382 ShortCommunications and Commentaries [Auk, Vol. 110

The Auk 110(2):382-385, 1993

Factors Influencing Numbers of Syntopic House Sparrows and Eurasian Tree Sparrows on Farms

PEDROJ. CORDERO InstitutoPirenaico de Ecologœa,Consejo Superior de InvestigacionesCient•ficas, Apartado202, 50080 Zaragoza,Spain

The House Sparrow(Passer domesticus) and the Eur- was mappedand classifiedinto agriculturalland, in- asianTree Sparrow(P. montanus)show differencesin cluding: (FORAGE)orchards, cereals and associated habitat use. The former is predominantly an urban natural vegetationwhere sparrowsforaged; (FOR- and suburbanspecies, while the EurasianTree Spar- EST)woodlands, including Mediterranean oak, pine row is more rural (Summers-Smith 1963, Pinowski forestand Mediterranean scrub;and (GARDEN) gar- 1967, Lack 1971, Cody 1974, Dyer et al. 1977). How- dens, including groves of several tree speciesand ever, they often coexistalong suburban-ruralgradi- ornamentalshrubs where the sparrowsdid not forage. ents (Cody 1974), where extensivediet overlap (An- Open fields supportedagricultural activities, mostly derson1978, 1984) and nest-sitesegregation (Anderson intensivevegetable growing (89%),with the rest in 1978,Cordero and Rodriguez-Teijeiro1990) have been alfalfa, wheat, barley and maize. found. Some ecological(Pinowski 1967, Anderson To evaluate livestock I considered cattle equiva- 1978)and geographical(Summers-Smith 1963, 1988) lents of domestic animal biomass (LIVESIND; Table evidence suggeststhat habitat differences,in part, 1). To obtain this, total masswas evaluatedusing the may be modeled by competitive interactionsduring following estimatedvalues: 0.325 kg for a domestic which the House Sparrow is dominant. Assuming pigeon(Columba livia); 2.5 kg for a domesticfowl; 3.5 that EurasianTree Sparrowsare nest-sitelimited (An- kg for a rabbit;27 kg for a dog; 100kg for a pig; 450 derson 1978), I have analyzed the habitat character- kg for a cowor bull; and 850kg for a horse.Livestock isticsof House and Eurasian Tree sparrowson farms exceptpigeons, dogs, and occasionallydomestic fowl where nest boxes were introduced to increase the were in enclosedbuildings; with the presenceof live- available nest sites. stock,an abundantfood supplywas providedfor the My objectivewas to determine the effectsof dif- sparrows(both vegetablesand insects;Dyer et al. 1977, ferent habitat variableson the breeding numbers of Lowther 1983). both House and Eurasian Tree sparrows on farms I usedthe possiblepredictor variables summarized and at rural residences. I have examined the rela- in Table 1. Other possible variables, such as year of tionships between variables related to food, nest site, investigationin the plot (in which nest boxeswere and the number of individuals of each species,both introduced), elevation, distance to coast, distance to in natural holes and introduced nest boxes. nearestfarm or rural house,nest box support(tree or I studied 14 farms or houses in rural areas of Bar- wall), and climatic features were excluded because celona Province in northeastern Spain. All siteswere the range of variation was minimal, or earlier de- used by man and were isolated, surroundedby cul- scriptiveanalyses indicated that the variable was not tivatedfields or forest.Other hole-nestingspecies were important for this analysis. absent except for a few pairs of European Starlings Linear-multiple-regressionanalyses (Norusis 1985) (Sturnusvulgaris) at one of the farms. I put up 10 to were used to determine which of the factors best ex- 131nest boxes (Table 1) on buildingsor on the nearest plained the observed numbers of House Sparrows trees adjacent to them. In the breeding season I and EurasianTree Sparrowsin the plots. Before the searchedintensively for occupiednests (as in Cordero analyses,I transformedthe variableslogarithmically and Rodriguez-Teijeiro1990) and checkednest boxes (log) sothat they were normallydistributed and had at three- to four-day intervals to determine whether homoscedasticvariances (e.g. Zar 1984). The factors they containedcompleted clutches. The number of that could affect the numbers of birds (Table 1) were sparrowswas determined basedon the total number entered into the analyseswith forward stepwise in- of nest sites occupiedand for which there was evi- clusion of variables (Norusis 1985). Four regression denceof breeding.Also, cavities potentially available models were developed. In two, the dependent vari- for nesting were noted. ableswere the total numberof House•parrow nests Variablesassociated with food availability were re- (ALL-HS) and the total number of Eurasian Tree lated to vegetationcover and livestock(Dyer et al. Sparrownests (ALL-ETS). In the other two models, 1977, Lowther 1983). The vegetationalcover in the the dependent variableswere the number of House 100 m around eachplot was measured,assuming that, Sparrownests in nestboxes (HSBOX) and the number normally, foraging trips were within that distance of EurasianTree Sparrow nests in nest boxes (ETS- (Summers-Smith1963, pers. observ.).The vegetation BOX).For eachfinal equation,the regressionresiduals April 1993] ShortCommunications andCommentaries 383

