Fisheries science JSFS , tập 76, số 2, 2010 3

Juvenile and adult mackerel were sampled along with their young anchovy prey field in 2004 juvenile mackerel and larval anchovy and 2005 adult mackerel and juvenile anchovy off the Pacif

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

Identification of facultative anaerobic bacteria isolated

from the intestine of the minke whale Balaenoptera acutorostrata

by 16S rRNA sequencing analysis

Go Ogawa•Masami Ishida•Hidehiro Kato •

Yoshihiro Fujise•Naoto Urano

Received: 23 July 2009 / Accepted: 11 November 2009 / Published online: 22 January 2010

Ó The Japanese Society of Fisheries Science 2010

Abstract There are few reports in the literature about the

isolation of bacteria from whale intestine In this report, we

counted colony-forming units in the feces obtained from

three female common minke whales (Balaenoptera

acut-orostrata) The number of colony-forming units ranged

from (2.2 ± 0.4) 9 105 to (8.9 ± 2.0) 9 108 per gram

(wet weight) of excrement 16S rRNA gene sequences of

141 isolates were determined These strains were identified

as Enterococcus faecalis, Enterococcus sp., Enterobacter

cloacae, Enterobacter sp., Escherichia coli, Edwardsiella

ictaluri or Clostridium sp The data suggested that the

facultative anaerobic population of the intestinal bacterial

flora of the minke whale was similar to that of ground

mammals

Keywords Balaenoptera acutorostrata

Edwardsiella ictaluri
Enterobacteriaceae

Intestinal flora
Minke whale

Introduction

There have been few reports about the isolation and

iden-tification of bacteria from whale intestine Only the

Bru-cella strains have been reportedly isolated from whale

skins [1,2] In recent years, 16S rRNA sequence analysis

has become the standard method for identification or

classification of bacteria However, commercial whalingwas prohibited after a decision by the International Whal-ing Commission in 1982 [3], before Woese et al [4]reported their rRNA classification method Therefore, thesampling of whale excrement is a valuable way of studyingthe microbiology of whales Bacterial infection is one ofthe most frequent causes of death of marine animals inJapan [5] Additionally, whale meat contamination andfood poisoning have been reported [6 8] Studying theintestinal flora of the whale is also important for under-standing sickness in whales and potential pathogens thatcould be passed on to humans

We obtained excrement from wild common minkewhales Balaenoptera acutorostrata during research whal-ing carried out in 2008 In this report, we take the first step

in the study of the whale intestinal flora, by counting ony-forming units in the whale excrement and identifyingfacultatively anaerobic bacteria

col-Materials and methodsMinke whale intestines that contain excrement wereobtained from the Japanese Whale Research Programunder special permit in Japan’s Whale Research Program

in the western North Pacific (JARPN II); the details areshown in Table1 The samples were transported within aday, on ice, to our laboratory in a sterilized plastic bag Theexcrement from the intestine was suspended in saline andspread onto each type of agar plate Media, containing 2%agar, were prepared for each of the following: nutrientbroth (Eiken Chemical, Tochigi, Japan), GAM broth(Nissui Pharmaceutical, Tokyo, Japan), MRS (Merck,Darmstadt, Germany) and desoxycholate (Wako PureChemical industries, Osaka, Japan) Agar plates were

Tokyo University of Marine Science and Technology, Konan,

Minato-ku, Tokyo 108-8477, Japan

incubated at 35°C for a week The GAM broth plates were

incubated anaerobically with Aneropack (Mitsubishi Gas

Chemical, Tokyo, Japan) After incubation, bacterial

col-onies were counted and single colcol-onies were picked at

random The cells isolated were stored at 10°C and

identified

The 16S rRNA sequence analysis was carried out

according to our previous study [9] The BLASTN program

[10] was then used to determine homology with other

organisms The 16S rRNA sequence data obtained in this

study (141 sequences) were registered in the DNA Data

Bank of Japan (DDBJ) (accession no AB496751–

AB496891)

Results and discussion

The viable bacterial counts of each sample are listed in

Table2 Bacterial counts ranged from 1.3 9 105 to

8.9 9 108cfu/g These results corresponded to previous

reports looking at the number of facultative anaerobic

bacteria in human and fish intestines [11–13] Therefore, we

propose that the number of bacteria in the whale intestine is

similar to that of others animals However, a higher viable

bacterial count was reported in human feces under

completely anaerobic conditions [14] A higher bacterialcount may be confirmed if the feces are handled underanaerobic conditions Additionally, the viable bacterialcount of feces was similar to those previously reported inwhale stomach [15, 16] We predict that about 105–108(cells/g or ml) viable bacteria live in the digestive tract ofwhales The viable bacterial count in whale 08NPCS-M020was lower than that found in the other whales This may just

be an individual difference among the whales High bers of bacteria on the GAM agar suggested that this agar(the composition of nutrients and/or the anaerobic incuba-tion conditions) is more suitable for the isolation of bacteriafrom whale feces compared with other media

num-Identification of bacterial strains isolated from whale08NPCS-M020 are shown in Table3 Twenty-two strainswere isolated and identified as Enterococcus faecalis orEnterococcus sp from the nutrient agar and GAM agar.These media are used for the growth of a wide range ofspecies, indicating that only a few species of facultativelyanaerobic bacteria may inhabit the whale intestine E.faecalis is often isolated from the feces of animals and isknown to be potentially pathogenic [17] Therefore, whaleexcrement may be a potential human pathogen

The 71 strains isolated and identified from whale08NPCS-M038 are shown in Table 4 These included

Each value is mean ± SEM (n = 3)

Enterobacter cloacae, Enterobacter sp., Escherichia coli,

E faecalis and Edwardsiella ictaluri E coli and

Entero-bacter sp are known to inhabit the animal intestine [17]

However, there have been few reports on the occurrence of

E ictaluri in mammal excrement E ictaluri has generally

been isolated from fish or freshwater environments [18–

20] Isolation of E ictaluri from whale excrement is a

novel finding

The 48 strains isolated and identified from whale

08NPCS-M050 are shown in Table5 These included E

faecalis, E coli and Clostridium sp Clostridium sp are

anaerobic endospore-forming bacteria [17], and we

sup-pose that a clostridial spore was in the specimen and

ger-minated in anaerobic culture Since several species of

Clostridium are human pathogens [17], whale feces sibly contain pathogenic Clostridium spores

pos-Among the three whale samples, different kinds ofbacterial genera were found Only Enterococcus specieswere isolated from 08NPCS-M020, whereas species fromthree different genera, Enterobacter, Enterococcus andEscherichia, were isolated from 08NPCS-038 Therefore,

we suggest that individual differences occur in the tinal bacterial flora of whales, at least with respect to fac-ultatively anaerobic bacteria In contrast, there was nodifference in the species level isolated from excrementtaken from different regions of the large intestine.Among the facultative anaerobic bacteria isolated in thisstudy, most species, except E ictaluri, are representative of

Similarity (%)

DDBJ accession numbers of these strains are AB496751–AB496891 (22 sequences)

G GAM, D desoxycholate, N nutrient broth, M MRS

Similarity (%)

DDBJ accession numbers of these strains are AB496773–AB496843 (71 sequences)

G GAM, D desoxycholate, N nutrient broth, M MRS

the intestinal bacteria of ground mammals Therefore, we

supposed that the intestinal flora of whales would be

sim-ilar to that of many ground mammals However, the

microbial diversity in the intestine of the whale was lower

than that of other mammals [21] Such low diversity seems

to be related to the predatory behavior of minke whales

Ley et al [22] point out that the diversity of intestinal flora

of predatory animals is lower than that of herbivorous or

omnivorous animals Additionally, there was a difference

between the bacterial species in the feces in this study and

those found in the stomach of whales [16] In this study,

Enterobacteriaceae were mainly isolated from the faces

The bile acid of whale’s liver [23] may affect the bacterial

flora of the stomach by acting as a selective pressure

Saltwater fish have intestinal bacterial flora composed of

Vibrio/Photobacterium, Micrococcus, Corynebacterium

[24] or Enterobacteriaceae, Aeromonas, Pseudomonas [25]

or mainly Vibrio [26] Enterobacteriaceae was a common

species between whales (E coli and Enterobacter) and salt

fishes There was no other common species This result

suggests that whales have a particular intestinal bacterial

flora in seawater environments [27,28]

E ictaluri isolated from the whale intestine in this study

is a characteristic bacterium of fish [19, 20] Sakata

reported that fish intestinal flora was relatively simple

compared with that of ground mammals [29] These facts

suggest that the whale intestinal flora have characteristic

properties of not only ground mammals, but also of fish

The 16S rRNA clone library method [14, 30] or

PCR-DGGE method [31] may be useful for detailed analysis

the Japanese Whale Research Program, who operated under special

permit in Japan’s Whale Research Program in the western North

Pacific (JARPN II), for preparing whale feces samples and gathering data about whales.

References

1 Clavareau C, Wellemans V, Walravens K, Tryland M, Verger

JM, Grayon M, Cloeckaert A, Letesson JJ, Godfroid J (1998) Phenotypic and molecular characterization of a Brucella strain isolated from a minke whale Balaenoptera acutorostrata Microbiology 144:3267–3273

2 Tryland M, Kleivane L, Alfredsson A, Kjeld M, Arnason A, Stuen

S, Godfroid J (1999) Evidence of Brucella infection in marine mammals in the North Atlantic Ocean Vet Rec 144:588–592

3 Fujise Y (2006) Background and an outline of the Japanese whale programs under special permit in the Antarctic and North Pacific (in Japanese) In: Kato H, Ohsumi S (eds) A guide to cetacean population biology Seibutsu Kenkyusha, Tokyo, p 30

4 Woese CR, Kandler O, Wheelis ML (1990) Towards a natural system of organisms: proposal for the domains Archaea, Bacteria, and Eucarya Proc Natl Acad Sci USA 87:4576–4579

5 Enoshima Marine (2005) The investigation report on infectious disease of captive cetaceans in Japan (in Japanese with English abstract) J Jpn Assoc Zool Gard Aquariums 46:6–20

6 Aoki K, Hidema T, Kobayashi K, Matushima K (1984) Food poisoning induced by Welch’s bachillus, Clostridium perfringens, Hoobs type 3 (in Japanese) Annu Rep Fukui Prefect Inst Public Health 22:1–4

7 Toyokawa Y, Otomo Y (1987) Two cases of food poisoning caused by Welch’s bacillus and Salmonella enteritidis occurred in Aomori Prefecture in 1986 fiscal year (in Japanese) Annu Rep Aomori Prefect Inst Public Health Environ 24:35–38

8 Omori S, Kobayashi T, Suzuki K, Ono J, Yoshida Y, Yokota H, Shimizu Y, Okada T, Takasugi N (1989) An outbreak of food poisoning caused by Salmonella due to whale meat (in Japanese) Annu Rep Sapporo City Inst Public Health 16:147–150

9 Ogawa G, Ishida M, Urano N (2009) Isolation and identification

of dibutyl phthalate-degrading bacteria from hydrospheres in Tokyo J Gen Appl Microbiol 55:261–265

10 Altschul SF, Thomas LM, Alejandro AS, Jinghui Z, Zheng Z, Webb M, David JL (1997) Gapped BLAST and PSI-BLAST: a

Similarity (%)

DDBJ accession numbers of these strains are AB496844–AB496891 (48 sequences)

G GAM, D desoxycholate, N nutrient broth, M MRS

new generation of protein database search programs Nucleic

Acids Res 25:3389–3402

11 Gorbach SL, Nahas L, Lerner PI, Weinstein L (1967) Studies of

intestinal microflora Effects of diet, age, and periodic sampling

on numbers of fecal microorganisms in man Gastroenterology

53:845–855

12 Mata LJ, Carrilo C, Villatoro E (1969) Fecal microflora in

healthy persons in a preindustrial region Appl Microbiol 17:596–

602

13 Sugita H, Miyajima C, Kobiki Y, Deguchi Y (1989) Daily

changes of the fecal flora of tilapia Sarotherodon niloticus.

Nippon Suisan Gakkaishi 55:1865

14 Hayashi H, Sakamoto M, Benno Y (2002) Phylogenetic analysis

of the human gut microbiota using 16S rDNA clone libraries and

strictly anaerobic culture-based methods Microbiol Immunol

46:535–548

15 Herwig RP, Staley JT (1984) Baleen whales: preliminary

evi-dence for forestomach microbial fermentation Appl Environ

Microbiol 47:421–423

16 Olsen MA, Aagnes TH, Mathiesen SD (1994) Digestion of

her-ring by indigenous bacteria in the minke whale forestomach.

Appl Environ Microbiol 60:4445–4455

17 Holt JG, Krieg NR, Sneath PHA (eds) (1984) Bergey’s manual of

systematic bacteriology Williams & Wilkins, London, pp 423,

467, 489, 1064, 1141

18 Sakazaki R, Tamura K (2002) Enterobacteria under volume (in

Japanese) Kindai Shuppan, Tokyo, p 7

19 Yuasa K, Kholidin B, Panigoro N, Hatai K (2003) First isolation

of Edwardsiella ictaluri from cultured striped catfish Pangasius

hypophthalmus in Indonesia Fish Pathol 38:181–183

20 Nagai T, Iwamoto E, Sakai T, Arima T, Tensha K, Iida Y, Iida T,

Nakai T (2008) Characterization of Edwardsiella ictaluri isolated

from wild ayu Plecoglossus altivelis in Japan Fish Pathol

43:158–163

21 Mitsuoka T (2001) Methods for isolation of bacteria from

intes-tine (in Japanese) In: Yamasato K, Kodama T, Udagawa S,

Morichi T (eds) Methods for isolation of bacteria NTS, Tokyo,

pp 259–274

22 Ley RE, Hamady M, Lozupone C, Turnbaugh PJ, Ramey RR, Bircher JS, Schlegel ML, Tucker TA, Schrenzel MD, Knight R, Gordon JI (2008) Evolution of mammals and their gut microbes Science 320:1647–1651

23 Nishiwaki M (1954) Body of whales (in Japanese) Douwa Shunju Press, Tokyo, p 139

24 Ishimaru K, Yamamoto S, Murata O, Kumai H (2000) Intestinal bacterial flora of farmed Japanese parrotfish, Oplegnathus fasci- atus at larval and juvenile stages (in Japanese with English abstract) Bull Fish Lab Kinki Univ 7:35–42

Skrupskelis K (2008) Phylogenetic analysis of intestinal bacteria

of freshwater salmon Salmo salar and sea trout Salmo trutta trutta and diet Fish Sci 74:1307–1314

26 Ward NL, Steven B, Penn K, Methe BA, Detrich WH III (2009) Characterization of the intestinal microbiota of two Antarctic notothenioid fish species Extremophiles 13:679–685

27 Mcfell-Ngai M (2006) Love the one you’re with: vertebrate guts shape their microbiota Cell 127:247–249

28 Rawls JF, Mahowald MA, Ley RE, Gordon JI (2006) Reciprocal gut microbiota transplants from zebrafish and mice to germ-free recipients reveal host habitat selection Cell 127:423–433

29 Sakata T (2001) Methods for isolation of bacteria from fish (in Japanese) In: Yamasato K, Kodama T, Udagawa S, Morichi T (eds) Methods for isolation of bacteria (in Japanese) NTS, Tokyo, pp 274–285

30 Hayashi H, Sakamoto M, Kitahara M, Benno Y (2003) Molecular analysis of fecal microbiota in elderly individuals using 16S rDNA library and T-RFLP Microbiol Immunol 47:557–570

31 Yang JL, Cheng AC, Wang MS, Pan KC, Luo QH, Zhu DK, Chen XY, Qi XF (2009) New strategies for electrophoresis analysis of enterobacterial repetitive intergenic consensus PCR in animal intestinal microflora J Microbiol Methods 77:63–66

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

Predation dynamics of mackerel on larval and juvenile anchovy:

is capture success linked to prey condition?

