Comparative neurobiologists have provided ample evidence that in vertebrates small animals have proportionally larger brains: in a double-logarithmic plot of brain weight versus body weight all data points conform quite closely to a straight line with a slope of less than one. Hence vertebrate brains scale allometrically, rather than isometrically, with body size. Here we extend the phylogenetic scope of such studies and the size range of the brains under investigation to the insects, especially ants. We show that the principle of (negative) allometry applies as well, but that ants have considerably smaller brains than any ant-sized vertebrate would have, and that this result holds even if the relatively higher exoskeleton weights of ants (as compared to endoskeleton weights of mammals) are taken into account. Finally, interspecific comparisons within one genus of ants, Cataglyphis, show that species exhibiting small colony sizes (of a few hundred individuals) have significantly smaller brains than species in which colonies are composed of several thousand individuals.

Wood ants Formica japonica can steer their outbound (foraging) and inbound (homing) courses without using celestial compass information, by relying exclusively on landmark cues. This is shown by training ants to run back and forth between the nest and an artificial feeder, and later displacing the trained ants either from the nest (when starting their foraging runs: outbound full-vector ants) or from the feeder (when starting their home runs: inbound full-vector ants) to various nearby release sites. In addition, ants that have already completed their foraging and homing runs are displaced after arrival either at the feeder (outbound zero-vector ants) or at the nest (inbound zero-vector ants), respectively, to the very same release sites. Upon release, the full-vector ants steer their straight courses by referring to panoramic landmark cues, while the zero-vector ants presented with the very same visual scenery immediately search for local landmark cues defining their final goal. Hence, it depends on the context, in this case on the state of the forager's round-trip cycle, what visual cues are picked out from a given set of landmarks and used for navigation.

A new training and testing paradigm for walking sheep blowflies, Lucilia cuprina, is described. A fly is trained by presenting it with a droplet of sugar solution on a patch of coloured paper. After having consumed the sugar droplet, the fly starts a systematic search. While searching, it is confronted with an array of colour marks consisting of four colours displayed on the test cardboard (Fig. 1). Colours used for training and test include blue, green, yellow, orange, red, white and black. Before training, naive flies are tested for their spontaneous colour preferences on the test array. Yellow is visited most frequently, green least frequently (Table 2). Spontaneous colour preferences do not simply depend on subjective brightness (Table 1). The flies trained to one of the colours prefer this colour significantly (Figs. 5 and 9-11). This behaviour reflects true learning rather than sensitisation (Figs. 6-7). The blue and yellow marks are learned easily and discriminated well (Figs. 5, 9, 11). White is also discriminated well, although the response frequencies are lower than to blue and yellow (Fig. 11). Green is discriminated from blue but weakly from yellow and orange (Figs. 5, 9, 10). Red is a stimulus as weak as black (Figs. 8, 9). These features of colour discrimination reflect the spectral loci of colours in the colour triangle (Fig. 14). The coloured papers seem to be discriminated mainly by the hue of colours, but brightness may also be used to discriminate colour stimuli (Fig.13).

No abstract

No abstract