WESTERN Western Specialty: Pinyon Jay Photo by © Brian E. Small of Los Angeles, California: Pinyon Jay ( Gymnorhinus cyanocephalus ) Deschutes National Forest, Lake County, Oregon, 9 August 2006. Volume 40, Number 4, 2009 Recent Purple Martin Declines in the Sacramento Region of California: Recovery Implications Daniel A. Airola and Dan Kopp 254 Further Decline in Nest-Box Use by Vaux’s Swifts in Northeastern Oregon Evelyn L. Bull and Charles T. Collins 260 Use of a Nesting Platform by Gull-billed Terns and Black Skimmers at the Salton Sea, California Kathy C. Molina, Mark A. Ricca, A. Keith Miles, and Christian Schoneman 267 Birds of Prey and the Band-tailed Pigeon on Isla Guadalupe, Mexico Juan-Pablo Gallo-Reynoso and Ana-Luisa Figueroa-Carranza 278 Food Habits of Wild Turkeys in National Forests of Northern California and Central Oregon Greta M. Wengert, Mourad W. Gabriel, Ryan L. Mathis, and Thomas Hughes 284 Seasonal Variation in the Diet of the Barn Owl in Northwestern Nevada Abigail C. Myers, Christopher B. Goguen, and Daniel C. Rabbers 292 NOTES First Record of a Mangrove Yellow Warbler for Arizona Nathan K. Banfield and Patricia J. Newell 297 Prey Remains in Nests of Four Corners Golden Eagles, 1998- 2008 Dale W. Stahlecker, David G. Mikesic, James N. White, Spin Shaffer, John P DeLong, Mark R. Blakemore, and Craig E. Blakemore 301 Book Reviews Dave Trochlell and John Sterling 307 Featured Photo: Juvenal Plumage of the Rufous-crowned Sparrow Brad Schram 310 Index Daniel D. Gibson 314 Front cover photo by © Todd Easterla of Rancho Cordova, California: Streaked Shearwater ( Calonectris leucomelas), Monterey Bay, Monterey County, California, 12 October 2008. A species of the western Pacific, the Streaked Shearwater has been substantiated once in Oregon, 18 times in California (most frequently on Monterey Bay but once inland), and once in Wyoming. Back cover “Featured Photo” by © Brad Schram of Arroyo Grande, California: juvenile Rufous-crowned Sparrow ( Aimophila ruficeps ), Deer Canyon, San Luis Obispo County, California, 6 August 2006. Note the pencil-streaked upper breast and sides over buff, streaked crown, and facial pattern, a paler version of the adult’s. Western Birds solicits papers that are both useful to and understandable by amateur field ornithologists and also contribute significantly to scientific literature. Particularly desired are reports of studies done in or bearing North America west of the 100th meridian, including Alaska and Hawaii, northwestern Mexico, and the northeastern Pacific Ocean. Send manuscripts to Kathy Molina, Section of Ornithology, Natural History Muse- um of Los Angeles County, 900 Exposition Blvd., Los Angeles, CA 90007 ; kmolina@ nhm.org. For matters of style consult the Suggestions to Contributors to Western Birds (at www.westernfieldornithologists.org/docs/journal_guidelines.doc). WESTERN BIRDS Volume 40, Number 4, 2009 RECENT PURPLE MARTIN DECLINES IN THE SACRAMENTO REGION OF CALIFORNIA: RECOVERY IMPLICATIONS DANIEL A. AIROLA, Airola Environmental Consulting, 2700 6th Ave., Sacramento, California 95818; d.airola@sbcglobal.net DAN KOPP, 8295 La Riviera Drive, Sacramento, California 95677 ABSTRACT: We monitored the Purple Martins ( Progne subis) breeding in the Sacramento region, California, in 2008 and 2009, following similar monitoring from 2002 to 2007. This bridge-nesting population is the last remnant of the formerly widespread Central Valley population. All 11 colonies occupied in 2007 remained active in 2008, but the number of occupied colonies declined to 9 in 2009. From 2007 to 2009 the number of breeding pairs declined by 34%, from 106 to 70. Since 2004, the population has declined by 60% (103 pairs), and the annual rate of decline has increased. Probability calculations suggest the decline has been caused by factors other than chance fluctuation. If the decline continues at its current rate, the population will be extirpated within 22 years. Causes of the decline are unknown, but our previous studies have implicated mortality during the breeding season from vehicle collisions. Sacramento-area populations also are threatened by disturbance from a variety of construction projects and land-use changes that may affect habitat suitability. Recent losses in Sacramento represent a 6-11% decline in California’s estimated Purple Martin population. Increased effort is needed to stabilize the Sac- ramento population, which would serve as the likely source for any future recovery within the Central Valley. The Purple Martin ( Progne subis) is recognized by the California Depart- ment of Fish and Game as species of special concern because of reductions in its geographic range and numbers (Airola and Williams 2008). The martin population in the Sacramento area (mostly within the city of Sacramento, as well as limited areas in adjacent Placer and Yolo counties) is a remnant of a population once more widespread in California’s Central Valley. This population apparently has survived competition from the European Starling { Sturnus vulgaris) by nesting in overpasses and elevated freeways (collec- tively, bridges), which are entered through “weep holes” on the undersides of the structures (Airola and Grantham 2003). The Sacramento population has, until recently, constituted approximately 9-21% of the total estimated nesting in California (Airola and Williams 2008). 254 Western Birds 40:254-259, 2009 RECENT PURPLE MARTIN DECLINES IN SACRAMENTO, CALIFORNIA We monitored the Purple Martins breeding around Sacramento from 2002 to 2007 (Airola et al. 2008) after conducting similar surveys in the early 1990s (Airola and Grantham 2003). Here, we report the population’s status in 2008 and 2009 and implications of recent trends for conservation of the species. STUDY AREA AND METHODS We studied Purple Martins at bridges in the Sacramento region (Sacra- mento, Yolo, and western Placer counties) previously identified as occupied or suitable for use by the species. Locations of colonies and other suitable sites, and the criteria used to define them, were described by Airola and Grantham (2003), Leeman et al. (2003), and Kopp and Airola (2007). Survey methods included mapping of holes in which martins nested, re- cording diagnostic nesting behaviors [carrying food to nests, removal of fecal sacs, juveniles begging, or seeing nestlings in holes (Airola and Grantham 2003, Leeman et al. 2003)], and video monitoring of a sample of accessible nest sites (see Airola et al 2008 for methods). These methods provide a consistent and repeatable basis for estimating numbers. We confirmed breeding by 93% and 96% of pairs in 2008 and 2009, respectively, through observation of diagnostic nesting behaviors and video monitoring at 35% (2008) and 38% (2009) of nest sites. Remaining nests were inferred on the basis of pairs entering holes frequently over the nesting season (see Airola and Grantham 2003 and Airola et al. 2008). RESULTS In 2008, Purple Martins bred in the same 11 sites they used in 2007 (Table 1). The newest site, at Highway 65 and Taylor Road in Placer County (Kopp and Airola 2007), was reoccupied by a single pair of Purple Martins that contained a male in typical dark after-second-year plumage (Pyle 1997), suggesting the return of the male that bred there in second-year plumage in 2007. As in 2007, the formerly used Marconi, Airbase, and Pole Line sites remained unoccupied again in 2008 and 2009. In 2009, martins bred at only nine colonies (Table 1), the first time since extensive surveys of all suitable sites were initiated in 2003 that the number of colonies has dropped below 1 1 . Second-year birds were seen at both the Marconi and Pole Line colonies but did not nest. Fifteen other sites previously identified as suitable (Leeman et al. 2003, Kopp and Airola 2007) but never known to be occupied also were surveyed in 2008 and 2009 and did not support nesting martins. The number of pairs of Purple Martins nesting around Sacramento de- clined during both 2008 and 2009. The 2009 total of 70 pairs represents a decline by 34% from the number in 2007, by 60% from the peak in 2004 (Table 1), and the fifth consecutive year of decline. At eight colonies, numbers in 2009 decreased from those in 2007, while they remained the same at one colony and increased at two. No obvious causes for the changes were evident. As in previous years, colonies in the more urbanized downtown areas (I St., 20 th St., Broadway, 35 th St., S St.) declined from 2007 to 2009 255 RECENT PURPLE MARTIN DECLINES IN SACRAMENTO, CALIFORNIA Table 1 Number of Breeding Pairs of the Purple Martin around Sacra- mento, California, 2002-2008 Colony 2002 2003 2004 2005 2006 2007 2008 2009 I Street 37 29 35 32 17 11 6 5 20 th Street 14 21 23 23 16 15 6 5 Sutterville 4 6 8 5 6 6 5 6 Broadway 8 7 7 7 5 1 1 0 S Street 14 14 16 14 18 9 7 6 35 th Street 29 19 15 14 6 3 3 1 Redding Rd. 0 3 12 10 14 14 15 17 El Camino ns a 15 23 21 21 20 11 5 Marconi ns 1 4 3 0 0 0 0 Roseville Rd. 29 39 27 24 24 17 17 13 Arden ns 0 3 6 13 9 11 12 Airbase ns 0 0 1 1 0 0 0 Pole Line ns 2 0 0 0 0 0 0 Hwy. 65/Taylor ns ns ns ns ns 1 1 0 Total 135 156 173 160 141 106 83 70 a ns, not surveyed. at a higher rate (59%) than did more outlying colonies (21%), although this differences is only marginally significant statistically (x 2 1 = 3.05, P = 0.08). Since 2004, when the size of the nesting population peaked, downtown colonies have declined by the significantly greater rate of 82% (from 96 to 17 pairs), while outlying colonies have declined by 31% (77 to 53 pairs; X 2 i = 79.7, P< 0.0001). DISCUSSION The 2008 and 2009 declines in the number of Purple Martins nesting in Sacramento represent the fifth consecutive year of decline following two years of increases (Table 1). The 2008 and 2009 populations are lower than any recorded previously in 10 surveys since 1992 (Airola and Gran- tham 2003, Airola et al. 2008), despite surveys since 2003 being more comprehensive (Table 1). How likely is it that recent declines represent a trend or a population fluctuation? We address this question in several ways. First, the annual rates of population decline appear to be accelerating, from 8% in 2004 and 12% in 2005 to 16-25% from 2007 through 2009. Second, we calculated the probability that by chance alone the population could have declined five years in a row over our eight years of monitoring. If the probability of an increase or decline in any one year is 0.5, the probability of five consecutive annual declines (one of six possible outcomes over seven years) is P = 6(0. 5) 7 = 0.047, a figure strongly suggesting that a cause other than chance is responsible. Notwithstanding the relatively large declines in 2008 and 2009, most recent colonies remained occupied. Several colonies that supported a single pair in 2008 (Broadway and Highway 65/Taylor Rd) were abandoned in 2009, and the 35 th St. colony declined to a single pair. Maximizing the 256 RECENT PURPLE MARTIN DECLINES IN SACRAMENTO, CALIFORNIA number of colonies may be important for conservation because the Purple Martin’s colonial habits may discourage reoccupation of sites once they are abandoned, especially because declining populations likely reduce pressure on new breeders to disperse to unoccupied sites. The observation in 2009 of nonbreeding birds at two sites previously abandoned (Marconi and Pole Line), however, suggests the possibility of recolonization of these sites, and perhaps others, when the population is healthier. A reduction in the number of active colony sites presumably increases the risk that a future disturbance or habitat changes at an occupied site could affect the population disproportionately. Such disturbance is a substantial threat, as 60% of the 2009 population is breeding at sites that have been recently affected by, or are proposed for, construction projects that may disturb the bridges, the lands below them, or alter the habitat immediately adjacent to sites (updated from Airola et al. 2008). Protecting habitat at current colonies, however, also requires considerable understanding of local patterns of habitat use such as flight paths to nests, sources of nest material, and perch sites. Also, appropriate protection from construction disturbance requires recognition that Purple Martins tolerate considerable activity as long as nest sites remain secure (Airola et al. 2009). Excluding Purple Martins from their nesting areas during construction (as is typically done for other bridge-nesting swallows) reduced the subsequent nesting population at one Sacramento colony (Airola and Grantham 2003, Airola et al. 2004). Because such exclusion and displacement also likely reduces reproductive success and risks long-term abandonment of colonies (Airola and Grantham 2003) it should be used cautiously. CONSERVATION IMPLICATIONS The Purple Martin’s nesting population in the Sacramento area has declined despite considerable efforts to protect and enhance it, including removing feral cats where they preyed on Purple Martins collecting nest material, reducing collision mortality with trains, reducing nestlings’ falling out of nests, rehabilitating fallen nestlings, removing encroaching vegetation that blocked access to nest sites, and protecting colony sites from construc- tion and from land uses beneath and adjacent to sites (Airola and Grantham 2003, Airola et al. 2008). Causes of the decline remain uncertain. Previous studies have indicated that the Sacramento bridge-nesting population is not limited by lack of suitable nesting habitat, infection by West Nile virus or other diseases, high predation rates, or competition with the European Starling (Airola and Grantham 2003, Leeman et al. 2002, Airola and Kopp 2007, Airola et al. 2008). Airola et al. (2008) suggested that an increase in traffic and other distur- bance associated with urban redevelopment may be increasing mortality from vehicle collisions during the nesting season. The higher rate of decline since 2004 at downtown colonies, where traffic volumes are higher, is consistent with this hypothesis. Notwithstanding evidence suggesting anthropogenic causes of decline, it remains possible that declines are related to short-term variation in weather, which we have not examined rigorously, or other unknown causes. 