sign in sign up
<<
Home , Archilochus (bird), Common nighthawk, Cyrtonyx, Eremophila (bird), List of birds of Chile, Virginia rail

The Condor 106:873-886 8 The Cooper Ornitholog~calSociety 2004

SUMMER DIET OF THE IN FAUNISTICALLY RICH AND POOR ZONES OF ARIZONA ANALYZED WITH CAPTURE-RECAPTURE MODELING

DAVID H. ELLIS~J~,CATHERINE H. ELL IS^, BETH ANN SAB03, AMADEO M. REA4, JAMES DAWSON~,JAMES K. FACKLER~,CHARLES T. LA RUE^, TERYL G. GRUBB~, JOHN SCHMITT~,DWIGHT G. SMITHIOAND MARC URY" 'USGS Southwest Biological Science Center, Sonoran Desert Field Station, HC I Box 4420, Oracle, AZ 85623 21nstitute for Raptor Studies, HC I Box 4420, Oracle, AZ 85623 'Otus Shelf; 1715 N. MacArthur Blvd., Irving, TX 75061 4Anthropology Program, University of Sun Diego, 5998 Alcala Park, Sun Diego, CA 92110 SLiberty Wildlife, 15825 N. Roadrunner Ridge Lane, Catalina, AZ 85739 65888 Inez Street, Bow, WA 98232 '3.525 W. Lois Lane, Flagstafl AZ 86001 8Rocky Mountain Research Station, Southwest Forest Science Complex, 2500 S. Pine Knoll Dr., Flagsta8 AZ 86001 9P.0. Box 9, Wofford Heights, CA 93285 l0Biology Department, Southern Connecticut State University, 501 Crescent Street, New Haven, CT 06515 "Swiss Ornithological Institute, CH-6204 Sempach, Switzerland

Abstract. We collected prey remains from 25 Peregrine Falcon (Falco peregrinus) terri- tories across Arizona from 1977 to 1988 yielding 58 eyrie-years of data. Along with 793 individual (107 and six additional genera). we found seven mammals and nine insects. In addition, two nestling peregrines were consumed. We found a larger dependence upon White-throated Swifts (Aeronautes saxatalis) and birds on migration in northern Arizona, while in southeastern and central Arizona average prey mass was greater and columbiforms formed the largest dietary component. In northern, central, and southeastern Arizona, 74, 66, and 56 avian prey taxa, respectively, were recorded. We used capture-recapture modeling to estimate totals of 111 ? 9.5, 113 ? 10.5, and 86 i 7.9 (SE) avian taxa taken in these same three areas. These values are counterintuitive inasmuch as the southeast has the richest avi- fauna. For the entire study area, 156 2 9.3 avian taxa were estimated to be taken by peregrines. Key words: Arizona, diet. Falco peregrinus, Peregrine Falcon, prey.

Dieta Estival de Falco peregrinus en Arizona Cornparando Zonas Ricas y Pobres en Avifauna Mediante un Modelo de Captura-Recaptura Resumen. Desde 1977 a 1988 colectamos restos de presas en 58 nidos de Falco pere- grinus a travts de Arizona. Conjuntamente con 793 aves individuales (107 especies y seis gtneros adicionales), encontramos siete mamiferos y nueve insectos. Ademas, fueron con- sumidos dos pichones de Falco peregrinus. En la zona norte encontramos una mayor de- pendencia sobre Aeronautes saxatalis y aves en migraci6n, mientras que en las zonas sureste y central la masa promedio de presa fue mas grande y 10s columbiformes constituyeron el componente principal de la dieta de Falco peregrinus. En las zonas norte, central y sureste se registraron 74, 66 y 56 taxa de aves presa, respectivamente. Para estimar el ndmero total de taxa capturados por Falco peregrinus usamos un modelo de captura-recaptura. Los va- lores calculados fueron 111 f 9.5, 113 5 10.5 y 86 i: 7.9 (EE) taxa para las zonas norte, central y sureste, respectivamente. Estos valores no reflejan 10s que esperabamos, ya que la zona sureste tuvo una avifauna mas rica. Se estim6 que 156 i 9.3 taxa fueron capturados por Falco peregrinus en la totalidad del Area de estudio.

INTRODUCTION data have been collected on diet for the Pere- rin With the exception of the Barn (Tyto alba; g e (Falco peregrinus) than for any Clark et al. 978, Marti 1988), perhaps other raptor (Ratcliffe 1980. Porter et al. 1987, White et al. 2002). Unlike the , which concentrates on one or a few species of locally Manuscript received 17 December 2003; accepted abundant the peregrine 22 June 2004. takes a wide variety of prey, except when nest- l2 Corresponding author. E-mail: [email protected] ing on islands and feeding on seabirds (e.g., An- 876 DAVID H. ELLIS ET AL.

