125

for evaluating harvest impacts on this E$$protected stock. Without this inf orinati oiicontinuation of the entire fishery co�ldb

e iiijeopardy. The relatively small cost of igratory processing approximatel514 00/fish! and high precision of the +

a ey

group contribution estimates makes GSIcostwffecti ve rnanagernent tool for this fisbery

~ Selective fisheries ancl inass markingMost innovative methods used to nianaymixed-stock fisheri es are intended to protectweak stacks from over-exploitation by limit-ing harvest rate to that appropriate for theweakest stock in the co-mingled aggregation,However, it is not possible to utilize hatcheryproduction fully using existing stock ID pro-cedures such as GSI because the hatchery ornatural origin of individual fish cannot bedetermined at the time and location of cap-ture. Thus, despite weak stock managemeiitintent, natural stock production goals areoften not achieved. Without changes to iniiii-agement, it is likely there will be a co»nu<deterioration of both mixed-stock fisheriesand resource status.

The fishery was managed for inany yearsusing coded-wire tag CWT! recoveries.However, the expense of tagging adequatenuinbers of hatchery fish to provide enoughtagged fish in samples froin the fishery andthe difficulty and expense of tagging wildstocks led to the application of GSI tech-niques in the late 1980s. For the past severalyears, decisions about whether to extend orterminate the fishery affer the first week offishing have been based solely on in-seasanGSI estimates of upper river niid-Columbiaand Snake rivers! stock contribution, Figure3 illustr~ estimated stock-group contribu-tions to this fishery in three different timeperiods in one year Fig. 3a! and in threedi8'erent years Fig. 3b!. The substantialintra- and inter-annual variation of stockgroup contributions to this fishery makesannual in-season' -season monitoring a necessity.Additionally, because the GSI technique hasreasonable

Snake Rivere power to estimate the presence f0nake River spring chinook stocks Shaklee

1991! the!, technique provides inforination

~ ~innhia River spring chinookgll4eet fishery

An example of the power of genetic stockidentification is provided by the managementof the lower Columbia River winter gill-netfishery for spring-run chinook salmon. Thisfishery targets the abundant hatchery stocksfrom the Willamette River and other lowerColumbia River tributaries, but is con-strained by both manageinent intent and fed-eral court mandate from harvesting excessivenumbers of fish from bath mid-Columbia

and Snake River stocks Fig. I!. The recentU.S. governnient listing of Snake Riverspring- and summer-run chinook salmon asa threatened species under the EndangeredSpecies Act has further constrained this fish-

Similar in-season GSI fishery estimates areconducted to optimize the nianageinent ofchurn salmon fisheries in Puget Souiid Baker and Bishop 1993! and to inana,fisheries in British Columbia for odd-yearpink salinon PSC 1990!.

It would be»ghly desirabharvest of robust hatch yseparated from the inad"e~pacts on weaker naturallyh h ! st ks. Mass marking ofstocks shows same pr<this goal. Applicatio" afish produced in hatcheries m~

fisheries cou

t,~ genetics pr raini for t'acitic Salmon126

~ Skag, it River «niiner chinook sahaonThe Skagit Rive' is a major river system innorthern Puget Sound Fig. 1! that supportsspring-, summer- and fall-run chinookstocks. The WDF Skagit Hatchery haa, formany years, cultured all three races of chi-nook, The fall stock cultured at the hatcheryactually originated in the Green River incentral puget Sound! and is intended to aug-ment fishery harvests. The spring stock inthe hatchery was derived from chinookspawning in tributaries to the Skagit Riverand is, therefore, presumed to represent thenative stock. The primary giurpose of thehatchery was to increase the production ofspring-run fish in the region, The summerchinook program at the hatchery was initi-ated to augment the numbers of fish producednaturally by the healthy wild stock in theLlppef river,per river, Heziuse the fish in this stockreturn as adults at a large size and in primecondition, they are highly regarded by bothcommercial and sport fishers. This was am or factor contributing to the initiation ofthe summer chinook program at themalor actor con

hatch-

An ambitious and labor-intensive broodstockcollecting program was conducted yucted b WDF

staff from 1975 through 1979 to obtainenough spawners to establish a hatchery

instrument for achieving increased Protec-pf weak wild and hatchery stocks by

~cing their harvest in mixed-stock fisher-ies.

Managers must also consider the geneticimplications of the increased exploitation oftarget hatchery stocks. The benefits gainedthrough improved harvest access must beviewed in the context of long term geneticstability of the hatchery-based resource com-ponent. For example, it has been postulatedthat mean size of chinook and coho stocksharvested in British Columbia fisheries hasbeen reduced due to the selective pressuresapplied by the commercial fisheries Rickerl 980, Ricker and Wickett 1980!. Addition-ally, the negative effects of hooking andhandling mortality on the fish from the pro-tected, unmarked stocks that would be re-leased in such selective fisheries must also

be considered.

Mass marking proposals are being activelypursued in Washington as well as Idaho,Oregon and British Coluinbia. Many of theproposals are aimed toward avoiding thefishery constraints posed by weak stock man-agement, but the tool can be viewed just aseasily as a potential way to reduce overallexploitation of natural stocks in order to meet

escapement goals consistently and increaseproduction, thereby reducing the ri sk of ex-tinction or loss of genetic diversity.

