The sequestration of soil organic carbon (SOC) in agricultural systems has the potential to partially mitigate climate change by removing carbon dioxide (CO2) from the atmosphere. Furthermore, incorporating forage legumes into a grazing system may aid in carbon (C) storage, as mature pasture soils are typically nitrogen (N) limited. In Queensland, Leucaena leucocephala, a leguminous shrub is often incorporated into grazing systems to provide a source of protein for beef cattle production. Additionally, leucaena is known to improve the N fertility of grazing lands. However, there is limited research on the C and N dynamics beneath leucaena-grass pastures, particularly at depth in the soil profile (>0.15 m). Likewise, there is a need to quantify the size and turnover of physically uncomplexed soil organic matter (SOM) and elucidate the mechanisms operating to limit decomposition through physical protection, chemical stabilization and biochemical recalcitrance in grazed leucaena pastures. This PhD investigated the sequestration (t/ha/yr), turnover and stabilization of SOC and total soil N (TSN) beneath seasonally grazed, leucaena-grass pastures of subtropical Queensland, Australia. Fieldwork was conducted at Brian Pastures Research Station, 17 km east south-east of Gayndah on a chronosequence of leucaena stands (9, 22, 34, 40 years) and paired grass sites (0 years). The research was undertaken at a number of scales, starting at the site level and progressing to aggregate, particulate and chemically resistant fractions. Soil C and N dynamics were quantified vertically to 1 m and spatially in relation to leucaena hedgerows. The origin and quantity of C (C3-leucaena or C4-grass) and N stocks were determined with the aid of stable isotope analysis (δ13C, δ15N). The first study investigated TSN and δ15N dynamics in a chronosequence of leucaena-grass pastures. Leucaena leaves recorded a high N concentration (2.8%), low C:N ratio (16:1) and slightly enriched δ15N value (0.7‰) compared to atmospheric N2 (~0‰). Based on a mixing model, it was estimated that 73% of leucaena N was derived from diazotrophy with the remaining 27% sourced from the mineral N pool. Total soil N stocks were 45% higher beneath leucaena rows compared to grass sites in the upper 0.1 m and this trend was significant to 0.6 m in the profile. Leucaena stand age was also important. Total soil N stocks increased by 32% in the 0-0.3 m zone over time, reaching a maximum in the 40 year stand. The N dynamics beneath row and mid-row strips suggest a tight cycling of available N. Soil δ15N values were enriched compared to plant values and displayed a marked increase then decline with depth in the profile. This trend is typical for a N saturated legume system where labile foliage is rapidly mineralised by soil microbes leading to a zone of δ15N enrichment. This study highlighted the importance of incorporating N2 fixing legumes into grazing systems to increase plant available N and minimise pasture rundown. Additionally, δ15N values provided a useful insight into biological N fixation and soil N dynamics in tree-legume pastures. The second study investigated SOC and δ13C dynamics in a chronosequence of leucaena-grass pastures. Pasture type affected SOC stocks within the upper 0.2 m (grass 22 years) highlights the need to maintain soil fertility in grazed pastures to ensure the maintenance and further accretion of SOC. The third study investigated light fraction (LF) SOC and δ13C dynamics in a chronosequence of leucaena-grass pastures. Light fraction C stocks were significantly higher beneath leucaena rows compared to mid-rows and paired grass sites in the upper 0.1 m. A similar trend was evident at 0.2 m. On average, the LF formed 20% of SOC stocks in the upper 0.1 m of leucaena rows. This decreased to 9% in the 0.1-0.2 m zone. Leucaena stand age also affected LF-C stocks, with the highest values noted in the 22 year old pasture, followed by the 34, 40 and 9 year old stands. The LF δ13C values in leucaena stands were enriched by 4.3-7.9‰ compared to leucaena foliage (-28.7‰). On average, 37% of LF-C was derived from C3 plants in the upper 0.1 m of leucaena rows. By comparison, leucaena mid-row and grass sites contained just 7 and 9% of C3-C in the LF. Irrespective of pasture type or depth, the majority of LF-C (>60-97%) was derived from C4 pasture-grass species. The LF is typically considered a fast-turnover pool of labile C with a short mean residence time (MRT). However, this study suggests that LF-C can form a significant portion of total SOC, especially beneath leucaena-grass pastures. Furthermore, if leucaena productivity is maintained and the system remains at equilibrium the LF may represent a longer-term, albeit still labile reservoir of SOC. The fourth study investigated the mechanisms operating to stabilize SOC in a chronosequence of leucaena-grass pastures. Leucaena stands were found to promote aggregation and in particular the formation of coarse macroaggregates in the upper 0.3 m. This trend was attributed to (i) higher levels of free particulate organic matter (POM) beneath leucaena stands that could act as a nucleating and binding agent in the formation of macroaggregates and (ii) higher soil N levels which may promote bacterial and fungal activity and thus contribute to aggregate stability. There was no significant difference in the concentration of C within different aggregate size fractions suggesting that (i) traditional aggregate hierarchy was less relevant in this Vertisol and (ii) the majority of C was associated with the silt and clay sized fractions. The stock of SOC was significantly higher beneath leucaena compared to grass in the upper 0.2 m. Likewise, macroaggregates contained a greater SOC stock than microaggregates, principally as a function of the soil mass distribution rather than a greater C concentration per se. Irrespective of pasture type, the majority of SOC carried a C4 signature. In general, microaggregates contained a higher proportion of occluded POM (oPOM); however this fraction was relatively depleted in C and N compared to other aggregates and formed a smaller proportion of total soil mass. While the majority of SOC was stored in more labile macroaggregates, this size class may also slow the rate of plant residue decomposition and facilitate the formation of more stable microaggregates. Hence, aggregates appeared to both physically protect SOC per se and also act as an intermediate phase before more recalcitrant SOC was stabilized by clay and silt particles. This PhD provided insight into the SOC and TSN dynamics operating within the seasonally grazed leucaena-grass pastures at Brian Pastures Research Station. The most significant finding was that SOC stocks were tightly coupled with soil fertility and in particular the N status of grazed pastures. Hence incorporating leucaena into a grazing system has the potential to significantly enhance SOC sequestration. While certain trends are likely to be less defined in commercial leucaena-grass pastures, this thesis and the chapters it contains still have broad implications for the long-term management of SOC in subtropical pastures and grazing lands.