257 RECENT PURPLE MARTIN DECLINES IN SACRAMENTO, CALIFORNIA More detailed evaluation of the timing of mortality (breeding season versus migratory and wintering period) is underway, which may help to illuminate critical causes of mortality. Additional evaluation of reproduction and mor- tality rates also is needed to clarify the role of these factors in determining the sizes of breeding populations. Regardless of causes of declines, if the 16.5% average annual population decline from 2004 to 2009 were to continue, the Sacramento Purple Martin population would be eliminated within 22 years. However, the martin’s colo- nial nature suggests benefits from group living (i.e. , for increased fitness from extra-pair copulation and predator detection; Stutchbury 1991). Therefore, it is possible that as martin numbers at individual colonies decrease, the rate of population decline may accelerate further. The abandonment of two small colonies in 2009 may illustrate this process. The recent decline in Sacramento’s Purple Martin population is also important in a larger context, because of the small size and declining trend of the population statewide (Airola and Williams 2008). The recent decline by 103 pairs in Sacramento represents a loss of 6 to 11% of California’s estimated population of 900-1850 pairs, which was based primarily on surveys conducted in the 1990s but updated to reflect the 2004 census for Sacramento (Airola and Williams 2008). Unfortunately, few other martin populations in California have been monitored systematically since Williams’ (1998) surveys in the early 1990s, so it is unknown whether the recent de- clines in Sacramento are isolated or part of a broader trend. The population at Shasta Lake, the only other site of relatively consistent long-term monitor- ing in northern California, appears to be stable (Lindstrand 2008). A pilot survey initiated in 2009 by Airola and volunteers at a sample of northern California sites occupied in the 1990s may provide a broader indication of the martin’s status in this region. The continued decline in the number of Purple Martins breeding in Sacra- mento argues for higher conservation concern for this remnant population. Therefore, additional effort is warranted to continue monitoring the popula- tion to reveal the causes of recent declines, to minimize effects of construc- tion projects, land-use changes, and vehicular traffic at martin colonies, and to establish a population nesting in boxes (Kostka et al. 2008, Elwood et al. 2009) to bolster the existing population and reduce risk of extirpation. Loss of the Sacramento population would eliminate the primary source of potential colonists for new bridges or nest boxes and so could preclude the possibility of population recovery in the Central Valley. The results of moni- toring in 2008 and 2009 further support previous suggestions (Airola et al. 2008) that protection of the bridge-nesting martin population in Sacramento is critical as a strategy for recovery. ACKNOWLEDGMENTS We especially thank the Sacramento Audubon Society for funding. Kevin Thomas, Stan Kostka, Lynn Schmidt, and Ed Whisler assisted with monitoring. Mike Gleck- ler and other staff of the California Department of Parks and Recreation, State Railroad Museum, provided access and valuable assistance. Gabe Avila and Sharon Fultz (Sacramento Regional Transit), Bob Sleppy (California Department of General Services), and Lee Ann Lambirth (Caltrans) provided access to colony sites. Ron 258 RECENT PURPLE MARTIN DECLINES IN SACRAMENTO, CALIFORNIA Schlorff, California Department of Fish and Game, assisted with permit acquisition. Len Lindstrand and Tim Manolis provided useful comments. Airola thanks Fereshteh (Angie) Raygani and Donya and Layla Airola for tolerating the extensive volunteer field time required for this study. LITERATURE CITED Airola, D. A., and Grantham, J. 2003. Purple Martin population status, nesting habitat characteristics, and management in Sacramento. W. Birds 34:235-251. Airola, D. A., and Kopp, D. 2005. Results of the 2005 survey for breeding Purple Martins in the Sacramento region. Central Valley Bird Club Bull. 8:37-44. Airola, D. A., and Kopp, D. 2007. Breeding population status and mortality assess- ment of Purple Martins in Sacramento during 2006. Central Valley Bird Club Bull. 10:34-44. Airola, D. A., Kopp, D., and Kostka, S. 2004. Purple Martin population status and colonization patterns in the Sacramento region in 2004. Central Valley Bird Club Bull. 7:71-77. Airola, D. A., Kopp, D., and Thomas, K. 2008. Breeding population status, repro- ductive success, and mortality of Purple Martins in Sacramento in 2007. Central Valley Bird Club Bull. 11:25-36. Airola, D. A., Kopp, D., Thomas, K, and Kostka, S. 2009. Effects of construction activities on a Purple Martin nesting colony in Sacramento, California. Central Valley Bird Club Bull. 12:8-16. Airola, D. A., and Williams, B. D. C. 2008. Purple Martin (Progne subis ), in California Bird Species of Special Concern: A ranked assessment of species, subspecies, and distinct populations of birds of immediate conservation concern in California (W. D. Shuford and T. Gardali, eds.), pp. 293-299. Studies of Western Birds 1. W. Field Ornithol., Camarillo, CA, and Calif. Dept. Fish and Game, Sacramento. Elwood, C., Wickland, S., Kostka, S., and Airola, D. A. 2009. Success continues for pilot nest box program at Shelter Cove, California. Purple Martin Update 18(2):27-29. Kopp, D., and Airola, D. A. 2007. First documented nesting by the Purple Martin in Placer County, California, in nearly a century. Central Valley Bird Club Bull. 10:68-70. Kostka, S., Airola, D. A., and Switzer, G. 2008. Recent use of nest boxes by Purple Martins in northern California. Purple Martin Update 17(3): 13-15. Leeman, T. S., Airola, D. A., and Kopp, D. 2003. 2003 status of breeding Purple Martins in Sacramento. Central Valley Bird Club Bull. 6:61-68. Lindstrand, L. III. 2008. Purple Martin distribution and nesting habitat at Shasta Lake, California. W. Birds 39:166-170. Pyle, P. 1997. Identification Guide to North American birds, Part 1. Slate Creek Press, Bolinas, CA. Stutchbury, B. J. 1991. Coloniality and breeding biology of Purple Martins ( Progne subis hesperia) in saguaro cacti. Condor 93:666-675. Williams, B. D. C. 1998. Distribution, habitat associations, and conservation of Purple Martins in California. M.S. thesis, Calif. State Univ., Sacramento. Accepted 13 Mag 2009 259 FURTHER DECLINE IN NEST-BOX USE BY VAUX’S SWIFTS IN NORTHEASTERN OREGON EVELYN L. BULL, U.S. Lorest Service, PNW Research Station, La Grande, Oregon 97850; ebull@fs.fed.us CHARLES T. COLLINS, Department of Biological Sciences, California State Uni- versity, Long Beach, California 90840; ccollins@csulb.edu ABSTRACT: Populations of the Vaux’s Swift ( Chaetura vauxi), a species of conservation concern, are declining in the Pacific Northwest. We compared the number of swifts nesting in boxes 2007-2008 to those nesting in the same boxes 1999-2002 to determine if numbers had changed. There were 51 nest attempts in the earlier 4-year period but only two to five nest attempts in the later 2-year period, an average decline of 72% in nest-box use. The cause of the decline is unknown. Northern Plying Squirrels (Glaucomas sabrina), Bushy-tailed Woodrats ( Neotoma cinerea), and Red Squirrels ( Tamiasciurus hudsonicus ) usurped some of the nest boxes. Seven of seven swifts tested negative for the antibodies of West Nile virus. Conservation measures that protect existing nest and roost sites and create additional nesting sites (nest boxes and chimneys) would help ensure that habitat is not limiting Vaux’s Swift populations. From 1980 to 1999, on the basis of Breeding Bird Surveys, numbers of Vaux’s Swifts ( Chaetura vauxi) declined across their breeding range in the United States with a significant and steep decline in Washington (Bull and Collins 2007). Although the Vaux’s Swift has no official status as threatened or endangered in any state within its breeding range, it is listed as a species of greatest conservation need in Washington’s Comprehensive Wildlife Conservation Strategy (Washington Department of Fish and Wildlife 2005) and as a species of special concern in California (Hunter 2008). In 2003 in northeastern Oregon, Bull (2003a) recorded the swifts in only 46% of the forest stands they occupied in 1991 and significantly fewer swifts on survey transects. The cause for the decline in numbers of Vaux’s Swifts in the Pacific Northwest is unknown, although loss of forest habitat, weather patterns, and disease are all potential factors. In addition, modification of their forest habitat, in the form of the harvest of large-diameter hollow trees and snags, eliminates the swifts’ potential nest and roost sites. Other forest management that decreases the incidence of heartrot and reduces aerial insects could also reduce the number of the swifts’ nesting and roosting sites as well as their prey (Bull 2003a). Currently, there is no quantitative information about the effect of adverse conditions, including weather and pesticide use, during migration or in wintering areas. West Nile virus ( Flavivirus ) has spread rapidly across North America, caus- ing deaths in humans, birds, mammals, and reptiles. Tens of thousands of birds have died, and local declines have been documented (McLean 2006, Clark et al. 2006), although it is difficult to attribute region-wide declines to West Nile virus specifically (Marra et al. 2004). In New York, 33% of the birds tested in 2000 were positive for West Nile virus (Kramer and Bernard 2001). In Ohio, Marshall et al. (2006) found 33% of Northern Cardinals {Cardinalis cardinalis ) to be seropositive. Migratory birds are likely agents of the rapid spread of the disease (Rappole et al. 2000, McLean 2006). 260 Western Birds 40:260-266, 2009 FURTHER DECLINE IN NEST-BOX USE BY VAUX’S SWIFTS IN OREGON Populations of the Chimney Swift ( Chaetura pelagica ), closely related to the Vaux’s Swift, have been adversely affected by West Nile virus (Komar 2003). The arrival of West Nile virus in the Pacific Northwest may be yet another factor increasing mortality and depressing Vaux’s Swift populations. The 103 nest boxes erected for swifts in northeastern Oregon in 1998 (Figure 1; Bull 2003b) provide a unique opportunity for the breeding activities of Vaux’s Swifts to be quantified and for adults and nestlings to be captured and tested for the virus (Bull 2003a). In addition, these boxes provided us the opportunity to compare numbers of swifts using them at different times. The objectives of this study were to compare the numbers of swifts nesting in boxes 1999-2002 and 2007-2008 and to determine if West Nile virus antibodies are present in nesting Vaux’s Swifts. METHODS We monitored 103 nest boxes erected for Vaux’s Swifts at 12 localities in the Wallowa-Whitman and Umatilla national forests within 50 km of La Figure 1. Nest box designed for Vaux’s Swift, installed near La Grande, Oregon, in 1998 . Photo by Evelyn L. Bull 261 FURTHER DECLINE IN NEST-BOX USE BY VAUX’S SWIFTS IN OREGON Grande in northeastern Oregon. All boxes within a 5.4-km radius were con- sidered to be at a single locality because 5.4 km was the maximum distance Bull and Beckwith (1993) found that radio-tagged swifts traveled from the nest while rearing young in northeastern Oregon. Nest boxes were 3.5 m deep and 30 cm square and put in trees 10-15 m above the ground. Ap- proximately one-third of the boxes were installed in each of three habitat types: late-seral stands of Grand Fir ( Abies grandis), harvested stands of Grand Fir, and stands of Ponderosa Pine ( Pinus ponderosa). We determined use of the boxes by climbing to them 1999-2002 during an earlier study (Bull 2003b) and again in 2007 and 2008 for this study. In late June or early July we inspected the walls of each box with a flash- light for stick nests (Figure 2) or swifts and collected its contents, which we inspected for eggshell fragments, sticks, and the swifts’ fecal material. The presence of small white eggshell fragments indicated that eggs had been laid in the box. The presence of more than 300 cm 3 of the swifts’ distinctive fecal material, containing lots of insect chitin, was a good indication that young had fledged successfully (Bull 2003b). The presence of only small sticks suggested that a nest had been started, but we did not consider these boxes to have been active if they contained no eggshell fragments. Some of the original 103 boxes had fallen or the trees to which they were attached had died and were unsafe to climb; only 92 and 86 boxes were checked in 2007 and 2008, respectively. We netted swifts at active nests in early August 2007 and 2008 to col- lect a blood sample (<0.2 ml) from either the brachial vein or a toe nail. Figure 2. Nest of Vaux’s Swift in a box mounted on a tree near La Grande, Oregon. Photo by Evelyn L. Bull 262 FURTHER DECLINE IN NEST-BOX USE BY VAUX’S SWIFTS IN OREGON The blood sample was put on filter paper containing preservative and dried and sent to Orange County Vector Control (Orange County, California) for testing for West Nile virus antibodies by the blocking ELISA technique (Jozan et al. 2003). RESULTS We found substantially fewer swift nests in 2007 (2 active nests) and 2008 (5 active nests) than during the earlier study (Bull 2003b) (Table 1). In the earlier 4-year period, 1999-2002, there were 51 nest attempts (in 30 different boxes). From 1999 to 2002 there were 12.75 nest attempts per year and a mean of 13.1% of the boxes were used each year. In 2007 and 2008 these figures were 3.2 and 4.0%, respectively. In 2007, one box contained six live young swifts and a second box con- tained three dead nestlings; on the basis of the amount of fecal material in the box we presumed some additional nestlings had fledged. In 2008, the five active nest boxes contained (1) at least four nestlings, (2) one nestling (accumulated fecal material indicated that other young had already fledged); (3) eggshells and a detached wing; (4) four whole eggs on the box’s floor; and (5) eggshells and a destroyed nest. These findings suggest that both nest attempts in 2007 and two of the five in 2008 were successful in hatching and probably fledging young. We made no direct observations of Vaux’s Swifts using nest boxes from 2003 to 2007. Ten additional boxes, however, had fecal material and/ or egg shell fragments implying that swifts had nested in them sometime between 2003 and 2006. Nine of these contained >300 cm 3 of fecal mate- rial, suggesting that young could have fledged from them (Bull 2003b). The one box with only eggshell fragments did not contain any fecal material, indicating this attempt failed. There were two boxes with 2000 and 2400 cm 3 of the swifts’ fecal material that we assumed represented at least 2 years of successful nesting. Table 1 Variation by Year in Vaux’s Swift Use of Nest Boxes in Northeastern Oregon 1999-2002 Year(s) 2007 2008 Swift nests Boxes with swift eggs or nestlings 30 2 5 Nest attempts (boxes used) 51 2 5 Active nests each year 10-15 2 5 Nest attempts per year (mean) 12.75 2 5 Percent of active nests that fledged young 53% 100% 40% Squirrel nests Boxes with lichen or grass nests 58 73“ 49 Years of monitoring 4 1 1 Boxes checked 103 92 86 “Accumulated from 2003 to 2007. 