Carotenoid pigments have become a central subject of research on animal signaling systems during the past decade. Thus, measurement of plasma carotenoid levels is widespread in the literature. Many plasma biochemical parameters tend to vary with the hour of sampling, which may be an important source of sampling error. However, little is known about this kind of variation for circulating carotenoids. With a sampling protocol that allowed us to separately analyze the effect of sampling hour and repeated blood extraction at the within-individual level, we evaluated the effect of these two parameters on plasma carotenoid concentration in captive red-legged partridges (Alectoris rufa). The hour of sampling did not show a significant influence on carotenoid variability. Nevertheless, carotenoid levels significantly decreased as a result of repeated sampling. In fact, carotenoid variability was positively related to hematocrit values, which also decreased throughout the experiment. Furthermore, the effect was evident from the second sampling event. These results suggest that blood samples for carotenoid analysis may be obtained at random during the daytime with no serious risk of adding variance to data obtained. However, the effect of hemodilution associated with repeated blood extraction should be carefully considered in those studies involving repeated sampling.