The Wilson Journal ofOrnithology 19(l):23-27, 2007 1 PARENTAL INVESTMENT IN SWAN GEESE IN AN URBAN ENVIRONMENT CHRISTOPH RANDLER' — ABSTRACT. I studied brood-rearing behavior of introduced Swan Geese (Anser cygnoides) in Heidelberg, Germany during 2002 and 2003. Two hypotheses were tested: (1) division of labor between males and females is similar to that of wild Anser species, and (2) parental investment (vigilance behavior) is adjusted for brood size. I used 10-min sessions of focal animal sampling during which I simultaneously recorded the behavior of the male, the female, and a majority ofthejuveniles every 15 sec. Division oflaborwas similartothatobserved in wild Anser populations: males were more vigilant whereas females spent more time feeding during the first 4 weeks of brood-rearing. As brood-rearing progressed, vigilance and agonistic behavior by both males and females decreased, whereas juveniles decreased feeding and increased vigilance. Adults (males and females combined) adjusted vigilance for brood size. A general linear model showed a significant influence of both brood size and brood age on parental vigilance. Received 12 February 2004. Accepted 12 July 2006. During brood rearing, females of most spe- Behavior of introduced geese has rarely cies of wild Anser geese usually spend more been studied (e.g.. Randier 2003a, 2003b), and time feeding than males to compensate for en- little is known about their brooding behavior. ergy loss during incubation. Males spend Studies of introduced geese in an urban en- more time being vigilant, i.e., looking for vironment, where most natural predators are predators (Afton and Paulus 1992). This di- absent, may clarify the complimentary hy- vision of labor by gender was found in time potheses of parental care in geese. Further- budget studies of many goose species (Afton more, parental investment in neither wild nor and Paulus 1992). introduced Swan Geese {Ansercygnoides) has Concerning brood size, larger groups of been examined. Since this species is critically goslings should receive more vigilance by endangered (Goroshko 2001), studies ofintro- their parents than smaller groups as parental duced populations may be of conservation in- investment is considered to be “shared” (Les- terest. The objectives ofthis study were to ex- sells 1987). This hypothesis suggests that pa- amine (1) whether parental care (time budgets rental care might be adjusted for brood size and division of labor between males and fe- by devoting more time to vigilance as brood males) of introduced Swan Geese is similar to size increases (“shared” parental investment). that of wild populations, and (2) the relation- The “unshared” parental investment hypoth- ship between brood size (number of goslings) esis suggests that parental vigilance should and parental care (as measured by vigilance). not be adjusted for brood size, since any time METHODS devoted to vigilance benefits all goslings si- multaneously, regardless of brood size. Some The Swan Goose is a non-native species in empirical tests found an adjustment of paren- Europe, having been introduced in the 18th tal investment (e.g., the level of vigilance) to century (Delacour 1954). The study flock in brood size (Sedinger and Raveling 1990, For- Heidelberg, southwestern Germany (8°41' E, slund 1993, Siriwardena and Black 1999) and 49° 25' N) was established in the 1990s. The others did not (Lazarus and Inglis 1978, Les- birds breed on an island in the Neckar River sells 1987, Schmutz and Laing 2002). How- and soon after hatching, families move to feed ever, gosling age is another important variable on a lawn which extends 1.1 km along the since mortality of goslings is highest during river. In 2002 and 2003, I studied 13 families the first 2-3 weeks of life (Owen 1980, For- of Swan Geese (140 individuals in 2002 and slund 1993). 