Middle American Frogs of the Hyla microcephala Group

Pattern and duration of notes.—In all five taxa the basic pattern consists of a call-group made up of one primary note followed by a series of shorter secondary notes. In some species the secondary notes differ from the primary in other characteristics. Both subspecies of Hyla microcephala have a long, unpaired primary note followed by 0 to 18 (usually about 4) somewhat shorter paired secondary notes. In calls of Hyla m. microcephala the mean duration of the primary is 0.131 (0.10-0.16) second and that of the secondaries is 0.101 (0.05-0.14) second, whereas in H. m. underwoodi the mean duration of the primary is 0.018 (0.05-0.15) second and that of the secondaries is 0.086 (0.06-0.11) second.

Hyla robertmertensi has a reverse of this pattern in that the primary note is paired and the secondaries are unpaired. In the sample studied a call-group contains 0-28 secondary notes (generally about 3). The mean duration of the primary is 0.091 (0.07-0.11) second and that of the secondaries is 0.040 (0.025-0.06) second.

Hyla phlebodes and sartori have call-groups composed of a rather short, unpaired primary and several short, unpaired secondaries (0-28 in phlebodes, 0-23 in sartori). The mean duration of the primary of phlebodes is 0.105 (0.07-0.16) second and that of the secondaries is 0.067 (0.035-0.12) second. The mean duration of the primary of sartori is 0.080 (0.07-0.09) second and that of the secondaries is 0.053 (0.035-0.07) second.

The two subspecies of H. microcephala are identical in call pattern and agree closely in duration of notes, although those of the nominate subspecies tend to be slightly longer. Hyla robertmertensi is distinctive in call pattern in that it is the only species having a paired primary; the duration of the primary is completely overlapped by that in the other species, but the secondaries tend to be the shortest in the group. The call patterns of H. phlebodes and H. sartori are identical and the range of duration of notes of phlebodes completely overlaps that of sartori, although both the primary and secondary notes of the latter tend to be somewhat shorter (Table 5, Pl. 16).

Fundamental frequency.—This parameter was analyzed for the primary notes. It was measured for the secondaries as well and was found to differ in magnitude in the same way as the primary note. In a few examples of both subspecies of H. microcephala a high primary note, in which the fundamental frequency is exceptionally high, is sometimes emitted (Fouquette, 1960b). None of these notes was used in this analysis; only the fundamental frequencies of normal primary notes are compared (Table 5, Fig. 7).

The two subspecies of H. microcephala agree closely in fundamental frequency. There is considerable overlap, but the difference between the means is significant at the 0.001 level of probability (t = 4.2406). The call of H. robertmertensi does not overlap that of H. sartori or either subspecies of H. microcephala in this parameter; but it does overlap that of H. phlebodes, although again the difference between the means is significant at the 0.001 level (t = 9.360). Hyla phlebodes and sartori have the lowest fundamental frequencies, and there is some overlap, but here too the difference between the means is significant at the 0.001 level (t = 4.923).

Dominant frequency.—A dominant band of of frequencies cuts across the harmonics of the fundamental, obscuring the harmonic pattern and generally shifting upward in frequency. The midpoint of this band is measured at the terminal border as the dominant frequency. As with the fundamental frequency, only the normal primary notes were utilized in the comparisons (Table 5, Fig 8).

The two subspecies of H. microcephala agree more closely in this parameter than in fundamental frequency. The overlap is great, but the difference between the means is significant at the 0.001 level (t = 3.658). The calls of both subspecies completely overlap that of robertmertensi in this parameter, but the difference between the means is significant at the 0.001 level. The calls of H. phlebodes and H. sartori overlap considerably in this characteristic, although the difference between the means is significant at the 0.001 level (t = 7.504) (Fig. 9). The call of neither species overlaps those of H. microcephala and robertmertensi.

Repetition rate.—The repetition rate of the secondary notes, in calls consisting of more than one secondary, was measured for each form. A considerable amount of variation in this parameter was found in all of the taxa (Table 5). This variation probably is due in part to the effect of temperature differences. Repetition rate is the only parameter analyzed for which there is a correlation with the air-temperature, but even here the correlation is weak, probably due to the microenvironmental effects of humidity, air-movement, and other factors in addition to the ambient air temperature that influences the body temperature of the frogs. These rates are nearly alike in both subspecies of H. microcephala and in phlebodes. The repetition rates in H. robertmertensi and H. sartori are considerably faster than in the other three taxa. Hyla sartori has the fastest repetition rate of the group.

