Understanding the characteristics of short rotation woody crops (SRWC), like shrub willow, that affect feedstock quality and its variability is a priority as this source of biomass is expanded. Because of its relationship to ash content, the allotment of bark and wood within the stem is often highlighted as a factor impacting feedstock quality. Bark-to-wood ratios are frequently determined in willow by measuring this ratio for the centroid and assuming it represents the entire stem. The objectives of this study were to determine how site and genetic factors influenced the allometry of whole stem bark (WSB%) content on a dry weight basis and if the centroid bark percentage (CB%) on a dry weight basis adequately represents WSB%. A multiple linear regression approach was used to model WSB% and centroid bark content (CB%) using diameter, cultivar, and site factors. Five candidate models were evaluated ranging from parsimonious to complex. The simplest models estimated WSB% from CB% (R 2 = 0.76) and stem diameter (R 2 = 0.44). The most complex model included all factors and had an R 2 of 0.90. Two key relationships demonstrated by several models are that (1) CB% adequately predicts but underestimates WSB%, particularly for larger stems, and (2) WSB% increases as diameter decreases. However, empirical models can be enhanced with the inclusion of diameter, site, and cultivar information. The overall approaches can be useful for making relative comparisons between cultivars and sites. In a 3-year-old stand of willow stems less than 20 mm had high WSB%, the maximum being 23.8 %, but accounted for only 5 to 15 % of the total biomass. Large diameter stems accounted for the majority of the total and bark biomass, but the WSB% in large diameter stems is as low as 11.5 %. WSB% ranged from 12.9 to 14 % across the two cultivars at the stand level.