Owls on Islands and Mainland Sites
Total Page:16
File Type:pdf, Size:1020Kb
J. Raptor Res. 21(1):3-7 ¸ 1987 The Raptor Research Foundation, Inc. PREY SELECTION OF COMMON BARN-OWLS ON ISLANDS AND MAINLAND SITES DAVID W. JOHNSTONAND JAMESM. HILL ABSTRACT.--Datafrom the literature and a recentcollection of CommonBarn-Owl (Tyto alba) pellets from Block Island, Rhode Island, were used to assessthe relative numbers of birds and non-avian vertebratestaken by this owl on islandsand mainland sites.Our analysissupports the hypothesisthat barn-owl diets include proportionatelymore birds (both speciesand individuals) on islands than at mainland sites.The percentof bird speciesand individual birds in the diet decreasesfrom the equator to 54øN. Possible causes for island vs. mainland diets and latitudinal trends are discussed. The interaction between the Common Barn-Owl the percentof bird speciesamong all the vertebratespecies (Tyro alba) and its prey is well documented.Pub- captured, and 2) by consideringthe percent of all indi- vidualbirds vs. individuals of all non-avianvertebrate prey. lishedstudies deal with economicaspects (Bendire 1895; Errington 1932; Wallace 1950), population RESULTS AND DISCUSSION ecology(Davis 1959; Otteni et al. 1972; Herrera Prey Selectionon Islands.Most previousdietary and Jaksi• 1980), and range extensionsof mam- studies of barn-owls from mainland sites have shown malian prey (Kirkpatrick and Conway 1947; Stickel and Stickel 1948; Baker 1953; Parmalee 1954). This a preponderanceof mammalianprey. Mammal prey speciesfrom the 50 mainland sites examined here owl is believedto specializeon mammalian prey, constituteda mean of 92.4% (SD = 8.29) of the total butJohnston (1974: 172) reporteda high percentage vertebratediet. Despite this preponderanceof small of bird speciesin barn-owl pelletsfrom Grand Cay- mammal prey species,the mainland barn-owls took man Island, BWI. From that study and two other somesmall birds and, evenless frequently, reptiles, island reportsavailable then, he proposedthat "on amphibians,and insects.By contrast,on 23 island some islands ... where small mammalian prey is sites mammal speciesconstituted a mean of only reducedin diversity and total numbers, the barn- 60.5% (SD = 25.47). owl becomesalternatively a significantpredator of birds and other non-mammalian vertebrates." We The number of bird speciesas a percent of the totalvertebrate prey species from islands (.• = 38.6) now test that hypothesisby using data from addi- is significantlygreater than values from mainland tional publishedaccounts of barn-owl diets on is- landsand mainland sites.While examiningthe data sites(-• = 19.9)(Table 1). Althoughthe number of from those locations,we developedan additional bird speciesper se doesnot differ significantlybe- tween islands and the mainland (Table 1), barn- hypothesisthat barn-owl prey on more northerly islands includes fewer birds than on islands closer owls took fewer mammal specieson islandsthan on the mainland, thus making the proportion of bird to the equator. speciestaken on islandshigher. We also examined MATERIALS AND METHODS published data on the total number of individual A searchof the literature from the northernhemisphere birds and non-avianvertebrates extracted from pel- yieldedquantitative pellet analysesfrom 23 island sites lets. The number of individual birds identified as a occurringfrom the GalapagosIslands (0 ø) to SheppeyIsle (British Isles, 54øN) and from 50 mainland sites,mostly percentof all vertebratesis greateron islands(,• = in the United States,but alsofrom somelocalities in Spain, 10.5)than on the mainland(-• = 4.0). Thus,barn- Poland,and Italy. Unpublisheddata from severalislands owls on islandsprey proportionatelymore on birds in the British Isles were obtained from David E. Glue, as than other vertebrates(mainly mammals), than at were unpublisheddata from Martha's Vineyard, Mas- mainland sites. sachusetts(G. Jonesand K. Driscoll). In Appendix I are unpublisheddata from BlockIsland, RhodeIsland. Data Some publishedaccounts are of interest because extracted from these accounts are used in our statistical of the extremes(0-100%) of avian prey taken by analyses(Mann-Whitney U-Test) and in the regression barn-owls.For example,mainland areasfrom which analyses. no birds were reported include California (Foster To test the hypothesissuggested by Johnston(1974) that barn-owlstake proportionatelymore birds on islands 1927:6 of 11 sites;Hawbecker 1945; Fitch 1947), than on the mainland, pellet data from the literature were Massachusetts(Boyd and Shriner 1954), SouthCar- examinedand comparedin two ways: 1) by considering olina (Tedards1963: oneof four seasonalsamples), 4 ,JOHNSTONAND HILL VOL. 21, No. 1 Table 1. A comparisonof vertebrateprey of GommonBarn-Owls betweenmainland and island