After decades of warnings by environmental scientists, population biologists, and demographers (see especially Osborn 1948; Carson 1962; Ehrlich 1968; Meadows et al. 1972), it is becoming increasingly apparent to an enlarging group of people that the Earth is entering a sixth spasm of mass extinction of life, at least comparable to and, perhaps, exceeding in scope the five episodes that have preceded it (Ehrlich and Ehrlich 1981, 1996; E. Wilson 1988, 1992; E. Wilson and Perlman 2000). What has come to be known as biodiversity decline is best documented in areas where the flora and fauna are most completely understood, e.g., the United States, and correspondingly less well understood in the areas of the world supporting the greatest amount of biodiversity—the tropics.

To use as an example the country that has been the focus of our research for more than three decades—Honduras—and the group upon which we have specialized—the herpetofauna, it is evident that the modern study of the Honduran herpetofauna began with the research of John R. Meyer that led to his dissertation, which appeared in 1969. Meyer's (1969) study documented a known herpetofauna of 196 species, including 53 amphibians and 143 reptiles. The current tally is 334 species, including 117 amphibians and 217 reptiles (McCranie and Wilson 2002; Wilson and McCranie 2002). With respect to the total count, there has been an increase of 138 species or 41.3% in the 33 years since 1969 to the present (although Meyer did not include five marine turtles species then known to occur in Honduran waters, nor five species of reptiles known in Honduran territory only from the Swan Islands, which are included in the total count of 334). Meyer (1969) included 35 species in the cloud forest herpetofauna of Honduras, although one species included by him (Ungaliophis continentalis) is not so included by us. Presently, we can document the presence of 122 species in one or more cloud forest regions of Honduras. This increase of 88 species (or 72.1% of the total now known) is largely a result of our field work in the country. Forty-two of these 88 species (47.7%) have been described as new species since 1979. In addition, populations of two species reported from cloud forest by Meyer (1969) have been described as new species (Ptychohyla spinipollex and Ninia lansbergi cloud forest populations of Meyer equal P. hypomykter and N. espinali, respectively).

There is still significant mountainous terrain in Honduras supporting cloud forest that has been incompletely sampled herpetofaunally. Such is the case with the Yoro Highlands, the Agalta Highlands, and the Santa Barbara Highlands. Given the frequency with which new taxa have been added to the Honduran cloud forest herpetofauna (2.3 taxa per year since 1972), it can be expected that additional forms await discovery in these yet poorly known ranges.

Acting in contraposition, however, is a more recent trend toward decline of herpetofaunal populations, which has been documented in Honduras by Wilson and McCranie (1998, 2003 a and b) and McCranie and Wilson (2002). This trend has been most evident in regions of the country in excess of 900 m in elevation and has most obviously affected the species composing the most distinctive group, i.e., those that are endemic to Honduras or otherwise restricted in distribution to Nuclear Middle America. Of the 125 species belonging to this group, 52 or 41.6% are considered to have declining populations, to be extinct, or to be extirpated in Honduras. This trend is extremely alarming, given the fact that the 125 species involved do not occur outside of Nuclear Middle America.

In light of the importance of the cloud forest environments of Honduras as centers of herpetodiversity and the accumulating evidence of the decline and disappearance of a significant amount of this diversity, it is the purpose of this paper to update our current understanding of the composition and distribution (both geographic and ecological) of this herpetofauna, to discuss its biodiversity significance, to examine its current conservation status, and to speculate on the future for this segment of the Honduran herpetofauna.

Fieldwork upon which this paper is based has been conducted by one or both of us since 1968. The material collected has been reported in a number of publications written by one or both of us since 1971 and summarized in Meyer and Wilson (1971, 1973), Wilson and Meyer (1985), and McCranie and Wilson (2002, in preparation).

The Coefficient of Biogeographic Resemblance algorithm (Duellman 1990) was used to demonstrate herpetofaunal relationships among the cloud forest ecophysiographic areas examined in this study. The formula is CBR = 2C/(N₁ + N₂), where C is the number of species in common to both formations, N₁ is the number of species in the first formation, and N₂ is the number of species in the second formation.

