Fisheries science JSFS , tập 77, số 6, 2011 11

O R I G I N A L A R T I C L E FisheriesHydroacoustic survey of fish density, spatial distribution, and behavior upstream and downstream of the Changzhou Dam on the Pearl River, China Xic

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

Hydroacoustic survey of fish density, spatial distribution,

and behavior upstream and downstream of the Changzhou

Dam on the Pearl River, China

Xichang Tan• Myounghee Kang• Jiangping Tao•

Xinhui Li•Daoming Huang

Received: 9 March 2011 / Accepted: 18 August 2011 / Published online: 22 September 2011

Ó The Japanese Society of Fisheries Science 2011

Abstract Hydroacoustic surveys were conducted to

understand the relationship between fish density, spatial

distribution, and behavior upstream and downstream of the

Changzhou Dam on the Pearl River, China, and the

con-dition (open/closed) of the spillways When the spillways

were open on 24 June 2010, numerous fish were observed

to be densely distributed in the forebay upstream of the

dam, with an average fish density was 0.22 fish m-3 When

the spillways were closed on 25 June 2010, the fish

upstream of the dam dispersed, and the average fish density

decreased to 0.007 fish m-3 Prior to operating the

spill-ways on 24 May 2010, the average fish density downstream

of the dam was 0.28 fish m-3; in comparison, on 26 June,

immediately following closure of the spillways, the

aver-age fish density downstream of the dam was 0.08 fish m-3

Fish were more active on June 24 than on 25 June: they

swam faster and their positions in the water column varied

greatly On 26 June, fish did not to swim as freely in the

water column as those measured on 24 May Based on

these observations, we conclude that a large number of fish

are able to swim to the upstream side of the dam while the

spillways are open

Keywords Hydroacoustic survey Fish density  Fishspatial distribution Fish behavior  Pearl River Changzhou Dam

IntroductionThe Pearl River is the largest river in southern China, withthree major tributaries—the Xi Jiang River, the Bei JiangRiver, and the Dong Jiang River—that all flow into theSouth China Sea The Pearl River is 2,218 km long with450,000 km2 of catchment areas and 3.3 9 1011m3 ofmean annual discharge The Pearl River exhibits a seasonalflow, with a high flow during the summer and a low flow inthe spring and winter The water temperature in the rivervanes between approximately 10 and 35°C, depending onthe season At one time the Pearl River was home to 321fish species, with approximately 208 of these species beingfreshwater and approximately 113 being estuary species.The construction of the Changzhou Dam began in 2003and was completed in 2006 The dam spans the mainstream

of Xi Jiang River, close to Wu Zhou City, Guangxi ZhuangAutonomous Region, China It stretches 3.47 km across the

Xi Jiang River and is 34.6 m high When operating at fullcapacity, it can generate 630 MW from 15 turbines Thedam has two powerhouses and 41 spillways; however, only

33 spillways are currently functional due to operationalproblems During flooding events, when the dischargereaches 22,000 m3s-1, the spillways of the ChangzhouDam are open to release the floodwater The ChangzhouDam is the first dam on the mainstream of the Xi JiangRiver—i.e., it is nearest to the South China Sea Down-stream of the dam is a habitat that is extremely important inthe life cycles of many commercially valuable fish speciesthat are native to the Pearl River However, fish distribution

X Tan  X Li

Pearl River Fishery Research Institute, Chinese Academic of

Fishery Science, 510380 Guangzhou, China

Institute of Hydroecology, Ministry of Water Resources &

Chinese Academy of Science, 430079 Wuhan, China

Fish Sci (2011) 77:891–901

DOI 10.1007/s12562-011-0400-5

and behavior above and below the dam have not been

investigated despite the geographical importance

Avail-able information comes only from interviews of local

fishermen and surveys of their catches [1]

A great concern among ecologists is that the dam

obstructs fish migration Commercially important fish

species, such as black carp Mylopharyngodon piceu, grass

carp Ctenopharyngodon idellus, silver carp

Hypophthal-michthys molitrix, and bighead carp Aristichthys nobilis,

should be able to swim upstream of the Changzhou for

spawning, and their larvae should be able to float

down-stream to the estuary for early development It can be

assumed that fish can swim upward when the spillways are

open during the flood period However, due to having to

take into account the fluctuations of water level upstream

and downstream of the dam, the spillways were open for

only about 10 days per year in the period 2007–2009 This

raises the question of whether such short periods of time

are sufficient to enable the majority of fish to migrate

upward while the spillways are open? There is only one

fish pass and this has not been properly functional since it

was built This fish pass is managed by a private company,

and no regulations are in place to allow the company to run

the fish pass In addition, a recent investigation on fish

migration around the Gezhouba Dam on the Yangtze River

concluded that ship locks can not be used as channels for

fish migration [2] Fish migration toward upstream regions

through spillways is an extremely important issue in

fish-eries science in general and, more specifically, in dam

management in the river basin Therefore, it is exceedingly

important to collect data on fish density and their spatial

distribution as well as their behavior both upstream and

downstream of the dam in the light of the condition of the

spillways Here, the phrase ‘‘condition of the spillways’’

refers to the state of whether the spillways are open or

closed

The development and improvement of scientific

acoustic instruments and new monitoring techniques

during the last decade now enable researchers to monitor

fish migration both efficiently and non-obtrusively [3,4]

Acoustic techniques can also be used in a practical

context, such as for monitoring ecosystems in deep lakes

or reservoirs, where traditional shallow-water netting

techniques (e.g., gill nets, traps net, and seining) are

difficult to employ, enabling a precise characterization of

fish spatial distribution and furthering our understanding

of fish behavior Although hydroacoustic studies have

been performed in the reservoir, these mostly focused on

fish abundance and distribution; little research has been

carried out to date on the influence of the functional

aspects of the dam [4] The results of these studies

suggested that fish behavior at two dams was different

according to the use of the turbine units In China,

hydroacoustic surveys have been widely used for fishresource research in inland waters, such as speciesidentification of the Chinese sturgeon Acipenser sinensis[5], fish distribution in the forebay of the Three GorgeReservoir [6], and spatial and temporal distribution ofthe naked carp Gymnocypris przewalskii in Qinghai Lake[7] However, in the Pearl River, a hydroacoustic surveyhas only been performed as a means to observe thespawning aggregation of black Guangdong bream Meg-alobrama hoffmanni in the two spawning grounds [8].Therefore, it is imperative to apply the hydroacousticmethod for monitoring fishery and fish ecology in thePearl River, especially around the Changzhou Dam Wehave therefore undertaken a hydroacoustic survey withthe aim of acquiring reliable information on the density,spatial distribution, and behavior of fish both upstreamand downstream of the Changzhou Dam We also sam-pled using a gill net

The aim of the study was to gain an understanding of thedensity, spatial distribution, and behavior of fish upstreamand downstream of the Changzhou Dam in relation to thecondition of the spillways in order to provide valuableinformation for (1) the effective management and operation

of spillways in terms of fish migration, (2) scientificassessment of the dam as an obstacle to fish migration and(3) the development of a methodology for estimating fishresources in the Pearl River

Materials and methodsStudy area

The chosen research region upstream of the ChangzhouDam was 4 km in length, 1.7–3 km in width, and had anaverage depth of 12.4 m; the region chosen for studydownstream of the dam was 22 km in length, approxi-mately 1.5 km in width, and had an average depth of14.6 m (Fig.1) A downstream fish conservation zone,established in December 2009, and its core area are indi-cated in Fig.1 Two spillways, with 33 gates, are locatedacross two branches of the river Each gate is 16 m highand 16.6 m wide During the period when the hydroacou-stic survey was being conducted in the downstream region(23–25 May 2010), the amount of discharge was low(11,200 m3s-1) When the second survey was carried out

in both the upstream and downstream of the dam tounderstand the dynamic changes in the fish populationduring 24–26 June 2010, the discharge was relatively high(22,400 m3s-1) The data on water levels upstream anddownstream of the dam were obtained from the Wu ZhouChannel Authority’s website (http://wzhd.gxgh.cn) in theGuangxi Zhuang Autonomous Region

Fish sampling

From 19 to 24 May 2010, 23 gillnettings with various mesh

sizes (5–8 cm) were carried out to sample fish downstream

of the Changzhou Dam All catches were identified for

species, each fish was measured (body length, in

millime-ters) and weighed (grams)

Hydroacoustic surveys

Two hydroacoustic surveys were conducted in two separate

time periods The first survey started downstream of the

area marked D in Fig.1on 23 May, proceeded upstream,

reaching area D on 24 May, and then moved back

down-stream on 25 May The second hydroacoustic survey

star-ted upstream of the area marked U in Fig.1on the first day,

on 24 June The survey then proceeded downstream to area

U in Fig.1 on 25 June, and then was conducted in area

marked D in Fig.1on 26 June A calibrated Simrad EY60

portable echosounder (Horten, Norway) and a circular

split-beam 120 kHz transducer (7° of nominal angle) were

used in the surveys The pulse duration was 64 ls and the

pulse interval was 128 ls The transducer was mounted on

the fore-port side of the ship at the depth of 0.5 m below

the water surface The EY60 echosounder was connected to

a laptop computer running the ER60 software in Windows

XP in order to store data, and the computer was also linked

to a Garmin GPS (Olathe, USA) A 25-m vessel was

uti-lized for the survey, and the cruising speed was

approxi-mately 7–8 knots The total recorded acoustic data was

about 11 GB, which included 537 sections, and each

sec-tion is approximately 300 m in length All survey data were

stored during daylight hours because of the dangerinvolved in navigating at night The acoustic surveys werecarried out following a zigzagging route, and the degree ofcoverage was calculated for each survey using the Aglen’sformula [9] The coverage degree in this study was in therange of 7.1–16.2, which exceeds the level recommended

by Aglen [9]

Data analysisThe raw data files of the EY60 sounder were directly addedinto Echoview (ver 4.90), which is a fisheries acousticsdata analysis software (Myriax, Hobart, Australia) Thevolume backscattering strength (SV) echograms were pre-sented with a 20 log R time varied gain (TVG), where R isthe range to a target from the transducer The targetstrength (TS) echograms were displayed with a 40 log

R TVG The angular position (angle) echograms were alsoshown There was considerable noise at the level of thewater surface due to the movement of the vessel There-fore, a straight line, either 0.8 or 1 m deep, was drawn toexclude any data samples considered as water surface noise

or plankton-like organisms However, some noise echoesappeared up to the middle of the water column Theseechoes were manually selected and defined as bad data,which means that they were removed from further dataanalysis The river bottom line was defined using the Bestbottom candidate algorithm in Echoview [10] and wasmanually edited when it was necessary Accordingly, anydata above the water surface line, below the river bottomline, and regions defined as bad data were eliminated fromthe data analysis

Fig 1 Map of river system in

the upstream and downstream of

the Changzhou Dam in the Pearl

River, China U Upstream study

area, D downstream study area,

x spillway locations, a fish

conservation zone, b center of

the fish conservation zone

To detect fish tracks for evaluating fish density and

behavior in studies like the one presented here, a single

target detection echogram should be created since the fish

track technique should be performed on a single target

detection echogram There are several single target

detec-tion methods in Echoview, each based on different

echo-sounder algorithms Among these, the single target

detection split beam method 2 [11] used by the algorithm

from the Simrad EK60 echosounder was assessed to be

appropriate to detect single targets using split beam data

from the EY60 echosounder The parameters used in single

target detection split beam method 2 were: TS

thresh-old = -70 dB; pulse length determination level =

-6 dB; minimum normalized pulse length = 0.4;

maxi-mum normalized pulse length = 1.4; maximaxi-mum beam

compensation = 10 dB; maximum standard deviation of

minor-axis angles = 0.6; maximum standard deviation of

major-axis angles = 0.6

The fish track technique is generally used to identify

groups of single targets which show a pattern of systematic

movement The targets grouped into a fish track are

assumed to have been generated by a single object moving

through space Echoview’s a–b Fish Tracker algorithm

implements a fixed coefficient filtering method, as

pre-sented in Blackman [12] The parameters used for fish

tracking in this study are shown in Table1 Output from

the fish track technique was exported in the

comma-sepa-rated values format to investigate fish density and fish

behavior The fish tracks were extracted with a focus on

four separate days of acoustic data: 24 June, 25 June, 24

May and 26 June 2010, since the condition of spillways are

unique In the subsection Fish behavior in the Results

section, fish tracks extracted from the acoustic data on 24

June 2010 are called ‘‘24 June fish’’; those from the

acoustic data on 25 June 2010, ‘‘25 June fish’’; those from

data on 24 May, ‘‘24 May fish’’; those from data on 26 June

‘‘26 June fish’’

The spatial distribution of fish upstream and

down-stream of the dam was described using Surfer software

(ver 8; Golden Software, Golden, USA) with a minimumcurvature gridding method

ResultsGillnet sampling

A total of 642 fish were caught by gillnetting, and theidentify of 35 fish species was confirmed The dominantspecies in the catch were barbel chub Squaliobarbus cur-riculus (23.7%), mud carp Cirrhina molitorella (16.9%),black Guangdong bream (12.5%), white amur breamParabramis pekinensis (7.1%), and common carp Cyprinuscarpio (4.9%) These five species made up 82.4% of thetotal biomass of the catch in the gill nets The majority ofthe catch had body lengths ranging from 9.2 to 58 cm.Spillways operation

The difference of water levels between the study regionsupstream and downstream of Changzhou Dam duringJanuary through October 2010 is shown in Fig.2 Fewfloods occurred due to the unusual drought in the spring of

2010 in the Pearl River area Therefore, the spillways wereopened only for 4 days—from 21 to 24 June 2010 Withinthat time frame, information on the density, spatial distri-bution, and behavior of fish were compiled and laterexamined with reference to the condition of the spillways.The acoustic data collected on 24 June was used since thespillways were open on that date For comparison purposes,acoustic data collected on 25 June, when the spillwayswere closed, were also used Both data sets were collectedupstream of the dam Two more acoustic data sets werederived from the study area downstream of the dam on 24May and 26 June, respectively The first of these datesprecedes by a long time the opening date of the spillways,and the second date is immediately after the spillways wereclosed

Table 1 Setting parameters for

detecting fish tracks using

acoustic data from the 200-kHz

echosounder

Target gates Exclusion distance for major and minor axes and range 2 m

Fish density and their spatial distribution

Echogram analysis showed that fish communities chiefly

consisted of many individual fish Figure3 shows an

example of numerous single fish detected as fish tracks

The three-dimensional fish track (Fig.3c) is an example of

the orientation of the fish track in the beam

On 24 June 2010, the average fish density upstream of

the Changzhou Dam was 0.22 fish m-3, with a maximum

density of 0.71 fish m-3 A large number of fish found

upstream were distributed adjacent to the dam When the

spillways of the dam were closed on 25 June 2010, the fish

quickly dispersed, and the average fish density decreased to

0.007 fish m-3, with a maximum value of 0.03 fish m-3

On 24 May 2010, average fish density downstream of the

Changzhou Dam was 0.28 fish m-3, with a maximum

density of 1.57 fish m-3 A large number of fish were

distributed in the fish conservation area On 26 June 2010,

average fish density downstream of the Changzhou Dam

was 0.08 fish m-3, with a maximum density of

0.27 fish m-3 Comparison of the fish densities upstream

and downstream of the dam demonstrated that the

down-stream fish density on 26 June was nearly tenfold higher

than the upstream fish density upstream on 25 June 2010

The summary of the average fish density upstream and

downstream of the dam based on the condition of the

spillways (open/closed) is shown in Table2

The fish communities upstream of the dam were not

evenly distributed, as shown in Fig.4 On 24 June 2010,

during the time the spillways were open, few fish were

observed in an open-water area (black arrow in Fig.4a)

On 25 June 2010, when the spillways were closed, most

fish had shifted (right side of the figure, gray arrow in

Fig.4b) A small number of fish were present in front ofthe spillways In comparison with the fish density on 24June, on 25 June nearly all fish had disappeared during theone night upstream of the dam The difference in distri-butions between 24 and 25 June (as shown in Fig.4a, b)may be due to the condition of the spillways Regarding thefish spatial distribution downstream of the dam, the fishcommunities were not uniformly distributed (Fig.5) On

24 May 2010, a higher density was found between111.31°300E and 111.37°480E along the river, with a greatnumber of fish being detected in the fish conservation zone(Fig.5a) On 26 June 2010, most of the fish were distrib-uted rather evenly downstream, although an area of densedistribution was found further down from the dam(Fig.5b)

Fish behavior

A variety of fish behavioral descriptors, such as TS, depth,depth change, tortuosity, vertical direction, and speed, wereused to understand fish behavior in the study areasupstream and downstream of the dam The behavioraldescriptors are explained in Table3 Here, TS is nottechnically a behavioral descriptor However, it can be afactor of body length since the body length of fish can beestimated from TS and two other species-specific constants

Fig 2 Difference in water levels between the study areas upstream

and downstream of Changzhou Dam between January and October

2010

Fig 3 Fish tracks in the Changzhou Dam on 25 June 2010 a SVechogram, b single target detection echogram, c three-dimensional (3D) single targets in the beam from a fish track circled in black in

b The region on the top of SVand single target echograms is defined

as a bad region and excluded from further analysis The line at the bottom of the figure indicates the river bottom Note that weak signals

in the SVechogram do not appear on the single target echogram

[13] Therefore, TS is also included in these descriptors.

