athomas m. harder , gordon g. gotsch & robert c ......spawning season (mcginty and hodson 2008)....

This article was downloaded by: [Department Of Fisheries]On: 27 October 2013, At: 20:37Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

North American Journal of AquaculturePublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/unaj20

Survival and Condition of Cold-Banked JuvenileWalleyesThomas M. Harder a , Gordon G. Gotsch a & Robert C. Summerfelt ba McGraw Wildlife Foundation , Post Office Box 9, Dundee , Illinois , 60118 , USAb Department of Natural Resource Ecology and Management , Iowa State University , Ames ,Iowa , 50011-3221 , USA

To cite this article: Thomas M. Harder , Gordon G. Gotsch & Robert C. Summerfelt (2013) Survival and Condition of Cold-Banked Juvenile Walleyes, North American Journal of Aquaculture, 75:4, 512-516, DOI: 10.1080/15222055.2013.812588

To link to this article: http://dx.doi.org/10.1080/15222055.2013.812588

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of theContent. Any opinions and views expressed in this publication are the opinions and views of the authors, andare not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon andshould be independently verified with primary sources of information. Taylor and Francis shall not be liable forany losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use ofthe Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/loi/unaj20

http://www.tandfonline.com/action/showCitFormats?doi=10.1080/15222055.2013.812588

http://dx.doi.org/10.1080/15222055.2013.812588

http://www.tandfonline.com/page/terms-and-conditions


North American Journal of Aquaculture 75:512–516, 2013C© American Fisheries Society 2013ISSN: 1522-2055 print / 1548-8454 onlineDOI: 10.1080/15222055.2013.812588

COMMUNICATION

Survival and Condition of Cold-Banked Juvenile Walleyes

Thomas M. Harder* and Gordon G. GotschMcGraw Wildlife Foundation, Post Office Box 9, Dundee, Illinois 60118, USA

Robert C. SummerfeltDepartment of Natural Resource Ecology and Management, Iowa State University, Ames,Iowa 50011-3221, USA

AbstractOver their broad latitudinal distribution in nature, Walleye

Sander vitreus have adapted to survive long winters in ice-coveredlakes and for more than 5 months when water temperatures areless than required for growth (10◦C). Thus, our hypothesis isthat hatchery-reared fall-fingerling Walleyes may be maintainedin hatchery confinement at low temperature for relatively long in-tervals with minimal requirement for feed and maintenance, i.e.,cold banked (CB). Little information has been published, how-ever, on survival and condition of CB fish to judge its merits. Inthis study, survival and condition of three cohorts (N = 1,538) offeed-trained juvenile Walleyes (181–192 mm total length, 47–58 g)are described after overwintering for 125–153 d in indoor culturetanks that were supplied with lake water of ambient winter tem-perature and pelleted feed. The mean for survival of three tanks ofCB fish in each cohort was 85.8, 91.7, and 98.6% at correspondingmean water temperatures of 7.6, 8.3, and 6.8◦C, where the lowesttemperature ranged from 3.3◦C to 4◦C. The difference in meanlength, weight, and condition from start to finish was small andstatistically significant in only 1 of 3 years. Accelerated growth wasobserved in recaptured CB Walleyes 2 and 3 years after they werestocked into an 8-ha lake. The findings of this study suggest thatcold banking can be used to hold Walleyes overwinter for subse-quent spring stocking, with minimum loss of stock, or for holdingback stock in a commercial facility for grow out.

Cold banking (CB) refers to holding fingerling fish at alow temperature to temporarily arrest further growth (Daniels2005). Walleye Sander vitreus and other percids endemicto northern latitudes are inherently adapted for survival atwater temperatures ≤4◦C for as long as 8 months (Colletteet al. 1977; Hokanson 1977). The Walleye’s natural ability tosurvive such long intervals at low temperatures on a subsistencediet suggests that they may be held in the hatchery at low

*Corresponding author: [email protected] March 6, 2013; accepted June 1, 2013

temperatures for several months with minimum feeding.Holding juvenile Yellow Perch Perca flavescens at a lowtemperature for up to 180 d had no significant effect ongrowth rates after reintroduction to the culture system withtemperatures suitable for growth (Shewmon 2005). Althoughthere have been extensive evaluations on the poststockingsurvival of fry and fingerling Walleyes, we did not find researchdescribing the use of CB to carryover fall fingerlings for springstocking.

Cold banking also may be used to provide a continuoussupply of stock in a commercial aquaculture setting to com-pensate for the lack of availability of fingerlings year-round(Sullivan 2007). Other applications in aquaculture are holdingcaptive broodfish to satisfy a biological requirement for a cool-ing interval before spawning (Hokanson 1977) or extending thespawning season (McGinty and Hodson 2008). A combinationof CB and feed deprivation improved the feed conversion ofhybrid striped bass (White Bass Morone chrysops × StripedBass Morone saxatilis) by as much as 30% with no loss in totalbiomass, providing a practical method for reducing productioncosts (Picha 2007).

There is limited information on the survival and condition ofWalleyes held for long intervals at temperatures similar to whatWalleyes endure in nature, where fish may be considered in atorpid metabolic state. Such information is needed to determinethe utility of CB for enhancement stocking or for use in intensivecommercial culture systems. The objective of our study was todescribe the survival and condition of three cohorts of juvenileWalleyes after they were maintained 125–153 d in indoor tanksthat were supplied with ambient winter lake water, with temper-atures as low as 3.3–4◦C. A secondary objective was to obtaininformation on the growth of CB fish after stocking.

512

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:37

27

Oct

ober

201

3

COMMUNICATION 513

METHODSFish.—Walleye fingerlings for this study were produced from

gametes stripped from feral broodstock captured in McGawLake at the Max McGraw Wildlife Foundation in Dundee,Illinois, in the spring of 2008 (April 11) and 2009 (April 10).In 2010 (April 8) gametes were obtained from feral broodstockcaptured in Wolf Lake, Boulder Junction, Wisconsin. BecauseWalleye is considered susceptible to viral hemorrhagic sep-ticemia, water-hardened eggs from the Wisconsin source wereimmersed in povidine-iodine (100 mg/L) for 15 min prior totransport to our hatchery in Illinois. At the hatchery, incubationwas carried out in McDonald jars.

Fingerlings were produced by the pond (phase I) to tank(phase II) tandem method. Fry were stocked in an earthen pondat the McGraw Wildlife Foundation during the first week of Mayeach year for growth to fingerling size, then harvested after a44–53-d pond culture interval by seining. Initial fish size at theend of phase I pond culture for each of the 3 years was 55, 51,and 56 mm, respectively. Fish were transported from the pondto the hatchery building in a 0.5% NaCl solution, stocked inholding tanks for 1–2 d until they were enumerated and stockedinto other tanks for habituation to dry diet (phase II).

Phase II Culture.—The pond-cultured fingerlings were ha-bituated to dry diets over a 30–40-d interval each year in 0.56-m3

rectangular tanks that were covered with black plastic landscapecloth except for an opening for a horizontal solenoid feeder thatwas suspended above each tank. Each culture tank had a single,low intensity, centrally located submerged light that provideda 24-h photoperiod. Commercial starter diets (Otohime C1 andC2, Reed Mariculture, San Diego, California) were initially fedat 8% body weight per day (BW/d) at 30 min intervals 24 h/d.Fish were weaned from Otohime diets to a less expensive 1-mmWalleye grower diet (WG 9206, Skretting Silver Cup, Tooele,Utah) over 12–14 d.

After habituation, fish were offered WG 9206 diet in progres-sively increasing sizes from 1-mm, to 2-mm, and 3-mm pelletsize, and then transitioned to a sinking 3-mm then 4-mm SilverCup steelhead diet. As the incoming ambient lake water temper-ature decreased, feeding rates were reduced from 8% BW/d to<1% BW/d of steelhead diet during CB, which was invariablyin excess of their consumption.

Culture systems.—The three replicate culture tanks used forCB were rectangle in shape (3 m × 0.6 m), each having a totalvolume of 0.56 m3. Tanks were supplied with ambient lake waterwith mean inflow rate of 14 L/m, an hourly exchange rate of 1.5.Cold bank tanks were covered and had a single, submerged light(Summerfelt et al. 2011) on a 24-h photoperiod. A second set ofthree covered tanks (0.56 m3 volume per tank) with submergedlighting were part of a water reuse aquaculture system (WRAS)that had water temperatures suitable for growth (20–23◦C) andwere also on a 24-h photoperiod.

Performance characteristics.—Specific growth rate (SGR)as percent/d was calculated from a formula of Hopkins (1992)as the following: SGR = [loge (wt) – [loge (wi)/d]/100, wherewt is weight after time (t) and wi is initial weight. Condition

was expressed using relative weight (Wr), which was calcu-lated using the standard weight (Ws) formula of Anderson andNeumann (1996) as follows: log10Ws = −5.453 + (3.180 ×log10TL), where TL is total length (mm). Temperature unitgrowth is expressed as length increase in mm/◦C.

Cold banking.—Intervals for CB were December 11, 2008, toMay 14, 2009 (153 d), December 15, 2009, to May 12, 2010 (147d), and October 27, 2010, to March 3, 2011 (125 d). Because the2010 year-class fish were going to be used in another study, theirculture interval at ambient water temperatures was the shortestof the three trials. Initial stocking densities ranged from 7.0to 30.7 kg/m3 or 33–260 fish per tank or tank compartment.Stocking densities were determined each year by the number offish remaining after stocking into WRAS experimental tanks.A total of 1,538 fish were cold banked over the three differentculture intervals. At the time of tank stocking, the mean totallengths were 181–192 mm. A sample of fish was taken fromeach tank at 2-week intervals to obtain individual lengths andweights, and the total weight used to recalibrate the amountof feed to add during the next interval. Feeding was always inexcess of what was consumed. Tanks were partially dewateredand swept down daily to remove excess feed and feces.

Water Quality monitoring.—The water supply was drawninto the tank building from an adjacent 2-ha lake with a pHof 7.4–8.0, alkalinity of 220–230 mg/l, and hardness of 280–300 mg/l. We measured dissolved oxygen, dissolved oxygensaturation, and water temperature (YSI model 550, YSI, YellowSprings, Ohio) weekly throughout the culture intervals.

Observations on growth after stocking.—Yearling CB andWRAS fish were stocked on the same dates, May 15, 2009,and May 21, 2010, in Lake Bruning, an 8-ha gravel quarry lakeon the McGraw Wildlife Foundation property. The lake had anaverage depth of 2.8 m and a maximum depth of 6.7 m. Prior tostocking all fish received an injection in the lower mandible ofvisible implant elastomer dye (Northwest Marine Tech, ShawIsland, Washington), on the right side for CB fish or the leftside for WRAS fish. Examples of marking are right-red-dye for2008 CB fish and left-red-dye for 2008 WRAS fish. The WRASfish from the 2010 year-class were not available for stocking.

A 2012 spring stock assessment was carried out to collectWalleye broodstock for spawning using multiple fyke net setsand 1-h boom electrofishing with pulsed DC. All Walleyes cap-tured were weighed, measured, and examined for elastomermarks to determine their identity, i.e., the year they were stockedand whether they were CB or WRAS fish.

RESULTS

Water QualityMean water temperatures were 7.6, 8.3, and 6.8◦C for the

3 years of CB (Table 1); the lowest temperature ranged from3.3◦C to 3.9◦C, but a few days reached 10.4◦C. Mean dissolvedoxygen ranged from 11.6 to 13.2 mg/l, but saturation averaged99% (Table 1). Water temperature in the WRAS system was∼21.8◦C (Harder et al. 2012).

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:37

27

Oct

ober

201

3

514 HARDER ET AL.

TABLE 1. Temperature (Temp), dissolved oxygen (DO), and dissolved oxygen saturation conditions (% sat) during the cold-banking intervals.

2008–2009 2009–2010 2010–2011

Statistics Temp (◦C) DO (mg/l) DO (% sat) Temp (◦C) DO (mg/l) DO (% sat) Temp (◦C) DO (mg/l) DO (% sat)

Mean ± SE 7.6 ± 0.5 13.2 ± 0.8 102 ± 1.9 8.3 ± 0.5 11.6 ± 0.2 96 ± 0.6 6.8 ± 0.3 12.3 ± 0.1 100 ± 0.5High 16.2 17.3 139 17.6 14.0 107 10.9 14.2 109Low 4.0 9.0 89 3.9 8.2 83 3.3 9.6 87Median 5.3 12.4 98 5.0 12.3 96 7.6 12.0 100

Survival during Cold BankingIn the 3-year study, a total of 1,501 (92%) of the 1,538

fish stocked in tanks survived the 125-d to 153-d CB interval.Walleye fingerling survival for seven of nine tanks during thethree overwinter culture intervals averaged 98%. The lowestsurvival was 68% for one tank in 2008 and 75% for one tankin 2009. The 68% survival in one of three tanks in 2008 waspartially due to accidental loss of fish during tank dewateringfor cleaning purposes. Fish were not graded before stocking;therefore, cannibalism may have contributed to some loss.

Growth and Fish Condition during Cold BankingIn two of the three year-classes of CB fish, there was a positive

increase in length and weight during the CB interval, whichwas statistically significant in only 1 year (Table 2). In 2008–2009 there was positive change in mean individual weight, butbiomass declined because there were fewer fish at final inventory(68% survival in one of the tanks). In the 2009–2010 year-class,fish increased 8.2 mm in length and 9.6 g in weight and theirrelative weight (Wr) averaged 94 (Table 2). In the 2010–2011trial, a slight decrease occurred in both mean length and weightduring CB. Growth rates (mm/d, g/d) were trivial but positive intwo of the three trial years as was SGR (Table 2), but the data givethe impression of a positive relationship to daily temperatureunits (Table 2).

Fish Growth after StockingAfter-stocking growth of both WRAS and CB fish was com-

pared for the 2008 and 2009 year-classes (Table 3). After 3 yearsat large in the lake, the length at capture of the CB fish of the2008 year-class in April 2012 was only 3.2 mm shorter than theWRAS fish that were 78.3 mm longer than the CB fish whenthey were stocked (Table 3). The mean total length of CB fishin the 2009 year-class 2 years after stocking (392.5 mm) was2.3 mm longer than the WRAS fish that were 83.6 mm longerat the time of stocking. Differences in survival as representedby percent recapture between the CB and WRAS fish in thespring stock assessment were <1.5%, but recaptures of CB fishin both year-classes (5.3% and 2.2%) were greater than that ofthe WRAS fish (4.5% and 1.1%).

Over both at-large intervals (700–1,067 d), WRAS fish grewat a mean rate of approximately 0.16 mm/d and CB fish growthrate was 0.23–0.29 mm/d. The mean individual increase in

length for the 2 years’ stocking assessments was 173.2 and114.3 mm for WRAS fish and 248.3 and 200.2 mm for CB fish(Table 3).

DISCUSSIONOur results show that feed-trained, advanced fingerling

Walleyes can be held for up to 5 months in a single-pass systemusing ambient temperature lake water with high survival andgood body condition. Expenses for feeding (<1% body weightper day) and maintenance to clean tanks to remove excess feedand feces was minimal. Monitoring of water quality includedonly DO and temperature, thus a minimum cost for supplies andlabor.

Based on our experience, holding fish overwinter seems prac-tical because tank space is available and overwinter tank main-tenance and feeding is minimal. In a commercial aquaculturesetting, especially in the production of food fish, the key to asuccessful business is to be able to have a constant supply of fishto “grow out” to market size. Cold banking may be an effectiveoption to out-of-season spawning to provide a constant supply offingerlings for a commercial production facility seeking to raisefish to market size. Holding Yellow Perch fingerlings at 10◦C forup to 180 d had no negative effect on the subsequent growth rateor feed conversion of the fingerlings after stocking into a WRASfor grow out (Shewmon 2005). In that experiment, cold-bankedYellow Perch fingerlings actually had positive growth from 2.3to 3.2 g during the CB interval.

In Oneida Lake, New York, 97–100% of the first year growthof Walleyes was completed by October 1 and growth did notresume for 226 d until water temperature exceeded 10◦C in mid-to-late May (Forney 1966). Although growth is not expected ofWalleyes during temperatures used in our study, positive growthin both length and weight occurred in 2 of the 3 years (Table 2).

The return of stocked Walleyes to angler harvest is the ul-timate goal of fisheries managers and is dependent on survivalafter stocking, environmental conditions, fishing pressure, andharvest regulations (Laarman 1978). Kurzawski and Heidinger(1982) voiced the view that the size of a fish after its first year isimportant in determining growth in subsequent years and that arapid initial growth rate would enable a fish to prey on a widerrange of prey species, therefore itself being less vulnerable topredation. Compared with minimum growth performance in theCB fish, we calculated Walleye growth in the WRAS over a

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:37

27

Oct

ober

201

3

COMMUNICATION 515

TABLE 2. Size and growth of three cohorts of juvenile Walleyes that were cold banked for 125–153 d. Columns are culture intervals for each of the 3 years.Values of each cohort are means ± SE of three tanks of fish.

Culture intervals

VariableDec 11, 2008 toMay 14, 2009

Dec 15, 2009 toMay 12, 2010

Oct 27, 2010 toMar 2, 2011

Interval (days) 153 147 125Temperature (◦C) 7.6 ± 0.51 8.3 ± 0.48 6.8 ± 0.26Daily temperature units (DTUs) 1,162 1,220 850Total length (mm)

Initial 181.2 ± 1.5 184.1 ± 1.4 191.7 ± 0.7Final 185.2 ± 2.7 192.3 ± 0.7 188.0 ± 0.6P-value for ANOVA 0.27 <0.01 0.02

Weight (g)Initial 47.4 ± 1.4 51.5 ± 1.3 57.6 ± 0.4Final 50.0 ± 2.1 61.1 ± 1.3 53.4 ± 1.1P-value for ANOVA 0.34 <0.01 0.02

Relative weight (Wr)Initial 87 ± 0.3 88 ± 0.9 90 ± 0.9Final 85 ± 0.9 94 ± 1.2 88 ± 1.2P-value for ANOVA 0.1 0.02 0.32

Growthmm/d 0.03 ± 0.01 0.06 ± 0.01 −0.03 ± 0.01Temperature unit growth (mm/DTU) 0.0034 0.0067 −0.0044g/d 0.02 ± 0.01 0.07 ± 0.01 −0.03 ± 0.01SGR 0.04 ± 0.02 0.12 ± 0.01 −0.06 ± 0.02

Biomass density (kg/m3)Initial 30.7 ± 2.37 7.3 ± 0.15 12.1 ± 0.34Final 27.8 ± 4.01 7.9 ± 0.95 11.1 ± 0.21

P-value for ANOVA 0.57 0.52 0.05Survival (%) 85.8 91.7 98.6

TABLE 3. Comparative growth of two year-classes of Walleyes after stocking in Lake Bruning Bay. Fish were cold banked (CB) or cultured in a water reuseaquaculture system (WRAS) at temperatures suitable for growth (21.8◦C). Growth is for the interval between stocking and recapture during the 2012 spring stockassessment (SSA).

Year-class

2008 2009

Variables WRAS CB WRAS CB

Cold banking before stocking 153 147Temperature (◦C) during CB 7.6 8.3DTU in CB interval 1,163 1,220Number stocked 246 245 473 90Date stocked May 15, 2009 May 15, 2009 May 21, 2010 May 21, 2010Days fish at large in lake 1,067 1,067 700 700Total length at stocking (mean) 263.5 185.2 275.9 192.3Total length in 2012 sample (mean) 436.7 433.5 390.2 392.5Growth (mm) after stocking 173.2 248.3 114.2 200.2Growth rate (mm/d) 0.16 0.23 0.16 0.29Number recaptured in SSA 11 13 5 2Recaptures in SSA as percent of N stocked 4.5 5.3 1.1 2.2

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:37

27

Oct

ober

201

3

516 HARDER ET AL.

131-d overwinter culture interval to be 0.65 mm/d and 1.18 g/dto attain a final mean length of 282 mm and weight of 218 g(Harder et al. 2012).

The timing of stocking and the size of fish stocked may havea strong influence on survival of the stocked fish. The majorityof intensively reared Walleye fingerlings (120–140 mm) stockedin September had empty stomachs up to 4 months after stock-ing (Olson et al. 2000). During a fall electrofishing survey ofMcGaw Lake, we also observed that intensively reared Walleyefingerlings (140–160 mm) stocked in October appeared emaci-ated 2–3 weeks after stocking. If fish are emaciated and haveempty stomachs going into winter, their chance for survival toyearlings would certainly be reduced. In East Okoboji Lake,Iowa, overwinter survival of Walleye fingerlings 100–150 mmTL that were produced both intensively (fed manufactured feed)or extensively reared (fed minnows in ponds) and stocked inSeptember was less than 6%; most mortality occurred withinthe 2–5 weeks after they were stocked (Larscheid 1995). Stud-ies also show a lesser amount of food in the stomach ofWalleyes in the fall than in the summer (Wagner 1972) and oftenmore empty stomachs of Walleyes in October than in May toAugust (Kelso 1973). In Oneida Lake, New York, 78% ofWalleye stomachs were empty in the winter (Galligan 1960).Forage species that spawn in early spring, however, would pro-vide prey for yearling Walleyes that are stocked in the spring,basically serving as a starter diet in the natural setting. This isalso a time when yearling prey species of small size migrateback into the shallows and become available prey as tempera-tures warm. Thus, CB not only has commercial applications butalso application in fisheries management in that fish held over-winter in a controlled environment will have already passed thecritical “first-winter-survival” test.

Comparative results of the poststocking survey in BruningBay (Table 3) suggest that CB Walleye fingerlings exhibit ac-celerated growth once released into a natural setting. Recapturesof CB Walleyes 2 and 3 years after they were stocked into an8-ha lake were of similar size to that of Walleyes stocked andrecaptured on the same date that were ∼80 mm longer than theCB fish at the time of stocking. Though present observationson growth and survival after stocking have obvious weakness interms of sample sizes and lack of statistical analysis, they forma consistent pattern that suggests that CB fish show a growthresponse after stocking that exceeds that of fish that were muchlarger at stocking. These observations suggest that CB can beused to hold Walleyes overwinter for subsequent spring stockingwith minimum loss of stock or used for holding back stock in acommercial facility for grow out. This method may also be use-ful in holding fish for researchers attempting to conduct consec-utive trials with same-age fish. Evaluations should be conductedto assess survival, growth, and condition of Walleye fingerlingstransferred from a CB condition for on-growing in a WRAS.

ACKNOWLEDGMENTSWe thank the McGraw Wildlife Foundation board of directors

for their support in this project. Kyle Rumpel, Lucas Brown, and

Karl Klimah, summer interns from University of Wisconsin–Stevens Point, assisted with culture.

REFERENCESAnderson, R. O., and R. M. Neumann. 1996. Length, weight and associated

structural indices. Pages 447–482 in B. R. Murphy and D. W. Willis, editors.Fisheries techniques, 2nd edition. American Fisheries Society, Bethesda,Maryland.

Collette, B. B., M. A. Ali, E. F. Hokanson, M. Nagiec, S. A. Smirnov, J. E.Thorpe, A. H. Weatherley, and J. Willemsen. 1977. Biology of the percids.Journal of the Fisheries Research Board of Canada 34:1878–1889.

Daniels, H. 2005. Accelerated growth and domestication of Yellow Perch.North Carolina State University, North Carolina Sea Grant Project 03-AM-02,Raleigh.

Forney, J. L. 1966. Factors affecting first-year growth of Walleyes in OneidaLake, New York. New York Fish and Game Journal 13:146–167.

Galligan, J. P. 1960. Winter habits of pikeperch in Oneida Lake. New York fishand Game Journal 7:156–157.

Harder, T. M., G. G. Gotsch, and R. C. Summerfelt. 2012. Effect of photoperiodon growth and feed efficiency of fingerling Walleye. North American Journalof Aquaculture 74:547–552.

Hokanson, K. E. F. 1977. Temperature requirements of some percids and adap-tations to the seasonal temperature cycle. Journal of the Fisheries ResearchBoard of Canada 34:1524–1550.

Hopkins, K. D. 1992. Reporting fish growth; a review of basics. Journal of theWorld Aquaculture Society 23:173–179.

Kelso, J. R. M. 1973. Seasonal energy changes in Walleye and their diet inWest Blue Lake, Manitoba, Canada. Transactions of the American FisheriesSociety 102:363–368.

Kurzawski, K. F., and R. C. Heidinger. 1982. The cyclic stocking of parentalin a farm pond to produce a population of male Bluegill × female hybridsunfish F1 hybrids and male Redear Sunfish × female Green Sunfish F1hybrids. North American Journal of Fisheries Management 2:188–192.

Laarman, P. W. 1978. Case histories of stocking Walleyes in inland lakes,impoundments, and the Great Lakes-100 years with Walleyes. Pages 254–260 in R. L. Kendall, editor. Selected coolwater fishes of North America.American Fisheries Society, Special Publication 11, Bethesda, Maryland.

Larscheid, J. G. 1995. Development of optimal stocking regime for Walleyesin East Okoboji Lake, Iowa. Pages 472–483 in H. L. Schramm Jr. and R. G.Piper, editors. Uses and effects of cultures fishes in aquatic ecosystems.American Fisheries Society, Symposium 15, Bethesda, Maryland.

McGinty, A. S., and R. G. Hodson. 2008. Hybrid Striped Bass: hatcheryphase. Southern Regional Aquaculture Center, Publication 301, Stoneville,Mississippi.

Olson, M. H., T. E. Brooking, D. M. Green, A. J. VanDeValk, and L. G. Rudstam.2000. Survival and growth of intensively reared large Walleye fingerlings andextensively reared small fingerlings stocked concurrently in small lakes. NorthAmerican Journal of Fisheries Management 20:337–348.

Picha, M. E. 2007. Characterization of compensatory growth in hybrid StripedBass Morone chrysops × Morone saxatilis: hormones and mechanisms.Doctoral dissertation. North Carolina State University, Raleigh.

Shewmon, L. N. 2005. Culture methods for growth enhancement and off-seasonproduction of Yellow Perch Perca flavescens. Master’s thesis. North CarolinaState University, Raleigh.

Sullivan, C. V. 2007. Finfish aquaculture: improved cultivars, farming practices,and production systems. North Carolina State University, Special ResearchReport, Raleigh.

Summerfelt, R. C., J. A. Johnson, and C. P. Clouse. 2011. Culture of Walleye,Sauger, and hybrid Walleye. Pages 451–548 in B. A. Barton editor. Biology,management, and culture of Walleye and Sauger. American Fisheries Society,Bethesda, Maryland.

Wagner, W. C. 1972. Utilization of alewives by inshore piscivorous fishes inLake Michigan. Transactions of the American Fisheries Society 101:55–63.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:37

27

Oct

ober

201

3



The Safety of Aquaflor (50% Florfenicol) Administeredin Feed to Fingerling Yellow PerchJames D. Bowker a , Dan Carty a & Molly P. Bowman aa U.S. Fish and Wildlife Service, Aquatic Animal Drug Approval Partnership Program , 4050Bridger Canyon Road, Bozeman , Montana , 59715 , USA

To cite this article: James D. Bowker , Dan Carty & Molly P. Bowman (2013) The Safety of Aquaflor (50% Florfenicol)Administered in Feed to Fingerling Yellow Perch, North American Journal of Aquaculture, 75:4, 517-523






http://dx.doi.org/10.1080/15222055.2013.815676



North American Journal of Aquaculture 75:517–523, 2013American Fisheries Society 2013ISSN: 1522-2055 print / 1548-8454 onlineDOI: 10.1080/15222055.2013.815676

ARTICLE

The Safety of Aquaflor (50% Florfenicol) Administeredin Feed to Fingerling Yellow Perch

James D. Bowker,* Dan Carty, and Molly P. BowmanU.S. Fish and Wildlife Service, Aquatic Animal Drug Approval Partnership Program,4050 Bridger Canyon Road, Bozeman, Montana 59715, USA

AbstractAquaflor is an aquaculture feed premix containing 50% florfenicol and is approved for use in more than 50

countries to control mortality in a variety of cultured fishes caused by diseases associated with infectious bacterialpathogens. As part of an effort to expand the current approval in the United States, we conducted a study to evaluatethe safety of Aquaflor to Yellow Perch Perca flavescens when administered in feed at 0 × (0 mg/kg), 1 × (15 mg/kg), 3 ×(45 mg/kg), or 5 × (75 mg/kg) the proposed maximum therapeutic treatment dose of 15 mg florfenicol·kg fish−1·d−1

for 20 consecutive days, 2 × the proposed therapeutic treatment duration of 10 consecutive days. Fingerling YellowPerch (7.8 ± 1.6 cm and 5.0 ± 3.4 g; mean ± SD) were stocked into flow-through test tanks at 15 fish per tank, andtreatments were randomly assigned to tanks in triplicate. At the end of the 20-d exposure period, mean cumulativemortality in the 0 × and 3 × groups (6.7% for both) was greater than that in the 1 × and 5 × groups (2.2% and 0.0%,respectively); however, differences among the groups were not significant (P = 0.3741). Throughout the study, generalfish behavior was characterized as normal, and fish consumed virtually all feed offered. Fish health and histologyassessments revealed no signs or lesions associated with toxicity of florfenicol. In conclusion, there is an adequatemargin of safety associated with administering Aquaflor-medicated feed to fingerling Yellow Perch at the proposedtherapeutic treatment regimen of 15 mg florfenicol·kg fish−1·d−1 for 10 d.

Bacterial disease outbreaks can cause significant losses ofcaptive-reared fish (Clarke and Scott 1989; Frerichs and Roberts1989; Bjørndal 1990). Often, such outbreaks can be preventedor minimized by, for example, disinfecting and oxygenating in-coming water, implementing appropriate nutrition, rearing, andhealth management practices, and regularly disinfecting equip-ment (Piper et al. 1982; Post 1987; Jeney and Jeney 1995; Wede-meyer 2001). In addition, there are ongoing efforts to developefficacious vaccines (e.g., Bebak and Wagner 2012; Burbanket al. 2012; Shoemaker et al. 2012), but until then antimicro-bials are needed.

Several antimicrobials, including three oral antibiotics, areapproved by the U.S. Food and Drug Administration (FDA) foruse to control mortality in captive-reared fish associated with avariety of diseases (Matthews et al. 2013). However, their use isrestricted to specific disease indications and treatment regimens(FDA 2012). These restrictions limit the ability of fish culturists

*Corresponding author: jim [email protected] March 28, 2013; accepted June 7, 2013

to control bacterial disease outbreaks, and thus there is a needfor new antimicrobials or expanded uses of the antimicrobialscurrently approved in the United States.

Florfenicol {[R-(R*, S*)]-2, 2-dichloro-N-[1-(fluoro-methyl)-2-hydroxy-2-[4-(methylsulfonyl) pheny] ethyl-acetamide} is a potent, broad-spectrum, antibacterial agent withbacteriostatic and bactericidal properties that is active against avariety of Gram-positive and Gram-negative bacteria (Horsberget al. 1996). Because of its high potency and because it is notused in human medicine, florfenicol has become an importantveterinary therapeutic drug, especially when administered infeed. Florfenicol can control mortality caused by furunculosis inAtlantic Salmon Salmo salar (Nordmo et al. 1994; Samuelsenet al. 1998), pseudotuberculosis in Yellowtail (buri) Seriolaquinqueradiate (Yasunaga and Yasumoto 1988), columnaris inLargemouth Bass Micropterus salmoides, and Bluegill Lepomismacrochirus (Matthews et al. 2013), and streptococcal disease

517

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:38

27

Oct

ober

201

3

518 BOWKER ET AL.

in sunshine bass (female White Bass Morone chrysops ×male Striped Bass M. saxatilis) (Darwish 2007; Bowker et al.2010) and Nile Tilapia Oreochromis niloticus (Gaunt et al.2010). In addition, florfenicol caused no mortalities, changesin fish growth, or clinical changes in Channel Catfish Ictaluruspunctatus (Gaikowski et al. 2003) when fed for 20 d at dosesup to 34.9 mg florfenicol·kg fish−1·d−1 or in sunshine bass(Straus et al. 2012) when fed for 20 d at doses up to 75 mgflorfenicol·kg fish−1·d−1.

Aquaflor (Merck Animal Health, Summit, New Jersey) isan aquaculture feed premix containing 50% florfenicol that isapproved for use in more than 50 countries to control mortal-ity in a variety of cultured fishes caused by diseases associatedwith infectious bacterial pathogens. In the United States, theFDA has approved its use to control mortality in (1) freshwater-reared salmonids affected by furunculosis disease associatedwith Aeromonas salmonicida and coldwater disease associ-ated with Flavobacterium psychrophilum (10 mg florfenicol·kgfish−1·d−1 for 10 d), (2) catfish affected by enteric septicemiaassociated with Edwardseilla ictaluri (10–15 mg florfenicol·kgfish−1·d−1 for 10 d), (3) freshwater-reared warmwater finfish af-fected by streptococcal septicemia associated with Streptococ-cus iniae (10–15 mg florfenicol·kg fish−1·d−1 for 10 d), and (4)freshwater-reared finfish affected by columnaris disease associ-ated with F. columnare (10–15 mg florfenicol·kg fish−1·d−1 for10 d for warmwater finfish and 10 mg florfenicol·kg fish−1·d−1

for 10 d for all other finfish).The U.S. aquaculture community would like to expand the

Aquaflor label such that all freshwater finfishes can be treated atup to 15 mg·kg fish−1·d−1 for 10 d to control mortality causedby a variety of diseases. To obtain such an approval, data mustbe generated to show that this treatment regimen is safe to rep-resentative target animals. Consequently, we conducted a targetanimal safety (TAS) study to evaluate the safety of Aquafloradministered in feed to a representative coolwater finfish, Yel-low Perch Perca flavescens, at 0 × (0 mg/kg), 1 × (15 mg/kg),3 × (45 mg/kg), or 5 × (75 mg/kg) the proposed maximumtherapeutic dose of 15 mg florfenicol·kg fish−1·d−1 for 20 con-secutive days, which is 2 × the proposed therapeutic treatmentduration of 10 consecutive days. This exposure scheme allowedus to establish a margin of safety, which is herein defined as thedosage at which chronic or acute toxicity becomes evident. Awater temperature of 23◦C was selected as the test temperaturebecause it was considered the upper end of the range at whichoral antibiotic treatments (e.g., Terramycin 200 for Fish andAquaflor) were administered to coolwater finfish under autho-rization of the U.S. Fish and Wildlife Service (USFWS) Inves-tigational New Animal Drug (INAD) exemption (B. Johnson,USFWS, personal communication).

METHODSTesting facility, test fish, and test article.—The Yellow Perch

used in the study were approximately 10 months of age and

were hatched from wild-collected eggs incubated at the USFWSBozeman Fish Technology Center (BFTC; Bozeman, Montana)in April 2009. After hatching, the resultant fry were reared un-der standard hatchery conditions by BFTC staff. Sex of fish wasneither determined nor considered; however, it was assumedthat males and females were present in roughly equal propor-tions. The reference population fish were held in one fiberglasscircular tank (water volume, 756 L) with a water inflow of 30L/min (single-pass, flow-through water), which produced a wa-ter exchange rate of 2.4 exchanges/h. One week before exposurefish were moved to test tanks to begin the acclimation period;30 fish were collected from the reference population and mea-sured for TL (7.8 ± 1.6 cm, mean ± SD) and weight (5.0 ±3.4 g). During the 6-d acclimation period, fish were fed non-medicated Silver Cup No. 3 Salmon/Trout Crumbles (Nelsonand Sons, Murray, Utah) at 1% body weight (BW)/d via beltfeeders (Zeigler Brothers, Gardners, Pennsylvania).

Aquaflor premix was provided by Merck Animal Health.Control and medicated feeds were prepared at the BFTC in aMarion model SPS-1224 Mixer (Marion Mixers, Marion, Iowa).Medicated feeds were prepared by top-coating the commercialfeed with appropriate amounts of Aquaflor and fish oil (0.5%w:w) to administer doses of 0, 15, 45, and 75 mg florfenicol·kgfish−1·d−1 when fish were fed at 1% BW/d (representing 0 × ,1 × , 3 × , and 5 × the proposed dose). Control feed was top-coated with fish oil only. Immediately after test feeds were pre-pared, one sample was collected from each of the top, middle,and bottom (n = 3 samples per batch total) to verify homo-geneity of florfenicol in each batch of feed. On study days 1, 7,14, and 20, one sample of feed was collected from each batchto verify drug stability. Control feed samples were collected toverify that it was not contaminated with florfenicol. Forfenicolconcentrations were determined via HPLC by Eurofins Lan-caster Laboratories, Portage, Michigan. We tested for no otherantibiotics or contaminants in the feed.

Experimental design and procedures.—A completely ran-domized design procedure was used to (1) assign treatments totanks (n = 3) and (2) stock 15 fish into each test tank. Hence, 12test tanks and 180 test fish were used in the study. Test tanks were19-L plastic buckets (water volume, 17.4 L). Three additionalnonstudy tanks were also stocked with 15 fish per tank so thatwe could monitor growth and make weekly adjustments to theamounts of feed administered to test tanks. Feed amounts werealso adjusted daily to account for mortality. Water flow (single-pass, flow-through water) to each test tank was 3.8 L/min, whichproduced a water exchange rate of 13.1 exchanges/h.

Fish were observed daily for mortality, general behavior, andfeeding behavior. During the acclimation and exposure peri-ods, fish were fed twice daily. Feeding behavior was assessedonce daily during the acclimation period and during each of thetwo feeding events during the exposure period. The followingfive-point scale was used to score feeding behavior: 0 = ap-proximately no feed was consumed, and fish show no interest infeeding; 1 = approximately 25% of the feed was consumed, and

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:38

27

Oct

ober

201

3

SAFETY OF FLORFENICOL TO YELLOW PERCH 519

fish showed little interest in feed; 2 = approximately 50% ofthe feed was consumed, and fish showed a moderate interest infeeding; 3 = approximately 75% of the feed was consumed, andfish showed moderate interest in feeding; and 4 = approximately100% of the feed was consumed, and fish fed aggressively.

Water temperature (23.3 ± 0.4◦C) and dissolved oxygen(DO) concentration (6.0 ± 0.2 mg/L) were measured oncedaily in each tank with a YSI model 550 dissolved oxygen andtemperature meter (YSI, Yellow Springs, Ohio). Water alkalin-ity (276 ± 23 mg/L as CaCO3) and hardness (168 ± 4 mg/Las CaCO3) were measured with Hach reagents and equipment(Hach, Loveland, Colorado), and pH (7.9 ± 0.25) was mea-sured with a YSI EcoSense pH pen four times during the study(once during the acclimation period and three times during theexposure period). Overhead lights were on for 9–10 h/d.

Fish health and histology.—Before the study started, 20 ref-erence fish were collected and necropsied to characterize base-line fish health and histopathology associated with routine fishculture and handling procedures. After collection, fish were se-dated in an ice–water slurry and then euthanized by spinal sev-erance. Each necropsy consisted of visual examination of skin,gills, and internal organs and tissues for gross lesions or ab-normalities. Ten of the 20 fish were randomly selected with acompletely randomized design procedure, fixed in Davidson’sfixative solution, stored in 70% ethyl alcohol, and later pro-cessed for histology.

At the end of the in-life phase, all live fish in all test tankswere collected, measured for TL and weight, sedated in anice–water slurry, and then euthanized by spinal severance andnecropsied. If a test tank held 10 or more live fish, then 10 fishwere randomly selected with a completely randomized designprocedure and processed for histology. If a test tank held fewerthan 10 fish, then all fish were processed for histology. All buttwo fish that died during the study were too decomposed beforebeing collected to be necropsied or used for histology.

Selected tissues were dissected and then processed in FisherOmnisette tissue cassettes (Fisher Scientific, Pittsburgh, Penn-sylvania). The tissues were infiltrated with paraffin by means ofa Leica ASP 300 Advanced Smart Processor (Leica Microsys-tems, Nussloch, Germany), and the paraffin-infiltrated tissuesamples were embedded in paraffin blocks by means of a LeicaEG 1160 tissue embedding system. Tissues in selected paraffinblocks were sectioned with a Leica RM2255 rotary microtome.The 5-µm tissue sections were mounted on glass microscopeslides, stained with hematoxylin and eosin using a Leica Au-toStainer XL, and evaluated microscopically. As per FDA Cen-ter for Veterinary Medicine (CVM) guidelines, gill, liver, ante-rior kidney, posterior kidney, brain, heart, muscle, skin, spleen,pyloric intestine, and rectal intestine tissues were evaluated fromtwo of the fish randomly selected from each tank for histology.Histological evaluations of the remaining fish were only for gill,liver, anterior kidney, and posterior kidney.

Tissues were submitted for histopathologic evaluation offlorfenicol-induced toxicity. Tissues were scored under a six-

point ordinal severity scale: 0 = no change; 1 = normal (<5%of the tissue affected); 2 = mild (5–15% of the tissue affected);3 = moderate (15–25% of the tissue affected); 4 = marked(25–50% of the tissue affected); or 5 = severe (>50% of thetissue affected). Only scores of 4 or 5 were considered severeenough to have adversely affected fish health. As per CVM guid-ance, to minimize the number of histological images needingto be scored, images from the 0 × and 5 × treatment groupswere evaluated first. If significant differences were not detectedbetween these two groups, then we were not required to eval-uate differences between the 0 × and 3 × exposure groups orbetween the 0 × and 1 × exposure groups.

Statistical analysis.—Percent cumulative mortality and his-tology data (5 × exposure group versus 0 × exposure grouponly) were analyzed separately with SAS (2008) version 9.2,Proc Glimmix-based models (logit link). In both analyses, thetest tank was the experimental unit. Inadvertent mortality oc-curred in one of the 3 × treatment tanks on exposure day 4when the water supply line was inadvertently disconnected for24 h. Hence, data from this tank were excluded from analysisand there were only two replicates for this treatment group. Be-fore the histology data were analyzed, lesions scored as 0, 1, 2,or 3 were coded as “0” (not biologically important), and lesionsscored as 4 or 5 were coded as “1” (biologically important).Treatment effect on mortality and histology was tested at α =0.10 (two-sided). Mean length and weight of fish at the end of thestudy were analyzed with a one-way ANOVA (SYSTAT 2012).The treatment effect on fish size was tested at the significancelevel of α = 0.05. Feeding behavior was summarized by addingthe feeding score across replicates in each exposure group foreach feeding event (e.g., day 1, first feeding) and plotting theresults in a mosaic plot using Microsoft Office Excel software,2010 version (Bowker et al. 2013).

