Spatial environmental gradients can promote adaptive differences among conspecific populations as a result of local adaptation or phenotypic plasticity. Such divergence can be opposed by various constraints, including gene flow, limited genetic variation, temporal fluctuations, or developmental constraints. We focus on the constraint that can be imposed when some populations are found in locations characterized by low levels of an essential nutrient. We use scales of wild fish to investigate phenotypic effects of spatial variation in a potentially limiting nutrient—calcium. If scale calcium (we use “scalar” calcium for consistency with the physiology literature) simply reflects environmental calcium availability, we expect higher levels of scalar calcium in fish from calcium‐rich water, compared to fish from calcium‐poor water. To consider this “passive response” scenario, we analyzed scalar calcium concentrations from three native fish species (
Lepomis gibbosus
,
Percina caprodes
, and
Perca flavescens
) collected at multiple sites across a dissolved calcium gradient in the Upper St. Lawrence River. Contradicting the “passive response" scenario, we did not detect strong or consistent relationships between scalar calcium and water calcium. Instead, for a given proportional increase in water calcium across the wide environmental gradient, the corresponding proportional change in scalar calcium was much smaller. We thus favor the alternative “active homeostasis” scenario, wherein fish from calcium‐poor water are better able to uptake, mobilize, and deposit calcium than are fish from calcium‐rich water. We further highlight the importance of studying functional traits, such as scales, in their natural setting as opposed to only laboratory studies.