• host alternation;
  • migration;
  • overwintering;
  • Rhopalosiphum padi;
  • sympatric speciation

1. For a full assessment of explanations for the evolution of host-specificity it is necessary to estimate the probability that a dispersing parasite finds a host. We develop a method of estimating this success rate from samples of dispersing parasites and populations resident on hosts.

2. Applying this method to data on the bird cherry-oat aphid, Rhopalosiphum padi (L.), from southern Scotland in 1984–92, we estimate that 0·6% of the autumn migrants find hosts.

3. With such a low success rate, there should be selection for a broadening of host range, to include any host on which the colonist's fitness is more than about 0·6% of that on the normal hosts. We argue that neither nutrition nor the need for ‘enemy-free space’ are sufficient explanations of the host-specificity of this animal, and propose instead that it is the host's role as a rendezvous for mating that constrains the migrants to their costly host-specificity.

4. We also discuss the implications of this low success rate for the hypothesis that aphids speciate sympatrically through the formation of host races.