RESULTS

ExposuresAt the end of the 20-d exposure period, mean cumulative mor-

tality in the 0 × and 3 × groups (6.7% for both) was greater thanthat in the 1 × and 5 × groups (2.2% and 0.0%, respectively).However, differences among groups were not significant (P =0.3741). Throughout the study, general fish behavior was char-acterized as normal. Fish consumed all the feed that they weregoing to consume within 10–20 s of it being offered. Fish in the1 × , 3 × , and 5 × groups appeared to consume approximately100% of the feed offered in all but three instances (Table 1). Onstudy day 10, fish in two of the 1 × treatment tanks appearedto consume approximately 75% of the feed offered, and on thelast day of the study, fish in one of the 1 × tanks appeared toconsume approximately 75% of the feed offered. Fish in the 0 ×and nontrial groups appeared to consume less feed than fish intanks offered medicated feed.

At the end of the 20-d exposure period, no significant differ-ences were detected in mean length (P = 0.642) or mean weight

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:38

27

Oct

ober

201

3

520 BOWKER ET AL.

TABLE 1. Sum of feeding scores (based on five-point ordinal scale) across the three replicate tanks per exposure group of Yellow Perch during the first andsecond feeding periods on each study day. Areas with no shading indicate that fish in each of the replicate tanks in an exposure group appeared to consumeapproximately 100% of the feed offered. Sequentially darker shades of gray indicate less feed consumed. Note that in the 3 × exposure group, there were onlytwo replicates.

Exposure group

0 × 1 × 3 × 5 × Nontrial

Study day First Second First Second First Second First Second First Second

123456789

1011121314151617181920Summary of feeding scores:For 0 × , 1 × , 5 × , and nontrial groups

White area = 12Very light gray area = 11Dark gray area = 10

For 3 × groupWhite area = 8

(P = 0.750) in fish from among the four exposure groups. Testfish had grown an average of 0.8 cm and 2.3 g, and mean TL andweight (n = 164 fish in 12 test tanks) were 8.6 cm ( ± 1.8 cm)and 7.3 g ( ± 4.8 g). Water temperature, DO concentration, wa-ter hardness, alkalinity, and pH were within acceptable rangesfor Yellow Perch culture (Hart et al. 2006).

Mean measured florfenicol concentration in the 1 × , 3 × ,and 5 × feed samples indicated that fish were treated with 15.3± 0.5 ( + 3% from the target dose), 45.6 ± 2.6 ( + 1% fromtarget), and 77.7 ± 2.1 ( + 4% from target) mg florfenicol·kgfish−1·d−1, respectively, at the beginning of the experiment. Noflorfenicol was detected in the 0 × feed samples.

Fish Health and HistologyReference population.—External and internal tissues ap-

peared normal, although skeletal or opercular deformities werenoted in 95% of the 20 fish sampled. In the 10 fish evaluated for

histology, no lesions were observed in the skin, muscle, or py-loric intestine tissues. Lesions observed in other tissues were (1)mild to marked gill epithelial separation, mild to moderate pro-liferation of gill epithelium at the base of lamellae, and mild tomoderate telangiectasia (aneurysms in lamellar blood vessels);(2) mild to moderate liver glycogen vacuolation, indicating theamount of carbohydrate storage in fish at time of sampling; and(3) presence of mild, moderate, or marked nephrocalcinosis andcellular changes (e.g., degeneration and necrosis of tubule ep-ithelium) associated with this condition. The observed lesionsdid not appear sufficient to adversely affect fish health.

0 × exposure.—All fish (n = 42) appeared healthy at theend of the study, although skeletal or opercular deformities werenoted in 57% of the fish. Lesions observed in the reference pop-ulation fish were also observed in the 0 × fish. In addition, mildor moderate (Table 2), or marked (Table 3), melanomacrophagecenters were observed in the anterior kidney of 18 fish and spleen

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:38

27

Oct

ober

201

3


TABLE 2. Percentage of fingerling Yellow Perch treated with 0 × or 5 ×of the standard dose of 15 mg florfenicol·kg fish−1·d−1 for 20 d and evaluatedhistologically where lesions were observed and scored as mild or moderate.These lesions were not considered biologically important. Where two numbersare listed, the first number refers to the number of 0 × fish evaluated and thesecond number represents the number of 5 × treatment fish evaluated.

TreatmentNumber of

Tissue lesion samples 0 × 5 ×Spleen – melanomacrophage

centers6 50% 0%

Heart – inflammation 6 17% 0Liver – degeneration 30 7% 17%Liver – melanomacrophage

centers30 0% 3%

Liver – vacuolation 30 80% 83%Gill – epithelial lifting 30 80% 79%Gill – proliferation 30/29 53% 35%Gill – aneurysms 30/29 7% 10%Anterior kidney –

melanomacrophage centers18/14 89% 78%

Posterior kidney – proliferation 24/23 12% 4%Posterior kidney – degeneration

of tubule epithelium24/23 63% 69%

Posterior kidney – necrosis oftubule epithelium

24/23 96% 61%

Posterior kidney – inflammation 24/23 12% 13%

TABLE 3. Percentage of fingerling Yellow Perch treated with 0 × or 5 ×the standard dose of 15 mg florfenicol·kg fish−1·d−1 for 20 d and evaluatedhistologically where lesions were observed and scored as marked. These lesionswere considered biologically important. Note that no lesions were observed thatwere scored as “severe.” Where two sample numbers are listed, the first numberrefers to the number of 0 × fish evaluated and the second number representsthe number of 5 × treatment fish evaluated.

TreatmentNumber of

Tissue lesion samples 0 × 5 ×Spleen – melanomacrophage

centers6 17% 0%

Liver – degeneration 30 7.5% 5%Liver – glycogen vacuolation 30 3% 0%Gill – epithelial lifting 30/29 20% 14%Anterior kidney –

melanomacrophage centers18/14 11% 21%

Anterior kidney – inflammation 18/14 6% 0%Posterior kidney – degeneration

of tubule epithelium24/23 0% 4%

Posterior kidney – necrosisof tubules

24/23 0% 4%

of four fish, and mild inflammation of posterior kidney was ob-served in one fish (Tables 2, 3). No lesions considered severeenough to affect fish health were observed in other tissues.

1 × exposure.—All fish (n = 44) appeared healthy at theend of the study, although skeletal or opercular deformities werenoted in 69% of the fish. No tissues from fish in this exposuregroup were examined histologically.

3 × exposure.—All fish (n = 33) appeared healthy at theend of the study, although skeletal or opercular deformities werenoted in 64% of the fish. No tissues from fish in this exposuregroup were examined histologically.

5 × exposure.—All fish (n = 45) appeared healthy at theend of the study, although skeletal or opercular deformities werenoted in 71% of the fish. Histological examinations showed thatchanges observed in the 5 × exposure group fish were similarto those described for 0 × exposure group fish (Tables 2, 3).

Observed lesions that were marked in the 0 × and 5 × expo-sure groups included (1) liver degeneration, (2) anterior kidneymelanomacrophage centers, and (3) gill epithelial lifting (Ta-ble 3). Marked anterior kidney and spleen melanomacrophagecenters, liver glycogen vacuolation, and anterior kidney inflam-mation were observed in fish from the 0 × group. Each of theselesions was observed in a different fish. Marked posterior kidneydegeneration and necrosis of tubules was observed in one fishfrom the 5 × exposure group. No severe lesions were detected.Differences between prevalence of marked lesions in the 0 ×and 5 × exposure groups were not significant (P-values > 0.1).

DISCUSSIONOur results indicated that the margin of safety for florfenicol

administered in feed to fingerling Yellow Perch extends to atleast 75 mg florfenicol·kg fish−1·d−1 for 20 d. This statementwas based on the facts that there was no mortality among fish inthis group and no dose–response trend in mortality was evident.Regardless of treatment there was no difference between groupsin medicated feed consumption, behavior, or fish size. No fishhealth, lesions, or histological changes that indicated the highestflorfenicol dosage was not safe to Yellow Perch were detected.The relative degree of skeletal or opercular deformities washigher than anticipated, and we speculate that it might havebeen attributed to feeding fish a commercial salmon–trout diet.

Our results are consistent with those found in similarly con-ducted studies to evaluate the safety of Aquaflor administeredin feed to other freshwater-reared finfishes. Straus et al. (2012)observed no mortality, gross lesions, or microscopic lesionswhen sunshine bass were fed florfenicol doses ranging from 0to 75 mg florfenicol·kg fish−1·d−1 for 20 d. In addition, fishconsumed 100% of feed offered, often breaking the surface ofthe water while feeding. Similar results were reported in whichfingerling Rainbow Trout Oncorhynchus mykiss were fed flor-fenicol doses ranging from 10 to 50 mg florfenicol·kg fish−1·d−1

for 20 d (FDA 2007). Rainbow Trout in each group consumed

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:38

27

Oct

ober

201

3

522 BOWKER ET AL.

>99% of the feed that was offered, and there was no mortality orclinically observable changes detected in fish behavior amongthe treated fish relative to the controls. In addition, no grossabnormalities of the internal organs were observed on necropsyand no morphological differences were detected during the his-tological examination. Similarly, Inglis et al. (1991) reportedthat no histological changes were observed in the kidneys ofAtlantic Salmon parr when exposed to 100 mg florfenicol·kgfish−1·d−1 for 10 d. Gaunt et al. (2003) reported no mortality,microscopic lesions, histological changes, palatability issues,or adverse behavior among 5-month-old Channel Catfish as-sociated with florfenicol doses ranging from 10 to 100 mg·kgfish−1·d−1 when administered for 10 d. In another study withChannel Catfish fed florfenicol at doses ranging from 10 to50 mg florfenicol·kg fish−1·d−1 for 20 d, Gaikowski et al. (2003)reported no mortality, but did observe signs of inappetance andhistological changes attributed to prolonged exposure to flor-fenicol. Those authors reported an increase in the amount ofuneaten feed among fish exposed to 30 or 50 mg florfenicol· kgfish−1·d−1 and a “minimal to mild decrease” in hematopoietic–lymphopoietic (H&L) tissue in the anterior kidney, posteriorkidney, and spleen. However, because of insufficient data, theywere unable to determine whether the decrease in H&L tissuewas an adverse effect. In a study with tilapia Oreochromis sp.fed florfenicol-medicated feed at the same dosages as adminis-tered in the our study, Gaikowski et al. (2013), reported a total ofthree mortalities that were considered incidental and evidence ofinappetance among the 45- and 75-mg florfenicol·kg fish−1·d−1

groups during exposure days 11–19. Histopathological findingsamong the tilapia in the Gaikowski et al. (2013) study weremore extensive than previously reported for other fish speciesand included lesions observed in gills, liver, anterior kidney,and posterior kidney. The authors concluded that these changeswere likely to be of minimal clinical importance given the lackof mortality, but that feeding florfenicol-medicated feeds at 45and 75 mg florfenicol·kg fish−1·d−1 for an extended period (>10d) will cause significantly decreased feed consumption and fishgrowth.

Although deleterious effects of Aquaflor treatment are pos-sible, e.g., in untested species or sensitive life stages, we specu-late that such events would be rare. As noted previously, severaltarget animal safety studies have shown little to no effect ofexposing fish to florfenicol concentrations well beyond the in-tended therapeutic dose of 15 mg florfenicol·kg fish−1·d−1 for10 d. Further, over 100 million fish have been treated withAquaflor since 2001 without adverse effects under the auspicesof the USFWS National INAD Program (B. Johnson, USFWS,personal communication) and many more have been treated inother countries with existing approvals for this product. Accord-ingly, we conclude that Aquaflor-medicated feed administeredat 15 mg florfenicol·kg BW−1·d−1 for 10 d is safe for use on Yel-low Perch and is likely to be safe for use on all freshwater-rearedfinfishes.

ACKNOWLEDGMENTSMerck Animal Health supplied the Aquaflor and paid for

analysis of feed samples. Diane Sweeney (Merck AnimalHealth) helped analyze mortality and histology data. Mi-randa Dotson (Aquatic Animal Drug Approval PartnershipProgram [AADAP]) helped collect data, and Niccole Wande-lear (AADAP) prepared medicated feeds and collected feedsamples for dose verification. Beth MacConnell (HeadwatersFish Pathology LLC) evaluated histology samples. Kurt Borgewas the independent quality assurance officer. Jesse Trushen-ski (Southern Illinois University) and Richard Endris (MerckAnimal Health) reviewed manuscript drafts. Mention of tradenames or commercial products in this article is solely for thepurpose of providing specific information and does not implyrecommendation or endorsement by the USFWS.

REFERENCESBebak, J., and B. Wagner. 2012. Use of vaccination against enteric septicemia

of catfish and columnaris disease by the U.S. catfish industry. Journal ofAquatic Animal Health 24:30–36.

Bjørndal, T. 1990. The economics of salmon aquaculture. Blackwell ScientificPublications, Oxford, UK.

Bowker, J. D., D. Carty, and M. P. Bowman. 2013. The safety of SLICE (0.2%emamectin benzoate) administered in feed to fingerling Rainbow Trout. NorthAmerican Journal of Aquaculture 75:455–462.

Bowker, J. D., V. E. Ostland, D. Carty, and M. P. Bowman. 2010. Effectiveness ofAquaflor (50% florfenicol) to control mortality associated with Streptococcusiniae in freshwater-reared subadult sunshine bass. Journal of Aquatic AnimalHealth 22:254–265.

Burbank, D. R., S. E. LaPatra, G. Fornshell, and K. D. Cain. 2012. Isolationof bacterial probiotic candidates from the gastrointestinal tract of RainbowTrout, Oncorhynchus mykiss (Walbaum), and screening for inhibitory activityagainst Flavobacterium psychrophilum. Journal of Fish Diseases 35:809–816.

Clarke, R., and D. Scott. 1989. An overview of world salmon production andrecent technology developments. Bulletin of the Aquaculture Association ofCanada 89(4):31–48.

Darwish, A. M. 2007. Laboratory efficacy of florfenicol against Streptococcusiniae infection in sunshine bass. Journal of Aquatic Animal Health 19:1–7.

FDA (Food and Drug Administration). 2007. Aquaflor, florfenicol type Amedicated article freshwater-reared salmonids, for the control of mortality infreshwater-reared salmonids due to furunculosis associated with Aeromonassalmonicida. FDA, Freedom of Information Summary, Supplemental NewAnimal Drug Application, NADA 141-246, Center for Veterinary Medicine,Rockville, Maryland. Available: www.fda.gov/downloads/AnimalVeterinary/Products/ApprovedAnimalDrugProducts/FOIADrugSummaries/UCM051489.pdf. (March 2013).

FDA (Food and Drug Administration). 2012. Aquaflor, florfenicol type A med-icated article freshwater-reared warmwater finfish, freshwater-reared finfish.FDA, Freedom of Information Summary, Supplemental New Animal DrugApplication, NADA 141-246, Center for Veterinary Medicine, Rockville,Maryland. Available: www.fda.gov/downloads/AnimalVeterinary/Products/ApprovedAnimalDrugProducts/FOIADrugSummaries/UCM299146.pdf.(March 2013).

Frerichs, G. N., and R. J. Roberts. 1989. The bacteriology of teleosts. Pages 289–319 in R. J. Roberts, editor. Fish pathology, 2nd edition. Bailliere Tindall,London.

Gaikowski, M. P., J. C. Wolf, R. G. Endris, and W. H. Gingerich. 2003. Safetyof Aquaflor (florfenicol, 50% type A medicated article), administered in feedto Channel Catfish, Ictalurus punctatus. Toxicologic Pathology 31:689–697.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:38

27

Oct

ober

201

3


Gaikowski, M. P., J. C. Wolf, S. M. Schleis, D. Tuomari, and R. G. Endris.2013. Safety of florfenicol administered in feed to tilapia (Oreochromis sp.).Toxicologic Pathology 41:639–652.

Gaunt, P. S., R. Endris, L. Khoo, A. T. Leard, S. Jack, T. Santucci, T. Katz,S. V. Radecki, and R. Simmons. 2003. Preliminary assessment of the tol-erance and efficacy of florfenicol against Edwardsiella ictaluri adminis-tered in feed to Channel Catfish. Journal of Aquatic Animal Health 15:239–247.

Gaunt, P. S., R. Endris, A. McGinnis, W. Baumgartner, A. Camus, J. Steadman,D. Sweeney, and F. Sun. 2010. Determination of florfenicol dose rate in feedfor control of mortality in Nile Tilapia infected with Streptococcus iniae.Journal of Aquatic Animal Health 22:158–166.

Hart, S. D., D. L. Garling, and J. A. Malison, editors. 2006. Yellow Perch(Perca flavescens) culture guide. North Central Regional Aquaculture Center,Culture Series 103, Iowa State University, Ames.

Horsberg, T. E., K. A. Hoff, and R. Nordmo. 1996. Pharmacokinetics of flor-fenicol and its metabolite florfenicol amine in Atlantic Salmon. Journal ofAquatic Animal Health 8:292–301.

Inglis, V., R. H. Richards, K. J. Varma, I. H. Sutherland, and E. S. Brokken.1991. Florfenicol in Atlantic Salmon, Salmo salar L., parr: tolerance andassessment of efficacy against furunculosis. Journal of Fish Diseases 14:343–351.

Jeney, Z., and G. Jeney. 1995. Recent achievements in studies on diseases ofCommon Carp (Cyprinus carpio L.). Aquaculture 129:397–420.

Matthews, M. D., J. D. Bowker, D. G. Carty, N. Wandelear, M. P. Bow-man, J. C. Sakmar, and K. Childress. 2013. Efficacy of Aquaflor (50%florfenicol)–medicated feed to control mortality associated with Flavobac-

terium columnare infection in Florida Largemouth Bass and Bluegill. NorthAmerican Journal of Aquaculture 75:385–392.

Nordmo, R., K. J. Varma, I. H. Sutherland, and E. S. Brokken. 1994. Florfeni-col in Atlantic Salmon, Salmo salar L.: field evaluation of efficacy againstfurunculosis in Norway. Journal of Fish Diseases 17:239–244.

Piper, R. G., I. B. McElwain, L. E. Orme, J. P. McCraren, L. G. Fowler, and J. R.Leonard. 1982. Fish hatchery management. U.S. Fish and Wildlife Service,Washington, D.C.

Post, G. W. 1987. Textbook of fish health, revised and expanded edition. TFHPublications, Neptune City, New Jersey.

Samuelsen, O. B., B. Hjeltnes, and J. Glette. 1998. Efficacy of orally adminis-tered florfenicol in the treatment of furunculosis in Atlantic Salmon. Journalof Aquatic Animal Health 10:56–61.

SAS (Statistical Analysis Systems). 2008. SAS, version 9.2. SAS Institute, Cary,North Carolina.

Shoemaker, C. A., B. R. LaFrentz, and P. H. Klesius. 2012. Bivalent vacci-nation of sex reversed hybrid tilapia against Streptococcus iniae and Vibriovulnificus. Aquaculture 354–355:45–49.

Straus, D. L., J. D. Bowker, M. P. Bowman, D. Carty, A. J. Mitchell, and B.D. Farmer. 2012. Safety of Aquaflor-medicated feed to sunshine bass. NorthAmerican Journal of Aquaculture 74:1–7.

SYSTAT. 2012. SigmaPlot 12.3 graphical and statistical software. Systat Soft-ware, San Jose, California.

Wedemeyer, G. A., editor. 2001. Fish hatchery management, 2nd edition. Amer-ican Fisheries Society, Bethesda, Maryland.

Yasunaga, N., and S. Yasumoto. 1988. Therapeutic effect of florfenicol on exper-imentally induced pseudotuberculosis in Yellowtail. Fish Pathology 23:1–5.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:38

27

Oct

ober

201

3



Out-of-Season Spawning Method for FloridaLargemouth Bass to Produce Advanced-SizedFingerlings by Early SpringMichael D. Matthews a & Richard B. Stout aa Florida Fish and Wildlife Conservation Commission , Florida Bass Conservation Center ,3583 County Road 788, Webster , Florida , 33597 , USA

To cite this article: Michael D. Matthews & Richard B. Stout (2013) Out-of-Season Spawning Method for Florida LargemouthBass to Produce Advanced-Sized Fingerlings by Early Spring, North American Journal of Aquaculture, 75:4, 524-531






http://dx.doi.org/10.1080/15222055.2013.824943




COMMUNICATION

Out-of-Season Spawning Method for Florida LargemouthBass to Produce Advanced-Sized Fingerlings by Early Spring

Michael D. Matthews* and Richard B. StoutFlorida Fish and Wildlife Conservation Commission, Florida Bass Conservation Center,3583 County Road 788, Webster, Florida 33597, USA

AbstractThe objective of this research was to establish a production-scale

“out-of-season” (OS) spawning protocol yielding Florida Large-mouth Bass Micropterus salmoides floridanus swim-up fry in a 3-to 4-week period in the fall. The OS photothermal manipulationbegan each year on approximately June 18 and was completed bySeptember 20. The intent was to simulate winter to spring temper-atures and day length over a 90-d period to naturally induce gonaddevelopment without the use of hormones. Adult bass were stockedat a 1:1 or 2:3 male-to-female sex ratio and spawned in late Septem-ber through mid-October. Total numbers of spawns collected for2009, 2010, and 2011 were 193, 205, and 199, representing an av-erage 1.4, 2.0, and 1.0 spawns per female, respectively. Spawningduration required to achieve production numbers were 26, 31, and23 d, respectively. Off-season spawning was accomplished withoutphotothermal manipulation in December 2008–January 2009, butcold pond temperatures resulted in 10% survival. The OS spawningtechnique allows for biannual production seasons from the samebroodstock population and the production of large numbers of100-mm bass by March. Increased production, spawning latency,broodstock behavior, hatchery efficiency, and temporal limitationsare discussed.

Spawning techniques for and behaviors of Largemouth BassMicropterus salmoides and Florida Largemouth Bass M. s. flori-danus have been investigated for more than 100 years (Lamkin1900; Snow 1968; Isaac et al. 1998). Bass typically spawn inearly spring into the summer, displaying some latitude variationbetween northern and southern climates (Kramer and Smith1962; Brauhn et al. 1972; Heidinger 1975; Rogers 2006), butspawning beds and fry have been observed in the fall in Alabama(Swingle 1956) and Florida (R. Stout, Florida Fish and WildlifeConservation Commission [FWC], personal communication).Historically, Florida Largemouth Bass fry are commonly pro-duced outdoors by pairing male and female broodstock intoclean earthen ponds, free of forage. The fry are collected and

*Corresponding author: [email protected] March 26, 2013; accepted July 8, 2013

restocked into fingerling rearing ponds (Simco et al. 1986). Cul-turists began development of new spawning and rearing methodsbecause of the long-standing popularity of bass as a freshwatergame fish and the demand for greater numbers and larger-sizedindividuals for stocking.

Spawning largemouth bass beyond the natural spawningseason has been investigated by several researchers. Brauhnet al. (1972) delayed spawning until August using pellet-trainedbroodstock with temperature manipulation in indoor concreteraceways at 17◦C in Missouri. The technique was an effortto delay spawning and avoid unpredictable spring temperaturechanges and maximize hatchery production. Carlson (1973) suc-cessfully laboratory spawned Largemouth Bass held indoors inMinnesota from December through February and May throughJuly by manipulating photoperiod and temperature. Naturalspawning was delayed by holding broodfish below 14◦C for12 continuous months initially at naturally occurring photope-riods, then changing to a 12-h photoperiod. Spawning involvedone to three pairs of broodfish in an effort to produce smallnumbers of embryos for research. Jackson (1979) successfullydelayed pond spawning in New York by holding LargemouthBass broodstock in indoor raceways at an average water tem-perature of 16◦C. Nests and fry were observed in ponds stockedwith 6–10 pairs of these temperature-manipulated broodfish inJune through early September at an average water tempera-ture of 23◦C. Initial efforts at the Florida Bass ConservationCenter (FBCC) in Webster, Florida, to produce early spawnswere attempted by spawning in December and stocking fry infertilized nursery ponds in January. These efforts were unsuc-cessful due to >85% fry mortality caused by unpredictable coldoutside water temperatures. These “off-season” spawning tech-niques successfully manipulated individual bass to spawn inmonths outside the natural season observed in their specific lo-cation, but required substantial preparation time (>10 months)

524

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

COMMUNICATION 525

or simply extended the natural spawning period. Acclimationtime and hormone treatment of both sexes with human chori-onic gonadotropin successfully controlled Florida LargemouthBass spawning during the natural spawning season in San Mar-cos, Texas, but was not tested for producing spawns off-season(Mayes et al. 1993).

Development of advanced fingerling bass (≥100 mm) culturetechniques in indoor, concrete raceways using formulated feedshas allowed researchers to investigate stocking advanced-sizedfingerlings. While Largemouth Bass are traditionally reared inponds (Snow and Maxwell 1970), protocols involving grow outin raceways enable culturist to intensively produce larger quan-tities of similar-sized bass (Nelson et al. 1974; Isaac et al. 1998).Hearn (1977) showed that bass reared intensively in raceways onformulated diets survived at greater percentage compared withsimilar-sized bass pond-reared on Fathead Minnows Pimephalespromelas when transferred to ponds, in the absence of largepredators. Stocking larger-sized bass should increase survivalbased on reduced predation vulnerability and increased preyavailability (Loska 1982; Wahl et al. 1995). Starting in 2007,all bass spawning was conducted indoors in 45,400 L, 24-mconcrete raceways at the FBCC. Spawning bass indoors allowsmore control in timing of spawns, is more efficient, and producesgreater numbers of eggs from few broodfish (Mayes et al. 1993;Isaac et al. 1998). Spawning dates coincided with natural springspawning periods in central Florida (February to April). Con-trolled spawning’s primary objective was the production of 100-mm fingerlings by early June for stocking in depressed systemsin order to bypass forage constraints. Poststocking evaluationsfound that advanced-sized, pellet-reared hatchery fish stockedon top of age-0 natural populations of similar-size fish experi-enced high initial stocking mortality (Porak et al. 2011). Pouderet al. (2010) reported a significantly higher frequency of emptybass stomachs 7 d poststocking, supporting Porak’s evaluationof an early inability to transition to natural prey as consequentlycontributing to reduced survival. Additional advanced finger-ling evaluations showed minimal gain when stocking 55-, 100-,150-, and 200 mm fish against similar-sized wild bass (Dianaand Wahl 2009). The inability to stock 100-mm bass earlierin the spring when naturally occurring age-0 bass are unableto compete for forage required spawning bass earlier in theyear to enable the FBCC to produce 100-mm fish by March.December–January spawning was initially investigated basedon confirmed reports of spawning beds recorded in hatcheryponds during mild winters. Unfortunately, pond temperaturesin central Florida will not permit consistent fry growth duringthis period, so spawning dates between late September throughearly October were believed to be more realistic.

We developed a biannual, full-scale production spawn-ing protocol that included an “out-of- season” (OS) spawn(September–October) and a typical spring spawn (February–March) utilizing the same broodfish due to an increased de-mand for fingerlings earlier in the year and of larger size thanpond techniques can produce. The expected end result would

yield fingerlings (35–40-mm TL) in the spring and advancedfall-spawned fingerlings (85–110-mm TL) stocked the follow-ing spring. Photoperiod and temperature were manipulated in aneffort to simulate daylight and temperatures experienced duringa typical central Florida winter in an effort to induce an indoorfall spawn without hormone injection. Comparisons betweenspring and OS spawning end results (fry per spawn, spawns perfemale, and total number of spawns) were recorded but couldnot be statistically compared in all cases, and are not the mainfocus of this work.

METHODSAdult Florida Largemouth Bass stocks (wild-caught) were

maintained indoors all year in concrete 45,400-L (24 × 2.5 ×1-m) raceways. Broodstock must be wild caught, F1 pellet orforage-reared broodstock being strictly prohibited as detailedin an internal agency genetics policy (Tringali et al. 2007) ad-dressing the loss of genetic variability. Both males and femalesremained separated in the raceways for reduced acclimation timeduring spawning. Male broodfish ranged in size from 30.5- to51.0-cm TL (600–2,050 g), 35 cm or larger being the preferredsize. Female broodfish ranged in size from 35.5- to 76.2-cmTL (800–5,450 g), with a preferred size range of 40.6–61.0-cmTL. Larger females frequently consumed smaller males duringfailed courtship attempts. Each broodfish was genetically testedand authenticated as pure Florida Largemouth Bass using mi-crosatellite DNA testing methods described by Lutz-Carrilloet al. (2006). All broodfish were implanted with a 12-mm,125-kHz glass PIT tag (Biomark, Boise, Idaho) “naming” in-dividual bass with a 10-digit code. Individual PIT tags wereutilized to cull intergrades (bass not 100% Florida LargemouthBass) and for pairing different male and female groups each newspawning event for maintaining genetic diversity as describedin Austin et al. (2012).

The OS manipulation procedure began each year on approx-imately June 18 and was completed by September 20. The goalwas to simulate winter–spring temperatures and day length overa 90-d period to naturally induce spawning without the use ofhormones. To our knowledge, this was the first time tempera-ture and photoperiod manipulation without hormone injectionshad been used to produce bass on a production scale for the pur-pose of producing large quantities of advanced-sized fingerlingsfor early spring. Co-mingled bass broodstock were stocked into24 × 2.5 × 1-m concrete raceways with a 0.6-m standpipe re-ducing water volume to 35,900 L. Each raceway was equippedwith a 20-ton heater–chiller (Trane air-cooled heat pump [460 V,three phase]; Lacrosse, Wisconsin) designed to run at 454 L/minusing a 3-hp water pump (Model JX22SS [208 V, three phase];BJM Pumps, Old Saybrook, Connecticut). Generally, a maxi-mum of 150 broodfish were stocked per raceway during pres-pawning conditioning. The water temperature was slowly re-duced from ambient temperatures at a rate of 1–2◦C every 2 to 3d over a 4-week period until water temperatures reach 10–12◦C.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

526 MATTHEWS AND STOUT

FIGURE 1. Temperature (◦C) and daylight (H) manipulated for a 90-d periodin June–September each year to exaggerate natural winter conditions experi-enced normally in Webster, Florida.

Water temperatures were continuously recorded using HOBOdata loggers (HOBOware version 3.3; Onset Computer, Bourne,Massachusetts). Broodstock were held at 10–12◦C for approxi-mately 3 to 4 weeks. The sequence was then reversed and watertemperatures increased 1–2◦C every 2 to 3 d over a 4-weekperiod to ambient temperatures (Figure 1).

Concurrent with chilling the broodfish, day length was ma-nipulated to exaggerate shorter daylight periods experienced inwinter. Each raceway was covered with a single sheet of blackplastic to exclude all ambient light. Exaggerated shortened daylengths and complete darkness was crucial to manipulating fishto spawn without hormone injection (J. Chappell, Auburn Uni-versity, personal communication). Initially, raceways had a 10-hphotoperiod for the first 4 weeks, dropping to an 8-h photope-riod during the 3 to 4 weeks spent at 10–12◦C. Photoperiod wasincreased to 10 h for 2 weeks and then increased to 14 h forthe last 2 weeks of conditioning through the end of the spawn-ing period (Figure 1). The bass were then ready for placementinto a spawning raceway. Bass were fed live prey, Koi Cypri-nus carpio, once a week at 3% body weight during the first8 weeks of the conditioning process. The Koi, raised on site,were required since wild adult bass will not accept a pelletedration. Feed amounts were increased to satiation twice a week

FIGURE 2. Image of a 24-m concrete spawning raceway containing (A)spawning mats spaced 1.8 m apart and displaying (B) the black plastic coverused to exaggerate periods of darkness. [Figure available in color online.]

thereafter until initiation of spawning in an effort to maximizespawning condition. Bass broodstock were not fed during theactual spawning period.

Brood bass were redistributed in spawning raceways (sameraceways used for the photothermal manipulation) at a 1:1 ora 2:3 male-to-female sex ratio (20 males to 20 or 30 females).Water level in the raceway was set to include 0.3–0.5 m of free-broad from the top of the raceway. Spawning efficiency was notaffected by the deceased water level, but this helped prevent thebass from jumping out of the tank during initial acclamation tothe “dark” period. Twenty, 55 × 61-cm Spawntex mats, 10 matsper side, were placed 1.8 m apart against the interior walls of theraceway. Raceways received well water, filtered and treated withultraviolet light from a lined reservoir at a 190 L/min flow rate.No other structures were used in the raceways (Figure 2). Matscontaining eggs were removed and immediately replaced eachmorning. The mats were hung in 9.2-m fiberglass raceways forincubation and treatment as described by Matthews et al. (2012).The 5–7-d-old swim-up fry were stocked into fertilized outdoorponds at ≤197,600/ha (80,000/acre). Bass fry fed on naturalzooplankton in ponds and were harvested 25–40 d after stocking(35–40 mm). Monitored afternoon outdoor pond temperaturesat the FBCC range from 32◦C in early October to 20◦C in mid-November.

RESULTS AND DISCUSSION

Off-Season Natural Spawning, 2008–2009Initial off-season annual spawning attempted without

photoperiod and temperature manipulation in six concreteraceways in December 2008–January 2009 produced 200spawns, collected between December 11 through January 27

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

COMMUNICATION 527

TABLE 1. Production totals and averages of unmanipulated Florida Largemouth Bass spawned off-season December 2008–January and March 2009 in indoorraceways. Broodfish are wild-caught and maintained on live forage at the FBCC. The use of F1 generation broodstock is strictly prohibited.

Total Dead Spawns per Fry Total Fry SpawningRaceway spawns spawns female Fry/g per spawn fry per female females

2008R-1 72 3 3.1 310 6,848 472,504 15,122 23R-2 55 1 2.5 307 6,762 365,125 16,597 22R-3 45 0 2.0 311 7,045 317,041 13,784 23R-4 52 0 2.4 311 6,601 343,229 15,601 22R-5 9 0 0.4 316 7,768 69,914 3,040 23R-6 26 0 1.1 315 6,697 174,125 7,571 23

259 4 1.9 312 6,954 1,741,938 11,953 23

(48 d), and an additional 59 spawns March 8 through March 17(Table 1). The first spawn was collected 11 d after introductioninto the raceways. Longer latency periods in initial spawningsuccess of nonacclimated broodstock to raceway environmentswas demonstrated by Mayes et al. (1993) when they experi-enced a 7-d period until spawning. The 1.2 million swim-up fryproduced in December–January were stocked in 21 fertilizedearthen ponds at ≤197,600/ha from December 18 to February3. Low mean survival at harvest (10 ± 16.5%) from ponds wasdue to unpredictable coldwater temperatures and resulting re-duced zooplankton production. Water temperatures in fry pondsdropped as low as 10.3◦C and constantly remained below 15◦Cfor an 11-d period, January 15–25. Early life stages of bass areless tolerant to severe cold temperatures (Fullerton et al. 2000).Carmichael et al. (1988) reported Florida Largemouth Bass ad-vanced fingerlings (10.3 ± 3.1 g) survived temperatures cooledto 1◦C for less than 2 h, but populations experienced 48% mor-tality when held at 2◦C for 5 d.

The extended 48-d spawning off-season period, reducedforage availability, and unpredictable Florida winter tempera-tures terminated further investigation of off-season spawning.We determined that bass needed to be spawned in the fall totake advantage of warmer Florida climate suitable for adequatezooplankton production in outdoor ponds. A technique thatproduces all Florida Largemouth Bass fry required in a shortwindow of time would decrease losses from cannibalism andmaximize tank space. We determined fry would need to bestocked in fertilized ponds no later than mid-October formore consistent warmer water temperatures. Delaying springspawning by holding broodstock below spawning temperaturefor periods ≥8 months has been successfully accomplished(Brauhn et al. 1972; Carlson 1973) but was impractical at theFBCC when other species production needs were considered.

Out-of-Season Photothermal-Manipulated Spawning,Fall 2009–2011

The first attempt to manipulate brood Florida LargemouthBass to re-spawn in the fall was successfully accomplishedin October 2009. The OS technique was successfully repeated

in September–October of 2010 and 2011 (Figure 3). Out-of-season spawning resulted in consistently producing approx-imately 1,000,000 swim-up fry in a short 3–4-week timeframe (Table 2). Total numbers of spawns collected each year

FIGURE 3. Total number of spawns collected each morning and daily temper-ature (◦C) corresponding to the 26-, 31-, and 23-d spawning period needed toproduce required swim-up fry during each OS production cycle from three con-tinuous years. The photothermal periods displayed in Figure 1 combined withthe respective data shown here depict the complete manipulation and spawningduration incurred for each of the three OS periods.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3


TABLE 2. Production totals and averages of photoperiod and temperature-manipulated wild-caught Florida Largemouth Bass spawned out of season in indoorraceways, September–October 2009–2011. Data represents fry production in 26-, 31-, and 23-d time periods for their respective year.

Total Dead Spawns Fry Total Fry SpawningRaceway spawns spawns per female Fry/g per spawn fry per female females

2009R-1 34 1 1.4 308 5,386 177,728 7,405 24R-2 37 0 1.4 317 4,986 184,480 7,095 26R-3 17 0 1.2 319 5,460 92,817 6,630 14R-4 36 0 1.4 313 5,541 199,473 7,979 25R-5 36 0 1.4 317 5,524 198,866 7,955 25R-6 33 1 1.3 312 5,093 162,964 6,519 25

193 2 1.4 314 5,332 1,016,328 7,264 23

2010R-1 30 0 2.0 319 6,271 188,115 12,541 15R-2 37 0 1.9 317 6,566 242,951 12,148 20R-3 22 0 1.5 317 6,712 147,654 9,844 15R-4 47 0 2.6 316 6,765 317,973 17,665 18R-5 47 0 2.6 317 6,681 313,996 17,444 18R-6 22 0 1.4 315 6,656 146,424 9,152 16

205 0 2.0 317 6,608 1,357,113 13,132 17

2011R-1 36 0 1.5 300 4,048 145,742 6,073 24R-2 7 0 0.6 300 5,286 37,004 3,084 12R-3 20 0 0.5 300 4,921 98,419 2,289 43R-4 52 0 1.1 300 5,177 269,212 5,728 47R-5 48 0 1.2 300 4,669 224,119 5,747 39R-6 36 0 1.2 300 4,807 173,049 5,768 30

199 0 1.0 300 4,818 947,545 4,781 33

(2009–2011) were 193, 205, and 199 spawns, respectively, andrepresented an average 1.4, 2.0, and 1.0 spawns per female.Parental contribution from individual broodfish was unknown,but a previous study conducted at the FBCC using nine mi-crosatellite loci determined 70% of 120 bass broodstock con-tributed offspring (Austin et al. 2012). Fry counts averaged 314and 317 swim-up fry/g in fall 2009 and 2010, respectively, andwere similar to average fry counts obtained from spring. For sim-plification fry counts in 2011 were standardized at 300 fry/g andremains the set standard at the FBCC for spring and fall produc-tion (Tables 2, 3). Mean values for OS measured 5,332; 6,608;

and 4,818 fry per spawn, respectively, for the 2009 through 2011OS cycles.

Spawning OS (2009–2011) yielded 7,264; 13,132; and 4,781fry per female, respectively (Table 2). Amounts were com-parable to reports of 4,680–18,510 eggs per female (Jackson1979) and fecundity ranges established for Largemouth Bass(Carlander 1977). Recording mean number of eggs per femalewas not standard procedure due to the increased labor and eggmortality caused by removing the adhesive eggs from the spawn-tex mats. Efforts to determine total number of eggs per spawn infall 2010 and spring 2011 utilized sodium sulfate, 45 g/L, and

TABLE 3. Production summary of wild-caught Florida Largemouth Bass spawned under normal spring spawning conditions in indoor raceways in 2010–2012following OS spawning seasons.

Total Dead Spawns Fry Total Fry SpawningSpring spawns spawns per female Fry/g per spawn fry per female females

2010 109 8 1.1 328 11,828 1,191,092 12,325 972011 126 0 1.2 300 5,820 730,910 5,984 1222012 91 0 0.7 300 11,018 1,001,715 7,269 138

1.0 9,555 8,526

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

COMMUNICATION 529

mild water spray to remove eggs from 50 and 19 spawning mats.Mean number of eggs ( ± SD) per mat were 5,540 ( ± 784) and8,270 ( ± 981), respectively. Percent hatch was less than 50%from both trials and was likely the result of egg damage occurredduring removal.

Spawning LatencyCo-mingled bass broodstock prior to spawning reduced the

time to initiate spawning. In 2009, broodstock were separated bysex in the photothermal-manipulated raceways. The separationwas thought necessary to prevent unwanted spawning as seenin outdoor ponds. Spawning began 17 d after placement intothe spawning raceways. In 2010 and 2011, broodfish were co-mingled for the duration of the OS photothermal manipulationand spawning was initiated in 5 and 3 d upon placement in thespawning raceways. Bass allowed to acclimate with each otherand to the raceway environment needed less time to producespawns. Similar findings were reported by Jackson (1979) andMayes et al. (1993). Hormone injection produced spawns in 48 h(Mayes et al. 1993) but would be labor intensive and expensiveconsidering the numbers of broodfish (300: 180 females and 120males) utilized in a typical FBCC production season. Handlingstress was initially thought the cause of the delayed spawningin 2009 (Ostrand et al. 2004). Identical handling procedures in2010 and 2011 dismissed this initial hypothesis.

Spawning BehaviorAll spawns collected in the 2009–2011 OS events were prod-

ucts of afternoon and night spawning. The 14-h light phase be-gan at 0600 hours. No spawning activity was observed between0600 and 1200 hours. Males displayed courtship behavior asearly as 1200 hours after selecting a spawning mat and contin-ued until attracting a female or “dark” out at 2000 hours. Isaacet al. (1998) suggested males always used the same spawningsite over an 11-d period. Observations from FBCC spawningraceways showed males continued to produce spawns at a sin-gle site for 1–3 d but would move to new mats over the durationof the spawning period. Males continued to fan and protect re-placed mats after the removal of the previous day’s spawn foran unknown duration. Individual courtship and spawning timeand duration varied. All spawning, with the exception of sixspawns in 2010, were initiated and completed between 1300and 2000 hours, and mimicked observations reported by Isaacet al. (1998). Six spawns were initiated and completed at nightbetween 2000 and 0800 hours the next day. Our data agree withafternoon and evening spring spawning occurring in the wildand in concrete raceways observed by Kelley (1968) and Isaacet al. (1998). Duration of each spawning event was not recordedeach year, but singular observed events lasted up to 180 min.

