J. Field Ornithol., 67(1):86-99

FOOD HABITS OF DMNG 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 tamafiosimilar.

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 -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 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 ). 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 aver- aged for all individuals;Swanson et al. 1974) were calculatedfor each waterfowl species.When the upper GI tract contained only 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 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), 15 Buffleheads (nine and six), six Common Golden- eyes (three and three), 11 Canvasbacks(five and five; one in winter), 10 Redheads (six and four), and one Greater Scaup (one and none) had food in their upper GI tracts.Each duck with food in its esophagusor proventriculuswas included in the analyses. Springand fall diets.--Ninety-sevenpercent (31 of 32) of LesserScaup upper GI tractscontained at least one zebra mussel(Table 1); therefore, no statisticalcomparisons of frequency of occurrence between seasons, agesor sexeswere made. Frequencyof occurrenceof zebra musselsdid not differ between Lake Erie and Lake St. Clair (Fisher's Exact, n = 32, P = 0.156). The aggregatepercent zebra musselsin the upper GI tract of Lesser 90] c. M. Custerand T. W. Custer J.Field Ornithol. Winter 1996

T,•I•E 1. Frequencyof occurrenceand aggregatepercent of foods in the upper gastroin- testinaltract (esophagusand proventriculus)of five speciesof diving ducksfrom the westernend of Lake Erie and Lake St. Clair during fall and spring,1992 and 1993. Species,location, Frequencyof Aggregate% food taxa (n) occurrence (%) ,• SE LesserScaup (32) Lake Erie (27) zebra mussels 27 (100) 98.6 0.94 Amnicola walkeri 1 (4) 0.2 0.19 Physellaintegra 1 (4) 0.9 0.93 Sphaeriidae 1 (4) (0.1 Isopoda Caecidoteasp. 2 (7) 0.2 0.13 Annelida 2 (7) 0.1 0.11 Lake St Clair (5) Mollusca zebra mussels 4 (80) 54.4 19.92 Amnicola walkeft 3 (60) 2.6 1.58 Pleuroceraacuta 2 (40) 20.2 19.95 Physellaintegra 1 (20) 1.0 1.00 Valvatasp. 1 (20) 0.8 0.80 Gyraulussp. 1 (20) 1.2 1.20 Sphaeriidae 1 (20) 1.2 1.20 Ferrissiaparallela 1 (20) 0.1 0.10 Amphipoda Hyalella/Gammarusspp. 4 (80) 14.4 12.21 Isopoda Caecidoteasp. 1 (60) 3.0 1.48 Diptera 1 (40) 1.3 1.18 Fish 1 (20) (0.1 Bufflehead (15) Lake Erie Mollusca zebra mussels 9 (60) 23.5 8.83 Amnicolawalkeri 1 (7) 0.6 0.60 Amphipoda Hyalella/Gammarusspp. 10 (67) 46.6 10.42 Isopoda Caecidoteasp. 3 (20) 8.8 6.72 Trichoptera Hydropsychesp. 1 (7) 4.8 4.80 Tricopteraadult 1 (7) (0.1 Odonata Enallag•nasp. 1 (7) 3.6 2.45 Crustacea 1 (7) 1.1 1.13 wildcelerywinterbuds 1 (7) 6.1 6.13 Polygonurnsp. seeds 1 (7) 6.7 6.67 (6) Lake Erie Mollusca zebra mussels 5 (83) 79.2 16.35 Physellaintegra 1 (17) 0.3 0.33 Vol. 67, No. 1 FoodHabits of Diving Ducks [91

T•d•LE 1. Continued.

