The Condor98:61-74 0 The CooperOrnithological Society 1996

AGE-SPECIFIC SURVIVAL AND PHILOPATRY IN THREE SPECIES OF EUROPEAN DUCKS: A LONG-TERM STUDY ’

PETERBLIJMS,~ AIVARS MEDNIS,~LMARS BAUGA Instituteof Biology,Latvian Academy of Sciences,Micra 3, LV-2169, Salaspils,Latvia

JAMES D. NICHOLS AND JAMES E. HINES PatuxentEnvironmental Science Center, National Biological Service, Laurel, MD 20708

Abstract. Capture-recaptureand band recovery modelswere usedto estimateage-specific survival probabilities for female Northern Shovelers (Anasclypeata), Common Pochards (Aythyaferina), and Tufted Ducks (.fuZigula)at Engure Marsh, Latvia, in 1964-1993. We banded more than 65,100 day-old ducklings of both sexesand captured 10,211 incu- bating females (3,713 new bandings and 6,498 recaptures).We developed a set of 3-age capture-recapturemodels to estimate annual survival rates for female ducklings,yearlings (SY), and adults (ASY) using programs SURGE and SURVIV and selectedparsimonious models usina a method develoned bv Akaike (1973). Survival rates of SY and ASY females were highest-forTufted Ducks intermediate for Common Pochards,and lowest for Northern Shovelers.Survival rates of SY and ASY females varied in parallel for shovelers and po- chards. We believe that much of the difference in survival estimates between SY and ASY was caused by mortality rather than permanent emigration. Estimates of day-old duckling survival, reflecting both mortality and permanent emigration, were 0.12 for shov- eler, 0.06 for pochard, and 0.03 for . For all species,duckling survival varied over years, but the pattern of variation was not similar to that of the other age classes. Estimatesof survival usingband recovery data for SY + ASY female pochardsand Tufted Ducks were similar to the capture-recaptureestimates, suggestingthat surviving females returned to the breeding marsh with probabilities approaching 1. Key words: age-speciJicfemale survival;; Anas clypeata;Aythya ferina; Aythya fuligula; breedingphilopatry;permanent emigration; Latvia.

INTRODUCTION , Bucephala clangula (Dow Capture-recaptureand band recovery models are and Fredga 1984) or island nesting Common Ei- becoming increasingly important for estimation der, Somateria mollissima (Coulson 1984), all of of survival rates of birds, including waterfowl. which are relatively easy to capture. Traditionally, band recovery models have been Most evidence on breeding philopatry in fe- used extensively to estimate survival rates of male waterfowl comes from estimates of return ducksand geese(reviewed by Johnsonet al. 1992), rate, however, this statistic incorporates three but capture-recapturemodels have received little probabilities and can be only used to draw ten- attention, primarily becauseof the lack of long- tative inferences about homing (Anderson et al. term studieswith marked birds. There have been 1992, Johnsonet al. 1992). If both band recovery few long-term capture-recapture studies of and capture-recapture survival estimates can be breeding ducks that estimated survival proba- obtained for breeding females banded in a par- bilities basedon modem statisticalmethods. Most ticular location then it is possible to directly es- ofthese have focusedon box-nesting speciessuch timate unconfounded homing probability (see as Wood Duck, Aix sponsa (Hepp et al. 1987, Methods). Becauseof the lack of adequate data Dugger 1991, Hepp and Kennamer 1993) and no suchestimates are currently available (but see Hepp et al. 1987). In this paper we use 27 years of band recovery and 18 years of capture-recapture data to esti- mate survival and breeding probabilities of fe- LReceived 31 March 1995. Accepted 9 November male ground-nesting and over-water nesting 1995. * Presentaddress: Gaylord Memorial Laboratory,The ducks on a single study area in Latvia. Aging of School of Natural Resources,University of Missouri- breeding females and banding of day-old duck- Columbia, Puxico, MO 63960, USA. lings with special oval bands (Blums et al. 1994),

1611 62 PETER BLUMS ET AL.

