-I quantified parental effort in biparentally caring Razorbills (Alca torda) to test whether variation in male effort was related to variation in their mates' receptivity to extrapair copulations. On average, males contributed approximately equally to their mates in chick feedings and overnight nest attendance. There was, however, marked variation in relative male effort, with the proportion of male feeding relative to their mates' feeding ranging from 16 to 72%. No significant portion of this variation was explained by direct and indirect measures of males' confidence of paternity. Support was equivocal for an alternative hypothesis, that variation in male effort is caused by differential male ability. Male care increased significantly with fighting ability and mating arena attendance, but not with five other measures of male vigor. I propose a second alternative hypothesis-that Razorbills and other long-lived, breeding-site-faithful species may reduce their social status over the long term by provisioning poorly, and that lowered status may decrease fitness. Thus, circumstances may exist under which it is beneficial for males to provide care for offspring that may not be their own. Received 22 April 1991, accepted 10 February 1992. ALTHOUGH many species of monogamous birds perform extra-pair copulations (EPCs; McKinney et al. 1984, Birkhead et al. 1987, Westneat et al. 1990), males in most species provide parental care (Wittenberger and Tilson 1980, Oring 1982). The growing evidence that EPCs often succeed in fertilizing eggs (Sherman and Morton 1988, Morton et al. 1990, Westneat 1990) poses a problem in that individuals should be adapted to avoid investing in unrelated offspring (Trivers 1972, Werren et al. 1980). It has been suggested that when males risk losing paternity, they should invest parentally in proportion to their probability of siring their mates' young (Trivers 1972, Maynard Smith 1977, Craig and Jamieson 1985), a prediction I refer to as the confidence-of-paternity hypothesis. I tested this hypothesis with monogamous Razorbills (Alca torda) in which a high percentage of females actively seek EPCs (Wagner 1992a), even though females require parental assistance from the male. Biparental care is necessary in Razorbills because the chick cannot be left unattended, as it requires brooding (Harris and Birkhead 1985), and because in open sites the chick is very vulnerable to predation (Hudson 1982). Evidence that a single parent 1 Present address: Department of Zoological Research, National Zoological Park, Smithsonian Institution, Washington, D. C. 20008, USA. may be unable to provide sufficient food for a chick was suggested in an experiment that showed that two parents had difficulty feeding two young (Lloyd 1977). Confidence of paternity may be estimated directly by quantifying relevant behaviors in which the male participated or witnessed (M01ler 1988), such as the number of EPC attempts involving his mate in his presence. The number of within-pair copulations (WPCs) a male achieves also may affect his confidence of paternity (M0ller 1988), because the outcome of sperm competition (in domestic fowl) is known to be influenced by the relative quantity of sperm of each male stored by females (Martin et al. 1974). Confidence of paternity also may be estimated indirectly by analyzing the receptivity of females to EPCs witnessed by the observer, but not necessarily by the male, such as the number of EPCs the female accepted, or the total number of EPC attempts males made with the female. Female Razorbills were found to vary considerably in the number of EPCs they accepted (0-7; Wagner 1992a), with 50% of females accepting at least one EPC (Wagner 1991a). Females also varied markedly in their receptivity to EPC, accepting EPCs during 0 to 100% of their opportunities (Wagner 1992a). Two measurable forms of parental effort, chick feeding and overnight nest attendance, appear to be costly to Razorbills. The expense of over-