Journal of Fish Biology (2009) 75, 1459–1472

doi:10.1111/j.1095-8649.2009.02393.x, available online at www.interscience.wiley.com

Elevated ability to compete for limited food resourcesby ‘all-fish’ growth hormone transgenic common carp

Cyprinus carpio

M. Duan*†, T. Zhang*‡, W. Hu*, L. F. Sundstrom§, Y. Wang*,Z. Li *‡ and Z. Zhu*

*State Key Laboratory of Freshwater Ecology and Biotechnology, Institute of Hydrobiology,Chinese Academy of Sciences, Wuhan 430072, China, †Graduate University of Chinese

Academy of Sciences, Beijing 100049, China and §Department of Fisheries and Oceans,University of British Columbia, Centre for Aquaculture and Environmental Research, 4160

Marine Drive, West Vancouver, British Colombia V7V 1N6, Canada

(Received 6 December 2008, Accepted 2 July 2009)

Food consumption, number of movements and feeding hierarchy of juvenile transgenic common carpCyprinus carpio and their size-matched non-transgenic conspecifics were measured under conditionsof limited food supply. Transgenic fish exhibited 73·3% more movements as well as a higher feedingorder, and consumed 1·86 times as many food pellets as their non-transgenic counterparts. After the10 day experiment, transgenic C. carpio had still not realized their higher growth potential, whichmay be partly explained by the higher frequency of movements of transgenics and the ‘sneaky’feeding strategy used by the non-transgenics. The results indicate that these transgenic fish possessan elevated ability to compete for limited food resources, which could be advantageous after anescape into the wild. It may be that other factors in the natural environment (i.e. predation risk andfood distribution), however, would offset this advantage. Thus, these results need to be assessedwith caution. © 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles

Key words: competition; foraging; growth hormone (GH); transgene.

INTRODUCTION

The use of growth hormone (GH) treatment (McLean & Donaldson, 1993) and alteredGH gene expression in fishes by transgenesis (Zhu et al ., 1985; Gross et al ., 1992;Devlin et al ., 1994) have been explored as a way to increase aquaculture productionand efficiency. Fast-growing fishes also have been considered as the best candidatesfor the marketing of transgenic animals for human consumption (Zbikowska, 2003).Released or escaped fast-growing transgenic fishes, however, pose an ecological riskthat should be of concern (Devlin et al ., 2006). Assessment of the environmentaleffects of fast-growing transgenic fishes is therefore urgently needed (Kapuscinskiet al ., 2007).

‡Authors to whom correspondence should be addressed. Tel.: +86 27 6878 0369; fax: +86 27 6878 0063;email: tlzhang@ihb.ac.cn; zhongjie@ihb.ac.cn

1459© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles

1460 M . D UA N E T A L .

Under natural conditions, acquiring large size through rapid growth can providefitness advantages (Arendt, 1997). Rapid growth reduces vulnerability to gape-limitedpredators, allows earlier exploitation of novel resources (Werner & Gilliam, 1984)and can result in sexual maturity at a younger age (Alm, 1959). The increase ingrowth rate observed after GH treatment or transgenesis comes primarily from anelevated food intake acting through increased appetite (Johnsson & Bjornsson, 1994)and foraging activity (Jonsson et al ., 1996), leading to higher competitive ability, atleast in salmonids (Devlin et al ., 1999). GH-transgenic fishes also appear to be betterat using low-quality food and have higher feed-conversion efficiency (Fu et al ., 1998;Venugopal et al ., 2004; Raven et al ., 2006). GH-transgenic fishes, however, do notalways benefit from elevated GH production. In coho salmon Oncorhynchus kisutch(Walbaum), transgenic individuals were prone to spending more time and to investingmore energy on unprofitable prey due to reduced discrimination (Sundstrom et al .,2004a). They experienced lower survival in the presence of predators (Sundstromet al ., 2004b, 2005), and may experience reduced disease resistance (Jhingan et al .,2003) and impaired swimming ability (Farrell et al ., 1997; Li et al ., 2007). Distincteffects observed among species and strains under different conditions reveal theneed for risk assessments to be performed on a case-by-case basis (Devlin et al .,2001, 2006).

Many natural water bodies are diverse and complex, in which the resources areoften unevenly distributed. Individuals of the same species require similar resourcesand therefore compete when these resources are in limited supply (Begon et al .,1996). The prevalence of dominant–subordinate relationships in animal societiessuggests that there is an adaptive significance to achieving a high rank when in agroup (Huntingford & Turner, 1987). Feeding rank within a group has been usedas an indirect measure of social rank in a number of studies involving salmonidsflatfishes and tilapias (McCarthy et al ., 1992, 1999; Carter et al ., 1994; Jobling &Baardvik, 1994; Shelverton & Carter, 1998). Foraging theory predicts that individualsattempt to maximize their net energy gain by foraging in patches with high densitiesof preferred prey (Stephens & Krebs, 1996). This energy maximization premiseassumes that the forager adopts a strategy, largely depending on the ambient foragingconditions, which maximizes its long-term fitness (Townsend & Winfield, 1985).Motivation to feed increases with hunger, which also affects short-term behaviouralchanges such as increased search rate (Colgan, 1973; Dill, 1983; Hojesjo et al ., 1999;Vehanen, 2003).