T^n•,E1. Designation and descriptionfor variables included in multiple-regressionanalyses, with means, standard deviations and ranges observed at 14 sites.

Code Description œ+ SD Range FORAGE Percent cultivated land and 62.43 + 22.72 30-99 associatedvegetation cover FOREST Percent forest + scrub cover 29.28 + 24.25 0-60 GARDEN Percent grove and scrub 4.29 + 7.28 0-20 garden cover TREES Percent FOREST + GARDEN 33.57 + 22.57 0-60 LIVE$IND Livestock index a 11.86 + 26.93 0-98 CAVITY No natural cavities 22.00 + 21.72 3-70 BOXES No nest boxes 31.21 + 29.47 10-131 HSCAV No House Sparrow nests in 12.71 + 15.05 0-43 natural cavities ETSCAV No EurasianTree Sparrow nest in 1.14 + 1.23 0-4 natural cavities HSBOX No House Sparrow nestsin 3.78 + 6.09 0-22 nest boxes ETSBOX No Eurasian Tree Sparrow nests in 5.35 + 7.99 0-31 nest boxes RHSCAV. Ratio HSCAV/CAVITY 0.45 + 0.76 0.0-0.9 RETSCAV Ratio ETSCAV/CAVITY 0.13 + 0.17 0.0-0.5 RPSCAV Ratio (HSCAV + ETSCAV)/CAVITY 0.57 + 0.23 0.0-0.9 ALL-HS No House Sparrow nestsb 18.00 + 18.34 0-50 ALL-ETS No Eurasian Tree Sparrow nestsb 6.71 + 9.23 0-36 • Cattle equivalentsof domesticanimal biomasscalculated as the sum of estimatedmean massand total numbersof all domesticanimal species in plot divided by estimated mean massof cattle (450 kg). • Nests included that were in open (i.e. not in cavities or nest boxes).

