The difference between serum levels from 36 captive and 56 free‐ranging polar bears (Ursus maritimus) for 25‐hydroxyvitamin D (25‐OH‐D) was found not to be significant (mean ± SD = 348 ± 215 nmol/L [captive], 360 ± 135 nmol/L [free‐ranging], t = 0.30, df = 52.8, P = 0.76), whereas the difference for retinol and α‐tocopherol was significant (retinol, 1.37 ± 0.67 μmol/L [captive] 1.89 ± 0.63 μmol/L [free‐ranging], t = 3.88, df = 72.4, P <0.001, α‐tocopherol, 18.56 ± 18.56 μmol/L [captive], 48.76 ± 13.92 μmol/L [free‐ranging], t = 7.85, df = 61.9, P < 0.001). Due to the high fat content in the polar bear diet, seal blubber may be the source of these fat‐soluble vitamins. Six skin biopsies were analyzed from captive polar bears at the Denver Zoological Gardens for 7‐dehydrocholesterol levels and found to contain 0.11 ± 0.03 nmol/cm2. This finding also helps to support the contention that the source of vitamin D for polar bears may be ingestion and not cutaneous production. Vitamin D content in the milk from one captive sow in the den (0.14 nmol/g) and 10 free‐ranging sows with cubs of the year out on the ice pack (0.0042 ± 0.0073 nmol/g) were also evaluated. It would be helpful to evaluate additional milk samples from denning and non‐denning sows with cubs to see whether vitamin D content varies according to the stage of lactation. Zoo Biol 17:285–293, 1998. © 1998 Wiley‐ Liss, Inc.