1 BULLETIN OF THE FLORIDA STATE MUSEUM BIOLOGICAL SCIENCES Volume 5 Number 5 PATTERN VARIATION IN THE FROG ELEUTHERODACTYLUS NUBICOLA DUNN Coleman J. Goin UNIVERSITY OF FLORIDA Gainesville . 1 1960 . b \ The numbers of THE BULLETIN OF THE FLORIDA STATE MUSEUM, BIOLOGICAL SCIENCES, are published at irregular intervals. Volumes contain about 300 pages and are not necessarily completed in any one calendar year. OLIVER L. AUSTIN, JR., Editor WILLIAM J. RIEMER, Managing Editor All communications concerning purchase or exchange of the publication should be addressed to the Curator of Biological Sciences, Florida State Museum, Seagle Building, Gainesville, Florida. Manuscripts should be sent to the Editor of the BULLETIN, Flint Hall, University of Florida, Gainesville, Florida. Published 17 August 1960 Price for this issue $0.35 1 PATTERN VARIATION IN THE FROG ELEUTHERODACTYLUS NUBICOLA DUNN COLEMAN J.COIN 1 SYNopsm: An analysis'is made of the cdlor pattern variation in the Jamaican frog, Eleutherodactylus nubicola Dunn. A basic pattern, Mottled, may be over- laid by no less tlian six modifying patterns as follows: Dorsolateral stribes, Mid- dorsal stripe, Broad middorsal stripe, Picket, Interocular bar, and Pelvic spots. It is possible to have more than one of these modifying patterns present in a single individual; one specimen had, at one and the same time, Dorsolateral stripes, Middorsal stripe, and Picket. Evidence is presented that these mbdifying patterns are inherited in a mendelian manner and Dorsolateral stripes and Mid- dorsal stripe seem to be dominant. Evidence is not yet available that will permit us to determine if the other pattern modifiers are dominant or recessive. It is pointed out that certain Rhacophoridae from Ceylon and Microhylidae in the Papuan region not only have developed terrestrial breeding habits like those of Eleutherodactylus, but that they have several of the pattern modifications described for E. nubicola, The significance of the parallelism is at present un- known. With the possible exception of the African genus Hyperollus, the neotropical frog genus Eleutherodactylus is the most "difficult" tax. onomically of any of the large anuran genera. The reason seems to be threefold-the phenomenal amount of speciation, the morphologi- cal similarity of the species, and the intraspecific diversity of color pattern. In earlier papers (Coin, 1947, 1950) I presented evidence that much of this yariability in color pattern is due to a series of genes that are inherited in a mendelian manner. So little is known of amphibian genetics that,the determination of the genetic basis of color pattern inheritance in these species is of interest in itself, but more importantly, an understanding of it may help to throw light on the evolution and relationship of species within this complex genus, and possibly on the even more fundamental problems of evo- lution in natural populations. Eleutherodactylus contains nearly 250 recognizable forms. Many of the species are rather variable in color pattern, with the same basic pattern modifications occurring in species after species. The evi- dence, discussed below, suggests that these pattern variations reflect the segregations of a number of different genes (and their alleles) which are homologous in the different species. i The author is Professor of Biological Sciences, University of Florida, and ' Research Associate, Carnegie Museum. Manuscript submitted 16 February.1960. 244 BULLETIN FLORIDA STATE MUSEUM Vol. 5 My interest in this problem first centered on E. ricordi planirostris, a species with two phenotypes, which was introduced into Florida and has been abundant locally for a number of years. At the sugges- tion of my friend, W. Gardner Lynn, I later turned my attention to E. nubicola of Jamaica, because this species is known to have a large number of phenotypes and because its eggs can be collected easily - along the mountain trails in the Blue Mountains of Jamaica. In. the summer of 1948 I was able to collect 25 clutches of nubicola.eggs which produced 1002 hatchlings. Analysis of the data (Coin, 1950) indi- cated that, as in planirostris, Variability in nubicola is produced by genes that modify the color pattern and