyoming Upper Cretaceous, Wyoming Upper Cretaceous, Wyoming Oligocene, Oregon Eocene, Wyoming, Utah Upper Cretaceous, Wyoming Eocene, central Wyoming Eocene, central Wyoming Eocene or Oligocene, Montana Eocene or Oligocene, Montana Helicinid snails with apertural barriers: Greater Antilles, Mexico, Venezuela, Ecuador, Peru, Bolivia, Brazil (Boss and Jacobson 1975); Laos, Szechuan (Wenz 1938; Bishop 1980) Hendersonia: E North America (Solem 1979) Lucidella: Antilles, Central America (Wenz 1938) Note 1 Note 1 Note 1 Cyclophoridae (sensu lato): American tropics, Andes; E and SE Asia; Poly- nesia; E Africa, Malagasy Is. (Solem 1979) Rhiostoma: SE Asia (Wenz 1938) Carychium: North America except NE Cordillera; American tropics; Eu- rope, Asia, Philippines, Indonesia (Pilsbry 1948; Zilch 1959) Tornatellinidae: Polynesia and Juan Fernandez Is. (Solem 1979) Subgenus Gastrocopta: South Dakota to SW U.S.; American tropics; Afri- ca, Mascarene Is., Ceylon, Philip- pines (Pilsbry 1948) Ischnopupoides: South Dakota to Ari- zona (no Montana or central Wyo- ming records); Mexico to Argentina, Chile (this paper) P. (I.) hordaceus: SE Wyoming to New Mexico (Bequaert and Miller 1973) Strobilopsidae, Strobilops: E North America; American tropics; E Asia (Pilsbry 1948); Baja California (Mil- ler and Christensen 1980) 262 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 TABLE 2. CONTINUED. Taxon Fossil occurrence Holocene range Strobilops sp. (Taylor 1975) Family Clausiliidae Genus and sp. indet. (Taylor 1975) Family Subulinidae Pseudocolumna spitzia Tozer, 1956 P. vermicula (Meek and Hayden, 1857) P. haydeniana (Cockerell, 1906) P. spp. (Taylor 1975) Family Urocoptidae Holospira dyeri Tozer, 1956 H. grangeri Cockerell, 1914 #.(?) sp. (La Rocque 1960; McKenna, Robinson, and Taylor 1962; Dorr 1969) H. leidyi (Meek, 1873) //.(?) adventicia Russell, 1955 Genus and sp. indet. (Taylor 1975) Family Bulimulidae Oreoconus jepseni (Russell, 1931) O. planispira Taylor, 1962 O. spp. (Oriel 1962; Taylor 1975) Family Charopidae aff. Charopa (Taylor 1975) Family Discidae Anguispira russelli Tozer, 1956 A. holroydensis Russell, 1956 Family Arionidae Early Tertiary, Wyoming Eocene, Wyoming Upper Cretaceous, Alberta Paleocene, North Dakota Paleocene, Alberta, North Dakota; Eocene, Wyoming Paleocene and Eocene, Wyoming Upper Cretaceous, Alberta Paleocene, New Mexico Paleocene and Eocene(?), Utah; Eocene, Wyoming Eocene, Wyoming Eocene or Oligocene, British Colum- bia Eocene, Montana, Wyoming BulimulusC?) sp. (La Rocque 1960) Eocene, Utah Eocene, Wyoming Eocene, Wyoming Eocene and Oligocene, Wyoming; Eocene, Utah Paleocene, Wyoming Paleocene, Alberta Miocene, Wyoming Clausiliidae: Eurasia; Andes; 2 spp. in Greater Antilles (Solem 1979) Subulinidae: Tropics except Polynesia and Micronesia; S Africa; Mediterra- nean region; high diversity in Africa and tropical South America (Zilch 1959) Note 2 Note 2 Urocoptidae: southern U.S.; American tropics (Zilch 1960) Holospira: Texas, New Mexico, Arizo- na; Mexico (Bequaert and Miller 1973) Note 3 Bulimulidae: In North America, only one species north of southern tier of states (Pilsbry 1946); Central and South America; Australasia Bulimulus (sensu lato): southern U.S.; Mexico to South America (Zilch 1960) Charopidae: Australia, New Zealand, New Caledonia; South Africa; Cen- tral and South America; Idaho to Arizona (Solem 1979) Anguispira: North America, mainly S of U.S.-Canadian border; questiona- bly, Alberta (La Rocque 1953); no Wyoming records ROTH: EARLY TERTIARY LAND MOLLUSKS 263 TABLE 2. CONTINUED. Taxon Fossil occurrence Holocene range Binneya antiqua Russell, 1955 Craterarion pachyostracon Taylor, 1954 Family Zonitidae "Gastrodonta" coryphodontis Cockerell, 1914 "G." imperforata Hanna, 1920 "Omphalina" laminarum Cockerell, 1906 "0." oreodontis Cockerell and Henderson, 1912 VentridensV) lens (Gabb, 1864) K.(?) sp. (Russell 1955) Family Polygyridae Polygyra(>) petrochlora Cockerell, 1914 P.(?)sp. (Taylor 1975) "P." veternior (Cockerell, 1915) "P." expansa Hanna, 1920 "P." martini Hanna, 1920 TriodopsisC?) spp. (Taylor 1975) Vespericola(>) dalli (Stearns, 1885) Family Oleacinidae Genus and sp. indet. (Taylor 1975) Family Camaenidae Caracolus aquilonaris Bishop, 1979 Hodopoeus crassus Pilsbry and Cockerell, 1945 H. hesperarche (Cockerell, 1914) Kanabohelix kanabensis (White, 1876) Pleurodonte (Pleurodonte) wilsoniRoth, 1984 P. (Dentellaria)(>) sp. (Roth 1984) "Helix" spatiosa Meek and Hayden, 1861 "Oreohelix" steini Cockerell, 1914 Genus and sp. indet. (Roth 1984) Eocene or Oligocene, British Colum- bia Miocene, southern California Eocene, Wyoming Oligocene or Miocene, Oregon Oligocene, Colorado Oligocene, Colorado Upper Cretaceous, California Eocene or Oligocene, British Colum- bia Eocene, New Mexico Eocene, Wyoming Eocene, Wyoming Oligocene or Miocene, Oregon Oligocene or Miocene, Oregon Eocene and Oligocene, Wyoming Oligocene or Miocene, central Ore- gon Eocene, Wyoming Oligocene, Nebraska Paleocene(?), SW U.S. Paleocene(?), Texas Upper Cretaceous, Utah Eocene and Oligocene, W Texas Eocene, W Texas Paleocene, Alberta, North Dakota; Paleocene and Eocene, Wyoming; Eocene,, Texas Paleocene, New Mexico Eocene, W Texas Binneya: California Channel Is.; Isla de Guadalupe, Baja California (Pils- bry 1948) Craterarion: possibly Holocene of cen- tral California (Taylor and Roth, MS) Note 4 Note 4 Note 4 Note 4 Note 5. Ventridens: E North America (Pilsbry 1946) Polygyra (including Daedalochila): SE North America, Mexico, Antilles, Bermuda (Pilsbry 1940) Note 6 Note 6 Triodopsis: Washington, Oregon, Ida- ho; E and midwestern U.S. (Vagvol- gyi 1968) Vespericola: Alaska to California; in Oregon, west of Cascade crest (Pils- bry 1940) Oleacinidae: SE North America to Texas; American tropics; Mediterra- nean region (Pilsbry 1948; Zilch 1960) Camaenidae: India to Australia and Solomon Is.; Costa Rica to Peru; Antilles (Solem 1978) Caracolus: Greater Antilles (Bishop 1979) Note 7 Note 7 NoteS Subgenus Pleurodonte: Lesser Antilles (Roth 1984) Subgenus Dentellaria: Jamaica (Roth 1984) Note 9 Note 9 264 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 TABLE 2. CONTINUED. Taxon Fossil occurrence Holocene range Family Ammonitellidae Ammonitella lunata (Conrad, 1871) Polygyrella amplexa (Meek and Hayden, 1857) P. parvula (Whiteaves, 1885) P. venerabilis (Russell, 1937) P. sp., cf. P. polygyrella (this paper) P. sp. (Ross 1959; Taylor 1975) P. polygyrella (Bland and Cooper, 1861) Family Oreohelicidae RadiocentrumC?) anguliferum (Whiteaves, 1885) R. thurstom (Russell, 1926) R. grangeri (Cockerell and Henderson, 1912) R. laevidomus n. sp. R. taylori n. sp. R. hendersoni '(Russell, 1938) Family Helminthoglyptidae Glypterpes rotundatus (Russell, 1931) G. veternus (Meek and Hayden, 1861) Helminthoglypta bozemanensis n. sp. H.(l) sp. (Taylor 1975) Hemitrochus(7) sp. (Taylor 1975) Lysinoe breedlovei 'Roth, 1984 Mesoglypterpes sagensis Yen, 1952 Monadenia antecedens (Stearns, 1900) M. dubiosa (Steams, 1902) M. (ShastelixT) marginicola (Conrad, 1871) Polymita texana Roth, 1984 Xerarionta waltmilleri Roth, 1984 "Helix" adapts White, 1886 "H." nacimientensis White, 1886 Undescribed genus and sp. (Taylor 1975) Oligocene or Miocene, Oregon Upper Cretaceous, Montana Upper Cretaceous, Alberta Upper Cretaceous, Alberta Eocene, Montana Eocene, Montana, Wyoming Oligocene or Miocene, central Ore- gon Upper Cretaceous, Alberta Paleocene, Alberta Eocene, Wyoming Eocene(?), Montana Eocene or Oligocene, Montana Oligocene, Colorado Paleocene, Alberta Eocene, Wyoming Eocene or Oligocene, Montana Oligocene, Wyoming Oligocene, Wyoming Eocene and Oligocene, W Texas Upper Cretaceous, Wyoming Oligocene or Miocene, central Ore- gon Oligocene or Miocene, central Ore- gon Oligocene or Miocene, central Ore- gon Eocene, W Texas Oligocene, W Texas Paleocene, New Mexico Paleocene, New Mexico Eocene, Wyoming Ammonitella: Sierra Nevada, Califor- nia (Pilsbry 1939) Polygyrella: NE Oregon to W Montana (Pilsbry 1939) P. polygyrella: NE Oregon to W Mon- tana (Pilsbry 1939) Radiocentrum: SW U.S. and N Mexi- co; Baja California Sur (this paper) Helminthoglyptidae: Alaska to Califor- nia; SW U.S.; Mexico, Central America, Florida Keys, Antilles; An- des from Ecuador to W Argentina (Pilsbry 1939) Helminthoglypta: S Oregon to N Baja California (this paper) Hemitrochus: S Florida, Antilles (Turner 1958) Lysinoe: Chiapas, Mexico; Central America (Roth 1984) Monadenia: Alaska to California; in Oregon, W of Cascade crest (Roth 1981*) Shastelix: Klamath Mountains, N Cal- ifornia (Roth 1981ft) Polymita: Oriente Province, Cuba (Zilch 1960) Xerarionta: S California to Baja Cali- fornia (Roth 1984) Note 10 Note 10 Notes 1. Solem (1979) suggested relationship to West Indian Camaenidae; now shown to belong to one (Bishop 1980) or more (Solem in press) genera of Helicinidae. ROTH: EARLY TERTIARY LAND MOLLUSKS 265 tions and faunal replacements in many groups of organisms beyond that evident in the first two- thirds of the Eocene. Continental aridity in- creased in interior regions of North America and general world-wide climatic equability de- creased" (Lillegraven 1979:344). Axelrod (1981, table 1) noted that the period around 40 million years before present was one of spreading dry climate in southwestern North America. As a generalization (admittedly much simpli- fied), organisms in the western interior of North America that were most sensitive to cooling tem- peratures should have undergone southward dis- placement: the general trend of isotherms is lat- itudinal. Organisms most sensitive to drought (annual or seasonal) should have been displaced to east or west: away from the complex topog- raphy of the Great Basin and Rocky Mountains, the general trend of isohyets is longitudinal (U.S. Dept. Commerce 1968). What one in fact sees is a mixture of displacements, both among Boze- man Group taxa and among North American land mollusk groups in general (Table 2). At the family level, the late Eocene and early Oligocene was the time of greatest moderniza- tion of the worldwide land mammal fauna, ar- chaic kinds generally adapted to warmer climates giving way to modern varieties more tolerant of the temperate climate of the late Cenozoic (Lil- legraven 1979). For North American land mol- lusks, at least, the time seems not to have been one of evolutionary innovation so much as local extinction and biogeographic rearrangement. ACKNOWLEDGMENTS I am indebted to Dwight W. Taylor for intro- ducing me to the mollusk fauna of the Bozeman Group and for a critical reading of the manu- script. I am also grateful to John H. Hanley for lending material from the U.S. Geological Sur- vey collections, to Glenn Goodfriend for his sug- gestions on the generic allocation of certain fossil species, to Emmett Evanoff for advice on strati- graphic nomenclature, and to Alan Solem for a preprint of his work in press. Scanning electron microscope photographs were taken in the SEM facility of the Department of Entomology, CAS; I thank Mary Ann Tenorio for her competent instruction and assistance. LITERATURE CITED ALBRITTON, C. C, JR. AND K. BRYAN. 1939. Quaternary stratigraphy in the Davis Mountains, Trans-Pecos Texas. Geol. Soc. America Bull. 50:1423-1474. AXELROD, D. I. 1976. History of the coniferous forests, Cal- ifornia and Nevada. Univ. California Publ. Bot. 70. 62 pp. . 1981. Role of volcanism in climate and evolution. Geol. Soc. America Spec. Pap. 185. 59 pp. BABRAKZAI, N., W. B. MILLER, AND O. G. WARD. 1975. Cy- totaxonomy of some Arizona Oreohelicidae (Gastropoda: Pulmonata). Bull. Amer. Malacol. Union (1974):4-1 1. BEQUAERT, J. C. AND W. B. MILLER. 1973. The mollusks of the arid southwest, with an Arizona check list. Univ. Arizona Press, Tucson. BISHOP, M. J. 1 979. A new species of Caracolus (Pulmonata: Camaenidae) from the Oligocene of Nebraska and the biotic history of the American camaenid land snails. Zool. J. Linn. Soc. 67:269-284. . 1980. Helicinid land snails with apertural barriers. J. Moll. Stud. 46:241-246. Boss, K. J. AND M. K. JACOBSON. 1975. Catalogue of the taxa of the subfamily Proserpininae (Helicinidae: Prosobranchia). Occ. Pap. Moll. 4:93-102. BRANSON, B. A. AND D. L. BATCH. 1970. An ecological study of valley-forest gastropods in a mixed mesophytic situation in northern Kentucky. Veliger 12:333-350. CHEATUM, E. P. AND R. W. FULLINGTON. 1973. The aquatic and land Mollusca of Texas. Part two: the Recent and Pleis- tocene members of the Pupillidae and Urocoptidae (Gas- tropoda) in Texas. Dallas Mus. Nat. Hist. Bull. 1(2): 1-67. 2. These Pseudocolumna "can be interpreted as smooth-shelled bulimulids equivalent in shape and size to South American Bostryx (Peronaeus) rather than to any African subulinids" (Solem 1979:281). 3. The Paleocene Holospira sites in New Mexico are about 370 km north of the modern range of the genus (Bequaert and Miller 1973). 4. Genus of original proposal almost certainly incorrect; allocation uncertain, possibly helicoid (Pilsbry 1946; Solem 1979). 5. Pilsbry (in Stewart 1926) assigned this species to Ventridens but later (1946:436) expressed reservations about the as- signment. 6. "Not . . . certainly referable to any recent West Coast genera of Polygyridae" (Pilsbry 1940:893). 7. Solem (1978, 1979) noted resemblance between Hodopoeus and the extant South American camaenid genus Isomeria. 8. Regarded as helicinid by Bishop ( 1 980) but shown by Solem (in press) to be stylommatophoran; possibly camaenid (Solem 1978, 1979). 9. New genus of Camaenidae (D. W. Taylor, personal communication). Related Paleogene species occur in New Mexico and southern California. 10. Allocation uncertain, possibly helminthoglyptid (Solem 1979, H. adapis; Taylor 1975, H. nacimientensis). 266 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 1 1 CHRISTENSEN, C. C. AND W. B. MILLER. 1976. A new Ra- diocentrum (Pulmonata: Oreohelicidae) from Baja Califor- nia, Mexico. Veliger 18:378-380. COCKERELL, T. D. A. 1914. Tertiary Mollusca from New Mexico and Wyoming. Bull. Amer. Mus. Nat. Hist. 33:101- 107. DALEY, B. 1972. Some problems concerning the early Ter- tiary climate of southern England. Palaeogeog., Palaeocli- matol., Palaeoecol. 1 1:177-190. DORR, J. A., JR. 1969. Mammalian and other fossils, early Eocene Pass Peak Formation, central western Wyoming. Contrib. Univ. Michigan Mus. Paleontol. 22:207-219. DOUGLASS, E. 1 903. New vertebrates from the Montana Ter- tiary. Ann. Carnegie Mus. 2:145-199. DURHAM, J. W. 1950. Cenozoic marine climates of the Pacific coast. Geol. Soc. America Bull. 61:1243-1264. EVERNDEN, J. F., D. E. SAVAGE, G. H. CURTIS, AND G. T. JAMES. 1 964. Potassium-argon dates and the Cenozoic mammalian chronology of North America. Amer. J. Sci. 262:145-198. FRANZEN, D. S. AND A. B. LEONARD. 1947. Fossil and living Pupillidae (Gastropoda=Pulmonata) in Kansas. Univ. Kan- sas Sci. Bull. 31:311-411. HANNA, G D. 1920. Fossil mollusks from the John Day Basin in Oregon contained in the Condon Museum of the University of Oregon. Univ. Oregon Publ. 1(6): 1-8. HOCHBERG, F. G., JR., B. ROTH, AND W. B. MILLER. In press. Rediscovery of Radiocentrum avalonense (Pilsbry, 1905) (Gastropoda: Pulmonata). Bull. South. California Acad. Sci. KOBELT, W. 1876-81. Illustriertes Conchylienbuch. Bauer und Raspe, Nurnberg. KONIZESKI, R. L. 1 96 1 . Paleoecology of an early Oligocene biota from Douglass Creek Basin, Montana. Geol. Soc. America Bull. 72:1633-1642. KUENZI, W. D. AND R. W. FIELDS. 1971. Tertiary stratigra- phy, structure, and geologic history, Jefferson Basin, Mon- tana. Geol. Soc. America Bull. 82:3373-3394. LA ROCQUE, A. 1953. Catalogue of the Recent Mollusca of Canada. Nat. Mus. Canada Bull. 129, Biol. Ser. 44. 406 pp. — . 1960. Molluscan faunas of the Flagstaff Formation of central Utah. Geol. Soc. America Mem. 78. 100 pp. LEONARD, A. B. 1950. A Yarmouthian molluscan fauna in the midcontinent region of the United States. Univ. Kansas Paleontol. Contrib., Mollusca 3:1-48. LILLEGRAVEN, J. A. 1979. A biogeographical problem in- volving comparisons of later Eocene terrestrial vertebrate faunas of western North America. Pp. 333-347. in Historical biogeography, plate tectonics, and the changing environ- ment, J. Gray and A. J. Boucot, eds. Oregon State Univ. Press, Corvallis. LILLEGRAVEN, J. A. AND A. R. TABRUM. 1983. A new species of Centetodon (Mammalia, Insectivora, Geolabididae) from southwestern Montana and its biogeographical implications. Contrib. Geol. Univ. Wyoming 22:57-73. MCKENNA, M. C, P. ROBINSON, AND D. W. TAYLOR. 1962. Notes on Eocene Mammalia and Mollusca from Tabernacle Butte, Wyoming. Amer. Mus. Novitates 2102. 33 pp. MEEK, F. B. 1876. A report on the invertebrate Cretaceous and Tertiary fossils of the upper Missouri country. Rept. U.S. Geol. Geog. Surv. Terr. (Hayden) 9. 629 pp. METCALF, A. L. 1980. A new fossil Radiocentrum (Pulmo- nata: Oreohelicidae) from northern Coahuila, Mexico. Nau- tilus 94:16-17. METCALF, A, L. AND W. JOHNSON. 1971. Gastropods of the Franklin Mountains, El Paso County, Texas. Southwest. Nat. 16:85-109. MILLER, W. B. 1 973a. A Recent Oreohelix (Gastropoda: Pul- monata) from Baja California Sur, Mexico. Veliger i 5:332- 334. . 1973/7. Saltational speciation in American Helmin- thoglyptidae (Gastropoda: Pulmonata). Bull. Amer. Mala- col. Union 38:44. -. 198 la. A new genus and a new species of helmin- thoglyptid land snails from the Mojave Desert of California. Proc. Biol. Soc. Washington 94:437-444. -. 1 98 1 b. Helminthoglypta reederi spec. nov. (Gastrop- oda: Pulmonata: Helminthoglyptidae), from Baja California, Mexico. Veliger 24:46-48. MILLER, W. B. AND C. C. CHRISTENSEN. 1980. A new Stro- bilops (Mollusca: Pulmonata: Strobilopsidae) from Baja Cal- ifornia Sur, Mexico. Proc. Biol. Soc. Washington 93:593- 596. ORIEL, S. S. 1 962. Main body of Wasatch Formation near La Barge, Wyoming. Bull. Amer. Assoc. Petrol. Geol. 46: 2161-2173. PALMER, A. R. 1983. The Decade of North American Ge- ology 1983 Geologic Time Scale. Geology 1 1:503-504. PILSBRY, H. A. 1889-90. Helicidae:— Vol. iii. Manual of Conchology, ser. 2, 5:1-216. . 1893-95. Guide to the study of helices. Manual of Conchology, ser. 2, 9:1-336. . 1905. Mollusca of the southwestern states, I: Uro- coptidae; Helicidae of Arizona and New Mexico. Proc. Acad. Nat. Sci. Philadelphia 57:211-290. . 1916-18. Pupillidae (Gastrocoptinae). Manual of Conchology, ser. 2, 24:1-380. -. 1 920-2 1 . Pupillidae (Vertigininae, Pupillinae). Man- ual of Conchology, ser. 2, 26:1-254. . 1 922-26. Pupillidae (Orculinae, Pagodulinae, Acan- thinulinae, etc.). Manual of Conchology, ser. 2, 27:1-369. . 1 939. Land Mollusca of North America (north of Mexico). Acad. Nat. Sci. Philadelphia Monog. 3, 1(1): 1-573. . 1 940. Land Mollusca of North America (north of Mexico). Acad. Nat. Sci. Philadelphia Monog. 3, 1(2):574- 994. . 1946. Land Mollusca of North America (north of Mexico). Acad. Nat. Sci. Philadelphia Monog. 3, 2(1): 1-520. . 1948. Land Mollusca of North America (north of Mexico). Acad. Nat. Sci. Philadelphia Monog. 3, 2(2):521- 1113. . 1953. Inland Mollusca of northern Mexico. II. Uro- eoptidae, Pupillidae, Strobilopsidae, Valloniidae, and Cio- nellidae. Proc. Acad. Nat. Sci. Philadelphia 105:133-167. PREECE, R. C. 1982. The land Mollusca of the British lower Tertiary. Malacologia 22:731-735. PROTHERO, D. R., C. R. DUNHAM, AND H. G. FARMER. 1982. Oligocene calibration of the magnetic polarity time scale. Geology 10:650-653. ROBINSON, G. D. 1961. Origin and development of the Three Forks Basin, Montana. Geol. Soc. America Bull. 72:1003- 1014. . 1 963. Geology of the Three Forks Quadrangle, Mon- tana. U.S. Geol. Surv. Prof. Pap. 370. 143 pp. ROLLO, C. D. AND W. G. WELLINGTON. 1977. Why slugs squabble. Nat. Hist. 86:46-51. . 1979. Intra- and inter-specific agonistic behavior among terrestrial slugs (Pulmonata: Stylommatophora). Canad. J. Zool. 57:846-855. Ross, C. P. 1959 [I960]. Geology of Glacier National Park and the Flathead region, northwestern Montana. U.S. Geol. Surv. Prof. Pap. 296. 125 pp. ROTH: EARLY TERTIARY LAND MOLLUSKS 267 ROTH, B. 198 la. Shell color and banding variation in two coastal colonies of Monadenia fidelis (Gray) (Gastropoda: Pulmonata). Wasmann J. Biol. 38:39-51. . 1981ft. Distribution, reproductive anatomy, and variation of Monadenia troglodytes Hanna and Smith (Gas- tropoda: Pulmonata) with the proposal of a new subgenus. Proc. California Acad. Sci. 42:379-407. . 1 984. Lysinoe (Gastropoda: Pulmonata) and other land snails from Vieja Group, Eocene-Oligocene of Trans- Pecos Texas, and their paleoclimatic significance. Veliger 27:200-218. ROTH, B. AND A. E. BOGAN. 1 984. Shell color and banding parameters of the Liguus fasciatus phenotype (Mollusca: Pulmonata). Amer. Malacol. Bull. 3:1-10. RUSSELL, L. S. 1937. New non-marine Mollusca from the upper Cretaceous of Alberta. Trans. Royal Soc. Canada, ser. 3, 31(4):61-67. . 1955. Additions to the molluscan fauna of the Kish- enehn Formation, southeastern British Columbia and ad- jacent Montana. Bull. Nat. Mus. Canada 136:102-1 19. SAVIN, S. M. 1977. The history of the earth's surface tem- perature during the past 100 million years. Ann. Rev. Earth Planet. Sci. 5:319-355. SAVIN, S. M., R. G. DOUGLAS, AND F. G. STEHLI. 1975. Ter- tiary marine paleotemperatures. Geol. Soc. America Bull. 86:1499-1510. SHIMER, H. W. AND R. R. SHROCK. 1944. Index fossils of North America. John Wiley and Sons, New York. 837 pp. SMITH, A. G. 1943. Mollusks of the Clearwater Mountains, Idaho. Proc. California Acad. Sci., ser. 4, 23:537-554. SOLEM, A. 1978. Cretaceous and early Tertiary camaenid land snails from western North America (Mollusca: Pul- monata). J. Paleontol. 52:581-589. . 1979. Biogeographic significance of land snails, Pa- leozoic to Recent. Pp. 277-287 in Historical biogeography, plate tectonics, and the changing environment, J. Gray and A. J. Boucot, eds. Oregon State Univ. Press, Corvallis. . In press. Lost or kept internal whorls: ordinal dif- ferences in land snails. J. Moll. Stud., Suppl. 12a. STEWART, R. B. 1926 [1927]. Gabb's California fossil type gastropods. Proc. Acad. Nat. Sci. Philadelphia 78:287-447. TAYLOR, D. W. 1954. Nonmarine mollusks from the upper Miocene Barstow Formation, California. U.S. Geol. Surv. Prof. Pap. 254-C67-80. . 1 960. Late Cenozoic molluscan faunas from the High Plains. U.S. Geol. Surv. Prof. Pap. 337. 94 pp. . 1975. Early Tertiary mollusks from the Powder River Basin, Wyoming-Montana, and adjacent regions. U.S. Geol. Surv. Open-file Rep. 75-331. 515 pp. 1985. Evolution of freshwater drainages and mol- luscs in western North America. Pp. 265-321 in Late Ce- nozoic history of the Pacific Northwest, C. J. Smiley, ed. Pac. Div. Amer. Assoc. Adv. Sci., San Francisco. TOZER, E. T. 1956. Uppermost Cretaceous and Paleocene non-marine molluscan faunas of western Alberta. Geol. Surv. Canada Mem. 280. 125 pp. TURNER, R. D. 1958. The genus Hemitrochus in Puerto Rico. Occ. Pap. Moll. 2:153-178. U.S. DEPARTMENT OF COMMERCE. 1968. Climatic atlas of the United States. Government Printing Office, Washington, D.C. 80pp. VAGVOLGYI, J. 1 968. Systematics and evolution of the genus Triodopsis (Mollusca: Pulmonata: Polygyridae). Bull. Mus. Compar. Zool. 136:145-254. WENZ, W. 1938-44. Gastropoda, Teil 1, Allgemeiner Teil und Prosobranchia. Handbuch der Palaozoologie 6(1):1-1505 [1938]; 1506-1639 [1944]. WHITE, C. A. 1883. A review of the non-marine fossil Mol- lusca of North America. U.S. Geol. Surv., Ann. Rep. 3:403- 550. WOLFE, J. A. 1978. A paleobotanical interpretation of Ter- tiary climates in the Northern Hemisphere. Amer. Sci. 66: 694-703. . 1979. Temperature parameters of humid to mesic forests of eastern Asia and relation to forests of other regions of the northern hemisphere and Australasia. U.S. Geol. Surv. Prof. Pap. 1106.37pp. YEN, T.-C. 1 952. Freshwater molluscan fauna from an upper Cretaceous porcellanite near Sage Junction, Wyoming. Amer. Jour. Sci. 250:344-359. ZILCH, A. 1959-60. Gastropoda, Teil 2, Euthyneura. Hand- buch der Palaozoologie 6(2): 1-400 (1959); 401-834 (1960). CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 12, pp. 269-282, 18 figs. May 6, 1986 LATE CENOZOIC MARINE MOLLUSKS FROM TUFF CONES IN THE GALAPAGOS ISLANDS By William D. Pitt Department of Geology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 Matthew J. James1 Department of Paleontology, University of California, Berkeley, California 94720 Carole S. Hickman Department of Paleontology, University of California, Berkeley, California 94720 Jere H. Lipps Department of Geology, University of California, Davis, California 95616 Lois J. Pitt2 ABSTRACT: Palagonite tuff cones on Isla Santa Cruz in the Galapagos Archipelago have yielded fossil marine mollusks, preserved both as individual shells with tuff infillings and as larger fossiliferous limestone inclu- sions. Twenty species (19 gastropods and 1 bivalve) are reported from the Cerro Gallina tuff cone, and representative specimens are illustrated. Two of the nominal species are known only as fossils from the Galapagos; the remaining nominal species are living today, although not necessarily in the archipelago. The two modes of preservation in Galapagos tuff cones reflect two different age relationships between the fossils and the enclosing pyroclastic rock: the individual fossils are more or less contemporaneous with the tuff matrix, while the larger fossiliferous inclusions are incorporated from an older limestone formation. The subaqueously formed and subsequently uplifted tuff cones represent an earlier phase of Galapagos volcanism than the younger Bruhnes-age volcanoes and subaerial flows that dominate the emergent surfaces of the islands today, although geologic evidence suggests that they may have formed no earlier than about 3 million years ago. INTRODUCTION serve fossil records of their contemporaneous . . c .. biotas. Although the geologic history of the Ga- Igneous rocks seldom contain fossils, and ,, , . , , , , • • , . f , • •• ,. lapagos Archipelago is dominated by volcanic oceanic islands of volcanic origin seldom pre- . . iC j- *• +• A- activity, there are at least five distinctive sedi- mentary settings in which remains of marine or- 1 Current address: Department of Geology, Sonoma State J •,,-,• TT- University, Rohnert Park, California 94928. gamsms have been preserved (Lipps and Hick- 2 2444 38th Avenue, Sacramento, California 95822. man 1982; Hickman and LippS 1985). In [269] 270 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 12 addition, there is one volcanic setting that has preserved marine fossils in an unusual manner: palagonite tuff cones, massive topographic struc- tures formed from the products of submarine pyroclastic volcanism but incorporating occa- sional isolated shells and larger fossiliferous limestone inclusions. In this paper we provide the first systematic documentation and illustra- tion of tuff cone faunas in the Galapagos, along with a brief discussion of their occurrence, dis- tribution, and mode of fossilization. This report (contribution number 366 of the Charles Darwin Foundation) is based on a pa- leontological reconnaissance expedition to the Galapagos during February 1982, organized by the senior author and including the remaining authors as participants (see also Pitt and James 1983; Pitt 1984). We thank the Galapagos Na- tional Park Service, the Charles Darwin Re- search Station and its former directors, D. C. Duffy and F. Koster, and Ecuadorian military officials for their cooperation and assistance. We are grateful to A. and J. DeRoi for calling our attention to the presence of marine shells in the tuff cone at Cerro Gallina. Our research has been supported in part by the Committees on Re- search, University of California, Berkeley and Davis, and the Museum of Paleontology, Uni- versity of California, Berkeley. GEOLOGIC SETTING There are six major tuff cones and tuff cone complexes ringing Isla Santa Cruz, a large, low island near the center of the young, active vol- canic archipelago (Fig. 1). The cones are pri- marily of submarine origin and represent an ear- lier, subsequently uplifted phase of Galapagos volcanism than the younger cinder cones, vol- canoes, and subaerial flows (McBirney and Wil- liams 1 969; Bow 1 979). The uplifted cones are now deeply eroded and dissected, emerging as islands from more recent subaerial basalt flows. Although one of the cones is now situated 1 km inland from the modern shoreline (Bow 1979), the others have conspicuous wave-cut exposures and stand out as landmarks along the coast. The absolute age and contemporaneity of the tuff cones on Santa Cruz have not been dem- onstrated. These cones occur below Bruhnes paleomagnetic-age flows of the late Pleistocene and Recent and may represent Matuyama-age volcanism. However, the oldest radiometric date that has been obtained in the archipelago is a potassium-argon date of 4.8 ± 1.87 mybp on the palagonite tuff on South Plazas, part of the Cerro Colorado cone complex (Cox and Dalrymple 1966; Cox 1983). Examination of fossiliferous inclusions in the tuff and consideration of Matuyama-age dates associated with the inferred source of the inclusions (Cox and Dalrymple 1966; Cox 1983) lead us to reject the Plazas date (which is inherently questionable in its large stan- dard deviation). We examined and collected fossils from two of the tuff cones, Cerro Gallina and Cerro Col- orado. The more abundant and diverse material was from Cerro Gallina, and it is this fauna that is treated systematically and illustrated below. The geology of Cerro Colorado is more com- plicated. Here we also observed and collected fossil mollusks from a prominent, richly fossil- iferous limestone bed, originally reported by Durham (1965), that crops out north of and in faulted contact with the main tuff cone complex (see Hickman and Lipps 1985). The faunas of the fossiliferous limestone and tuffaceous sand- stone beds on Santa Cruz and Isla Baltra merit separate consideration and are not treated in this report. Earlier California Academy of Sciences collections from these settings on Baltra and San- ta Cruz were described by Dall (1924), Dall and Ochsner (1928), Hertlein and Strong (1939), and Hertlein (1972); Durham (1979) described a new species of Haliotis from the limestone at Cerro Colorado. Durham's collections are housed in the Museum of Paleontology, University of Cal- ifornia, Berkeley. CERRO GALLINA.— Cerro Gallina stands out as a dissected red hill of bedded lapilli tuffs rising approximately 100 m above sea level on the southwest coast of Isla Santa Cruz (00°42'50"S; 90°29'50"W). Fossils occur as isolated whole shells within the tuff from the base of the exposed cone at sea level to its eroded summit. A fauna of 20 species (19 gastropods and 1 bivalve) was collected. CERRO COLORADO.— Cerro Colorado (Fig. 2), a reddish brown eroded tuff cone remnant, is a conspicuous landmark on the eastern coastline of Santa Cruz (00°34'30"S; 90°10'20"W). It is part of a larger tuff cone complex that includes at least two distinct vents (Bow 1 979). The Island of South Plazas is an offshore remnant of this complex. Although fossils are less abundant in the tuff at Cerro Colorado than at Cerro Gallina, PITT ET AL.: GALAPAGOS TUFF CONE FAUNAS 271 91° GALAPAGOS ISLANDS 90° Pinta Marchena San Salvador I. Genovesa . Baltra CERRO COLORADO -1 San Cristobal :;>^k Santa Marfa '• <3) Espanola FIGURE 1 . Map of the Galapagos Islands showing locations of palagonite tuff cones (dots) on Isla Santa Cruz, including Cerro Gall ina and Cerro Colorado. Some additional tuff cones on other islands are also shown (triangles). their distribution and mode of preservation is more interesting. In some of the well-sorted fa- cies more distal to the main Cerro Colorado vent (Fig. 3), we found individual fossil shells, al- though the proximal tuff on the mainland is largely unfossiliferous. On South Plazas, in poor- ly sorted fades proximal to the second vent, fos- sils also occur as isolated individual shells (Fig. 4), but they are more prominent as large fossil- iferous limestone inclusions (Fig. 5) that have been incorporated from an older fossiliferous limestone that is lithologically and faunally sim- ilar to limestone cropping out in the cliffs im- mediately north of Cerro Colorado. OTHER FOSSILIFEROUS TUFF CONES.— More detailed examination of tuff cones in the Gala- pagos may expand the fauna reported here. Bow (1979) studied several of the tuff cones on Santa Cruz that we did not visit and reported incor- poration of exotic blocks of limestone coquina up to several meters across. Although Darwin apparently did not personally observe or collect fossil material from tuff cones in the Galapagos, he reported (1844) receiving shell fragments imbedded in tuff from an officer who collected them "several hundred feet" above sea level on San Crist6bal. Darwin did, however, study the tuffs from the San Crist6bal cones as well as those on Santiago and Isabela, and was the first geol- ogist to recognize that they had formed sub- aqueously (Darwin 1 844). Fossils have not been reported from the tuff cones on Isla Isabela, and we saw no trace of shell material in our explo- ration of the Tagus cone complex. FOSSILIZATION IN VOLCANIC ROCKS Because volcanic rocks form from molten ma- terial, they do not normally contain fossils, al- 272 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 12 though there are scattered reports of fossils pre- served in predominantly igneous settings. There are peculiar tectonic environments in which marine mollusks have been preserved within thick sequences of oceanic basalt, notably in Oregon and Washington where Tertiary vol- canic sequences have been accreted to the North American continent over a subduction zone (Hickman and Lindberg 1984). Hickman (1976) described two new species of Pleurotomaria from the Siletz River Volcanics in Oregon and figured fragments of two additional species from the Crescent Formation, a thick (at least 5,000 m) Eocene volcanic sequence in northwestern Washington. Individual fossils are not, however, encased in basalt, but occur characteristically in thin limestone lenses or tuffaceous agglomerates within the basalt. Snavely and Baldwin (1948) reported marine mollusks from tuffaceous in- terbeds in the Siletz River Volcanics; tuffaceous agglomerates in the Crescent Formation also pre- serve foraminifera (Berthiaume 1938) and corals (Durham 1 942), as well as marine mollusks. Direct incorporation of organic remains into volcanogenic rocks is more difficult and less common. Subaerial ash falls provide one mech- anism for rapidly burying organisms in a me- dium that has cooled sufficiently to be nonde- structive. Pompeii and Herculaneum are modern examples of the same process that enveloped suc- cessive forests of tree trunks in the Eocene of Yellowstone National Park. Closer to the molten state, tree trunks do oc- casionally leave molds in rapidly cooling basalt flows, where total destruction of volatile organic matter is not complete until the lava is sufficient- ly chilled to preserve the empty space as a hollow tube. There is also the famous mold of a bloated rhinoceros in Miocene Columbia River Basalts in eastern Washington (Beck 1937; Chappell et al. 1949, 1951). Again, this fossil was preserved under very special circumstances involving rapid chilling of the lava. Finally, there are several accounts of fossils preserved in volcanic rocks as inclusions. Late Quaternary fossiliferous xenoliths in the subaer- ially deposited tephra of Surtsey have been dis- cussed in a series of reports dealing with this historic volcanic event (Alexandersson 1970, 1972; Simonarson 1974). Fossiliferous inclu- sions also occur on adjacent Heimaey in the older Vestmann Islands (Jakobsson 1968; Simonarson 1982), and were apparently carried upward in the hot magma to be ejected to their current elevation as these volcanoes emerged from the sea. The individual fossil shells and the fossilifer- ous limestone inclusions in the subaqueously formed Galapagos tuff cones represent yet another mode of preservation of organic remains in a volcanic setting. Our knowledge of the physical and chemical process by which basaltic magmas are palagonitized and subaqueous tuff cones fomed is based primarily on studies in the Ga- lapagos (McBirney and Williams 1969; Simkin 1984). It is therefore appropriate to consider the organic component of these otherwise well- known volcanic structures. GALAPAGOS TUFF CONE FORMATION AND INCORPORATION OF SHELLS AND FOSSILS Explosive submarine volcanism of the type that produced the Galapagos tuff cones is atypical of the current eruptive mode in the archipelago, being more typical of convergent plate bound- aries than of spreading centers and hot spots (see Simkin 1984, for a review of eruptive styles and products in the Galapagos). On Santa Cruz we observed two main types of fragmental deposits resulting from explosive volcanism: relatively young, steep-sided, cinder cones that formed subaerially; and older, consolidated, tuff deposits with more subdued, broader profiles that formed subaqueously when gaseous magma erupted from shallow submarine vents. There are two different ways that molluscan shells have become incorporated into the Gala- pagos tuffs: (1) as living or recently dead indi- viduals that were on the surface or in unconsol- idated sediment adjacent to the vent at the time of eruption, or (2) as blocks of older fossiliferous rock that were incorporated into the tuff as it formed. At Cerro Gallina, fossils occur as isolated shells with infillings of palagonite tuff. These shells were therefore empty at the time of eruption and were infilled and incorporated into the cones during the episodes of cold-water quenching, volumet- ric expansions and fragmentation, chemical al- teration, and cone building. The palagonite tuffs associated with both vents in the Cerro Colorado tuff cone complex contain isolated infilled shells similar to those at Cerro Gallina. Examination of a thin section of one of PITT ET AL.: GALAPAGOS TUFF CONE FAUNAS 273 FIGURES 2-6. Galapagos tuff cones and mode of occurrence and preservation of fossil material. Figure 2. Cerro Colorado, eroded tuff cone remnant on the east coast of Isla Santa Cruz. Figure 3. Seacliff exposure of well-bedded distal facies of a portion of the eroded tuff cone at Cerro Colorado. Figure 4. Individual fossil shell in tuff matrix (at arrow) on South Plazas, Cerro Colorado tuff cone complex. Figure 5. Fossiliferous limestone inclusion in tuff on South Plazas. Figure 6. Thin section of fossil strombid gastropod from the Cerro Colorado tuff cone. a. Outer shell layers showing loss of original microstructure. b. Inner shell layers showing well-preserved crossed-lamellar microstructure. Scale bar = 1 60 nm. these shells (Fig. 6) shows alteration of both the innermost and outermost layers where they are in contact with the tuff matrix. The alteration consists of a loss of shell microstructure (Fig. 6a) in contrast to the well-preserved original crossed- lamellar structure of the interior layers (Fig. 6b). In addition to isolated shells (Fig. 4), which are most evident in the better-sorted and better- bedded distal facies at Cerro Colorado, there are older xenoliths in the form of fossiliferous lime- stone boulders (Fig. 5). The boulders were in- corporated into the tuff as the magma erupted through older rocks, and they are analogous to those found on Surtsey except that they were incorporated subaqueously rather than blown out subaerially. PALEOECOLOGY AND TAPHONOMY The tuff cone mollusks are primarily assigned to species that are alive today and restricted to 274 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 12 FIGURES 7-18. Fossil mollusks from Cerro Gallina, Isla Santa Cruz. Figure 7. Trigonocardia! sp. CAS Geology 61411, CAS Loc. 61227, length 1 1.5 mm. Figure 8. Turritella broderipiana marmorata Kiener, 1843, CAS Geology 61412, CAS Loc. 61226, length 61.4 mm. Figure 9. Turritella rubescens Reeve, 1849, CAS Geology 61413, CAS Loc. 61233, length 11.4 mm. Figure PITT ET AL.: GALAPAGOS TUFF CONE FAUNAS 275 depths of less than 100 meters. Many of the shells in the Cerro Gallina tuff show signs of post- mortem infestation by boring organisms (Fig. 7- 9, 1 1-13, 16), suggesting that they were exposed for a period of time prior to burial. Most of the shells are entire, and we did not encounter frag- mented shell debris suggestive of extensive ex- posure and transportation. The fauna is, how- ever, dominated by relatively thick-shelled species with morphologies resistant to post- mortem destruction. SYSTEMATIC PALEONTOLOGY The specimens upon which this study is based are deposited in the Department of Geology, Cal- ifornia Academy of Sciences (CAS). Voucher specimen numbers are assigned only to figured specimens. All specimens bear Academy locality numbers, and complete locality descriptions are provided in the Appendix and in the locality register maintained in the Department of Inver- tebrates and Geology. Representative fossil spec- imens from the tuff cones will also be deposited in the reference collection at the Charles Darwin Research Station, Isla Santa Cruz, Galapagos. Comparative discussion of taxa treated below is based on examination of material in Academy collections. Class PELECYPODA Subclass HETERODONTA Order VENEROIDA Superfamily CARDIACEA Family CARDIIDAE Subfamily FRAGINAE Genus Trigonocardia Dall, 1 900 Trigonocardia? sp. (Figure 7) DISCUSSION.— This taxon is represented by a single worn partial valve. Sculpture consists of flattened, scaled ribs with narrow, finely cross- threaded interspaces as in Trigonocardia bian- gulata (Broderip and Sowerby, 1 829). This spec- imen is not as convex as in typical Trigonocardia and lacks all of the hinge region, making positive generic and specific allocation impossible with- out additional material. DISTRIBUTION.— Galapagos Islands, fossil. GEOLOGIC OCCURRENCE.— Pleistocene. LOCALITY.-CAS Loc. 61227. FIGURED SPECIMEN (incomplete). — CAS Geology No. 6 1 4 1 1 . Length 1 1.5 mm; width 8.6 mm. Class GASTROPODA Order MESOGASTROPODA Superfamily TURRITELLACEA Family TURRITELLIDAE Subfamily TURRITELLINAE Genus Turritella Lamarck, 1799 Turritella broderipiana marmorata Kiener, 1 843 (Figure 8) Turritella broderipiana Orbigny, 1840:388. Turritella marmorata Kiener, 1843:23, pi. 8, fig. 1. Turritella broderipiana marmorata Kiener. Hertlein 1972:41- 42, fig. 26-28. Discussion. — Turritella broderipiana was originally described but not illustrated, leading to difficulties recognizing the taxon and assessing its relationship to the subsequently described T. marmorata on one hand and T. gonostoma Va- lenciennes, 1832 on the other. Our fossil speci- mens from the Galapagos are most similar to Kiener's (1843) illustrations of T. marmorata, but the above synonymy does not resolve all the problems attending use of the three available names. We summarize these problems below. Reeve (1849, species 6, pi. 2, fig. 6A, B) fig- ured specimens that he assigned to Turritella broderipiana and T. marmorata and placed Kie- ner's name in synonymy with T. broderipiana. Keen (1958:290, fig. 183, and 197 1:392, fig. 438) figured the holotype of T. broderipiana and came to a conclusion counter to that of Reeve: that T. marmorata was a synonym of the more northern 10. Polinices uber (Valenciennes, 1832), CAS Geology 61414, CAS Loc. 61235, length 19.9 mm. Figure 11. Bursa caelata (Broderip, 1833), CAS Geology 61415, CAS Loc. 61225, length 45.1 mm. Figure 12. Cantharus sanguinolentus (Duclos, 1833), CAS Geology 61416, CAS Loc. 61225, length 30.3 mm. Figure 13. Phos laevigatus (A. Adams, 1851), CAS Geology 61417, CAS Loc. 61235, length 30.0 mm. Figure 14. Strombina lanceolata (Sowerby, 1832), CAS Geology 61406, CAS Loc. 61225, length 17.4 mm. Figure 15. Latirus centrifugus (Dall, 1915), CAS Geology 61407, CAS Loc. 61225, length 29.2 mm. Figure 16. Columbella castanea Sowerby, 1832, CAS Geology 61408, CAS Loc. 61225, length 24.1 mm. Figure 17. Conus (Asprella) arcuatus Broderip and Sowerby, 1829, CAS Geology 61409, CAS Loc. 61235, length 25.3 mm. Figure 18. Conus (Cylindrus) lucidus Wood, 1828, CAS Geology 61410, CAS Loc. 61225, length 25.2 mm. 276 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 12 species T. gonostoma rather than the southern T. broderipiana. Hertlein (1972) elected another alternative when he recognized the subspecies Turritella broderipiana marmorata for specimens from Peru as well as specimens from Late Cenozoic deposits on Isla Baltra, Galapagos Islands. We have fol- lowed Hertlein in our identification of specimens from Cerro Gallina because they compare most closely with Hertlein's material from the adja- cent Isla Baltra as well as with CAS specimens from Peru. Both Turritella broderipiana and T. gonosto- ma have been characterized as highly variable (Merriam 1941:9), and evaluation of the species complex is beyond the scope of this paper. In Recent populations, color pattern has been used to separate the two species. The taxonomic sig- nificance of pigmentation has not been evalu- ated, however, and cannot be used to distinguish fossil specimens. Additional material from mainland Ecuador and Peru, where the geo- graphic ranges of the two color-forms overlap, may eventually help resolve this problem. DISTRIBUTION.— Ecuador to Peru, living and fossil. GEOLOGIC OCCURRENCE. — Miocene(?)-Recent. LOCALITIES. -CAS Locs. 61225, 61226, 61238. MATERIAL COLLECTED.— Five specimens. FIGURED SPECIMEN. -CAS Geology No. 61412 (Loc. 61226). Length 61.4 mm; width 32.8 mm. Turritella rubescens Reeve, 1 849 (Figure 9) Turritella rubescens Reeve, 1849, vol. 5, pi. 11, sp. 63; Keen 1958:290, fig. 187, as synonym of T. nodulosa King and Broderip 1832; Keen 1971:394, fig. 445. DISCUSSION.— Specimens from Cerro Gallina match the original figure of Reeve (1 849) and the lower figure of Keen (1971), which illustrates a syntype from the British Museum (Natural His- tory). The four figures of Keen (1971:445) in- dicate the variability of this species. DISTRIBUTION.— Gulf of California to Colombia, living; Ga- lapagos Islands, fossil. GEOLOGIC OCCURRENCE. — Pleistocene-Recent. LOCALITIES. -CAS Locs. 61225, 61233. MATERIAL COLLECTED.