DDE-INDUCED EGGSHELL-THINNLNG IN THE AMERICANKESTREL: A COMPARISON OF THE FIELD SITUATION

AND LABORATORY RESULTS

BY JEFFREY L. LINGER

Mote Marine Laboratory, 9501 Blind Pass Road, Sarasota, Florida 33581, U.S.A.

INTRODUCTION

Because of the stable and lipophilic nature of many organochlorines (Noakes & Benfield1965; Ecobichon & Saschenbrecker 1968), they show both environmental persistence(Herman, Garrett & Rudd 1968; Dimond & Sherburne 1969) and 'biological magnifica-tion' (Hickey, Keith & Coon 1966). By these mechanisms many species of birds of preyhave become highly contaminated with organochlorines. Reports of this contaminationwere paralleled by field observations of eggshell-thinning and consequent reproductivefailure (Ratcliffe 1967, 1970; Hickey 1969; Berger et al. 1970; Cade & Fyfe 1970). Cadeet al. (1971) reported a strong correlation between eggshell-thinning and egg p,p'-DDE •content and indicated the need for experimentation on causal relationships. j

Although DDE-induced eggshell-thinning has been shown experimentally with both Ipredacious (Wiemeyer & Porter 1970; McLane & Hall 1972) and non-predacious birds!(Heath, Spann & Kreitzer 1969; Longcore, Samson & Whittendale 1971), the dose-response relationship has not been characterized nor have captive and wild predatorsbeen compared.

In choosing a species of predacious bird to study a phylogenetically related species tothose showing population decline, such as the peregrine falcon (Falco peregrinus Tun-stall) was selected. The American kestrel or sparrow hawk (F. sparverius L.), in additionto being congeneric with the peregrine, is more numerous and readily available and hasbeen Successfully bred in captivity (Willoughby & Cade 1964; Porter & Wiemeyer 1969,1970; Wiemeyer & Porter 1970).

The objectives of this study using the American kestrel were to elucidate the dose-response relationship between dietary DDE and eggshell-thinning, and to compare theselaboratory results with egg residues and eggshell-thinning found in a north-easternpopulation of the same species. In view of the recent controversy concerning the 'real'relationship between DDE and eggshell-thinning (Hazeltine 1972; Wiemeyer & Porter1972; Switzer, Wolfe & Lewin 1972; Risebrough 1972; Blus et al. 1972), the followingdata appear timely.

METHODSField

To establish an easily accessible source from which eggs for residue monitoring andyoung for laboratory and aviary experiments could be obtained, wild kestrels wereencouraged to use nest boxes. During the late winters and early springs of 1969 and 1970eighty-five wooden nest boxes were constructed and placed in an approx. 155-km2 area,which is delineated by the United States Geological Survey map Ithaca East, N. Y.

781

782 Eggshell-thinning in wild and DDE-dosed American kestrels

Outside dimensions of nest boxes were 29 cm wide x 32-5 cm deep with a sloping andremovable roof providing a height of 38 cm at the rear to 30-5 cm at the front. A hole,8-2 cm wide x 7-2 cm high, was located in the front under a 6-4 cm roof overhang. Nestboxes were placed in open-field habitats usually on lone trees, or along hedgerows andforest edges. Trees were selected for their sturdiness and proximity to presumed huntingareas. Nest boxes were placed 5-13 m off the ground. An average of 25% of these weretaken over by kestrels during 1969-71. Approximately 65% were taken over by starlingsand the remainder divided by red squirrels (Tamiasciurus hudsonicus (Erxleben)), greysquirrels (Sciurus carolinensis Gmelin)), honey bees (Apis meliifera L.), empty boxes andscreech owls (Otus asio (L.)) in descending order (see Lincer 1972).

Field observations varied in frequency and duration depending on objectives and stageof reproductive behaviour. Nest boxes were checked and cleaned out once in late winterand then checked once a week starting in mid-April. As soon as an occupant becameestablished, it was either left alone if not a kestrel or observed more frequently (two orthree times a week) if a kestrel. On several occasions, especially early in the spring,starlings were removed if the nest box appeared to have some chance of attracting kestrels.Incubating birds were disturbed to varying degrees. Some would leave the nest box before.the ladder touched the tree; others had to be lifted in an incubation position while theeggs were checked and then replaced to continue incubation.

AviaryExperimental birds for the laboratory breeding experiment were trapped along the

New York Atlantic coast during autumn migration.During 1971, kestrels were maintained in the Raptor Ethology Building, Cornell

University Laboratory of Ornithology. Individual juxtaposed cages, separated by ply-wood walls, measured approx. 3-04 m wide, 6-08 m long and a sloping roof provided a

'.height of 3-96-5-48 m. Each cage contained one male and one female. Kestrels were fed,30-g, day-old cockerels daily and ad libitum. Feeding of dosed prey to kestrels began inmid-March and continued through late August. The cages were rarely entered for foodremoval or for any other purpose. A rear door in the nest box allowed direct observationsand removal of eggs. Generally the number, condition, and relative warmth of eggswere inspected daily. All other observations were made through a one-way glass

AnalyticalDosing of food was accomplished as follows. A concentrated mixture of DDE was

made up in a small volume of acetone which was added to A.C.S. (American ChemicalSociety) grade sesame oil, gently heated and shaken until the acetone had evaporated andthen brought up to volume in a glass-stoppered bottle. Dosing was done daily by injecting0-2 ml of the mixture into dead cockerels in the breast region. Control food was injectedwith 0-2 ml sesame oil only. Dietary doses were one of the following: control, 0-3, 3-0,6-0 or 10-0 ppm, based on wet weight (WW).