Species,location, Frequencyof Aggregate% food taxa (n) occurrence (%) • SE Amphipoda Hyalella/Gammarusspp. 2 (33) 5.8 5.44 Isopoda Caecidoteasp. 2 (33) 14.8 11.03 Canvasback(11) Lake Erie (3) and Lake St. Clair (8) Mollusca zebra mussels 1 (9) 9.1 9.09 Plants wildcelerywinterbuds 10 (91) 90.9 9.09 Redhead (10) Lake Erie (4) Mollusca zebra mussels 4 (100) 50.5 28.58 Plants Potamogetonsp. 2 (50) 49.5 28.58 Lake St. Clair (6) Mollusca zebra mussels 6 (100) 1.3 0.98 Plants Potamogetonsp. 6 (100) 98.7 0.98

Scaupsin Lake Erie (98.6%) wasgreater than in Lake St. Clair (54.4%) (Table 1, F•,29= 17.97, P < 0.001). Aggregatepercent of zebra mussels in the upper GI tract did not differ betweenages or seasons(age, season, and interaction,F•,2s = 0.56, < 0.01, 2.39; P = 0.463, 0.993, 0.136, re- spectively)or betweensexes (F•,25 = 1.48, P = 0.235) for LesserScaups in Lake Erie. Food habits of LesserScaups in Lake St. Clair were more variable than food habits in Lake Erie; six taxa comprised>95% of the food itemsin Lake St. Clair comparedwith only one taxon,zebra mussels, comprising>95% of the food items in Lake Erie. Sixtypercent of Buffleheadscontained one or more zebramussels (Ta- ble 1). Frequencyof occurrenceof zebra musselsdid not vary by sex (Fisher's Exact, P = 1.00, n = 15) or season (P = 0.136, n-- 15). No one food item dominated in the upper GI tracts of Buffleheads (Table 1). Approximately47% of Buffleheadfood itemsconsisted of am- phipods.Nearly 24% of Buffleheaddiet wascomposed of zebra mussels; aggregatepercent did not differ by sexor season(sex, season, and inter- actionF1, 5 = 0.08, 0.29, < 0.01; P = 0.793, 0.613, 0.954, respectively). Five of six Common Goldeneyescontained at least one zebra mussel (Table 1). Zebra musselsdominated in the diet (79%); isopodsand am- phipodsalso were consumed. Wildcelerywinterbuds (Vallisneria americana) were the dominantfood (91%) in Canvasbackupper GI tracts (Table 1). Only one Canvasback, collected from Lake Erie, contained zebra musselsin its upper GI tract. 92] c.M. Custerand T. W. Custer J.Field Ornithol. Winter 1996

TA•I•Ir 2. Size (mm) of zebra musselsconsumed by six speciesof divingducks in the western end of Lake Erie and Lake St. Clair during fall and spring, 1992 and 1993.

Locationand species • SE na Minimumb Maximumb

Lake Erie Greater Scaup 13.4 1 2.4 23 Common Goldeneye 12.0 Ac 1.55 5 1.6 23 LesserScaup 10.7 A 0.60 27 0.8 23 Canvasback 9.7 1 2.8 15 Bufflehead 7.7 AB 1.45 9 0.8 18 Redhead 6.3 B 2.61 4 0.2 18

Lake St. Clair LesserScaup 4.4 0.22 4 0.8 9 Redhead 3.2 0.25 6 1.2 7

a Number of ducksused to calculatemean zebra musselshell lengths. b Minimum and maximum sizesdetermined from all zebra musselsconsumed by each species. cMeans with sameletter are not significantlydifferent (F = 3.51; df -- 3,42; P = 0.023, Bonferroni mean separation).

