The study of nocturnal dreaming has been mainly a matter of correlational studies. This article brings together and reviews those scattered studies in which an attempt was made to actively influence dreaming. Despite a great diversity of methods, aims, and results, it is apparent that dreaming can be experimentally manipulated, and thus it may be possible to develop a functional experimental approach in which dreams become more immediate and observable aspects of behavior, rather than only memories.Scientific research is carried out with two basic kinds of procedures. One is the observational or correlational approach, where we observe "what happens naturally" and try to make sense of it by finding relations or correlations between various aspects of our observations. This approach leads to statements of the order, "subjects classified as Sensitizers report recalling dreaming significantly more frequently than those classified as Repressers" on the MMPI (Tart, 1962). The other approach is the functional method, where we actively manipulate one factor or variable and observe the effect of some dependent variable. This approach leads to statements of the order, "A large dose of alcohol reduces the amount of EEC stage 1-REM time and inferred dreaming in normal Subjects . . ." (Gresham, Webb, & Williams, 1963). Both approaches are used in most fields of science, as they are complementary.Research into the nature of dreaming has been, by and large, a matter of correlational research (ignoring for the moment the last decade's research using the electroencephalo-1 This investigation was supported (in part) by a Public Health Service postdoctoral fellowship (1-F2-MH-14,622-01) from the National Institute of Mental Health. I also wish to thank Leslie M. Cooper and Ernest R. Hilgard for their helpful comments during the preparation of this manuscript.