INFLUENCE OF RESOURCE ABUNDANCE ON USE OF TREE-FALL GAPS BY BIRDS IN AN ISOLATED WOODLOT JOHN G. BLAKE• AND WILLIAM G. HOPPES2 Departmentof Ecology,Ethology, and Evolution, 606 East Healey Street, University of Illinois, Champaign,Illinois 61820 USA ABSTR^CT.--Theoccurrence of birds in forestunderstory and tree-fallgaps during spring and fall migrationperiods was determined in an isolatedwoodlot. We usedmist-net captures to testthe hypothesisthat birds are attractedto gapsbecause of higher resourcelevels. We captured1,010 birds (74 species)in spring and 458 (44 species)in fall. Total capturesand capturesper net were higher (P < 0.001) in gapsduring spring and fall. Mean number of speciesper net washigher in gaps(P < 0.001)during both seasons,but total speciesin gaps (69 spring,43 fall) was not significantlyhigher than in forestunderstory (60 spring, 28 fall). Of 44 speciesrepresented by adequatesample sizes (n > 5) in spring,9 were significantly (P < 0.05) more commonin gapsand 2 were more commonin forest understory.Nine of 17 specieswere capturedmore often (P < 0.05) in gapsduring fall. During spring, flycatchers, ground insectivores,foliage insectivores,and granivore-omnivoreswere capturedmore fre- quently(P < 0.05) in gaps.Flycatchers showed no differencein fall, but other trophicgroups, including frugivores,were capturedmore frequently (P < 0.05) in gapsthan in forestunder- story sites.Bark foragersshowed no statisticalpreference for gapsor forestunderstory in springor fall. Total speciesper net and total capturesper net correlatedpositively (P < 0.05) with densityof foliagein the lower canopyand negativelywith densityof upper canopy foliage in both spring and fall. Total speciesand capturescorrelated positively (P < 0.05) with insectabundance in spring and with fruit abundancein fall. Foliageinsectivores cor- related positively with low canopy foliage and insectabundance in both spring and fall. Capturesof frugivorescorrelated with fruit abundancein fall. Thesedata supportthe hy- pothesisthat birds are attractedto tree-fall gapsbecause of higher resourceabundance and provide further evidence of the importance of habitat heterogeneity to the structure and compositionof bird communities.Received 7 November1984, accepted 28 October1985. GAPScaused by tree fails contribute to the 1986) have demonstrated differences in assem- creation of a habitat mosaic in many forests blages of birds captured in forest gaps and (Williamson 1975; Hartshorn 1978; Whitmore understory.Results of thesestudies suggest that 1978; Runkle 1981, 1982; Brokaw 1985). Tree- birds might be attracted to gaps becauseof fall gaps ["a vertical 'hole' in the forest extend- higher resourcelevels. Here, we test the pre- ing to within 2 m of the forest floor" (Brokaw diction that abundanceof individuals in gaps 1982)] influence abundance and distribution of and forest understory correlates with abun- bird speciesby maintaining habitat heteroge- dance of resources in these locations. We used neity and by affecting abundance and distri- mist nets to obtain concurrentsamples of birds bution patterns of food resources(e.g. fruit and in gaps and forest understory sites and com- insects).Resource levels may be higher in gaps pared number and speciescomposition of cap- becauseof greater primary productivity asso- tures with estimates of insect and fruit abun- ciatedwith increasedlight levels (Fogden 1972, danceand with measuresof vegetationstructure Halle et al. 1978). Previous studies in east-cen- for the same sites. tral Illinois (Willson et al. 1982, Martin and Karr Migration is energetically expensive, and many birds must replenish fat reservesperiod- ically (Berthold 1975, Graber and Graber 1983, • Present address: Department of Zoology, Birge Hall, University of Wisconsin, Madison, Wisconsin Walsberg1983). Food suppliesoften are low or 53706 USA. unpredictable during migration (Walsberg 2 Present address: Center for Earth and Environ- 1983),particularly in springwhen migrantsare mental Science,Hudson Hall 102, StateUniversity of moving toward areaswhere local weather con- New York, Plattsburgh,New York 12901 USA. ditions may be severeand unpredictable. Find- 328 The Auk 103: 328-340. April 1986 April 1986] Useof Tree-fallGaps by Birds 329 ing adequatefood after a flight may be difficult known, all were well establishedand at least 3-5 yr becausethe landing area often is unknown to old. migrants. Consequently,it would be advanta- We nettedbirds 3 days/week,except when rain or geousfor birds to recognizeand quickly select strong winds precludeduse of nets, from about 15 min before sunrise until about noon. We standard- foraging sites that are profitable (Martin and ized netting effort among daysto avoid problemsas- Karr 1986). This may be particularly