Raptor Research A Quarterly Publication of The Raptor Research Foundation, Inc. Volume 20, Number 3/4, Fall/ Winter 1986 fLSN GGGG-OGGG) Contents Reproductive Biology of Northern Hawk-Owls In Denali National Park, Alaska, Kenneth Kcrtcl! 91 Roost Tree Characteristics and Abundance of Wintering Vultures at a Communal Roost in South Central. PENNSYLVANIA, Anthony L. Wright. Hit-hard II. Vahner and Ccralri R. Stnrm . , ,,,,,,, 302 The Barn Owl Ego: Weight Loss Characters. Fresh Weight Prediction AND INCUBATION Period. Janies D. Marshall, Chine II. Hager and Gw)'n McKee I OH Prey and Tropic Ecology of Great Horned Owls in Western South A m e rig a : An In dication of Latttudi n a l T ren ds . Fabian M. Jaksic, Jusc L. Van:/ and Jaime R. Ran 113 Impact of a High- Voltage Transmission Line on a Nesting Pair of Southern Bald Eagles in Southeast Louisiana. David A. Dell and Phillip J Zwank ... 11? Food of the Booted Eagle (Hierafietus pennafrts) in Central Spain. JomP. Veiga 120 Food of Nesting Bald Eagles in Louisiana. Joseph A, Dm^uii. Philip J. 7 . wank and Gary C. Furman . , , 12^ Male Food Provisioning and Female Reproduction in American Kestrels, Timothy J. Cboup 128 Short Communications Sueecw EUtcof (he Peregrine Filccm {Fake pfirtgrinus) Hisnring Dunlin {Cotidm atpk wU During Winter. Jitscpli D. Buchanan, Steven 6. I If imen and Tnd M. Jiihniinn 130 Gulden F.a^le Capture rjf an American Cntit. Daniel J. Severson . 1.3 1 ^Bilatei Ld Bmnhlelonl in a Wild Red-tailed Hawk. Kevin L. EUlis- 1 32 Dissertation Abstracts ***«.***♦*********..•., 136 News and Reviews 119,133-136 Dedication r T 136 The Raptor Research Foundation, Inc. Provo, Utah THE RAPTOR RESEARCH FOUNDATION, INC. (Founded 1966) OFFICERS PRESIDENT: Jeffrey L. Lincer, Office of the Scientific Advisor, 4718 Dunn Drive, Sarasota, Florida 33583 VICE-PRESIDENT : Richard Clark, York College of Pennsylvania, Country Club Road, York, Pennsylvania 1 7403- 3426 SECRETARY: James D Fraser, Virginia Polytechnic Institute and State University, Cheatham Hall, Blacksburg, Virginia 24061 TREASURER: Jim Fitzpatrick, Carpenter Nature Center, 12805 St. Croix Trail, Hastings, Minnesota 55033 BOARD OF DIRECTORS EASTERN DIRECTOR: James A. Mosher, Savage River Consulting, P.O. Box 71, Frostburg, Maryland 21532 CENTRAL DIRECTOR: Patrick T. Redig, Department of Veterinary Biology, 295 Animal Science/ Veterinary Medicine Building, University of Minnesota, St. Paul, Minnesota 55108 MOUNTAIN & PACIFIC DIRECTOR: Al Harmata, Department of Biology, Montana State University, Bozeman, Montana 59717 EAST CANADA DIRECTOR: David M. Bird, Macdonald Raptor Research Centre, Macdonald Campus of McGill University 21,111 Lakeshore Road, Ste. Anne de Bellevue, Quebec H9X ICO WEST CANADA DIRECTOR: Lynn Oliphant, Department of Veterinary Anatomy, University of Saskatchewan, Saskatoon, Saskatchewan S7M 0W0 INTERNATIONAL DIRECTOR: Martin Bottcher, Postfach 2164, Blankenheimner Strass 3, 5372 SCHLEIDEN, Federal Republic of Germany, GERMANY DIRECTOR AT LARGE # 1 : Michael Collopy, University of Florida, School of Forest Resources and Conservation, 118 Newins-Ziegler Hall, Gainesville, Florida 32601 DIRECTOR AT LARGE #2: Gary Duke, Department of Veterinary Biology, 295/K Veterinary Medicine Building, University of Minnesota, St. Paul, Minnesota 55108 DIRECTOR AT LARGE #3: Richard P. Howard, U.S. Fish and Wildlife Service, 4696 Overland Road, Room 566, Boise, Idaho 83705 ** ** sfofc * ** ****** * * ** * Persons interested in predatory birds are invited to join The Raptor Research Foundation, m Inc. Dues are $15 per year in the U.S., $17 per year outside the U.S., $13 per year for U.S. students, and $15 per year for students outside the U.S. Add $2 to dues if membership is received after 15 February. The Foundation’s journal Raptor Research is distributed quarterly to all current members. Subscription price to institutions and nonmembers is the same as regular membership. Single copies and back issues are available from the Treasurer. A Contributing Membership if $25, a Sustaining Membership is $100, and a Life Membership is $500. All contributions to The Raptor Research Foundation, Inc., are tax-deductible. Send requests for information concerning membership, subscriptions, special publications, or change of address to the Treasurer. Other communications may be routed through the appropriate Officer or Board member. All inquiries concerning the journal should be addressed to Clayton M. White, Editor, Raptor Research, Department of Zoology, 161 WIDB, Brigham Young University, Provo, Utah 84602, U.S. A 3jc Published quarterly by The Raptor Research Foundation, Inc. Business Office: Jim Fitzpatrick, Carpenter Nature Center, 12805 St. Croix Trail, Hastings, Minnesota 55033, U.S.A. RAPTOR RESEARCH A QUARTERLY PUBLICATION OF THE RAPTOR RESEARCH FOUNDATION, INC. Vol. 20 Fall/Winter 1986 No. 3/4 REPRODUCTIVE BIOLOGY OF NORTHERN HAWK-OWLS IN DENALI NATIONAL PARK, ALASKA Kenneth Kertell Abstract — Two nesting pairs of the Northern Hawk-Owk (Sumia ulula) were studied in 1980 in Denali National Park, Alaska. Observations began during the incubation phase and ended when the young left the nest and could no longer be found. During this period information was gathered on food habits and breeding biology. Owls did not return to breed in the study area until 1984 when a pair layed eggs atanestused in 1980. Failure to breed, at least in 1981, was apparently the result of a substantial decrease in the microtine population. Surprising little is known about the status and biology of the Northern Hawk-Owl (Sumia ulula), particularly in North America. Walker (1974) claimed that hawk-owls have been reduced consid- erably in North America but offered no explana- tion to account for the reduction. Fyfe (1976) de- scribed it as rare to low in abundance in eastern Canada and low to moderate in abundance in cen- tral and western Canada. In Europe, Mikkola (1972) believed that hawk-owls had suffered a gen- eral population reduction in Finland, Norway, and Sweden, based on small recent invasions. Adequate raptor data are hard to obtain because of the gener- ally low densities of raptors and their habit of nest- ing in remote and inaccessible places (Newton 1976). The fact that owls are secretive and noctur- nal further compounds the problems of obtaining adequate data. Bent (1938) summarized most early information available on the hawk-owl in North America, and Gabrielson and Lincoln (1959) summarized infor- mation on their breeding biology in Alaska. More recently Smith (1970) published information on various aspects of the reproductive habits of hawk- owls near Ottawa, Canada. Information on the hawk-owl in Europe is more extensive (Mikkola 1983). Although no studies have provided detailed de- scriptions of hawk-owl breeding behavior, similarities in appearance and behavior between hawk-owls and the diurnal falconiforms are appa- rent. According to Sparks and Soper (1970), the hawk-owl is an ecological vicariate of a diurnal fal- con or accipiter, and behaves like a falconid even though it is primarily a predator of small mammals. Harrison (1973) speculated that the hawk-owl may be filling a vacant diurnal niche. Here I describe aspects of the breeding biology and behavior of Northern Hawk-Owls nesting in Alaska. Study Area and Methods Two hawk-owl nests were studied; both were on the north slope of the Hines Creek drainage at about 670 m elevation in Denali National Park, Alaska. The 2 nests were west of park headquarters (R7W,T14S,S7 and R7W,T14S,S12) and 1.8 km apart. Both were within 100 m of the park road. The nests were located in open needleleaf forest (Vie- reck et al. 1982) dominated by white spruce ( Picea glauca). Aspen (Populus tremuloides) and balsam poplar (Populus balsamifera) occurred uncommonly. Ground cover con- sisted largely of willow (Salix spp.) in wet areas and dwarf birch ( Betula nana and B. glandulosa ) in dry areas. Lab- rador tea (Ledum palustre) , blueberry ( Vaccinium vitisidaea), and crowberry (Empetrum nigrum ) also occurred. Sphag- num was thick in places. Annual rainfall at Denali Park headquarters averages about 37.5 cm, with summer rains and occasional summer snow accounting for most of the total. Daylength varies from 12 hrs in late March and September to 22 hrs. in late June. Field Observations - Hawk-owls were observed for 137 hrs between 12 May and 5 July 1980. A 20-45x zoom lens spotting scope and 9x binoculars were used to observe all activities. Owls appeared to habituate to observer pre- sence allowing observations to be made from a distance of 91 Raptor Research Vol. 20 (3/4): 9 1-101 92 Kenneth Kertell Vol. 20, No. 3/4 less than 60 m from the nest. To reach nests, trees were climbed directly or with the aid of an aluminum ladder. Additional observations of hawk-owls were made in 1977 and 1984. In 1977, a nest from which 5 young fledged was visited twice between 1 1 June and 2 July, and a different group of 5 fledglings was located on 27 June. In 1984, a pair of adults was observed between 24 March and 7 April near a nest used in 1980. Data from 1976 were from the park observation files. Clutch sizes in 1980 were measured by climbing to the nest, while brood sizes were determined from the fact that all known eggs hatched and young fledged. In 1976 and 1977 the number of young fledged served as an index of both the minimum clutch sizes and brood sizes. I noted plumage differences that enabled me to recog- nize sexes of adults at both nests following observations of copulation, prey exchange, and egg-laying. Males had grayish-brown or blackish-brown barring while in females barring was a lighter chestnut-brown. In the male, the border between the upper breast and foreneck was de- marcated by a contrasting blackish band, while the transi- tion in the female was less distinct. The differences were more apparent in one pair than in the other and, accord- ing to Mikkola (1983), these kinds of differences can be attributable to age. Food Habits. Information on food habits was obtained from the analysis of 387 pellets, by direct observation of prey brought to young, and from discarded prey remains. Analysis of pellets provided over 95% of cricetid, 100% of soricid, and about 10% of sciurid and avian prey data. All remaining data was obtained by observation of prey deliv- ery and the location of discarded remains. Pellets were collected beginning on 16 May and it was assumed that all pellets were cast during the 1980 breeding season. Mic- rotines were identified to species on the basis of dentition (Bee and Hall 1956; Hall and Kelson 1959), and a collec- tion of dentition was sent to the University of Alaska for verification. When dentition was lacking or badly frag- mented, prey remains were placed in higher taxonomic categories. Most pellets were collected at scattered and often previ- ously unsearched locations. Since the date when they were cast could not be determined accurately, trends in food habits were determined by direct observations of prey brought to young and by discarded prey remains. Numerical abundance of prey from pellets was deter- mined by counting pairs of small mammal jaws and by examining skeletons of larger mammals and birds. The biomass contribution of each species was calculated by multiplying numbers of individuals found by mean prey wt. Average prey wts were determined from specimens in the University of Alaska Museum (Appendix 1). Results and Discussion Food Habits. A total of 651 prey remains was recovered from the 2 nests, including at least 4 species of birds and at least 8 species of mammals (Table 1). Mammalian prey comprised over 94% of the combined total biomass, with birds contributing the remainder. Diets of both pairs of owls were similar qualitatively, but differed quantitatively, especially in relative use of Clethrionomys rutilus. Pellets from the 2 nests averaged 1.53 and 1 .72 prey items, respectively, for an overall average of 1.61 prey items/pellet (range 1 to 4) at both nests. Mik- kola (1972) found an average of 1.7 prey items/ pellet in Finland. Microtine voles, particularly C. rutilus and Mi- crotus sp., were the most important prey of hawk- owls, contributing at least 70% of the total prey biomass. Mikkola (1972) found that voles, particu- larly Clethrionomys sp. and Microtus sp. were ex- tremely important in the diet of hawk-owls in Fin- land, Norway, and Russia, contributing 94.8, 98.3, and 97.7% respectively, of the total prey items. Clethrionomys sp. was numerically most important in all countries except Finland, where Microtus sp. was most prevalent. Although infrequently rep- resented in European studies, the Water Vole (Aruicola terrestris) comprised 99,4% of the prey ta- ken by 2 pairs of hawk-owls nesting on Ulkokrunni Island, Finland in 1977 (Pulliainen 1978). Thus, use of microtines by hawk-owls in this study is com- parable to other areas. The Varying Hare (Lepus americanus) and Red Squirrel (Tamiasciurus hudsonicus) comprised over 20% of the total prey biomass, a surprisingly large percentage considering that the biomass contribu- tions of sciurids and leporids have not been quan- tified previously, although hawk-owls are known to prey on them. Dixon (1938) claimed that the Great Horned Owl (Bubo virginianus) and hawk-owls were important predators of Varying Hares in Denali National Park. Henderson (1919) observed hawk- owls carrying remains of Varying Hare, but con- cluded that they probably had been scavenged. On 27 May the wing of an adult Willow Ptarmi- gan (Lagopus lagopus) was found near a pile of hawk-owl pellets. Flesh remaining on the wing was extremely dessicated, indicating that the ptarmigan had not been captured recently. Ptarmigan, and other grouse, apparently are not important prey items during the breeding season (Table 1), al- though they are reportedly taken during winter (Fisher 1893). Birds, especially L. lagopus, were ta- ken 30 times more frequently during winter than summer in Finland (Mikkola 1972). During the time hawk-owls are confined to the vicinity of their nests, the Gray Jay (Perisoreus canadensis) is probably Fall/ Winter 1986 Northern Hawk-Owl in Alaska 93 Table 1 . Relative frequency of occurrence and relative biomass of prey in the diet of 2 pairs of Northern Hawk-Owls in Denali National Park, Alaska. Total number of prey items=651; total prey biomass— 20.641 kg. Species % Numbers % Biom^ Bird Tetraonidae Lagopus lagopus 0.15 2.60 L. lagopus or Canachites canadensis 0.31 0.39 Corvidae Perisoreus canadensis 0.92 2.09 Fringilidae Spizella arborea 0.31 0.07 Zonotrichia leucophrys 0.15 0.12 Small bird 0.92 0.64 Mammal Soricidae Sorex cinereus 1.39 0.18 Sorex hoyi 0.15 0.01 Leporidae Lepus americanus 0.92 9.43 Sciuridae Tamiasciurus hudsonicus 2.15 10.85 Cricetidae Clethrionomys rutilus 49.00 35.54 Microtus miurus 5.84 4.97 M. miurus or Microtus pennsylyanicus 5.53 4.71 Microtus oeconomus 19.82 18.13 Microtus sp. 5.22 4.45 Lemmus sibiricus 0.46 0.49 unidentified microtine 6.76 5.33 Total 100.00 100.00 a more important source of food than grouse (Ta- ble 1). Trends in Predation. Hawk-owls exploited hares, squirrels, and birds in late May and con- tinued to do so until observations ended on 5 July. Predation on these larger animals was related to the availability of large numbers of easily captured young. Predation by hawk-owls on Varying Hares was restricted entirely to juvenile hares, taken between 31 May and 24 June. O’Farrell (1965) estimated that first litters of hares were born in late May and that the breeding season ended in late July near Fairbanks, Alaska. Red Squirrels were taken by hawk-owls between 17 May and 2 July. Although owls preyed predom- inantly on juvenile squirrels, they also took adults. Since Red Squirrel populations do not fluctuate as widely as those of hares, Red Squirrels probably represent a more uniform food source from year to year than do hares. Juvenile Gray Jays were taken by hawk-owls be- tween 25 May and 19 June. Young Gray Jays are generally available as early as 15 April; thus they may have been taken more frequently prior to the beginning of observations. Most migrant birds ar- rived in late May or early June, and fledglings of migrant species generally appeared during the 2nd wk of July. Other than the nestlings and occasional adults of a few migrant, ground-nesting species, 94 Kenneth Kertell Vol. 20, No. 3/4 Figure 1 . Portion of the bog where male hawk-owl from nest A frequently hunted in 1980. such as the American Tree Sparrow (Spizella ar- borea) and White-crowned Sparrow ( Zonotrichia