Spring Production versus Out-of-SeasonProduction Comparisons

Recorded spring and OS spawning periods (1300–2000hours), latency of spawning (4–7 d), and broodstock ratios (1:1

and 2:3, respectively) were similar between spring and fall. Nodifference (not statistically substantiated) between egg viability,percent hatch, and survival after approximately 30 d reared onzooplankton in ponds (spring: 60–100%, 80–100%, and 65 ±9% [ ± SD]; fall: 65–95%, 80–100%, and 60 ± 17% [ ± SD])was observed. More variability in pond survival was noted inponds stocked after mid-October.

Mean OS fry per spawn values (4,000–7,000) were lowerthan spring FBCC production totals that averaged 6,000–12,000fry per spawn. The difference between OS and spring in fry perspawn was observed in 2 of the 3 years (Tables 2, 3). Springproduction of Florida Largemouth Bass at the Texas Freshwa-ter Fisheries Center reported similar numbers of fry per spawn(8,000–12,000; J. Martinez, Texas Parks and Wildlife Division,personal communication). Given that the same group of brood-stock was spawned in successive fall–spring–fall–spring fre-quency, the influence from repeated spawning events may havehad a negative effect on fry per spawn values during the spring2011 period, but reduced fry numbers were not evident in the2010 and 2012 spring seasons respawning bass from the previ-ous fall (Table 3).

The fry per spawn value is of significant importance. It isused to estimate the number of spawns required to reach produc-tion goals and is used to set the number of required broodstock.Raceway space and the number of raceways needed for photope-riod and temperature manipulation are determined using thesevalues as well. Annual reproductive cycles and gonadosomaticindices (GSIs) have been characterized for natural spring spawnof Florida Largemouth Bass and Largemouth Bass. The effectsof spawning the same broodfish twice a year for concurrentyears remains to be evaluated.

The annual reproductive cycle of hatchery-raised FloridaLargemouth Bass—focusing on gonadal development, sexsteroid concentrations, and vitellogenin—was characterizedover a 12-month period starting in January (Gross et al.2002). Florida Largemouth Bass GSI in both sexes peaked inFebruary–March with lowest indices recorded in July–October.Rosenblum et al. (1994) reported similar findings and observedfully regressed gonads in September–October in hatchery bassraised on forage or pelleted feeds. We did not sacrifice brood-stock to characterize OS GSI, but Brauhn et al. (1972) reporteda mean GSI of 3.9% (range = 1.7–8.4%) in Largemouth Bassheld at 17◦C and off-season spawned in August. The August GSIdata reported by Brauhn et al. (1972) was similar to GSI valuesrecoded in February–April reported by Rosenblum et al. (1994)and Gross et al. (2002), a month in which both report lowest GSIlevels. These reports show that bass spawned naturally or off-season have a unimodal GSI peak, suggesting a distinct annualreproduction cycle. It may be of value to recreate the Gross et al.(2002) study to characterize our wild-caught and permanentlyretained hatchery bass that are spawned biannually. Evidenceof a forced secondary bimodal GSI peak that showed gonaddevelopment again during the 90-d photothermal manipulationperiod could help resolve the issue of reduced numbers of fry

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3


collected per spawn in the fall or may suggest bass that do notspawn in the spring spawn out of season.

CONCLUSIONEstablishing a consistent, production-scale OS spawning pro-

tocol to produce sufficient quantities of Florida LargemouthBass fry and fingerlings to support a stocking program in a de-fined period was the objective of this project. Our efforts at theFBCC have been successful for the previous four OS spawningcycles, including fall 2012. Florida’s mild fall climate allows forrearing bass fry in fertilized outside ponds utilizing zooplank-ton. Application of this OS spawning technique may be limitedto certain geographic locations due to negative effects of lowerambient water temperature extremes. Procedures outlined forOS spawning allows for biannual production seasons using thesame broodstock population. Producing advanced fingerlingsin the fall and winter months when pond and indoor racewayspace is readily available increased hatchery efficiency and totalannual production. This biannual spawning also permits cul-ture and stocking of 100-mm TL advanced-sized fingerlings byMarch. Characterizing biannual reproductive cycles (spring andfall), fry per spawn between seasons from the same females, andtotal egg and fry production over a complete biannual seasonneeds further investigation.

ACKNOWLEDGMENTSThe authors would like to acknowledge all the regional FWC

staff for their assistance in collecting broodfish. We also want toexpress our gratitude to hatchery staff members Joshua Sakmar,Justin Elkins, Rick Thompson, and Dewayne Thomas for all thelong hours and hard work they provided. Finally, we would liketo thank Paul Cardeilhac, Kelly Winningham, Juan Martinez,and two anonymous reviewers for their insightful comments onthe manuscript.

REFERENCESAustin, J. D., A. Johnson, M. Matthews, M. D. Tringali, W. F. Porak, and

M. S. Allen. 2012. An assessment of hatchery effects on Florida Bass (Mi-cropterus salmoides floridanus) microsatellite genetic diversity and sib-shipreconstruction. Aquaculture Research 43:628–638.

Brauhn, J. L., D. Holz, and R. O. Anderson. 1972. August spawning of Large-mouth Bass. Progressive Fish-Culturist 34:207–209.

Carlander, K. D. 1977. Handbook of freshwater fishery biology, volume 2. IowaState University Press, Ames.

Carlson, A. R. 1973. Induced spawning of Largemouth Bass [Micropterussalmoides (Lacepede)]. Transactions of the American Fisheries Society102:442–444.

Carmichael, G. J., J. H. Williamson, C. A. Caldwell Woodward, and J. R.Tomasso. 1988. Responses of Northern, Florida, and hybrid LargemouthBass to low temperature and low dissolved oxygen. Progressive Fish-Culturist50:225–231.

Diana, M. J., and D. H. Wahl. 2009. Growth and survival of four sizes ofstocked Largemouth Bass. North American Journal of Fisheries Management29:1653–1663.

Fullerton, A. H., J. E. Garvey, R. A. Wright, and R. A. Stein. 2000. Overwintergrowth and survival of Largemouth Bass: interactions among size, food,origin, and winter severity. Transactions of the American Fisheries Society129:1–12.

Gross, T. S., C. M. Wieser, M. S. Sepulveda, J. J. Wiebe, T. R. Schoeb, and N.D. Denslow. 2002. Characterization of annual reproductive cycles for pond-reared Florida Largemouth Bass Micropterus salmoides floridanus. Pages205–212 in D. P. Philipp and M. S. Ridgway, editors. Black bass: ecology,conservation, and management. American Fisheries Society, Symposium 31,Bethesda, Maryland.

Hearn, M. C. 1977. Post-stocking survival of Largemouth Bass reared in race-ways on an artificial diet. Progressive Fish-Culturist 39:126–127.

Heidinger, R. C. 1975. Life history and biology of the Largemouth Bass. Pages11–20 in R. H. Stroud and H. E. Clepper, editors. Black bass biology andmanagement. Sport Fishing Institute, Washington, D.C.

Isaac, J., Jr., T. M. Kimmel, R. W. Bagley, V. H. Staats, and A. Barkoh. 1998.Spawning behavior of Florida Largemouth Bass in an indoor raceway. Pro-gressive Fish-Culturist 60:59–62.

Jackson, U. T. 1979. Controlled spawning of Largemouth Bass. ProgressiveFish-Culturist 41:90–95.

Kelley, J. W. 1968. Effects of incubation temperature on survival of LargemouthBass eggs. Progressive Fish-Culturist 30:159–163.

Kramer, R. H., and L. L. Smith Jr. 1962. Formation of year classes in LargemouthBass. Transactions of the American Fisheries Society 91:29–41.

Lamkin, J. B. 1900. The spawning habits of the Large-mouth black bass in thesouth. Transactions of the American Fisheries Society 29:129–138.

Loska, P. M. 1982. A literature review on the stocking of black basses (Mi-cropterus spp.) in reservoirs and streams. Georgia Department of NaturalResources, Federal Aid in Sport Fish Restoration, Project SW-1, Final Re-port, Atlanta.

Lutz-Carrillo, D. J., C. C. Nice, T. H. Bonner, M. R. J. Forstner, and L. T.Fries. 2006. Admixture analysis of Florida Largemouth Bass and NorthernLargemouth Bass using microsatellite loci. Transactions of the AmericanFisheries Society 135:779–791.

Matthews, M. D., J. C. Sakmar, and N. Trippel. 2012. Evaluation of hydrogenperoxide and temperature to control mortality caused by Saprolegniasis and toincrease hatching success of Largemouth Bass eggs. North American Journalof Aquaculture 74:463–467.

Mayes, K. B., P. M. Rosenblum, and T. M. Brandt. 1993. Raceway spawning ofFlorida Largemouth Bass: effects of acclimation time and hormone treatmenton spawning success. Progressive Fish-Culturist 55:1–8.

Nelson, J. T., R. G. Bowker, and J. D. Robinson. 1974. Rearing pellet-fedLargemouth Bass in a raceway. Progressive Fish-Culturist 36:108–110.

Ostrand, K. G., S. J. Cooke, and D. H. Wahl. 2004. Effects of stress on Large-mouth Bass reproduction. North American Journal of Fisheries Management24:1038–1045.

Porak, W. F., N. A. Trippel, B. Thompson, and S. Bisping. 2011. Stock enhance-ment of Largemouth Bass. Florida Fish and Wildlife Conservation Commis-sion, Federal Aid in Sport Fish Restoration Act, Grant F-132-R, CompletionReport, Tallahassee.

Pouder, W. F., N. A. Trippel, and J. R. Dotson. 2010. Comparison of mortalityand diet composition of pellet-reared advanced-fingerling and early-cohortage-0 wild Largemouth Bass through 90 days poststocking at Lake Semi-nole, Florida. North American Journal of Fisheries Management 30:1270–1279.

Rogers, M. W., M. S. Allen, and W. F. Porak. 2006. Separating genetic andenvironmental influences on temporal spawning distributions of LargemouthBass (Micropterus salmoides). Canadian Journal of Fisheries and AquaticSciences 63:2391–2399.

Rosenblum, P. M., T. M. Brandt, K. B. Mayes, and P. Hutson. 1994. Annualcycles of growth and reproduction in hatchery-reared Florida LargemouthBass, Micropterus salmoides floridanus, raised on forage or pelleted diets.Journal of Fish Biology 44:1045–1059.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

COMMUNICATION 531

Simco, B. A., J. H. Williamson, G. J. Carmichael, and J. R. Tomasso. 1986.Centrarchids. Pages 73–89 in R. R. Stickney, editor. Culture of nonsalmonidfreshwater fishes. CRC Press, Boca Raton, Florida.

Snow, J. R. 1968. Production of six- to eight-inch Largemouth Bass for specialpurposes. Progressive Fish-Culturist 30:144–152.

Snow, J. R., and J. I. Maxwell. 1970. Oregon moist pellet as a production rationfor Largemouth Bass. Progressive Fish-Culturist 32:101–102.

Swingle, H. S. 1956. Appraisal of methods of fish population study: IV. deter-mination of balance in farm fish ponds. Transactions of the North AmericanWildlife and Natural Resources Conference 21:298–322.

Tringali, M. D., T. M. Bert, F. Cross, J. W. Dodrill, L. M. Gregg, W. G. Halstead,R. A. Krause, K. M. Leber, K. Mesner, W. Porak, D. Roberts, R. Stout, andD. Yeager. 2007. Genetic policy for the release of finfishes in Florida. FloridaFish and Wildlife Conservation Commission, Florida Fish and Wildlife Re-search Institute, Publication IHR-2007-001, St. Petersburg.

Wahl, D. H., R. A. Stein, and D. R. DeVries. 1995. An ecological frame-work for evaluating the success and effects of stocked fishes. Pages 176–189 in H. L. Schramm Jr., editor. Uses and effects of cultured fishes inaquatic ecosystems. American Fisheries Society, Symposium 15, Bethesda,Maryland.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3



Determining Optimum Temperature for Growth andSurvival of Laboratory-Propagated Juvenile FreshwaterMusselsCaitlin S. Carey a , Jess W. Jones b , Eric M. Hallerman c & Robert S. Butler da Department of Fish and Wildlife Conservation , Virginia Polytechnic Institute and StateUniversity, Cheatham Hall, Room 149 , 310 West Campus Drive, Blacksburg , Virginia ,24061 , USAb U.S. Fish and Wildlife Service, Department of Fish and Wildlife Conservation , VirginiaPolytechnic Institute and State University, Cheatham Hall, Room 106a , 310 West CampusDrive, Blacksburg , Virginia , 24061 , USAc Department of Fish and Wildlife Conservation , Virginia Polytechnic Institute and StateUniversity, Integrated Life Sciences Building, Room 1021 , 1981 Kraft Drive, Blacksburg ,Virginia , 24061 , USAd U.S. Fish and Wildlife Service , Asheville Field Office , 160 Zillicoa Street, Asheville , NorthCarolina , 28801 , USA

To cite this article: Caitlin S. Carey , Jess W. Jones , Eric M. Hallerman & Robert S. Butler (2013) Determining OptimumTemperature for Growth and Survival of Laboratory-Propagated Juvenile Freshwater Mussels, North American Journal ofAquaculture, 75:4, 532-542






http://dx.doi.org/10.1080/15222055.2013.826763




ARTICLE

Determining Optimum Temperature for Growthand Survival of Laboratory-Propagated JuvenileFreshwater Mussels

Caitlin S. Carey*Department of Fish and Wildlife Conservation, Virginia Polytechnic Institute and State University,Cheatham Hall, Room 149, 310 West Campus Drive, Blacksburg, Virginia 24061, USA

Jess W. JonesU.S. Fish and Wildlife Service, Department of Fish and Wildlife Conservation,Virginia Polytechnic Institute and State University, Cheatham Hall, Room 106a,310 West Campus Drive, Blacksburg, Virginia 24061, USA

Eric M. HallermanDepartment of Fish and Wildlife Conservation, Virginia Polytechnic Institute and State University,Integrated Life Sciences Building, Room 1021, 1981 Kraft Drive, Blacksburg, Virginia 24061, USA

Robert S. ButlerU.S. Fish and Wildlife Service, Asheville Field Office, 160 Zillicoa Street, Asheville,North Carolina 28801, USA

AbstractThe effects of temperature on growth and survival of laboratory-propagated juvenile freshwater mussels of two

federally endangered species, the Cumberlandian combshell Epioblasma brevidens and oyster mussel E. capsaeformis,and one nonlisted species, the wavy-rayed lampmussel Lampsilis fasciola, were investigated to determine optimumrearing temperatures for these species in small water-recirculating aquaculture systems. Juveniles 4–5 months oldwere held in downweller buckets at five temperatures. Growth and survival of juveniles were evaluated at 2-weekintervals for 10 sampling events. At the end of the 20-week experiment, mean growth at 20, 22, 24, 26, and 28◦Cwas, respectively, 0.75, 2.22, 3.27, 4.23, and 4.08 mm for Cumberlandian combshell; 1.35, 3.73, 3.81, 4.90, and4.70 mm for oyster mussel; and 2.09, 3.96, 4.99, 5.13, and 4.87 mm for wavy-rayed lampmussel juveniles. Generally,temperature was positively correlated with growth of juveniles. Final mean maximum growth occurred at 26◦C for allthree species, although no significant differences in growth were detected between 26◦C and 28◦C. The relationshipbetween temperature and survival of juveniles was less clear. Final survival was 82.5, 89.0, 91.0, 89.5, and 93.5%for Cumberlandian combshell; 73.0, 83.5, 78.0, 78.0, and 68.1% for oyster mussel; and 75.0, 89.5, 87.0, 86.5, and89.5% for wavy-rayed lampmussel juveniles at the five temperature treatments, respectively. Based on the speciesused in this study, results indicate that 26◦C is the optimum temperature to maximize growth of juvenile mussels indownweller bucket systems. The ability to grow endangered juveniles to larger sizes will improve survival in captivityand upon release into the wild and will reduce time spent in hatcheries. As a result, hatcheries can increase theiroverall production and enhance the likelihood of success of mussel population recovery efforts by federal and stateagencies, and other partners.


532

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

LABORATORY-PROPAGATED JUVENILE MUSSELS 533

Because of significant declines of mussel populations in re-cent decades (Williams et al. 1993; Neves et al. 1997; Neves1999), and with the culture and release of laboratory-propagatedmussels into the wild being used as a recovery method (USFWS2003, 2004; Jones et al. 2005, 2006; Eckert and Pinder 2010),there is a growing need to improve culture methods, particu-larly grow-out of propagated juveniles. Water temperature isa significant environmental variable affecting growth and sur-vival of juvenile mussels in captivity and in the wild, which alsoaffects various reproductive processes in adults, such as game-togenesis, spawning, and larval brooding (Krebs 1972; Hastieet al. 2000; Zimmerman and Neves 2002; Gosling 2003; Hastieet al. 2003; Zimmerman 2003; Jones et al. 2005; Negishi andKayaba 2010; Pandolfo et al. 2010b). Temperature can affectmussel developmental and physiological processes and also hasspecific effects on different life stages (Krebs 1972; Negishiand Kayaba 2010; Pandolfo et al. 2010b). Efforts to propagateand culture mussels in captivity require an understanding of theenvironmental factors that influence growth and survival at dif-ferent life stages. Thus, defining the optimum temperature forproduction of laboratory-propagated juvenile mussels is criticalfor optimizing propagation and culture success and thereby hasimportant implications for their conservation.

In the past few years, it has become clear that larger and olderlaboratory-propagated juveniles have a significantly increasedchance of survival when released in the wild than do newlymetamorphosed juveniles (Sarrazin and Legendre 2000; Huaet al. 2011). Though methods have been developed to producethousands of newly metamorphosed juveniles, refinement ofculture methods to grow these species to larger sizes is needed.The ability to grow juveniles of imperiled species to larger sizesimproves survival of individuals while captive and upon releaseto the wild by decreasing the incidence of predation in bothsettings. In addition, enhancing grow out of cultured musselsincreases detection probabilities for subsequent monitoring and,most importantly, improves population recovery (Zimmermanet al. 2003; Hua et al. 2011).

The purpose of this study was to determine the effect of tem-perature on the growth and survival of juvenile (>4 months oldand ≥1.5 mm) mussels of two federally endangered species,Cumberlandian combshell Epioblasma brevidens and oystermussel E. capsaeformis, and one nonlisted species, wavy-rayedlampmussel Lampsilis fasciola, in captivity. The intent of thisresearch was to determine optimum rearing temperatures tomaximize growth and survival of juvenile mussels of these threemussel species in captivity.

METHODSGravid mussel collection.—Juveniles were produced by the

Freshwater Mollusk Conservation Center (FMCC) at VirginiaPolytechnic Institute and State University (Virginia Tech) inBlacksburg, and Virginia Department of Game and Inland Fish-eries’ Aquatic Wildlife Conservation Center (AWCC) near

Marion, Virginia, following standard propagation and cul-ture methods for these organisms. Gravid females of eachspecies were collected in May 2011 by snorkeling and us-ing view scopes in the lower Clinch River, Hancock County,Tennessee. Gravid individuals were held and transported tothe FMCC and AWCC in coolers containing river water withaeration.

After arriving at the facilities, gravid females were placed inholding systems with maintained water temperatures of 15◦C toprevent early glochidial release before infestation of host fishescould be conducted. The holding system at the FMCC contained50–80 mm of river substrate (pebble, gravel) and water from thefacility’s pond; the holding system at the AWCC contained 50–80 mm of coarse limestone gravel substrate and water sourcedfrom the South Fork Holston River. Mussels were fed daily witha premixed commercial algae diet (Nanno 3600 and ShellfishDiet 1800 from Reed Mariculture, Campbell, California).

Host fish collection and care.—Based on the results of pre-vious studies (Zale and Neves 1982; Yeager and Saylor 1995),Black Sculpin Cottus baileyi were used as the host fish forthe Cumberlandian combshell and oyster mussel, and Large-mouth Bass Micropterus salmoides were used as the host forthe wavy-rayed lampmussel. Black Sculpin were collected us-ing a backpack electrofisher (Model LR24, Smith-Root, Van-couver, Washington) and Largemouth Bass were obtained froma regional fish farm in Arkansas.

Black Sculpin were held and transported to each facility in140-L coolers containing local stream water. Salt was added tocoolers to increase salinity to 0.7‰ in order to reduce fish stressduring transport. Water temperature was maintained at ambientstream levels during transportation, and dissolved oxygen wasmaintained using an aerator. Transport time ranged from 1 to2 h. After arrival at culture facilities, fish were acclimated tolaboratory conditions, regarding temperature and salinity, andwere quarantined for 2–3 d at a salinity of 3.0‰ prior to beinginfested with glochidia.

Infestation with mussel glochidia and juvenile mussel col-lection.—Host fish were infested with mussel glochidia follow-ing FMCC established nonlethal laboratory protocols (Zale andNeves 1982; Neves 2004). At the FMCC, 180 Black Sculpinwere separated into groups of 45 fish, which were placed intoone of four 16-L containers with 3.5 L of conditioned waterat 21◦C under continuous aeration. Glochidia from two gravidoyster mussels were mixed into each of the four containers(eight gravid oyster mussels in total) and allowed 45 min toattach to host fish. After infestation, host fish were moved intowater-recirculating aquaculture holding systems. Water qualityvariables were monitored bi-weekly in the host-fish holding sys-tems. Similar host-fish infestation methods were used at AWCCto produce juvenile mussels.

Once juveniles began to excyst from host fish, tank waterwas siphoned daily through 300-µm and 150-µm mesh sieves.Collected juveniles were rinsed into a petri dish, counted, andplaced into 18-L downweller-bucket culture systems for growth

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

534 CAREY ET AL.

FIGURE 1. Top view of recirculating downweller bucket culture system andchambers. [Figure available in color online.]

and development (Barnhart 2006; Figure 1). Buckets were filledwith 18 L of filtered (<5 µm) pond (FMCC) or river (AWCC)water maintained at 20–24◦C, and bucket water was exchangedonce per week. At each water exchange interval, buckets werecleaned and standard water quality parameters were tested. Ju-veniles were fed continuously with a premixed commercial al-gae diet. Young juveniles experience a mortality bottleneck at4–8 weeks of age and are susceptible to flatworm predation atsmall sizes (Henley et al. 2001; Jones et al. 2005); therefore,to remove any confounding factors, juveniles were cultured for4–5 months to the desired initial size of 1–2 mm before theculture experiment was initiated. A summary of gravid musselcollection, captive holding conditions, and host fish infestationprotocols are given for each species in Table 1.

Test conditions.—Juvenile mussels were acclimated to 20◦Cbefore testing and then allowed to acclimate to treatment tem-peratures gradually over a 24-h period. Temperature was con-trolled by a water bath surrounding the buckets that was heldconstant ( ± 0.5◦C) through the use of heaters or chillers, andmonitored daily using a temperature data logger (Onset Com-puter Corporation, HOBO Pendant Logger Model UA-001-08).Water quality testing in each bucket was conducted biweeklyfor ammonia (salicylate method, Hach Method 8155), nitrite(diazotization method, Hach Method 8507), nitrate (cadmiumreduction method, Hach Method 8171), dissolved oxygen, pH,and specific conductivity (YSI Professional Plus Multiparam-eter Meter). Total hardness (mg of Ca/L as CaCO3 plus mgof Mg/L as CaCO3) via the titration method (Hach Method8213) and total alkalinity (Hach Model AL-AP; mg of phe-nolphthalein alkalinity/L plus mg total methyl orange alka-linity/L as CaCO3) were tested on the source water once aweek.

Mussels in each bucket were fed 500 mL daily (21 mL/h) ofa premixed commercial algae formula (mean cell concentration,about 1.0–2.0 × 106 µm3/mL) delivered continuously from a 1-L water bottle through a drip valve. Eighty water samples weretaken randomly from the buckets over the course of the exper-iment in order to quantify the algal cell concentrations using aCoulter counter (Beckman Coulter, Multisizer 3) located at theAWCC. Algal concentrations also were measured by a hemo-cytometer and compared with those from the Coulter counter.Feeding bottles were cleaned, juvenile mussel holding cham-bers were rinsed, and bucket water was completely exchangedonce a week. Air bubbles were removed from culture chambersand pumps and power sources and water levels were checkeddaily, as per the FMCC protocols.

Testing of Cumberlandian combshell and oyster mussel ju-veniles began 18 November 2011 and finished 4 April 2012.Testing of wavy-rayed lampmussel juveniles began 22 Novem-ber 2011 and finished 12 April 2012. Mussels in buckets weresampled at 2-week intervals for 20 weeks to provide a total of10 sampling events. Random samples of 10 of the 40 juvenilesin each chamber were measured under a microscope (Olym-pus American, Model SZ40) to assess mean growth (i.e., meanlength at time t minus initial length). All live individuals andshells within a chamber were counted to assess survival ratessince the start of experiment. Shells of dead mussels were re-moved and documented. A summary of test conditions is givenin Table 2.

Experimental design and statistical analyses.—Five tem-perature treatments were tested (20, 22, 24, 26, and 28◦C),covering the range of normal (20–24◦C) to upper (26–28◦C)temperatures that mussels experience in the wild during thewarmer months of the annual growth period in the ClinchRiver. The test was conducted in recirculating downwellerbucket aquaculture systems, in which each bucket was inde-pendent of others and served as one experimental unit (EU)(Barnhart 2006).

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3


TABLE 1. Summary of gravid female mussel, host-fish collection, and host-fish infestation methods at the Freshwater Mollusk Conservation Center (FMCC)and Aquatic Wildlife Conservation Center (AWCC) in 2011 used to produce juveniles used in this study. All gravid females were collected from the Lower ClinchRiver, Tennessee.

Cumberlandian Wavy-rayedExperimental detail combshell Oyster mussel Oyster mussel lampmussel

Facility AWCC AWCC FMCC AWCCMussel collection month June June May JulyMussel holding system 150-L circular

fiberglass tank150-L circular

fiberglass tank300-L living stream 150-L circular

fiberglass tankHost fish species Black Sculpin Black Sculpin Black Sculpin Largemouth BassFish collection site Middle Fork

Holston, VirginiaMiddle Fork

Holston, VirginiaSouth Fork Holston,

VirginiaRegional Fish Farm,

ArkansasFish holding system AHABa AHABa Quarantine tank RPSb

Infestation month June June May JulyNumber gravid mussels used 1 4 8 4Number of fish used 64 84 180 78Infestation temperature (◦C) 22–24 22–24 21 22–24Duration of infestation (min) 60 60 45 60Infested fish recirculating

aquaculture holding systemAHABa AHABa 76-L Tanksb RPSc

Days till first excystments 12 12 13 12Number available 1,000 500 500 1,000

aAHAB = Aquatic Habitats, Inc. Z-Hab System.bA 2,000-L closed recirculating system made up of twenty 76-L tanks and two sumps.cRPS = recirculating propagation system.

Following a power analysis that determined the appropriatesample size needed to achieve a minimum of 80% power atα = 0.05, each temperature treatment was assigned five in-dependent downweller buckets that served as replicates. Eachbucket contained a total of six juvenile mussel holding cham-bers. The three species were tested alongside one another withinbuckets, with a single chamber containing juveniles of only onespecies, while the other three chambers remained unoccupied(Figure 1). Juveniles of each species were randomly separatedinto chambers containing 40 individuals and placed in 1 of the25 buckets. The EUs were randomly assigned to one of the fivedifferent treatment temperatures (Table 2).

Growth and survival of juveniles for each species were com-pared among temperature treatments using a mixed-model anal-ysis with repeated measures. Growth and survival of juvenileswithin EUs were treated as random effects. Temperature treat-ment, time, and temperature × time interaction were treated asfixed effects. Treatment means were estimated and comparedat each sampling event for further analyses. Survival data (pro-portion survival) for each species was arcsine(x)-transformedbefore being compared among temperature treatments in or-der to meet the normality assumption. Additionally, a one-wayANOVA was used to determine whether algal concentrationsdiffered among treatments. Analyses were conducted using SASsoftware (version 9.2, SAS institute, Inc., Cary, North Carolina,)and were considered significant at the α = 0.05. Unless other-wise stated, all significant results were P < 0.01.

RESULTS

Cumberlandian CombshellFor the five temperature treatment conditions we tested, final

growth at 138 d for Cumberlandian combshell juveniles rangedfrom 0.75 mm at 20◦C to 4.23 mm at 26◦C (Table 3; Figure 2).Analysis of simple effects (i.e., separating the data by samplingevents and conducting one-way ANOVAs at each time step)revealed significant differences in growth between temperaturetreatments at each of the 10 sampling events. Results of themixed model analysis for growth indicated that the fixed effectsof temperature, time, and temperature × time interaction wereall significant.

Contrasts of differences in treatment means (effect size) forfinal growth (i.e., mean shell length at final sample minus initialmean shell length) revealed that growth at 20◦C was signifi-cantly lower than that of all of the other treatment tempera-tures. Growth at 22◦C was significantly lower than growth at24, 26, and 28◦C, and growth at 24◦C was significantly lowerthan growth at 26◦C and 28◦C. No significant difference injuvenile growth was detected between 26◦C and 28◦C (P =0.36).

Cumberlandian combshell juvenile survival ranged from82.5% to 93.5% (Table 3; Figure 3). Examination of simpleeffects for temperature treatments at individual sampling eventsshowed some significance (P = 0.05) of temperature on survivalat the fourth (day 54) sampling event; however, survival was not

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

536 CAREY ET AL.

TABLE 2. Experimental design and test conditions for culture temperature tests of Cumberlandian combshell, oyster mussel and wavy-rayed lampmusseljuveniles at the Freshwater Mollusk Conservation Center, November 2011–April 2012.

Statistical analysis Test conditions

Experimental design Randomized, repeated measuresTest system Downweller buckets with six chambersTest duration (weeks) 20Test bucket volume (L) 18Water renewal Every 7 dInitial age of juveniles Cumberlandian combshell: 4.5 months

Oyster mussel: 5 monthsWavy-rayed lampmussel: 4.5 months

Initial size of juveniles (mm; mean ± SE) Cumberlandian combshell: 2.2 ± 0.03Oyster mussel: 1.5 ± 0.03Wavy-rayed lampmussel: 1.8 ± 0.03

Chambers/species/bucket 1Juveniles/chamber 40Buckets (replicates)/ treatment 5Juveniles/treatment 200Feeding (each bucket/day) 0.05 mL Nanno 3600: 0.15 mL shellfish diet

1800: 500 mL conditioned waterAlgal cell concentration in bucket Mean range: 1.0 – 2.0 × 106 um3/mLFlow Submersible pumpa, maximum flow = 590 L/hTest water Pond water filtered to <5 µm, mean hardness = 200 mg/L as CaCO3,

alkalinity = 184 mg/L as CaCO3

Test temperatures (◦C) 20, 22, 24, 26, or 28Water quality Bi-weekly testing of ammonia, nitrite, nitrate, dissolved oxygen, pH,

and specific conductivitySampling interval (d) 14Endpoints Growth (mean length at time t minus mean initial length) and survival

(proportion survival)aAquarium Systems Mini-jet model MN-606.

affected by treatment temperature at any other sampling event.Survival was not affected by temperature (P = 0.13), while theeffects of time and temperature × time interaction were sig-nificant. Contrasts of differences in treatment means for final

survival showed that survival was significantly lower at 20◦Cthan at 24◦C (P = 0.05) and 28◦C (P = 0.03). The remainingfinal survival estimates were not significantly different betweenother treatment temperatures.

TABLE 3. Final growth and survival (mean ± SE) of Cumberlandian combshell (CC), oyster mussel (OM), and wavy-rayed lampmussel (WL) juvenilescultured in five temperature treatments. Values followed by different subscripts are significantly different (P < 0.05); the lowercase letters z–w indicate differenceswithin species, the capital letters A–C differences within temperature treatments. The final sampling event occurred at 138, 138, and 141 d for CC, OM, and WLjuveniles, respectively.

Temperature (◦C)

Species 20 22 24 26 28

Growth (mm)CC 0.75 ± 0.04 zA 2.22 ± 0.13 yA 3.27 ± 0.16 xA 4.23 ± 0.16 wA 4.08 ± 0.11 wAOM 1.35 ± 0.09 zB 3.73 ± 0.10 yB 3.81 ± 0.14 yB 4.90 ± 0.10 xB 4.70 ± 0.22 xBWL 2.09 ± 0.12 zC 3.96 ± 0.13 yB 4.99 ± 0.09 xC 5.13 ± 0.14 xB 4.87 ± 0.21 xB

Survival (%)CC 82.50 ± 3.45 zA 89.00 ± 4.37 zyA 91.00 ± 3.76 yA 89.50 ± 1.70 zyA 93.50 ± 1.70 yAOM 73.00 ± 2.00 zyA 83.50 ± 3.02 yA 78.00 ± 4.96 zyB 78.00 ± 6.02 zyB 68.13 ± 4.25 zBWL 75.00 ± 8.44 zA 89.50 ± 3.20 yA 87.00 ± 3.98 zyAB 86.50 ± 3.10 zyAB 89.50 ± 2.89 yA

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3


FIGURE 2. Mean growth versus time for (a) Cumberlandian combshell,(b) oyster mussel, and (c) wavy-rayed lampmussel juveniles cultured in oneof five temperature treatments. Growth measurements were taken at 2-weekintervals for 20 weeks to provide a total of 10 sampling events.

FIGURE 3. Mean survival versus time for (a) Cumberlandian combshell,(b) oyster mussel, and (c) wavy-rayed lampmussel juveniles cultured in one offive temperature treatments. Survival was assessed at 2-week intervals for 20weeks to provide a total of 10 sampling events.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

538 CAREY ET AL.

Oyster MusselFinal growth increment at 138 d for oyster mussel juveniles

ranged from 1.35 to 4.90 mm across all temperature treatments(Table 3; Figure 2). Analysis of simple effects of temperature ateach sampling event revealed significant differences in growthbetween temperature treatments at each of the 10 samplingevents. Mixed-model analysis for growth indicated that the ef-fects of temperature, time, and temperature x time interactionwere all significant.

Contrasts of differences in treatment means for final growthrevealed that growth at 20◦C was significantly lower than growthat all other treatment temperatures. Significantly lower growthalso was observed for the 22◦C and 24◦C treatments thangrowth at 26◦C and 28◦C. Growth was similar between 22◦Cand 24◦C (P = 0.63) and between 26◦C and 28◦C (P = 0.25).

Substantial mortality was observed in one of the EUs of the28◦C treatment during the fourth sampling event (day 54), caus-ing it to be a significant outlier for the survival analysis andviolating the homogeneity of variance assumption. It is unlikelythat the mortality observed was caused by temperature because(1) no other EU within the treatment experienced similar mortal-ity, (2) mortality occurred in a single early sampling event, and(3) mortality ceased in this bucket for all further sampling events(sample events 5–10). It is possible that this single mortalityevent may have been induced by human error during samplingefforts (e.g., handling stress). Data from this outlier unit wereremoved from the mixed-model analysis of survival from thefourth sampling event forward (days 54–138) to reduce modelvariance and thereby to meet the assumption of homogeneity ofvariance. Final survival of oyster mussel juveniles ranged from68.1% to 83.5% (Table 3; Figure 3). Analysis of simple effectsfor temperatures by time under this model revealed no signifi-cant differences in survival between any temperature treatments.Time was significant, whereas the effects of temperature (P =0.16) and temperature × time interaction (P = 0.71) were notsignificant. Contrasts of differences in treatment means for fi-nal survival uncovered significantly lower survival at 22◦C thanat 28◦C (P = 0.03). Final survival contrasts between all othertemperature treatments were not significant.

Wavy-Rayed LampmusselFinal growth at 141 d for wavy-rayed lampmussel juveniles

ranged from 2.09 to 5.13 mm (Table 3; Figure 2). Analysisof simple effects for temperatures by time under this modelrevealed significant differences in growth between temperaturetreatments at each sampling event. All fixed effects for growthwere significant.

Contrasts of differences in treatment means for final growthrevealed that growth at 20◦C was significantly lower than growthat all other temperature treatments. Significantly lower growthwas observed at 22◦C than at 24, 26, and 28◦C. Growth at 24◦Cdid not differ statistically from growth at 26◦C (P = 0.44) and28◦C (P = 0.51). No significant differences in growth weredetected between 26◦C and 28◦C (P = 0.15).

Wavy-rayed lampmussel juvenile final survival ranged from75.0% to 89.5% (Table 3; Figure 3). Statistically significant dif-ferences in survival between some temperature treatments weredetected at the second (day 29, P = 0.04) and sixth (day 86,P = 0.03) sampling events. No significant differences in sur-vival between temperature treatments at other sampling eventswere revealed by examination of simple effects under this model.Survival was not affected by temperature (P = 0.10), while timeand the temperature × time interaction effects were significant.Contrasts of differences in treatment means for final survival re-vealed significantly lower survival at 20◦C than at 22◦C (P =0.04) and 28◦C (P = 0.02). Final survival means were not signif-icantly different between any of the other temperature treatmentcomparisons.

Algal Concentrations and Water QualityAlgal cell concentrations within buckets ranged from 1.54

to 2.06 × 106 µm3/mL (mean = 1.80 × 106 µm3/mL) anddid not differ among temperature treatments (P = 0.23). Tem-peratures within treatments did not vary greatly from targettemperatures ( ±0.2◦C). Ammonia (mean, 0.01 mg/L as NH3),nitrite (0.005 mg/L as NO2), and nitrate (0.2 mg/L as NO3)concentrations within buckets stayed within acceptable levels.Water in buckets had a mean dissolved oxygen concentration of7.33 mg/L, pH of 8.46, and specific conductivity of 393 µS/cm.The source pond water had a total hardness as CaCO3 range from193.76 to 209.09 mg/L (mean = 201.42 mg/L) and alkalinity asCaCO3 was 174.76–193.68 mg/L (mean = 184.22 mg/L).

DISCUSSIONPrevious experimental and observational studies have exam-

ined the direct effects of numerous factors affecting growth andsurvival rates of freshwater bivalves in captivity and the wild.Factors that have been found to correlate with mussel growth andsurvival rates include, but are not limited to, substrate type andsize (Hinch et al. 1986; Liberty et al. 2007), flows and sedimentload (Beaty 1997; Zimmerman 2003; Jones et al. 2005; Libertyet al. 2007; Rypel et al. 2008), toxicant exposure (Pandolfo et al.2010a), mussel density (Hanson et al. 1988; Beaty 1997; Beatyand Neves 2004; Negishi and Kayaba 2009), food availability(Hanlon 2000), sampling frequency (Beaty 1997; Zimmerman2003; Liberty et al. 2007), maturity of larvae (Jones et al. 2005),and temperature (Hanson et al. 1988; Buddensiek 1995; Beaty1997; Hanlon 2000; Zimmerman and Neves 2002; Zimmerman2003; Liberty 2004; Hanlon and Neves 2006; Pandolfo et al.2010a, 2010b; Negishi and Kayaba 2010). These studies havehelped define requirements for mussel propagation and cultureby advancing understanding of factors affecting growth and sur-vival and have shown that mussels are useful biological indica-tors of environmental change. Providing optimal temperaturesfor laboratory-propagated mussels is critical for propagation andculture success.

Due to their small size (<10–20 mm), juvenile mussels aredifficult to detect in the wild, restricting field investigations to

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3


adult life stages and making it difficult to examine effects oftemperature and other factors on early life stages (Negishi andKayaba 2010). Though growth rates of juveniles from field-based studies are uncommon, researchers have begun to closethis knowledge gap by utilizing laboratory-propagated juvenilesfor experimental studies, as we have done. To our knowledge, noother studies have been published that directly tested the effectsof temperature on the growth and survival of older and larger (>4months, ≥1.5 mm) laboratory-propagated endangered juvenileswith the goal of determining an optimal rearing temperature formaximizing culture success.

We found that temperature had a positive correlation withgrowth of Cumberlandian combshell, oyster mussel, and wavy-rayed lampmussel juveniles, which agreed with conclusionsfrom previous studies regarding the effect of temperature onjuvenile mussel growth (Buddensiek 1995; Beaty 1997; Hanlon2000; Hanlon and Neves 2006). Further, the positive relation-ship between temperature and growth and the magnitude ofgrowth varied between juveniles of these three species. Theseobserved differences in juvenile mussel growth demonstratedthat growth among these species varies in relation to water tem-perature. In contrast to previous studies, temperature was neitherpositively nor negatively associated with survival (Buddensiek1995; Beaty 1997). The relationship between temperature andsurvival of juveniles was less clear within the time-scale andtemperature treatments of this study. Even though survival didnot statistically differ over time between treatments for all threespecies of juveniles, a few significant treatment comparisonsbetween final survivals (at sampling event 10) were detected.Generally, it appeared that lowest survival occurred at 20◦C,although some pairwise comparisons were not significant.

Prior to our study, we set a biologically important effect sizefor final growth between temperature treatments at 1 mm. Formonitoring release and population success, juveniles are indi-vidually tagged in the laboratory with a Hallprint shellfish tagbefore being released into the wild. This tagging procedure re-quires that individuals be a minimum size of 10 mm becauseof the size of the tags (8 × 4 mm oval tag size). Thus, a dif-ference in 1 mm between individuals can influence how soonjuveniles can be tagged in the laboratory and then released tosites selected for population restoration. In addition to size influ-encing when juveniles can be released, survival of overwinteringjuveniles may be directly correlated with size, significantly im-proving the survival of individuals when released to the wild(Buddensiek 1995; Hanlon 2000; Sarrazin and Legendre 2000;Hanlon and Neves 2006; Hua et al. 2011). Greater size willalso increase detection probability during monitoring efforts ofreleased individuals and enhance the overall likelihood of pop-ulation recovery success (Hua et al. 2011).

One goal of this study was to determine the optimum tem-perature for maximizing growth of juveniles in captivity. Wefound that maximum growth in shell length after approximately4.5 months for Cumberlandian combshell, oyster mussel, andwavy-rayed lampmussel juveniles occurred at 26◦C. However,

growth at 26◦C did not differ statistically nor biologically (dif-ference < 1.0 mm) from growth at 28◦C. Therefore, differencesin final survivals within species were assessed to make evalua-tions between these two temperatures.