STATISTICAL ANALYSIS widely used heterogeneity model M,, (Otis et al. Proportion of prey by numbers. To test if avian 1978) implemented in program SPECRICH prey taxa were taken in similar proportions (Hines et al. 1999) to estimate the total number across the three regions, we used a log-linear of avian taxa taken in each zone as well as model (McCullagh and Nelder 1989) imple- across Arizona. Although this is a new applica- mented in the statistical package GenStat (Payne tion of capture-recapture modeling, similar mod- et al. 1993). A log-linear model is an example eling has been used many times to estimate the 6 of a generalized linear model where the log of size of populations (Otis et al. 1978, the expectation of a count (p) is related to a sum Burnham and Overton 1979, Boulinier et al. of covariate effects, just as in ordinary regres- 1998). Our estimates for prey taken in each re- sion analysis (including ANOVA): log(pijk)= a gion are provided as estimates t SE with sig- + b (taxon,) + c (regionj) + d (taxon X regionq) nificance level set at 0.05. + e,. A significant interaction taxon X regionq Proportion of prey by mass. We assigned pub- would indicate that the distribution of taxa taken lished values (Dunning 1993) of average adult as prey is not the same across the three regions. mass for species recorded as prey. This practice As in ordinary ANOVA, this interaction can be ignores the fact that juveniles, often taken as partitioned into contrasts to further study differ- prey, typically weigh less than adults. Therefore, ences among regions. this practice overestimates biomass (Rosenfield To avoid having low frequencies of occur- et al. 1995) but would not affect prey variety rence invalidate the analysis (i.e., for species de- and should have a minimal effect on our com- tected as prey only once or a few times), we parisons between zones. pooled taxa into the following groups: ( only); Odontophoridae ( only); RESULTS (White-faced [Plegadis TOTAL NUMBER OF TAXA TAKEN chihi] only); (Sharp-shinned Our list of prey taken for all of Arizona (Ap- and only); Rallidae; pendix) included 113 avian taxa (i.e., 107 spe- small shorebirds (20-100 g); large shorebirds cies, plus six additional genera that could not be (>100 g); ; ; Strigidae; Ca- identified to species), three species of mammals, primulgidae; Apodidae; Trochilidae; Picidae; and three insect taxa. For the entire study area, Corvidae; Alaudidae (Homed [Eremophila capture-recapture modeling estimated that 156 + alpestris] only); Hirundinidae; (Sialia 9.3 avian taxa were taken by peregrines during spp.); Orioles (Icterus spp.); Icteridae (excluding the study period. In our lists, we recorded 113 orioles); other small Passeriformes (5-30 g); and species, 72% of this value. Comparing the esti- other mid-sized Passeriformes (31-100 g). mated total prey taxa taken in Arizona with our Estimating the variety of prey taken. Because list of potential prey taxa, we estimated that per- our collections constitute only a sample of the egrines preyed on 71% (156 of 221) of the total prey taken by peregrines in Arizona and because available avian prey taxa. of imperfect detectability, not every taxon ac- In the northern, central, and southeastern tually taken appears in our samples. This con- zones, we recorded 73, 66, and 55 avian prey dition results in an underestimate of prey rich- taxa, respectively. Using capture-recapture mod- ness, and this underestimate is inversely propor- eling, we estimated that 111 ? 9.5, 113 + 10.5, tional to the size of the prey list. and 86 2 7.9 avian taxa were taken in these To compensate for this undersampling bias, same three zones. This suggests that among the we used capture-recapture modeling. This ap- available avian taxa (Table I), peregrines in proach assumes that not all species actually tak- northern, central, and southeastern Arizona en will show up in prey lists, but that the em- preyed on 67%, 56%, and 42% of the available pirical species-abundance distribution contains avian taxa, respectively. information about the probability with which a species will be detected. Because there were no PREY USE BY NUMBER AND MASS repeated sampling occasions in our study (i.e., In Arizona, the three most common groups of only one sample was taken from an eyrie in any birds taken were swifts, columbids, and shore- year), we used the limiting form of the jacknife birds (Table 2). Although White-throated Swifts estimator (Burnham and Overton 1979) of the were consistently important in the diet by num- 874 DAVID H. ELLIS ET AL. cient Murrelets [Synthliboramphus antiquus], TABLE 1. Number of species that are potential Nelson 1990; Heermann's [Larus heer- Peregrine Falcon prey in three zones of Arizona. manni], Velarde 1993). Total In Arizona, incidental records of peregrines Zone Breeders Migrants species were recorded by various ornithologists from 1884-1975 (Ellis and Monson 1989). Organized North 112 54 166 Central 150 52 202 studies of the Peregrine Falcon began in Arizona Southeast 140 67 207 4 in 1975 (Ellis 1988). An intensive survey of old sites (begun in 1975) and a search for new eyries (begun in 1977) documented a general trend to- ward reoccupancy and high productivity (Ellis this purpose, we assembled lists of likely prey and Glinski 1988, Brown et al. 1992). This find- species in each zone. Because the largest prey ing is consistent with moderate pesticide levels item we collected in Arizona was a Northern for prey and minimal eggshell thinning (Ellis et Pintail (Anas acuta, 1011 g, Dunning 1993) and al. 1989). because Paine et al. (1 990) found that the largest In this long-term study, we compared dietary prey for peregrines at Tatoosh Island, Washing- differences for three regions of Arizona, includ- ton, was the Common Murre (Uria aalge; 1010 ing southeastern, where the diversity of avian g), we used 1100 g as the upper limit for prey. prey is 20% greater than for northern Arizona The smallest birds found as prey in the eyrie (Table 1). We also make a new application of middens were two species of capture-recapture models to adjust for differen- (Black-chinned [Archilochus ale- tial undersampling bias resulting from prey lists xandri] 4 g, and Broad-tailed Hummingbird [Se- that vary greatly in size. lasphorus platycercus] 3 g). Hummingbirds are so small and were so seldom taken (n = 3) that METHODS they were judged unimportant in the diet, by ei- For this study, we included Arizona and a nar- ther number or mass, and excluded from our row swath of habitat in New Mexico and Utah lists of potential prey in each zone. Our lists where peregrines nesting within those states included all species (except as stated below) ly- must, at least sometimes, hunt in Arizona. Be- ing between 5 g and 1100 g. To avoid inflating cause no Peregrine Falcons were known to breed our list, we excluded rare or accidental birds, along the lower Colorado River or elsewhere in even if they appeared once or twice as prey. We southwestern Arizona during the study, we gath- also excluded (except the diurnal Burrow- ered diet only for northern, central, and south- ing Owl [Athene cunicularia]) and all but the eastern Arizona, three broad geographic zones smallest Falconiformes (i.e., we included only (see map in Ellis 1988) based on topography and the American Kestrel [Falco spawerius] and the vegetation (Brown and Lowe 1980). The north- Sharp-shinned Hawk [Accipiter striatus]). Larg- ern zone consisted of the canyons and mesas of er raptors were excluded because attacks on the Colorado Plateau including the drainages of them were considered territorial defense, not the Virgin, Colorado, Little Colorado, San Juan, . Nestling peregrines (recorded as prey and Paria Rivers with their myriad associated twice; Ellis, Oliphant, and Fackler 2002) were canyons. The central zone consisted of a broad likewise excluded from our analyses as not hav- swath of forests and mountains in the topograph- ing been captured elsewhere and brought to the ically complex region dropping away south and eyrie. Cannibalization of young peregrines has west from the Mogollon Rim and lying north of occasionally been reported (White and Cade the Gila River. The southeastern zone consisted 197 1, Court et al. 1988, Bradley and Oliphant of basin-and-range habitats with conifer- and 199 1). live oak-covered mountain ranges surrounded in In compiling potential prey lists, we included the east by Chihuahuan Desert grasslands and in only prey which were likely to have been taken the west by Sonoran Desert plant communities. during the breeding season (1 March to 1 Au- gust). The birds present within this time frame POTENTIAL PREY were divided into two categories: breeding birds To measure food preference, we must compare (including both resident breeders and migratory what is actually taken with what is available. For breeders) and migrants. Table 1 provides totals PREY OF THE PEREGRINE FALCON IN ARIZONA 875