met the assumptionsfor linear-multiple-regression explained 57% of total variance (Table 2). The nega- analysis(e.g. Norusis 1985). Results are given as )• + tive association with TREES is related to food avail- SD. ability. Food overlap between House and Eurasian Both House and EurasianTree sparrowsoccurred Tree sparrowsmay be considerablein the breeding in 12 (86%)of the plots (n = 14), but the HouseSpar- season (Anderson 1984). However, some differences row was more numerousthan the EurasianTree Spar- in foraging ecology (a consequenceof greater attach- row (Table 1; Wilcoxon matched-pair signed-ranks ment of House Sparrowsto humans)should exist be- test, z = -2.118, two tailed, P = 0.034). The number tween the two species.The House Sparrow could ob- of available natural cavities (CAVITY) was the most tain food both from agricultural areas (including important factor affecting the total number of breed- associatednatural vegetation)and from areasclose to ing House Sparrows(ALL-HS) explaining 54%of the human dwellings (e.g. food waste, livestock),while variation (Table 2). Variables related to food avail- the EurasianTree Sparrowmay obtain mostof its food ability (i.e. livestockindex, percent of foraging areas) from agricultural areasin open fields where natural were not determinantsfor ALL-HS or, at least,they vegetation is common. An increaseof woodland and were not so within the range of variation found in gardens with tree plantation reduces the potential the plots investigated(Table 1). LIVESIND, although foraging areas (r = -0.99, P < 0.001), and this may approaching significance,did not enter in the re- limit the carrying capacityof the habitat for the Eur- gression equation. Although the relationship be- asian Tree Sparrow. This result does not contradict tween the House Sparrow and presenceof livestock the findings of another study on habitat selectionby is well documenteddue to the additional food sup- the Eurasian Tree Sparrow, in which forests repre- plies (e.g. Summers-Smith 1963, Dyer et al. 1977, sented a secondarybreeding area for the species(Pi- Lowther 1983),agricultural land alsosupported high nowski 1967). In Pinowski's study, nest boxes were densitiesof sparrows.The sparrowsconcentrated in put up in villagesand forests,while in mine the nest isolatedbuildings amid orchardsand croplands;old boxeswere placed only near buildings. In both sit- roofs (with pottery tiles) in these areasprovided nu- uations, EurasianTree Sparrows needed to increase merous nest sites. foragingdistance from the nestswhen forestcoverage The total numberof EurasianTree Sparrows(ALL- was extensive. ETS)was positively related to the number of available Although the numbers of nest boxesoccupied by nest boxes(BOXES) and negatively to the percent of the two specieswere similar (Wilcoxon test,z = - !.27, total tree and scrub cover (TREES),which together two tailed,P = 0.20),proportionally, the EurasianTree 384 ShortCommunications and Commentaries [Auk, Vol. 110

