SUMMARY
The knowledge about the architecture of sugarcane root system is limited due to difficulties associated to methodological aspects. The correct way to assess the root system is still not clearly established. This study aimed to compare two direct methods of root biomass evaluation: soil monoliths and root sampling from probes. Three experimental plots (7.2 m2) were randomly established in an area under sugarcane planting on Typic Quartzipsamments located in São Pedro, Sao Paulo state, Brazil. One plant in each experimental plot was selected for root system assessment. The sampling with probes was performed in nine points inside half of the Voronoi polygon and the root density was estimated in nine combinations of sampling points. The root samples collected by the two methods were separate from soil particles by washing in a set of sieves. Afterwards, the roots were oven-dried and weighed. Significant differences (p<0.05) in the density root were found among the soil layers in each treatment. A trend of decreasing root density with soil depth was found in both methods. The probes method tended to underestimate the amount of root found in the soil top layer. Only two sampling combinations with probe indicated to be equivalent to the monolith method.