Honduras contains within its borders a major segment of the mountains of Nuclear Middle America (West 1964). Many of the ranges found within the country have portions high enough to support cloud forest (Fig. 1). Descriptions of the physiography of Honduras have appeared in Wilson and Meyer (1985) and McCranie and Wilson (2002), so this description will be limited to only those mountain ranges upon which cloud forest vegetation occurs.

Figure 1 Generalized map of the cloud forest areas of Honduras.

Elevations high enough to support cloud forest are distributed throughout the Serranía, the mountainous interior of Honduras, which is a portion of the Nuclear Middle American highlands (Fig. 1). The Serranía is traditionally divided into the Northern Cordillera and the Southern Cordillera, the latter distinguishable from the former by an overlay of Pliocene volcanic ejecta deposits (Wilson and Meyer 1985). Both of these cordilleras are interrupted by an irregular graben, called the Honduran depression, traceable from north to south through the Ulúa-Chamelecón Plain, the Valley of Humuya, the Comayagua Plain, and the Valley of Goascorán (Wilson and Meyer 1985). In effect, these physiographic features divide the mountainous interior of Honduras into four sectors, three of which are recognized as ecophysiographic areas on the basis of this division. They are the Northwestern Highlands, the Southwestern Highlands, and the Southeastern Highlands. The fourth sector is significantly larger than any of the other three and is broken into four ecophysiographic areas (see below).

Savage (2002), in his opus on the amphibians and reptiles of Costa Rica, noted “the term cloud forest is often applied to forests that develop at an altitude where the temperature (6 to 10°C) causes water condensation that produces clouds, fog, and rain. This zone may be at any elevation, and its degree of development is related to the amount of water vapor in the air. Cloud forests usually occur where there are prevailing onshore winds that have their air masses uplifted along ocean-facing mountains. In Central America, cloud forests develop principally on the windward slopes affected by the northeast trade winds. In the Holdridge (1967) system, cloud forests are regarded as atmospheric association within bioclimates that, in Central America, usually develop in the lower portion of the lower montane life zone under the influence of strong prevailing winds. During much of the year these forests receive precipitation in the form of light mists. In the drier seasons, much of the time they are enveloped in dense, dripping fog.”

Areas supporting cloud forest in Honduras are generally subject to a Highland Wet climatic regime (Wilson and Meyer 1985). This climatic type is broadly characterized by annual rainfall of >1500 mm and a mean annual temperature of <18°C. The cloud forest regions occurring in the Southern Cordillera generally receive less rainfall than do those in the Northern Cordillera, part of the general effect of the dissipation of moisture in clouds carried by the prevailing winds arising over the Caribbean Sea as they sweep inland.

Climatic data are available for the nuclear zone and the buffer zone of Parque Nacional El Cusuco, a cloud forest reserve in the Sierra de Omoa in northwestern Honduras (Fundación Ecologista “Hector Rodrigo Pastor Fasquelle” 1994). Annual precipitation in the nuclear zone is 2995 mm and in the buffer zone 2580 mm. The rainiest months, in both cases, are October, November, and December, accounting for 45.1% of total rainfall in both zones. The least rainiest months are March, April, and May, when only 12.1% of rainfall occurs in both zones. Monthly temperatures range from 12.9°C in December to 20.2°C in April, with a mean of 16.7°C, in the nuclear zone and from 17.5°C in December to 23.1°C in April, with a mean of 20.6°C, in the buffer zone.

The vegetation of the Honduran cloud forests is referable to two forest formations, as slightly modified from the work of Holdridge (1967), which differ from one another on the basis of the amount of annual precipitation (Wilson and Meyer 1985). The formation characteristic of the cloud forests of the Northern Cordillera is the Lower Montane Wet Forest formation. It is characterized by annual precipitation of >2000 mm. The formation typical of the cloud forests of the Southern Cordillera is the Lower Montane Moist Forest formation. It features an annual precipitation of <2000 mm.

