THE PEARCE­SELLARDS Series NUMBER 12 Geographic Variation in Brazilian Species of Hyla by BERTHA LUTZ October, 1968 TEXASMEMORIALMUSEUM/24TH&TRINITY/AUSTIN, TEXAS W. W. NEWCOMB, JR., DIRECTOR Contents Abstract 3 New Subspecies ofHylacatharinae H. c. hocainensis n. subsp. 3 H. c. simplex n. subsp. 5 H. c. opalina n. subsp. 6 A New Subspecies of Hyla perpusilla H. p. v-signata n. subsp. 7 Variation in Hyla polytaenia Cope H. p. cipoensis n. subsp. 9 H. p. goiana n. subsp. 11 Literature Cited 13 Acknowledgments 13 Illustrations FIGURE PAGE 1. Hyla catharinae bocainensis, female 4 2. Hyla catharinae opalina, female 6 3. Hyla perpusilla v-signata 8 4. Hyla polytaenia from Upper Itatiaia 9 5. (a) Hyla polytaenia cipoensis, female allotype (b) Hijla polytaenia cipoensis, lightphase, male 10 6. Hyla polytaenia goiana, male 12 GeographicVariationinBrazilian SpeciesofHyla * BERTHA LUTZ ABSTRACT Until now, practically all Brazilian species of Hyla have been described or discussed exclusively under the nominative form, without regard for geographical variation. The subspecies is almost unknown in the systematicsof Brazilian frogs. This paper is concerned with local variation in two difficult groups, that of Hyla catharinae and of the species Hyla polytaenia. Variationbecomes apparent on extension of the known distribution, going from al­titudinal forms to subspecies in isolated or peripheral parts of the range.Altitudinal variation was first pointed out by the author (1952) for Hylacircumdata Cope when specimens of this montane coastal species were collected above 1,800 meters on the Itatiaia Mountains. They were robust,and lacking the specific pattern of dark violet circles on a lighter blue-violet background on thigh parts which were concealed in repose. These colors were substituted by a uniform purplish-black, or slate colored surface. A newsubspecies was not set up because some robust specimens from slightly lowerelevationsatItatiaia,andelsewhere, showedvestigesofthecircles. New Subspecies ofHyla catharinae The typical form of Hyla catharinae was described by Boulenger (1888)from the state of Santa Catarina in southern Brazil. This species is individu­ally variable, but the nominative form is recognizable in the mountains of that state and of Rio Grande do Sul. The subspecies described below come from more northern serras (mountains) belonging to the same coastalorographical system. The specimens conform to much more than 75 percentofdifferentiationfrom othergeographicalraces. Hyla catharinae hocainensis n. subsp.-This form was first seen by A. Lutz in 1925 at his brother’s Fazenda do Bonito in the Serra da Bocaina, at approx­imately 1,100 meters altitude. He selected the name long ago, but did notformallydescribetheform.Allbuttwoofthe specimensobtainedthereagreewith this form. One of these is aberrant, the other, a small specimen takenabove timberline, agrees with the form I call simplexwhich is dominant , abovetimberlineontheItatiaiaBange, acrosstheParaibaValley. H. c. hocainensis (Fig. 1) is best characterized by a light, scalloped and * Museu Nacional, Rio de Janeiro, Brazil. Fig. 1. Hyla catharinae bocainensis female; natural size, 44 mm. festooned network surrounding and enclosing, or limiting, the dark dorsal spots and dorsolateral areas. The females are large and robust, 40 mm. or more from snout to vent. Type and Type-Locality.—There are 11 specimens, including three largefemalesintheA.Lntzcollection and oneattheDepartmentofZoologyofthe state of Sao Panlo. The following description is based on the latter (No. 56) as it is more recent and in perfect condition. It was collected at the Fazenda do Bonito by Mrs. Maria Aparecida Volcano of the DZSP, in January 1960. The female paratypes are numbers 2517, 2518, and 2535 in the A. Lntz col­lection at the Musen Nacional, Rio. Type-locality: Fazendo do Bonito,22° 22' 30" S.; 44° 35' 30" W. Size.