TABLE2. Coefficients for multiple-regression equa- cavities(ETSCAV) was small (Table 1), nonbreeding tions usedto predict total numbersof occupiednests individuals also were around; they readily occupied and numbers of occupied nest boxes by the two nest boxes when they were put up. House Sparrow speciesof sparrows.For total numbersof occupied nests in natural cavities (HSCAV) also could affect the nests, independent variables entered in the anal- number of Eurasian Tree Sparrow nests in nest boxes yses were: BOXES, CAVITY, FORAGE, FOREST, GARDEN, TREES, and LIVESIND; also ALL-HS and by competitivedisplacement and differential nest-site ALL-ETSused as independent variables for analysis preference. of the other species.For numbers of occupied nest In rural areas,where the building of open nestsis boxes by the two species,independent variables uncommon(e.g. Cordero and Rodriguez-Teijeiro1990), were: BOXES, CAVITY, HSCAV, ETSCAV, numbers of breeding House and Eurasian Tree spar- RHSCAV, RETSCAV, RPSCAV, FORAGE, FOREST, rows on farms are primarily limited by the presence GARDEN, TREES, and LIVESIND; also HSBOX and of the available nest sites (e.g. Kendeigh 1934, ETSBOXused for analysesof other species.Variable MacKenzie 1946, von Haartman 1971). I did not find codes as in Table 1. a negative relationship between the numbers of the two species.Notice, however, that the aim of my study Indepen- Signif- Cumu- dent Regres- icance lative was not to investigate interspecificcompetition be- variable sion level ad- tween the speciesunder conditions of abundant nest Dependent or coeffi- (two- justed sites; such a study would require a quite different variable intercept cient (b) tailed) R2 approach. ALL-HS CAVITY 0.158 0.006 0.54 Some additional and unexplained variation in the Intercept 0.705 0.001 number of House and Eurasian Tree sparrows may be due to uncontrolled variables. Certain cultural fac- ALL-ETS BOXES 0.012 0.003 0.32 tors and/or local predation rates may influence TREES -0.012 0.029 0.57 whether sparrowsoccupy available nests sites on farms Intercept 0.704 0.002 or rural housessupplied with nest boxes. HSBOX RPSCAV 1.895 0.002 0.63 I am grateful to those who let me work on their BOXES 0.005 0.044 0.79 properties and supplied me with facilities. I thank T. Intercept 0.641 0.001 R. Anderson, R. F. Johnston, P. Lowther, J.P. Mar- ETSBOX BOXES 0.187 0.000 0.79 tinez-Rica,J. Pinowski, J. D. Rising, J. C. Senar and ETSCAV 1.825 0.003 0.85 J. D. Summers-Smithfor revising and improving ear- FOREST - 0.087 0.007 0.90 lier versionsof the manuscript.I thank P. L6pez de HSCAV 0.103 0.027 0.94 Vilar for her unconditional assistance in different as- Intercept - 1.880 0.088 pectsof the investigation.This study was carried out during a postdoctoralfellowship given by the C.S.I.C. Sparrowsused more nest boxes (75 of 94 total nest at the Instituto Pirenaicode Ecologiaat Jaca. sites;80%) than House Sparrows (53 of 252; 21%; X 2 = 101.2,P < 0.0001).The number of House Sparrow LITERATURE CITED nestsin nest boxes(HSBOX) was affectedby the ratio of natural cavities occupied by both sparrow species ANDERSON,T. R. 1978. Population studies of Euro- (RPSCAV) and by the number of available nest boxes pean sparrows in North America. Occas. Pap. (BOXES), which together explained most of the vari- Univ. Kans. Mus. Nat. Hist. 70:1-58. ation (79%;Table 2). This I interpreted to be the result ANDERSON,T. R. 1984. A quantitative analysis of of nest-site preference and the dynamics of nest-site overlap in nestling diets of village populations occupation.The House Sparrow doesnot readily oc- of sparrows(Passer spp.) in Poland. Ekol. Pol. 32: cupy nest boxes when natural cavities are available; 693-707. however, when natural cavities are scarce or the re- COPY,M.L. 1974. Competition and the structure of source is saturated, nest-box occupation increasesif bird communities. Princeton Univ. Press, Prince- nest boxes are abundant. ton, New Jersey. Occupationof nest boxes by Eurasian Tree Spar- CORDERO,P. J., AND J. D. RODRIGUEZ-TEIJEIRO.1990. rows (ETSBOX), was related to the number of avail- Spatial segregation and interaction between able nest boxes(BOXES), the number of Eurasian Tree House Sparrowsand Tree Sparrows(Passer spp.) Sparrow nestsin natural cavities(ETSCAV), the per- in relation to nest site. Ekol. Pol. 38:443-452. cent of forest cover (FOREST;negative relationship), DYER, M. l., J. PINOWSKI, AND B. PINOWSKA. 1977. and the number of House Sparrow nests in natural Population dynamics.Pages 53-105 in Granivo- cavities(HSCAV). Together, these explained 94% of rous birds in ecosystems(J. Pinowski and S.C. the variation (Table 2). Nest boxes were a limiting Kendeigh, Eds.). Cambridge Univ. Press, Cam- factor for the Eurasian Tree Sparrow in the plots col- bridge. onized mostly by House Sparrows. Although the KENDEIGH, S.C. 1934. The role of environment in number of Eurasian Tree Sparrow nests in natural the life of birds. Ecol. Monogr. 4:299-417. April 1993] ShortCommunications andCommentaries 385

L^CK,D. 1971. Ecologicalisolation in birds. Black- SUMMERS-SMITH,D. 1963. The House Sparrow. Col- well Scientific Publications, Oxford. lins. London. LOWTHER,P. 1983. Breedingbiology of House Spar- SUMMErS-SMITH,J. D. 1988. The sparrows. Poyser rows: Intercolony variation. Occas.Pap. Univ. Ltd., Calton, United Kingdom. Kans. Mus. Nat. Hist. 107:1-17. VONHAARTMAN, L. 1971. Populationdynamics. Pages MACKENZIE,J. M.D. 1946. Somefactors influencing 392-461 in Arian biology, vol. 1 (D. S. Farner and woodland birds. Q. J. For. 40:82-88. J. R. King, Eds.). Academic Press, New York. NORUSiS,M.J. 1985. SPSSXadvanced statistics guide. Z^R, J.H. 1984. Biostatisticalanalysis, 2nd ed. Pren- McGraw-Hill, New York. tice-Hall, EnglewoodCliffs, New Jersey. PtNOWSKt,J. 1967. Die Auswahl desBrutbiotops beim Feldsperling (Passerm. montanusL.). Ekol. Pol. 15: Received3 August 1992, accepted29 November1992. 1-30.

