Temporal stability of δ2H in insect tissues: Implications for isotope-based geographic assignments

Hydrogen isotope geolocation of insects is based on the assumption that the chitin in the wings of adult migratory insects preserves the hydrogen isotope composition ( δ 2 H) of the larval stages without influence of adult diet. Here, we test this assumption by conducting laboratory feeding experiments for monarch butterflies ( Danaus plexippus ) including: (1) a starvation treatment where adults were not fed and (2) an enriched treatment where adults were fed a diet isotopically enriched in deuterium (~ +78‰) compared to the larval diet. The δ 2 H values of adult wings were measured at different time steps along the 24-day experiment. We also investigated intra-wing differences in δ 2 H values caused by wing pigmentation, absence of wing scales, and presence of major wing veins. We conclude that, although the magnitude of the changes in δ 2 H values are small (~6‰), wing δ 2 H values vary based on adult diet and insect age, particularly early after eclosion (i.e., 1–4 days). We found that wing shade, wing pigmentation, and the presence of wing scales do not alter wing δ 2 H values. However, wing samples containing veins had systematically higher δ 2 H values (~9‰), suggesting that adult diet influences the hemolymph that circulates in the wing veins. We hypothesise that there is a stronger influence of adult diet on the isotope signal of wings during early adult life relative to later life because of increased metabolic and physiologic activity in young insect wings. We argue that the influence of the isotopic contribution of adult diet is generally small and is likely minimal if the wings are carefully sampled to avoid veins. However, we also demonstrated that wings are not inert tissues, and that adult feeding contributes to some of the intra-population δ 2 H variance. We conclude that δ 2 H geolocation using insect wings remains valid, but that adult feeding, butterfly age and wing vein sampling generate an inherent uncertainty limiting the precision of geolocation.