There are numerous publications on the feeding behaviour of GH-transgenicsalmonids (Abrahams & Sutterlin, 1999; Devlin et al ., 1999; Sundstrom et al .,2004a, 2007a; Tymchuk et al ., 2005). As far as is known, however, there have notbeen any reports on the feeding behaviour of the ‘all-fish’ GH-transgenic commoncarp Cyprinus carpio L. Here, in a short-term experiment, the ability of hatchery-reared juvenile GH-transgenic C. carpio to compete for limited food resources wascompared with size-matched non-transgenic counterparts. Such information wouldbe useful for understanding the survival component of fitness of these transgenic fishunder competitive conditions and also for assessing potential ecological risk of thetransgenic individuals should they escape or be released into the wild.

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

F E E D I N G C O M P E T I T I O N I N T R A N S G E N I C C Y P R I N U S C A R P I O 1461

MATERIALS AND METHODS

S O U R C E O F F I S H

Fast-growing genotypes of P0 ‘all-fish’ GH-transgenic C. carpio were initially producedby microinjection of the pCAgcGHc into the fertilized eggs of C. carpio (Yellow Rivervariety). The all-fish gene construct pCAgcGHc was a recombinant construct of grass carpCtenopharyngodon idella (Valenciennes) growth hormone cDNA (gcGHc), whose expressionis driven by the β-actin gene promoter of C. carpio (pCA) (Wang et al ., 2001). The F1, F2and F3 generation were, respectively, 1·6 times (Wang et al ., 2001), 1·8–2·5 times (W. Hu,unpubl. data) and 1·4–1·9 times (Li et al ., 2007) the body mass of non-transgenic counterpartsunder hatchery-reared conditions, showing how the growth enhancement remains relativelystable across generations. The F5 generation transgenic and non-transgenic fish were producedfrom crosses between a wild-type female and an F4 hemizygous transgenic male of a fast-growing transgenic strain on 25 April 2007. Frequencies of transgene transmission to F5progeny from this line were c. 50% (W. Hu, unpubl. data). Siblings were used to minimizeeffects of genetic differences and maternal and paternal effects. After their emergence inDuofu Technology Farm, Wuhan, China, the first-feeding fry, containing a mix of the twogenotypes, were transferred to four concrete rectangular pools (8 m3) with a rearing densityof 100 individuals m−3 each. A month later, the mixed populations were transferred to anindoor re-circulating system and reared in four circular fibreglass tanks (diameter 150 cm,volume 1000 l) at the Institute of Hydrobiology, Chinese Academy of Sciences. Thereafter,the pCAgcGHc transgene-positive fish were identified by the polymerase chain reaction (PCR)method following Wang et al . (2001). The two genotypes then were reared separately in twoof the fibreglass tanks as described above. Non-transgenics were fed to satiation with frozenchironomid Chironomus sp. larvae twice daily, whereas the transgenics were fed roughly thesame amount every other day to allow size-matching and avoid the possible effects of sizeon competitive ability (Huntingford et al ., 1990). This feeding regime was maintained untilthe fish gape sizes matched the size of the artificial food pellets (mean ± s.e. 4·6 ± 0·3 mg;comprised of 33·2% protein, 9·1% lipid, 11·4% ash and 18·1 J mg−1 energy).

E X P E R I M E N TA L P RO C E D U R E S

The non-commercial research facility at the Fish Behavioral Ecology Laboratory of theInstitute of Hydrobiology had multiple containment screen systems and was specially designedto prevent transgenic fish from escaping to nature. The experiment consisted of two main parts:(1) an acclimation period (days 1–7) and (2) an observation period (days 8–10). The waterinlet to the research facilities was at 10 m depth in the Liangzi Lake near the laboratory.Incoming water was aerated and filtered with an adjustable internal filter. During the exper-iment, the water temperature was 26·1 ± 0·3◦ C, the air temperature was 25·2 ± 0·9◦ C, thedissolved oxygen was 6·6 ± 0·1 mg l−1 and the pH was 6·9 ± 0·01 (mean ± s.e.).

