Foraging time and dietary intake by breeding Ross s and Lesser Snow Geese

Oecologia (2001) 127:78 86 DOI 10.1007/s004420000577 Mark L. Gloutney Ray T. Alisauskas Alan D. Afton Stuart M. Slattery Foraging time and dietary intake by breeding Ross s and Lesser Snow Geese Received: 5 September 1997 / Accepted: 16 October 2000 / Published online: 19 January 2001 Springer-Verlag 2001 Abstract We compared foraging times of female Ross s (Chen rossii) and Lesser Snow Geese (Chen caerulescens caerulescens) breeding at Karrak Lake, NT, Canada and examined variation due to time of day and reproductive stage. We subsequently collected female geese that had foraged for known duration and we estimated mass of foods consumed during foraging bouts. Female Ross s Geese spent more time foraging (mean % ± SE = 28.4 ± 1.3%; P = 0.0002), on average, than did female Lesser Snow Geese (21.5 ± 1.4%). Foraging time by female geese differed among reproductive stages, but differences were not consistent among time periods (stage-bytime block interaction, P=0.0003). Females spent considerably more time foraging during prelaying and laying than during incubation. Ross s Geese also spent a greater percent of time feeding (83.0±2.8%) during incubation recesses than did Lesser Snow Geese (60.9±3.6%). Consumption of organic matter during foraging bouts was minimal; estimated consumption averaged 9.6±4.0 and 12.4±4.6 g (mean ± SE) dry mass/day before incubation and 5.9±2.0 and 5.7±2.1 g dry mass/day during incubation for Lesser Snow and Ross s Geese, respectively. Diets consisted primarily of mosses (bryophytes), Chickweed (Stellaria spp.) and Sedges (Carex spp.). Before incubation, eggshell consumption was estimated as M.L. Gloutney ( ) R.T. Alisauskas S.M. Slattery Department of Biology, University of Saskatchewan, 112 Science Place, Saskatoon, SK, S7 N 5E2, Canada e-mail: M.gloutney@ducks.ca Tel.: +1-902-6678726 R.T. Alisauskas Canadian Wildlife Service, Prairie and Northern Wildlife Research Centre, 115 Perimeter Rd. Saskatoon, SK S7 N 0X4, Canada A.D. Afton U.S. Geological Survey, Biological Resources Division, Louisiana Cooperative Fish and Wildlife Research Unit, Louisiana State University, Baton Rouge, LA 70803, USA M. Gloutney Present address: Ducks Unlimited Canada, PO Box 430, Amherst, NS, B4H 3Z5, Canada 4.3±3.2 and 0.4±0.3 g dry mass/day for Lesser Snow and Ross s Geese, respectively; neither species consumed eggshell during incubation. We conclude that eggshell from nests of previous years is likely an important source of dietary calcium used to meet mineral demands of eggshell formation at Karrak Lake. Our findings of wide disparities between foraging time and food intake indicate that results from studies that do not directly measure intake rates remain equivocal. Finally, we propose four hypotheses accounting for foraging effort that evidently yields little nutritional or energetic benefit to geese nesting at Karrak Lake. Keywords Chen caerulescens caerulescens Chen rossii Female foraging behaviour Lesser Snow Goose Ross s Goose Introduction It has been widely accepted that females of larger species of arctic-nesting geese spend little time feeding on breeding areas before their eggs hatch, but instead rely primarily upon endogenous reserves accumulated during winter or spring migration (e.g., Ryder 1970; Ankney 1977; Newton 1977; Raveling and Lumsden 1977; Ankney and MacInnes 1978; Alisauskas and Ankney 1992). Several recent studies, however, suggest that exogenous nutrients are important during the pre-incubation period in goose populations that are on breeding grounds for several weeks before laying (Budeau et al. 1991; Bromley and Jarvis 1993; Gauthier 1993; Ganter and Cooke 1996). At most Lesser Snow Goose (Chen caerulescens caerulescens, hereafter Snow Geese) and Ross s Goose (Chen rossii) colonies, egg laying typically begins within a few days of arrival (Ryder 1970; Ankney 1977; Newton 1977; Ankney and MacInnes 1978; Raveling 1979; Thomas 1983) and thus, females have a limited number of days to exploit food resources on nesting grounds before laying eggs. However, females are expected to exploit any energy resource on the breeding

grounds given the substantial nutrient and mineral requirements of egg production and incubation. Arcticnesting colonial geese typically arrive on the breeding grounds when there is still extensive snow cover; thus, availability of food is limited. Our preliminary observations, however, indicated that Snow and Ross s Geese spend considerable time foraging at Karrak Lake. Few studies have directly assessed either allocation of time to foraging by breeding geese or rates of nutrient acquisition during foraging bouts (reviewed by Astrom 1993). Aside from energy and nutrients, calcium can limit egg production (e.g., Drent and Woldendorp 1989; Perrins 1996). Although calcium can be drawn from reserves in medullary and cortical bone (Alisauskas and Ankney 1992; Houston et al. 1995), this source may be insufficient to supply large quantities required for eggshell synthesis. Within goose colonies, considerable quantities of calcium may be available as eggshells from previous years; however, the extent to which geese exploit this resource is unknown. Body size is an important factor affecting interspecific variation in the use of exogenous resources (the body size hypothesis; Skutch 1962; Afton 1980). Females of smaller waterfowl species generally allocate proportionately more endogenous nutrients to eggs (Ankney 1984), and so rely more on exogenous nutrients during incubation (Afton and Paulus 1992) because of shorter fasting endurance and higher mass-specific metabolic rates (Calder 1974). Consequently, there may be a minimum size below which females are too small to store sufficient nutrient reserves to meet the energetic demands of egg production and incubation (Afton 1979; Ankney 1984; Afton and Paulus 1992). Both Snow and Ross s Geese rely on endogenous reserves (Ankney and MacInnes 1978; Hobson et al. 1993; Bon 1997) to complete incubation, but Ross s Geese are about 20% smaller than Snow Geese (Slattery and Alisauskas 1995). According to the body size hypothesis, Ross s Geese