Fish Sci (2012) 78:1153–1162
DOI 10.1007/s12562-012-0557-6

ORIGINAL ARTICLE

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 otoliths, the von Bertalanffy growth curves in male and
females were determined as FLt ¼ 1049½1 À expfÀ0:835
ðt þ 6:975  10À14 ÞgŠ and FLt ¼ 938½1 À expfÀ1:029ðtþ
6:975  10À14 ÞgŠ, respectively, where FLt is 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

S. Furukawa (&) Á S. Tomoe

Graduate School of Science and Technology, Nagasaki
University, Bunkyo-machi, Nagasaki 852-8521, Japan
e-mail:
S. Ohshimo Á T. Shiraishi
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-minamimachi, Okayama 702-8024, Japan
N. Nakatsuka Á R. Kawabe
Graduate School of Fisheries Science and Environmental
Studies, Nagasaki University, Taira-machi, Nagasaki 851-2213,
Japan

index and histological observations. Therefore, based on
the relationship between the fork length and the developmental stage of the testes or ovaries, male and female
dolphinfish were found to reach sexual maturity by the
following spawning season after hatching in the northern
East China Sea.
Keywords Dolphinfish Á Growth Á Scale Á Otolith Á
Reproduction Á Gonad histology

Introduction
Dolphinfish Coryphaena hippurus is a highly migratory
oceanic pelagic fish found worldwide in tropical, subtropical, and temperate waters [1]. In East Asia, dolphinfish

support economically important recreational and commercial fisheries, and are a shared resource among multiple
countries, such as Taiwan and Japan [2, 3]. Dolphinfish
feed on several important commercial fishery species of the
East China Sea, including anchovy Engraulis japonicas,
flying fish (Exocoetidae), and other small pelagic prey,
including squid [4].
The removal of predator biomass during commercial
fishing can have profound effects on pelagic ecosystems
because of the removal of predation pressure and topdown, 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, how
these dolphinfish traits will change in the future remains to
be determined. Therefore, it is necessary to clarify the
current biological characteristics of dolphinfish so that we
may understand how they will change with time and how
we should manage this species.

123


1154

Fish Sci (2012) 78:1153–1162

Some studies of the biological characteristics of dolphinfish 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 frequency 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 reproductive 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.

120 E

130 E

140 E

150 E

50 N

Sea of Japan

40 N

30 N


Pacific Ocean

East
China
Sea

Sea of Japan

Taiwan

36 N

Tsushima
Islands

34 N

Goto
Islands

126 E

128 E

32 N

130 E

132 E


Fig. 1 Sampling areas for dolphinfish in the waters off west Kyushu
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 specimens 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 occasionally 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 calculated as follows:
GSI ¼

GW
 100:
ðBW À GWÞ


123

Because dolphinfish body weight and length are correlated with GSI values and dolphinfish size varied significantly with month, GSIs were analyzed by a generalized
linear model (GLM) using a gamma distribution with a log
link function to test for a month effect with fork length
(FL) as the covariate for males and females separately. For
histological observations, small pieces of the gonad were
fixed in 10 % formalin.
Age determination
For age determination, we used sagittal otoliths and scales
obtained from small-sized specimens and large-sized
specimens, respectively. The deposition of increments in
dolphinfish otoliths begins on the hatching date, and rings
are laid down daily [12, 19]. Thus, no adjustment was
required to estimate age from incremental counts of sagittae, and it was assumed that rings were formed daily.
Previous studies on the microstructure of sagittal otoliths of
dolphinfish from the western Mediterranean Sea had found
that the daily ages of larger dolphinfish ([650 mm FL)
appeared to be underestimated [12]. Furthermore, in this
study, daily rings of sagittal otoliths were unclear in large
dolphinfish (C412 mm FL). Therefore, our daily ring
determination was restricted to small dolphinfish. To
determine the ages of small dolphinfish in days (herein
‘‘daily ages’’), otoliths were removed under a dissecting
microscope and embedded in resin on a glass slide. The


Fish Sci (2012) 78:1153–1162

1155


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. However, only small fishes were used for age determination in
this study, and we did not section and thin polish the otoliths to determine the daily ages of the fish.
Annual marks are not detectable on otoliths of dolphinfish [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 validation based on marginal increment analysis was used. The
marginal increment (MI) was determined using the following equation:
MI ¼

ðR À rn Þ
;
ðrn À rnÀ1 Þ

and the growth coefficient for sex s, respectively. The
hypothetical age corresponding to a fork length of zero is
t0. We defined FL?,s and ks in the following ways:
8

FL1;s ¼ L1 þ L2 Â S; Ks ¼ K1 þ K2 Â s ðcase 1Þ
>
>
>
<
ðcase 2Þ
FL1;s ¼ L1 þ L2 Â S; Ks ¼ K1
;
>
ðcase 3Þ
FL1;s ¼ L1 ; Ks ¼ K1 þ K2 Â s
>
>
:
ðcase 4Þ
FL1;s ¼ L1 ; Ks ¼ K1
where L1, L2, K1, and K2 are unknown constants and s is a
binary parameter (if male, s = 0; if female, s = 1). We
assumed that t0 is not influenced by sex, because larvae and
juveniles could not be sexed, and we used FL and age data
for larvae and juveniles of both sexes. The variance of the
fork length at age t and for sex s is given by
Vðt; sÞ ¼ r20 þ

r2
f1 À expðÀ2Ks tÞg;
2Ks

ð2Þ


where r20 and r2 represent the variance of the fork length at
age 0 and the intensity of white noise, respectively [21].
The log likelihood can be represented by the following
equation:
"
#
N
1X
fFLi À lðti ; si Þg2
ln L ¼ À
lnf2pVðti ; si Þg þ
; ð3Þ
2 i¼1
Vðti ; si Þ

where R is the overall radius from the focus to the outer
edge of the scale. Rn is 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.

where ln L is the log likelihood and i ¼ 1; 2; . . .; N. FL?,s,
ks ; t0 ; r0 and r were estimated by maximizing Eq. 3 using
the gosolnp function in the Rsolnp package of R. To select
the best-fit model from among cases 1–4, we used the
Akaike information criterion (AIC). The model which
yielded the minimum AIC was selected as the best model.