263 FURTHER DECLINE IN NEST-BOX USE BY VAUX’S SWIFTS IN OREGON Two adult swifts and four nestlings in one nest box in 2007 and one nest- ling in 2008 were tested for WNV antibodies; all tests proved to be negative. While checking the nest boxes, we saw or heard free-flying swifts at only 4 of the 12 localities in 2007 and at 3 of the 12 localities in 2008. A variety of other birds and mammals also used the nest boxes, making some of them unavailable to swifts. Northern Flickers ( Colaptes auratus) and Pileated Woodpeckers (, Dryocopus pileatus) roosted in the boxes regularly, leaving fecal material and wood chips. Pileated Woodpeckers had excavated additional entrance holes in five boxes by 2002 and in 19 boxes by 2008 (53% in stands of pine, 31% in harvested stands of fir, and 16% in late-seral stands of fir). We consider nest boxes with additional holes excavated in them unsuitable for swift nesting. In 2007, seven boxes contained live Northern Flying Squirrels ( Glauco- mas sabrinus), and three boxes contained Bushy-tailed Woodrats ( Neotoma cinerea). In 2008, Northern Flying Squirrels again occupied seven boxes, and unweaned Red Squirrels ( Tamiasciurus hudsonicus) occupied one box. Nests of lichen or grass, presumably constructed by squirrels or woodrats, were found in 73 boxes (79%) in 2007. In 2008, rodents had built new nests in 49 of the boxes (56%). This one-year accumulation is greater than the total of 58 boxes with squirrel nests during the longer 1999-2002 study period. During that time, the percentage of the 58 swift nests that fledged young was 63% in boxes without a squirrel nest (n = 46) and 17% in boxes with a squirrel nest (n = 12). The use of boxes in the various forest types changed over time. The swifts’ use of late-seral stands of Grand Fir decreased but their use of stands of harvested Grand Fir and Ponderosa Pine increased. In all three forest types squirrels’ use of boxes increased (Bull pers. obs.). Unfortunately, there are no data on squirrel population sizes during the study period. DISCUSSION The number of swifts nesting in boxes in northeastern Oregon decreased by 72% since the previous 1999-2002 study. In addition, in 2007 and 2008 Vaux’s Swifts were seen or heard in only 4 and 3 of the 12 localities, respectively, whereas from 1997 to 2002 they were seen or heard in 11 of the 12 localities. Numerous factors could influence Vaux’s Swifts’ use of nest boxes in north- eastern Oregon. Some of these would include the size of the swift population in the area, the presence of squirrels in the boxes, and the condition of the boxes. Although all seven of the swifts captured tested negative for antibod- ies to West Nile virus, it might be argued that swifts that contacted the virus might have already succumbed to the disease, thus reducing the overall swift population in the study area. A wider program of West Nile virus testing in Vaux’s Swifts would be informative. Squirrels occupying a box may make it less attractive, as evidenced by the success rate in boxes with squirrel nests being lower than in those without squirrel nests. As the number of squirrel nests in the boxes increased, the number of swift nests decreased. In addi- tion to just appropriating the nest boxes, squirrels and woodrats could also be predators of eggs and hatchlings. 264 FURTHER DECLINE IN NEST-BOX USE BY VAUX’S SWIFTS IN OREGON Among the 92 and 86 boxes that could be checked in 2007 and 2008, respectively, some were not suitable for swift nesting for several reasons. In 58 hollow trees used by Vaux’s Swifts the average depth of natural cavities was 4.1m (Bull and Collins 2007), and the nest was attached to the chamber wall 2.3 m below the entrance hole. When given a choice of nest boxes 1 .2 m, 2.4 m, and 3.5 m deep, swifts frequently nested in the boxes 3.5 m deep, nested only once in a box 2.4 m deep, and never nested in the boxes 1.2 m deep (Bull 2003b). In the earlier study period (1999-2002), nest boxes were cleaned annually, so the accumulation of lichen and grass that squirrels brought in as nest material was minimal — usually less than 0.3 m deep on the bottom of the box. In 2007, after 5 years with no box maintenance, squirrel-nest material was more than 1.3 m deep in some of the boxes. By 2007 the accumulation of such large amounts of nest material may have made these boxes less at- tractive to the swifts by concealing their depth. The cleaning of the boxes in 2007 could have been partially responsible for the increase in nesting pairs in 2008; two of the swift nests that year were in boxes that had accumulated squirrel-nest material in them in 2007 and were thus not suitable. Another explanation for some of the reduction in the swifts’ use of the boxes is the higher incidence of Pileated Woodpecker entrance holes: in 19 boxes in 2008, in only five boxes 1999-2002. Additional entrance holes may make the box less attractive to swifts, perhaps because it is less secure from predators and the increased light may be undesirable. We had not an- ticipated that Pileated Woodpeckers would use these nest boxes for roosting. The excavation of additional holes in the boxes provides the woodpeckers with multiple avenues of escape should a predator enter the box. Multiple holes are also characteristic of Pileated Woodpecker roosts in hollow trees (Bull and Jackson 1995). Only one stand of forest, containing four boxes, had been altered by fuel reduction since 2002, so it is unlikely that habitat alteration since 2002 significantly influenced the swift’s population. Conservation measures to prevent further declines of the Vaux’s Swift population could include increasing opportunities for breeding by putting up nest boxes (Bull 2003b) or building mock chimneys as nest sites, as has been done for the Chimney Swift (Kyle and Kyle 2005). Cleaning of nest boxes just prior to nesting may also increase their use. It is also important to protect existing nest and roost sites (hollow trees and chimneys) in both the breeding range and at migration stop-over locations; their loss might be yet another factor contributing to the apparent decline of Vaux’s Swift we have documented. ACKNOWLEDGMENTS We thank Ed Baird and Jim Harris for climbing to the boxes to check nest contents. Michael Snider and Paul and Georgean Kyle, Tom Ryan, Kathy Molina, and John Hunter made helpful comments on earlier drafts of the manuscript. Robert Cummings at Orange County Vector Control arranged for the testing of blood samples for West Nile virus antibodies. Funding was provided by the California State University Long Beach Foundation and the U.S. Forest Service Pacific Northwest Research Station. 265 FURTHER DECLINE IN NEST-BOX USE BY VAUX’S SWIFTS IN OREGON LITERATURE CITED Bull, E. L. 2003a. Declines in the breeding population of Vaux’s Swifts in northeastern Oregon. W. Birds 34:230-234. Bull, E. L. 2003b. Use of nest boxes by Vaux’s Swifts. J. Field Ornithology 74:394- 400. Bull, E. L., and Beckwith, R. C. 1993. Diet and foraging behavior of Vaux’s Swifts in northeastern Oregon. Condor 96:1016-1023. Bull, E. L., and Collins, C. T. 2007. Vaux’s Swift ( Chaetura vauxi), in The Birds of North America Online (A. Poole, ed.), no. 77 (revised). Cornell Lab of Ornithol- ogy, Ithaca, NY. http://bna. birds. Cornell. edu/beta/species/077. Bull, E. L., and Jackson, J. A. 1995. Pileated Woodpecker (. Dryocopus pileatus), in The Birds of North America (A. Poole and F. Gill, eds.), no. 148. Acad. Nat. Sci., Philadelphia. Clark, A. B., Robinson, D. A., Jr., and McGowan, K. J. 2006. Effects of West Nile virus mortality on social structure of an American Crow ( Corvus brachyrhynchos ) population in upstate New York. Ornithol. Monogr. 60:65-78. Hunter, J. E. 2008. Vaux’s Swift ( Chaetura vauxi), in California Bird Species of Special Concern (W. D. Shuford and T. Gardali, eds.), pp. 254-259. Studies of W. Birds 1,.W. Field Ornithol., Camarillo, CA. Jozan, M., Evans, R., McLean, R., Hall, R., Tangredi, B., Reed, L., and Scott, J. 2003. Detection of West Nile virus infection in the United States by blocking ELISA and immunohistochemistry. Vector Borne Zoonotic Diseases 3:99-109. Komar, N. 2003. West Nile virus: epidemiology and ecology in North America. Advances in Virus Research 61:185-199. Kramer, L. D., and Bernard, K. A. 2001. West Nile virus infection in birds and mam- mals. Annals N. Y. Acad. Sci. 951:84-93. Kyle, P. D., and Kyle, G. Z. 2005. Chimney Swift towers: New habitat for America’s mysterious birds. Texas A. & M. Univ. Press, College Station, TX. McLean, R. G. 2006. West Nile virus in North American birds. Ornithol. Monogr. 60:44-64. Marra, P. P., Griffing, S., Caffrey, C. A., Kilpatrick, M., McLean, R., Brand, C., Saito, E., Dupuis, A. P, Kramer, L., and Novak, R. 2004. West Nile virus and wildlife. BioScience 54:393-402. Marshall, J. S., Zuwerink, D. A., Restifo, R. A., and Grubb, T. C., Jr. 2006. West Nile virus in the permanent-resident bird community of a fragmented Ohio landscape. Ornithol. Monogr. 60:79-85. Rappole, J. J., Derrickson, S. R., and Hubalek, Z. 2000. Migratory birds and spread of West Nile virus in the Western Hemisphere. Emerging Infectious Diseases 6:319-328. Washington Department of Fish and Wildlife. 2005. Washington’s Comprehensive Wildlife Conservation Strategy. Wildlife Division, Wash. Dept. Fish and Wildlife, Olympia, WA. Accepted 1 7 September 2009 266 USE OF A NESTING PLATFORM BY GULL-BILLED TERNS AND BLACK SKIMMERS AT THE SALTON SEA, CALIFORNIA KATHY C. MOLINA, Section of Ornithology, Natural History Museum of Los Angeles County, 900 Exposition Blvd., Los Angeles, California 90007; kmolina@nhm.org MARK A. RICCA and A. KEITH MILES, U.S. Geological Survey, Western Ecological Research Center — Davis Field Station, 1 Shields Avenue, University of California, Davis, California 95616 CHRISTIAN SCHONEMAN, Sonny Bono Salton Sea National Wildlife Refuge, 906 W. Sinclair Road, Calipatria, California 92233 ABSTRACT: In 2006, we constructed an elevated nesting platform at the Salton Sea, California, and monitored its use by Gull-billed Terns and Black Skimmers over three subsequent breeding seasons. Black Skimmers were the first to colonize the platform with a total of five nests in 2006. In 2007 Gull-billed Terns colonized the platform with a total of 28 nests and the number of Black Skimmer nests increased to 20. Neither species nested on the platform in 2008. Low success for both species was probably influenced by at least two factors. First, when both species nested on the platform, nest densities were higher than is typical of their colonies on larger, earthen islands, and colony success may have been reduced by overcrowding. Second, lack of access to water may have reduced chicks’ ability to thermoregulate effectively in the hot environment of the Salton Sea. Refinements to the size, design, and location of artificial nesting habitats are necessary to enhance productivity of colonial ground- nesting birds at the Salton Sea successfully. Artificial or constructed nesting habitats have often been employed to increase the number of suitable breeding sites for ground-nesting colonial birds such as terns (Lampman et al. 1996, Quinn and Sirdevan 1998, Spear et al 2007, Jenniges and Plettner 2008). Such artificial nesting habitats have been provided to (1) accommodate the growth of newly colonizing populations, (2) enhance reproductive success, (3) create novel sites that are relatively free of predators or that reduce competition for space with more aggressive species (such as gulls; Molina 2004), and (4) augment the number of alternative sites when the suitability of existing natural habitats declines with the encroachment of vegetation or when the protective isola- tion of islands diminishes as lake levels fall. Constructed habitats are typically located in lakes, impoundments, or other shallow and protected waters. These nesting habitats often take the more traditional form of sand, rock, and earthen islands (usually formed with bottom sediments). Floating platforms such as rafts (Dunlop et al. 1991, Lampman et al. 1996) also have been used when the deposition of dredged sediments is infeasible. When natural nesting sites are limited, rafts and barges have been used to promote the colonization of alternative nesting habitats when land-use conflicts arise and colony relocation is desired (Collis et al. 2002). In some cases breeding larids have spontaneously (i.e., without the aid of audio-visual attractants such as decoys or call broadcasters) colonized the surfaces of existing floating or fixed platforms such as moored barges (Molina pers. obs.), the surfaces of navigational buoys (Karwowski et al. 1995), or even the flat gravel roof tops of buildings (Fisk 1975, Coburn et al. 1997). Western Birds 40:267-277, 2009 267 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS The Salton Sea is the largest terminal lake in California; this highly dynam- ic and productive saline ecosystem lies below sea level and lacks an outlet. Over the past decade, Gull-billed Terns ( Gelochelidon nilotica ) and Black Skimmers (Rynchops niger) breeding at the Salton Sea have experienced the loss of isolated nesting sites (islands) with declines in water level (Molina 2004). Since 2001, the continued decline of water levels has resulted in additional losses of the traditional nesting islands of Obsidian Butte, Morton Bay, and Elmore Desert Ranch by increasing mammalian predators’ access to these sites and further intensifying waterbirds’ demand for undisturbed nesting and resting sites (Molina 2004, 2007). There is no precedent in California for breeding Gull-billed Terns and Black Skimmers nesting on habitats other than earth- or gravel-based islands (Molina 2008a, b) or rafts that mimic islands (Molina 2007). However, Gull- billed Terns (Molina and Erwin 2006), Least Terns ( Sternula antillarum ; Krogh and Schweitzer 1999), and Black Skimmers (Coburn et al. 1997) have spontaneously colonized elevated sites such as flat gravel roofs in coastal Louisiana and Florida. Given the propensity of larids in general to use artificial sites in other parts of North America, in 2006 we constructed a fixed and elevated platform at the south end of the Salton Sea to address the decline in the number of suitable colony sites and to enhance nesting for the Gull-billed Tern and its close nesting associate, the Black Skimmer. In this paper we report the number of breeding pairs, nest attempts, and young of Gull-billed Terns and Black Skimmers on this novel habitat and summarize the phenology of its use in 2006, 2007, and 2008. We also discuss the disadvantages and potential refinements of artificial nesting habitats at the Salton Sea. METHODS In April 2006, United States Geological Survey personnel in cooperation with the U. S. Fish and Wildlife Service constructed an elevated platform at the southeast end of the Salton Sea approximately 400 meters north of the northern seawall boundary of the Sonny Bono Salton Sea National Wildlife Refuge (Figure 1). Our criteria for the platform’s location were that it be (1) placed on refuge property, (2) sufficiently isolated by water over a period of years to deter access by mammalian predators, (3) observable from a distance to minimize disturbance to nesting birds, and (4) in close proximity to traditionally active colony sites such as the islands within one of the refuge’s impoundments (D pond) adjacent to the northern seawall. The wooden platform measured 4.9 x 7.3 m with approximately 35 m 2 of potential nesting area (Figure 2). The platform was supported by twelve 10 cm x 10 cm x 3.7 m posts driven about 1.5 m into the sediment, with the deck, consisting of %-inch tongue-and-groove plywood, placed approxi- mately 1.2 m above the water’s surface. The average elevation of the Salton Sea in April 2006 was -69.2 m (-227 ft), decreasing to -69.4 m (-227.8 ft) in April 2008 (Imperial Irrigation District); water depth at this location during the study was about 1 m. The deck was covered with a layer of about 2.5 cm of mixed sand and gravel which was then covered with a 10-cm layer of crushed barnacles collected from the Salton Sea’s shoreline. Three plywood 268 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS Figure 1. Geographical location of the study area and nesting platform. Rock Hill ^ Nesting platform 2008 nesting site 269 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS Figure 2. Nesting platform at completion and close up of nesting substrate. Note solar panel at upper center and Gull-billed Tern decoys at left. Photos by Mark Ricca ramps measuring 3.6 x 1.2 m and extending below the water’s surface were attached to the north, south, and east ends of the deck to facilitate the return of nonvolant young to the deck. The platform was equipped with visual and audio attractants in the form of 20 Gull-billed Tern decoys (Mad River Decoys, Waitsfield, VT) and a solar-powered sound system (MurreMaid, Bremen, ME) to continuously broadcast the Gull-billed Tern’s vocalizations (Kress 1983). A motion-activated camera was initially installed but removed 270 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS in mid-May 2006 along with ten decoys to increase the amount of potential nesting area. The solar-powered sound system was