Hatcliery Operations�Selected Case Histories

Although many of WDF's fisheries po i~~are specifically designed to use culturo"atchery stocks to augment harvest op%nities, several of the department's ne ~hatchery programs are intended to enha ceor supplement local native stocks.

latter context, the hatchery operations shouldbe designed and carried out in such a manneras to have minimal or no impact on theoriginal genetic and biological character ofthe stock being cultured. Genetic evalu-ations are being conducted to determinewhether or not the Department's hatcheryoperations are achieving this goal, Fourspecific examples of recent or ongoing stud-ies illustrate how this is being done and whatthe results have been,

127

TaMe 2. Allele frequesicies at tert inforraative loci in tltree chinook stocks propagated at QWOF Skaait River Hatchery. year collected; N = number of fish collected!. Locus arsd ameled~6 k4 Sh kk etaL >99O

1987' N ~101986' H ~102

AT-3

0.995'1 00 0.935

0.0000.0050.065

0,986'100 0.935 0. 990

0.065 0.010

ADA-1

0,9580.910

0. 049

'1000 9110 825

0.0940.'I 60 0. 089'116 0 015 0 000

sMEP-I

7000.4200,384

0.5470 380 0 611'105

0 0330 020 0 005PCK-2

0.6680,470 0 765

0 2350.530 0 332

0 965 0 8'I20.907

0.035 0,1 le0. 093

'1000 '>170.613

'730 0 340 0,46 I0.387

'-100 0.730 0.5700.6230.255 0.3880 3280.015

0 0420 049

0.9650 9950 995

0.035~ IOI

Upper Skamit RiverWikl Summer Chinook

Skagit Hatchery"Summer" Chinook

Skay't HatcheryFall Chinook

gonscrvatian Genetics Pi rains for Paci7ic Salmon 128

stock that would have the same or verysiniilar! genetic characteristics as the wildsurnrner-run stock, The actual numbers offish used to establish the hatchery stocktotaiol approximately 560 fish ranging fromg2 fish in 1975 to 142 in 1977!. Since 1980,tire hatchery has produced and reared largenumbers of this presumed summer chinookstock. Indeed, the hatchery program was sosuccessful that this hatchery stock was cho-sen as a CWT indicator stock to be used to

represent the performance of wild summercliinook from northern Puget Sound in thePacific Salmon Treaty process.

A genetic evaluation of this program wasinitiated in 1986 by comparing the charac-teristics of the summer chinook stock cul-tured in the hatchery, the wild upper SkagitRiver summer chinook stock and the fallcinnook stock cultured in the hatchery.Horizontal starch-gel electrophoresis Shak-lee and Keenan 1986; Aebersold et al. 1987!vras used to screen samples from each of thethree groups for approximately 40 variablegene ltrci. The allele frequencies at severalinformative loci in each of these three groupsof fish are shown in Table 2, Surprisingly,this analysis clearly indicated that the "sum-rner" stock being cultured in the hatcherywas significantly different p < 0. 001! fromits source the wild summer stock in theupper Skagit River!. Furthermore, thehatchery "summer" stock was similar to the'«troduced fall stock reared at the hatchery.This analysis waa repeated two years laterwith a second collection of the "summer"shck at the hatchery with basically the sameresult.

Subsequent examination of hatchery spawn-ing records suggested that there was likelysubstantial albeit unintentional! interbreed-ing bahrein the fall hatchery stock and the

wild summer fish brought into the hatchesyto be the source for establishing a summerstock program. Because this direct geneticevaluation of the Skagit Hatchery sununerchinook program showed it was not achiev-ing its goal of propagating pure summerchinook, the hatchery stock was dropped asa CWT indicator stock for northern PugetSound summer chinook in 19g7 and theentire hatchery program far summer chinookat this hatchery is being phased out.

~ Snake River fall chinook salmorr

Legislation was passed in the mid-1970s forhatchery mitigation to compensate for fallchinook losses caused by four dams on thelower Snake River in Washing', A hatch-ery site was chosen at Lyorrs Ferry, abovethe lower two dams on the Snake River Fig1!. The fall chinook run was so low at thattime, however, that a temporary hatcheryoperation, called an egg4ank program, wasbegun while the new hatchery was beingbuilt. Adult fall chinook were trapped in theSnake River, but to avoid dam pasarge rnor-talities, their progeny were reared and re-leased at a downriver location.

The program began in 1976 and from 1977onward adults were trapped at Jc» HarborDam and spawned. The resulting juvenileswere all marked by finAipping! and werereleased from WDF's Kalama Falls Hatch-ery, which is on a Colunibta River tributarybelow all mainstem dams Fig, 1!. Egg-bankfish begrm returning to Kalarna Falls in 1980,and were used as broodstock along with fishtrapped at Ice Harbu. A similar but muchsmaller operation was conducted by the U.S.Fish and Wildlife Service at two federalhatcherie in Idaho. The eggkank programended as the Lyons Ferry Hatchery becameoperational. Adults frcnrr Ice Harbor were

129

Understandably, concerns were raised aboutthe genetic impact of the egg bank prograni;specifically the effect of temporarily trans-planting fish several hundred miles dowii-river. What effects the egg-bank programhad on quantitative genetic variation in tliestock will never be known, but a cotnparisonusing allele frequencies at thirty loci wasmade between adults returning to KalaniaFalls N = 100! in 1986 and fish collected ItIce Harbor and Lyons Ferry N = 100! in1986. Significant differences p < O.OS by

sAHG-test! were found at only two loci,overall the collections were not signifi-4

cantly different p < 0.3! Seidel et al.1988!.

The intent of the Lyons Ferry HatcheryAf pi

program has always been to culture thenative Snake River fall chinook. How-

ever, genetic purity of the Lyons Ferrystock has been a central issue for severalyears, and there have been two cori-cerns. The first was the questionablewi adorn of collecting broodstock for the

siirDH-B1,2

~ -~W~c 0,$f I.a+ ~ .7jo.

