Diets of Insectivorous Birds Along the Colorado River in Grand Canyon, Arizona

The Condor 106:106±115 ᭧ The Cooper Ornithological Society 2004

DIETS OF INSECTIVOROUS BIRDS ALONG THE COLORADO RIVER IN GRAND CANYON, ARIZONA

HELEN K. YARD1,4,CHARLES VAN RIPER III2,BRYAN T. B ROWN3 AND MICHAEL J. KEARSLEY1 1Department of Biology, Northern Arizona University, Flagstaff, AZ 86011 2USGS Southwest Biological Science Center, Colorado Plateau Field Station, P.O. Box 5614, Northern Arizona University, Flagstaff, AZ 86011 3SWCA Environmental Consultants, Inc., 230 South 500 East, Suite 380, Salt Lake City, UT 84102

Abstract. We examined diets of six insectivorous bird species (n ϭ 202 individuals) from two vegetation zones along the Colorado River in Grand Canyon National Park, Arizona, 1994. All bird species consumed similar quantities of caterpillars and beetles, but use of other prey taxa varied. Non-native leafhoppers (Opsius stactagolus) speci®c to non-native tamarisk (Tamarix chinensis) substantially augmented Lucy's Warbler (Vermivora luciae) diets (49%), while ants comprised 82% of Yellow-breasted Chat (Icteria virens) diets. Yel- low Warbler (Dendroica petechia) diets were composed of 45% aquatic midges. All bird species consumed the non-native leafhopper speci®c to tamarisk. Comparison of bird diets with availability of arthropod prey from aquatic and terrestrial origins showed terrestrial insects comprised 91% of all avian diets compared to 9% of prey from aquatic origin. Seasonal shifts in arthropod prey occurred in diets of three bird species, although no seasonal shifts were detected in arthropods sampled in vegetation indicating that at least three bird species were not selecting prey in proportion to its abundance. All bird species had higher prey overlap with arthropods collected in the native, mesquite-acacia vegetation zone which contained higher arthropod diversity and better prey items (i.e., Lepidoptera). Lucy's Warbler and Yellow Warbler consumed high proportions of prey items found in greatest abundance in the tamarisk-dominated vegetation zone that has been established since the construction of Glen Canyon Dam. These species appeared to exhibit ecological plasticity in response to an anthropogenic increase in prey resources. Key words: anthropogenic, arthropods, avian diets, Colorado River, Grand Canyon Na- tional Park, insectivorous birds, Neotropical migrants.

Dieta de Aves InsectõÂvoras a lo largo del RõÂo Colorado en el Gran CanÄon, Arizona Resumen. Examinamos la dieta de seis especies de aves insectõÂvoras (n ϭ 202 indivi- duos) de dos zonas de vegetacioÂn a lo largo del RõÂo Colorado en el Parque Nacional del Gran CanÄon, Arizona, en 1994. Todas las especies de aves consumieron cantidades similares de orugas y escarabajos, pero el uso de otras presas fue variable. Los Cicadellidae (HomoÂptera) exoÂticos (Opsius stactagolus) especõÂ®cos del tamarisco exoÂtico (Tamarix chinensis) comprendieron una parte sustancial de la dieta de Vermivora luciae (49%), mientras que las hormigas representaron el 82% de la dieta de Icteria virens. La dieta de Dendroica petechia incluyoÂ un 45% de dõÂpteros acuaÂticos de la familia Cecidomyiidae. Todas las especies de aves consumieron el cicadellide exoÂtico especõÂ®co del tamarisco. La compara- cioÂn de las dietas de aves con la disponibilidad de presas de artroÂpodos de origen acuaÂtico y terrestre mostroÂ que los insectos terrestres comprendieron el 91% de todas las dietas de aves, comparado con un 9% de presas de origen acuaÂtico. Se registraron cambios estacionales en los artroÂpodos de la dieta de tres especies de aves, aunque no se detectaron cambios estacionales en los artroÂpodos muestreados en la vegetacioÂn, indicando que al menos tres especies de aves no estuvieron seleccionando las presas en proporcioÂn a su abundancia. Todas las especies de aves presentaron mayor superposicioÂn de presas con los artroÂpodos colectados en la zona de vegetacioÂn nativa de mesquite (Prosopis)yAcacia, la cual contuvo mayor diversidad de artroÂpodos y presas de mejor calidad (i.e., Lepidoptera). V. luciae y D. petechia consumieron altas proporciones de presas encontradas en mayor abundancia en la zona de vegetacioÂn dominada por el tamarisco, que ha sido establecida desde la construccioÂn de la Presa Glen Canyon. Estas especies parecieron exhibir plasticidad ecoloÂgica en res- puesta a un incremento antropogeÂnico en las presas como recuso alimenticio.

