Fisheries science JSFS , tập 78, số 3, 2012 5

The horizontal wave force was related to the height and width of the netting panel, wave height, wave length, twine diameter, bar length of the mesh, and sloping angle relative to the wa

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

Experimental study on the effect of waves on netting panels

at a range of angles to the wave direction

Wei-hua Song•Zhen-lin Liang• Rong Wan•

Fen-Fang Zhao• Kinoshita Takeshi•

Liu-yi Huang•Jia-zhi Ma•Bo-hai Chen

Received: 31 July 2011 / Accepted: 13 February 2012 / Published online: 28 March 2012

 The Japanese Society of Fisheries Science 2012

Abstract The effect of horizontal waves on flexible

netting panels placed at angles to the wave direction is

studied with the aim of evaluating the testing method of

pre-tensioned mooring and radial systems and flexible

netting structures The netting was hung on a frame at a

specific hanging ratio for ten types of polyethylene panels

Regular waves were experimentally generated with a wave

period varying from 0.8 to 2.0 s and wave height ranging

from 50 to 250 mm The force on the netting structure was

recorded by a tension transducer and a digital signal

recorder The results showed that the horizontal wave force

on the netting panel changed periodically and

asymmetri-cally when it was back and left or right declinate to the

wave direction; similar results were found for the surface

wave elevation The opposite results were obtained when

the sloping angle declinated front to the wave direction,

with two obvious crests during each period The horizontal

wave force was related to the height and width of the

netting panel, wave height, wave length, twine diameter,

bar length of the mesh, and sloping angle relative to the

wave direction Using dual series relations, the least squareapproximation, and multiple stepwise regression analysis,the formula for estimating the maximum value of the waveforce on the netting was obtained

Keywords Netting panel Horizontal wave force Sloping angle Flume experiment

IntroductionHydrodynamics of the fishing gear has been a main focus

of researchers in fisheries science for many years and can

be traced back to the work of Tauti [1,2] and Baranow [3]

in the 1930s, who used traditional formulae to calculateforces exerted on fishing gear and/or their components.Various calculation methods have been developed duringthe intervening years based on a combination of empiricalformulae and numerical computations, as described in She[4] and Matuda [5] Although physical experiments onfishing gear hydrodynamics have been conducted for manyyears, and many theories have evolved, most work hasfocused on the effect of water currents, with the effect ofwaves on fishing gears receiving less attention This isparticularly the case of netting panels positioned not nor-mal to the wave direction

Fish cages are composed of relatively rigid parts, i.e.,the frame, and more flexible parts, i.e., netting panels andmooring lines—with both rigid and flexible componentsconsidered to be fixed fishing gear These two main com-ponents react differently in terms of hydrodynamic prop-erties to the actions of waves and currents In addition,research methods are different for currents and waves Thedynamic behavior of cage frames usually belongs to theresearch field of ocean engineering, while the net falls

W Song ( &)  J Ma

Department of Marine Fisheries, Zhejiang Ocean University,

Zhoushan 316004, China

e-mail: whsong6806@163.com

W Song  Z Liang  R Wan  F.-F Zhao  L Huang

College of Fisheries, Ocean University of China,

Qingdao 266003, China

W Song  F.-F Zhao  K Takeshi

Institute of Industrial Science, University of Tokyo,

within the scope of fishing gear hydrodynamics A number

of studies have been conducted in which the cage system

was considered as a whole and the hydrodynamic forces of

the spar cage determined using measurements made in the

ocean and flume tanks, respectively, and by numerical

simulations [6 8] Song et al [9 11] presented the iterative

calculating method for the main cage’s individual

compo-nents under different wave conditions and proved the

validity of the method by wave flume experiments The

results indicated that it is possible to separate the cage wave

force into stiff and flexible forces Forces on the rigid

frames have been calculated by the Morison equation [12],

and forces on the flexible netting panels and ropes are

tackled according to traditional formulae in fishing gear

mechanics The nets on the fish cage are often the largest

component and have been the focus of drag force research

[13] Using theoretical and model test methods, Aarsnes

et al [14] determined the force and blockage characteristics

for individual net types Colbourne and Allen [15]

con-ducted a field experiment measuring waves and the load and

motion response of a gravity type net-pen and then

com-pared the results with physical models; however, they

focused on fishing gear mechanics only in terms of current

response In contrast, some researchers have suggested that

the wave forces could be calculated using the theory of

fishing gear mechanics through the conversion of the water

particle velocity according to different wave conditions

[13] Song et al [16] also have carried out experiments on

wave force on netting structures at positioned normal to the

wave direction in the wave flume and determined the

rela-tionships between the wave force and various parameters

under regular wave conditions and net panel dimensions

The aim of the study reported here was to examine the

effect of horizontal waves on netting panels located so as

not to be normal to the wave direction and to examine the

effect of horizontal waves on netting panels positioned at

various angles to the wave direction and flexible netting

structures The purpose of this study is to determine the

expression of maximum wave force in terms of wave

length, wave height, twine diameter, among other factors

Materials and methods

Regular wave theory

Regular wave tests were conducted using physical

model-ing methods for seven monochromatic waves with

fre-quencies spanning the expected forcing band The

characteristics of these waves were approximated using

linear wave theory as described by Dean and Dalrymple

[17] and were characterized by the following velocity

potential (u)

u¼ gH4pf

cosh kðz þ d0Þcoshðkd0Þ sinðkx  2pftÞ ð1Þand surface elevation (g)

g¼ A0cosðkx  2pftÞ ¼H

2cosðkx  2pftÞ; ð2Þwhere A0is the wave amplitude (equals to H/2, H the waveheight), g is the gravitational acceleration, f is the fre-quency (equals the reciprocal of T, which is the waveperiods), k is the wave number (equal to 2p/L), L is thewavelength, d0 is the water depth, z is the vertical position

in the water column, and x is the horizontal position

In view ofð2pf Þ2¼ gk tanhðkd0Þ; sinhðkd0Þ ¼ coshðkd0Þ  tanhðkd0Þ;

L¼gT2

2p tanh kd

0;

Equation1 can be derived as

u¼pHTk

cosh kðz þ d0Þsinhðkd0Þ sinðkx  2pftÞ ð3Þwith

ux¼ou

ox; uz¼ou

ozTherefore, the fluid element velocity of the position (x,z) is obtained as:

ux¼pHT

cosh kðz þ d0Þsinhðkd0Þ cosðkx  2pftÞ

uz¼pHT

sinh kðz þ d0Þsinhðkd0Þ sinðkx  2pftÞ

of the wave is defined as the following [12,18–20],

According to the experiments conducted by Kuznetsov[13] and Song et al [16], we consider that the apex value ofthe horizontal wave force is drag force when the fluidelement velocity is at its maximum value in a periodbecause the physical volume of the netting is smaller thanthat of the frame volume, and the inertial force is smaller

than that of the drag force [1 3] Combined with the

dimensional analysis method, the horizontal wave forces

on the differential area (DA) of the netting could be defined

as

X

dF¼X

where Cx is the pending coefficient related to the netting

characteristic coefficient d/a (d is diameter of the mesh

twine, and a is bar length of the mesh), KC (waves periods

parameters), Re(Reynolds number), and H/L [16]

Physical modeling

In our study, there were ten experimental netting panels(Table1), each of which needs to be fixed to a framebecause the netting is flexible Each netting structure wassloped at a different angle to the wave direction under theregular wave condition (Fig.1) [16,21–27]

Based on the motions of the aquaculture fish cage in theopen ocean field [6,7] and the experimental conditions ofKuznetsov on the netting wave force (as described in [13]),the netting structures produced similar motions to the cageunder our wave conditions [20,21], including six compo-nents of responses (surge, sway, heave, roll, pitch, andyaw) To ensure that the netting structures could not pro-duce any motion during the wave experiments, the method

of pre-tension and radial moorings were selected to fix thenetting structures onto the wave flume [22–25] (Fig.1).The wave experiments were conducted under a pre-tensionwhich was approximately equal to the maximum horizontalwave force every time, and the mooring line was made ofthin steel wire which could withstand tension withoutelongation (Fig.1) The horizontal distance between themooring position of the front wave height meter and one ofthe netting panels was 5.0 m (Fig.1), and the distancebetween the mooring position of the rear wave height meterand one of the netting panels was 4.0 m; a and b are thefront–back sloping angles and the right–left sloping angles,which were adjusted through the variation of four tensionleg length Based on the wave tank conditions, these were

Table 1 Characteristics of the ten netting panels used as physical

Twine diameter (mm)

d/a The netting characteristic coefficient (d is diameter of the mesh

twine, and a is bar length of the mesh)

Fig 1 Installation of netting structure in the wave flume and experimental apparatus

a1= 30, 45, 60, 75, 90 (forward sloping), a2= 90, 105,

120, 135, 150 (backward sloping), and b = 30, 45, 60, 75,

90 (right–left sloping), respectively The sensor can record

the horizontal wave signal The wave force on the netting

panel and the wave parameter were recorded

synchro-nously so that it was possible to calculate the phase

dif-ference between the force and the wave

Physical model tests using regular waves were

con-ducted in the 65 9 1.2 9 1.7-m flume tank of the wind–

wave–current Flume Laboratory of the Ocean University of

China The working water depth was 0.70 m, the wave

period was 0.8–2.0 s, and the wave height was 50–

250 mm The netting structures were moored 20 m away

from the wave maker, which produced regular waves The

instrument for measuring force consisted of a transducer,

computer, and digital recorder, and it was able to measure

forces ranging from 0 to 40 ± 0.01 N The measuring

technique was gradually improved in the field [13,16,25,

26]

The bottom of the netting structure has a stated distance

with the bottom of the pool When the sloping angle a1(or

a2) was 30 (or 150), the distance between the center point

of the frame and the bottom of the pool was the same as

that between the point where the tension line and the frame

junct and the bottom of pool Then the distance between

the lower side of the frame and the bottom of the pool is

500 mm For example, when the wave height is 250 mm,

the water surface cannot overflow the top of frame (i.e., the

highest position of the water surface is 825 mm, and the

highest position of the frame top is 900 mm)

Song et al [9 11, 16, 25] showed that the horizontal

wave force is influenced by the netting structures

dimen-sions In their studies, the wave force had a linear

rela-tionship with the netting structure width, but an exponential

relationship with the structure height sinking into the water,

with the force decreasing dramatically with the sinking

depth of the netting structure from the surface, being nearly

equal to zero when the height was equal to the four-fifths of

the flume depth from the surface Considering the above

results and flume dimensions, the test frame was designed

at 0.9 9 0.8 m in order to eliminate the interaction

between the flume and the test structure The frame was

made of high-strength polyvinyl chloride pipe (Qingdao

ShenBon Sea cage engineering and technology Co.,

Qingdao, Shandong, China) The top of the frame was

150 mm above the water surface to ensure that the wavesurface would not submerge the entire netting structures atmaximum wave height The nets were constructed usingknotted high-strength polyethylene (China National Fish-ery Yantai Marine Fisheries Corp, Yantai, ShandongProvince, China) with the specifications shown in Table1.The hanging ratio of the netting structures was 0.707.Experimental data analysis

Data were collected as part of the experiment program,including experiment surface waves, the horizontal waveforce on the netting structures, and dimensions under dif-ferent sloping angles The wave parameters were different

in each experiment because the wave conditions werelimited by the wave maker The wave heights in ten nettingpanels and the frame experiments were different, but thedifference was small In order to attain the wave force onthe netting panel, data processing techniques included thecubic spline method because the force recorded waseffected by all the netting panel and the frame

For example, in the wave experiment during the sameperiod, the testing force Fnon the frame corresponding tothe experiment wave height Hn(H1\ H2\ H3\ \ Hn)

is given by

F¼ 3

h2 i

ðHiþ1 HÞ2 2

h3 iðHiþ1 HÞ3

Fi

h2 i

ðH  HiÞ2 2

h3 i

ðHiþ1 HÞ2 1

h3 iðHiþ1 HÞ3

Fi0

þ hi 1

h2 i

ðH  HiÞ2 1

h3 i

i = 1, 2,…, n - 1; n is the wave height number in eachperiod The critical condition is defined as F10 ¼ 0; F0

iþ1

¼ ðFiþ1 FiÞ=hi; i¼ 1; 2; ; n  1, and F00

1 ¼ 0 becausethe pre-tension force is constant Fi0 (i = 1, 2,…,n) differ-ential coefficient is given by

Table 2 Wave length, wave period, and wave height number

According to the above analysis, the wave force on the

netting panels under different angles can be dealt with

through the iterative calculation method [9 11,16]

Results

Variation of the horizontal wave force on netting panel

Wave force on the netting panel is similar to the variation

of the experiment wave Table2 presents data on the

relationship between the wave length and the wave periodand also gives the wave height number Taking nettingpanel F as the subject of our experiment, the conduct waveexperiment was conducted according to the front-and-backslope and right-and-left slope, respectively

Since the wave is two-dimensional sine wave, the sloping and right-sloping waves are consistent (if theexperimental components are symmetrical) Figures2, 3,and 4 show wave forces on the netting panel F at angle

left-a1= 30 and wave height 185.2 mm, angle a2= 150 and

Fig 2 Wave force on the netting panel (F) under the condition that

the angle a1is 30, the wave height is 185.2 mm, and the wave period

is 1.6 s

The horizontal wave force of net-F structures in period 1.6s and

heights 185.4mm

-2 -1 0 1 2 3

The horizontal wave force of net-F structures in period 1.6s and

heights 189.5mm

-1 -0.5 0 0.5 1 1.5 2 2.5

F10

F20

Fn0

266666

377777

¼

f2f3

.fn1

266666

377777

ð8Þ

wave height 185.4 mm, and b = 30 and wave height

189.5 mm, respectively under the same period 1.6 s

Tables3 and 4 presented the relationships between the

wave force on the netting panel F and its angles of a or b

According to data presented in these tables, the change

in wave force was clearly different when the angle is

mainly formed by forward–backward sloping (e.g., the

angles a130–75 corresponds to the angles a2105–150)

If each wave force during a cycle at a1= 30 has two

amplitudes, the wave force on the netting panel is fairly

complicated Based upon our calculations, Figs.5 and 6

show the relationship between the horizontal wave force on

the netting panel and wave height and length The

rela-tionship between wave force of the netting panel and wave

height is fairly evident when there is an angle of attack, but

the relationship between wave force of the netting panel

and wave length cannot be clearly determined

However, the reasons for such complicated variation

were that the horizontal wave force on the netting panel

from top to bottom was not in phase, and was even

sometimes in reverse, which is different from the state

when the netting panel is positioned so as not to be normal

to the wave direction Therefore, the value for the wave

force on the whole netting panel exhibits several crests andasymmetry that are related to the sloping angle forms(angles a1, a2, and b), the measurement of the nettingpanel, wave length, among others

