EASTERN MEADOWLARKS NESTING in Rangelands and CONSERVATION RESERVE PROGRAM FIELDS in KANSAS

J. Field Ornithol., 67(2):222-235

EASTERN MEADOWLARKS NESTING IN RANGElANDS AND CONSERVATION RESERVE PROGRAM FIELDS IN KANSAS

DIANE A. Gv•c•o• 1 Departmentof Range and Wildlife TexasTech University Lubbock, Texas 79409 USA

KEVIN E. CHURCH Wildlife ResearchSection Nebraska Game and Parks Division Lincoln, Nebraska USA

LOREN M. SMITH Departmentof Range and Wildlife TexasTech University Lubbock, Texas 79409 USA

Abstract.---EasternMeadowlark (Sturndla magna)nesting habitat was studied to make man- agementrecommendations for fields enrolled in a federal land retirement program.We comparedavailable microhabitat, nest-site selection, and nest successon rangelandsand ConservationReserve Program (CRP) fields in easternKansas. Daily nest survivalrates and numbersfledged per femaledid not differ significantlybetween land-use types, but the power of thesetests was low. Predationwas the primary sourceof nest failure throughoutincuba- tion, hatching,and nestlingstages; abandonment, trampling, inviability, and unknowncauses alsowere importantduring incubation.Mowing CRP fieldswas a sourceof nest failure and alsoinduced adultsto abandonsome fields. CRP fieldshad a significantlyhigher percent, depth, and densityof litter cover;a taller herbaceouscanopy; less herbaceous cover; and more standingdead cover than rangelands.Differences in habitat structureindicate that CRP has increasedthe diversityof availablenesting habitats. Eastern Meadowlarks selected nestsites with significantlygreater litter cover,higher proportion of grass,more uncompacted litter, and more structuralhomogeneity than availableon randomplots. Delay of mowing and prescribedburning are recommendedto enhanceand maintainhabitat suitability for nestingEastern Meadowlarks in CRP fields.

STURNELIA MAGNA ANIDANDO EN EXTENSIONES DE TIERRA Y TERRENOS DEL PROGRAMA DE RESERVAS DE CONSERVACI6N EN KANSAS Sinopsis.--Seestudi6 el habitatde anidamientode Sturneliamagna para producirrecomen- dacionesde manejo para terrenosenlistados en un programafederal de retraer fierras.Se compararonel 6xito de los nidos, el microhabitatdisponible ¾ la selecci6nde lugar para anidaren las extensionesde tierra ¾en terrenosdel Programade Reservasde Conservaci6n (PRC) en el estede Kansas.Ni las tasasdiarias de supervivenciade nido ni el ndmero de volantonespot hembravariaron significativamente entre terrenosde usodiferente, pero la potenciade la pruebaestadistica fu6 baja. La depredaci6nfu6 la causaprincipal de la mor- talidad en nidos durante las etapasde incubaci6n,eclosi6n ¾ de crecimientoen el nido, pero el abandono,el pisoteo,la infertilidad, ¾ otras causasdesconocidas tambi6n rueton importantesdurante la incubaci6n.Desyerbar los campos del PRC caus6mortandad de nidos

• CurrentAddress: Department of Wildlifeand Fisheries, South Dakota State University, Brookings, South Dakota 57007 USA.

222 Vol.67, No. 2 EasternMeadowlark Nesting [223 y tambifin indujo al abandono de los predios. Los terrenos del PRC tuvieron mayorespor- centaje,profundidad y densidadde cubiertade materiales,un mayordosel herbficeo, menos cubiertaherbacea y m/rscubierta erecta muerta que las extensionesde tierra. Las diferencias en la estructurade hfibitat indican que el PRC ha aumentado la diversidadde habitatsde anidamiento. Los individuosde Sturnellamagma seleccionaron lugaree de anidaje con una cubierta de materialessignificativamente mayor, mayor proporci6n de hierbas, materiales poco compactosy mayor homogeneidadestructural que laos disponiblesen parcelasfortui- tas. Se recomiendaretrasar el corte de cespedy de quema prescritapara estimulary man- tener la adecuidadde habitat para el anidaje de Sturnellamagma en los prediosdel PRC. The ConservationReserve Program (CRP) of the 1985 Food Security Act converted >14.3 million ha of highly erodible cropland in the United Statesto a variety of permanent cover typesfor 10 yr. In Kansas,>1.1 million ha were enrolled in CRP with >90% planted in native grasses (U.S. Department of Agriculture 1993). The CRP wasintended to accom- modatewildlife and agriculturalconcerns and mayprovide quality habitat for grasslandbird populations. Quality of CRP habitat has been evaluatedby comparingpasserine den- sitiesin CRP versuscropland (Johnsonand Schwartz1993) and by apply- ing Habitat SuitabilityIndex (HSI) models (Hays et al. 1989). Although thesestudies use practicalmethodologies for regional scales,the assump- tion that densityor HSI alone reflect habitat quality may not be valid (Van Horne 1983, Vickery et al. 1992). Severalstudies have used produc- tivity as an indicator of CRP habitat quality for game species(e.g., Ber- thelsen et al. 1990, Kantrud 1993). Given high use of CRP by passerines (johnson and Schwartz1993), reproductive successof non-game species also needs to be investigated. We assessedthe quality of habitat in CRP fields by comparing the reproductive ecology of Eastern Meadowlarks (Sturnella maffna) in CRP fields and grazed, native grassrangelands. These habitats had similar plant speciescomposition, but hayingand grazingwere not permitted in CRP. Our objectiveswere to: (1) compare Eastern Meadowlark nest successin CRP and rangeland habitats,(2) identify causesof nest failure, (3) compare nest-sitemicrohabitat selection between CRP fields and rangelands, and (4) examine the importance of microhabitat to nest success.