While Cumberlandian combshell and wavy-rayed lampmus-sel juveniles experienced highest final survival at 28◦C, oystermussel juveniles had the lowest survival at this temperaturetreatment. It is not clear whether high mortality at 28◦C wasdue to approaching an upper thermal limit for oyster musseljuveniles, sampling stress, or factors other than temperature.Sampling procedure involves handling juveniles to obtain shellmeasurements and to estimate survival data; this requires short-term exposure to air, which can cause stress (Liberty et al. 2007).Several studies have reported lower mortality in juveniles thatwere sampled less frequently (Beaty 1997; Zimmerman 2003;Liberty et al. 2007). Considering that differences in survival ofjuveniles over time were not significant for the three speciesin our study, perhaps sampling frequency or other factors con-tributed to final mortality rather than temperature alone. Withno statistical or biological difference detected between the 26◦Cand 28◦C in growth and survival within species, and due tothe unknown source of additional mortality at 28◦C for oys-ter mussel juveniles, we incorporated conclusions of previousstudies on water temperature relationships into our assessmentof optimum rearing temperature for these species.

Water temperature is one of the most important environmen-tal variables affecting growth and survival of juvenile musselsin captivity (Zimmerman 2003; Jones et al. 2005; Pandolfo et al.2010a, 2010b). Several laboratory experiments have describedthe effects of temperature on growth and survival of freshwaterbivalves during early life stages (i.e., newly transformed juve-niles and <1-year-old juveniles; Buddensiek 1995; Beaty 1997;Hanlon 2000; Zimmerman 2003; Hanlon and Neves 2006; Pan-dolfo et al. 2010a, 2010b). Buddensiek (1995) found that growthrates and mortality of juvenile eastern pearlshell Margaritiferamargaritifera were positively correlated with temperature. Simi-larly, Beaty (1997) reported a positive relationship between tem-perature and growth and survival of newly transformed rainbowVillosa iris. Hanlon (2000) also reported a positive relation-ship between temperature and growth in juvenile wavy-rayedlampmussels but showed seasonal variation in survival to sug-gest temperature is negatively associated with mortality. Hanlon(2000) further suggested that the relationship between temper-ature and survival is not always clear and that opposing studyresults may be due to resource availability at different experi-mental scales (i.e., streams are less likely to be food-limiting athigher temperatures than in a laboratory-scale experiment).

Two other studies examined temperature effects on survivalduring early life stages and determined acute lethal temperatures(LT50s) for glochidia and laboratory-propagated juveniles. Theaims of these studies were to determine upper thermal limits ofearly life stages to provide insight into any effects that risingmaximum water temperatures—due to global climate change—may have on mussel populations. Pandolfo et al. (2010b)

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

540 CAREY ET AL.

reported acute lethal thermal tolerances for glochidia of eightspecies and juveniles of seven species, ranging in age from <1–8 weeks old. They reported that mean LT50s in 96-h tests were34.7◦C for juveniles, and 31.6◦C in 24-h tests for glochidia.Pandolfo et al. (2010b) concluded that the survival of theseearly life stages can decline significantly with small increasesin temperature. Dimock and Wright (1993) also reported acutethermal tolerances for 1-week old juveniles of two freshwatermussel species and reported LT50s between 31.5◦C and 33◦C.Because our study goal was to determine optimum productiontemperatures, our experiment did not cover the upper tempera-ture ranges (i.e., >30◦C) considered in these studies, suggest-ing why we likely did not observe a clear relationship betweentemperature and survival for Cumberlandian combshell, oystermussel, and wavy-rayed lampmussel juveniles.

Temperature has a significant effect on aquatic organismgrowth and survival rates in hatchery settings due to its in-fluence on physiological processes such as respiration, filtra-tion, and excretion rates (Zimmerman and Neves 2002; Spooner2007; Spooner and Vaughn 2008; Pandolfo et al. 2010a, 2010b;Fitzgibbon and Battaglene 2012). These metabolic activities ofmussels generally increase with higher water temperatures, i.e.,within the natural range (Hanlon 2000; Spooner and Vaughn2008; Vaughn et al. 2008). Typically, oxygen and food resourcescan become limiting with increasing water temperature (Han-lon 2000). The availability of dissolved oxygen in a systemis negatively related to temperature and dependent on the wa-ter system (Hastie et al. 2003). The availability of food in aclosed system is limited by the amount of supplemental dietdispensed to individuals exhibiting higher feeding rates in sys-tems cultured at higher temperatures. Therefore, a combinationof increased dissolved oxygen demand and feeding rates withlower availability of these resources at higher temperatures canstrongly influence growth and survival. In addition, total ammo-nia concentrations have been shown to increase with increasedexcretion rates of mussels due to higher temperatures (Spoonerand Vaughn 2008). Although ammonia toxicity (total ammo-nia) from increased water temperatures is negligible between3◦C and 30◦C for fish in freshwater systems, early life stages ofmussels are more sensitive to total ammonia concentrations thanare other aquatic organisms (USEPA 1998, cited by Randall andTsui 2002; Wang et al. 2007a, 2007b).

In healthy non-degraded streams, juveniles generally do notface issues with food and oxygen availability or ammonia tox-icity because of the continuous influx of freshwater and highturnover rate. Conversely, experiments that are confined to smallrecirculating aquaculture systems, compared with streams, havea higher likelihood of encountering (if not managed properly)limited food and oxygen or increased ammonia levels at highertemperatures because of their lack of a continuous influx offreshwater (Hanlon 2000). As a consequence, juveniles may ex-perience increased levels of mortality. Because of the possibleoccurrence of food and oxygen limitations and sublethal am-monia levels in small recirculating systems, researchers have

been cautious about culturing juveniles at higher temperatures.These general temperature relationships were taken into consid-eration, even though food quantity was not a limiting factor inour experiment, and our experimental culture systems did notexperience any abnormal dissolved oxygen or total ammonialevels.

Based on our analyses of final growth and survival and pre-viously described temperature relationships, we believe that theoptimal rearing temperature for maximum growth and survivalin captivity is around 26◦C for Cumberlandian combshell, oys-ter mussel, and wavy-rayed lampmussel juveniles. We believeour findings can be applied by researchers to improve labora-tory culture methods for juveniles of other species of mussels.Present culture temperatures for juvenile mussels are set basedon research manager discretion and source water temperatures,and sometimes overwintering juveniles in captivity are held be-low growing temperatures (i.e., <15◦C). Researchers also havebeen cautious about culturing juveniles, particularly those of en-dangered species, at temperatures consistently exceeding 24◦Cbecause of concern of increased mortality. Results suggestedthat a simulated winter season is not necessary for continuedmussel growth or survival. However, because some biologistsbelieve laboratory-propagated mussels need to experience loweroverwintering temperatures to be better adapted to natural con-ditions upon release, further investigation is needed to determinewhether long-term survival after release is affected by the ab-sence of a cooling-off period in captivity.

Determination of an optimum rearing temperature has clearimplications for culturing of laboratory-propagated juveniles,and ultimately for conservation efforts. The culture and releaseof laboratory-propagated juveniles has been identified by fed-eral species recovery plans and other documents as an approachto increasing the viability of existing populations or reintroduc-ing species within their historical ranges (Williams et al. 1993;Neves et al. 1997; Neves 1999; USFWS 2003, 2004). Optimiz-ing temperature to maximize growth and survival of mussels inhatchery settings reduces the length of time juveniles are held inthe laboratory, allowing biologists to grow endangered juvenilesto larger sizes more quickly and maximizing production levelsrelative to costs. Decreasing holding time is important because itreduces mortality in captivity (i.e., subjects them to less handlingstress) and frees up space in hatcheries, thereby increasing theoverall number of individuals produced for population recoveryefforts by resource managers.

Understanding the relationship between temperature andmussel growth and survival across all life stages is importantfor optimizing propagation and culture success and, by exten-sion, recovery of imperiled species. Our findings support previ-ous conclusions that higher temperatures increase growth ratesbut neither supported nor contradicted conclusions on the rela-tionship between temperature and survival. Upper thermal lim-its (i.e., >50% mortality over the duration of this experiment)were not observed for juveniles of the three species in our study.This experiment should be repeated with newly transformed

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3


juveniles to determine if temperature affects growth and survivaldifferently for younger and smaller juveniles. Furthermore, ad-ditional testing of growth and survival of juveniles within thesetemperature ranges (20–28◦C) over a larger temporal scale andat higher temperature ranges (>28◦C), is needed to reveal a clearrelationship between temperature and survival and to understandand predict the potential effects of persistent high water tem-peratures on mussel populations due to global climate change(Hastie et al. 2003; Pandolfo et al. 2010b).

ACKNOWLEDGMENTSThis project was funded by the U.S. Fish and Wildlife Service

(USFWS), Asheville, North Carolina, Field Office. We thankVirginia Department of Game and Inland Fisheries’ (VDGIF)Aquatic Wildlife Conservation Center and the Freshwater Mol-lusk Conservation Center for providing the juvenile mussels forthis study. In addition, we thank the many people who helped uscomplete the field and laboratory work for the project: AmandaDuncan, Mike Pinder, and Joe Ferraro, VDGIF; Gale Heffinger,USFWS; and Kasey Ewing, Tim Lane, Jennifer Rogers, An-drew Phipps, Hua Dan, Lee Stephens, Brian Parks, and DanielSchilling, Virginia Polytechnic Institute and State University.The views expressed in this article are the authors and do notnecessarily represent those of the USFWS.

REFERENCESBarnhart, M. C. 2006. Buckets of muckets: a compact system for rearing juvenile

freshwater mussels. Aquaculture 254:227–233.Beaty, B. B. 1997. Development of juvenile culture techniques and testing of

potential biomarkers of environmental stress in freshwater mussels (Bivalvia:Unionidae). Doctoral dissertation. Virginia Polytechnic Institute and StateUniversity, Blacksburg.

Beaty, B. B., and R. J. Neves. 2004. Use of a natural river water flow-throughculture system for rearing juvenile freshwater mussels (Bivalvia: Unionidae)and evaluation of the effects of substrate size, temperature, and stockingdensity. American Malacological Bulletin 19:15–23.

Buddensiek, V. 1995. The culture of juvenile freshwater pearl mussels Margar-itifera margaritifera L. in cages: a contribution to conservation programmesand the knowledge of habitat requirements. Biological Conservation 74:33–40.

Dimock, R. V., Jr., and A. H. Wright. 1993. Sensitivity of juvenile freshwatermussels to hypoxic, thermal and acid stress. Journal of the Elisha MitchellScientific Society 109:183–192.

Eckert, N. L., and M. J. Pinder. 2010. Freshwater mussel survey of ClevelandIsland, Clinch River, Virginia: augmentation monitoring site: 2008. FinalReport to the Virginia Department of Game and Inland Fisheries, Bureau ofWildlife Resources, Richmond.

Fitzgibbon, Q. P., and S. C. Battaglene. 2012. Effect of water temperature on thedevelopment and energetics of early, mid and late-stage phyllosoma larvae ofspiny lobster Sagmariasus verreauxi. Aquaculture 344/349:153–160.

Gosling, E. 2003. Bivalve molluscs: biology, ecology and culture. Fishing NewsBooks, Oxford, UK.

Hanlon, S. D. 2000. Release of juvenile mussels into a fish hatchery raceway: acomparison of techniques. Master’s thesis. Virginia Polytechnic Institute andState University, Blacksburg.

Hanlon, S. D., and R. J. Neves. 2006. Seasonal growth and mortality of juve-niles of Lampsilis fasciola (Bivalvia: Unionidae) released to a fish hatcheryraceway. American Malacological Bulletin 21:45–49.

Hanson, J. M., W. C. Mackay, and E. E. Prepas. 1988. The effects of water depthand density on the growth of a unionid clam. Freshwater Biology 19:345–355.

Hastie, L. C., P. J. Cosgrove, N. Ellis, and M. J. Gaywood. 2003. The threat ofclimate change to freshwater pearl mussel populations. Ambio 32:40–46.

Hastie, L. C., M. R. Young, and P. J. Boon. 2000. Growth characteristics offreshwater pearl mussels, Margaritifera margaritifera (L.). Freshwater Biol-ogy 43:243–256.

Henley, W. F., L. L. Zimmerman, R. J. Neves, and M. R. Kidd. 2001. Designand evaluation of recirculating water systems for maintenance and propa-gation of freshwater mussels. North American Journal of Aquaculture 63:144–155.

Hinch, S. G., R. C. Bailey, and R. H. Green. 1986. Growth of Lampsilis radiata(Bivalvia: Unionidae) in sand and mud: a reciprocal transplant experiment.Canadian Journal of Fisheries and Aquatic Sciences 43:548–552.

Hua, D., J. Rogers, J. Jones, and R. Neves. 2011. Propagation, culture, andmonitoring of endangered mussels for population restoration in the Clinchand Powell rivers, Tennessee, 2006–2010. Final Report to the TennesseeWildlife Resources Agency, Nashville.

Jones, J. W., E. M. Hallerman, and R. J. Neves. 2006. Genetic managementguidelines for captive propagation of freshwater mussels (Unionoidea). Jour-nal of Shellfish Research 25:527–535.

Jones, J. W., R. A. Mair, and R. J. Neves. 2005. Factors affecting survival andgrowth of juvenile freshwater mussels cultured in recirculating aquaculturesystems. North American Journal of Aquaculture 67:210–220.

Krebs, C. J. 1972. Ecology: the experimental analysis of distribution and abun-dance. Harper and Row, New York.

Liberty, A. J. 2004. An evaluation of the survival and growth of juvenile and adultfreshwater mussels at the Aquatic Wildlife Conservation Center (AWCC),Marion, Virginia. Master’s thesis. Virginia Polytechnic Institute and StateUniversity, Blacksburg.

Liberty, A. J., B. J. Ostby, and R. J. Neves. 2007. Determining a suitablesubstrate size and sampling frequency for rearing juvenile rainbow musselsVillosa iris. North American Journal of Aquaculture 69:44–52.

Negishi, J. N., and Y. Kayaba. 2009. Effects of handling and density on thegrowth of the unionoid mussel Pronodularia japanensis. Journal of the NorthAmerican Benthological Society 28:821–831.

Negishi, J. N., and Y. Kayaba. 2010. Size-specific growth patterns and estimatedlongevity of the unionid mussel (Pronodularia japanensis). Ecological Re-search 25:403–411.

Neves, R. J. 1999. Conservation and commerce: management of freshwatermussel (Bivalvia: Unionoidea) resources in the United States. Malacologia41:461–474.

Neves, R. J. 2004. Propagation of endangered freshwater mussels in NorthAmerica. Pages 69–80 in I. J. Killeen and M. B. Seddon, editors. Molluscanbiodiversity and conservation. Conchological Society of Great Britain andIreland, Special Publication 3, Reading, UK.

Neves, R. J., A. E. Bogan, J. D. Williams, S. A. Ahlstedt, and P. W. Hartfield.1997. Status of aquatic mollusks in the southeastern United States: a down-ward spiral of diversity. Pages 43–85 in G. W. Benz and D. E. Collins, editors.Aquatic fauna in peril: the southeastern perspective. Southeast Aquatic Re-search Institute, Lenz Design and Communications, Special Publication 1,Decatur, Georgia.

Pandolfo, T. J., W. G. Cope, and C. Arellano. 2010a. Thermal tolerance ofjuvenile freshwater mussels (Unionidae) under the added stress of copper.Environmental Toxicology and Chemistry 29:691–699.

Pandolfo, T. J., W. G. Cope, C. Arellano, R. B. Bringolf, M. C. Barnhart,and E. Hammer. 2010b. Upper thermal tolerances of early life stages offreshwater mussels. Journal of the North American Benthological Society 29:959–969.

Randall, D. J., and T. K. N. Tsui. 2002. Ammonia toxicity in fish. MarinePollution Bulletin 45:17–23.

Rypel, A. L., W. R. Haag, and R. H. Findlay. 2008. Validation of annual growthrings in freshwater mussel shells using cross dating. Canadian Journal ofFisheries and Aquatic Sciences 65:2224–2232.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3

542 CAREY ET AL.

Sarrazin, F., and S. Legendre. 2000. Demographic approach to releasing adultsversus young in reintroductions. Conservation Biology 14:488–500.

Spooner, D. E. 2007. An integrative approach to understanding mussel com-munity structure: linking biodiversity, environmental context and physiology.Doctoral dissertation. University of Oklahoma, Norman.

Spooner, D. E., and C. C. Vaughn. 2008. A trait-based approach to species’ rolesin stream ecosystems: climate change, community structure, and materialcycling. Oecologia 158:307–317.

USEPA (U.S. Environmental Protection Agency). 1998. Addendum to “ambientwater quality criteria for ammonia – 1984.” USEPA, National TechnicalInformation Service, Springfield, Virginia.

USFWS (U.S. Fish and Wildlife Service). 2003. Endangered and threatenedwildlife and plants; proposed designation of critical habitat for five endan-gered mussels in the Tennessee and Cumberland river basins. Federal Register68:106(3 June 2003):33234–33282.

USFWS (U.S. Fish and Wildlife Service). 2004. Recovery plan for Cumber-land elktoe (Alasmidonta atropurpurea), oyster mussel (Epioblasma capsae-formis), Cumberlandian combshell (Epioblasma brevidens), purple bean (Vil-losa perpurpurea), and rough rabbitsfoot (Quadrula cylindrica strigillata).USFWS, Southeast Region, Atlanta.

Vaughn, C. C., S. J. Nichols, and D. E. Spooner. 2008. Community and foodwebecology of freshwater mussels. Journal of the North American BenthologicalSociety 27:409–423.

Wang, N., C. G. Ingersoll, I. E. Greer, D. K. Hardesty, C. D. Ivey, J. L. Kunz, W.G. Brumbaugh, F. J. Dwyer, A. D. Roberts, T. Augspurger, C. M. Kane, R. J.

Neves, and M. C. Barnhart. 2007a. Chronic toxicity of copper and ammoniato juvenile freshwater mussels (Unionidae). Environmental Toxicology andChemistry 26:2048–2056.

Wang, N., C. G. Ingersoll, D. K. Hardesty, C. D. Ivey, J. L. Kunz, T.W. May, F. J. Dwyer, A. D. Roberts, T. Augspurger, C. M. Kane,R. J. Neves, and M. C. Barnhart. 2007b. Acute toxicity of copper,ammonia, and chlorine to glochidia and juveniles of freshwater mus-sels (Unionidae). Environmental Toxicology and Chemistry 26:2036–2047.

Williams, J. D., M. L. Warren Jr., K. S. Cummings, J. L. Harris, and R. J. Neves.1993. Conservation status of freshwater mussels of the United States andCanada. Fisheries 18(9):6–22.

Yeager, B. L., and C. F. Saylor. 1995. Fish hosts for four species of freshwatermussels (Pelecypoda: Unionidae) in the upper Tennessee River drainage.American Midland Naturalist 133:1–6.

Zale, A. V., and R. J. Neves. 1982. Fish hosts of four species of lampsilinemussels (Mollusca: Unionidae) in Big Moccasin Creek, Virginia. CanadianJournal of Zoology 60:2535–2542.

Zimmerman, L. L. 2003. Propagation of juvenile freshwater mussels (Bivalvia:Unionidae) and assessment of habitat suitability for restoration of mussels inthe Clinch River, Virginia. Master’s thesis. Virginia Polytechnic Institute andState University, Blacksburg.

Zimmerman, L. L., and R. J. Neves. 2002. Effects of temperature on duration ofviability for glochidia of freshwater mussels (Bivalvia: Unionidae). AmericanMalacological Bulletin 17:31–35.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:39

27

Oct

ober

201

3



Performance of a Recirculating Aquaculture SystemUtilizing an Algal Turf Scrubber for Scaled-Up CaptiveRearing of Freshwater Mussels (Bivalvia: Unionidae)Zhitao Huang a b , Jess Jones c , Junye Gu b , Eric Hallerman b , Timothy Lane b , Xiefa Song a

& Rong Wan aa Department of Fisheries , Ocean University of China , Qingdao , 266003 , Chinab Freshwater Mollusk Conservation Center, Department of Fish and Wildlife Conservation ,Virginia Polytechnic Institute and State University , 100 Cheatham Hall, 310 West CampusDrive, Blacksburg , Virginia , 24061 , USAc U.S. Fish and Wildlife Service, Department of Fish and Wildlife Conservation , VirginiaPolytechnic Institute and State University , 100 Cheatham Hall, 310 West Campus Drive,Blacksburg , Virginia , 24061 , USA

To cite this article: Zhitao Huang , Jess Jones , Junye Gu , Eric Hallerman , Timothy Lane , Xiefa Song & Rong Wan (2013)Performance of a Recirculating Aquaculture System Utilizing an Algal Turf Scrubber for Scaled-Up Captive Rearing ofFreshwater Mussels (Bivalvia: Unionidae), North American Journal of Aquaculture, 75:4, 543-547






http://dx.doi.org/10.1080/15222055.2013.826762




NOTE

Performance of a Recirculating Aquaculture SystemUtilizing an Algal Turf Scrubber for Scaled-Up CaptiveRearing of Freshwater Mussels (Bivalvia: Unionidae)

Zhitao HuangDepartment of Fisheries, Ocean University of China, Qingdao 266003, China;and Freshwater Mollusk Conservation Center, Department of Fish and Wildlife Conservation,Virginia Polytechnic Institute and State University, 100 Cheatham Hall, 310 West Campus Drive,Blacksburg, Virginia 24061, USA

Jess JonesU.S. Fish and Wildlife Service, Department of Fish and Wildlife Conservation,Virginia Polytechnic Institute and State University, 100 Cheatham Hall, 310 West Campus Drive,Blacksburg, Virginia 24061, USA

Junye Gu, Eric Hallerman,* and Timothy LaneFreshwater Mollusk Conservation Center, Department of Fish and Wildlife Conservation,Virginia Polytechnic Institute and State University, 100 Cheatham Hall, 310 West Campus Drive,Blacksburg, Virginia 24061, USA

Xiefa Song and Rong WanDepartment of Fisheries, Ocean University of China, Qingdao 266003, China

AbstractTo develop a system supporting rapid growth of juvenile fresh-

water mussels, a recirculating aquaculture system was designedand built at the Laboratory for Conservation Aquaculture andAquatic Ecology, Virginia Polytechnic Institute and State Uni-versity. The system included a rectangular trough tank, a sump(biofilter), a pump, a microalgae drip feeder, and an air deliverysystem. An algal turf scrubber (ATS) was evaluated for its potentialto maintain and improve water quality within the system. Growthand survival rates of juvenile rainbow mussels Villosa iris after 90 dwere compared between system units with and without ATSs. Flowrate through the culture units was approximately 23.3 L/min. Re-sults showed no statistically significant differences between thegrowth and survival rates of juvenile mussels reared in systemswith ATSs and those reared in systems without ATSs. Ammoniaand nitrite levels were low and did not differ among treatments.However, systems with ATSs exhibited significantly lower levelsof nitrate and phosphate than systems without ATSs. Our resultsshow that freshwater mussel culture systems can be scaled up toincrease production and that the use of ATSs may help to maintain

*Corresponding author: [email protected] February 4, 2013; accepted July 12, 2013

water quality in recirculating aquaculture systems during long-term culture of freshwater mussels.

North America contains the greatest diversity of freshwatermussels in the world—nearly 300 species (Williams et al. 1993;Neves 1999). However, about 70% (213) of the North Ameri-can species are listed as endangered, threatened, or of specialconcern, and nearly 35 of these species are considered extinct(Williams et al. 1993; Neves 1999). Causes of population de-cline include habitat loss and destruction from impoundmentof rivers, excessive sedimentation, water pollution, dredging,and other anthropogenic factors that affect the natural structureand function of free-flowing rivers (Neves et al. 1997; Neves1999; Jones et al. 2005). Conservation of freshwater musselshas become a priority in the United States, and conservationmeasures include the propagation and culture of endangeredmussel species in order to augment existing populations and

543

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3

544 HUANG ET AL.

reintroduce mussels into historical sites of occurrence (Joneset al. 2005).

Approaches for improving the survival and growth of cul-tured juvenile and adult mussels have included rearing in cages,ponds, raceways, and tanks (Gatenby et al. 1996; Dunn andLayzer 1997; Farris et al. 1999; Hanlon and Neves 2006). Morerecently, recirculating aquaculture systems (RASs) have beenused to rear freshwater mussels (O’Beirn et al. 1998; Layzeret al. 1999; Henley et al. 2001; Kovitvadhi et al. 2006, 2008).Culture units for producing freshwater mussels tend to be small(on the order of 30 × 30 cm, with a volume of a few liters), thuslimiting the number of mussels that can be produced. Up-scalingof the culture vessels could allow for increased mussel produc-tion and the establishment of flow regimes that mimic those ofrivers, presenting advantages for providing food to filter-feedingmussels. However, scaling up may also lead to water quality is-sues (e.g., ammonia or nitrite accumulation) or other technicalproblems. In this study, we designed and evaluated a relativelylarge RAS to culture freshwater mussels and we assessed theutility of an algal turf scrubber (ATS) to help maintain waterquality in the RAS. An ATS utilizes filamentous algae to takeup excess nutrients, such as nitrate and phosphate, which tend toaccumulate in aquatic systems (Adey et al. 1993, 1996). As thealgae grow, they assimilate nutrients, such as inorganic nitrate,inorganic phosphate, nitrite, ammonia, and ammonium, therebyimproving water quality (Veraart et al. 2008). Thus, the purposeof our study was to evaluate water quality in scaled-up RASswith and without ATSs and to assess the survival and growthrates of freshwater mussels reared in these systems.

METHODSConstruction of recirculating aquaculture systems and algal

turf scrubbers.—Recirculating aquaculture systems for rearingfreshwater mussels were developed at the Laboratory for Con-servation Aquaculture and Aquatic Ecology, Virginia Polytech-nic Institute and State University (Virginia Tech). The RASdesign (Figure 1) included (1) a plastic stock-watering troughthat was utilized as the container for substrate and culturedmussels, (2) a sump that also served as a biofilter, (3) a pump,

FIGURE 1. Schematic diagram of the recirculating aquaculture system usedin the study of juvenile freshwater mussel growth and survival.

(4) a microalgae drip feeder (1-L volume), (5) an air deliverysystem, and (6) an ATS. The mussel culture trough was madeof polyethylene and was 300 cm long, 68 cm wide, and 27 cmdeep along the midline; the trough held 330–373 L of waterat a depth of 16–18 cm. A magnetically driven pump (ModelNH-100PX-X; Pan World Co., Ltd.) generated water flow inthe RAS. The tank water volume was exchanged approximatelyfour times per hour (once every 15 min) via a total system flowof 23.3 L/min. Water velocity at the surface along the centerline of the trough was 0.77 m/s. Fine sand (<2 mm in diameter)and limestone gravel (<4 mm in diameter) were mixed andused as substrate for the mussels; substrate was placed evenlythroughout the trough to a depth of 4–5 cm. Water was recircu-lated through the trough and sump by using a 3.08-cm polyvinylchloride (PVC) pipe (1-in schedule-40 PVC) and other plastictubing. Plastic biomedia (Dynamic Aqua Science, Inc., LagunaBeach, California) was added into the sump tank so that it wouldfunction as a biofilter. The nitrification function of the biofilterwas not established before the experiment. The ATS was madefrom plastic mesh (60 cm long × 60 cm wide; mesh size =1.3 × 1.3 cm) and received recirculated water through a bypasspipe (Figure 2). During the experiment, six RASs were used:three with ATSs (treatment) and three without ATSs (control).Aeration in the system was provided by a Sweetwater regener-ative blower (Aquatic Eco-Systems, Inc., Apopka, Florida) andwas delivered through PVC pipes, flexible tubing, and an airdiffuser. All systems were located in a greenhouse and receivednatural light—no shade cloth was used. A 60-W lamp was usedfor nighttime illumination of systems with ATSs.

Food and feeding.—Mussels were fed a commercial algalmix (1:3 ratio) consisting of Nanno 3600 (Nannochloropsis) ata concentration of 0.02 mL/L and Shellfish Diet 1800 (Isochrysissp., Pavlova sp., Thalossiosira weissflogii, and Tetraselmis sp.;Reed Mariculture, Inc., Campbell, California) at 0.007 mL/L.The feed densities were approximately 136,000 cells/mL forNanno 3600 and 14,000 cells/mL for Shellfish Diet 1800. Thealgal diet was delivered into the system over each 24-h periodby using a 1-L drip bottle mounted over the sump. Fresh algalmix was placed in each drip bottle daily at 0900 hours. EachRAS contained a 1:1 mix of pond water and well water, 50% ofwhich was replaced each week.

Experimental design and analyses.—In total, six RASs wereused (three with ATSs and three without ATSs). Juvenile rain-bow mussels Villosa iris (∼8 months old; average shell length =17.3 mm) were reared for 13 weeks during the experiment. Mus-sels were produced at the Freshwater Mollusk ConservationCenter, Virginia Tech. Three-thousand mussels were randomlyassigned to troughs (500 mussels/trough); initial stocking den-sity was 245 mussels/m2. Thirty of the mussels in each troughwere tagged (Hallprint, Ltd., Hindmarsh Valley, South Aus-tralia) on the shell surface. Tagged mussels were sampled andmeasured for length once per week to monitor growth.

Water quality.—Data on ammonia, nitrite, nitrate, phosphate,conductivity, salinity, temperature, pH, and dissolved oxygen

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3

NOTE 545

FIGURE 2. Algal turf scrubber with growing algae; the turf scrubber was evaluated for its ability to maintain water quality in recirculating aquaculture systemsused for rearing juvenile freshwater mussels. [Figure available in color online.]

were collected from each RAS every other day. Ammonia-nitrogen, nitrite-nitrogen, nitrate-nitrogen, and phosphate weredetermined using a Hach DR2400 spectrophotometer (HachCompany, Loveland, Colorado). Temperature and dissolvedoxygen were measured with a YSI Model 550A dissolved oxy-gen meter (YSI, Inc., Yellow Springs, Ohio). Conductivity wasdetermined by use of a YSI Professional Plus conductivity meter.Salinity was measured with a salinometer (Model TDS-4TM;HM Digital, Inc., Korea), and pH was determined with a pHmeter (Thermo Electron Corp., Waltham, Massachusetts).

Data analyses.—Statistical analyses were performed usingJMP version 9 for Windows. Mussel growth was analyzed withrepeated-measures multivariate ANOVA (MANOVA) at a sig-nificance level α of 0.05. Survival rate of the mussels and waterquality in the RASs were analyzed using one-way ANOVA.

RESULTS AND DISCUSSIONScaling up of freshwater mussel production units can increase

the mussel numbers and biomass produced. In the present study,we designed and demonstrated the suitability of an RAS forthe grow-out of freshwater mussels, and we obtained excellentsurvival and growth of juvenile rainbow mussels. Furthermore,suitable water quality was maintained in our RASs, especiallyin systems that were equipped with ATSs.

Water QualityTemperature, dissolved oxygen, pH, ammonia, and nitrite

did not differ significantly (P > 0.05) between RASs with ATSsand those without ATSs (Table 1). At the laboratory site, highpH (∼8.0) and conductivity (∼420 µS/cm) are characteristic

of the well water, which is drawn from a karst aquifer. Afterthe first 3 weeks of the study, ammonia concentrations in bothtypes of system were less than 0.04 mg/L (Figure 3), whichis considered safe for freshwater mussels (Layzer et al. 1999).It took approximately 25–33 d for the biofilters and ATSs tobecome biologically functional—that is, fully capable of pro-cessing ammonia, nitrite, and nitrate. For example, ammoniaand nitrite levels were maintained below 0.05 and 0.01 mg/Lafter 1 month in RASs with ATSs and in those without ATSs,respectively (Figure 3). Therefore, our data demonstrate thatthe RAS and biofilter together were sufficient for nitrification ofammonia and nitrite to nitrate. However, the RASs with ATSswere much better at eliminating nitrate, with concentrations

TABLE 1. Water quality in recirculating aquaculture systems (RASs) withalgal turf scrubbers (ATSs) and in those without ATSs over the 13-week studyperiod (mean ± SE). Within a row, values with different lowercase letters aresignificantly different (P < 0.05).

RASs with RASs withoutVariable ATSs ATSs

Temperature (◦C) 22.1 ± 1.60 z 22.6 ± 1.24 zDO (mg/L) 8.40 ± 1.61 z 8.51 ± 1.74 zpH 8.72 ± 0.30 z 8.69 ± 0.31 zConductivity (µS/cm) 437.1 ± 69.1 z 403.1 ± 45.8 yAmmonia (mg/L) 0.034 ± 0.02 z 0.029 ± 0.03 zNitrite (mg/L) 0.028 ± 0.06 z 0.056 ± 0.112 zNitrate (mg/L) 0.46 ± 0.28 z 1.82 ± 1.07 yPhosphate (mg/L) 0.71 ± 0.32 z 1.2 ± 0.54 y

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3

546 HUANG ET AL.

0

0.05

0.1

0.15

0.2

0 10 20 30 40 50 60 70 80 90

RAS withATS

RAS withoutATS

Amm

onia

(mg/

l)

Time(days)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40 50 60 70 80 90

Time(days)

Nitr

ite (m

g/l)

00.5

11.5

22.5

33.5

44.5

0 10 20 30 40 50 60 70 80 90

Time(days)

Nitr

ate

(mg/

l)

0

0.5

1

1.5

2

2.5

0 10 20 30 40 50 60 70 80 90

Time(days)

Phos

phat

e (m

g/l)

FIGURE 3. Concentrations of ammonia, nitrite, nitrate, and phosphate overthe 13-week study period in recirculating aquaculture systems (RASs) equippedwith algal turf scrubbers (ATSs) and in those without ATSs.

remaining less than 1 mg/L over the course of the study; incontrast, nitrate increased from 0.2 to 4 mg/L in RASs withoutATSs (Figure 3). From day 26 to the end of the experiment, ni-trate was significantly greater (P < 0.05) in RASs without ATSs,

whereas nitrate in the RASs with ATSs did not increase. Fur-ther, phosphate was also significantly lower (P < 0.05) in RASswith ATSs, increasing from 0.2 to 1.09 mg/L; in RASs withoutATSs, phosphate increased from 0.19 to 1.97 mg/L. Thus, theATSs effectively utilized filamentous algae (including speciesof Eunotia and Melosira) to absorb both nitrate and phosphateas nutrient sources.

Our results are broadly convergent with those of earlier stud-ies. Algal turf scrubbers have been used successfully to treatmultiple types of pollution, including agriculture runoff, excessnutrient accumulation in lakes, and manure effluents (Adey et al.1993, 1996; Craggs et al. 1996; Mulbry et al. 2008). This wasaccomplished simply by allowing native algae to grow attachedto a screen in a shallow, flowing-water system and then regularlycropping the algae from the screens to permanently remove se-questered nutrients and promote continued algal growth (Adeyet al. 1993).

Results from our study showed that water quality variableswere maintained below known effect levels for mussels. Ata temperature of 25◦C and a pH of 8, the acute and chroniccriteria for total ammonia nitrogen concentration are 2.9 and0.26 mg/L, respectively, for freshwater mussels (USEPA 2009).In a study of juvenile fatmucket mussels Lampsilis siliquoidea,Myers-Kinzie (1998) reported 48-h LC50 values (concentrationlethal to 50% of test organisms) of 0.09 mg/L for ammonia and0.19 mg/L for nitrite. We found no references on the toxicity ofnitrate to adult or juvenile freshwater mussels, but MacMillanet al. (1994) reported that for marine bivalves, nitrite should notexceed 0.01 mg/L and nitrate should not exceed 19.16 mg/L.

The time required to turn over the entire water volume in aculture trough was 15 min given an approximate flow rate of23.3 L/min for the trough. Although flow requirements for fresh-water mussels are likely quite variable among species, manythreatened and endangered freshwater mussels are found in rif-fle habitat, where water velocities are high. Higher flow veloc-ities are likely required for many mussel species, emphasizingthe need for further research on flow requirements for culturedfreshwater mussels.

Growth and SurvivalAt the conclusion of the study, mean mussel length ( ± SE)

was 22.5 ± 0.46 mm for RASs with ATSs and 20.8 ± 0.29 mmfor RASs without ATSs. The growth data were analyzed usingMANOVA; the results indicated a significant time × treatmentinteraction effect (P = 0.001) on the mean length of rainbowmussels. Time refers to the time of culture in the RASs; treatmentrefers to the RASs with and without ATSs. Mean growth of mus-sels in the two treatments overlapped for much of the experimentbut began to diverge after week 10 (Figure 4), when growth be-came faster in the RASs with ATSs than in those without ATSs.This divergence in juvenile mussel growth may indicate (1) achronic impact of excess nitrate or phosphate, despite the oc-currence of both nutrients at relatively low levels in the RASs;or (2) differing levels of an unmeasured water quality variable.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3

NOTE 547

16

17

18

19

20

21

22

23

24

1 2 3 4 5 6 7 8 9 10 11 12 13

RAS withATS

RAS withoutATS

Time/week

Leng

thm

m

FIGURE 4. Mean growth curves (shell length, mm) for juvenile rainbowmussels that were reared in recirculating aquaculture systems (RASs) equippedwith algal turf scrubbers (ATSs) and in RASs without ATSs.

The survival rate of juvenile mussels was 96.2% in RASswith ATSs and 96.8% in RASs without ATSs. Hence, no signif-icant difference in survival was observed between RAS types(P > 0.05). Generally, survival rates of the juvenile mussels inour study were higher than or similar to those reported in othercomparable studies. For example, O’Beirn et al. (1998) reporteda survival rate of 26.8% for juvenile rainbow mussels after 22weeks in a recirculating trough system. Gatenby et al. (1996)obtained survival rates ranging from 2.7% to 66.5% for juvenilerainbow mussels after 45 d in aerated glass culture dishes con-taining different types of sediment. Layzer et al. (1999) rearedthree species of freshwater mussel in a closed recirculating sys-tem, and survival rates were over 83%. The high survival ratesobserved in RASs with and without ATSs during our study maybe attributable to (1) the maintenance of suitable water quality ineach system type or (2) greater robustness of the older musselswe used.

We regard the results of this pilot-scale trial of relatively largeculture vessels for producing freshwater mussels as promising.We recommend evaluation of such systems at higher biomassloadings and over longer time periods, which would allow morerigorous assessment of system design and biofilter and ATScapacity. Development of high-capacity systems will promoteproduction of imperiled species at such a scale that augmentationor restoration of mussel populations of conservation interest canbe realized.

ACKNOWLEDGEMENTSWe thank Dan Hua (Department of Fish and Wildlife Con-

servation, Virginia Tech) for her assistance during the study.The views expressed in this article are the authors’ and notnecessarily those of the U.S. Fish and Wildlife Service.

REFERENCESAdey, W. H., C. Luckett, and K. Jensen. 1993. Phosphorus removal from natural

waters using controlled algal production. Restoration Ecology 1:29–39.

Adey, W. H., C. Luckett, and M. Smith. 1996. Purification of industrially con-taminated groundwaters using controlled ecosystems. Ecological Engineering7:191–212.

Craggs, R. J., W. H. Adey, K. R. Jenson, M. S. St. John, F. B. Green, andW. J. Oswald. 1996. Phosphorus removal from wastewater using an algal turfscrubber. Water Science and Technology 33:191–198.

Dunn, C. S., and J. B. Layzer. 1997. Evaluation of various holding facilitiesfor maintaining freshwater mussels in captivity. Pages 205–213 in K. S.Cummings, A. C. Buchanan, C. A. Mayer, and T. J. Naimo, editors. Conser-vation and management of freshwater mussels II: initiatives for the future.Upper Mississippi River Conservation Committee, Onalaska, Wisconsin.

Farris, J. L., C. D. Milam, and J. L. Harris. 1999. Zebra mussel impacts on fresh-water mussels in Arkansas. Arkansas Game and Fish Commission, NongameAquatic Section, 1999 Summary Report, Little Rock.

Gatenby, C. M., R. J. Neves, and B. C. Parker. 1996. Influence of sedimentand algal food on cultured juvenile freshwater mussels. Journal of the NorthAmerican Benthological Society 15:597–609.

Hanlon, S. D., and R. J. Neves. 2006. Seasonal growth and mortality of juve-niles of Lampsilis fasciola (Bivalvia: Unionidae) released to a fish hatcheryraceway. American Malacological Bulletin 21:45–49.

Jones, J. W., R. A. Mair, and R. J. Neves. 2005. Factors affecting survival andgrowth of juvenile freshwater mussels cultured in recirculating aquaculturesystems. North American Journal of Aquaculture 67:210–220.

Kovitvadhi, S., U. Kovitvadhi, P. Sawangwong, and J. Machado. 2008. Alaboratory-scale recirculating aquaculture system for juveniles of freshwaterpearl mussel Hyriopsis (Limnoscapha) myersiana (Lea, 1856). Aquaculture275:169–177.

Kovitvadhi, S., U. Kovitvadhi, P. Sawangwong, A. Thongpan, and J. Machado.2006. Optimization of diet and culture environment for larvae and juvenilefreshwater pearl mussels, Hyriopsis (Limnoscapha) myersiana Lea, 1856.Invertebrate Reproduction and Development 49:61–70.

Layzer, J. B., L. M. Madison, J. R. Khym, and R. D. Quinn. 1999. Developingtechnology for long-term holding of mussels in captivity. 1998 Annual Reportto the U.S. Fish and Wildlife Service, Asheville, North Carolina.

MacMillan, R. J., R. J. Cawthorn, S. K. Whyte, and P. R. Lyon. 1994. Designand maintenance of a closed artificial seawater system for long-term holdingof bivalve shellfish. Aquacultural Engineering 13:241–250.

Mulbry, W., S. Kondrad, C. Pizarro, and E. Kebede-Westhead. 2008. Treatmentof dairy manure effluent using freshwater algae: algal productivity and re-covery of manure nutrients using pilot-scale algal turf scrubbers. BioresourceTechnology 99:8137–8142.

Myers-Kinzie, M. L. 1998. Factors affecting survival and recruitment of unionidmussels in small Midwestern streams. Doctoral dissertation. Purdue Univer-sity, West Lafayette, Indiana.

Neves, R. J. 1999. Conservation and commerce: management of freshwatermussel (Bivalvia: Unionoidea) resources in the United States. Malacologia41:461–474.

Neves, R. J., A. E. Bogan, J. D. Williams, S. A. Ahlstedt, and P. W. Hartfield.1997. Status of aquatic mollusks in the southeastern United States: a down-ward spiral of diversity. Pages 43–85 in G. W. Benz and D. E. Collins, editors.Aquatic fauna in peril: the southeastern perspective. Southeast Aquatic Re-search Institute, Special Publication 1, Decatur, Georgia.

O’Beirn, F. X., R. J. Neves, and M. B. Steg. 1998. Survival and growth of ju-venile freshwater mussels (Unionidae) in a recirculating aquaculture system.American Malacological Bulletin 14:165–171.

