1. Lipid‐rich animal tissues have low δ¹³C values, which can lead to inaccurate ecological inferences. Chemical lipid extraction (LE) or correction models account for this depletion, but the need for LE or correction is tissue‐ and species‐specific. Also, LE can alter δ¹⁵N values, increasing labour and costs because bulk samples must be analysed for δ¹⁵N values separately.
2. We studied the effects of LE on δ¹³C and δ¹⁵N values in liver, muscle and skin of common bottlenose dolphins Tursiops truncatus and West Indian manatees Trichechus manatus, two ecologically important species that occupy different trophic levels. We fit lipid‐correction models to each species. We also performed a meta‐analysis to more broadly determine the effects of LE across taxa, tissues and trophic groups (carnivores, omnivores and herbivores) and to fit lipid‐correction models to different taxonomic and trophic groups.
3. Lipid extraction increased the δ¹³C values in dolphin tissues but had little effect on manatee tissues and no effect on the δ¹⁵N values in either species. A mass balance lipid‐correction model best fit the data from all dolphin tissues, and a linear model best fit data for manatee liver while null models best fit data from manatee muscle and skin. Across 128 terrestrial and aquatic species, the effects of LE varied among tissues and were lower for herbivores compared to carnivores. The best‐fitting lipid‐correction models varied among tissue, taxa and trophic groups. Finally, the δ¹⁵N values from muscle and liver were affected by LE.
4. Our results strengthen the growing body of evidence that the need for LE is tissue‐ and species‐specific, without a reliable C:N ratio predictive threshold. The prediction errors of lipid‐correction models generally decreased with taxonomic and trophic specificity. The smaller effects of LE in herbivores may be due to differences in diet composition or the physiology of lipid synthesis in members of this trophic group. These results suggest that researchers should use the most species‐, tissue‐ and trophic group‐specific information on LE available and, if not available, perform LE on a subset of samples prior to analysis to determine effects.

中文翻译：

---

脂质提取对δ13C和δ15N值的影响以及跨组织，分类群和营养族的脂质校正模型的使用

1. 富含脂肪的动物组织具有低δ ¹³ C值，这可能会导致不准确的生态推论。化学脂质提取（LE）或校正模型解决了这种消耗问题，但LE或校正的需求因组织和物种而异。另外，LE可以改变δ ^15个N个值，增加劳力和成本，因为散装样品必须分析δ ¹⁵分开的N个值。
2. 我们研究了LE的影响δ ¹³ C和δ ¹⁵在肝脏，肌肉和常见的宽吻海豚皮肤的N值宽吻海豚和西印度海牛Trichechus manatus，占据不同营养水平的两个重要生态价值的物种。我们将脂质校正模型适合每个物种。我们还进行了荟萃分析，以更广泛地确定LE在整个分类群，组织和营养类群（食肉动物，杂食动物和草食动物）中的作用，并使脂质校正模型适合不同的分类学和营养族。
3. 脂质提取增加了δ ¹³ C值在豚组织，但对海牛组织的影响很小并且在δ无影响¹⁵中任一物种的N个值。质量平衡脂质校正模型最适合所有海豚组织的数据，线性模型最适合海牛肝的数据，而空模型最适合海牛肌肉和皮肤的数据。在128种陆生和水生物种中，LE的影响在各组织之间各不相同，与食肉动物相比，草食动物的LE效应较低。最合适的脂质校正模型在组织，分类群和营养组之间有所不同。最后，δ ¹⁵从肌肉和肝脏的N值由LE的影响。
4. 我们的结果加强了越来越多的证据表明，对LE的需求是针对组织和物种的，而没有可靠的C：N比预测阈值。脂质校正模型的预测误差通常随着分类学和营养学特异性的降低而降低。LE对草食动物的影响较小，可能是由于该营养族成员的饮食组成或脂质合成的生理机制不同。这些结果表明，研究人员应使用有关LE的种类，组织和营养最多的特定群体信息，如果无法获得，则应在分析前对一部分样品进行LE以确定其作用。