permitted inferencesthat have not been possible tions, many artificial elevated islands were con- using previous methods (i.e., we estimated sur- structed on the flooded sections of two natural vival probabilities for females of three age class- islands during 198 1-1983 (Blums and Mednis es). To determine whether females are philopa- 199 1). Beginning in 1984,82 islands totaling 14.3 tric, we compared survival estimates from band ha in area were available for nesting within the recovery and capture-recapturemodels, and also previous island territory. We believe that frag- evaluated band recovery locations of females mentation of large islands increasedthe carrying captured on the nests. capacity of island breeding habitats despite the substantial decreasein total surface area. This METHODS was confirmed by the highest ever number of STUDY AREA AND BREEDING duck nests recorded on all islands in the early POPULATIONS OF DUCKS 1990s (Blums et al. 1993). A long-term capture-recapturestudy of Northern In addition to the natural and artificial islands, Shoveler (Anus clypeuta, hereafter shoveler), nest searcheswere expanded to three isolated Common Pochard (Aythyu ferina, hereafter po- areas of persistent emergent marsh in 1972, to- chard), and Tufted Duck (Aythyu fuligulu) was taling approximately 111 ha of reed-beds and conducted from 1958 to 1993 on Engure Marsh, cattail stands, excluding open water. Thus, per- Latvia, Eastern . The 35-km2 Engure manent sampling areas included natural and ar- Marsh is a shallow, permanently-flooded pal- tificial islands, 1958-1993, and emergent marsh- ustrine marsh (Cowardin et al. 1979) on the east es, 1972-1992. coast of the Baltic Sea (57”15’N, 23”07’E). The During the last 20 years, the marsh supported marsh has gradually changed from an open to a about 2,000 breeding pairs of ducks, with about hemi-marsh (Weller and Spatcher 1965) domi- 60% consisting of pochards, Tufted Ducks, and nated by tall, robust hydrophytes such as Com- shovelers.The averagenumbers of breeding pairs mon Reed (Phrugmitesuustralis) and cattail (Ty- on the entire marsh during 1977-1993 were as phu spp.). Human activities are prohibited on follows (Blums et al. 1993): Common Pochard islands and areas of emergent vegetation during 900 (range 560-1640), Tufted Duck 280 (160- the breeding season but most of the marsh is 360), and Northern Shoveler 33 (range 19-59). open to during early August Of thesenumbers 99% of shovelers,42% of Tuft- through early November. Different management ed Ducks, and 23% of pochards nested within activities (such as construction of artificial is- permanent sampling areas. Shovelers nested al- lands, vegetation and predator control, and at- most exclusively on the islands within perma- traction of gulls and terns) were conducted on all nent sampling areas, thus the entire breeding sampling areas throughout the study that pre- population was monitored each year during rou- vented the decline of carrying capacity of breed- tine nest searches.The breeding populations of ing habitats (Mihelsons et al. 1976, Blums and Tufted Ducks and shovelers were fairly stable Viksne 1990, Blums and Mednis 1991). Preda- throughout the study period however, the num- tors were systematically controlled for two to ber of shovelers increased substantially during three months beginning with the break-up of ice, the last five years (Blums et al. 1993). Pochards and an average of 11 (range 1-17) American increased during the last 16 years. Minks (Musteluvison), 83 (44-l 44) Marsh Har- riers (Circus aeruginosus),21 (O-63) Ravens FIELD METHODS (Corvuscorux), 17 (6-34) Hooded Crows (Corvus We conducted two to three complete searchesfor corone),14 (2-25) Herring Gulls (Larusurgen- duck nestson the permanent sampling areasfrom tutus),etc., were removed or relocated out of the mid-May to mid-June. All breeding habitats marsh during eachbreeding season(records from within permanent sampling areas were system- 1978 through 1993). atically searchedto locate nests by walking par- Permanent sampling areas included five nat- allel transects.We adjusted the distance between ural islands with a total surfacearea at low water transectsfrom 1.5 to 3.0 m in relation to vege- of approximately 20 ha, from 1958 to 198 1. Large tation density and height. Additional intensive portions of theseislands were flooded during high effortswere made each breeding seasonto locate water conditions and were not always suitable new nests by flushing females and watching lone for nesting. To maintain stable breeding condi- individuals of both sexes.We believe the effec- SURVIVAL AND PHILOPATRY OF FEMALE DUCKS 63 tivenessof nest searcheswas high becausecertain ings or a narrow brownish ring at the inner part sampling areas were assigned to the biologists of iris. who worked on the same areas for lo-30 con- Newly hatched ducklings were captured by secutive field seasonsand knew potential nest hand at nests and banded with plasticine-filled sites. That knowledge was especially important oval aluminum bands (Blums et al. 1994) rather for finding nestsin reed-beds.Experimental burns than with the web tagscommonly usedby Amer- on islands, after the last nests hatched, verified ican biologists(Grice and Rogers 1965, Haramis that more than 95% of all nests were found each and Nice 1980). Band loss for ducklings was es- year. We captured incubating females on nests timated to be extremely low, ~0.5% (Blums et during the last week of incubation using drop- al. 1994). Oval bands were replaced with con- door traps (Blums et al. 1983) or dip nets; un- ventional round bands at the first recapturewhen marked females were banded with conventional birds returned to breed. legbands. Beginning in 1976, we obtained a sample of STATISTICAL METHODS known-age females using two different methods. Birds marked in this study were reobserved in First, more than 65,000 day-old ducklings were two different ways. First, some birds were re- individually marked using plasticine-filled leg- covered after having been shot or found dead bands (Blums et al. 1994). Subsequentrecaptures and their band numbers reported by investiga- of these birds as breeding females allowed us to tors or the general public. Resulting banding and assignthem an exact age. Second, unmarked in- recovery data were analyzed using band recovery cubating femaleswere agedas either yearlings (l- models of Brownie et al. (1985). These models year-old, SY) or adults (?2-years-old, ASY) us- permit estimation of annual survival rates and ing wing feather characteristics.There were a rel- band recovery rates. Annual survival rates reflect atively small fraction of unmarked birds (no the proportion of birds alive at the time of band- plasticine-filled or conventional bands) in the ing in one year that are still alive at the time of sample of females that were used for capture- banding in the subsequentyear. The complement recapture analyses (24, 25, and 32% of Tufted of survival rate estimates obtained from band Ducks, shovelers, and pochards, respectively). recovery data and models (1 - L?,where 3 is a The shapeand coloration of greater secondary survival estimate) typically reflects only mortal- coverts (GSC) were key identification characters ity. usedsingly or in combination for all three species Second,many femaleswere recapturedat nests (see Boyd et al. 1975, Palmer 1976). The 11th during the breeding seasoneach year by inves- or 12th, rather than 13th or 14th, GSC were tigators. Resulting data were analyzed using cap- better indicators for both diving ducks because ture-recapture models for open populations, de- thesecoverts usually had not been molted during fined as populations in which birds can enter or the breeding season.The age-specificpattern in leave the population between sampling period shape (Tufted Duck, pochard) of GSC was sim- (Seber 1982, Pollock et al. 1990, Lebreton et al. ilar to that describedby Dane and Johnson(1975) 1992). Unlike band recovery data in which a for Redheads (Aythyu americana), and vermic- banded can be recovered as dead only once, ulation of upper wing feathers (pochard), to that a singlebird may be recapturedin multiple years. describedby Serie et al. (1982) for Band recoveries can occur over a broad geo- (Aythya valisineria), respectively, in North graphic area, but recapture data typically come America. The coloration of first GSC and distal from restricted study areas (in our case the En- parts of middle and lessersecondary coverts were gure Marsh). As a result, the complement of “sur- the best indicators for shovelers. Eye color also vival” estimates from capture-recapture models was used to age female Tufted Ducks (seeTraug- includes both death and permanent emigration, er 1974), although there was some overlap in since marked birds may depart the study area color between SY and ASY females. The iris and go to different breeding areas in subsequent color of SY femaleswas brownish-yellow or dull- years. This difference in the survival estimates yellow, or rarely yellow with a blurred brownish computed from band recovery and capture-re- inner zone adjacent to the pupil. The iris color capture data and models permits inference about of ASY females was typically orange-yellow or homing and permanent emigration (Hepp et al. yellow, or rarely with scatteredbrownish mark- 1987). If all birds home back to the breeding 64 PETER BLUMS ET AL. marsh each year, then survival estimates hased data, whereas a nonsignificant test statistic in- on band recovery and capture-recapture data dicatesthat the more general model is not needed should be similar, as they estimate the same to explain the data. Goodness-of-fit testsprovide quantity. Conversely, if many birds do not return evidence about the adequacyof a particular mod- to the breeding marsh but shift breeding loca- el to describe the variation in the data set. A tions, then the survival estimate from band re- significant x2 goodness-of-fit test statistic indi- covery data will be larger than the estimate from cateslack of fit of the model to the data, whereas capture-recapture data, with the difference re- a nonsignificant test statistic provides no evi- flecting the amount of permanent emigration dence of lack of fit. After obtaining these esti- (Hepp et al. 1987). mates, we compared the two speciesusing pro- Band recoverymodels. Survival rates were es- gram CONTRAST (Sauer and Williams 1989, timated from band recovery data using the mod- Hines and Sauer 1989). els described by Brownie et al. (1985). Band re- Capture-recapturemodels. Capture-recapture covery data were limited, and numbers of band- data from the permanent sampling areas at En- ings and recoveries of female ducks were fewer gure Marsh were available for all three species. than thosetypically usedin banding studiesfrom Because ducklings were banded with legbands . Northern Shoveler data were in- (Blums et al. 1994) and becauseolder birds could sufficient for reasonable use with these models be identified as SY (l-year-old) or ASY (r2- but numbers of bandings and recoveries were year-old), we were able to consider models in adequate for pochardsand Tufted Ducks, so we which survival and capture probabilities differed estimated survival rates because of the virtual among three age classesof ducks (Pollock 198 1). absenceof other survival estimatesfor thesediv- The only complication in this modeling was that ing ducks. For the Common Pochard, there were not all ducklings were sexed. We dealt with this enough data for SY and ASY females to compare by assuming 50% females among banded duck- model HO (year-specific survival and recovery lings. Preliminary analysis revealed no signifi- rates; no age-specificity) against Hl (year- and cant deviation from a 50:50 sex ratio (6,495 day- age-specific survival and recovery rates) using old ducklings sexed)at hatching for all three spe- the contingency table test of Brownie et al. (1985) cies (Blums and Mednis 1996). So if n, is the for age-specificityof survival and recovery rates. total number of ducklings banded and released We conducted this test for two separate time in year t and r, is the number of these that are periods, 1974-1982 and 1984-1991 (data were recaptured as SY or ASY females in subsequent scarce for 1983), and summed the resulting x2 years, then we computed the number of female values (and their associateddegrees of freedom) ducklings released but never recaptured as: (n,l to obtain an overall test of the null hypothesis 2) - r,. of no difference between the two age classesof We developed three-age capture-recapture pochards.Tests for age-specificityof survival and models for each speciesusing programs SURGE recovery rates of pochards did not indicate that (Pradel et al. 1990) and SURVIV (White 1983). the null hypothesis should be rejected (x*~~ = Our approach was to use SURGE in some of the 35.9, P = 0.38). initial screeningof models and then to use SUR- Band recovery analyses were therefore con- VIV in the final analyses. We developed the ducted on combined SY and ASY females for models to investigate two possible sources of pochards(1974-1992) and Tufted Ducks (1964- variation in survival and capture probabilities, 199 1). We used single-age models of program age and time (calendar year). Model parameters ESTIMATE, selectinga parsimonious model us- included survival probability, d,, the probability ing likelihood ratio and goodness-of-fit tests. that an of age a (day-old duckling, SY, Likelihood ratio testswere conducted between a ASY) in the breeding season of year t survived more general model (representingthe alternative until the breeding seasonof year t + 1 and did hypothesis)with more parameters and a simpler not permanently emigrate from Engure Marsh, model (representing the null hypothesis) with and capture probability, pat, the probability that fewer parameters. A significant (e.g., P < 0.10) an animal of age a in the breeding seasonof year likelihood ratio x2 test statisticprovides evidence t was captured in that season. that the additional parameters of the more gen- Different models are defined by the subscripts eral model are required to adequately model the on survival and capture probability parameters. SURVIVAL AND PHILOPATRY OF FEMALE DUCKS 65

