Atelopus longirostris Cope 1868

Atelopus longirostris is distinguished easily from A. lynchi and A. carauta by having discrete yellow spots on the dorsal surfaces, and an elongate marking behind each eye. A. longirostris has a long, narrow snout, more so than that of A. lynchi (see Cannatella 1981, p. 137). The tip of the snout is not curved in A. longirostris, as it is in A. carauta (where it is curved toward the venter). The canthus is distinctly concave in A. longirostris but straight in A. carauta and A. lynchi (Cannatella 1981).The type specimen was described by Cope (1868) but has apparently been lost (Peters 1973). A. longirostris has been confused with other species (A. coynei, A. carauta, A. lynchi); specimens from Colombia in particular belong to a different species (Peters 1973; Lötters 1996). Cannatella (1981) has described specimens from northern Ecuador and southern Colombia as a new species A. lynchi, rather than A. longirostris, and pointed out that some Ecuadorian Atelopus specimens tentatively assigned to A. longirostris (Peters 1973) are in fact A. coynei. The genus Atelopus, found in Central and South America, has experienced dramatic declines due to amphibian chytrid fungus. Of 113 described and putative species, at least 30 species appear to be extinct, having been missing from all known localities for at least 8 years (La Marca et al. 2005). Only 52 of the surviving species have sufficient data with which to evaluate population trends; of these, 81% (42 of 52) have population sizes that have been reduced by at least half (La Marca et al. 2005). Only 10 of the 52 species appear to have stable populations (La Marca et al. 2005). Higher-elevation species (those living at least 1000 m a.s.l.) have been hit the worst, with 75% (21 of 28) having disappeared entirely (La Marca et al. 2005). Most Atelopus species are restricted to very limited areas (no more than two localities) and occur along mid- to high-elevation streams (1500-3000 m a.s.l., though the maximum vertical range is from sea level to permanent snow; Lötters 2007), a habitat preference frequently associated with the co-occurrence of chytridiomycosis (La Marca et al. 2005). Habitat loss has occurred within the ranges of many Atelopus species, but does not appear to be a major factor in the decline of most Atelopus species; 22 species declined despite occurring in protected areas (La Marca et al. 2005). Many Atelopus species are local endemics, putting them at particular risk of extinction, with at least 26 species known only from a single population inhabiting a narrow altitudinal range (La Marca et al. 2005).

Atelopus longirostris occurred on the north-west side of the Ecuadorian Andes from the Provinces of Esmeraldas, Imbabura, Cotopaxi and Pinchincha, at elevations from 200 to 2,500 m a.s.l. Its main habitat was montane and lowland tropical rainforest (Stuart et al 2008).

This species bred in streams (Stuart et al 2008).

This species was not seen in Ecuador from 1989 until 2017 despite extensive searches, and was considered extinct after twenty years without a sighting (Stuart et al 2008; Tapia et al 2017). The cause of extinction or population crash has not been determined, but possibilities include the disease chytridiomycosis, climate change, pollution, and habitat destruction. Chytridiomycosis is believed to have been a factor in extinction, as for many species of Atelopus, but this is uncertain; A. longirostris occurred at lower elevations where this disease has not previously been considered to be a threat. Its range lay partly within at least one protected area, the Reserva Ecológica Cotacachi-Cayapas and may have extended to the northern limit of the Reserva Ecológica Los Illinizas (Stuart et al 2008).In March 2016, four adults were rediscovered in a newly documented site, Junín, Provincia Imbabura (1250–1480 m) in patches of native forest otherwise heavily disturbed by agriculture and pastures. One of the patches of forest is in a protected area, Junín Community Reserve, which the authors urge to be expanded to ensure the species persistence (Tapia et al 2017). Interestingly, Tapia et al (2017) found a high prevalence of Bd in the amphibians of Junín Community Reserve but A. longirostris tested negative.

Atelopus longirostris Cope

Atelopus longirostris Cope, 1868:116 [type locality: “Valley of Quito,” Ecuador].—Orton, 1871:693.—Espada, 1875:155.—Nieden, 1926:81.—Guibe, 1948:32.

Atelopus varius [not Stannius].—Boulenger, 1880:46.

Phryniscus longirostris.—Boulenger, 1882:153.

Phryniscus boussingaulti Thominot, 1889:28 [type locality: “Entre Latacunga et Guayaquil, au sud de Quito,” Ecuador].

HOLOTYPE.—Apparently lost. Not located in either the Academy of Natural Sciences of Philadelphia or the National Museum of Natural History.

