Reproductive Success, Growth and Survival of Black-Crowned Night-Heron (Nycticorax Nycticorax) and Snowy Egret (Egretta Thula) Chicks in Coastal Virginia

The Auk 113(1):119-130, 1996

REPRODUCTIVE SUCCESS, GROWTH AND SURVIVAL OF BLACK-CROWNED NIGHT-HERON (NYCTICORAX NYCTICORAX) AND SNOWY EGRET (EGRETTA THULA) CHICKS IN COASTAL VIRGINIA

R. MICHAEL ERWIN, JOHN G. HAIG, DANIEL B. $TOTTS,AND JEFF$. HATFIELD NationalBiological Service, Patuxent Environmental Science Center, 11410American Holly Drive, Laurel,Maryland 20708, USA

ABSTRACT.--Westudied reproductive success, growth, and survivalof Black-crownedNight- Heron (Nycticoraxnycticorax) and Snowy Egret (Egrettathula) chicksin two mixed-species heronries on marsh islands in ChincoteagueBay, AccomackCounty, Virginia in 1992 and 1993. We attachedradio transmitterswith mortality sensorsto the oldest chicks (A-chicks) in 11 to 22 nestsof both speciesto monitor survival during the mid- to late nestlingperiod and into the postnestingdispersal period. For both species,we found significantdifferences between 1992and 1993in growth ratesand survival. Massgrowth ratesof chickswere higher in 1993 than in 1992for both species.Culmen-length growth ratesvaried significantlydue to year-colonyeffects for night-herons,but only for hatching order for egrets.Differences in survival rates due to hatching order were found for the egrets in both years, but were found only in 1992for night-herons.As with massgrowth rates, survival of chickswas higher in 1993 than 1992.Survival of radio-markedA-chicks did not differ between speciesor years for the period from hatchingto fledging or from fledging through the end of the study (ca. two monthspostfledging). Survival ranged from 0.80 to 1.00from the time radio transmitters were attached(ca. two weeksof age) until dispersalage (53-55 daysfor egrets;55-60 days for night-herons).After birds left the colony,survival rateswere lower during the next 40 to 55 days,ranging from 0.25 to 0.60. Theseresults suggest that, at leastfor Snowy Egrets, A-chicks may be buffered from annual variations in food conditions,but that growth and survival of other brood membersmay provide a barometerof local conditions.Despite prob- lems of variability in measurement,some of theseparameters show promiseas bioindicators of estuarineconditions. We recommendthat a cost-efficientwading-bird monitoring program would include:(1) estimatingnumbers of nestingbirds of selectedspecies (e.g. Snowy Egrets) at particular estuaries,and (2) monitoring "initial brood sizes"and survival of young to at least two weeks of age in a sample of nests for each focal species.However, additional biomonitoring (e.g. marsh forage fish, contaminantloads) is necessaryto evaluatehow well top trophic-levelorganisms such as fish-eatingbirds respondto changesin estuarinepro- duction or quality. Received12 December1994, accepted 18 March 1995.

RELIABLEBIOLOGICAL INDICATORS are being indicators, we investigated the reproductive soughtby a number of federal and stateagen- success,growth and survival of young Black- ciesto monitor estuarineecosystem health and crowned Night-Herons (Nycticoraxnycticorax) integrity (Gray 1980,Diamond and Filion 1987, and Snowy Egrets(Egretta thula) at two salt- Fox and Weseloh 1987, Peakall and Shugart marsh coloniesin coastalVirginia in 1992 and 1993).Colonially nestingwading birds (Cicon- 1993.Nesting studieshave been done on these lifotrees) have been identified as one potential speciesin other areas(e.g. Frederickand Col- bioindicatorgroup by a number of researchers lopy 1989a,b, Custerand Peterson1991, Fred- (Custer and Osborn 1977, Kushlan 1993, Par- erick et al. 1992, Parsons1994), thus providing sons1994, Custer in press,Erwin and Custer in a basis for year and geographiccomparison. press),although others (Morrison 1986,Temple These other studies, however, were not de- and Wiens 1989) have expressedreservations signedto evaluatehow survivalrelates to growth about the usefulness of birds as environmental and success,since investigators did not follow indicators. birdsthrough the entire nestingperiod (but see As part of an evaluation of wading birds as Frederick et al. 1993), nor did they attempt to 119 120 ERWINET AL. [Auk,Vol. 113