— Three specimens. FIGURED SPECIMEN. -CAS Geology No. 6 1 4 1 3 (Loc. 61233). Length 1 1.4 mm; width 6.8 mm. Superfamily NATICACEA Family NATICIDAE Genus Polinices Montfort, 1810 Subgenus Polinices sensu stricto Polinices (Polinices) uber (Valenciennes, 1832) (Figure 10) Natica uber Valenciennes, 1832:266. Polinices uber (Valenciennes, 1832). Carpenter 1857:452-453; Dall and Ochsner 1928:96-97; Hertlein and Strong 1939: 370. Polinices (Polinices) uber (Valenciennes, 1832). Keen 1958: 323, fig. 272; Keen 1971:480, fig. 882. DISCUSSION.— Two incomplete specimens of this species were collected from Cerro Gallina. On one specimen the spire is low, the body whorl globose and smooth, the columellar callus thin, and the umbilicus deep. This specimen does not have a funicle, and the outer lip is missing. Ma- rincovich (1977) discussed the complex relation- ships of .P. uber, P. intemeratus (Philippi, 1853) and P. unimaculatus (Reeve, 1855). DISTRIBUTION.— Cedros Island, western Baja California, throughout the Gulf of California, and south to the Galapagos and Paita, Peru, living; Imperial Formation of California and Galapagos Islands, fossil. GEOLOGIC OCCURRENCE. — Pliocene-Recent. LOCALITIES. -CAS Locs. 61235, 61238. MATERIAL COLLECTED.— Two specimens. FIGURED SPECIMEN. — CAS Geology No. 6 1 4 1 4 (Loc. 61235). Length 19.9 mm; width 17.2 mm. Superfamily CYMATIACEA Family BURSIDAE Genus Bursa Roding, 1798 Bursa caelata Broderip, 1833 (Figure 11) Ranella caelata Broderip, 1833:179. Bursa caelata (Broderip, 1833). Keen 1958:347, fig. 327; Keen 1971:508, fig. 964. DISCUSSION.— Four incomplete fossil speci- mens from the Cerro Gallina tuff cone are most similar morphologically to specimens of the Re- cent Bursa caelata (Broderip, 1833). The fossils have three nodes between varices, four on some of the earlier whorls. Spiral sculpture is worn, but there are indications of possible secondary nodes. Varices are too worn to show sculpture pattern. The typical Recent specimen of B. cae- lata has numerous primary nodes at the shoul- der, with rows of secondary nodes above and below the shoulder, and several rows below the shoulder on the body whorl. Some specimens in lots from Panama and Costa Rica have only three nodes between varices and have few secondary spirals and nodes. The fossil specimens are not complete enough to obtain accurate measure- ments. However, proportions are very close to PITT ET AL.: GALAPAGOS TUFF CONE FAUNAS 277 those of Recent conspecific specimens in the CAS collections. DISTRIBUTION.— Gulf of California to Peru, living; Galapa- gos Islands, fossil. GEOLOGIC OCCURRENCE. — Pleistocene-Recent. LOCALITIES.-CAS Locs. 61225, 61234. MATERIAL COLLECTED. — Five specimens. FIGURED SPECIMEN.— CAS Geology No. 61415 (Loc. 61225). Length 45.1 mm; width 31.8 mm. Superfamily BUCCINACEA Family BUCCINIDAE Genus Cant bar us Roding, 1798 Subgenus Gemophos Olsson and Harbison, 1953 Cantharus (Gemophos) sanguinolentus (Duclos, 1833) (Figure 12) Purpura sanguinolentus Duclos, 1833, pi. 22, fig. 1. Cantharus sanguinolentus (Duclos, 1833). Keen 1958:400, fig. 539; Keen 197 1:561, fig. 1115. DISCUSSION. —The single specimen collected at Cerro Gallina is incomplete, lacking the early whorls, outer lip, and anterior canal. A compar- ison was made between our specimen and Can- tharus janellii (Kiener, 1835-36). Our specimen has low elongated nodes at the shoulder, while C. janellii has more pronounced and rounder nodes. The spiral sculpture is variable and should not be considered a diagnostic feature. Differ- ences between these two species are in the node at the shoulder, the columellar markings, and the color (C. sanguinolentus has columellar pustules while C. janellii has columellar plications). Also, C. sanguinolentus has a pink columella whereas C. janellii has a black columella. DISTRIBUTION.— Outer coast of Baja California through the southern part of the Gulf of California to Guaymas, Mexico, and south to the Ecuadorian mainland, living; Galapagos Is- lands, fossil. GEOLOGIC OCCURRENCE. —Pleistocene-Recent. LOCALITY.-CAS Loc. 61225. FIGURED SPECIMEN.— CAS Geology No. 61416. Length 30.3 mm; width 21.5 mm. Genus Engina Gray, 1839 Engina pyrostoma (Sowerby, 1832) (Not figured) Columbella pyrostoma Sowerby, 1 832: 116-117 (not illustrat- ed). Engina pyrostoma (Sowerby, 1832). Keen 1971:565, fig. 1 128; Hertlein 1972:29. DISCUSSION. —The single specimen collected at Cerro Gallina is worn and incomplete. Never- theless, it exhibits sufficient morphological sim- ilarity to Recent specimens of Engina pyrostoma to warrant recognizing it as a fossil representative of this endemic Galapagos taxon. DISTRIBUTION.— Galapagos Islands, living and fossil. GEOLOGIC OCCURRENCE.— Pliocene-Recent. LOCALITY. -CAS Loc. 61236. Genus Phos Montfort, 1810 Subgenus Metaphos Olsson, 1 964 Phos (Metaphos) laevigatus A. Adams, 1851 (Figure 13) Phos laevigatus Adams, 1851:155 (not figured, but see Emer- son 1967, for discussion of subsequent figuring of the type specimen). Phos chelonia Dall, 1917:578. Strong and Lowe 1936:310, pi. 22, fig. 3 (holotype). Metaphos laevigatus (Adams, 1851). Emerson 1967:99-102, pi. 13, fig. 1-8. Phos (Metaphos) laevigatus (Adams, 1851). Keen 1971:569, fig. 1145. DISCUSSION.— The single specimen from Cerro Gallina lacks early whorls and the anterior canal, and fine shell sculpture details are worn. When whole, it had approximately eight whorls, 14 rounded axials, and a weakly tabulate shoulder sloping to the suture giving the effect of being slightly noded; spirals numerous, whorls straight- sided, body whorl tapering on anterior one-third. DISTRIBUTION.— Galapagos Islands, living and fossil. GEOLOGIC OCCURRENCE. — Pleistocene-Recent. LOCALITY.— CAS Loc. 61235. FIGURED SPECIMEN.— CAS Geology No. 61417. Length 30.0 mm; width 15.5 mm. Family COLUMBELLIDAE Genus Columbella Lamarck, 1799 Columbella cf. C. strombiformis Lamarck, 1822 (Not figured) DISCUSSION.— The single specimen from Cerro Gallina is missing about one-quarter turn from the outer lip and part of the anterior canal. In general outline, the specimen resembles Col- umbella strombiformis. Aperture shape is also similar to that of C. strombiformis when the missing portion of the shell is taken into consid- eration. DISTRIBUTION.— Galapagos Islands. GEOLOGIC OCCURRENCE.— Pleistocene. LocALiTY.-CAS Loc. 61225. 278 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 12 Columbella castanea Sowerby, 1832 (Figure 16) Columbella castanea Sowerby, 1832:1 18; Keen 1971:574, fig. 1154. Discussion. — Columbella castanea differs from other Panamic columbellids in its turreted whorl profile. The single specimen from Cerro Gallina has a second slight angulation at the su- ture that is more pronounced than on living spec- imens. Columbella major Sowerby, 1832, which we also collected as a Pleistocene fossil in terrace deposits on Isla Santa Fe, is distinguished by its rounder periphery and straighter-sided spire pro- file. DISTRIBUTION.— Galapagos Islands, living and fossil. GEOLOGIC OCCURRENCE.— Pleistocene-Recent. LOCAUTY.-CAS Loc. 61225. FIGURED SPECIMEN.— CAS Geology No. 61408. Length 24.1 mm; width 14.2 mm. Genus Anachis H. and A. Adams, 1853 Anachis? sp. indet. (Not figured) DISCUSSION.— An incomplete specimen, ten- tatively assigned to the genus Anachis, was col- lected from Cerro Gallina. The specimen has 14 low, axial ribs that become obsolete below the periphery, where they are replaced by numerous fine, raised spirals. The columella is smooth with a light callus, the aperture is narrow with a rather deep posterior notch, and the outer lip is lirate within. DISTRIBUTION.— Galapagos Islands, fossil. GEOLOGIC OCCURRENCE.— Pleistocene. LOCALITY. -CAS Loc. 61225. Genus Strombina Morch, 1852 Subgenus Strombina sensu stricto Strombina (Strombina) lanceolata Sowerby, 1 832 (Figure 14) Columbella lanceolata Sowerby, 1832:1 16 (not illustrated). Strombina lanceolata (Sowerby, 1832). Keen 1958:394. Strombina (Strombina) lanceolata (Sowerby, 1832). Keen 1971: 601, fig. 1275. Strombina recurva Sowerby. Dall and Ochsner (1928:96) [not Strombina recurva (Sowerby, 1832)]. Strombina gibberula Sowerby. Hertlein (1972:29) [not Strom- bina gibberula (Sowerby, 1832)]. DISCUSSION.— This is one of the most abun- dant species in the fauna at Cerro Gallina. Spec- imens compare favorably both with modem rep- resentatives of the species and with specimens from Isla Baltra that were originally assigned by Dall and Ochsner (1928) and Hertlein (1972) to other species of Strombina (see synonymy). DISTRIBUTION. —Ecuadorian mainland to Galapagos Islands, living and fossil. GEOLOGIC OCCURRENCE.— Pliocene-Recent. LocALrriES.-CAS Locs. 61225, 61234, 61235, 61236, 61237. MATERIAL COLLECTED.— Eighteen specimens. FIGURED SPECIMEN.— CAS Geology No. 61406. Length 17.4 mm; width 8. 1 mm. Family NASSARIIDAE Genus Nassarius Dum6ril, 1 806 Nassarius caelolineatus Nesbitt and Pitt, 1986 (Not figured) Nassarius caelolineatus Nesbitt and Pitt, 1986:294-295, fig. 1, 2, 17a. DISCUSSION.— Abundant specimens of a nas- sariid gastropod at Cerro Gallina compare fa- vorably with both living and fossil specimens from the Galapagos that have been assigned to Nassarius nodicinctus (A. Adams, 1852). Mate- rial from the Galapagos does represent an en- demic taxon, but a new name was required be- cause specimens conspecific with the syntypes of N. nodocinctus have never been collected in the archipelago. DISTRIBUTION.— Galapagos Islands, living and fossil. GEOLOGIC OCCURRENCE.— Pliocene-Recent. LOCALITIES.-CAS Locs. 61225, 61234, 61236. MATERIAL COLLECTED.— Twenty-five specimens. Family FASCIOLARIIDAE Subfamily FASCIOLARIINAE Genus Latirus Montfort, 1810 Latirus centrifugus (Dall, 1915) (Figure 15) Fusinus centrifugus Dall, 1915:56 (not figured). Latirus centrifugus (Dall). Keen 1971:613, fig. 1327. DISCUSSION.— Two fasciolariid specimens col- lected from different parts of the Cerro Gallina tuff cone have the proportions and characteristic ornamentation of Latirus centrifugus. This is the first report of this species as a fossil in the Ga- lapagos. Fasciolariids described by Dall and Ochsner (1928) under Latirus from the older Pliocene limestone bed north of Cerro Colorado have shorter anterior canals and different orna- mentation. PITT ET AL.: GALAPAGOS TUFF CONE FAUNAS 279 DISTRIBUTION.— Galapagos Islands, living and fossil. GEOLOGIC OCCURRENCE.— Pleistocene-Recent. LOCALITIES. -CAS Locs. 61225, 61234. MATERIAL COLLECTED.— Two specimens. FIGURED SPECIMEN. —CAS Geology No. 6 1 407 (Loc. 6 1 225). Length 29.2 mm; width 14.9 mm. Superfamily CONACEA Family CONIDAE Genus Conus Linnaeus, 1758 Subgenus Asprella Schaufuss, 1869 Conus (Asprella) arcuatus Broderip and Sowerby, 1829 (Figure 17) Conus arcuatus Broderip and Sowerby, 1829:379. Conus (Lithoconus) arcuatus Broderip and Sowerby. Keen 1958: 458, fig. 936. Conus (Asprella) arcuatus Broderip and Sowerby. Keen 1971: 663, fig. 1496. DISCUSSION.— Specimens collected over a range of 40 m elevation in the Cerro Gallina tuff cone preserve the slender profile and turreted, faintly nodulose spire diagnostic of this species. DISTRIBUTION.— Gulf of California to Panama, living; Costa Rica to Galapagos Islands, fossil. GEOLOGIC OCCURRENCE.— Pleistocene-Recent. LOCALITIES. -CAS Locs. 61225, 61234, 61235. MATERIAL COLLECTED.— Three specimens. FIGURED SPECIMEN. -CAS Geology No. 6 1 409 (Loc. 6 1 235). Length 25.3 mm; width 12.8 mm. Subgenus Chelyconus Morch, 1852 Conus (Chelyconus) orion Broderip, 1833 (Not figured) Conus orion Broderip, 1833:55; Keen 1958:483 (as a synonym of Conus vittatus Bruguiere, 1792). Conus (Chelyconus) orion Broderip, 1833. Keen 1971:664, fig. 1499. DISCUSSION.— The single specimen from Cerro Gallina is worn but retains the characteristic pro- file and features of this species. DISTRIBUTION.— Mexico to Ecuador, living; Galapagos Is- lands, fossil. GEOLOGIC OCCURRENCE.— Pleistocene-Recent. LOCALITY.— CAS Loc. 16237. Subgenus Cylindrus Deshayes, 1824 Conus (Cylindrus) lucidus Wood, 1828 (Figure 18) Conus lucidus Wood, 1828:8, pi. 3, fig. 4; Hanna 1963:56-58, pi. 6, fig. 1. Conus loomisi Dall and Ochsner, 1928:103, pi. 2, fig. 6. Conus (Cylindrus) lucidus Wood, 1828. Keen 1958:484, fig. 933; Keen 1971:664, fig. 1503. DISCUSSION.— Conus loomisi Dall and Ochs- ner, from Pleistocene terrace deposits on Isla Is- abela, is here considered a synonym of C. (C.) lucidus because fossil specimens in the CAS col- lections clearly show the color pattern of living C. (C.) lucidus. Although the fossil specimens from Cerro Gallina do not preserve color pat- terns, the raised spiral threads distinguish it from cones of similar profile, such as Conus vittatus. For complete synonymy and discussion, see Hanna (1963:56-58). DISTRIBUTION.— Baja California, Mexico, to mainland Ec- uador and the Galapagos Islands, living; Galapagos Islands, fossil. GEOLOGIC OCCURRENCE.— Pleistocene-Recent. LocALmES.-CAS Locs. 61224, 61225, 61234, 61235, 61237. MATERIAL COLLECTED.— Seven specimens. FIGURED SPECIMEN. —CAS Geology No. 6 1 4 1 0 (Loc. 6 1 225). Length 25.2 mm; width 14.3 mm. Family TEREBRIDAE Genus Terebra Bruguiere, 1789 Terebra armillata Hinds, 1 844 (Not figured) Terebra armillata Hinds, 1844:154; Keen 1958:490, fig. 956; Keen 1971:672, fig. 1522; Bratcher and Burch 1971:556- 557, fig. 27. DISCUSSION.— Two fragmentary terebrid spec- imens from Cerro Gallina, each consisting of ap- proximately two whorls, have proportions and sculpture that place them within the range of variation that Bratcher and Burch (1971) de- scribed for this species. DISTRIBUTION.— Baja California, Mexico, to Peru and the Galapagos Islands, living; Galapagos Islands, fossil. GEOLOGIC OCCURRENCE. — Pleistocene-Recent. LOCALITY.-CAS Loc. 61225. Terebra plicata Gray, 1834 (Not figured) Terebra plicata Gray, 1834:61; Keen 1971:682, fig. 1556. DISCUSSION.— This species is represented in our collections by a single specimen consisting of one whorl. The sculpture, although worn, is suffi- ciently distinctive to place it within the range of variation in CAS specimens of living represen- tatives of the species. DISTRIBUTION.— Central America to the Galapagos Islands, living; Galapagos Islands, fossil. GEOLOGIC OCCURRENCE.— Pleistocene-Recent. LOCALITY. -CAS Loc. 61236. 280 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 12 Family TURRIDAE? (Not figured) DISCUSSION.— Two poorly preserved speci- mens from Cerro Gallina are tentatively referred to the Turridae. LOCALITY. -CAS Loc. 61236. APPENDIX Locality Data CERRO GALLINA, ISLA SANTA CRUZ (00°42'50"S; 90°29'50"W) CAS 61224. Low tuff cliff just above beach level, west side of small cove on west side of Cerro Gallina. (Field no. G-l- 82.) Collected by W. D. Pitt and J. H. Lipps, 2 Feb. 1982. CAS 6 1 225. Southeast side of Cerro Gallina at beach on east side of small cove bounded by tuff cliffs, including all exposures along sea cliff from gully running into head of cove eastward to last accessible cliffs (about 30 m). Fossils occur as isolated shells in tuff. (Field no. G-4-82.) Collected by W. D. Pitt, L. J. Pitt, C. S. Hickman, J. H. Lipps, M. J. James, 2 Feb. 1982. CAS 61226. Stratified, water- worked unit in massive tuff, at head of beach in tuff beds gently sloping across top of beach immediately east of gully entering head of cove. (Field no. G-5-82.) Collected by C. S. Hickman and W. D. Pitt, 2 Feb. 1982. CAS 61227. Small ridge trending south toward beach, just above sea cliff on east side of cove and just above CAS 61226 (Field no. G-5-82). No field number. Collected by J. H. Lipps, 2 Feb. 1982. CAS 61233. Approximately 12m above sea level on second ridge north of beach in gully entering head of cove. (Field no. G-9-82.) Collected by C. S. Hickman, 2 Feb. 1982. CAS 61234. Halfway up gentle ridge to sharp break in slope leading up to peak of Cerro Gallina; ridge is third one north of beach in gully and causes a bend in the gully where it in- tersects it. Approximately 20 m above sea level. (Field no. G-6-82.) Collected by J. H. Lipps, 2 Feb. 1982. CAS 61235. Massive outcrop of tuff approximately 40 m above sea level on same ridge described in CAS 61234. (Field no. G-6-82.) Collected by J. H. Lipps and C. S. Hickman, 2 Feb. 1982. CAS 61236. On same ridge as CAS 61234 (Field no. G-6- 82) in massive tuff above principal break is slope leading to top of Cerro Gallina. (Field no. G-8-82.) Collected by J. H. Lipps and C. S. Hickman, 2 Feb. 1982. CAS 61237. On ridge trending south from top of Cerro Gal- lina, approximately 20 m below summit. (Field no. G- 10-82.) Collected by J. H. Lipps and C. S. Hickman, 2 Feb. 1982. CAS 61238. Tuff cliff on southeast side of Cerro Gallina, on northwest side of sandy beach approximately 1 Vz m up sea cliff. (Field no. G-3-82.) Collected by C. S. Hickman and W. D. Pitt, 2 Feb. 1982. CERRO COLORADO, ISLA SANTA CRUZ (00°34'30"S; 90°10'20"W) CAS 61228. Red tuff hill opposite South Plazas Island, northeastern coast of Santa Cruz. Isolated fossils collected on north and west slopes of cone. (Field no. G-56-82.) Collected by W. D. Pitt, C. S. Hickman, J. H. Lipps, 1 1 Feb. 1982. LITERATURE CITED ADAMS, A. 1851. A monograph of Phos, a new genus of gasteropodous Mollusca. Proc. Zool. Soc. Lond., Pt. 1 8 ( 1 850): 152-155. . 1 852. Catalogue of the species ofNassa, a new genus of gasteropodous Mollusca belonging to the family Buccin- idae, in the collection of Hugh Cuming, Esq., with the de- scription of some new species. Proc. Zool. Soc. Lond., Pt. 19:94-114. ADAMS, H. AND A. ADAMS. 1 853-1 854. The genera of Recent mollusca; arranged according to their organization. Vol. I. John Van Voorst, London, xl + 484 pp. ALEXANDERSSON, T. 1 970. The sedimentary xenoliths from Surtsey: marine sediments lithified on the sea-floor— a pre- liminary report. Surtsey Res. Progr. Rep. V:83-89. . 1972. The sedimentary xenoliths from Surtsey: tur- bidites indicating shelf growth. Surtsey Res. Progr. Rep. VI: 101-116. BECK, G. F. 1937. Remarkable west American fossil, the Blue Lake rhino. Mineralogist 5(8):7-8; 20-21. BERTHIAUME, S. A. 1938. Orbitoids from the Crescent For- mation (Eocene) of Washington. J. Paleontol. 12(5):494- 497. Bow, C. S. 1979. The geology and petrogenesis of lavas of Floreana and Santa Cruz Islands: Galapagos archipelago. Unpubl. Doctoral Dissertation, University of Oregon, Eu- gene. 308 pp. BRATCHER, T. AND R. D. BURCH. 1971. The Terebridae (Gas- tropoda) of Clarion, Socorro, Cocos, and Galapagos Islands. Proc. Calif. Acad. Sci., Ser. 4, 37(21):537-566. BRODERIP, W. J. 1833. Characters of new species of Mollusca and Conchifera, collected by Mr. Cuming. Proc. Zool. Soc. Lond. (for 1832):50-61; 173-179. BRODERIP, W. J. AND G. B. SOWERBY. 1829. Observations on new or interesting Mollusca contained, for the most part, in the museum of the Zoological Society. Zool. J., London 4:359-379. BRUGUIERE, J. 1782-1832. Tableau encyclopedique et me- thodique des trois regnes de la nature. Vers, coquilles, mol- lusques, et polypiers. 196 vols. in 186. Paris: [For collation and dates of the zoological portion of "Encylopedie me- thodique," see C. D. Sherborn and B. B. Woodward, Proc. Zool. Soc. Lond. (for 1893):582-584; ibid, (for 1899):595; Ann. Mag. Nat. Hist., Ser. 7, 17:577-582, 1906; E. V. Coan, Veliger9:132-133, 1966.] CARPENTER, P. P. 1857. Catalogue of the collection of Ma- zatlan shells in the British Museum. Oberlin Press, War- rington (reprinted 1967 by the Paleontological Research Institution, Ithaca, New York), i-vi + ix-xvi + 552 pp. CHAPPELL, W. M., J. W. DURHAM, AND D. E. SAVAGE. 1949. Rhinoceros mold in basalt. Bull. Geol. Soc. Am. 60:1949 (abstr.). . 1951. Mold of a rhinoceros in basalt, lower Grand Coulee, Washington. Bull. Geol. Soc. Am. 62:907-918. Cox, A. 1983. Age of the Galapagos Islands. Pp. 11-23 in Patterns of evolution in Galapagos organisms, R. I. Bow- man, M. Berson, and A. E. Leviton, eds. American Asso- ciation for the Advancement of Science, Pacific Division, San Francisco. Cox, A. AND G. B. DALRYMPLE. 1966. Paleomagnetism and Potassium-argon ages of some volcanic rocks from the Ga- lapagos Islands. Nature (London) 209:776-777. DALL, W. H. 1900. Contributions to the Tertiary fauna of PITT ET AL.: GALAPAGOS TUFF CONE FAUNAS 281 Florida with especial reference to the silex-beds of Tampa and the Pliocene beds of the Caloosahatchie River, including in many cases a complete revision of the generic groups treated of and their American Tertiary species. Part V. Te- leodesmacea: Solen to Diplodonta. Trans. Wagner Free Inst. Sci. Philadelphia 3(5):949-1218, + pis. 36-47. . 1915. Notes on the west American species of Fusinus. Nautilus 29(5):54-57. . 1917. Summary of the mollusks of the Alectrionidae of the west coast of America. Proc. U.S. Nat. Mus. 5 1 (2 1 66): 575-579. . 1924. Note on fossiliferous strata on the Galapagos Islands explored by W. H. Ochsner of the Expedition of the California Academy of Sciences in 1905-1906. Geol. Mag. 61:428-429. DALL, W. H. AND W. H. OCHSNER. 1 928. Tertiary and Pleis- tocene Mollusca from the Galapagos Islands. Proc. Calif. Acad. Sci., Ser. 4, 17(4):89-139. DARWIN, C. R. 1844. Geological observations on volcanic islands. Smith, Elder and Co., London. 175 pp. DESHAYES, G. P. 1824-1837. Description des coquilles fos- siles des environ de Paris. Vols. 1-2. Paris. DUCLOS, P. L. 1833. (various genera) Mag. Zool. Paris Yr. 3, cl. 5, Oliva, pi. 20 and text; Purpura, pi. 22 and text. DUMERIL, A. M. C. 1806. Zoologie analytique, ou method naturelle de classification des animaux. Paris. 344 pp. DURHAM, J. W. 1942. Eocene and Oligocene coral faunas of Washington. J. Paleontol. 16(1):84-104. . 1965. Geology of the Galapagos. Pacif. Discov. 18: 3-6. . 1979. A fossil Haliotis from the Galapagos Islands. Veliger 2 1:369-372. EMERSON, W. K. 1967. On the identity of Phos laevigatus A. Adams, 1851 (Mollusca: Gastropoda). Veliger 10(2):99-102. GRAY, J. E. 1834. Enumeration of the species of the genus Terebra, with characters of many hitherto undescribed. Proc. Zool. Soc. Lond., Pt. 2:59-63. . 1839. Molluscous animals and their shells. In The zoology of Capt. Beechey's voyage to the Pacific and Behr- ing's Straits in his Majesty's ship Blossom, F. W. Beechey. London, i-xii + pp. 103-155, pis. 33-44. HANNA, G D. 1963. West American mollusks of the genus Conus— II. Occas. Papers Calif. Acad. Sci., No. 35, 103 pp. HERTLEIN, L. G. 1972. Pliocene fossils from Baltra (South Seymour) Island, Galapagos Islands. Proc. Calif. Acad. Sci., Ser. 4, 39(3):25-46. HERTLEIN, L. G. AND A. M. STRONG. 1939. Marine Pleis- tocene mollusks from the Galapagos Islands. Proc. Calif. Acad. Sci., Ser. 4, 22(24):367-380. HICKMAN, C. S. 1976. Pleurotomaria (Archaeogastropoda) in the Eocene of the Northeastern Pacific: a review of Ce- nozoic biogeography and ecology of the genus. J. Paleontol. 50:1090-1102. HICKMAN, C. S. AND D. R. LINDBERG. 1 984. Relationship of molluscan biogeographic anomalies to changing settings on an active continental margin. Geol. Soc. Am., Abstracts with Programs 16:289 (abstr.). HICKMAN, C. S. AND J. H. LIPPS. 1985. Geologic youth of Galapagos Islands confirmed by marine stratigraphy and paleontology. Science 227(4694): 1578-1 580. HINDS, R. B. 1 844. Descriptions of new shells, collected dur- ing the voyage of the Sulphur, and in Mr. Cuming's late visit to the Philippines. Proc. Zool. Soc. Lond. (for 1843) 1 1:149- 159. JAKOBSSON, S. 1968. The geology and petrography of the Vestmann Islands— a preliminary report. Surtsey Res. Progr. Rep. IV: 11 3- 130. KEEN, A. M. 1958. Sea shells of tropical west America. 1st ed. Stanford Univ. Press, Stanford, California, xi + 642 pp. . 1971. Sea shells of tropical west America. 2nd ed. Stanford Univ. Press, Stanford, California, i-xvi -I- 1064 pp. KIENER, L. C. 1834-1870. Species general et iconographie des coquilles vivantes. Vols. 1-1 1, livr. 1-165. Paris. KING, P. P. AND W. J. BRODERIP. Description of the Cirrhi- peda, Conchifera and Mollusca, in a collection formed by the officers of H. M. S. Adventure and Beagle employed between the years 1826 and 1830 in surveying the southern coasts of South America, including the Straits of Magalhaens and the coast of Tierra del Fuego. Zool. J., London 5:332- 349. LAMARCK, J. P. B. 1 799. Prodrome d'une nouvelle classifi- cation des coquilles. Soc. Hist. Nat. Paris, Mem. 1:63-90. . 1815-1822. Histoire naturelle des animaux sans ver- tebres. Vols. 1-7. Paris. [For the dates of issue of this work, see C. D. Sherborn, 1922, Index Animalium, Sect. 2, p. Ixxvii, and T. Iredale, 1922, Proc. Malacol. Soc. Lond. 15: 85.] LINNAEUS, C. 1758. Systema naturae per regna tria naturae. Editio decima, reformata. Vol. 1, Regnum animale. Stock- holm. 824 pp. LIPPS, J. H. AND C. S. HICKMAN. 1982. Paleontology and geologic history of the Galapagos Islands. Geol. Soc. Am., Abstracts with Programs 14(7): 548 (abstr.). MCBIRNEY, A. R. AND H. WILLIAMS. 1969. Geology and pe- trology of the Galapagos Islands. Geol. Soc. Am. Memoir 118. 197pp. MARINCOVICH, L. N., JR. 1977. Cenozoic Naticidae (Mol- lusca: Gastropoda) of the Northeastern Pacific. Bulls. Am. Paleontol. 7(294): 17 1-494. MERRIAM, C. W. 1941. Fossil turritellas from the Pacific Coast region of North America. Univ. Calif. Pubs., Bull. Dept. Geol. Sci. 26(1): 1-2 14. MONTFORT, D. DE. 1810. Conchyliologie systematique et classification methodiques des coquilles. Vol. 2. Paris. 767 pp., 161 figs. MORCH, O. A. L. 1852-1853. Catalogus conchyliorum quae reliquit D. Alphonso D'Aguirra et Gadea Comes de Yoldi. Copenhagen. Fasc. 1 , 1 70 pp; Fasc. 2, 74 pp. NESBITT, E. A. AND W. D. PITT. 1986. Nassarius (Gastropoda: Neogastropoda) from the Galapagos Islands. Veliger 28(3): 294-301. OLSSON, A. A. 1964. Neogene mollusks from northwestern Ecuador. Paleontol. Res. Inst. 256 pp. + 38 pis. OLSSON, A. A. AND A. HARBISON. 1953. Pliocene Mollusca of southern Florida, with especial reference to those from North Saint Petersburg. Monographs, Acad. Nat. Sci. Phila- delphia No. 8. viii + 457 pp. PHILIPPI, R. A. 1853. Descriptiones naticarum quarundam novarum ex collectone Cummingiana. Proc. Zool. Soc. Lond. (for 1851):233-234. ORBIGNY, ALCIDE D'. 1 840. Voyage dans 1'Amerique M6ri- dionale. Mollusques. Paris 5(3):I-XLH + 1-758. PITT, W. D. 1984. Late Cenozoic invertebrate paleontology of the Galapagos Islands. Annual Report of the Charles Dar- win Research Station (for 1 982) (Informe Anual 1 .982):258- 268. PITT, W. D. AND M. J. JAMES. 1983. Late Cenozoic marine 282 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 12 invertebrate paleontology of the Galapagos Islands. Annu. Rep. Western Soc. Malacol. 15 (for 1982): 14- 15 (abstr.). REEVE, L. 1 846-1 849. Initiamenta conchologica or elements of conchology. London. 160 pp. . 1855. Account of the shells collected by Captain Edward Belcher, C.B., North of Beechey Island. Vol. 2, Pp. 392-399 in The last of the arctic voyages; being a narrative of the expedition in HMS ASSISTANCE, under the com- mand of Captain Sir Edward Belcher, C.B., in search of Sir John Franklin, during the years 1852-53-54, with notes on the natural history, Sir John Richardson, ed. London. RODING, P. F. 1798. Museum Boltenianum; pars seconda contiens Conchylia. J. C. Trappii, Hamburg, i-vii + 109 pp. SCHAUFUSS, L. W. 1869. Molluscorum systema et catalogus. System und Aufz&hlung sammtlicher Conchylien der Samm- lung von Fr. Paetel. Oscar Weiske, Dresden. 1 1 9 pp. SIMKIN, T. 1984. Geology of Galapagos. Pp. 15-41 in Ga- lapagos (Key Environments), R. Perry, ed. Pergamon Press, Oxford, England, x + 321 pp. SIMONARSON, L. A. 1974. Fossils from Surtsey— a prelimi- nary report. Surtsey Res. Progr. Rep. VII:80-82. . 1982. Fossils from Heimaey, Iceland. Surtsey Res. Progr. Rep. IX: 152-1 54. SNA VELY, P. D., JR. AND E. M. BALDWIN. 1948. Siletz River Volcanic Series, northwestern Oregon. Bull. Am. Assoc. Pe- trol. Geol. 32:805-812. SOWERBY, G. B. 1 832. Characters of new species of Mollusca and Conchifera, collected by Hugh Cuming. Proc. Zool. Soc. Lond. for 1832:113-120. STRONG, A. M. AND H. N. LOWE. 1936. West American species of the genus Phos. Trans. San Diego Soc. Nat. Hist. 8(22):305-320. VALENCIENNES, A. 1821-1833. Coquilles univalves. In Voy- age aux regions equinoxiales du Nouveau Continent. Recueil d'observations de zoologie et d'anatomie comparee, F. H. A. Von Humboldt and A. J. A. Bonpland. Paris. 2:262-339. WOOD, W. 1828. Supplement to the Index Testacoelogius; or a catalogue of shells, British and foreign. London, i-vi, 1-59 (privately published). CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 13, pp. 283-334, 20 figs. February 25, 1987 THE CERAMBYCIDAE, OR LONGHORNED BEETLES, OF SOUTHERN TEXAS: A FAUNAL SURVEY (COLEOPTERA) By Frank T. Hovore Placenta Canyon Nature Center, 19152 W. Placenta Canyon Road, Newhall, California 91321 Richard L. Penrose California Department of Food and Agriculture, 1220 N Street, Sacramento, California 95814 Raymond W. Neck Texas Parks and Wildlife Department, 4200 Smith School Road, Austin, Texas 78744 ABSTRACT: An annotated species list of the longhorned wood-boring beetles (Coleoptera: Cerambycidae) is presented for southern Texas. The area surveyed roughly corresponds to the Texas portions of the Matamoran and Nuecian districts of the Tamaulipan Biotic Province, including all of the lower Rio Grande valley. Data given for the 178 species include original author citation, range, adult activity period, confirmed larval hosts, and anecdotal collecting and locality information. We propose no taxonomic changes, and nomenclature corresponds to the most recent literature. The species list is ordered according to the monographic revision of the family Cerambycidae (Linsley 1962o, b, 1963o, 1964; Linsley and Chemsak 1976, 1985), excepting that portion of the subfamily Lamiinae not yet treated by those authors, which is ordered according to the checklist of the Cerambycidae (Chemsak and Linsley 1982). Brief accounts of the biological, ecological, and historical aspects of the fauna are discussed. Prior literature on southern Texas Cerambycidae is summarized and collated. Species reared from selected native plants are listed by host, with an updated account of species known to infest Citrus in southern Texas. The origins and phyletic relationships of the fauna are briefly discussed, with a summary of some of the taxonomic limitations complicating faunal analyses of Neotropical Cerambycidae. Literature cited includes all original species descriptions. INTRODUCTION reduced in size. Variation within the family is extreme; North American genera range in av- Adult Cerambycidae are characteristically erage length from 3 mm (Cyrtinus) to over 70 elongate, subcylindrical beetles with long anten- mm (Derobrachus) and vary in appearance from nae, fully developed hind wings (numerous obscure, drab ground-dwelling forms to brightly species, however, are flightless), and five- colored, contrastingly patterned insect jewels, ca- segmented tarsi with the fourth segment greatly pable of swift flight. They are equally diverse [283] 284 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 ecologically and behaviorally, occupying thou- sands of forest ecosystem microhabitats and niches, and partitioning resources to permit com- mon use of limited amounts of suitable host plant material. Gosling (1981) documented the pres- ence of active populations of over 100 species of cerambycids in an 80-ha woodlot in Michigan, and the number of species and population den- sities are even greater in Neotropical forest eco- systems. Chemsak and Linsley (1970) reported collecting 55 species of cerambycids at a light one August evening in the thorn forest north of Mazatlan, Sinaloa, Mexico. This figure repre- sents no more than one-third of the total long- horned beetle fauna at that locality, since it does not include species taken other nights, nonpho- totropic species, species not active during that particular season, or any of the numerous diurnal species known to occur there. Cerambycids are phytophagous, and as a fam- ily they utilize their host plants from rootlets to buds. Adult beetles may feed upon flowers, leaves, pine cones and needles, fruit, sap, fungi, or bark; while larval Cerambycidae feed externally upon roots, or bore within living, dying, or dead trunks, branches, stems, bark, floral stalks, or roots of both herbaceous and woody plants. Most species utilize existing suitable host materials for larval development, but a few genera create a larval habitat by girdling (either externally as adults, or internally as larvae) portions of living plants. No cerambycids are known to be truly predaceous, but adults of certain mimetic species of Elytro- leptus have been observed feeding upon their lycid beetle models. The larval habits of this ge- nus are unrecorded, but adults of two species have been taken from pupal cells in dead twigs, and the larvae probably feed upon dead wood (see Eisner et al. 1962, for a discussion of Ely- troleptus predation upon Lycidae). The only oth- er account of Cerambycidae as predators, by Bit- tenfeld (1948), shows adult Aromia moschata (Linnaeus) eating young spiders, but it is gen- erally regarded with suspicion due to its lack of detailed observations and the absence of any sub- sequent corroboration. Cerambycid larvae are whitish or yellowish, elongate, cylindrical or subquadrate in cross- section, with rounded heads and powerful chew- ing mouthparts. Growth and development may be quite rapid, with several generations maturing annually, or very slow, extending over several years. Larval feeding may be confined to a spe- cific part of the host, particularly in species uti- lizing living plants, or the larvae may tunnel throughout the woody portions of the host, carv- ing galleries several meters long. A few genera degrade or destroy large volumes of harvested timber; others attack and weaken shade, fruit, and forest trees. Most species of Cerambycidae, however, breed in shrubs and trees of little current economic importance. Overall, longhorned beetles are essential to forest decom- position, recycling vast amounts of dead plant material. Larval feeding activities may alter a considerable volume of dead host material; Ho- vore and Penrose (1982) found that larval work- ings resulted in a wood-mass reduction of up to 70% in dead Leucaena in Texas. Additionally, larval galleries and adult emergence holes permit access into the wood for water, fungus, and soft- bodied insects such as termites and ants. Many adult cerambycids feed upon pollen and other portions of flowers, thereby serving as pollinators for many plant species. And both adult and im- mature life stages are a major food source for a broad spectrum of arthropod and vertebrate predators. Based upon extant study material, the Neo- tropical Cerambycidae are both the most evo- lutionarily diversified and least-studied portion of the world longhorned beetle fauna. Over 5,000 Neotropical species have already been character- ized, with many more thousands awaiting de- scription or discovery. Unfortunately, the New World tropics are rapidly disappearing before the onslaught of unregulated land and resource usage; in many regions little remains of the original tropical forests. At this writing the estimated ex- tinction rate for tropical organisms is one per day, with forests being cut at a rate of between 25 and 1 00 ha per minute (Wiley 1982). As forest tree species become extinct, their obligate faunas also disappear, altering or destroying many long- established interrelationships and trophic pat- terns. By virtue of the inseparable and often narrowly circumscribed relationships between Cerambycidae and their host plants, the popu- lation dynamics of these insects may well reflect the general health, or decline, of an overall forest ecosystem. In North America, the forest habitats in great- est jeopardy are those combining small geo- graphical size with accessible or economically desirable resources, climates, or soils. Thus, the semitropical regions of southern Florida and HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 285 southern Texas are North America's most crit- ically threatened major ecosystems, with only fractional remnants of the original biota persist- ing in either area. HISTORICAL ACCOUNTS FROM LITERATURE Although primarily confined to remnant hab- itats in the extreme southern portion of Texas, the Neotropical cerambycid fauna is remarkable for its species diversity and abundance. The low- er Rio Grande valley (Lower Valley) has received considerable entomological attention, and since the appearance of the first brief species account by Schwarz in 1896, no fewer than six lists of the longhorned wood-boring beetles have been published. Wickham (1898) recorded the collection of 26 species of cerambycids from the Lower Valley, and Townsend (1902) presented an annotated record of over 40 species from Texas and adja- cent Tamaulipas, Mexico. Results of the 1 904 and 1905 Kansas entomological expeditions to the Texas Gulf Coast were catalogued by Snow ( 1 906), and included 22 species from the vicinity of Brownsville. Schaeffer (1 908), in his extensive list of Cerambycidae from Brownsville, recorded 78 species and commented upon the validity of some of the previous accounts (not including Snow's list). Discrepancies in data citations, and identifications based upon outdated or synony- mized names preclude a precise collation of data from these earliest accounts. We have, wherever possible, updated and explained changes in sta- tus or nomenclature. Of greatest value to this study were the excel- lent species accounts of Linsley and Martin (1933) and Vogt ( 1 949a). The former gave an annotated record of the results of two highly successful col- lecting trips to the Lower Valley region in 1930 and 1932, while the Vogt paper provided accu- rate host and habitat information for 83 ceram- bycid species. Linsley and Martin estimated that their list of 65 species brought the southern Texas total to 88, and their figure, combined with Vogt's account, boosted the total to approximately 100 species. Given the relatively small geographical area covered, the collection methods available to these workers, and the fact that only Vogt collected in the fall, this is a most remarkable figure. For the present study, seven cerambycid col- lecting trips were made to southern Texas be- tween 1972 and 1980, concentrating upon the spring and fall activity periods. Dates and col- lectors include the following: 12-18 May 1972, F. T. Hovore (FTH), E. F. Giesbert (EFG); 5-15 October 1975, FTH, EFG, R. L. Penrose (RLP); 2-6 May 1976, FTH, RLP; 9-19 May 1977, FTH, EFG, RLP; 10-13 May 1978, FTH and family; 21-28 October 1978, FTH, RLP; 10-16 May 1980, FTH, RLP, D. C. Carlson. The results of the individual surveys varied considerably due to the vagaries of weather, methodologies, lo- calities visited, and the length of each stay. In total, 136 species of Cerambycidae were collect- ed. GEOGRAPHIC BOUNDARIES OF THE STUDY AREA In order to reflect the ecological limits of the semitropical elements of the Texas cerambycid fauna, our list encompasses a slightly greater geo- graphical area than did prior accounts. Specimen data indicate that the northernmost limits of the true semitropical fauna in Texas extend into the Nueces River drainage near Corpus Christi along the southeastern coastal strand, and northwest up the Rio Grande valley to the vicinity of Eagle Pass. These distributional limits correspond closely with the general parameters of the semi- tropical flora and fauna as expressed by Schwarz (1888, citing C. S. Sargent, Report of Forest Trees of North America). The Nueces River also marks the southernmost region of general distribution of the floral and faunal elements of the eastern woodlands, with the ranges of a number of wide- spread North American tree species extending south to the Corpus Christi-Kingsville area. Blair (1950), in discussing and redefining the concepts of biotic provinces in Texas, considered the re- gion south of the Balcones Escarpment (below the Edwards Plateau) on the west, and the line between pedocal and pedalfer soils on the east (roughly corresponding to the drainage basin of the Nueces River), to comprise the Texan portion of the Tamaulipan Biotic Province. Within this province he later united the extreme southern counties (Starr, Hidalgo, Willacy, and Cameron), plus portions of adjacent Tamaulipas, into the Matamoran Biotic District, with the remainder regarded as the Nuecian District (Blair 1952). Our study region (Fig. 1) more or less corre- sponds to Blair's limits for the Texan portions of the Tamaulipan Biotic Province, although available records do not extend as far north along 286 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 • AUSTIN WELDER WILDLIFE REFUGE BROWNSVILLE FIGURE 1. Southern Texas: Dotted line indicates the ap- proximate boundaries of the study area; solid line indicates the northern limits of the Tamaulipan Biotic Province. the Rio Grande. Our data citations include spec- imens from that portion of Texas south of Eagle Pass on the western border, and Copano Bay Gust north of Corpus Christi) on the Gulf Coast. Most data, however, come from material gathered in the drainage basin of the Rio Grande from Fal- con Lake (Zapata County) to Boca Chica (Cam- eron County); the Southmost sector of Browns- ville; and from Lake Corpus Christi State Park and Welder Wildlife Refuge (both in San Patricio County). Much of the habitat within the re- mainder of the study area is dry upland Tamau- lipan thorn-scrub (Texas Chaparral), mixed overstory brush savanna (overgrazed potential grassland), or cultivated land. Cerambycid species diversity is relatively low in these areas, and little collecting has been conducted beyond cursory beating and sweeping. CLIMATE AND TOPOGRAPHY OF STUDY AREA The lower Rio Grande valley is comprised largely of a deltaic plain, often quite narrow, ex- tending from extreme southern Starr County through southern Hidalgo County and expanding to encompass all of Cameron County, southern Willacy County, and portions of adjacent Ta- maulipas. The inland portions of southern Texas are of diverse geologic origins, with a patchwork of soil types and subsurface formations of vary- ing depths and ages, many of which exert direct controlling influences upon surface vegetational types. Most of the older sandstone formations of the upland portions of the Lower Valley are cov- ered with brushlands or mesquite/huisache sa- vannas, while the terrace deposits along the river valley proper support most of the substantial gal- lery forests. The deltaic plain is also covered with brushy plant formations, but they are denser and more luxuriant than those found in the drier up- lands. Near the Gulf Coast, the plain gives way to open salt marshes and low, Fwcoz-dominated ridges, while to the north along the coast and inland there are deep, wind-blown sand deposits covered with prairie grass and scattered oak com- munities. For more complete discussions of the geology, soils, and vegetational characteristics of southern Texas, see: Coffey (1909); Hawker et al. (1925); Sellards et al. (1932); Trowbridge (1932); Clover (1937); Wynd ( 1 9440, b)\ LeBlanc (1958); Box (1961); Thompson et al. (1972); and Williams etal. (1977). The climate of southern Texas is generally rather mild, with warm dry summers and mod- erately cool winters. Winter frosts are not un- common, but rarely last more than a few days; temperatures usually remain above -4°C. Ac- cording to Clover (1937:42), "Killing frosts are rare, but frequent enough to make commercial growing of bananas and other tropical fruits im- possible." These periodic hard frosts might also be a primary constraint upon the northward ad- vance of the semitropical flora and fauna, and more severe winters undoubtedly result in tem- porary dieback of more cold-sensitive organ- isms, along with high rates of mortality among winter-active species. The record low tempera- ture for Brownsville is — 1 1°C, while tempera- tures at Welder Wildlife Refuge have gone as low as -12.7°C (Box and Chamrad 1966). Precipi- tation may occur during any month, with max- ima in April-May and September. According to Porter ( 1 977:30), ". . . there is actually great vari- ation from month to month and from year to year. Protracted droughts are common but some years may have more than 1,000 mm of rain." March is the driest month (precipitation averages 26 mm at Brownsville), and September the wet- test (precipitation averages 124.8 mm), with an- nual precipitation averages of 669 mm at Brownsville and about 800 mm at Welder Wild- life Refuge. Sudden, violent thunderstorms, common during spring and fall, drop several HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 287 hundred millimeters of rain in a few hours. The entire southern portion of the state is subject to occasional hurricane activity in late summer and early fall, with accompanying high winds, tor- rential rain, and coastal lowland inundation. Se- lected climatological studies of the Lower Valley include: Gunter and Hildebrand (1951); Had- dock (1963); Carr (1967); Orton et al. (1967). HABITAT INFORMATION— HISTORICAL PERSPECTIVES Agriculture and other forms of land develop- ment have more or less confined modern col- lecting in the lower Rio Grande valley to parks and sanctuaries. The remnants of the native flo- ral communities are largely restricted to pre- serves, many of which are small and totally sur- rounded by developed land. Isolated stands of trees and brush do persist along resacas (old river channels) and roadsides, and these often contain surprisingly dense populations of insects. We cannot determine accurately how many, or what percentage, of the original floral and faunal com- ponents of the Lower Valley have been lost to land conversion, as much of the region (the river delta in particular) was cleared for agriculture by the late 1800s. Most of the original native vege- tation was removed between 1 880 and 1 930, with little regard (and sometimes open contempt) for the unique ecosystems destroyed in the process. The typically narrow perspective of Lower Val- ley pioneers was typified by Kerbey (1939:52) in National Geographic Magazine Westward-faring pioneers in the early days of the United States had to chop down forests of sizeable trees to earn their land. . . . Here