VRITION IN PEREGRINE FLCON EGGS WILLIM. BURNHM, JMES H. ENDERSON, 2 ND THOMS j. BORDMN 3 The Peregrine Fund, Crnell University, 1424 N. E. Frntage Rad, Frt Cllins, Clrad 80524 US; 2Department f Bilgy, The Clrad Cllege, Clrad Springs, Clrad 80903 US; and 3Department f Statistics, Clrad State University, Frt Cllins, Clrad 80523 US BSTRCT.--Eggs cllected frm captive and wild Peregrine Falcns (Falc peregrinus) were used t examine variatin in eggshell thickness, length, breadth, and initial weight t reslve questins abut eggshell data frm wild falcns. Fr captive falcns, shell thickness f first clutches did nt change ver the years a falcn laid r with embrynic develpment. Eggshells in third clutches, but nt secnd clutches, were significantly thinner than thse frm first clutches. Greatest variatin in shell thickness existed between eggs within a clutch and did nt differ significantly between wild and captive eggs. Entire clutches f wild falcns shuld be represented in future studies t maximize the chance f btaining a representative sample in regard t shell thickness. Egg size (L, B, and fresh weight) decreased ver the years a captive falcn laid. significant decrease in size (B and fresh weight) als ccurred in secnd and third clutches laid the same year. Ratcliffe's Index generally appeared t be a reliable indicatr f shell thickness in captive-laid eggs. Received 30 September 1983, accepted 23 February 1984. EGGSHELL thickness has been reduced in a va- riety f predatry birds (ndersn and Hickey 1972), a phenmenn shwn t be clsely crrelated t the presence f DDE (Cade et al. 1971, Lincer 1975, Peakall and Kiff 1979). Variatin in shell thickness relating t embrynic develpment, the laying female's age, and prductin f multiple clutches has smewhat cluded an elucidatin f the precise relatinships between DDE and eggshell thinning. Papers n nnraptrial birds have fcused n variatin in eggshell thickness, and sme have defined sampling schemes t maximize the detectin f thickness changes. Thickness decreases with embrynic develpment have been reprted in Cedar Waxwings (Btnbycilla cedrrutn; Rthstein 1972), Japanese Quail (Cturnix japnica; Kreitzer 1972), and White-faced Ibises (Plegadis chihi; Capen 1977). In Cedar Waxwings, but nt in the White-faced Ibis, eggs frm smaller clutches als have thicker shells. In a study f museum eggshells f Clapper Rails (Railus lngirstris), Black-crwned Night Herns (Nycticrax nycticrax), White Ibises (Eudcitnus albus), Nrthern Mckingbirds (Mitnus plygtts), and Lggerhead Shrikes (Lanius ludvicianus), variatin in shell thickness amng eggs within clutches cntributed as much t the sample variance as variatin between clutches frm different females (Klass et al. 1974). The large captive ppulatin f breeding Peregrine Falcns (Falc peregrinus) develped by The Peregrine Fund (Frt Cllins, Clrad and Ithaca, New Yrk) t prduce falcns fr release prvided a surce f eggs frm females f knwn age and rigin. dditinally, the substitutin f captive-reared yung fr wild Peregrine Falcn clutches in Clrad prvided us with the pprtunity t cmpare variatin in egg parameters amng wild-laid eggs with thse prduced in captivity (Burnham et al. 1978). The wild birds were subject t DDEinduced eggshell thinning and prduced shells averaging 13% thinner than thse f the captive birds (Endersn et al. 1982). In this study we examined variatins in shell thickness, egg breadth (B), and egg length (L) f wild-laid eggs within clutches, between clutches, amng sites (females) by year, and between all sites ver several years. We cmpared shell-thickness measurements fr captive-laid eggs with embrynic develpment. Thickness, L, and B in captive-laid eggs were als related t female age and t secnd and third clutches laid within a year. Egg length and breadth were cnsidered because they are used t calculate Ratcliffe's Index, a cmmn measure fr eval- 578 The uk 101: 578-583. July 1984