cleaner and on body down with small amounts of brightly coloredfingernail polish for individual iden- tification.Aluminum leg bandsand radiotransmitters were attached(see section on late-nesting-periodsur- vival) when the birds' tarsi were of sufficient size (8 to 14 days)to retain an adult-sizedband. We assigned hatchingorders to chicksbased on observedhatching dates or, for unknown dates, we used relative size (length of the bird's culmen; Custer and Peterson 1991). A-chicks were the first hatched, B-chicks the secondhatched, and so on, through E-chicks.Chicks were followed only if they were seen alive at least ortce. Nestsuccess.--To sample nests in closeto a random manner, we selectedstudy nests from 6 to 10 areas of the colonieseach year. We then comparedthe av- erageclutch size and average"initial broodsize" (i.e. number hatched) for each speciesby year. We used Fisher's exact test (Lehmann 1975) with Monte-Carlo Fig. 1. Location of nesting coloniesin Chinco- P-valueestimate statistics (Cytel 1991)to test for dif- teagueBay, Accomack County, Virginia. ferencesbetween years. The traditionalMayfield method for estimatingnest determine survival rates beyond the fledging success(Mayfield 1975, Hensler and Nichols 1981) period. This is a major gap in the demography was not appropriate to use in our study. Instead, nest successwas calculatedin two ways. First, we com- of almostall migrantbird species,although some pared the number of birds per nest that survived 14 recent data exist for a few other waterbirds, days (i.e. the age up to which we were certain of including RoseateSpoonbills (Ajaia ajaja;Bjork finding all young on a visit); second,we compared and Powell 1994) and Great White Herons (Ar- the number of successful (i.e. •1 chick survived 14 deaherodias occidentalis; Powell and Bjork 1990) days)versus unsuccessful nests using Fisher'sexact in Florida. tests.Since brood size had a significanteffect on the Our objectivesfocused on the following ma- number of chickssurviving to 14 days(see Results), jor questions:(1) How closely coupled are re- we analyzedthe differencein the number of chicks productive success,growth rates, and survival surviving to 14 days for nestsin which at least one chick hatched for each brood size. "Initial brood size" of young for the two speciesof wading birds? was defined as the total number of chicks known to How sensitiveare they to annual changes?(2) have hatched in a nest. All chicks were included, Which speciesand nestingparameters might be regardlessof whether their hatching dates were usefuland costeffective in a biomonitoringpro- known precisely. When we found chicks whose gram? hatching(or pipping) wasnot observed,we estimated the age/hatching date by comparing culmen length STUDY AREA AND METHODS to Custerand Peterson's(1991) equation and by comparing the chick'sinitial size to thoseof its siblings. We conductedour studiesat two large mixed-spe- Samplesizes for most brood sizeswere unequalfor cies heronries (800 to 2,000 nests) on marsh islands the two years.Only for brood sizesof three for both near Chincoteague, Accomack County, Virginia speciesin both yearswere there sufficientsamples to (37ø56'N,75ø25'W) in 1992and 1993(Fig. 1). In 1992, make meaningful comparisons;however, we ana- we restrictedour studyto the northerncolony (Cause- lyzed all brood sizes, using Fisher'sexact test with way site),but in 1993expanded the studyto include Monte-Carlo P-value estimates(Cytel 1991). Thus, the southern (Willis site) colony. Since ardeid nests comparisonswith smallnumbers of nestsmay not be are mostsusceptible to disturbancein the nest-build- very powerful. If it is assumedthat a randomsample ing to egg-layingstage (Tremblay and Ellison 1979), of brood sizes in both years was obtained,we can we limited our visitsto one early in the period (mid- combine all brood sizes and evaluate whether the May) and then one or two per week from late incu- average number of chicks produced per nest is the bation (>23 May) onward. During late incubation, samein both yearsfor both species. we marked20 to 45 egretand night-heronnests with Growth rates.--On each visit, we measured chick plasticvinyl flaggingin marshelder (Iva frutescens) mass(to nearest0.1 g) usingan Ohauselectronic bal- shrubs,and checkednests once or twiceweekly. Chicks ancethat was calibrateddaily. We alsomeasured cul- were markedon the tarsuswith piecesof coloredpipe men length (to nearest0.1 ram) from the edge of the January1996] Growthand Survival of Night-Heronand Egret Chicks 121 feathersto the tip of the culnaen.Birds whose exact ciesseparate. Then, differencescaused by year-colony hatching dateswere unknown were excludedfrom weretested for eachhatching-order combination. First, the analyses.In addition, to insure a linear relation- the two 1993 colonieswere comparedseparately, and ship,we followed Custerand Peterson's(1991) meth- then the 1992 colony was added. odologyof only includingmeasurements from birds To estimate survival of a sample of birds during between ages5 and 18 days when calculatingage- the late-nestingperiod until dispersal,we attached massrelationships, and between ages3 and 18 days small (10-g) oval radios with mortality sensorsto for calculatingage-culnaen relationships. To avoidau- A-chicks between the ages of 7 and 21 days (œ= 2 tocorrelation,we calculatedindividual chick growth weeks). Radio transmitters were attached to blank alu- slopesfor the relationshipsbetween mass and age, minum leg bandsapplied above the tarsonaetatarsus and between culnaenlength and age. Any bird with bone. We attachedthe radio to the band using both two or more measurementsin the age period of in- epoxy and naonofilarnentfishing line. In only one terestwas included in the studyregardless of whether casewas the radio thought to have contributed to the the bird survived or died; hence, some negative death of the bird, so that bird was not included in growthswere recorded(i.e. masslosses). Means were the study. In 1992, 20 radio transmitters were used then generatedfor individual results.For eachspecies on Snow Egrets, 10 on night-heron chicks.In 1993, separately,we conductedtwo-way ANOVA and Tu- 20 were used on each species.When mortality oc- key tests(Sokal and Rohlf 1981)to assessyear-colony curred, we were able to reuse transmitters from sev- and hatching-ordereffects, with interaction terms. eral chicks. We visited the colonies one to three times When we comparedgrowth ratesfor both parameters per week during late nestingto monitorbird activity for both speciesbetween coloniesin 1993, no signif- with portablereceivers and hand-heldYagi antennas. icant model effects were obtained (in ANOVA, all P When a mortality signalwas heard (50% higher pulse > 0.05);therefore, any differencesfound were judged rate), we usuallylocated the bird within a day. A bird to be year effects. wasconsidered to havefledged on the midpointdate SurvivaL--Becauseyoung wading birds become betweenthe last day recordedin the colonyand first mobile and harder to capture after about two weeks day it was missed. The age of death was similarly of age,we conductedsurvival analyses in three parts: calculated; if the bird had not died, it became cen- (1) duringthe periodwhen all chickswere very young soredat the age at which it fledged. Daily survival (<20 days old) and easy to find during nest checks; rates for the two speciesin the two years were cal- (2) from hatchingthrough fledging for radio-marked culatedas Kaplan-Meier statistics(Lawless 1982) and, A-chicksonly; and (3) for radio-markedyoung after again, the log rank statisticwas used to test for dif- they dispersed(i.e. postfledging) from colony.For the ferencesamong strata. first two parts,only birds with hatching datesknown We were able to make accuratecalculations of age to within one day were included. of dispersalfor most of the radio-marked birds in our For the early-nestingperiod, chicks that were found study. In addition, we used ANOVA to determine dead, were missingfrom their nest at a young age whether, eachyear, the two specieshad the sameage (<5 days old), or were missing after having been of dispersal,and if birdsof the samespecies dispersed noted to be weak and losing masswere assumedto at the sameage in both years. have died at the midpoint between the two visits, We analyzedpostdispersal survival using capture- unless a more accurate determination of date of death recapturemodels (Pollock et al. 1990). After almost was made in the field. Birds that appeared healthy all the birds had fledged,we followed them using but disappearedat the next visit became"censored" motorvehicles and fixed-wingaircraft at intervalsof (see the Lifetest Procedure in SAS for a discussionof 3 to 12 days from 12 August through 24 September censored data; SAS Institute 1987) for the survival in 1992and from 27 July through 2 Octoberin 1993. analysison the day they were last known to be alive. We flew a regular searchpattern on every searchthat The averageage of censoringfor birds that were not covered most of the coastalareas from: Cape May, radioedwas approximately 16 days.The Lifetestpro- New Jerseyto Salem,New Jersey;the Atlantic coast gram calculatesKaplan-Meier probabilitiesof surviv- from the Chesapeakeand Delaware Canal, Delaware al throughtime, incorporatingthe censoreddata, and to Cape Charles, Virginia; and the eastern shore of alsotests for honaogeneityamong strata.In this case, the ChesapeakeBay from CapeCharles to Kent Island, stratarefer to either year and colony("year-colony"), Maryland. Additional searcheswere made several or hatchingorder. We usedthe log rank statistic(Law- timeswhile crossingthe Delnaarvapeninsula, by fly- less 1982) to test for differencesamong strata.First, ing up the Delaware River 15 kna north of Philadel- the number of chicks in a brood was used as a co- phia and along the mid- and southernNew Jersey variable to evaluate the effect on survival of chicks shorenorth to Manasquan.One searchwas made along of eachspecies-hatching order combination.Since our the western shore of the ChesapeakeBay in August P-values exceeded 0.05, we combined all brood sizes. 1993.The radioshad a range of 5 to 18 knafrom the We testedfor differencesin survival causedby hatch- aircraftand about1 to 2 knaon the ground.When we ing orderfor eachyear-colony stratum, keeping spe- heard a mortality signal, we attemptedto locate the 122 ERWINher AL. [Auk,Vol. 113

