PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024 Number 3316, 15 pp., 6 figures, 1 table January 30, 2001 Generic Recognition for a Neglected Lineage of South American Pitvipers (Squamata: Viperidae: Crotalinae), with the Description of a New Species from the Colombian Choco´ RONALD L. GUTBERLET, JR.1 AND JONATHAN A. CAMPBELL2 CONTENTS Abstract ....................................................................... 2 Introduction .................................................................... 2 Materials and Methods .......................................................... 2 Systematic Account ............................................................. 3 Bothrocophias, new genus ....................................................... 4 Bothrocophias myersi, new species ............................................... 7 Discussion .................................................................... 10 Acknowledgments ............................................................. 12 References .................................................................... 12 Appendix: Bothrocophias Specimens Examined ................................... 15 1DepartmentofBiology,The UniversityofTexasatTyler,3900 UniversityBoulevard,TylerTX75799. 2Research Associate, Division of Vertebrate Zoology (Herpetology), American Museum of Natural History; De- partmentofBiology,Box 19498,The UniversityofTexasat Arlington,ArlingtonTX76019. 2 AMERICAN MUSEUM NOVITATES NO. 3316 ABSTRACT The name Bothrops campbelli Freire-Lascano (1991) is currently applied to two distinct species.Werestrictuseofthenamecampbellitothespeciesthatoccursonthewesternslopes of the Andes in Ecuador, and we describe the species that occurs in the Chocoan rainforest of western Colombia. Besides B. campbelli, the closest living relatives of the new species appear to be B. hyoprora and B. microphthalmus of the Atlantic slopes and lowlands of northern South America. Together these four species comprise a distinctive clade for which we propose the new generic name Bothrocophias. INTRODUCTION al. (1999) for additional comments on this issue. TheChoco´ regionofsouthwesternColom- Herein, we demonstrate that the name bia and northwestern Ecuador isnotedforits Bothrops campbelli applies to a highland heavy rainfall, remarkable biological diver- species known only from the Pacificslopeof sity, and high number of endemic species. theEcuadorianAndes,andweproposeanew During a trip to the Colombian Choco´ in name for the lowland species from the Col- 1973,CharlesMyersandJohnDalycollected ombian Choco´. Also, in a continuing effort a seriesof pitvipersthatobviouslyrepresents to resolve the paraphyly of the genus Both- an undescribed species; these specimens rops, we recognize a new genus for a dis- were deposited in the American Museum of tinctivelineagefromnorthernSouthAmerica Natural History (AMNH). Additional mate- that includes Bothrops campbelli, B. hyopro- rialofthisspeciesishousedintheFieldMu- ra, B. microphthalmus, and the new species. seum of Natural History (FMNH), The Uni- versity of Texas at Arlington Collection of MATERIALS AND METHODS Vertebrates(UTA),andtheUnitedStatesNa- Comparative data for diagnosing the new tional Museum (USNM). species were obtained through examination Recent treatments of Neotropical pitvipers of museum specimens (appendix I) and from (e.g.,CampbellandLamar,1989,1992)have information provided by Campbell and La- discussed specimens of the new species to- mar (1989, 1992), Freire-Lascano (1991), getherwithspecimensofBothropscampbelli and Nice´foro-Mar´ıa (1975). Freire-Lascano (1991) under the name Both- The following features were measured rops pulcher (Peters, 1862). Scha¨ttiandKra- with dial calipers or with a meter stick and mer (1993) discovered that the holotype of string: snout-vent length (SVL), tail length Bothrops pulcher (Peters) actuallyrepresents (TL), distance from anteroventral corner of what had been referred to in much of the eye to caudal border of pit (EP), distance literature as Bothriopsis albocarinata from anterodorsal corner of eye to center of (Shreve,1934)andthereforethenamepulch- naris (EN), horizontal distance across eye er has priority. After receiving several com- (ED), distance from tip of snout to angle of ments and recommendations regarding this jaw (HL), height of rostral taken at midline taxonomic confusion (Gutberlet and Harvey, (RH), width of