Wilson Bulletin, 110(1), 1998, pp. 56-64 EFFECTS OF MALE REMOVAL ON FEMALE REPRODUCTIV3E BIOLOGY IN ROSS’ AND LESSER SNOW GEESE CRAIG R. LeSCHACK,13 ALAN D. AFTON,14 AND RAY T. ALISAUSKAS — ABSTRACT. We studied effects of mate removal on nesting and hatching success, incubation behavior, body mass, and post-hatch dispersal distance of female Ross’ (Chen rossii) and Lesser Snow Geese (C. caerulescens caerulescens at Karrak Lake, N.W.T., Canada. Male geese were removed during early incubation (days 1-8), ) and widowed and paired control females were monitored through post-hatch dispersal. Nesting and hatching success did not differ between species or treatments (widowed vs paired) and averaged 77.5 ± 3.8% and 64.0 ± 3.6% (±SE), respectively. Paired females spent more time with their bills tucked (23.7 ± 3.3% vs 9.1 ± 4.0%) and less time alert (8.6 ± 2.9% vs 22.9 ± 3.5%) while on nests than did widowed females. Snow widowed females (31.1 ± 4.7%) and Ross’ widowed females (20.6 ± 6.0%) generally spent more time each day in head- up alert than did Snow paired females (7.1 ± 3.8%), Snow paired males (11.8 ± 3.8%), Ross paired females (9.4 ± 3.6%), and Ross’ paired males (7.9 ± 3.6%). Body mass of paired and widowed female Ross Geese did not differ at hatch or at time of post-hatch recapture; however, mean distance recaptured from the breeding colony was greater for paired (50.9 ± 6.1 km) than for widowed females (27.3 i 6.6 km). Total mass gain (276 ± 19 g) and rate of mass gain (8.4 ± 0.5 g/day), from hatch until post-hatch recapture (33.1 ± 1.2 days), were similar for widowed and paired female Ross’ Geese. Male removal experiments in monogamous, precocial We species generally have produced few effects on female nesting success or incubation behavior. suggest that male parental care in arctic-nesting geese is more critical during laying and the post-hatch period than during incubation. Received 21 June 1996, accepted 29 June 1997. Swans and geese (Anserini), and whistling Oring and Sayler 1992). Although most geese ducks (Dendrocygnini) form long-term pair- pair for life, opportunities for extra-pair mat- bonds in which both parents care for young. ings exist; thus, effective mate guarding and Perennial monogamy is associated with large territorial defense by males should enhance body size, high probability of mate survival, their certainty of paternity and may deter pre- lack of renesting opportunities, and obligate dation. Consequently, mate loss, intrusion of brood-rearing (Oring and Sayler 1992). Male other males, and costs of re-pairing reduce geese defend territories and mates during in- lifetime reproductive success (Owen et al. cubation, and subsequently protect broods 1988, Forslund and Larsson 1990) and, there- while females replenish nutrient reserves used fore, could select for male attendance during during laying and incubation (Ryder 1975; incubation (Paine 1992). Lazarus and Inglis 1978; Ankney 1977, 1979; Ross’ (Chen rossii) and Lesser Snow Geese Sedinger and Raveling 1990; Afton and Pau- (C. caerulescens caerulescens hereafter Snow ; lus 1992; Paine 1992). Arctic-nesting geese Geese) nest sympatrically in the central Ca- face ecological conditions that may favor ex- nadian Arctic, providing a unique opportunity tended male parental care, including colonial for comparative studies of similar species. nesting, herbivory, and short breeding seasons Ross’ Geese are smaller than Snow Geese (Martin et al. 1985, Gauthier and Tardif 1991, [mean body mass of nesting adults (sexes combined) = 1356 g and 2029 g, respectively; 1 U.S. Geological Survey, Biological Resources Di- Maclnnes et al. 1989J. Daily energy require- vision, Louisiana Cooperative Fish and Wildlife Re- ments, clutch size, and vulnerability to pre- search Unit, Louisiana State Univ., Baton Rouge, LA dation are influenced by body mass via stored 70803-6202; e-mail: [email protected]. nutrient reserves (Barbault 1986). Female 2 Canadian Wildlife Service, Prairie and Northern Wildlife Research Centre, 115 Perimeter Road, Sas- geese generally feed little during egg-laying katoon, Saskatchewan, S7N 0X4, and Dept, of Biolo- and incubation, relying primarily on endoge- gy, Univ. of Saskatchewan, Saskatoon, Saskatchewan nous reserves to complete their clutches (Ry- S7N 0W0, Canada. der 1970a, Ankney and Maclnnes 1978). 