of potential prey species for the Peregrine Fal- when comparing it with a series of museum con in the three zones. As expected, the list for specimens. northern Arizona was much smaller (i.e., the In documenting food habits, we collected avifauna was less diverse) than the list for the most items from eyries, but some prey items southeast. However, contrary to our expecta- were gathered from plucking perches in the im- tions, the central Arizona list was nearly as large mediate vicinity of an eyrie. However, only as for the southeast (Table 1). It might appear those items were counted for which enough ma- that these latter two zones could be combined, terial was present, or for which feathers showed but the actual species in these zones were dif- bill marks, to indicate that the material was in- ferent enough that we decided to treat the two deed falcon prey. For example, although molted zones separately. raven (Cowus corax or C. cryptoleucus) feathers were frequently found beneath cliffs used by PREY SAMPLING peregrines, the only raven tabulated as prey was We collected Peregrine Falcon prey remains for represented by a sizable tuft of feathers con- 58 eyrie-years representing 25 territories from nected by flesh and found directly in one eyrie. 1977 through 1988. Of these, 21, 20, and 17 Also, White-throated (Aeronautes saxata- eyrie-years were from the northern, central, and lis) feathers were regularly found (sometimes southeastern zones, respectively. Because some copiously) at the base of some eyrie cliffs: all sites had no young in some years (Ellis 1988), of these we ignored. and because some eyries were under overhung One peregrine eyrie was evidently used pre- ledges and others were on inaccessible spires viously by Spotted Owls (Strix occidentalis): (Ellis 1982), only a few sites were sampled in others had been used by ravens or Prairie Fal- any one year (mean 5.3 sites per year, range = cons (Falco mexicanus). To avoid the risk of 1-10), and the number of times a site was sam- tabulating as prey an item which was collected pled during the study varied from 1-6 (mean 2.3 in an earlier year by another species, the bony + 1.0 [SD] samples). Of 18 sites sampled more remains found in falcon eyries were counted than once, the interval between samples at the only if they showed clinging tendons or flesh, same site ranged from 1-4 years (mean 1.8 2 evidence of their having been depredated during 1.2 years). We visited eyries only after the that breeding season. Bony remains from a par- young had fledged to avoid adversely affecting ticular species were subtracted from the feather reproduction of this endangered subspecies (F. remains of that species (or vice versa) to avoid p. anatum; now listed as threatened). counting an individual twice. Feather remains were analyzed using refer- When a large number of feathers in one eyrie ence collections at the University of Arizona, the sample were attributed to one species, we used U.S. National Museum (Smithsonian Institu- tion), and the Western Foundation for Vertebrate diagnostic feathers, bills, or feet to determine the Zoology. Nearly all bones were identified using minimum number of individuals present. For the collection at the San Diego Natural History doves and swifts, the outermost left or right pri- Museum and the A. M. Rea collection (now at mary most often provided our minimum count. the University of New Mexico, Albuquerque). Regurgitated pellets found in the eyries were Insects were identified only to family as sug- not analyzed in this study. Our primary reason gested by Rosenberg and Cooper (1990). for not using pellets was that five pellets may For prey identification purposes, we prepared represent one prey item fed to five young. Also, a series of flat skins (i.e., skins with minimal it is easier to identify a prey species from clean body stuffing and with the tail and one wing feathers than from a regurgitated pellet. Further spread). Comparisons of feather patterns, espe- logic for not using pellet analysis is that vege- cially in the wing, were thereby facilitated and table and insect remains often found in pellets damage to the reference specimen minimized. may actually come from the stomach contents of Our methods of identifying prey were described avian prey, not from insects captured by the fal- by Ellis, Sabo et al. (2002). To prevent loss of cons. In fact, the only insects tallied in our lists sample feathers and for ease in handling, we rec- were those for which large wings or large legs ommend use of a hemostat to grasp a feather were found intact in the eyries. TABLE 2. Proportions of avian groups commonly found in the diet of Peregrine Falcons. Of the total, 793 items statewide, 390 were from northern, 194 from central, and 209 from southeastern Arizona. Numbers are minimum counts of individuals recovered from eyries and plucking perches. Values for mass are from Dunning (1 993).

North Central Southeast Taxonomic % by % by % by % by % by % by Group n number mass n number mass n number mass Anatidae Odontophoridae Threskiornithidae Falconiformesa Rallidae Shorebirds, 20-100g Shorebirds,> 100g Laridae Columbidae Strigidae Caprimulgidae Apodidae Trochilidae Picidae Corvidae Alaudidae Hirundinidae Turdidae Bluebirds (Sialia spp.) Icteridae Orioles (Zcterus spp.) Other Icteridae Other Passeriformes, 5-30 g Other Passeriformes, 30-100 g Unidentified birds

