A better understanding of the soil aggregation process is needed to address a variety of concerns, including soil quality and erosion, agricultural sustainability, soil C sequestration, the mobility of hazardous chemicals and remediation of contaminated sites. We used data from a chronosequence of tallgrass prairie restorations and a path analysis approach to evaluate how the lengths of two dia size classes of fibrous roots, the length of external mycorrhizal hyphae, microbial biomass C, hot-water soluble carbohydrate C and soil organic C interact to promote the stabilization of soil aggregates. The measured binding agents accounted for 88% of the variation in macroaggregates >212 μm diameter and goodness-of-fit indexes indicated a good practical fit of the path model to the data. The restoration of macroaggregate structure in this system was apparently driven by the direct and indirect effects of roots and external hyphae, with lesser relative contributions by the three measured C pools. Although the two root size classes had similar direct effects on the percentage of macroaggregates, their indirect contributions differed substantially. Fine roots (0.2–1 mm diameter) exerted considerable indirect effects via their strong influences on external hyphae and microbial biomass C. Very fine roots (<0.2 mm dia) made a stronger contribution to soil organic C than fine roots, but their overall indirect contribution to aggregation was minimal. In addition, the relative importance of each binding agent varied for different size fractions of macroaggregates and generally supported the hypothesis that the effectiveness of various binding mechanisms depends on the physical dimensions of the binding agents relative to the spatial scales of the aggregate planes of weakness being bridged.