inoperable in 2007 and removed from the platform prior to the 2008 breeding season. All of the re-entry ramps were damaged or destroyed during high winds in April 2006, and subsequent attempts at repair and redesign were unsuccessful. We monitored the presence or absence of terns and skimmers sitting or standing on the platform from at least early April through August of each year at roughly weekly intervals. In 2006 multiple censuses were made in most weeks, but there were gaps in observations of up to 14 days in July and August. In 2007 and 2008, we censused the platform at least every other week, often more frequently. Less regular observations continued into mid- October in 2006 and mid-September in 2007 and 2008. Because of lands’ semi-precocial nature and sensitivity to disturbance, studies of birds nesting on rafts or elevated sites such as roof tops and our platform confront inherent dif- ficulties because disturbance by investigators often reduces a colony’s success (e.g., Krogh and Schweitzer 1999). Therefore, we observed from elevated locations from either the overlook at Rock Hill (about 500 m southwest of the platform) or from the refuge’s northern seawall boundary (Figure 1). Our observations were distributed throughout the daylight period. The duration of observations ranged from 15 to 30 minutes, and for each observation we recorded the peak number of adult Gull-billed Terns and Black Skimmers on the platform as well as the platform’s use by other species. We briefly visited the platform once or twice each season to confirm the number of suspected nests. To minimize the potential of older chicks being flushed from the plat- form, our first visit of the season was timed to precede expected hatching. When feathered young were not observed when expected, we made a second visit to confirm their absence. We did not uniquely mark individual birds or systematically monitor fates of individual nests, which precluded distinguish- ing renesting attempts from those of new pairs or robustly estimating nesting and fledging success. Instead, we report (1) the frequency of platform use by both species over time and (2) the peak numbers of active nest attempts and observed young on any single observation to estimate nesting use and reproductive success. Because in both species both male and female invest heavily in incubation and chick-rearing duties, and pairs raise only one brood per season, we assumed that the peak number of nest attempts accurately reflected a minimum number of nesting pairs for the platform. RESULTS Peak Numbers and Phenology of Platform Use In 2006, the platform’s first year, we observed only one adult Gull-billed Tern briefly in early April (Figure 3). In contrast, Black Skimmers were on the platform continuously from 17 June to 13 October with an average of 9 adults present (SD = 3, n = 14 observations). In any single day’s observa- tion, the maximum number of adults was 14 on 24 June (Figure 3). Four skimmer nests were first noted on 5 July, although exact initiation dates were unknown. A fifth nest was established by mid-September, inferred from the presence of four small downy young on 13 October. 271 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS -B- GBTE -a- BLSK Figure 3. Phenology of platform use in 2006, 2007, and 2008 by Gull-billed Terns and Black Skimmers. Data reported are the peak numbers of adults present. 272 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS In 2007, Gull-billed Terns increased interest in the platform with an aver- age of 17.2 (SD = 11, n = 22 observations) adults present per visit from 7 May through 11 July (Figure 3). The maximum number of adult terns was 45 on 1 June. The first Gull-billed Tern nests were established between 4 and 7 May. However, by 15 June the number of adult terns present on the platform declined to seven. Over the next 10 days, when young should have been approaching fledging age and colony activity (adults attending and feeding chicks) should have been at its peak, the number of adults on the platform remained low (Figure 3). In contrast, Black Skimmers were continuously present from 10 May to 24 August (mean number and SD = 44 ± 20 adults, n = 23 observations). In 2007 the maximum number of adults on the platform was 100 on 24 June (Figure 3) and nests were first established between 24 May and 1 June. In 2008, Gull-billed Terns were noted on the platform in early April (Figure 3), but use was brief and limited to the period around dusk. A maximum of 43 terns was present on the platform between 3 and 1 1 April. The terns were always absent on return visits at first light on the following mornings. Black Skimmers were never noted to land on the platform in 2008 (Figure 3). Other species that occasionally rested on the platform when it was unoc- cupied by terns and skimmers were two Black-necked Stilts ( Himantopus mexicanus), two Great Blue Herons ( Ardea herodias), several Double- crested Cormorants ( Phalacrocorax auritus), mainly on the platform’s underlying cross beams, and, in 2008 only, as many as 26 Brown Pelicans (. Pelecanus occidentalis). Nesting Activity The platform was used for nesting in two of the three years of our study by at least one of the two target species. In the first season of its availability (2006), the Black Skimmer was the only species to nest on the platform, establishing at least five nests. The number of active nests peaked at four (0.11 nests/m 2 ) on 5 July. The peak number of five young was observed on 13 October; their size and development ranged from one nearly fledged chick to four small, downy chicks. In 2007, Gull-billed Terns colonized the platform with a total of 28 nests. A maximum of three tern chicks, all small and downy, was observed between 1 and 15 June, but no young or eggs were present on a visit on 3 July. In 2007, the number of nest attempts by Black Skimmers also increased from the previous year. The number of active skimmer nests peaked at 11 (0.31 nests/m 2 ) on 27 July. A maximum of eight young (four large and feathered, four small and downy) was observed on 27 July. The number of active tern and skimmer nests taken together peaked at 32 (0.9 nests/m 2 ) on 1 June. In 2008 neither species nested on the platform. DISCUSSION Black Skimmers immediately colonized the nesting platform in 2006 with at least four pairs (representing <1% of the Salton Sea’s breeding population in that year) and increased to 20 pairs in 2007 (representing ~7% of the 273 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS annual population). In contrast, at least 28 pairs of Gull-billed Terns (repre- senting nearly 22% of the annual breeding population) initiated nesting on the platform in 2007. The lack of nesting by either species on the platform in 2008 may have been due to the proximity of a Great Horned Owl ( Bubo virginianus ) nest, active from early May through June, at Rock Hill. This owl is known to disrupt nesting colonies of Common ( Sterna hirundo ) and Least Terns in Maryland (Erwin et al. 2007). Gull-billed Terns and Black Skimmers colonized other breeding sites at the Salton Sea in 2008: terns nested on earthen islands (available since 2006) in a series of impoundments about 3 km to the east of the platform, and late that season skimmers nested on a small, newly available raft in the adjacent D pond (Figure 1). Terns also nested on a natural islet about 2.7 km to the southwest of the platform, a site where the species has nested occasionally since 1992 (Molina 2004). At the Salton Sea, Gull-billed Terns typically initiate nesting between late April and early May and Black Skimmers generally do so between late May and early June (Molina unpubl. data). In the year that each species colonized the platform, 2006 for the skimmer and 2007 for the tern, the timing of nest initiations on the platform was later than at other sites used in those years, suggesting the platform was less attractive than other sites and used by pairs whose earlier attempts had failed or by late nesters that could not find space elsewhere. The ratio of young to nest attempts, which may be a useful indicator of colony success, was low for both species, particularly the tern. In 2007, the number of adult Gull-billed Terns diminished rapidly at a time when chicks near fledging age and a corresponding high level of nest-site attendance by adults were expected (Figure 3). These observations, accompanied by adult terns’ delivering food to the platform infrequently, all indicated poor colony success. The causes of the low success for both species are conjectural. The concomitant increase in the larger, more aggressive skimmer at a vulnerable period (hatching) for the tern may have caused nest failure, chick deaths, and early abandonment of the site. Other factors possibly contributing to low colony success were (1) the platform’s size potentially constraining nest densities above the optimum and (2) its unprotected location with respect to wind-driven waves on the surface of the Salton Sea coupled with the region’s harsh environment. Few studies of raft- or barge-nesting larids have specifically addressed the causes of mortality of young terns on these artificial sites. In a comparison of the relative proportions of Common Terns banded as fledglings in Britain and subsequently resighted in Africa, Norman (1987) found terns nesting on elevated platforms to be marginally more successful than those breeding at other sites, concluding that the main benefit of the platform was its inaccessi- bility to mammalian predators. Studies of raft-nesting Caspian and Common terns on Lake Ontario noted dead juveniles washed up on the mainland shore (Lampman et al. 1996), suggesting that some missing chicks, after falling into the water, were unable to navigate the ramps and return to the raft (Dunlop et al. 1991). We routinely searched along the northern seawall, the closest point on the shore and downwind from prevailing winds, for dead or live chicks displaced from the platform but never encountered any. Nearest-neighbor distances of Gull-billed Tern nests on constructed 274 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS islands at the Salton Sea are highly variable, ranging from 0.8 to 3.85 m (mean = 2.32, SD = 1.1, n = 14; Molina unpubl. data), but within the range of 0.3-20 m reported by Gochfeld and Burger (1996). Although we did not measure nearest-neighbor distances for nests on the platform, the density of concurrent tern and skimmer nests on it in 2007 (maximum = 0.9 nests/m 2 ) was greater than the average density of 0.09 nests/m 2 (SD = 0.06, n = 9 colony sites) for Gull-billed Tern nests on constructed islands in 1999 and 2001 (Molina unpubl. data). The nesting asynchrony of Gull- billed Terns and Black Skimmers that often promotes a degree of temporal separation at shared colony sites at the Salton Sea (Molina unpubl. data) was not observed on the platform in 2007; nest densities and the likely nearest-neighbor distances we observed for terns and skimmers combined may have exceeded the Gull-billed Tern’s levels of tolerance and suggests that the platform’s area may have been insufficient to support the typically sized colonies of 30-50 pairs for both species (Molina and Erwin 2006, Gochfeld and Burger 1994). In this unprotected location, our experimental platform was challenged immediately after construction by high wind-driven waves that briefly but completely swept the platform’s deck. During successive periods of wind the platform’s ramps were eventually washed away and a nonvolant chick’s only means of return was eliminated. The loss of the ramps also prevented routine access to water by chicks and brooding adults for thermoregulation. At the Salton Sea, colonial birds that nest in exposed locations experience ambient temperatures that often exceed 42°C and corresponding high rates of insolation. Frequent foot and body soaking are important thermoregula- tory behaviors for adult larids and their young in maintaining normal body temperature via evaporative cooling. These behaviors are also performed by incubating and brooding adults to maintain the eggs and nest at adequate temperature and humidity. Unrelieved heat stress may result in adults aban- doning nests; chicks that lack access to water and constant brooding may die (Grant 1982, Molina pers. obs.). Our platform’s elevated design (Figure 2) more closely approximated the snags and trees in which herons and cormorants commonly nest at the Salton Sea (Molina and Sturm 2004) rather than the islands or rafts (which more closely mimic islands) typically used by larids. Such a design may be more appropriate for nesting by species with altricial or less precocial young and in regions where breeding adults and young are not exposed to high ambient temperatures that require birds have regular access to water for thermoregulation. Even if nesting larids do not use the platform in the future, the smaller-bodied terns and skimmers may still benefit if the platform attracts resting congregations of other large waterbirds, such as the White {P erythrorhynchos) and Brown pelicans and Double-crested Cormorants, away from the terns’ and skimmers’ island nesting sites. In past years, resting flocks of these larger species have destroyed substantial numbers of tern and skimmer eggs by trampling them (Molina 2004, 2007). More study is required for the size and design of constructed sites to be optimized for terns and skimmers in the region. We did find that audio at- tractants appeared unnecessary for colonization. Although artificial nesting substrates such as rafts and islands are usually installed in protected waters 275 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS (Dunlop et al. 1991, Lampman et al. 1996, Molina 2007), our placement of the nesting platform was constrained by logistical and environmental considerations, including a paucity of such protected impoundments and a fluctuating sea-surface elevation and shoreline. If at the Salton Sea plat- forms can be successfully adapted to mimic islands more closely, that is, to maintain a degree of direct connectivity between the deck’s surface and surrounding water (e.g., with permanent ramps or “skirts”), they will most certainly require placement in protected waters. ACKNOWLEDGMENTS Platform construction and initial monitoring costs were funded by the U. S. Fish and Wildlife Service Migratory Bird Office, Portland, Oregon. The Sonny Bono Salton Sea National Wildlife Refuge funded monitoring of nesting colonies and provided additional support for platform construction. We thank Nanette Seto for project support and Tom Anderson, Meg Harper, Susan Kaveggia, Cody Massing, Elizabeth Morgan, Keegan Ramey, and Amy Story for field assistance. We thank Brian Brinegar and Amy Story for construction assistance and Scott Story for construction design. Manuscript drafts were thoughtfully reviewed by Josh Ackerman, Tom Anderson, Charles Collins, Mike Erwin, Kimball Garrett, and MaryEllen Mueller, and Robert Gill provided editorial guidance. The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the U.S. Fish and Wildlife Service or U.S. Geological Survey. Use of trade names does not imply governmental endorsement. LITERATURE CITED Coburn, L., Cobb, D., and Gore, J. 1997. Management opportunities and techniques for roof- and ground-nesting Black Skimmers in Florida. Final performance report. Florida Game and Fresh Water Fish Commission, 620 S. Meridian St., Tallahassee, FL 32399. Collis, K., Roby, D. D., Thompson, C. W., Lyons, D. E., and Tirhi, M. 2002. Barges as temporary breeding sites for Caspian terns: Assessing potential sites for colony restoration. Wildlife Soc. Bull. 30:1140-1149. Dunlop, C. L., Blokpoel, H., and Jarvie, S. 1991. Nesting rafts as a management tool for a declining Common Tern (Sterna hirundo) colony. Colonial Waterbirds 14:116-120. Erwin, R. M., Miller, J., and Reese, J. 2007. Poplar Island Environmental Restoration Project: Challenges in waterbird restoration on an island in Chesapeake Bay. Ecol. Restor. 