Dominique Robert•Akinori Takasuka•Sayaka Nakatsuka •

Hiroshi Kubota•Yoshioki Oozeki •Hiroshi Nishida•

Louis Fortier

Received: 3 June 2009 / Accepted: 16 November 2009 / Published online: 16 January 2010

Ó The Japanese Society of Fisheries Science 2010

Abstract We tested whether the predation dynamics of

chub mackerel Scomber japonicus and spotted mackerel

S australasicus on young anchovy Engraulis japonicus

relates to individual growth characteristics of the prey and

could account for the growth-selective survival predicted

by recruitment hypotheses Juvenile and adult mackerel

were sampled along with their young anchovy prey field in

2004 (juvenile mackerel and larval anchovy) and 2005

(adult mackerel and juvenile anchovy) off the Pacific coast

of Honshu, Japan The recent 5-day mean growth rate of

larval and juvenile survivors and prey found in the stomach

of mackerel was estimated from the otolith microstructure

No significant difference was found between the recent

growth of larval or juvenile survivors and that of preyed

individuals We conclude that despite a relatively small

body size, the high activity level and predation skills

dis-played by mackerel prevent fast-growing larvae and early

juveniles from benefitting in terms of the expected survival

advantage over slow-growers Hence, growth-selective

predation mortality of larval fish would depend on the

feeding ecology of the predator rather than predator size

Selection for fast growth is more likely to occur under

predation pressure from invertebrate organisms and smallpelagic fish specialized on zooplankton, such as herringand anchovy

Keywords Growth rate
Growth-selective predation
Larval and juvenile anchovy
Mackerel

Otolith microstructure
Predation mortality

Introduction

In marine fish, strong year classes are often associated to ahigh survival rate during the first weeks or months of life[1 3] Early life survival variability would in turn be driven

by variable growth performance Fast-growing individualsgenerally experience lower cumulative vulnerability topredation, explaining their higher recruitment potentialrelative to slower growing conspecifics [4 6] Fast growersmay also gain a survival advantage due to their larger bodysize (‘‘bigger is better’’ hypothesis [6]) and the shorterduration of the larval stage (‘‘stage duration’’ hypothesis[7]) during which mortality is maximal Moreover, a highgrowth rate can directly translate into decreased vulnera-bility to predation (‘‘growth-selective predation’’ hypothe-sis [5]) In these mechanisms, predation is assumed to bethe major and direct source of mortality Such a potentiallinkage of growth to survival led to the assumption that thedetection of strong selection for fast growth during earlylife of a given cohort is a symptom of high predationpressure [8 10] However, the typical survival advantage

of fast growers under planktivorous predation pressure(e.g., small pelagic fish) tends to disappear when young fishface large piscivorous predators [11] As survival proba-bility at a given growth rate is predator-specific, thepredator field encountered during early life stages would

De´partement de Biologie, Que´bec-Oce´an,

Pavillon Alexandre-Vachon, Universite´ Laval,

Quebec, QC G1V 0A6, Canada

e-mail: dominique.robert@qo.ulaval.ca

National Research Institute of Fisheries Science,

Fisheries Research Agency, Yokohama,

Kanagawa 236-8648, Japan

Tokyo University of Marine Science and Technology,

Minato, Tokyo 108-8477, Japan

Fish Sci (2010) 76:183–188

DOI 10.1007/s12562-009-0205-y

regulate the characteristics of survivors through selection

on growth traits It would therefore appear to be

impor-tant to describe predation dynamics of the main larval fish

predators to determine whether their impact on survival

can be resolved by the analysis of selection for fast

growth [11]

Mackerel Scomber spp are often depicted as one of the

main predators of young stages of marine fish [12–15]

Their occurrence in most temperate coastal areas of the

world implies predation pressure on the larvae and

juve-niles of a large number of commercially important species

Despite representing a major potential source of mortality

for young fish, our knowledge of the predation dynamics of

mackerel remains limited to the results of a few qualitative

[16] and laboratory [12, 17–19] studies The objective of

this study was to evaluate the role of mackerel in driving

growth-selective survival in young fish Accordingly, we

provide a field-based test of the ‘‘growth-selective

preda-tion’’ hypothesis [5], proposing that growth rate directly

impacts vulnerability to predation, independent of body

size or stage duration Two types of predator–prey

inter-actions were assessed: (1) juvenile mackerel preying on

larval anchovy and (2) adult mackerel preying on juvenile

anchovy in the western North Pacific offshore Japan

Materials and methods

Japanese anchovy Engraulis japonicus larvae and

juve-niles, as well as their chub mackerel Scomber japonicus

and spotted mackerel S australasicus predators, were

collected concurrently offshore Honshu (Japan) in the

western North Pacific (Fig.1) Larval and juvenile anchovy

of standard length (SL) 20–65 mm (n = 47) and predatory

juvenile mackerel of fork length (FL) 52–113 mm

(n = 28) were simultaneously sampled using an otter trawl(mesh size 10 mm) The trawl was towed for 30 min at aspeed of 3.5 knots in the surface layer (depth \27 m) aftersunset on May 20, 2004 Adult mackerel (219–293 mm FL,

n = 61) were collected from two consecutive drift netdeployments (mesh size 19–157 mm), each followed by aconcurrent kite trawl (mesh size 5 mm) or MOHT [20] net(mesh size 1.59 mm) tow to sample potential anchovyprey The two different gear types used to capture younganchovy allowed us to sample the whole body lengthinterval (11–64 mm SL, n = 180) All of these deploy-ments were realized during nighttime hours on 7–8 June

2005 along the trajectory of a drifting buoy to optimize theprobability of repeatedly sampling the same potential preyand predator populations All samples were frozen at-20°C immediately following capture until being processed

in the laboratory

After taking the FL measurement of juvenile and adultmackerel, we dissected and identified the gut content,sorting out larval and juvenile Japanese anchovy Anchovyprey found in the gut of mackerel were digested to variousdegrees, but sagittal otoliths were retained in the head ofmost individuals In total, the otoliths of 13 and 163 indi-viduals ingested by juvenile and adult mackerel, respec-tively, were available for further analyses Young anchovycollected concurrently with predatory mackerel wereregarded as the original populations of these ingestedindividuals Anchovy prey were pooled among the twomackerel species

Sagittal otoliths were extracted from all young anchovy(prey and original populations) and mounted on a glassslide with enamel resin under a binocular microscope.Maximum otolith radius (OR) and daily growth incrementwidth were measured to the nearest 0.1 lm along a transectfrom the nucleus to the outermost margin Otolith

collected along with their young anchovy prey field

measurements were conducted using an otolith

measure-ment system (RATOC System Engineering, Bunkyo-ku,

Tokyo, Japan) that consists of a transmitting light

micro-scope equipped with a video camera connected to a

com-puter and monitor

Following Searcy and Sponaugle [21], we did not

reconstruct each individual somatic growth trajectory or

growth rate history via back-calculation techniques in order

to avoid potentially important sources of error In

partic-ular, the relation between otolith and somatic growth

changes through ontogeny in young anchovy (see below),

which makes it difficult to derive growth history from a

single reliable OR–SL equation Instead, OR and increment

width were considered to be proxies for somatic size and

daily growth rate, respectively A detrended growth index

was computed to reduce potential bias due to sudden

departures in the growth trajectory and allow the inclusion

of all individuals in the analysis, independent of their age

[22,23]:

where DGijis the detrended growth of individual I at age

j; and IW and SD are the increment width and standard

deviation, respectively This index provides a relative

measure of growth performance realized by a given

individual at a certain age, with mean growth achieved at

the same age for all individuals used as a reference point

It thus yields an estimate of growth remaining

invul-nerable to any ontogenetical shifts in growth patterns

Here, we defined recent growth performance of a given

individual as the average detrended growth achieved

during the last 5 full days of life (i.e., excluding the edge

of the otolith) Recent 5-day mean detrended growth

was compared between the ingested anchovy and their

original population by analysis of variance (ANOVA) or

analysis of covariance (ANCOVA) with the OR as a

covariate

Results

The relationship between OR and SL of the original larval

populations was expressed by strong allometric functions,

indicating that otolith growth is a reliable proxy for somatic

growth in young anchovy (Fig.2) The slope of the

rela-tionship changed at an OR of approximately 280 lm

(SL 30 mm), which corresponds to the transition between

the larval and juvenile stages [24] Due to this ontogenetic

change in the nature of the OR–SL relationship, as well as

the difficulty of precisely measuring the SL of ingested

larvae, we considered otolith growth directly for further

analyses

Juvenile mackerel preyed preferentially on larvalanchovy (Fig.3a), and we observed only one occurrence ofpredation on a juvenile anchovy (OR [ 280 lm) A posi-tive linear relation was found between recent 5-day meandetrended growth and OR over the larval size intervalconsumed by juvenile mackerel (Fig.3b) Both slopes andintercept did not differ between larval anchovy prey(n = 12) and original larval population (n = 34)(ANCOVA, P = 0.15), indicating that juvenile mackerelcaptured their larval prey independent of prey growth.Contrary to juvenile mackerel, adult mackerel mainlyconsumed juvenile anchovy and seldom preyed on thelarvae (OR \ 280 lm) available in the environment(Fig.4a) No difference was observed between the recent5-day mean detrended growth of juvenile anchovy prey(n = 141) and that of the original juvenile population(n = 114) over the predated size interval (Fig.4b; ANOVA,

r2= 0.01, P = 0.06)

DiscussionThe key role of mackerel Scomber spp as a regulator of fishpopulations through predation on early ontogenetic stageswas proposed nearly three decades ago [13] This hypoth-esis has been strongly supported by the strong negativerelationships between small pelagic fish abundance anddemersal fish recruitment reported in recent studies [15,25].However, while clupeid predation dynamics on larval fishhas been relatively well documented [26–28], field reports

length (SL) in young anchovy Allometric regressions for the larval and juvenile stages are Log(SL) = 0.101 Log(OR) ? Log(6.79)

of mackerel predation on ichthyoplankton remain few in

number and only qualitative in nature [16] Our study is the

first field-based assessment of both late juvenile and adult

mackerel predation dynamics on young fish and the

influ-ence of potential prey condition (given by recent growth) on

capture success

The possibility of a sampling bias in our measurements

of the prey field encountered by juvenile mackerel can not

be ruled out The 10-mm mesh net probably only retained a

partial fraction of the larval anchovy encountered It is also

likely that sampling efficiency increased with larval size A

possible consequence of such gear selection towards large

individuals is the simultaneous selection of the fastest

growing larvae The sampled prey field, however, more

than adequately covered the size range of larvae found in

the gut of juvenile mackerel, and no significant difference

in growth was found between ingested larvae and original

populations The similarity between growth of the prey and

that of the sampled prey field provides evidence that,

despite the risks of gear selection for fast-growing

indi-viduals, mackerel did not select for slow-growing anchovy

Another aspect that needs consideration is that the reported

absence of growth selection may partly be attributable to

sampling location The intensity of selection may vary with

spatial changes in larval growth potential (likely superior in

highly productive coastal areas) and with the predator field

[10] For example, Takahashi et al [29] observed an

inshore–offshore negative growth rate gradient in theKuroshio–Oyashio transition region, which larvae possiblyexperienced as a result of transport and migration pro-cesses In our study, the reported absence of growthselection in two distinct areas nevertheless suggests thatmackerel is not an important agent of growth-selectivemortality in young anchovy

Despite a relatively small body size, mackerel alwayscaptured their anchovy prey independent of growth rate.Mackerel thus strongly diverged from the general patternsuggested in a previous study [11] in which we classifiedthe potential fish predators of young anchovy in twocategories based on size In that study, small pelagic fish(anchovy, round herring Etrumeus teres, jack mackerelTrachurus japonicus, white croaker Pennahia argentatus)were identified as growth-selective predators, while largepredatory fish (sea bass Lateolabrax japonicus, greateramberjack Seriola dumerili, skipjack tuna Katsuwonuspelamis) were identified as non-growth-selective predatorsconsuming larvae randomly, independent of growth rate.Such a discrepancy can be resolved by shifting the focusfrom predator size to feeding strategy The presence ofmackerel in the non-growth-selective group along withlarge predatory fish may be attributable to the high activitylevel displayed by scombrids relative to other families Inparticular, mackerel adopt a raptorial feeding mode in thepresence of potential fish prey and increase their speed as a

recent 5-day mean otolith

growth rate (a) and recent

detrended 5-day mean growth

rate (b) over the ingested body

size range, and otolith radius for

larval anchovy survivors and

prey found in the gut of juvenile

mackerel

the recent 5-day mean otolith

growth rate (a) and recent

detrended 5-day (5d) mean

growth rate (b) over the

ingested body size range,

and otolith radius for juvenile

anchovy survivors and prey

found in the gut of adult

mackerel

function of prey size [19] Combined with particularly

acute vision [17] and persistence in attack [18], their high

foraging speed allows mackerel to capture large fish prey

independent of physiological condition, relative to other

small pelagic fish The propensity of a given predator to

induce growth-selective mortality in larval fish would thus

be attributable to its feeding ecology rather than to body

size alone

While we found that mackerel captured young anchovy

independent of growth rate, predation clearly remained a

size- or stage-selective process Juvenile mackerel preyed

almost exclusively on larval anchovy (with the exception

of one juvenile prey), while adult mackerel ingested only

juveniles, despite the systematic presence of both larvae

and juveniles in the prey assemblage In the sea, prey

selection is driven by the ratio of potential energy gain over

mean capture cost for a given prey type [30] This ratio

varies principally with two size-dependant factors:

detec-tion potential and probability of a successful attack [31] In

juvenile mackerel, even if the capture of a juvenile

anchovy constitutes a possible event (Fig.3), the high

probability that the potential prey escapes due to

well-developed anti-predator behavior would depress average

energy gain The larger and faster adult mackerel, however,

likely detect juvenile anchovy more easily than larvae and,

therefore, would be able to efficiently capture juvenile

anchovy due to their higher attack speed and mouth

aper-ture This result corroborates those from previous

labora-tory experiments which predicted that adult mackerel

would select the largest potential fish prey available in

the assemblage over the small, first-feeding individuals

[12,18]

We found that both juvenile and adult mackerel preyed

on large anchovy relative to their own body size Takasuka

et al [11] reviewed the characteristics of several fish

pre-dators of young Japanese anchovy From their Table 1, we

investigated the relationship between prey and predator

size among the species captured in the different surveys

(Fig.5) It appeared that both mackerel juveniles and

adults consumed relatively large anchovy and exhibited the

highest predator–prey size ratios observed to date While

this result may partly be attributable to differences in

ori-ginal prey populations, it remains consistent with those

obtained in the laboratory by Folkvord and Hunter [18],

who compared the performance of adult northern anchovy

Engraulis mordax and juvenile mackerel foraging on larval

and juvenile anchovy At an SL of 20 mm, larval anchovy

systematically escaped from the attacks of adult anchovy

In contrast, only 20% of these larvae could escape from

juvenile mackerel attacks, and an escape rate of 50% was

reached only at the onset of the juvenile stage, at a body

length of approximately 29 mm Contrary to most small

pelagic fish specialized on zooplankton (e.g., anchovy and

herring), mackerel bears the characteristics of a piscivorouspredator which may alternatively shift to filter-feeding onzooplankton when potential fish prey are scarce

The results from our study corroborate the conclusions

of laboratory studies [12, 18, 19] and provide the firstfield-based insight into the predation dynamics of mack-erel on early stages of fish Based on our findings,mackerel appears to be an efficient predator of late larvaland early juvenile anchovy, which they ingest indepen-dent of condition (recent growth) at anytime throughontogeny When preying on young fish, the switching bymackerel to a raptorial feeding mode generates anincrease in evacuation rate [19] This implies that satia-tion is seldom reached and that the potential impact ofmackerel predation on larval fish survival could increasewith larval density Despite demonstrating the charac-teristics of a major larval fish predator, mackerel capturesits prey independent of growth, which means that suchpredation pressure cannot be detected through an analysis

of growth selection The numerous reports of selection forfast growth during the larval stage of marine fish [9,10,32]would instead be attributable to predation from smallpelagic fish specialized on zooplankton or invertebrateorganisms

Hokuho-maru, Kaiun-maru, and Soyo-maru and to M Takahashi,

A Yatsu, and N Yamashita for their support during the sampling at sea We also thank S Shindo and A Matsuura for their help with otolith preparation, and M Saito for otolith reading DR was funded

by a joint scholarship of the Japan Society for the Promotion of Science (JSPS) and the Natural Science and Engineering Research Council of Canada (NSERC).

mackerel (results from this study) and the various fish predators

ingested by mackerel was estimated using the OR–SL relationships

1 Anderson JT (1988) A review of size dependent survival during

pre-recruit stages of fishes in relation to recruitment J Northwest

Atl Fish Sci 8:55–66

2 Cushing DH (1990) Plankton production and year-class strength

in fish populations: an update of the match/mismatch hypothesis.