on the semitropical frontier there was work again, and lots of it, before the land could be put to use. ... I was amazed when I saw for the first time the dense tangle of virgin growth that still covers parts of the region. The sight of it gave me a healthy respect for the early comers who had imagination and energy enough to peel off this ugly and tenacious "rind" to get to the rich, productive earth beneath. . . . Clearing land in the Rio Grande Delta, and thereby transmitting vir- tually worthless wilderness areas into valuable farms . . . has been a slow and expensive process. Since a little after the turn of the century, about 450,000 acres have been cleared ... and between 50,000 and 100,000 acres of good irrigable land still remain to be cleared. The earliest accounts of the coleopterous fauna of the Lower Valley referred to the native tropical forests as occurring in small, isolated "islands" or "little jungles" (Wickham 1897:97). E. A. Schwarz (1896:3) remarked that, "the Texan semi-tropical flora and fauna are doomed to al- most complete extinction by the progress of ag- riculture, and already at the time of my visit, flourishing sugar-cane fields and corn-fields cov- ered the major part of the area once occupied by the semi-tropical forest." Thirty-three years lat- er, H. F. Schwarz (1929:426) sounded a similar, if less ominous, note, "The region will still con- tinue verdant and attractive, but it will be with the blossom of citrus growth and other market products, and less and less with the bloom of the cactus, the huisache, and the Mexican mahog- any. Let us hope that amid all the changes . . . representative groups of wild life may succeed in surviving, even if in diminished numbers." Linsley and Martin (1933:178) noted that by the early 1930s the spread of land conversion had reduced the original habitat to "half a dozen such thickets . . . and few of these are more than an acre or two in size," and that the Sabal Palm grove was being used as a public picnic ground, "where one may collect upon payment of the twenty-five cent admission price." In describing his 1946 and 1947 collecting localities, Vogt ( 1 949b) noted that portions of several floral as- sociations were then being cleared for agriculture and stated that practically all of the land north of the alluvial plain and delta region of the Lower Valley was under cultivation. In 1977, Vogt (pers. comm. to F. T. Hovore) further remarked that, "aside from Santa Ana Refuge and Bentsen (Rio Grande Valley State Park), almost no natural areas remain. Also weed cover has changed, ap- parently due to invasion of more exotic species. Even in the hills of Starr County farming and pasture improvement with exotic (South Afri- can) grasses has changed the ecology exten- sively." He concluded that, "Since many of the vegetation formations I studied have vanished, I would expect a complete faunal change in the Cerambycidae." Neck (1980), discussing the in- vertebrate fauna of the Lower Valley, stated, "There can be no doubt that the invertebrate fauna has been devastated by massive land clear- ing. However, a healthy fauna can be found in remnant tracts of brush. . . . There is no room for complacency, however; all remaining native brush tracts on the left bank should be preserved. As bleak as the situation is on the left (American) bank of the Rio Grande, native brush tracts are almost non-existent on the right (Mexican) bank." 288 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 ECOLOGICAL CONSIDERATIONS: THE LOWER VALLEY REGION At present, agriculture has overwhelmed all but a few stands of native forest and scrub brush. In the delta of the Lower Valley, most of the remaining native flora lies along the deltaic plain and on the periodically inundated low coastal strand between Brownsville and Boca Chica beach. The National Audubon Society's Palm Grove Sanctuary at Southmost sector (near Brownsville), Resaca de la Palma State Park near Brownsville, Santa Ana National Wildlife Ref- uge near Alamo, and Anzalduas Park and Bent- sen-Rio Grande Valley State Park, both near Mission, contain the major portions of the re- maining Lower Valley hardwood forests. Al- though theoretically protected from further en- vironmental destruction, these preserves are nevertheless subject to considerable unnatural stress from such factors as drift and seepage from application of agricultural chemicals on adjoin- ing fields, irregularly fluctuating water tables (af- fected by irrigation and controlled river flow), and even the format of the environmental pro- tection itself. Anzalduas and Bentsen parks are managed to varying extents for recreational uses (picnicking, camping, sports, etc.), in some cases with regular chemical and mechanical vegetation control. Ce- rambycid collecting in the parks has frequently been well below our expectations (based upon observable floral elements and subjective as- sessment of potential), and in recent years, de- spite excellent collecting at other nearby locali- ties, our results from park areas have been relatively poor. It may be that the ecological in- tegrity of these communities has been dimin- ished by continuously manicuring the vegetation in the natural areas. Efforts have recently been made by the Texas Parks and Wildlife Depart- ment to guard against ecosystem decline, and remedial measures (e.g., the prohibition of burn- ing or removing dead wood, and restrictions on the use of topical pesticides) have been instituted in state parks and preserves. Another factor that may contribute to the slow decline of faunas in relictual habitats, and one that would be most difficult to mitigate at this late date, is ecological isolation. Parks and pre- serves are separated from one another by broad zones of radically altered habitat, so there is little or no genetic exchange between populations of organisms with limited mobility. Sanctuaries (where there is less vegetation removal and al- teration) appear to be ecologically healthier than parks but are also geographically isolated and are gradually declining. Recent studies of avifaunal regimes in ecological "islands" among the rem- nant woodland tracts of the eastern United States (MacClintock et al. 1977; Whitcomb 1977; Sim- berloff 1978) concluded that regional extinctions of Neotropical migrants would occur in habitats which were either too small (minimum size based upon an aggregate of territorial, trophic, and oth- er needs) or lacking the necessary biotic diversity. Similar studies involving birds and mammals in tropical ecosystems and "habitat islands" (Ter- borgh and Winter 1980; Wilcox 1980) predicted variable rates of population decline and extinc- tions based upon general and species-specific cri- teria, but overall these studies concluded that rates and percentages of species extinctions in- crease exponentially as habitat size decreases. Because they are relatively small, reproduce rapidly, and utilize minimal amounts of host ma- terial, insects are less vulnerable to some popu- lation pressures, particularly stochastic popula- tion death resulting from diminished territory size and insufficient gene pool size. We know of no detailed studies on habitat requirements for maintenance of population viability in Ceram- bycidae; but overall population vigor probably relates, in part, to the general biotic condition of the ecosystem and, more specifically, to the di- versity, abundance and serai status of the woody plants. Overstoried and senescent communities with decreasing floristic diversity, or with heavy invasion of exotic species, would be expected to lose some more narrowly specialized phytoph- agous insect species. Most sanctuary areas in the Lower Valley are old-growth Tamaulipan interior swamp and ri- parian hardwood forest, Tamaulipan semidecid- uous forest, or overstoried Tamaulipan thorn- scrub (plant formation terminology adapted from Brown et al. 1980), dominated by a mature cli- max sere; all appear to be losing floristic diversity to senescence. Few of the requisite cyclical and successional processes of growth and decompo- sition occur at natural rates. Periodic flooding no longer occurs because of artificial levee systems. Fire, essential to vegetational succession in many plant communities, particularly in arid or semi- arid regions, is suppressed within parks and pre- serves; their small size and lack of adjacent re- HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 289 fugia for wildlife practically precludes the use of fire to stimulate new growth and increase species diversity. The preserves are, in fact, extremely vulnerable to some of the very factors which once served to keep them vigorous and dynamic. Fire, flood, drought, and severe frosts, which together historically combined to influence the compo- sition and distribution of much of the Texas semi- tropical habitat, could alone or in combination eliminate large numbers of species, with virtually no hope of any natural recolonization. Simberloff (1978:10) discussed the probabilities of species extinction in preserves, and noted that an inter- connecting system of refuges would be a "bet- hedging strategy against catastrophes such as fire or epidemic diseases." Such a refuge network would no doubt also provide broader avenues of genetic exchange. Natural habitats in southern Texas have been so drastically reduced, separated, and altered by human activity that many unique ecosystems have already been lost or radically reduced. And, despite the fact that most remnant forests of the Lower Valley are now in some type of preserve, urban and agricultural pressures on undeveloped land continue to mount. At this writing, the Low- er Valley region has the most rapid rate of pop- ulation growth in Texas. It seems then, that de- spite noble (if belated) attempts to preserve intact representatives of the original biota of the lower Rio Grande valley, the sad predictions of E. A. and H. F. Schwarz will at last be realized. THE UPLAND REGION Vast tracts of Tamaulipan thorn-scrub (chap- arral) and a variety of savanna-woodland plant formations remain in the upper Rio Grande val- ley and northern portions of the study area, where it is still possible to find limited areas of more or less undisturbed habitat. Botanists (Clover 1937; Johnston 1963; Inglis 1964) have indicat- ed that present upland chaparral regions are much more extensive now than they were prior to the introduction of livestock. Recent grazing has vis- ibly altered formations and spatial relationships of many brushland plants, and in many areas exotic grasses and disturbed land-favoring gen- era of Compositae (= Asteraceae) grow in dense formations surrounding native trees and shrubs. Although not as species-productive as the Lower Valley habitats, xeric upland communi- ties nevertheless have strong representations of certain cerambycid tribes (e.g., Purpuricenini, Acanthocinini), particularly genera associated with either herbaceous rangeland shrubs or the dominant leguminous tree species, mesquite (Prosopis glandulosa) and huisache (Acacia far- nesiand). Interestingly, beetles in the purpurice- nine genera Tylosis, Lophalia, Parevander, and Crossidius, adults of which are found on fall- blooming herbaceous or woody subshrubs (Abu- tilon, Haplopappus, Viguiera, Verbesina, and Helianthus), are apparently increasing in distri- bution and relative abundance. Of these ceram- bycids, only Tylosis had been previously record- ed from the study area; Vogt (1949#) reported the presence of Tylosis, in the only other paper with records from the fall season. He encoun- tered Tylosis in only two localities, despite the fact that he spent considerable time collecting from flowers in areas where these beetles are now very abundant. Vogt is a most capable and ob- servant entomologist/collector, and it is improb- able that he would have overlooked these large, brightly colored cerambycids. Their absence from previous accounts is more likely either a reflec- tion of their recent advance (along with their hosts) into now-suitable disturbed land habitats, or an artifact of some sort of environmental phe- nomenon. Cyclical population fluctuations of more "tropical" species may occur as a result of unusual pluvial cycles, and temporary popula- tion retreat or dieback may follow repeated freez- es. The plants with which these cerambycid gen- era are associated, either as adult food sources or as larval hosts (only Tylosis and Crossidius have actually been reared), are primarily "weedy" forms that are sensitive to environmental changes, quick to invade disturbed substrates, and coin- cidentally nurtured by agriculture. A converse effect of accelerating land conver- sion and the attendant increase in herbaceous vegetation is the reduction or elimination of tree species, and this is nowhere more evident than in southern Texas. Habitat and host plant re- duction may lead to decline and extinction in associated insects, with oligophagous species most vulnerable. Recent rearings of southern Texas Cerambycidae (Hovore and Giesbert 1976; Ho- vore et al. 1978; Turnbow and Wappes 1978, 1981; Hovore and Penrose 1982) have, fortu- nately, indicated considerable polyphagy in a number of deadwood-boring species (see Select- ed Rearings from Deadwood). Dean (1953) and Manley and French (1976) 290 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 FIGURE 2. Lower Rio Grande valley collecting localities: 1) Bentsen-Rio Grande Valley State Park, 2) Santa Ana National Wildlife Refuge, 3) Audubon Society Palm Grove Sanctuary, 4) 1 6 km west of Boca Chica. recorded rearing 1 8 species of Cerambycidae (and one species each of Bostrichidae and Bupresti- dae) from Citrus grown in the Lower Valley. Dil- lon and Dillon (1946) additionally list Oncideres pustulatus as having been taken on (but not nec- essarily infesting) Citrus. The breadth of host plant preferences shown by these rearings suggests that the net effect of tree species elimination may be mitigated by the abilities of entomofaunas to uti- lize alternative or introduced hosts, including or- namental and agricultural plants. Fox and Mor- row (1981:889) stated that host plant selection "may have a strong genetic basis, controlled by only one locus or polygenic region, so that shifts in preferences for particular host plants can be very rapid." The potential genetic significance of colonization of new or introduced hosts by insect populations was hypothesized by Mayr (1954); populations on new hosts may quickly begin to function as sibling species, morphologically, re- productively (and therefore genetically) isolated from the parent populations (Fox and Morrow 1981). Based upon experiments with Drosophila, Templeton (1979) concluded that colonizing a new genetic environment can quickly lead to unique and possibly isolating changes in the mor- phology, ontogeny, physiology, and behavior of a species. Shifts in host plants might in turn lead to discernible phenotypic differences between original, naturally occurring cerambycid popu- lations and new-host pioneering populations. Should such changes occur in Cerambycidae in southern Texas, we will be better able to detect and quantify them if we take care to preserve adequate voucher samples of all species from all native hosts. DESCRIPTIONS OF COLLECTING LOCALITIES (Figures 1 and 2) Audubon Society Palm Grove Sanctuary, South- most sector, Brownsville, Cameron County Clover (1937) catalogued the floral compo- nents of this unique remnant of the original trop- ical palmetto forests, listing the plant species en- countered in a transect line from the margin of the Rio Grande River into the densest stand of palms. A later survey of the grove was conducted by Da vis (1942). The 425-ha sanctuary contains a subriparian gallery forest of mixed hardwood trees (Ulmus, Celtis, Leucaena, Pithecellobium, etc.) and ma- ture Sabal texana, with elements of semidecid- uous forest, festooned along the margins with vines of Clematis, Serjania, Melothria, and Cis- sus. Beneath the tree canopy, understory vege- HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 291 tation is limited by shading, excessive dampness, and a rank layer of rotting fronds. Openings in the forest (most recently created by limited brush clearing and a small fire) exhibit luxuriant and varied herbaceous growth, interspersed with Baccharis and seedling Leucaena. For the few years that these clearings remain (before the Leu- caena and other tree species shade out the her- baceous vegetation) they will provide excellent insect collecting. Unfortunately, a bamboolike grass (Arundo donax) has invaded the sanctuary along the southern and northern margins and appears to be spreading along paths and service roads. A large stand of Salix grows along the margins of the sanctuary's crescent-shaped resaca, and as one moves away from the lower portion of the intermittent pond, there is a narrow row of large Celtis and Sapindus between the Salix and the adjoining cultivated lands. Another row of Cel- tis, Ulmus, Fraxinus, and herbaceous plants grows along the levee of the Rio Grande, ex- tending about 2 km out from the main portion of the grove. Along the north margin of the re- saca, and well above the water table, there is an extensive stand ofProsopis, Condalia, Zizyphus, and Celtis. According to a recent publicity note from the Audubon Society (Line 1978), long range plans call for expansion of the palm grove to its former size, and restoring the native shrubs that were cleared for agriculture. In 1980 the Nature Con- servancy conveyed about 900 ha of palm jungle habitat to the U.S. Fish and Wildlife Service for inclusion in a future refuge, the "Boscaje de la Palma" (The Nature Conservancy 1981). Collecting in the sanctuary has been remark- ably productive, yielding over 75 species of Cer- ambycidae, including several as yet unknown outside the grove area. 16 km west of Boca Chica on Rt. 4, Cameron County This locality consists of a few acres of open brushland atop a loma, or clay dune, along the highway leading from Brownsville to Boca Chica beach. Dominant plants include Zizyphus, Bac- charis, Karwinskia, Yucca, Opuntia, and Hap- lopappus, with a few scattered Prosopis and Aca- cia. A detailed study of the salt-flat-clay-dune coastal lowland area was presented by Johnston (1952). Collecting at this site was particularly fruitful in the fall, yielding over 20 species of longhorned beetles. Santa Ana National Wildlife Refuge, 9.7 km south of Alamo, Hidalgo County Santa Ana National Wildlife Refuge consists of approximately 1 2,000 noncontiguous hectares of Tamaulipan semideciduous brush and mature forests along the Rio Grande. Significant areas of vegetation were altered during the construc- tion of several artificial intermittent ponds as waterfowl enhancement projects. Fleetwood (1973) presented detailed information regarding plant formations on the refuge. We collected near the refuge headquarters in the 5,200-ha Santa Ana tract, in an overmature Pithecellobium/ Cel- tis/Ulmus forest. Most material was attracted to lights placed along the wildlife drive or beaten from slash and downhanging branches along the west margin of Willow Lake. Bentsen-Rio Grande Valley State Park, 4.8 km west of Mission, Hidalgo County This 1,440-ha state park contains dense for- mations of most of the major native plant com- munities of the Lower Valley, including Tamau- lipan semideciduous forest, Tamaulipan interior swamp, riparian hardwood forest, and mature thorn scrub. Portions of the present protected area were substantially altered by human use prior to 1953, and a major section of the resaca bank is maintained as a grass-lawn picnic and camping area. Collecting techniques included ultraviolet and mercury vapor lights in the camping and picnic areas, beating and sweeping along roads and trails, and searching slash piles at night. Species totals were excellent in 1972 but decreased in succes- sive visits, most notably in light-collected ma- terial. Nevertheless, a number of species taken during our survey are known only from the park. Rob and Bessie Welder Wildlife Foundation Ref- uge, 11.3 km north of Sinton, San Patricio County Welder Wildlife Refuge is managed in part as a working cattle ranch and experimental range. According to Box and Chamrad ( 1 966), the prop- erty has been grazed for more than a century but has never undergone formal cultivation. There are 1 6 recognized plant formations on the prop- 292 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 FIGURE 3. Male Oncideres pustulatus on girdled branch of tepehuaje (Leucaena pulverulenta). erty, most of which are open-range habitats. Communities on arid sites are generally char- acterized by mixed grasses, cactus, and rangeland shrubs, while on more mesic sites riparian, semi- aquatic, or aquatic plants predominate. Some grassland communities are interspersed with thickets of leguminous trees, and most are bor- dered by large tracts of almost impenetrable brush. We conducted most of our collecting in four habitats: the dense hackberry/anaqua and wood- land/spiny aster communities at the extreme eastern edge of the refuge; the chaparral/bristle- grass community along the railroad right-of-way south of the main refuge entrance; and the live oak/chaparral community adjacent to the head- quarters buildings (community terminology after Box and Chamrad 1966). Lake Corpus Christ i State Park, 8 km southwest ofMathis, San Patricia County Dominant vegetation formations around Lake Corpus Christi are upland Tamaulipan semide- ciduous forest and thorn scrub, with minor in- fluences from the more northern Balconian Biot- ic Province. Scrub communities are characterized by a mixture of Condalia, Zanthoxylum, Dios- pyros, Leucophyllum, and Yucca, with scattered invasions of Prosopis. Drainages are wooded mainly with Ulmus, Celtus, and Ehretia. Original bottomland communities were inundated when the Nueces River was impounded in the 1930s to form the lake. FIGURE 4. Male Lochmaeocles cornuticeps cornuticeps on tepehuaje. Larval frass may be seen protruding from ruptures in the bark. Collecting techniques consisted primarily of beating and sweeping roadside vegetation, searching slash piles at night, and light collecting near the park maintenance area. PHENOLOGY Adult cerambycid activity in southern Texas is distinctly bimodal, the spring and fall peaks coinciding with seasonal patterns of moderate temperatures and increased precipitation. These activity peaks generally agree with those docu- mented by Fuchs and Harding (1976) for ar- thropod predators in the Lower Valley. Although a number of species have been collected through the hot summer months, there is a general hiatus in cerambycid activity during July and August. Most summer records are for nocturnal Sonoran species that are apparently better able to tolerate high temperatures and low humidity. We have not seen enough material from the winter months to draw any meaningful conclusions regarding general activity, but it appears that a few species (such as Placosternus difficilis, Euderces reichei exilis, and Anelaphus spurcus) may be encoun- tered during any month of the year. Adult activity within peak seasons fluctuates, with both species-abundance and rates of move- ment generally increasing in response to rises in temperature and humidity. Periods of extended drought may delay adult emergence. Unseason- ably cool temperatures tend to suppress activity, especially of nocturnal species. Once emergence has occurred, rainfall has no more than a tran- HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 293 sitory effect upon beetle activity, except at night. Light collecting is rarely profitable during or im- mediately after heavy rainfall, although sudden rises in air temperature (and consequently in am- bient humidity and numbers of mosquitoes) can initiate surges of dispersal activity. During the severe drought in 1980, ceram- bycids breeding in living host plants (e.g., Cal- lona in mesquite, Mecas spp. in Compositae) emerged somewhat earlier in spring than normal, while deadwood-feeding species were weeks or months behind normal activity patterns. Ex- amination of a variety of infested wood revealed high rates of larval and pupal mortality of all woodboring insects, and numerous dead adult longhorns were found within their pupal cham- bers. The protracted drought of the summer and fall of 1982 had even more severe effects upon col- lecting, and a half-day's beating in the palm grove yielded only two beetles. Extended dry periods must exert considerable selective pressure upon the insect fauna of the region, affecting the "trop- ical" species most dramatically. INTRODUCTION TO SPECIES ACCOUNTS Distributional ranges in the following accounts were drawn from recent literature and from spec- imen data. Activity periods for species with widespread distributions outside the study area include dates from other localities only where insufficient data were available from southern Texas. Cited larval hosts represent rearing records or reliable immature associations from original lit- erature sources. Some host listings cited from Linsley (\962a, b, 19630, 1964) refer to records of adult collection, and do not represent larval hosts. Host citations that refer to specimen data follow the format of collector and institution ab- breviations in the acknowledgments. Uncredited host citations are from our rearings, recorded for the first time herein. We have attempted to up- date and emend pertinent data citations from older literature, by including discussions of ques- tionable records in the species accounts. Common collecting localities are abbreviated in text as follows: Audubon Society Palm Grove Sanctuary (PG), 1 6 km west of Boca Chica (BC), Bentsen-Rio Grande Valley State Park (BRG), Lake Corpus Christi State Park (LCC), Rob and Bessie Welder Wildlife Foundation Refuge (WWR), Santa Ana National Wildlife Refuge (SAR), Falcon State Park and Falcon Heights (these two localities are contiguous) (FSP). Lo- cality data taken from specimens are cited as given on labels with metric equivalents in brack- ets following mileages. The arrangement of species corresponds to Linsley (\962a, b, 1963a, 1 964) and Linsley and Chemsak (1976, 1985), except that portion of the subfamily Lamiinae not yet treated in the Linsley monograph series, genera and species of which are ordered according to the Checklist of Cerambycidae: the Longhorned Beetles (Chem- sak and Linsley 1982). Literature Cited includes all original species descriptions. See the Linsley monograph series for more complete taxonomic references, generic and species keys, species de- scriptions, and general bionomic information. SPECIES ACCOUNTS Parandrinae Parandra (Archandra) polita Say, 1835:192 RANGE.— Central America to Indiana, Ohio, and northern Florida. ADULT ACTIVITY.— May to July. LARVAL HOSTS.— Fagus, Carya, Liriodendron (Linsley 1962a), Pinus (Chemsak et al. 1980). DISCUSSION. — Snow ( 1 906) recorded collecting this species at Galveston, Galveston County, and Brownsville, Cameron County, and Linsley (1962a, fig. 2) shows a locality near Houston, Harris County. Adults were collected from be- neath bark of decaying trunks of the larval hosts and at lights. Prioninae Archodontes melanopus serrulatus LeConte, 1854a:82 RANGE. — Southwestern U.S. from Texas to Arizona. ADULT ACTIVITY.— June to September. LARVAL HOSTS.— Populus spp., Prosopis (Linsley 1962a), Citrus (Dean 1953). DISCUSSION.— The nominate subspecies bores within root crowns of living or dying Quercus in the southeastern U.S., and oak may also serve as a larval host for serrulatus in the oak-savanna habitats of southcoastal Texas. Adults are at- tracted to lights. NEW LOCALITIES.— Flour Bluff, Nueces County, 26 Septem- ber (TAI); Padre Island. Stenodontes (Orthomallodon) dasytomus dasytomus (Say, 1 824: 326) (Figure 5) 294 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 FIGURE 5. Male (left) and female (right) Stenodontes (Or- thomallodori) dasytomus dasytomus. RANGE.— Southeastern U.S. to eastern Mexico. ADULT ACTIVITY.— April to October. LARVAL HOSTS. — Platanus, Celtis, Quercus, Salix, Liquid- ambar, Bursera, Acer (Linsley 1962a), Carya (Riley 1880), Citrus (Dean 1953), Ulmus. DISCUSSION.— Larvae feed in decaying stumps and logs; adult beetles congregate under loose bark, often retreating into their emergence holes during the day. Pupae and teneral adults were taken from heartwood portions of a rotting Celtis stump in the Palm Grove Sanctuary, Cameron County, in May. Adults are commonly attracted to lights. NEW LOCALITIES.— Anzalduas Park, Hidalgo County; SAR; BRG; WWR; LCC. Derobrachus geminatus LeConte, 1853:233 RANGE. — Southern California to Arizona, Texas, and north- ern Mexico. ADULT ACTIVITY. — May to September. LARVAL HOSTS.— Populus, Quercus, Prosopis (Linsley 1962a), Ulmus, Cercidium, Morus, Citrus (Moore and Little 1967), Vitis (Thomas 1951). DISCUSSION.— This is an upland species, as- sociated with mesquite and paloverde; the larvae feed upon roots of living trees. Vogt (1949a) col- lected six males at lights in Rio Grande City, Starr County, in May, and we took specimens at street lights at Falcon State Park in September. Prionus (Neopolyarthron) imbricornis Linnaeus, 1767a:622 RANGE.— Atlantic states south to Florida and west to Ne- braska and south-central Texas. ADULT ACTIVITY.— March to September. LARVAL HOSTS. — Quercus, Costarica, Pyrus, Vitis, maize, and FIGURE 6. laevicollis. Male (left) and female (right) Rhopalophora a wide variety of hardwoods and herbaceous shrubs (Linsley 1962a). DISCUSSION.— One specimen in the TAI col- lection is labeled as this species (identification not verified) from Kingsville, Kleberg County. Adults are common at lights throughout the species range. Prionus (Antennalia) fissicornis Haldeman, 1845:125 RANGE.— Great Plains east of the Rocky Mountains, from Montana and Minnesota south to Texas. ADULT ACTIVITY.— May to July. LARVAL HOSTS.— Grasses (Linsley 1962a). DISCUSSION.— Linsley (\962a, fig. 16) showed this species from near Corpus Christi, Nueces County, and we have seen specimens from near Austin, Travis County, collected in early May (RWN). Cerambycinae Smodicum cucujiforme (Say, 1826:277) RANGE.— Eastern North America to Florida and Texas. ADULT ACTIVITY.— April to August. LARVAL HOSTS.— Robinia, Carya, Fagus, Celtis, Salix, Pop- ulus (Linsley 19626). DISCUSSION.— Linsley' s (1962&) account of this species included material later described as Smo- dicum texanum Knull ( 1 966), and the record of Salix as a larval host was probably based upon observations of the latter taxon in the Lower Valley (Linsley and Martin 1933). Characters cit- ed by Knull for separating the two species are difficult to interpret in material from southern HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 295 Texas, and it appears that texanum differs from cucujiforme only by its slightly paler coloration, more lightly pubescent dorsum, and minor dif- ferences in antennal proportions. Martins (1975), who examined only two male specimens of tex- anum, suggested that it may be a subspecies of cucujiforme but made no formal status change. Because of the difficulty in denning the taxonom- ic parameters and status of texanum, we are con- sidering as cucujiforme only a single specimen from WWR (RHT), determined by R. H. Turn- bow. All other south Texas specimens of Smo- dicum are referred to texanum pending resolu- tion of the status of the two names. Smodicum texanum Knull, 1966:137 RANGE.— Southern Texas. ADULT ACTIVITY. — March to June. LARVAL HOSTS.— Salix! DISCUSSION.— The taxonomic status of this form is uncertain (see S. cucujiforme, above), and material from southern Texas cannot be placed with certainty. Adults referred herein to texanum were collected at lights in several localities, and other south Texas specimens (presumably tex- anum) were collected from beneath bark ofSalix (Linsley and Martin 1933) and Celtis (Vogt 19490). NEW LOCALITIES. -LCC; PG; SAR. Malacopterus tenellus (Fabricius. 1801:335) RANGE.— Southern California and the southern Great Basin to Texas, Mexico, and Central and South America. ADULT ACTIVITY.— May to October. LARVAL HOSTS.— Salix, Populus, Celtis (Linsley 19626). DISCUSSION.— In southern Arizona, this species was cut from pupal cells in moist, punky trunks of dead willow (FTH, EFG), and J. E. Wappes beat an adult from Celtis foliage in the palm grove. The host record cited from Linsley (1 962b) for Celtis referred to adults taken beneath bark. We collected specimens at lights in May and again in October. NEW LOCALITIES.— BRG; Brownsville, Cameron County. Methia const ricticollis Schaeffer, 1908:351 RANGE.— Southeastern Texas to Mexico. ADULT ACTIVITY.— April, May, and September. LARVAL HOSTS. — Celtis (Turnbow and Wappes 1978), Zan- thoxylum (Turnbow and Wappes 1981). DISCUSSION.— Adults were reared from twigs of dead hackberry and colima and have been taken at lights. NEW LOCALITIES. -BRG, SAR; WWR; FSP. Styloxus fuller! fuller! (Horn, 1880:138) RANGE.— South-central Texas. ADULT ACTIVITY.— July to October. DISCUSSION.— The larval habits of this sub- species are not recorded, but they are probably similar to those of the subspecies / californicus (Fall) which girdles oak twigs. Vogt (19490: 140) recorded collecting "Styloxus sp." from Pharr, Hidalgo County, and stated that it was neither fulleri nor texanus (now considered a synonym offulleri). Linsley (19626), perhaps based upon a reassessment of Vogt's material, recorded ful- leri from Hidalgo County. This beetle is evi- dently most active during the summer and is uncommon in collections. Adults are attracted to lights. NEW LOCALITIES. — LCC; SAR; FSP. Achryson surinamum (Linnaeus, 1767a:632) RANGE.— Southern California, Baja California to Arizona and Texas, Mexico, Central and South America, and the West Indies. ADULT ACTIVITY. — March to November. LARVAL HOSTS.— Aspidosperma, Cercidium, Ficus, Prosopis, Acacia, Schnopsis, Pithecellobium, Ulmus, Celtis, Inga, Nec- tandra, Robinia, Tamarindus, Chlorophora, Brya (Linsley 1 9626), Leucaena (Hovore and Penrose 1 982). DISCUSSION.— This species is very abundant on almost any sort of deadwood at night, and adults are readily attracted to lights. The larvae mine extensively within the dry sapwood and heartwood of branches and trunks of dead host plants. NEW LOCALITIES. — BRG; Anzalduas Park, Hidalgo County; LCC; WWR; SAR. Geropa concolor (LeConte, 1873:176) RANGE.— Southern Texas to southern Mexico. ADULT ACTIVITY. — March to November. LARVAL HOSTS.— Ulmus, Acacia, Mimosa (Linsley 19626), Pithecellobium (Linsley and Martin 1933), Leucaena (Hovore and Penrose 1982). DISCUSSION.— This nondescript species was abundant on two-year dead Acacia trees at Weld- er refuge in May and October. The larval habits are similar to those of Achryson surinamum. NEW LOCALITIES.— SAR; BRG; BC; LCC; Sinton, San Pa- tricio County; 10 mi (ca. 16 km) E. jet. of Hwy. 4 and 1419, Cameron County (RHT). Gracilia minuta (Fabricius, 1781:235) RANGE.— Europe, Africa, introduced into North America. ADULT ACTIVITY. — May to July. LARVAL HOSTS.— Salix, Quercus, Rhamnus, Corylus, Aes- culus, Betula, Ceratonia, Rubus, Rosa (Linsley 1 9626), Citrus (Manley and French 1976). 296 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 DISCUSSION.— This cosmopolitan species is often injurious to wood products. Specimens were reared from Citrus in the Lower Valley (Manley and French 1976). Hypexilis pallida Horn, 1885:173 RANGE.— Southeastern Arizona to western and southern Texas, northern Mexico. ADULT ACTIVITY.— April to July. LARVAL HOSTS.— Salix (Turnbow and Wappes 1981), VI- musl (Hovore et al. 1978). DISCUSSION. — Specimens were beaten and reared from willow, and beaten from elm. Adults are most commonly collected at lights. NEW LOCALITIES.— SAR. Eburia stigmatica Chevrolat, 1834:fasc. 3, no. 60 RANGE.— Southern Texas to central Mexico. ADULT ACTIVITY.— March to May, October. LARVAL HOSTS. — Celtis (Hovore et al. 1978). DISCUSSION.— Adults have been collected from beneath loose bark of Celtis (Vogt 1949a), Salix, and Acacia (Linsley and Martin 1933) and reared from dry limbs of sugar hackberry. At night we collected numerous adults from recently felled hackberry and attracted several specimens to lights. NEW LOCALITIES. — PG; BRG; Anzalduas Park, Hidalgo County; FSP; SAR. Eburia ovicollis LeConte, 1873:180 RANGE.— Northern Mexico to southern Texas. ADULT ACTIVITY.— May to September. LARVAL HOSTS.— Prosopis (Hovore and Giesbert 1976). DISCUSSION. — Townsend (1902) collected adults from dead guava and by beating foliage. Linsley and Martin (1933) took a few specimens on ebony, and the senior author (in Hovore and Giesbert 1976) collected an adult male as it emerged from a branch of dead mesquite. Adults are attracted to lights and are most numerous in early summer. NEW LOCALITIES. — PG; BRG; SAR; Kingsville, Kleberg County; WWR. Eburia mutica LeConte, 1853:233 RANGE.— Central Texas to northern Mexico. ADULT ACTIVITY.— April to June, October. LARVAL HOSTS. — Celtis (Hovore et al. 1978), Citrus (Dean 1953), Prosopis, Pithecellobium (Turnbow and Wappes 1978), Leucaena (Hovore and Penrose 1 982), Ulmus. DISCUSSION.— Previous lists variously record- ed this species as Eburia mutica, E. mutica var. manca LeConte, or E. tumida LeConte. Adults are abundant on dead limbs of the larval hosts at night and come to lights. Numerous specimens were taken from the trunk of a wind-thrown hackberry in Bentsen State Park, and pupae and teneral adults were cut from dead branches of that host. NEW LOCALITIES.— WWR; Sinton, San Patricio County; LCC; PG; SAR; FSP. Eburia haldemani LeConte, 1850:102 RANGE.— Arizona to the southeastern U.S. and Florida, south to northern Mexico. ADULT ACTIVITY. — May to July. LARVAL HOSTS. — Celtis (Rice et al. 1985). DISCUSSION.— Numerous adults were attracted to fermenting molasses bait in western and cen- tral Texas. The host record of Celtis is based upon collections from decayed hackberry in western Texas. Linsley and Martin (1933) took an adult beneath bark of Salix near Brownsville, Cameron County, and Vogt (19490) collected a specimen under bark of Ulmus. It is occasionally attracted to lights. NEW LOCALITIES. — BRG. Tylonotus bimaculatus Haldeman, 1847:38 RANGE.— Eastern North America, south to Florida, south- west to southern Texas, and west to Arizona. ADULT ACTIVITY.— May to August. LARVAL HOSTS.— Fraxinus, Betula, Juglans, Carya, Lirio- dendron, Ulmus, Ligustrum (Linsley 1 962ft). DISCUSSION.— A single specimen of this com- mon eastern species was attracted to light in May at Bentsen-Rio Grande Valley State Park, Hi- dalgo County (FTH). In other parts of the range this species is often abundant on living trees, particularly ash. Mannophorus laetus LeConte, 1854ft:442 RANGE.— Western and southern Texas, northern Mexico. ADULT ACTIVITY. — May, September to November. DISCUSSION.— Adults frequent blossoms of Compositae, especially Helianthus, Viguiera, and Verbesina, but the larval host is unknown. It is an upland species, most commonly encountered along roadsides in thornscrub communities. NEW LOCALITIES.— 1 .5-2 mi [ca. 2.4—3.2 km] E Sullivan City, Starr County; 6-8.5 [ca. 9.7-13.7 km] and 13-14 mi [ca. 21- 22.6 km] E El Sauz, Starr County; Hwy. 755, 2.5 mi [ca. 4 km] NE Jet. 490, Starr County; Sam Fordyce Road, 0.5 mi [ca. 0.8 km] S Hwy. 83, Starr County; 16 mi [ca. 26 km] N, 1 mi [ca. 1 .6 km] W Rio Grande City, Starr County. Taranomis bivittata bivittata (Dupont, 1838:58) RANGE. — New Mexico and Texas to central Mexico. ADULT ACTIVITY.— May, July to November. HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 297 LARVAL HOSTS.— Ficus (Townsend 1902), Acacia (Linsley 1940), Prosopis (Rogers 1977a), "cosabe" (Duffy 1960), Ulmus (Turnbow and Wappes 1 978), Leucaena (Hovore and Penrose 1982). DISCUSSION.— Adults, which are common on new growth of mesquite and on freshly cut Aca- cia, may also be collected from a variety of blos- soms, including Jatropha, Eysenhardtia, Sphaeralcea, and Prosopis. On some early lists, this species appeared in the genus Ischnocnemis. NEW LOCALITIES. — PG; 1 mi [ca. 1.6 km] SE Los Indios, Cameron County; BRG; SAR; 3 mi [ca. 4.8 km] S Pharr, Hidalgo County; 5.3 mi [ca. 8.5 km] E Rio Grande City, Stan- County; Hwy. 281, 1.6 mi [ca. 2.6 km] S 83 BR, Hidalgo County; Hwy. 649, 1-6 mi [ca. 1 .6-9.7 km] N Jet. Rt. 83, Stan- County; 7 mi [ca. 1 1.3 km] SW El Sauz, Starr County; WWR; 3-7 mi [ca. 4.8-1 1.3 km] N Sinton, San Patricio County. Lophalia cyanicollis (Dupont, 1838:59) RANGE.— Arizona to Texas and southern Mexico. ADULT ACTIVITY.— October and November. DISCUSSION.— This species was abundant in October on foliage and blossoms of a variety of herbaceous and woody plants, including Verbesi- na, Karwinskia, and Baccharis. Specimens from Mexico (Sinaloa) are more elongate and may rep- resent a different subspecies. NEW LOCALITIES. — PG; SAR; Pharr, Hidalgo County; Hwy. 4, 6.8-7.2 mi [ca. 1 1-1 1.6 km] E Jet. 1419, Cameron County; BC. Gnaphalodes trachyderoides Thomson, 1860:236 RANGE.— Central America to southern Texas. ADULT ACTIVITY.— April to October. LARVAL HOSTS.— Acacia, Pithecellobium, Prosopis, Parkin- sonia (Linsley 1962ft), Celtis (Hovore and Giesbert 1976), Cit- rus (Manley and French 1976), Ulmus (Turnbow and Wappes 1978), Leucaena (Hovore and Penrose 1982). DISCUSSION. —Adults are very common at night on freshly cut wood, and are readily attracted to lights. Thus far, this species has not been en- countered outside the Lower Valley region. NEW LOCALITIES.— PG; SAR; Hwy. 4, 10 mi [ca. 16 km] E Jet. 1419, Cameron County; Anzalduas Park, Hidalgo County; 2 mi [ca. 3.2 km] S Pharr, Hidalgo County; Brownsville, Cam- eron County. Stenaspis verticalis insignis Casey, 1924:262 RANGE.— Southcentral Texas to northern Mexico. ADULT ACTIVITY.— June to November. DISCUSSION.— Adults of this species, like those of its congener, S. solitaria (Say), are strongly attracted to certain plant exudates. Aggregations of beetles, including many mating pairs, were encountered on stems of Baccharis in San Pa- tricio County in October. Many of the Stenaspis, along with other insects, appeared to be feeding at oozing lesions created by scarab beetles (Co- tinis mutabilis Gory and Percheron). Adults were also found on Baccharis foliage, and on blossoms and foliage of Acacia, Serjania, Clematis, Cissus, Jatropha, Condalia, and Haplopappus. Despite the abundance of adults of this large, red and blue species, the larval habits are unknown. Specimens from central Texas (Comal County) have very little black coloration on the underside and pronotum and represent the typical subspe- cies phenotype, while material from further south shows varying degrees of character intermediacy with the nominate form or the western subspe- cies, arizonicus Casey. NEW LOCALITIES.— 6 mi [ca. 9.7 km] E Eagle Pass, Maverick County; PG; BC; 16 mi [ca. 26 km] N, 9 mi [ca. 14.5 km] W Rio Grande City, Starr County; 3-7 mi [ca. 4.8-1 1.3 km] N Sinton, San Patricio County; WWR. Stenaspis solitaria (Say, 1824:410) RANGE.— Southwestern U.S. to south Texas and northern Mexico, Baja California. ADULT ACTIVITY.— May to October. LARVAL HOSTS.— Prosopis, Acacia (Linsley 1962ft). DISCUSSION.— Vogt (\949a) encountered this species in the uplands in May and June, and the host record for Prosopis (cited above from Lins- ley) was based upon observations of larvae he tentatively assigned to this genus. Pupae and adults of S. solitaria were taken from pupal cells in root crowns of dying Acacia in Arizona and western Texas (FTH). Adults frequent foliage and stems of Acacia, Condalia, and Baccharis in the southwestern U.S. and are very abundant on fo- liage of Melochia in the Cape Region of Baja California. Linsley and Cazier (1962) reported this species feeding upon, and apparently becom- ing intoxicated by, fermenting exudates of Se- necio in Arizona. NEW LOCALITIES. — 3 mi [ca. 4.8 km] W, 5 mi [ca. 8 km] N Roma, Starr County. Callona rimosa Buquet, 1840:142 RANGE.— Central Texas to northern Mexico. ADULT ACTIVITY.— April to June. LARVAL HOSTS.— Prosopis, Acacia (Vogt 1949a). DISCUSSION.— Vogt (19490) discussed the lar- val habits of this bright metallic green species, commenting that adults were rarely collected ex- cept from their pupal chambers in bases of living mesquite and huisache. We took adults from fo- liage of infested mesquite and from foliage of Baccharis and other nonhost shrubs growing amongst the host trees, 2 mi [ca. 3.2 km] S Pharr, 298 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 Hidalgo County. We took several specimens in the Palm Grove Sanctuary, including an adven- titious specimen beaten from hackberry foliage (RLP). NEW LOCALITIES.— La Gloria, Starr County; Kingsville, Kle- berg County; WWR. Knulliana cincta cincta (Drury, 1773:85) RANGE.— Eastern North America to western Texas, south to northern Mexico. ADULT ACTIVITY.— March to October. LAR VAL HOSTS.— Juglans, Carya, Castanea, Quercus, Celtis, Pyrus, Sapindus, Salix (Linsley 1 962b), Prosopis (Hovore and Giesbert 1976), Citrus (Dean 1953), Leucaena (Hovore and Penrose 1982). DISCUSSION.— This widespread species was abundant at night on limbs and trunks of newly felled hackberry, huisache, and tepehuaje in May and October. Adults are occasionally attracted to lights. NEW LOCALITIES. — PG; SAR; BRG; Rio Grande City, Stan- County; FSP; LCC; WWR. Tragidion coquus (Linnaeus, 1758:393) RANGE.— Eastern North America to Arizona, western and southern Texas. ADULT ACTIVITY.— August to November. LARVAL HOSTS. — Quercus (Linsley 19626), Prosopis (Swen- son, 1969). DISCUSSION.— A single male of this variably colored species was collected on blossoms of Haplopappus 6 mi [ca. 9.7 km] E Eagle Pass, Maverick County, in October (FTH). Bat vie suturalis cylindrella Casey, 1893:587 RANGE.— Western and southern Texas. ADULT ACTIVITY. — May to July. DISCUSSION.— Adults of this entirely red sub- species were taken on Opuntia and Helianthus blossoms 2 mi [ca. 3.2 km] S Pharr, Hidalgo County in May, and it is common on roadside flowers throughout the upland portions of south- ern Texas. NEW LOCALITIES. -PG; BRG; 1-5 mi [ca. 1.6-8.1 km] N of the Jet. of Hwy. 35 on Hwy. 83, Webb County; 3 mi [ca. 4.8 km] N Sarita, Kenedy County. Plionoma suturalis (LeConte, 1858a:25) RANGE.