Moreover, the experiment results revealed that the izontal wave force on the netting panel had two or morewave apex values in a wave period at angle a1 formsbecause the wave force was affected by the effect of theout-phase wave force (referring to its drag and its inertia),and its wave apex value was less than the angle a2ones.Multiple stepwise regression analysis

hor-In general, the horizontal wave force on the netting wasprimarily related to netting dimensions, including the netwidth (l) and height (h), the twine diameter, bar length,wave characteristics, such as wave height and wave length,and the sloping angle relative to the wave direction Thedimension of the net width was fixed, and the dimension ofthe net height should be attributed to that of the waveheight Therefore, the dimensions of wave height H, length

L, twine diameter d, bar length a, and sloping angle aespecially were examined

Table 3 Relationships between the wave force on the netting panel (F) and angle b at wave period 1.6 s

Angle b Wave height (mm)

Table 4 Relationships between the wave force on the netting panel (F) and angle a at wave period 1.6 s

Angle a Wave height (mm)

Based on the experimental data and Eqs.5 and 6, the

apex value expression of the horizontal force on the netting

where Fmax is the apex value of the wave force, m is

number of components of the wave force, C0, and ai, ai?1,

ai?2, ai?3, ai?4, ai?5are pending coefficients Because the

netting’s mass is small in the water, the inertial force which

is affected by the mass is small compared with the drag

force [1 3] According to the experimental results of

Kuznetsov [13] and Song et al [16], the apex value of the

horizontal wave force may be considered as drag force only

in each period, with Eq 9derived asFmax¼ C0Ha 1La2aa3da4la5aa 6 ð10ÞFollowing logarithmic transformation, the aboveequation is expressed as

lnðFmaxÞ ¼ lnðC0Þ þ a1lnðHÞ þ a2lnðLÞ þ a3lnðaÞ

þ a4lnðdÞ þ a5lnðlÞ þ a6lnðaÞ ð11ÞThe coefficients were analyzed using the method of theleast square approximation and multiple stepwise regressionanalysis based on actual experiments [16,19] The formulafor calculating the maximum value of the wave force on thenetting is:

Relations between the wave force and the wave height

slopping angles slopping angles

Fig 5 Relations between the wave force on the net (F) and wave

height at a period of 1.6 s under different the sloping angles

0 1 2 3 4

Slopping angle (°)

Wave length 2.17m Wave length 2.80m Wave length 3.43m Wave length 4.03m Wave length 4.62mRelations between the wave force and the front-and-back

0 1 2 3 4

Slopping angle (°)

Wave length 2.17m Wave length 2.80m Wave length 3.43m Wave length 4.03m Wave length 4.62m

Relations between the wave force and the

right-and-left slopping

Fig 6 Relations between the wave force on the net (F) and the wavelength at wave height 150 mm under different sloping angles

F¼ expa0ðHÞa1L0:79a1:48d1:06l1:05aa6 ð12Þ

The wave force on the netting panels was related to

these three sloping angle forms in Table5, which is a

negative relation to the wave force at angle a2 increased

from 90 to 150 The relationship between the wave

force on the netting panel at these sloping angles and the

wave height showed 2.5–3 power, with the smaller the

sloping angle (only about an acute angle), the lower the

coefficient of relationship The relationship between the

wave force and the wave length showed 0.8 power, whichwas related to the wave length and the measure of thenetting panel

Therefore, the calculation formula for the maximumcoefficient of the wave force is as follows when the nettingpanel is perpendicular to the wave direction,

F¼ 2669:4ðH=2Þ2:62L0:79a1:48d1:06l1:05: ð13Þ

DiscussionAccording to Figs.2, 3, and 4, components of the waveforce on netting F should be analyzed The fast Fouriertransform (FFT) and integral were used, and the component

of the wave force at the sloping angle (a1, a2, and b) of 30˚when the the netting was not normal to the wave directions

is shown in Table6.The relationship between the wave force ingredients onthe net F and the wave height at wave periods 1.6 s andthree sloping angles of 60 is shown in Fig.7, which

is clearly a square relationship However, it is a linear

Table 5 Coefficient of relationships between wave force on the net

and the two parameters of wave height and angles

a1, a6, Coefficients of relationship between the wave force and the

wave height or sloping angle, respectively

Table 6 Components of the wave force on netting panel (F) when the panels were in a position not normal to the wave directions at wave period 1.6 s

Forms of sloping angle a The number of wave height

relationship when the wave height is more than 150 mm

and the wave excitation forces has a fourth-order and

fifth-order relationship in addition to including three

excitation forces and one drifting force (Fig.7b) For

example, Fig.8 showed five pulses of excitation wave

forces on the netting F by the FTT method FFT and

integral [16, 25]

Based on the FFT analysis, the simulation formula for

the wave force is as follows,

first-The comparison between the wave force on the netting(F) obtained by the simulation method and those obtained

by the experimental method at period 1.6 s, wave height189.5 s, and angle b 30 is shown in Fig.9, which illus-trates that the simulation wave force by FFT was similar tothe results obtained by the experimental method

These phenomena and variation are related to the ing forms of the netting panels, the sloping angles, thewave periods, the measurements of the netting panels,among others

slop-The horizontal wave force on the netting panel changedperiodically and asymmetrically when it was not normal tothe wave direction, which was similar to the surface waveelevation The horizontal wave force is related to the fol-lowing parameters: netting panel height (l) and width (h),wave height (H), wave length (L), twine diameter (d), barlength (a) of the mesh, and inclination angle (a or b) to thewave direction At the back slope, the wave force isaffected by one wave apex value However, at the front andside slope, the wave force is affected by more than twowave apex values The variation resulted from the change

in the height of netting panels, sloping angles, and wavelengths and has many components, such as excursion for-ces and excitation forces

The different kinds of excitation forces were beyond thescope of this study In the future, we will be studying therules and reasons for variations between the wave force

The relation between drifting forces f0

and wave height

0.0 1.0 2.0 3.0

The relation between excitation forces f1

and wave height

0.0 1.0 2.0 3.0

The relation between excitation forces f

and wave height

0.0 0.5 1.0 1.5 2.0

The relation between excitation forces f3

and wave height

0 0.4 0.8 1.2

Fig 7 Relationship between

the wave force ingredients on

the net (F) and the wave height

at a wave period of 1.6 s under

the sloping angle 60

Fig 8 Five pulses of excitation wave forces on the netting (F) at

period 1.6 s, height 204 mm, and angle a260

components and the experimental parameters by the

non-dimension and the comparison method Only then will it be

possible to calculate and simulate wave forces on the

net-ting panels under random conditions

Acknowledgments This research was supported by funding from the

National Natural Science Foundation of China (NSFC: 40876049,

40776052, 30972256), National Project ‘‘863’’ China Grant No.

2006AA100301, Zhejiang Province Science and Technology

Depart-ment Important Project of China No 2008C12065-1, Zhejiang Province

Teaching Department Important Project (No Z200803912), the Open

Foundation from Ocean Fishery Science and Technology in the Most

Important Subjects of Zhejiang, Key Physics Ocean Laboratory,

Min-istry of Education, China The authors thank Q Liu, Y Xue (Ocean

University of China, China), C.T Guan (Yellow Sea Fisheries Research

Institute, China), X.W She, H.C Yu (Zhejiang Ocean University,

China), and Pingguo He (University of Massachusetts Dartmouth,

School for Marine Science and Technology, USA) for their support.

References

1 Tauti M (1934) A relation between experiments on model and

full scale of fishing net Nippon Suisan Gakkaishi 3:171–177

2 Tauti M (1934) The force acting on the plane net in motion

through the water Nippon Suisan Gakkaishi 3:1–4

3 Baranow FI (1948) Theory and assessment of fishing gear Fish

Industry Press, Moscow, pp 67–95

4 She XW (2001) Instruction to calculation of fishing gear

mechanics Shanghai Science and Technology Documents Press,

Shanghai, pp 218–247

5 Matuda K (2001) Fishing gear physics Seizando Press, Tokyo,

pp 21–42

6 Fredriksson DW (2001) Open ocean fish cage and mooring

sys-tem dynamics PhD thesis University of New Hampshire,

Dur-ham, pp 138–154

7 Fredriksson DW, Swift MR, Irish JD, Tsukrov I, Celikkol B

(2003) Fish cage and mooring system dynamics using physical

and numerical models with field measurements Aquac Eng

27:117–146

8 Tsukrov I, Ozbay M, Fredriksson DW, Swift MR, Baldwin K,

Celikkol B (2000) Open ocean aquaculture engineering:

numer-ical modeling Mar Technol Soc J 34:29–40

9 Song WH, Liang ZL, Guan CT, Zhao FF, Huang LY, Zhu LX

(2004) Calculation and experiment of horizontal wave and proved

by experiment in wave pool for a square sea-cage Oceanol Limnol Sin 35:202–208

10 Song WH, Liang ZL, Guan CT, Zhao FF, Huang LY, Zhu LX (2005) Calculation on horizontal wave force using two wave theories for a square sea-cage Fish Sci China 12:91–98

11 Song WH, Liang ZL, Zhao FF, Huang LY, Zhu LX (2003) Approximate calculated on waving-force for a square sea-cage hydrodynamics J Zhejiang Ocean Univ 23:211–220

12 Morison JR, Johnson JW, O’Brien MP, Schaaf SA (1950) The forces exerted by surface waves on piles Pet Trans Am Inst Min Eng 189:149–157

13 Kuznetsov AH (1988) The wave resistance to the netting panel Fishing gear theory and design (translated from Russian by Hou

EH and Gao QL) Ocean Press, Beijing, pp 36–47, 57–63

14 Aarsnes JV, Rudi H, Loland G (1990) Current forces on cage, net deflection Engineering for offshore fish farming Thomas Tel- ford, London, pp 137–152

15 Colbourne DB, Allen JH (2001) Observations on motions and loads in aquaculture cages from full scale and model scale measurements Aquac Eng 24:129–148

16 Song WH, Liang ZL, Chi HF, Huang LY, Zhao FF, Zhu LX, Chen BH (2006) Experimental study on the effect of horizontal waves on the netting panels Fish Sci 72:1379–1386

17 Dean RG, Dalrymple RA (1991) Water wave mechanics for engineers and scientists World Scientific Publ Singapore,

21 Linfoot BT, Hall MS (1986) Analysis of the motions of model sea-cage systems Autom Data Process Aquac Trondheim Nor 8:18–21

scale-22 Baldwin K, Celikkol B, Steen R, Michelin D, Muller E, Lavoie P (2000) Open aquaculture engineering: mooring and net pen deployment Mar Technol Soc J 34:53–67

23 Kimura H (1998) Numerical calculation of horizontally fully immersed and two points moored cylindrical buoy in wave Fish Eng 35:167–176

24 Turner R (1997) Mooring system design Offshore mariculture CIHEAM-Options Mediterrane´ennes, Scotland, pp 168–169

25 Song WH, Liang ZL, Zhao FF, Huang LY, Zhu LX, Chen BH (2005) Wave experiment on netting structure with single point moorings Oceanol Limnol Sin 36:202–208

The wave force on the netting F at period 1.6s and height 189.5mm under angle 30 degree

-1 -0.5 0 0.5 1 1.5 2 2.5

experimental value simulation value

Fig 9 Compare between the

simulation and experimental

value of wave force on the

netting (F) at wave period 1.6 s,

height 189.5 mm, and angle b

30

26 Swift MR, Palczynski M, Kestler K, Michelin D, Celikkol B,

Gosz M(1998) Fish cage physical modeling for software

devel-opment and design applications In: Proc 26th US–Japan

Aqua-culture Symp UJNR Technical Report No 26 UJNR

Aquaculture Panel, Sarasota, pp 199–206

27 Fujita S, Fukahori K, Nishinokubi H (1991) Tension of the mooring ropes on the net cage model of raft type in regular waves Nippon Suisan Gakkaishi 57:2229–2235

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

Relationship between light and diel vertical migration of Diaphus

theta and Euphausia pacifica off eastern Hokkaido

Tomohiko Matsuura• Kazuhisa Uchikawa•

Kouichi Sawada

Received: 14 October 2011 / Accepted: 14 February 2012 / Published online: 22 March 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract Underwater irradiance in sound-scattering

lay-ers (SSLs) during diel vertical migration (DVM) of marine

organisms was investigated by a drifting vessel at locations

in the western North Pacific during the period August

24–27, 2008 The inhabitants of the layers were identified

via sampling as Diaphus theta, a type of myctophid fish,

and Euphausia pacifica, a species of zooplankton

Under-water irradiance was estimated by the Beer–Lambert law

using irradiance measurements on the deck and the diffuse

attenuation coefficients obtained in the water During

DVM, the mean irradiances at 490 nm at the lower and

upper boundaries of the SSL composed of E pacifica were

-54.0 and -45.3 dB, respectively In contrast, the values

for the boundaries of the D theta SSL were -82.5 and

-59.3 dB, respectively It was possible to distinguish

between D theta and E pacifica based on these ranges and

the differences between the mean volume-backscattering

strength (DMVBS) at 120 and 38 kHz It was also possible

to distinguish between D theta/E pacifica and other causes

of scattering more clearly using these two parameters

together than when using the DMVBS parameters

separately Our results suggest that underwater irradiance is

an important parameter for discriminating among isms in DVM

organ-Keywords Diel vertical migration  Underwaterirradiance Diaphus theta  Euphausia pacifica  Speciesdiscrimination Mean volume-backscattering strength

IntroductionMesopelagic micronekton are an important component ofoceanic ecosystems because of their abundance and uni-versal distribution in the world’s oceans [1, 2] Diaphustheta and Euphausia pacifica are dominant species in themicronektonic fish and zooplankton communities, respec-tively, in the subarctic and transitional waters of the wes-tern North Pacific [3 5] These species migrate upwardevery night to feed in the productive epipelagic zone.Hence, information on the abundance of these species isimportant Acoustic methods are widely used for estimat-ing fish abundance [6] because they can be used to extracthigh-resolution data on the distribution and temporal var-iation of fish abundance, making it possible to assess largeareas in less time compared to net sampling methods.However, only a few acoustical surveys have been con-ducted on myctophids [4, 7,8], mainly because the lattertend to occur in multi-species groups and acousticalmethods have difficulty in differentiating among species insuch conditions, but also because myctophids providelimited target strength (TS) information In particular,multi-species conditions can lead to significant errors inbiomass estimation because it is difficult to differentiateamong species based on mean volume-backscatteringstrength (MVBS) [9] The ability to distinguish among

T Matsuura ( &)