Estimation of the von Bertalanffy growth parameters


Influence of water temperature on growth

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 dolphinfish. Therefore, the birth date of every large individual
was assumed to be 1 July, which approximately corresponded 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: 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 represented by the following von Bertalanffy growth equation:

Asymptotic fork lengths, estimated from von Bertalanffy
growth function fits using size and age data collected in
different regions of the world [10, 11], were used as a regionspecific growth index. We did not use the maximum size as
an index for growth in order to avoid local sampling bias. To
examine the influence of water temperature and sex on
asymptotic fork length estimated from the previous studies
and the present study, a generalized linear mixed model
(GLMM) assuming a gamma distribution and a log-link
function were used. Differences in catch region were defined
as a random effect, since our objective was not to test for
unknown regional effects. The gamma GLMM was conducted using the GLMM function in the ‘‘repeated’’ package
of R. As an index for water temperature, we used satellitederived sea surface temperatures [SST, 11 km resolution
advanced very high resolution radiometer (AVHRR):
Pathfinder V5] obtained from the Ocean Watch webpage
( and


lðt; sÞ ¼ FL1;s ½1 À expfÀKs ðt À t0 ÞgŠ;

ð1Þ

where l(t, s) represents the mean fork length at age t and
for sex s. FL?,s and ks represent the asymptotic fork length

123


1156

Fish Sci (2012) 78:1153–1162

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.

3.

4.

Histological observations
The fixed gonads were dehydrated and embedded in paraffin, 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 microscope 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 characteristics (Figs. 2, 3).
The five testis stages were as follows:
1.

2.

5.

The six stages of oocytes were as follows:
1.

2.
3.

Spermatogonial proliferation stage (Sp; Fig. 2a): only
spermatogonia (sg) are abundant in the seminal lobule.

a

Early spermatogenesis stage (Es; Fig. 2b): spermatogonia (sg) and spermatids (st) are organized in the
seminal lobules.
Late spermatogenesis stage (Ls; Fig. 2c): spermatogenesis proceeds in the testis. Spermatids (st) of the
seminal lobules increase, and spermatozoa (sz) are
found in the lumina of the seminal lobules.
Functional maturation stage (Fm; Fig. 2d): spermatozoa (sz) are abundant in the lumina of the seminal
lobules and main sperm duct. Spermatogonial division

and further spermatogenesis proceeds in the seminal
lobules.
Postspawning stage (Ps; Fig. 2e): spermatogonia (sg)
are found in the seminal lobules, although spermatozoa
(sz) occur in the lumina of the seminal lobules.

Immature stage (Im; Fig. 3a): only previtellogenic (pn)
oocytes are present, including those in the perinucleolus and yolk vesicle stages.
Developing stage (D; Fig. 3b): the most advanced
oocytes are at the early yolk (ey) or mid-yolk (my) stages.
Vitellogenic stage (Vi; Fig. 3c): the most advanced
oocytes are at the late yolk (ly) stage, which marks the
end of vitellogenesis.

b

c

st
sz

sg
80µm

st

sg

d


e

sz

Fig. 2 Photomicrographs of testes at different developmental stages
in dolphinfish. a Spermatogonia proliferation stage, b early spermatogenesis stage, c late spermatogenesis stage, d functional

123

maturation stage, and e postspawning stage. sg spermatogonial,
st spermatid, sz spermatozoon


Fish Sci (2012) 78:1153–1162

1157

Fig. 3 Photomicrographs of ovaries at different developmental stages
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

4.

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

4 and 53 days, respectively. Scales were collected from
136 large-sized dolphinfish, and the rate of agreement
between readers of the number of annual ring marks was
64.2 % (88 of the 137 specimens in total). A total of 69
scales were classified as age 1 or older, and the remaining
scales (n = 19) were estimated to be age 0. The estimated
maximum ages for males and females were five years old.
MIs from [age 0 dolphinfish (n = 69) were greatest in
October, November, and December, dropped in January,
and stayed low during the winter months and the spawning
season (see subsequent results) (Fig. 4). There was a significant difference in marginal increment width per month
(gamma GLM, p \ 0.05).
The von Bertalanffy growth parameters were estimated
for cases 1–4 (Table 1). When we used case 1, the minimum AIC was obtained, and the DAIC value of the next
most parsimonious case (case 2) was more than 2 [22]. The
relationship between age and FL of the dolphinfish is
shown in Fig. 5a, where most of the data are within the
95 % prediction interval for both sexes (Fig. 5b, c),

5.

6.

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.
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.
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 determination. Unfortunately, we could not collect the fish
between 237 and 412 mm because this size range of dolphinfish does not support economically important commercial fisheries in this study area. However, we obtained a
sufficient wide size range to describe the general growth

123


1158

Fish Sci (2012) 78:1153–1162

3.2

8

2.4

MI

8
1.6


11

0.8

7

5

15

3

7

4

Fork length (mm)

1200

1
0

900

600

Male
Female

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

300

0

0
Dec

Nov

Oct

Sep

Aug

Jul

Jun

May

Apr

Mar


Feb

Jan

0
0

1

Month

2

3

4

5

Age

Fig. 4 Box plot of the marginal increment width (mm) for dolphinfish (sampled from May 2008 through May 2010) pooled by month
(January–December). Sample sizes are given above the box for each
month

Fig. 5 Relationship between age and fork length for male (black) and
female (gray) dolphinfish. The mean growth curves and 95 %
prediction intervals are indicated by solid lines and dotted lines,
respectively


