Fisheries science JSFS , tập 78, số 6, 2012 11

Some studies of the biological characteristics of dol-phinfish have been reported in several regions, and include the determination of their age and growth characteristics in North Carol

O R I G I N A L A R T I C L E Fisheries

Age, growth, and reproductive characteristics of dolphinfish

Coryphaena hippurus in the waters off west Kyushu,

northern East China Sea

Seishiro Furukawa•Seiji Ohshimo• Seitaro Tomoe•

Tetsuro Shiraishi•Naoyuki Nakatsuka•Ryo Kawabe

Received: 27 December 2011 / Accepted: 27 August 2012 / Published online: 13 October 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract The growth and reproductive characteristics of

dolphinfish Coryphaena hippurus collected in the waters

off western Kyushu from May 2008 to April 2011 were

determined based on scale and otolith readings and gonad

histological examinations, respectively Based on annual

increments in scales and daily increments in sagittal

oto-liths, the von Bertalanffy growth curves in male and

females were determined as FLt¼ 1049½1  expf0:835

ðt þ 6:975  1014Þg and FLt¼ 938½1  expf1:029ðtþ

6:975 1014Þg, respectively, where FLtis the mean fork

length (mm) at age t The spawning period was found to

last from June to August for dolphinfish, based on an

examination of the monthly changes in the gonadosomatic

index and histological observations Therefore, based onthe relationship between the fork length and the develop-mental stage of the testes or ovaries, male and femaledolphinfish were found to reach sexual maturity by thefollowing spawning season after hatching in the northernEast China Sea

Keywords Dolphinfish Growth  Scale  Otolith Reproduction Gonad histology

IntroductionDolphinfish Coryphaena hippurus is a highly migratoryoceanic pelagic fish found worldwide in tropical, subtrop-ical, and temperate waters [1] In East Asia, dolphinfishsupport economically important recreational and commer-cial fisheries, and are a shared resource among multiplecountries, such as Taiwan and Japan [2, 3] Dolphinfishfeed on several important commercial fishery species of theEast China Sea, including anchovy Engraulis japonicas,flying fish (Exocoetidae), and other small pelagic prey,including squid [4]

The removal of predator biomass during commercialfishing can have profound effects on pelagic ecosystemsbecause of the removal of predation pressure and top-down, trophic-cascade effects [5 7] Intense harvesting(i.e., overexploitation) may select for biological traits such

as slow growth [8] or early maturity [9]; however, howthese dolphinfish traits will change in the future remains to

be determined Therefore, it is necessary to clarify thecurrent biological characteristics of dolphinfish so that wemay understand how they will change with time and how

we should manage this species

Graduate School of Science and Technology, Nagasaki

University, Bunkyo-machi, Nagasaki 852-8521, Japan

e-mail: seishirou07@gmail.com

Seikai National Fisheries Research Institute, Fisheries Research

Agency, Taira-machi, Nagasaki 851-2213, Japan

S Tomoe

Japan Overseas Cooperation Volunteers, Japan International

Cooperation Agency (JICA), Tokyo, Japan

S Tomoe

Service De´partemental de Peˆche et de la Surveillance de Mbour,

Mbour, Republic of Senegal

T Shiraishi

Okayama Fisheries Promotion Foundation,

Urayasu-minami-machi, Okayama 702-8024, Japan

Graduate School of Fisheries Science and Environmental

Studies, Nagasaki University, Taira-machi, Nagasaki 851-2213,

Japan

DOI 10.1007/s12562-012-0557-6

Some studies of the biological characteristics of

dol-phinfish have been reported in several regions, and include

the determination of their age and growth characteristics

in North Carolina [10], Gulf of Mexico [11], and the

Mediterranean [12]; their feeding habits in the eastern

Pacific Ocean [13] and Mediterranean [14]; and their

swimming behavior in natural conditions in the northern

East China Sea [15] The reproductive characteristics of

dolphinfish have been reported from North Carolina [10],

the Gulf of Mexico [11], Taiwan [16], and the Gulf of

Tehuantepec [17]

Previous studies on dolphinfish in Japan reported their

age and growth characteristics based on fork length

fre-quency data from the Sea of Japan, and estimated spawning

periods from seasonal changes in oocyte diameter [4]

However, little is known about the growth of dophinfish

from Japanese waters using hard parts and their

repro-ductive characteristics using histological techniques The

objective of this study was to determine age using otolith

and scale readings, and to examine annual reproductive

cycle and sexual maturity using histological techniques, for

dolphinfish in the northern East China Sea

Materials and methods

Collections

Both small and large specimens were used for aging while

large dolphinfish were used for reproduction Large

spec-imens were collected monthly from May 2008 to July 2010

(except in August 2008, April, May, September, and

October 2009, and January, February, and March 2010)

and in April 2011, which were caught predominantly by set

net along the coast of the Goto Islands, Japan, but

occa-sionally using troll and long line gear in the coastal waters

off of Tsushima Island and the Goto Islands (Fig.1) Small

specimens were caught by neuston net [18] with a mesh

size of 2 mm at sampling sites distributed in coastal waters

off West Kyushu and in the Tsushima Strait in June to

September 2005 The neuston net was towed through the

surface water for 10 min, and specimens were sorted

onboard and frozen immediately at -35°C We did not

use small specimens caught by neuston net Specimens

were measured to the nearest millimeter in total length and

fork length and to the nearest gram of body weight (BW)

For reproductive characterization, the gonad weight (GW)

was measured to the nearest 0.1 g after determining the

sex, and the gonadosomatic index (GSI) value was

corre-Age determinationFor age determination, we used sagittal otoliths and scalesobtained from small-sized specimens and large-sizedspecimens, respectively The deposition of increments indolphinfish otoliths begins on the hatching date, and ringsare laid down daily [12, 19] Thus, no adjustment wasrequired to estimate age from incremental counts of sag-ittae, and it was assumed that rings were formed daily.Previous studies on the microstructure of sagittal otoliths ofdolphinfish from the western Mediterranean Sea had foundthat the daily ages of larger dolphinfish ([650 mm FL)appeared to be underestimated [12] Furthermore, in thisstudy, daily rings of sagittal otoliths were unclear in largedolphinfish (C412 mm FL) Therefore, our daily ringdetermination was restricted to small dolphinfish Todetermine the ages of small dolphinfish in days (herein

‘‘daily ages’’), otoliths were removed under a dissectingmicroscope and embedded in resin on a glass slide The

China Sea

50 N

30 N

Tsushima Islands

Sea of Japan

Taiwan

Goto Islands

in the northern East China Sea

otolith increments were counted under a light microscope.

Since the daily rings of sagittal otoliths for large fish were

unclear, sectioning and a thin polish were required

How-ever, only small fishes were used for age determination in

this study, and we did not section and thin polish the

oto-liths to determine the daily ages of the fish

Annual marks are not detectable on otoliths of

dol-phinfish [10] Thus, the ages of the dolphinfish in years

(herein ‘‘annual ages’’) were estimated from their scales

Scales were taken from above the lateral line, washed with

water, and placed between two slide glasses Numbers of

annual scale rings were counted under a digital microscope

(E-LV100D, Nikon, Tokyo, Japan) with transmitted light

The ring radii of the scales were measured using an otolith

measurement system (ODRMS, RATOC, Tokyo, Japan)

Each scale was examined two times, with a minimum of

one month between examinations, by two independent

readers If two or more examinations of the scales of the

individual agreed in terms of the number of ring marks, this

number was recorded and used for the analyses To validate

the annual marks in dolphinfish scales, an indirect

valida-tion based on marginal increment analysis was used The

marginal increment (MI) was determined using the

fol-lowing equation:

MI¼ ðR  rnÞ

ðrn rn1Þ;

where R is the overall radius from the focus to the outer

edge of the scale Rnis the radius from the focus to each

annulus MIs were analyzed by a GLM using a gamma

distribution with a log link function to test for a month

effect

Estimation of the von Bertalanffy growth parameters

The von Bertalanffy growth curve was fitted to daily ages

for small dolphinfish and annual ages for large dolphinfish

We could not determine the birth date of the large

dol-phinfish Therefore, the birth date of every large individual

was assumed to be 1 July, which approximately

corre-sponded to the middle of the spawning period (see

‘‘Results’’) The von Bertalanffy growth parameters were

estimated using maximum likelihood estimation (MLE)

with the R 2.13 software package [20] (The R Project for

Statistical Computing: http://www.r-project.org/) We

assumed a normal distribution for fork length (FL) at age t

and for sex s, with a mean of l(t, s) and a variance of V(t,

s) The mean fork length at age t and for sex s is

repre-sented by the following von Bertalanffy growth equation:

lðt; sÞ ¼ FL1;s½1  expfKsðt  t0Þg; ð1Þ

where l(t, s) represents the mean fork length at age t and

for sex s FL?,sand ksrepresent the asymptotic fork length

and the growth coefficient for sex s, respectively Thehypothetical age corresponding to a fork length of zero is

t0 We defined FL?,s and ksin the following ways:

ln L¼ 1

2

XN i¼1

where ln L is the log likelihood and i¼ 1; 2; ; N FL?,s,

ks; t0;r0 and r were estimated by maximizing Eq 3usingthe gosolnp function in the Rsolnp package of R To selectthe best-fit model from among cases 1–4, we used theAkaike information criterion (AIC) The model whichyielded the minimum AIC was selected as the best model.Influence of water temperature on growth

Asymptotic fork lengths, estimated from von Bertalanffygrowth function fits using size and age data collected indifferent regions of the world [10,11], were used as a region-specific growth index We did not use the maximum size as

an index for growth in order to avoid local sampling bias Toexamine the influence of water temperature and sex onasymptotic fork length estimated from the previous studiesand the present study, a generalized linear mixed model(GLMM) assuming a gamma distribution and a log-linkfunction were used Differences in catch region were defined

as a random effect, since our objective was not to test forunknown regional effects The gamma GLMM was con-ducted using the GLMM function in the ‘‘repeated’’ package

of R As an index for water temperature, we used derived sea surface temperatures [SST, 11 km resolutionadvanced very high resolution radiometer (AVHRR):Pathfinder V5] obtained from the Ocean Watch webpage(http://las.pfeg.noaa.gov/oceanWatch/oceanwatch.php), and

satellite-we used mean SSTs obtained from 2003 to 2010 in the area

where dolphinfish were caught in previous studies To test

the relative importance of water temperature and sex, we

compared GLMMs that included terms for SST alone, sex

alone, and for both SST and sex, and used the AIC to assess

the best-fit model

Histological observations

The fixed gonads were dehydrated and embedded in

par-affin, and sections (thickness 4 lm) were obtained and

stained by Mayer’s hematoxylin and eosin method, or were

dehydrated and embedded in resin (Historesin) and sections

were stained with 2 % toluidine blue and 1 % borax The

stained sections were observed under an optical

micro-scope and the most advanced testis and oocyte stages were

recorded The developmental stages of testes and ovaries

were classified into five and six stages of maturity,

respectively, based on the development of the most

advanced testes and oocytes and their histological

charac-teristics (Figs.2,3)

The five testis stages were as follows:

1 Spermatogonial proliferation stage (Sp; Fig.2a): only

spermatogonia (sg) are abundant in the seminal lobule

2 Early spermatogenesis stage (Es; Fig 2b): gonia (sg) and spermatids (st) are organized in theseminal lobules

spermato-3 Late spermatogenesis stage (Ls; Fig.2c): genesis proceeds in the testis Spermatids (st) of theseminal lobules increase, and spermatozoa (sz) arefound in the lumina of the seminal lobules

spermato-4 Functional maturation stage (Fm; Fig.2d): zoa (sz) are abundant in the lumina of the seminallobules and main sperm duct Spermatogonial divisionand further spermatogenesis proceeds in the seminallobules

spermato-5 Postspawning stage (Ps; Fig.2e): spermatogonia (sg)are found in the seminal lobules, although spermatozoa(sz) occur in the lumina of the seminal lobules.The six stages of oocytes were as follows:

1 Immature stage (Im; Fig.3a): only previtellogenic (pn)oocytes are present, including those in the perinucle-olus and yolk vesicle stages

2 Developing stage (D; Fig.3b): the most advancedoocytes are at the early yolk (ey) or mid-yolk (my) stages

3 Vitellogenic stage (Vi; Fig.3c): the most advancedoocytes are at the late yolk (ly) stage, which marks theend of vitellogenesis

80µm

in dolphinfish a Spermatogonia proliferation stage, b early

sper-matogenesis stage, c late spersper-matogenesis stage, d functional

maturation stage, and e postspawning stage sg spermatogonial,

st spermatid, sz spermatozoon

4 Mature stage (M; Fig.3d): the most advanced oocytes

are at the hydration (hy) stages The degenerated old

postovulatory follicles (pof) appear in some ovaries at

the germinal vesicle migration

5 Spawning stage (Sp; Fig.3e): yolked oocytes and new

pof are present Most pofs disappear from the ovaries

before the developing oocytes attain the germinal

vesicle migration stage

6 Resting stage (Re; Fig.3f): all yolked oocytes are

degenerating (atretic stage, at) and non-yolked oocytes

are present

Results

Growth

A total of 278 specimens including small dolphinfish (total

length, TL 9.5–237.0 mm, n = 141) and large dolphinfish

(FL 412–1124 mm, n = 137) were used for age

determi-nation Unfortunately, we could not collect the fish

between 237 and 412 mm because this size range of

dol-phinfish does not support economically important

com-mercial fisheries in this study area However, we obtained a

sufficient wide size range to describe the general growth

curve for dolphinfish A total of 141 otoliths from sized dolphinfish were examined Sex could not be deter-mined for the juvenile dolphinfish whose sagittae wereexamined (mean FL = 25.5 mm, range 9.5–237.0 mm);however, these dolphinfish were still used in the von Ber-talanffy analysis Minimum and maximum daily ages were

small-4 and 53 days, respectively Scales were collected from

136 large-sized dolphinfish, and the rate of agreementbetween readers of the number of annual ring marks was64.2 % (88 of the 137 specimens in total) A total of 69scales were classified as age 1 or older, and the remainingscales (n = 19) were estimated to be age 0 The estimatedmaximum ages for males and females were five years old.MIs from [age 0 dolphinfish (n = 69) were greatest inOctober, November, and December, dropped in January,and stayed low during the winter months and the spawningseason (see subsequent results) (Fig.4) There was a sig-nificant difference in marginal increment width per month(gamma GLM, p \ 0.05)

The von Bertalanffy growth parameters were estimatedfor cases 1–4 (Table1) When we used case 1, the mini-mum AIC was obtained, and the DAIC value of the nextmost parsimonious case (case 2) was more than 2 [22] Therelationship between age and FL of the dolphinfish isshown in Fig.5a, where most of the data are within the