METHODS Our study was conducted 10 km north of Emporia in Lyon County, Kansas(38ø30'N, 96ø20'W). Topography was gently rolling with elevations ranging from 323-466 m. Winter wheat, row crops,rangeland, and hay- fields Were primary land uses.Climate was continental averaging0 C in winter and 25 C in summer,with a 6-mo growingseason. Annual precip- itation rangedfrom 63-114 cm with 73% fallingfrom April throughSep- tember (U.S. Departmentof Agriculture,Soil ConservationService 1981). To minimize variabilityassociated with different CRP practices,we examined conversionto native grassbecause it promoted vegetationwith which Eastern Meadowlarks evolved and was the most common practice in Kansas(U.S. Department of Agriculture 1993). We used three CRP fields in 1990 and added one field in 1991. Field size ranged from 18-24 224] D. A. Granforset al. J.Field Ornithol. Spring 1996 ha. Mowing, without removal of cut material, was conductedin mid-June 1990 and in previousyears to control noxious weeds. One field was mowed only in selectedstrips, and mowingwas completed in late August. Only one of the four CRP fields was mowed in mid-June of 1991. We used five rangeland sitesin 1990. Stockingrates were -<1.63 animal unit months (AUM)/ha (cow/calf) and field size was 14-63 ha. Range siteshad a history of spring burning, but were not burned in 1990. We used four range sites in 1991, but only one 1990 site was reused. This field was not burned in 1991 and was stocked with cows and calves at 0.54 AUM/ha. Three replacementrange sites(178-259 ha) were burned in April 1991 and stockedwith steersat 1.85-3.09 AUM/ha. CRP fields were planted in 1988 or 1989 with a mix of big bluestem (Andropogongerardi•), little bluestem ( Schizachyriumscoparium), sideoats grama (Bouteloua curtipendula) , Indiangrass ( Sorghastrumnutans) , and switchgrass(Panicum virgatum). Cool-seasongrasses (Bromus spp.) also were common. The most common forbs were western ragweed (Ambrosia psilostachya)and mare'stail (Conyzacanadensis) in 1990 and lettuce (Lac- tucaspp.) and sweetclover(Melilotus spp.) in 1991. The abovenative grass species,tall dropseed(Sporobolus asper), and Kentuckybluegrass (Poa pra- tensis)were also characteristicof rangelands.Clumping of native grasses, which occurred in CRP fields, was prevented in rangelandsby grazing and burning. The most common forbs in rangeland were western ragweed and Baldwin's ironweed (Vernonia baldwin•). Nest success.--Wecaptured female Eastern Meadowlarkswith mist nets and by night-lighting(Drewien et al. 1967) in April-June 1990 and 1991. We placed mist nets in areasof meadowlarkactivity and near nests.We moved nets within each field to avoid concentratingthe sample in one area. We night-lighted by walking through fields and dragging a rope to flush meadowlarks.We determined sex on the basisof plumage, bill color, and body size (Kemmerer 1981). We fitted femaleswith 2.5-g transmitters(-<3% body mass), each at- tached to the back with an elasticwing-loop harness.We obtained daily locationsby circling with a hand-held antenna systemto within 30-40 m of the radio-taggedbird. We found nestsby flushing femalesthat had the same location for >--2d. When we could not estimate incubation stage from the female's location record, we floated eggsto estimatehatch date (Westerskov1950). We also used any nests of unmarked birds that we found. When a radio-taggedfemale was off the nest, we checked the site to confirm nest status.We examined nests of unmarked birds about every third day.To prevent human-inducedpremature fledging, we did not visit nestsbetween the seventhand 12th day posthatch (Brown 1988). On the 13th day posthatch,we determinedif the attemptwas successful (>-1 nestling fledged; Nice 1957) by finding feather sheathsat the bottom of the nest (Roseberryand Klimstra1970). Successfulnests were alsocharacter- ized by an enlarged nest bowl and the presenceof fecal sacs.We consid- ered nests unsuccessful if feather sheaths were few or absent, or there Vol.67, No. 2 EasternMeadowlark Nesting [225 were signsof nest destruction.We placed nest lossesinto five categories: (1) depredated, (2) abandoned, (3) trampled, (4) inviable, and (5) unknown.Nests of radio-taggedbirds composed70% of the sample,so for these nests the date of nest failure was known. We estimated daily nest survivaland failure rates using the Mayfield (1961) method as adapted by Heiseyand Fuller (1985). Becauseof small sample size, we estimated daily survival rates for 1990 over the entire nestingperiod. For 1991, we estimateddaily survivalfor incubation (1- 12 d), hatching (13-14 d) and nestling (15-24 d) periods. We treated complete clutchesof inviable eggs as survivinguntil the twelfth day of incubation and thereafter deleted them from analyses.We delimited seasons (early,middle, and late) by grouping first, second,and subsequent nestingattempts. We used a z-testfor comparisonof daily survivaland failure rates (Johnson1979) and adjustedoverall nest successrates for non-normal bias (Heisey and Fuller 1985). We tested power for non-significant (P > 0.05) results as in White and Garrott (1990:28-35) using an effect size equal to observeddifferences. Microhabitat.roWe measured microhabitat characteristics at nests when nestlingsfledged or 24 d after initiation of incubation if the attempt was unsuccessful.In 1990, we measuredrandom plots within the nest field (1/ha) during 1-wk samplingperiods in June and July. On the basisof subsamplevariation in the 1990 data,we sampled20 randomplots/field/ mo in 1991 and measuredrandom plots on the same date as nests (e.g., if four nestsin one field were due to fledgewithin a month, we measured five random plots per nest). To determine potential nest-siteselection over a larger area, we sampledvegetation at four plots 5 m from nestsin the cardinal directions. We visually estimated herbaceous canopy cover, proportion of grass, standingdead cover, and litter coverto the nearest5% within a 75 x 75- cm quadrat (Hays et al. 1981). We measuredherbaceous canopy height to the nearestcm and litter depth in four cardinaldirections 20 cm from the quadrat center. In 1991, we ranked litter density (1-5, 5 = densest) on the basisof gap size among litter materialsand the difficulty of in- sertinga metal ruler to the soil surface.We also ranked litter type (1 = predominantly fine materials [<3 mm diameter], 5 = predominantly coarse materials [>10 mm diameter]) in 1991. We recorded 24 visual obstructionreadings at each plot (Robel et al. 1970). We estimatedmi- crohabitatheterogeneity using the covarianceof the 24 readingsfor each plot (Wiens 1974). We measureddistance from neststo an edge (fence- line, cover type change,road, or pond). Some microhabitatvariables were not normally distributed or violated assumptionsof homogeneousvariances. We still used ANOVA for uni- variate analysesbecause the F-testis robust to theseviolations (Cochran 1947, Pearson1931) and resultscould be interpreted as a comparisonof actual means.Calculated probability levels are not exact however(Coch- ran 1947). We assessedpotential violations of sphericityfor blocked (subplot) databy comparingcomputed F-values to an F-distributionwith 1 df. 226] D. A. Granforset at. J.Field Ornithol. Spring 1996

