Adult vs tadpole sampling: which perform better for estimation of species richness?

Species richness is a key response variable in ecological studies. However, species richness may be susceptible to sampling biases. Two main methods have been used to survey anurans and to assess species richness: visual and acoustic searching and tadpole dipping net. These methods are usually employed to sampling adult and larvae, respectively. Does the survey based on larvae or adults affect the species richness estimation? We addressed this question by using a meta-analysis. We search for papers in Web of Science, Google Academic and Scielo that sampled both adults and tadpoles to assess richness and diversity. Then, we constructed a matrix of binary data, which were assigned event values (presence) rather than event (absence). The size of the effect and its variation were calculated from Relative Risk, computed on a logarithmic scale. The relationship between the size effect and the moderators of total amphibian richness, number of sample units and year of publication were performed through meta-regression analyzes and used to explore the heterogeneity between the studies. The mean richness of adults was 25.70 ± 7.93 species, while the tadpole richness was 19.45 ± 7.30 species. Eight studies recorded total richness only through adults and none through only tadpoles. The use of methodologies that include adult anurans had a positive effect on species richness (cumulative effect = 0.23, p <0.01). Ten studies have no effect size. Ten studies have a positive effect, while none had a negative effect. The number of analyzed units negatively influenced the effect size of the studies. The results indicate that adult-focused studies can better detect species richness as adults become more exposed and vocalization facilitates detection. Tadpole sampling is probably more effective when adults are explosive breeders. In this case, adults vocalize for a short period and they do not co-occur with tadpoles.