The Condor 86:7-l 1 0 The CooperOrnithological Society 1984

EGG DIMENSIONS AND NEONATAL MASS OF SHOREBIRDS

ROBERT E. RICKLEFS

ABSTRACT. -Eggs of 10 speciesof shorebirds(Charadriidae, Scolopacidae,Phal- aropodinae) collected at Churchill, Manitoba, were measured and artificially in- cubated. Neonates were weighed within 6 h of hatching. This paper reports vari- ation in egg size and neonate mass among clutches,and the relationship between egg dimensions and neonate mass. Coefficients of variation were 4 to 9% for egg volume and 6 to 15% for neonate mass. In most species,a significant proportion (48-84%) of the variance in eggvolume could be attributed to differencesbetween clutches.Neonate mass was significantly related to egg size in six of nine species.

Coefficients of variation in the volume of eggs Whimbrel (Numeniusphaeopus); Hudsonian within populations of shorebirds (Charadri- Godwit (Limosa haemastica);Lesser Yellow- idae, Scolopacidae,and their allies) are on the legs(Tringaflavipes); Stilt Sandpiper (Calidris order of 4 to 7% (Vaisanen et al. 1972, Vais- himantopus);Short-billed Dowitcher (Lim- anen 1977). Much of this variation arisesfrom nodromusgriseus); Dunlin (Calidris alpina); differencesin the sizesof eggslaid by different Least Sandpiper (C. minutilla); Red-necked females, which are detected as between-clutch Phalarope(Phalaropus Zobatus). These were the variation (e.g., Vaisanen et al. 1972, Miller speciesmost readily available in the study area 1979, Baker and Cadman 1980). The relation- and, fortuitously, represent much of the vari- ship between the sizesof eggsand neonateshas ation in size and habits among arctic shore- been described for many speciesof birds, in- birds. cluding some members of Charadriiformes Clutches were collected at various stagesof (e.g., Ricklefs et al. 1978, Lundberg and Vais- incubation from wet and dry tundra habitats anen 1979, Lloyd 1979), but not for shore- to the east of Churchill, Manitoba (50”N, birds. Because shorebird chicks must feed 94”W), during mid- and late June 1979. Eggs themselvesshortly after hatching, often in cold were marked with indelible ink and their environments, larger size, more advanced de- lengths(L) and breadths (B) were measuredto velopmental state, and greater energy reserves the nearest 0.1 mm with dial calipers. I used of neonates presumably carry a selective pre- the product of the length and the square of mium, all else being equal. Indeed, the relative breadth (LB2)as an index to volume (Vaisanen egg size (percent adult mass) of shorebirds is 1969, Vaisanen et al. 1972). Regressionsof large compared to most other groups of birds volume on LB2 had coefficients of determi- (Rahn et al. 1975). If egg size is under strong nation of between 88 and 98% in eight species selection (i.e., most unfavorable genes have of waders studied by Vaisanen (1977). Eggs been removed), variation within populations were incubated in a small incubator (Hova- must be either nongenetic in origin or com- bator, Inc.) at 37°C and approximately 60% pensatedby variation in other, balancingtraits relative humidity. Neonates were individually that are correlated with egg size (e.g., Rose marked and weighed soon after hatching, usu- 1982). The large variance between clutches ally within 6 h. I could match neonatesto their found in most populations suggestsgenetic eggsunambiguously in most cases.Where there variation in egg size, although age and devel- was doubt, the data were omitted from the opmental effects have not been ruled out. pertinent analyses. As a step towards understanding the evo- Additional data on lengths and breadths of lution of large egg size in shorebirds, and in- eggs in 5 clutches of Hudsonian Godwits, 8 terpreting variation in egg size among birds in clutches of Whimbrels, 7 clutches of Short- general, I report here on variation in egg size billed Dowitchers, 24 clutches of Least Sand- and neonate massamong clutches,and on the pipers, and 33 clutchesof Stilt Sandpipers(all relationshipbetween egg size and neonate mass from the Churchill area) were generously pro- in 10 speciesof shorebirds at Churchill, Man- vided by J. R. Jehl. I analyzed these measure- itoba. mentsseparately, rather than poolingthem with my data, because neonatal masses were not MATERIALS AND METHODS available. I used the SAS statistical package Ten specieswere included in this study: Lesser (Helwig and Council 1979), primarily the Golden-Plover (Pluvialisdominica); Semipal- MEANS and GLM procedures,to analyze the mated Plover (Charadrius semipalmatus); data. Univariate statistics presented are the [71 8 ROBERT E. RICKLEFS mean (x), standard deviation (SD), standard Whimbrel (4.52), Hudsonian Godwit (2.62) error of the mean (SE), and coefficient of vari- Dunlin (2.44), and Least Sandpiper (2.15). ation (CV). The statistical significanceof dif- Analyses of variance in egg volume indices ferences between the CVs was based upon