leucophrys), owls did not regularly prey on migrant birds. Hunting Habitat. Hawk-owls in Denali National Park frequently hunted in open areas with scat- tered trees. The male at nest A, for example, hunted a white spruce bog where 60% of 25 ob- served hunting strikes took place (Fig. 1). The bog, located at 0.60 km NW of the nest, was in an area of widely spaced, stunted white spruce < 4 m tall. The sparse open understory was composed of willow, labrador tea, and blueberry. Poor drainage prom- oted the growth of a thick sphagnum ground layer. The open understory and sphagnum substrate apparently enabled the male owl to hunt easily. The male at nest B was observed also to hunt an area with short white spruce and a ground cover of scat- tered shrubs and thick sphagnum. Foraging Behavior. Hawk-owls captured prey by pouncing from an elevated perch (Table 2). Ele- vated perches were always spruce trees, and 92% (N=25) of the perches were at the top of a tree. When scanning for prey, owls leaned forward so that the body and tail were nearly horizontal, and the head was tilted downward, presenting a very kestrel-like silhouette. When prey was located the owl’s head “snapped” into a fixed position and the body became rigid. When making a strike, owls launched into a gliding dive. If the strike distance was great (Table 2), owls flapped their wings a few times before beginning their descent. Roughtly 2/3 of the hunting strikes of male hawk-owls were suc- cessful (Table 2). When potential prey was not properly situated, hawk-owls leaned far forward while engaged in exaggerated tail pumping, a kes- trel-like behavior. In extreme cases owls opened their wings and appeared as if to pounce, almost falling off the perch before regaining their balance. At other times owls glided to a lower perch and waited. On 3 July, for example, a male was perched atop a 6 m spruce when he apparently located prey below and immediately flew 3 m and perched at the top of a 2 m spruce. After 20 sec, he glided to a perch 0.60 m high and pounced onto a vole. Hov- ering by hawk-owls has been noted (C. Collins pers. obs.; Mikkola 1983), but was not observed in this study. The young of ground-nesting birds were cap- tured on the ground. On 22 June a male owl dropped from its perch atop a 5 m spruce and took a tree sparrow nestling from the nest. Twenty min later the male owl returned, descended to the same nest, and took the remaining nestling. I did not observe the manner in which owls captured fledgl- ings or adult birds. Hawk-owls may take arboreal prey in a different way. On 15 June, a perched male turned to face a tree about 7 m away and launched into a rapid glide directly toward a young Red Squirrel climbing the trunk. The owl flew directly toward the trunk, and hit a branch, but the squirrel moved out of range before contact was made. Feeding Behavior. Hawk-owls generally “pre- pared” prey before feeding. Microtines were evis- cerated prior to, or sometimes after, removal of the head. Prey items were eviscerated by a tear in the side, which opened the peritoneal cavity just an- terior to the hindlegs. Owls pulled out and dis- carded the intestines and the stomach. The re- Fall/ Winter 1986 Northern Hawk-Owl in Alaska 95 Table 2: Hunting success, perch height, and strike distance of hawk-owls in Denali National Park, Alaska. No. OF Observations Mean Success (%) Mean (M) Range S.D. Hunting success male 28 68 female 3 5 20 total 33 61 Perch height (male) 25 5.41 0.61—10.61 2.61 Strike distance (male) 18 8.10 0.91—21.21 5.47 a 80% of the female’s strikes occurred while her tail feathers were molting. mainder of the organs were eaten, and a few times the intestines were swallowed as they were pulled from the rodent. Large prey items were not eviscer- ated, at least not immediately, but the organs may have been discarded or consumed at a later time. Varying Hares, Red Squirrels, and Gray Jays often were partially plucked before they were eaten. Except for very small prey, such as fledgling sparrows and young microtines, which were swal- lowed entire, hawk-owls always began feeding by removing and eating the head, including the rela- tively large heads of Red Squirrels. In the case of microtines, after the head was removed the re- mainder was usually swallowed intact. Prey larger than Microtus sp. were dismembered more com- pletely and eaten in several pieces. Adult owls did not always completely consume large prey. At vari- ous perch sites I found the discarded tails and hindlegs of Red Squirrels, and the legs of Gray Jays. Owlets were observed swallowing the legs and tails of squirrels on occasion. Food Caching. Hawk-owls cached excess food 47 times during observations. Food was cached more frequently after owlets left the nest than when they were in the nest. During incubation and brooding, when the female remained at the nest, the male conducted all caching and food retrieval. When brooding of the young was completed the female also cached and retrieved prey. The male was twice observed caching prey in a favorite hunt- ing area about 0.60 km from the nest. Prey almost always was cached at least 3 m above the ground on spruce boughs or spruce brooms (caused by the rust Chrysomyxa arctostaphylii) . All sizes of prey were cached. Some large prey items were fed upon periodically for up to 24 hrs. Three rodents were retrieved and consumed 5 hr, 1 hr, and 15 min, respectively, after being cached. Smith (1922) first observed food caching by a hawk-owl during the breeding season, and Collins (1976) and Ritchie (1978) described the food cach- ing behavior of captive and wild hawk-owls, respec- tively. Nest Trees. In addition to the 2 nests studied in 1980, single nests were found in 1977 and 1982. All nests were located inside the hollow tops of white spruce trunks 2 to 10 m above the ground (Table 3). In Europe, nests were usually 4 to 5 m above the ground (range 2 to 13) (Glutz von Blotzheim and Bauer 1980). All nest trees were dead, and in all cases nest cavities probably formed when the tops of diseased trees blew off, exposing the hollow upper trunk (Fig. 2). The nest cavities were characterized by sections of old trunk projecting 0.3 to 0.9 m above the nest. Owls entered the nest, cavity over low points in this shell. Eggs were laid directly on decomposed sapwood. Nesting Chronology. In 1980, owls were seen near nest A on 17 April by park employees. On 18 April, a rodent was passed from one adult to another at a habitually used perch. In 1984, when nest B was first reoccupied, owls called near the nest tree on 24 March. On 27 March, one adult was perched at the nest cavity entrance and a microtine was exchanged nearby. According to Eckert (1974), hawk-owls begin breeding (presumably selecting nest sites) in mid-March, and sometimes as early as February. Henderson (1919) observed a pair “in 96 Kenneth Kertell Vol. 20, No. 3/4 < c u P .£ 49 ! -C X3 H eo p CM CM oq X 00 uO to C/3 co irj oq d to OO oo s 2 % z w M z H oO « 2 z ~ H Z W s w M t3 < W s PQ o OO £ 00 05 X! o "tt* CM : n w 05 05 d o o o z Tf CM OO OO CM CM CM X! oO CM 05 to f-H CM to lO 00 CM lO cm 2 to cm 1 fl & V 2 i o JS 0j 6 a •S' i -*• S' > W> ‘5 a -o w to d - ” « - S OO $ £ 00 l> 05 00 a |d 3 JS ■M in C .g P *g JJ & ?3 P u u oi v 4) "3 a a I I 8 £ r £ H (5 ■So *3 J5 97 Fall/ Winter 1986 Northern Hawk-Owl in Alaska Figure2. Hawk-owl nest tree (nest B) discovered in 1980. the act of breeding” on 19 February in Alberta, Canada. Mikkola (1972) found that they began calling as early as 17 February in Finland, and after the beginning of March in Russia, with territories being established a “few weeks” before nesting be- gan. The initiation of breeding apparently can be- gin as late as early May (Harrison 1973). Assuming an incubation period of 28 d (Harrison 1973; Terres 1980), and back-dating from the date of egg-hatching, the mean date of clutch initiation in 1980 was calculated to be about 19 April (range 13 to 24 April). Elsewhere in Alaska nests contain- ing eggs range from 16 April to 18 May (Gabrielson and Lincoln 1959). In Alberta and central to south- ern Canada, eggs normally were found between 30 March and 5 June, and in Labrador and New- foundland between 9 May and 11 June (Eckert 1974). Extreme dates when eggs were found in Lapland and Finland range from 30 March to 23 June. The mean date of hatching in 1980 was 17 May (11 May to 22 May). This estimate was based on the condition of the young at nest A. On 29 May the nest contained 4 downy, white young, all with their eyes closed. Spotted Owls and Short-eared Owls open their eyes at 8 to 9 and 7 to 8 d after hatching (Clark 1975; Forsman 1981). Assuming that hawk-owls open their eyes at about 7 to 9 d, and considering the different sizes of the young at the time the nest was examined, I estimated the oldest young to be about 1 wk old when I first examined the brood in late May. Hawk-owls left the nest in early and mid-June (1 to 5 June and 11 to 15 June). If calculations of hatching dates were correct, the young left the nest when approximately 20 to 22 d old (Fig. 3). Roles of Adults During Incubation. Incubation was performed entirely by the female, while the male did all the foraging. Mikkola (1972) also found that females did all incubation. The female at nest A remained on the nest except for short periods when she left to receive food, preen, cast, or defecate. When not foraging the male perched in the tops of nearby trees about 30 m from the nest. Food was usually exchanged away from the nest. Generally the female did not respond immediately when the male arrived with food and he either cached the food or, more commonly, flew to the nest and perched at the cavity entrance. The male frequently flew to the nest entrance several times before the female left the nest and accepted food at a nearby perch. Roles of Adults During the Nestling Period. The female at nest A brooded the young almost constantly for the first 10 d to 2 wks after eggs hatched. During this time, foraging was conducted entirely by the male. Until the young were about 2 wks old the female received all food at the nest. After the 2nd wk the female left the nest to receive food at nearby perches. Toward the end of the nestling period the female spent almost all of her time perched outside the nest. At this time the male visited the nest only to deliver prey and, when not foraging, usually perched at least 100 m from the female. Roles of Adults During the Post-Nestling Period. During the first 10 d after the young left the nest females perched nearby constantly. When not foraging, males continued to perch about 1 00 m from females. Ten to 1 1 d after leaving the nest, owlets moved 98 Kenneth Kertell Vol. 20, No. 3/4 Figure 3. Owlets approximately 17 days after leaving the nest, 37 days old on 28 June. further from nest trees, flying up to 30 m horizon- tally and frequently landing on the ground. When they landed on the ground near a potential perch, they usually would climb. At this time males began to perch nearer the young and even brought food directly to them on occasion. Males were observed to offer only small intact prey to the young, while females often fed owlets pieces of prey. The female at nest A was first seen hunting about 2 wks after owlets left the nest and by 27 June, 3 days later, roles of the sexes had changed drasti- cally. The female was now absent for periods of at least 5 h and, although presumably hunting part of the time, seldom brought food to the owlets. The male fed and guarded the young in the absence of the female, and owlets were left alone for varying lengths of time when the male foraged. On 5 July, the last day of observation, the male at nest A con- tinued to perch near the young and provided al- most all their food. The female at nest B was not observed hunting. Clutch and Brood Size. Clutch and brood sizes of nests in this study (Table 4) were similar to those reported elsewhere. According to Bent (1938), hawk-owls lay between 3 to 9 eggs, usually 7. Mik- kola (1972) recorded a mean clutch size of 6.31 (range 3 to 13), and a mode of 5 for 135 completed clutches in Europe. Nest Success. Both nesting attempts in 1980 were successful, with no infertile eggs or nestling mortality. Hawk-owls also nested successfully in 1976 and 1977. Virtually no quantitative informa- tion is available on nest success or reasons for nest failure in hawk-owls. Tail Molt. Mayr and Mayr (1954), and Collins (1961) summarized information on tail molt of sev- eral species of small owls, although tail molt of the hawk-owl has not been well described. Wheelwright (1863:8443) stated that “the old birds may be seen in deep moult, without tails, even before the young are flyers.” Only the female at nest A molted her tail during the nesting period. The pair at nest B dispersed before tail molt was initiated by either adult. Tail feathers of the female at nest A were first noticed Table 4. Productivity of hawk-owls in Denali National Park, Alaska. Year Nesting Attempts Clutch Sizes Brood Sizes -+- Fledglings / Successful Nest 1976 a 2 5,6 5,6 5.5 (2) 1977 a 2 5,5 5,5 5.0 (2) 1980 2 4,5 4,5 4.5 (2) a nesting attempts, clutch sizes, and brood sizes in 1976 and 1977 are represented by minimum numbers, based on family groups located. Fall/Winter 1986 Northern Hawk-Owl in Alaska 99 missing on 24 June, and only the 2 central tail feathers remained on 26 June, indicating that the molt was centripetal; the innermost rectrices were last to molt. On 29 June all her rectrices were mis- sing. By 1 2 July her new tail feathers appeared to be about 20.0 mm long, or about 12% of their total length (Eckert 1974). Among smaller owls (those with wing lengths < 210 mm) the tail molt is simultaneous, while among larger owls (those with wing lengths of > 230 mm) it is usually gradual or irregular (Mayr and Mayr 1954). Wing lengths of male and female hawk-owls average 220.9 mm and 226.0 mm, respectively (Earhart and Johnson 1970). Simultaneous tail molt in the hawk-owl, then, would extend the upper limit of wing lengths of owls predicted by Mayr and Mayr to undergo simultaneous molt. Since the tail feath- ers of small owls usually are shed over a period of several days to several weeks, Forsman (1981) has suggested that the word “simultaneous” be used sparingly. Nest Defense and Natural Enemies. Of the in- terspecific encounters witnessed, a male hawk-owl defended its nest most vigorously against a North- ern Goshawk (Accipiter gentilis). On 20 May the owl attempted to intercept a goshawk that was flying directly toward the nest tree. The goshawk was 200 m away, and flying rapidly about 35 m above the ground when the owl left its perch and flew toward it. The hawk-owl flew past the goshawk without striking it, and then banked and pursued the goshawk until the accipiter was about 40 m beyond the nest. Other than the goshawk encounter, hawk-owls remained perched when other raptors flew into view. The Golden Eagle (Aquila chrysaetos), for example, soared high over the nest at least once every 2 observation days, but hawk-owls only watched until the eagle disappeared from view. Other raptors elicited a more vigorous response. On 2 1 May, a perched male hawk-owl stiffened as a Red-tailed Hawk (Buteo jamaicensis harlani) sailed rapidly over the nest. Although it remained perched, the owl called several times and was visibly agitated. I observed no instances of hawk-owls being pur- sued by other raptors and no instances of predation on adults or young were recorded. Hawk-owls, however, often were harassed by other birds, par- ticularly the Gray Jay, American Robin (Turdus migratorius) , and Varied Thrush (Ixoreus naevius). Robins and Varied Thrushes attacked hawk-owls vigorously, diving from above and in 3 to 4 in- stances struck perched owls. These attacks dis- rupted the activities of hunting owls, and on several occasions males flew to the nest area with Robins or Varied Thrushes in pursuit. A male hawk-owl once responded aggressively when it was attacked by an American Kestrel (Falco sparverius). During the des- cent phase of each of the kestrel’s 10 pendulum attacks, the owl jumped from its perch into the air and presented its talons to the falcon. Cryptic Posture. On 2 different occasions, once in response to the approach of a goshawk and once in the presence of a low-soaring Golden Eagle, male hawk-owls assumed vertically elongated postures. The owl stiffened and the feathers of the breast, belly, and back were drawn tightly to the body. The wings also were pulled tightly against the body and the leading edge was aligned vertically. The feath- ers in the facial disc above the eyes were raised, making the eyes appear very large. The posture was identical to the “concealing pose” of the Northern Saw-whet Owl (Aegolius acadicus) and the Boreal Owl (Aegolius funereus) as described by Catling (1972), and apparently is the same posture assumed by several other small strigids, including the Eastern Screech Owl (Otus asio), Long-eared Owl (Asio otus), and Elf Owl (Mi- crathene whitneyi) (Bent 1938; Ligon 1968). 