Food web structure is paramount in regulating a variety of ecologic pattern
s and processes, although food web studies are limited by poor empirical de
scriptions of inherently complex systems. In this study, stable isotope rat
ios (delta(15)N and delta(13)C) were used to quantify trophic relationships
and food chain length (measured as a continuous variable) in 14 Ontario an
d Quebec lakes. All lakes contained lake trout as the top predator, althoug
h lakes differed in the presumed number of trophic levels leading to this s
pecies. The presumed number of trophic levels was correlated with food chai
n length and explained 40% of the among-lake variation, food chain length w
as most closely related to fish species richness (r(2) = 0.69) and lake are
a (r(2) = 0.50). However, the two largest study lakes had shorter food chai
ns than lakes of intermediate size and species richness, producing hump-sha
ped relationships with food chain length. Lake productivity was not a power
ful predictor of food chain length (r(2) = 0.36), and we argue that product
ive space (productivity multiplied by area) is a more accurate measure of a
vailable energy. This study addresses the need for improved food web descri
ptions that incorporate information about energy flow and the relative impo
rtance of trophic pathways.