95 % prediction interval for both sexes (Fig.5b, c),

in dolphinfish a Immature stage, b developing stage, c vitellogenic

stage, d mature stage, e spawning stage, and f resting stage at atretic

oocyte, ey early yolk oocyte, hy hydration oocyte, ly late yolk oocyte,

my mid-yolk oocyte, pn perinucleolus oocyte, pof postovulatory follicle

supporting the model for case 1 The von Bertalanffy

growth curves are thus shown separately for males and

females The mean growth curve of dolphinfish was

esti-mated in males and females as follows:

lðt; maleÞ ¼ 1049 1  exp 0:835ðt þ 6:975  10  14Þ

lðt; femaleÞ ¼ 938½1  expf1:029ðt þ 6:975  1014Þg:

Comparison of the AIC scores (Table2) revealed that

the model that included a term for SST but not sex

pro-vided the best explanation of the variation in asymptotic

fork length The relationship between the asymptotic fork

length and SST was plotted (Fig.6), and SST ranged from

19.7 to 27.4°C There was a significant positive ship between FL1and SST (p \ 0.05) The best-fit modelwas

relation-FL1¼ expð5:87 þ 0:053  SSTÞ:

Annual reproductive cycle

A total of 329 large dolphinfish (FL 412–1124 mm, 112male and 217 female) were used for reproductive charac-terization (137 of 329 specimens were used for the agingstudy) Length-adjusted mean gonad weights varied sig-nificantly with month (gamma GLM, p \ 0.001) for bothmale and female dolphinfish (Fig 7) The mean value of

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

8

8

3

dolphin-fish (sampled from May 2008 through May 2010) pooled by month

(January–December) Sample sizes are given above the box for each

month

3 2

1 0

Age

Male Female Small Male mean Male 95 % PI Female mean Female 95 % PI

female (gray) dolphinfish The mean growth curves and 95 % prediction intervals are indicated by solid lines and dotted lines, respectively

GSI in male dolphinfish was high (GSI [0.7) from May to

August (Fig.7a), and in female dolphinfish (GSI [4.5)

from June to August (Fig.7b) The maximum values of the

mean GSI for males and females were 1.9 in May and 16.4

in July, respectively The mean GSI value became low in

September, and was below 0.4 in males and 1.5 in females

from September to March Immature males (Sp, Ls, and Es

stages) of the dolphinfish were observed from October to

May (Fig.8a) Males with testes at the Fm stages appeared

in June (100 %) and July (54.5 %), although it is important

to note that only one and two specimens were collected,

respectively Immature females (Im and D stages) of

dol-phinfish were observed from September to June (Fig.8b)

Females with ovaries at the Vi stage appeared from June

(12.1 %) and August (12.5 %) Specimens collected in

June to August had ovaries at the Vi or M stages, and

females with ovaries at the Sp stage were also observed

The proportion of Sp-stage females was highest in July.Females with ovaries at the Re stage were found fromAugust and October

Spawning size and GSISexually mature males were defined as individuals withtestes at the Fm stage Sexually mature females weredefined as individuals with ovaries with Vi, M, or Sp stageoocytes GSI values for immature stages (Sp to Ls) and themature stage (Fm) overlapped, and these stages rangedfrom 0.5 to 0.9 (Fig.9a) Individuals with testes at the Fmstage were also larger than 524 mm FL (Fig.9b) GSIvalues of female individuals were less than 0.8 in theimmature (Im) stage, and the values for the oocytes rangedfrom 0.2 to 4.0 in the developing (D) stage (Fig.9c) GSIvalues for females in the Vi, M, and Sp stages ranged from3.3 to 11.5 The minimum fork length of females in themature (from Vi to Sp) stages was 514 mm FL (Fig.9d)

24 22

20

SST (ºC)

Mediterranean

Present study

North Carolina

Florida

Puerto Rico

Male Female

the gray and light gray zones indicate the 50 and 95 % prediction

intervals, respectively

explain variation in asymptotic fork length in terms of sea surface

temperatures (SST), sex, and their interaction The cases are listed

from best to worst based on AIC and DAIC

2.4

a

3

7 54

9

3 12

1

2

35 25 6

Month

b

2 11 4

6 28 41 5

8

15

42

29 26

pooled by month (January–December) Sample sizes are given above the box for each month

Age and growth

This study is the first to use sagittal otoliths and scales to

determine daily and annual ages of dolphinfish from the

northern East China Sea The von Bertalanffy growth

parameters were elucidated and the FL1 values of male

and female dolphinfish were estimated to be 1049 and

938 mm, respectively, while the k values in males and

females were 0.835 and 1.029, respectively The growth

parameters of dolphinfish in the southwestern Sea of Japan

adjacent to the northern East China Sea were analyzed

using length frequency [4] According to that study, the

FL? and k values were 1750 and 0.22 mm, respectively,

which were pooled in males and females The initial

growth rate of the dolphinfish examined here was faster

than that of those from the southwestern Sea of Japan, but

the maximum sizes of both sexes in this study were smallerthan those from the southwestern Sea of Japan Kojima [4]did not have small fish in his sample This study is the first

to use sagittal otoliths to determine daily ages of phinfish near the East China Sea Therefore, we suggestthat the growth parameters estimated in this study are moreuseful for examining and comparing growth among otherregions

dol-The asymptotic fork length of dolphinfish in the ern East China Sea shows a greater similarity to theasymptotic fork length of western Mediterranean Sea dol-phinfish [12] than to the asymptotic fork lengths of dol-phinfish in other regions (Fig.6) However, to the best ofour knowledge, the first-year growth for dolphinfish fromthe northern East China Sea is the smallest in the world It

north-is well known that differences in estimated growth betweenregions [5,6] can be related to environmental conditions(i.e., water temperature, food availability, exploitationlevels) Moreover, temperature appears to be the mostimportant environmental factor affecting growth in fish.Because of the importance of temperature as a controllingfactor [23], the physiological literature is replete withexamples of studies evaluating thermal effects on fish [24–

26] Although differences in the growth of dolphinfishamong regions have been found [10, 11], few measure-ments of the effect of temperature on the growth of dol-phinfish have been performed Our study is unique in that itshows that there were clear distinctions in asymptotic forklength with respect to water temperature (Fig.6) However,additional bioenergetic data are required to parameterizemodels that attempt to depict patterns of growth observed

in dolphinfish Of particular importance are data describingthe effects of water temperature, body size, and feeding onmetabolism

The asymptotic fork length of dolphinfish was cantly larger in males than in females in this study, whichreflects results from Florida and North Carolina (Fig.6).However, the asymptotic fork length of dolphinfish waslarger in females than in males from the Mediterranean andPuerto Rico (Fig.6) Differences in growth features due todifferent laboratory methods can not be excluded Forexample, age determination of dolphinfish based solely onotoliths was found to underestimate the ages of older,larger fish [12] Obviously, one of the main ways toidentify the factors responsible for this inter-region vari-ability in growth would be to standardize age and growthmethods

signifi-ReproductionThere are no reports regarding dolphinfish maturation inthe East China Sea that utilized histological techniques.Kojima [4] estimated the spawning period by examining

maturation stages of a testes and b ovaries in dolphinfish For males:

Sp spermatogonial proliferation stage, Es early spermatogenesis stage,

Ls late spermatogenesis stage, Fm functional maturation stage,

Ps postspawning stage For females: Im immature stage, D developing

stage, Vi vitellogenic stage, M mature stage, Sp spawning stage,

Re resting stage

seasonal changes in oocyte diameter obtained from

dol-phinfish in the southwestern Sea of Japan, which is

adja-cent to the northern East China Sea We used histological

techniques and examined the relationship between the most

advanced oocyte stage and GSI values Previous studies of

the reproductive characteristics of dolphinfish revealed that

dolphinfish spawned throughout the year, with

reproduc-tive activity peaking in February to March in the southern

East China Sea on the east coast of Taiwan [16] On the

other hand, in the northern East China Sea, the GSI values

in both sexes were high, and oocytes at the spawning stage

in dolphinfish occurred from June to August in our study

These results suggest that peak spawning in dolphinfish in

the northern East China Sea occurs from June to August

However, it is not clear whether these differences in

spawning season in the East China Sea occur due to

geo-graphic differences in dolphinfish distribution (i.e.,

respective latitudes and physical conditions) or genetic

differences among the dolphinfish In the future, controlled

experiments to examine how environmental conditions

affect the reproduction of dolphinfish and to detect the

genetic population structure [28,29] are needed to clarify

any differences in growth and reproduction among

differ-ent areas

Generally, FL at 50 % maturity (L50) was used as the

index of size at maturity for dolphinfish by fitting a logistic

function to the frequency of mature fish for each body size

class [10, 16, 27] The L50 determined in the Mexican

Pacific (L50 = 483.8 and 505.7 mm for females and males,respectively) [17], Taiwanese waters (L50= 510 mm forboth sexes) [16], and off the coast of North Carolina(L50 = 458 and 476 mm for females and males, respec-tively) [10] agree with values for fish that are less thanone year old, regardless of sex We were unable to estimate

L50 for dolphinfish in both sexes from the East China Seabecause of a lack of reproductive characterization of small-sized individuals during the spawning season Neverthe-less, we determined that the smallest individuals withmatured testis and oocytes were 524 and 514 mm,respectively, with an estimated age of less than one year inboth sexes Hence, dolphinfish reach sexual maturity intheir first year of life in the East China Sea, which is similar

to other regions [10, 16, 27] Clearly, it is necessary tomonitor variations in reproductive characteristics in futurestudies, and to further determine the growth and maturityprocesses of fish that are yet to reach one year of age, inorder to elucidate size at sexual maturation in the northernEast China Sea

Fisheries Research Institute for his cooperation during the study,

Mr E Kusaba, D Tawara, Y Mori and other members of the Takahama Fisherman’s Association, and H Tsubakiyama of Wakamatsu Fisherman’s Association for collecting the samples We also thank Dr G.N Nishihara, who assisted with the interpretation and the English of the manuscript This study was supported by the Fisheries Research Agency.

Im D Vi M Sp Re

1000 800

600 400

Fork length (cm)

Im D Vi M Sp Re

12 10 8 6 4 2 0

GSI

Sp Es Ls Fm Ps

1200 1000 800 600 400

Fork length (cm)

Sp Es Ls Fm Ps

1.4 1 0.7 0.4 0

GSI

testes and gonadosomatic index

(GSI), b the five maturation

stages of testes and fork length,

ovaries and GSI, and d the six

maturation stages of ovaries and

fork length of dolphinfish Refer

and ovarian stage, respectively.

Crosses in c and d indicate that

postovulatory follicles were

observed

1 Palko BJ, Beardsley GL, Richards WJ (1982) Synopsis of the

biological data on dolphin fishes, Coryphaene hippurus and

Coryphaene equiselis Linnaeus FAO Fish Synop 130:1–28

2 Taiwan Fisheries Bureau (1996) Fisheries yearbook of Taiwan

area Department of Agriculture and Forestry Provincial

Gov-ernment of Taiwan, Taipei

3 Sakamoto R, Kojima S (1999) Review of dolphinfish biological

and fishing data in Japanese waters Sci Mar 63:375–385

4 Kojima S (1966) Fishery biology of the common dolphin,

Coryphaena hippurus L., inhabiting the Pacific Ocean Bull

Shimane Pref Fish Exp Stn 1:1–108

5 Shiomoto A, Tadokoro K, Nagasawa K, Ishida Y (1997) Trophic

relations in the subarctic North Pacific ecosystem: possible

feeding effect from pink salmon Mar Ecol Prog Ser 150:75–85

6 Estes J, Tinker AMT, Williams TM, Doak DF (1998) Killer

whale predation on sea otters linking oceanic and nearshore

ecosystems Science 282:473–476

7 Verheye HM, Richardson AJ (1998) Long-term increase in

crustacean zooplankton abundance in the southern Benguela

upwelling region (1951–1996): bottom-up or top-down control?

ICES J Mar Sci 55:803–807

8 Conover DO, Munch SB (2002) Sustaining fisheries yields over

evolutionary time scales Science 297:94–96

9 Trippel EA (1995) Age at maturity as a stress indicator in

fish-eries Bioscience 45:759–771

10 Schwenke KL, Buckel JA (2008) Age, growth, and reproduction

of dolphinfish (Coryphaena hippurus) caught off the coast of

North Carolina Fish Bull 106:82–92

11 Oxenford HA (1999) Biology of the dolphinfish (Coryphaena

hippurus) in the western central Atlantic: a review Sci Mar

63:277–301

12 Massutı´ E, Morales-Nin B, Moranta J (1999) Otolith

micro-structure, age, and growth patterns of dolphin, Coryphaena

hip-purus, in the western Mediterranean Fish Bull 97:891–899

13 Olson RJ, Galvan-Magan˜a F (2002) Food habits and comparison

rates of common dolphinfish (Coryphaena hippurus) in the

eastern Pacific Ocean Fish Bull 100:279–298

14 Massutı´ E, Deudero S, Sa´nchez P, Morales-Nin B (1998) Diet and

feeding of dolphin (Coryphaena hippurus) in western

Mediter-ranean waters Bull Mar Sci 63(2):329–341

15 Furukawa S, Kawabe R, Ohshimo S, Fujioka K, Nishihara GN,

Tsuda Y, Aoshima T, Kanehara H, Nakata H (2011) Vertical

movement of dolphinfish Coryphaena hippurus as recorded by

acceleration data-loggers in the northern East China Sea Environ

Biol Fish 92:89–99

16 Wu CC, Cheng SW, Kawasaki T (2001) Reproductive biology of the dolphin fish Coryphaena hippurus on the east coast of Tai- wan Fish Sci 67:784–793

17 Alejo-Plata C, Dı´az-Jaimes P, Salgado-Ugarte IH (2011) Sex ratios, size at sexual maturity, and spawning seasonality of dol- phinfish (Coryphaena hippurus) captured in the Gulf of Tehu- antepec, Mexico Fish Res 110:207–216

18 Oozeki Y, Kimura R, Kubota H, Ishida M (2001) Modified neuston net for collecting larvae and juveniles of Pacific saury, Cololabis saira Suisan Kaiyo Kenkyu 65:1–5

19 Uchiyama JH (1986) Growth of dolphins, Coryphaena hippurus and C equiselis, in Hawaiian waters as determined by daily increments on otoliths Fish Bull 84:186–191