1981 Breeding Season. Hawk-owls were seen occasionally in 1981, and did not nest in the study area. Other researchers have noted similar declines in hawk-owl numbers and reproductive success in interior Alaska (Dixon 1938; Murie 1963). Even though hawk-owls were not observed to breed in 1981, there were 1 4 sightings of single owls between 24 March and 15 September, Twice owls were ob- served < 2 km, and once only 0.3 km from 1980 nest sites. Although hawk-owls feed on birds, squirrels, and young hares, they apparently depend on microtines for successful nesting, thus resembling other strigid rodent specialists which also respond to low rodent densities by failing to breed. Among 10 species of Fenno-Scandian owls, hawk owls were second only to Snowy Owls (Nyctea scandiaca) in the proportion of Microtinae in the diet (Mikkola 1983). Hawk-owls did not breed again in the study area until 1984 when a pair layed eggs at a nest used in 1980. It was not determined if the owls bred suc- cessfully. 100 Kenneth Kertell Vol. 20, No. 3/4 Acknowledgements Permission to do research in Denali National Park was granted by John Dalle-Molle. During the Field work I was aided by several park employees, especially Karen Laing, Rick McIntyre, and Rick Sladick. Weights of prey species were provided by Daniel D. Gib- son and Stephen O. Macdonald of the University of Alaska Museum. Stephen O. MacDonald also provided valuable assis- tance in the identification of shrews and microtines. James R. Koplin, Humboldt State University provided guidance and freely imparted his knowledge of raptor biology throughout the study. Literature Cited Bee, J. W. and E. R. Hall. 1956. Mammals of northern Alaska. Lawrence, Kansas. Univ. of Kansas. Bent, A. C. 1938. Life histories of North American birds of prey. Part 2. Smithsonian Inst., U.S. Natl. Mus. Bull. 170. Catling, P.M. 1972. A behavioral attitude of Saw-whet and Boreal Owls. Auk 88: 195-196. Clark, R.J. 1975. A field study of the Short-eared Owl Asio flammeus Pontoppidan in North America. Wildl. Mono. 47:1-67. Collins, C.T. 1961. Notes on the feeding behavior, metabolism, and weight of the Saw-whet Owl. Condor 65:528-530. Collins, C.T. 1976. Food caching behavior in owls. .Rap- tor 7?^. 10: 74-76. Dixon, J.S. 1938. Birds and mammals of Mount McKin- ley National Park, Alaska. Wash., D.C. Earhart, C.M. and N.K. Johnson. 1970. Sizedimorph- ism and food habits of North American owls. Condor 72:251-264. Eckert, A.W. 1974. The owls of North America. Doubleday and Co., Inc., Garden City, New York. Fisher, A.K. 1893. The hawks and owls of the United States and their relation to agriculture. U.S. Dept, of Agric. Div. Ornithol. and Mammal, Bull. 3:1-210. Forsman, E.D. 1981. Molt of the Spotted Owl. Auk 98:735-742. Fyfe, R.W. 1976. Status of Canadian raptor populations. Can. Field-Nut. 90:370-375. Gabrielson, I.N., and F.C. Lincoln. 1959. The birds of Alaska. Stackpole Co., Harrisburg, Pennsylvania and the Wildl. Manage. Inst., Wash., D.C. Glutz von Blotzheim, U.N. and K.M. Bauer. 1980. Handbuch der Vogel Mitteleuropas. Vol. 9. Akademische Verlagsgesellschaft, Wiesbaden. Hall, R.E. and K.R. Kelson. 1959. The mammals of North America. 2 vols. New York. The Ronald Press Co. Harrison, C. 1973. Hawk owls, pp. 147-163. In Burton, J.A. (Ed.) Owls of the world. New York. E.P. Dutton Co. Inc. Henderson, A.D. 1919. Nesting of the American Hawk Owl. Oologist 36:59-63. Ligon,J.C. 1968. The biology of the Elf Owl, Micrathene whitneyi. Misc. Publ. Mus. Zool. Univ. Mich. No. 136. Mayr, E. and M. Mayr. 1954. The tail molt of small owls. Auk 71: 172-178. Mikkola, H. 1972. Hawk Owls and their prey in north- ern Europe. Br. Birds 65: 452-460. Mikkola, H. 1983. Owls of Europe. Vermillion, S.D. Buteo Books. Murie, A. 1963. Birds of Mount McKinley, Alaska Mount McKinley Nat. Hist. Assoc. Newton, I. 1976. Population limitation in diurnal rap- tors. Can. Field-Nat. 90:274-300. O’Farrell, T.P. 1965. Home range and ecology of snowshoe hares in interior Alaska. J. Mammal. 46:406-418. Pulliainen, E. 1978. Nesting of the Hawk Owl, Surnia ulula, and Short-eared Owl, Asio flammeus, and the food consumed by owls on the island of Ulkokrunni in the Bothnian Bay in 1977. Aquilo Ser. Zool. 18: 17-22. Ritchie, R.J. 1978. Food caching of nesting wild hawk owls .Raptor Res. 14:59-60. Smith, D.A. 1970. Observations on nesting Hawk Owls at the MerBleue, near Ottawa, Canada. Can. Field-Nat. 84:377-383. Smith, F.N. 1922. The American Hawk Owl. Can Field-Nat. 36:68-71. Sparks, J. and T, Soper. 1970. Owls: their natural and unnatural history. Newton Abbot, England. David and Charles. Terres, J.K. 1980. The Audubon society encyclopedia of North American birds. New York. Alfred A. Knopf Viereck, L.A., C.T., Dyrness and A.R. Batten. 1982 1982 revision of preliminary classification for vegeta- tion of Alaska. Unpubl. PNW-106, 1980. Inst, of N. Forestry, Univ. of Alaska, Fairbanks. Walker, L.W. 1974. The book of owls. New York. Alfred A. Knopf. Wheelwright, H. 1963. Notes on the Hawk Owl (Strix funerea ), and Tengmalm’s Owl (Strix tengmalmi) as ob- served in Lapland. Zoologist 21:8442-8444. U.S. Fish and Wildlife Service, Alaska Office of Fish and Wildlife Research, 1011 E. Tudor Road, Anchorage, Alaska 99503. Present address: 211 La Vida Way, Davis, California 95616 Received: 16 December 1985; Accepted: 25 June 1986. Fall/Winter 1986 Northern Hawk^Owl in Alaska 101 Appendix 1. Weights of prey species used to compute biomass consumption by hawk-owls. Species No. OF Specimens Mean Weight (g) Source Birds Lagopus lagopus 60 550 UA a L. lagopus or Canachites canadensis — 40 estimated mean juvenile wt. Perisoreus canadensis 33 72 UA a Spizella arborea — 7 estimated mean juvenile wt. Zonotrichia leucophrys 26 25 UA a Small birds — 22 estimated mean juvenile wt. Mammals Sorex cinereus 25 4 UA a Sorex hoyi 25 3 UA a Lepus americanus 24 325 mean juvenile wt., UA a T amiasciurus hudsonicus 29 160 b mean wt., UA a Clethrionomys rutilus 25 23 UA a Microtus miurus 25 27 UA a M. miurus or Microtus pennsylvanicus 20 27 UA a Microtus oeconomus 25 29 UA a Microtus sp. — . 27 estimated Lemmus sibiricus 25 34 UA a Unidentified microtine — 25 estimated Specimens in University of Alaska Museum b Mean weight from a combination of adult and juvenile weights Third World Conference on Birds of Prey, 1987. An International Conference will be held 22-27 March 1987 at Eilat, Israel. The Conference will be organized by the World Working Group on Birds of Prey in conjunction with the Israel Raptor Information Center and the U.S. Hawk Mountain Sanctuary Association. The Conference will consist of seven paper sessions, each of which may occupy up to one whole day. The themes and organizers are as follows: 1) Conservation and biology of rare raptors — U.-Meyburg and N. Collar; 2) Conservation and biology of rare owls — R. J. Clark and H. Mikkola; 3) Raptors on migration and wintering grounds — M. Fuller and J. M. Thollay; 4) Population biology and breeding — I. Newton; 5) Raptors in polluted environments — R. Risebrough and J. Ledger; 6) Educa- tion — Y. Leshem and J. Brett; 7) Legislation — P. Robinson. Contributions to these different themes can also take the form of poster papers. The Conference will take place within the framework of an international festival, which will include a raptor photography competition (under the patronage of Eric Hosking), a painting and drawing competition (patron, Roger Tory Peterson), a film festival and competition, and ornithological and cultural excursions and tours. During this season, the famous and massive migration movement of raptors over Eilat is in full swing, and in 1985 included 1.1 million raptors of 30 species. For further information, write to the Honorable Secretary of the World Working Group, Mr. R. D. Chancellor, 15 Bolton Gardens, London SW5 OAL, UNITED KINGDOM. ROOST-TREE CHARACTERISTICS AND ABUNDANCE OF WINTERING VULTURES AT A COMMUNAL ROOST IN SOUTH CENTRAL PENNSYLVANIA Anthony L. Wright, Richard H. Yahner and Gerald L. Storm ABSTRACT — Roost-tree characteristics and abundance of the Black Vulture ( Coragyps atratus) and the Turkey Vulture 0 Cathartes aura) were studied during 2 winters at a communal roost in southcentral Pennsylvania. Vultures selected large conifers for roosting, which were easily accessible and probably offered a nocturnal microenvironment favorable for energy conservation. Turkey Vultures left the roost earlier in the morning than Black Vultures. Numbers of vultures were highest during mid-winter, and Turkey Vultures outnumbered Black Vultures during both winters. Recommen- dations are to preserve forest stands containing conifers in the vicinity of the roost and minimize human disturbances near roosts. Although roosts and perching areas used by vultures have been described (Coles 1938; Davis 1974; Stewart 1978; Rabenold 1983), quantitative descriptions of habitat used by vultures during winter in the northeastern United States are lack- ing. We examined winter roost trees and abun- dance of the Black Vulture (Coragyps atratus) and the Turkey Vulture ( Cathartes aura) at a large com- munal winter roost at the Gettysburg National Military Park, Adams Co., Pennsylvania. Our ob- jectives were to determine (1) characteristics of roost trees used by vultures at the Big Round Top (BRT) roost, and (2) within- and between-year changes in abundance of both species at the roost during 2 winters. Study Area and Methods The study was conducted from 7 December 1982 to 5 March 1983 and from 27 December 1983 to 7 March 1984 at the BRT roost, which was used nightly by vultures during both winters (Wright 1984). The Harpers Hill and the Gettysburg Quarry roosts, used infrequently by vultures, were located within 5 km of the BRT roost (Wright 1984). The BRT roost is in the Gettysburg Basin, which is a wide, level plain, except for low ridges (Socolow 1962). The city of Gettysburg (population 7,200) lies 3 km from the roost. Forests cover 32% of Adams County and are composed of 6% conifer ( Pinus spp., Picea spp.), 81% oak ( Quercus spp.), and 13% northern hardwood (Be- tula spp., Acer spp., Fagus grandiflora) forest types (Considine and Powell 1980). Mean temperature from December to February at Gettysburg is (FC. Annual snowfall averages 73.7 cm, and pre- cipitation from December to February averages 22.7 cm (Ruffiner 1980). Description of the Roost. — Trees with at least 25% of the ground beneath the crown whitewashed by vulture excreta were defined as roost trees. All roost trees were white pine (Pinus strobus) located at the base of BRT. Control trees were those receiving little or no night use by vultures, as indicated by fewer than 2 large splashes of excreta beneath the tree. Control trees were chosen by following a 2-m wide transect in a random direction from each roost tree until an overstory white pine was encountered. Fifteen variables (Table 1) were compared between roost trees and control trees with either single-classification analyses of var- iance or median tests (Daniel 1978; Sokal and Rohlf 1981). Step- wise logistic regression (BMDPLR, Dixon 1981) was used to pre- dict use of a tree for roosting based on variables measured at each tree. The logistic model used was E(s/N) = exp (U)/(l + exp (U)), where U is the linear combination of one or more independent variables, s is the sum of the binary (0, 1) dependent variable, and N is the total sample size. The maximum likelihood method of estimating variables with default options for remove limit (P > 0. 15) and enter limit (P < 0. 10) was used to build the model. Counts at the Roosts. — Counts of vultures at the BRT roost were conducted 2 to 6 d/wk on mornings without measurable precipitation (< 0.25 mm), beginning 35 min before sunrise and cominuing until 100 min after sunrise. A cutoff of 100 min was chosen arbitrarily as birds that did not leave by this time typically remained in the roost for most of the day. When possible vultures flying out of the roost were counted and identified to species from a vantage point that was 280 m from the main roost. A correction factor (2.2 ± 0.8) was determined to account for birds that did not leave the roost during a given count. This factor, based on 5 counts during 1982-83, was the mean ratio of birds flushed to those visible in the roost before flushing. The number of vultures visible (both species combined) in the roost at the end of a count was multiplied by the correction factor to estimate the number remaining in the roost. When large numbers (2= 60) of vultures were visible in the roost at the termination of a count, the count was considered unsuccessful; unsuccessful counts (N — ■ 16/68) were discarded from analyses. The total number of vul- tures in a roost/count was equal to the number of birds leaving plus the estimated number remaining in the roost ( X = 24 birds/suc- cessful count). Winter counts were divided into 3 winter periods: early winter, mid-winter and late winter (see Table 3). Results Comparison of Roost Trees with Control Trees. . . Vultures roosted only in white pines at BRT, although hardwoods made up to 58% of the over- story within the roost and 92% of the overstory within 0.5 km of the roost. Six variables related to tree size and amount of evergreen foliage were significantly great (P < 0.05) for roost trees than for control trees, whereas distance to the nearest roost tree was less for roost trees than for control 102 Raptor Research Vol. 20 (3/4); 102-107 Fall/ Winter 1986 Vulture Roosts in Pennsylvania 103 T able 1 . V ariables measured at roost trees of Black and Turkey Vultures and at control trees at Big Round T op roost Adams Co., Pennsylvania (from Wright 1984). Variable Description Diameter at breast height Height of tree Diameter (cm) of tree measured at breast height (1.5 m) with tree diameter tape. Height (m) of tree measured with Abney level and tape. Height to lowest limb Height (m) from ground level to lowest living limb greater than 6 cm in diameter at base, measured with Abney level and tape. Maximum crown diameter Maximum horizontal distance (m) between the ends of living limbs of trees measured by ocular tube with plumb-bob and tape. Mid-tree crown diameter Horizontal distance (m) between the ends of living limbs measured midway between ground level and tree top. Method of measurement same as crown diameter. Distance to nearest roost tree Distance (m) from roost or control tree to nearest roost tree measured with a 50-m Distance to nearest clearing tape or taken from a 1:1,600 aerial photo. Distance (m) from roost or control tree to nearest area of over 200 m essentially free of overhead vegetation. Measured by same method as distance to nearest roost tree. Number of overstory trees Understory stem density Number of overstory trees in a 0.04-ha circular plot. Density (100’s of stems/ha) of shoulder height non-overstory, woody stems in 2 perpendicular 22.8-m transects in a 0.04-ha circular plot. Percent evergreen canopy cover Evergreen canopy coverage (%) based on 56 ocular tube readings evenly spaced on lines running in 8 main compass directions from center tree of a 0.04-ha circular plot. Slope Maximum ground slope (degrees) from tree to edge of a 0.04-ha circular plot, measured with Abney level. Elevation Canopy height Elevation (m) taken from USGS 1:24,000 topographic map. Mean height (m) of trees in a 0.04-ha circular plot. These are the center tree and the tree with the greatest diameter at breast height in each quarter. Total basal area Basal area (m) of all overstory trees in a 0.04-ha circular plot. Basal area of white pine Same as basal area, but only for white pine. trees (Table 2). Basal area of white pine, understory (understory stem density) + 0.28 (height of tree), stem density, and tree height were the best variables The model gave 81.2% correct classification of for predicting use of a tree for roosting: U = trees. — 10.89 + 6.22 (basal area of white pine) — 0.01 104 Wright et. al. Vol 20, No. 3/4 Table 2. Means ^ and standard deviation (SD) of 15 variables measured at roost trees of Black and Turkey