a and neonate masseswithin and among clutch- value of F calculated as the square of the ratio es are also presented in Table 1. Much of the of the larger to the smaller (Lewontin 1966). variation in egg volume index and neonate The statistical significance of a range of CVs masswas related to differencesin the averages was based on the F,,, test (Sokal and Rohlf of eggs and neonates between clutches. The 1969:37 1). Variance among clutchesin eggdi- massof the neonate varied significantlyamong mensions and neonate mass was determined clutches in five species(R*, 50-89%) but not by a one-way analysis of variance (ANOVA) in the Whimbrel, Lesser Yellowlegs, Least with clutch as the main effect. Sandpiper, and Northern Phalarope (R* < Relationshipsbetween neonate massand egg 50%). The coefficient of determination among volume were determined by linear regression clutches in neonate mass was not clearly re- of both original measurementsand log-trans- lated to values of R2 for the volume index of formed values. The slope of the logarithmic the egg. regression(b) defines the allometric relation- In most speciesof birds, large eggsproduce ship between measurements Y and 1, accord- largechicks, and shorebirdsare consistentwith ing to the expression Y = aXb. Values of b this general pattern. I calculated the regression significantlygreater than or lessthan 1 indicate of the log of neonate mass against the log of departure from linearity or direct proportion- LB2 to determine the relationship between ality between increase in neonate mass and neonate size and egg size (Table 2). The rela- increasein eggvolume. The slope of the arith- tionship was significant in all species except metic regression(b) in the expression Y = a + the Lesser Yellowlegs, Least Sandpiper, and bX representsthe grams of increasein neonate Northern Phalarope. For the significant rela- massper cubic centimeter increasein eggvol- tionships, R2 varied between 2 1 and 75%. The ume index. Regressionswere calculated over slopes(b) of the significant regressionsvaried the sample as a whole for each species, al- between 0.69 (0.30 SE) in the Dunlin and 1.45 though the similarity of eggdimensions within (0.27 SE) in the Whimbrel. Except for the Dun- clutches suggeststhat measurements of indi- lin, all the slopesexceeded 1, but none signif- vidual eggsare not independently distributed icantly so (t-test, where t = [b - 11/s, with y1 and therefore that sample sizes used for sta- - 2 degreesof freedom). Although within some tistical inference may be inflated. Regressions specieslarge eggsmay produce disporportion- also were calculated within clutches, using an ately large chicks (b > l), my samplesof neo- analysisof covariance in which the effectswere nates were too small to quantify the relation- clutch and LB*. By entering clutch as the first ship with much precision. effect, differences between clutches are re- Because the relationship between neonate moved and the regressionof mass upon vol- massand eggvolume index may result in part ume expressestheir relationship within clutch- from differences among clutches, I performed es. a second analysis of variance with clutch as a main effect in order to determine the relation- RESULTS ship of neonate size to eggsize within clutches. Univariate statisticsfor the egg volume index The logarithmic regressionswere significant (LB2) and neonate massare presentedin Table only for the Whimbrel (F[1,12] = 9.42, P = 1. Coefficients of variation in volume index 0.01, b = 1.58 + 0.5 1 SE) and Stilt Sandpiper differed significantlyamong speciesand ranged (F[1,15] = 20.49, P < 0.001, b = 0.89 * 0.20 from 4.2 to 9.1% (F,,,,, = 4.6, P < 0.05). CV’s SE), although they were close to being signif- of neonate massalso differed significantly,with icant for the Semipalmated Plover (P = 0.068) arangeof5.9 to 15.1%(F,,,= 6.6, P < 0.01). and Northern Phalarope (P = 0.056). Signifi- The largest CVs belonged to the two largest cant arithmetic regressions were: Semipal- species, Whimbrel (15.1%) and Hudsonian mated Plover (F[1,15] = 4.83, P = 0.044, b = Godwit (12.0%). Among the remaining species 0.43 * 0.20 SE); Whimbrel (F[1,12] = 6.69, (dowitcher excluded), CVs did not differ sig- P = 0.024, b = 0.59 + 0.23); Stilt Sandpiper nificantly. (fll,15] = 21.61, P < 0.001, b = 0.31 * 0.07); In all speciesexcept the Lesser Yellowlegs, and Northern Phalarope (F[ 1,121 = 4.54, P = the CV among neonate massesexceeded that 0.055, b = 0.31 * 0.15). among egg volume indices; the squaresof the ratios of CVs (F-values) significantly exceeded DISCUSSION 1 (P < 0.05) in five of nine species(dowitcher Samples of eggs collected in this study had excluded): Semipalmated Plover (3.47) average volumes and coefficients of variation EGG DIMENSIONS AND NEONATAL MASS 9

TABLE 1. Statisticaldistributions of sizesof eggsand neonates,and analysisof variation within and between clutches, of shorebirdsat Churchill, Manitoba.

