Fisheries science JSFS , tập 78, số 5, 2012 9

The main ideas behind the sorting box concept are 1 to move the window as far back as possible to improve the release of cod but prevent the accumulating Norway lobster catch from coming

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

Improving the effectiveness of escape windows in directed Norway

lobster Nephrops norvegicus trawl fisheries

Niels Madsen•Rene´ Holst •Rikke Petri Frandsen•

Ludvig A Krag

Received: 14 December 2011 / Accepted: 22 May 2012 / Published online: 30 June 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract A substantial improvement in the bycatch

selectivity of Norway lobster Nephrops norvegicus trawls

is required, particularly with respect to cod Gadus morhua,

whose stocks are at low levels in several areas

Conven-tional escape windows are not adequate to properly release

cod and other bycatch species caught in the trawls To

address this issue, we developed a novel sorting box

con-cept consisting of a four-panel section with a window on

the top in order to improve the escape of cod and other

bycatch species through an escape window while retaining

the target catch of Norway lobster The concept was tested

on a commercial trawler in Kattegat and Skagerrak Two

different window mesh sizes and two different sorting box

heights were tested using a traditional codend cover and a

dual codend cover We observed greatly reduced bycatches

of both cod and other fish species compared to a standard

codend The reduction in bycatch decreased with

decreas-ing mesh size and increasdecreas-ing height of the sortdecreas-ing box

Escape of Norway lobster through the escape window was

limited A modified version of the sorting box concept was

implemented in the Kattegat fishery from 2009 onwards

Keywords Cod Fishery management  Norway lobster 

Plaice Selectivity  Trawl  Bycatch

IntroductionNorway lobster Nephrops norvegicus is a major targetspecies in several areas in the North East Atlantic, andrelatively small codend mesh sizes are used compared tothe codends used in whitefish fisheries Consequently,Norway lobster fisheries are characterized by a relativelyhigh bycatch of juvenile fish species and high discard rates.This problem is rather universal among North East AtlanticNorway lobster fisheries, and trawls with improved selec-tivity properties have recently been tested in Portuguesewaters [1], the Bay of Biscay [2], the North Sea [3 8], theIrish Sea [9], Icelandic waters [10], and the Kattegat andSkagerrak [11–15]

Bycatch of Atlantic cod Gadus morhua is a particularproblem, since stocks have decreased in several areas [16,

17] In the Kattegat, the cod stock is at a critically low level[14], and measures have been taken to rebuild it, includingdesignating seasonally protected areas where only veryselective fishing gear is allowed [14] The use of a sortinggrid is an option in the Norway lobster fishery under cur-rent legislation in Skagerrak and Kattegat [12–14] and hasalso been tested recently in other Norway lobster fisheries[2, 3, 5, 8] While sorting grids are very effective atallowing cod to escape [12–14], they are more difficult tohandle onboard the small vessels that typically operate inthis area, and fish and debris can block the grid Further-more, losses of Norway lobster, particularly the larger andmore valuable individuals, have been observed [13] Ingeneral, Danish vessels in Kattegat and Skagerrak have notused the Norway lobster grids that have been permitted bythe legislation since 2005, even though the use of thesegrids allows unlimited days at sea, whereas there have beensevere restrictions on using less selective gear The square-mesh escape window (henceforth window) is one of the

N Madsen ( &)  R P Frandsen  L A Krag

DTU Aqua, National Institute of Aquatic Resources,

North Sea Science Park, P.O Box 101,

9850 Hirtshals, Denmark

e-mail: nm@aqua.dtu.dk

R Holst

Institute of Regional Health Service Research,

University of Southern Denmark, J.B Winsløwsvej 9b,

5000 Odense C, Denmark

Fish Sci (2012) 78:965–975

DOI 10.1007/s12562-012-0525-1

most widely used selective devices in European fisheries A

120 mm window was implemented in the Kattegat and

Skagerrak fisheries beginning in 2005 [11], but it did not

produce a marked improvement in selectivity for cod [13]

A first step in developing a new window to improve the

selectivity of gear, and thereby reduce cod bycatches in the

Norway lobster fishery without incurring major losses of

the target species, was tested by Madsen et al [17] The

results demonstrated very high reductions in cod bycatch

without any apparent increased losses of Norway lobster

compared to values quoted in the literature [17] The

possibility that some Norway lobster escaped through the

window could not be excluded from these experiments

alone [17] A better estimate of possible escape through the

window can be achieved by using a window cover which

only collects fish that escape from the window [17]

This paper presents the second stage in our research:

developing and testing a window design that can be used

commercially, and is simple to install and control in a

variety of trawl designs We developed a flexible solution

that allows the selectivity of a window to be adjusted

according to a particular need We assessed the effects of

different window mesh sizes and heights of the sorting box

We compared our results against a standard codend, as a

baseline, to assess the full effect of the new design concept

Materials and methods

Development of the sorting box concept

The basic objective of the sorting box concept is to

improve and control the gear’s selectivity for cod and other

bycatch species without influencing its selectivity and

catch efficiency for Norway lobster The initial

develop-ment of this novel concept is described in more detail in

Madsen et al [17] Here we discuss the further

develop-ment and testing of the concept, focusing in particular on

its commercial use

The main ideas behind the sorting box concept are (1) to

move the window as far back as possible to improve the

release of cod but prevent the accumulating Norway lobster

catch from coming into contact with the window; (2) to

place the window in a relatively narrow four-panel section

in order to promote stable performance and control over the

up/down orientation; (3) to place the window in a relatively

narrow section to reduce the escape route length; (4) to use

larger mesh sizes in the window than those currently in use;

and (5) to allow very quick adjustment of the selectivity so

that changes that are needed to protect cod stocks can be

made rapidly

The sorting box developed and used in these sea trials is

shown in Figs.1and2 The mesh size in the window was

increased from about 290 mm (as used in the design that

we tested previously [17]) to a nominal 370 mm becauseobservations showed that the smaller mesh prevented thelargest cod escaping This is an important consideration inthe Kattegat area, where the specific aim is to protect cod inits spawning grounds An optional smaller nominal

150 mm mesh size was also tested, which would be ciently large to release cod below the MLS but to retainmore cod and other relevant commercial bycatch speciesfor use during periods when the landing of cod was per-mitted The net in the window was made of a coatedpolyamide thread produced by Carlsen Net (http://www

ensures a more stable opening, and has been very effective

in other square mesh window applications The coatedtwine was white, which should reduce the visual contrastbetween the meshes in the window and the light fromabove [18] The window was fastened to the codend oneach side with a zipper that was the same length as the boxsection (about 3 m) (Fig 2) and was joined to the forwardand aft ends of the codend with thread This means ofattachment allows the net selectivity to be quickly adjusted

by replacing the window with one with a different meshsize, and takes only about 10–15 min The remainder of thecodend was made of nominal 90 mm double-twine (4 mm)

PE netting, which was chosen because it is widely usedcommercially and is stronger than the single twine nettingused in our earlier prototypes [17]

The sorting box was tested and adjusted in the Hirthals(Denmark) flume tank prior to the sea trials, and inspectedand measured after the sea trials at a speed of 1.8 knots.The sorting box shape was found to be more unstable (lessstretched) when made from stiff double twine, than fromthe single twine used in the initial experiments [17] Thisproblem was solved by adding leaded rope to each of thelower selvages of the four-panel section Different weightswere attached to the two sorting boxes tested to assess theeffect of the height of the sorting box The two sortingboxes were termed the ‘‘low’’ and the ‘‘high’’ sortingboxes, and the effect is illustrated in Fig.1 The low andhigh sorting boxes had 5.8 and 6.8 kg of leaded rope (about

1 kg/m) added, respectively, along the lower two selvages

in the box section

Test codendsThree final versions of the sorting box were tested: (1) alow sorting box with a nominal 370 mm window; (2) ahigh sorting box with a nominal 370 mm window; and (3)

a high sorting box with a nominal 150 mm window

A conventional 90 mm two-panel standard diamondmesh codend was used as a comparison to the sortingboxes This codend was on one side of a twin trawl rig in

some separate experiments conducted prior to this test of

the sorting boxes; the same vessel was used in the same

areas and the same methodology was employed for the

measurements The codend was made from the same type

of netting as the sorting box trawls: 4 mm double-PE

netting, 92 open meshes in circumference and 6 m long

(stretched) It was attached to a 3 m long diamond

elon-gation made of the same netting as the codends that was the

same number of meshes in circumference This was done to

facilitate the handling of the codend, and furthermore the

codend section had about the same length as the sorting

box codends The 90 mm mesh is the minimum legal mesh

size, and is the preferred mesh size of most of the Kattegatand Skagerrak fleet

Codend coversData on the selectivities of the various codends were obtainedusing the covered codend methodology The single coveredcodend methodology [19] was used to test the standard co-dend and the low sorting box The cover was made of nominal

40 mm (inside mesh opening) netting with a combination ofkites, chains, and floats to keep the covers from touching orblocking the meshes of the test codends [19]

Fig 1 Illustration of the low

sorting box, with the high

sorting configuration indicated

by the hatched line The high

To measure the numbers of Norway lobster and fish

escaping through the escape window, we developed a novel

dual codend cover (Fig.3) which was used to test both of

the high sorting boxes The cover is divided into two

col-lecting bags; the upper one collects escapees through the

escape window, while the lower one collects fish escaping

from the rest of the sorting box system A 40 mm

hori-zontal separator net panel is attached to the upper selvage

of the sorting box (Fig.3) To reduce the risk of

influ-encing the performance of the sorting box, some slack was

introduced into the separator net panel to ensure that it did

not limit the movement of the sorting box at increased

speed Furthermore, it was placed in a position where it

could not be detected by fish looking through the window

from the inside of the sorting box, and where it was

unli-kely to influence the water flow inside the sorting box The

bottom panel of the sorting box was able to move freely

with additional weight attached The top panel in front of

the window was blinded to ensure that the only way to

reach the upper cover was through the window Floats were

attached to the lower collecting bag in the area where most

of the escaping fish were expected to make their exit, to

ensure that this part was clear of the codend The dual

cover was tested and adjusted in a flume tank prior to the

sea trials, and inspected after the sea trials

Sea trials and data collection

Sea trails were carried out in the Kattegat and Skagerrak,

and general hauling information is provided in Table1 A

commercial stern trawler (RS30, Mette Amalie, 386 kW,

20 m in length) was used to conduct sea trials in August

2007 It was rigged for twin trawling with two identical,

combined fish and Norway lobster trawls that were fishing

simultaneously The trawls had a circumference of 460

meshes in the fishing circle and a nominal mesh opening of

100 mm No codend extension was attached to the trawl

since this could have closed the sorting box The sorting

boxes were attached directly to the tapered end section of

the trawl, where it had a circumference of 100 open

meshes A three-warp towing system with a 550 kg chain

clump and two 194 cm Welle otter boards was used to tow

the gear The low and high sorting boxes with the 370 mmmesh size were tested together on each side of the twintrawl rig The 150 mm high sorting box was tested on oneside of the twin trawl rig The other side of the twin trawlrig was used for another experiment to be describedelsewhere

To obtain the weight of the total catch, the entire codendand cover fraction was weighed using a crane scale ondeck, and then the weight of the netting was subtracted.Overall length measurements of cod, plaice Pleuronectesplatessa, and Norway lobster were taken, as these are themost important commercial species in this fishery.All cod were measured, although only subsamples ofplaice were measured for a few very large catches Norwaylobster were caught in higher numbers and often subsam-pled Other commercial fish species caught by the sortingbox were measured because they have some influence onthe economics of this fishery For species where there is astipulated MLS, only individuals above the MLS weremeasured, but very few individuals smaller than the MLSwere observed Only cod, plaice, and Norway lobster weremeasured in the standard codend Fish were measured tothe nearest cm A subsample of the Norway lobster catchwas taken when catches were large and measured to thenearest mm with an electronic caliper The midpoints of thelength classes of fish and Norway lobster were used in thesubsequent analysis Mesh sizes were measured with theOMEGA gauge [20]

The meshes of the sorting box codends (N = 200), thestandard codend (N = 300), and windows (N = 50) weremeasured in wet conditions

Statistical modeling and analysisStatistical models were constructed for the catch data forcod, plaice, and Norway lobster The analysis followedFrandsen et al [17] and Madsen et al [13], and theapproach is briefly recapped here

Individuals entering the codend were potentiallyexposed to two selective [21] mechanisms due to theactions of the sorting box window and the codend Thefollowing composite effective selection curve applies:

Fig 3 Illustration of the dual

cover

u ‘ð Þ ¼ 1  c  1  r½ ð windowð Þ‘ Þ  rcodendð Þ;‘

where rwindowð Þ and r‘ codendð Þ denote the proportion of the‘

cod (of length ‘) retained by the window and codend,

respectively, given that they have been in contact with the

window/codend, and c denotes the proportion of the fish

that enter the codend and make contact with the window

We used logistic curves to model rwindowð Þ and r‘ codendð Þ.‘

The proportions of the total catch retained in each

compartment when testing the low sorting box with a

traditional single cover are described by:

ucodendð Þ ¼ 1  c  1  r‘ ½ ð windowð Þ‘Þ  rcodendð Þ;‘

ucoverð Þ ¼ 1  1  c  1  r‘ ½ ð windowð Þ‘ Þ  rcodendð Þ:‘

For the three compartments when testing the high sorting

boxes with a dual cover, these proportions are:

ucodendð Þ ¼ 1  c  1  r‘ ½ ð windowð Þ‘Þ  rcodendð Þ;‘

uwindow coverð Þ ¼ c  1  r‘ ð windowð Þ‘ Þ;

ucodend coverð Þ ¼ 1  c  1  r‘ ½ ð windowð Þ‘ Þ  1  rð codendð Þ‘ Þ:

For cod in all of the sorting boxes and plaice in the

150 mm high sorting box, the composite selection curve

provided the best fit because of the considerable numbers

that escaped through both the window and the codend In

the standard codend, the traditionally used logistic function

[22] was used for all species This was also the case for

Norway lobster in all of the sorting boxes where escape

through the window was limited and the goodness of fit for

individual hauls, as assessed by deviance residuals and the

deviance statistic [23], suggested that the logistic function

generally provided a good fit

The catch data for plaice in the two sorting boxes with a

370 mm window suggested a bell-shaped effective

reten-tion curve This reflects the product of codend selecreten-tion

and increasing escape through the window with increase in

plaice length This window mesh size is too large in

rela-tion to the length of the plaice to detect any increase in

retention with increasing fish length Comparing the mal, log-normal, and gamma curves [14], the normal curve

nor-rð Þ ¼ x  exp ‘ ð‘ ‘0Þ

2

2r2

!

provided the best fit in terms of least deviance, where, ‘0,

r, and x are the modal length, the spread, and the modalvalue, respectively; see Frandsen et al [14] for furtherdetails

The selectivity model fitting process failed to convergefor several of the hauls for plaice and cod Consequently,the use of Fryer’s model of between-haul variation [24]would have been limited to a subset of the data for cod andplaice, thus leading to a loss of efficiency and potentiallybiased estimates Therefore, instead, we utilized all data forthese two species by fitting the SELECT method to the set

of data from all hauls stacked into a single data set [14,25].Standard errors for the parameter estimates were subse-quently adjusted by the REP [23,25], which allowed theextra variation due to sampling from replicate hauls to beaccommodated For Norway lobster, it was possible to fitselection curves for individual hauls, except for one haul ofthe standard codend where only four individuals werecaught Residual maximum likelihood (REML) estimateswere obtained using a fixed and random effects model,taking the between-haul variation into account [22,24] Weused Wald tests to analyze differences between parameters

of the codends, with significant differences corresponding

to nonoverlapping 95 % confidence limits

ResultsExperiments with 370 mm windowsThe average codend and window mesh sizes (with standarddeviations) of the low sorting box were 95.7 mm (2.7) and367.1 mm (5.7), respectively For the high sorting box, the

Table 1 Operational conditions

during the sea trials, as

indicated by average values

Average per haul with the

Depth (m) Speed

(knt)

Codend catch (kg) August 2007

370 mm, low sorting box

Single 11 3.27 ± 0.18 64.4 ± 33.8 2.8 ± 0.28 147 ± 219

370 mm, high sorting box

Dual 11 3.27 ± 0.18 64.4 ± 33.8 2.8 ± 0.28 219 ± 114

September 2007

150 mm, high sorting box

Dual 10 3.35 ± 0.41 72.0 ± 30.5 2.7 ± 0.17 402 ± 199

August 2007

90 mm, standard codend

Single 16 3.58 ± 0.79 52.0 ± 31.9 2.6 ± 0.11 222 ± 129

codend mesh measured 95.2 mm (2.5) and the window

mesh 365.6 mm (6.9) The mesh size of the standard

co-dend was 94.9 mm (3.4)

Sorting box heights were measured following the sea

trials and were found to be around 55 cm (low sorting box)

and 75 cm (high sorting box) at a speed of 1.8 knots, which

is the maximum speed for the flume tank The flume tank

test showed that the codend and the escape window were

totally clear of the dual cover and the separator net panel

There was some slack in the separator net panel that made

movements of the sorting box possible upon increasing the

speed

A total of 11 paired hauls were conducted with the two

370 mm sorting boxes The total codend catch of the low

sorting box was lower than that for the high sorting box

(Table1) Table2 provides detailed information on the

recorded total catches and retention and the escape of cod,

plaice, and Norway lobster above and below the MLS The

total retention rate of cod was very low (8.9 %) in the low

sorting box and low (23.7 %) in the high sorting box

compared to that in the standard codend (58.2 %) The

proportion of cod above the MLS retained in the standard

codend (84.2 %) was high compared to those in the low

(10.4 %) and high (34.1 %) sorting boxes The dual cover

showed that very few cod above the MLS escaped through

the codend of the high sorting box, and that the main

escape route was through the window

The total retention of plaice above MLS was very low inthe low sorting box (8.3 %) compared to that in the stan-dard codend, where almost all of the plaice were retained.The retention was higher (29.5 %) in the high sorting box.The dual cover of the high sorting box showed that plaiceabove the MLS only escaped through the window Only29.5 % of the plaice below the MLS escaped through thewindow