Because of the possible effect of embryonic development on the eggshell thickness(Kreitzer 1972), all eggs were collected five days after the last egg of the respective clutchwas laid. This also allowed adequate embryonation for determination of viability.

Eggs collected from the field or aviary experiments were checked for cracks or dentsand weighed to the nearest 0-001 g and then measured for breadth and length to thenearest 0-1 mm with a Mitutoyo No. 505-633 caliper equipped with a vernier. They werethen either stored temporarily at 5° C or blown to remove contents. After the shell was

i sloping androot. A hole,erhang. Nestdgerows andimed huntingof these werer by starlingsleben)), greyty boxes and

ves and stagen late winterpant became.-ntly (two ori the spring,ting kestrels,st box beforec /hile the

'd along the

ing, Cornellated by ply-f provided arels were fed•els began inred for foodDbservationsmth of eggsa1" window.

•f DDE wasin Chemicalporated andby injecting

was injected-ol, 0-3, 3-0,

11 thickness:ctive clutchbility.;ks or dentsngth to the. They werehe shell was

JEFFREY L. LINGER 783broken at the equator, the interior was gently washed with lukewarm water to removealbumen while not disturbing the membranes attached to the shell. The shell was allowed 'to drain for approximately 30 s on a paper towel before measuring the shell thickness tothe nearest 0-005 mm at the equator with an Ames No. 25ME thickness gauge. Theprocedure of not letting the shell completely dry before measuring might have differedfrom that of other investigators. At least four measurements were taken on each shell,half. High and/or low extremes were disregarded in calculating an average shell thickness.A second measure of eggshell thickness was Ratcliffe's Index (RI), which equalled theweight of the dried eggshell (mg) divided by the length times the breadth in millimetres(Ratcliffe 1970).

The analytical procedure for organochlorine residues was similar to that of Cade et al.(1971). The clean-up procedure followed that outlined in the Pesticide Analytical Manual(U.S.D.H.E.W. 1970, Section 2.2! A, 3-4). The determinative step involved using a VarianAerograph Model 2100 gas chromatograph, with 2 Ni63 electron capture detectors. Thegas chromatographic columns and operating conditions used were: (A) 1% QF-1 onChrom. G. A/W DMCS 70/80, 4 mm (i.d.)x2 m, injector—238° C, column—203° C,detector—281° C, nitrogen flow—30 ml/min. (B) 1% silicone GE XE-60 on Varaport#30 100/120, 2 mmx 1-66 m, injector—238° C, column—203° C, detector—281 ° C,nitrogen—25 ml/min. (C) 2% SE-30 on Chrom. G A/W DMCS 70/80, 4 mm x 2 m,injector—250° C, column—225° C, detector—281 ° C, nitrogen—90 ml/min. Two of thethree columns, filled with liquid phases of different polarities, were used simultaneously.

For the first florisil fraction of the field samples containing DDE, columns A and Cwere used. For the second florisil fraction of the samples possibly containing dieldrincolumns A and B were used. The residues in the sample were calculated for both columnsand the lower value chosen for expressing ppm. Agreement was usually within 15%.Because of the unknown residue nature of field-collected sample residues (Lincer 1973),an additional column clean-up step (Snyder & Reinert 1971) was employed to separateorganochlorine pesticides from polychlorinated biphenyls (PCBs). Recoveries of common l

organochlorines (e.g. /?,p'-DDE, -TDE and -DDT) exceeded 88%. No corrections were jmade for recovery.

RESULTS AND DISCUSSION

DDE residues in kestrel eggs collected in the field averaged 35 ppm, based on oven dry jweight (OD) for five pairs in 1969,42 ppm for twenty-two pairs in 1970, 33 ppm for seven ;pairs in 1971 and 37 ppm for five pairs in 1972 (Table 1). Observed variation is, at least,,partially explained by differential migration and choice of prey species (Lincer & Sher-burne 1974; Mueller 1974). Respective RIs for 1969-72 were 0-999, 0-960, 0-968 and0-910 which gives an unweighted average of 10% thinner than the pre-DDT thicknessgiven by Anderson & Mickey (1972) for interior-northern North America. Since it islikely that the above eggshell measurement methodology was slightly different thanAnderson & Hickey's, it is .unwise to compare the two on a millimetre thickness basis.The most extensive sampling of eggs was done in 1970 for DDE. Twenty-two nests weresampled and, with one exception, at least two eggs per clutch were analysed for DDE andmeasured for eggshell thicknesses to estimate within-clutch variation. Average within-clutch coefficients of variation for DDE and RI were 12-1 and 4-93%, respectively.