rostral taken at widest point 1998; Kuch, 1997; Scha¨tti and Smith, 1997; (RW). Wu¨ster, 1998), the InternationalCommission Terminology for most scales follows on Zoological Nomenclature(1999)resolved Klauber(1972).Interoculabialswerecounted theissueasfollows(Opinion,1939):thecor- in a vertical line from below the pupil of the rect name for the highland, arboreal pitviper eye to the mouth, including one subocular recently known as Bothriopsis albocarinata and one supralabial. Prefoveals are the small isBothriopsispulchra(Peters),andthesnake scales bounded by the supralabials, nasals, referred to as Bothrops pulcher by Campbell loreals, lacunals, and subfoveals (if present). and Lamar (1989) and many previous work- The level of the anterior edge of the pit de- ers is now appropriately known as Bothrops marcates the separation between prefoveals campbelli Freire-Lascano. SeeMcDiarmidet and subfoveals. Canthalsare thescalesalong 2001 GUTBERLET AND CAMPBELL: SOUTH AMERICAN PITVIPERS 3 the lateral edge of the dorsal surface of the riopsisrenderstheremainingterrestrialBoth- head,andaresituatedbetweentheinternasals rops paraphyletic. Including the species of andthesupraoculars.Canthalsarecountedin Bothriopsis in the genus Bothrops, as sug- a single row, and the small scalesadjacentto gested by Saloma˜o et al. (1997), would at the supraocularsarealsoincludedinthecan- least partially rectify this problem. An alter- thal series, even when these scales are ex- native is to partition the large and unwieldy cluded from the canthus rostralis proper by genusBothrops,ascurrentlyrecognized,into the dorsalmost preocular. This method of several genera and continue to recognize counting canthals is the same one used by Bothriopsis (Campbell and Lamar, 1992; Gutberlet (1998a) and is consistent with McDiarmid et al., 1999). We prefer thelatter Klauber’s (1972) definition. Intersupraocu- approach, which allows continued recogni- larsarecountedinastraightlinebetweenthe tion of the distinctive Bothriopsis clade and two supraoculars, which are not included in which will in our opinion better describe the the series. Gulars were counted in a straight evolutionary history of South American pit- line between the chin shields and the first vipers. scale on the underside of the head, which is In phylogenetic analyses using characters widerthanlong.Interrictalcountsincludethe from gross anatomy, Gutberlet and Harvey last supralabial on each side. Dorsal scale (in press) identified a clade that includes rows were counted one head length behind Bothrops campbelli, B. microphthalmus, and the occiput, at midbody, and one headlength B. hyoprora. This group was recovered in anterior to the vent. The first ventral is con- every analysis with bootstrap support rang- sidered to be the first scale on the underside ing between 75 and 100%. In a separate of the head that is bordered on both sides by study based on mitochondrialDNAevidence the first row of dorsals(Dowling,1951).The (Parkinson et al., in press), the sister rela- last ventral is the scale anterior to the anal tionship of B. hyoprora and B. microphthal- plate. We use slashes to denote paired struc- mus was strongly supported; B. campbelli tures on the left/right sides of the body. Ter- was not included in that study. A fourth spe- minology for hemipenes is that of Dowling cies,whichwedescribebelow,sharesseveral and Savage (1960). unique derived character states with these The decision to recognize a new genusfor species,indicatingthatitispartofthisclade. thecladewediscussbelowisbasedprimarily Could this monophyletic group be accom- on a suite of shared, derived charactersiden- modated within an existing genus? Though tifiedinphylogeneticanalysesofNewWorld the clade was recovered in every analysis pitviper species (Gutberlet and Harvey, in conducted by Gutberlet and Harvey (in press; Parkinson et al., in press). All other press), its position relative to other South recent phylogenetic studies of pitvipers, American lineages varied among analyses though they do not address the monophyly (fig. 1). Seven of eight analyses indicated of the clade in question, have presented hy- that the B. hyoprora clade is more closely potheses of pitviper relationships that are related to Lachesis than it is to Bothrops (cid:49) consistent with the classification we propose Bothriopsis. If this hypothesis of relation- (see Gutberlet, 1998a; Kraus et al., 