Carbon storage by forests has potential for contributing to ‘Kyoto Protocol’ greenhouse gas emission reduction targets, but evidence about C-storage and loss below ground is conflicting. The study addresses why soil carbon stocks are increased by land use change from forest to pasture, but are reduced by planting conifer trees, though not broadleaf trees, onto prior pasture. Can species differences in fine root production and litter input play a role? Addressing that hypothesis, a 1-year pot experiment was established as a model system in a glasshouse. Two tree species, pine (Pinus radiata) and Blue Gum (Eucalyptus globulus), and two grass species, Kangaroo Grass (Themeda triandra) and Wallaby Grass (Austrodanthonia racemosa), were sown in pots of soil taken from a native pasture with Kangaroo Grass dominant or from an adjacent pine plantation forest. After 3 months, half of the grass pots were defoliated monthly to 7 cm above ground to test for any cutting effect. Fine root production and turnover was monitored via minirhizotrons, with a destructive harvest after 1 year. Fine root oven dry mass at the end of the year varied between species: Kangaroo Grass (17 g/pot), pine (13 g/pot), Blue Gum (8 g/pot), Wallaby Grass (4 g/pot). Cutting significantly reduced fine root production in Kangaroo Grass but not in Wallaby Grass. From minirhizotron monitoring, 70% of the fine root length produced by pine during the year had disappeared, presumably by decomposition, before the final harvest. The equivalent loss for Wallaby Grass was only 26%, for Kangaroo Grass 36%, and for Blue Gum 45%. Despite the faster fine root disappearance under pines than under Kangaroo Grass, soil C declined under pines but increased under Kangaroo Grass as found in the field. Thus the experiment did not support the idea that a lower dead fine root production was the source of decline in soil C under pine. There was only a weak correlation between soil C change and the net amount of live fine root mass produced. However, the soil carbon changes in this study were positively correlated with live fine root length density in the soil. Kangaroo Grass maintained large lengths of very fine roots while pines produced a small length of thick fine roots. Accordingly, it is hypothesized that the increase of soil carbon under Kangaroo Grass compared with pine may be caused more by the activity of live fine roots than by decomposition of fine root mass to humus. This hypothesis, formulated from the model experimental system, needs to be evaluated for field sites where pines have been planted onto pastures.