—Type 44 mm. snout to vent length; other females from 41.5 to 46.5 mm. Build robust. Males much smaller, less robust, 33-36.5 mm. snout to vent length. Leg length average, the tibiotarsal articulation reaching be­tween the eye and the nostril when the hind limb is adpressed, head is oval and as wide as long. Snout truncate between the nostrils, rounded in front, with moderately distinct canthus rostralis and concave loreal region flaring to the mouth. Eye large and very prominent, its horizontal diameter greater than the distance from its anterior corner to the nostril, equal to the distance to tip of the snout. Tympanum moderate, one-third the diameter of the eye. Interorbital space slightly wider than the upper eyelid; internarial space one-fourth less than the interorbital space. Tongue large, round, and 4 emarginate. Vomerine teeth in two small groups between the choanae, con­tiguous in the type. Disks short, wide, the largest slightlyrounded in front, all constricted behind. A rudiment of web between the lateral fingers. Web on foot oblique and narrow between first and second toes, slightly wider be­tween second and third, wide and quite symmetrical between third, fourth and fifth; inserted below the disk on the outer sides of second, third, and on fifth, wide to the second subarticular tubercle on the inner side of the third and on both sides of the fourth toe. A callous under the first finger and an­other, wider one under the third and fourth; an elongate inner and a minute outer metatarsal tubercle. Palmer and plantar tubercles as in typical 11. catherinae. Skin slightly warty on the head, granular on the ventral aspect.Pattern.—ThelightmantleofthemantledphaseofH. c.catharinaefrom Santa Catarina is replaced by a conspicuous light network with scallopedand festooned edges surrounding and invading the dark dorsal areas. The interocular spot retains its form best; it has a slight dent in front, between the eyes, and is prolonged backward, narrowing gradually to a blunt pointwhich is bipartite in some specimens. The sacral spot is broken up into a number of fragments. In the tvpe, the largest fragment is narrow, roughlychevron-shaped and verv much prolonged in front; it is followed by two slightly eccentric ocelli and a quite irregular and eccentric part. The sacral spotalso maybebroken upintoinsularspots.In someoftheotherspecimensonly vestiges of the chevron remain. The network is prolonged onto the topof the head and nostrils as a light cap, and may contain one or two large dark spots. The sides of the head below the canthi are dark, with four or even fivelightobliquebarsfrombelow the eye tothe edgeoftheupper jaw.Light areas on the limbs similar to the network, enclosing broad dark spots, which are sometimes spoolshaped. Ventralaspect similarto thatofthe typicalform. Pattern on concealed parts of the limbs not marked nor visible ventrally.Edges of limbs dark. Hyla catharinae simplex n. subsp. This is the form of Hyla catharinae from the Serra da Mantiqueira, which is parallel to and slightly inland from the Serra do Mar. This subspecies differs from the nominative form and other subspecies by the slightly smaller size, narrow, elongate build, simplifieddorsalpattern,andthe absenceofvividflashcolors. Type-Locality.—Brejo da Lapa, Alto Itatiaia, 2,200 meters above the sea(app. 22° 31' 15" S.; 44° 39' 00" W.). Male type MNR No. 4035 collected by B.Lutz,Nov. 1954;femaleallotypeNo.4036collectedbyA.L.Carvalhoand H. Berla, Dec. 1962. Size.—Variable, generally fairly small. Female 35-40 mm., male 2.5-28,rarely more. Build narrow, elongate. Morphological characters of the species:dorsal pattern very simple. A triangular interocular spot prolonged pos­teriorly, and a pair of rather straight longitudial lateral stripes on the back. Sacral spot indistinct. Mid-dorsal region plain, lighter than the spots. Flash colors dull, grayish-violet. Ventral aspect stippled, darker than usual. 5 In some specimens the interocular spot is pointed posteriorly, in others blunt, sometimes bifid, in others assymetrical; lateral stripes sometimes slightly deflected inward. The mid-dorsal region may be slightly ornate. At Brejo do Lapa a large population has been under observation for many years. They call, in the weak voice of the whole species, on the ground or on the herbs at the edge of an artificial pool in which they breed. Specimens from lower elevations at the Itatiaia and from Cidade Azul,Campos do Jordao, also on the Mantiqueira Range, are paler and slightlymore oranate. A number of juveniles of other more ornamented races are similartoH. c.simplex.This suggeststhatthesimplexpatternmayrepresentthe specific pattern in its simplest form, devoid of glands and a sacral spot,whereasintheotherraces thepatternismodifiedandbecomesmorecomplexduring growth. Hyla catharinae opolina n. subsp.