The presenceor absenceof a subscript indicates enough parameters to adequately describe the that the sourceof variation (age or time) does or variation in the data, then estimates are likely to does not apply to the parameter, respectively. be biased. AIC can be computed for all models With respectto age, we considered models with under consideration for a given data set. The all three age classes(denoted by subscript a) and model with the lowest AIC is optimal with re- models with only two age classes,ducklings and spect to this tradeoff between adequate fit and SY + ASY (these models are denoted by sub- precision. script a’). Thus model (&., p,.,) indicates the We used the Pearson goodness-of-fit test with general 3-age model with temporal and age-spe- the cell pooling algorithm of SURVIV to assess cific variation in both survival and capture prob- the fit of each model to the data. Simulations, abilities. This model contains separate parame- however, indicated that the pooled Pearson fit ters for each age-time combination for both sur- statistic rejected the null hypothesis of reason- vival and capture probability. Capture proba- able model fit more frequently than the nominal bility cannot be estimated for the initial ageclass, level when the null hypothesis was true. Thus, so model (+,., p,.,) has the following age-specific in caseswhere the goodness-of-fitstatistic for the parameters: 9,,, &, &, pzrand& Model (&, P) low-AIC model indicated lack of fit (P < 0.05), indicates a model in which survival differs for we investigated model fit in more detail via Mon- the two age classes(ducklings, SY + ASY), cap- te Carlo simulation. We generated 500 data sets ture probability applies to the combined SY + using the actual sample sizes and model param- ASY class, and neither capture nor survival eter values set equal to the actual estimates. For probabilities vary over time (they are constants the simulated data sets,we computed the distri- for all years of the study). bution of the goodness-of-fitstatistics divided by In instances where both age and time were their degreesof freedom, $/df, and then noted important sourcesof variation in model param- the location of the actual x*/df relative to this eters, we tested models in which age-specificpa- distribution. rameters exhibited “parallelism” (sensu Lebre- Although our capture-recapturemodeling was ton et al. 1992) over time. For example, in model directed at making inferences about survival (I#J), (&.,, p,,,), where the subscript a+t denotes par- we also used an approachbased on the reasoning allelism, capture probability for SY females at of Clobert et al. (1990, 1994) to investigate age- time t is modeled as pZ1= y”p,,. So capture prob- specificbreeding proportions. Clobert et al. (1990, ability varies by age but the pattern of temporal 1994) have noted that in speciesfor which in- variation is similar for the two age classes.For dividuals are only available for capture in years survival, subscript a + t denotes full parallelism when they attempt breeding, age-specificdiffer- (&, = Y& & = r’d,, &), whereas subscript ences in capture probability can often be attrib- (a + t) ’ denotes parallelism between SY and ASY uted to differencesin age-specificbreeding pro- survival probabilities but not with duckling sur- portions. Specifically, if SY and ASY females vival ($,,, & = y’&, &). We used likelihood that nest in a given year are equally likely to be ratio teststo test for specificsources of variation captured (a reasonableassumption in this study), using nested models. Two models are nested if then the ratio of SY to ASY capture probabilities, one model can be obtained by constraining the pJp,,, should estimate the ratio of SY to ASY parameters of the other. breeding probabilities. When the breeding prob- We usedAkaike ’s Information Criterion (AIC, ability for ASY approaches1 (i.e., when virtually Akaike 1973) as a means of selecting the most all ASY females breed), then the ratio of capture parsimonious model for each data set. The Prin- probabilities estimates the proportion of SY fe- ciple of Parsimony (see Bumham and Anderson males that breed (Clobert et al. 1990, 1994). 1992) is based on the idea that our model should We choseto estimate the ratio of capture prob- reflect a compromise between “overfitting” and abilities directly using models with age-specific “underfitting” our data. Variances of estimates parallelism in capture probabilities, i.e., the P,+~ increase as the number of model parameters in- models. The estimates of 4” from such models creases,so if our model contains more param- directly estimate the ratio of SY to ASY breeding eters than are needed, we will sacrificeprecision proportions and, if all ASY females breed, the needlessly,by having larger than necessaryvari- proportion of SY females that breed. We com- ance estimates. Conversely, if we do not include pared the age-specificratios of breeding propor- 66 PETER BLUMS ET AL.

TABLE 1. Band recovery model goodness-of-fitand likelihood ratio tests for SY and ASY female Common Pochardsand Tufted Ducks banded at Engure Marsh, Latvia.

Common Pochard, 1974-1992 Tufted Duck, 1964-1991 Type of test Model(s) x1 df P Yl df P

Goodness-of-fit Model 1 30.9 24 0.16 19.5 8 0.01 Model 2 39.9 37 0.34 46.2 33 0.06 Model 3 96.8 63 co.01 85.6 61 0.02 Likelihood ratio Model 3 vs. 1 65.3 35 co.01 82.3 53 co.01 Model 2 vs. 1 18.0 17 0.39 36.0 26 0.09

tion, T”, as well as survival rates, 6, among spe- CAPTURE-RECAPTURE MODELS cies using program CONTRAST (Sauer and Wil- Northern Shoveler. The model with the lowest liams 1989, Hines and Sauer 1989). AIC for female shovelers was model ($(.+I)‘, p,) (Table 2). The pooled goodness-of-fit statistics RESULTS from program SURVIV indicated poor fit of this BANDING AND CAPTURE DATA model to the data (P = 0.03; Table 2); the sim- ulation-based estimate, however, indicated an A sample of 65,122 day-old ducklings of both acceptable fit (P = 0.09). Under model (&a+ry, sexeswas banded and 10,2 11 incubating females p,) capture probability varied by age (SY and were captured (3,713 new bandings and 6,498 ASY) but not by time. Estimated capture prob- recaptures) at the marsh during the study, but ability for ASY (0.80) was higher than that for data were broken down by periods and species SY females(0.5 1; Table 3). We usedmodel (+(n+ty, foranalyses.Samplesof 17,802(748), 1,351(668), p,+J to directly estimate the ratio of SY to ASY and 3,783 (2,229) releases(recaptures) of duck- breeding proportions as +” = 0.70 (S% = ling, SY, and ASY females, respectively, were 0.08 1). used for estimation of age-specificsurvival prob- Annual survival probability varied by age and abilities with programs SURGE and SURVIV time. Temporal variation in survival probabili- for the periods 1976-1993 (shoveler) and 1976- ties of SY femalesparalleled that of ASY females, 1992 (pochard, Tufted Duck). New bandings (n but variation in duckling survival probabilities = 4,064) of incubating females and associated did not parallel that of the other two age classes. band recoveries (n = 5 15) provided data for es- Thus the same factors probably are responsible timation of survival probabilities for diving ducks for year-to-year variation in survival probabili- with program ESTIMATE for the periods 1964- ties of the two older age classes,but other factors 1992 (Tufted Duck), and 1974-1993 (pochard). are responsiblefor temporal variation in survival during the first year of life. The mean annual BAND RECOVERY MODELS survival probability for ASY females was 0.58. Common Pochard and Tufted Duck. Results of The estimate off’ = 0.65 for SY females indi- goodness-of-fit and likelihood ratio tests (Table cates that average SY survival probabilities 6, 1) led us to select Model 2 (Brownie et al. 1985) were about 0.38 (Table 3). The average survival for both speciesof diving ducks. Model 2 con- estimate for ducklings was 0.12 and reflects the tains a single time-constant survival parameter probability that a newly-hatched duckling sur- and year-specific recovery rates. The estimated vived the subsequent year and returned to the sxrvi_val rate for female pochards ($_ = 0.59; study area on Engure Marsh the following breed- Sz[S] = 0.022) and Tufted Ducks (S = 0.71; ing season(i.e., did not permanently emigrate to SE[S] = 0.020) were different (x2, = 15.4, P < another breeding area). 0.01). Average estimated reco_veT rates for the Common Pochard. Numbers of pochard cap- two speciesweref = 0.056 (sE[fj = 0.003) for tures exceeded those of shovelers and Tufted pochardsandf= 0.056 (SE[fl = 0.005) for Tuft- Ducks, and required a general (i.e., incorporating ed Ducks. There was some evidence of different many sources of variation) model (+(a+tr,p,.,; recovery rates for the two species,but test results Table 4). Capture probabilities varied by both were equivocal (x21 = 3.3, P = 0.07). age and time, and the pattern of temporal vari- SURVIVAL AND PHILOPATRY OF FEMALE DUCKS 67

