A total of 33 parrots of four different species completed this study. The study participants included eight African grey parrots, Psittacus erithacus, and three macaw species. The macaws included nine great green macaws (also known as “Buffon’s macaw” in American aviculture), Ara ambiguus; eight blue-throated macaws (also known as the “Canindé macaw” in American aviculture), Ara glaucogularis

; and eight blue-headed macaws, Primolius couloni. All study parrots were captive-bred and hand-raised, and they lived in eight aviaries at the Max-Planck Comparative Cognition Research Station at Loro Parque Fundación.

A total of 33 parrots of four different species completed this study. The study participants included eight African grey parrots, Psittacus erithacus

, and three macaw species. The macaws included nine great green macaws (also known as “Buffon’s macaw” in American aviculture), Ara ambiguus; eight blue-throated macaws (also known as the “Canindé macaw” in American aviculture), Ara glaucogularis; and eight blue-headed macaws, Primolius couloni. All study parrots were captive-bred and hand-raised, and they lived in eight aviaries at the Max-Planck Comparative Cognition Research Station at

Loro Parque Fundación.

Using a “Marshmallow Test” that was specially modified for parrots, Dr. Krasheninnikova and her collaborators taught the study subjects the value of three different tokens, each associated with one of three differently valued treats: a low-value reward (a sunflower seed), a more expensive medium-value treat (a dry piece of corn) or a high-value food reward (a piece of walnut). The researchers conducted a series of tests to ascertain whether the parrots had learned to associate different tokens with different types of food and determined whether they would select a token over an immediately available food of lesser value. If the parrots were capable of deciding economically by maximising their payoffs, Dr. Krasheninnikova and her collaborators predicted they would choose the most profitable option (food or token), whilst avoiding unnecessary effort.

The researchers then tested the parrots’ economic sense and decision-making abilities by presenting them with an either-or choice between an immediately available food reward or one of the three types of tokens. The parrots were tested under a total of six conditions: under three conditions, the subject could increase its payoff by choosing the token over the immediately available food; under one condition, the subject would get a lower payoff by choosing the token rather than the immediately available food; and under two conditions, there was no difference between immediate and future gain, and so the most economical decision was to consume the food immediately.

The researchers found that the different species performed differently and proposed this may stem from their different feeding ecologies. For example, the two large macaw species, the great green macaws and the blue-throated macaws, made the most economically beneficial decisions significantly more often than would be expected by chance in the tests. These two macaws are food specialists: great green macaws rely heavily on mountain almond trees whilst blue-throated macaws specialize on motaçu palm fruits (ref and ref), both of which are ephemeral food resources.

In contrast, African grey parrots are generalist feeders that consume a variety of seeds and nuts (ref). The other macaw species, the blue-headed macaw, has relatively unknown food preferences in the wild but are thought to be nomadic, which suggests no particular dependance upon any one food resource, and thus, they too, might be generalists. In view of these results, Dr. Krasheninnikova and her collaborators proposed that food specialists might have a greater need for optimal decision-making than food generalists, who, being generalists, probably opportunistically accept any foods they encounter.