—This is a very handsome race of H. catharinaefromthenorthernblockofthe SerradoMar. Type Specimens and Type-Locality. —Serra dos Orgaos at Teresopolis,stateofRio de Janeiro(22° 26'12"S.; 42° 52'42"W.),800-900meterseleva­tion. MNR numbers 4037 (female, March 1945) and 4038 (male, March 1948)from thecollection ofBerthaLutz. Hylacatharinaeopalina.—This differsfrom thenominativeformbyhavinga more northern distribution, a generally very light and glandular dorsal aspect, and an opaline flash color on the flank and upper concealed aspects of the thigh. Fig. 2. Hyla catharinae opalina female; natural size, 44 mm. Morphological characters are those of the group. Build robust, size rather large: female (Fig. 2) 40-45 mm., male 23-34 mm. snout to vent. Dorsal parts very glandular, generally light in color, pattern indistinct. Interocular spot shallow, W-shaped; sacral spot absent; dorsolateral stripes marked onlyanteriorly. Flash color opaline, light greenish blue on bluish green. Bold black blotches at the flank and groin very conspicuous from beneath. In some specimens the interocular spot is cresent-shaped, concave anter­iorly but also shallow. Some specimens are more pigmented, and have a more pronounced dorsal pattern. This subspecies is often seen in or at the entrance of the Organ Mountains National Park. It calls above the mountain streams and breeds in shallows. The voice is feeble, as in the whole group. The tadpole has a white spot between the nostrils. Similar specimens have been collected at Nova Friburgo, and also in the northern block (22° 16' 42" S.; 42° 21' 34" W., 900 meters elevation). A New Subspecies of Hyla perpusilla H. perpusilla v-signata n. subsp.—SometimeafterdiscoveringH. perpusillain the ground-dwelling bromeliads of the coastal plain, we began to find specimens in the bromeliads, both ground-growing and epiphytic, of the ad­jacent coastal sierras. These specimens, while conforming to the general col­oring, morphology, voice, habitat and ecology of the nominative form,differ from it by attaining a slightly larger size and more intense pigmenta­tion, especially on the gula. Nor is the proportion of the different patternsquite the same. This form is considered to be an altitudinal race which we call Hyla perpusilla v-signata (Fig. 3). DiagnosticCharacters.—Anaccumulationofpigment onthe gulaformsa thick, V-shaped figure, either fragmented into the two arms, or slightly out of shape, and sometimes broken up into dark flecks. Size slightly larger,flash color deeper and more abundant; montane distribution. Types and Type-Locality.-More than 100 specimens were examined. The types were chosen from the Organ Mountains population, whose sample is more numerous than the others and also the most characteristic. Type:MNR 3607; four paratypes, MNR numbers 3608, 3609 from the OrganMountains National Park at Teresopolis were collected by the author in November and December 1956. Size.—Slightly larger than that of the nominative form; female 27 mm.;others 25, 26, and 24 mm., snout to vent. Build rather elongate and robust for the size. Hind limb variable, the tibiotarsal articulation reaching between eve and nostril in three, only to the front of the eve in others. Feet slightly more webbed than H. p. perpusilla, otherwise morphological characters in agreement with that subspecies. Dorsal aspect rather moss-like or lichenous. Pattern as in the H. catharinae group of species, more complete than usual in the nominative form. Interocular spot very shallow with lighthalo in front. Fig. 3. Hyla perpusilla v-signata; natural size, 27 mm. Sacral spot roughly crescent-shaped, somewhat longer and slightly eccentric in the largest syntype. Dorsolateral spots very distinct, like inverted (lefttoright), pointedparentheses,theapexturnedupward