Fish behavior was compared based on the condition of the

spillways (open/closed), such as fish behavior on 24 June

versus that on 25 June fish, and fish behavior on 24 May

versus that on 26 June (Figs.6,7) Figure 6compares fishbehavior on 24 June and 25 June in a box plot The TS ofthe 24 June fish was larger than that of the 25 June fish Forexample, 50% of the 24 June fish (the first quartile throughthe third quartile) had the TS range of -60.0 and -51.1 dB, although that of the 25 June fish had a range of -66.6 to –60.3 dB The depth of the 24 June fish was slightlyshallower than that of the 25 June fish; however, themaximum depth of the 24 June fish was deeper than that ofthe 25 June fish The depth change of the 24 June fish wassomewhat larger than that of the 25 June fish These resultscould lead to the interpretation that the 24 June fish swammore dynamically in a vertical direction The tortuosity ofthe 24 June fish school was marginally larger than that ofthe 25 June fish school The middle half of the 24 June fishhad a tortuosity range of 1.03–1.6 and that of the 25 Junefish had a range of 1.01–1.28 In terms of vertical direction,

up to the third quartile of the 24 June fish had a minus sign(i.e., -1.44), which meant that 75% of the fish headeddown In comparison, half of the 25 June fish swam in amixed direction (some fish heading up and other fishheading down) Fifty percent of the 24 June fish swamapproximately 1 m s-1 faster than the 25 June fish Tosummarize, in comparison to the 25 June fish, the 24 Junefish had larger TS, were distributed at a shallower waterdepth, showed relatively large changes in water depth,were swimming more dynamically, were heading slightly

Table 2 Average fish density upstream and downstream of the Changzhou Dam based on the condition of the spillways (open/closed)

Fig 4 Fish spatial distribution upstream of Changzhou Dam on 24

June 2010 when the dam spillways were open (a) and on 25 June

2010 when the spillways were closed (b) This location is the same as

that marked U in Fig 1 Note that different density (fish m-3) scales

are used The vertical bar in dark gray on the right side of b indicates the dam Arrows indicate the area where the spillways are being operated

Fig 5 Fish spatial distribution downstream of Changzhou Dam on

24 May 2010, which is prior to the opening of the spillways (a), and

on 26 June 2010, which is after the spillways have been closed (b).

This location is as that marked D in Fig 1 Note that different density

(fish m -3 ) scales are used

down, and were swimming relatively fast It can be

pre-sumed that the reason why 24 June fish were more actively

moving around compared to 25 June fish was that a great

number of fish had migrated from the downstream area

while the spillways were open Therefore, the fish were

acting more vigorously because the fish which were

already upstream would be stimulated by the presence of

new fish and fish newly arrived from the downstream area

would be in a new environment

A comparison of fish behavior on 24 May and 26 Juneare shown in Fig.7 Half of the 24 May fish had a TS range

of –66.70 and –60.79 dB, and the 26 June fish had a TSrange of –65.86 to –58.85 dB Fifty percent of the 24 Mayfish were distributed at a depth approximately 1.9 m shal-lower than that of the 26 June fish Both groups of fish didnot show a significant depth change, although the range ofthe depth change in the 26 June fish was larger than that inthe 25 May fish The tortuosity of the 24 May fish was

Table 3 Delineation of each

behavioral descriptor

TS, Target strength

Name (unit) Definition

TS (dB re 1 m-2) Mean TS value of the targets in the track Depth (m) Average depth over all targets in the track Depth change (m) The depth of the first target minus the depth of the last target in a track The depth

change with a plus sign means that the fish swims toward the water surface, while the depth change with a minus sign shows that the fish swims toward the river bottom

Tortuosity The sum of the distances between adjacent targets in a track (that is, the total distance

traveled) divided by the straight line distance between the first and last targets in a track

Vertical direction It is calculated from a line drawn from the first to the last target in a track The

vertical direction of 0° describes a direction parallel to the X–Y (horizontal) plane (i.e., no vertical direction), -90° describes a downward direction, and 90° an upward direction

Speed (m s -1 ) The accumulated distance between targets divided by the total time

Fig 6 Box plot of various

behavioral descriptors of the 24

June fish and 25 June fish Fifty

percent of the samples (box) are

within the first quartile (bottom

of the box) and the third quartile

(top of the box) The minimum

and maximum values (bars) are

shown However, the upper

inner fence is replaced with the

maximum value for target

strength (TS), tortuosity, and

speed Each descriptor is

precisely explained in Table 3

slightly larger than that of the 26 June fish; for example, the

third quartile was 1.94 compared to the 1.31 of the 26 June

fish The 24 May fish had a tendency to swim upwards

since the first through to the third quartile of the vertical

direction had all plus signs, that is 2.6–10.1° On the other

hand, the middle half of the 26 June fish had a narrow

range of -2.6 and 1.4° in the first and third quartiles The

swimming speed of the 26 June fish was approximately

1.8 m s-1faster than that of the 24 May fish In summary,

compared to the 24 May fish, the 26 June fish had larger TS

values, they were distributed at a deeper depth, their

movement did not vary along the depth axis, and they

swam relatively straighter and faster Based on these

results, it can be assumed that after the spillways had

closed, the 26 June fish preferred to stay at a one level in

relatively deep water

Discussion

In general, spillways operate to maximize the storage of

water upstream The overriding rule when operating

spill-ways is that the safety of the dam is paramount, and the rate

of outflow must not exceed the rate of inflow during the

rising stage of an incoming flood [14] Little consideration

given to the fish community during the operation of ways In this study, density, spatial distribution, and thebehavior of fish communities upstream and downstream ofthe Changzhou Dam were discovered to have altered,depending on the operation of spillways In particular, thedifference in fish density owing to the operation of spill-ways demonstrated that the dam—or the poor operation ofthe spillways—created a discontinuity that shifts fishcommunities from downstream to upstream An nonfunc-tional fish pass also contributed to the dam being anobstruction for fish migration The fish pass must operate asdesigned and remain fully operational at all times Ablocked or damaged fish pass is simply a waste of preciousresources

spill-Ours observations demonstrated that most fish arrivingupstream of the dam between 21 and 24 June remained infront of the spillways after having swum through them.Based on the fish density difference between 24 and 25June, the number of fish that swam upstream can be esti-mated to be about 4.5 million fish This implies that 43 fishmigrated through the spillways every second during theperiod 21–24 June 2010 Using similar reasoning, on 24May 2010, the fish biomass can be loosely estimated as134.9 million fish, indicating that only 10.7% of fishdownstream of the dam were able to migrate up through the

Fig 7 Box plot of various

behavioral descriptors of the 24

May fish and 26 June fish Fifty

percent of the samples (box) are

within the first quartile (bottom

of the box) and the third quartile

(top of the box) The minimum

and maximum values (bars) are

shown However, the upper

inner fence is replaced with the

maximum value for TS,

tortuosity, and speed Each

descriptor is precisely explained

in Table 3

spillways during the flood period These estimations

sug-gest that the operating days and the operating frequency of

the spillways must be increased for the purpose of fish

resource protection and ecology management, since only

one operation per year may not be adequate for fish

migration in the Pearl River

A detailed understanding of fish behavior in dam

envi-ronments is of great importance; for example, mechanical

structures designed to divert or guide juvenile migrant

salmon at dams often work less effectively than expected

This likely results from a lack of understanding of fish

behavior [15] Biologically substantiated techniques for

studying fish behavior in the dam environment require a

deeper knowledge of fish behavior in a broad ecological

context For example, the most appropriate selection

method for a specific location, such as a fish conservation

area, should take into account the spatial fish distribution

and fish behavior This report is the first to provide

infor-mation on fish behavior using the acoustic method Little

information on fish behaviors in the study area is currently

available Therefore, it is nearly impossible to directly

compare our results with other research However, Our

study may contribute valuable information on the

com-plexity of the ecological systems for sustainable dam

management and initiate further study We found the

spa-tial fish distribution downstream of the dam to be within

the range of the fish conservation region However, the

coverage of survey area should be extended farther

downstream in order for ecologists to acquire a larger

picture of spatial distribution throughout the entire

down-stream area of the Pearl River For mobile organisms such

as fish, micro- (response to local hydrodynamic conditions)

and macro- (life-cycle features) movement and migrations

are critical for maintaining viable populations [16–18] The

life cycle of each and every fish species can be considered

to be a sequence of residential and migratory phases Thus,

the operation of spillways should consider the spatial and

temporal patterns of fish distribution and dispersal and

migratory activity in terms of fish behavior Our results

show that the key to successfully operating the spillways

and the fish pass lies in a better understanding of fish

behavior and their response to hydrodynamic conditions at

obstructions Although our results only show fish

micro-movement around the dam, they can be used as a

foun-dation for the comprehensive study on macro-movement

and migrations Ongoing monitoring should be mandatory

to increase knowledge on how to improve the ecological

status of the regions upstream and downstream of the dam

Therefore, information (fish density, spatial distribution

and behavior) from monitoring should be collected for a

relatively long time period during regular time intervals

Two methods, namely, catch statistics and fish larvae

sampling, have been used to estimate fish resources in the

Pearl River However, these may not provide quantitativelyrepresentative results from fish communities in a targetedarea We have demonstrated that a hydroacoustic surveycan provide valuable information on fish density, fishspatial distribution, and fish behavior upstream anddownstream of the Changzhou Dam This kind of infor-mation is very difficult to obtain using conventionalmethods Accordingly, the hydroacoustic method provides

a simple, yet reliable way to assess the impact of the dam

on fish migration This method can also be used to estimatefish resources relatively quickly and easily This study wasthe first trial of the hydroacoustic method to estimate fishresources in the Pearl River and to demonstrate the effect

of operating spillways on fish migration However, themethod does not directly provide data on species compo-sition, which requires a ground-truth method In our study,gillnetting was utilized as the ground-truth method Thestandardized sampling of fish in the majority of freshwaterenvironments is done with gillnets [19] However, gill-netting can only be performed easily in pelagic areas; thus,results from gillnetting most likely are not representative ofthe entire fish community of that area In freshwater,fishery science trawling is even less frequently applied as it

is associated with a relatively high sampling effort andrequires sophisticated equipment Trawl net sampling hasless selectivity than gillnet sampling, which means that theformer method can be utilized to catch representative fishsamples more reliably than the latter method Thus, com-bined surveys of hydroacoustics and trawling methods cancomplement each other by balancing their individualdrawbacks; the hydroacoustics method provides data onabsolute fish density and the trawling method provides data

on species composition and length distribution In species circumstances, where information on speciescontributions and detailed size spectra are needed, thecombination of hydroacoustics with trawl catches is thefavored approach, although manpower and finances remainmajor constraints The Pearl River has been seriously over-fished, and fish communities require protection, especiallywith respect to spawning and growth [20] Certain fishspecies, such as the black carp, grass carp, silver carp, andbighead carp, urgently need to migrate upstream to finishspawning (unpublished information provided by fisheriesscientists) Hereby, the enhanced reliability of the combi-nation approach can provide essential data on Pearl Riverpopulations that will enable researchers to obtain an eco-logical understanding of these species and, consequently,

multi-to protect them This combined method is flexible enough

to be conducted with other observation and experimentalsystems In the distant future it would be ideal to take intoaccount complicated factors such as biological (foodavailability, predation risk and competition), environmen-tal (oxygen concentration and water temperature) and

physical (water turbulence, visual and hydrodynamic

hab-itat heterogeneity) information when the methodology of

fish resource estimation and dam obstruction assessment

has been developed and enhanced [15–18,21,22]

In comparison, horizontal beaming is considered to be a

crucial component of acoustic fish stock assessment in

freshwater lakes and reservoirs [23, 24] A conventional

hydroacoustic method is limited by the low volume of

water sampled near the vertically aimed transducer, which

can be problematic when the aim is to detect fish located

near a surface or at a relatively shallow water depth Thus,

to sample the upper layers of the water column or the entire

layers in relatively shallow waters, horizontal beaming has

been employed in freshwater lakes and reservoirs During

our survey, the hydrological characteristics upstream of the

Changzhou dam were similar to those downstream of the

dam since there was no change of the water column

structure related to temperature during this period The

sound speed and absorption coefficient, which are directly

associated with acoustical strengths (SV and TS), are

greatly affected by water temperature [25, 26]

Accord-ingly, a uniform temperature in a horizontally beamed area

ascertains that these given values in the acoustic data

set-ting are valid in both streams Therefore, these sites would

be suitable for the horizontal beaming technique The

technique can be used to understand fish density and their

behavior along the length of the river and would enable the

accurate monitoring of fish passing through the spillways

In general the estimation of fish length, even roughly, is

beneficial to any understanding of an aquatic ecosystem

Fish length from TS data can be calculated based on an

equation that expresses the relationship [13] In the

multi-species situation, one of the most widely used equations for

the TS–fish length relationship is that of Love [27], which

is TS = 19.4 Log(L) - 0.9 Log(f) - 63.7, where L is the

body length and f is frequency Using the above equation, a

mean TS of -55.6 dB was observed on 24 June upstream

of the dam and -61.6 dB was observed on 26 June

downstream of the dam; these data correspond to a fish

length of 3.3 and 1.6 cm, respectively However, this is not

a realistic result because fish of such a small size have not

been caught in the river Frouzova et al [28] defined a

mean TS–fish length equation in an ex situ experiment

using pooled data on five species (trout Salmo trutta, perch

Perca fluviatilis, bream Abramis brama, roach Rutilus

ru-tilus, common carp, bleak Alburnus alburnus) Their

equation is TS = 20.45 Log(L) - 96.13, where L is

the body length Using this equation, a TS of -55.6 and

-61.6 dB corresponds to a fish length 28.4 and 13.0 cm,

respectively Such values are closer to the length of the fish

sampled from the gillnetting It is important that the TS–

fish length equation for fish, particularly Chinese fish

species, be known Hence, a plan for future research on the

Pearl River is to measure in situ TS for dominant speciesand to use a ground-truth method to obtain a better TS–fishlength relationship

It is worth stating that hydraulic engineers and dammanagers need to work closely with fisheries scientists inthe common aim of sustaining fish resources in the damarea Such a cooperative effect should assist in convertingtechnical knowledge into policy and to help disadvantagedsectors in obtaining compensation However, the degree towhich the results from this study are being used in man-agement can not yet be determined It is certain that sci-entific journals reporting results from on-going researchwill impact on dam management and relevant policies.Effective measures can best be achieved by compilingdatabases from countrywide projects that document andanalyze fish density, spatial distribution and behavior invarious dam, river and lake environments in China

