The allometry of axis length, diameter, and taper is described for the trunk, rachis, and rachilla of nonbranching ramets of Aralia spinosa. Significant log-linear relationships were found between length and diameter for all axis categories, and in all cases, scaling was negatively allometric. Linear models best described the relationship between length and diameter for the rachis and rachilla, while a quadratic model best described this relationship for the trunk. During the trunk-building stage, the safety factors for trunk height were size dependent, with larger trunks exceeding their predicted critical buckling height. Taper was described by a linear relationship between diameter and position along the axis for all axis categories. All rachises and rachillas sampled exhibited taper along the length of the axis, however, only 51% of the trunks showed continuous taper. The trunk was less tapered than the rachis, but no differences in taper were found between the trunk and the rachilla, or the rachis and the rachilla. In unbranched ramets the large bipinnately compound leaves occupy the space normally occupied by lateral branches. We suggest that the rachis and rachilla are functionally equivalent to branches, that is, acting as axes of exploration and exploitation of the environment.

During the crown-building phase, the mechanical architecture of the trunk of Aralia spinosa exhibits considerable ontogenetic variation. All trunks were tapered along their length, and taper was dependent on both ramet size and age; older, larger trunks were more tapered than younger, smaller trunks. Trunk specific gravity, % bark, wood, and pith exhibited considerable inter- and intra-ramet variation. Specific gravity increased with both increasing ramet size and age, and declined acropetally in the majority of ramets sampled. Wood specific gravity was generally unrelated to ramet size, age, or position along the length of the trunk. Percent wood increased while % pith decreased with increasing ramet size and age. There was no relationship between % bark and either ramet size or age. Both % bark and % wood tended to decline acropetally, while % pith increased acropetally. On average, 47% of the variation in specific gravity could be attributed to % wood, while 77% could be attributed to % pith. Percent bark accounted for only 14% of the variation in specific gravity. We suggest that the relatively pithy trunk of Aralia spinosa (average range: 4-15%) allows for rapid height growth, but imposes severe constraints on crown architecture and the maximum size attainable by this species.