Numerous crop species and other economically important plants are amenable to techniques of regeneration from tissue culture and to transformation technique. "Somaclonal variation" occurring among callus-derived transformed and non-transformed somaclones, as well as traits introduced by transformation provide genetic variation. Genetic variation introduced by transformation is certainly novel, while that due to somaclonal variation per se may be novel. Our knowledge of somaclonal variation is inadequate, necessitating empirical inferences about breeding procedures that might be most proficient. Strategies that might be most productive are examined for crops having different types of reproductive systems, and whether or not regeneration techniques are applicable to elite germplasm or only a few specific genotypes. Consideration is given to the fact that variants might exhibit "qualitative" or "multigenic" traits, beneficial and(or) deleterious. A tabular analysis showed that in most instances backcross breeding is requisite. The form of backcross breeding that would be most desirable, however, is highly dependent on both the number of variant loci and the "gene action" governing the new traits. Genetic information on these aspects of somaclonal variation is direly needed from the practical standpoint of designing suitable breeding procedures. Backcross-inbred development is recommended as one means of gaining such information in the course of applied breeding. Being able to regenerate a wide variety of genotypes seemed relatively unimportant, though in some instances it could reduce the amount of backcrossing. However, if variation of multigenic traits is extensive, then extensive backcrossing will be requisite. Backcross inbreds, or perhaps products from intermated backcrosses, are perceived as optimal for cotton. Where feasible, somaclones and somaclonal derivatives should be used as pollen parents during backcrossing. Until the possibility of recessive control of somaclonal traits can be eliminated, early generation small-scale self-progeny testing is recommended as a selection aid.