Vultures and at control trees at Big Round Top roost, Adams Co. , Pennsylvania, during winters 1 982-83 and 1 983-84. Roost Tree (N - 33) Control Tree (N - 31) Variable X SD X SD Diameter at breast height 3 57.42* 10.0 48.6 15.1 Height of tree 3 28.8* 2.7 25.8 5.2 Height to lowest limb 17.1 2.5 15.4 4.0 Crown diameter 9.4 2.1 8.2 3.0 Perpendicular crown diameter 3 7.5*** 1.8 5.6 2.7 Distance to nearest roost tree 3 7_g*** 7.1 63.4 40.8 Distance to nearest clearing 109.7 27.7 130.7 89.4 Number of overstory trees 9.7 2.9 8.4 3.0 Understory stem density 97.8 75.1 114.6 111.7 Percent ever- green canopy cover 3 38.3*** 9.0 26.9 9.0 Slope 9.7 2.1 8.9 3.7 Elevation 167.1 0.3 164.7 0.9 Total basal area 3 1.47*** 0.35 1.16 0.36 Basal area of white pine 3 0.90*** 0.33 0.43 0.26 a Means or distribution of means varied between roost trees and control trees; *P = 0.05, ***P = 0.001, based on single-classification analyses of variance or median tests (Daniel 1978; Sokal and Rohlf 1981). Counts at Big Round Top Roost. — The number of both vulture species combined was greater in winter 1982-83 compared to winter 1983-84 (Table 3). Mean number/count varied significantly among the 5 winter periods (F — 45.3; df = 4, 47; P < 0.001). Paired comparisons of means between winter periods were significantly different (P < 0.03), except for the comparison of late winter 1982-83 and late winter 1983-84 (Table 3). As a general trend, numbers increased in early winter, peaked and remained stable in mid-winter, and declined in late winter. Several large day-to-day changes in numbers at the roost also were documented (Wright 1984). Turkey Vultures were more common than Black Vultures at the BRT roost based on all winter periods combined (Wilcoxon paired-rank test, Z = - 6.7, n = 63, P < 0.001). The mean percentage of both Black and Turkey Vultures observed at the roost differed among periods (F — 7.2; df = 4.58; P < 0.001); pairwise comparisons of mean percen- tages of each species observed at the roost were significantly different between most periods (Table 4). Fall/ Winter 1986 Vulture Roosts in Pennsylvania 105 Table 3. Means, SD, and coefficients of variation (CV) for counts (N) of Black Vultures, Turkey Vultures, and vultures of unknown species combined at Big Round Top roost, Adams Co., Pennsylvania, during winter periods of 1982-83 and 1983-84. Period Dates of Counts N Means ± SD CV 1982 - 83 : Early winter 10 Dec 1982-27 Dec 1982 9 517 ± 239 46.1 Mid-winter 28 Dec 1982-16 Feb 1983 15 719 ± 85 11.8 Late winter 17 Feb 1983-5 Mar 1983 7 199 ± 82 41.4 1983 - 84 : Early winter a a a Mid-winter 28 Dec 1983-6 Feb 1984 10 420 ± 74 17.8 Late winter 6 Feb 1984-6 Mar 1984 113125 ± 76 361.0 a A total of 427 and 501 vultures was counted at the roost on 8 December and 17 December, respectively (E. Daniels, pers. comm.) Numbers of individual birds departing the BRT roost/ 15-min time interval in the morning were de- pendent on species (G = 1,082; df = 8; < 0.001). Turkey Vultures tended to leave earlier than Black Vultures (Table 5). Discussion BRT, Harpers Hill, and Gettysburg Quarry roosts are associated with ridges (Wright 1984), which presumably modify winds (Geiger 1965). Be- cause both vulture species often use winds when soaring, ridges may have an effect on roost location by creating updrafts that were used as travel lanes (Wright 1984). Topography is known to affect the distribution of different species of African vultures according to their flight characteristics and body sizes (Houston 1975). Vultures selected mature white pines rather than hardwoods as roost trees at BRT. Coles (1938) ob- served that vultures in Virginia abandoned a hardwood roost site and moved to a conifer roost site after leaf fall; a similar shift took place at BRT (J. Coleman, pers. comm.) Both white pines and hardwoods were used as roost trees at Harpers Hill; Table 4. Mean ± SD of percent composition of Black and Turkey Vultures between winters and among winter periods at Big Round Top Roost, Adams Co., Pennsylvania, 1982-83 and 1983-84. Winter 1982-83 Winter 1983-84 Winter period Black Turkey Black Turkey Early 20.5 ± 8.3 a 79.5 ± 8.3 no data no data no data Mid- 29.4 ± 7.2 70.6 ± 7.2 40.2 ± 11.6 59.8 ± 11.6 Late 33.8 ± 16.0 66.2 ± 16.0 21.3 ± 13.4 78.7 ± 13.4 Combined 28.0 ± 10.5 72.0 ± 10.5 32.5 ± 15.4 67.5 ± 15.4 a All pairwise comparisons for each species were significantly different except between mid-winter 1982-83 and late winter 1982-83, and between all winter 1982-83 periods combined and all winter 1983-84 periods combined; Wilcoxon two-sample and Wilcoxon signed-rank tests (Sokal and Rohlf 1981). 106 Wright et. al. Vol 20, No. 3/4 Table 5. Percentages (numbers) of individual Black and Turkey Vultures departing from the Big Round Top roost, Adams Co., Pennsylvania, during 9, 15-min morning time intervals in winters 1982-83 and 1983-84 combined. Time interval (Relative to sunrise) Percentages (Numbers) of Individual Birds Black Vultures Turkey Vultures 35 to 20 min before 0.4 (17) a 1.5 (167) 20 to 5 min before 9.1 (392) a 21.3 (2471) 5 min before to 10 min after 24.7 (1065) a 40.5 (5012) 10 to 25 min after 22.4 (965) a 12.7 (2203) 25 to 40 min after 15.2 (565) a 8.2 (1453) 40 to 55 min after 14.6 (628) a 8.1 (1417) 55 to 70 min after 9.4 (406) a 5.1 (903) 70 to 85 min after 3.5 (151) a 2.1 (360) 85 to 100 min after 0.7 (31) 0.5 (81) a Numbers of departures per time interval varied between species; P < 0.001, based on 2 x 2 G-tests of independence, where rows are numbers of vultures/time interval of interest versus numbers/all other time intervals combined and columns are the 2 species (Sokal and Rohlf 1981). 3 Virginia pines ( Pinus virginiana) were the major roost trees at Gettysburg Quarry where the forest type was > 95% hardwood (Wright 1984). Conifers reduce both wind velocity and nightly drops in am- bient temperature during winter, suggesting that vultures lower daily energy requirements by roost- ing in clusters of large conifers (Francis 1976; Kelty and Lustick 1977. Stalmaster and Gessaman 1984; Walsberg 1986). Further strong temperature inver- sions form in mature forest stands on calm nights (Geiger 1965); therefore, a perch on an upper limb in a full conifer would afford a warm microenvi- ronment to a roosting vulture. Finally, widely- spaced, horizontal limbs on dominant white pines enabled vultures to easily alight. Numbers using the BRT roost may vary by year according to weather conditions. For example, mid- winter 1982-83 (January mean temperature, — 0. 1°C; monthly snowfall, 3.8 cm) was less rigorous than mid-winter 1983-84 (January mean tempera- ture, — 3.8°C; monthly snowfall, 18 cm). Numbers of vultures using the BRT roost were much lower in winter 1983-84, perhaps due to more vultures mi- grating farther south than in 1982-83. The BRT roost presumably provides a favorable microclimate in mid-winter, but other factors (e.g.; information centers, Rabenold 1983, 1986; protec- tion from predation, Weatherhead 1983; abundant winter food resources, Yahner et al. 1986), also may be important in explaining high use of this com- munal roost. Communal roosting by both species has been observed during summer months (Stewart 1978) and at southerly latitudes (Bent 1937; Coles 1938). Although our results are based primarily on 1 roost in southcentral Pennsylvania, we recommend that forest stands containing conifers should be preserved near communal winter roosts. Efforts should be made to minimize human disturbances (e.g., road construction, forest clear-cutting) within a reasonable distance of a roost. In addition, large trees at pasture — woodland interfaces within 1 km of the roost were used readily by vultures at Gettys- burg National Military Park (Wright 1984) and, thus, should be retained near roosts. Acknowledgments Thanks are given to P. Rabenold, R. Shipmen, K. Steenhof, P. Stewart, J. Swenson and P. Weatherhead for reviewing an earlier draft of the manuscript; to H. Greenlee and J. Karish, National Park Service, for providing logistical support; to E. Daniels and J. Coleman for sharing information about vultures; to M. Fuller, U.S. Fish and Wildlife Service, for advice on field techniques; and J. Grimm for help with statistical analyses. This research was funded by the Pennsylvania Agricultural Experiment Station, the U.S. Fish and Wildlife Service, the National Park Service, and the Fall/ Winter 1986 Vulture Roosts in Pennsylvania 107 Max McGraw Wildlife Foundation. This is Scientific Journal Series Number 7149 of the Pennsylvania Agricultural Experiment Station, The Pennsylvania State University, University Park. Literature Cited Bent, A. C. 1937. Life histories of North American birds of prey. Park I. New York. Dover Publications, Inc. Coles, V. 1938. Studies in the life history of the Turkey Vulture. PhD Thesis, Cornell Univ., Ithaca, New York. Considine, T.J. and D.S. Powell. 1980 Forest statistics for Pennsylvania — 1980. Northeast For. Exp. Stn., Broomall. Daniel, W.W. 1978. Applied nonparametric statistics. Boston. Houghton Mifflin Co. Davis, D. 1974. Roosting behavior of the Turkey Vul- ture. MS Thesis, Idaho State Univ., Pocatello. Dixon, W.J. Ed. 1981. BMDP statistical software. Berkeley. Univ. California Press. FRANCIS, W.J. 1976. Micrometeorology of a blackbird roost. J. Wild. Manage. 40:132-136. GEIGER, R. 1965. The climate near the ground. Cam- bridge. Harvard Univ. Press. Houston, D.C. 1975. Ecological isolation of African scavenging birds. Ardea 63:55-64. Kelty, M.P. and S.L. Lustick. 1977. Energetics of the starling in a pine woods. Ecology 58: 1 181-1185. Rabenold, P.OP. 1983. The communal roost in Black and Turkey Vultures — an information center? Pages 303-329 In S.R. Wilbur and J.A. Jackson 1 Eds.], Vul- ture biology and management. Berkeley. Univ. of California Press. 1986. Family associations in commun- ally roosting Black Vultures, Auk 103:32-41. Ruffiner, J.H. 1980. The climate of the states. Vol. 2. Detroit. Gale Research Co. Socolow, A.A. 1962. Geology and the Gettysburg cam- paign. Harrisburg. Pennsylvania Topographic and Geologic Survey. Sokal, R.R. and F.J. Fohlf. 1981. Biometry. San Fran- cisco. W.H. Freeman and Co. Stalmaster, M.V. and J.A. Gessaman. 1984. Ecological energetics and foraging behavior of overwintering Bald Eagles, Ecol. Monogr. 54:407-428. Stewart, P.A. 1978. Behavioral interactions and niche separation in Black and Turkey Vultures. Living Bird 17:79-84. Walsberg, G.E. 1986. Thermal consequences of roost- site selection: the relative importance of three modes of heat conservation. Auk 103:1-7. Weatherhead, P.J. 1983. Two principal strategies in avian communal roost. Aw. Naturalist 121:237-243. Wright, A.L. 1984. Winter habitat use and abun- dance of Black and Turkey Vultures at Gettys- burg. MS Thesis, The Pennsylvania State Univ., University Park. Yahner, R.H., G.L. Storm and A.L. Wright. 1986. Winter diets of vultures in southcentral Pennsylvania. Wilson Bull. 98:157-160. School of Forest Resources, The Pennsylvania State University, University Park, Pennsylvania 16802 USA. Address of third author: Pennsylvania Cooperative Fish and Wildlife Research Unit, The Pennsylvania State University, University Park, Pennsylvania 16802 USA. Received 1 March 1986; Accepted 1 June 1986. THE BARN OWL EGG: WEIGHT LOSS CHARACTERS, FRESH WEIGHT PREDICTION AND INCUBATION PERIOD James D. Marshall, Claire H. Hager and Gwyn McKee ABSTRACT. — A total of 177 Common Barn-Owl ( Tyto alba pratincola) eggs produced by 14 captive pairs were studied during the spring of 1985, Initial egg parameters for 75 eggs were fresh weight (26.6 ± 1.4 g), length (43.07 ± 1.24 mm) and breadth (33.67 ± 0.70 mm). Using these data, a coefficient (K w ) unique to the barn owl egg was calculated for Hoyt’s (1978) equation for predicting the fresh weight of an egg. (K w = ° .0005453) For 50 artificially incubated eggs (hatchability = 93.5%) the lay to pip (LP) interval was 28.2 ± 1.4 d, the pip to hatch (PH) interval was 2. 1 ± 0.5 d and the overall incubation period was 30 ± 1 .5 d. Variance in the latter period (range: 27-35 d) may have been due to an observed delay in initial embryonic development of from 1-7 d. During incubation, several externally quantifi- able changes occur in the avian egg. These include: 1) the relatively steady reduction in weight due mainly to loss of water vapor by diffusion from the embryonic chorioallantois through the porous shell and its evaporation at the eggshell surface (Romanoff and Romanoff 1949; Ar and Rahn 1980); and 2) the equal exchange of O 2 and CO 2 gases through the eggshell by the chorioallantois - a process not affecting weight loss (Wagensteen and Rahn 1970, 1971). The mean percentage of fresh egg weight (Wo) lost during the incubation period for many avian species ranges from 12-18% (Drent 1970). Proper weight loss is correlated with hatcha- bility and normal embryonic development (Walsberg 1980). During artificial incubation, accu- rate regulation of egg weight reduction is possible through a variety of methods; (Burnham 1983; Weaver and Cade 1983). A mathematical equation (1) based upon egg length (L) and breadth (B), parameters which are invariant during incubation, was developed by Hoyt (1978) to predict avian Wo. Wo = K W LB 2 (1) The coefficient (K w ) of this equation interrelates shell measurements, and may be adjusted to ac- commodate a single species for accurate Wq for Peregrine Falcon ( Falco peregrinus ) eggs, and also observed a reduction in Wq of 15 ± 2% during incubation of normal eggs. However, our study of the incubation of common Barn-Owl ( Tyto alba practincola ) eggs indicates that they cannot be pre- cisely characterized by values developed for Pereg- rine Falcon eggs. Our objective was to measure barn owl egg weight loss and incubation period, and algin Hoyt’s equation for this species. Materials and Methods The barn owl breeding colony of the Raptor Rehabilitation and Propagation Project, Inc., Eureka, Missouri, was established in 1979 and produced more than 150 juvenile owls yearly through 1986 for release into Missouri. The colony contained non-sibling breeding pairs collected from eastern North America. Each pair was housed in an outdoor mew in a natural setting and was fed daily a diet of fresh rodents ad libitum. Human disturbance was normally limited to 2 short intervals. Barn Owls will naturally produce > 1 clutch of 6-8 eggs during favorable seasons (Eckert and Karalus 1974), and often breed repeatedly all year in captivity (Mendenhall, pers. comm.). Thus, 2 clutches/pair of owls were assured. The first clutch produced by each pair was removed for artificial incubation and subsequent clutches were left with the parents for natural incubation. Beginning in early January, approximatley 2 wks before initia- tion of barn owl breeding, each mew was entered daily by 1 or 2 workers and the nest boxes were checked for eggs. This procedure was completed at a prescribed time every morning through April to ensure that no egg was older than 24 hr when initially mea- sured, and to minimize non-random disturbance of the adult owls. As each freshly laid (4- 0-24 hr) egg was discovered, it was weighed on an electric field balance to determine Wo, and the dimensions were measured with a Vernier caliper. Additionally, each fresh egg was marked with a graphite letter corresponding to its sequ- ence in the clutch. No egg was ever fully removed from the nest box and adults were kept at a distance during measurement. During the subsequent incubation period, each egg was weighed every other day using similar methods. To reduce parental stress and promote successful copulation, no eggs were collected from nest boxes prior to clutch completion (W.C. Crawford, Jr. pers. comm.), Egg laying interval was ap- proximately 1 egg every 2-3 d, thus eggs were from 1 to 16 d old when removed from the nest for artificial incubation. Eggs were incubated in Roll-X RX2A automatic rolling incubators with a constant temperature of 37.5° C, and relative humidity of 48%. Each egg was rolled manually ISO 9 3x/d to supplement automatic rolling. Throughout the lay-to-pip (LP) interval, each egg was weighed and candled every other day to determine both weight loss and corresponding embryonic development. Once an embryo had pipped its shell, the egg was placed pipped side upwards in another Roll-X RX2A set at a lower temperature (35° C) but higher relative humidity (60%). Pipped eggs were not turned. During the pip-to-hatch (PH) interval, no weight measurements were made due to shell fragility and difficulty in determining weight at the instant of hatching. Infertile eggs or eggs containing dead em- bryos were removed from the incubators to inhibit bacterial growth. Eggs undergoing natural incubation were weighed similarly through pipping, but only occasionally candled to reduce nest disturbance. No extra care was provided for these clutches (i.e. cleaning of nest boxes, bad egg removal, etc.) unless a shell failed 108 Raptor Research Vol. 20 (3/4): 108-1 12 Fall/ Winter 1986 The Barn Owl Egg 109 Table 1 . Mean total fraction of grams Wo lost over the 28 d lay to pip interval for Barn Owl eggs a incubated and naturally. artifically Incubation N b X SD min/max CASES C r d Artificial 39 0.11 0.02 0.07-0.14 441 0.95 Natural 23 0.14 0.04 0.10-0.24 249 0.87 a Only fertile, successful hatching eggs represented, k Number of eggs. c Number of points used in generating r values and regression lines give Figure 1. ^ Correlation coefficient relating cumulative fraction of Wo lost to day of incubation. in a fertile egg; such eggs were removed for artificial incubation and excluded from the study. To prevent cannibalism, an occa- sional aspect of barn owl adult-chick behavior, the amount of food provided for each mew was increased considerably following the hatch of each egg (W.C. Crawford Jr., pers. comm.). Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS) (Nie et. al. 1975). A regression line developed by the least squares fit was generated plotting the cumulative fraction of Wo lost by corresponding interval day. The resulting linear equation was used as a model (assuming 28 d LP interval) to predict the total fractional weight loss for all cases in each of the 2 incubation type categories. Other SPSS options were used to generate F-Test, t-Test, Pearson’s r and Chi-squared (X 2 ) values and probabilities. Results The mean total fraction of Wo lost during the LP interval was significantly different (F — 07.05 df = P < 0.001) between artificially and naturally incu- bated eggs which hatched successfully (Table 1). High degrees of correlation were found between cumulative reaction of Wo lost and interval day within each incubation group, implying that eggs dehydrated similarly in their respective categories although a wide range of total fraction of Wo lost by individuals was noted. We defined hatchability as the percent of fertile eggs successfully hatched. The hatchability of naturally incubated eggs was 80.9% (n — 62). Hatchability between incubation types was signific- antly different ( x 2 = 4.56; df = P < 0.05). The relationship between day of incubation and cumulative fraction of Wo lost was examined (Fig. 1). An increase in the spread of points (statistically indicated by increasing standard deviations of re- siduals) from the regression line (Table 2), and corresponding decrease in correlation coefficients as incubation progressed through consecutive seg- ments of LP interval were found. Both incubation types had this characteristic. A species specific coefficient (K w = 0.0005453) was determined using equation (1) for Wo predic- tion and the measured values of Wo, L and B col- lected from 75 barn owl eggs (Table 3). Using this K w a strong correlation was found between directly Table 2. Increasing deviation of points from regression lines indicated by increasing standard deviation of residuals and decreasing correlation between fresh Wo lost and incubation day. Incubation cases a r P RESIDUAL SD Artificial 0-10 days 136 0.83 <0.001 0.7906 11-19 days 133 0.77 <0.001 1.0272 20-30 days 172 0.67 <0.001 1.2954 Natural 0- 1 0 days 101 0.75 <0.001 1.3191 11-19 days 85 0.66 <0.001 2.2498 20-30 days 56 0.29 <0.01 3.9209 a Number of points used in generating the r values and regression lines given Fig. 1. CUMULATIVE FRACTION OF FRESH WEIGHT LOST 110 Marshall et. al. Vol. 20, No. 3/4 Figure 1. Regression of cumulative fresh weight lost in barn owl eggs by day of incubation. Fall/Winter 1986 The Barn Owl Egg 111 Table 3. Summary of physical parameters from natural incubation and period of incubation for common Barn-Owl (T.a.pratincola) eggs incubated artifically. Parameter N X SD min/max Length (1) (mm) 75 43.07 1.24 39.95-47.95 Breadth (B) (mm) 75 33.67 0.70 32.50-35.40 Fresh Weight (Wo) (g) 75 26.6 1.4 24.6 -29.9 Lay to Pip (LP) Interval (days) 50 28.2 1.4 25-33 Pip to Hatch (PH) Interval (Days) 50 2.1 0.5 1-4 Incubation Period (days) 50 30.3 1.5 27-35 measured and calculated values of W 0 (r = 0.917; P < 0.001); the 2 group means were similar (t-Test = 0.39; P = 0.701). When the coefficient K w ; 0.0005474 developed by Burnham (1983) was used in equation (1), strong correlation (r = 0.917; P < 0.001) was also evident between measured and calculated values of W G , although statistical confidence in the similarity bet- ween the 2 group means was decreased (t-Test = 1.86; P= 0.067). Discussion The total incubation period of the barn owl can be generalized from the literature as 30-33 d, with extremes of 29 and 34 d (Eckert and Karalus 1974; Bunn et. al. 1982). Our study indicatd a similar mean incubation period and range. The mean Wo value (Table 3) of the barn owl eggs studied is inconsistent with the mean (Wo) de- veloped from the single random sample collected (from the wild) by Sumner (1929), and his values were reported in other works (Drent 1970; Ar and Rahn 1980). However, Hoyt (1978) noted that in- traspecific variability in the values of Wo, L and B could be expected and we have attempted to account for such deviation through relatively large samples collected from many pairs of owls within the subspecies T. a. pratincola. Careful, frequent illumination of eggs with cool, high intensity light provided good visual tracking of embryonic development. A small fraction of em- bryos did not achieve the visible blastodisc stage (indicative of fertility) for up to 7 d following the date of laying. However, most embryos apparently began their development immediately, and showed a blastodisc within 24 hr. A sharp increase in the rate of egg weight loss in conjunction with abrupt initialization of embryonic development in dor- mant-fertile eggs was routinely observed. After an extremely low rate of daily weight loss, these eggs suddenly achieved a relatively constant rate of weight loss which continued for about 28 d until a normal fraction of Wo was lost. The chicks then pipped the eggshell. Thus, a specific weight loss rate occurred for the latter portion of the LP inter- val, although this interval may have been initially extended by the dormant-fertile condition. Since the PH interval was fairly constant, with variance probably due to observational error, nearly all de- viation in the barn owl incubation period was due to the initial dormant-fertile egg. It was unclear whether the dormant-fertile condition was random or relative to other eggs’ development within clutches, but eggs generally hatched in sequence of their laying. Quantification of this embryological characteristic was not possible using their sample and further study is required. Although hatchability and mean total fraction of Wo lost was related to incubation type (natural vs artificial), the 2 incubation methods are very diffe- rently affected. Factors inherent only during natural incubation include frequent variation in nest microclimate and ambient temperature and humidity, high bacterial exposure, and violent movement of delicate eggs by disturbed adult owls. Such relatively uncontrollable variables may have caused natural incubation weight loss rates to occur which do not parallel those of eggs in undisturbed 112 Marshall et. al. Vol. 20, No. 3/4 nests. These adverse factors undoubtedly contri- buted to the lower hatchability of fertile eggs un- dergoing natural incubation, although the sample analyzed includes many eggs from undisturbed nests. Regression of weight lost by interval day reveals an increase in deviation between predicted and ac- tual egg weights during the LP interval. Since weight loss is due to expired water vapor, as previ- ously cited, this unexpected trend may reflect diffe- rential individual respiratory function, effected by the chorioallantois in conjunction with the eggshell, which was not subject to purely passive diffusion. This result contrasts with recent literature which cites simple diffusion down concentration gra- dients as the single force moving gases across the eggshell (Wangensteen and Rahn 1970, 19721). Inferences drawn from these results are in- teresting to both the ecologist and the conser- vationist propagating this species artifically. Tyto alba supp. possess extremely favorable reproduc- tive capabilities. Developmental flexibility is re- flected in the variable egg weight losses achieveable during incubation and in the dormant-fertile con- dition which allows extension of incubation period. These factors may contribute to the high hatchabil- ity evident from the data in this study. Acknowledgments We thank the Raptor Rehabilitation and Propagation Project, Inc., and W. C. Crawford, Jr., who has reintroduced over 500 barn owls in Missouri; Literature Cited Ar, A. andH. Rahn. 1980. Water in the avian egg overall budget of incubation. Amer. Zool. 20:373-384. Bunn, D.S., A.B. Washburton, and R.D.S. Wilson. 1982. The Barn Owl. Buteo Books, Inc. Vermillion, South Dakota, p. 220. Burnham, W. Artificial incubation of falcon eggs. J. Wildl. Manage. 47:158-168. Drent, R. 1970. Functional aspects of incubation in the Herring Gull. Behaviour Suppl. 17: 1-32. Eckert, A.W. and K.E. Karalus. 1974. The Owls of North America. Doubleday and Co., Inc. Garden City, New York. pp. 15-16. Hoyt, D.F. 1978. Practical methods of estimating vol- ume and fresh weight of birds’ eggs. Quk 96:73-77 Nie, N.H., C.H. Hull, J.G. Jenkins, K. Steinbrenner and D.H. Bent. 1975. Statistical Package for the So- cial Sciences, 2nd Ed. McGraw-Hill Book Co., New York. Ramanoff, AL. and A.J. Romanoff. 1949. The Avian Egg. John Wiley and Sons, Inc. New York, p. 378. Sumner, E.L. Jr. 1929. Comparative studies in the growth of young raptors. Condor 31:85-111. Walsberg, G.E. 1980. The gaseous microclimate of the avian nest during incubation. Amer. Zool. 20:363-372. Wangensteen, O.D. and H. Rahn. 1970/71. Respiratory gas exchange in the avian embryo. Respiration Physiol. 11:1-45. Weaver, J.D. and T.J. Cade. 1983. Falcon propagation: a manual on captive breeding. The Peregrine Fund, Inc. Ithaca, New York. Raptor Rehabilitation and Propagation Project Inc., Box 193, Eureka, Missouri 630025. Current address of first author: 1017 Broadway, New Orleans, Louisiana 70118. Current address of second author: Box 193, Eureka, Missouri 63025. Current ad- dress of third author: Box 2007, Bartlesville, Oklahoma 70432. Received 1 February 1986; Accepted 3 November 1986. Fall/Winter 1986 Ecology of South American Owls PREY AND TROPHIC ECOLOGY OF GREAT HORNED OWLS IN WESTERN SOUTH AMERICA: AN INDICATION OF LATITUDINAL TRENDS Fabian M. Jaksic, Jose L. Yanez, and Jaime R. Rau Abstract — Quantitative information on the diet of three Great Horned Owl (Bubo virginianus) popula- tions along 18 lat. degrees in western South America (Chile) is compared with that of Great Horned Owls in comparable latitudes along western North America. In Chile, owls preyed mainly on small mammals, with proportion of birds decreasing, and that of insects increasing, toward southern latitudes. Mean prey size and diet breadth declined toward southern Chile. These latitudinal trends closely mirror those documented in western North America. Although the Great Horned Owl ( Bubo vir- ginianus) is distributed throughout the Americas, its food habits have received considerable study mainly in North America (Burton 1973). The only published quantitative information on their food habits in South America comes from central Chile (approximately latitude 33° to 38°; see Jaksic and Yanez 1980; Jaksic and Marti 1984). Except for a preliminary report by Jaksic et ah (1978), no dietary information was previously available from their southernmost distribution (see Humphrey et al. 1970). Here we report the prey identified in 125 fresh pellets collected in September (austral spring) 1977 and in 14 other pellets collected in July (winter) 1978, from under the same nest located at Torres del Paine National Park (approximately 51° 01'S, 72°54'W; 142 km north of Puerto Natales). For purposes of comparison we report earlier diet- ary data published by Reise and Venegas ( 1 974) in a Chilean journal of very local circulation. Their study material (an unreported number of fresh pellets, ±55) was collected under one nest, located 10 km north of Puerto Ingeniero Ibanez (46° 18' S; 71°55'W), in January (summer) 1971. For com- parative purposes we also use Jaksic and Yanez’s (1980) report on the prey of the Great Horned Owl at La Dehesa (33°21'S, 7(L32'W; 20 km east of Santiago), based on 98 fresh pellets collected dur- ing September (spring) 1979, beneath one nest. Although the information analyzed is based on very small sample sizes, we believe it is useful in con- solidating new and old information fragmented in the Chilean literature and not readily available to ornithologists elsewhere. Methods Considering that ca. 95% of the pellets analyzed reflect spring and summer diet, and that this dietary information covers ap- proximately 18° latitude, a quantitative comparison seems war- ranted. We use the following trophic metrics: (a) Geometric mean prey weight in the diet — essentially the back-transformation of the mean prey size obtained with log-transformed weight data, weighted by their relative occurrence in the diet (see Jaksic and Braker 1983 for formula, justification, and assumptions of this trophic statistic). Prey sizes are mean weights of small mammals in Table 1. (b) Diet breadth — the diversity of prey in the diet as computed by Levins’ (1968) index: Bobs = l/(Spi^)> where/?* is the relative occurrence of prey taxon i in a given population’s diet This index generates values between 1 andn (whenn resources are used equally). Because Levins’ index increases with the number of prey taxa, a standardization is necessary when comparing popula- tions in different localities, where the availability of prey taxa may differ. Colwell and Futuyma (1971) provide a standardized ver- sion of Levins’ index: Bsta - (Bobs - Bmin)/(Bmax - Bmin), where Bobs is the observed niche breadth (= Levins’ index), Bmin is the minimum niche breadth possible (= 1), and Bmax is the maximum niche breadth possible (= n), which is the number of prey taxa actually taken by a given owl population (i.e., each of the taxa that receives a separate line in Table 1; generally, species for mammals and orders for insects). This standardized index renders values between 0 and 1 (i.e., between a comparatively narrow niche breadth, with disproportionately high representation of one or a few prey items, and a broad one, with a more even consumption of the available prey categories, respectively). Results and Discussion Results are summarized in Table 1, and are here discussed in a north-south succession. In La De- hesa, the owls preyed upon all small mammals known to occur in the locality (see Jacks! c et al. 1981), with the exception of the rodents Octodon degus (a semi-fossorial species) and S palac opus cy anus (a truly fossorial one). Jaksic and Yanez (1980) 113 Raptor Research Vol, 20 (3/4): 113-116 114 Jaksic et, al. Vol. 20, No. 3/4 Table 1. Prey of Great Horned Owls in La Dehesa (33° S), Puerto Ibanez (46° S), and Torres del Paine (51° S), Chile. Figures are percentages by number of prey individuals; subtotals are in brackets. Prey Categories WEIGHT(g)* 33°S 46° S 51°S Mammalia [88.6] [86.0] [87.5] Lagomorpha Lepus capensis 2.000.0 — 5.3 0.6 Oryctolagus cuniculus 1,300.0 15.8 — — Marsupialia Marmosa elegans 40.0 3.5 — — Rodentia Abrocoma bennetti 219.0 18.4 — — Akodon lanosus 32.5 — — 4.8 Akodon longipilis 76/41.0** 16.7 8.7 3.0 Akodon olivaceus 40.0 0.8 — — Akodon xanthorhinus 21.5 — 5.3 9.5 Ctenomys cf. magellanicus 271.8 — 15.8 — Eligmodontia typus 26.5*** — — 0.6 Euneomys chinchilloides 87.8*** — 26.3 0.6 Notiomys macronyx — — — 2.4 Oryzomys longicaudatus 45/29.8** 4.4 1.8 39.8 Phyllotis darwini 66.0 4.4 7.0 — Phyllotis micropus 75.0 — 12.3 — Phyllotis sp. — — 3.5 — Rattus rattus 158.0 19.3 — — Reithrodon physodes 81.7 — — 25.6 Unidentified — 5.3 — 0.6 Aves [11.4] [5.3] [2.4] Unidentified — 11.4 5.3 2.4 Insecta [0.0] [8.7] [10.1] Coleoptera — — 8.7 8.9 Hymenoptera — — — 0.6 Orthoptera — — — — Unidentified — — — 0.6 No. pellets 98.00 55? 139. 0C No. prey 114.00 57.00 168. 