are predicted to rely more on exogenous nutrients to meet the nutritional demands of incubation than are Snow Geese if both species begin incubation with proportionally similar energy reserves. However, an analysis of nutrient and energy demands during laying is complicated by potential interspecific differences in allocation strategies of endogenous versus exogenous nutrients. If we assume that both species allocate the same relative proportions of their endogenous reserves to eggs, then massspecific metabolic differences should result in Ross s Geese spending more time foraging than Snow Geese in order to meet the energetic costs of reproduction. Our objectives were to compare foraging times and nutrient-intake rates of female Snow and Ross s Geese nesting at Karrak Lake, NT. In addition, we measured eggshell availability and consumption, and predicted that Snow Geese should consume more eggshell than Ross s Geese because Snow Geese lay larger clutches (Snow =3.7±0.8 eggs, Ross s =3.2±0.7 eggs: Slattery and Alisauskas 1993) and eggshell constitutes a greater fraction of egg mass than that in Ross s Geese (Snow =8.4%, Ross s =7.8%, calculated from Slattery and Alisauskas 1995). Materials and methods Study area and methods 79 We estimated foraging time of breeding female geese at Karrak Lake (67 14 N, 100 16 W) from 6 to 24 June 1993; observations were made while concealed in rock outcrops on the largest island in Karrak Lake; observations of individual foraging bouts and collections of female geese were made between 2 and 30 June 1995 on the mainland immediately east of this island. Habitats in both areas were similar. Detailed descriptions of the general region surrounding Karrak Lake were given by Ryder (1969, 1972) and McLandress (1983). Foraging time analysis In 1993, we tested the prediction that female Ross s Geese spend more time foraging than female Snow Geese. We made behavioural observations throughout the 24-h daylight cycle in 4-h time blocks, using instantaneous sampling of focal paired females (Altmann 1974). Time blocks were randomly sampled by each observer without replacement, such that all time blocks were sampled every 3 days. Random selections of time blocks were repeated every 3 days until the study was completed. Within a time block, each observer completed sixteen 10-min sampling periods. We alternated species of geese under observation such that eight 10-min sample periods were completed for each species. This scheme of alternating species was not possible in the first few days because Ross s Geese typically arrive on the colony about 4 days after Snow Geese (Slattery and Alisauskas 1993). We also terminated observations of Snow Geese earlier because of their slightly earlier nesting schedule (see below). Geese were observed with 15 60 spotting scopes at distances from 100 to 1,000 m depending on weather conditions and observation locations. We selected the first paired female which came into the field of view after randomly selecting an observation location and direction. Following completion of the first sample period, the scope was moved to the right until a paired female of the alternate species was located. This process was repeated until the end of the time block. We changed observation locations after each 4-h time block to avoid repeated observations of individual females. During sampling periods, we recorded behaviours of females at 10-s intervals using a metronome (Wiens et al. 1970). We categorised behaviours as: (1) foraging head below horizontal, either grazing, grubbing or searching for food (see Gauthier and Tardif 1991) or (2) other activities. We classified the reproductive stage for each species as follows: (1) prelaying period between first arrival and mean first egg date; (2) laying period between mean first egg date and mean clutch completion date; and (3) early incubation period between mean clutch completion date and mean 8th day of incubation. Mean first egg dates during 1993 at the Karrak Lake Colony were 8 June and 12 June, mean clutch completion dates were 13 June and 16 June, and mean 8th days of incubation were 22 June and 25 June for Snow and Ross s Geese, respectively (Alisauskas, unpublished data). In order to evaluate further incubation recess behaviours, after the mean 8th day of incubation, we used ad libitum sampling (Altman 1974) to quantify activities of focal females during complete incubation recesses (i.e., periods in which females were away from nests; see Afton and Paulus 1992). We continuously scanned a group of nests (5 20 nests depending on nest density) until a female initiated a recess. Following protocol outlined above, we recorded activities of focal females until she returned to the nest. Groups of birds scanned were switched after a complete recess was recorded.

80 Food intake rate analysis In 1995, we evaluated food and eggshell consumption and tested the prediction that Ross s Geese consume more food than Snow Geese. Because of small samples, we divided the breeding season into pre-incubation (prelaying and early laying combined) and incubation periods. During pre-incubation, observers scanned foraging areas for sleeping pairs of geese. Only pairs sleeping for more than 20 min were selected for observation in order to increase the likelihood that food in the oesophagi and gizzard was ingested during the observed foraging bout. Five to 15 sleeping pairs were usually under observation. Behavioural observations began as soon as the first pair woke up. Behaviours were classified as for the foraging time analysis above. Behaviours were recorded from a distance of 100 300 m every 10 s during an entire foraging bout with a 15 60 spotting scope. Focal females were observed until the end of foraging bouts (when females returned to sleep) at which time females were shot with a rifle. During incubation, we observed areas within a radius of 100 300 m where at least 200 females were incubating. We selected females that were seen leaving the nest after covering eggs. Behaviours were recorded until the female resettled on the nest, at which time she was shot. Number of eggs and incubation stage (estimated by candling; Weller 