A random sample of eggs collected during 1970 was also analysed for DDT, DDD,|dieldrin and PCBs (Table 1). PCBs, quantified as Aroclor 1254, averaged 37 ppm (OD)1

784 Eggshell-thinning in wild and DDE-dosed American kestrels

, while TDE and DOT averaged less than 1 ppm. The apparently high average of almost 3ppm for dieldrin may be biased by one high sample (9-70 ppm), as seven samples collected

'during 1971 averaged only 0-15 ppm dieldrin. Jn addition, only one of these containedmore than trace amounts of heptachlor epoxide (0-15 ppm),

Heavy metal concentrations in eggs collected in 1970 were: mercury, 0-17; copper,iO-38; cadmium, 0-08; lead, 0-72 ppm on a calculated fresh weight basis (Lincer &McDuffie 1974).

Table 1. Organochlorine residues in American kestrel eggs collected in thefield during 1969-72, Ithaca, New York (values refer to x ppm, OD±s.d,

[range]; number of nests sampled in parentheses*}Year

Compound

DDE

DOT

ODD (TDE)

Dieldrin

Heptachlor epoxide

PCBsJ

.1969

34-6 + 28-1 (5)[7-02-74-2]

-t

0-05 ±0-00(1)

1970

41-9 ±25-8 (22)[8-78-100]

0-87 ±1-05 (5)[Tr.§-2-34]

0-26 ±0-57 (5)[NDf-1-28]

2-70±4-19(5)[ND-9-70]

37-0-t 56-0 (5)[6-72-137]

1971

33-2+16-4(6)[18-3-64-9]

0-15±0-22 (7)[Tr.-0-64]

0-02 ±0-06 (7)[ND-0-15]

1972

36-8 + 8-37(5)[12-6-74-9]

* Numbers of eggs/pair sampled for DDE each year: 1969, two eggs/nest; 1970, one from one nest,two from nineteen, four from one and five from one; 1971 and 1972, one egg/nest. One egg from eachnest was sampled for other organochlorines during 1969, 1970 and 1971.

t Not analysed for.t Quantified as Aroclor 1254 on the basis of two peak heights which did not interfere with DDE.§ Trace (i.e. present but below detector's range of linearity).H None detected (i.e. usually less than 0-05 ppm, OD).

Wiemeyer & Porter (1970) indicated that dietary DDE at 2-8 ppm (WW) was capableof inducing eggshell-thinning in captive kestrels. Fig. 1 reflects the dose-response relation-ship which is important from a toxicological point of view. The upper curve illustrates theexperimentally induced percent thinning compared to the control, while the lower curvedisplays this relationship to pre-DDT thickness. Kestrels on 0-3 ppm DDE producedeggshells that were not different from controls, however, dietary levels of 3-0, 6-0 and 10ppm resulted in average shell-thinnings of 14-0, 17-4 and 21-7%, respectively, as reflectedby RL In terms of a pre-DDT RI of 1-06 (from Anderson & Hickey 1972), the 'control',0-3, 3, 6 and 10 ppm diets resulted in 13-2, 6-0, 25-5, 28-3 and 32-1% thinning, respect-

i ively. Eggshell breakage began to occur when eggshells became more than approximately'22% thinner than pre-DDT eggshells. However, some eggs collected from the 10 ppmj group were up to 38% thinner than pre-DDT eggshells.

Figure 2 shows that DDE residues increased as a direct function of dietary DDE levels.Interestingly (Fig. I) the initial steep decline in shell thickness resulting from dietary

i DDE in excess of 0-3 and less than 3-0 ppm is followed by a levelling off. These data are.similar to those of Porter & Wiemeyer (1969) and Wiemeyer & Porter (1970), and theresulting negatively decelerating curve, reflecting the response of the kestrel, parallels the

^e of almost 3iples collectedese contained

0-17; copper,>is (Lincer &

in the

1972

*± 8-37 (5)2-6-74-9]

om one nest,gg from each

c with DDE.

capable

llustrates the: lower curve>E produced0,6-0 and 10', as reflectedthe 'control1,ing, respect->proximatelythe 10 ppm

DDE levels.Yom dietaryicse data are70), and theparallels the

JEFFREY L. LINGER 785

published field observations on raptor eggshell-thinning and egg DDE residues fromEurope (Ratcliffe 1970) and North America (Cade et al. 1971).

Table 2 shows the experimentally induced eggshell-thinning. In view of the eggshell-thinning controversy based on statistical analyses, an intensive statistical treatment wascarried out on the data. First, these values were subjected to a hierarchical analysis ofvariance (Myers 1972), from which significant differences between pairs within the samedietary group (f(ll,46) = 13-00, P<0-01) and between dosage levels (F(4,ll) = 7-65,/*<0-05) were obtained. The differences between parents receiving the same diet indicates

0(3)

7-88(5,2) 84-5(7,5) 154 (4.2) 245 (2,1) 1390(3,?)