Manuscript received 22 November 2002; accepted 18 September 2003. 4 Present address: 2720 W. Lynette Drive, Flagstaff, AZ 86001. E-mail: [email protected]

[106] INSECTIVOROUS BIRD DIETS IN GRAND CANYON, ARIZONA 107

INTRODUCTION in¯uencing arthropod composition along the Food exploitation strategies are central to our Colorado River after the completion of the dam knowledge of habitat use by birds, yet direct ex- was the invasion of non-native tamarisk (Ta- marix chinensis), now the most common shrub amination of Neotropical migratory bird diets is found in the vegetation zone established along generally underrepresented in avian ecology the river's edge. Tamarisk supports a high abun- studies. This is especially true of riparian areas dance of a host-speci®c, non-native leafhopper throughout the southwestern United States (Opsius stactogalus; Stevens 1985; Yard and where arthropods on vegetation adjacent to riv- Cobb, unpubl. data), yet it is unknown if these ers and streams are important food resources for changes to the arthropod community have af- insectivorous birds (Rosenberg et al. 1991, Gray fected the way birds forage as no previous stud- 1993). Previous studies have found that arthro- ies have examined diets of insectivorous birds pods from aquatic origins play a key role as prey along the Colorado River in Grand Canyon. for insectivorous birds in riparian habitats (Jack- The goals of our study were to quantify diets son and Fisher 1986, Gray 1993). Aquatic midg- of six insectivorous, Neotropical migrant pas- es (Chironomidae), for example, have been dem- serines, to determine if seasonal dietary shifts onstrated to be an important prey item for some occurred between spring (March±May) and riparian bird species (Busby and Sealy 1978). summer (June±July), and to examine arthropod Riparian habitat in the southwestern United availability between vegetation zones and sea- States supports a disproportionately high density sons. We compared proportions of prey taxa and diversity of birds (Johnson et al. 1977, Ro- found in stomach samples among six bird spe- senberg et al. 1982). Szaro and Jakle (1985) cies, determined the origin of the prey (aquatic have shown that southwestern riparian wood- versus terrestrial) consumed by each bird spe- lands may support up to 10 times the number of cies, and qualitatively related prey items to ar- Ϫ1 birds ha when compared to adjacent upland thropod availability in the tamarisk-dominated habitats due primarily to greater availability of versus the mesquite-acacia zone. This informa- food and cover. Before the completion of Glen tion may provide insight into the in¯uence of Canyon Dam, riparian habitat along the Colo- anthropogenic change on avian ecology in dam- rado River in Grand Canyon, Arizona, was controlled river systems and highlight manage- sparse (Turner and Karpiscak 1980). Riparian ment options which could affect birds. breeding bird abundance in the last 40 years has increased commensurate with vegetation chang- METHODS es along the Colorado River through Grand Can- yon (Carothers and Brown 1991). Following STUDY AREA dam construction, more than 500 ha of new ri- The Colorado River ¯ows 478 km through parian vegetation has been distributed linearly Grand Canyon National Park, Arizona, starting along the shores of the Colorado River from the in the northeast corner of the park near Glen dam to Lake Mead (478 km, Brown and Trosset Canyon Dam on the Arizona-Utah border and 1989). stretching to Lake Mead in northwestern Ari- Two changes in arthropod availability have zona. The postdam vegetation zone adjacent to resulted from the completion of Glen Canyon the river is composed predominantly of non-na- Dam. Prior to dam construction, an abundant tive tamarisk (the tamarisk zone). The predam and diverse assemblage of aquatic arthropods vegetation zone adjacent to and upslope from was found in the river (Ward and Stanford the tamarisk zone is predominantly native honey 1979). Presently, midges are the most abundant mesquite (Prosopis glandulosa) and catclaw arthropod of aquatic origin emerging from the acacia (Acacia greggii; Carothers and Brown river (Shannon 1993). Aquatic diversity in and 1991; the mesquite-acacia zone). emerging from the river is presently low when We selected four representative study sites at compared to unregulated rivers (Blinn and Cole intervals along the river to document bird diets 1991). Postdam regulated water ¯ows have all and to estimate arthropod prey availability: Riv- but eliminated seasonal variability in water tem- er km (rk) 1.6 (Paria Creek) was entirely dom- perature and restricted sediment loads affecting inated by tamarisk (2.3 ha, mesquite-acacia ϭ 0 aquatic arthropod productivity. Another factor ha). The other three sites contained both vege- 108 HELEN K. YARD ET AL. tation zones: rk 75.1 (Saddle Camp; tamarisk ϭ variable-power dissecting microscope (Borror et 3.4 ha, mesquite-acacia ϭ 0.8 ha), rk 318.6 (Pa- al. 1989). Only stomach samples containing Ն5 rashant Camp; tamarisk ϭ 1.5 ha, mesquite-aca- items were used for analyses. A possible bias cia ϭ 1.4 ha), and rk 329.0 (Spring Canyon; may have been that birds with one or a few large tamarisk ϭ 1.8 ha, mesquite-acacia ϭ 2.5 ha). prey items were omitted from the analysis. Min- Eight to 10 mist nets (36-mm mesh, 12 m) were imum numbers of prey individuals in each stom- erected at each site for 2 days per month from ach sample were determined from diagnostic late March±July 1994 in each vegetation zone. fragments (e.g., head capsules, caterpillar man- The two vegetation zones were immediately ad- dibles, elytra, wings) and compared with our ref- jacent to each other with no discernible space or erence collection (Chapman and Rosenberg barrier between them. Logistical constraints lim- 1991, Sillett 1994). Proportions of the number ited us to 10 days of bird netting and insect col- of prey items found in stomach contents of each lections at each site. bird species pooled were calculated for the following nine taxa categories: Araneae (spiders); COLLECTION TECHNIQUES FOR DIET Hemiptera (bugs); Homoptera (primarily leaf- AND ARTHROPODS hoppers); Coleoptera (beetles); Diptera (¯ies and We selected six insectivorous passerine bird spe- aquatic midges); Hymenoptera (subclassi®ed as cies for dietary analysis: Ash-throated Flycatch- either wasps or ants); Lepidoptera (primarily er (Myiarchus cinerascens), Bewick's Wren caterpillars) and other (Thysanoptera, Neurop- (Thryomanes bewickii), Bell's Vireo (Vireo be- tera, Acari, and unknown). To determine if die- llii), Lucy's Warbler (Vermivora luciae), Yellow tary shifts occurred within the breeding season, Warbler (Dendroica petechia), and Yellow- we compared diet information collected from the breasted Chat (Icteria virens). These six species six bird species in spring (March, April, and are among the most abundant breeding riparian May) with that collected in summer (June and birds along the Colorado River in Grand Canyon July). (Brown et al. 1987) and elsewhere in the