Adv Mar Biol 26:249–294

3 Houde ED (2008) Emerging from Hjort’s shadow J Northwest

Atl Fish Sci 41:53–70

4 Chambers RC, Leggett WC (1987) Size and age at

metamor-phosis in marine fishes––an analysis of laboratory-reared winter

flounder (Pseudopleuronectes americanus) with a review of

variation in other species Can J Fish Aquat Sci 44:1936–1947

5 Takasuka A, Aoki I, Mitani I (2003) Evidence of

growth-selec-tive predation on larval Japanese anchovy Engraulis japonicus in

Sagami Bay Mar Ecol Prog Ser 252:223–238

6 Miller TJ, Crowder LB, Rice JA, Marschall EA (1988) Larval

size and recruitment mechanisms in fishes: toward a conceptual

framework Can J Fish Aquat Sci 45:1657–1670

7 Houde ED (1987) Fish early life dynamics and recruitment

variability Am Fish Soc Symp 2:17–29

8 Beyer JE, Laurence GC (1980) Stochastic-model of larval fish

growth Ecol Model 8:109–132

9 Meekan MG, Fortier L (1996) Selection for fast growth during

the larval life of Atlantic cod Gadus morhua on the Scotian Shelf.

Mar Ecol Prog Ser 137:25–37

10 Robert D, Castonguay M, Fortier L (2007) Early growth and

recruitment in Atlantic mackerel: discriminating the effects of

fast growth and selection for fast growth Mar Ecol Prog Ser

337:209–219

11 Takasuka A, Aoki I, Oozeki Y (2007) Predator-specific

growth-selective predation on larval Japanese anchovy Engraulis

japo-nicus Mar Ecol Prog Ser 350:99–107

12 Pepin P, Pearre SJ, Koslow JA (1987) Predation on larval fish by

Atlantic mackerel, Scomber scombrus, with a comparison of

predation by zooplankton Can J Fish Aquat Sci 44:2012–2018

13 Lett PF (1980) A comparative study of the recruitment

mecha-nisms of cod and mackerel, their interaction, and its implication

for dual stock management Can Tech Rep Fish Aquat Sci 988.

St Andrews

14 Darbyson E, Swain DP, Chabot D, Castonguay M (2003) Diel

variation in feeding rate and prey composition of herring and

mackerel in the southern Gulf of St Lawrence J Fish Biol

63:1235–1257

15 Swain DP, Sinclair AF (2000) Pelagic fishes and the cod

recruitment dilemma in the Northwest Atlantic Can J Fish Aquat

Sci 57:1321–1325

16 Rosel PE, Kocher TD (2002) DNA-based identification of larval

cod in stomach contents of predatory fishes J Exp Mar Biol Ecol

267:75–88

17 Macy WK, Sutherland SJ, Durbin EG (1998) Effects of zooplankton size and concentration and light intensity on the feeding behavior of Atlantic mackerel Scomber scombrus Mar Ecol Prog Ser 172:89–100

18 Folkvord A, Hunter JR (1986) Size-specific vulnerability of northern anchovy, Engraulis mordax, larvae to predation by fishes Fish Bull US 84:859–869

19 Pepin P, Koslow JA, Pearre S (1988) Laboratory study of foraging by Atlantic mackerel, Scomber scombrus, on natural zooplankton assemblages Can J Fish Aquat Sci 45:879–887

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

21 Searcy SP, Sponaugle S (2000) Variable larval growth in a coral reef fish Mar Ecol Prog Ser 206:213–226

22 Pepin P, Dower JF, Benoit HP (2001) The role of measurement error on the interpretation of otolith increment width in the study

of growth in larval fish Can J Fish Aquat Sci 58:2204–2212

23 Robert D, Castonguay M, Fortier L (2009) Effects of preferred prey density and temperature on feeding success and recent growth in larval mackerel of the southern Gulf of St Lawrence Mar Ecol Prog Ser 377:227–237

24 Takahashi M, Watanabe Y (2004) Staging larval and early juvenile Japanese anchovy based on the degree of guanine deposition J Fish Biol 64:262–267

25 Frank KT, Petrie B, Choi JS, Leggett WC (2005) Trophic cascades in a formerly cod-dominated ecosystem Science 308:1621–1623

26 Daan N, Rijnsdorp AD, Vanoverbeeke GR (1985) Predation by North Sea herring Clupea harengus on eggs of plaice Pleuro- nectes platessa and cod Gadus morhua Trans Am Fish Soc 114:499–506

27 Ko¨ster FW, Mo¨llmann C (2000) Trophodynamic control by clupeid predators on recruitment success in Baltic cod? ICES J Mar Sci 57:310–323

28 Last JM (1989) The food of herring, Clupea harengus, in the North Sea, 1983–1986 J Fish Biol 34:489–501

29 Takahashi M, Watanabe Y, Kinoshita T, Watanabe C (2001) Growth of larval and early juvenile Japanese anchovy, Engraulis japonicus, in the Kuroshio-Oyashio transition region Fish Oceanogr 10:235–247

30 Robert D, Castonguay M, Fortier L (2008) Effects of intra- and inter-annual variability in prey field on the feeding selectivity of larval Atlantic mackerel (Scomber scombrus) J Plankton Res 30:673–688

31 Buskey EJ, Coulter C, Strom S (1993) Locomotory patterns of microzooplankton: potential effects on food selectivity of larval fish Bull Mar Sci 53:29–43

32 Plaza G, Ishida M (2008) The growth-mortality relationship in larval cohorts of Sardinops melanostictus, revealed by using two new approaches to analyse longitudinal data from otoliths J Fish Biol 73:1531–1553

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

Scale calcification in the goldfish in vitro: histological

and quantitative analysis

Nobuhiro Ogawa•Kazuhiro Ura•Yasuaki Takagi

Received: 21 July 2009 / Accepted: 16 November 2009 / Published online: 26 December 2009

Ó The Japanese Society of Fisheries Science 2009

Abstract The external layer of a teleost fish scale is

composed of type I collagen, an amorphous matrix

sub-stance and hydroxyapatite crystals Calcification of this

layer can be inhibited in the scale-regenerating process

under calcium- and phosphate-deficient (CaDPD)

condi-tions, and can be facilitated by incubation in physiological

saline The aim of this study was to evaluate this model of

calcification using histological and quantitative analysis in

order to promote further understanding of the mechanism

of calcification in fish scales We found that the external

layer of the scales produced under CaDPD conditions

contained more densely aligned collagen fibrils with a

small amount of the amorphous matrix substance The

CaDPD scale contained only one-third of the amount of

calcium and phosphate present in the control fish After 4

hours of incubation, a two- to threefold increase in calcium

content and a 1.5-fold increase in phosphate content were

observed Calcification proceeded in the external layer, and

mineral deposits grew only in the electron-dense matrix

substance We conclude that this model would be suitable

for studying the early process of fish scale calcification that

occurs in the noncollagenous substance The

electron-dense substance may contain key molecules that promote

calcification

Keywords Biomineralization
Carasssius auratus
Goldfish
In vitro calcification
Scale regeneration

IntroductionThe innermost layer of the teleost fish scale, the basal plate,

is a collagenous plate that has a unique fibril alignmentsimilar to a mammalian cornea In both the basal plateand the cornea, multiple collagen layers are stacked, andthe fibrils in one layer are parallel but those in differentlayers are orthogonal [1 4] As a result, the layers form

a plywood-like structure Because of the extremely highregenerating ability of the scale, the process of scaleregeneration is a good model with which to study thecellular mechanism employed to construct cornea-likecollagenous extracellular matrix [5] Clarification of themechanism is a prerequisite for bioinspired fabrication ofartificial collagenous matrices to replace an injured cornea

in regenerative medicine [5] Our research is now cerned with the fabrication of such matrices using scalecollagen, since it has low risk of zoonotic infection com-pared with the current major collagen source (bovine andporcine) for medical use

con-Another layer of the scale, the external layer, has astructure similar to that of fibrous bone or dentine Itconsists of calcium phosphate hydroxyapatite crystals,collagen (mainly type I), and amorphous noncollagenoussubstances [6 9] Therefore, the process of regenerating theexternal layer would be a good model with which to studythe cellular mechanism employed to construct a bone/dentine-like calcified collagenous matrix Bone is a tissuethat has a limited ability to regenerate, and so bone sub-stitutes are often required for clinical treatment Althoughthe fabrication of artificial materials suitable for such a

Graduate School of Fisheries Sciences, Hokkaido University,

3-1-1 Minato-cho, Hakodate, Hokkaido 041-8611, Japan

Present Address:

Biomaterial Center, National Institute for Materials Science,

1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan

e-mail: OGAWA.Nobuhiro@nims.go.jp

Fish Sci (2010) 76:189–198

DOI 10.1007/s12562-009-0197-7

purpose has long been attempted, at present there is no

satisfactory one that has excellent biocompatibility, high

osteoconductivity and bioactivity In order to fabricate

high-performance bone substitutes, it is important to mimic

the composition and organization of bone extracellular

matrix, which is generally an important determinant of

cellular behavior such as cellular adhesion, migration,

proliferation and differentiation [10] Regulating the size

and organization of the hydroxyapatite crystals is quite an

important task, as can be inferred from the fact that gene

expression patterns related to osteoblast adhesion,

prolif-eration, extracellular matrix synthesis and differentiation

are regulated in part by the size of the sieved

hydroxy-apatite in MC3T3-E1 mouse osteoblastic cells [11]

However, studies on the mechanism of scale regeneration

including early mineral deposition are scarce, except for

histological observations of the regenerating process For

this reason, our understanding of this process from a

physiological, biochemical and molecular perspective

remains incomplete

Iguchi [12] and Yamane et al [13] reported that fish

reared in distilled water and fed a calcium- and

phosphate-deficient diet (CaDPD conditions) regenerate uncalcified

scales, and that these uncalcified regenerating scales can be

subsequently artificially calcified in vitro This technique is

useful for investigating the precise physiological and

molecular mechanisms underlying early mineral

deposi-tion However, the structure of the scales regenerated under

CaDPD conditions, the calcification levels of the

uncalci-fied and artificially calciuncalci-fied scales, and the ultrastructure

of the mineral deposits and their surrounding areas after

incubation are unknown For these reasons, this model

cannot yet be accurately applied to studies of calcification

mechanisms

In this study, we have re-evaluated this mineralization

model First, we quantified the calcium and phosphate

levels of scales regenerated under the CaDPD condition

Second, we evaluated the effects of incubation in

physio-logical saline on the calcium and phosphate levels of the

regenerating scales Thereafter, light and electron

micro-scopic observations were conducted to compare the

struc-tures and mineralization states of scales regenerated under

control and CaDPD conditions, as well as those of

regen-erated scales that were subsequently calcified in vitro

Materials and methods

Fish

Goldfish (Carassius auratus; weighing about 48.4 g) were

purchased from a commercial dealer They were kept in tap

water at 26°C and fed a commercial diet (4C; Nippon

Formula Feed Mgf Co., Ltd., Kanagawa, Japan) 4 times aday (total 4% of body weight/day) for at least 1 weekbefore the experiment

Experimental diet

A calcium- and phosphate-containing diet (control diet)and a calcium- and phosphate-deficient diet (CaDPD diet)were prepared largely by following the method of Takagi

et al [14] The compositions of the two diets are listed inTable1 The calcium contents of the control diet and theCaDPD diet were 15.64 and 0.49 mg/g dry weight,respectively The phosphate contents of the control andthe CaDPD diet were 42.3 and 1.78 mg/g dry weight,respectively

Experimental protocolAfter the fish had been anesthetized in a 0.02% solution of2-phenoxyethanol, every second scale was removed fromeach entire flank Next, the fish were divided into twogroups, each containing eight fish Fish in the first group(‘‘control fish’’) were reared in tap water at 26°C and fedthe control diet four times a day (a total of 4% of bodyweight/day) Fish in the second group (‘‘CaDPD fish’’)were reared in deionized water at 26°C and fed the CaDPDdiet four times a day (a total of 4% of body weight/day).After 7 days, the fish were again anesthetized and all of the

then defatted with ethyl ether

CoCl 2
6H2 O (0.420%)

regenerating scales were collected; these scales were

called the ‘‘first-trial scales.’’ The fish were then reared

for another 7 days in each condition, and all of the

regenerating scales were collected again (‘‘second-trial

scales’’) The fish were reared for a further 7 days in each

condition, and the regenerating scales were obtained

(‘‘third-trial scales’’) The regenerating scales collected

were subjected to calcium and phosphate content analysis

as well as light and electron microscopy A portion of the

regenerating scales was also used for the in vitro

calcifi-cation experiment

In vitro calcification

A portion of the regenerating scales was incubated in

physiological saline (Na?: 154 mM; K?: 7 mM; Ca2?:

2.5 mM; Mg2?: 1.5 mM; PO43-: 2 mM; glucose: 10 mM;

HEPES: 10 mM; pH 7.4) for 4 h at 37°C After this

incubation, the scales were subjected to calcium and

phosphate content analysis, as well as light and electron

microscopy

Ca and Pi content analysis

The scales were dried overnight at 80°C, weighed and

solubilized in concentrated HNO3at 40°C Calcium levels

were measured by a Hitachi Z-6100 polarized Zeeman

atomic absorption spectrophotometer (Tokyo, Japan)

Inorganic phosphate levels were determined by following

the method of Goldenberg and Fernandez [15] using a

Hitachi U-2000 spectrophotometer (Tokyo, Japan)

Light microscope observation

The regenerating scales were fixed in 70% ethanol and

divided into two groups The scales in one group were

immersed in a 1% silver nitrate solution and illuminated

with fluorescent light for 10 min to detect calcification

sites Next, they were rinsed in deionized water, mounted

on glass slides and examined with a Nikon eclipse E800

light microscope (Tokyo, Japan) The scales in the other

group were dehydrated in a graded series of ethanol and

embedded in JB-4 resin (Polysciences, Inc.) Frontal

sec-tions (3 lm thickness) were cut with a Reichert–Jung 2050

microtome (Heidelberg, Germany), stained with 1% silver

nitrate and a mixture of methylene blue and azure II, and

observed under a light microscope

Transmission electron microscope (TEM) observation

The scales were immersed in a mixture of 2%

parafor-maldehyde and 2.5% glutaraldehyde in 0.05 M cacodylate

buffer (pH 7.6) for 2 h at room temperature After a short

rinse in buffer, the samples were postfixed for 1 h in 1%OsO4in 0.05 M cacodylate buffer (pH 7.6) The sampleswere then rinsed in buffer, dehydrated in a graded series ofethanol, and embedded in Spurr resin (Agar Scientific Ltd.,Stansted, UK) Ultrathin sections were cut with a Reichert–Nissei Ultracut N ultramicrotome (Heidelberg, Germany)using a diamond knife Sections were unstained or stainedwith uranyl acetate and lead citrate and examined with aHitachi H-7000 transmission electron microscope (Tokyo,Japan)

StatisticsStudent’s t test for unpaired data was used to test the sig-nificance of the means between the control and CaDPDgroup Student’s t test for paired data was used to test thesignificance of the means of the calcium and phosphatecontent of the scales before and after incubation in the

in vitro calcification experiments Statistical significancewas set at P \ 0.05 All data are reported as the mean ±standard error All analyses were performed with Stat View5.0 for Windows (SAS Institute Inc.)

