PASTURE PRODUCTION UNDER A SERIES OF PINUS RADIATA-PASTURE AGROFORESTRY SYSTEMS IN SOUTH-WEST VICTORIA, AUSTRALIA

Net pasture production has been monitored over time on an on-going Pin us radiata-pasture agroforestry experiment located at Carngham, Victor ia. Established in 1983 by thinning a 2-year-old plantation, the study contains five treatments (systems): open pasture (no trees), 60 trees /ha, 200 trees/ha wide-spaced, 200 trees/ha 5-row belt and 1360 trees/ ha. Net pasture production was only weakly affected by the proximity o f trees in the 60 trees/ha and 200 trees/ha wide-spaced systems. Howev er, in the 200 trees/ha 5-row belt system, pasture production was ofte n significantly less within the tree zone and at 1.5 m from the tree e dge compared with production at 10.5 m from the tree edge. While there was a trend for greater pasture production at 10.5 m and 18 m from th e tree line than occurred in the open pasture system, the width of pas ture adjacent to the 5-row belts was too narrow (36 m) to demonstrate any shelter benefit on pasture production. From 1990 to 1992, pasture production (kg ha(-1)) for each system was similar. although productio n tended to decrease with increasing tree stocking. Following the cess ation of pruning (1992), pasture production was significantly reduced by increased tree stocking, such that pasture production in the 200 tr ees/ha wide-spaced and 200 trees/ha 5-row belt systems was 60% and 69% and 23% and 67% of open pasture in 1993 and 1994 respectively.