Treatment periodOne hundred transgenic and 100 non-transgenic individuals were placed separately into

two rectangular aquaria (210 × 35 × 35 cm in size) on 1 September 2007, and were fed 2%of total body mass with the artificial food pellets twice daily as described above. On day 1 (5September), 12 individuals of each genotype (size-matched within pairs) were anesthetizedwith eugenol. Body mass (M , to the nearest 0·01 g) and total length (LT nearest 0·01 cm)were measured. Individual fish in each pair were identified visually through randomly markingthem with white (a saturated solution of white titanium oxide and water was used) (Southwood& Henderson, 2000) or blue (alcian blue) dye (Nicieza & Metcalfe, 1999; Westerberg et al .,2004) by subcutaneous injection. Next, the 12 pairs were transferred to 12 tempered glassaquaria measuring 35 × 35 × 35 cm with a water depth of 20 cm to acclimate. Two fluorescentlights were fixed 1 m above each aquarium [a 12 L:12 D (on 0700 and off 1900 hours)photoperiod was used]. A black plastic screen was used to cover all sides of each aquariumto minimize disturbance and to maintain similar levels of illumination. In addition, each

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

1462 M . D UA N E T A L .

aquarium had a white gravel substratum and a white-bottomed Petri dish (diameter 10 cm)located at the centre of the aquarium bottom. The dish marked the feeding area and clumpedthe food (i.e. prevented food pellets from drifting away). A tube from the outside of thescreen was fixed above the feeding area 2 cm under the water surface allowing sinking foodpellets to be administered in the exact location each time without any disturbance to the fish(Abrahams, 1989; Tymchuk et al ., 2005).

During the acclimation period, a pilot study was undertaken to ascertain the maximumnumber of food pellets consumed by a single fish at each feeding trial. An individual wasestimated to be satiated when it refused to grab, or did not swallow five consecutive pelletswhich then sank gently down to the bottom of the aquarium [the method was modified fromJohnsson & Bjornsson (1994)]. Neither of the two genotypes would consume >30 food pelletsduring a 10 min interval. In the light of these results, each pair was fed with 10 pellets atone time, but twice daily (0800–1100 and 1400–1700 hours) during the 6 day acclimation.Uneaten food items and faeces were removed by a glass siphon after 10 min. The fish weredeprived of food for 1 day (day 7) before the observations commenced on day 8.

ObservationsThe same feeding procedures described above were repeated during the 3 day behavioural

observations (from 13 September to 15 September).Behaviours of the fish were recorded by a digital videocassette recorder (Sony, HDR-HC1E

Handy Camera; www.sony.net) placed 50 cm above the aquarium. The order of observationswas randomly allocated among the aquaria so as to remove any bias and order effects. Foreach aquarium, the video camera was started, and then 10 food pellets were poured into thePetri dish through the feeding tube at once (no fish fed during this period). Behaviours ofthe two individuals then were recorded for 10 min. Observations were made twice daily, at0800–1100 and 1400–1700 hours during the feeding periods. Consequently, this procedureyielded six repeated measures for competitive ability for each of the 12 pairs during the3 day observations. Four main behavioural variables of the two genotypes were noted fromthe video: feeding order, number of food pellets consumed, number of movements (whena fish moved more than its LT following by a less active period of at least 1 s) and thetime spent in the feeding area (when the entire fish remained over the Petri dish). After thelast observation on day 10 (15 September), LT and M of the 24 individual fish were againmeasured under light anaesthesia.

DATA A NA LY S I S

The specific growth rate for LT was calculated as GL = 100 (ln LT2− ln LT1)t−1(Ricker,

1979), where LT2 and LT1 were the final and initial LT (cm) and t is the experimental period(10 days, consisting of the 7 day treatment period and the 3 day observations). The conditionfactor (K) was calculated as K = 100ML−3

T .Most data were normally distributed and had homogeneous variance, so paired t-tests and a

general linear model (GLM) were used. There was some variance heterogeneity for the num-ber of movements, so these data were square-root transformed prior to statistical analysis.Rank data on feeding order were tested with the Wilcoxon signed ranks test. For evaluatingcorrelations between two behavioural variables, Spearman’s rank correlation (rs) and partialcorrelation analysis were used. Differences were regarded as significant when P < 0·05. Allthe data were analysed with SPSS 15.0 (www.spss.com) and described as mean ± s.e.

RESULTS

G ROW T H

At the beginning of the experiment, there was no significant difference in LTbetween the two genotypes (paired t-test, d.f = 11, P > 0·05), but transgenic fish

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

F E E D I N G C O M P E T I T I O N I N T R A N S G E N I C C Y P R I N U S C A R P I O 1463

Table I. Total length (LT), body mass (M) and condition factor (K) of transgenic and non-transgenic Cyprinus carpio before the experiment (mean ± s.e.)

Genotype n LT (cm) M (g) K

Transgenic 12 6·75 ± 0·04 4·39 ± 0·15∗ 1·42 ± 0·03∗Non-transgenic 12 6·76 ± 0·03 3·71 ± 0·08 1·20 ± 0·02

n, number of fish.∗, A significant difference between the two genotypes, P < 0·05 (paired t-test).

were heavier (paired t-test, d.f. = 11, P < 0·001), and consequently had a higher K

(paired t-test, d.f = 11, P < 0·001) (Table I). During the 10 day experiment, therewas not a statistically significant difference (ANCOVA, with initial M as a covari-ate, d.f = 1, 21, P > 0·05. adjusted means: transgenics, 0·52 ± 0·09% day−1, non-transgenics, 0·49 ± 0·09% day−1; Fig. 1) although the GL of transgenic C. carpio(0·53 ± 0·07% day−1) was higher than that of non-transgenic carp (0·48 ± 0·08%day−1). Initial M had no significant effect on the GL (ANCOVA, d.f = 1, 21, P >

0·05). Compared with the growth potential of these two genotypes under rich (natu-ral and artificial foods) hatchery conditions (Li et al ., 2007), transgenic fish attainedon average 9·0% of their growth potential and non-transgenic fish attained 29·4% oftheir growth potential under these competitive conditions.