ResultsCalcium and phosphate contents of the regeneratingscales

As shown in Fig 1, the dry weight of the regeneratingscales collected from CaDPD fish was about 80% of thatfrom control fish in all trials Figure2 shows the calcium

0 1.0 2.0 3.0 4.0 5.0

condi-tions on the dry weight of scales after 7 days of regeneration Values are mean ± SE for 7 samples *Significantly different (P \ 0.05) by unpaired t test

and phosphate contents of the regenerating scales before and

after incubation in the in vitro calcification experiments

Before the incubation, the calcium and phosphate contents of

the first-trial scales from CaDPD fish were about 30 and

33%, respectively, of those of first-trial scales from control

fish Moreover, the calcium and phosphate contents of the

second- and third-trial scales from CaDPD fish were even

smaller than those of the first-trial scales After incubation,

the calcium and phosphate contents of the regenerating

scales collected from control fish increased only slightly: thecalcium content was only significantly higher after incuba-tion in the third-trial scales The calcium content of thescales from CaDPD fish was increased by two- to threefoldafter the incubation The phosphate content of the regener-ating scales from CaDPD fish was increased by about1.5-fold after the incubation However, the calcium andphosphate contents of the scales from CaDPD fish did notreach the levels seen in the control fish after the incubation

0 1.0 2.0 3.0

scales after 7 days of

regeneration before and after

incubation in physiological

saline for 4 h at 37°C.

Regenerating scales were

collected from calcium- and

phosphate-sufficient (control) or

-deficient (CaDPD) conditions.

Values are mean ± SE for 7

samples *Significantly different

(P \ 0.05) by paired t test.

(P \ 0.05) by unpaired t test

Light microscopy

Figure3 shows a surface view of the first-trial scales

stained with silver nitrate, before and after the incubation

Before the incubation, the control scales showed strong

staining with silver nitrate except for the marginal narrow

area and grooves In the scales from CaDPD fish, by

con-trast, silver nitrate staining was weaker than that in the

scales from control fish In particular, the marginal wide

area did not react with silver nitrate The results were

almost the same for the second- and third-trial scales After

the incubation, both the control and the CaDPD scales

reacted well with silver nitrate

In the histological sections (Fig.4), the regenerating

scales from control fish comprised an outer flat-cell layer,

an external layer that reacted with silver nitrate, the basal

plate, and an innermost cubic-cell layer The structure was

the same in the scales collected from CaDPD fish;

how-ever, the external layer of the regenerating scales from

CaDPD fish was scarcely stained with silver nitrate and had

more pores and concavities containing several cells

Around the pores and concavities, the silver nitrate stainingwas even weaker than in the other parts of the externallayer After the incubation, silver nitrate reacted only withthe external layers of scales collected from both the controland CaDPD fish The extent of the reaction with the scales

of the CaDPD fish increased after the incubation

Transmission electron microscopy

In the regenerating scales collected from the control fish,the external layer was morphologically divided into fivelayers, mainly depending on the diameter of the collagenfibrils (Fig.5) The outermost layer, to which flat cellswere attached, was composed of scattered collagen fibrils(about 10 nm in diameter) with an electron-light substance(the A layer) The next layer comprised relatively assem-bled but scattered collagen fibrils with a diameter of about

20 nm (the B layer) The spaces between the fibrils werefilled with electron-dense and electron-light amorphoussubstances Next to the B layer, a layer with thick collagenfibrils (about 50 nm diameter) was observed (the C layer)

7 days of regeneration a A scale that regenerated under control

conditions; b a scale that regenerated under control conditions and

was incubated for 4 h in physiological saline; c a scale that

regenerated under CaDPD conditions; d a scale that regenerated

under CaDPD conditions and was incubated for 4 h in physiological saline Arrows in a and c indicate marginal narrow areas, which were not stained with silver nitrate Scales were stained with silver nitrate.

G, groove Bars 50 lm

The collagen fibrils were aggregated and formed thick

bundles Electron-dense and electron-light substances filled

the spaces between the collagen fibrils This layer was in

part patchily distributed in the innermost of the external

layer Under the C layer, the D layer, whose structural

features were similar to the B layer, was observed The

innermost layer was composed of collagen fibrils (about

50 nm in diameter) that were most aggregated and

rela-tively aligned (the E layer) In the CaDPD scales, the

amount of noncollagenous matrix in the external layer was

much less than that in the control scales, resulting in

densely packed collagen fibrils in the external-layer matrix

Moreover, only the B- and E-like layers were observed

The B-like layer was composed of two parts One had a

small amount of electron-dense substances between the

fibrils, whereas the other had electron-light substances

None of the boundaries between the layers was clear

To investigate the mineral deposits in the external layer,

unstained sections were examined (Fig.6) In the unstained

sections of the control scales, the external layer comprised

amorphous electron-dense and electron-light substances,

and electron-dense spindle-shaped particles were observed

in the electron-dense substance The central region of theexternal layer (where the B and C layers may be distrib-uted) were rich in the electron-dense substance and parti-cles, as observed in the stained sections In the externallayer of the CaDPD scales, by contrast, the electron-denseparticles and substances were scarcely observed Afterincubating the CaDPD scales, the electron-dense substancewas observed more frequently in the central region of theexternal layer, where B layer was distributed in the stainedsections, and the spindle-shaped particles appeared withinthis substance However, the particles were smaller thanthose in the control scales The amount of particles in theCaDPD scales was also less than in the control scales

Discussion

We have shown here, using both histological and tative methods, that fish reared under CaDPD conditionsregenerated scales in which calcification was inhibited It

7 days of regeneration a A scale that regenerated under control

conditions; b a scale that regenerated under control conditions and

was incubated for 4 h in physiological saline; c a scale that

regenerated under CaDPD conditions; d a scale that regenerated

under CaDPD conditions and was incubated for 4 h in physiological saline Arrows in c and d indicate pores and concavities All sections were stained with silver nitrate and methylene blue–azure II EL, external layer; BP, basal plate Bars 50 lm

was also observed that calcification of these scales in

physiological saline could only proceed in their external

layers, as observed in the control regenerating scales in

vivo Moreover, electron-dense, spindle-shaped particles,

which morphologically resembled mineral deposits present

in fish scales [2, 7, 16–18], grew in an electron-dense

substance in both the regenerated scales from control fish

and the in vitro incubated scales from CaDPD fish,

although the amount of this substance was smaller in thelatter These findings suggest that the mechanisms of cal-cification induced in the CaDPD scales in vitro and thecontrol scales in vivo are comparable, and so the CaDPDscale is a good model with which to study the mechanismunderlying the early calcification of teleost scales.The CaDPD scale is composed of an external layer and abasal plate, as is a control regenerating scale In the CaDPD

of scales after 7 days of regeneration Five layers (A–E) were

distinguished in scales regenerated under control conditions (a),

whereas only two layers (B, E) were distinguished in scales

regenerated under CaDPD conditions (b) Stained with uranyl acetate

and lead citrate Bars 1 lm Higher magnification images of each layer that were used to compare collagen fibril diameters are shown in c–g c The A layer; d the B layer; e the C layer; f the D layer; g the E layer Stained with uranyl acetate and lead citrate Bars 0.2 lm

scale, however, the dry weight was lower and the density of

collagen fibrils was higher than that in the control scale

These results indicate that deposition of the scale matrix

components, especially that of noncollagenous substances,

is inhibited under CaDPD conditions In addition, Iguchi

[12] and Yamane et al [13] reported that goldfish reared

under CaDPD conditions resulted in hypophosphatemia but

not hypocalcemia Under calcium-deficient conditions,

tilapia Oreochromis niloticus induces calcium mobilization

from its scales and bones and prevents hypocalcemia [14,

19] Thus, inhibition of calcification in the CaDPD scales is

caused by both a smaller amount of noncollagenous matrix

and hypophosphatemia, but not by hypocalcemia

Matrix vesicles are the membrane-bound particles

located in the extracellular matrix of various calcified

tissues They contain a putative calcification-inducing

enzyme, alkaline phosphatase, and are generally accepted

as being calcification-inducing structures in various

calci-fied tissues in mammals [20–23] In previous studies of

regenerating scales, some researchers have observed

matrix-vesicle-like structures with hydroxyapatite-like

crystals in the marginal scale area, where calcification

initiates [2,7,16–18] However, a mineral deposit has also

been observed in the external scale layer, independent of

such structures Olson and Watabe [24] showed that the

mineral deposits of developing scales were not related to

observable substances or structures in the sheepsheadminnow Cyprinodon variegatus In the developing scales

of chum salmon Oncorhynchus keta, goldfish, guppy istes reticulatus and Poecilia reticulata, tilapia Tilapiarendalli and rainbow trout Salmo gairdnerii irideus, min-eral deposits are observed in the interfibrillar electron-dense substance [2,6,8,16,25] Thus, the matrix vesiclesare not prerequisites for scale calcification, as suggested byBereiter-Hahn and Zylberberg [2] In this study, using the

Leb-in vitro system, we succeeded Leb-in showLeb-ing that calcificationduring goldfish scale regeneration is initiated in the elec-tron-dense substance, independent of matrix vesicles.Therefore, the present in vitro system is suitable forstudying the mechanism of matrix-vesicle-independentcalcification

Using histochemical methods, Maekawa and Yamada[6] suggested that electron-dense particles, containingsulfated mucopolysaccharides and protein–mucopolysac-charide complexes, play a role in the nucleation of themineral phase in the scales of juvenile rainbow trout Theelectron-dense substances that we observed in this studymay also be calcification-inducing proteoglycan compos-ites, as suggested by Maekawa and Yamada [6] It is dif-ficult to imagine that a 4-h incubation could induce therapid synthesis of calcification-related organic matrices byscale-forming cells We therefore propose that these

micrographs of the external

layers of scales after 7 days of

regeneration a A scale

regenerated under control

conditions; b a scale

regenerated under CaDPD

conditions; c a scale regenerated

under CaDPD conditions and

incubated for 4 h in

physiological saline.

Arrows indicate putative

mineral deposits Unstained.

Bars 0.5 lm

molecules may already exist in the CaDPD scale matrix,

dispersed in the interfibrillar spaces, although there may be

less of them than in normal scales Incubation in

physio-logical saline may induce these molecules to aggregate and

show electron density, and the process of calcification may

be induced by the aggregated macromolecules The

importance of aggregated macromolecules in the initiation

of calcification was also reported for mammalian bone

Recently, it has been reported that local extracellular

nucleation complexes called ‘‘biomineralization foci,’’

which are independent structures with matrix vesicles,

consist mainly of bone acidic glycoprotein-75 (BAG-75)

and bone sialoprotein in UMR106-01 osteoblastic cultures

[26] BAG-75 has the ability to self-associate into 10–

20 lm supramolecular spherical complexes composed of

microfibrils [27,28] Immunohistological analysis using a

BAG-75 aggregation-specific monoclonal antibody showed

that a BAG-75 complex exists in vivo at the site of new

primary bone formation [28], suggesting that

matrix-vesi-cle-independent calcification also occurs in mammalian

bone We assume that the electron-dense matrix substance

observed in this study may comprise aggregated

macro-molecules similar to the biomineralization foci in

mammalian bone, and may play an important role in the

matrix-vesicle-independent calcification of scales

Calcification-related organic matrices that can promote

hydroxyapatite nucleation have long been studied in

mammalian calcified tissues Nevertheless, only a few

proteins have been identified as crystal nucleators [29] The

method developed in this study will be a key tool for

identifying organic matrices that promote scale

calcifica-tion, especially in the early stages of calcification Further

biochemical and molecular biological investigations will

reveal the molecular composition of the electron-dense

substance and the mechanism underlying matrix

calcifica-tion in detail Such informacalcifica-tion will be applicable to the

development of artificial bone using scale collagen

Grant-in-Aid for Exploratory Research (No 16658077) and a Twenty-First

Century COE program awarded by the Ministry of Education,

Cul-ture, Sports, Science and Technology of Japan The study was also

supported in part by a Grant-in-Aid for Creative Scientific Research

(No 17GS0311) and a Research Grant for Young Scientists (No

18-04520), awarded by the Japan Society for Promotion of Science.

References

1 Zylberberg L, Ge´raudie J, Meunier FJ, Sire J-Y (1992)

Biomin-eralization in the integumental skeleton of the living lower

vertebrates In: Hall BK (ed) Bone 4 Bone metabolism and

mineralization CRC, Boca Raton, pp 171–224

2 Bereiter-Hahn J, Zylberberg L (1993) Regeneration of teleost fish

scale Comp Biochem Physiol 105A:625–641

3 Huysseune A, Sire J-Y (1998) Evolution of patterns and cesses in teeth and tooth-related tissues in non-mammalian ver- tebrates Eur J Oral Sci 106:437–481

pro-4 Sire J-Y, Akimenko M-A (2004) Scale development in fish: a review, with description of sonic hedgehoc (shh) expression in the zebrafish (Danio rerio) Int J Dev Biol 48:223–247

5 Takagi Y, Ura K (2007) Teleost fish scales: a unique biological model for the fabrication of materials for corneal stroma regen- eration J Nanosci Nanotechnol 7:757–762

6 Maekawa K, Yamada J (1970) Some histochemical and fine structural aspects of growing scales of the rainbow trout Bull Fac Fish Hokkaido Univ 21:70–77

7 Sire J-Y, Ge´raudie J (1984) Fine structure of regenerating scales and their associated cells in the cichlid Hemichromis bimaculatus (Gill) Cell Tissue Res 237:537–547

8 Zylberberg L, Nicolas G (1982) Ultrastructure of scales in osts (Carassius auratus L.) after use of rapid freeze-fixation and freeze-substitution Cell Tissue Res 223:349–367

tele-9 Sauk JJ, Cocking-Johnson D, Cervenka VA, Van Kampen CL (1984) Noncollagenous phosphoprotein derived from teleostean fish-scales Biochim Biophys Acta 798:199–203

10 Ekblom P, Vestweber D, Kemler R (1986) Cell–matrix tions and cell adhesion during development Ann Rev Cell Biol 2:27–47

interac-11 Xie J, Baumann MJ, McCabe LR (2004) Osteoblasts respond to hydroxyapatite surfaces with immediate changes in gene expression J Biomed Mater Res 71A:108–117

12 Iguchi M (1985) Gyoruikousoshiki (hone uroko) no seichou to sekkaika ni okeru karusiumu to rin no kyodou ni kansuru kenkyu (Studies on the actions of calcium and phosphate in the growth and calcification of fish bones and scales) (Ph.D dissertation) Hokkaido University, Hakodate (in Japanese)

13 Yamane S, Iguchi M, Ogasawara T, Nakamura Y (1982) Effects

of blockage of exogenous calcium and phosphorus on the calcium regulatory systems in goldfish Comp Biochem Physiol 72A:709– 713

14 Takagi Y, Hirano T, Yamada J (1989) Scale regeneration of tilapia (Oreochromis niloticus) under various ambient and dietary calcium concentrations Comp Biochem Physiol 92A:605–608

15 Goldenberg H, Fernandez A (1966) Simplified method for the estimation of inorganic phosphorus in body fluids Clin Chem 12:871–882

16 Yamada J (1971) A fine structural aspect of the development

of scales in the chum salmon fry Nippon Suisan Gakkaishi 37:18–29

17 Fouda MM (1979) Studies on scale regeneration in the Common goby, Pomatoschistus microps (Pisces) J Zool Lond 189: 503–509

18 Sire J-Y, Ge´raudie J (1983) Fine structure of developing scale in the Chichlid Hemichromis bimaculatus (Pisces, Teleostei, Perci- formes) Acta Zool 64:1–8

19 Takagi Y, Yamada J (1992) Effect of calcium deprivation on the metabolism of acellular bone in tilapia, Oreochromis niloticus Comp Biochem Physiol 102A:481–485

20 Anderson HC (1984) Mineralization by matrix vesicles Scan Electron Microsc 2:953–964

21 Anderson HC (1995) Molecular biology of matrix vesicles Clin Orthop Relat Res 314:266–280

22 Anderson HC (2003) Matrix vesicles and calcification Curr Rheumatol Rep 5:222–226

23 Arsenault AL, Frankland BW, Ottensmeyer FP (1991) Vectorial sequence of mineralization in the turkey leg tendon determined

by electron microscopic imaging Calcif Tissue Int 48:46–55

24 Olson OP, Watabe N (1980) Studies on formation and resorption

of fish scales IV Ultrastructure of developing scales in newly

hatched fry of the Sheepshead minnow (Cyprinodon variegatus).