174 in 2003) during brood rearing (Randier 2003a, 2003b). ' University of Leipzig, Institute of Biology I, Jo- I used instantaneous focal animal sampling hannisallee 21-23, D-04103 Leipzig, Germany;e-mail: (Altmann 1974) to detect differences between [email protected] males and females, and identify rare behaviors 23 24 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 119, No. 1, March 2007 that may be overlooked during flock scans studied, I also sampled 17 non-breeders ofun- (Baldassarre et al. 1988). Sampling sessions known gender between 31 May and 26 June were 10-min/family, during which I recorded 2002 using the same focal-animal sampling the behavior of the male, female, and a ma- method. jority of the juveniles. Goslings could not be Gender of adults was assigned by knob identified individually and I recorded the be- size, bill size, body size, and behavior. Fe- havior displayed by the majority of the brood males had shorter and thinner bills, and short- at each instantaneous sample (Schmutz and er necks (Madge and Bum 1988, Ogilvie and Laing 2002). The order of sampling families Young 1998); males swam behind broods was random. I used 15-sec sampling intervals (Bauer and Glutz von Blotzheim 1968, Rut- because this interval provides data that are schke 1997). Family sample sizes by periods close to continuous observations (Pdysa were 13 (weeks 1-4), 8 (weeks 5-8), and 9 1991). I recorded the following behavioral (weeks 9-12); I sampled each family between categories (adapted from McWilliams and one and five times during each period. I first Raveling 1998): feeding, resting, walking, calculated the mean for each family (male, fe- comfort behavior (preen, stretch, shake, or male, juveniles) by sampling period and then scratch), vigilance (neck stretched upward to calculated the means ofthe three sampling pe- full length), and agonistic interaction (intra- riods. Parental vigilance is the mean of male specific aggressive encounters). Sampling was and female vigilance. Some post-hatching conducted between 0900 and 1600 hrs (Cen- brood amalgamation took place, at times tral European Summer Time) and only when forming families including 3 adults. One families were on land. If disturbed during “family” of 13 juveniles was led by 4 adults. sampling (e.g., by dogs; Randier 2()03a), fam- I did not use amalgamated families in the ilies escaped into the water and sampling con- analyses. tinued (if necessary) after the geese returned I expressed behaviors (e.g., feeding, vigi- to land. Some bias may be present because lance, etc.) as percentages oftotal time budget data collection was only done during certain (square-root arcsine transformed). To compare daytime periods. This seems unlikely, because percentages between different groups, I used time of day does not strongly influence be- the Mann-Whitney U test on untransformed havior of families with goslings (Lazarus and data and, to compare dependent variables, I Inglis 1978, Forslund 1993, Schmutz and used Wilcoxon matched-pairs signed rank test. Laing 2002). 1 used Pearson’s correlation to examine the re- 1 separated samplings into 4-week periods lationship between vigilance and brood size of brood rearing (weeks 1-4, 5-8, and 9—12), (logio transformed). I used a general linear because parental investment may differ be- model (GLM) with year and period as fixed tween these periods (Forslund 1993) and gos- factors, number of juveniles (logu,) as a co- lings were more prone to predation during variate and parental care (vigilance) as the de- their first weeks of life (Owen 1980). 