25:256-262. Fisk, E. J. 1975. Least Tern: Beleaguered, opportunistic and roof-nesting. Am. Birds 129:15-16. Gochfeld, M., and Burger, J. 1994. Black Skimmer (Rynchops niger), in The Birds of North America (A. Poole and F. Gill, eds.), no. 108. Birds N. Am., Philadel- phia. Gochfeld, M., and Burger, J. 1996. Family Sternidae (Terns), in Handbook of the Birds of the World (J. del Hoyo, A. Elliott, and J. Sargatal, eds.), vol. 3, pp. 624-667 Lynx Edicions, Barcelona. Grant, G. S. 1982. Avian incubation: Egg temperature, nest humidity, and behavioral thermoregulation in a hot environment. Ornithol. Monogr. 30. 276 USE OF A NESTING PLATFORM BY TERNS AND SKIMMERS Jenniges, J. J., and Plettner, R. G. 2008. Least Tern nesting at human created habitats in central Nebraska. Waterbirds 31:274-282. Karwowski, K., Gates, J. E., and Harper, L. H. 1995. Common Terns nesting on navigational aids and natural islands in the St. Lawrence River, New York. Wilson Bull. 107:423-436. Kress, S. W. 1983. The use of decoys, sound recordings, and gull control for re- establishing a tern colony in Maine. Col. Waterbirds 6:185-196. Krogh, M. G. , and Schweitzer, S. H. 1999. Least Terns nesting on natural and artificial habitats in Georgia, USA. Waterbirds 22:290-296. Lampman, K. P., Taylor, M. E., and Blokpoel, H. 1996. Caspian Terns ( Sterna caspia ) breed successfully on a nesting raft. Col. Waterbirds 19:135-138. Molina, K. C. 2004. Breeding larids of the Salton Sea: Trends in population size and colony site occupation. Studies Avian Biol. 27:92-99. Molina, K. C. 2007. The breeding of terns and skimmers at the Salton Sea, 2007. Final report to Sonny Bono Salton Sea National Wildlife Refuge, 906 W. Sinclair Rd., Calipatria, CA 92233. Molina, K. C. 2008a. Gull-billed Tern (Gelochelidon nilotica), in California Bird Species of Special Concern: A ranked assessment of species, subspecies, and distinct populations of birds of immediate conservation concern in California (W. D. Shuford and T. Gardali, eds.), pp. 187-1972. Studies of Western Birds 1. W. Field Ornithol., Camarillo, CA, and Calif. Dept. Fish and Game, Sacramento. Molina, K. C. 2008b. Black Skimmer (Rynchops niger), in California Bird Species of Special Concern: A ranked assessment of species, subspecies, and distinct populations of birds of immediate conservation concern in California (W. D. Shuford and T. Gardali, eds.), pp. 199-204. Studies of Western Birds 1. W. Field Ornithol., Camarillo, CA, and Calif. Dept. Fish and Game, Sacramento. Molina, K. C., and Erwin, R. M. 2006. The distribution and conservation status of the Gull-billed Tern ( Gelochelidon nilotica) in North America. Waterbirds 29:271-295. Molina, K. C., and Sturm, K. K. 2004. Annual colony site occupation and patterns of abundance of breeding cormorants, herons and ibis. Studies Avian Biol. 27:42-51. Norman, D. 1987. Are Common Terns successful at a man-made nesting site? Ring- ing and Migration 8:7-10. Quinn, J. S., and Sirdevan, J. 1998. Experimental measurement of nesting substrate preference in Caspian Terns, Sterna caspia, and the successful colonization of human constructed islands. Biol. Conserv. 85:63-68. Spear, K. A., Schweitzer, S. H., Goodloe, R., and Harris, D. C. 2007. Effects of management strategies on the reproductive success of Least Terns on dredge spoil in Georgia. Southeastern Naturalist 6:27-34. Accepted 4 June 2009 277 BIRDS OF PREY AND THE BAND-TAILED PIGEON ON ISLA GUADALUPE, MEXICO JUAN-PABLO GALLO-REYNOSO, Centro de Investigation en Alimentacion y Desarrollo, A.C. Unidad Guaymas, Carretera a Varadero Nacional km 6.6, Col. Las Playitas, Guaymas, Sonora 85480, Mexico; jpgallo@ciad.mx ANA-LUISA FIGUEROA-CARRANZA, Area de Proteccion de Flora y Fauna Islas del Golfo de California, Oficina Regional Sonora, CONANP-SEMARNAT, Calle Isla del Peruano esquina con Calle Isla de la Rasa, Col. Lomas de Miramar, Guaymas, Sonora 85450, Mexico ABSTRACT : We noted eight species of birds of prey at Isla Guadalupe during ten visits from 1991 to 2003. The most abundant species was the Burrowing Owl ( Athene cunicularia), found throughout the island; second most numerous was the American Kestrel (Falco sparuerius ), widespread but uncommon. The frequency of the kestrel paralleled the population of mice, peaking 1992, a year of El Nino. We observed the Red-tailed Hawk ( Buteo jamaicensis), Osprey (Pandion haliaetus), and Peregrine Falcon ( Falco peregrinus) two or three times each, the Prairie Falcon {F. mexicanus) once. Our records of the Northern Harrier ( Circus cyaneus) and Band-tailed Pigeon (. Patagioenas fasciata) are the first for Isla Guadalupe. Early publications on the birds of Isla Guadalupe (Gaylord 1897, Anthony 1925, Hanna 1925) were devoted largely to the endemic species that “fas- cinated biologists” (Jehl and Everett 1985). Thus the history of some of the taxa now extinct, such as the Guadalupe Caracara ( Caracara lutosa) are well documented, but some others remain scarcely known. Luna-Mendoza et al. (2005) and Quintana-Barrios et al. (2006) summarized the island’s avifauna most recently. Here, on the basis of 10 visits to Isla Guadalupe from 1991 to 2003, we report our observations of the island’s raptors and owls, their relationships with their prey, and an accidental occurrence of the Band-tailed Pigeon ( Patagioenas fasciata). STUDY SITE AND METHODS Isla Guadalupe (29° 00' N, 1 18° 20' W), of volcanic origin, lies within the waters of the California Current, 260 km off Baja California. Northwesterly winds predominate (Berdegue 1957). The orientation of the island and its elongated shape (37 km long and 6. 5-9. 5 km wide) act as a barrier against the prevailing winds and generate updrafts at various altitudes. Fog covers the mid-northern portion of the island almost daily. The climate is semi-arid to temperate, the annual mean temperature is 28° C, and winters are cool, mean temperature in the coldest month (January) being <17° C (Montanez et al. 2000). The northern part of the island is high, culminating at Monte Augusta (137 0 m). This area of the island has a relict forest of cypress trees ( Callitropsis guadalupensis), an almost extinguished stand of pines ( Pinus radiata var. binata), fan palms ( Brahea edulis), island oaks ( Quercus tomentella), grass- lands dominated by several introduced species ( Hordeum murinum, Avena barbata , A. fatua, Bromus spp.), and the introduced tree tobacco (Nicotiana glauca). Most of the middle part of the island is a large grass-covered plateau, 278 Western Birds 40:278-283, 2009 BIRDS OF PREY AND THE BAND-TAILED PIGEON ON ISLA GUADALUPE culminating at Monte Esther (1250 m). To the south, the terrain is lower, with numerous small cinder cones, and vegetated largely by shrubs such as Ambrosia camphorata and introduced grasses (Rico-Cerda 1983). Our observations were made during expeditions to census the Guadalupe Fur Seal ( Arctocephalus townsendi), California Sea Lion ( Zalophus califor- nianus), and Northern Elephant Seal ( Mirounga angustirostris) along the east- ern coast of the island and while censusing Laysan Albatrosses ( Phoebastria immutabilis) at Punta Sur (Gallo-Reynoso and Figueroa-Carranza 1996, Pit- man et al. 2005). Aided with 10 x 50 binoculars, our observations were made from a 24-foot fiberglass skiff and spread over 213 days and 10 expeditions: 11-21 February, 16 June-22 July, and 27 November-7 December 1991, 3-25 February, 21 June-16 August, and 1 1-26 November 1992, 21 June-26 August 1993, 6-12 January 2000, 12-23 February, 2003, and 6-18 July 2003. Birds were also observed from our campsites, one on the southeastern coast of the island in the area called “Sealer’s mins’’ or Corralitos, the other at the barracks of Campo Norte, and while we were traveling by the dirt roads from Punta Sur to Campo Norte, from Caleta del Oeste (west anchorage) to the airstrip in the island’s middle and from Campo Norte to El Aguaje. ANNOTATED LIST Osprey ( Pandion haliaetus). Visitor. Two birds observed flying together near Dos Arroyos (Twin Canyons) on the northeastern coast of the island in February and July 1992. No nests were observed, and no other birds were observed in 2000 and 2003. The Osprey has not been reported from Isla Guadalupe since specimens were collected on 11 July 1922 (Anthony 1925, Hanna 1925) and 25 July 1941 (Bond and Meyer de Schauensee 1944). Karl W. Kenyon visited Guadalupe in 1965 and saw no ospreys but reported two presumed nests near the north end of the island (Jehl and Everett 1985), near where we observed the species. Red-tailed Hawk ( Buteo jamaicensis). Howell and Cade (1954) considered the Red-tailed Hawk “apparently resident until at least 1932.'’ Anthony (1925) reported several in 1922. We observed the Red-tailed Hawk on three occasions. The first was in July 1991 at Monte Augusta, where two individuals ascended in circling flight. One captured a White-winged Dove (, Zenaida asiatica). The White-winged Dove, known from Isla Guadalupe from only two previous reports (Howell and Cade 1954, Mellink and Palacios 1990), is a less likely prey for the Red-tailed Hawk than the Mourning Doves (Z. macroura), now a common resident of the island (population estimated at >2000; Barton et al. 2004, 2005). The second observation, in November 1991, was of a lone hawk flying over the mid-southern portion of the island. The third, also of a single individual, was in July 1993 near the airstrip in the middle of the island, eleva- tion -1000 m. In April 2007 S. Gallo-Corona (pers. comm.) observed a Red-tailed Hawk attack a fledgling Rock Wren ( Salpinctes obsoletus guadeloupensis) that had been banded in the open area between the cypress forest and the pine grove. Guadalupe Caracara ( Caracara lutosa). Extinct (Anthony 1925, Jehl and Everett 1985). No report of living birds in the past 100 years. The history of this species, its taxonomy and extinction, were described by Abbott (1933) and Brown and Amadon (1968). We propose that its extinction was precipitated by the decimation of the fur and elephant seals, eliminating the pups, placental tissue, and carcasses that probably sustained these predators/carrion eaters. American Kestrel ( Falco sparverius). Resident. Bond (1943) described the Gua- dalupe population as a subspecies ( F. s. guadalupensis), though other authors have 279 BIRDS OF PREY AND THE BAND-TAILED PIGEON ON ISLA GUADALUPE not recognized this status. Jehl and Everett (1985) reported its population as small, and Barton et al. (2004, 2005) estimated a population of >15. Quintana-Barrios et al. (2006) suggested that on Guadalupe kestrels nest primarily in rock crevices. We observed kestrels regularly at all seasons from our campsite at Corralitos and along the road from the south end of the island to the cypress forest. In 1991, 1992, and 1993 we noted kestrels feeding on introduced House Mice (Mus muscu/us). Because of the abundance of grass seeds, the mouse population grew exponentially during the years of El Nino (1992-1993). In the summer of 1991, we observed only one mouse at our campsite at Corralitos, but in 1992 and 1993 mice were abundant. At our Corralitos campsite, we captured an average of 14 per night (standard deviation ±5, range 6-21, n = 25 nights) in 1992, 10 ± 6 (range 4-18, n = 27 nights) in 1993, with a simple trap consisting of a 25-liter plastic bucket filled with 6 liters of sea water. We ran a steel wire through the middle of an empty tin can and secured it to the bucket’s wall. For bait, we tied a bit of flour dough to the outside of the can. The trap was set near the rocks surrounding the campsite with a piece of wood to serve as a bridge and make it easier for the mouse to jump to the dough. When the mouse jumped, the can turned, and the mouse fell into the water and drowned. The number of kestrels observed per day during the summer increased significantly from 1991 (mean 0.4 ± 0.5, range 0-1, n = 20 days) to 1992 (mean 1.27 ± 0.4, range 1-2, n = 30 days) then decreased in 1993 (mean 0.9 ± 0.7, range 0-2, n = 40 days). Peregrine Falcon ( Falco peregrinus). Possibly resident. In June 1991 we observed one bird flying and one bird standing on the cliffs of the mid-eastern coast of the island, at an estimated height of 45 m above sea level. In July 1992 we saw a solitary bird flying north along the coast at the Corralitos campsite. The only previous record for Isla Guadalupe is the observation of a single bird on 19 September 1896 by Gaylord (1897). Subsequently, the species has been noted in both summer and winter at the south end and on the west side of the island by Barton et al. (2004, 2005), Luna- Mendoza et al. (2005), and R. A. Erickson (pers. comm.), suggesting that it may now occur regularly and breed. Prairie Falcon (Falco mexicanus). Vagrant. On 18 July 1992, near the Corralitos campsite, we noted a lone individual, identifying it as a Prairie Falcon by its size, generally pale color, and darker axillars. Bryant (1889) reported seeing the species on “two or three occasions” in 1886. Although Jehl and Everett (1985) suspected that Bryant had misidentified the Peregrine Falcon, the coloration of the individual we observed was distinct and corresponded to the Prairie Falcon. Northern Harrier ( Circus cyaneus). Vagrant. We noted a solitary male on 18 July 2003 at an elevation of 500 m in an area of basaltic rocks and grasses along the road from the airstrip to Campo Oeste. The bird was perched on a basaltic boulder then flew away, in its typical low slow flight with wings angled up in a shallow V. We identified it by its grayish color, white rump patch, facial disk, black wingtips and trailing edge of the wing, and long tail. This is the first record of this species at Isla Guadalupe. Barn Owl (Tyto alba). Status uncertain. Sweet et al. (2001) found a flank feather near the airstrip and deposited it at Coleccion Nacional de Aves, Instituto de Biologia, Universidad Nacional Autonoma de Mexico, providing the first documentation of the Barn Owl on the island. In 1991 we found large pellets, larger than those of the Bur- rowing Owl, on the navigational light tower of Caleta Melpomene. The two pellets analyzed contained cat and mouse hairs and storm-petrel ( Oceanodroma sp.) and Black-vented Shearwater (Puf finus opisthomelas ) feathers. Because of their large size and containing cat hair these pellets may represent the Barn Owl. The navy personnel stationed at the military base at Punta Sur and fishermen at Caleta del Oeste have reported seeing large white owls or tecolotes on the island, although their reports have been sporadic (1991, 1993 and 2003). 