Nest Predation and Its Relationship to Habitat and Nest Density in Dickcissels

The Condor 86:68-72 0 The Cooper Ornithological Society I984

NEST PREDATION AND ITS RELATIONSHIP TO HABITAT AND NEST DENSITY IN DICKCISSELS

JOHN L. ZIMMERMAN

ABSTRACT.-Analysis of the histories of over 500 Dickcissel (Spiza americana) nests found in eastern Kansas showed that those in old-field habitats suffered more predation than those in prairies. Predation rates on the prairie, were not correlated with the weeks of the nesting season, but those in old-fields varied significantlywith time. Although both predation rates and nest densitiesincreased concurrently to a peak during the middle of the nesting seasonin old-fields, an analysis of the relationship between nest densities per week and both daily predation rates and the percent of nests lost to predators each week indicated that predation was not density-dependent. Predation rates are higher in old-fields than in prairies, not because of greater nest densities, but perhaps becausepredators are more abundant in old-fields. Snakesare the most probable nest predator, and their method of hunting, by chance encountersrather than by directed search, is suited to the absence of a density-dependent effect of predation on Dickcissel nests.

Overall mortality rates in avian populations ity of Dickcissels,and both the frequency and may be density-dependent(Ricklefs 1973) but intensity of parasitism are inversely related to “it is not a popular notion that [predation on the density of available nests (Zimmerman nest contents] . . . operatesin a density-depen- 1983). Hence, Dickcissels nesting at low den- dent fashion to regulatepopulation size” (Cody sity in prairie habitat suffer high parasitism 197 1:486). Indeed, von Haartman (197 1) con- (Elliott 1978). Productivity per nest, however, cluded that there was no evidence that nest is similar in prairies and old-fields (Zimmer- lossdue to predation was greater in densepop- man 1982). This suggeststhat predation rates ulations. Since the publication of thesereviews in the high-density old-fields are greater, off- some data have accumulated that support the setting the lowered effect of cowbird parasit- idea that nest predation is density-dependent ism, while in the low-density prairies, the high in both cavity-nesting (Krebs 197 1) and open- loss caused by cowbird parasitism is offset by nesting species(Fretwell 1972, Goransson et low predation rates. I present this analysis of al. 1975). An experiment specifically designed nesting successin the Dickcissel in order to to test this hypothesis, however, did not sup- test the hypothesis that nest predation is den- port the density-dependent conclusion (Gott- sity-dependent, as these data suggest. fried 1978). Although Fretwell and Shipley (198 1) argued that Gottfried’s failure to dem- METHODS onstrate density-dependence was an effect of I obtained nest history data from 1965 through sample size rather than a biological reality, 1979 from over 500 nests in populations on both Gottfried (1978) and Best (1978) sug- the Ft. Riley Military Reservation and the gested that the kind of predator may be im- Konza Prairie ResearchNatural Area in Riley portant in determining whether or not nest and Geary countiesof Kansas.Most nestswere predation is density-dependent. In the studies found by watching the females. Except for a showing density-dependence, the predators few populations where nestswere visited every (Mustela, Corvus, Larus, Cyanocitta) hunted three or four days, most data were gathered by visually. For suchanimals the development of weekly visits to the nests. For nests expected a search image (permitting goal-oriented for- to fledge young between visits, I determined aging) probably depends upon the density of the fate of the nests by the female’s behavior. the prey items (Tinbergen 1960). In Gottfried’s At successfulnests the female remained near- old-field, however, the major predators were by, feeding the young and displaying aggres- snakes(Gottfried and Thompson 1978) which sively towards me. If a nest was lost since the probably find nests randomly (Best 1978). last visit, the female was always gone. Nesting The density of Dickcissels (Spiza ameri- successdata are presented as survival rates cana) varies accordingto habitat, being higher (Mayfield 196 1, 1975) with the 95% confi- in old-fields and lower in prairies (Zimmerman dence limits calculated by Johnson’s (1979) 197 1). Parasitism by the Brown-headed Cow- method. For three old-field and two prairie bird (Molothrus ater) depressesthe productiv- populations, concurrent studies of territorial- 1681 NEST PREDATION IN DICKCISSELS 69

