Are there exceptions to the “foxes scatter cache rule” and, if so, what are the benefits of scatter caching?


Sometimes foxes put several items in a single cache (larder caching), but this is rare and tends to be associated with a seasonal glut of prey. Spreading your food around puts more pressure on a fox’s memory, but makes it less likely they’ll lose all their cached food to thieves.

The Details

If you look hard enough, there are exceptions to pretty much any “rule” in nature – it’s what makes biological systems so frustrating and, at the same time, so alluring. Indeed, not all fox species even to cache food; the Blandford’s fox (Vulpus cana), for example. Nonetheless, while scatter caching seems more common than larder caching (or larder hoarding) both behaviours have been observed in Red foxes (Vulpes vulpes).

The role of caching

First, a little background. Caching seems to be a flexible behavioural response to a surfeit of food and we tend to see it in predators that rely on prey whose abundance fluctuates. J David Henry observed foxes as young as six weeks old caching food, suggesting some genetic predisposition to hiding leftovers. Conventionally, it was believed that larder hoarding was more efficient for territorial species, the larder would be clumped and located within the area of greatest activity, while non-territorial animals would scatter cache because scatter caches tend not to be defended. Foxes, however, seem to employ scatter caching as the ‘norm’ and larder caching as the exception, despite being territorial.

A "cross phase" Red fox with an Arctic ground squirrel in Denali National Park. Photographer Gregory Smith described how the fox caught three squirrels over the course of 10 minutes, caching each one before returning to hunt. - Credit: Gregory "Slobirdr" Smith

Foxes are, as a function of a small stomach and reduction in skeletal mass, much lighter than canids of similar physical dimensions. The Irish terrier and English foxhound, for example, have similar physical dimensions to Red foxes, but are nearly twice as heavy. Studies on the stomach capacities of foxes have shown that they have proportionally smaller stomach sizes compared to many other canids. For example, a Red fox can consume a maximum of about 10% of its bodyweight in a single sitting, while a Grey wolf (Canis lupis) may consume as much as 20%. Thus, the average fox can consume only about 500g (1.1 lbs.) of meat in a single sitting and any excess food is carried away and buried for later use.

Choosing a cache

The type of cache foxes use, as well as the extent to which they attempt to conceal the cache, varies between individuals, improves with age and is related to both the availability of food and its value to the fox. In their book Urban Foxes, Steve Harris and Phil Baker note that the urban foxes of Bristol would often cache their quarry rather haphazardly, leaving feathers and wings sticking out from the ground. Conversely, in his 1986 book, David Henry reported that, of the hundreds of fox cachings he witnessed in the boreal forests of Canada, each site was carefully chosen and the caching process meticulously implemented – one fox was observed to walk backwards away from his cache hole, carefully erasing his footprints from the snow as he went! Indeed, it appears that the care with which food remains are cached reflects the fox’s preferences.

Foxes tend to cache high-value food items, such as meat, more carefully than they do lower value items such as fruit, vegetables and cleaned bones. Similarly, in a 1976 paper to the German journal Zeitschrift fur Psychologie, David MacDonald reported foxes caching preferred prey such as field voles (Microtus agrestis) more consistently than less-preferred prey, like bank voles (Myodes glareolus).

A Red fox with a small rodent (possibly a vole). The value of the food item to the fox is significant in determining whether it is cached. - Credit: Kelly Colgan Azar

Most detailed observations on fox caching behaviour have revealed a tendency to scatter cache their leftovers. Economically, this seems to make good sense: don’t put all your eggs in one basket (or all your food in one hole). During his many hours of fox observation in Canada, David Henry undertook a series of impressively ingenious experiments to assess the benefits of scatter caching. Henry wandered around the forests, burying small amounts of tinned dog food; first in a scattered pattern and then in a larder caching fashion. He made detailed maps and notes of where each hoard was buried and returned to each a short time later to see if any had been raided.

Henry found that when he larder cached his meat, an average of six caches (out of his total of 15) remained, while seven of the scatter cached hoards (again out of 15) were still buried. These results didn’t suggest much until he calculated the standard deviation, a statistical indication of how much a set of numbers vary from the mean. It transpired that the standard deviation of the larder hoards was much higher than it was for the scatter caches (6 + 4.39 compared with 7 + 1.67). In other words, scatter caching doesn’t reduce the average number of caches that are likely to be discovered; instead it makes these inevitable losses more regular, more uniform and ultimately more predictable. Ergo, it would seem from Henry’s results that scatter caching increases the chances that at least a portion of the caches will still be there when the fox needs it.