20 Ihaka R, Gentleman RR (1996) A language for data analysis and graphics J Comput Graph Stat 5:299–314

21 Tanaka E, Tanaka S (1990) A method for estimating position from length–frequency by using stochastic growth equation Nippon Suisan Gakkaishi 56:1209–1218

age-com-22 Burnham KP, Anderson DR (1998) Model selection and ence: a practical information-theoretic approach Springer, New York

infer-23 Fry FEJ (1957) The aquatic respiration of fish In: Brown ME (ed) The physiology of fishes, vol I Academic, New York,

pp 1–63

24 Brett JR, Groves TDD (1979) Physiological energetics In: Hoar

WS et al (eds) Fish physiology, vol VIII Academic, New York,

pp 279–352

25 Houde ED, Zastrow CE (1993) Ecosystem- and taxon-specific dynamic and energetic properties of larval fish assemblages Bull Mar Sci 53:290–335

26 Keckeis H, Kamler E, Bauer-Nemeschkal E, Schneeweiss K (2001) Survival, development and food energy partitioning of nase larvae and early juveniles at different temperatures J Fish Biol 59:45–61

27 Zu´n˜iga-Flores MS, Ortega-Garcı´a S, Rodrı´guez-Jaramillo MDC, Lo´pez-Martı´nez J (2011) Reproductive dynamics of the common dolphinfish Coryphaena hippurus in the southern Gulf of Cali- fornia Mar Biol Res 7:677–689

28 Dı´az-Jaimes P, Uribe-Alcocer M, Ortega-Garcı´a S, Durand JD (2006) Spatial and temporal mitochondrial DNA genetic homo- geneity of dolphinfish populations (Coryphaena hippurus) in the eastern central Pacific Fish Res 80:333–338

29 Tripp-Valde´z M, Garcı´a de Leo´n FJ, Ortega-Garcı´a S, Cota D, Lo´pez-Martı´nez J, Cruz P (2010) Population genetic structure of dolphinfish (Coryphaena hippurus) in the Gulf of California, using mirosatellite loci Fish Res 105:172–177

Lluch-O R I G I N A L A R T I C L E Fisheries

Food habits of introduced brown trout and native masu salmon

are influenced by seasonal and locational prey availability

Koh Hasegawa•Chitose Yamazaki•

Tamihisa Ohta•Kazumasa Ohkuma

Received: 19 May 2012 / Accepted: 28 August 2012 / Published online: 26 September 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract A knowledge of food habits is important for

evaluating interspecific competition and predation between

sympatric species Data on food availability should be

combined with data on food habits in this type of survey

Although food availability differs between habitats or

seasons, these differences had never been considered in

previous studies We conducted year-round field surveys

throughout a stream to compare the food habits of an

introduced salmonid, brown trout Salmo trutta, and a

native salmonid, masu salmon Oncorhynchus masou Masu

salmon did not constitute a large proportion of the diet of

brown trout and vice versa Thus, predation will likely not

affect the population level of either species The dietary

overlap between brown trout and masu salmon varied

depending on the presence of Gammaridae and terrestrial

invertebrates; i.e., the intensity of interspecific competition

for food resources may differ according to food conditions

Keywords Competition Gammaridae Terrestrial invertebrates  Predation

IntroductionBecause of their importance for commercial and recrea-tional fishing, salmonids have been introduced into manyregions outside their natural range [1] Their impacts onnative species and communities through interspecificinteractions have been well studied (e.g., [2, 3]) In par-ticular, brown trout Salmo trutta and rainbow trout On-corhynchus mykiss have been introduced into many regionsthroughout the world and are listed in the list of 100 of theworld’s worst invasive alien species, published by theInternational Union for the Conservation of Nature andNatural Resources [4]

Introduced brown trout have become a significantproblem in streams in Hokkaido, northern Japan Thesestreams contain masu salmon Oncorhynchus masou, anendemic salmonid in far-eastern Asia and an importantfishery resource In general, there is worldwide concernthat brown trout are detrimental to native salmonid popu-lations because of predation by large individuals (particu-larly those larger than 300 mm) and competition [4] Forexample, Takami et al [5] reported that brown troutreplaced native white-spotted charr Salvelinus leucomaenis

in a mountainous stream in Hokkaido, and Hasegawa andMaekawa [6] suggested that interspecific competition is theprimary mechanism driving replacement However, therehave been few direct studies of interspecific interactionsbetween brown trout and masu salmon These includelimited case studies of predation (e.g., [7, 8]) Hasegawa

et al [9] clarified that a difference in species-specificontogenetic habitat-shift patterns at the fry stage reduced

Hokkaido National Fisheries Research Institute,

Fisheries Research Agency, Nakanoshima, Toyohira,

Sapporo, Hokkaido 062-0922, Japan

e-mail: hasekoh@affrc.go.jp

K Ohkuma

e-mail: ohkuma@affrc.go.jp

C Yamazaki

Graduate School of Environmental Science, Hokkaido

University, N10 W5, Kita, Sapporo, Hokkaido 060-0810, Japan

e-mail: yamyama@fsc.hokudai.ac.jp

T Ohta

Tomakomai Experimental Forest, Hokkaido University,

Takaoka, Tomakomai, Hokkaido 053-0035, Japan

e-mail: tami-hisa@fsc.hokudai.ac.jp

DOI 10.1007/s12562-012-0554-9

competitive interactions in a natural stream Hasegawa

et al [10] demonstrated that brown trout dominated masu

salmon in interference competition for foraging habitat at

the parr stage However, Hasegawa and Maekawa [11]

demonstrated that they had different habitat preferences at

parr stage (both species preferred pool habitats, though

masu salmon preferred the surface whereas brown trout

preferred benthic areas) Furthermore, habitat use by both

species was unchanged by the presence of other species,

such that interspecific competition for foraging habitat was

unlikely to occur in an artificial stream [11] However,

these studies [10,11] were conducted in artificial streams,

and knowledge of interspecific interactions at the parr stage

under natural conditions is still inadequate

To improve our understanding of the interspecific

inter-actions (competition and predation) between brown trout and

masu salmon at the parr stage, we need to account for the

food habits of both species This is because of the following

reasons First, stream-dwelling salmonids compete for food

resources [12], though species-specific preferences for food

resources differ, due to species-specific microhabitat

pref-erences in some cases [13] Based on the observations of

Hasegawa and Maekawa [11], therefore, brown trout likely

forage primarily on benthic invertebrates, whereas masu

salmon likely forage on terrestrial invertebrates However,

the composition of benthic and terrestrial invertebrates

var-ies dramatically with habitat and season [13–15] Thus, food

habits and dietary overlaps may vary depending on prey

composition Second, a knowledge of food resource

avail-ability is essential in order to accurately estimate the

occurrence of predation [16] Due to the current lack of

information on brown trout predation on masu salmon, we

tried to identify the environment where such predation is

common To address these questions, we conducted field

surveys throughout the year at various places in the stream

Materials and methods

Study site

The field survey was conducted in Mamachi stream, a

trib-utary of the Chitose River, Hokkaido, northern Japan (Fig.1)

between the end of May 2009 and the end of March 2010

The stream is spring-fed and has a generally homogeneous

pebble substrate [17] We set up 12 sites (length 90–185 m)

in the stream (Fig.1) Sites 1–4 were situated in an urban

zone with little riparian vegetation, whereas sites 5–12 run

through mixed forest Salmonid species tended to stay in the

same local habitat (e.g., pool habitat) with the occurrence of

intra- and interspecific competition [18] Thus, we assumed

that fish movements among sites due to competition did not

occur The stream is located between 10 and 70 m above sea

level, and the water temperature ranged from 4 to 12°Cduring the survey

Brown trout were first found in Mamachi stream in themid-1980s, and are now distributed throughout the stream([17]; Saneyoshi, personal communication) Native stream-dwelling salmonids include masu salmon and white-spottedcharr However, white-spotted charr was replaced bybrown trout in the 1990s (Saneyoshi, personal communi-cation), most likely as a result of hybridization [19] andinterspecific competition [6] In addition, native sculpinCottus nozawae and stone loach Noemacheilus barbatulus

as well as the introduced rainbow trout were present in thestream With the exception of stone loaches at sites 1–3, weobserved very low numbers of these species

Population estimation, diet analysis, and foodavailability

We estimated the population size and collected stomachcontent samples every two months between the end of May

2009 and the end of March 2010 However, we wereunable to access sites 9–12 in January because of heavysnow To estimate the population size at each site, weperformed three removal passes during each sampling periodusing a model 12 backpack electrofisher (Smith-Root Inc.,

N

Hokkaido Chitose River

Lake Shikotsu

Sea of Japan

1 km

12345

678

9101112

Chitose River

Mamachi Stream

G-A area

G-P area

‘‘Gam-maridae-present’’ and ‘‘Gammaridae-absent’’ areas, respectively

Vancouver, WA, USA) To estimate fish density, the

dimensions of each study site were calculated following the

method of Hasegawa et al [20] using river width data from

summer 2009 We assumed the dimensions were constant

throughout the year, as the water level in Mamachi stream

was stable Following capture, the fish were anesthetized

using ethyl 3-aminobenzoate methanesulfonic acid, and

then weighed and measured (fork length) to the nearest

0.1 g and 1 mm, respectively In addition, we haphazardly

selected up to about 20 individuals each of masu salmon

and brown trout and sampled their stomach contents by

gastric lavage Hasegawa et al [9] described that fry

appeared in March, May, and July (although not in March

for brown trout fry) However, a clear definition of the

boundary line between the fry and parr stages does not

exist Thus, we used fishes that were apparently in the parr

stage (i.e., larger than 70 mm in fork length for both

spe-cies) for stomach content sampling Sampling times of

stomach contents at each site and during each month were

random in the daytime The concentration of anesthetic

was relatively low, such that anesthetized fish recovered

within about 3 min after processing

We also collected samples of benthic and drifting

ter-restrial invertebrates at each site to evaluate the availability

of food resources We collected six samples of benthic

invertebrates using a Surber sampler (25 9 25 cm

quad-rat) Drifting terrestrial invertebrates were collected once at

the upstream margin of each site using three drift nets

(25 9 25 cm opening) The nets were placed in a riffle for

30 min during the daytime We measured the current

velocity at the center of the net opening to calculate the

volume of water passing through the net

The samples of stomach contents and benthic and drifting

terrestrial invertebrates were preserved in 70 % ethanol in

the field then sorted and identified to the order level in the

laboratory (except for the terrestrial invertebrates, which

were treated as a single category) The sorted samples were

dried at 60°C for 24 h and weighed to the nearest 0.0001 g

We used the mean density of six samples for benthic

inver-tebrates (g/m2) and three samples for drifting terrestrial

invertebrates (g/m3) as an index of the availability of food

resources at each site during each survey period

Data analysis

The overlap in diet composition between masu salmon and

brown trout was quantified using a proportional similarity

index [21,22]:

PS¼ 1  0:5Xm

i¼1

jMSi BTi

where MSiand BTi represent the dry mass proportions of

prey category i (among m categories) for masu salmon and

brown trout, respectively The index ranged from 0 (nooverlap) to 1 (complete overlap) We evaluated differences

in PS between sites with (G-P area) and without (G-A area)Gammaridae using a two-way repeated measures ANOVA(see ‘‘Fish density and food availability’’ in ‘‘Results’’).Then we used Pearson’s correlation test to evaluate therelationship between PS and stomach fullness (wet mass ofstomach contents/body weight) in masu salmon and browntrout for each month to determine the effect of competitionfor food resources on foraging efficiency in both species Atwo-way ANOVA followed by Scheffe’s test was used tocompare mean fork length between the G-P and G-A areas,and among months for both brown trout and masu salmon.The alpha level was set at 0.05

ResultsFish density and food availabilityThe abundance (fry and parr in total) of masu salmon wasgenerally higher than that of brown trout throughout theyear in Mamachi stream (Fig.2)

We observed the typical spatial differences in benthicinvertebrate composition corresponding to the presence ofGammaridae throughout the year at sites 1–4 (Fig.3).Hereafter, these sites are referred to as the ‘‘Gammaridae-present’’ (G-P) area, whereas the remaining sites (5–12) arereferred to as the ‘‘Gammaridae-absent’’ (G-A) area(Fig.1) Trichoptera and Ephemeroptera were widely dis-tributed in the stream (Fig.3) Although the abundance ofeach category fluctuated dramatically from month tomonth, the presence/absence pattern of each category ateach site was almost same during the survey (Fig.3).The dynamics of the abundances of drifting terrestrialinvertebrates were similar for the G-P and G-A areas

parr) at 12 sites (except in January, when 8 sites were monitored) Circles with a solid line and triangles with a dashed line indicate masu salmon and brown trout, respectively

(Fig.4) We observed relatively few drifting terrestrial

invertebrates in winter (November to March) In contrast,

we captured terrestrial invertebrates at all sites during the

summer (May to September), though the density of

ter-restrial invertebrates was quite variable

Dietary overlap between masu salmon and brown trout

We included 931 brown trout (mean ± SD

165.0 mm ± 59.2, range 70–500 mm fork length) and

1176 masu salmon (111.4 mm ± 20.4, 73–194 mm fork

length) in the diet analysis In detail, both brown trout and

masu salmon in the G-P area tended to be larger than those

in the G-A area, with mean fork length varying among

months (Table1; Fig.5)

A two-way repeated-measures ANOVA revealed that

months and months 9 presence of Gammaridae

(interac-tion) had significant effects (Table2; Fig.6) During May

and July, when terrestrial invertebrates were abundant(Fig.4), PS was smaller in G-P than in G-A (Fig 6) TheGammaridae comprised a higher proportion of the diet ofbrown trout than the diet of masu salmon, and vice versafor the proportion of terrestrial invertebrates, during Mayand July in G-P (Fig 7) Conversely, both species shared asimilar diet in the G-A reaches Trichoptera, Ephemerop-tera, and terrestrial invertebrates dominated their diets inMay, whereas terrestrial invertebrates were dominant inJuly (Fig.7) When terrestrial invertebrates were scarce(between September and March) (Fig.4), the difference in

PS between G-P and G-A was smaller than the sponding difference during May and July (Fig.6) In G-P,both species preyed primarily on Gammaridae, thoughmasu salmon also preyed upon terrestrial invertebrates inSeptember (Fig.7) In G-A, both species preyed primarily

corre-on terrestrial invertebrates in September, and Trichopteraand Ephemeroptera in January and March (Fig.7)

Gammaridae Trichoptera Ephemeroptera

0 1 2 3

November

0 1 2 3

January

0 1 2 3

sites without Gammaridae The y-axis scale used for May is different from that used for the other months

However, the dietary contribution from Trichoptera and

Ephemeroptera differed for the two species in November,

leading to the large difference in PS (Fig.7)

