ABSTRACT Paleodiet research traditionally interprets differences in collagen isotopic compositions (δ13C, δ15N) as indicators of dietary distinction even though physiological processes likely play some role in creating variation. This research investigates the degree to which bone collagen δ13C and δ15N values normally vary within the skeleton and examines the influence of several diseases common to ancient populations on these isotopic compositions. The samples derive from two medieval German cemeteries and one Swiss reference collection and include examples of metabolic disease (rickets/osteomalacia), degenerative joint disease (osteoarthritis), trauma (fracture), infection (osteomyelitis), and inflammation (periostitis). A separate subset of visibly nonpathological skeletal elements from the German collections established normal intraindividual variation. For each disease type, tests compared bone lesion samples to those near and distant to the lesions sites. Results show that normal (nonpathological) skeletons exhibit limited intraskeletal variation in carbon‐ and nitrogen‐isotope ratios, suggesting that sampling of distinct elements is appropriate for paleodiet studies. In contrast, individuals with osteomyelitis, healed fractures, and osteoarthritis exhibit significant intraskeletal differences in isotope values, depending on whether one is comparing lesions to near or to distant sites. Skeletons with periostitis result in significant intraskeletal differences in nitrogen isotope values only, while those with rickets/osteomalacia do not exhibit significant intraskeletal differences. Based on these results, we suggest that paleodiet researchers avoid sampling collagen at or close to lesion sites because the isotope values may be reflecting both altered metabolic processes and differences in diet relative to others in the population. Am J Phys Anthropol 153:598–604, 2014. © 2013 Wiley Periodicals, Inc.