280 BIRDS OF PREY AND THE BAND-TAILED PIGEON ON ISLA GUADALUPE Burrowing Owl ( Athene cunicularia). Resident; widespread and common (Jehl and Everett 1985, Barton et al. 2004, 2005). Common around our campsite at Corralitos year round and at our campsite at Campo Norte in winter. Burrowing Owls feed on Leach’s Storm-Petrel (Oceanodroma leucorhoa) and on Xantus’s Murrelet ( Synthli - boramphus hypoleucus ), as we found the wings and some other body parts of these birds in the entrances of four owl burrows at the southern end of the island (near the lighthouse in 1991), at some burrows near our campsite at Corralitos (1991-1993), and at one place where the owls perch at the Campo Norte campsite (2003). Barton et al. (2004, 2005) also observed such predation at Islote Negro and Islote Zapato. Of 3 1 Burrowing Owl pellets recovered from a rocky perch near the navigational tower at Campo Norte in February and July 2003 (probably representing food items of a pair of owls over several months), 11 contained crickets, moths, and other insects, including parts such as shields, antennae and legs, 10 contained hair and bones of House Mice, three contained feathers of storm-petrels, one contained the introduced cockroach Periplaneta americana, one contained feathers of the Guadalupe Junco (< -Junco insularis), two contained Burrowing Owl feathers, one contained sand, and, unexpectedly, two contained skin and hair of the elephant seal, which are shed each spring during the seals’ molt on the beach. We infer that the owls scavenge the elephant seals’ shed skin. Band-tailed Pigeon {Patagioenas fasciata). Vagrant. On 12 July 2003, we saw one adult in the canyon that comes down from Aguaje to Campo Norte. It was sitting on a rock overlooking a cliff; when we approached to take photographs it took flight. The bird flew over us several times before coming to rest far away on a towering cliff. We identified it by the white band on the nape, green iridescence on the neck, purplish head and chest, yellow beak, grayish tail with a darker gray band, and grayish wings with darker primaries. It was similar in size to the Rock Pigeon (Columba liuia), found at Punta Sur, where it was introduced by the military personnel and fishermen. This constitutes the first record of the Band-tailed Pigeon for Isla Guadalupe. DISCUSSION The American Kestrel and Burrowing Owl are the only birds of prey defi- nitely known to breed at Isla Guadalupe currently. The status of the Osprey, Red-tailed Hawk, Peregrine Falcon, and Barn Owl, possibly resident in very small numbers, requires further investigation. Most studies of the avifauna of Isla Guadalupe have focused on seabirds and wooded areas, not the grasslands and lowlands where potential prey for raptors are more numerous. Certain insects (principally moths, crickets, and cockroaches), mice, doves, feral cats, and even feral dogs supply prey for raptors throughout the year. Still, there are no detailed studies of these birds’ feeding habits, their relation to the vegetation, or their general ecology on the island. With the eradication of the feral goats responsible for extinctions of several plants (Ezcurra et al. 2005) and the ecosystem’s nascent recovery, such studies should be encouraged. Mexico has taken a major step toward the island’s conservation by declaring Isla Guadalupe a biosphere reserve (Diario Oficial de la Federacion 2005). ACKNOWLEDGMENTS We are indebted to the Secretaria de Marina and to fishermen of “Cooperativa de langosteros y abuloneros de Ensenada,” which provided transportation to and logistic support on the island. Our work was conducted under permits 0596, 2538, 281 BIRDS OF PREY AND THE BAND-TAILED PIGEON ON ISLA GUADALUPE and 4933 from the Secretaria de Desarrollo Urbano y Ecologia and 7095 from the Secretaria de Medio Ambiente y Recursos Naturales, Mexico. We thank the Educa- tion Abroad Program of the University of California and the Direccion General de Intercambio Academico of the Universidad Nacional Autonoma de Mexico for funding three expeditions. The National Geographic Society partially funded one expedition, and the University of California Institute for Mexico and the United States funded two expeditions (1992 and 2003). Thanks to Rex Passion, Dan Crocker, Octavio Maravilla, Carlos Nino, Horacio Cabrera, and David Larsen for their help in the field. Thanks to Sandra Gallo-Corona for allowing us to use her recent observations and commenting on the manuscript. The manuscript benefited from the comments of Dan Anderson, Philip Unitt, and Richard A. Erickson. LITERATURE CITED Abbott, C. G. 1933. Closing history of the Guadalupe Caracara. Condor 35:10-14. Anthony, A. W. 1925. Expedition to Guadalupe Island, Mexico in 1922. The birds and mammals. Proc. Calif. Acad. Sci. (Ser. 4) 14:277-320. Barton, D. C., Lindquist, K. E., Henry, R. W. Ill, and Luna-Mendoza, L. M. 2004. Landbird and waterbird notes from Isla Guadalupe, Mexico. W. Birds 35:186- 196. Barton, D. C., Lindquist, K. E., Henry, R. W. Ill, and Luna-Mendoza, L. M. 2005. Notas sobre las aves terrestres y acuaticas de Isla Guadalupe, in Isla Guadalupe, Restauracion y Conservacion (K. Santos del Prado and E. Peters, compilers), pp. 103-113. Instituto Nacional de Ecologia, Mexico. Berdegue, J. 1957. La Isla de Guadalupe, Mexico. Contribucion al conocimiento de sus recursos naturales renovables. Secretaria de Marina, Direccion General de Pesca e Industrias Conexas, Mexico. Bond, R. M. 1943. Variation in western Sparrow Hawks. Condor 45:168-185. Bond, J., and Meyer de Schauuensee, R. 1944. Fifth George Vanderbilt Expedition (1941). Acad. Nat. Sci. Philadelphia Monogr. 6. Brown, L. , and Amadon, D. 1968. Eagles, Hawks and Falcons of the World. McGraw- Hill, New York. Bryant, W. E. 1889. A catalogue of the birds of Lower California, Mexico. Proc. Calif. Acad Sci. (Ser. 2) 2:237-320. Diario Oficial de la Federacion. 2005. Decreto por el que se declara area natural protegida con la categoria de reserva de la biosfera, la zona marina y terrestre que incluye a la Isla Guadalupe. Lunes 25 de abril. Secretaria de Medio Ambiente y Recursos Naturales, Mexico. Ezcurra, E., Aguirre-Munoz, A., Salas-Flores, L. M., Santos del Prado, K., Garcia- Gutierrez, C., Luna-Mendoza, L., and Peters, E. 2005. Plan de erradicacion de especies introducidas en Isla Guadalupe, in Isla Guadalupe, Restauracion y Conservacion (K. Santos del Prado and E. Peters, compilers), pp. 263-277. Instituto Nacional de Ecologia, Mexico. Gallo-Reynoso, J. P., and A. L. Figueroa-Carranza. 1996. The breeding colony of Laysan Albatrosses on Isla de Guadalupe, Mexico. W. Birds 27:70-76. Gaylord, H. A. 1897. Notes from Guadalupe Island. Nidologist 4:41-43. Hanna, G. D. 1925. Expedition to Guadalupe Island, Mexico, in 1922. General report. Proc. Calif. Acad. Sci. (Ser. 4) 14:217-275. 282 BIRDS OF PREY AND THE BAND-TAILED PIGEON ON ISLA GUADALUPE Howell, T. R., and T. J. Cade. 1954. The birds of Guadalupe Island in 1953. Condor 56:283-294. Jehl, J. R., Jr., and Everett, W. T. 1985. History and status of the avifauna of Isla Guadalupe, Mexico. Trans. San Diego Soc. Nat. Hist. 20:313-336. Luna-Mendoza, L. M., Barton, D. C., Lindquist, K. E., and Henry, R. W. III. 2005. Historia de la avifauna anidante de Isla Guadalupe y las oportunidades actuales de conservacion, in Isla Guadalupe, Restauracion y Conservacion (K. Santos del Prado and E. Peters, compilers), pp. 115-133. Instituto Nacional de Ecologia, Mexico. Mellink, E., and Palacios, E. 1990. Observations on Isla Guadalupe in November 1989. W. Birds 21:177-180. Montanez, L., Mellink, E., and Gallo-Reynoso, J. P. 2000. Isla Guadalupe, in Areas de Importancia para la Conservacion de las Aves en Mexico (M. C. Arizmendi and L. Marquez-Valdelamar, eds.), pp. 365-366. Sociedad para el Estudio y Conservacion de las Aves en Mexico (CIPAMEX), Mexico. Pitman, R. L., Walker, W. A., Everett, W. T., and Gallo-Reynoso, J. P. 2004. Popula- tion status, foods and foraging of Laysan Albatrosses Phoebastria immutabilis nesting on Guadalupe Island, Mexico. Marine Ornithology 32:159-165. Quintana-Barrios, L., Ruiz-Campos, G., Unitt, P., and Erickson, R. A. 2006. Update on the birds of Isla Guadalupe, Baja California. W. Birds 37:23-36. Rico-Cerda. J. 1983. Mapa de vegetacion de Isla Guadalupe. Chapingo 40:46-54. Sweet, P. R., Barrowclough, G. E., Klicka, J. T., Montanez-Godoy, L., and Escalante- Pliego, P. 2001. Recolonization of the flicker and other notes from Isla Guada- lupe, Mexico. W. Birds 32:71-80. Accepted 24 September 2009 Burrowing Owl Sketch by George C. West 283 FOOD HABITS OF WILD TURKEYS IN NATIONAL FORESTS OF NORTHERN CALIFORNIA AND CENTRAL OREGON GRETA M. WENGERT, MGW Biological, 102 Larson Heights Road, McKinleyville, California 95519; greta@mgwbio.com MOURAD W. GABRIEL, MGW Biological, 102 Larson Heights Road, McKinleyville, California 95519 RYAN L. MATHIS, National Wild Turkey Federation, Eureka, California 95501 THOMAS HUGHES, National Wild Turkey Federation, Edgefield, South Carolina 29824 ABSTRACT: We studied the diet of the Wild Turkey (Meleagris gallopauo) in five national forests in northern California and two national forests in central Oregon by collecting turkey droppings and analyzing them for specific food items. In all national forests the diet included insects; in all but one it included grasses. We analyzed the diet by sex and season and found that it varied seasonally and that females from California consumed more insects than did males. Seeds made up a small percentage of the diet in most national forests but constituted a majority of the diet in the Tahoe National Forest in California. The Wild Turkey ( Meleagris gallopauo) ranged historically from south- eastern Canada and the eastern and southwestern United States to northern Mexico (Kennamer et al. 1992). It is not native to California or Oregon, although a similar fossil species (M. californica) occurred prehistorically in California until about 11,500 years before present (Bochenski and Camp- bell 2006). Introductions by various agencies of three subspecies, the Rio Grande (M. g. intermedia), Eastern (M. g. siluestris), and Merriam’s (M. g. merriami), to west coast states as early as 1877 and into the 20 th century have proven successful, and the turkey’s range and numbers in these states are increasing (Wunz 1992, Keegan and Crawford 1999, Delgado 2004, California Department of Fish and Game 2004). The Wild Turkey is a generalist omnivore whose diet consists largely of plant material and insects (Hurst 1992). The literature on the food habits of the Wild Turkey in California and Oregon is limited, addressing mostly macro- and microhabitat selection for foraging (Delgado 2004). The single published study addressing the diet of the Wild Turkey in California, focus- ing on San Luis Obispo County on the central coast found that turkeys for- aged primarily on grasses, forbs, and hardwood mast (Smith and Browning 1967). In southern California, turkeys select as macrohabitat grasslands, mixed coniferous, and hardwood forest; as microhabitat for feeding they use primarily meadows (Delgado 2004). Because Wild Turkeys are not native to California and Oregon, the ques- tion whether they consume animals and plants of conservation concern has been raised (California Department of Fish and Game 2004, Oregon Department of Fish and Wildlife 2004). To date, there is no documentation of Wild Turkeys’ potential effects on native species in California and Oregon, whether through direct consumption or indirect modification of habitat. In this study, we provide the first data on the diet of Wild Turkeys in previously 284 Western Birds 40:284-291, 2009 FOOD HABITS OF WILD TURKEYS IN CALIFORNIA NATIONAL FORESTS unstudied areas of northern California and central Oregon. We describe the sexual and seasonal differences of the Wild Turkey’s diet in selected national forests within this region. METHODS From 2003 through 2005, samples of turkey feces were collected op- portunistically through the volunteer efforts of United States Forest Service personnel and independent researchers. Areas sampled in California were the El Dorado, Modoc, Plumas, Tahoe, and Mendocino national forests; those in Oregon were the Deschutes and Ochoco national forests (Figure 1). Samples were collected only when turkeys were observed in the area, ensuring that only fresh samples were collected and seasonal differences in diet composition could be estimated accurately. As many samples as possible were gathered at each site and stored in paper bags until analysis. When possible, sex of the turkey that deposited the sample was determined by direct observation. Typically, however, the sex of the bird that left the sample could not be identified, so we determined the sex from the shape and configuration of the sample itself, as feces of the sexes differ (Bailey 1956). In most cases, age class of the turkey could not be determined from the sample. Samples were collected during March, April, May, July, August, September, and October. The samples were analyzed microhistologically for diet composition at the Wildlife Habitat Laboratory at Washington State University, under director Bruce Davitt, by the methods of Sparks and Malachek (1967). For quan- titative analyses we examined both insect and vegetation matter similarly, though for lack of resources we were unable to classify insects more precisely than to the broad category “insects.” We categorized the samples by sex, season, and site and calculated the percentage cover of each type of food item on microscope slides made with a subsample of the pooled sample of each category. For the purposes of this study, we defined spring samples as those collected in March, April, or May, summer samples as those collected in July, August, or September, and fall samples as those collected in October. Using NCSS (Kaysville, UT), we compared the percentages of various food items in the diet of males and females with paired t tests. RESULTS At least 435 individual droppings were analyzed for this study. Because we were uncertain how many individual droppings some collections included, we based quantitative analyses only on collections in which the total number of droppings was known. Collections of droppings in which the total number was unknown were used for descriptive analyses only. In all the national forests the turkey’s diet consisted of various combinations of the following general categories: grasses, sedges/rushes, forbs, coniferous trees, shrubs, mosses, seeds, roots, ferns, and insects. Within these groups, we distinguished only grasses, rushes/sedges, forbs, coniferous trees, and shrubs further, in most cases to genus (Appendix A). Some of the conifer and shrub parts were identified to species. A group consisting of agricultural 285 FOOD HABITS OF WILD TURKEYS IN CALIFORNIA NATIONAL FORESTS Figure 1. National forests in California and Oregon where feces of the Wild Turkey were sampled for this study of food habits, 2003-2005. grains (barley and oats) were distinguished from grasses and forbs since the turkeys presumably obtained these foods from agricultural stocks rather than from grasses and forbs growing in the wild. For the plant foods listed (Appendix A), most of the parts found in the droppings were from seeds and flowers rather than from stems, roots, or leaves. Grasses were present in fecal samples from all national forests except Tahoe and ranged from 1.3% of the diet in spring droppings from males 286 FOOD HABITS OF WILD TURKEYS IN CALIFORNIA NATIONAL FORESTS in the Modoc National Forest (NF) to 66.8% in summer droppings from males in the Plumas NF. Insects were present in samples from all national forests and ranged from 0.6% in spring droppings from the Modoc NF to 39.9% in summer droppings from the Ochoco NF. Agricultural grains were only found in droppings from the Plumas NF (a maximum of 36.3% of food items in fall droppings from females), Deschutes NF (maximum 32.7% in spring droppings from males), and Mendocino NF (maximum 0.6% in fall droppings from females). Conifer seeds and “berries” constituted a majority of food items found in droppings from spring males in the El Dorado NF (61.9%) and from spring females in the Deschutes NF (79.9%). In most areas rushes/sedges