Trophic discrimination factors and stable isotope variability in a captive feeding trial of the southern rock lobster Jasus edwardsii () (Decapoda: Palinuridae) in Tasmania, Australia

Abstract Trophic discrimination or fractionation factors (TDFs), such as ∆15N and ∆13C, are used in stable isotope mixing models to account for differences between source tissues (diet/prey) and consumer tissues (predator). We aimed firstly to obtain TDF values for a spiny lobster, the southern rock lobster Jasus edwardsii (Hutton, 1875), to better understand lobster diet in the wild and secondly to investigate variability in isotope signature within tissues of individuals and across a temporal scale to test if non-lethal sampling can be used in an ecological context. We conducted an 18-mo captive feeding trial with juvenile lobsters using three diet treatments and analysed dorsal and leg muscle, along with dorsal and leg exoskeleton for δ13C and δ15N values. Average TDFs for the three diet treatments were 3.86 ± 0.98‰ (∆13C) and 5.06 ± 0.65‰ (∆15N) for leg muscle, and 4.45 ± 1.04‰ (∆13C) and 4.36 ± 0.6‰ (∆15N) for dorsal muscle. When tested against wild lobsters and prey, these TDFs outperformed multi-taxa TDFs found in the literature. Isotope values from lobster leg muscle were not identical to associated dorsal muscle but the two were highly correlated, indicating that non-lethal sampling is acceptable. Values for exoskeleton isotope were significantly different from muscle, likely due to the exoskeleton not being in a constant state of growth and replacement, unlike the muscle tissue, which constantly incorporates new material. We conclude that our experimentally derived TDFs are suitable for mixing model analysis for J. edwardsii and when tested on a wild sample of lobsters they outperformed other TDFs reported in the literature. We show that non-lethal sampling using leg muscle is an appropriate sampling method, since this tissue is highly correlated to the commonly used dorsal muscle. This option for non-lethal sampling enhances the potential to widely sample wild populations or sample during industrial processing without the need to sacrifice whole animals.