This review examines how birds use the annual cycle in photoperiod to ensure that seasonal events--breeding, molt, and song production--happen at the appropriate time of year. Differences in breeding strategies between birds and mammals reflect basic differences in biology. Avian breeding seasons tend to be of shorter duration and more asymmetric with respect to changes in photoperiod. Breeding seasons can occur at the same time each year (predictable) or at different times (opportunistic), depending on the food resource. In all cases, there is evidence for involvement of photoperiodic control, nonphotoperiodic control, and endogenous circannual rhythmicity. In predictable breeders (most nontropical species), photoperiod is the predominant proximate factor. Increasing photoperiods of spring stimulate secretion of gonadotropin-releasing hormone (GnRH) and consequent gonadal maturation. However, breeding ends before the return of short photoperiods. This is the consequence of a second effect of long photoperiods--the induction of photorefractoriness. This dual role of long photoperiods is required to impart the asymmetry in breeding seasons. Typically, gonadal regression through photorefractoriness is associated with a massive decrease in hypothalamic GnRH, essentially a reversal to a pre-pubertal condition. Although breeding seasons are primarily determined by photoperiodic control of GnRH neurons, prolactin may be important in determining the exact timing of gonadal regression. In tropical and opportunistic breeders, endogenous circannual rhythmicity may be more important. In such species, the reproductive system remains in a state of "readiness to breed" for a large part of the year, with nonphotic cues acting as proximate cues to time breeding. Circannual rhythmicity may result from a temporal sequence of different physiological states rather than a molecular or cellular mechanism as in circadian rhythmicity. Avian homologues of mammalian clock genes Per2, Per3, Clock, bmal1, and MOP4 have been cloned. At the molecular level, avian circadian clocks appear to function in a similar manner to those of mammals. Photoperiodic time measurement involves interaction between a circadian rhythm of photoinducibility and, unlike mammals, deep brain photoreceptors. The exact location of these remains unclear. Although the eyes and pineal generate a daily cycle in melatonin, this photoperiodic signal is not used to time seasonal breeding. Instead, photoperiodic responses appear to involve direct interaction between photoreceptors and GnRH neurons. Thyroid hormones are required in some way for this system to function. In addition to gonadal function, song production is also affected by photoperiod. Several of the nuclei involved in the song system show seasonal changes in volume, greater in spring than in the fall. The increase in volume is, in part, due to an increase in cell number as a result of neurogenesis. There is no seasonal change in the birth of neurons but rather in their survival. Testosterone and melatonin appear to work antagonistically in regulating volume.