that are inherited in a simple mendelian manner. Additional data gathered 6n two related species, alticola and pantoni, indicate.a similar mode of pattern inheritance in these forms. I returned to Jamaica in 1957 to gather additional ma- terial on nubicola; 48 clutches, comprising 1341 hatchlings, were col- lected on this trip. These, together with the previous material, make a total of 2,848 young nubicola from 68 clutches of eggs now available for analysis. ACKNOWLEDGMENTS In the course of this work I have become indebted to a great many institutions and people. My field work in 1948 was financed by the American Academy of Arts and Sciences, that in 1957 by the American Philosophical Society. I obtained data on the variability of other sp, ecies in Jamaica in 1952 by field work done with the aid of a grant from George R. Cooley. As is true of anyone doing natural history work in Jamaica, I am deeply indebted to C. Bernard Lewis and his staff at the Science Museum, Institute of Jamaica. The Forest De- partment of Jamaiea, through its Conservators, E. W. March, E. M. Brown, and J. A. N. Burra, made the Forest Hut at Clydesdale avail- able to me as a base of operations. My neighbor in the Blue Moun- tains, Bonner E. C. Shekell, formerly of Chester Vale, St. Andrew, rendered me a number of courtesies that made my stay in the moun- tains more pleasant and profitable. My Beld companions, Dennis E. Miller, Byrum W. Cooper, and Robert H. Barth in 1948, Albert M. Laessle in 1952, and my wife and two children in 1957, helped appre- ciably in the collecting of the eggs. My research assistant, Ronald Baxter, aided in tabulating the results. My colleague, Kenneth W. = Cooper, contributed materially by his help in analyzing the data and Henry W. Wallbrunn assisted in some of the statistical calculations. Curt Stern most kindly took time from his busy schedule to go over the manuscript and has made several helpful suggestions. As usual, I 1960 COIN: PATTERN IN ELEUTHERODACTYLUS 245 am deeply obligated to my wife, Olive Bown Coin, for her skillful assistance in typing and editing the manuscript. This work was com- pleted while I was under tenure of a Grant (G-5628) from the National Science Foundation. To all of the above I am deeply grateful. Eleutherodact!/lus nubicola Dunn Eleutherodactylus nubicola is a montane species found in the Blue Mountains of Jamaica between the altitudes of 4000 and 6000 feet. Lynn (1942) described in detail its life history and embryonic development. The slopes of the Blue Mountains are so steep it is prac- tical to collect only beside the foot trails. The female· lays from 26 to 75 eggs which are fertilized by a single male at the time of deposi- tion. The eggs are not scattered, but are gathered together in a single, compact bunch under a rock. Thus it is certain that each clutch is the product of a single set of pafents. The female remains with the clutch during intraovular development. It is thus possible for one to determine the phenotype of the female parent of many clutches of eggs collected in the field, although at times the female escapes be- fore her pattern can be noted. A study of live individuals in the field and preserved specimens in the laboratory indicates that there is a basic pattern, Mottled, with at least six distinctive modifications as follows: Dorsolateral stripes, Middorsal stripe, Broad middorsal stripe, Picket, IntePocular bar, and Pelvic spots. Data compatible with a simple mendelian interpreta-, tion are now available for all of these patterns except Broad mid- dorsal stripe and Pelvic spots. In the discussion that follows, the 16cus for genes affecting Dor- solateral stripe is indicated by S, that for Picket by P, that for Mid- dorsal by' M. A dominant gene at the S locus is designated by S, its recessive allele by s, and correspondingly P and p, and M and m, for the other loci. The following data are now available for pattern inheritance j analysis: (I) For the phenotypic ratios in 'the wild population, 800 specimens collected along the mountain trails in the same area where the eggs Were taken showed the following patterns or pattern combinations (females with eggs ar6 included): 'Mi*tled-222 Dorsolateral stripes-32 Picket-21 