A sampling study was conducted in sugarcane to estimate infestation levels of the sugarcane borer (SCB), Diatraea saccharalis (F.). A single sample consisted of five stalks per plant from five plants spaced 3 m apart. A mean of 7 min was required for one scout to take one sample in cane 0.6–1.2 m tall; 16 min was required in cane 3.0–3.6 m tall. Variance to mean ratios, Green’s coefficient of dispersion, and a simple linear regression of log10( s 2)against log10( x ¯) showed that SCB populations usually were aggregated in sugarcane. Number of SCB larvae per sample followed the negative binomial distribution. In 0.4- to 0.8-ha areas of sugarcane, aggregation was not detected at low population densities but was pronounced at high densities. Estimated SCB infestation levels (mean number of borers per 25 stalks × 100) ranged from 0.4 to 48.0% in the different fields sampled. Based on 90% CL and a mean of nine samples per 0.4- to 0.8-ha area, accuracy of estimates for SCB infestation levels ranged from ±0.8 to ±26.6% with an overall average of ±7.4% based on simple variance or ±7.1% based on a regression estimate of variance. At infestation levels <10%, estimates of SCB infestation levels were generally within ±4.5% of the true levels based on simple variance or ±3.9% based on regressed variance. Accuracy of infestation level estimates generally decreased as SCB infestation levels increased.