Abstract. As part of the Cicerone Project's whole-farmlet experiment on the Northern Tablelands of New South Wales, Australia, the fat scores and reproductive performance of ewes were measured to assess the effect of different management systems on these important production parameters over time. The three farmlets (each of 53 ha) included one (farmlet B) subjected to 'typical' district management consisting of moderate levels of inputs and a target stocking rate of 7.5 dse/ha, with flexible grazing management across eight paddocks. A second farmlet (A) was managed in a similar fashion to farmlet B with respect to number of paddocks and grazing management, but modified by high rates of pasture renovation and higher levels of soil fertility, with a target stocking rate of 15 dse/ha. The third farmlet (C) was managed at the same level of moderate inputs as farmlet B but employed intensive rotational grazing over 37 paddocks and also had a high target stocking rate of 15 dse/ha. The experiment was conducted over 6.5 years from July 2000 to December 2006. In spite of the fact that target levels of stocking rate were chosen at the beginning of the experiment, stocking rate, together with fat scores and reproduction were treated as emergent properties of each farmlet system. Joining took place in April-May and lambing occurred in September-October of each year. Over the first 2 years of the experiment, there were few differences among farmlets in ewe fat scores or reproductive performance. From 2003 onwards, while the percentage of ewes pregnant was similar between farmlets, the average proportion of multiple births (ewes scanned in late July, with twins) was 30%, 16% and 12%, respectively, on farmlets A-C. However, lamb losses were greater on farmlet A, with average lamb mortalities recorded on farmlets A-C of 29%, 10% and 19%, respectively. Over the duration of the experiment, ewes on farmlets A and B were more often above a fat score level of 3, and less often below 2.5, than were ewes from farmlet C. Differences among farmlet ewes in fat score were found to be significant in 7 of the total of 13 assessments over the duration of the experiment. A generalised additive model applied to whole-farmlet data showed that green digestible herbage, legume herbage, stocking rate, the amount of supplement fed and especially the proportion of each farmlet grazed at any one time all influenced fat scores of ewes. While fat scores and conception rates tended to be highest on farmlet A, farmlet B had slightly better reproductive outcomes due to less lambing losses, whereas ewes on farmlet C tended to have somewhat lower fat scores and levels of reproduction. These farmlet-scale findings highlighted the importance for livestock managers to focus not only on grazing management, stocking rate and stock density during lambing, but also on the availability of sufficient green, and especially legume herbage, and the difficulty of overcoming a deficit in quality herbage with supplementation.