tion exceeded 50% during the week of 13-19 July, although the maximum proportion of weekly nest loss occurred the previous week. The variances in the temporal distribution of the proportions of nests lost to predators in the two habitats, however, were not significantly different. g 20- The daily mortality rates due to predation 0” showed a similar peak in mid-July in the old- % 15- field, coming soon after the peak period when new nestswere started (Fig. 2). The quadratic relationship between week of nesting and pre- 10 - dation rate is significant (F = 18.96; df = 2,8; P < 0.01). Best (1978) also showed a positive 5 relationshipbetween the number of active nests in an old-field and the intensity of predation. 1 0 L In the prairies,daily predation ratesalso peaked m” r soon after the period of most frequent nest starts, but the daily predation rate showed no temporal pattern. WEEK OF NESTING RELATIONSHIP OF PREDATION AND FIGURE 1. Percent frequencydistribution of predation OTHER CAUSES OF LOSS TO THE nest lossesby week of loss in prairie (open bars) and old- NESTING CYCLE field (cross-hatchedbars) habitats. Number at top of each histogram is the number of nests lost. In both old-fields and prairie, the daily mortality rate attributed to causesother than predation, principally cowbird parasitism, was greatestduring the egg-layingperiod (Table 1). ityand matingpattemsin 1967, 1968 and 1979 These rates exceededpredation rates during all provided an intensity of coverage that makes phasesin the nesting cycle. In the prairie this me confident that most nestswere found. For value was significantly different from non- these populations it was then possible to cor- predatory lossesduring incubation (F = 5.28; relate nest densities, determined on a weekly df = 10, 10; P < O.Ol), while in the old-fields basis, with predation. this value was significantlydifferent from loss- RESULTS es due to other causesduring both incubation (F = 28.4; df = 34,3 1; P < 0.0 1) and brooding PHENOLOGY OF NEST PREDATION (F = 31.5; df = 42,7; P < 0.01). Comparing Nesting in Kansas old-field populations began habitats, mortality rates due to other causes in the third week of May and continued until during the incubation and brooding phases mid-August. The cumulative frequency of nest were not significantly different. lossesdue to predation surpassed50% by the The highestdaily predation rate occurred in week of 6-12 July; the weekly proportion of the old-field during incubation. This value was predation lossesreached a maximum the fol- significantly different from the daily predation lowing week and then dropped to very low rate during incubation in the prairie (F = 2.13; levels by August (Fig. 1). Nesting begana week df = 24, 136; P < 0.01) and during egg de- later and ended two weeksearlier in the prairie position in the old-field (F = 3.36; df = 16,136; (Fig. 2). The cumulative frequency of preda- P -c O.Ol), but similar to all other predation

TABLE 1. Daily mortality rates f 95% C.L. due to predation and all other causes.*

Egg-laying Incubation Brooding

Old-field Predation 0.039 & 0.0064 0.064 -c 0.010’ 0.042 f 0 0121.5 All other 0.084 i 0.0284 0.014 f 0.0056 0.006 f O:OOS6 Prairie Predation 0.062 f 0.054’” 0.038 i 0.015* 0.043 i o.0171J~5 All other 0.123 + 0.0731+ 0.016 i 0.010’.6 0.002 + 0.004’”

* Data are based on 3,624 days of exposure for 40 I old-field nests and 1,244 days of exposure for I34 prairie nests. Values with the same superscript are not sigmficantly different (P > 0.05) from each other. 70 JOHN L. ZIMMERMAN

0.06 E 2 0.05 E a n nn

/i 0.01 O1

t /

WEEKS - PRAIRIE

FIGURE 2. Daily predation ratesin old-fields (solid circlesand dashedcurve) and prairies (solid triangles)and percent frequency distribution of new nest starts in old-fields (open circles) and prairies (open triangles)according to week of nesting season.