PS was higher in winter (November, January, and

March) than in summer (May, July, and September) in the

G-P reach, whereas PS was similar during all months

except for November in the G-A reaches (Fig.6) There

was no relationship between PS and stomach fullness in

masu salmon or brown trout, except for in masu salmon

during May (Pearson’s correlation test: P [ 0.096; masu

salmon in May: r = -0.598, P = 0.040)

Brown trout predation on masu salmon

Of the 931 brown trout we examined, 28 individuals had

consumed 46 masu salmon Some individuals also preyed

on chum salmon fry, brown trout fry, sculpin, and stone

loach More than half of the brown trout were smaller than

300 mm in fork length, which was regarded as the

threshold body size for the occurrence of piscivory [4]

(Fig.8) In general, however, fish contributed very little to

the diet of the brown trout in Mamachi stream throughoutthe year (Fig.7)

DiscussionAlthough the body sizes of brown trout and masu salmonwere different for the G-P or G-A areas, and amongmonths, the difference did not appear to be large enough toproduce a significant difference in the food habits ofindividuals Thus, we assumed that the food habit differ-ences shown in this study were caused by differences infood conditions

The degree of dietary overlap (PS) differed among areasand seasons Gammaridae and terrestrial invertebrates play

(a)

(b)

a,ba

a

b

aa

used for diet analysis in each area (black bars G-P area; white bars G-A area) each month Bars with different letters are statistically significantly different based on Scheffe’s test among months The sample sizes for G-P and G-A were the sums of sites 1–4 and 5–12,

on the fork lengths of brown trout and mosu salmon

Jul Sep Nov Jan Mar.

with (G-P, circles with solid line) and without (G-A, triangles with a

dashed line) Gammaridae during each study period

a key role in determining the degree of dietary overlap In

May and July, when terrestrial invertebrates were

abun-dant, the proportion of Gammaridae in the stomach

contents of brown trout was higher than the proportion ofGammaridae in the stomach contents of masu salmon at thesites where Gammaridae were present Masu salmon, inturn, had a higher percentage of terrestrial invertebrates intheir diet than did brown trout This is likely important, asthe dietary overlap was smaller in the Gammaridae-presentarea than in the Gammaridae-absent area The smallerdietary overlap in the Gammaridae-present area may not bedue to interspecific competition With regards to species-specific habitat preferences, brown trout typically occupypositions close to the substrate, whereas masu salmonoccupy the surface or midrange of the water column [11].Inoue et al [13] demonstrated that species occupying thebottom range prey on Gammaridae, whereas speciesoccupying other ranges tend to prey on terrestrial inverte-brates, thus avoiding interspecific competition Therefore,the presence of Gammaridae may mitigate competition forfood resources between masu salmon and brown trout.Microhabitat use in the Gammaridae-present and Gam-maridae-absent areas must be evaluated to confirm thisidea

(G-P, circles with a solid line) and without (G-A, triangles with a

dashed line) Gammaridae

each food item in the diets of

brown trout and masu salmon at

each site during each study

period Numbers on each bar

indicate the number of fish used

in the diet analysis Sites 1–4

represent areas with

Gammaridae (G-P) and 5–12

represent sites without

Gammaridae (G-A)

Between the summer and winter seasons, we observed

an increase in the degree of dietary overlap in the

Gam-maridae-present area as the abundance of terrestrial

invertebrates decreased Thus, the difference in dietary

overlap between the Gammaridae-present area and the

Gammaridae-absent area was reduced In November, the

interspecific difference in the proportion of the diet sponding to Trichoptera explained the decline in the dietaryoverlap in the Gammaridae-absent area and the large dif-ference in the overlaps for the Gammaridae-present areaand the Gammaridae-absent area However, there was nodifference in the composition of benthic invertebrates inNovember relative to the remaining winter months, so it isunclear why the diets and dietary overlap of brown troutand masu salmon differed during this month

corre-Overall, the degree of dietary overlap obviously depended

on prey composition; it did not correlate with stomach ness in masu salmon and brown trout, though the randomsampling times for the stomach contents may have obscuredthe relationships, because salmonid species have a diurnalfeeding rhythm [23] This suggests that competition for foodresources does not occur intensely, because each individualhas access to adequate food resources, even in instanceswhere masu salmon and brown trout prey on similar fooditems However, we may need to evaluate growth rates underdifferent food conditions in order to detect the detailed effect

brown trout individuals that preyed on masu salmon

It is generally recognized that piscivory is performed by

larger brown trout individuals, particularly those [300 mm

in total length [4] The larger sizes of these individuals

confer several advantages, including larger gapes and

higher mobilities However, we noted that trout smaller

than 300 mm were also able to prey on masu salmon This

is consistent with the observations of Mayama [7], who

concluded that the body size of the brown trout did not

influence the occurrence of piscivory Instead, the relative

difference in body size between the brown trout and the

masu salmon is more important However, our results

indicate that native masu salmon do not constitute a large

proportion of the diet of brown trout under food conditions

such as those present in Mamachi stream This implies that

predation by brown trout may not lead to a dramatic

decline in the masu salmon population, as previous studies

have pointed out [4,24]

Our results showed that prey composition (presence of

Gammaridae and terrestrial invertebrates), which varies

depending on the conditions present (which in turn are

dependent on the location and season), affects the dietary

overlap between introduced brown trout and native masu

salmon However, competition between these species for

food resources is unlikely to be intense Predation was also

rare Although these interactions may not lead to the

replacement of masu salmon by brown trout in Mamachi

stream at present, further studies are required to understand

their interactions For example, we may also need to

evaluate details of the interspecific competition, such as the

growth rate under each food condition In addition, it is

necessary to compare sympatric and allopatric situations to

confirm the occurrence of interspecific competition

Fur-ther studies should also determine how the seasonal

vari-ation in dietary overlap at a particular locvari-ation affects the

outcome of interspecific competition throughout the year

Resource Center for their support during this study We thank

members of the following groups at Hokkaido University for their

assistance with the fieldwork: the Field Science Center of Northern

Biosphere, the Laboratory of Animal Ecology, and Tsuri-Aikoukai.

We also thank Motohiro Kikuchi from Chitose Salmon Aquarium as

well as Takashi Teramoto and his students from the Chitose Institute

of Science and Technology for their kind support during the study.

Dr Mineo Saneyoshi kindly provided us with information on the

species replacement of white-spotted charr by brown trout in

Ma-machi stream This study was supported in part by Grants-in-Aid for

Postdoctoral Research Fellows to KH from the Japan Society for the

Promotion of Science.

References

1 Fausch KD (1988) Tests of competition between native and

introduced salmonids in streams: what have we learned? Can J

Fish Aquat Sci 45:2238–2246

2 Krueger CC, May B (1991) Ecological and genetic effects of salmonid introductions in North America Can J Fish Aquat Sci 48(suppl 1):66–77

3 Baxter CV, Fausch KD, Murakami M, Chapman PL (2004) Fish invasion restructures stream and forest food webs by interrupting reciprocal prey subsidies Ecology 85:2656–2663

4 Lowe S, Browne M, Boudjelas S, De Poorter M (2000) 100 of the world’s worst invasive alien species Invasive Species Specialist Group of the World Conservation Union, Auckland

5 Takami T, Yoshihara T, Miyakoshi Y, Kuwabara R (2002) Replacement of white-spotted charr Salvelinus leucomaenis by brown trout Salmo trutta in a branch of the Chitose River, Hokkaido Nippon Suisan Gakkaishi 68:24–28 (in Japanese with English abstract)

6 Hasegawa K, Maekawa K (2009) Role of visual barriers on mitigation of interspecific interference competition between native and nonnative salmonids Can J Zool 87:781–786

7 Mayama H (1999) Predation of juvenile masu salmon corhynchus masou) and brown trout (Salmo trutta) on newly emerged masu salmon fry in the Chitose River Bull Natl Salmon Resour Center 2:21–27 (in Japanese with English abstract)

(On-8 Aoyama T, Kudo S, Shimoda K, Ueda M (2002) Ribbon tags from brown trout stomach! Uo to Mizu 38:39–41 (in Japanese)

9 Hasegawa K, Yamazaki C, Ohkuma K, Ban M (2012) Evidence that an ontogenetic niche shift by native masu salmon facilitates invasion by nonnative brown trout Biol Invasions 14:2049– 2056

10 Hasegawa K, Yamamoto T, Murakami M, Maekawa K (2004) Comparison of competitive ability between native and introduced salmonids: evidence from pairwise contests Ichthyol Res 51:191–194

11 Hasegawa K, Maekawa K (2006) The effect of introduced monids on two native stream-dwelling salmonids through inter- specific competition J Fish Biol 68:1123–1132

sal-12 Nakano S (1995) Individual differences in resource use, growth and emigration under the influence of a dominance hierarchy in fluvial red-spotted masu salmon in a natural habitat J Anim Ecol 64:75–84

13 Inoue M, Miyata H, Tange Y, Taniguchi Y (2009) Rainbow trout (Oncorhynchus mykiss) invasion in Hokkaido streams, northern Japan, in relation to flow variability and biotic interactions Can J Fish Aquat Sci 66:1423–1434

14 Nakano S, Murakami M (2001) Reciprocal subsidies: dynamic interdependence between terrestrial and aquatic food webs Proc Natl Acad Sci USA 98:166–170

15 Mori T, Mukakami M, Saitoh T (2010) Latitudinal gradients in stream invertebrate assemblages at a regional scale on Hokkaido Island, Japan Freshw Biol 55:1520–1532

16 Liesenjohann M, Liesenjohann T, Trebaticka L, Haapakoski M, Sundell J, Ylo¨nen H, Eccard JA (2011) From interference to predation: type and effects of direct interspecific interactions of small mammals Behav Ecol Sociobiol 65:2079–2089

17 Hasegawa K, Maekawa K (2006) Effect of habitat components on competitive interaction between native white-spotted charr and introduced brown trout J Freshw Ecol 21:475–480

18 Nakano S (1995) Competitive interactions for foraging habitats in a size-structured interspecific dominance hierarchy of two sympatric stream salmonids in a natural habitat Can J Zool 73:1845–1854

micro-19 Kitano S, Hasegawa K, Maekawa K (2009) Evidence for specific hybridization between native white-spotted charr and non-native brown trout on Hokkaido island, Japan J Fish Biol 74:467–473

inter-20 Hasegawa K, Yamamoto T, Kitanishi S (2010) Habitat niche separation of the nonnative rainbow trout and native masu salmon

in the Atsuta River, Hokkaido, Japan Fish Sci 76:251–255

21 Feinsinger P, Spears EE, Poole RW (1981) A simple measure of

niche breadth Ecology 62:27–32

22 Nakano S, Fausch KD, Kitano S (1999) Flexible niche partitioning

via a foraging mode shift: a proposed mechanism for coexistence in

stream-dwelling charrs J Anim Ecol 68:1079–1092

23 Young MK, Rader RB, Belish TA (1997) Influence of vertebrate drift and light on the activity and movement of Col- orado River cutthroat trout Trans Am Fish Soc 126:428–437

macroin-24 Townsend CR (1996) Invasion biology and ecological impacts of brown trout Salmo trutta in New Zealand Biol Conserv 78:13–22

O R I G I N A L A R T I C L E Fisheries

A comparative study of sexual product quality in F1 hybrids

of the bream Abramis brama 3 the silver bream Blicca bjoerkna

Billy Nzau Matondo•Michae¨l Ovidio•

Jean-Claude Philippart•Pascal Poncin

Received: 23 July 2012 / Accepted: 20 September 2012 / Published online: 18 October 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract The gonadosomatic index at spawning,

abso-lute fecundity, and egg size for the female sexual products

as well as the density and consistency of semen for the

male sexual products were examined in cultured Abramis

brama 9 Blicca bjoerkna F1 hybrids and compared with

the parental species at their first sexual maturity Females

ovulated under environmental conditions, and their eggs

were weighed, counted and measured Semen of males was

macroscopically examined and spermatozoa counted using

a hemocytometer Results revealed that hybridization

affected the quality of female and male gametes but with

an overlap between hybrids and parents The

gonadoso-matic index and fecundity were significantly lower than

those of parental species Egg sizes in hybrids showed a

parental effect but to the benefit of hybrids Semen of

hybrids was more diluted which was classified into two

groups: the white semen overlapping slightly with parents

and the aqueous without any overlap with parents

Over-lapped areas between hybrids and parents in term of quality

of sexual products could translate that females and males of

these hybrids have the biological capacity to produce high

quality gametes and thus, a greater chance to produce F2

and backcross generations in rivers

Keywords Semen Eggs  Quality  Hybrids  Breams 

Fish

IntroductionCultured F1 hybrids of Abramis brama 9 Blicca bjoerknapresent successful viability in terms of age and size, fer-tility, and sexual activity at first maturity [1] The observedreproductive success of hybrids raised an important ques-tion about the exact nature of the phylogenetic relationshipbetween these two species We and others have suggestedthat it would be better to combine these two species, cur-rently belonging to a different genus status, in the samegenus [2 4] Several studies on F1 hybrids of A bra-

ma 9 Rutilus rutilus and B bjoerkna 9 R rutilus showed

a real weakness in their fertility, particularly in the duction of F2 generations [5 10] This prompted us to testthe quality and quantity of the gametes of these twohybrids, with results showing a depression of the absolutefecundity and sperm density Whereas a slight overlap inthe fecundity was found between these hybrids and theirparents, no possible overlap was observed concerningsperm density

pro-In hybrid as in nonhybrid fish species, the quality andquantity of gametes play a significant role in the develop-mental success of the crosses [11–13], and the study of thequantity and quality of eggs and sperm should be a majorfocus of all crossbreeding programs In hybrids of A bra-

ma 9 B bjoerkna, a good quality of eggs and spermresulted in the high survival of offspring [1], but the levels

of gametes produced by these hybrids compared to theirparents in terms of sperm density, fecundity, and egg sizeremain unknown It is well known that high sperm density,high fecundity, and large egg size are considered as anadvantage Indeed, high sperm density is often associatedwith high fertilization rates [14], large eggs with importantreserves of nutrients useful for survival and growth afterhatching [15, 16], and high fecundity with better

Biology of Behaviour Unit, Laboratory of Fish Demography

and Hydroecology, University of Lie`ge, 10 Chemin de la Justice,

4500 Tihange, Belgium

e-mail: bnmatondo@ulg.ac.be

P Poncin

Biology of Behaviour Unit, Laboratory of Fish Ethology,

University of Lie`ge, 22 Quai Van Beneden, 4020 Lie`ge, Belgium

DOI 10.1007/s12562-012-0564-7

recruitment in natural populations after hatching [17].

