The Auk 110(1):49-56, 1993 GENETIC DIVERGENCE AMONG POPULATIONS OF THE HAWAIIAN DUCK, LAYSAN DUCK, AND MALLARD ROBERT A. BROWNE,• CURTICE R. GRIFFIN, 2 PAUL R. CHANG, 2 MARK HUBLEY, • AND AMY E. MARTIN s •Departmentof Biology,Wake Forest University, Winston-Salem, North Carolina27109, USA; and 2Departmentof Forestryand Wildlife Management, University of Massachusetts,Amherst, Massachusetts 01003, USA ABSTRACr.--Allozymicvariation at 20 gene loci was estimatedfor populationsof the Laysan Duck (Anaslaysanensis) and the Hawaiian Duck (A. wyvilliana)from the Hawaiianarchipelago, as well as for Mallard populations(A. platyrhynchos)from Hawaii and North America. The Laysan Duck and Hawaiian Duck are endemic, have experiencedsevere bottlenecks,and are listed as endangered species.Alternative alleles are fixed at six loci for Mallards versus Hawaiian anatids (Hawaiian and Laysan ducks). In contrast,every allelic variant found in the Laysan Duck was present in the Hawaiian Duck (but not vice versa), suggestingthe formeris an offshootof the latter.The geneticdistance (Nei's D) betweenLaysan and Hawaiian ducks is lessthan 0.01, while that between both Hawaiian and Laysan ducks and Mallards is greater than 0.45. The allozymic evidence also suggeststhat there has been extensive hy- bridization between Mallards and Hawaiian Ducks on Oahu, with the near disappearanceof Hawaiian Duck alleles.However, there is only slight evidenceof Mallard genic introgression into the Hawaiian Duck population on Kauai. Finally, the allozymic data suggestthat the Hawaiian Duck is a distinctspecies from the Mallard, but that little geneticdivergence has occurredbetween Hawaiian and Laysanducks. Received 25 July 1991,accepted 8 March 1992. THERESULTS of evolutionary processeson oce- and maintenance of populations of Hawaiian anic islands are evident in the Hawaiian avi- Ducks on Oahu and Hawaii relied on progeny fauna, which exhibits striking examples of from relatively few captive-reared birds. Thus, adaptive radiation. Hawaii provides a model genetic diversity of these speciesmay be sub- systemto examinethe geneticdivergence of an stantially reduced (Griffin et al. 1989). endemic, insular waterbird fauna. Several of The Hawaiian Duck, Laysan Duck and Mar- these waterbird specieshave gone through se- iana Mallard (A. oustaleti)are monochromatic, vere population bottlenecksand have remained insular endemicsthat presumablyevolved from at chronically small populations (Scott et al. straymigratory Mallard (A. platyrhynchos)stocks 1986).Of the relatively few avian families that (Weller 1980). Currently, the AOU Check-list colonized the remote Hawaiian archipelago, (1983) classifiesthe Hawaiian Duck as a distinct waterbirds were further restrictedby the scar- species,but birds in this taxon sometimes have city of coastal and inland wetlands in the is- been classified as a subspeciesof the Mallard lands (Griffin et al. 1989).Today, three endemic (Delacour 1956, Weller 1980). The Hawaiian waterbird speciessurvive: the Hawaiian Goose Duck is listed as endangered by the U.S. De- (Nesochensandivicensis), Hawaiian Duck (Anas partment of the Interior and the stateof Hawaii. wyvilliana),and LaysanDuck (A. laysanensis).All Formerly, the Hawaiian Duck occupied all are believed derived from North American spe- the main Hawaiian Islands except Lanai and cies. However, the phylogeny of the Hawaiian Kahoolawe(Perkins 1903;see Fig. I). Numbers waterbirds is not well established and is based declinednoticeably after the turn of the century primarily on characterssuch as bill size, plum- (Swedberg unpubl. manuscript). By 1960, the age color, and plumage pattern. specieswas found only on Kauai with an esti- Approximately 3,200 Hawaiian Ducks occur mated 3,000 ducks on the island in the mid- on the islands of Hawaii, Oahu and Kauai, 1960s (Swedberg unpubl. manuscript). A cap- whereasthe LaysanDuck hasan estimatedpop- tive propagation and release program for the ulation of 500 (Griffin et al. 1989) and is limited Hawaiian Duck was initiated in 1958, and birds solely to Laysan Island (Fig. 1). Both species subsequentlywere releasedon Hawaii and Oahu historically have experiencedsevere popula- (Paton 1981, U.S. Fish Wildl. Serv. 1985). Over tion bottlenecks. Furthermore, reestablishment 500 captive-reared Hawaiian Ducks have been 49 50 BROWNEET AL. [Auk, Vol. 110 • MidwayIslands • Pearl & Hermes Atoll tions to the wild population have been made . LislanskiIsland from captive-rearedstock (J. G. Giffin pers. comm.)ß • 'Laysan IslandßNecker Islandß Nihoa To date, Hawaiian and Laysanduck popu- o4• KAUA• • • OA.U lations have been characterizedsolely by clas- sical taxonomic methods. The objectivesof this ß • MAUl study are to: (1) quantify the degree of genetic variation and differentiation within and among North /•*•"• WAll endangered Hawaiian and Laysan ducks, and their presumed ancestral populations (repre- ,!