rates.Higher lossesduring the incubation phase DENSITY AND PREDATION have been recorded for the Dickcissel (Harme- son 1974) as well as for various other species The density-dependenthypothesis predicts that (Best and Stauffer 1980). Daily predation rates predation rates will be greater in denser than of prairie nests,however, did not differ among in sparser prey populations. To test this ex- the three phasesin the nesting cycle. pectation, I used data from three old-field and Using the separatedaily predation rates for two prairie populations in which the number nestsin the egg-laying,incubation and brood- of nests found was thought to be close to the ing stages(Table 1) and the mean number of actual number present. For the old-field pop- days for each of these stages(3, 12 and 9, re- ulations the mean daily nest density per week spectively), I calculated the overall probability was related to the daily predation rate as well for a nest being taken by a predator during the as to the percentageof available nests lost to entire 24-day period after nestcompletion. The predators that week. I analyzed the prairie result for these independent, daily probabili- populations in a similar manner. I found no ties was 104 times greater for old-field nests significant relationships between nest density than for prairie nests. Owing to the signifi- and the percent of nests depredated in either cantly greater predation rate on old-field nests prairies (Spearman R = 0.010, y1= 20, P > during the long incubation period, predation 0.05) or old-fields (Spearman R = 0.125, n = affected nesting successmore severely in old- 34, P > 0.05). Neither could I demonstrate a field than in prairie populations. significant relationship between nest density NEST PREDATION IN DICKCISSELS 7 1 and daily predation rates in prairies (Spearman 197 1). An alternate hypothesis that explains R = 0.016, n = 20, P > 0.05) or in the old- these results is simply that predators are also fields(Spearman R = 0.269, y1= 34, P > 0.05). more abundant in old-fields. The predation These results match those from my earlier rate on Dickcissel nestsdepends upon the den- analysisof a smaller sample of Dickcissel nests sity of the predators, not the nests. (Zimmerman 197 1) as well as those of Best I agree with Best (1978) and Gottfried and (1978) and Gottfried (1978). Thompson (1978) that snakes are the most probable predator in old-field habitats and per- DISCUSSION haps in the prairie as well. Indeed, the only In the old-fields studied, the density of nests observed predation on a Dickcissel nest was increased32-fold from May to July, and a sig- by a snake (Facemire and Fretwell 1980). I nificant relationship between predation rate suspect that snakes are not primarily visual and time of nesting was demonstrated. In the predators, but rather probably depend addi- prairie, however, densities increasedonly two tionally upon olfaction. Furthermore, I doubt to six times during this period and predation that hunting by snakes is facilitated by ac- rates followed no regular pattern. Although the quiring a search image, a behavioral devel- prediction that predation was greater in old- opment that depends upon the density of the fields than in prairies is correct and nest den- prey (Tinbergen 1960). Snakes find nests by sities and predation rates in old-fields in- random close encounters(Best 1978), and the creased concurrently, I found no support for finding of one Dickcissel nest does not appear the hypothesis that there is a density-depen- to increasethe chanceof finding another. Snake dent response.No significant associationsbe- populations can be expected to be higher in tween nest densities and the percent of nests habitats where prey are more plentiful; hence, depredated or daily predation rates in either I suggestthat the increase in predation during habitat were shown. By analyzing data from the nesting season simply reflects a seasonal populations in which actual prey (nest) density addition of young snakesin the resident pop- was closely approximated, using a time period ulation. (one week) that was sensitiveto changesin prey population densities, estimating predation by ACKNOWLEDGMENTS measuring two variables, and having a suffi- This work was funded by grant GB-6087 from the Na- ciently-sized sample, I am convinced that pre- tional Science Foundation and NSF Undergraduate Re- dation of these Dickcissel nests was not den- search Participation programs during several successive sity-dependent. summers,as well as supportedby the Chapman Memorial Fretwell (1977), on the other hand, con- Fund of the American Museum of Natural History. Per- mission to conduct this work on the preferred study sites tended that prey densitiesand predation rates was generouslygranted by the Commander, Headquarters were significantly related in Dickcissels. He Fort Riley, and the Director of the Konza Prairie Research plotted the density of females in different hab- Natural Area. Such an extensive sample of nest records itats against the percent of nests lost to pred- could never have been accumulatedwithout the diligent efforts of Gil Blankespoor,Jeff Fergen, Elmer Finck, Scott ators in each of these habitats using some data Hatch. Harvard Townsend,and SteveWienert. Elmer Finck of his own, some of mine (Zimmerman 1966, also piovided the quadratic regressionanalysis. Stephen 197 1) and some of Harmeson’s (1974). Pre- Fretwell and Louis Bestmade criticalcomments that helped dation rates increasewith an increasein female in the preparation of this manuscript. density, but I suggestthat the correlation is spurious.The variation in predation rate is not LITERATURE CITED dependentupon variation in female (nest) den- BEST,L. B. 1978. Field Sparrow reproductive success sity in a way suggestinga negative feedback and nesting ecology.Auk 95~9-22. effect. My analysisshows that predation is two BEST,L. B., AND D. F. STAUFFER.1980. Factors affecting nestingsuccess in riparian bird communities. Condor orders of magnitude greater in old-fields, a 82:149-158. habitat in which female density and nest den- CODY, M. L. 1971. Ecological aspectsof reproduction, sity are also greater, compared to predation in p. 46 1-S 12. In D. S. Farner and J. R. King [eds.], prairies, where female and nest densities are Avian biology. Vol. 1. Academic Press,New York. lower. These data would fit Fretwell’s plot ELLIOTT,P. F. 1978. Cowbird parasitism in the Kansas tallgrassprairie. Auk 95: 161-167. nicely, but I found that predation rates in Dick- FACEMIRE,C. F., AND S. D. FRETWELL.1980. Nest pre- cisselpopulations are not density-dependent. dation by the speckledking snake. Wilson Bull. 92: The data that Fretwell used and those re- 249-250. ported here both show that predation is greater FRETWELL,S. D. 1972. Populations in a seasonalenvi- ronment. Monogr. Popul. Biol. 5. in habitats where female density is also higher, FRETWELL,S. D. 1977. Is the Dickcissel a threatened namely, in old-field habitats where there is species?Am. Birds 3 1~923-932. more structural heterogeneity (Zimmerman FRETWELL,S. D., AND F. S. SHIPLEY. 1981. Statistical 72 JOHN L. ZIMMERMAN