TABLE 2. AIC values,goodness-of-fit test statistics and likelihoodratio (LR) teststatistics for femaleNorthern Shovelerscaptured and recapturedat EngureMarsh, Latvia, 1976-1993.

Goodness-of-fit P.)* LR test with &+q> Source of variation Model NP’ AIC x2 df P x ’ df P tested by LR (&+r).,Pa)* 37 422.9 33.3 20 0.03 - - (&+r).,P,.J 67 434.3 - - - 48.6 ii 0.02 Full- temporalvariation in P. (&+t).,P*+J 52 428.8 28.5 7 co.01 24.0 15 0.06 Parallel temporal variation in p. c#&D)P ’ 36 440.0 42.3 21 10.01 19.2 3: <0:0110 01 Age-specificvariation in p 5 441.9 116.7 54 co.01 83.0 Temporalvariation in 4. (&>“p., 21 449.0 94.8 38 co.01 58.2 16 KO.01 Parallel temporal variation in &, ti2 and 4, (@J,.,,P.) 53 428.4 13.7 4 co.01 26.5 16 0.05 Full temporal variation in 9. (@J,,PJ 19 605.2 276.7 48 co.01 218.4 18 CO.01 Age-svecific variation in Q, * Model selectedfor usein estimation. ** Number of parametersestimated. ation was not parallel for SY and ASY females. birds, but variation in duckling survival was not Average capture probability was higher for ASY parallel to that of the other two ageclasses. Mean (j3 = 0.69) than for SY females (j2 = 0.54; Table annual probability of surviving and not perma- 5). We used model (rj~(~+~).,p,+J to directly esti- nently emigrating was estimated at 0.06 for duck- mate the ratio of SY to ASY breeding propor- lings and 0.65 for ASY females (Table 5). The tions as 4” = 0.70 (& = 0.045). estimated +’ of0.85 indicates an averagesurvival Survival probabilities also varied by age and rate for SY females of about 0.55. time (Table 4). Parallel temporal variation was Tufted Duck. Tufted Ducks were best mod- observedin survival probabilities of SY and ASY eled using model (&,.,, p,) in which capture probabilities varied by age but not time (Table 6). Estimated capture probability for ASY fe- TABLE 3. Parameter estimates under model males (0.80) was higher than that for SY (0.73; C#J@+*).>pa)*for female Northern Shovelerscaptured and Table 7). We used model ($I(~,.,,p,+J to directly recapturedat Engure Marsh, Latvia, 1976-1993. estimate the ratio of SY_to ASY breeding pro- portions as +” = 0.89 (SE = 0.047). ASY* -.survival The mean estimated probability that a duck- Year(t) 6vCWb.l) ling survived until the next breeding seasonand 1976 0.11 (0.039) 0.61 (0.128) did not permanently emigrate from EngureMarsh 1977 0.11 (0.043) 0.52 (0.124) was 0.032, the lowest value for the three species. 1978 0.21 (0.054) 0.38 (0.105) There was no evidence of a difference in annual 1979 0.04 (0.021) 0.48 (0.105) 1980 0.22 (0.059) 0.30 (0.107) survival rate between SY and ASY Tufted Ducks 1981 0.04 (0.022) 0.49 (0.126) (Table 6), and the average annual survival rate 1982 0.05 (0.028) 0.57 (0.124) estimate (0.7 1) for thesebirds was very high (Ta- 1983 0.08 (0.040) 0.66 (0.145) ble 7). 1984 0.07 (0.032) 0.78 (0.130) 1985 0.03 (0.024) 0.63 (0.132) INTERSPECIFIC COMPARISONS 1986 0.15 (0.051) 0.59 (0.131) 1987 0.23 (0.058) 0.47 (0.124) The ratios of SY to ASY breeding proportions 1988 0.18 (0.053) 0.79 (0.131) varied among the three species(x2* = 9.6, P -C 1989 0.07 (0.027) 0.62 (0.107) 0.0 l), being large for Tufted Ducks and smaller, 1990 0.23 (0.046) 0.54 (0.092) but nearly identical, for shovelersand pochards 1991 0.10 (0.026) 0.69 (0.094) 1992 0.09 (0.030) 0.72 (0.098) (Table 8). Survival rates of ducklings varied 0.12 (0.013) 0.58 (0.030) among the three species(x2* = 76.1, P -C O.Ol),

* Other parameterestimates un er this model are:ratio of SY to ASY with estimates for shovelers being largest and survival (7’ = &,&,), Ev’= 0.65, d (4’) = 0.076 SY captureproba&lity, estimates for Tufted Ducks being smallest. Sur- bO;O$:l, %kX,) = 0.054; ASY captureprobability, b,, = 0.80, (B,,) vival estimatesalso varied among speciesfor SY 68 PETER BLUMS ET AL.

TABLE 4. AIC values,goodness-of-fit test statistics and likelihoodratio (LR) teststatistics for femaleCommon Pochardscaptured aad recapturedat EngureMarsh, Latvia, 1976-1992.

LR test with model Goodness-of-fit (@,..+!I?p..J* Sourceof variation Model NP* ’ AIC x’ df P X2 df P tested by LR 63 816.8 46.2 41 0.27 - - - 35 828.9 114.3 72

(x22 = 71.0, P < 0.01) and ASY females (x22 = to reject the null hypothesis.The ratio of the two 23.1, P < O.Ol), but the ordering was opposite estimates was 1.O, indicating that the estimated that for duckling survival, with the largest esti- probability of permanent emigration was zero. mates for Tufted Ducks and the smallest for This means that the capture-recapture survival shovelers (Table 8). estimate reflects the true survival for female Tufted Ducks. INFERENCESABOUT PHILOPATRY The capture-recapture survival estimates for We were interested in the null hypothesis that pochards were clearly age-specific.To obtain a the conditional probability of a breeding female single survival estimate for pooled age groups duck returning to the same breeding area, given (SY + ASY) we weighted both values for their that the bird is alive, equals 1. The capture-re- relative importance in the population (SY/ASY capture survival estimate (0.7 1) for ASY female = 0.26/0.74) and calculated the weighted mean Tufted Ducks was identical to the band recovery survival rate estimate, 0.62. The single capture- survival estimate (0.71), providing no evidence recaptureestimate for pochards(0.62) was slight-

TABLE 5. Parameterestimates under model (@ (a+tr, pa.,)* for femaleCommon Pochards captured and recaptured at EngureMarsh, Latvia, 1976-1992.

