Interseasonal and Intersexual Resource Partitioning in Hairy and White-headed Woodpeckers

--Foraging behavior in the closely related Hairy (Picoides villosus) and Whiteheaded (P. albolarvatus) woodpeckers was examined in an area of sympatry to evaluate interseasonal and intersexual resource use. Similar foraging heights were observed for each species-sex sample during summer, but significant differences were evident during winter. Male and female White-heads maintained similar relative foraging heights between seasons, whereas male and female Halties foraged relatively higher during winter. Use of tree species differed significantly for each class between seasons, except for male Halties. Differences in foraging substrates and tree health also were noted interseasonally. All foraged at similar times of day during summer, but negative relationships occurred between times of foraging during winter; foraging times were significantly different between male Halties and Whiteheads. A general trend toward decreased overlap in foraging behaviors during winter was a reflection of concentration of foraging activities on live incense cedar (Calocedrus decurrens), especially by female White-heads. This change apparently was due to the presence of an abundant and accessible prey (incense cedar scale, Xylococculus macrocarpae) on cedar. Behavioral shifts in foraging activities may be related to the differential ability to extract prey as a function of bill morphology; such an idea can be extended to intersexual as well as interseasonal considerations. Thus, segregation of foraging activities in these two woodpecker species may be attributed to morphological differences and habitat complexity rather than to competitive interactions dictated by resource limititions. Received 16 May 1986, accepted 13 October 1986. THE utilization and partitioning of resources by closely related, syrupattic species of birds has attracted much study (e.g. see reviews by Schoener 1974, Eckhardt 1979, Rotenberry 1985). As a group, co-occurring species of woodpeckers (family Picidae) have received frequent attention (e.g. Kisiel 1972; Williams 1975; Conner 1980, 1981), possibly owing to the ease of observation and the occurrence of sexual dimorphisrn. Thus, woodpeckers make good subjects for evaluation of resource partitioning on both interspecific and intersexual levels (e.g. Kilham 1965, Ligon 1968, Austin 1976, Jenkins 1979, Williams 1980, Peters and Grubb 1983). During preliminary studies of bird communities in the western Sierra Nevada, California, we noted the co-occurrence of two Picoides species in a mixed-conifer forest: the Hairy (P. villosus) and White-headed (P. albolarvatus) woodpeckers (Morrison et al. 1985, 1986). Sym•Present address: Department of Biological Sciences, Northern Arizona University, Flagstaff, Arizona 86011 USA. patric species of woodpeckers differ greatly in size and in foraging habits (Short 1971). The two species considered here were described by Short (1971) as "ecologically separate" when sympatric, with albolarvatus occuring in "pine" and villosus in "other coniferous forest and mixed woods." We evaluated modes of resource partitioning by these two similar-size (by body mass; Dunning 1984) and closely related (Short 1971, 1982) species in an area typified by a ubiquitous mixture of five coniferous and one deciduous tree species. These data were analyzed on an intersexual and interseasonal basis to determine if methods of resource use and the magnitude of resource partitioning varied by sex, season, or both. STUDY AREA AND METHODS The study area was the Blodgett Forest Research Station (administered by the Department of Forestry and Resource Management, University of California, Berkeley), El Dorado Co., California. This 1,200-ha forest is located in the mixed-conifer zone (see Grif225 The Auk 104: 225-233. April 1987 226 MORRISON AND WITH [Auk, Vol. 104 fin and Critchfield 1972) at about 1,350 m elevation in the west-central Sierra Nevada. The forest is predominated by incense cedar (Calocedrus decurrens; 25% of total basal area, unpubl. data), white fir (Abies concolor; 21%), ponderosa pine (Pinus ponderosa; 19%), Douglas fir (Pseudotsuga rnenziesii; 15%), sugar pine (P. larnbertiana; 10%), and California black oak (Quercus kelloggii; 8%). The forest has been divided into 5-40ha compartments managed under various silvicultural systems. The forest is now mostly mature (>70 yr old) conifer. During summer (early May to late July) 1983 and 1984, 24 compartments totaling about 420 ha were selected for study. During summer 1985 we selected 9 compartments, 7 of which differed from those used during 1983-1984; the 1985 study area totaled about 210 ha. Because