TABLE1. Clutch-sizea and brood-sizeb analysesbe- TABLE2. Summaryof differencesbetween yearsin tween years (1992 and 1993) for Black-crowned "final brood sizes"a for Black-crownedNight-Her- Night-Heron and Snowy Egret chicks. on and Snowy Egret nests• in 1992and 1993.

No. nests with clutch No. nestsby No. (orbrood size) Param- final-brood-size Year nests 1 2 3 4 5 andeter No. categories year nests 0 1 2 3 4 5 Black-crownedNight-Heron clutchsize (F/• = 2.70% 3 df) Black-crownedNight-Heron 1992 15 0 1 11 3 0 Brood size• 1 (no test) 1993 44 0 1 25 15 3 1992 3 2 1 1993 0 -- -- Snowy Egret clutch size (FI= 25.22'*, 2 df) Brood size 2 (FI = 3.08,•' 2 dr) 1992 22 0 5 13 4 0 1992 3 1 0 2 1993 48 0 0 11 37 0 1993 6 0 3 3 Black-crowned Night-Heron initial brood size Brood size 3 (FI = 2.39,•' 2 df) (FI = 9.31', 4 df) 1992 8 2 0 1 5 1993 25 2 0 2 21 1992 15 3 3 8 1 0 1993 44 0 6 25 11 2 Brood size 4 (FI = 4.582, "s 3 df) 1992 1 0 0 1 0 0 Snowy Egret initial brood size (F/= 21.11'*, 3 df) 1993 11 1 0 1 5 4 1992 22 2 9 8 3 0 Brood size 5 (no test) 1993 48 0 4 13 31 0 1992 0 1993 2 0 0 0 0 1 1 % P > 0.05; *, P < 0.05; **, P < 0.01. ' Clutch-sizecalculations based on greatestnumber of eggsseen in Snowy Egret nest that hatchedat least one egg. b Initial brood size was initial number of chicks seen alive in nests Brood size 1 (no test) 1992 2 1 1 that hatchedat leastone egg. • Fisher's exact test (Lehmann 1975) and Monte Carlo P-value esti- 1993 0 mates (Cytel 1991). Brood size 2 (FI = 1.295,"' 2 df) 1992 9 2 1 6 1993 4 0 1 3 Brood size 3 (FI = 14.59,** 3 df) bird within a day or two, and determine probable 1992 8 1 5 0 cause of death. 1993 13 0 0 8 We computedsurvival rates using program JOLLY, Brood size 4 (FI = 5.09,•' 3 df) 1992 3 0 0 2 1 0 which suggesteda number of modelsto fit our data 1993 31 1 0 8 20 2 while incorperatingan open-populationmodel (i.e. allowing both deathsand emigration;Pollock et al. Final broodsize definedas no. birdssurviving >_ 14 days. 1990).The data were not sufficientto run goodness- Nestsrequired -> 1 hatchlingto qualify. Refersto initial broodsize (at hatching). of-fit teststo comparemodel efficiency,so we useda simplemodel (Model D) that yieldeda singleconstant survival rate over the entire period, rather than day- or week-specificrates (as in the nestlingsurvival data). in the colony at the time, we compareddifferences Time-period-specificmodels did not fit the data sig- in A- versus C-chicks with survival data. Because dif- nificantlybetter. In addition,we usedprogram CON- ferenceswould be expectedto accentuatewith time, TRAST (Hines and Sauer 1989),which incorporates we used the differences in mass of chicks on the last associated variance and covariance estimates to con- day both were weighed. We combined these values duct Z-testson survival rate datato comparethe sur- for all nestsin all year-colonies,and then testedwith vival ratesbetween species and years. linear regressionfor a relationshipbetween these dif- Nest success,growth, and survivalrelationships.--We ferencesand the number of birds surviving per nest examined whether any growth statisticsor indices to 14 days of age. We conducteda permutation test were associated with survival or nest-success mea- (Manly 1991) using 1,000 permutationsto calculate sures.The LIFETESTprocedure in SAS produceslog the P-valueassociated with the regressionslope and, rank statisticsand marginal test statisticsfor the sur- thus,avoided the assumptionof normallydistributed vival and covariateparameters (SAS Institute 1987). residuals. We usedthe slopesof individual growth asa covariate Colonycensus.--We were interestedin examining of survival rate. For each species,we examinedall the relationshipbetween the reproductivemeasures combinationsof broodorder and year-colonyfor the describedabove and the numbersof nestingbirds in effect of growth as a covariate. the study coloniesand the region. We made counts To test whether the differencein massof siblings of the number of birdsat both heronriesin late May is a reliable indicator of general feeding conditions both years by walking the perimeter of the colonies January1996] Growthand Survival of Night-Heron and Egret Chicks 123