1996; Parkinson, 1999; Saloma˜o et al., 1997; Vidal ships is accurate, a new generic nameforthe and Lecointre, 1998; Vidal et al., 1997; Wer- hyoprora clade is needed to rectify the par- man, 1992; Werman et al., 1999). We care- aphyly of Bothrops, regardless of how the fully considered each of these studies in species of Bothriopsis are treated. One of reachingourdecision,whichwediscussfully eight analyses indicated that the hyoprora below. clade is sister to Bothrops (cid:49) Bothriopsisand thatLachesisfallsoutsidethisclade.Inevery analysis conducted by Parkinson et al. (in SYSTEMATIC ACCOUNT press), the hyoprora clade was found to be As pointed out by Campbell and Lamar sister to Bothrops (cid:49) Bothriopsis. Thus, even (1992),recognitionofthedistinctivecladeof if the species of Bothriopsis were to be in- South American arboreal pitvipers as Both- cluded in Bothrops, the generic partition we 4 AMERICAN MUSEUM NOVITATES NO. 3316 Fig. 1. Alternative hypotheses of relationships among species of Bothrops, Bothriopsis, Lachesis, and Bothrocophias new genus. (A) Recovered in six of eight different analyses of anatomicaldataonly (Gutberlet and Harvey, in press); (B) recovered in one reanalysis of Kraus et al.‘s ND4 sequence data (Gutberlet, 1998b); (C) recovered in one of eight analyses of anatomical data only (Gutberlet and Harvey, in press) and in all analyses of sequence data from four mitochondrial genes (Parkinson et al., in press); (D) recovered in one of eight analyses of anatomical data only (Gutberlet and Harvey, in press). propose represents progress toward a natural America. Members of the four species in- classification of New World pitvipers. cluded in thislineageareofmoderatelength, relatively stout-bodied, and terrestrial, lack- Bothrocophias, new genus ing a prehensile tail. Females attain greater size than do males. The snout is weakly (B. TYPE SPECIES: Bothrops hyoprorusAmaral campbelliandthenewspeciesdescribedsub- (1935), by present designation. sequently in this paper) to strongly (B. hy- ETYMOLOGY: The generic name is derived oprora and B. microphthalmus) elevated from the Greek words bothros, meaning pit, with a rostral scale that is approximately as and kophias, meaning snake or adder; the high as broad ordistinctlyhigherthanbroad. gender of this name is masculine. CONTENT: The genus Bothrocophias con- Dorsal coloration within the genus consists tains four species: campbelli known only mainly of darker shades of brown and red- from the western slopes of the Ecuadorian dish-brown. A patternofdorsalbandingmay Andes; hyoprora distributed throughout the be clearly evident or subdued, and in some western lowlands of the Amazonian rainfo- specimens the bands do not meet evenly at rest; microphthalmus, found along the east- the middorsal line and are staggered as large ern versant of the Andes in Colombia, Ec- lateral blotches. uador, Peru, and possibly Bolivia; and a new Derived characteristics shared by the four species described below that is restricted to species of Bothrocophias include small, lowland rainforest in western Colombia (fig. smooth intersupraocular scales; distinctive 2). white spots on gular and infralabial scales; DEFINITION AND DIAGNOSIS: A lineage of and tuberculate keels on scales on thecaudal crotaline snakes occurring in northern South portion of the dorsum. Many, though not all, 2001 GUTBERLET AND CAMPBELL: SOUTH AMERICAN PITVIPERS 5 6 AMERICAN MUSEUM NOVITATES NO. 3316 specimens of B. hyoprora and B. microph- diagnoses Bothrocophias from all other New thalmus exhibit tiny scales between the ros- World pitviper genera: (1) 7–8 supralabials; tral and internasals that have not been ob- (2)smallbutunkeeleddorsalheadscales,not served in any other New World pitviper. arranged in a nine-plate, colubrid-like pat- There are 124–177 ventrals; 38–64 sub- tern; (3) tubercular keels on dorsal scales on caudals, most of which are entire in B. hy- posterior half of body; (4) white spots with oprora, but mostly divided in the other three dark borderson somegularsandinfralabials; species; 21–25 middorsal scale rows; 2–9 (5) 124–177 ventrals; (6) 4–5 palatine teeth; smooth intersupraoculars; 7–8 supralabials; (7)12–15 pterygoid teeth;(8)14–16dentary 8–11 infralabials; 1–11 prefoveals; 1–3 can- teeth; (9) maxillary fang approximately 1.5 thals; and 3–4 interoculabials. timeslongerthan heightofmaxilla;(10)me- Theosteologyofthegenusremainspoorly