South- west. Immediately upon capture, stomach con- STATISTICAL ANALYSES tents were obtained by lavage (Moody 1970, We used multivariate analysis of variance (MA- Laursen 1978, Rosenberg and Cooper 1990). A NOVA) to test for overall difference in mean 6-cm-long plastic tubing was slowly inserted arthropod proportions between the two vegeta- into the bird's beak and then guided gently down tion zones. Arthropods collected were pooled by the esophagus. A syringe attached to the tubing vegetation zone for all sites, and proportions released saline solution into the bird's stomach, were arcsine transformed to meet normality as- causing the bird to regurgitate its stomach con- sumptions (Sokal and Rohlf 1995, SPSS Inc. tents. Stomach contents were collected in a plas- 2000). The Paria site was excluded from be- tic dish, then transferred into a glass vial and tween-zone analysis because it had no mesquite- preserved with 70% ethyl alcohol. acacia vegetation zone. One-way analysis of Arthropods were collected from the tamarisk variance (ANOVA) was used to test for differ- and mesquite-acacia vegetation zones one day ences in mean proportions of prey categories each month, concurrent with bird diet-sample present between vegetation zones (Sokal and collections. We used sweep-net and beat-sheet Rohlf 1995, SPSS Inc. 2000). We also used an sampling to obtain representative collections of analysis of similarity (ANOSIM; Clarke 1993) arthropods present in each vegetation zone to compare arthropod prey items in each zone (Cooper and Whitmore 1990). Collections were between seasons. This nonparametric, permuta- made by taking 25 sweeps in the vegetation tion-based procedure compares mean ranks of branches with a standard 37-cm-diameter net, at dissimilarities of samples within and between 1±3 m above ground. On undisturbed adjacent groups. When groups of samples are distinct vegetation, we beat 25 branches to dislodge ar- from each other, the compositional dissimilari- thropods onto a beating canvas. Arthropods col- ties between samples within a group are smaller lected were preserved in 70% ethyl alcohol. than dissimilarities between samples from dif- Prey items from stomach contents and vege- ferent groups. The ANOSIM test statistic, R, tation sampling were all identi®ed to order, then varies between Ϫ1 and 1, reaching its maximum to the lowest taxonomic level possible using a value when all between-group dissimilarities are INSECTIVOROUS BIRD DIETS IN GRAND CANYON, ARIZONA 109 greater than all within-group dissimilarities. Sta- ity of a Type I error is the proportion of runs tistical signi®cance is determined by comparing from random data with higher IndVal scores the sample R with those produced by randomly than were found in the actual data set. DufreÃne assigning samples to groups. The proportion of and Legendre (1997) set a cutoff of greater than random arrangements with R-values higher than 25, which we followed, with the additional con- the sample value is the signi®cance level of the straint that the probability of the indicator value test (Clarke and Gorley 2001). The two-way lay- be less than 0.10. Diet samples for each species out is described in Clarke and Warwick (2001). pooled were divided into two groups based on MANOVA was used to compare proportions period (spring and summer). Statistical signi®- of prey taxa found in the diet samples among cance was accepted at P Յ 0.10. bird species (stomach samples were pooled for each species) and to compare proportions of ar- DIET AND ARTHROPOD thropods between vegetation zones. All propor- AVAILABILITY OVERLAP tions were arcsine transformed to meet normal- Foraging observations were not conducted; ity assumptions. We used ANOVA to test for therefore we could not determine foraging lo- differences in mean proportions of prey cate- cations for birds. Indices of overlap were used gories present within each bird species and be- as an indicator of foraging location or zone of tween vegetation zones (Sokal and Rohlf 1995). foraging preference (tamarisk or mesquite-aca- A nonparametric Kruskal-Wallis test was used cia) used by the six bird species. These quali- to compare mean ranks of percentages of aquatic tative indices have proven useful to ecologists arthropods found in diets of the six bird species in comparative studies of diet and habitat pref- (Sokal and Rohlf 1995, SPSS Inc. 2000). erence as well as in descriptions of dietary sim- We used an ANOSIM to compare arthropod ilarity between bird species (Horn 1966). We as- prey items in diet samples between seasons to sessed foraging preference using Pianka's (1974) identify possible seasonal shifts in the diets of index for overlap: three bird species with adequate sample sizes. Diet composition data were divided between 22 Oa ϭ (PPia ja)΋ͱ΂΃΂΃PPia ja , spring and summer. Dissimilarities between ͸͸͸ samples were calculated using the Bray-Curtis where Pia and Pja are the proportions of prey distance because it handles compositional infor- category a in the diets of each bird species i and mation well (Faith et al. 1987). We performed the habitat j respectively. We made the assump- an ANOSIM with periods as treatments for Be- tion that birds were foraging in relation to the wick's Wren, Bell's Vireo, and Lucy's Warbler. availability of prey in each habitat. The three remaining species were excluded from RESULTS analyses because few or no diet samples were obtained during spring (Clarke and Gorley ARTHROPOD ABUNDANCE 2001.) We found a signi®cant difference in mean pro- To determine whether individual prey taxa portions of arthropods collected in tamarisk ver- were differentiated in the spring and summer di- sus mesquite-acacia zones (MANOVA: Wilks' ets of each bird species, we used indicator spe- lambda F1.64 ϭ 11.7, P Ͻ 0.01; Fig. 1). Flies and cies analysis (DufreÃne and Legendre 1997). The midges were collected in overall highest fre- method combines information on the speci®city quency followed by leafhoppers, ants, caterpil- and ®delity of prey taxa to one time period or lars, spiders, hemipterans, beetles, and wasps. another to calculate the test statistic, IndVal, ANOVA results showed higher proportions of which ranges from 0 to 100. Speci®city is de- ¯ies and midges (F1,64 ϭ 16.8, P Ͻ 0.01) and ®ned as the average abundance of a prey taxon leafhoppers (F1,64 ϭ 8.3, P Ͻ 0.01) in the tam- per sample in a group compared to that in other arisk vegetation zone. Higher proportions of groups. The ®delity of a taxon is de®ned as the beetles occurred in the mesquite-acacia zone proportion of samples within a group which con- (F1,64 ϭ 21.6, P Ͻ 0.01). Mean proportions of tain the taxon. Values for IndVal from sample caterpillars, hemipterans, ants, spiders, and data are compared to those calculated from ran- wasps were not signi®cantly different between domly assigning samples to groups (Monte Car- vegetation zones. No seasonal difference was lo test of signi®cance, 999 trials). The probabil- detected between arthropods in each zone be- 110 HELEN K. YARD ET AL.