SY captureprobability ASY captureprobability Year (t) Bu(salBd) ba(saA,l) 1976 0.10 (0.018) 0.54 (0.056) 1977 0.03 (0.011) 0.64 (0.060) 0.57 (0.096) 0.63 (0.068) 1978 0.04 (0.013) 0.55 (0.058) 0.43 (0.152) 0.58 (0.057) 1979 0.05 (0.014) 0.64 (0.055) 0.63 (0.156) 0.55 (0.058) 1980 0.06 (0.016) 0.72 (0.061) 0.71 (0.161) 0.8 1 (0.049) 1981 0.07 (0.016) 0.54 (0.052) 0.43 (0.118) 0.6 1 (0.058) 1982 0.05 (0.015) 0.70 (0.061) 0.68 (0.121) 0.77 (0.053) 1983 0.05 (0.015) 0.59 (0.059) 0.37 (0.127) 0.63 (0.058) 1984 0.06 (0.014) 0.65 (0.049) 0.40 (0.131) 0.67 (0.059) 1985 0.14 (0.020) 0.60 (0.049) 0.57 (0.117) 0.79 (0.048) 1986 0.04 (0.009) 0.62 (0.046) 0.46 (0.071) 0.73 (0.052) 1987 0.04 (0.009) 0.73 (0.048) 0.54 (0.115) 0.66 (0.045) 1988 0.06 (0.012) 0.78 (0.046) 0.72 (0.130) 0.63 (0.044) 1989 0.11 (0.017) 0.75 (0.041) 0.58 (0.103) 0.68 (0.043) 1990 0.03 (0.007) 0.72 (0.043) 0.25 (0.056) 0.81 (0.037) 1991 0.76 (0.138) 0.78 (0.041) Means 0.06 (0.004) 0.65 (0.014) 0.54 (0.033) 0.69 (0.014)

l Ratio of SY to ASY survival y’ = &,/@ll underthis model is 4’ = 0.85 (S&‘] = 0.036). SURVIVAL AND PHILOPATRY OF FEMALE DUCKS 69

TABLE 6. AIC values, goodness-of-fit test statistics and likelihood ratio (LR) test statistics for female Tufted Ducks captured and recaptured at Engure Marsh, Latvia, 1976-1992.

Goodness-of-fit LR test with (&.,, p.)* Sourceof variation Model NP* AIC x’ df P x2 df P testedbv LR

563.6 33.1 42 0.83 - - i4 600.1 27.5 0.07 19.6 ii 0.88 Full temporal variation in pa 48 579.8 32.1 :; 0.46 11.8 14 0.62 Parallel tepmoral variation in pa 33 566.0 35.1 42 0.76 4.3 1 0.04 Age-specific variation in p 4 617.7 146.4 79 co.01 114.1 30 co.01 Full temporal variation in @., 20 609.3 107.2 58 co.01 73.6 14 co.01 Parallel temporal variation in

18 613.8 114.1 58 10.01 82.2 16 co.01 Parallel temporal variation in & and (6, = 6,) 50 585.9 24.1 28 0.68 9.8 16 0.88 Full age-specific variation in 4, * Model selectedfor usein estimation. ** Number of parametersestimated.

ly higher than the band recovery estimate (0.59), were similar or even higher than those estimated providing no evidence to reject the null hypoth- from long-term studies of box-nesting Wood esis. Ducks in the United States (0.78, Hepp et al. 1987; 0.65, Dugger 1991). DISCUSSION Estimated average capture probability of ASY CAPTURE PROBABILITIES was higher than that of SY females for all three Average capture probabilities of ASY females species. Because almost all females were cap- were relatively high (0.69-0.80) for all three duck tured on nests, differencesin capture probabili- species, suggesting that effectiveness of nest ties most likely reflect lower probabilities of nest- searchesand capture rates of breeding females ing by SY females. Indeed, our estimates of SY have been generally high despite the large terri- to ASY breeding proportions were 0.70 for shov- tories and difficult terrain. Capture probabilities eler and pochard,and 0.89 for Tufted Duck. These estimates were based on the assumption that all ASY (z2-year-old) females breed. However, as TABLE 7. Parameter estimates under model (&.,),p3* indicated by our previous research(Mihelsons et for female Tufted Ducks captured and recaptured at al. 1986) and some North American studies (Af- Engur Marsh, Latvia, 1976-92. ton 1984), possibly not all 2-year-old females breed. Therefore, our estimates of breeding pro- Duc_yggyvival SY +. ASY^” sunivd Year (1) O&W,,I) &,(sWd portions of SY birds may have been biased slight- ly high. 1976 0.082 (0.0140) 0.69 (0.052) The pattern of temporal variation of capture 1977 0.046 (0.0092) 0.84 (0.038) 1978 0.016 (0.0055) 0.70 (0.040) probabilities was not parallel for SY and ASY 1979 0.056 (0.0108) 0.75 (0.042) pochards.This may suggestheterogeneity in cap- 1980 0.020 (0.0065) 0.73 (0.039) ture rates of females of different age classesin 1981 0.025 (0.0078) 0.76 (0.041) particular years. Possibly, SY and ASY females 1982 0.03 1 (0.0083) 0.72 (0.044) 1983 0.0 17 (0.0064) 0.70 (0.046) responded differentially to changes in breeding 1984 0.003 (0.0028) 0.59 (0.049) habitats, and different proportions from these 1985 0.039 (0.0103) 0.60 (0.052) two age classesmay have temporarily emigrated 1986 0.019 (0.0072) 0.72 (0.049) to suitable areasoutside permanent sampling ar- 1987 0.003 (0.0030) 0.79 (0.048) eas. 1988 0.032 (0.0100) 0.67 (0.054) 1989 0.021 (0.0079) 0.67 (0.053) Capture probabilities of pochards also varied 1990 0.051 (0.0133) 0.69 (0.058) over time. Potential explanations for this lack of 1991 0.049 (0.0 136) 0.71 (0.060) uniformity in capture rates were: (1) pochards Means 0.032 (0.0022) 0.71 (0.011) showed the most flexible nesting patterns and l Other pammete estimatesunder this model are: SY captureprob- bred at a wide variety of different habitats within ability, &, = 0.73, &@J = 0.037; ASY captureprobability, fi,, = 0.80, SF&) = 0.013. the marsh (e.g., in years with high water levels 70 PETER BLUMS ET AL.

TABLE 8. Interspecific comparisonsof mean estimates for different age classesof females in three speciesof European ducks. See text for !%.