USEPA (U.S. Environmental Protection Agency). 2009. Draft 2009 update:aquatic life ambient water quality criteria for ammonia–freshwater. USEPA,Office of Water, EPA-822-D-09-001, Washington, D.C.

Veraart, A. J., A. M. Romanı, E. Tornes, and S. Sabater. 2008. Algal responseto nutrient enrichment in forested oligotrophic stream. Journal of Phycology44:564–572.

Williams, J. D., M. L. Warren Jr., K. S. Cummings, J. L. Harris, and R. J. Neves.1993. Conservation status of freshwater mussels of the United States andCanada. Fisheries 18(9):6–22.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3



The Effect of Stocking Different Ratios of NileTilapia Oreochromis niloticus, Striped Mullet Mugilcephalus, and Thinlip Grey Mullet Liza ramada inPolyculture Ponds on Biomass Yield, Feed Efficiency,and Production EconomicsAl-Azab Tahoun a , Ashraf Suloma b , Yasser Hammouda c , Hanan Abo-State c & Ehab El-Haroun a da Aquaculture Department, Fish Resources College , Suez Canal University , El-Arish , Egyptb Fish Nutrition Lab, Animal Production Department, Faculty of Agriculture , CairoUniversity, El-Gamma Street, 12613 Giza , Egyptc Animal Production Department , National Research Center , Cairo , Egyptd Biology Department, Faculty of Science , Taibah University , Al-Madinah Al-Munawwarah30001, Saudi Arabia

To cite this article: Al-Azab Tahoun , Ashraf Suloma , Yasser Hammouda , Hanan Abo-State & Ehab El-Haroun (2013) TheEffect of Stocking Different Ratios of Nile Tilapia Oreochromis niloticus, Striped Mullet Mugil cephalus, and Thinlip GreyMullet Liza ramada in Polyculture Ponds on Biomass Yield, Feed Efficiency, and Production Economics, North American Journalof Aquaculture, 75:4, 548-555






http://dx.doi.org/10.1080/15222055.2013.826764




ARTICLE

The Effect of Stocking Different Ratios of Nile TilapiaOreochromis niloticus, Striped Mullet Mugil cephalus, andThinlip Grey Mullet Liza ramada in Polyculture Ponds onBiomass Yield, Feed Efficiency, and Production Economics

Al-Azab TahounAquaculture Department, Fish Resources College, Suez Canal University, El-Arish, Egypt

Ashraf SulomaFish Nutrition Lab, Animal Production Department, Faculty of Agriculture, Cairo University,El-Gamma Street, 12613 Giza, Egypt

Yasser Hammouda and Hanan Abo-StateAnimal Production Department, National Research Center, Cairo, Egypt

Ehab El-Haroun*1

Fish Nutrition Lab, Animal Production Department, Faculty of Agriculture, Cairo University,El-Gamma Street, 12613 Giza, Egypt

AbstractA 180-d trial was done to evaluate the effect of stocking different ratios of Nile Tilapia Oreochromis niloticus,

Striped Mullet Mugil cephalus, and Thinlip Grey Mullet Liza ramada on production and economic return in brackish-water ponds. The trial was conducted in twelve 3,000-m2 earthen ponds with four treatments (3 ponds/treatment):Nile Tilapia alone (100% Nile Tilapia, monoculture group [MG]); 75% Nile Tilapia and 25% Thinlip Grey Mullet(polyculture group 1 [PG1]); 75% Nile Tilapia, 12.5% Thinlip Grey Mullet, and 12.5% Striped Mullet (PG2); and75% Nile Tilapia and 25% Striped Mullet (PG3). Fish were fed a commercial diet containing 25% crude protein twiceper day. The efficiency of feed utilization was determined in terms of feed conversion ratio (FCR; feed intake/weightgain). Polyculture group 3 had the best FCR, whereas the FCR results did not differ significantly among the MG,PG1, and PG2 treatments. In addition, PG3 had the highest total yield (3,350.3 kg/pond), followed in descendingorder by PG2 (3,234.4 kg/pond), PG1 (3,078.4 kg/pond), and MG (2,856.3 kg/pond). The PG3 treatment achieved thehighest net financial return, followed by PG2, PG1, and MG. This study indicates that tilapia–mullet polyculture mayimprove the efficiency with which natural food resources are used within the system, resulting in better environmentalquality, system sustainability, feed utilization, and net financial return.

Aquaculture is capable of increasing the total fish productionand fulfilling the high demand for fish protein (FAO 2009). Thegrowth of aquaculture in Egypt has steadily increased over thepast few years (Suloma and Ogata 2006). Egypt’s national de-

*Corresponding author: [email protected] address: Biology Department, Faculty of Science, Taibah University, Al-Madinah Al-Munawwarah 30001, Saudi Arabia.Received March 14, 2013; accepted July 2, 2013

velopment plan includes the objectives of increasing the annualper-capita consumption of fish and ensuring the availability ofaffordable fish to the consumer by increasing domestic produc-tion (Suloma and Ogata 2006). One of the policies in the plan is

548

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3

EFFECTS OF POLYCULTURE POND STOCKING RATIOS 549

to develop aquaculture systems and technologies that increasethe efficiency of natural food resource utilization and reduce theenvironmental impact of aquaculture practices, resulting in foodsecurity and sustainability (Bakeer et al. 2008).

Polyculture is one approach to developing aquaculture. Mostof the aquaculture production in Egypt is pond-based using poly-culture farming techniques (GAFRD 2010). Polyculture refersto the farming of several fish species in the same pond, with thespecies differing in feeding behavior, habits, and ecological re-quirements. A freshwater species, the Nile Tilapia Oreochromisniloticus, is the most promising candidate species for increasingaquaculture production over short and intermediate time scales(Suloma and Ogata 2006). In addition, Egypt is the second-largest producer of tilapia and the world’s top producer of mul-let (FAO 2009). Nile Tilapia raised in earthen pond polyculturesystems have shown very good production and feed efficiencyresults (Abdel-Hakim et al. 2000). To increase tilapia productionin polyculture systems, it is important to select the right com-bination of fish species and the appropriate stocking densities(Cruz and Laudencia 1980) since this increases the synergisticfish–fish or fish–environment interaction and decreases aggres-sion. Synergistic relationships can be achieved by enhancingutilization efficiency of natural feed resources, optimizing theuse of the aquatic food web, and improving environmental con-ditions in the pond (Milstein 1995; Greglutz 2003; Billard andBerni 2004; Kotiya Anil et al. 2010; Yuan et al. 2010). NileTilapia are commonly reared in polyculture systems with speciessuch as Common Carp Cyprinus carpio, Striped Mullet Mugilcephalus, catfish, Milkfish Chanos chanos, shrimp, and Euro-pean Eel Anguilla anguilla (Guerrero and Guerrero 1976; Jameset al. 1984; Papoutsoglou et al. 1992; Abdel-Hakim et al. 2000;Dankwa et al. 2004; Kotiya Anil et al. 2010; Biswas et al. 2012).

Earlier studies have focused on tilapia and Striped Mulletin polyculture; however, no research has been performed to as-sess feed utilization efficiency and production of Nile Tilapia inpolyculture with Thinlip Grey Mullet Liza ramada. Therefore,we sought to investigate the effect of various stocking combi-nations of Nile Tilapia and mullet in brackish-water ponds ontotal yield and net return.

METHODSExperimental fish and location.—The trial was carried out in

12 rectangular earthen ponds (3,000 m2; 1.2-m average depth)located at a commercial farm at Riyadh Province, Kafr El-Sheikh Governorate, Egypt. The study was conducted for 180d, during which the ponds were supplied with brackish wa-ter (salinity = 5–15‰). Three percent of the water was re-placed daily during the trial. Monosex (all male) Nile Tilapiafingerlings (20.6 g, overwintered), Thinlip Grey Mullet finger-lings (24.9 g), and Striped Mullet fingerlings (22.2 g) wererandomly assigned to the 12 ponds at a fixed total stocking den-sity of 12,000 fish/pond. Four stocking combinations (Table 1)were randomly assigned to 12 ponds (3 ponds/treatment): 100%Nile Tilapia (monoculture group [MG]); 75% Nile Tilapia and25% Thinlip Grey Mullet (polyculture group 1 [PG1]); 75%Nile Tilapia, 12.5% Thinlip Grey Mullet, and 12.5% StripedMullet (PG2); and 75% Nile Tilapia and 25% Striped Mullet(PG3).

Feed and growth indices.—To determine growth perfor-mance and feed intake, about 20% of the fish were harvestedbiweekly with cast nets. Fish were fed a supplemental diet(3-mm pellet size) at 3% of body weight/d; the diet was admin-istered twice daily at 0800–0900 hours and 1300–1400 hours.Daily feed intake was adjusted at 15-d intervals in response toweight gain. The composition and proximate analysis of theexperimental diet are presented in Table 2. Organic fertilizer(dry poultry manure at 21.43 kg·pond−1·week−1) and inorganicfertilizer (triple superphosphate at 11.12 kg·pond−1·week−1)were used to enhance phytoplankton and zooplanktonproductivity.

Growth and feed efficiency parameters.—Growth rate wasexpressed as specific growth rate (SGR), which was calculatedas {[loge(final body weight) − loge(initial body weight)] ×100}/(number of days). Feed conversion ratio (FCR) was calcu-lated for each pond as (dry matter feed intake)/(live weight gain).Average daily gain (ADG) was calculated as (final biomass −initial biomass)/(number of days). Survival rate (%) was calcu-lated as [(number of harvested fish)/(number of fish stocked intothe ponds)] × 100.

TABLE 1. Experimental treatments and stocking ratios used in the study of Nile Tilapia–mullet polyculture.

Treatment DescriptionStocking rate (total = 12,000

fish/pond)

Monoculture group (MG) 100% Nile Tilapia 12,000 Nile TilapiaPolyculture group 1 (PG1) 75% Nile Tilapia + 25% Thinlip

Grey Mullet9,000 Nile Tilapia + 3,000 Thinlip

Grey MulletPolyculture group 2 (PG2) 75% Nile Tilapia + 12.5% Thinlip

Grey Mullet + 12.5% StripedMullet

9,000 Nile Tilapia + 1,500 ThinlipGrey Mullet + 1,500 StripedMullet

Polyculture group 3 (PG3) 75% Nile Tilapia + 25% StripedMullet

9,000 Nile Tilapia + 3,000 StripedMullet

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3

550 TAHOUN ET AL.

TABLE 2. Formulation and proximate analysis of the experimental diets ad-ministered to ponds containing Nile Tilapia, Thinlip Grey Mullet, and StripedMullet in polyculture.

Ingredient or component Percentage

Fish meal (72% crude protein) 7.0Soybean meal (44% crude protein) 25.0Cottonseed meal 10.0Corn grain 32.25Wheat bran 15.25Vegetable oil 5.0Dicalcium phosphate 2.5Molasses 2.5Anti-aflatoxins 0.1Vitamin C 0.1Mineral and vitamin mixturea 0.3Total 100Analyzed composition (dry matter

basis)Dry matter (%) 89.0Crude protein (%) 25.6Lipid (%) 6.2Ash (%) 9.6Total carbohydrate (%) 58.6

aEach 3-kg quantity of the mineral and vitamin mixture contains the following: vitaminA, 10,000,000 IU; vitamin D3, 2,500,000 IU; vitamin E, 10,000 mg; vitamin K, 1,000 mg;vitamin B1, 1,000 mg; vitamin B2, 5,000 mg; vitamin B6, 1,500 mg; vitamin B12, 10 mg;niacin, 30,000 mg; pantothenic acid, 10,000 mg; folic acid, 1,000 mg; biotin, 50 mg;iodine, 300 mg; iron, 30,000 mg; manganese, 60,000 mg; copper, 4,000 mg; cobalt,100 mg; selenium, 100 mg; zinc, 50,000 mg; and calcium carbonate, 3,000 mg.

Economic and budget analysis.—Economic analyses forNile Tilapia monoculture and Nile Tilapia–mullet polyculturewere calculated using the enterprise budget method. Fixedcosts were based on the requirements of an average Egyptianfarm size of about 1 feddan (0.42 ha). Variable costs includedfingerlings, hired labor, fuel, irrigation, feed, and maintenance.Output prices (in Egyptian pounds, E£) were calculated at 10E£/kg for Nile Tilapia, 14 E£/kg for Thinlip Grey Mullet, and16 E£/kg for Striped Mullet.

Total income (E£/pond) was calculated for each species as(price [E£] of 1 kg of fish) × (fish yield, kg/pond). Total cost(E£/pond) was calculated as (fixed cost) + (variable cost). Netreturn (E£/pond) was computed as (total income) − (total cost).

Statistical analysis.—One-way ANOVA was used to comparewater quality, growth, and production variables among treat-ments by using the GLM procedure in the Statistical AnalysisSystem (SAS Institute 1985). Tukey’s multiple range test wasused to compare differences among individual means. Treat-ment effects were considered significant at P-values less than orequal to 0.05.

Water chemistry analysis.—Dissolved oxygen (DO), pH, to-

tal alkalinity (TA), phosphate (PO−34 ), and total ammonia nitro-

gen (TAN) levels in the water were monitored weekly. Water

temperature was monitored daily at 1500 hours by using a stan-dard mercury thermometer. Dissolved oxygen and pH were de-termined using a waterproof polarographic DO meter (Hanna In-strument, Inc. [HI] 9146-04) and a digital pH meter (HI 98103),respectively. Total alkalinity (CaCO3, mg/L) and phosphate (or-thophosphate, mg/L) were determined with a TA test kit (HI38014) and a phosphate test kit (HI 3833). The TAN levels weredetermined by use of an ammonia test kit (HI 38049). Sec-chidisk readings were recorded weekly to measure water trans-parency. The diet and ingredients were analyzed for dry matterand ash content according to standard methods (AOAC 1990);crude protein (N × 6.25) was determined by using the Kjel-dahl method and a Kjeltech auto-analyzer (Model 1030; Tecator,Hoganas, Sweden); and total lipids were measured by using thepetroleum–ether extraction method and an Ankom XT-20 fat ex-tractor (Ankom, Macedon, New York). The gross energy contentof samples was measured using an automated bomb calorimeter(Model 1272; Parr Instruments, Moline, Illinois); nitrogen-freeextract content was determined as 100 − (crude protein % +crude fat % + crude fiber % + total ash %).

RESULTS AND DISCUSSION

Water QualityNo significant differences in temperature, pH, TAN, or TA

were observed among the treatment ponds (Table 3). The DOlevels were within the desirable range as described by Boyd(1979). Values of TAN ranged between 0.15 and 0.18 mg/L andthus were lower than the maximum recommended concentra-tion of TAN (1.0 mg/L) as reported elsewhere (Meade 1989;Lawson 1995; Ridha and Cruz 1998). There were no signif-icant differences in TA concentration among the ponds (P ≥0.05; Table 3). The TA of the pond water remained within theacceptable limits (10–400 mg/L as CaCO3) reported by Meade(1989). Various stocking ratios significantly affected orthophos-phate levels (P ≤ 0.05). Phosphate (mg/L) values for the MGponds (1.40 mg/L) were significantly (∼20%) higher than thosefor the polyculture ponds (1.24 mg/L for PG1, 1.25 mg/L forPG2, and 1.28 mg/L for PG3), which were not significantlydifferent from each other (P ≥ 0.05; Table 3). These resultsare consistent with those of Tian et al. (2001), who found thatthe phosphate content in the sediments of polyculture systemscontaining Chinese shrimp Penaeus chinensis, Taiwanese redtilapia (Mozambique Tilapia O. mossambicus × Nile Tilapia),and constricted tagelus Sinonovacula constricta (a species ofclam) was lower than the phosphate level in monoculture sedi-ments. The mean Secchi depth in the MG ponds (39.3 cm) wassignificantly greater (P ≤ 0.05) than those in the Nile Tilapia–mullet polyculture ponds (34.6 cm for PG1, 34.0 cm for PG2,and 34.3 cm for PG3), which did not differ (P ≥ 0.05; Table 3).Observation of the lowest Secchi depth readings in the NileTilapia–mullet polyculture ponds can be explained by the factthat mullet mostly consume zooplankton, benthic nutrients, anddetritus before releasing a portion of the consumed nutrients

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3


TABLE 3. Effect of the different stocking treatments (defined in Table 1) on water quality variables (mean ± SD) measured in earthen ponds (DO = dissolved

oxygen; TAN = total ammonia nitrogen; TA = total alkalinity; PO−34 = phosphate). Within a given row, means with different letters are significantly different

(P ≤ 0.05).

Treatment

Water quality variable MG PG1 PG2 PG3

Temperature (◦C) 26.97 ± 0.07 27.03 ± 0.06 26.96 ± 0.06 27.00 ± 0.05pH 7.91 ± 0.04 7.95 ± 0.08 8.03 ± 0.02 7.93 ± 0.01DO at bottom (mg/L) 5.44 ± 0.07 5.30 ± 0.15 5.4 ± 0.05 5.83 ± 0.18TAN (mg/L) 0.15 ± 0.02 0.17 ± 0.00 0.17 ± 0.01 0.18 ± 0.01TA (mg/L) 136.30 ± 9.70 150.00 ± 6.8 143.00 ± 9.80 143.30 ± 4.90

PO−34 (mg/L) 1.40 ± 0.03 z 1.25 ± 0.02 y 1.24 ± 0.01 y 1.28 ± 0.01 y

Secchi depth (cm) 39.3 ± 0.33 z 34.60 ± 0.88 y 34.00 ± 0.57 y 34.30 ± 1.20 y

as excreta. This may provide an important source of nutrientsto stimulate phytoplankton growth, which is mainly consumedby the filter-feeding Nile Tilapia (Brusle 1981; Abdel-Tawwabet al. 2005; Kang and Xian 2008). These results suggest that apolyculture of “noncompetitive and compatible” fish with dif-ferent feeding habits is one way of achieving ecosystem stabilitythrough a variety of processes, including productivity, decom-position, and nutrient cycling (Balvanera et al. 2006; Douglasset al. 2008; Martınez-Porchas et al. 2010). Similarly, Beltonand Little (2008) confirmed that polyculture severely affects the

ecosystem and corroborated the hypothesis that integrated poly-culture systems are good alternatives for achieving aquaculturesustainability by enhancing water quality, reducing contamina-tion, and promoting environmental sustainability.

Survival RateSurvival rates of Nile Tilapia and the two mullet species

(95.5–99.0%; Table 4) were not significantly different amongtreatments. These results are consistent with the findings of ear-lier studies by Cruz and Ridha (1995) and Hengsawat et al.

TABLE 4. Effect of the different stocking treatments (defined in Table 1) on growth performance and survival rate (mean ± SD) of Nile Tilapia, Thinlip GreyMullet, and Striped Mullet reared in earthen ponds (ADG = average daily gain; SGR = specific growth rate). Within a given row, means with different letters aresignificantly different (P ≤ 0.05).

Treatment

Variable MG PG1 PG2 PG3

Nile TilapiaInitial body weight (g) 20.5 ± 0.3 20.5 ± 0.3 20.7 ± 0.2 20.8 ± 0.4Final body weight (g) 321.6 ± 2.2 z 279.0 ± 3.1 y 282.6 ± 1.2 y 279.5 ± 1.9 yADG (g/d) 1.7 ± 0.0 z 1.4 ± 0.0 y 1.5 ± 0.0 y 1.4 ± 0.0 ySGR (%/d) 1.5 ± 0.0 z 1.5 ± 0.0 y 1.5 ± 0.0 y 1.4 ± 0.0 ySurvival (%) 98.7 ± 0.0 99.0 ± 0.0 98.3 ± 0.0 98.7 ± 0.0

Thinlip Grey MulletInitial body weight (g) 24.8 ± 0.4 25.5 ± 0.3Final body weight (g) 203.0 ± 6.2 205.6 ± 3.0ADG (g/d) 0.99 ± 0.0 1.2 ± 0.0SGR (%/d) 1.2 ± 0.0 1.0 ± 0.0Survival (%) 95.5 ± 0.0 96.0 ± 0.0

Striped MulletInitial body weight (g) 22.2 ± 0.4 22.3 ± 0.9Final body weight (g) 304.0 ± 1.7 261.6 ± 3.1ADG (g/d) 1.6 ± 0.0 1.3 ± 0.2SGR (%/d) 1.5 ± 0.0 1.4 ± 0.1Survival (%) 96.9 ± 0.0 98.1 ± 0.0

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3

552 TAHOUN ET AL.

(1997), who observed no significant differences in survivalamong fish species in polyculture systems. The high survivalrates observed in the present study could be attributed to thesynergistic relationship between the fish species, the optimizeduse of the aquatic food web in the water column, decreased ag-gression, and the accelerated turnover of nutrients in the ponds,which resulted in a lower microbial load and waste accumulationand hence higher water quality (Kotiya Anil et al. 2010; Yuanet al. 2010). In contrast, significant differences in survival ratesamong fish species reared under varying stocking densities havebeen reported by others (Miao 1992; Eid and Magouz 1995; Eidand El-Gamal 1996; El-Sagheer 2001; Bakeer et al. 2008).

Growth and Feed UtilizationThe growth of Nile Tilapia slowed significantly (P ≤ 0.05)

in the presence of mullet. The highest Nile Tilapia growth per-formance was obtained in the MG treatment, followed by PG2,PG3, and PG1; Nile Tilapia growth performance did not sig-nificantly differ between the polyculture treatments (Table 4).These results are consistent with the findings of Lovshin (1980)and Teichert-Coddington (1996), who reported that weight gainin Nile Tilapia decreased as the ratio of mullet to Nile Tilapiaincreased in the polyculture system. Similarly, the growth ofNile Tilapia reared in different polyculture systems (e.g., glassaquaria, tanks, floating cages, and concrete ponds) slowed as thestocking rate per species and the weight of stocked fish increased(Cruz and Ridha 1995). However, the growth of Nile Tilapia inour study was greater than or similar to that observed in suc-cessful semi-intensive commercial farms (Milstein 1995) andwas even greater than the values reported by Eid and Magouz(1995). The 14% decline in ADG of Nile Tilapia in polyculturecompared with monoculture (P ≤ 0.05; Table 4) could not be at-tributed to competition for food, since the feed input to all poly-culture and monoculture treatments was adjusted in response toweight gain and the survival of fish was high in all polyculturetreatments. Thus, the feed input in our study was probably suf-ficient to avoid observable competition. The present growth re-

sults were probably caused by a combination of mechanisms, in-cluding (1) the detrimental effect of stocking ratios on the phys-iological adaptive responses of fish to stressors and (2) the neg-ative social interactions among individuals of the same speciesand different species (Wolhfarth et al. 1985; Eid and El-Gamal1996). Among the three tested species, Nile Tilapia attainedthe highest ADG and SGR in comparison with Striped Mulletand Thinlip Grey Mullet;there were no significant differencesin growth of Thinlip Grey Mullet or Striped Mullet stocked atdifferent densities (P ≥ 0.05; Table 4). This may be attributableto the fact that Nile Tilapia grow faster than mullet. In a 150-dNile Tilapia–Striped Mullet polyculture trial, Nile Tilapia exhib-ited SGRs of 1.76–1.98% per day, whereas the SGRs for StripedMullet were 1.03–1.31% per day (Abou Zeid et al. 2005). More-over, Milstein (1995) found that Nile Tilapia demonstrated su-perior production and higher ADG in a polyculture system whenNile Tilapia constituted the majority of cultured organisms inthe pond and were farmed with shrimp and mullet.

The feed utilization efficiency and total yield are presentedin Table 5. Stocking ratios significantly influenced the amountof feed consumed (P ≤ 0.05). Consumption was 17,148.6 kg/hafor MG; 18,097.5 kg/ha for PG1; 18,108.5 kg/ha for PG3; and18,680.9 kg/ha for PG2. The FCR was significantly affected bydifferent Nile Tilapia : mullet stocking ratios (Table 5). The bestFCR was observed for PG3 (P ≤ 0.05), whereas the FCRs forthe other three groups were not significantly different. Abdel-Tawwab et al. (2005) attributed the positive effect of polycultureto the trophic divergence between species: the favorite foods ofmullet are diatoms, while Nile Tilapia consume mostly green al-gae and cyanobacteria. Furthermore, mullet excrete a significantamount of nutrients that promote the growth of phytoplankton,which as stated above is a primary food source for filter-feedingNile Tilapia (Brusle 1981; Abdel-Tawwab et al. 2005; Kangand Xian 2008). This synergy in the tilapia–mullet polyculturesystem leads to an optimized use of the aquatic food web, a re-duction in interspecific and intraspecific competition for food,and improvements in feed efficiency.

TABLE 5. Effect of the different stocking treatments (defined in Table 1) on feed efficiency and total yield (mean ± SD) of Nile Tilapia, Thinlip Grey Mullet,and Striped Mullet reared in earthen ponds (FCR = feed conversion ratio). Relative total yield is reported in relation to the MG treatment yield (i.e., MG yield =100%). Within a given row, means with different letters are significantly different (P ≤ 0.05).

Treatment


Feed intake (kg/ha) 17,148.6 ± 252.1 y 18,097.5 ± 521.1 zy 18,680.9 ± 302.5 z 18,108.5 ± 142.5 zyFCR 1.8 ± 0.0 z 1.8 ± 0.1 z 1.7 ± 0.0 z 1.6 ± 0.0 yNile Tilapia yield (kg/pond) 2,856.3 ± 99.1 z 2,486.1 ± 68.4 y 2,501.6 ± 20.1 y 2,482.1 ± 47.7 yThinlip Grey Mullet yield (kg/pond) 592.3 ± 17.7 z 290.9 ± 17.0 yStriped Mullet yield (kg/pond) 441.9 ± 7.6 y 868.2 ± 9.9 zTotal yield (kg/pond) 2,856.3 ± 9.9 w 3,078.4 ± 40.1 x 3,234.4 ± 20.0 y 3,350.3 ± 25.5 zRelative total yield (%) 100 108 113 117

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3


Relative Yield and Production EfficiencyTotal pond production in the PG1, PG2, and PG3 treatments

was higher by 108, 113, and 117%, respectively, than productionin the MG treatment (Table 5), indicating that the three combi-nations of species are favorable. The high biomass production isattributable mainly to high-quality rearing conditions through-out the trial, such as feed and water quality provided by the dailywater exchange and the weekly application of fertilizers. Fur-thermore, owing to the positive influence of Striped Mullet andThinlip Grey Mullet on water quality, the yield of Nile Tilapiaand the overall conversion rate of input nutrients were high in thepolyculture ponds (Table 5), consistent with other studies thathave reported improvements in Nile Tilapia and Striped Mulletproduction in polyculture systems (Milstein 1995; Eid and El-Gamal 1996; El-Sagheer 2001; Abdel-Tawwab et al. 2005). Thehigh stocking rate of Striped Mullet in PG3 (3,000 fish/pond)demonstrated that with an adequate food supply, it is possible toproduce 868.2 kg of Striped Mullet per pond during one growingseason in a polyculture production system (Table 5). Similarly,Scorvo-Filho et al. (1995) found that the total biomass yieldof Striped Mullet grown in a polyculture system with CommonCarp was greater (P ≤ 0.05) than that of Striped Mullet reared ina monoculture system. Dankwa et al. (2004) found that the rear-

ing of Striped Mullet for 270 d in a monoculture system yielded336.5 kg/ha of mullet in ponds with feed and 287.0 kg/ha inponds without feed. The yield of Striped Mullet in the study byDankwa et al. (2004) was lower than that obtained for StripedMullet and Thinlip Grey Mullet in the present study. Increasedgrowth and total yield in polyculture systems are related to betterfood utilization rather than to higher food consumption, sincethe FCR of Nile Tilapia was about 10.8% higher in MG thanin the polyculture treatments. Consequently, the species farmedtogether in one pond tolerate each other’s presence, and eachspecies also improves the growth and feed efficiency of the otherspecies (Papoutsoglou et al. 1992). The economic efficiency andnet return showed that the total income from polyculture washigher than that from the MG treatment (Table 6). The net re-turn analysis was based on total cost and total income, whichdepends on the price and grades of the polycultured fish (AbouZeid et al. 2005). Based on a budget analysis, the polyculturesystems we examined produced an increase in net income of114.9% (PG1), 132.7% (PG2), and 153.3% (PG3) in relationto the MG treatment. Nile Tilapia farmed with Striped Mulletat a ratio of 75% : 25% (i.e., PG3) had the highest production,income, and net return among the four treatments, mainly dueto the increased yield of Striped Mullet (Table 6).

TABLE 6. Economic analysis and net return of the different stocking treatments (defined in Table 1) of Nile Tilapia, Thinlip Grey Mullet, and Striped Mulletreared in earthen ponds. All values are given in Egyptian pounds (1 E£ = US$6.06). Selling price of 1 kg of fish was 10.0 E£ for Nile Tilapia, 14.0 E£ for ThinlipGrey Mullet, and 16.0 E£ for Striped Mullet; the price of 1 kg of feed was 2.5 E£.

Treatment


Total income (E£/pond)a

Nile Tilapia 28,652.7 24,861.3 25,015.8 24,821.0Thinlip Grey Mullet 8,293.0 4,071.9Striped Mullet 7,070.0 13,891.3Total income 28,652.7 33,154.3 36,157.7 38,712.3Relative total income (%) 100.0 115.7 126.2 135.1Variable costs (E£/pond)FingerlingsNile Tilapia 1,200.0 900.0 900.0 900.0Thinlip Grey Mullet 2,250.0 1,125.0Striped Mullet 1,500.0 3,000.0Labor 250.0 250.0 250.0 250.0Irrigation 500.0 500.0 500.0 500.0Feeds 12,866.0 13,578.0 14,015.0 13,586.0Other 500.0 500.0 500.0 500.0Total variable cost (E£/pond) 15,316.0 17,978.0 18,790.0 18,736.0Fixed cost (E£/pond) 1,000.0 1,000.0 1,000.0 1,000.0Total cost (E£/pond)b 16,316.0 18,978.0 19,790.0 19,736.0Relative total cost (%) 100.0 116.3 121.3 120.9Net return (E£/pond)c 12,786.7 14,176.3 16,367.7 18,976.3Relative percentage of net return 100.0 114.9 132.7 153.3

aTotal income (E£/pond) = price of 1 kg of fish (E£) × fish yield (kg/pond).bTotal cost (E£/pond) = fixed cost + variable cost.cNet return (E£/pond) = total income – total cost.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3

554 TAHOUN ET AL.

ConclusionsThe importance of introducing mullet species into tilapia

ponds is clear, as our results indicate that polyculture improvesfeed utilization efficiency, enhances water quality, and increasestotal yield and profit. Economic and budget analyses revealedthat tilapia–mullet polyculture is a feasible technique and canbe ecologically friendly as well as economically attractive fortilapia farmers.

ACKNOWLEDGMENTSWe thank the students and staff of Cairo University’s Fish

Nutrition Research Unit for their assistance with laboratory andfield work during the research.

REFERENCESAbdel-Hakim, N. F., M. N. Bakeer, and M. A. Soltan. 2000. [Growth perfor-

mance of Eel (Anguilla anguilla), Nile Tilapia (Oreochromis niloticus), andGrey Mullet (Mugil cephalus) cultured in cages under two feeding systems].Pages 329–335 in Proceedings 3rd all Africa conference on animal agricul-ture and 11th conference of the Egyptian society animal production. EgyptianSociety of Animal Production, Giza. (In Arabic).

Abdel-Tawwab, M., A. M. Eid, E. Ali, A. E. Abdelghany, and H. I. El-Marakby.2005. The assessment of water quality and primary productivity in earthenfishponds stocked with Striped Mullet (Mugil cephalus) and subjected todifferent feeding regimes. Turkish Journal of Fisheries and Aquatic Sciences5:1–10.

Abou Zeid, R. M., A. M. Abd El-Maksoud, and A. A. Ali. 2005. Effect ofstocking rates of Nile Tilapia Oreochromis niloticus L. and Grey MulletMugil cephalus L. on their performance in polyculture earthen ponds. Annalsof Agricultural Science (Moshtohor) 43:1057–1066.

AOAC (Association of Official Analytical Chemists). 1990. Official methods ofanalysis, 15th edition. AOAC, Washington, D.C.

Bakeer, M. N., M. A. A. Mostafa, and I. M. A. Samra. 2008. Effect of Mugilcephalus size and density at initial stocking on growth performance andfish marketable size at harvest. Journal of the Arabian Aquaculture Society3:16–32.

Balvanera, P., A. B. Pfisterer, N. Buchmann, J. S. He, T. Nakashizuka, D.Raffaelli, and B. Schmid. 2006. Quantifying the evidence for biodiversityeffects on ecosystem functioning and services. Ecology Letters 9:1146–1156.

Belton, B., and D. Little. 2008. The development of aquaculture in centralThailand: domestic demand versus export-led production. Journal of AgrarianChange 8:123–143.

Billard, R., and P. Berni. 2004. Trends in cyprinid polyculture. Cybium 28:255–261.

Biswas, G., R. Ananda Raja, T. K. Sundaray, J. K. Ghoshal, S. Kumar, S. Pani-grahi, A. Thirunavukkarasu, and A. G. Ponniah. 2012. Evaluation of produc-tions and economic returns from two brackish water polyculture systems intide fed ponds. Journal of Applied Ichthyology 28:116–122.

Boyd, C. E. 1979. Water quality in warm-water fish ponds. Auburn UniversityAgricultural Experimental Station, Auburn, Alabama.

Brusle, J. 1981. Food and feeding in Grey Mullet. Pages 185–217 in O. H. Oren,editor. Aquaculture of Grey Mullet. Cambridge University Press, Cambridge,UK.

Cruz, E. M., and I. L. Laudencia. 1980. Polyculture of Milkfish (Chanos chanosFurskal), all male Nile Tilapia (Oreochromis niloticus), and Snakehead (Oph-icephalus striatus) in fresh water ponds with supplementary feeding. Aqua-culture 20:231–237.

Cruz, E. M., and M. T. Ridha. 1995. Survival rates of tilapia, Oreochromisspilurus (Gunther), fingerlings reared at high densities during winter usingwarm underground sea water. Aquaculture Research 26:307–309.

Dankwa, H. R., J. Blay, and K. Yankson. 2004. Potential for culture of GreyMullets (Pisces: Mugilidae) in Ghana. Ghana Journal of Science 44:19–27.

Douglass, J. G., J. E. Duffy, and J. F. Bruno. 2008. Herbivore and predatordiversity interactively affect ecosystem properties in an experimental marinecommunity. Ecology Letters 11:598–608.

Eid, A., and A. A. El-Gamal. 1996. Effects of stocking density on growthperformance of Nile Tilapia Oreochromis niloticus reared in three differentculture systems. Journal of Animal Production 33(Supplemental Issue):485–495.

Eid, A., and F. Magouz. 1995. Effect of stocking density and feeding rate ongrowth performance of Nile Tilapia Oreochromis niloticus. Journal Agricul-tural Research Tanta University 21:229–236.

El-Sagheer, F. H. M. 2001. Effect of stocking densities, protein levels andfeeding frequencies on growth and production of tilapia mono-sex in earthenponds. Doctoral dissertation. Alexandria University, Alexandria, Egypt.

FAO (Food and Agriculture Organization of the United Nations). 2009. Thestate of world fisheries and aquaculture (SOFIA). FAO, Rome.

GAFRD (General Authority for Fish Resources Development). 2010. The 2009statistical yearbook. Ministry of Agriculture and Land Reclamation, Cairo,Egypt.

Greglutz, C. 2003. Polyculture: principles, practices, problems and promise.Aquaculture Magazine (3–4):1–5.

Guerrero, R. D., and L. A. Guerrero. 1976. Culture of Tilapia niloticus andMacrobrachium species separately and in combination in fertilized freshwater fish ponds. Philippine Journal of Fisheries 14:232–235.

Hengsawat, K., F. J. Ward, and P. Jaruratjamorn. 1997. The effect of stockingdensity on yield, growth and mortality of African Catfish Clarias gariepinusBruchell 1822 cultured in cages. Aquaculture 152:67–76.

James, P. R., A. Raju, and V. Rengaswamy. 1984. Further observations on poly-culture of finfishes and prawns in saltwater ponds and a net pen at Mandapam.Indian Journal of Fisheries 11:31–46.

Kang, B., and W. Xian. 2008. C, N and P regeneration by a detritivorous fish,Liza haematocheila T. and S.: effects of temperature, diet and body size.Aquaculture International 16:319–331.

Kotiya Anil, S., B. Gunalan, K. L. Jetani, G. Kuldeep Trivedi, and P. Soundara-pandian. 2010. Determine the economic feasibility of the polyculture system(giant tiger shrimp and mullet). African Journal of Basic and Applied Sciences2(3–4):124–127.

Lawson, T. B. 1995. Fundamentals of aquaculture engineering. Chapman andHall, New York.

Lovshin, L. L. 1980. Progress report on fisheries development in northeastBrazil. Alabama Agricultural Experiment Station, Auburn.

Martınez-Porchas, M. L. R., M. A. Martınez-Cordova, J. A. Porchas-Cornejo,and P. Lopez-Elıas. 2010. Shrimp polyculture: a potentially profitable, sus-tainable, but uncommon aquacultural practice. Reviews in Aquaculture 2:73–85.

Meade, J. W. 1989. Aquaculture management. Van Nostrand Reinhold, NewYork.

Miao, O. S. 1992. Growth and survival model of red tail shrimp Penaens peni-cillates (Alock) according to manipulating stocking density. Bulletin of theInstitute of Zoology Academia Sinica (Taipei) 31:1–8.

Milstein, A. 1995. Fish-management relationships in Israeli commercial fishfarming. Aquaculture International 3:292–314.

Papoutsoglou, S. E., G. Petropoulos, and R. Barbieri. 1992. Polyculture rearingof Cyprinus carpio (L.) and Oreochromis aureus (St.) using a closed circulatedsystem. Aquaculture 103(3–4):311–320.

Ridha, M. T., and E. M. Cruz. 1998. Observations on the seed production of thetilapia Oreochromis spilurus (Gunther) under different spawning conditionsand with different sex ratios. Asian Fisheries Science 10:201–210.

SAS Institute. 1985. SAS/STAT: guide for personal computers. SAS Institute,Cary, North Carolina.

Scorvo-Filho, J. D., L. M. S. Ayroza, P. F. Colherinhas Novato, and E. R.Almeida Dias. 1995. Efeito da densidade de estocagem sobre o crescimentoda Tainha listrada (Mugil platanus) criada em mono e policultivo Com Carpa

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3


comum (Cyprinus carpio) na regiao do Vale do Ribeira, SP. [Effect of den-sity on the growth of Striped Mullet (Mugil platanus) reared in mono andpoliculture with Common Carp (Cyprinus carpio), in Vale do Ribeira re-gion.] Boletim do Instituto de Pesca 22(2):85–93. (In Portuguese with Englishabstract.)

Suloma, A., and H. Y. Ogata. 2006. Future of rice-fish culture, desert aqua-culture and feed development in Africa: the case of Egypt as the leadingcountry in Africa. JARQ (Japan Agricultural Research Quarterly) 40:351–360.

Teichert-Coddington, D. R. 1996. Effect of stocking ratio on semi-intensivepolyculture of Colossoma macropomum and Oreochromis niloticus inHonduras, Central America. Aquaculture 143:291–302.

Tian, X., D. Li, S. Dong, G. Liu, Z. Qi, and J. Lu. 2001. Water quality of closedpolyculture of penaeid shrimp with tilapia and constricted tagelus. ChineseJournal of Applied Ecology 12:287–292.

Wolhfarth, G., W. G. Hulata, I. Carplus, and A. Halevy. 1985. Polyculture offreshwater prawn Macrobrachium rosenbergii in intensively manured ponds,and the effect of stocking rate of prawns and fish on their production charac-teristics. Aquaculture 46:143–156.

Yuan, D., Y. Yi, A. Yakupitiyage, K. Fitzimmons, and J. S. Dian. 2010. Effectsof addition of Red Tilapia Oreochromis spp. at different densities and sizes onproduction, water quality and nutrient recovery of intensive culture of whiteshrimp Litopenaeus vannamei in cement tanks. Aquaculture 298(3–4):226–238.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:40

27

Oct

ober

201

3



Suppression of Cannibalism during Larviculture ofBurbot through Size GradingJames M. Barron a b , Nathan R. Jensen c d , Paul J. Anders c a , Joshua P. Egan c d & KennethD. Cain d ea Department of Fish and Wildlife Resources , University of Idaho , Post Office Box 441136,Moscow , Idaho , 83844-1136 , USAb U.S. Fish and Wildlife Service , Abernathy Fish Technology Center , 1440 Abernathy CreekRoad, Longview , Washington , 98632 , USAc Cramer Fish Sciences , 317 West 6th Street, Suite 204, Moscow , Idaho , 83843 , USAd Department of Fish and Wildlife Resources and Aquaculture Research Institute , Universityof Idaho , Post Office Box 441136, Moscow , Idaho , 83844-1136 , USAe National Centre for Marine Conservation and Resource Sustainability , University ofTasmania , Locked Bag 1370, Launceston , Tasmania , 7250 , Australia

To cite this article: James M. Barron , Nathan R. Jensen , Paul J. Anders , Joshua P. Egan & Kenneth D. Cain (2013)Suppression of Cannibalism during Larviculture of Burbot through Size Grading, North American Journal of Aquaculture, 75:4,556-561






http://dx.doi.org/10.1080/15222055.2013.829146




TECHNICAL NOTE

Suppression of Cannibalism during Larviculture of Burbotthrough Size Grading

James M. BarronDepartment of Fish and Wildlife Resources, University of Idaho, Post Office Box 441136, Moscow,Idaho 83844-1136, USA; and U.S. Fish and Wildlife Service, Abernathy Fish Technology Center,1440 Abernathy Creek Road, Longview, Washington 98632, USA

Nathan R. JensenCramer Fish Sciences, 317 West 6th Street, Suite 204, Moscow, Idaho 83843, USA;and Department of Fish and Wildlife Resources and Aquaculture Research Institute,University of Idaho, Post Office Box 441136, Moscow, Idaho 83844-1136, USA

Paul J. AndersCramer Fish Sciences, 317 West 6th Street, Suite 204, Moscow, Idaho 83843, USA;and Department of Fish and Wildlife Resources, University of Idaho,Post Office Box 441136, Moscow, Idaho 83844-1136, USA

Joshua P. EganCramer Fish Sciences, 317 West 6th Street, Suite 204, Moscow, Idaho 83843, USA;and Department of Fish and Wildlife Resources and Aquaculture Research Institute,University of Idaho, Post Office Box 441136, Moscow, Idaho 83844-1136, USA

Kenneth D. Cain*Department of Fish and Wildlife Resources and Aquaculture Research Institute, University of Idaho,Post Office Box 441136, Moscow, Idaho 83844-1136, USA;and National Centre for Marine Conservation and Resource Sustainability, University of Tasmania,Locked Bag 1370, Launceston, Tasmania 7250, Australia

AbstractThe survival and percentage of North American Burbot Lota

lota maculosa larvae and metamorphosing larvae presumed to havebeen cannibalized during a 15-d period immediately following asize-grading event were compared with those of fish in nongradedcontrol groups. In larvae (mean TL, 11.8 mm), grading imme-diately produced a size distinction, as the group that passed thegrader was significantly narrower and shorter than the group re-tained by the grader. The mean coefficient of variation of the lengthof larvae in the retained group was significantly lower than that ofthe control group, indicating that grading reduced size heterogene-ity. Grading significantly increased larval survival, which averaged74.3% and 93.3% for the passed and retained fish, respectively,compared with 59.3% in the control. Increased survival was linked

*Corresponding author: [email protected] May 15, 2013; accepted July 22, 2013

to a reduction in the percentage of larvae presumed cannibalized inthe graded groups. In metamorphosing larvae (mean TL, 21.0 mm),grading did not significantly change TL, width, or the coefficientof variation of length, nor did it improve survival or reduce pre-sumed cannibalism. This study provides initial empirical evidencethat size grading can be an effective way to significantly reducecannibalism when done at the onset of cannibalism in larval-stageBurbot.