FIGURE 1. Proportion Ϯ SE of arthropod taxa collected in the tamarisk and mesquite-acacia vegetation zones along the Colorado River, Grand Canyon, Arizona, spring and summer 1994. tween spring and summer (ANOSIM, R ϭ 0.1, Bewick's Wren a higher mean proportion of spi-

P ϭ 0.20). ders (F5,195 ϭ 2.6, P ϭ 0.01), Bell's Vireo more hemipterans (F5,195 ϭ 5.3, P Ͻ 0.01), Lucy's BIRD DIET Warbler more leafhoppers (F5,195 ϭ 5.9, P Ͻ Arthropod prey items were identi®ed from 202 0.01), Yellow Warbler more diptera (¯ies and bird diet samples, with 98% identi®ed to order. midges, F5,195 ϭ 2.9, P ϭ 0.01), and Yellow- The ``other'' category was excluded as it rep- breasted Chat more ants (F5,195 ϭ 20.8, P Ͻ 0.01, resented only 2% of prey items. A signi®cant Table 1). All six bird species consumed similar difference in prey items was found among diets proportions of beetles (F5,195 ϭ 2.2, P ϭ 0.06) of the six bird species (MANOVA: Wilks' lamb- and caterpillars (F5,195 ϭ 2.2, P ϭ 0.06). All six da F5,195 ϭ 5.0, P Ͻ 0.01). ANOVA showed that bird species consumed non-native leafhoppers. Ash-throated Flycatchers consumed an overall We classi®ed 95% of arthropods from stom- higher proportion of wasps when compared to ach samples as aquatic or terrestrial in origin. the other ®ve species (F5,195 ϭ 8.0, P Ͻ 0.01); When diets of all six bird species were pooled,

TABLE 1. Bird species, sample size, number of prey items, and mean percent Ϯ SE of prey items (in paren- theses) in diet samples of six insectivores (n ϭ 202 individuals) along the Colorado River in Grand Canyon, Arizona, spring and summer 1994.

Ash-throated Lucy's Yellow Arthropod Flycatcher Bewick's Wren Bell's Vireo Warbler Warbler Yellow-breasted orders (n ϭ 17) (n ϭ 33) (n ϭ 39) (n ϭ 77) (n ϭ 18) Chat (n ϭ 18) Araneae 15 (21 Ϯ 3) 54 (28 Ϯ 5) 28 (12 Ϯ 2) 76 (18 Ϯ 2) 12 (8 Ϯ 3) 16 (14 Ϯ 3) Hemiptera 3(3Ϯ 2) 14 (6 Ϯ 2) 98 (17 Ϯ 3) 61 (7 Ϯ 2) 27 (4 Ϯ 2) 7(4Ϯ 2) Homoptera 9 (11 Ϯ 5) 76 (16 Ϯ 3) 20 (7 Ϯ 5) 555 (27 Ϯ 3) 32 (12 Ϯ 4) 2(1Ϯ 1) Coleoptera 7(9Ϯ 3) 41 (13 Ϯ 2) 82 (23 Ϯ 3) 202 (20 Ϯ 2) 40 (15 Ϯ 4) 18 (14 Ϯ 3) Diptera 8 (10 Ϯ 4) 43 (10 Ϯ 3) 23 (12 Ϯ 4) 56 (6 Ϯ 1) 173 (24 Ϯ 6) 4(3Ϯ 1) Hymenoptera Wasps 22 (26 Ϯ 7) 18 (4 Ϯ 1) 40 (8 Ϯ 2) 70 (6 Ϯ 1) 73 (25 Ϯ 5) 18 (13 Ϯ 4) Ants 6 (7 Ϯ 3) 57 (13 Ϯ 3) 15 (4 Ϯ 1) 33 (7 Ϯ 2) 10 (5 Ϯ 2) 343 (42 Ϯ 7) Lepidoptera 10 (13 Ϯ 4) 36 (10 Ϯ 2) 80 (17 Ϯ 2) 75 (9 Ϯ 1) 18 (7 Ϯ 4) 12 (9 Ϯ 2) INSECTIVOROUS BIRD DIETS IN GRAND CANYON, ARIZONA 111