T,•I•I,œ3. Resultsof testscomparing mass and culmen-length growth rates (• + SD, with n in parentheses) for Black-crownedNight-Heron and Snowy Egret chicksin 1992 and 1993.

Mass(g) Tukey Culmen-length (mm) Year Colony Chick growth per daya testb growth per dayc Black-crowned Night-Heron 1992 Causeway A 31.28 + 10.54(8) A a 1.87 + 0.32 (8) 1992 Causeway B 35.94 + 9.70 (7) A a 1.90 + 0.24 (7) 1992 Causeway C 7.41 + 10.18(4) B a 1.59 + 0.57 (4) 1993 Causeway A 39.33 + 5.31 (11) A a 2.23 + 0.18 (10) 1993 Causeway B 34.88 + 7.35 (9) A a 2.21 + 0.21 (9) 1993 Causeway C 33.85 + 11.99(9) A b 2.17 + 0.36 (9) 1993 Causeway D 26.83 + 3.94 (2) A a 1.76 + 0.30 (2) 1993 Causeway E 4.10 (1) A 1.58 (1) 1993 Willis A 39.18 + 12.17 (15) A a 2.27 + 0.26 (15) 1993 Willis B 43.88 + 9.22 (16) A a 2.24 + 0.39 (16) 1993 Willis C 31.33 + 14.30 (15) A b 2.04 + 0.26 (13) 1993 Willis D 41.34 + 30.35 (3) A a 2.22 + 0.53 (4) Snowy Egret 1992 Causeway A 15.53 + 4.51 (15) A a 1.99 + 0.66 (17) 1992 Causeway B 10.43 + 7.21 (12) AB a 2.14 + 0.57 (13) 1992 Causeway C 0.28 + 2.07 (5) B a 1.48 + 0.92 (6) 1992 Causeway D -1.83 (1) AB a 2.07 (1) 1993 Causeway A 14.23 + 3.91 (11) A a 1.80 + 0.16 (11) 1993 Causeway B 15.32 + 3.17 (13) A a 1.94 + 0.19 (13) 1993 Causeway C 9.44 + 12.80(7) A ab 1.52 + 0.40 (8) 1993 Causeway D 17.03 + 12.76(3) A a 1.28 + 0.41 (5) 1993 Willis A 17.65 + 12.76 (15) A a 1.84 + 0.36 (17) 1993 Willis B 16.99 + 5.23 (16) A a 1.87 + 0.28 (17) 1993 Willis C 14.38 + 7.80 (14) AB b 1.67 + 0.39 (19) 1993 Willis D 5.01 + 11.46 (7) B a 1.55 + 0.43 (9) ßMain effectsfor night-herons:year-colony, F = 8.71'**, 2 df; hatching order, F = 2.47*, 4 df; interaction,F = 2.32*, 5 df. Main effectsfor egrets:year-colony, F = 6.68'*, 2 df; hatching order, F = 8.72'**, 3 df; interaction,F = 3.36**, 6 df; m,p > 0.05; *, very closeto P = 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001. bResults of Tukey multiple-comparisonprocedure (overall a = 0.05), where uppercaseletters comparegrowth ratesamong chicks of different hatchingorders within sameyear and colony,while lowercaseletters compare years and coloniesof chicksof samehatching order. Growth-rate meansthat sharea given letter of samecase are not significantlydifferent (Tukey comparison). ' Main effectsfor night-herons:year-colony, F = 9.85'**, 2 df; hatching order, F = 1.77m, 4 df; interaction,F = 0.87n', 5 df. Main effectsfor egrets:year-colony, F = 1.82•, 2 df; hatchingorder, F = 5.43**, 3 df; interaction,F = 0.84•, 6 df. Statisticalsymbols as in footnote a.

with three to six observersand periodically flushing both specieshad larger initial broodsin 1993 all adults.The colony was divided into sectionsgen- than in 1992 (Table 1). erally 50 to 100 m long (5 to 40 m wide) using local When comparing chick production per nest landmarks. Each person would count the number of (basedon successfullyhatched nests), we found individuals of one or two speciesthat would flush for both speciesin both years that the initial from the colony.In addition, population estimatesof brood size had a significanteffect on the num- other coloniesfrom the mid-Atlantic region were so- licited from the Virginia CoastReserve (The Nature ber of chicksthat survived 14 days (for night- Conservancy)and biologistsfrom the statesof Mary- herons in 1992 [F! = 13.08, P = 0.043] and 1993 land and Virginia. [F! = 41.68, P < 0.001];for egretsin 1992 [F! = 15.94,P = 0.008]and 1993[FI = 15.50,P = 0.017]). We performed separateanalyses by speciesand RESULTS brood size, but small sample sizes limited our interpretations(Table 2). Egretswith broodsize Nest success.--Acomparison of clutch sizes 3 producedsignificantly more chicksper brood for the two speciesusing Fisher's exact tests to 14 days in 1993 than in 1992;night-herons indicatesa year effectfor egrets(in 1993,clutch- with brood size 3 showed no significantyear es were larger)but not for night-herons(Table effect (Table 2). 1). Comparingmean initial broodsizes for each When we compared successfulwith unsuc- speciesbetween the two years,we found that cessfulnests, we againfound yearly differences. 124 ERWINET AL. [Auk,Vol. 113

TABLE4. Effect of hatching order (by year-colony) TABLE5. Effectsof year-colony on survival rates in on survival rates of Black-crowned Night-Heron Black-crowned Night-Heron and Snowy Egret and Snowy Egret chicksthrough 20 days. chicks through 20 days for given hatching orders.