Middorsal stripe-21 Dorsolateral stripes plus Middorsal stripe-2 Picket plus Midd6rsal stripe-2 246 BULLETIN FLORIDA STATE MUSEUM Vol. 5 (2) From mothers of'known pattirn 1544 hatchlings from 45 clutches of eggs: Mother-Mottled- (84 clutches) Collected in 1948 Collected in 1957 26 Mottled, 20 Dorsolateral 38 Mottled, 4 Dorsolateral stripes stripes 5 Mottled, 8 Dorsolateral stripes 25 Mottled, 28 Middorsal stripe 19 Mottled, 12 Dorsolateral stripes 25 Mottled, 26 Middorsal stripe 8 Mottled, 4 Dorsolateral ,stripes 26 Mottled 43 Mottled, 19 Dorsolateral stripes 48 Mottled 26 Mottled, 14 Middorsal stripe 58 Mottled 16 Mottled, 21 Middorsal stripe 47 Mottled 53 Mottled, 8 Picket 42 Mottled 18 Mottled, 11 Picket 53 Mottled 22 Mottled 21 Mottled 21 Mottled 51 Mottled 13 Mottled 83 Mottled 14 Mottled 41 Mottled 29 Mottled 26 Mottled 86 Mottled 29 Mottled 89 Mottled 41 Mottled 11 Mottled 87 Mottled Mother-Dorsolateral stripes (5 clutches) 1 Collected in 1948 Collected in 1957 31 Dorsolatefal stripes 6 Mottled, 16 Dorsolateral stripes 9 Mottled, 8 Dorsolateral < 21 Mottled, 80 Dorsolateral stripes stripes 5 Mottled Mother-Middorsal stripe (4 clutches) Collected 'in 1948 Collected in 1957 29 Mottled, 16 Middorsal stripe 5 Mottled, 5 Middorsal stripe, 84 Mottled, 89 Middorsal stripe 10 Picket 16 Mottled, 19 Middorsal stripe Mother-Picket (2 clutches) Collected dn 1948 Collected in 1957 7 Mottled, 4 Dorsolateral stripes, 18 Mottled, 11 Picket 3 Picket, 4 Middorsal stripe, 8 Dorsolateral stripes plus Picket, 5 Picket plus Mid- dorsal stripe, 5 Dorsolateral stripe5 plus Picket plus Mid- dorsal stripe 1960 COIN: PATTERN IN ELEUTHERODACTYLUS 247 (3) From parents of unknown phenotype 799 hatchlings from 23 clutches of eggs: Collect6d in 1948 Collected in 1957 5 Mottled, 4 Dorsolateral 4 Mottled, 4 Dorsolateral stripes stripes 11 Mottled, 15 D6rsolateral stripes 23 Mottled, 22 Middorsal stripe 15 Mottled, 29 Dorsolateral stripes 48 Mottled 81 Mottled, 3 Dorsolateral stripes 60 Mottled 20 Mottled, 14 Dorsolateral stripes, 84 Mottled 10 Interocular bar, 11 Dorso- 39 Mottled lateral stripes plus Interocular bar 18 Mottled, 16 Midorsal,stripe 26 Mottled, 17 .Picket 25 Mottled, 25 Picket· 21 Mottled 29 Mottled 85 Mottled 41 Mottled 27 Mottled 20 Mottled 29 Mottled 26 Mottled 47 Mottled Mottled The Mottled pattern consists of a brown background overlaid by a dark brown 61, black mottling. The most consistent element is a dark W-Shaped mark in the suprascapular region, extending fr6m above the arm on one side, to above the arm on the opposite side. The apex of the middle portion of the W is directed anteriorly and located on the midline of the dorsum. The area immediately caudad to the two posteriorly directed apices of the W is generally somewhat lighter in color than the surrounding dorsum. A faint, wavy line of dark pig- ment is discernible caudad of this lighter area. A dark interocular bar is present. Intensity of pigmentation shows a great deal of indi- vidual variation and generally tends to increase with age. In hatch- lings the pattern is immediately apparent, but in large, mature indi- viduals it is often obscured by a general darkening of the dor- sum. The main features of this basic pattern may be ~ discerned in the majority of the species of this genus. In E. nubicola it seems to be the basic wild pattern upon which the pattern modifications discussed below are superimposed. 248 BULLETIN FLORIDA STATE MUSEUM Vol. 5 ·thilli!. Figure 1. Diagrammatic sketch of Figure 2. Diagrammatic sketch of Mottled pattern. Dorsolateral stripes pattern. Dorsolateral stripes In this pattern two broad cream-colored bands, one on each side, originate on the posterior margins of the upper eyelids, 'pass poster- iorly above the tympani directly backward through the arms of the W, and terminate above the insertions of the hind limbs. These cream-colored bands obliterate the arms of the W leaving only the A-shaped portion between the two stripes in the Suprascapular region. This is one of the most widely distributed pattern modiRcations in the genus. I have demonstrated fairly conclusively (Coin, 1947) that in E. ricordi planirostris this pattern is determined by a dominant gene that is inherited in a