Bmdscvy PropOtiO~ capture-recapture.suwival of yearlings SY + ASY SY ASY that breed

Northern Shoveler 0.12 0.38 0.58 0.70 Common Pochard 0.59 0.06 0.55 0.65 0.70 Tufted Duck 0.71 0.03 0.71 0.71 0.89

* Ducldhgs.

some females temporarily emigrated from per- found no evidence that annual variation in sur- manent sampling areas and bred in adjacent vival rates can be explained by differential pre- flooded reed-bedsthat were seldom used in years dation on nesting females. Two major predators with normal and low water levels), and (2) nest were known to kill nesting females at Engure successof pochard was the lowest of all three Marsh; the American Mink (73% of all killings) speciesand exhibited the highest annual varia- and the Marsh Harrier (15%). We estimated that tion (Blums et al. 1993); this evidently resulted on averageonly 2.8% (range0.0-l 3.6), 2.0% (O.O- in lowered capture rates in some years. 4.6), and 1.3% (range 0.0-4.0) of female shov- Capture probabilities did not appear to vary elers, pochards, and Tufted Ducks, respectively, by time for shovelers and Tufted Ducks. Both were found dead each year from the number of species had higher nest successthan pochard, available breedersat permanent sampling areas. showed relatively conservative nesting patterns We believe most dead females were found, and and were constrained to particular breeding hab- thus breeding-seasonmortality of thesethree duck itats, such as grassy islands (shoveler, Tufted speciesat Engure Marsh should have been much Duck) and Black-headedGull (Larus r&bun&s) lower than the 20-40% reported for female Mal- colonies located on floating mats of emergent lards in North American prairies (reviewed by vegetation (Tufted Duck). As a result, capture Sargeant and Raveling 1992, Greenwood et al. probabilities of thesetwo specieswere higher and 1995). Becausepredator control at Engure Marsh relatively constant during the entire study peri- was consistently effective, apparent nest success od. was high (long-term average 0.8 1,0.78, and 0.69 for Tufted Duck, shoveler, and pochard, respec- SURVIVAL OF ADULTS AND YEARLINGS tively) and breeding-seasonmortality of female Our data provide some of the first estimates of low. survival ratesfor European ducksthat were based The only direct comparison of annual survival on modem statistical models. Becauseour aging estimates obtained in this study can be made techniques permitted captured females to be with mark-resighting survival estimates of identified as SY and ASY, we estimated age- Northern Shoveler in Canada (Arnold and Clark, specific survival probabilities for SY, a topic unpubl.), but this survival rate (0.51) included about which little is known. pooled age groups of breeding females. No es- Annual survival probabilities of SY and ASY timates of Tufted Duck and pochard survival are females were the highest for Tufted Ducks (0.71 available. A review by Johnson et al. (1992) and for both age classes),intermediate for Common other recent studies(Hepp et al. 1987, Szymczak Pochards(0.55 and 0.65), and lowest for North- and Rexstad 199 1, Anderson et al., unpubl., Ar- em Shovelers (0.38 and 0.58). This pattern also nold et al., unpubl., Arnold and Clark, unpubl.), was reflected in maximum longevity of breeding provided survival estimates of the adult females femalesrecorded on Engure Marsh: Tufted Ducks (SY + ASY) for 11 speciesof North American and pochardsreached 2 14 years, whereasshov- ducks marked during either the breeding or pre- elers only reached age 10. seasonperiod. These estimatesranged from 0.47 Removal of predators increased nest success to 0.70 and were derived from band recovery, and it is possible that survival rates of female capture-recapture, and mark-resighting data. ducks at Engure Marsh were higher than those Thus, average survival probability (0.71) of of ducks in other breeding areas. However, we breeding female Tufted Ducks on Engure Marsh, SURVIVAL AND PHILOPATRY OF FEMALE DUCKS 7 1

Latvia, was among the highest ever recorded for within sampling areas and within all territory of female dabbling or diving ducks when estimates the marsh and found no evidence that SY fe- were based on modem statistical models. males moved significantlyfarther than ASY birds, Evidently some sea ducks exhibit even higher either after successfulor failed nesting. We con- survival with estimatesof 0.90 reported by Coul- cluded that much of the difference between sur- son (1984) for Common Eider in . An- vival estimates of SY vs. ASY females for shov- other long-term eider study (Baillie and Milne eler and pochard in this study concernsmortality 1982) from reported extremely high av- rather than emigration. eragelife expectancyof 26 years that should also be associatedwith very high survival rate. Dow SURVIVAL OF DUCKLINGS and Fredga (1984) obtained a high survival es- Survival estimates for ducklings, which reflect timate (0.77) for female Common Goldeneyes in both survival and permanent emigration, were Sweden. typically low, and averagesranged from 0.03 to Survival probabilities varied significantly by 0.12. The specieswith the highest SY and ASY age for pochards and shovelers (P < 0.01) but survival probability, the Tufted Duck, had the not for Tufted Ducks (P = 0.88). The only in- lowest duckling survival probability, and, con- formation on annual survival probabilities of SY versely, the specieswith the lowest SY and ASY and ASY females is from a long-term capture- survival probability, the Northern Shoveler, had recapture study of Wood Ducks in Missouri the highest duckling survival probability. Cap- (Dugger and Fredrickson, unpubl.). Dugger and ture-recapture models for annual survival esti- Fredrickson used program JOLLYAGE and mates of ducklings have not been used by other found some evidence that SY females exhibit researchers,so comparable data are not avail- lower survival than ASY females (0.45 vs. 0.60), able. There are two main reasons for these low however, the overall contingency chi-square test survival probabilities of ducklings, (1) ducks ex- was not significant. The test result may reflect hibit much higher mortality during the first year low power becauseof small sample sizes rather of life than later (Sargeant and Raveling 1992, than a true lack of age-dependence.Reynolds et Johnson et al. 1992), and (2) natal philopatry in al. (1995) found that SY female (Anus ducks is much weaker than breeding philopatry platyrhynchos)exhibited higher survival during (reviewed by Rohwer and Anderson 1988, An- spring-summer than did older females, but this derson et al. 1992). inference pertained to seasonsurvival and most Survival probabilities of shoveler ducklings likely resulted from differences in reproductive were higher than for both speciesof diving ducks. behavior of the two age groups (a large portion We believe shoveler ducklings have higher ab- of SY females do not breed and thus avoid mor- solute survival during the first year of life and tality by nest predators). In North Dakota, ASY this may be true also for other dabbling ducks and ATY female Gadwalls (Anus strepera) had (see Bengtson 1972). An alternative explanation significantly higher return rates than SY females might be that some proportion of female diving but the factors causing this difference were not ducks may have emigrated permanently for their identified (Lokemoen et al. 1990). Return rates breeding to other suitable areaswithin the marsh include the probabilities of surviving, returning or out of the marsh, and that this emigration is to the study area, and being recaptured or reob- responsiblefor the lower survival probability es- served, and are thus not always easy to interpret. timates for pochard and Tufted Duck ducklings. There is no reason to suspectthat permanent Although some evidence supports this notion emigration within the Engure Marsh may have (Blums et al. 1989, this study), we believe the affected survival estimates for any age category relative importance of emigration in the com- of breeding female shovelers (see below). How- plement of survival was minor for all three spe- ever, some SY and ASY female pochards did cies. Circumstantial inferencesabout emigration emigrate permanently out of sampling plots to can be drawn from the data on population com- other areas of the marsh, and, if this emigration position reportedby Mihelsons et al. (1986). They is age-specific,survival probability estimates for estimated that 9 1% of female shovelersbreeding SY birds may have been biased downwards. We at Engure Marsh were of local origin and only examined 1,787 short-distance(< 10 km) breed- 9% were immigrants. Most of the latter were ing dispersal movements of female pochards young first breeding females hatched outside En- 12 PETER BLUMS ET AL. gure Marsh. There is no evidence that emigration able to different mortality agents, most impor- within the marsh may have affected survival es- tantly to adverse weather. timates of shoveler ducklings becausenearly the entire nesting population was captured each year PHILOPATRY during the study. Similarity between survival rate estimates ob- Survival probabilities of Tufted Duck duck- tained using band recovery and capture-recap- lings were half that of pochard and may result ture models demonstrates that female Tufted either from higher post-hatch mortality, more Ducks and pochards at Engure Marsh are ex- permanent emigration out of sampling areas, or tremely philopatric and nearly all surviving birds a combination of these. Indeed, Blums et al. return to breeding sites used previously. Hepp et (1989) demonstrated that the degreeof long-dis- al. (1987) used this method to test philopatry in tance natal dispersal (“gross natal dispersal,” Wood Ducks but they compared survival rates sensu Greenwood 1980) seems to be higher for of two different populations that may have ex- Tufted Ducks than pochards.Blums et al. (1989) hibited different survival. The numbers of band- also reported that for all indirect recoveries of ings and recoveries used for our band recovery bands during April-November of ducklings analyses, however, were smaller than data sets banded in Latvia that were never detected re- commonly used by North American biologists. turning to their native wetland, revealed signif- Our data provide some evidence that not all fe- icantly more recoveriesfor Tufted Duck than for male Tufted Ducks are strongly philopatric: two pochard classified as long-distance natal dis- females immigrated to Engure Marsh from persal movements (65 vs. 23%, P < 0.001). How- breeding areas located as far as 170 and 290 km ever, the sex of emigrating birds in most cases to the north-east (Estonia) and north (Finland), was unknown. respectively. Both birds were marked and recap- Yearly variation in survival probabilities of tured on the nestsand, to our knowledge, are the ducklings did not parallel that of adults for any only confirmed casesof effectivebreeding in ducks of the three species,suggesting that different fac- after long-distance breeding dispersal. tors affectsurvival of ducklingsversus older birds. We did not test philopatry for shovelers be- Prefledgedducklings are more vulnerable to dif- cause banding and recovery data were too lim- ferent mortality agents, and the brood-rearing ited for reasonableuse with band recovery mod- period is especially important to the dynamics els. There is little evidence on permanent emi- of waterfowl populations (Sedinger 1992, John- gration out of Engure Marsh for breeding female son et al. 1992). shovelers (Mihelsons et al. 1986, Blums et al. Most mortality occurs during the first two 1989). Return rates to previous breeding sites weeks after hatching, when ducklings are still have been relatively high both at Engure Marsh, small and thermoregulatory ability is incom- Latvia (Blums and Mednis 1986) and Canada pletely developed (reviewed by Sargeant and (Sowls 1955; Arnold and Clark, in prep.). This Raveling 1992). We believe much of the varia- information provides evidence that female shov- tion in survival probabilities of ducklings during elers may be strongly philopatric. their first year of life can be linked to the 5-10 Recent advancesin development of models for day period immediately after hatch as reported combined analysis of band recovery and recap- by other investigators (Bengtson 1972, Hill and ture data (Szymczak and Rexstad 199 1, Bum- Ellis 1984, Savard et al. 199 1, Wayland and ham 1993) offer the potential for efficient use of McNicol 1994, Sayler and Willms, in press). all available band encounter information. The Mihelsons et al. (1986) suggestedthat not only use of these new models can likely improve pre- recaptures of females which return to breed on cision and permit separateestimates of survival the native marsh but also band recoveries of and permanent emigration probabilities if re- fledgedyoung of either sex, banded at hatch and covery and recapture data are incorporated in a never recorded returning to the birth place, can single model. Such models are efficient and pro- provide useful information on survival. It is pos- vide a formal test for philopatry. sible that the use of all band recovery informa- tion will permit additional inferences about ACKNOWLEDGMENTS duckling survival rates and the time period (soon We dedicatethis publication to the memory of H. Mih- after hatching) when ducklings are most vulner- elsons,a wonderful man of vision and complete ded- SURVIVAL AND PHILOPATRY OF FEMALE DUCKS 73

