Figure 5.
Deconstructed time-series analysis of perch disparity. Perch disparity was calculated as the sum of females perching minus the sum of males perching. Data were collected in 0.93 m3 mating cages (n = 4 treatments, n = 3 replicates) held in an indoor rearing environment from 07:00 h to 18:00 h for 6 days. Each cage had an initial population of 90 males and 90 females. Gaps in observed data were filled by taking a 13-hour moving average. Observed data are the sum of “trend” + “seasonal” + “random” data. Respectively, these represent the following: (a) declining perch disparity with each day and a secondary post-mating peak; (b) cyclical pattern of sex bias to peak in the early hours and decline throughout the day; and (c) random effects not otherwise explained.