Investigating the concentration within a research community using joint publications and co-authorship via intermediaries

Given extensive research collaboration in modem science, both at the national and international level, one might wonder whether the network of researchers within each discipline is now sufficiently meshed that a large proportion of contributors to peer-reviewed journals in a given field could either share joint publications or, more realistically, be connected through chains of co-authorships. Such is not the case yet in the fields of probability and statistics, however, as shown here using a large data base covering 9 reknown journals from each of these two areas over the period 1986-1995.