Parameters of the deleterious mutation process can be estimated using the data on genotypes, phenotypes, or fitnesses. These data can be on long-term evolution, on short-term changes, or on the properties of equilibrium populations. The two most important parameters at the genomic level, the total deleterious mutation rate U and the mutational pressure on fitness P, remain poorly known. Reliable data on the rates of presumably neutral evolution, together with less certain estimates of the functionally important fraction of the genome, suggest that in mammals U > 1. The magnitudes of inbreeding depression in populations of selfers imply U approximately 1 in flowering plants. The straightforward way to estimate P is to assay the decline of fitness in populations with relaxed selection. The relevant data are contradictory, possibly because the results of the measurement of fitness depend strongly on the environmental conditions.