F O O D C O N S U M P T I O N

Overall, out of 60 pellets provided, the number of food pellets consumed by trans-genic C. carpio was 38·5 ± 2·9 during the 3 day observation period, which was sig-nificantly more (repeated measures ANOVA, between-subject genotype; d.f. = 1, 22,P < 0·001) than that consumed by non-transgenic fish (20·7 ± 3·1). This higher food

0·65

0·60

0·55

0·50

0·45

GL (

% d

ay−1

)

0·40

0·35Transgenics Non-transgenics

Fig. 1. Mean ± s.e. specific growth rate for total length (GL) of transgenic (n = 12) ( ) and non-transgenic(n = 12) ( ) Cyprinus carpio during the 10 day experiment.

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

1464 M . D UA N E T A L .

01 2 3 4

Trial

5 6

2

4

Num

ber

of f

ood

pelle

ts c

onsu

med

6

8

10

Fig. 2. Mean ± s.e. amount of food consumed by transgenic (n = 12) ( ) and non-transgenic (n = 12) ( )Cyprinus carpio during each of six consecutive trials. All of the differences between the two genotypesin each trial were significant (P < 0·05).

intake was consistent over time (within-subject trial; d.f = 5, 110, P > 0·05), withno significant interaction between trial and genotype (d.f = 5, 110, P > 0·05), indi-cating that the transgenic fish maintained a higher consumption rate during the sixconsecutive observations (Fig. 2).

F E E D I N G O R D E R

The individuals were ranked according to their scores obtained for feeding order.The first-feeding fish obtained a score of 2, the last (second-feeding) one was givena score of 1 and if the fish did not feed it was given a score of 0. Overall, the scoresobtained for transgenics (1·6 ± 0·1) were significantly higher (Wilcoxon signed rankstest, z = −2·149, P < 0·05) than those for the non-transgenics (1·1 ± 0·2), indicatingthat transgenics were faster to grasp the food pellets.

T I M E S P E N T I N T H E F E E D I N G A R E A

Transgenic C. carpio spent >50% of their time in the feeding area (306·5 ±42·6 s), which was significantly more (repeated measures ANOVA, between-subjectgenotype; d.f = 1, 22, P < 0·01) than that for the non-transgenic C. carpio (143·7 ±25·4 s).

Trial had no significant effect on the time spent in the feeding area (within-subject trial, d.f = 5, 110, P > 0·05), nor was the interaction between trial andgenotype significant (d.f. = 5, 110, P > 0·05; Fig. 3). These data indicate that trans-genic C. carpio often monopolized the feeding areas, and thus they were dominant,whereas the non-transgenic fish were subordinate. In addition, dominant C. carpiowere often observed to swim freely in aquaria, whereas the subordinate individualswere residing in the corners.

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

F E E D I N G C O M P E T I T I O N I N T R A N S G E N I C C Y P R I N U S C A R P I O 1465

01 2 3 4

Trial5 6

20

40

Tim

e sp

ent i

n th

e fe

edin

g ar

eas

(s)

60

80

100

Fig. 3. Mean ± s.e. amount of time spent in the feeding areas for transgenic (n = 12) ( ) and non-transgenic(n = 12) ( ) Cyprinus carpio during each of six consecutive trials. All of the differences between thetwo genotypes in each trial were significant (P < 0·05).

M OV E M E N T L E V E L

Overall, the number of movements by transgenic fish averaged 353·2 ± 32·1,which was significantly higher (repeated measures ANOVA, between-subject geno-type, d.f = 1, 22, P < 0·01) than that of non-transgenic fish, 203·8 ± 29·5.

Trial had a significant effect (within-subject trial, d.f = 5, 110, P = <0·01) onthe number of movements, but the pattern was similar for the two genotypes (d.f =5, 110, P > 0·05; Fig. 4).

R E L AT I O N S H I P S B E T W E E N F O O D C O N S U M P T I O N , F E E D I N GO R D E R A N D M OV E M E N T L E V E L

There was a positive correlation between the food consumption and the frequencyof movement of the two genotypes (rs = 0·437, n = 24, P < 0·05), indicating thatthe higher the number of movements, the more food was consumed. Moreover, therewas a significant positive correlation between feeding order and food consumption(rs = 0·547, n = 24, P < 0·01), indicating that the earlier the food acquisition, themore food was consumed. In order to determine which factor among frequency ofmovement and feeding order more strongly affected the quantity of food consumed,partial correlation was used. After controlling for feeding order, no clear correlationwas found between frequency of movement and the number of food pellets con-sumed (rs = −0·017, n = 21, P > 0·05). On the other hand, when movement levelwas controlled for, there was a significant positive correlation between feeding orderand the number of the food pellets consumed by the two genotypes (rs = 0·664,n = 21, P = <0·001). These results indicate that feeding order had a much strongereffect on the food consumption than did the frequency of movement (Fig. 5).