3 Present address: Florida Game and Fresh Water However, small species of geese (e.g., Ross’ Fish Commission, 2690-E South Ponte Vedra Blvd., Ponte Vedra Beach, FL 32082. Geese) cannot store as much endogenous re- 4 Corresponding author. serves as large species (but see Bromley and 56 LeSchack et al. • EFFECTS OF MALE REMOVAL IN GEESE 57 Jarvis 1993); consequently small species gen- the transect (left or right) in which to select nests. A erally take more recess time during incubation total of 120 nests (60 Ross’, 60 Snow) were selected for the experiment and individually marked with small to feed than do large species (Aldrich and colored flags. We subsequently recorded final clutch Raveling 1983; Thompson and Raveling size, embryo age (Weller 1956), and estimated first egg 1987; Alton and Paulus 1992; Afton, unpubl. date (assuming a laying rate of 1 egg every 1 .3 days data). for both species; Ryder 1970b). Each egg was num- We investigated the importance of male bered in several places and lines were drawn around Ross’ and Snow Geese to nesting and hatch- the egg with a permanent marker to help determine ing success, and to incubation behavior, body hatching success (see below). Nests were observed ev- We ery 3—5 days with a spotting scope to determine mass, and post-hatch dispersal of females. whether females were incubating their clutches. If a predicted that widowed females would have female was absent, we visited her nest immediately to lower nesting and hatching success than determine condition of the clutch.— would paired females in both species. Snow Nesting and hatching success. After parents and Geese are larger and can store more nutrient broods left the colony, we estimated nesting and hatch- reserves than can Ross’ Geese; consequently, ing success of all nests. A nest was classified as suc- we reasoned that widowed Snow Geese would cessful if at least 1 egg hatched. Hatching success was defined as percentage of eggs within a clutch that be capable of incubating for relatively longer hatched. We classified an egg as hatched ifany marked periods and, therefore, should have greater fragment (number or line) of eggshell, pieces of egg success than should widowed Ross’ Geese. membrane, and/or gosling down were present in or We also predicted that behavior during incu- near the nest. — bation would differ between paired and wid- Male removals. Experimental nests were randomly owed geese, if paired females benefit from assigned to one of two treatments; paired controls or widowed. Thirty-five nests of each species were allo- mate-guarding. Benefits of male protection cated to control groups, and 25 nests of each species could include predator deterrence, reduced were assigned to widowed groups. Males assigned to disturbance by neighboring females (i.e., nest the widow treatment were shot during early incubation parasitism), reduced harassment from males (days 1-8) and used for other studies. We were unable attempting forced extra-pair copulations, and to collect the entire sample of male Ross’ Geese be- increased foraging efficiency during incuba- cause of time constraints; consequently, we had 37 tion recesses. If male protection is beneficial conItnrcoulbaatnidon23bewhiadvoiwoerd—fWemealoebsseforrvetdhiswsipdeociweesd. and during incubation, then paired females might . paired birds using focal-animal sampling techniques be in better physiological condition at hatch (Altmann 1974). We divided the day into 6 time pe- than are widowed females. Finally, if parental riods: 00:01-04:00. 04:01-08:00, 08:01-12:00, 12:01— care by male geese is important during brood 16:00, 16:01-20:00, and 20:01-24:00 CST. We ran- rearing, then removing males at hatch should domly selected nests for observation during 2 of the 6 affect female condition, and gosling survival periods each day. We recorded activities of widowed and growth. We predicted that if widowed fe- f2e1m-ales or paired females and their mates every 10 s during 15-min observation periods. Activities were males assumed sole responsibility for protect- categorized as alert (head-up or extreme head-up pos- ing their broods while concurrently replenish- t2u2r—es; Lazarus and Inglis 1978), foraging, walking, ing nutrient reserves and molting, they would swimming, flying, comfort movements (preening, be in poorer condition than would paired fe- stretching, etc.; McKinney 1965), aggression (threats, males during the post-hatch period. calls, chases, and forced copulations), nest attendance (female only), and absent from territory. Behavior of STUDY AREA AND METHODS females while on nests was subdivided into 4 catego- ries: head-low (non-alert behavior; Lazarus and Inglis We studied Ross’ and Snow Geese from 27 May to 1978), alert (head-up or extreme head-up), bill-tucked, 9 August 1994 at Karrak Lake, N.W.T., Canada (67° and comfort movements. 14' N, 100° 16' W). The area is typical tundra habitat Body—mass and dispersal distance offemale Ross' with numerous shallow lakes and streams (Ryder Geese. Wc trapped a random sample of paired con- 1972). Both species nest in areas of rock, heath, moss, trol females (N = 16) at the end of incubation (days or a mixture of these habitats located on islands or the 23) to compare body mass with that of widowed mainland (McLandr—ess 1983). females. Females were captured using remote-con- Nest selection We placed transects randomly trolled, modified bow-net traps (B. M. Grand, pers. . through a 1-km2 section of the colony located on the comm.). Widowed females (N = 10) were shot at days mainland and selected nests every 30 m along tran- 23 of incubation and used for other studies. All sects. We randomly selected species and direction from trapped and collected geese were weighed (± g) and 1 n n 58 THE WILSON BULLETIN • Vol. 110, No. 1, March 1998 measured (mid-wing, tarsus bone, and head length; ries: early (days 1-8), mid (days 9-15), and late (days ±0.1 mm; Dzubin and Cooch 1992). 