04a 4.6

0.5ill RQXarear.eaa ~ ae III

PGK-2O.IO.'F0.%

< QA

1cOQI 0.84 D7 ~ *l a.a

o.a cc cl 0

j 0A ~4.l

LUPI RRRI ILCI ~

PBP-Br

I. ~

4 OTj e.I «oe

jo.~ t r ~ .szaC r;r ~ I Ere ~ -a llew tc3

siOHP-r,2 PEP< T

eoIA

I DTjoI

05>'CFICCRIIl RS ~ t I X ~ ~ llRfi8ure 4, Temporal pattern of allele frequenciest'100! in n~4mbia and Snake River fall chi-nook Solid circles ~ Ice Harbor/t.yons Ferry Hatch-ery' pure l.yons ferry stock Open circh. ~ 'mixed"collection of snita88ed fish that returned to thet.yons ferry Hatchery ai 1990 and contained anunknown nrrrnber of Uruatilla Hatchery strays,Solid triangles ~ niicLCohrnbia River HanfordReach and Ptiea4 Rapids Hatchery slock!. Collectionsiaes rangecl froru 6S to 300.

spawned at Lyons Ferry in 1984, and re-leases there began in 1985. Adults beganreturning to the Lyons Ferry Hatchery in1987. Until 1990, when operations at thehatchery were changed in response to con-cerns raised by the petitioning and sub-sequent listing of Snake River fall chinookas threatened under the ESA, the LyonsFerry broodstock consisted of adults trappedat Ice Harbor Darn and volunteers enteringthe hatchery itsel f,

hatchery program at lce Harbor D~.Although the d~ is far downsl ~ ofthe remaining natural spawning groundsin the upper Snake River, fish werecollected at this location because this siteallowed access to more fish than anymore upstream alternative due to darnpassage mortalities. Nevertheless, be-cause the dam is only about 16 kilorne-ters above the Columbia River, therewas also a possibility of trapping niid-Coluinbia "dip-ins" � Columbia Riverorigin fish that had entered the SnakeRiver, but actually would have droppedback to the Columbia River and spawnedthere. The second concern focused onstrays in the broodstock. A low fre-quency of CTW strays from other hatch-

gooseivation Genetics Pr rarns for Pacific Salrnori

z,es had always been noted at the hatchery.However, in l989, strays from a singlehatchery operation on the nearby UmatillaRiver in Oregon accounted for an estimated30/o of the Lyons Ferry broodstock. Nativefall chinook had been extirpated from theUmatilla long ago; consequently, the stockused in this hatchery was derived from fishcollected at Bonneville and Priest RapidsJams.

Unfortunately, there are no existing electro-pboretic data characterizing unequivocal lypure Snake River fall chinook. There are,however, electrophoretic data collected byNMFS and WDF dating back to 1977 fromlee Harbor Dam, Lyons Ferry Hatchery andthe tnid&olumbia. Figure 4 summarizesmuch of these data to display temporal trendsin allele frequencies at eight genetic systemsin lce Harbor/Lyons Ferry fish and the mid-Columbia fish including the UmatillaHatchery fish!, In years for which data frommultiple coll ecti ons are available, compositeallele frequencies were calculated as themean of the individual collection frequen-cies, with one exception: in 1990, two col-lections of samples were taken at I yonsFerry, one of "known" Lyons Ferry stock CWT-tagged fish! and one of untagged fi sh.Dani from the untagged fish are plotted sepa-rately in Figure 4. Note that locus and alleledesignations throughout this report followShaklee et al. �990a!. Frequencies at~H-B1,2' and at LDH-B2~; LDH-C~;and PB'A' � data not shown! are tooamilar between the two stocks to be infor-niative. Although frequencies at sAHs andG<-2' are too erratic to indicate trends, it

is dear that the untagged 1990 Lyons Ferryfuh are inore similar to mid&olumbia fishat >0' than the tagged fish are. The twosenes are near]y parallel for MPI' frequen-

cies, except that the 1990 untagged LyotmFerry collection is again more similar to thernid-Columbia collections than to the true

Lyons Ferry collections. The series ofPKPB-I ~ frequencies aI.il.ear approximateiyparallel, and in both the frequency of the'100 allele appears to be dedining sligMy.Lyons Ferry frequencies at sSOD-1~,slDHP-I, 2s and PEP-J.T exhibit de6mtetrends in the direction of the mibColurnbia

series, which for each locus remains rela-tively stable. The frequency of the F00allele at sSOD-1' in the 1990 untaggedLyons F erry collection is more similar to the1990 mid-Columbia collection, than it is tothe 1990 tagged Lyons F erry collection.

Two time periods are of interest in examin-ing these trends, before UtnatiIIa straying before 1984! and idler. Four systems pro-vide some insight into the question ofwhether the increasing similarity betssmmSnake River and mid&olumbia stocks was

coincident with the eggkenk program andcollection of broodstiock at Ice Harbor Dain

or the Umatilla Hatchery straying. Three ofthese MPI', PUB-I ~ and sSOD-/'! exhibitlittle convergence in allele frequency prior to1984. The isolocus pair slDHP-1,2' showsa slight convergence in 1980 and 198L Theexisting data are too limited to draw stronginferences, but they provide little evidencethat significant directional chiinges in allelefrequency of the Lyons Ferry stock occurredbefore the time when substantial Umatilla

Hatchery straying was erst notol.

Although a genetic impact from mid-Colunt-bia fish through straying and possible dip-incapture is evident, the genetic distinctionbetween Snake and mid-Columbia fall chi-

nook remains. A canparison of allele fre-quencies for the 1990 tagged Lyons Ferryand the 1990 mibColutnbia Priest Rapids

131

Dain! collections by chi-square heterogene-ity test is highly significant p < 0.00002!.A similar comparison of the l990 taggedLyons Ferry collection and the 1986 LyonsFerry collection is not significant p < 0. 71!,indicating the effect of heavy Umatilla stray-ing in recent years could be diminished byrestricting the broodstock to known LyonsPerry fish.

