We can only speculate about the benefits of this ability. (a) What is the benefit for humans and animals to be able to “subitize” object numbers up to four precisely? This seemingly innate ability of humans, of non-human vertebrates and of an insect, the honeybee, to recognize and memorize up to four objects correctly without sequential counting raises the following questions: In summary, object numbers of more than three or four consequently have the meaning of “many” for honeybees and for other, rigorously examined animals. However, under rigorous testing conditions animals “subitize,” i.e., recognize and memorize object numbers without the ability of real counting. Moreover, many such studies, especially with “counting” animals, do not differentiate strictly enough between genuine counting, “subitizing” and pattern recognition. 1 Partly those results were obtained under less stringent conditions because the very rigorous “delayed match-to-sample” setup had not been used or could not be used for all species. Similar numerical capacities have previously been observed for a number of vertebrates. We were especially careful to avoid any possibility of pattern recognition which would render our results meaningless.

Thus, up to 3 objects were memorized for about 5 sec during the flight from the entrance to the decision chamber of our setup, but 4 objects were recognized with less precision and with some difficulties, 1 indicating that the borderline of numerical capacity in this species is between 3 and 4 objects. Even under variable and complex conditions, the insects were able to choose the correct object numbers in more than 80% of the decisions after only 4 rounds of training. Using a “delayed match-to-sample” setup in a Y-maze we determined the numerical capacity of honeybees.

3-d setup square by jevons