So, Henry’s findings provide some insight into why foxes may be more prone to scatter caching leftover food than larder hoarding it. Nonetheless, larder caching does still occur. J Sande observed a Red fox in Sweden repeatedly putting prey into the same hole and examination of the cache found it to contain a hare, ten field mice and a grouse. Similar observations of both Red and Arctic foxes on Baccalieu Island off Newfoundland by Bohdan Sklepkovych, currently at the University of Stockholm in Sweden, have shown larder-caching tendencies in both species. In a 1994 paper, Sklepkovych reported that seabirds, especially Leach’s storm petrels (Oceanodroma leucorhoa), captured and hoarded prior to the onset of winter formed the largest proportion of the Red fox’s winter diet. Indeed, a subsequent study reported that storm petrels made-up a minimum of 53% to 63% of fresh scat collected in February, despite the birds’ absence from the island between November and April.

Foxes will often urinate on and around caches and this may serve as a kind of "book-keeping system" helping the fox know the cache has been excavated even when it can still smell food. - Credit: Tony McLean

A 1996 paper to the journal Arctic, Sklepkovych and his colleague William Montevecchi, at the Memorial State University in Canada, looked at food hoarding behaviour of Red and Arctic foxes on Baccalieu. The biologists concluded that larder hoarding was associated with a superabundance of food and that it appeared to represent a flexible response to changing environmental conditions. In one example cited by Sklepkovych and Montevecchi, a larder discovered in 1985 had 16 holes and 385 (97%) of the carcasses inside were Leach’s storm petrels. They observed both larder and scatter caches from their foxes, witnessing two distinct ‘types’ of larder hoard: (1) one or more holes naturally occurring or actively dug beneath boulders; and (2) several holes dug vertically into the soil to a depth of about one metre (3ft). Interestingly, no attempt was made to conceal the hoard at ground level, suggesting a superabundance of prey.

Sklepkovych and Montevecchi also found that decomposition in the caches was often reduced because the microclimate in the larder was noticeably cooler than the ambient. Indeed, three freshly-killed petrels placed under a rock in July 1986 showed little sign of dehydration and were all well preserved when checked some four months later.

Returning to caches

Tracks in the snow leading to and from a utilised cache (top) and a fox in the process of retrieving stored food. Photos originally published in the journal Arctic in 1996 (10.14430/arctic1199) and reproduced here with permission of the author. - Credit: Bohdan Olaf Sklepkovych

We know that both larder and scatter caching are known in Red foxes, and that scatter caching serves to make losses more uniform, but what is the likelihood of a fox ever returning to its cache? Much of the data available suggest that foxes have a good spatial memory, aided by the use of urine and faeces to mark cache sites.

In their 1996 paper, Sklepkovych and Montevecchi found that, of 67 petrels caught by two foxes, 22 (33%) were eaten there and then, while all but two of the remaining birds were partially eaten before being cached in one of 45 scatter hoards. Of those cached, 12 (28%) had been recovered within 24 hours, 17 (41%) within two days, 22 (53%) within one week, 26 (62%) within two weeks and 1 within three weeks. This suggests that at least 62% of the caches were recovered during the study period and, whether or not foxes recovered the remaining caches at a later date is unknown. Oxford University’s David Macdonald and three colleagues reported higher recovery rates in a 1994 paper to the Journal of Mammalogy. During their study in Dalyan, on the coast of Southwest Turkey, Macdonald and his co-workers found that Red foxes retrieved 94% of scatter cached loggerhead turtle (Caretta caretta) eggs (no larder hoarding was observed), 80% on the subsequent night.

Fox cache sites may be marked with either urine or faeces, apparently depending on their contents. In his book Red Fox: The Catlike Canid, and in a 1977 paper to the journal Behaviour, Henry notes that where foxes cache more durable items, like bones, they mark the cache with more durable scent (i.e. faeces). Generally, however, foxes will only mark caches after they’ve been emptied and, as such, he considered this urine/faeces marking to be a sort of ‘bookkeeping’ method. Henry observed that the foxes would urinate on the cache after the food had been recovered and consumed (or moved elsewhere), possibly to prevent wasting time and energy looking for food that is no longer there. If the smell of food was sufficiently strong, however, the fox would ignore the smell of urine and excavate the ground anyway.

In conclusion, although foxes scatter cache with greatest frequency, larder caching has been observed. Scatter caching appears to serve as a method for regulating losses of hoarded food, making them more regular and uniform. Where larder hoarding is observed, it appears to be a flexible behavioural response to a superabundance of available prey.