Table 1 Estimated parameters and comparison of AIC scores for the four von Bertalanffy growth curve cases
Sex

Parameter

AIC

DAIC

FL?,s

ks

t0

1048.7

0.84

-4.7 9 10-14

-3007.0

0

937.6

1.03


1010.7

0.93

-2.5 9 10-12

-3018.6

11.5

995.8

0.92

-4.2 9 10-10

-3019.7

1.1

994.4

0.90

-5.2 9 10-11

-3020.9

1.2


Case 1
Male
Female
Case 2
Male
Female

979.0

Case 4
Male
Female
Case 3
Male
Female

0.94

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 estimated in males and females as follows:

19.7 to 27.4 °C. There was a significant positive relationship between FL1 and SST (p \ 0.05). The best-fit model
was
FL1 ¼ expð5:87 þ 0:053 Â SSTÞ:

Â

È
ÉÃ

lðt; maleÞ ¼ 1049 1 À exp À0:835ðt þ 6:975 Â 10À14 Þ
lðt; femaleÞ ¼ 938½1 À expfÀ1:029ðt þ 6:975  10À14 ÞgŠ:

Annual reproductive cycle

Comparison of the AIC scores (Table 2) revealed that
the model that included a term for SST but not sex provided 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

A total of 329 large dolphinfish (FL 412–1124 mm, 112
male and 217 female) were used for reproductive characterization (137 of 329 specimens were used for the aging
study). Length-adjusted mean gonad weights varied significantly with month (gamma GLM, p \ 0.001) for both
male and female dolphinfish (Fig. 7). The mean value of

123


Fish Sci (2012) 78:1153–1162

1159

Table 2 Comparison of the AIC scores for alternative models to
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

131.9

1.8


137.9

6.0

Sex

139.2

1.3

7

5

35

4

6
25

2

3
0
Jan

SST, sex, SST 9 sex
Null model


0.6

Dec

0.1

Nov

130.1

Oct

SST, sex

1

Sep

1.3

Jul

130.0

3
1.2

Aug


SST 9 sex

12

Jun

0

May

128.7

Apr

SST

Mar

DAIC

9
1.8

Feb

AIC

a

Male GSI


Fixed effect

2.4

Month
1800

Florida

20

North
Carolina

1400

1200

41

28

10

6

2

11


Feb

4

Jan

1000

8
42

5

Mediterranean

800

5

15

Female GSI

29 26

15

20


Dec

Nov

Oct

Sep

Aug

Jul

Jun

May

Present
study

Apr

0

Male
Female

Mar

Asymptotic fork length (mm)


b
Puerto
Rico

1600

Month
22

24

26

28

SST (ºC)
Fig. 6 Relationship between SST and asymptotic fork length, as
reviewed by Oxenford [11]. The models are indicated by solid lines,
the gray and light gray zones indicate the 50 and 95 % prediction
intervals, 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 dolphinfish 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.

Fig. 7 Box plots of gonadosomatic index (GSI) for a male and
b female dolphinfish (sampled from May 2008 through May 2010)
pooled by month (January–December). Sample sizes are given above
the box for each month

The proportion of Sp-stage females was highest in July.
Females with ovaries at the Re stage were found from
August and October.
Spawning size and GSI
Sexually mature males were defined as individuals with
testes at the Fm stage. Sexually mature females were
defined as individuals with ovaries with Vi, M, or Sp stage
oocytes. GSI values for immature stages (Sp to Ls) and the
mature stage (Fm) overlapped, and these stages ranged
from 0.5 to 0.9 (Fig. 9a). Individuals with testes at the Fm
stage were also larger than 524 mm FL (Fig. 9b). GSI
values of female individuals were less than 0.8 in the
immature (Im) stage, and the values for the oocytes ranged
from 0.2 to 4.0 in the developing (D) stage (Fig. 9c). GSI
values for females in the Vi, M, and Sp stages ranged from
3.3 to 11.5. The minimum fork length of females in the
mature (from Vi to Sp) stages was 514 mm FL (Fig. 9d).


123


1160

a

Fish Sci (2012) 78:1153–1162
n=

3

6

5

0

2

2

11

1

2

39


9

5

100
Ps
Fm
Ls
Es
Sp

Frequency (%)

80
60
40

Dec

Oct

Nov

Sep

Jul

Aug


Jun

May

Apr

Feb

Mar

0

Jan

20

1

9

Month

b

n=
100

2

11


4

0

3

33

4

8

15 44

Re
Sp
M
Vi
D
Im

Frequency (%)

80
60
40
20

Dec


Oct

Nov

Sep

Aug

Jul

Jun

May

Apr

Mar

Feb

Jan

0

Month
Fig. 8 Monthly variations in the frequency of occurrence of various
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

Discussion
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

123

the maximum sizes of both sexes in this study were smaller
than 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 dolphinfish near the East China Sea. Therefore, we suggest
that the growth parameters estimated in this study are more
useful for examining and comparing growth among other
regions.

The asymptotic fork length of dolphinfish in the northern East China Sea shows a greater similarity to the
asymptotic fork length of western Mediterranean Sea dolphinfish [12] than to the asymptotic fork lengths of dolphinfish in other regions (Fig. 6). However, to the best of
our knowledge, the first-year growth for dolphinfish from
the northern East China Sea is the smallest in the world. It
is well known that differences in estimated growth between
regions [5, 6] can be related to environmental conditions
(i.e., water temperature, food availability, exploitation
levels). Moreover, temperature appears to be the most
important environmental factor affecting growth in fish.
Because of the importance of temperature as a controlling
factor [23], the physiological literature is replete with
examples of studies evaluating thermal effects on fish [24–
26]. Although differences in the growth of dolphinfish
among regions have been found [10, 11], few measurements of the effect of temperature on the growth of dolphinfish have been performed. Our study is unique in that it
shows that there were clear distinctions in asymptotic fork
length with respect to water temperature (Fig. 6). However,
additional bioenergetic data are required to parameterize
models that attempt to depict patterns of growth observed
in dolphinfish. Of particular importance are data describing
the effects of water temperature, body size, and feeding on
metabolism.
The asymptotic fork length of dolphinfish was significantly larger in males than in females in this study, which
reflects results from Florida and North Carolina (Fig. 6).
However, the asymptotic fork length of dolphinfish was
larger in females than in males from the Mediterranean and
Puerto Rico (Fig. 6). Differences in growth features due to
different laboratory methods can not be excluded. For
example, age determination of dolphinfish based solely on
otoliths was found to underestimate the ages of older,
larger fish [12]. Obviously, one of the main ways to

