Procedures for monitoring larval populations of the Douglas-fir tussock moth and the western spruce budworm are recommended based on many years experience in sampling these species in eastern Oregon and Washington. It is shown that statistically reliable estimates of larval density can be made for a population by sampling host trees in a series of permanent plots in a geographical monitoring unit. The most practical method is to estimate simultaneously densities on a plot of both insect species by the nondestructive sampling of foliage on lower crown branches of host trees. This can be done either by counting all larvae on sample branches or by estimating the frequency of occurrence of a selected threshold number of larvae in samples. Statistics are given on the expected within- and between-plot variances and the number of sample plots needed in different sized monitoring units. In large monitoring units, plot densities of tussock moth and budworm larvae usually are not normally distributed, but they can be normalized by logarithmic transformation to predict the probability of subpopulations of any given density occurring somewhere in the unit. It is urged that sampling methods be consistent and that monitoring be done annually to accumulate continuous databases that reflect the behavior of defoliator populations over a long period.