PMC

Effect of changing guarding inefficiency, ε_G, on the ASR at which the dominant strategy switches from multiple mating to guarding, i.e. when guarding begins to take over the population. When ε_G is just greater than zero, so guarding is slightly less than perfectly effective, the ASR required for guarding to take over from multiple mating rises immediately from 0.5 to 0.57 and the takeover threshold ASR for guarding increases, becoming more male biased with decreasing effectiveness of guarding. Other parameters used are β = 0.8, c = 0.2 and ε_P = 1. When ε_G = 1, all guarded females' offspring are susceptible to extra-pair paternities by multiple-mating males, and the sex ratio at which guarding begins to take over is 0.62. (Online version in colour.)

(A) Time-course for the mate-guarding behavior observation. One female and two males were placed in an aquarium in the morning (~10:00) or in the evening (~20:00). Lights were turned off at 22:00. (B) One example of typical mate-guarding behavior. One male tended to maintain its position between the female and the other male and prevent that male from approaching the female. We defined this behavior as “mate-guarding behavior”. (C) A procedure for calculating the guarding index representing the degree of mate-guarding. As fish with the same genotype were used, we discriminated the two male fish according to the mean distance for 100 s from the female (near male and far male). We measured the relative locations of the three fish and calculated the probability of the near male being in the guarding circle with center (1/2, 0) and radius 1/2 when the female and far male positions were defined as (0, 0) and (1, 0), respectively. We defined this probability as the “guarding index of the near male” and compared the groups in the guarding test. (D) Relative positions of the three fish (one female and two males). Dots indicate the relative positions of the near male over 100 s (1 dot/5 s). Red and green small circles indicate positions of the female and far male, respectively. Guarding indices are shown in the lower right. “Merge” indicates superimposed data of three fish moving independently, and thus were used as a negative control. Each n = 21 dots x 13 trials. (E) Comparison of guarding indices of the near male in the experimental group with those in the merged control to examine the effect of female condition on the guarding test. Mean ± SEM. Each n = 13, Student’s t-test: ***P<0.001. Before: guarding test before spawning in the morning. After: guarding test after spawning in the morning. Evening: guarding test in the evening. (F) Comparison of guarding indices of the near male in the experimental group with those of the merged control to examine the effect of fish properties on guarding test. Mean ± SEM. Each n = 13, Student’s t-test. Zebrafish: one zebrafish female and two medaka males. Female: two females and one male. Please see . for detail procedures for individual conditions.

a. Duration of guarding (minutes/h) and b. number of guarding bouts/h by M ₁ by M ₂ and by F ₁.