Course of Applied Marine Environmental Studies,

Tokyo University of Marine Science and Technology,

4-5-7 Konan, Minato-ku, Tokyo 108-8477, Japan

e-mail: mtsr@affrc.go.jp

K Uchikawa

Japan Sea National Fisheries Research Institute,

Fisheries Research Agency, 1-5939-22 Suido-cho,

Chuo-ku, Niigata, Niigata 951-8121, Japan

K Sawada

National Research Institute of Fisheries Engineering,

Fisheries Research Agency, 7620-7 Hasaki, Kamisu,

Ibaraki 314-0408, Japan

DOI 10.1007/s12562-012-0481-9

species is critical for stock assessment Although species

allocation has been performed mainly by net sampling [10],

Kang et al [9] were able to discriminate between E

pacifica and walleye pollock Theragra chalcogramma

based on differences in MVBS values at two frequencies

While their method (the DMVBS method) is simple and

effective, additional species-specific information is needed

to distinguish among species more decisively In this

context, behaviors such as diel vertical migration (DVM)

are a key factor for discriminating among species The

causes of DVM include temporal changes in light,

tem-perature, and growth [11–17], but it is generally accepted

that DMV is primarily controlled by light [18–21] If the

relationship between light and DVM could be used to

distinguish among species in conjunction with acoustic

information (DMVBS), it would be possible to discriminate

among species more precisely

In the study reported here, we investigated the

rela-tionship between underwater irradiance levels and the

depths of different sound-scattering layers (SSLs) during

DVM of D theta and E pacifica and successfully

dis-criminated between these two species based on a

combi-nation of each species’ preferred irradiance range and their

DMVBS values

Materials and methods

Study site and data collection

Acoustic and irradiance data were collected on board the

R/V Wakataka-maru (692 tons, Tohoku National Fisheries

Research Institute, Fisheries Research Agency) during the

period August 24–27, 2008, with the vessel drifting at fixed

locations (Fig 1) in the western North Pacific (42°330N,144°000E) A quantitative echosounder (KFC3000; SonicCo., Tokyo, Japan) equipped with 38- and 120-kHz trans-ducers was used to observe the DVM of mesopelagicorganisms The beam width of both transducers was 8.4° Apulse width of 0.6 or 1.2 ms was used throughout thesurvey Interference-noise was observed on the 120-kHzchannel echogram because an acoustic current profileroperated at 130 kHz was actuated all day for automaticship-position control during drifting To filter out thisnoise, a denoising method (Echoview 4.90 Help file;

http://support.echoview.com/WebHelp/Echoview.htm) wasapplied using Echoview ver 4.10 software (Myriax,Hobart, Australia) After we had confirmed that the deno-ising method was effective, it was applied to all measuredacoustic data at 120 kHz A conductivity, temperature, anddepth (CTD) (SBE-9; Sea-Bird Electronics, Bellevue, WA)cast was made down to 400 m, and vertical temperatureand salinity profiles were obtained once per day throughoutthe survey To identify species composition and bodylength composition at a desired layer, net sampling wasconducted using a framed trawl net (Matsuda–Oozeki–Hu-Trawl [22]; MOHT) with a 5-m2 mouth opening and a4-mm mesh size The net was towed horizontally at2.5–4.0 kt (log speed) at the target depth where theobserved volume-backscattering strength (SV) values werehigh Net samples were categorized into groups of fish,euphausiids, and other zooplankton The total number wasmeasured for fish and subsamples, and approximately1/10th of the total in wet weight was used to estimate thebiomass of euphausiids and other zooplankton The numberand wet weight of each subsample were measured, and thetotal number was estimated by multiplying the number ofsubsamples by 10 The standard length (SL) of fish and thetotal length (TL) of dominant euphausiids were measuredfrom 200 and 100 randomly selected individuals, respec-tively Irradiance measurements were conducted from16:00 to 19:00 (Japan Standard Time, JST) on August24–26, corresponding to the period between dusk and night

at the measurement location, and from 17:00 to 19:00 JST

on August 27 On-deck (sea surface) and underwater diance were measured at six wavelengths (412, 443, 490,

irra-520, 565, and 670 nm) at a 2- to 3-Hz sampling frequencyusing a profiling reflectance radiometer (PRR-600/610;Biospherical Instruments, San Diego, CA) The underwaterunit (PRR-600) and on-deck unit (PRR-610) measuredunderwater irradiance, Iz, and sea surface irradiance, Iz=0,respectively, where z is the depth from the sea surface.During the irradiance measurements, the observation side

of the vessel was always oriented to face the sun The

PRR-600 was lowered to 100 m at a constant speed of 0.2 m/sand was pulled up to the sea surface at a speed of0.5–1.0 m/s

Fig 1 Observation locations (filled diamonds) in the western North

Pacific

Irradiance data processing

Assuming that seawater is optically homogeneous, the

underwater irradiance at z ? Dz (m) is predicted by the

Beer–Lambert law using underwater irradiance, Iz (lmol/

m2/nm), and diffuse attenuation, K, as follows:

where z (m) is the distance from the sea surface and

Dz (m) is water column thickness Diffuse attenuation

coefficients were estimated by using Iz and Iz ? Dz

mea-sured at an interval of Dz = 1 m in Eq.1 Because tilting a

radiometer causes errors in irradiance measurements, data

measured within a tilt angle range of ±5° were used At

deeper depths where irradiance could not be measured, the

mean value of K actually measured at shallower depths was

used to extrapolate irradiance In this way, a time-series of

irradiance vertical profiles at six wavelengths was

esti-mated from 16:00 to 19:00 JST on August 24–26, and from

17:00 to 19:00 JST on August 27

To investigate the irradiance level at the depth of a

vertically migrating SSL, outlines of the SSL were traced

on echograms by hand, and the depths of the traces were

extracted using software (Echoview ver 4.10; Myriax) By

comparing SV levels of the SSL at 38 kHz and at 120 kHz,

we were able to use the echogram with the stronger SSL to

trace the outlines At that time, the lower threshold level of

SV was set to -80 dB for both echograms [9] When the

SV values were smaller than the given SV threshold, we

assumed that there was no echo

Irradiance at the traced depths was calculated using Eq.1,

and the mean value was obtained The mean irradiances at the

upper and lower traced depths are the upper and lower

iso-lumes that the SSLs follow, respectively The depths

corre-sponding to each isolume were compared to the traced depths

using root-mean-square errors (RMSEs) to confirm whether

the organisms followed the isolume The RMSEs of both

depths in meters were calculated as follows:

where di is the depth at the traced SSL outline, zi is the

depth at the isolume, and N is the total number of

obser-vations This equation evaluates the dissimilarity between

the depths at the traced SSL outline and the depths at the

isolume

DMVBS calculation at the selected layer and a species

discrimination trial

The TS values of animals vary by frequency, which differs

by target size relative to wavelength and whether

swimbladders are present This frequency information can

be used to discriminate among species [8] The difference

in the MVBS at two frequencies equals the difference in

TS Kang et al [9] described DMVBS using two cies of 38 and 120 kHz as follows:

frequen-DMVBS¼ MVBS120kHz MVBS38kHz: ð3ÞDMVBS values at the selected regions were calculated inthe following ways The echo integration cell size wasselected as 5 m of depth and 90 s of time on the SVechogram Following the interference–noise reduction,the MVBS value at each cell was calculated and anMVBS threshold of -90 dB was applied to each fre-quency Mean DMVBS values within the regions sur-rounded by the upper and lower outlines of each SSL werecalculated using MATLAB software (The MathWorks,Natick, MA)

We distinguished among species in SSLs on August 26

by using the DMVBS range parameters and the irradiancerange at each layer obtained on the other 3 days (August

24, 25, and 27) The weighted mean and standard tions (SD) of DMVBS according to the number of obser-vations each day were calculated to obtain the DMVBSrange that was specific for a species Using these values,

devia-we then expressed the range of DMVBS at each SSL as themean ± 1.96 SD, which indicates a confidence interval forDMVBS Mean irradiance at the upper and lower outlineswas obtained at the wavelength where the estimated RMSEexpressed by Eq.2was the lowest Finally, the mean area-backscattering strength (SA), which is a decibel value ofthe product of fish density and mean target strength in thelinear domain, was estimated within the determined range

by three different methods: (1) using only the DMVBSrange (DMVBS method); (2) using only the estimatedirradiance range (irradiance method); (3) using bothmethods within the determined region

ResultsVertical distribution of SSLs and species compositionTwo layers that showed periodic DVM were observed onechograms over the 4 days The SV values at 120 kHzwere stronger than those at 38 kHz for the upper layer.Conversely, the SV values at 38 kHz were stronger for thelower layer Therefore, the upper and lower layers (here-after referred to as Layer 1 and Layer 2) were traced by120- and 38-kHz echograms, respectively (Fig.2) Layers

1 and 2 were distributed at 90–110 and 150–250 m,respectively, in the daytime Layer 1 migrated to the seasurface during the 4-day study period, while Layer 2migrated to the sea surface on August 24 and to 30–70 m

on the other days Layers 1 and 2 were encountered at the

surface on August 24, from 16:00 to 17:00 JST on August

25, from 17:00 to 18:00 JST on August 27, and just before

the end of the DVM on August 26

In total, 3,447 fish were caught, and 3442 D theta were

caught (Table1) Among the euphausiids, E pacifica was

the most abundant species throughout the 4 days; on

August 24, this species made up [90% of the total catch

collected from Layer 1, and no fish were caught Therefore,

E pacifica was the dominant species in Layer 1 Although

the ratio of the number of D theta to that of total catch,

including euphausiids, was \30% in Layer 2, it was [99%

of the total fish catch throughout the 4 days In particular,

the maximum ratio of E pacifica to D theta was 72.4

(=12677/175 in Table1) on August 25 The body length

compositions of these species were similar among samples

(Table2)

Environmental conditions

The vertical temperature and salinity profiles were similar

during the 4 sampling days (Fig.3) Water temperatures at

the sea surface were 14–17°C and decreased sharply to

1.8–2.5°C from the sea surface to 150 m Salinity varied

from 33.1 to 33.7 psu, except near the sea surface The

largest change in temperature and salinity values occurred

between depths 50 and 150 m

The measured irradiances at the six wavelengths areshown in Fig.4 The maximum measurable depths at 670and 490 nm were 20 and 92 m, respectively The irradi-ance level attenuated exponentially at an approximatelyconstant rate at a depth of[20 m Thus, diffuse attenuationcoefficients per 1 m at five wavelengths were calculated by

Eq 1using irradiance levels obtained below 20 m As themaximum measurable depth at 670 nm was only 20 m,

670 nm was not used for further analysis Figure5 showsthe means and SD of the diffuse attenuation coefficients;the SD were small for all diffuse attenuation coefficients atapproximately 490 nm

Estimated irradiance at SSL depthsThe mean and SD of the estimated irradiance at the upperand lower outlines of Layers 1 and 2 for the 4-day studyperiod were calculated (Fig.6) The strongest among thefive wavelengths at the two layers was the mean irradiance

at 490 nm (Table3) The SD of Layer 1 was similar at 490,

520, and 565 nm and that of Layer 2 at 490 nm wassmaller than the SD of the other wavelengths The SD ofthe estimated irradiance at the lower outline was largerthan that at the upper outline on each layer at each wave-length (Table3)

Figure7compares the RMSEs at five wavelengths using

Eq 2 The RMSEs at Layer 1 were smaller than those at

Fig 2 Echograms during diel

vertical migration (DVM) and

the traced outlines of sound

scattering layers (SSLs) from

day to night over the 4-day

study period Upper horizontal

bar is the

volume-backscattering strength (SV)

scale Traced outlines, which

correspond to the top and

bottom edges of the SSLs, are

shown on the echograms by

dotted lines (Layer 1: upper

layer, traced by 120-kHz

echograms) and solid lines

(Layer 2: lower layer, traced by

38-kHz echograms)

Layer 2 In addition, those at the upper outline were mostlysmaller than those at the lower outlines, and this was mostclearly observed for Layer 2 Therefore, the irradiancerange was narrower for Layer 1 than for Layer 2 Amongthe wavelengths, the RMSE at approximately 490 nm wasthe lowest and that at 412 nm was the highest This was acommon characteristic of both layers, except for the upperoutline of Layer 2 at 490 nm on August 25 and the loweroutline of Layer 2 at 412 nm on August 24 In summary,the depths at a certain isolume, particularly at about

490 nm, agreed well with the depths at Layers 1 and 2during DVM Taking the results shown in Fig.6 intoaccount, we used the mean irradiance at 490 nm to dis-criminate between species The 490-nm irradiance ranges

at Layers 1 and 2 for species discrimination on August 26were -53.7 to -45.3 dB and -84.1 to -59.5 dB,respectively (Table4)

Table 1 Towed depth, time, and number of micronektonic animals sampled by the Matsuda–Oozeki–Hu trawler during the 4-day study period

a Towing applied a stair-step method at 240, 220, 200, and 170 m

Table 2 Measured body lengths of Diaphus theta and Euphausia pacifica

a All variables for body length are given in millimeters Mean body length is presented ± standard deviation (SD); n = number of individuals

D theta was measured as standard length (SL) and E pacifica was measured as total length (TL)

Fig 3 Vertical temperature and salinity profiles obtained on August

24–27

DMVBS at the selected layer and species

discrimination trial

The mean DMVBS values at Layers 1 and 2 were ?7.7 and

-4.8 dB, respectively, throughout the 4-day study period

(Table5) The positive DMVBS indicates that the MVBS

values at 120 kHz were higher than those at 38 kHz The

ranges of DMVBS for species discrimination on August 26

at Layers 1 and 2 were 1.7–14.6 and -9.8 to 1.1 dB,respectively (Table4)

The regions of Layers 1 and 2 were determined by theDMVBS range and the irradiance range at 490 nm (Fig.8)

Fig 4 Measured underwater

irradiance at six wavelengths

versus depth on August 24–27.

Underwater irradiance was

measured before sunset.