sites. PREY MAINLAND a ISLANDSb All vertebratespecies 10.7; SD = 4.84 ,,• = 8.6; SD = 5.40 (U=743; P=0.30) All bird species 2.6; SD =2.57 •=3.9; SD =4.39 (U=482; P=0.318) All mammal species 8.0; SD = 3.26 •'=4.4; SD = 1.82 (U = 941; P < 0.0005) Number of bird speciesas percentof all 19.9;SD = 16.52 •' = 38.6; SD = 24.63 vertebrate species (U=312; P< 0.01) Number of individual birds as percentof 4.0; SD = 6.51 ,,• = 10.5;SD = 14.34 all vertebrate individuals (U= 284; P=0.008) Data from 50 sites: references1, 2, 7, 10, 14, 16, 17, 18, 20, 22, 23, 25, 26, 27, 28, 29, 30, 33, 36, 37, 38, 39, 41, 42, 43, 44, 46, 47, 49, 50, 51, 52, 53, 54, 55; D. Glue (unpubl. data). Data from 23 sites:references 4, 5, 8, 11, 12, 13, 15, 21, 34, 56; D. Glue (unpubl.data); G. Jonesand K. Driscoll(unpubl. data), Appendix I. and Ireland (Fairley 1966: two of 10 seasonal By examiningonly mainlanddata for barn-owls samples). Island studiesreporting no birds were in Europe, Herrera (1974) proposeda latitudinal thosefrom Skomer(Brown and Twigg 1971), Bute effectand useda modificationof the now-question- (D. E. Glue, pers.comm.), and Martha's Vineyard able (Pielou 1977) Shannon-Wienerdiversity index, (Choate 1972). At the other extreme, on someislands namely"trophic diversity in relationto biomass"of where extensivecolonies of seabirdsoccur, barn-owls prey captured.Herrera's report, althoughnot strict- fed exclusivelyon birds (Bonnot 1928). The "out- ly comparableto the presentstudy which focuseson lier" or anomalouspoints in Figure I (60 units at avianprey selection,noted a significantnegative cor- 25øN) and Figure 2 (51 units at 25øN) came from relation betweentrophic diversityin relation to bio- the small and perhapsinadequate sample of Banks mass and north latitude. (1963) wherein the "remains of at least six Craveri Although the effectsof latitude were the samefor Murrelets [Brachyramphuscraveri] and at leastfour islands and mainland sites, the greater percentages wood rats ..., were identified." of bird speciestaken on islandscompared with main- Latitudinal Variation. We searched for relation- land sites at the same latitude merit comment. We shipsbetween latitude and bird prey in barn-owl believethat this differenceis due, at least in part, to diets using covarianceanalysis of the transformed the different relative numbers of available prey percentagesof vertebrateprey speciesthat were birds species,especially birds vs. small mammals.Such (Fig. 1). ANCOVA indicatedthat a simplelinear comparisonsare often impossibleto documentand modelis an adequatedescription for both the island quantifybecause of thelack of publishedinformation and mainland data (P > 0.05). The slopesof the on numbersof availablespecies. On Grand Cayman regressionlines were significantlydifferent from zero at 19øN where barn-owl diets included about 60% (P = 0.05), but there was no evidencethat the two bird species,only 5 smallmammal speciesincluding slopeswere different from each other even though 3 bats occur,whereas about 70 passerinebird species the y-axisintercepts were significantlydifferent (P < havebeen identified (Johnston 1974 and pers.obs.) 0.05). Consideringnumbers of speciesfound in pel- Regressionlines for percentagesof individualbirds lets, effects of latitude were, therefore, the same on in the prey items also show a negativecorrelation islandsand the mainland.Toward the equatorbird with latitude but the y-intercepts are not signifi- speciescomprised a significantlygreater percentage cantlydifferent from eachother (Fig. 2). The best of vertebrate prey speciesthan at higher northern estimate for a commonslope is a decreasein indi- latitudes,although no comparabledata were avail- vidual birds of approximately 7.3% for each 10ø able for mainland sites from 0ø-25øN. The best es- latitude northward. timate of a commonslope (islands and mainland) Our analysisdemonstrates that barn-owlsprey was that bird speciesin barn-owl diets decreaseby 1) on proportionatelymore avian speciesand in- approximately6.2% for each10 ø latitude northward. dividual birds on islands than on the mainland and SPRING 1987 COMMON BARN-OWL PREY SELECTION 5 60' IS•I.. ANDS 60 o o •.• ,•o_ o ß Z 4o- o o•••ß ß • 30 • ß o• ß • ß 20. -'?":):' ' ._,,i o ..,o• øo 2o ß o • ß •'- 10• o • ß ß •- o • o O o• ß • ß lo I ß ;-. o ,;' 2;' 3;' ,o' •o' NORTH LATITUDE ,'o' io' 3;' , o' 5o' 6'o' Figure 2. The relationshipsbetween degreesof north NORTH LATITUDE latitude and the percentof total birds among Figure 1. The relationshipbetween degrees of north lat- all vertebrateprey found in Common Barn- itude and the percentof bird speciesamong Owl pellets from islandsand mainland sites. all vertebratespecies of preyfound in Common Percentageshave bccn transformed(arc-sine Barn-Owl pelletsfrom islandsand mainland of •). Individualdata pointsarc from