Acknowledgments This study was carried out with the tion of the Wuzhou Detachment of Guangxi Zhuang Autonomous Region, Fengkai Detachment, Guangdong Fishing Administrative Brigade, and financed by Science and Technology Item of China (No 2005DIB3J023) and Special Fund for Agro-scientific Research in the Public Interest We would like to thank Professor Yong-Zhen Li and students Xi-Yong Hu and Li-Na Dong for their participation and great help We thank Suenor Woon and Alison Wilcox for assistance with the English in the paper We are grateful to the anonymous reviewers for their contribution in improving the manuscript with their valuable comments.

participa-References

1 Li YF, Li XH, Tan XC (2008) Fisheries resources of the Zhaoqing section the Pearl River Chin J Reserve Fish 28:80–83 (in Chinese)

2 Tao JP, Gao Y, Qao Y (2010) Hydroacoustic observation of fish behavior and spatial patterns in the ship lock and adjacent areas

of the Gezhouba Dam, Yangtze River Acta Ecol Sin 30:233– 239

3 Thorne RE, Johnson GE (1993) A review of hydroacoustic studies for estimation of salmon downriver migration past hydroelectric facilities on the Columbia and Snake Rivers in the 1980s Rev Fish Sci 1:27–56

4 Steig TW, Iverson TK (1998) Acoustic monitoring of salmonid density, target strength, and trajectories at two dams on the Columbia River, using a split-beam scanning system Fish Res 35:43–53

5 Tao JP, Qiao Y, Tan XC (2009) Species identification of Chinese sturgeon using acoustic descriptors and ascertaining their spatial distribution in the spawning ground of Gezhouba Dam Chin Sci Bull 54:3972–3980

6 Tan XC, Chang JB, Tao JP, Li XH (2008) Survey on fish tribution in the forequarter of the three gorge reservoir using EY60 split-beam echosounder Chin J Ecol Sci 27:329–334

dis-7 Chen DQ, Zhang X, Tan XC (2009) Hydroacoutic study of spatial and temporal distribution of Gymnocypris prezewalskii in Qing- hai Lake, China Environ Biol Fish 84:231–239

8 Tan X, Li X, Chang J, Tao J (2009) Acoustic observation of the spawning aggregation of Megalobrama hoffmanni in the Pearl River J Freshw Ecol 24:293–299

9 Aglen A (1983) Random errors of acoustic fish abundance

esti-mates in relation to the survey grid density applied FAO Fish

Rep 300:293–298

10 Myriax (2011) Echoview Version 4.90 Myriax Software Pty

Ltd, Hobart

11 Soule M, Barange M, Solli H, Hampton I (1997) Performance of

a new phase algorithm for discriminating between single and

overlapping echoes in a split-beam echo sounder ICES J Mar Sci

54:934–938

12 Blackman SS (1986) Multiple-target tracking with radar

appli-cations Artech House, Dedham

13 Simmonds J, Maclennan D (2005) Fisheries acoustics: theory and

practice, 2nd edn Blackwell, Oxford

14 Allen P (2009) Operation of spillway gates—how to avoid the

problems and pitfalls In: 34th Annual Qld Water Industry

Operations Workshop Caloundra, pp 98–105

15 Kemp PS, Gessel MH, Williams JG (2008) Response of

down-stream migrant juvenile Pacific salmonids to accelerating flow

and overhead cover Hydrobiologia 609:205–217

16 Scheuerell MD, Schindler DE (2004) Changes in the spatial

distribution of fishes in lakes along a residential development

gradient Ecosystems 7:98–106

17 Prchalova M, Kubecka J, Cech M, Frouzova J, Drastık V,

Ho-hausova E, Juza T, Kratochvıl M, Matena J, Peterka J, Rıha M,

Tuser M, Vasek M (2009) The effect of depth, distance from dam

and habitat on spatial distribution of fish in an artificial reservoir.

Ecol Freshw Fish 18:247–26

18 Warner DM, Schaeffer JS, O’Brien TP (2010) The Lake Huron

pelagic fish community: persistent spatial pattern along biomass

and species composition gradients Can J Fish Aquat Sci

66:1199–1215

19 CEN (European Committee for Standardization) (2005) Water quality—sampling of fish with multi-mesh gill nets, EN 14757 CEN, Brussels

20 Liao GZ, Lu KX, Zhao XZ (1989) Fisheries resources of the Pearl River and their exploitation In: Dodge DP (ed) Proc Int Large River Symp (LARS) Can Spec Publ Fish Aquat Sci 106:561–568

21 Drastık V, Kubecka J, Tuser M, Cech M, Frouzova J, Jarolım O, Prchalova M (2008) The effect of hydropower on fish stocks: comparison between cascade and non-cascade reservoirs Hyd- robiologia 609:25–36

22 Pavlov DS, Mikheev VN, Lupandin AI, Skorobogatov MA (2008) Ecological and behavioural influences on juvenile fish migrations in regulated rivers: a review of experimental and field studies Hydrobiologia 609:125–138

23 Kubecka J, Wittingerova M (2003) Horizontal beaming as a crucial component of acoustic fish stock assessment in freshwater reservoirs Aquat Living Resour 16:325–331

24 Djemali I, Toujani R, Guillard J (2008) Hydroacoustic fish mass assessment in man-made lakes in Tunisia: horizontal beaming importance and diel effect Aquat Ecol 43:1121–1131

bio-25 Mackenzie KV (1981) Nine-term equation for sound speed in the ocean J Acoust Soc Am 70:807–812

26 Francois RE, Garrison GR (1982) Sound absorption based on ocean measurements: part I: pure water and magnesium sulfate contributions J Acoust Soc Am 72:896–907

27 Love RH (1977) Target strength of a fish at any aspect J Acoust Soc Am 62:1397–1403

28 Frouzova J, Kubecka J, Balk H, Frouz J (2005) Target strength of European freshwater fish and its dependence on fish body parameters Fish Res 75:86–96

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

Ex situ and in situ measurements of juvenile yellowfin tuna

Thunnus albacares target strength

Hsueh-Jung Lu•Myounghee Kang•

Hsing-Han Huang•Chi-Chang Lai•

Long-Jin Wu

Received: 10 March 2011 / Accepted: 18 August 2011 / Published online: 17 September 2011

Ó The Japanese Society of Fisheries Science 2011

Abstract To provide target strength (TS) information for

estimating the body length of yellowfin tuna Thunnus

al-bacares and its abundance around fish aggregating devices,

TS was measured ex situ and in situ In the ex situ TS

measurements, two cameras synchronized with a 200 kHz

echosounder were used to obtain the precise orientation of

the yellowfin tuna under free swimming conditions The ex

situ TS (dB re 1 m2)–fork length (FL, cm) regression was:

TS = 27.06 log (FL) - 85.04 Ex situ TS was found to

reach its maximum in the tilt angle range of -15° to -20°

after excluding TS samples with insignificant correlation to

the tilt angle The angle between the vertebra and the swim

bladder was approximately 25° according to X-ray images,

supporting the above tilt range The relationship between

the swim bladder volume (VSB, ml) and the fork length

was: VSB= 0.000213 FL3 The results from the in situ TS

measurements indicated that the tilt angle was highly

concentrated between -10° and 15° The results from a

calculation using the ex situ TS–FL equation with the forklength from biological sampling agreed strongly with theaverage in situ TS

Keywords Target strength Fork length Yellowfin tuna FAD  Swim bladder

IntroductionYellowfin tuna Thunnus albacares (YFT) was once one ofthe most abundant resources for tuna fisheries around theworld; however, stocks of YFT have been depleted bymany commercial fisheries in the three main oceans sincethe 1950s According to the 2007 FAO Yearbook, theglobal catch of YFT began to decline in 2003 becausefishing efforts have increased and advanced fishing tech-nologies have been commonly applied to the tuna fishingindustry [1] One important reason for the downturn inYFT stocks could be the mortality of juvenile YFT caught

as by-catch Documentation from the Western and CentralPacific Fisheries Commission pointed out that 80% of theYFT caught by purse seine using a fish aggregating device(FAD) were immature in the Western Central PacificOcean (WCPO) [2] This was true not only in the WCPObut also in other oceans For example, 42 anchored FADs

in the coastal waters of Taiwan were set up by the TaiwanFisheries Research Institute (TFRI) Most of the YFTaround the FADs in Taiwan were found to be immature.The annual catch of migratory fish per FAD was estimated

to be about 200–300 tons in Taiwan, so the areas where theFADs are effective have become extremely importantfishing grounds for coastal fishermen [3] However, FAD-aided fishing has become a controversial issue in domesticand sustainable fisheries management Therefore, some

H.-J Lu  H.-H Huang

Department of Environmental Biology and Fisheries Science,

National Taiwan Ocean University, 2 Pei-Ning Rd, Keelung 202,

Taiwan, ROC

H.-J Lu

Center of Excellence for Marine Bioenvironment

and Biotechnology, National Taiwan Ocean University,

2 Pei-Ning Rd, Keelung 202, Taiwan, ROC

Coastal and Offshore Resource Research Center, Taiwan Fishery

Research Institute, No.1-1, Yugang N 1st Rd., Cianjhen District,

Kaohsiung 80672, Taiwan, ROC

DOI 10.1007/s12562-011-0401-4

strategies have been developed or adapted for regional tuna

purse seine fishery management For example, the use of

FADs has been banned in certain seasons or zones when/

where high ratios of by-catch are common More

specifi-cally, it is more likely that a modified FAD with deeper

webbing will attract larger YFT A special sorting grid

under the corkline of the purse seine aids in the release of

small YFT [4,5] Nevertheless, juvenile YFT by-catch is

still a serious issue among fisheries managers and

fisher-men Therefore, a quick and efficient method of estimating

body length and even the approximate abundance of YFT

that can be applied prior to a fishing operation is urgently

needed

Hydroacoustic techniques have been widely used for fish

stock assessment and in fish ecological research [6] They

provide great tools, as they allow the simultaneous

collec-tion of qualitative and quantitative data on various aquatic

organisms and environmental information They also

rep-resent one of the most effective and nonintrusive methods

of obtaining information on YFT around FADs A standard

echo integration method using data from a scientific

echo-sounder is used to convert acoustic energy into estimates of

fish density To use this method, the target strength (TS) of a

representative fish must be known The TS is a measure of

the acoustic reflectivity of a fish, which depends on the

presence of a swim bladder, the size of the fish, its behavior,

its morphology, and its physiology [6] It is well known that

a swim bladder and the orientation of the fish are the

greatest influences on the TS, and are strongly associated

with the backscattering cross-section [7] The TS of a fish

with a swim bladder is much higher than that of a fish

without a swim bladder, since the reflection from the swim

bladder contributes approximately 90–95% of the

back-scattered energy [8] In the past, several acoustic surveys of

Atlantic cod Gadus morhua, Atlantic mackerel Scomber

Scombrus, and bay anchovy Anchoa mitchilli have proven

that a swim bladder contributes considerably to the TS [9

11] Moreover, the TS plays a role not only as a decisive

scaling factor for fish abundance estimation, but also as a

tool for fish species identification [12] However, published

research studies on the TS of tuna are scarce, except those

on the in situ TS of YFT and big eye tuna Thunnus obesus

from French Polynesia performed by Bertrand et al [13]

and Josse and Bertrand [14] No measurements of the TS of

tuna off the coast of Taiwan have been performed Hence,

precise information on the TS around FADs is urgently

required, in order to calculate a representative body length

of YFT and to estimate the abundance of YFT Three

techniques have been used to estimate the TS: (1) in situ

measurements of free swimming in the natural habitat [13,

15,17]; (2) ex situ measurements of dead or live fish in a

controlled environment [16]; (3) numerical or theoretical

backscattering models based on fish anatomy [7,18] Ex

situ TS measurements are especially closely examined todetermine how variables such as species, length, and ori-entation influence individual fish TS values

To identify YFT and to estimate its body length andabundance are vastly important, in order to mitigate oravoid as much as possible the by-catch of juvenile YFT inFAD aided fishing operation Therefore, this study wasperformed to obtain accurate TS information on YFTaround FADs using ex situ and in situ measurements In the

ex situ TS experiment, optical systems synchronized to theacoustic system were used to obtain the orientation of YFTunder free swimming conditions Thus, the aims of thisstudy were to: (1) obtain the ex situ TS–fork length rela-tionship; (2) obtain the swim bladder–fork length rela-tionship; and (3) examine the ex situ TS and in situ TS ofYFT Eventually, these results will contribute to (1) thequantitative assessment of YFT and (2) by-catch mitigation

in FAD-aided tuna fishing

Materials and methods

Ex situ TS experimental system and YFT sampled

In order to measure the dorsal ex situ TS of live YFT undernatural swimming conditions, an EY60 echosounder[Simrad (Horten, Norway), 200 kHz] was linked to twooptical devices [Sony (Tokyo, Japan), DCR-PC115 cam-corders] and installed in a circular tank 12 m in diameterand 4 m deep (Fig.1) The transducer and two cameraswere attached to a steel frame tied onto two life buoys.Camera 1 was mounted such that it recorded the top-downview, enabling the exact fork length of YFT to be calcu-lated Camera 2 was placed 3 m deep in order to record theside view, enabling the tilt angle of the YFT to be assessed.The tilt angle was defined as the angle between the bodyaxis of YFT and the horizontal line (the xy plane), where apositive tilt angle indicates that the fish is facing up and anegative angle means that the fish is facing down Everyping from the echosounder was synchronized with the twocameras, which were both connected to a computer clock.The paired images from the two cameras were selected andcaptured by a VGA recorder [Epiphan (Ontario, Canada),VGA Recorder StandardTM] The recorder was capable ofdiscerning a paired image corresponded to a ping from theechosounder The echosounder used was the EY60 systemwith a split-beam transducer that had a nominal angle of 7°

It was calibrated according to the standard calibrationmethod [19] using a 32 mm diameter copper sphere Theacoustic parameters used for the ex situ TS experiment areshown in Table1

Juvenile YFT that aggregated around the FADs insouthwestern Taiwan (120°22.2130E and 22°12.0610N)

were caught by troll lines during 25–30 June 2010 using a

research vessel of the TFRI The catch was kept alive in a

tank onboard and shipped with special caution to the

cir-cular tank on land Approximately 50 live YFT were

selected for the ex situ experiment after they had adapted to

the new environment for several days In this study, a

juvenile YFT is defined as a YFT with a body length of

25–80 cm

Measurement of fork length

The fork length and tilt angle were measured using the

footage from the cameras and image measurement

share-ware (Meazure 2.0) The image size decreases with the

distance from the camera To calculate the fork length of

the YFT in the optimal zone (depth of 2.6–3.4 m; Fig.1),

the relationship between the image size and the depth needs

to be considered Accordingly, the actual size and the

image size were compared using a standard ball The

standard ball used was 9.45 cm in diameter, and it was

positioned at depths of between 2 and 3.4 m The size of

the ball was estimated from the number of pixels on the

video image The linear relation between the number of

pixels corresponding to the standard ball (BPD) and the

depth (D) determined by the echosounder can be derived

from

where m is the slope and n is the intercept for the

regression The length of the YFT (number of pixels) and

the tilt angle (h, degrees) were obtained from the paired

video image More precisely, the image from camera 1 was

utilized to extract the projected length (FPDin Eq.2), andthat from camera 2 was used to extract the tilt angle (h in

Eq 2) A tilt angle of 0° describes the direction parallel tothe xy (horizontal) plane, -90° describes the downwarddirection, and 90° the upward direction Finally, the actualfork length (FL, cm) can be calculated as follows:

where 9.45 is the diameter of the standard ball, and FPDisthe length of the YFT measured from the video image fromcamera 1

Fig 1 Experimental setup for

ex situ TS and tilt angle

measurements The

echosounder and two cameras

were synchronized Two life

buoys floated on the surface of

the water in a 430 m3tank;

these buoys were fixed to a steel

frame, and were used to support

the 200 kHz split beam

transducer and cameras Ex situ

TS data samples were analyzed

when a yellowfin tuna was

within the optimal zone, at a

depth of between 2.6 and

Single target detection

Minimum echolength ratio with pulse duration 0.8 Maximum echolength ratio with pulse duration 1.8

