TheCondor88:365-370 0 The Cooper Ornithological Society I986

GROWTH OF LESSER SNOW GEESE ON ARCTIC BREEDING GROUNDS

A. E. AUBIN GravenhurstHigh School,P.O. Box 370, Gravenhurst,ON POC IGO, Canada

E. H. DUNN AND C. D. MACINNES~ Ontario Ministry of Natural Resources,Wildlif ResearchSection, Maple, ON LOJ lE0, Canada

Abstract. Wild Snow Geese(Chen caerulescens)at the McConnell River, Northwest Territories, Canada, grew rapidly for their size, requiring only 30 days to grow from 10 to 90% of asymptotic weight (1,475 g). Fledging occurred at about 72% of adult weight. Culmen and tarsus were close to adult size at fledging, but primaries were only 65% grown. Fat and water indices resembled those of altricial . Captive geesediffered in growth patterns from wild birds, particularly in feather and tarsal growth and in water index, but showed the same pattern of caloric density. Snow Geese and other anatids grow more rapidly than other precocial speciesand do not appear to support Ricklefs’ hypothesisconcerning the inverse relationship between growth rate and propor- tion of functionally mature tissue at hatch. Key words: LesserSnow ;growth rates;anatids; body composition.

INTRODUCTION tarsus-from the epiphysisto the distal end of the tarsometatarsus, measured with vernier Despite much recent research on Lesser Snow calipers to the nearest 0.1 mm; primaries- Geese (Chen c. caerulescens),we know of no numbered 1 through 10 (proximal to distal), published data on the development of wild measured unflattened from tip to point of in- goslings.The few data for growth of geese of sertion to the nearest 1.0 mm with a ruler; any species are primarily for captives (Rob- body weight-weighed on spring scales, with erson and Francis 1965, Yocum and Harris birds < 1 kg weighed to the nearest 1.0 g, and 1966, Morehouse 1974, Wiirdinger 1975, those heavier weighed to the nearest 10.0 g. Ankney 1980). Such growth rates, however, Additional information was taken from col- can differ markedly from those of wild goslings lected birds: Gizzard-opened and cleaned or- (e.g., Wiirdinger 1975). We report here on gan was weighed on a top-loading electric bal- growth of wild Snow Geese at the McConnell anceto 0.1 g; intestine-portion extendingfrom River, Northwest Territories, Canada the gizzard to the cloaca was cleaned and (60”50’N, 94”25’W), giving comparative data weighedto 0.1 g; eye lens-whole eyeballswere from semicaptive goslingsraised at the same removed from all collected birds and stored in site. 10% formalin for 2 to 4 months, then lenses METHODS were removed and dried in an oven at 75°C for 24 to 40 hr. Four thousand wild goslingswere marked at All collected birds were weighed and were hatching in 197 1 with individually numbered measured externally. After removal of gizzard web tags. Growth was studied in web-tagged contents, the birds were chopped into smaller goslingswhich were reencounteredeither when pieces and ground in a grinder until ho- captured in massbanding drives or when shot mogenization was complete. Ground birds from towers located on the feeding grounds. were frozen and analyzed further after the field Data from only the first encounter were used, season,all within a few months. to avoid statistical problems of repeated mea- A subsample (usually at least 100 ml, and sures from the same individual, and because comprising at least 10% of the homogenate) growth may be slowed for a few days after was freeze-dried to determine water content. recapture (A. Dzubin, pers. comm.). Data were On average, 9% of live weight was lost during also collected from a few non-web-taggedgos- storage, and this was assumed to be water. lings, whose ageswere estimated from eye lens Lipid was extracted from at least two, and usu- weight. The following measurementswere tak- ally three or four aliquots of the dried sample en: culmen - the exposed culmen, measured (about 0.5 g each), usingthe Goldfisch method with vernier calipers to the nearest 0.1 mm; with petroleum ether asthe solvent. Live weight caloric content was estimated by converting I Received20 September198 5. Final acceptance2 1 April tissue composition to calories (9.00 kcal/g fat 1986. and 4.78 kcal/g lean dry; Ricklefs 1974). Oxy- 2 Reprint requeststo third author. gen-bomb calorimetry on about half the sam-

[3651 366 A. E. AUBIN, E. H. DUNN AND C. D. MAcINNES

parison to results of other growth studies (Ricklefs 1983). The log of each side of the equation was taken, to equalize residuals at each end of the growth curve and to reduce the effect of extreme values on estimation of the asymptote. Comparison of captive growth to that of wild birds was made by fitting polynomials to data for both groups combined, as above, but including a dummy variable for captivity status. Because repeated measurements were taken on captives, data were included only for those birds with at least 12 measures taken over at least 35 days (6 males and 6 females). 400 t ab Data for captive goslings were weighted