Symbols are identified at the top

of the figure

Fig 5 Mean and standard deviation (SD) values of the diffuse

attenuation coefficients at five wavelengths Each set of four circles

with bars at each wavelength correspond, from left to right, to values

measured on August 24, 25, 26, and 27, respectively

Fig 6 The estimated mean irradiance at Layers 1 and 2 at five wavelengths on August 24–27 Filled inverted triangles, filled triangles mean irradiance on the upper and lower outlines of Layer

1, respectively Open inverted triangles, open triangles irradiance on the upper and lower outlines of Layer 2 Vertical bars SD Each set of four symbols correspond, from left to right, to the mean diffuse attenuation coefficients on August 24, 25, 26, and 27, respectively

The corresponding estimated mean SA values are

sum-marized in Table6 The mean SA value estimated by the

DMVBS method was the largest among the three methods,

and that estimated by the irradiance method was slightly

larger than that using both methods together

Discussion

We confirmed that the underwater irradiance values at

certain depths, particularly irradiance at 490 nm, agreed

well with the depths of each species during DVM, asshown in Figs.6 and 7 In the following sections, wediscuss potential errors associated with the use of theirradiance range to distinguish among species Theseerrors include contamination from other species, theselected threshold level, and estimated irradiance inthe SSLs We also demonstrated that the species dis-crimination method using a suitable irradiance rangefor E pacifica and D theta together with traditionalparameters, such as the DMVBS range, reducesuncertainty

Table 3 Estimated irradiance

at depths of Layers 1 and 2 over

the 4-day study period

Data are presented as the mean

± SD Units: lmol/m2/nm

Layer Wavelength (nm)

Layer 1 Upper -68.8 (±9.4) -57.6 (±4.6) -45.3 (±3.1) -48.2 (±3.1) -51.1 (±2.8) Lower -85.5 (±12.2) -70.0 (±5.8) -54.0 (±3.7) -58.1 (±3.5) -61.9 (±2.9) Layer 1

Upper -95.9 (±13.8) -77.4 (±9.1) -59.3 (±4.8) -64.1 (±5.9) -68.5 (±6.3) Lower -140.9 (±23.1) -110.1 (±18.2) -82.5 (±9.1) -90.3 (±12.6) -98.2 (±14.0)

Fig 7 Root-mean-square errors

of depth for the estimated mean

irradiance and the traced

outlines at five wavelengths on

August 24–27 Symbols are the

same as those in Fig 6

Table 4 Estimated species

Species allocation in Layers 1 and 2

Although net sampling in Layer 1 was conducted only

once, our results show that Layer 1 was composed mainly

of E pacifica (Table1) The catches from Layer 2 may

have included some organisms from Layer 1 because the

MOHT net cannot be closed at a given target layer and thus

may have collected some organisms from other layers

during shooting and hauling The ratio of the number of

D theta to that of E pacifica on August 25 was particularly

low compared to that on the other days (Table1)

There were only a few E pacifica in Layer 2 with

D theta It will be important to know whether the MVBS

of E pacifica at 38 kHz can be neglected compared to that

of D theta The difference in MVBS (DMVBS) between

D theta and E pacifica was estimated using the ratio of E.pacifica to D theta and model-based TS About 72.4-foldmore E pacifica than D theta were caught on August 25(Table1) The TS of D theta and E pacifica were calcu-lated based on known TS-length models for D theta [23]and E pacifica [24], respectively; the predicted TS valueswere -58.0 and -101.8 dB at 38 kHz and -59.0 and-87.9 dB at 120 kHz, respectively At 38 kHz, the MVBS

of E pacifica was still lower than that of D theta by25.2 dB, indicating that the linear MVBS of D theta was333-fold larger than that of E pacifica The effect ofdensity estimation was small The model calculationshowed that the expected DMVBS was -1.1 dB for D.theta, becoming -0.7 dB when E pacifica was contami-nated; this result means that the effect of using the DMVBS

Table 5 Mean and standard deviation of DMVBS for Layers 1 and 2 over the 4 days (in dB)

a Data are presented as the mean ± SD

Fig 8 Example of species

discrimination on August 26

using the estimated mean

volume-backscattering strength

(DMVBS) and irradiance on

August 24, 25, and 27 Upper

horizontal bar MVBS scale

Table 6 Estimated mean

area-backscattering strength at each

layer on August 26 using

DMVBS and irradiance

SA Area-backscattering strength

method was also small Thus, it is not a problem to

con-sider D theta as the acoustically dominant species in Layer

2 throughout the study period

Layer 2 remained at 150–250 m during the daytime and

ascended to 0–70 m at night in the slope area This was

slightly different from the observed depth (300–500 m) of

D theta during the daytime in the Oyashio–Kuroshio

transition waters reported by Watanabe et al [25] The

difference in daytime distribution depths may be caused by

differences in the observation area Layers 1 and 2 underwent

DVM at almost the same time during the 4 days of the study

(Fig.2) These behaviors suggest that D theta fed on

E pacifica when the two layers were in contact and is

sup-ported by a report that D theta actively feeds on euphausiids

and copepods during the day and night, respectively, in the

subarctic area of the western North Pacific [3]

Water temperatures appropriate for D theta are 2–4°C

during the daytime and 6–8°C at night [25]; those for adult

E pacifica are 2–8°C and 7–14°C, respectively, for the

entire water column off Onagawa [5] The CTD

observa-tions in the present study revealed that the water

temper-ature was 12–16°C at 0–20 m, where E pacifica was

distributed, and 7–12°C at 30–70 m, where D theta was

mainly distributed at night on August 24–27 Although E

pacifica was distributed within the reported range of water

temperatures, D theta was found in a higher temperature

range than reported previously

SV threshold level for tracing outlines

The upper and lower outlines of SSLs were traced by hand

on the echogram by setting the minimum SV threshold to

-80 dB To evaluate the validity of this threshold level at

120 kHz, we compared the SA values of Layer 1 calculated

at no SV threshold and those calculated at the minimum SV

threshold of -80 dB The difference between them was

only 2%, indicating that the SV threshold of -80 dB could

be used for tracing the SSL outlines An objective method

that does not rely on hand-tracing would be desirable to

obtain more precise outlines more quickly The width

between the upper and lower outlines of the SSL varies

with the SV threshold value and the acoustic properties of

aquatic animals Although the SV threshold was selected

by trial and error in the present study, a more systematic

procedure is desirable

Examination of the calculated diffuse attenuation

coefficients

The light attenuation coefficient is expressed by the sum of

absorption and scattering by seawater, suspended

sub-stances, and dissolved substances [26] Because the

wavelength characteristics of the attenuation differ with the

structure of the suspended and dissolved substances, thischaracteristic varies by water body The diffuse attenuationcoefficient is an apparent optical property, and it is influ-enced by the angular structure of the underwater light field[27] Namely, the diffuse attenuation coefficient varieswith oceanographic conditions and solar elevation [28].However, the dependence of irradiance on solar elevation

is a minor effect that cannot always be distinguished fromexperimental error in irradiance observations [29] Bakerand Smith [29] found that the diffuse attenuation coeffi-cient for downward irradiance at several wavelengths at thesurface layer varied by \20% throughout the day in aninland impoundment In our study, the diffuse attenuationcoefficients were calculated based on underwater irradi-ance measured at sunset The results of the studies citedabove suggest that it is not a problem to use the samediffuse attenuation coefficient at sunset provided that themeasurement period is short, as in the present study.Mean diffuse attenuation coefficients were used toestimate irradiance at depths deeper than the maximummeasurable depth As the measured diffuse attenuationcoefficients were nearly constant from 20 to 100 m, whichwas the maximum measurable depth in the present study,the estimated diffuse attenuation coefficients will not havediffered from the actual diffuse attenuation coefficients.Estimated preferred irradiance

Many myctophids, including D theta, have similar visualpigments that have maximum visual sensitivity at wave-lengths of 485–490 nm [30], and E pacifica has maximumvisual sensitivity at approximately 462 nm [31] Thus,myctophids react to visual stimuli at these wavelengths.These observations strongly support the results shown inFig.7 Namely, the DVMs of D theta and E pacifica werehighly correlated with the underwater irradiance at about

490 nm With some exceptions, the RMSEs were thelowest and highest at about 490 and 412 nm, respectively(Fig.7) The RMSE of the upper Layer 2 outline at 490 nm

on August 25 was the highest and that of the lower Layer 2outline at 412 nm on August 24 was lower than that at 443,

520, and 565 nm The former was due to the smallernumber of D theta compared to other days and the fact thatthe echo of D theta was dispersed during DVM The lattereffect may have been caused by the traced lower outline ofLayer 1 on August 24, which may not have been appro-priate, or by the lower outline of Layer 1 having a weakrelationship with the isolume

The mean irradiances at the upper and lower limits ofLayer 2 and Layer 1 at 490 nm were -59.3 and -82.5, and-45.3 and -54.0 dB, respectively, during the 4-day studyperiod (Table3) Thus, D theta and E pacifica were dis-tributed over the ranges of 23.2 and 8.7 dB in the

irradiance level In particular, D theta was widely

dis-tributed in the mesopelagic layer compared to E pacifica

Roe [32] reported that mesopelagic euphausiids and fish

are distributed in a wide irradiance range of at least three

orders of light-intensity magnitude Although the targeted

species and the observation methods differ between that

study and ours, ambient light strength differed by [20 dB

between the upper and lower limits of the scattering layer,

where there were still mesopelagic organisms in both

studies

Widder and Frank [19] suggested that individuals

migrating vertically may follow different isolumes due to

differences in internal modifying factors, such as sex and

age, in addition to conflicting drives to optimize foraging

while minimizing predator impact Therefore, while a

hungry individual might push the upper limit of the

iso-lume to have a better chance of spotting prey, a well-fed

individual might remain deeper to reduce the chances of

being spotted by a predator [19] The RMSEs at the upper

outlines of Layer 2 were smaller than those at the lower

outlines over the 4-day period (Fig.7) Although this can

be partially explained by errors in the diffuse attenuation

coefficient, the main reason may be that the influence of

light on the organisms was weaker at the lower limit than at

the upper limit

Species discrimination trial using DMVBS

and irradiance

The mean DMVBS values of Layers 1 and 2 were ?7.7 and

-4.8 dB, respectively, throughout the 4 days (Table4)

The TS–length relationships of D theta [23] and E

pacifica [24] were derived based on a theoretical model

calculation, assuming a tilt angle distribution The

differ-ence in the TS of D theta at 38 and 120 kHz (DTS) was

-1.1 dB for 55.1 mm in SL; in comparison, the DTS of

E pacifica was 13.9 dB for 17.8 mm in TL Our measured

DMVBS for D theta was lower than the above predicted

DTS Sawada et al (unpublished data, 2011) measured the

mean SA of D theta (SL 53.1 ± 7.0 mm) at 38 and

120 kHz during DVM in the North Pacific and found a

mean difference of -5.0 dB Although the DMVBS

pre-diction based on Yasuma’s formula [23] is slightly larger

than the results of our study, our results agree well with the

DMVBS reported by Sawada et al (unpublished, 2011)

The DMVBS values of E pacifica were slightly lower than

the predicted DTS by Matsukura et al [24], which may

have been caused by the actual tilt angle distributions,

which change TS drastically

Although the DMVBS method discriminated well

between E pacifica and D theta, the echoes of D theta

during DVM were not extracted clearly, probably due to

the presence of other species (Fig.8) The irradiance

method worked well with respect to differentiatingbetween the two species and other organisms By usingthe irradiance parameters together with the DMVBS, wewere able to extract the SSL composed mainly of D thetamore clearly The noise and non-migrant layer at150–200 m within the region estimated by the DMVBSmethod were removed by the irradiance method Thespecies composition at 150–200 m is unknown becausenet sampling was not conducted in this layer, but if thedominant species in this layer is D theta, then the irra-diance method underestimated standing stock However,the habitat of D theta clearly differs between day andnight [25] and, therefore, D theta may not be a constantresident of this layer

Finally, using DMVBS and irradiance for species crimination has some limitations The detection range ofthe DMVBS method is generally limited by the higherfrequency that is used, and TS varies by the actual tilt angledistribution, which is usually unknown, which the variation

dis-in TS affectdis-ing DMVBS The mean SSL irradiance must bemeasured at sunset, and the result can be applied only tospecies that undergo DVM However, combining these twomethods is effective for precisely distinguishing amongspecies of fishes and euphausiids, whose behaviors arestrongly affected by light

Future studies should attempt to improve species crimination using a combination of optical and acousticinformation, including the use of a map of attenuationcoefficients and a calibrated photometer

dis-Acknowledgments We express our sincere thanks to Drs H saki, K Anraku, and K Amakasu for their valuable comments and suggestions We also thank Mr T Ogawa and Mr J Shigeno for help with sampling, and the captain and crew of the R/V Wakataka-maru for facilitating sampling onboard the vessel The comments of two reviewers greatly improved this manuscript and were appreciated.

Sugi-References

1 Beebe W, Vander Pyl M (1944) Eastern Pacific expeditions of the New York Zoological Society XXXIII Pacific Myctophidae (Fishes) Zoologica 29:59–95

2 Kawaguchi K (1974) Ecological study on the micronektonic fishes of the western North Pacific Ocean (in Japanese) J Ocea- nogr Soc Jpn 30:312–318

3 Moku M, Kawaguchi K, Watanabe H, Ohno A (2000) Feeding habits of three dominant myctophid fishes, Diaphus theta, Stenobrachius leucopsarus and S nannochir, in the subarctic and transitional waters of the western North Pacific Mar Ecol Prog Ser 207:129–140

4 Yasuma H, Ohshima T, Honda S, Miyashita K, Aoki I (2010) Acoustic biomass estimate of California headlightfish Diaphus theta, off eastern Hokkaido (in Japanese) Japan Nippon Suisan Gakkaishi 76:362–369

5 Taki K, Kotani Y, Endo Y (1996) Ecological studies on hausia pacifica Hansen and seasonal change of its environment off Onagawa, Miyagi Prefecture III Distribution and diel vertical

migration of Euphausia pacifica Bull Tohoku Natl Fish Res Inst

58:89–104 (in Japanese)

6 MacLennan DN, Simmonds EJ (1992) Fisheries acoustics.

Chapman & Hall, London

7 Koslow A, Kloser RJ, Williams A (1997) Pelagic biomass and

community structure over the mid-continental slope off

south-eastern Australia based upon acoustic and midwater trawl

sam-pling Mar Ecol Prog Ser 146:21–35

8 McClatchie S, Dunford A (2003) Estimated biomass of vertically

migrating mesopelagic fish off New Zealand Deep-Sea Res I

50:1263–1281

9 Kang M, Furusawa M, Miyashita K (2002) Effective and accurate

use of difference in mean volume backscattering strength to

identify fish and plankton ICES J Mar Sci 59:794–804

10 Simmonds EJ, Williamson NJ, Gerlotto F, Aglen A (1992)

Acoustic survey design and analysis procedure: a comprehensive

review of current practice International Council for the

Explo-ration of the Sea (ICES) cooperative research report no 187.

ICES, Copenhagen

11 Forward RB Jr (1988) Diel Vertical migration: zooplankton

photobiology and behavior Oceanography 26:361–393

12 Clarke TA (1974) Some Aspects of ecology of stomiatoid fishes

in Pacific Ocean near Hawaii Fish Bull 72:337–351

13 Frost BW, McCrone LE (1979) Vertical distribution, diel vertical

migration, and abundance of some mesopelagic fishes in the

eastern subarctic Pacific Ocean in Summer Fish Bull 76:751–770

14 Huntley M, Brooks ER (1982) Effects of age and food

avail-ability on diel vertical migration of Calanus pacificus Mar Biol

71:23–31

15 Stearns DE (1986) Copepod grazing behavior in simulated

nat-ural light and its relation to nocturnal feeding Mar Ecol Prog Ser

30:65–76

16 Rudjakov JA (1970) Possible causes of diel vertical migrations of

planktonic animals Mar Biol 6:98–105

17 Boscarino BT, Rudstam LG, Eillenberger JL, O’Gorman R

(2009) Importance of light, temperature, zooplankton and fish in

predicting the nighttime vertical distribution of Mysis diluviana.