Wilson and McCranie (in preparation a) presented information on the vegetation of Parque Nacional El Cusuco (Lower Montane Wet Forest formation), as follows: “Fundación Ecologista ‘Hector Rodrigo Pastor Fasquelle’ (1994) indicated that this forest formation, called ‘Zona de Vida Bosque Muy Húmedo Montano Bajo Sub-Tropical,’ is characterized by the presence of three strata. The uppermost stratum consists of a closed canopy of trees attaining heights of 35 to 40 m of the following species: Quercus spp.; Podocarpus oleifolius; Clusia massoniana; and Liquidambar styraciflua. The middle stratum is composed of the forgoing species lying in the shade of the taller conspecifics mixed with Persea vesticula and Myrica cerifera. The lowermost stratum is comprised of seedlings of the species in the middle and uppermost strata intermixed with palms such as Chamaedorea costaricana and C. oblongata, as well as Geonoma congesta and a great variety of ferns. Many epiphytic orchids, bromeliads, and mosses are present, as well as lianas and vines.”

Espinal et al. (2001) presented similarly limited data on floristic composition at two sites (at 1570 and 1650 m) in Parque Nacional La Muralla (both in Lower Montane Wet Forest formation), located in the Ocote Highlands of the northwestern portion of the department of Olancho. They stated the following (p. 102): “At the 1570 m site, of 38 species with chest-high diameters of 5 cm or more, seven species were considered most important [based on numerical prevalence]. These species, in order of importance, are: Persea sp. (agua-cate); Calatola mollis (nogál); Quercus sapotaefolia (encinillo); Calophyllum brasiliense (aceite de maría); Elaeagia auriculata (oreja de macho); Quercus skinneri (bel-lota); Chamaedorea neurochlamys (palma pacaya).

At the 1650 m site, a group of 6 species (out of 30) were judged most important, based on frequency of occurrence. These species, in order of importance, are: Calophyllum brasiliense (aceite de maría); Quercus sapotaefolia (encinillo); Persea sp. (aguacate); Quercus skinneri (bellota); Elaeagia auriculata (o[reja] de macho); Alchornea latifolia (amargoso). Five of these six species are of greatest importance as well at the 1570 m site.”

Wilson and McCranie (in preparation b) included data on the floristic makeup of the vegetation of the Lower Montane Moist formation in Parques Nacionales de Celaque and La Tigra, as follows: “The undisturbed forest is composed of three strata. The upper stratum is composed of trees 25 to 30 m in height, principally of the species Quercus skinneri (bellota), Liquidambar styraciflua (liquidámbar), Pinus pseudostrobus (pinabete), P. oocarpa (ocote), and Persea sp. (aguacate sucte). These trees carry a moderate amount of epiphytic mosses, orchids, bromeliads, and aroids. The middle stratum consists of Quercus sp. (curtidor), Q. oleoides (encino), Clethra macrophylla (alámo blanco), Cedrela oaxacensis (cedro), Inga sp. (guama), and various species of laurals. The lower stratum consists of shrubs belonging to the families Compositae, Myrsinaceae, Rubiaceae, Saurauiaceae, and Verbenaceae and the genera Cleyera, Miconia, Piper, Psidium, and Vismia.”

The herpetofauna of the cloud forests of Honduras is known to consist of 122 species (Table 1), including 18 salamanders (14.8% of total), 38 anurans (31.1%), 27 lizards (22.1%), and 39 snakes (32.0%). The salamanders are all members of the family Plethodontidae. The anurans belong to six families, including the Bufonidae (4 species), Centrolenidae (1 species), Hylidae (17 species), Leptodactylidae (12 species), Microhylidae (2 species), and Ranidae (2 species). The lizards are members of four families, the Anguidae (6 species), Phrynosomatidae (1 species), Polychrotidae (18 species), and Scincidae (2 species). The snakes belong to five families, including the Typhlopidae (1 species), Boidae (1 species), Colubridae (31 species), Elapidae (3 species), and Viperidae (3 species).

Table 1 Geographic and ecological distribution, relative abundance, and conservation status of the cloud forest herpetofauna (122 species) of Honduras.

The species distributed in cloud forests fall into three distributional categories, viz., restricted, widespread, and peripheral (Table 1). Restricted species are those whose distribution is limited to a particular cloud forest formation. Widespread species are those that are widespread in distribution in a particular cloud forest formation or both cloud forest formations, as well as, perhaps, outside those forest formations. Finally, peripheral species are those whose distribution is largely peripheral to a particular cloud forest formation.

