Fluctuating asymmetry (FA) as a measure of environmental stress was studied in polar bear (Ursus maritimus) skulls from East Greenland (n = 300, collected during the period 1892-2004) and Svalbard (n = 391, collected during the period 1950-2004). Nine bilateral metric traits in skull and lower jaw were measured twice. The measured levels of FA for each trait were compared between sex/age groups (subadult, adult females, adult males), time periods (¡Ü1960 and >1960), and localities (East Greenland and Svalbard). The time period before 1960 was chosen to represent a period prior to the appearance of organohalogens (PCBs, HCB, HCHs, DDTs, dieldrin, CHLs, PBDEs, mirex) in the Arctic, while the time period after 1960 represents the period when the polar bears have been subjected to such organohalogens. One of the traits had a significantly higher degree of FA (for all sex/age groups) during the first period, while in two other traits, the subadults had a significantly higher FA during the second period. The Svalbard polar bears had a significantly higher FA in one trait than did the bears from East Greenland. This latter trend was reflected in the investigation of a pooling of four of the traits. Levels of fluctuating asymmetry of each trait were furthermore assessed over time on bears born during the period 1950-2004. The polar bears were divided into sex/age groups, but not according to locality. All significant results from these analyses had a negative slope, with FA decreasing over the period of time. Correlation analyses of FA for each trait versus the levels of contaminants in adipose tissue samples was performed for a subsample of the East Greenland polar bears, collected recently (1999-2002). The majority of the significant results of these correlation analyses showed a negative trend in the data, with higher levels of FA corresponding to lower levels of contaminants. Overall, the present study showed no substantial evidence of a relation between levels of FA and organochlorines. Instead, the indications were of subpopulations with generally declining levels of FA over time, suggesting the existence of positive population level effects powerful enough to overrule the negative influence of stress caused by factors such as global warming, pollution, and overharvesting. Or, alternatively, a) that the individuals with the highest levels of stress and thus ditto values of FA are aborted or die as neonates, and thus are not included in the collections of polar bears examined in the present study, b) that FA of metric skulls traits is not a suitable predictor of stress in polar bears, c) that the organohalogen toxic loads are not at high enough levels to elicit a measurable positive response in the FA of the examined polar bears.