Aquat Biol 5:263–279

18 Clarke GL (1970) Light conditions in the sea in relation to the

diurnal vertical migrations of animals Maury Center for Ocean

Science Department of the Navy, Washington, D.C., pp 41–50

19 Widder EA, Frank TM (2001) The speed of isolume: a shrimp’s eye view Mar Biol 138:669–677

20 Balino BM, Aksnes DL (1993) Winter distribution and migration

of the sound scattering layers, zooplankton and micronekton in Masfjorden, western Norway Mar Ecol Prog Ser 102:35–50

21 Staby A, Aksnes DL (2011) Follow the light-diurnal and seasonal variations in vertical distribution of the mesopelagic fish Mau- rolicus muelleri Mar Ecol Prog Ser 422:265–273

22 Oozeki Y, Hu F, Kubota H, Sugisaki H, Kimura R (2004) Newly designed quantitative frame trawl for sampling larval and juve- nile pelagic fish Fish Sci 70:223–232

23 Yasuma H, Sawada K, Miyashita K, Aoki I (2008) Swimbladder morphology and target strength of myctophid fishes in the Northwestern Pacific Mar Acous Soc Jpn 35:17–28 (in Japanese)

24 Matsukura R, Mukai T, Ando Y, Iida K (2009) The variation of density and sound speed contrasts and theoretical target strength estimation of Euphausia pacifica Nippon Suisan Gakkaishi 75:38–44 (in Japanese)

25 Watanabe H, Moku M, Kawaguchi K, Ishimaru K, Ohno A (1999) Diel vertical migration of myctophid fishes (Family My- ctophidae) in the transitional waters of the western North Pacific Fish Oceanogr 8:115–127

26 Morinaga T (1986) Ocean studies of turbidity distribution (as measured by the beam attenuation coefficient) and its movement (in Japanese) J Tokyo Univ Fish 73:57–96

27 Tyler JE, Shaules A (1964) Irradiance on flat object underwater Appl Optics 3:105–110

28 Jerlov NG (1976) Marine optics Elsevier Scientific, Amsterdam

29 Baker KS, Smith RC (1979) Quasi-inherent characteristics of the diffuse attenuation coefficient for irradiance Soc Photo-Opt In- strum Eng 208:60–63

30 Douglas RH, Partridge JC, Marshall NJ (1998) The eyes of sea fish I: lens pigmentation, tapeta and visual pigments Prog Retin Eye Res 17:597–636

deep-31 Kampa EM (1955) Euphausiopsin, a new photosensitive pigment from the eyes of euphausiid crustaceans Nature 175:996–998

32 Roe HSJ (1983) Vertical distributions of euphausiids and fish in relation to light-intensity in the Northeastern Atlantic Mar Biol 77:287–298

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

Reproductive characteristics of the orbiculate cardinalfish

Sphaeramia orbicularis in the Chuuk Lagoon, Micronesia

Young-Ung Choi•Heung-Sik Park•

Soo-Jin Heo

Received: 17 May 2011 / Accepted: 15 December 2011 / Published online: 19 February 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract Detailed knowledge of the reproductive

char-acteristics of fish in tropical areas is currently lacking We

have performed a histological study of the reproductive

aspects of Sphaeramia orbicularis, a tropical fish found on

the Weno Islands of the Chuuk Lagoon, Micronesia, and

further estimated the gonadosomatic indices (GSI) of this

fish The sex ratio was approximately 1:1 The standard

length of males and females at 50% maturity was 49.0 and

46.9 mm, respectively Histological observation of

S orbicularis ovaries and testes revealed that the fish

exhibits multiple spawner characteristics, including

coex-istence of primary yolk stages, migratory nucleus or mature

stages, and asynchronism of the spermatogenesis process

Monthly variations in the GSI and gonadal maturity stages

demonstrated that reproductive activity occurs throughout

the year The findings in this study provide important

information on the spawning activity and season of S

orbicularis in Micronesia

Keywords Gonadal development  Gonadosomatic

indices Tropical fish

Introduction

In tropical regions, fish generally exhibit multiple

spawn-ing behaviors that include an extended spawnspawn-ing period

[1] The reproductive activity of such fish is influenced by

wind, rainfall, and the subsequent inflow of nutrients from

streams into coastal areas [2, 3] Another environmental

cue, the lunar cycle, is an important factor in the ductive cycle of pelagic and benthic spawning coral reeffish [4,5]

repro-It has been estimated that Apogonidae (Cardinalfish), afamily of bright-colored tropical marine fish, includes 207species in 22 genera These fishes are distributed world-wide, but they are most abundantly found on reefs in Indo-Pacific oceans [6,7] They generally live in shallow water

of the reefs and near-reef habitats in tropical regions,although a few species have been found in temperateregions [8] This family forms a major component of reeffish assemblages in terms of both abundance and diversity[9] It is also known for its notable reproductive behaviors,including mouthbrooding involving oral incubation bymales [10] The typical reproductive activity of cardinal-fish, which is related to the lunar cycle, has been reportedfor some fishes from both tropical and subtropical regions[11–13], and it has been shown that the lunar spawningpatterns are influenced by tidal cycles [6, 11, 13] It hasalso been suggested that the synchrony of the reproductiveactivities of some tropical reef fish is influenced by themoon, which has a beneficial effect on mating opportunity,dispersal of released eggs, and reduction of predation [5,6,

14] However, our understanding of the reproductive terns of such fish with a multiple spawning style over along spawning period in tropical areas is still limited.The orbiculate cardinalfish Sphaeramia orbicularis iswidespread throughout the tropical and subtropical areas ofthe Indo-Pacific oceans This species exhibits the typicalCardinalfish reproductive pattern involving mouthbrooding

pat-by males [15–17] S orbicularis lives mostly in the shallowand sheltered shoreline waters along mangroves in the ChuukLagoon, Micronesia As such, this species is a suitable modelfor examining the reproductive patterns of Cardinalfish intropical areas The reproductive patterns of this fish, including

Y.-U Choi  H.-S Park ( &)  S.-J Heo

Korea Ocean Research and Development Institute (KORDI),

Ansan, P.O Box 29, Seoul 426-744, Republic of Korea

e-mail: hspark@kordi.re.kr

Fish Sci (2012) 78:515–523

DOI 10.1007/s12562-012-0473-9

the courtship period, brood size, and spawning activity,

have been reported by Allen [15] However, no information

exists on the evolution of gonadal development of the

oocytes and testes or on the spawning season of this fish In

the study reported here, we investigated the reproductive

characteristics of S orbicularis in the Chuuk Lagoon

Specifically, we focused on examination of the fish sex ratio

and fish size at sexual maturity, and we investigated the

spawning season by histological examination of the gonads

and the calculation of gonadosomatic indices (GSI) We

also examined the environmental cues of reproductive

activity in S orbicularis

Materials and methods

Sampling and collection of biological data

Each collection of S orbicularis samples included 15–25

individuals, which were captured between the 20th and 30th

day of each month from August 2008 to July 2009 Fish were

caught with hand nets during a 3-h fishing period

(14:00 hours to 17:00 hours) along a shallow shoreline in a

point section of Weno Island (7°270N, 151°540E) of the

Chuuk Lagoon (Fig.1), at depths ranging from 2.7 to 3.2 m

below the surface, depending on the tide The monthly mean

bottom water temperature (at a depth of 2 m below the

sur-face) in the sampling area was investigated by collecting

weekly (every Monday) sea surface temperature

measure-ments with a SEACAT Profiler (SBE 19; Sea-Bird

Elec-tronics, Bellevue, WA) To obtain information on the daily

length and monthly mean of rainfall, data were acquired from

the websites of the Time and Date website (TADWeb:

http://www.timeanddate.com/worldclock/; accessed 1 Oct

2010) and the National Oceanic and Atmospheric

Admin-istration website (NOAAWeb: http://www.prh.noaa.gov/

peac/update.php/; accessed 20 Oct 2010), respectively

Collected fish were anesthetized in a solution of 0.05%2-phenoxyethanol and then preserved in buffered 10%formalin in a plastic bag and plastic packaging The sam-ples were transported to the laboratory for further analysis

In the laboratory, the standard length (SL) of each fish wasmeasured to the nearest 0.1 cm, the body wet weight(W) was measured to the nearest 0.1 g, and the gonad wetweight (GW) was measured to the nearest 0.001 g Allexperiments complied with the laws of the MacronesianMaritime Authority and the Federated States of Microne-sia, according to the Arrangement on Cooperation in theField of Ocean Science and Technology The sex ratios ofthe collected fish per sampling date and per size class wereexamined using a chi-square (v2) test to assess the statis-tical significance of the deviation from the expected pro-portion of males and females (1:1)

Histological observation of gonadsFor the histological examination, gonads were fixed inBouin’s solution, dehydrated in ethanol, embedded inparaffin wax, and sectioned into slices at a thickness of5–7 lm All sections were stained with hematoxylin andeosin following standard histological procedures described

by Garcı´a del Moral [18] The development stages ofoocytes were classified based on the standard criteriashown in Table 1, which have been modified from theoriginal terminology proposed by Yamamoto and Yama-zaki [19] The testes were classified into maturity stagesbased on observation of spermatocytes or spermatozoa,according to a method proposed by Kume et al [20] Fishsize at maturity (SL at which 50% of the fish had becomemature) was estimated from 112 females and 91 malescollected from October 2008 to July 2009 (the spawningseason) Sexually mature individuals were defined as thosewith gonads at mature stages (stages III and IV) The size

at 50% maturity was defined by the cumulative percentage

Fig 1 Location of the

sampling site in the Weno

Islands of the Chuuk Lagoon,

Micronesia Large black dot

Sampling point

of mature individuals per size class from each sex by the

non-linear least square method to a logistic curve [21] A

logistic function was fitted to the fraction of mature fish per

10-mm length interval using logistic regression models

with MicroCal Origin software (OriginLab, Northampton,

MA) The logistic equation is

YSL= 100/1 ? exp{(a – SL)}/b, where YSL is the

cumulative percentage of mature fish at the SL, a is the SL

at 50% maturity, and b is the slope

Spawning season

The spawning period of fish was determined based on

monthly evaluation of the GSI and the maturity stage of

gonads The GSI was calculated using the formula

GSI = GW (g) 9 100/W (g), and the condition factor (CF)

was calculated using the formula CF = (W (g)/SL3) 9 100

[22] The monthly changes in the GSI and CF were

determined for specimens using the maturity scale, which

were established by the size of maturity specimens for each

sex The GSI and CF data were tested for statistical

sig-nificance using one-way analysis of variance (ANOVA)

Statistical analyses were performed with SPSS ver 14.0 for

Windows (SPSS, Chicago, IL)

Results

Sex ratio

A total of 243 specimens were collected, of which 130

(56.2%) were females and 113 (43.7%) were males The

SL of the collected fish ranged from 28 to 86 mm for

females, and from 29 to 92 mm for males The sex ratio

was 1:1.15 (M:F), and v2analysis showed that this actual

ratio did not significantly differ from 1:1 throughout the

sampling period (v2test, P [ 0.05) The sex ratio by size

class revealed that there was a significant predominance of

females in the SL size classes 45–50 mm and 70–75 mm,respectively (v2test, P \ 0.05) (Fig.2)

Gonadal development

In general, ovarian maturation was divided into four stages:Stage I—immature stage (Fig.3a): previtellogenicoocytes were present This stage included the perinucle-olus and cortical alveolus phase Oocyte diameter rangedfrom 19 to 116 lm (n = 346)

Stage II—developing stage (Fig.3b): primary yolk phaseand the secondary yolk were observed in the ovaries.Oocyte diameter ranged from 48 to 146 lm (n = 191).Stage III—ripe and spawning stage: oocytes reached thetertiary yolk, migrating nucleus, and mature phases(Fig.3c, d) oocyte diameter ranged from 129 to 609 lm(n = 257) Various types of oocytes were observed,including postovulatory follicles

Table 1 Histological features of Sphaeramia orbicularis oocytes at different developmental stages

Developmental stage Diameter (lm) Microscopic appearance

Perinucleolus 19–58 Nucleoli were located in the peripheral part of the nucleus

Cortical alveolus 48–72 Cortical alveolus appeared in the peripheral area of the cytoplasm

Secondary yolk 96–146 The oocytes were increasing in size Cortical alveolus spread throughout the

cytoplasm and began to fuse around the nucleus Tertiary yolk 129–191 Yolk globules were greatly increased in size Lipid droplets distributed around the nucleus Migratory nucleus 195–589 Nucleus migrated toward the pole Lipid droplets fused and became larger

Development stages of oocytes were adapted from Yamamoto and Yamazaki [ 19 ]

0 4 8 12 16 20

Stage IV—spent and resting stage: atretic oocytes and

primary yolk oocytes were observed Also of note, the

zona radiate of atretic oocytes broke down and

disap-peared during this stage (Fig.3e, f)

Similarly, testicular maturation was divided into four

stages:

Stage I—immature stage: the seminiferous tubules were

filled with spermatogonia, which coexisted with a few

spermatocytes and spermatids (Fig.4a)

Stage II—developing stage: spermatids and spermatozoa

made up most of the seminiferous tubules (Fig.4b)

Stage III—ripe and spawning stage: the seminiferoustubules were filled with spermatozoa (Fig.4c, d).Stage IV—spent and resting stage: the seminiferoustubules were empty, and residual spermatozoa wereabsorbed in the early stage and filled with spermatogoniaand spermatocytes in the late stage (Fig.4e, f)

Size at maturityThe minimum size (i.e., SL) at maturity was found to be

44 mm for males and 46 mm for females However, the SL

at 50% maturity, based on the cumulative percentage of

Fig 3 Photomicrographs of the developmental stages of the ovaries

of Sphaeramia orbicularis: a immature stage, b developing stage, c,

d ripe and spawning stage, e, f spent and resting stage AO Atretic

oocyte, Ca cortical alveolus, LD lipid droplets, M maturity stage

oocyte, MN migratory nucleus stage oocyte, PF postovulatory follicle, PN perinucleous stage oocyte, PY primary yolk stage oocyte,

SY secondary yolk stage oocyte, TY tertiary yolk stage oocyte, ZR zona radiate Scale bar: 100 lm

mature males and females, was estimated to be 49.0 and

46.9 mm, respectively (Fig.5) Finally, the mean SL of all

fish at full maturity was determined to be 60 and 61 mm for

males and females, respectively

Spawning season

From August 2008 to July 2009, the mean monthly water

temperature fluctuated from 28.3 to 30.6°C The day length

was shortest in December 2008 (11.4 h) and progressively

increased until June 2009, when it was the longest (12.3 h)

The mean monthly rainfall in August 2008 (83.1 mm) and

in March and June 2009 (93.2 and 143.0 mm, respectively)

was lower than those of the other months (Fig.6)

The mean GSI of females began to increase inNovember 2008, subsequently reached a peak in January

2009, remained at a high level between January and March

of 2009, and then decreased rapidly (Fig.7) FromNovember 2008 to March 2009, the GSI was higher in thelast quarter and the first quarter However, it was not sig-nificantly different throughout the year (P [ 0.05) Incontrast, the mean GSI of males did not reveal any seasonalfluctuation (Fig.7)

Similar to the GSI in females, the mean CF of femalesreached a peak in January 2009 and subsequently remained

at a low level from February 2009 (Fig 7) On the otherhand, the CF of males reached a peak in November 2008and was subsequently maintained at a low level from

(e)

RST

ST

SZ (c)

Fig 4 Photomicrographs of the developmental stages of the testes of

S orbicularis a Immature stage, b developing stage, c, d ripe and

spawning stage, e, f, spent and resting stage RST Residual spermatid,

SC spermatocyte, SG spermatogonia, ST spermatid, SZ spermatozoa Scale bar: 200 lm