July 1984] Variatin in Peregrine Falcn Eggs 579 uating shell thinning (Ratcliffe 1980). These analyses clarify histrical data n Peregrine Falcn eggshell thickness and suggest sample techniques fr cllecting peregrine eggs that will prvide the mst accuratestimate f shell thinning resulting frm DDE residues. METHODS falcns and lst in the lfts. n analysis f variance was perfrmed t determine pssible linear and quadractichanges ver years f laying n length, breadth, and calculated initial weight. dditinally, clutch averages were calculated fr length, breadth, and initial weight f first, secnd, and third clutches f eggs laid within a year by 4 captive peregrines, and the averages were cmpared by t-tests. Measurements f 299 eggs laid in first clutches thrughut a perid as lng as 11 yr by 14 captive peregrines were used t assess the ptential effect f female age n shell thickness. We perfrmed an analysis f variance n the yearly means t determine differences. Eighty-eight first clutches frm 16 captive females ver as many as 11 yr prvided data n egg breadth, length, and initial weight. Frm 1 t 4 eggs were available per clutch; smetimes eggs were brken by RESULTS ll eggs were treated similarly. Egg length and breadth were measured t the nearest 0.1 mm with vernier calipers. We calculated fresh egg weights fr Variatin amng eggs within clutches accaptive-laid eggs by using K.v(LB2), where cunted fr tw-thirds (67.1%) f the variatin Kw(0.0005474) is the bserved cefficient fr pere- in shell thickness f eggs laid by wild peregrine eggs, as described by Burnham (1983). Infertile grines (Table 1). Differences between clutches eggs and thse with dead embrys were stred in a accunted fr 26.2% f the variatin. Little refrigeratr fr up t several days befre they were variatin in shell thickness existed amng eggs pened and emptied. When the cntents were re- at a nest-site lcatin ver years (2.4%) r bemved, r when fertile eggs hatched, the shells were tween all nest-site lcatins ver years (4.3%). rinsed with tap water and allwed t dry at lw am- Within-clutch variatin in shell thickness bient humidity. Eggshell thickness was measured p- (with membranes) was great amng bth wild tically t +0.004 mm frm fresh chips taken frm three places at the shell equatr. The cular scale and captive eggs. The mean cefficient f variemplyed was calibrated with a Bausch and Lmb atin (C V = SD/œ) f shell thickness f captive 0.001-mm stage micrmeter (Endersn et al. 1982). eggs (0.046, n = 71) was nt significantly dif- The 153 wild-laid eggs measured fr shell thickness and the 146 measured fr breadth and length ferent (P > 0.10, t-test) frm that f wild eggs (0.054, n = 29) when nly 3- and 4-egg clutches were frm first and secnd clutches laid at 14 terri- were cnsidered. tries in Clrad and nrthern New Mexic ver as Thickness f captive-laid eggshells prduced many as 5 yr in the perid 1977-1981. In mst in- by females that were cpulating with their stances, we culd nt verify that the same female laid mates and thse that were artificially insemiat a particular territry in successive years. Usually, nated had similar average