Quantifying errors associated with using prey skeletal structures from fecal samples to determine the diet of Steller's sea lion (<em>Eumetopias jubatus</em>)

We examined the digestion and passage times of bones and other hard parts from pollock, herring, salmon, and sandlance recovered from two juvenile captive Steller's sea lions (Eumetopias jubatus) subjected to varying activity levels. Key bones that could be identified to species were distributed over an average of 3.2 scats (range 1-6) following a single meal, with pollock remains Occurring in significantly more scats than other species. Relying on otoliths alone to determine the presence of prey resulted in significantly fewer prey being identified than if other structures were also used (such as vertebrae, jaw bones, and teeth), particularly for salmon. Using either technique, there were significant differences in the likelihood that bones would be recovered from the series of scats produced following a meal, with pollock recovery exceeding herring (by three-fold) and sandlance (by eight-fold). Differences between species were reduced when recovery was calculated on a per scat basis rather than over multiple scats. Active animals passed greater numbers of bones, but the overall effect on prey recovery estimates was not significant. Defecation times of prey structures from a meal were variable and ranged from an initial 2-56 h to a final 28-148 h. The time interval to pass 95% of recovered structures varied by a factor of two among prey species, and was highest for pollock due to retention beyond 65 h.