identify the factors responsible for this inter-region variability in growth would be to standardize age and growth
methods.
Reproduction
There are no reports regarding dolphinfish maturation in
the East China Sea that utilized histological techniques.
Kojima [4] estimated the spawning period by examining


Fm
Ls
Es
Sp

a
0.4

0.7

1

Testis maturation stage

Ps

0

Ps
Fm
Ls
Es

Sp

1.4

b
400

Re
Sp
M
Vi
D
Im

c
0

2

4

6

8

GSI

seasonal changes in oocyte diameter obtained from dolphinfish in the southwestern Sea of Japan, which is adjacent 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 reproductive 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 geographic 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 different 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

600

800

1000

1200

Fork length (cm)


10

12

Oocyte maturation stage

GSI

Oocyte maturation stage

Fig. 9 Relationships between
a the five maturation stages of
testes and gonadosomatic index
(GSI), b the five maturation
stages of testes and fork length,
c the six maturation stages of
ovaries and GSI, and d the six
maturation stages of ovaries and
fork length of dolphinfish. Refer
to Figs. 2 and 3 for each testis
and ovarian stage, respectively.
Crosses in c and d indicate that
postovulatory follicles were
observed

1161

Testis maturation stage


Fish Sci (2012) 78:1153–1162

Re
Sp
M
Vi
D
Im

d
400

600

800

1000

Fork length (cm)

Pacific (L50 = 483.8 and 505.7 mm for females and males,
respectively) [17], Taiwanese waters (L50 = 510 mm for
both sexes) [16], and off the coast of North Carolina
(L50 = 458 and 476 mm for females and males, respectively) [10] agree with values for fish that are less than
one year old, regardless of sex. We were unable to estimate
L50 for dolphinfish in both sexes from the East China Sea
because of a lack of reproductive characterization of smallsized individuals during the spawning season. Nevertheless, we determined that the smallest individuals with
matured testis and oocytes were 524 and 514 mm,
respectively, with an estimated age of less than one year in
both sexes. Hence, dolphinfish reach sexual maturity in

their first year of life in the East China Sea, which is similar
to other regions [10, 16, 27]. Clearly, it is necessary to
monitor variations in reproductive characteristics in future
studies, and to further determine the growth and maturity
processes of fish that are yet to reach one year of age, in
order to elucidate size at sexual maturation in the northern
East China Sea.
Acknowledgments We thank Dr. H. Tanaka at the Seikai National
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.

123


1162

References
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 Government 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 fisheries. 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 microstructure, age, and growth patterns of dolphin, Coryphaena hippurus, 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 Mediterranean 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

123

Fish Sci (2012) 78:1153–1162
16. Wu CC, Cheng SW, Kawasaki T (2001) Reproductive biology of
the dolphin fish Coryphaena hippurus on the east coast of Taiwan. 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 dolphinfish (Coryphaena hippurus) captured in the Gulf of Tehuantepec, 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 age-composition from length–frequency by using stochastic growth
equation. Nippon Suisan Gakkaishi 56:1209–1218
22. Burnham KP, Anderson DR (1998) Model selection and inference: a practical information-theoretic approach. Springer, New
York
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 California. 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 homogeneity 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, LluchCota 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


Fish Sci (2012) 78:1163–1171
DOI 10.1007/s12562-012-0554-9

ORIGINAL ARTICLE

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.

K. Hasegawa (&) Á K. Ohkuma
Hokkaido National Fisheries Research Institute,
Fisheries Research Agency, Nakanoshima, Toyohira,
Sapporo, Hokkaido 062-0922, Japan
e-mail:
K. Ohkuma
e-mail:
C. Yamazaki
Graduate School of Environmental Science, Hokkaido
University, N10 W5, Kita, Sapporo, Hokkaido 060-0810, Japan
e-mail:

T. Ohta
Tomakomai Experimental Forest, Hokkaido University,
Takaoka, Tomakomai, Hokkaido 053-0035, Japan
e-mail:

Keywords Competition Á Gammaridae Á
Terrestrial invertebrates Á Predation

Introduction
Because of their importance for commercial and recreational fishing, salmonids have been introduced into many
regions outside their natural range [1]. Their impacts on
native species and communities through interspecific
interactions have been well studied (e.g., [2, 3]). In particular, brown trout Salmo trutta and rainbow trout Oncorhynchus mykiss have been introduced into many regions
throughout the world and are listed in the list of 100 of the
world’s worst invasive alien species, published by the
International Union for the Conservation of Nature and
Natural Resources [4].
Introduced brown trout have become a significant
problem in streams in Hokkaido, northern Japan. These
streams contain masu salmon Oncorhynchus masou, an
endemic salmonid in far-eastern Asia and an important
fishery resource. In general, there is worldwide concern
that brown trout are detrimental to native salmonid populations because of predation by large individuals (particularly those larger than 300 mm) and competition [4]. For
example, Takami et al. [5] reported that brown trout
replaced native white-spotted charr Salvelinus leucomaenis
in a mountainous stream in Hokkaido, and Hasegawa and
Maekawa [6] suggested that interspecific competition is the
primary mechanism driving replacement. However, there
have been few direct studies of interspecific interactions
between brown trout and masu salmon. These include

limited case studies of predation (e.g., [7, 8]). Hasegawa
et al. [9] clarified that a difference in species-specific
ontogenetic habitat-shift patterns at the fry stage reduced