simple mendelian fashion. No clutches of eggs of E. nubicoZa were taken for which the color patterns of both parents were known, but indirect evidence for domi- nance or recessivity may be sought by use of the Hardy-Weinberg equilibrium formula q2 SS : 2*1-q) Ss : (1-q)2 ss to determine the genotypic ratio in the wild population. If we assume Dersolateral stripes to be dominant, then in a pop- ulation in which 34/300 showed Dorsolateral stripes, 1960 GOIN: PATTERN IN ELEUTHERODACTYLUS 249 2q(1-q) + q2 - 84/300 - 0.1188 therefore (1-q)2 = 0.8867 and 1-q 0.9416 hence q = 0.0584 in which case SS = 0.0084 SS = 0.1100 ss == 0.8867 If the assumptions necessary for the application of the Hardy- Weinburg law are made, we can calculate the phenotypic constitu- tion of the clutches of the offspring. Thus of 68 clutches of eggs only approximately 0.46 clutches would be expected to be made up of offspring with Dorsolateral stripes only, 14.08 clutches should con- tain offspring both with and without Dorsolateral stripes, and 58.46 clutehes should be made up of nonstriped individuals only. Going through the same steps, but assuming Dorsolateral stripes to be recessive rather than dominant, it can be calculated that of 68 clutches approximately 0.87 should contain only offspring with Dor- solateril stripes, 20.44 clutches. should contain both types of offspring, and 46.69 clutches should contain no offspring having Dorsolateral stripes. These two sets of estimates are compared with the actual results in table 1. TABLE 1 Theoretical clutches Theoretical clutches Dorsolateral stripes Dorsolateral stripes Actual dominant recessive clutches Ddsolateral· stripes only 0.46 0.87 1 Mixed 14.08 20.44 16 Without Dorsolateral stripes only 58.46 46.69 51 Total 68.00 68.00 68 xi O,530 r 2.140 P <0.5>0.8 P <0.2>0.1 * Lines 1 and 2 are summed for calculation of x2. While these results give no certain statistical basis for choosing - between the two aSsumptions, I believe that it logieally may be as- sumed that the gene for Dorsolateral stripes is dominant in E. nubicola because it has been shown to be dominant in both E. alticola (Goin, 250 BULLETIN FLORIDA STATE MUSEUM Vol. 5 1950) and E. r. planirostris (Goin, 1947). As Sturtevant (1948: 280) has concluded, "there can be no doubt that, in general, related species have essentially the same complements of genes. The pattern of Dorsolateral stripes has been recorded in at least the following species of Eleutherodactylus and is probably present in many others as well. albipes interinedius polytychus alticola junon[ portoricensis armstrongi Zentus rhodopis atkinsi ockendeni richmondi audanti orcutti ricordi bogotensis orientalis ruhfemoratis cimeatus pantoni turquinensis gossei pictissimus wienlandi gundlachi planirostris Picket This pattern consists of a light .11!iliti/•· area on the dorsum, bounded an- teriorly by the posterior margin of the middle section of the W mark. The lateral margins of this light area apparently coincide with the median margins of the cream- ~ colored bands in individuals with ; Dorsolateral stripes. When both Dorsolateral stripes and Picket are present in the same individual, the entire area between the stripes pos- terior to the A-mark is occupied by the light picket. On the basis· that 28 out of the sample of 800 individuals of the wild population had the Picket pat- tern and 277 lacked it, an analysis similar to that made for the inheri- Figure 8. Diagrammatic sketch of Picket pattern. tance of Dorsolateral stripes may be made for Picket (see table 2). Here the evidence is consistent with the assumption that the gene for Picket pattern is dominant. I ha~e found this pattern in only three other species: gossei, pan- toni, and polvtuchus. 