200 (weight = l/total cases for that bird) to give t % 00 d each individual equal representation, whether 1 / I / 1 I 1 I 0 5 10 15 20 25 10 15 40 45 wild or captive. The equations given in this AGE(days) paper were fit to data for wild birds alone. FIGURE 1. Weights of individual wild Snow Goosegos- RESULTS lings. Circles indicate males, triangles females, and aster- isks unknown sex. Open symbols show birds whose age Snow Geese developed rapidly, with scapulars was estimated from eye lens weight. Values for hatchlings first showing at 10 to 12 days and primaries are means for 9 males and 9 females. at 18 days. Most fledging in 197 1 took place only 10 days before the start of migration. A ples gave a value of 4.93 kcal/g lean dry (n = logistic curve fit to weights of wild goslings 42). The micro-Kjeldahl method was used to (Fig. 1) gave a t,O_so(days between 10% and measurenitrogen content (Schmidt 196 1). The 90% of asymptotic weight, 1478 g) of 30 days samemethods were usedin analyzingyolk from and a K (growth rate constant) of 0.15 1. The 68 fresh eggsand 9 yolk sacsremoved from sexesdid not differ in the course of weight gain hours-old hatchlings. (P > O.OS),although adult males outweighed Thirty-five goslingstaken from nestsduring adult females, and just-hatched males weighed hatching were held for a week in cardboard more than females (83.4 g for 9 males vs. 77;4 cartons placed near an oil burner. Dog food g for 9 females, t-test P < 0.01). and water were supplied freely. After a week, The equation for culmen on agein wild birds the goslingswere placed in a 0.6-ha enclosure (Fig. 2A) was Y,, = 17.542 + 0.862 age - made of 95cm high fence with 2.5 0.0001 age3(r* = 0.98, SE ofregression = 1.58, cm mesh, placed around a traditional goose n = 112). Average length of primaries 1 to 10 feeding area. Although the goslingsfed mainly of wild goslings(Fig. 2C) was linear with re- on vegetation, largely Carex spp., they were spect to age (Y,, = -149.419 + 6.993 age; provided with a supplement of dog food (Mir- SE = 7.60, it = 70). The equation for eye lens acle brand, minimum 25% protein), given at growth (Fig. 2D) was Y,, = 36.130 + 1.577 the rate of 0.09 g per g of gooseper day. Water age + 0.120 age* - 0.002 age3(r* = 0.98, SE = was freely available from a stream flowing 7.33, n = 63). Sexes differed only in tarsus through the area. All birds were measuredtwice length (Fig. 2B, P < 0.05), with males growing a week as described above. A sample of these faster (Y,, = 30.126 + 2.063 age - 0.001 age3 (usually four) was collected weekly for other for females, Y,, = 3 1.892 + 2.063 age - analyses(see above). 0.001 age3for males (r2 = 0.98, SE = 2.82, y1= Equations describing weight and measure- 114). Sexesare combined in Figure 2B because ments of wild birds during growth were deter- of the smallnessof difference between them. mined by fitting polynomials with linear Growth of the gizzard and intestine in wild regression(SPSS Statistical Package),allowing birds was examined relative to overall body stepwiseentry (after inclusion of age) of age*, growth. A regression of log organ weight on age3and a dummy variable for sex. This meth- log body weight shows whether organ growth od does not impose a curve of any particular is relatively more rapid (slope of regression, form onto the data and offers a readily ob- b > 1.0) or slower (b < 1.0; Ricklefs 1983). tained, statistically meaningful indication of Gizzards grew rapidly in the first week after the effect of sex. Weight data were further fit hatching (b = 1.3 l), then more slowly (b = to a logistic curve (SAS NLIN program) to 0.89 for days 8 to 35). After 35 days, actual obtain standardparameters (K, t,,_,,) for com- gizzard weight appeared to decrease(Fig. 3A). SNOW GOOSE GROWTH 361

A CULMEN

C PRIMARIES 200

01 IO 20 30 40 50

AGE (days1 AGE I days) FIGURE 2. Mean culmen (2A), tarsus (2B), primary length (2C), and eye lens weights (2D) of known age goslings, sexescombined. Values for primaries are average length of Primaries 1 to 10. Solid symbols = wild goslings;open symbols = captives.