123


1164

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 interactions (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 preferences 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 varies 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 availability 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 tributary 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

123

Fish Sci (2012) 78:1163–1171

N


Sea of
Japan
Hokkaido
42° N

Chitose River

141° E
Lake Shikotsu

1
Chitose River

4 3

2

5
Mamachi Stream
6
7
8
9
10
11

G-P area

G-A area


12

1 km

Fig. 1 Locations of the study sites. G-P and G-A indicate ‘‘Gammaridae-present’’ and ‘‘Gammaridae-absent’’ areas, respectively

level, and the water temperature ranged from 4 to 12 °C
during the survey.
Brown trout were first found in Mamachi stream in the
mid-1980s, and are now distributed throughout the stream
([17]; Saneyoshi, personal communication). Native streamdwelling salmonids include masu salmon and white-spotted
charr. However, white-spotted charr was replaced by
brown trout in the 1990s (Saneyoshi, personal communication), most likely as a result of hybridization [19] and
interspecific competition [6]. In addition, native sculpin
Cottus nozawae and stone loach Noemacheilus barbatulus
as well as the introduced rainbow trout were present in the
stream. With the exception of stone loaches at sites 1–3, we
observed very low numbers of these species.
Population estimation, diet analysis, and food
availability
We estimated the population size and collected stomach
content samples every two months between the end of May
2009 and the end of March 2010. However, we were
unable to access sites 9–12 in January because of heavy
snow. To estimate the population size at each site, we
performed three removal passes during each sampling period
using a model 12 backpack electrofisher (Smith-Root Inc.,



Fish Sci (2012) 78:1163–1171

Data analysis
The overlap in diet composition between masu salmon and
brown trout was quantified using a proportional similarity
index [21, 22]:
m
X
PS ¼ 1 À 0:5
jMSi À BTi
i¼1

where MSi and 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 (no
overlap) 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 the
relationship between PS and stomach fullness (wet mass of
stomach contents/body weight) in masu salmon and brown
trout for each month to determine the effect of competition
for food resources on foraging efficiency in both species. A
two-way ANOVA followed by Scheffe’s test was used to
compare 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.


Results
Fish density and food availability
The abundance (fry and parr in total) of masu salmon was
generally higher than that of brown trout throughout the
year in Mamachi stream (Fig. 2).
We observed the typical spatial differences in benthic
invertebrate composition corresponding to the presence of
Gammaridae throughout the year at sites 1–4 (Fig. 3).
Hereafter, these sites are referred to as the ‘‘Gammaridaepresent’’ (G-P) area, whereas the remaining sites (5–12) are
referred to as the ‘‘Gammaridae-absent’’ (G-A) area
(Fig. 1). Trichoptera and Ephemeroptera were widely distributed in the stream (Fig. 3). Although the abundance of
each category fluctuated dramatically from month to
month, the presence/absence pattern of each category at
each site was almost same during the survey (Fig. 3).
The dynamics of the abundances of drifting terrestrial
invertebrates were similar for the G-P and G-A areas
0.9

0.7

density (ind./m2)

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 species) 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 terrestrial 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 quadrat). 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 invertebrates (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.


1165

0.5

0.3

0.1

-0.1

May

Jul.

Sep.

Nov.

Jan.

Mar.

Fig. 2 Seasonal changes in mean (±SD) total fish density (fry and
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

123



1166

Fish Sci (2012) 78:1163–1171

6

Gammaridae

Trichoptera

Ephemeroptera

Plecoptera

Diptera

Coleoptera
3

May

November
2

4

1

2


0

0
1

2

3

4

5

6

7

8

9

10

11

1

12


3

2

3

4

5

6

7

8

9

10

11

12

3

4

5


6

7

8

9

10

11

12

3

4

5

6

7

8

9

10


11

12

3

July

January

2

g/m2

g/m2

Others

2

1

1

0

0
1

2


3

4

5

6

7

8

9

10

11

1

12

3

2

3

September


March

2

2

1

1

0

0
1

2

3

4

5

6

7

8


9

10

11

12

Study sites

1

2

Study sites

Fig. 3 Abundance of benthic invertebrates at each site during each study period. Sites 1–4 represent areas with Gammaridae and 5–12 represent
sites without Gammaridae. The y-axis scale used for May is different from that used for the other months

(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 terrestrial 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 (Table 1; Fig. 5).
A two-way repeated-measures ANOVA revealed that
months and months 9 presence of Gammaridae (interaction) had significant effects (Table 2; Fig. 6). During May

123

and July, when terrestrial invertebrates were abundant
(Fig. 4), PS was smaller in G-P than in G-A (Fig. 6). The
Gammaridae comprised a higher proportion of the diet of
brown trout than the diet of masu salmon, and vice versa
for the proportion of terrestrial invertebrates, during May
and July in G-P (Fig. 7). Conversely, both species shared a
similar diet in the G-A reaches. Trichoptera, Ephemeroptera, and terrestrial invertebrates dominated their diets in
May, whereas terrestrial invertebrates were dominant in
July (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 corresponding difference during May and July (Fig. 6). In G-P,
both species preyed primarily on Gammaridae, though
masu salmon also preyed upon terrestrial invertebrates in
September (Fig. 7). In G-A, both species preyed primarily
on terrestrial invertebrates in September, and Trichoptera
and Ephemeroptera in January and March (Fig. 7).


Fish Sci (2012) 78:1163–1171


1167

0.025

(a)

a

a

a

a

a,b

b

g/m3

0.015

-0.005

May

Jul.

Sep.


Nov.

Jan.

