Data from: Using survival models to estimate invertebrate prey identification times in a generalist stream fish

Estimates of predator feeding rates are important for understanding
trophic dynamics. One common method for quantifying feeding rates in
fishes combine mass-based diet data with gastric evacuation times to
estimate prey mass consumed per predator. An alternative approach is to
estimate rates of prey individuals consumed using prey identification
times − the time period over which prey remain identifiable in a
predator’s stomach. One challenge with the analysis of prey identification
times, however, is that the response variable is likely to be censored
because the ‘true’ prey identification time cannot be observed directly.
Here, we applied survival analysis that can incorporate censored data to
estimate the effects of predator body size, water temperature, and prey
characteristics (type, count and body size) on identification times in
reticulate sculpin (Cottus perplexus). We focused on seven types of prey
that are common in this generalist predator’s diet: mayflies
(Ephemeroptera), caddisflies (Trichoptera), stoneflies (Plecoptera), flies
(Diptera), beetles (Coleoptera), worms (Annelida), and sculpin eggs. An
information-theoretic model comparison approach indicated that an
accelerated failure-time Weibull model with all five covariates provided
the best relative fit to the full dataset. Prey type had a strong effect
on prey identification times, with annelid worms having the shortest times
(< 1 hr) and caddisflies having the longest times (> 15
hrs). Water temperature decreased prey identification times (7.5% per 1°C
increase), whereas prey count (i.e. meal size) increased prey
identification times (15.5% per additional prey item). Sculpin body size
had a weak negative effect on prey identification times (0.04% per 1 mm
increase). Body sizes of some taxa, including mayflies, caddisflies, and
stoneflies, increased prey identification times, leading to an interaction
between prey type and prey size. Our study highlights the utility of
survival analysis to quantify variation in prey identification times in
the diets of generalist predators.