sial spines present on hemipenial lobes; (11) known, though we have been able to exam- moderate number (ca. 18–30 perlobe)oflat- ine one skull each of Bothrocophias hyopro- eral spines on hemipenes; (12) hemipenial ra and B. microphthalmus. In these two spe- lobes only slightly longer than organ’s base; cies, the dorsal surface of the frontal bones (13) choanal process of palatine attenuate is predominantly flat. The postfrontals are distally; (14) ectopterygoid and base of pter- large, contributing more to the dorsal perim- ygoid approximately equal in length; (15) eter of the orbit than does the parietal. The dorsal surface of frontal bones predominant- posterolateral edges of the dorsal surface of ly flat; (16) postfrontal bones large, contrib- the parietalbearasmalllateralshelfofbone. uting more to dorsal perimeter of orbit than The ectopterygoid has a single pit for the at- does parietal; (17) tail not prehensile and tachment of the ectopterygoid retractormus- lacks a rattle. cle, is approximately equal in length to the Because the species of Bothrocophias base of the pterygoid, and has a broad, flat have been previously included in the genus shaft. The choanal process is positioned me- Bothrops, it is significant that skull mor- dially on the palatine bone and is attenuate phology in the two genera is markedly dif- distally. The angular and splenial are partial- ferent. In Bothrops (sensu stricto), including ly fused, and the Meckellian foramen is di- such species as B. asper and B. atrox, the vided by a thin extensionofbone.Speciesof skull is relatively narrow and elongate (Wer- Bothrocophias have 4–5 palatine teeth, 12– man, 1992, 1999; Gutberlet, 1998a; Gutber- 15 pterygoid teeth, and 14–16 dentary teeth. let and Harvey, in press), whereas the skull The maxillary fang is approximately 1.5 in Bothrocophias species is broad and heavi- times longer than the height of the maxilla. lyossified.Specificcharacteristicsthattypify WehaveseenhemipenesofBothrocophias this divergence in skull morphology are hyoprora and the species described in this (condition of Bothrops in parentheses): dis- paper. Hemipenes of these species are calyc- tance across frontals equal to (less than) ulate distally.Wallsofthemoreproximalca- width of skull at anterior end of supratem- lyces are spinulate. Proximal to the calyces, porals, shaft of ectopterygoid wide and flat small mesial spines and a moderate number (slender and round in cross section), and (ca. 18–30 per lobe) of lateralspinesarepre- shaft of ectopterygoid equal to (greaterthan) sent. Though some lateral spines are notice- length of base of pterygoid. ably larger than the mesial spines, none can The holotype of Bothrocophias campbelli be characterized as basal hooks. The hemi- (INHMT 1956) is from Provincia de Chim- penes of B. hyoprora differ slightly from borazo, Ecuador. As Campbell and Lamar thoseofthenewspeciesinhavinglongerand (1992) noted, all specimens from Colombia more slender lateral spines. previously assigned to B. campbelli possess The following combination of characters a lower number of ventrals and coloration ‹ Fig. 2. Map of northwestern South America, showing geographic distribution of species of Both- rocophias. Ranges of B. hyoprora and B. microphthalmus are based on Campbell and Lamar (1989). 2001 GUTBERLET AND CAMPBELL: SOUTH AMERICAN PITVIPERS 7 different from those from further south. coloration in life; (8) distinctive white spots These differences, in addition tootherslisted on some gular and infralabial scales. below, clearly indicate that the Colombian COMPARISONS: Bothrocophias myersi and snakes represent a distinct species, and this B. campbelli are easily distinguished from may be known as their Amazonian congeners by the presence ofalacunolabial(prelacunalnotfusedtosec- ond supralabial in B. microphthalmus and B. Bothrocophias myersi, new species hyoprora). Additionally, B. hyoprora is the only member of the genus with undivided Botrops lanceolatus—Garc´ıa, 1896: 22 [not of Bonnaterre, 1790; placement of this name here subcaudals. Bothrocophias myersi differs is tentative]. from B. campbelli in having fewer ventrals Bothrops pulcher [not of Peters, 1862]—Pe´rez- (139–151 vs. 162–177), fewer subcaudals Santos and Moreno, 1988: 358 [in part, also (44–52 vs. 48–64), fewer intersupraoculars included undetermined taxa from Amazonian (3–6 vs. 6–8), and fewer prefoveals (1–2 vs forests of