Mar.

fork length (mm)

0.005

(b)

Fig. 4 Mean (±SD) abundances of terrestrial invertebrates in areas
with (G-P, circles with solid line) and without (G-A, triangles with a
dashed line) Gammaridae during each study period

a

a

b

b

b

b

Table 1 Two-way ANOVA results for the effects of area and month

on the fork lengths of brown trout and mosu salmon
df

F

p

Brown trout
Area (G-P or G-A)

1

Month

5

Area 9 month

5

8.66
54.4
1.26

\0.001
\0.001
0.281

Masu salmon
Area (G-P or G-A)


1

15.1

\0.001

Month
Area 9 month

5
5

11.1
9

\0.001
\0.001

Fig. 5 Mean (±SD) fork lengths of a brown trout and b masu salmon
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,
respectively (see Fig. 7)

Table 2 Results of repeated-measures ANOVA

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).

Effect

df

F

P

Month

5

5.61

0.001

Month 9 (G-P or G-A)

5


3.36

0.016

G-P or G-A

1

2.22

0.187

Brown trout predation on masu salmon

Although the body sizes of brown trout and masu salmon
were different for the G-P or G-A areas, and among
months, the difference did not appear to be large enough to
produce a significant difference in the food habits of
individuals. Thus, we assumed that the food habit differences shown in this study were caused by differences in
food conditions.
The degree of dietary overlap (PS) differed among areas
and seasons. Gammaridae and terrestrial invertebrates play

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 throughout
the year (Fig. 7).

Discussion

123


1168

Fish Sci (2012) 78:1163–1171

contents of brown trout was higher than the proportion of
Gammaridae in the stomach contents of masu salmon at the
sites where Gammaridae were present. Masu salmon, in
turn, had a higher percentage of terrestrial invertebrates in
their diet than did brown trout. This is likely important, as
the dietary overlap was smaller in the Gammaridae-present
area than in the Gammaridae-absent area. The smaller
dietary overlap in the Gammaridae-present area may not be
due to interspecific competition. With regards to speciesspecific habitat preferences, brown trout typically occupy
positions close to the substrate, whereas masu salmon
occupy the surface or midrange of the water column [11].
Inoue et al. [13] demonstrated that species occupying the
bottom range prey on Gammaridae, whereas species
occupying other ranges tend to prey on terrestrial invertebrates, thus avoiding interspecific competition. Therefore,
the presence of Gammaridae may mitigate competition for
food resources between masu salmon and brown trout.
Microhabitat use in the Gammaridae-present and Gammaridae-absent areas must be evaluated to confirm this

idea.

Fig. 6 Mean (±SD) proportional similarity index (PS) in areas with
(G-P, circles with a solid line) and without (G-A, triangles with a
dashed line) Gammaridae

a key role in determining the degree of dietary overlap. In
May and July, when terrestrial invertebrates were abundant, the proportion of Gammaridae in the stomach
Fig. 7 Mean proportions of
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)

Fish

Gammaridae

Plecoptera

Diptera

Terrestrials


Others

Masu salmon

Brown trout
100%

13 14 11 19 20 20 20 15 14 20 14 3

100%

May

40%

40%
20%

0%

0%
1

2

3

4

5


6

7

8

4

7

6 17 1 20 20 16 20 21 14 6

1

9 10 11 12

100%

80%

2

3

4

5

6


7

8

9 10 11 12

19 20 8 20 3 17 19 15 10 15 15 15

80%

60%

60%

Jul.

40%
20%

40%
20%

0%

0%
1

100%


5

60%

60%

100%

20 20 16 19 11 7 10 9 10 10 7

80%

80%

20%

2

3

7 12 6

4

5

6

7


8

9 10 11 12

5 11 20 15 20 21 19 9

1

9

100%

80%

2

3

4

5

6

7

8

9 10 11 12


20 21 21 20 20 20 21 13 13 11 22 13

80%

60%

60%

Sep.

40%
20%

40%
20%

0%

0%
1

2

3

4

5

6


7

8

9 10 11 12

Study sites

123

Ephemeroptera

Trichoptera

1

2

3

4

5

6

7

8


9 10 11 12


Fish Sci (2012) 78:1163–1171

1169

Fig. 7 continued

Fish

Gammaridae

Plecoptera

Ephemeroptera

Trichoptera

Diptera

Terrestrials

Others

Brown trout
100%

13 17 4


Masu salmon

4 14 17 14 19 19 15 20 15

100%
80%

80%

60%

60%

Nov.

40%

0%

0%
1

2

3

4

5


6

7

8

20 8

7

7 11 13 11 17

9 10 11 12

1

100%

80%

2

3

4

5

6


7

8

9 10 11 12

20 21 23 20 20 13 12 16

80%

60%

60%

Jan.

40%
20%

40%
20%

0%

0%
1

100%


40%
20%

20%

100%

22 25 21 19 20 18 20 19 20 19 20 20

2

3

23 18 6

4

5

6

7

8

9 10 11 12

1

7 18 20 21 21 20 20 20 15


100%

80%

2

3

4

5

6

7

8

9 10 11 12

18 23 19 23 20 21 21 9 20 20 20 20

80%

60%

Mar.

60%


40%

40%

20%

20%

0%

0%
1

2

3

4

5

6

7

8

9 10 11 12


1

2

3

4

5

6

7

8

9 10 11 12

Study sites
10

frequency

8
6
4
2
0
50


100

150

200

250

300

350

400

450

500

550

mm
Fig. 8 Frequency distribution of the body sizes (fork length, mm) of
brown trout individuals that preyed on masu salmon

Between the summer and winter seasons, we observed
an increase in the degree of dietary overlap in the Gammaridae-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 corresponding to Trichoptera explained the decline in the dietary
overlap in the Gammaridae-absent area and the large difference in the overlaps for the Gammaridae-present area
and the Gammaridae-absent area. However, there was no
difference in the composition of benthic invertebrates in
November relative to the remaining winter months, so it is
unclear why the diets and dietary overlap of brown trout
and masu salmon differed during this month.
Overall, the degree of dietary overlap obviously depended
on prey composition; it did not correlate with stomach fullness in masu salmon and brown trout, though the random
sampling times for the stomach contents may have obscured
the relationships, because salmonid species have a diurnal
feeding rhythm [23]. This suggests that competition for food
resources does not occur intensely, because each individual
has access to adequate food resources, even in instances
where masu salmon and brown trout prey on similar food
items. However, we may need to evaluate growth rates under
different food conditions in order to detect the detailed effect
of interspecific competition.