This individual was collected while it was dabbling along the shoreline during a period of low water causedby a seiche (Bookhout et al. 1989). One Canvasbackand one LesserScaup were collectedfrom a mixed feed- ing flock of LesserScaups and Canvasbacks.The Canvasbackhad only wildcelery winterbuds and the Lesser Scaup had only zebra musselsin their upper GI tracts. All 10 Redheadscontained at leastone zebra musselin their upper GI tracts (Table 1). The difference in aggregatepercent of zebra mussels betweenLakes Erie and St. Clair approachedsignificance (F•,8 = 3.51, P = 0.098). This lack of difference resultedfrom the large varianceof Lake Erie Redheads;two of four Redheadscontained 99% vegetation,whereas the other two Redheads contained 100% zebra mussels. The zebra mus- selsthat eight of 10 Redheadswere eating (six of six Redheadsfrom Lake St. Clair and two of four Redheads from Lake Erie) were attached to the stemsof vegetation. Only one Greater Scaup(immature female from Lake Erie) contained food in its upper GI tract. Ninety-nine percent of the food in its upper GI tract waszebra musselsand 1% wasa snail,Pleurocera sp. Size of zebra mussels.--Sizeof zebra musselsconsumed did not differ betweenmale and female LesserScaups (Table 2, F•,•8= 2.69, P = 0.118) or betweenmale and female Buffleheads(Table 2, F•,7= 1.12, P -- 0.325). There were no seasonaldifferences in sizeof zebra musselseaten by Less- er Scaups(Table 2, Fl,18= 2.57, P = 0.126) or Buffleheads(F1,7 = 1.21, P = 0.308). LesserScaups from Lake Erie ate larger zebra musselsthan Lesser Scaupsfrom Lake St. Clair (Table 2, F•,•0 = 16.19, P < 0.001). Using Lake Erie data only, LesserScaups, Buffleheads and Common Gol- deneyesconsumed zebra musselsof similar size, and LesserScaups and Vol.67, No. I FoodHabits of Diving Ducks [93

bufflehead

size = 5.5 mm) (mldianzebramussei-

lesser scaup lO

(m©dlan zebra muaael alz© = 10.3 mm)

1-2 3-4 5-6 7-8 9-10 11-12 13-14 15-16 17-18

Mean zebra mussel shell length (mm)

FIGURE2. Size distributionof zebra musselsfrom the upper gastrointestinaltracts of Lesser Scaupsand Buffleheadscollected from Lake Erie during fall 1992 and spring 1993.

Goldeneyesconsumed larger zebra musselsthan Redheads(F•,42 -- 3.51, P: 0.023). The size of zebra musselsconsumed by Redheads did not differ be- tweenLake Erie and Lake St. Clair (Table 2, F•,8-- 0.70, P = 0.426). The averagesize of zebramussels eaten by Redheadswhen zebramussels were the only food presentin the upper GI tract was 10.1 ___0.27 mm (n = 2). Averagesize of zebra musselsattached to vegetationin the Redhead's upper GI tract was 2.9 - 0.43 mm (n = 8). Although the mean size of zebra musselseaten by Buffleheadsand LesserScaups did not differ, the size distribution of zebra musselscon- sumedby LesserScaups was larger than thoseconsumed by Buffleheads (Kolmogorov-Smirnov,P < 0.05) (Fig. 2). The median sizeof zebra mus- 94] C. M. Custerand T. W. Custer J.Field Ornithol. Winter 1996 selseaten by Buffleheadswas 5.5 mm compared with 10.3 mm for Lesser Scaups. Winter season.--Eightof nine LesserScaups collected in an area of open water influenced by the hot water dischargeof a power plant con- tained 100% gizzardshad (Dorosoma cepedianum); one LesserScaup con- tained 100% zebra mussels. One Canvasback collected from the same area contained100% gizzardshad also.

