Finch-1990-Con-House Wrens and Tree Swallows-1.Pdf

The Condor92674487 Q The CooperOrnithological Society 1990

EFFECTS OF PREDATION AND COMPETITOR INTERFERENCE ON NESTING SUCCESS OF HOUSE WRENS AND TREE SWALLOWS

DEBORAH M. FINCH ’ Rocky Mountain Forest and Range Experiment Station, 222 South 22nd Street, Laramie, WY 82070

Abstract. I examined the relationships among brood survival in House Wrens (Trog- lodytesaedon) and Tree Swallows (Tachycinetabicolor) and rates of nest-box use, species interference, and nest predation. Tree Swallows nested in boxes in one of three woodlands occupied by House Wrens. Over a 4-year period, clutch mortality rates in swallows were significantlyhigher than those in wrens, but wrens on swallow-free plots had lower failure rates than wrens that coexisted with swallows. Though conspecificsinterfered at 9 of 99 (9%) wren nests, predation was the major causeof wren nest failure, accountingfor 70% of 27 unsuccessfulattempts. Increasednest failure in wrens was associatedwith increasedrates of box use. In contrast, clutch mortality in Tree Swallows was related to nest interference by wrens. Of 29 swallow nests, 13 (45%) showed signs of interference by wrens, and eight of these 13 (62%) were initiated in boxes containing empty “dummy” nests built earlier by wrens. HouseWrens are interferencecompetitors because they exclude swallows from boxes by destroyingswallow nests. Destroying nests of otherspecies has advantages for wrensif high concentrationsof empty nests,including dummy nests,inhibits searchefficiency of predators.An experimentalapproach is recommendedfor testingthe hypothesisthat House Wrens build dummy nestsand destroyheterospecific nests because empty nestsdeceive predators. Key words: Interferencecompetition; clutch mortality; nestpredation; box-use rate; dummy nest;riparian woodlands;House Wren; Troglodytesaedon; Tree Swallow; Tachycineta bicolor.

INTRODUCTION populations (Stutchbury and Robertson 1987a), The availability of nest holes may limit abun- brood parasitism and egg dumping (Lombard0 dances of secondary cavity-nesting birds (Von 1988, Picman and Belles-Isles 1988, Price et al. Haartman 1957; Holroyd 1975; Gustafsson 1988; 1989) bigamy and extra-pair copulations (Quin- Brawn and Balda 1987,1988). House Wrens and ney 1983, Johnson and Kermott 1989) infan- Tree Swallowsare secondarycavity-nesters whose ticide (Belles-Isles and Picman 1986b, Robert- abundances often increase after introduction of son and Stutchbury 1988) and killing of adult nest boxes (Low 1933, Drilling and Thompson conspecifics (Lombard0 1986, Belles-Isles and 1984, Yahner 1983/1984). Male House Wrens Picman 1987). build dummy nests and defend multiple cavities Under conditions of nest-site limitation, the within their territory boundaries (Kendeigh 194 1, consequencesof interspecificcompetition for nest Belles-Isles and Picman 1986a), thereby limiting boxes and natural tree holes may overshadow breeding opportunities for other nesting pairs. In the effectsof intraspecific competition (Van Bal- Tree Swallows, territorial defense in the form of en et al. 1982, Gustafsson 1988, Ingold 1989) aggressiveattacks can prevent conspecificsfrom resulting in reduced reproductive successof the breeding even when unoccupied boxes are pres- subordinate competitor (Gustafsson 1987, In- ent (Harris 1979, Robertson and Gibbs 1982). gold 1989). In House Wrens, both sexespuncture Intraspecific competition for nest sitesand mates eggs and destroy nests of open-nesting birds in these speciescan apparently lead to nest usur- (Belles-Isles and Picman 1986b) and other cav- pation (Leffelaar and Robertson 1985) floating ity-nesting species(Gardner 1925, Sherman 1925, Kendeigh 1941). Belles-Islesand Picman (1986b) speculatedthat interspecificcompetition for food I Received 27 November 1989. Final acceptance12 or nest sites may explain nest-destroying behav- April 1990. ior by House Wrens. If House Wrens dominate NESTING SUCCESS OF HOUSE WRENS AND TREE SWALLOWS 675 interactions, then interference competition by ability in nesting productivity and focusing on wrens may result in lowered reproductive success correlative relationships among the residuals of in other species.Yet, no studies have compared rates of box use, interference, and nesting failure. reproductive successin House Wrens to success To determine if availability and frequency of use in hole-nesting speciesexposed to wren interfer- of nest boxes influenced rates of nest destruction ence. Unfortunately, effectsof predation on nest and competition, I used the ratio of used to total distribution and brood survival can mask or be nest boxesas an index. If the number of potential mistaken for results of competition (Nillson sites that remain unused or empty is as infor- 1984). To determine if interspecific competition mative to search-strategypredators or compet- is assymetrical, interference effectsmust first be itors as the number of used sites(e.g., Watts 1987, isolated from nesting lossescaused by factorslike Martin and Roper 1988), then this ratio may be predation. a more relevant predictor of “density-depen- In the central Rocky Mountains, House Wrens dent” interactions than abundances of birds or and Tree Swallowsfrequently co-occur in foothill nests. riparian woodlands (Finch 1989a); House Wrens numerically dominate cavity-nesting avifaunas STUDY AREA in these habitats, with population levels that are Three study plots were established in May 1982 two to three times higher than Tree Swallows in streamsidehabitats in Carbon County, south- (Finch 1987, Finch 1989b). During a study of eastern Wyoming, at elevations ranging between nest-siteselection by House Wrens (Finch 1989~) 2,050 and 2,250 m. One plot was established at I observed that swallows and wrens were syn- Rock Creek 5 km northeast of Arlington, and topic in one of three woodlands containing nest two plots, named Foote Camp and Treasure Is- boxes. Here, I postulate that temporal, spatial, land, were spaced 34 km apart along the North and interspecificdifferences in frequenciesof box Platte River near Saratoga.House Wrens nested use influence reproductive successof wrens and in boxes at all three plots, but Tree Swallows swallows through effectson probabilities of nest usedboxes at Foote Camp only. Woodlands were destruction and competitive interference. For ex- dominated by narrowleaf cottonwood (Populus ample, rates of nest predation were positively angustijblia), a variety of shrub species, espe- associated with densities of box-nesting Great cially willows (Salix spp.),and herbaceousspecies. Tits (Purus major) (Krebs 197 1, Dunn 1977) but Shortgrass prairie interspersed with sagebrush were unrelated to densities of other cavity-nest- (Artemisia tridentata) bordered riparian wood- ing species(Tompa 1967, Brawn 1987). Varia- lands. Vegetation composition of the study areas tion in the densities or spacing of box-nesters is described in greater detail in Finch (1987, may also influence competition for nest sites, 1989~). dictating rates of box intrusion (Kendeigh 194 1, Muldal et al. 1985) and subsequent nesting suc- METHODS cess. On each of the three plots, 21 to 22 nest boxes In this study, I predicted the following: (1) rates (n = 65 boxes) were mounted at heights of 2 m of nesting failure in wrens and swallows vary in on live deciduous trees > 10 cm dbh. The num- relation to probabilities of nest predation or in- ber of boxes on each plot was limited by habitat terspecificinterference, (2) spatial and temporal patch size. Boxes were spacedat intervals of 30- changes in frequencies of nest destruction and 35 m in grids conforming to the shape and size interspecific nest intrusion are associated with of the riparian corridor. Nest boxes were built differences in box availability and rates of box ofcedar 1.7 cm thick, 14 x 14 x 28 cm in outside use, (3) in areas of syntopy, House Wrens out- dimension, with latchable top doors and en- compete Tree Swallows for boxes as reflected by trances3.8 cm diameter. Boxeswere labeled with interspecific differences in rates of interference grid coordinates. Although natural cavities were and nesting success,and (4) wrens benefit by the available, spot-map checks indicated that most absence or exclusion of nesting swallows such cavity-nesters shifted to boxes after they were that wren productivity is higher in areas without erected. swallowsthan in areaswith swallows. These pre- I determined status(empty or occupied) of box dictions were addressed by first assessingand and progressof nesting attempts by checking all then controlling for spatial and temporal vari- nest boxes early in the afternoon every 24 days 616 DEBORAH M. FINCH from mid-May to early August of 1983 through Tree Swallows and House Wrens exclusively. In- 1986. Becauseof the short period of temperate terferenceby House Wrens at Tree Swallow nests weather, only a few boxes were nested in more was determined iftwigs were found in boxesafter than once in a single breeding season,either by swallowsconstructed grass nests, or if wrens were the same nesting pair, by different pairs, or by observed at swallow nests. Pierced or cracked different species.Based on observations of entry eggs, partial clutch losses, or emptied swallow by other speciesor additional conspecificsinto nests were sometimes discovered at, or follow- occupied boxes, and records of egg destruction, ing, the time of box intrusion by wrens. Inter- followed by subsequent egg laying, it appeared ference by swallows at wren nests was recorded that some “second” clutches were products of if twig nests lined with feathers were built first, usurping birds. Therefore, to indicate competi- followed by construction of grassnests. Interfer- tive pressure for boxes, estimates of box-occu- ence by conspecificswas assumed when three or pancy rates were based on all nests. Use rates of more birds were observed simultaneously using specieswere pooled because concurrent use of an occupied box, when a singing male chased multiple boxes by both species(1) reducesavail- another singer from the box, or when, on the ability of nest sites,fostering conditions that may swallow-free plots, wren eggswere punctured or favor interspecific and intraspecific competition, lost between successivenest checks from nests and (2) increasesabundance of active box nests, that otherwise remained unaltered and active. possibly encouraging search-strategypredation. Some casesof destructionof entire clutches(Freed Occupancy rate, defined as the ratio of the num- 1986a) were probably overlooked because I ber of nests to the total number of boxes avail- scoredthem as predation by mammals and snakes able, was computed by plot and year for com- (see Finch 1989~); thus, my estimates of nest parisons with rates of interspecific nest interference are conservative. If partial clutch interference and nesting failure. Occupancy rate losseswere recorded prior to total loss, egg de- was also calculated using wren nests only. Boxes struction by wrens or swallows was designated. were cleaned out each September so that new Interference rate was defined as the number of twigs and nestswere not confused with box con- nests showing signs of interference divided by tents from the previous year. the total number of nests. The likelihood of in- Because individual male House Wrens fre- terference at nests of wrens, swallows, or both quently stuff multiple cavities with twigs (Ken- speciesmay be related to availability of nest sites, deigh 194 l), the appearance of the first egg was herein indicated by box-occupancy rate. I first used as an index to nesting. Nests were consid- assessedthe influence of the factors PLOT (l-3) ered abandoned if adults no longer attended the and YEAR (1983-l 986) on occupancy rate and nest, and nest contents failed to hatch or fledge. interference rate at nestsof House Wrens or both Predation was assumed if nest material was dis- speciesby applying two-way analysis of variance turbed, whole clutchesdisappeared, eggs or nest- (procedure MANOVA, SPSS/PC+). Differences lings were partially eaten, predator feces were in occupancy rates between plots were detected found in the nest, boxes were damaged or un- using a Student-Newman-Keuls (SNK) test of latched, or nestswere empty before nestlingswere multiple comparisons. To determine if interfer- due to fledge. Partial losses resulted from egg ence rate was associatedwith occupancy rate, I puncturing by intruders, hatching failure, and controlled for the effectsof the extraneous factors nestling starvation. Reproductive outcome was PLOT and YEAR by entering the casewise re- classified as successful,even in reduced broods, siduals listed by MANOVA in subsequenttests. if at least one offspring fledged from the nest. Pearson’s product-moment correlations were House Wrens, which are known to destroynests used to compare the residuals of occupancy rate of their own and other species,may interfere at with the residualsof either of the dependent vari- nests of nearest neighbors. Observations of be- ables, interference rate at wren nests or interfer- havioral interactions between species were re- ence rate at all nests. Number of plots occupied corded during nest checks.Because the materials by Tree Swallows was insufficient for statistical used to construct nests differ between Tree Swal- analyses of interference rates at swallow nests. lows and House Wrens, I inferred entry into an The probability of clutch mortality was cal- occupied box by another species when foreign culated using Mayfield’s (1961, 1975) method nesting material appeared. Boxes were used by based on the number of days of exposure (EXP) NESTING SUCCESS OF HOUSE WRENS AND TREE SWALLOWS 617