In all characteristics of the mating calls the two subspecies of H. microcephala agree closely, as might be expected, although the differences are statistically significant. Hyla robertmertensi is distinctive in call pattern and seems to be closer to microcephala in dominant frequency but closer to H. phlebodes in fundamental frequency. Thus, it is somewhat intermediate between microcephala and phlebodes. The identical pattern and similarity in fundamental and dominant frequencies of the calls of H. phlebodes and H. sartori possibly indicate close relationship.

Geographic variation in call.—Hyla m. microcephala has higher fundamental and dominant frequencies in Costa Rica than in Panamá. In Costa Rican H. m. underwoodi the fundamental and dominant frequencies are lower than in other parts of the range. Frogs of this subspecies recorded in Nicaragua and Honduras have slightly lower dominant frequencies and higher fundamental frequencies than those recorded in Guatemala or Oaxaca. The duration of both primary and secondary notes decreases to the south; samples from Nicaragua and Costa Rica have the shortest notes. Comparison of duration of notes in the two subspecies shows that the Panamanian H. m. microcephala have slightly longer notes than do any H. m. underwoodi; the more northern populations of H. m. underwoodi from México most closely approach H. m. microcephala in this characteristic.

The calls of H. robertmertensi in Oaxaca have higher dominant and fundamental frequencies and longer secondary notes than do those in Chiapas.

The calls of H. phlebodes recorded at Puerto Viejo, Costa Rica, have slightly lower dominant frequencies than do those recorded at Turrialba, Costa Rica, and in Panamá, whereas those recorded at Turrialba have lower fundamental frequencies than in other samples. The duration of notes is slightly shorter in both Costa Rican samples than in those recorded in Panamá.

The frogs of the Hyla microcephala group breed in shallow grassy ponds. In some places they breed in permanent ponds, but usually congregate around temporary pools, such as depressions in forests, flooded fields, and roadside ditches. At the height of their breeding season, usually in the early part of the rainy season, the congregations are made up of large numbers of individuals. In April, 1961, and in June, 1966, the senior author noted nearly continuous choruses of H. m. microcephala in roadside ditches along the 75 kilometers of road between Villa Neily and Palmar Sur, Puntarenas Province, Cost Rica; on June 20, 1966, at Puerto Viejo, Heredia Province, Costa Rica, he estimated approximately 900 Hyla phlebodes in one pond, and two nights later noticed that the number of individuals had increased substantially. Other observations by the first author on size of breeding congregations include nearly continuous choruses of H. m. underwoodi between Villahermosa and Teapa, Tabasco, in July of 1958, an estimated 400 Hyla robertmertensi in a road side ditch 7.2 kilometers west-northwest of Zanatepec, Oaxaca, on July 13, 1956, and approximately 150 Hyla sartori around a rocky pool in a riverbed, 11.8 kilometers west-northwest of Tierra Colorada, Guerrero, on June 28, 1958.

The length of the breeding season seemingly is more dependent on climatic conditions in various parts of Middle America than on behavioral differences in the various species. Thus, Fouquette (1960b) found in the Canal Zone that H. m. microcephala formed breeding choruses from May through January, the entire rainy season in that area. In the wetter coastal region of Puntarenas Province, Costa Rica, the species breeds as early as mid-March, whereas in the drier region encompassing Guanacaste Province, Costa Rica, and southwestern Nicaragua breeding activity is initiated by the first heavy rains of the season, usually in June.

Hyla phlebodes inhabits regions having rainfall throughout the year. Although large breeding congregations are most common in the early parts of the rainy season, males probably call throughout the year. At Puerto Viejo in Costa Rica the senior author has heard Hyla phlebodes in February, April, June, July, and August. Charles W. Myers noted calling males of this species in the area around Almirante, Bocas del Toro Province, Panamá, in September, October, and February. An exception to the correlation between rainfall and breeding activity was noted by the junior author in Hyla phlebodes in the Canal Zone, where he noticed a decrease in activity of that species in October and November, when the rains are heaviest and most frequent. Furthermore, independent observations made by both of us indicate that H. phlebodes does not reach peaks of activity during or immediately after heavy rains, but instead builds up to peaks of activity two or three days after a heavy rain. This is in contrast to the other species, all of which characteristically inhabit drier environments than does H. phlebodes. Peaks of breeding activity in the other species occur immediately after, or even during, heavy rains.