Cell Tissue Res 211:303–316

25 Scho¨rnbo¨rner AA, Boivin G, Baud CA (1979) The mineralization

processes in teleost fish scales Cell Tissue Res 202:203–212

26 Huffman NT, Keightley JA, Chaoying C, Midura RJ, Lovitch D,

Veno PA, Dallas SL, Gorski JP (2007) Association of specific

proteolytic processing of bone sialoprotein and bone acidic

gly-coprotein-75 with mineralization within biomineralization foci.

J Biol Chem 282:26002–26013

27 Gorski JP, Kremer EA, Chen Y (1996) Bone acidic

glycoprotein-75 self-associates to form large macromolecular complexes.

Connect Tissue Res 35:137–143

28 Gorski JP, Kremer EA, Chen Y, Ryan S, Fullenkamp C, scio J, Jensen K, McKee MD (1997) Bone acidic glycoprotein-75 self-associates to form macromolecular complexes in vitro and

Delvi-in vivo with the potential to sequester phosphate ions J Cell Biochem 64:547–564

29 Hunter GK, Hauschka PV, Poole AR, Rosenberg LC, Goldenberg

HA (1996) Nucleation and inhibition of hydroxylapatite tion by mineralized tissue proteins Biochem J 317:59–64

forma-30 Halver JE, Coastes JA (1957) A vitamin test diet for long-term feeding studies Prog Fish Cult 19:112–118

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

Mg-calcite, a carbonate mineral, constitutes Ca precipitates

produced as a byproduct of osmoregulation in the intestine

of seawater-acclimated Japanese eel Anguilla japonica

Miyuki Mekuchi•Tamao Hatta •Toyoji Kaneko

Received: 11 September 2009 / Accepted: 20 November 2009 / Published online: 4 February 2010

Ó The Japanese Society of Fisheries Science 2010

Abstract Marine teleosts are known to produce white

feces, which is often referred to as Ca precipitates Ca

precipitates have been suggested to be a product of

osmoregulation In the present study, we examined the

physicochemical nature of Ca precipitates, and possible

involvement of Ca precipitate formation in

hyposmoregu-latory processes in seawater-acclimated Japanese eel

Whereas Ca precipitates were not produced in eel

accli-mated to freshwater, Ca precipitates were seen in eel

acclimated to seawater in a salinity-dependent manner

According to X-ray diffraction analysis, Ca precipitates

were a mixture of carbonate minerals: Mg-calcite and its

amorphia Quantitative analysis showed that the molar ratio

between Ca and Mg was approximately 7:2 Ca precipitate

formation was reduced in eel exposed to low-Ca2?or

low-Mg2? seawater, indicating that Ca and Mg in Ca

precipi-tates were derived from seawater

Keywords Ca precipitate
Carbonate mineral

Japanese eel
Mg-calcite
Osmoregulation

Introduction

Marine teleosts inhabiting hypertonic environments

main-tain their blood plasma osmolality at levels equivalent to

about one-third of seawater osmolality by exerting posmoregulatory ability [1] Marine teleosts are continu-ously in danger of dehydration, which is caused by anosmotic gradient between the blood and ambient seawater.Since the early work by Smith [2], it has become wellaccepted that marine teleosts drink a large amount ofambient seawater to compensate for osmotic water lossacross the body surface

hy-Water and ion movements during the passage of ted seawater along the gastrointestinal tract have been welldescribed in marine or seawater-adapted teleosts (forreview, see Marshall and Grosell [3]) In seawater-accli-mated Japanese eel Anguilla japonica and European eelAnguilla anguilla, Hirano and Mayer-Gostan [4] haveshown that ingested water is desalted in the ion-permeableand water-impermeable esophagus, and to a lesser extent inthe stomach, by passive diffusion of ions into the blood.The subsequent water absorption is considered to takeplace in the intestine following active absorption of Na?and Cl- [3, 5] In the posterior part of the intestine,including the rectum, osmolality of ingested seawater isreduced to isotonic or hypotonic levels [6], followed bypassive absorption of water through aquaporin, a waterchannel, located at the apical membrane of intestinal epi-thelial cells [6 9] For passive water absorption in theintestine, it is thus critical to remove ions from ingestedseawater and to reduce its osmolality

inges-Mechanisms of monovalent ion (Na?and Cl-) tion in the intestine have become increasingly clear Acurrently accepted model for NaCl absorption in intestinalepithelial cells consists of the cooperative actions of iontransport proteins, including basolaterally located Na?/K?-ATPase and apically located cation-chloride cotransporterssuch as Na?/K?/Cl-cotransporter (NKCC) and Na?/Cl-cotransporter (NCC) [10] Recently, Cutler and Cramb [5]

Department of Aquatic Bioscience, Graduate School of

Agricultural and Life Sciences, The University of Tokyo,

Bunkyo, Tokyo 113-8657, Japan

e-mail: mekuchi@marine.fs.a.u-tokyo.ac.jp

T Hatta

Japan International Research Center for Agricultural

Sciences, Tsukuba, Ibaraki 305-8686, Japan

Fish Sci (2010) 76:199–205

DOI 10.1007/s12562-009-0199-5

have shown that NKCC2b and NCCb are importantly

involved in active absorption of monovalent ions in the

intestine of European eel

Meanwhile, mechanism of removing divalent cations

(Ca2? and Mg2?) from ingested seawater is less

under-stood Marine teleosts are known to produce white feces,

which are often referred to as Ca precipitates [11] It has

been suggested that Ca precipitates are produced in the

intestine as a result of seawater adaptation, or

hypos-moregulation Evacuation of Ca precipitates is typically

observed in fasting Japanese eel acclimated to seawater

The Ca precipitates are soft to the touch and often coated

with mucus It has been known that Ca precipitates contain

calcium carbonate [12] In gulf toadfish Opsanus beta,

Walsh et al [11] found Ca precipitates to be calcium

kutnohorite, a carbonate mineral containing Ca, Mg, and

Mn It has been indicated that Ca precipitates are likely

associated with bicarbonate secretion and water absorption

[13,14]

The present study elucidates the relationship between

osmoregulation and Ca precipitate formation First, we

compared the production of Ca precipitates in eel

accli-mated to environmental waters with different salinities

Secondly, physicochemical characteristics of Ca

precipi-tates were analyzed by means of X-ray diffraction (XRD)

and X-ray fluorescence (XRF) analyses Finally, excretion

of Ca precipitates was examined in eel exposed to artificial

seawater with lowered Ca2?or Mg2? concentration

Materials and methods

Fish

Japanese eel cultivated in freshwater and weighing

approximately 200 g were obtained from a commercial

dealer in Hamamatsu, Shizuoka Prefecture, Japan Fish

were acclimated to laboratory conditions in a 1-t tank with

recirculating freshwater (dechlorinated Tokyo tap water) at

20°C For preparation of seawater-acclimated eel, fish were

first acclimated in 50% seawater for 2 days, and then

transferred to natural full-strength (100%) seawater The

fish were maintained in seawater for at least 2 weeks

before use for the following experiments To prepare 30%

and 60% seawater-acclimated eel, freshwater fish were

directly transferred to 30% and 60% seawater, and

main-tained in respective environmental waters for 2 weeks

Water temperature was maintained at 20°C, and the fish

were not fed during the acclimation period To prepare

dilute seawater, natural seawater (1050 mOsm/kg: Na?,

Collection of Ca precipitates for physicochemicalanalysis

For collection of Ca precipitates for physicochemicalanalyses, seawater-acclimated eel were placed in 60-lglass tanks containing seawater Half the water waschanged every other day to assure water quality Ca pre-cipitates excreted from the eel were periodically collected

by wide-mouthed pipette from the bottom of the tanks.Collected Ca precipitates were washed in distilled waterthree times (1 min each) and air-dried overnight at roomtemperature

X-ray diffraction analysisThe crystal structure of Ca precipitates was identified byXRD analysis (RAD-X; Rigaku, Tokyo, Japan) Everycrystal gives a characteristic diffraction spectrum with aset of diffraction peaks, whose position and intensitycorrespond to a specific crystalline structure The dif-fraction data were obtained by the powder diffractionmethod with Cu Ka radiation at scanning angle 2h from2° to 50° Then, obtained diffraction peaks were com-pared with those deposited in the XRD database main-tained by the International Centre for Diffraction Data byusing a computer program to determine the crystalstructure

X-ray fluorescence analysisQuantitative elemental analysis of Ca precipitates was per-formed by XRF spectrometry with an XRF analyzer (JSX-

3220, JEOL) XRF analysis allows accurate quantification;

however, it cannot detect low-atomic-number elements such

as C and O Ca precipitates were manually powdered with a

pestle and mortar, uniformed Ca precipitates were subjected

to XRF analysis, and the molar ratio between Ca and Mg in

Ca precipitates was determined

Excretion of Ca precipitates in eel exposed to artificial

seawater with lowered Ca2?or Mg2? concentration

Effects of Ca2? and Mg2? concentrations in

environ-mental seawater on Ca precipitate excretion were

exam-ined in seawater-acclimated eel exposed to artificial

seawater with lowered level of Ca2? or Mg2? Three

seawater-acclimated eels were placed in a 60-l glass tank

containing control artificial seawater and maintained for

3 days Then, the environmental water was changed

sequentially to low-Mg2? seawater (Mg concentration

one-tenth of control seawater), control seawater, low-Ca2?

seawater (Ca concentration one-tenth of control seawater),

and back to control seawater Fish were maintained in

respective waters for 3 days During the experiment, Ca

precipitates excreted from three eels were collected daily

by wide-mouthed pipette from the bottom of the tank On

the day of water change, the environmental water was

replaced shortly after collection of Ca precipitates The

experiment was conducted in triplicate Fish were not fed

during the experiment The dry weight of Ca precipitates

collected daily was measured as described above, and data

are expressed as mg Ca precipitates per kg fish per day

The molar ratio between Ca and Mg in Ca precipitates

was also determined by means of XRF analysis as

described above The Ca ratio was calculated as the molar

ratio of Ca to the sum of Ca and Mg, representing the

relative abundance of Ca

Control artificial seawater was prepared according to a

modified Herbst method: control seawater consisted of

460 mM NaCl, 55 mM MgCl2
6H2O, 10 mM CaCl2
2H2O,

10 mM KCl, and 3 mM NaHCO3, and osmolality was

adjusted to 1050 mOsm/kg H2O by adding NaCl For the

preparation of low-Ca2?seawater and low-Mg2?seawater,

CaCl2
2H2O and MgCl2
6H2O, respectively, were reduced

to one-tenth of control seawater levels, and osmolarity was

adjusted to the same level as control seawater

(1050 mOsm/kg H2O) with NaCl

Statistical analysis

Data are expressed as mean ± standard error of mean

(SEM) Significant differences at P \ 0.05 were

deter-mined by one-way analysis of variance (ANOVA),

fol-lowed by Tukey’s multiple-comparison test

seawater a Clumps of Ca precipitates can be seen in the intestine through the intestinal wall b Ca precipitates of amorphous shape are found along the intestine, which was longitudinally cut and opened.

Arrows and asterisks indicate Ca precipitates and the anus, tively Scale bars a and b, 10 mm; c, 2 mm

precipitates was significantly higher in 100% seawater than

in 30% and 60% seawater (Fig.2)

Crystal structure of Ca precipitates

X-ray diffraction analysis revealed the crystal structure of

Ca precipitates Figure3 shows a representative X-ray

diffraction pattern obtained from Ca precipitates A

con-spicuous diffraction peak was detected at 30°, showing that

Ca precipitates were composed of Mg-calcite However,

the diffraction peak characteristic of Mg-calcite was

somewhat broad This implied that Mg-calcite was not

completely but rather partially crystallized

Quantitative analysis by XRF

XRF spectrometry was utilized to quantify the elements

present in Ca precipitates (Fig.4a) Large peaks

repre-senting Ca Ka1 and Kb1 were detected at 3.690 and

4.012 keV, respectively, which were followed by that for

Mg Ka1at 1.253 keV Ca and Mg accounted for more than

99% of the elements that could be detected in Ca

precip-itates by XRF The calculated molar ratio between Ca and

Mg was approximately 7:2 (Fig.4b) Although small peaks

for P and S were also detectable, calculated abundance

ratios of P and S were 0.36% and 0.14%, respectively

Excretion of Ca precipitates in eel exposed to artificial

seawater with lowered Ca2?or Mg2? concentration

The total dry weight of Ca precipitates excreted from eel

exposed to control seawater for the first 3 days was about

85 mg/kg fish/day (Fig.5a) During subsequent exposure

to low-Mg2? seawater for 3 days, the weight of Ca

pre-cipitates was slightly reduced by about 20%, which was

followed by a recovery to the initial level after return tonormal seawater When eel were further exposed to low-

Ca2? seawater, Ca precipitates were drastically decreased

to less than 10% of the initial level In fish returned tocontrol seawater, the total weight of Ca precipitates wasincreased and showed a recovery trend, although theweight of Ca precipitates was about half the initial level

Japanese eel acclimated to 0% (freshwater, FW), 30%, 60%, and

100% seawater (SW) Values are mean ± SEM (n = 5) N.D., not

detectable Significant differences at P \ 0.05 are indicated by

different letters

30 40 20

θ

10 0 20 40

graphs show the conspicuous peaks; b Ca precipitates, c Mg-calcite Mg-calcite peaks were obtained from the XRD database maintained

by the International Centre for Diffraction Data

6000 8000 10000

Mg 22·61%

P 0·36%

S 0·14%

b

76.89%

Ca Ka

Ca

fluorescence (XRF) spectrometry b Elemental composition of Ca precipitates expressed on a molar basis Ca and Mg are the major elements of Ca precipitates, whereas small amounts of P and S are also present

