Nearly one-third of amphibian species worldwide are threatened and many are already extinct (Stuart et al., Reference Stuart, Chanson, Cox, Young, Rodrigues, Fischman and Waller2004). A large proportion of these threatened amphibians inhabit tropical America, where habitat degradation and loss, infectious diseases and climate change are the major threats (Stuart et al., Reference Stuart, Chanson, Cox, Young, Rodrigues, Fischman and Waller2004; Lips et al., Reference Lips, Burrowes, Mendelson and Parra-Olea2005). While most attention has focused on the interactions between disease and climate change because of a link with dramatic declines (Lips et al., Reference Lips, Brem, Brenes, Reeve, Alford and Voyles2006; Pounds et al., Reference Pounds, Bustamante, Coloma, Consuegra, Fogden and Foster2006; Lawrence, Reference Lawrence2008), little research and few conservation efforts have focused on the effects of habitat change on tropical amphibians (Becker et al., Reference Becker, Fonseca, Haddad, Batista and Prado2007; Gardner et al., Reference Gardner, Barlow and Peres2007). About 50% of the glassfrog species (family Centrolenidae) are declining and 40% are threatened (Bustamante et al., Reference Bustamante, Ron and Coloma2005; IUCN, 2008). Diseases and global warming have been linked to some of these declines (Pounds et al., Reference Pounds, Fogden and Campbell1999; Lips et al., Reference Lips, Brem, Brenes, Reeve, Alford and Voyles2006) but the causes of most remain poorly understood (IUCN, 2008).

Cochranella mache is a recently described glassfrog endemic to Ecuador, categorized as Endangered on the IUCN Red List based on the limited information available from its original description (Guayasamin & Bonaccorso, Reference Guayasamin and Bonnacorso2004; IUCN, 2008). To provide further information on the species we undertook visual encounter surveys and standardized visual transect sampling (≥ 25 person-hours per site) across 14 localities in western Ecuador over 2004–2007 (Appendix 1) and examined museum collections (Appendix 2).

The species was previously known only from its type-locality, Bilsa (Site 6; Table 1, Fig. 1; Guayasamin & Bonaccorso, Reference Guayasamin and Bonnacorso2004; Cisneros-Heredia & McDiarmid, Reference Cisneros-Heredia and McDiarmid2007). We found C. mache in three new localities and a museum specimen from one additional locality (Site 15): one male at Monte Saino (Site 5) after c. 96 person-hours of searching, one female at Canandé (Site 4) after c. 40 person-hours, and a male near Quinindé (Site 7) after c. 25 person-hours (Table 1). At Bilsa a male was found previously after 85 person-hours (G. Vigle, pers. comm.) and two males after 40 person-hours (Guayasamin & Bonaccorso, Reference Guayasamin and Bonnacorso2004) but we did not record any individuals in 40 person-hours of searching in the same location in December 2006.

Fig. 1 Ecuador, showing the 14 localities surveyed for Cochranella mache in western Ecuador (Appendix 1) and the four where the species was found (shaded squares; Table 1): 4, Reserva Canandé; 5, Monte Saino, Punta Galeras area; 6, Reserva Bilsa (type locality); 7, Quinindé. Site 15, Río La Carolina, is the location of a museum specimen.

Table 1 Details of all known records of Cochranella mache (see numbered sites in Fig. 1). All are in the Province of Esmeraldas, Ecuador.

* See Appendix 2

C. mache is known to deposit egg clutches on the top of leaves over well-oxygenated streams and its tadpoles fall into the water and burrow in debris (Guayasamin & Bonaccorso, Reference Guayasamin and Bonnacorso2004; Cisneros-Heredia & McDiarmid, Reference Cisneros-Heredia and McDiarmid2007; Cisneros-Heredia et al., Reference Cisneros-Heredia, Delia, Yánez-Muñoz and Ortega-Andrade2008) but otherwise little is known of its biology. All known records are from riverine areas in primary and old secondary forests. It has not been found in recent secondary forest, small isolated forest patches, or agricultural/suburban areas. All known localities (Table 1) are restricted to the Cordillera Mache-Chindul and surrounding areas in the Province of Esmeraldas. This is an isolated mountain range in the northernmost portion of Cordillera de la Costa, the mountain chain that runs parallel and independently from the Andes along the Pacific coast of Ecuador. The species is restricted to altitudes of 100–640 m. All records are from seasonal evergreen forests, a moist forest formation endemic to the West Ecuadorian region (the biogeographical area between the humid non-seasonal Chocó and xeric highly-seasonal Tumbesian regions; Cisneros-Heredia, Reference Cisneros-Heredia2006, Reference Cisneros-Heredia2007). The northernmost record is consistent with the distribution patterns of most endemic species of the West Ecuadorian region (Cisneros-Heredia, Reference Cisneros-Heredia2006, Reference Cisneros-Heredia2007). Based on the distribution of seasonal evergreen forests along the Cordillera Mache-Chindul, C. mache's range may extend to the south, reaching the Mache-Chebe-Tabiaza Rivers, the southernmost limit of the Cordillera.