Ex situ TS–fork length relationship with regard

to tilt angle

The dorsal ex situ TS of YFT was collected when the

geometric center of a single target (i.e., a YFT) detected by

the echosounder was within the optimal zone (Fig.1) For

each YFT, the video image length and the tilt angle were

obtained from the paired images from cameras 1 and 2 The

fork length was then obtained using the Eq.2 Empirically

regressing TS against fork length is a common method of

investigating their relationship The ex situ TS–fork length

relationships (i.e., linear regression equations using a free

slope and a restricted slope of 20) were established in this

study (Eqs.3,4) The free slope and the restricted slope of

20 were used because backscattering is expected to be

proportional to the cross-sectional area (L2) of the target

However, in some species, slope values deviating from 20

provide a better TS–length fit McClatchie et al [20]

insisted that the regression slope should not be set to 20,

and recommended fitting an empirical TS–length

rela-tionship for each species McQuinn and Winger [21] and

Rose [11] modified the standard regression according to

fish behavior, especially the diel orientation pattern

Free slope: ex situ TS¼ a logðFLÞ þ b ð3Þ

Fixed slope: ex situ TS¼ 20 logðFLÞ þ B20: ð4Þ

Entire TS samples were used to examine the ex situ

TS–fork length relationship with regard to tilt angle To

observe this more precisely, the total number of TS

samples were grouped into 5° intervals of tilt angle

Moreover, the relationship between the ex situ TS and the

tilt angle with regard to fork length was examined using

5 cm intervals of fork length

Measurements of the swim bladder volume

The volumes of the swim bladders of 45 YFT were

mea-sured These YFT were sampled around the same FADs as

used in the ex situ TS experiment These fish died onboard,

but were brought to the laboratory within 24 h to perform

measurements The fork length range of the YFT was 32.5–

129 cm The procedure used to measurie the volume of a

swim bladder was to: (1) clean the abdominal cavity except

the swim bladder, and (2) using a syringe, inject water

into the swim bladder until it was full During the

mea-surement, the air in the swim bladder was released and

replaced with the injected water The injected water volume

was equal to the volume of the swim bladder The

rela-tionship between fork length and swim bladder volume

(VSB, which is also the water volume) can be obtained using

while the precise location of the swim bladder, especially itsangle, was determined from anatomic and X-ray pictures

In situ TS measurement and data analysis

A stationary acoustic survey was conducted using the sameEY60 echosounder around the same FADs around whichthe live YFT were sampled for the ex situ TS experiment,

in order to understand the characteristics of the in situ TS.The location of the stationary survey was selected based onthe confirmation of the presence of YFT schools by amobile survey conducted in a star pattern On 22 December

2010, the acoustic survey was carried out between 6 and

7 p.m as the YFT migrated to feed The acoustic eters used for the in situ TS measurements were same asthose used in the ex situ TS experiment (Table1) The trollline was employed as a ground truth method Therefore,fish species and body lengths were gained from thesamples

param-The acoustic data were analyzed to detect single targetsand fish tracks in Echoview (Myriax, version 4.90) Theparameter settings for single target detection were equal tothose used for single target detection in the ex situ TSexperiment (Table1) Fish tracks were extracted only if thenumber of single targets in the fish track was higher thanthree consecutively (Fig.2) A total of 567 fish tracks wereextracted from the 7825 single targets that were detected.For each fish track, fish behavior information, such asswimming speed, orientation, and so forth, was exported.The difference in TS between the first and the last singletarget in a fish track was regarded as the difference due tothe orientation of the fish; i.e., the tilt angle

Results

Ex situ TS–fork length relationshipThe linear regression (Eq.1) between the video imagelength of the standard ball (BPD) and its depth (D) wasestablished as

BPD¼ 13:06 D þ 61:12 ðr2¼ 0:89; p\0:01Þ: ð6Þ

Ex situ TS from 185 single YFT targets were detected inaccordance with paired images from cameras 1 and 2 Thefork length of the YFT was computed using Eq.2 Thefrequency distributions of ex situ TS, fork length, and tiltangle are shown in Fig 3 The fork length distribution hasthe range of 25.2–68.1 cm Two major modes (A and C) and

a minor mode (B) in the fork length were observed (Fig.3a).The ex situ TS distribution exhibited a range of -50 to-30 dB, with two slightly different modes (Fig.3b) The

mode (B) in Fig.3b may be shifted from small-sized fish,

which is mode (B) in Fig.3a, with a negative tilt angle The

most frequent orientation (approximately –5° to 1°) of the

fish that passed through the optimal zone was a

horizontal-like aspect (Fig.3c) Using the overall datasets, a significantpositive correlation between ex situ TS (dB) and logarithmicfork length (cm) was found, as summarized in the followingtwo linear equations:

Fig 2 Echogram of a stationary acoustic survey, with ping number

plotted on the x-axis and depth (in m) along the y-axis The fish track

echograms are included The first and last single targets were used to

examine the orientation of the yellowfin tuna, and are marked in the expanded fish track echogram

Fig 3 The frequency

distributions of fork length (a),

ex situ TS (b), and tilt angle

(c) for yellowfin tuna The

normal distribution is

represented by the black

straight line SD standard

deviation, N number of samples

Free slope: ex situ TS¼ 27:06 logðFLÞ  85:04

ðr2¼ 0:42; p\0:01Þ ð7ÞFixed slope: ex situ TS¼ 20 logðFLÞ  73:69

ðr2¼ 0:39; p\0:01Þ: ð8Þ

Ex situ TS variation in relation to tilt angle and fork

length

The variation observed in the ex situ TS observations when

fitting the linear regression (Fig.4) may be caused by the

presence of various tilt angles Therefore, ex situ TS was

thoroughly examined by grouping the targets into 5°

intervals of tilt angle from -30° to 25° The linear

regression for each group was calculated and is presented

in Table 2 The largest difference in B20 (the intercept ofTS–fork length regression, using a fixed slope), which is2.58 dB, is found between the groups corresponding to 5°

to 10° and -20° to -15° Most of the ex situ TS withnegative tilt angles are higher than those with positiveones When the groups with insignificant correlations wereexcluded, the highest B20was obtained for the group cor-responding to -15° to -20° The result, excluding groupswith low correlation, is plotted in Fig.5 The ranges of thex- and y-axes are clearly narrower The standardized TS(B20) was found to be highly correlated with the tilt angle.Meanwhile, plots of the ex situ TS versus orientation fordifferent fork lengths are illustrated in Fig.6 The depen-dence of ex situ TS on tilt angle changes greatly as the forklength increases The difference in TS is larger than

*10 dB in each group, indicating that the orientationdirectly affects the difference Long YFT seem to swimmore actively than short YFT, since the ranges of ex situ

TS values for the groups representing longer fish are widerthan those for the groups representing shorter fish How-ever, the ranges of ex situ TS by tilt angle for groups withfork lengths of 25–35 cm do not appear to be as wide asthose for other length groups In general, the ex situ TSappears to be higher when the YFT faces down (when thetilt angle has a negative sign)

Relationship between swim bladder volume and forklength

The anatomical and X-ray images were used to estimate theprecise location of the swim bladder based on the position

of the vertebra An example is shown in Fig.7 One reason

Fig 4 The relationship between ex situ TS and fork length based on

linear regression The dependence of the ex situ TS on the logarithm

of the fork length is clearly shown, although the ex situ TS values are

rather scattered

Table 2 Summary of the parameters from the linear regression of TS versus orientation (tilt angle) for each 5° interval of tilt angle

that the ex situ TS was constantly high when the YFT were

heading downwards is the angle between the vertebra and

the swim bladder (approximately 25°) This angle implies

that if the YFT head down at an angle of approximately25°, their swim bladders become flat, yielding the maxi-mum backscattering cross-section Hence, the maximum

TS appears when the YFT swim downward Figure8

represents the relationship between the volume of the swimbladder (ml) and the fork length (cm) This relationshipwas described using Eq.5, and acts as the growth equation.Figure8clearly shows that the volume of the swim bladderincreases more rapidly with fork length in the fork lengthrange 50–72 cm than at other lengths, based on Eq 10andFig.8 Note that, in terms of the sample data used to derivethe equations, the number of samples that provided the dataused to derive Eq.9was only one more than the number ofsamples that contributed the data used to construct Eq 10.The reason that a data point ([100 cm) was used was toview the trend of the relationship between the volume ofthe swim bladder and the fork length for the adult YFT.Using all YFT: VSB¼ 0:000213 FL3

ðr2¼ 0:81; n ¼ 45; p\0:01Þ ð9ÞExcluding adult YFTði:e:; [ 100 cmÞ:VSB

¼ 0:000144 FL3ðr2 ¼ 0:95; n ¼ 44; p\0:01Þ: ð10Þ

Fig 5 Scatter plot of B20 versus tilt angle after excluding some

groups (each TS group represented a 5° interval of tilt angle) that

exhibited insignificant correlations between TS and tilt angle A circle

indicates that the TS is significantly related to the logarithm of the

fork length A triangle means an insignificant relationship Relatively

high TS values tend to occur on the left side of the plot (negative tilt

angles), which means that the heads of the YFT tend to be pointing

down

-55 -50 -45 -40 -35 -30

-25

45-50 cm

-55 -50 -45 -40 -35 -30

Tilt angle (°)

60-65 cm

-55 -50 -45 -40 -35 -30 -25

In situ TS and tilt angle of YFT

The tilt angle frequency distribution obtained from in situ

TS measurements performed around the FAD is illustrated

in Fig.9 Most of the tilt angles occur in the interval –20°

to 20° However, there is a high concentration of tilt angles

between -10° and 15° Approximately 25% of the fish

were found to be swimming downward, leading to an

increase in TS This phenomenon is similar to what was

observed in the ex situ TS experiment The frequency

distribution of the in situ TS is shown in Fig.10 Two

modes of in situ TS were found; the first mode was

approximately -40 dB and the second mode was about

-28 dB This implies that two length classes of YFT were

aggregated under the FAD According to biological

sam-pling performed in parallel with the acoustic survey, the

fork length range of the first TS mode in Fig.10is around

39–46 cm (mean = 42.4, SD = 3.7) A TS of -41.01 dB

was obtained using the ex situ TS–fork length relation

(Eq 7) with a mean fork length of 42.4 cm gained from thetroll lines This value was very close to the average in situ

TS value of -39.66 dB, which was the mean value of theleft mode in Fig.10 However, no biological samplescorresponding to the second TS mode (mean TS =-25.38 dB) were obtained

DiscussionExperimental design and setup

It was extremely difficult to keep the YFT samples alive onboard Hence, the number of YFT used in the ex situ TSand in the in situ TS experiments differed In the caged andtethered experiments, it was possible to monitor an

Fig 7 X-ray image profile The position, size, and angle between the

vertebra and the swim bladder can be obtained from the image

Fig 8 Regressions between the swim bladder volume and the fork

length of yellowfin tuna The solid line was created using the entire

dataset, while the dashed line was generated by excluding adult

yellowfin tuna ([100 cm)

Fig 9 The tilt angle frequency distribution extracted from the fish tracks The data were obtained from in situ measurements of the TS of yellowfin tuna distributed around the FAD SD standard deviation,

N number of samples

Fig 10 The frequency distribution of the average in situ TS obtained from the extracted fish tracks SD standard deviation, N number of samples

individual fish over a long time period in order to obtain a

considerable amount of data [16,22] However, this would

be difficult to achieve with free swimming YFT owing to

their physiological characteristics Moreover, a great

number of data samples where the geometric centers of the

YFT were not completely included in the optimal zone

were excluded from the data processing Also, when

pro-cessing the video images, YFT pixel images without clear

edges of the head and tail (i.e., indicating targets that were

swimming towards or away from camera 2) were excluded

However, the total number of juvenile YFT monitored in

the ex situ experiment (185) should be sufficient to

esti-mate an accurate value for the ex situ TS and to precisely

measure the fork length

Ex situ TS values were concentrated in the range of tilt

angles from -10° to 10° (Fig.3c) It was assumed that

YFT may not move very easily due to the relatively

shal-low depth of the water tank However, the YFT in the tank

seemed to swim naturally, based on visual observations

For example, they swam freely during the vertical

migra-tion when feeding One concern should be noted though:

some large fish might not have been completely insonified

Hence, the experimental environment—particularly that of

a relatively deep water tank—should be adapted to provide

a sufficient depth for the YFT to swim around without

encountering obstacles

There is another important issue about the nonlinear loss

that should be mentioned [23, 24] The loss caused by

nonlinear acoustic propagation may be important,

espe-cially in the case of high-power, high-frequency, and

highly directive sources It is important to consider the

nonlinear loss when measuring TS The transmitted power

and the range of the target significantly affect it In this

study, a power of 1,000 W was used in both the tank and

field experiments (unintentionally) Nonlinear effects have

been studied using simulation methods For example,

Tichy et al [23] simulated nonlinear loss using a Bergen

code simulation with three power settings (83, 593, and

927 W), along with 200 kHz Their simulation extended

out to 12 m from the transducer They found that the

nonlinear losses for a simulated power of 927 W at 2.6 and

3.4 m (the depth range for ex situ TS measurements in this

study) were -1.9 and -2.2 dB, respectively Pedersen [24]

simulated this loss using numerical simulation with various

power settings (50, 150, 1,000, and 1,500 W) at both 120

and 200 kHz His simulation extended out to 300 m from

the transducer Using his simulation with settings of

1000 W and 200 kHz, the nonlinear losses at 2.6 and 3.4 m

were -1.8 and -2.2 dB, respectively The losses at 40 and

60 m (the depth range for in situ TS measurements in this

study) were -4.6 and -5 dB, respectively This nonlinear

loss range can be understood as the range of possible errors

in TS measurement in this study The mean ex situ TS was

-39.06 dB (Fig.3b), and the mean in situ TS at the firstmode was -39.66 dB (Fig.10) Based on the compensated

TS values obtained using the simulated nonlinear loss fromthese works, they should be in the proximity of -37.06 and-34.86 dB, respectively Note that both simulations wereperformed under simulated freshwater conditions Thenonlinear loss in seawater may be smaller than that infreshwater [24] Thus, the TS may be slightly lower ifseawater conditions are assumed It is important to statethat accurate measurements of TS should ideally be per-formed using low power (approximately less than 100 W),and at a shallow depth, as used in an experimental tankenvironment

Relevant research on the TS of YFT and its swimbladder

A few hydroacoustic groups have performed research onthe TS of tuna Using a Simrad EK500 echosounder at

38 kHz, the in situ TS–L relationship for YFT of length60–120 cm in French Polynesia was obtained by Bertrand

et al [13] and Bertrand and Josse [25] as

In situ TS¼ 25:26 logðFLÞ  80:62: ð11ÞThe fork length in Eq.11 was derived from knowninformation, since in situ TS values were measured for fourYFT that were tagged in water depths of up to 200 m Theyalso found that the TS increased with depth [14] Forexample, a TS of -50 dB beyond 270 m was found,compared to a TS of -34 dB at 500 m Thus, highlysignificant differences in tuna TS values with depth wereobserved It should be mentioned that this result wasderived from two tuna species: bigeye tuna Thunnus obesusand yellowfin tuna For the YFT, a significant effect ofdepth on TS was not observed because the tuna weredistributed over a rather narrow depth range (approximately

up to 200 m) A trend in TS with depth was apparent,although it was not significant [14]

In this study, the volume of the swim bladder for YFT oflength 60 cm was 46.01 ml, and that for YFT of length

90 cm was 155.28 ml, based on Eq.9 Bertrand and Josse[25] found that the volume of the swim bladder for YFT oflength 60 cm was 80 ml, and that for YFT of length 90 cmwas 130 ml The TS increased logarithmically with thevolume of the swim bladder in both studies While thesetwo sets of results did not match perfectly, they did showgood agreement The development of the swim bladder inthe early stages of the growth of juvenile YFT appears tofollow a linear relationship rather than a logarithmic one,

up until the YFT reach the adult stage Note that thenumber of adult YFT sampled in the present research wasnot sufficient Thus, more adult YFT need to be sampled inorder to examine the relationship between YFT swim

bladder volume and fork length At least 25 samples are

needed to achieve a statistically meaningful result

When the swim bladder axis is near horizontal, TS

reaches its maximum value [6] In this study, the TS

obtained with negative angles, particularly those between

-15° and -20°, appeared to give the best B20values, since

the angle between the swim bladder and the vertebra in

such cases was approximately -25° A similar angle of 20°

was found for bigeye tuna Thunnus obesus, and the highest

TS was observed when the tuna was descending at that

angle [13]