Studying these characteristics may help to evaluate the

reproductive performance of these hybrids compared to

their parents, but also compared with other hybrids such as

F1 hybrids of A brama 9 Rutilus rutilus and B

bjoer-kna 9 R rutilus produced and reared in similar conditions

as these hybrids, and may help to better understand the

ecological impact of these hybrids in natural populations of

parental species

It is therefore important to establish whether these

hybrids are capable of producing gametes of the same

quality as those in the parental species Thus, in this study,

we aimed to evaluate the reproductive capacity of F1

hybrids of A brama 9 B bjoerkna compared to parental

species at their first sexual maturity with the specific

objectives of: (1) analyzing the female sexual products

with regard to the gonadosomatic index at spawning (GSI),

absolute fecundity, and egg size, and (2) estimating the

quality of semen by analyzing the density and consistency

of the sperm

Materials and methods

Production of F1 generation

The mature fish examined in this study (Table1) were

produced from an experimental hybridization made using

mature specimens of A brama and B bjoerkna These

mature specimens were captured in a fish pass at the Lixhe

Dam (Belgian River Meuse, 50°450N; 5°400E) [18] and

were morphologically identified following the descriptions

made by Spillman [19] The experimental hybridization

program was conducted to obtain the specimens of F1

generations: hybrids of AB (from A brama male 9 B

bjoerkna female) and BA (B bjoerkna male 9 A brama

female), and parental species of A brama and B bjoerkna

[1] Hybrids and parental species were reared in captivity at

20°C until their first sexual maturity at the Tihange

aquaculture station in Belgium

Female sexual product analysis

To evaluate the GSI, absolute fecundity, and egg size, ten

gravid females for each type of F1 hybrid and parental

species were selected and placed to reproduce with their

corresponding fish males used in the sperm examination

Ovulation was observed under environmental conditions at

20°C, 16 L/8 D photoperiod, spawning substrate

simu-lating vegetation, in a 0.92 9 0.40 9 0.40-m experimental

nylon basket installed in a 6.00 9 1.00 9 0.67-m tank,

which was linked to an isolated recirculating system The

GSI was expressed as the percentage of egg weights

spawned per total body weight, the absolute fecundity wasconsidered as the total number of eggs spawned per femaleand calculated from two samples of eggs (1 g), and the eggsize was determined from samples of 50 eggs per femalethat were individually measured using a microscope fittedwith an ocular micrometer [9,10]

Sperm evaluation

To estimate the density and consistency of sperm, semen often selected males for each type of F1 hybrid and parentalspecies was individually extracted by abdominal pressure.The consistency of sperm was determined as lactic whensemen was a white liquid or as watery when semen was agray liquid by macroscopic examination The sperm con-centration was back-calculated after counting the sperma-tozoa in milt extracted with a syringe and diluted 200-foldwith an extender, a bicine solution at pH 7.8 [9,10, 20].Spermatozoa were counted in 30 random cases(0.0025 mm2) for a hemocytometer (Bu¨rker’s cell) on aphase contrast microscope (9400)

Statistical analysisComparisons of GSI, fecundity, egg size, and spermdensity between hybrids and parental species were madeusing the Kruskal–Wallis (KW) test followed by multiplepaired comparisons tests, the Mann–Whitney U-test.Fisher’s exact probability (FEP) test was used to com-pare the relative frequency of sperm consistency inhybrids and parents, and the Chi-square (v2) test tocompare lactic and watery sperm in hybrids For allstatistical analyses, a probability level of p \ 0.05 wasconsidered significant

species used in sexual product analysis

A, Abramis brama; B, Blicca bjoerkna; AB and BA, F1 hybrids;

n number of fish studied For the same feature, mean values sharing the same letter in superscripts are not significantly different (Mann– Whitney U-test, p \ 0.05)

Female sexual product

Analysis of the sexual products from the selected females

revealed (Figs.1, 2) that hybridization of A brama 9 B

bjoerkna affected the quality of their sexual products in

terms of GSI (KW test: df = 3, H = 17.651, p = 0.0005),

absolute fecundity (KW test: df = 3, H = 15.272,

p = 0.0016), and egg diameters (KW test: df = 3,

H = 562.077, p \ 0.0001), but with an overlap between

hybrids and parents The Mann–Whitney U-test showed a

significantly lower GSI and fecundity (U-test, p \ 0.05) in

AB and BA hybrids (median values 12.2 and 10.8 % for

GSI, and 6.5 9 103 and 4.7 9 103 eggs for fecundity,

respectively) than those of the parents of B bjoerkna

(15.6 % and 8.8 9 103 eggs) (Fig.1a, b) However, the

parents of A brama (14.6 % and 11.9 9 103 eggs) were

not found to be significantly different (U-test, p [ 0.05) to

hybrids and to B bjoerkna species The frequency

distri-bution of egg sizes showed the paternal effect in hybrids

but to their advantage (Fig.2a) AB hybrids (mean value of

egg size: 1.4 mm) were closer to the male parental species,

the A brama (1.3 mm), and the BA hybrids (1.6 mm) were

closer to B bjoerkna (1.5 mm) (Fig.2b) In AB hybrids,

two peaks of egg size were observed around 1.3 and 1.5

mm, respectively The difference in egg size between

hybrids and their parents was found to be significant

(U-test, p \ 0.05)

Semen analysisThe comparison of semen (Fig.3) also revealed a signifi-cant effect of the hybridization process of A brama 9 B.bjoerkna concerning sperm density (KW test: df = 3,

H = 28.898, p \ 0.0001) and sperm consistency (FEPtest, p \ 0.05) AB and BA hybrids (median values =0.2 9 1010 and 0.3 9 1010 spermatozoa ml-1, respec-tively) showed substantially lower sperm density (U-test,

p\ 0.05) than parental species (1.2 9 1010and 1.4 9 1010spermatozoa ml-1 for B bjoerkna and A brama, respec-tively) (Fig 3a) A slight overlap was observed betweenthe BA hybrids and their parents The difference was notsignificant between hybrids, but between parental species,the sperm density of A brama was found to be significantlyhigher than B bjoerkna For the consistency of sperm, thepercentage of males with sperm of lactic consistency wassignificantly higher (FEP-test, p \ 0.05) in both parentalspecies, accounting for 100 %, whereas that of AB and BAhybrids accounted only for 50 % and 60 %, respectively(Fig.3b) In these hybrids, no significant difference wasfound (v2-test, p [ 0.05) between lactic sperm and waterysperm in terms of proportion of fish However, the spermconsistency of hybrids significantly affected their sperma-tozoa concentration (KW-test: df = 3, H = 16.012,

p = 0.0011) A significantly higher spermatozoa tration (U-test, p \ 0.05) was observed in lactic sperm(0.47 9 1010 and 0.46 9 1010 spermatozoa ml-1 for ABand BA hybrids, respectively) than in watery sperm(0.17 9 1010 and 0.10 9 1010 spermatozoa ml-1 for ABand BA hybrids, respectively) (Fig 3c)

concen-DiscussionThe GSI and fecundity of hybrids were found to be low butwith an overlap with parents In these hybrids as in other B.bjoerkna hybrids, neither GSI nor fecundity can clearlydifferentiate between hybrids and parents [9] The egg sizealso showed the same trend but with a slight advantage forthe hybrids This can be considered as a benefit for hybrids,since larger eggs are associated with more nutrient reservesuseful for survival and growth after hatching [15,16] Forthese three criteria characterizing the female sexual productanalysis, an overlap observed between hybrids and theirparents could mean that these hybrids have a high repro-ductive capacity and a successful gametogenesis process.Thereby, our criteria may be inadequate for analyzingsexual products of females to distinguish the hybrids of A.brama 9 B bjoerkna from their parents

In the first sexual maturity period, the GSI and fecundity

of these hybrids were low compared to those of R lus 9 B bjoerkna F1 hybrids [9], but egg sizes were veryHybrids and parental species

a a b ab

fecundity, and egg diameters for hybrids and parent species A,

Abramis brama; B, Blicca bjoerkna; AB and BA, F1 hybrids Values

of GSI and fecundity are median, percentiles 5, 25, 75, and 95, the

horizontal line inside the box marks the position of the median and

circles indicate minimal and maximal values; n = 10 fish for GSI and

fecundity; hybrids or parental species marked with the same letter are

not significantly different (Mann–Whitney U-test, p \ 0.05)

similar between these two types of hybrids On the

con-trary, the fecundity of these hybrids was found to be higher

than that of R rutilus 9 A brama F1 hybrids [10] The

GSI and fecundity of hybrids was not significantly different

from the parent species of A brama, indicating that these

hybrids can match the high reproductive performance of

their parents, which, according to Karjalainen et al [17],

means more chances of recruitment in rivers The GSI and

fecundity of hybrids were closer to A brama, a finding that

is not surprising, as in our previous study we showed thatthese hybrids were also closer to the parent species of A.brama with regard to age, size, and reproductive tactics atfirst sexual maturity [1] In contrast, in F1 hybrids of R.rutilus 9 A brama the absolute fecundity was significantlower than in parental species [10], suggesting a lowgametogenesis efficiency In F1 Clariidae hybrid fish,considerably lower GSI and fecundity than in parentalspecies were also found [21] In AB hybrids, two peaks of

Hybrids and parental species Egg diameter, (mm) 1

1.2 1.4 1.6

1.8

1.6

1.4 1.3

100 200 0 100 200 0 100 200

Egg size, (mm)

mean, and range values of egg

sizes for hybrids and parent

species A, Abramis brama; B,

Blicca bjoerkna; AB and BA,

F1 hybrids n = 500 eggs from

10 females; hybrids or parental

species marked with the same

letter are not significantly

AB/L AB/W BA/L BA/W

parent species A, Abramis brama; B, Blicca bjoerkna; AB and BA,

F1 hybrids; L, lactic or white sperm; W, watery or aqueous sperm.

Values of sperm density are median, percentiles 5, 25, 75 and 95, the

horizontal line inside the box marks the position of the median and

circles indicate minimal and maximal values; values of sperm consistency are in percentage (n = 10) hybrids or parental species marked with the same letter are not significantly different (Mann– Whitney U-test, p \ 0.05); *p \ 0.05 (FEP-test)

egg size observed matching mean values of parental

spe-cies could be well related with the biological capacity of

hybrid offsprings to produce gametes with different ploidy

levels which may be confirmed according to Liu et al [22,

23] by genetic analysis

The significantly lower sperm density in hybrids of A

brama 9 B bjoerkna than in parents is a common finding

with other hybrid cyprinid fish [9, 10], meaning a low

efficiency of spermatogenesis in hybrids Overall, the

spermatozoa concentration could thus contribute toward

differentiating these hybrids from their parents

Accord-ing to Stoumboudi et al [24], the spermatozoa index may

be a more accurate indicator of both testicular activity and

the timing of reproductive activity than the GSI In terms

of median values, the sperm density of these hybrids was

found to be very similar to that observed in F1 hybrids of

R rutilus 9 A brama and R rutilus 9 B bjoerkna [9,

10] Undoubtedly, sharing a common parental species

between these three types of hybrids and the similar

rearing conditions could well explain this finding

How-ever, in hybrids of A brama 9 B bjoerkna, higher

maximal values and white semen overlapping with

par-ents were observed, which again means a high

repro-ductive capacity for these hybrids, and, according to

Leong [14], a high fertilization rate and, thus, a higher

chance of reproducing F2 and backcross offsprings in

rivers The translucent or aqueous semen, extremely

diluted sperm, that we observed is not limited to these

hybrids It has already been found in other hybrid fish

belonging to the Clariidae family [21]

This study has demonstrated that F1 hybrids of A

bra-ma 9 B bjoerkna have a high reproductive capacity, and

the quality of their sexual products shows an overlap with

the parental species This high reproductive performance

could translate into a higher chance of these hybrids

reproducing their post-F1 generations in natural

popula-tions of parental species The new reproductive success of

these hybrids again raises the question about the

phyloge-netic relationship between their parental species, which

would fit better within the same genus rather than in two

genera as is currently the case The overlap observed was

more significant for the sexual products of female hybrids

in terms of GSI, fecundity, and egg size than for the sexual

products of the males Using sexual products of males with

criteria such as semen density or consistency could be more

useful for hybridization analyses of these hybrids than the

analysis of female sexual products

Y Neus and A B Nlemvo for their help with the field and laboratory

work We greatly appreciate the comments and suggestions made by

two anonymous reviewers that led to improvement of this manuscript.

Financial support for this research was provided by F.R.F.C grants

N°1482 and 1.5.120.04.

References

1 Nzau Matondo B, Ovidio M, Philippart JC, Poncin P (2008) Sexual maturity, reproductive behaviour and fertility in the first- generation hybrids of Blicca bjoerkna (L.) x Abramis brama L Cybium 32:286–289

2 Freyhof J (1998) Riffle spawning white bream Abramis bjoerkna (Cyprinidae) in a regulated river: a case of unsuitable habitat choice? Ital J Zool 65:441–444

3 Wolter C (1999) Comparison of intraspecific genetic variability

in four common cyprinids, Abramis brama, Abramis bjoerkna, Rutilus rutilus and Scardinius erythrophthalmus, within and between lowland river systems Hydrobiologia 394:163–177

4 Ha¨nfling B, Brandl R (2000) Phylogenetics of European nids: insights from allozymes J Fish Biol 57:265–276

cypri-5 Wood AB, Jordan DR (1987) Fertility of roach 9 bream hybrids, Rutilus rutilus (L.) 9 Abramis brama (L.) and their identifica- tion J Fish Biol 30:249–261

6 Pitts CS, Jordan DR, Cowx IG, Jones NV (1997) Controlled breeding studies to verify the identity of roach and common bream hybrids from a natural population J Fish Biol 51:686–696

7 Yakovlev VN, Slyn’ko YuV, Grechanov IG, Krysanov EYu (2000) Distant hybridization in fish J Ichthyol 40:298–311

8 Hayden B, Pulcini D, Kelly-Quinn M, O’Grady M, Caffrey J, McGrath A, Mariani S (2010) Hybridisation between two cyp- rinid fishes in a novel habitat: genetics, morphology and lifehi- story traits BMC Evol Biol 10:169

9 Nzau Matondo B, Nlemvo AB, Ovidio M, Philippart JC, Poncin P (2008) Fertility in first-generation hybrids of roach, Rutilus ru- tilus (L.), and silver bream, Blicca bjoerkna (L.) J Appl Ichthyol 24:63–67

10 Nzau Matondo B, Ovidio M, Philippart JC, Poncin P (2011) Reproductive behaviour and sexual production in the first-gen- eration hybrids of roach Rutilus rutilus L 9 common bream Abramis brama L J Appl Ichthyol 27:859–867

11 Vuthiphandchai V, Yonathan Zohar Y (1999) Age-related sperm quality of captive striped bass Morone saxatilis J World Aquac Soc 30:65–72

12 Brown Kim H, Gardner-Brown Tara M, Thorgaard Gary H, Quattro JM (2004) Equivalent survival and different development rates in reciprocal apache trout 9 rainbow trout hybrids Copeia 2:378–382

13 Durland ER, Quintero HE, Davis DA, Dunham RA (2010) Influence of forage fish and dietary lipid supplements on egg quality and fry production in channel catfish (Ictalurus puncta- tus) 9 blue catfish (Ictalurus furcatus) hybridization Aquacult Nutr 16:153–162

14 Leong RJH (1989) Sperm concentrations and egg fertilization rates during spawning of captive anchovy, Engraulis mordax CalCOFI Rep 30:136–139

15 Bagenal TB (1969) Relationship between egg size and fry vival in brown trout Salmo trutta L J Fish Biol 1:349–353

sur-16 Berg OK, Hebdry AP, Svendsen B, Bech C, Arnekleiv JV, Lohrmann A (2001) Maternal provisioning of offspring and the use of those resources during ontogeny: variation within and between Atlantic Salmon families Funct Ecol 15:13–23

17 Karjalainen J, Auvinen H, Helminen H, Arjoma¨ki TJ, Niva T, Sarvala J, Vilhanen M (2000) Unpredictability of fish recruit- ment: interannual variation in young–of–the–year abundance.

