This note concerns theoretical and experimental studies of multifactorial traits, especially fitness and its components, in which (1) the loci studied are only a subset of those relevant to the character of interest and (2) the genotypes at the loci studied are in nonrandom association (linkage disequilibrium) with genotypes at the loci ignored. In these cases, phenotypic differences between cis and trans double heterozygotes can occur even though no linkage phase effects are inherent in the genetic determination of the trait. Examples are drawn from both theoretical and experimental work, and implications in both areas are discussed.