•90 (7,31 Mean DOE residues in eggs (ppm, OD) (sar«ple size (eggs, pairs) in parentheses)__|_____________I I_________;_____________I________________________________I_________________________________________i

0 - 3 2 3 6 I O,'.tO (1-13) ( 7 ' 5 6 ) ( l l - 3 ) (22-6) (37-7)

Dietary DOE. ppm, based on WWof food (estimated ppm, OD assuming 7

FIG. 1. Relationship between dietary DDE and egg DDE residues and resulting eggshell-thinning. 1971 aviary experiment, % of control (•), pre-DDT eggshell thickness (x). Points(encircled) o and ® are from other experiments and have not been considered for curvesdrawn. (1) Porter & Wiemeyer (1969), (2) Wiemeyer & Porter (1970), (3) this study, pilotexperiment, 1970. Residue data refer to this study. Experimentally induced events ( - - - )

are relative to pre-DDT thickness data.

that genetic and/or cage effects may be involved in producing reliable differences ineggshell thickness and, therefore, the parental and/or cage source of variation ought to betaken into consideration as an important potential variable. The effects of dosage levelconfirm the reliability of the trends displayed in Table 2. Fisher's two-tailed LeastSignificant Differences at P< 0-05 and 0-01 were 0-0202 and 0-0285, respectively (Li 1964).Mean differences greater than or equal to these values were statistically significant. Thus,with the exception of the 0-3 ppm group, all other experimental groups differed signi-

786 Eggshell-thinning in wild and DDE-dosed American kestrels

ficantly from the controls. In addition, there were differences (/><0-01) between the 0-3ppm group and higher dietary groups (for both thickness (mm) and Rl).

To understand more about the relationship between the DDE content (ppm, OD) ofthe eggs and their shell-thinning, correlations between DDE levels and the dependentmeasures were calculated for .both the laboratory and the field data. The Pearson Product

25O -

200

0-3(1-13)

10(377)

Dietary ODE, pprn. WWof food (ppm. OD. assuming 73-5 % wuter in prpy)

FIG. 2. Relationship between dietary DDE fed to American kestrels and resulting DDEresidues in eggs produced. Numbers in parentheses refer to numbers of eggs and pairs,

respectively. Y= 6-485*,0D) + 4-325.

Table 2. Eggshell-thinning effect of dietary DDE on the American kestrelTreatment Mean shell thickness (mm)§ % Change Significance!

±S.D. from control

Control0-3 ppm DDE3-0 ppm DDE6-0 ppm DDE

10-0 ppm DDE

0-171 ±0-012 (15,4)t0-175 ±0-000 (8,2)0-145 + 0-007 (18,5)0-135±0-018 (4,2)0-126 ±0-00* (4,2)

0+ 2-1

-15-1-22-8-29-2

NS******

t Number of eggs, number of pairs.J Based on Fisher's two-tailed LSD, experimental group significantly different

from controls at: * P<0-05; ** P<Q-Ol; *** P<0-001.§ Includes membrane.

Moment correlations (Hays & Winkler 1970) between experimentally induced egg DDE(ppm, OD) and eggshell thickness were -0-840 and -0-890, respectively, for mm and RT.Each of these values differed (/><0-05) from zero. Similar correlation coefficients were

/een the 0-3

>m, OD) ofdependent

on Product

JEFFREY L. LINGER 787

DDEpairs,

ttrel

nt

I egg DDEimand RI.;ients were

also determined for the field data (Table 3). Based" on an egg by egg comparison, whichcould be criticized for looking more powerful than it really is, the correlations betweenDDE and thickness and between DDE and RI were, respectively, -0-365 and -0-400.When the clutch means were used these respective correlations were elevated slightly to-0-413 and -0-470 (P = 0-05 and /><0-05, respectively). Nevertheless, all of thesecorrelations differed from zero. Thus, for both the laboratory and the field observations,!significant inverse relationships were obtained between DDE content and eggshell-1

thinning. These relationships were obtained for actual measurements of thickness and,those reflected by RI; these two measures being also highly correlated (r = 0-791, basedon forty-six eggs collected in 1970).

Because it has been suggested that the Product Moment correlation is inappropriatefor comparing egg DDE content and thinning (Hazeltine 1972; Blus el al. 1972), Spear-man's Rank Difference correlations were also calculated and were essentially equivalent

Table 3. Statistics on field observations on shell-thinning and egg DDE content in theAmerican kestrel

Testing for

rxy against sH0: Pxy = 0, wherex = DDE in egg (ppm, OD) andy = shell thickness (mm)Same as above, where y = RI

Correlation between DDE andshell thicknessCorrelation between DDE* andRICorrelation between logIO DDEand shell thicknessSame as row 3 above

Test used*

Student's two-tailed /, df = 45

Same as above, df = 44

Pearson Product MomentCorrelation, N = 47Same as above, N = 46

Same as above, A' = 47

Spearman's Rank DifferenceCorrelation, N = 47

Result and significance

Correlation differs fromzero (P< 0-05)

Correlation differs fromzero(/><0-01)r = -0-365; /><0-05

r = -0-400; /><0-01

r = -0-364; /»<0-05

r = -0-371; /»<005

* Treating data egg by egg (vs. clutch means).

to the values obtained with the other method, as they should be (see Hays & Winkler1970). For example, the Spearman Rank Difference correlation between DDE and shellthickness was -0-371, a value which is only marginally different from the ProductMoment value of —0-365, and is attributable to ties in the rankings.