Future research plan

The effective distance from the FAD could be employed as

a scale to quantitatively estimate the effectiveness of the

FAD Fre´on and Dargorn [26] introduced the terms

‘‘ex-tranatants’’ (i.e., fish that remain within 10–50 m of a

FAD) and ‘‘circumnatants’’ (i.e., fish that remain within

50 m to several nautical miles from a FAD) Doray et al

[27] found that most of the fish biomass was concentrated

within a radius of 400 m around the two moored FADs

Mobile acoustic surveys were designed by Josse et al [28]

for the purpose of quantitative estimation based on the

FAD An eight-branch star pattern would be optimal to

estimate the effectiveness of the FAD This pattern was

actually employed in this study, but the quality of the

resulting acoustic data was found to be too low Therefore,

a stationary acoustic survey was carried out instead A

future plan is to use the eight-branch star pattern as a

survey design to investigate the effective distance from the

FAD along the Taiwan coast

Fish tracking techniques can enhance our understanding

of the behavior of fish species, including YFT The results

can elucidate swimming behavior such as speed and

direction (vertical and horizontal displacements) This is

essential information for fish ecologists, FAD managers,

and fisheries resource researchers Therefore, a more

thor-ough analysis using tracked fish is needed in the near future

A study has focused on the distribution of the tuna

school, as well as its interaction with its prey (i.e.,

micro-nekton) [13] Current acoustic data analysis applications

allow fish schools to be detected automatically, and they

can provide data on various characteristics (morphological,

positional, and energetic descriptors) of the school in

comma-separated value format Schools of YFT can be

investigated more precisely using the abovementioned

characteristics The multi-frequency technique—the

dif-ference in the mean volume backscattering strengths at two

frequencies, also called the dB difference—has been

widely used, especially to distinguish fish schools from

krill or plankton and to classify fish of the same species

into age groups [29,30] This method can also distinguish

micronekton from tuna-like fish schools It is worth usingthis technique to study the interaction between tuna and itsprey, as well as to identify tuna schools

Acknowledgments We are grateful to the captains and crews of the TFRI research vessels Hai Fu and Hai Jian for their help during the acoustic surveys and fish sampling This study was supported by the Council of Agriculture [96 AST-15.1.2-FID-02(14)] and the National Science Council (NSC95-5611-M-019-007) of the Republic of China (Taiwan) We thank Alison Wilcox for improving the English of the paper We are also grateful to both referees, whose comments helped improve this paper.

3 Wu LJ, Weng CH (2006) The comparison of the aggregating efficacy of mid-water fish aggregating device off south-western and eastern Taiwan Fish Res Inst Newsl 16:19–23 (in Chinese)

4 Nelson PA (2004) Reducing juvenile bigeye tuna mortality in FAD sets (working paper) In: 17th Standing Committee of Tuna and Billfish, South Pacific Commission, Majuro, Marshall Islands, 8–18 Aug 2004

5 IATTC (2008) Annual report of the Inter-American Tropical Tuna Commission for 2006 IATTC, La Jolla

6 Simmonds J, MacLennan D (2005) Fisheries acoustics: theory and practice, 2nd edn Blackwell, Oxford

7 Hazen EL, Horne JK (2004) Comparing the modeled and sured target-strength variability of walleye pollock, Theragra chalcogramma ICES J Mar Sci 61:363–377

mea-8 Foote KG (1980) Importance of the swimbladder in acoustic scattering by fish: a comparison of gadoid and mackerel target strengths J Acoust Soc Am 67:2084–2089

9 Bosewll KM, Wilson CA (2008) Side-aspect target strength measurements of bay anchovy (Anchoa mitchilli) and Gulf menhaden (Brevoortia patronus) derived from ex situ experi- ments ICES J Mar Sci 65:1012–1020

10 Nesse TL, Hobaek H, Korneliussen RJ (2009) Measurements of acoustic-scattering spectra from the whole and parts of Atlantic mackerel ICES J Mar Sci 66:1169–1175

11 Rose GA (2009) Variations in the target strength of Atlantic cod during vertical migration ICES J Mar Sci 66:1205–1211

12 Jurvelius J, Knudsen FR, Balk H, Marjomaki TJ, Peltonen H, Taskinen J, Tuomaala A, Viljanen M (2008) Echo-sounding can discriminate between fish and macroinvertebrates in fresh water Freshw Biol 53:912–923

13 Bertrand A, Josse E, Masse J (1999) In situ acoustic strength measurement of bigeye (Thunnus obesus) and yellowfin tuna (Thunnus albacares) by coupling split-beam echosounder observations and sonic tracking ICES J Mar Sci 56:51–60

target-14 Josse E, Bertrand A (2000) In situ acoustic target strength surements of tuna associated with a fish aggregating device ICES

17 Sawada K, Takahashi H, Abe K, Ichii T, Watanabe K, Takao Y

(2009) Target-strength, length, and tilt-angle measurements of

Pacific saury (Cololabis saira) and Japanese anchovy (Engraulis

japonicus) using an acoustic-optical system ICES J Mar Sci

66:1212–1218

18 Jech JM, Horne JK (2002) Three-dimensional visualization of

fish morphometry and acoustic backscatter J Acoust Soc Am

3:35–40

19 Foote KG, Aglen A, Nakken O (1986) Measurements of fish

target strength with a split-beam echosounder J Acoust Soc Am

80:612–621

20 McClatchie S, Macaulay GJ, Coombs RF (2003) A requiem for

the use of 20log10Length for acoustic target strength with special

reference to deep-sea fishes ICES J Mar Sci 60:419–428

21 McQuinn IH, Winger PD (2003) Tilt angle and target strength:

target tracking of Atlantic cod (Gadus morhua) during trawling.

ICES J Mar Sci 60:575–583

22 Jørgensen R, Olsen K (2002) Acoustic target strength of capelin

measured by single target tracking in a controlled cage

experi-ment ICES J Mar Sci 59:1081–1085

23 Tichy FE, Solli H, Klaveness H (2003) Non-linear effects in a

200-kHz sound beam and the consequences for target-strength

measurement ICES J Mar Sci 60:571–574

24 Pedersen A (2006) Effects of nonlinear sound propagation in fisheries research (Ph.D thesis) University of Bergen, Bergen

25 Bertrand A, Josse E (2000) Tuna target-strength related to fish length and swimbladder volume ICES J Mar Sci 57:1143–1146

26 Fre´on P, Dagorn L (2000) Review of fish associative behaviour: toward a generalisation of the meeting point hypothesis Rev Fish Biol Fish 10:183–207

27 Doray M, Josse E, Gervain P, Reynal L, Chantrel J (2006) Acoustic characterization of pelagic fish aggregations around moored fish aggregating devices in Martinique (Lesser Antilles) Fish Res 82:162–175

28 Josse E, Bertrand A, Dagorn L (1999) An acoustic approach to study tuna aggregated around fish aggregating devices in French Polynesia: methods and validation Aquat Living Resour 12:303–313

29 Kang M, Furusawa M, Miyashita K (2002) Effective and accurate use of difference in mean volume backscattering strength to in- dentify fish and plankton ICES J Mar Sci 59:794–804

30 Kang M, Honda S, Oshima T (2006) Age characteristics of walleye pollock school echoes ICES J Mar Sci 63:1465–1476

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

Fishery biology of mud crabs Scylla spp at Iriomote Island,

Japan: species composition, catch, growth and size at sexual

maturity

Cynthia Yuri Ogawa•Katsuyuki Hamasaki•

Shigeki Dan• Shuichi Kitada

Received: 10 June 2011 / Accepted: 30 August 2011 / Published online: 25 September 2011

Ó The Japanese Society of Fisheries Science 2011

Abstract The fishery biology of mud crabs Scylla spp

was examined using baited traps and gill nets from

Sep-tember 2001 to August 2005 at Iriomote Island, Japan To

elucidate the growth of the crabs, artificially produced

S serrata juveniles were released and recaptured at the

study site The sizes at which 50% of females and males of

S serrata reached sexual maturity (SM50) were estimated

as an external carapace width (ECW) based on the

mor-phology of the abdomen and the chela respectively Two

species, S serrata and S olivacea, were identified in the

area with S serrata being the dominant species ([95% of

the catch) Changes in the mean ECW and the results of

the release and recapture experiments suggested that the

recruitment of young crabs to the fishery occurred from

December/January to April/May The SM50of females and

males occurred at 132.4 and 150.7 mm ECW respectively

The body size composition of S serrata revealed that

immature crabs comprised approximately 40 and 65% of

the catch for females and males respectively To maintain a

sustainable fishery for S serrata, a minimum landing size

based on the SM50 estimates should be implemented as a

fishing regulation

Keywords Baited traps Chela allometry  CPUE Generalised additive model Gill net  Minimum landingsize  Juvenile release

IntroductionMud crabs of the genus Scylla de Hann are large portunidsthat live in estuaries and coastal waters throughout thetropical to warm temperate zone in the Pacific and IndianOceans [1 3] Mud crabs are acknowledged for theircommercial importance as food, and overfishing of mudcrab species has been observed at varying levels in dif-ferent countries in accordance with the development ofboth national and international markets [4] Mud craboverfishing has resulted in the decreased size and abun-dance of mud crabs in many fisheries, making them anincreasingly scarce resource throughout the Indo-Pacificregion [4 16] Fishing regulations, such as prohibiting thecapture of immature crabs, should be indispensable forsustainable utilisation of mud crab resources [5,8,14].Traditionally, mud crabs were grouped into one species,Scylla serrata [17] However, the species identification ofthe genus Scylla has been controversial [18, 19], and theresearchers have reported that mud crabs include severalspecies/morphs in many places including the Philippines[20], Vietnam [21], Malaysia [22], India [23] and Japan[3,18,24] Recently, taxonomy of the genus Scylla has beenresolved as four distinct species, i.e., S serrata (Forska˚l),

S tranquebarica (Fabricius), S olivacea (Herbst) and

S paramamosain Estampador by Keenan et al [2] based

on both morphometric and genetic characteristics Sincethen, biological and ecological studies in relation to fisher-ies have focused on individual mud crab species in severalcountries and regions [3 16], including Japanese waters

Electronic supplementary material The online version of this

article (doi: 10.1007/s12562-011-0408-x ) contains supplementary

material, which is available to authorized users.

C Y Ogawa  K Hamasaki (&)  S Kitada

Department of Marine Biosciences, Tokyo University of Marine

Science and Technology, Minato, Tokyo 108-8477, Japan

e-mail: hamak@kaiyodai.ac.jp

S Dan

Tamano Laboratory, National Research Institute of Fisheries

and Environment of Inland Sea, Fisheries Research Agency,

Tamano, Okayama 706-0002, Japan

Fish Sci (2011) 77:915–927

DOI 10.1007/s12562-011-0408-x

[3,18,19,24–26] In Japan, S serrata, S paramamosain

and S olivacea inhabit coastal inlets and support

commer-cially important fisheries on a local scale in warm temperate

zones, including Lake Hamana in Shizuoka Prefecture and

Urado Bay in Kochi Prefecture, where several studies have

been conducted to reveal the species composition of the

catch and growth of mud crabs [25,26] Furthermore, mud

crabs are commercially fished in subtropical waters in

Ryukyu Archipelago, which is located in southern Japan

[24] However, little is known about the fishery biology, such

as the species composition of the catch, catch per unit effort

(CPUE) and growth, of mud crab species in Japanese

sub-tropical waters

Therefore, we studied the fishery biology of mud crabs

in Ryukyu Archipelago and selected Iriomote Island as the

study area Iriomote Island is located at the southernmost

end of the Ryukyu Archipelago (24°200N and 123°490E)

(Fig.1) The island has several river estuaries with

man-grove forests where mud crabs are abundant We

charac-terised the fishery of mud crabs for fishing gear used as

well as the species composition, CPUE, sex ratio and body

size (carapace width) of the mud crabs caught in the area

To elucidate the growth of mud crabs before and at

recruitment to the fishery, release and recapture

experi-ments using artificially produced juveniles of the dominant

mud crab species S serrata were conducted Juveniles

were released into an open natural tidal flat and in a net

enclosure set on a natural tidal flat in Iriomote Island The

mud crabs were then recaptured and measured

Further-more, the size of S serrata at sexual maturity was

exam-ined as the basis for setting a minimum landing size as a

fishing regulation for this species in Iriomote Island

Materials and methods

Collection of fishery data

In Iriomote Island, mud crabs were caught using baited

traps and gill nets Commercial rectangular traps (45 cm

wide 9 60 cm long 9 20 cm high) with two slit entrances

were used throughout the year during the period of this

study The galvanised rod frame of the trap was covered

with a black polyethylene square-shaped mesh net with a

stretched mesh diameter of 2 cm Bait consisted of pieces

of fish resulting from the by-catch and skip jack The gill

nets were made of monofilament nylon (dimension of one

unit being 1.8 m wide 9 25 m long) with a stretched mesh

size of 6 cm Local legislation restricted the use of gill nets

from October to May Each trap was connected to a

polyethylene rope, marked with a buoy, and individually

deployed alongside the river margin near the mangrove

forest, and gill nets were placed at the mouth of the river

An overnight soaking time was utilised for both types offishing gear

The most active crab fisherman, who employed bothfishing gears to catch mud crabs in the northwestern fishinggrounds, including the Nakara and Kuira Rivers and theiradjacent waters as well as the Amitori and Sakiyama Bays

in Iriomote Island (Fig.1), was selected to collect thefishery data Fishing activity was monitored for four suc-cessive years from September 2001 to August 2005 Afisherman was trained to record the following data on adaily basis: the type of fishing gear and the number offishing gear units as well as the species, sex and externalcarapace width (ECW) including the anterolateral spines tothe nearest 0.5 cm of each individual captured mud crab.Species identification was performed according to Keenan

et al [2] and Keenan [27]

Analyses of fishery dataThe following data were summarised on a monthly basis:the number of fishing gear units, the number of operationdays for traps and gill nets, and species composition Catchdata from each fishing gear were standardised as the CPUE

on each operation day (number of crabs gear unit-1day-1) The sex ratio was calculated as the number ofmales divided by the total number of crabs caught in eachmonth of each year To detect the yearly and seasonalfluctuations of the monthly species composition, the dailyCPUE, the monthly sex ratio and the individual ECW datafor mud crabs caught by each type of fishing gear, gener-alised additive models (GAMs) using the ‘‘mgcv’’ package[28] for the R language [29] were employed A GAM is anonparametric extension of the generalised liner model.This modelling approach allows the relationship betweenexplanatory variables and the response variable to beanalysed by a linear and/or nonlinear smoothing function,and this approach is flexible regarding distributions of data[28,30] GAMs have been applied to temporal datasets todetect yearly and seasonal trends in biological data [30] Inthe GAM analysis for this study, the proportion data onspecies composition and the sex ratio in addition to thenumber of crabs (daily catch data) and ECW values ofindividual males and females were used as the responsevariables The explanatory variables included two temporalvariables: year and month The year variable had integervalues between 1 and 4 (corresponding to the first to fourthsurvey years with each year being defined as the periodfrom September to the following August), and the monthvariable had values between 1 (September) and 12 (fol-lowing August) GAMs consisting of a binomial errordistribution with a logistic-link function, Poisson errordistribution with a log-link function and a Gaussian errordistribution with an identical link function were assumed

for proportion data, catch data and ECW data respectively.