access to much of the forest was limited during winter, a subset of 4 large compartments used during summer totaling about 100 ha was selected for study during winter (early November to mid-March) 1982-1983 and 1983-1984. Summer and winter sites were of similar age and tree-species composition. The abundance of Hairy and White-headed woodpeckers is similar during winter (0.19 and 0.17 birds/ count, respectively) and summer (0.13 and 0.08 birds/ count, respectively) (Morrison et al. 1986). Other woodpecker species that occurred at Blodgett were Northern Flickers (Colaptes auratus), Red-breasted Sapsuckers (Sphyrapicus tuber), and Downy (P. pubescens) and Pileated (Dryocopus pileatus) woodpeckers (Morrison et al. 1986). Adequate data were not available, however, to analyze these species interseasonally or intersexually. About 1,250 person-hours during summer and 700 person-hours during winter were spent observing foraging behavior. An observer walked systematically through a compartment recording data on birds as encountered; I0 different observers spent varying amounts of time recording data during the study. We divided our activities so that roughly equal time was spent in each compartment during each season. Data on only one individual of a species were recorded at a particular place and time regardless of the number of birds present. Data on only one or two species were recorded when observing flocking birds to minimize the potential problem of correlated activities of co-occurring individuals. Except for winter 1982-1983, foraging activities were recorded for a minimum of 10 s to a maximum of about 30 s. For each individual we recorded date, time of day, species, sex, species of plant, foraging substrate (e.g. limb, trunk), time spent on each substrate, perch height, plant height, and type of foraging motion (e.g. glean, peck). This same information was recorded during winter 1982-1983, but observations were made on an individual every 30 s to a maximum of about 5 rain (i.e. sequential observations). We modified our methods because analysis of sequential observations has certain statistical problems. Results of both techniques are comparable, however, given adequate sample sizes; our data exceeded this requirement (Morrison 1984). The presence of observers likely affected foraging activities. Because the initial sighting of a bird probably was biased toward conspicuous individuals, using only the first observation of an individual would not solve the problem associated with observer effects on foraging activities. We assumed that these problems influenced the data in a similar manner for all species and sexes. Because of problems concerning independence of data recorded sequentially, we were espe~ cially careful when interpreting results at P near 0.05. We calculated resource overlap, where Px, and py, are the proportional use of the ith resource state by the xth and yth species. C ranges from 0 (total dissimilarity) to 1 (total similarity) (Colwell and Futuyma 1971). A two-dimensional measure was calculated from a foraging tree-foraging substrate matrix. Calculating multidimensional resource overlap in this manner is preferable to using additive or multiplicative values (Alatalo 1982). We did not calculate three-dimensional matrices because cellular values would become too small for calculation of meaningful results (Alatalo 1982). These indices present simple summaries of data, but do not indicate statistical significance (although they may be biologically important). Therefore, we compared (for independence) the distribution of certain foraging activities for the various resource states between species and sexes using a Chi-square analysis (Fisher's exact test was used for small samples; see Norusis 1983). Further, through computer simulations, Ricklefs and Lau (1980) found that a difference of 0.1-0.2 was required to reject the null hypothesis that two overlap indices were drawn from the same sampling distribution. Therefore, we adopted the general rule that two indices must differ by at least 0.2 to merit discussion. Resource partitioning may occur if species forage at different times (Schoener 1974). For this study we used a simple index of foraging time: assuming observers encountered birds at random, we calculated speciesand sex-specific times of activity by determining the frequency at which birds were observed in 1-h periods. We then used Spearman's rank correlation (rs) to determine if times of foraging activities were similar. Observers systematically surveyed birds throughout the day, which reduced potential biases in foraging activity times. The external morphology of Hairy and WhiteApril 1987] Woodpecker Resource Partitioning 227