DISCUSSION

Zebra musselsare now the dominant food consumedby LesserScaups and Common Goldeneyesin western Lake Erie and northern Lake St. Clair. Zebra musselswill probably remain an abundant food sourcefor waterfowl in Lakes Erie and St. Clair in the future (Leach 1993). Zebra musselsare lessimportant in the diets of Buffleheadsand Redheads,and of little importance to Canvasbacks.The dominance of zebra musselsin the diet of LesserScaups was expected because of the wide varietyof food items and the predominanceof molluscsin the diet of migrating and winteringLesser Scaups in somegeographic areas. Nearly 100% of Lesser Scaupdiet was dwarf surfclams(Mulinia lateralis)in Louisiana (Harmon 1962), 79% was fingernail clams (Sphaeriumsp.) in Iowa (Thompson 1973), and 50% wasAsiatic clams (Corbiculafluminea) in South Carolina (Hoppe et al. 1986). In Canada and Minnesota, 33-85% of the diet was amphipods (Afton et al. 1991, Dirschl 1969), whereaschironomids dom- inated in California (44%, Gammonleyand Heitmeyer 1990) and Loui- siana (46%, Afton et al. 1991). Food habitsof LesserScaups consisted of gizzardshad during periods when the lakeswere iced covered,usually during January and February (Assel et al. 1983). Waterfowl were forced to leave or move into the re- maining areasof open water (CMC, unpubl. data) at the mouthsof rivers, power plant water dischargepoints, and ice leads. Shad were plentiful and may have been the only food availablein the power plant discharge areas (CMC, pers. obs). Only about 25% of Bufflehead diets was zebra musselsin Lake Erie; amphipodsdominated (47%). Buffleheadsin Californiaconsumed main- ly seedsof Scirpusrobustus, Polygonum lapathifolium and Potamogetonpec- tinatus (34% of diet), corixids (water boatman, 29%) and chironomids (22%); molluscscomprised -•11% of the diets (Gammonleyand Heit- meyer 1990). Cottam's (1939) compilationof waterfowldiets (mainly from gizzards)indicated that (41%), (17%) and mol- luscs (16%) were the main foods in the diet of Buffleheads. Nearly80% of CommonGoldeneye diet waszebra mussels in Lake Erie. Prior to the invasionof the zebra mussels,Common Goldeneyediets in the nearbyDetroit Riverwere comprisedof 37% wildcelerywinterbuds, 9% oligochaetesand 7% decapods;molluscs were found in only trace amounts (Jonesand Drobney 1986). Cottam (1939) indicated that crus- taceans(32%) and insects(28%) were the dominant componentsin Vol.67, No. 1 FoodHabits of Diving Ducks [95

Common Goldeneyediets (mainly from gizzards)with molluscscompris- ing <10%. Canvasbacksfed primarily on wildcelery winterbudsin Lakes Erie and St. Clair, probably because this aquatic vegetation was widely available (CMC, pers. obs.). Few individualsconsumed zebra mussels.Canvasbacks switchedto a diet of Balticmacoma (Macoma balthica, 82%) (data mainly from gizzardsamples, Perry and Uhler 1988) when the submersedaquatic vegetationdeclined in ChesapeakeBay (Bayleyet al. 1978, Haramis and Carter 1983, Orth and Moore 1981). In North and South Carolina aquatic vegetationpredominated in the diet of Canvasbackwhen available,oth- erwise Baltic macoma dominated (gizzard data, Perry and Uhler 1982). On the upper MississippiRiver Canvasbacks consumed ->99% wildcelery winterbuds and Sagittaria rigida tubers (Korschgen et al. 1988) even though invertebratefoods were available(C. E. Korschgen,pers. comm.). The Canvasbackthat was collected while feeding alongside a Lesser Scaupwas eating 100% wildcelerywinterbuds whereas the LesserScaup was eating 100% zebra mussels.This may suggesta preference for wild- celery winterbudsby Canvasbacksand a preferencefor zebra musselsby Lesser Scaups.The one Canvasbackthat was eating zebra musselswas apparentlytaking advantageof zebra musselsthat were exposeddue to low water conditions created by a seiche. In Lakes Erie and St. Clair eight of 10 Redheads consumed mainly above-groundportions of Potamogetonspp. and incidentlyingested zebra musselswhich were attached to the vegetation stems.In two of six Red- heads in Lake Erie, however,zebra musselswere the only food in the GI tract. This dominance of vegetation in the diet is consistentwith other studiesof Redhead food habits.Redheads in North Carolina ate predom- inatelyshoalgrass (Halodule wrightis) (from gizzarddata mainly,Perry and Uhler 1982) as did Redheadsin Louisiana (83% & 67% of diet as deter- mined from esophagiand proventriculi) (Michot and Nault 1993) and Texas (Koenig 1969). On the MississippiRiver in Wisconsin,65% of Re- dhead's diet was wildcelery winterbuds (Korschgenet al. 1988). Zebra musselsconsumed by ducks,except for the eight Redheadsmen- tioned above,were presentas individualmussels. We did not find clumps of zebra musselsstill held together by their byssalthreads nor zebra mus- selsstill attached to a substratesuch as vegetationor rock. This was true even for Buffleheadsthat consumedvery small zebra mussels.This ob- servation suggeststhat zebra musselswere either plucked individually from the substrateor if pulled off in clumps then the clumpswere sep- arated in the duck's mouth prior to swallowing. Waterfowl from both lakesthat were specificallyforaging on zebra mus- selsconsumed zebra musselsthat averaged8-12 mm long. Musselsthat were attachedto vegetation,primarily from Lake St. Clair, and inciden- tally consumedwere smaller (• = 3 mm). The size distributionof zebra mussels in the areas where we collected waterfowl is unknown, but water- fowl consumedsizes close to the size deemed most profitable for female (9-10 mm) and male (13-14 mm) Tufted Ducks (de Leeuw and Van 96] C. M. Custerand T. W. Custer J.Field Ornithol. Winter 1996

