Annual Variation of Primary Molt with Age and Sex in Cassin's Auklet

Abstract

-In Cassin's Auklet (Ptychoramphus aleuticus) on Southeast Farallon Island, California, 1979-1984, we found significant annual and seasonal variation in timing and rate of primary molt in adult males, adult females, and subadults (1-2 years old). Except in 1979, adult males began to molt at least 10 days before adult females, and males molted at a slower rate. Subadult birds initiated molt later and molted faster, but showed less annual variation than adults. Timing of breeding and breeding success varied annually but were not correlated directly with molt. In years of extended breeding, molt rates were slower apparently in response to the energetic demands caused by the overlap of breeding and molt. Received 17 October 1989, accepted 12 April 1990. ANNUAL and seasonal variation in molt, as well as its overlap with other phenological events (e.g. breeding, migration), is rarely studied in wild birds. Only the broadest patterns of molt are known for most pelagic seabirds because they molt at sea, away from breeding colonies where they are most accessible. Minimal data are available for alcids because wing molt for most species is synchronous, which renders them flightless at sea. Consequently, most of the information on molt in alcids is based on museum specimens (Salomonsen 1944, Storer 1952) or on captive birds (Birkhead and Taylor 1977, Swennen 1977, but see Harris and Wanless 1990). The auklets (Aethia spp., Cyclorrhynchus, Ptychoramphus), however, molt their primaries over an extended period that overlaps with breeding (Payne 1965, Bedard and Sealy 1984). Cassin's Auklet (Ptychoramphus aleuticus) lays a single egg and has an average 38-day incubation and 41-day nestling period. It is the only auklet with a subarctic distribution in the eastern Pacific (Manuwal 1974a, Ainley and Boekelheide 1990). A large breeding colony on Southeast Farallon Island (SEFI), California, is attended by auklets nearly year-round. The accessibility of this colony has accommodated numerous studies on the natural history and breeding biology of Cassin's Auklets (for a summary, see Ainley and Boekelheide 1990). The Cassin's Auklet also is the only alcid known to double brood, though most Cassin's Auklets do so unsuccessfully (Ainley and Boekelheide 1990). In birds that double brood, molt may overlap extensively with breeding. In addition, this species undergoes significant annual variation in the timing of breeding and the amount of double brooding relative to oceanographic conditions (Manuwal 1974a, Ainley and Boekelheide 1990). Payne (1965) showed that the progress and rate of molt are slower in Cassin's Auklets that have expended the most energy for breeding, although he was unaware that these birds can double brood. From 1979-1984, we studied the primary molt of Cassin's Auklets on SEFI to investigate further the effects of breeding on molt in consideration of the potentially confounding variables of individual age, sex, and body weight. STUDY AREA AND METHODS Southeast Farallon Island is located 42 km west of San Francisco and supports the largest breeding colony of seabirds in the continental United States (Ainley and Boekelheide 1990). Manuwal (1974a, b, 1979) described the breeding colony and habitats used by Cassin's Auklets on SEFI. We used a 20 x 5 m fish net (1.27-cm2 mesh) to capture and band birds as described by Ralph and Sibley (1970). The net was opened 2 to 3 times per month ca. 1 h before dawn when birds began to leave the colony for the day. Sampling dates varied each month until approximately 100 birds were captured. We recorded the following information on each captured or recaptured bird: 1. Weight-Measured to nearest gram. 2. Eye color-To estimate age we modified the methods of Manuwal (1978) to include nine (instead of four) pigmentation categories: from 1.0 (white eye) to 5.0 (dark eye) in scoring increments of 0.5. Birds 689 The Auk 107: 689-695. October 1990 This content downloaded from 157.55.39.26 on Sun, 04 Sep 2016 04:35:19 UTC All use subject to http://about.jstor.org/terms 690 EMSLIE, HENDERSON, AND AINLEY [Auk, Vol. 107 with dark eyes (>3.5) were considered subadults (assumed nonbreeders, 1-2 yr old), and those with light eyes ('3.0) were considered adult (assumed breeders). Accuracy of these criteria was tested by examination of known-age breeders on SEFI in 1988 and 1989. 3. Bill dimensions-Width and depth (in mm) were greatest at anterior edge of nares (Nelson 1981). Birds with a bill depth (or average bill depth from multiple captures) of > 10.3 mm were classified as males, and those with a depth of <9.5 mm, females. Although Nelson (1981) found most females to have a bill depth of <10.3 mm, we used a stricter definition to account for multiple-observer and measurement error. Subadult birds were not sexed. 4. Primary molt-Molting feathers were scored as either old (score 0), missing, or growing. Growing feathers were classified into one of six stages of growth: pin or 1/10, 3/10 5%, 7/10, or 9/10 grown, or new. These categories were recoded with standardized methods (e.g. Newton 1966, Ginn and Melville 1983). Accordingly, primaries (P1 to P10) in pin were scored as 1, ?/ to 3/lo grown were scored 2, 5/lo was scored 3, 7/lo to 9/10 were scored 4, and new primaries were scored 5. Summation of scores for all 10 primaries provided the molt score for each captured bird. We did not collect data on body or secondary molt, and hereafter molt refers to primary molt. We used molt score rather than feather mass (see Summers et al. 1983) in our analysis. In Cassin's Auklet, P10 is only 25-29% longer than P1, and P7-P10 are nearly equal in length (within 2 mm; n = 9). Thus, we assume that a change in the amount of feather mass replaced during molt in Cassin's Auklet probably does not affect molt rate as it does in other Charadriiformes (Summers et al. 1983, Underhill and Zucchini 1988, Underhill et al. 1990). We measured birds that were captured before, during, and after the molt period. Birds recaptured within the same month were excluded. These Type II data provided estimates of molt parameters with the smallest bias (see Underhill and Zucchini 1988). We did not use Underhill and Zucchini's (1988) numerical algorithm to estimate molt parameters because it relies on a linear relationship of molt rate with time. This relationship is not always linear in Cassin's Auklet (Payne 1965, and data here). We used linear regression of Julian date on molt score (Pimm 1976) for all birds captured from April through October to estimate molt parameters for each year. The slopes of regression lines produced by this method do not give molt rate, but an inversion of molt rate (rate-'). Although our statistical comparisons are based on the regression values, for clarity in discussing relationships of molt with age and sex, we refer to actual molt rates. All regression analyses were completed with SPSS/PC+ (Nurusis 1986). Annual variation in molt among sex and age classes was assessed with an analysis of covariance (ANCOVA) on SPSS/PC+. The relationships of molt with body weight, timing of breeding, and breeding success were compared using Spearman's rank correlation (re, P < 0.05). We determined annual mean (? SD) laying dates (MLD) and breeding success (BS) of Cassin's Auklets from a sample of monitored nest sites (Ainley and Boekelheide 1990). Although MLD implies a normal distribution in laying dates, in Cassin's Auklets this distribution is often skewed towards the earlier part of the laying period (Ainley and Boekelheide 1990). The use of median laying dates resulted in a ranking of years similar to that by mean laying dates. We used data from first clutches only (excluding relays) to calculate MLD. Breeding success is the average number of chicks fledged per pair laying.