Growth of the intestine (Fig. 3B) resembled A GIZZARD that of the gizzard (b under 8 days was 1.45, 160 and 0.85 from 8 to 35 days). I *- f Water index (Fig. 4A) decreasedlinearly af- ter the first week. Lipid indices were extremely low (Fig. 4B), except at hatch when yolk sacs ranged from 5 to 15% of fresh body weight. Even without yolk sacs, the lipid index for hatchlingswas high. This reflected extra body fat reserves that were, like the yolk sac, used up within a few days. Caloric content of the live bird increased linearly from 0.8 to 1.8 kcal/g from time of yolk sac absorption to B INTESTINE fledging (Fig. 5). Water and fat index and ca- 80c . loric density were not affected by sex (P < 0.05). Nitrogen content was measured to see if it would provide a good index of physical con- dition, as it is related to the amount of protein in the body. Multiple regressionshowed that log nitrogen content was directly proportional to log lean dry weight (n = 42, r2 = 0.68, b = 0.94, P < 0.001) with no additional effects of age, sex, feather length, or log fat content. It FIGURE 3. Gizzard (3A) and intestine weight (3B) of therefore provided no more information on individual wild goslings(sexes combined, solid symbols). condition than did less-tediouslyobtained val- Means for captive birds at each age are shown for com- ues such as weight. parison (open symbols). 368 A. E. AUBIN, E. H. DUNN AND C. D. MAcINNES

A WATER INDEX

50-

06. B- LIPID INDEX .

It 5 04- ;

0 * z- 0 0 iii oz- 00 w t I 8 o :. , o . l. . 4q;;,.u . “g ,**o 00

0 IO 20 30 40 50 AGE (days) AGE (days) FIGURE 5. Live weight caloric content of individual FIGURE 4. Water (4A) and lipid (4B) indices of indi- known age goslings, calculated from composition (see vidual known-agegoslings (sexes combined). Index is water Methods). Symbols as in Figure 4. or lipid content divided by lean dry weight. Solid sym- bols = wild goslings;open symbols = captives. Asterisks denote lipid indices of hatchlingswith yolk sac removed. ney (1980) and other results are summarized in Table 1. Growth of captive goslingswas the same as for wild birds only in weight gain (to the age DISCUSSION of fledging) and in eye lens growth (see Meth- Weight gain of wild Snow Geese was as rapid ods). Captives continued to gain weight after as for altricial speciesof the same size (Fig. 6). the age of fledging (42 days), and it is assumed However, data were not available after the age that wild birds did also. Culmen, tarsus, in- of fledging. Because many anatids gain more testine, and gizzard growth laggedsignificantly weight after fledgingand all other growth curves in captives for the first month (Figs. 2 and 3), for anatids but one were fit to data from cap- but reached final sizes similar to those of wild tives that could be measured beyond the age goslings.In captive birds, the decreasein giz- offledging (Ricklefs 1973) t,o-gofor Snow Geese zard weight late in growth was clearer. Primary may have been underestimated.Nearly all oth- growth of captives (Fig. 3C) laggedconsistently er precocial specieswere also studied in cap- behind that of wild birds. tivity (Ricklefs 1973) however, and it should Captive birds had higher water content than be safe to conclude from Figure 6 that anatids wild birds (Fig. 4A, P < O.OS),but the rate of grow more rapidly than other precocial birds. dehydration (excluding hatchlings)did not dif- By the age of fledging, Snow Goose body fer. Lipid indices were higher (P < 0.05) after weight was about 70 to 75% of that of adults, the first week (Fig. 4B), perhaps as a result of and primaries were about 65% grown (Table the supplemented diet. Caloric density, how- 2). These proportions are at the low extreme ever, was the same in captives as in wild gos- of the rangesgiven in Owen (1980) for several lings (Fig. 5). goose species(75 to 96% and 70 to 85%, re- Composition of eggsand yolk sacswere in- spectively). vestigated(Table 1) to determine whether rap- Canvasback (Aythyu valisineria) and Lesser id development in wild Snow Geese might re- Scaup (A. afinis) showed patterns of gizzard sult from particularly large energy reserves at and intestine growth similar to those of the hatching. Yolk was larger in larger (P < Snow Goose (Lightbody and Ankney 1984). A 0.001, b = 0.3 13, r2 = 0.24) as found by Ank- decreasein gizzard weight in Red-throated Bee- SNOW GOOSE GROWTH 369

TABLE 1. Composition and energy in fresh yolks and hatchling yolk sacsof Snow Geese. Sample sizes are in parentheses.