Figure5b shows the changes in the molar ratio of Ca to the

sum of Ca and Mg (Ca ratio) in Ca precipitates The Ca

ratio was about 70% in Ca precipitates excreted from eel

exposed to control seawater The relative amount of Ca

was increased in eel transferred to low-Mg2?seawater, the

ratio being 96% Following returned to control seawater,

the Ca ratio showed a recovery trend Conversely, the Ca

ratio dropped to 59% in eel exposed subsequently to

low-Ca2?seawater The Ca ratio was then restored after return

to control seawater

Discussion

Ca precipitates have been observed in several marine and

seawater-adapted teleosts, such as southern flounder

Paralichthys lethostigma [12], rainbow trout

Oncorhyn-chus mykiss [13,15], European eel [16], gulf toadfish [11],

European flounder Platichthys flesus [14], and mefuguTakifugu obscures [17] In the present study, production of

Ca precipitates was examined in eel acclimated to 0%(freshwater), 30%, 60%, and 100% seawater Plasmaosmolality was kept within a physiological range in fishacclimated to the different salinities, suggesting successfuladaptation to those environments Sasai et al [18]observed a similar osmoregulatory response in Japanese eeltransferred from freshwater to seawater Ca precipitateswere not found in the intestine or rectum of fish acclimated

to freshwater, while at higher salinities fish produced Caprecipitates to a varying degree (Fig.2) The weight of Caprecipitates became greater with increasing environmentalsalinity, indicating that Ca precipitate formation is closelyrelated to seawater adaptation, or hyposmoregulation, ineel A similar salinity-dependent formation of Ca precipi-tates was also reported in rainbow trout [15]

Ca precipitates have been reported to be calcium bonate in southern flounder [12] Further analysis con-ducted by Humbert et al [16] revealed the crystal structure

car-of the calcium carbonate to be calcite In the present XRDanalysis, the crystal structure of Ca precipitates of Japaneseeel was found to be Mg-calcite and its amorphia On theother hand, Walsh et al [11] analyzed Ca precipitates ofgulf toadfish and found that the crystal structure of Caprecipitates was kutnohorite containing Mg with a smallamount of Mn The difference may be accounted for byspecies difference or difference in environmentalconditions

Our XRF analysis also revealed that the ratio of Ca to

Mg in Ca precipitates was approximately 7:2 (Fig.4).Considering the results of XRD and XRF analyses, Mgdetected in Ca precipitates can be attributed to that in Mg-calcite In addition to the major elements of Ca and Mg,small amounts of P and S were detected, with abundanceratios of 0.36% and 0.14%, respectively Small amounts of

P and S detected by XRF are likely derived from mucus orother contaminants It has been reported that an organicmatrix composed of glycoproteinaceous mucous fibersfunctions as the framework on which crystals grow [19] It

is therefore difficult to remove mucus even if Ca tates are washed in distilled water before elemental anal-ysis The abundance molar ratio of magnesium carbonate inMg-calcite is generally about 20%, and the other 80% iscalcium carbonate Mg-calcite has been found as a meta-stable mineral in modern shallow marine carbonate andbiominerals produced by algae and halophilic bacteria [20,21] Mg-calcite in marine sediments are easily dissolvedand recrystallized by fresh meteoric and/or marine water[22] During Mg-calcite formation, calcite is first formed,then Ca in calcite is replaced by Mg under high Mg2?concentration [23] In fact, Mg2? concentration in theintestine was about three times higher than that in seawater

control seawater (Cont 3) for 3 days each a Changes in dry weight, b

changes in molar ratio of Ca to the sum of Ca and Mg of excreted Ca

precipitates The experiment was conducted in triplicate Values are

mean ± SEM (n = 3) Significant differences at P \ 0.05 are

indicated by different letters

(50 mM) [2] The replacement ratio of Mg varies

depending on temperature and Mg2? concentration [23,

24] Walter and Morse have demonstrated that the

solu-bility of Mg-calcite is higher than that of calcite, and the

thermodynamic stability changes according to the amount

of Mg in Mg-calcite [25]

Results of exposure to low-Ca2?or low-Mg2?seawater

showed that ambient Ca2?and Mg2?concentrations directly

affected the formation of Ca precipitates and the Ca/Mg ratio

(Fig.5a, b) The weight of excreted Ca precipitates in

arti-ficial control seawater was equivalent to that in natural

seawater Ca precipitates were decreased by about 20% in

low-Mg2?seawater, compared with that in control seawater

On the other hand, the weight of Ca precipitates was

decreased to less than 10% of the initial level in low-Ca2?

seawater Decreases in excreted Ca precipitates can be

attributed to low-Mg2?and low-Ca2?conditions, in which

Ca2?and Mg2?concentrations were reduced to one-tenth of

control seawater levels, indicating that Ca and Mg in Ca

precipitates are derived from seawater Similarly, Shehadeh

and Gordon [15] reported, in rainbow trout, that Ca

precip-itates increased as Ca concentration in media increased The

reduction in Ca precipitate excretion was apparently greater

in low-Ca2?seawater than in low-Mg2?seawater This is in

accordance with our finding that the Ca content is about three

times larger than that of Mg in Ca precipitates produced and

excreted in eel exposed to normal artificial seawater This is

also supported by the formation process of Mg-calcite In the

process of Mg-calcite formation, calcite is firstly formed

Under low-Ca condition, it is difficult to produce calcite As

a result, low-Ca condition reduced the amount of Mg-calcite

formation Even under the low-Ca2?condition, although Ca

precipitates were remarkably reduced to less that 10% of the

initial value, the Ca content (59%) exceeded the Mg content

(41%) On the other hand, Mg seemed to be difficult to

replace with Ca under the low-Mg2?condition The Ca ratio

in Ca precipitates was increased to as high as 96% The

weight of Ca precipitates failed to recover fully in control

seawater following low-Ca2? seawater exposure This is

probably because it may require considerable time to restore

Ca precipitate formation and evacuation reduced in

low-Ca2?seawater to the initial state

Marine teleosts drink ambient seawater to compensate

for osmotic water loss It is essential for water absorption in

the intestine to reduce osmolality of ingested seawater to

isotonic or subisotonic levels, since the major driving force

for intestinal water absorption is an osmotic gradient

cre-ated between the intestinal and body fluids [6, 8] In the

present study, we clearly showed that Ca precipitate

for-mation in seawater-acclimated eel contributed greatly to

the removal of Ca2?, and of Mg2?to a lesser extent, from

ingested seawater, whereas Na? and Cl- are absorbed

along the gastrointestinal tract [1,3,12]

Whereas natural seawater typically contains mately 10 mM Ca2?and 50 mM Mg2?, Smith [2] reportedthat Ca2? and Mg2? concentrations of ingested seawaterbecame 8.2 and 167 mM, respectively, in the intestine ofAmerican eel Anguilla rostrata Those changes in ionconcentration are considered to be a combined result ofprecipitation of Ca2? and Mg2? as Ca precipitates,absorption of Na? and Cl-, and concomitant waterabsorption during the passage of ingested seawater alongthe gastrointestinal tract Preferential precipitation of Ca2?during the formation of Ca precipitates can account for therelatively high concentration of Mg2? in the intestinalfluid Although it had been reported that almost all Mg2?iningested seawater was passed out of the gastrointestinaltract [2], we confirmed that Mg2? as well as Ca2? wasprecipitated and excreted as Ca precipitates [26]

approxi-Most calcium carbonates are synthesized by a chemicalreaction between Ca2?and bicarbonate [27] Whereas thesource of divalent cations (Ca2? and Mg2?) for Ca pre-cipitates was revealed to be ambient seawater, bicarbonatehas been reported to be secreted from the intestine [26,28] Recently, a molecular mechanism of bicarbonatesecretion from intestinal epithelial cells has been proposedfor the formation of Ca precipitates in mefugu, consisting

of Cl-/HCO3-exchangers (slc26a6A and slc26a6B) in theapical membrane and Na?/HCO3- cotransporter (NBCe1)

in the basolateral membrane [17] In addition to theintestine, the pancreas is presumed to be another organresponsible for bicarbonate secretion In particular, bicar-bonate secretion from the pancreas has been known toplay an important role in buffering gastric acid Thesecretion of bicarbonate into the intestinal lumen could beimportantly linked not only to gastric-acid-neutralizingaction but also to Ca precipitate formation in seawater-adapted teleosts

References

1 Evans DH (1993) Osmotic and ionic regulation In: Evans DH (ed) The physiology of fishes CRC, Boca Raton, pp 315–341

2 Smith HW (1930) The absorption and excretion of water and salts

by marine teleosts Am J Physiol 93:480–505

3 Marshall WS, Grosell M (2006) Ion transport, osmoregulation, and acid–base balance In: Evans DH, Claiborne JB (eds) The physiology of fishes CRC, Boca Raton, pp 179–230

4 Hirano T, Mayer-Gostan N (1976) Eel esophagus as an regulatory organ Proc Natl Acad Sci USA 73:1348–1350

osmo-5 Cutler CP, Cramb G (2008) Differential expression of absorptive cation-chloride cotransporters in the intestinal and renal tissues of European eel (Anguilla anguilla) Comp Biochem Physiol B 149:63–73

6 Kim YK, Ideuchi H, Watanabe S, Park SI, Huh MD, Kaneko T (2008) Rectal water absorption in seawater-adapted Japanese eel Anguilla japonica Comp Biochem Physiol A 151:533–541

7 Cutler CP, Cramb G (2002) Branchial expression of an

aquapo-rin-3 (AQP-3) homologue is downregulated in the European eel

Anguilla anguilla following seawater acclimation J Exp Biol

205:2643–2651

8 Aoki M, Kaneko T, Katoh F, Hasegawa S, Tsutsui N, Aida K

(2003) Intestinal water absorption through aquaporin 1 expressed

in the apical membrane of mucosal epithelial cells in

seawater-adapted Japanese eel J Exp Biol 206:3495–3505

9 Mackenzie FT, Bischoff WB, Bishop FC, Loijens M (1983)

Magnesian calcites: low-temperature occurrence solubility and

solid-solution behavior In: Reeder RJ (ed) Carbonates Mineral

Soc Am, Michigan, pp 97–144

10 Delpire E, Mount DB (2002) Human and murine phenotypes

associated with defects in cation-chloride cotransport Annu Rev

Physiol 64:803–843

11 Walsh PJ, Blackwelder PK, Gill A, Danulat E, Mommsen TP

(1991) Carbonate deposits in marine fish intestines: a new source

of biomineralization Limnol Oceanogr 36:1227–1232

12 Hickman CP Jr (1968) Ingestion, intestinal absorption, and

elimination of seawater and salts in the southern flounder,

Paralichthys lethostigma Can J Zool 46:457–466

13 Wilson RW, Gilmour KM, Henry RP, Wood CM (1996)

Intes-tinal base excretion in the seawater-adapted rainbow trout: a role

in acid–base balance? J Exp Biol 199:2231–2343

14 Wilson RW, Grosell M (2003) Intestinal bicarbonate secretion in

marine teleost fish-source of bicarbonate, pH sensitivity, and

consequence for whole animal acid–base and divalent cation

homeostasis Biochim Biophys Acta 1618:163–193

15 Shedadeh ZH, Gordon MS (1969) The role of intestine salinity

adaptation of the rainbow trout, Salmo gairdneri Comp Biochem

Physiol 30:397–418

16 Humbert W, Kirsch R, Simonneaux V (1986) Is mucus involved

in biocrystallization? Study of the intestinal mucus of the

sea-water eel Anguilla anguilla L Cell Tissue Res 245:599–604

17 Kurita Y, Nakada T, Kato A, Doi H, Mistry AC, Chang MH,

Romero MF, Hirose S (2008) Identification of intestinal

bicarbonate transporters involved in formation of carbonate cipitates to stimulate water absorption in marine teleost fish Am J Physiol 294:R1402–R1412

pre-18 Sasai S, Kaneko T, Tsukamoto K (1998) Extrabranchial chloride cells in early life stages of the Japanese eel, Anguilla japonica Ichthyol Res 45:95–98

19 Humbert W, Voegel JC, Kirsch R, Simonneaux V (1989) Role of intestinal mucus in crystal biogenesis: an electron-microscopical, diffraction and X-ray microanalytical study Cell Tissue Res 255:575–583

20 Sa´nchez-Roma´n M, Rivadeneyra MA, Vasconcelos C, McKenzie

JA (2007) Biomineralization of carbonate and phosphate by moderately halophilic bacteria FEMS Microbiol Ecol 61:273– 284

21 Rivadeneyra MA, Delgado R, Pa´rraga J, Ramos-Cormenza A, Delgado G (2006) Precipitation of minerals by 22 species of moderately halophilic bacteria in artificial marine salts media: influence of salt concentration Folia Microbiol 51:445–453

22 Moore CH (1989) The classification and nature of carbonate porosity In: Moore CH (ed) Carbonate diagenesis and porosity Elsevier, Amsterdam, pp 21–40

23 Pilkey OH, Goodell HG (1963) Trace elements in recent mollusk shells Limnol Oeceanogr 8:137–148

24 Chave KE (1954) Aspects of the biogeochemistry of magnesium

1 Calcareous marine organisms J Geol 62:266–283

25 Walter LM, Morse JW (1984) Magnesian calcite stabilities: a reevaluation Geochim Cosmochim Acta 48:1059–1069

26 Wilson RW, Wilson JM, Grosell M (2002) Intestinal bicarbonate secretion by marine teleost fish—why and how? Biochim Bio- phys Acta 1566:182–193

27 Kitano Y, Kanamori N (1966) Synthesis of magnesian calcite at low temperatures and pressures Geochem J 1:1–10

28 Grosell M (2006) Intestinal anion exchange in marine fish osmoregulation J Exp Biol 209:2813–2827

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

The impact of nigorobuna crucian carp larvae/fry stocking

and rice-straw application on the community structure

of aquatic organisms in Japanese rice fields

Masatsugu Yamazaki•Taisuke Ohtsuka •Yasushi Kusuoka•

Masayoshi Maehata•Hiroyuki Obayashi• Kiyoyuki Imai•

Fujiyoshi Shibahara•Makoto Kimura

Received: 7 July 2009 / Accepted: 26 November 2009 / Published online: 20 January 2010

Ó The Japanese Society of Fisheries Science 2010

Abstract We examined the effects of predation by

Nigorobuna Carassius auratus grandoculis larvae and fry,

a crucian carp endemic to Lake Biwa, Japan, on the

com-munity structure of aquatic organisms in rice fields Six

experimental plots with three different rice-straw

treat-ments in the presence/absence of stocked larvae were

prepared In each plot, the number of aquatic organisms

ranging in size from 30 lm to 5 mm in the water, as well

as those from 63 lm to 5 mm in size in the surface

sedi-ments, was surveyed 6, 13, 20, 26, 34, and 41 days after the

onset of irrigation Three-day-old fish larvae were released

on day 10 Undigested organisms in the gut contents of the

larvae or fry were identified on days 20, 26, 34, and 41,

respectively Ten-day-old larvae mainly preyed on

Clado-cera, but the fry thereafter shifted to Diptera as their main

prey While Cladocera and Podocopida decreased in

fish-stocked plots, Euglenales and Halteriida became more

abundant there Top-down or bottom-up effects of fish

seemed to control these changes in community structure

Keywords Carassius auratus grandoculis
Cladocera
Indirect bottom-up effect
Rice field