a Excludes 2 cannibalized nestling peregrines. 878 DAVID H. ELLIS BT AL. ber (26%, 19%, and 16% in the north, central, nithidae (Table 2). In central Arizona, Colum- and southeast, respectively), they made up only bidae contributed 46% of the biomass, with 5-10% of the diet by biomass. By far the most Anatidae and Picidae also important. In the important family by mass for the three zones southeast, Corvidae was second, but far behind combined was Columbidae. Columbids were Columbidae. Across Arizona, the most impor- most important by numbers in the central and tant taxa by mass, in order of increasing impor- southeastern zones. In the north, large shorebirds tance, were Laridae, Rallidae, Anatidae, Picidae, were most significant in the diet by mass, a fact Apodidae, Corvidae, large shorebirds, and Co- reflecting a greater dependence oi migrants. lumbidae, with columbids representing over Prey importance by number. There were 390, 40% of prey mass. 194, and 209 items recorded in the northern, central, and southeastern zones, respectively. DISCUSSION Prey taxa were not taken in even propbrtions In METHODS OF COMPILING PREY LISTS the northern, central, and southeastern zones (log-linear analysis: taxon X region interaction, Each method of gathering dietary information xZd2= 114.2, P < 0.001). When we further par- has advantages and disadvantages (Rosenberg titioned this interaction by combining zones, we and Cooper 1990, Redpath et al. 2001, Booms found that the north + central versus southeast and Fuller 2003). Where direct observations of contrast contributed x2,, = 52.9, while the north hunting are not available, pellet analysis may be versus central + southeast contrast contributed necessary to estimate the importance of insects xZ2,= 94.1 to the total chi-square. This indicates and small mammals (Bradley and Oliphant that central and southeastern zones were more 199 1, Morimando et al. 1997). Another potential similar in terms of the distribution of peregrine bias in dietary studies using prey remains is that prey among taxa than were northern adcentral rare items can be overemphasized. For example, Arizona. 1000 swift feathers from 100 individuals may A comparison of observed and expected num- tally only 10 birds, but one feather will ber of captures per taxon group and zone (Table tally one bluebird. Remnants of rare birds, rather 3) shows that in northern Arizona, Rallidae, than being discarded, are emphasized. For ex- small and large shorebirds, Laridae, Apodidae, ample, the first Arizona specimen record for the Alaudidae, and Turdidae among others, were Common Grackle (Quiscalus quiscula) was re- more frequently taken than the average for all covered from a peregrine eyrie (LaRue and Ellis of Arizona. In central Arizona, the following 1992), just as the first specimen record for the taxa were most prominent: Anatidae, Falconi- Wood Lark (Lullula arborea) on the Isle of Man formes, Columbidae, and Picidae. In southeast- was a presumed peregrine kill (Oxenham 1979). ern Arizona, Columbidae, among others, was Surely the most accurate method for assessing prominent. the diet of breeding raptors would be to follow Prey importance by mass. Ignoring taxonomic the adults and observe the species and size of group and focusing only on mass, significantly prey taken. But following a predator long term heavier prey were caught in the central and is nigh impossible logistically. We know of only southeastern zones than in northern Arizona one such study; White and Nelson (199 1) used (ANOVA, F,,,,, = 6.1, P = 0.002). Mean mass a helicopter to follow two peregrines and two of prey caught was 56.6 g (95% CI: 50.6-63.5) Gyrfalcons (Falco rusticolus) on hunting forays. in the north, 72.4 g (95% CI: 61.5-85.1) in the Another way to lessen sampling biases is long- central zone, and 70.6 g (95% CI: 60.4-82.4) in term observations from very near the eyries (in- the southeast. Here, once again, central and cluding the use of videocameras). This method southeastern Arizona were more similar than should more accurately show the importance of were the northern and central zones. nestling birds in the diet (as found for pere- By mass, Columbidae was the most important grines, Rosenfield et al. 1995; and Gyrfalcons, taxon in the central and southeastern zones and Booms and Fuller 2003). As an alternative, data was second in importance in northern Arizona, can be taken from lengthy watches at prey con- where shorebirds made the greatest contribution centration sites (Dekker 1980, 1999). However, to total prey mass. Other prey taxa important in some bias is also associated with these labor- the north were Laridae, Rallidae, and Threskior- intensive methods. Most of the literature on rap- TABLE 3. Proportions of avian taxa in prey remains of Peregrine Falcons in three zones of Arizona. Excepted number of occurrences are based on a log-linear model that assumes prey are caught in equal proportions in the three regions of Arizona. Discrepancy is the percent difference between observed and expected captures and shows percent over or underrepresentation of a taxon in a region relative to the expectation when lumping all three zones.