Most Norway lobster above the MLS were retained inthe standard codend (82.3 %), and slightly more wereretained in the high sorting box (86.4 %), whereas fewer(72.0 %) were retained in the low sorting box The dualcover of the high sorting box showed that 2.0 % of theNorway lobster caught escaped through the window.The numbers (or proportions) of escapees for othercommercial species above the MLS from the sorting boxesare indicated in Table2 Saithe Pollachius virens werecaught in higher numbers than the other species Very fewsaithe were retained at either codend, with only a minordifference (2.3 %) observed between the low sorting boxand the high sorting box Catches of other species werelower (Table3) Fewer haddock Melanogrammus aeglefi-nus were caught in the low sorting box and none wereretained, whereas 11 % were retained in the high sortingbox and all escapes were through the window Only a fewwitch flounder Glyptocephalus cynoglossus were retained

in the low sorting box, whereas more than half were

Table 2 Recorded catches,

divided into total numbers or

percentages that entered, were

retained, and escaped through

the codend and window

MLS cod = 30 cm,

plaice = 27 cm, Norway

lobster = 40 mm; UMLS under

minimum landing size

Cod Plaice Norway lobster UMLS MLS Tot UMLS MLS Tot UMLS MLS Tot

370 mm, low sorting box Entered (no.) 1156 1942 3098 2323 205 2528 24601 4087 28688 Retained (no.) 75 202 276 121 17 139 9840 2943 12795 Retained (%) 6.5 10.4 8.9 5.2 8.3 5.5 40.0 72.0 44.6

370 mm, high sorting box Entered (no.) 1743 1880 3623 1622 210 1832 20768 3918 24685 Retained (no.) 216 641 859 144 62 205 12648 3268 15922 Retained (%) 12.4 34.1 23.7 8.9 29.5 11.2 60.9 83.4 64.5 Escaped codend (%) 29.8 2.1 15.4 61.7 0.0 54.6 36.9 15.2 33.5 Escaped window (%) 57.8 63.8 60.9 29.5 70.5 34.2 2.2 1.5 2.0

150 mm, high sorting box Entered (no.) 1599 2141 3740 333 213 546 6172 1371 7543 Retained (no.) 168 1355 1522 69 179 248 4438 1185 5627 Retained (%) 10.5 63.3 40.7 20.7 84.0 45.4 71.9 86.4 74.6 Escaped codend (%) 52.2 2.2 23.6 65.2 13.6 45.1 24.8 11.6 22.4 Escaped window (%) 37.3 34.5 35.7 14.1 2.3 9.5 3.3 2.0 3.0

90 mm, standard codend Entered (no.) 444 333 777 2984 263 3247 22711 4626 27337 Retained (no.) 181 280 452 830 262 1042 14830 3807 18644 Retained (%) 40.8 84.2 58.2 27.8 99.6 32.1 65.3 82.3 68.2

retained in the high sorting box, where all escapes were

through the window The total retention of brill

Scoph-thalmus rhombus was the same between the low sorting

box and the high sorting box, and most escapes from the

high sorting box were through the window

Experiments with the 150 mm window

The codend mesh size of the sorting box with the 150 mm

window was the same as that of the high sorting box with

the 370 mm window The window mesh size (with

stan-dard deviation) was 148.6 ± 4.8 mm A total of ten hauls

were performed (Table1) Very few cod below the MLS

were retained (10.5 %), with most escaping through the

codend meshes (Table2) About 30 % more cod above

the MLS were retained than in the large mesh (370 mm

window) high sorting box, with most escapes being

through the window More than twice as many plaice

below the MLS were retained compared to the large mesh

high sorting box The retention of plaice above the MLS

was high compared to those for the two large mesh

sorting boxes, and very few plaice escaped through the

window Most of the escaping plaice that were below the

MLS left through the codend Compared to the large mesh

high sorting box, 3 % more Norway lobster above the

MLS and 11 % below the MLS were retained The escape

rates through the window were 2.0 and 3.3 % for Norway

lobster above and below the MLS, respectively Fewer

individuals escaped through the codend compared to the

large-mesh high sorting box

The numbers (or proportions) of escapees for other

commercial species are shown in Table3 Many more

individuals were retained than for the other sorting box

codends: almost all witch flounder and brill, almost

two-thirds of the saithe, and almost half of the haddock Saithe,

haddock, witch flounder, and hake escaped only throughthe window

Selectivity estimatesThe selectivity parameters for cod, plaice, and Norwaylobster are shown in Table4, and the resulting selectioncurves are depicted in Fig.4 The estimated proportion ofcod that came into contact with the window of the sortingbox (c) was high (89 %) for the 370 mm low sorting box.The probability of window contact was statistical signifi-cantly lower (61 %) in the 370 mm high sorting box, andstatistical significantly lower still in the 150 mm highsorting box (43 %) The estimated selectivity parametersfor cod in all codends (‘50%;codend, SRcodend) did not differstatistically significantly These parameters were similarfor both of the high sorting boxes, whereas the high stan-dard errors for the low sorting box indicated a less preciseestimate The resulting selection curves are compared inFig.4 The selection curves for cod in the tested sortingbox codends are clearly shifted to the right of that for thestandard codend, and there is a marked effect of the

370 mm window size, particularly for the low sorting box,which shows a very low retention, even for larger cod Thedifference between the selection curves for cod in the two

370 mm sorting boxes increases with fish length up toaround 50 cm and then decreases However, there were notmany observations of fish above 80 cm The retention ofthe 370 mm low sorting box rose steeply for fish above

80 cm

The modal top of the selection curve for plaice (r) forthe 370 mm low sorting box is shifted to the left comparedwith that for the 370 mm high sorting box, and the peak(x) is about half the height (Table 4; Fig.4) The modelestimates that 14 % of the plaice made contact with the

Table 3 Recorded catches of

other commercial species at or

above the MLS, divided into

total numbers or percentages

that entered, were retained, and

escaped through the codend and

Saithe Haddock Witch flounder Brill Hake

370 mm, low sorting box

150 mm, high sorting box

Retained (%) 63.9 45.7 96.7 100.0 64.3 Escaped codend (%) 0.0 0.0 0.0 0.0 0.0 Escaped window (%) 36.1 54.3 3.3 0.0 35.7

window (c) in the 150 mm low sorting box Most

obser-vations of plaice related to fish 10–40 cm long

There were no statistically significant differences in the

selectivity parameters of Norway lobster between any of

the four codends, and the resulting selection curves areshown in Fig.4 However, the standard errors were rela-tively large (Table4) This can largely be explained by thehigh variation that was observed between hauls Very few

Fig 4 Selectivity curves for the tested codends

Table 4 Estimated selectivity

parameters, with the standard

error given in parentheses

DF degrees of freedom,

DEV deviance

Parameter Estimate Parameter Estimate Parameter Estimate

370 mm, low sorting box

Norway lobster were observed outside the range 25–

50 mm, and a substantial part of the selection curve was

therefore not covered Inspection of the 95 % confidence

limits around the selection curves indicated that there was

no statistically significant difference between any of the

codends within the range of observations

Discussion

There was no statistically significant difference between

the selectivity parameters for Norway lobster between the

codends However, these parameters must be regarded with

caution since there was a high variation between the hauls,

as indicated by relatively large variances In general,

considerable variation in the results between experiments

for Norway lobster is common [26] The use of a dual

cover is important in order to enable the identification of

the very small number of escapes through the window, as

observed in the present experiment The dual cover

indi-cated a very low escape (2–3 %) of Norway lobster through

the window in the high sorting box, but this may also be the

case through the top panel of a codend without a window

[27] Because variances are high, and escape rates through

the window are likely low, it is impossible to detect any

significant potential loss of Norway lobster from the low

sorting box Genuinely low losses seem possible since

fewer individuals above the MLS were retained than for the

other codends Further investigation of this issue is

important for reducing losses of Norway lobster as much as

possible, and thus maximizing the economic viability of

the Norway lobster fishery

Caution is required when comparing the selectivities of

the high and low sorting boxes Although it was possible to

measure the difference in height in a flume tank, the

situ-ation may be different during sea trials In the flume tank,

the height was measured at a water speed (i.e., 1.8 knt) that

is lower than is usual during commercial fishing activity

(around 2.5 knt), and without a catch in the trawl The

considerable differences in catch rates, which were highest

in the high sorting boxes, could likely influence the height

and the water flow inside the sorting box, causing

addi-tional differences The height of the sorting box will also

depend on the height of the trawl at the aft end to which the

sorting box is attached Our main theory is that the height

of the sorting box was actually lower than measured in the

flume tank, which particularly influences the selectivity of

the window in the low sorting box This is supported by the

very high proportion of escaping cod and plaice and a

tendency (albeit not a statistically significant one) for lower

retention of Norway lobster, which are more likely to make

contact with the window A lower height and geometry and

a narrower section will likely change the water flow inside

the sorting box section, which could be decreased or couldincrease the flow through the window The dual cover didnot influence the performance of the high sorting box in theflume tank test It is not likely that increasing the speedwould change the performance of the side panel in a sig-nificant way The very high escapes of saithe and haddock(around 90 %) through the window of the large-mesh highsorting box suggest that it window penetration was not aproblem Saithe are relatively large, strong swimmers, andshow very active net escape behavior, as do haddock [13],suggesting that the height (distance) is a driving factor Theescape of cod was considerable in both of the high sortingboxes, considering that little or no effect of using an escapewindow has been detected in the past [11,13] The contactprobability in the low sorting box is very high, with close

to 90 % of the cod entering the codend coming into contactwith the window These findings suggest that the height is avery important influence on the escape of fish through awindow as it affects the escape distance, and a narrowsection will increase escape panic However, the heightwill also influence species other than cod, leading to eco-nomic losses if these are landed This experiment, as well

as other experiments [13,27], shows that most commercialspecies above the MLS will be retained in a standard

90 mm codend, and most that escape through the windowwill be lost This issue deserves continued investigation onorder to optimize the design further In addition, it isimportant to try to keep the performance of the sorting boxstable We foresee several ways in which this could beachieved: (1) the use of four-panel sections throughout;(2) the use of single twine with the same diameter as thatused in the first design [17]; (3) the insertion of a framemade of plastic or nylon into each end of the box section

to keep the opening fixed; (4) optimizing the opening inthe aft part of the trawl to which the sorting box section isattached and avoiding codend extensions; and (5) usingsquare meshes instead of diamond meshes in the boxsection with a fixed mesh opening, of a size that makes itpossible to retain Norway lobster more selectively [27,

28] All of our experiments were conducted during thesummer and during the day, but cod may react differently

at night [29], or when the water temperature is lower andswimming performance is reduced [30,31] Furthermore,

it is important to assess if escapes occur when the trawl ishauled back, since this may cause additional mortalitycompared to escapes during towing [32, 33] Underwaterobservations of fish behavior in relation to the gear would

be valuable, but several previous attempts to do this havefailed because of heavy mud clouds which cause limitedvisibility in this fishery

The escape rate of plaice was very high with the

370 mm windows, similar to those seen in our initial totype trials [17] The vertical distribution of plaice is

relatively even at the aft end of a trawl [34], and plaice

seem to particularly seek out the window as a means of

escape The bell-shaped selection curve of plaice is caused

by smaller plaice primary escaping through the codend

meshes, whereas swimming performance in general

improves [30, 31] with size, and general behavior might

also be size dependent This leads to an increased

possi-bility of penetrating the window meshes, either initially

when passing under the window or by swimming the

limited distance (about 3 m) forward from the catch

building up at the end of the codend and then escaping

through the window Size-dependent differences in flatfish

selection have also been reported elsewhere [35] The

plaice were not large enough to be physically retained by

the 370 mm window

The limited results for witch and brill indicate that other

flatfish species might behave differently In future studies,

it will be important to gather data on sole, because it is an

economically important species in several areas Whether it

is possible to develop this design concept further to

increase the retention of marketable flatfish should be

investigated to make the concept usable in mixed fisheries

This process would benefit from including fishermen and

netmakers in the process of development

Nearly 100 % of the cod above 50 cm are retained in

the 120 mm window used by the Kattegat and Skagerrak

fishery [13], which is far too many considering the need to

rebuild cod spawning stocks in several areas Our

exper-iments demonstrated that it is possible to develop trawl

designs that can substantially improve the selectivity of

gear for cod A high proportion of the cod come into

contact with the window, as shown in our initial

experi-ments with the sorting box concept [17] The

improve-ments in gear selectivity and the proportion of cod that

escape obtained with the use of the sorting box are far

greater than those achieved by merely increasing the

co-dend mesh size to 120 mm, as has been done in some

North Sea consumption fisheries, a strategy which also

reduces the catch of Norway lobster above the MLS by

about one-third [11] Compared to the estimates for a

35 mm grid in a nominal 90 mm codend [13], the

reten-tion of small cod (\30 cm) is lower in the sorting box

For larger cod ([30 cm), the selectivity in the sorting box

was lower than that of the grid, which rejects all cod

above a given size [3,12,13] The proportion of cod that

came into contact with the window decreased when the

window had a mesh size of 150 mm Since the escape rate

of Norway lobster through the window was about the

same for each, one theory could be that the smaller

meshes are more clearly visible to cod, resulting in fewer

escape attempts However, an increased catch rate might

also increase the height of the sorting box and lower the

or according to a particular vessel’s specific quota rights(on cod, for example) The concept is general in nature andcould be used in other areas (e.g., the North Sea, Irish Sea)and fisheries where similar problems exist A disadvantage

of it when compared to a grid is that the effectiveness ofthe window requires that the target fish show active escapebehavior, whereas a grid works mechanically to sort thecatch and to reject fish above a given size

Attaching the window using a zipper provides a simplemeans of adjusting the selectivity by changing the windowmesh, and this was found to work well The high propor-tion of cod that comes into contact with the window offersthe possibility of adjusting the selectivity for cod in caseswhere it is desirable to retain fish over the MLS This is animportant aspect of future fisheries management, whereelectronic monitoring will make it possible to change fromlandings to total catch quotas, which will give fishers agreater incentive to avoid discards [36] Furthermore, dis-card bans are likely to be integrated into European Unionfisheries management in the future, and have already beenannounced for Skagerrak In a study described elsewhere[28], we developed the concept of a four-panel codendfurther in order to improve the selectivity for Norwaylobster by inserting square meshes into the bottom panel,since their main escape route is through the lower part ofthe codend [27] Furthermore, we added conventionaldiamond meshes to the trawl sides to ensure optimal meshselection for plaice, which improved the selectivity [27]

A modified version of the sorting box (minimum

300 mm window meshes) used in the present experiment(named the SELTRA trawl) was introduced in the Kattegatfishery from 2009 in a bilateral agreement between Swedenand Denmark as an alternative to the sorting grid, for use inareas and spawning seasons when the catchability of cod ishigh [14] From 2011, the modified sorting box (minimum

180 mm window meshes) will be mandatory in Kattegat,and from 2012 this will be the case in Skagerrak Our work

in assessing and improving the performance of the sortingbox concept continues The sorting box system imple-mented by legislation is only allowed to have a maximum

of 100 open meshes in circumference, compared to 140meshes in the present design This difference is due to thegeneral EU legislation requirements, and the smaller cir-cumference could potentially increase losses of Norwaylobster but also increase the escape of bycatch through thewindow Furthermore, the implemented sorting box isinserted into a four-panel section that extends all the way to

the codline The effect on the cod stock in the Kattegat will

be evaluated in the near future

Acknowledgments Thanks are particularly due to Mogens

Ander-sen and Kurt HanAnder-sen (SINTEF), who assisted in development work

and sea trials, and to the skipper crew of the Mette Amalie This work

was done within the SELTRA project, which was carried out with the

financial support of the European Union and the Danish Ministry of

Food, Agriculture, and Fisheries.