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

1466 M . D UA N E T A L .

01 2 3 4

Trial

5 6

20

40

Num

ber

of m

ovem

ents

60

80

100

Fig. 4. Mean ± s.e. number of movements of transgenic (n = 12) ( ) and non-transgenic (n = 12) ( )Cyprinus carpio during each of six consecutive trials. All of the differences between the two genotypesin each trial were significant (P < 0·05).

0

2

1Feeding order 0

0

20

Numbe

r of m

ovem

ents

4060

801002

4

Num

ber

of f

ood

pelle

ts c

onsu

med

6

8

10

Fig. 5. Correlations between mean number of movements, social status (feeding order) and food consumption(number of food pellets consumed) for transgenic (n = 12) ( ) and non-transgenic (n = 12) ( ) Cyprinuscarpio across six consecutive trials.

DISCUSSION

In the present study, expression of the all-fish GH transgene altered a numberof feeding-related behaviours in C. carpio, mainly increasing food consumption andfeeding order, time spent in the feeding areas and the frequency of movement. These

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

F E E D I N G C O M P E T I T I O N I N T R A N S G E N I C C Y P R I N U S C A R P I O 1467

results are highly consistent with previous studies on growth hormone function insalmonids, both from GH injection (Johnsson & Bjornsson, 1994; Jonsson et al .,1996, 1998) and GH transgenesis (Abrahams & Sutterlin, 1999; Devlin et al ., 1999;Tymchuk et al ., 2005). The present study suggests that GH influences competitionfor food through increased feeding motivation in C. carpio in a similar way.

The GH transgene induces over-expression of GH, which may elevate metabolismand appetite in GH-transgenic C. carpio (Fu et al ., 2007), which in turn enhancesfeeding motivation and induces short-term behavioural changes such as increasedsearch rate (Colgan, 1973; Dill, 1983; Hojesjo et al ., 1999; Vehanen, 2003). Quitesimilarly, GH injection in rainbow trout Oncorhynchus mykiss (Walbaum) was foundto increase foraging due to hunger (Jonsson et al ., 1998), it may be that the higherhunger levels may translate to transgenic fishes being less choosey when it comes toforaging. Constant searching for prey may increase activity and consequently resultin higher energy costs for the individual. Fu et al . (2007) observed an increasedfood intake for the F2 GH-transgenic C. carpio of the same strain as employed inthe present study, and suggested that their elevated food intake partly compensatefor a higher metabolism originating from greater activity. In Fu et al . (2007) study,they cultured the two genotypes separately, and sufficient food resources were sup-plied, and thus the growth of the two genotypes did not interfere competitively, andthe transgenic individuals may not significantly express their agonistic behaviourpotential towards cohorts (Devlin et al ., 2004). Besides, the M of the fish used inthe study of Fu et al . (2007) was >138 g, whereas in the present study, the M ofjuvenile transgenic fish was <5 g, so the behaviours of these two different animalswere probably different. In addition, the transgenic fish were fed with a restrictedration before the acclimation period (i.e. feeding every other day), which may also bea factor that could affect the transgene effects on the feeding motivation due to somephysiological alterations. In addition, the short duration of the present experimentmay affect this result. The transgenic C. carpio, however, did not realize their growthpotential under restricted food availability conditions during the 10 day experiment,which parallels results of studies on transgenic O. kisutch where transgenics did notreach their highest growth potential under semi-natural environments (Sundstromet al ., 2004a, 2007b).

According to Metcalfe (1986), less competitive individuals may use a feeding strat-egy other than that used by stronger competitors. Instead of attempting to maximizefood intake, less competitive individuals may try to minimize energetic expenditureby, for example, decreasing their activity. The energetic cost to obtain one food pelletmay therefore be significantly different between the two C. carpio genotypes. Underconditions of limited food supply, the benefits of the increased energy intake by trans-genic fish, however, were diminished by their increased movements and metabolism,which may partly explain why the juvenile GH-transgenic C. carpio did not expressfast growth potential during the short-term (10 days) experiment. The results furthersuggest that the initial size differences between the two genotypes are not the maincause of the higher competitive ability found in transgenic C. carpio.