1960 COIN: PATTERN IN ELEUTHERODACTYLUS, 251 TABLE 2 Theoretical clutches Theoretical clutches Actual Picket dominant Picket recessiye clutches Picket only 0.22 0.40 0 Mixed 9.81 15.07 7 Without Picket only 57.97 52.52 61 Tdtal 68.00 67.99 68 1.074 6.008 P 0.8 P <0.02>0.01 Middorsal stripe In adults showing this pattern, the dorsum is marked by a narrow, cream-colored. median line which extends from the snout to the vent and there branches to continue along the posterior margin of each leg to.the· sole of the foot, where it terminates at the juncture of the fourth and fifth toes. Ventrally a narrow, cream-colored, median line extends from the tip of the chin to the · crotch. This is the only one of the patterns that is manifest on the ven- tral side of the body. The midven- „11'i 111:„' tral portion of this line tends to be- come obliterated with increasing 95 4" size (hence age) but even in the largest specimens a median line on the chin and throat is still evi- dent. In hatchlings the middorsal portion of the stripe extends only as far forward as the anteriorly- directed apex of the W but in ma- ture individuals it continues on to the snout. - Using the same method of ~a~) analysis, on the basis that 25 in a sample of 800 individuals of the wild population had the Middorsal h, /5 stripe pattern and 275 lacked it, similar calciilations were made. Figure 4. Diagrammatic sketch of' Middorsal stripe pattern.The results are shown in table 8. Some additional evidence may be obtained from the clutches of females having Middorsal stripe. Of five such clutches, each con- tained offspring showing Middorsal stripe. This is to be expected if 252 BULLETIN FLORIDA STATE MUSEUM Vol. 5 the gene for Middorsal stripe is dominant; if Middorsal stripe is re- cessive, each of these females must have mated with a male carrying the recessive gene. , Since only 25 out of 800 adults in the population showed the pattern, then 25/300 - 0.0888 = (1-42 and (1-q) = 0.2886 and q = 0.7114. TABLE 8 Theoretical clutches Theoretical clutches Middorsal striDe Middorsal stripe Actual dominant recessive clutches Middorsal stripe only 0.24 0.47 0 Mixed 10.62 16.12 11 Without Middorsal stripe only 57.14 51.40 57 Total 68.00 67.99 68 0.0021 x 2.4915 P >0.95 > P >0.1<0.2 The distribution of the genbtypes in the population may accord- ingly be estimated: MM - 0.506 Mm = 0.411 mm = 0.088 1.000 Still assuming the gene for Middorsal stripe to be recessive, each of the five females must have been mm in genotype. Since each female produced both Middorsal striped and non-Middorsal striped progeny, all five females must have mated with heterozygous males' (Mm). But genotypically Mm frogs make up only 0,411 of the popu- lation, and the lik61ihood that, on chance alone, a given set of five females would each be mated with an Mm male is only (0.411)5, or a P of 0.012-a poor likelihood. It seems most probable, then, that Middorsal stripe is also dominant. This Middorsal stripe pattern occurs in at least the following additional species: abbotti bakeri lanciformis atticola cundalli matudai antillensis dimidiatus pantoni auriculatoides gossei portoricensis avocalis junori . 1960 COIN: PATTERN IN ELEUTHERODACTYLUS 258 Independent Assortment For considering Whether the three modifications of the basic Mot- tled phenotype that have been discussed so far are determined by- genes that are multiple alleles, or whether these genes occur in differ- ent loci, evidence is provided by the presence of three modifications of the pattern in a single clutch. The most complex series of offspring from a single dutch is the set of 81 young from a Picket mother which contains 7 of the 8 combinations of patterns possible with these three modifiers. The presence in one clutch of all three modifications plus the basic pattern precludes the possibility that the, three pattern modifiers are all allelomorphic inter se or to the basic pattern gene. Let us assume then that three different pairs of genes are involved, and that the gene modifying the color pattern is dominant. Since there are unmodified Mottled offspring in this clutch, it. would not have been possible for either parent to have been homozy- gous for Dorsolateral stripes, or Picket, or Middorsal stripe. As both borsolateral stripes and Middorsal stripe occur among the offspring in approximately al:1 ratio, and as the mother showed neither, the dominant genes for both of these patterns should have been present in the father. Since Picket is present in the mother and in approxi- mately one-half of the young, the dominant gene for this character should not have been present in the fathef. Thus, on the basis of the argument given above, the mother must have been heterozygous for