Ecuador, Peru, and Colombia]; 2–4). All known specimens of B. myersi ex- Campbell and Lamar, 1989: 221 [in part, in- cept one (UTA R-21689, a small specimen cluded B. campbelli]; Campbell and Lamar, with a banded pattern) have a distinctive 1992: 11 [in part, included B. campbelli];Va´z- brown-bronze dorsum in preservative,which quez de Kartzow, 1995: 86. differs from the distinctly banded dorsum of Porthidium almawebi Scha¨tti and Kramer, 1993: juvenile and adult B. campbelli. Because the 258. Holotype: MHNG 2248.12. Type locality: ‘‘San Francisco de las Pampas (0(cid:56)26(cid:57) S x holotype of B. myersi was reddish-brown in 78(cid:56)57(cid:57) W, Cotopaxi, ca. 1’800 m u¨ M.)’’ [In lifeandisnowbrown-bronzeinpreservative, part, included B. campbelli]. we presume that all the brown-bronze spec- Porthidium almawebi—Golay et al., 1993: 83 [in imens were reddish-brown in life. Table 1 part, included B. campbelli]. summarizes additional variation among the Bothrops campbelli—McDiarmid et al., 1999: four species of Bothrocophias. 259 [in part, included B. campbelli]. Bothrocophias myersi is potentially sym- HOLOTYPE: American Museum of Natural patric with five pitviper species: Bothriechis History (AMNH) 109812 (original field schlegelii, Bothriopsis punctata, Bothrops number CWM 11878); an adult female (figs. asper, Lachesis stenophrys, and Porthidium 3, 4) from Quebrada Guangu´ı, 0.5 km above nasutum (Campbell and Lamar, 1989; Za- the R´ıo Patia, Department of Cauca, Colom- mudio and Greene, 1997). Bothriechisschle- bia, 100–200 m, collected by C. W. Myers gelii and Bothriopsis punctata have prehen- and J. W. Daly, 16–17 February 1973. sile tails (tail not prehensile in B. myersi). PARATYPES: All from Colombia: Cauca: Also, B. schlegelii has undivided subcaudals QuebradaGuangu´ı,ca.0.5kmabovetheR´ıo and spinelike superciliary scales, and B. Patia,100–200m(AMNH107919–20);Val- punctata has many more ventrals (175–213) le del Cauca: Pacific coast on road Buena- than does B. myersi. Bothrops asper and ventura–R´ıo Calima (FMNH 165586–96); Lachesis stenophrys attain considerably Caimancito, S of Buenaventura, on bank of greater lengths and also have higher ventral R´ıo Cajambre, 75 m (UTA R-21689); Bue- scale counts (161–240 and (cid:46)199,respective- naventura,nearR´ıoRaposo(USNM151708, ly). The twice-divided distal subcaudals of 154051). Lachesis are unique to that genus. Porthi- DIAGNOSIS: A moderately stout, terrestrial dium nasutum bears 24–41 undivided sub- pitviper reaching a maximum known total caudals, lacks a lacunolabial, has a rostral length of 756 mm that may be distinguished scale that is markedly higher than broad,and from all other New World rattlelesspitvipers has fewer ventrals (123–145). by the following combination of characters: DESCRIPTION OF HOLOTYPE: An adult fe- (1)139–151ventrals;(2)44–52dividedsub- male; rostral almost rectangular, but slightly caudals; (3) 21–23 dorsal scale rows at mid- wider ventrally, almost as wide as high (4.2 body; (4) prelacunal fused with second su- (cid:51) 4.3 mm); nasal incompletely divided pralabial; (5) 3–6 smooth intersupraoculars; above naris, distinctly divided below naris; (6) 7 supralabials; (7) reddish-brown dorsal loreal single, bordered anterodorsally by 8 AMERICAN MUSEUM NOVITATES NO. 3316 Fig. 3. Bothrocophias myersi, holotype (AMNH 109812), 756 mm total length. larger canthal and by upper preocular pos- posterior canthal tiny, separating anterior teriorly, longer than high; prefoveals 2/2; canthal from supraocular; intercanthals 3, subfovealsabsent;prelacunalfusedtosecond immediately posterior to internasals; inter- supralabial; interoculabials 3/3; postfoveals canthals 5, between posterior canthals; inter- 1/1, longer than high, extending posteriorly supraoculars 6; all dorsal head scales, from beyond posteroventral corner of pit, contact- level of posterior of supraoculars forward, ing lower preocular posteriorly and lacuno- smooth; supraoculars longer than wide but labial anteriorly; preoculars 3/3; upper preo- distinctly wider than adjacent intersupraocu- cular longer than high, contributing to can- lars; dorsal scale rows 25–23–19; ventrals thus; middle preocular much shorter, con- 148, ultimate scale in series extending only tacting but not fused to supralacunal; lower partially across venter; anal entire; subcau- preoculartiny,almostexcludedfromorbitby dals 48, all divided; tail spine