summer (n ϭ 23). Bell's Vireo averaged approx- imately ®ve times more hemipterans in their diet during spring than in summer (n ϭ 12, IndVal ϭ 63, P ϭ 0.001) but twice as many ants during summer (n ϭ 27, IndVal ϭ 29, P ϭ 0.09). Lucy's Warbler diets contained a signi®cantly higher number of caterpillars in the spring than in summer (n ϭ 22, IndVal ϭ 22, P ϭ 0.001), with twice as many ants (n ϭ 55, IndVal ϭ 33, P ϭ 0.02), and ®ve times as many leafhoppers (IndVal ϭ 71, P ϭ 0.01) in their diets during the summer period, leafhoppers being the most FIGURE 2. Mean Ϯ SE percent of aquatic arthropods in bird diets (n ϭ 202) along the Colorado River, abundant prey item in their diet. Grand Canyon, Arizona, in spring and summer 1994. ATFL ϭ Ash-throated Flycatcher; BEWR ϭ Bewick's DIET AND ARTHROPOD Wren; BEVI ϭ Bell's Vireo; LUWA ϭ Lucy's War- AVAILABILITY OVERLAP bler; YWAR ϭ Yellow Warbler; YBCH ϭ Yellow- Overlap index values for arthropods in the diets breasted Chat. of the six bird species more closely matched arthropods collected in the mesquite-acacia zone the resulting combination exhibited 91% terres- than the tamarisk zone. Arthropods in all six trial and 9% aquatic arthropods. The mean per- species had overlap values of 0.77 or greater centage of arthropods from aquatic origins dif- with the mesquite-acacia zone (Table 2). The 2 Yellow Warbler had the highest overlap with ar- fered among bird diets (Kruskal-Wallis, ␹ 5 ϭ 18.5, P Ͻ 0.01; Fig. 2). Yellow Warblers had a thropods in the tamarisk zone when compared higher mean percent of arthropods from aquatic with the other ®ve species of birds (0.77). origin (16% Chironomidae [midges]) than the other ®ve bird species. Midges comprised 45% DISCUSSION of their total diet. The six bird species we studied consumed sim- The diet similarity analyses revealed signi®- ilar proportions of caterpillars and beetles but cant seasonal shifts in diets of the three species appeared to partition other prey, with each bird most adequately sampled. Bewick's Wren diets species consuming a higher proportion of one differed between spring and summer (R ϭ 0.18, particular prey taxon. The similar proportions of P ϭ 0.02), as did Bell's Vireo (R ϭ 0.11, P ϭ caterpillars and beetles found in the diets of six 0.03), and Lucy's Warbler (R ϭ 0.22, P ϭ 0.01). insectivores are consistent with dietary data Indicator species analysis revealed Bewick's from birds on the Lower Colorado River (Ro- Wren diets were not characterized by any one senberg et al. 1991) and studies conducted on particular prey item in either spring (n ϭ 10) or migrant species in the tropics (Sillett 1994, Pou- lin et al. 1994, Poulin and Lefebvre 1996). Cat- erpillars have been reported to be the single TABLE 2. Overlap values (Pianka 1974) between ar- most important prey item for breeding birds, as thropods found in diets of six bird species and arthropods sampled from foliage in two vegetation zones caterpillars are markedly seasonal and most along the Colorado River, Grand Canyon, Arizona, abundant in spring and summer (Thiollay 1988, spring and summer 1994. A value of 1.0 represents Greenberg 1995). Productivity of breeding birds 100% overlap between diets and available arthropods. is a function of the abundance of large, soft- bodied arthropods (i.e., caterpillars) identi®ed as Mesquite- ``breeding currency'' which are needed as food acacia Tamarisk Bird species zone zone for growing nestlings (Greenberg 1995). Dietary differences may partially be explained Ash-throated Flycatcher 0.77 0.58 Bewick's Wren 0.86 0.63 by different bill sizes and shapes among the six Bell's Vireo 0.89 0.53 passerines since bill size was likely analogous Lucy's Warbler 0.84 0.61 to prey size (Newton 1972, Schoener 1974, Yellow Warbler 0.81 0.77 Wiens and Rotenberry 1987). Other possible ex- Yellow-breasted Chat 0.78 0.33 planations for dietary differences may have in- 112 HELEN K. YARD ET AL. cluded differences in foraging height, use of the breeding season and are perhaps very im- plant species, habitat selection, and competition portant food for adults feeding ¯edglings and for (MacArthur 1958, Morse 1968, Perrins and ¯edglings learning to capture prey (Lucy's War- Birkhead 1983). Foraging tactics must also be bler, Johnson et al. 1997). We found ants to be considered when examining dietary differences most abundant in the mesquite-acacia zone. Ants among species. Gleaners (Bell's Vireo, Bewick's were reported in relatively high proportions in Wren, Lucy's Warbler, and Yellow-breasted diets of migrant Neotropical species sampled in Chat) all preyed primarily on arthropods