The Lower Montane Wet Forest formation is inhabited by 26 restricted species (26.5% of the total of 98 in this formation), including six salamanders, seven anurans, ten lizards, and three snakes. Thirty-two species (32.7%) are widespread in this formation, including six salamanders, ten anurans, four lizards, and 12 snakes. Finally, 40 species (40.8%) are peripherally distributed in this formation, including three salamanders, 11 anurans, eight lizards, and 18 snakes.

The Lower Montane Moist Forest formation is home to ten restricted species (22.2% of the total of 45 in this formation), including three salamanders, three anurans, two lizards, and two snakes. Twenty-nine species (64.4%) are widespread in this formation, including one salamander, 11 anurans, five lizards, and 12 snakes. Finally, there are six species (13.3%) peripherally distributed in this formation, including one anuran, three lizards, and two snakes. Notably, there are proportionately more peripheral and widespread species than restricted species in the Lower Montane Wet Forest formation. In the Lower Montane Moist Forest formation, most species are widespread ones, followed by relatively few restricted and peripheral species. The relative prevalence of peripheral species in the Lower Montane Wet Forest formation apparently is due to the grading of this type of cloud forest into highland rain forest (Premontane Wet Forest formation) at elevations usually around 1500 m, whereas the Lower Montane Moist Forest formation grades into upland pine forest (Premontane Moist Forest) typically.

As noted above, 21 species are distributed in both cloud forest formations (Table 1). The largest number of these species (17) are widespread in both formations. Two species are peripheral in distribution in both formations, and, finally, two species are widespread in one formation and peripheral in the other.

The restricted species, as a group, range from 1340 to 2700 m. The mean elevational range for this group of 36 species is 209.6 m. The species that are widespread in at least one of the two cloud forest formations, as a group, range from sea level to 2744 m. The mean elevational range for this group of 44 species is 1000.4 m. The species that occur peripherally in at least one of the two cloud forest formations, as a group, range from sea level to 2200 m. The mean elevational range for this group of 44 species is 1260.3 m (two species are peripheral in one formation and widespread in the other).

• Northern terminus of the range in the United States (or Canada) and southern terminus in South America.
• Northern terminus of the range in the United States and southern terminus in Central America south of the Nicaraguan Depression.
• Northern terminus of the range in the United States and southern terminus in Nuclear Middle America.
• Northern terminus of the range in Mexico north of the Isthmus of Tehuantepec and southern terminus in South America.
• Northern terminus of the range in Mexico north of the Isthmus of Tehuantepec and southern terminus in Central America south of the Nicaraguan Depression.
• Northern terminus of the range in Mexico north of the Isthmus of Tehuantepec and southern terminus in Nuclear Middle America.
• Northern terminus of the range in Nuclear Middle America and southern terminus in South America.
• Northern terminus of the range in Nuclear Middle America and southern terminus in Central America south of the Nicaraguan Depression.
• Restricted to Nuclear Middle America (exclusive of Honduran endemics).
• Endemic to Honduras.

The data on broad distributional patterns in Table 1 are summarized in Table 2. These data indicate that the largest number of species (48 or 39.3% of the total of 122 species) fall into the J category, i.e., that containing the species endemic to Honduras. The next largest category is I, with 35 species (28.7%), containing those species not endemic to Honduras but restricted in distribution to Nuclear Middle America. Together, these two categories contain 68.0% of the cloud forest species. The other eight categories contain from one to ten species and harbor, as a group, 32.0% of the total number. These data again point to the biogeographic and conservation importance of the Honduran cloud forest herpetofauna.

Table 2 Summary of numbers of taxa exhibiting various Broad Patterns of Geographic Distibution (see text for explanation of categories).

In terms of vertical positioning within the primary microhabitats, 49 species (40.2%) were usually found only in arboreal situations, 62 species (50.8%) only in terrestrial situations, and 11 (9.0%) in both. With respect to occurrence in the three major microhabitats (forest proper, streamside, pondside), 76 species (62.3%) were found exclusively in the forest proper, 26 (21.3%) only along streams, eight (6.6%) only around ponds, seven (5.7%) in the forest and along streams, three (2.5%) around ponds and along streams, and two (1.6%) in the forest and around ponds (Table 1).