that a known nest is at risk. Estimates of daily mortality rate ofwhole clutches(clutch m), probability of survival for one day (clutch S), and probability of survival through the incubation and nestling phases of length t days (clutch at) were calculated for each speciesby plot and year.

I used Mayfield’s formula based on egg or nest- .I ’ ,:^.._, . ’ ling days of exposure to compute the daily rate ,,.’ ._ ,: ’ ,: ’ .,“/ ’ ofpartial losses(partial m) and the survival probability ofreduced clutchesthrough the entire nest

period (partial St).Total probability of mortality - , * J for the entire nest period (total mt) is equal to 1 - (clutch rht x partial mt x hatching rate). Nests -cmp T-bhd Rock cti were visited frequently so I used Mayfield’s mid- FIGURE 1. Ratesof boxuse and ratesof interference point assumption (Johnson 1979). at uoolednests of HouseWrens and Tree Swallowsin threestudy areas from 1983 through1986. To obtain a large-sample estimate of the variance (v) of m, I used Johnson’s (1979) estimator EXP3/((EXP - losses)x losses).To test for vari- per plot were insufficient to analyze effects of ation in daily mortality ratesbetween speciesand occupancy rate and interference on swallow off- study areas, I used the statistic (a, - s,)/( V(5,) + spring survival. But if probabilities of interaction V(S,))l’Z (Johnson 1979), where the subscript with predators or competitors are dependent on specifieseach treatment group (note that m, - the total number of boxes occupied or empty, n-l2 = 9, - 8,). If the resulting value is greater then occupancy patterns of swallows may not than the statistic, z,,~, then the null hypothesis explain variation in swallow productivity. I that m, = m, is rejected. I pooled wren nests pooled wren and swallow nests to determine if from the two box grids that had no swallows overall rates ofbox use or interference influenced becausewren fledging successby year was similar probabilities of predation, competition, or over- between these grids (Finch 1989~). Using z-sta- all nest failure. tistics, nesting mortality in this pool ofwren nests RESULTS was then compared by year to mortality of wren or swallow nestsfound on the grid sharedby both BOX USE AND NEST INTERFERENCE species.Probability values obtained from inde- Of 128 nestsfound in 65 boxesfrom 1983 through pendent testsby year were combined using Fish- 1986, 77.3% were initiated by House Wrens and er’s procedure (Sokal and Rohlf 198 1) for deriv- 22.7% by Tree Swallows. Most nests (44%) were ing an overall significancetest based on multiple at the Foote Camp plot, while 35% were at Rock outcomes. The first study year was regarded as Creek, and only 21% at Treasure Island. Al- a transitional period during which returning mi- though Tree Swallows nested primarily at Foote grants learned the locations of nest boxes, and Camp in all study years(86% of all swallow nests), therefore, 1983 results were excluded from com- the speciesbuilt four nestsin boxesat Rock Creek bined-year tests to reduce bias. in 1985. Effectsof rates of box use and nest interference Box occupancy rates varied somewhat by on nesting successof wrens or both specieswere PLOT (F2,6= 4.52, P = 0.064) but not by YEAR assessedusing partial correlational analyses. I (F,,, = 1.99, P = 0.272) (Fig. 1); over the 4-year first used two-way ANOVAs to remove the ef- period, 64% of all boxes at Foote Camp were fects of PLOT and YEAR from each continuous used, whereas 32% were occupied at Treasure variable and then correlated resultant residuals Island, and 49% at Rock Creek. Total occupancy of clutch m and partial m to residuals of inter- rate differed between Foote Camp and Treasure ference rate and occupancy rate entered concur- Island (SNK tests, P < 0.05) but not between rently. To evaluate the independent influences other pairs of plots. Wren occupancyrate did not of interference rate and occupancy rate on nest- vary among plots or years (P > 0.1). ing success,I separated out the effect of either Rates of interference by House Wrens (esti- variable and reported the resulting partial cor- mated by direct observation, and from occur- relation coefficient(v,). Numbers of swallow nests rencesof punctured eggsand displaced or added 678 DEBORAH M. FINCH nest material) at active nests of conspecificswas After effects of PLOT and YEAR were re- similar among plots and years (P > 0.10 for all moved, correlational analysis showed a signifi- comparisons), but wren interference at nests of cant positive relationship between residuals of both speciesvaried by PLOT (F2,6= 5.68, P = interference rate (both species) and total occu- 0.041) but not by YEAR (F3,6= 0.83, P = 0.522) pancy rate (Y = 0.66, P = 0.020). That is, inter- (Fig. 1). Over the 4-year period, interference was ference rates increased at nests of both species detected at 9 of 99 (9%) wren nests and 13 of 29 as more boxes were occupied. At House Wren (45%) swallow nests. At Foote Camp, the plot nests only, interference rate was not related to shared by both speciesin all years, wren inter- total occupancy rate (r = -0.0 13, P = 0.969) or ference was evident at only 3 of 3 1 (10%) wren wren occupancy rate (r = -0.34, P = 0.273), nor nests, whereas 12 of 25 (48%) nest boxes used was interference rate at nests of both speciesre- by swallows were invaded, primarily by House lated to wren occupancy rate (r = 0.014, P = Wrens depositing twigs. Although 5 of 4 1 (12%) 0.964). of all House Wren nestsat Rock Creek displayed signs of intrusion by conspecifics,only 1 of 27 NESTING FAILURE IN WRENS AND (4%) of all wren nests at Treasure Island showed SWALLOWS signs of interference. Of 99 nesting attempts by House Wrens, 70 (7 1%) At least 20 of 22 (90%) records of intrusion at fledgedone or more offspring, two had unknown nestsof either speciesinvolved House Wrens. Of outcomes,and 27 failed. Predation accounted for the nine wren nests that suffered conspecific in- 19 (70%) of the 27 failures, while five (19%) were trusions, five were totally destroyed,probably by destroyed by conspecifics,and three (11%) were conspecifics.Circumstantial evidence (chasingof desertedduring the egg-layingor incubation stages conspecifics;new twigs deposited onto feather- for reasonsunknown. Fifty successfulnests (7 1% lined nest; and egg destruction, followed by the of 70) lost one to four eggsor nestlings before appearance of new eggs) suggestedthat wrens fledging. The mean (&SE) number of lossesfrom usurped at least one conspecific box at Rock successfulnests was 1.2 * 0.1 young/nest. Tree Creek. In addition, an adult wren was found dead, Swallows lost 20 of 29 nests (71%), IO-12 (50- but uneaten, in a failed Rock Creek nest with 60%) due to wren destruction (two nests were intact eggs, suggestingeither conspecific killing abandoned after interference), and eight (40%) (Freed 1986a, Belles-Isles and Picman 1987) or by predation. The outcome of one swallow nest mortality for reasons other than predation. was unknown. Of eight successfulswallow nests, Clear signs of interference by Tree Swallows four lost one to four eggs(X t- SE = 1.00 + 0.35 at nests of conspecifics or wrens were not ob- losses/nest). served, possibly because they were obscured by In House Wrens at the shared plot, the prob- effectsofwren intrusions. Of the 13 swallow nests ability of clutch mortality over the entire nesting having signsof wren interference, 12 failed, either period (i.e., partial m’ x clutch Ifit) ranged from at the time of interference or later. At Foote a low of 50.3% in 1983 to a high of 89.1% in Camp, wrens apparently usurped two boxes set- 1984 (Table 1). The overall Ifi’ of Tree Swallows tled by Tree Swallows by first depositing twigs, ranged from 3 1.6% in 1983 to 99.3% in 1986. then laying eggsin new nests after the swallow Although daily mortality rates of whole clutches nests were destroyed. On four occasions, Tree (clutch m) were statistically similar between Swallows were observed chasing House Wrens speciesat the shared plot in the first three study away from swallow-occupied boxes. Tree Swal- years, Tree Swallow young suffered higher daily lows built nests and laid eggsin eight boxes that mortality than House Wrens in 1986 (Tables 1 contained twigs or dummy nestsplaced by wrens, and 2). Daily rates of partial clutch losses(partial and four additional swallow nests had twig nests rh) in wrens were similar to those in swallows in constructed over them. Eleven of these 12 swal- each year except 1986, a year in which partial low nests failed, and six failed during the egg- losseswere lower in swallows only because no laying stage.A second swallow clutch laid in one swallow clutches survived. Considering nests of thesetwig nestswas also destroyedafter failure pooled over all years, overall m’ at the shared of an earlier incomplete clutch. At Rock Creek, plot was 87.3% in swallows and 67.1% in wrens wren(s) intruded at one unsuccessfulswallow nest (Table 1). Results of combined-probability tests demonstrated that daily clutch m in swallows NESTING SUCCESSOF HOUSE WRENS AND TREE SWALLOWS 679

TABLE 1. Mean daily rates ? SE of whole clutch and partial clutch mortalities(m), numbersof days of exposure(EXP), and probabilitiesof clutchmortality through the nestingperiod (@) in 1983, 1984, 1985,and 1986in House Wrens and Tree Swallowsusing the same (shared)or separate(unshared, swallow-free) box areas.*

