Dragonflies are often included in aquatic food web studies because they are relatively long-lived, generalist predators, but sacrificing them to obtain information about their trophic ecology might be unnecessary if nonliving tissues such as exuviae (final-instar larval exoskeletons) and adults are stoichiometrically similar. To determine the degree of intraspecific, interspecific, and between-site variation in adult and exuvial %C, %N, and C:N ratios, I collected groups of newly-emerged adults and their respective exuviae for four species of libellulid dragonflies emerging at three pond sites in Lawrence, KS (USA). Exuvial %C values were always more variable than adult %C, with means ranging from 28.51-46.32% (± 1.09-9.09) for groups of exuviae vs. 45.52-48.13% (± 0.32-1.93) for their respective adults. %N values were also more variable in exuviae (6.85-11.14%, ± 0.31-2.29) than in adults (11.48-12.49%, ± 0.14-0.51). In contrast, mean C:N ratios were remarkably similar within groups and between exuviae and their respective adults regardless of species or site, with consistently slightly higher values in exuviae (4.16-4.31, ± 0.04-0.11) than in adults (3.78-4.05, ± 0.05-0.10). The small and predictable differences in C:N ratios between exuviae and adults (0.17-0.53, ± 0.09-0.12) suggests that the stoichiometric homeostasis commonly observed in predatory aquatic insects can be reflected in their discarded exoskeletons. Therefore, substituting non-living exuviae for whole dragonflies in stoichiometric analyses can provide equivalent information while minimizing negative impacts on these insects’ sensitive populations and fragile aquatic habitats.