Sugarcane borer, Diatraea saccharalis (F.), damage is complex and can be measured by multiple variables, i.e. visual damage and percentage of internodes bored. Cluster analysis is a statistical technique to summarize similarities among many objects that have multiple measured variables and was used to assess early sugarcane, Saccharum interspecific hybrids, line trials for resistance to sugarcane borer. A selection series of 44 varieties was evaluated by their response to sugarcane borer; four times under artificial infestation and once under natural infestation. The composition of varieties partially changed from year-to-year as low-yielding or poor-performing varieties were dropped from the program and as new varieties were added. Visual injury estimates and percentage of bored internodes at harvest were subjected to traditional analysis of variance with subsequent multiple comparisons and to hierachical cluster analysis. Classification of varieties by analysis of variance was made difficult by the multivariate data collected and multiple comparison test assigning variety means to overlapping groups. With cluster analysis, information from multiple variables can be used to assign varieties to discrete classes. Labels (resistant, susceptible) were placed on these discrete classes with the aid of commercial varieties of known to reaction sugarcane borer included as industry standards. The four-cluster analysis explained more variation among the varieties than the three-cluster and including an additional cluster group provided insight into the mechanisms of resistance to sugarcane borer. Groups with high visual injury ratings and low percent internodes bored may lack tolerance to sugarcane borer feeding. Conversely, groups with low ratings and high percent internodes bored may possess tolerance to sugarcane borer.