thicknesses (F = 2.07, 3 r 4 eggs were available fr measurement frm each clutch. nested analysis f variance was used t de- P > 0.15). These means are 0.34 mm (n = 143) termine surces f variatin (Skal and Rhlf 1969). and 0.35 mm (n = 111), respectively. Fr captive peregrines, we determined shell thickness fr eggs fertilized as a result f bth cpulatin and artificial inseminatin and perfrmed an analy- Embrynic develpment had n apparent effect n thickness f captive-laid eggshells. The mean thicknesses f clutches f hatched eggs sis f variance t determine whether r nt a differ- (0.281 mm) were nt significantly different (P > ence in thickness ccurred. We measured 15 eggs that 0.50, df = 27, t-test) frm thse f clutches f were infertile r develped fr less than 1 week and infertile eggs r eggs dying within 7 days (0.282 14 hatched eggs frm the same clutches laid by 11 mm). Membranes frm eggs with full-term emfemales. The means f shell thickness fr each grup were cmpared by a paired t-test. Shell-thickness brys (0.077 mm, n = 65) were nt significantly means, usually fr 3 r 4 eggs, were calculated fr different in thickness (P > 0.05, t-test) frm first, secnd, and third clutches and cmpared by membranes frm eggs with n develpment t-tests. In all, 13 sets f three clutches prduced by 4 (0.079 mm, n = 27). females were used. The average f mean thicknesses (with membranes) f captive-laid eggs was 0.357 in first clutches, 0.350 in secnd clutches, and 0.342 mm in third clutches, with 95% cnfidence intervals f +0.014, 0.016, and 0.014, respective- ly. The differences in clutch means f shell thickness (with membranes) between first and secnd clutches, secnd and third clutches, and first and third clutches were cmpared by paired t-tests. The latter was significant (P <
580 BURNHM, ENDERSON, ND BORDMN [uk, Vl. 101 TBLE 1. nalysis f variance f measurements f eggs frm wild peregrines in Clrad and nrthern New Mexic, 1977-1981. Shell Surce f Percentage variatin df Mean squares Variance cmpnent variability thickness a mng sites 13 0.001015 0.000022 4.30 mng years 19 0.000815 0.000012 2.41 mng clutches 13 0.000732 0.000131 26.16 mng eggs 99 0.000336 0.000336 67.51 Egg length mng sites 13 19.7885 1.3167 31.72 mng years 19 6.4774 0.5818 14.01 mng clutches 12 3.7716 0.7080 17.05 mng eggs 102 1.5450 1.5450 37.22 Egg breadth mng sites 13 8.1097 0.3724 20.92 mng years 19 4.3833 0.8374 48.17 mng clutches 12 0.5631 0.0060 0.34 mng eggs 102 0.5441 0.5441 30.57 144 eggs in 46 clutches with mean thickness f 0.305 mm with shell membrane. 0.025, n = 13), and there was an apparent trend t thinner eggshells after the first clutch. The number f years a captive falcn laid eggs had n significant effect upn r apparent trend in shell thickness (Fig.! and Table 2). The amunt f variatin in length and breadth f wild eggs was much mre evenly dispersed ver the surces f variatin than was variatin in shell thickness (Table 1). pprximately ne-third f the variatin existed amng eggs within a clutch. Mst f the remaining variatin ccurred amng years and sites. We analyzed measured averages f clutches f captive-laid eggs and fund trends tward narrwer eggs f lighter weight when first,.35 With Membrane.30 E E.25 Withut Membrane.20 29 37.15 0 2 N. 11 11 11 14 12 9 9 9 5 2 2 Females 35 45 43 27 27 27 15 6 6 N. Eggs I I I I I I I [ I I 4 6 8 10 12 Year Fig. 1. Peregrine Falcn eggshell thickness ver years f laying. The mean shell thickness was 0.354 mm with membranes and 0.275 mm withut membranes. Laid