References

1 Campos A, Fonseca P, Erzini K (2002) Size selectivity of

dia-mond and square mesh cod ends for rose shrimp (Parapenaeus

longirostris) and Norway lobster (Nephrops norvegicus) off the

Portuguese south coast Fish Res 58:281–301

2 Loaec H, Morandeau F, Meillat M, Davies P (2006) Engineering

development of flexible selectivity grids for Nephrops Fish Res

79:210–218

3 Catchpole TL, Revill AS, Dunlin G (2006) An assessment of the

Swedish grid and square-mesh codend in the English (Farn

Deeps) Norway lobster fishery Fish Res 81:118–125

4 Revill A, Dunlin G, Holst R (2006) Selective properties of the

cutaway trawl and several other commercial trawls used in the

Farne Deeps North Sea Nephrops fishery Fish Res 81:268–275

5 Graham N, Fryer RJ (2006) Separation of fish from Nephrops

norvegicus into a two-tier cod-end using a selection grid Fish

Res 82:111–118

6 Catchpole TL, Tidd AN, Kell LT, Revill AS, Dunlin G (2007)

The potential for new Nephrops trawl designs to positively effect

North Sea stocks of cod, haddock and whiting Fish Res

86:262–267

7 Revill AS, Catchpole TL, Dunlin G (2007) Recent work to

improve the efficacy of square-mesh panels used in a North Sea

Nephrops norvegicus directed fishery Fish Res 85:321–327

8 Drewery J, Bova D, Kynoch RJ, Edridge A, Fryer RJ, O’Neill FG

(2010) The selectivity of the Swedish grid and 120 mm square

mesh panels in the Scottish Nephrops trawl fishery Fish Res

106:454–459

9 Briggs RP (2010) A novel escape panel for trawl nets used in the

Irish Sea Nephrops fishery Fish Res 105:118–124

10 Ingo´lfsson O ´ A (2011) The effect of forced mesh opening in the

upper panel of a Nephrops trawl on size selection of Nephrops,

haddock and whiting Fish Res 108:218–222

11 Krag LA, Frandsen RP, Madsen N (2008) Evaluation of a simple

means to reduce discard in the Kattegat–Skagerrak Norway

lobster (Nephrops norvegicus) fishery: commercial testing of

different codends and square-mesh panels Fish Res 91:175–186

12 Valentinsson D, Ulmestrand M (2008) Species-selective Norway

lobster trawling: Swedish grid experiments Fish Res 90:109–117

13 Frandsen RP, Holst R, Madsen N (2009) Evaluation of three

levels of selective devices relevant to management of the Danish

Kattegat–Skagerrak Nephrops fishery Fish Res 97:243–252

14 Madsen N, Valentinsson D (2010) Use of selective devices in

trawls to support recovery of the Kattegat cod stock: a review of

experiments and experience ICES J Mar Sci 67:2042–2050

15 Hilborn R, Litzinger E (2009) Causes of decline and potential for

recovery of Atlantic Cod populations Open Fish Sci J 2:32–38

16 Lilly GR, Wieland K, Rothschild B, Sundby S, Drinkwater K,

Brander K, Ottersen G (2008) Decline and recovery of Atlantic

cod (Gadus morhua) throughout the Atlantic In: Resiliency of gadoid stocks to fishing and climate change Alaska Sea Grant College Program, University of Alaska, Fairbanks, pp 39–67

17 Madsen N, Frandsen RP, Holst R, Krag LA (2010) Development

of new concepts for escape windows to minimise cod catches in Norway lobster fisheries Fish Res 103:25–29

18 Glass CW, Wardle CS, Gosden SJ (1993) Behavioural studies of the principles underlying mesh penetration by fish ICES Mar Sci Symp 196:92–97

19 Madsen N, Hansen KE, Moth-Poulsen T (2001) The kite cover: a new concept for covered codend selectivity studies Fish Res 49:219–226

20 Fonteyne R, Buglioni G, Leonori I, O’Neill FG, Fryer RJ (2007) Laboratory and field trials of OMEGA, a new objective mesh gauge Fish Res 85:197–201

21 Zuur G, Fryer RJ, Ferro RST, Tokay T (2001) Modelling the size selectivities of a trawl codend and an associated square mesh panel ICES J Mar Sci 58:657–671

22 Wileman DA, Ferro RST, Fonteyne R, Millar RB (eds) (1996) Manual of methods of measuring the selectivity of towed fishing gears ICES Coop Res Rep 215:126

23 McCullagh P, Nelder JA (1989) Generalized linear models, 2nd edn Chapman & Hall, London

24 Fryer RJ (1991) A model of the between-haul variation in selectivity ICES J Mar Sci 48:281–290

25 Millar RB, Broadhurst MK, Macbeth WG (2004) Modelling between-haul variability in the size selectivity of trawls Fish Res 67:171–181

26 ICES (2007) Report of the Workshop on Nephrops Selection, ICES CM 2007/FTC: 01 ICES, Copenhagen

27 Frandsen RP, Madsen N, Krag LA (2010) Selectivity and escapement behaviour of five commercial species in a square mesh and a diamond mesh codend ICES J Mar Sci 67:1721–1731

28 Frandsen RP, Herrmann B, Madsen N, Krag LA (2011) opment of a codend concept to improve size selectivity of Nephrops (Nephrops norvegicus) in a multi-species fishery Fish Res 111:116–126

Devel-29 Krag LA, Holst R, Madsen N, Hansen K, Frandsen R (2010) Selective haddock (Melanogrammus aeglefinus) trawling— avoiding cod (Gadus morhua) bycatch Fish Res 101:20–26

30 He P (1993) Swimming speeds of marine fish in relation to fishing gears ICES Mar Sci Symp 196:183–189

31 Videler JJ, He P (2010) Swimming in marine fish In: He P (ed) Behavior of marine fishes—capture processes and conservation challenges Wiley–Blackwell, Ames, pp 5–23

32 Madsen N, Skeide R, Breen M, Krag LA, Huse I, Soldal AV (2008) Selectivity in a trawl codend during haul-back operation—

an overlooked phenomenon Fish Res 91:168–174

33 Grimaldo E, Larsen RB, Sistiaga M, Madsen N, Breen M (2009) Selectivity and escape percentages during three phases of the towing process for codends fitted with different selection systems Fish Res 95:198–205

34 Krag LA, Holst R, Madsen N (2009) A study of vertical bution of fish in the aft end of a demersal trawl ICES J Mar Sci 66:772–777

distri-35 Walsh SJ (1992) Size-dependent selection at the footgear of a groundfish survey trawl North Am J Fish Manag 12:625–633

36 Kindt-Larsen L, Kirkegaard E, Dalskov J (2011) Fully mented fishery: a tool to support a catch quota management system ICES J Mar Sci 68:1606–1610

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

Effect of a stepwise lighting method termed ‘‘stage reduced

lighting’’ using LED and metal halide fishing lamps

in the Japanese common squid jigging fishery

Yoshiki Matsushita•Yukiko Yamashita

Received: 10 May 2012 / Accepted: 9 July 2012 / Published online: 26 July 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract Lighting systems combining light-emitting

diodes (LEDs) and metal halide lamps (MHs) are expected

to be energy-saving tools in Japan’s squid jigging fishery

Previous research has shown the need for light stronger

than LEDs (9 kW) and 36 MHs (108 kW) to catch the

Japanese common squid Todarodes pacificus We tested a

stepwise lighting method termed ‘‘stage reduced lighting’’

in the Tsushima Strait in January and February 2010 using

nine fishing boats LEDs (9 kW) and 50 MHs (150 kW)

were lit for 3.9 h on average, and then the number of MHs

was reduced to either 30 or 36 until the end of fishing

(7.3 h on average) This method reduced fuel consumption

by 22–25 % compared to the continuous use of all fishing

lamps (159 kW) We carried out a catch analysis of nine

experimental boats and 21 commercial boats during the

experimental period Generalized linear modeling analysis

suggested that the squid catch can be explained by the

illuminated fraction of the moon and monthly changes in

squid abundance, and the lighting method The stage

reduced lighting using LEDs and MHs has the potential to

reduce fuel consumption while maintaining the squid catch

Keywords Catch performance Fishing light  Fuel

saving Japanese common squid Todarodes pacificus 

Light-emitting diode Squid jigging  Stepwise lighting

method

Introduction

A Japanese coastal squid jigging boat of 19 gross tons(GT) typically consumes approximately 60 l of fuel perhour while jigging with lamps during the night [1]; thus,this type of fishing is rather energy intensive We con-ducted a series of fishing experiments on this fishery byequipping boats with arrays of light-emitting diode panels(LEDs, 9 kW in total) and different numbers of conven-tional metal halide lamps (MHs, 3 kW each) in order tooptimize the economic balance between reduction in fuelconsumption and squid catch The largest catch ofswordtip squid Photololigo edulis was observed when 24MHs and LEDs were employed, but the optimum com-bination of MHs and LEDs was unclear for the Japanesecommon squid Todarodes pacificus, because the largestcatch was observed with the maximum number of MHs(36 MHs) and LEDs [2] A greater catch may be obtained

if more MHs are used, but the number of MHs must berestrained, as increasing the number of MHs obviouslyincreases fuel consumption Accordingly, we tested astepwise lighting method termed ‘‘stage reduced lighting’’for Japanese common squid fishing in the Tsushima Strait

in winter

In the study described in this paper, we examine theeffectiveness of stage reduced lighting in terms of reduc-tion in fuel consumption and catch performance Wetherefore measured the fuel consumption of experimentalboats and compared the observed fuel consumption withthe estimated fuel consumption of commercial boats Togauge the catch performance, we analyzed the catch ofexperimental boats that employed stage reduced lightingand the catch of commercial boats by a generalized linearmodel (GLM), taking into account other factors that poten-tially affect the catch

Y Matsushita ( &)  Y Yamashita

Graduate School of Fisheries Science and Environmental

Studies, Nagasaki University, Nagasaki 852-8521, Japan

e-mail: yoshiki@nagasaki-u.ac.jp

DOI 10.1007/s12562-012-0535-z

Materials and methods

Nine 19 GT squid jigging boats (the same boats as

previ-ously reported [2]) participated in experimental fishing for

43 days between January 9 and February 24, 2010 in the

Tsushima Strait (Fig.1) The boats were equipped with

9 kW blue LEDs (Takagi Corporation, Kagawa, Japan) in

addition to 46–50 MHs (3 kW each) and positive

dis-placement flowmeters (Oval Corporation, Tokyo, Japan,

LS4976-460A for the main engines and LSF40PO-M1 for

the auxiliary engines) We obtained data on time, position,

amount of fuel consumed in each operational process (e.g.,

until arrival at the fishing ground, until the start of the stage

reduced lighting, until the decrease in lighting, until the

end of the lighting period) and the catch of squid (number

of boxes) from each captain’s log books

All MHs and LEDs (total 147–159 kW) were switched

on at the beginning of the lighting period for several hours

(Fig.2a, hereafter referred to as ‘‘full lighting’’), and then

the number of MHs was reduced to either 30 (90 kW) or 36

lamps (108 kW) until the end of fishing (Fig.2b; hereafter

referred to as ‘‘30 MHs’’ or ‘‘36 MHs’’) in stage reducedlighting Thus, boats during the full lighting period con-sumed an equal amount of fuel to that used in commercialoperations, but fuel is then saved by subsequently reducingthe amount of lighting

Fuel consumption data for commercial boats were notavailable The fuel consumption of commercial boats wasestimated based on the average fuel consumption rateduring the full lighting period and the typical time schedule

of fishing operations by the experimental boats

We compared catches of experimental boats to theaverage catch of 21 commercial boats between January 9and February 24, 2010 Experimental boats had to operateoffshore (12 nautical miles from the coastline of the Iki andTsushima islands; Fig.1) due to local regulations [lightingpower must be \60 kW (20 MHs) within 12 nautical milesfrom the coastline] On the other hand, commercial boatswere able to choose their fishing locations at will Inaddition, commercial boats sometimes reduce the number

of MHs for about 1–2 h during the middle of the night toencourage squid to rise to a shallower layer Catches of

Fig 1 Map of the fishing

ground Solid circles designate

positions where stage reduced

lighting with 30 MHs was

carried out in January 2011,

gray circles indicate stage

reduced lighting with 30 MHs in

February, solid squares refer to

stage reduced lighting with 36

MHs in January, and gray

squares label stage reduced

lighting with 36 MHs in

February

Fig 2 Squid jigging boat

lighting LEDs and MHs Full

lighting (a) and lighting with

reduced numbers of MHs (b)

commercial boats are consequently influenced by

differ-ences in location and various lighting conditions

(conven-tional lighting) However, we consider that commercial

data can be useful for comparing catches, because

fisher-men generally try to maximize their fishery earnings in

commercial operations

GLM analysis was conducted for catch analysis The catch,

C, is generally expressed as the product of the catchability

coefficient q, the fishing effort E, and the abundance of squid

in the fishing ground N:

where E is the fishing effort expressed as one operation

For N, we observed a clear difference in the catch of

squid for the experimental and commercial boats between

data for January and data for February We therefore set

a two-level categorical variable (January and February)

for N

We considered that the catchability coefficient q can be

extracted as the product of several factors: q is influenced

by the lighting method, direct and indirect impacts of the

lunar phase, and the fishing power (ability) of each boat

(which derives from the fishermen’s skills):

where qMis the fraction of the catchability coefficient that

is governed by the lighting method We defined qM as a

categorical variable, because it showed a nonlinear

rela-tionship between the catch and number of MHs used in the

previous study [2] qLis the direct and indirect influence of

the lunar phase on the catchability coefficient, and qBis the

fraction of the catchability coefficient that originates from

the difference in fishing power of each boat In this study,

we obtained catch data from 30 boats (9 experimental and

21 commercial boats), but we set ten levels: nine for the

experimental boats and one for the mean fishing power of

the 21 commercial boats, because too many levels require

too many dummy variables, which reduces the degrees of

freedom for analysis

From these assumptions, we selected four factors that

explain the catch C: light as a three-level categorical

var-iable (30 MHs, 36 MHs, and conventional lighting), lunar

as a continuous variable (the illuminated fraction of the

moon, which took a value between 0 and 1), boat as a

ten-level categorical variable (nine experimental boats and the

mean fishing power of the commercial boats), and month as

a two-level categorical variable that reflects the change in

squid abundance (January and February)

Catch amounts were analyzed as a function of the

fac-tors mentioned above by GLM We assume that the catch

Ci(i.e., the number of boxes of squid caught during the ith

operation) is a random variable with a negative binomial

distribution [2];

where h ([0) is a potential dispersion parameter to beestimated [3] The expected values of C, E(C), and itsvariance var(C) are then expressed as

varðCÞ ¼ l þ l2

Overdispersion is expressed as the multiplicative factor

1 ? l/h, which depends on l C is modeled as

ln C¼ b0þ b1lightþ b2boatþ b3lunarþ b4monthþ e;

ð6Þwhere b0is the intercept (constant), and b1, b2, b3, and b4are the coefficients for light, boat, lunar, and month,respectively, while e is the error Parameter estimation wasperformed by the maximum likelihood method (glm.nbfunction in the MASS package [4] in R v.2.13.0, RDevelopment Core Team) We used the stepwise forwardentry method for parameter estimation, following Ya-mashita et al [2] At each stage of the forward entry,Akaike’s information criteria (AIC) was computed forevery candidate model, and the model with the lowest AICwas chosen

Results

We obtained catch and fuel consumption data from the nineexperimental boats, consisting of a total of 114 operations(57 operations with 30 MHs and 57 operations with 36MHs) between January 9 and February 24, 2010 We alsocollected catch data for a total of 466 operations conducted

by 21 commercial boats in the same period

The three box-and-whisker plots in Fig.3 show tions of lighting, amounts of fuel consumed, and fuelconsumption rates versus the lighting method The averageduration of full lighting was 3.9 h, regardless of the num-ber of lamps used after the full lighting period Theexperimental boats then reduced the number of MHs toeither 36 or 30 and continued fishing for 7.3 h on average.Experimental boats consumed approximately 242 l of fuelduring the average 3.9 h of full lighting, and during theensuing stage—the reduced lighting period—an average of

dura-294 l of fuel were consumed when 36 MHs were used, and

273 l when 30 MHs were used From these data, the averagefuel consumption rates were calculated to be 61.5 l/h duringthe full lighting period, 40.4 l/h with 36 MHs, and 37.6 l/hwith 30 MHs

The daily catches of the boats that used stage reducedlighting and conventional lighting are summarized inFig.4 Skewed distributions and wide variations were

observed in the catch amounts, regardless of lighting

method The median catch values were 72.5, 105, and 105

boxes for conventional lighting, 30 MHs, and 36 MHs,

respectively Nonparametric multiple comparisons for all

catch data suggested that the catch obtained using

con-ventional lighting was less than that obtained by stage

reduced lighting with 30 and 36 MHs (Steel–Dwass test,

P\ 0.05), and the number of MHs used after the full

lighting period did not affect the catch amount (P [ 0.05)

When the time sequence was considered, the catch amounts

obtained via stage reduced lighting in two different ments and conventional lighting were similar (Fig.5); allboats stopped fishing for several days around the full moon(January 30), and catch amounts peaked around the newmoon (January 15 and February 14)

treat-GLM analysis revealed that the AIC value was smallestwhen the parameters lunar, month, and light were included

in the model, as shown in Tables 1and2 The results of theGLM analysis demonstrated that light is less significant(P \ 0.05 only for 36 MHs), and that lunar exerted asignificant negative effect (P \ 0.001), suggesting that thecatch decreases during fishing operations around the fullmoon period In addition, the catch was significantly larger

in January than in February (P \ 0.001) Thus, in the GLManalysis, we adopted a model (Model 3-1 in Table2) inwhich catch amount significantly depended on lunar andmonth, with less of an influence of light Expected catchamounts from the adopted model are plotted against theobserved catch amounts in Fig.6 The catch amounts cal-culated from this model ranged between 41 and 206 boxes,and this model did not produce catch amounts that wereoutside of the range that was frequently observed duringthe experimental period

DiscussionOne advantage of stage reduced lighting is that it has thepotential to save fuel When a commercial operation isconducted under full lighting conditions (159 kW), which

is close to the maximum lighting power specified by thevoluntary regulation (160 kW) for 11.2 h (the averagelighting duration in this study), the total fuel consumptiondue to the lighting was estimated as 690 l However, thefuel consumption during commercial operations is

Full lighting After full lighting Full

lighting

Full lighting

36 MHs

Fig 3 Box-and-whisker plots showing the durations of the full

lighting period and the lighting period with reduced numbers of MHs

for the nine experimental boats when stage reduced lighting was used

(a), the amounts of fuel consumed during the full lighting period and

the lighting period with reduced numbers of MHs (b), and the fuel

consumption rates during the full lighting period and the lighting

period with reduced numbers of MHs (c) The band in the box is the median value and the bottom and top of the box indicate the lower and upper quartiles, respectively The ends of the whiskers represent the 1.5 interquartile range, and points depicted as open circles designate outliers

36 MHs

30 MHs Conventional

Fig 4 Box-and-whisker plots of the daily catch amounts (boxes) of

boats that used different lighting approaches (stage reduced lighting

with 30 MHs, stage reduced lighting with 36 MHs, or conventional

lighting) The band in the box is the median value and the bottom and

top of the box are the lower and upper quartiles, respectively The

ends of the whiskers represent the 1.5 interquartile range, and points

depicted as open circles designate the outliers These outliers were all

included in the catch analysis in the study

sometimes less than this value, because fishermen

occa-sionally reduce the number of MHs for about 1–2 h

(per-sonal communication with a captain who participated in the

experiment) In this case, the fuel reduction achieved by

this procedure is at least 123 l, which is the approximate

amount of fuel consumed during full lighting for 2 h When

typical durations of the full lighting period (3.9 h) and the

lighting period with 30 MHs or 36 MHs (7.3 h) are taken

into account, the stage reduced lighting is estimated to

consume 516 l for 30 MHs during jigging with lamps and

536 l for 36 MHs These estimated values are 174 and

154 l (22 and 25 %) less than the estimated maximumamount of fuel consumed by the conventional lighting.Unlike the summer fishing season [1], squid fishinggrounds in the Tsushima Strait in winter are relatively close

to the base ports, so fuel saving during jigging with lamps

is a management priority to improve profitability

Several studies have demonstrated that squid around ajigging boat are generally hooked in the shadow zonecreated by the boat hull [5 8] These findings suggest thatsquid shelter from strong light around the boat, eventhough this fishing technique applies the principle thatsquid are attracted to light We consider that full lightinginitially delivers light over a broad area, attracting squidschools to the area around the boat at the start of fishing.Once the squid schools are close to the boat, strong light(such as the maximum lighting power specified by thevoluntary regulation, 160 kW) may not be necessary.Stage reduced lighting appeared to lead to larger catchesthan conventional lighting (Fig.4), but the difference incatch between the conventional and the stage reducedlighting with 30 MHs was not significant in the GLManalysis (Table 2, P = 0.086) Therefore, stage reducedlighting is considered to potentially have the same catch

