Patterns of phenotypic change across environmental gradients (e.g., latitude, altitude) have long captivated the interest of evolutionary ecologists. The pattern and magnitude of phenotypic change is determined by the covariance between genetic and environmental influences across a gradient. Cogradient variation (CoGV) occurs when covariance is positive: that is, genetic and environmental influences on phenotypic expression are aligned and their joint influence accentuates the change in mean trait value across the gradient. Conversely, countergradient variation (CnGV) occurs when covariance is negative: that is, genetic and environmental influences on phenotypes oppose one another, thereby diminishing the change in mean trait expression across the gradient. CnGV has so far been found in at least 60 species, with most examples coming from fishes, amphibians, and insects across latitudinal or altitudinal gradients. Traits that display CnGV most often involve metabolic compensation, that is, the elevation of various physiological rates processes (development, growth, feeding, metabolism, activity) to counteract the dampening effect of reduced temperature, growing season length, or food supply. Far fewer examples of CoGV have been identified (11 species), and these most often involve morphological characters. Increased knowledge of spatial covariance patterns has furthered our understanding of Bergmann size clines, phenotypic plasticity, species range limits, tradeoffs in juvenile growth rate, and the design of conservation strategies for wild species. Moreover, temporal CnGV explains some cases of an apparent lack of phenotypic response to directional selection and provides a framework for predicting evolutionary responses to climate change.

Ecology Letters (2012) Abstract The geography of adaptive genetic variation is crucial to species conservation yet poorly understood in marine systems. We analyse the spatial scale of genetic variation in traits that broadly display adaptation throughout the range of a highly dispersive marine species. We conducted common garden experiments on the Atlantic silverside, Menidia menidia, from 39 locations along its 3 000 km range thereby mapping genetic variation for growth rate, vertebral number and sex determination. Each trait displayed unique clinal patterns, with significant differences (adaptive or not) occurring over very small distances. Breakpoints in the cline differed among traits, corresponding only partially with presumed eco‐geographical boundaries. Because clinal patterns are unique to each selected character, neutral genes or those coding for a single character cannot serve as proxies for the genetic structure as a whole. Conservation plans designed to protect essential genetic subunits of a species will need to account for such complex spatial structures.

The coastal marine environment of the Northwest Atlantic contains strong environmental gradients that create distinct marine biogeographic provinces by limiting dispersal, recruitment, and survival. This region has also been subjected to numerous Pleistocene glacial cycles, resulting in repeated extirpations and recolonizations in northern populations of marine organisms. In this study, we examined patterns of genetic structure and historical demography in the Atlantic silverside, Menidia menidia, an annual marine fish with high dispersal potential but with well-documented patterns of clinal phenotypic adaptation along the environmental gradients of the Northwest Atlantic. Contrary to previous studies indicating genetic homogeneity that should preclude regional adaptation, results demonstrate subtle but significant (FST = 0.07; P < 0.0001) genetic structure among three phylogeographic regions that partially correspond with biogeographic provinces, suggesting regional limits to gene flow. Tests for non-equilibrium population dynamics and latitudinal patterns in genetic diversity indicate northward population expansion from a single southern refugium following the last glacial maximum, suggesting that phylogeographic and phenotypic patterns have relatively recent origins. The recovery of phylogeographic structure and the partial correspondence of these regions to recognized biogeographic provinces suggest that the environmental gradients that shape biogeographic patterns in the Northwest Atlantic may also limit gene flow in M. menidia, creating phylogeographic structure and contributing to the creation of latitudinal phenotypic clines in this species.Electronic supplementary materialThe online version of this article (doi:10.1007/s00227-010-1577-3) contains supplementary material, which is available to authorized users.

The Atlantic silverside, Menidia menidia (Pisces: Atherinidae), exhibits an exceptionally high level of clinal variation in sex determination across its geographic range. Previous work suggested linear changes in the level of temperature-dependent sex determination (TSD) with increasing latitude. Based on comparisons at 31 sites encompassing the entire species' range, we find that the change in level of TSD with latitude is instead highly nonlinear. The level of TSD is uniformly high in the south (Florida to New Jersey), then declines rapidly into the northern Gulf of Maine where genotypic sex determination (GSD) predominates and then rebounds to moderate levels of TSD in the northern-most populations of the Gulf of St. Lawrence. Major latitudinal breakpoints occur in central New Jersey (40(o)N) and the northern Gulf of Maine (44(o)N). No populations display pure TSD or GSD. Length of the growing season is the likely agent of selection driving variation in TSD with a threshold at 210 days. Because gene flow among populations is high, such distinct patterns of geographic variation in TSD/GSD are likely maintained by contemporary selection thereby demonstrating the adaptive fine tuning of sex determining mechanisms.

a b s t r a c tAtlantic salmon (Salmo salar) are endangered anadromous fish that may be exposed to feminizing endocrine disrupting compounds (EDCs) during early development, potentially altering physiological capacities, survival and fitness. To assess differential life stage sensitivity to common EDCs, we carried out short-term (4 day) exposures using three doses each of 17 ␣-ethinylestradiol (EE2), 17 ␤-estradiol (E2), and nonylphenol (NP) on four early life stages; embryos, yolk-sac larvae, feeding fry and 1 year old smolts. Differential response was compared using vitellogenin (Vtg, a precursor egg protein) gene transcription. Smolts were also examined for impacts on plasma Vtg, cortisol, thyroid hormones (T 4 /T 3 ) and hepatosomatic index (HSI). Compound-related mortality was not observed in any life stage, but Vtg mRNA was elevated in a dose-dependent manner in yolk-sac larvae, fry and smolts but not in embryos. The estrogens EE2 and E2 were consistently stronger inducers of Vtg than NP. Embryos responded significantly to the highest concentration of EE2 only, while older life stages responded to the highest doses of all three compounds, as well as intermediate doses of EE2 and E2. Maximal transcription was greater for fry among the three earliest life stages, suggesting fry may be the most responsive life stage in early development. Smolt plasma Vtg was also significantly increased, and this response was observed at lower doses of each compound than was detected by gene transcription suggesting plasma Vtg is a more sensitive indicator at this life stage. HSI was increased at the highest doses of EE2 and E2, and plasma T 3 was decreased at the highest dose of EE2. Our results indicate that all life stages are potentially sensitive to endocrine disruption by estrogenic compounds and that physiological responses were altered over a short window of exposure, indicating the potential for these compounds to impact fish in the wild.