Statlstlcs~ ANOVk Species and lll~~S”T~lll.E”t* n 1 SD cv df P R ’

LesserGolden-Plover LB2 24 51.79 3.35 6.47 5,18

*LB’ in cubic centimeters; mass in grams. b n = number of eggs in sample, SD = standard deviatmn, CV = coetlicient of vanation. r df = degrees of freedom (clutches, eggs withm clutches), P = probablbty of F 2 I. R2 = coetlicmt of determmatmn. d Measurements provided by J. R. Jehl. in volume similar to the samplescollected from (Spearman rank correlation coefficient r, = the same area by: Jehl and Smith (1970) for 0.52, P > 0.05). Within a sample, increased the Whimbrel, Hudsonian Godwit, Short- variation in neonate massover that of eggvol- billed Dowitcher, and Least Sandpiper; Jehl ume could be causedby several factors. First, (1973) for the Stilt Sandpiper;and Hagar (1966) not all chicks were weighed at the same length for the Hudsonian Godwit. Jehl (1973) re- of time after hatching. All were fully dried out ported the average mass of nine chicks of the and their down fluffed up before they were Stilt Sandpiper less than 24 h old as 8.1 g, weighed, but some, especially those which which is very close to the average of 8.15 g hatched overnight, remained up to 6 h longer obtained for 22 incubator-hatched chicks in in the incubator, during which period they lost this study. The average massof 16 Hudsonian an unknown amount of mass.Second, neonate Godwit chicks hatched in this study (X = 23.0 mass may be determined by yolk size, rather g) was similar to the value of 23.4 g for 10 than total egg contents, which may be more chicks weighed by Jehl and Smith (1970). variable than eggvolume. Third, embryos may Variability in the massof neonateswas gen- differ in the manner in which they utilize egg erally greater than the variability among egg contents, perhaps hatching with more or less volumes; CVs of the two measureswere pos- yolk, more or lessaccumulated body fat, or at itively correlated, but not significantly so a greater or lesser stage of development (see, 10 ROBERT E. RICKLEFS

TABLE 2. Relationship between neonatal mass and egg (1972) were position in laying sequence, se- volume index in shorebirdsat Churchill, Manitoba. quence of clutch within years, and year. Few studies have related egg dimensions to Speciesand type Statisti& of regression’ df F P R ’ b % the characteristicsof the female parent. Vais- anen et al. (1972) found small but significant LesserGolden-Plover correlations between egg volume and female Logarithmic 1,12 5.73 0.034 0.32 1.01 0.42 massin the Dunlin (R2 = 0.08) Common Red- 0.023 0.36 0.33 0.13 Arithmetic 1,12 6.79 shank (Tringa totanus;0.07), and the Com- Semipalmated Plover mon Ringed Plover (Charadrius hiaticula; Logarithmic 1,20 24.01 Stint (Calidristem- Arithmetic 1,20 27.61