Feeding rank within a group has been used as an indirect measure of social rank ina number of studies involving salmonids, flatfishes and tilapiines (McCarthy et al .,1992, 1999; Carter et al ., 1994; Jobling & Baardvik, 1994; Shelverton & Carter,1998). Once individuals recognize a conspecific’s rank (Utne-Palm & Hart, 2000)or when the ranks are settled (Tindo & Dejean, 2000; Forkman & Haskell, 2004),

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

1468 M . D UA N E T A L .

costly overt aggression may not be necessary, thereby decreasing the risk of fatalinjury or stress that may lead to increased susceptibility to pathogens (Huntingford& Turner, 1987). In the present study, the treatment period was 6 days, and therewas adequate time to establish a feeding hierarchy. Therefore, the feeding orderand the time spent in the feeding area were used as metrics to evaluate the feedinghierarchy (McCarthy et al ., 1999; Bailey et al ., 2000). Transgenic C. carpio weremuch faster to occupy the feeding areas and to obtain the food pellets. Furthermore,they spent more time in the feeding areas, also indicating that the juvenile transgenicC. carpio were dominant over the non-transgenic individuals. Similarly, dominantthree-spined sticklebacks Gasterosteus aculeatus L. consumed more food and grewfaster, but subordinate individuals could also gain body mass by sneaking access tofood (Sneddon et al ., 2006). Similarly, non-transgenic fish often were observed tocarefully approach the feeding area and sneakily fed upon the remaining food whilethe transgenic individuals were approaching fast and swallowing several food pelletswhile staying in the area for a considerable time.

Environmental conditions are likely to determine whether it would pay for a fishto be highly aggressive and monopolize the food resource or to avoid aggressiveencounters and obtain food in a sneaky manner (Sneddon et al ., 2006). In the presentstudy, fish were kept in small and simple tank environments where it presumablywas easy for dominant transgenic fish to dominate the food source, but in nature con-ditions would be more complex (Devlin et al ., 2004). Hojesjo et al . (2005) observedthat subordinate Atlantic salmon Salmo salar L. fed by briefly invading the spaceoccupied for most of the time by the dominant brown trout Salmo trutta L. in acontrolled stream channel environment, and the S. salar using such sneaky strat-egy fed no less than did the dominants. Under natural conditions, food availabilitywould be less predictable and predation risk may influence feeding decisions so thatthe advantage of the transgenic fish could be lessened or increased depending ontheir responses under such complex environments. Arendt (1997) argued that therewere many functions that may be compromised when intrinsic growth rates wereaccelerated to meet evolutionary challenges. Conversely, if selection acts to improveany of these functions, intrinsic growth rates might be reduced. Thus, further workon the behavioural and developmental alteration of the transgenic fish in complexenvironments is needed.

The present study is a first approximation of answering whether transgenicC. carpio compete effectively for food. Cyprinus carpio, however, are different fromsalmonids in many important aspects of foraging behaviour, often spending muchtime foraging together for food items such as aquatic invertebrates. Hence, furtherresearch should focus on how much time the fish spend foraging, how likely they areto deplete food resources and how likely they are to engage in agonistic behaviourif food is limiting.

In the present study, the all-fish GH-transgenic C. carpio exhibited elevated feed-ing motivation, higher frequency of movement, earlier food capture tendency andhigher feeding rank, and thus increased the ability to compete for limited foodresources in comparison with non-transgenic conspecifics. These results suggest thatsuch altered feeding behaviour would probably benefit the transgenic individuals innatural ecosystems. Environmental complexity such as food abundance and predationrisk levels, however, generally reduces hormonal effects on behaviour (Sloman &Armstrong, 2002) and can greatly alter the phenotypic effects of the GH transgene

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

F E E D I N G C O M P E T I T I O N I N T R A N S G E N I C C Y P R I N U S C A R P I O 1469

(Sundstrom et al ., 2007b). Muir & Howard (2001), however, presented a comprehen-sive method by using the wild-type and transgenic Japanese medaka Oryzias latipes(Temminck & Schlegel) to determine environmental risk of transgenic fish that theycalled the ‘net fitness approach’. They grouped various aspects of an organism’slife cycle into six net fitness components: juvenile viability, adult viability, age atsexual maturity, female fecundity, male fertility and mating advantage. The presentstudy only examined a single fitness-related trait, the ability to engage in interferencecompetition. The entire life cycle, however, needs to be studied to determine overallfitness of transgenic C. carpio because once released into the natural environment,the impact may be permanent and irreversible (Muir & Howard, 1999, 2001). Fur-thermore, genotype-by-environment interactions will play a crucial role determiningrisk levels posed by transgenic fishes to natural ecosystems (Sundstrom et al ., 2005;Devlin et al ., 2006). Thus, the results of this study should be used with caution whenassessing the risk of genetically modified C. carpio.

The authors wish to thank K. V. Radhakrishnan and S. Ye for their constructive com-ments on the manuscript and also thank W. Li, B. Guan and D. Li for their help duringthe experiment. This work was financially supported by the Development Plan of the StateKey Fundamental Research of China (Grant No. 2007CB109205) and the National NaturalScience Foundation (Grant Nos. 30770377 and 30830025).

References

Abrahams, M. V. (1989). Foraging guppies and the ideal free distribution: the influence ofinformation on patch choice. Ethology 82, 116–126.

Abrahams, M. V. & Sutterlin, A. (1999). The foraging and antipredator behaviour of growth-enhanced transgenic Atlantic salmon. Animal Behaviour 58, 933–942.