If the two sets of categories, vertical positioning in primary habitat and microhabitats, are combined, it can be demonstrated that 94 species (77.0%) fall into four groups, as follows (Table 1): 40 terrestrial forest inhabitants (32.8%); 31 arboreal forest inhabitants (25.4%); 12 arboreal streamside inhabitants (9.8%); and 11 terrestrial streamside inhabitants (9.0%). The terrestrial forest inhabitants include four salamanders, four anurans, four lizards, and 28 snakes. The arboreal forest inhabitants are eight salamanders, two anurans, 19 lizards, and two snakes. The arboreal streamside inhabitants are one salamander, nine anurans, and two snakes. The terrestrial streamside inhabitants are ten anurans and one lizard.

Table 3 Distribution of the Honduran cloud forest herpetofauna within eight ecophysiographic areas.

• The numbers of species in these eight areas range from four (Agalta Highlands) to 60 (Northwestern Highlands).
• Significantly more species are known from the Northern Cordillera areas (98 or 80.3% of total) than those in the Southern Cordillera (45 or 36.9%) areas. Only 20 species (16.4%) are distributed in both cordilleras (the Rana berlandieri listed in Table 3 from the Southern Cordillera are considered R. berlandieri x R. forreri hybrids—see McCranie and Wilson 2002).
• The above pattern is seen in each of the major herpetofaunal groupings. Only four salamanders are found in the Southern Cordillera cloud forests, compared to 15 in the Northern Cordillera cloud forests. Only a single species (Bolitoglossa conanti) is distributed in both cordilleras (although the population in the Southern Cordillera likely represents an undescribed species). Fifteen species of anurans occur in the Southern Cordillera cloud forests, as opposed to 28 in the Northern Cordillera forests. Only four species (Plectrohyla guatemalensis, P. matudai, Ptychohyla hypomykter, and Rana maculata; the Rana berlandieri listed in Table 3 from the Southern Cordillera are considered R. berlandieri x R. forreri hybrids—see McCranie and Wilson 2002) occur in both regions. Ten species of lizards are distributed in the Southern Cordillera forests, whereas 22 are in the Northern Cordillera forests. Only five species (Mesaspis moreletii, Sceloporus malachiticus, Norops laeviventris, N. tropidonotus, and Sphenomorphus cherriei) are found in both areas. Finally, 16 species of snakes occupy the Southern Cordillera cloud forests and 33 the Northern Cordillera forests. Ten species (Dryadophis dorsalis, Drymobius chloroticus, Geophis fulvoguttatus, Leptodeira septentrionalis, Ninia espinali, Rhadinaea godmani, R. kinkelini, R. montecristi, Bothriechis thalassinus, and Cerrophidion godmani) are distributed in both areas.
• Most of the 122 species (102 or 83.6%) occur in only one or two cloud forest ecophysiographic areas. The most broadly-distributed species occur in seven ecophysiographic areas (Sceloporus malachiticus) or in five ecophysiographic areas (Plectrohyla guatemalensis, Smilisca baudinii, Rana maculata, and Dryadophis dorsalis). The average area occurrence is 1.6.

A greatest shared species diagram of the eight cloud forest physiographic areas is presented in Figure 2. The areas are abbreviated as follows: Southeastern Highlands - SEH; Southwestern Highlands - SWH; North-Central Highlands -NCH; Yoro Highlands - YH; Ocote Highlands - OH; Agalta Highlands - AH; Northwestern Highlands - NWH; Santa Bárbara Highlands - SBH. The number of species shared between areas ranges from two to 18. In general, the greater the total herpetofaunas of any two compared areas, the greater is the number of species shared.

Figure 2 Greatest shared species diagram of eight cloud forest areas in Honduras.

Generation of Coefficient of Biogeographic Resemblance (CBR) values allows for a more robust analysis of herpetofaunal resemblances. Thus, a matrix of CBR values for the eight ecophysiographic areas is summarized in Table 4, and these values are used to produce a CBR diagram (Fig. 3) indicating highest values for each ecophysiographic area. These values indicate that the herpetofauna of a given ecophysiographic area most closely resembles that of another area occupied by the same forest formation and/or lying in close geographic proximity. For example, the Southeastern Highlands and the Southwestern Highlands are both occupied by the Lower Montane Moist Forest formation and they share 13 species. Also, as an example, the Northwestern Highlands and the Southwestern Highlands are in close geographic proximity and share 18 species. Geographic proximity, however, appears to be the more important determinant of the degree of herpetofaunal resemblance, inasmuch as Figure 3 illustrates a western and southern grouping of areas (NWH, SBH, SWH, and SEH) and a northern and eastern grouping of areas (NCH, OH, YH, and AH). These two groups are connected by a relatively high CBR value between SEH and OH.