The calling location of the males generally is on vegetation above, or at the edge of, the water. Hyla microcephala and H. phlebodes call almost exclusively from grasses and sedges; phlebodes usually calls from taller and more dense grasses than does microcephala. Except for some minor differences in calling location observed by the junior author (Fouquette, 1960b) in the Canal Zone, the differences in density and height of grasses utilized for calling-locations probably is dependent primarily on the nature of the available vegetation. Although bushes and broad-leafed herbs are usually present at the breeding sites, males of these species seldom utilize them for calling locations. Both H. robertmertensi and H. sartori have been observed calling from grasses, herbs, bushes, and low trees. Calling males of robertmertensi have been found two meters above the ground in small trees.

Daytime retreats in the breeding season sometimes are no more than shaded clumps of vegetation adjacent to a pond or in clumps of grass in a pond. Individuals of H. m. underwoodi were found by day under the outer sheaths of banana plants next to a water-filled ditch. Dry season refuges are unknown.

Amplexus is axillary in all four species. Egg deposition has been observed in H. m. microcephala, m. underwoodi, and phlebodes. In all three the eggs are deposited in small masses that float near the surface of the water and usually are at least partly attached to emergent vegetation. Each clutch does not represent the entire egg complement of the female.

Tadpoles are definitely known of only H. m. microcephala and phlebodes; these have been described in the preceding accounts of the species. The tadpoles of these two species can be distinguished readily (Pl. 15). The tadpole of H. microcephala has a uniformly white venter and nearly transparent tail, whereas in H. phlebodes the venter is flecked anteriorly and the tail is mottled. In life, H. microcephala is easily recognized by the orange posterior half of the tail, whereas the tail in H. phlebodes is mottled tan and grayish brown.

As mentioned in the account of Hyla m. microcephala, the species microcephala possibly is subspecifically related to Hyla misera, a frog widespread in the Amazon Basin. Hyla misera resembles microcephala in coloration, external structure, and cranial characters. The frontoparietals are equally poorly ossified, and the frontoparietal fontanelle is extensive. Our principal reason for not considering the two taxa conspecific at this time is our lack of knowledge concerning the color of living H. misera, the structure of the tadpoles, and the characteristics of the mating call. Even with the absence of such data that we think essential to establish the nomenclature status of the taxa, we are confident that the two are sufficiently closely related that any discussion of the phylogenetic relationships of one species certainly must involve consideration of the other.

Hyla misera possibly is allied to other small yellowish tan South American Hyla that lack dark pigmentation on the thighs. Probable relatives are Hyla elongata, minuta (with goughi, pallens, suturata, velata, and possibly others as synonyms), nana, and werneri. The consideration of the interspecific relationships of these taxa is beyond the scope of this paper, but we can say that each of these species has a pale yellowish tan dorsum, relatively broad dorsolateral brown stripe, and narrow longitudinal brown lines or irregular marks on the dorsum. Furthermore, examination of the skulls of elongata, nana, and werneri reveals that they are like misera and microcephala in the nature of the frontoparietal fontanelle and in having a greatly reduced quadratojugal. Thus, on the basis of cranial and external characters the Hyla microcephala group can be associated with Hyla misera and its apparent allies in South America. This association can be only tentative until the mating calls, tadpoles, and chromosome numbers of the South American species are known.

Among the Middle American hylids, only the Hyla microcephala group and H. ebraccata have a haploid number of 15 chromosomes (Duellman and Cole, 1965). All other New World Hyla, for which the number is known, have a haploid number of 12; the only other Hyla having 15 is a Papuan Hyla angiana (Duellman, 1967).

Hyla ebraccata occurs in the humid tropical lowlands of Middle America and the Pacific lowlands of northwestern South America. It is the northernmost, and only Central American, representative of the Hyla leucophyllata group, which is diverse (about 10 species currently recognized) and widespread in tropical South America east of the Andes. This group is characterized by having broad, flat skulls with larger nasals and more ossification of the frontoparietals than in the Hyla microcephala group. The quadratojugal is present as a small anteriorly projecting spur that does not connect with the maxillary. Externally, the Hyla leucophyllata group is characterized by having a well-developed axillary membrane, uniformly yellow thighs, and a dorsal color pattern in many species consisting of a dark lateral band, a pale dorsolateral band or dorsal ground color, and a large middorsal dark mark. In some species, the dorsal pattern consists of small dark markings or is nearly uniformly pale. At least in the Central American Hyla ebraccata, the mating call consists of a single primary note followed by a series of shorter secondary notes, the tadpoles have xiphicercal tails and lack teeth, and the haploid number of chromosomes is 15. On the strength of these observations it seems imperative to consider the Hyla leucophyllata group as a close ally to the Hyla microcephala group. Successful artificial hybridization supports the close relationship of H. m. microcephala and phlebodes; partial success of artificial hybridization of these two with ebraccata (Fouquette, 1960b) provides further evidence for close relationship between the Hyla leucophyllata and Hyla microcephala groups.