Unsharedarea Mortality rate HouseWren Tree Swallow HouseWren

1983 Clutch EXP 125 116 430 Partial EXP 758 464 2,219 Clutch m 0.008 + 0.008 0.000 + 0.000 0.005 + 0.003 Partial m 0.012 + 0.004 0.011 + 0.005 0.005 + 0.001 Total m’ 0.503 0.316 0.286 1984 Clutch EXP 127 94 529 Partial EXP 607 394 2,361 Clutch m 0.055 -t 0.020 0.106 k 0.032 0.006 f 0.003 Partial m 0.007 k 0.003 0.003 * 0.003 0.010 f 0.002 Total m’ 0.891 0.983 0.427 1985 Clutch EXP 239 102 313 Partial EXP 1,393 479 2,038 Clutch m 0.017 + 0.008 0.039 + 0.019 0.013 + 0.006 Partial m 0.004 + 0.002 0.002 * 0.002 0.006 t 0.002 Total mt 0.524 0.77 1 0.473 1986 Clutch EXP 86 30 493 Partial EXP 433 121 2,642 Clutch m 0.023 t- 0.016 0.133 k 0.062 0.006 f 0.004 Partial m 0.009 k 0.005 0.000 * 0.000 0.006 + 0.001 Total m’ 0.683 0.993 0.338 Total years Clutch EXP 577 342 1,765 Partial EXP 3,191 1,458 9,260 Clutch m 0.024 t 0.006 0.053 + 0.012 0.007 + 0.002 Partial m 0.007 * 0.002 0.005 + 0.002 0.007 * 0.001 Total rh’ 0.671 0.873 0.375 ’ The standarderror (SE) of IR is computedusing EXP, the numberof daysof exposurethat a clutchor individual egg/nestlingrisks mortality was significantly higher than that in wrens, but differed in 1984 and 1986, and in the all-year partial rh was similar between species(Table 2). test. Overall rh in wrens over all 4 years was 37.5% at the swallow-free plots compared to 67.1% at RELATIONSHIPS AMONG RATES OF BOX the swallow-used plot (Table 1). The greatest USE, INTERFERENCE, AND MORTALITY within-year difference was in 1984 when clutch The daily mortality rate (m) of House Wren rh through the entire nesting period was 89.1% clutches was positively associatedwith total oc- in wrens on swallow-used plots and only 42.7% cupancy rate (partial Y, = 0.67, P = 0.025) (Fig. in wrens on swallow-free plots (Table 2). The 2), but was unrelated to interference rate at wren outcome of the combined-probability test indi- nests (yp= 0.32, P = 0.335). That is, wren nests cated that daily clutch m in wrens on the shared were more likely to fail as box-occupancy rates, plot was significantly greater than that in wrens rather than interference rates, increased. Daily on the swallow-free plot. In comparisons be- clutch rh was not associatedwith wren occupancy tween swallows at Foote Camp and wrens oc- rate (P > 0.05). Incomplete clutch losses(partial cupying the swallow-free plots, daily clutch m m) in wren nests were negatively related to total differed significantly in 1984 and 1986, as well occupancy rate (Y, = -0.80, P = 0.003) and un- as in the combined-year test, while partial m related to interference rate at wren nests (r, = 680 DEBORAH M. FINCH

TABLE 2. Z-statistics and combined-probability tests comparing daily mortality rates (rh) between nests of (1) House Wrens and Tree Swallowsin the area shared by both species,(2) wrens on areasused by swallowsand wrens on swallow-free areas, and (3) wrens on swallow-free areasand swal10ws.~

Z-WlWS Compansonb 1983 1984 1985 1986 Combined-Ptest=

Shared area Wren vs. swallow Clutch rh 1.01 1.36 1.07 1.72* 11.14* Partial m 0.18 0.98 0.84 2.02* 10.34 Shared vs. unshared Wren vs. wren Clutch m 0.39 2.41* 0.38 1.03 12.09* Partial m 1.68* 0.92 0.44 0.74 5.80 Wren vs. swallow Clutch m 1.40 3.15** 1.31 2.05* 22.45*** Partial m 1.16 2.41* 1.25 4.01*** 30.39***

‘*Pr0.1,**P~0.0l,***P~0.001. bClutch I% meansdaily mortality rate of whole clutches;partial a meansdaily rate of clutch reductionat neststhat remain active. r Comparisonswere made between groups by combiningprobabilities across years using Fisher ’s procedure(S&al and Rohlf 198I). The teststatistic is distributedas x2 with 6 degreesof freedom(Zk, wherek = the numberof yearsand probabilities).Probabilities computed for 1983 wereexcluded from thesetests because 1983 was a transitionyear for nestingbirds.

-0.01, P = 0.973). In sum, wren clutches were m correspondingly declined (Y, = -0.94, P = more likely to be totally destroyed than partially 0.0001). Because variation in partial-loss rates lost as box-occupancy rates increased. may be contingent on variation in failure rates Although Tree Swallow nesting was primarily of whole nests, effects of interference rate and on one plot, preventing statistical comparisons occupancyrate on partial lossesshould probably of rates of interference and nesting failure, the be interpreted with caution. positive relationship between wren interference DISCUSSION rates and clutch mortality at swallow nests is clear (Fig. 3). I combined Tree Swallow nests EFFECTS OF CONSPECIFIC INTERFERENCE with wren nests to determine if the addition of AT WREN NESTS swallow nests influenced relationships among Conspecifics, rather than swallows, were pri- clutch mortality, interference rate, and total oc- marily responsible for interference at boxes occupancy rate. Daily clutch m of pooled species cupied by wrens. Some wren nests lost one or was positively related to interference rate (rp = more eggsdue to interference, but were successful 0.92, P = 0.0001) (Fig. 4) but was unrelated to in producing fledglings. House Wrens achieved total occupancy rate (rp = -0.50, P = 0.117). highestrates of brood survival on the plots where Partial m at all nests was not associated with box use by swallows was absent or rare. Preda- either interference rate (r, = -0.19, P = 0.569) tion, rather than nest interference, explained high or occupancyrate (yp= -0.53, P = 0.093). Thus, rates of total nest failure in House Wrens at the due to the strongeffect of interference at swallow swallow-usedplot. In contrast, Freed (1986a) re- nests,the independent variable affecting survival ported that conspecific destruction of whole ofwhole broods shifted from box-occupancy rate clutchesby tropical House Wrens was common. (wren nests only) to interference rate at pooled Although box-use rates were higher at Foote nests of swallows and wrens. Camp than at the swallow-free plots, interference Factors related to failure appeared to affect rates at wren nests did not vary among plots, probability of partial lossesand mortality rates suggestingthat interference was not directly ex- of whole clutches in opposite directions. In a plained by reduced availability of nest sites. In partial correlational analysis of these two mor- densepopulations of tropical House Wrens, mates tality components with interference rate at all rather than nest sites per se are in short supply; nests, interference rate increased as daily clutch unmated wrens vigorously compete for mates by m increased (T,,= 0.98, P = O.OOOl),and partial destroying eggs,usurping territories, and mating NESTING SUCCESS OF HOUSE WRENS AND TREE SWALLOWS 681

FO.67, p=o.o26. Partial -r=om, -P=O.oOl. Pam61 2-

2 . . . z 1 1 1 l- .