ication to waterfowl science,a mentor and supervisor Breeding populations of ducks on Engure Lake, of the study during 1958-1981. We thank Y. Kats, J. Latvia, for 35 years. Ring 15:165-169. Kazubiemis, M. Kazubieme, V. Klimpins, P. Leja, G. BLUMS,P. N., V. K. REDERS,A. A. MEDNIS,AND J. A. Lejins (deceased),D. Spals,A. Stipniece,and J. Viksne BAUMANIS.1983. Automatic drop-door traps for for the assistancewith field work, and many other peo- ducks. J. Wildl. Manaae. 47:199-203. ple who have contributed to the Latvian duck study. Bovn, H., J. HARRISON, & A. ALLISON. 1975. Duck T. W. Arnold. L. H. Fredrickson. G. R. Heoo. R. D. wings: a study of duck production. WAGBI Con- Drobney, J. RI Longcore,D. G. McAuley, P. x Magee, servation Publication. Caxton and Holmesdale and an anonymous reviewer provided helpful com- Press, Sevenoaks,UK. ments on the manuscript. The senior author was sup- BROWNIE,C., D. R. ANDERSON,K. P. BURNHAM,AND ported by Gaylord Memorial Laboratory (The School D. S. ROBSON. 1985. Statistical inference from of Natural Resources,University of Missouri-Colum- band recoverydata: a handbook. 2nd ed. U.S. bia, and Missouri Department of Conservation coop- and Wildl. Serv. Res. Publ. 156. erating) during the data analysesand preparation of BURNHAM.K. P. 1993. A theorv for combined anal- the manuscript. This is Missouri Agricultural Experi- ysis of ring recovery and recapture data, p. 199- ment Station project 183, Journal Series #12,373. 213. In J.-D. Lebreton and P. M. North [eds.], Marked individuals in the study of bird popula- LITERATURE CITED tion. Birkhauser,Basel, Switzerland. BURNHAM,K. P., ANDD. R. ANDERSON.1992. Data- AF~ON, A. D. 1984. Influence of age and time on basedselection ofan appropriatebiological model: reproductiveperformance of female LesserScaup. the key to modem data analysis,p. 16-30. In D. Auk 101:255-265. R. McCullough and R. H. Barret feds.], Wildlife AKAIKE,H. 1973. Information theory and an exten- 2001: populations. Elsevier, New York, NY. sion of the maximum likelihood principle, p. 267- CLOBERT.J., J.-D. LEBRETON.AND G. MARZOLIN. 281. In B. N. Petran and F. Csaki [eds.], Inter- 1990. The estimation of local immature survival national Symposium on Information Theory. 2nd rates and of age-specificproportions of breeders ed. Akademiai Kiadi, Budapest,Hungary. in bird populations,p. 199-2 13. In J. Blondel, A. ANDERSON,M. G., J. M. RHYMER, AND F. C. ROHWER. Gosler, J.-D. Lebreton, and R. H. McCleery [eds.], 1992. Philopatry, dispersal,and the geneticstruc- Population biology of Passerine birds: an inte- ture of waterfowl populations, p. 365-395. In grated approach. Springer-Verlag,Berlin. B.D.J. Batt et al. [eds.], Ecologyand management CLOBERT,J., J.-D. LEBRETON,D. ALLA~NE,AND J.-M. of breeding waterfowl. Univ. Minnesota Press, GAILLARD. 1994. The estimation of age-specific Minneapolis. breeding probabilities from recapturesor resight- BAILLIE,S. R., AND H. MILNE. 1982. The influence ings in vertebrate populations: II. Longitudinal of female age on breeding in the Eider Somateria models. Biometrics 50:375-387. mollissima. Bird Study 2955-66. COULSON,J. C. 1984. The population dynamics of BENGTSON,S.-A. 1972. Reproduction and fluctua- the Eider Duck Somateria mollissima and evi- tions in the size of duck populations at Lake My- dence of extensive non-breeding by adult ducks. vatn, Iceland. Oikos 23:35-58. Ibis 126:525-543. BLUMS,P., ANDA. MEDNIS. 1986. Dynamics of pop- COWARDIN,L. M., V. CARTER,F. C. GOLET,AND E. L. ulation composition and return rate in breeding LAROE. 1979. Classificationof wetlandsand deep- duckson EngureMarsh, Latvia. Var Fagelv. Suppl. water habitats of the United States. U.S. Dept. 11:17-21. Interior, Fish and Wildl. Serv. FWSOBS-79/3 1. BLUMS,P., AND A. MEDNIS. 1991. Management of DANE, C. W., AND D. H. JOHNSON.1975. Age deter- islands for breeding waterfowl on Engure Marsh, mination of female Redheadducks. J. Wildl. Man- p. 128-134. In C. M. Finlayson and T. Larsson age. 39~256-263. [eds.],Proc. of the workshopwetland management Dow, H. AND S. FREDGA. 1984. Factors affecting re- and restoration. Swedish Environmental Protec- productive output of the Goldeneye Duck Buce- tion Agency Report, Solna. phala clangula. J. Anim. Ecol. 53~679-692. BLUMS,P., AND A. MEDNIS. 1996. Secondarysex ratio DUGGER,K. M. 1991. Nesting parametersand pop- in Anatinae. Auk 113:in press. ulation dynamics of box-nesting female Wood BLUMS,P., ANDJ. VIKSNE. 1990. Managing of breed- Ducks in Mingo Swamp. M.S.thesis, Univ. Mis- ing habitats for waterfowl in Latvia. D. 196. In souri-Columbia, MO. G.V.T. Matthews [ed.], Managing waterfowl pop- GREENWOOD,P. J. 1980. Mating systems,philopatry ulations, IWRB Spec. Publ. No. 12. IWRB, Slim- and dispersalin birds and mammals. Anim. Be- bridge, UK. hav. 28:1140-l 162. BLUMS,P., A. MEDNIS,AND J. BAUMAN~S.1989. Long- GREENWOOD,R. J., A. B. SARGEANT,D. H. JOHNSON, distance dispersal in some European ducks. La L. M. COWARDIN,ANDT. L. SHAFFER.1995. Fac- Sauvagine309:44-5 1 (French). tors associatedwith duck nest successin the Prai- BLUMS,P., A. MEDNIS,AND J. D. NICHOLS. 1994. Re- rie Pothole Regionof Canada.Wildl. Monogr. 128. tention of web tags and plasticine-filled leg bands GRICE,D., ANDJ. P. ROGERS.1965. The Wood Duck applied to day-old ducklings. J. Wildl. Manage. in Massachusetts.Massachusetts Div. Fish. and 58:76-81. Game, Westboro. BLUMS,P., I. BAUGA,P. LEJA,AND A. MEDNIS. 1993. HARAMIS,G. M., AND A. D. NICE. 1980. An im- 14 PETER BLUMS ET AL.