December 2008 to July 2009 (Fig.7) The CF values of

both sexes were not significantly different throughout the

year (P [ 0.05)

The monthly evolution of the gonadal maturity stages

(GMS) observed in this study is shown in Fig.8 In general, the

ripe and spawning stage of females (stage III) was observed

from October 2008 to July 2009 The spent and resting stage of

females (stage IV) occurred from April to July 2009 In

con-trast, males with developing testes were collected from

Sep-tember 2008 to July 2009, but not in June 2009 The ripe and

spawning stage (stage III) of testes was observed to begin in

October 2008 and to extend until July 2009 The spent and

resting stage of testes (stage IV) was observed in July 2009

Discussion

S orbicularis has been reported to have a principal

spawning peak between the first quarter and the full moon

and a lesser peak between the last quarter and the new

moon [15], leading to the suggestion that this fish has a

multiple spawning strategy during the breeding period

However, our understanding of such a multiple spawning

strategy over a relatively long spawning period is limited

due to a lack of information on fish reproductive biological

aspects To determine such reproductive characteristics, we

describe for the first time an analysis of the GSI and a

histological study of the GMS of S orbicularis collected in

the field during a 1-year period

In this study, the observed sex ratio of our sample,collected from a natural population of fish, was approxi-mately equal to the expected ratio (1:1) However, therewas a predominance of females in two size classes, namely,45–50 mm and 70–75 mm Reay [23] reported that varia-tions in the sex ratio can occur when one of the sexes has aparticular advantage In the Apogonidae family of fish,males may be provided more opportunities to mate withfemales after they participate in cannibalism, especially ifthe operational sex ratio (OSR; the ratio of sexually com-peting males to females ready to mate) is more femalebiased than male biased during the breeding season [24,

25] Okuda [26] reported that the OSR was female biased(i.e., more females than males) in Apogon doederleini,whereas it was male biased (i.e., more males than females)

in Apogon notatus, based on field observations of the ratio

of sexually active males to females at Murote Beach,Shikoku Island of Japan In our, the sex ratio of S orbi-cularis exhibited a female bias in two size classes greater

Fig 5 Cumulative percentages of mature specimens (presented by

fish length) in both male (open circles) and female (filled squares) S.

orbicularis See section Histological observation of gonads for

explanation of formula

26 27 28 29 30 31

(a)

10.5 11.0 11.5 12.0 12.5 13.0

(b)

0 100 200 300 400 500 600

than the size at sexual maturity This result revealed a

similar trend with the bias previously reported for

A doederleini females [26] We suggest that in such

con-ditions involving a female bias in S orbicularis in the

Chuuk Lagoon, it is quite possible that the males of this

population have more opportunities to mate multiple times

during the brooding season

During the spawning stage of S orbicularis, the primary

yolk stages occurred in the fish ovaries, including

coexis-tence of the migratory nucleus and mature stages, and

spermatogenesis was asynchronous This coexistence of

different developmental stages of oocytes and an

asyn-chronous process of spermatogenesis have been reported in

other Apogonidae species, including Apogon lineatus in the

Tokyo Bay of Japan during the spawning season [20] The

gonads of multiple spawners usually coexist with different

developmental stages of oocytes and an asynchronous

process of spermatogenesis during the spawning season

[20,27] Therefore, our results suggest that S orbicularis is

a multiple spawner

In a previous study of fish size at sexual maturity, Mees

et al [17] reported that both females and males had maturesizes (i.e., SL) of [40 mm based on an observation of theexternal shapes of fish in the Gazi Bay of Kenya In anotherstudy that was conducted by rearing the fish in an artificialpool in the Western Caroline Islands Palau Archipelago,the size of fish at maturity (i.e., SL) was considerablysmaller than 60–70 mm [15] Our results indicate a maturefish size (SL, 49.0 mm in males, 46.9 mm in females,respectively) in between those previously reported Theseresults may compensate for the wide range of size at sexualmaturity seen in previous studies

In tropical zones, the reproductive activities of reef fishhave increased due to variations in the tide and moonlight,particularly because the range of seasonal variation intemperature and photoperiod is less [28] Damselfish, such

as Pomacentrus flavicauda, P wardi, Abudefduf saxatilis,and A troschelii (Pomacentridae) have peaks in repro-ductive activity near the new and full moon [29, 30].Serranidae species, such as Epinephelus guttatus,

(a)

2009

a ab

b

ab

ab ab

a

ab ab ab

Fig 7 Monthly variation in the gonadosomatic index (GSI) and the

condition factor (CF) for S orbicularis All values are presented as

the mean ± standard error (SE) Different letters indicate

signifi-cantly different values (P \ 0.05) Moon phases are shown: filled

circles new moon, left-half-filled circles first quarter moon, open

circles full moon, right-half-filled circles last quarter moon

0 20 40 60 80 100

Month

Stage Stage Stage Stage

2008 2009

5 8 5 8 9 8 6 9 7 12 9

MALES

n =

Fig 8 Monthly percentage frequency of the gonadal maturity stages

in female and male S orbicularis Stages: I Immature stage, II developing stage, III ripe and spawning stage, IV spent and resting stage The numbers (n) along the top of the figure represent the sample sizes Of note, there is a lack of histological analysis data for August, and there were no mature females collected in September 2008

E striatus, and Plectropomus leopardus, have been

shown to exhibit migration and aggregation behaviors in

and to the spawning ground within a specific lunar phase

[31, 32] A member of the Apogonidae, Pterapogon

kauderni, shows peak spawning activity during the full

moon and the last quarter [13] We found no significant

monthly variation in the GSI and CF throughout the year

In addition, ripe and spawning stages in the GMS of

males and females were observed throughout the year

As such, our results show that the reproductive activity

of S orbicularis occurs throughout the year in Chuuk

Lagoon, Micronesia We have also defined the difficult

relationship between GSI, CF, GMS, and environmental

factors Allen [15] reported that the reproductive activity

of S orbicularis is closely related to the lunar cycle,

which led us to expect that this type of lunar

cycle-related reproductive activity would also occur in this

study area Further studies should examine lunar-related

environmental changes in reproductive activity to gain a

clear understanding of the reproductive patterns of

S orbicularis in the Chuuk Lagoon, Micronesia

In conclusion, we describe the reproductive

character-istics of S orbicularis in a tropical region Our results

indicate that the spawning strategy of this species includes

multiple spawning over a relatively long spawning period

The proposed characteristics of the reproduction of this

species may be useful as guidelines in future studies of the

reproductive patterns of other Cardinalfish and various

other species in the Chuuk Lagoon

Acknowledgments We thank the staff of the Korea-South Pacific

Ocean Research Center for their help in collecting samples We are

also grateful to the Chuuk State Government for their cooperation in

the study process and their kind support This study was supported by

KORDI (PE98785) and by research grants from the MLTM, Republic

of Korea (PE56642 and PE56560).

References

1 Hilder ML, Pankhurst NM (2003) Evidence that temperature

change cues reproductive developmental in the spiny damselfish,

Acanthochromis polyacanthus Environ Biol Fish 66:187–196

2 Tyler WA, Stanton FG (1995) Potential influence of food

abun-dance on spawning patterns in a damselfish, Abudefduf

abdomi-nalis Bull Mar Sci 57:610–623

3 Srinivasan M, Jones GP (2006) Extended breeding and

recruit-ment periods of fishes on a low latitude coral reef Coral Reefs

25:673–682

4 Robertson DR, Peterson CE, Brawn JD (1990) Lunar

reproduc-tive cycles of benthic-brooding reef fishes: reflection of larval

biology or adult biology? Ecol Monogr 60:311–329

5 Robertson DR (1991) The role of adult biology in the timing of

spawning of tropical reef fishes In: Sale PF (ed) The ecology of

coral reef fishes Academic Press, New York, pp 356–386

6 Thresher RE (1984) Reproduction of reef fishes TFH Publ,

Neptune City

7 Nelson JS (1994) Fishes of the world, 3rd edn Wiley, New York

8 Thacker CE, Roje DM (2009) Phylogeny of cardinalfishes (Teleostei; Gobiiformes: Apogonidae) and the evolution of vis- ceral bioluminescence Mol Phylogenet Evol 52:735–745

9 Bellwood DR, Wainwright PC (2002) The history and raphy of fishes on coral reefs In: Sale PF (ed) Coral reef fishes: dynamics and diversity in a complex ecosystem Academic Press, San Diego, pp 33–55

biogeog-10 Blumer LS (1982) A bibliography and categorization of bony fishes exhibiting parental care Zool J Linn Soc 76:1–22

11 Allen GR (1972) Observation on a commensal relationship between Siphamia fuscolineata (Apogonidae) and the crown- of-thorns starfish Acanthaster planci Copeia 1972:595–597

12 Okuda N, Ohnishi N (2001) Nocturnal hatching timing of mouthbrooding male cardinalfish Apogon niger Ichthyol Res 48:207–212

13 Vagelli AA, Volpedo AV (2004) Reproductive ecology of apogon kauderni, an endemic apogonid from Indonesia with direct development Environ Biol Fish 70:235–245

Pter-14 Johannes RE (1978) Reproductive strategies of coastal marine fishes in tropics Environ Biol Fish 3:65–84

15 Allen GR (1975) The biology and taxonomy of the cardinalfish, Sphaermia orbicularis (Pisces; Apogonidae) J Roy Soc West Aust 58:86–92

16 Shao K-T, Chen J-P (1986) Ten new records of cardinalfishes from Taiwan, with synopsis of the family Apogonidae J Taiwan Mus 39:61–104

17 Mees J, Mwamsojo GU, Wakwabi EO (1998) Aspects of the biology and feeding ecology of the orbiculate cardinalfish Sph- aeramia orbicularis (CUVIER, 1828) (TELEOSTEL: APOG- ONIDAE) in a KENYAN mangrove forest Biol Jaarb Dodonaea 66:134–145

18 Garcı´a del Moral R (1993) Te´cnicas de laboratorio de anatoml´a pato´logica Interamericana-McGraw Hill, Madrid, Spain,

pp 405–425

19 Yamamoto A, Yamazaki F (1961) Rhythm of development in the oocyte of the gold fish, Carassius auratus Bull Fac Fish Hok- kaido Univ 12:93–110

20 Kume G, Yamaguchi A, Aoki I, Taniuchi T (2000) Reproductive biology of the cardinalfish Apogon lineatus in Tokyo Bay, Japan Fish Sci 66:947–954

21 Yoneda M, Futagawa K, Tokimura M, Horikawa Matsuura H, Matsuyama SM (2002) Reproductive cycle, spawning frequency and batch fecundity of the female whitefin jack Kaiwarinus equula in the East China Sea Fish Res 57:297–309

22 Nielson LA, Johnson DL (1983) Fisheries techniques American Fisheries Society, Bethesda

23 Reay PJ (1989) Reproductive tactics: a non-event in aquaculture? In: Potts GW, Wooton MN (eds) Fish reproduction: strategies and tactics Academic Press, London, pp 291–309

24 Kvarnemo C, Ahbnesjo¨ I (1996) The dynamics of operational sex ratios and competition for mates Trends Ecol Evol 11:404–408

25 Okuda N, Yanagisawa Y (1996) Filial cannibalism in a paternal mouthbrooding fish in relation to mate availability Anim Behav 52:307–314

26 Okuda N (2000) Interspecific differences in male cannibalistic behavior between two sympatric cardinalfishes (Pisces: Apog- onidae) J Ethol 18:5–10

27 Hunter JR, Goldberg SR (1980) Spawning incidence and batch fecundity in northern anchovy Engraulis mordax Fish Bull 77:641–652

28 Takemura A, Rahman SMD, Nakamura S, Park YJ, Takano K (2004) Lunar cycles and reproductive activity in reef fishes with particular attention to rabbitfishes Fish Fish 5:317–328

29 Doherty PJ (1983) Diel, lunar and seasonal rhythms in the reproduction of two tropical damselfishes: Pomacentrus flavic- auda and P wardi Mar Biol 75:215–222

30 Foster SA (1987) Diel and lunar patterns of reproduction in the

Caribbean and Pacific sergeant major damselfishes Abudefduf

saxatilis and A troschelii Mar Biol 95:215–222

31 Collin PL, Shapiro DY, Weiler D (1987) Aspects of the

repro-ductive of two groupers, Epinephelus guttatus and E striatus in

the West Indies Bull Mar Sci 40:220–230

32 Ferreira BP (1995) Reproduction of the common coral trout Plectropomus leopardus (Serranidae: Epinephelinae) from the central and northern Great Barrier Reef, Australia Bull Mar Sci 56:653–669

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

Discovery of a spawning area of the common Japanese conger

Conger myriaster along the Kyushu-Palau Ridge in the western

North Pacific

Hiroaki Kurogi• Noritaka Mochioka•Makoto Okazaki•

Masanori Takahashi• Michael J Miller•Katsumi Tsukamoto•

Daisuke Ambe•Satoshi Katayama• Seinen Chow

Received: 21 November 2011 / Accepted: 4 January 2012 / Published online: 23 February 2012

Ó The Author(s) 2012 This article is published with open access at Springerlink.com

Abstract The common Japanese conger Conger

myrias-ter is an important commercial coastal fisheries species in

East Asia, but its spawning area has not been determined

A larval sampling survey was conducted in September

2008 along 136°E between 13°N and 22°N, which roughly

followed the Kyushu-Palau Ridge in the western North

Pacific Twenty larval specimens were confirmed to be

C myriaster using DNA analysis Two were newly hatched

larvae (preleptocephali) 5.8 and 7.8 mm in total length(TL), which were caught at 17°N The 5.8 mm TL larvawas estimated to be 3–4 days after hatching, the youngestpreleptocephalus (i.e., the earliest stage) of this speciesever collected Eighteen other leptocephali were caught at18°N and 21°N, and these ranged from 18.6 to 40.0 mm

TL Based on these collections, we discerned that there is aspawning area of C myriaster in the area along the Kyu-shu-Palau Ridge approximately 380 km south of Okino-torishima Island Similar to the Japanese eel spawning areaalong the West Mariana Ridge, the Kyushu-Palau Ridgemay play an important role as a landmark of the spawningarea The discovery of this offshore spawning area shouldlead us to a better understanding of the recruitmentmechanisms of C myriaster, and help to facilitate futureinternational management efforts

Keywords Conger myriaster Spawning area Preleptocephali  Kyushu-Palau Ridge  North EquatorialCurrent Meso-scale eddies

IntroductionThe common Japanese conger Conger myriaster (alsocalled the whitespotted conger) is an important commercialfish species in the seas around Japan, Korea [1,2] and theEast China Sea [3] It mainly inhabits shallow coastalwaters to the edge of the continental shelf (e.g., the EastChina Sea) in temperate regions [4 6], and it is absent insubtropical areas such as the Ryukyu Islands in southernJapan, where other Conger species are present Becausemature individuals have not been collected from inshorewaters to the continental margin, where many adults arecaught commercially, the spawning area of this species had