— Southern California and northern Baja California to Texas and northern Mexico. ADULT ACTIVITY.— May to July, September to November. LARVAL HOSTS.— Prosopis (Linsley 19626). DISCUSSION.— This species was encountered in the fall on fresh-cut mesquite and huisache, and in early summer on blossoms of leguminous trees. Some earlier lists recorded this species in the genus Sphaenothecus. NEW LOCALITIES. — Brownsville and Los Indies, Cameron County; LCC. Tylosis oculatus LeConte, 1850:9 RANGE. — Western and southern Texas to southern Mexico. ADULT ACTIVITY.— September to November. LARVAL HOSTS. —Abutilonl DISCUSSION.— As is typical of the genus Ty- losis, adults frequent blossoms and foliage of malvaceous plants. Vogt ( 1 949a) collected a small series of adults from roadside and canalbank stands of Abutilon, and we found these insects to be very abundant on a tall, red-flowered mal- low in the Lower Valley. In the uplands, near Eagle Pass, we found large numbers of adults on a yellow-flowered, prostrate species of mallow. Larvae of Tylosis jiminezi Casey bore within roots of dead mallow (Sphaeralcea sp.) in western Tex- as (FTH, RLP), and the habits of T. oculatus are probably similar. NEW LOCALITIES.— PG; SAR; BRG; Mission, Hidalgo Coun- ty; 10 mi [ca. 16 km] SE Los Indies, Cameron County; BC; Sarita, Kenedy County (TAI); Kingsville, Kleberg County; 6 mi [ca. 9.7 km] E Eagle Pass, Maverick County. Crossidius humeralis quadrivittatus Penrose, 1974:251 RANGE.— Southern Texas. ADULT ACTIVITY.— September to November. LARVAL HOSTS.— Haplopappus (sometimes listed as Iso- coma) (Hovore and Giesbert 1976). DISCUSSION.— This subspecies is widespread and abundant on blossoms of its larval host in October. Collections were made at a number of localities along the coastal strand from San Pa- tricio County south into Cameron County. A series from 6 mi [ca. 9.7 km] E Eagle Pass, Mav- erick County suggests that vitiate populations of humeralis may be distributed with the larval host in suitable habitats throughout southern Texas. NEW LOCALITIES.— 10-14 mi [ca. 16-22.6 km] W Boca Chi- ca, Cameron County; Arroyo City, Willacy County; Riviera Beach, Kleberg County; Kingsville, Kleberg County; WWR; Laguna Salada, Brooks County. Crossidius suturalis melanipennis Penrose in Giesbert and Penrose, 1984:62 RANGE.— Coastal portions of southern Texas to extreme northern Mexico. ADULT ACTIVITY.— October to December. LARVAL HOSTS.— Haplopappus drummondi. DISCUSSION.— This highly melanic suturalis phenotype occurs with C. humeralis quadrivit- tatus in coastal habitats, where both may use the same species of host plant (Haplopappus drum- HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 299 mondi [T & G Greene]). Although the two species are microsympatric, they appear to be largely allochronic, peak numbers of C. suturalis me- lanipennis occur in November, when C. humer- alis quadrivittatus activity wanes. Similar tem- poral stratification between the nominate forms of both species was observed in southeastern New Mexico (RLP). NEW LOCALITIES. — Riviera Beach and Kingsville, Kleberg County; Laguna Salada, Brooks County. Crossidius pulchellus LeConte, 1861:356 RANGE. — Alberta, Canada southward through the Great Plains to western Kansas, southern California, southern Texas, and northern Mexico. ADULT ACTIVITY.— August to November. LARVAL HOSTS. — Gutierrezia spp. (Linsley and Chemsak 1961), Gymnosperma (also listed as Xanthocephalum). DISCUSSION.— A small series of a highly me- lanic population was collected in October from blossoms of Gymnosperma glutinosa (Spreng.) Less., 8 mi [ca. 13 km] SE Beeville, Bee County (RLP, FTH). Elytroleptus divisus (LeConte, 1884:23) RANGE. — North-central to southern Texas. ADULT ACTIVITY.— April to July. DISCUSSION.— Vogt (1949a) collected adults from blossoms and foliage of Karwinskia and Condalia in the upland regions of Hidalgo and Starr counties in April and May, and we have likewise found it to be relatively common on those plants. Larval habits of the genus Elytro- leptus are unrecorded. NEW LOCALITIES. — 1 mi [ca. 1.6 km] and 3 mi [ca. 4.8 km] W Roma, Starr County; 3 mi [ca. 4.8 km] N Laredo, Webb County (AEL); 45-55 mi [ca. 73-89 km] E Carrizo Springs, Hwy. 83, in Webb County (AEL). Parevander hovorei Giesbert in Giesbert and Penrose, 1984:59 RANGE. — Southern Texas to central Mexico. ADULT ACTIVITY.— September to December. DISCUSSION.— This orange and black species was previously recorded from southern Texas as P. xanthomelas (Guerin) (Hovore and Giesbert 1976). Earlier faunal accounts did not include this species, and it may be that it has only re- cently colonized, or recolonized, southern Texas. Austin (1880:60) in his North American checklist stated, ". . . add Evander Thorns. 9730 xanthomelas (Guer.)," but Leng (1886) later stat- ed that "Evander" had not been found within our faunal limits, and he dropped it from his checklist. (The name Evander was incorrectly ap- plied to New World species and was later re- placed by Parevander Aurivillius). Parevander is a Neotropical genus, with closely related species distributed into Central America; the adults are associated with disturbed-land plants (larval hosts are unknown). It therefore may be sensitive to long- or short-term environ- mental phenomena, experiencing population fluctuations or extinctions during droughts or frosts. Parevander hovorei was taken in abundance in several previously well collected localities, often with Mannophorus laetus, from several mem- bers of the Compositae, including Viguiera, He- lianthus, and Verbesina. Trachyderes (Dendrobias) mandibularis (Audinet-Serville, 1834:42) RANGE.— Southern Texas and northern Mexico. ADULT ACTIVITY. — March to November. LARVAL HOSTS. — Celt is (Hovore and Giesbert 1976), Leu- caena (Hovore and Penrose 1982), Pithecellobium, Vlmus, Acacia. DISCUSSION.— This species is extremely abun- dant throughout southern Texas, frequenting a variety of blossoms, utilizing numerous types of freshly cut wood for mating and ovipositing ac- tivities, and often aggregating in large numbers on stems of Baccharis. Specimens from the study area are assignable to the subspecies virens Casey, although not all material at hand matches the phenotype char- acterization given by Linsley ( 1 9626). As he not- ed, south Texas specimens vary considerably in coloration; some individuals exhibit an expand- ed elytral pattern similar to that of the nominate subspecies. In our field-collected and reared specimens about 70% of the males show the typ- ical reduced elytral pattern, while the remainder of the males and all the females have a greatly expanded elytral pattern, with the basal and me- dian dark fasciae broadly united along the lateral margins. Thus there are two distinct patterns, each different from that of any other population seen, with a few individuals resembling the light- ly marked males of the nominate taxon. Further, all of our specimens have the third antennal seg- ment wholly black (it is yellow-annulated in typ- ical mandibularis Audinet-Serville from western Texas and southeastern Arizona, and in m. re- ductus Casey, from the lower Colorado River Valley of Arizona and California), and there is distinctive allometric reduction of the develop- 300 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 ment of the male mandibles. The largest male virens (26 mm long) have mandibles approxi- mately 30% smaller than those of comparably sized mandibularis from southeastern Arizona and Mexico and nearly 40% smaller than those of similarly sized reductus. Both virens and reductus appear to represent distinctive local phenotypes, but a study of over- all species variability, including analysis of Mex- ican and Baja Californian material, is needed to resolve their taxonomic status. Chemsak and Linsley (19750, 1982) list reductus and virens as synonyms of mandibularis, but J. A. Chemsak (pers. comm.) informed us that these were ty- pographical errors, not synonymies. Hudepohl (1985) placed Dendrobias as a subgenus of Trachyderes Dalman, and eliminated all subspe- cies of mandibularis. NEW LOCALITIES.— PG; BC; BRG; Pharr and Mission, Hi- dalgo County; LCC; WWR; 3-7 mi [ca. 4.8-1 1 .3 km] N Sinton, San Patricio County. Lissonotus flavocinctus puncticollis Bates, 1885:333 RANGE.— Northern Mexico and Baja California to southern Texas. ADULT ACTIVITY.— April to November. LARVAL HOSTS.— Acacia (Vogt 1949a), Leucaena (JEW). DISCUSSION. — Vogt (1949a) collected this species from goldenrod blossoms and on freshly cut Acacia, and J. E. Wappes reared these insects from tepehuaje collected in the Palm Grove Sanctuary. Adults have also been taken at light (AEL) and in pitfall traps in a cotton field (Huff- man and Harding 1980). As with Dendrobias (discussed above), the described subspecies of Lissonotus flavocinctus are difficult to define geo- graphically, and further study of Mexican pop- ulations is needed to clarify the relationships of the various phenotypes. NEW LOCALITIES. — Pharr, Mercedes, and Mission, Hidalgo County. Psyrassa texana Schaeffer, 19056:160 RANGE.— Southern Texas. ADULT ACTIVITY.— May to August. DISCUSSION.— Linsley and Martin (1933) beat adults from Acacia and attracted them to lights. Several specimens were beaten from Celtis (FTH) and Fraxinus (JEW) in the palm grove. Psyrassa texana is very close to, if not synonymous with, the Mexican species P. castanea Bates. NEW LOCALITIES.— LCC. Psyrassa pertenuis (Casey, 1924:248) RANGE. — Eastern North America from New York to Florida, west to southern Texas. ADULT ACTIVITY.— April to July. LARVAL HOSTS.— Magnolia, Prunus, Carya (Linsley 1963a). DISCUSSION.— Numerous specimens of this common Austroriparian species were collected at lights at Welder Wildlife Refuge in May. Psyrassa brevicornis Linsley, 1934:164 RANGE.— Lower Rio Grande valley and lower Gulf Coast to Kleberg County. ADULT ACTIVITY. — May to September. DISCUSSION. — Linsley (19630) stated that adults were captured on dead branches of Acacia and Pithecellobium, and numerous specimens were taken at lights. NEW LOCALITIES. -BRG; PG; SAR; FSP; Kingsville, Kleberg County (TAI). Psyrassa sallaei Bates, 1885:255 RANGE.— Southern Texas to north-central Mexico. ADULT ACTIVITY.— September to October. DISCUSSION.— The original description of this species is rather general, and may be applied to a number of Mexican species of Psyrassa, some of which are as yet undescribed. We therefore refer south Texan material to sallaei by com- parison with determinations by Linsley (19630) and Vogt (19490, determined by Linsley). Texas specimens were taken at light (JEW) and by beat- ing Sapindus (Vogt 1 9490) and Cordia (Turnbow and Wappes 1978). NEW LOCALITIES. — PG. Stenosphenus notatus (Olivier, 1795:61) RANGE.— Eastern North America to southern Texas. ADULT ACTIVITY.— April to July. LARVAL HOSTS. — Carya, Celtis (Linsley 1963a). DISCUSSION.— J. E. Wappes reared a single specimen of this Alleghenian species from wood of an unidentified legume from Santa Ana Ref- uge. Stenosphenus lugens LeConte, 1862:41 RANGE.— Southern Texas to Mexico. ADULT ACTIVITY.— August to November. LARVAL HOSTS.— Acacia (Linsley 1963a), Celtis (Turnbow and Wappes 1978), Leucaena (Vogt \949a; Hovore and Pen- rose 1982), Zanthoxylum. DISCUSSION.— Although rare in collections, adults of this species were extremely abundant on dead branches of their hosts and on blossoms HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 301 and foliage of Baccharis, Serjania, Cissus, and Clematis. NEW LOCALITIES. -PG; BC; BRG; SAR. Stenosphenus dolosus Horn, 1885:179 RANGE.— Central and southern Texas. ADULT ACTIVITY.— April to June, September to November. LARVAL HOSTS.— Prosopis, Acacia (Linsley 1963a), Leu- caena (Hovore and Penrose 1 982). DISCUSSION.— This beetle is relatively abun- dant in both spring and fall on blossoms and stems of Helianthus (Linsley and Martin 1933), Solidago and Baccharis (Vogt 1 9490), Aster, and Cissus. Snow's record (1 906) for S. novatus Horn (a Baja Californian species) may almost certainly be referred to this species. NEW LOCALITIES: PG; BRG; Rio Grande City, Starr County; 1.4 mi [ca. 2.3 km] SE Carrizo Springs, Dimmit County; 1.8 mi [ca. 3 km] ESE Eagle Pass, Maverick County; Pharr, Hidalgo County; 25 mi [ca. 40 km] S Sarita, Kenedy County; Kingsville, Kleberg County; WWR. Aneflus sonoranus Casey, 1924:241 RANGE.— Southern California to Sonora, Mexico, and south- ern Texas. ADULT ACTIVITY. — May to September. LARVAL HOSTS.— Acacia (WHT). DISCUSSION.— Vogt (19490) collected a single adult in June from decadent Condalia in Stan- County, and R. H. Turnbow took specimens at lights in Bentsen State Park, Hidalgo County, and in Zapata County. W. H. Tyson (pers. comm.) stated that larvae bore within living branches and trunks of catclaw acacia. NEW LOCALITIES. — FSP. Aneflus prolixus insoletus Chemsak and Linsley, 1963:88 RANGE.— Southern Texas to east-central Mexico. ADULT ACTIVITY.— May to September. LARVAL HOSTS.— Acacia (Rice et al. 1985). DISCUSSION.— Turnbow and Wappes (1978) recorded collecting adults at lights and from slash piles in September. The larvae breed in living roots and stem bases of Acacia berlandieri (Rice etal. 1985). Aneflus protensus protensus (LeConte, 18586:82) RANGE. — Southeastern Arizona to Baja California, northern Mexico and southern Texas. ADULT ACTIVITY.— June to September. LARVAL HOSTS.— Prosopis (Linsley 1963a). DISCUSSION.— Vogt (1949a) collected adults from dead mesquite branches in Starr County in June and July. In Arizona this species is com- monly attracted to lights. NEW LOCALITIES.— FSP. Aneflomorpha tenuis (LeConte, 1854a:81) RANGES.— Southwestern Texas to northern Mexico. ADULT ACTIVITY. — May to September. DISCUSSION.— Adults have been taken on Aca- cia (Linsley and Martin 1933), on blossoms of Karwinskia (Turnbow and Wappes 1981), and at lights. NEW LOCALITIES. -FSP; BRG; LCC; WWR, SAR. Aneflomorpha seminuda Casey, 1912:294 RANGE. — Western to southern Texas. ADULT ACTIVITY.— April to July. DISCUSSION.— This nocturnal longhorn, which is not uncommon at lights in western Texas, was recently recorded from the Lower Valley region (Turnbow and Wappes 1978). NEW LOCALITIES. -BRG (AEL). Aneflomorpha opacicornis Linsley, 19576:285 RANGE.— Western to southern Texas. ADULT ACTIVITY.— July to September. DISCUSSION.— Specimens tentatively assigned to this species were collected at lights in Falcon Heights, Zapata County, in September (RHT, JEW). Axestinus obscurus LeConte, 1873:177 RANGE. — Southeasten Arizona to western and southern Tex- as and northern Mexico. ADULT ACTIVITY.— May to July. DISCUSSION.— Although we have not seen any south Texas specimens of this Sonoran elaphi- diine, we include it herein by the type locality: "Rio Grande Valley?" (fide Linsley 19630). Adults are common at lights in western Texas, southern New Mexico, and southeastern Arizo- na, but larval habits are unknown. Specimens in the University of California, Berkeley, collection are from "La Gloria, south of Monclova," Coa- huila, Mexico, approximately 200 km southwest of Laredo, Webb County. Sphaerion exutum (Newman, 1841:93) RANGE.— Argentina and Brazil to southern Mexico and southern Texas (based upon records from Blackwelder [ 1 946] and Linsley [196 la]). ADULT ACTIVITY.— May. DISCUSSION.— This tropical species was first recorded from the U.S. on the basis of eight spec- imens collected on dead ebony at Bentsen-Rio 302 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 Grande Valley State Park in May, 1 972 and May, 1 973 (Giesbert and Hovore 1 976). An additional male specimen was taken at that locality on dead Acacia in May, 1980 (FTH). Enaphalodes hispicornis (Linnaeus, 1767a:634) RANGE.— North America from California to Idaho, Min- nesota, New Jersey, Florida, Texas, and extreme northern Mexico. ADULT ACTIVITY.— June to October. LARVAL HOSTS. — Quercus (Linsley 1963a). DISCUSSION.— Linsley (1963a, fig. 23) showed a locality for this widely distributed species near Corpus Christi, Nueces County. The larval host, oak, occurs in dense formations on the sand- sheets of Kleberg County and sporadically over much of the northern portion of the study area. Enaphalodes taeniatus (LeConte, 1854a:81) RANGE.— Central to southern Texas and extreme northern Mexico. ADULT ACTIVITY.— April to September. LARVAL HOSTS. — Citrus (Dean 1953). DISCUSSION.— This attractive beetle is never particularly common; a few specimens have been taken under loose bark of willow (Linsley and Martin 1933) and at lights. NEW LOCALITIES.— PG; BRG; SAR. Enaphalodes rufulus (Haldeman, 1847:32) RANGE. — Eastern North America from Canada to Florida, western and southern Texas. ADULT ACTIVITY.— June to August. LARVAL HOSTS. — Quercus, Acer (Linsley 1963a). DISCUSSION.— Several specimens ofE. rufulus, the red oak borer, were taken at lights at Welder Wildlife Refuge, and we observed evidence of heavy infestation in oak near the refuge head- quarters. The southern limits of E. rufulus in Texas probably correspond to those of the pri- mary host, oak. Enaphalodes atomarius (Drury, 1773:93) RANGE.— Eastern North America from Canada to Florida, west to Texas, Arizona, and Central America. ADULT ACTIVITY. — May to September. LARVAL HOSTS. — Quercus, Castanea, Celtis, Juglans, Carya, Chamaerops (Linsley 1963a). DISCUSSION.— We have not seen any speci- mens from the study area, but Linsley (1963a, fig. 26) showed the species as occurring in the Lower Valley, and Townsend (1902) reported taking a specimen in a mail sack from Alice, Jim Wells County. Chemsak et al. (1980) recently recorded specimens from Honduras, and there is a specimen in the TAI collection (determi- nation not verified) of either this species or its cryptic sibling, E. cortiphagus (Craighead), from Welder Wildlife Refuge. Eustromula validum (LeConte, 18586:82) RANGE.— Southern California to southern Texas, northern Mexico and Baja California. ADULT ACTIVITY. — May to August. LARVAL HOSTS.— Prosopis, Cercidium, Parkinsonia (Linsley 1963a), Salix (Hovore and Giesbert 1976). DISCUSSION. — This nondescript species is commonly attracted to lights in the desert regions of the American southwest. Vogt (\949a) took a single specimen at a light in Starr County in May. Elaphidion linsleyi Knull, 1960:7 RANGE.— Western to southern Texas. ADULT ACTIVITY. — March to October. LARVAL HOSTS.— Salix (Turnbow and Wappes 1981), Bac- charis, Ungnadia (Rice et al. 1985). DISCUSSION.— The presence of a somewhat in- termediate condition in the development of the femoral spines in southern Texas material, par- ticularly female specimens, suggests that this tax- on may only be a western subspecies of the wide- spread E. mucronatum (Say). In the south Texas hypodigm, femoral spines range from short and rounded to prolonged and acute, but they are never as pronounced as in typical mucronatum. It has also been suggested (Turnbow and Wappes 1981, based upon analysis of two separate reared series of specimens displaying intergrading char- acters) that linsleyi may be hybridizing with E. mimeticum on Salix in the Brownsville (Cam- eron County) area. Vogt's (1949a) record of Elaphidionoides in- certus from willow may be based in part upon specimens of this species, or "linsleyi x mime- ticum" hybrids. Larvae, pupae, and adults of E. linsleyi were cut from injured Baccharis near Del Rio, Val Verde County (FTH, RLP). Elaphidion mimeticum Schaeffer, 19050:132 RANGE. — Southern Texas and extreme northeastern Mexico. ADULT ACTIVITY.— April to October. LARVAL HOSTS.— Salix. DISCUSSION.— Adults of this species hide dur- ing the day beneath loose bark of willow, acacia, hackberry, and ash (Linsley and Martin 1933) and may be found at night on dead host trees. They have also been taken in molasses bait and have been attracted to lights. Elaphidion mi- meticum was recorded on some earlier lists as HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 303 the very similar-appearing E. irroratum Lin- naeus, an Antillean species. NEW LOCALITIES. -PG; BRG; FSP; WWR; SAR. Elaphidionoides villosus (Fabricius, 1792:302) RANGE. — Eastern North America to Texas and Arizona. ADULT ACTIVITY. — March to July. LARVAL HOSTS. — Quercus, Carya, Castanea, Prunus, Vitis, Abies, Mains, Tilia, Wisteria, Cladrastis, Gleditsia, Celtis, Acer, Juglans, among others (Linsley 1963a), Citrus (Dean 1953). DISCUSSION. — This is a common eastern species, utilizing a broad variety of larval hosts. Recent rearings from Rio Grande valley Citrus provide the only known southern Texas records, and we have not been able to verify the deter- mination. Elaphidionoides incertus (Newman, 1840:28) RANGE. — Eastern North America to southern Texas. ADULT ACTIVITY.— May to September. LARVAL HOSTS.— Moms, Quercus, Carya (Linsley 1963a). DISCUSSION.— Vogt (1949a) recorded the cap- ture of two specimens from beneath bark of wil- low (see discussion of E. linsleyi, above), and we took a few adults, including a mating pair, from dead Celtis at night in Bentsen State Park and Santa Ana Refuge. Adults are attracted to lights and fermenting molasses baits. NEW LOCALITIES. -PG; WWR. Elaphidionoides aspersus (Haldeman, 1847:32) RANGE.— Atlantic states to Iowa and Texas. ADULT ACTIVITY. — May to August. LARVAL HOSTS. — Carya, Quercus (Linsley 1963a). DISCUSSION.— There appear to be more than two species involved in the material examined in the incertus-aspersus species complex, and de- terminations of specimens listed herein follow the concepts of Linsley (1963#). Verified records include a specimen from Brownsville, Cameron County (RWN) and another from Kingsville, Kleberg County (TAI). Anelaphus niveivestitus (Schaeffer, 1905a:132) RANGE. — Southern Texas. ADULT ACTIVITY.— April to July. DISCUSSION.— This diminutive species is com- monly attracted to lights, and has been beaten from branches of ash (Linsley and Martin 1933), hackberry, and tepehuaje. R. H. Turnbow took specimens in fermenting molasses bait in the Palm Grove Sanctuary. NEW LOCALITIES. -BRG; WWR. Anelaphus debilis (LeConte, 1 854/>:442) RANGE.— Central Texas to northeastern Mexico. ADULT ACTIVITY. — March to June, October. LARVAL HOSTS. — Prosopis (Hovore and Giesbert 1 976), Bac- charis, Pithecellobium, Celtis (Turnbow and Wappes 1978), Leucaena (Hovore and Penrose 1982), Acacia. DISCUSSION. —Adults of this species were com- monly collected at lights, at fermenting molasses bait (RHT), and by beating dead branches of larval hosts. Linsley and Martin's (1933) record of"Anoplium truncatum LeConte," and Vogt's (1949a) " 'Anelaphus truncatus (Hald)" probably referred to A. debilis, A. spurcus, or A. inermis. All three are similar in coloration and form and were consistently misidentined in material ex- amined during this study. NEW LOCALITIES. — FSP; Rio Grande City, Starr County; 6- 7 mi [ca. 9.7-11.3 km] NE Roma, Starr County; Zapata, Zapata County; PG; LCC; WWR. Anelaphus spurcus (LeConte, 18546:442) RANGE.— Central Texas to northeastern Mexico. ADULT ACTIVITY.— April to June. DISCUSSION.— Adults were attracted to lights in spring and early summer and were taken from beneath loose bark of dead ebony and tepehuaje. NEW LOCALITIES. — PG; SAR; BRG; Zapata and Lopeno, Za- pata County; LCC; WWR. Anelaphus inermis (Newman, 1840:29) RANGE. — Southeastern U.S. to Texas, West Indies, and Mex- ico. ADULT ACTIVITY.— April to June, September to November. LARVAL HOSTS. — Citrus, Quercus, Carya, Ichyomethia (Linsley 1963a). DISCUSSION.— This widespread species was collected at lights and by beating freshly fallen Yucca trunks in Starr County in May. Hubbard (1885) and Manley and French (1976) reported rearing adults from Citrus. Specimens from the Antillean faunal region differ slightly from Texan and Mexican specimens and may prove to be a separate subspecies. NEW LOCALITIES. — 3 mi [ca. 4.8 km] W, 5 mi [ca. 8 km] N Roma, Starr County; PG; BRG; LCC; WWR; FSP. Anelaphus moestus moestus (LeConte, 18546:442) RANGE.— Western Arizona to Texas and northern Mexico. ADULT ACTIVITY.— April to October. LARVAL HOSTS.— Juglans (Linsley 1963a), Quercus (Hovore and Giesbert 1976), Celtis (Turnbow and Wappes 1978), Rhus (Riceetal. 1985). DISCUSSION. — This beetle is abundant throughout its range, commonly coming to lights and fermenting molasses bait. Vogt (\949a) took 304 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 adults beneath Celtis bark and on fire-killed Opuntia. NEW LOCALITIES. -PG; BRG; 2 mi [ca. 3.2 km] S Pharr, Hidalgo County; FSP. Elaphidionopsis fasciatipennis Linsley, 1936:467 RANGE.— Western and southern Texas to northern Mexico. ADULT ACTIVITY. — May to September. DISCUSSION.— This attractive species is very rare in collections; the few specimens we saw were collected at lights in western Texas. A single specimen was attracted to building lights at Fal- con Heights, Zapata County, in September (RHT). Heterachthes ebenus Newman, 1840:9 RANGE.— Eastern portions of North and South America; Mexico; West Indies. ADULT ACTIVITY.— January to August. LARVAL HOSTS.— Pinus (Craighead 1923). DISCUSSION. — If the larval association with Pi- nus is valid, then other plants must also serve as hosts, since pines are not found over most of the range of this species. Two specimens were taken at Welder Wildlife Refuge, one from dead huis- ache in May (FTH) and one in a UV light trap in August (RHT). Heterachthes nobilis LeConte, 1862:41 RANGE.— Southern Texas. ADULT ACTIVITY.— April to August. LARVAL HOSTS. —Prosopis (Linsley 1 963a; Hovore and Gies- bert 1976). DISCUSSION.— Adults are not common in col- lections, most specimens having been taken at lights or on decadent mesquite. A few adults were reared from fire-killed branches of this host. NEW LOCALITIES. -BRG; 2 mi [ca. 3.2 km] S Pharr, Hidalgo County; Rio Grande City, Starr County; SAR; WWR. Neocompsa exclamationis (Thomson, 1860:201) RANGE. — Southern Texas to Chiapas, Mexico. ADULT ACTIVITY.— April to August. LARVAL HOSTS.— Mimosa (Craighead 1923), Leucaena (Ho- vore and Penrose 1982), Zanthoxylum. DISCUSSION.— This large ibidionine has been taken during the day from rotten branch stubs of Acacia, Mimosa, and Celtis (Linsley and Mar- tin 1933), from beneath loose bark, and on slash of colima and ebony. Adults were common at lights in the palm grove and were also found at night on tepehuaje blossoms. NEW LOCALITIES.— BRG; Anzalduas Park, Hidalgo County. Neocompsa mexicana (Thomson, 1865:573) RANGE.— Southern Texas to Guatemala and Costa Rica. ADULT ACTIVITY. — March to November. LARVAL HOSTS.— Acacia (Craighead 1923), Pithecellobium (Linsley 1 963a), Celtis (Turnbow and Wappes 1978), Leucaena (Hovore and Penrose 1982), Zanthoxylum. DISCUSSION.— This species appeared on pre- vious lists as "Ibidion townsendi Linell," and was also once referred to as Neocompsa hippopsioides (Bates) (Martins and Chemsak 1 966); both names are now considered synonyms of N. mexicana. Large numbers of adults were reared and taken from dead tepehuaje branches in the palm grove. NEW LOCALITIES. -BRG; SAR. Neocompsa intricata Martins, 1970:1088 RANGE.— Eastern Mexico to southern Texas. ADULT ACTIVITY. — May to October. DISCUSSION.— This species was previously col- lected in Texas, but earlier material was recorded as either "Compsa textilis var. alacris Bates" (Linsley and Martin 1933), or "Compsa alacris" (Linsley 1963a). According to Martins (1970), Neocompsa alacris (Bates) is distributed primar- ily along the Pacific slope of Mexico and Central America, and the occurrence of this species in Texas is very doubtful. Linsley (1963a) recorded "Compsa quadriplagiata (LeConte)" (=Neo- compsd) from southern Texas, based upon the type locality (Brownsville, Cameron County) of a junior synonym, Ibidion pubescens Casey. Martins considered the Casey holotype to be mislabelled, and extant distributional data for Neocompsa quadriplagiata restricts it to Baja California Sur and the Pacific slope of Mexico. Single specimens of N. intricata were beaten from Baccharis (RLP) and taken from herba- ceous foliage (JEW) in October. NEW LOCALITIES. -BC; WWR. Neocompsa puncticollis orientalis Martins and Chemsak, 1966: 466 RANGE.— Southern Texas to Oaxaca, Mexico. ADULT ACTIVITY. — May to August. DISCUSSION.— Vogt (1949a) collected a single specimen, tentatively referred to this subspecies by Martins (1970), at lights in Pharr, Hidalgo County, in August. Piezocera serraticollis Linell, 1896:394 RANGE. — Southern Texas, and perhaps also southern Mex- ico to Panama. ADULT ACTIVITY.— April to June. LARVAL HOSTS. — Celtis (Turnbow and Wappes, 1978). HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 305 DISCUSSION.— Specimens of this peculiar bee- tle were beaten from dead branches of Celtis, Pithecellobium, and Prosopis. Martins (1976) suggested that serraticollis and P. monochroa Bates may be conspecific; due to insufficient ma- terial, Martins retained the two taxa as distinct, tentatively assigning several specimens from Mexico and Central America to serraticollis. NEW LOCALITIES. -PG; BRG. Obrium rufulum Gahan, 1908:142 RANGE. — Eastern North America to Texas. ADULT ACTIVITY.— April to July. LARVAL HOSTS. — Fraxinus (Linsley, 1963a). DISCUSSION. — This Alleghenian species is known from the study area by material from Kingsville, Kleberg County, collected in April (TAI). Obrium maculatum (Olivier, 1795:32, 39) RANGE.— North America from eastern Canada to Florida and southern California, south to Costa Rica. ADULT ACTIVITY. — March to October. LARVAL HOSTS. — Carya, Quercus, Castanea, Celtis, Morus, Madura, Cercis, Acacia (Linsley 1963a), Ficus (Townsend 1 902), Citrus (Manley and French 1 976), Leucaena (Vogt 1 949a; Hovore and Penrose 1982), Sapindus (Vogt 1949a). DISCUSSION.— Adults are abundant on dead twigs and branches of the larval hosts, and are also readily attracted to lights. NEW LOCALITIES. — PG; 8 mi [ca. 13 km] SE Zapata, Zapata County; BRG; FSP; LCC; WWR. Obrium mozinnae Linell, 1896:395 RANGE.— Southern Texas to Tamaulipas, Mexico. ADULT ACTIVITY.— April to June. LARVAL HOSTS.— Leucaena (Hovore et al. 1978), Prosopis (Turnbow and Wappes 1978). DISCUSSION.— This tiny, bicolored species is often abundant on blossoms of leguminous trees and shrubs and is also attracted to lights. NEW LOCALITIES. -PG; BRG; Anzalduas Park, Hidalgo County; SAR; La Lomita Park, Hidalgo County; Southmost sector, Brownsville, Cameron County. Nathriobrium methioides Hovore, 1980:1 16 RANGE.— Southern Texas. ADULT ACTIVITY.— November to January. LARVAL HOSTS.— Pithecellobium (Hovore 1980), Diospyros (Turnbow and Wappes 1981), Zanthoxylum (Rice etal. 1985). DISCUSSION.— This unusual, monotypic genus appears most closely related to genera from southern South America. The few known spec- imens, all reared, emerged from ebony, Texas persimmon, and colima. Turnbow and Wappes (198 1) described and figured the larval workings in persimmon. NEW LOCALITIES.— PG. Plinthocoelium suaveolens plicatum (LeConte, 1853:233) RANGE.— Central Texas to Arizona and northern Mexico. ADULT ACTIVITY. — May to August. LARVAL HOSTS.— Bumelia (Linsley 1964). DISCUSSION.— R. H. Turnbow (pers. comm.) reported the collection of a single specimen in a light trap at Welder Wildlife Refuge in August. This is the only light collection record we have seen for the species, but some tropical Calli- chromatini readily come to UV lights. Adults were collected from foliage of the larval host in fermenting baits. Plinthocoelium schwarzi (Fisher, 1914:97) RANGE.— Southern Texas. ADULT ACTIVITY.— March to May. DISCUSSION.— This metallic green species fades postmortem to deep cobalt blue. Adults frequent blossoms of Condalia and Cissus in the upland regions of the Lower Valley, and are strong, swift flyers, making capture quite difficult. When dis- turbed they emit a milky substance described by Vogt (1949#) as having an odor like that of bu- tyraldehyde. This substance may act as an alarm pheromone, as many individuals will take flight when one is captured. NEW LOCALITIES.— PG; 3 mi [ca. 4.8 km] W, 5 mi [ca. 8 km] N Roma, Starr County. Ornithia mexicana mexicana (Sturm, 1843:354) RANGE. — Southern Texas to Panama. ADULT ACTIVITY.— April to August. DISCUSSION.— The only North American record for this striking species is Vogt's (19490) collec- tion of a single specimen from beneath bark of Celtis. We have not seen Vogt's specimen and so have listed it as the nominate subspecies. Lins- ley's figure (1964:10, fig. 3) is of the form des- ignated as zapotensis Tippmann, from Guate- mala and Sinaloa, Mexico. Adults of both subspecies are common on a variety of deadwood and on blossoms in Mexico and Central Amer- ica. Hylotrupes bajulus (Linnaeus, 1758:396) RANGE. — Europe, Asia, North and South America, Asia Mi- nor, eastern Mexico, and Texas. ADULT ACTIVITY.— July to September. LARVAL HOSTS.— Pinus, Picea, Abies, Populus, Alnus, Cor- 306 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 ylus, Quercus, Genista, Conium, Acacia, Tamarix, among oth- ers (Duffy 1960). DISCUSSION.— This is the "Old House Borer" of economic literature— a species capable of causing considerable structural damage to a wide variety of wood products, including framing tim- bers, roofing, and flooring. It has been spread into many areas in North America in imported wood; a single record from Brownsville, Cam- eron County (RWN) has been seen from south- ern Texas. Megacyllene caryae (Gahan, 1908:141) RANGE.— Eastern North America to Texas and northern Mexico. ADULT ACTIVITY.— September to November (Texas and northern Mexico only). LARVAL HOSTS. — Carya, Juglans, Morus, Celtis, Madura, Vitis, Ulmus, Fraxinus, Gleditsia, Prosopis (Linsley, 1 964). DISCUSSION.— Adults of this large clytine are common in the fall on freshly cut mesquite, being most active during the late afternoon. This species was reared from burned mesquite logs gathered near Rio Grande City, Starr County, and spec- imens were collected in San Patricio County from stems and foliage of Baccharis, in company with Stenaspis, Placosternus, and Dendrobias. Inter- estingly, M. caryae is active only during the spring months over most of its range, but is a fall-active species in southern Texas. NEW LOCALITIES. — BRG; LCC; WWR; Corpus Christi, Nueces County; 3-7 mi [ca. 4.8-11.3 km] N Sinton, San Patricio County. Placosternus difficilis (Chevrolat, 1862:263) RANGE. — Florida and the West Indies, northern Mexico, Texas, and southern California. ADULT ACTIVITY.— February to November. LARVAL HOSTS.— Prosopis, Acacia, Pithecellobium, Platanus (Linsley 1964), Citrus (Manley and French 1976), Leucaena (Hovore and Penrose 1982). DISCUSSION.— Adults are active day and night, running rapidly along freshly cut branches of their host plants and feeding on the blossoms of Koe- berlinia, Acacia (Vogt 1949#), Baccharis, Bu- melia, Clematis, and Solidago. They are readily attracted to lights, and are common in both the spring and fall activity periods. This is the Cyl- lene crinicornis of older lists. NEW LOCALITIES. — PG; BRG; Anzalduas Park, Hidalgo County; SAR; 27 mi [ca. 44 km] S Sarita, Kenedy County; Kingsville, Kleberg County; Sinton, San Patricio County; WWR; LCC. Placosternus erythropus (Chevrolat, 1835:fasc. 4, no. 95) RANGE.— Texas to Guatemala. ADULT ACTIVITY. — March to October. LARVAL HOSTS.— Acacia, Prosopis (Duffy 1960). DISCUSSION.— In the fall, adults were abundant on Baccharis stems and on a variety of blossom- ing vines in the palm grove and at Welder Wild- life Refuge, and on Condalia and Bumelia flow- ers in the uplands near El Sauz, Starr County. Duffy cited host records for this species (listed as "Megacyllene [Cyllene] erythropa") by quot- ing older references, and we have seen no reared material. NEW LOCALITIES. — Hwy. 649, 1.6 mi [ca. 2.6 km] N Jet. Rt. 83, Starr County; 3-7 mi [ca. 4.8-11.3 km] N Sinton, San Patricio County. Ochraethes citrinus Chevrolat, 1860:474 RANGE.— Western Texas to southern Mexico. ADULT ACTIVITY.— September to November. DISCUSSION.— This species is included on the basis of several old specimens labelled only as having come from Hidalgo or Cameron County. We have not verified the determination and therefore list these specimens as citrinus, follow- ing Linsley (1964). Valid citrinus localities seen include 1 7 km S Saltillo, Coahuila, Mexico (FTH), and Big Bend National Park, Brewster County, Texas (MER). Most specimens were taken from blossoms of Compositae. Tanyochraethes tildeni Chemsak and Linsley, 1965:148 RANGE.— Southern Texas to extreme northern Mexico. ADULT ACTIVITY.— October to November. DISCUSSION.— Adults of this species were tak- en from inflorescences of Eriogonum and Soli- dago growing on the sandsheets of Kenedy Coun- ty in October (Hovore and Giesbert 1976). The yellow elytral vestiture is typically arranged into humeral, antemedian, median, and postmedian bands that have internally coalesced with the su- tural vitta. In our material, however, many in- dividuals have portions of the pattern, or even the entire elytral surface, suffused with yellow pubescence. In some specimens the patterns were altered or obliterated by abrasion. Neoclytus mucronatus vogti Linsley, 1957a:35 RANGE. — Southern Arizona to southern Texas and northern Mexico. ADULT ACTIVITY. — March to October. LARVAL HOSTS. — Celtis, Ulmus, Parkinsonia (Vogt 1949a), Prosopis (Turnbow and Wappes 1978). DISCUSSION.— A series of this colorful subspe- cies was taken from fresh-cut Celtis in May and again in October, in Bentsen State Park, Hidalgo County. Adults were also collected on Baccharis, and several beetles were found at night on dead elm trees. Larvae heavily infest dead trunks and HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 307 branches of the hosts, reducing most of the heart- wood to frass and fecula. Typically, N. m. vogti is lighter in coloration and more strikingly marked than the widely distributed nominate subspecies; however, as Linsley (19570) noted, there is con- siderable intermediacy in coloration in material from central and eastern Texas. Specimens from near San Antonio, Bexar County, cannot be placed with certainty in either subspecies; most of these specimens closely resemble material from the eastern U.S. NEW LOCALITIES. — 2 mi [ca. 3.2 km] S Pharr, Hidalgo Coun- ty; WWR; LCC. Neoclytus acuminatus hesperus Linsley, 19356:163 RANGE.— Colorado, New Mexico, southern Texas. ADULT ACTIVITY. — March to October. LARVAL HOSTS. — Quercus (Linsley 1964), Acacia (Linsley and Martin 1933), Citrus (Manley and French 1976), Bac- charis, Prosopis (Turnbow and Wappes 1978), Zanthoxylum (Turnbow and Wappes 1981), Celtis. DISCUSSION.— Adults are wary and quick to fly at the slightest disturbance, making capture dif- ficult. The nominate subspecies, often called the "red-headed ash borer," is a well-known pest that breeds on a variety of hardwood trees in the eastern U.S. Lighter integumental coloration, the primary separating character for the subspecies hesperus, is variable and difficult to quantify in material examined from the total species range. The subspecies was originally defined from a sin- gle specimen from Colorado, and uniformly red- dish coloration is found in a number of popu- lations peripheral to the range of N. a. acuminatus, including those from southern Tex- as material. NEW LOCALITIES. -PG; BRG; 2 mi [ca. 3.2 km] S Pharr, Hidalgo County; LCC; WWR. Neoclytus augusti Chevrolat, 1835:fasc. 4, no. 73 RANGE.— Southern Texas to northern Mexico. ADULT ACTIVITY. — March to October. LARVAL HOSTS.— Sapindus (Vogt 1949a), Citrus (Manley and French 1976, 1977), Prosopis, Ulmus, Celtis (Turnbow and Wappes 1978). DISCUSSION.— Vogt (19490) collected a series of adults from weakened Baccharis, and Hovore (1983) discussed augusti-like material from Bac- charis in western Texas. On earlier lists (except Manley and French 1976, 1977) this species ap- peared as N. abbreviatus Schaeffer, a junior syn- onym. Euderces reichei exilis Casey, 1893:591 RANGE.— Southern Texas to Tamaulipas, Mexico. ADULT ACTIVITY.— March to October. LARVAL HOSTS.— Sapi ndus (Vogt 1949a), Celtis, Prosopis (RHT), Acacia, Zanthoxylum. DISCUSSION.— Although this tiny ant-mimick- ing beetle was recorded as having been beaten (Vogt 19490) or reared (Linsley 1940) from a variety of shrubs and trees, specific host data were rather scant. Our Acacia specimens emerged from a branch, 3 cm in diameter, girdled by On- cideres pustulatus LeConte at Kingsville, Kleberg County. Adults are very common on deadwood and at blossoms of a variety of woody and her- baceous plants. Linsley (1964) cited the distri- bution of this subspecies as Hidalgo and Cam- eron counties, but specimens from Zapata County on the west side of the state, and San Patricio County on the Gulf Coast, based upon the rel- ative development of the antennal spines, are also referable to exilis. The nominate taxon is distributed throughout the southcentral U.S. NEW LOCALITIES. -BRG; PG; SAR; LCC; 8 mi [ca. 13 km] SE Zapata, Zapata County; 5.3 mi [ca. 8.5 km] E Rio Grande City, Starr County; WWR; La Lomita Park, Hidalgo County; 3 mi [ca. 4.8 km] S Mission, Hidalgo County. Tetranodus niveicollis Linell, 1896:396 RANGE.— Southern Texas south to Oaxaca, Mexico. ADULT ACTIVITY. — May to June. LARVAL HOSTS.— Pithecellobium (Turnbow and Wappes 1981). DISCUSSION.— Adults were beaten from Mi- mosa, Acacia (Linsley and Martin 1933), and Prosopis (FTH), and two specimens were reared from dead ebony gathered near Boca Chica, Cameron County. NEW LOCALITIES. — PG; LCC; Brownsville, Cameron Coun- ty- Pentanodes dietzii Schaeffer, 1904:222 RANGE.— Southern Texas. ADULT ACTIVITY.— Unknown. DISCUSSION.— The unique holotype and allo- type were reportedly collected at Brownsville, Cameron County, with no further data supplied by their describer. No other specimens are known. Dihammophora dispar Chevrolat, 1859:52 RANGE.— Southern Texas to Mexico. ADULT ACTIVITY. — Unknown for Texas; one specimen seen from Oaxaca, Mexico in August. DISCUSSION.— This species is occasionally col- lected from blossoms and on deadwood in Mex- ico. Aside from Schaeffer's (1908) record from Brownsville, Cameron County, based upon ma- terial in the Dietz collection (which contains a number of unduplicated records), we know of no other Texas specimens. 308 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 Rhopalophora angustata Schaeffer, 1905/>:162 RANGE. — Southern Texas and northern Mexico. ADULT ACTIVITY. — March to October. LARVAL HOSTS. — Citrus (Manley and French 1976), Pithe- cellobium, Eysenhardtia (Turnbow and Wappes 1978), Zan- thoxylum, Diospyros (Turnbow and Wappes 1981). Discussion.-This graceful species was abun- dant on flowering Baccharis at Welder Wildlife Refuge in October, and adults were also beaten from dead twigs of granjeno and ebony (Hovore and Giesbert 1976). Specimens have been col- lected from blossoms ofMonarda (Vogt 1949a) and Clematis. NEW LOCALITIES. -PG; LCC; 5.3 mi [ca. 8.5 km] SE Rio Grande City, Starr County; 3 mi [ca. 4.8 km] N Roma. Rhopalophora laevicollis (LeConte, 1873:193) (Figure 6) RANGE.— Southern Texas to southern Mexico. ADULT ACTIVITY.— May to October. LARVAL HOSTS. — Citrus (Manley and French 1976), Pithe- cellobium, Diospyros, Zanthoxylum (Turnbow and Wappes 1981). DISCUSSION.— Adults often are common on fresh-cut limbs of larval hosts, on Celtis, and at blossoms of Clematis, Cissus, Serjania, Sam- bucus, Helianthus, Mimosa, Baccharis, and Haplopappus. NEW LOCALITIES.— PG; Rio Grande City, Starr County; LCC; WWR. Rhopalophora rugicollis (LeConte, 18586:83) RANGE.— Texas and northern Mexico to northern Arizona and the Cape Region of Baja California. ADULT ACTIVITY.— March to June. LARVAL HOSTS. — Celtis (Tyson 1 970), Pithecellobium (Turn- bow and Wappes 1978). DISCUSSION.— Linsley and Martin (1933) took this species on willow (Linsley and Martin 1933) in the Lower Valley, and in other portions of the species range, adults have been collected from blossoms of Mimosa, Acacia, Lupinus, and Ce- anothus. NEW LOCALITIES.— LCC. Rhopalophora longipes longipes (Say, 1823:426) RANGE. — Eastern North America to Kansas and Texas. ADULT ACTIVITY.— May to June. LARVAL HOSTS. — Cercis, Cornus (Linsley 1964). DISCUSSION.— Two specimens, which are ten- tatively referred to this common eastern species, were collected from white Compositae growing along the roadside 1 1 mi [ca. 1 8 km] S Three Points, Webb County, in May (FTH). Although R. I. meeskei Casey is known from as near as montane western Texas, the relative pronotal proportions of the two specimens preclude their placement with that subspecies. Rhopalophora longipes rather closely resembles R. bicolorella Knull, from southern Arizona, but it differs by having slightly coarser pronotal punctures, very slightly sparser elytral punctation, and a less- pubescent dorsal surface. Other longipes-\ike specimens have been seen from central Mexico and the Cape Region of Baja California, and the Neotropical species of Rhopalophora need a comprehensive taxonomic review before a de- finitive determination can be made on our ma- terial. Agallisus lepturoides (Chevrolat, 1849:12) RANGE. — Southern Texas(?) to Honduras. ADULT ACTIVITY.— Unknown for Texas. DISCUSSION.— This exotic species has been list- ed from Texas several times, but we have been unable to locate or collect any U.S. material. The genus is structurally similar to other Agallisini (Zagymnus and Osmopleurd), species of which breed in dead fronds and floral scapes of Pal- maceae. A similar host association for Agallisus would restrict its range in Texas to remnant sabal palmetto groves in the Lower Valley. Ancylocera bicolor (Olivier, 1795:32) RANGE.— Southeastern North America to western Texas. ADULT ACTIVITY.— April to July. LARVAL HOSTS. — Carya, Quercus (Fattig 1947), Celtis (Turnbow and Wappes 1978), Acacia (Turnbow and Wappes 1981). DISCUSSION.— Vogt (1949#) collected this pe- culiar-looking beetle on Acacia, Baccharis, and fresh-cut Leucaena. We took numerous speci- mens in southern Texas from cedar elm slash and from blossoms of Verbesina. We collected an adult female from roadside Compositae near Uvalde, Uvalde County, in western Texas (FTH). In Florida, Turnbow and Hovore (1979) en- countered numerous adults feeding on fungus growing on old stumps and logs of oak. NEW LOCALITIES. -BRG; LCC; WWR; Resaca de la Palma State Park, Cameron County; Anzalduas Park, Hidalgo Coun- ty; 1 1 mi [ca. 18 km] S Three Points, Webb County. Lepturinae Strangalia virilis LeConte, 1873:212 RANGE.