Eerden 1992). The modal sizeof zebra musselseaten by six ducksat Point Pelee, Ontario was 11-13 mm (Hamilton et al. 1994). This sizewas larger than the modesin our study;however, Hamilton et al. (1994) disregarded all mussels<5 mm in length when calculating modes, so their modal value is biasedhigh. In our study,38% of zebra musselseaten by Lesser Scaupand 50% of zebra musselseaten by Buffleheadwere --•5 mm. numbers, especiallyLesser Scaups and Canvasbacks,and their duration of stay have increased on Lake Erie and Lake St. Clair since the arrival of zebra mussels(Wormington and Leach 1992; E. N. Kafcasand J. Weeks,unpubl. data). Canvasbackseem to have benefitted indirectly from the arrival of zebra musselsbecause the increasedwater clarity (Holland 1993) has allowedwildcelery to proliferate (E. N. Kafcas, pers. commun.). This alteration of distribution patterns could increase survivalrates for wintering diving ducks becausesome do not travel the additional 500-1300 km to the Atlantic or Gulf Coastsand some migrate late, after the hunting seasonshave closed.In contrast,more than 15,000 diving ducksdied of starvationin Switzerlandand the Netherlandswhen their feeding areasfroze over for an extended period in late winter (Suter and Van Eerden 1992). Body reservesof these were not sufficient to survivea 29-d period when ice coveredzebra musselbeds and physi- ological condition prevented them from moving south so late in the win- ter. Body massesof Lesser Scaupsand Buffleheads in this study (CMC, unpubl. data) were similar to body massesreported from springmigrants in California (Gammonley and Heitmeyer 1990) and wintering waterfowl in New York (Ryan 1972). Lipid contentof LesserScaups from Lake Erie (5-46%, CMC, unpubl. data) wasalso similar to lipid levelsin fall migrant LesserScaups in Manitoba (Austin and Fredrickson1987). These mass comparisonsmay indicatethat zebra musselsare an adequatewinter food. Energetic constraintsof zebra musselsas a food should be examined ex- perimentally, however.Waterfowl may need to consume greater volumes of and spend more time feeding when consumingzebra mussels,a low- energy food (4389 cal/g dry mass [Stanczykowskaand Lawacz 1976]) comparedwith other higher energyfoods (Driver et al. 1974,Jorde et al. 1995, Perry et al. 1986). The low energycontent of zebra musselsmay be partially compensatedfor, however,by their high abundance and easy availability. Because zebra mussels are known to accumulate contaminants such as polychlorinatedbiphenyls (Brieger and Hunter 1993), it will be important to determine whether waterfowlare being affected by increasedcontam- inant exposure,especially in areasof known pollution suchas the Detroit and Maumee Rivers (Hoke et al. 1990, Smith et al. 1985). Reduced re- production has been identified in Tufted Ducks fed zebra musselsfrom a contaminated basin in the Netherlands (de Kock and Bowmer 1993).