significance and density-dependent nest predation. ing biology of birds. National Academy of Sciences, Wilson Bull. 93541-542. Washington, DC. GORANSSON, G., J. KARLSSON, S. G. NILSSON, AND S. TINBERGEN,L. 1960. The natural control of insects in ULFSTRAND. 1975. Predation in birds’ nests in re- pinewoods. I. Factorsinfluencing the intensity of pre- lation to antipredator aggressionand nest density: an dation by songbirds. Arch. Neerl. Zool. 13:265-343. experimental study. Oikos 26: 117-l 20. VONHAARTMAN, L. 1971. Population dynamics, p. 39 l- GOTTFRIED,B. 1978. An experimental analysis of the 459. In D. S. Famer [ed.], Breedingbiology of birds. interrelationship between nest density and predation National Academy of Sciences,Washington, DC. in oldfield habitats. Wilson Bull. 90:643-646. ZIMMERMAN,J. L. 1966. Polygyny in the Dickcissel.Auk GOTTFRIED, B., AND C. F. THOMPSON. 1978. Experi- 83:534-546. mental analvsis of nest aredation in oldfield habitat. ZIMMERMAN.J. L. 1971. The territory and its densitv Auk 95:304-312. - dependent effect in Spiza americana. Auk 88:591- HARMESON,J. P. 1974. Breeding ecology of the Dick- 612. cissel.Auk 91:348-359. ZIMMERMAN,J. L. 1982. Nesting successof Dickcissels JOHNSON,D. H. 1979. Estimating nest success:the May- (Spiza americana) in preferredand lesspreferred hab- field method and an alternative. Auk 96:65l-66 1. itats. Auk 99:292-298. KREBS, J. R. 1971. Territory and breedingdensity in the ZIMMERMAN,J. L. 1983. Cowbird parasitism of Dick- Great Tit, Parus major. Ecology 52:2-22. cisselsin different habitats and at different nest den- MAYFIELD,H. 1961. Nesting successcalculated from ex- sities. Wilson Bull. 95:7-22. posure. Wilson Bull. 73~255-261. MAYFIELD,H. 1975. Suggestionsfor calculatingnest suc- Division of Biology, Kansas State University, Manhattan, cess.Wilson Bull. 87~456-466. Kansas 66506. Received 15 January 1983. Final accep- RICKLEFS,R. E. 1973. Fecundity, mortality and avian tance 8 July 1983. demography,p. 366-435. In D. S. Famer [ed.], Breed-

The Condor 86:72 0 The CooperOrnithological Society 1984

RECENT PUBLICATIONS

The Foraging Behavior of Mountain Bluebirds/with em- Population Ecology of the Dipper (Cinch mexicanus) in phasis on sexual foraging differences.-Harry W. Power. the Front Range of Colorado.-Frank E. Price and Carl E. 1980. Ornithological Monographs No. 28, American Or- Bock. 1983. Studiesin Avian Biology No. 7, Cooper Or- nithologists’ Union, Washington DC. 72 p. Paper cover. nithologicalSociety [Los Angeles,CA]. 84 p. Paper cover. $8.50 ($7.50 to AOU members). Source: Assistant to $9.00. Source:Allen Press. P.O. Box 368. Lawrence. KS the Treasurer of the AOU, Dr.’ Glen E. Woolfenden, 66044. Owing to their relatively simple ecology, dippers Dept. of Biology, University of South Florida, Tampa, FL appearto be ideal subjectsfor studiesofanimal population 33620. This monographexamines the possibleevolution- dynamics. Such a study is reported here, focussingon a ary and ecologicalcauses of sexualdifferences in foraging particular population, its resources,social interactions, re- behavior of insectivorouspasserines. Mountain Bluebirds productive success,and the potential effectson it ofcertain (Sialia currucoides) were chosenfor the study not because abiotic factors. Three years of field work produced an there is anything unique about their foraging,but because abundanceof data which, after statisticalanalysis, yielded the author was already familiar with their habits and the seventeen major factors impinging on the population at speciesmet his criteria for the investigation. Arising from different seasons.All of this is solid new information. The observationsduring two breeding seasons,several exper- major conclusion that emerges, however, is a reinforce- iments were run in order to test alternative hypotheses. ment of the view from other recent studies that popula- Although sampleswere necessarilysmall, the rigorousde- tions are not regulatedby only one or a few environmental sign and careful analysis of the tests yielded some clear factors.“Instead, there are many interactingvariables that insightsabout the roles of vegetation, foragingtactics, and operate with differing intensities to influence the major work load. The report deservesattention not only for its population processes . . .” While not novel, this message findings but also for its demonstration of the value of is perhapsmore widely important than the study’s findings experiments in field studies. Illustrations, references. and methodology. Illustrations, references.