Picket and homozygous recessive for Dorsolateral stripes and Midi dorsal stripe, and the father must have been heterozygous for both Dorsolateral stripes and Middorsal stripe, and homozygous recessive for Picket. There should be, according to the laws of chance, a nearly equal distribution of the eight possible phenotypes in the offspring. The theoretical'phenotypical composition of· a clutch from this cr6ss is compared below with the actual. ratio of the 31 offspring of the female with Picket pattern. ss Pp mm X Ss pp Mm Offspring Theoretical Actual Mottled 8.875 7 Dorsolateral stripes 8.875 · 4 Picket 3.875 3 Middorsal stripe 8.875 4 Dorsolateral stripes plus Picket 8.875 8 Dorsolateral stripes, plus Middorsal stripe 8.875 0 Dorsolateral stripes plus Picket plus Midaorsal stripe 3.875 - 5- Picket plus Middorsal stripe 8.875 5 . 254 BULLETIN FLORIDA STATE MUSEUM Vol. 5 Moreover, if the parents were ss Pp mm X Ss pp Mm, each of the three modifiers (Dorsolateral stripes, Picket, and Middorsal stripe), when considered individually, should be equally distributed among the offspring. Below the theoretical number is compared with the actual number of each of the three traits considered independently: Theoretical Act~al P' Dorsolateral stripes present , 15,5 121 0.14 Ddrsolateral stripes absent 15.5 19 J Picket present 15.5 161 O.50 Picket absent 15.5 15 J Middorsal stripe present 15.5 141 } 0.36 Middorsal stripe absent 15.5 17 J The actual ratios obtained for Middorsal stripe and Picket pat- terns are obviously very close to the expected ratios, and a ratio as divergent or worse than 12 to 19, as represented in the Dorsolateral stripes series, would be expected in a random sample about 14 per- cent of the time. The data are thus consistent with the hypothesis that there are three pairs of independ- ent genes. Interocular bar A sharply defined, cream-col- ored, slightly convex bar passes across the top of the head from the upper eyelid of one side to the upper eyelid of the other side. Both anteriorly and posteriorly it Figure 5. Diagrammatic sketch of is margined by buff-brown or black. Interocular bar pattern. I had no data on this character 2 Binomial estimate of likelihood, on sampling error, of getting as bad or worse fit to 1: 1 ratio. 1960 COIN : PATTERN IN ELEUTHERODACTYLUS 255 when I first published on inheritance in nubicola (Coin,'1950). When I discussed the evolution of the gossei group (Coin,· 1954) I reported its occurrence in gossei and junori, and later (Coin, 1957) I recorded , it in nubicola. I am now able to report that Interocular bar in nu- bicola, like other pattern modifications, is apparently mendelian in nature. In a clutch of 56 eggs from an unknown parent; 55 of the eggs hatched. The hatchlings showed the following patterns: 20 Mottled, 10 Interocular bar, 14 Dorsolateral stripes, and 11 Dorsolateral stripes plus Interocular bar. Thus it would seem that Interocular bar well might be the phenotypic effect of a simple dominant, or recessive, color-pattern gene-21 with the bar, 84 without-and also that it assorts mdependently of Dorsolateral stripes. Such a thesis is eer- tainly consistent with what we know concerning the inheritance of the other color pattern modifiers. The presence of the character in two other species in the group is also suggestive. The fact that Inter- ocular bar has not yet been found with color patterns other than Dor- solateral stripes is, I suspect, due simply to the rarity of the gene in the population. Br6ad middorsal stripe A sharply defined, broad, cream-colored stripe runs from .'}11 Ii, back of the head to the vent along the median dorsal line. For most %„, 1 '441'' of its length it is about as wide as the greatest diameter of the eye. Its margins are sharply defned by black or very dark brown. This is the only pattern modi- fication so far known that might completely mask another. It is quite possible that the presence of Broad middorsal stripe would pre- vent the phenotypic expression of Middorsal stripe. There are at present no data on the inheritance of this charac- ter. The possibility that it may be an allele to the gene for Middorsal Figure 6. Diagrammatic sketch of stripe should not be overlooked. Broad' middorsal stripe pattern. 