made up a small percentage of food items, with a maximum of 9.0% in fall droppings from females in the Modoc NF. Forbs were found in samples from all areas and consistently made up a large proportion of food items found in droppings, from 3.3% in summer males in the Plumas NF to 95.8% in spring males in the Modoc NF. Mosses were present in samples from a few national forests but always in low proportions (maximum 2.7% in summer droppings from males in the Modoc NF). Ferns were present in samples from only Modoc NF and Plumas NF, always in small proportions (maximum 1.4% in fall samples from females in the Plumas NF). Roots were uncommon in fecal samples and were not found in those from the El Dorado NF, Tahoe NF, or Deschutes NF. The maximum percentage of roots was 4.4% in summer samples from the Modoc NF. In most areas seeds of unknown origin made up a small percentage of the diet (<5.6%), but in the Tahoe NF they constituted 88.9% of food items found in the samples. Finally, in many of the forests, material from shrubs made up a moderate percentage of food items in the samples, from 0.1% in spring females from the Ochoco NF to 29.4% in fall females from the Modoc NF. The turkeys’ diet varied by season and site (Figure 2). The only national forest from which samples were collected in all three seasons (spring, summer, fall) was the Modoc NF. In this area, the percentage of forbs decreased from 71.5% to 47.3% to 45.0% from spring to summer to fall. The percentage of grasses fluctuated from 24.9% to 33.0% to 5.3% in spring, summer, and fall, respectively. In fall, shrubs were important in the diet at 29.4% but con- tributed only 1.7% and 2.0% in the spring and summer, respectively. Sedges/ rushes were less common in the spring diet (0.1% of droppings) than in the summer (8.6%) and fall (9.0%). At all seasons conifers, mosses, seeds, roots, ferns, and insects each constituted a small percentage of the diet; of these the maximum percentage was for insects at 6.3% in the summer. The only significant difference between the diets of the sexes was in the percentage of insects in California. In all California samples combined, by season, the mean percentage of fecal samples containing insects was greater for female turkeys (7.3 ± 2.9%) (mean ± standard error) than for males (4.2 ± 2.2%, t6 = 3.3, P = 0.02). But this relationship did not hold when samples from both California and Oregon were pooled (ti = 1.7, P > 0.05). For all other food items, the sexes’ percentage composition of the fecal samples was not statistically different (P > 0.05). Because of the large variation among the sites in habitat and anthropo- genic influence, we did not analyze differences among the sites statistically; results would likely be biased by vastly different availabilities of food items. 287 FOOD HABITS OF WILD TURKEYS IN CALIFORNIA NATIONAL FORESTS a) c CD O <5 CL 100.0 80.0 60.0 40.0 20.0 0.0 ■ El Dorado H Plumas □ Deschutes □ Modoc i L b) O' / <9 & & c ^ & & Food Item Category C g 0) CL 100.0 80.0 60.0 40.0 20.0 0.0 non H Plumas □ Modoc m Ochoco eO & rv O ■fr & v ■ <9 & £ & 93% of pellets. Bird remains were found in 15.5% and 11.1% of pellets in summer and winter, respectively. Remains of giant water bugs (family Belostomatidae) were present in 7.7% of summer pellets but absent in winter. Although the diet was dominated by the same five categories of prey (four mammal genera and birds) at both seasons, the proportions of Microtus and Peromyscus declined during the winter, while those of Reithrodontomys and Dipodomys increased. Regurgitated pellets have commonly been used to examine the compo- sition of and both spatial and temporal variation in the diet of owls. The Barn Owl ( Tyto alba) has been the subject of many diet studies owing to its broad distribution in North America, concern over its declining numbers in some regions, and its tendency to nest in artificial sites (e.g., buildings or nest boxes) that allow easy collection of pellets (Marti 1992). Most studies have found that small mammals, particularly rodents and shrews, dominate the Barn Owl’s diet, although the primary prey species vary substantially both by location (reviewed in Marti 1992) and by season or year at a single location (Otteni et al. 1972, Smith et al. 1972, Marti 1973, 1988, Franzreb and Laudenslayer 1982, Gubanyi et al. 1992). We studied the diet of the Barn Owl at the Stillwater National Wildlife Refuge in northwestern Nevada during both summer and winter. Although the Barn Owl’s diet is well studied in some regions of North America, little information on it is available for the Great Basin of Nevada, a region where populations appear to be relatively stable (Floyd et al. 2007). The only two published accounts of the owl’s diet in Nevada were based on 89 (Alcorn 1942) and 14 ( Bogiatto et al. 2006) pellets. Our objective was to describe the composition of this Barn Owl population’s prey and to evaluate if the diet varied seasonally. METHODS We collected pellets from six sites in or near Stillwater National Wildlife Refuge within Churchill County, Nevada. The sites were centered primarily on nest boxes that the owls used for breeding in summer and roosting in winter, except for one in the attic of a deserted building. Prior to our study, 292 Western Birds 40:292-296, 2009 SEASONAL VARIATION IN THE DIET OF THE BARN OWL pellets had been regularly removed from these sites, so that the “summer” pellets we collected were deposited between May and September 2007, the “winter” pellets between October 2007 and February 2008. The surround- ing landscape consisted of a mixture of agricultural fields of alfalfa ( Medicago sativa) and other hay crops and upland desert scrub dominated by black greasewood ( Sarcobatus uermiculatus) and four-wing saltbush ( Atriplex canescens). Irrigation ditches with cattails (Typha spp.) were present at all sites, often with associated stands of Fremont cottonwood ( Populus fre- montii). One site was near several large lakes edged by marshes of cattails and tule ( Scirpus spp.). Before dissecting them, we dried the pellets for ~1 hr at 150° C to kill potential pathogens and then weighed them. We dissected the pellets by standard methods, identifying skull remains to the genus or species level by means of skull keys and a reference collection of local species. For each spe- cies, we used the maximum count of crania, left dentaries, or right dentaries to determine the number of individuals in each pellet. Finally, we used G tests of independence to determine if the proportional importance of each of the five most abundant categories of prey differed by season. RESULTS We identified 387 individual items of vertebrate prey in 143 summer pellets (total dry mass of all pellets, 1.004 kg) and 409 prey items in 163 winter pellets (total dry mass, 1 .035 kg). We found mammal remains in 135 summer pellets (94%) and 152 winter pellets (93%) and identified seven genera of mammals in summer pellets and eight genera in winter pellets (Table 1). We found bird remains in 15.5% and 1 1 . 1% of pellets in summer and winter, re- spectively. We were unable to identify all bird remains to family or species, but several skulls were of icterids, likely the Red-winged ( Agelaius phoeniceus ) or Yellow-headed (Xanthocephalus xanthocephalus) blackbirds, or sparrows (family Emberizidae), and one pellet contained the legs of a small rail, probably a Sora ( Porzana Carolina). Although we found some insect parts in associa- tion with bird remains, particularly within gizzards, we identified the remains of giant water bugs (family Belostomatidae, likely Lethocerus americanus) independent of bird remains in 7.7% of summer pellets, suggesting that the owls occasionally targeted these large insects as prey. During both summer and winter, the diet of this owl population was dominated by the same 5 categories of prey; voles ( Microtus spp.), white- footed mice ( Peromyscus spp.), harvest mice ( Reithrodontomys spp.), kangaroo rats ( Dipodomys spp.), and birds. The importance of the various taxa in the diet, however, varied by season (Table 1). Voles were the most common prey in both seasons, but their importance was greater in the summer. The proportion of pellets that contained Peromyscus mice did not differ by season, but these mice made up a greater proportion of the total vertebrate prey items consumed during summer than winter. Harvest mice showed the greatest seasonal shift and were significantly more important in the owls’ diet during winter. Kangaroo rats showed a similar pattern of increased importance during winter. The importance of birds in the diet did not differ by season. 293 SEASONAL VARIATION IN THE DIET OF THE BARN OWL Table 1 Prey contents of Barn Owl pellets collected during summer (May- Sept 2007; n = 143) and winter (Jan-Feb 2008; n = 163) at Stillwater National Wildlife Refuge, Churchill County, Nevada Prey and season Total individuals 0 Individuals per pellet fa Proportion of pellets 0 Proportion of total vertebrate prey items d Voles (Microtus spp.) Summer 144 1.02(1.06) 0.620 P= 0.09 0.372 P= 0.06 Winter 126 0.77 (0.94) 0.528 0.308 White-footed mice (Peromyscus spp.) Summer 140 0.98(1.59) 0.415 P= 0.64 0.362 P= 0.001 Winter 105 0.64(1.00) 0.387 0.257 Harvest mice ( Reithrodontomys spp.) Summer 35 0.30 (0.62) 0.176 P= 0.003 0.090 P< 0.001 Winter 90 0.55 (0.99) 0.319 0.220 Kangaroo rats ( Dipodomys spp.) Summer 26 0.18(0.45) 0.155 P= 0.06 0.067 P= 0.006 Winter 51 0.31 (0.64) 0.239 0.124 Botta’s pocket gopher (Thomomys bottae) Summer 10 0.07 (0.26) 0.070 0.026 Winter 6 0.04 (0.19) 0.037 0.015 Mountain cottontail ( Sylvilagus nuttalli ) Summer 4 0.03 (0.17) 0.028 0.010 Winter 2 0.01 (0.11) 0.012 0.005 Woodrats (Neotoma spp.) Summer 0 0.00 0.000 — Winter 3 0.02 (0.13) 0.018 0.005 Muskrat ( Ondatra zibethicus) Summer 3 0.02 (0.14) 0.021 0.008 Winter 0 0.00 0.000 — Bats ( Myotis spp.) Summer 0 0.00 0.000 — Winter 1 0.01 (0.08) 0.006 0.002 Birds Summer 25 0.18(0.40) 0.155 P = 0.15 0.065 P = 0.84 Winter 25 0.15(0.54) 0.110 0.061 Invertebrates Summer 12 0.08 (0.27) 0.081 Winter 0 0.00 0.000 “Minimum total number of individuals identified in all pellets for each season and category of prey. fa Mean number of individuals per pellet on the basis of all pellets for the season (standard devia- tion). “Proportion of all pellets for the season that contained the category of prey. Probability values are from G tests of independence comparing the proportions in winter and summer of the five principal categories of prey. “Probability values are from G tests of independence comparing the proportions in winter and summer of the five principal categories of prey, on the basis of 387 and 409 total vertebrate prey items detected during summer and winter, respectively. 294 SEASONAL VARIATION IN THE DIET OF THE BARN OWL DISCUSSION As most published studies elsewhere have found (Marti 1992), small mam- mals dominated the diet of the Barn Owl population we studied. Although the dominant prey species taken by Barn Owls vary by location (Marti 1992), the four main mammalian prey in our study, voles, Peromyscus mice, harvest mice, and kangaroo rats, have all been documented as common components of Barn Owl diet in at least some western habitats (Jones 1949, Smith et al. 1972, Marti 1973, 1988, Gubanyi et al. 1992, Van Vuren et al. 1998, Lockwood and Jones 2000). Some seasonal variation in diet was apparent, but it mainly took the form of a complementary shift in use of these same four prey; use of voles and Peromyscus mice declined during winter, while use of harvest mice and kangaroo rats increased. Whether these shifts in diet relate to seasonal changes in abundance of these species, and whether these shifts are typical of most years, is unknown. At Stillwater, Barn Owls rarely took larger prey (adults >100 g). Pocket gophers ( Thomomys spp.) have been documented as a major constituent of the Barn Owl’s diet at some western sites, particularly during summer when small, dispersing young are available (e.g., Knight and Jackman 1984, Gubanyi et al. 1992, Van Vuren et al. 1998). The owls we studied may have taken gophers infrequently because they were rare in the region or, more likely, because other smaller prey items were more available and easily captured. Other studies have also documented cottontail rabbits ( Sylvilagus spp.) and muskrats ( Ondatra zibethicus) as rare but regular components of the Barn Owl’s diet (e.g., Marti 1973, Franzeb and Laudenslayer 1982, Van Vuren et al. 1998). The size of adults of these species may limit the owls’ using them as prey; on the basis of skull size, all cottontail and muskrat remains in our pellets were of juveniles. Birds were taken regularly, representing just over 6% of all vertebrate prey items in both seasons. Birds have often been documented in the Barn Owl’s diet, although most studies in the western United States have reported that they represent <5% of all prey items (Jones 1949, Marti 1973, 1988, Gubanyi et al. 1992, Van Vuren et al. 1998, Lockwood and Jones 2000). Otteni et al. (1972) suggested that Barn Owls may opportunistically increase their use of birds, particularly abundant colonially nesting or roosting birds, when the availability of rodents is low. Although we found no seasonal pat- tern in use of birds within our population, our study and another of the Barn Owl’s diet at a site nearby in Nevada (Alcorn 1942) found most bird prey items to consist of colonial marsh birds, especially icterids. Large invertebrates have occasionally been reported in the diet of some Barn Owl populations (Marti 1992). For example, in western North America, Jerusalem crickets (family Stenopelmatidae) are eaten at many locations, and in some areas are regular prey (Smith and Hopkins 1937, Maser et al. 1980, Knight and Jackman 1984). To our knowledge, our study is the first to document the giant water bug as Barn Owl prey. These insects are appar- ently large enough (50-60 mm; Bland and Jaques 1978) to be worth pursuit when encountered. However, as in most Barn Owl populations, invertebrates likely represented a tiny proportion of the total biomass of prey taken. 295 SEASONAL VARIATION IN THE DIET OF THE BARN OWL ACKNOWLEDGMENTS We thank Raymond Bogiatto, Juliana Bosi del Almeida, and Motti Charter for providing comments that improved the manuscript. Support for this research was provided by the Stillwater National Wildlife Refuge and by Penn State University- Hazleton. LITERATURE CITED Alcorn, J. R. 1942. Food of the Barn Owl at Soda Lake, Nevada. Condor 44:128- 129. Bland, R. G., and H. E. Jaques. 