J. Field Ornithol., 67(1):86-99 FOOD HABITS OF DMNG DUCKS IN THE GREAT LAKES AFTER THE ZEBRA MUSSEL INVASION CHRISTINE M. CUSTER 1 National BiologicalService Northern Prairie Science Center P.O. Box 818 La Crosse,Wisconsin 54602 USA THOMAS W. CUSTER 1 National BiologicalService Patuxent Environmental Science Center PO. Box 818 La Crosse,Wisconsin 54602 USA Abstract.--Zebramussels (Dreissena polymorpha) invaded the Great Lakesin the mid-1980s and quicklyreached high densities.The objectiveof this studywas to determine current consumptionof zebramussels by waterfowlin the Great Lakesregion. Feeding Lesser Scaups (Aythyaaffinis), Greater Scaups (A. marila),Canvasbacks (A. valisineria),Redheads (A. amer- icana), Buffleheads(Bucephala albeola) and Common Goldeneyes(B. clangula)were col- lected in western Lake Erie and in Lake St. Clair between fall and spring, 1992-1993 to determinefood habits.All 10 Redheads,97% of LesserScaups, 83% of Goldeneyes,60% of Buffleheadsand 9% of Canvasbackscontained one or more zebra musselsin their upper gastrointestinaltracts. The aggregatepercent of zebra musselsin the diet of LesserScaups washigher in Lake Erie (98.6%) than in Lake St. Clair (54.4%). Zebra mussels(aggregate percent) dominatedthe diet of Common Goldeneyes(79.2%) but not in Buffleheads (23.5%), Redheads(21%) or Canvasbacks(9%). LesserScaups from Lake Erie fed on larger zebra mussels(• = 10.7 +_0.66 mm SE) than did LesserScaups from Lake St. Clair (• = 4.4 +_0.22 mm). LesserScaups, Buffleheads and Common Goldeneyesfrom Lake Erie con- sumed zebra mussels of similar size. it/•rros ALIMENTICIOS DEL PATOS ZAMB•ORF• F_aNLOS GRANDES LAGOS LUEGODE LA INVASIONDE LA ALMEJADREISSENA POLYMORPHA Sinopsis.--Laalmeja Dreissena polymorpha invadi6 los Grandes Lagos a mediadosde la dficada del 1980 yen poco tiempo alcanz6densidades sumamente altas. E1 objetivo de esteestudio fue determinarel consumode estaalmeja por parte de los patosque utilizan estoslagos. Se coleccionaronindividuos de Aythia affinis,A. marila,A. valisineria,A. americana,Bucephala albeolay de B. clangulaentre el otofio y la primaverade 1992-1993, en los lagosErie y St. Clair,para determinarlos h•bitosalimenticios de estasespecies. E1 100% de los especimenes de A. amehcana,97% de A. affinis,83% de B. clangula,60% de B. albeola,y el 9% de A. valisineriaingirieron almejas. E1 porciento agregado de almejasen la dieta de Aythiaaffinis fue mayoren el Lago Erie (98.6%) queen el Lago St. Clair (54.4%). Las almejas(porciento agregado)dominaron la dieta de B. clangula(79.2%) pero no aslla de A. affinis,A. amer- icana, o A. valisineria.Los individuosde A. affinisdel Lago Erie se alimentaronde almejas milsgrandes (• = 10.7 _+0.66 mm ES) que losindividuos del Lago St. Clair (• = 4.4 +_0.22 mm). Los individuosde A. affinis, B. albeolay de B. clanguladel Lago Erie se alimentaron de almejasde tamafio similar. Currentaddress: National BiologicalService, Upper Mississippi Science Cent• P.O. Box 818, La Crosse,Wisconsin 54602 USA. 86 Vol.67, No. 1 FoodHabits of Diving Ducks [87 Zebra mussels(Dreissena polymorpha) have reached high densities,es- pecially in Lake Erie and Lake St. Clair (Leach 1993), since their intro- duction into the Great Lakes in the mid-1980s. Densities of zebra mussels over 700,000/m2 havebeen reported at powerplants on Lake Erie (Ko- valak et al. 1993) and as many as 342,000/m 2 on fish-spawningreefs in Lake Erie (Leach 1993). In Europe, Tufted Ducks (Aythyafuligula), Greater Scaups(A. marila), Common Pochards(A. ferina) and Common Goldeneyes(Bucephala clan- gula) commonly consumezebra mussels(de Vaate 1991, Olney 1963). In the United States,Mitchell and Carlson (1993) reported that 19 of 21 LesserScaups (Aythya affinis) entrained into a power plant in Michigan containednearly 100% zebra musselsin their esophagiand/or proven- triculi. Hamilton (1992a) and Wormington and Leach (1992) found piec- es of zebra musselshell in eight waterfowl gizzardsfrom Lake Erie (four Buffleheads [Bucephalaalbeola], two LesserScaups, one Greater Scaup and one Common Goldeneye). The sample size in that study was small, however, and gizzard data are biased becauseof differential digestion rates of soft and hard food items (Swansonand Bartonek 1970). Zebra musselshave the potential to affect waterfowldistribution and abundance (Stanczykowskaet al. 1990, Wormington and Leach 1992) in North America. Additionally, the bioaccumulation capacitiesof zebra mussels(Brieger and Hunter 1993, Buschand Schuchardt1991, Mersch et al. 1992) may enhance the transfer of contaminantsto waterfowl (de Kock and Bowmer 1993). Contaminants can negatively affect waterfowl reproduction (de Kock and Bowmer 1993) and may have secondaryef- fects as a contaminant source for Bald Eagles (Heliaeetusleucocephalus) and