0 100 200 300 400 500 600

Full moon

Fig 5 Average catches of

boats that used the different

lighting approaches (with 30

MHs after full lighting, with 36

MHs after full lighting, and

conventional lighting) during

January and February 2010

Table 1 Models considered in the study, and fit results

Formula Null deviance Null df Residual deviance Residual df AIC h SE Model 0 C * 1 708.92 579 708.92 579 6464.4 0.5792 0.0329 Model 1-1 C * light 719.95 579 708.82 577 6457.4 0.5896 0.0337 Model 1-2 C * lunar 741.02 579 708.71 578 6434.9 0.6096 0.0350 Model 1-3 C * boat 725.88 579 708.79 570 6465.5 0.5952 0.0340 Model 1-4 C * month 736.54 579 708.75 578 6439.2 0.6053 0.0347 Model 2-1 C * lunar ? light 748.29 579 708.73 576 6432.1 0.6165 0.0355 Model 2-2 C * lunar ? boat 754.93 579 708.76 569 6439.9 0.6228 0.0359 Model 2-3 C * lunar ? month 750.66 579 708.78 577 6427.9 0.6187 0.0356 Model 3-1 C * lunar ? month ? light 758.67 579 708.83 575 6424.5a 0.6264 0.0362 Model 3-2 C * lunar ? month ? boat 765.24 579 708.88 568 6432.5 0.6326 0.0366 Model 4 C * lunar ? month ? light ? boat 765.24 579 708.88 567 6434.5 0.6326 0.0366

a Adopted as the model

Table 2 Estimated coefficients

Parameter Estimate SE Wald statistic P value

a Coefficient for catch obtained using conventional lighting during

the new moon in February

performance as conventional lighting The observed increase

in the catch with stage reduced lighting (Fig.4) may be due

to the difference in the light sources used by the experimental

and commercial boats Experimental boats partially

employed LEDs, which emit a certain range of wavelengths

(blue–blue green, 450–500 nm) that show good penetration

into the water [7, 9], whereas commercial boats use only

MHs, which emit light within other wavelength ranges Light

from LEDs penetrates into the water better than light from

MHs and reaches squid that are distributed further from the

boat and/or are in a deeper area

Unlike the results obtained in our previous study, in which

we performed a GLM analysis of the squid catch in summer

[2], we detected an influence of the illuminated fraction of the

moon Lunar rhythmicity in the catch amount is recognized

among fishermen who fish under artificial light, and they

generally suspend their fishing for several nights around the

full moon Our results demonstrated an influence of the lunar

phase on the catch amount, based on an analysis of the catch

data during two lunar phase cycles Regardless of the lighting

method, the catch amount tended to increase from the full

moon to the new moon, while the opposite tendency was seen

from the new moon to the full moon (Fig.5) This tendency

may be due to the direct influence of light in the environment,

but it may also be due to the internal rhythm of squid

(gov-erned by the lunar phases) [10,11] In addition, we did not

detect any influence of fishing power (due to differences in the

fishing skills of the fishermen) This result probably reflects

the unique characteristics of the fishing ground in Tsushima

Strait in winter Japanese common squid migrates from the

north for spawning in this season and forms dense

distributions in limited areas [12] Many squid jigging boatsconcentrate on limited areas in this season to capture squidwhile maintaining a sufficient distance (at least two nauticalmiles; personal communication with a captain who partici-pated in the experiment) from the next boat so as not to affectthe area influenced by its lighting Under such conditions, thechoice of fishing position may be restricted, and choosing agood fishing position is one of the most important fishing skills(and therefore one of the greatest influences on the fishingpower)

There may be other factors that we did not take intoaccount that help to explain the catch amount, becausethere was a much narrower range of expected catchamounts in the GLM analysis (41–206 boxes) than theactual observed range of catch amounts (0–659 boxes,Fig.6) One possible reason for this may be the influence

of the weather and sea conditions In a preliminary GLManalysis, we assumed that the hours of sunshine (JapanMetrological Agency Web:http://www.data.jma.go.jp/obd/

start-ing the operation described the weather conditions on theday, and when the catch amounts were compared with theweather, the weather appeared to have a significant influence

on the catch However, the number of hours of sunshineincidentally exhibited a positive correlation with the lunarphase, so we did not take this factor into account Furtherresearch involving the accumulation of data for longer periods

is necessary to evaluate the influence of the weather and seaconditions In addition, the foraging behavior of dolphins(generally Pacific white-sided dolphins Lagenorhynchusobliquidens and bottlenose dolphins Tursiops truncatus) in thefishing grounds is a concern to fishermen, as the squid aroundthe boat tend to disperse when dolphins are nearby Smallcatch datasets that are sometimes recorded by commercialboats may show the impact of dolphin behavior

Thus, stage reduced lighting using LEDs and MHs has thepotential to save fuel consumption by up to 25 % whencompared to conventional commercial lighting, while stillmaintaining the squid catch in Tsushima Strait in winter.Fishermen should therefore note that stage reduced lighting is

a promising method for improving profitability by saving fuel

Acknowledgments We are grateful to members of the Katsumoto Fisheries Cooperative for their help in collecting data We also thank the captains and crews of the squid jigging boats who participated in this experiment This study was carried out as a part of the Project on Promoting Energy Saving Technology, Fisheries Agency, Govern- ment of Japan.

References

1 Matsushita Y, Azuno T, Yamashita Y (2012) Fuel reduction for small squid jigging boats by equipping conventional metal halide lamps with combinations of LED panels Fish Res 125:14–19

Observed catch (boxes)

Experimental boats

Conventional boats

Fig 6 Expected catch (in boxes) plotted against the observed catch

(in boxes) for the adopted model (3-1) A solid line designates equal

values

2 Yamashita Y, Matsushita Y, Azuno T (2012) Catch performance

of coastal squid jigging boats using LED panels in combination

with metal halide lamps Fish Res 113:182–189

3 Venables WN, Dichmont CM (2004) GLMs, GAMs and

GLMMs: an overview of theory for applications in fisheries

research Fish Res 70:319–337

4 Venables WN, Ripley BD (2002) Modern applied statistics with

S Springer, New York

5 Inada H, Ogura M (1988) Historical change of fishing light and

its operation in squid jigging fisheries Rep Tokyo Univ Fish

24:189–207 (in Japanese, with English abstract)

6 Inada H (1996) Retinomotor response and retinal adaptation of

Japanese common squid Todarodes pacificus at capture with jigs.

Fish Sci 62:663–669

7 Arakawa H, Choi S, Arimoto T, Nakamura Y (1998)

Relation-ships between underwater irradiance and distribution of Japanese

common squid under fishing lights of squid jigging boat Fish Sci

64:553–557

8 Shikata T, Shima T, Inada H, Miura I, Daida N, Sadayasu K,

Watanabe T (2011) Role of shaded area under squid jigging boat

formed by shipboard fishing light in the process of gathering and capturing Japanese common squid, Todarodes pacificus Nippon Suisan Gakkaishi 77:53–60 (in Japanese, with English abstract)

9 Arakawa H, Choi S, Arimoto T, Nakamura Y (1996) Underwater irradiance and distribution of fishing lights used by small-type squid jigging boat Nippon Suisan Gakkaishi 62:420–427 (in Japanese, with English abstract)

10 Scho¨n PJ, Sauer WHH, Roberts MJ (2002) Environmental influences on spawning aggregations and jig catches of chokka squid Loligo vulgaris reynaudii: a ‘‘black box’’ approach Bull Mar Sci 71:783–800

11 Postuma FA, Gasalla MA (2010) On the relationship between squid and the environment: artisanal jigging for Loligo plei at Sa˜o Sebastia˜o Island (24°S), southeastern Brazil ICES J Mar Sci 67:1353–1362

12 Sakurai Y, Kiyofuji H, Saitoh S, Goto T, Hiyama Y (2000) Changes in inferred spawning areas of Todarodes pacificus (Cephalopoda: Ommastrephidae) due to changing environmental conditions ICES J Mar Sci 57:24–30

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

antiporter genes in the marine red alga Porphyra yezoensis

Toshiki Uji•Ryota Monma•Hiroyuki Mizuta•

Naotsune Saga

Received: 23 April 2012 / Accepted: 29 May 2012 / Published online: 20 June 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract Na?/H?antiporters are known to play a crucial

role in pH and Na?homeostasis In the present study, we

characterized the molecular structures and expression

pat-terns of two Na?/H?antiporters from the marine red alga

Porphyra yezoensis (designated PySOS1 and PyNhaD)

The full-length cDNAs of PySOS1 and PyNhaD were 5122

and 1804 bp, and contained open reading frames (ORFs) of

4773 and 1275 bp, respectively The deduced amino acid

sequences showed high similarity to SOS1 and NhaD from

the higher plant Arabidopsis thaliana PySOS1 and

PyN-haD contained conserved sequences found in the cation–

proton antiporter Quantitative real-time PCR analysis

revealed that both antiporter genes were expressed in both

the gametophyte and sporophyte of P yezoensis In

addi-tion, mRNA expression of PySOS1 and PyNhaD was

simultaneously upregulated by light irradiation, suggesting

that coordinated activity between the two is important in

pH and Na?homeostasis under light conditions Moreover,

the expression levels of both genes were partially reduced

by the photosynthetic inhibitors DCMU and DBMIB,

suggesting that upregulation is linked to

photosynthesis-related metabolism These findings provide an initial step

towards understanding Na?/H? antiporters in marine red

algae

Keywords Marine macroalga Na?/H?antiporter 

Porphyra yezoensis Red alga

Introduction

Na?/H?antiporters are found in all domains of life and havebeen shown to play important roles in cell homeostasis,including pH regulation, Na?tolerance, and osmoregulation,and in vesicle trafficking and the control of the cell cycleand cell proliferation [1 3] In land plants, three types of

Na?/H? antiporter have been identified One SOS1-type

Na?/H? antiporter, AtSOS1, was initially identified as agene locus required for salt tolerance in A thaliana [4], andloss-of-function mutations in AtSOS1 were shown to result

in extreme salt sensitivity and overaccumulation of Na?inshoots under salt stress compared with the wild type [5].AtSOS1 transcription is specifically upregulated upon NaClstress [5] and AtSOS1-GFP fusion proteins localize at theplasma membrane [6] The Nhx1-type Na?/H?antiporter islocated at the vacuolar membrane, and the expression of thisgene is induced by salt and osmotic stress [7] Moreover,transgenic plants that overexpress NHX1 are highly tolerant

of salt stress [8] NhaD-type Na?/H? antiporters, whichwere first identified in Vibrio parahaemolyticus [9], arepotential candidates for Na?transport in plant organelles Inplants, NhaD transporters have been characterized in Pop-ulus euphratica [10] and Physcomitrella patens [11], and in

P patens they have been localized in chloroplasts [11].However, while data on Na?/H?antiporters in land plantshave been accumulating, little is known about their coun-terparts in marine macroalgae

The marine red alga Porphyra yezoensis has been posed as a model organism for physiological and molecularbiological studies of marine algae because of its biologicaland economical importance [12,13] This species inhabitsthe intertidal zone where the physiological environmentrapidly changes with the turning tides, resulting in expo-sure to abiotic stresses such as salinity, pH, temperature,

pro-T Uji  H Mizuta  N Saga (&)

Faculty of Fisheries Sciences,

Hokkaido University, Hakodate 041-8611, Japan

e-mail: nsaga@fish.hokudai.ac.jp

R Monma

Graduate School of Fisheries Sciences,

Hokkaido University, Hakodate 041-8611, Japan

Fish Sci (2012) 78:985–991

DOI 10.1007/s12562-012-0520-6

and light To elucidate the molecular mechanisms

regu-lating the responses to environmental stresses in marine red

algae, we previously performed expressed sequence

anal-ysis (EST) analanal-ysis [14,15] and developed transient gene

expression systems in P yezoensis [16–18]

In the P yezoensis EST database (http://est.kazusa.or.jp/

homol-ogous to SOS1- and NhaD-type Na?/H? antiporters

(designed as PySOS1 and PyNhaD) were found by

Barre-ro-Gil et al [11,19]; however, the full-length cDNAs have

not yet been obtained The aim of the present study,

therefore, was to clone and characterize the full-length

cDNAs of PySOS1 and PyNhaD to further our

under-standing of Na?/H? antiporters in marine macroalgae

Subsequently, expression analyses of two Na?/H?

anti-porter genes were carried out To our knowledge, this is the

first report to perform expression analyses of Na?/H?

antiporters in marine macroalgae

Materials and methods

Algal material and abiotic stress treatment

The leafy gametophytes and filamentous sporophytes of

P yezoensis strain TU-1 were cultured as previously

described [18] The cultured algae were used to clone

PySOS1 and PyNhaD, and for transcriptional analysis of the

two genes under salt stress, acid stress, and the light–dark

cycle Salt and acid stresses were applied by treating

vege-tative gametophytes and sporophytes with enriched sealife

(ESL) medium containing 0.3 M NaCl or adjusted to pH 6.0

through the addition of 6 N HCl under continuous light

conditions Light treatment was carried out by incubating at

a light intensity of 80 lmol m-2s-1 The light/dark cycle

consisted of 10 h light and 14 h dark

3-(3,4-Dichloro-phenyl)-1,1-dimethylurea (DCMU) and

2,5-dibromo-3-methyl-6-isopropylbenzoquinone (DBMIB) treatments were

performed from 30 min prior to light irradiation Algal

material was placed directly into liquid nitrogen and stored

at -80°C until it was used in rapid amplification of 50

cDNA ends (50RACE) and transcription analysis

Cloning of full-length cDNAs encoding Na?/H?

antiporters from P yezoensis

Based on the partial sequences of PySOS1 (EST accession

no AU192232) and PyNhaD (EST accession no AV438

891), cDNA 50ends were obtained by 50RACE Total RNA

extraction from the gametophytes was carried out with an

RNeasy plant mini kit (Qiagen, Hilden, Germany), and the

resultant total RNA was further purified with an RNase-Free

DNase Set (Qiagen) First-strand cDNA was synthesized

with a SMART RACE cDNA amplification kit (Clontech,Mountain View, CA, USA) and SuperScript II reversetranscriptase (Invitrogen, Carlsbad, CA, USA) according tothe manufacturers’ instructions The cDNA 50 ends wereamplified by polymerase chain reaction (PCR) with TaKaRa

LA Taq (GC buffer type) (TaKaRa Bio, Shiga, Japan) withthe primer SOS1-RACE-R1 (Table1) The 50 end of PyN-haD cDNA was also amplified by PCR in the same mannerwith the primer NhaD-RACE-R1 The PCR products wereseparated by electrophoresis with 1.0 % agarose gel, and theamplification products were excised and purified with aQIAquick gel extraction kit (Qiagen) Purified PCR productswere cloned into the pT7Blue vector (Merck, Darmstadt,Germany), and nucleotide sequences were determined with

an ABI 3130xl genetic analyzer (Applied Biosystems/LifeTechnologies, Carlsbad, CA, USA)

Transcript analysis of PySOS1 and PyNhaDTotal RNA extraction from gametophytes and sporophyteswas carried out with an RNeasy plant mini kit (Qiagen), andthe resultant total RNA was further purified with a TURBODNA-free kit (Applied Biosystems/Life Technologies).Total RNA (0.5 lg per reaction) was used for first-strandcDNA synthesis with a PrimeScript II 1st strand cDNASynthesis Kit (TaKaRa Bio) Real-time PCR was performedwith an ABI Prism 7300 sequence detection system andsoftware (Applied Biosystems/Life Technologies) All real-time PCR was performed under the following conditions:

30 s at 95°C followed by 40 cycles of 5 s at 95 °C and 31 s

at 60°C The gene-specific primers used for real-time PCR(SOS1-q-F/R, NhaD-q-F/R and 18SrRNA-q-F/R) are listed

in Table1 The specificity of the PCR products was firmed by analyzing the dissociation curve at the end of eachreaction (15 s at 95°C, 1 min at 60 °C, and 15 s at 95 °C).The reaction mixture (20 ll) contained 10 ll SYBR Premix

con-Ex Taq GC (TaKaRa Bio), 0.8 ll of each forward andreverse primer (5 lM), 0.4 ll of ROX reference dye, and

4 ll of cDNA template (100-fold dilution) The Py18SrRNAgene, whose transcriptional activity does not significantlyfluctuate over the course of a whole day, was used as aninternal control to normalize the amount of mRNA in eachreaction The mRNA amounts of the PySOS1 and PyNhaDgenes were calculated based on a standard curve Thestandard curve for each primer set was prepared by plottingserial cDNA dilution (1:10–1:105) against CT (thresholdcycle) All experiments were conducted in triplicate.Phylogenetic analysis of the Na?/H?antiporters

A phylogenetic tree was constructed using the joining method with MEGA version 5.0 (http://www

Na?/H? antiporters were retrieved from the Genbank

database

In silico analyses of Na?/H?antiporter

The BLAST program was used to identify homologous

sequences in the GenBank database Hydrophobicity

pro-files of PySOS1 and PyNhaD were analyzed using SOSUI

of PySOS1 and PyNhaD

Results

Characterization of PySOS1 and PyNhaD

50RACE was performed to obtain the full-length cDNAs of

PySOS1 and PyNhaD The full-length cDNA of PySOS1

was 5122 bp in length and contained an open reading frame

(ORF) of 4773 bp, while that of PyNhaD was 1804 bp with

an ORF of 1275 bp PySOS1 and PyNhaD were shown to

encode putative proteins of 1591 aa and 425 aa,

respec-tively The complete cDNA sequences and deduced amino

acid sequences were deposited into GenBank under the

accession numbers AB694755 and AB694756, respectively

BLAST analysis revealed that the deduced amino acid

sequences of PySOS1 and PyNhaD share 55 % similarity

with SOS1 from A thaliana and 76 % similarity with

NhaD from A thaliana, respectively The hydrophobicity

profiles of the amino acid sequences of PySOS1 and

PyNhaD predicted 12 and 10 membrane-spanning regions,

respectively In addition, we used WoLF PSORT to predict

the subcellular localization of PySOS1 and PyNhaD based

on their amino acid sequences PSORT predicted the

localization of PySOS1 in the plasma membrane, but was

unable to localize PyNhaD

Several characteristic sequences have been identified in

SOS1- and NhaD-type Na?/H?antiporters For example,

Hamada et al [20] reported that Asp 138 in the Na?/H?