1956) were recorded for each female. Geese were dissected within 4 h of collection. We measured tarsus length, skull length, skull height, and keel length (±0.05 mm) after Dzubin and Cooch (1992) and mass of abdominal fat (±0.1 g). In addition, we counted number of ovulated and developing follicles of pre-incubation geese to determine potential clutch size (after Ankney and MacInnes 1978). We estimated days from ovulation of first egg based on: (1) the number of ovulated follicles, or (2) for females which had not yet laid, the mass of the largest developing follicle was matched to a size hierarchy derived from follicular masses of females in early laying. A negative number indicates that a female had yet to ovulate. Oesophageal and gizzard contents were saved and initially dried in the field. Samples were redried to constant mass in the laboratory and composition of contents determined. Dry mass of consumed vegetation and eggshell was recorded, and vegetation was identified to species (Burt 1991), and structure (e.g., berry, leaf, root). Estimation of daily intake rates Vegetation generally clears the goose gut in 2 4 h (Ziswiler and Farner 1972; Prop and Vulink 1992), although staging Snow Geese have a mean retention time of 1.37±0.33 h (Hupp et al. 1996). We assumed that all forage from previous foraging bouts (usually greater than 20 min from start of observed bout) would have been ground in the gizzard and only unground vegetation in the gizzard was included in our diet analyses. We also assumed that no food was lost from the gizzard because material collected from the gizzard represented on average 33.0±35.9 and 4.8±8.5% of the total mass ingested (oesophageal and gizzard contents) during pre-incubation and incubation foraging bouts, respectively. We calculated total food intake for foraging bouts as total dry mass of oesophageal contents plus unground vegetation in gizzards. Total intake of eggshell included eggshell fragments within oesophagi as well as relatively large fragments (>5 mm) in gizzard. Gizzards typically also contained small eggshell fragments (<2 mm), so we assumed that this break in the distribution of eggshell sizes (i.e., no 3 4 mm fragments) suggested that small fragments were from previous foraging bouts. We calculated intake rates of food and eggshell separately for each female as: intake rate (g/h) = total dry mass ingested (g)/ time foraging (h) (1) where time foraging is the number of minutes that geese were observed foraging during a bout. Using estimates of daily foraging time obtained in 1993, we calculated total daily food and eggshell intake rates as: estimated daily intake (g/day) = intake rate (g/h) predicted daily foraging time (h/day) (2) Availability of eggshell In 1996, we collected all eggshell present within a 1.5 m radius of 56 randomly selected nests from previous years, located in areas where intake rates were determined in 1995. Eggshell fragments were sorted from soil and other debris and dried to constant mass. Statistical analyses Throughout, all ANOVA, ANCOVA and MANCOVA were based on Type III (i.e., orthogonal) SS (SAS 1993, pp 120 122). Foraging time Two experienced observers made behavioural observations of female geese. Simultaneous observations by both observers of the same female (1993: n=32 females, 1995: n=22 females) revealed no differences in observed behaviours; consequently, we assumed that variance due to observer effects was negligible and, thus, was not considered in the analyses. Females were the sampling unit in all statistical analyses. We believe that probability of repeated observations of individual females was small because of the large number of pairs typically within view (n=3,300 Ross s Goose pairs, n=1,700 Snow Goose pairs), and our changing observation locations. We used 3-way ANOVA (SAS 1993) to examine variation in daily foraging data collected in 1993 that was due to species, reproductive stage, time block and all their interactions. Data on foraging time during incubation recesses was analysed separately using two-way ANOVA to test for differences due to species, year (1993 or 1995) and their interactions. Angular transformations were applied to proportions before analysis to improve homogeneity of variances (Sokal and Rohlf 1969). Least-squares means (± SE) were obtained from reduced models containing only significant explanatory variables. We used a stepwise hierarchical procedure, in which the highest level, non-significant interactions were deleted and the analysis was redone (Alisauskas and Ankney 1994). Final models contained only significant effects. Analyses of raw and transformed data in final models yielded similar results; consequently, least-squares means (± SE) of raw data are presented. Interspecific analysis of diets and intake rates Structural size We indexed body size using the first principal component from a principal components analysis (PCA) on the correlation matrix of the four morphometric measurements (Reyment et al. 1984). Species were pooled in the PCA to index body size on the same scale. Food intake rates Data were restricted to females for which both food consumption and foraging time were known. We compared food and eggshell intake rates between Snow and Ross s Geese using t-tests. Analyses were performed separately for pre-incubation and incubation periods because dietary composition differed between periods (see below). Eggshell was never found in incubation period samples and was omitted from the incubation analysis.