Development of culture techniques has recently begunfor the Eurasian Burbot Lota lota lota and North AmericanBurbot L. lota maculosa subspecies in response to decliningpopulations and burgeoning commercial interest (Wolnicki

556

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3

SUPPRESSION OF CANNIBALISM IN BURBOT 557

et al. 2001, 2002; Harzevili et al. 2003, 2004; Jensen et al.2008; Zarski et al. 2009; Trebelsi et al. 2011; Wocher et al.2011). Burbot larvae are inherently difficult to culture, haveexhibited low survival in captivity, and may require live feedsfor up to 5 weeks (Harzevili et al. 2003; Jensen et al. 2008).Cannibalism in Burbot has been reported in intensive culturesettings (Trebelsi et al. 2011) and the natural environment(Kahilainen and Lehtonen 2003). Kahilainen and Lehtonen(2003) reported cannibalism in Burbot that were 21.1 mm long,which is within the size range in which metamorphosis occurs(18–30 mm; Eloranta 1985; Ryder and Pesendorfer 1992;Fischer 1999; McPhail and Paragamian 2000). Metamorphosisfrom larvae to juveniles in Burbot includes the developmentof external morphological features analogous to those in adultfish, including fin differentiation, formation of a single chinbarbel, and full pigmentation (Donner and Eckmann 2011).During this time, previously pelagic larval Burbot adopt abenthic orientation (Ryder and Pesendorfer 1992; Fischer 1999;McPhail and Paragamian 2000). In a weaning trial that spannedthe larval and juvenile life stages, Trebelsi et al. (2011) reportedthat up to 45% of the Burbot population was lost to cannibalismover 36 d. In juveniles of a related species, Atlantic CodGadus morhua, reduced size heterogeneity following gradingeffectively reduced cannibalism (Folkvord and Ottera 1993).

We speculated that grading could be an effective strategy forreducing cannibalism during early-life-stage rearing of Burbot ifthey, like numerous other teleost fishes, require size differencesfor cannibalism to occur (Smith and Reay 1991). Thus, the goalsof this study were to (1) develop a grading method and determinewhether size grading of two early life stages of Burbot couldseparate a single population into two populations of distinct sizeswith reduced body length and width variation and (2) determinewhether size grading could reduce cannibalism and increase thesurvival of larval and metamorphosing Burbot.

METHODSTwo grading trials were conducted at the University of Idaho–

Aquaculture Research Institute (UI–ARI), Moscow, to investi-gate whether a grading device could reduce length variation andthe magnitude of cannibalism, thereby increasing survival dur-ing Burbot larviculture. Trial 1 began when larval Burbot firstexhibited cannibalistic behavior. Trial 2 began after trial 1 wascompleted, when a commercial weaning diet was first intro-duced to larval Burbot undergoing metamorphosis. Both trialswere conducted using the same municipal water source, whichreceived carbon filtration and sodium thiosulfate treatments witha Reefdoser Quadro Pump (Aqua Medic, Fort Collins, Colorado)to neutralize chlorine.

Experimental animals.—All experimental fish used in bothtrials were produced at the UI–ARI from captive North Ameri-can Burbot broodstock obtained from Moyie Lake in southeast-ern British Columbia. The fish used in both trials originated fromthe same mixed-family population. Trail 1 used 1,100 larval Bur-bot at 80 d posthatch (DPH), 800 of which were graded. Initially,

the TL for this population was 11.8 ± 2.4 mm (mean ± SD)and the width was 2.3 ± 0.4 mm (n = 90). Trial 2 used a sourcepopulation of 650 Burbot at 100 DPH (i.e., when they were en-tering metamorphosis), of which 500 were graded. Initially thispopulation had a TL of 21.0 ± 2.4 mm and a width of 3.8 ±0.5 mm (n = 90).

Experimental apparatus.—A small experimental grader wasconstructed for this study using a rectangular fish net breeder(13.3 cm tall, 16.4 cm wide, and 12.5 cm deep; Rolf C. HagenCorp., Mansfield, Massachusetts) as a frame. For trial 1 a nylonmesh screen (2.0-mm square mesh) was hot-glued to the graderframe on five sides, providing an open-topped box with a totalscreen area of 788 cm2 and a volume of approximately 2.7 L.For trial 2 the same grader frame was refitted with larger mesh(3.0-m × 3.5-mm rectangular mesh) to accommodate larger fish.

Grading.—In each trial, fish were fed live enriched Artemiafranciscana 16 h prior to grading. Burbot were graded by ran-domly capturing groups of 100 fish from the source populationand pouring them gently into the grader, which was 95% sub-merged in a second tank. This design provided approximately1 cm of freeboard to prevent fish from jumping out of the grader.Water was gently added to the center of the grader (1.0 L/minat 12◦C) to encourage passage through the mesh. To completegrading with sufficient time to initiate a culture trial the sameday, the grader was removed 15 min after the group was pouredinto the grader and all retained fish were transferred to a separatetank. This process was repeated until both graded populations(passed and retained) contained enough fish to stock 300 fish intrial 1 and 150 fish in trial 2.

Culture trials.—Upon completion of grading, replicated cul-ture trials (trials 1 and 2) were conducted to assess the effectson TL, width, and the coefficient of variation of TL among thegraded (passed and retained) and ungraded control groups, aswell as whether grading affected the percentage presumed can-nibalized and survival 15 d postgrading. Each of the three groups(passed, retained, and control) represented treatments in a com-pletely randomized experimental design with three replicatesper treatment.

After grading, three random samples of 100 fish each for trial1 and 50 fish each for trial 2 were captured from the treatmentgroups and transferred to a flat-bottomed container with a waterdepth of 1 cm and a scale of measure. Digital photographs werethen taken (Kodak EasyShare C330 digital camera; EastmanKodak Co., Rochester, New York). The resulting electronicimages enabled accurate measurement of the TLs and widthsof 30 randomly subsampled fish per replicate through digitalimage analysis software (ImageJ 1.40; Wayne Rasband, opensource). An additional group weight estimate was obtained intrial 2 to enable calculation of feeding rates (% body weight/d).Each replicate of 100 (trial 1) and 50 (trial 2) fish was then trans-ferred to a 6.5-L rectangular flow-through aquarium (26.7 ×15.9 × 16.5 cm) with a flow rate of 360 mL/min at 13.9 ±0.9◦C in trial 1 and 350 mL/min at 15.2 ± 0.8◦C in trial 2. Theaquaria were continuously illuminated via overhead fluorescent

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3

558 BARRON ET AL.

lighting. All tanks were monitored in the morning and eveningfor mortalities, which were promptly removed and recorded.Tanks were siphoned daily in the morning prior to feeding.

In trial 1, the fish in each tank were fed A. franciscana nau-plii (enriched for 24 h posthatch with Nannochloropsis sp. paste)three times per day at a density of 2/mL. In trial 2, the fish ineach tank were fed an amount equal to 3.1% of their initialweight four times per day. Weaning was initiated at the begin-ning of trial 2, when live and artificial feed was co-fed from day0 through day 7. The feeding regime consisted of enriched A.franciscana nauplii as described for trial 1 followed 1 h later bya commercially available diet (Otohime B2; Reed Mariculture,Campbell, California) equal to the dry weight of the A. francis-cana offered. This was done twice per day, in the morning andevening. The enriched A. franciscana were assumed to weigh2 µg (Evjemo and Olsen 1999). Co-feeding was terminated afterthe first 8 d of trial 2, when both A. franciscana feedings werereplaced by feedings of Otohime B2. All trials were terminatedon day 15, when all of the remaining fish in each replicate tankwere captured and enumerated. All fish unaccounted for duringthe 15-d trials were assumed to have been cannibalized.

Metrics and statistical analysis.—To assess the effects ofgrading, the three variables of interest—TL, width, and the co-efficient of variation of TL (CV; 100 × SD/mean)—were deter-mined for each replicate on day 0 of each trial. Survival and thepercentage presumed cannibalized were calculated on day 15 ofthe trial and expressed as percentages of the initial population.Thus, survival was calculated as 100 × (final population/initialpopulation) and the percentage presumed cannibalized was cal-culated as 100 × (number cannibalized/initial population).

All data were expressed as means ± SEs. One-way analysisof variance (ANOVA) was used to compare postgrading TL,width, and CV, the transformed (arcsine square root) proportionsurviving, and the transformed proportion presumed cannibal-ized among the treatments. A Tukey post hoc analysis was usedto test for differences among treatments. Data analysis was per-formed using SAS 9.2 (SAS Institute, Inc., Cary, North Car-olina). Statistical significance was defined as P < 0.05 for allcomparisons.

RESULTS AND DISCUSSIONGrading successfully separated one population of larval

Burbot into two distinct size-groups, as fish that passed throughthe grader were significantly narrower and shorter than fish thatwere retained in the net (Table 1). Of the 800 fish subjected to

grading in trial 1, 41.8% passed and 58.2% were retained. Thepassed and retained groups both differed significantly from thecontrol fish in width but not in TL. Grading also significantly re-duced size heterogeneity, as the retained fish had a significantlylower CV than the fish in the control group (Table 1). The meanCV for passed fish was lower than that for the control, but notsignificantly so.

In trial 1, survival was significantly higher in the gradedlarvae than the control group; on average, 34.0% and 15.0%more fish survived in the retained and passed groups, respec-tively, than in the control group (Table 1). The retained groupalso experienced significantly higher survival than the passedgroup (93.3% versus 74.3%). Increased survival in the passedand retained groups was associated with reductions in presumedcannibalism.

The amount of presumed cannibalism in larval Burbot wassignificantly reduced by grading (Table 1). On average, 1.0% ofthe retained larvae were presumed to have been cannibalized, aquantity significantly less than in the control (29.0%) and passed(14.3%) groups. Passed fish were also cannibalized significantlyless than control fish. As indicated by their CV, retained larvaerepresented the most homogenous group in terms of length dis-tribution, which likely reduced the opportunities for cannibalismwithin this treatment. The positive correlation between the per-centage of cannibalized fish and size heterogeneity suggests thata size differential between predator and prey is required for can-nibalization to occur, as has been reported for other teleost fishes(Baras and Jobling 2002). Folkvord (1997) has suggested thatfor Atlantic Cod undergoing metamorphosis a 25% differencein TL is required for successful ingestion; otherwise, the gapeof the predator’s mouth will be too small to accommodate theprey’s body width. Although the mouth gape : body width ra-tio required for ingestion in Burbot and how it changes as theygrow are currently unknown, the effectiveness of grading intrial 1 (resulting in significantly lower cannibalization in groupswith less size variation) suggests that a size differential is alsoa prerequisite for cannibalism among Burbot.

Failed attacks resulting in mortal wounding of the attackedindividuals were observed during both trial 1 and trial 2; how-ever, the number of failed attacks could not be quantified. Mortalwounding by cannibals has also been reported in larval AtlanticCod, occurring when an attacked fish was too large for inges-tion (Øiestad 1985). Estimates of the percentage cannibalizedare likely conservative in populations that exhibit this mortalwounding phenomenon without subsequent scavenging of the

TABLE 1. Total length, width, the coefficient of variation of TL (CV), survival, and cannibalism (means ± SEs) of larval Burbot (trial 1) after grading (passedand retained) and in a nongraded control. Total length, width, and the CV were assessed at day 0 postgrading and survival and cannibalism at day 15. Withincolumns, means denoted by different lowercase letters are significantly different (P < 0.05) as determined by ANOVA and Tukey’s comparison of mean values.

Treatment TL (mm) Width (mm) CV (%) Survival (%) Cannibalism (%)

Control 11.8 ± 0.7 zy 2.3 ± 0.1 y 18.1 ± 1.2 z 59.3 ± 0.3 x 29.0 ± 0.6 zPassed 10.0 ± 0.3 y 2.0 ± 0.0 x 16.1 ± 1.3 zy 74.3 ± 1.8 y 14.3 ± 2.6 yRetained 13.8 ± 0.5 z 2.6 ± 0.1 z 11.7 ± 0.8 y 93.3 ± 1.2 z 1.0 ± 0.6 x

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3


TABLE 2. Total length, width, the coefficient of variation of TL (CV), survival, and cannibalism (means ± SEs) of metamorphosing Burbot (trial 2) aftergrading (passed and retained) and in a nongraded control. See Table 1 for additional details.

Treatment TL (mm) Width (mm) CV (%) Survival (%) Cannibalism (%)

Control 21.0 ± 0.1 z 3.9 ± 0.0 z 11.4 ± 0.5 z 39.3 ± 4.7 z 10.0 ± 9.0 zyPassed 20.4 ± 0.6 z 3.8 ± 0.1 z 10.1 ± 0.8 z 38.7 ± 2.9 z 0.7 ± 0.7 yRetained 21.1 ± 0.3 z 3.9 ± 0.1 z 13.5 ± 1.1 z 28.7 ± 9.6 z 52.7 ± 21.0 z

mortalities because the mortalities generated from this behaviorcannot be directly attributed to cannibalism.

Unlike in trial 1, the grading of postlarval Burbot in trial 2 didnot separate the population into two distinct size-groups; TL andwidth did not differ significantly among treatments (Table 2). Ofthe 500 fish subjected to grading in trial 2, 65.6% were passedand 34.4% were retained. Although the passed group exhibitedslightly lower TL and width relative to the control and retainedgroups, these differences were not significant. The mean CV washighest in the retained group (13.5%) and lowest in the passedgroup (10.1%); however, CV did not differ significantly amongthe treatments (Table 2). The control fish in trial 2 had a lowermean CV (11.4%) than the control fish in trial 1 (18.1%). Theseresults suggest that intense cannibalism selectively removed thesmallest individuals in the general population prior to trial 2,effectively reducing size heterogeneity, a phenomenon seen inpopulations of other fishes during intense cannibalism (Barasand Jobling 2002).

Grading also did not significantly improve survival in trial 2(Table 2). Survival was lower among the metamorphosing fish intrial 2 than among the larvae in trial 1, which may have been dueto weaning to artificial diets, a process typically characterized byhigh mortality in Burbot (Jensen et al. 2008). Starvation duringweaning may have also contributed to an increase in the num-ber of consumptive and nonconsumptive attacks, collectivelyincreasing mortality.

The percentage of fish presumed cannibalized in trial 2was not significantly different between the graded and controlpopulations (Table 2); however, retained fish exhibited a sig-nificantly higher percentage of presumed cannibalization thanpassed fish, with means of 52.7% and 0.7%, respectively. Ineach of the tanks in trial 2 in which substantial (≥28%) losses tocannibalism were observed, one or two disproportionately largeindividuals were present. These individuals were discernibleon day 1 of trial 2 (Figure 1A), and the size distinction wasmagnified over the course of the trial (Figure 1B). As observed

FIGURE 1. Metamorphosing Burbot from the (A) retained and (C) control groups on day 0 of trial 2 and the (B) retained and (D) control groups on day 15.The two larger, circled fish in panels (A) and (B) illustrate the growth advantage of cannibalism. The retained group lost 86% of its fish to cannibalism, with 10%surviving to the end of the trial. The control group lost no fish to cannibalism, with 48% surviving to the end of the trial.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3

560 BARRON ET AL.

in other cannibalistic fishes, an individual cannibal’s diet accel-erates its growth relative to that of its potential prey, resultingin increasing size differences that further promote cannibalisticbehavior in a positive feedback loop (Folkvord and Ottera1993; Folkvord 1997; Baras and Jobling 2002). Therefore,size grading before the onset of substantial cannibalism isrecommended.

The potential for losses to be inflicted by fish that have gaineda large growth advantage was highlighted in trial 2, in whichone replicate of the retained treatment contained two of theselarger individuals and lost 86% of its population (presumablyto cannibalism) over the 15-d trial (Figure 1A, B). In replicatesthat did not have any of these disproportionately large individ-uals, the occurrence of presumed cannibalism was minimal tononexistent (Figure 1C, D). The presence of these few dispro-portionately large cannibals may have also contributed to thehigh variation in survival and the percent presumed cannibal-ized within certain treatments. In these cases, it may have beenmore effective to use a larger mesh, isolating only the largest5% of the population to effectively isolate the disproportionatelylarge fish. In this study, however, such a grading strategy wouldnot have yielded sufficient quantities of retained fish to conductreplicated trials.

As observed with the larval Burbot in trial 1, increased sizeheterogeneity was associated with an increase in the percentagepresumed cannibalized in trial 2. The mean CV was lower intrial 2 (11.4%) than in trial 1 (18.1%), which may have alsocontributed to a lower percentage presumed cannibalized in thecontrol group of trial 2 (10.0% as opposed to 29.0% in trial 1).Further refinement and modification of grading methods, includ-ing grading duration and frequency, environmental conditions,fish size and density, and grader design may further increase theeffectiveness with which grading separates fish by size, reducessize heterogeneity and the percent cannibalized, and increasessurvival.

In conclusion, this study revealed significant impacts of can-nibalism associated with the culturing of early life stages ofBurbot and provided an initial grading method to suppress thisbehavior and thus increase survival. The effectiveness of grad-ing young Burbot in intensive culture settings may be further in-creased if grading is done multiple times during the early devel-opmental stages to more effectively decrease the CV. The CVsfor each trial provided initial guidelines with which to assessthe relative risk of cannibalism among larval and metamorphos-ing Burbot populations. The technique and grading apparatusintroduced herein has the potential to be scaled up and adaptedfor use with other larval fish species. Based on this research,size grading should occur early in Burbot larviculture (i.e., be-fore substantial cannibalism has occurred) to increase larval sur-vival. It is expected that reductions in cannibalism following sizegrading of the magnitude reported here, along with the recom-mended population-specific optimization of grading regimes,would significantly improve the success of Burbot production

in a wide variety of conservation and commercial aquaculturesettings.

ACKNOWLEDGMENTSThis project was funded in part by the U.S. Fish and Wildlife

Service (grant 14330-7-H067). Special thanks to Ray Jones forprogram funding and coordinating. This project was also sup-ported by the Kootenai Tribe of Idaho and the Bonneville PowerAdministration (project 198806400; contract 46821). We extendour deepest gratitude to the Kootenai Tribe of Idaho, the BritishColumbia Ministry of Forests, Lands, and Natural Resource Op-erations, the Idaho Department of Fish and Game, Cramer FishSciences, the U.S. Fish and Wildlife Service, and the Universityof Idaho Aquaculture Research Institute (UI–ARI). We extendour gratitude to Susan Ireland of the Kootenai Tribe of Idahofor her support and dedication. Without these collaborators thiswork would not have been possible. Finally, we thank the UI–ARI staff members who assisted during this study. The findingsand conclusions in this article are those of the authors and donot necessarily represent the views of the U.S. Fish and WildlifeService.

REFERENCESBaras, E., and M. Jobling. 2002. Dynamics of intracohort cannibalism in cultured

fish. Aquaculture Research 33:461–479.Donner, M. T., and, R. Eckmann. 2011. Diel vertical migration of larval and

early juvenile Burbot optimizes survival and growth in a deep, pre-alpinelake. Freshwater Biology 56:916–925.

Eloranta, A. 1985. Observations on the development and growth of young Bur-bot (Lota lota L.). Jyvaskylan Yliopiston Biologian Laitoksen Tiedonantoja43:73–107.

Evjemo, J. O., and Y. Olsen. 1999. Effect of food concentration on the growthand production rate of Artemia franciscana feeding on algae (T. iso). Journalof Experimental Marine Biology and Ecology 243:273–296.

Fischer, P. 1999. Otolith microstructure during the pelagic, settlement, andbenthic phases in Burbot. Journal of Fish Biology 54:1231–43.

Folkvord, A. 1997. Ontogeny of cannibalism in larval and juvenile fishes withspecial emphasis on Atlantic Cod. Pages 251–278 in R. C. Chambers andE. A. Trippel, editors. Early life history and recruitment in fish populations.Chapman and Hall, London.

Folkvord, A., and H. Ottera. 1993. Effects of initial size distribution, day length,and feeding frequency on growth, survival, and cannibalism in juvenileAtlantic Cod (Gadus morhua L.). Aquaculture 114:243–260.

Harzevili, A. S., D. De Charleroy, J. Auwerx, J. Van Slycken, P. Dhert, andP. Sorgeloos. 2003. Larvae rearing of Burbot (Lota lota) using Brachionuscalyciflorus rotifer as starter food. Journal of Applied Ichthyology 19:84–87.

Harzevili, A. S., I. Dooremont, I. Vught, J. Auwerx, P. Quataert, and D. DeCharleroy. 2004. First feeding of Burbot, Lota lota, larvae under differenttemperatures and light conditions. Aquaculture Research 35:49–55.

Jensen, N., S. C. Ireland, J. T. Siple, S. R. Williams, and K. D. Cain. 2008.Evaluation of egg incubation methods and larval feeding regimes for NorthAmerican Burbot. North American Journal of Aquaculture 70:162–170.

Kahilainen, K., and H. Lehtonen. 2003. Piscivory and prey selection of fourpredator species in a whitefish dominated subarctic lake. Journal of FishBiology 63:659–672.

McPhail, J. D., and V. L. Paragamian. 2000. Burbot biology and life history.Pages 11–23 in V. L. Paragamian and D. W. Willis, editors. Burbot: biology,

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3


ecology, and management. American Fisheries Society, Fisheries Manage-ment Section Publication 1, Spokane, Washington.

Øiestad, V., P. G. Kvenseth, and A. Folkvord. 1985. Mass production of AtlanticCod juveniles Gadus morhua in a Norwegian saltwater pond. Transactions ofthe American Fisheries Society 114:590–595.

Ryder, R. A., and J. Pesendorfer. 1992. Food, growth, habitat, and communityinteractions of young-of-the-year Burbot, Lota lota L., in a PrecambrianShield lake. Hydrobiologia 243/244:211–227.

Smith, C., and P. Reay. 1991. Cannibalism in teleost fish. Reviews in FishBiology and Fisheries 1:41–64.

Trebelsi, A., J. Gardeur, F. Teletchea, and P. Fontaine. 2011. Effects of 12 factorson Burbot Lota lota (L., 1758) weaning performances using a fractionalfactorial design experiment. Aquaculture 316:104–110.

Wocher, H., A. Harsanyi, and F. J. Schwarz. 2011. Husbandry conditions inBurbot (Lota lota L.): impact of shelter availability and stocking density ongrowth and behavior. Aquaculture 315:340–347.

Wolnicki, J., and L. R. Kaminski. 2001. Influence of water temperature ongrowth, survival, condition, and biological quality of juvenile Burbot Lotalota (L.). Archives of Polish Fisheries 9:79–86.

Wolnicki, J., R. Kaminski, and L. Myszkowski. 2002. Temperature-influencedgrowth and survival of Burbot Lota lota (L.) larvae fed live food undercontrolled conditions. Archives of Polish Fisheries 10:109–113.

Zarski, D., W. Sasinowski, D. Kucharczyk, M. Kwiatkowski, S. Krejszeff,and K. Targonska. 2009. Mass initial rearing of Burbot Lota lota (L.) lar-vae under controlled conditions. Polish Journal of Natural Sciences 24:76–84.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3



Effect of Short-Term Holding without Feedingafter Capture on Reduction in Oxidative Stress andMaintenance of Lipid and Amino Acid Contents inDecapterus maraudsiRyusuke Tanaka a , Hideto Fukushima b , Toshimichi Maeda b , Taisei Kumazawa c , YoshimasaSugiura b , Teruo Matsushita b , Hideo Hatate a & Yutaka Fukuda ba Department of Marine Biology and Environmental Sciences, Faculty of Agriculture ,University of Miyazaki , Gakuen-kibanadai-nishi-1-1, Miyazaki , 889-2192 , Japanb Department of Food Science and Technology , National University of Fisheries , 2-7-1Nagata-Honmachi, Shimonoseki , Yamaguchi , 759-6595 , Japanc Research and Development Department , Nichimo Company, Ltd. , 2-3-17, Ozuki-Kojima,Shimonoseki , Yamaguchi , 750-1136 , Japan

To cite this article: Ryusuke Tanaka , Hideto Fukushima , Toshimichi Maeda , Taisei Kumazawa , Yoshimasa Sugiura , TeruoMatsushita , Hideo Hatate & Yutaka Fukuda (2013) Effect of Short-Term Holding without Feeding after Capture on Reductionin Oxidative Stress and Maintenance of Lipid and Amino Acid Contents in Decapterus maraudsi , North American Journal ofAquaculture, 75:4, 562-571






http://dx.doi.org/10.1080/15222055.2013.831003




ARTICLE

Effect of Short-Term Holding without Feeding after Captureon Reduction in Oxidative Stress and Maintenance of Lipidand Amino Acid Contents in Decapterus maraudsi

Ryusuke Tanaka*Department of Marine Biology and Environmental Sciences, Faculty of Agriculture,University of Miyazaki, Gakuen-kibanadai-nishi-1-1, Miyazaki 889-2192, Japan

Hideto Fukushima and Toshimichi MaedaDepartment of Food Science and Technology, National University of Fisheries, 2-7-1 Nagata-Honmachi,Shimonoseki, Yamaguchi 759-6595, Japan

Taisei KumazawaResearch and Development Department, Nichimo Company, Ltd., 2-3-17, Ozuki-Kojima, Shimonoseki,Yamaguchi 750-1136, Japan

Yoshimasa Sugiura and Teruo MatsushitaDepartment of Food Science and Technology, National University of Fisheries, 2-7-1 Nagata-Honmachi,Shimonoseki, Yamaguchi 759-6595, Japan

Hideo HatateDepartment of Marine Biology and Environmental Sciences, Faculty of Agriculture,University of Miyazaki, Gakuen-kibanadai-nishi-1-1, Miyazaki 889-2192, Japan

Yutaka FukudaDepartment of Food Science and Technology, National University of Fisheries, 2-7-1 Nagata-Honmachi,Shimonoseki, Yamaguchi 759-6595, Japan

AbstractChanges in hydroxyl lipids (L-OHs; a stress level indicator), body weight, triglycerides, fatty acids, and free amino

acids (FAAs) were investigated in Decapterus maraudsi caught by purse seine from two different ocean regions nearJapan to determine the effect of short-term holding without feeding on the reduction in oxidative stress. In experiment1, the L-OHs in the dorsal muscle significantly decreased by 9 d, whereas those in the liver decreased significantlyby the end of the 33-d holding period. The body weight, lipid content, and triglyceride composition did not changesignificantly for 9 d and then decreased significantly by 33 d. Fatty acid composition of the dorsal muscle and liver didnot change significantly within the holding period. The FAA content in the dorsal muscle did not change until 9 d andthen decreased by 33 d. In experiment 2, the L-OHs in the dorsal muscle had significantly increased at 33 d, whereasthe levels in the liver of the fish after transport to port were twice as high as those in the fish that had just been caught,but then they decreased significantly by the end of the holding period. The change in body weight, lipid content, andtriglyceride composition were the same as in experiment 1. The fatty acid compositions changed at 33 d, and the totalFAA content did not change until 20 d and then decreased at 33 d. Both results suggest that to avoid stress and toobtain a higher quality fish product, fish captured by purse seine should be held without feeding for 8–9 d rather than


562

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3

EFFECT OF SHORT-TERM HOLDING WITHOUT FEEDING ON JAPANESE SCAD 563

immediately shipped after capture. The results of this study may be useful for improving the quality and commodityvalue of fish captured in commercial fisheries, such as by purse seine, and those held in aquaculture operations, suchas in sea cages.

Popular fishes such as Horse Mackerel Trachurus japonicas(also known as Japanese Scad), Pacific Chub Mackerel Scomberjaponicus, and Japanese Sardine Sardinops sagax melanostictus(also known as Japanese Pilchard) are phototactic fish and arecaught by purse seine in Japan. However, since purse-seinefishing catches pelagic fishes, which are highly active and fastswimmers, the fish encircled by the net may attack the net in anattempt to escape (Machii et al. 1990). Therefore, fish caughtby purse seine experience stress in the sea and in the air atthe time of capture (Marcalo et al. 2006, 2008). Although thestressed fish are frozen or chilled and then distributed for sale,this does not prevent deterioration in the quality of the fish.Therefore, to obtain a higher-quality fish product it is necessaryto reduce the stress caused by capture and transport. A goodmethod to reduce stress is to capture one fish at a time and killit immediately. However, this method requires time and effortand is economically disadvantageous on a fishing vessel.

In recent years, short-term holding (1–2 weeks after cap-ture) of fish without feeding has been performed for JapaneseScad Decapterus maraudsi (known as Whitetip Scad in NorthAmerica) and Pacific Chub Mackerel, which support majorcoastal fisheries on the west side of Japan. Fish are capturedby purse seine and held without feeding for a fixed period un-til shipment. Because holding without feeding helps adjust thetiming of shipment and sale of live fish in the market, this sys-tem has an economic advantage and also helps fish recover fromstress. According to fishermen, fish held without feeding tendto show decreased stress and their lipid content does not de-crease immediately but rather increases or remains constant fora certain period of time and then decreases.

Many researchers have studied the stress response in fish. Inboth commercial fisheries and aquaculture operations, capture,handling, and transportation are traumatic procedures and maycause considerable physiological responses in fish, which canaffect flesh quality (Sigholt et al. 1997). The stress inducesreactive oxygen species, and the reactive oxygen species canreact with all biological molecules, i.e. lipids, proteins, nucleicacids, and carbohydrates (Martinez-Alvarez et al. 2005). Es-pecially, the reacted lipids generated during an early encounterwith an oxidant add molecular oxygen to produce lipid radicals.This prooxidant abstracts an allylic hydrogen from anotherunsaturated side chain, producing lipid peroxide (LPO).Polyunsaturated fatty acids (PUFAs) contained in the fish areeasily changed to LPO, and the LPO might lead to oxidativestress in the fish. Moreover, the LPO leads to deteriorated foodflavor, texture, and appearance in fish products (Bess et al. 2013;Jimenez-Colmenero et al. 2013; O’ Dwyer et al. 2013). The

LPO is readily decomposed to produce a variety of volatile com-pounds that are the main cause of undesirable odor in fish muscle(Tanaka et al. 2013a, 2013b). In our laboratory, the stress in livefish (Tanaka et al. 1999, 2002, 2006, 2012) and the quality of fishproducts (Tanaka et al. 2013a, 2013b) were evaluated by levelsof LPO. However, LPO levels only reflect the current physiolog-ical condition in live fish. The LPOs are immediately reducedto more inactive hydroxyl lipids (L-OHs) by low-molecular-weight antioxidants or enzymes such as catalase, superoxidedismutase, and glutathione peroxidase (Harman 1981; Sevanianet al. 1983). Therefore, estimating L-OH levels should provideanother helpful index of oxidative stress in vivo. We developeda method to determine L-OH levels using a fluorescent reagentand high-performance liquid chromatography (HPLC) (Tanakaet al. 1999, 2002). In these reports, L-OH levels were highestin the liver and then in red muscle followed by white muscle,based on activity in these tissues, and levels in cultured fishwere significantly higher than those in wild fish, reflecting thedegree of oxidative stress (Tanaka et al. 1999). We also verifiedL-OH accumulation in various diseased and exercise-stressedfish (Tanaka et al. 2002; Tanaka and Nakamura 2012). Theseresults suggest that L-OH is a good index of internal stress infish.

The quality of a marine product is determined not only bylipids but also by other components. In particular, free aminoacids (FAAs) were implicated in the characteristic taste ofseafood (Konosu et al. 1974; Komata 1990). In addition, FAAsplay important roles in physiological functions, such as os-moregulation and buffering capacity in the tissues of aquaticanimals (Love 1970, 1980; Abe et al. 1987; Van Waarde 1988).Therefore, the FAA content may be affected by oxidative stress.A sharp decline in histidine occurs in Skipjack Tuna Katsuwonuspelamis, Japanese Eel Anguilla japonica, Rainbow Trout On-corhynchus mykiss, and Japanese Dace Tribolodon hakonensisduring starvation (Abe and Ohmama 1987).

In this study, to confirm the effect of short-term holding with-out feeding after capture on the reduction in oxidative stressto obtain a higher quality fish product, L-OH levels were in-vestigated as a marker of stress in Japanese Scad caught witha purse seine, transported by vessel, and held without feed-ing. Also, the change in lipid components (lipid content, lipidclasses, vitamins A and E, fatty acids, and LPO) and aminoacids related to fish product quality were investigated. On thewest side of Japan, short-term holding for 1–2 weeks after cap-ture of fish without feeding is used for Japanese Scad. There-fore, we selected two regions that use this fishery style for thisstudy.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3

564 TANAKA ET AL.

METHODSSampling and culture of Japanese Scad without feeding after

capture.—Japanese Scad were caught by commercial vesselsusing a purse seine offshore of Kanesaki, Fukuoka Prefecture,Japan, in July 2010 (experiment 1). The seawater temperaturewas 22.0◦C. Fish were captured by a typical seine operationused in commercial fishing. The fish captured by vessels withthe seine were directly transferred from the seine to tanks(2 m × 3 m × 1.7 m) on the transport vessels. On the transportvessels, 10 samples of the transported fish were randomlycollected to analyze. Immediately after decapitation andexsanguination, the dorsal muscle and liver were excised. Thenthe tissues were dipped in 30 mL of chloroform:methanol (2:1,volume to volume [v/v]) and kept in an ice box. The accurateweight of these tissues was not determined because of unstableconditions on the vessel. The other fish were transported for 3 hto Kanesaki port. On arrival at the port, 10 samples of the trans-ported fish from the vessels were randomly collected and thedorsal muscle and liver from each fish were excised to analyze.The other transported fish (∼500 kg) were moved to a fish cage(10 m × 10 m × 8 m, mesh size of 43.3 mm), which was locatedinside the port. The fish were held in the fish cage for a holdingperiod of 33 d without feeding; the mortality of the fish and theseawater temperature were monitored every morning through-out the holding period. The seawater temperature ranged from21.3◦C to 24.2◦C for the holding period. The held fish weresampled after 3, 9, and 33 d, and at each of the sampling days,20 fish were randomly selected and the dorsal muscle andliver from each fish were excised to analyze. The experimentalprotocol was approved by the Institutional Animal Care andUse Committee, National University of Fisheries. The collecteddorsal muscle and liver were kept in an ice box and transportedto our laboratory. In the laboratory, the dorsal muscle ofeach sample (∼50 g) was homogenized with a commercialfood processor (TK410; Tescom, Tokyo, Japan) for 30 s. Theposthomogenized sample and whole liver were treated.

A similar study using Japanese Scad was conducted offshoreof Uwajima, Ehime Prefecture, Japan, in July 2010 (experiment2). The size of the purse seine, fishing procedure, and samplingon the vessel were the same as for experiment 1, althoughthe time in transport from fishing point to port differed. Thecaptured fish were transported by transport vessel for about1.5 h. On arrival at the port, the transported fish (∼500 kg) weremoved to a fish cage (8 m × 8 m × 8 m, mesh size of 33.7 mm)located inside the port. These fish were also cultured in the fishcage for a holding period of 33 d without feeding; the mortalityof the fish and the seawater temperature were monitored everymorning throughout the holding period. The seawater tempera-ture ranged from 22.5◦C to 25.1◦C for the holding period. Theheld fish were sampled after 0.5, 8, 20, and 33 d, and at eachof the sampling days, 20 fish were randomly selected and thedorsal muscle and liver from each fish were excised to analyze.The subsequent procedures were the same as those used inexperiment 1.

Analysis of lipid-related substances.—Lipids were extractedby mixing the posthomogenized dorsal muscle or whole liversample (∼5 g) with chloroform:methanol (2:1, v/v) contain-ing 0.005% butylated hydroxytoluene under ice-cold conditions(Folch et al. 1957; Nakamura et al. 1991). Then 10 mL ofchloroform:methanol (2:1, v/v) was added, and the sampleswere stored at −40◦C. Triglycerides and phospholipids wereanalyzed by thin-layer chromatography with flame ionizationdetection using IATROSCAN-MK5 (Mitsubishi Chemical Me-dience Corporation, Tokyo, Japan) (Moriya et al. 2007). Foranalysis of fatty acids, the extracted lipids were transmethylatedby saponification, followed by BF3-catalyzed methylation (Ichi-hara et al. 2010), and the resultant fatty acid methyl esters wereanalyzed by gas–liquid chromatography with flame ionizationdetection (G-6000; Hitachi, Tokyo, Japan). Lipid peroxide levelswere determined by the triphenylphosphine reduction methodwith HPLC (Nakamura and Maeda 1991). The L-OH contentwas determined by measuring anthroyl esters with a fluorescentHPLC (Tanaka et al. 1999). Levels of vitamin A and vitamin Ewere estimated by measuring the amount of retinol and alpha-tocopherol, respectively. The vitamins E and A were extractedfrom lipids (Kramer 1997). The extraction was dissolved in1 mL of ethyl acetate:n-hexane (10:90, v/v) for HPLC analysis.Vitamins E and A were measured with a fluorescent HPLC sys-tem. The Japanese Scad samples mainly contained α-tocopherol,with small amounts of β-, γ-, and δ-tocopherol; therefore, weassessed the vitamin E content by the α-tocopherol contentfrom a logarithmic calibration curve of authentic α-tocopherolstandards.

Analysis of FAAs and taurine.—Free amino acids were im-plicated in the characteristic taste of seafood, and they mayhave been affected by oxidative stress and holding withoutfeeding. Taurine and FAAs were extracted by homogenizingthe dorsal muscle of Japanese Scad (∼5 g) with 20 mL of 1-Mperchloric acid under ice-cold conditions (Ahimbisibwe et al.2010). The extracted FAA and taurine contents were analyzedby the AccQ-Tag amino acid with a fluorescent HPLC system(Millipore 1993).

Statistical analyses.—In experiment 1, the sample size takenafter catch and after transport was 10 fish but was 20 fish forsamples taken at 3, 9, and 33 d, as summarized in Table 1. Inexperiment 2, the sample size taken after catch was 5 fish, aftertransport was 10 fish, and at 0.5, 8, 20, and 33 d was 20 fish,as summarized in Table 2. These data were analyzed using one-way analysis of variance. Statistical differences between eachsampling period were assessed using Tukey’s test, and a P-value<0.05 was considered significant. The statistical analyses wereperformed by Ekuseru-Toukei 2010 (Social Survey ResearchInformation Co., Japan)

RESULTSThe weight of the fish did not change significantly for 9

d (mean ± SD, 137.34 ± 28.74 g) without feeding; however,

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3


TABLE 1. Change in body weight and lipid-related substances in the muscle and liver of Japanese Scad without feeding after capture (experiment 1). The valuesare mean ± SD. The values of each parameter in the same row with different letters are significantly different (P < 0.05).

Measurement After catch After transport 3 d 9 d 33 d

Sample size 10 fish 10 fish 20 fish 20 fish 20 fishBody weight (g) 145.58 ± 26.79 z 140.17 ± 27.69 z 137.34 ± 28.74 z 120.1 ± 15.06 y

Muscle (dorsal)Lipid contents (%) 4.90 ± 1.87 z 4.20 ± 1.34 zy 5.45 ± 1.12 z 3.55 ± 0.99 yTriglyceride composition (%) 76.96 ± 9.18 zy 81.37 ± 9.02 zy 70.05 ± 8.81 zy 85.30 ± 9.68 z 69.98 ± 8.20 yPhospholipid composition (%) 27.69 ± 2.87 27.26 ± 3.86 36.53 ± 4.66 29.47 ± 4.48 34.91 ± 3.74Vitamin E (µg/mg lipid) 0.05 ± 0.02 0.05 ± 0.01 0.07 ± 0.03 0.07 ± 0.02 0.08 ± 0.05Vitamin A (µg/mg lipid) tracea trace trace trace traceLPO (nmol/mg lipid) 2.72 ± 1.14 z 2.31 ± 1.04 z 1.71 ± 0.81 z 3.94 ± 1.54 y 2.09 ± 1.08 zL-OH (nmol/mg lipid) 4.69 ± 3.12 z 4.11 ± 1.31 z 4.20 ± 3.00 z 2.68 ± 1.25 y 3.98 ± 2.19 z

LiverLipid contents (%) 7.95 ± 1.68 7.24 ± 0.92 7.72 ± 0.65 8.11 ± 1.92Triglyceride composition (%) 32.65 ± 18.87 35.18 ± 4.86 46.93 ± 12.66 44.26 ± 4.48 47.54 ± 5.74Phospholipid composition (%) 66.27 ± 18.81 63.09 ± 10.53 51.39 ± 10.18 54.49 ± 9.47 51.10 ± 5.30Vitamin E (µg/mg lipid) 0.20 ± 0.23 z 0.19 ± 0.20 z 0.17 ± 0.13 z 0.15 ± 0.17 z 0.06 ± 0.07 yVitamin A (µg/mg lipid) 3.65 ± 2.37 3.75 ± 3.75 7.99 ± 6.30 6.95 ± 3.63 3.51 ± 2.29LPO (nmol/mg lipid) 8.53 ± 6.41 z 9.91 ± 8.53 z 8.47 ± 3.56 z 10.73 ± 8.99 z 20.25 ± 9.00 yL-OH (nmol/mg lipid) 14.45 ± 9.64 z 15.67 ± 5.42 z 11.99 ± 5.44 y 9.58 ± 2.63 yx 8.07 ± 3.84 x

aTrace indicates only a small amount of vitamin A was found in the sample.

the weights decreased significantly (P < 0.05) by 33 d(120.10 ± 15.06 g; rate of weight loss was 12.5%) in experi-ment 1. The change in fish weight in experiment 2 was similarto that in experiment 1. The weight did not change for 8 d(140.17 ± 20.83 g) and then decreased significantly (P < 0.05)at 20 d (115.48 ± 14.60 g; rate of weight loss was 17.9%) and 33d (111.28 ± 15.88 g; rate of weight loss was 20.8%) (Tables 1, 2).