In any experimentation dealing with the possible effects of environmental contaminantson wildlife, there is a need to relate experimentally induced changes to the field situation.As mentioned earlier, the DDE residues in the field-collected kestrel eggs and their shell-thinning showed an inverse relationship as did those eggs produced under conditions ofcaptivity (Fig. 3). These values were subjected to regression analyses to permit directcomparisons of the relationship of DDE to eggshell-thinning in both groups of birds.Both the wild and the experimental birds show a logarithmic relationship between thick-ness and DDE concentration. The linear regressions for both the wild birds (F(l,19) =7-05, /><0-05), and for the experimental birds (F(},\ 1) = 42-8, /»<0-OI) were significant,and the slopes for these curves were not different (f(l,30) = 0-017). Thickness of theeggshells of the experimental birds, however, were lower than those of the wild birds(F(l,31) = 31-9, P<0-01) (Snedecor & Cochran 1967). The variances of the residualsabout each plotted regression line did not differ significantly (F(19,l 1) = 1-52, /J>0-05).These comparisons indicate that both the experimental and wild birds show the same

788 Eggshell-thinning in wild ami DDE-dosed American kestrels

response to DDE. However, all of the experimental birds la id th inner-shel led eggs, whichmight be at t r ibutable to genetic differences and/or to effects un ique ly associated withcaptivity.

Perhaps the most important findings from this research are the elucidat ion of the dose-response relationship between dietary DDE and eggshell-thinning in a raptor, and, to theauthor's knowledge, this is the first time that the eggshell- thinning relationship has been

.shown not to be significantly difierent for the same species in both the laboratory and thefield.

100

80

70

x • •

x • x

1 I I.10 20 30 4O

DDE residues in eggs [ppm, 00}

1 i L i I _.60 SO I GO 2OO 3OO

FIG. 3. Relationship between mean clutch shell-thickness and DDE residue of kestrel eggscollected in Ithaca, New York during 1970 (•) and same relationship experimentally induced

with dietary DDE(x) .

This is confirmed when DDE residues in various raptor eggs are compared and offersthe additional advantage of correlative eggshell-thinning data (Fig. 4). Gyrfalcons(Falco rusticolus L.) on the Seward Peninsula feeding, for the most part, on relativelyuncontaminated resident prey produced eggs containing 1-10 ppm DDE (OD) andeggshells no different in thickness than pre-DDT values. Rough-legged hawks (Buteolagopus (Pontoppidan) on the Colville River in Alaska produced eggs with slightly higherresidues (6-45 ppm DDE) and a small, but discernible, degree of eggshell-thinning(-3-3%) (Cade et al. 1971). A population of prairie falcons (Falco mexicanus Schlegel)in Alberta and Saskatchewan produced eggs with 17 ppm DDE and eggshells 11%thinner than pre-DDT thickness (Fyfe et al. 1969). Another population of prairie falcons,

eggs, which)ciated with

of the dose-, and , to thelip has beenory and the

:1 eggsduced

and offersGyrfa Iconsi relatively(OD) and

•vks (Buteohtly higher•11-thinnings Schlegel)•hells 11%rie falcons,

JEFFREY L. LINGER 789

from Colorado and Wyoming, produced eggshells-14% thinner than pre-DDT thicknessand their eggs contained 27 ppm DDE (Enderson & Berger 1970). Four pairs of merlins(F. columbarius L.) breeding in Newfoundland produced eggs conlaining 40 ppm. Theeggshells of the four pairs of merlins were 11 % thinner than eggs collected before 1947(Temple 1972). Snyder et al. (!973) found a highly significant inverse correlation betweenCooper's hawk (Accipiter cooperii (Bonaparte) (egg DDE residue and eggshell thickness.

22 -

21 -

16

12

Peregrine(Unqavo. Canaaa)

PeregrineO (Colville Rv.

Alaska)

Peregrine (Taiga, Aiasho)G

OCooper's Hawk t (Anz.-N. Mex

G Prairie falcon (Col-Wyorn.)

Merlin (Newfoundland)G O Proine falcon

(Aiberto-Saskatch.

O Peregrine (Amchitka. Alaska)O Arner kestrel (Irhoca, N.Y.)

G Cooper's Hawks* (Ariz.- N. Mex}

O Rough-legged hawk (Colville Rv., Alaska)

Gyrfalcon (Seward Penn., Alaska)^_____I I I i I i 1 I I

10 ?0 30 40Mean ppm (00) DOF! in eggs

60 80 lOO 2OO 500

FIG. 4. Relationship between mean DDE content of eggs and mean percentage decrease ineggshell thickness in North American raptors. See Table 4 and text for references to variousraptor species. * refers to fertile eggs, t refers to broken eggs. Conversion of residue levels toppm (OD), for consistency, based on lipid and/or percentage water given by respective

authors or Food and Drug Administration.