true for sociatedwith diurnal variability in capturerates (Karr migrants in east-centralIllinois, where natural 1981).We nettedbirds on 18 daysin spring(19 April habitat is limited and existsas isolatedpatches to 5 June) and on 24 days in fall (23 August to 31 in an agriculturalsetting (Blake 1986).Recent October). Nets were checked every hour, and cap- work by Graberand Graber(1983) demonstrat- tured birds were identified, sexed (when possible), ed that migrant warblersin spring experience weighed,banded with U.S. Fish and Wildlife Service a net lossof energywhile foragingin woodlots aluminum leg bands, and releasedat point of cap- in east-centralIllinois. By demonstratingdif- ture. ferencesin resourcelevels between gapsand We comparedthe number of capturesin gap and forest understory sites, and by demonstrating forest understory nets using Chi-square tests (for samplesizes >_30) or FishersExact Test (for n < 30; correlationsbetween capturefrequency and re- Sokal and Rohlf 1981).We correlatednumber of cap- source level, we support the prediction that tures per net with measuresof habitat structure and during migration birds selectas foraging sites resource abundance (see below) using Spearman's microhabitats with abundant food resources. rank correlation (Sokal and Rohlf 1981). We divided speciesinto six major trophic groups STUDY AREA AND METHODS in spring (flycatchers,ground insectivores,bark for- agers,foliage insectivores, omnivore-granivores, and Study site.--We mist-netted birds in William Tre- nectarivores)and seven in fall (above six plus fru- lease Woods, an upland forest tract of about 24 ha givores)based on literature (Martin et al. 1951, Will- located 6.5 km northeast of Urbana (Champaign son 1974, Willson et al. 1982) and personal observa- County), Illinois. Principal trees in the woods are tions (seeAppendix). Nectarivoreswere represented sugarmaple (Acersaccharum), hackberry (Celtis occi- by one species(Ruby-throated Hummingbird, Ar- dentalis),white ash (Fraxinusamericana), slippery elm chilochuscolubris), and we did not include nectari- (Ulmus rubra), basswood (Tilia americana), red oak vores in the analysesof trophic groups.We catego- (Quercusrubra), and buckeye (Aesculusglabra). The rized frugivoresas primary [those that dependheavily understorysupports a variety of species,including on fruit, i.e. more than 75%of foragingobservations young individuals of the above species, pawpaw in fall (Hoppespers. obs.)] and secondaryfrugivores (Asiminatriloba), and spicebush(Lindera benzoin). [thosethat consumefruit regularlybut lessoften than Many speciespresent in Trelease Woods produce primary frugivores,i.e. 25-50% of foraging observa- fruits that are eaten by birds in late summerand fall. tionsin fall (Hoppespers. obs.)] and comparednum- Theseinclude spicebush, hackberry, Virginia creeper bers of capturesof all frugivoresand number of cap- (Parthenocissusquinquefolia), wild grape(Vitus vulpina), turesof primaryand secondaryfrugivores separately. moonseed(Menispermum canadensis), catclaw (Smilax Becauseconsiderable attention has been placed on hispida),poison ivy (Rhusradicans), and wahoo (Eu- the useof gapsby frugivoresin fall (e.g. Thompson onymusatropurpureus). and Willson 1978, 1979), we also examined use of Birds.--We used mist nets (12 m length, 2.6 rn gapsand forestunderstory during springby species height, 36 mm mesh)to samplebirds using the lower that are frugivorous in fall. levels(bottom 2-3 m) of gapsand forestunderstory Vegetation.--We recorded vertical distribution of during spring and fall 1983. We believe that the ad- vegetationat eachnet site following the methodsof vantagesgained by simultaneoussampling of all sites Karr (1971).Presence or absenceof vegetationin each outweigh any biases associatedwith mist-net sam- of 14 height intervalswas noted at 40 points per net pies (Karr 1979, Martin and Karr 1986; but see Rem- (Fig. 1). Height intervals (meters)were 0-0.3, 0.3-0.9, sen and Parker 1983). Also, comparisonswith pre- 0.9-1.5, 1.5-2.1, 2.1-2.7, 2.7-3.4, 3.4-4.0, 4.0-4.6, 4.6- vious tree-fall gap studies are facilitated because 6.1, 6.1-9.1, 9.1-12.2, 12.2-15.2, 15.2-18.3, and >18.3. similar techniqueswere used (Schemskeand Brokaw Percentagecover for a given height interval was cal- 1981, Willson et al. 1982, Martin and Karr 1986). Six culatedas the percentageof all points with vegeta- nets were placed in gaps (1/gap) and 6 in forest tion presentat that height. We sampledvegetation understory. Gap nets were positioned to minimize profileseach week in spring and once every three visibility while still remainingwithin the gaps.We weeksin fall, when foliagedistributions change less did not placenets in forestunderstory between gaps rapidly. that were less than 40 rn apart or within