The influence of sexual experience gained prior to castration, on the degree of retention of sexual behavior after removal of the testes was investigated in an experiment involving 15 male cats raised from weaning in the laboratory. During a series of 20-minute tests, nine males were permitted between 11 and 81 intromissions per animal (Group I). Six males were permitted only minimal sexual experience (Group II). Four of these cats were tested until the first mount was performed, at which time they were gently separated from the female. The other two males were not permitted any sexual behavior. After castration the males of both groups were tested weekly until sexual behavior had ceased to appear, or in the case of "long persisters", until a low level of sexual behavior was observed. After castration, the maximally experienced males of Group I were superior in their performance of sexual behavior to those of Group II when this superiority was measured by (1) presence or absence of intromission, (2) frequencies of intromissions and mounts per test, and (3) the number of weeks after castration that any elements of sexual behavior were observed. All of the experienced males performed at least one intromission after castration, and when these ceased, most of these males continued to mount for many weeks or months before becoming unresponsive to the receptive female. In contrast to this, only two of the six minimally experienced males of Group II achieved intromission after castration. These appeared in the first postoperative test, and none were observed thereafter. These two males continued to mount the female for several months. The remaining four males showed no interest in the female except for a single male that performed two mounts. The results of this experiment led to the conclusion that prior sexual experience functioned to facilitate the continuation of sexual behavior after castration. The pattern of sexual behavior having been differentiated and practiced by the males of Group I during the preoperative tests, it could be elicited after castration by stimuli present in the testing situation, especially the receptive female, despite the low level of testicular hormone. In the absence of sexual experience and under the influence of the depressing effects of castration on sexual behavior, most of the males of Group II were unable to respond to stimulation provided by the receptive female, and failed to develop a sexual pattern or even perform the introductory elements of the mating pattern. An analysis of the decline in sexual behavior after castration of the maximally experienced males (Group I) revealed a steady decrease in the number of males showing intromissions, until by the third month only two males were performing this behavior regularly. Mounting behavior also declined after castration, both as to the average frequency of mounts per test and in the number of males performing this behavior. The sequence in which elements of the sexual pattern disappeared was consistent. Intromission dropped out first. This was followed by a rise in the number of long mounts. Eventually the males performed only short mounts without stepping or pelvic thrusting. These continued for varying periods and eventually declined and became infrequent or nearly absent in most of the animals. An analysis of the postcastrate performance of individual males showed that they could be divided into three classes as follows: Type A - Following an initial decline in frequency of intromission per test, sexual behavior including intromissions persisted for 8 months to 31/2 years. Mounting continued indefinitely. T y p e B - Intromissions ceased between the 2nd and 3rd months after castration, but mounting behavior persisted for many months thereafter. T y p e C - All sexual behavior ceased shortly after castration. Further analysis of mounting and intromissive behavior of three males of type A revealed a gradual loss in the ability of these animals to achieve and maintain adequate erections. As a result of this, four periods in the postcastrate history of these animals can be recognized as follows: Period A - Lowered frequency of intromission. Period B - Mounts leading to intromissions became prolonged as erections weakened. Period C - Long mounts not leading to intromission became more frequent. Period D - Intromissions no longer achieved, but long mounts still frequent. The three different patterns of decline in sexual behavior described above, which were derived by analysis of the performance of individual males were compared with results previously reported for mammals, where the performance of the group as a whole was the chief consideration. A number of difficulties arose when we attempted to interpret differences among these three modes of decline in sexual behavior in terms of the concept "levels of sexual arousal". These difficulties were discussed and an alternative interpretation of the data was presented. Thus, we attribute the loss of intromissive behavior by the males of types A and B to failure of these animals to develop adequate erections resulting from morphological and physiological changes occurring primarily in the penis. Likewise, the inability of the males to sustain mounts probably resulted from a reduction in afferent stimulation normally derived from contact with the female. The decline in general responsiveness to the receptive female characteristic of type C males and the more or less complete loss of the concomitant anticipatory behavior related to the test situation, were attributed to broader physiological and psychiological changes following withdrawal of hormone. Even with respect to these latter changes, the concept "sexual arousal" appears to be inadequate as a working hypothesis. Emphasis was placed on the need for a more analytical approach which would incorporate research on physiological mechanisms, experiential factors and ontogenetic processes.