s O

[ -l-

-21 ’ ’ ’ ’ ’ ’ ’ 1 -2 -1 0 1 2 -2 I 0 1 2 I- nts al all nesta

FIGURE 2. Standardized partial regressionplot for FIGURE 4. Standardized partial regressionplot for daily mortality rate of House Wren clutchesand ratio daily mortality rate in nests of both speciesand inter- of the number of all nests to the number of boxes. ference rate at all nests. Residualsof the two variables were plotted after removing the effectsof PLOT, YEAR, Residuals of the two variables were plotted after re- and box occupancyrate. moving the effects of PLOT, YEAR, and interference rate at House Wren nests. Nor was the probability of partial wren losses with territory owners (Freed 1986a, 1986b). In related to interference rates, possibly becausein- this study, not all box-nesters were conspecifics terference occurred at failed whole nests as well which may explain the lack of relationship be- as at nests that remained active after clutch re- tween box-use rates and wren-nest interference. duction. In addition, partial losseswere caused I found little evidence that swallows intervened by multiple factors (e.g., nestling starvation, egg at wren nests, and therefore, I did not expect and nestling mortality at hatching, predation) so increasedrates of box use (due in part to presence that the effectsof any single factor were obscured. of swallowsat Foote Camp) to result in increased RELATIONSHIP BETWEEN BOX-USE interference at wren nests. Interference effectson RATE AND PREDATION AT wrens may be more important when wrens alone WREN NESTS occupy a high percentage of boxes, resulting in Total nest failure in House Wrens was strongly increased opportunities for intraspecific inter- associated with rates of box occupancy. Clutch actions like infanticide, brood parasitism, extra- mortality rates in wrens were especially high, pair copulations, bigamy, nest confiscation, and ranging from 50-89% over the 4-year period, at mate theft. the box-plot with highest nest densities. Nests that may have survived with partial lossescaused by interference or starvation at low box-use rates iz were likely to be totally destroyed by predators E *O‘ at high occupancy rates. Predators usually de- 0 0.6 stroy whole clutches rather than leave any sur- 5 viving offspring (Brown and Brown 1988). Pre- 8 r 0.6 dation was the major cause of nest failure, accounting for 70% of all unsuccessful nesting E 0.4 attempts by wrens. I detected signs of nest pre- 6 dation by red squirrels (Tamiasciurus hudsoni- 9 OP cus), raccoons(Procyon fotor), and small rodents 0 (unknown species)(Finch 1989~). Weasels (MUS- 1983 1234 13% 1336 tela frenata) and bullsnakes (Pituophis melano-

mcunp leucussayi), other predators that destroy House FIGURE 3. Interference rates and probability of Wren nests in riparian woodlands of Wyoming overall clutch mortality of Tree Swallows at Foote (L. Scott Johnson, pers. comm.), were also ob- Camp. served in my study areas. 682 DEBORAH M. FINCH

Field experiments have demonstratedthat high EFFECTS OF WREN INTERFERENCE AT nest densities resulted in heightened predation SWALLOW NESTS rates (Tinbergen et al. 1967, Goransson et al. Male House Wrens build and guard multiple 1975, Page et al. 1983). Andersson and Wiklund empty nests within their territories, but each fe- (1978) suggestedthat spacing out of nests was a male selects one site for egg laying (Kendeigh predator avoidance strategy, but Picman (1988) 194 1). Multiple-nest defense by males may be a concluded that dispersing nests widely apart was tactic for attracting mates (Kendeigh 1941), for only advantageous in habitats with high preda- increasing opportunities for bigamy (Freed tion rates. Because deciduous riparian wood- 1986b), or for ensuring the availability of surplus lands typically have higher bird densities than nest sites if the first nesting attempt fails. But surrounding habitats (Hehnke and Stone 1978, high densities of dummy nests may also protect Johnson and Haight 1985) rates of nest preda- actual nest sites from search-strategypredators, tion may be high. For instance, Janzen (1978) as is evident in Marsh Wrens (Leonard and reported higher rates of egg mortality in decid- Picman 1987). If Tree Swallows nest in boxes uous woodlands of Costa Rica than in other hab- defended by male House Wrens, then a dummy- itat types. Thus, spacingnests out by minimizing nest strategy is negated by the presenceof swal- occupancyrate of boxes could have adaptive val- low eggs.Wrens began building nests in boxes ue in the cottonwood woodlands of this study. on my study areas 6 days earlier on average than Reduced predation rates are associated with swallows (two-way ANOVA with factors year increased numbers of empty nests in Northern and species;species effect: P < 0.05). Although Cardinals (Cm-din& cardinalis) (Watts 1987) Tree Swallows generally arrive on North Amer- and Marsh Wrens (Cistothoruspalustris) (Leon- ican breeding grounds earlier than House Wrens ard and Picman 1987). Likewise, a high propor- (G. Holroyd, pers. comm.), territory-settlement tion of empty boxes may inhibit predator search dates in swallows are not related to timing of egg efficiency.In this study, destructionrates of House laying (Stutchbury and Robertson 1987b). Late- Wren nests declined as the proportion of vacant breeding swallowsnested unsuccessfullyin seven boxes increased. If predators use nest boxes as of eight boxes that contained dummy nests built search images based on previous rewards, then by wrens. If swallows nest in an “empty” box risk of predation may be highest when the ratio that is defended by a multiple-nest House Wren, of empty-to-occupied boxes is low. Martin and then the wren may interfere with the swallow Roper (1988) similarly concluded that the pre- nest during a territorial visit. If some or all eggs dation rate on open nests of Hermit Thrushes are destroyed by wrens, swallows may abandon (Cuthurus guttutus) was influenced by the asso- the nest site, and then the box will return to its ciation between unused sites and used sites. original condition, that of a dummy nest. In this study, I detected signsthat a predator(s) In this study, wren interferenceat swallow nests sequentially destroyed nests in multiple years at often preceded or accompanied whole-clutch the plot shared by both species.Its dexterity in failure. Destroying active swallow nests has ad- unlatching boxes, ripping off box lids, and tearing vantages for wrens if the function of a dummy up nestsimplicates the raccoon, a known search- nest is to deceive predators. Just as predation strategy predator (Bowman and Harris 1980). rates are reduced in areas where old nests have Raccoon tracks were observed along the river accumulated (Watts 1987), so too might search banks bounding the box grid, and claw marks efficiency of predators be inhibited by high con- were visible on some nest boxes. Serial predation centrations of empty swallow nests and wren in response to high reward rate set by the high dummy nests. Becausecavity-nest predators are density of box nests may explain increased nest- also likely to prey upon the contents of open ing mortality in wrens at Foote Camp. Following nests, high densities of empty nests of any type 1984, the year of peak mortality, fewer boxes may benefit House Wrens. By destroying the eggs had nests at Foote Camp, suggestingeither that of open-nesting and cavity-nesting birds (Ken- box-nesters adopted a predator-avoidance strat- deigh 194 1, Belles-Isles and Picman 1986b) and egy (i.e., shifting to new cavities, Sonerud 1985) forcing pairs to nest further away, House Wrens or that bird populations were lower due to high may causereduced predation rates in the locality nest failure the year before. of their own nest sites. NESTING SUCCESS OF HOUSE WRENS AND TREE SWALLOWS 683