adriiforms other than shorebirds.Lloyd (1979) ACKNOWLEDGMENTS determined that, in the Razorbill (Alca to&u), The support and hospitality of the Churchill Northern neonate mass and egg volume were related, StudiesCentre, especiallythat ofits Director, W. Erickson, with coefficients of determination of 29 and are nratefullv acknowledged.K. Baker. L. Clark. and D. 66% in two years (n = 45, 25, P < 0.001). Gol&tein assistedin the%eld. D. Brotman perfoimed the Lundberg and V&s&en (1979) found a similar computer analyses. Computer time was provided by a grant from the Faculty of Arts and Sciences,University relationshipin the Common Black-headedGull of Pennsylvania.J. C. Coulson,J. R. Jehl, R. A. VBisBnen, (Lams ridibundus;R2 = 0.59).The slope ofthe and J. B. Williams provided helpful comments on drafts linear regression between neonate mass and of the manuscript. Jehl kindly supplied his original mea- eggvolume was 0.83 g.cm-3. In the Laughing surementsof sandpipereggs. The study was supportedby Gull, linear relationshipsbetween neonate mass National ScienceFoundation grant DPP79-0857 1. and egg mass had slopes of 0.92 g.g-l (yolk sacincluded) or 0.72 (yolk sacexcluded; Rick- LITERATURE CITED lefs et al. 1978). BAKER,A. J., AND M. CADMAN. 1980. Breedingschedule, The information discussedabove yields the clutch size and egg size of American Oystercatchers following conclusionsconcerning egg variation Haematopuspalliatus. Wader Study Group Bull. 30: in shorebirds. The standard deviation in egg 32-33. volume is approximately 6% of the mean, per- HAGAR,J. A. 1966. Nesting of the Hudsonian Godwit hapswith some differencesamong species.Ap- at Churchill, Manitoba. Living Bird 5:5-43. HELWIG.J. T.. AND K. A. COUNCIL. 1979. SAS users’ proximately 50 to 80% of the variation in egg guide. SAS Institute, Gary, NC. volume may be attributed to differencesamong HOYT, D. F. 1979. Practical methods of estimating vol- females and probably hasa stronggenetic com- ume and fresh weight of bird eggs.Auk 96:73-77. ponent, although relatively little of this vari- INNES,J. L. 1980. Wader eggs,and growth ratesof wader chicks,p. 80-95. In J. L. Innes [ed.], Cambridge Nor- ation (<20%) is related to measurements of wegian Expedition 1978 Report. Cambridge Norwe- female size. gian Expedition, Cambridge. England. The percentageof variation in neonate mass JEHL~J. R., Jo. 1973. Breed& biology and systematic distributed between clutchesvaries widely be- relationshipsof the Stilt Sandpiper. Wilson Bull. 85: tween about 30 and 90%. The within-clutch 115-147. JEHL,J. R., JR., AND B. A. SMITH. 1970. Birds of the variation ranged between CVs of 4.6% and Churchill region, Manitoba. Spec. Publ. No. 1, Man- 13.5%, with six of the nine speciesin this study itoba Museum of Man and Nature, Winnipeg. falling between 6.2 and 8.5%. The low ex- LEWONTIN,R. C. 1966. On the measurementof relative tremes were the Hudsonian Godwit, in which variability. Syst. Zool. 14:141-142. LLOYD,C. S. 1979. Factors affecting breeding of Razor- most neonate variation was distributed among bills Alca tordu on Skokholm. Ibis 121:165-176. clutches, and the Stilt Sandpiper, which had LUNDBERG,C.-A., AND R. A. V;~IS.&NEN.1979. Selective low variability in neonate mass in the sample correlation of egg size with chick mortality in the as a whole. Most of the variation in neonate Black-headedGull (Larus ridibundus).Condor 8 1:146- 156. massin the extremely high species,the Whim- MILLER, E. H. 1979. Egg size in the Least Sandpiper brel, occurred within clutches. Neonate mass Calidrisminutilla on SableIsland, Nova Scotia.Omis is directly related to egg size, but values of R2 Stand. lO:lO-16. vary widely between perhaps20 and 75%. The RAHN, H., C. V. PAGANELLI,AND A. AR. 1975. Relation variation in quality of the chick related to mass ofavian eggweight to body weight. Auk 921750-765. RICKLEFS,R. E., C. D. HAHN, AND W. MONTEVECCHI.1978. is unknown as is the effect of this variation on The relationship between egg size and chick size in the survival and future fecundity of the indi- the LaughingGull and JapaneseQuail. Auk 95: 135- vidual. 144. This study supplements previous work in ROSE, M. R. 1980. Antagonistic pleiotropy, dominance, documenting levels of variation in eggsize and and genetic variation. Heredity 48:63-78. SOKAL,R. R., AND F. J. ROHLF. 1969. Biometry. W. H. neonate mass within populations of shore- Freeman, San Francisco. birds, confirming that a large part of the vari- WISANEN, R. A. 1969. Evolution of the Ringed Plover ation is associated with differences between (Charadriushiaticula L.) during the last hundred years clutches and, probably, between females, and in Europe. A new computer method based on egg dimensions. Ann. Acad. Sci. Fenn. A IV 149:1-90. demonstrating that neonate mass is related to VAFANEN, R. A. 1977. Geoarauhic variation in timine egg size both within the samples and within of breeding and egg size in eight European speciesoyf clutches. Further work is required to deter- waders. Ann. Zool. Fenn. 14:l-25. mine whether speciesdiffer in overall egg-size VXIS;~NEN, R. A., 0. HILDEN, M. SOIKKELI,AND S. VUOLANTO. 1972. Egg dimension variation in five variability, apportionment of variability with- wader species: the role of heredity. Omis Fenn. 49: in and between clutches, and the relationship 25-44. of eggsize to neonate massand quality. Studies are also needed to determine the consequence of egg size, hence neonate size, for postnatal Departmentof Biology, Universityof Pennsylvania,Phila- growth and survival of both the chicks and the delphia,Pennsylvania 19104. Received 11 June 1982. Fi- female parent. nal acceptance13 June 1983.