ontotheback.Barson limbs distinct, with wide light intervals, and extensive flash color; light area on the flanks between the dorsolateral dark areas and the sacral spots; con­cealed pattern perceptible from beneath. Warts rather numerous, especiallyontheheadand edgesofthebody. Dorsal colors olive browns and grays. Flash color orange to cadmium orange. Pupil slightly tawny, iris as in nominative form. Underside: gulagrayish, belly with whitish glandules, the viscera showing through dark blue. Ventralaspect ofthighspaler,slightlyviolaceous. Diagnostic V. —The montane subspecies was first separated from the low­land form because of the pigmented gula. At its best, the V of color on thethroat is wide, complete, composed of two thick, not very oblique branches covering the upper part of the throat. In some specimens the branches arethicker at the top, in others they are almost horizontal; they may also be broken into separate branches. Occasionally the V-shaped part is prolongeddownward by a double or single streak of pigment approximating a Y. Inothers the V is substituted by a single squarish blotch. Occasionally there isanother pair of blotches below the first pair, or again the pigment is dissolved into a number of flecks and dots. Whereas the V is very striking in living specimensit is,unfortunately,liable tofadeafterdeath.Byvarying the inci­dence of light it can, however, frequently be sufficiently illuminated to stand out. Voice, ecology, and breeding site is the same as in H. p. perpusilla. Bromeliads constitute a habitat with an independent microclimate that emancipates their inhabitants from the general environment. The coolerclimate of the serras seems to favor growth and intense flash colors. Fig. 4. Hyla polytaenia from Upper Italiaia. Variation in Hyla polytaenia Cope Hyla polytaenia (Fig. 4) is a montane coastal Brazilian species of mod­erate size originally described from the southeastern mountains. Its typicalpattern consists of a series of longitudinal stripes down the back. Cochran (1955) points out that in large specimens (over 40 mm. long) this patternis reduced to three stripes, one central and two lateral. She calls this the adult patterns and considers simplification as a function of size and perhaps of sex, since all the large specimens are females. This is true to a certain extent for the southern part of the Serra do Mar which is the terra typica of the nominative form. There is, however, a geographical factor involved. On the parallel, slightly inland, at Serra da Mantiqueira (Campos do Jordao and especially Upper Itatiaia) this trend becomes evident in males also. The Upper Itatiaia populations living on the Plateau above 2,000 meters, where there are a few pools derived from mountain brooks, are unusually large.Males attain the length of the females from elsewhere and females mayreach 50 mm. from snout to vent instead of the usual 39-40 (or quite ex­ceptional 43 mm.). Simplification of pattern is not uniform. Some specimenshave only the three main stripes but many show two longitudinal series of dotsonthehindpart oftheback. Hyla polytaenia goes into Minas Gerais. The samples from Juiz de Fora and from outside Belo Horizonte are small to average; the secondary stripes are present but generally fragmented. At Pecos de Caldas the margin of the three main stripes is brilliant. Two populations, H. p. cipoensis, distant and peripheral from Goias, and H. p. goiana, from the Serra do Cipo in Minas Gerais diverge even further. H. p.cipoensis n.subsp.—Thisisaseriesoffairlysmalltreefrogs(Fig.5),collected in the Serra do Cipo, Minas Gerais. While similar to 11. p. poly­taenia, they are sufficiently different to warrant considering them as sub-specifically distinct. Differential diagnosis.—Morphologically like H. p. polytaenia but differing Fig. 5. (a) Hyla polytaenia cipoensis, dark phase. Female allotype, natural size 37 mm. (b) H. p. cipoensis, light phase, male, natural size 26 mm. greatly in color, dorsal pattern and pattern on the gala. Lateral pattern more similar, size and build moderate. Hind limb variable but rather short. Head mostly as long as wide. Tympanum small. Group of vomerine teeth round and short. Types and Type-Locality.