Monday Pets: How Anteaters Decide

"But wait," you say. "Anteaters aren't pets!" Well, I didn't think so either. But Salvador Dali had a pet anteater. And that's good enough for me.

The Giant Anteater, Myrmecophaga tridactyla, only eats ants and termites, making it a myrmecophage. (Hey, Alex Wild, now I get what Myrmecos means!) In 1984, a researcher named Kent Redford was interested in the foraging behaviors of the giant anteater, and the relationship between these anteaters and their prey, colonies of highly social insects. So Redford went to Brazil to study a group of anteaters at the Brasilia Zoo, as well as in the wild at Emas National Park.

Since the giant anteater and its evolutionary ancestors have hunted ants and termites for nearly 60 million years, Redford hypothesized that there would be "appreciable co-evolution between anteater feeding preferences and the biology of their prey." More specifically, ants and termites are highly social, consisting of three main castes: the reproductive, the worker, and the soldier castes. The defenses of the soldier castes in termites and ants vary from entirely chemically based, in which they secrete toxic or repellent chemicals, to the fully mechanical, in which they use their mandibles to pierce the skin of the attacker, and occasionally to draw blood. So he hypothesized that the foraging behavior of the anteaters would vary according to the type of defense behavior employed by the soldier castes of the ant and termite colonies on which they fed.

The first part of the study was conducted on three giant anteaters that were kept at the Brasilia Zoo. Two shallow trays were placed in a small cage adjacent to the main cage where the anteaters were kept, which each contained broken up pieces of a termite mound. Each tray, therefore, contained thousands of living and active termites. In a series of two-way forced choice trials, Redford tested the foraging behaviors of the anteaters for eight species of termites: Grigiotermes metoecus, Armitermes euamignathus, Cornitermes cumulans, Cortaritermes silvestri, Nasutitermes, Procorniterrnes araujoi, Velocitermes paucipilis, and Orthognathotermes gibberorum. Each individual anteater was tested with all possible two-way combinations.

The researchers recorded the sequence in which the individuals ate the different species, as well as the number of times the anteater sniffed but did not snack on a particular species. Since it was impossible to quantify the number of termites eaten, they recorded the duration of the feeding session as well, per termite species. The anteaters' foraging behaviors were compared with three termite variables: the size of the termites, their nutritional value, and the type of defensive system that they used (chemical or mechanical).

The first main finding was that the anteaters did not feed on the different species equally. In 24 of the 28 two-way comparisons, there were clear preferences for one termite species over the other, and all three individual anteaters showed the same preferences. If you were wondering, the two tastiest termite species for these anteaters were Cornitermes and Procornitermes.

These findings alone suggest that the anteaters are making explicit foraging decisions; the next question is, on which variables do they base their decisions?

Prey size did not correlate significantly with feeding decisions, so that can be safely ruled out, and neither did nutritional value. What of the defense-related variables? First, the percentage of the colony that comprised the soldier caste was not correlated with feeding decisions, either. Nor was the aggressiveness of the termite soldiers. Despite their clear preferences, Redford was unable to determine the variables that were behind those preferences. (But where's the data on the type of defense used? Chemical versus mechanical? It's coming, I promise.)

The second part of the study was done with wild giant anteaters, in the field. Animals located from a car were then followed on foot, so that the experimenters could observe and record data on feeding sessions. Data was collected on the location and duration of feedings, as well as (where possible) the prey species. In total, data was collected and analyzed from thirty-three hours and thirty-nine minutes of feeding. In all, the anteaters spent 23.1% of their time (across 40 observations and 1487 individual feeding sessions) engaged in feeding behaviors. In all, eight species of termites and six species of ants were identified across the observed feeding sessions.

Again, it was clear that the anteaters displayed clear preferences. Critically, however, the preferences shown by the wild anteaters were different from the preferences shown by their captive brethren. For example, the favorite termite species of the captive anteaters, Cornitermes, accounted for only 12% of the feeding time in the wild and was rated only fourth in preference. And the least favorite genus of the captive anteaters, Velocitermes, was the favorite in the wild.

What can explain the differences in terms of species preference? One possibility is the mound structure that the different species use. In the tests at the zoo, the anteaters never had to break into a termite mound; instead, the crumbled up mound pieces were presented to them. For example, Syntermes, one of the favorites in the zoo, has a mound structure that goes up to 1.5 meters into the ground, allowing the ants to retreat into deeper ground when attacked from above. It is also possible that certain defenses (particularly, the chemical defenses) may have been significantly less effective in the broken-up mounds offered to the captive anteaters.

Additionally, in the wild, the anteaters may feed in such a way as to avoid the soldiers, and face a lower soldier-to-worker ratio than in the samples offered in the zoo. Specifically, at the beginning of an anteater feed, more workers and fewer soldiers are present than is the case after even a few seconds. Previous studies have shown that once the ratio of soldiers-to-workers becomes unfavorable (due to the retreat of workers and the recruitment of soldiers), anteaters will take a break from eating. Converging evidence for this comes from studies comparing the soldier-to-worker ratio in stomach samples.

Another possible explanation for the discrepancy is that, in many cases, multiple species of termites or ants share the same mound. This may mean that certain species that were particularly palatable under the controlled environment at the zoo, were less favorable in the wild due to their proximity to other less desirable or more aggressive species.

Yet another explanation for the discrepancy comes from the possibility that anteaters may be not specialized for ants or termites, per se, but rather on small arthropods that live socially. In Venezuela, anteaters prey mostly on ants, while in Brazil, they feed mostly on termites. The diet differences are probably related to differences in the relative availability of various species of ants and termites in the different regions. Thus, Redford argues:

Anteater preference may be only relative to the social insects available to them and not based on factors intrinsic to particular prey species. This flexibility would be necessary to allow [anteaters] to range through so many different habitats with their different social insect communities, and may account for the lack of agreement between wild and captive preferences reported in this study.

Very likely, each of these possibilities probably plays a role in explaining the observed discrepancies in prey preferences between the captive and wild anteaters.

But what about the larger issue of defense strategy? Chemical versus mechanical soldier defensive tactics? And nutritional value? Well, its not so straightforward, but when combining data from both the captive experiments and the wild observations, there is evidence that nutritional value plays at least some role, at least for Grigiotermes, Orthognathotermes and Armitermes. There is also an effect of defense type: the termites that utilized a mechanical defense were more highly preferred, compared with those that use a chemical defense. It may be that the species (such as Procornitermes and Cornitermes) that rely on using their mandibles to slash or slice tend to be effective defenses for small predators (like mice), but are generally ineffective against larger predators such as anteaters, that pick them up with their long saliva-covered tongues.

Overall, three major variables emerge that seem to govern the feeding decisions of giant anteaters:
(1) the type of soldier-based defense (chemical warfare is way more effective),
(2) the type of mound structure used, and
(3) the nutritional value.

Each of these factors combine in different ways for each species of ant or termite, resulting in different overall defensive success.

One other important finding emerged from this study. Giant anteaters feed on all of the different prey species in a similar way, irrespective of nutritional value, mound structure, or defense system: they feed for short periods of time, across multiple different locations. Previously, it was suggested that this was simply prudent predation - that the anteaters could avoid totally destroying any particular mound and losing a valuable source of food. Instead, these data suggest that the anteaters' predatory patterns emerge by the nature of defensive strategies employed by their prey.

Kent H. Redford (1985). Feeding and food preference in captive and wild Giant anteaters (Myrmecophaga tridactyla) Journal of Zoology, 205, 559-572.