The lipid content and triglyceride composition in the dorsalmuscle did not change significantly for 9 d and then decreasedsignificantly (P < 0.05) by 33 d without feeding in experiment 1(Table 1). The change in lipid content in the dorsal muscle cor-related with that of triglyceride composition from after transportto the end of the 33-d holding period (r2 = 0.9015). However,levels in the liver did not change significantly for 33 d. In exper-iment 2, the lipid content in the dorsal muscle and liver did notchange for 8 d but then decreased significantly (P < 0.05) by20 d and 33 d. Triglyceride composition in the dorsal muscle didnot change for 8 d but then decreased significantly (P < 0.05)by 20 d and 33 d. The change in lipid content in the dorsal mus-cle correlated with that of triglyceride composition from aftertransport to the end of the 33-d holding period (r2 = 0.9914).However, triglyceride composition in the liver increased and de-creased during the 33-d holding period and was not correlatedwith lipid content (Table 2).

Lipid peroxide significantly (P < 0.05) increased by 9 din the dorsal muscle (3.94 ± 1.54 nmol/mg lipid) and by 33 din the liver (20.25 ± 9.00 nmol/mg lipid) in experiment 1. TheL-OHs significantly (P < 0.05) decreased by 9 d (2.68 ± 1.25nmol/mg lipid) in the dorsal muscle, whereas that in liver de-

creased significantly (P < 0.05) by the end of the holding period.In experiment 2, the level of LPO significantly (P < 0.05) in-creased by 20 d in the dorsal muscle (3.04 ± 1.19 nmol/mglipid) and that in the liver increased significantly (P < 0.05) by20 d (33.87 ± 10.43 nmol/mg lipid). The L-OHs significantlyincreased by 33 d in the dorsal muscle (15.14 ± 6.03 nmol/mglipid), whereas those in liver decreased significantly (P < 0.05)by the end of the holding period (Table 2).

Vitamin A and vitamin E were estimated in this study by mea-suring the amount of retinol and alpha-tocopherol, respectively.A small amount of vitamin A was found in the dorsal muscleduring both experiments. Although vitamin A was found in theliver, the change in its level did not correlate with the holdingperiod. In contrast, vitamin E was found in the dorsal musclebut it did not change during the holding period. However, in theliver the vitamin E content decreased during the holding periodand was inversely correlated with the LPO level. The change invitamin E content correlated inversely with that of LPO fromafter transport to 33-d holding (r2 = 0.940 in experiment 1, r2 =0.981 in experiment 2) (Tables 1, 2).

The fatty acid composition of the dorsal muscle during ex-periment 1 did not change significantly with the holding period(Table 3). However, the composition of 14:0, 16:0, 18:1(n-9),and 20:5(n-3) fatty acids decreased and that of 22:6(n-3) fattyacid increased significantly (P < 0.05) by 33 d in the dorsalmuscle during experiment 2 when compared with levels at thebeginning of holding (Table 4). If (n-3) fatty acid content isexpressed in terms of milligrams per gram tissue, the level didnot change until 8 d. In the case of the liver, the composition of

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3

566 TANAKA ET AL.

TABLE 2. Change in body weight and lipid-related substances in the muscle and liver of Japanese Scad without feeding after capture (experiment 2). The valuesare mean ± SD. The values of each parameter in the same row with different letters are significantly different (P < 0.05).

Measurement After catch After transport 0.5 d 8 d 20 d 33 d

Sample size 5 fish 10 fish 20 fish 20 fish 20 fish 20 fishBody weight (g) 146.54 ± 26.54 z 144.62 ± 28.43 z 140.17 ± 20.83 z 115.48 ± 14.60 y 111.28 ± 15.88 y

Muscle (dorsal)Lipid contents (%) 2.57 ± 1.28 z 2.75 ± 1.27 z 2.29 ± 0.31 z 1.34 ± 0.68 y 0.88 ± 0.46 xTriglyceride

composition (%)37.44 ± 3.90 z 53.40 ± 6.16 y 55.71 ± 4.87 y 51.88 ± 5.17 y 28.16 ± 6.90 x 16.81 ± 2.01 w

Phospholipidcomposition (%)

61.51 ± 5.60 z 45.41 ± 5.42 y 43.58 ± 5.54 y 48.1 ± 4.97 y 70.42 ± 5.46 x 81.18 ± 6.60 w

Vitamin E (µg/mglipid)

0.35 ± 0.10 0.37 ± 0.17 0.35 ± 0.14 0.49 ± 0.07 0.39 ± 0.13 0.54 ± 0.20

Vitamin A (µg/mglipid)

tracea trace trace trace trace trace

LPO (nmol/mg lipid) 0.67 ± 0.57 z 0.75 ± 0.49 z 1.13 ± 0.69 z 1.88 ± 0.65 z 3.04 ± 1.19 y 2.34 ± 1.03 zL-OH (nmol/mg lipid) 6.10 ± 1.40 z 7.38 ± 4.25 z 7.70 ± 2.54 z 11.49 ± 3.03 zy 9.46 ± 3.69 z 15.14 ± 6.03 y

LiverLipid contents (%) 5.53 ± 1.74 z 5.13 ± 0.94 z 5.29 ± 1.54 z 3.20 ± 0.30 y 2.88 ± 1.31 yTriglyceride

composition (%)21.21 ± 2.60 z 20.16 ± 2.49 z 14.23 ± 2.99 y 17.69 ± 3.43 z 10.30 ± 2.95 y 24.28 ± 3.80 z

Phospholipidcomposition (%)

75.97 ± 5.36 z 75.22 ± 8.01 z 81.23 ± 9.21 y 79.85 ± 7.47 z 87.34 ± 7.47 y 73.47 ± 6.45 z

Vitamin E (µg/mglipid)

1.74 ± 1.10 z 1.95 ± 0.52 z 1.15 ± 0.47 z 1.09 ± 0.09 z 0.09 ± 0.10 y 0.07 ± 0.08 y

Vitamin A (µg/mglipid)

2.22 ± 3.10 1.93 ± 2.46 8.34 ± 4.25 5.44 ± 8.21 8.96 ± 8.64 4.22 ± 5.14

LPO (nmol/mg lipid) 1.21 ± 0.28 z 1.22 ± 0.29 z 2.65 ± 0.67 z 2.46 ± 0.31 z 33.87 ± 10.43 y 21.46 ± 17.20 xL-OH (nmol/mg lipid) 17.41 ± 0.70 z 33.04 ± 2.17 y 25.47 ± 1.54 x 18.35 ± 5.76 z 14.29 ± 2.08 w 10.48 ± 1.37 w

aTrace indicates only a small amount of vitamin A was found in the sample.

20:4(n-6) fatty acid increased and 20:5(n-3) fatty acid decreasedsignificantly (P < 0.05) by the end of the holding period in ex-periment 1 (Table 5). In experiment 2, the composition of 16:0,18:2(n-6), 20:1(n-9), 18:3(n-3), 22:5(n-3), and 22:6(n-3) fattyacids were not stable for the holding period. However, the com-position of 20:4(n-6) fatty acid increased and 20:5(n-3) fattyacid decreased significantly (P < 0.05) during the holding pe-riod, the same as in experiment 1 (Table 6).

Total FAA content in the dorsal muscle did not change un-til 9 d (441.79 ± 52.55 mg/100 g tissue) and then decreasedby 33 d (233.64 ± 74.65 mg/100 g tissue) in experiment 1. Inexperiment 2, the total FAA content did not change until 20 d(286.68 ± 47.20 mg/100 g tissue) and then decreased by 33 d(198.40 ± 31.95 mg/100 g tissue). For both experiments, histi-dine made up a large percentage of the total FAAs (60–70%)so that changes in the total FAA content were correlated withchanges in the histidine content. Change in taurine content fol-lowed changes in the total FAAs and histidine content. In ex-periment 1, aspartic acid (influences umami taste) and alanineand leucine (both of which influence sweet taste) did not changeuntil 9 d and then decreased by 33 d. However, in experiment 2,

glycine and alanine (both of which influence sweet taste), valineand leucine (both of which are concerned with bitter taste), andaspartic acid and glutamic acid (both of which are concernedwith umami taste) did not change for 33 d (Tables 7, 8).

DISCUSSIONIn both experiments, fish body weight, which was related to

the lipid content and triglyceride-to-total-lipid ratio in the dor-sal muscle, did not change for 8–9 d but then decreased by theend of the experiments. However, the initial lipid content of thedorsal muscle in both experiments was different because of dif-ferent growth environments. The lipid content of the JapaneseScad was easily changed by season, water temperature, speedof oceanic current, etc. The oceanic current at Uwajima (exper-iment 2) was faster than at Kanesaki (experiment 1); therefore,the initial lipid content of Uwajima’s Japanese Scad might havebeen higher than that of Kanesaki’s Japanese Scad due to theuse of lipid for getting energy.

To confirm the effect of short-term holding without feeding,we examined L-OH as a stress marker in the dorsal muscle

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3


TABLE 3. Change in fatty acid composition (%; mean ± SD) in the dorsal muscle of Japanese Scad without feeding after capture (experiment 1).

Fatty acidsa After catch After transport 3 d 9 d 33 d

14:0 3.78 ± 1.55 4.73 ± 1.26 4.02 ± 1.35 4.14 ± 1.22 3.28 ± 0.6616:0 27.34 ± 2.59 27.32 ± 1.81 26.09 ± 3.09 27.18 ± 1.73 24.91 ± 2.2718:0 8.72 ± 0.97 8.48 ± 0.93 9.53 ± 3.41 8.40 ± 0.82 8.48 ± 1.9918:1(n-9) 21.7 ± 4.64 21.45 ± 3.31 19.2 ± 6.25 22.99 ± 2.58 25.10 ± 4.1318:2(n-6) 0.97 ± 0.22 0.99 ± 0.18 0.93 ± 0.18 0.96 ± 0.16 0.95 ± 0.2418:3(n-3) 0.62 ± 0.09 0.54 ± 0.23 0.56 ± 0.20 0.59 ± 0.17 0.53 ± 0.1420:1(n-9) 0.92 ± 0.44 0.59 ± 0.31 0.80 ± 0.52 0.77 ± 0.27 0.67 ± 0.3520:4(n-6) 2.07 ± 0.63 1.98 ± 0.42 2.05 ± 0.85 1.88 ± 0.47 1.92 ± 0.7920:5(n-3) 8.91 ± 1.75 8.89 ± 1.10 8.89 ± 0.97 7.97 ± 2.08 6.99 ± 1.0122:5(n-3) 2.90 ± 0.53 3.21 ± 0.55 3.37 ± 0.62 3.06 ± 0.42 3.09 ± 0.8322:6(n-3) 22.05 ± 4.55 21.83 ± 2.89 24.56 ± 6.83 22.06 ± 2.62 24.06 ± 4.27Saturated 36.07 ± 2.30 35.80 ± 1.61 35.62 ± 4.58 35.51 ± 1.57 33.09 ± 2.34Unsaturated 64.38 ± 9.30 65.24 ± 10.61 65.18 ± 9.58 65.28 ± 8.57 67.23 ± 9.11(n-6) PUFA 3.04 ± 0.62 2.97 ± 0.52 2.98 ± 0.79 2.77 ± 0.53 2.90 ± 0.86(n-3) PUFA 34.49 ± 5.00 34.47 ± 3.51 37.38 ± 6.80 34.33 ± 3.56 34.84 ± 4.94(n-3) PUFAb (mg/g tissue) 12.67 ± 2.34 11.77 ± 1.25 14.03 ± 1.02 10.28 ± 0.85

aFor the fatty acids, the number to the left of the colon is the number of carbon atoms in the compound, the number immediately to the right of the colon is the number of doublebonds, and the number after the hyphen indicates the position of the first double bond from the methyl end.

bNote that these (n-3) PUFAs are expressed in terms of milligrams of fatty acid per gram of tissue.

and liver of Japanese Scad during holding. In both experiments,the L-OH levels in the liver decreased during short-term hold-ing without feeding. However, the L-OH levels in the liver offish sampled immediately after transport were twice as highas those of fish sampled immediately after capture in experi-ment 2 but not in experiment 1. One possible reason for thiswas that the captured fish in experiment 2 might have been

stressed during transport. The L-OH levels in the muscle didnot change during short-term holding without feeding but thosein the liver decreased. In a previous report, we confirmed thatL-OH levels were highest in the liver, then in red muscle fol-lowed by white muscle, based on activity in these tissues andthat L-OH levels in lower-activity tissues, such as white muscle(dorsal muscle), were not always in agreement with the stress

TABLE 4. Change in fatty acid composition (%; mean ± SD) in the dorsal muscle of Japanese Scad without feeding after capture (experiment 2). The values ofeach parameter in the same row with different letters are significantly different (P < 0.05).

Fatty acids After catch After transport 0.5 d 8 d 20 d 33 d

14:0 3.20 ± 1.08 z 3.10 ± 1.05 z 4.37 ± 1.18 z 1.62 ± 0.76 y 1.84 ± 1.07 y 1.11 ± 0.98 y16:0 22.67 ± 1.72 z 23.33 ± 2.24 z 24.79 ± 1.83 z 20.07 ± 2.61 y 20.15 ± 1.76 y 18.70 ± 2.71 y18:0 9.94 ± 1.07 10.57 ± 0.75 11.81 ± 4.93 10.66 ± 0.77 10.95 ± 0.86 11.30 ± 1.1418:1(n-9) 10.26 ± 4.67 z 13.31 ± 2.55 z 17.15 ± 3.63 y 10.08 ± 1.95 z 12.96 ± 1.51 z 8.42 ± 1.06 x18:2(n-6) 1.04 ± 0.18 1.29 ± 0.47 1.15 ± 0.19 1.05 ± 0.46 0.95 ± 0.36 1.31 ± 1.0420:1(n-9) 2.39 ± 2.20 2.36 ± 0.99 2.58 ± 1.01 1.50 ± 0.67 0.95 ± 0.65 1.36 ± 0.5218:3(n-3) 0.71 ± 0.25 0.79 ± 0.11 0.56 ± 0.30 1.20 ± 0.83 0.49 ± 0.26 1.00 ± 0.7520:4(n-6) 2.50 ± 0.69 2.24 ± 0.40 2.27 ± 1.10 3.27 ± 0.84 2.88 ± 0.44 3.46 ± 0.6020:5(n-3) 7.68 ± 0.82 z 8.20 ± 0.73 z 8.45 ± 0.62 z 6.11 ± 0.91 z 6.16 ± 0.76 z 5.54 ± 1.15 y22:5(n-3) 3.46 ± 0.33 3.51 ± 0.43 3.53 ± 0.73 3.28 ± 0.48 3.97 ± 0.80 3.35 ± 0.4622:6(n-3) 36.13 ± 6.41 z 31.32 ± 5.94 z 23.32 ± 5.57 y 41.15 ± 5.41 zx 38.69 ± 9.06 zx 44.47 ± 9.16 xSaturated 32.62 ± 1.09 33.90 ± 1.75 36.61 ± 3.86 30.73 ± 2.52 31.10 ± 1.87 30.00 ± 2.53Unsaturated 68.12 ± 10.09 66.92 ± 9.75 64.86 ± 8.23 70.73 ± 9.84 69.19 ± 9.88 70.00 ± 10.45(n-6) PUFA 3.54 ± 0.52 3.53 ± 0.82 3.42 ± 1.17 4.32 ± 0.70 3.83 ± 0.57 4.77 ± 1.01(n-3) PUFA 47.99 ± 6.73 zx 43.81 ± 5.61 z 35.86 ± 5.05 y 51.75 ± 4.62 x 49.31 ± 8.45 x 54.35 ± 8.44 x(n-3) PUFAa (mg/g tissue) 8.44 ± 1.23 z 7.40 ± 1.61 z 8.89 ± 1.25 z 4.96 ± 0.61 y 3.59 ± 0.28 y

aNote that these (n-3) PUFAs are expressed in terms of milligrams of fatty acid per gram of tissue.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3

568 TANAKA ET AL.

TABLE 5. Change in fatty acid composition (%; mean ± SD) in the liver of Japanese Scad without feeding after capture (experiment 1). The values of eachparameter in the same row with different letters are significantly different (P < 0.05).

Fatty acids After catch After transport 3 d 9 d 33 d

14:0 2.60 ± 1.10 2.81 ± 0.87 2.34 ± 1.17 2.65 ± 1.84 1.95 ± 0.8516:0 23.08 ± 6.02 25.60 ± 4.35 23.58 ± 6.54 27.38 ± 4.81 23.47 ± 5.9918:0 9.85 ± 6.49 8.19 ± 1.28 8.02 ± 1.81 7.27 ± 0.79 6.87 ± 1.2718:1(n-9) 17.87 ± 8.08 19.04 ± 6.99 17.90 ± 5.19 19.31 ± 3.43 20.81 ± 5.0318:2(n-6) 0.85 ± 0.28 0.88 ± 0.21 0.73 ± 0.44 0.87 ± 0.73 0.94 ± 0.4718:3(n-3) 0.72 ± 0.30 0.78 ± 0.50 0.77 ± 0.08 0.26 ± 0.31 0.55 ± 0.3920:1(n-9) 3.30 ± 1.38 3.45 ± 0.93 0.80 ± 0.98 0.46 ± 0.57 0.49 ± 0.2420:4(n-6) 0.25 ± 0.30 z 0.37 ± 0.45 z 3.74 ± 1.88 y 3.08 ± 0.68 y 3.84 ± 2.04 y20:5(n-3) 9.09 ± 2.04 z 8.26 ± 1.21 z 7.55 ± 2.62 zy 7.14 ± 1.30 zy 6.61 ± 2.19 z22:5(n-3) 3.15 ± 1.25 2.75 ± 0.70 3.56 ± 0.86 3.32 ± 1.09 3.78 ± 1.3022:6(n-3) 29.23 ± 6.49 27.85 ± 9.47 31.02 ± 7.03 28.27 ± 6.19 30.71 ± 6.64Saturated 32.93 ± 3.80 33.79 ± 4.73 31.59 ± 5.34 34.65 ± 5.37 29.94 ± 7.08Unsaturated 68.54 ± 7.20 67.20 ± 7.73 69.11 ± 8.26 66.11 ± 7.37 71.25 ± 7.87(n-6) PUFA 1.10 ± 0.42 1.25 ± 0.46 4.47 ± 1.81 3.95 ± 1.09 4.90 ± 2.19(n-3) PUFA 42.2 ± 7.74 39.65 ± 9.69 42.9 ± 7.77 38.98 ± 6.59 42.09 ± 7.64(n-3) PUFAa (mg/g tissue) 23.64 ± 4.23 23.29 ± 3.45 22.57 ± 2.45 25.60 ± 3.78


level (Tanaka et al. 1999). Therefore, the L-OH levels of dor-sal muscle in experiment 1 might not accurately reflect stress,and we have determined accurate stress levels on live fish us-ing L-OHs in higher activity tissues, such as the liver (Tanakaet al. 1999, 2002). In this study, the change of L-OH levelsas a stress marker was not correlated with LPO levels. Thisresult is in agreement with an experiment on Yellowtail Seri-ola quinqueradiata (also known as Buri) in which fish wereexposed to stress by artificial exhaustive exercise (Tanaka andNakamura 2012). Since LPOs are unstable and hazardous, they

are reduced to more stable and nonpoisonous products like L-OHs shortly after their formation. In other words, the LPO levelmay reflect current or differential peroxidative conditions andthe L-OH level may reflect historical or integrated peroxida-tive conditions during the recent past in the living body. Forthese reasons, the L-OH level is estimated to be superior to theLPO level as an index of peroxidative conditions in the livingbody. Therefore, changes in the L-OH and LPO levels in livefish were not always in agreement in the present and previousstudies.

TABLE 6. Change in fatty acid composition (%; mean ± SD) in the liver of Japanese Scad without feeding after capture (experiment 2). The values of eachparameter in the same row with different letters are significantly different (P < 0.05).

Fatty acids After catch After transport 0.5 d 8 d 20 d 33 d

14:0 1.75 ± 0.83 1.28 ± 0.55 1.57 ± 0.87 1.39 ± 0.78 1.18 ± 1.08 1.03 ± 0.5216:0 19.57 ± 9.22 z 24.81 ± 2.71 y 24.71 ± 2.11 y 26.53 ± 3.03 y 24.57 ± 4.19 y 20.59 ± 1.74 z18:0 10.11 ± 1.78 9.96 ± 0.53 10.73 ± 1.02 10.34 ± 1.11 8.43 ± 0.91 8.39 ± 1.8318:1(n-9) 10.29 ± 2.79 13.12 ± 4.96 15.15 ± 3.30 10.50 ± 2.34 11.87 ± 2.92 10.86 ± 1.9918:2(n-6) 1.59 ± 1.66 z 1.22 ± 0.85 z 0.88 ± 0.22 z 1.17 ± 0.82 z 3.26 ± 0.15 y 0.82 ± 0.73 z20:1(n-9) 1.23 ± 1.14 z 0.54 ± 0.41 y 0.95 ± 0.81 z 1.26 ± 1.01 z 1.62 ± 0.95 z 3.41 ± 3.60 x18:3(n-3) 0.70 ± 0.48 z 0.46 ± 0.16 z 0.53 ± 0.21 z 3.22 ± 4.39 y 1.43 ± 1.68 z 4.97 ± 3.67 y20:4(n-6) 3.97 ± 0.85 z 4.22 ± 0.80 z 3.87 ± 0.95 z 5.32 ± 1.74 y 5.13 ± 1.42 y 7.02 ± 2.78 x20:5(n-3) 9.49 ± 0.51 z 8.80 ± 0.61 z 7.84 ± 1.55 z 4.54 ± 0.57 y 4.97 ± 0.91 y 5.36 ± 1.38 y22:5(n-3) 4.31 ± 1.03 z 3.34 ± 0.33 z 3.37 ± 0.65 z 2.62 ± 0.42 y 3.45 ± 1.06 z 3.77 ± 1.39 z22:6(n-3) 37.00 ± 6.63 z 32.23 ± 4.83 z 30.40 ± 3.64 y 33.12 ± 7.00 z 34.08 ± 6.02 z 33.79 ± 5.59 zSaturated 29.67 ± 7.58 34.78 ± 2.40 35.44 ± 2.93 36.87 ± 4.08 33.00 ± 4.63 28.98 ± 2.87Unsaturated 71.38 ± 7.58 66.54 ± 2.75 65.94 ± 8.93 64.11 ± 4.95 66.54 ± 5.63 72.12 ± 3.10(n-6) PUFA 5.56 ± 2.07 5.44 ± 1.55 4.75 ± 0.98 6.49 ± 1.81 8.40 ± 4.50 7.84 ± 2.16(n-3) PUFA 51.5 ± 7.74 z 44.83 ± 5.04 y 42.14 ± 5.09 y 43.5 ± 5.39 y 43.93 ± 7.29 y 47.89 ± 4.76 z(n-3) PUFAa (mg/g tissue) 26.73 ± 2.77 22.88 ± 1.98 25.19 ± 2.01 25.44 ± 1.47 29.13 ± 2.05


Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3


TABLE 7. Change in free amino acid content (mg/100 g tissue; mean ± SD) in the dorsal muscle of Japanese Scad without feeding after capture (experiment1). The values of each parameter in the same row with different letters are significantly different (P < 0.05).

Amino acids After transport 3 d 9 d 33 d

Aspartic acid 1.78 ± 1.26 z 1.60 ± 1.13 z 2.35 ± 1.89 z 0.50 ± 0.39 ySerine 4.82 ± 8.93 z 2.27 ± 0.87 y 2.28 ± 1.01 y 1.09 ± 0.47 xGlutamic acid 9.62 ± 3.89 7.32 ± 3.54 10.31 ± 4.31 7.47 ± 3.14Glycine 6.88 ± 2.73 5.37 ± 2.40 7.62 ± 3.40 6.91 ± 2.90Histidine 371.52 ± 57.54 z 285.3 ± 70.44 z 338.09 ± 44.35 z 175.84 ± 62.15 yArginine 2.05 ± 1.81 2.36 ± 1.88 3.43 ± 1.45 1.95 ± 1.64Taurine 44.64 ± 19.77 z 39.34 ± 14.77 z 38.22 ± 25.80 z 22.19 ± 13.60 yThreonine 0.80 ± 0.59 1.30 ± 1.36 1.08 ± 1.34 0.30 ± 0.58Alanine 11.58 ± 5.74 z 10.64 ± 3.77 z 13.20 ± 4.33 z 5.83 ± 1.46 yProline 1.33 ± 0.64 z 1.25 ± 0.77 z 0.80 ± 0.40 z 0.33 ± 0.26 yGamma amino butyric acid nda nd nd ndCysteine 7.52 ± 7.09 z 4.71 ± 1.61 z 7.44 ± 1.88 z 2.80 ± 1.25 yTyrosine 1.36 ± 1.28 0.85 ± 0.29 1.34 ± 0.34 0.51 ± 0.23Valine 1.70 ± 1.16 2.14 ± 0.87 2.44 ± 0.46 1.20 ± 0.65Methionine 1.03 ± 0.53 0.93 ± 0.42 0.81 ± 0.24 0.18 ± 0.19Ornithine nd nd nd ndLysine 14.01 ± 7.10 z 7.90 ± 2.98 y 7.71 ± 3.11 y 4.80 ± 2.81 yIsoleucine 1.37 ± 0.52 z 1.40 ± 0.57 z 1.40 ± 0.28 z 0.53 ± 0.38 yLeucine 2.11 ± 0.73 z 2.16 ± 0.90 z 2.13 ± 0.47 z 0.76 ± 0.58 yPhenylalanine 1.17 ± 0.24 1.27 ± 0.37 1.14 ± 0.27 0.45 ± 0.23

Totals484.28 ± 91.81 z 378.12 ± 90.49 z 441.79 ± 52.55 z 233.64 ± 74.65 y

aThe term “nd” means levels were not detected.

The ratio of the fatty acid composition in the dorsal mus-cle during experiment 2 showed that 14:0, 16:0, 18:1(n-9), and20:5(n-3) fatty acids decreased and 22:6(n-3) fatty acid signifi-cantly increased during the holding period. When the fatty acidlevels are expressed in terms of milligrams of fatty acid pergram of tissue, the (n-3) fatty acids (such as 20:5 and 22:6) inthe dorsal muscle were constant for 8–9 d during holding inspite of a decrease in the percent of 20:5(n-3) fatty acid in ex-periment 2 (Tables 3, 4). However, the fatty acid composition inthe dorsal muscle did not change significantly in experiment 1.The change in fatty acid composition might be related to L-OHlevels or phospholipid composition. In experiment 2, the changein L-OH levels in the muscle and liver ranged from 6.02 ± 1.40to 15.14 ± 6.03 and from 10.48 ± 1.37 to 33.04 ± 2.17, respec-tively, which was higher than the change in experiment 1. Sincethe L-OHs were generated by the oxidation of fatty acids, thefatty acid composition in experiment 2 might have significantlychanged during holding. Moreover, the phospholipid composi-tion of the muscle and liver in experiment 2 was higher than thatin experiment 1. Generally, PUFAs contained in fish are easilyoxidized and exists as phospholipids more than as triglycerides.Therefore, fatty acid composition in experiment 2 might havesignificantly changed during holding. The lipid and (n-3) PUFAlevels in muscle affect the quality of fish as a raw food orfood material. The (n-3) PUFAs, such as DHA, in fish meat en-

hance umami taste. Many researchers report that oil rich in (n-3)PUFAs, such as tuna Thunnus spp. oil, suppresses the bitternessof quinine sulfate and enhances the umami components, suchas inosine 5’-monophosphate disodium salt and monosodiumglutamate monohydrate (Koriyama et al. 2002; Nakaya et al.2006; Thakur et al. 2009). Therefore, the maintenance of thelipid level and (n-3) PUFA content in the dorsal muscle for 8–9 d may have improved the quality of fish as a raw food or foodmaterial.

The quality of a marine product is determined not only bylipids but also by other components. In particular, FAAs havebeen implicated in the characteristic taste of seafood (Konosuet al. 1974; Komata 1990). In addition, FAAs play importantroles in physiological functions, such as osmoregulation andbuffering capacity in the tissues of aquatic animals (Love 1970,1980; Abe and Ohmama 1987; Van Waarde 1988). In this study,the change in total FAAs in the dorsal muscle was in agreementwith the changes in body weight, lipid content in the muscle, andtriglyceride-to-total-lipid ratio. Histidine in the dorsal muscleof Japanese Scad comprised 70% of the total FAAs. Therefore,the change in this amino acid affected the total FAA content.Compared with other fish and shellfish, Japanese Scad have aFAA profile similar to migratory high-speed swimmers such asChum Salmon Oncorhynchus keta, Pacific Chub Mackerel, andYellowfin Tuna Thunnus albacares, which also possess a very

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3

570 TANAKA ET AL.

TABLE 8. Change in free amino acid content (mg/100 g tissue; mean ± SD) in the dorsal muscle of Japanese Scad without feeding after capture (experiment2). The values of each parameter in the same row with different letters are significantly different (P < 0.05).

Amino acids After transport 0.5 d 8 d 20 d 33 d

Aspartic acid 3.17 ± 2.03 1.87 ± 1.98 2.63 ± 2.43 2.84 ± 2.00 1.34 ± 1.04Serine 1.54 ± 0.36 z 5.56 ± 7.60 y 1.29 ± 0.33 z 1.12 ± 0.52 z 1.37 ± 0.40 zGlutamic acid 4.79 ± 1.72 6.03 ± 1.58 8.51 ± 3.52 7.05 ± 2.41 5.70 ± 1.80Glycine 5.24 ± 1.39 z 4.79 ± 1.34 z 10.02 ± 3.30 y 8.38 ± 2.99 y 8.56 ± 3.42 yHistidine 217.11 ± 52.48 z 212.27 ± 41.22 z 273.51 ± 74.42 z 197.43 ± 38.70 z 139.14 ± 29.74 xArginine 0.42 ± 0.59 1.03 ± 1.40 0.63 ± 1.18 1.31 ± 1.42 0.39 ± 0.71Taurine 45.75 ± 15.64 z 30.60 ± 22.00 z 59.48 ± 48.55 z 43.25 ± 22.69 z 16.95 ± 3.57 yThreonine 0.24 ± 0.35 0.93 ± 1.69 0.86 ± 1.15 0.40 ± 0.71 1.78 ± 1.79Alanine 6.27 ± 1.57 7.35 ± 1.34 7.60 ± 2.45 8.60 ± 3.33 6.43 ± 1.59Proline 0.95 ± 0.37 1.34 ± 0.96 0.48 ± 0.21 0.42 ± 0.35 0.81 ± 0.58Gamma Amino butyric acid nda nd nd nd ndCysteine 4.88 ± 1.62 5.71 ± 0.74 6.00 ± 1.52 4.13 ± 1.33 4.30 ± 1.02Tyrosine 0.88 ± 0.29 1.03 ± 0.13 1.08 ± 0.28 0.75 ± 0.24 0.78 ± 0.18Valine 1.89 ± 0.35 1.86 ± 0.42 2.50 ± 0.51 1.86 ± 0.64 1.90 ± 0.62Methionine 0.75 ± 0.21 0.88 ± 0.21 0.87 ± 0.29 0.57 ± 0.35 0.96 ± 0.35Ornithine nd nd nd nd ndLysine 6.30 ± 2.02 z 9.92 ± 3.22 z 3.61 ± 2.69 y 5.56 ± 3.50 z 5.38 ± 1.91 zIsoleucine 1.09 ± 0.27 1.07 ± 0.18 1.31 ± 0.30 0.96 ± 0.32 1.02 ± 0.35Leucine 1.64 ± 0.36 1.60 ± 0.41 1.93 ± 0.41 1.43 ± 0.50 1.56 ± 0.56Phenylalanine 0.85 ± 0.19 1.01 ± 0.18 0.94 ± 0.14 0.63 ± 0.21 0.82 ± 0.26

Totals303.76 ± 56.64 zy 294.86 ± 52.58 zy 383.25 ± 95.76 z 286.68 ± 47.20 zy 198.40 ± 31.95 x

aThe term “nd” means levels were not detected.

high level of histidine in their white muscle (Abe and Ohmama1987; Abe et al. 1995; Hiraoka et al. 2007). The histidine-relateddipeptides in fish white muscle are metabolically inert comparedwith free histidine. It seems obvious that a good biologicalstrategy would be to maintain high levels of these dipeptides inwhite muscle in order to maintain the proton buffering capacity,whereas histidine could be used in case of emergencies suchas starvation (Abe and Ohmama 1987). Therefore, the histidinein the dorsal muscle of Japanese Scad in this study decreasedduring the 33-d period of holding because the fish used histidinewhen they were not getting food.

In summary, short-term holding without feeding after cap-ture decreased stress in captured fish following transport fromthe fishing point to the port. The lipid content in the dorsalmuscle did not change for 8–9 d but then decreased by day 33.The lipid content during holding without feeding did not in-crease significantly, although it did decrease, but not linearly; itwas constant for a certain period and then decreased. This resultwas also observed with regard to body weight, triglyceride-to-total-lipid ratio, and FAAs. The short-term holding without feed-ing may have an effect on not only fish caught by purse seine butalso on aquacultured fish. In aquaculture, feeding the fish leadsto an exhaustion of physical strength, an increase in oxygenconsumption, and an eutrophication of ocean space. In orderto prevent these problems, the fish could be cultured without

feeding, although the lack of feeding may lower the nutritionalvalue of the cultured fish. In this study, we confirmed that theshort-term holding without feeding after capture reduced oxida-tive stress but the nutritional components of the dorsal muscledid not change for 8–9 d. Therefore, in aquaculture suspendingfeeding for the short-term may provide an effective method forthe improvement of declining fish conditions.

In conclusion, to avoid the physiological effects of cumula-tive stress in captured fish and to obtain a higher quality fishproduct, fish caught in commercial fisheries, such as by purseseine, or maintained in aquaculture facilities, such as in large-scale marine cages, can be held without feeding for 8–9 d ratherthan shipped immediately if they cannot be killed at the timeof capture. Fish captured and held also have an economic ad-vantage because they can be sold at any time, even during poorweather conditions when additional fish cannot be captured. Asystem of holding fish without feeding after capture can be intro-duced with few initial costs because of the absence of feeding.

ACKNOWLEDGMENTSWe thank Ebisumaru (experiment 1) and Shouseimaru (ex-

periment 2) for facilitating shipboard observations and sam-pling, together with the owners, skippers, and crew members ofthe purse-seine vessels. This study was part of a Japanese na-tional project (project number 2002) “Quality Improvement of

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3


Coastal Fish and Marine Invertebrates—Achievement by Short-Term Holding and Associated Systems for Transportation andMarketing.” This study was funded by Research and Devel-opment projects for application in promoting new policy ofAgriculture Forestry and Fisheries.

REFERENCESAbe, H., P. W. Hochachka, and T. P. Mommsen. 1995. Histidine-related dipep-

tides: distribution, metabolism, and physiological function. Pages 309–333in H. Abe, P. W. Hochachka, and T. P. Mommsen, editors. Biochemistry andmolecular biology of fishes. Elsevier, New York.

Abe, H., and S. Ohmama. 1987. Effect of starvation and sea-water acclimationon the concentration of free l-histidine and related dipeptides in the muscleof eel, Rainbow Trout and Japanese Dace. Comparative Biochemistry andPhysiology Part B 88:507–511.

Ahimbisibwe, J., K. Inoue, T. Shibata, and T. Aoki. 2010. Effect of bleeding onthe quality of Amberjack Seriola dumerili and Red Sea Bream Pagrus majormuscle tissues during iced storage. Fisheries Science 76:389–394.

Bess, K. N., D. D. Boler, M. A. Tavarez, H. K. Johnson, F. K. McKeith,J. Killefer, and A. C. Dilger. 2013. Texture, lipid oxidation and sensorycharacteristics of ground pork patties prepared with commercially availablesalts. LWT - Food Science and Technology 50:408–413.

Folch, J., M. Lees, and G. H. S. Stanley. 1957. A simple method for the isolationand purification of total lipids from animal tissues. Journal of BiologicalChemistry 226:497–509.

Harman, D. 1981. The aging process. Proceedings of the National Academy ofSciences of the USA 78:7124–7128.

Hiraoka, Y., Y. Sasaki, T. Takai, A. Kawakubo, S. Miyoshi, and S. Nakajima.2007. Fortification of niboshi with histidine prepared from katsuobushi pro-tein digest. Fisheries Science 73:1383–1387.

Ichihara, K. I., and Y. Fukubayashi. 2010. Preparation of fatty acid methyl estersfor gas-liquid chromatography. Journal of Lipid Research 51:635–640.

Jimenez-Colmenero, F., M. Triki, A. M. Herrero, L. Rodrıguez-Salas, and C.Ruiz-Capillas. 2013. Healthy oil combination stabilized in a konjac matrixas pork fat replacement in low-fat, PUFA-enriched, dry fermented sausages.LWT - Food Science and Technology 51:158–163.

Komata, Y. 1990. Umami taste of seafoods. Food Reviews International 6:457–487.

Konosu, S., K. Watanabe, and T. Shimizu. 1974. Distribution of nitrogenousconstituents in the muscle extracts of eight species of fish. Bulletin of theJapanese Society of Scientific Fisheries 40:909–915.

Koriyama, T., S. Wongso, K. Watanabe, and H. Abe. 2002. Fatty acid com-positions of oil species affect the 5 basic taste perceptions. Journal of FoodScience 67:868–873.

Kramer, J. K. G. 1997. A rapid method for the determination of vitamin Eforms in tissues and diet by high-performance liquid chromatography usinga normal-phase diol column. Lipids 32:323–330.

Love, R. M. 1970. Depletion. Pages 222–257 in R. M. Love, editor. Chemicalbiology of fishes. Academic Press, New York.

Love, R. M. 1980. Feeding and starving. Pages 166–229 in R. M. Love, editor.Chemical biology of fishes. Academic Press, New York.

Machii, T., and Y. Nose. 1990. Method to determine the mesh size of a purseseine and its application to Peruvian Jack Mackerel. Nippon Suisan Gakkaishi56:7–10.

Marcalo, A., L. Mateus, J. Correia, P. Serra, R. Fryer, and Y. Stratoudakis. 2006.Sardine (Sardina pilchardus) stress reactions to purse seine fishing. MarineBiology 149:1509–1518.

Marcalo, A., P. Pousao-Ferreira, L. Mateus, J. H. Duarte Correia, and Y.Stratoudakis. 2008. Sardine early survival, physical condition and stress afterintroduction to captivity. Journal of Fish Biology 72:103–120.

Martinez-Alvarez, R., A. E. Morales, and A. Sanz. 2005. Antioxidant defensesin fish: biotic and abiotic factors. Reviews in Fish Biology and Fisheries15:75–88.

Millipore. 1993. Waters AccQ-Tag chemistry package instruction manual. Mil-lipore Corporation, Milford, Massachusetts.

Moriya, H., T. Kuniminato, M. Hosokawa, K. Fukunaga, T. Nishiyama, andK. Miyashita. 2007. Oxidative stability of salmon and herring roe lipids andtheir dietary effect on plasma cholesterol levels of rats. Fisheries Science73:668–674.

Nakaya, K., T. Kohata, N. Doisaki, H. Ushio, and T. Ohshima. 2006. Effect ofoil droplet sizes of oil-in-water emulsion on the taste impressions of addedtastants. Fisheries Science 72:877–883.

Nakamura, T., and H. Maeda. 1991. A simple assay for lipid hydroperoxidesbased on triphenylphosphine oxidation and high-performance liquid chro-matography. Lipids 26:765–768.

O’ Dwyer, S. P., D. O’ Beirne, D. Nı Eidhin, A. A. Hennessy, and B. T. O’Kennedy. 2013. Formation, rheology and susceptibility to lipid oxidation ofmultiple emulsions (O/W/O) in table spreads containing omega-3 rich oils.LWT - Food Science and Technology 51:484–491.

Sevanian, A., S. F. Muakkassah-Kelly, and S. Montestruque. 1983. The influenceof phospholipase A2 and glutathione peroxidase on the elimination of mem-brane lipid peroxides. Archives of Biochemistry and Biophysics 223:441–452.

Sigholt, T., U. Erikson, T. Rustad, S. Johansen, T. S. Nordtvedt, and A. Seland.1997. Handling stress and storage temperature affect meat quality of farmed-raised Atlantic Salmon (Salmo Salar). Journal of Food Science 62:898–905.

Tanaka, R., H. Hatate, M. Ito, and T. Nakamura. 2006. Elevation of lipid peroxidelevel and production of hydroxy lipids in cultured Hepa-T1 cells by oxidativestressors. Fisheries Science 72:665–672.

Tanaka, R., Y. Higo, H. Murata, and T. Nakamura. 1999. Accumulation ofhydroxy lipids in live fish with oxidative stress. Fisheries Science 65:796–797.

Tanaka, R., Y. Higo, T. Shibata, N. Suzuki, H. Hatate, K. Nagayama, and T.Nakamura. 2002. Accumulation of hydroxy lipids in live fish infected withfish diseases. Aquaculture 211:341–351.

Tanaka, R., K. Naiki, K. Tsuji, H. Nomata, Y. Sugiura, T. Matsushita, andI. Kimura. 2013a. Effect of antioxidative treatments on lipid oxidation inskinless fillet of Pacific Saury Cololabis saira in frozen storage. Journal ofFood Processing and Preservation 37:325–334.

Tanaka, R., and T. Nakamura. 2012. Effects of exhaustive exercise on lipidperoxide and hydroxy lipids in Yellowtail. North American Journal of Aqua-culture 74:164–168.

