When a limiting resource (e.g. food) varies drastically between years, and population density is measured in the conventional way as individuals per area, demographic processes such as productivity and survival may erroneously be considered density-independent. We tested the hypothesis that if the variation in a limiting resource is not taken into account in the density measure, this may lead to erroneous conclusions about the density-dependence of demographic variables. We studied the food-related variation in productivity of bramblings Fringilla montifringilla, an insectivorous passerine bird, using 19 years of standardised insect censusing, bird censusing and mist-netting of birds in subalpine birch forest in Swedish Lapland. The yearly variation in our measure of brambling per capita productivity (numbers of juveniles per adult trapped) was explained to 30-40% by the larvae abundance of the moth Epirrita autumnata. Taking larvae density into account, no other environmental variable (inferred predation pressure, breeding phenology, and summer temperature) was significantly related to variation in reproductive output. There was no effect of brambling population density on per capita productivity, that is, when density was measured the conventional way, productivity seemed density-independent. However, per capita productivity was significantly and negatively correlated to the food-related population density (population density divided by larval density), supporting the hypothesis that not including a limiting resource into the density measure may indeed lead to erroneous conclusions about the density-dependence of demographic variables.