In addition, they found significantly higher DDE levels in eggs from unsuccessful nests1

than in eggs from successful nests. In the present study, kestrels around Ithaca, New York,,contained an average of 36 ppm DDE and showed 10% shell-thinning. Peregrines (Falcoperegrinus Tunstall) on Amchitka, a fairly non-migratory race, produced eggs containing48 ppm DDE and eggshells 7-5% thinner than pre-DDT eggshells. From the same study(Cade et al. 1971), eggs from the more migratory peregrines inhabit ing the taiga regionand the Colville River of Alaska contained 192 and 252 ppm DDE, respectively andpossessed eggshells 17 and 22% thinner than pre-DDT data. Peregrine eggs collected inUngava, Canada by Berger et al. (1970) contained 80 ppm DDE and were, surprisingly,21% thinner than pre-DDT eggshells. Since Berger and his co-workers emphasized that

Table 4. Eggshell-thinning and population status of North American falconiformes^Reference

Hickey & Anderson (1968)

Hickey & Anderson (1968)

Hickey & Anderson (1968)Hickey & Anderson (1968)

Cadeetal. (1971)

Cade et al. (1971)

This study

Fyfe et al. (1969)Enderson & Berger (1970)

Temple (1972)

Snyder et al. (1973)

Hickey & Anderson (1968)Hickey & Anderson (1968)

Hickey & Anderson (1968)Hickey & Anderson (1968)Cade et al. (1971)Cade era!. (1971)Cade et at. (1971)Berger et al. (1970)Hickey & Anderson (1968)Hickey & Anderson (1968)Hickey & Anderson (1968)

* Much of the eggshell-thinning data supplied by Anderson & Hickey (1970).t See text for scientific names.

Region or population

California

California

New JerseyMd-Va.

Seward Penn.

Colville Rv., Alaska

Ithaca, New York

Alb. and Sask., CanadaCol. and Wyom.

Newfoundland

Ariz. and N. Mexico

Brevard Co., FlaOsceola Co., Fla

VermontB.C., CanadaAmchitka, AlaskaTaiga, AlaskaColville River, AlaskaUngava, CanadaCaliforniaMass.New Jefsey

Period

1953-67

1947-65

19571957

1968-69

1967-69

1969-72

1967-681967-68

1969

1969-71

1947-521959-62

19461947-531969-701968-691967-6919671947-5219471950

Eggshell-thinning*(% change from pre-DDT)

Red-tailed Hawk+ 2-7

Golden Eagle-2-8

Osprey-25-2-8

GyrfalconNo change

Rough-legged Hawk-3-3

American kestrel-8-0

Prairie falcon— 11-14

Merlin-9

Cooper's Hawk-3 to -18Bald Eagle

-18-20

Peregrine-1-8-1-4-7-5-17-22-21-18-21-26

Populationstatus

Stationary

Stationary

DecliningStationary

Stationary

Stationary

Stationary?

QuestionableStationary

Stationary

Stationary

DecliningDeclining

StationaryStationaryStationaryQuestionableDecliningDeclining?DecliningExtirpatedExtirpated

bb

JEFFREY L. LINGER 791

= CO O

It <L)

= C

•J ocir

u ^ij (5a a< xJ UJ

l

T3C

aa3

T300

•— IflC D.5 6f C=3 O

- o<~*~0

their sample was biased towards broken or cracked eggs, it is interesting to note that thissample is somewhat inconsistent with the other raptor populations (Fig. 4).

The inverse correlation between DDE in North American raptor eggs and eggshellthickness (Fig. 4) is clear but does not prove a causal relationship since other chemicalsor factors could be involved. On the other hand, experiments carried out with the |American kestrel (Lincer 1972; Peakali & Lincer 1972) have shown that those toxicchemicals most widely distributed in North America, like PCBs and mercury, do notproduce eggshell-thinning by themselves in a raptor, a causal relationship exists between ,dietary DDE and eggshell-thinning under controlled conditions (Table 2) which resultsin a diagnostic correlative relationship between DDE in the resulting eggs and eggshell-thinning (Fig. 1), and this latter, experimentally produced relationship between kestrelegg DDE and eggshell-thinning parallels that observed in the field for the same species(Fig. 3). The relationship between eggshell-thinning in North American raptors and DDEin their eggs has now been established (Fig. 4) and is supported by experimental evidence,(Porter & Wiemeyer 1969; Wiemeyer & Porter 1970; this study, Figs 1, 2 and 3).

The documentation on raptor eggshell-thinning and population status is more thanample (Table 4). Pertinent to the logical conclusion of this discussion are three invariablefacts: (1) small decreases in eggshell thickness are correlated with predator species feedingprimarily on low trophic levels and/or resident northern prey; (2) more drastic decreasesin shell thickness are associated with predators preying on higher trophic level preyquite often associated with the aquatic habitat; (3) not one single North American raptor'population exhibiting 18% or more eggshell-thinning has been able to maintain a stable,self-perpetuating population. Consistent with these findings, a population analysis of,selected avian species (Henny 1972) revealed that the kestrel, along with the brown pelican '(Pelecanus occidenlalis L.), osprey (Pandion hal/oetus (L.)), Cooper's hawk (Accipitercooperii (Bonaparte) and red-shouldered hawk (Buteo lincatus (Gmelin)), has been shownto be producing fewer young than it did before DDT was made widely available. Despite ,the recent controversy, there can be little doubt now as to the causal relationship betweenthe global contaminant DDE and the observed eggshell-thinning and consequentpopulation declines in several birds of prey (Ratcliffe 1967, 1970; Mickey 1969; Bergeret al. 1970; Cade & Fyfe 1970). '*