123


1170

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. Further studies should also determine how the seasonal variation in dietary overlap at a particular location affects the
outcome of interspecific competition throughout the year.
Acknowledgments We thank the staff of the National Salmon
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 Mamachi 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

123

Fish Sci (2012) 78:1163–1171
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 (Oncorhynchus 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)
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 salmonids on two native stream-dwelling salmonids through interspecific competition. J Fish Biol 68:1123–1132
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 microhabitats in a size-structured interspecific dominance hierarchy of
two sympatric stream salmonids in a natural habitat. Can J Zool
73:1845–1854
19. Kitano S, Hasegawa K, Maekawa K (2009) Evidence for interspecific hybridization between native white-spotted charr and
non-native brown trout on Hokkaido island, Japan. J Fish Biol
74:467–473
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


Fish Sci (2012) 78:1163–1171
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

1171
23. Young MK, Rader RB, Belish TA (1997) Influence of macroinvertebrate drift and light on the activity and movement of Colorado River cutthroat trout. Trans Am Fish Soc 126:428–437

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

123


Fish Sci (2012) 78:1173–1178
DOI 10.1007/s12562-012-0564-7

ORIGINAL ARTICLE

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, absolute 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 gonadosomatic 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. Overlapped 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
Fish

Semen Á Eggs Á Quality Á Hybrids Á Breams Á

B. Nzau Matondo (&) Á M. Ovidio Á J.-C. Philippart
Biology of Behaviour Unit, Laboratory of Fish Demography
and Hydroecology, University of Lie`ge, 10 Chemin de la Justice,
4500 Tihange, Belgium
e-mail:
P. Poncin
Biology of Behaviour Unit, Laboratory of Fish Ethology,
University of Lie`ge, 22 Quai Van Beneden, 4020 Lie`ge, Belgium

Introduction
Cultured F1 hybrids of Abramis brama 9 Blicca bjoerkna
present successful viability in terms of age and size, fertility, and sexual activity at first maturity [1]. The observed
reproductive success of hybrids raised an important question about the exact nature of the phylogenetic relationship
between these two species. We and others have suggested

that it would be better to combine these two species, currently belonging to a different genus status, in the same
genus [2–4]. Several studies on F1 hybrids of A. brama 9 Rutilus rutilus and B. bjoerkna 9 R. rutilus showed
a real weakness in their fertility, particularly in the production of F2 generations [5–10]. This prompted us to test
the quality and quantity of the gametes of these two
hybrids, with results showing a depression of the absolute
fecundity and sperm density. Whereas a slight overlap in
the fecundity was found between these hybrids and their
parents, no possible overlap was observed concerning
sperm density.
In hybrid as in nonhybrid fish species, the quality and
quantity of gametes play a significant role in the developmental success of the crosses [11–13], and the study of the
quantity and quality of eggs and sperm should be a major
focus of all crossbreeding programs. In hybrids of A. brama 9 B. bjoerkna, a good quality of eggs and sperm
resulted in the high survival of offspring [1], but the levels
of gametes produced by these hybrids compared to their
parents in terms of sperm density, fecundity, and egg size
remain unknown. It is well known that high sperm density,
high fecundity, and large egg size are considered as an
advantage. Indeed, high sperm density is often associated
with high fertilization rates [14], large eggs with important
reserves of nutrients useful for survival and growth after
hatching [15, 16], and high fecundity with better

123


1174

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. bjoerkna 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

Fish Sci (2012) 78:1173–1178
Table 1 Fork length in mm and weight in g of hybrids and parent
species used in sexual product analysis
Females

A

Males


n

Features

Mean

Range

Mean

Range

10

Fork length

173a

114–220

181c

150–230

Weight

a

118


23–239

116b

60–216

AB

10

Fork length
Weight

152a
58a

108–160
26–79

128a
36a

105–155
17–57

BA

10

Fork length


151a

147–155

140b

128–156

Weight
B

10

Fork length
Weight

a

a

56

53–62

41

33–62

139a


145–157

131ab

120–145

55a

51–62

37a

27–46

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)

spawned per total body weight, the absolute fecundity was
considered as the total number of eggs spawned per female
and calculated from two samples of eggs (1 g), and the egg
size was determined from samples of 50 eggs per female
that were individually measured using a microscope fitted
with an ocular micrometer [9, 10].
Sperm evaluation

The mature fish examined in this study (Table 1) 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°450 N; 5°400 E) [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.

To estimate the density and consistency of sperm, semen of
ten selected males for each type of F1 hybrid and parental
species was individually extracted by abdominal pressure.
The consistency of sperm was determined as lactic when
semen was a white liquid or as watery when semen was a
gray liquid by macroscopic examination. The sperm concentration was back-calculated after counting the spermatozoa in milt extracted with a syringe and diluted 200-fold
with 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 a
phase contrast microscope (9400).

Female sexual product analysis

Statistical 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 simulating 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

Comparisons of GSI, fecundity, egg size, and sperm
density between hybrids and parental species were made
using the Kruskal–Wallis (KW) test followed by multiple
paired comparisons tests, the Mann–Whitney U-test.
Fisher’s exact probability (FEP) test was used to compare the relative frequency of sperm consistency in
hybrids and parents, and the Chi-square (v2) test to
compare lactic and watery sperm in hybrids. For all
statistical analyses, a probability level of p \ 0.05 was
considered significant.