16-23). We used means of behavioral observations for We attached radio transmitters to random samples the few nests that were observed more than once. Sig- MANOVA of paired control females (n = 16) and paired females nificant explanatory variables from were whose mates subsequently were removed at hatch ( used in ANOVAs to examine effects on individual be- = 13) to obtain post-hatch dispersal distances. Trans- haviors. Lor females, four behaviors were used in the mitters weighed an average of 26 g and were attached analysis: nest attendance (i.e„ incubation constancy), using a modified backpack harness (Dwyer 1972). foraging, absent from territory, and alert posture (while Captured females also were fitted with neck collars and off nests). These behaviors accounted for more than standard leg bands. Goslings ( = 33 of paired control 98% of female activities. We further analyzed head females, n = 32 of females widowed at hatch) were posture of females while on nests (head-low, comfort web-tagged through the eggshell during the pipping movements, bill-tucked, and alert). In addition, we stage (Alliston 1975) or at hatch using metal web-tags compared time spent alert (head-up or extreme head- - to allow comparisons of gosling growth and survival up) each day among experimental female groups (k rates. Geese were relocated 29-42 days after hatch us- 4 groups) and male geese (k = 2). We compared LSMs ing telemetry; aerial antennae were mounted on a hel- in behavioral analyses because of unbalanced designs icopter, and flightless geese were captured by driving (PDILL option, SAS Institute 1990). Analyses of raw them into a portable net corral. Lemales and web- and transformed proportion data yielded similar results tagged goslings were banded, collared, weighed and in final models; therefore, we report results from anal- measured. Recapture distance from the breeding col- yses of raw data (LSM ± SE). ony was determined using a Global Positioning Sys- Because body mass often is related to structural size tem. As a result of time, permit and funding con- and body size is positively correlated with post-hatch straints, we were unable to trap, radio track or collect dispersal distance in Ross’ Geese (Slattery 1994), we female Snow Geese for analysis ofbody mass and dis- indexed female body size using Principal Components persal distance. — Analysis (PCA; Proc PRINCOMP, SAS Institute Statistical analysis. We used analysis of variance 1990). Body size was defined as the first principal (ANOVA) to test for differences in first egg date, component (PC computed from the correlation ma- clutch size, predicted hatch date, and hatching success trix of mid-wing1,) tarsus bone, and head length mea- between species, treatments, and their interaction (Proc surements. We ran separate PCAs to estimate body size GLM, SAS Institute 1990). We compared least-square for females at hatch and during post-hatch dispersal; muneeaqnusal((LPSDMI)LLusoipntgion/-,teSstAsSwIhnesntitustaemp1l9e90)siazensd wuesreed all loadings were positive and explained 80—84% of the original cumulative variance. Tukey's studentized range test to compare unadjusted We used analysis of covariance (ANCOVA) to test means when sample sizes were equal (Sokal and Rohlf 1981 Because of time constraints, we were unable to for differences in body mass and post-hatch dispersal ). distance between treatments, with PCI (i.e., body size) accurately estimate first egg date for 18 of the 120 used as a covariate in models. We used ANCOVA to experimental nests. These included four paired Ross’, test for differences in mass change (from hatch to post- five widowed Ross’, three paired Snow, and six wid- hatch recapture) between treatments, with days elapsed owed Snow Goose nests. Accordingly, these nests since hatch, recapture distance, and PC as covariates. were excluded from analyses of first egg date and pre- We also used ANCOVA to test for diff1erences in rate dicted hatch date. We tested whether nest success differed between of mass change between treatments, with recapture dis- tance and PC as covariates. Pearson correlation co- species, treatments, and their interaction using maxi- 1 mum-likelihood ANOVA (Proc CATMOD, SAS Insti- efficients (r) were used to describe relationships be- tute 1990). We present apparent nest success (number tween PCI and various response variables (Proc of successful nests/total nests) because we began mon- CORR, SAS Institute 1990). Linally, we tested for sta- itoring all nests during early laying. We also used max- tistical significance of the combined results ofours and imum-likelihood ANOVA to test whether number of other goose studies, to determine whether nest success of widows was lower than that ofcontrols, using