DESCRIPTION.—Snout from above strongly projecting, with comparatively sharp point, distinct inward curve from snout tip to nostril, canthi almost parallel from nostril to point just before corner of eye where each canthus flares outward; area from snout tip along canthus, outer eyelid border, and to corner of head raised and fleshy. Dorsum of head flat, shagreened; eyes prominent, protrusive. Snout from side strongly protrusive beyond lower jaw, with upper lip distinct from protrusion as well; nostril lateral, posteriorly directed, slightly below fleshy canthus, and almost directly above end of lower lip. Loreal region sunken, eye overhung by fleshy eyelid margin; vertical ridge behind eye and horizontal ridge from eye to corner of head; side of head and tympanic region shagreened, with few larger, white-topped pustules. Dorsum and sides of body heavily shagreened, with scattered white-topped pustules on sides only, belly almost smooth or with finely wrinkled surface. Vertebral column forms distinct, raised middorsal ridge from shoulder level to coccyx.

Forelimb very slender, forearm more fleshy than humeral region, pustulate on shoulder and dorsum of humeral region, shagreened elsewhere. Entire hand somewhat fleshy, heaviest around first digit, which is distinct although almost totally enmeshed in fleshy webbing. Tips of digits slightly swollen; second digit shorter than fourth; second, third, and fourth digits almost free of webbing. Subarticular tubercles present only at digit base, weakly developed; large palmar, small plantar tubercle.

Hind limb long, slender, shagreened but without pustules; tibiotarsal articulation reaches points between nostril and posterior corner of eye when carried forward along body; heels overlap or touch when tibiofibulae are parallel to femora held at right angles to body. Fleshy webbing of foot extends to swollen tips of digits, but is incised between digits; first digit completely hidden in webbing, tip of second distinct. Subarticular tubercles present but weak; outer metatarsal tubercle raised, prominent; inner metatarsal tubercle elongate, flattened, kidney shaped.

COLOR IN ALCOHOL.—Dorsum brown, from reddish brown to very dark brown, snout tip light brown, sharply distinct from top of head, which usually has few scattered white spots, characteristically one on back of each eyelid. Dorsum with scattered, often bilaterally paired light spots; dorsolateral area over shoulder usually with elongate white spot, almost forming stripe in some individuals, usually followed dorsolaterally by one or two additional elongated spots or streaks. Sides usually unicolor, occasionally lightly vermiculated, abruptly changing into very light, immaculate yellow of belly, from chin to anus, with occasional individual showing brown vermiculations ventrally. Legs as body, usually unicolor brown above, spots laterally and dorsolaterally, light yellow below, posterior surface of thigh usually vermiculated. Tips of digits yellow, soles of feet usually dark, but some with yellow areas.

COLOR IN LIFE.—Dorsum blackish brown, shading into a light chocolate brown low on sides. Spots on head, back, sides, and dorsum of legs dull yellow. Those on back of femur white. Tip of snout yellowish brown, brighter than rest of spots. Series of tiny white stipples along sides and between eye and arm, behind temple. Eye black all over. Belly white, not flesh colored. Tips of digits very light orange. Stripe along upper lip yellow as dorsal spots, but grades into white of belly near corner of mouth. In a female heavy with eggs, the eggs show brightly through the belly flesh, giving a strong orange tint to ventral surfaces.

ECOLOGY.—I have collected this species in several places on the western slope, and always found the specimens in rocky areas on the edges of streams. I have found individuals under rocks and also out walking about in and over the rocks in full sunlight. The habitat is not usually a shaded one, but is very exposed. A pair in amplexus was taken at Apuela on April 17, 1959, which was at the end of the rainy season. The female was heavy with eggs.

COMPARISONS.—An obviously close relative of longirostris is A. varius glyphus, described by Dunn (1931). The type and paratype series are in the National Museum of Natural History, and I have compared them directly with the series of longirostris available from Ecuador. In addition to the general similarity in color pattern, both species have rather slender bodies, elongate, thin limbs, and a comparatively prominent, overhanging snout. It is possible that the relationships of glyphus are not with varius at all, and it should be regarded as a distinct species. The two paratypes from Porto Obaldia, noted as quite different by Dunn (1931:396), are very reminiscent of A. varius loomisi Taylor (1952).

I have examined the holotype of Phryniscus boussingaulti Thominot, and see no reason not to continue its synonymy with longirostris. The specimen is badly faded dorsally, retaining no trace of pattern. There is a series of clear, rounded, whitish spots laterally, which, although not typical of longirostris, is almost duplicated in some of my specimens (e.g., USNM 193833). Photographs of the holotype were taken and are on deposit at the National Museum of Natural History.