0C Geometric mean prey weight (g) 181.9 104.5 41.1 Twice standard error 0.61 0.83 0.31 Sample size (= prey with weight) 95.00 47.00 142. 0C Diet breadth (Bobs) 6.90 7.18 4.07 Standardized diet breadth (Bsta) 0.66 0.62 0.24 * Weights with no decimal places are from Jaksic and Marti (1984); all the remaining (except for those marked with asterisks) are from Jaksic et al. (1983). **There is a strong latitudinal cline in body size for this species (see Yanez et al. 1978, and Palacios 1982): the first figure corresponds to its mean weight in central Chile; the second, to its mean in southern Chile. ***From Greer (1965). Fall/Winter 1986 Ecology of South American Owls 115 suggested that the absence of these 2 species from the Great Horned Owl diet was due to their diur- nal-crepuscular activity pattern. In Puerto Ibanez, owls preyed on essentially all small mammal species trapped by Reise and Venegas (1974) in the same locality, and on 2 additional rodents: Euneomys chinchilloides (a scansorial species) and Ctenomys cf. magellanicus (a fossorial one). These 2 made up more than 40% of the owls’ diet (Table 1), but were neither trapped nor seen in the area (Reise and Venegas 1974; Yanez et al. in press). In Torres del Paine, owls preyed on all small mammal species known to occur there, as well as on 3 other rodents hitherto not recorded (Rau et al. 1978): the terres- trial Eligmodontia typus and Akodon lanosus, and the semi-fossorial Notiomys macronyx. In general, the three owl populations studied preyed mainly on small mammals (averaging 87% of their prey). With increasing latitude, the proportion of birds in the diet decreased, with the opposite trend seen in the insect prey (from no insect consumption at all in La Dehesa, to 10% of the diet in Torres del Paine). The geometric mean weight of prey declined monotonically from north to south, with no indica- tion of a corresponding trend in owl body size (Johnson 1965; Humphrey et al. 1970). A similar (but not so consistent) decline in mean prey weight away from the equator was reported by Knight and Jackman (1984) for Great Horned Owls along the Pacific coast of the United States. Comparing areas at latitudes 30° to 40? between the two hemispheres, Jaksic and Marti ( 1 984) showed that central Chilean and California Great Horned Owls did not differ significantly in body size (1,227 g vs. 1,166 g, re- spectively), but mean prey weight of California owls was 59% of Chilean ones. Knight and Jackman ( 1 984) reported mean prey weight of Great Horned Owls in central Washington (46° N), which coin- cides with the latitude of Puerto Ibanez. Because Knight and Jackman (1984) used an arithmetic es- timate of mean prey weight, we recalculated from their raw data the geometric estimate, thus making their results comparable to ours. Washington owls exhibited a geometric mean prey weight of 22.9 ± 0.21 g (mean ± 2 s.e.; sample size = 872) which amounted to only 22% of the value reported for southern Chilean owls at the equivalent latitude (Table 1). It is difficult to assign causal relations to these patterns without knowing prey sizes available to owls in these different localities. Knight and Jackman (1985), following Herrera and Hiraldo (1976), speculated that the decrease in mean prey weight taken by owls at higher latitudes may be related to smaller prey becoming more abundant as latitude increases. We have no data to substantiate this claim. Diet breadth in Chile also decreased with in- creasing latitude, in agreement with trends re- ported by Knight and Jackman (1984) for the Great Horned Owl along the Pacific coast of the United States and by Herrera and Hiraldo (1976) for the Eagle Owl (Bubo bubo) in Europe. Jaksic and Marti (1984) reported that central Chilean and California Great Horned Owls have a similar diet breadth at the class level of prey identification (H’NGG in their Table 3), but that the former have significantly narrower diet breadth at the species level of mam- malian prey H’NM in their Table 3). Knight and Jackman (1984) documented a diet breadth of 4.12 (which amounts to a standardized diet breadth = 0.12; because Bmax = 26, and Bmin= 1) for Washington Great Horned Owls. These values amount to 57% and 19% (respectively) of those computed for owls at the equivalent latitude in Chile, and are in fact more similar to observations 5 latitudinal degrees south, in Torres del Paine (Ta- ble 1). Apparently, both South and North Ameri- can Great Horned Owls exhibit narrower diets to- ward higher latitudes, but the latter prey heavily on relatively few items. In fact, only two rodents (Thomomys talpoides and Perognathus parvus ) ac- counted for 73% of the items in the diet of Washington owls. A similar value in the diet of Chilean owls was accounted for by the six most preyed upon rodent species in Puerto Ibanez, and by three in Torres del Paine (Table 1). The de- creasing diet diversity away from the equator might be related to a decreasing number of potential prey species which is consistent in both hemispheres. Latitudinal trends in the trophic niche of Great Horned Owls along the Pacific coast of southern South America closely mirror trends documented in northern North America (and of the congeneric Eagle Owl in Europe). Local estimates of trophic statistics for latitudinally-matched localities in the two hemispheres, however, show some marked dif- ferences. The pattern of decreasing diet diversity away from the equator could have been expected, but 'a similar trend in mean prey weights at corres- ponding latitudes, both related to the local availa- bility/vulnerability of prey, was unlikely to hold within/between the two hemispheres. 116 Jaksic et. al. Vol. 20, No. 3/4 Acknowledgments We thank Richard J. Clark, Richard L. Knight, M. Ross Lein, Carl D. Marti, Martin K. McNicholl, Karen Steenhof, and an anonymous reviewer, for critically reading different versions of this paper. Jaksic acknowledges the support of grants DIUC 202/83 and 076/85 (awarded by the Pontificia Universidad Catolica de Chile), and INT-8308032 (awarded by the U.S. Na- tional Science Foundation) during the several stages of prepara- tion of the manuscript. Literature Cited Burton, J. A. [ed.]. Owls of the World, E.P. Dutton, New York. Colwell, R.K., and D.J. Futuyma. 1971. On the mea- surement of niche breadth and overlap. Ecology 52:567-576. Greer, J.K. 1965. Mammals of Malleco province, Chile. Publ. Mus., Michigan State Univ., Biol. Ser. 3:49-152. Herrera, C.M., and F. Hiraldo. 1976. Food-niche and trophic relationships among European owls. Ornis Scand. 7:29-41. Humphrey, P.S., D. Bridge, P.W. Reynolds, and R.T. Peterson. 1970. Birds of Isla Grande (Tierra del Fuego). Preliminary Smithsonian Manual, Smithso- nian Institution, Washington, D.C. Jaksic, F.M. and H.E. Braker. 1983. Food-niche re- lationships and guild structure of diurnal birds of prey: competition versus opportunism. Can. J. Zool. 61:2230-2241. Jaksic, and J. L. Yanez. 1980. Differential utilization of prey resources by Great Horned Owls and Barn Owls in central Chile. Awk 97:895-896. Jaksic, F.M., H.W. Greene, and J.L. Yanez. 1981. The guild structure of a community of predatory verteb- rates in central Chile. Oecologia 49:21-28. Jaksic, F.M., J. Rau, andJ. Yanez. 1978. Ofertade presas y predacion por Bubo virginianus (Strigidae) en el Par- que Nacional “Torres del paine.” anales del Instituto de la Patagonia, Punta Arnas (Chile) 9:199-202. jAksic, F.M., and C.D. Marti. 1984. Comparative food- habits of Bubo owls in Mediterranean-type ecosystems. Condor 86:288-296. Jaksic F.M., J.L. Yanez, and J.R. Rau. 1983. Trophic relations of the southernmost populations of Dusicyon in Chile./. Mamm. 64:693-697. Johnson, A.W., 1965. The birds of Chile and adjacent regions of Argentina, Bolivia and Peru: volume II. Platt Establecimientos Graficos, Buenos Aires. Knight, R.L., and R.E. Jackman. 1984. Food-niche re- lationships between Great Horned Owls and Common Barn-Owls in eastern Washington. Auk 101:175-179. Levins, R. 1968. Evolution in changing environments: some theoretical explorations. Princeton Univ. Press, Princeton, New Jersey. Palacios, O.V. 1982. Morfometria y sistematica de Oryzomys longicaudatus (Rodentia: Cricetidae). Thesis, Universidad de Chile, Santiago, 91 pp. Rau, J.,J. Yanez and F. Jaksic. 1978. Confirmacion de Notiomys macronyx alleni O. y Eligmodontia typus typusQ.,y primer registro de Akodon ( Abrothrix ) lanosus T. (Rodentia: Cricetidae) en la zona de Ultima Esperanza (XII Region, Magallanes). Anales del Instituto de la Patagonia, Punta Arenas (Chile) 9:203-204. Reise, D., and W. Venegas. 1974. Observaciones sobre el comportamiento de la fauna de micromamiiferos en la region de Puerto Ibanez (Lago General Carrera), Aysen, Chile. Boletin de la Sociedad de biologia de Concepcion (Chile) 47:71-85. Yanez, J., W. Sielfeld, J. Valencia, and F. Jaksic. 1978. Relaciones entre la sistematica y la morfometria del subgenero Abrothrix (Rodentia: Cricetidae) en Chile. Anales del Instituto de la Patagonia, Punta Arenas (Chile) 9:185-197. Yanez, J.L., J.C. Torres-Mura, J.R. Rau, and L.C. Con- treras. In press. New record and current status of Euneomys (Cricetidae) in southern South America. Fieldiana (Zoology). Departamento de Biologia Ambiental, Universidad Catolica de Chile, Casilla 1 14-D Santiago, Chile. Address of second au- thor: Museo Nacional de Historia Natural, Casilla 787 San- tiago, Chile. Address of third author; Estacion Biologica de Donana, Apartado 1056, 41080 Sevilla, Spain. Received 28 March 1986; Accepted 25 July 1986. IMPACT OF A HIGH-VOLTAGE TRANSMISSION LINE ON A NESTING PAIR OF SOUTHERN BALD EAGLES IN SOUTHEAST LOUISIANA David A. Dell and Phillip J. Zwank Abstract — To evaluate the impact of a 500th kv power transmission line on a pair of nesting bald eagles. (Haliaeetus leucocephalus) pre- and post-installation observations of eagle area-use were recorded. The mean of the daily proportion of eagle activity spent in the vicinity of the powerline decreased (P = 0.02) from pre-installa- tion ( X = 27.6%) to post-installation (X= 18.7%) seasons, indicating that activity patterns were changed after installation of the powerline. No serious physical threat to nesting eagles could be ascertained. The eagles regularly flew over and under the powerline, and perched and foraged near it. They never used the powerline itself for perching. Wilcox (1979) reported on the success of a pair of Southern Bald Eagles (Haliaeetus leucocephalus leucocephalus ) nesting 50 m from a 240th kv power line, however, quantitative data are unavailable on the effects of power transmission lines on territory use by nesting Southern Bald Eagles. The con- struction of a transmission line through the nesting territory of a pair of eagles in southeast Louisiana provided an opportunity to compare area-use by the eagles within the powerline zone before and after construction. Study Area and Methods The Waterford-Churchill 500-kV line passes through Salvador Wildlife Management Area (SWMA), St. Charles Parish, Louisiana, at the northwest shore of Lake Cataouatche, approxi- mately 14 km south of New Orleans International Airport. The line consists of steel self-supporting towers of an “H” design. Each tower is 30.5 m tall and supports 3 phase conductors 9.6 m apart. The conductors vary from 11 to 21 m above marsh level. Two smaller static lines are strung approximately 9 m above the phase conductors. Distances between towers vary, but they are 265-274 m apart in the study area. The powerline is approximately 600 m north of the eagle nest studied and centered in a corridor ap- proximately 60 m wide that has been cleared of all trees. Con- struction occurred during summer (when eagles are absent from SWMA) 1983. The eagle nest is in a living bald cypress (Taxodium distichum), 32.9 m high and 107.4 cm in diameter above the swelling at the base (Dugoni 1980). An observation blind was placed approxi- mately 320 m north of the nest during 1983-84, between the nest tree and powerline. In 1979-80, the blind was approximately 100 m closer to the nest (Fig. 1). From the blind, we could observe eagles flying over an area of about 810 ha. This area was a non- tidal, permanently flooded, palustrine system (Cowardin et al. 1979) occupied by forested wetland (cypress and Nyssa aquatica), aquatic bed (Bidens laevis, Eleocham spp., and Sagittaria lancifolia on floating turf; Nelumbo lutea and Eichhornia crassipes were free- floating), and unconsolidated organic bottom habitats. We observed eagles twice weekly from dawn to dusk and re- corded total minutes spent in various activities and areas. To analyze the effect of the powerline on the eagles’ area-use, a “powerline zone” extending 400 m to the south and up to 1000 m north of the powerline was defined within the study area. The boundaries of the zone were chosen to include the perch trees and foraging areas close to the powerline, and because the eagles had to cross the powerline to reach the most frequently-used foraging area visible from the blind. The proportion of “eagle-minutes” (combined number of minutes that both adults were observed) spent within the powerline zone each day was used as the depen- dent variable in a randomized-block design analysis of variance to test for differences between pre-and post-installation seasons (treatments) and among periods of the nesting season (blocks). The periods of the nesting season we blocked on were brooding, pre-fledging (eaglets still in nest, but not brooded), and post- fledging (eaglets out of nest). Results and Discussion Pre-installation observations were conducted from 3 January 29 to April 1980. During that sea- son, 25 observation days were completed and 30,651 eagle-minutes were recorded (Shealy and Zwank 1981). Due to a lawsuit, construction of the powerline was delayed until summer 1983. Post-in- stallation observations were for 4 January to 3 May 1984. Thirty-two observation days were completed, and 45,784 eagle-minutes were recorded. The mean of the daily proportion of eagle- minutes spent by the adult eagles in the powerline zone decreased (P = 0.02) from pre-installation (x = 27.6x) to post-installation (x=18.7%). Also, activity varied among the periods of the nesting season (P = 0.0004) (Table 1). The eagles spent more of the brooding and pre- fledging periods in the powerline zone before in- stallation than after. In the post-fledging period during both years, the eagles spent almost the same 117 Raptor Research Vol. 20 (3/4): 1 17-1 118 Dell and Zwank Vol. 20, No. 3/4 Area visible from eagle observation blind, Salvador WMA Figure 1. proportion of time in the powerline area. Activities within the powerline zone consisted of perching, soaring, foraging, or straight-line flight between perches. The eagles often flew over and under the conductors while going between the nest and various foraging areas. Herrick (1924) re- ported that one nesting pair of eagles regularly flew past “wires by the railroad.” We saw an eagle react to the powerline only once. While flying in circles 20-40 m above the marsh, an adult approached the wires several times, then banked quickly to avoid them. None of the eagles were ever seen perching on the transmission lines or towers. Relocation of the observation blind in 1983 closer to a perch tree appeared to affect behavior. Use of this perch tree accounted for 1.3% of total activity in the pre-installation season (Shealy and Zwank 1981), but was never used during the post-installa- tion season. Table 1. Average daily proportion of eagle-minutes spent in the powerline zone in 1979-80 and 1983-84 and averages by nest period. Period X SE N a cv% 1979-80 0.276 0.0350 25 63.3 Post-fledging 0.400 0.2006 4 100.3 Pre-fledging 0.254 0.251 14 37.0 Brooding 0.250 0.0386 7 40.9 1983-84 0.187 0.0418 32 126.7 Post-fledging 0.411 0.0806 11 65.0 Pre-fledging 0.082 0.0315 13 139.0 Brooding 0.048 0.0151 8 88.4 Observation days. Fall/ Winter 1986 Powerline Impacts on Bald Eagles 119 Changes in area-use observed may have resulted from removal of potential perch trees from the powerline corridor, blocking of forage flights by transmission wires or changes in prey availability or distribution. Replacement of one or both members of the adult pair could also have influenced be- havior; we cannot be certain that the same pair nested in both 1980 and 1983. Also, relocation of the observation blind changed perching habits, but its influence on use of the powerline zone could not be determined. Based on our observations of eagles during flight, we do not think the powerline poses a serious physical threat to the nesting adults. Also, nesting attempts were successful before and after power- line installation. Possibly, however, awkward fledglings could collide with the powerline. Eagle electrocutions are unlikely because phase conduc- tors are widely spaced (9.6 m) and we never ob- served perching on the powerline or towers. Funding This article is a contribution from the La. Coop. Wildl. Res. Unit (USFWS, La. Dept. Wildl. and Fisheries, La. State Univ., Wildl. Manage. Inst., and School of Forestry, Wildl., and Fisheries, LSU cooperating). Funding was provided by Louisiana Power and Light Co. Literature Cited Cowardin, L.M., V. Carter, F.L. Golet, and E.T. LaRoe. 