Experimental evidence demonstrating how freeze-thaw patterns affect spoilage of perishable cached food

For the small number of temperate and boreal species that cache perishable food, previous research suggests that increasing freeze-thaw events can have a negative impact on fitness by degrading the quality of cached food. However, there is no experimental evidence that directly links freeze-thaw events to cache quality. To examine how the timing, frequency, duration, and intensity of freeze-thaw events influenced cached food mass loss, a proxy for caloric content, we conducted a series of month-long laboratory experiments by placing simulated caches (raw chicken placed between two pieces of black spruce Picea mariana bark) in programmable freezers. Freeze-thaw treatments were modelled after weather data from Algonquin Provincial Park, Ontario, where a population of Canada jays (Perisoreus canadensis), a species that caches perishable food for overwinter survival and to support late-winter breeding, has declined by >  70% since the 1980s. First, we found no evidence that an increased frequency of freeze-thaw events influenced mass loss, suggesting that microstructural damage caused by crystal reformation does not significantly influence cache quality. Instead, our experimental results demonstrated that mass loss was positively influenced by longer individual thaws, which likely reflects increased microbial growth, oxidation, and progressive drip loss. We also found that caches lost more weight when subjected to early freeze-thaw events compared to late freeze-thaw events. Finally, we show that milder freezes led to less mass loss and, unexpectedly, warmer than average thaws post-freeze also led to less mass loss. Our results suggest that longer thaw periods post-freezing and milder freezes cause or lead to significantly increased spoilage of perishable cached food. All of these temperature-related conditions are closely associated with long-term changes in climate and, thus, the effects on cache degradation reported in these experiments should be applicable to species caching perishable food in the wild.