Atelopus longirostris marmorata Werner

This taxon was described by Werner in 1910, page 600, from “Ecuador,” without more detailed information on the type locality. He noted that the females in his collection appeared different from the males of longirostris, and used the trinomen above for them. Nieden (1926:81) included the taxon in the synonymy of Atelopus longirostris, and the few citations since that time (Rivero, 1963; Cochran and Goin, 1970) accepted Nieden’s action as having done away with the taxon. Careful examination of all of Nieden’s work in this volume of Das Tierreich, however, shows that he included all trinomials in existence prior to his work as synonyms of a binomial, and then mentioned those he considered to be recognizable in the text below the synonymy I find only three subspecies mentioned in the entire volume, all in the text material. One of these is marmorata. It is, therefore, not particularly valid to cite Nieden as the authority for the synonymy of marmorata with longirostris. I think it entirely possible that Werner had specimens of a completely different species in hand. Since elegans occurs on the western slopes in Ecuador, and may actually be sympatric with longirostris in part of its range, I would suggest that Werner may not have recognized specimens of elegans among his longirostris material. Unfortunately, the type specimens were destroyed during World War II, and marmorata becomes a nomen dubium.

Zoogeography and Evolution

Rivero (1963:112) has remarked, “Frogs of the genus Atelopus seem to be very sensitive to environmental factors.” This is certainly true of Ecuadorian species, particularly at altitudes above the subtropical zone. The Ecuadorian terrain makes exploration by herpetologists difficult, and not too much work has been done on the outward facing slopes of the Andes as a consequence. Access to areas between 1500 and 3300 meters has always been and still generally is available only on foot or horseback, although road construction on both east and west slopes continues to make more areas accessible by car. Most of my personal collecting efforts in Ecuador have been focussed on these slopes. I have spent comparatively little time in the Pacific lowlands, and even less in the Amazonian rain forest. Herpetologically, the latter has been the most popular and thoroughly explored area in the country, and, while it continues to produce both novelties and surprises, it still does not hold the same promise for investigation of ecological and evolutionary phenomena as do the Andean slopes.

In 1958–1959 I made a series of seven transects from the inter-Andean valleys to the Pacific lowlands, and in 1962 my group made five transects on the eastern slopes, usually, although not always, penetrating to the Amazonian lowlands. Such transects perforce follow available routes, and of the total of twelve transects, five were by mule and on foot, six were by road in truck or Landrover, and one was by rail. Since the routes taken were either mule trails several hundred years old, or roads that follow such old trails and replace them, it is to be expected that each trail would be separated by some distance and usually some fairly substantial physical or physiographic barrier from the next trail, since arrieros follow paths of least resistance, and do not develop and follow duplicate trails ending at the same place very often (except as smugglers). The physiographic barriers are deep river valleys, sheer cliffs, or too heavily dissected slopes, all of which can and do act as biological barriers to species movement equally well. In addition, rain shadow slopes, where the vegetation and moisture available change radically within a few thousand feet, occur on both east and west slopes, and represent a major biotic barrier to horizontal migration of mountainside species. Since these various types of barriers are common, and are most often encountered between the traditionally used trails, one would expect each transect to produce different taxa, each evolving in an isolation produced by the barrier of altitude on one hand, and the many horizontal barriers mentioned above on the other hand. This proved to be particularly true of frogs of the genus Atelopus and lizards of the genus Proctoporus. Since representatives of both genera are fairly often found even if the total collecting time available is slight, they become good indicators of evolutionary patterns, effects of isolation, adaptive radiation, and other phenomena often associated with the study of insular faunas. In point of fact, mountains such as the Andes of Ecuador can be considered “continental islands,” and many of the phenomena long discussed by students of insular biology are equally observable there, with the added advantage that the mountains are not dependent upon waifs and drifters for replenishment of the supply of evolutionary materials.