Top-down trophic cascade

IntroductionRice-fish farming is practiced in quite a few countries,particularly in Asia, and its aquatic production is the mostreadily available, most reliable and cheapest source ofanimal protein and fatty acids for both farming and landlesshouseholds [1] Although carp production in rice fields inJapan reached 3400 t in 1943 due to war-time food pro-duction subsidies, it was not practiced on a significant scale

in the latter half of the twentieth century [2,3] Changes inJapanese food habits due to the high economic growthexperienced by Japan from the 1960s may be the reasonfor this

However, there has been a recent reawakening of fish farming used as a nursery for the production ofcrucian carp in Japan Carassius auratus grandoculis(Cyprinidae; nigorobuna in Japanese) is a crucian carpendemic to Lake Biwa, and it is used as an ingredient inFuna-zushi (lacto-fermented fish) that has been the spe-cialty food of Shiga prefecture, Japan In recent years,the catch of C a grandoculis has declined drasticallyfrom more than 100 ton year-1 in 1985 to 31 ton year-1

rice-in 2004 [4] The rice-increase rice-in carnivorous fishes of foreignorigin (Micropterus salmoides and Lepomis macrochi-rus), reed community decline, and artificial water levelregulation are possible causes of the marked decline of

C a grandoculis in Lake Biwa [5] With the aim ofreviving the stock of C.a grandoculis as a fisheriesresource, its fry are released into rice fields around LakeBiwa

National Agricultural Research Center, Tsukuba,

Ibaraki 305-8666, Japan

e-mail: yamasatu@affrc.go.jp

Lake Biwa Museum, Kusatsu, Shiga 525-0001, Japan

Agricultural Technology Promotion Center of Shiga Prefecture,

Azuchi, Shiga 521-1301, Japan

M Kimura

Graduate School of Bioagricultural Sciences,

Nagoya University, Nagoya, Aichi 464-8601, Japan

DOI 10.1007/s12562-009-0200-3

After producing the fish seeds of C a grandoculis at the

Shiga prefectural fisheries experiment station, the larvae of

this fish in the paddy rice fields usually grew into fry with

total lengths of [25 mm within 30 days of hatching [6]

From the viewpoint of survival rate, fry of this size are

considered appropriate to release into Lake Biwa [7]

Although the breeding of C a grandoculis in rice fields

only started in 2003, the number of the larvae grown in rice

fields (1.1 9 107 individuals) exceeded those grown in

other facilities (3.4 9 106 individuals) in 2005 [8] This

resulted in a massive run-down of crucian carp fry,

including C a grandoculis, from rice fields to Lake Biwa

in the middle of the rice cultivation period [9]

Because C a grandoculis mainly spawns in May [10],

controlling the prey population in rice fields during the first

half of the rice cultivation period (May and June) is

especially important for their stable initial growth Wild fry

of C a grandoculis mainly prey on Chydorus (Cladocera)

in the macrophyte zone of Lake Biwa, although the range

of foods available to them becomes wider as they grow

[11] They are also presumed to prey on Moina (Cladocera)

in rice fields since they are also abundant there during the

first half of the rice cultivation period [12–14]; however,

there are no concrete data on the feeding habits of

C a grandoculis in rice fields

In the present study, we clarify the feeding habits of

larvae/fry of C a grandoculis and their impact on the

plankton community in rice fields To do this, we regularly

surveyed the gut contents of C a grandoculis and the

community structure of aquatic organisms in the

experi-mental rice fields in the presence/absence of larvae stocking

In addition, the effect of rice-straw plowing on the

com-munity was examined collaterally In our previous study, the

application of rice-straw compost to rice fields increased the

population of Mona during the first half of the rice

culti-vation period [15] This suggests that rice-straw plowing,

which is now widely practiced in rice fields around Lake

Biwa, can improve the supply of prey for C a grandoculis

Materials and methodsOutline of field study sitesThe study site was located at the Shiga PrefecturalAgricultural Experiment Station in Azuchi town, ShigaPrefecture, Japan (35°1002500N, 136°705200E) Six experi-mental plots were established in a paddy field there(Fig.1), each covering 80 m2 The six treatments inclu-ded all combinations of rice-straw plowing in eitherautumn or spring or no such plowing, and the presence orabsence of stocked fish larvae There was only one plotfor each combination of conditions; the plot with no fishand no rice-straw plowing served as the control The soil

in the field was Typic Endoaquept, i.e., fine-textured andgley (low redox potential) The total carbon and totalnitrogen (TN) contents of the soil were 22.4 and2.08 g kg-1, respectively In each plot, rice Oryza sativa

L cv kinuhikari was cultivated by conventional methods,

as shown in Table1 The depth of the irrigation waterwas kept at ca 5 cm, although at collection times itranged between 3 and 7 cm We released three-day-oldlarvae into the three fish-stocking plots 13 days after theonset of irrigation (day 13), at a stocking density of

20 individuals m-2.Measurement and determination of environmentalparameters

We set up a temperature logger (Tinytag plus 2: GeminiData Loggers, UK) in each plot 8 days after flooding andretrieved it at the mid-season draining; the measurementinterval was 10 min The concentrations of TN and totalphosphorus (TP) in paddy water that had been passedthrough a 60-lm sieve were also determined for each plot

on days 6, 14, 21, 27, 34, and 41 using ion column matography (Prominence HIC-NS: Shimadzu, Kyoto,Japan)

chro-N

10 m s(+) a(-)

10 m n(+)

10 m s(-)

10 m a(+)

)

10 m m

n(-)

10 m

8 m

plowing; a, the plot where rice straw was plowed in autumn last year;

s, the plot where rice straw was plowed in spring this year; addition

sign, the plots with fish stocking; minus sign, the plots without fish stocking; filled circles, sampling points; open squares, inlets of irrigation water; filled squares, outlets of floodwater

Characterization of aquatic organism communities

To determine the population density of aquatic organisms

in each plot, we inserted an acrylic cylinder (3.9 cm in

inner diameter, 20 cm in height) into the soil and

sep-arately collected the water column supernatant and the

water–sediment interface using a 20 ml pipet For the

latter, after removing the water column sample, 10 ml of

Chalkley’s solution (NaCl 0.1 g, KCl 0.004 g, CaCl2

0.006 g, distilled water 1000 ml [16]) was poured into

the cylinder and the resulting turbid fluid was collected;

this procedure was repeated three times for each sample

The water-column samples were stored in an insulated

box at approximately 20°C The water–sediment

inter-face samples were poured through a 63 lm sieve, and

the sieve residue was fixed in 10% neutral buffered

formalin and stored as above in an insulated box On

each occasion, samples were collected from the same

three points in each plot, as shown in Fig.1 Sampling

was conducted between 9:00 a.m and 10:30 a.m on

days 6, 13, 20, 26, 34, and 41 It took approximately 2 h

to deliver the living and preserved samples to the

laboratory

In the laboratory, aquatic organisms [30 lm in size

were picked out Identification and classification was done

following Mizuno [17], Lee et al [18], and Canter-Lund

and Lund [19] for Cyanophyta, Heterokontophyta, and

Chlorophyta; Lee et al [18] and Mizuno and Takahashi

[20] for Sarcomastigophora, some species of Ciliophora,

and Rotifera; Inoki [21] and Patterson [22] for some

spe-cies of Ciliophora; Uchida and Uchida [23] for

Gastrotri-cha, Platyhelminthes, Nematoda, and Oligochaeta; Uchida

and Uchida [24] for Archipora, Acari, and Mollusca; andUeno [25] for Insecta We further identified the Cladoceraand Podocopida to the family or genus level, identifyingthe dominant species based on Mizuno and Takahashi [20].Because it was difficult to identify Turbellaria and Nema-toda even to the order level, no detailed taxonomicbreakdown for these groups was attempted Division,phylum, class, and order names were mostly adopted fromthe Iwanami Encyclopedia of Biology, 4th edition [26], but

we referred to Lee et al [27] for the order-level cation of ciliates

classifi-The population density of each taxonomical group in thewater column and at the water–sediment interface wascalculated as the total numbers of individuals observed atthree collecting points combined divided by the triple ofthe inner area of the acrylic cylinder (=3.584 9 10-3m2).Analysis of fish gut contents

After sampling the aquatic organisms, three fish fry werecollected from each plot on each sampling occasion using asmall scoop net (17 cm 9 15 cm) Collecting these fishcaused turbidity in the water, so we also used the samescoop net to stir the water in the unstocked plots withoutfish as well The fish fry were immediately cooled with iceuntil we returned to the laboratory

The gut contents were stored in 70% ethanol Weidentified only undigested organisms in the gut contents

To evaluate food selectivity, we used Manly’s b, asgiven by

Pi¼PIbiAih¼1bhAh

;

where Pi is the proportion of individuals belonging to theith item in the fish gut contents, biis the selectivity indexfor the ith item, Ai is the number of available preyindividuals belonging to the ith item in the ambientpopulation, and I is the number of items Values of bithat are greater than 1/I indicate positive selection [28].All individuals in both the water column and water–sediment interface samples were regarded as beingavailable as prey because the fry can eat benthos as well

as plankton The selectivity index was estimated for fourabundant groups of arthropods: Cladocera, Cyclopoida,Podocopida, and Diptera (excluding all adults as well asNematocera pupae, which were not found in the ambientsamples) These groups usually comprised the main prey

in terms of quantity Many individuals from other nomic groups in the water–sediment interface sampleswere rendered unusable when the sieve and formalinwere used

Statistical analyses

We estimated Manly’s selectivity index b for each

taxo-nomic group on each sampling day as the maximum

like-lihood estimator given by

^

bijk¼PInijk=aij

h¼1nhjk=ahj;

where i, j, and k, respectively, denote the categories of food

item, experimental plot, and individual fish, nijk is the

observed number of individuals of class ijk in the gut

con-tents, and aijis the observed number of individuals of class ij

in the ambient sample [28] An approximate 95% confidence

interval for b for each food item is given by a

t-approxi-mation of the arcsine-square-root-transformed ^bik[29]

To test differences in (a) nutrient concentration, (b) the

number of taxonomic groups of aquatic organisms, and (c)

the abundances of the common groups (i.e., those present

in at least one-third of the tested samples) among plots after

the release of the fish larvae, a three-way analysis of

var-iance (ANOVA) with two between-subject factors and one

within-subject factor (repeated measures) without

replica-tion was conducted using SPSS 15.0 (SPSS Inc., Chicago,

IL, USA) The analyses were based on the following model

that explains variations in dependent variables after the

introduction of larvae:

Yijk¼ m þ Aiþ Bjþ Ai Bjþ Ckþ Ai Ck

þ Bj Ckþ eijk;

where Yijkdenotes the dependent variable, m is the mean, Ai, Bj

and Ck, respectively, are the effects of fish stocking (presence/

absence of fish), rice straw treatments (in spring, in autumn,

and none) and sampling days (days 20–21, 26–27, 34, and 41),

and Ai9 Bj, Ai9 Ck and Bj9 Ck are their interactions

Between-subject effects, Aiand Bj, were tested by Ai9 Bj

Within-subject effects, Ck, Ai9 Ckand Bj9 Ck, were tested

by eijk, with the Huynh–Feldt adjustment The dependent

variables, except for pH and the number of taxonomic groups,

were converted into common logarithms before the test

because we assumed that each factor has a multiplicative

(log-linear) effect One was added to the number of individuals of

each taxonomical group before taking the logarithm because it

often became zero Because the sample size was small in the

present study, we did not apply any multiple-test adjustments

to avoid inflating the Type II error [30]

Results

Temporal variation in water quality

The water temperature from day 2 to day 41 was 37.7°C at

the maximum, 22.3°C on average, and 8.9°C at the

minimum The differences in mean temperature betweenplots were \1°C The daily mean and minimum tempera-tures both showed significant uptrends (Page’s trend test,

X2= 33.01 and 106.23, p \ 0.001), while the maximumtemperature did not (X2= 0.24, p = 0.624)

Total nitrogen drastically decreased by day 21, but itfluctuated at a somewhat higher level after day 27 (Fig.2).After day 21, it was significantly different between days(F3,6= 4.77, p = 0.050), but the effects of both fishstocking and the manner of rice-straw plowing were notsignificant (F1,2= 11.95, p = 0.074, and F2,2= 10.83,

p = 0.085, respectively) The interactions between daysand fish, and between days and manner of rice strawplowing, were also not significant (F3,6= 1.23, p = 0.378,and F3,6= 0.61, p = 0.717, respectively)

Total phosphorus rapidly decreased by day 14 After day

21, it tended to decrease in the plots without fish, butincrease in the fish-stocked plots (Fig.2) It was not sig-nificantly different between days (F3,6= 0.95, p = 0.475),but both the effect of fish and the interaction betweendays and fish were significant (F1,2= 23.99, p = 0.039,and F3,6= 5.15, p = 0.043, respectively) In contrast, theeffect of the manner of rice-straw plowing and the

the samples in the plots with and without fish, respectively Circles, triangles, and squares indicate the samples from the plot n, the plot a, and the plot s, respectively Fish larvae were released on day 13

interaction between this and days were not significant

(F2,2= 3.27, p = 0.234, and F3,6= 0.35, p = 0.885,

respectively)

Gut contents and food selection of larvae/fry

Representatives of at least 15 orders of aquatic organisms

were found in the examined fish guts Taking size into

account, Cladocera, Cyclopoida, Podocopida, and Diptera

(mainly Chironomidae larvae) appeared to be the main

food items Ten-day-old fish larvae 11.4 mm in mean total

length always had Cladocera, mainly Moina and Bosmina,

as the dominant component of their gut contents

Subse-quently, the main food item of 16- and 24-day-old fry (16.3

and 24.0 mm in mean total length, respectively) became

variable, differing among individuals The 31-day-old fry

29.7 mm in mean total length usually had Chironomidae

larvae as the main component of their gut contents

Ploi-mida (Rotifera: mainly benthic species of Lecane) and

benthic algae were often numerous in the guts of fry over

16 days old (Table2)

The confidence interval for Manly’s b was usually wide,

indicating variability in food selectivity among individuals

Podocopida were always negatively selected throughout

the experiment Cladocera were positively selected by

10-day-old larvae, but less so as the fry grew, and they

were not eaten by 31-day-old fry Similarly, selectivity for

Cyclopoida appeared to decrease as the fish grew, althoughthey were in general usually less strongly selected thanCladocera anyway On the contrary, chironomid larvae andpupae (Diptera), which were selected against by 10-day-oldfish larvae, gained in selectivity as the fry grew, and werefinally strongly selected for by 31-day-old fry (Fig.3)

found in the guts of Carassius

auratus grandoculis larvae

and fry

The minimum–mean–maximum

numbers of individuals eaten by

a fish are shown for each

taxonomic group on four

sampling days Nine larvae or

fry were examined on each

occasion

Spirogyra (Zygnematales) and

Oedogonium (Oedogoniales)

Mean total length

of fish (mm)

10 16 24 31 10 16 24 31 10 16 24 31 10 16 24 31

Fish age in days 0.00

0.25 0.50 0.75 1.00

confidence interval for each item on each sampling day All adults and Nematocera pupa in the fish guts were excluded from Diptera because they were not found in the ambient samples

Effect of fish stocking on the community structure

of aquatic organisms

Twenty-nine taxonomic groups of aquatic organisms were

indentified, mainly at the order level, in the water column

samples (Table3) Among the eight common taxonomic

groups besides arthropods in the water-column samples

(Fig.4), Euglenales and Halteriida were significantly more

abundant in plots with fish than in those without fish

(p \ 0.05; Table4) No group showed any significant

difference with respect to the manner of rice-straw plowing

(p [ 0.05; Table4) The orders Euglenales, Halteriida,

Prorodontida, and Ploimida showed significant temporal

changes (p \ 0.05; Table4) The interaction between days

and fish was significant in Euglenales and Prorodontida(p \ 0.05; Table4); after day 26, these groups increased inplots with fish, but not in plots without fish