Lish- nests that had partial clutch reduction during incuba- tion (but still hatched >1 egg) differed between spe- er’s inverse x2 method (Hedges and Olkin 1985:37). cies, treatments or their interaction. We computed proportion of time spent in various RESULTS behavioral activities by dividing the frequency that each behavior was recorded by the maximum number First egg—date, clutch size, and predicted of behaviors possible per 15-min sampling period (n hatch date. First egg date did not differ be- = 90). We analyzed raw and arcsine square-root trans- tween species (F — 2.82; df = 1, 98; P > tfiovramrieadtepraonaployrstiisonosf(vSaorkiaalncaend(RMoAhlNfOV19A8)1 )tuosienxgammiunl-e 0>.05) or treatments (F = 0.62; df = 1, 98; P 0.05); the species-by-treatment interaction differences in overall time-activity budgets by species, treatment, stage of incubation, and associated interac- also was not significant (F = 0.52; df = 1, tions. Stage of incubation was divided into 3 catego- 98; P > 0.05). Modal first egg date for Ross’ — FF F F F F LeSchack et al. • EFFECTS OF MALE REMOVAL IN GEESE 59 TABLE 1. Fate of experimental nests at Karrak Lake, N.W.T., Canada, 1994. Ross’ Geese Snow Geese Control Widow Control Widow Fate ofnest n % n % n % n % LInsuccessfuF 7 18.9 8 34.8 6 17.1 6 24.0 Successful 30 81.1 15 65.2 29 82.9 19 76.0 Clutch reduction11 8 21.6 3 13.0 9 25.7 6 24.0 Total nestsc 37 23 35 25 aNests abandoned, orpreyed upon and entire clutch lost. bReduction occurred during incubation, but a1 egg hatched in nest. cUnsuccessful + successful nests. — (39% of 51 nests) and Snow Geese (41% of Female behavior during incubation. We 51 nests) was 31 May. collected focal-animal samples from 44 fe- Mean clutch size of Snow Geese (3.95 ± male Ross’ (n = 27 control, 17 widow) and 0.08; n = 60) was larger (F = 16.64; df = 1, 46 female Snow Geese (n — 27 control, 19 116; P < 0.001) than that of Ross’ Geese widow) during incubation. Overall percent (3.43 ± 0.08; n = 60), but the treatment main time spent in nest attendance, alert (off nest), effect (F = 0.14; df = 1, 1 16; P > 0.05) and absent from territory, and foraging by females = treatment-by-species interaction were not sig- did not differ between species (F 0.50; df nificant ( = 0.64; df = 1, 116; P > 0.05). = 4, 129; P > 0.05) or treatments ( = 1.38; Predicted hatch date did not differ between df = 4, 129; P > 0.05), or among incubation species ( = 0.01; df = 1, 98; P > 0.05) or stages (F = 1.60; df = 8, 258; P > 0.05; Table treatments ( = 1.25; df = 1, 98; P > 0.05), 2); none of the interactions of various main and the species-by-treatment interaction also effects were significant (all P > 0.05). Nest was not significant (F = 0.18; df = 1, 98; P attendance for both species combined aver- > 0.05). Modal predicted hatch date for Ross’ aged 91.8 ± 3.0%. (55% of 51 nests) and Snow Geese (45% of Head postures of incubating females dif- 51 nests) was 27 June. First egg and predicted fered between treatments (F = 4.72; df = 4, hatch dates were 8-16 days earlier in 1994 as 130; P < 0.002) and among incubation stages compared to previous years (1991-1993) at (F = 3.09; df = 8, 260; P < 0.003), but were Karrak Lake (Slattery and Alisausk—as 1993). similar between species (F = 0.80; df = 4, Nesting and hatching success. Nesting 122; P > 0.05); none of the interactions were success did not differ between species (x2 = significant (all P > 0.05). Paired females 0.53, df = 1, P > 0.05) or treatments (\2 = spent more time with their bills tucked (F = 1.98, df = 1, P > 0.05); the species-by-treat- 8.32; df = 1, 133; P < 0.005), and less time ment interaction also was not significant (x2 alert while on nests than did widows (F — = 0.21, df = 1, P > 0.05; Table 1). Nesting 10.62; df = 1, 133; P < 0.002; Table 2). Time success for both species combined was 77.5 spent with bill tucked under the wing differed ± 3.8% (N = 120). Nests with partial clutch among incubation stages (F = 9.32, df = 2. reduction did not vary by species (x2 = 1.04, 133, P < 0.001). Females spent more time df = 1, P > 0.05) or treatments (x2 = 0.54, with bill tucked during late incubation than df = 1, P > 0.05), and the species-by-treat- during early or mid incubation (Table 2). ment interaction was not significant (x2 = Alert behavior of males and females. 