ACKNOWLEDGMENTS

This research was supported by PHS grant ES-00306, T. J. Cade, Principal Investigator.Since we all stand on the shoulders of those who have gone before us, a reverent kind ofappreciation and respect is extended to Dr Cade. I should also like to extend thanks toR. Risebrough, D. Nelson and R. Heath for statistical assistance, to D. B. Peakall andW. Walker for reviewing the manuscript, and to L. Erdoesy for typing and editorialassistance. Computer facilities were provided by the Canadian Wildlife Service, Ottawa.This paper was prepared while the author held a William G. Selby and Marie SelbyFoundation Environmental Health Fellowship.

SUMMARY

(1) DDE residues in kestrel eggs collected from the Ithaca, New York area averaged35, 42, 33 and 37 ppm for the years 1969, 1970, 1971 and 1972, respectively.

792 Eggshell-thinning in wild and DDE-dosed American kestrels

(2) Based on Ratcliffe's Index, eggshells of the local population averaged 10% thinnerthan pre-DDT eggshells.

(3) A dose-response relationship is established for dietary DDE and eggshell-thinningin a captive kestrel population.

(4) Statistical analysis revealed that the correlative relationship between DDE in theegg and eggshell-thinning is the same for both captive experimental birds and the wildpopulation.

(5) A discussion of organochlorines, eggshell-thinning and the decline of severalpopulations of North American raptors concludes that a causal relationship existsbetween the ingestion of prey highly contaminated with DDE and the consequent egg-shell-thinning and eggshell breakage. The breeding failure that follows and subsequentpopulation declines of several raptor populations proceeds in a straightforward, logicaland well-documented sequence.

REFERENCESAnderson, D. W. & Hickey, J. J. (1972). Eggshell changes in certain North American birds. Proceedings of

the XV International Ornithological Congress, The Hague, 1970, pp. 514—40. E. J. Brill, Leiden.Berger, D., Anderson, D. W-, Weaver, J. & Risebrough, R. W. (1970). Shell thinning in eggs of Ungava

peregrines. Can. Fid. Nat. 84, 265-7.Blus, L. J., Gish, C. D., Belisle, A. A. & Prouty, R. M. (1972). Further analysis of the logarithmic relation-

ship of DDE residues to eggshell thinning. Nature, Lond. 240, 164-6.Cade, T. J. & Fyfe, R. (1970). The North American peregrine survey, 1970. Can. Fid Nat. 84, 231-45.Cade, T. J., Lincer, J. L., White, C. M., Roseneau, D. G. & Swartz, L. G. (1971). DDE residues and

eggshell changes in Alaskan falcons and hawks. Science, 172, 955-7.Dimond, J. B. & Sherburne, J. A. (1969). Persistence of DDT in wild populations of small mammals.

Nature, Lond. 221, 486-7.Ecobichon, D. J. & Saschenbrecker, P. W. (1968). Pharmacodynamic study of DDT in cockerels. Can.

J. Physiol. Pharm. 46(5), 785-94.Enderson, J. H. & Berger, D. D. (1970). Pesticides: Eggshell thinning and lowered production of young in

prairie falcons. BioSct'ence, 20, 355-6.Fyfe, R. W., Campbell, J., Hayson, B. & Hodson, K. (1969). Regional population declines and organo-

chlorine insecticides in Canadian prairie falcons. Can. Fid Nat, 83 (3), 191-200.Hays, W. L. & Winkler, R. L. (1970). Statistics: Probability, Influence and Decision, Vol. 2, pp. 245-9.

Holt, Rinehart & Winston, New York.Hazeltine, W. (1972), Disagreement on why brown pelican eggs are thin. Nature, Lond. 239, 410-11.Heath, R. G., Spann, J. W. & Kreitzer, J. F. (1969). Marked DDE impairment of mallard reproduction in

controlled studies. Nature, Lond. 224 (5214), 47-8.Henny, C. J. (1972). An analysis of the population dynamics of selected avian species; with special

reference to changes during the modern pesticide era. Wildl. Rex. Rep. 1, 1-99.Herman, S. G., Garrett, R. L. & Rudd, R. L. (1968). Pesticides and the western grebe. Chemical Fallout

(Ed. by M. S. Miller & G. G. Berg), pp. 24-53. Charles C. Thomas, Springfield, Illinois.Hickey, J. J. (Ed.) (1969). Peregrine Falcon Populations: Their Biology and Decline. University of Wis-

consin Press, Madison.Hickey, J. J. & Anderson, D. W. (1968). Chlorinated hydrocarbons and eggshell changes in raptorial and

fish-eating birds. Science, 162, 271-3.Hickey, J. J., Keith, J. A. & Coon, F. B. (1966). An exploration of pesticides in a Lake Michigan eco-

system. J. appl. Ecol. 3 (suppl.), 141-54.Kreitzer, J. F. (1972). The effect of embryonic development on the thickness of eggshells of coturnix

quail. Poult. Set. LI (5), 1764-5.Li, J. C. R. (1964). Statistical Influence, Vol. 1. Edwards, Ann Arbor, Michigan.Lincer, J. L. (1972). The effects of organochlorines on the American kestrel (Falco sparverius Linn.).