found the tropics (Poulin and Lefebvre 1996) and may on vegetation (spiders, beetles, leafhoppers, he- be an important year-round prey item for some mipterans, and ants). Yellow Warblers consumed migrant birds. a high proportion of aquatic midges and could Our data indicate that anthropogenic changes have gleaned them from vegetation or captured to the Colorado River ecosystem downstream of them aerially by sallying from high perches. Glen Canyon Dam have resulted in an increased Ash-throated Flycatchers used aerial foraging abundance of two arthropod prey taxa: tamarisk- tactics that would explain their higher consump- speci®c leafhoppers and aquatic midges. Midg- tion of wasps and bees. These ®ndings are con- es, a food source not abundant before the dam sistent with Ash-throated Flycatcher diets in was built but which are the most common prey California, where wasps and bees were the most type in the tamarisk-dominated zone, comprised prominent prey item (Cardiff and Dittmann 45% of the total number of prey items consumed 2002). by the Yellow Warbler (mean 16 Ϯ 3% per in- Arthropods of terrestrial origin provided the dividual). An unexpected ®nding of our study primary food base for ®ve of the six riparian was that all diet samples from the six bird spe- bird species, comprising 91% of their cumula- cies contained varying proportions of the non- tive diets in contrast to only 9% of prey items native leafhopper, a prey item found exclusively of aquatic origin. The exception was the Yellow on tamarisk (Carothers and Brown 1991). Leaf- Warbler, whose total stomach samples were 45% hopper presence in bird stomachs was likely re- aquatic midges (mean 16% per individual). lated to the increased availability of these abun- Aquatic midges were the most abundant arthro- dant and easy to detect insects. Lucy's Warblers, pod collected in the tamarisk during the breed- whose total diet consisted of 49% leafhoppers ing season. The ®nding that Yellow Warbler (mean 27 Ϯ 3% per individual), have recently consumed a high percentage of midges is con- increased during the breeding season along the sistent with dietary information collected in Ca- Colorado River (Brown et al. 1987, Felley and nada (Busby and Sealy 1978) and in Utah, Sogge 1997). This increase may be directly re- where midges were among the most frequent lated to the increased abundance of this non-na- prey items found in diet samples (Frydendall tive leafhopper. Yellow Warbler and Lucy's War- 1967). bler diets directly re¯ected arthropod abundance A high percentage of ants (82% total, mean in the tamarisk vegetation zone, while the other 42 Ϯ 7% per individual) was found in Yellow- four bird species did not. breasted Chat stomach samples, much greater All six bird species showed higher overlap than suggested in existing literature. Few quan- values of prey items with arthropods collected titative data are available on chat diets; however from the mesquite-acacia zone. Ongoing studies ants occurred in 11 out of 14 stomachs collected have found a higher diversity of arthropods and from Yellow-breasted Chats at various locations a higher proportion of pro®table prey types, throughout North America (although quantities those considered to be better ``breeding curren- were not reported; Eckerle and Thompson cy'' (large soft-bodied arthropods, Greenberg 2001). Chat stomach samples from the Lower 1995) in the native mesquite-acacia zone (Yard Colorado River (n ϭ 4) contained ants as one of and Cobb, unpubl. data). In studies conducted four prey items, but again, proportions were not on arthropods and bird abundance in acacia ver- reported (Rosenberg et al. 1991). sus adjacent vegetation in the tropics, a higher Leafhoppers and ants were found in highest abundance of foraging migrant birds was found proportions in diets of two bird species during in acacia presumably due to better quality prey summer. Leafhoppers were the most common availability (Greenberg et al. 1997, Greig-Smith prey type in the tamarisk vegetation zone during 1978). South African acacia is reported to have INSECTIVOROUS BIRD DIETS IN GRAND CANYON, ARIZONA 113 high palatability to arthropods (Coe and Coe of data collection and (2) concurrent foraging 1987, Cooper et al. 1988). In addition to foliage studies to determine where prey are obtained. palatability for arthropods, Coe and Coe (1987) These recommendations would allow a more reported that leguminous seeds produced by aca- complete understanding of dietary patterns and cia supported high numbers of beetles (Bruchi- seasonal differences. dae). Similarly, we also found high proportions