Only eight taxonomic groups (Turbellaria, Nematoda,Haplotaxida, Cladocera, Cyclopoida, Coleoptera, Diptera)were also found in the water–sediment interface samples.This was due to the fact that many taxonomic groups wereexcluded from capture and/or good preservation due to theuse of a 63 lm sieve and 10% neutral buffered formalin.Cladocera and Podocopida were also significantly moreabundant in the plots without fish than in those with fish(p \ 0.05; Table4), but Cyclopoida and Diptera were not(p [ 0.05; Table4) Also, Cladocera and Podocopidashowed significant temporal changes (p \ 0.05; Table4)

organisms found in the present

experimental rice paddy

Underlined names denote the

taxonomic groups whose

population densities were

surveyed in the water column

samples Groups with asterisks

were observed in fish gut

Harpacticoida were not

observed in the water column

samples

Micrasterias

Spirostomum

12 Stichotrichina (subclass)

Unidentified

Harpacticoida*

Podocopida tended to be more abundant in plots without

fish only on days 34 and 41 After day 20, Cyclopoida

tended to increase in the fish-stocked plots, whereas they

decreased in plots without fish Thus, the interaction

between days and fish was significant for the Podocopida

and Cyclopoida (p \ 0.05; Table4) Although the

abun-dances of any four taxonomical groups were not

signifi-cantly affected by the manner of rice-straw plowing

(p [ 0.05; Table4), temporal changes in the population

density of Diptera depended on the manner of rice-straw

plowing (p \ 0.05; Table4; Fig.5)

Six groups of Cladocera (Bosmina, Ceriodaphnia,

Daphnia, Moina, Chydoridae, and Macrothricidae) were

not found in the fish-stocked plots on days 26, 34, and 41,

when they were not particularly rare in the plots without

fish (Fig.5) Ilyocypris (Podocopida) was not extirpated,but was significantly less numerous in the fish-stockedplots than in those without fish (p \ 0.05) The populationdensities of Scapholebelis (Cladocera) and Cyprididae(Podocopida) were not significantly affected by fishstocking (p [ 0.05; Table4) Rice-straw plowing both inautumn and spring resulted in a significantly higher pop-ulation density of Moina (Student–Newman–Keuls posthoc test, a = 0.05; Fig.5)

Biodiversity of aquatic organisms in paddy watersAlthough the number of taxonomic groups present was notsignificantly different between plots with and without fish(F1,2= 16.00, p = 0.057), a significant interaction between

the eight common taxonomic

groups besides arthropods

observed in the respective

floodwater column samples.

Closed and open symbols

indicate the samples in the plots

with and without fish,

respectively Circles, triangles,

and squares indicate the

samples from the plot n, the plot

a, and the plot s, respectively.

Fish larvae were released

on day 13

days and fish (F3,6= 7.943, p = 0.016) resulted from the

marked decrease in taxa number in the plot without fish after

day 26, when there was scarcely any decrease in the

fish-stocked plots (Fig.6)

Discussion

Crucian carp Carassius auratus grandoculis fry changed

their main prey from Cladocera to benthic animals

(par-ticularly Diptera) as they grew in rice fields This change

in food habit with growth has been also reported in other

water environments [11, 31] The selectivity for

Clado-cera by the larvae on day 20 and the marked decline of

Cladocera in the fish-stocked plots thereafter indicate that

the larvae had mostly depleted Cladocera in both the

water column and the water–sediment interface We

concluded that the C a grandoculis fry were forced to

change their food habits by the change in aquatic

organism community structure that they had caused

Kanao et al [6] have already presumed this causal

relationship

The dominant genera of Cladocera in the gut contents of

C a grandoculis were Moina and Bosmina As observed inour previous study [14], the application of rice straw sig-nificantly increased the abundance of Moina These resultssuggest that rice-straw plowing provides larvae with astable prey population However, we were not able tocontrol the number of Bosmina The future challenge is toincrease the abundance of a wide variety of Cladocera.Chironomidae, which was a dominant family of Diptera,was not reduced in fish-stocked plots irrespective of thestrong selective predation by the fry after day 34 This waspossibly due to continuous oviposition by adult Chiro-nomidae, which were very abundant around the rice fields

If this is correct, surrounding paddy fields or similar waterenvironments may favor the stable supply of chironomidlarva as a food for fish in paddy fields

Interestingly, the stocking of crucian carp increased thepopulations of the three groups of protozoa (Euglenales,Halteriiida and Prorodontida), whose dominant generawere Euglena, Halteria, and Coleps, respectively In gen-eral, fish can increase protozoa via two different pathways:nutrient enrichment caused by fish-induced sediment

replication performed to analyze the effects of (A) fish stocking, (B)

rice-straw treatments, (C) sampling days, and (A 9 C, B 9 C)

interactions on the population density of common groups of aquatic organisms after the release of larvae of Carassius auratus grandoculis into the experimental plots

resuspension [32], and by reducing the zooplankton ulation that feeds on protozoa [33] In the present study, thehigher TP in fish-stocked plots (Fig.2) supports theoccurrence of the former pathway Such phosphorusenrichment can enhance the growth of bacteria andEuglena, which can utilize phosphate directly [34], andthen the growth of Halteria, which feeds on bacteria [35].This phenomenon is called an indirect bottom-up effectinduced by top-down control of the food web [36] On theother hand, we guessed that Coleps could increase its uti-lization of undecomposed residue in fish feces; the genus isknown to feed on dead or dying animal tissue [22], and wealso observed that it swarmed around dead bodies ofCladocera and Cyclopoida Meanwhile, the decline ofMoina may support the latter notion Moina macrocopa,which lives in Japanese rice fields [20], feeds efficiently on

pop-a wide rpop-ange of sizes of plpop-ankton ppop-article, from the bpop-ac-terium Escherichia coli (2 lm long and 0.5 lm wide) tothe ciliate Tetrahymena pyriformis (50 lm long and 30 lmwide; [37, 38]) As Euglena, Halteria, and Coleps aremainly below 50 lm in cross-sectional diameter [20],Moina seemed to be able to feed on these protozoa That is,

Days after onset of paddy flooding

arthropod groups Sporadic

groups observed in less than

four samples were omitted.

Closed and open symbols

indicate the samples in the plots

with and without fish,

respectively Circles, triangles,

and squares indicate the

samples from the plot n, the

plot a, and the plot s,

respectively Fish larvae

were released on day 13

Days after onset of paddy flooding

observed in the respective floodwater column samples Closed and

open symbols indicate the samples in the plots with and without fish,

respectively Circles, triangles, and squares indicate the samples from

the plot n, the plot a, and the plot s, respectively Fish larvae were

released on day 13

C a grandoculis (the highest trophic level) controlled the

abundance of Moina (the next trophic level), and then the

reduction of Moina may have caused Euglena, Halteria

and Coleps (the lowest trophic level) to prosper This

phenomenon is called a top-down trophic cascade [39]

In many countries, crucian carp Carassius auratus

fingerlings are raised to an edible size though the rice

cultivation period [1] This may also also possible for

C a grandoculis However, culturing the larvae in rice

fields is only part of the fish recovery project in Lake

Biwa; at the same time, the conservation and expansion of

reed beds where the larvae grow up and the extermination

of carnivorous fishes of foreign origin that feed on the

larvae are being promoted in Lake Biwa [4] Recently,

young people from the local community around Lake

Biwa have begun to help to release the larvae into Lake

Biwa, and so their awareness of the conservation of the

water environment surrounding Lake Biwa has increased

Lake Biwa has a number of aquatic organisms that are not

found elsewhere, such as groups of the catfish Silurus

biwaensis, the pond snail Heterogen longispira and the

planarian Bdellocephala annandalei The fish recovery

project will also lead to biodiversity conservation in Lake

Biwa

of Environmental and Information Sciences, Yokkaichi University,

for his advice regarding the identification of the phytoplankton,

Dr Mark J Grygier of Lake Biwa Museum for his advice on English

composition, and the late Prof Tadashiro Koyama, Prof Arata

Katayama, Dr Susumu Asakawa, and Dr Jun Murase, all of Nagoya

University, and Dr Koki Toyota of the Tokyo University of

Agri-culture and Technology, for their valuable comments This work was

partially supported by Lake Biwa Museum Comprehensive Research

Program 07-01 and Grant-in-Aid for Scientific Research (C)

21580243.

References

1 Halwart M, Gupta MV (eds) (2004) Culture of fish in rice fields.

WorldFish Center, Penang

2 Kuronuma K (1980) Carp culture in Japanese rice fields In:

Pullin RSV, Shehadeh ZH (eds) Proceedings of the

ICLARM-SEARCA conference on integrated agriculture aquaculture

farming systems, Manila, Philippines, 6–9 Aug 1979, pp 167–174

3 Pillay TVR (1990) Aquaculture principles and practices Fishing

New Books, Oxford

4 Shiga Prefecture (2006) Shiga Prefecture recovery planning of

stock of Nigorobuna Shiga Prefecture Fisheries Management

Division (in Japanese)

5 Fujioka Y (2006) Nigorobuna In: Shiga Prefecture

Bioassess-ment Committee (ed) Wildlife to care about in Shiga

Prefec-ture—Shiga Prefecture Red Data Book Sunrise Publishing,

Shiga, p 468 (in Japanese)

6 Kanao S, Ohtsuka T, Maehata M, Suzuki N, Sawada H (2009)

Effectiveness of rice fields as an initial growth environment of

larval and juvenile nigorobuna Carassius auratus grandoculis.

Nippon Suisan Gakkaishi 75:191–197 (in Japanese)

7 Fujiwara K, Usuki T, Ujiie S, Mizutani E (1994) Survival rates of nigorobuna larva/fry in natural water area depending on devel- opment stages Accomplishment Rep Shiga Prefect Fish Exp Stn 1992:24–25 (in Japanese)

8 Shiga Prefecture Fisheries Promotion Association (2007) Outline

of seed production and release project results from 1983 to 2005 Shiga Prefecture Fisheries Promotion Association, Kusatsu (in Japanese)

9 Nishikawa S (2006) The activities of Midori Net Aisei In: The Organizing Committee of the Symposium on Wetland Restora- tion 2006 (ed) Proceedings of the international symposium on wetland restoration 2006—restoration and wise use of wetlands Shiga Prefecture Government, Otsu, pp 135–140 (in Japanese)

10 Nakamura M (1969) Cyprinid fishes of Japan—studies on the life history of cyprinid fishes of Japan Research Institute for Natural Resources, Tokyo (in Japanese)

11 Hirai K (1972) Ecological studies on fry and juvenile of fishes at aquatic plant areas in a bay of Lake Biwa III Relationship of the food habits to the habitat of nigorobuna (Carassius carassius grandoculis) larvae Jpn J Ecol 22:69–93 (in Japanese)

12 Yamamoto K (1987) Succession of plankton in the rice fields Suisanzoshoku 34:261–268 (in Japanese)

13 Kuwabara R (1999) Dynamic of water quality and planktonic community in a paddy of northeastern Hokkaido along with the growth of rice plant In: Proceedings of international seminar on development of agribusiness and its impact on agricultural pro- duction in south east Asia (DABIA III), Tokyo, Japan, 11–16 Nov

1996, pp 434–442

14 Yamazaki M, Hamada Y, Kamimoto N, Momii T, Aiba Y, Yasuda N, Mizuno S, Yoshida S, Kimura M (2003) Changes in the community structure of aquatic organisms after midseason drainage in the floodwater of Japanese fields Soil Sci Plant Nutr 49:125–135

15 Yamazaki M, Hamada Y, Ibuka T, Momii T, Kimura M (2001) Seasonal variations in the community structure of aquatic organisms in a paddy field under a long-term fertilizer trial Soil Sci Plant Nutr 47:587–599

16 Chalkley HW (1930) Stock cultures of amoeba Science 71:442

17 Mizuno T (1964) Illustrations of the freshwater plankton of Japan Hoikusha, Osaka (in Japanese)

18 Lee JJ, Hutner SH, Bovee EC (eds) (1985) An illustrated guide to the protozoa Society of Protozoologists, Lawrence

19 Canter-Lund HJ, Lund WG (1995) Freshwater algae Their microscopic world explored Biopress Ltd., Bristol

20 Mizuno T, Takahashi E (eds) (1991) An illustrated guide to freshwater zooplankton in Japan Tokai University Press, Tokyo (in Japanese)

21 Inoki S (1981) An illustrated protozoa book Kodansha Scientific, Tokyo (in Japanese)

22 Patterson DJ (1996) Free-living freshwater protozoa: a color guide University of New South Wales Press, Sydney

23 Uchida K, Uchida T (1965) New illustrated encyclopedia of the fauna of Japan (I) Hokuryukan, Tokyo (in Japanese)

24 Uchida K, Uchida T (1965) New illustrated encyclopedia of the fauna of Japan (II) Hokuryukan, Tokyo (in Japanese)

25 Ueno K (1986) The late Tamiji Kawamura freshwater biology of Japan Hokuryukan, Tokyo (in Japanese)

26 Iwatsuki K (1996) The classified table of organisms In: Yasugi

R, Ozeki H, Furutani M, Hidaka T (eds) Iwanami encyclopedia of biology, 4th edn Iwanami Shoten, Tokyo, pp 1535–1617 (in Japanese)

27 Lee JJ, Leedale GF, Bradbury P (eds) (2000) An Illustrated guide

to the protozoa, 2nd edn Society of Protozoologists, Lawrence, Kansas

28 Manly BFJ (1974) A model for certain types of selection experiments Biometrics 30:281–294

29 Sokal RR, Rohlf FJ (1995) Biometry, 3rd edn Freeman, New

York

30 Moran MD (2003) Arguments for rejecting the sequential

Bon-ferroni in ecological studies Oikos 100:403–405

31 Tomoda Y (1965) Morphological studies on the development of

the crucian carp in Lake Biwa I Ontogenetic development of

nigoro-buna, and some remarks on the differentiation of

gengoro-buna and nigoro-gengoro-buna Bull Osaka Mus Nat Hist 18:3–30 (in

Japanese)

32 Breukelaar A, Lammens E, Breteler J, Tatrai I (1994) Effects of

benthivorous bream (Abramis brama) and carp (Cyprinus carpio)

on sediment resuspension and concentrations of nutrients and

chlorophyll-a Freshw Biol 32:113–121

33 Attayde JL, Hansson LA (2001) The relative importance of fish

predation and excretion effects on planktonic communities.

Limnol Oceanogr 46:1001–1012

34 Jansson M (1988) Phosphate uptake and utilization by bacteria

and algae Hydrobiologia 170:177–189

35 Bick H (1972) Ciliated protozoa: an illustrated guide to the species used as biological indicators in freshwater biology World Health Organization, Geneva

36 Benndorf J (1995) Possibilities and limits for controlling phication by biomanipulation Int Rev Gesamten Hydrobiol 80:519–534

eutro-37 Stuart CA, McPherson M, Cooper HJ (1931) Studies on riologically sterile Moina macrocopa and their food require- ments Physiol Zool 4:87–100

bacte-38 Kumar R, Hwang JS (2008) Ontogenetic shifts in the ability of the Cladoceran, Moina macrocopa Straus and Ceriodaphnia cornuta Sars to utilize ciliated protists as food source Int Rev Hydrobiol 93:284–296

39 Hrba´cˇek J, Dvorˇakova M, Korˇı´nek V, Procha´zko´va L (1961) Demonstration of the effect of the fish stock on the species composition of zooplankton and the intensity of metabolism of the whole plankton association Verh Int Ver Theor Gewandte Limnol 14:192–195