Species conservation with fragmented and endangered populations must be based on a prior and thorough knowledge of the structure and population dynamics. Oophaga lehmanni is a dendrobatid species endemic of Colombia and is restricted to its type locality. This species has a fragmented distribution and is considered as critically endangered mainly due to habitat destruction and overexploitation. Oophaga lehmanni exhibits phenotypic variation in the dorsal color pattern (red and yellow morphs). We reconstructed the overexploitation history that this species has faced in the last 40 years. In addition, we collected genetic and morphological data for the first time in natural populations to describe genetic diversity between and within populations, and to evaluate morphological and genetic differences between red and yellow morphs. Overexploitation data suggest that more than 80.000 (Min=60.047 - Max=102.236) frogs were extracted from the field in the last four decades, probably generating the local extirpation or population decline from the type locality. Genetic data showed reduced genetic diversity. Observed heterozygosity (mean±s.d.=0.599±0.165) is lower than expected (mean±s.d.=0.867± 0.082). We did not find differences in body size and heterozygosity between the two morphs; however, individuals analyzed were assigned to two genetic clusters, which corresponded to the O. lehmanni-yellow and O. lehmanni-red. In addition, F_ST (0.209) and Nei genetic distance (0.18) values indicated genetic differentiation between the two morphs; therefore, red and yellow morphs should be treated as independent management units. This information will help to define appropriate and long-term conservation units, as a useful tool to mitigate the extinction risk of this species.

We obtained genomic DNA from wild populations of O. lehmanni-red from the FCNP (N=54) and of O. lehmanni-yellow from areas near the type locality (N=15). We do not provide precise coordinates given the species' threat levels (Fig. 1A). Additionally, samples from O. lehmanni with yellow and red phenotypes were obtained from specimens kept in captivity at the Cali Zoo (N=15). These animals were confiscated between 2006 and 2009. We collected tissue samples through a noninvasive technique of mouth swabs (Goldberg et al. 2003; Pidancier et al. 2003; Beebee 2008) for all individuals. In the field, samples were stored dry at room temperature. In the laboratory, these were stored at 4ºC until extraction. DNA was extracted and purified using the DNeasy tissue extraction kit (QIAGEN, Valencia, CA), following the manufacturer’s protocol. Extracted DNA samples were diluted to 12ng/µl and used as template in polymerase chain reactions (PCR) for ten di-, tri-, and tetranucleotide microsatellite loci previously designed: Dpum 110, Dpum 14, Dpum 63, Dpum 44 (Wang and Summers 2009) and Oop C11, Oop F1, Oop B9, Oop G5, Oop H5, Oop E3 (Hauswaldt et al. 2009). Microsatellites are suitable to identify genetic differences between individuals and populations, and these have been widely used in conservation genetics (Bruford and Wayne 1993; Jarne and Lagoda 1996). Forward primers for each PCR amplification were labeled with a 5′- fluorescent tag (6-FAM, NED, VIC, or PET) for later visualization. We amplified the loci individually and confirmed the PCR amplification using an agarose gel electrophoresis for all samples. The PCR reactions were performed using this thermal cycling program: following an initial denaturation at 95ºC for 3 min, 35 cycles were run with 95ºC for 30 s, 54ºC for 30s and 72ºC for 4s. This was followed by a 6-min extension at 72ºC and 25 min incubation at 60ºC. This amplification cycling was used for seven markers (Oop B9, Oop G5, Oop H5, Oop E3, Dpum 14, Dpum 63, and Dpum 44). For Dpum 110 annealing temperatures were at 58°C and Oop C11 and Oop F1 amplified at 64°C. For genotyping, we performed a multiplex of three markers. Fragments were sized with LIZ- 500 size standard and then scored genotypes using GeneMapper V4.1 (Applied Biosystems). We repeated the scoring procedure two times on all samples to check that alleles were assigned appropriately. PCR products were run on an ABI3500 Genetic Analyzer (Applied Bio-systems) in the Sequentiation Center at the Universidad de los Andes.

missing values = 0