123


Fish Sci (2012) 78:1173–1178

1175

Results

Semen analysis

Female sexual product


The comparison of semen (Fig. 3) also revealed a significant 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 (FEP
test, p \ 0.05). AB and BA hybrids (median values =
0.2 9 1010 and 0.3 9 1010 spermatozoa ml-1, respectively) showed substantially lower sperm density (U-test,
p \ 0.05) than parental species (1.2 9 1010 and 1.4 9 1010
spermatozoa ml-1 for B. bjoerkna and A. brama, respectively) (Fig. 3a). A slight overlap was observed between
the BA hybrids and their parents. The difference was not
significant between hybrids, but between parental species,
the sperm density of A. brama was found to be significantly
higher than B. bjoerkna. For the consistency of sperm, the
percentage of males with sperm of lactic consistency was
significantly higher (FEP-test, p \ 0.05) in both parental
species, accounting for 100 %, whereas that of AB and BA
hybrids accounted only for 50 % and 60 %, respectively
(Fig. 3b). In these hybrids, no significant difference was
found (v2-test, p [ 0.05) between lactic sperm and watery
sperm in terms of proportion of fish. However, the sperm
consistency of hybrids significantly affected their spermatozoa concentration (KW-test: df = 3, H = 16.012,
p = 0.0011). A significantly higher spermatozoa concentration (U-test, p \ 0.05) was observed in lactic sperm
(0.47 9 1010 and 0.46 9 1010 spermatozoa ml-1 for AB
and BA hybrids, respectively) than in watery sperm
(0.17 9 1010 and 0.10 9 1010 spermatozoa ml-1 for AB
and BA hybrids, respectively) (Fig. 3c).

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 distribution 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).

(a) 20
bc

3

16

GSI, (%)


14
12
10

ab

8

a

6

30

20
15

2

0
B

BA

b

10
5

Hybrids and parental species


ab

25

4
A AB

Discussion

35

c

Fecundity, (eggs × 10 )

18

(b) 40

a

a
A AB

B

BA

Hybrids and parental species


Fig. 1 Comparison of gonadosomatic index at spawn (GSI), absolute
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)

The GSI and fecundity of hybrids were found to be low but
with an overlap with parents. In these hybrids as in other B.
bjoerkna hybrids, neither GSI nor fecundity can clearly
differentiate between hybrids and parents [9]. The egg size
also showed the same trend but with a slight advantage for
the hybrids. This can be considered as a benefit for hybrids,
since larger eggs are associated with more nutrient reserves
useful for survival and growth after hatching [15, 16]. For
these three criteria characterizing the female sexual product
analysis, an overlap observed between hybrids and their
parents could mean that these hybrids have a high reproductive capacity and a successful gametogenesis process.
Thereby, our criteria may be inadequate for analyzing
sexual 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. rutilus 9 B. bjoerkna F1 hybrids [9], but egg sizes were very

123



1176

Fish Sci (2012) 78:1173–1178

Fig. 2 Frequency distribution,
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
different (Mann–Whitney
U-test, p \ 0.05)

200

(a)

(b)
B

100

BA

AB

100


200

A
100
0
0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

(c)
-1

10

3
100

a
0

-1
AB

B

BA

Hybrids and parental species

Watery


60

1
AB

BA

B

1.2
1.0

c

0.8

c
0.6
0.4

b
0.2

a

40
0
20
A


AB

BA

B

Hybrids and parental species

BA/L

a

80

1.2

BA/W

1

A

*
Lactic

b

Consistence, (%)

c


*

1.4

1.3

1.4

AB/L

2

c
1.5

1.6

AB/W

Spermatozoa, (cells × 1010 ml-1)

4

(b)

1.4

a


1.6

Hybrids and parental species

Spermatozoa, (cells × 10 ml )

5

b
1.6

A

Egg size, (mm)

(a)

d

1.8

0

1.2 – 1.7

200

1.4 – 1.8

0


1.0 – 1.8

0.9 – 1.7

100

Egg diameter, (mm)

Absolute frequencies, (number)

0
200

Sperm consistency of hybrids

Fig. 3 Comparison of sperm density and consistency for hybrids and
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)

similar between these two types of hybrids. On the contrary, 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 that
these hybrids were also closer to the parent species of A.
brama with regard to age, size, and reproductive tactics at
first sexual maturity [1]. In contrast, in F1 hybrids of R.
rutilus 9 A. brama the absolute fecundity was significant
lower than in parental species [10], suggesting a low
gametogenesis efficiency. In F1 Clariidae hybrid fish,
considerably lower GSI and fecundity than in parental
species were also found [21]. In AB hybrids, two peaks of

123


Fish Sci (2012) 78:1173–1178

egg size observed matching mean values of parental species 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. According 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. However, in hybrids of A. brama 9 B. bjoerkna, higher
maximal values and white semen overlapping with parents were observed, which again means a high reproductive 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. brama 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 populations of parental species. The new reproductive success of
these hybrids again raises the question about the phylogenetic 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.

Acknowledgments Authors are grateful to D. Sony, G. Rimbaud,
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.

1177

References
1. Nzau Matondo B, Ovidio M, Philippart JC, Poncin P (2008)
Sexual maturity, reproductive behaviour and fertility in the firstgeneration 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 cyprinids: insights from allozymes. J Fish Biol 57:265–276
5. Wood AB, Jordan DR (1987) Fertility of roach 9 bream hybrids,
Rutilus rutilus (L.) 9 Abramis brama (L.) and their identification. 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 cyprinid fishes in a novel habitat: genetics, morphology and lifehistory 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 rutilus (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-generation 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 punctatus) 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 survival in brown trout Salmo trutta L. J Fish Biol 1:349–353
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 recruitment: 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

123