Egg yolk Yolk sac

Fresh egg or hatchling wt. (g) 124.3 (68) 80.9 (10) Yolk or yolk sac wt. (g) 47.0 (68) 8.3 (10) Yolk or yolk sac as % egg or hatchling wt. 37.8 10.2 % water of fresh wt. 47.1 (68) 48.0 (11) % lipid 29.5 (6) 25.3 (11) kcal/g fresh wt. 4.06 3.80

eaters (Merops bulocki) has been attributed to I I I I a switch in food source late in growth (Fry 500 1000 2000 4000 6000 1972), but this cannot be the explanation in ASYMPTOTIC WEIGHT (9) waterfowl. Decreasesin internal organ weights FIGURE 6. Days between 10% and 90% of weight gain late in growth may instead facilitate fledging. from hatch to asymptotic weight in anatids. Solid regres- Gizzard and intestine development in water- sion line is for altricial speciesand dashed line for other are more similar to those of seven altricial precocialbirds (Tetraonidae,Phasianidae, Meleagridae and speciesthan to those of the few semiprecocial Rallidae; Ricklefs 1973). Solid symbols for data from and precocial speciesstudied, e.g., Herring Gull Ricklefs (1973), open symbols from Lightbody and Ank- ney (1984), and asteriskfor Snow Goose (this paper). Lams argentatus and domestic fowl Gallus gallus (Dunn 1975), in that digestive organs in the latter grow more slowly than the body velopment. The values for live weight caloric as a whole throughout the post-hatch period. density of wild Snow Geese (Fig. 5) are very If water index of wild goslingsin the first 10 similar to thoseof altricial species(Dunn 1975), days after hatching is close to that of captive mainly becauseof low lipid accumulation. ones (Fig. 4A), then Snow Geese adhere to the Caloric value for fresh egg yolk (Table 1) pattern seen in Wood Aix sponsa(Clay was similar to those for three domestic galli- et al. 1979) and Herring Gull (O’Connor 1984). form species,but lower than in domestic wa- In these,water index first increasesafter hatch, terfowl (4.39 kcal/g in domesticduck and goose, instead of decreasing steadily throughout Ricklefs 1974). Hatchling yolk sacsaveraged growth as is typical of other precocial species. 11.4% of yolk-free body weight; within the Although high body fat at hatching has been range found both in galliform and anseriform found in other precocialspecies (Ricklefs 1974), species(10 to 25%, Ricklefs 1974). Thus, rapid the low lipid level seen in wild Snow Geese growth of goslingsdid not appear to be con- later on (Fig. 4B) is unusual. Indices reached nected to qualities of the egg. values of 0.30 and higher in some species(Bris- Except for the eye lens, measurements of bin and Talley 1973, Clay et al. 1979) but may captive birds differed enough from those of have been artifacts of those birds being raised wild birds that they could not be used to pre- in captivity on an artificial diet. Alternatively, dict agesof wild goslings.In our captives, early the low lipid storagein Snow Geese could re- growth tended to lag behind that of wild birds, flect high allocation of energy to growth in an but some conditions of captivity can produce environment in which food (grassesand sedges) unusually rapid development. For example, is in relatively uniform supply throughout de- captive Snow Geese raised at the McConnell

TABLE 2. Size of wild Snow Goose goslingsat fledging (taken as 42 days) with respect to adult size. Gosling data from equations fit to data in Figures 1 and 2 (see text). Sample sizes for adults in parentheses.

Weight Culmen TXSUS Primary Age Sex w (mm) (mm) (mm) Fledgling M 1,475 46 81 144 F 1,475 46 80 144 Adultb M 2,186 (207) 57.3 (35) 83.9 (35) 235 (20) F 1,912 (243) 55.4 (49) 79.9 (49) 223 (38) Fledgling as % of adult M 67 97 61 F 77 :: 100 65

’ ’ Average length of Primaries I to IO. b Data for McConnell River adults in late summer 1970 (Lieff 1973; Maclnnes, unpubl.). 370 A. E. AUBIN, E. H. DUNN AND C. D. MAcINNES