81 Dietary composition All females collected before incubation were used in these analyses, including three females for which we did not collect behavioural data. We calculated frequency of occurrence of each dietary component as proportion of females in which a dietary component was found. Proportional dry mass was transformed to square root arcsine values before multivariate analyses of covariance (MANCOVA) to test if overall diets differed in relation to species. Differences in diet diversity precluded the pooling of diets between the two periods. Covariates used in MANCOVA were body size (from PCA), mass of abdominal fat (g), clutch size estimated from examination of follicles (Ankney and MacInnes 1978), and days from ovulation of first egg. Statistical protocol follows the hierarchical procedure outlined above. F-values based on Type III SS reported from MANCOVA were determined using Wilks criterion (SAS 1993). Dietary diversity In order to evaluate whether diet diversity differed between species, we calculated Shannon-Wiener diversity indices that consider the relative abundance of each component in the diet for each female and is calculated based on the mass of each dietary component consumed. We then used ANOVA to test for differences in diet diversity between species and period. Results Analysis of 24-h foraging time Female Ross s Geese spent more time foraging (mean %± SE=28.4±1.3%: F=12.9, df=1, 1164, P=0.0003), on average, than did female Snow Geese (21.5±1.4%). Controlling for species effects, foraging time differed among reproductive stages, but differences were not consistent among time blocks (reproductive stage-by-time block interaction, F=3.1, df=10, 1164, P=0.0007). Females spent considerably more time feeding during prelaying and laying than during incubation (Fig. 1). None of the other interactions were significant (Ps >0.07). Fig. 1 Mean foraging time (x% ± SE) of Snow and Ross s Geese females by stage of reproduction and time block, Karrak Lake, NT, 1993 Foraging during incubation recesses Our final model indicated that species differences in foraging time during incubation recess were not consistent between years (year by species interaction: F=6.2, df=1, 24, P<0.0201). Ross s Geese spent more time foraging (LSM± SE=83.0%±2.8: P=0.0002) than did Snow Geese (60.9%±3.6) in 1993, but species did not differ (P=0.56) in 1995, although the means varied in the same direction as in 1993 (Ross s Geese=83.3%±3.6; Snow Geese= 80.0%±3.6). Pre-incubation: estimated food and eggshell intake We collected 10 Snow and 13 Ross s Geese during the pre-incubation period for which foraging bout behaviours were recorded. Food intake (g/h) and daily food intake (g/day) generally were low and did not differ be- Table 1 Mean ± SE (range) foraging data for preincubation and incubating female Snow and Ross s Geese at Karrak Lake, NT. t-test of difference between species in food intake rates and daily food consumption Snow Geese Ross s Geese t (P) Preincubation: (n=10) (n=13) Foraging bout (min) 22.2±1.2 (19.7 31.7) 24.2±3.3 (6.7 49.8) 0.52 (0.61) Time spent foraging (min) 14.8±1.5 (7.0 20.5) 19.1±3.3 (3.8 46.7) 1.07 (0.3) Food intake (g/h) 1.37±0.57 (0 4.98) 1.35±0.5 (0.01 5.98) 0.02 (0.99) Eggshell intake (g/h) 0.61±0.45 (0 4.46) 0.05±0.03 (0 0.45) 1.43 (0.17) Mean daily foraging time (h) 7.06±0.43 9.17±0.6 Daily food intake (g/day) 9.6±4.0 12.4±4.6 2.08 (0.65) Daily egg shell intake (g/day) 4.3±3.2 0.4±0.3 2.07(0.18) Incubation recess: (n=6) (n=6) Recess length (min) 13.3±3.6 (4.5 26.8) 15.6±1.9 (11 24) 0.56 (0.59) Time spent foraging (min) 10.8±3.0 (2.8 20.7) 12.9±1.5 (8 18.7) 0.64 (0.53) Food intake (g/h) 6.51±2.20 (0.14 16.4) 2.91±1.07 (0 6.26) 1.24 (0.24) Eggshell intake (g/h) 0 0 Mean daily foraging time (h) 0.91±0.29 1.97±0.26 Daily food intake (g/day) 5.9±2.0 5.7±2.0 2.23 (0.95)

82 Fig. 2 Eggshell mass as a proportion of the diet of Snow and Ross s Geese females in relation to days from ovulation of first egg, where negative ovulation dates correspond to a female which is yet to ovulate, Karrak Lake, NT, 1995 tween species during either pre-incubation or incubation periods (Table 1). Both species appear to increase their consumption of eggshell with the onset of laying (Fig. 2). Eggshell intake rates (g/h) and daily eggshell intake (g/day) did not differ between species (Table 1). Diet Diets of 10 Snow and 16 Ross s Geese collected during pre-incubation were analysed (Table 2). While controlling for other significant effects (body size and mass of abdominal fat), overall dietary composition differed between species (Wilks λ=0.52, F=3.3, df=5, 18, P=0.028; Table 3). During foraging bouts Ross s Geese consumed less eggshell than did Snow Geese (Tables 2 and 3). There were no differences between species in the proportion of any other dietary components (Tables 2 and 3). Six Snow and six Ross s Geese were collected following incubation recesses. Overall diets during incubation recesses did not differ between species (Wilks λ = 0.3; F=1.3, df=7, 4, P=0.41, Tables 2 and 3). Dietary diversity Our final model indicated that species differences in Shannon-Wiener diversity indices were not consistent between periods (species by period interaction, F=5.3, Table 2 Mean total dry mass ± SE (range) of oesophageal and gizzard contents per foraging bout in preincubation and incubating female Snow and Ross s Geese, and mean percent composition ± SE of each dietary component. Frequency of occurrence shown in parenthesis Snow Geese Ross s Geese Preincubation Incubation Preincubation Incubation (n=10) (n=6) (n=16) (n=6) Total dry mass (g) 0.46±0.18 1.21±0.47 0.39±0.12 0.69±0.25 (0.0 1.61) (0.01 3.04) (0.01 1.48) (0 1.39) Eggshell 16.7±11.0 (30) 0 (0) 9.7±6.2 (39) 0 (0) Moss a 20.0±13.3 (20) 55.5±18.2 (67) 50.8±11.9 (56) 82.5±16.5 (83) Chickweed b 10.0±10.0 (10) 0 (0) 25.4±10.6 (31) 0.5±0.5 (17) Sedge c 22.8±12.5 (30) 1.1±1.1 (17) 12.5±8.5 (11) 0 (0) Cranberry d leaves Trace (10) 11.1±8.0 (50) 0 (0) 0 (0) Cranberries 10.5±9.9 (20) 0.7±0.7 (17) 0 (0) 0 (0) Bearberry e leaves 0 (0) 9.6±8.4 (33) Trace (6) 0.1±0.1(17) Labrador tea f leaves Trace (10) 0.2±0.2 (33) Trace (6) 0.1±0.1(17) Crowberry g leaves 0 (0) 0 (0) 0 (0) Trace (17) Arctic heather h 0 (0) 0.9±0.6 (33) 0 (0) 0 (0) Roots 0 (0) 24.0±13.4 (50) 0 (0) 0 (0) a (Bryophytes); b (Stellaria spp.); c (Carex spp.); d (Vaccinium vitis-idaea); e (Arctostaphylos spp.); f (Ledum decumbens); g (Empetrum nigrum); h (Cassiope tetragona) Table 3 P values from analyses of covariance (ANCOVA: Type III SS) test for species differences in diets of Snow and Ross s Geese during preincubation and incubation periods. Significant (P<0.05) explanatory variables were species (Spec), body size (PC1), abdominal fat (Abdo) Eggshell Moss Chickw a Sedge CrnFrt b CrnLf c BearLF d LabTLf e Heather Root Pre-incubation Spec 0.016 0.78 0.24 0.18 0.91 NI f NI NI NI NI PC1 0.045 0.77 0.3 0.17 0.57 Abdo 0.015 0.21 0.97 0.36 0.45 Incubation Spec NI 0.25 0.34 0.34 0.34 0.13 0.26 0.67 0.16 0.073 a Chickweed; b Cranberries; c Cranberry leaves; d Bearberry leaves; e Labrador tea leaves; f Not present in any sample, thus, not included in analyses