In 1990, radio-taggedfemales nested in different rangeland siteseach month. Therefore, we used a factorial split-plotANOVA with land use X month (June and July) as whole-plot factors and location (random plots, nest sites,and plots surrounding the nest) as the subplot factor. For 1991 data, we analyzed microhabitat characteristicsin a split-plot, repeated measuresANOVA with land use as the whole-plot factor, fields as experimental units, location as the subplot factor, and seasonas the repeated measure. We used a t-test to separate means when a single factor F-test was significant (P < 0.05). If an interaction was present (P < 0.1), we used a t-test to compare means within levels of interacting factors. We followed Milliken and Johnson's (1984:377-407) recommendationsfor split-plotdesigns with missingdata. We compared microhabitats of successfuland unsuccessfulnests in 1991 using a split-plot design with treatments as the whole plot factor, fields as the experimental unit, and successas the subplot factor. We included only seasonX field combinationsfor which both successfuland unsuccessfulnests were found and pooled characteristicsover time periods.We used a similar analysisfor distanceto edge, but included all nests. We excluded trampled nestsand inviable clutchesfrom these analyses. We used DESIGN (Dallal 1988) to test power for consistenttrends that yielded nonsignificant (P > 0.05) results.

RESULTS Nestsuccess and mortalityfactors.--Dailynest survivalrates tended to be lower in CRP fields in 1990 (CRP: 0.93 ___0.022 [SE], n = 10 nests; rangeland:0.95 +_0.022, n -- 11 nests)and for all nestingperiods in 1991 (Fig. 1), but the trend wasnot significant,perhaps as a result of low power (1990: z = 0.48, P = 0.63, power = 0.07; 1991: z = 0.40-0.78, P = 0.44- 0.69, power = 0.06-0.13). Overall nest successrates in 1990 were 16.5% in CRP and 24.6% in rangelands;1991 successrates were 9.6% in CRP and 19.8% in rangelands.Daily nest successof radio-taggedbirds (0.93 + 0.011, n = 49 nests) was not different (z = 0.72, P = 0.47) from unmarked birds (0.91 +_ 0.022, n -- 22 nests). Daily nest survivalrates were higher (z = 2.49, P = 0.013) for late versusearly nestsin rangelands in 1991. We were unable to detect a difference (F = 0.95, P -- 0.37) in nestlings fledged per radio-tagged female between CRP fields (1.92 -+ 0.30, n -- 12 females)and rangelandsites (0.71 + 0.28, n = 11 females), again perhaps as a result of low power (0.08). There were more (z = 2.26, P = 0.024) nest failures due to inviability in CRP than rangelands,but no difference (P > 0.1, power = 0.05-0.14) among other mortality factorsbetween land usesin 1991 (Fig. 2). Daily predation rateswere higher (CRP: z = 2.11, P = 0.035; rangeland:z = 2.10, P = 0.036) during hatching(overall CRP: 37.3%; rangeland: 25.5%) than during incubation(CRP: 24.4%; rangeland:15.5%). In 1990, mowing causedthe lossof three nestsin two CRP fields. A female remained to re-nestin the field that wasspot mowed,but completemowing of the other field causedabandonment by three radio-taggedfemales. Similarly, Vol.67, No. 2 EasternMeadowlark Nesting [227

0.95

0.90 0.85 ' 0.80

0.75 CRP Rangeland--'--- 0.70

0.65 I I I I I I Eggs Hatch Brood Early Mid Late • Nest Stage• Season FIGURE1. Daily nest survivalrates (i ___SE) of Eastern Meadowlarksin Conservation Re- serve Program fields and rangelandsin 1991. Eggs = 1-12 d, Hatch = 13-14 d, Brood = 15-24 d, Early = incubationinitiated <21 May,Mid = initiated21 May-14Jun.,Late = initiated >14 Jun. Means with the same letter are not different (P > 0.1, z-test) betweenland usesor among nest stagesand seasons. a female with four-day-oldfledglings abandoned a CRP field when it was mowed completelyon 12 Jun. 1991. Microhabitat.--Ninenest sitesand 204 random plotswere measuredin five rangelands,and sevennest sitesand 190 random plots in three CRP fields in 1990. Herbaceouscover was greater in rangelandsthan in CRP (F1,5.77= 12.3, P = 0.014) and decreasedfrom June to july (F•,5.87= 64.1, P < 0.001) in both land use types.Available standing dead cover was greater in CRP than in rangelands(t = 3.14, df = 13.48, P = 0.007) and waslower at nest sitesthan random plots in CRP (t = 4.48, df = 11, P < 0.001). In 1991, 44 nestsand 215 random plotswere measuredon four rangeland sites,and 30 nests and 160 random plots were measured on four CRP fields.At least one nest wasfound in each rangelandsite for each seasonalperiod, but in CRP,one field wasrepresented by one early-season nest and another field had nestsfor the first two seasonsonly. 228] D. A. Granforset al. J. FieldOrnithol. Spring 1996

0.30

0.25 CRP Rangeland 0.20

0.15

0.10

0.05

0.00 Pr Ab Tr In Un Pr Pr Un Eggs Hatch Brood FIGURE2. Daily failure (mortality) rates (i - SE) for Eastern Meadowlarknests in Conser- vation ReserveProgram fields (CRP) and rangelandsin 1991. Eggs= 1-12 d, Hatch = 13-14 d, Brood = 15-24 d, Pr = predation,Ab = abandonment,Tr = trampling,In = inviable clutch, Un -- unknown.

T•I•F. 1. Microhabitatcharacterisitics of ConservationReserve Program fields (CRP) and rangelandsfor Eastern Meadowlarksthrough three seasonsin Lyon County, Kansas, 1991, for variableswithout land use x seasoninteractions (P > 0.1).