1978. How to Know the Insects, 3 rd ed. Wm. C. Brown, Dubuque, IA. Bogiatto, R. J., Broughton, J. M., Cannon, V. I., Dalton, K., and Arnold, S. 2006. Fish remains dominate Barn Owl pellets in northwestern Nevada. W. N. Am. Nat. 66:395-396. Floyd, T., Elphick, C. S., Chisholm, G., Mack, K., Elston, R., Ammon, E. M., and Boone, J. D. 2007. Atlas of the Breeding Birds of Nevada. Univ. of Nevada Press, Reno. Franzreb, K. E., and Laudenslayer, W. F. 1982. Composition and seasonal variation of the Barn Owl (Tyto alba ) diet in Arizona. Raptor Res. 16:36-39. Gubanyi, J. A., Case, R M., and Wingfield, G. 1992. Diet and nesting success of Barn Owls breeding in western Nebraska. Am. Midland Nat. 127:224-232. Jones, J. K., Jr. 1949. Notes on the small mammal content of pellets of the Barn Owl (Tyto alba pratincola) in Nebraska. Nebraska Bird Rev. 17:4-5. Knight, R. L., and Jackman, R. E. 1984. Food-niche relationships between Great Horned Owls and Common Barn-Owls in eastern Washington. Auk 101:175- 179. Lockwood, M. W., and Jones, C. 2000. Vertebrate remains found in Barn Owl pellets from Crosby County, Texas. Texas J. Sci. 52:169-173. Maser, C., Shaver, S., Shaver, C., and Price, B. 1980. A note on the food habits of the Barn Owl in Malheur County, Oregon. Murrelet 61:78-80. Marti, C. 1973. Ten years of Barn Owl prey data from a Coloradan nest site. Wilson Bull. 85:85-86. Marti, C. 1988. A long-term study of the food-niche dynamics in the Common Barn-Owl: Comparisons within and between populations. Can. J. Zool. 66:1803-1812. Marti, C. 1992. Barn Owl (Tyto alba), in The Birds of North America (A. Poole and F. Gill, eds.), no. 1. Acad. Nat. Sci., Philadelphia. Otteni, L., Bolen, E., and Cottam, C. 1972. Predator-prey relationships and repro- duction of the barn owl in southern Texas. Wilson Bull. 84:434-448. Smith, C. F. and Hopkins, C. L. 1937. Notes on the Barn Owls of the San Francisco Bay region. Condor 39:189-191. Smith, D. G, Wilson, C. R., and Frost, H. H. 1972. Seasonal food habits of Barn Owls in Utah. Great Basin Nat. 32:229-234. Van Vuren, D., Moore, T. G., and Ingels, C. A. 1998. Prey selection by Barn Owls using artificial nest boxes. Calif. Fish and Game 84:127-132. Accepted 28 July 2009 296 NOTES FIRST RECORD OF A MANGROVE YELLOW WARBLER IN ARIZONA NATHAN K. BANFIELD, 8 Rainbow Circle, Montgomery City, Missouri 63361; nathankbanfield@yahoo . com PATRICIA J. NEWELL 170 E Green St., Warnell School of Forestry and Natural Resources, University of Georgia, Athens, Georgia 30602-2152; pattijean . newell@gmail . com On 31 July 2004 at 1130 MST we captured a Mangrove Yellow Warbler (Dend- roica petechia erithachorides subspecies group), the first to be recorded in Arizona and the northernmost ever. The bird was captured in a mist net along Tonto Creek (33.75415°N, 111.22273° W), which flows into the north end of Roosevelt Lake , in Gila County about 90 km northeast of Phoenix. The bird was netted in early succes- sional riparian habitat consisting of Goodding’s Willow (Salix gooddingii) and tama- risk (Tamarix spp.) that is established along the lake’s inflows as a result of periodic fluctuations of the lake’s level. To document this significant record, we took 15 digital photographs that clearly show diagnostic features such as a chestnut face, throat, and forecrown, thin chestnut streaks on the breast, and yellow spots on the inner webs of the rectrices (Figures 1,2). These diagnostic photos led the Arizona Bird Committee to accept this record as the first of a Mangrove Warbler in Arizona (Rosenberg et al. 2007). Although the subspecies of the Mangrove Warbler we captured could not be ascertained, the subspecies rhizophorae breeds the closest to Arizona. It is resident in coastal mangroves north to the vicinity of Bahia Kino, Sonora, Mexico (Russell and Monson 1998), about 550 km south of Roosevelt Lake. Mangrove Warblers are currently recognized as a group of subspecies within the Yellow Warbler (Browning 1994, Dunn and Garrett 1997, AOU 1998, Lowther et al. 1999). The Yellow Warbler currently contains 43 subspecies divided among three groups, the Yellow Warbler in the narrow sense ( aestiua group), the Golden Warbler (petechia group), and the Mangrove Warbler (erithachorides group), which differ in plumage, breeding strategy, distribution, and habitat (Lowther et al. 1999). Genetic studies of Yellow Warblers suggest these may constitute more than one species, espe- cially a northern migratory species distinct from a southern species largely restricted to mangroves (Klein and Brown 1994). While all male Yellow Warblers share the generally yellow plumage, chestnut streaking on the breast, and yellow tail spots, the color of the adult male’s head distinguishes the three Yellow Warbler groups (Mennill 2000). In most populations of the Mangrove group, the male has a distinct chestnut hood (Dunn and Garrett 1997). The individual we captured had an overall yellow appearance with a distinct chestnut face and chestnut streaks on the upper breast (Figure 1). Chestnut on the face extended from the throat along the side of the face behind the eyes to the top of the head. The extent of chestnut on the head indicated it was a second-year male; after-second-year males have a complete chestnut hood. J. Salgado-Ortiz, who has studied Mangrove Yellow Warblers on the Yucatan Peninsula, Mexico, confirmed the bird’s age and sex. The bird was molting extensively, and its feathers were worn and in poor condition, giving it a ragged appearance. The bird was actively molting in the eye ring, face, head, throat, and breast (Figure 1) as well as the back, primaries, tertials, upper and underwing coverts, and rectrices (Figure 2). The six remaining unmolted rectrices were a brownish olive with distinct pale yellow spots on the inner webs. These spots Western Birds 40:297-300, 2009 297 NOTES Figure 1. Mangrove Yellow Warbler caught at Roosevelt Lake, Gila Co., Arizona on 31 July 2004, showing active molt and extensive chestnut on the head. Photo by Nathan K. Banfield were larger and more noticeable on the outer rectrices, but most of the inner rectrices were missing or in extremely poor condition. Bright new feathers were noticeable on the tertials, upper wing coverts, and inner primaries (Figure 2). Old wing feathers had slight pale yellow edges, whereas newer feathers had distinct brighter and wider yellow-olive edges. This contrast was especially noticeable on the new tertials, which had exceptionally wide and bright edges (Figure 2). Mangrove Warblers are generally considered to be nonmigratory and are found along the coasts of Mexico and Central America south to the northern coasts of South America and in the Galapagos Archipelago (AOU 1998). In most of their range, Man- grove Warblers are restricted almost exclusively to coastal mangroves, especially the red mangrove ( Rhizophora mangle) (Lowther et al. 1999). On the Yucatan peninsula, Salgado-Ortiz (pers. comm.) noted the use of habitats other than coastal mangroves only twice, both times within 5 km of mangroves. The riparian habitat in which we captured the Mangrove Warbler, consisting of dense willow and tamarisk groves, is structurally similar to the mangrove habitat this subspecies group occupies. This warbler’s appearance in Arizona corresponded to the schedule of the North American monsoon system, on the northern fringe of which Arizona is located. Dur- ing the monsoon, from the end of June through September (Ropelewski et al. 2005), winds originate primarily from the Gulf of California and the Gulf of Mexico (Adams and Comrie 1997, Ellis et al. 2004). Patten and McCaskie (2004) reported a pattern of northward dispersal of subtropical waterbirds into southern California during the monsoonal period from May to October, but no such correlation with vagrancy of landbirds is known. During the monsoonal season post-breeding dispersal has brought into Arizona a number of subtropical landbirds such as the Eared Quetzal ( Euptilotis neoxenus), Aztec Thrush [Ridgwayia pinicola), and Sinaloa Wren {Thryothorus si- 298 NOTES Figure 2. The same individual, showing worn and missing rectrices and active molt on the wings. Photo by Nathan K. Banfield naloae) (G. H. Rosenberg pers. comm.), but monsoonal winds have probably played little or no role in such occurrences. Prior records of Mangrove Warblers in the United States have all come from south- coastal Texas and are presumably of the Gulf of Mexico subspecies oraria (Lockwood and Freeman 2004). In the winters of 2003 to 2007, ten or more male and female Mangrove Warblers were documented in patches of the black mangrove ( Auicennia germinans ) on barrier islands off South Padre Island (S. Colley pers. comm.). 299 NOTES Since our discovery, the Mangrove Warbler has occurred twice in southern California, once at the mouth of the Alamo River south end of the Salton Sea (18 December 2007, O. Johnson; McCaskie and Garrett 2008) and once at San Diego (13 January-27 March 2009, M. J. Billings; McCaskie and Garrett in press); both identifications are supported by photographs. First authorship was determined by a coin toss. Funding for the work to which our observation was incidental was provided by the U.S. Bureau of Reclamation, U.S. Geological Survey, and the U.S. Forest Service. We thank Javier Salgado-Ortiz for verification of the bird’s age and sex and Scarlet Colley for information on Mangrove Warblers she observed in Texas. Eben H. Paxton, Mark K. Sogge, Talima Pearson, and Gary H. Rosenberg reviewed early drafts of the manuscript. LITERATURE CITED American Ornithologists’ Union. 1998. Check-list of North American Birds, 7 th ed. Am. Ornithol. Union, Washington, D.C. Adams, D. K., and Comrie, A. C. 1997. The North American monsoon. Bull. Am. Meteorol. Soc. 78:2197-2213. Browning, R. M. 1994. A taxonomic review of Dendroica petechia (Yellow Warbler) (Aves: Parulinae). Proc. Biol. Soc. Washington 107: 27-51. Dunn, J. L., and Garrett, K. L. 1997. A Field Guide to Warblers of North America. Houghton Mifflin, Boston. Ellis, A. W., Saffell, E. M., and Hawkins, T. W. 2004. A method for defining mon- soon onset and demise in the southwestern United States. Int. J. Climatol. 24:247-265. Klein, N. K., and Brown, W. M. 1994. Intraspecific molecular phylogeny in the Yel- low Warbler ( Dendroica petechia) and implications for avian biogeography in the West Indies. Evolution 48:1914-1932. Lockwood, M. W., and Freeman, B. 2004. Texas Ornithological Society Handbook of Texas Birds. Texas A&M Univ. Press, College Station, TX. Lowther, P. E., Celada, C., Klein, N. K., Rimmer, C. C., and Spector, D. A. 1999. Yellow Warbler ( Dendroica petechia), in The Birds of North America (A. Poole and F. Gill, eds.), no. 454. Birds N. Am., Philadelphia. McCaskie, G., and Garrett, K. L. 2008. Southern California. N. Am. Birds 62:302- 306. McCaskie, G., and Garrett, K. L. 2009. Southern California. N. Am. Birds 63: in press. Mennill, D. J. 2000. Song characteristics and singing behavior of the Mangrove Warbler ( Dendroica petechia bryanti). J. Field Ornithol. 72:327-337. Patten, M. A., and McCaskie, G. 2004. Patterns and processes of the occurrence of pelagic and subtropical waterbirds at the Salton Sea. Studies in Avian Biology 27:33-41. Ropelewski, C. F. D., Gutzler, S., Higgins, R. W., and Mechoso, C. R. 2005. The North American monsoon system. Proceedings of the Third International Work- shop on Monsoons (IWM-III), Hangzhou, China, 2-6 Nov. 2004. Rosenberg, G. H., Radamaker, K., and Stevenson, M. M. 2007. Arizona Bird Com- mittee report, 2000-2004 records. W. Birds: 38:74-101. Russell, S. M., and Monson, G. 1998. The Birds of Sonora. Univ. Ariz. Press, Tucson. 300 Accepted 10 June 2009 NOTES PREY REMAINS IN NESTS OF FOUR CORNERS GOLDEN EAGLES, 1998-2008 DALE W. STAHLECKER, Eagle Environmental, Inc., 30 Fonda Road, Santa Fe, New Mexico 87508 DAVID G. MIKESIC, Navajo Natural Heritage Program, P. O. Box 1480, Window Rock, Arizona 86515; dmikesic@hotmail.com JAMES N. WHITE, Jicarilla Game and Fish Department, P. O. Box 313, Dulce, New Mexico 87528 (current address: Colorado Division of Wildlife, 151 E. 16 th Street, Durango, Colorado 81301) SPIN SHAFFER, P. O. Box 4084, Truckee, California 96160 JOHN P. DeLONG, Eagle Environmental, Inc., 2314 Hollywood Ave. NW, Albuquer- que, New Mexico 87103 (current address: Department of Ecology and Evolutionary Biology, Yale University, New Haven, Connecticut 06520) MARK R. BLAKEMORE, Jicarilla Apache Utility Authority, P. O. Box 916, Dulce, New Mexico 87528 CRAIG E. BLAKEMORE, P. O. Box 1048, Lake City, Colorado 81235 The Golden Eagle ( Aquila chrysaetos ) is among the most studied of raptors (Watson 1997, Kochert et al. 2002). The easiest way to monitor its diet during the breeding season has been to sample remains in active nests (Collopy 1983). In western North America, Golden Eagle nests contain remains primarily of leporids (rabbits and hares) and sciurids (ground squirrels) but also of many other species (Olendorff 1976, Palmer 1988, Kochert et al. 2002). Published information for the southwestern U. S. is limited to prey collected from 41 nests in southeastern New Mexico and western Texas in the 1960s (Mollhagen et al. 1972), nine nests in central Arizona in 1985 (Eakle and Grubb 1986), and time-lapse photography and prey collected at four nests over two years in the trans-Pecos region of Texas (Lockhart 1976). Mammals, mostly hares, rabbits, and sciurids, dominate these samples. From 1998 to 2008 we rappelled into 182 active Golden Eagle nests in 90 ter- ritories in northeastern Arizona, southeastern Utah, and northwestern New Mexico, visiting each territory one to five times for a total of 191 nest checks. The primary reason for these May and June visits was to band nestlings at an age of 4 to 6 weeks. The nests studied from 1998 to 2001 were distributed throughout the 65,000-km 2 Navajo Nation; those studied from 2003 to 2008 were in the much smaller 3550-km 2 Jicarilla Nation. All nests were on rock substrates (cliffs and buttes) within a variety of habitats ranging from nearly barren deserts at 1250 m elevation near the Colorado River in Arizona to rugged mesas and 2500-m mountains of the Navajo Nation, woodlands and forests near the continental divide, and the 2500-m mountains of the Jicarilla Nation in New Mexico. Great Basin desert scrub and desert grasslands dominate the lower desert habitats, while ponderosa pine (Pinus ponderoso ) forests cover the mountainous areas; woodlands of pinyon pine ( P. edulis) and juniper {Juniperus spp.) and flats with sagebrush ( Artemisia tridentata, A. frigida) occupy the middle elevations (Brown 1982). While habitats varied, all nests were within the Colorado Plateau ecoregion. We did not census populations of prey, but lagomorphs were generally widespread and sometimes abundant, and ground squirrels and Gun- nison’s prairie dogs ( Cynomys gunnisoni ) were locally common. Large avian prey was lacking throughout the region, except for waterfowl at mountain lakes and Wild Turkeys ( Meleagris gallopavo) in montane forests and openings. During visits to nests we documented 660 prey items of 24 species (Table 1). We enumerated all species as the minimum number based on full bodies or counting of Western Birds 40:301-306, 2009 301 NOTES o o • *— < X 0) (0 1 t2 a o o H 00 CO CM CM CM CM 00 Tt m CO CO CD Z TD C (0 (0 C O N • *— < U-. < _C (0 T— 1 D c o • »— < co CD c£ CD Sh CD c S-H O o S-H 3 O Uh to - 4— 1 CO CD 2 CO (0 UJ c CD IS o O o -t-l co 00 o o CO o o CO VO o o CO LO o o CO o o CO CO o o CO o o CO o o o CO Ov cv cv 00 CTv cv LO CO CO OM> CO VO 0^ [>■ CO CO 'Cf 'Cf 'Cf CO LO 00 'cf CO co ^ vo VO CO CO T-H CO CO CO vD ’ — i CO [>■ CO © © © © © © CM © CO co r> m CO CO 00 to • »— I > T— I Oc T3 c 3 o Uh CD c • »— < 3 CD cc CD S-H Oh 00 o o 1-1 CM ,© Oc ON H T-H (0 a CD s— Oh (0 CD 3 o E S-. .O a o CD 3 a CD l ~j a . CD is s ■a a 2 2 -X •=• CD -2 hS a ~ co -a , _cd ~ '5 -2 v c -x o r ) -*—* JS o CQ o .g c o a cr O oo . ~"0 ' o’ C o 3 a o .2 o ' il-. 0 h +3 C y> to CD 3 -+H a CD .2 'C a 3 CD J3 IE U o a 2 o 03 CD 3 C o CD a, oo c E o v cd c 'c CD c 2 3 O o o hC CD a £ O -g £ c r a-x co g a CD CD CD CD r- a 3 5 Ur cu g 3 ~ ■SS 3 O -C CD CD ■S CD O £ a~a uj "5 O „ r > uh- 2 „ CD .2 X o J3 co 3 Sh CM LO LO CO o o CM LO CO CM O O O O O CM lo CM LO O CM CO CM CM o o "a 0) -t-1 c c o u a 1 » CO S-. CL (0 • X co > -a o co Q "co 3*C V O TO o a> O CQ DC co X r- a.S _2 -c O c L-C £ a '■=, o <-> x a E & o o o O 9| 15 o tJL 0* £- £ § -e =o O c 2 Cl C3 v O 3 3 3 £ o zO CO „ 3S C c> a> CO > (0 - o CC 3 O JS o a .c/s 53 -5 g 8 - ~ § § •ti s ^ a u™ jx +j- » JS 3 « 3 £ 2 0 ^ 8 ^ L0 *c » C > ^ 'q, 5 -4-/ il g- o tf) .3 o £ a> DQ O DQ 2 > .13