In response to the high stray rate in 1989,WDP instituted two important hatchery m an-agement changes. One was to tag 10PYo ofthe 1989 Lyons Ferry brood with CWTs! inorder to exclude all of these fish froin sub-sequent broods at the hatchery. The secondwas to modify broodstock collection opera-tions for the Lyons Ferry Hatchery for 1990and beyond. The normal broodstock collec-tion procedure at Ice Harbor Dam and LyonsFerry Hatchery was continued, but these fishwere augmented by trapping approximatelySPY' of the tagged adults that reached LowerGranite Dam, the uppermost of the fourdarns on the Snake River and the last beforethe spawning grounds Fig. 1!. In additionto providing fi sh, this allowed monitoring ofthe adult run composition at the dam for thefirst tiine, The second change in inanae-ment was a screening of spawner s for stockorigin. All CWTs were read as the fish werespawned, so that the fish could be spawnedin three groups; only Lyons Ferry taggedfish, orily forogn tagged fish, and all un-tagged fish Progeny of untagged adultswere to be used in the Lyons Ferry programonly if the stray rate was deemed to be belowan acceptiible level. In 1990, stray levelswere appreciable, so only the progeny of theLyons Ferry tagged adults were retained forprogram use, Current broodstock proce-dures for Lytsis Ferry involve use of onlytagged Lyons Ferry fish in order to exclude

fish of unknown origin from the gene pool.To sustain this program, 100 Jo of the LyonsFerry releases are now tagged, While theflow of foreign genes into the hatchery stockhas been stopped by this tagging and tagreading effort, untagged strays cannot beprevented from continuing upriver to thenatural spawning grounds. The Lyons PerryHatchery stock may well turn out to be abetter representation of the original SnakeRiver fall chinook than the natural spawnersnow protected under the ESA. In this regard,the Lyons Ferry Hatchery fish have recentlybeen determined by NMFS to be part of theSnake River fall chinook "ESU."

~ Minter Creek chum salmon

Despite a stated intent to manage churnsalmon fisheries in south Puget Sound on thebasis of local wild stocks, chum productionat the Minter Creek Hatchery Fig. I! hasutilized a stock that was originally derivedfrom the Hood Canal Hatchery. In 1986, thestock being propagated at the Minter CreekHatchery was electrophoretically charac-terized and determined to have a geneticprofile that was basically identical to that ofthe Hood Canal stock. Because GSI is theprimary method for estimating stock cotitri-butions to churn salmon fisheries in PugetSound, the high degree of genetic similaritybetween the fish produced at the MinterCreek facility and the large numbers of chumproduced at the state, tribal and federalhatcheries in Hood Canal made it impossibleto estimate contributions froin all southPuget Sound sources including MinterCreek! accurately. This, in turn, inade ef-fective harvest management of south PugetSound chum extremely difficult Further-niore, the propagation of Hood Canal type"'sh in south Puget Sound and their possible~t~aying from the facility to spawn naturally

Gaeervstion Genetics I'r rsrr» for Pacitie Salmon 't 32

in adjacent south Puget Saund tributaries andthe use of these fish for volunteer enhance-

ment projects in south Puget Sound clearlyviolated the department's goal of maintainingamong-stock genetic variability,

Far the reasons outlined above, the Depart-ment of Fisheries and tribal co-managersimplemented a program to replace the HoodCanal original chum stock at the MinterCreek Hatchery with a native south PugetSound stock Elson Creek!. Each year, be-ginning in 1988, adults returning to theInter Creek Hatchery presumably fromthe Hood Canal type stock! were removedfrom the system by a wipe-out fishery outsidethe mouth of Minter Creek and their produc-tice was replaced at the hatchery with fertil-ized eggs from the Elson Creek stock ob-tained from the Squaxin tribal facility. Theeggs from Elson werc taken from femalesthrcasghout the run and a total of approxi-mately 1300 females and 1300 males werespawned to provide fertilized eggs for theMinter Creek Hatchery. These practiceswere fdlowed in order to meet the produc-tive goals for Minter Creek and to assure thatthe recipient hatchery stock Minter Creek!would be genetically representative of thedonor stock Elson!. Thc program proceededin this manner for four years � thc averageduration of the life cycle of the Hood CanalHatchery stock. In the fish year �992!,balue eely the small fraction of fish return-ing as five-year-olds were of the Hood CanalHatrkevy type, we siinply had to identifythese fish by scale reading at the hatchery!and remove them prior to spawning toachieve camplete removal of all Hood CanalHatchery type chum from the facility, Be-ynning in 1993, all broods at the MinterCreek Hatchery should be the native southPuget Sound El son! type.

This conversion of hatchery production to anative south Puget Sound stock would nothave been succimful without tribal partici-pation to furnish the fertilized eggs neceslaryto replace production losses resulting fromelimination of the Hood Canal origin spawn-ers, Joint state/tribal alterations in harvest

management regimes were also necessiiry tosupport the progrln. WDF salmon cultureand research stafF were responsible for ac-complishing the conversion at the hatcheryand for identifying the nature and extent ofthe problem and contributing to its solutionby aging thc large fish returning in the finalyear of conversion to allow identification andremoval of five-year-olds.

The situation at a small volunteer chum

salmon enhancement program at DonkeyCreek also in south Puget Sound! run by theGig Harbor Fisherman's Civic Club, closelyparalleled the Minter Creek Hatchery pro-gram. The Donkey Creek program was in-itiated 20 years ago using fertilized eggs Hood Canal churn stock! originally obtainedfrom the Minter Creek facility. Althoughthis pragram had been sua~sful in estab-lishing a run of chum back to Donkey Credc,it was inconsistent with the department'sintent to use local stocks becisuse it was

founded using a foreign gene pool, Using asimilar approach to that employed at thcMinter Creek Hirtcheiy, the department andthe coop replaced the potential productionfrom all adults returmng to the DonkeyCreek p~~ aver the crxirse of the lastfive years with fertihzed eggs from the Elsanstock, This remedial action should have

successfully replaced the Hood Canal originpopulation in Donkey Creek with the moreappropriate Hson stock.

