We have developed a model to study the influence of conditioning on sexual partner preference in the rat. In this model, pairing a neutral odor (almond) with copulation to ejaculation produces a subsequent preference to ejaculate with females bearing that odor. We refer to this phenomenon as a conditioned ejaculatory preference (CEP). The present study investigates the nature of the conditioned response that mediates CEP. Given equal mount and intromission distributions, but an unequal ejaculation distribution, we hypothesized that two mechanisms could account for CEP: facilitated ejaculation or selective ejaculation. To test these hypotheses, we examined the effect of omitting the olfactory conditioned stimulus (CS) or applying it to a nonreceptive female. Males trained with the CS paired with copulation (paired-trained males) failed to display evidence of delayed ejaculation when copulating with unscented females. Conversely, paired-trained males displayed CS-elicited copulatory behavior with CS-bearing nonreceptive females. In addition, we re-analyzed the data from the copulatory preference tests of previous experiments for CEP-displaying males. Again, we failed to find evidence of facilitated ejaculation with the scented female. However, the time between the last mount or intromission and ejaculation was increased if either occurred with the scented female. Furthermore, more mounts were directed towards the scented female near the end, but not at other points, of an ejaculatory series. These findings suggest that the paired-trained males attend to the scented female near the point of ejaculation, and are consistent with the hypothesis that CEP is mediated by selective ejaculation.