1979. Classification of wetlands and deep- water habitats of the United States. U.S. Fish and Wildl. Serv., Off. Biol. Serv. 103 pp. Dugoni, J.A. 1980. Habitat utilization, food habits, and productivity of nesting southern bald eagles in Louisiana. M.S. Thesis, Louisiana State Univ., Baton Rouge. 151 pp. Herrick, F.H. 1924. The daily life ofthe American eagle: late phase. Auk 41:389-422. Shealy, P.M. and P J. Zwank. 1981. Activity patterns and habitat use of a nesting pair of southern bald eagles in southern Louisiana. Pages 127-135. In: R.R. Odom and J. W. Guthrie, eds. Proc. nongame and en- dangered wild, symp., Georgia Dept. Nat. Res., Game and Fish Div. Tech. Bull. WL5. Wilcox, J.R. 1979. Florida power and light company and endangered species: examples of coexistence. Pages 451-454. In G.A. Swanson, tech, coord. The mitiga- tion sym.: a national workshop on mitigating losses of fish and wild, habitats. U.S. For. Serv., Rocky Mt. Forest and Range Exp. Sta. Gen. Tech. Rep. RM-65. Louisiana Cooperative fish and wildlife Research Unit, School of Forestry, wildlife and fisheries, Louisiana State Uni- versity, Baton Rouge, Louisiana 70803 Received 15 December 85; Accepted 15 June 1986. Third New England Regional Hawk Conference - The New England Hawk Migration Committee wishes to announce the Third New England Regional Hawk Conference will be held 4 April 1987 at the Holiday Inn, Holyoake, Massachusettes. Registration forms are available from HAWKS, P.O. Box 212, Portland, Connecticut 06480. There are special rates available for lodging at the Conference center. Registration will be limited. FOOD OF THE BOOTED EAGLE (HIERAAETUS PENNATUS) IN CENTRAL SPAIN Jose P. Veiga Abstract. — The identification of 202 prey remains of the Booted Eagle (Hieraaetus pennatus) shows that mammals (41.6% of prey items identified), birds (36.6%) and reptiles (21.8%) are important prey in Central Spain. Most mammals captured were young rabbits, and the majority of the bird prey were fledglings or juveniles. Lizards were adult or subadult individuals. Over 90% of the prey captured weighed between 27 and 243 g. Little is known about the biology of the Booted Eagle (Hieraaetus pennatus), as it occurs in countries with little ornithological activity. Most published accounts of food habits are single enumerations of prey remains recorded mainly during sporadic vis- its to nests (Val verde 1967; Araujo 1973; Garzon 1973; Iribarren 1975). This procedure provides an inaccurate picture of diet, since prey that are large and leave persistent remains are over-represented in samples (e.g., Valverde 1967; Delibes 1975). In spite of this, several recent papers dealing with the trophic relationships between members of various raptor communities have made use of such data (Jaksic and Soriguer 1981; Jaksic 1983; Jaksic and Braker 1983). In my opinion this has led to errone- ous conclusions regarding the ecological position of the Booted Eagle in Mediterranean environments. The present paper presents more accurate infor- mation about the diet of this raptor, obtained using a more systematic data collection procedure. I also take into consideration some attributes of prey, such as size and age, that have been overlooked. Study Area and Methods This study was carried out in 3 areas, each about 35 km^ in size, located on the northern slope of the Sierra de Guadarrama mountains (4(f 35' -4(f 60' N, OP 5'-(f 60' W). Area 1 is about 60% pasture interspersed with thick scrub. The only arboreal forma- tions present are 3 small pine groves of between 1 and 5 ha. Area 2 is 1 0 km away and about 40% covered with mature natural pine trees (Pinus silvestris) over 15m tall. The rest of area 2 is made up of a sparse evergreen oak grove ( Quercus rotundifolia) with extensive clearings in v/hich low scrub mixes with pasture land. Area 3, 15 km from area 2 and 30 km from area 1 , is similar to area 1 in that it has only 2 arboreal formations, one of 2 ha and the other of 25 ha. Area 1 was visited from 1 978 to 1 98 1 . One pair of Booted Eagles used the same nest year after year. Area 2 was also visited from 1979 to 1981. In 1979 2 pairs of nesting eagles were present, but in 1980 to 1981 no nests were found. Area 3 was also visited from 1979 to 1981. In both 1979 and 1980 1 pair of eagles was located, but no eagles were seen in 1981. Visits were made approximately every 15 d from shortly before incubation (mid-late April) until after the young left the nest (mid-late August). During the feeding period nests were occasionally visited every 7 d. Pellets and prey remains were sought in and around nests and below perches which were usually within a 200 m radius of the nests. A total of 110 pellets, containing 130 identifiable prey items, and 72 prey remains were collected. Each species found in any one pellet was counted as 1 individual unless it was possible to show that more than 1 was represented. Therefore, it was necessary to count pieces of remains such as nails, beaks, teeth, etc. Weight and approximate age of the prey were estimated by comparing re- mains with material from zoological collections and with speci- mens collected in the study areas. In order to establish a frequency distribution for prey, weight classes were established whose limits followed a geometric progression (Fig. 1). This insured that the resulting distribution would be more or less normal (Schoener 1 969; Hespenheide 1971). Only some prey identified in the pellets could be assigned to one of the established weight categories, particularly in the case of species, like rabbits and ocellated lizards, whose weights vary a great deal. Results and Discussion Mammals, birds, and reptiles, in decreasing or- der of capture frequency, comprised the diet of the Booted Eagle in the study area. Percentage differ- ences of these taxa in the diet increased considera- bly when biomass was taken into account. (Table 1). Among mammal prey, rabbits were the most im- portant prey species. Birds captured were primarily species that forage on the ground. The Ocellated Lizard (Lacerta lepida) was the only reptile prey, although other lizards are common in the study area. The weight of prey items varied between 1 0 and 800 g. However, most were in the 27 to 243 g range (Fig. 1). A major part of the diet consisted of prey in the 81 to 243 g weight-class (Fig. 1). Prey-size dis- tributions do not appear to be the same for the 3 taxa present in the diet: most mammal and lizard prey weighed between 8 1 and 243 g. Avian prey was 120 Raptor Research Vol. 20 (3/4): 120-123 Fall/Winter 1986 Booted Eagle Diet 121 Figure 1 . Diet of the Booted Eagle. Thick line histogram: percent of the total biomass supplied by the prey-items; thin line histogram: percent of the total number of prey-items. Sample size = 165. predominantly between 27 and 81 g (Fig. 2). The majority of birds in this class were the Spotless Starling, (Sturnus unicolor ) weighing 70 g. Nearly all rabbits captured were very young individuals. Of 27 bird prey items of known age, the number of fledgling and juveniles was greater than the number of adults (22 young vs. 5 adults). All Ocel- lated Lizards identified were adults or sub-adults. Prey-size distribution could merely reflect the size distribution of available prey, assuming Booted Eagles on the study area selected prey randomly with respect to size. Nevertheless, the lack of insects, amphibians, and small reptiles in the diet of some other raptors of similar size such as the Common Buzzard (Buteo buteo ), Black Kite (Milvus migrans) and Red Kite (M. milvus) in the same study area (Veiga 1982) suggests that prey below a certain weight were avoided. Prey might also be selected according to age and experience. This may be par- ticularly true for avian prey, since the poor flying abilities of young birds make this age class more vulnerable to predation by Booted Eagles. It has been reported that the analysis of pellets and prey remains for Order Falconiformes tends to underestimate the amount of some prey while overestimating others (Valverde 1967; Delibes 1975; Collopy 1983). The absence of small prey such as insects, amphibians or small reptiles in the Booted Eagles’ diet could be due to these methodological biases. However, using the same methodology, these small prey have been found in the diet of other similar sized raptors in the same areas in which the Booted Eagle was studied, Fur- thermore, by sampling prey remains regularly and at relatively short intervals the potential bias possi- bly caused by the greater detectability of certain prey when collected at longer intervals would be diminished. The fact that the material to be analyzed was collected from the nests as well as from the perches of the adults reduces the possibil- ity of obtaining a distorted image of diet if it is assumed that food taken to the nestlings is different from that of the adults. I have not been able to demonstrate this in the Booted Eagle. Earlier studies of Booted Eagle feeding habits carried out in the Palearctic and in South Africa describe them as a hunter of small birds and, to a lesser degree, lizards (Valverde 1967; Araujo 1973; Garzon 1973; Iribarren 1975; Steyn and Grobler 1981). It is worth noting that although the scarcity of mammals in the South African Booted Eagles’ diet could be due to a lack of appropriate sized individuals in the field, the low representation of this taxon in reports from Spain where rabbits abound in a variety of sizes is surprising. My results suggest that the Booted Eagle behaves, in my study Figure 2. Distribution of the prey remains in the prey- weight classes in each taxonomic group., Black circles = mammals; open circles = birds; squares = reptiles. Sample sizes: mammals = 61; birds = 64; reptiles = 40. 122 Jose P. Veiga Vol. 20, No. 3/4 Table 1. Prey of the Booted Eagle in central Spain. Species Number Occurrence Biomass of Items Percent Percent Reptiles Ocellated Lizard (Lacerta lepida) 44 21.8 14.3 Total 44 21.8 14.3 Birds Common Kestrel (Falco tinnunculus) 2 0.99 1.0 Quail ( Cotumix cotumix) 2 0.99 0.46 Unidentified Phasianidae 1 0.49 0.46 Little Bustard ( Otis tetrax) 1 0.49 1.76 Stone Curlew (Burhinus oedicnemus) 1 0.49 1.05 Wood Pigeon ( Columbia palumbus ) 1 0.49 1.08 Unidentified Columbidae 3 1.48 2.54 Swift ( Apus apus) 1 0.49 0.09 Hoopoe (Upupa epops) 7 3.46 1.03 Green Woodpecker (Picus viridis) 1 0.49 0.39 Unidentified Alaudidae 1 0.49 0.08 Mistle Thrush (Turdus viscivorus) 1 0.49 0.27 Spotless Starling (Stumus unicolor) 28 13.86 5.13 Magpie (Pica pica) 9 4.45 4.33 Jackdaw ( Corvus monedula) 4 1.98 2.1 Carrion Crow ( Corvus corone) 1 0.49 1.18 Unidentified 10 4,95 1.83 Total 74 36.6 24.8 Mammals Common White-toothed Shrew (Crocidura russula) 1 0.49 0.03 Blind Mole ( Talpa caeca) 2 0.99 0.19 Rabbit ( Oryctolagus cuniculus) 65 32.18 48.71 Hare (Lepus granatensis) 2 0.99 5.0 Unidentified Lagomorpha 1 0.49 0.75 Water Vole (Arvicola sapidus) 8 3.96 5.0 Weasel (Mustela nivalis) 3 1.48 0.94 Unidentified 2 0.99 0.19 Total 84 41.6 60.8 Total Items 202 Fall/Winter 1986 Booted Eagle Diet 123 area, like a taxa-generalist that concentrates on ter- restrial prey weighing between 70 and 240 g. It is probable that the general decrease of the rabbit in Iberian ecosystems in the last decades, resulting from the effect of mixomatosis, has influenced the composition of the Booted Eagle’s diet. However, there are no detailed studies of the population dynamics of the rabbit and other prey species, which would be necessary before this could be seri- ously discussed. Acknowledgments I am indebted to G. Bortolotti, C. Griffin and B. Millsap for their critical comments of an earlier draft. Literature Cited Araujo, J. 1973. Falconiformes del Guadarrama sur- occidental. Ardeola 19: 257-278. Collopy, M.W. 1983. A comparison of direct observa- tions and collections of prey remains in determining the diet of Golden Eagles./. Wildl. Mange. 47: 360-368. Delibes, M. 1975. Alimentacion del Milao Negro (Milvus migrans) en Donona (Huelva, Espana). Ardeola 21: 183- 207. Garzon, J. 1973. Contribucion al estudio del status, alimentacion y proteccion de las falconiformes en Es- pana central. Ardeola 19: 279-330.’ Hespenheide, H.A. 1971. Food preference and the ex- tent of overlap in some insectivorous birds, with special reference to the Tyrannidae. Ibis 113: 59-72. Iribarren, J.J. 1975. Biologla del aguila calzada (Hieraaetus pennatus) durante el periodo de nidifica- cion en Navarra. Ardeola 25 (Vol. Esp.): 305-320. Jaksic, F.M. 1983. The trophic structure of sympatric assemblages of diurnal and nocturnal birds of prey. Amer. Mid. Natur. 109: 152-162. Jaksic, F.M. and Braker, H.E. 1983. Food niche re- lationships and guild structure of diurnal birds of prey: competition versus opportunism. Can. f. Zool. 61:2230-2241. Jaksic, F.M. and Soriguer, R.C. 1981. Predation upon the European rabbit (Oryctolagus cuniculus) in Mediter- ranean habitats of Chile and Spain: a comparative analysis./. Anim. Ecol. 50: 269-281. Schoener, TW. 1969. Models of optimal size for solitary predators. Am. Nat. 103:277-313. Steyn, P. and Grobler, J.H. 1981. Breeding biology of the Booted Eagle in South Africa. Ostrich 52: 108-118. Valverde, J.A. 1967. Estructura de una comunidad mediterranea de vertebrados terrestres. C.S.I.C. Mad- rid. Veiga, J.P. 1982. Ecologia de las rapaces de un ecosistema Mediterraneo de montana. Aproximacion a su estructura comunitaria. Ed. Univ. Compl., Mad- rid. Museo Nacional de Ciencias Naturales C.S.I.C. Jose Gutierrez Abascal, 2. Madrid-28006. SPAIN. Received 30 March 1985; Accepted 8 April 1986. FOODS OF NESTING BALD EAGLES IN LOUISIANA Joseph A. Dugoni, Phillip J. Zwank, and Gary C. Furman Abstract — During the summer of 1979, remains of 243 vertebrates comprising 3 1 species were collected from 10 nests that had fledged young during the previous spring to determine the food habits of nesting Bald Eagles {Haliaeetus leucocephalus) in Louisiana. American Coots (Fulica americana) and freshwater catfish ( Ictalurus spp.) were the most abundant species, but fish probably constituted a greater portion of the diet than results indicate, due to more complete digestibility of piscian skeltons. The Bald Eagle (Haliaeetus leucocephalus) nests in swamps of southcentral and southeastern Louisiana. Portions of this habitat are being lost or altered due to drainage, channelization conversion of land to agriculture, and industrial development (Yancey 1970). Loss of swamp habitat may harm nesting eagles by reducing the availability or abun- dance of prey. Support for this hypothesis is pro- vided by McEwan (1977) who found that Bald Eagles in Florida rely primarily on fish and wetland birds for food. Foods of nesting Bald Eagles in Louisiana have not been previously documented. Study Area and Methods Fieldwork was conducted in coastal southeastern and south- central Louisiana, including Terrebone, Jefferson, St. Charles, St. Tammany, and Assumption Parishes. Climate is subtropical maritime. Wetlands of 0-2 m elevation predominate; relief is pro- vided by levees and spoilbanks. Much of the region consists of permanently or annually flooded baldcypress (! Taxodium distichum) - tupelogum ( Nyssa aquatica) forests. Dominant land uses include gas and oil production and industrial development, as well as hunting, fishing and trapping. Area vegetation and other charac- teristics are further described by Bahr et. al, (1983) and Chabreck and Condrey (1979). Bald Eagle nest locations were determined in 1977 and 1978 by interviews with private citizens and by using helicopter surveys. In June and July 1979, immediately following fledging of young and seasonal departure of parents, prey remains were collected from 9 nests. Additional remains were collected in July from a nest after it was downed by a hurricane. To ensure as much as possible that prey remains were those left by 1979 nesters, we collected only those remains on or near the nest surface immediately after eagles vacated the nest, prior to possible nest use by other species. Results Prey species of nesting Bald Eagles were deter- mined from remains found in 10 nests during the summer of 1979. We collected remains of 243 ver- tebrates, including 4 classes and 31 species (Table 1). Birds comprised the highest percentage of prey animals (42.4%), followed by fish (41.5%), mam- mals (15.7%), and a reptile (0.4%). American Coots (Fulica americana) comprised 40 (47.6%), of the 103 birds, while freshwater catfish (Ictalurus sup.) ac- counted for 53 (52%) of 101 fish. Muskrat (Ondatra zihethicus) and Nutria( (Myocastor coypus) combined comprised 82.2% of mammals, and the reptile re- mains were those of a Mud Turtle (Kinosternon sub- rubrum). Discussion Remains of 31 vertebrate prey species may sup- port claims that Bald Eagles are opportunistic feed- ers (Retfalvi 1970; Todd et. al. 1982; fielder 1982). However, American Coots and catfish made up nearly 42% of prey animals, indicating that a pre- ference for these species may exist. Our findings agree with those of McEwan (1977), who found that American Coots and catfish comprised the major portions of the diet of Bald Eagles in Florida. Fiel- der (1982) reported that American Coots were the major prey animal of Bald Eagles at a study site in Washington, but concluded that availability of prey dictated usage. Haywood and Ohmart (1986) found in Arizona that, while catfish and other benthic-feeding fish comprised the majority of prey, American Coots were the major avian prey of Bald Eagles. Benthic fish are common prey proba- bly because of their high vulnerability to aerial pre- dators (Todd et. al. 1982). Bald Eagle consumption of benthic fish, American Coots, and dabbling waterfowl makes obvious the importance of shallow wetlands within foraging distance of nest sites. Be- cause of this importance, proposals to alter such wetlands should be carefully studied. A bias toward nonfish prey species probably exists in our study, because fish skeletal parts can be more completely digested than those of other ver- tebrates (Todd et. al. 1982). For instance, although we observed over 20 Gizzard Shad (Dorosoma cepedianum) brought to nests and consumed, the remains of only 2 were recovered. 