Colony Year n • X 2 Hatching order n s" X 2 Black-crowned Night-Heron Causeway 1992 25 0.706 8.21' Black-crowned Night-Heron Causeway 1993 40 0.947 1.63ns 1 42 0.902 2.17 n' Willis 1993 60 0.765 5.77 ns 2 38 0.892 0.05 n' 3 36 0.725 5.34 n' Snowy Egret 4 8 0.000 7.76* Causeway 1992 47 0.484 13.89'* Causeway 1993 52 0.691 18.97'* Snowy Egret Willis 1993 81 0.618 19.39'* I 51 0.823 2.45 n• 2 51 0.750 8.03* n., p > 0.05; *, P < 0.05; **, P < 0.01. 3 48 0.415 10.40'* aMean survival(daily) in broodsof four (3 df) calculatedfrom Kaplan- Meier estimatesduring period from hatchingto mid-nestling(15 to 20 4 26 0.168 3.96 •s days) period. Only initial broodsof four were used. n', P > 0.05; *, P < 0.05; **, P < 0.01. a Mean survival (daily) for each hatching order calculatedfor the three year-colonies.Means basedon Kaplan-Meierestimates during For night-herons in 1992, we recorded 5 un- periodfrom hatchingto mid-nestling(15 to 20 days). successful and 10 successful nests, while in 1993 there were only 3 unsuccessfulnests and 41 successfulnests in two colonies (FI = 5.93, P = more obvious,even though samplesizes are not 0.02). For egrets,we recorded 4 unsuccessful adequatefor testing.The singleegret chick from and 18 successfulnests in 1992,and only 1 un- 1992 that survived long enough to permit successfuland 47 successfulnests in 1993 (FI = weighing had a negative growth rate. In con- 3.53, P = 0.07). trast, in 1993, D-chicks grew an average of 5 Growth.--We first evaluated whether brood and 17 g/day at the Willis and Causewaysites, size had an effect on the growth rate of chicks respectively. of either speciesfor all year-colony combina- Survival.--For the early-nesting period, we tions. No resultswere significant (all P > 0.05), first examined the data on chick survival to 20 so data were combined from all brood sizes for days for a speciesdifference. For each brood the rest of the analyses. member order (A-D), we found no differences Mass differences were more variable and more in survival between speciesfor any combina- divergent among brood members than were tion of year and colony (all P > 0.05). However, culmen differencesfor both species(Table 3). the order in which chicks hatched had a sig- For mass growth in night-herons, a two-way nificant effect on survival in all three cases for ANOVA indicated that both the main effects of egrets,but only at the Causewaycolony in 1992 year-colony and hatching order were signifi- for night-herons(Table 4). When year effects cant (hatching order was at critical limit, P = are considered for each brood order, differences 0.0501),and that for the year-hatchorder in- were noted for both speciesbut most strongly teractionalso was significant(Table 3). For cul- for egrets (Table 5, Fig. 2). A-chicks of both men length, only the year-colonyeffect was sig- specieswere the only ones consistentin show- nificant (Table 3). ing no year effect in survival to 20 days (Table For egrets, mass growth rates were signifi- 5). Figure 3 illustrates the differencesamong cantly different due to the main effectsof year- years and colony sites for each of the brood colony and hatching order, and their interac- ordersseparately. In general, D-chickssuffered tions (Table 3). Massgain of C- and D-chicksin suchlow survival that statisticalpower was low. 1992 was much reduced, compared to A- and During the late-nestingperiod, radio-marked B-chicks (Table 3). For culmen, the main effect A-chicks of both specieshad very high survival of year-colonywas not significant,but hatching ratesfrom the time of marking (ca. two weeks order was (Table 3). old) until fledging age in both years (Fig. 4). The year-colonyeffect was already apparent Within both years, there were no significant in egret C-chicksthat grew significantlyfaster differences (P > 0.05) in survival rates between at Willis in 1993 than at Causewayin 1992 (Ta- species. ble 3); for D-chicks, the differences are even In examining dispersal age, we were ham- January1996] Growthand Survival of Night-Heron and Egret Chicks 125

Black-crownedNight-Heron SnowyEgret 1.0 i.•-- Causeway1992 0.8 ,

0.6 0.6

0.4

0.2 0.2

0.0 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20

1.0 Causeway 1993

0.8 0.8 '

0.6 0.6

0.4

0.2 0.2

0.0 ' •' •' •' • ' ' ' • ' ' ' 0.0 ' • ' • ' • ' • '•'0'•'2'½.'•'•'•'8'2'0

Willis 1993 1.0

...... •":'•c...!;...i,..;...i...• .;...•...;...•...;...•...•...i 0.8 ' 0.8

0.6 0.6

0.4

0.2 0.2

0.0 2 4 6 8 10 12 14 16 18 20 0'00' ? ' • ' 6 8 1•01•2 1•4 1•6 1•82•0

Days from hatching date Days from hatching date

A-chick .... B-chick ...... C-chick .... D-chick

Fig. 2. Survivalfunction estimatesfor Black-crownedNight-Heron and Snowy Egretchicks during early growth period (to 20 days)at eachcolony site. 126 ERWIN ET AL. [Auk, Vol. 113