— Texas and Oklahoma. ADULT ACTIVITY. — May to June. LARVAL HOSTS. — Quercus (Linsley and Chemsak 1976). DISCUSSION.— Adults of this striking species have been recorded as visiting blossoms of a va- HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 309 riety of plants (Linsley and Chemsak 1976), and it is particularly abundant on horsemint (Monar- da punctata) in central and eastern Texas. A sin- gle specimen was seen from Lake Corpus Christi State Park, San Patrick) County, in June (H. Flaschka). Pseudostrangalia cruentata (Haldeman, 1847:64) RANGE.— Eastern North America from Canada to Texas. ADULT ACTIVITY.— April to June. DISCUSSION.— One specimen has been seen from southern Texas, labelled "Brownsville, VII- 2-65" (JC). I.eptura (Stenura) gigas LeConte, 1873:223 RANGE.— Texas and northern Chihuahua, Mexico. ADULT ACTIVITY.— April to August. LARVAL HOSTS.— Salix (Vogt 1949a). DISCUSSION.— Adults of this large red-and- black species are difficult to capture, being strong and agile flyers and spending much of their time high in the foliage of their host trees. By their color, form, and swift, buzzing flight these insects closely resemble pompilid wasps of the genus Pepsis, which they may mimic. Adults are at- tracted to fermenting molasses bait, and occa- sionally come to lights. Larvae bore in decaying logs or rotting portions of living willow trees, particularly wind-broken branch butts and healed-over scars (Hovore 1983). NEW LOCALITIES. — PG; Anzalduas Park, Hidalgo County. Cyphonotida laevicollis laevicollis (Bates, 1880:39) RANGE.— Southern Texas to El Salvador. ADULT ACTIVITY.— October. DISCUSSION. — Vogt (1949a) collected five specimens on flowers of Bumelia, and we took numerous specimens on blossoms of Clematis, Serjania, and Cissus in the Palm Grove Sanc- tuary in the fall. NEW LOCALITIES. — SAR; Brownsville, Cameron County; BRG; Mission, Hidalgo County. Lamiinae Parmenosoma griseum Schaeffer, 1908:344 (Figure 7) RANGE.— Southern Texas. ADULT ACTIVITY. — March to November. LARVAL HOSTS. — Opuntia (Mann 1969), Yucca (Rice et al. 1985). DISCUSSION.— Most specimens of this flightless species were collected by beating basal rosettes of fallen Yucca and Agave, both of which prob- ably serve as larval hosts. FIGURE 7. (right). Parmenosoma griseum (left) and Ataxia tibialis NEW LOCALITIES. — 3 mi [ca. 4.8 km] W, 5 mi [ca. 8 km] N Roma, Starr County; Lopeno, Zapata County. Moneilema armatum LeConte, 1853:234 RANGE. — Southern portions of the Great Plains from Col- orado and Kansas south to Mexico (distribution given for all forms of armatum). ADULT ACTIVITY. — May to October. LARVAL HOSTS. — Opuntia. DISCUSSION. — Raske (1971) considered south- ern Texas armatum to belong to the subspecies punctatum Psota, 1930:133, distinguished from more northern populations by the more coarsely punctate dorsal surface. This feature varies cli- nally from north to south in populations of ar- matum, reaching its highest degree of develop- ment in the form rugosipenne Fisher from central Mexico (also considered by Raske to be a sub- species of armatum). Linsley and Chemsak (1 985) did not recognize subspecies in Moneilema ar- matum. Moneilema larvae bore in stems and root col- lars of living cactus; M. armatum larvae show a preference for the prickly pear cactus Opuntia (Raske 1971). NEW LOCALITIES. — 3 mi [ca. 4.8 km] N Roma, Starr County; Lopeno, Zapata County; 10 mi [ca. 16 km] N Laredo, Webb County; 14 mi [ca. 22.5 km] SE Three Points, Webb County. Moneilema blapsides ulkei Horn, 1885:188 (Figure 8) RANGE.— Central Texas to northern Mexico. ADULT ACTIVITY.— April to December. LARVAL HOSTS.— Opuntia (Mann 1969). DISCUSSION.— This species is both dimorphic 310 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 FIGURE 8. Male (left) Moneilema blapsides ulkei, female (middle) Moneilema mundelli, female (right) Moneilema blap- sides ulkei. See species accounts for a discussion of the rela- tionships of these taxa. and dichromatic. Males have a black, densely punctate dorsum, usually with a finely reticulated pattern of whitish pubescence intermixed with indistinct brownish hairs. Females are usually wholly black, glabrous, and at most very sparsely punctate. Moneilema mundelli Fisher, 1931:200 (Fig. 8) may only be a morph of this species, differing primarily by the white-reticulated dor- sal pubescence of the females and the more clear- ly denned pubescent pattern of the males. Linsley and Chemsak (1985) synonymized mundelli un- der M. b. ulkei. Adults of all forms of M. b. ulkei were abun- dant on Opuntia atop the so-called Yucca Ridges northeast of Brownsville, Cameron County, while the typical form was found at a number of upland localities. NEW LOCALITIES.— 1 1 mi [ca. 18 km] S Three Points, Webb County; 10 mi [ca. 16 km] N Laredo, Webb County. Neoptychodes trilineatus (Linnaeus, 17676:532) RANGE.— Southern U.S. to northern South America, West Indies, Tahiti, Baja California. ADULT ACTIVITY. — May to October. LARVAL HOSTS.— Ficus, Alnus, Morus (Dillon and Dillon 1941), Chlorophora, Spondias, Inocarpus (Duffy 1960), Salix (Linsley et al. 1961), Celtis (JC), Juglans. DISCUSSION. — Horton (1917) recorded the species' life history on fig trees in Louisiana, and Linsley et al. (1961) stated that N. trilineatus is a primary borer in willow and mulberry in south- eastern Arizona. Dillon and Dillon (1941) listed N. trilineatus from Brownsville, Cameron Coun- ty; and its occurrence in the Lower Valley would be expected, but we have not encountered it dur- ing the course of this study. Plectrodera scalator (Fabricius, 1792:278) RANGE. — Eastern North America, from Great Lakes states west to New Mexico and south to Texas. ADULT ACTIVITY.— April to July. LARVAL HOSTS.— Populus, Salix (Milliken 1916). DISCUSSION.— Adults of this boldly patterned species were collected from yard and tree lawn plantings of Populus in Kingsville, Kleberg County, in May. This is the southernmost record that we are aware of for the species; this species may have been introduced in ornamental plant- ings of the host tree. Adults frequent foliage and trunks of larval hosts. The larvae mine the living root crown, often seriously damaging the plant. Milliken (1916), Craighead (1950), and Solomon (1980) described the immature stages and dis- cussed the life history in other portions of the species range. Goes fisheri Dillon and Dillon, 1941:122 RANGE.— Western to southern Texas. ADULT ACTIVITY.— June to August. DISCUSSION. — Originally described from Uvalde in western Texas, the few specimens we saw were from the Balcones Escarpment region of the state. A single south Texas specimen is known, labeled " Raymond ville, Willacy County, VIII- 1969." J. E. Wappes (pers. comm.) stated that this specimen "was in alcohol UVL material along with some Oncideres pustulata." Although the pubescence of the specimen is rubbed and matted, it compares well with the original char- acterization of G. fisheri. Goes tesselatus (Haldeman, 1847:51) RANGE.— Eastern North America south to Florida, west to Texas. ADULT ACTIVITY. — May to July. LARVAL HOSTS. — Quercus, Castanea, Amelanchier (Dillon and Dillon 1941), Ulmus (Linsley and Chemsak, 1985). DISCUSSION.— A single specimen was seen from Lake Corpus Christi State Park, San Patricio County, collected in mid-June by R. Heitzman (TCM). Goes pulverulentus (Haldeman, 1847:51) RANGE. — Eastern North America south to northern Florida, west to Texas. ADULT ACTIVITY. — May to July . LARVAL HOSTS.— Betula, Carpinus, Ostrya, Quercus, Ulmus, Platanus, Fagus (Craighead 1923), Prunus (Knull 1946). DISCUSSION.— Dillon and Dillon (1941) re- corded this eastern monochamine from Corpus Christi, Nueces County. Solomon (1972) gave HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 311 details of its bionomics on oak in Mississippi, and based upon his observations, it is probable that the species occurs on oak in the sandsheet regions south of Corpus Christi and Kingsville, Kleberg County. Dorcaschema wildii Uhler, 1855:417 RANGE. — Eastern North America to southern Texas. ADULT ACTIVITY. — May to August. LARVAL HOSTS.— Morus, Toxylon (=Maclura) (Craighead 1923). DISCUSSION.— A single specimen of this Alle- ghenian species was taken at Welder Wildlife Refuge in July (RHT). Adults are often common on foliage and infested branches of mulberry and come to lights. Solomon (1968) detailed the life history of D. wildii on Morus in Mississippi. Dorcaschema alternatum Say, 1823:405 RANGE.— Eastern North America to Florida and Texas. ADULT ACTIVITY.— April to October. LARVAL HOSTS.— Morus (Craighead 1923), Toxylon (=Ma- c/«ra)(Knull 1946). DISCUSSION.— Dillon and Dillon (1948) as- signed specimens from southern Texas (Kings- ville, Kleberg County) to the subspecies D. a. octovittata Knull (described in 1937 from the Davis Mountains, Jeff Davis County, Texas). Material at hand from the type locality of octo- vittata differs markedly in coloration and pubes- cent pattern from all other alternatum popula- tions examined, and material from southern Texas definitely does not belong with the west Texan form. Linsley and Chemsak (1985) did not recognize subspecies in D. alternatum. NEW LOCALITIES. — Pharr, Hidalgo County; Brownsville, Cameron County; WWR; Nueces and Lavaca counties (TAI). Parmenonta wickhami Schaeffer, 1908:350 (Figure 9) RANGE.— Southern Texas. ADULT ACTIVITY. — May to December. DISCUSSION.— Two specimens of this flightless longhorn were swept from herbaceous vegetation at Welder Wildlife Refuge in May, and numerous adults were beaten from Celtis, Condalia, and Clematis in the palm grove. Adetus brousi (Horn, 1880:137) (Figure 9) RANGE.— Kansas to northern Mexico. ADULT ACTIVITY. — May to July. LARVAL HOSTS. — Cucumis (=Cucurbita) (Horn 1880). DISCUSSION.— This species breeds in dried FIGURE 9. Parmenonta wickhami (left), Adetus brousi (middle), and Desmiphora aegrota (right). stems of wild gourd and possibly other Cucur- bitaceae; adults have been taken from foliage of the larval host. In southern Texas, specimens were beaten from tangles of vines in the palm grove and were attracted to lights. Dorcasta cinerea (Horn, 1860:571) RANGE.— Texas. ADULT ACTIVITY.— May to October. LARVAL HOSTS.— Datura, Nicotiana, Solanum, Gossypium, Verbesina (Linsley and Chemsak 1985), Matelea (Rice et al. 1985). DISCUSSION.— Adults were collected by sweep- ing or beating the larval hosts; a few specimens were attracted to lights. We took numerous spec- imens from stems of sunflower (Helianthus) near Kingsville, Kleberg County, and we found the species infesting Nicotiana trigonophylla at Fal- con Heights, Zapata County, in May. Turnbow and Wappes (1978) took a female beetle on an Oncideres-girdled Acacia twig at Bentsen-Rio Grande Valley State Park, Hidalgo County. Huff- man and Harding (1980) took a single specimen in a pitfall trap in a Citrus grove. NEW LOCALITIES. — PG; SAR; Lopeno, Zapata County; Ar- royo Salado at Hwy. 83, Starr County; Jet. Hwys. 649 and 2686, Starr County; San Ygnacio, Zapata County; LCC; 4 mi [ca. 6.5 km] S Agua Dulce, Nueces County. Ataxia huhbardi Fisher, 1924:253 (Figure 10) RANGE.— Southern U.S. from Arizona to Louisiana. ADULT ACTIVITY. — March to October. LARVAL HOSTS.— Xanthium, Helianthus, Ambrosia, Sil- phium, Vernonia, Cirsium, Erigeron, Gossypium, Smilax, 312 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 FIGURE 10. Ataxia hubbardi. Thurberia, Verbesina (Linsley and Chemsak 1985), Apocynum (Williams 1941). DISCUSSION.— Adults are collected at lights and by beating or sweeping dead stems of the larval hosts. Rogers (1977 b) presented life history data for A. hubbardi on sunflowers in Texas. NEW LOCALITIES. — BRG; Pharr and Mission, Hidalgo Coun- ty; San Ygnacio, Zapata County; FSP; WWR; 4 mi [ca. 6.5 km] S Agua Dulce, Nueces County. Ataxia crypta (Say, 1831:5) RANGE. — Eastern North America south into northern Mex- ico. ADULT ACTIVITY. — March to November. LARVAL HOSTS. — Quercus, Castanea, Pyrus, Xanthium, Verbesina, Ambrosia, Thurberia, Smilax, Gossypium (Craig- head 1923), Salix (Hovore et al. 1978), Acer, Celtis (Leng and Hamilton 1896), Acacia (Tumbow and Wappes 1981), Prunus (Linsley and Chemsak 1985). DISCUSSION.— This species breeds in a wide variety of host plants, and adults are abundant on dead branches of hardwood trees. A few spec- imens were beaten from dead Yucca near El Sauz, Starr County (FTH). Earlier host records for this species in herbaceous plants are considered er- roneous, referring to the more recently described Ataxia hubbardi. In material examined during this study, the two species were consistently mixed and misidentified. Adults are readily attracted to lights. NEW LOCALITIES.— PG; BRG; Anzalduas Park, Hidalgo County; Lopeno, Zapata County; FSP; 8 mi [ca. 1 3 km] SE Zapata, Zapata County; BC; Kingsville, Kleberg County; LCC; WWR. Ataxia tibialis Schaeffer, 1908:348 (Figure 7) RANGE. — Brownsville, Cameron County, and vicinity. ADULT ACTIVITY. — May and June. DISCUSSION.— We know of only seven speci- mens, all from the palm grove; some specimens were collected from dead Zanthoxylum, some by miscellaneous beating and some at lights. Desmiphora hirticollis (Olivier, 1795:1 1) (Figure 11) RANGE. — Southern Texas to Mexico and South America (Ar- gentina), West Indies. ADULT ACTIVITY.— March to October. LARVAL HOSTS.— Sapium (Duffy 1960). DISCUSSION.— Vogt (1949a) found this species feeding upon terminal shoots of Cordia in June and September, and a few adults have been beat- en from this shrub. Specimens have also been taken at lights. In Central America this species is common at night on dead trunks and branches of a variety of hardwood trees. In southern Mex- ico it is abundant on healthy green leaves of an undetermined species of nettle (FTH, EFG). NEW LOCALITIES.— PG; Anzalduas Park, Hidalgo County; SAR; BRG; 4 mi [ca. 6.5 km] W Sullivan City, Starr County; 10 mi [ca. 16 km] E Rio Grande City, Starr County; Pharr, Hidalgo County; LCC. Desmiphora aegrota Bates, 1880:1 16 (Figure 9) RANGE.— Southern Texas to Panama. ADULT ACTIVITY.— April to October. LARVAL HOSTS.— Malvaviscus (Rice et al. 1985). DISCUSSION.— This tropical species was only recently recorded from North America (Turn- bow and Wappes 1981), and the oldest record seen is a specimen labeled "Southmost, Cameron County, 20-X-74" (UCB). We collected adults from the vinelike stems of turk's cap (Malvavis- cus arboreus var. drummondi} in the palm grove both day and night, and M. Rice subsequently reared it from dead stems of this plant. In Central America D. aegrota has been beaten from dead branches of hardwood trees. Eupogonius pauper LeConte, 1852:159 RANGE. — Eastern North America south to Florida and Mex- ico. ADULT ACTIVITY.— April to June. LARVAL HOSTS.— Morus, Cornus, Juglans, Cercis, Celastrus, Acer, Fraxinus, Asimina, Zanthoxylum, Carpinus, Carya, Castanea, Gleditsia, Hamamelis, Prunus, Quercus, Rhus, Til- ia, Ulmus (Linsley and Chemsak 1985). HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 313 FIGURE 12. Ecyrus arcuatus (left) and Ecyrus penicillatus (right). FIGURE 1 1 . Desmiphora hirticollis. DISCUSSION.— On previous lists, this eastern species was erroneously identified as E. fulvo- vestitus Schaeffer, or recorded as E. vestitus (Say), an unavailable name due to homonymy (Dillon and Dillon 1953; Breuning 1974). Specimens from southern Texas differ from typical material from the eastern U.S. by having yellowish pu- bescence (whitish, gray, or cinereous in eastern populations) and a reddish-brown integument (typically dark brown to piceous). Adults were beaten from Fraxinus (Linsley and Martin 1 933) and Ulmus (FTH, RLP) at several localities in southern Texas. NEW LOCALITIES. -PG; BRG; SAR; LCC; WWR. Eupogonius fulvovestitus Schaeffer, 1 905a: 1 34 RANGE.— Southern Texas. ADULT ACTIVITY. — March to May. DISCUSSION.— In addition to Schaeffer's orig- inal record, the only specimens we saw were col- lected by D. J. and J. N. Knull, labeled simply "Hidalgo County" (Knull published this record in 1954 but added no further data). Pygmaeopsis viticola Schaeffer, 1908:348 RANGE.— Southern Texas. ADULT ACTIVITY.— May to September. DISCUSSION. -Schaeffer (1908:348) stated that this species was taken from "heavy dead stems of vines inside the palmetto grove. . . ." Linsley and Martin (1933) beat specimens from jungle vines at the same site, and Vogt (1949fl) took single specimens by sweeping weeds and at lights. We have seen no recently collected material. Callipogonius cornutus (Linsley, 1930:86) RANGE.— Southern Texas to Veracruz and Jalisco, Mexico. ADULT ACTIVITY.— April to June, October to November. LARVAL HOSTS.— Salix (Hovore et al. 1978). DISCUSSION.— This cryptically colored pogo- nocherine was abundant on fresh broken willow during spring and early summer in the palm grove, and adults were later reared from this host. Ho- vore et al. (1978) discussed the larval habits of this species and listed ecologically associated Co- leoptera. Callipogonius cornutus is very closely related to C. hircinus (Bates) from Veracruz, Mexico, and the two may prove to be conspecific. Ecyrus penicillatus Bates, 1880:137 (Figure 12) RANGE.— Southern Texas to Veracruz and Sinaloa, Mexico. ADULT ACTIVITY.— April to August, October. LARVAL HOSTS.— Pithecellobium (Rice et al. 1985). DISCUSSION.— This beetle resembles a bird dropping when resting on dead twigs or in a death- feigning posture (legs and antennae drawn tight to the body) on the beating sheet. This species is uncommon; most material has been beaten from dead branches of Celtis or Salix or has been at- tracted to UV lights. The species appeared as E. fasciatus Hamilton on some previous lists. NEW LOCALITIES.— PG; Sam Fordyce Road, 0.5 mi [ca. 0.8 km] S Hwy. 83 (N. M. Downie). 314 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 FIGURE 13. Male (left) and female (right) Lochmaeocles cornuticeps cornuticeps. Ecyrus arcuatus Gahan, 1892:259 (Figure 12) RANGE.— Central Texas to Guatemala. ADULT ACTIVITY. — March to October. LARVAL HOSTS.— Acacia (Linsley 1935a), Prosopis. DISCUSSION.— This species is abundant on dead branches of its hosts, particularly in the fall. Nu- merous adults were taken at night from trunks and limbs of uprooted second-growth mesquite and huisache at Welder Wildlife Refuge in Oc- tober, and from fresh-cut mesquite at Bentsen- Rio Grande Valley State Park in May. Linsley (1 940) recorded rearing it from leguminous plants which had been girdled by Oncideres pustulatus, and we bred it from dead mesquite. Adults oc- casionally come to lights. Some earlier authors regarded arcuatus as a subspecies (texanus Schaeffer, 1908:347) of the eastern Ecyrus dasycerus (Say). Chemsak and Linsley ( 1 91 5b) cited specimens of arcuatus from X-Can, Quintana Roo, Mexico, and Peten, Ti- kal, Guatemala. NEW LOCALITIES. — PG; BC; Mission, Hidalgo County; 8 mi [ca. 1 3 km] SE Zapata, Zapata County; LCC; SAR. Lochmaeocles cornuticeps cornuticeps (Schaeffer, 1906:20) (Figures 4, 1 3) RANGE. — Southern Texas and northern Mexico. ADULT ACTIVITY.— April to October LAR VAL HOSTS.— Salix (Hovoreetal. 1978), Leucaena (Vogt 1 949a; Hovore and Penrose 1 982), Celtis, Acacia (Knull 1 937). DISCUSSION.— This large onciderine, abundant on dead tepehuaje and hackberry in the palm grove, was not encountered in any other Lower Valley habitat. Vogt (1949a) took adults at Pharr, Hidalgo County, and we saw two specimens la- beled "Raymondville" (UCB), so it does occur in other areas, but apparently less commonly than in the grove. Adults come to lights, and a single female came to molasses bait in October. Hovore and Penrose (1982) discussed the larval habits in Leucaena and gave comparative characters for separating larvae of L. c. cornuticeps from larvae of Oncideres pustulatus. Oncideres pustulatus LeConte, 1854a:82 (Figure 3) RANGE.— Texas and northeastern Mexico to southern Ari- zona(?). ADULT ACTIVITY.— August to December. LARVAL HOSTS.— Acacia, Pithecellobium, Prosopis, Parkin- sonia, Mimosa (Linsley 1 940), Leucaena (Vogt 1 949a; Hovore and Penrose 1982), Citrus (Dillon and Dillon 1946), Albizzia (Thomas in Ferris 1980). DISCUSSION.— The life history of this species, commonly called the huisache girdler, has been recorded by High (1915, as O. putator Thom- son), Linsley and Martin (1933), Vogt (19490), Duffy (1960, as O. putator), Thomas in Ferris (1980), and Hovore and Penrose (1982). The gir- dling habits of adult beetles can be very destruc- tive to smaller trees, and severe growth deform- ities can result from pruning distal portions of the trunk and lateral branches. Thomas in Ferris (1980), however, stated that at least one host (Albizzia julibrissin, introduced) gains increased longevity by regular prunings, suggesting a mu- tualistic relationship between O. pustulatus and its host. Leucaena saplings girdled near the base may grow into shorter, more compact trees than ungirdled saplings; a compact shape could be advantageous to a soft-wood species during se- vere storms. Dillon and Dillon (1946) and Linsley and Chemsak (1985) stated that O. pustulatus is con- fined to Texas and adjacent Mexico, but Papp (1959) recorded it from New Mexico and Ari- zona, based upon material from the LACM col- lection (Ramsey Canyon and Huachuca Moun- tains, Arizona; Santa Fe, New Mexico; "Rio Grande Canyon, south of Taos, New Mexico" data fide R. R. Snelling). We have seen no other Arizona or New Mexico collections. NEW LOCALITIES. — PG; Brownsville, Cameron County; SAR; BC; Kingsville, Kleberg County; WWR. Oncideres cingulata texana Horn, 1885:195 (Figure 15) RANGE.— Texas. ADULT ACTIVITY. — May to November. HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 315 LARVAL HOSTS.— Prosopis, Acacia, Pithecellobium (Linsley 1940), Citrus (Manley and French 1976), Gliditala [sic] (Gle- ditsia) (Dillon and Dillon 1 946), Parkinsonia, Celtis. DISCUSSION. — Rogers (1977a) reported on the bionomics of O. cingulata ssp. in north-central Texas, and stated that a small percentage of ma- ture larvae pass a second winter in the host, pu- pating and emerging the following spring. If this pattern applies to texana in the study area, it might account for what appear to be two distinct broods each year. Specimens taken in May are, on average, slightly smaller in size and are less densely pubescent dorsally than those found in the fall; the pubescence difference is not attrib- utable to abrasion. Habits of O. c. texana were recorded in older literature under "O. cingulatus," "O. cingula- tor," and "O. texana." Determining which sub- species of cingulata was being discussed in older papers on biology is difficult, as most works did not differentiate records geographically. It ap- pears that few of the early bionomic reports at- tributed to texana actually refer to the taxon as currently recognized. Adults are extremely abundant on mesquite and huisache throughout the southern portion of the state. We have observed girdling and ovi- positing in species of trees not known to actually serve as larval hosts (retama, hackberry), but we have not reared any specimens from these plants. Adult beetles are commonly attracted to lights. NEW LOCALITIES. -PG; BC; BRG; LCC; WWR. Cacostola salicicola (Linsley, 1934:184) (Figure 14) RANGE.— Southern Texas, western Mexico. ADULT ACTIVITY.— April to October. LARVAL HOSTS.— Leucaena, Salix (Turnbow and Wappes 1981). DISCUSSION.— Adults were beaten from dead willow twigs in the palm grove in May and Oc- tober, and the species was subsequently reared from this host. Turnbow and Wappes (198 1) re- ported rearing a single specimen from Oncideres- girdled tepehuaje, also from the palm grove. Linsley (1934) noted that adults pose with the mesothoracic legs and antennae oriented at an angle to the linear axis of the body and the ab- domen raised, perhaps mimicking the appear- ance of a spider or a broken twig. Similar pos- turing was observed in the ataxiine Epectasis hiekei Breuning in Mexico (FTH). Specimens of either this or a very closely re- lated species were taken from dead shrubs (but FIGURE 1 4. Cacostola lineata (left) and Cacostola salicicola (right). not Salix or Leucaena) near Mazatlan, Sinaloa, Mexico (FTH, EFG), and the species may be widely distributed in Mexico. On older lists this species was placed in the genus Cylindrataxia Linsley. NEW LOCALITIES.— "Hidalgo" (American Museum of Nat- ural History); Southmost sector, Brownsville, Cameron Coun- ty; LCC. Cacostola lineata (Hamilton, in Lengand Hamilton, 1896:142) (Figure 14) RANGE.— Southern Texas. ADULT ACTIVITY.— April to October. DISCUSSION.— A rare species in collections, C. lineata appears to be confined to the extreme Lower Valley region. We collected numerous adults from dead Baccharis growing on the low hills west of Boca Chica beach, and we beat ad- ditional specimens from Salix, Celtis, Condalia, and tangles of vines and shrubs in the palm grove. Linsley and Martin (1933) recorded C. lineata as a new species of Aporataxia, listing the then undescribed C. salicicola as lineata. NEW LOCALITIES.— 10 mi [ca. 16 km] W Boca Chica, Cam- eron County. Hippopsis lemniscata (Fabricius, 1801:330) RANGE. — Eastern North America to Central and South America. ADULT ACTIVITY.— April to September. LARVAL HOSTS.— Melothria, Coreopsis, Bidens, Ambrosia (Leng and Hamilton 1896), Vernonia, Xanthium (Schwitzgebel and Wilbur 1942), Erigeron (Harris and Piper 1970), Erech- tites, Ageratum, Sesamum (Duffy 1960), Helianthus (Rogers \977b), Amaranthus, Desmodium, Glycine, Rudbeckia (Lins- ley and Chemsak 1985). 316 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 FIGURE 15. Oncideres cingulata texana. DISCUSSION.— Larvae bore in stems of herba- ceous plants, principally Compositae. Craighead (1923) described the larva, and Piper (1977) gave a fully referenced account of the life history and habits of this species. Adults are readily attracted to lights and may be swept from their hosts dur- ing the day. A single specimen was beaten from Aster spinosus at Anzalduas Park, Hidalgo Coun- ty, in company with Mecas linsleyi Knull. NEW LOCALITIES. -PG; SAR; BRG; LCC; WWR. Spalacopsis texana Casey, 1891:146 RANGE.— Southern Texas. ADULT ACTIVITY. — May to October. DISCUSSION. — Tyson (1973) collected this species from "Hostelezkya" [sic] and lantana. R. L. Penrose swept numerous specimens from grasses and understory vegetation at Welder Wildlife Refuge in May and beat a mating pair from Baccharis at that site in October. The larval host is not known, but S. texana probably breeds in dead stems of grasses, annual Compositae, or other pithy plants. NEW LOCALITIES.— BC; South Padre Island, Cameron Coun- ty (PAU); Sarita, Kenedy County (TAI). Thryallis undatus (Chevrolat, 1834:fasc. 3, no. 61) (Figure 16) RANGE.— Southern Texas to Mexico and Guatemala. ADULT ACTIVITY.— April to October. LARVAL HOSTS.— Leucaena (Vogt 1949a; Hovore and Pen- rose 1982), Celtis (Turnbow and Wappes 1981), Pithecellobi- um, Acacia (Rice et al. 1985). DISCUSSION.— This rotund beetle is common in the palm grove on dead branches of its larval hosts, and adults also were beaten from dead branches of willow and ebony. Vogt's (19490) collection of a single specimen near Mission, Hi- dalgo County, is the only Texas record for T. undatus outside the Palm Grove Sanctuary. Aegomorphus quadrigibbus (Say, 1835:195) RANGE.— Southern Texas and Mexico. ADULT ACTIVITY.— April to July. LARVAL HOSTS. — Castanea, Ficus, Fagus, Tilia, Acer, Car- ya, Cercis, Ulmus, Quercus, Betula, Celtis (Linsley and Chem- sak 1985). DISCUSSION. — We did not encounter this species in the study area, but we collected several specimens matching Knull's (1958) description of the form lucidus from dead Celtis and Acer in Goliad and Bastrop counties in south-central Texas. The genera Acanthoderes and Aegomor- phus contain over 30 species north of Panama and many more in South America, and these species are difficult to separate or define by older descriptions. Aegomorphus quadrigibbus occurs in Mexico and may have been recorded there under other specific names. Knull (1944) col- lected adults from Prosopis near Brownsville, Cameron County (recorded as Psapharochrus). Graphisurus triangulifer (Haldeman, 1847:45) (Figure 17) RANGE.— Ohio to Alabama and Texas. ADULT ACTIVITY. — May to October. LARVAL HOSTS. — Celtis (Leng and Hamilton 1896). DISCUSSION.— Specimens were collected from hackberry and at lights at Welder Wildlife Ref- uge. Schwarz (in Leng and Hamilton 1896) re- ported larvae boring under the bark of Celtis, and Riley (1890) and Craighead (1923) also listed the same host. We found numerous adults on dead and dying hackberry in Goliad and Bastrop counties, and several beetles subsequently emerged from dead Celtis gathered at those sites (FTH). Adults occasionally come to lights. An- tecrurisa apicalis (Bates) from Mexico may be conspecific with G. triangulifer, differing only slightly in the extent of the elytral maculations. HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 317 FIGURE 16. Thryallis undatus. Lagocheirus texensis Dillon, 1956:139 (Figure 17) RANGE.— Western to southern Texas. ADULT ACTIVITY. — May to October. DISCUSSION.— This species was originally de- scribed from material labeled "Dimmit Coun- ty." Vogt (I949a, as L. procerus Casey) recorded two specimens, presumably of texensis, beaten from cut Yucca in Starr County; we also beat this species from dead Yucca, 1 mi [ca. 1 1 km] SW El Sauz, Starr County (FTH). Specimens were taken at lights at Falcon Heights, and A. E. Lewis (pers. comm.) collected it at light near Uvalde, Uvalde County, in western Texas. Dillon did not include texensis in his generic revision (1 957), so the taxonomic position of this species is somewhat uncertain. Lagocheirus tex- ensis is very closely related to, if not synonymous with, L. undatus Voet from Mexico and Central America. NEW LOCALITIES.— Rio Grande City, Starr County. Astylidius parvus (LeConte, 1873:234) RANGE. — Mississippi to southern Texas. ADULT ACTIVITY.— April to August. LARVAL HOSTS.— Ficus (Townsend 1902), Pithecellobium (Turnbow and Wappes 1978), Zanthoxylum (Tumbow and Wappes 1981). wy FIGURE 17. Lagocheirus texensis (left) and Graphisurus triangulifer (right). DISCUSSION.— Most specimens of this greenish longhorn were beaten from dead branches of ebony, persimmon, and hackberry. Vogt (1949a, as A. leiopinus Casey) took three specimens at lights in Pharr, Hidalgo County. NEW LOCALITIES. -PG; LCC; WWR. Leptostylus transversus ssp. DISCUSSION.— A single specimen of this wide- spread, polytypic species was taken from a light trap at Welder Wildlife Refuge in June (RHT), and several similar appearing specimens from dead Acer and Celtis at Goliad, Goliad County, in May (RLP, FTH). They exhibit the general facies of the subspecies dietrichi Dillon (from the southeastern U.S.), but their coloration is more like that of the subspecies asperatus (Haldeman) from central and western Texas. Leptostylus gibbulosus vogti Dillon, 1956:141 RANGE.— Southern Texas, Mexico(?). ADULT ACTIVITY.— December to May. LARVAL HOSTS. — Fruit of Sapindus (Vogt 1949a). DISCUSSION. —The unusual larval habits of this species were reported in detail by Vogt (1949a), who discovered the host to be mature fruits of soapberry. Although he reared large numbers of adults from the fruits, he did not collect them by any other method, and the few additional spec- imens seen by us bear no collecting data. The nominate subspecies occurs from northern Mex- ico to Colombia (Dillon 1 962) where it is com- monly beaten from deadwood. 318 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 FIGURE 18. Sternidius texanus (left), Sternidius mimeticus (middle), and Sternidius wiltii (right). Leptostylopsis luteus Dillon, 1956:147 RANGE.— Southern Texas. ADULT ACTIVITY.— October. DISCUSSION.— This species is very rare in col- lections, and we have seen only two specimens: one from dead Acacia at Welder Wildlife Ref- uge (EFG), the other beaten from dead Bac- charis near Boca Chica, Cameron County (RLP), both in October. The type specimen reportedly came from "Esper Ranch" (Espe- ranza Ranch), near Brownsville, Cameron County. Sternidius wiltii (Horn, 1880:124) (Figure 18) RANGE.— Southern Texas and northern Mexico. ADULT ACTIVITY.— April to October. LARVAL HOSTS.— Acacia (Linsley 1940). DISCUSSION.— Adults of this relatively large Sternidius were beaten from Oncideres-gird\ed twigs and branches, and Linsley (1940) reported rearing them from unspecified legumes girdled by O. pustulatus and from Acacia pruned by O. cingulata texana. We collected numerous spec- imens from girdled huisache at Welder Wildlife Refuge and from drought-stressed mesquite near Boca Chica, Cameron County. Adults have been attracted to lights. NEW LOCALITIES. — PG; Southmost sector, Brownsville, Cameron County; FSP; LCC. Sternidius mimeticus (Casey, 1 89 1 :49) (Figure 18) RANGE.— Texas. ADULT ACTIVITY. — May to October. LARVAL HOSTS.— Leucaena (Hovore and Penrose 1982), Acacia, Celtis. DISCUSSION.— Although S. mimeticus and S. texanus are both abundant throughout the study area on a variety of hosts, published accounts are difficult to correlate with current nomencla- ture. Examination of type specimens has shown that "Liopus houstoni" Casey was correctly placed by Dillon (1956) as a synonym of mimeticus, but L. texanus Casey, also synonymized under mi- meticus, is distinctly different. Texas records for S. crassulus LeConte (a Baja California species) no doubt refer to mimeticus. This species may also be the Leptostylus biustus of Townsend (1902), recorded as infesting fig twigs and dead cotton. Adults were commonly beaten from known larval hosts and numerous other woody plants and were attracted to lights. NEW LOCALITIES. — PG; BC; SAR; Mission, Hidalgo County; Resaca de las Palmas State Park, Cameron County; LCC; WWR; 4 mi [ca. 6.5 km] S Pharr, Hidalgo County; 1 mi [ca. 1.6 km] E Los Indios, Cameron County. Sternidius texanus (Casey, 1913:315) (Figure 18) RANGE.— Southern Texas and northeastern Mexico. ADULT ACTIVITY. — March to October. LARVAL HOSTS.— Leucaena (Hovore and Penrose 1982), Acacia. DISCUSSION.— This species is very closely re- lated to described taxa in the Sternidius alpha complex and may ultimately prove to be syn- onymous with S. naeviicornis Bates from Mexico or S. alpha misellus (LeConte) from the eastern U.S. The genus Sternidius needs taxonomic re- view before names can be applied with certainty to the various phenotypes, particularly from the southeastern U.S. and Mexico. Variation in body coloration and elytral vestiture is extreme in our long series from southern Texas; this series en- compasses most of the phenotypic diversity re- corded for both of the aforementioned species and intergrades broadly with material from west- ern and southern Mexico. Previous .listings of Leiopus alpha (Say) prob- ably refer to this species. NEW LOCALITIES. -PG; 2 mi [ca. 3.2 km] S Pharr, Hidalgo County; Brownsville, Cameron County; WWR. Astyleiopus variegatus (Haldeman, 1847:47) RANGE. — Eastern North America to southern Texas, Utah, and southern Arizona. ADULT ACTIVITY. — May (in the study area). LARVAL HOSTS. — Castanea, Juglans, Morus, Ulmus, Robin- ia, Celastrus (Craighead, 1923), Celtis. DISCUSSION. —A single female was beaten from Celtis at Welder Wildlife Refuge in May (FTH), and numerous adults were taken from fresh-cut logs of this host at Goliad (RLP, FTH). A spec- HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 319 imen was subsequently reared from a larva taken from beneath dead Celtis bark at this locality. Valenus inornate Casey, 1891:50 RANGE. — Northwestern Arizona to northern Mexico. ADULT ACTIVITY. — May to October. DISCUSSION.— No larval habits have been re- corded for this species, but adults are generally associated with Yucca. Numerous specimens were beaten from dead, persistent foliage of Yucca in Zapata and Starr counties, and a series of beetles was collected in western Texas from freshly trimmed leaves of ornamental Agave (FTH). Adults are attracted to lights. NEW LOCALITIES. — 3 mi [ca. 4.8 km] W Roma and 5-7 mi [ca. 8-11 km] SW El Sauz, Starr County; FSP; 8 mi [ca. 13 km] SE Zapata, Zapata County. Dectes texanus aridus Casey, 1913:343 RANGE.— Southern Texas to central Mexico. ADULT ACTIVITY.— April to August. LARVAL HOSTS.— Helianthus. DISCUSSION.— The five subspecies of D. tex- anus are poorly defined and of dubious taxo- nomic value; southern Texas material is phe- notypically intermediate between aridus and the nominotypical taxon. Our placement follows that of Dillon (1956). Larvae of D. texanus sensu latu girdle stems of Compositae, and we beat adults from sunflowers near Mission, Hidalgo County, and from various herbaceous plants along the margins of the palm grove. Vogt (1949a) took specimens from Solidago south of Pharr, Hi- dalgo County; R. H. Turnbow swept a series from Parthenium at Santa Ana National Wildlife Ref- uge, Hidalgo County; and Townsend ( 1 902) beat a specimen from Abutilon near Brownsville, Cameron County. Lepturges angulatus canus Casey, 1913:317 (Figure 19) RANGE.— Eastern to southern Texas, northern Mexico. ADULT ACTIVITY. — March to October. LARVAL HOSTS.— Ficus (Townsend 1902), Pithecellobium, Celtis (Turnbow and Wappes 1978), Leucaena (Hovore and Penrose 1982), Acacia. DISCUSSION.— Linsley and Martin (1933:182) stated that this longhorn was "abundant on every type of tree and shrub," and this beetle is indeed exceedingly common in certain habitats, most notably Celtis-dominated semideciduous wood- lands. Adults are rapid runners and are difficult to collect from the beating sheet with appendages intact. This species, and possibly also the follow- FIGURE 19. Lepturges infilatus (left) and Lepturges angu- latus canus (right). ing, appeared on earlier lists as Lepturges sym- metricus (Haldeman). NEW LOCALITIES. -PG; BC, BRG; 2 mi [ca. 3.2 km] S Pharr, Hidalgo County; SAR; LCC; WWR. Lepturges infilatus Bates, 1872:216 (Figure 19) RANGE.— Southern Arizona and southern Texas to southern Mexico and Panama. ADULT ACTIVITY. — May to October. LARVAL HOSTS.— Leucaena (Hovore and Penrose 1982), Moms. DISCUSSION.— This tropical species was only recently reported from the U.S. (Marqua 1976) from specimens collected at light in southeastern Arizona. South Texan material is lighter in color and more heavily maculate than specimens from Arizona, and it is possible that more than one taxon is being included under the name infilatus. Larvae mine the cambium layer of dead branches of tepehuaje. We took adult specimens on dead mulberry at night (SAR). Adults come to lights. Lepturges infilatus is very similar to the pre- ceding species in general coloration and form, which may account for its having been omitted from previous lists. NEW LOCALITIES. — PG. Lepturges vogti Hovore and Tyson, 1983:349 RANGE.— Southern Texas. ADULT ACTIVITY. — March to October. LARVAL HOSTS.— Yucca (Hovore and Tyson 1983). DISCUSSION.— This is the species Vogt (1949a) recorded from Yucca treculeana in the uplands as "Lepturges sp. near confluens." It is more 320 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 FIGURE 20. Mecas (Dylobolus) rotundicollis (left) and Me- cas (Mecas) marginella (right). closely related to L. yucca Schaeffer (western Texas to Arizona) and L. subglaber Casey (Du- rango, Mexico), from which it differs conspicu- ously by the distinctly patterned elytra and more slender form. Larvae mine dead, persistent leaves of Yucca. Adults were reared and beaten from this host, and collected at lights. NEW LOCALITIES. -BC; LCC; WWR; FSP. Urgleptes celtis (Schaeffer, 19056:168) RANGE. — Southern Texas. ADULT ACTIVITY.— April to October. LARVAL HOSTS.— Leucaena (Hovore and Penrose 1982), Celtis. DISCUSSION.— Schaeffer (1905&) and Linsley and Martin (1933) collected this species from hackberry, and we reared it in large numbers from this host and from tepehuaje. NEW LOCALITIES. -PG; BRG. Urgleptes knulli Dillon, 1956:337 RANGE. — Southern Texas to central Mexico. ADULT ACTIVITY. — May to August. DISCUSSION.— Specimens were taken in the palm grove by beating dead Celtis and miscel- laneous vegetation. This is probably "Lepturges minutus" of Linsley and Martin (1933). Cyrtinus pygmaeus (Haldeman, 1847:42) RANGE. — Eastern North America to Texas. ADULT ACTIVITY. — March to May. LARVAL HOSTS. — Quercus, Carya, Cornus, Liriodendron, Robinia, Acer (Craighead 1923). DISCUSSION.— Vogt (1949a) swept one speci- men from succulent vegetation and took another on Sapindus, both in Hidalgo County. We have not seen those specimens and so have not been able to verify the determination. Mecas (Dylobolus) rotundicollis (Thomson, 18686:196) (Figure 20) RANGE.— Oklahoma to Arizona, Texas, and Mexico, south to Costa Rica. ADULT ACTIVITY. — May to June in southern Texas. DISCUSSION.— Adults of this lampyrid-mimic were common on foliage of capitana ( Verbesina micropterd) at Welder Wildlife Refuge in May (Hovore et al. 1978); frostweed, the common name cited by Hovore et al. for V. microptera, was incorrect. Chemsak and Linsley (1973) recorded a single specimen from Brownsville, Cameron County, and two specimens from Tamaulipas, Mexico. Mecas (Mecas) marginella LeConte, 1873:239 (Figure 20) RANGE.— Southeastern U.S. to New Mexico. ADULT ACTIVITY. — March to May. DISCUSSION.— Several specimens were swept from roadside vegetation 27 mi [ca. 43.5 km] S Catarina, Webb County (AEL). We took nu- merous adults from Compositae in western Tex- as. NEW LOCALITIES.— 1 5 mi [ca. 24 km] SE Three Points, Webb County. Mecas (Mecas) confusa Chemsak and Linsley, 1973:163 RANGE.— Kansas to Texas. ADULT ACTIVITY.— April to June. LARVAL HOSTS.— Heterotheca. DISCUSSION.— We swept adults of this all-gray species from roadside vegetation at Lake Corpus Christi State Park, San Patricio County, and along Highway 77, 41 mi [ca. 66 km] N Raymond ville, Kenedy County (RLP, FTH). We collected lar- vae and adults of M. confusa and M. pergrata from pupal chambers in dead root crowns of Heterotheca sp. (probably subaxillaris), 10 mi [ca. 16 km] S Sarita, Kenedy County, and we took both species from foliage of this plant at a number of localities in south-central Texas. NEW LOCALITIES. — 6 mi [ca. 9.7 km] E Riviera, Kleberg County (TAI); 40 mi [ca. 64.5 km] N Pharr, Hidalgo County (AEL). Mecas (Mecas) cineracea Casey, 1913:360 RANGE.— Southeastern and Great Plains states to southern Rockies, Texas, and northern Mexico. ADULT ACTIVITY.— April to June. LARVAL HOSTS.— Helenium, Baileya (Chemsak and Linsley 1973). DISCUSSION.— This species is common in road- side stands of Compositae throughout central Texas, but Vogt's (1949a) collection from He- HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 321 lenium near Mission, Hidalgo County, is the only south Texas record under the name cineracea. Mecas inornata of Townsend (1902) and Linsley and Martin (1933) may also be this species, but we have not examined their material. Older ac- counts, as well as more recent biological papers (Rogers 1977 b), have broadly applied the no- mina dubia, Mecas inornata (Say), to several different species and may therefore have variously referred to M. confusa, M. cineracea, or M. cana saturnina (see Chemsak and Linsley 1973, for a full discussion of this problem). Mecas cineracea probably utilizes a variety of plants as larval hosts. NEW LOCALITIES. — 3 mi [ca. 4.8 km] N Eagle Pass, Maverick County; 7 mi [ca. 11.3 km] N San Ygnacio, Zapata County; 1-5 mi [ca. 1.6-8 km] NW Jet. Hwy. 35 on Rt. 83; WWR; 24 mi [ca. 39 km] S Sarita, Kenedy County. Mecas (Mecas) pergrata (Say, 1824:407) RANGE.— Great Plains to southeastern U.S., New Mexico, Texas, and northern Mexico. ADULT ACTIVITY.— April to June. LARVAL HOSTS.— Aster (Craighead 1923), Helianthus (Chemsak and Linsley 1973), Heterotheca. DISCUSSION.— Although it is one of the most widespread species of Mecas, M. (M.) pergrata is not particularly common in collections. Craig- head (1923) described the larva and its feeding habits in stems and roots of Aster. Adults were swept from roadside vegetation near Refugio, Refugio County, at Lake Corpus Christi State Park, San Patrick) County, and 11-14 mi [ca. 17.7-22.5 km] S Three Points, Webb County, and a few beetles were taken from pupal cells in roots of Heterotheca 10 mi [ca. 16 km] S Sarita, Kenedy County. Townsend (1902) and Linsley and Martin (1933) recorded collecting pergrata near Brownsville, Cameron County, but we have not seen their material, and they may in part refer to the then-undescribed Mecas linsleyi. NEW LOCALITIES. — Kingsville, Kleberg County (TAI); 24 mi [ca. 39 km] S Sarita, Kenedy County; Freer, Duval County (TAI). Mecas (Mecas) linsleyi Knull, 1975:130 RANGE.— Southern Texas. ADULT ACTIVITY. — March to May. DISCUSSION.— Adults ofM. linsleyi were taken at several localities, always in association with spiny aster (Aster spinosus), which is probably the larval host. It may be distinguished from the similar-appearing M. pergrata by its larger size, longer, all-black antennae, and more elongate prothorax. The type locality is Bentsen-Rio Grande Valley State Park, Hidalgo County. NEW LOCALITIES.— Anzalduas Park, Hidalgo County; 3 mi [ca. 4.8 km] E Rio Grande City, Starr County. Mecas (Mecas) cana saturnina (LeConte, 1859:21) RANGE.— Great Plains to Alabama, Texas, and northern Mexico. ADULT ACTIVITY.— April to August. LARVAL HOSTS.— Ambrosia, Xanthium, Helianthus, Gail- lardia (Chemsak and Linsley 1973). DISCUSSION. — Specimens were taken from roadside stands of Ambrosia and mixed herba- ceous plants at several localities in southern Tex- as, often in company with one or more other Mecas species. NEW LOCALITIES. — 3 mi [ca. 4.8 km] W, 5 mi [ca. 8 km] N Roma, Starr County; 11-14 mi [ca. 17.7-22.5 km] S Three Points, Webb County; 6 mi [ca. 9.7 km] E Riviera, Kleberg County; Padre Island, Kleberg County(?) (TAI). Tetraopes discoideus LeConte, 1858a:26 RANGE.— Rocky Mountain states to Kansas, south to Texas and El Salvador. ADULT ACTIVITY. — May to October. LARVAL HOSTS.