humans. Our objectiveswere to determine which diving ducks cur- rently consumezebra musselsin the U.S. portion of the Great Lakesand what proportion of their diet is now comprisedof zebra mussels. METHODS We collecteddiving ducksduring fall (prior to ice coverageof the lakes, November-December), mid-winter (lakes frozen, January-February) and spring (beginning of ice break-up, March-April) from January 1992 to April 1993 in three general areas of Lakes Erie and St. Clair (Fig. 1). These three general collection areas represented the major waterfowl concentrationsin the U.S. portions of these two lakes (Bookhout et al. 1989). We collectedducks by shooting(Federal Permit PRT-673019,Mich- igan Permit SC- 785, Ohio Permits249, 167) either from shore or from boats,generally after we had observedthe ducksfeeding. We immediately removed contentsof the esophagusand proventriculus(upper GI tract) and stored them separatelyin 95% ethanol. Esophagealand proventric- ular contentswere combined for analysisto maximize samplesize (Afton et al. 1991). We determined the age and sex of waterfowl using plumage and cloacal characteristics(Carney 1964). All food items were identified to genus or speciesand number of in- dividual items counted. Frequencyof occurrence (number of waterfowl 88] C. M. Custerand T. W. Custer J.Field Ornithol. Winter 1996 Lake St, Clair Toledo,OH••],? •] LakeErie F[GUV•1. Three generallocations where six speciesof divingducks were collectedduring fall, winter, and spring,January 1992-April 1993. with a particular food taxon divided by the total number of waterfowl) and aggregatepercent (proportion of each food item in each bird aver- aged for all individuals;Swanson et al. 1974) were calculatedfor each waterfowl species.When the upper GI tract contained only animal ma- terial (n = 74), we calculatedaggregate percent based on numericaldata, rather than dry massor volume, becausefood items were generally the samesize (averagelength of zebra mussels= 8.1 mm, isopods= 7.8 mm, amphipods = 5.5 mm, caddisflies= 8.8 mm) and becausemolluscs were an important componentof the diet. When upper GI tract samplescon- tained both plant matter and animal matter (n = 8 Redheads,2 Buffle- heads), volumetric measurements of food items were taken and the vol- umes used to calculateaggregate percent (Bartonek and Hickey 1969, Gammonleyand Heitmeyer 1990). Aggregatepercent basedon numbers is preferable to aggregatepercent basedon dry masswhen molluscsare an important component of the diet becausedry massinflates the im- portanceof molluscsin the diet due to the proportionatelylarge massof Vol.67, No. 1 FoodHabits of Diving Ducks [89 undigestibleshell material.Additionally, numerical countsare lessprone to measurementerror than either the dry massor volume methods,which is important when measuringsmall volumes or dry massesas in our study. Aggregatepercent basedon dry massdoes facilitate making energyand nutritional inferences.Dry massof zebra musselsin our study can be estimatedbased on their shell length (Draulans 1982). Even though methodsto determinepercent composition of diet differsbetween stud- ies,we follow the commonpractice of comparingour data to other studies (Afton et al. 1991,Dirschl 1969, Gammonleyand Heitmeyer1990, Hoppe et al. 1986). Shell length of ingested zebra mussels-•5 mm was measured to the nearest mm and those •5 mm were measured to the nearest 0.2 mm. All zebra musselswere measuredalong the longestaxis of the shell (Hamil- ton 1992b). Average size of zebra musselswas calculated for each duck and thosemeans used for statisticalcomparisons. An individualduck was the measurementunit for zebra musselsize comparisonsamong species, ages, sexesand locations. Data were analyzedin a step-wisemanner. We used analysisof variance (ANOVA) on aggregatepercent and zebra musselsize data. Bartlett'stests were used to testfor homogeneityof variancesprior to each ANOVA. If varianceswere not equal, data were rank transformed,which in all cases equalizedthe variances.Untransformed means _ 1 SE are presentedin text and tables.When main factorsand the interaction term in 2-way ANOVAswere non-significant,these factors were combinedin subsequent 2-wayor 1-wayanalyses. Similarly, sequential chi-square and/or Fisher's Exact testswere used on frequency-of-occurrencedata; sexes,ages, and seasonswere combinedin subsequentanalyses as these factors were found to be not significantlydifferent. Differences in size distributions of zebra musselsconsumed by waterfowl were tested with Kolmogorov-Smirnov two-sampletests. RESULTS We collected 60 LesserScaups, 20 Buffleheads,11 Common Golden- eyes,14 Canvasbacks,10 Redheadsand four Greater Scaups.Of these,41 LesserScaups (the numbers collectedin fall and spring were nine and 23; eight in winter),