256 BULLETIN FLORIDA STATE MUSEUM ' Vol. 5 I know of this character in the following species: alticola intermedius Thodopis gossei nubicola Pelvic spots A rather small but conspicuous, black, more or less rounded patch is present on each side of the back above the groin. This pattern differs in two major respects from those so far dis- cussed. It varies greatly in its conspicuousness and it is expressed by the addition 6f pigment rather than by deletion. (All of the other pat- tern modifications in nubicola are brought about by diminution or•,1/;:ilil, absence of pigment, resulting in the presence of pale, crearn-col- ored areas.) In some mature in- dividuals the pigmented pelvic spots are fairly conspicuous, in others they are rather dim. This variability seemi to be due in part t6 variation in intensity of the spots themselves and in part to variation in the dorsal ground color of the individual. Very dark, conspicuous Pelvic spots are invariably present in E. andrewsi, and I have seen paler, less conspicuous Pelvic spots in pantoni and gossei as Well as nubi- - cola. Judging from published de- Figure 7. Diagrammatic sketch of scriptions, this character seems to Pelvic spots pattern. be fairly widespread in the genus. Corollary Discussion The genus Eleutherodact!/lus is remarkable not> only for the extent of its pattern variability but also for its life history and mode of em- bryonic development. . Mating takes place and the eggs are laid on land. ' The young hatch, not as tadpoles, but as miniature replicas of the adult. It is not simply a case of metamorphosis before hatch- ing, for the embryo fails to develop many of the typical tadpole char- 1960 GOIN : PATTERN IN ELEUTHERODACTYLUS 257 acters (e.g., functional gills, mouth parts, sucking disk) and does de- velop a flattened, respiratory tail. A most remarkable instance of parallelism is shown by certain members of two other unrelated families of frogs (Rhacophoridae and Microhylidae) on the opposite side of the world. They have inde- pendently developed similar terrestrial - breeding habits and similar patterns of embryonic development (tadpole 1structures lacking, res- piratory tail present) and show, moreover, series of color pattern mod- ifications essentially similar to those found in Eleutherodactylus. Rhacophorus microt!/mpanum (Gunther), a member of the family Rhacophoridae, lives in the mountainous regions of Ceylon. It lays about 20 eggs on land in situations similar to those occupied by E. nubicola, and its embryo is of the eleutherodactyloid type. Kirti- singhe (1957: 11, 66-67) has pointed out the striking parallelism be- tween this species and E. nubicola of Jamaica. The following pat- terns, as described for nubicola, also occur in R. microtympanum: (1) mottled pattern with dark pigment arranged in the shape Qf a W-mark on the back in the suprascapular region, (2) a pair of pale, dorsolateral stripes, one on each iide from the eye to the groin, (3) a narrow middorsal stripe, (4) a pale interocular bar bordered posteriorly and anteriorly by darker pigment. (See Kirtisinghe, 1957, fig. 50) In the Papuan region live a number of species belonging, to sev- eral genera (Sphenophryne, Oreophryne, Asterophrvs, etc.) of the family Microhylidae. Many of these species are highly variable in color pattern. Indeed, Parker's. (1984: 168) descripti6n of color pat- tern variation in Oreophryne uariabilis (Boulenger) from Mt. Bon- thain, southern Celebes, reads like the description of some species of Eleutherodactylus. The following patterns are described: (I) a mottled pattern with a distinct, dark W-shaped mark ·on the back in the suprascapular -region, (2) a pair of broad, yellow, dorsolateral light stripes··from eye to groin, (3) a narrow middorsal stripe from snout to groin. Parker also describes interocular light bars and ocellar dark spots in the groin in Other species of this genus. These Papuan microhylid genera are all probably rather closely related and are the only genera of Microhylidae anywhere in the 258 BULLETIN FLORIDA STATE MUSEUM Vol. 5 world known to have not only terrestrial breeding habits but also an eleutherodactyloid pattern of embryonic development with the young hatching as tiny frogs. Thus members of three different families of frogs, from three dis- tinct geographic regions, have independently developed very similar reproductive habits and embryos and at the same time show the same kinds of color patterns and pattern variability. It is perhaps signifi- cant that all three types are found in mountainous regions, and while not at present confined to mountains, they may well have evolved as montane forms. But the evolutionary significance of this striking parallelism is still not clear. LITERATURE CITED Coin, Coleman Jett 1947. Studies on the life history of Eleutherodact!