slightly com- subocular and middle preocular; suboculars pressed laterally, curving slightly dorsad, as 1/1, each partially divided anteriorly; post- long as adjacent 3 subcaudals, tip blunt; dor- oculars 2/2; supralabials7/7;infralabials9/9, sals extending onto tail spine only slightly first infralabial on left side partially fused to fartherthansubcaudals;dorsalscalesofmid- mental, first pair of infralabials barely con- dorsum at midbody twice as long as wide tact each other medially; mental approxi- anteriorly, becoming wider posteriorly such mately as wide as long; chin shields longer that length only slightly exceeds width; dor- than wide, contacting first four infralabials sal body scales distinctly keeled, keel does oneachside;gularsbetweenchinshieldsand not extend to terminus of scale, keels dis- first scale on underside of head that is wider tinctly tuberculate on posterior portion of than long 4/5; 4 internasals, middle two body; paraventrals smooth over most of smallest and partially fused medially; can- body,veryweaklykeeledposteriorly,similar thals 2/2, anterior canthal distinctly larger, in size and shape to adjacent dorsals; para- 2001 GUTBERLET AND CAMPBELL: SOUTH AMERICAN PITVIPERS 9 patterned, pale brown to bronze in color; a wide, pale, brown-bronze postocular stripe extending from the posteroventral corner of eye to cover supralabial 7, all but antero- ventral corner of supralabial 6, and postero- dorsal edge of supralabial 5; pale postocular stripe bordered dorsally by a thin, dark brownstripe,andanteroventrallybyathicker stripeofsamedarkbrowncolor;mostofside of head and rostral dark brown, but with a small triangle of pale brown-bronze occu- pying anteroventral corner of supralabial 4 and posteroventral corner of supralabial 3; dorsum brown-bronze with a greater amount of dark brown or dark gray coloration later- ally; most paraventrals partly dark gray- brown and partly cream-colored; no clear pattern discernible on dorsum; tail plain brown dorsally, darker than dorsal body col- oration; ventral scales predominantly cream but also with a small amount of brown pig- ment, mostly concentrated laterally; subcau- dalsalsocreambutwithslightlymorebrown pigment than ventrals; ventral surface of Fig. 4. Bothrocophias myersi, paratype headcreambutwithdistinctivecolorationon (FMNH 165593), dorsal view of head, head several scales; mental and first 2 infralabials length 35 mm. on each side dark brown; 5 infralabials on each side and a total of 8 gular scales con- tainingwhitespotswithirregulardarkbrown subcaudals weakly keeled anteriorly, smooth edges. posteriorly. VARIATION: On each side of the head, Measurements of holotype (in millime- specimens of Bothrocophias myersi have 1– ters): SVL 597, TL 92, EP 1.7, EN 6.7, ED 2 prefoveals, 1–2 canthals, 3–4 interocula- 3.9, HL 30.1, RH 4.3, RW 4.2. bials, 1 postfoveal, 3 preoculars, 1–3 subo- In hisfieldnotes,Myersdescribedthecol- culars, 1–2 postoculars, 7 supralabials, 8–10 or in life of the holotype as follows: ‘‘Dark infralabials, and 3–5 gulars between the chin red-brown, posteriorly turning light blue- shield and first ventral. Additional variation gray onsides.Palegrayishblueobliquelines in features of scalation includes 3–4 inter- from eye, and small spots of same color un- nasals, 2–3 anterior intercanthals, 4–7 pos- derneath head. Under head and venter or- terior intercanthals, 3–6 intersupraoculars ange, except turning white down center on (fig.4),21–26anteriordorsalscalerows,21– posterior half of belly.’’ From our inspection 23 middorsal scale rows, 16–19 posterior of color slides of the holotype (AMNH dorsalscalerows,139–151ventrals,and44– 109812) taken before preservation, we can 52 divided subcaudals. Coloration is similar add thefollowing:the dorsalcolorationisal- in all type specimens, with the exception of most uniform, but slightly darker bands are one small male (UTA R-21689) with fairly discernible against the red-brown ground distinct bands on the dorsum. Measurements color(fig.3).Fromthesupraocularsforward, (in millimeters) for specimensin thetypese- the top of the head is dark brown, as is most riesvaryasfollows.Males:totallength376– of the side of the head. The iris appears to 677, SVL 311–577, tail length 48–100. Fe- be red-brown, paler above. males: total length 327–756, SVL 280–655, Color in preservative (70% ethanol after tail length 47–120. Table 1 summarizes var- 10% formalin): Dorsal surface of head un- iation within B. myersi. 