/•,,Kahoolawe• sentedby samplesof Mallards from Oahu and Fig. 1. The Hawaiian Islands. California);(2) evaluatethe degreeof interspe- cific hybridization among Hawaiian Ducks and Mallards;(3) assessthe potentialsignificance of releasedon Oahu since1968 (Bostwick unpubl. introgressive hybridization by Mallards on the manuscript).Fewer than 300 have been counted species integrity of the Hawaiian Duck; (4) during semiannual statewide counts since 1980 quantify the degree of genetic variability with- (Griffin et al. 1989). There is potential for hy- in and among captive flocks of Hawaiian and bridization of Hawaiian Ducks with feral Mal- Laysan ducks. lards on the three islands where Hawaiian Ducks are believed to occur (Griffin et al. 1989). MATERIALS AND METHODS Having the mostrestricted distribution of any extant speciesof duck, the Laysan Duck is re- Samplesof tissueswere taken from January1988 through November 1989 from wild birds and captive stricted to the 400-ha Laysan Island in the birds, and nonlethal collection of blood and feather northwestern Hawaiian Islands (Fig. 1). It was pulp from captive birds. Sample sizeswere limited reportedto alsoinhabit LisianskiIsland in 1828, due to the endangeredstatus of the Hawaiian anatids. a 170-ha island over 230 km northwest of Lay- Sampleswere obtainedas follows: (1) For Mallards, san but it no longer occursthere (Berger 1981). tissuesamples (heart, kidney and liver) were obtained The number of LaysanDucks has fluctuated dra- from six wild Mallards from Oahu collected from Par- matically, from 7 birds in 1912 to 500 in the adise Park and Kahuku Meadow Gold Dairy. Tissue 1960s(Berger 1981). However, asfew as25 were sampleswere alsotaken from six wild, suspectedHa- seen in 1973 and as many as 350 were counted waiian Duck/Mallard hybrids collectedfrom the Kii Unit of James Campbell National Wildlife Refuge in 1978 (Sincock and Kridler unpubl. manu- (NWR) in Oahu. Body size and plumage character- script),suggesting that severalpopulation bot- istics(especially head colorationfor males)were the tlenecks occurred. In 1979, the Laysan Duck criteria usedto identify suspectedhybridsß Tissue was population was estimatedat 500 (Moulton and also obtained from 11 eggs collectedfrom 11 wild Weller 1984). In addition, a captive population duck nestsat the Kii Unit of JamesCampbell NWR. has been maintained since the early 1960sand The parentalpairs which laid the eggswere suspected currently numbers 22 ducks. No reintroduc- to be Hawaiian Duck/Mallard hybridsßTissue sam- TABLE1. Allele frequenciesof at 11 loci for LaysanDuck, Hawaiian Duck and Mallard populations. Est-1 Idh-1 Lap-1 Ldh-I Pep-1 Pep-2 75 100 50 100 100 NP a 75 100 80 100 88 94 LaysanDucks (Cb) 1.00 1.00 1.00 1.00 1.00 0.29 0.71 Hawaiian Ducks (Cb) 0.12 0.88 1.00 1.00 0.38 0.62 0.17 0.83 0.33 0.63 Hawaiian Ducks (W c) 1.00 1.00 1.00 1.00 0.19 0.81 0.06 0.94 NC Mallard (O) 0.50 0.50 1.00 1.00 1.00 1.00 1.00 CA Mallard (We) 0.96 0.04 1.00 1.00 1.00 1.00 0.08 0.92 Oahu Mallards (W e) 0.75 0.25 1.00 1.00 1.00 1.00 1.00 Oahu possiblehybrids (Wc) 0.75 0.25 0.67 0.33 0.33 0.67 1.00 0.75 0.25 0.25 0.75 Eggs(possibly hybrids) 0.50 0.50 0.25 0.75 • NP = not present. • C = captive. ß W = wild. January1993] DuckGenetic Divergence 51 pleswere taken from 13 wild Mallards collectedfrom For each electrophoreticallydetectable locus, the two different sitesapproximately 25 km apart in cen- mobility (distancetraveled) of the fastest-movingal- tral California (Sutter and Delevan NWR). Tissuesam- lelic productfound was used as the standardand ples were alsoobtained from one Mallard from a pri- designated100. Other alleles were designated by the vate captiveflock in North Carolina.(2) For Hawaiian migrationdistance of their proteinproducts to that Ducks, tissuesamples were taken from 12 Hawaiian of allele 100. Multiple loci encodingthe sameenzy- Ducksfrom the PohakuloaEndangered Species Prop- maticactivity were numberedsequentially beginning agation Facility on Hawaii, as well as from a recently with the form migratingclosest to the origin. The deceasedHawaiian Duck on Kauai.Feather pulp and percentageof polymorphicloci was calculated as the blood sampleswere obtained from seven wild Ha- numberof loci polymorphicdivided by the totalnum- waiian Duckscaptured at Hanalei NWR on Kauai that ber of loci examined.Heterozygosity was defined as were in captivity at Olinda EndangeredSpecies Prop- the number of heterozygousgenotypes recorded in agation Facility on Maui. (3) For LaysanDucks, tissue a sampledivided by the productof the numberof sampleswere obtainedfrom sevenLaysan Ducks from loci and the number of