124 Raptor Research Vol. 20 (3/4); 124-127 Table 1. Species identified from remains collected from 10 Louisiana Bald Eagle nests after the 1978-1979 nesting season. Fall/ Winter 1986 Foods of Bald Eagles 125 & u T}H o . 00 d Tf GM CM CM LTO UTO m CD QO O d <3 ■'t x> 3 W5 d TOP d d QO o Class Aves Class Osteichthyes Class Reptilia 126 Dugoni et. al. Vol. 20, No. 3/4 o 2 CJ '0 C/3 d 2 d d o 2 o 2 Fall/ Winter 1986 Foods of Bald Eagles 127 Acknowledgments Contribution of the Louisiana Cooperative Fish and Wildlife Research Unit; Louisiana State University, U.S. Fish and Wildlife Service, Louisiana Department of Wildlife and Fisheries, and Wildlife Management Institute, cooperating. We wish to thank Dr. John V. Conner, Professor, Louisiana State University, and Dr. Royal B. Suttkus, Director, Tulane University Museum of Natural History, for help in identification of specimens. We thank Mr. John D. Newsom, Leader, Louisiana Cooperative Wildlife Unit, retired, for guidance during the early stages of this study, and thank the U.S. Army Corps of Engineers for financial support. Literature Cited Bahr, L.M. Jr., R. Costanza, J.W. Day Jr., S.E. Bailey, C. Neill, S. G. Leibowitz and J. Fruci. 1983. Ecologic characterization of the Mississippi Deltaic Plain region; a narrative with management recom- mendations, FWS/OBS-82/69. 189pp. Chabreck, R.H., and R.E. Condrey. 1979. Common vascular plants of the Louisiana marsh. La. State Univ. Center for Wetland Res. Sea Grant publ. LSU-T-79- 003. 116pp. Fielder, P.C. 1982. Food habits of Bald Eagles along the mid-Columbia River, Washington. Murrelet 63:46-50. Haywood, D.D., and R.O.Ohmart. 1986. Utilization of benthic-feeding fish by inland breeding Bald Eagles. Condor 88:35-42. McEwan, L.C. 1977. Nest site selection and the produc- tivity of the Southern Bald Eagle. MlS. Thesis, Univ. of Florida, Gainesville. 63pp. Retfalvi, L.I. 1970. Food of nesting Bald Eagles on San Juan Island, Washington. Condor 72:358-361. Todd, C.S., L.S. Young, R.B. Todd, C.S., L.S. Young, R.B. Owen, J.R., and F.J. Gramlich. 1982. Yancey, R.K. 1970. Our vanishing delta hardwoods. La. Conserv. 22:30-36. Address of first and second authors: Louisiana Cooperative Wildlife Research Unit, Louisiana State University, Baton Rouge, Louisiana 70803. Address of third author: School of Forstry, Wildlife and Fisheries, Louisiana State University, Bat- ron Route, Louisiana 70803. Received 1 February 1986; Accepted 31 October 1986. MALE FOOD PROVISIONING AND FEMALE REPRODUCTION IN AMERICAN KESTRELS Timothy J. Coonan While the effects of male raptor nest provisioning on clutch quality have been documented (Drent and Daan 1980; Wink et. al. 1980), the effect of provisioning on later nest success is less well estab- lished. Male provisioning ability should affect hatching and fledging success, since the female and young of many raptor species depend to a degree on the male for food delivery until fledging (Bal- gooyen 1976; Snyder and Wiley 1976; Newton 1978; Mueller et. al 1981; Rudolph 1982; and Vil- lage 1983). The purpose of this study is to document the relationship between male provisioning perfor- mance and pair reproductive success, beyond clutch size, in the American Kestrel (Falco spar- verius). Effects of differential male provisioning performance should be seen in number of young hatched and number of young fledged from each nest. MATERIALS AND METHODS Six kestrel pairs in wooden nestboxes (Gary and Morris 1980) were observed from the pre-hatching to fledging stage in the Coconino National Forest near Flagstaff, Arizona, in June and July 1982. The study area was primarily ecotonal within the pon- derosa pine (Pinus ponderosa) forest of the Transition Life-zone (Lowe 1964). Stands of ponderosa pines were interspersed with more open areas of one-seed juniper (Juniperus monosperma ), Gambel’s Oak ( Quercus gambelli), squawbush (Rhus trilobata), prickly, pear cactus (Opuntia spp.), Parry rabbitbrush ( Chrysothamnus par- ry i), and blue grama grass (Bouteloua gracilis ). Elevation in the study area ranged from 2070 to 2160 m. Observations were made with 7-15x binoculars or 20-60x spot- ting scope, 100 to 200 m from each nest. Nests were observed in 2 to 8 h shifts between 0800 and 1800 h. Kestrels did not forage appreciably before 0800 h, perhaps due to the inactivity of or- thopterans, their principal prey. Wind velocity can affect kestrel foraging strategies and success (Rudolph 1982; Village 1983). My observations were confined to relatively calm days, since windy days added too many variables to adequately measure its effect on provisioning. Two hundred fifty-nine male and female provisioning trips to the nest were documented. Percent male and female provisioning and male and female prey deliveries/hr were calculated for each nest. Clutch size (determined just prior to hatching), number of young hatched and number of young fledged were recorded for each nest. Male and female prey deliveries/chick/hr were calcu- lated for each nest. Spearman’s rank correlation procedure (Zar 1974) was used to analyze data. A significance level of 0.05 was used in all tests. Results and Discussion Males contributed an average of 44.32% of the food deliveries to the nest from hatching to fledg- ing. Individual males, however, varied in their contributions relative to the female (Table 1). Per- cent male provisioning ranged from 18.0 to 67.1%. Male prey deliveries/hr ranged from 0.60 to 2.80 and was significantly correlated with percent male provisioning (r § = 0.943, P < 0.025). Clutch size, number of young hatched and number of young fledged were each significantly correlated with male prey deliveries/hr (r s = 0.843, P < 0.05; r s = 0.843, P < 0.05; r s - 0.929, P < 0.025). Balgooyen (1976) found that the female pro- vided 71.1% of the food deliveries to one nest dur- ing the period after hatching when both male and female hunt. Females in this study during the same period provided an average of 54.02% of the prey deliveries (n = 6), though there was high individual variation in female provisioning. Female prey de- liveries/hr ranged from 0.64 to 2.73 (Table 1). Neither female prey deliveries/hr nor total prey deliveries/hr correlated significantly with number of young fledged (r s = 0.500, P > 0.10; r s = 0.014 P > 0.25). Spearman r g between total prey deliveries/chick/hr and persent fledged was 0.629. While not significant (P > 0.10), this suggests that higher feeding rates result in higher chick survival. Spearman’s r g between male and female prey deliveries/hr was -0.657 (P = 0.10), suggestive of a negative correlation. Males and females of indi- vidual pairs may adjust their prey delivery rates relative to their mate’s abilities. The individual variation in male provisioning correlated with several measures of nest reproduc- tive success. Those males which delivered a greater number of prey/hr appeared to realize a higher immediate reproductive success of young hatched and fledged. Other factors could account for the observed re- sults. First, data were insufficient to evaluate quality of prey delivered by males or females. Some males may have provided relatively more vertebrates to their mates and broods than did others. The 128 Raptor Research Vol, 20 (3/4): 128-129 Fall/ Winter 1986 American Kestrels 129 Table 1. Male and female nest provisioning performance and female reproduction for six American Kestrel nests, northern Arizona, 1982. Percent Provisioning Prey Deliveries/hr Prey Deliveries/chick/hr Nest Male Female Male Female Total 1 67.1 22.9 1.31 0.64 1.95 2 52.4 47.6 0.85 0.77 1.62 3 18.0 82.0 0.60 2.73 3.33 4 25.0 75.0 0.71 2.13 2.84 5 43.8 56.2 0.82 1.05 1.87 6 59.6 40.4 2.80 1.90 4.70 energetic advantage this could confer is substantial; Bird et. al. (1982) estimated 1 vole contained about 125 times more energy than 1 grasshopper. Second, clutch size and brood size were each cor- related with male prey deliveries/hr. Perhaps males adjusted their respective rates of prey delivery ac- cording to the brood size for which they had to provide. Testing for such a functional response would require comparing male delivery rates be- fore and after hatching. Data were insufficient to evaluate this possibility. However, male prey de- liveries/chick/hr was not correlated with brood size (r s -0.129, P > 0.50), indicating that male prey delivery rates to individual chicks were the same for large and small broods. This suggests that males respond functionally to larger broods by providing prey items at a higher rate. In contrast, female prey deliveries/hr was not significantly correlated with brood size (r s = 0.554, P > 0.10), indicating that females do not respond functionally to larger broods by increasing their rates of prey delivery. The variability in male and female prey delivery rates could have resulted from other factors, such as age or experience of the bird (see Newton 1979) or territory quality (Newton 1976; Rudloph 1982). A final possibility is that smaller size and lower wing-loading of some males provided them with greater energetic efficiency (Balgooyen 1976; von Schantz and Nilsson 1981). These factors were not measured. Literature Cited Balgooyen, T.G. 1976. Behavior and ecology of the American Kestrel (Falco sparverius L.) Univ. Calif. Publ. Zool. 103:1-83. Bird. D.M., S. Ho and D. Pare. 1982 Nutritive values of three common prey items of the American Kestrel. Comp. Biochem. Physiol. 73A: 513-515. Male Female Total Clutch Size Young Hatched Young Fledged 0.262 0.128 0.390 5 5 4 0.170 0.154 0.324 5 5 3 0.200 0.910 1.110 4 3 2 0.355 1.065 1.420 3 2 2 0.273 0.350 0.623 4 3 2 0.560 0.380 0.940 5 5 4 Drent, R.H., and S. Dann. 1980. The prudent parent: energetic adjustments in avian breeding. 2Ardea 68: 225-252. Gary, H.C., and M.J. Morris. 1980. Constructing wooden boxes for cavity-nesting birds. Research Note RM-381, USD A Forest Service. Lowe, C.H. 1964. Arizona’s natural environment. Tus- con; Univ. of Ariz. Press. Mueller, H.C., N.S. Mueller, and P.G. Parker. 1981. Observation of a brood of Sharp-shinned Hawks in Ontario, with comments on the functions of sexual dimorphism. Wilson Bull. 93: 85-92. Newton, I. 1976. Breeding of sparrowhawks in diffe- rent environments. /. Anim. Ecol. 45: 831-849. 1978. Feeding and development of Spar- rowhawk Accipiter nisus nestlings. J. Zool. Lond. 184: 465-487. . 1979. Population ecology of raptors. Ver- million, South Dakota; Buteo Books. Rudolph, S.G. 1982. Foraging strategies of American Kestrels during breeding. Ecology 63: 1268-1276. Schantz, T. von, and I. N. Nilsson. 1981. The reversed size dimorphism in birds of prey: a new hypothesis. Oikos 36: 129-132. Snyder, N.F.R., and J.W. Wiley. 1976. Sexual size di- morphism in hawks and owls of North America. Or- nithol. Monogr. 20: 1-96. Village, A. 1983. Seasonal changes in the hunting be- havior of Kestrels. Ardea 71: 117-124. Wink, M., C. Wink, and D. Ristow. 1980. Biology of Eleanora’s Flacon (Falco eleonorae). 8. Clutch size in relation to hunting success and weight of the parent falcons./. Ornithol. 121: 387-390. Zar, J. 1974. Biostatistical analysis. Englewood Cliffs, N.J.; Prentice-Hall, Inc. Department of Biology, Northern Arizona University, Flagstaff, AZ 86011. Present addresss: San Antonio Missions National Historical Park, 2202 Roosevelt Ave., San Antonio, TX 78210. Received 8 October 1985; Accepted November 1986. 130 Short Communications Vol. 20, No. 3/4 Short Communications Success Rates of the Peregrine Falcon ( Falco peregrinus) Hunting Dunlin ( Calidris alpina) During Winter Joseph B. Buchanan, Steven G. Herman and Tod M. Johnson The Peregrine Falcon ( Falco peregrinus) utilizes a wide variety of prey types (Ratcliffe 1979; Cade 1982) throughout its nearly cos- mopolitan distribution. In addition, it also exhibits a huge range (7-83%) of success rates for hunting flights (see Roalkvam 1985 for review). Success rates of hunting flights dur- ing winter periods have been reported by Lindberg (1975), Clunie (1976) and Roalkvam (1985); however, data on success rates for specific prey species or prey type are lacking. Here we present data for rates of success when Peregrines hunted Dunlin ( Calidris alpina ) during the winter (December-March). During the winters of 1979-1981, while studying the ecology of the Dunlin in western Washington, we observed hunting flights di- rected at this species by peregrines. Hunting flights were observed at the Samish River De- lta, in northern Puget Sound, and at Bower- man Basin and 2 other estuarine sites located in Grays Harbor on the outer coast. Subadult and adult male and female Peregrines of 2 subspecies, F. p . pealei and F. p . anatum, were observed hunting Dunlins. A description of behavioral interactions between Dunlins and their falcon predators [Peregrines and Mer- lins (F. columbarius)] will be presented elsewhere. We define a hunting flight as a perch-to- perch flight involving one or more capture attempts at suitable prey. A capture attempt is defined as an individual effort to capture a specific individual during a hunting flight. We observed 17 hunting flights directed at Dunlins, 15 of which had known outcomes. Peregrines were successful at capturing Dun- lins from a stoop and when pursuing indi- viduals in a low direct chase. Stoops were used in 11 (65%) hunting flights. In 2 flights these were high stoops, originating from heights of 1000 m or more. Other stoops originated from 50-80 m. Feint stoops were observed only twice. Of 47 capture attempts, 33 (70%) were stoops at flocks, 6 (13%) were low chases of single Dunlins and 8 (17%) were horizontal attacks of flocks. The success rate for hunting flights was 47% (n= 7) while the rate for capture attempts was 14.6%. Five flights (33%) were successful on the first capture attempt. A stoop was used in 3 of these hunts, and low chase after a single Dunlin in the other 2. Seventy-one percent of the successful hunts were successful on the first capture attempt. All but one successful hunt involved in-flight prey capture. The success rate (47%) which we observed was significantly greater than the winter rate of 9.6% reported by Clunie (1976) (x^ = 7.9, df = 1; 0.001 < P < 0.005) or the 13.7% reported by both Lindberg (1975) and Roalkvam (1985) (x 2 = 1 1.6; df = 2; 0.001