The Auk 110(2):385,1993

Song Differences Between North American and European White-winged Crossbills( Loxia leucoptera)

JOHAN ELMBERG Departmentof AnimalEcology, Swedish University of AgriculturalSciences, S-901 83, Umer•,Sweden

Accordingto my experience,the songof the White- bill have received little attention (cf. Griscom 1937). winged Crossbill (Loxialeucoptera; referred to as the Thus, better knowledge of the song (and other vo- Two-barred Crossbill in Europe) in eastern Canada is calizations)of the White-winged Crossbillis needed distinctly different from that of Fennoscandianbirds. from all parts of its range. I especially urge birders I heard about 20 singing individuals in New Bruns- and professionalzoologists to record and describe wick, Canada, in August 1991. They all had a song singing White-winged Crossbillsfrom easternRussia, similar to that generally describedin North American western North America, and Hispaniola, where the field guides.This songincludes a long seriesof loud geographicallyisolated L. I. megaplagaoccurs. trills on different pitches,each trill being monoto- I thank Jeff Groth and Krister Mild for comments nous and reminiscent, for example, of a singing on a draft of this note. I thank Claude Chappuis,Jean Greenfinch(Carduelis chloris), a calling CommonRed- C. Roch• and Richard Ranft (National Sound Archive poll (Acanthisfiammea), or a singing Arctic Warbler in London, United Kingdom) for information con- (Phylloscopusborealis). Four recordings of singing cerning recordings.Special thanks are due to Andrea White-winged Crossbillsfrom Ohio and Maine, sup- L. Priori and RobertaSwan at the Cornell Laboratory plied by the Cornell Laboratoryof Ornithology, all of Ornithology,Ithaca, New York,for providingNorth featurethe samesong type. American recordings. In contrast,the White-winged Crossbillsthat I have heard in Sweden and Finland (June-October) all have LITERATURE CITED had a very different song. It is rich and varied, con- sistingmainly of clear(metallic) and slurredwhistles, ELMBERG,J. 1991. Flight calls of Two-barred Cross- but also of chattersand wheezes. This song is fre- bill. Br. Birds 84:344-345. quently interspersedwith any of the three basicflight GRISCOM,L. 1937. A monographic study of the Red calls (Elmberg 1991). This song type, thus, is remi- Crossbill. Proc. Boston Soc. Nat. Hist. 41:77-210. niscentof the songof the Red Crossbill(L. curvirostra) GROWH,J. G. 1988. Resolutionof cryptic speciesin or the Parrot Crossbill(L. pytyopsittacus).Note that Appalachian Red Crossbills.Condor 90:745-760. the recordingof a singing White-winged Crossbillon KNOX, A. G. 1990. Identification of Crossbill and All the Bird Songsof Europe(Roch• 1990) is also a Scottish Crossbill. Br. Birds 83:89-93. Canadianbird (JeanC. Roch• pets. comm.). ROCHE,J.C. 1990. All the bird songsof Europe.Four Recentsuggestions of the occurrenceof crypticspe- compact discs. Sittelle, La Mute, France. ciesin the Red Crossbill(Groth 1988)highlights the TYRBERG,T. 1991. Crossbill(genus Loxia) evolution importanceof knowingwhether the song of theWhite- in the West Palearctic--A look at the fossil evi- winged Crossbill differs consistentlybetween the dence. Ornis Svecica 1:3-10. Palearcticand the Nearctic.In comparisonwith its closestrelatives (see Knox 1990, Tyrberg 1991), the Received27 April 1992, accepted15 June1992. taxonomyand phylogenyof the White-wingedCross-