Other possible advantages to destroying eggs may be naturally selected if the trait results in in hole nests include: (1) cannibalism and diet fewer behavioral errors than one involving a rea- augmentation (Belles-Isles and Picman 1986b, soning process(i.e., deciding to destroy or not). Brown and Brown 1988), (2) freeing the cavity In other words, harming othersmay benefit wrens so that a House Wren pair can readily renest in more often than not. But, if harming others does a surplus site if their first attempt fails; (3) for not render manifest or potential benefits to the males, attracting females by displaying superior harmer, it can be interpreted as spiteful behavior mate quality as advertised by his ability to de- (sensu Hamilton 1970). Though spiteful behav- fend surplus cavities (Kendeigh 1941); (4) re- ior is feasible in interspecific relationships where ducing risks of nest usurpation by spacing nests kinship is zero (Hamilton 1970) it is an unlikely of secondary cavity-nesters more widely apart; reason for nest destroying behavior by wrens be- (5) reducing risks of eggdestruction, eggdumping causedestruction is nonselective, i.e., directed at (Price et al. 1989) extra-pair copulations (John- conspecificsand heterospecificswho may or may son and Kermott 1989), and mate theft by in- not possessthe trait; it seems to be inherent to creasingdistances between conspecificpairs (e.g., all wrens (Belles-Islesand Picman 1986b); it car- Tree Swallows, Muldal et al. 1985); (6) reducing ries the potentially high cost of retaliation; and nest-mate competition in broods parasitized by it affords many plausible advantages. Possibly, conspecifics(sensu Brown and Brown 1988, see no single factor or advantage is responsible for also Lombard0 et al. 1989), and (7) reducing the selection of egg puncturing by male and fe- local competition for other resources like food male wrens. or foraging substrate (Belles-Isles and Picman Interspecific competition for limited numbers 1986b). The seventh point is also an alternative of nest sitesis yet another factor that might cause explanation for destruction of open nests(Belles- House Wrens to interfere with Tree Swallow nests. Isles and Picman 1986b). Unlike the empty-nest If swallows occupy a high proportion of boxes, hypothesis, however, the food-competition hy- then wrens have fewer nest sites from which to pothesisdoes not account for indiscriminate de- choose.Male House Wrens have high return rates struction of open nests by House Wrens regard- to breeding sites that they occupied the year be- less of whether their diet or foraging habits are fore (Drilling and Thompson 1988). Returning similar to that of the “competing” species.Nor wrens may try to usurp occupied boxes within does it explain why nest-destroying behavior is their territories. Despite intrusions at swallow suppressedin males after pairing and in females nests, however, only a few swallow boxes were after egglaying when energy requirements should actually confiscatedby nesting wrens, suggesting be higher (Belles-Isles and Picman 1986b). that wrens did not destroy most swallow nests As no wrens have ever been observed to eat for purposes of unsurpation. By building nests the eggsthat they punctured or ejected (Belles- earlier, House Wrens seemingly avoided agonis- Isles and Picman 1986b), it is safe to rule out tic interactions with swallows. Wrens may build diet supplementation as a major reason for egg nestsearlier than swallowsto reducecompetition destruction. Though House Wrens may benefit for nest cavities among conspecifics or among from the destruction of hole nests for the other species.Stutchbury and Robertson (1987b) sug- six reasons listed, these reasons do not readily gestedthat swallows in Ontario timed their egg explain the evolution of eggdestruction because laying for mid-May to benefit from synchronous they do not account for the destruction of open breeding and favorable environmental condi- nests. Factors two, four, and five explain intra- tions. Also, because swallows typically choose specific interactions only. Nor does factor two nest sites in open meadows and woodland edges explain why female wrens (not just males) de- (Munro and Rounds 1985), interior woodlands stroy nests. And factor six does not answer why may be less preferred by them. Hence, swallows wrens destroythe nestsofTree Swallows,a species that breed later than average in woodlands may that foragesin the air on a different food supply. be floaters using suboptimal sites (e.g., Stutch- Perhaps House Wrens are genetically “pro- bury and Robertson 1987a). Lack of swallow grammed” to destroy a neighbor’s eggson sight nesting on two of three study areas supports the even when there is no immediate advantage. In- interpretation that swallows avoided nesting in discriminate destruction of open and hole nests interior woodlands where wrens were abundant, 684 DEBORAH M. FINCH even though they were observedin adjacent open and (3) having higher nesting success.Interfer- areas. ence competition, defined as immediate exclu- In conclusion, interference by House Wrens at sion of a competing individual from a resource Tree Swallow nests may be more related to ter- (Ricklefs 1979), was the most obvious type of ritorial defense of dummy nests and nest-pro- interspecific competition in this study. House tection tacticsthan to attempts to usurp nest sites. Wrens excluded Tree Swallows from nest boxes Although I have listed numerous possible expla- by destroying swallow nests,forcing them to nest nations, an experimental approach is needed to elsewhere (see also Belles-Isles and Picman factor out the underlying reason(s) for wren in- 1986b). terference and egg destruction. To test the hy- After the first year, interference rates were pothesisthat high concentrations of empty nests greater at swallow nests than at wren nests and discourage predation at active nests, I recom- were associated with low survival of swallow mend that densities of empty nests (in boxes or clutches. This evidence suggeststhat swallows in the open) be experimentally increasedin study were less successfulthan nesting wrens at de- areas containing multiple active nests of House fending nests from intruders (but see Kuerzi Wrens or an alternate species.The null hypoth- 194 l), or were less likely or less able to invade esis of no effect on predators is rejected if nest wren nests.Although Tree Swallows destroy eggs predation is (1) higher in study areas with no and usurp territories of conspecifics(Harris 1979), treatment (spatial control) than in treatment they have not been reported to regularly destroy areas, and (2) higher in the same areas in year(s) the nests of other species (but see Butler and prior to treatment (temporal control) than in Campbell 1987). Furthermore, swallows appar- treatment year(s). Both temporal and spatial con- ently do not defend multiple surplus nests from trols are recommended to account and adjust for other species (Robertson and Gibbs 1982, but natural variation in nesting successby area and seeRendell and Robertson 1989), although they year. Testing the reverse hypothesis, that pre- can prevent conspecific accessto empty boxes dation at wren nests is increased in areas having through territorial defense of active nests (Rob- high densities of nestswith prey, will not support ertson et al. 1986). By generalizing nest-destroy- or