July 1984] Variatin in Peregrine Falcn Eggs 581 TBLE 2. nalysis f variance f measurements f eggs frm captive Peregrine Falcn taken ver the years f laying. Shell thickness with membrane Significance Sum f squares df Mean square F f F Linear 1.427!!.427 0.4272 0.514 Quadratic!.499!!.499 0.4489 0.504 Shell thickness withut membrane Linear 3.684! 3.684!.247 0.266 Quadratic 3.803! 3.803 1.287 0.258 Length Linear 16.144! 16.144 19.447 <0.005 Quadratic 5.527! 5.527 6.658 0.012 Breadth Linear 18.852! 18.852 38.784 <0.00! Quadratic 5.585! 5.85 11.490 0.00! Initial weight Linear 188.244! 188.244 53.068 <0.00! Quadratic 6.904! 6.904!.946 0.168 secnd, and third clutches laid in the same year were cmpared (Table 3). We als fund a significant decrease in length, breadth, and initial weight ver years f laying (Fig. 2 and Table 2). DISCUSSION The large variatin (within clutches) that we have fund amng wild-laid eggs suggests that samples based n ne egg in a clutch r a few fragments may insufficient!y represent the average shell thickness prduced by a falcn. If eggs r shell fragments were frm secnd r third clutches, a further increase in variatin wuld be pssible because f the trend tward thinner eggshells after the first clutch. Because little variatin exists in mean shell thickness amng clutches f different females within a year r f the same female (nest site) in different years, accuracy may nt be enhanced by cllecting eggs annually frm many females within a ppulatin. The great within-clutch variatin in shell thickness seen in wild-laid eggs may nt be due t DDE cntaminatin, because the much thicker captive-laid shells vary similarly. The average shell thickness f first clutches laid thrughut the entire reprductive life f captive falcns des nt change significantly TBLE 3. Variatin in egg size between clutches frm captive peregrines. Initial weight (g) Length (mm) Breadth (mm) Vari- Vari- Varidf ance Mean t P df ance Mean t P df ance Mean t P Clutches C 9 3.11 C2 9 2.70 C 9 7.00 Cntrasts C vs. C2 18 C2 vs. C 18 C vs. Cs 18 47.64 1! 2.55 51.56 10 0.91 40.91 44.99 1! 1.95 51.17 10 0.93 40.00 42.46 1! 3.74 50.71 10 1.24 38.92 3.49 0.0! 22 0.61 0.5 20 2.20 0.05 2.58 0.02 22 0.64 0.5 20 2.26 0.02 5.15 0.001 22 1.12 0.2 20 6.51 0.00!
582 BURNHM, ENDERSON, ND BORDMN [uk, Vl. 101 50 45 49 44 58 57 Length [] Breadth ' Weight 48 43 47 42 56 55 46 41 45 40 44 39 43 38 54 53 52 51 [] [] D D [] [] 0 D [] 42 37 50 0 41 36 49 40 35 N= 48 13 14 14 13 8 6 6 6 4 2 2 0 2 4 6 8 10 12 Year Laid Fig. 2. Peregrine Falcn egg measurements ver years f laying. Each datum represents the average f 2-4 eggs laid in the first clutch.year-.female-h TBLE 4. The effect n Ratcliffe's Index values f the number f years a female has laid. Change in index Change in index Rat- frm frm Year Initial Number f Shell cliffe's Year 1 Year 2 laid Breadth a Length a weighp clutches b weight ½ Index a (%) (%) 1 40.41 54.56 48.76 13 3.90 1.77-2.2 2 41.08 53.28 49.52 14 3.96 1.81 +2.3 3 41.02 52.61 48.73 14 3.90 1.82 +2.8 +0.6 4 40.90 52.16 47.88 13 3.83 1.80 + 1.7-0.6 5 41.06 52.28 47.62 8 3.81 1.77 0.0-2.2 6 40.89 51.74 47.36 6 3.79 1.79 + 1.1-1.1 7 40.30 51.78 46.08 6 3.69 1.77 0.0-2.2 8 39.64 51.59 44.61 6 3.57 1.75-1.1-3.3 9 39.28 50.79 42.95 4 3.44 1.72-2.8-5.0 10 39.43 50.14 43.23 2 3.46 1.75-1.1-3.3 11 38.05 50.85 40.28 2 3.22 1.66-6.2-8.3 Mean values f clutch averages (graphed n Fig. 2). Number f clutches and females; all clutches represented are the first laid each year. Eight percent f initial weight. Ratcliffe (1980).
