The spatial pattern and probability distribution of Mexican bean beetle, Epilachna varivestis Mulsant, egg masses were studied in four dry bean, Phaseolus vulgaris L., fields near Scottsbluff, NE, during the 1994 and 1995 growing seasons. Sampling was conducted in a regular and uniform grid to guarantee coverage of the field. The experimental unit in all fields was 0.30 row-m, and five measurements were made at 0.76-m intervals. A total of 12,290 locations was sampled. Geostatistics and discrete statistics were used to describe the egg mass distribution. Regression was used to detect and separate macroscale trends from the microscale variation. The presence of macroscale variation indicated a significant edge effect with ovipositing females moving into dry bean fields from their overwintering sites. The microscale variation estimated from the estimated residuals from regressions to estimate trends was studied using semivariograms for all fields. Semivariograms strongly indicate that Mexican bean beetle eggs are randomly distributed across the field and that this random distribution holds across 10-fold differences in population densities. As a result of the significant macroscale trends observed, we conclude that egg mass densities may be greater in areas of the field near edges adjacent to beetle overwintering sites. The probability distribution that best fit the data was the negative binomial. Our results illustrate the importance of scale in discussing and characterizing distribution. Although there is evidence of edge effect at the field level, there is no evidence of spatial dependence between egg samples at the sampling region level. At an even lower level, the egg mass itself, eggs are aggregated.