~ Dungeness River chinook captivebnzodstock program

The Dungeness River once supported alarge, productive chinook salmon popula-tion, but in recent years the number of adultsreturning to spawn has decreased to approxi-mately 200 fish per year C. Smith, WDF;unpublished data!. The depressed run sizeof this population threatens its long termsurvival due to the increased risk of a geneticbottleneck or ex6nction from an environ-

mental ~ophe. In response to the criti-cal status of Dungeness chinook, state, tribaland federal fisheries biologists and con-cerned citizens have joined forces in a resto-ration effort. The long term goal of thisrecovery program is to increase the numberof naturally spawning fish in the river whilemaintaining the genetic characteristics di-versity, pattern and amount of variation! ofthe existing stock However, all parties in-volved agreed that the critically low numbersof returning adults place this stock in suchjeopardy that priority be given to increasingpopulation size as quickly as possible.

After considering several approaches forachieving an immediate increase in fishnum-bers, the group decided to implement a cap-tive braodstock program. This relativelynew approach � rearing normally anadro-mous salmon in captivity throughout theirentire life cycle � has the potential to in-crease population numbers drainatically in asingle generation because of the high fecun-dity of the species approximately 3,500 eggsper female! and the low mortalities expectedfrom hatchery propaganon. Because the de-sign of this program was driven by concernto minimize inbreeding and genetic drilt maintain the amount and pattern of geneticvariation characteristic of the natural popu-l ation!, an effecti ve number of breeder s Nb!

goal of 50 per year in each of four successiveyears was established, for an Ne of approxi-mately 200 over the average four-year gen-eration time of this stock,

Two concerns made achieving this goal viathe traditional capture of pre-spawnirigadults seem unrealistic or undesirable. First,it was doubtful that the 25 pairs of adultsneeded for hatchery spawning to achieve theNb goal could be obtained because of thesmall size, low density and protracted fresh-water maturation schedule of this populationSecond, there was a strong desire to retain ahigh level of natural production in the Dun-geness River during the captive broodstockprogram. Removing 50 adults from the riverfor the captive broodstock would have rmmoved approximately one-quarter of thenatural production from the system, andwould have yielded far more eggs than nec-essary for the program.

Because of these concerns, a novel experi-mental approach was devised for establishingthe captive broodstock program based on thecollection of fry rather than adults. Thetwo-component approach involves the col-lection of pre-emergent fry by hydraulicsampling of redds salmon nests in streambed gravel! and the capture of post~ergentfry by electroshocking or seining in the river.The intent is to collect approximately 200 fryfrom each of 25 redds and up to 2500 Ayfrom throughout the river. Redd samplingwill yield representatives of a known nuinberof essentially discrete family groups, yet willremove less than 10'/o of the production ofeach family from the river. Electroshock-ingj'seining has the potential af capturingrepresentatives from all families produced inthe system, while also having only a smallitnpact on the total production. These twoapproaches together should provide adequate

Conservation Genetics Pr rams for Pacific Salmon

genetic representation of the iiatural popula-tion and yield enough fish to establish thecaptive broodstock program. The fish ineach of the above lots will be tagged in sucha wsy that specific crosses of individuals ofknown origin can be made when the fishmatiire,

The Dungeness chinook captive broodstockprograin was initiated in 1993. Fourteenfamily groups N = 2,600 fish! were success-fully sampled from redds and approximately1300 postmmergent fry were obtained byelectroshocking, All fish in the broodstockare presently being reared at the WDF HurdCreek Hatchery in the lower Dungeness wa-tershed.

Mother notable feature of this captivebroodstock program is that one-half of thefish will be reared to maturity in freshwatertanks at the hatchery! and one-half in salt-water net pens in the Strait of Juan de Fuca!.This approach will allow evaluation of therelative merits of freshwater and marine cap-tive rearing and will also minimize the riskof ~ophic failure of the program due tohaving all broodstock in a single facility.

The current plan calls for this captive brood-stock program to be conducted for an eight-year period and monitoring and evaluation tocontinue for an additional four years. Thiswill represent two generations of chinookproduction, will provide enough tagged fishto assess fishery impacts on the stock andshould allow enough time for the litnitingfactor s! responsible for the depressed statusof this stock to be identified and correctivemeasures to be initiated. Restricting theprogram to two generations should also limitinadvertent domestication selection on thestock. While the captive broodstock pro-gram is expected to dramatically increase

fish nuinbers in the- short term, the long termsuccess of Dungeness chinook restoration isentirely dependent on identifying and over-coming habitat and/or harvest snastagemcsttimpacts that have driven the stock to itscurrent critical state.

Hatchery htonitoring, Kvaltsation andResearch Pro rams

~ Tucanssoss River spring Ckaaok nsorsitor-asg and evaiaatim

The Tucannon River, in sottthetltern Wash-ington, is a tributary of the lower SnakeRiver. As part of the Lower Snake RiverFish and Wildlife Compensation Plan, thewild spring chincok stock in this river wastargeted for hatchery supplementation to pro-vide compeimsation for salsnon productionlost because of hydroelectric development inthe Snake River Basin, Because the Tucan-

non River spring chinook population repre-sented one of the last wild chinook stocks inWashington with no history of significantexposure to hatchery origin fish, WDF rec-ognized this as a unique opportunity to moni-tor and evaluate the effects of a Mw hatcheryon the genetic and biological characteristicsof the native stock. The Tucannon Hatcheryoperation was designed to incorporate thegenetic principles and concerns recognizedin the mid-1980s-, Hatchery operations com-menced in 1986, with an initial year's spawn-ing of 48 females and 43 males, and continue with similar numbers! to the present,