The taxa in the genus Atelopus lend themselves to the study of these phenomena because of the sensitivity to environmental factors already noted by Rivero. But, although they are sensitive to, and often limited by, very slight environmental changes, they are also capable of adaptation and divergence, which permits them to conquer new environments. Suitable habitat for the taxa in the lowlands of Amazonian Ecuador as well as in the lower altitudes on the Pacific side appears to be continuous, with few barriers to gene flow, and lowland species tend to have a considerable latitudinal range. A. pulcher ranges throughout the southern half of Amazonian Ecuador and a considerable distance in Peru, from whence it was originally described. It stops at about the Río Pastaza, and is replaced by A. palmatus, the northern limit of which is currently unknown. A. elegans is a lowland species on the west side of the Andes in Ecuador, and is found north well into Colombia and south to the point where the wet areas are replaced by the dry coastal scrub, under the influence of the Humboldt Current. Figure 31 is an attempt to show how the situation changes when the species found on the slopes at slightly higher altitudes are examined, and the opportunities for horizontal migration are diminished by the multiplicity of barriers. The extremely narrow distribution of the species found at altitudes above 2000 meters is in part an artifact of the collecting opportunities, which are restricted to a few hundred yards laterally to the trails, but it is fairly certain that there is a high degree of endemism on the higher slopes, and it is very likely that the endemic species are totally allopatric. Finally, after an altitude sufficient to permit the inter-Andean highlands to be reached, the picture is reversed. The highly successful species A. ignescens is capable of living at very high altitudes under most inhospitable conditions. As a result, it is the most widespread amphibian species in the Andean highlands, ranging from well into Colombia practically to the Peruvian border. The ability to live in paramo and subparamo climates has eliminated all physical barriers to movement by the species, and it is ubiquitous over a considerable range. It is the only species that may in several cases be sympatric with those species with restricted ranges on the outer slopes. I have found A. ignescens within 2 to 3 kilometers of populations of A. pachydermus, and it appears that the same is true for A. bomolochos, much farther to the south. A. ignescens is, of course, sympatric with A. bufoniformis, which has been taken as high as 3400 meters, and is known from both slopes of the Andes, but only in the most northern provinces of Ecuador. A. bufoniformis is the end point in the evolutionary sequence discussed below.

While the evolutionary sequences are not entirely clear, and the evidence used here is primarily morphological and zoogeographic and thus could be easily disproven by data from cytological, genetic, biochemical, and other more modern tests, it still seems legitimate to construct hypotheses concerning what has happened to Atelopus phylogeny in Ecuador. The potential absurdity and essential sterility of such hypotheses can only become evident after considerable research on the taxa concerned, research which I do not anticipate being done for many years—and which might even verify my guesses here.

In any event, the evidence derived from an examination of the distribution and morphology of the two groups of species indicates the direction that evolution has taken. With regard to distribution, the taxa of the longirostris group are almost entirely low altitude species. The only exception may be A. planispinus, which is found on Mount Sumaco, which reaches an altitude of 3900 meters. The type locality for the species, at the foot of the mountain, is at 500 meters, and specimens in the American Museum of Natural History are recorded as coming from the top of Sumaco, which gives the species a possible altitudinal range of 500 to 3900 meters, a most unlikely situation. Since this taxon is part of the longirostris group, and since it has not been taken elsewhere in the Amazonian lowlands, as would be expected if it actually occurred at San José Nuevo (the modern name for San José de Moti, or Mote), I predict it will eventually be found to occupy the slopes of Sumaco at lower to intermediate altitudes (perhaps 1000 to 2000 meters) and probably on the eastern slope only (because of habitat change on the back or western side of the mountain). Other than A. planispinus, the taxa in the longirostris group do not reach altitudes above 2000 meters, and enjoy their greatest abundance below 1000 meters.

The taxa of the ignescens group, on the other hand, are almost all high altitude forms. The one exception, A. mindoensis, is an enigma. As I point out frequently in this paper, this taxon is intermediate between the two groups in many ways. For example, it tends to fall halfway between the groups in almost all of the analyses of proportional measurements. Leaving that species to one side, the remaining members of the igenscens group all show a common zoogeographic pattern. They are very restricted horizontally, and appear to be equally restricted vertically to a fairly narrow altitudinal band. They have not, and I believe they are unable to, spread widely on a horizontal plane, although the horizontal distribution of each is certainly greater than my data show. The vertical restrictions are for the most part real, and I think the various taxa show sharp response to and are closely confined by the changing conditions which accompany increasing altitude. All of these species are living in a comparatively young area, geologically speaking. The surrounding lowlands are much older, and it seems plausible that the ancestral species in the genus Atelopus are the lowland taxa, and the highland taxa have been repeatedly derived from the lowland stocks. I would further suggest that each of the highland taxa (with the probable exception of A. ignescens and the possible exception of A. bufoniformis) is the product of a separate, successful invasion of the highlands by lowland stock. This, of course, means that the ignescens group is clearly not monophyletic, but simply represents a series of separately derived species all differentiating from the ancestral stock along similar morphological lines.

The morphology of the two groups of species tends to verify the hypothesis. The adaptive changes evolved in response to higher altitudes and colder temperature include shorter, more stocky bodies and a reduction of the limbs from long and slender to short and stout. This suggests, of course, the development of mechanisms to reduce total surface and concomitant heat loss. The highland species are much less likely than the lowland taxa to be found out and actively moving about except during breeding periods, when they become abundant.