In the GAMs, the number of crabs caught was standardised

by the offset variable of the number of fishing gear units

[30] The models were fitted using quasi-frameworks for

binomial and Poisson GAMs with overdispersion of the

data taken into consideration [30] The maximum number

of degrees of freedom of a smooth term for the month

variable was set at four in each GAM to avoid overfitting

and to infer the seasonal fluctuation The significance of

each explanatory variable in the model was evaluated by

the approximate F test using the ‘‘mgcv’’ package

Release and recapture experiments

Release and recapture experiments using artificially

produced juveniles were conducted for S serrata, which

was the dominant mud crab species in the study area

Broodstock management and larval rearing to produce

juvenile crabs were conducted according to the methods

of Hamasaki [31] and Hamasaki et al [32]

Approxi-mately 270,000 juveniles in the first to second crab

stages were stocked in five net pens with bottom nets

(4-mm mesh size; 10 9 10 9 2 m) set at a tidal flat

(Fig.1) on May 27, 2003 Crabs were fed with minced

krill After 1 month of culture, juvenile crabs grew to a

mean ECW of 23 mm, and 18,330 crabs were released

into an open tidal flat at the mouth of the Nakara River

on June 26, 2003 (Fig.1) Moreover, 5,390 crabs that

were each injected with a coded wire tag (CWT) [33]

were released at the same site of the Nakara River onJune 27, 2003 A total of 970 crabs with CWTs werestocked in a net enclosure (5-mm mesh size;

20 9 20 9 2 m) without a bottom net set at a tidal flat

of the Nakara River (Fig.1) on June 27, 2003 TheCWT consisted of a small piece of stainless steel wirewith a diameter of 0.25 mm and a length of 1.1 mm, and

it was detected in animals using a magnetic detectiondevice (Northwest Marine Technology) After the initia-tion of the experiments, the mud crabs were recaptured

by hand near the release site at low tide during the day

or at night on June 26, June 29, July 10, August 22 andSeptember 24, 2003 Sampling was also conducted in thenet enclosure at low tide during the day on September

24 and October 22, 2003 The ECW of the capturedcrabs was measured to the nearest 0.1 mm, and theirCWTs were detected with a magnetic detection device.Estimation of size at sexual maturity of mud crabsThe size at sexual maturity was estimated for the dominantmud crab species, S serrata A total of 54 females and 64males were obtained from the above mentioned fishermanfrom May to September 2010 For each crab, the ECW andinternal carapace width (ICW), excluding the anterolateralspines, were measured using a Vernier calliper to thenearest 0.1 mm The highest point of the crushing chela,excluding dorsal spines of males (CH), and the widestportion of the fifth abdominal segments of females (AW)

Fig 1 Iriomote Island The white, black and grey circles indicate the sites for rearing juveniles before release, releasing juveniles, and rearing crabs in the net enclosure after release at the tidal flat in the Nakara River respectively

were also determined for the analysis of maturity classes

(see below)

The abdomen of female brachyuran crabs becomes larger

in width to accommodate an egg incubation chamber at the

pubertal moult [34] Female mud crabs can be assigned to

one of the following three maturity classes based on

abdominal shape: mature females with broad U-shaped

abdominal flaps, immature females with narrow abdominal

flaps, and prepubertal females with intermediate abdominal

shapes between immature and mature forms [8] However,

Hamasaki et al [14] suggested that the intermediate-form

females should be treated as premature because of their low

reproductive ability Therefore, to estimate the maturity size

of S serrata in this study, only two groups (mature and

immature) were considered, and prepubertal females were

considered immature Allometric relationships between

ECW (x) and AW (y) were examined for females at

immature and mature stages by a linear regression equation

using log-transformed data as follows: lny = alnx ?

lnb Parameters were estimated by the ordinary least squares

method Analysis of covariance (ANCOVA) was performed

to detect the differences in slopes and intercepts of the

regression equations between different maturity stages

Although the sexual maturity of male crabs is not easily

determined by external characteristics, allometric change in

the growth of chelae has been analysed to detect the size at

sexual maturity of males [34] The chelae of males have an

important role in pre- and/or post-copulatory mate guarding

of female crabs The chelae markedly increase relative to

body size growth at the pubertal moult Pre- and

post-cop-ulatory mate guarding of females by males also occurs in

mud crabs [35] The morphometric maturity of males was

estimated by the ratio of chela height to body size [8,14,36]

Allometric relationships between ECW (x) and CH (y) were

examined for males according to Sampedro et al [37] and

Corgos and Freire [38] Using the log-transformed data of

ECW and CH, a principal component analysis (PCA) was

performed resulting in two distinguishable groups, which

corresponded to immature and mature crabs With the use of

a non-hierarchical classification procedure (k-means cluster

with two predetermined groups) based on the scores on the

two axes of the PCA, animals were classified as immature or

mature, and the parameters of the linear regression equations

(lny = alnx ? lnb) for the two male groups were then

esti-mated These two regression equations were compared with

ANCOVA

For both females and males, a generalised linear model

consisting of a binomial error distribution with a

logistic-link function [30] was applied to determine the size at

which 50% of the crabs underwent allometric change of

AW or CH The allometric relationship (lny = alnx ? lnb)

between ECW (x) and ICW (x) was also analysed to

cal-culate the ICW from ECW measurements for comparisons

between different studies All analyses were performedusing the R language [29], and the level of significance wasassessed at a = 0.05

ResultsFishery dataTwo species of the mud crab, S serrata and S olivacea,were identified in the fishing activity at Iriomote Island.The total number of crabs caught by traps and gill nets was3,490 and 671 respectively

S serrata was the dominant species, accounting for 95.8and 99.4% of the catches by traps and gill nets respectively.The sex ratio (males to total crabs) was 0.56 for S serrata

in both gear types, and the sex ratio was 0.91 and 1.0 for

S olivacea caught in traps and gill nets respectively A nullhypothesis (H0; sex ratio = 0.5) was rejected by thebinomial test for both species excluding gill nets for

S olivacea with a small sample size (S serrata caught

by trap, P \ 0.0001; S serrata caught by gill net,

P = 0.0011; S olivacea caught by trap, P \ 0.0001; and

S olivacea caught by gill net, P = 0.125) In both species,berried females were not collected The ECW of most ofthe S serrata crabs was between 100 and 170 mm, and theECW of most of the S olivacea crabs was between 100 and

150 mm (Fig 2)

Only four S olivacea crabs were captured using gillnets, so the GAM analysis for species composition wasperformed using trap data The composition of speciesfluctuated yearly and seasonally The number of S olivaceacrabs decreased towards the end of the survey with thenumber largely decreasing after May (Fig.3)

The number of gear units used for both traps and gillnets in addition to the number of fishing operation daysseasonally fluctuated, and they tended to peak betweenJune and August (Fig.4) Data analyses were performedfor the dominant S serrata species In general, the CPUEsincreased when fishing activities were high (Fig.5a) TheGAM analysis detected similar fluctuations in catches of S.serrata by both fishing gear types with higher catchesfound in the first survey year (Fig.5b, d) The catchestended to be low from January/February to March/April;then, they increased each month, peaking in August andMay in traps and gill nets respectively (Fig 5c, e).There were no evident modal progressions in themonthly size-frequency distributions for S serrata (sup-plementary material; Fig S1) with the modes for bothsexes remaining around ECW values of 120 and 160 mmthroughout the study period However, the mean ECWvalues of the females and males seasonally fluctuated(Fig.6a, b) The GAM analysis detected the crab growth

trend (Fig.6c–j) The ECW values of the females and

males caught by traps were smaller in the second and third

survey years when compared with the other years (Fig.6c,

g), and there was no yearly fluctuation of growth in the

crabs caught by gill nets (Fig.6e, i) Similar seasonal

fluctuations were detected in females and males caught

with both gear types (Fig.6d, f, h, i) The ECW of crabs

caught in both gear types decreased from

December/Jan-uary to April/May, and the ECW of crabs caught in the

traps increased after this time peaking in November/

December

The sex ratios of S serrata with sample sizes are shown

in Fig.7a No significant fluctuations by year were found in

the sex ratios of crabs caught with traps, and no significantfluctuations by year and season were found in the sex ratios

of crabs caught with gill nets (Fig.7b, d, e) The proportion

of males caught by traps increased from February to June(Fig.7c)

Release and recapture experiments

At the tidal flat of the Nakara River, a total of 115 mudcrabs were captured Crabs with CWTs were recaptured onJune 26, July 26 and August 22, 2003 with two, four andtwo crabs captured on each date respectively Crabs grewexponentially with ECW values increasing from 21 to

70 mm in 3 months (Fig.8a) In the net enclosure at thetidal flat, the mean ECW of stocked S serrata reached 62and 71 mm on September 24 and October 22, 2003respectively At both survey dates, the retention rate ofCWTs was 84% Crabs with ECW values greater than

100 mm, which was the approximate minimum size caughtwith traps and gill nets, were found in the size-frequencydistributions for animals sampled at the end of the exper-iments in September and October (Fig.8b)

Size at sexual maturityBased on abdomen morphology, immature females hadECW values between 97.3 and 133.1 mm, and maturefemales had ECW values between 128.1 and 195.2 mmrespectively The linear regression equations between log-transformed ECW and AW data for immature and maturefemales had significantly different slopes (F1,50= 6.74 and

P = 0.0123) (Fig.9a) As analysed by ECW, the size atwhich 50% of females reached maturity (SM50) was132.4 mm (95% CI 128.6–136.2 mm) (Fig.10)

Males were classified by PCA and k-means tion analysis into two groups corresponding to morpho-logically immature and mature stages (Fig.9b) Immaturemales had ECW values ranging between 101.0 and160.6 mm, and mature males had ECW values rangingbetween 139.6 and 193.5 mm Mature males had chelaethat were larger than 51.5 mm The slopes of the linearregression equations between log-transformed data corre-sponding to the ECW and CH for immature and maturemales were not significantly different (F1,59= 0.336 and

classifica-P = 0.564), but the intercepts were significantly different(F1,60= 119.7 and P \ 0.0001) As analysed by ECWvalues, the SM50 occurred at 150.7 mm (95% CI146.5–155.0 mm) (Fig.10)

The allometric equations between ECW (x) and ICW(y) were determined for females and males as follows:females, lny = 1.0270lnx - 0.1987 (n = 54, R2= 0.9979,and P \ 0.0001); and males, lny = 1.0434lnx - 0.2707(n = 64, R2= 0.9964, and P \ 0.0001)

Fig 2 Size-frequency distributions for Scylla serrata females and

males caught using traps (a) and gill nets (b) and for Scylla olivacea

females and males caught with both fishing gear types (c)

70 75 80 85 90 95 100

-1.0 -0.5 0.0 0.5 1.0

Fig 3 Species composition of

the monthly catches by traps

(a), and the relationships

between the smoothed

component (solid line) of the

explanatory variables (x-axis,

year and month) used in the

fitted generalised additive

model and the following

response variable: species

composition (b year, c month).

The y-axis is the normalised

effect of the variable with zero

corresponding to no effect of the

covariate on the estimated

response Values greater than

zero indicate a positive

correlation, and values less than

zero indicate a negative

correlation The estimated

degree of freedom (edf) and

F value with probability are

shown in the graph Dashed

lines indicate ±2 SE or

approximately 95% CI

0 100 200 300 400 500 600 700 800 900

0 5 10 15 20 25

Number of gears (trap) Number of gears (gill net) Number of operation days (trap) Number of operation days (gill net)

Fig 4 Fishing activity

represented as the total number

of operation days and the

number of gear units used for

both traps and gill nets

In trap and gill net fisheries at Iriomote Island, two mud

crab species, S serrata and S olivacea, were found

S serrata was the dominant species, accounting for 95.8

and 99.4% of the mud crabs caught with traps and gill netsrespectively This was consistent with the predominance of

S serrata over other mud crab species in tropical oceanicislands, including Palau [12] and Kosrae [13] Keenan et al.[2] suggested that S serrata is the most widespread and

0.1 0.2 0.3 0.4 0.5 0.6

Trap Gill net

(a)

-0.2 -0.1 0.0 0.1 0.2

Month

(c)

-0.5 0.0 0.5

Year

(d)

-0.5 0.0 0.5

Fig 5 Changes in mean (±SE)

catch per unit effort (CPUE) on

a monthly basis for Scylla

serrata caught by traps and gill

nets (a), and the relationships

between the smoothed

component (solid line) of the

explanatory variables (x-axis,

year and month) used in the

fitted generalised additive

model and the following

response variables: CPUE

values for traps (b year,

c month) and gill nets (d year,

e month) More information is

provided in Fig 3

dominant species in oceanic conditions with full salinity

([34%), and other species are commonly associated with

estuarine habitats that have seasonal reductions in salinity

due to rainfall S olivacea is also closely associated with

mangroves and typically occupies burrows within the

mangrove habitat [5, 20, 24] No fishing activities were

observed inside the mangroves located at Iriomote Island

Consequently, the species composition obtained through

fishing activities in the present study may underestimate

the abundance of S olivacea in mud crabs inhabiting

Iriomote Island

Yearly and seasonal fluctuations of the species

composi-tion of the mud crabs caught by traps in Iriomote Island were

found S olivacea decreased year by year with a large

decrease after May of each survey year In the Mekong Delta

located in Vietnam, S paramamosain dominated and

com-prised over 95% of the catch of mud crabs [8] S olivacea

was also present but at a low frequency with a yearly ation [8]; the frequency of S olivacea decreased from 3.2 toless than 0.1% when the mean monthly salinity decreasedfrom 7 to 4 ppt Walton et al [8] also suggested that rela-tively high salinity conditions may be the reason why

vari-S olivacea is the dominant species on the southeast coast butnot in the Mekong Delta in Vietnam The annual amount ofrainfall at Iriomote Island was 1,355, 1,622, 1,957 and3,087 mm for each of the four survey years (September2001–August 2005) respectively, and the mean monthlytotal amount of rainfall largely increased after May as shown

in Fig.11 (data from the Japan Meteorological Agency;

http://www.data.jma.go.jp) Thus, salinity fluctuations may

be one of the causes for yearly and seasonal variations of

S olivacea occurrence at Iriomote Island

The CPUE values for S serrata caught with both types

of fishing gear were low during the period from January/

10 11 12 13 14 15 16 17

(f)

Month

-1.0 -0.5 0.0 0.5 1.0

Oct Nov Dec Jan Feb Mar Apr May

-0.5 0.0 0.5

(i)

-0.5 0.0 0.5

-0.4

(h)

-0.2 0.0 0.2 0.4 0.6

Sep Nov Jan Mar May Jul

Fig 6 Changes in mean (±SE) external carapace width (ECW) on a

monthly basis for female (a) and male (b) Scylla serrata caught by

traps and gill nets and the relationships between the smoothed

component (solid line) of the explanatory variables (x-axis, year and

month) used in the fitted generalised additive model and the following

response variables: ECWs for females caught by traps (c year,

d month) and gill nets (e year, f month); and ECWs for males caught

by traps (g year, h month) and gill nets (i year, j month) More information is provided in Fig 3

February to March/April, but the values increased after this

period There were no available data for water temperature

at Iriomote Island, but the temperature data at the

neigh-bouring Ishigaki Island (Fig.1) (data from the Japan

Oceanographic Data Center;http://www.jodc.go.jp) could

represent the temperatures present at Iriomote Island

(Fig.11) Thus, the catch rate of the crabs was generally

low during the winter when temperatures were lower, and itincreased during the summer when temperatures increased.The increased catch rate in the summer has been previouslyreported for mud crab species [6,39]

Modal progressions in the monthly size-frequency tributions of S serrata caught with both fishing gear types

dis-in Iriomote Island were not observed However, clear

0 20 40 60 80 100 120 140 160 180 200

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

(a)

-0.2 -0.1 0.0 0.1 0.2 0.3

(b)

Year

-0.2 -0.1 0.0 0.1 0.2 0.3

Month

(c)