0.29, df = 1, P > 0.05; Table 1). Overall time spent alert each day differed Hatching success did not differ between among control Ross’ females, widowed Ross’ species (F = 0.12, df = 1, 1 16, P > 0.05) or females, control Snow females, widowed treatments (F = 0.12, df = 1, 116, P > 0.05); Snow females, Ross’ males, and Snow Goose the species-by-treatment interaction also was males ( = 3.08; df = 10, 390; P < 0.001); not significant (F = 0.10, df — 1, 116, P > however, stage of incubation ( = 0.96; df = 0.05). Hatching success for both species com- 4, 390; P > 0.05) and stage-by-group inter- bined averaged 64.0 ± 3.6% (N = 120). action were not significant (F = 0.99; df = n ; i 60 THE WILSON BULLETIN • Vol. 110 No. I, March 1998 , 0) C3 TABLE 3. Percent time spent in 1head-up alert (least-square means ± SE) during incubation by ex- CQ C3 < < < < < < m— perimental group at Karrak Lake, N.W.T., Canada, c o Toj- o >on ooi dind 1994. *6 +1 +1 +1 +1 +1 +1 +1 QJ ion oi oo~ on- o(N <dN o— Group n Head-up C</D3 Ross’ widow females 28 20.6 ± 6.0ABa <CDD Snow widow females 32 31.1 ± 4.7A Ross’ paired females 40 9.4 ± 3.6C v<o <in <m o<o o< ><n <O) Snow paired females 37 7.1 ± 3.8C oi oi oi IN d d oi Ross’ paired males 40 7.9 ± 3.6C u00 iN+d-1 md+1 md+1 9d+1 id+1n oo+1o id+o1 SnaLoeawst-psaqiuarreedmemaanslewsithdifferent le3tt7ersaresig1ni1f.i8can±tly3di.ff8erBenCt(P< C3 0.05). < < < < < < < X — — o or- ovo ovO (oN —i or~ — 20, 390; P > 0.05). Time spent head-up alert -cco3 c—+o1 mo+1 i—+n1 oo+1i No+1 dO+01di+1t edfac=h d5,ay19d6i;ffPere<d 0a.m00o1n)g, wghreoruepass(tFim=e s4p.e42n;t in extreme head-up alert was similar among <£L> groups (F = 1.40; df = 5, 196; P > 0.05) and a < < < < < < < averaged 0.9 — 0.2%. Widowed Ross and <D o o o— o<N OoI d—HdOl Snow Goose females generally spent more CoO +1 +1 +1 +1 +1 +1 +1 time in head-up alert than did geese in other T> vo in 0o0 drodin <Odu o o o o groups (Table 3). C/3 Body mass and post-hatch dispersal dis- X — tance offemale Ross’ Geese. Body mass at W C+/01 dq< d9< cd<q odcq iCQn rvCoqQ ^<<n h(aFtc=h 9w.a4s8;podsfit=ive1l,y2(3r;=P0<.540).0r0e6l)a,tebduttomPaCsIs = = P > C3Oe/J3 oi+n1 rd+-1; rnt+o-1 O—+1n dv+oi od+\i d+i 0d.i0d5)nobtetdiwfefeern (pFaired0.(0919;5 d±f 18 1g,, 2«3;= 16) <u -h OI and widowed females (997 ± 23 g, n = 10). £ 3 Overall mean mass at hatch, unadjusted for cCCC<//I3u33r do<i d—< (COQNN rI</oO o<dv d—<< dq< sizeB,ecwaausse99o6f l±imi1t6edgh(enli=cop2t6e)r.time, we cap- tured only 13 radio-tagged females (5 paired CUo/.3 i+1n O+1) V+O1 O+N1 9+! o+i1 o+i1 controls, 8 widowed at hatch). Only one fe- ’c>d d d oo r<in d d d male (paired control) recovered in banding XJ = X<u drives had web-tagged goslings (n 2 out of C3/3 4 goslings hatched); one other control female o < < < < < < < escaped before being weighed. Body mass at odo d rdnf (IFN3 dd dd Td recapture was positively (r — 0.82) related to +1 +1 +1 +1 +1 +1 +1 PCI (F = 19.22; df = 1, 9; P < 0.002), but c<CCL/l>3, o0ii0n Td>fn otdTfn) riP--n’ didn vvdo di9n m>as0s.05d)idbentowtedeinffpearir(eFd =(120.6067;±d4f0=g; 1n, =9; 4P) 0£/ and widowed females (1279 ± 28 g; n = 8). C 44 46 54 36 26 63 47 Overall mean mass at recapture was 1275 ± Tj- = <£D OOnN 37 g (n 12). <D Mass change of females was not related to CL C3 ± "O days elapsed since hatch (33.1 1.2 days. aSc3 Range = 29—42 days, n = 12; F = 0.20; df CH<Q oD '56cc//33 2Oc c<scsD3 Ub2co 5-£*oo <0D0 wS s 0=.011;, d7;f =P 1>, 70;.0P5)>, r0e.c0a5p)tuorrePdCiIst(aFnce= (0F.75=; CO/)- CO <u c/5 df = 1, 7; P > 0.05), and mass change did P LeSchack et al. • EFFECTS OF MALE REMOVAL IN GEESE 61 TABLE 4. Summary of studies investigating the influence of male removals on reproductive parameters in precocial birds. Impact ofMale-removal Nest Brood Brood Female Female Species (Citation) success survival growth survival body mass — Lesser Snow Geese (Martin et al. 1985) —Varied0 Varied*5 C — —Decline Bar-headed Geese (Schneider and Lamprecht 1990) Decline —NSd — Canada Geese (Paine 1992) NS —Decline — — —NS Wood Duck (Hipes and Hepp 1993) NS — — Willow Ptarmigan (Hannon 1984) —Varied' —NS — Decline Willow Ptarmigan (Martin 1984) — NS Increase Willow Ptarmigan (Martin and Cooke 1987) NS —NS — —NS —Varied* Western Sandpiper (Erckmann 1983) Decline — — — Ross’ Geese (This study) NS — — — N—S Lesser Snow Geese (This study) NS aAt Boas River, paired females had higher (P < 0.001) nest success than did widows whose mates were removedduring early laying, but nestsuccess was similar when males were removed during late laying ( > 0.051) and during incubation (P > 0.056). At La Perouse Bay, nest success was similar for paired females and widows whose males were removed in early incubation (P > 0.05). bSuggested decline at La Perouse Bay in 1983, but no difference in 1984. cData not collected. dData not significant. eNest success declined in 1981, but did not differ in 1979 and 1980. 1 Widows were heavier than paired females at hatch in 1982; trends in 1981 and 1983 were similar, but not significantly different. Widows lost more weight than did paired females in the post-hatch period. not differ (F = 0.31; df = 1,7; P > 0.05) would have to be doubled (i.e., 120 nests for between paired (282 ± 58 g; n = 4) and wid- each species) in order for observed estimates owed females (273 ± 12 g; n = 8). Mass of nest success to be statistically different (P change averaged 276 ± 19 g (Range = 192- < 0.05). Other male removal studies in pre- — 448 g; n 12) for all females combined. Rate cocial species also detected little or no differ- of mass change was not related to recapture ences in nest success between widowed and distance (F = 0.03; df = 1, 8; P > 0.05) or paired females (Table 4). However, among the PCI (F = 1.18; df = 1, 8; P > 0.05), and did 6 statistical tests on geese, nest success varied not differ between paired (7.5 ±1.2 g/day; n in the predicted direction in 4 tests and was n==4)8;anFd=wi0.d7o5w;eddf f=em1a,l8e;sP(8>.80.