of Bruchidae in both mesquite and acacia (Yard ACKNOWLEDGMENTS and Cobb, unpubl. data) and in most bird diet This study would not have been possible without fund- samples (Table 1). ing and logistical support from the U.S. Bureau of Examples of ecological plasticity or the ten- Reclamation, Glen Canyon Environmental Studies, and David Wegner. We thank the USGS Southwest dency of bird species to exploit new, abundant Science Center Colorado Plateau Field Station and food resources (Greenberg 1990), illustrate how Northern Arizona University, especially Mark K. Sog- dietary opportunism has the potential to enhance ge and Russell P. Balda, for providing valuable support or maintain ®tness in certain passerines. In- and guidance during this study. We also thank the fol- creased abundance of breeding birds, speci®cal- lowing individuals who contributed to arthropod identi®cation, sample collections, mist netting, and trip lo- ly Lucy's and Yellow Warblers, along the Col- gistics and organization: Neil Cobb, Charles Drost, orado River in Grand Canyon may be related to Patti Hodgetts, David Felley, and Michael Yard. John their ability to exploit new food resources estab- Blake, David Dobkin, Bette Loiselle, and Scott Sillett lished because of Glen Canyon Dam's alteration made productive comments on this manuscript. of the ecosystem. Our data show both vegetation LITERATURE CITED zones have much to offer insectivorous birds in terms of arthropod prey. The tamarisk zone ap- ANDERSON,B.W.,AND R. D. OHMART. 1977. Vegeta- pears more useful than previously thought, but tion structure and bird use in the lower Colorado the mesquite-acacia zone is clearly more impor- River Valley, p. 23±33. In R. R. Johnson and D. A. Jones [TECH. COORDS.], Importance, preserva- tant when considering all bird species together. tion and management of riparian habitat: a sym- Tamarisk, though low in arthropod diversity, posium. USDA Forest Service General Technical offers two new and abundant prey resources as Report RM-GTR-43. well as cover for nesting birds (Brown and Tros- BLINN,D.W.,AND G. A. COLE. 1991. Algal and invertebrate biota in the Colorado River: compari- set 1989). The salty exudate from tamarisk sons of pre- and post-dam conditions, p. 102±123. leaves has been reported as a deterrent to ar- In G. R. Marzolf [ED.], Colorado river ecology and thropod diversity, while native vegetation has dam management, proceedings of a symposium, been suggested to exhibit a more diverse arthro- May 24±25, 1990. National Academy Press, pod assemblage and better bird food (i.e., large, Washington, DC. BORROR, D. J., C. A. TRIPLEHORN, AND N. F. JOHNSON. soft bodied insects) for foraging birds (Anderson 1989. An introduction to the study of insects. 6th and Ohmart 1977). Management decisions for ed. Saunders College Publishing, Philadelphia, riparian vegetation should consider the impor- PA. tance of both vegetation zones to foraging Neo- BROWN, B. T., S. W. CAROTHERS, AND R. R. JOHNSON. 1987. Grand Canyon birds. University of Arizona tropical birds. Present tamarisk eradication pro- Press, Tucson, AZ. jects throughout the Southwest have the poten- BROWN,B.T.,AND M. W. TROSSET. 1989. Nest-habitat tial for negative effects on at least two species relationships of riparian birds along the Colorado of Neotropical migrant birds, unless replacement River in Grand Canyon, Arizona. Southwestern by native woody vegetation is rapid. Naturalist 34:260±270. BUSBY, D. G., AND S. G. SEALY. 1978. Feeding ecology Two bird species examined for seasonal die- of a population of nesting Yellow Warblers. Ca- tary shifts showed signi®cant differences in prey nadian Journal of Zoology 57:1670±1681. items found in stomach contents when compared CARDIFF, S. S., AND D. L. DITTMANN. 2002. Ash-throat- between the spring and summer periods. In con- ed Flycatcher (Myiarchus cinerascens). In A. Poole and F. Gill [EDS.], The birds of North Amer- trast, arthropods collected in the vegetation did ica, No. 664. The Birds of North America, Inc., not re¯ect seasonal shifts in abundance. This Philadelphia, PA. clearly demonstrates that two bird species were CAROTHERS,S.W.,AND B. T. BROWN. 1991. The Col- seeking different prey items seasonally and that orado River through Grand Canyon. University of their prey selection was not based on prey abun- Arizona Press, Tucson, AZ. CHAPMAN, A., AND K. ROSENBERG. 1991. Diets of four dance. Future dietary studies of Southwestern ri- sympatric Amazonian woodcreepers (Dendroco- parian birds should consider (1) multiple years laptidae). Condor 93:904±915. 114 HELEN K. YARD ET AL.