ACKNOWLEDGMENTS We thank A. E Boysen,T. Cichecki,G. Chandler,R. Hamric, C. O. Kochanny,E. N. Kafcas, K. Lee and L. W. Tischfor help collectingwaterfowl; N. R. Adams,R. W. Ainslie,C. W. Blair, Vol.67, No. 1 FoodHabits of Diving Ducks [97

J. Bresciami,Buds Fine Dining, Burr Marine,Schmidt Marine, Selfridge Air NationalGuard Base,Surf North Restaurant,Toledo EdisonPower Plant and U.S. Army Corpsof Engineers for accessto their properties;J.R. P.French, III, M. T. Lesinskiand B. A. Mannyfor assistance in identifyingfood items, and J. E. Austin,J. H. Gammonley,C. E. Korschgen,W. L. Green, M. E. Heitmeyer,G. A. Swanson,and four anonymousreviewers for commentson the manu- script.

LITERATURE CITED

A_vrON,A.D., R. H. HIER,•4D S. L. PAULUS.1991. Lesserscaup diets during migrationand winterin the Mississippiflyway. Can. J. Zool. 69:328-333. ASSEL,R. A., R. H. QUINN,G. A. LESHKEVICH,AND S.J. BOLSENGA.1983. Great Lakesice adas. Natl. Ocean. Atmos. Admin. GLERL Contrib. No. 299. AUSTIN,J. E., ANDL. H. FREDRICKSON.1987. Bodyand organmass and bodycomposition of postbreedingfemale Lesser Scaup. Auk 104:694-699. BARTONEK,J. C., ANDJ. J. HICKEY.1969. Food habitsof canvasbacks,redheads, and lesser scaupin Manitoba.Condor 71:280-290. BA'mEY,S., V. D. STOTTS,P. F. S?RINGER,ANDJ. STEENIS. 1978. Changesin submergedaquatic macrophytepopulations at the headof ChesapeakeBay, 1958-1975. Estuaries 1:171- 182. BOOKHOUT,T. A., K. E. BEDNAmI•,AND R. W. K•OLL. 1989. The Great Lakes Marshes.Pp. 131-156, in L. M. Smith, R. L. Pederson,and R. M. Kaminski,eds. Habitat management for migratingand wintering waterfowl in North America.Texas Tech Univ. Press, Lub- bock, Texas. BRIEGER,G., ANDR. D. HUNTER.1993. Uptakeand depurationof PCB77, PCB 169, and hexachlorobenzenebyzebra mussels (Dreissena polymorpha). Ecotoxicol. Environ. Safety 26:153-165. BUSCH,D., ANDB. SCHUCH2UU)T.1991. The useof the freshwatermussel Dreissena polymorpha (Pallas)for biomonitoringheavy metals in limnic ecosystems:the Weser(FRG). Verh. Int. Verein. Limnol. 24:2261-2262. C•2qv.y, S. M. 1964. Preliminarykeys to waterfowlage and sex identificationby meansof wingplumage. U.S. Fishand Wildl. Serv.Spec. Sci. Rep. 82. COTT)d, C. 1939. Food habitsof North Americandiving ducks. U.S. Dept. Agr. Tech. Bull. 643. DE KocK, W. C., ANDC. T. BOWMER.1993. Bioaccumulation,biological effects, and food chaintransfer of contaminantsin the zebramussel (Dreissena polymorpha). Pp. 503-533, in T. F. Nalepaand D. W. Schloesser.eds. Zebra mussels biology, impacts, and control. Lewis Publishers, Boca Raton, Florida. DELEEUW, J. j., ANDM. R. VANEERDEN. 1992. Sizeselection in divingTufted Ducks Aythya fuligulaexplained by differentialhandling of smalland large mussels Dreissena polymor- pha. Ardea 80:353-362. DEV•TF., A. B. 1991. Distributionand aspectsof populationdynamics of the zebramussel, Dreissenapolymorpha (Pallas, 1771), in the lakeIJsselmeer area (the Netherlands).Oe- cologia86:40-50. DIRSCHL,H.J. 1969. Foodsof lesserscaup and blue-wingedteal in the SaskatchewanRiver delta.J. Wildl. Manage.33:77-87. Dv•uI•d•s,D. 1982. Foragingand sizeselection of musselsby the tuftedduck, Aythyafuli- gu/a.J. AnimalEcol. 51:943-956. Dramre,E. A., L. G. SUCr)EN,ANt) R.J. KOVACI4.1974. Calorific,chemical and physicalvalues of potentialduck foods. Freshwat. Biol. 4:281-292. G•vmONI•E¾,J. H., ANDM. E. HE•X•tE'mR.1990. Behavior,body condition, and foodsof Buffleheadsand LesserScaups during springmigration through the KlamathBasin, California. Wilson Bull. 102:672-683. I4_xUm•:ON,D.J. 1992a.The relationshipbetween two predator groups, diving ducks and fish,and a novelprey item, the zebramussel (Dreissena polymorpha), in LakeErie at Point Pelee, Ontario. M.Sc. thesis,Univ. Western Ontario, London, Ontario. 98] C. M. Custerand T. W. Custer j. FieldOrnithol. Winter 1996