10 AMERICAN MUSEUM NOVITATES NO. 3316 HEMIPENES: One paratype of Bothroco- (table1).BothrocophiashyoproraandB.mi- phias myersi (UTA R-21689) was preserved crophthalmus do not have lacunolabial with its hemipenes everted. The total length scales; rather, in these species the prelacunal of the right hemipenis is equivalent to the and the second supralabialarediscretescales length of the first five subcaudals. The lobes that are separated either by a suture or by bifurcate at a length of two subcaudals from tiny subfoveal scales. In many, but not all, the base, and the sulcus spermaticus divides specimens of B. hyoprora and B. microph- at a distance approximately one subcaudal thalmus, tiny scales are present between the from the base. As in most other crotalines, rostral scale and the internasals-canthals(fig. the organ is spinous proximally and calycu- 5). These unusual scales have not been re- late distally. There are 18–20 lateral spines ported in any other New World pitviper spe- per lobe, including 6–8 that are relatively cies and are here termed ‘‘canthorostrals.’’ large (approximately 1 subcaudal in length). Canthorostrals do not seem to be homolo- Small mesial spines are also present on the gouswiththesmallscalesseparatingtheros- proximalhalfofeachlobe.Themesialspines tralandprenasal(nasorostrals)thatarefound extend as far distally as the lateral spines. in some populations of Atropoides nummifer The distal half of each lobe is entirelycalyc- (Burger, 1950; Werman, 1984). The synapo- ulate. Many of the proximal calyces have morphy of canthorostrals strongly suggests spinulate walls, but walls of the more distal that the two eastern species are sister taxa, calyces are smooth. but available data do not allow resolution of DISTRIBUTIONANDHABITAT:Bothrocophias the branching sequence of the two western myersi is known only from the Pacific low- species (fig. 6). lands (ca. 75–200 m) of Cauca and Valle del The phylogenetic evidence suggests that Cauca,Colombia. Theoriginalhabitatinthis speciation in Bothrocophias may have been region consists of very wet rainforest (prob- precipitated when formation of the Andes ably (cid:46)5000 mm of rain per year). Myers et isolatedtheChocoanrainforestfromtheAm- al. (1978: 321–324) provided a detailed de- azonian rainforest, resulting in a geographic scription and photographs of the habitat at barrier to gene flow between eastern and the type locality. This type of habitat for- western populations of the common ancestor merly occupied lowland areas of southwest- of this clade. This is consistent with Dixon’s ern Colombia and adjacent areas in north- (1979) hypothesis that theChocoanandAm- western Ecuador. It is possible that B. myersi azonian rainforests were connected until the occurs (or used to occur) throughout the Upper Pliocene, when the Huancabama De- Choco´ region, although we know of no re- flection was closed by uplifting in the south- cords of this species from Ecuador. Unfor- ern Andes. Speciation within the easternand tunately, extensive habitat destruction western components of this clade may have throughout the Ecuadorian Choco´ region been the result of elevational segregation:B. makes itunlikelythatB.myersicurrentlyoc- microphthalmus and B. campbelli are upland curs in that country. species, whereas B. hyoprora and B. myersi ETYMOLOGY: The specific epithet, a noun are lowland species (fig. 2, table 1). in the genitive case, is a patronym in honor The identification of monophyletic groups of Dr. Charles W. Myers, who collected the within the large radiation of New World pit- new species and who has made many signif- vipers has proceeded steadily in recent years icant contributions to Neotropical herpetolo- (Campbell and Lamar, 1989, 1992; Crother gy through dedicated work in both the field et al., 1992; Werman, 1992; Gutberlet, and laboratory. 1998a). However, relationships among these monophyletic groups have been less tracta- ble: intergeneric relationships among New DISCUSSION Worldpitvipersarelargelyunresolved.Com- The two species of Bothrocophias thatoc- pare, for example, the findings of Werman cur on the Amazonian side of the Andes (1992), Kraus et al. (1996), and Parkinson share several derived characterstatesthatare (1999). Whereas Werman’s analysis identi- not found in the species west of the Andes fied Ophryacus as sister to Bothriechis, Par-