CLARKE, K. R. 1993. Non-parametric multivariate anal- JACKSON, J. K., AND S. G. FISHER. 1986. Secondary yses of changes in community structure. Austra- production, emergence and export of aquatic in- lian Journal of Ecology 18:117±143. sects of a Sonoran Desert stream. Ecology 67: CLARKE,K.R.,AND R. N. GORLEY. 2001. Primer v5: 629±594. user manual and tutorial. PRIMER-E Ltd., Plym- JOHNSON, R. R., L. T. HAIGHT, AND J. M. SIMPSON. outh, UK. 1977. Endangered species vs. endangered habitats: CLARKE,K.R.,AND R. M. WARWICK. 2001. Change in a concept, p. 68±74. In R. R. Johnson and D. A. marine communities: an approach to statistical Jones [TECH. COORDS.], Importance, preservation analysis and interpretation. 2nd ed. PRIMER-E, and management of riparian habitat: a symposium. Plymouth, UK. USDA Forest Service General Technical Report COE, M., AND C. COE. 1987. Large herbivores, acacia RM-GTR-43 trees and bruchid beetles. South African Journal JOHNSON, R. R., H. K. YARD, AND B. T. BROWN. 1997. of Science 83:624±635. Lucy's Warbler (Vermivora luciae). In A. Poole COOPER,J.C.,AND R. C. WHITMORE. 1990. Arthropod and F. Gill [EDS.], The birds of North America, sampling methods in ornithology. Studies in Avi- No. 664. The Birds of North America Inc., Phil- an Biology 13:29±37. adelphia, PA. COOPER, S. M., N. OWEN-SMITH, AND J. P. BRYANT. LAURSEN, K. 1978. Interspeci®c relationships between 1988. Foliage acceptability to browsing ruminants some insectivorous passerine species illustrated by in relation to seasonal changes in the leaf chem- their diet during spring migration. Ornis Scandi- istry of woody plants in a South African savanna. navica 9:178±192. Oecologia 75:336±442. MACARTHUR, R. H. 1958. Population ecology of some DUFREÃ NE, M., AND P. L EGENDRE. 1997. Species assem- warblers of northeastern coniferous forests. Ecol- blages and indicator species: the need for a ¯ex- ogy 39:599±619. ible assymetric approach. Ecological Monographs MOODY, D. T. 1970. A method for obtaining food sam- 67:345±366. ples from insectivorous birds. Auk 87:579. ECKERLE,K.P.,AND C. F. THOMPSON. 2001. Yellow- MORSE, D. H. 1968. A quantitative study of foraging breasted Chat (Icteria virens). In A. Poole and F. of male and female spruce-wood warblers. Ecol- Gill [EDS.], The birds of North America, No. 575. ogy 49:779±784. The Birds of North America, Inc., Philadelphia, PA. NEWTON, I. 1972. Finches. Collins Publishing, London. FAITH, D. P., P. R. MINCHIN, AND L. BELBIN. 1987. PERRINS, C. M., AND T. R. BIRKHEAD. 1983. Avian ecol- Compositional dissimilarity as a robust measure ogy. Blackie and Sons Ltd., Glasgow. of ecological distance. Vegetatio 69:57±68. PIANKA, E. R. 1974. Niche overlap and diffuse com- FELLEY, D. L., AND M. K. SOGGE. 1997. Comparison petition. Proceedings of the National Academy of of techniques for monitoring riparian birds in Sciences 71:2141±2145. Grand Canyon National Park, p. 73±84. In C. van POULIN, B., AND G. LEFEBVRE. 1996. Dietary relation- Riper III and E. T. Deshler [EDS.], Proceedings of ships of migrant and resident birds from a humid the third biennial conference of research on the forest in central Panama. Auk 113:277±287. Colorado Plateau, National Park Service Trans- POULIN, B., G. LEFEBVRE, AND R. MCNEIL. 1994. Diets actions and Proceedings Series NPS/NRNAU/ of land birds from northeastern Venezuela. Con- NRTP-97/12, Flagstaff, AZ. dor 96:354±367. FRYDENDALL, M. J. 1967. Feeding ecology and terri- ROSENBERG,K.V.,AND R. J. COOPER. 1990. Approach- torial behavior of the Yellow Warbler. Ph.D. dis- es to avian diet analysis. Studies in Avian Biology sertation, Utah State University, Logan, UT. 13:80±90. GRAY, L. J. 1993. Response of insectivorous birds to ROSENBERG, K. V., R. C. OHMART, AND B. W. ANDER- emerging aquatic insects in the riparian habitats of SON. 1982. Community organization of riparian a tallgrass prairie stream. American Midland Nat- breeding birds: response to an annual resource uralist 129:288±300. peak. Auk 99:260±274. GREENBERG, R. 1990. Ecological plasticity, neophobia, ROSENBERG, K. V., R. C. OHMART,W.C.HUNTER, AND and resource use in birds. Studies in Avian Biol- B. W. ANDERSON. 1991. Birds of the lower Colo- ogy 13:29±37. rado River Valley. University of Arizona Press, GREENBERG, R. 1995. Insectivorous migratory birds in Tucson, AZ. tropical ecosystems: the breeding bird currency SCHOENER, T. W. 1974. Resource partitioning in eco- hypothesis. Journal of Avian Biology 26:260± logical communities. Science 185:27±39. 263. SHANNON, J. P. 1993. Aquatic ecology of the Colorado GREENBERG, R., P. BICHIER, AND J. STERLING. 1997. River through Grand Canyon National Park, Ari- Acacia, cattle and migratory birds in southeastern zona. M.Sc. thesis, Northern Arizona University, Mexico. Biological Conservation 80:235±237. Flagstaff, AZ. GREIG-SMITH, P. W. 1978. The formation, structure and SILLETT, T. S. 1994. Foraging ecology of epiphyte- function of mixed species insectivorous bird searching insectivorous birds in Costa Rica. Con- ¯ocks in West African savanna woodlands. Ibis dor 96:863±877. 120:284±295. SOKAL, R. R., AND F. J. ROHLF. 1995. Biometry. 3rd HORN, H. S. 1966. Measurement of ``overlap'' in com- ed. W. H. Freeman and Co. Publishers, New York. parative ecological studies. American Naturalist SPSS INC. 2000. Systat version 10. SPSS Inc., Chica- 100:419±424. go. INSECTIVOROUS BIRD DIETS IN GRAND CANYON, ARIZONA 115

STEVENS, L. E. 1985. Invertebrate herbivore commu- TURNER, R. M., AND M. M. KARPISCAK. 1980. Recent nity dynamics on Tamarix chinensis Loueiro and vegetation changes along the Colorado River be- Salix exigua Nuttal in the Grand Canyon, Arizona. tween Glen Canyon Dam and Lake Mead, Ari- M.Sc. thesis, Northern Arizona University, Flag- zona. U.S. Geological Survey Professional Paper staff, AZ. 1132, U.S. Government Printing Of®ce, Washing- SZARO, R. C., AND M. JAKLE. 1985. Avian use of a ton, DC. desert riparian island and its adjacent scrub habi- WARD,J.V.,AND J. A. STANFORD. 1979. The ecology tat. Condor 87:511±519. of regulated streams. Plenum Press, New York. THIOLLAY, J. M. 1988. Comparative foraging success WIENS,J.A.,AND J. T. ROTENBERRY. 1987. Diet rela- of insectivorous birds in tropical and temperate tionships among North American grassland and forests: ecological implications. Oikos 53:17±30. shrubsteppe birds. Oecologia 42:253±292.