[Experimental studies of the capacity of ants for the addition and subtraction of small numbers].

In the laboratory experiments Formica polyctena ants had to transmit the information about the position of a "branch" situated on a long "trunk". Each branch ended in an empty trough, save one filled with syrup. The food was placed on different branches with different frequencies; on the preliminary chosen special branches the troughs were filled substantially more frequently than on the others. For example, in 1993 we chose two special branches, nos. 10 and 20, on which the food was placed with a probability of 1/3, while for any of the other 28 branches the probability was 1/84. When the ants had learnt this, they changed the way of information transmission about the coordinates of the branch containing food and spent less time to transmit the messages about branches nos. 10 or 20. Careful analysis of time required for transmitting messages about trough coordinates suggests that ants "present" the number of a branch in a way similar to that for the Roman figures, and the "special" numbers (10 and 20 in the case) have the same significance as the basic Roman figures V, X, L, etc. Thus, the ants seem to be able, first, to put the duration of a message into agreement with its frequency and, second, to add and subtract small numbers, as people do when use the Roman figures.