proved web-tagging technique for waterfowl. J. REYNOLDS,R. E., R. J. BLOHM,J. D. NICHOLS,AND J. Wildl. Manage. 44:898-899. E. HINES. 1995. Spring-summer survival rates HEPP,G. R.. AND R. A. KENNAMER. 1993. Effects of of yearling versusadult females. J. Wildl. age and experience on reproductive performance Manage. 59:691-696. of Wood Ducks. Ecology 74:2027-2036. ROHWER,R. C., AND M. G. ANDERSON. 1988. Fe- HEPP, G. R., R. T. HOPPE,AND R. A. KENNAMER. male-biased philopatry, monogamy, and the tim- 1987. Population parameters and philopatry of ing ofpair formation in migratory waterfowl. Curr. breeding female Wood Ducks. J. Wildl. Manage. Omithol. 5: 187-22 1. 51:401404. SARGEANT,A. B., ANDD. G. RAVELING. 1992. Mor- HILL, D. A., AND N. ELLIS. 1984. Survival and age tality during the breeding season,p. 396422. In related changesin the behavior and time budget B.D.J. Batt et al. [eds.], Ecologyand management of Tufted ducklingsAythya fuligula. Ibis 126:544- of breeding waterfowl. Univ. Minnesota Press, 550. Minneapolis. HINES,J. E., ANDJ. R. SAUER. 1989. Program CON- SAUER,J. R., ANDB. K. WILLIAMS. 1989. Generalized TRAST: a generalprogram for the analysisof sev- proceduresfor testing hypothesesabout survival eral survival or recovery rate estimates.U.S. Fish or recovery rates. J. Wildl. Manage. 53: 137-142. and Wildl. Serv., Tech. Rep. 24. SAVAFCD. J.-P. L.. G.E.J. SMITH. AND J.N.M. SMITH. JOHNSON,D. H., J. D. N~crrors,AND M. D. SCHWARTZ. 1991. Duckling mortality in Barrow’s Goldeneye 1992. Population dynamics of breeding water- and Buffleheadbroods. Auk 108:568-577. fowl, p. 446485. In B.D.J. Batt et al. [eds.], Ecol- SAYLER,R. D., AND M. A. WILLMS. In press. Brood ogyand managementofbreeding waterfowl. Univ. ecology of Mallards and Gadwalls nesting on is- Minnesota Press, Minneapolis. lands in deep-water reservoirs.J. Wildl. Manage. LEBRETON,J.-D., K. P. BURNHAM,J. CLOBERT,AND D. SEBER,G.A.F. 1982. The estimation of animal abun- R. ANDERSON. 1992. Modeling survival and test- dance and related parameters.2nd ed. Macmillan ing biological hypothesesusing marked : Publ. Co., New York. a unified approach with case studies. Ecol. Mon- SEDINGER,J. S. 1992. Ecology of prefledgingwater- ogr. 62:67-l 18. fowl, p. 109-127. In B.D.J. Batt et al. [eds.], Ecol- LOKEMOEN, J. T., H. F. DUEBBERT,AND D. E. SHARP. ogyand managementof breedingwaterfowl. Univ. 1990. Homing and reproductive habits of Mal- Minnesota Press, Minneapolis. lards. Gadwalls. and Blue-winaed Teal. Wildl. SERIE. J. R., D. L. TRAUGER,H. A. Don, AND D. E. Monbgr. 106. SHARP.. 1982. Age-class determination of Can- MIHELSONS,H., J. VIKSNE,AND A. MEDNIS. 1976. The vasbacks.J. Wildl. Manage. 46:894-904. improvement of waterfowl nesting sites, p. 345- Sowrs, L. K. 1955. Prairie ducks: a study of their 348. In Proc. Int. Conf. on the Conservation of behavior, ecology and management. Stackpole Wetlands and Waterfowl. IWRB, Slimbridge,UK. Books, Harrisburg, PA. MIHELSONS,H., A. MEDNIS, AND P. BLUMS. 1986. SZYMCZAK,M. R., ANDE. A. REXSTAD. 1991. Harvest Population ecology of migratory ducks in Latvia. distribution and survival of a Gadwall population. Publ. House Zinatne, Riga (Russian,with English J. Wildl. Manage. 55:592-600. summarv). TRAUGER,D. L. 1974. Eye color of female Lesser PALMER,R. S: 1976. Handbook of North American Scaupin relation to age. Auk 91:243-254. birds. Vol. 2:498-5 15. Yale Univ. Press.New Ha- WAYLAND. M.. AND D. K. MCNICOL. 1994. Move- ven, CT. ments andsurvival of Common Goldeneyebroods POLLOCK,K. H. 1981. Capture-recapturemodels al- near Sudbury,Canada. Can. J. Zool. 72:1252-1259. lowing for age-dependent survival and capture WELLERM. W.. AND C. S. SPATCHER.1965. Role of rates. Biometrics 37:521-529. habitat in the distributionand abundanceof marsh POLLOCK,K. H., J. D. NICHOLS,C. BROWNIE,AND J. birds. Iowa State University Agric. and Home E. HINES. 1990. Statisticalinference for capture- EconomicsExp. StationSpec. Rep. 43. Ames, Iowa. recantureexperiments. Wildl. Monow. 107. WHITE,G.C. 1983. Numerical estimation of survival PRADEL,R. J., J. &BERT, ANDJ. D. LEBGTON. 1990. rates from band recovery and biotelemetry data. Recent developments for the analysisof capture- J. Wildl. Manage. 47:716-728. recapturemultiple data sets:an example concem- ing two Blue Tit populations. Ring 13:193-204.