antiporter from Synechocystis sp PCC 6803 (that is highly

homologous to AtSOS1 and named SynNhaP), which isinvolved in exchange activity, is conserved in the SOS1-type Na?/H? antiporter The conserved amino acid wasalso found in PySOS1 (Fig.1) In NhaD-type Na?/H?antiporters, Habibian et al [21] reported that the residuesSer 150, Asp 154, Asn 155, Asn 189, Asp 199, Thr 201,Thr 202, Ser 389, Asn 394, Ser 428, and Ser 431 of NhaDfrom Vibrio cholera (named VcNhaD) are involved in thereaction of the cation–proton antiporter These amino acidswere also strongly conserved in PyNhaD (Fig.2) Phylo-genetic tree analysis revealed that PySOS1 and PyNhaDwere located in SOS1-type and NhaD-type clusters,respectively (Fig 3)

Expression patterns of PySOS1 and PyNhaD

P yezoensis has a heteromorphic life cycle with a scopic leafy gametophyte and microscopic filamentoussporophyte The expression patterns of PySOS1 and PyN-haD in gametophytes and sporophytes were thereforeanalyzed by real-time PCR As shown in Fig.4, the mRNAtranscripts of the two Na?/H? antiporters were expressed

macro-in both generations

Next, we examined whether PySOS1 and PyNhaD wereresponsive to salt stress, acid stress (pH 6.0), and the light–dark cycle using the gametophyte Expression patterns ofthe two genes in response to salt stress and acid stress (pH6.0) did not fluctuate under continuous light conditions(data not shown) On the other hand, the transcript level ofPySOS1 rapidly increased after 4 h of light, peaked at 2 h

of dark, and subsequently decreased to its baseline level at

14 h of dark (Fig.5) The expression pattern of PyNhaDwas similar to that of PySOS1 (Fig.5) The maximumexpression levels of PySOS1 and PyNhaD were approxi-mately 5.0-fold and 65.2-fold higher than the baseline,respectively In addition, in the sporophyte, the expressionpatterns of both genes were similar to the gametophyteresults (data not shown)

To elucidate whether photosynthesis affects the dependent accumulation of PySOS1 and PyNhaD tran-scripts, the effect of the photosynthetic inhibitors DCMU

light-Table 1 List of used PCR

primers Primer name Sequence (5

0 –30) Description SOS1-RACE-R1 GATGTAGTCCTGCCACTCGTG 50RACE NhaD-RACE-R1 TGATCTCGCGGCTAGCAATATGCG 5 0 RACE SOS1-q-F TCAATGTCGTGGAGCTGGG Real-time PCR SOS1-q-R CGCCTTGAAAAGGTCCTCG Real-time PCR NhaD-q-F GGTCGCTGTCTACCTTGATGC Real-time PCR NhaD-q-R ACCAGGCCAATGACTGCAG Real-time PCR 18SrRNA-q-F TGATAGTCCTGGGTCGGAAG Real-time PCR 18SrRNA-q-R TGATGACCTGCGCCTACAAG Real-time PCR

Fig 1 Structural characteristics of PySOS1 Alignment of the

deduced amino acid sequences of SOS1-type Na ? /H ? antiporters

and SynNhaP The sequences were aligned using ClustalW The

conserved Asp (Asp 138 in SynNhaP) is shown by an arrow Black

shading indicates 100 % conserved amino acid residues, dark gray

indicates 80 % conservation, and light gray 60–80 % conservation.

Bars represent gaps Numbers correspond to amino acid positions

from the first methionine residue PySOS1, Porphyra yezoensis (AB694755); AtSOS1, Arabidopsis thaliana (NM_126259); PeSOS1, Populus euphratica (DQ517530); OsSOS1, Oryza sativa (AY785147); PpSOS1, Physcomitrella patens (AM707025); SynN- haP, Synechocystis sp PCC 6803 (NP_441245) The amino acid sequences were retrieved from NCBI ( http://www.ncbi.nlm.nih.gov/ )

Fig 2 Structural characteristics of PyNhaD Alignment of the

deduced amino acid sequences of NhaD-type Na?/H? antiporters.

The sequences were aligned using ClustalW The amino acids

involved in the reaction of the cation–proton antiport (VcNhaD) are

shown by arrowheads Black shading indicates 100 % conserved

amino acid residues, dark gray indicates 80 % conservation, and

light gray 60–80 % conservation Bars represent gaps Numbers

correspond to amino acid positions from the first methionine residue PyNhaD, Porphyra yezoensis (AB694756); AtNhD1, Arabidopsis thaliana (NP_566638); PeNhaD1, Populus euphratica (AJ561195); PpNhaD, Physcomitrella patens (AM491807); VcNhaD, Vibrio cholerae (AAG48354) The amino acid sequences were retrieved from NCBI ( http://www.ncbi.nlm.nih.gov/ )

and DBMIB was investigated We used 10 lM DCMU and

10 lM DBMIB, since it has been reported that

0.05–10 lM DCMU and 8–10 lM DBMIB effectively

inhibited the photosynthetic activity of marine macroalgae

[22, 23] Figure6 shows that the levels of the two scripts were partially reduced by both inhibitors and thatPyNhaD is more sensitive to the inhibitors than PySOS1

tran-Discussion

In the present study, we cloned full-length cDNAs of two

Na?/H?antiporters, PySOS1 and PyNhaD Both possessedconserved amino acid sequences found in Na?/H? anti-porters (Fig.1), showing them to be functional proteins.Computational analysis suggests that PySOS1 localizes tothe plasma membrane, similar to SOS1 in higher plantssuch as A thaliana and P euphratica [24], whereas thesubcellular localization of PyNhaD could not be predicted.Previous subcellular localization analysis using GFPshowed that NhaD from P patens, PpNhaD1, localizes tothe chloroplast [11], raising the possibility that PyNhaDdoes so too As shown in Fig.6, the effect of photosyn-thetic inhibitors on light-dependent accumulation of thePyNhaD transcripts was more pronounced than that of thePySOS1 transcript, indicating that PyNhaD expression ismore closely related to chloroplasts than PySOS1 expres-sion However, this close relationship does not necessarilyindicate the localization of PyNhaD in the chloroplasts Toconfirm this, subcellular localization analysis of PyNhaDusing GFP is now required

SOS1- and NhaD-type Na?/H?antiporters are known to

be upregulated in response to salt stress [5,24,25]; however,transcripts of PySOS1 and PyNhaD were not upregulated inresponse to salt stress under continuous illumination(Fig.5) The results raise the possibility that the amounts ofthe two gene transcripts reached their maximum values uponlight irradiation, regardless of salt stress Interestingly, it wasreported in Porphyra leucosticta that an increase in theintracellular Na? content was observed after the onset oflight [26] Also, in P yezoensis cells, the Na? content ispredicted to increase under illumination with light Thus, thesimultaneous upregulation of PySOS1 and PyNhaD afterlight irradiation may play an important role in Na?homeostasis under light conditions However, we need toinvestigate PySOS1 and PyNhaD expression under salt stress

in the continuous dark further

In land plants and green algae, it has been shown thatphotosynthesis is responsible for the regulation of geneexpression by light [27–29] For example, light-dependentaccumulation of superoxide dismutase genes from a liver-wort was inhibited by both DCMU and DBMIB [30],which inhibits the flow of electrons from PSII to plasto-quinone and the flow after plastoquinone by binding to thecytochrome b6/f complex, respectively [31] As shown inFig.6, levels of the PySOS1 and PyNhaD transcripts werepartially reduced during incubation in the presence of

Fig 3 Phylogenetic tree of Na ? /H ? antiporters The phylogenetic

tree was constructed based on the amino acid sequences of the ORF of

Na?/H? antiporters using the neighbor-joining method Branch

lengths indicate evolutionary distance, with a scale of 0.2 Species

and accession numbers (in alphabetical order): AtNhD1, Arabidopsis

thaliana; AtNhx1, A thaliana (NP_198067); AtSOS1, A thaliana;

OsNhaD, O sativa (BAD17583); OsNhx1, O sativa (AB021878);

OsSOS1, O sativa; PeNhaD1, Populus euphratica; PeNhx1,

P euphratica (AJ853472); PeSOS1, P euphratica; PpNhaD,

P patens; PpSOS1, P patens; PySOS1, Porphyra yezoensis;

PyNhaD, P yezoensis; ScNha1, Saccharomyces cerevisiae

(NP_013239); ScNhx1, S cerevisiae (NP_010744); SynNhaP,

Syn-echocystis sp PCC 6803; VcNhaD, Vibrio cholerae; VpNhaD, Vibrio

parahaemolyticus (BAA25994)

Fig 4 Expression of PySOS1 and PyNhaD transcripts in Porphyra

yezoensis gametophytes and sporophytes RNA prepared from

gametophytes cultured at 15 °C (G) and sporophytes cultured at

15 °C (S) was used for quantitative real-time PCR The 18S rRNA

gene from P yezoensis (Py18SrRNA) was used as an internal control.

Results represent the relative expression compared to that of the

PySOS1 gene in the gametophyte Data are presented as the

mean ± standard deviation (n = 3)

DCMU and DBMIB under light irradiation, respectively,

suggesting that their upregulation is linked to

photosyn-thesis-related metabolism However, little is known about

the transcriptional regulatory mechanisms by light in red

algae Promoter analyses of PySOS1 and PyNhaD will

therefore contribute to our understanding of the

mecha-nisms regulated by light in red algae

Acknowledgments This study was supported, in part, by the

Regional Innovation Cluster Program (Global Type) of the Ministry

of Education, Culture, Sports, Science and Technology of Japan,

awarded to N.S.

References

1 Brett CL, Donowitz M, Rao R (2005) Evolutionary origins of

eukaryotic sodium/proton exchangers Am J Physiol Cell Physiol

288:223–239

2 Mitsui K, Yasui H, Nakamura N, Kanazawa H (2005) merization of the Saccharomyces cerevisiae Na ? /H ? antiporter Nha1p: implications for its antiporter activity Biochim Biophys Acta 1720:125–136

Oligo-3 Bassil E, Ohto MA, Esumi T, Tajima H, Zhu Z, Cagnac O, Belmonte M, Peleg Z, Yamaguchi T, Blumwald E (2011) The Arabidopsis intracellular Na?/H?antiporters NHX5 and NHX6 are endosome associated and necessary for plant growth and development Plant Cell 23:224–239

4 Wu SJ, Ding L, Zhu JK (1996) SOS1, a genetic locus essential for salt tolerance and potassium acquisition Plant Cell 8:617–627

5 Shi HZ, Ishitani M, Kim CS, Zhu JK (2000) The Arabidopsis thaliana salt tolerance gene SOS1 encodes a putative Na?/H?antiporter Proc Natl Acad Sci USA 97:6896–6901

6 Shi HZ, Quintero FJ, Pardo JM, Zhu JK (2002) The putative plasma membrane Na ? /H ? antiporter SOS1 controls long-dis- tance Na ? transport in plants Plant Cell 14:465–477

7 Yokoi S, Quintero FJ, Cubero B, Ruiz MT, Bressan RA, gawa PM, Pardo JM (2002) Differential expression and function

Hase-of Arabidopsis thaliana NHX Na ? /H ? antiporters in the salt stress response Plant J 30:529–539

8 Apse MP, Aharon GS, Snedden WA, Blumwald E (1999) Salt tolerance conferred by overexpression of a vacuolar Na?/H?antiport in Arabidopsis Science 285:1256–1258

9 Nozaki K, Kuroda T, Mizushima T, Tsuchiya T (1998) A new

Na?/H?antiporter, NhaD, of Vibrio parahaemolyticus Biochim Biophys Acta 1369:213–220

10 Ottow EA, Polle A, Brosche M, Kangasjarvi J, Dibrov P, Zorb C, Teichmann T (2005) Molecular characterization of PeNhaD1: the first member of the NhaD Na?/H? antiporter family of plant origin Plant Mol Biol 58:75–88

11 Barrero-Gil J, Rodriguez-Navarro A, Benito B (2007) Cloning of the PpNHAD1 transporter of Physcomitrella patens, a chloroplast transporter highly conserved in photosynthetic eukaryotic organisms J Exp Bot 58:2839–2849

12 Saga N, Kitade Y (2002) Porphyra: a model plant in marine sciences Fish Sci 68(Suppl):1075–1078

13 Waaland JR, Stiller JW, Cheney DP (2004) Macroalgal dates for genomics J Phycol 40:26–33

candi-14 Asamizu E, Nakajima M, Kitade Y, Saga N, Nakamura Y, Tabata

S (2003) Comparison of RNA expression profiles between the two generations of Porphyra yezoensis (Rhodophyta), based on expressed sequence tag frequency analysis J Phycol 39:923–930

15 Nikaido I, Asamizu E, Nakajima M, Nakamura Y, Saga N, Tabata S (2000) Generation of 10,154 expressed sequence tags

Fig 5 Expression of PySOS1 and PyNhaD transcripts in Porphyra

yezoensis gametophytes under the light/dark cycle Algal material was

sampled every 4 h under a photoperiod of 10 h L:14 h D The

Py18SrRNA gene was used as an internal control Results represent

the relative expression compared to that at the start of the experimental culture period Data are presented as the mean ± stan- dard deviation (n = 3)

Fig 6 Effect of DCMU and DBMIB on the expression of PySOS1

and PyNhaD in Porphyra yezoensis gametophytes Algal material

grown under a photoperiod of 10 h L:14 h D was sampled 8 h post

light irradiation and then treated with 0.01 % DMSO (Control),

10 lM DCMU, or 10 lM DBMIB from 30 min prior to light

irradiation The Py18SrRNA gene was used as an internal control.

Results represent the relative expression compared to that in 0.01 %

DMSO (Control) Data are presented as the mean ± standard

deviation (n = 3)

from a leafy gametophyte of a marine red alga, Porphyra

yezo-ensis DNA Res 7:223–227

16 Fukuda S, Mikami K, Uji T, Park E-J, Ohba T, Asada K, Kitade

Y, Endo H, Kato I, Saga N (2008) Factor influencing efficiency of

transient gene expression in the red macrophyte Porphrya

yezo-ensis Plant Sci 174: 329–339

17 Mikami K, Uji T, Li L, Takahashi M, Yasui H, Saga N (2009)

Visualization of phosphoinositides via the development of the

transient expression system of a cyan fluorescent protein in the

red alga Porphyra yezoensis Mar Biotechnol 11:563–569

18 Uji T, Takahashi M, Saga N, Mikami K (2010) Visualization of

nuclear localization of transcription factors with cyan and green

fluorescent proteins in the red alga Porphyra yezoensis Mar

Biotechnol 12:150–159

19 Barrero-Gil J, Garciadeblas B, Benito B (2005) Sodium,

potas-sium-atpases in algae and oomycetes J Bioenerg Biomembr

37:269–278

20 Hamada A, Hibino T, Nakamura T, Takabe T (2001) Na?/H?

antiporter from Synechocystis species PCC 6803, homologous to

SOS1, contains an aspartic residue and long C-terminal tail

important for the carrier activity Plant Physiol 125:437–446

21 Habibian R, Dzioba J, Barrett J, Galperin MY, Loewen PC,

Dibrov P (2005) Functional analysis of conserved polar residues

in Vc-NhaD, Na?/H?antiporter of Vibrio cholerae J Biol Chem

280:39637–39643

22 Barros MP, Necchi O Jr, Colepicolo P, Pederse´n M (2006)

Kinetic study of the plastoquinone pool availability correlated

with H2O2release in seawater and antioxidant responses in the

red alga Kappaphycus alvarezii exposed to single or combined

high light, chilling and chemical stresses Biochim Biophys Acta

1757:1520–1528

23 Hsu YT, Lee TM (2012) Modulation of gene expression of otene biosynthesis-related protein by photosynthetic electron transport for the acclimation of intertidal macroalga Ulva fasciata

car-to hypersalinity and excess light Physiol Plant 144:225–237

24 Wu Y, Ding N, Zhao X, Zhao M, Chang Z, Liu J, Zhang L (2007) Molecular characterization of PeSOS1: the putative Na?/H?antiporter of Populus euphratica Plant Mol Biol 65:1–11

25 Kurz M, Brunig AN, Galinski EA (2006) NhaD type sodium/ proton-antiporter of Halomonas elongata: a salt stress response mechanism in marine habitats? Saline Syst 2:10

26 Escassi L, Aguilera J, Figueroa FL, Ferna´ndez JA (2002) sium drives daily reversible thallus enlargement in the marine red alga Porphyra leucosticta (Rhodophyta) Planta 214:759–766

Potas-27 Maxwell DP, Laudenbach DE, Huner NPA (1995) Redox lation of light-harvesting complex II and cab mRNA abundance

regu-in Dunaliella salregu-ina Plant Physiol 109:787–795

28 Tullberg A, Alexciev K, Pfannschmidt T, Allen JF (2000) tosynthetic electron flow regulates transcription of the psaB gene

Pho-in pea (Pisum sativum L.) chloroplasts through the redox state of the plastoquinone pool Plant Cell Physiol 41:1045–1054

29 Harashima S, Takano H, Ono K, Takio S (2004) Chalcone synthase-like gene in the liverwort, Marchantia paleacea var diptera Plant Cell Rep 23:167–173

30 Sakaguchi S, Fukuda T, Takano H, Ono K, Takio S (2004) Photosynthetic electron transport differentially regulates the expression of superoxide dismutase genes in liverwort, March- antia paleacea var diptera Plant Cell Physiol 45:318–324

