We explore the claim that citation rates for astronomers should not be compared with those of other scientists because astronomers put more references in their papers, causing their citation rates to be artificially high. We scanned 2079 papers with 59,659 references in 13 journals: five American astronomical journals, three non-American astronomical journals, and five journals in physics, chemistry, and geophysics. We find a tight relation between mean numbers of references, <R>, and mean paper lengths,, expressed in normalized 1000-word pages; it is <R> 9.9 + 2.18 . Except for review papers and papers filled with many new data, this relation holds true within 2-3 σ for 12 of the journals. The reason why American papers in astronomy and geophysics have more references than papers in physics and chemistry is because the former are approximately twice as long; for papers of the same lenghts, the average numbers of references per paper are the same among all four sciences.