Field observations and laboratory experiments were carried out to determine the influence of body length of preys on the acceptance rate by spiders. Feeding experiments with 13 spider species and a model prey (crickets) reveal a decreasing acceptance rate with increasing prey size. Prey sizes of 50-80% of the spiders' size yielded the highest acceptance rates, crickets of double the spiders' size were accepted by two species only. By fitting the acceptance rate Y versus prey size X by Y(x)=Y(0) (1-βx2), two coefficients could be calculated: Y(0), the size-independent palatibility of the prey and β, a coefficient of size-induced refusal of the prey. These values describe the degree of specialisation towards (a) crickets and (b) large prey, respectively. Further comparison showed (a) that labidognath (= araneomorph) spiders do not necessarily subdue larger prey items than orthognath (=mygalmorph) spiders and (b) that webbuilding spiders are superior to non-webbuilding spiders in respect of catching large prey. A modified model of the generalized pattern of the length relations of predator and prey is given with special reference to spiders and compared to other polyphagous predator groups.