Variation in female-biased sexual size dimorphism of northern pike (Esox lucius) associated with environment and life history

Abstract

Sexual size dimorphism (SSD) is a widespread phenomenon in the animal world resulting from differential selection on the sexes. The northern pike (Esox lucius) is a freshwater apex predatory fish species that exhibits female-biased SSD, but the degree to which SSD varies among populations and what variables might dictate variation in SSD in this species remain poorly understood. We sought to quantify the degree of variation in SSD among northern pike populations across a large portion of their North American range, as well as evaluate associations between the magnitude of SSD in northern pike populations with environmental variables and life history traits of populations. We quantified SSD in 102 populations of northern pike across the province of Ontario, Canada, using a standardized gillnetting database, and investigated the degree to which both environmental variables (cisco [Coregonus artedi] abundance as catch-per-unit-effort, lake surface area, and latitude) and northern pike life-history traits (early growth and mortality rates) explained variation in female-biased SSD using linear models. Female-biased SSD in mean weight of northern pike increased with increasing cisco abundance, and the difference in female and male mean age increased with increasing latitude. Furthermore, SSD was greater in populations with lower female mortality and early growth rates. These results indicate that slow-growing, long-lived populations of northern pike should exhibit greater female-biased SSD, and that these conditions may be facilitated by the availability of large, energy-dense prey and cooler temperatures at northern latitudes.

Appendix

Descriptive statistics and correlation tables for response and predictor variables, example calculation of S and Z estimates, model outputs for northern pike SSD analyses with environmental and life-history variables, and the relationships between female and male Z and ω estimates (Figs. 6, 7, 8 and Table 1, 2, 3, 4, 5, 6, 7).

Fig. 6

Example of the descending limb of a catch curve used to estimate instantaneous total mortality for female northern pike sampled in Lake Nipissing, Ontario. Solid dots represent the log number of individuals caught in each age class on the descending limb of the catch curve. Open dots represent the log number of individuals caught in each age class on the ascending limb of the catch curve and were not used in the estimate of instantaneous total mortality. The solid line is a regression line

Fig. 7

Relationship between female northern pike early growth rate estimates (ω) and size-at-age 3 (mm) (LM, y = 0.57x + 38.11, df = 94, t = 4.73, P < 0.0001, R² = 0.18). Shading represents 95% confidence intervals around the regression line

Fig. 8

Relationship between male and female northern pike instantaneous total mortality estimates (Z; LM, y = 0.81x + 0.17, df = 100, t = 8.12, P < 0.0001, R² = 0.39). The dashed line indicates a 1:1 relationship

Table 1 Descriptive statistics for northern pike SSD response variables and environmental and life-history predictor variables included in linear models

Table 2 Pearson correlation coefficients among northern pike SSD response variables

Table 3 Example calculation of S and T estimates to support Chapman–Robinson instantaneous mortality estimates for female northern pike in Lake Nipissing, Ontario

Table 4 Pearson correlation coefficients among predictor variables included in the analysis

Table 5 Summary statistics for the mean fork length (mm) at age for populations with at least three individuals in each age class

Table 6 Linear model outputs for the effects of environmental predictor variables on northern pike SDI for L₁₀ (df = 98, F = 0.71, P = 0.55, R² = 0.02), mean weight (df = 98, F = 4.69, P = 0.004, R² = 0.13), and mean age (df = 98, F = 5.60, P = 0.001, R² = 0.15)

Table 7 Linear model outputs for the effects of female northern pike Z and ω on northern pike SDI for L₁₀ (df = 98, F = 12.49, P = 1.48e−05, R² = 0.20), mean weight (df = 98, F = 34.76, P = 3.89e−12, R² = 0.42), and mean age (df = 98, F = 31.91, P = 2.13e−11, R² = 0.39)