While the hatchery operations were begin-ning, genetic characteriztttion of the native,pre-facility spring chittook population wasinitiated. In 1985, 100 outmigrating smoltsfrom the 1983 brood were sampled from theriver. In subsequent years, sasnples of natu-rally produced fish were obtained, both from

genetic profiles of the stock before and afterartificial production was initiated.

returning adults and from pre-smolts andsmolts from the river. Each of these collec-

tions was electraphoreti cally analyzed at ap-proximately 35 variable loci to provide agenetic characterization of the native Tucan-non spring chinook stock.

sA4 TW IIDHP-1

4Nlt

sSOD-4

ogII 'I

I4 IS IC rr II la A I7 IO CI l0 ~ I

Figure S. ANele fequency trajectories at four loci Ii juvenile and adult I'ucannon River springchinook before and after initiation of hatchery supplementation activities. Vears shown alongthe x~at the broom of the fiyire refer to the year Ihefish were pnxluced juvenile coNecti~'!or the year fish returned to spawn adult collections!. Solid symbols = frequencies for nataralyprodisceil fish; open sIpnbols = freqiiencies for hatchery produced fish, Squares show tlsefrequency af the most conasion aNele $00! at each locsis, triangles show the frequency of thesecond most abundant allele at each locua N = 50 - 100 per collection.

Beginning in l990, ofFspring from the hatch-ery program began returning to the systemin significant numbers. Because all indi-viduals produced in the hatchery weremarked with CWTs, it was possible to sepa-rate the returning adults into either hatchery-produced or wilbarigin 6slt Electrophoreticanalysis of returning adults and of pre-smolt/smolt collections each year has al-lowed monitoring of the genetic charac-teristics of the stock and comparison of the

Figure 5 shows examples of the temporalpatterns in allele frequency seen at four lociin this stock. Two important patterns wereevident from this analysis. First, there wassignificant year-to-year variation in allelefrequencies at many loci. Indeed, this vari-ation was sufficient to make overall G-testsamong pairs of annual collections statisti-cally significant in many cases. The magni-tude of the inter-annual variation in allelefrequencies may be a consequence of thesmall eff'ective population size of this stockSecond, despite the significant annual vari-ability, there does not seem to be any cleardirectional shift in allele frequency, either at

Careervation Genetics Pr res for PaciRe Salinon

the four loci shown in F igure 5 or any of the3 I other variable loci being screened.

Because we have less than one completecycle's worth of data for the stock aAer thehatchery operation began, it is premature toconclude that there have been no measurable

changes at the loci monitored by electropho-resis. Although, as shown in Figure 5, thereis no evidence of substantial directional

changes. Furthermore, because electropho-resis only allows us to monitor a very smallproportion of genes in the genome, therecould well be substantial changes at otherloci that we would never detect by the elec-trophoretic screening. Nevertheless, theelectrophoretic monitoring represents an at-tempt to evaluate geneti c efFects of the hatch-cny operation. As such, this electrophoreticmonitoring and evaluation is an importantaspect of the supplementation effort becauseit has the potential to provide an early warn-ing of genetic problems in the hatchery.However, because hatchery-produced fishare now being allowed to spawn naturallyupstream and their progeny are unmarked!,our ability to distinguish between fish withand without past hatchery influence witt endwithin the next three years and subsequentmonitoring can only be done for the com-bined hatchery and naturally-spawningstock.

Hal~ conservation

As mentioned above, there are numerousquestions about conservation genetics thatneed to be answered for efFective, rationalprograms to control genetic risk. Somequestions can be answered experimentallyand some by modelling. In cixperation withWDF's Salmon Culture Division, two smallreearch programs at WDF hatcheries are

currently underway to address some of thisunceftiunty.

One study, at Tucannon, is designed toevaluate the genetic iinpact of a single gen-eration of hatchery rearing on performance,an opportunity created by the fact that we arestill in the first cycle of returns from thehatchery. The method used ia to make interse rnatings of known hatchery HxH! andwild WxW! returning adults, and evaluatetheir progeny. Family lots are reared indi-vidually until the families are combined inrearing ponds = "ponded"!, so early per-formance by family can be readily evaluated.Although family identity is lost when the fishare ponded, they are ponded by treatmentgroup HxH or WxW! and tagged by treat-ment group upon release for finther evalu-ation of group performance, Started inI 990, this study has so far revealed strikingdifFerences between hatchery and wild fe-rnales in prespawning survival of adults and.early survival of progeny, but the cause isunclear. Hatchery females tend to beyounger and smatter than wild females, andthis may account for much of the perform-ance difFerence between the two groups.The data are currently being analyzed toevaluate this efFect.

At the Methaw Hatchery Complex on themid-Columbia River! the retationatup Mtween census and efFective population size isbeing studied. Typically hatchery brood-stock guidelines as|anne that one fish equalsone efFective spawner, and that the mostseriois deptirtures from this are due to un-equal sex ratios. In reahty, probaly themeat impcrtmt determinant of efFective sizein Pacific salmon is variance of family size.At Method, full-sib families are being indi-vidlatty reared «nd marked befca.e release.Upon their return we will be able to calculate

137

the variance of family size. This study wasinitiated in 1992 with 21 families and will be

replicated in future years.

e Future monitoring and research at WOFhatcheries

A major issue surrounding the increasinglypervasive use of hatcheries is to what extentthey domesticate the fish, making them lessfit in the wild. Questions about domestica-tion selection need to be answered, but thetype of monitoring conducted in the Tucsn-

The Washington Department of Fisherieshas, like other agencies in the Pacific North-west, conducted a number of programs andpursued policies that have almost certainlyhad negativ e impacts on the genetic integrity,productivity and survival of salmon stocksunder its stewardship. Some of these eff'ectsare irreversible. However, if the departmentlearns from past mistakes, if current prac-tices and programs are evaluated and cor-rected aggressively and if enlightened poli-cies and programs are developed and imple-inented in the future, it is reasonable topresume that the continuing erosion of bio-diversity that threatens our fish and fisheryresources can be halted or even reversed.

ln conclusion, the following actions for fishmanagement agencies sre recommended:

Establish stock management poli-cies based on sound genetic andecological! principles,

non operation to date is unlikely to providethe information needed. The issue of dmneii-tication selection is currently being studied,but monitoring programs should be devel-oped to evaluate changes at quantitative lociand changes in demographic profiles. Thisis considerably more expensive, logisticallydemanding and difficult to design titan anelectrophoresiskased monitoring program,but needs to be done.