— Asclepias spp. (Chemsak 1963). DISCUSSION.— Knull (1948) recorded the col- lection of this widespread and common species on low milkweed in May at Brownsville, Cam- eron County, and Chemsak (1963) listed it from San Benito, Cameron County. We have seen no other south Texas material. Tetraopes texanus Horn, 1878:49 RANGE. — Eastern Oklahoma to western and southern Texas. ADULT ACTIVITY.— April to June. DISCUSSION. — In his review of the genus Tetra- opes, Chemsak (1963) cited no larval hosts for this species, but all Tetraopes species are consid- ered host specific on Asclepias. Adults of T. tex- anus have been taken from foliage and blossoms of several different species of milkweed in both lowland and montane habitats in western and central Texas, but we have not collected it in the study area. Chemsak (1963) listed texanus from Boca Chica, Cameron County and "mouth of Rio Grande." Tetraopes thermophilus Chevrolat, 1861:190, 254 RANGE.— Southern Texas to El Salvador along the tropical belt (fide Chemsak 1963). ADULT ACTIVITY.— August to October. LARVAL HOSTS.— Asclepias. DISCUSSION.— This species was encountered on stems and foliage of milkweed on disturbed sub- strates along roadsides and railroad rights-of-way in Mission, Hidalgo County, in October. Vogt's (19490) record of T. femoratus from Pharr, Hi- 322 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 dalgo County, probably refers to thermophilus. Chemsak (1963) examined specimens from Bee- ville, Bee County; Victoria, Victoria County; and Brownsville, Cameron County. Tetraopes femoratus LeConte, 1847:93 RANGE.— Western and central states to Ohio, south to Mis- sissippi, Texas, and Central America. ADULT ACTIVITY.— June to September. LARVAL HOSTS. —Asclepias (Chemsak 1 963). DISCUSSION.— Two specimens of this wide- spread, polytypic species were swept from road- side vegetation 5-8 mi [ca. 8-13 km] S Guerra, Jim Hogg County, in September (JEW, RHT). This is the southernmost record for femoratus in Texas by about 800 km. Phenotypically, this ma- terial best fits Chemsak's (1963:64) character- ization of the "Great Plains series" of T. femo- ratus. Cathetopteron amoena (Hamilton in Leng and Hamilton, 1 896: 161) RANGE.— Southern Texas. ADULT ACTIVITY.— April to October. LARVAL HOSTS. — Celtis (Turnbow and Wappes 1978). DISCUSSION.— Adults of this beautiful lamiine sun themselves on the upper surfaces of hack- berry leaves. Numerous specimens have been beaten from Celtis foliage or swept from her- baceous plants growing nearby. Portions of the head and thorax described by Hamiton (in Leng and Hamilton 1896) as white are, in living spec- imens, delicate peach-pink (fading postmortem to white). NEW LOCALITIES. — BRG; Anzalduas Park, Hidalgo County; PG. Hemierana marginata (Fabricius. 1798:48) RANGE. — Eastern North America to southern Texas. ADULT ACTIVITY.— April to June. LARVAL HOSTS.— Vernonia (Schwitzgebel and Wilbur 1942). DISCUSSION.— This species is frequently col- lected by sweeping herbaceous vegetation, and many so-called host records represent collections of adults from plants that may not actually serve as larval hosts. Adults were common in May on Verbesina at Welder Wildlife Refuge (Hovore et al. 1978) and Lake Corpus Christi State Park. Schwitzgebel and Wilbur (1942) recorded details of the larval biology in ironweed in Kansas. Hemierana suturalis Linell, 1896:398 RANGE.— Southern Texas, Florida(?). ADULT ACTIVITY. — May to July. LARVAL HOSTS.— Bernardia. DISCUSSION. — Most specimens seen were tak- en by beating or sweeping miscellaneous vege- tation. Townsend (1902) collected several adults by beating tangles of Clematis and Ehretia in the palm grove in June. Specimens in the USNM collection bear data indicating that they were reared from larvae collected in roots and stems of myrtle croton (Bernardia myricaefolid) at Brownsville, Cameron County. We saw specimens of this species that were labeled as coming from Central Florida, and Blatchley (1930) reported beating it from oak in the Everglades (as Amphionycha). If these rec- ords are accurate, the species either has a very unusual distribution or is distributed across the Gulf Arc and simply has not been collected in intermediate areas. QUESTIONABLE RECORDS The following species, which have either been previously recorded from southern Texas or have been encountered in curated material examined during this study, appear to represent either ad- ventitious, misidentified, or mislabeled material. Ergates spiculatus neomexicanus Casey, 1890:491 Linsley (\962a, fig. 8) showed a locality in the Lower Valley, probably based upon specimens from structural timber. Pinus is the larval host. Megaderus bifasciatus Dupont, 1836:5 A single specimen in the Carnegie Museum, Pittsburgh, Pennsylvania labeled as coming from Brownsville, Cameron County, is the only record we saw for southern Texas. This locality may be erroneous, as the remaining six specimens in the Carnegie series are labeled "El Paso." According to Riley (1880) and Beutenmuller (1896), Mega- derus was collected from cedar timber (Junip- erus) in Comal County, Texas, in December and it was recently taken from beneath bark of rotting Pinus in Honduras (Chemsak et al. 1980). We have seen specimens from Comal (USNM) and Bastrop (UCB) counties in central Texas, and from Chihuahua, Mexico (UCB). Callidium texanum Schaeffer, 1917:185 Vogt ( 1 949a) collected two specimens on "ce- dar" fence posts in Hidalgo County, noting that the wood had been imported from northern Tex- as. This beetle breeds in juniper, and the prob- ability of its becoming established in southern HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 323 Texas is therefore very remote. The nearest nat- ural occurrence of C. texanum that we saw was in Sutton County, 8 mi [ca. 1 3 km] SW Roosevelt (FTH, RLP). Strangalepta abbreviate (Germar, 1824:523) Papp (1955) listed this common eastern lep- turine as S. vittata (Olivier) from Brownsville, Cameron County. Linsley and Chemsak (1976) recorded the distribution as reaching only as far south as Georgia. Taricanus truquii Thomson, 1868a:74 Leng and Hamilton (1896), and Dillon and Dillon (1946) recorded this Mexican onciderine as occurring in Texas or the southern U.S., with- out further data. The nearest Mexican locality we have seen thus far is in Veracruz. Ataxia spinicauda Schaeffer, 1904:224 Chemsak and Linsley (1982) listed this Antil- lean species from Florida and Texas. It has often been collected in Florida (Schaeffer 1908; Turn- bow and Hovore 1979), but we have seen no material from Texas. SELECTIVE REARINGS FROM DEADWOOD During the course of this project, F. T. Hovore, R. L. Penrose, R. H. Turnbow, and J. E. Wappes conducted a series of selective rearings of Cer- ambycidae from host plant material gathered at Lake Corpus Christi State Park, Bentsen-Rio Grande Valley State Park, the Palm Grove Sanc- tuary, and in Southmost sector, Brownsville, Cameron County. The results of these rearings, along with a compilation from literature of species utilizing Citrus, are presented below. Species cit- ed as having been taken "on Citrus" have not been included, as these may not be actual rearing records. Both English and Spanish common names, where known, are listed for the host plants. LEGUMINOSEAE Leucaena pulverulenta (Schlect.) Benth. — Lead Tree, Tepe- huaje Achryson surinamum, Geropa concolor, Eburia mutica, Gnaphalodes trachyderoides, Taranomis b. bivittata, Den- drobias mandibularis virens, Stenosphenus lugens, Anela- phus debilis, Neocompsa exclamationis, N. mexicana, Ob- rium maculatum, O. mozinnae, Lochmaeocles c. cornuticeps, Cacostola salicicola, Oncideres pustulatus, Sternidius mi- meticus, S. texanus, Lepturges angulatus canus, L. injilatus, Urgleptes celtis, Thryallis undatus. Prosopis glandulosa Torr. — Mesquite Eburia mutica, E. ovicollis, Knulliana c. cincta, Stenosphe- nus dolosus, Anelaphus debilis, Heterachthes nobilis, Ob- rium mozinnae, Placosternus difficilis, Megacyllene caryae, Neoclytus acuminatus Hesperus, N. augusti, N. mucronatus vogti, Oncideres cingulata texana, Sternidius wiltii, Ecyrus texanus. Acacia farnesiana (L.) Willd.— Sweet Acacia, Huisache Achryson surinamum, Geropa concolor, Gnaphalodes trachyderoides, Taranomis b. bivittata, Stenosphenus dolo- sus, Neocompsa mexicana, Obrium maculatum, Placoster- nus difficilis, Neoclytus acuminatus Hesperus, Oncideres pus- tulatus, O. cingulata texana, Sternidius wiltii, S. mimeticus, S. texanus, Ecyrus texanus. RUTACEAE Citrus paradisi Macf., and C. sinensis L. (Osbeck)— Grapefruit and Sweet Orange (compiled from Dean [1953] and Manley and French [1976]) Archodontes melanopus serrulatus, Stenodontes d. dasyto- mus, Eburia mutica, Gracilia minuta, Gnaphalodes trachy- deroides, Knulliana c. cincta, Dendrobias mandibularis ssp., Enaphalodes taeniatus, Elaphidionoides villosus, Anelaphus inermis, Obrium maculatum, Placosternus difficilis, Neocly- tus acuminatus Hesperus, N. augusti, Euderces reichei exilis, Rhopalophora angustata, R. laevicollis, Oncideres cingulata texana. SALICACEAE Salix nigra Marsh— Black Willow Hypexilis pallida, Elaphidion linsleyi, E. mimeticum, Lep- tura gigas, Lochmaeocles c. cornuticeps, Cacostola salici- cola, Callipogonius cornutus, Ataxia crypto. ULMACEAE Celtis pallida Torr.— Spiny Hackberry, Granjeno Methia constricticollis, Stenosphenus lugens, Piezocera ser- raticollis, Neoclytus augusti, Ancylocera bicolor, Lepturges angulatus canus, Cathetopteron amoena, Urgleptes celtis. Celtis laevigata Willd. — Sugar Hackberry, Palo Blanco Stenodontes d. dasytomus, Eburia stigmatica, E. mutica, Gnaphalodes trachyderoides, Knulliana c. cincta, Dendro- bias mandibularis virens, Neocompsa mexicana, Obrium maculatum, Neoclytus acuminatus Hesperus, Lochmaeocles c. cornuticeps, Oncideres cingulata texana, Sternidius mi- meticus, Urgleptes celtis, Thryallis undatus. Ulmus crassifolia Nutt.— Cedar Elm, Olmo Gnaphalodes trachyderoides, Taranomis b. bivittata, Neo- clytus augusti. ORIGINS AND AFFINITIES OF THE SOUTH TEXAS CERAMBYCIDAE: PALEOECOLOGICAL INFORMATION Although south Texas Cerambycidae are pre- dominantly neotropical in origin, the overall fau- na is a composite of genera derived from nu- merous geographic regions. The complex array of probable routes and times of movement im- 324 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 plies that present faunal concepts may have to be further refined before we completely under- stand the mosaic pattern of species origins and distribution. Independent analyses of the origins of other faunal elements (i.e., reptiles, amphib- ians, and birds) have yielded somewhat differing theories regarding centers of generic differentia- tion and boundaries of present faunal regions. In the following analysis we utilized the faunal sys- tems defined and discussed by Linsley (1939, 1958, 19616, 1 9636) and Halffter ( 1976), which we feel represent the most useful zoogeographic assessments thus far applied to neotropical Co- leoptera. We have combined or modified their concepts only where necessitated by more recent taxonomic and distributional information. (For a more thorough discussion of faunal affinities, regional definitions, and global relationships for North American Cerambycidae, see Linsley [19616].) During the early Cenozoic, tropical flora and fauna from Middle and South America extended over much of what is now North America; south- ern Texas was within the extensive Neotropical Tertiary Geoflora. Insect populations spread along a number of environmental corridors, with many Neotropical forms reaching the southern and eastern portions of the continent via the low- lands of the Gulf Arc corridor. Subsequent periods of glacial maxima, with mesic and xeric interglacial episodes, forced many Neotropical organisms to retreat southward into refugia in the hot, humid lower valleys and del- taic plains of the major river systems along the southern boundary of the North American con- tinental land mass. Relatively mild climate dur- ing the present interglacial period has permitted many species to extend (or reextend) their ranges away from refugial areas, northward into the me- sic eastern deciduous forests, eastward and west- ward across the lowlands of the Gulf Arc, or northward from southern Mexico through the subtropical Mexican forests into Texas. The arid climate now dominating much of eastern Mexico appears to have disrupted the southern portion of the Gulf Arc corridor, and may serve as an interposed ecological limit to the Austroriparian, Tamaulipan, and Mexican Tropical faunal re- gions. Thus, a number of Neotropical genera and species found in southern Texas also occur in the vicinity of Veracruz, Mexico, but do not seem to be present in the intervening portions of the Mexican coastal plain. In addition to repeated terrestrial movements of Neotropical floras and faunas through Texas during shifting climatological regimes, Gulf Stream currents have undoubtedly introduced a number of Central American or Antillean Cer- ambycidae into the Texas fauna via infested driftwood. A number of species with limited dis- tributions in South, Central, or North America and in the Antillean faunal region may have been dispersed into portions of their present ranges by this method (e.g., Desmiphora hirticollis, Pla- costernus difficilis, Anelaphus inermis, Heter- achthes ebenus). Although the rate at which ce- rambycids are transported by floating wood is not known, the oceanic corridor may provide for constant introduction and reintroduction of Neotropical species to the south-coastal portions of North America. The lower Rio Grande valley, with its rich deltaic soils and comparatively hot, humid sub- climate surrounded by more xeric habitats, has repeatedly been a refugium for mesic-adapted neotropical organisms— a retreat from glacial ad- vances as well as the northernmost extension of tropical forms. According to Porter (1977, dis- cussing mesostenine Ichneumonidae), Pleisto- cene climatic alterations produced in southern Texas a multiple overlap of northern and south- ern Neotropical ichneumonid species, and this pattern appears to pertain to the cerambycid fau- na as well. TAXONOMIC AND ANALYTICAL PROBLEMS Genera are often rather subjective taxonomic entities. In certain cerambycid tribes many gen- era appear transitional (e.g., Aneflomorpha and Psyrassd) or polyphyletic (e.g., Deltaspis and Anelaphus). Others have not been treated taxo- nomically since their original description. At- tempting to analyze composite or poorly defined genera can produce confused results, but it is beyond the scope of this project to redefine ge- neric concepts for Neotropical Cerambycidae. Further, genera regarded as arising in a par- ticular faunal region may belong in generic com- plexes with origins or relatives in South America, Eurasia, Africa, or Micronesia. In the following discussion we have, where possible, employed a species aggregate and re- lated genera method of analysis. A concentration of modern species in a single faunal region may suggest that the region represents the ancestral HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 325 home of the genus, particularly in genera with taxonomically and bionomically well-defined species. The distribution of closely related genera may also provide clues to the origin and devel- opmental direction of a genus. The genus An- cylocera Serville is an example: according to Viana (1971), there are six species of Ancylocera in Mexico (one of which, A. macrotela Bates, ex- tends southward to Nicaragua), one species in Colombia, and two species in central and south- ern South America (Brazil, Argentina, and Uru- guay). Viana incuded six other genera in the An- cylocerini, but Chemsak (1967) had removed one of them, Championa, to the Sphaerionini (Ela- phidionini [=Elaphidini] of Chemsak and Lins- ley 1982). Of the remaining five genera, Cera- locyna Viana has two Mexican and six South American species, and Lallancyca Viana has three species, one each in Panama, Brazil, and Argen- tina. The other three genera (Cercoptera Spinola, Callancyla Aurivillius, Corallancyla Tippman) are entirely confined to South America, mainly in Brazil and Argentina. Ancylocera bicolor has in the past been considered as Austroriparian in origin, and as a species, it may well have evolved in one of the southeastern North American re- fugia. However, it is apparent from its congeneric and tribal affinities that the genus is present in the Austroriparian region as a relict of Neotrop- ical Tertiary expansion from Central and South America. As noted earlier, certain genera that appear to be well defined and taxonomically compact with- in a limited geographical region such as North America may, in other portions of their ranges, intergrade so evenly with one another that they form supergeneric phenoclines. Intermediate character states exhibited in some Neotropical generic complexes preclude absolute placement of certain species in any genus; in taxonomically homogeneous groups such as the Elaphidionini (sensu Linsley 1963a), there may exist a virtual continuum of character transition between even the most seemingly disparate genera. For ex- ample, several southwestern species of Ena- phalodes are structurally similar to species in Elaphidion, which is, at one character extreme, close to certain species ofElaphidionoides, which in turn shows intermediacy with Aneflus, Ane- laphus, and Aneflomorpha. These genera are in turn related to Psyrassa, Micropsyrassa, and Stenosphenus (via Aneflomorpha); Meganeflus, Micraneflus, and Neaneflus (via Aneflus); or Gymnopsyra, Peranoplium, Anopliomorpha, and Elaphidionopsis (via Anelaphus). Because there is a transformation series between more deriv- ative genera in a number of tribes, and because relatively little is known concerning the biologies and immature life stages of most Neotropical ce- rambycids, determinations regarding the rate and direction of phylogenetic progression must for now be viewed as speculative. We have, there- fore, used the species aggregate analysis method rather conservatively. Further, certain elements of the Neotropical Cerambycidae are as yet so poorly known tax- onomically as to preclude any meaningful as- sessment without systematic revision. The value of taxonomic refinement in cerambycid faunal analysis was made obvious by the separation of Leptostylopsis from Leptostylus (Dillon 1956). A clear zoogeographic division appeared when the characters used to segregate North American species were applied to Middle American forms. Of the 64 species remaining in Leptostylus (fide Chemsak and Linsley 1982), all but 16 are from North or Central America, while all of the 24 species reassigned to Leptostylopsis are restricted to the West Indies, Florida, or the southern U.S. GENERIC ORIGINS AND AFFINITIES A number of genera from Texas and Mexico either have species distributions that display no distinct faunal affinities, or have their nearest relationships within Old World generic com- plexes. Centers of origin will only be determined by careful study of the beetles and their host plants. Even this approach may not fully resolve the question of generic origin, as Chemsak (1 963) explained in his monographic study of Tet- raopes. Genera distributed widely over a number of faunal regions include Crossidius, Dectes, Hemi- erana, and Mecas; one common attribute of these genera is that they utilize as larval hosts such plant genera as Gutierrezia, Gymnosperma, Haplopappus, Chrysothamnus, Helianthus, As- ter, Ambrosia, and Heterotheca, most species of which are primary invaders of disturbed sub- strates. The very broad distributions of some species may be an artifact of the recent spread of their host plants along road and railway grades and into agriculturally altered habitats. None of the naturally occurring genera in southern Texas are of recent northern origin, but 326 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 one Nearctic genus, Leptura, is represented by a single species, L. (Stenurd) gigas. The presence of a Nearctic lepturine may be explained by an early austral ancestry for the subgenus Stenura. This taxon contains three very closely related species, together displaying a tricentric pattern of Miocene-Pleistocene relictual distributions. L. (S.) emarginata Fabricius is widespread in east- ern North America, apparently having spread from the Carolinian faunal region northward and westward to the Great Plains and New England states, and south to central Florida and east cen- tral Texas. Leptura gigas is confined to the south- ern two-thirds of Texas, and the rarely collected L. (S.) splendens Knull is apparently localized in southeastern Arizona. All three species are very similar in form and coloration, and the known larval habits are nearly identical. Leptura emar- ginata breeds in decaying portions of living hard- wood trees or in old stumps and snags, while L. gigas infests rotting scars, branch butts, and stumps of the riparian tree genus Salix, and to a lesser extent Populus and perhaps Quercus. Thus, it appears that the present species of Ste- nura arose from a common progenitor that be- came dispersed into Pleistocene refugia in the southeastern U.S., Texas, and Arizona. Isolation led to species differentiation, with the derivative taxa redistributing themselves into suitable hab- itats during the recent postglacial (or interglacial) period. Leptura gigas and L. splendens appear to be constrained by the extreme aridity of the re- gions surrounding their present ranges, but L. emarginata has undergone considerable range expansion, spreading through the mesic forests of the eastern and central U.S. The time of arrival of the Stenura progenitor is somewhat proble- matical, but a Holarctic ancestor would probably have been an early entrant to the Neotropical fauna. The only known fossil species of Leptura are found in the Florissant shales of Colorado, indicating that the genus sensu latu was present along the southern boundary of the Arcto-Ter- tiary Geoflora by at least the mid-Oligocene. The only other Nearctic genera recorded from southern Texas (Ergates and Callidium) have been taken only as adventitious emergences from imported coniferous fencing and wood products. Two species with very broad host preferences, Gracilia minuta and Hylotrupes bajalus, are more or less cosmopolitan in distribution, having been spread by commerce into numerous regions in both the Old and New Worlds. The Alleghenian fauna (in the restricted sense defined by Linsley [19616]) is rather poorly rep- resented in southern Texas, with only a single species each in Tylonotus, Pseudostrangalia, and Astyleiopus, and two in Dorcaschema; however, a number of "Alleghenian" species (in otherwise Neotropically distributed genera) such as Tragi- dion coquus, Enaphalodes rufulus, E. atomarius, Elaphidionoides spp., Obrium rufulum, Mega- cyllene caryae, Rhopalophora longipes, Cyrtinus pygmaeus, and Eupogonius pauper reach the study area from the northeast. The Sonoran fauna enters southern Texas from the west, through the arid portions of the north- ern Mexican plateau. Many genera of Sonoran origin are more or less restricted to the Chihua- huan Desert, ranging through northern Mexico into extreme southeastern Arizona, southern New Mexico, and east to western and southern Texas. Other genera are more broadly distributed, rang- ing over the Chihuahuan, Sonoran, and Colo- radan desert regions from Texas to California. Most species in Sonoran genera are associated with hardwoods and leguminous trees and shrubs, many of which are derivatives of several vege- tation types found in the Madro-Tertiary Geo- flora (Axelrod 1 958). According to Halffier (1 976: 8), "the Sonoran cenocron has a two-fold phy- letic-biogeographic origin (ancient South Amer- ican and Paleoamerican) resulting in adaptation to aridity and marked endemism, both of which indicate a strong degree of in situ evolution." Sonoran representatives in the south Texas fauna include Aneflus (sensu stricto), Styloxus, Monei- lema, Eustromula, Taranomis, Plionoma, Vale- nus, and perhaps also Methia andAneflomorpha. The latter two genera have species in the Cali- fornian, Mexican Montane, Mexican Tropical, Austroriparian, and (Methia only) Antillean fau- nal regions. The remainder of the southern Texas Cer- ambycidae are clearly Neotropical, with regions and probable times of phyletic origin ranging from ancient South American to more recent Mexican Plateau faunas. Three monobasic genera— Pyg- maeopsis, Cathetopteron, and Nathriobrium—are presently known only from southern Texas. Na- thriobrium appears to be an isolated relative of the South American genera Necydaliella, Para- leptidea, and Cambaia. The other two genera, although considered to be Tamaulipan endem- ics, are closely related to Middle American (Cathetopteron to various Hemilophini) or Flor- HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 327 idan-Antillean (Pygmaeopsis to Zaplous) genera. Both would be expected to occur in suitable hab- itat in adjacent portions of Mexico. In all, 6 1 of the Neotropical species listed herein are currently known to occur in North America only in the Tamaulipan Biotic Province. Several wide-ranging monobasic genera (Geropa, Gna- phalodes, Ornithid) reach their northernmost distributional limits in southern Texas, and al- together over 70 of the 172 species now known to occur naturally in the study area have their general ranges extending only south into the American tropics. A few genera, such as Elaphidion, Spalacopsis, Cyrtinus, and Leptostylopsis, have the largest number of their species in the Antillean faunal region, and Pentanodes presently contains but two species, dietzii from Texas, and albofasciata Fisher from Cuba. Austroriparian faunal elements extending southwest into southern Texas include species in Archodontes, Plectrodera, Graphisurus, Astyli- dius, and possibly also Knulliana. The majority of the Neotropical cerambycid species in Texas belong to genera extending north from the Mexican Montane or tropical faunal regions of southern Mexico, and Central and South America. Genera reaching southern Texas from the Mexican Montane fauna often have one or more species in the Sonoran region. In some species, such as Cyphonotida laevicollis, allopat- ric subspecies of a single species are found in several different faunal regions. Prionus, Hypex- ilis, Elytroleptus, Ochraethes, Tylosis, Lophalia, Mannophorus, Tetraopes, and possibly Cypho- notida appear to have originated in the Mexican Montane faunal region. Genera with primarily Mexican Tropical (Me- soamerican region of Halffter [1976]) distribu- tions include Psyrassa, Parevander, Ancylocera, Parmenonta, Thryallis, Strangalia, and possibly also Obrium, Euderces, and Stenosphenus. The latter three genera have species in the Vancou- veran (Obrium only), Sonoran, Mexican Mon- tane, Austroriparian, and Alleghenian faunal re- gions. Central and South American genera extending northward through Mexico or across the Carib- bean into southern Texas (and rarely into other faunal regions as well) include Sphaerion, Pie- zocera, Neocompsa, Tetranodus, Dihammo- phora, Dendrobias, Lissonotus, Megaderus, Dor- casta, Desmiphora, and Cacostola. The genera Parandra, Stenodontes, Smodicum, Achryson, Eburia, Heterachthes, Neoclytus, Rhopalophora, Neoptychodes, Adetus, Ataxia, Eupogonius, On- cideres, Lepturges, Urgleptes, Hippopsis, and Leptostylus are Pan-American in distribution, occurring collectively from the Californian and Sonoran regions to the Floridan-Antillean re- gion, through the West Indies and into portions of Central and South America. SUMMARY The longhorned wood-boring beetles (Coleop- tera: Cerambycidae) of southern Texas have been the subjects of entomological investigations for nearly a century, beginning with a brief collecting account by E. A. Schwarz in 1896. Since then, no fewer than seven species lists have been com- piled, providing records and distributional data for approximately 1 00 species. With the addition of the data contained in the present list, the total number of species naturally occurring in the southern portion of Texas stands at 178. The study area considered herein encompasses a larger geographical area than did most prior accounts; it roughly corresponds to the Texan portions of the Matamoran and Nuecian districts of the Tamaulipan Biotic Province. Collections and field observations for the project were con- centrated in the drainage of the Rio Grande Riv- er from Zapata County to Cameron County (with particular emphasis upon remnant forest habi- tats), and the southern Gulf Coast woodlands (most notably at Lake Corpus Christi State Park and Welder Wildlife Foundation Refuge, San Pa- tricio County). Remnant forest habitats in the lower Rio Grande valley are now almost entirely restricted to parks and preserves, the bulk of most original native floral communities having been elimi- nated by agriculture and urbanization. Although the sanctuaries are protected from further direct environmental degradation, most are continual- ly subjected to unnatural stress, from outside ele- ments such as agricultural chemical drift and fluctuating water tables, and from within by cer- tain resource management practices. Older in- terior swamp and hardwood forests, as well as some semideciduous forests and brushland com- munities, are overmature and appear to be de- clining. Normal cyclical and successional pro- cesses no longer occur within the refugia, and many areas exhibit community senescence and 328 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 lowered species diversity and abundance. Geo- graphical isolation in most of the refugia also contributes to the apparent loss of species di- versity, by limiting genetic exchange and by mak- ing species vulnerable to ecological catastrophes arising from otherwise natural successional and cyclical events. Fire, flood, protracted drought, or severe frost could alone or in combination eliminate sensitive species from isolated habi- tats, with no natural pathways available for re- colonization from other refugia. Vast tracts of brushland and savanna-wood- land habitat remain in the upland portions of southern Texas, but even these have been sub- stantially altered by long-term cattle grazing. This, combined with the introduction of exotic grasses, has changed community compositions and spa- tial relationships, favoring the spread and in- creased density of disturbed-land plant genera such asAbutilon, Haplopappus, and Viguiera. In- creases in relative abundance and overall distri- butions of longhorned beetles associated with these plants have also been observed. Reduction in the number and diversity of host plant species has no doubt already led to the decline or extinction of certain oligophagous species of insects. Mitigating this situation, how- ever, is the polyphagy exhibited by a number of Cerambycidae, which permits exploitation of al- ternative native hosts and introduced plant species. Twenty species of cerambycids have been reared from girdled limbs of tepehuaje (Leu- caena pulverulentd), indicating that such natural polyphagy exists, and 1 5 genera of Cerambycidae have been recorded as breeding in wood of grape- fruit and sweet orange (Citrus spp.). Adult cerambycid periodicity is distinctly bi- modal; peak spring and fall activity coincides with moderate temperatures and increased pre- cipitation. Summer-active forms are typically nocturnal genera of Sonoran faunal origin. Drought may inhibit or delay emergence of adults of deadwood-boring species, but appears to have less effect upon species breeding within living hosts. Distributional data indicate that 170 ceram- bycid species, in 103 genera, naturally occur in southern Texas. The subfamily Cerambycinae comprises over half the fauna with 95 species in 56 genera; Lamiinae are represented by 65 species in 38 genera; Prioninae by 5 species in 4 genera; Lepturinae by 4 species in 4 genera; and Par- andrinae by a single species of Parandra. The genus containing the greatest number of south Texan species (7) is Mecas, while 1 7 genera have 3 or more species in the study area. Although primarily Neotropical in overall or- igin, southern Texas Cerambycidae also show elements of the Nearctic, Alleghenian, and So- noran faunal regions. Endemism is fairly pro- nounced at the species level, with 6 1 taxa (about 35% of the total) known in the U.S. only from the Tamaulipan Biotic Province. There are 3 monobasic genera, which are at present known only from the lower Rio Grande valley: Pyg- maeopsis, Cathetopteron, and Nathriobrium. The Sonoran fauna enters southern Texas from the west via arid portions of the northern Mex- ican Plateau and is represented by 1 1 species in 7 genera. Only 4 Alleghenian genera occur in the region, but a number of "Alleghenian" species in otherwise clearly Neotropical genera extend into southern Texas from the east. The Nearctic fauna is represented by a single species of Lep- tura, displaying a relictual Miocene-Pleistocene distributional pattern. Gracilia minuta and Hy- lotrupes bajulus are widely distributed by com- merce and have essentially cosmopolitan distri- butions. The Neotropical species have arisen from a number of apparent centers of evolutionary di- versification, including the Mexican Montane, Mexican Tropical, and Central and South Amer- ican faunal regions, and to a lesser extent, the Antillean region. Seventeen genera are thought to be Pan-American in overall distribution. ACKNOWLEDGMENTS We wish to express our sincere appreciation to the following persons and institutions for their assistance with this project: E. Gorton Linsley and John Chemsak, University of California, Berkeley (UCB), for advice, criticism, manu- script review, and generosity with specimens and data; David H. Riskind, Texas Department of Parks and Wildlife, for permission to survey park properties; John Anderson and Ernest Ortiz, Na- tional Audubon Society, for permission to work in the Palm Grove Sanctuary; W. C. Glazener, Eric Bolen, and Lynn Drawe, Welder Wildlife Foundation, for hospitality and numerous cour- tesies extended to us during our studies at the refuge; John Beall, U.S. Fish and Wildlife Ser- vice, for permission to collect in the Santa Ana Refuge; J. E. Gillaspy, Texas A&I University HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 329 (TAI); H. R. Burke, Texas A&M University, Col- lege Station (TAM); R. Lonard and T. Allison, Pan-American University, Edinburg, Texas (PAU); G. Wallace, Carnegie Museum, Pitts- burgh, Pennsylvania; T. J. Spilman and R. Gor- don, Smithsonian Institution, U.S. National Mu- seum (USNM); R. R. Snelling, Los Angeles County Natural History Museum (LACM), for access to specimens and data from collections in their care; Edmund Giesbert (EFG) and Dave Carlson (DCC) for data, camaraderie, and in- valuable field assistance; James Cope (JC), Ru- dolph Lenczy (RL), Art Lewis (AEL), Ted C. Macrae (TCM), Marlin Rice (MER), William Tyson (WHT), and George Vogt for field notes and specimens from their collections; and Robert H. Turnbow, Jr. (RHT) and James E. Wappes (JEW) for specimen, field, and rearing data and for comments and helpful suggestions generously contributed. LITERATURE CITED AUDINET-SERVILLE, J. G. 1834. Nouvelle classification de la famille des longicornes (suite). Ann. Soc. Entomol. France (Ser. 1), Vol. 3:5-110. AUSTIN, E. P. 1880. Supplement to the checklist of the Co- leoptera of America north of Mexico. Boston. 67 pp. AXELROD, D. I. 1958. Evolution of the Madro-Tertiary Geo- flora. Bot. Rev. 24:433-509. BATES, H. W. 1872. On the longicorn Coleoptera of Chon- tales, Nicaragua. Trans. Entomol. Soc. Lond. 1872(3): 163- 238. . 1879-1886. Biologia Centrali-Americana, Insecta, Coleoptera, Longicornia. Vol. 5. xii + 436 pp. BEUTENMULLER, W. 1896. Food habits of North American Cerambycidae. J. New York Entomol. Soc. 4:73-81. BITTENFELD, H. W. H. VON. 1948. Aromia moschata (L.) als Spinnen-eter. Entomol. Berl. 12:232. BLACKWELDER, R. E. 1946. Checklist of the Coleopterous insects of Mexico, Central America, the West Indies and South America, Pt. 4. U.S. Nat. Mus. Bull. No. 185:551- 763. BLAIR, F. W. 1950. The Biotic Provinces of Texas. Texas J. Sci. 2:93-117. . 1952. Mammals of the Tamaulipan Biotic Province in Texas. Texas J. Sci. 2:230-250. BLATCHLEY, W. S. 1930. Notes on the distribution of Co- leoptera in Florida with new additions to the known fauna of that state. Can. Entomol. 62:28-35. Box, T. W. 1961. Relationships between plants and soils of four range plant communities in south Texas. Ecology 42: 794-810. Box, T. W. AND A. D. CHAMRAD. 1966. Plant communities of the Welder Wildlife Refuge. Contribution 5, Ser. B, Weld- er Wildlife Foundation. 28 pp. BREUNING, S. 1974. Revision des Rhodopinini Americains. Studia Entomol. 17:1-210. BROWN, D. E., C. H. LOWE, AND C. P. PASE. 1980. A digitized systematic classification for ecosystems with an illustrated summary of the natural vegetation of North America. U.S.D.A. For. Serv. Gen. Tech. Rept. RM-73. 93 pp. BUQUET, J. B. L. 1840. Description de quelques coleopteres longicornes appartenant aux genres Calocosmus, Stenaspis et Galissus. Rev. Zool., p. 142. CARR, J. T. JR. 1967. Hurricanes affecting the Texas Gulf Coast. Texas Water Dev. Bd. Rept. 49:1-58. CASEY, T. L. 1890. Coleopterological notices, II. Ann. New York Acad. Sci. 5:307-504. . 1891. Coleopterological notices, III. Ann. New York Acad. Sci. 6:9-214. . 1893. Coleopterological notices, V. Ann. New York Acad. Sci. 7:28 1-606. . 1912. Studies in the Longicornia of North America. Memoirs on the Coleoptera 3:215-375. . 1913. Further studies among the American Longi- cornia. Memoirs on the Coleoptera 4:193-400. . 1 924. Additions to the known Coleoptera of North America. Memoirs on the Coleoptera 1 1:1-347. CHEMSAK, J. A. 1963. Taxonomy and bionomics of the genus Tetraopes. Univ. Calif. Publ. Entomol. 30(1): 1-90. . 1967. Review of the genus Championa Bates. Pan- Pacific Entomol. 43(l):43-48. CHEMSAK, J. A. AND E. G. LINSLEY. 1963. Synopsis of the known Mexican species ofAneflus. Bull. Brooklyn Entomol. Soc. 63:80-96. . 1965. New genera and species of North American Cerambycidae. Pan-Pacific Entomol. 41(3):141-153. . 1970. Death-feigning in North American Ceram- bycidae. Pan-Pacific Entomol. 46(4):305-307. . 1973. The genus Mecas LeConte. Proc. Calif. Acad. Sci., 4th Ser. 39(12):141-184. . 1975a. Checklist of the beetles of Canada, United States, Mexico, Central America and the West Indies. Vol. 1, Pt. 6, The longhorn beetles and the family Disteniidae. ("Red Version") Biol. Res. Inst. Amer. Latham, New York. 224 pp. . 19756. Mexican Pogonocherini. Pan-Pacific Ento- mol. 51(4):271-286. . 1 982. Checklist of Cerambycidae, the longhorned beetles. Checklist of the Cerambycidae and Disteniidae of North America, Central America and the West Indies. Plex- us Publ. Inc., Medford, New Jersey. 138 pp. CHEMSAK, J. A., E. G. LINSLEY, AND J. V. MANKINS. 1980. Records of some Cerambycidae from Honduras. Pan-Pacific Entomol. 56(l):26-37. CHEVROLAT, L. A. A. 1834. Coleopteres du Mexique. Fasc. 3: 48 pp. . 1835. Coleopteres du Mexique. Fasc. 4: 70 pp. . 1 849. In d'Orbigny, Dictionnaire universe! d'histoire naturelle Vol. 12. 816 pp. . 1859. Description d'un genre nouveau etabli aux depens de plusieurs especes de Rhopalophora de Dejean. Arcana Naturae 1:50-54. . 1860. Description d'especes de Clytus propres au Mexique. Ann. Soc. Entomol. France, Ser. 3, 8:451-504. . 1861. Reflexions et notes synonymiques sur le travail de M. J. Thomson sur les cerambycides avec descriptions de quelques nouvelle especes. J. Entomol. 1:185-192, 245- 254. . 1862. Coleopteres de 1'Ile de Cuba. Notes, synony- mies et descriptions d'especes nouvelles. Families des ce- rambycides et des parandrides. Ann. Soc. Entomol. France, Ser. 4, 2:245-280. 330 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 CLOVER, E. U. 1937. Vegetational survey of the lower Rio Grande valley, Texas. Madrono 4:41-66, 77-100. COFFEY, G. L. 1909. Reconnaissance soil survey of south Texas. Pp. 1029-1 130 in Field operations of the Bureau of Soils, U.S.D.A., Washington, D.C. CORRELL, D. S. AND M. C. JOHNSTON. 1970. Manual of the vascular plants of Texas. Texas Research Foundation, Ren- ner, Texas. 1881 pp. CRAIGHEAD, F. C. 1 923. North American cerambycid larvae. Can. Dep. Agric. Bull, (n.s.) 27:1-239. . 1950. Insect enemies of eastern forests. U.S.D.A. Misc. Pub. 657:1-679. DAVIS, A. M. 1942. A study of bosque de la palma in Cam- eron County, Texas, and ofSabal texana. M.S. Dissertation (unpubl.), Univ. Texas, Austin. DEAN, H. A. 1953. Long-horned beetles that attack Citrus in the lower Rio Grande valley of Texas. J. Econ. Entomol. 46(1): 174. DILLON, L. S. 1956. The nearctic components of the tribe Acanthocinini Pts. 1, 2 and 3. Ann. Entomol. Soc. Am. 49: 134-167,207-235,332-355. . 1957. Revision of the neotropical Acanthocinini, II. The genus Lagocheirus. Bull. British Mus. (N.H.) Entomol. 6(6):141-168. . 1962. Additional notes on nearctic Acanthocinini. Coleop. Bull. 16:31-32. DILLON, L. S. AND E. S. DILLON. 1941. The tribe Mono- chamini in the western hemisphere. Reading Public Mus. and Art Gallery, Sci. Publ. 1:1-135. . 1946. The tribe Onciderini, Pt. II. Reading Public Mus. and Art Gallery, Sci. Publ. 6:189-413. . 1 948. The tribe Dorcaschematini. Trans. Am. Ento- mol. Soc. 73:173-298. . 1953. A change of names in the Cerambycidae with other notes. Entomol. News 64:260-261. DRURY, D. 1773. Illustrations of natural history (exotic in- sects), Vol. 1 . London. 1 30 pp. DUFFY, E. A. J. 1960. A monograph of the immature stages of neotropical timber beetles (Cerambycidae). British Mu- seum (N.H.), London. 327 pp. DUPONT, H. 1836. Monographic des Trachyderides. Mag. Zool. 6(cl. 9): 1-51. . 1838. Monographic des Trachyderides de la famille des longicornes. Mag. Zool. 8:29-59. EISNER, T., F. C. KAFATOS, AND E. G. LINSLEY. 1962. Lycid predation by mimetic adult Cerambycidae. Evolution 1 6(3): 316-324. FABRICIUS, J. C. 1781. Species Insectorum, Vol. 1. 552 pp. . 1792. Entomologica systemica, Vol. 1. 330 + 538 pp. . 1798. Supplementum entomologiae systematicae. 572 pp. . 1801. Systema eleutheratorum, Vol. 2. 687 pp. FATTIG, P. W. 1947. The Cerambycidae or long-horned bee- tles of Georgia. Emory Univ. Mus. Bull. 5:1-48. FERRIS, T. 1980. Hopeful prophet who speaks for human aspiration. Smithsonian Mag. 1 1(1):127-142. FISHER, W. S. 1914. A new species of Callichroma from Texas. Proc. Entomol. Soc. Wash. 16(3):97-98. . 1 924. A new species of Ataxia from the United States. Can. Entomol. 56:253-254. . 1931. New cactus beetles, III. Proc. Entomol. Soc. Wash. 33(8): 197-201. FLEETWOOD, R. J. 1973. Plants of Santa Ana National Wild- life Refuge. Santa Ana NWR, Alamo, Texas. 55 pp. Fox, L. R. AND P. A. MORROW. 1981. Specialization: species property or local phenomenon? Science 21 1:887-893. FUCHS, T. W. AND J. A. HARDING. 1976. Seasonal abundance of arthropod predators in various habitats in the lower Rio Grande valley of Texas. Environ. Entomol. 5(2):288-290. GAHAN, C. J. 1892. Additions to the longicornia of Mexico and Central America, with notes on some previously- recorded species. Trans. Entomol. Soc. Lond. 1892:255-274. . 1908. Notes on North American longicornia, with descriptions of some new species. Ann. Mag. Nat. Hist., Ser. 8, 1:140-145. GERMAR, E. F. 1824. Insectorum species novae aut minus cognitae, descriptionibus illustratae. Halae. 624 pp. GIESBERT, E. F. AND F. T. HOVORE. 1976. Records and de- scriptions of some southwestern Cerambycidae. Coleop. Bull. 30(l):95-99. GIESBERT, E. F. AND R. L. PENROSE. 1984. Two new pur- puricenine longhorns from the Tamaulipan Biotic Province. Coleop. Bull. 38:(l):59-65. GOSLING, D. C. L. 1981. Ecology of the Cerambycidae in a southwestern Michigan woodland. Ph.D. Dissertation (un- publ.), Univ. Michigan. 142 pp. GUNTER, G. AND H. H. HILDEBRAND. 1951. Destruction of fishes and other organisms on the south Texas coast by the cold wave of January 28-February 3, 1951. Ecology 32:731- 736. HADDOCK, D. J. 1963. The recurrent threat of cold to crops in the lower Rio Grande valley: an interpretation of records. J. Rio Grande Valley Hortic. Soc. 17:178-184. HALDEMAN, S. S. 1845. On several genera and species of insects. Proc. Acad. Nat. Sci. Phila. 3:124-128. . 1847. Materials towards a history of the Coleoptera Longicornia of the United States. Trans. Am. Philos. Soc. 10:27-66. HALFFTER, G. 1976. Distribucion de los insectos en la zona de transicion Mexicana. Relaciones con la entomofauna de Norteamerica. Folia Entomol. Mexicana 35:1-64 [pp. 1-4 unnumbered as printed]. HARRIS, P. AND G. L. PIPER. 1970. Ragweed (Ambrosia spp.: Compositae): its North American insects and the possibil- ities for its biological control. Tech. Bull. 13, Common- wealth Inst. Biol. Control. Pp. 117-140. HAWKER, H. W., M. W. BECK, AND R. E. DEVEREUX. 1925. Soil survey of Hidalgo County, Texas. Bur. Chem. & Soils, Texas Agric. Exp. Stn. College Station, Texas. 59 pp. HiGH,M.M. 1915. The huisachegirdler. U.S.D.A. Bull. 184: 1-9. HORN, G. H. 1860. Descriptions of new North American Coleoptera, in the cabinet of the Entomological Society of Philadelphia. Proc. Acad. Nat. Sci. Phila. 12:569-571. . 1878. Notes on some genera of Cerambycidae of the United States. Trans. Am. Entomol. Soc. 7:41-50. . 1880. Notes on some genera of Cerambycidae with description of new species. Trans. Am. Entomol. Soc. 8:115- 138. . 1885. Descriptions of some new Cerambycidae with notes. Trans. Am. Entomol. Soc. 12:173-197. HORTON, J. R. 1917. Three-lined fig-tree borer. J. Agric. Res. 11(8):37 1-382. HOVORE, F. T. 1980. A new genus and species of Ceram- bycidae from southern Texas. Coleop. Bull. 34(1 ):1 15-1 19. HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 331 . 1983. Taxonomic and biological observations on southwestern Cerambycidae. Coleop. Bull. 37(4):379-387. HOVORE, F. T. AND E. F. GIESBERT. 1976. Notes on the ecol- ogy and distribution of western Cerambycidae. Coleop. Bull. 30(4):349-360. HOVORE, F. T. AND R. L. PENROSE. 1982. Notes on Cer- ambycidae co-inhabiting girdles of Oncideres pustulata LeConte. Southwestern Nat. 27(l):23-27. HOVORE, F. T., PENROSE, R. L., AND E. F. GIESBERT. 