1 chose pendent variable. 1 used as a measure of these sampling periods because at 8 weeks, explained variance. I used SPSS version 11.0 most juveniles were close to fledging (i.e., ca- to analyze the data (Biihl and Zofel 2002) and pable of sustained flight; Kolbe 1999). Family set statistical significance at P < 0.05. bonds extend over the brood rearing period RESULTS and sampling during week 9—12 allows com- parison of the brood rearing period with the The first goslings appeared on the feeding post-fledging period. grounds during the last 10 days ofMay. Brood Some studies report between-year differ- sizes did not remain stable during the study ences in time budgets (Schmutz and Laing period because of predation. I observed two 2002). Because vigilance did not differ be- unsuccessful predation attempts, one by Car- tween years 1 pooled years. An unknown rion Crows {Corvus c. corone) and one by numberofindividuals may have been sampled Yellow-legged Mediterranean Gulls {tarns in both years and my data may include rep- niichahellis), both of which bred nearby. licated observations of the same birds. In ad- Brood sizes were 4.3 ± 1.4 (jc ± SE) in 2002 dition to the 6 (2002) and 7 (2003) families and 3.5 ± 0.7 in 2003 during week 1-4. Randier* PARENTAL INVESTMENT IN SWAN GliESE 25 TABLE 1. Time budgets (%, means ± SE) ofmale, female, andJuvenile Swan Geese during tliree different periods ofbrood rearing in 2()()2 and 2003 in Heidelberg, Germany. Each family was sampled between one and five times during each period. The mean of each family per period was used to calculate percentages to not over-represent some families. Differences between either males or females and goslings are expressed as *R < 0.05; < 0.01. Feeding Resting Walking Comfort Vigilant Agonistic Families Weeks 1-4 n = 13 Male 27.7 ± 3 3.2 H- 1.6 2.5 ± 0.8 10.6 ± 4.0 51.1 ± 4 1a** 4.2 -F 1.0 Female 41.4 ± 4 4.8 -H 2.0 2.2 -F 0.7 7.0 ± 2.1 39.9 -F 4 3.8 -F 1.1 Juvenile 75.5 ± 5.0 1 1.2 -F 3.9 5.7 -F 1.8 5.9 + 1.9 0.6 -F 0.2 0.0 0 Parental 34.6 ± 3.4 4.0 -F 1.5 2.4 -F 0.7 8.8 2.7 45.5 -F 3.7 4.0 -F 0.9 Families Weeks 5-8 /7 = 8 Male 35.1 ± 7.6 ± 3.4 4.4 ± 1.5 15.4 ± 4.5 34.5 -F 5.9“* 1.5 ± 0.7 Female 29.4 ± 7.3'’* 9.6 ± 3.0 4.4 ± 1.5 23.6 + 5.0 29.5 -F 4.1“* 1.7 ± 0.7 Juvenile 62.8 ± 12.0 10.1 + 3.8 2.6 -F 1.1 18.8 ± 8.7 4.1 + 0.8 0.0 ± 0 Parental 32.2 ± 7.5 8.6 — 3.0 4.4 — 1.5 19.5 4.7 32.0 -F 3.8 1.6 0.7 Families Weeks 9-12 n = 9 Male 31.8 ± 6.9'’** 14.7 -F 6.5 3.4 -+- 1.3 17.2 + 5.2 29.8 -F 4.0^** 1.5 ± 0.7 Female 37.4 ± 15.3 -F 5.7 3.8 ± 1.4 13.7 ± 5.0 24.4 ± 2.6a** 3.2 ± 1.3 Juvenile 57.1 ± 9.2 23.1 ± 8.8 3.6 ± 1.0 12.2 ± 5.8 2.8 ± 1.1 0.1 ± 0.1 Parental 34.6 ± 5.3 15.0 6.0 3.6 — 1.3 15.4 4.4 27.1 -F 2.6 2.3 0.7 Families Weeks 1-12 n = 30 Male 30.9 ± 3.3 7.8 ± 2.3 3.3 ± 0.6 13.9 ± 2.6 40.3 ± 3.1 2.7 -F 0.6 Female 37.7 ± 3.3 9.3 2.1 3.3 ± 0.6 13.4 ± 2.4 32.2 + 2.7 3.0 -F 0.6 Juvenile 66.6 ± 4.8 14.5 ± 3.3 4.3 ± 0.9 1 1.2 ± 3.0 2.2 -+• 0.4 0.0 -F 0.0 Parental 34.0 ± 2.8 8.5 ± 2.1 3.3 ± 0.6 13.6 ± 2.2 36.4 ± 2.2 2.8 -F 0.5 Non-breeders 39.1 ± 6.1 22.1 ± 6.8 1 1.3 2.2 14.4 ± 4.1 12.8 2.3 0.0 ± 0.0 /7 = 17 ^Valuehigherthangoslings. Non-breedersdepictedforcomparison. Valuelowerthangoslings. Non-breedersdepictedforcomparison. Means per period (both years pooled) were ing brood rearing, males reduced their vigi- 3.9 ± 0.7 (week 1-4), 3.6 ± 1.1 (week 5-8), lance between periods 1 and 3 (Wilcoxon test: and 3.6 ± 1.0 (w—eek 9-12). Z = -2.201, P = 0.028, n = 7), and their