--, 1992b. A method for reconstructionof zebra mussel(Dreissena polymorpha) length from shell fragments.Can. J. Zool. 70:2486-2490. --, c. D. ANiCNE¾,AND R. C. BA•LE¾.1994. Predation of zebra musselsby diving ducks: an exclosurestudy. Ecology 75:521-531. HARAMIS,G. M., ANDV. CA•,TEP,.1983. Distribution of submersedaquatic macrophytesin the tidal PotomacRiver. Aquat. Bot. 15:65-79. HAP,MON, B. G. 1962. Mollusksas food of lesserscaup along the LouisianaCoast. Trans. N. Am. Wildl. Conf. 27:132-138. HOKE, R. A., J.P. GIESY,G. T. ANKLEY,J. L. NEWSTED,AND J. R. ADAMS.1990. Toxicity of sediments from western Lake Erie and the Maumee River at Toledo, Ohio, 1987: im- plicationsfor current dredgedmaterial disposal practices. J. Great LakesRes. 16:457- 470. HOLLAND,R. E. 1993. Changesin planktonicdiatoms and water transparencyin Hatchery Bay,Bass Island Area, westernLake Erie sincethe establishmentof the zebra mussel.J. Great Lakes Res. 19:617-624. HOPPE,R. T., L. M. SMITH, AND D. B. WESTER.1986. Foods of wintering diving ducks in South Carolina.J. Field Ornithol. 57:126-134. JONES,J. j., ANDR. D. DP,OBNEY. 1986. Winter feeding ecologyof scaupand common gold- eneye in Michigan.J. Wildl. Manage. 50:446-452. JomgE,D. G., G. M. •lS, C. M. BUNCK,AND G. W. PENDLETON.1995. Effectsof diet on rate of body massgain by winteringcanvasbacks. J. Wildl. Manage.59:31-39. KOENIG,R. L. 1969. A comparisonof the winter food habitsof three speciesof waterfowl from the upper LagunaMadre, of Texas.M.Sc. thesis,Texas A&I Univ., Kingsville,Texas. KORSCHGEN,C. E., L. S. GEORGE,AND W. L. GREEN.1988. Feeding ecologyof canvasbacks stagingon Pool 7 of the upper MississippiRiver. Pp. 237-249, in M. W. Weller, ed. Waterfowlin winter. Univ. MinnesotaPress, Minneapolis, Minnesota. KOVALAK,W. P., G. D. LONGTON,AND R. D. SMITHEE.1993. Infestafionof powerplant water systemsby the zebramussel (Dreissena polyrnovpha Pallas). Pp. 359-380, in T. E Nalepa and D. W. Schloesser,eds. Zebra musselsbiology, impacts, and control.Lewis Publishers, Boca Raton, Florida. LF•C•, J. H. 1993. Impactsof the zebra mussel(Dreissena polyrnorpha) on water qualityand fish spawningreefs in the westernLake Erie. Pp. 381-397, in T. F. Nalepa and D. W. Schloesser,eds. Zebra musselsbiology, impacts, and control. LewisPublishers, Boca Ra- ton, Florida. MERSCH,J., A. JEANJEAN,H. SPOP,,ANt)J.-C. PIHAN. 1992. The freshwatermussel Dreissena polymorphaas a bioindicatorfor trace metals,organochlorines and radionuclides.Lim- nologieAktuell 4:227-244. MICHOT,T. C., ANDA. J. NAULT. 1993. Diet differencesin redheadsfrom nearshore and offshorezones in Louisiana.J. Wildl. Manage.57:238-244. MITCHELL,C. M., ANDJ. CARLSON.1993. Lesser Scaup forage on zebra musselsat Cook Nuclear Plant, Michigan.J. Field Ornithol. 64:219-222. OLNEY,P.J.S. 1963. The food and feeding habits of tufted duck Aythyafuligula. Ibis 105: 55-62. OATH,R.J., ANDK. A. MooRE. 1981. Submergedaquatic vegetation of the ChesapeakeBay: past,present, and future. Trans.N. Am. Wildl. Nat. Resour.Conf. 46:271-283. PEP,•Y,W. C., W.J. KUENZEL,B. K. WILLIAMS,ANDJ. A. SERAFIN.1986. Influence of nutrients on feed intake and conditionof captivecanvasbacks in winter.J. Wildl. Manage.50:427- 434. , ANt)E M. UHLEP,.1982. Food habitsof diving ducksin the Carolinas.Proc. South- east.Assoc. Fish Wildl. Agencies36:492-504. , ANt) • 1988. Food habits and distribution of wintering canvasbacks,Aythya valisineria,on ChesapeakeBay. Estuaries 11:57-67. RYAN,R. A. 1972. Body weight and weight changesof wintering diving ducks.J. Wildl. Manage.36:759-765. SMIT•, V. E.,J. M. SvuP,X,J. C. FILKINS,ANt)J.J. JoNES. 1985. Organochlorinecontaminants of winteringducks foraging on Detroit River sediments.J. Great LakesRes. 11:231-246. Vol.67, No. 1 FoodHabits of Diving Ducks [99