—Three males, type and paratypes, were col­lected in December 1947by the author, one female allotype and seven other paratypes in the same year by Prof. Amilcar Vianna Martins. All were taken near Alto do Palacio, in the Serra do Cipo, municipality of Jaboticatubas,between Lagoa Santa and Ferros, Minas Gerais, at approximately 1,300meters elevation. Types in the Museu Nacional, paratvpes in the Adolpho Lutz Collection, also at the Museu: Nos. 4039, male tvpe; 4040, female allo­type; 4041, 4042, 4043-49, maleparatypes. Description. Size small. Males 29 (type), 27, and 26 mm. from snout to vent. Hind limb 48, 47, 46 mm., the adpressed tibiotarsal articulation reach­ing in front of the eye. Head as long as wide, except for the largest one, in which it is barely wider. Snout short, loreal region concave, high; canthus rostralis distinct; interorbital region wider than the upper eyelid. Eye longerthan the distance from its anteriorcorner to the nostril. Tympanum less than half the diameter of the eye. Tongue variable, unnotched, slightly emargin­ate in the third, very slightly free. Vomerine teeth in small, rounded, sep­arate groups between the choanae. A short web on the lateral fingers. Toes abouthalfwebbed;palmsand solesfulloftubercles;pollexrudimentpresent, knob-like; an oval, inner metatarsal tubercle. Skin of abdomen and mid-ventral part of the thigh very granular. Female (allotype) much larger than male, 37 mm., the hind limb (64 and 63 mm.) relatively shorter, the tibiotarsal articulation reaching only to the eye. Build more robust. Pollex rudiment not noticeable. Paratypes 36 mm. (female); 33 and 28 mm. (males); 23 and 22 mm. juven­iles, from snout to vent. Pattern.—Dorsalpatternreduced tothe threemaindarkstripes, oneverte­bral, two lateral, the first ending on the head before the tip of the snout, the others ontheinnerposterioredgeofthe uppereyelid.Novestigeoftheother stripes from which the name of 11. p. polytaenia is derived. Permanently vis­ible dorsal stripe of the limbs similar in color to the dorsal surface, light edged.Main lateral stripe passing under the canthus rostralis and ending before the tipofthesnout(as in 11.p.polytaenia), enclosingthe eyeand thetympanum,andendingbefore the groin.Upperlight stripeno morethanthe dorsolateral light edge; lower light area wide, about as long as the dark stripe, brilliant white on the sides and on the maxilla, except in front, where it takes the color of the dorsal background. A brilliant white curved stripe on each side of the gula. Color.—ln life completely different from the browns and buffs of H. p.polytaenia. Two phases, dark more common; background dull green, lighter on the limbs and the outer digits visible in repose. Stripes purplish-brown.Light phase: background a clear luminous yellow, stripes grayish lavender. Well preserved specimens show vestiges of pigment, especially the greenone which acquires a bluish tint. Length of hind limb variable (as in other forms and in most Brazilian hylas), but on the whole short. Head wider than long in one paratype. Shapeof the tongue also variable. Additional short, light and dark lateral areas in the large female paratype. Sexual dimorphism not marked in our series. Vocal sac of male not large;claw-shaped point of the pollex rudiment not palpable; only the medium syntype has thickened forearms. The short stay of only two days and nights in the type-locality, mostlyunder torrential rain, did not yield much information on points such as be­havior and voice. Ecology. The Serra do Cipo constitutes a most interesting environment. The vegetation belongs to the type called “Cerrado” in Brazil, i.e., openscrub with tortuous bushes and many open spaces. A brook runs between banks below Alto do Palacio, where the water table overflows much of the higher ground. Low herbaceous vegetation grows in marshy places which are spongy underfoot. Rocks are plentiful and are used for erecting stone walls. Distribution not known. The typical form, 11. p. polytaenia, occurs 134 kilometers from the Serra do Cipo, and some 500 meters lower. Hylapolytaeniagoianan.subsp.