In México and northern Central America two small species, Hyla picta and Hyla smithi, comprise the Hyla picta group. These frogs resemble members of the Hyla microcephala group by having a yellowish tan dorsum with a dorsolateral white stripe and uniformly yellow thighs. Furthermore the mating call is not unlike those of the species in the Hyla microcephala group. Despite these similarities, the Hyla picta group differs from the Hyla microcephala group by having a well-developed quadratojugal that connects to the maxillary, tadpoles with teeth present and caudal fins completely enclosing the caudal musculature, and a haploid number of 12 chromosomes. In all of these characteristics the frogs of the Hyla picta group more closely resemble other Middle American Hyla than they do the Hyla microcephala group. Therefore, it can best be presumed that the superficial resemblances of coloration and the mating call are the result of convergence.

Since the Hyla microcephala and leucophyllata groups apparently are related and since the greatest diversity of these frogs is in South America (if Hyla misera and its relatives are placed with the Hyla microcephala group), it seems appropriate to place the centers of origins of these groups in South America. Therefore, the Hyla microcephala group and Hyla ebraccata of the Hyla leucophyllata group either have immigrated into Central America, or they are representatives of those groups that were isolated in Central America during most of the Cenozoic when South America was separated from Central America.

The interspecific relationships of the species in the Hyla microcephala group are not clear. On the basis of coloration, H. m. microcephala and H. robertmertensi are close, and H. m. underwoodi and H. phlebodes are nearly identical. The mating calls of H. phlebodes and sartori closely resemble one another, whereas the call of robertmertensi is intermediate between these and microcephala.

In most respects Hyla microcephala is distinct from the other species, and with the exception of the amount of ossification of the frontoparietals, the other species can be easily derived from a microcephala-like ancestor. Possibly the slightly increased ossification of the frontoparietals in robertmertensi, phlebodes, and sartori is secondary, or possibly after differentiation of the species the amount of ossification was further reduced in microcephala. If so, the species fall into a reasonable phylogenetic scheme that has microcephala as the extant species most like the ancestral stock.

We visualize the evolutionary history of the group to have followed a course that began with the invasion of Central America by a microcephala ancestral stock that differentiated into two populations in lower Central America—a microcephala-like frog on the Pacific lowlands and a phlebodes-like frog on the Caribbean lowlands. Differentiation could have been brought about by isolation by montaine or marine barriers. The population on the Pacific lowlands either was preadapted for subhumid conditions or became so adapted and dispersed northward onto the Pacific lowlands of northern Central America. Simultaneously the frogs on the Caribbean lowlands, which were adapted to humid environments, dispersed northward in the humid forested regions to southern México and crossed the Isthmus of Tehuantepec onto the Pacific slopes of Oaxaca and Guerrero northward to Jalisco. Subsequent development of arid conditions, possibly in the Pliocene, Pleistocene, or even as late as the Thermal Maximum in post-Wisconsin time, resulted in a restriction of the ranges in northern Central America, thereby isolating part of the phlebodes-stock on the Pacific slopes of México, where it adapted to drier conditions and evolved into sartori. The rest of the phlebodes-stock was restricted to the humid forests on the Caribbean lowlands of lower Central America. The increased aridity on the Pacific lowlands eliminated the microcephala-stock from southern Honduras and northwestern Nicaragua and in so doing left an isolated population on the lowlands of Chiapas and Guatemala, which differentiated into robertmertensi. The original stock on the Pacific lowlands of Panamá and southeastern Costa Rica became microcephala.

If the microcephala-stock was, as we believe, better adapted for existence under subhumid conditions than was the phlebodes-stock, the development of subhumid conditions in much of the lowland region of northern Central America and southern México would have permitted the expansion of the range of microcephala into the area now inhabited by H. m. underwoodi, while phlebodes was being eliminated from this area by climatic conditions that were unsuited to its survival there. Perhaps the similarity in coloration of H. m. underwoodi and phlebodes is the result of convergence or possibly hybridization occurred at the time the former was expanding its range and the latter's range was being restricted. If hybridization did occur, the differences in mating call subsequently were enhanced, thereby providing a valid isolating mechanism in sympatric populations.

Hyla microcephala and phlebodes range into northern South America. Probably both species entered South America in relatively recent times after they had differentiated from one another in Central America.

Transmitted July 11, 1967.