Herbaceous % litter Litter Litter CV visual Landuse height(cm) cover depth(cm) type(1-5) obstruction or season a • SE • SE • SE • SE • SE

Rangeland 32.4Ab 2.43 -- -- 2.4A 0.31 41.8A 2.97 CRP 42.1B* 2.52 • • 2.3A 0.33 47.6A 3.15 Early 36.2A 1.84 67A 6.1 2.5A* 0.25 2.8A 0.24 43.4A 2.34 Mid 39.2B* 1.90 59A 6.4 2.3A** 0.29 2.2B** 0.25 45.1A 2.45 Late 36.5AB 1.98 64A 6.9 3.6B 0.35 2.lB** 0.27 45.6A 2.60

Early -- <21 May;Mid -- 21 May-14June;Late = >14June. Means followed by the sameletter are not different (t-test), *P < 0.05, **P < 0.01. Vol.67, •o. 2 EasternMeadowlark Nesting [229

Microhabitat structure of CRP differed from rangelandsfor sevenof 10 variables.The only consistentdifference over locationsand through seasonswas taller herbaceouscanopy in CRP versusrangelands (Table 1). Land use X seasoninteractions indicated differencesin vegetationstruc- ture developmentbetween CRP and rangeland.In CRP,herbaceous cover and the proportion of grassdecreased through the seasonin contrastto the increasein rangeland (Table 2). Randomplots in CRP had more litter and standingdead cover than in rangelands(Table 3). Microhabitat at nest sitesdid not differ betweenCRP and rangelands (P • 0.09), except for canopy height and litter density,which differed over all locations.In both land use types,vegetation structure was most homogeneousat nest plots and litter wasfiner at plots surrounding nests than at random plots (Table 4). Plots 5 m from nestswere more like nest sitesthan random plots for all variablesassociated with litter. In rangelands, meadowlarksselected nest siteswith greater litter cover and depth than was availableat random plots (Table 3). In CRP, selectionfor a greater proportion of grassand lessdense litter was evident. The only vegetation characteristicsthat may have differentiated suc- cessfulfrom unsuccessfulnests were higher proportion of grass (P -- 0.092, power = 0.22) and greater standingdead cover (P = 0.095, power = 0.28). Successfulnests were not farther from edge than unsuccessful nests (P-- 0.92, power = 0.14).

DISCUSSION We were unable to detect differences in Eastern Meadowlark repro- ductivesuccess between CRP and rangelands,despite differences in overall habitat structure.Unfortunately, low nest successresulted in few nest daysfor determinationof daily success,high variance,and low testpower for all factorsassociated with nest survivaland numbersfledged per nest. Small differencesin daily survivalrates result in large differencesin overall nest success,which may be biologicallysignificant. Our data suggest lower nest survivalrates but higher reproductiveoutput per female in CRP fields. Much larger sample sizes (100-200 nests) would be needed to determine (power = 0.8) if theseapparent differences really existand the degree to which they may be compensatory. Nestsuccess rates were lower (• 25%) than previouslyreported for East- ern Meadowlarks(e.g., 33% in Illinois [Lanyon 1957], 30% for monog- amousfemales and 52% for polygynousfemales in Ontario [Knapton 1988]). In the precedingstudies, it wasassumed that all nestswere found, and nestsuccess calculated directly may haveover-estimated success (May- field 1961). Radio-taggingbirds did not appear to affect nest successin our study,but could have affected the propensityto re-nest (Rotella et al. 1993), thereby lowering the number of late-seasonnests. Because the life expectancyof Eastern Meadowlarksis not known, the impact of low nest successon the population cannot be evaluated.Eastern Meadowlarks were persistentre-nesters in our study,which would lessenthe impact of low nest successon the population. 230] D. A. Granforset al. j. Field Ornithol. Spring 1996 Vol.67, •o. 2 EasternMeadowlark Nesting [231 232] D. A. Granforset al. J.Field Ornithol. Spring 1996

TABLE 4. Microhabitat characterisitics without land use X location interaction (P > 0.1) in ConservationReserve Program fields and rangelandsat random plots, Eastern Mead- owlark nest sites,and plots 5 m from nestsin Lyon County, Kansas,1991.

Visual % herbaceous Herbaceous Litter obstruction CV visual cover height (cm) type (1-5) a (din) obstruction Location 5• SE 5• SE 5• SE 5• SE 5• SE

Random 67A b 2.3 37.8A 2.01 2.7A 0.26 2.5A 0.28 47.0A 2.46 Nest 69A 2.3 37.8A 2.01 2.3AB 0.26 2.3A 0.28 39.4B* 2.46 5-m 65A 2.3 36.2A 2.01 2.0B* 0.26 2.3A 0.28 47.7A 2.46

1 = most materials <3 mm diam.; 5 = most matehals >10 mm diam. Meansfollowed by the sameletter are not different (t-test),*P < 0.01.