refute the empty-nest hypothesis. But, if this ing behavior to include other species(Belles-Isles second hypothesis is also supported, then nest- and Picman 1986b), possibly in relation to mul- destroyingbehavior hastwo advantagesfor House tiple nest control, House Wrens outcompete Wrens, i.e., creating empty neststhat deter pred- swallows for boxes. ators and eliminating active nests that facilitate Exploitation competition, in which resource predation. The ratio of empty to active nestsmay availability is diminished for some animals then act as a dynamic continuum that selectsfor through use by others (Alatalo et al. 1987) may the evolution of eggdestruction by House Wrens. operate secondarily in wren-swallow relation- The study design for the active-nest hypothesis ships. For example, by occupying boxes earlier would parallel that used for the empty-nest hy- and by having higher population levels than pothesis, expect that real eggs (e.g., quail eggs) swallows (Finch 1989b), House Wrens may in- would be added to experimental nests. directly reduce the availability of nest sites for Tree Swallows. Establishment of dummy nests ASYMMETRICAL COMPETITION BETWEEN by wrens also may deter swallows from nest-box WRENS AND SWALLOWS settlement. Results from this study cannot sup- By destroying swallow nests and controlling sur- port or refute the concept of exploitation com- plus nest sites, House Wrens prevent accessand petition because they are subject to other inter- successfulbreeding by swallows; such interac- pretations; however, the results from this study tions are usually consideredcompetitive (Dhondt do offer new directions for future experimental and Eyckerman 1980). Asymmetrical competi- tests. tion occurswhen one speciesis more negatively affected than the other (Connell 1983, Schoener MANAGEMENT IMPLICATIONS 1983, Persson 1985). In this study, House Wrens Interspecific competition may be minimized dominated interactions with swallows by (1) de- when cavity resourcesare partitioned in space. stroying eggsand nests of swallows, (2) having When boxes are distributed in a wide variety of few or no nestsdamaged or usurped by swallows, habitats, Tree Swallows readily choosethose that NESTING SUCCESS OF HOUSE WRENS AND TREE SWALLOWS 685 are in or adjacent to open areas (Rustad 1972, spatialheterogeneity on ground-nestdepredation. Munro and Rounds 1985), whereasHouse Wrens J. Wildl. Manage. 44:806-8 13. BRAWN,J. D. 1987: Density effects on reproduction prefer boxes that are in tree and shrub habitats of cavity nesters in northern Arizona. Auk 104: (Willner et al. 1983, Munro and Rounds 1985). 783-787. House Wrens also selectnatural cavities in short- BRAWN,J. D., ANDR. P. BALDA. 1987. Investigations er trees and snagsthan those preferred by Tree of density interactions among breeding birds in Swallows (Stauffer and Best 1982). By introduc- ponderosa pine forests: correlative and experimental evidence. Oecologia 72:348-357. ing nest boxes, competitive interactions between BRAWN,J. D., AND R. P. BALDA. 1988. Population speciesmay be artificially increased, particularly biology of cavity nestersin northern Arizona: do when boxes are placed at the same height or in nest sites limit breeding densities?Condor 90:6 l- habitats where both speciescoexist. Boxes may 71. BROWN,C. R., AND M. B. BROWN. 1988. The costs be more attractive or more easily found than and benefits of egg destruction in colonial Cliff natural nest sites, resulting in shifts away from Swallows. Auk 105:737-748. use of natural cavities (Gustafsson 1988). More- BUTLER,R. W., AND C. A. CAMPBELL.1987. Nest over, population densities of secondary cavity- appropriation and interspecific feeding between nesters often increase after introduction of nest Tree Swallows, Tachycineta bicolor, and Barn Swallows, Hirundo rustica. Can. Field-Nat. 101: boxes (Gustafsson 1988). Thus, establishment of 433434. nest-box plots is an excellent means of experi- CONNELL,J. H. 1983. On the prevalenceand relative mentally inducing interspecific competition. On importance of interspecificcompetition: evidence the other hand, if the purpose of nest-box estab- from field experiments. Am. Nat. 127:819-834. DHONDT,A. A., AND R. EYCKERMAN. 1980. Com- lishment is to enhance habitats for cavity-nesting petition between the Great Tit and Blue Tit out- species, then standardized placement of boxes side the breedingseason in field experiments.Ecol- may not provide suitable nest-site resourcesfor ogy 61:1291-1298. targeted species, especially if some species ex- DRILLING,N. E., AND C. F. THOMF~ON. 1984. The clude boxesfrom others. By spacingboxes widely use of nest boxes to assessthe effect of selective loggingon House Wren populations, p. 188-196. apart in a broad range of habitat patches and In W. C. McComb [ed.], Proceedingsof the work- heights, managers may increase box accessand shopon managementof nongamespecies and eco- speciesuse. logical communities. Univ. of Kentucky, Lexing- ton. ACKNOWLEDGMENTS DRILLING,N. E., AND C. F. THOMPSON.1988. Natal and breedingdispersal in House Wrens (Troglody- I would like to thank P. F. Gutzwiller, G. J. Sherman, tes aedon). Auk 105:480-49 1. K. A. Conine, C. L. Canaday, and R. D. Greer for DUNN, E. 1977. Predation by weasels (Mustela ni- assistancein checking nest boxes and recording data. valis) on breeding tits (Parus spp.)in relation to I am grateful to A. L. Ward for his advice and en- the densityof tits and rodents.J. Anim. Ecol. 46: couragement during the study. For reviewing the 634-652. manuscript, I thank W. J. Arendt, J. D. Brawn, R. N. FINCH, D. M. 1987. Bird-habitat relationshipsin ri- Conner, N. E. Drilling, S. J. Hejl, G. L. Holroyd, H. parian communities of southeastern Wyoming. W. Power, and B. J. Stutchbury. Ph.D.diss., Univ. of Wyoming, Laramie. LITERATURE CITED FINCH,D. M. 1989a. Habitat use and habitat overlap of riparian birds in three elevational zones. Ecol- ALATALO,R. V., D. ERIKSSON,L. GUSTAFSSON,AND K. ogy 70:866-880. LARSSON.1987. Exploitation competition influ- FINCH,D. M. 1989b. Speciesabundances, guild dom- encesthe use of foragingsites by tit&experimental inance patterns, and community structure of evidence. Ecology 68:284-290. breedingriparian birds, p. 629-645. In R. R. Shar- ANDERSSON,M., ANDC. G. WIKLUND. 1978. Clump- itz and J. W. Gibbons [eds.], Freshwater wetlands ing versus spacingout: experiments on nest pre- and wildlife: perspectives on natural, managed, dation in fieldfares(Turdus nilaris). Anim. Behav. and degradedecosystems. DOE-CONF 8603 101. 22:521-567. - ’ Office Sci. and Tech. Inform., U.S. Dept. Energy, BELLES-ISLES,J. D., AND J. PICMAN. 1986a. Nesting Oak Ridge, TN. lossesand nest site preferencesin House Wrens. FINCH, D. M. 1989~. Relationships of surrounding Condor 88:483-486: riparian habitat to nest-box use and reproductive BELLES-ISLES,J. D., AND J. PICMAN. 1986b. House outcome in House Wrens. Condor 9 1:848-859. Wren nest-destroyingbehavior. Condor 88: 190- FREED,L. A. 1986a. Territory takeover and sexually 193. selected infanticide in tropical house wrens. Be- BELLES-ISLES,J. D., ANDJ. PICMAN. 1987. Suspected hav. Ecol. Sociobiol. 19:197-206. adult intraspecifickilling by House Wrens. Wilson FREED,L. A. 1986b. Usurpatory and opportunistic Bull. 991497-498. bigamy in tropical housewrens. Anim. Behav. 34: BOWMAN,G. B., AND L. D. HARRIS. 1980. Effects of 1894-1896. 686 DEBORAH M. FINCH