H Kurogi  S Katayama  S Chow

Arasaki Marine Biological Station, National Research Institute

of Fisheries Science, Fisheries Research Agency,

Yokosuka, Kanagawa 238-0316, Japan

Present Address:

H Kurogi ( &)  S Chow

Coastal Fisheries and Aquaculture Division, National Research

Institute of Aquaculture, Fisheries Research Agency,

Yokosuka, Kanagawa 238-0316, Japan

e-mail: hkuro@affrc.go.jp

N Mochioka

Faculty of Agriculture, Kyushu University,

Hakozaki, Fukuoka 812-8581, Japan

M Okazaki  M Takahashi  D Ambe

National Research Institute of Fisheries Science, Fisheries

Research Agency, Yokohama, Kanagawa 236-8648, Japan

M J Miller  K Tsukamoto

Atmosphere and Ocean Research Institute, The University

of Tokyo, Kashiwa, Chiba 277-8564, Japan

Present Address:

S Katayama

Faculty of Agriculture, Tohoku University,

Sendai, Miyagi 981-8555, Japan

DOI 10.1007/s12562-012-0468-6

been presumed to occur somewhere in deeper offshore

waters [5 7] However, until recently, C myriaster larvae

(leptocephali) had been only caught in relatively shallow

coastal areas of East Asia (Japan and Korea) [7 15],

including brackish waters of the upper estuaries of rivers

[16] and over the continental shelf in the East China Sea

[17], and most of them were large [80–130 mm in total

length (TL)] in the late leptocephalus or metamorphic

stages

Recently, the distribution of larvae of C myriaster was

studied using molecular genetic analysis of collections in

the East China Sea along the shelf break near the Kuroshio

mainstream [18,19] and south of the Ryukyu Islands [19]

In addition, a likely spawning area in the North Equatorial

Current region was discovered, based on the collection of

six genetically identified preleptocephali at one station and

by the historical collection of possible C myriaster

lepto-cephali in this region that were not genetically identified

[20] This was considered to be in an analogous location to

the offshore spawning area of the American conger eel,

C oceanicus, in the western North Atlantic [20,21] The

offshore distribution of these larvae showed that C

myri-aster spawned in an offshore area far south of its growth

habitats in East Asia, and that its adults and larvae perform

long-distance migrations similar to catadromous eel

spe-cies such as the Japanese eel Anguilla japonica, which has

been found to spawn in a narrow area along the West

Mariana Ridge [22–25] However, further confirmation

beyond those larvae caught at one station is needed to help

determine the location of the spawning area of C myriaster

and identify the possible landmarks used to find the

spawning area

Understanding the precise location of the spawning area

of this species is of critical importance, because landings of

C myriaster in Japanese waters declined from 13,000 to

6,300 metric tons during 1995–2008 (source: Ministry of

Agriculture, Forestry and Fisheries of Japan, Tokyo),

despite the fact that various stock management approaches

were implemented during this period, such as a reduction

in fishing efforts targeting smaller individuals in major

local fishing grounds in Japanese coastal areas [26, 27]

The landings in Korea declined from 20,000 to 8,000 tons

during 1985–2001, and then rapidly increased to nearly

20,000 tons (source: FAO, Rome) These resource

fluctu-ations and the lack of statistical information about the East

China Sea landings of C myriaster in China are triggering

concern about the decline of this important fisheries

spe-cies In order to implement an efficient and stable stock

management plan, the spawning area, recruitment

mecha-nisms, and resource structure of C myriaster throughout

East Asia needs to be much better understood

In the present study, we describe the collection of

genetically identified larvae of C myriaster in relation to

seafloor structure and ocean current flow in the survey areawhich indicate the presence of a spawning area of thisspecies in this particular area along the Kyushu-PalauRidge, and we discuss its larval migration toward EastAsia, which could lead to further research on the C myri-aster recruitment mechanism

Materials and methodsResearch cruise and collection of C myriaster larvaeThe C myriaster collections were made as a part of theexpedition to study the spawning ecology of the Japaneseeel that was conducted by the R/V Kaiyo Maru of theFisheries Agency of Japan (KY-08-4, 20 August to 15September 2008) [24] along a 12°N, 135.5°E to 13°N,136°E line and a 136°E line between 13°N and 22°N thatroughly followed the Kyushu-Palau Ridge in the NorthPacific (Fig 1) from 5 to 8 September 2008 Larval col-lections were conducted at ten stations at 1°N latitudinalintervals (Fig.1) by a 3 m Isaacs–Kidd midwater trawl(IKMT), which had a 0.5 mm mesh and a mouth opening

of 8.7 m2 Tows of the IKMT fished from a depth of 300 m

to the surface with step towing ranging from 150 to 50 m indepth, and took approximately 60 min (including the

45 min step towing) during both day and night graphic observations consisting of conductivity, tempera-ture, and depth measurements (CTD) up to 1000 m werecarried out at the ten stations of the 136°E line using aSBE-911plus sensor (Sea-Bird Electronics, USA) Multi-level current velocities were observed along the ship track

Hydro-by a shipboard acoustic Doppler current profiler (ADCP);

38 kHz narrow band (RD Instruments, USA) The velocitydata were horizontally gridded into 0.05° boxes, where thevelocities included were ensemble averaged

Identification of Conger myriaster larvaeAnguilliform larvae (leptocephali and preleptocephali)were sorted out of the plankton samples on board andmeasured to the nearest 0.1 mm in total length (TL) Thetotal number of myomeres (TM) and the position of the lastvertical blood vessel on the myomeres (LVBV) wererecorded using a binocular microscope unless the specimenwas too damaged Conger leptocephali were morphologi-cally identified to the genus level based on body shape,pigmentation pattern, and/or number of TM (i.e., 138–149)and LVBV (i.e., 50–59) [28, 29], and then preserved in99% ethanol In the laboratory, the samples of Congerleptocephali and preleptocephali were identified to species

by analyzing nucleotide sequences of approximately 550

base pairs (bp) containing the mitochondrial DNA 16S

rRNA region using the primers 50-GGTCCWRCCTGCCC

AGTGA-30 and 50-CCGGTCTGRACYAGATCACGT-30

There were sequence differences of 3–15% in the

frag-ments among each of the four Conger species present in the

western north Pacific (C myriaster, C japonicus, C

ere-bennus, C cinereus [4]), and a maximum difference of 1%

within the same Conger species [18], that were compared

to the known sequences of C myriaster [18,19,30]

ResultsCollection of Conger myriaster leptocephaliand preleptocephali

A total of 20 Conger myriaster larvae were caught at threestations at 17°N, 18°N, and 21°N along the 136°E line(Fig.1; Table 1) from 6 to 7 September 2008, and thesewere confirmed to be C myriaster by analyzing mito-chondrial DNA fragments of approximately 550 bps con-taining the 16S ribosomal RNA region in the laboratory.Comparisons of the sequences of these 20 specimens(DDBJ/EMBL/GenBank accession numbers AB617683–AB617702) with those of known C myriaster specimensfound sequence identities of 99.2–100% The two that werecollected at 17°N (Figs.1, 2) were 5.8 and 7.8 mm TL(Table1) and were newly hatched larvae (preleptocephali).The 5.8 mm TL larva had no teeth or jaws, and had anearly pigmented eye with an oil globule (Fig.2a), whichwas the youngest preleptocephalus (i.e., earliest stage) ofthis species ever collected The other 18 larvae collected at18°N and 21°N ranging from 18.6 to 40.0 mm TL hadmyomere counts that ranged from 141 to 149 TM and 51 to

57 LVBV (Table1), and had an elongate body, pigmentspots on the gut, and a crescentic patch of pigment beneaththe eye, which are typical morphological features of allknown species of Conger leptocephali [18,19,28,29].Oceanographic features

Current flow along the 136°E lineThe shipboard ADCP observations near the surface at 40 m(Fig.3a) and in the subsurface at 112 m (Fig.3b) along theship track of the transect from 12°N, 135.5°E to 13°N,136°E and the 136°E line between 13 and 22°N (Fig.3)showed alternating east- and west-flowing currents andcountercurrents or eddies, which indicated a mostly con-tinuous steady westward current (the North EquatorialCurrent, NEC) between 12°N and 14°N, a strong eastwardcurrent at around 17°N, and an eastward current (possiblythe Subtropical Countercurrent, STCC) between 21°N and22°N (Fig 3) The current at 17°N, 136°E, where newlyhatched larvae were caught (Fig.1), was flowing eastward

or northeastward at about 0.45–0.24 m/s near the surfaceand in the subsurface layer (Fig.3)

Hydrographic structure at 17°N, 136°EFigure4 shows the CTD profile down to a mean depth of

1000 m at 17°N, 136°E, where newly hatched cephali were caught Water temperature at the netdeployment depths (0–300 m) ranged from approximately

prelepto-Fig 1 Survey area along the Kyushu-Palau ridge on the 136°E line

between 13°N and 22°N in September 2008 Closed circles indicate

the sampling stations at which an Isaacs-Kidd midwater trawl (IKMT)

net survey and conductivity–temperature–depth (CTD) observations

were conducted Stars with circles indicate stations where Conger

myriaster larvae were collected The black square shows the station

where preleptocephali were collected in June 2008 [ 20 ]

16 to 30°C, and within the 50–150 m depth layer where the

step tow was conducted, the temperature ranged from 23 to

29°C The salinity profile indicated that the survey areas

were in the North Pacific Tropical Water (NPTW), which

showed a salinity maximum ([35.0 PSU) in the subsurface

layer at a depth of around 175 m The density drastically

increased from rt = 21–24.5 in the 50–200 m depth layerassociated with pycnocline

DiscussionSpawning area of Conger myriaster along the Kyushu-Palau Ridge

The collection of newly hatched larvae of Conger aster as well as larger specimens in the present studyindicates that this species had spawned in the open oceannear the survey area around 17°N, 136°E, approximately

myri-205 nautical miles (380 km) south of OkinotorishimaIsland (20°250N, 136°040E), in early September 2008 and

in previous months The location at which these cephali were collected in September (17°N, 136°E) wasremarkably close to the location (16°N, 137°E) at whichsix preleptocephali had been collected several monthsearlier, in June 2008, and these were also geneticallyconfirmed to be C myriaster [20] The presence of largergenetically identified larvae to the north of this area in thepresent study and the inferred presence of larger lepto-cephali of this species in July of 1991 and 1995 that couldnot be genetically identified [20] are strong evidence that

prelepto-C myriaster spawns in this particular area of the western

Table 1 List of Conger myriaster preleptocephali and leptocephali collected along the 136°E transect

Pre preleptocephalus stage, Lepto leptocephalus stage, PAM number of pre-anal myomeres, LVBV the position of the last vertical blood vessel

Fig 2 Newly hatched larvae of Conger myriaster collected at 17°N,

136°E a Early-stage larvae: 5.8 mm in total length with no teeth or

jaws, and early eye pigmentation with an oil globule; b 7.8 mm in

total length with teeth, jaws, and eye pigmentation Scale bars are 1

mm

North Pacific Combining these data, it can also be

con-cluded that C myriaster spawns in this area between at

least June and September, which is several months earlier

than previous estimations of the spawning season obtained

by back-calculation based on otolith daily increments of

large leptocephali of this species, which peak from about

September to February [11,17]

In the present study, the preleptocephali had only

recently hatched, so they must have been collected

rela-tively close to their spawning location The smallest larva

(5.8 mm TL), which had no teeth or jaws, and early

pig-mented eyes with an oil globule, was similar to the

Anguilla japonica preleptocephali estimated to be 3–4 days

after hatching based on their eye pigmentation state and

otolith analyses [22] They were also similar in shape and

size to the preleptocephali collected in this area in June

2008 (5.6–6.9 mm), which mostly had pigmented eyes and

early teeth [20] Reared larvae of Conger myriaster from

artificial fertilization were reported to develop such thatthey reached about 5.8 mm TL at 4–5 days after hatchingand mouth opening was observed at 7 days after hatching[31]; thus, the smallest larva (5.8 mm TL) in the presentcollection was equivalent to 4–6 days after hatching of thereared larvae of C myriaster However, their rearingtemperature was significantly lower (12–14°C) than thetemperature profile of the survey area (approximately16–30°C in the depth range of net deployment—the upper

300 m) Within the 50–150 m depth layer, where the steptow was conducted and the larvae were likely captured, thetemperatures ranged from 23 to 29°C Therefore, it ispossible that the age in days of the smallest larva (5.8 mmTL) was about 3–4 days after hatching and that it hadexperienced rapid growth caused by the warmer watertemperature environment

The easterly current in which the newly hatched larvaewere found (17°N, 136°E) was flowing at about 0.24–0.45

Fig 3 Acoustic Doppler

current profiler (ADCP) current

vectors at depths of 40 m

(a) and 112 m (b) along the line

from 12°N, 135.5°E to 13°N,

136°E and the 136°E line

between 13°N and 22°N, which

were obtained from 4 to 7

September 2008

m/s (Fig.3), so the smallest larva (estimated to be

3–4 days after hatching) could have been transported,

depending on its depth, 62–156 km after hatching in the

western area near the longitudinal axis of the Kyushu-Palau

Ridge around 135°E (Fig.3) This suggests that C

myri-aster had spawned in September and probably also in June

quite near to or just to the east of the Kyushu-Palau Ridge

This oceanic ridge could play an important role as a

landmark for the spawning area, as appears to be the case

for the Japanese eel that spawns just to the west of the West

Mariana Ridge [22–25] Since the oceanic ridge crest

within the present survey area (13–22°N) is quite deep

(approximately 2,500–3,500 m deep), except for

Okinoto-rishima Island, C myriaster spawning appears to occur in a

pelagic environment similar to that of the Japanese eel,

which has been studied through the collection of mature

adults [23–25], naturally spawned eggs, and newly hatched

preleptocephali [25] along the West Mariana Ridge The

Kyushu-Palau Ridge is at the boundary between the

Philippine Basin (5,000–6,000 m deep) and the WestMariana Basin (approximately 4,000–5,000 m deep), andmay provide longitudinal cues for C myriaster adults toaggregate and spawn due to possible geomagnetic anom-alies if they have a magnetic sense like anguillid eels [25,

32] Other cues such as olfactory cues could also be sible if they migrate to the spawning area at deeper depthsthan anguillid eels The Japanese eel and other anguillideels migrate at depths of less than about 800 m and showvertical migrations to shallower depths at night [33].However, the maturation of C myriaster appears to occur

pos-at much colder temperpos-atures of about 6°C [34] than thematuration temperatures of about 19–22°C seen for theJapanese eel [35, 36] This suggests the possibility that

C myriaster migrates at deeper depths than the Japaneseeel, and that the Kyushu-Palau Ridge could be a landmarkused by this species to locate the area of aggregation andspawning Based on the molecular phylogenetic relation-ships of anguilliform fishes, conger eels and freshwater eelsevolved through different lineages, even though theirgeneral body morphologies are quite similar, so conger eelsappear to be most related to the shelf and slope eels of theNettastomatidae and may have retained characteristics ofdeep-benthic eels such as the cutthroat eels of theSynahobranchidae, while freshwater eels are most related

to the midwater pelagic deep-sea eels such as the gulpereels of the Eurypharyngidae and the saw-tooth eels of theSerrivomeridae [37]

