The interaction between predation landscape and phenotypic variability within prey populations is of substantial significance in evolutionary biology. Extending from several decades of studies at a remote freshwater lake on Haida Gwaii, western Canada, we analyze the incidence of predator-induced sub-lethal injuries in 8,069 wild-captured threespine stickleback (Gasterosteus aculeatus) and using cohort analyses test whether the distribution of injuries informs the selective landscape influencing the bell-shaped frequency distribution of the traits. Our results indicate that (1) the incidence of injuries varies among phenotypes differing in the number and position of lateral plates, (2) these differences occur only among younger fish, (3) the incidence of injuries is inversely related to the estimated population frequencies of plate phenotypes, with the modal phenotype generally having the fewest injuries, (4) direct estimates of selective differentials and relative fitness based on analyses of 1,735 fish from 6 independent yearly cohorts indicates statistically informative elevated differentials in phenotypes with greater number of plates and elevated relative fitness of non-modal phenotypes, and (5) there are significant differences among yearly cohorts in strength and direction of selection, and an increased prevalence of diversifying versus stabilizing selection despite longer-term stasis (4 decades) in trait means. We conclude that the presence of multiple “optimal” phenotypes complements the renewed interest in quantifying short-term temporal or spatial variation in ecological processes in studies of fitness landscapes and intrapopulation variability.

Stickleback were sampled (1980-1983, 1985-1990, 2006, 2011-2013, 2015-2016) with minnow traps baited with 1cc old cheddar cheese (total N=8069). For each collection prior to 2006, all fish with any evidence of injury as well as an equivalent sub-sample of uninjured fish were retained (details in Reimchen 1995) while from 2006 onwards, sampling effort was greatly reduced but all fish (injured and uninjured) were retained. Fish were anesthetized with MS-222 or Clove Oil preserved in 10% formalin and later transferred to isopropyl alcohol or ethanol. Collections were made under the Ministry of Environment (Gov’t of British Columbia) permits (SM09-51584 and SM10-62059) and the University of Victoria Aquatic Unit facility Standard Operating Procedure OA2003.

We measured standard length (anterior edge of the jaw to the posterior edge of the caudal peduncle +/- 1mm), sex (examination of gonads), number and position of left lateral plates, length of the ventral plate of the pelvic girdle, and skeletal and integumentary injuries (details in Reimchen 1983, 1988, Reimchen et al., 1985). All traits, apart from SL, were measured under a dissecting scope.