Early nighttime testosterone peaks are correlated with GnRH-induced testosterone in a diurnal songbird.

Experimental manipulation has established testosterone as a potent, pleiotropic regulator coordinating morphology, physiology and behavior. However, the relationship of field-sampled, unmanipulated testosterone concentrations with traits of interest is often equivocal. Circulating testosterone varies over the course of the day, and recent reports indicate that testosterone is higher during the night in diurnal songbirds. Yet, most field studies sample testosterone during the morning. Sampling at times when levels and individual variation are low may be one reason relationships between testosterone and other traits are not always observed. Testosterone is regulated by the hypothalamic-pituitary-gonadal axis, with gonadotropin-releasing hormone (GnRH) initiating the endocrine cascade. Research has examined GnRH-induced testosterone levels with traits of interest, yet the relevance of these induced levels and their relationship with endogenously produced levels are not fully clear. Using photostimulated male great tits (Parus major) we tested the hypotheses that circulating testosterone levels peak during the night and that GnRH-induced testosterone concentrations are positively related to nightly testosterone peaks. Blood was sampled during first, middle or last third of night. One week later, baseline and GnRH-induced testosterone levels were sampled during mid-morning. Morning baseline testosterone levels were low compared with night-sampled levels that peaked during the first third of the night. Further, GnRH-induced testosterone was strongly positively correlated with levels observed during the first third of the night. These data suggest that morning testosterone samples likely do not reflect an individual's endogenous peak. Instead, GnRH-induced testosterone levels do approximate an individual's nightly peak and may be an alternative for birds that cannot easily be sampled at night in the field. These findings are likely to have implications for research aimed at relating traits of interest with natural variation in sex steroid hormone levels.