Experimental photoperiods, presented either once only or repeatedly, were used to assess the oscillatory and hourglass properties of the photoperiodic clock in Japanese quail. Gonadectomized quail on 8-hr daylengths respond to a single skeleton photoperiod consisting of two 8-hr light pulses separated by 2 hr of darkness (i.e., LDLD 8:2:8:6) with a marked increase in secretion rate of luteinizing hormone (LH). This response suggests that the second light pulse interacts with a "photoinducible phase" (phi i) lying some 10-16 hr from "dawn" (start of the first light pulse). If, however, groups of quail maintained on 8-hr daylengths are transferred to continuous darkness (DD), and the position of the phi i is sought by a single 8-hr light pulse applied at various times on the first or third day of DD, then an increase in circulating LH is, at best, barely detectable. It would appear that a strongly responsive phi i does not recur rhythmically in DD. Instead, the light pulse apparently acts primarily as a "dawn" signal that triggers a single cycle of photoinducibility, since a second 8-hr light pulse, placed to begin 2 hr after the end of the first, induces a large increase in plasma LH. Similar results are obtained if any single 8-hr light pulse presented to animals held in darkness is preceded, 10 hr earlier, by a short "dawn" light signal. Such dawn signals can be effective when very short; a pulse of only 30 sec can cause a subsequent phi i. The dawn pulse is effective at any circadian phase and leads to a single cycle in photoinducibility. In contrast, a much longer light pulse (perhaps not less than 4 hr) is needed to interact with phi i if significant gonadotropin secretion is to be stimulated. In confirmation of the findings described above, we found that Nanda-Hammer lighting schedules have remarkably little effect in stimulating gonadotropin secretion in gonadectomized quail. There is, for example, a very marked difference between the effectiveness of "resonating" schedules such as LD 6:6, which stimulates a high LH secretion rate since each "inductive" light pulse is preceded by an appropriate "dawn" signal, and a theoretically effective schedule such as LD 6:30, which induces a very small response by comparison. Such schedules (even theoretically noninductive ones) can, however, be made very highly inductive if alternate light pulses are preceded by an appropriately positioned 15-min light pulse to act as "dawn."