J Fish Biol 56:837–857

18 Nzau Matondo B, Ovidio M, Poncin P, Kakesa TA, Wamuini LS, Philippart JC (2007) Hybridization success of three common European cyprinid species, Rutilus rutilus, Blicca bjoerkna and Abramis brama and larval resistance to stress Fish Sci 73:1136–1145

19 Spillman CJ (1961) Fauna of France, n° 65 Freshwater fish In:

Paul Le Chevalier (ed), Paris (In French)

20 Moore A (1996) Use of semen extenders for walleye In:

Sum-merfelt RC (ed) Walleye culture manual NCRAC Culture Series

101 North Central Regional Aquaculture Center Publications

Office, Iowa State University, Ames, pp 51–53

21 Legendre M, Teugels GG, Cauty C, Jalabert B (1992) A

com-parative study on morphology growth rate and reproduction of

Clarias gariepinus (Burchell, 1822), Heterobranchus longifilis

Valenciennes, 1840, and their reciprocal hybrids (Pisces,

Clarii-dae) J Fish Biol 40:59–79

22 Liu S, Liu Y, Zhou G, Zhang X, LUO C, Feng H, He X, Zhu G,

Yang H (2001) The formation of tetraploid stocks of red crucian

carp x common carp hybirds as an effect of interspecific hybridization Aquaculture 192(2–4):171–186

23 Liu S, Qin Q, Xiao J, Lu W, Li W, Liu J, Duan W, Zhang C, Tao

M, Zhao R, Yan J, Liu Y (2007) The formation of the polyploidy hybrids from different subfamily fish crossing and its evolution- ary significance Genetics 107:1023–1034

24 Stoumboudi MT, Villwock W, Sela J, Abraham M (1993) nadosomatic index in Barbus longiceps, capoeta damascina and their natural hybrid (Pisces, Cyprinidae), versus spermatozoan index in the parental males J Fish Biol 43:865–875

Go-O R I G I N A L A R T I C L E Biology

Migration route of Pacific saury Cololabis saira inferred

from the otolith hyaline zone

Satoshi Suyama• Masayasu Nakagami•

Miyako Naya•Yasuhiro Ueno

Received: 12 October 2011 / Accepted: 31 July 2012 / Published online: 4 September 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract In order to establish the migration route of Pacific

saury Cololabis saira, we measured the radius of otolith

annual rings (ROA) in fish collected from areas off the

Japa-nese coast up to 165°W in June and July (pre-fishing season)

and from fishing grounds in August–November (fishing

sea-son) The average ROA for six sea areas that each spanned 10°

of longitude sampled during the pre-fishing season were

compared with data obtained during each month of the fishing

season The average ROA decreased from west to east and

also decreased monthly from August to November The

average ROA of fish caught after October at the peak of the

fishing season was equivalent to that of the fish caught in the

areas east of 160°E or 170°E We conclude that Pacific saury

caught by Japanese fishing vessels during the peak of the

fishing season migrate from an area east of 160°E

Keywords Pacific saury  North Pacific  Otolith 

Age determination Fishing season  Annual ring

Introduction

Pacific saury Cololabis saira are distributed from subarctic

to subtropical regions of the North Pacific Ocean, and from

the coast of Japan to the western coast of North America[1] The life span of this fish is approximately 2 years [2],and their spawning period lasts from September to June [3].They migrate northward from the subtropical to the sub-arctic region in May and July, then start to return to thesouth between July and August [4] Although the habitat ofthis species is broad, genetic variance is low, with verylittle genetic structuring [5]

Pacific saury is an important commercial fish in Japan,Russia, Korea and Taiwan, and the total landings by thesecountries increased from 180,973 tons in 1998 to 622,119tons in 2008 (FAOWeb: http://www.fao.org/fi/statist/statist.asp Accessed 28 March 2011) Japanese fishingvessels mainly catch saury in the region between the KurilIslands and the Japanese coast from August to Decemberusing stick-held dip nets, but Taiwanese fishing vesselsoperate over a wider area, including the open sea mainlywest of 160°E, from June to December [6,7] It has alsobeen reported that Taiwanese fishing vessels operated untilOctober in the sea around 170°E in years when fishabundance was high, such as in 1997 [6], suggesting thatsome fish migrate south in autumn via a route that is farfrom the Japanese coast

Stock assessment surveys were carried out by theTohoku National Fisheries Research Institute during Juneand July (pre-fishing season) in areas near Japan and inoffshore regions to the west of 165°W using a sea surfacetrawl net [8] The results of these surveys demonstrated thatthe abundance of the fish west of 150°E was low in the pre-fishing season, indicating that Pacific saury caught byJapanese fishing vessels in autumn had in June and Julybeen distributed far from Japan (east of 150°E) The fishinggrounds of the Taiwanese fishing vessels in the open seaoff Japan move from east of 155°E to the area west of155°E between June and November, also indicating that the

Hachinohe Laboratory, Tohoku National Fisheries Research

Institute, Fisheries Research Agency Japan, 25-259

Shimomekurakubo, Same, Hachinohe, Aomori 031-0841, Japan

e-mail: suyama@affrc.go.jp

Present Address:

Y Ueno

National Research Institute of Fisheries Science,

Fisheries Research Agency Japan, 2-12-4 Fukuura,

Kanazawa, Yokohama, Kanagawa 236-8648, Japan

DOI 10.1007/s12562-012-0546-9

shoals migrate westwards [6] Knowledge of the migration

route of Pacific saury caught in Japanese fishing grounds is

important for determining the appropriate stock abundance

and management measures of this resource In addition, if

the migration process of the shoal of pacific saury from

their pre-fishing season location far from the coast to their

fishing season location near the coast of Japan can be

deter-mined, forecasts of the fishery situation (e.g arrival time at the

fishing grounds and seasonal changes in abundance during the

fishing season) will become more accurate However, at

present the migration route has only been investigated in the

fishing grounds and during the fishing season [4]

Suyama et al [9] indicated that body length frequency

distributions and the radius of otolith hyaline (the distance

from the otolith core to the area where the annual rings

begin to form; ROA; seeMaterials and methods) of age-1

Pacific saury vary geographically, based on samples

col-lected in 2006 They reported that the modes of body

length frequency distributions and ROA were larger for fish

caught west of 160°E than for fish caught east of 160°E By

comparing ROA of age-1 fish collected in June and July

from each of these areas with those of fish caught by

fishing vessels later in the year, it will be possible to

determine whether Pacific saury caught by Japanese fishing

vessels originate in the area east of 160°E

The first objective of this study was to confirm that thegeographical variation in ROA observed by Suyama et al

in June and July 2006 [9] is also seen in the other years.Second, we compared the ROA of fish collected in June toJuly with those caught in the fishing season, betweenAugust and December, in order to infer the migration route

of Pacific saury from offshore areas to the coastal fishinggrounds

Materials and methodsSample collection during the pre-fishing season(June–July)

Age-1 Pacific saury were collected from 150 samplingstations in the central North Pacific Ocean and Japanesecoastal waters by the research vessels Omi-maru, Hokuho-maru and Wakataka-maru between June and July in 2002,

2003, 2004 and 2006 (Fig.1; Table1) using a sea surfacetrawl net The trawling duration (from setting to haulingthe net) was 1 h

The survey region was divided into six areas that eachspanned 10° of longitude: 140°E–150°E, 150°E–160°E,160°E–170°E, 170°E–180°, 180°–170°W, 170°W–160°W;

Aug Sep Oct Nov

Aug

Sep Oct Nov

Aug Sep Oct Nov

Aug Sep Oct Nov

140°E 150°E 160°E 170°E 180° 170°W 160°W

35°N 40°N 45°N

140°E 145°E 150°E

Pacific Ocean

Pacific Ocean

(June and July) (a) and fishing

season (August to November)

(b) sampling stations in the

North Pacific off Japan in 2002,

2003, 2004 and 2006 The

pre-fishing season survey area was

divided into six sub-areas, each

spanning 10° of longitude

these were denoted A140E, A150E, A160E, A170E,

A170W and A160W, respectively

Following capture, samples were frozen and transported

to the laboratory, where the length of each individual was

measured (knob length: the distance from the tip of the

lower jaw to the posterior end of the muscular knob on the

caudal peduncle [10]) to the nearest 0.1 cm At sampling

stations where fewer than 80 individuals were caught, all

individuals were measured and their otoliths extracted At

stations where more than 80 individuals were caught, 80

fish were randomly selected for measurement and otolith

extraction

Sample collection during the fishing season

(August–November)

Samples provided by fishing vessels between August and

November in 2002, 2003, 2004 and 2006 were caught using

stick-held dip nets (Fig.1; Table1) Age-1 Pacific saury

were collected from 97 sampling stations It was not

possible to collect sufficient samples of age-1 fish inDecember, and most of the landed fish consisted of age-0fish in the late fishing season Samples were frozen orrefrigerated following capture and transported to the lab-oratory Measurement of body length and otolith weremade in the same way as for samples collected in the pre-fishing season

Age determinationThe right otolith of each fish was embedded in epoxy resinand examined for age determination by light microscopyfollowing the method described previously [2,9,11,12].Otoliths of age-0 fish do not have annual rings; there isonly an opaque area that appears white when viewed bylight microscopy (type I [12]) In contrast, otoliths of age-1fish have annual rings that appear as a black or dark-grayzone surrounded by an opaque zone(s) We classified eachannual ring as one of three types, as reported in otherstudies [13–16] In the first type, the annual ring was

Pre-fishing season

Area sampled

Area sampled

Unless indicated otherwise, data are presented as the number (n)

sandwiched between two opaque zones and appeared black

when viewed by light microscopy The hyaline zone was

clearly demarcated at both edges This type of otolith was

classified as type III [12] (Fig.2a) In the second type, the

commencement point of the annual ring was clear, but the

end was not; under the microscope, the dark-gray,

trans-lucent area continued from the outer side of opaque zone to

the edge of otolith This type of otolith was classified as

type II [12] (Fig.2b) The third type of otolith, rarely

observed in our study (see Table2), had two hyaline zones

on the outer side of opaque zone (type IV; [12])

Differ-ences between type III and type II and between type II and

type IV, respectively, were often not clear because the

hyaline zone is not distinct Otoliths with no clear hyaline

zone were classified as type II Type I was clearly

different from types II, III and IV due to the lack of ahyaline zone or translucent area The first hyaline zonebegins forming in the autumn, and a hyaline zone ortransparent area under formation can be observed in alltype I samples in February [12, 13] However, the for-mation of a second hyaline zone often occurs in otherseasons [12], and for this reason we did not consider asecond hyaline zone as an annual ring We classified fishwith type I otoliths as age-0 fish, and those with types II,III and IV otoliths as age-1 fish

Fish with otoliths of two additional types were excludedfrom the study: (1) those with otoliths that had more thanthree hyaline zones (seen in three fish) and (2) those withotoliths that were partially or extremely thin, with atransparent part that was different from the hyaline zone Inthe latter case, it is considered that vaterite has replacedaragonite [17]

Radius of annual ring measurementThe ROA, defined as the distance from the otolith core tothe area where the annual ring begins to form [9,11], wasmeasured in age-1 fish with type II, III or IV otoliths(Fig.2a, b) collected in the 2002–2004 pre-fishing andfishing seasons and the 2006 fishing season Measurementswere made using an optical micrometer To measure theROA from fish collected in the pre-fishing season of 2006,otolith images were captured on a PC at 329 magnificationand the ROA calculated using image analysis software(Image X Earth 3.0; Kikuchi Optical Co., Ltd Nagano,Japan)

The average ROA was determined for each of the sixlongitude areas in the pre-fishing season and for eachmonth of the fishing season When the total number of age-

1 fish for each area or each month did not reach 80 viduals, they were excluded from the results

indi-Statistical analysisThe averages and standard deviations (SD) for each 10°longitude area sampled in the pre-fishing season and foreach month sampled in fishing seasons were calculated.Comparisons were made of ROA averages for each lon-gitude area in the pre-fishing season in the same year, andalso between months of the same year in the fishing season

In addition, the average ROA of the longitude areas in thepre-fishing season were compared with those of all months

in the fishing season in all combinations within each year.Differences between mean ROA were assessed usinganalysis of variance (ANOVA) followed by Tukey’s mul-tiple comparisons test Significant differences were deter-mined at the 5 % level

a

b

sandwiched between opaque areas (seen here as white) The edges of the

hyaline zone, i.e the annual ring, are clear (between white arrows).

dark gray) that stretches from the outer side of the opaque zone to the

edge of the otolith The commencement point of the annual ring is distinct

(outside of white arrow), but the end is hard to distinguish Short black

arrows indicate the otolith core (center of otolith), long black arrows

indicate the radius of the annual ring (ROA), from the otolith core to the

point where the annual ring begins Scale bars 1 mm

Geographical variation in ROA in the pre-fishing

season

Average ROA of fish collected in the western areas were

larger than those of fish collected in the eastern areas in all

years The average ROA of the westernmost area exceeded

0.56 mm; in comparison, the average value of the

east-ernmost area was less than 0.53 mm (Fig.3; Table 2)