GARDNER,A. F. 1925. A summer of bluebird tra- LEONARD,M. L., ANDJ. PICMAN. 1987. The adaptive gedies. Bird-lore 27:24 1. significanceof multiple nestbuilding by male marsh GORANSSON,G., J. KARLSSON, S. G. NILSSON,AND S. wrens. Anim. Behav. 35~271-277. ULFSTRAND. 1975. Predation on birds’ nests in LOMBARE+ M. P. 1986. A possible case of adult relation to antipredator aggressionand nest den- intraspecific killing in the Tree Swallow. Condor sity: an experimental study. Oikos 26: 117-120. 88:112. GUSTAFSSON,L. 1987. Interspecificcompetition low- LOMBARDO,M. P. 1988. Evidence of intraspecific ers fitness in Collared Flycatchers Ficedulu albi- brood parasitism in the Tree Swallow. Wilson. collis:an experimental demonstration.Ecology 68: Bull. 100:126-128. 291-296. LOMBARDO, M. P., L. C. ROMAGNANO, A. S. GUSTAFSSON,L. 1988. Inter- and intraspecific com- HOFFENBERG,P.C. STOUGGER,AND H. W. POWER. petition for nest holes in a population of the Col- 1989. Egg removal and intraspecific brood par- lared Flycatcher Ficedulu albicollis. Ibis 130:1 l- asitism in the European Starling (Sturnus vulgur- 16. is). Behav. Ecol. Sociobiol. 241217-223. HAMILTON,W. D. 1970. Selfish and spiteful behav- Low, S. H. 1933. Further notes on the nesting of the iour in an evolutionarv model. Nature 228: 12 18- Tree Swallows. Bird-Bandine 3:76-87. 1220. MARTIN,T. E., ANDJ. J. ROPER. l-988. Nest predation HARRIS,R. W. 1979. Aggression,superterritories, and and nest-site selection of a western population of reproductivesuccess in Tree Swallows.Can. J. Zool. the Hermit Thrush. Condor 90:51-57. 57:2072-2078. MAYFIELD,H. F. 1961. Nesting successcalculated HEHNKE,M., AND C. P. STONE. 1978. Value of ri- from exposure. Wilson Bull. 73:255-261. parian vegetation to avian populations along the MAYFIELD,H. G. 1975. Suggestionsfor calculating Sacramento River System, p. 228-235. In R. R. nest success.Wilson Bull. 87:456466. Johnson, and J. F. McCormick [eds.], Strategies MULDAL,A., H. L. GIBBS,AND R. J. ROBERTSON.1985. for protection and managementof floodplain wet- Preferred nest spacingofan obligatecavity-nesting landsand other riparian ecosystems.USDA Forest bird, the Tree Swallow. Condor 87:356-363. Service,Gen. Tech. Rep. WOL12, Washington,DC. MUNRO, H. L., AND R. C. ROUNDS. 1985. Selection HOLROYD,G. L. 1975. Nest site availability as a fac- of artificial nest sitesby five sympatric passerines. tor limiting populationsize of swallows.Can. Field- J. Wildl. Manage. 491264-276. Nat. 89:60-64. NILLSON,S. G. 1984. The evolution of nest-site se- INGOLD,D. J. 1989. Nesting phenology and com- lection among hole-nesting birds: the importance petition for nest sitesamong Red-headed and Red- of nest predation and competition. Omis Stand. bellied woodpeckersand European Starlings.Auk 15:167-175. 106:208-217. PAGE,G. W., L. E. STENZEL,D. W. WINKLER,AND C. JANZEN,D. H. 1978. Predation intensity on eggson W. SWARTH. 1983. Spacing out at Mono Lake: the ground in two Costa Rican forests.Am. Midl. breeding success,nest den&y, and predation in Nat. 100:467470. the Snowy Plover. Auk 100:13-24. JOHNSON,D. H. 1979. Estimating nest success:the PERSSON,L. -1985. Asymmetrical competition: are Mayfield method and an alternative. Auk 96:65 l- larger animals competitively superior?Am. Nat. 661. 126:261-266. JOHNSON,L. S., AND L. H. KERMOTT. 1989. Terri- PICMAN,J. 1988. Experimental study of predation on torial intrusions in the House Wren Troglodytes eggsof ground-nestingbirds: effectsof habitat and aedon: evidence for the sperm competition hy- nest distribution. Condor 90: 124-l 3 1. pothesis. Omis Stand. 20:89-92. PICMAN,J., AND J. C. BELLES-ISLES.1988. Evidence JOHNSON,R. R., AND L. T. HAIGHT. 1985. Avian use for intraspecific brood parasitism in the House of xeroriparian ecosystemsin the North American Wren. Condor 90:5 13-5 14. warm deserts, p. 156-160. In R. R. Johnson, C. PRICE, D. K., G. E. COLLIER,AND C. F. THOMPSON. D. Ziebell, D. R. Patton, P. F. Ffolliott, and R. H. 1989. Multiple parentagein broodsof housewrens: Hamre [eds.],Riparian ecosystemsand their man- genetic evidence. J. Hered. 80: l-5. agement: reconciling conflicting uses.First North QUINNEY,T. E. 1983. Tree Swallowscross a polygyny American rinarian conference.USDA Forest Ser- threshold. Auk 100:750-754. vice, Rocky-Mountain Forest and Range Experi- RENDELL,W. B., AND R. J. ROBERTSON.1989. Nest- ment Station, Gen. Tech. Rep. RM-120. site characteristics,reproductive successand cav- KENDEIGH, S. C. 1941. Territorial and mating be- ity availability for Tree Swallowsbreeding in nat- havior of the House Wren. Ill. Biol. Monogr. 18: ural cavities. Condor 91:875-885. l-20. RICKLEFS,R. E. 1979. Ecology. 2nd ed. Chiron Press, KREBS,J. R. 1971. Territory and breeding density in New York. the great tit, Parus major L. Ecology 52~2-22. ROBERTSON,R. J., AND H. G. GIBBS. 1982. Super- KUEP~, R. G. 1941. Life history studies of the tree territoriality in Tree Swallows: a reexamination. swallow. Proc. Linn. Sot. N.Y. 52-53: 1-52. Condor 84:313-316. LEFFELAAR,D., AND R. J. ROBERTSON.1985. Nest ROBERTSON,R. J., H. L. GIBBS,AND B. J. STUTCHBURY. usurpation and female competition for breeding 1986. Spitefulness, altruism, and the cost of opportunities in the Tree Swallow. Wilson Bull. aggression:evidence against superterritoriality in 97~221-224. Tree Swallows. Condor 88:104-105. NESTING SUCCESS OF HOUSE WRENS AND TREE SWALLOWS 687