Recognize the strengths and lirnita-tions of hatchery production andensure that it is used in appropriatesituations only,

Develop a detailed inventory of thelocations, characteristics and statusof all stocks, both natural and cul-tured,

implement adequate monitoring andevaluation programs for both nativestocks and for hatchery programs,

Develop an aggressive program ofpublic education that emphasizesgenetic principles, the importanceof native storks and the critical roleof habitat,

Recognize that, in the long run,there is no substitute for adei~habitat.

Literature Cited

Acbcisold, P.B,, G.A,Winans, D.J. Teel, G.B,and F.M Utter. 1987, Maned for

starch gel clectrophoresis: a method for thedctectiixi of genetic variation. NOAA Tech-nical ReporL NMFS 61, 19pp.

Alicndorf, F.W. and N, Ryean. 1987, Geneticnuuiailenictu of hatchery stocks. Pages 141-159/n: N. Ryman md F, Utter eds.!, Popu-lation Genetics and Fishery Management.Washington Sea Grant Program, Seattle, 420PP.

Baker, B,M, and S. Bishop. 1993. Genetic stockidentification estimates of chum salmonstocks contributing to 1992 weekly test fish-crics at Apple Cove Point, Washington-Areas 9 and 10. unpublished WDF and.NWIFC repoit. Olympia, WA.

Bcacluun, T.D., R.E. Withler and A,P. Gould,1985a. Biochemical genetic stock identifica-tion of pink salmon Oncarhynchus gor-buscha! in southern British Columbia andPugct Sound. Gni. J, Fish Aquar. Sci, 42:1474-1483.

Beset~, T.D., R.E, Withler and A.P, Gould.1985b. Biochmiical genetic stock idcntifica-tioa of chum sahnoa r"ineorirynchiis @ra! insouthern British Coluinbia. Can. J Fish.A~r. Sci, 42;437~.

Berat, A. H, and R. C, Simon, cds. 1981.Pmcecdiatls of the Stock Concept Interna-

Symposium SIQCS!. Can, J Fish.~ Set 38:1457-1923.

hisack C.A. 1990. Yahntn%lickim Preduc-genctic risk asacanncnt. Jn:iM Production Project Prelinu-

y Dengn Report, Appendix A, Boiug:villePower Administration Report DOE/BP-00245-2. 20pp

Cuneus, K.P. 1993, Genetic vulnerability of thcYakima Fishery Project A risk assessmcnt diail! Washington llicparttneiit of Fisheries, 83pp,

Fournier, D.A., T.D. Beacham, B < Raddetand C A Busack, 1984, Estimating stockcomposition in mixed stock fisheries usictmorphometric, ineristic, and electrophoretiecharacteristics. C'a n. J Fisk Agent. Sci.41:400408.

Fraidenburg, M,E. and R. H. Lincoln. 19$5Wild chinook salmon management An inter-national conservation challenge. K Amer, J.Hsh. Mgmr. 5; 311-329,

Frankel, O.H. and M.E. Soule. 198l, Conter-vation and Evolution. Cambridge UnivcripPress, Cambridge.

Franklin, l.R. 1980. Evolutionary change bsmall populations. pp. 135-149 Jn; M.E.Soule and B.A, Wilcox, eds,! ConservaticaBiology: An Evolutionary-Ecological Per-spective. Sinaucr Associates, Inc., Sunder.land, MA.

Hershberger, W.K. and R.N, lwamoto. 1981,Genetics Manual and Guidelines for the h.cific Salmon Hatcheries of Washingtaa.WDF. Olyinpia, WA. unpublished, 83pp.

Hill, W,G. 1981. Estimation of effective popa-hition size from data on linkage discquili-briuin. Gener. Res. Cambridge! 31:209-216,

Holtby, L,B., R.K, Kadowaki and K.S, %spson. 1992. Factors aRecting the vulneinbiktyof juvenile coho Oncorhynchrrs kisurch! aidchinook salmon O. rshawyrscIaa! in a ~ter sport fishery. Can, J Fish. Aqxai. Sa.49: 2164-2178,

Hynes, J.D., E,H. Brown, Jr, J.H, Hello, N.Ryman and D.A. Webster, 1981. Guidelisetfor the culture of fish stocks for resourcemanagement. Can. J. Fish. Aqrarr. Sci38: 1867-1876,

~inski, A.R and L.M, Miller. 1993.&nctic hatchery guidelines forYakima/Klickitat Fisheries Project dratlj.

goiusrvation t enetics Pr rams for Pacific Salmon

Washington Department of Fisheries,75pp.+ appendices.

~, R and G.F, Barrowclough. 1987. EfIec-tive population size, genetic variation, andtheir use in population management, pp,87-123. Iri: M. Soule, ed.! Viable Populabonsfor Conservation, Cambridge UniversityPress, New York.

Mace, G.M, and R Lande, 1991, Assessingextinenon threats: Toward a reevaluation of

IUCN threatened species categories. Cortrerv.Biol 5:14$-157.

Millar, R.B. 1987. Maximum likelihood estima-tidi ofmixed stock fishery composition. Can.J Fish. 4quar. Sci. 44;583-590.