Deforestation in western Ecuador is extensive (Dodson & Gentry, Reference Dodson and Gentry1991) and < 100 km² of primary or old secondary forests remain on the Cordillera (18–20% of the original forested area). The remnant forests are highly fragmented, with the largest single block < 16–18 km², and deforestation rates are c. 3–5% per year (Dodson & Gentry, Reference Dodson and Gentry1991; Mudd, Reference Mudd1991; Paredes & Tapuyo, Reference Paredes and Tapuyo1998; Conservation International, 2001; Kvist et al., Reference Kvist, Skog, Clark and Dunn2004; and remote-sensing analyses by DFC-H based on Hansen et al., Reference Hansen, DeFries, Townshend, Carroll, Dimiceli and Sohlberg2006 and Mulligan, Reference Mulligan2007). Although the range of C. mache is partially within the Mache-Chindul Ecological Reserve, most conservation measurements are ineffective because of institutional and funding restrictions and a lack of law enforcement. Some of the larger fragments are preserved by private organizations but many remain unprotected. Habitat degradation is mainly caused by unsustainable timber extraction, uncontrolled expansion of the agricultural frontier, and replacement by non-native plantations.

The categorization of C. mache as Endangered (IUCN, 2008) underestimates its threatened status. Our data suggest that, inferred from the destruction of its habitat, C. mache has suffered a large reduction in its range since the mid 1990s. The current known range is small and even if it extends across the entire Cordillera Mache-Chindul it will be < 100 km². We recommend that C. mache be categorized as Critically Endangered based on criteria A2c, B1ab(i,ii,iii,iv) (IUCN, 2001) as its range is extensively fragmented and continued declines of its extent, habitat quality and number of localities and subpopulations are inferred.

While many lowland glassfrogs are conspicuous members of riverine communities, C. mache is scarce even in well-preserved areas. It may be naturally rare or not easily detected because of sampling bias (common survey methods fail to record canopy specialists, D.F. Cisneros-Heredia, pers. obs.), or it may have been affected by gradual population declines such as those reported by Whitfield et al. (Reference Whitfield, Bell, Philippi, Sasa, Bolaños and Chaves2007). Lack of long-term data hinders drawing further conclusions but data available for two glassfrogs sympatric with C. mache, Centrolene prosoblepon and Hyalinobatrachium fleischmanni, suggest that lowland glassfrog populations in western Ecuador may have suffered gradual population declines in the past 3 decades. C. prosoblepon and H. fleischmanni were the most abundant glassfrogs in surveys by KU and USNM (see Appendix 2 for museum abbreviations) in the late 1970s and early 1980s, and also in surveys by us, DHMECN and QCAZ in the 2000s. Combined relative abundances in the earlier surveys were 0.4–2.0 per person-hour in three localities, whereas they were 0.1–0.4 per person-hour in recent surveys in the same localities (Bustamante et al., Reference Bustamante, Ron and Coloma2005; R.W. McDiarmid & K. Miyata, unpubl. data; Cisneros-Heredia et al., unpubl. data; DHMECN, unpubl. data; QCAZ, unpubl. data).

Deforestation in lowland areas has been found to modify micro- and meso-scale climate through changes in albedo, evapotranspiration, roughness, cloudiness, rainfall and seasonality patterns (Lawton et al., Reference Lawton, Nair, Pielke and Welch2001; Durieux et al., Reference Durieux, Toledo Machado and Laurent2003; Ray et al., Reference Ray, Welch, Lawton and Nair2006; Pielke et al., Reference Pielke, Adegoke, Beltrán-Przekurat, Hiemstra, Lin and Nair2007). As > 70% of forests across western Ecuador have been felled (Dodson & Gentry, Reference Dodson and Gentry1991; Sierra, Reference Sierra1999; Reference Kvist, Skog, Clark and DunnKvist, 2004) this may have induced changes in local climate patterns of nearby well-preserved areas and thus affected amphibians. Further studies are required to test this hypothesis. However, most areas with rich amphibian diversity are undergoing high rates of habitat degradation (Gallant et al., Reference Gallant, Klaver, Casper and Lannoo2007), the effects of which may be as deleterious as the extirpations caused by disease and global warming. In situ conservation actions are urgently needed and should include reinforcement of existing protected areas, establishment of new ones, and development of mitigation strategies, including habitat restoration and creation of incentives to foster conservation. Future research on the distribution, habitat preferences, population ecology, home ranges and dispersal capacity of glassfrogs is required, along with knowledge of the impacts of edge effects, habitat modification, and micro- and meso-scale climate changes. Researchers at DHMECN have begun these studies for West Ecuadorian endemics.