North Central Southeast Taxonomic Discrepancy Discrepancy Discrepancy group Observeda Expectedd (%) Observeda Expectedd (a) Observeda Expecteda (%) Anatidae I 2.5 -60 3 1.2 147 I 1.3 -24 Odontophoridae 0 2.0 -100 2 1 .0 106 2 1.1 90 Threskiornithidae 2 1 .0 102 0 0.5 -100 0 0.5 -100 Falconiformesa 5 5.4 -8 4 2.7 50 2 2.9 -3 1 Rallidae 4 3.0 3.5 2 1.5 37 0 1.6 -100 Shorebirds, 20-100g 11 7.4 48 1 3.6 -73 3 4.0 -24 Shorebirds, > 100g 30 19.3 56 4 9.5 -5 8 5 10.3 -5 1 Laridae 6 4.5 35 3 2.2 37 0 2.4 -100 Columbidae 53 89.9 -4 1 52 44.1 18 77 47.9 6 1 Strigidae 0 I .0 -100 I 0.5 106 1 0.5 90 Caprimulgidae 7 5.9 18 1 2.9 -66 4 3.2 27 Apodidae 102 85.0 20 37 41.7 -1 1 33 45.3 -27 Trochilidae I 1.5 -33 1 0.7 37 I 0.8 27 Picidae 11 17.8 -38 1.5 8.7 72 10 9.5 6 Corvidae 26 23.7 10 12 11.6 3 10 12.6 -2 1 Alaudidae 14 10.4 35 3 5.1 -4 1 4 5.5 -28 Hirundinidae 9 9.9 -9 5 4.9 3 6 5.3 14 Turdidae Bluebirds (Siuliu spp.) I I 7.9 39 4 3.9 3 1 4.2 -76 Icteridae Orioles (lcterus spp.) 0 3.0 -1 00 3 1.5 1 06 3 1.6 90 Other lcteridae 19 17.8 7 8 8.8 -8 9 9.5 -5 Other Passeriformes 5-30 g 43 33.6 28 I I 16.5 -33 14 18.0 -22 31-100 g 26 28.7 -9 1.5 14.1 7 17 15.3 11 " Units are number of individuals. 880 DAVID H. ELLIS ET AL. tor food habits derives from methods known to to properly deal with unequal probability of taxa be biased (namely, pellet analyses and prey re- being included in prey lists of different lengths. mains from eyries; Sherrod 1978). Some studies In our study the concept of trapping occasions P have even relied only on materials collected at was not important; we used instead the frequen- the base of breeding cliffs (Oro and Tella 1995). cy with which each taxon was captured. From Prey remains sometimes adequately document this, an estimator of a widely used, closed-pop- the proportion of each prey species in the diet ulation, capture-recapture M, model was com- of certain raptors (e.g., Collopy 1983, Golden puted (Bumham and Overton 1979). In other [Aquila chrysaetos]). Other studies sug- studies, a sampling occasion might be defined as gest that prey remains alone are inadequate; see an individual eyrie visit, or a sample may be all Simmons et al. (1991) for the African Marsh- prey gathered in a year. More sophisticated mod- (Circus ranivorus), Bielefeldt et al. els may allow for the inclusion probability to (1992) for the Cooper's Hawk (Accipiter coo- vary by occasion, or may directly estimate the perii), Redpath et al. (2001) for the Hen Harrier number of species that are common to prey lists (C. cyaneus), Booms and Fuller (2003) for the fop two regions (using the methods of Boulinier Gyrfalcon, and Marti (1987) for a general dis- et al. 1998, Nichols et al. 1998). cussion. For the Peregrine Falcon in Greenland, Rosenfield et al. (1995) found similar frequen- PREY CLASSES cies and rankings for six major prey species when prey remains were compared to direct ob- The preponderance of birds in the diet of the servations of prey deliveries. However, they not- Peregrine Falcon worldwide is well known (Rat- ed that prey remains frequently did not ade- cliffe 1980, Mearns 1983); some studies report quately show the age of the prey (with the result avian prey exclusively (e.g., Rosenfield et al. of skewing biomass estimates too high through 1995 for Greenland). In our study of prey re- the use of adult weights). Although this bias is mains from Arizona, 98% were birds. present in our study, we believe that we Few mammals (chipmunk [Eutamius spp.], achieved good zonal comparisons in both the cottontail [Silvilagus auduboni], and unidenti- biomass and the variety of avian prey in the diet fied) were reported from prey remains in each of the Peregrine Falcon in Arizona from prey of the three zones in Arizona. Predation on remains. We are confident, however, from direct was probably significant based on hunting ob- observations of hunting behavior (Ellis and servations (Glinski 1998), but the lack of bats in LaRue, unpubl. data), that insects and bats are our prey remains in eyries may stem from the grossly underrepresented in our samples. As a fact that pellets were not analyzed (Bradley and result, our regional comparisons must be viewed Oliphant 199 1). as predation rates for avian prey only. Ratcliffe (1980) speculated that insect remains Our analysis was strengthened by the fact that found in pellets could derive from insects cap- feather and osteological remains were almost al- tured on the eyrie ledge, meaning that the insects ways (98% of the time) identified to species. were not being brought as prey by the adults. Greene and JaksiC (1983) found that a precise Insects are seldom reported as food items in per- level of prey identification (to the generic or spe- cific level) versus ordinal identification (to fam- egrine prey studies, but Snyder and Wiley ily level only) was necessary to accurately rep- (1976) estimated that insects represented 20% of resent food-niche parameters. They also sug- the diet (by number) in a sample of 116 stom- gested that taxonomic identification of prey achs. Also, Oro and Tella (1995) found arthro- items is more relevant than size class in docu- pod remains important in pellets, but not in prey menting competitive interactions among preda- remains, under eyries in Spain, and Ritchie tors. (1982) reported insects in the diet of Alaskan peregrines. Dekker (1999) concluded that insects ESTIMATION OF THE TRUE LENGTH OF PREY are commonly taken by recently fledged pere- LISTS grines. Visual observations of peregrines hunt- As far as we are aware, this is the first appli- ing in Arizona (Ellis and LaRue, unpubl. data) cation of capture-recapture modeling to estimate showed that insects were important in the diet, the total number of taxa taken by a raptor and at least numerically. PREY OF THE PEREGRINE FALCON IN ARIZONA 881