Black-crownedNight-Heron Snowy Egret

1.0 A-chick 1.0

0.8 0.8

E o.6 0.6

'• 0.4 0.4

0.2 0.2

o.o 0.0 0

B-chick 1.0 1.0

0.8 0.8 '•0.6 0.6

0.4

0.2 0.2

0.(• 0

C-chick 1.0 -'-• ...... ,

0.8 i...... 0.8

0.6 -.>_ l •=0A -- 0,4

0.2 0.2

, 2I , 4• , 6• , 8• , 10• , 12• , 1•4' 1•6' 1•8 ' 20

Days from hatching date Days from hatching date

Causeway1992 .... Causeway1993 ...... t44•s 1993

Fig. 3. Survival-functionestimates (to 20 days)for eachbrood member of nestlingBlack-crowned Night- Heronsand SnowyEgrets at eachstudy colony.Because of limited samplesize, D- and E-chickswere not included. January1996] Growthand Survival of Night-Heronand Egret Chicks 127

Black-crownedNight-Heron Snowy Egret

1.00 1.00

0.95 0.95

•0.90 0.90

0.85 0.85

0.80 0.80 [,092(._--,,) 1992(n=22)1993(n=20) 0'750' 10 2'0 3'0 40 5'0 0'750 10 2•0 3•0 40 5'0 Days from hatchingdate Days from hatching date

Fig.4. Survival-functionestimates of radio-markedBlack-crowned Night-Heron and Snowy Egret A-chicks from hatchinguntil dispersalfrom colonyin 1992and 1993. pered becauseof chick mortality and lack of Virginia, on the upper EasternShore of Virgin- preciseinformation concerning hatching dates; ia. for these reasons, we did not include all radio- Growth, success,and survivaL--We conducted marked birds in the analysis.With a reduced separateanalyses for influencesof growth(mass) sample,no year effectsfor either specieswere rate on survival (to 20 days) for each combi- found(night-herons, F = 3.68,P > 0.05;egrets, nation of hatchingorder, species,and year-col- F = 2.70, P > 0.10), but we did detecta signif- ony; none of these were significant(log rank icant speciesdifference in 1993 (F = 4.19, P = X 2, all P > 0.1). However, sample sizes were 0.05; Table 6). limited for someof the hatchingorders. When After dispersal,survival was estimated using we combined all hatching orders, survival in capture-recapturemodels. We found survival of egrets showed a significant associationwith A-chicksto be lower during the postdispersal growth ratesat two of three year-colonycom- phasethan during the late-nestlingphase (Fig. binations(Causeway 1992, X 2= 16.34,P < 0.001; 5 and Table 7). Approximately six weeks after Willis 93, X 2 = 12.16,P = 0.002).When all year- dispersalbegan, survival rates ranged only from colony and brood orders were combined, we 0.35 to 0.65 for the two speciesin both years (Fig. 5), comparedto ratesof 0.80to 1.00during the nestlingperiod (Fig. 4). The probablecauses T^BLE6. Average age (ñ SD) of dispersalfor Black- of mortality included predation (9), collision crowned Night-Heron and Snowy Egret chicks in Virginia in 1992 and 1993." with powerline wires (1), and unknown (2). On most occasions,we recovered the radio only; Year nb Age (days) typically,only a few featherswere present.Most Black-crownedNight-Heron of these radio recoveries(with no carcass)were 1992 7 (11) 60.3 _ 7.1 locatedin woodedareas under a largetree where 1993 14 (20) 55.8 ñ 3.1 a Great Horned Owl (Bubovirginianus) perched. Snowy Egret Without direct evidence or observations, how- 1992 12 (22) 56.3 + 5.6 ever, we cannot confirm whether owls, other 1993 13 (20) 53.0 ñ 3.7 predators,or scavengerscarried the corpsesto ' ANOVA between speciesfor 1992 (F = 1.66, 1 df, P = 0.215) and their recovery location. Recovery locations 1993 (F = 4.19, 1 df, P = 0.051). ranged from a powerline site near Salem, New bSample size of radio-markedbirds. Number in parenthesesindicates total number of birds radio-marked(including reusedradios); birds Jersey,east to a salt-marshnear Cape Henlopen, were removedfrom sampleif hatchingdate was uncertain,or in a few Delaware,and a woodedcemetary near Saxis, casesof mortality. 128 ERWINI/T ^L. [Auk,Vol. 113

TABLE7. Weekly estimatedsurvival rates' (•) of im- 1.0 mature Black-crownedNight-Herons and Snowy Egretsdispersing from Chincoteaguecolonies, 1992 and 1993.

