Anita. Behav., 1987, 35, 1-6 The amorous Gammarus: size assortative mating in G. pulex ROBERT ELWOOD, JULIE GIBSON & SUSAN NEIL Department of Zoology, The Queen's University of Belfast, Belfast, N. Ireland Abstract. The hypothesis that size assortative mating in Gammarus pulex is due to mal~male competition was examined in three experiments. When males were given a choice of two mates they preferred to enter into precopula with the female that offered the greater utility in terms of eggs fertilized per reproductive effort. When a receptive female was simultaneously encountered by two males it was usually the larger male that entered into precopulatory amplexus. There was more male-male c~nflict in the presence of a receptive female than in that of an unreceptive female. Few take-overs occurred but those that did were by males larger than the male in precopula. These data are consistent with the hypothesis that competition results in size assortative mating; however, it is suggested that the optimal time (relative to female moult) of entry into precopula varies as a function of both male and female size. Thus large males enter into precopula earlier than small males and prefer to take a large female to a small female. Mal~male competition may determine some of the costs and benefits associated with size assortative mating but it is not the mechanism by which it occurs. In populations with a sex ratio of unity, males may be able to fertilize eggs at a faster rate than they can be produced so that at any one time the operational sex ratio may be biased towards males. Inevitably this will lead to competition among the males for access to receptive females and one form this competition may take is mate guarding (Parker 1970). Precopulatory guarding occurs in many species but is particularly common in the Crusta- cea, a group in which females are often receptive only for a short period after moulting and males may therefore guard females prior to their moult. Such precopulatory guarding occurs in Gammarus pulex, a freshwater amphipod in which males carry females for several days prior to mating (Hynes 1955; Birkhead & Clarkson 1980; Ward 1984). Guarding by males is expensive both in energy, since the mate carries the passive female when swimming (Adams & Greenwood 1983), and time, during which the male cannot mate with other females. Some tangible benefit must therefore accrue to a male that performs this activity; this benefit is his presence at the female's moult which affords the male the opportunity to mate and fertilize her next batch of eggs. The time for which a male should guard a female has been investigated theoretically (Grafen & Ridley 1983), and it has been shown that it may be advantageous for a male to guard a female for all or for only a part of her moult cycle. If males guard for only a part of the moult cycle they would be expected to take females closest to their moult and selection should favour males that detect the time remaining before a female's moult (Birkhead & Clarkson 1980; Ward 1984; Hunte et al. 1985). The optimal time for which a male should guard is influenced by many factors including the operational sex ratio, search time, costs of searching per unit time, cost of guarding, number of eggs likely to be fertilized, and the possibility of take-overs. The precise mathema- tical solution to this question is complex even in unrealistically simple situations. For example, Grafen & Ridley (1983) assumed that either both sexes did not vary in size or only males varied in size resulting in possible take-overs by larger males. However, in G. pulex both sexes vary in size and this may be important in mating decisions since large females carry more eggs than do small ones (Birkhead & Clarkson 1980). Size assortative mating occurs in G. pulex, with large males mating with large females, usually with a male: female size ratio of about 1.3 in this sexually dimorphic species (Birkhead & Clarkson 1980; Adams & Greenwood 1983; Ward 1983, 1984). However, the mechanism by which this size assor- tative mating occurs is a matter of some contro- versy. Three main hypotheses have been proposed to account for it. First, Birkhead & Clarkson (1980) have sug- gested that males select females simply on the basis of the time before their moult and that size assortative mating results from spatial heteroge-