31 Fey V, Wagner R, Bra¨utigam K, Pfannschmidt T (2005) synthetic redox control of nuclear gene expression J Exp Bot 56:1491–1498

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

Culture method and growth characteristics of marine benthic

dinoflagellate Ostreopsis spp isolated from Japanese coastal

waters

Haruo Yamaguchi• Yuko Tanimoto•Takamichi Yoshimatsu•

Shinya Sato•Tomohiro Nishimura•Keita Uehara •

Masao Adachi

Received: 24 January 2012 / Accepted: 15 June 2012 / Published online: 7 July 2012

Ó The Japanese Society of Fisheries Science 2012

Abstract Blooms of toxic dinoflagellates of the genus

Ostreopsis, which is known as a producer of palytoxin

(PTX) analogs, may pose a threat to human health in

tropical, subtropical, and temperate regions around the

world In the present study, we established a suitable

cul-ture method for Ostreopsis spp isolated from Japanese

coastal waters and characterized their growth potential

using the method to discuss their bloom dynamics Each

clonal strain of Ostreopsis cf ovata, Ostreopsis sp 1,

Ostreopsis sp 5, and Ostreopsis sp 6 was incubated in

25 9 150 mm test tubes with a flat bottom containing

various kinds of medium Since Ostreopsis spp strains

grew well in IMK and/or f/2 media, we selected these

media for cultivation of all the Ostreopsis spp isolates

Growth rates of O cf ovata (0.834 divisions/day),

Ostre-opsis sp 1 (0.619 divisions/day), and OstreOstre-opsis sp 6

(1.04 divisions/day) that produce PTX analogs

signifi-cantly differed (p \ 0.05) respectively and are clearly

higher than those of other reported epiphytic dinoflagellate

Gambierdiscus toxicus, Prorocentrum lima, and Coolia

monotis cultures, which suggest that these species haveecological advantages to predominate through the algalsuccession in Japanese coastal waters, resulting in apotential risk to human health in this region

Keywords Culture method Growth characteristics Ostreopsis Palytoxin

IntroductionPalytoxin (PTX) and its analogs in marine organisms such

as the toxic zoanthid Palythoa [1] and dinoflagellate treopsis [2] can be bioaccumulated in finfish [3 5], crabs[6], sea urchins [7], and shellfish [8] via food webs [9,10]and also be spread as certain aerosols [11] These PTXanalogs may affect humans through seafood poisoning [4

Os-7, 12–16] and even through exposure to the poisonousaerosols [11, 17] In Japan, the risk profile of PTX-likepoisoning summarized by the Ministry of Health, Labor,and Welfare (MHLW; http://www.mhlw.go.jp/topics/syok

2012) shows 36 incidents of the occurrence of PTX-likepoisoning which have involved 6 human fatalities and 116reported illnesses since 1953 [5] In the case of Italy,aerosols containing PTX analogs occurred along thecoastline of Genoa in 2005 and caused serious respiratorydistress and skin irritation in people around the regionthrough inhalation of the aerosols [11, 18] These eventspotentially caused by PTX analogs have posed a threat topublic health

A marine benthic dinoflagellate, genus Ostreopsis,capable of synthesizing PTX and PTX analogs is widelydistributed in coastal environments in tropical, subtropical,and temperate regions of the world [15,19–22] They are

H Yamaguchi  T Yoshimatsu  K Uehara  M Adachi ( &)

Faculty of Agriculture, Kochi University, 200-Otsu,

Monobe, Nankoku, Kochi 783-8502, Japan

e-mail: madachi@kochi-u.ac.jp

Y Tanimoto  T Nishimura

The United Graduate School of Agricultural Sciences,

Ehime University, 3-5-7 Tarumi, Matsuyama,

usually present as benthic cells on seaweed, sand, rocks, and

invertebrates [15,19,20] and also as planktonic cells in the

water column [15,23] Blooms of toxic Ostreopsis spp have

been recognized to be the potential causative agent for some

outbreaks of PTX-like poisonous fish [5,13,14] and

aero-sols [11] Taniyama et al [5] have found PTX-like activities

in parrotfish Scarus ovifrons collected from a coast of

wes-tern Japan after the increase of abundance of Ostreopsis sp

and found nonpoisonous fish when the dinoflagellate

abun-dance was very low in the region Around the coast of Genoa

in 2005, aerosols containing PTX analogs occurred, being

associated with massive blooms of Ostreopsis ovata [11]

Laboratory experiments have found high contents of PTX

analogs in cultures of O ovata collected from the coastal

environments during blooms [11,18] These observations

imply that blooms of toxic Ostreopsis spp may be

respon-sible for human health problems Thus, it is important to

clarify the bloom dynamics of Ostreopsis to assess and

minimize the risk to human health

Recent works have clarified that the genus Ostreopsis

consists of various ‘‘morphospecies’’ as well as

phylo-groups (=clades): O ovata/O cf ovata, Ostreopsis

len-ticularis, and Ostreopsis siamensis clades [21,22] Around

the coast of south-western Japan, Ostreopsis labens [22,

24], O ovata, O siamensis [2, 6,19,25], and Ostreopsis

sp [5] have been reported so far Recently, we conducted

phylogenetic analyses of strains from Japanese coasts using

sequences of large subunit (LSU) ribosomal DNA (rDNA)

(D8-D10) as well as the internal transcribed spacer (ITS)

region of the nuclear rDNA and found 5 clades: O cf

ovata (clade A), Ostreopsis sp 1 (clade B), Ostreopsis

sp 2, Ostreopsis sp 5 (clade C), and Ostreopsis sp 6

(clade D) clades [26] Moreover, we detected PTX analogs

in several strains of O cf ovata, Ostreopsis sp 1, and

Ostreopsis sp 6 (Suzuki et al., submitted data) Since the

growth potential (=maximal growth rate and cell yield) of

an Ostreopsis sp plays a key role in the predomination

process of the organism over other benthic algal species, it

is important to clarify the growth characteristics of the

toxic Ostreopsis which are responsible for a potential risk

to human health in Japan

To our knowledge, studies quantifying and

characteriz-ing the growth potential of the genus Ostreopsis by culture

experiments have been limited to some species: Ostreopsis

lenticularis from the Puerto Rico coast [27], Ostreopsis

heptagona and O siamensis from the Florida coast [28],

and O ovata/O cf ovata from the Mediterranean Sea [29–

32] and from the coastal waters of Japan [33] and Rio de

Janeiro [34] Recently, Vidyarathna and Grane´li [33]

reported the maximal growth rate (0.7 divisions/day) and

cell yield (4 9 103cells/ml) of a strain of Japanese O cf

ovata using f/10 medium The nutrient contents are much

lower than other media such as f/2 and SWM-3 used

generally in planktonic dinoflagellate cultures We inferthat the growth potential of O cf ovata cultures reported

by them [33] could not be estimated precisely, because thelack of nutrients in f/10 medium probably limited thegrowth of the organism Since suitable culture methods forthe organisms have not been established yet, the growthpotential of Ostreopsis spp collected from the Japanesecoast has been poorly understood

In the present experiments, we selected suitable media forcultivation of Ostreopsis spp from Japanese coastal areasand then estimated the growth rates, cell yields, and bio-masses under laboratory conditions Using the data obtainedhere, we quantified and comparatively characterized thegrowth potential of O cf ovata, Ostreopsis sp 1, Ostreopsis

sp 5, and Ostreopsis sp 6 to discuss the bloom dynamics ofOstreopsis spp in coastal environments of Japan

Materials and methodsStrains

Four clonal strains belonging to clades A–D [26]: O cf.ovata s0662 (clade A), Ostreopsis sp 1 s0716 (clade B),Ostreopsis sp 5 O07421-2 (clade C-2), and Ostreopsis sp 6s0587 (clade D-2), were used for the present experiments.Strain s0662 was isolated from the coast of Tei (33°31010 N,133°45020 E), Kochi Prefecture, Japan Strains s0716 andO07421-2 were isolated from the coast of Otsuki(32°47053 N, 133°42030 E), Kochi Prefecture, Japan Strains0587 was isolated from the coast of Haemida, IriomoteIsland, Okinawa Prefecture, Japan (24°15011 N, 123°510E)

Culture conditionsStock cultures of these strains were maintained in polypro-pylene (PP)-capped test tubes (18 9 120 mm) with a flatbottom containing 10 ml Daigo IMK medium (‘‘Daigo’’ isomitted below; Nihon Pharmaceutical Co., Ltd., Japan)which contained 2.35 mM NaNO3, 50.1 lM NH4Cl,9.86 lM Na2HPO4, 28.7 lM K2HPO4, 12.3 lM Fe–ethy-lenediamine tetraacetic acid (EDTA), 0.871 lM Mn–EDTA, 111 lM Na2EDTA, 0.01 lM CuSO45H2O,0.03 lM Na2MoO42H2O, 0.08 lM ZnSO47H2O, 49.8 nMCoSO47H2O, 0.91 lM MnCl24H2O, 0.0132 lM H2SeO3,0.001 lM vitamin B12, 0.006 lM biotin, and 0.593 lMthiamine–HCl at 25°C under 90–100 lmol photons/m2/s ofcool-white fluorescent illumination on a 12:12 h L:D cycle.Natural sea water with salinity 30.8 collected from Susaki,Kochi Prefecture, Japan was used for the IMK medium andthe other media as below The IMK medium was autoclaved

at 121°C for 20 min

Selection of suitable medium for Ostreopsis spp.

cultures

To select a suitable medium for cultivation of the tested

strains, the present experiments examined four kinds of

media: Provasoli enriched seawater (PES) [35, 36], f/2

(Tris-free and Se-free) [37,38], IMK (Nihon

Pharmaceu-tical Co., Ltd.), and SWM-3 [39,40] Fifteen milliliters of

each medium was added to 25 9 150 mm PP-capped test

tubes with a flat bottom These tubes containing the media

were autoclaved at 121°C for 20 min Stock cultures

growing in IMK medium were inoculated into duplicate

test tubes containing medium Inoculum size was adjusted

giving 1/50–1/100 (v/v) precultures These cultures were

incubated in duplicate at 25°C under 90–100 lmol

pho-tons/m2/s of cool-white fluorescent illumination on a

12:12 h L:D cycle as mentioned above

Cells in a portion of culture medium sampled from the

tube were counted under a microscope (IX-70; Olympus,

Japan) The time course of cell density was expressed as a

growth curve During incubation, in vivo chlorophyll (chl.)

a fluorescence (=autofluorescence) in the tubes was

mea-sured as an indicator of biomass using a Turner fluorometer

(model 10-100 R; Turner Designs, USA) [41] Before

counting and measurement, the culture tube was stirred

using a vortex mixer (Vortex-Genie 2Ò, model G560;

Scientific Industries, USA) at vortexing speed of 6 for

5–10 s to suspend and separate each of cells on the bottom

of the tube into the water column

Specific growth rate (/day) was calculated using data

(n = 3) of cell density from the exponential portion of the

growth curve by least-squares regression of the natural

logarithm of cell density on the number of days [42]

Dividing the specific growth rate by ln 2 value, we

deter-mined the growth rate (divisions/day) [42,43] The highest

value of cell density and of chl a fluorescence obtained

during the incubation were expressed as cell yield and as

biomass yield, respectively The growth rates of Ostreopsis

spp grown in IMK, f/2, PES, and SWM3 media were

averaged, respectively The cell and biomass yield of the

Ostreopsis spp strains in the media were also averaged

Differences among the averaged growth rate, the averaged

cell and biomass yield obtained in various media were

calculated, and the significance of pairwise comparisons

was analyzed by using the replication of a Tukey test for

determination of the suitable media that give significant

high growth rate and/or yield for the tested Ostreopsis spp

strains

Growth potential of Ostreopsis spp strains

Growth rates, cell yields, and biomass yields of each

Os-treopsis strain obtained in the suitable media (f/2 and IMK)

were averaged, respectively Differences among the aged growth rates, cell yields, and biomass yields of fourstrains were calculated, and the significance of pairwisecomparisons was analyzed by using replicate Tukey tests.Based on these results, the growth potential of O cf ovata,Ostreopsis sp 1, Ostreopsis sp 5, and Ostreopsis sp 6isolated from Japanese coastal waters and their differenceswere estimated

aver-Relationship between chlorophyll fluorescence and celldensity

Correlation efficiencies between cell density and chl afluorescence in the selected medium were calculated, andthen a regression function of cell density by chl a fluores-cence was expressed by the least-squares method

of O cf ovata, Ostreopsis sp 1, and Ostreopsis sp 6grown in IMK and f/2 media clearly showed a tendency

to be higher than those in PES and SWM3, althoughthere were no significant differences (p [ 0.05) in thecell and biomass yields among the kinds of medium(data not shown) From these results, IMK and f/2 wereselected as the suitable media for cultivation of thetested strains of Ostreopsis spp in common, althoughPES may also be a suitable medium for Ostreopsis sp 5culture

The averaged growth rates and cell and biomass yields

of each strain obtained in the f/2 and IMK media aresummarized in Table 3 Results of Tukey tests showed thatthe growth rates significantly (p \ 0.05) differed amongthe strains (Table4) The growth rate of Ostreopsis sp 6

(1.04 divisions/day) was the highest (p \ 0.05) among

those of the strains tested (Tables3,4) The growth rate of

O cf ovata (0.834 divisions/day) was significantly higher

(p \ 0.05) than that (0.619 divisions/day) of Ostreopsis

sp 1 (Tables3, 4) and that (0.373 divisions/day) of

Os-treopsis sp 5 which showed the lowest rate (Tables3,4)

Cell yield (3.31 9 104cells/ml) and biomass yield (328

rel fluor.) of O cf ovata were significantly higher than

those of other strains (Tables3,4) In the suitable medium,

the cells of Ostreopsis spp., especially O cf ovata andOstreopsis sp 6, grew not only on the bottom of the tubebut in the surface-intermediate waters as planktonic cellsduring the exponential and stationary phases (data notshown)

Cell densities of the four strains cultivated in IMK andf/2 media significantly correlated (r [ 0.97, p \ 0.001)with their chl a fluorescence and showed significantregressions (Fig.2)

Fig 1 Growth data of

Ostreopsis cf ovata, Ostreopsis

sp 1, Ostreopsis sp 5, and

Ostreopsis sp 6 cultures

growing in each medium Bars

of symbols denote range of

values of the cell density in the

media measured in duplicate

Table 1 Growth rates of four Japanese strains of Ostreopsis spp in IMK, f/2, PES, and SWM3 media

Species (strain) Growth rate (divisions/day)ain each medium

Ostreopsis cf ovata (s0662) 0.802 0.866 0.608 0.181 Ostreopsis sp 1 (s0716) 0.629 0.608 0.384 0.431 Ostreopsis sp 5 (O07421-2) 0.394 0.353 0.364 0.285 Ostreopsis sp 6 (s0587) 1.09 0.918 0.468 0.347

a Averaged value of two determinations

Table 2 Averaged growth rates of four Japanese strains of Ostreopsis spp in IMK, f/2, PES, and SWM3 media with the standard deviation and differences among the averaged growth rates in various kinds of medium

Medium Averaged growth rate

This study is the first report to quantify and characterize

precisely the growth potential of O cf ovata, Ostreopsis

sp 1, Ostreopsis sp 5, and Ostreopsis sp 6 isolated from

Japanese coastal waters [26] using a suitable culture

method

For cultivation of benthic dinoflagellate Ostreopsis spp.,

various kinds of media have been used [28–30,44] Grane´li

et al [30] cultivated an O ovata isolate from the

Tyrrhe-nian Sea by using f/10 medium Guerrini et al [29] and

Pezzolesi et al [32] used f/10 medium for culturing

O ovata/O cf ovata isolates from the Tyrrhenian [29] and

Adriatic Seas [29, 32] Nascimento et al [34] used L-2/2

medium for cultivation of O cf ovata from the Rio de

Janeiro coast Although these isolates from the ranean Sea belong to the Mediterranean Sea and WestAtlantic clade of O cf ovata [21] to which the Japaneseisolate also belongs [26], growth rates and cell yields of theMediterranean and Brazilian strains are lower than those ofthe Japanese O cf ovata strain presented in this study(Table3) Vidyarathna and Grane´li [33] cultivated Japa-nese O cf ovata strain s0662 under the conditions oftemperature 25°C, salinity 31, and light intensity

Mediter-140 lmol photons/m2/s using f/10 medium and found themaximal growth rate (0.7 divisions/day) and cell yield(4 9 103cells/ml) [33] Especially the cell yield is anorder of magnitude lower than our data (3.3 9 104cells/ml), which can presumably be attributed to the differencebetween the kind of medium used by Vidyarathna and

Table 3 Growth rates and cell and biomass yields of various Ostreopsis strains

Species Strain Cladea Country Growth rateb

(divisions/day)

Yieldb References Cell

(9103cells/ml)

Biomass (rel fluor.) Ostreopsis cf ovata s0662 A Japan 0.834 (±0.063) 33.1 (±8.82) 328 (±152) Present study Ostreopsis sp 1 s0716 B Japan 0.619 (±0.034) 8.28 (±2.46) 28.7 (±1.60) Present study Ostreopsis sp 5 O07421-2 C-2 Japan 0.373 (±0.067) 1.24 (±0.450) 32.8 (±19.5) Present study Ostreopsis sp 6 s0587 D-2 Japan 1.04 (±0.125) 8.05 (±2.85) 96.0 (±11.5) Present study Ostreopsis cf ovata s0662 A Japan 0.7 4 [ 33 ] Ostreopsis ovata OOAN0601 A Italy 0.53 10 [ 29 ] Ostreopsis cf ovata OOAN0601 A Italy 0.707 16 [ 32 ] Ostreopsis ovata KAC85 A Italy 1.07 0.6 [ 30 ] Ostreopsis cf ovata ECA-B7 A Brazil 0.14 19.6 [ 34 ] Ostreopsis cf ovata ECA-E7 A Brazil 0.22 10.1 [ 34 ] Ostreopsis lenticularis Puerto Rico 0.97c [ 27 ]

a Sato et al [ 26 ]

b The averaged values of growth rates and cell and biomass yields of each Japanese strain cultivated in f/2 and IMK media are shown; standard deviation (SD) is shown in parenthesis

c Value calculated form the shortest mean generation observed in Tosteson et al [ 27 ]

Table 4 Differences among the averaged growth rates, cell yields, and biomass yields of four Japanese strains of Ostreopsis spp.