df=2, P=0.01), with Snow Geese having more diverse diets during incubation (pre-incubation mean ± SE, 0.11±0.04, incubation 0.23±0.05) and Ross s geese having more diverse diets during pre-incubation (pre-incubation 0.17±0.03, incubation 0.02±0.01). Eggshell availability There were about 90,000 nests of each goose species on 62.2 km 2 of land surface within the Karrak Lake colony perimeter in 1994 (Alisauskas, unpublished data). Average clutch sizes within the colony are 3.65 and 3.32 eggs for Snow and Ross s Geese, respectively (Slattery and Alisauskas 1993). This yields, on average, 38.2 and 23.9 g of eggshell/nest of Snow and Ross s, respectively (eggshell mass from Slattery and Alisauskas 1995). Therefore, in 1994, an estimated 5,589 kg (0.89 kg/ha) of eggshell was deposited on the colony. We found that previous years nests contained 9.2±7.7 g, (n=56, range 0 27.5 g) dry mass of eggshell resulting in about 1,656 kg (0.26 kg/ha) of eggshell available in spring. These estimates are biased low because eggshell was collected while geese were on the breeding grounds and geese had probably consumed some eggshell before our eggshell collections. Discussion Snow and Ross s Geese breeding at Karrak Lake provide a unique opportunity to examine the body size hypothesis as it relates to endogenous reserve use during reproduction. The two species are closely related (Avise et al. 1992) and differ in body size (MacInnes et al. 1989; Slattery and Alisauskas 1995). Both species also generally nest within the same habitats and have similar nesting chronology (Slattery and Alisauskas 1993; LeSchack et al. 1998). Thus, regional variation in microclimate is controlled in species comparisons made at the colony. However, the two species use different winter and spring staging areas (Bellrose 1980); thus, factors affecting nutrient reserve storage and use before arrival at Karrak Lake are not controlled in our comparisons. Given interspecific differences in body size and mass, Ross s Geese should have higher mass-specific metabolic rates (kj/day per kg, Calder 1974) and larger surface area/volume ratios than Snow Geese. Assuming that maximum endogenous reserves are proportional to body size, fasting endurance of Ross s Geese should be less than that of Snow Geese because of greater mass-specific energy requirements and higher rates of heat loss in Ross s Geese (Calder 1974). As predicted by the body size hypothesis, Ross s Geese spent more time foraging from arrival through early incubation than did Snow Geese. We suspect that species differences in foraging time would have been greater if our sampling had included the entire incubation period. Although we did not detect species differences in female 83 nest attendance during early incubation, female Ross s Geese are less attentive to their nests later in incubation than are Snow Geese (Afton, unpublished data) and, thus, may spend more time foraging during incubation. We also found that Ross s Goose females spent a greater proportion of incubation recesses foraging than Snow Geese females in 1993. This difference was not significant during the cooler 1995 season (Alisauskas unpublished data). Thus, during cold weather conditions, female Snow Geese may have catabolised reserves at a greater rate than during warm weather and apparently adjusted their foraging time in response to cooler conditions. The lack of response by Ross s Geese may have arisen if they are already investing maximum time to foraging. Alisauskas and Ankney (1992) argued that reliance on endogenous reserves during laying and incubation is greater in Chen spp. than Branta spp., in part because Chen spp. nest in dense colonies where interspecific competition for food is greater. Branta spp. and Whitefronted Geese (Anser albifrons) generally show a longer lag between arrival and laying (10 13 days) than do Chen spp. (<5 days), allowing for increased foraging time and replenishment of reserves depleted during migration (Raveling 1978; Ely and Raveling 1984; Budeau et al. 1991; Bromley and Jarvis 1993). In contrast to other Chen spp. Greater Snow Geese (Chen caerulescens atlantica) also show a long lag between arrival and laying, during which time females forage intensively (Gauthier and Tardif 1991; Gauthier 1993; Choiniere and Gauthier 1995; but see Bon 1997) and are reported to rely very little on endogenous reserves during prelaying and laying. So reliance on endogenous reserves appears related to the time lag between arrival on the breeding ground and initiation of laying provided that forage is available on the breeding grounds. Our results show that although Snow and Ross s Geese dedicate a substantial amount of time to foraging, intake rates of organic matter were small. Furthermore, much of the ingested food was moss with low nutritive value (Prop and Vulink 1992). Taken together, these suggest that intake of nutrients during nesting was trivial. This disparity between foraging effort and energy intake has important implications for studies of goose nutrition that have not directly measured intake rates (i.e., Gauthier and Tardif 1991; Ganter and Cooke 1996) as we have done. Furthermore, given low intake rates, we question why geese at Karrak Lake devoted such effort to foraging? Several other species of arctic-breeding geese also consume low quality foods while in the Arctic (Derksen et al. 1982; Sedinger and Raveling 1984; Madsen et al. 1989; Loonen et al. 1991; Prop and Vulink 1992). We propose four hypotheses: (1) facultative exploitation of any external energy source during breeding is beneficial, but habitats at Karrak Lake currently are so degraded that intake rates are low, (2) foraging is a means of defining/signalling/defending territories, (3) geese need to ingest a minimum amount of food to maintain gut microflora and (4) geese are searching for calcium rather than energy or organic nutrients.