±050)..7Ovge/rdaaly;l osifgnciofmicbainntleyddPif-fvearleunetsinfr1omtestt.heMe6tat-eastnsaliynsdiis- rate of mass gain averaged 8.4 ± 0.5 g/day cated a significant overall difference between (Range = 5.1-10.7 g/day; n = 12). Distance treatments (x2 = 23. 1 1, df = 12, P = 0.0268); recaptured from the colony was not related to however, when the P-value from the single PCI (F = 1.42; df = 1, 10; P > 0.05), but significant test was excluded, the combined p=1,ai51r)0e;dwePfreem<arl0ee.c0aR2po1ts)usr’ferdGoefmeursttehhee(rc5o0(.lF9on=±y7t6..h5a41n;kdwmfe;r=en a1n0a,lyPsis>w0a.s05)n.otMasirgtniinficeatntal.(x(21=9859).30a,lsodfre=- widowed females (27.3 ± 6.6 km; n — 8). ported that rates of egg loss or intraspecific nest parasitism did not differ between wid- DISCUSSION owed and paired Snow Geese breeding at a — Female reproductive success. We predict- smaller sub-arctic colony. ed that removal of male Ross’ and Snow Several factors may explain why male re- Geese during early incubation would nega- movals during incubation have little or no ef- tively affect several components of female re- fect on nest success in colonial geese. By ear- productive success. However, we did not de- ly incubation, territories have been estab- tect statistically significant differences in nest- lished, laying is completed, and females can ing or hatching success between treatments, no longer be fertilized via extra-pair copula- although nest success varied in the predicted tions (EPCs). Accordingly, forced EPCs rarely directions for both species (Table 1). Simula- occurred during incubation at Karrak Lake tion modeling indicated that sample size (LeSchack, unpubl. data; Afton, unpubl. data; — 62 THE WILSON BULLETIN • Vol. 110 No. 1. March 1998 , but see Mineau and Cooke 1979). Paine paired females (Martin et al. 1985, Schneider (1992) rejected the hypothesis that non-colo- and Lamprecht 1990). We also found that nest nial Canada Goose females benefit from male attendance was similar between widowed and protection against predators; however, he paired females, although the trend was in the questioned his results because nest success predicted direction. Similar trends also were was high (>82%) and many males frequently reported for Snow Geese (Martin et al. 1985) were away from their nests. and Canada Geese (Paine 1992). We found that early snow melt at Karrak We found that presence of their mates re- Lake did not result in increased overall nest sulted in paired females spending more time success. Although egg-laying began 8-16 with their bills tucked and less time alert days earlier than in the previous three years, while on nests than did widowed females. nest success of Ross’ (75%) and Snow Geese Widowed female Snow Geese spent twice as (80%) were similar to rates recorded at Karrak much time in the extreme head-up posture Lake in 1993 (82% Ross’, 76% Snow; Slat- while incubating compared to paired individ- A tery and Alisauskas 1993). Adverse weather uals (Martin et al. 1985). similar observa- conditions upon arrival at breeding grounds tion was reported for widowed Canada Geese delay reproduction and reduces clutch size. (Paine 1992). Martin et al. (1985) suggested Atlantic Brant Branta bernicla hrotd) ex- that an increase in alert behavior might result ( pended more energy searching for nest sites in an energy cost (i.e., decrease in body mass) and food resources, and had lower productiv- to widowed birds. We did not detect a signif- ity when nesting was delayed by late snow icant difference in body mass at hatch be- melt (Barry 1962). We suggest, that during a tween paired and widowed female Ross’ delayed breeding season, widowed females Geese, although the trend was in the predicted might have significantly lower nest success direction; however, our sample size was small. than do paired females as a result of declines Body mass of widowed ptarmigan also did not in physiological condition caused by an in- differ from that of paired females at the end crease in energy expenditure. Widows also of incubation (Martin 1984). might take more or longer incubation recesses Our study is the first to examine effects of than would paired females during late breed- male removal on post-hatch dispersal dis- ing seasons to search for food, thereby in- tance. Paired Ross’ females were recaptured creasing the risk of both nest predation and a nearly twice as far from the breeding colony lengthened incubation period (Aldrich and as were widowed females. We suggest that Raveling 1983, Madsen et al. 1989). We ob- paired females are able to travel further and served that time spent off nests by widowed to better brood-rearing areas than are widowed and paired females of both species, although females because of presence of their mates; not statistically