(Fig.6). Thisistheoutmostpopulation Fig. 6. Hyla polytaenia goiana, male paratype; natural size 30 mm. of H. polytaenia known. While immediately reminiscent of the species, it diverges not only by the distance of its locality from the remainder of theknown range, but also in color by having dull, powdery spots and pigments in contrast to the usually brilliant colors of H. polytaenia, and bv having longerlegs. Unlike most forms described by the writer, I had not seen the specimensfrom the Highlands of Goias alive. They were given subspecific rank because of the characters mentioned above, albeit with some doubts about the valid­ity of this race. Type-Locality.—Jatobasinho, Sao Joao da Alianca, highland of Goias,collected by A. L. Carvalho, May 1956. Type, MNR No. 3235; male para­types, numbers 3233, 3234,3236, and 3237. A series of nine males in the Museu Nacional was collected by Mr. Antenor Leitao de Carvalho in the county of Sao Joao da Alianca, above 1,000meters of altitude; these specimens diverge from the typical form of H.polytaenia. Four, obtained in a camp called Jatoba, are very dark, whereas the other five, from another camp, Jatobasinho, are pale. Sizeand Appearance.-Similartothoseofthetypicalform-30to31mm.from snout to vent—one 29, another 33, both from Jatoba Camp. Build similar to nominative form; head apparently narrower, length and width subequal. Hind limb relatively long, the adpressed tibiotarsal articulation reachingalmost to the nostril, or even beyond the snout in all but one shorter-leggedspecimen. Vomerine teeth forming a straight row; vocal sac puckered in themiddle, in some specimens, with two lateral oblique folds. Forearm thick inthe nuptial condition; rudiment of pollex present in all, but no sharp pointperceptible. Dorsal pattern simplified; one entirely devoid of pattern, the others with the three main stripes only; uniform in width, not fragmented or forming drops, the two lateral near the central stripe. Vestiges of secondary stripes fragmented into drops present only in the largest specimen. Dark bordersoflimbsnoticeableonforearmandtarsus asinthenominativeform;a light, upper border over the anal and the tarsal folds. Vestiges of rounded spots made up of minute dots, occupying the permanently visible dorsal strip of the femur and corresponding area of the tibia, in large specimensonly. In all nine specimens spots and pigment have a powdery aspect. LITERATURE CITED Boulenger, G. A., 1888. A list of Batrachians from the Province of Santa Catha­ rina, Brazil. Ann.and Mag. Nat. Hist. 6(1): 415-417. Cochran, D. M., 1955. Frogs from Southeastern Brazil. U. S. National Museum Bull 206: 89-92. Lutz, Bertha, 1952. New Frogs from Itatiaia Mountain, Brazil. Copeia (1):27-28. ACKNOWLEDGMENTS Theauthorwishestothankherbrother,ProfessorG.A.Lutz, forprovidingphotographs of specimens from his original color slides. ThePearce-Sellards Series The Pearce-Sellards Series consists of occasional papers published by the Texas Memorial Museum, 24th & Trinity, Austin, Texas. Other publications include the Bulletin series, Notes, and mimeographed information circulars. A complete list will be sent upon request. No. 1. Fossil Bears from Texas, by Bjorn Kurten, 1963 35 No. 2. Post-Pleistocene Raccoons from Central Texas and their Zoogeo­ graphic Significance, by Thomas Wright & E. L. Lundelius, Jr., 1963 ' .40 . No. 3. A New Fossil Tortoise From the Texas Miocene, by Walter Auffenberg, 1964 25 No. 4. The Osteology and Relationships of the Pliocene Ground Squir­ rel, Citellus dotti Hibbard, from the Ogallala Formation of Bea­ ver County, Okla., by Margaret Skeels Stevens, 1966 75 No. 5. The Status of Bootherium brazosis, by Clayton E. Ray, 1966 . .25 . No. 6. Geologic Reconnaissance of the Ft. Davis National Historic Site,Texas, by Gordon Everett, 1967 35 No. 7. Mammalian Remains from Rattlesnake Cave, Kinney County,Texas, by Holmes Semken, 1967 35 No. 8. Development of Terminal Buds in Pinyon Pine and Douglas-firTrees, by Charles L. Douglas & James A. Erdman, 1967 35 No. 9. Toxotherium (Mammalia: Rhinocerotoidea) from Western JeffDavis County, Texas, by John M. Harris, 1967 35 No. 10. New Brazilian Forms of Hyla, by Bertha Lutz, 1968 35 14