84 Where examined, nutrient storage and use during egg formation and incubation is virtually a ubiquitous phenomenon in temperate and arctic-nesting waterfowl (Alisauskas et al. 1990; Ankney et al. 1991; Afton and Paulus 1992; Alisauskas and Ankney 1992, 1994). However, species vary in their use of endogenous reserves and arctic-nesting geese likely have been selected for extreme reliance on stored nutrients. Indeed, failure to store sufficient reserves can result in starvation during incubation (Ankney and MacInnes 1978). Given high variability in arctic weather, geese may exploit any available food resource in order to minimise the consumption of endogenous reserves. Long-term dynamics in forage availability are likely linked to local population changes in arctic-nesting geese (Kerbes et al. 1990). Over the past 25 years, populations of Snow and Ross s Geese nesting at Karrak Lake have increased from 17,000 nesting geese in 1965 1967 (Kerbes 1994) to 480,000 in 1995 (Alisauskas, unpublished data). Increased breeding populations at Karrak Lake have resulted in extensive devegetation within the colony (Alisauskas, unpublished data). Consequently, foraging female geese at Karrak Lake presently have little opportunity to ingest significant amounts of food. Barnacle Geese (Branta leucopsis) increased digestive efficiency by increasing gut retention times (Prop and Vulink 1992). This has been proposed by Prop and Vulink (1992) as a strategy used by Barnacle Geese that enables them to exploit mosses on the breeding grounds. If Snow and Ross s Geese at Karrak Lake are employing a similar strategy, then this may account for the observed consumption of relatively poor quality forage. Foraging behaviour could be a means of defining/defending territories. Ryder (1975) proposed that male Ross s Geese defended territories that were large enough to provide sufficient forage for themselves while attending nesting females during incubation. In contrast, Inglis (1976) proposed that the main function of territoriality in Pink-footed Geese (Anser brachyrhynchus) was to ensure a supply of food around the nest for the female, especially early in nesting. In 1995, Snow and Ross s Geese on average had 3.0±0.2 (mean ± SE, n=63, range =0 6) and 2.8±0.2 (n=77, range= 0 7) neighbouring nesting within a 10-m radius. Thus, at Karrak Lake this explanation does not appear plausible because territories were as small as 52 and 45 m 2 for Snow and Ross s Geese, respectively. We believe it is unlikely under current degraded conditions that there is sufficient forage within territories for any significant nutritional or energetic benefits to be achieved. However, benefits may be derived from territoriality if extra pair copulations or nest parasitism are reduced (Mineau and Cooke 1979). When food intake rates and food quality are reduced, digestive efficiency may be assisted by a complex gut microfloral community (Prop and Vulink 1992; Bedford 1996) including cellulolytic bacteria (Buschbaum et al. 1986; Prop and Vulink 1992). If this community is eliminated during periods of starvation, forage conversion efficiencies may be depressed when females most need to regain body condition (Ankney 1982; LeSchack et al. 1998). Therefore, geese may consume low quality food during incubation to maintain gut microflora, but further research is required to confirm this. A calcium appetite may motivate foraging behavior. Not only are large quantities of calcium required during laying, but calcium must be rapidly mobilised (Perrins 1996). Calcium used in eggshell formation originates from the diet, or from medullary and cortical bone sources (Raveling 1978; Alisauskas and Ankney 1992; Houston et al. 1995). In most species, medullary bone acts as a short-term calcium reserve (Alisauskas and Ankney 1992). However, Ankney and Scott (1980) predicted that medullary bone might serve as a significant calcium reserve in species that feed little during laying. Ingestion of calcareous resources during laying has long been known (i.e., Simkiss 1961; Perrins 1996). Our results clearly demonstrate exploitation of exogenous calcium and this may be an important supplement to declining mineral reserves as egg formation proceeds in both Snow Geese (Ankney and MacInnes 1978) and Ross s Geese (Bon 1997). Indeed, eggshell consumption increased with onset of laying (Fig. 2). Mean eggshell mass of Snow and Ross s Geese is 10.46 and 7.21 g, respectively (Slattery and Alisauskas 1995). Therefore, by consuming 4.5 g of eggshell/day and assuming a laying rate of 1 egg/1.3 days (Ryder 1971), and a 100% conversion efficiency, Snow Geese, obtained 56% of the 41.8 g of calcium required for a four egg clutch, from exogenous eggshell in 1995. On the other hand, daily consumption of 0.4 g of eggshell provides only 7.2% of the 28.8 g of calcium required for a clutch of four Ross s Goose eggs. Further evidence of a calcium appetite for egg formation is suggested by the absence of eggshell consumption during incubation. Eggshell is readily available to arriving geese in snowfree areas, and is probably exploited to offset rapid calcium mobilisation during laying (Fig. 2). Therefore, we suggest that during laying, birds forage more for calcium rather than for energy or other organic nutrients. Ross s Geese arrive at the colony and initiate nesting 2 4 days later than Snow Geese, by which time a significant proportion of Snow Geese are already laying (Slattery and Alisauskas 1993). Seasonal differences in arrival may influence availability of eggshell, potentially explaining interspecific differences in eggshell consumption. However, eggshell availability varies temporally, with old nests continually being exposed as snow recedes. As far as we are aware, this is the first study reporting detailed dietary composition of pre-incubation and incubating Snow and Ross s Geese. Our results for dietary composition reflect feeding within a highly degraded colony. Growth in number and size of Snow and Ross s Goose colonies throughout arctic Canada (Kerbes 1994) suggests that poor foraging conditions may be widespread and probably will prevail until density dependent mechanisms lead to population declines and habitats recover. Potential consequences may be similar to those observed on the subarctic colony at La Pérouse Bay, where habitat degradation has been linked with declining