July 1984] Variatin in Peregrine Falcn Eggs 583 (Fig. 2). Therefre, this statistic frm wild ppulatins, which may have skewed age distributins, may nt be biased tward thinner r thicker shells. Shell-thickness data frm hatched, addled, r infertile eggs are equally representative. The secnd cmmn analysis f shell thinning is Ratcliffe's Index, which is the dry weight f the shell (mg) divided by the prduct f its length (mm) and breadth (mm) (Ratcliffe 1980). decrease in length r breadth ver the years f laying culd bias the Index dwnwards. T test this, we held shell weight t be 8% f initial egg weight and determined the Index values ver the years f laying (Table 4). The change in the Index did nt exceed +2.8% frm the first t seventh years f laying, whether the average Index frm the first r secnd year f laying was used as the base. fter the seventh year f laying, smaller Index values result, even thugh shell thickness (Fig. 1) may remain cnstant. ppulatin with an age distributin skewed tward very ld birds culd prduce eggs with Indecies biased dwnwards. If many egg samples frm a ppulatin were frm secnd r even third clutches, a further bias culd be intrduced, because a significant reductin in egg breadth and initial weight ccurred between clutches laid in a single year by a falcn (Table 3). reductin f length was als evident, but nt significant. CONCLUSIONS Our results frm wild-laid eggs suggested little need t cllect eggs frm a large number f females each year within an eclgically similar ppulatin. T maximize the chance f btaining a representative sample f shellthickness changes, all eggs in either first r secnd clutches shuld be represented, r the number f cllectin sites r cllectin years shuld be increased. Eggshells frm hatched, infertile, r addled eggs were equally representative. Ratcliffe's Index appears t be a reliable indicatr f shell thickness unless measure- ments are frm secnd r third clutches r frm eggs laid by very ld females. CKNOWLEDGMENTS We are grateful t G. Craig, D. Berger, and W. Heinrich fr cllectin f a prtin f the wild eggs and t C. Sandfrt, D. Knkel, J. Weaver, and W. Heck fr cllectin f many captive eggs. The manuscript was reviewed by and imprved by cmments frm D. ndersn, T. Cade, R. Ryder, W. Graul, and R. Stabler. P. Burnham typed the manuscripts, and B. Cade and J. Zum Brunnen prvided statistical cnsultatin. The research was supprted by The Peregrine Fund, Inc. (President, T. J. Cade). LITERTURE CITED NDERSON, D. W., & J. J. HICKEY. 1972. Eggshell changes in certain Nrth merican birds. Prc. 15th Intern. Ornithl. Cngr.: 514-540. BURNHM, W. 1983. rtificial incubatin f falcn eggs. J. Wildl. Mgmt. 47: 158-168. -, J. CRIG, J. H. ENDERSON, r W. H. HEINRICH. 1978. rtificial increase in reprductin f wild Peregrine Falcns. J. Wildl. Mgmt. 42: 625-628. CDE, T. J., J. L. LINCER, C. M. WHITE, D. G. ROSENEU, L. G. SWRTZ. 1971. DDE residues and eggshell changes in laskan falcns and hawks. Science 172: 955-957. CPIN, D. E. 1977. Eggshell thickness variability in the White-faced Ibis. Wilsn Bull. 89: 99-106. ENDERSON, J. H., G. R. CRIG, W.. BURNHM, r D. D. BERGER. 1982. Eggshell thinning and rganchlrine residues in Rcky Muntain peregrines and their prey. Can. Field-Naturalist 96: 255-264. KLSS, E. E., H. M. OHLENDORF, r R. G. HETH. 1974. vian eggshell thickness: variability and sampling. Wilsn Bull. 86: 156-164. KREITZER, J.F. 1972. The effect f embrynic devel- pment n thickness f eggshells f Cturnix Quail. Pulry Sci. 51: 1764-1765. LINCER, J. 1975. DDE-induced eggshell thinning in the merican Kestrel: a cmparisn f field and labratry results. J. pp1. Ecl. 12: 781-793. PEKLL, D. B., & L. F. KIFF. 1979. Eggshell thinning and DDE residue levels amng Peregrine Falcns (Falc peregrinus): a glbal perspective. Ibis 121: 200-204. RTCLIFFE, D. 1980. The Peregrine Falcn. Vermillin, Suth Dakta, Bute Bks. ROTHSTEIN, S. I. 1972. Eggshell thickness and its variatin in the Cedar Waxwing. Wilsn Bull. 84: 469-474. SOKL, R. R., & F. J. ROHLF. 1969. Bimetry. San Francisc, W. H. Freeman and C.