Table 4 CBR matrix of herpetofaunal relationships for the eight ecophysiographic areas supporting cloud forest.

Figure 3 Coefficient of Biogeographic Resemblance diagram for the eight cloud forest areas in Honduras.

Averaging all CBR values provides a gauge of herpetofaunal distinctiveness, as follows: SEH (0.23); SWH (0.21); NCH (0.21); YH (0.16); OH (0.25); AH (0.09); NWH (0.20); SBH (0.08). The most distinctive herpetofauna is that of the SBH (average CBR value of 0.08), the least that of the OH (average CBR value of 0.25). The distinctiveness of the SBH herpetofauna is an artifact of being poorly known. The fewer the species known from a given area, the fewer there are to be shared with other areas.

As noted in the Introduction, the herpetofauna of Honduras is being subjected to the same anthropogenic pressures as have been demonstrated to be in effect elsewhere in the tropics. The most substantial pressure is created by habitat loss as a result of deforestation (Wilson et al. 2001; Wilson and McCranie 2003 a and b). Also significant is a threat of unsubstantiated origin (but see Duellman 2001, for a discussion of events elsewhere in the tropics) that is decimating amphibian populations in the country occurring at elevations in excess of 900 m (Wilson and McCranie 1998; McCranie and Wilson 2002), thus conceivably impacting all cloud forest areas.

That these threats are impinging on herpetofaunal populations at 900 m and above is especially poignant, inasmuch as the herpetodiversity of greatest significance is distributed in these regions, especially those supporting cloud forest. This most significant herpetodiversity consists of those species endemic to Honduras and those otherwise restricted to Nuclear Middle America. Of the 334 species now known to constitute the Honduran herpetofauna (including six marine reptiles), 78 are country endemics (23.4% of total) and 47 are Nuclear Middle American-restricted species (14.1%). A greater percentage of the amphibian species fall into these two categories than do the reptilian species. There are 41 amphibian Honduran endemics (35.0% of total of 117 species) and 25 Nuclear Middle American-restricted amphibian species (21.4%), compared to 37 (17.1% of total of 217 species) and 22 (10.1%) such reptilian species, respectively. Thus, a total of 125 species of amphibians and reptiles (37.4%) are either endemic to Honduras or otherwise restricted to Nuclear Middle America.

Of these 125 species, 83 or 66.4% are distributed in cloud forests in Honduras (Table 2). Of the remaining 209 Honduran species not found in cloud forests, only 42 species or 20.1% are Honduran endemics or Nuclear Middle American-restricted. It is obvious that the large majority of the species of greatest biodiversity significance is found in cloud forests.

As indicated above, deforestation is eroding forest resources throughout the country. Wilson and McCranie (2003 a) presented estimates, based on a computer model in E. Wilson and Perlman (2000), suggesting that the current deforestation rate is −2.3%, giving rise to a halving rate of 30.1 years. At this rate, only a half a million hectares of forest will remain in Honduras by the year 2085 and none will remain by the end of the current century.

This trend has been affecting cloud forests in Honduras, just as it has everywhere else in the country, and continues to the present day. It has been abated somewhat by the establishment of biotic reserves in several of the ranges supporting cloud forest (Wilson et al. 2001). This establishment largely has been the result of an effort to secure water supplies for populated areas. As noted by Wilson et al. (2001), however, most of these reserves are incompletely developed, such that deforestation still proceeds in many, if not all of them (e.g., Espinal et al. 2001), as a result of illegal logging and subsistence farming.