different, varied in the pre- however, this hypothesis needs to be tested. dicted direction even though it was an early Slattery (1994) found that body size of Ross’ breeding season. We suspect that our sampling Geese varied positively with recapture dis- effort was not adequate to detect small differ- tance from the breeding colony at Karrak ences in reproductive success that could im- Lake; however, we found no relationship be- pinge on fitness differences between treatment tween body size and recapture distance. Thus, groups. pair status appears more important than body Incubation behavior, body mass and post- size in determining post-hatch dispersal from — hatch dispersal distance. In our study, nest the colony. attendance of Ross’ and Snow Geese during Role of male geese in biparental care. early incubation was similar to that recorded Male removal experiments in monogamous, in June 1993 at Karrak Lake (Afton, unpubl. precocial species generally have produced few data). Others who did male removal experi- effects on nest success or female incubation ments in Snow Geese (Martin et al. 1985) and behavior. In our study, widowed female geese Canada Geese (Paine 1992) reported no dif- spent increased time alert while on nests, but ferences in incubation constancy between this apparently did not entail major physiolog- widowed and paired geese. However, widows ical costs because body mass at hatch was were harassed and displaced more often than similar for paired and widowed females. LeSchack et al. • EFFECTS OF MALE REMOVAL IN GEESE 63 We suggest that male parental care in Arc- #WSNWT-04/93. We thank S. L. Koch and R. Bon for tic-nesting geese is more critical during laying assistance in data collection, and C. F. Bryan, P. O. than during incubation. Removal of males Dunn, P. A. Gowaty, G. R. Hepp, K. G. McCracken, and an anonymous reviewer for providing helpful during laying may cause many females to comments on the manuscript. abandon nests because of increased sexual and physical harassment (e.g., forced EPCs) from LITERATURE CITED conspecihcs. Nest success differed between Afton, A. D. and S. L. Paulus. 1992. Incubation and widowed and paired Snow Geese in a sub- brood care. Pp. 62-108 in Ecology and manage- arctic colony when males were removed dur- ment of breeding waterfowl (B. D. J. Batt, A. D. ing early laying (eggs 1-2; Martin et al. Afton, M. G. Anderson, C. D. Ankney, D. H. 1985). Lone Snow Goose females were sub- Johnson, J. A. Kadlec, and G. L. Krapu, Eds.). Univ. of Minnesota Press, Minneapolis. jected to frequent harassment from neighbor- Aldrich, T. W. and D. G. Raveling. 1983. Effects of ing males (Martin et al. 1985); however, these experience and body weight on incubation behav- authors did not determine whether this direct- ior of Canada Geese. Auk 100:670-679. ly caused nest failure. Alliston, W. G. 1975. Web-tagging ducklings in We also suggest that male parental care is pipped eggs. J. Wildl. Manage. 39:625-628. more critical after hatch than during incuba- Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour 49:227—267. tion because males are primary providers of Ankney, C. D. 1977. The use of nutrient reserves by brood defense while females replenish nutri- breeding male Lesser Snow Geese Chen caerules- ent reserves utilized during laying and incu- cens caerulescens. Can. J. Zool. 55:1984-1987. bation (Ankney 1977, 1979; Lazarus and In- Ankney, C. D. 1979. Does the wing molt cause nu- glis 1978; Sedinger and Raveling 1990; Afton tritional stress in Lesser Snow Geese? Auk 96:68— and Paulus 1992). Female geese lose 11-32% 72. of their body mass during incubation (Ankney Ankney, C. D. and C. D. MacInnes. 1978. Nutrient reserves and reproductive performance of female and Maclnnes 1978, Aldrich and Raveling Lesser Snow Geese. Auk 95:459-471. 1983, Thompson and Raveling 1987, Afton Barbault, R. 1986. Body size, ecological constraints, and Paulus 1992). Brood survival generally is and the evolution of life-history strategies. Evol. reduced by male removal (Table 4). For ex- Biol. 22:261-286. ample, Schneider and Lamprecht (1990) re- Barry, T. W. 1962. Effects of late seasons on Atlantic ported higher gosling feeding and survival Brant reproduction. J. Wildl. Manage. 26:19—26. Bromley, R. G. and R. L. Jarvis. 1993. The ener- rates and fewer interruptions of brooding for getics of migration and reproduction of Dusky paired females than for widows in a semicap- Canada Geese. Condor 95:193-210. tive flock of Bar-headed Geese (Anser indi- Dwyer, T. J. 1972. An adjustable radio-package for cus). We conclude that, under normal breeding ducks. Bird-banding 43:282-284. conditions, male removal during incubation is Dzubin, A. X. and E. G. Cooch. 