Alm, G. (1959). Relation between maturity, size and age in fishes. Report of the Institute ofFreshwater Research, Drottningholm 40, 4–145.

Arendt, J. D. (1997). Adaptive intrinsic growth rates: an integration across taxa. The QuarterlyReview of Biology 72, 149–177.

Bailey, J., Alanara, A. & Brannas, E. (2000). Methods for assessing social status in Arcticcharr. Journal of Fish Biology 57, 258–261.

Begon, M., Harper, J. L. & Townsend, C. R. (1996). Ecology , 3rd edn. Oxford: BlackwellScientific.

Carter, C. G., McCarthy, I. D., Houlihan, D. F., Johnstone, R., Walsingham, M. V. &Mitchell, A. I. (1994). Food consumption, feeding behaviour and growth of triploidand diploid Atlantic salmon, Salmo salar L., parr. Canadian Journal of Zoology 72,609–617.

Colgan, P. W. (1973). Motivational analysis of fish feeding. Behaviour 45, 38–66.Devlin, R. H., Yesaki, T. Y., Biagi, C. A. & Donaldson, E. M. (1994). Extraordinary salmon

growth. Nature 371, 209–210.Devlin, R. H., Johnsson, J. I., Smailus, D. E., Biagi, C. A., Jonsson, E. & Bjornsson, B. T.

(1999). Increased ability to compete for food by growth hormone-transgenic cohosalmon Oncorhynchus kisutch (Walbaum). Aquaculture Research 30, 479–482.

Devlin, R. H., Biagi, C. A., Yesaky, T. Y., Smailus, D. E. & Byatt, J. C. (2001). Growth ofdomesticated transgenic fish: a growth-hormone transgene boosts the size of wild butnot domesticated trout. Nature 409, 781–782.

Devlin, R. H., D’Andrade, M., Uh, M. & Biagi, C. A. (2004). Population effects of growthhormone transgenic coho salmon depend on food availability and genotype by envi-ronment interactions. Proceedings of the National Academy of Sciences of the UnitedStates of America 101, 9303–9308.

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

1470 M . D UA N E T A L .

Devlin, R. H., Sundstrom, L. F. & Muir, W. M. (2006). Interface of biotechnology and ecol-ogy for environmental risk assessments of transgenic fish. Trends in Biotechnology 24,89–97.

Dill, L. M. (1983). Adaptive flexibility in the foraging behaviour of fishes. Canadian Journalof Fisheries and Aquatic Sciences 40, 398–408.

Farrell, A. P., Bennett, W. & Devlin, R. H. (1997). Growth-enhanced transgenic salmon canbe inferior swimmers. Canadian Journal of Zoology 75, 335–337.

Forkman, B. & Haskell, M. J. (2004). The maintenance of stable dominance hierarchies andthe pattern of aggression: support for the suppression hypothesis. Ethology 110,737–744.

Fu, C., Cui, Y., Hung, S. S. O. & Zhu, Z. (1998). Growth and feed utilization by F4 humangrowth hormone transgenic carp fed diets with different protein levels. Journal of FishBiology 53, 115–129.

Fu, C., Li, D., Hu, W., Wang, Y. & Zhu, Z. (2007). Growth and energy budget of F2 “all-fish”growth hormone gene transgenic common carp. Journal of Fish Biology 70,347–361.

Gross, M. L., Schneider, J. F., Moav, N., Moav, B., Alvarez, C., Myster, S. H., Liu, Z.,Hallerman, E. M., Hackett, P. B., Guise, K. S., Faras, A. J. & Kapuscinski, A. R.(1992). Molecular analysis and growth evaluation of northern pike (Esox lucius)microinjected with growth hormone genes. Aquaculture 103, 253–273.

Hojesjo, J., Johnsson, J. & Axelsson, M. (1999). Behavioural and heart rate responses to foodlimitation and predation risk: an experimental study on rainbow trout. Journal of FishBiology 55, 1009–1019.

Hojesjo, J., Armstrong, J. D. & Griffiths, S. W. (2005). Sneaky feeding by salmon in sympatrywith dominant brown trout. Animal Behaviour 69, 1037–1041.

Huntingford, F. A. & Turner, A. K. (1987). Animal Conflict . London: Chapman & Hall.Huntingford, F. A., Metcalfe, N. B., Thorpe, J. E., Graham, W. D. & Adams, C. E. (1990).

Social dominance and body size in Atlantic salmon parr, Salmo salar L. Journal ofFish Biology 36, 877–881.

Jhingan, E., Devlin, R. H. & Iwama, G. K. (2003). Disease resistance, stress response andeffects of triploidy in growth hormone transgenic coho salmon. Journal of Fish Biology63, 806–823.

Jobling, M. & Baardvik, B. M. (1994). The influence of environmental manipulations oninter-and intra-individual variation in food acquisition and growth performance ofArctic charr, Salvelinus alpinus. Journal of Fish Biology 44, 1069–1087.