Tanaka, R., Y. Sugiura, and T. Matsushita. 2013b. Simultaneous identi-fication of 4-hydroxy-2-hexenal and 4-hydroxy-2-nonenal in foods bypre-column fluorigenic labeling with 1,3-cyclohexanedione and reversed-phase high-performance liquid chromatography with fluorescence detec-tion. Journal of Liquid Chromatography and Related Technologies 36:881–896.

Thakur, D. P., K. Morioka, N. Itoh, M. Wada, and Y. Itoh. 2009. Muscle biochem-ical constituents of cultured amberjack Seriola dumerili and their influenceon raw meat texture. Fisheries Science 75:1489–1498.

Van Waarde, A. 1988. Biochemistry of non-protein nitrogenous com-pounds in fish including the use of amino acids for anaerobic energyproduction. Comparative Biochemistry and Physiology Part B 91:207–228.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:41

27

Oct

ober

201

3



Phenotypic Comparisons among Natural-Origin,Hatchery-Origin, and Captive-Reared Female SpringChinook Salmon from the Tucannon River, WashingtonMichael P. Gallinat a & Wan-Ying Chang a ba Washington Department of Fish and Wildlife , 600 Capitol Way North, Olympia ,Washington , 98501-1091 , USAb 4201 Wilson Boulevard, Suite 965, Arlington , Virginia , 22230 , USA

To cite this article: Michael P. Gallinat & Wan-Ying Chang (2013) Phenotypic Comparisons among Natural-Origin, Hatchery-Origin, and Captive-Reared Female Spring Chinook Salmon from the Tucannon River, Washington, North American Journal ofAquaculture, 75:4, 572-581






http://dx.doi.org/10.1080/15222055.2013.837126




ARTICLE

Phenotypic Comparisons among Natural-Origin,Hatchery-Origin, and Captive-Reared Female SpringChinook Salmon from the Tucannon River, Washington

Michael P. Gallinat* and Wan-Ying Chang1

Washington Department of Fish and Wildlife, 600 Capitol Way North, Olympia,Washington 98501-1091, USA

AbstractWe examined the effects of hatchery rearing on FL, weight, egg size, fecundity, relative fecundity, and reproductive

mass of female spring Chinook Salmon Oncorhynchus tshawytscha from a population that had been in captivity for 0(natural-origin), 18 (hatchery-origin), and 48 (captive-reared broodstock) months. Age-4 captive-reared broodstockfemales that were reared for their entire life in the hatchery environment had significantly lower mean FL, weight,fecundity, relative fecundity, and reproductive mass, but had significantly larger eggs than age-4 females from theother groups after correcting for body size. Hatchery-origin females had significantly lower fecundity than natural-origin fish. Our findings illustrate a phenomenon of lower overall reproductive potential for hatchery-reared fish inthe form of reduced fecundity that decreases as time spent in the hatchery environment increases. We also observedthat progeny of captive-reared broodstock parents, released as smolts and recaptured as returning age-4 adults, havea size and fecundity distribution that is similar to the hatchery-origin adults, suggesting that the decrease in fecunditywas not a genetically linked trait.

Considerable controversy exists over the use of hatchery sup-plementation programs due to the potential for increased risksof adverse effects to the natural fish population (see reviews byWaples 1991; Brannon et al. 2004; Araki et al. 2008; Kostow2009). One concern is that the hatchery environment exposesfish to different developmental and evolutionary forces or do-mestic selection that may shape their phenotype (Fleming et al.1994). This may change the direction of selection and cause ge-netic divergence from the wild population (Lynch and O’Hely2001; Ford 2002). In most organisms, progeny phenotypes tendto be influenced more by the genotype or environment of theirmother than by the genotype or environment of their father(Heath et al. 1999). This large effect of maternal (relative topaternal) genotype or environment is referred to as a maternaleffect, or when mediated by maternal environmental conditions,as an inherited environmental effect (Heath et al. 1999).

*Corresponding author: [email protected] Address: 4201 Wilson Boulevard, Suite 965, Arlington, Virginia 22230, USAReceived May 20, 2013; accepted August 13, 2013

Studies that have examined hatchery environmental effectson salmonids have tended to focus on Coho Salmon On-corhynchus kisutch and steelhead O. mykiss (Swain et al. 1991;Kostow 2004; Campbell et al. 2006). Knudsen et al. (2008)compared reproductive traits of wild-origin female springChinook Salmon O. tshawytscha with first-generation hatchery-origin females to determine whether their reproductive traitshad diverged after a single generation of artificial propagation.Their findings suggested that a single generation of conservationhatchery propagation using wild broodstock produces hatcheryfish with reproductive traits similar to wild fish, given compara-ble body size (Knudsen et al. 2008). However, most integratedhatchery programs use both hatchery and natural-origin brood-stock, not 100% wild broodstock, so the findings from Knudsenet al. (2008) would probably not apply to the majority of hatch-ery programs.

572

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3

PHENOTYPIC COMPARISONS AMONG FEMALE CHINOOK SALMON 573

FIGURE 1. Tucannon River spring Chinook Salmon groups used for phenotypic comparisons. Possible spawning crosses are shown in brackets. Spawningcrosses for fish spawned in the hatchery environment are listed in order of priority.

Our study examined the environmental effects of hatcheryrearing on FL, weight, egg size, fecundity, relative fecundity,and reproductive mass on female spring Chinook Salmon thathad been in captivity for 0, 18, and 48 months. This allowedfor the examination of specific phenotypic traits that may beexpressed and selected for in the hatchery environment. Specif-ically, studies conducted on populations that have a commongenetic background provide a chance to understand the mech-anisms behind changes caused by the hatchery environment.The three reproductive groups in the Tucannon River of Wash-ington State are described as follows: (1) natural-origin fishthat were the product of natural reproduction of natural- andhatchery-origin fish spawning in the Tucannon River and werein captivity for 0 months; (2) hatchery-origin fish used for sup-plementation stocking that were the product of artificial repro-duction in a hatchery, but were released after 18 months asyearling smolts and trapped as returning adults and whose par-ents represented approximately a 50:50 mixture of hatchery-and natural-origin adults trapped in the Tucannon River; and(3) captive-reared broodstock that represented a subsample ofthe hatchery-origin group that were not released but, instead,reared to sexual maturity in captivity (48 months). We focusedon the single-generation effects of the hatchery environment onthe phenotypic expression of size and reproductive (e.g., egg sizeand fecundity) traits rather than the multigenerational effects ofartificial propagation. We also compared the phenotypic traitsbetween a sample of age-4 females identified as captive-rearedbroodstock progeny and a sample of age-4 hatchery-origin fe-males that returned in 2008. Both groups were released at sim-ilar sizes after 18 months of hatchery rearing and differed onlyin parentage. This comparison was to determine whether bothgroups had similar phenotypic traits or if there was evidence of

phenotypic divergence. The groups used for phenotypic com-parisons are described in Figure 1.

METHODSStudy population.—The Tucannon River is a third-order

stream in southeastern Washington that flows into the SnakeRiver between Little Goose and Lower Monumental dams ap-proximately 622 river kilometers (rkm) from the mouth of theColumbia River (Figure 2). Spring Chinook Salmon adults mi-grate to the Tucannon River basin in the spring and spawn duringthe early fall. The adults generally spawn and the juveniles rearupstream from rkm 35 in the river. Natural-origin smolts leav-ing the system are about 18 months old, have a mean FL of105–113 mm, and rear in the ocean for 1–3 years until mature.

The Tucannon River spring Chinook Salmon populationsteadily declined after the construction and operation ofthe federal Columbia and Snake river hydropower system(USACE 1975; Nehlsen et al. 1991). The decline has beenattributed to mortalities of adults and juveniles during migra-tion through four hydropower dams on the Columbia River andtwo hydropower dams on the Snake River (USACE 1975), andhabitat loss or degradation in the Tucannon River along withother environmental factors such as variable ocean conditions,drought, and floods (Columbia Conservation District 2004). Thepopulation is currently listed as “threatened” under the U.S. En-dangered Species Act as part of the Snake River Spring/SummerChinook Salmon evolutionary significant unit (March 25, 1999;FR 64(57):14517–14528).

In 1985, the Washington Department of Fish and Wildlife(WDFW) began a spring Chinook Salmon hatchery supplemen-tation program in the Tucannon River by trapping wild endemic

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3

574 GALLINAT AND CHANG

FIGURE 2. Location of the Tucannon River, adult salmon trap, and Lyons Ferry Hatchery Complex facilities within the Snake River Basin, Washington.

adults for broodstock. The first hatchery smolts were releasedin 1987. Since 1989, the hatchery broodstock has consisted ofboth natural- and hatchery-origin fish (Table 1).

The hatchery program is a fully integrated conservation pro-gram, designed to allow gene flow between the hatchery andnatural components both in the hatchery and on the spawninggrounds. Recent genetic analysis looking at 14 microsatelliteloci (13 coast-wide Genetic Analysis of Pacific Salmon [GAPS]loci plus Ssa-197) found that the genetic diversity of springChinook Salmon in the Tucannon River has not significantlychanged as a result of the hatchery supplementation or captivebrood programs (Kassler and Dean 2010).

In 1994, the total adult escapement declined severely to fewerthan 150 fish, and in 1995 was estimated at 54 fish (Table 1). TheWDFW and tribal comanagers determined the risk of extinctionwas high enough to warrant aggressive intervention beyond theexisting hatchery supplementation program in the form of acaptive broodstock program. Captive broodstock programs dif-

fer from conventional hatchery supplementation programs inthat fish are held in the hatchery environment throughout theirlife to ensure a readily available gamete source.

With the two hatchery programs operating concurrently wewere able to examine the effects of different levels of hatcheryrearing on the same stock. “Natural” is used throughout thispaper to describe fish that are progeny of parents spawned andreared in a natural environment, regardless of the origin of theparents (Figure 1).

Hatchery operations.—Tucannon Fish Hatchery is located atrkm 59 on the Tucannon River and uses an adult trap to collectbroodstock from throughout the run. Broodstock are transportedto the Lyons Ferry Hatchery, located on the Snake River at itsconfluence with the Palouse River in southeastern Washington(Figure 2). Lyons Ferry Hatchery is used for broodstock hold-ing and spawning, egg incubation, and early life rearing. Wellssupply constant temperature (11◦C) water to the hatchery. Af-ter juveniles are coded-wire-tagged at Lyons Ferry Hatchery,

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3


TABLE 1. River escapement of natural-origin, hatchery-origin, and captive-reared broodstock (CRB) progeny of Tucannon River spring Chinook Salmon andbroodstock collected from the adult trap and spawned for the 1985–2008 run years. The last column shows numbers of fish spawned in the captive-reared broodstockprogram. Numbers represent a combination of sexes and brood years.

River escapement by origin Broodstock spawned by origin

Run year Natural Hatchery CRB progeny Total run Natural Hatchery CRB progeny CRB program

1985 591 0 0 591 8 0 0 01986 636 0 0 636 91 0 0 01987 582 0 0 582 83 0 0 01988 410 19 0 429 90 0 0 01989 336 109 0 445 55 67 0 01990 494 260 0 754 30 32 0 01991 260 268 0 528 40 31 0 01992 418 335 0 753 37 45 0 01993 317 272 0 589 45 42 0 01994 98 42 0 140 35 34 0 01994 21 33 0 54 9 30 0 01996 165 85 0 250 33 42 0 01997 160 191 0 351 38 51 0 01998 85 59 0 144 45 41 0 01999 3 242 0 245 3 118 0 02000 82 257 0 339 8 65 0 202001 718 294 0 1,012 52 52 0 2492002 350 655 0 1,005 42 51 0 2042003 248 196 0 444 41 34 0 3452004 400 170 3 573 48 40 0 3472005 289 117 14 420 47 46 2 2002006 140 109 4 253 36 52 0 862007 198 127 19 344 51 28 3 02008 534 417 240 1,191 40 35 39 0

they are transferred to Tucannon Fish Hatchery to rear throughthe winter on a mixture of well, spring, and river water beforebeing transferred to the Curl Lake Rearing Pond (Figure 2) toacclimate. Fish are acclimated in river water for 6 to 8 weeks inthe spring before being volitionally released. Hatchery fish werereared according to the comanagers Salmonid Disease ControlPolicy and Integrated Hatchery Operation Team fish health pol-icy (Peck 1993; Watson 1996).

The hatchery supplementation broodstock goal was for up to100 adults trapped from the river composed of both natural- andhatchery-origin returns (1:1 ratio). Natural- and hatchery-originfish are used in the broodstock for two reasons: to achieve thehatchery production goal without excessively reducing the abun-dance of natural spawning fish, and to decrease the possibilityof inadvertently creating separate populations of the TucannonRiver spring Chinook Salmon population through a steady in-fusion of naturally produced, endemic adults. Returning hatch-ery fish used in the hatchery supplementation broodstock wereverified to have come from the Tucannon River population byreading each fish’s coded wire tag after it was spawned.

The captive broodstock was started by collecting 80 sac fryfrom 15 family groups (1,200 fish total) from the hatchery

supplementation program for five brood years between 1997and 2001. Each of the family groups was subsequently reducedto 30 fish per family (450 fish per brood year) after the first yearof rearing. Captive broodstock were tagged by family groupwith an alphanumeric visible implant tag placed behind the eyeof each fish and a coded wire tag in its snout. Tags were usedto verify family groups during subsequent spawning in orderto prevent full- or half-sibling matings. After tagging, the cap-tive brood families were combined by brood year for rearing.Complete details on the collection and rearing procedures forthe captive brood and hatchery supplementation programs areprovided in Gallinat et al. (2009).

The captive broodstock were reared outdoors at Lyons FerryHatchery under natural photoperiod conditions. During lateJune and early July, captive-reared broodstock age-2 or olderwere examined for signs of sexual maturation. Sexually ma-ture captive-reared broodstock were transported to broodstockholding raceways in common with, but separated by screens,from broodstock (hatchery- and natural-origin) collected fromthe Tucannon River.

During spawning, the total number of eggs from two fe-males were divided into two groups and fertilized by two males

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3


following a 2 × 2 factorial spawning matrix approach. Due tothe relatively small size of the population, this mating strategywas used to increase the effective population size and to main-tain genetic diversity (Busack and Knudsen 2007). The priorityon each spawn day was for natural-origin fish to be crossed withhatchery-origin and captive-reared broodstock.

Data collected from spawned fish included age at maturityfrom scale samples (natural-origin) or coded wire tag informa-tion (hatchery-origin and captive-reared broodstock), FL, andweight (combined gonad and somatic tissue). Acetate impres-sions were made from collected scale samples and aged at theWDFW Scale Lab by experienced personnel. Ages were deter-mined as the number of years from fertilization (brood year) tospawn year.

Heath et al. (1999) found a highly significant effect of ma-ternal (but not paternal) size on larval Chinook Salmon bodysize at 45 d postfertilization. After 45 d postfertilization, theeffect of maternal body size (relative to paternal size) began todecrease (Heath et al. 1999). At the eyed egg stage (26–28 dpostfertilization) in our study, eggs were shocked, water wasdrained from the egg mass, and dead eggs were counted andremoved. Based on Heath et al. (1999) we assumed that differ-ences in egg size at 26–28 d postfertilization are attributed tomaternal effects. A random sample of live eggs collected witha 100-count egg counter (www.marisource.com) was weighedand the mean weight per egg was used to define egg size. Thetotal number of live eggs was estimated using the total weightof live eggs divided by egg size. This estimate was decreasedby 4% to compensate for water adherence to the eggs (WDFWSnake River Lab, unpublished data). The total numbers of liveand dead eggs were combined to estimate fecundity. Partiallyspawned fish were excluded from our data set.

Relative fecundity was calculated by dividing fecundity bybody weight (kg) (Knudsen et al. 2008). Relative fecundity wasused to correct for the effect of body size on the number of eggsproduced by each female.

Female salmon may allocate similar amounts of reproduc-tive effort but partition it differently (e.g., small eggs and highfecundity may be equal in energy expenditure to large eggs andlow fecundity). To account for differences in fecundity causedby egg size, reproductive mass was calculated by multiplyingfecundity by egg size to provide total reproductive contributionin grams.

Any phenotypic differences observed among the two groupsof hatchery fish may also be expressed in their progeny as aheritable trait, or may simply be a result of the length of timethat fish are exposed to the hatchery environment. The progenyfrom both the hatchery-origin and captive-reared broodstockprograms were reared in similar environments and at similarrearing densities, and were the same size at release (Gallinatet al. 2009). All juveniles released from both programs weretagged by group with coded wire tags, but were not fin-clippedto prevent their inclusion in the sport fishery. Based on 1985–2004 brood year coded wire tag recoveries, the sport, commer-

cial, and treaty ceremonial harvest combined accounted for lessthan 6% of the adult fish recovered (Gallinat and Ross 2009).Therefore, any observed phenotypic differences between thegroups in this study should not have been caused by selec-tive fishing mortality. In 2008, age-4 female progeny from thehatchery-origin and captive-reared broodstock programs (2004brood year) were collected and examined in a similar manner aspreviously described.

Statistical analysis.—Six phenotypic traits—FL, weight, eggsize, fecundity, relative fecundity, and reproductive mass—werecompared among the three groups of varying levels of hatcheryrearing using the dominating portion of returning females (age4) where sufficient data were available for all groups. Hatchery-origin and natural-origin groups had at least five observationsper group for each year from 2001 to 2008 while captive-rearedbroodstock data were available from 2001 to 2006. The three-group comparisons are based on data collected during 2001–2006. We expected environmental factors, such as ocean condi-tions and weather, could potentially affect the phenotypic traitsand that the magnitude of the effect was likely to vary fromyear to year. For each group we initially performed multiplecomparisons with P-values adjusted by permutation resamplingto test for differences of means among years for each variable.The results showed no clear patterns or significant trends forall variables by group. We proceeded to compare each traitamong groups using a linear mixed model containing a singlefixed group effect and a random year effect. The mixed modelapproach takes into account the clustering nature of the data.Samples collected for a given year are likely to experience asimilar level of environmental influence. The model equation isexpressed as Yijk = µ + αi + bij + eijk, where µ, αi representunknown fixed intercept and group effects respectively, and bij

and eijk are random variables representing the year effect anderror, respectively.

Theoretically, morphometric variables such as fecundity andbody size measurements are closely related due to allomet-ric growth. It is also expected that both body size and fe-cundity traits were affected by hatchery rearing (Thorpe 2004;Campbell et al. 2006). We further developed two linear mixedmodels in which egg size and fecundity were the response vari-ables. In each of these mixed models, the fixed effects includednot only the group effect, but also relevant phenotypic traits ascovariates, and the year effect remained random. These modelsexplored the relationship between fecundity traits and body size,and revealed additional hatchery-rearing effects on the fecun-dity traits that cannot be explained through the affected bodysize. Fork length, weight, and fecundity were log transformed toimprove linearity assumptions. The model-building process ini-tially included all covariates for fixed effects and then removedthose, except for group effect, that were not significant by typeIII F-tests. Interaction terms between group and each remainingcovariate were tested one at a time by the order of the magnitudeof the type III F statistics. Those significant by type III F-testwere added to determine the final models. Model fitting was

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3


done using SAS version 9.1.3, procedure MIXED with REMLas the estimating method (SAS 2004).

High correlations were expected and observed (r > 0.88)among fecundity, weight, and FL. To avoid collinearity prob-lems in model building, the normalization technique by Lleonartet al. (2000) was applied to the above variables.

We applied the technique by keeping weight unchanged andhaving FL and fecundity normalized according to a standardvalue of weight. A standard value of weight of 2,915 g was cho-sen by comparing 95% CIs of mean weight for each group andselecting the midpoint of the intersection portion of those CIs.The normalized FL and fecundity have the influence of weightremoved and retain the unique individual shape deviation.

In 2008, a sample of age-4 females was identified as progenyof the captive-reared broodstock. We compared the six pheno-typic traits of that group with a sample of age-4 hatchery-originfemales that returned in 2008. We performed two-sample t-testsfor differences in means with P-values adjusted by permutationresampling for multiplicity of tests.

RESULTS

Phenotypic Trait ComparisonThe sample sizes, means, and SEs of the data collected for the

selected phenotypic traits are presented in Table 2. The differ-ences of FL and weight among the three groups are illustrated inFigure 3. The linear mixed models containing group as the onlyfixed effect yielded significant results for all six traits. Overallgroup effect for FL, weight, fecundity, and reproductive masswas highly significant (P = 0.000). The group effect was alsosignificant for egg size and relative fecundity (P = 0.011 and0.002, respectively). Comparisons of means between each pairof groups are summarized in Table 3. Mean values of all six traitsof the captive-reared broodstock group were significantly differ-ent from means of the hatchery-origin and natural-origin groups.The captive-reared broodstock group had smaller means for alltraits except egg size. Differences between the hatchery-originand natural-origin groups were mostly not significant except

FIGURE 3. Relationship between log (weight) and log (FL) for age-4 Tucan-non River hatchery-origin (black circles), natural-origin (gray triangles), andcaptive-reared broodstock (white squares) spring Chinook Salmon females.

for fecundity. Natural-origin females had higher fecundity, onaverage, than did hatchery-origin females.

Fecundity and Egg Size Related to Other Size MeasuresResults of the two final models including other phenotypic

traits as covariates are summarized in Table 4. Egg size was posi-tively associated with weight and negatively associated with nor-malized FL and normalized fecundity. The group effect was stillsignificant with the inclusion of the covariates. Given the samesize, shape, and fecundity values, the captive-reared broodstockgroup had larger eggs than did the hatchery-origin and natural-origin groups. The difference between the hatchery-origin andnatural-origin groups was not significant and none of the interac-tion terms were significant. Fecundity was positively associatedwith weight and negatively associated with egg size (P = 0.000).Given the same weight and egg size, there was no significantdifference in fecundity among groups. Since normalized FL was

TABLE 2. Sample size (n), mean, and SE for the selected phenotypic traits for age-4 captive-reared broodstock (CRB), hatchery-origin (HOR), and natural-origin(NOR) female Tucannon River spring Chinook Salmon. Collection years are shown for each group.

CRB HOR NOR2001–2006 2001–2008 2001–2008

Phenotypic trait n Mean SE n Mean SE n Mean SE

FL (cm) 708 52.8 0.19 135 69.0 0.32 141 70.6 0.32Weight (kg) 708 2.17 0.02 135 3.53 0.06 141 3.90 0.05Fecundity (eggs/female) 708 1,664 20.05 135 2,982 56.20 141 3,419 56.08Egg size (g) 708 0.256 0.002 135 0.231 0.003 141 0.230 0.002Relative fecundity (eggs/kg) 708 779.3 6.24 135 846.1 10.21 141 879.6 9.70Reproductive mass (g) 708 425.9 5.78 135 687.3 14.46 141 787.4 13.86

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3


TABLE 3. Results from mixed models containing a group effect and a random year effect for age-4 female captive-reared broodstock (CRB), hatchery-origin(HOR), and natural-origin (NOR) Tucannon River spring Chinook Salmon (2001–2006 data). Type III tests of fixed effects are significant for all models (anasterisk denotes contrast of group means is significant at the 0.05 level).

Comparisons between group means: F statistics (P-value)

Phenotypic trait CRB versus HOR CRB versus NOR HOR versus NOR

FL (cm) 209.85 (0.000)* 258.51 (0.000)* 1.43 (0.247)Weight (kg) 49.64 (0.000)* 77.34 (0.000)* 2.28 (0.150)Fecundity (eggs/female) 60.53 (0.000)* 115.04 (0.000)* 6.84 (0.018)*Egg size (g) 6.05 (0.025)* 9.73 (0.006)* 0.30 (0.588)Relative fecundity (eggs/kg) 4.69 (0.043)* 17.01 (0.001)* 2.77 (0.106)Reproductive mass (g) 33.84 (0.000)* 62.27 (<0.000)* 3.36 (0.085)

removed from the model, the data suggested that fecundity wasmore closely related to weight than length.

Comparison of Progeny from Captive-Reared andHatchery Programs

The ranges of both fecundity and length of age-4 captive-reared broodstock progeny females were more similar to age-4hatchery-origin females than to the captive-reared broodstockfemales that they were derived from (Figure 4). Results of thetwo-sample t-tests with adjustment for multiple testing showedno significant difference in mean values of all six traits betweenthe two groups. Descriptive statistics of the data are given inTable 5.

DISCUSSIONDecreased body size has been associated with decreased fe-

cundity, smaller eggs, lower reproductive success, and lower sur-vival of progeny (Kostow 2009). However, despite the smallersize on average, the age-4 captive-reared Chinook Salmonbroodstock females had significantly larger eggs even after ac-counting for size, shape, and fecundity. Fleming and Gross(1992) reported that they also found hatchery-reared CohoSalmon had larger eggs than did wild females. In contrast, re-search by Heath et al. (2003) found that hatchery rearing relaxes

natural selection favoring large eggs, allowing fecundity selec-tion to drive rapid evolution of small eggs. They stated that thesesmall eggs could lead to reduced survival and limit the successof hatchery programs. However, Heath et al. (2003) may haveincorrectly attributed an ocean environmental effect and femalevariation on egg size to a genetic change as a result of hatch-ery enhancement (Beacham 2003; Fleming et al. 2003). Thebroodstock they studied was also developed to satisfy a “niche”market, and matures at a much smaller size and has unusuallysmall eggs (Beacham and Murray 1993; Beacham 2003).

Egg size can have important fitness consequences, so thereis a selective advantage for producing large eggs even withinthe hatchery environment (Heath et al. 1999). Kinnison et al.(2001) also found that egg size is strongly correlated with initialoffspring fry size in salmonids and offspring size is, in turn,correlated with survival in salmon. Large egg size was insuffi-cient to compensate for other deficiencies and did not appearto increase survival in our study since mortality to eye-up was49% for captive-reared broodstock eggs compared with hatch-ery eye-up mortalities of 4% and 3% for hatchery-origin andnatural-origin fish, respectively. The high egg mortality from thecaptive-reared broodstock group may be related to environmen-tal, physiological, dietary, or other unknown factors. Pattersonet al. (2004) found that captive Sockeye Salmon O. nerka also

TABLE 4. Summary of final models for egg size and fecundity for age-4 female captive-reared broodstock (CRB), hatchery-origin (HOR), and natural-origin(NOR) Tucannon River spring Chinook Salmon (2001–2006 data). The directions of fixed effects are summarized using plus “ + ” and minus “−” symbols, whichrefer to the signs of the slopes in the regression (adj: denotes variables are normalized to the standard weight).

Type III F-test

Response Fixed effects Direction F statistic (P-value) Significant contrast

Egg size (g) Group CRB > HOR > NOR 12.35 (0.000) CRB > HOR, CRB > NORWeight + 108.75 (0.000)FL, adj − 7.40 (0.007)Fecundity, adj − 659.89 (0.000)

Log (fecundity) Group No difference 1.74 (0.190)Log (weight) + 2292.76 (0.000)Log (egg size) − 565.89 (0.000)

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3


0

1000

2000

3000

4000

35 45 55 65 75 85

Fork length (cm)

Fec

undi

ty (

eggs

/fem

ale)

FIGURE 4. Comparison of FL and fecundity of age-4 captive-reared brood-stock progeny females (black circles) and age-4 hatchery-origin (white circles)Tucannon River spring Chinook Salmon females that returned to the TucannonRiver in 2008. Captive-reared broodstock females (gray squares) from the 2004brood year that were the parents of the age-4 captive-reared broodstock progenyare included for comparison.

had significantly lower mean egg survival than did natural-originfish and attributed it to confinement stress.

Tucannon River spring Chinook Salmon migrate 622 kmfrom the mouth of the Columbia River to the mouth of the Tu-cannon River (Marvin and Nighbor 2009). This long migrationmay explain the difference in egg size between the migratingfish (natural-origin and hatchery-origin) and the nonmigrating,captive-reared broodstock group. A reduction in ovary invest-ment due to migration costs has been noted in several salmonspecies (Kinnison et al. 2001; Campbell et al. 2006). The en-ergy costs of migration are expected to be reflected by a variationin egg size, since eggs continue to gain mass until just prior tospawning and egg number is determined well in advance of finalmaturation (Kinnison et al. 2001). Beacham and Murray (1993)and Healey (2001) suggested that a limited amount of energy isexpended on egg production in more northern stocks and stockswith long freshwater migrations. Thus, because of a readilyavailable food supply and protected hatchery environment, thecaptive-reared broodstock may be able to allocate more energyinto producing larger eggs than the migrating natural-origin andhatchery-origin fish.

We found that hatchery rearing of Tucannon River springChinook Salmon resulted in a phenomenon of lower over-all reproductive potential in the form of reduced fecunditythat decreased further as time spent in the hatchery environ-ment increased. Because the fish in our study had geneticallysimilar backgrounds, but were reared under different growth–environmental conditions, the differences observed were mostprobably environmentally induced. Hard et al. (2000) notedthat morphometric development in hatchery fish is highly plas-tic and probably stems from differences between the hatcheryand wild environment. The question remains whether the re-duced fecundity attributed to hatchery rearing is inherited inTucannon River spring Chinook Salmon. Our limited data, af-ter adjusting for multiple testing, has provided some evidencethat suggests fecundity was not significantly different betweenthe captive-reared broodstock progeny and hatchery-origin fish.The captive-reared broodstock progeny that were reared in asimilar manner to the hatchery-origin fish were more similar inphenotypic attributes to their hatchery cohorts than their captivebroodstock parents. We hypothesize that the offspring of thehatchery-origin females that spawn in the natural environmentwould also follow the phenotypic pattern of the natural-originfish, although the program currently does not have the meansto test this as we are unsure which natural-origin fish were pro-duced by hatchery-origin fish.

Phenotypic differences by themselves do not provide suf-ficient evidence to conclude that genotypic divergence has oc-curred (Knudsen et al. 2006). However, as Kostow (2004) stated,even if the phenotypic differences of the hatchery fish are notinherited they probably influence the relative fitness of the hatch-ery fish when they are in the natural environment and this couldlead to eventual genetic divergence between the groups. Re-gardless of whether the observed differences were caused bygenetics, environmental differences, or a mixture of the two,current hatchery practice does not produce hatchery-origin fishthat are reproductively equivalent to the natural-origin fish (i.e.,lower fecundity).

Araki et al. (2009) provided evidence that the genetic ef-fects of a hatchery supplementation program were not easilyerased by a full generation of fish in the wild, suggesting that

TABLE 5. Descriptive statistics for selected phenotypic traits for age-4 captive-reared broodstock progeny (CRB progeny) and hatchery-origin (HOR) femaleTucannon River spring Chinook Salmon (2008 data, n = sample size). The P-values are two-sample t-test adjusted for multiple testing.

CRB progeny (n = 20) HOR (n = 19)

Phenotypic trait Mean SE Mean SE P-value

FL (cm) 67.9 0.74 68.9 0.78 0.722Weight (kg) 3.32 0.10 3.53 0.12 0.509Fecundity (eggs/female) 2,847 81.1 3,215 131.1 0.076Egg size (g) 0.217 0.007 0.215 0.006 0.999Relative fecundity (eggs/kg) 861.8 17.21 910.7 19.17 0.206Reproductive mass (g) 621.5 32.45 693.4 35.81 0.401

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3


recovery will probably not be immediate after a supplementa-tion program is terminated. However, Lynch and O’Hely (2001)stated that, in principle, a natural population could recover froman excess segregation load after being isolated from a supple-mentation program because beneficial wild-type alleles are stillpresent and can be returned to high frequency by natural selec-tion. More research is needed in this area to better understandthe influence hatchery supplementation has on wild populationsand whether the natural environment will shift the phenotype,and eventually genotype, back to the natural population norm.However, hatchery supplementation programs are not neces-sarily recovery programs on their own. Without addressing theunderlying mechanism that put the population in the positionwhere intervention was necessary, there is little chance the pop-ulation will be able to maintain any demographic boost providedby the hatchery program if the program is stopped.

No discussion on the risks of hatchery supplementationwould be complete without also taking into account the risksof doing nothing (e.g., extinction, reduced effective populationsize, potential for inbreeding). We believe the risk of losing theTucannon River spring Chinook Salmon population to extinc-tion without the hatchery supplementation program is greaterthan the genetic or demographic risks posed by the hatcheryprogram (Gallinat et al. 2008). Lande and Shannon (1996) con-cluded from their work that genetic variability is often less crit-ical in the short term than other determinants of population per-sistence (e.g., habitat destruction, predators, competitors), but inthe long term, it can play a decisive role in allowing a populationto persist and adapt to a changing environment. Hatchery sup-plementation programs will need to balance possible adversegenetic risks while attempting to maintain population persis-tence. In hatchery supplementation programs, such as that forthe Tucannon River spring Chinook Salmon, abundance is notonly demographically important, it is also legally important inorder to fulfill the requirements of the U.S. Endangered SpeciesAct that will support delisting of the population. The TucannonRiver spring Chinook Salmon captive broodstock program hada specific endpoint from the beginning as it was designed to lastfor only one generation (five brood years) to limit genetic risksassociated with captive broodstock programs. Once the conven-tional hatchery supplementation program also ends, the naturalenvironment will eventually determine which phenotypes andgenotypes are best suited to persist if the population is to survive.

ACKNOWLEDGMENTSWe thank the Lyons Ferry Hatchery Complex staff and the

Snake River Lab Hatchery Monitoring and Evaluation staff, forwithout their dedication and hard work over the years this studywould not have been possible. This paper was greatly improvedduring the early stages of our analysis by discussions with ToddPearsons. We thank Andrew Weiss for the map and John Snevafor aging the scale samples. Yong-Woo Lee, Andrew Murdoch,Mark Schuck, and two anonymous reviewers provided construc-

tive comments on the manuscript. The captive broodstock pro-gram (Project 2000-019-00) was funded largely by BonnevillePower Administration, U.S. Department of Energy. The con-ventional hatchery supplementation program is funded by theU.S. Fish and Wildlife Service through the Lower Snake RiverCompensation Plan Office. The use of trade names does notimply endorsement by the Washington Department of Fish andWildlife.

REFERENCESAraki, H., B. A. Berejikian, M. J. Ford, and M. S. Blouin. 2008. Fitness of

hatchery- reared salmonids in the wild. Evolutionary Applications 1:342–355.

Araki, H., B. Cooper, and M. S. Blouin. 2009. Carry-over effect of captivebreeding reduces reproductive fitness of wild-born descendants in the wild.Biology Letters 5:621–624.

Beacham, T. D. 2003. Comment on “rapid evolution of egg size in captivesalmon” (II). Science 302:59d.

Beacham, T. D., and C. B. Murray. 1993. Fecundity and egg size variation inNorth American Pacific salmon (Oncorhynchus). Journal of Fish Biology42:485–508.

Brannon, E. L., D. F. Amend, M. A. Cronin, J. E. Lanna, S. LaPatra, W. J.McNeil, R. E. Noble, C. E. Smith, A. J. Talbot, G. A. Wedemeyer, and H.Westers. 2004. The controversy about salmon hatcheries. Fisheries 29(9):12–31.

Busack, C., and C. M. Knudsen. 2007. Using factorial mating designs to increasethe effective number of breeders in fish hatcheries. Aquaculture 273:24–32.

Campbell, B., B. R. Beckman, W. T. Fairgrieve, J. T. Dickey, and P. Swanson.2006. Reproductive investment and growth history in female Coho Salmon.Transactions of the American Fisheries Society 135:164–173.

Columbia Conservation District. 2004. Tucannon subbasin plan. Available:http://www.nwcouncil.org/fw/subbasinplanning/tucannon/plan. (May 2013).

Fleming, I. A., S. Einum, B. Jonsson, and N. Jonsson. 2003. Comment on “rapidevolution of egg size in captive salmon” (I). Science 302:59b.

Fleming, I. A., and M. R. Gross. 1992. Reproductive behavior of hatcheryand wild Coho Salmon (Oncorhynchus kisutch): does it differ? Aquaculture103:101–121.

Fleming, I. A., B. Jonsson, and M. R. Gross. 1994. Phenotypic divergence ofsea- ranched, farmed, and wild salmon. Canadian Journal of Fisheries andAquatic Sciences 51:2808–2824.

Ford, M. J. 2002. Selection in captivity during supportive breeding may reducefitness in the wild. Conservation Biology 16:815–825.

Gallinat, M. P., J. D. Bumgarner, M. L. Schuck, and G. W. Mendel. 2008.Supplementation of an ESA-listed spring Chinook population with manage-ment limitations. Pages 967–975 in J. L. Nielsen, J. J. Dodson, K. Friedland,T. R. Hamon, J. Musick, and E. Verspoor, editors. Reconciling fisheries withconservation: proceedings of the Fourth World Fisheries Congress. AmericanFisheries Society, Symposium 49, Bethesda, Maryland.

Gallinat, M. P., J. D. Bumgarner, D. Maxey, S. Roberts, R. Rogers, L. A.Ross, and M. A. Varney. 2009. Tucannon River spring Chinook Salmon cap-tive broodstock program. Final Project Completion Report to the BonnevillePower Administration, Project 2000-019-00, Portland, Oregon. Available:www.efw.bpa.gov/searchpublications. (May 2013).

Gallinat, M. P., and L. A. Ross. 2009. Tucannon River spring Chinook Salmonhatchery evaluation program 2008 annual report. Washington Department ofFish and Wildlife, report prepared for the U.S. Fish and Wildlife Service,Cooperative Agreement 1411-08-J011, Olympia.

Hard, J. J., B. A. Berejikian, E. P. Tezak, S. L. Schroder, C. M. Knudsen, andL. T. Parker. 2000. Evidence for morphometric differentiation of wild andcaptively reared adult Coho Salmon: a geometric analysis. EnvironmentalBiology of Fishes 58:61–73.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3


Healey, M. C. 2001. Patterns of gametic investment by female stream- andocean-type Chinook Salmon. Journal of Fish Biology 58:1545–1556.

Heath, D. D., C. W. Fox, and J. W. Heath. 1999. Maternal effects on offspringsize: variation through early development of Chinook Salmon. Evolution53:1605–1611.

Heath, D. D., J. W. Heath, C. A. Bryden, R. M. Johnson, and C. W. Fox. 2003.Rapid evolution of egg size in captive salmon. Science 299:1738–1740.

Kassler, T. W., and C. A. Dean. 2010. Genetic analysis of natural-origin springChinook and comparison to spring Chinook from an integrated supplemen-tation program and captive broodstock program in the Tucannon River. Re-port prepared for the Bonneville Power Administration, Project 2000-019-00,Portland, Oregon.

Kinnison, M. T., M. J. Unwin, A. P. Hendry, and T. P. Quinn. 2001. Migratorycosts and the evolution of egg size and number in introduced and indigenoussalmon populations. Evolution 55:1656–1667.

Knudsen, C. M., S. L. Schroder, C. A. Busack, M. V. Johnston, T. N. Pearsons,W. J. Bosch, and D. E. Fast. 2006. Comparison of life history traits be-tween first-generation hatchery and wild upper Yakima River spring ChinookSalmon. Transactions of the American Fisheries Society 135:1130–1144.

Knudsen, C. M., S. L. Schroder, C. Busack, M. V. Johnston, T. N. Pearsons, andC. R. Strom. 2008. Comparison of female reproductive traits and progenyof first- generation hatchery and wild upper Yakima River spring ChinookSalmon. Transactions of the American Fisheries Society 137:1433–1445.

Kostow, K. 2009. Factors that contribute to the ecological risks of salmon andsteelhead hatchery programs and some mitigating strategies. Reviews in FishBiology and Fisheries 19:9–31.

Kostow, K. E. 2004. Differences in juvenile phenotypes and survival betweenhatchery stocks and a natural population provide evidence for modified se-lection due to captive breeding. Canadian Journal of Fisheries and AquaticSciences 61:577–589.

Lande, R., and S. Shannon. 1996. The role of genetic variation in adaptation andpopulation persistence in a changing environment. Evolution 50:434–437.

Lleonart, J., J. Salat, and G. J. Torres. 2000. Removing allometric effects of bodysize in morphological analysis. Journal of Theoretical Biology 205:85–93.

Lynch, M., and M. O’Hely. 2001. Captive breeding and the genetic fitness ofnatural populations. Conservation Genetics 2:363–378.

Marvin, D., and J. Nighbor, editors. 2009. 2009 PIT tag specification document:Columbia Basin PIT tag information system. Available: www.ptagis.org.(May 2013).

Nehlsen, W., J. E. Williams, and J. A. Lichatowich. 1991. Pacific salmon at thecrossroads: stocks at risk from California, Oregon, Idaho, and Washington.Fisheries 16(2):4–21.

Patterson, D. A., J. S. Macdonald, S. G. Hinch, M. C. Healey, and A. P. Farrell.2004. The effect of exercise and captivity on energy partitioning, reproductivematuration and fertilization success in adult sockeye salmon. Journal of FishBiology 64:1039–1059.

Peck, L. 1993. Integrated hatchery operations team: operation plans for anadro-mous fish production facilities in the Columbia River Basin, volume IV,annual report 1992. Report to Bonneville Power Administration, Project 92-043, Portland, Oregon.

SAS (SAS Institute, Inc.). 2004. SAS/STAT 9.1 user’s guide. SAS Institute,Inc., Cary, North Carolina.

Swain, D. P., B. E. Riddell, and C. B. Murray. 1991. Morphological differ-ences between hatchery and wild populations of Coho Salmon (Oncorhynchuskisutch): environmental versus genetic origin. Canadian Journal of Fisheriesand Aquatic Science 48:1783–1791.

Thorpe, J. E. 2004. Life history responses of fishes to culture. Journal of FishBiology 65(Supplement A):263–285.

USACE (U.S. Army Corps of Engineers). 1975. Special report, lower SnakeRiver fish and wildlife compensation plan: lower Snake River, Washingtonand Idaho. USACE, Walla Walla, Washington.

Waples, R. S. 1991. Genetic interactions between hatchery and wild salmonids:lessons from the Pacific Northwest. Canadian Journal of Fisheries and AquaticSciences 48:124–133.

Watson, M. 1996. Hatchery evaluation report Lyons Ferry Hatchery – springChinook: an independent audit based on integrated hatchery operations team(IHOT) performance measures. Report to the Bonneville Power Administra-tion, Project 95-2, Portland, Oregon.

Dow

nloa

ded

by [

Dep

artm

ent O

f Fi

sher

ies]

at 2

0:42

27

Oct

ober

201

3

athomas m. harder , gordon g. gotsch & robert c ......spawning season (mcginty and hodson 2008)....

Documents