1978. Notes on North American Cerambycidae. Entomol. News 89(2&3):95-100. HOVORE, F. T. AND W. H. TYSON. 1983. A new species of Lepturges from southern Texas. Coleop. Bull. 3 7(4): 349- 352. HUBBARD, H. G. 1885. Insects affecting the orange. U.S.D. A. Div. Entomol. Publ., Washington, D.C. 227 pp. HUDEPOHL, K. E. 1985. Revision der Trachyderini. Entomol. Arbeiten Mus. Frey, 33/34, 1985:1-167. HUFFMAN, F. R. AND J. A. HARDING. 1980. Pitfall collected insects from various lower Rio Grande valley habitats. Southwestern Entomol. 5(l):33-46. INGLIS, J. M. 1964. A history of vegetation on the Rio Grande plain. Texas Parks and Wildlife Dept. Bull. 45:1-122. JOHNSTON, M. C. 1952. Vegetation of eastern Cameron County, Texas. Masters Thesis (unpubl.), Univ. Texas, Aus- tin. 127 pp. . 1963. Past and present grasslands of southern Texas and northeastern Mexico. Ecology 44:456-466. KERBEY, M. 1939. The Texas delta of an American Nile. National Geographic Mag. 75(l):51-88. KNULL, J. N. 1937. New southwestern Buprestidae and Cer- ambycidae with notes. Ohio J. Sci. 37(5):30 1-308. . 1 944. New Coleoptera with notes. Ohio J. Sci. 44(2): 90-93. . 1946. The long-horned beetles of Ohio. Ohio Biol. Survey Bull. 39, 7(4): 133-354. . 1 948. New genus and species of Cerambycidae with notes. Ohio J. Sci. 48(2):82-83. . 1954. A new North American Eupogonius with note. Entomol. News 65:127-128. -. 1958. One new species and one subspecies of Cer- ambycidae from Texas. Ohio J. Sci. 58(5):282. . 1 960. A new species ofElaphidion from Texas. Ohio J. Sci. 60(1):7. . 1966. Two new North American species of Smodi- cum Lacordaire. Entomol. News 77(5): 136-1 39. . 1975. A new species of Mecas from Texas. Ohio J. Sci. 75(3): 130-1 31. LEBLANC, R. J. 1958. Sedimentology of south Texas. Gulf Coast Association of Geological Society Handbook, Hous- ton. 59 pp. LECONTE, J. L. 1 847. Fragmenta entomologica. J. Acad. Nat. Sci. Phila. (n.s.) 1:71-93. . 1850-1852. An attempt to classify the longicorn Co- leoptera of the part of America north of Mexico. J. Acad. Nat. Sci. Phila., Ser. 2, 2:5-38, 99-112, 139-178. -. 1853. Descriptions of twenty new species of Coleop- tera inhabiting the United States. Proc. Acad. Nat. Sci. Phila. 6:226-235. . 1854a. Notice of some coleopterous insects, from the collections of the Mexican Boundary Commission. Proc. Acad. Nat. Sci. Phila. 7:79-85. . 1854ft. Descriptions of some new Coleoptera from Texas, chiefly collected by the Mexican Boundary Commis- sion. Proc. Acad. Nat. Sci. Phila. 6, 1853(1854):439-448. -. 1858a. Catalogue of Coleoptera of the regions ad- jacent to the boundary line between the United States and Mexico. J. Acad. Nat. Sci. Phila., Ser. 2, 4:9-42. -. 1858ft. Descriptions of new species of Coleoptera, chiefly collected by the United States and Mexican Boundary Commission, under Major W. H. Emory, U.S.A. Proc. Acad. Nat. Sci. Phila. 10(1858):59-89. . 1859. The Coleoptera of Kansas and eastern New Mexico. Smiths. Contrib. to Knowledge 1 1:1-58. -. 1861. New species of Coleoptera inhabiting the Pa- cific district of the United States. Proc. Acad. Nat. Sci. Phila. 13:338-359. . 1862. Note on the classification of Cerambycidae, with descriptions of new species. Proc. Acad. Nat. Sci. Phila. 14:38-43. . 1873. New species of North American Coleoptera prepared for the Smithsonian Institution. Smiths. Misc. Coll. 264(2): 169-348. . 1 884. Short studies of North American Coleoptera (No. 2). Trans. Am. Entomol. Soc. 12:1-32. LENG, C. W. 1886. Synopses of the Cerambycidae (cont.). Entomol. Am. 2:27-32. . 1 890. Synopses of the Cerambycidae (cont.). Ento- mol. Am. 6:9-13. LENG, C. W. AND J. HAMILTON. 1896. Synopsis of the Cer- ambycidae of North America, Pt. Ill, the Lamiinae. Trans. Am. Entomol. Soc. 23:101-178. LINE, L., ED. 1978. The National Audubon Society Sanctu- aries: Sabal Palm Grove. Audubon Mag. 80(1): 198 [no au- thor credited]. LINELL, M. L. 1896. Descriptions of new species of North American Coleoptera in the families Cerambycidae and Scarabaeidae. Proc. U.S. Nat. Mus. 19(1 1 13):393^401. LINNAEUS, C. 1758. Sy sterna naturae par regna tria naturae secundum classes, ordines, genera, species, cum character- ibus, differentiis synonymis, locis. Ed. 10, Vol. 1. Holmiae. 823 pp. . 1767a. Systema naturae, Vol. 1 , pars 2, editio duo- decima reformata. Holmiae. Pp. 533-1327. . 1767ft. [also listed as 1771] Mantissa plantarum. Holmiae. 588 pp. LINSLEY, E. G. 1930. New Pogonocherus and Ecyrus with notes concerning others. Pan-Pacific Entomol. 7(2):77-90. [A correction notice to this paper appeared Ibid. 7(3): 106, E. G. Linsley.] . 1 934. Notes and descriptions of west American Cer- ambycidae. Entomol. News 45:161-165, 181-185. . 1935a. A revision of the Pogonocherini of North America. Ann. Entomol. Soc. Am. 28(1):73-103. -. 1935ft. Notes and descriptions of west American Cerambycidae, Pt. II. Entomol. News 46:161-166. -. 1936. Preliminary studies in the North American Phoracanthini and Sphaerionini. Ann. Entomol. Soc. Am. 29(3):46 1-479. -. 1939. The origin and distribution of the Cerambyc- idae of North America, with special reference to the fauna of the Pacific slope. Proc. 6th Pac. Sci. Congr. 4:269-282. . 1940. Notes on Oncideres twig girdlers. J. Econ. Entomol. 33(3):56 1-563. . 1 957a. New subspecies of Cerambycidae mostly from the southwestern United States. Coleop. Bull. 1 1:33-36. 332 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 . 1957ft. Some new genera and species of North Amer- ican Cerambycidae. Can. Entomol. 89(6):283-287. -. 1958. Geographical origins and phylogenetic affini- ties of the cerambycid beetle fauna of western North Amer- ica. Publ. Am. Assoc. Adv. Sci. (Zoogeography) 5 1 :299-320. . 196 la. A reclassification of the described Mexican and Central American Sphaerionine Cerambycidae. Pan-Pa- cific Entomol. 18:1-97. . 1 96 1 ft. The Cerambycidae of North America, Pt. 1 , Introduction. Univ. Calif. Publ. Entomol. 18:1-97. . 1 962a. The Cerambycidae of North America, Pt. 2, Taxonomy and classification of the Parandrinae, Prioninae, Spondylinae and Aseminae. Univ. Calif. Publ. Entomol. 19: 1-102. -. 1962ft. The Cerambycidae of North America, Pt. 3, Taxonomy and classification of the subfamily Cerambyci- nae, tribes Opsimini through Megaderini. Univ. Calif. Publ. Entomol. 20:1-188. . 1963a. The Cerambycidae of North America, Pt. 4, Taxonomy and classification of the subfamily Cerambyci- nae, tribes Elaphidionini through Rhinotragini. Univ. Calif. Publ. Entomol. 21:1-165. . 1963ft. The characteristics and history of the North American fauna: longhorned beetles. Proc. 1 6th Int. Congr. Zool. 4:20-27. -. 1964. The Cerambycidae of North America, Pt. 5, Taxonomy and classification of the subfamily Cerambyci- nae, tribes Callichromini through Ancylocerini. Univ. Calif. Publ. Entomol. 22:1-197. LINSLEY, E. G. AND M. A. CAZiER. 1962. A note on the attraction of Stenaspis solitaria (Say) and other insects to Senecio longilobus, a range plant highly toxic to livestock. Can. Entomol. 94(7):745-748. LINSLEY, E. G. AND J. A. CHEMSAK. 1961. A distributional and taxonomic study of the genus Crossidius. Misc. Publ. Entomol. Soc. Am. 3(2):25-64. . 1976. The Cerambycidae of North America, Pt. 6, No. 2, Taxonomy and classification of the subfamily Lep- turinae. Univ. Calif. Publ. Entomol. 80:1-186. . 1985. The Cerambycidae of North America, Pt. 7, No. 1 , Taxonomy and classification of the subfamily La- miinae, tribes Parmenini through Acanthoderini. Univ. Cal- if. Publ. Entomol. 102:1-258. LINSLEY, E. G., J. N. KNULL, AND M. STATHAM. 1961. A list of Cerambycidae from the Chiricahua Mountain area, Coch- ise County, Arizona. Am. Mus. Nat. Hist. Novitates 2050: 1-34. LINSLEY, E. G. AND J. O. MARTIN. 1933. Notes on some longicorns from subtropical Texas. Entomol. News 44:178- 183. MACCLINTOCK, L., R. F. WHITCOMB, AND B. L. WHITCOMB. 1977. II: Evidence for the value of corridors and minimi- zation of isolation in preservation of biotic diversity. Am. Birds 31(1):6-12. MANLEY, G. V. AND J. V. FRENCH. 1976. Wood boring beetles inhabiting Citrus in the lower Rio Grande valley of Texas, Pt. 1, Cerambycidae. J. Rio Grande Valley Hortic. Soc. 30: 45-53. . 1977. A Neoclytus new to the United States. Ento- mol. News 88:39-40. MANN, J. 1969. Cactus-feeding insects and mites. U.S. Nat. Mus. Bull. 256:1-158. MARQUA, D. G. 1976. A new record of Acanthocinine ce- rambycid from America north of Mexico. Pan-Pacific Ento- mol. 52(2): 137. MARTINS, U. R. 1970. Monografia da tribo Ibidionini, Pts. 4 and 5. Arquivos de Zool. Vol. 16, fasc. 4:879-1 149; fasc. 5:1151-1342. . 1975. A taxonomic revision of the world Smodicini. Arquivos de Zool. Vol. 26, fasc. 4:319-359. . 1976. Sistematica e evolucao da tribo Piezocerini Arquivos de Zool. Vol. 27, fasc. 3/4:165-370. MARTINS, U. R. AND J. A. CHEMSAK. 1966. Synopsis of the known Mexican Ibidionini. J. Kansas Entomol. Soc. 39(3): 454_467. MAYR, E. 1954. Change of genetic environment and evolu- tion. Pp. 157-180 in Evolution as a process, J. Huxley, A. C. Huxley, and E. B. Ford, eds. Allen and Unwin, London. MILLIKEN, F. B. 1916. The cottonwood borer. U.S.D.A. Bur. Entomol. Bull. 424:1-7. MOORE, L. AND J. W. LITTLE. 1967. The paloverde borer or Prionid. Coop. Ext. Survey, Univ. Arizona Folder 134:1-4. NECK, R. W. 1 980. Invertebrates of the lower Rio Grande valley of Texas, with special references to the Southmost, Cameron County, area. Report to Natural Area Survey (un- publ.), Texas Conservation Foundation, Austin, Texas. 54 pp., unnumbered. NEWMAN, E. 1 840- 1841. Entomological notes. Entomologist (1):1-16, 17-32, 33-37, 67-80, 90-95, 110-112, 169-171, 220-223. OLIVIER, A. G. 1795. Entomologie, ou histoire naturelle des insectes, avec leurs caracteres generiques et specifiques, leur description, leur synonymic, et leur figure enluminee. Co- leopteres. Vol. 4(68). Paris. [Genera separately paged.] ORTON, R., D. J. HADDOCK, E. G. BICE, AND A. C. WEBB. 1967. Climatic guide. The lower Rio Grande valley of Tex- as. Texas Agric. Exp. Sta. Misc. Publ. 241:1-108. PAPP, C. S. 1955. New records for North American Cer- ambycidae and a new subspecies ofLeptura. Entomol. News 66(8):2 17-220. . 1959. Notes on some cerambycid beetles from the southwest United States and Mexico. Bull. So. Calif. Acad. Sci. 58(2):84-94. PENROSE, R. L. 1974. A new subspecies of Crossidius hu- meralis LeConte from Texas, with a redescription of the species. Pan-Pacific Entomol. 50(3):248-254. PIPER, G. L. 1977. Biology and habits of Hippopsis lemnis- cata. Coleop. Bull. 31(3):273-278. PORTER, C. C. 1977. Ecology, zoogeography and taxonomy of the lower Rio Grande valley mesostenines. Psyche 84(1): 28-91. PSOTA, F. J. 1930. The Moneilema of North America and Mexico, I. Coleop. Contrib. 1(2):1 11-144. RASKE, A. G. 1971. Bionomics and taxonomy of the genus Moneilema Say. Ph. D. Dissertation (unpubl.), Univ. Cali- fornia, Berkeley. 270 pp. RICE, M. E., R. H. TURNBOW, JR., AND F. T. HOVORE. 1985. Biological and distributional observations on Cerambycidae from the southwestern United States. Coleopt. Bull. 39(1): 18-24. RILEY, C. V. 1880. Food habits of the longicorn beetles or wood borers. Am. Entomol. 1:237-239, 270-271. . 1 890. Insects injurious to hackberry. Pp. 60 1-622 in Fifth Report of the U.S. Entomol. Comm. Washington. ROGERS, C. E. 1977a. Bionomics ofOncideres cingulata on mesquite. J. Kansas Entomol. Soc. 50:222-228. . 1977ft. Cerambycid pests of sunflower: distribution and behavior in the southern plains. Environ. Entomol. 6(6): 833-838. SAY, T. 1823-1824. Descriptions of coleopterous insects col- HOVORE, PENROSE, AND NECK: CERAMBYCIDAE OF SOUTHERN TEXAS 333 lected in the late expedition to the Rocky Mountains, per- formed by order of Mr. Calhoun, Secretary of War, under the command of Major Long. J. Acad. Nat. Sci. Phila., 3, Pt. 1:139-216; Pt. 2:239-282, 298-331, 403-462. . 1826. Descriptions of new species of coleopterous insects inhabiting the United States. J. Acad. Nat. Sci. Phila. 5(2):237-284, 293-304. . 1831. Description of new species of North American insects found in Louisiana by Joseph Barabino. 19 pp. In- diana [from Horn]. . 1835. Descriptions of new North American coleop- terous insects, and observations on some already described. Boston J. Nat. Hist. 1(2): 15 1-203 [from LeConte]. SCHAEFFER, C. F. A. 1 904. New genera and species of Co- leoptera. J. New York Entomol. Soc. 12:197-236. . 1905a. Additions to the Coleoptera of the United States with notes on some known species. Sci. Bull. Mus. Brooklyn Inst. Arts Sci. 1(6):123-140. . 1 905ft. Some additional new genera and species of Coleoptera found within the limit of the United States. Sci. Bull. Mus. Brooklyn Inst. Arts Sci. 1(7):141-179. . 1 906. Two new Oncideres with notes on some other Coleoptera. Can. Entomol. 38:18-27. . 1908. List of the longicorn Coleoptera collected on the Museum expeditions to Brownsville, Texas, and the Huachuca Mountains, Arizona, with descriptions of new genera and species and notes on known species. Sci. Bull. Mus. Brooklyn Inst. Arts Sci. l(12):325-352. . 1917. On Merium and some blue Callidium. J. New York Entomol. Soc. 25:183-187. SCHWARZ, E. A. 1888. The insect fauna of semi-tropical Flor- ida with special regard to the Coleoptera. Proc. Entomol. Club of the A.A.A.S. In Entomol. Am. 4(9): 165-1 73. . 1 896. Semi-tropical Texas. Proc. Entomol. Soc. Wash. 4:1-3. SCHWARZ, H. F. 1929. Honey wasps. Nat. Hist. 29:421-426. SCHWITZGEBEL, R. B. AND D. A. WILBUR. 1942. Coleoptera associated with ironweed, Vernonia interior Small in Kan- sas. J. Kansas Entomol. Soc. 15:37-44. SELLARDS, E. H., W. S. ADKJNS, AND F. B. PLUMMER. 1932. The geology of Texas. Univ. Texas Bull. No. 3232. SIMBERLOFF, D. 1978. Our fragile evolutionary heritage. Is- lands and their species. The Nature Conservancy News, July- August:^ 10. SNOW, F. H. 1906. Some results of the University of Kansas entomological expeditions to Galveston and Brownsville, Texas, in 1904 and 1905. Trans. Kansas Acad. Sci. 20:136- 154. SOLOMON, J. D. 1968. Cerambycid borer in mulberry. J. Econ. Entomol. 61(4):1023-1025. . 1972. Biology and habits of the living beech borer in red oaks. J. Econ. Entomol. 65(5): 1307-1 3 10. . 1980. Cottonwood borer (Plectrodera scalator)—a guide to its biology, damage, and control. U.S.D.A. For. Serv. Res. Paper SO-157:1-10. STURM, J. 1843. Catalog der Kaefer-Sammlung von Jacob Sturm. Nuremberg. 386 pp. SWENSON, W. H. 1969. Comparisons of insects on mesquite in burned and unburned areas. Masters Thesis (unpubl.), Texas Technological College. 62 pp. TEMPLETON, A. R. 1 979. Genetics of colonization and estab- lishment of exotic species. Pp. 41-49 in Genetics in relation to insect management. Working Papers: the Rockefeller Foundation. New York. TERBORGH, J. AND B. WINTER. 1980. Some causes of extinc- tion. Pp. 1 19-133 in Conservation biology, M. E. Soule and B. A. Wilcox, eds. Sinauer Assoc. Inc., Sunderland, Mas- sachusetts. 395 pp. THE NATURE CONSERVANCY. 1981. Annual Report 1980, Projects 1980, Texas: Lower Rio Grande. The Nature Con- servancy News 3 1(3): 24. THOMAS, S. L. 1951. Derobrachus geminatus on grape roots. Pan-Pacific Entomol. 27(1):35. THOMPSON, C. M., R. R. SANDERS, AND D. WILLIAMS. 1972. Soil survey of Starr County, Texas. U.S.D.A. Soil Conserv. Serv. and Texas Agric. Exp. Sta. 62 pp. THOMSON, J. 1860. Essai d'une classification de la famille des cerambycides et materiaux pour servir une monographic de cette famille. Paris. 404 pp. . 1865. Diagnoses d'especies nouvelles qui seront de- crites dans 1'appendix du systema cerambycidarum. Mem. Soc. Roy. Sci. Liege 19:541-578. -. 1 868a. Revision du groupe des onciderites (Lamites, Cerambycides, Coleopteres). Physis Recueil d'Hist. Nat. 5(2): 41-92. . 18686. Materiaux pour servir a une revision des lamites (cerambycides, Col.). Physis Recueil d'Hist. Nat. 2(6): 146-200. TOWNSEND, C. H. T. 1902. Contribution to a knowledge of the coleopterous fauna of the lower Rio Grande valley in Texas and Tamaulipas, with biological notes and special reference to geographical distribution. Trans. Texas Acad. Sci. 5:49-101. TROWBRIDGE, A. C. 1932. Tertiary and quaternary geology of the lower Rio Grande region, Texas. U.S. Geol. Surv. Bull. 837:1-260. TURNBOW, R. H. JR. AND F. T. HOVORE. 1979. Notes on southeastern Cerambycidae. Entomol. News 90(5):2 1 9-229. TURNBOW, R. H. JR. AND J. E. WAPPES. 1978. Notes on Texas Cerambycidae. Coleop. Bull. 32(4):367-372. . 1981. New host and distributional records for Texas Cerambycidae. The Southwestern Entomol. 6(2): 7 5-80. TYSON, W. H. 1970. Notes on the host, larval habits, and parasites of two Texas cerambycids. Proc. Entomol. Soc. Wash. 72:93. . 1973. The Spalacopsis of the West Indies and Amer- ica north of Mexico. Coleop. Bull. 27(3):1 17-137. UHLER, P. R. 1855. Descriptions of a few species of Coleop- tera, supposed to be new. Proc. Acad. Nat. Sci. Phila. 7:415- 418. VIANA, M. J. 1971. Las especies Argentinas de Ancylocerini Thompson y catalogo bibliographico de la tribu. Rev. Museo Argentino de Ciencias Nat., Entomol. 3(3): 149-205, 3 lam. VOGT, B. G. 1 949a. Notes on Cerambycidae of the lower Rio Grande valley, Texas. Pan-Pacific Entomol. 25(3): 137- 144;(4):175-184. . 1 949ft. A biologically annotated list of the Bupres- tidae of the lower Rio Grande valley, Texas. Ann. Entomol. Soc. Am. 42:191-202. WHITCOMB, R. F. 1977. Island biogeography and "habitat islands" of eastern forests, I. Introduction. Am. Birds 31(1): 3-5. WICKHAM, H. F. 1897. Coleoptera of the lower Rio Grande valley, I. Bull. Lab. Nat. Hist. St. Univ. Iowa 4:96-1 15. . 1 898. Recollections of old collecting grounds, III, the lower Rio Grande valley (cont.). Entomol. News 9:39-41. WILCOX, B. A. 1980. Insular ecology and conservation. Pp. 95-117 in Conservation biology, M. E. Soule and B. A. Wilcox, eds. Sinauer Assoc. Inc., Sunderland, Massachu- setts. 395 pp. 334 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 13 WILEY, J. P. JR. 1982. Phenomena, comment and notes. Smithsonian Mag. 12(12):37-43. WILLIAMS, D., C. M. THOMPSON, AND J. L. JACOBS. 1977. Soil survey of Cameron County, Texas. U.S.D.A. Soil Conserv. Serv. and Texas Agric. Exp. Sta., Washington, D.C. 94 pp. WILLIAMS, R. W. 1941. Notes on the bionomics of Ataxia hubbardi Fisher in Illinois. Entomol. News 52(10):27 1-273. WYND, F. L. 1944a. The soil series represented in Hidalgo County, Texas. Am. Midland Nat. 32:181-199. . 1 944ft. The geologic and physiographic background of the soils in the lower Rio Grande valley, Texas. Am. Midland Nat. 32:200-235. APPENDIX A List of Common and Scientific Names of Plants Cited in Species Accounts (Terminology according to Correll and Johnson 1970) annqua—Ehretia anacua (Berl.) Johnston black willow— Salix nigra Marsh brasil— Condalia Hookeri M. C. Johnston cedar elm— Ulmus crassifolia Nutt. cenizo—Leucophyllumfrutescens (Berl.) I. M. Johnston colima—Zanthoxylumfagara (L.) Sarg. coyotillo— Karwinskia humboldtiana (R. & S.) Zucc. ebony— Pithecellobium flexicaule (Benth.) Coulter granjeno— Celtis pallida Torr. honey mesquite— Prosopis glandulosa Torr. huisache— Acacia farnesiana (L.) Willd. lotebush ("lote")— Zizyphus obtusifolia (Hook.) Weberb. Mexican ash, fresno— Fraxinus berlandieriana A. D. C. (Lower Valley) Mexican olive— Cordia boisseri D. C. red ash— Fraxinus pennsylvanica v. subintegerrima (Vahl.) Fem. (Welder Wildlife Refuge) red mulberry— Morus rubra L. relama—Parkinsonia aculeata L. Spanish dagger, yucca— Yucca treculeana Carr. sugar hackberry— Celtis laevigata \. texana (Scheele) Sarg. tepehuaje, lead tree—Leucaena pulverulenta (Schlect.) Benth. Texas palmetto— Sabal texana (Cook) Becc. Texas kidneywood— Eysenhardtia texana Scheele Texas persimmon— Diospyros texana Scheele western soapberry, jaboncillo— Sapindus saponaria v. drum- mondi (H. & A.) Benson CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 14, pp. 335-342, 3 figs. June 30, 1987 RHINOBATOS PUNCTIFER, A NEW SPECIES OF GUITARFISH (RHINOBATIFORMES: RHINOBATIDAE) FROM THE RED SEA, WITH NOTES ON THE RED SEA BATOID FAUNA By Leonard J. V. Compagno1 J. L. B. Smith Institute of Ichthyology, Grahamstown 6140, South Africa and John E. Randall Bernice P. Bishop Museum, Honolulu, Hawaii 96817 ABSTRACT: A new species of guitarfish, Rhinobatos punctifer, is described from a single 705-mm specimen from the Gulf of Aqaba, Red Sea. Apparently the specimens reported as K. schlegelii by Gohar and Mazhar (1968) from Suez were the same species. Rhinobatos punctifer belongs in the subgenus Rhinobatos. It is characterized by a mod- erately long, angular, blunt-tipped snout (preoral snout 2.8 times mouth width); broad rostral ridges well sepa- rated along their length; large eyes (greatest eye diameter 1.3 times interorbital space); oblique broad nostrils, their width 1.3 in mouth width; mouth nearly straight, its width 6.7 in distance from snout to anus; origin of dorsal fin posterior to pelvic bases by a distance 1.5 in interdorsal space; regularly spaced, small white spots on head, disc, pelvic fins, and tail; no pale edge on snout. INTRODUCTION Norman (1926), in a revision of the guitarfish genus Rhinobatos Linck, 1790, reported two spe- cies from the Red Sea: R. halavi (Forsskal 1775), and doubtfully R. thouin (Anonymous 1798). In Fishes of the Red Sea and Southern Arabia, Fowler (1956) accepted these two species and listed also R. schlegelii Miiller and Henle, 1841 and R. granulatus Cuvier, 1829. Fowler based his inclu- sion of R. schlegelii on a listing by Zugmayer (1913), who reported the species from Oman, not from the Red Sea. (Norman [1926] gave the dis- tribution of R. schlegelii only as China and Ja- pan.) Fowler (1956) specifically listed the Red Sea 'Research Associate, Department of Ichthyology, California Academy of Sciences, Golden Gate Park, San Francisco, Cali- fornia 941 18. among the localities for R. granulatus; but he ex- amined no Red Sea material, and neither of the references he gave with the species included the Red Sea. Nor could Fowler have been citing the Red Sea record of R. granulatus by Bamber (1915) because Fowler followed Norman in con- sidering this a misidentification of R. halavi. We, therefore, regard Fowler's (1956) record of R. schlegelii from the Red Sea as false and that of R. granulatus as very doubtful. Gohar and Mazhar (1964) reported four white- spotted specimens of Rhinobatos, "ranging from 62 to 80.5 cm in length," from the Suez market as R. schlegelii. Apparently their specimens were not retained. The junior author obtained a specimen from fishermen in the Gulf of Aqaba, Red Sea, which appears to be the same species as that reported as R. schlegelii by Gohar and Mazhar (1964). Com- 335 336 CAMPAGNO AND RANDALL: RHINOBATUS PUNCTIFER parison of this specimen with published accounts and material of species of Rhinobatos convinced us that it is not R. schlegelii but a new species that we name R. punctifer. The holotype has been de- posited in the Bernice P. Bishop Museum, Hono- lulu (BPBM). Specimens of related species were examined at the British Museum (Natural His- tory), London (BM [NH]). Photographs are pro- vided herein of the holotype (Fig. 1) and of speci- mens of two other species of the genus that have been recorded from the Red Sea, R. halavi (Fig. 2) and R. thouin (Fig. 3). (The photo oiR. thouln is of an Indonesian specimen; we have not seen Red Sea material of this species.) The new species falls in Norman's (1926) subge- nus Leiobatus Rafinesque, 1810 of the genus Rhinobatos. However, because of the inclusion of R. rhinobatos (Linnaeus, 1758) in Leiobatus and the assignment of R. rhinobatos as type species of Rhinobatos by absolute tautonymy, Leiobatus of Norman should be considered a junior synonym of the subgenus Rhinobatos Linck, 1790. Norman (1926) listed seven species in Leiobatus (Rhino- batos), R. schlegelii, R. rhinobatos, R. holcorhyn- chus Norman, 1922, R. formosensis Norman, 1926, R. annandalei Norman, 1926, R. lionotus Norman, 1926, and/?, hynnicephalus Richardson, 1846. Additional species include R. albomacula- tus Norman, 1930, R. irvinei Norman, 1931, and R. punctifer. All of the species in Norman's subge- nus Leiobatus agree in having a moderately long, pointed, angular snout and anterior nasal flaps ex- tending medially onto the internasal space but not nearly meeting on the midline of the snout. Rhinobatos punctifer can be distinguished from all other species in this group by a combination of characters including its broad but elongated and angular snout, broad, well-separated rostral ridges, reduced spination, and white spots. Char- acters distinguishing R. punctifer are presented in the diagnosis below. Rhinobatos punctifer, new species Figure 1 HOLOTYPE. — BPBM 20843, 705 mm total length, adolescent male, Red Sea, Gulf of Aqaba, from fishermen through Coral World, Eilat, J. E. Randall, 8 August 1976. DIAGNOSIS. — A Rhinobatos with a moderately elongate, broad and bluntly round-tipped, angu- lar snout, with a slightly concave margin towards tip; tip of snout not laterally expanded; preoral snout 2.8 times mouth width; preorbital snout 2.3 times distance between spiracles; distance from tip of snout to anterior edge of eye 1.5 in distance Figure 1. Holotype of Rhinobatos punctifer, BPBM 20843, 705 mm, Gulf of Aqaba. PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 14 337 Figure 2. Rhinobatos halavi, BPBM 28364, 825 mm, Jeddah. Figure 3. Rhinobatos thouin, BPBM 26591, 390 mm, Jakarta. 338 CAMPAGNO AND RANDALL: RHINOBA TUS PUNCTIFER from posterior edge of eye to pectoral axil; rostral ridges of snout broad, thick, widely separated from each other along their lengths, slightly diver- gent basally but then somewhat convergent ante- riorly, not fused together or touching each other over precerebral cavity of rostrum; eyes large, length of eyeball 1.3 times interorbital space, 3.2 in preorbital snout; interorbital space slightly con- cave; distance from front of eye to rear edge of spiracle about equal to distance between spira- cles; spiracles with two moderately strong poste- rior ridges; nostrils oblique, at about a 57° angle to longitudinal axis of snout; nostrils moderately broad, their width 1.3 in mouth width, 1.9 times internarial space; anterior nasal flaps with medial folds extending onto internarial space but not me- dial to the excurrent apertures; anterior nasal flap with a long, broad lobe at its midlength; posterior and posterolateral nasal flaps very broad; hori- zontal distance from lateral edge of incurrent ap- erture to lateral margin of snout 4.6 in preoral snout; mouth nearly straight, its width 6.7 in dis- tance from snout to vent; first dorsal fin with ori- gin posterior to pelvic bases by distance of 1.5 in interdorsal space, its base 2.5 in interdorsal space, its height about 1.2 times its length; enlarged den- ticles or thorns obsolete on dorsal surface of body, absent on snout tip and rostral ridges; denticles on scapular region, midline of back, and between and behind dorsal fins minute, blunt, and incon- spicuous; rostrum 1.3 times nasobasal length of cranium (from base of rostrum to occipital con- dyle), its width across nasal capsules 1.3 times na- sobasal length; nasal capsules oblique; pectoral fin with 71 radials (including 33 propterygial radi- als); 179 free vertebral centra behind synarcual; back with regular, symmetrical, wide-spaced, small white spots on head, disc, pelvic fins and tail; no light stripes on snout edge. DESCRIPTION. — Proportional dimensions of holotype, 705 mm total length, as percentages of total length, are as follows. Snout to: nostrils, 10.8; eyes, 12.6; mouth 15.9, fifth gill openings, 24.7; pectoral apices, 28.4; pec- toral rear tips, 40.9; first dorsal origin, 53.0; sec- ond dorsal origin, 72.3; pelvic origins, 35.0; vent, 37.9; upper caudal origin, 64.4. Distance between: front edge of eye and rear margin of spiracle, 5.4; eyeball to pectoral axil, 18.9; outer edge of nostril to rim of disc, 3.5; first and second dorsal bases, 13.2; pectoral and pelvic bases, 0.6; pelvic and first dorsal origins, 17.7; pelvic and first dorsal bases 14.1; second dorsal base and upper caudal origin, 7.4; pelvic bases and lower caudal origin, 44.0. Eye: length of eyeball, 4.0; length of cornea, 3.0; interorbital space, 34.0. Nostril: diagonal width, 4.5; length, 3.0; in- ternarial, 2.4. Spiracle: width, 3.3; interspiracular, 5.5. Mouth: width, 5.7; length, 3.0. Gill openings: width of first, 1.4; second, 1.6; third, 1.6; fourth, 1.4; fifth, 1.1. Width between first, 12.5; width between fifth, 8.7. Height of: head at eyes, 3.4; trunk at pectoral insertions, 4.8; trunk at pelvic insertions, 5.0. Width of trunk at: pectoral insertions, 12.3; pel- vic insertions, 8.8. Pectoral disc width: 33.8. Pelvic fin: anterior margin length, 10.1; height, 5.5; base length, 9.4; inner margin length, 7.3; length of fin from origin to free rear tip, 16.7. First dorsal fin: anterior margin length, 11.9; height, 9.5; base length, 5.4; inner margin length, 2.7; length of fin from origin to free rear tip, 8.1. Second dorsal fin: anterior margin length, 10.8; height, 8.5; base length, 5.5; inner margin length, 2.4; length of fin from origin to free rear tip, 7.8. Caudal fin: dorsal margin length, 15.2; preven- tral margin, 7.4. Snout broadly wedge-shaped, angle in front of eyes 66°; fifth gill openings about 2/3 length of first 4; posterolateral nasal flaps extending from poste- rior margin of incurrent apertures to inner third of excurrent aperture; tooth row counts 76/22 or 37- 1-38/34-38; teeth with low, oval, transversely elongated crowns, indistinct cutting edges, no transverse ridges, strong basal ledges and grooves, and small roots, regularly increasing in size from symphysis to mouth corners and not abruptly enlarged in symphyseal region; disc width 87% of disc length; tail from vent to cau- dal tip 1.6 times snout-vent length, nearly flat be- low, rounded above, and tapering to caudal fin, its width at pelvic insertions 1.6 times distance be- tween spiracles; tail with lateral dermal folds orig- inating slightly anterior to free rear tips of pelvics and reaching just behind lower caudal origin, widths of folds opposite interdorsal space about V? of eyeball length. First dorsal fin slightly larger than second, both triangular, with slightly convex anterior margins, narrowly rounded or pointed apices, concave, nearly vertical posterior margins, angular, slightly pointed free rear tips, and convex inner margins; inner margins of dorsal fins 2/s to l/2 length of ba- PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 14 339 ses; interspace between second dorsal base and upper caudal origin 1.3 times length of second dorsal; pelvic fins with slightly convex anterior margins, narrowly rounded apices, convex poste- rior margins, narrowly rounded free rear tips, straight inner margins, and free rear tip angles of about 128°; pelvic lengths from origins to free rear tips 1.8 times base lengths; caudal fin with upper origin slightly anterior to lower origin, dorsal margin convex and with length about 1.2 times in- terdorsal space, broadly convex preventral mar- gin, broadly rounded ventral apex, undulated postventral margin, and angular dorsal apex; cau- dal fin without ventral lobe, with axis at about a 16° angle above body axis; epaxial lobe of caudal as high as hypaxial lobe. Dermal denticles minute, close-set, covering entire body except for area behind posterior nasal flaps on snout, upper lip, and chin, and at pecto- ral, pelvic, and dorsal fin axils; lateral trunk denti- cles above the pelvic fin bases with wedge-shaped crowns, low but strong medial ridges, sometimes low lateral ridges, and broad, blunt medial cusps; one or two small, inconspicuous, blunt denticles or thorns present on scapular region; similar den- ticles at front edges of eyes and along supraorbital ridges. Rostral cartilage broad, its shaft nearly uni- formly wide behind rostral node; rostral appendi- ces broadly expanded and rounded, not angular; rostrum enclosing a broad precerebral cavity that tapers only slightly to rostral node; dorsal edges of precerebral cavity (rostral ridges on surface of snout) broadly separated along their lengths; na- sal capsules large, their transverse axes anterola- terally directed; width across nasal capsules 1.3 times nasobasal length of cranium (base of ros- trum to occipital condyles); length of nasal cap- sules about equal to their width; basal plate nar- row, its width at anterior ends of orbits 0.2 times in nasobasal length; cranial roof with small, keyhole-shaped frontal fenestra, well behind an- terior fontanelle; antorbital cartilage triangular, broad, and wedge-shaped posteriorly, without an anterior lobe extending past nasal capsules; post- orbital processes large and bifurcate; preorbital processes poorly differentiated on supraorbital crests; width across postorbital processes 0.6 times nasobasal length; width across otic capsules 0.4 times nasobasal length. Pectoral fin skeleton with 33 propterygial, 6 mesopterygial, 2 neopterygial, and about 30 me- tapterygial radials; anteriormost radials of pro- pterygium reaching in front of base of nasal cap- sules by about 0.08 of rostral length; pelvic girdle medially arched, with short, broad lateral prepu- bic processes and narrow, falcate iliac processes; pelvic fin with about 26 radials. Vertebral column with cervicothoracic synar- cual having 15 centrum-free segments and 14 cen- tra (29 total), 27 monospondylous precaudal cen- tra behind synarcual (most with elongate, slender ribs), 104 diplospondylous precaudal centra, and 48 caudal centra; total segments 208 and total cen- tra 193; intestinal valve of spiral type, with 11 turns. Color in preservative medium brown on dorsal surface of disc and tail, cream below; rostral ridges darker but with a light area on either side of rostrum; small light spots, the largest about 5 mm wide, mostly arranged in sparse, transverse rows on dorsal surface of head, disc, pelvic fins, and tail in front of second dorsal base; underside of pre- oral snout with a dusky blotch. DERIVATION OF NAME. — Latin punctifer, bearer of spots, for the prominent regular pattern of white spots on the dorsum. COMPARISON WITH OTHER SPECIES OF RHINO- BATOS. — As noted above Rhinobatos punctifer is closest to seven species of Eastern Hemisphere Rhinobatos included by Norman (1926, 1930, 1931) in the subgenus Leiobatus ( = Rhinobatos). Of these, Rhinobatos rhinobatos occurs in the Mediterranean Sea and eastern Atlantic. It differs from R. punctifer in having a more angular, narrow-tipped, bottle-shaped snout; rostral ridges closer together; nostrils smaller, with widths 1.1 to 1.3 times intern arial space, 1.7 in mouth width; supraorbital, scapular, and mid- dorsal thorns well developed; distance from first dorsal origin to pelvic bases 1.1 in interdorsal space; and no white spots. Two specimens of Rhinobatos rhinobatos BM(NH) 1935.3.5.1, a 487-mm female, and BM(NH) 1936.4.14.44, a 478-mm immature male, were examined for this study. Two West African species of this group, Rhino- batos albomaculatus and R. irvinei (descriptions by Norman 1930, 1931), have white spots like R. punctifer; the holotypes (R. albomaculatus, BM[NH] 1930.3.24.2, 566-mm female; R. irvinei, BM[NH] 1930.8.26.3, 569-mm adult male) were examined. These two species differ from R. punc- tifer in having more acutely angular, narrow- tipped snouts; narrower, more closely confluent rostral ridges; smaller eyes, 4 to 4.8 times in pre- 340 CAMPAGNO AND RANDALL: RHINOBATUS PUNCTIFER orbital snout; smaller, more widely spaced nos- trils, 1.3 times internarial space and 1.7 to 1.9 in mouth width; and first dorsal base 3.1 to 3.3 in in- terdorsal space. Rhinobatos irvinei also differs from R. punctifer in having dark spots on the in- terorbital space and small but prominent supraor- bital, scapular, and middorsal denticles. Both R. albomaculatus and R. irvinei were placed in the genus Rhynchobatus by Bigelow and Schroeder (1953) because of their supposedly notched tails with ventral caudal lobes, but both holotypes of these species proved to belong to Rhinobatos, having damaged, artificially notched tails. Rhinobatos holcorhynchus is an Indian Ocean, South African species similar to R. punctifer and redescribed by Norman (1926) and Wallace (1967). It differs from R. punctifer in having a longer, narrower snout with the preorbital length 2.8 times the interspiracular space; the distance from first dorsal origin to pelvic bases 1.3 in inter- dorsal space; large supraocular, scapular, and middorsal thorns; and no white spots. Rhinobatos annandalei and R. lionotus are two similar species described by Norman (1926) from the Bay of Bengal. They are close to R. punctifer but differ from it in having narrower snout tips; rostral ridges much closer together; nostrils smaller and more widely separated, 1.7 in mouth width and 1.3 times internarial space; and no white spots. Rhinobatos annandalei additionally differs by having conspicuous, sharp-tipped su- perocular, scapular, and middorsal thorns, and R. lionotus by having the first dorsal origin posterior to the pelvic bases by a distance equal to the inter- dorsal space. Three western North Pacific species, Rhinoba- tos schlegelii, R. hynnicephalus, and R. formosen- sis are similar to R. punctifer, but all differ in hav- ing narrower-tipped snouts with rostral ridges close together; smaller nostrils, 1.2 to 1.5 times in- ternarial space and 1.4 to 1.7 in mouth width; and origin of first dorsal posterior to pelvic bases by 1.0 to 1.3 times in interdorsal space. Rhinobatos schlegelii and R. formosensis additionally differ in their much longer snouts, with the preorbital snout 3.1 to 3.3 times interspiracular, preoral snout 3.3 to 3.7 times mouth width, and plain col- oration; R. schlegelii in its more bottle-shaped snout and weak spiracular ridges; and R. hynni- cephalus in its smaller eyes, 4.7 to 5.8 in preorbital snout, and dorsal color pattern with rosettes of dark spots but no white spots. Apparently, there are no confirmed records of R. schlegelii in the Red Sea or even the Indian Ocean, that of Gohar and Mazhar (1964) from Egypt being based on R. punctifer and that of Fowler (1956) from Oman uncertain. According to Norman (1926), R. schle- gelii has been confused with R. lionotus, as well as with the western Pacific R. formosensis and R. hynnicephalus. Among other species of Rhinobatos in the Red Sea, R. halavi was recorded as very common off Egypt (Gohar and Mazhar 1964) and was col- lected in 1982 by the junior author at Jeddah, Saudi Arabia. Rhinobatos halavi differs from R. punctifer in its shorter, more acutely angular snout; rostral ridges closely adjacent to each other along their lengths; smaller eyes; lower spiracular ridges; anterior nasal flaps not extending onto the internasal space; enlarged rostral, supraorbital, scapular and middorsal thorns; and plain colora- tion. The other Red Sea species, R. thouin, has not been recently reported from the Red Sea and its presence there requires confirmation. It is eas- ily separable from R. punctifer by its extremely elongate, attenuate snout (the preorbital snout 3.3 to 3.7 times the interspiracular space) with lat- erally expanded tip (unlike that of any other living rhinobatoid). It also has narrow, closely spaced rostral ridges; weak spiracular ridges; narrower nostrils, with very small and narrow anterior, pos- terior, and posterolateral nasal flaps ;• anterior na- sal flaps not extending medially onto the interna- rial space; enlarged rostral, supraorbital, scapular, and middorsal thorns; and plain colora- tion. NOTES ON THE RED SEA BATOID FAUNA We preface remarks on the Red Sea batoid fauna with a checklist of species, mostly compiled from available literature (including Fowler 1956; and Gohar and Mazhar 1964). The ray fauna of the Red Sea is poorly known, more so than the shark fauna, and the following list is tentative: Order RHINOBATIFORMES Family RHYNCHOBATIDAE Rhina ancylostoma (Bloch and Schneider, 1801), Rhynchobatus djiddensis (Forsskal, 1775). Family RHINOBATIDAE Rhinobatos halavi (Forsskal, 1775), R. punctifer Compagno and Randall, new species, and R. thouin (Anonymous, 1798). PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 14 341 Order PRISTIFORMES Family PRISTIDAE Anoxypristis cuspidata (Lathan, 1794), Pristis pectinata Latham, 1794, Pristis zijsron Bleeker, 1851 (note, photos labeled Pristis pectinata in Go- har and Mazhar, 1964 apparently are of P. zij- sron, hitherto not known from the Red Sea). Order TORPEDINIFORMES Family TORPEDINIDAE Torpedo panthera Olfers, 1831, T. sinuspersici Olfers, 1831. Order MYLIOBATIFORMES Family DASYATIDIDAE Himantura gerrardi (Gray, 1851), H. itnbricata (Bloch and Schneider, 1801), H. uarnak (Forsskal, 1775), Hypolophus sephen (Forsskal, 1775), ?Taeniura grabata (St. Hilaire, 1809), T. lymma (Forsskal, 1775), T. melanospilos Bleeker, 1853, Urogymnus asperrimus (Bloch and Schneider, 1801). Family GYMNURIDAE Aetoplatea tentaculata Valenciennes in Miiller and Henle, 1841, Gymnura poedlura (Shaw, 1804). Family MYLIOBATIDAE Aetobatus narinari (Euphrasen, 1790), Aetomy- laeus milvus (Valenciennes, in Miiller and Henle, 1841). Family MOBULIDAE Manta ehrenbergi (Miiller and Henle, 1841) or M. birostris Walbaum, 1792), Mobula kuhlii (Valen- ciennes in Miiller and Henle, 1841) or M. diabolus (Shaw, 1804). Like the Red Sea shark fauna, the batoid fauna of the Red Sea is relatively depauperate with fewer species than the western Indian Ocean and with the fauna comprising coastal-benthic, coastal-pelagic, and epipelagic species. There are no deep-water Red Sea rays and no Red Sea mem- bers of the order Rajiformes, although deep-sea rays including rajoids occur in the Gulf of Aden. Of the 24 species listed above, 22 are also found in the western Indian Ocean. The ones not known from this region are Rhinobatos punctifer and the dubiously recorded Taeniura grabata (otherwise known from the Mediterranean Sea and eastern Atlantic). Three of the Red Sea batoids are cir- cumtropical in distribution: Pristis pectinata, Aetobatus narinari, and Manta birostris (provid- ing M. ehrenbergi is a junior synonym of it). If the West African Urogymnus africanus is a junior synonym of the Indo-Pacific U. asperrimus, then it too ranges beyond the Indo-West Pacific region. Compared to the Red Sea shark fauna, the Red Sea batoids have a much lower proportion of epi- pelagic and circumtropical species and more Indo-West Pacific species. Rhinobatos punctifer is currently the only known endemic Red Sea elas- mobranch, but it may eventually be collected in the northwestern Indian Ocean. On the other hand, it may prove to be confined to the cooler northern part of the Red Sea. Taeniura grabata is a species otherwise known from the Mediterra- nean Sea and eastern Atlantic, but records of it from the Red Sea are apparently doubtful (Krefft and Stehmann 1973). The nature of the Red Sea batoid fauna may be due to restrictive conditions in the Red Sea envi- ronment, limiting inshore species that can live there and barring deep-water species. Presum- ably, the Red Sea batoid fauna originated by dis- persal from the western Indian Ocean. ACKNOWLEDGMENTS LJVC thanks the following for assistance dur- ing the preparation of this paper: Alwyne C. Wheeler and Oliver Crimmen of the British Mu- seum (Natural History), for their help with speci- mens; and Walter Fischer of Food and Agricul- ture Organization of the United Nations, Rome, for providing the funding to enable him to revisit the BM(NH) in 1979. JER acknowledges the sup- port of the U.S. -Israel Binational Science Foun- dation for his field work in the Red Sea and thanks David Fridman and Eitan Levy of Coral World, Eilat, for procuring the holotype of R. punctifer for our study. LITERATURE CITED ANONYMOUS. 1798. Naturgeschichte Paris, B. Plossan, Histoire naturelle des Poissons par le Cit, La Cepede. 1, 532 pp. Allg. lit. Zeit., Jena 3 (287): 673-678, (288): 68 1-685. BAMBER,R. C. 1915. Reports on the marine biology of the Suda- nese Red Sea, from collections made by Cyril Grassland, M.A., D.Sc., F.L.S.— XII. The Fishes. J. Linn. Soc., Zool. 31:477-485. BIGELOW, H. B. and W. C. SCHROEDER. 1953. Chap. 1, saw- fishes, guitarfishes, skates and rays. Mem. Sears Found. Mar. Res. l(pt. 2)2:1-514. BLEEKER, P. 1851. Vijfde bijdrage tot de kennis der ichthyolo- gische fauna van Borneo, met beschrijving van eenige nieuwe soorten van zoetwatervisschen. Nat. Tijdschr. Ned. Ind. 2:415-442. 342 CAMPAGNO AND RANDALL: RHINOBA TVS PUNCT1FER — . 1853. Diagnostische beschrijvingen van nieuwe of weinig bekende vischsoorten van Batavia. Tiental I- VI. Nat. Tijdschr. Ned. Ind. 4:451-516. BLOCH, M. E. and J. G. SCHNEIDER. 1801. Systema Ichthyolo giae. Iconibus CX Illustratum. Sanderiano Commissum, Berlin. 584pp. CUVIER, G. L. C. F. D. 1829. Le regne animal distribue d'apres son organisation, pour servir de base a 1'histoire naturelle des animaux et d'introduction a 1'anatomie comparee. Vol. 2. Poissons. Paris. EUPHRASEN, B. A. 1790. Raja [ALtobatis narinari] beskrifven. Handl. K. Vetensk. Akad. 11:217-219. FORSSKAL, P. 1775. Descriptiones animalium avium, amphi- biorum, piscium, insectorum, vermium; quae in itinere orien- tali observavit. Post mortem auctoris edidit Carsten Niebuhr. Molleri, Hauniae. 164 pp. FOWLER, H. W. 1956. Fishes of the Red Sea and southern Ara- bia. Vol. 1. Branchiostomida to Polynemida. Weizmann Sci- ence Press, Israel. 240 pp. GOHAR, H. A. F. and F. M. MAZHAR. 1964. The elasmobranchs of the northwestern Red Sea. Pub. Mar. Biol. St., Al- Ghardaqa (Red Sea) (13):1-144. GRAY, J. E. 1851. List of the specimens of fish in the collection of the British Museum. Part I. Chondropterygii. London. 160 pp. KREFFT, G. and M. STEHMANN. 1973. Dasyatidae. Pp. 70-73 in Check-list of the fishes of the northeastern Atlantic and of the Mediterranean, J. C. Hureau and T. Monod, eds. UNESCO, Paris. Vol. 1. LATHAM, J. 1794. An essay on the various species of sawfish. Trans. Linn. Soc. London 2:273-282. LINCK, H. F. 1790. Versuch einer Entheilung der Fische nach den zahnen. Mag. Neuste Phys. Naturgesch. 1789, 6 (pt. 3):28-38. LINNAEUS, C. 1758. Systema Naturae, tenth edition. Laurentii Salvii, Holmiae. 