While not as sharply distinctive between groups as the body and limb configuration, the thickening of the skin in the webbing between the digits, and the overall fleshiness of the highland species in general, may well be another modification toward heat conservation. It reaches maximum development in A. bufoniformis, the taxon which has both the maximum development of the characteristics discussed here and reaches, with A. ignescens, maximum altitudes.

McDiarmid (1971:58) has discussed the evolutionary relationships of Atelopus with other genera of the family Bufonidae, and demonstrated that the genus is quite advanced rather than primitive in a large number of its characteristics. He pointed out several phenomena concerning the genus Atelopus that are of interest in understanding evolutionary events in Ecuador. He stated:

With the initial uplifts of the Andes in Peru and Ecuador in the Late Cretaceous…. entire new areas became available to and were exploited by Atelopus. The ancestral stock probably came from southeastern portions of the Guiana shield where the more primitive Atelopus flavescens is found today. However, it is possible that Atelopus initially invaded the montane habitat in the southern Andes and moved north along their eastern slope. It appears also that the higher elevations have been invaded several times.

I would agree completely with the concept of multiple invasion of the higher elevations, but think it unlikely that there has been very much movement north along the eastern slope of the Andes from the south. The evidence presented above seems to indicate fairly strongly that lateral or horizontal movement is not a common phenomenon in Ecuador, probably as a consequence of multiple barriers. McDiarmid (1971:58) has himself provided an additional argument against horizontal movement, for he said:

The apparent close association of species of Atelopus with mountain streams has restricted species to particular drainage systems. The resulting isolation (drainage system and mountain top) probably accounts for the striking differences among many populations and probably has been an important factor contributing to speciation.

The only species of the ignescens group which has been able to break out of the limitations noted above for other species is A. ignescens itself. It has been a highly successful invader of the inter-Andean valleys in Ecuador, and its vertical range is such that the nudos or knots between the valleys are not a barrier to horizontal movement. The species ranges from the southernmost provinces in Ecuador well into Colombia, a horizontal range almost as great as that of any lowland species. Only A. bufoniformis shows greater differentiation than A. ignescens along the morphological sequences to high altitude adaptation, as can be seen in the graphs of proportional measurements. A. ignescens represents an explosive movement taking place, because its ancestral stock continued the upward expansion of its range until it spilled over the wall of the Andes, opening up the whole inter-Andean plateau to settlement.

The success and widespread distribution of A. ignescens leads one to a second look at the other members of its group. It is entirely possible that they represent repeated successful invasions of the outer slopes by A. ignescens, adapting and differentiating in different areas. A. ignescens definitely occurs on the eastern slope within a few kilometers of A. pachydermus, and the relationship between them seems to be close. The same is true for A. bomolochos, which is equally close to A. ignescens both geographically and morphologically. Why, then, is it not possible that A. ignescens is the direct ancestor to these restricted populations?

Perhaps the strongest argument against this hypothetical ancestry is the fact that A. ignescens tends to maintain its specific identity throughout a considerable distance in the inter-Andean plateau. Although quite variable, and perhaps with recognizeable subspecies in several places, it still remains one species throughout, in my estimation. It is uncertain how much contact there is between populations separated by nudos, but I suspect there is less contact between them than within such populations, certainly a situation to encourage local populational differentiation, if the species is plastic enough to respond. I suggest that A. ignescens is a species well advanced evolutionarily, and highly adapted to a very inhospitable environment, and therefore has comparatively little evolutionary potential. The likelihood of such a species invading a series of isolated drainages on both sides of the Andes of Ecuador and giving rise to such a large number of very distinct species as are found there seems very slight to me.

On the other hand, I must reverse that argument with regard to A. bufoniformis. This species has gone beyond A. ignescens in every one of the adaptive characteristics discussed here. It is clearly the terminal species in the differentiation sequence. It is difficult for me to visualize any other ancestor for this species than A. ignescens, and this causes some considerable discomfort. At one time I had thought A. bufoniformis to be an inhabitant of the high eastern slope, found slightly below the range of A. ignescens. But material in the UKMNH collections show it from 3400 meters, and well into the higher parts of the inter-Andean Plateau. It does not seem to occur in the inter-Andean valleys, although it has been found low enough on the outer slopes that altitude would not be a barrier. It appears to have been a comparatively recent derivative of A. ignescens stock, and is perhaps even more thoroughly adapted to the cold, wind-swept, high altitude paramo than the parent stock. It is probably in the process of expanding its range at those high altitudes, but may face problems in the fact that the habitat it utilizes is not sufficiently continuous to permit it to reach all existent paramo.