Roseberryand Klimstra(1970) attributedan 18% nestsuccess rate for EasternMeadowlarks in a re-seeded,ungrazed pasture in Illinois to high predatordensities. Predation rates of 53% (CRP) and 42% (rangeland) are high, but similar to thosefound for other ground-nestingspecies in grasslandecosystems (Martin 1993, Miller and Knight 1993). Most likely, a varietyof potential predators,(e.g., bullsnake [Pituophismelanoleucus], Common Grackle [ Quiscalusquiscula], Corvids, opossum [Didelphis vir- giniana], badger [ Taxideataxus], striped skunk [ Mephitismephitis], rac- coon [Procyonlotor], coyote [ Canis latrans]) lead to high predation rates (Miller and Knight 1993). Other causesof nest failure were as important as predationduring incubation,indicating that the sum of many factors was responsiblefor the low successrates we observed. We often found nestswith less than the full clutch taken, especiallyin rangelands,sug- gestingdisturbance by small predatorsand Brown-headedCowbirds (Mo- luthrusater), which may have increasedabandonment rates. Meadowlarks also have been shown to remove eggs from nests,which could indicate possibleinterference competition (Picman 1992). Brown (1988) and Roseberryand Klimstra(1970) did not find differences in vegetation between successfuland unsuccessfulEastern Mead- owlark nestsites, whereas we found an indication of more grassand standing dead coverat successfulnests. Lanyon (1957) hypothesizedthat an observedincrease in nest successthrough the seasonwas the resultof an increase in vegetativecover. The increase in nest successin rangeland correspondsto an increasein herbaceouscover, but we did not find more herbaceous cover at successful versus unsuccessful nests. Characteristicsimportant at nest sitescould be ascertainedwhen man- agementaffected habitat structure.Annual grazingand burning reduced litter in rangelandsresulting in nest-siteselection for more litter cover. Greater litter cover in CRP wasa positivefactor, but high litter densityas a result of the lack of removalof mowedmaterial may have lowerednest- site quality.Lack of disturbancein CRP (e.g., fire, grazing) wasalso likely to have perpetuated forbs, which provided little ground cover and re- suited in nest-siteselection for a higher proportion of grass.Risser et al. Vol.67, •o. • Easte• MeadowlarkNesting [233

(1981) found higher EasternMeadowlark densities on moderatelygrazed versus ungrazed grasslands,possibly reflecting their co-evolution with grazersand a need for periodic habitat perturbation.Grazing can have negativeimpacts, however, such as nest trampling and possibleattraction of Brown-headed Cowbirds (Friedmann 1929, Rothstein et al. 1987) lead- ing to increasedbrood parasitismrates (Granfors 1992). Differencesin habitat structurebetween CRP fields and rangelandsin- dicate an increasein the diversityof habitatsavailable for nestingEastern Meadowlarksand other ground-nestingbirds. An increasein breeding habitat diversitymitigates adverse effects of stressin other habitats (Rug- giero et al. 1988) and could provide a variety of micrositeswhich may vary in quality within and between seasonsdepending on the relative number and types of predators. The similarity of meadowlark nest-site structurebetween land usesreflects the meadowlark'sability to find suit- able nest sites in spite of overall differences in habitat structure. CRP provided nestinghabitat for meadowlarks,but qualitywas not consistent through the season.

MANAGEMENT IMPLICATIONS Whereasconstraints of forage productionand grazingmay limit rangeland managementoptions, CRP fields can be managedfor grassland-nesting birds. Rotenberry and Wiens (1980) found that Eastern Meadowlark densitieswere correlated with whole-field characteristicsequivalent to nest-sitecharacteristics selected in our study (e.g., grasscover, litter cover and depth, and structuralhomogeneity). Thus, microhabitatselection at nest-sitesappears to be similar to macrohabitat selection. Mowing and burning are often recommendedfor controlling noxious plantsin CRP.Mowing causesundesirable litter build up and, dependent on seasonalityand extent, may cause abandonment of fields and direct failure of nests.Frawley and Best (1991) found no reduction in Western Meadowlark (Sturnella neglecta)densities when part of a breeding terri- tory was mowed. Similarly,females did not abandon fields in our study when only portions of the field were mowed. Therefore, we recommend spot mowing to control noxious plants. Mowing should be conductedas late as possible(after 15 July) to avoid destructionof nests. Grazing is an alternative to mowing, and it would decreaselitter build- up (Risser et al. 1981), but increase the probability of trampling and attracting cowbirds (Friedmann 1929, Granfors 1992, Rothstein et al. 1987). When conductedat the proper time of year,prescribed burning can reduce litter and increasethe proportion and vigor of nativegrasses, while decreasingthe proportion of cool-seasongrasses and forbs (Wright and Bailey 1982). Periodic burns may increaseinvertebrate populations (Risseret al. 1981) and increasedvegetation cover can lead to increased nest successfor ground-nestinggrassland species (Johnson and Temple 1990). Annual burning could increasesoil erosion (Wright and Bailey 1982) and conflict with other conservationobjectives of CRP.Thus, we recommend burning in spring before significantgreen-up every 3-4 yr 234] D. A. Granforset al. J.Field Ornithol. Spring 1996 to approximate natural fire frequency (Risseret al. 1981, Wright and Bailey 1982), but specificprescriptions will depend on litter build-up, grassvigor, and local plant populations.This practicemay temporarily reduce early-seasonhabitat quality in the year of burning, but increased grasscover and somereduction of litter shouldenhance quality later in the seasonand in subsequentyears.

