This work corroborates the hierarchical conceptual model for soil aggregate structure presented by Tisdall and Oades (1982), extends it to North American grassland soils, and elaborates on the aspects relating to the influence of cultivation upon losses of soil organic matter. Aggregate size distributions observed for our soils are very similar to those of Australian soils, indicating that the microaggregate‐macroaggregate model may hold for a wide array of grassland soils from around the world. The use of two wetting treatments prior to sieving demonstrated that the native sod soil had the same general structural characteristics as cultivated soil but the macroaggregates were more stable. When slaked, native and cultivated soil present in the microaggregate size classes had less organic C, N, and P than the soil remaining as macroaggregates, even when expressed on a sand‐free basis. Moreover, the C/N, C/P, and N/P ratios of microaggregates were narrower than those of macroaggregate size classes. Much more C and N was lost than P under the conditions of this study. Element‐specific differences in microbial catabolic processes vs. extracellular enzyme activity and its attendant feedback controls are postulated to account for this difference. When the macroaggregates were crushed to the size of microaggregates, mineralizable C as a percent of the total organic C was generally greater for macro‐ than microaggregates early in the incubation for the cultivated soil and throughout most of the incubation for the native sod soil. Mineralizable N as a percent of the total organic N was greatest in macroaggregates even when the macroaggregates were not crushed. The macroaggregate‐micraggregate conceptual model is applied to help explain accumulation of soil organic matter under native conditions and its loss upon cultivation.