clutch and fledgling body size and declining first year survival (Cooke et al. 1995). Ganter and Cooke (1996) reported that Snow Geese at La Pérouse Bay in the subarctic fed extensively before and during laying by grubbing in salt marshes, and speculated that food consumption comprised a substantial part of the energetic costs of egg formation. However, at an arctic colony we found that food consumption was trivial despite substantial time spent foraging by geese. At the La Pérouse Bay colony geese may stage and feed for two weeks on nesting areas before laying eggs. In contrast, at Karrak Lake geese arrive and commence nesting within a few days of arrival, when there is little or no green vegetation. Additionally, habitat at the Karrak Lake colony consists of less productive arctic tundra habitats (see Ryder 1972 for description of habitats) than the coast wetlands of La Pérouse Bay. Therefore, Ganter and Cooke s (1996) implication that food on nesting areas is important for successful reproduction does not apply to geese nesting at Karrak Lake. Instead, we believe that in addition to endogenous sources of fat, protein and mineral, exogenous calcium from eggshell produced in previous years may be a more critical resource than are exogenous nutrients in constraining clutch size and nesting success. In summary, Snow and Ross s Geese spent significant amounts of time foraging during pre-incubation and incubation at Karrak Lake. However, food intake rates were minimal. Our results support the prediction that Ross s Geese forage more than Snow Geese. However, our results did not support the prediction that Ross s Geese would consume more food than Snow Geese. We suggest that this similarity in food consumption results from extensive devegetation as the colony expanded over the past three decades. Furthermore, we suggest that observed foraging behaviours also result from a calcium appetite as females attempt to supplement calcium requirements for eggshell formation. Acknowledgements We thank Delia Person, Stephanie Koch, Dana Kellett, Doug Stern, Carma Woof, Mellissa Cunningham and Jeff Danielson for help with fieldwork at Karrak Lake and Bobby Cox and Dave Shutler for comments on manuscript. Our project received financial and logistical support from California Department of Fish and Game, Canadian Wildlife Service, Polar Continental Shelf Project, Louisiana Cooperative Fish and Wildlife Research Unit, School of Forestry, Wildlife and Fisheries at Louisiana State University, Louisiana Department of Wildlife and Fisheries, and NSF/LaSER Grant (1993)-HRD-01. References Afton AD (1979) Time budget of breeding Northern Shovelers. Wilson Bull 91:42 49 Afton AD (1980) Factors affecting incubation rhythms of Northern Shoveler. Condor 82:132 137 Afton AD, Paulus SL (1992) Incubation and brood care. In: Batt BDJ, Afton AD, Anderson MG, Ankney CD, Johnson DH, Kadlec JA, Krapu GL (eds) Ecology and management of breeding waterfowl. University of Minnesota Press, Minneapolis, pp 62 108 85 Alisauskas RT, Ankney CD (1992) The cost of egg laying and its relationship to nutrient reserves in waterfowl. In: Batt BDJ, Afton AD, Anderson MG, Ankney CD, Johnson DH, Kadlec JA, Krapu GL (eds) Ecology and management of breeding waterfowl. University of Minnesota Press, Minneapolis, pp 30 61 Alisauskas RT, Ankney CD (1994) Nutrition of breeding female Ruddy Ducks: the role of nutrient reserves. Condor 96:878 897 Alisauskas RT, Eberhardt RT, Ankney CD (1990) Nutrient reserves of breeding Ring-necked Ducks (Aythya collaris). Can J Zool 68:2524 2530 Altmann J (1974) Observational study of behavior: sampling methods. Behaviour 49:227 267 Ankney CD (1977) Feeding and digestive organ size in breeding Lesser Snow Geese. Auk 94:275 282 Ankney CD (1982) Annual cycle of body weight in Lesser Snow Geese. Wildl Soc Bull 10:60 64 Ankney CD (1984) Nutrient reserve dynamics of breeding and molting Brant Auk 101:361 370 Ankney CD, MacInnes CD (1978) Nutrient reserves and reproductive performance of female Lesser Snow Geese. Auk 95:459 471 Ankney CD, Scott DM (1980) Changes in nutrient reserves and diet of breeding Brown-headed cowbirds. Auk 97:684 696 Ankney CD, Afton AD, Alisauskas RT (1991) The role of nutrient reserves in limiting waterfowl reproduction. Condor 93:1029 1032 Astrom M (1993) Time allocation in Canada Geese during the breeding season. Ornis Fenn 70:89 95 Avise JA, Alisauskas RT, Nelson WS, Ankney CD (1992) Matriarchal population genetic structure in an avian species with female natal philopatry. Evolution 46:1084 1096 Bedford MR (1996) Interaction between ingested feed and the digestive system in poultry. J Appl Poultry Res 5:86 95 Bellrose FC (1980) Ducks, geese and swans of North America. Stackpole Books, Harrisburg Bon RL (1997) Spring nutritional ecology of migrating and breeding Ross s Geese Chen rossii. M.Sc. Thesis, University of Saskatchewan, Saskatoon Bromley RG, Jarvis RL (1993) The energetics of migration and reproduction of Dusky Canada Geese. Condor 95:193 210 Budeau DA, Ratti JT, Ely CR (1991) Energy dynamics, foraging ecology, and behavior of prenesting Greater White-fronted Geese. J Wildl Manage 55:556 563 Burt P (1991) Barrenland beauties. Outcrop Ltn. Yellowknife Buschbaum R, Wilson J, Valiala I (1986) Digestibility of plant constituents by Canada Geese and Atlantic