ACKNOWLEDGMENTS

We thank M. E. Eisenbarth, M. L. Gonzales and S. T. Meiman for assistance with field work, S. L. Brown for technical advice and D. B. Wester for statisticalguidance. Kansas Department of Wildlife and Parksprovided field and office equipment. S. L. Brown,J. L. Zimmerman, L. B. Best,P. D. Vickery and anonymousreviewers provided commentson the manuscript.Research was funded by TexasTech University(College of AgriculturalSciences and Natural ResourcesMS No. T-9-739), KansasDepartment of Wildlife and Parksand the Federal Aid in Wildlife Restoration Fund.

LITERATURE CITED

BERTHELSEN,P.S., L. M. SMITH,AND R. R. GEORGE.1990. Ring-neckedpheasant nesting ecologyand productionon CRP landsin the TexasSouthern High Plains.Trans. N. Am. Wildl. Nat. Resour. Conf: 55:46-56. BROWN,S. L. 1988. The maintenanceand consequencesof polygynyin easternmeadowlark (Sturnella magna). Ph.D. thesis,Purdue Univ., West Lafayette,Indiana. COCHP•N,W. G. 1947. Someconsequences when the assumptionsfor the analysisof variance are not satisfied. Biometrics 3:22-38. D•L•L, G. E. 1988. DESIGN: a supplementarymodule for SYSTATand SYSGRAPH.SYS- TAT, Inc. Evanston,Illinois. 23 pp. DP,•WIEN,R. C., H. M. R•EV•S,P. E SPRINGER,AND T. L. KucK. 1967. Back-packunit for capturingwaterfowl and uplandgame by night-lighting.J. Wildl. Manage.3:778-783. FR•WLE¾,B.j., •Nt) L. B. BEST.1991. Effectsof mowingon breeding bird abundanceand speciescomposition in alfalfa fields.Wildl. Soc. Bull. 19:135-142. FreEr)MANN,H. 1929. The cowbirds:a studyin the biologyof socialparasitism. Charles C. Thomas Publ., Springfield,Illinios. 421 pp. G}•2qFORS,D. A. 1992. The impact of the ConservationReserve Program on easternmeadowlark production and validationof the eastern meadowlarkhabitat suitabilityindex model. M.Sc. thesis, Texas Tech Univ., Lubbock, Texas. I-M¾s,R. L., C. SUMMERS,AND W. SEITZ.1981. Estimatingwildlife habitatvariables. U.S. Fish and Wildl. Serv.,Biol. Serv.Prog. FWS/OBS-81/47. Washington, D.C. --, R. P. WEBB,ANt) A. H. F•,MER. 1989. Effectsof the ConservationReserve Program on wildlife habitat: resultsof the 1988 monitoring. Trans. N. Am. Wildl. Nat. Resour. Conf. 54:365-376. HEISEY,D. M., ANDT. K. FULLER.1985. Evaluationof survivaland cause-specificmortality ratesusing telemetry data. J. Wildl. Manage.49:668-674. JOHNSON,D. H. 1979. Estimatingnest success: the Mayfieldmethod and an alternative.Auk 96:651-661. --, ANt)M.D. SCHW•U•TZ.1993. The ConservationReserve Program and grasslandbirds. Conserv. Biol. 7:934-937. JOHNSON,R. G., ANDS. A. TEMPLE.1990. Nest predation and brood parasitismof tallgrasss prairie birds.J. Wildl. Manage. 54:106-111. IZ•d'4TRUt),H. A. 1993. Duck nest successon ConservationReserve Program land in the Prairie Potholeregion. J. Soil Water Conserv.48:238-242. KEMMERER,M. R. 1981. Physiologicalcondition and sex determinationof easternmeadowlarks in north central Oklahoma. M.Sc. thesis, Oklahoma State Univ., Stillwater, Okla- homa. KN•PTON,R. W. 1988. Nesting successis higher for polygynouslymated femalesthan for rnonogamouslymated femalesin the EasternMeadowlark. Auk 105:325-329. Vol.67, No. 2 EasternMeadowlark Nesting [235

