In the absence of well-resolved food webs that include both information on predators and parasites, body-size has been proposed as the biological mechanism underlying the assumption of a trophic hierarchy in the cascade model and its variants. We show that parasites consume prey larger than themselves, and therefore the logical conclusion is that the trophic hierarchy cannot be justified on the basis of body-size when food webs include parasites. Two other assumptions central to the cascade model and its variants, the link-species scaling law and the probability of predation, are not supported by the empirical data presented here. We show that linkage density is scale-dependent rather than independent, but that the relationship between linkage density and web size is far from clear. Specifically, the webs that exclude parasites examined here, overestimate the values of linkage density and connectance by more than one order of magnitude. There is little evidence for a relationship between predator generalisation (the number of prey exploited by a predator), prey vulnerability (the number of predators exploiting a prey) and trophic position within each web, when trophic position corresponds to an ordering by size. Therefore the probability of predation neither remains constant as assumed under the original cascade model, nor increases with trophic position (increasing body-size) as assumed under some of its variants. The present study however, does show some support for the notion that the cascade model's assumption of constant predation probability should be replaced by heterogeneous predation probabilities. We also examined the relationship between the morphological niche of consumers, a metric where resources falling within a defined size are always taken, and a parameter that may be closely linked to the connectance of whole food webs, but rarely examined in empirical webs. Although we could find no evidence for trends in morphological niche with trophic position, further investigation of the consequences of individual or species-level feeding constraints on the connectance of webs may repay further investigation. We contend that the organisms often most neglected in food web studies e.g. parasites and pathogens, must be included in food webs if we are able to make more realistic descriptions of the trophic structure of natural communities, and to develop mechanistic explanations for the structure of natural food webs.