Doctoral dissertation, Cornell University, Ithaca, New York.Lincer, J. L. (1973). Poly chlorinated biphenyls: Their potential interference with pesticide residue analysis

and present analytical status. Chemical Analysis of the Environment and Other Modern Techniques(Ed. by S. Ahuja, E. M. Cohen, T. J. Kneip, J. L. Lambert & G. Zweig), pp. 109-31. Progress inAnalytical Chemistry, Vol. 5. Plenum Press, New York.

Lincer, J. L. & McDuffie, B. (1974). Heavy metal residues in the eggs of wild American kestrels (Falcosparverius Linn.). Bull. Envir. Contam. Toxicol. 12(2), 227-32.

thinner

binning

5 in thehe wild

several3 exists;nt egg-sequentlogical

edings ofc1

relation-

, 231-45.lues and

animals.

s. Can,

young in

organo-

>. 245-9.

10-11.jc*" 'n in

i special

/ Fallout

of Wis-

trial and

.jan eco-

^oturnix

; Linn.).

analysischniquesigress in

JEFFREY L. LINGER 793

Lincer, J. L. & Sherburne, J. A. (1974). Organochlorines in kestrel prey: A nor th south dichotomy. J.WiUtl. Mgml, 38(3), 427-34.

Longcore, J. R., Samson, F. B. & Whittendale, T. W., Jr (1971). DDE th ins eggshells and lowers repro-ductive success of captive black duck. Bull. Envir. Contam, Toxicol. 6(6), 485-90.

McLane, M. A. & Hall, L. C. (1972). DDE thins screech owl eggshells. Bull. Envir. Contam. Toxicol.8(2), 65-8.

Mueller, H. C. (1974). Factors influencing prey selection in the American kestrel. Auk, 91(4), 722-7.Myers, J. L. (1972). Fundamentals of Experimental Design, pp. 226-58. Al lyn & Bacon, Boston, Mass.Noakes, D. N. & Benfield, C. A. (1965). Tissue accumulation of DDT and its metabolites in the domestic

fowl. J. Sci. Fd Agric. 16, 693-7.Peakall, D. B, & Lincer, J. L. (1972). Methyl mercury: Its effect on eggshell thickness. Bull. Envir.

Contam. Toxicol. 8(2), 89-90.Porter, R. D. & Wiemeyer, S. N. (1969). Dieldrin and DDT; Effects on sparrow hawk eggshells and

reproduction. Science, 165, 199-200.Porter, R. D. & Wiemeyer, S. N. (1970). Propagation of captive American kestrels. J. WildL Mgmt,

34(3), 594-604.Ratcliffe, D. A. (1967). Decrease in eggshell weight in certain birds of prey. Nature, Land. 215, 208-10.Katctiffe, D. A. (1970). Changes attributable to pesticides in egg breakage frequency and eggshell thickness

in some British birds. J. appl. Ecol. 7, 67-115.Risebrough, R. W. (1972). Reply to Hazeltine. Nature, Lond. 240, 164.Snedecor, G. W. & Cochran, W. G. (1967). Statistical Methods, 6th edn. Iowa State Universi ty Press,

Ames.Snyder, D. AReinert, R.(1971). Rapid separation of polychlorinated biphenyls from DDT and analogues

on silica gel. Bull. Envir. Contam. Toxicol. 6, 385-90.Snyder, N. F. R., Snyder, H. A., Lincer, J. L. & Reynolds, R. T. (1973). Organochlorines, heavy metals,

and the biology of North American accipiters. BioScience, 23(5), 300-5.Switzer, B. C., Wolfe, F. H. & Lewin, V. (1972). Eggshell t h inn ing and DDE. Nature, Land. 240, 163.Temple, T. B. (1972). Chlorinated hydrocarbon residues and reproductive success in eastern North

American merlins. Condor, 74(1), 105-6.U.S.D.H.E.W. (1970). food and Drug Administration. Pesticide Analytical Manual, Vol. 1, Sect. 211 15

(revised April 1971). U.S. Government Printing Office, Washington D.C.Wiemeyer, S. N. & Porter, R. D. (1970). DDE thins eggshells of captive American kestrels. Nature,

Lond. 227, 737-8.Wiemeyer, S. N. & Porter, R. D. (1972). Eggshell thinning and DDE. Nature, Lond. 240, 162-3.Willoughby, E. J. & Cade, T. J. (1964). The breeding behavior of the American kestrel (sparrow hawk).

The Living Bird, pp. 75-96. Third Annual of the Cornell Laboratory of Ornithology, Cornell Univer-sity, New York.

(Received \5January 1975; revision received 20 March 1975)

s (Falco