It has been demonstrated in recent years that populations of many Honduran amphibians and reptiles are in decline or have disappeared altogether, as part of a global pattern (Duellman 2001). Wilson and McCranie (2003 a) have provided the most recent assessment of this trend for the Honduran herpetofauna. However, their assessment differs somewhat from the one undertaken here. Wilson and McCranie (2003 a) considered the range as a whole for each species when classifying whether a given species had stable populations somewhere in their range. However, a few species may have stable populations at some low elevation localities, but may be extirpated from their known cloud forest localities (e.g., Eleutherodactylus charadra). Thus, the conservation status categories in this paper refer only to cloud forest populations. Table 1 lists the conservation status for each of the 122 species at their known cloud forest localities. These data indicate that 40 species (32.8%) have populations that are in decline, eight species (6.6%) have disappeared altogether from cloud forests, and 16 species (13.1%) are too poorly known to determine their status in Honduran cloud forests. Fifty-eight species (47.5%) appear to have stable populations in at least one cloud forest locality.

When one considers only the two most important components of the Honduran cloud forests (the Honduran endemics and the Nuclear Middle American-restricted species), then 15 of the 48 Honduran endemics (31.3%) have declining populations, six endemics (12.5%) have disappeared, five endemics (10.4%) are too poorly known to determine their status, and 22 endemics (45.8%) appear to have stable populations in at least one cloud forest locality. Of the 35 Nuclear Middle American-restricted species, 20 (57.1%) have declining populations, one (2.9%) has disappeared, two (5.7%) are too poorly known, and 12 (34.3%) appear to have stable populations in at least one cloud forest locality. Thus, about one half of the 83 Honduran endemics or Nuclear Middle American-restricted species have declining populations (35 species or 42.2%) or have disappeared from Honduran cloud forests (seven species or 8.4%). Of the six Honduran endemic species that have disappeared from cloud forests, five are feared extinct. These are shocking statistics, considering the importance of these species not only biologically, but also from conservation and ecotourist standpoints.

From simply a biological standpoint, the systematics of the majority of the 83 cloud forest notables (Honduran endemics and Nuclear Middle American-restricted species) are insufficiently understood to be subjected to cladistic analysis, a requirement for reconstructing their phylogenies, and, beyond this, their biogeographic histories. These species are particularly important in our effort to understand the general patterns of evolution of the herpetofauna and to take that understanding beyond the work done on this subject to date.

From the perspective of conservation biology, we have demonstrated here and elsewhere (Wilson and McCranie 1998, 2003 a and b; Wilson et al. 2001; McCranie and Wilson 2002, in press) that the herpetofauna is anything but the pedestrian compendium alluded to in Lynch and Fugler's (1965, p.15) conclusions when they wrote that, “The anuran fauna seems to be derived from largely widespread species and species with northern affinities.” Quite to the contrary, the work that has been accomplished since Lynch and Fugler published their paper 38 years ago has shown that slightly more than a third (37.4%) of the Honduran herpetofauna is composed of endemics or otherwise Nuclear Middle American-restricted species. Our work in cloud forests has provided the major support for that conclusion.

The economic value of the Honduran cloud forests for ecotourism is only beginning to be calculated. It is stunningly evident to us, however, based on the several decades of our field work in the country, that efforts to develop an ecotourist-generated component to the Honduran economy is likely to be doomed by the uncontrolled human population growth that continues to stymie efforts to conserve the considerable biodiversity of the country.

Over the decades that we have worked on the cloud forest herpetofauna in Honduras, we have become indebted to innumerable individuals and organizations. First and foremost, this work could not have been accomplished without the support of the personnel of the governmental agencies Recursos Naturales Renovables and Corporación Hondureña de Desarrollo Forestal (COHDEFOR), who made the necessary collecting and export permits available to us over the years. In addition, our inestimable friend Mario R. Espinal has acted as our agent in Honduras in assuring that the garnering of these permits went as smoothly as possible, that vehicles were ready when we needed them, and that other myriad details upon which the success of our field work depended were appropriately handled. Finally, it has been our great and continuing pleasure to work with uncounted Hondurans who have made us gringos feel at home in Honduras. Without the unflagging efforts of these friends, we would never have unearthed the secrets of the amphibians and reptiles to which we have devoted our scientific lives.

We also extend our gratitude to Nidia Romer, who graciously translated the Abstract into Spanish for use as the Resumen. We are also grateful to Louis Porras and Jay M. Savage, the reviewers of this paper, for their improvements.