Year fi _+SE (95% confidence interval) ff o.• •_ '. Black-crownedNight-Heron 1992 0.846 _+0.068 (0.712-0.979) 0.4 1993 0.845 _+0.040 (0.766-0.924) Snowy Egret 0.2 10 2•0 3•0 4•0 5'0 6•0 1992 0.919 + 0.038 (0.846-0.993) Days f•om first dispersal location 1993 0.833 + 0.042 (0.750-0.916) Fig. 5. Survival ratesof dispersedBlack-crowned ßUsing computer program JOLLY (Pollock et al. 1990). Night-Heron and SnowyEgret chicks from a Virginia colonyin 1992and 1993as determined by aerialradio- telemetry.Note that X-axisis daysfrom firstdispersal locationin both years.The first locationflight in 1992 found a highly significant association(X 2 = was 16 dayslater than first flight in 1993. 27.43,P = 0.001).This wasnot the casefor night- herons (P > 0.10). For both species,we examined whether the proportional massof a "subordinate" chick (C) g/day reported, averaged for brood members; compared to A-chicks influenced the number Wolford and Boag1971). In 1993,at the Virginia of chicks per nest surviving to 14 days. For colonies, we found culmen-length growth of egrets,we found an effect (r2 = 0.15, P = 0.006); A-chicksof 2.23 to 2.27 mm/day at the two sites, however, for night-herons, no effect was evi- not significantlyhigher than the 2.1 mean re- dent (r 2 < 0.001, P = 0.80). ported for A-chicks in Texas (Custer and Peter- Colonycensuses.--Black-crowned Night-Her- son 1991), but markedly higher than the 1.97 ons,Snowy Egrets,and other wading-bird spe- reported from Alberta in 1964 (Wolford and cies nested in greater numbers in both study Boag 1971). coloniesand in the region in 1993 than in 1992 For Snowy Egrets,comparisons are more lim- (Table 8). The other species included Great ited. Growth rates in massof 17.7 g/day for Egrets(Casmerodius albus), Cattle Egrets(Bubul- A-chicksin 1993in Virginia at the Willis colony cus ibis), Little Blue Herons (Egretta caerulea), exceededthe reported 14.1 g/day for A-chicks TricoloredHerons (E. tricolor),and GlossyIbises in Texas (Custer and Peterson 1991). Culmen- (Plegadisfalcinellus). However, the patternwith- length growth rates were similar between the in coloniesin the region was inconsistent,with two studieswith rangesfrom 1.80 to 1.99 mm/ somecolonies decreasing in 1993 (Table 8). This day in Virginia in the two years,compared to indicates probable immigration into the Chin- 2.0 mm/ day in Texas(Custer and Peterson1991). coteague colonies in 1993 from other coastal We were surprisedto find lower survivalrates coloniesnearby. The total numbersof all wad- (range 0.25 to 0.60) during the 40- to 60-day ing birds in these mixed-speciescolonies in- postfledgingperiod than during the late-nest- creasedfrom 1992 to 1993 in the region. ling (ca. 14- to 40-day)period (range 0.80 to 1.0) for both species.In the Great White Heron, Powell and Bjork (1990) found that, of 42 im- DISCUSSION mature herons that dispersedto southern Flor- ida from the Keys, only 10% survived the first Growth rates in mass and culmen length six months. In contrast, in American Crows found at our coloniesare the highestreported (Corvusbrachyrhynchos), Caffrey (1992) reported in North America for Black-crownedNight- that young only had a 70%survival to fledging, Herons.Mass increases of greaterthan 39 g/day but from fledging to two monthspostfledging for A-chicksexceed those reported from Cape had an 89% survival rate. Because crows are res- Cod,Massachusetts (œ = 33.5g/day, range26.2- ident species,they may not be faced with the 37.6 among brood members; Parsonsand Bur- hazards encounteredby naive first-year mi- ger 1981),Texas (38.4 g / dayfor A-chicks;Custer grant birds moving over great distances into and Peterson 1991), or Alberta, Canada (31.3 unknown areasshortly after fledging. January1996] Growthand Survival of Night-Heronand Egret Chicks 129

TABLE8. Differencesin numbersof nestingwading birdsat majorMaryland-Virginia heronries, 1992 and 1993.

Black-crowned Night-Heron Snowy Egret All speciesb Colony 1992 1993 1992 1993 1992 1993 Willis Marsh, Virginia a 47 90 308 391 742 1,182 Causeway,Virginia a 32 37 233 636 872 2,032 Chimney Pole, Virginia -- -- 76 102 226 294 Cobb Island, Virginia 36 61 180 28 672 251 Wreck Island, Virginia 41 80 232 150 788 584 FishermanIsland, Virginia 260 235 64 44 480 484 CoardsMarsh, Virginia 12 14 30 80 82 166 Heron Island, Marylandc -- -- 208 329 874 1,759 South Point, Marylandc -- -- 72 70 713 1,035 œ(including study sites) 62 74 140 183 546 779 œ(excluding study sites) 71 78 108 100 480 572 ßIncludes Great Egrets, Cattle Egrets,Little Blue Herons,Tricolored Herons, and GlossyIbises. bStudy sites at Willis Marshand Causeway. ' Basedon flight-countdata, with conversionsto nestestimates (D. Brinkerpets. comm.).

From the perspectiveof using wading birds stronger affinities to brackish and salt-marsh aspotential bioindicatorsin estuarinewetlands, habitatsfor feeding. our data suggestthat Snowy Egretsare a better "indicator species" candidate than are Black- ACKNOWLEDGMENTS crowned Night-Herons. When comparing growth rates, reproductive success,chick sur- We thank the staff of the ChincoteagueNational vival, and total nesting populations, Snowy Wildlife Refugefor their logisticalsupport during the project. We also thank Dr. and Ms. Robert Baker for Egretsrevealed stronger relationships than did their fieldassistance and help with photography.Nu- Black-crownedNight-Herons. Survival was sig- merous staff members from Patuxent Wildlife Re- nificantlyassociated with massgrowth ratesfor search Center also assisted us on occasion. Tom Custer egretsbut not night-herons.For both species, and Kathy Parsonsprovided suggestionsand ideas A-chicksseemed relatively immune to yearly during the developmentof the project.We appreciate variations,while B-, C-, and D-chick growth and the useful commentsof T. Bancroft, T. Custer, P. Fred- survival seemed sensitive to between-year erick, G. M. Haramis, R. Hines, G. Schnell, and an changes.Our data suggestthat food conditions anonymous reviewer on earlier drafts of the manu- in the estuarymay have been inferior in 1992 script. comparedto 1993. Unfortunately, no fisheries data were being collectedin the region to eval- LITERATURE CITED uateindependently estuarine productivity. Such BIORK,R., ANDG. V. N. POWELL.1994. Relationships data are vital to validate the utility of "bioin- between hydrologic conditionsand quality of dicators." foraginghabitat for RoseateSpoonbills and other Black-crowned Night-Herons may not be wading birds in the C-111 basin.Final report to good "bioindicators" becauseof their general- EvergladesNat. Park, Homestead, Florida. ized feeding in both freshwater and estuarine CAFFREY,C. 1992. Female-biaseddelayed dispersal environments, both in Massachusetts (Kornhis- and helping in American Crows. Auk 109:609- er and McColpin 1994) and Virginia. Young 619. birds,small mammals, frogs, and fishcomprised CUSTER,T. W. In press. Environmental contami- the diet of young Black-crownedNight-Herons nants.In Statusand conservationof herons(J. A. Kushlan and H. Hafner, Eds.). Academic Press, during the nestingseason at our studysite (Er- London. win and Custerin press).For estuarineindicator CUSTER,T. W., ANDR. G. OSI•ORIq.1977. Wadingbirds species,both SnowyEgrets and TricoloredHer- asbiological indicators: 1975 colony survey. U.S. ons (Frederick et al. 1992, Ramo and Busto 1993) FishWildl. Serv.,Spec. Sci. Rep. Wildl. 206. are probably good candidatesbecause of their CUSTER,T. W., ANDD. W. PETERSON,JR. 1991. Growth 130 ERW•NET ^L. [Auk, Vol. 113

