Soil pH decline and net acidification inputs were determined for a long-term crop rotation experiment at Rutherglen in north-eastern Victoria. The rotations utilized were continuous wheat (WW), a 1 : 1 wheat-lupin sequence (WL) and continuous lupins (LL), and each rotation was cropped from 1975-1989. The soil at the site had an initial pH (0.01 mol/LCaCl2) of 6.0 (0-10 cm depth), sandy loam texture, and had a past use of grape vines and then lucerne pasture. The soil pH (0-10 cm) declined for each rotation with time (1977/78-1988/89), decreasing by about 0.8 units for WW and further decreasing with the inclusion of lupin in the rotation. Compared with the WW soil, the WL soil pH was 0.7 and 0.4 units lower at 5-10 cm and 10-15 cm depth and the LL soil pH was 1.0 and 0.8 units lower at 5-10 and 10-15 cm depth. There was no difference in pH between WW and WL below 20 cm depth, but the LL soil had a significantly lower pH to 40 cm depth. Acidification rates were calculated for the period of cropping and for the 3 rotations, with rates of 3.22, 4.11 and 5.26 kmols H+/ha.yr as net acid input for WW, WL and LL rotations. These values represent a rapid rate of soil acidification. The removal of alkalinity in grain accounted for between 15-21% of the overall calculated acidification rate for the 3 rotations. Therefore, it is likely in this cropping system that the acidification largely results from progressive nitrate leaching.