-0.5 0.0 0.5

Fig 7 Sex ratio (SR) of the

monthly catches for Scylla

serrata with number of crabs

(N) caught by traps and gill nets

(a) and the relationships

between the smoothed

component (solid line) of the

explanatory variables (x-axis,

year and month) used in the

fitted generalised additive

model and the following

response variables: SRs for

traps (b year, c month) and gill

nets (d year, e month) More

information is provided in

Fig 3

seasonal variations were detected in the mean ECW values

of S serrata caught with both gear types, indicating a high

growth rate during the summer from April to May In

contrast, decreases in the mean ECW values occurred

during the winter from December to January The

follow-ing reasons may explain the decreased mean ECW values:

natural death of larger (older) crabs and/or recruitment of

young crabs to the fishery In the release and recapture

experiments at the natural tidal flat in this study, some of

the juveniles, which moulted to the first and second crab

stages in late May, reached a recruitment size (ECW) of

100 mm after 4 months Captive S serrata females

obtained from Iriomote Island spawn year round according

to laboratory experiments, but main spawning is observed

in the warmer season from April to November, peaking

around August [40, 41] Furthermore, the larval

develop-mental period from hatching to the first crab stage ranges

between 22 and 42 days at water temperatures ranging

between 23 and 32°C [31] Together, these findings suggest

that the decreased mean ECW values found from

December/January to April/May are due to the recruitment

of young crabs to the fishery

Yearly fluctuations were observed in the catch rate andbody size of S serrata in Iriomote Island As compared tothe other survey years, the catch rate was higher in the firstsurvey year with both fishing gear types Compared to the

Fig 8 Results of release and recapture experiments using artificially

produced Scylla serrata juveniles Mean (±SE) external carapace

width (ECW) of crabs captured over time after being released at the

natural tidal flat in the Nakara River (a) The size-frequency

distributions of the crabs recovered in September and October from

the natural tidal flat and the net enclosure set at the natural tidal flat in

the Nakara River (b) Numbers in the graph (a) indicate the number of

samples captured and measured

Immature: y = 1.6403x - 4.1569 (n = 42, R2= 0.8983, P < 0.0001)

Mature: y = 1.3209x - 2.4464 (n = 12, R2= 0.9694, P < 0.0001)

3 3.2 3.4 3.6 3.8 4 4.2 4.4 4.6 4.8

(a)

Immature Mature

Immature: y = 1.3595x - 2.9848 (n = 36, R2= 0.8488, P < 0.0001)

Mature: y = 1.2005x - 1.9526 (n = 28, R2= 0.7414, P < 0.0001)

3 3.2 3.4 3.6 3.8 4 4.2 4.4 4.6

car-0 0.5 1

External carapace width (mm)

Female Male Female Male

Fig 10 Maturity or immaturity of Scylla serrata females and males according to their external carapace width

first and fourth survey years, the body size of crabs caught

with the trap fishing gear was smaller in the second and

third survey years Water temperature affects the catch rate

of mud crabs, and it may affect the growth rate of the crabs

[42] Water temperature data at Ishigaki Island are

avail-able from January 2002 onward, but they are not availavail-able

between September and December in 2001 Consequently,

to compare water temperature data between different

sur-vey years in this study, the mean water temperature from

January to August was used and calculated to be 21.6, 21.7,

22.0 and 21.7°C for each year in the four survey years

respectively According to these calculations, the mean

water temperature did not vary greatly between years, and

it did not affect yearly fluctuations of catch and growth of

S serrata at Iriomote Island during the study period

However, there may have been another possible cause of

the smaller ECW values found in the second and third

survey years Artificially produced S serrata juveniles

were stocked in the Nakara and Kuira Rivers at Iriomote

Island in late July 2001 (7,789 crabs with a mean ECW of

29 mm), early June 2002 (35,410 crabs with a mean ECW

of 25 mm), and late June 2003 (28,650 crabs with a mean

ECW of 23 mm) (K Hamasaki; unpublished data from

2001, 2002 and 2003), showing higher numbers of releasedjuveniles in 2002 and 2003 than in 2001 Crabs in the first

to second stage were recruited to the fishery after 4 months.Therefore, the stock enhancement programme may havedecreased the mean ECW of the crabs by increasing therecruitment of small crabs Yearly fluctuations of S serratacatch numbers and body size should be evaluated usingmore long-term data that includes variable ocean envi-ronmental conditions

The overall sex ratio of S serrata caught in traps andgill nets in Iriomote Island was slightly male-biased (0.56)

In addition, seasonal variation of the sex ratio was observed

in traps with high values from February to June, indicatingthat females were inactive and/or males were active andcould be caught in baited traps in this period Similarly,overall and seasonal male-biased sex ratios have beenreported for S olivacea in the Andaman Sea (Thailand)depending on offshore migration of females for spawning[7] In this study, no ovigerous females were caught

S serrata females are expected to migrate offshore forspawning, which has been reported for mud crab species[7,43,44] all year round, especially in the main spawningperiod from April to November [40, 41] Female feedingactivity may decrease before offshore spawning and/ormales may become active for copulating with females Inthis study, the sex ratio of S olivacea, which was mainlycaught using traps, was highly biased towards males (0.96).Such a heavy male-biased sex ratio may be a specific trait

of the local fishery because it has not been previouslyreported for mud crab species [6,7,11]

The discontinuity of the two regression lines of carapacewidth (ECW) and AW in females demonstrated theoccurrence of a puberty moult over the ECW range

of 128.1–133.1 mm (ICW ranging between 119.7 and124.5 mm, which was calculated from a regression equa-tion between log-transformed data of ECW and ICW in thepresent study), and the ECW at which 50% of femalesreached morphological maturity (SM50) was estimated at132.4 mm (ICW of 123.8 mm) for S serrata at IriomoteIsland The SM50value in the present study was larger thanthat estimated for mud crab females from South Africa(ECW of 123 mm) [35] where S serrata is the only speciespresent [2] The SM50 of S serrata females in the presentstudy was also larger than that estimated for S parama-mosain (ICWs of 110.5 and 105.6 mm in Bandon Bay,Thailand [5,14]; and ICW of 102.3 mm in Mekong Delta,Vietnam [8]) and S olivacea (ICW of 91.2 mm in BandonBay [5]; and ECW of 95.5 mm in Klong Ngao mangroveswamp in Andaman Sea, Thailand [15]) As for malesexual maturity, the ECW/CH relationships suggested thatthe two different morphometric forms (mature and imma-ture) overlapped between ECWs of 139.6 and 160.6 mm

Fig 11 Mean (±SE) monthly total amount of rainfall from

Septem-ber 2001 to August 2005 at Iriomote Island (a) and mean (±SE) water

temperature each month from January 2002 to December 2005 at

Ishigaki Island (24°20 0 N and 124°08 0 E) (b) (data from the Japan

Meteorological Agency; http://www.data.jma.go.jp )

(ICWs between 132.0 and 152.7 mm), and the SM50was

estimated at an ECW of 150.7 mm (ICW of 142.9 mm)

Robertson and Kruger [35] evaluated the functional

maturity size of South African S serrata males by the

presence of mating scars on the front of the first pair of

walking legs, which are formed by the rubbing of the

female carapace on the legs of the male during the

pre-copulatory embrace They reported that mating scars on

males were most frequently found on larger animals with

ECWs greater than 135 mm Knuckey [36] estimated the

morphological maturity size at ICW values between 146

and 149 mm by the ratio of chela height to the ICW of

males Moreover, in mud crabs from Australia, where the

S serrata species is dominant [2,4], mating scars are more

common in males with ICWs greater than 140 mm [36]

Thus, the maturity size of S serrata males at Iriomote

Island was similar to that of Australian mud crabs, which

have a larger maturity size than the South African mud

crabs The SM50 of S serrata males was larger than that

estimated for S paramamosain (ICW of 101.9 mm in

Mekong Delta [8] and ECW of 106.4 mm in Bandon Bay

[14])

Immature females and males, which had overall

size-frequency distributions less than the SM50 classes of

S serrata (Fig.2), comprised approximately 37 and 66%

of the catch by traps respectively, and 41 and 64% of the

catch by gill nets respectively Both fishing gear types

caught more immature male crabs than female crabs

Prohibitions on the capture of immature crabs based on the

SM50values estimated for S serrata females and males in

the present study should be implemented in the fishery

management of S serrata to sustain the resources of

Irio-mote Island Thus, this study suggested that the minimum

size for capture, as measured by ECWs, should be 140 mm

for females and 160 mm for males (higher than the upper

limit of 95% CI for the estimates of SM50) to allow almost

all individuals to achieve a functional maturity size

(Fig.10)

Acknowledgments We gratefully acknowledge the help of

Mr Choji Ishigaki for collecting fishery data during the surveys.

References

1 Macnae W (1968) A general account of the fauna and flora of

mangrove swamps and forests in the Indo-West-Pacific region.

Adv Mar Biol 6:73–270

2 Keenan CP, Davie PJF, Mann DL (1998) A revision of the genus

Scylla de Haan, 1833 (Crustacea: Decapoda: Brachyura:

Portu-nidae) Raffles Bull Zool 46:217–245

3 Imai H, Cheng JH, Hamasaki K, Numachi K (2004) Identification

of four mud crab species (genus Scylla) using ITS-1 and 16S

rDNA markers Aquat Living Res 17:31–34

4 Le Vay L (2001) Ecology and management of mud crab Scylla spp Aisan Fish Sci 14:101–111

5 Overton JL, Macintosh DJ (2002) Estimated size at sexual maturity for female mud crabs (genus Scylla) from two sympatric species within the Ban Don Bay, Thailand J Crust Biol 22:790–797

6 Pillans S, Pillans RD, Johnstone RW, Kraft PG, Haywood MDE, Possingham HP (2005) Effects of marine reserve protection on the mud crap Scylla serrata in a sex-biased fishery in subtropical Australia Mar Ecol Prog Ser 295:201–213

7 Koolkalya S, Thapanand T, Tunkijjanujij S, Havanont V, Jutagate

T (2006) Aspects in spawning biology and migration of the mud crab Scylla olivacea in the Andaman Sea, Thailand Fish Manag Ecol 13:391–397

8 Walton ME, Le Vay L, Truong LM, Ut VN (2006) Significance

of mangrove-mudflat boundaries as nursery grounds for the mud crab, Scylla paramamosain Mar Biol 149:1119–1207

9 Walton ME, Le Vay L, Lebata JH, Binas J, Primavera JH (2006) Seasonal abundance distribution and recruitment of mud crabs (Scylla spp.) in replanted mangroves Estuar Coast Shelf Sci 66:493–500

10 Walton ME, Le Vay L, Lebata JH, Binas J, Tapper J, Primavera

JH (2007) Effectiveness of mangrove replanting in the restoration

of fisheries: abundance and distribution of mud crabs, Scylla olivacea in intact, degraded and rehabilitated mangroves Biol Conserv 138:180–188

11 Lebata MJHL, Le Vay L, Primavera JH, Walton ME, Bin˜as JB (2007) Baseline assessment of fisheries for three species of mud crabs (Scylla spp.) in the mangroves of Ibajay, Aklan, Philip- pines Bull Mar Sci 80:891–904

12 Ewel KC (2008) Mangrove crab (Scylla serrata) population may sometimes be best managed locally J Sea Res 59:114–120

13 Bonine KM, Bjorkstedt EP, Ewel CK, Palik M (2008) Population characteristics of the mangrove crab Scylla serrata (Decapoda: Portunidae) in Kosrae, Federated States of Micronesia: effects of harvest and implications for management Pac Sci 62:1–19

14 Hamasaki K, Matsui N, Nogami M (2011) Size at sexual maturity and body size composition of mud crabs Scylla spp caught in Don Sak, Bandon Bay, Gulf of Thailand Fish Sci 77:49–57

15 Jirapunpipat K (2008) Population structure and size at maturity of the orange mud crab Scylla olivacea in Klong Ngao mangrove swamp, Ranong Province, Thailand Kasetsart J (Natl Sci) 42:31–40

16 Jirapunpipat K, Yokota M, Watanabe S (2009) The benefits of species-based management of sympatric mud crabs migrating to a common fishing ground ICES J Mar Sci 66:470–477

17 Stephenson W, Campbell B (1960) The Australian portunids (Crustacea: Portunidae) IV: remaining genera Aust J Mar Freshw Res 11:73–122

18 Fuseya R, Watanabe S (1996) Genetic variability in the mud crab genus Scylla (Brachyura: Portunidae) Fish Sci 62:705–709

19 Fushimi H, Watanabe S (2000) Problems in species identification

of the mud crab genus Scylla (Brachyura: Portunidae) In: Tamaru CC-T et al (eds) Spawning and maturation of aquacul- ture species UJNR Tech Rep 28:9–13

20 Estampador EP (1949) Studies on Scylla (Crustacea: Portunidae).

I Revision of the genus Philipp J Sci 78:95–108

21 Sere`ne R (1952) Les espe`ces du genre Scylla a` Nhatrang nam) Proc IPFC 3:133–137 (in French with English abstract)

(Viet-22 Ong KS (1964) The early developmental stages of Scylla serrata Forska˚l (Crustacea: Portunidae), reared in the laboratory Proc IPFC 11:135–146

23 Joel DR, Raj PJS (1980) Taxonomic remarks on two species of the genus Scylla de Haan (Portunidae: Brachyura) from Pulicat Lake J Inland Fish Soc India 12:39–50

24 Oshiro N (1991) Mangrove crabs (Scylla spp.) In: Shokita S et al

(eds) Aquaculture in tropical areas Midori Shobo, Tokyo,

pp 218–229

25 Ito M (2000) Some observations on the release and catch of mud

crabs (Scylla spp.) in Lake Hamana, Pacific coast of central Japan

on the basis of past records Saibai Giken 28:57–64 (in Japanese)

26 Obata Y, Imai H, Kitakado T, Hamasaki K, Kitada S (2006) The

contribution of stocked mud crabs Scylla paramamosain to

commercial catches in Japan, estimated using a genetic stock

identification technique Fish Res 80:113–121

27 Keenan CP (1999) The fourth species of Scylla In: Keenan CP,

Blackshaw A (eds) Mud crab aquaculture and biology ACIAR

Proceedings no 78 Australian Centre for International

Agricul-tural Research, Canberra, pp 48–58

28 Wood SN (2006) Generalized additive models: an introduction

with R Chapman and Hall/CRC, Boca Raton

29 R Development Core Team (2011) R: a language and

environ-ment for statistical computing R Foundation for Statistical

Computing, Vienna

30 Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009)

Mixed effects models and extensions in ecology with R Springer,

New York

31 Hamasaki K (2003) Effects of temperature on the egg incubation

period, survival and developmental period of larvae of the mud

crab Scylla serrata (Forska˚l) (Brachyura: Portunidae) reared in

the laboratory Aquaculture 219:561–572

32 Hamasaki K, Suprayudi MA, Takeuchi T (2002) Mass mortality

during metamorphosis in the seed production of mud crab

Scylla serrata (Crustacea, Decapoda, Portunidae) Fish Sci

68:1226–1232

33 Jefferts KB, Bergman PK, Fiscus HF (1963) A coded wire tag

identification system for macroorganisms Nature 198:460–462

34 Hartnoll RG (1974) Variation in growth pattern between some

secondary sexual characters in crabs (Decapoda Brachyura).