rates of Great Egret, Snowy Egret, and Black- for lifetime data. J. Wiley & Sons,New York. crowned Night-Heron chicks.Colon. Waterbirds LEm•N, E. L. 1975. Nonparametrics. Holden-Day, 14:46-50. Inc., San Francisco. CYrEL. 1991. Statxact,vers. 2. Cytel SoftwareCorp., MANLY, B. F.J. 1991. Randomization and Monte Car- Cambridge,Massachusetts. lo methodsin biology. Chapman and Hall, New DmMOND, A. W., •a•v F. L. FItaON (Eds.). 1987. The York. value of birds. Int. Counc. Bird Preserv. Tech. MAYFIELD,H. 1975. Suggestionsfor calculatingnest Publ. 6. success. Wilson Bull. 87:456-466. ERWIN,R. M., AND T. W. CUSTER.In press. Herons MORRISoN,M.L. 1986. Bird populationsas indicators as indicators. In Status and conservation of her- of environmentalchange. Curr. Ornithol. 3:429- ons (J.A. Kushlan and H. Hafner, Eds.).Academic 451. Press, London. PARSONS,K. 1994. The Arthur Kill oil spills:Biolog- Fox, G. A., • D. V. WESELOH. 1987. Colonial wa- ical effectsto birds. Pages215-237 in Before and terbirds as bio-indicators of environmental con- after an oil spill: The Arthur Kill (J. Burger, Ed.). tamination in the Great Lakes. Int. Counc. Bird Rutgers Univ. Press,New Brunswick, New Jer- Preserv. Tech. Publ. 6:209-216. sey. FREDERICK,P., R. BIOPAC,G. T. BANCROFT,AND G. V. N. PARSONS,K., AND J. BURGER.1981. Nestling growth POWELL.1992. Reproductive successof three in early- and late-nesting Black-crownedNight speciesof herons relative to habitat in southern Herons. Colon. Waterbirds 4:120-125. Florida. Colon. Waterbirds 15:192-201. PVa•IC•LL,D. B., AND L. R. SHUG•T (Eds.). 1993. Bio- FREDEmCK,P. C., AND M. COLLOP¾.1989a. Nesting markers:Research and application in the assess- successof five ciconiiform speciesin relation to ment of environmental health. Springer-Verlag, water conditionsin the Florida Everglades.Auk New York. 106:625-634. POLLOCK,K. H., J. D. NICHOLS,C. BROWNIE,AND J. E. FREDEmCK,P. C., AND M. COLLOP¾. 1989b. The role HINES. 1990. Statistical inferences for capture- of predationin determining reproductivesuccess recaptureexperiments. Wildl. Monogr. 107. of colonially-nesting wading birds (Ciconi- POWELL,G. V. N., III, AND R. BIORK. 1990. Studies of iformes) in the Florida Everglades.Condor 91: wading birds in Florida Bay:A biologicalassess- 860-867. ment of the ecosystem.National Audubon So- FREDERICK,P. C., M. SPALDING, AND G. V. N. POWELL, ciety report to E. Ordway Dunn Foundation,Tav- III. 1993. Evaluating methods to measurenest- ernier, Florida. ling survivalin TricoloredHerons. J. Wildl. Man- RAMO,C., ANDB. BUSTO.1993. Resourceuse by her- age. 57:34-41. ons in a Yucaton wetland during the breeding GR•¾, J. S. 1980. Why do ecologicalmonitoring? season. Wilson Bull. 105:573-586. Mar. Pollut. Bull. 11:62-65. SASINST•TUTE.1987. Statisticalanalysis system, vers. HENSLER,G., ANDJ. D. NICHOLS.1981. The Mayfield 6.04. SAS Institute, Inc., Cary, North Carolina. method of estimating nesting success:A model, SOK•L,R. R., AND F. J. ROHLF. 1981. Biometry, 2nd estimators, and simulation results. Wilson Bull. ed. W. H. Freeman and Co., San Francisco, Cal- 93:42-53. ifornia. HINES,J. A., ANDJ. R. SAUER.1989. Program Contrast. TEMPLE,S., ANDJ. WIENS. 1989. Bird populationsand U.S. Fish Wildl. Serv., Fish Wildl. Tech. Rep. 24. environmental changes;can birds be bio-indi- KORNHISER,L., AND A. MCCOLPIN. 1994. The Samp- cators? Am. Birds 43:260-270. son'sIsland mystery. Conserv.Sci. Quart. (Win- TP,EMSr•¾, J., • L. ELtaSON. 1979. Effectsof human ter):10-13. disturbance on breeding Black-crownedNight KUSHLAN,J. A. 1993. Colonial waterbirds as bioin- Herons. Auk 96:364-369. dicators of environmental change. Colon. Wa- WOLFORD,J. W., AND D. A. BOAG. 1971. Distribution terbirds 16:223-251. and biology of Black-crowned Night Herons in LAWLESS,J.F. 1982. Statistical models and methods Alberta. Can. Field-Nat. 85:13-19.