We counted references in about 200 research papers in each of 16 journals in six physical sciences. They show that for average papers, the number of references is a linear function of the paper length. In fact, it is the same function for journals in different sciences. The fact that various physical sciences all give the same reference frequencies for papers of the same length and impact factor tells us that citation counts in those sciences can be intercompared. There is a dependence upon impact factor and a general relation is derived. In addition, the number of references increases by about 1.5% per year, probably due to the increase in the literature pertinent to any paper. The average paper lengths differ among the six sciences and three possible explanations for that difference are given.