Crustaceana 27:131–136

35 Robertson WD, Kruger A (1994) Size at maturity, mating and spawning in the portunid crab Scylla serrata (Forska˚l) in Natal, South Africa Estuar Coast Shelf Sci 39:185–200

36 Knuckey IA (1996) Maturity in male mud crabs, Scylla serrata, and the use of mating scars as a functional indicator J Crust Biol 16:487–495

37 Sampedro MP, Gonza´lez-Gurriara´n E, Freire J, Muin˜o R (1999) Morphometry and sexual maturity in the spider crab Maja squi- nado (Decapoda: Majidae) in Galcia, Spain J Crust Biol 19:578–592

38 Corgos A, Freire J (2006) Morphometric and gonad maturity in the spider crab Maja brachydactyla: a comparison of methods for estimating size at maturity in species with determinate growth ICES J Mar Sci 63:851–859

39 Williams MJ, Hills BJ (1982) Factors influencing pot catches and populations estimates of the portunid crab Scylla serrata Mar Biol 71:187–192

40 Kaji S (1988) Broodstock management of mud crabs Scylla serrata Annual report of Japan Sea Farming Association in fiscal year 1986 Japan Sea Farming Association, Tokyo, pp 74–78 (in Japanese)

41 Tezuka N (1989) Broodstock management of mud crabs Scylla serrata Annual report of Japan Sea Farming Association in fiscal year 1987 Japan Sea Farming Association, Tokyo, pp 64–67 (in Japanese)

42 Leffler CW (1972) Some effects of temperature on the growth and metabolic rate of juvenile blue crabs, Callinectes sapidus, in the laboratory Mar Biol 14:104–110

43 Hyland SJ, Hill BJ, Lee CP (1984) Movement within and between different habitats by the portunid crab Scylla serrata Mar Biol 80:57–61

44 Hill BJ (1994) Offshore spawning by the portunid crab Scylla serrata (Crustacea: Decapoda) Mar Biol 120:379–384

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

Molecular characterization and expression analysis of heat shock

proteins 40, 70 and 90 from kuruma shrimp Marsupenaeus

japonicus

Threechada Danwattananusorn•Fernand F Fagutao•

Aiko Shitara•Hidehiro Kondo•Takashi Aoki•

Reiko Nozaki•Ikuo Hirono

Received: 24 May 2011 / Accepted: 19 July 2011 / Published online: 14 October 2011

Ó The Japanese Society of Fisheries Science 2011

Abstract Heat shock proteins (HSPs) are proteins that

are expressed more strongly when the cells are exposed to

physiological and stressful conditions In this study, the

full-length cDNAs of heat shock proteins 40 (MjHSP40),

70 (MjHSP70) and 90 (MjHSP90) were cloned from

kuruma shrimp Marsupenaeus japonicus The open reading

frames (ORFs) of the cDNA clones have lengths of 1,191,

1,959 and 2,172 bp and encode 396, 652 and 723 amino

acid residues, respectively The predicted MjHSP40 amino

acid sequence contains a J domain, a

glycine/phenylala-nine-rich region, and a central domain containing four

repeats of a CxxCxGxG motif, indicating that it is a type I

HSP40 homolog The signature sequences of the HSP70

and HSP90 gene families are conserved in the MjHSP70

and MjHSP90 amino acid sequences The deduced amino

acid sequences of MjHSP70 and MjHSP90 share high

identity with previously reported shrimp HSP70s and

HSP90s, respectively The expression of MjHSP90 mRNA

increased at 32°C Additionally, the expressions of

MjHSP40, MjHSP70 and MjHSP90 mRNAs increased in

defense-related tissues (i.e., hemocytes and lymphoid

organ) when the shrimp were challenged with white spot

T Danwattananusorn  F F Fagutao  A Shitara  H Kondo 

T Aoki  R Nozaki  I Hirono (&)

Laboratory of Genome Science, Tokyo University of Marine

Science and Technology, Konan 4-5-7, Minato-ku,

Tokyo 108-8477, Japan

e-mail: hirono@kaiyodai.ac.jp

DOI 10.1007/s12562-011-0394-z

to enable HSP70 to perform its essential cellular function

[9] Recent studies suggest a mechanism where the substrate

appears to be released from HSP40 and transferred to

HSP70 [10–12] Furthermore, HSP40 proteins can regulate

the activities of other chaperones, such as HSP90 [13]

HSP70s play a role in protein synthesis under normal

cel-lular conditions, fixing denatured proteins and preventing the

misfolding or aggregation of proteins [14–16] The HSP70

family is composed of several members, including

heat-inducible HSP70, constitutively expressed heat shock cognate

70 (HSC70), glucose-regulated protein (GRP78), and others

[17] HSC70s, unlike HSP70s, have no introns However, they

share common structural features, including a 44-kDa

N-ter-minal ATPase domain, an 18-kDa peptide-binding domain,

and a 10-kDa C-terminal substrate-binding domain [17,18]

HSP90 participates in the folding, maintenance of

struc-tural integrity, and the proper regulation of a subset of

cytosolic proteins [19], and accounts for 1% of the soluble

protein in most tissues, even in the absence of stress [20]

HSP90 also functions as a specialized chaperone for a set of

signaling proteins, including several protein kinases and

transcription factors [21] HSP90 has roles in cell growth and

differentiation, apoptosis, signal transduction and cell–cell

communication Eukaryotic HSP90 proteins consist of three

domains: a 25-kDa N-terminal ATP-binding domain, a

40-kDa middle domain, and a 12-kDa C-terminal

dimeri-zation domain The N-terminal ATP-binding domain is

connected to the middle domain by a ‘‘linker’’ of variable

length, and the C-terminal dimerization domain provides the

binding site for a set of co-chaperone molecules that function

with HSP90 as part of a multi-chaperone complex [21]

In recent years, there has been increasing interest in

shrimp HSPs (mostly HSP70 and HSP90), because of their

roles in shrimp immune response HSP70 genes and their

expressions have been studied in Chinese shrimp

Fenne-ropenaeus chinensis [22], Pacific white shrimp Litopenaeus

vannamei [23,24], and black tiger shrimp Penaeus

mon-odon [25,26], while the expressions of HSP90 genes have

been studied in F chinensis [27], P monodon [28], and

greasyback shrimp Metapenaeus ensis [29] In contrast, we

were unable to find reports on HSP40 in shrimp In the

present study, HSP40, HSP70 and HSP90 were cloned

from kuruma shrimp (Marsupenaeus japonicus), and their

expressions were examined after heat shock and being

challenged with white spot syndrome virus (WSSV)

Materials and methods

Shrimp

The kuruma shrimp used in this study were purchased from

a commercial shrimp farm in Miyazaki, Japan The shrimp

were analyzed for signs of infectious diseases, kept inartificial seawater maintained at 25°C and 30–32 ppt, andfed daily with commercial shrimp feed prior to all experi-mental procedures

Cloning of MjHSP40, MjHSP70 and MjHSP90The MjHSP40, MjHSP70 and MjHSP90 cDNAs wereamplified from a normal kuruma shrimp hepatopancreascDNA library prepared in our laboratory The pairs ofspecific primers designed based on the partial cDNAsequences of MjHSP40, MjHSP70 and MjHSP90 previ-ously identified in our laboratory were used for PCRamplification (Table1) Moreover, the pair of primersused to amplify MjHSP70 were also designed based onthe cDNA sequence of kuruma shrimp HSP70 (GenBankaccession no ABK76338) The PCR reaction was per-formed as follows: an initial denaturation at 95°C for

5 min, followed by 35 cycles of 95°C for 30 s, 55°C for

30 s, and 72°C for 1 min, and a final extension at 72°Cfor 5 min After electrophoresis on 1% agarose gel, thePCR products were subsequently purified The purifiedDNAs of the PCR products were ligated into the pGEM-Teasy vector (Promega, USA) and were then transformedinto Escherichia coli strain JM109 The positive cloneswere screened by colony PCR with M13 forward andreverse primers The subsequent PCR products weredirectly sequenced with an ABI 3130xl capillarysequencer using BigDye chemistry (Applied Biosystems,USA)

Sequence data analysisNucleotide and amino acid sequences were analyzed usingGENETYX WIN (v.7.0.3) software Homology analysisand cleavage site prediction were accomplished withBLASTP (see http://ncbi.nlm.nih.gov/) Conceptual trans-lation was performed and the characteristics of the proteinwere predicted using the ExPASy web server (http://www.expasy.ch) The motifs were predicted using SMART(http://smart.embl-heidelberg.de/) [30], and subcellularlocalization predictions were performed on the PSORT IIsever (http://psort.hgc.jp/form2.html) Multiple sequencealignments were created using ClustalW [31] The neigh-bor-jointing phylogenetic trees were then generated byMEGA4 software [32]

Expression analysis after heat shock and white spotsyndrome virus (WSSV) challenge experiments

In order to study the MjHSP40, MjHSP70 and MjHSP90expression after heat shock treatment, apparently healthy

kuruma shrimp, each weighing about 10 g, were acclimated

to a salinity of 30–32 ppt at 25°C for 7 days before

exper-imentation, and then heat shock treatment was performed at

32°C for 3 h Gills were dissected from three shrimp

sam-pled before the heat shock treatment as an initial control,

and at 1 and 3 h after heat shock After this, MjHSP40,MjHSP70 and MjHSP90 mRNA expression levels wereexamined by quantitative real-time PCR Total RNAs wereextracted using the RNAiso reagent (TaKaRa Bio Inc.,Japan), as described in the manufacturer’s protocol

Table 1 PCR primers and

primer sequences used for the

amplification of MjHSP40,

MjHSP70 and MjHSP90

cDNAs, and the quantitative

real-time PCR and RT-PCR

primers used in the experiment

HSP40-F2 GTTAAGGATGTCATCCATCAA Cloning of MjHSP40 partial cDNA sequence HSP40-R2 AATGCTAGTTTTCTGTATGCC Cloning of MjHSP40 partial cDNA sequence

HSP40-F4 TTCCCGGGGTCGACCCACGCGTC Cloning of MjHSP40 partial cDNA sequence HSP40-R4 CGTAAGCTTGGATCCTCTAGAG Cloning of MjHSP40 partial cDNA sequence

HSP70-F2 GTATCTTCGAAGTAAAGTCCA Cloning of MjHSP70 partial cDNA sequence HSP70-R2 TGTAGAAGTCGATACCTTCGA Cloning of MjHSP70 partial cDNA sequence HSP70-F3 GAAGTCACCTTCGACATCGAC Cloning of MjHSP70 partial cDNA sequence HSP70-R3 TCAACTGTCGACTTCATGTTG Cloning of MjHSP70 partial cDNA sequence HSP70-R4 TTAATCGACCTCCTCGATGGT Cloning of MjHSP70 partial cDNA sequence

HSP90-F2 ACAGTACATCTGGGAGTCGTC Cloning of MjHSP90 partial cDNA sequence HSP90-R2 TCAGGTACTCGGGGATCAGTT Cloning of MjHSP90 partial cDNA sequence

HSP90-R3 GCTTCTGGTTCTCCTTCATGC Cloning of MjHSP90 partial cDNA sequence HSP90-F4 ACCATGGGCTACATGGCCGCCA Cloning of MjHSP90 partial cDNA sequence HSP90-R4 CCTTCACAGACTTGTCGTTCT Cloning of MjHSP90 partial cDNA sequence

MjHSP40s-F CGGAGAAGTTCTAGACCAAGATGGT Quantitative real-time PCR and RT-PCR MjHSP40s-R GTGGGCTTCACCCCTAGGAT Quantitative real-time PCR and RT-PCR

MjHSP70s-R GAAGGGCCAGTGCTTCATGT Quantitative real-time PCR and RT-PCR

MjHSP90s-F TCGCTGAATTCCTCAGATACCA Quantitative real-time PCR and RT-PCR MjHSP90s-R GCACTCCTTGAGAGAAGACATATCG Quantitative real-time PCR and RT-PCR

The first-strand cDNAs were synthesized from 2 lg of the

total RNAs using Moloney murine leukemia virus reverse

transcriptase (Invitrogen, USA) Primer sets for MjHSP40,

MjHSP70 and MjHSP90 were designed with ABI Primer

Express Software v.3.0 (Applied Biosystems, USA), as

shown in Table1 Quantitative real-time PCR assays were

done in a 20 ll reaction volume consisting of 5 ll template

cDNA (2 lg/ml), 0.4 ll of both forward and reverse

primers (10 pM), 10 ll Power SYBR Green Master Mix

(Applied Biosystems), and 4.2 ll distilled water Real-time

PCR analysis was performed on an ABI7300 real-time PCR

system (Applied Biosystems), following the manufacturer’s

protocol Expression levels were measured by the 2DDCT

method [33] Elongation factor-1 a (EF-1a) mRNA

levels were used as an internal control Differences in

expression were measured by one-way analysis of variance

followed by the Tukey significant difference test using

SPSS 16.0 software P values of \0.05 were considered

significant

In addition, the MjHSP40, MjHSP70 and MjHSP90

mRNA expression levels were examined after the WSSV

challenge experiment Apparently healthy kuruma

shrimp, each weighing about 10 g, were acclimated to

laboratory conditions for 7 days and then injected with

50 ll of 1069 WSSV stock suspension This virus stock

dilution was used based on an earlier in vivo viral

titration assay which suggested that this particular

dilu-tion is optimal for use in challenge experiments

Hemocytes, lymphoid organ and hepatopancreas were

collected from three shrimp sampled before the WSSV

challenge experiment as an initial control and at 1, 3 and

5 days post-WSSV injection, and total RNA was

extracted and pooled together at each sampling time

Then the first-strand cDNA was synthesized as described

above RT-PCR was performed to analyze the expression

profiles of MjHSP40, MjHSP70 and MjHSP90 using

MjHSP40s-F, MjHSP40s-R, MjHSP70s-F, MjHSP70s-R,

MjHSP90s-F and MjHSP90s-R (see Table1) EF-1a was

amplified as an internal control using EFs-F and EFs-R

(see Table1) One microliter of the first-strand cDNA

was used as the template in the PCR amplification

The RT-PCR reaction was conducted with an initial

predenaturation step performed at 95°C for 5 min

followed by 25 and 30 cycles of 95°C for 30 s, 55°C for

30 s and 72°C for 30 s, and a final extension at 72°C for

5 min Ten microliters of the amplified products were

separated by electrophoresis with a 1% agarose gel and

visualized with ethidium bromide The mRNA bands

were semi-quantitatively assessed for their relative

expression following the method described by

Linden-strøm et al [34] using ImageJ software to measure light

intensity [35]

ResultsCharacterization of MjHSP40, MjHSP70 and MjHSP90genes in kuruma shrimp

The full-length MjHSP40, MjHSP70 and MjHSP90 cDNAsequences (GenBank accession no AB520825, AB520826and AB520827, respectively) from kuruma shrimp wereobtained by PCR amplification For MjHSP40, thesequence consists of 1,191 nucleotides of an open readingframe (ORF) encoding 396 amino acids with a calculatedmolecular weight of 44.42 kDa and a theoretical pI of6.62 The deduced amino acid sequence of MjHSP40contains an N-terminal conserved domain (J domain, aa5–60), a glycine/phenylalanine region (G/F domain, aa67–96), a central domain containing four highly con-served cysteine-rich repeats with a consensus sequence ofCxxCxGxG where x is any amino acid (CRR domain, aa122–207), and a C-terminal domain (C domain, aa220–344) (Fig 1) In a GenBank BLASTP search, thededuced amino acid sequence of MjHSP40 showed highhomology with those of other invertebrates: domesticsilkworm Bombyx mori (NP_001040292, 69% identity),red flour beetle Tribolium castaneum (XP_971446, 64%),and jewel wasp Nasonia vitripennis (XP_001607240,64%) MjHSP40 also had high similarities to DnaJ(Hsp40) homolog subfamily A member 1 from humanHomo sapiens (NP_001530, 66%), heat shock protein 40from American alligator Alligator mississippiensis(BAF94139, 65%), and DnaJ-like subfamily A member 4from zebrafish Danio rerio (XP_689328, 62%) (Fig.2).The deduced amino acid sequence of MjHSP70 was found

to be very similar to M japonicus HSP70s in the Bank (accession nos ABF83607 and ABK76338) Sinceboth submissions were unpublished, we first verified thesequences of these genes

Gen-The deduced amino acid sequences of MjHSP70 alsocontain HSP70 family motifs and signatures, including anadenosine triphosphate/guanosine triphosphate (ATP/GTP)-binding site, a bi-partite nuclear localization signal, a non-organellar motif, and a conserved EEVD motif MjHSP70displayed very high homology with shrimp HSP70s from

L vannamei (AAT46566, 99%), P monodon (AAQ05768,99%), and M ensis (ABF20530, 97%) Furthermore, MjHSP70 showed high similarities to HSP70 from Americanlobster Homarus americanus (ABA02165, 96%), marbledcrab Pachygrapsus marmoratus (ABA02164, 94%), andpearl oyster Pteria penguin (ABJ97377, 86%)

The MjHSP90 cDNA contains a 2,172-bp ORF thatencodes 723 amino acids with a calculated molecularweight of 83.6 kDa and a theoretical pI of 4.92 Thededuced amino acid sequence of MjHSP90 also contains