Here the groups of letters underlined are runs and the number of runs is totalled in parentheses.From the tables of Swed and Eisenhart (1943), testing the one-sided hypothesis that there are no more runs than would be expected by chance, each of these shows a significantly large number of runs (the first less than 5% significance, the others less than 2.5%). That is, each species tends to decrease following an increase and to increase following a decrease, proving population control non random. The mean periods of these fluctuations can easily be computed. F or a run of increases
followed by a run of decreases constitutes one oscillation. Thus, the periods of the oscillations of the three species are 13/5 = 2.6, 70/27 = 2.6,
and 46/19 = 2.4 years respectively. These fluctuations would require an unknown environmental cycle of period approximately 2.5 years if a
regularly recurring density independent event were controlling the populations. Thus, f~01l1 these data alone, it seems very probable that the three
species (myrtle, black-throated green. blackburn
ian) are primarily regulated by density dependent
events.