Growth parameter Strain Differences among the averaged growth rates, cell yields, and biomass yields of each straina

Growth rate (divisions/day) s0662 0.000

s0716 0.216* 0.000 O07421-2 0.461*** 0.245** 0.000 s0587 0.170* 0.386*** 0.631*** 0.000 Cell yield (9103cells/ml) s0662 0.000

s0716 24.8*** 0.000 O07421-2 31.8*** 7.04 0.000

* p \ 0.05; ** p \ 0.01; *** p \ 0.001

a Significance of the pairwise comparisons was analyzed by using the replication of a Tukey test

Grane´li [33] and that used in present study The former

used f/10 medium [33] whose nutrient contents are much

lower than that of f/2 medium used in latter study

Con-sidering the difference of methods and results between the

studies, we infer that the lack of nutrients in f/10 medium

could limit growth of O cf ovata cultures

Moreover, the growth rates and/or biomasses of O cf

ovata, Ostreopsis sp 1, and Ostreopsis sp 6 reported in

this study were almost equal to those of marine ‘‘red tide’’

flagellates [45, 46] and dinoflagellates [43] that present

high growth rates ([1 divisions/day) and massive

prolif-eration ([100 rel fluol.) in the field and under laboratory

conditions No adverse effect was caused by the stirring on

growth of Ostreopsis spp (data not shown) Based on our

culture conditions, use of IMK or f/2 medium is considered

to be suitable for elucidation of the growth potential of the

Ostreopsis strains

Growth rates and yields of the Ostreopsis spp cultures

in PES and SWM-3 showed lower values than those in

IMK and f/2 media, which may be attributable to growth

inhibition of the Ostreopsis cultures by a component in the

PES and SWM-3 media Both PES and SWM-3 media

contain Tris(hydroxymethyl)aminomethane as a pH buffer

reagent, which is not added into IMK and f/2 media Our

preliminary work also found worse growth of

Gambier-discus spp cultures in PES and SWM-3 media as well as

Ostreopsis cf ovata cultures in f/2 medium containing Tris

than those in Tris-free f/2 and IMK media (Yoshimatsu

et al., unpublished data, 2011) Preliminary we have not

achieved successful passage of Ostreopsis spp strains by

using PES and SWM-3 media (data not shown)

Consid-ering the results, we suggest the possibility that growth of

Japanese strains of Ostreopsis spp such as Ostreopsis sp 1and Ostreopsis sp 6 was affected by the pH buffer reagent,and thus recommend use of Tris-free media for culturingthe genus Ostreopsis

A Turner fluorometer has been applied to evaluation ofgrowth of various planktonic flagellates, dinoflagellates,and diatoms [43,45–47] However, growth of the benthicdinoflagellate Ostreopsis has not been analyzed by usingthe fluorometer partly because it seems to be difficult tosuspend the cells attached to the glassware into the culturemedium Since it was quite easy to suspend the cells ofOstreopsis spp attached to the bottom of test tubes by themethod described in this study, significant correlationsbetween chl fluorescence determined by using a Turnerfluorometer and cell density were found in the Ostreopsisspp isolates, which suggests that the present system usingthe Turner fluorometer is useful for rapid quantification ofgrowth rates and yields of other Ostreopsis cultures.The growth rate ([1.0 divisions/day) of Ostreopsis sp 6was significantly higher than those of O cf ovata, Ostre-opsis sp 1, and Ostreopsis sp 5, which suggests that theformer species may predominate over the latter species inJapanese coastal waters Growth rates of O cf ovata,Ostreopsis sp 1, and Ostreopsis sp 6 clarified in this studywere higher than those (ca.\0.6 divisions/day) of the otherbenthic dinoflagellates, Gambierdiscus toxicus [28,48,49],Prorocentrum lima, Prorocentrum concavum, Prorocen-trum mexicanum, and Coolia monotis [28], which suggeststhat O cf ovata, Ostreopsis sp 1, and Ostreopsis sp 6 arecapable of growing rapidly and would have ecologicaladvantages over other benthic dinoflagellates in field ben-thic environments

Fig 2 Correlations between

the cell density and the

chlorophyll a fluorescence in

the Ostreopsis cultures

A number of cells of O cf ovata and Ostreopsis sp 6 in

the cultures existed not only as benthic cells but also as

planktonic cells The planktonic cells must be the reason

why O cf ovata and Ostreopsis sp 6 cultures showed

large biomass yields ([90 rel fluor.) which were almost

equal to those of ‘‘red tide’’ phytoplankters such as the

flagellate Chattonella ovata [45,46] and the dinoflagellate

Heterocapsa circularisquama [47] Massive proliferation

seems to be a physiological feature of O cf ovata and

Ostreopsis sp 6 This feature would be supported by a

phenomenon that field populations of O ovata developed

into massive blooms in coastal waters in Genoa in 2005

[18] Further research is necessary to elucidate the

mech-anism of massive proliferation of the species of Ostreopsis

Recent work has shown that Ostreopsis sp 1

predomi-nates among Ostreopsis spp in Japanese coastal waters

[26] At the beginning of the present study, we predicted

that Ostreopsis sp 1 might be able to grow faster than

O cf ovata However, the growth rate of Ostreopsis sp 1

was significantly lower than those of O cf ovata and

Ostreopsis sp 6, which suggests that Ostreopsis sp 1

possesses physiological features that are not shared by the

other two species, such as tolerance for wide ranges of

temperature and salinity or a unique nutrient demand,

which play important roles in predomination of the

organism in the coastal environments of Japan

The present laboratory study succeeded in

character-ization of the growth potential of the strains belonging to

O cf ovata, Ostreopsis sp 1 Ostreopsis sp 5, and

Os-treopsis sp 6 collected from Japanese coastal waters by

using a suitable system that consists of a test tube

(25 9 150 mm) containing nutrient-enriched f/2 and IMK

media and a Turner fluorometer for culturing the organisms

and quantifying their growth The growth potential of O cf

ovata and Ostreopsis sp 6 may be of great ecological

advantage to their bloom development in benthic

envi-ronments along the coast of Japan Clarifying the growth

physiology such as temperature–salinity tolerance and

nutrition of these Ostreopsis strains with the present system

will be necessary to understand the mechanisms of bloom

dynamics of the genus Ostreopsis in coastal environments

Acknowledgments This study was supported by a grant-in-aid from

the Food Safety Commission, Japan (no 0904).

References

1 Moore RE, Scheuer PJ (1971) Palytoxin: a new marine toxin

from a coelenterate Science 172:495–498

2 Usami M, Satake M, Ishida S, Inoue A, Kan Y, Yasumoto T

(1995) Palytoxin analogs from the dinoflagellate Ostreopsis

siamensis J Am Chem Soc 177:5389–5390

3 Taniyama S, Mahmud Y, Terada M, Takatani T, Arakawa O,

Noguchi T (2002) Occurrence of a food poisoning incident by

palytoxin from a serranid Epinephelus sp in Japan J Nat Toxin 11:277–282

4 Taniyama S, Arakawa O, Takatani T, Noguchi T (2003) Food poisoning similar to Scarus ovifrons poisoning New Food Ind 45:55–61

5 Taniyama S, Arakawa O, Terada M, Nishio S, Takatani T, Mahmud Y, Noguchi T (2003) Ostreopsis sp., a possible origin of palytoxin (PTX) in parrotfish Scarus ovifrons Toxicon 42:29–33

6 Yasumoto T, Yasumura D, Ohizumi Y, Takahashi M, Alcala AC, Alcala LC (1986) Palytoxin in two species of xanthid crab from the Philippines Agric Biol Chem 50:163–167

7 Shears NT, Ross PM (2009) Blooms of benthic dinoflagellates of the genus Ostreopsis; an increasing and ecologically important phenomenon on temperate reefs in New Zealand and worldwide Harmful Algae 8:916–925

8 Aligizaki K, Katikou P, Nikolaidis G, Panou A (2008) First episode of shellfish contamination by palytoxin-like compounds from Ostreopsis species (Aegean Sea, Greece) Toxicon 51:418–427

9 Wiles JS, Vick JA, Christensen MK (1974) Toxicological uation of palytoxin in several animal species Toxicon 12:427–433

eval-10 Deeds JR, Schwartz MD (2010) Human risk associated with palytoxin exposure Toxicon 56:150–162

11 Ciminiello P, Dell’Aversano C, Fattorusso E, Forino M, glione L, Grillo C, Melchiorre N (2008) Putative palytoxin and its new analogue, ovatoxin-a, in Ostreopsis ovata collected along the Ligurian coasts during the 2006 toxic outbreak J Am Soc Mass Spectrom 19:111–120

Tarta-12 Parsons ML, Preskitt LB (2007) A survey of epiphytic flagellates from the coastal waters of the island of Hawai’i Harmful Algae 6:658–669

dino-13 Sagara T (2008) Profiles of palytoxin-like compounds from the dinoflagellate Ostreopsis sp isolated from the areas where poi- sonous fishes were collected Nippon Suisan Gakkaishi 74:913–914

14 Taniyama S (2008) The occurrence of palytoxin-like poisoning and ciguatera in parts of the mainland of Japan Nippon Suisan Gakkaishi 74:917–918

15 Totti C, Accoroni S, Cerino F, Cucchiari E, Romagnoli T (2010) Ostreopsis ovata bloom along the Conero Riviera (northern Adriatic Sea): relationships with environmental conditions and substrata Harmful Algae 9:233–239

16 Aligizaki K, Katikou P, Milandri A, Dioge`ne J (2011) Occurrence

of palytoxin-group toxins in seafood and future strategies to complement the present state of the art Toxicon 57:390–399

17 Mangialajo L, Ganzin N, Accoroni S, Asnaghi V, Blanfune´ A, Cabrini M, Cattaneo-Vietti R, Chavanon F, Chiantore M, Cohu S, Costa E, Fornasaro D, Grossel H, Marco-Miralles F, Maso´ M, Ren˜e´ A, Rossi AM, Sala MM, Thibaut T, Totti C, Vila M, Leme´e

R (2011) Trends in Ostreopsis proliferation along the Northern Mediterranean coasts Toxicon 57:408–420

18 Ciminiello P, Dell’Aversano C, Fattorusso E, Forino M, Magno

GS, Tartaglione L, Grillo C, Melchiorre N (2006) The Genoa

2005 outbreak Determination of putative palytoxin in ranean Ostreopsis ovata by a new liquid chromatography tandem mass spectrometry method Anal Chem 78:6153–6159

Mediter-19 Fukuyo Y (1981) Taxonomical study on benthic dinoflagellates collected in coral reefs Bull Jpn Soc Sci Fish 47:967–978

20 Faust MA, Morton SL, Quod JP (1996) Further SEM study of marine dinoflagellates: the genus Ostreopsis (Dinophyceae).

J Phycol 32:1053–1065

21 Penna A, Fraga S, Battocchi C, Casabianca S, Giacobbe MG, Riobo´ P, Vernesi C (2010) A phylogeographical study of the toxic benthic dinoflagellate genus Ostreopsis Schmidt J Biogeogr 37:830–841

22 Rhodes L (2011) World-wide occurrence of the toxic

dinofla-gellate genus Ostreopsis Schmidt Toxicon 57:400–407

23 Mangialajo L, Bertolotto R, Cattaneo-Vietti R, Chiantore M,

Grillo C, Lemee R, Melchiorre N, Moretto P, Povero P, Ruggieri

N (2008) The toxic benthic dinoflagellate Ostreopsis ovata:

quantification of proliferation along the coastline of Genoa, Italy.

Mar Pollut Bull 56:1209–1214

24 Faust MA, Morton SL (1995) Morphology and ecology of the

marine dinoflagellate Ostreopsis labens sp nov (Dinophyceae).

J Phycol 31:456–463

25 Nakajima I, Ochima Y, Yasumoto T (1981) Toxicity of benthic

dinoflagellates in Okinawa Bull Jpn Soc Sci Fish 47:1029–1033

26 Sato S, Nishimura T, Uehara K, Sakanari H, Tawong W,

Ha-riganeya N, Smith K, Rhodes L, Yasumoto T, Taira Y, Sudo S,

Yamaguchi H, Adachi M (2011) Phylogeography of Ostreopsis

along west Pacific coast, with special reference to a novel clade

from Japan PLoS One 6:e27983

27 Tosteson TR, Ballantine DL, Tosteson CG, Hensley V, Bardales

AT (1989) Associated bacterial flora, growth, and toxicity of

cultured benthic dinoflagellates Ostreopsis lenticularis and

Gambierdiscus toxicus Appl Environ Microbiol 55:137–141

28 Morton ST, Norris DR, Bomber JW (1992) Effect of temperature,

salinity and light intensity on the growth and seasonality of toxic

dinoflagellates associated with ciguatera J Exp Mar Biol Ecol

157:79–90

29 Guerrini F, Pezzolesi L, Feller A, Riccardi M, Ciminiello P,

Dell’Aversano C, Tartaglione L, Iacovo ED, Fattorusso E, Forino

M, Pistocchi R (2010) Comparative growth and toxin profile of

cultured Ostreopsis ovata from the Tyrrhenian and Adriatic seas.

Toxicon 55:211–220

30 Grane´li E, Vidyarathna NK, Funari E, Cumaranatunga PRT,

Scenati R (2011) Can increases in temperature stimulate blooms

of the toxic benthic dinoflagellate Ostreopsis ovata? Harmful

Algae 10:165–172

31 Pistocchi R, Pezzolesi L, Guerrini F, Vanucci S, Dell’Aversano

C, Fattorusso E (2011) A review on the effects of environmental

conditions on growth and toxin production of Ostreopsis ovata.

Toxicon 57:421–428

32 Pezzolesi L, Guerrini F, Ciminiello P, Dell’Aversano C, Iacovo

ED, Fattorusso E, Forino M, Tartaglione L, Pistocchi R (2012)

Influence of temperature and salinity on Ostreopsis cf ovata

growth and evaluation of toxin content through HR LC-MS and

biological assays Water Res 46:82–92

33 Vidyarathna NK, Grane´li E (2012) Influence of temperature on

growth, toxicity and carbohydrate production of a Japanese

Os-treopsis ovata strain, a toxic-bloom-forming dinoflagellate Aquat

Microb Ecol 65:261–270

34 Nascimento SM, Correˆa EV, Menezes M, Varela D, Paredes J,

Morris S (2012) Growth and toxin profile of Ostreopsis cf ovata

(Dinophyta) from Rio de Janeiro, Brazil Harmful Algae 13:1–9

35 Provasoli L (1968) Media and prospects for the cultivation of

marine algae In: Watanabe A, Hattori A (eds) Cultures and

collections of algae Jpn Soc Plant Physiol, Hakone, pp 63–75

36 Bold HC, Wynne MJ (1978) Introduction to the algae: structure and reproduction, 1st edn Prentice Hall of India, New Delhi,

pp 1–216

37 Guillard RRL, Ryther JH (1962) Studies of marine planktonic diatoms I Cyclotella nana Hustedt, and Detonula confervacea (Cleve) Gran Can J Microbiol 8:229–239

38 Guillard RRL (1975) Culture of phytoplankton for feeding ine invertebrates In: Smith WL, Chanley MH (eds) Culture of marine invertebrate animals Plenum, New York, pp 26–60

mar-39 Itoh K, Imai I (1987) Raphidophyceae In: Japan Fisheries Resources Conservation Association (ed) A guide for studies of red tide organisms Shuwa, Tokyo, pp 122–130

40 Imai I, Itakura S, Matsuyama S, Yamaguchi M (1996) Selenium requirement for growth of a novel red tide flagellate Chattonella verruculosa (Raphidophyceae) in culture Fish Sci 62:834–835

41 Brand LE, Guillard RRL, Murphy LS (1981) A method for the rapid and precise determination of acclimated phytoplankton reproduction rates J Plankton Res 3:193–201

42 Guillard RRL (1979) Methods for microflagellates and plankton In: Stein JR (ed) Handbook of phycological methods: culture methods and growth measurements Cambridge Univer- sity Press, Cambridge, pp 289–311

nanno-43 Yamaguchi M, Honjo T (1989) Effects of temperature, salinity and irradiance on the growth of the noxious red tide flagellate Gymnodinium nagasakiense (Dinophyceae) Nippon Suisan Gakkaishi 55:2029–2036

44 Ashton M, Tosteson T, Tosteson C (2003) The effect of elevated temperature on the toxicity of the laboratory cultured dinofla- gellate Ostreopsis lenticularis (Dinophyceae) Rev Biol Trop 51:1–6

45 Yamaguchi H, Sakamoto S, Yamaguchi M (2008) Nutrition and growth kinetics in nitrogen- and phosphorus-limited cultures of the novel red tide flagellate Chattonella ovata (Raphidophyceae) Harmful Algae 7:26–32

46 Yamaguchi H, Mizushima K, Sakamoto S, Yamaguchi M (2010) Effects of temperature, salinity and irradiance on growth of the novel red tide flagellate Chattonella ovata (Raphidophyceae) Harmful Algae 9:398–401

47 Yamaguchi H, Sakou H, Fukami K, Adachi M, Yamaguchi M, Nishijima T (2005) Utilization of organic phosphorus and pro- duction of alkaline phosphatase by the marine phytoplankton, Heterocapsa circularisquama, Fibrocapsa japonica and Chae- toceros ceratosporum Plankton Biol Ecol 52:67–75

48 Bomber JW, Guillard RRL, Nelson WG (1988) Roˆles of perature, salinity, and light in seasonality, growth, and toxicity of ciguatera-causing Gambierdiscus toxicus Adachi et Fukuyo (Dinophyceae) J Exp Mar Biol Ecol 115:53–65

tem-49 Lartigue J, Jester ELE, Dickey RW, Villareal TA (2009) Nitrogen source effects on the growth and toxicity of two strains of the ciguatera-causing dinoflagellate Gambierdiscus toxicus Harmful Algae 8:781–791

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

Daily growth rate model of Japanese anchovy larvae Engraulis

japonicus in Hiuchi-nada Sea, central Seto Inland Sea

Hiromu Zenitani•Naoaki Kono

Received: 22 December 2011 / Accepted: 28 June 2012 / Published online: 31 July 2012