1992. Measurements not greatly detrimental to nest success or fe- of geese: general field methods. California Water- fowl Association, Sacramento. male body condition; however, effects of male Erckmann, W. J. 1983. The evolution of polyandry removal during incubation in a late nesting in shorebirds: an evaluation of hypotheses. Pp. season should be studied. Finally, we believe 113-168 in Social behavior of female vertebrates that future research should focus on the im- (S. K. Wasser, Ed.). Academic Press. NY. portance of biparental care during the laying Forslund, P. and K. Larsson. 1990. The effect of and post-hatch periods. mate change and new partner's age on reproduc- tive success in the Barnacle Goose, Branta leu- ACKNOWLEDGMENTS copsis. Behav. Ecol. 2:1 16-122. Gauthier, G. and J. Tardif. 1991. Female feeding Our project received financial and logistical support and male vigilance during nesting in Greater from the Canadian Wildlife Service (CWS), Arctic Snow Geese. Condor 93:701-71 1. Goose Joint Venture, California Dept, of Fish and Hannon, S. J. 1984. Factors limiting polygyny in the Game, Polar Continental Shelf Project, Louisiana Co- Willow Ptarmigan. Anim. Behav. 32:153-161. operative Fish and Wildlife Research Unit, Louisiana Hedges, L. V. and I. Olkin. 1985. Statistical methods Dept, of Wildlife and Fisheries, School of Forestry, for meta-analysis. Academic Press, NY. Wildlife and Fisheries at Louisiana State Univ. (LSU), Hipes, D. L. and G. R. Hepp. 1993. Effect of male and NSF/LaSER Grant 1993)-HRD-01 Geese were removal on nest success of female Wood Ducks. ( . captured and collected under LSU Institutional Animal Condor 95:220-222. Care and Use Permit #A94-16 and CWS Permit Lazarus, J. and I. R. Inglis. 1978. The breeding be- 64 THE WILSON BULLETIN • Vol. 110 No. 1, March 1998 , haviour of the Pink-footed Goose: parental care portance of biparental care in Canada Geese. and vigilant behaviour during the fledging period. Ph D. diss., Southern Illinois Univ., Carbondale. Behaviour 65:62-88. Ryder, J. P. 1970a. A possible factor in the evolution MacInnes, C. D., R. K. Misra, and J. P. Prevett. of clutch size in Ross’ Goose. Wilson Bull. 82:5— 1989. Differences in growth parameters of Ross’ 13. Geese and Snow Geese: evidence from hybrids. Ryder, J. P. 1970b. Timing and spacing of nests and Can. J. Zool. 67:286-290. breeding biology of Ross’ Goose. Ph.D. diss., Madsen, J., T. Bregnballe, and L Mehlum. 1989. Univ. Saskatchewan, Saskatoon. Study of the breeding ecology and behaviour of Ryder, J. P. 1972. Biology of nesting Ross’ Geese. the Svalbard population of Light-bellied Brent Ardea 60:185-215. Goose Branta hemicla hrota. Polar Res. 7:1—21. Ryder, J. P. 1975. The signi—ficance of territory size Martin, K. 1984. Reproductive defence priorities of in colonial nesting geese an hypothesis. Wild- male Willow Ptarmigan (Lagopus lagopus): en- fowl 26:1 14-116. hancing mate survival or extending paternity op- SAS Institute. 1990. SAS/STAT user’s guide, 6.04 tions? Behav. Ecol. Sociobiol. 16:57-63. ed. SAS Institute, Cary, North Carolina. Martin, K., F. G. Cooch, R. F. Rockwell, and F. Schneider, J. and J. Lamprecht. 1990. The impor- Cooke. 1985. Reproductive performance in Less- tance of biparental care in a precocial, monoga- er Snow Geese: are two parents essential? Behav. mous bird, the Bar-headed Goose (Anser indicus). Ecol. Sociobiol. 17:257—263. Behav. Ecol. Sociobiol. 27:415-419. Martin, K. and F. Cooke. 1987. Bi-parental care in Sedinger, J. S. and D. G. Raveling. 1990. Parental Willow Ptarmigan: a luxury? Anim. Behav. 35: behavior of Cackling Geese during brood rearing: 369-379. division of labor within pairs. Condor 92:174- McKinney, F. 1965. The comfort movements ofAnat- 181. idae. Behaviour 25:120-220. Slattery, S. M. 1994. Neonate reserves, growth and McLandress, M. R. 1983. Temporal changes in hab- survival of Ross’ and Lesser Snow Geese gos- itat selection and nest spacing in a colony ofRoss’ lings. M.S. thesis, Univ. Saskatchewan, Saska- and Lesser Snow Geese. Auk 100:335—343. toon. Mineau, P. and F. Cooke. 1979. Rape in the Lesser Slattery, S. M. and R. T. Alisauskas. 1993. Studies Snow Goose. Behaviour 70:280-291. on the nutritional ecology and population biology Oring, L. W. and R. D. Sayler. 1992. The mating of Ross’ Geese nesting at Karrak Lake: progress systems of waterfowl. Pp. 190—213 in Ecology report. California Dept. Fish and Game, and Ca- and management of breeding waterfowl (B. D. J. nadian Wildlife Service, Saskatoon, Saskatche- Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, wan. D. H. Johnson, J. A. Kadlec, and G. L. Krapu, Sokal, R. R. AND F. J. Rohlf. 1981. Biometry, 2nd Eds.). Univ. of Minnesota Press, Minneapolis. ed. Freeman, New York. Owen, M., J. M. Black, and H. Liber. 1988. Pair Thompson, S. C. and D. G. Raveling. 1987. Incu- bond formation and timing of its formation in Bar- bation behavior of Emperor Geese compared with nacle Geese Branta leucopsis). Pp. 23-38 in Wa- other geese: interactions of predation, body size, ( terfowl in winter (M. W. Weller, Ed.). Univ. of and energetics. Auk 104:707-716. Minnesota Press, Minneapolis. Weller, M. W. 1956. A simple field candler for wa- Paine, C. R. 1992. Costs of parental care and the im- terfowl eggs. J. Wildl. Manage. 20: 1 1 1—1 13.