Significant differences were observed between the

average values of the westernmost and the easternmost

areas in each year (Tukey’s multiple comparison test,

p\ 0.05) For the two or three areas adjacent to each other

(e.g A170E, A170W, and A160W in 2006), significant

differences were rarely observed (p [ 0.05) However,

significant differences were always observed between areas

that were separated by at least two areas (p \ 0.05)

Monthly change in ROA during the fishing season

The average ROA decreased monthly from August to

November in all years Average ROA in August exceeded

0.56 mm in 2004 and 2006 Insufficient samples were

collected in August 2003, but the average ROA in

Sep-tember 2003 was 0.57 mm The monthly change in ROA

was relatively small in 2002 (Fig.3; Table2)

Significant differences were always observed between

the average ROA of August and November in 2002, 2004

and 2006 (p \ 0.05) Samples were not collected in August

2003, but a significant difference was observed between

the average ROA of September and November 2003

(p \ 0.05) The average ROA for November were slightly

larger than those for October in 2003 and 2004, but the

differences were not significant (p [ 0.05)

Comparisons of ROA between the fishing and the

pre-fishing seasons

The average values of ROA between each month of the

fishing season and each area for the pre-fishing season were

compared (Fig.3; Table2) The average ROA for area

A140E, which could be compared only in 2002 and 2006,

were significantly larger than that of any month of the

fishing season (p \ 0.05) The average ROA in August was

not significantly different to those of A150E and A160E in

2002, A150E in 2004 or A160E and A170E in 2006

(p \ 0.05) The ROA in October was not statistically

dif-ferent to values for A170E and A170W in 2002, A160E

and A170E in 2003 and 2004, and to A170E, A170W or

A160W in 2006 (p [ 0.05) The ROA in November was

not statistically different to values for eastern areas (except

in 2003), for 170W in 2002 and for A170E, A170W and

A160E in 2004 and 2006 (p [ 0.05) The ROA inNovember 2003 was not statistically different to that ofA160E (p [ 0.05), but was significantly larger than that ofA170E, A170W and A160W (p \ 0.05)

2004

0.2 0.3 0.4 0.5 0.6 0.7

A140E A150E A160E A170E A170W A160W

Aug Sep Oct Nov

2006

0.2 0.3 0.4 0.5 0.6 0.7

A140E A150E A160E A170E A170W A160W

Aug Sep Oct Nov

2003

0.2 0.3 0.4 0.5 0.6 0.7

A140E A150E A160E A170E A170W A160W

Aug Sep Oct Nov

2002

0.2 0.3 0.4 0.5 0.6 0.7

A140E A150E A160E A170E A170W A160W

Aug Sep Oct Nov

Pre-fishing season Fishing season

A150E A160E A160E

A150E A150E A160E

A160E

A170E A170E A170W A160W

A160E

A170E A170W

A170E A170W A160W A170E A170W

from Pacific saury caught during the pre-fishing and fishing seasons off the coast of Japan Results for each sea area in the pre-fishing season were compared with those of each month of the fishing season Parentheses show the combinations in each area in the pre-fishing season and in each month in the fishing season that were not found to

be significantly different using Tukey’s multiple comparison test (p [ 0.05) For the fishing season data, areas indicated under each bar show the combinations that were not statistically different for

sampling areas

Compared with the other three years, the geographical

and seasonal variation in ROA in 2006 had three distinctive

features First, the average values of ROA for areas west of

160°E exceeded the values in all months of the fishing

season Second, a significant difference in the average

ROA between A170E and August was not observed in

2006 (p [ 0.05) Finally, monthly changes in the average

ROA after September were too small to discern a statistical

difference (p [ 0.05) However, the average ROA at the

peak of fishing season, after October, was close to the

values in the areas east of 160°E in the pre-fishing season

Discussion

The average ROA gradually decreased from west to east in

the pre-fishing season, and in the fishing season, the ROA

decreased each month from August to November The

average ROA in August corresponded to the ROA of areas

west of 160°E or 170°E in the pre-fishing seasons in all

years except 2006 In 2006, the average ROA for the area

west of 160°E was larger than that in August The average

ROA after October corresponded to areas east of 160°E

These patterns were repeated every year As the total catch

per unit effort (one net haul) of Japanese fishing vessels

attains a maximum between late September and the

beginning of November, our results strongly suggest that

the main stock of Pacific saury caught off the coast of

Japan comes from the offshore area (i.e east of 160°E)

However, there were annual fluctuations in ROA that also

suggest it is not possible to predict the exact abundance of

stocks or the time when fish in the offshore area (east of

160°E) will migrate to the fishing grounds; both will

change in response to annual changes in oceanographic

conditions and the abundance of the fish in each area

The eastern boundary of the distribution area from

which Pacific saury migrate to the Japanese coast is not

currently known The average ROA in the 2004 and 2006

fishing seasons decreased from August to November, and

the ROA in November corresponded to those of the areas

east of 170°W Conversely, the average ROA in the late

fishing season in 2003 was higher than that of the areas east

of 180° Notably, the value for November corresponded to

those of the area between 160°E–170°E and was higher

than that of October, which corresponded to the value for

the area between 160°E–180° In that year, the number of

fish that migrated to the south through the offshore area

(east of 170°E) may have been high Huang [6] reported

that Taiwanese fishing vessels, especially vessels operating

around 170°E in years when fish abundance was high, such

as 1997, caught Pacific saury in the open sea after October

This indicates that some age-1 fish do not approach the

coast of Japan, instead migrating south via a route far from

the Japanese coast The proportion of the population thatmigrates via this route will likely fluctuate annually, beinginfluenced by distribution patterns and oceanographicconditions that also vary every year

Based on the occurrence pattern of the parasitic copepodPennella sp., the possibility that Pacific saury recruitsmigrate from west of 160°E to the coast has been previ-ously suggested [18–20] The otolith hyaline zone (ROA)

is a more reliable index for assessing migration because thehyaline zone remains once it has formed, whereas cope-pods may become detached from the fish during migration.The difference in ROA in fish collected between theareas east of 160°E and west of 170°E was clear in 2006.Suyama et al [9] concluded that this difference was caused

by the growth differences during the first year and not byage differences or differences in hatch period within sameage class Analysis of micro growth increments in the ot-oliths revealed that the growth difference between fishfrom these two areas occurred after the northward migra-tion in the first year [9], suggesting that Pacific saury col-lected in our surveys did not migrate widely between thetime when they started their northward migration in thefirst year and when they were surveyed in this study (i.e.June or July of their second year) If two groups that grew

up east of 160°E and west of 170°E were mixed duringtheir first year, the geographical difference in ROA valueswould not be maintained until June/July of their age-

1 year In the other survey years (2002 2003 and 2004), theaverage ROA in the western area was again greater thanthat of the eastern area These results support the twohypotheses that the geographical growth rate differenceoccurs every year and the fish that grow up in the easternand western areas do not mix in their first year

It was not possible to determine whether the hatchingareas of the eastern and western groups were clearly divi-ded However, it is possible that the hatching areas are notdifferent and that at least a part of the eastern group wastransported from the western spawning ground and grew upthere As saury larvae are distributed along the subtropicaland transitional waters from near the Japanese coast to atleast around the international date line [3], and are trans-ported eastward by the Kuroshio Extension [21], we pro-pose that the hatching areas of these two groups are notseparated

It is necessary to identify the eastern boundary of thearea where Pacific saury are distributed in the pre-fishingseason, before they migrate to the Japanese coast To dothis, more eastern areas of the North Pacific will need to beinvestigated in the pre-fishing season, in order to determine

if there is an area east of 165°W where distributiondecreases However, it is possible that any boundary linewould fluctuate annually Distribution patterns in the pre-fishing season, such as abundance, length frequency

distribution and age composition, in each area must be

compared with the respective patterns in the fishing season

to determine if this is indeed the case The first task should

be to survey the area east of 165°W and to compare the

abundance of Pacific saury in the areas east and west of

165°W

Our research reveals that Pacific saury caught off the

coast of Japan perform long migrations in the western half

of the North Pacific Further investigation of the offshore

area east of 160°E will allow for more effective resource

management of Pacific saury in the coastal fishing grounds

off Japan

from the Hokkaido Board of Education, T/S Omi-maru from

Yam-aguchi Prefecture and R/V Wakataka-maru from Tohoku National

Fisheries Research Institute We also thank Ms Junko Momosawa and

Ms Kanae Okabori of the Hachinohe station of the Tohoku National

Fisheries Research Institute for extracting the majority of otoliths and

taking light microcopy photographs This work was partially

sup-ported by the Fisheries Agency of Japan.

References

1 Hubbs CL, Wisner RL (1980) Revision of the sauries (Pisces,

Scomberesocidae) with descriptions of two new genera and one

new species Fish Bull US 77:521–566

2 Suyama S, Kurita Y, Ueno Y (2006) Age structure of Pacific

saury Cololabis saira based on observations of the hyaline zones

in the otolith and length frequency distributions Fish Sci

72:742–749

3 Watanabe Y, Lo NCH (1989) Larval production and mortality of

Pacific saury, Cololabis saira, in the northwestern Pacific Ocean.

Fish Bull US 87:601–613

4 Fukushima S (1979) Synoptic analysis of migration and fishing

conditions of Pacific saury in the northwest Pacific Ocean Bull

Tohoku Reg Fish Res Lab 41:1–70

5 Chow S, Suzuki N, Brodeur RD, Ueno Y (2009) Little population

structuring and recent evolution of the Pacific saury (Cololabis

saira) as indicated by mitochondrial and nuclear DNA sequence

data J Exp Mar Biol Ecol 369:17–21

6 Huang WB, Lo NCH, Chiu TS, Chen CS (2007) Geographical

distribution and abundance of Pacific saury, Cololabis saira

(Brevoort) (Scomberesocidae), fishing stocks in the northwestern Pacific in relation to sea temperatures Zool Stud 46:705–716

7 Huang WB (2010) Comparisons of monthly and geographical variations in abundance and size composition of Pacific saury between the high-seas and coastal fishing grounds in the north- western Pacific Fish Sci 76:21–31

8 Tohoku National Fisheries Research Institute (2010) Annual report

of the research meeting on Pacific saury resources (in Japanese) Tohoku National Fisheries Research Institute, Hachinohe

9 Suyama S, Nakagami M, Naya M, Ueno Y (2012) Comparison of the growth of age-1 Pacific saury Cololabis saira in the western and the central North Pacific Fish Sci 78:277–285

10 Kimura K (1956) The standard length of the Pacific saury, olabis saira (Brevoort) Bull Tohoku Reg Fish Res Lab 7:1–11

Col-11 Suyama S, Oshima K, Nakagami M, Ueno Y (2009) Seasonal change in the relationship between otolith radius and body length

in age-zero Pacific saury Cololabis saira Fish Sci 75:325–333

12 Suyama S, Sakurai Y (2000) Formation period of the otolith hyaline zones of Pacific saury, Cololabis saira (Brevoort) in the western North Pacific Ocean Tohoku Reg Fish Res Lab 63:97–108

13 Hotta H (1960) On the analysis of the population of the Pacific saury (Cololabis saira) based on the scales and the otolith char- acters, and their growth Bull Tohoku Reg Fish Res Lab 16:41–64

14 Sugama K (1957) Analysis of population of the saury (Cololabis saira Brevoort) on the basis of character of otolith-I Bull Hok- kaido Reg Fish Res Lab 16:1–12

15 Sugama K (1959) Analysis of population of the saury (Cololabis saira Brevoort) on the basis of character of otolith-II Bull Hokkaido Reg Fish Res Lab 20:175–191

16 Imai Y (1993) Otolith and scale characteristics of Pacific saury, Cololabis saira, based on otolith type, in the eastern waters off Hokkaido Sci Rep Hokkaido Fish Exp Stn 43:1–10

17 Gauldie RW (1986) Vaterite otoliths from chinook salmon corhynchus tshawytscha) New Z J Mar Fresh Res 20:209–217

(On-18 Nagasawa K, Imai Y, Ishida K (1985) Distribution, abundance, and effects of Pennella sp (Copepoda: Pennellidae), parasitic on the saury, Cololabis saira (Brevoort), in the western North Pacific Ocean and adjacent seas, 1984 Bull Biogeogr Soc Jpn 40:35–42

19 Watanabe Y, Kosaka A, Tanno Y, Takahashi S (1985) rence of parasitic Pennella sp on the Pacific saury, Cololabis saira in 1983 Bull Tohoku Reg Fish Res Lab 47:37–46

Occur-20 Kosaka A, Watanabe Y, Tanno Y, Takahashi S (1985) Ecology of the Pacific saury Cololabis saira Brevoort determined by the distribution of its parasite Pennella sp Bull Tohoku Reg Fish Res Lab 47:79–81

21 Oozeki Y, Takasuka A, Okamura H, Kubota H, Kimura R (2009) Patchiness structure and mortality of Pacific saury Cololabis sa- ira larvae in the northwestern Pacific Fish Oceanogr 18:328–345

O R I G I N A L A R T I C L E Biology

Dynamics of free amino acids in the hemolymph of Pacific

whiteleg shrimp Litopenaeus vannamei exposed to different types

of stress

Junpei Shinji•Marcy N Wilder

Received: 30 January 2012 / Accepted: 31 July 2012 / Published online: 24 August 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract We analyzed the dynamics of amino acids,

ammonia-N, and carbohydrates in hemolymph when

Pacific whiteleg shrimp Litopenaeus vannamei was

sub-jected to air-exposure and low-salinity stresses Glycine,L

-arginine, and D- and L-alanine levels in hemolymph were

increased under both stress conditions Ammonia-N, a

product of amino acid catabolism, also increased in level

These results suggest that the above-mentioned amino

acids are used as energy sources Levels of total

carbohy-drates, the sum of glucose and other carbohycarbohy-drates, in the

hemolymph showed different dynamics between the two

types of stress and were not always high This suggests that

other energy sources, such as amino acids, are important

when animals are subjected to stress It is possible that

particular amino acids act as energy sources under various

stress conditions

Keywords Amino acids
Crustacea
Hypoxia

Osmoregulation
Stress

Introduction

Crustacean aquaculture is a multibillion-dollar industry

worldwide that continues to show significant growth [1] In

this way, various shrimp species thus constitute important

commercial species targeted in aquaculture However,adverse environmental factors often contribute negatively

to the sustainability of shrimp aquaculture [2] It is fore considered to be important to maintain low levels ofenvironmental stress during rearing in order to preventdisease outbreak and achieve good growth rates

there-One significant representative stress response is themobilization of energy Like other animals, decapod crus-taceans need to produce additional energy in order to dealwith stress [3,4] Both the rate of oxygen consumption andglucose levels increase in the hemolymph when decapodcrustaceans are subjected to stresses such as handling [5,

