(4) An attacking predator may be confused by several prey animals and therefore be less effective

at catching any of them. This has been demonstrated by Neill & Cullen (1974) and

Major (1978) for predators on small fish singly or in shoals. It is much less likely to apply to

a lion attempting to catch one of the few widely spaced birds in an ostrich group.

(5) If only one prey animal is captured from a prey group, any particular individual is less likely

to be the victim the more companions it has. This protection by dilution (Hamilton 1971; Bertram

1978) inevitably operates provided that at least some of the prey group escape, and provided

that the rate of predator attacks does not increase faster than group size. Hamilton (1971)

demonstrated that grouping in this way could be selected for even if the predator's success rate

was higher than when hunting solitary prey. It is this aspect of grouping which appears to provide

ostriches with their greatest benefit against predators.

(6) A potential prey animal with companions may be able to situate itself in such a way, for

example in the middle of the group, that it is less likely than its companions to be the victim of a

predator attack (Hamilton 1971). Ostrich feeding groups are usually so small that a middle would

be hard to define, and are flexible and changing in shape as the birds drift about; there is no sign

of any seeking of the centre of the group. Clearly there may be other benefits from being

with companions apart from individual protection against predation. For example, another

important function of temporary grouping, in ostriches as in many other species, is of course

for social interactions such as display, mate selection, courtship, reproduction and social

learning.

Enhanced feeding rates as a result of having and presumably observing companions have

been demonstrated experimentally in captive great tits, Parus major, by Krebs et al. (1972). On

the other hand, Goss-Custard (1976) showed that redshanks (Tringa totanus) obtained food at

a slower rate when their companions were feeding nearby; the type of food, its distribution, and

the method by which it is found are obviously important. Murton (1971) showed that woodpigeons


(Columbapalumbus) copied one another's selection between food types, and suggested that

the subordinate birds at least benefited by doing so. In ostriches, peck rates while the head was

down could rarely be observed; when they could, they were variable and showed no relation to

group size. I assume that reduced vigilance was equivalent to a faster rate of intake of food. A

bird feeding alone was able to be pecking for 65 % of the time, a bird with one companion for

79 % of the time, and a bird in a group of three or four for 85 % of the time. Thus a first companion

produced a 21% increase in the rate of food intake, and the next one or two companions

produced a further 9 % increase. There was no indication that ostriches were gathering and

feeding in groups particularly in places where food was most abundant (as was shown for

great blue herons, Ardea herodias, by Krebs 1974), thereby causing a positive relationship

between feeding rate and group size. No data could be collected on ostrich food types in

relation to companions' food types. The interaction between anti-predator and

feeding benefits from grouping is a complicated one (Lazarus 1972), even without the additional

complications of food competition discussed by Rubenstein (1978). By reducing its own vigilance

when with companions, a bird is enabled to feed at a faster rate. If food or feeding time is limiting,

grouping can result in an individual's obtaining a greater total food intake (Drent & Swierstra

1977). If feeding time is not limiting, as appeared to be the case during the ostrich observations,

grouping does not increase the total food intake, but reduces the time during which the bird is

running an increased risk of predation, and reduces the predation risk during that time.

Depending on the relative risks while behaving in different ways or of being in different places,

predation can to some extent limit feeding time. By being in groups when feeding, an ostrich

increases its feeding rate by reducing its vigilance such that the group's vulnerability remains

almost the same. Nonetheless, there is a large decrease in its own individual vulnerability, and

an increase in its feeding rate. Until there are many more data it will not be possible to compare

these two benefits in any common currency of fitness, but that must remain the aim.

Acknowledgments

1 am indebted to the Government of Kenya for permission to carry out this research in Tsavo, to

the Royal Society for financial support, to Kate Bertram for her help in the field, and to Dan

Rubenstein, Richard Wrangham and an anonymous referee for their criticisms of an earlier

manuscript of this paper.


www.ostrichrdi.com