Many ecological interactions are not trophic and they do not represent competition. Until very recently, these interactions have been ignored in ecology although they could be very important in creating system stability.
But how do we find these interactions ? One way is by careful study of individual species interacting - but this is logistically expensive. Another way is to statistically examine large data sets and extract the interactions. In this seminar, i describe one method of identifying interactions in species extensive field abundance data. From these I create interaction webs - diagrams of all the trophic and non-trophic relationships between groups. I then compare the reliability of these webs and show that reliabilities for interaction webs are very high, at least as high as those for food-webs, and that they show similar relationships to other web characteristics. I demonstrate that, in nature, the reliability of interaction webs, and so also their contribution to mediating the biodiversity-ecosystem relationship, will depend on both web size and topology.