James I. Garnham

Estimates of sex ratio require the incorporation of unequal catchability between sexes

Abstract Context. Estimates of the sex ratio of a population are a common summary statistic used for ecological studies and conservation planning. However, methods to determine the sex ratio often ignore capture probability, which can lead to a perceived bias in the sex ratio when the sexes are detected at different rates. Aims. To illustrate the bias from conventional count-based analysis methods for determining sex ratio by comparison with analytical methods that include capture probability. Methods. Closed-population mark–recapture analysis was used to determine the population size of each sex within a population of green and golden bell frogs (Litoria aurea). This was then compared with the traditional count-based methods of estimating sex ratio to determine the effect of incorporating capture probability on the sex ratio estimate. Key results. More males than females were detected during surveys, producing a male-biased sex ratio when there was no incorporation of capture probability. Mark–recapture results indicated a similar population size between the two sexes, suggesting that the sex ratio is closer to even. Conclusions. Methods to estimate sex ratio that incorporate capture probability can significantly reduce the bias obtained from count data. Implications. We suggest that population studies must incorporate capture probability to determine the sex ratio of a population.

Evaluating monitoring methods to guide adaptive management of a threatened amphibian (Litoria aurea)

Prompt detection of declines in abundance or distribution of populations is critical when managing threatened species that have high population turnover. Population monitoring programs provide the tools necessary to identify and detect decreases in abundance that will threaten the persistence of key populations and should occur in an adaptive management framework which designs monitoring to maximize detection and minimize effort. We monitored a population of Litoria aurea at Sydney Olympic Park over 5 years using mark–recapture, capture encounter, noncapture encounter, auditory, tadpole trapping, and dip-net surveys. The methods differed in the cost, time, and ability to detect changes in the population. Only capture encounter surveys were able to simultaneously detect a decline in the occupancy, relative abundance, and recruitment of frogs during the surveys. The relative abundance of L. aurea during encounter surveys correlated with the population size obtained from mark–recapture surveys, and the methods were therefore useful for detecting a change in the population. Tadpole trapping and auditory surveys did not predict overall abundance and were therefore not useful in detecting declines. Monitoring regimes should determine optimal survey times to identify periods where populations have the highest detectability. Once this has been achieved, capture encounter surveys provide a cost-effective method of effectively monitoring trends in occupancy, changes in relative abundance, and detecting recruitment in populations.

Low disease-causing threshold in a frog species susceptible to chytridiomycosis

A simple diagnosis of the presence or absence of an infection is an uninformative metric when individuals differ considerably in their tolerance to different infection loads or resistance to rates of disease progression. Models that incorporate the relationship between the progression of the infection with the potential alternate outcomes provide a far more powerful predictive tool than diagnosis alone. The global decline of amphibians has been amplified by Batrachochytrium dendrobatidis, a pathogen that can cause the fatal disease chytridiomycosis. We measured the infection load and observed signs of disease in Litoria aurea Receiver operating characteristic curves were used to quantify the dissimilarity between the infection loads of L. aurea that showed signs associated with chytridiomycosis and those that did not. Litoria aurea had a 78% probability of developing chytridiomycosis past a threshold of 68 zoospore equivalents (ZE) per swab and chytridiomycosis occurred within a variable range of 0.5-490 ZE. Studies should incorporate a species-specific threshold as a predictor of chytridiomycosis, rather than a binary diagnosis. Measures of susceptibility to chytridiomycosis must account not only for the ability of B. dendrobatidis to increase its abundance on the skin of amphibians but also to determine how each species tolerates these infection loads.