6], salinity [7, 8], disease vectors, and pollutants [9, 10].Studies on stress response in decapod crustaceans havemost often focused on carbohydrate metabolism, whilethere have been only a few studies on other energy sourcessuch as amino acids

Amino acids are thought to play a role as a source ofenergy under stressful conditions because of the observa-tion of increased excretion rates of ammonia-N, a product

of amino acid catabolism [11] Levels of particular aminoacids in the hepatopancreas, such as arginine and proline,decline in Penaeus monodon under ammonia stress, sug-gesting that they are consumed as energy sources [12] Ithas also been reported that at least alanine is utilized ingluconeogenesis in the hepatopancreas under hyposmoticconditions; the hepatopancreas may take up amino acidssupplied from other tissues via the hemolymph [13] Thus,several studies have indicated that amino acids can act asenergy sources in Crustacea under stressful conditions

In this study, we investigated the hypothesis that, likecarbohydrates, amino acids in the hemolymph are impor-tant in energy mobilization as part of the stress response.Litopenaeus vannamei was chosen as an experimentalanimal This species is a penaeid shrimp native to the

Department of Global Agricultural Sciences,

Graduate School of Agricultural and Life Sciences,

The University of Tokyo, Bunkyo-ku, Tokyo 113-0032, Japan

Fisheries Division, Japan International Research Center

for Agricultural Sciences, Tsukuba, Ibaraki 305-8686, Japan

e-mail: marwil@jircas.affrc.go.jp

DOI 10.1007/s12562-012-0542-0

western coastal regions of the Western Hemisphere and is

distributed from northern Peru to Mexico, where

temper-ature and salinity ranges of 15–28°C and 5–45 % have

been reported [14] Culture of this species has become a

rapidly growing industry [15], and information on its

physiological responses to environmental stress is therefore

required as a fundamental component of aquaculture In

order to characterize amino acids that possibly function as

a common energy source under different types of stress, we

analyzed the dynamics of amino acids and relevant

catabolites in the hemolymph when L vannamei was

subjected to severely stressful conditions We chose air

exposure and low salinity as the experimental conditions,

as these parameters have often been utilized in the analysis

of stress response [16] Although the exact means of

implementing the stressors investigated in this study do not

necessarily represent a problem in aquaculture, the

responses observed to these stressors are likely

represen-tative of those that might occur in commercial farming—

for example, low-salinity conditions caused by heavy

rainfall and air exposure during transfer Our results

sug-gest that particular amino acids are commonly utilized as

energy sources during stress

Materials and methods

Experimental animals

Male subadult L vannamei were purchased from

Interna-tional Mariculture Technology Co Ltd (Myoko, Niigata,

Japan) Experimental animals were stocked in a 3,000-L

tank at 20°C, initially at a salinity level of 30–35 parts per

thousand (ppt) The rearing water utilized in this study was

prepared with dechlorinated Tsukuba City tap water [Na?

24.5 mg/L; Cl- 30.5 mg/L; total hardness 73.6 mg/L;

NH4–N nondetectable; heavy metals (Cu, Cd, Hg, Se, Cr,

Zn, Fe, Pb, and As) nondetectable] [17] and artificial

sea-water salt (Sea Life; Marinetech, Tokyo, Japan) Moreover,

these levels are expected not to affect the overall results

because the major ions in the tap water utilized in this

study constitute osmolality lower than 5 mOsm

Experimental procedures

Intermolt or premolt males were used Carapace length was

26.4 ± 0.2 cm (mean ± standard error, SE), and body

weight was 13.3 ± 0.3 g (mean ± SE) Before each stress

treatment, animals were reared in 60-L tanks at 28°C

under a salinity of 28 ppt for 7 days They were fed 6 g of

dry food pellets (Gold Prawn; Higashimaru Co Ltd.,

Kagoshima, Japan) per day, except in the 24 h before the

experiments Remaining feed was removed from the tank

5 h after feeding, and water quality was maintained bycirculating the water through filters

After the acclimation period, experimental animals weresubjected to air-exposure or low-salinity stress Three to sixexperimental animals were used for each treatment In the air-exposure experiment, animals were placed on top of a styro-foam box and subjected to air-exposure stress for 0 (control),

15, or 30 min Humidity was kept at 75 % and air temperaturewas maintained at 28°C using an air conditioner In the low-salinity experiment, animals were subjected to salinities of 0and 28 (control) ppt for 3 or 6 h at 28°C

After each stress treatment, experimental animals wereanesthetized by being frozen on ice, and 300–900 lL ofhemolymph was sampled by syringe from the side of theabdomen between the cephalothorax and tail The sampledhemolymph was mixed with 20 lL of 1 M sodium citrate,and the concentration of sodium citrate was then adjusted

to 0.1 M by adding 1 M sodium citrate until the volume ofsodium citrate reached 10 % of total volume (therefore, afinal concentration of 0.1 M sodium citrate) This yielded atotal volume of 300–1,000 lL of hemolymph in 0.1 Msodium citrate The above-prepared samples were stored at–80 °C until analysis of free amino acids, ammonia-N, andcarbohydrate levels

Extraction methodsOne hundred microliters of hemolymph sample was heated

at 105 °C for 10 min Heated samples were cooled on icefor 10 min, then centrifuged at 4°C and 20,000g for

10 min The supernatants were then collected and used insubsequent analysis

Analysis of free amino acids by HPLCFree amino acids were analyzed by high-performanceliquid chromatography (HPLC) A Shimadzu LC-10 HPLCsystem was used for precolumn derivatization, according toNimura and Kinoshita [18] A Develosil ODS-UG-5(6.0 9 200 mm; Nomura Chemical, Aichi, Japan) wasused as an analytical column, and a Develosil ODS-UG-5(6.0 9 35 mm; Nomura Chemical) was used as a guardcolumn A Shim-pack GRD-ODS (4 9 35 mm; Shimadzu,Tokyo, Japan) was used as a precolumn to guard the ana-lytical column, according to the manufacturer’s protocol.Eighty milligrams of o-phthalaldehyde (OPTA) and

100 mg of N-acetyl-L-cysteine (AcCys) were dissolved in

10 mL of methanol and used as a derivatization reagent.The samples, together with 0.1 M sodium tetraborate andOPTA-AcCys reagent, were mixed in a ratio of 1:3:2 andinjected into the HPLC system after 4 min of reaction atroom temperature Each derivatized amino acid was sepa-rated at 28°C by using a binary system consisting of

50 mM of sodium acetate (A) and methanol (B) as the

mobile phases The following time program was used:

0–12 % B from 0 to 3 min, 12–18 % B from 3 to 7 min,

18–24 % B from 7 to 18 min, 24–40 % B from 18 to

39 min, 40 % B from 39 to 50 min, 40–52 % B from 50 to

52 min, 52–67 % B from 52 to 64 min, and 80 % B from

64 to 69 min The flow rate was linearly increased from 1.2

to 1.5 mL/min between 0 and 69 min Amino acid

deriv-atives were made to emit fluorescence at a wavelength of

350 nm and were detected at a wavelength of 450 nm All

of the asparagine and lysine was regarded asL-amino acids,

becauseD-asparagine andD-lysine have not been found in

the tissues of crustaceans [19]

Analysis of ammonia-N concentration

The concentration of ammonia-N, i.e., the sum of the NH3

and NH4?concentrations, in the hemolymph was analyzed

by the salicylate method using a commercial kit and 25-mL

samples (HACH, Colorado, USA) The analytical

proce-dure mainly followed the protocol supplied with the kit

Salicylate and cyanurate reagents were prepared by

sepa-rately dissolving each reagent in the kit in 6.25 mL of

Milli-Q water Twenty microliters of sample and 10 lL of

salicylate reagent were mixed and incubated for 3 min at

room temperature Thereafter, 10 lL of cyanurate reagent

was added and the mixture was incubated for 15 min at

room temperature The absorbance at a wavelength of

655 nm was measured by spectrophotometer (NanoDrop

ND-2000C; Thermo Fisher Scientific, MA, USA)

Analysis of carbohydrate levels

Total carbohydrate levels in the hemolymph were analyzed

by using the anthrone method Anthrone reagent was

pre-pared by dissolving 0.5 g of anthrone and 10 g of thiourea in

100 mL of 75 % sulfuric acid at 80–90°C Five microliters

of each sample and 50 lL of anthrone reagent were mixed

and allowed to react for 5 min at 105°C Subsequently, the

reacted samples were cooled on ice for 20 min The

absor-bance at 620 nm was measured with a spectrophotometer

(NanoDrop ND-2000C; Thermo Fisher Scientific), and total

carbohydrate levels were determined against the standard

curve of glucose Glucose levels in the hemolymph were

analyzed with a commercial glucose assay kit

(Sigma-Aldrich Inc., MO, USA) Levels of carbohydrates other than

glucose were calculated from the difference in levels

between total carbohydrates and glucose

Statistical analysis

Statistical methods were selected according to the

homo-geneity of variance examined by the Bartlett test

Differences between mean concentrations in the sure stress experiments were compared by using a para-metric Dunnett’s method or nonparametric Bonferroni-typemultiple comparisons Differences between means in thelow-salinity stress experiments were compared by using aparametric or nonparametric Bonferroni-type multiplecomparisons Linear regression analysis of the molar con-centrations of ammonia-N in hemolymph as a function ofthose of total free amino acid-N in hemolymph was per-formed Fitness to linear models, a separated model, and acombined model between the experimental conditions,were compared based on Akaike’s information criterion(AIC), which is often utilized as a criterion for selectionamong competing statistical models [20] The Pearsonproduct-moment correlation coefficient was calculatedfrom these analyses, and the correlation coefficient wasexamined by t test

air-expo-ResultsSurvival ratesThe survival rate of animals subjected to the air-exposurestress was 100 % Animals exposed to air for 15 minexercised vigorously, but those exposed for 30 min wereless active The survival rate of the animals subjected tolow-salinity stress was 95 % One animal died when it wassubjected to low salinity for 6 h

Amino acids and ammonia-N levelsLevels of L-aspartic acid, L-arginine, and L-alanine in thehemolymph increased significantly when the animals wereexposed to air for 15 min; levels of L-glutamic acid, gly-cine, D-alanine, L-isoleucine, and lysine increased withexposure for 30 min (Table1; n = 5 each; P \ 0.05) Theincreases in the levels of these amino acids for the mostpart were highest between 15 and 30 min Increases inglycine, L-isoleucine, and D-alanine at 30 min were par-ticularly notable compared with those of the control.Levels of various amino acids in the hemolymphincreased when animals were exposed to low-salinitystress; these included L-aspartic acid, L-glutamic acid,glycine, L-arginine, and D- and L-alanine (Table2;

n = 4–6; P \ 0.05) The increases were largest forDnine and glycine Glycine became the most abundantamino acid when the animals were exposed to each stress,whereas taurine was most abundant in the controls.Ammonia-N, a product of amino acid catabolism, alsoincreased in the hemolymph under both stress conditions.Under air-exposure stress, ammonia-N levels in thehemolymph increased significantly after 30 min of

-ala-Table 1 Changes in free amino

exposure (Fig.1a; n = 7 each; P \ 0.05) Under

low-salinity stress, ammonia-N levels in the hemolymph

increased at each exposure time and were highest after 6 h

of exposure at 0 ppt (Fig.1b; n = 3–6 each; P \ 0.05)

Ammonia-N levels in the hemolymph were highly

corre-lated with total amino acid-N levels in the hemolymph

(Fig.2) Based on AIC, a single linear model that combines

air exposure and low salinity was selected (separated

model: AIC = 21.11; combined model: AIC = 19.06)

Pearson’s product-moment correlation coefficient was

significant (R = 0.912; P \ 0.05)

Carbohydrate levelsTotal carbohydrate levels in the hemolymph showed dif-ferent dynamics between air-exposure and low-salinitystress conditions Total carbohydrate levels in the hemo-lymph of animals subjected to air-exposure stress simplyincreased after 15 or 30 min of stress (Fig 3a; n = 5 each;

P\ 0.05) On the other hand, compared with 0 h afterexposure to low salinity, total carbohydrate levels showed

an increase under low-salinity conditions (0 ppt) for 3 h,but they recovered to control levels when subjected to

0 ppt for 6 h (Fig.4a; n = 3–6 each; P \ 0.05) Glucoselevels in the hemolymph showed similar dynamics to totalcarbohydrate levels in the hemolymph (Figs.3b, 4b).Glucose accounted for most of the total carbohydrates (up

to 80 %) The levels of carbohydrates other than glucoseshowed significant increases when animals were subjected

to air-exposure stress for 30 min (Fig 3c; n = 5 each;

P\ 0.05) The levels of other carbohydrates under salinity conditions showed a pattern similar to those of totalcarbohydrate, although changes were not significant(Fig.4c; n = 3–6 each; P \ 0.05)

low-DiscussionAir-exposure stress utilized in this study includes anoxicstress, dehydration stress, and handling stress Low-salinity

c

Exposure time (min)

exper-imental animals were subjected to air-exposure stress (a; n = 7 each)

or low-salinity stress (b; n = 3–6) Vertical lines indicate SE.

Asterisks indicate significant difference compared with control

(P \ 0.05)

Total free amino acid-N in hemolymph (mmol/L)

y= 0.063x+ 0.531

R= 0.912

0.0 1.0 2.0 3.0

Low salinity Air exposure

levels in hemolymph Units are shown in terms of molar tion of nitrogen, as the molar quantity of ammonia produced in the process of decomposing one amino acid molecule depends on the molar quantity of nitrogen within the molecular structure of the amino acid

concentra-stress subjects animals to the osmosis of unnecessary water

and loss of inorganic ions These stress conditions were

therefore completely different Nevertheless, the levels of

particular free amino acids, such as L-glutamic acid, cine, L-arginine, and D- and L-alanine, increased in bothstress conditions (Tables1,2)

carbohydrates other than glucose (c) in hemolymph when

experi-mental animals were subjected to air-exposure stress Vertical lines

indicate SE (n = 5 each) Asterisks indicate significant difference

compared with control (P \ 0.05)

0 40 80 120 160

10 20 30 40

b

a

a

a ab

carbohydrates other than glucose (c) in hemolymph when mental animals were subjected to low-salinity stress Vertical lines indicate SE (n = 3–6) Different letters indicate significant differ- ences (P \ 0.05)