General Herpetological Collecting is Size-Biased for Five Pacific Lizards
Accurate estimation of a species' size distribution is a key component of characterizing its ecology, evolution, physiology, and demography. We compared the body size distributions of five Pacific lizards (Carlia ailanpalai, Emoia caeruleocauda, Gehyra mutilata, Hemidactylus frenatus, and Lepidodactylus lugubris) from general herpetological collecting (including visual surveys and glue boards) with those from complete censuses obtained by total removal. All species exhibited the same pattern: general herpetological collecting undersampled juveniles and oversampled mid-sized adults. The bias was greatest for the smallest juveniles and was not statistically evident for newly maturing and very large adults. All of the true size distributions of these continuously breeding species were skewed heavily toward juveniles, more so than the detections obtained from general collecting. A strongly skewed size distribution is not well characterized by the mean or maximum, though those are the statistics routinely reported for species' sizes. We found body mass to be distributed more symmetrically than was snout–vent length, providing an additional rationale for collecting and reporting that size measure.Abstract
Empirical cumulative distribution functions for each of the five lizard species; heavy lines represent the census (total removal plots) and the thin lines represent GC (visual surveys and glue boards). The two-sample Kolmogorov–Smirnov test quantified the maximum distance between the empirical distribution functions and was used to test whether the two distributions differed (P-value) in either location or shape.
Size distributions obtained by census (heavy lines) and general collection sampling (thin lines) for five lizard species in the Mariana Islands. The plots are the first derivative of the cumulative distribution plots of Fig. 1 and display the instantaneous slopes of those smoothed functions. The size ranges in which the census line had the larger proportion of sample indicate relative undersampling by general collecting (most juvenile sizes) and the converse indicates relative overampling by general collecting (mostly smaller adults).
Mean relative bias for all species (± SE) in relation to size class (defined in Table 1).
Comparison of measures of central tendencies for SVL and mass distributions for Emoia caeruleocauda. All species had SVL means perched midway down the ascending slope of the size histogram at some distance from the median and the mode. For mass, the mean, median, and mode were more coincident in each species.
Contributor Notes