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

Abstract In the present study, we developed a larval

anchovy growth model in relation to sea temperature and

food availability via food consumption and metabolic

process terms, based on biological data from previous

laboratory experiments and field surveys from 2003 to

2006 in Hiuchi-nada Sea, central part of Seto Inland Sea,

Japan To investigate when food shortage for larval

anchovy and then recruitment failure occur in Hiuchi-nada

Sea, anchovy food requirements were estimated by using

the growth model, and we compared the food requirement

with anchovy food availability We applied an estimation

method for growth model parameters, Hewett–Johnson

p and Q10, by minimizing the sum of squares of difference

between mass-specific growth rates estimated by the

models and those by otolith growth analysis Parameter

p was 0.86, slightly higher than typical values, and Q10was

2.11, close to the value used for the biological model of

larval northern anchovy Food shortage for anchovy larvae

did not occur in Hiuchi-nada Sea, although it was indicated

that low food availability led to a low reproductive success

rate The newly developed growth model is considered

optimal at present and useful to link environmental

con-ditions and larval growth

Keywords Anchovy larvae Growth model 

Hiuchi-nada Sea Otolith growth analysis  Seto Inland Sea

IntroductionThe Seto Inland Sea is well known for its high fishproduction [1] Japanese anchovy Engraulis japonicus is

an important commercial species, accounting for 34 % ofthe total fish production of 67 9 103tons in the SetoInland Sea in 2009 Hiuchi-nada Sea is located in thecentral part of the Seto Inland Sea between KurushimaStrait and Bisan Strait, a semi-enclosed narrow sea(Fig.1) It has a size of about 50 9 30 km2 and anaverage depth of about 20 m, being a major spawning andfishing ground of Japanese anchovy Total catch ofanchovy in Hiuchi-nada Sea was 15 9 103tons in 2005,accounting for 27 % of the total anchovy catch in theentire Seto Inland Sea In this area, forecasting the degree

of recruitment populations of shirasu (larval anchovy inJapanese, body length: ca 20–35 mm) to the fishery stock

is required for efficient exploitation [2] Recruitmentforecasting however has generally not been successful [3],because it is difficult to forecast the survival rate (ormortality rate) from the prerecruitment stage to therecruitment stage

In recent years, individual-based models have beendeveloped to understand how physical or biological factorsaffect larval anchovy survival or recruitment [4 6] Thesemodels require algorithms for the growth process InZenitani et al [2], the authors already found that thegrowth rate for larval anchovy was dependent on temper-ature, food availability, and larval size However, a detailedgrowth model to describe the relation to sea temperatureand food availability has not been developed for larvalanchovy in the Seto Inland Sea The aims of the presentstudy are to (1) develop a larval anchovy growth model inrelation to sea temperature and food availability via foodconsumption and metabolic process terms in Hiuchi-nada

H Zenitani ( &)  N Kono

National Research Institute of Fisheries and Environment

of Inland Sea, Fisheries Research Agency, Hatsukaichi,

Hiroshima 739-0452, Japan

e-mail: zenitani@affrc.go.jp

DOI 10.1007/s12562-012-0532-2

Sea and (2) investigate if a food shortage for larval

anchovy occurs and detect any recruitment failure of

year-class in this study area We developed a growth model for

larval anchovy, based mainly on information from

Japa-nese anchovy and northern anchovy E mordax In the past,

anchovy were popular experimental organisms, cultured to

understand their growth [7,8], physiology [9], capacity to

survive starvation [10], and swimming behavior [2, 11]

Yokota et al [12] and Uotani [13] studied feeding behavior

in the Seto Inland Sea or waters along the Pacific coast of

Japan These studies provide a rich source of information

which can be used to formulate and calibrate the growth

model This manuscript presents a method for estimating

parameters of the growth model by linking otolith growth

analysis, laboratory experiments, and field survey data in

Hiuchi-nada Sea

A possible key factor in the regulation of anchovy

population levels is the fluctuations in abundance of the

copepod assemblage, and the crucial period for

recruit-ment of anchovy in Hiuchi-nada Sea would be the period

just before the anchovy recruitment to the shirasu fishery

[14] To investigate when food shortage for larval

anchovy and recruitment failure occur in Hiuchi-nada

Sea, anchovy food requirements were estimated by

using the growth model, and we compared the foodrequirement with anchovy food availability in Hiuchi-nada Sea

Materials and methodsOtolith growth

The mass-specific growth rate of an individual anchovy

l was determined by otolith growth analysis, as follows [2]:Anchovy were sampled from commercial catches (seinenet fishery) at Kanonji Port in Hiuchi-nada Sea (Fig.1) on 18July 2003, 6 July and 6 August 2004, 27 June 2005, and 24June 2006 A subsample of 50–100 anchovy larvae andjuveniles was taken randomly from each commercial catchand preserved in 99 % ethanol Each fish was measured forstandard length (SL) to the nearest 0.1 mm with digitalcalipers Sagittal otoliths of the larvae and juveniles(30.2–44.6 mm SL in 2003, 37.4–64.5 mm SL in 2004,17.7–33.0 mm SL in 2005, 23.7–32.0 mm SL in 2006) weredissected out, cleaned under a binocular dissecting micro-scope, and mounted on a glass slide with epoxy resin.Measurement of otolith daily increments was conducted

Hiuchi–nada Sea

44˚N

34˚N 39˚N

34º15’N

34º00’N

Port

Kanonji-Kurushima Strait

Bisan Strait

143˚E 133˚E

34˚N

33˚ N

Pacific Ocean

Sea of Japan

Seto Inland Sea

Fig 1 Survey stations of

plankton sampling in

Hiuchi-nada Sea, Seto Inland Sea,

Japan

along core–posterior margin axis using the otolith

measurement system (Ratoc System Engineering Inc.,

Tokyo, Japan) under a light microscope at 100–2009

magnification

When the relation between otolith radius and fish size is

predictable, growth rates can be calculated based on otolith

increment widths [2, 15–18] According to previous

experimental studies on Japanese anchovy, the daily

growth increment is deposited at the start of external

feeding, 3–4 days after hatching [8], and SL at completion

of yolk absorption is 5.6 mm [19] Daily age of each

anchovy was calculated, therefore, as the number of rings

plus 3 SL at each daily age was back-calculated by the

biological intercept method [20, 21], with SL at the first

deposition of the daily growth increment fixed at 5.6 mm

as the biological intercept at the individual level We

assumed that the relationship of otolith radius OR(l, m)

(lm) and SL [=La(l, m) (mm)] of individual anchovy l of

the mth ring formation can be expressed by an allometric

formula for the individual Hence, La(l, m) of each anchovy

was back-calculated by substituting OR(l, m) (lm) for La(l,

m) using the allometric relationship,

Laðl; mÞ ¼ aðlÞ  ORðl; mÞbðlÞ; ð1Þ

where a(l) and b(l) are parameters determined for each

anchovy by solving the two equations

Laðl; 1Þ ¼ aðlÞ  ORðl; 1ÞbðlÞ ð2Þ

and

Laðl; mmaxðlÞÞ ¼ aðlÞ  ORðl; mmaxðlÞÞbðlÞ; ð3Þ

where mmax(l) is the maximum number of rings of

indi-vidual anchovy l, and La(l, 1) of 5.6 mm, OR(l, 1), La(l,

mmax(l)), and OR(l, mmax(l)) are the SL at the first daily

growth increment deposition, measured radius of the first

daily ring, SL at sampling day ds(l), and measured radius at

sampling day ds(l), respectively

The relationship between body mass, Wa(l, m) (mgC),

and SL of Japanese anchovy larvae [22] has been found as

Waðl; mÞ ¼ 2:045  104 Laðl; mÞ3:385 0:43; ð4Þ

where 0.43 is a multiplier to convert to carbon weight from

dry weight [23]

The mass-specific growth rate Gðl; mði; lÞÞ (day^ -1)

for individual anchovy l at m(i, l)th ring formation

was estimated from the change in the back-calculated

weight for individual anchovy l in plankton survey cruise

3 days The metamorphosing stage of Japanese anchovy,defined as the stage of guanine deposition on the peritonealand trunk surface, occurs from 31 to 37 mm SL, while thejuvenile stage occurs from ca 40 mm SL [24] Based onthis study of anchovy development stages, we supposed alarval-stage SL range of 5.6–31.0 mm, and then estimatedthe growth rate in this range

Plankton surveyFrom spring to early summer, in 2003–2006, a total of 14daytime cruises of R.V Shirafuji maru, National ResearchInstitute of Fisheries and Environment of Inland Sea, wereconducted to sample plankton at 13 stations over 4-day surveyperiods, as follows (Fig.1; Table1, Zenitani et al [2]):

At each station, a hydrographic cast was made using aNiel Brown CTD (Cataumet, MA, USA) in 2003 and acompact CTD (ACL 208-PDK Alec Electronics, Kobe,Japan) in 2004–2006 Water samples for determiningcopepod nauplii density were collected with a 20-l VanDorn water sampler (Rigo, Tokyo, Japan) from 10 mdepth Since previous studies in Seto Inland Sea [25,26]have demonstrated that larval anchovies are distributedmainly in the 5–20 m depth strata, predominantly at 10 mdepth, we selected 10 m as a representative depth.For copepod nauplii abundance, 1 l of the water samplewas filtered aboard with a cellulose nitrate membrane filter(pore size 5.0 lm; Toyo Roshi Kaisha, Tokyo, Japan), andcopepod nauplii on the filters were resuspended and pre-served in 5–10 % formalin in a 10-ml test tube Afternauplii were stained by adding 0.1–0.2 ml Rose Bengal,they were enumerated on a counting plate Body lengthwas measured for all the nauplii

Copepods (copepodites and adults) were collected with

a plankton net (mouth diameter 0.45 m, length 1.95 m,mesh opening 100 lm, fitted with a Rigosha flowmeter;Rigo), which was pulled vertically from 2 m above the seabottom to the surface After the net was retrieved, thezooplankton were immediately preserved in 5–10 % for-malin-sea water solution Later the sample was split into 1/32–1/256 subsamples, from which copepods were enu-merated under a stereoscopic microscope To estimate thelength–frequency distribution from cruise i, up to 100copepods for each genus were randomly sorted from eachsample and their prosome length was measured to thenearest 0.1 mm

We classified the copepods into three stages

We calculated the mean body length Lc(i, 1) (mm) for a

copepod nauplii from cruise i, and then the carbon content

Wc(i, 1) (mgC) was calculated from the mean body length

as [27]

Wcði; 1Þ ¼ 1:51  1014 ðLcði; 1Þ  103Þ2:94: ð8Þ

Prosome length Lc(i, j, k) (mm) of taxonomic group k of

stage j (=2, 3) of copepod from cruise i was converted to

body carbon weight using the length–carbon weight

regression equation [23,28,29]

Wcði; j; kÞ ¼ 10cðkÞ 103 ðLcði; j; kÞ  103ÞdðkÞ: ð9Þ

Wc(i, j, k) (mgC) is the carbon weight of taxonomic

group k of stage j of copepod from cruise i The taxonomic

groups were k = 1, Acartia; k = 2, Calanus; k = 3,

Centropages; k = 4, Microsetella; k = 5, Oithona;

k = 6, Paracalanus; k = 7, others (Corycaeus, Oncaea,

etc.) c(k) and d(k) are parameters of the taxonomic group k

(Tables2; Zenitani et al [2])

We set the mean length of stage 2 copepods, Lc(i, 2, k),

at 0.37 mm, and calculated the mean length Lc(i, 3) (mm)

in stage 3 copepods and mean weight Wc(i, j) (mgC) instage 2 and 3 copepods by the following equations:

Lcði; 3Þ ¼

P7 k¼1ðLcði; j; kÞ  Acði; j; kÞÞ

P7

Wcði; jÞ ¼

P7 k¼1ðWcði; j; kÞ  Acði; j; kÞÞ

P7 k¼1Acði; j; kÞ ðj ¼ 2; 3Þ;

ð11Þwhere Ac(i, j, k) (103individuals m-3) is the mean abun-dance of taxonomic group k of stage j of copepod fromcruise i

ConsumptionThe mass-specific food consumption rate was determined

by an individual-based model for each individual anchovy

of given size at a temperature and abundance of prey asfollows (Fig.2; Zenitani et al [2]):

Copepods are the main prey items for anchovy larvae

At the first feeding stage, larval anchovy eat mainly

Table 1 Mean temperature, copepod nauplii abundance, small-size (0.25–0.50 mm prosome length) copepod abundance, and large-size (over 0.50 mm prosome length) copepod abundance in Hiuchi-nada Sea in 2003–2006

Cruise Date Mean temperature

(±SD) (°C)

Copepod nauplii Small-size copepods Large-size copepods mean concentration

(±SD) (103individuals m-3)

mean concentration (±SD)

(103individuals m-3)

mean concentration (±SD)

1 nauplius stageð\0:25 mm in body length)

2 small-size stageð0:250:50 mm in prosome length, mode 0:37 mm)

3 large-size stageð[ 0:50 mm in prosome length)

8

<

naupliar copepods, then with increasing size, the larvae

eat copepodites and adult copepods [12, 30] The food

availability Cc(l, m(i, l)) (mgC day-1) of individual

anchovy l on copepods at m(i, l)th ring formation was

where Em(i, j, l) is the number of encounters between stage

j copepods and individual anchovy l that resulted in

successful capture of stage j copepods during cruise i The

number of encounters E(i, j, l) (day-1) between stage j

copepods and individual anchovy l in cruise i was based on

the Gerritsen and Strickler [31] model for randomly

moving organisms in a three-dimensional space, such that

where DEis the proportion of daylight hours in a day (12/

24 h), CE is a conversion factor (8.64 9 10-4s day-1),

Ac(i, j) (individuals m-3) is the abundance of stage j

copepods from cruise i,

Acði; jÞ ¼X7

k¼1

and RT(i, j, l) (mm) is the total encounter radius Va(i, l)

(mm s-1) and Vc(j) (mm s-1) are the swimming speeds of

individual anchovy l in cruise i and stage j copepods,

respectively The total encounter radius was assumed to be

equal to the sum of the mean encounter radius Rc(i, j) (mm)

of stage j copepods and the mean encounter radius Ra(i, l)

(mm) of individual anchovy l in cruise i, as Bailey andBatty [32] estimated,

RTði; j; lÞ ¼ Rcði; jÞ þ Raði; lÞ; ð15Þ

in female adult stage, and 9 mm s-1 in male adult stage[33] According to Yen [34], Euchaeta rimana has a typicalswimming speed of roughly 7 mm s-1 We assumed thatthe swimming speed in the nauplii, small copepods, andlarge copepods was 1.5, 1.5, and 7 mm s-1, respectively.The number of encounters between stage j copepods andindividual anchovy l in cruise i, in which individual anchovy

l successfully captured stage j copepods, was determinedfrom a binomial distribution BNðEði; j; lÞ; Pði; j; lÞÞ; and themean number of encounters Em(i, j, l) (day-1) is

Emði; j; lÞ ¼ Eði; j; lÞ  Pði; j; lÞ; ð19Þwhere P(i, j, l) is the capture success probability ofindividual anchovy l for stage j copepods in cruise i

Pði; j; lÞ ¼ 1 for Lminði; lÞ  Lcði; jÞ  Lmaxði; lÞ

0 for Lcði; jÞ\Lminði; lÞ or Lmaxði; lÞ \ Lcði; jÞ:(

ð20ÞThe minimum and maximum length of copepods thatanchovy larvae were able to capture, Lmin(i, l) (mm) and

Lmax(i, l) (mm), were

Lminði; lÞ ¼ 0:15 for 5:6 mm Laðl; mði; lÞÞ\9:0 mm

0:20 for 9:0 mm Laðl; mði; lÞÞ(

ð21Þand

Lmaxði; lÞ ¼ 0:434  exp 0:105  Lð aðl; mði; lÞÞÞ; ð22Þafter Fig 5-1 in Yokota et al [12] and Fig 6 in Uotani [13].The mass-specific food consumption rate C(l, m(i, l))(day-1) of individual anchovy l on copepods at m(i, l)thring formation is defined as a function of feeding rate forcopepods and a proportionality constant parameter, Hew-ett–Johnson p value [35], having values of zero to one

Table 2 Parameters of regression equation of carbon weight against

length for copepods collected in the Seto Inland Sea

Cðl; mði; lÞÞ ¼ p  Ccðl; mði; lÞÞ  Waðl; mði; lÞÞ1: ð23Þ

Metabolic

In the growth model of larval northern anchovy, the

met-abolic rate, assumed as a function of an organism’s body

mass and temperature, is

where R (lg day-1) is the metabolic rate, Wd (lg) is dry

body weight, Q10is defined as the increase in the rate of a

physiological process resulting from a 10°C increase in

temperature, and T (°C) is temperature [36] We assumed a

simple and standard formulation for the mass-specific

metabolic rate U(l, m(i, l)) (day-1) of individual anchovy l

at m(i, l)th ring formation of

Uðl; mði; lÞÞ ¼ 0:187  Waðl; mði; lÞÞ  10

where 103/0.43 is a multiplier to convert to carbon weight

from dry weight Urtizberea et al [36] used the same Q10

(=2.2) as in bay anchovy Anchoa mitchilli larvae [37]

We estimated Q10 by the data fitting method (see ‘‘

Estimation of parametersThe mass-specific growth rate of an individual for nonre-producing fish is predicted as the weight increment perunit weight per time and defined by the following equation[38]

where the growth G over a time period is the differencebetween energy gained through food consumption C andthe sum of energy costs and losses through metabolism M,egestion F, and excretion E Metabolic costs arerepresented by standard metabolism U and the cost ofdynamic action SDA

Table 3 Back-calculated mass-specific growth rate and food availability of anchovy in Hiuchi-nada Sea in 2003–2006

Cruise Date Sampling

or calculation

back-Mass-specific growth rate as weight increment per unit weight (±SD) (day-1)

Mass-specific food availability per unit weight (±SD) (day-1)

calculated

Back-SL (±SD)

Ratio of food availability to minimum food requirement, k