However, further studies are needed to determine thelarger-scale distribution of spawning and to learn aboutpossible latitudinal cues for spawning, such as fronts [38]

or current flows, which have been suggested to be tant in determining the locations of Japanese eel spawning[25,39]

impor-Laval distribution and transportThe collections of C myriaster larvae were limited to thenorthern part of our survey area (17–21°N), which belongs

to the marginal area adjacent to the northernmost part ofthe NEC (south of approximately 15°N [40]) and coversthe STCC region (north of 19–20°N [40–43]) In contrast,the larvae were absent from the southern region (south of16°N), which belongs to the mainstream of the NEC as astable westward current [40]

Since the spawning area of the Japanese eel was found

to occur along the West Mariana Ridge within the latitudes

of the main part of the NEC (12–14°N) [22–25], with theirlarvae being transported westward by the NEC and thentransferring to the Kuroshio to recruit to East Asia [39,44–

46], it appears that C myriaster has a slightly differentspawning and recruitment strategy It has also been esti-mated that C myriaster larvae are transported westerly

Fig 4 Vertical profiles of water temperature (gray line), salinity

(black line), and density (dashed line) (rt) at 17°N, 136°E, where

newly hatched larvae of C myriaster were caught

toward their nursery areas along the coasts of East Asia

[18–20], but their transport may be more affected by eddies

and countercurrents at more northern latitudes C

myrias-ter larvae (22.3, 46.5 mm TL) were reportedly collected

near the western boundary of the western North Pacific,

south of the Ryukyu Islands (approximately 21°N, 125°E)

[19] However, it is unlikely that westerly larval transport

occurs via the simple, stable, westward flow of the NEC

because of the complex current structure, including the

likely influence of the eastward current of the STCC within

the area of 17–21°N This area is along the southern edge

of a region with much greater sea surface height (SSH)

variability due to the occurrence of many eddies [43,47]

The shipboard ADCP observations along the 136°E

transect showed alternating east- and west-flowing currents

and countercurrents (Fig.3), which may have been

asso-ciated with mesoscale eddies Similar alternating bands of

eastward and westward water flow were also seen in the

Doppler current profile along 137°E in July and August in

a previous study [41] Because both small and large

C myriaster larvae appear to have been collected mostly

north of 16°N in this area previously [20], it appears that

this spawning area of C myriaster may be typically located

just to the north of the main part of the NEC, within a band

of westerly flow that likely exists between two subtropical

fronts associated with eastward countercurrents in this

region [38] Various sizes of C myriaster larvae

(5.8–40.0 mm TL) were collected in the range of 17–21°N

along 136°E in the present study (Table1), indicating

larval retention in the eddies in this region However, SSH

observations show that mesoscale eddies propagate

west-ward through this area [48], which is just along the typical

southern edge of the STCC [43], through the influence of

the baroclinic Rossby waves that propagate westward

across the Pacific basin in 2.5–3 years [43] Therefore, the

C myriaster larvae trapped within the mesoscale eddies

could be transported to the western boundary in several

months, even if they were temporarily entrained into

eastward currents such as the STCC Mesoscale eddies are

also considered to have an influence on the larval migration

route of the Japanese eel [49], and the possible role of the

directional swimming of these migrating Conger and

Anguilla leptocephali to increase their westward transport

is not yet known Further studies on larval transport

sim-ulations using physical oceanographic models for the

western North Pacific subtropical current system are

required to elucidate the recruitment mechanism of

C myriaster, which should lead to international control of

the stock management of this important fisheries resource

Acknowledgments The authors thank Captain Nobuyuki Nagai and

the other crew members of the R/V Kaiyo Maru for supporting the

sample collections; Kyoichi Kawaguchi, Tokio Wada, and Misao

Arimoto for organizing this research cruise; Hiroto Imai, Kohichi Tahara, and Tsuyoshi Koga for their management of the R/V Kaiyo Maru cruises; Jun Aoyama, Akira Shinoda, and Atsushi Tawa for their enthusiastic sampling and sorting onboard; Yoichi Miyake for valuable comments on the oceanographic conditions; and Mayumi Sato for assisting with the DNA analysis We also thank the partici- pants in the Workshop on Conger-Eel Fisheries and Biology (ANA- KEN) for helpful discussions We gratefully acknowledge the financial support of the Fishery Agency of Japan and the Fisheries Research Agency.

Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, dis- tribution, and reproduction in any medium, provided the original author(s) and the source are credited.

3 Tokimura M (2001) Conger-eel fisheries and fisheries resources

in the East China Sea Nippon Suisan Gakkaishi 67:125–126 (in Japanese)

4 Hatooka K (2002) Congridae In: Nakabo T (ed) Fishes of Japan with pictorial keys to the species, English edn Tokai University Press, Tokyo, pp 227–234

5 Ochiai A, Tanaka M (1986) Maanago Conger myriaster In: Ichthyology, 2nd edn Koseisha-Kosei-kaku, Tokyo, pp 576–580 (in Japanese)

6 Yamada U (1986) Conger myriaster In: Okamura O (ed) Fishes

of the East China Sea and the Yellow Sea Seikai Reg Fish Res Lab, Nagasaki, pp 68–69 (in Japanese)

7 Takai T (1959) Studies on the morphology, ecology and culture

of the important apodal fishes, Muraenesox cinereus (Forsska˚l) and Conger myriaster (Brevoort) J Shimonoseki Coll Fish 8:209–555 (in Japanese with English summary)

8 Kubota S (1961) Studies on the ecology, growth and phosis of a conger eel, Conger myriaster (Brevoort) J Fac Fish Pref Univ Mie 5:90–370 (in Japanese with English summary)

metamor-9 Tanaka K, Tabeta O, Mochioka N, Yamada J, Kakuda S (1987) Otolith microstructure and ecology of the conger eel (Conger myriaster) larvae collected in the Seto Inland Sea, Japan Nippon Suisan Gakkaishi 53:543–549

10 Mochioka N (1988) Congridae In: Okiyama M (ed) An atlas of the early stage fishes in Japan Tokai University Press, Tokyo,

pp 44–54 (in Japanese)

11 Lee TW, Byun JS (1996) Microstructural growth in otoliths of conger eel (Conger myriaster) leptocephali during the meta- morphic stage Mar Biol 125:259–268

12 Onuma Y (1996) Metamorphosing location of Conger myriaster leptocephali Bull Fish Exp St Ibaraki-ken 33:103–107 (in Japanese)

13 Otake T, Ishii T, Ishii T, Nakahara M, Nakamura R (1997) Changes in otolith strontium:calcium ratios in metamorphosing Conger myriaster leptocephali Mar Biol 128:565–572

14 Mochioka N (2001) Early life history of Conger myriaster Nippon Suisan Gakkaishi 67:111–112 (in Japanese)

15 Katayama S, Shimizu Y (2006) Occurrence pattern of conger eel larva, Conger myriaster in south Tohoku area Bull Jpn Soc Fish Oceanogr 70:10–15

16 Yagi Y, Kodono N, Kinoshita I, Fujita S (2010) Late-stage

metamorphosing Conger myriaster leptocephali collected in a

river estuary of Ariake Bay, Japan Ichthyol Res 57:310–313

17 Kurogi H, Mochioka N, Takai Y, Tabeta O (2002) First offshore

record of Conger myriaster leptocephali in the East China Sea.

Fish Sci 68:1155–1157

18 Kurogi H (2006) Study on the inshore migration mechanism of

Conger myriaster leptocephali (Ph.D thesis) Kyushu University,

Fukuoka (in Japanese)

19 Ma T, Miller MJ, Aoyama J, Tsukamoto K (2007) Genetic

identification of Conger myriaster leptocephali in the East China

Sea region Fish Sci 73:989–994

20 Miller MJ, Yoshinaga T, Aoyama J, Otake T, Mochioka N,

Kurogi H, Tsukamoto K (2011) Offshore spawning of Conger

myriaster in the western North Pacific: evidence for convergent

migration strategies of anguilliform eels in the Atlantic and

Pacific Naturwissenschaften 98:537–543

21 McCleave JD, Miller MJ (1994) Spawning of Conger oceanicus

and Conger triporiceps (Congridae) in the Sargasso Sea and

subsequent distribution of leptocephali Environ Biol Fish 39:

339–355

22 Tsukamoto K (2006) Spawning of eels near a seamount Nature

439:929

23 Chow S, Kurogi H, Mochioka N, Kaji S, Okazaki M, Tsukamoto

K (2009) Discovery of mature freshwater eels in the open ocean.

Fish Sci 75:257–259

24 Kurogi H, Okazaki M, Mochioka N, Jinbo T, Hashimoto H,

Takahashi M, Tawa A, Aoyama J, Shinoda A, Tsukamoto K,

Hideki T, Gen K, Kazeto Y, Chow S (2011) First capture of

post-spawning female of the Japanese eels Anguilla japonica at the

southern West Mariana Ridge Fish Sci 77:199–205

25 Tsukamoto K, Chow S, Otake T, Kurogi H, Mochioka N, Miller

MJ, Aoyama J, Kimura S, Watanabe S, Yoshinaga T, Shinoda A,

Kuroki M, Oya M, Watanabe T, Hata K, Ijiri S, Kazeto Y,

Nomura K, Tanaka H (2011) Oceanic spawning ecology of

freshwater eels in the western North Pacific Nat Commun 2:179.

doi: 10.1038/ncomms1174

26 Gorie S, Tanda M, Nagasawa K (2010) Movement and growth of

white-spotted conger Conger myriaster in the Eastern Seto Inland

Sea, Japan Aquaculture Sci 58:233–242

27 Gorie S, Nagasawa K (2010) Habitat preference and feeding

habits of juvenile white-spotted conger Conger myriaster in the

Eastern Seto Inland Sea, Japan Aquac Sci 58:167–179

28 Tabeta O, Mochioka N (1988) Leptocephali In: Okiyama M (ed)

An atlas of the early life stage fishes in Japan Tokai Univ Press,

Tokyo, pp 15–64 (in Japanese)

29 Smith DG (1989) Family Congridae: leptocephali In: Bo¨hlke EB

(ed) Fishes of the Western North Atlantic, part 9, vol 2:

Lepto-cephali Sears Foundation for Marine Research, New Haven,

pp 723–763

30 Inoue JG, Miya M, Tsukamoto K, Nishida M (2001) Complete

Mitochondrial DNA sequence of Conger myriaster (Teleostei:

Anguilliformes): novel gene order for vertebrate mitochondrial

genomes and the phylogenetic implications for anguilliform

families J Mol Evol 52:311–320

31 Horie N, Utoh T, Yamada Y, Okamura A, Zhang H, Mikawa N,

Akazawa A, Tanaka S, Oka HP (2002) Development of embryos

and larvae in the common Japanese conger Conger myriaster.

Fish Sci 68:972–983

32 Nishi T, Kawamura G (2005) Anguilla japonica is already

magnetosensitive at the glass eel phase J Fish Biol 67:1213–1224

33 Manabe R, Aoyama J, Watanabe K, Kawai M, Miller MJ, kamoto K (2011) First observations of the oceanic migration of the Japanese eel using pop-up archival transmitting tags Mar Ecol Prog Ser 437:229–240

Tsu-34 Utoh T, Horie N (2011) Studies on the technique of seed duction of the common Japanese conger Conger myriaster Nippon Suisan Gakkaishi 77:775–778 (in Japanese)

pro-35 Ohta H, Kagawa H, Tanaka H, Okuzawa K, Hirose K (1996) Changes in fertilization and hatching rates with time after ovu- lation induced by 17,20b-dihydroxy-4-pregnen-3-one in the Japanese eel, Anguilla japonica Aquaculture 139:291–301

36 Horie N, Utoh T, Mikawa N, Yamada Y, Okamura A, Tanaka S, Tsukamoto K (2008) Influence of artificial fertilization methods

of the hormone-treated Japanese eel Anguilla japonica upon the quality of eggs and larvae (comparison between stripping– insemination and spontaneous spawning methods) Nippon Sui- san Gakkaishi 74:26–35 (in Japanese with English abstract)

37 Inoue JG, Miya M, Miller MJ, Sado T, Hanel R, Lo´pez JA, Hatooka K, Aoyama J, Minegishi Y, Nishida M, Tsukamoto K (2010) Deep-ocean origin of the freshwater eels Biol Lett 6: 363–366

38 Aoki Y, Suga T, Hanawa K (2002) Subsurface subtropical fronts

of the North Pacific as inherent boundaries in the ventilated thermocline J Phys Oceanogr 32:2299–2311

39 Kimura S, Tsukamoto K (2006) The salinity front in the North Equatorial Current: a landmark for the spawning migration of the Japanese eel (Anguilla japonica) related to the stock recruitment Deep Sea Res II 53:315–325

40 Hasunuma K, Yoshida K (1978) Splitting of the subtropical gyre

in the Western North Pacific J Oceanogr Soc Jap 34:160–172

41 Kaneko I, Takatsuki Y, Kamiya H, Kawae S (1998) Water property and current distributions along the WHP-P9 section (137–-142°E) in the western North Pacific J Geophys Res 103: 12959–12984

42 Qiu B, Chen S (2010) Interannual-to-decadal variability in the bifurcation of the North Equatorial current off the Philippines.

J Phys Oceanogr 40:2525–2538

43 Qui B, Chen S (2010) Interannual variability of the North Pacific Subtropical Countercurrent and its associated mesoscale eddy field J Phys Oceanogr 40:213–225

44 Tsukamoto K (1992) Discovery of the spawning area for nese eel Nature 356:789–791

Japa-45 Kimura S, Tsukamoto K, Sugimoto T (1994) A model for the larval migration of the Japanese eel: roles of the trade winds and salinity front Mar Biol 119:185–190

46 Shinoda A, Aoyama J, Miller MJ, Otake T, Mochioka N, Watanabe S, Minegishi Y, Kuroki M, Yoshinaga T, Yokouchi K, Fukuda N, Sudo R, Hagihara S, Zenimoto K, Suzuki Y, Oya M, Inagaki T, Kimura S, Fukui A, Lee TW, Tsukamoto K (2010) Evaluation of the larval distribution and migration of the Japa- nese eel in the western North Pacific Rev Fish Biol Fish doi:

10.1007/s11160-010-9195-1

47 Kobashi F, Kawamura H (2001) Variation of sea surface height at periods of 65–220 days in the subtropical gyre of the North Pacific J Geophys Res 106:26817–26831

48 Qiu B (1999) Seasonal eddy field modulation of the North Pacific subtropical countercurrent: TOPEX/Poseidon observations and theory J Phys Oceanogr 29:2471–2486

49 Kim H, Kimura S, Shinoda A, Kitagawa T, Sasai Y, Sasaki H (2007) Effect of El Nin˜o on migration and larval transport of the Japanese eel (Anguilla japonica) ICES J Mar Sci 64:1387–1395