The adaptive significance of dorsal spine variation in the fourspine stickleback, Apeltes quadracus. V. Temporal variation

We investigated temporal variation in polymorphisms for the number of dorsal spines in the fourspine stickleback, Apeltes quadracus, and in the ninespine stickleback, Pungitius pungitius, to complement studies based on geographic variation. The changes in spine number that occurred over a 10-year period at Daigle Inlet, New Brunswick, are small relative to geographic variation among sites in the Maritime Provinces. However, some statistically significant changes occur in both species: they take place at or near reproduction; there is no evidence that they are related to selection by predators or any other environmental factor that we measured; and they tend to be followed by reversals that damp the net change. Contrary to expectation, based on spatial relationships, spine numbers in A. quadracus and P. pungitius do not covary predictably over time. The results show that events at or near reproduction play a role in determining local morph frequencies. Our main finding is that spine number is surprisingly stable for both species, and we conclude that it is constrained to local equilibrium values.

The adaptive significance of dorsal spine variation in the fourspine stickleback, Apeltes quadracus, I. Geographic variation in spine number

The goal of our research is to investigate the adaptive significance of a polymorphism for the number of dorsal spines in Apeltes quadracus, the fourspine stickleback. One approach we take is to search for correlations between phenotypes and environments. To this end we collected Apeltes and scored environments at 570 sites in the Maritime Provinces of Canada. In this paper we describe geographic variation in spine number and evaluate how reliably it reflects genetic differentiation among sites. Morph frequencies are highly differentiated geographically. We describe four kinds of variation: relatively constant frequencies, gentle clines, steep clines, and remarkably abrupt changes (called "intrusions") where frequencies at some sites differ greatly from those at a larger number of surrounding sites. Most of the variation among sites is due to differences in the frequencies of the four- and five-spined morphs. However, a remarkable result is that the three-spined morph, which is rare or absent elsewhere in the range, reaches very high frequencies in Bras D'Or Lake. Our evidence suggests this variation among sites reflects substantial genetic differentiation. The differentiation is favorable for detecting selective agents, if indeed selection is responsible.

The adaptive significance of dorsal spine variation in the fourspine stickleback, Apeltes quadracus. II. Phenotype–environment correlations

The fourspine stickleback (Apeltes quadracus) is polymorphic for dorsal spine number, the variation is heritable, and morph frequencies are highly differentiated geographically. Our purpose here is to determine how spines vary in relation to environment. Dorsal spine number is significantly correlated (P < 0.001) with environmental differences among sites; the environmental correlates include predators, potential competitors, vegetation, physical environment, habitat, and geographical position. Based on these correlations we hypothesize that selection by predators favours the higher spined morphs and that selection by competitors favours the lower spined morphs. The correlations with other environmental variables probably reflect interactions with predation and (or) competition, but they may be concomitants of independent and unidentified selective agents. The observed patterns of geographic variation in spine number are in agreement with those expected if predators and competitors are selective agents. We conclude that selection acts on this polymorphism; the geographic differentiation in spine number is adaptive.

Nest site selection by the fourspine stickleback, Apeltes quadracus (Mitchill)

The fourspine stickleback, Apeltes quadracus (Mitchill), is commonly associated with aquatic vegetation on its intertidal breeding grounds, and male Apeltes build nests in the vegetation. We hypothesized that vegetation morphology is a proximate cue in Apeltes nest site selection, and that males nesting in preferred vegetation types are more successful than others in attracting females and acquiring clutches of eggs. In a laboratory experiment, one common vegetation type was rejected as a nest attachment site, but no clear preference was demonstrated among three other common and morphologically distinct types. Nests collected from these four vegetation types in the field showed that males nesting in the less preferred type had acquired fewer clutches and fewer eggs than males nesting in two of the other three types. There were no significant differences in the mean numbers of clutches or eggs acquired by males nesting in these other three types. We suggest that within fairly broad limits Apeltes is a generalist in nest site selection, and that a short breeding season in the highly variable intertidal environment may put a premium on temporal features of breeding rather than on nest location.

Ecology of the fourspine stickleback, Apeltes quadracus, with respect to a polymorphism for dorsal spine number

The subject of our research is the ecological genetics of a polymorphism for dorsal spine number in Apeltes quadracus. Here we present information on its ecology and on dorsal spine variation, as background for subsequent publications.Estimating age from otoliths is subject to several serious errors, and spine sections are questionable. We have estimated generation time at about 1 year using length frequency distributions. Dorsal spine number varies from one to seven, but the four-and five-spined morphs predominate. The polymorphism is ubiquitous; where samples are available for comparisons over time, morph frequencies have remained relatively constant for at least 50 years. Sex ratio is highly variable among sites and among samples at particular sites. However, morph frequencies are independent of sex and age. Morph frequencies are homogeneous among samples taken at differing times of day, tide levels, months within a year, and among years. They are also homogeneous among microhabitats we sampled in an estuary. The relevance of these results to our subsequent publications is the following: (1) we take 1 year as the generation time, (2) sexes and ages can be pooled to estimate morph frequencies, and (3) since morph frequencies remain relatively constant, at least within the time period of our survey, comparisons of morph frequencies among sites to detect geographical and ecological patterns of variation seems valid.