Brant. Ecology 67:368 393 Calder WA (1974) Consequences of body size for avian energetics. In: Paynter RA (ed) Avian energetics. Nuttall Ornithol Club Publ No 15, pp 86 144 Choiniere L, Gauthier G (1995) Energetics of reproduction in female and male Greater Snow Geese. Oecologia 103:379 389 Cooke F, Rockwell RF, Lank DB (1995) The Snow geese of La Pérouse Bay: Natural selection in the wild. Oxford University Press, Oxford Derksen DV, Eldridge WD, Weller MW (1982) Habitat ecology of Pacific Black Brant and other geese moulting near Testekpuk Lake, Alaska. Wildfowl 33:33 57 Drent PJ, Woldendorp JW (1989) Acid rain and eggshells. Nature 339:431 Dzubin AX, Cooch EG (1992) Measurements of geese: general field methods. California Waterfowl Association. Sacramento, Calif. Ely CR, Raveling DG (1984) Breeding biology of Pacific Whitefronted Geese. J Wildl Manage 48:823 837 Ganter B, Cooke F (1996) Pre-incubation feeding activities and energy budgets of Snow Geese: can food on the breeding grounds influence fecundity? Oecologia 106:153 165 Gauthier G (1993) Feeding ecology of nesting Greater Snow Geese. J Wildl Manage 57:216 223

86 Gauthier G, Tardif J (1991) Female feeding and male vigilance during nesting in Greater Snow Geese. Condor 93:701 711 Hobson KA, Alisauskas RT, Clark RG (1993) Stable-nitrogen isotope enrichment in avian tissues due to fasting and nutritional stress: implications for isotopic analyses of diet. Condor 95: 388 394 Houston DC, Donnan D, Jones PJ (1995) The source of the nutrients required for egg production in Zebra Finches Poephila guttata. J Zool Lond 235:469 483 Hupp JW, White RG, Sedinger JS, Robertson DG (1996) Forage digestibility and intake by Lesser Snow Geese: effects of dominance and resource heterogeneity. Oecologia 108:232 240 Inglis IR (1976) Agonistic behaviour of breeding Pink-footed Geese with reference to Ryder s hypothesis. Wildfowl 27:95 99 Kerbes RH (1994) Colonies and numbers of Ross s Geese and Lesser Snow Geese in the Queen Maud Gulf Migratory Bird Sanctuary. Can Wildl Serv Occ Paper No 81 Kerbes RH, Kotanen PM, Jeffries RL (1990) Destruction of wetland habitats by Lesser Snow Geese: a keystone species on the west coast of Hudsons Bay. J Appl Ecol 27:242 258 LeSchack CR, Afton AD, Alisauskas RT (1998) Effects of male removal on female reproductive biology in Ross s and Lesser Snow Geese. Wilson Bull 110:56 64 Loonen MJJE, Zijlstra M, van Eerden MR (1991) Timing of wing moult in Greylag Geese Anser anser in relation to the availability of their food plants. Ardea 79:253 260 MacInnes CD, Misra RK, Prevett JP (1989) Differences in growth parameters of Ross s Geese and Snow Geese: evidence from hybrids. Can J Zool 67:286 290 Madsen J, Bregnballe T, Mehlum F (1989) Study of the breeding biology and behaviour of the Svalbard population of Lightbellied Brent Goose Branta bernicla hrota. Polar Res 7:1 21 McLandress MR (1983) Temporal changes in habitat selection and nest spacing in a colony of Ross s and Lesser Snow Geese. Auk 100:335 343 Mineau P, Cooke F (1979) Territoriality in Snow Geese or the protection of parenthood-ryder s and Inglis s hypotheses re-assessed. Wildfowl 30:16 19 Newton I (1977) Timing and success of breeding in tundra-nesting geese. In: Stonehouse B, Perrins C (eds) Evolutionary ecology. University Park Press, Baltimore, pp 113 126 Perrins CM (1996) Eggs, egg formation and the timing of breeding. Ibis 138:2 25 Prop J, Vulink T (1992) Digestion by Barnacle Geese in the annual cycle: the interplay between retention time and food quality. Funct Ecol 6:180 189 Raveling DG (1978) The timing of egg laying by northern geese. Auk 95:294 303 Raveling DG (1979) The annual cycle of body composition of Canada Geese with special reference to control of reproduction. Auk 96:234 252 Raveling DG, Lumsden HG (1977) Nesting ecology of Canada Geese in the Hudson Bay lowlands of Ontario: evolution and population regulation. Ontario Ministry Nat Resour Fish Wildl Res Rep No 98 Reyment RA, Blackith RE, Campbell NA (1984) Multivariate morphometrics. Academic Press, London Ryder JP (1969) Nesting colonies of Ross s Goose. Auk 86:282 292 Ryder JP (1970) A possible factor in the evolution of clutch size in Ross s Goose. Wilson Bull 82:5 13 Ryder JP (1971) Distribution and breeding biology of the Lesser Snow Goose in central arctic. Wildfowl 22:18 28 Ryder JP (1972) Biology of nesting Ross s Geese. Ardea 60:185 215 Ryder JP (1975) The significance of territory size in colonial nesting geese an hypothesis. Wildfowl 26:114 116 SAS (1993) SAS/STAT User s guide. Version 6. SAS Institute, Cary, N.C. Sedinger JS, Raveling DG (1984) Dietary selectivity in relation to availability and quality of food for goslings of Cackling Geese Auk 101:295 306 Simkiss K (1961) Calcium metabolism and avian reproduction. Biol Rev 36:321 367 Skutch AF (1962) The constancy of incubation. Wilson Bull 74:115 152 Slattery SM, Alisauskas RT (1993) Studies on the nutritional ecology and population biology of Ross s Geese nesting at Karrak Lake, N.W.T. Prog Rep for Calif Dept Fish and Game, October 1993. Can Wildl Serv, Saskatoon, SK Slattery SM, Alisauskas RT (1995) Egg characteristics and body reserves of neonate Ross s and Lesser Snow Geese. Condor 97:970 984 Sokal RR, Rohlf FJ (1969) Biometry. Freeman, San Francisco Thomas VG (1983). Spring migration: the prelude to goose reproduction and a review of its implications. In: Boyd H (ed) First Western Hemisphere waterfowl and waterbirds symposium. Can Wildl Serv, Ottawa, pp 73 81 Weller MW (1956) A simple field candler for waterfowl eggs. J. Wildl Manage 20:111 113 Wiens JA, Martin AG, Holthaus WR, Iwen FA. (1970) Metronome timing in behavioural ecology studies. Ecology 51:350 352 Ziswiler V., Farner DS (1972) Digestion and the digestive system. In: Farner DS, King JR (eds) Avian biology, vol 2. Academic Press, New York, pp 343 430