Technical Abstract: Biomass partitioning to tree components is linked to tree function, ecosystem C pools and timber production. Patterns of biomass partitioning have been extensively studied in loblolly pine (Pinus taeda L.) plantations and natural stands, but not in silvopastoral practices which are being tested and promoted in southern United States. We conducted biomass harvests in 18 year-old loblolly pine growing in single, double and quadruple-row configurations at 2.1-m between trees in the row, and initial densities of 308, 568 and 932 trees ha-1. In the double and quadruple-row configurations, rows were 2.4 m apart, and alleys were 14.6-m wide in all cases. Trees were felled from single, double (east and west orientations) and, the exterior east (4-1) and one of the interior (4-2) quadruple rows, and separated into bole, branches, twigs (< 2 cm in diameter) and foliage. In total, sixty trees within the 25-29 cm DBH class were sampled (5 row/configurations x 3 trees per row x 2 pruning regimes x 2 replicates). Configuration affected biomass partitioning to boles (P < 0.01) which ranged, on average, from 60.0 % of total tree biomass in quadruple rows to 52.4 % in single rows, and foliage (P < 0.001) which varied from 12.5 % in single rows to 8.1 % in quadruple rows. Row affected biomass partitioned to boles (P < 0.01), and branches (P < 0.02). Proportion of total biomass in boles was highest in 4-2 trees at 62.8%. Single-row trees had 30.2 % of their biomass in branches versus 23.9 % in 4-2 trees. Bole DBH and available ground space per tree (as covariates) did not affect biomass partitioning as neither did pruning, although pruning markedly improved log quality. Timber production for the silvopastoral stand would likely be overestimated by 37-57% while woody biomass production would be overestimated by 6-10% when yields are estimated from forest simulators for loblolly pine.