STANCZYKOWSKA,A., •tND W. LAWACZ.1976. Caloricvalue of the Dreissenapolymorpha (Pall.) dry body weight in someMazurian lakes.Pol. Arch. Hydrobiol. 23:271-275. , P. Z•K•, A. DOMBROWSKI,H. KOT, AND E. Z•K•. 1990. The distribution of waterfowl in relation to molluscpopulations in the man-madeLake Zegrzyfiskie.Hydrobiologia 191:233-240. SUTER,W., ANDg. R. VAN EERDEN.1992. Simultaneousmass starvation of winteringdiving ducksin Switzerlandand the Netherlands:a wrong decisionin the right strategy?Ardea 80:229-242. SWANSON,G. A., ANDJ.C. BARTONEK.1970. Biasassociated with food analysisin gizzardsof blue-wingedteal. J. Wildl. Manage.34:739-746. •, G. L. K•a,u,J. C. B•v.TONEI•, j. R. SEPaE,ANr• D. H. JOHNSON.1974. Advantagesin mathematicallyweighting waterfowl food habitsdata. J. Wildl. Manage.38:302-307. THOMPSON,D. 1973. Feedingecology of diving duckson Keokuk Pool, MississippiRiver. J. Wildl. Manage.37:367-381. WORMINgTON,A., ANDJ. H. LE•CH. 1992. Concentrationsof migrant diving ducksat Point Pelee National Park, Ontario, in responseto invasionof zebra mussels,Dreissena poly- morpha.Can. Field Nat. 106:376-380. Received26Jan. 1995; accepted17 Apr. 1995.