Testing the randomness of the samples of body size in a replicated design in Drosophila melanogaster populations from the eastern Mediterranean
Ozsoy, E.D.; Bozcuk, A Nihat
Turkish Journal of Zoology 30(2): 221-224
Strong body size correlates, wing and thorax lengths, of Drosophila melanogaster populations from the Eastern Mediterranean were measured and tested for the presence of autocorrelation, i.e. nonrandom alternation of individual measurements within a sample. The experimental design was a laboratory setup aiming at overall homogeneity in fly culture conditions, which was achieved. A t-test, first proposed by Von Neumann, was performed for the autocorrelation analysis in each population. No significant autocorrelation was detectable, indicating very low bias in the experimental design for nonrandom picking of individuals for measurements. The test is suggested for routine use in studies before any inference is made about the state of a character chosen.