823 pp. MULLER, J. and F. G. J. HENLE. 1841. Systematische Bi- schreibung der Plagiostomen. Berlin. 200 pp. NORMAN, J. R. 1922. Three new fishes from Zululand and Natal, collected by Mr. H. W. Bell Marley; with additions to the fish fauna of Natal. Ann. Mag. Nat. Hist. London Ser. 9, 9:318- 322. . 1926. A synopsis of the rays of the family Rhinobatidae, with a revision of the genus Rhinobatus. Proc. Zool. Soc. Lon- don 1926:941-982. . 1930. A new ray of the genus Rhinobatus from the Gold Coast. Ann. Mag. Nat. Hist., Ser. 10, 6(32): 226-228. -. 1931. Four new fishes from the Gold Coast. Ann. Mag. Nat. Hist., Ser. 10, 7(40):352-359. OLFERS, I. F. J. M. 1831. Die Gattung Torpedo in ihren na- turhistorischen und antiquarischen Beziehungen erlautert. Berlin. 35 pp. RAFINESQUE, C. S. 1810. Caratteri di alcuni nuovi generi e nuove specie di animali [principalmente di pesci] e piante della Sicilia, con varie osservazioni sopra i medisimi. Pa- lermo. 105 pp. RICHARDSON, J. 1846. Report on the ichthyology of the seas of China and Japan. Rep. Brit. Assoc. Adv. Sci., 15th meeting, 1845:187-320. SAINT-HILAIRE, E. F. 1809. Poissons du Nil, de la Mer Rouge et de la Mediterranee. Pp. 1-52 in Description de 1'Egypte . . . Histoire naturelle, Vol. 1, Pt. 1, Paris. SHAW, G. 1804. General Zoology or Systematic Natural His- tory. Pisces, Vol. 5, Pt. 1. Thomas Davison, London. 250 pp. WALBAUM, J. J. 1792. Petri Artedi Sueci Genera piscium in quibus systema totum ichthyologiae proponitur cum classi- bus, ordinibus, generaum characteribus, specierum differen- tiis, observationibus plurimis. Redactis speciebus 242 ad gen- era 52. Ichthyologiae, pars iii. Grypeswaldiae. 723 pp. WALLACE, J. H. 1967. The batoid fishes of the east coast of Southern Africa. Part I: Sawfishes and guitarfishes. S. African Assoc. Mar. Biol. Res., Oceanogr. Res. Inst., Invest. Rep. (15):l-32. ZUGMAYER, E. 1913. Die Fische von Balutschistan mit einleiten- den Bemerkungen uber die Fauna des Landes. Abhandlungen der Koniglich Bayerischen Akademie der Wissenschaften Mathematisch-Physikalische. Klasse XXVI, Band 6: 1-35. ADDENDUM While this paper was in production, an additional specimen of Rhinobatos punctifer was collected in the Gulf of Aqaba. We add this specimen here as a paratype for the Hebrew University of Jerusalem (HUJ). It is HUJ 11733, 645 mm total length, taken with a gill net off Coral World, Eilat, at a depth of 240 m by Eli Kalmanson on 14 November 1986. PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 44, No. 15, pp. 343-371, 42 figs. August 17, 1987 REVIEW OF THE TENEBRIONID TRIBE ANEPSIINI (COLEOPTERA) By John T. Doyen University of California, Department of Entomology Berkeley, California 94720 ABSTRACT: The systematics of the 13 known species of Anepsiini is reviewed. Cladistic relationships support rec- ognition of four genera, including the new genus Batuliomorpha. Keys, descriptions, and diagnoses are provided for genera and species. New species are: Anepsius minutus; Batuliomorpha comata, B. imperialis, and B. tibioden- tata; Batuliodes wasbaueri, B. obesus, and B. spatulatus. Anepsius confiuens (Blaisdell) is newly transferred to Ha- tuliodes. In Anepsius and Batuliodes the more primitive species are surface dwellers, while the more derived ones are strongly modified for a psammophilous existence. Batulius and Batuliomorpha contain only psammophilous species. INTRODUCTION The Anepsiini comprise a small group of little- studied species that occupy arid or subarid habi- tats in western North America. Aside from the rather brief consideration of higher classification by Doyen and Lawrence (1979), previous system- atic treatments have consisted of isolated descrip- tions of species (LeConte 1851 ; Casey 1907; Blais- dell 1923) or very superficial analyses of generic interrelationships (LaCordaire 1859; Horn 1970; Casey 1907). During the last few years, a number of undescribed species have been collected from sand dunes in southeastern California and Baja California. Several of these are strongly modified for a psammophilous existence, and significantly increase the morphological variability within Anepsiini. Conversely, a number of names pro- posed by Casey (1907) need to be placed in synon- ymy. In addition, this paper reevaluates generic interrelationships and provides keys to genera and species. MATERIALS AND METHODS Anepsius delicatulus is commonly found under stones, but other members of Anepsiini are sel- dom noticed, due to their small size, cryptic color- ation, lethargic movements, and nocturnal activ- ity. They are most effectively collected in pitfalls, particularly those left in the substrate for ex- tended periods with a preservative such as ethyl- ene glycol. Dune-inhabiting species may be sifted from the sand about the bases of plants, but sev- eral of these species are small enough to pass through the mesh of standard window screening. Morphological terminology generally follows that in Doyen (1966). Orientation and terminol- ogy used in describing the legs follow Doyen (1984). Measurements were made with an ocular micrometer and grid. Elytral length (EL) is the distance from the posterior tip of the scutellum to the elytral apex, measured parallel to a frontal plane through the body. Pronotal length (PL) is measured along the midline; elytral and pronotal widths (EW, PW) are maximum dimensions per- pendicular to a sagittal plane. Type specimens of all described species were examined. Holotypes of species described by Ca- sey (1907) and considered in this paper are lo- cated in the United States National Museum; those described by Blaisdell (1923, 1943) are lo- cated in the California Academy of Sciences. [343] 344 DOYEN: REVISION OF ANEPSIINI BIOLOGY Life histories of Anepsiini are essentially un- known. Larvae have not been associated with any species. I was unable to obtain eggs or larvae from caged adults of Anepsius delicatulus and Batu- liomorpha comata, although some adults survived for many weeks or months. Anepsius delicatulus, Batuliodes confluens, and probably A. minutus and B. rotundicollis are surface-dwelling species that hide under stones and in other refuges. Anepsius delicatulus is occa- sionally encountered on the soil surface at night. All four species appear to be active throughout the year except during the coldest months in the northern portions of their ranges. For example in a pitfall survey of ground-dwelling arthropods at Owens Lake, Inyo County, California Anepsius delicatulus was recorded in all but three months, with peak numbers in May (Fig. 1) (F. Andrews, A. Hardy; personal communication). In warmer portions of its range, such as the San Joaquin Val- ley and the Los Angeles Basin, specimens have been collected in every month. A similar pattern was found for Batuliodes rotundicollis in the Eu- reka Valley (Mono County, California) (An- drews et al. 1979), where activity was almost en- tirely restricted to the period between May and September (Fig. 1). Presumably this reflects the relatively severe climate of the Eureka Valley at about 915 m. Once again, in warmer parts of B. rotundicollis' s range, collection records exist for most months of the year. These more generalized, surface-dwelling spe- cies range over a variety of sandy and rocky sub- strates in arid and subarid habitats. In their survey at Owens Lake, Andrews et al. recorded Anep- sius delicatulus principally from the Larrea, Fran- seria dumosa, Atriplex-Franseria, and Atriplex confertifolia associations recognized by Matson (1976). These plant associations occur on alluvial substrates that are not strongly alkaline. Only three of the beetles were recorded from alkali scrub associations, and none were taken from sand dune habitats. However, Anepsius delicatu- lus is common on the remnant sand hills near Antioch, Contra Costa County, California, and I have taken it from aeolian sand in the Ciervo Hills, Fresno County (J. Doyen Lot #75C4). In the Eureka Valley Batuliodes rotundicollis was common on rocky hillsides and alkali scrub, but was never found on the aeolian dunes. Specimens have been recorded from the Saline Valley Dunes (Inyo County, California), but ap- pear to be much more abundant on harder sub- strates. In contrast to the more generalized group of species discussed above, Anepsius montanus, Ba- tuliodes obesus, B. spatulatus, B. wasbaueri, Ba- tulius setosus, and the species of Batuliomorpha are all apparently restricted to aeolian sand. Adults shelter beneath the sand surface during the day, sometimes about the bases of vegetation, emerging nocturnally. Available collection re- cords indicate that these species are active during the winter and spring months, but no intensive surveys have been conducted throughout the year. The habits of A. valens are unknown. HIGHER CLASSIFICATION OF ANEPSIINI In the paper in which he originally described Anepsius and Batulius, LeConte (1851) did not use higher level categories. Thus LaCordaire (1859) made the initial tribal placements, includ- ing Anepsius in his Triboliides and Batulius in his Ulomides vrais. His placement of these genera in tribes possessing defensive glands was based on superficial characters. The fundamental differ- ences in abdominal structure were unrecognized at that time. Separation of Anepsius and Batulius was necessary in his classification because of their different mesocoxal configurations. LeConte (1862) included both genera in his Anepsiini, which he placed in his subfamily Tentyriinae. Horn (1870) removed Batulius to his tribe Batu- liini, which he left in Tentyriinae, but he transfer- red Anepsiini to the Asidinae. These changes, again based on the differences in mesocoxal struc- ture, were subsequently followed by LeConte and Horn (1883). Casey (1907) recognized the close relationship between Anepsius and Batulius, sug- gesting that all should form a single tribe, proba- bly without subdivisions. He stressed the strong overall similarity, emphasizing especially the unu- sual elytral sculpture shared by Anepsius and Ba- tuliodes, which he split from Batulius to receive rotundicollis LeConte. Casey was uncertain whether to apply Anepsiini or Batuliini as the proper name, but his comments regarding rela- tionship and classification were essentially cor- rect. Subsequent catalogers (Leng 1920; Gebien 1937; Papp 1961) continued to recognize two tribes, but heeded Casey by giving them adjacent positions in the subfamily Asidinae. Doyen and PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 15 150 345 •o c 100 50 II 50 25 I r M A M J JASOND ,1 I JFMAMJJASOND Month FIGURE 1 . Left: Seasonal distribution of Anepsius delicatulus at Owens Lake, Inyo County, California. (Based on data from Octo- ber, 1977 to January, 1979, courtesy of F. Andrews and A. Hardy, California Department of Food and Agriculture.) Right: Seasonal distribution of Batuliodes rotundicollis in the Eureka Valley, Mono County, California. (Based on data from October, 1977 to Janu- ary, 1979, from Andrews, et al. 1979.) Lawrence (1979) redefined Anepsiini to include Batuliini and suggested that Batuliodes should be placed in synonymy under Batulius. That classi- fication is followed here, except that I resurrect Batuliodes as a valid name for the reasons detailed below. Blackwelder (1945), and later Papp (1961), rec- ognized the subfamily Batuliinae without formal definition. From their checklists it is impossible to determine which taxa, if any, they intended to in- clude other than Batuliini. Relationships of Anepsiini to other groups of Tentyriinae are uncertain. The relatively small mentum is shared with tribes such as Stenosini, Coniontini, Lachnogyini, Cnemoplatiini, and Cryptochilini, as well as difficult to place genera such as Idisia Pascoe. Divided eyes are wide- spread in Stenosini (undivided in Stenosis) and also occur in Typhlusechus, Alaudes, and Boro- morphus. A further character shared by Anep- siini and most Stenosini is the integration of the elytral bases and the scutellum into the collarlike mesothorax, which is amplected into the pro- thorax. The anterior, amplected parts of the ely- tra and the scutellum are depressed below the level of the elytral disk. The scutellum is visible only if the prothorax is moved forward, relative to the elytra. In tribes such as Coniontini a relatively small portion of the elytra is incorporated into the mesothoracic collar. In Cryptoglossini, a rela- tively large part of the elytra forms the dorsolat- eral parts of the collar, but the scutellum remains exposed on the elytral disk. I have not surveyed these complex structures extensively in tentyrioid Tenebrionidae, and their taxonomic value is un- certain. Final resolution of the relationships of the tentyrioid tribes will require detailed morphologi- cal comparisons across the entire subfamily. At present the Anepsiini, Stenosini, and probably the Eurychorini (Koch 1955) may be considered to comprise a somewhat isolated clade within Tentyriinae. Relationships within Anepsiini were assessed by analyzing 27 characters across all the species. Character polarities were determined by out- group comparison with other tribes of Tenty- riinae. This procedure is rendered difficult by the large size and structural diversity of many of these tribes, which may themselves show more than one state for some characters. Characters and character-state polarities are discussed below and listed in Appendix 1. Antennal configuration (Characters 1-5). In Anepsius delicatulus the antennae are gradually enlarged to the 10th segment (Fig. 9), have no dis- tinct club, and are relatively long. In all other spe- 346 DOYEN: REVISION OF ANEPSIINI cies the last 3 or 4 segments are abruptly enlarged as a slight but distinct club (Fig. 7-8). Clavate an- tennae are commonplace in Tentyriinae and are therefore considered the primitive condition in Anepsiini. The two types of clubs probably devel- oped independently, since the lineages in which they occur differ in numerous other features. In burrowing species antennae are variably shortened. This derived condition, along with sev- eral other features correlated with a strongly psammophilous mode of life, has apparently arisen independently several times. Analogous shortening of antennae occurs in many other psammophilous Tenebrionidae. It is difficult to assign polarity to Character 5 (shape of apical antennal segment). Probably the subquadrate condition is associated with psam- mophily and shortening of the antennae. Tentorlal configuration (Character 6). In most Anepsiini the tentorium consists of subvertical lateral laminae, connected posteriorly by a trans- verse bridge. This primitive condition prevails through a great majority of Tenebrionidae. In Ba- tuliomorpha, the posterior space between the transverse bridge and the ventral part of the oc- cipital foramen is closed by a sheet of cuticle (Fig. 2). As far as known, this state is unique to Batulio- morpha. Possibly it is correlated to the function- ing of the mouthparts, whose muscles attach in part to the tentorium. Epistomal margin (Character 7). Both truncate and emarginate epistoma are common in Tenty- riinae. However, the configuration in Batuliodes, with rather prominent lateral lobes and a nearly straight medial portion is distinctive and consid- ered derived. Submental gland (Characters). This secondary sexual feature was discussed in some detail and il- lustrated by Doyen and Lawrence (1979). It is al- most certainly a synapomorphy in Anepsiini, but it is unclear whether its absence in Batuliodes rep- resents a primitive lack or a secondary loss. Pronotal configuration (Characters 9-12). In most Tentyriinae the pronotum has distinct angles and carinate lateral margins without fimbriation. This plesiomorphous condition pertains in Anep- sius. In Batuliodes, the posterior angles are strongly obtuse basally but are produced and acute or nearly right angled near the apex. In psammophilous species the pronotal margins be- come fimbriate, and the lateral carinae may be FIGURE 2. Tentorium of Batuliomorpha comata, posterior oblique aspect; surrounding cranium removed. faint or lacking. These latter modifications appear in diverse groups of psammophilous Tentyriinae. Sculpture of hypomeron (Characters 13-14). Polarity of these characters is uncertain. Elytral fimbriation (Character 15). The ple- siomorphous, glabrous state occurs in nonburro- wing forms. Fimbriae very likely developed inde- pendently in all psammophilous lineages. Elytral sculpture (Character 16). As noted by Casey (1907:503), Anepsiinae are characterized by a peculiar pattern of elytral sculpturing. In each row the punctures are intersected anteriorly by a narrow longitudinal carinule. In Anepsius de- licatulus the anterior edges of the punctures are elevated, producing a characteristic pattern. The same pattern, albeit somewhat modified, is dis- cernable in all except the strongly fossorial spe- cies, suggesting that it is primitively present in the anepsiine lineage. Ventral setation (Character 17) . Presence of se- tae is a derived condition correlated with psam- mophily. Mesocoxal closure (Characters 18-20). These characters were discussed in some detail by Doyen and Lawrence (1979). Mesocoxal cavities bounded laterally by the epimeron, with exposed trochantins, are primitive. Those closed by lobes of the meso- and metasterna are derived. How- ever, it should be noted that, even in those Anep- siini where the epimeron reaches the coxal cavity, PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 15 347 the sternite lobes are closely apposed, constrict- ing the epimeron to a narrow strip. In some other groups of Tenebrionidae, notably Diaperini, me- socoxal closure is variable, possibly related to re- duction in body size (Doyen 1984). Casey (1907:503) previously realized the instability of this character in Anepsiini. Metasternum length (Character 21). In flight- less Coleoptera, the metasternum commonly be- comes much shortened. Although wing loss is ple- siomorphous in Anepsiini, the metasternum is relatively long in all genera but Anepsius, which is relatively primitive in most other features. Protibial configuration (Characters 22 , 24). Ap- ical expansion of the tibiae is frequently associ- ated with psammophily in Tenebrionidae. Modi- fication of the tibial spurs is often an associated feature, and in extreme cases one spur may be en- tirely lost, as in Uniungulum Koch (1962:113). Likewise, fimbriation of the posterior protibial margin often accompanies these other changes. Aedeagal configuration (Character 25). The po- larity of this character is uncertain. The ventral surface (primitively dorsal in the noninverted ae- deagus) of the tegmen is often only lightly sclerotized; the amount of sclerotization some- times varies among individuals. Body proportions and setation (Characters 26, 27). Relatively slender, subglabrous bodies are associated with surface dwelling. Obese, fore- shortened bodies and long, projecting setae are associated with psammophilous, burrowing habits in many groups of tenebrionidae. These charac- teristics have probably evolved independently several times in Anepsiini. CLADISTIC RELATIONSHIPS One possible (hand-generated) cladistic ar- rangement of Anepsiini appears in Figure 3. This diagram includes the distribution of all characters for which the derived state is shared by at least two species even if the character polarity is not ab- solutely certain. Autapomorphous characters are described in the species treatments. At the outset it should be explained that psammophilous tene- brionidae that burrow in loose sand commonly share a syndrome of morphological specializa- tions. Important modifications, as discussed by Koch (1961), Doyen (1984:15), and others in- clude the following: enlargement of the foretibiae for digging; shortening of the antennae; increase in body pilosity, especially as lateral fimbriae; and development of obese, rotund body shapes. In ad- dition there is often development of protibial setal fringes, apparently to increase tibial area for dig- ging, and modification of the protarsi or protibial spurs. Parallelism in these characters is particu- larly evident in the Anepsiini, where most of the same apomorphies characterize psammophilous species of all four genera. Since the generalized body plan is similar throughout Anepsiini, the apomorphous states of these characters are dif- ficult or impossible to distinguish. However, sev- eral distinctive structural features, unrelated to psammophily, support the diagram in Figure 3. These are discussed where appropriate below. The characters subject to parallelism should ordi- narily be removed before computer analysis. I re- tain them here to emphasize the extensive level of homoplasy present in Anepsiini. The primary dichotomy separating Batuliodes from the remaining genera is based on differences in the mesocoxal region (Characters 18-20), an- tennae (Character 2), gular region (Character 8), and epistomum (Character 7). The sculpture of the hypomeron and the configuration of the poste- rior pronotal angles show derived states in the more generalized, surface-dwelling species. The absence of the derived states of these characters in some of the specialized burrowing species proba- bly represents a secondary absence. Exserted pos- terior pronotal angles also occur in Batulius seto- sus, but the majority of characters clearly show that B. setosus belongs in the Anepsius lineage. Batuliodes rotundicollis and B. confluens, the relatively generalized, surface-dwelling species, do not differ in important structural features. However, they differ in several details of cuticular sculpturing and diverge greatly in size and shape of the aedeagus (see species descriptions). The clade comprising B. wasbaueri, B. spatulatus, and B. obesus is united by a series of apomorphous character states related to a psammophilous mode of life. Batuliodes obesus is the most specialized member of this clade, having lost the pronotal ca- rina (Character 10), developed dense, long seta- tion on the venter (Character 17), and having en- tirely lost the elytral carinae (Character 16). In all other characters it is exceedingly similar to B. spa- tulatus. Batuliodes wasbaueri lacks apomorphies of the antennae (Character 5) and hypomeral 348 DOYEN: REVISION OF ANEPSIINI 5 | .2 CO CO 1 i = ' 1 ! i c i ! ? i i 1 i 4 0 3 1 2 5 ^ : 2 > > 0 ) £ : « > £ J C » c a i 1 1 T : C > 2 > * \ .< i l c 0 ; o 3 = 0 D C \ c : ' c ) c. 1 1 * C » i: 1 c ' a ) Z i "c i 1 i ! i « : i 4 1 ( ( ( i I !! • a ( > > >3 > > , | CO 5 1 2 4 4 4 4 4 4 4 4 >t 4 > < » < > < > 4 > I >j 4 »t > 4 > 4 > 4 > 4 } 4 > < \ { i 4 < < > > > > > > > 4 4 4 4 4 4 ANTENNAL LENGTH (4) TENTORIUM (6) POST.PRONOTAL ANGLES (12) VENTRAL SETATION (17) SETAL LENGTH (27) PRONOTAL MARGIN (9) APICAL ANTENNAL SEGMENT (5) HYPOMERON SCULPTURE (14) ANTENNAL LENGTH (3) ELYTRAL MARGIN (15) PROTIBIAL SHAPE (22) PROTIBIAL SETATION (24) PROTIBIAL SPURS (23) ELYTRAL SCULPTURE (16) BODY PROPORTIONS (26) ANTENNAL SHAPE (1) METASTERNUM LENGTH (21) SUBMENTAL GLAND (8) PRONOTAL CARINA (10) POST. PRONOTAL ANGLES (11) AEDEAGUS (25) EPISTOMAL MARGIN (7) STERNITE LOBES (19) ANTENNAL SHAPE (2) MESOCOXAL CLOSURE (18) MESOTROCHANTIN (20) HYPOMERON SCULPTURE (13) 1 minute, tuberculate 2 geographically variable 3 short, bristling 4 intermediate FIGURE 3. Distribution of character states across genera and species of Anepsiini. One possible (hand-generated) cladistic arrange- ment is indicated by the basal connecting lines. Black dots indicate character states presumed to be apomorphic. Note the numerous parallelisms among features adapted for life in aeolian sand. PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44, No. 15 349 sculpture (Character 14), but shares the remain- ing apomorphies. The alternate branch, comprising Anepsius, Batuliomorphus and Batulius, of the cladogram is unified by only a single character, the presence of a submental gland (Character 8) in males. This is shown as an apomorphy in Figure 3, but could also be a primitive feature of Anepsiini which has been lost in Batuliodes. Anepsius is distinguished by a single apo- morphy, the short metasternum (Character 21). Commonly, shortening of the metasternum is cor- related with winglessness in Coleoptera. Anep- siini are primitively wingless, and in most features Anepsius (especially A. delicatulus) is the most primitive member of the tribe. Unexpectedly the metasternum is longest in the most specialized burrowing species, and it is possible that the short condition is actually primitive to Anepsiini. Within Anepsius, A. valens and especially A. montanus show fossorial specializations of the protibiae and presumably represent a monophy- leticclade. Batulius and Batuliomorpha, which constitute the remaining lineage, share numerous apo- morphic states of characters involved in psam- mophilous life. On this basis one might recognize only a single genus. However, Batuliomorpha is further characterized by a distinctive, obese body silhouette (Character 26), reduction or loss of the posterior pronotal angles (Character 12), and the uniquely apomorphic tentorium (Character 6). In contrast, in Batulius the body shape is not mark- edly different from that in Anepsius, and the ten- torium is unmodified, while the posterior pronotal angles are strongly exserted, resembling those of Batuliodes. These characters suggest that the apo- morphic protibial, setal, and antennal character states shared with Batuliomorpha arose by con- vergence. An alternative arrangement might derive the lineage containing Batulius and Batuliomorpha from one of the fossorial species of Anepsius. Both A. montanus and A. valens show a few of the apomorphic features related to psammophilous life. However, such an arrangement would re- quire that metasternal length, presumably de- rived in Anepsius, be reversed in Batulius and Ba- tuliomorpha. It seems more likely that the specializations obviously related to psammophily arose independently in the two lineages. Within Batuliomorpha, the species comata and tibiodentata share a single, apomorphous feature, the smooth, subglabrous hypomeron. However, this is a common characteristic of psammophilous forms, and in the Anepsiini this characteristic oc- curs independently in Batuliodes. Apomorphies distinguish B. comata (loss of pronotal carina, Character 10) and B. tibiodentata (coarsely tri- dentate protibia; unique within Anepsiini, not shown in Fig. 3). Furthermore, B. comata occurs in the southern Mojave Desert, while B. tibioden- tata inhabits dunes in central and southern Baja California. Without more convincing evidence, it seems preferable to leave the relationships be- tween the three Batuliomorpha species unre- solved. Anepsiini LeConte Anepsiini LeConte, 1862:215; Horn 1870:276; LeConte and Horn 1883:367; Casey 1907:503; Arnett 1960:670; Doyen and Lawrence 1979:346. Batuliini Horn, 1870:270; Casey 1907:497; Arnett 1960:670; Doyen and Lawrence 1979:346 (synonymy). BatulinaePapp, 1961:105. BatuliniPapp, 1961:105. Wingless Tentyriinae 2-6 mm long with globu- lar pro thorax and oval abdomen. FEMALES. — Lateral epistomal sutures usually distinct, medial suture obliterated. Eye com- pletely divided by epistomal canthus and super- tended by low carina; antenna with 11 segments, clavate or with apical 3 or 4 segments forming club; labrum transverse with long, slender tormae with medial processes directed obliquely poste- riad; mandible with small, smooth mola remote from bidentate incisor lobe; maxilla with galea densely setose; lacinia with bidentate uncus; la- bium subhexagonal, moderate in size, exposing maxillary articulations; tentorium with sides short, bridge thick, or closed posteriorly (Fig. 2). Prosternal process unmargined, horizontal a short distance behind coxae, then abruptly decli- vous; mesosternum scarcely excavated. Elytra constricted basally, collarlike, amplected into pro thorax; scutellum angulate or rounded poste- riorly, retracted into prothorax, not visible on ely- tral disk. Mesocoxal cavities closed by sterna or nearly so; mesotrochantin exposed or concealed. Metasternum about one to two times length of mesocoxa. Metendosternite with short, thick stalk; long stout, tapering arms with tendons api- cal or subapical; median metasternal sulcus and 350 DOYEN: REVISION OF ANEPSIINI TROCHANTIN COXA MESOSTERNUM MESEPIMERON METASTERNUM FIGURES 4-6. Ventral aspect of pterothoraces, showing variation in mesocoxal structure in Anepsiini. 4) Batuliomorpha comata, 5) Anepsius valens, 6) Batuliodes rotundicollis. internal ridge absent. Femurs stout; foretibia di- lated, triangular; tarsi short with few spinose se- tae beneath. Ovipositor short, thick, with para- proct and coxite subequal; coxite lobes indistinct; gonostyles minute, papilliform, inserted dorsola- terally very near apex of coxite. MALES. — Aedeagus inverted with paramere 1.1-1.7 times longer than legmen; median lobe free, its lateral baculi fused proximally (Fig. 12- 15). Average length about 10% less than females; submentum perforated by circular opening with tuft of protruding setae (Doyen and Lawrence 1979; Fig. 16-19), except in Batuliodes. LARVAE . — Unknown . Key to Genera of Anepsiini 1. Middle coxal cavity with trochantin ex- posed; sternal lobes separated by nar- row space laterad of coxal cavity (Fig. 4,5) 2 Middle coxal cavity with trochantin con- cealed; lobes of sternites meeting laterad of coxal cavity (Fig. 6) Batuliodes Casey 2(1). Metasternum about twice as long as meso- coxa; pronotum and elytra fimbriate along lateral margins; apical antennal segment subquadrate or wider than long (Kg. 7, 8) 3 Metasternum about one to one and one- half times as long as mesocoxa; prono- tum and elytra subglabrous or with short, appressed setae; apical antennal seg- ment longer than broad (Fig. 9) Anepsius LeConte 3(2). Posterior pronotal angles obliterated or represented by minute tubercles; anten- nal club with three segments; lateral fim- briation long, flying (Fig. 31); venter with numerous long setae Batulio- morpha, new genus Posterior pronotal angles strongly angu- late; antennal club with four segments; lateral fimbriation short, stiff (Fig. 21); venter with few short setae Batulius LeConte Anepsius LeConte Anepsius LeConte, 1851:147, 1862:215; Horn 1870:277; Le- Conte and Horn 1883:367; Casey 1907:503. TYPE SPECIES. — Anepsius delicatulus LeConte; designated by Casey 1907:501. Relatively slender to moderately obese (0.55 0.72). 23. Protibial spurs: a) subequal, straight; b) mesial spur much 27. Setal length: a) pronotal and elytral nmbriae short (or ab- larger than lateral, curved posteriad (Fig. 17). sent); b) setae long, projecting. 24. Protibial setation: a) posterior margin glabrous or with short, appressed setae; b) posterior margin with three to seven long, erect or semi-erect setae (Fig. 17). INDEX TO VOLUME 44 (Compiled by Tomio Iwamoto) New names Acanthogilia 1 1 1-125 Asarcenchelys 12 Asarcenchelys longimanus 12-15 Cosmochilus cardinalis 3-7 Emoia campbelli 49-5 1 Emoia trossula 47-49 Gastrocopta cordillerae 244-245 Gastrocopta (Albinula) montana 241-243 Gastrocopta (Albinula) sagittaria 244 Helminthoglypta bozemanensis 255-256 Mixomyrophis 10 Mixomyrophis pusillipinna 10-11 Phenacostethus trewavasae 226-235 Pupoides (Ischnopupoides) tephrodes 246-247 Radiocentrum taylori 249-25 1 Radiocentrum laevidomus 251-252 New names in boldface type Acanthogilia 1 1 1-126 gloriosa 113-125 Adephaga 67-68, 73, 81, 83-92, 94-97 Agabus cordatus 98 Agonidae 17, 160 Agonopsis chiloensis 17 vulsa 19-22 Agonus 17, 31 cataphractus 17, 31 Ahlia 9 egmontis 15 Allophyllum 117 glutinosum 124 Ammonitella 252, 264 lunata 264 Ammonitellidae 238, 252, 264 Amphizoa 68-72, 78, 81, 83, 98-100, 102 carinata 67-68, 75-78 davidi 67-68, 70-73, 79-81, 98-102, 107 davidis 70 insolens 68, 70-75, 79-80, 82, 98-101, 103-106 josephi 70, 72 kashmirensis 68 lecontei 67-68, 70-80, 98-1 00, 103-104, 106-107 planata 75 striata 70-71, 74-75, 78-80, 82, 98-100, 103-106 Amphizoidae 67-69, 81, 84, 89, 100 Anableps 233 Anachis 278 Anachisl sp. indet. 278 Anchistoma parvulum 253 Anguispira 262 holroydensis 262 russelli 262 Anomia peruviana 59 Archostemata 83-84, 86-87, 89, 91, 94-97 Arionidae 262 Artediini 172 Artedius 157-163, 165-168, 170-175, 177, 182, 185, 188-189, 191, 194, 198-199, 213, 217, 221 corallinus 159, 161, 171-172, 188 creaseri 157-159, 161, 163, 166-168, 170-175, 177, 213-217, 221 fenestralis 157, 159, 161-168, 170-182, 184, 188 harringtoni 157-1 59, 161-162, 164-168, 170-175, 179, 181-185, 188 lateralis 157-164, 166-168, 170-175, 178-179, 182, 184-188 meanyi 157-159, 161, 163, 166-168, 170-175, 177,216-221 notospilotus 159, 161, 171-172, 188 type 2 158 type 3 157, 161-164, 168, 170, 172-175, 179, 182, 184, 188-191 Asarcenchelys 12 longimanus 9, 11-15 Aspidophoroides 38-39 bartoni 18-20, 22, 37 monopterygius 17 olriki 18-20, 22, 37-38 Asterotheca 18 Atherinomorpha 225 Avicula 60-61 Balanidae 60 Balaninae 60 Balanoidea 60 Balanus 60 calidus 55-56 poecilus 55-56, 59-61, 65 sp. cf. B. calidus 55-56, 59-60, 65 [285] 286 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44 tintinnabulum galapaganus 6 1 trigonus 55-56, 59-60, 65 Bathyagonus 38 alascanus 18-20, 22, 37 infraspinatus 18-20, 22, 37-39 nigripinnis 18-20, 22, 37-38 pentacanthus 18-20, 22, 37 Binneya 263 antiqua 263 Biomphalaria 258 glabrata 258 pseudammonius 239, 258 sp. cf. 5. pseudammonius 258 Bonplandia 117, 120-121 geminiflora 117, 123 (Peronaeus) 265 Bothragonus 38-39 swam 18-20, 22, 37 Brachinus 88 Buccinacea 277 Buccinidae 277 Bulimulidae 262 Bulimulus 262 sp. 262 Bulimus nltidulus 245 5wrra 276 caelata 275-276 Bursidae 276 Camaenidae 263, 265 Cantharus 111 (Gemophos) 211 (Gemophos) sanguinolentus 211 janellii 211 sanguinolentus 275, 277 Ca/tf«fl 111, 116-118, 120-121, 125 buxifolia 118-121, 125 candelilla 121-123, 125 pyrifolia 118, 123 quercifolia 121, 125 Cantueae 118, 125 Carabidae68, 71, 85 Carabinae 89 Carabini 87 Carabus 86 Caracolus 263 aquilonaris 263 Carbinae 83 Cardiacea 275 Cardiidae 275 Carlhubbsia kidderi 234 Carychium 261 sp. 276 Ceratostethus bicornis 228, 234 Charopa 262 Charopidae 262 Chelonibia 56, 58 testudinaria 55-58, 64 Chelonibiinae 56 Chitonotus 170, 173 Cicindelini 87 Clausiliidae 262 Clinocottus 157-1 60, 162-168, 170-175, 177, 199,201, 204,209,213,217, 221 acuticeps 157-162, 164-168, 170-172, 174-175, 199-204, 205, 207 analis 157, 159, 162, 164, 168, 170-171, 174- 175,203,205,209,212-213 embryum 157, 159, 162, 164, 166-168, 170-171, 174-175,203-208,210 globiceps 157, 159-162, 164, 166-168, 170-172, 174-175,203, 205,208-212 recalvus 157-159, 164, 168, 170-171, 174-175, 205, 209 Cobaea 117, 120-121, 125 biaurita 123 Codakia 65 Coleoptera 67, 85, 87, 92, 95 Collomia 117, 121 Collumbellidae 277 Columbella 211 castanea 275, 278 cf. C. strombiformis 211 lanceolata 278 major 278 pyrostoma 211 strombiformis 211 Colymbetinae 88 Conacea 279 Concavus 55-56, 61 (Arossia) 6 1 (Arossia) panamensis 56 (Arossia) panamensis eyerdami 55 (Arossia) sp. cf. C. panamensis eyerdami 55, 61- 62,65 aquila 6 1 henryae 6 1 panamensis eyerdami 6 1 panamensis panamensis 6 1 Conidae 279 Conus 279 (Asprella) 279 (Asprella) arcuatus 275, 279 (Chelyconus) 279 (Chelyconus) orion 279 (Cylindrus) 279 (Cylindrus) lucidus 275, 279 (Lithoconus) arcuatus 279 arcuatus 279 loomisi 219 lucidus 279 OA70H 279 vittatus 279 INDEX 287 Coptodava 96 Coptoclavidae 83 Coronula 58 diadema55, 57-58,61,64 Coronulidae 56 Coronulinae 58 Coronuloidea 56 Coscinodiscus 128 eccentricus 137 lacustris 128, 144 lineatus 131 Cosmochilus 1-7 cardinal is 1-7 falcifer 2-3, 6-7 harmandi 2-3, 6-7 Cottidae 157-158, 160, 173, 199, 204, 208 Coitus 167 Craterarion 263 pachyostracon 263 Cybister 88 Cyclocheilichthys 1 Cyclophoridae 261 Cyclopteridae 160, 168 Cymatiacea 276 Daedalochila 263 Dentatherina 225 merceri 226 Discidae 262 sahlbergii 72 Dytiscidae 68, 83-84, 89, 98, 101 Dytiscus 88 Ellobiidae261 Emoa samoensis 45 Emoia4l^4, 46, 48, 50 aneityumensis 42, 48 caeruleocauda 4 1 campbelli 41^2, 49-51, 52 concolor 41^43, 45-47, 49, 51-52 cyanogaster 4 1 cyanogaster tongana 46 cyanura 4 1 murphyi 41-43, 52 nigra4l^2, 49, 51 nigromarginata 42-43, 5 1 parked 41-43, 52 physicae group 42 resplendens 42, 45 samoense 43, 45, 47 samoensis 41-47, 49, 51-52 samoensis group 41-43, 48, 50-51 sanfordi 42, 49, 51 speiseri 42 trossula 41^4, 46-49, 51-52 Engina 111 pyrostoma 277 160, 168 Eodromeinae 83, 89, 97 Eohipptychia eohippina 27 1 £>*?tes 88 Eriastrum 117-118, 125 Eucalodium eophilum 260 Eumeces 44 samoensis 43, 47 Euprepes concolor 45 resplendens 45 samoensis 45 Fasciolariidae 278 Fasciolariinae 278 Fraginae 275 centrifugus 278 Gastrocopta 237, 240-241, 244, 258-259 (Albinula) 241, 243-244, 256-257, 259 (Albinula) contracta 241, 243, 256 (Albinula) dupuyi 244 (Albinula) falcis 241 (Albinula) holzingeri 241, 243, 256 (Albinula) montana 238, 241-244 (Albinula) proarmifera 241 (Albinula) sagittaria 238, 242-244, 257 (Albinula) sp. a 238, 242-244 (Albinula) tridentata 241 (Gastrocopta) 241 (Gastrocopta) cristata 243 (Gastrocopta) sp. 261 armifera 241-242 contracta 244, 256-257 cordillerae 238, 242, 244-245, 258 sagittaria 258 sp. cf. (7. montana 242, 258 Gastrodonta coryphodontis 263 imperforata 263 Gastropoda 237-238, 240, 275 Geadephaga 68, 81 Gehringia 88 C?i7/a 111, 116-117, 121-123, 125 gloriosa 111, 115, 117 j/jcwa 121-122 insignis 121 leptomeria 122 121-122, 125 121-122, 125 sect. Giliastrum 111, 118, 121-122, 125 sect. Leptodactylon 1 1 1 tricolor 122 288 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44 Gilieae 125 Glypterpes rotundatus 264 veternus 264 Gonyodiscus 248 Gymnosteris 117, 121 Gyrinidae68, 84, 88, 101 Haliotis 270 Haliplidae 89 Haplochromis elegans 229 Helicina cokevillensis 26 1 cretacea 26 1 vokesi 26 1 Helicinidae 26 1 Helix adapts 264-265 nacimientensis 248, 264-265 polygyrella 252 riparia 256 spatiosa 263 tudiculata 253 varians 255 Helminthoglypta 237, 253-259, 264 a#z 254 terry/ 259 bozemanensis 238, 254-256, 258-259, 264 californiensis 254-255, 259 JfeWi 259 hertleini 254 mailliardi 254, 259 nickliniana 255 nickliniana awania 259 obtusa 254 reederi 254 sp. 264 sp. cf. //. bozemanensis 255 tfocfa' 254 traskii 254 tudiculata 254 Helminthoglyptidae 238, 249, 255, 264 Hemilepidotus 166-167, 173 hemilepidotus 168 spinosus 160 Hemitrochus 255-256, 264 hoemastomus 255 sp. 264 Hendersonia 26 1 evanstonensis 26 1 oregona 26 1 Heterodonta 275 Hexagrammidae 160 Hexagrammos sp. 160, 168 86 Hodopoeus 265 crassus 263 hesperarche 263 Holospira 262, 265 adventicia 262 Jyen 262 grangeri 262 /«V/y/ 262 sp. 262 Horaichthys setnai 234 Humboldtiana 256 palmeri 256 //Hf/z/a 117-118, 120-121, 125 Hydaticus 88 Hydnophytum 50 Hydradephaga 67-68, 73, 81, 85-87, 91, 95-98, 101 Hydronebrius 68, 98 Hygrobiidae 68, 83, 88-89, 101 Hypsagonus 17, 30 quadricornis 17, 19-22, 30 Icelidae 158 Icelinae 158 157, 161-163, 166-167, 170-174, 177, 217, 221 sp. 160, 168 166, 173 Ipomopsis 111, 116-118, 123, 125 gloriosa 111, 115, 117 Isomeria 265 kanabensis 263 Labeo 1 Labeoinae 7 Laccophilinae 88 Langloisia 1 1 7 Latf/rws 278 centrifugus 275, 278 Leistus 86 Leptarionta 256 Leptocottus 166 armatus 160, 168 Leptodactylon 111, 117-118, 121, 123, 125 gloriosa 115, 117 gloriosum 111 pungens 124 Liadytidae 83, 89 Linanthus 1 1 7 dianthiflorus 124 grandiflorus 124 111, 116-118, 120, 123 amplectens 124 gloriosa 111, 115 INDEX 289 greggii 124 mexicana 1 1 8 purpusii 1 1 8 Luddella 261 buttsi 26 1 Lygosoma cyanogaster 45 cyanogaster tongana 45 samoense 43, 45, 47 Lymnaea 238 Lysinoe 256, 264 breedlovei 264 Malacocottus 166 Megabalaninae 6 1 Megabalanus 6 1 dippertonensis 6 1 galapaganus 55-56, 61-65 sp. indet. 62-64 tanagrae 6 1 Menesiniella 6 1 Mesogastropoda 275 sagensis 264 laevigatus 271 Microsteris 117, 121 gracilis 123 Mirophallus bikolanus 225-226, 232, 234-235 Mixomyrophis 10-11 pusillipinna 9, 10-11, 14 Monadenia 256, 260, 264 (Shastelix) marginicola 264 antecedens 264 dubiosa 264 chrysophekadion 1 Muraenichthys 9 puhioilo 1 1 Myoxocephalus 162-163, 172-174, 221 sp. 160, 168 Myrophinae 9 Myrophini 9 Myrophis 9-10, 15 vo/er 14 Nassariidae 278 Nassarius 278 caelolineatus 278 nodidnctus 278 w&?r 276 Naticacea 276 Naticidae 276 Navarretia 117, 121 fossalis 122 mitracarpa 122 86 Nebriini 83, 86-87 Necronectulus 84, 89-90, 94-96 Neenchelys 11, 14-15 Neostethidae 225 Neostethinae 225-226, 229, 231 Neostethus 229, 233 Noteridae 83, 89 Notiokasiini 83, 86 Notiophilini 87 dodecaedron 19-22 Odontopyxis 38-39 trispinosa 18-20, 22, 37-38 Oleacinidae 263 Oligocottinae 158, 171-172 Oligocottus 157-160, 162-168, 170-175, 177, 199, 208, 213, 217, 221 maculosus 157-164, 167-168, 170, 172-175, 191- 195, 197 rimensis 159, 163, 172 rubellio 159, 163, 172 snyderi 157, 159-164, 167-168, 170, 172-175, 194-199,221 Omma 88 Omphalina laminarum 263 oreodontis 263 Ophichthidae 9 Opisthiini 83, 87 jepseni 262 planispira 262 spp. 262 Oreohelicidae 238, 247-249, 264 Oreohelix 248-249, 251-252, 259-260 angulifera 248 chiricahuana 247 ste/'m 263 thurstoni 248 Orthonopias 173 Orthurethra 240 Osteochilus 1 Paleocydotus sp. 271 barbata 19-22 Parahygrobiidae 83 Paricelinus 173 Pelecypoda 275 Pelophila 86 japonicus 19-22 Phallostethidae 225, 233 Phallostethinae 225-226 290 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44 Phallostethus 225, 229, 231-233 dunckeri 225-226, 228-229, 232 Phenacostethus 226, 229, 231-232 posthon 225-228, 232-235 smithi 225-230, 232-233 thai 226 trewavasae 225-235 Phlox 117, 121 andicola 123 Phos 277 (Metaphos) 277 (Metaphos) laevigatus 211 chelonia 211 laevigatus 275, 277 P/zysa 239 Planorbis amplexus 253 Platymantis 41 Pleurodonte (Dentellarid) 263 (Dentellaria) sp. 263 (Pleurodonte) 263 (Pleurodonte) wilsoni 263 Pleurotomaria 212 Podothecus 17, 31 acipenserinus 19-22, 31 gilbert i 3 1 thompsoni 3 1 veternus 3 1 Polemoniaceae 111, 116-118, 121-123, 125 Polemonium 116, 121 Polinices 276 (Polinices) 276 (Polinices) uber 276 intemeratus 276 liter 275-276 unimaculatus 276 Polygyra 263 expansa 263 martini 263 petrochlora 263 sp. 263 venerabilis 253 veternior 263 Polygyrella 237, 252-253, 257-259, 264 amplexa 264 amplexus 253 parvula 253, 264 polygyrella 251-253, 257, 264 sp. 264 sp. cf. P. polygyrella 238, 251, 253, 257, 264 venerabilis 264 Polygyridae 263, 265 Polymita 264 texana 264 Polyphaga 97 Prosobranchia 261 Protorabinae 83, 89 Protornatellina isoclina 26 1 261 pupilla 26 1 uniplica 26 1 Pseudocolumna 265 haydeniana 262 spitzia 262 spp. 262 vermicula 262 Pseudomyrophis 11, 14-15 micropinna 14 Pteria beilana 6 1 peruviana 61 sterna 60 Pulmonata 237-238, 240, 261 Pupa acarus 240 contracta 241 hordacea 245 incolata 245 Pupillidae 238, 240-241, 243-244, 246-247, 261 Pupoides 237, 245-246 Pupoides (Ischnopupoides) 237, 258 (Ischnopupoides) 245-247 (Ischnopupoides) hordaceus 245-246, 261 (Ischnopupoides) inornatus 245 (Ischnopupoides) sp. 245, 247, 261 (Ischnopupoides) sp. cf. .P. (7. ) hordaceus 238, 246- 247,261 (Ischnopupoides) tephrodes 238, 245-247, 261 (Pupoides) albilabris 246 hordaceus 246-247, 258-259 inornatus 247, 259 tephrodes 246-247, 258 .Pwrpwra sanguinolentus 211 Radiocentrum 237, 247-249, 251-252, 256-258, 260 anguliferum 250, 252, 264 avalonense 257, 259 chiricahuanum 250, 252 chiricahuanum chiricahuanum 250 chiricahuanum obsoletum 249-250 grangeri 248, 264 hachetanum 251-252 hendersoni 248, 250-251, 264 laevidomus 238, 250-252, 257-258, 264 taylori 238, 249-252, 258, 264 thurstoni 250, 252, 264 Radulinus 166 asprellus 160, 168 caelata 276 INDEX 291 Rhiostoma 261 americana 26 1 Ruscariops 173 Salvia mellifera 257 Sarritor frenatus 19-22 Schizophoridae 83, 89 Scincidae 4 1 Scorpaenichthys 166 marmoratus 160, 168 Scorpaenidae 160 Scorpaeniformes 157 Sebastes 163 flavidus 160, 168 Shastelix 264 Sigmurethra 247 Sinilabeo 1 Spanglerogyrus 84, 89-91, 94-95 Stellerina 163 xyosterna 168 Stellerinna xyosterna 160 Strobilops 26 1 sp. 262 Strobilopsidae 261 Stombina 278 (Strombind) 278 (Strombind) lanceolata 278 gibberula 278 lanceolata 275, 278 recurva 278 Subulinidae 262 Systolosoma 87, 94, 96 Terebra 279 armillata 279 plicata 279 Terebridae 279 Tetradita 55-56, 58 milleporosa 55-61, 64-65 panamensis 58 porosa var. communis 58 rubescens 55, 60 rubescens rubescens 55, 58 sp. indet. 55, 58-60, 64 squamosa milleporosa 58 stalactifera 58-59 stalactifera confinis 58 stalactifera stalactifera 58 Tetraclitidae 58 Tetraclitinae 58 Thalassiosira 127-128, 130-132, 137-138, 142, 146, 151, 153-154 angstii 128 anguste-lineata 128, 146, 151, 153 decipiens 127-128, 137-138, 142, 144, 152-154 eccentrica 128, 137-138, 142, 153-154 endoseriata 128, 146, 151, 153 127-128, 130-132, 150, 153-154 128, 132, 146, 151, 153-154 lacustris 128, 146, 151, 153 leptopus 131 lundiana 128, 137, 142, 151, 153 minuscula 128, 132, 137-138, 153 nodulolineata 127, 130-132, 150-151, 153-154 nordenskioeldii 127-128, 130, 138, 153 oestrupii var. venrickae 130, 137, 153 paq/zca 127, 130, 138, 142, 153 punctigera 128, 130, 137-138, 142, 151, 153 /•0/w/a 127-128, 130, 146, 151, 153 simonsenii 130-132, 153 130, 146, 151, 153 130-132, 153 v«M/gis 127, 130, 138, 142, 146, 151-153 wongii 127, 130, 132, 136, 152-154 Tornatellinidae 26 1 Tozerpina douglasi 26 1 mokowanensis 26 1 rutherfordi 26 1 Trachypachidae 82, 85, 87 Trachypachinae 83, 89, 97 Trachypachus 94 Triadogyrus 90, 94 Triaplidae 89-90 Triglops 173 Trigonocardia 275 biangulata 275 Trigonocardia^ sp. 274-275 Triodopsis 263 spp. 263 Turridae? 280 Turritella 275-276 broderipiana 275-276 broderipiana marmorata 274-276 gonostoma 275-276 marmorata 275 nodulosa 276 rubescens 274, 276 Turritellacea 275 Turritellidae 275 Turritellinae 275 Urocoptidae 262 Veneroida 275 Ventridens 263, 265 /my 263 sp. 263 Vespericola 263 263 292 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 44 Xeneretminae 18 Xeneretmus 17-18, 30-36, 38 (Xeneretmus) 18, 36 (Xeneretmus) triacanthus 17-18, 32, 36-38 (Xenopyxis) 17-18, 32, 36, 38-39 (Xenopyxis) latifrons 17-18, 32-35 (Xenopyxis) leiops 17-18, 32, 35-36 (Xenopyxis) ritteri 17-18, 32, 36 infraspinatus 17 latifrons 18-21,31-35,37 leiops 17-21,32-34,36-37 ritteri 17-21, 30, 32-34, 36-38 triacanthus 18-38 Xenochirus 17, 31-32 alascanus 17 latifrons 1 7, 32 pentacanthus 17 triacanthus 17, 32, 36 Xenodexia ctenolepis 234 Xenopyxis 32 latifrons 32 36 Xerarionta 256, 264 waltmilleri 264 Zonitidae 263