REGIONAL COMPARISONS tors, JaksiC and Braker (1983) concluded that Major differences between the prey taken in local diet was influenced more by prey avail- northern, central, and southeastern Arizona were ability than by species-specific preferences. This found in both numbers and biomass. We ex- explanation is consistent with observations that pected that Peregrine Falcons in the species-rich even this supposedly obligate ornithophage is southeast would take a larger variety of prey able to shift to a diet high in mammals during than peregrines in the relatively depauperate microtine eruptions in the Arctic (Court et al. north. However, the opposite was true. One ex- 1988, Bradley and Oliphant 1991) or feast on planation is that peregrines in the north have insect swarms (Ellis and LaRue, unpubl. data). limited prey available and are therefore forced Arizona peregrines seem to focus on a few spe- to take whatever birds are available, both breed- cies but, through time, take a wide variety as ing birds and birds on migration. That White- opportunity allows. throated Swifts comprised a larger percentage of ACKNOWLEDGMENTS the diet in the north than in either of the other two zones is perhaps related to a smaller prey We express our appreciation to Lloyd Kiff (then at the base in the north, or at least limited vulnerability Western Foundation of Vertebrate Zoology), Phil An- gle and Richard Zusi at the U.S. National Museum of other species. Peregrines in the central and (Smithsonian Institution), and Steve Russell at the Uni- southeast may focus on fewer species of breed- versity of Arizona for providing access to facilities for ing birds near their eyries; they do take heavier the identification of bird remains. Roxy Laboume and prey. These results are consistent with the tenets Storrs Olson, both at the Smithsonian, helped by con- firming the identity of a few troubling samples. Bonnie of optimal foraging theory (Orians and Pearson Farmer (then at Smithsonian) identified mammalian 1979). hair. Carl Olsen (University of Arizona) and Gary He- Finally, a larger percentage of the species vel (Smithsonian) helped with insect remains. available was being used by Peregrine Falcons Fieldwork was financed primarily by the U.S. Fish in the north than in either the central or south- and Wildlife Service (especially Patuxent Wildlife Re- search Center), the National Forest Service, the Bureau eastern zones. This may not be due to the lower of Land Management, the Navajo Fish and Game diversity of prey in the north, but it may be a Branch, the U.S. Air Force, and the National Park Ser- response to a smaller number of birds of each vice; minor funding was provided by the Maricopa species (except swifts) being available and vul- Audubon Society, Tucson Audubon Society, and the Institute for Raptor Studies. Alan L. Hatch and E B. nerable (i.e., aloft, far from cover). The greater l? Trahan served as "safety" on most climbs. Bryan dependence of falcons in the north on migrating Brown and Scott Mills provided prey remains from birds was perhaps related to the assumption that one eyrie. Publication costs were paid by the National migrating birds are more vulnerable to predation Aeronautics and Space Administration. because they are forced to travel far from cover. LITERATURE CITED An abundance of a few favorite prey species (as in southeastern Arizona) may characterize an BIELEFELDT,J., R. N. ROSE~ELD,AND J. M. PAPP. ideal peregrine foraging area, whereas the use of 1992. Unfounded assumptions about diet of the a wide variety of prey at one eyrie or in one Cooper's Hawk. Condor 94:427-436. BOOMS,T. L., AND M. R. FULLER. 2003. Gyrfalcon diet zone, as in northern Arizona, may signify that in central west Greenland during the nesting pe- the area lacks a consistently available prey base riod. Condor 105:528-537. of one or a very few species. However, it is BOULINIER,T., J. D. NICHOLS, J. R. SAUER, J. E. HINES, AND K. H. POLLOCK. 1998. Estimating species noteworthy that, over the period from 1976 richness: the importance of heterogeneity in spe- through 1985 (when the bulk of our prey data cies detectability. Ecology 79: 1018-1028. were gathered), peregrines in the north produced BRADLEY, M., AND L. W. OLIPHANT.1991. The diet of more young per attempt than did pairs in either Peregrine Falcons in Rankin Inlet, Northwest Ter- ritories: an unusually high proportion of mam- central or southeastern Arizona (Ellis 1988). In malian prey. Condor 93:193-197. the diet of peregrines for areas where breeding BROWN, B. T., G. S. MILLS, R. L. GLINSKI,AN D S. W. density is (or has been) very high, such as in the HOFFMAN.1992. Density of nesting Peregrine Fal- Queen Charlotte Islands, one species, the An- cons in Grand Canyon National Park, Arizona. cient Murrelet, constitutes >90% of prey by Southwestern Naturalist 37:188-193. BROWN, D., AND C. H. LOWE.1980. Biotic communi- number (Nelson 1977). In evaluating these and ties of the Southwest [map]. U.S. Forest Service related concepts for the diet of 11 genera of rap- General Technical Report RM-78. 882 DAVID H. ELLIS ET AL.

BURNHAM,K. P., AND W. S. OVERTON.1979. Robust ing a capture-recapture approach. Bird Study estimation of population size when capture prob- 46(Suppl.):209-2 17. abilities vary among . Ecology 60:927- JAKSIC,E M., AND H. E. BRAKER.1983. Food-niche 936. relationships and guild structure of diurnal birds * CLARK, R. J., D. G. SMITH, AND L. H. KELSO.1978. of prey: competition versus opportunism. Cana- Working bibliography of owls of the world. NWF dian Journal of Zoology 61 :2230-2241. SciencetTechnical Series 1. National Wildlife Fed- LARUE,C. T,AND D. H. ELLIS. 1992. The Common eration, Washington, DC. Grackle in Arizona: first specimen record and 1 COURT, G. S., C. C. GATES, AND D. A. BOAG. 1988. notes on occurrence. Western Birds 23:84-86. Natural history of the Peregrine Falcon in the MARTI,C. D. 1987. Raptor food habits studies, p. 67- Keewatin district of the Northwest Territories. 80. In B. A. G. Pendleton, B. A. Millsap, K. W. Arctic 41 : 17-30. Cline, and D. M. Bird [EDS.],Raptor management COLLOPY, M. W. 1983. A comparison of direct obser- techniques manual. National Wildlife Federation, vations and collections of prey remains in deter- Washington, DC. mining the diet of Golden . Journal of Wild- MARTI,C. D. 1988. A long-term study of food-niche life Management 47:360-368. dynamics in the Common Barn-Owl: comparisons DEKKER,D. 1980. Hunting success rates, foraging hab- within and between populations. Canadian Journal its, and prey selection of Peregrine Falcons mi- of Zoology 66:1803 -18 12. grating through Central Alberta. Canadian Field- MCCULLAGH,I?, AND J. A. NELDER. 1989. Generalized Naturalist 94:371-382. linear models. Chapman and Hall, London, UK. DEKKER,D. 1999. Bolt from the blue: wild peregrines MEARNS,R. 1983. The diet of the peregrine Falco pe- on the hunt. Hancock House, Surrey, BC, Canada regrinus in south Scotland during the breeding and Blaine, WA. season. Bird Study 30331-90. DUNNING, J. B. JR. [ED.]. 1993. CRC handbook of avian MORIMANDO, E, E PEZZO,AND A. DRAGHI.1997. Food body masses. CRC press, Boca Raton, FL. habits of the Lanner Falcon (Falco biarnzicus fel- ELLIS, D. H. 1982. The Peregrine Falcon in Arizona: deggii) in central Italy. Journal of Raptor Research habitat utilization and management recommenda- 31:40-43. tions. Institute for Raptor Studies, Research Re- NELSON, R. W. 1977. Behavioral ecology of coastal port 1, Tucson, AZ. Peregrines (Falco peregrinus pealei). Ph.D. dis- ELLIS, D. H. 1988. Distribution, productivity, and sta- sertation, University of Calgary, Calgary, AB, Ca- tus of the Peregrine Falcon in Arizona, p. 87-94. nada. In T. J. Cade, J. H. Enderson, C. G. Thelander, NELSON, R. W. 1990. Status of the Peregrine Falcon, and C. M. White [EDS.],Peregrine Falcon popu- Falco peregrinus pealei, on Langara Island, lations: their management and recovery. The Per- Queen Charlotte Islands, British Columbia, 1968- egrine Fund, Boise, ID. 1989. Canadian Field-Naturalist 104: 193-199. ELLIS, D. H., L. R. DEWEESE,T. G. GRUBB, L. E Km, NICHOLS, J. D., T. BOULINIER,J. E. HINES, K. H. POL- D. G. SMITH, W. M. JARMAN,AN D D. B. PEAKALL. LOCK, AND J. R. SAUER.1998. Estimating rates of 1989. Pesticide residues in Arizona Peregrine Fal- local species extinction, colonization, and turn- con and prey. Bulletin of Environmental over in animal communities. Ecological Applica- Contamination and Toxicology 42:57-64. tions 8:1213-1225. ELLIS, D. H., AND R. L. GLINSKI. 1988. Population es- ORIANS, G. H., AND N. E. PEARSON.1979. On the the- timates for the Peregrine Falcon in Arizona: a hab- ory of central place foraging, p. 155-176. In D. itat inventory approach. National Wildlife Feder- J. Horn, G. R. Stairs, and R. D. Mitchell [EDS.], ation Scientific and Technical Series 11 : 191-196. Analysis of ecological systems. Ohio State Uni- ELLIS, D. H., AND G. MONSON. 1989. A historical re- versity Press, Columbus, OH. view of Peregrine Falcon breeding and summering ORO, D., AND J. L. TELLA. 1995. A comparison of two records for Arizona and the Navajo Indian Res- methods for studying the diet of the Peregrine Fal- ervation. Southwestern Naturalist 34565-569. con. Journal of Raptor Research 29:207-210. ELLIS, D. H., L. W. OLIPHANT, AND J. K. FACKLER. OTIS, D. L., K. I? BURNHAM,G. C. WHITE, AND D. R. 2002. Schizochromism in a Peregrine Falcon from ANDERSON. 1978. Statistical inference from cap- Arizona. Journal of Raptor Research 36:200-202. ture data on closed animal populations. Wildlife ELLIS, D. H., B. A. SABO, J. K. FACKLER, AND B. A. Monographs 62. MILLSAP.2002. Prey of the Peregrine Falcon (Fal- OXENHAM,I? 1979. Unusual peregrine kill: a co peregrinus cassini) in southern Argentina and at Cranstal. Peregrine 5:34. . Journal of Raptor Research 36:315-3 19. PAINE, R. T., J. T WOOTTON,AND P: D. BOERSMA. 1990. GL~SKI,R. L. [ED.]. 1998. The raptors of Arizona. Direct and indirect effects of Peregrine Falcon University of Arizona Press, Tucson, AZ. predation on seabird abundance. Auk 107: 1-9. GREENE, H. W., AND E M. JAKSIC.1983. Food-niche PAYNE, R. W., F! W. LANE, I? G. N. DIGBY,S. A. HAR- relationships among sympatric predators: effects DING, P: K. LEECH, G. W. MORGAN, A. D. TODD, of level of prey identification. Okios 40: 15 1-154. R. THOMPSON, G. TUNNICLIFFEWILSON , S. J. WEL- HWES,J. E., T. BOULINIER,J. D. NICHOLS, J. R. SAUER, HAM, AND R. I? WHITE. 1993. Genstat 5, Release AND K. H. POLLOCK. 1999. COMDYN: software 3, Reference manual. Clarendon Press, Oxford, to study the dynamics of animal communities us- UK. PREY OF THE PEREGRINE FALCON IN ARIZONA 883