LANYON,W. E. 1957. The comparativebiology of the meadowlarks(Sturnella) in Wisconsin. Pub. Nuttall Ornithol. Club 1, Cambridge,Massachusetts. 67 pp. MARTIN,T. E. 1993. Nest predationamong vegetation layers and habitat types:revising the dogmas.Am. Nat. 141:897-913. MAYFIELr),H. 1961. Nestingsuccess calculated from exposure.Wilson Bull. 73:255-261. MILLER,C. K., ANDR. L. KNIGHT.1993. Doespredator assemblage affect reproductive success in songbirds?Condor 95:712-715. MILLIKEN,G. A., ANDD. E. JOHNSON.1984. Analysisof messydata. Van Nostrand Reinhold Co., New York,New York.473 pp. NICE,M. M. 1957. Nestingsuccess in altricialbirds. Auk 74:305-321. PEARSON,E. S. 1931. The analysisof variancein casesof non-normalvariation. Biometrika 23:114-133. PICMAN,J. 1992. Egg destructionby EasternMeadowlarks. Wilson Bull. 104:520-525. POSSER,P. G., E. C. BIm,:EY,H. D. BLOCKER,S. W. M•y, W.J. PARTON,ANDJ. A. WIENS.1981. The true prairie ecosystem.U.S./IBP SynthesisSer. 16. Hutchinson RossPubl. Co. Stroudsburg,Pennsylvania. 556 pp. ROBEL,R. J., j. N. Bmc,GS, A.D. DAYTON,ANr• L. C. HULBERT.1970. Relationshipsbetween visualobstruction measurements and weight of grasslandvegetation. J. Range Manage. 23:295-297. ROSEBEP,•V,J. L., ANr•W. D. KLIMST•. 1970. The nestingecology and reproductiveperfor- mance of the Eastern Meadowlark. Wilson Bull. 82:243-267. ROTELI•, J. j., D. W. HOW•RTER,T P. SANKOWSKI,•_Nr• J. H. DEVRIES.1993. Nesting effort by wild mallardswith 3 typesof radio transmitters.J. Wildl. Manage.57:690-695. ROTENBEP,•V,J.T., ANr•J.A. WIENS.1980. Habitat structure,patchiness, and avian commu- nities in North American steppevegetation: a multivariateanalysis. Ecology 61:1228- 1250. ROTHSTEIN,S. I., J. VERNER,E. STEVENS,AND L. V. RITTER.1987. Behavioraldifferences among sex and age classesof the Brown-headedCowbird and their relation to the ef- ficacyof a controlprogram. Wilson Bull. 99:322-337. RUGGIERO,L. F., R. S. HOLTHAUSEN,B.G. MARCOT,K. B. AUBRY,J. w. THOMAS,AND E. C. MESLOW.1988. Ecologicaldependency: the conceptand its implicationsfor research and management.Trans. N. Am. Wildl. Nat. Resour.Conf. 53:115-126. U.S. DEPARTMENTOF AGraCULTURE.1993. Agricultural statistics1992. U.S. Govt. Printing Off., Washington,D.C. 517 pp. U.S. DEPARTMENTOF AGRICULTURE, SOIL CONSERVATIONSERVICE. 1981. Soil surveyof Lyon County,Kansas. U.S. Govt. Printing Off., Washington,D.C. 1981-293-883/95.96 pp. VANHORSE, B. 1983. Densityas a misleadingindicator of habitatquality. J. Wildl. Manage. 47:893-901. VICKERY,P. D., M. L. HUNTER,JR., ANDJ. V. WELLS.1992. IS densityan indicatorof breeding success? Auk 109:706-710. WESTE•,SKOV,K. 1950. Methodsfor determiningthe ageof gamebird eggs.J. Wildl. Manage. 14:56-57. WHITE, g. C., ANDR. a. GARROTT.1990. Analysisof wildlife radio-trackingdata. Academic Press,Inc. San Diego, California. 383 pp. WIENS,J. A. 1974. Habitat heterogeneityand aviancommunity structure in North American grasslands.Am. Midi. Nat. 91:195-213. WRIGHT,H. A., ANDA. W. BAILEY.1982. Fire ecology,United Statesand southern Canada. John Wiley and Sons,New York,New York. 501 pp. Received29 Dec. 1994; accepted23 May 1995.