Time Budgets. Non-breeding adults spent agonistic behavior between periods 1 and 2 (Z less time vigilant (Table 1) compared with = — 1.997, P — 0.046, /? = 6). Otherbehaviors male (Mann-Whitney U test: Z = -4.584, P did not change {P > 0.05). Females reduced < 0.001, n = 30), female (Z = -3.998, P < agonistic behavior between periods 1 and 2 (Z 0.001, n = 30), and parental vigilance (mean = —1.892, P = 0.05, n = 6) and vigilance of male and female in each pair: Z = —4.416, between periods 1 and 3 (Z = —2.197, P — P < 0.001, n = 30; based on the overall 0.028, n = 7). — means from weeks 1-12). Within families, fe- Parental Care. Mean parental care (vigi- males, spent more time feeding (Wilcoxon test lance) per period varied (Table 1). There was Z = —2.750, P = 0.006, n — 13) during weeks correlational evidence for an adjustment of 1-4 but not in weeks 5-8 and 9-12 {P > 0.05) parental vigilance to brood size for periods 1 and, a lower proportion of time vigilant than and 3 (period 1: r = 0.557, P = 0.048, n = males during weeks 1-4 (Z = -2.202, P - 13; period 2: r = 0.617, P = 0.10, n = 8; 0.028, n = 13), but not during periods 2 and period 3: r = 0.753, P = 0.019, n = 9). Vig- 3 (Wilcoxon-test, P > 0.05; Table 1). ilance was dependent on brood size and period Goslings fed more than both their parents but not year (Total model: F = 7.847, P < (,23 and their vigilance was lower (Table 1). Dur- 0.001; brood size F, = 16.599, P < 0.001; 26 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 119, No. 1, March 2007 period: F = 10.051, P = 0.001; year: F,= of Field Ornithologists’ program of editorial assis- 2 2.446, P = 0.13; all interaction terms: P > tance. Comments from three anonymous referees im- 0.10). Adults of larger broods were more vig- proved an earlier version ofthe manuscript. ilant and vigilance declined through the stages LITERATURE CITED of brood rearing. The total amount of ex- plained variance was high {R^ = 0.672, cor- Afton, A. D. and S. L. Paulus. 1992. Incubation and = rected 0.586). brood care. Pages 2-108 in Ecology and manage- DISCUSSION ment of breeding waterfowl (B. D. J. Batt, A. D. — Afton, M. G. Anderson, C. D. Ankney, D. H. Time Budgets. Non-breeders were less Johnson, J. A. Kadlec, and G. L. Krapu, Editors). vigilant, similar to the findings ofothers (Les- University of Minnesota Press, Minneapolis, sells 1987, Forslund 1993). Adult geese ofdif- USA. ferent species with broods usually spend be- Altmann, j. 1974. Observational study ofbehaviour: tween 15 and 45% of their time feeding and sampling methods. Behaviour 49:227-267. 40-45% vigilant to look for predators to pro- Austiilni,esJ.an1d99p0a.irCsoomfpwairnitseorningofCaacntiavdiatieGseewsiet.hiWnilfsaomn- tect and warn their goslings (Afton and Paulus Bulletin 102:536-542. 1992). Other studies also found marked dif- Baldassarre, G. a., S. L. Paulus, A. Tamisier, and ferences among adult males, females, and ju- R. D. Titman. 1988. Workshop summary: tech- veniles within broods (Austin 1990, Schmutz niques for timing activity ofwintering waterfowl. and Laing 2002). Females spend more time Pages 181-188 in Waterfowl in winter (M. W. Weller, Editor). University of Minnesota Press, feeding than males (Lazarus and Inglis 1978, Minneapolis, USA. Lessells 1987, Sedinger and Raveling 1990). Bauer, K. M. and U. N. Glutzvon Blotzheim. 1968. Males, in turn, spend more time vigilant, sim- Handbuch der Vogel Mitteleuropas. Bd. 2, Anser- ilar to the results of the present study. Juve- iformes 1. Teil. Akademische Verlagsgesellschaft. niles fed during a large part oftheir time sim- Franfkurt/Main, Germany. ilar to other goose species (Afton and Paulus Buhl, A. and P. Zofel. 