Suppression of Cortisol Levels in Subordinate Female Marmosets: Reproductive and Social Contributions Wendy Saltzman,* ,1 Nancy J. Schultz-Darken, Frederick H. Wegner, Daniel J. Wittwer, and David H. Abbott* , ² Wisconsin Regional Primate Research Center, *Endocrinology-Reproductive Physiology Program, and †Department of Obstetrics and Gynecology, University of Wisconsin, Madison, Wisconsin 53715-1299 Received August 26, 1997; revised November 21, 1997; accepted January 7, 1998 Socially subordinate female common marmosets ( Cal- lithrix jacchus) have markedly lower plasma cortisol levels than dominant females. Subordinate females also undergo hypoestrogenemic anovulation, and estrogen can elevate glucocorticoid levels. Therefore, we previously hypothe- sized that this cortisol difference is mediated by rank- related differences in reproductive hormones, probably es- tradiol. To test this possibility, we characterized the effects of the ovarian cycle and ovariectomy on plasma cortisol concentrations. Beginning in the early follicular phase, basal blood samples were collected from seven cycling female marmosets daily for 16 days and at 2- to 3-day intervals for another 16 days. Samples were collected iden- tically from seven anovulatory subordinate females and seven long-term ovariectomized females. Cortisol levels changed reliably across the ovarian cycle, with levels in the mid- to late follicular, peri-ovulatory, and early luteal phases higher than those in the remainder of the cycle. Cortisol levels of cycling females were significantly higher than those of subordinates at all parts of the cycle, but were significantly higher than those of ovariectomized fe- males only during the midcycle elevation. Unexpectedly, subordinates had significantly lower cortisol levels than ovariectomized females, as well as higher estradiol and estrone levels and lower progesterone and luteinizing hor- mone (LH) levels. These results confirm that circulating cortisol concentrations are modulated by reproductive function in female marmosets but also indicate that low cortisol levels in subordinate females cannot be attributed simply to hypoestrogenemia. Instead, other factors, such as direct effects of social subordination or suppression of LH levels, contribute to suppression of cortisol in subordi- nates. © 1998 Academic Press Measures of hypothalamo–pituitary–adrenal (HPA) activity have been found to correlate with social status in numerous species, with subordinate animals frequently evincing higher adrenal weights, basal glucocorticoid levels, or adrenal responses to stressors than their dom- inant counterparts (see reviews by Henry, Stephens, and Ely, 1986; Sapolsky, 1995). These findings have led to the widely held view that social subordination is inherently stressful and that the psychosocial stress generated by subordination directly elevates HPA activity. It is be- coming clear, however, that this simple socioendocrine model may not adequately explain the diversity of find- ings on endocrine function and social status. For exam- ple, an increasing number of investigations in a growing number of species have failed to demonstrate the ex- pected association of high chronic HPA activity with low social status. Instead, HPA activity may not be system- atically correlated with social status (e.g., Mays, Vleck, and Dawson, 1991; Bercovitch and Clarke, 1995; Ziegler, Scheffler, and Snowdon, 1995; Smith and French, 1997) or, in a few cases, may be lower in subordinate animals than in their dominant counterparts (Coe, Mendoza, and Levine, 1979; Schoech, Mumme, and Moore, 1991; Wing- field, Hegner, and Lewis, 1991; Saltzman, Schultz- Darken, Scheffler, Wegner, and Abbott, 1994; Creel, Creel, and Monfort, 1996). One reason for the absence of a consistent relationship between HPA activity and social status may be that HPA function is modulated by numerous behavioral and physiological variables in addition to psychosocial stress. Moreover, many of these variables may them- selves be influenced by social status. For example, phys- ical activity (Coleman, Garland, Marler, Newton, Swal- low, and Carter, 1998; Mallick, Stoddart, Jones, and Bradley, 1994), immune function (Hermus and Sweep, 1 To whom correspondence should be addressed at Wisconsin Regional Primate Research Center, University of Wisconsin, 1220 Capitol Court, Madison, WI 53715-1299. Fax: (608) 265-4729. E-mail: saltzman@primate.wisc.edu. Hormones and Behavior 33, 58–74 (1998) Article No. HB981436 0018-506X/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved. 58