Fluctuating asymmetry (FA), used as an indicator of developmental stability, has long been hypothesized to be negatively correlated with genetic variability as a consequence of more variable organisms being better suited to buffer developmental pathways against environmental stress. However, it is still a matter of debate if this is due to metabolic properties of enzymes encoded by certain key loci or rather to overall genomic heterozygosity. Previous analyses suggest that there might be a general difference between homeo- and poikilotherms in that only the latter tend to exhibit the negative correlation predicted by theory. In the present study, we addressed these questions by analysing roe deer (Capreolus capreolus) from five German populations with regard to FA in metric and non-metric skull and mandible traits as well as variability at eight microsatellite loci. Genetic variability was quantified by heterozygosity and mean d2 parameters, and although the latter did not show any relationship with FA, we found for the first time a statistically significant negative correlation of microsatellite heterozygosity and non-metric FA among populations. Because microsatellites are non-coding markers, this may be interpreted as evidence for the role of overall genomic heterozygosity in determining developmental stability. To test if the threshold character of non-metric traits is responsible for the metric vs non-metric difference we also carried out calculations where we treated our metric traits as threshold values. This, however, did not yield significant correlations between FA and genetic variability either.