ROBERTSON,R. J., ANDB. J. STUTCHBURY.1988. Ex- nest building and first egg dates reflect settlement perimental evidence for sexually selectedinfanti- patterns of females?Condor 89:587-593. cide in tree swallows. Anim. Behav. 36~749-753. TINBERGEN,N., N. IMPEKOVEN,AND D. FRANCK. 1967. RUSTAD,0. A. 1972. Tree swallow nesting study on An experiment on spacing-outas a defenseagainst a bluebird trail in south central Minnesota. Loon predation. Behaviour 28:307-32 1. - 44:100-105. TOMPA.F. C. 1967. Reoroductive successin relation SCHOENER,T. W. 1983. Field experiments on inter- to breeding density in Pied Flycatchers,Ficedula specific competition. Am. Nat. 70: 122-l 85. hypoleuca(Pallas). Acta Zool. Fenn. 118:3-28. SHERMAN.A. R. 1925. Down with the House Wren VAN BALEN,J. H., C.J.H. BOOY,J. A. FRANEKER,AND boxes. Wilson Bull. 37:5-l 3. E. R. OSIECK. 1982. Studieson hole nestingbirds SOKAL,R. R., ANDF. J. ROHLF. 1981. Biometry. 2nd in natural nest sites.I. Availability and occupation ed. W. H. Freeman and Co.. New York. of natural nest sites. Ardea 70: l-24. SONERUD,G. A. 1985. Nest hole shift in Tengmalm’s VON HAARTMAN,L. 1957. Adaptation in hole-nesting owl Aegoliusjiunereusas defenseagainst nest pre- birds. Evolution 11:339-347. dation involving long term memory of the pred- W~rrs, B. D. 1987. Old nest accumulationas a pos- ator. J. Anim. Ecol. 54: 179-l 92. sible protective mechanism against search-strat- STAUFFER,D. F., AND L. B. BEST. 1982. Nest-site egy predators. Anim. Behav. 35: 1566-1568. selection by cavity-nesting birds of riparian hab- WILLNER.G. R.. J. E. GATES.AND W. J. DEVLIN. 1983. itats in Iowa. Wilson Bull. 94:329-337. Nest box use by cavity-nesting birds. Am. Midl. STUTCHBURY,B. J., ANDR. J. ROBERTSON.1987a. Be- Nat. 109:194-201. havioral tactics of subadult female floaters in the YAHNER,R. H. 1983/1984. Avian use of nest boxes tree swallow. Behav. Ecol. Sociobiol. 20:4 13-4 19. in Minnesota farmsteadshelterbelts. J. Minn. Acad. STUTCHBURY,B. J., ANDR. J. ROBERTSON.1987b. Do Sci. 49: 1S-20.