A long term (1989-1994) experiment with sugar cane (RB 739735) was carried out, in Linhares, ES, Brazil, in order to evaluate the effect of two systems of sugar cane harvesting: the first with previous burning (burned cane), and the second with crop residue remaining on the soil surface (green cane), on some physical properties of a tableland Yellow Podzolic soil. The statistical design was in randomized blocks, replicated six times. After six years of cultivation, a process of soil degradation was observed in the burned cane system, with decreasing of aggregate mean diameter and increasing of soil bulk density at the 0-5 cm soil depth. There were also differences in total porosity and porous distribution at the 0-5 cm soil depth as a consequence of the harvesting system. The soil water infiltration, measured in the field, was greatest in the area under the green cane treatments. The same result was not observed when measuring saturated hydraulic conductivity with the steady state permeameter.
