Male zebra finches, Taeniopygia guttata castanotis, were normally-raised by zebra finches or were cross-fostered to Bengalese finch, Lonchura striata, foster-parents until 40 days of age. Following isolation until day 100, half the birds in each group were housed with a zebra finch female for seven days, isolated for three days and then housed with a Bengalese finch female for seven days. The other birds were exposed to females in the reverse order. Subsequent double-choice tests showed that all the normally-raised birds preferred zebra finch females whereas the preferences of cross-fostered males depended on the order of exposure to the two females: those exposed first to a Bengalese finch female preferred Bengalese finch females whereas of those exposed first to a zebra finch female, some preferred zebra finches, some preferred Bengalese finches and some showed no marked preference for either female. In order to examine the question of why the latter group showed such marked individual variation in their sexual preferences, a further group of males were cross-fostered to Bengalese finches and exposed to a zebra finch female and then to a Bengalese finch female and their behaviours were observed from day 21 until day 40 and for the two, seven-day periods with the females. The results show that, when comparing brothers within clutches, the one that begs and is fed more by by its foster-parents develops a stronger preference for Bengalese finch females and that the more song phrases a male directs to the zebra finch female during the first seven-day period, the stronger the sexual preference for zebra finch females in the double-choice tests. Hence, our results confirm and extend those of IMMELMANN et al. (1991) and KRUIJT & MEEUWISSEN (1991) that sexual imprinting may be a two step process. As a first step, information about the parents is learnt during a sensitive period early in life. In a second step, this information has to be tested for its validity for the selection of a sexual partner during first courtship encounters. It is this second step where the previously stored information is stabilized in memory. Giving conflicting information during the first and the second step, one can show that interactions between the young male and its parents as well as with its first sexual partner influene the final preference it shows in subsequent double choice tests.