River by Ankney (1980) reached weights at throughout is gratefully acknowledged. This is Ontario fledging (42 days) averaging 750 g more than Ministry of Natural Resources.Wildlife ResearchSection Contribution No. 84-05. our wild goslingsat the same age, and 170 g more than adults (Table 2); tarsuslengths were 2.5 mm larger than in adults. Retardation of LITERATURE CITED feather growth in captives (Fig. 2C) appearsto ANKNEY,C. D. 1980. Egg weight, survival, and growth be common, having been casually reported for of LesserSnow Goose goslings.J. Wildl. Manage. 44: geese(Hanson and Jones 1968) and well-doc- 174-182. BRISBIN,I. L., JR., AND L. J. TALLY. 1973. Age-specific umented in Ring-necked PheasantsPhasianus changesin the major body components and caloric colchicus(Etter et al. 1970). value of growing JapaneseQuail. Auk 90:624-635. Ricklefs (1979a, 1983) hypothesized that a CLAY. D. L.. I. L. BRISBIN.JR.. AND K. A. YOUNGSTROM. high proportion of maturely functioning tissue <979. Age-specificciang& in the major body com- ponents and caloric value of growing Wood . at hatch should lead to slow growth. In species Auk 961296-305. studied to date, the parts of the body that func- DUNN, E. H. 1975. Growth, body components and en- tion early (particularly the legsin precocial and ergy content of nestling Double-crested Cormorants. semiprecocial species)grow more slowly than Condor 77143l-438. the rest of the body. Because overall growth ETTER,S. L., J. E. WARNOCK,AND G. B. JOSELYN.1970. Modified wing molt criterion for estimating the ages rate should depend most on growth of body ofwild iuvenile Dheasants.J. Wildl. Manage. 34:620- parts composing a large proportion of total 626. - - mass(e.g., legs),Ricklefs (1979b) suggestedan FRY, C. H. 1972. The biology of African Bee-eaters. inverse correlation of growth rate with the rel- Living Bird 11:75-l 12. HANSON,fi. C., AND R. L. JONES. 1968. Use of feather ative size of legs in adults. Legs (including minerals as biological tracers to determine breeding bones) make up 16.5% of adult JapaneseQuail and molting groundsof wild geese.Ill. Nat. Hist. Surv. (Coturnixjaponica), which grow slowly, but Biol. Notes 60:3-g. only 4.5% of adults in the fast-growing Com- LIEFF, B. C. 1973. The summer feeding ecology of Blue mon Tern (Sternahirundo). Ducks, with legs and Canada geeseat the McConnell River, N.W.T. Ph.D.diss., Univ. Western Ontario, London, ON, making up 5 to 9% of adult body weight, have Canada. intermediate growth rates (Ricklefs 1979b, Fig. LIGHTBODY,J. P., AND C. D. ANKNEY. 1984. Seasonal 6). Geese are terrestrial and cursorial and influenceon the strategiesof growth and development should have proportionately larger leg mass of Canvasbackand LesserScaup ducklings. Auk 101: 121-133. than ducks. In Cackling Geese (Branta can- MOREHOUSE,K. A. 1974. Development, energeticsand adensis),leg muscle alone constitutes7 to 14% nutrition ofcaptive Pacific Brant (Branta berniclaori- of adult weight, dependingon season(Raveling entalis, Tougarinov). Ph.D.diss., Univ. of Alaska, 1979). Assumingproportional leg size of Snow Fairbanks. Geese is similar to that of Cackling Geese, O'CONNOR, R. J. 1984. The growth and development of birds. John Wiley & Sons, New York. Ricklefs’ hypothesis would predict an inter- OWEN,M. 1980. Wildgeese ofthe world: their life history mediate to slow growth rate for the Snow and ecology. B. T. Botsford Ltd., London, England. Goose. When the growth rate of Snow Geese RAVELING,D. G. 1979. The annual cycle of body com- (0.15 1) is scaled to the body size of quail and position of Canada Geese with special reference to control of reproduction. Auk 96:234-252. terns, however (K2 = K,(W2/W1)p0.28,Ricklefs RICKLEFS.R. E. 1973. Patterns of growth in birds. II. 1979b), it is as high (at 0.300) as that of the Gro\;th rate and mode of development. Ibis 115:17 I- tern. Growth patterns of geesecould still sup- 201. port Ricklefs’ hypothesis if a large proportion RICKLEFS,R. E. 1974. Energeticsof reproductionin birds. of leg growth is completed before hatch, but Publ. Nuttall Omithol. Club 15:152-292. RICKLEFS,R. E. 1979a. Adaptation, constraint,and com- goslingsare so small relative to adults that this promise in avian postnatal development. Biol. Rev. seemsunlikely. Geese should be studied more Camb. Philos. Sot. 54:269-290. closely with the objective of investigating RICKLEFS,R. E. 1979b. Patterns of growth in birds. 5. Ricklefs’ ideas. A comparative study of development in the Starling, Common Tern and JapaneseQuail. Auk 96: 10-30. ACKNOWLEDGMENTS RICKLEFS,R. E. 1983. A