This paper reports on variation in leaf area index (L) in five Eucalyptus globulus Labill. plantations in response to application of nitrogen, thinning at age 2 years and variation in climate wetness index (the ratio of rainfall to potential evaporation). Observed L is compared with: (i) L predicted to optimize net primary productivity for a given average annual temperature, annual water use and potential evaporation (L(opt)) and (ii) L calculated as a linear function of climate wetness index (L(eq)). L peaked in fertilized plots at between 4 and 5 years of age or immediately after canopy closure. The value of L from canopy closure to age 8 years was not strongly related to annual rainfall or climate wetness index. At two sites with total soil nitrogen <1.2 mg g(-)(1), L in fertilized plots was about two units greater than in unfertilized plots. This difference persisted until measurements ended in 2004 when the trees were 8 years old. The L of plots thinned to 300 and 600 stems ha(-)(1) at age 2 years recovered quickly and was not significantly different from L in unthinned plots when the trees were 8 years old. L(opt) was a good predictor of the leaf area index of 8-year-old plots of E. globulus when nitrogen and phosphorus were non-limiting (model efficiency (EF) was 0.5). For the same plots, L(eq) underestimated observed L by an average of two units, and the model efficiency was low (-3.25). Data from two nitrogen-limited sites demonstrated that for fertilized plots L(opt) (EF = 0.6) was a much better predictor of L than L(eq) (EF = -3.36). At the same sites, L(eq) (EF = 0.42) was a better model for predicting L of unfertilized plots than L(opt) (-3.59). These results provide evidence that comparing observed L with L(opt) can identify stands limited by factors other than growing climate.