2002. SPSS. Version 11. Ein- fuhrung in die moderne Datenanalyse unter Win- 1992), because juveniles have higher nutri- dows. Addison-Wesley, Miinchen, Germany. tional demands. Afton and Paulus (1992) also Delacour, j. 1954. The waterfowl of the world. Vol- present examples for decreasing vigilance dur- ume I. Country Life, London, United Kingdom. ing maturation of broods (also Lazarus and Forslund, P. 1993. Vigilance in relation to brood size Inglis 1978). I also found a decrease in vigi- and predator abundance in the Barnacle Goose, lance in Swan Geese. Thus, my study shows Branta leucopsis. Animal Behaviour45:965-973. that introduced geese have similar behavioral Goroshko, O. a. 2001. Swan Goose in eastern Trans- baikalia and Mongolia. Kasarka 7:68-98. patterns during br—ood rearing as wild geese. Kolbe, H. 1999. Die Entenvdgel der Welt. Ulmer, Parental Care. Parental vigilance during Stuttgart, Germany. the brood rearing period was related to brood Lazarus, J. and I. R. Inglis. 1978. Breeding behav- size and this relationship extended into the iour of the Pink-footed Goose: parental care and post-fledging period. These results support the vigilant behaviour during the fledging period. Be- haviour 65:62-88. “shared” parental investment hypothesis. Dif- Lessells, C. M. 1987. Parental investment, brood size ferences among studies about parental invest- and time budgets: behaviour of Lesser Snow ment and brood size may be caused by brood Goose families. Ardea 75:189-203. size as my study covered a wide range offam- Madge, S. and H. Burn. 1988. Wildfowl. Parey,Ham- ily sizes from 1 to 10 goslings. Age of gos- burg, Germany. lings may have an important role in affecting McWilliams, S. R. and D. G. Raveling. 1998. Hab- vigilance behavior (Forslund 1993). itat use and foragingbehaviorofCacklingCanada The major conclusion of the study is that athnedaRnoaslsys’isgeoefseecodluorgiincgalspdreitnge:rmiinmapnlticsatoifongsoofsoer introduced Swan Geese have similar parental social behavior. Pages 167-178 in Biology and care and division of labor between gender management of Canada Geese. (D. H. Rusch, M. compared to wild Swan Geese, and to other D. Samuel, D. D. Humburg, and B. D. Sullivan, Anser species. Editors). Proceedings ofthe International Canada ACKNOWLEDGMENTS Goose Symposium, Milwaukee, Wisconsin, USA. Ogilvie, M. and S. Young. 1998. Photographic hand- I appreciate the improvements in English usage book ofthe wildfowl ofthe world. New Holland, made by George Farnsworth through the Association London, United Kingdom. — Randier • PARENTAL INVESTMENT IN SWAN GEESE 27 Owen, M. 1980. Wild geese of the world. Fakenham ScHMUTZ, J. A. AND K. K. Laing. 2002. Variation in Press, Fakenham, United Kingdom. foraging behavior and body mass in broods of PoYSA, H. 1991. Measuring time budgets with instan- EmperorGeese {Chen canagica): evidence for in- taneous sampling: a cautionary note. Animal Be- terspecihc density dependence. Auk 19:996- haviour 42:317-318. 1 1009. Randcilnyegran,noiCud.rebs2a0ni0n3pHaoe.piudlReaeltabiceotnrigoon(fsSttWhoeGhSeuwrmamaannnyGd)oi.ostsAuecrtbAaannsOcere-rs Sedinbgeehra,vioJ.rSo.faCnacdklDi.ngG.CaRnaavdealiGnege.se19d9u0r.inPgarbernotoadl nithologica 38:47-52. rearing: division oflabor within pairs. Condor92: Randler, C. 2003b. Vigilance in urban Swan Geese 174-181. and their hybrids. Waterbirds 26:257-260. SiRiWARDENA, G. M. AND J. M. Black. 1999. Parent Rutschke,—E. 1997. Wildganse. Lebensweise and gosling strategies in wintering BarnacleGeese Schutz Nutzung. Parey, Berlin, Germany. Branta leucopsis. Wildfowl 49:18-26.