PORTER, R. D., M. A. JENKINS, AND A. L. GASKI.1987. Biases in diets determined from pellets and re- Working bibliography of the Peregrine Falcon. mains: correction factors for a mammal and bird- National Wildlife Federation Scientific and Tech- eating raptor. Journal of Raptor Research 25:63- nical Series 9, Washington, DC. 67. RATCLIFFE,D. 1980. The Peregrine Falcon. Buteo SNYDER, E N. R., AND J. W. WILEY. 1976. Sexual size Books, Vermillion, SD. dimorphism in and owls of North America. REDPATH,S. M., R. CLARKE, M. MADDERS, AND S. J. Ornithological Monographs 20. THIRGOOD. 2001. Assessing raptor diet: comparing VELARDE, E. 1993. Predation of nesting larids by Per- pellets, prey remains, and observational data at egrine Falcons at Rasa Island, Gulf of California, Hen Harrier nests. Condor 103: 184-188. Mexico. Condor 95:706-708. RITCHIE,R. J. 1982. Porcupine quill and beetles in Per- WHITE, C. M., AND T. J. CADE. 1971. Cliff-nesting rap- egrine castings, Yukon River, Alaska. Raptor Re- tors and ravens along the Colville River in arctic search 1659-60. Alaska. Living Bird 10: 107-1 50. ROSENBERG,K. V., AND R. J. COOPER. 1990. Approach- WHITE, C. M., N. J. CLUM, T. J. CADE, AND W. G. es to avian diet analysis. Studies in Avian Biology HUNT. 2002. Peregrine Falcon (Falco peregrinus). 13:80-90. In A. Poole and E Gill [EDS.],The birds of North ROSENFIELD,R. N., J. W. SCHNEIDER, J. M. PAPP, AND America, No. 660. The Birds of North America, W. S. SEEGAR.1995. Prey of Peregrine Falcons Inc., Philadelphia, PA. breeding in west Greenland. Condor 97:763-770. WHITE, C. M., AND R. W. NELSON. 1991. Hunting range SHERROD,S. K. 1978. Diets of North American Fal- and strategies in a tundra breeding peregrine and coniformes. Raptor Research 12:49-121. Gyrfalcon observed from a helicopter. Journal of SIMMONS, R. E., D. M. AVERY, AND G. AVERY. 1991. Raptor Research 25:49-62.

APPENDIX. Frequency of occurrence and mass (g) for prey species from Peregrine Falcon eyries in three areas of Arizona.abc

North Central Southeast Total Mass Anatidae Northern Pintail (Anas acuta) Green-winged Teal (Anas crecca) Teal spp. (Anas spp.) Ring-necked (Aythya collaris) Odontophoridae Quail spp. (Callipepla spp.) Northern Bobwhite (Callipepla spp.) (Cyrtonyx montezumae) Threskiornithidae White-faced Ibis (Plegadis chihi) Sharp-shinned Hawk (Accipiter striatus) American Kestrel(Fa1co sparverius) Peregrine Falcon (Falco peregrinus) Rallidae ( limicola) (Porzana Carolina) American (Fulica Americana) Killdeer (Charadrius vociferous) American (Recurvirostra Americana) Scolopacidae ( JEavipes) Solitary (Tringa solitaria) 884 DAVID H. ELLIS ET AL.