/lus ricordii planirost,is (Cope) in Florida with special reference to the local distribution of an allelo- , morphic color pattern. Univ. Florida Studies, Biol. Sci. Series, vol. 4, no. 2, xi t 66 pA, 6 pls., 7 text figs. 1950. Color pattern inheritance in some frogs Qf the genus Eleutherodactvlus. Bull. Chicago Acad. Sci., vol. 9, no. 1, pp. 1-15, 1 pl. 1954. Remarks on evolution of color pattern in the gossei group of the frog genus Eleutherodact!/lus. Ann. Carnegie Mus., art. 10, pp. 185-195, 2 figs. 1958. Further studies on color pattern inheritance in -the frog, Eleutherodacty- lus nubicola. Year Book Amer. Philo. Soc. for 1957, pp. 248-250. Kirtisinghe, P. 1957. The amphibia of Ceylon. Privately published, Ceylon, xiii + 112 pp., 74 figs. 1 Parker, H. W. 1934. A monograph of the frogs of the family Microhylidae. British Mus. (Nat. Hist.), London, viii + 208 pp., 67 fgs. Sturtevant, Alfred Henry 1948. The evolution and function of genes. Amer. Sci., vol. 86, no. 2, pp. 225-286. Contributions to the BULLETIN OF THE FLORIDA STATE MUSEUM may be in any field of biology. Manuscripts dealing with natural history or systematic problems involving the southeastern United State5 or the Caribbean area are · solicited especially. Manuscripts should be of medium length-12 to 200 printed pages. Examination for suitability is made by an Editorial Board. The BULLETIN is distributed worldwide through institutional subscriptions and exchanges only. It is considered the responsibility of the author to distribute his paper to all interested individuals. To aid in this, fifty copies are furnished the author without cost. PREPARATION OF MANUSCRIPT Highly recommended as a guide is the "Style shedt for the scientific serial publica- tions of the American Museum of Natural History," second edition, revised, 1953. Manuscripts should be typewritten with, double spacing, with ample margins, and on only one side of the paper. The author should keep a copy; the copy submitted must be the original. Tables, legends of figures, and all footnotes should be as- sembled separate from the text. Several legends or footnotes may be placed on a single sheet. Illustrations, including maps and photographs, should be referred to as "figures" wherever possible. All illustrations are reduced to a maximum of 444 by 71A inches. The scales, wherever it·is necessary, should be incorporated into the figure. All references to literature should conform with the bibliographic style used in recent numbers of the BULLETIN. Spell out in full the titles of non-English serials. Footnote material should be kept to a minimum. However, provide copy for a footnote detailing the title, affiliations, and address of the author (see recent numbers of the BULLETIN). Manuscripts must be accompanied by a synopsis-a brief and factual summary (not a mere description) of the contwts and conclusions, whith points out the presence of any new information and indicates its relevance. In it list all new Organisms described and give their ranges; indictite all taxonomic changes pro- posed. The synopsis, written in full sentences, should be concise, but completely intelligible in itself without reference to the paper, thereby enabling the busy reader to decide more surely than he can from the title alone whether the paper merits his reading. The synopsis will be published with the paper, hence it does not replace the usual conclusions or summary sections. It will also serve as copy for the abstracting services. Manuscripts and all editorial matters should be addressed to: Editor of the BULLETIN Flint Hall University of Florida Gainesville, Florida