Milner, G.B., D.J. Teel, F.M. Utter and G.A.Winans. 1985. A genetic method of stockidentification in inixed populations of Pacificsalmon, Oneorhynchris spp, Mar, Fish, Rev,47�!:1-8,

Nehlsen, W,, J,E. Williams and J.A. Licha-tnwich. 1991. Pacific salmon at the cross-

roads: Stocks at risk from California,Oregon, Idaho, and Washington. Fisheries16:4-2 1,

Nelscn, K. and M. Soule. 1987. Genetical con-servation of exploited fishes, pp, 345-368. In; N. Ryman and F. Utter, eds.! PopulationGenetics and Fishery Management. Wash-ington Sea Grant Program, Seattle, WA.

PSC. 1990. Report of the Fraser River Panel tothe Pacific Salmon Commission on the 1989Fraser River sockeye and pink salmon fishingseason. Pacific Salmon Commission. Van-

couver, B.C. 51pp.

Ricker, W,E. 1972. Hereditary and environ-rnental factors affecting certain salmonidpnpulations, pp. 19-160. In: R.C. Simon andP,A. Larkin, eds.! The Stock Concept inPacific Salmon. University of British Colum-bia. Vancouver.

+cker, W, E. 1980a. Changes in age and size ofchum salmon Oncorhynchns hera!. Can.Tech. Rpt, Fish. Aquat. Sci. No. 930. PacificBiological Station, Nanaimo, B.C. 99pp.

Ricker, W.E. 1980b, Causes of the <leered inage and size in chinook sahara Once~cirur rrinrwyecha!. Can. Tech. Rpt Fish.Aqua, Sei. No. 944. Pacific Biological Sta-tion, Nanaimo, B.C. 25pp.

Rieker, W.E. nad W.P, Wiebett. 1980. Gtusesof the decrease in size in coho sabttott Ort-cerhpvrcJsus kisureJr!. Can. Tech. Rpt. FishAquat. Sci. No. 971. Pacific Biological Sta-tion, Nanaimo, B,C. 63pp.

Rohlf, D.J. 1991. Six biolagic4reasrats wby theEndangered Specica Act doesn't work-andwhat to do about it, Conserv. Biology 5:273-282

Ryman. N. 199 L Coruiervation geiieties ~-erations in fishery maaiagemest. J Frsh. BfoL39 Supplement A!: 211-224.

Seeb, J.E., J.HHelle, L.W. Seeb, R.L. Wihnot,A.J. Gharrett and W,W, Smoker. eds. ~ceedings of the Internatiomd Sympasium onthe Genetics af Subarctic Fish and Sbellfish.Capr. J Fish. Aqrrar. Sci. Supplement; inpress!,

Seidel, P. 1983. Spawnmg Guidelines for Wash-ington Deparbinent of Fisheries Hatcheries,WDF, unpublished. 15pp.

Seidel, P., R. Bugert, P. LaRiviere, D. Mar-bach, S, Martin and L. Ross. 1988. lowerSnakc River campensation plan Lyons Ferryevaluation program l 987 annual report.WDF unpublished, l 06pp,

Shaklcc, J.B. 1991. Simulation and other analy-sis of thc 1991 Columbia River spring chi-nook GSI basdinc, Wash. Dept. Fi*. Tech.Rpt. I 15, 40pp

Shaklcc, J 8 and C.P. Kccnan 1986 A pracb-cal laboratory guide to thc techniques andmethodology of ckctrophoresis and its apph-

to rshrdk;t ~nr~~ C SIRO RptAust, 177:59pp

Shaklcc, J,B., F.W. Allcndorf, D.C. Morizotand G S. Whitt l99fh. Gne nomenclature

for protein-coding loci in fi.sh. 7'rans. Arrrer.Fry Sac. 119:2-15.

Shaldee, J.B., C. Busack, A. Marshall, M.Miller aud S.R. Phelps, 1990b. 'Ihe electro-phoretic analysis of mixed-stock fisheries ofPaci6c salmon, p.235-265. in: Z-I. Ogita andC.L, Markert ods,! Isezymes: Structure,Fuactim, and Use in Biology and MedicineProgress In Clinical and Biological Research.Vol. 344. Wiley-Liss, Inc, New York.

Shields, W.M. 1982. Philopatry, Inbreeding,and the Evolution of Sex. State Univ. of New

York Press, Albany, N.Y.

Sunon, LC. 1991. Manageinent techniques tominimize the loss of genetic variability intuitciiery Ssh populations. hillier. Fish. Soc.Sysip. I G: 487494.

Utter, F, and N. Roman. 1993. Genetic markersand mixed stock fisheries. Fisheyes

18 8!: 11-21.

Utter, F�D, Teel, G. Milner and D. Mclsaac.1987, Genetic estimates of stock composi-tions of 1983 chinook salmon harvests o IF the

Washington Coast and the Columbia River,Hsh. %/l. 85: 13-23.

Wallace, B. 1991. Coadaptation revisited. JHered. 82: 89-95.

Waples, R.S. 1990. Conservation g~ ofPacific salinon. II. EfFective population ateand the rate of loss of genetic variability, gHewed. 81: 267-276.

Waples, R S. 1991. Defmition of "species" in.der the Endangered Species Act Applicalsnto Pacific sabrion. U.S. Dept. ConMnereeNOAA Tech. Meino. NMFS N/NWC-194,29pp.

Waples, R.S., G.A. Winans, F.M. UtteraiidCMahnken. 1990. Genetic approaches to theinanagement of Pacific salmon. Hsheries15�!; 19-25.

Washington Departinent of Fisheries, Washing-ton Department of Wildlife, and WesternWashington Treaty Tribes. 1993 199?Washington state sahnon and steelhead stockinventory. Washington Department of Fish.eries, Olympia, WA.