APPENDIX. Continued.

North Central Southeast Total Mass Tringa spp. (Catoptrophorus semipalmatus) (Actitis macularia) Long-billed (Numenius americanus) Marbled (Limosa fedoa) (Calidris minutilla) Calidris spp. spp. (Limnodromus spp.) Common Snipe (Gallinago gallinago) Red-necked (Phalaropus lobatus) Laridae Franklin's (Larus pipixcan) Bonaparte's Gull (Larus Philadelphia) California Gull (Larus californicus) Black (Chlidonias niger) Columbidae Rock Pigeon (Columbia livia) Band-tailed Pigeon (Patagioenas fasciata) White-winged Dove (Zenaida asiatica) Mourning Dove (Zenaida macroura) Zenaida spp. Strigidae Flammulated Owl (Otus jlammeolus) (Athene cunicularia) .2 Caprimulgidae Lesser ( acutipennis) (Chordeiles minor) Common Poorwill (Phalaenoptilus nuttallii) Apodidae White-throated Swift (Aeronautes saxatalis) Trochilidae Black-chinned Hummingbird (Archilochus alexandri) Broad-tailed Hummingbird (Selasphorus platycercus) Picidae Acorn (Melanerpes formicivorus) Williamson's Sapsucker (Sphyrapicus thyroideus) Yellow-bellied Sapsucker (Sphyrapicus varius) Ladder-backed Woodpecker (Picoides scalaris) Hairy Woodpecker (Picoides villosus) Northern Flicker (Colaptes auratus) Tyrannidae Cassin's Kingbird (Tyrannus vociferans) Laniidae Loggerhead Shrike (Lanius ludovicianus) Vireonidae Plumbeous Vireo (Vireo plumbeus) Corvidae Steller's Jay (Cyanocitta stelleri) Western Scrub-Jay (Aphelocoma californica) Mexican Jay (Aphelocoma ultramarina) Pinyon Jay (Gymnorhinus cyanocephalus) Clark's Nutcracker (Nucifraga columbiana) Raven spp. ( spp.) PREY OF THE PEREGRINE FALCON IN ARIZONA 885

APPENDIX. Continued.

North Central Southeast Total Mass i Alaudidae (Eremophila alpestris) Hirundinidae Purple Martin (Progne subis) Tree (Tachycineta bicolor) Violet-green Swallow (Tachycineta thalassina) Northern Rough-winged Swallow (Stelgidopteryx serri- pennis) (Petrochelidon pyrrhonota) Paridae Poecile or Baeolophus spp. Aegithalidae Bushtit (Psaltriparus minimus) Sittidae Red-breasted Nuthatch (Sitta Canadensis) White-breasted Nuthatch (Sitta carolinensis) Pygmy Nuthatch (Sitta pygmaea) Troglodytidae Rock (Salpinctes obsoletus) Canyon Wren (Catherpes mexicanus) Bewick's Wren (Thryomanes bewickii) Turdidae (Sialia mexicana) (Sialia currucoides) Bluebird spp. (Sialia spp.) Townsend's Solitaire (Myadestes townsendi) Hermit Thrush (Catharus guttatus) American Robin (Turdus migratorius) Mimidae Northern (Mimus polyglottos) Sage (Oreoscoptes montanus) Crissal Thrasher (Toxostoma crissale) Sturnidae European Starling (Sturnus vulgaris) Parulidae Orange-crowned Warbler (Vermivora celata) Yellow-rumped Warbler (Dendroica coronata) Dendroica spp. American Redstart (Setophaga ruticilla) Yellow-breasted Chat (Icteria virens) Thraupidae Hepatic (Piranga JEava) Western Tanager (Piranga ludoviciana) Emberizidae Spotted Towhee (Pipilo maculates) Chipping Sparrow (Spizella passerina) Brewer's Sparrow (Spizella breweri) Spizella spp. Vesper Sparrow (Pooecetes gramineus) Lark Sparrow (Chondestes grammacus) Black-throated Sparrow (Amphispizu bilineata) Lark Bunting (Calarnospiza melanocorys) 886 DAVID H. ELLIS ET AL.

APPENDIX. Continued.

North Central Southeast Total Mass Melospiza spp. Dark-eyed Junco (Junco hyemalis) Cardinalidae Black-headed Grosbeak (Pheucticus melanocephalus) lndigo Bunting (Passerina cyanea) lcteridae Meadowlark spp. (Sturnella spp.) Yellow-headed Blackbird (Xanthocephalus xantho- cephalus) Brewer's Blackbird (Euphagus cyanocephalus) Common Grackle (Quiscalus quiscula) Great-tailed Grackle (Quiscalus mexicanus) Bronzed Cowbird (Molothrus aeneus) Brown-headed Cowbird (Molothrus ater) Hooded Oriole (Icterus cucullatus) Bullock's Oriole (Icterus bullockii) Scott's oriole (Icterus parisorum) Fringillidae Purple (Carpodacus purpureus) Cassin's Finch (Carpodacus cassinii) House Finch (Carpodacus mexicanus) Red Crossbill (Loxia curvirostra) Pine Siskin (Carduelis pinus) Lesser Goldfinch (Carduelis psaltria) Evening Grosbeak (Coccothraustes vespertinus) Passeridae House Sparrow (Passer domesticus) Unidentified bird Unidentified bird Unidentified bird Unidentified bird Unidentified bird Unidentified bird Unidentified bird Total birds Number of taxa 74

a Biomass data are from Dunning (1993). The following mammals were detected as prey: Colorado chipmunk (Eutamias umbrinus) once and uniden- tified chipmunk (Eutamias spp.) twice; desert cottontail (Silvilagus auduboni) once; and three other unidentified small mammals. The following insects were detected among prey remains: cicada (Cicadidae) twice; flying ant (Dolichod- erinae) once; grasshopper (Acrididae) six times. Nestling Peregrine Falcons (n = 2), included here, were not included in computations in Tables 1-3.