Johnsson, J. I. & Bjornsson, B. Th. (1994). Growth hormone increases growth rate, appetite,and dominance in juvenile rainbow trout, Oncorhynchus mykiss. Animal Behaviour 48,177–186.

Jonsson, E., Johnsson, J. I. & Bjornsson, B. Th. (1996). Growth hormone increases predationexposure of rainbow trout. Proceedings of the Royal Society B 263, 647–651.

Jonsson, E., Johnsson, J. I. & Bjornsson, B. T. (1998). Growth hormone increases aggressivebehavior in juvenile rainbow trout. Hormones and Behavior 33, 9–15.

Kapuscinski, A. R., Hayes, K. R., Li, S. & Dana, G. (2007). Environmental Risk Assessmentof Genetically Modified Organisms , Vol. III. Wallingford: CABI Publishers.

Li, D., Fu, C., Hu, W., Zhong, S., Wang, Y. & Zhu, Z. (2007). Rapid growth cost in “all-fish” growth hormone gene transgenic carp: reduced critical swimming speed. ChineseScience Bulletin 52, 1501–1506.

McCarthy, I. D., Carter, C. G. & Houlihan, D. F. (1992). The effect of feeding hierarchy onindividual variability in daily feeding of rainbow trout, Oncorhynchus mykiss (Wal-baum). Journal of Fish Biology 41, 257–263.

McCarthy, I. D., Gair, D. J. & Houlihan, D. F. (1999). Feeding rank and dominance in Tilapiarendalli under defensible and indefensible patterns of food distribution. Journal of FishBiology 55, 854–867.

McLean, E. & Donaldson, E. M. (1993). The role of growth hormone in the growth of poikilo-therms. In The Endocrinology of Growth, Development, and Metabolism in Vertebrates(Schreibman, M. P., Scanes, C. G. & Pang, P. K. T., eds), pp. 43–71. San Diego, CA:Academic Press.

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

F E E D I N G C O M P E T I T I O N I N T R A N S G E N I C C Y P R I N U S C A R P I O 1471

Metcalfe, N. B. (1986). Intraspecific variation in competitive ability and food intake insalmonids: consequences for energy budgets and growth rates. Journal of Fish Biology28, 525–531.

Muir, W. M. & Howard, R. D. (1999). Possible ecological risks of transgenic organism releasewhen transgenes affect mating success: sexual selection and the Trojan gene hypothesis.Proceedings of the National Academy of Sciences of the United States of America 96,13853–13856.

Muir, W. M. & Howard, R. D. (2001). Fitness components and ecological risks of transgenicrelease: a model using Japanese medaka (Oryzias latipes). The American Naturalist158, 1–16.

Nicieza, A. G. & Metcalfe, N. B. (1999). Costs of rapid growth: the risk of aggression ishigher for fast-growing salmon. Functional Ecology 13, 793–800.

Raven, P. A., Devlin, R. H. & Higgs, D. A. (2006). Influence of dietary digestible energycontent on growth, protein and energy utilization and body composition of growth hor-mone transgenic and non-transgenic coho salmon (Oncorhynchus kisutch). Aquaculture254, 730–747.

Ricker, W. E. (1979). Growth rates and models. In Fish Physiology , Vol. VIII (Hoar, W. S.,Randall, D. J. & Brett, J. R., eds), pp. 677–743. New York, NY: Academic Press.

Shelverton, P. A. & Carter, C. G. (1998). The effect of ration on behaviour, food consumptionand growth in juvenile greenback flounder (Rhombosolea tapirina: Teleostei). Journalof the Marine Biological Association of the United Kingdom 78, 1307–1320.

Sloman, K. A. & Armstrong, J. D. (2002). Physiological effects of dominance hierarchies:laboratory artefacts or natural phenomena. Journal of Fish Biology 61, 1–23.

Sneddon, L. U., Hawkesworth, S., Braithwaite, V. A. & Yerbury, J. (2006). Impact of envi-ronmental disturbance on the stability and benefits of individual status within domi-nance hierarchies. Ethology 112, 437–447.

Southwood, T. R. E. & Henderson, P. A. (2000). Ecological Methods . Oxford: BlackwellScience.

Stephens, D. W. & Krebs, J. R. (1996). Foraging Theory . Princeton, NJ: Princeton UniversityPress.

Sundstrom, L. F., Lohmus, M., Devlin, R. H., Johnsson, J. I., Biagi, C. A. & Bohlin, T.(2004a). Feeding on profitable and unprofitable prey: comparing behaviour of growth-enhanced transgenic and normal coho salmon (Oncorhynchus kisutch). Ethology 110,381–396.

Sundstrom, L. F., Lohmus, M., Johnsson, J. I. & Devlin, R. H. (2004b). Growth hormonetransgenic salmon pay for growth potential with increased predation mortality. Pro-ceedings of the Royal Society B 271, S350–S352.

Sundstrom, L. F., Lohmus, M. & Devlin, R. H. (2005). Selection on increased intrinsic growthrates in coho salmon Oncorhynchus kisutch. Evolution 59, 1560–1569.

Sundstrom, L. F., Lohmus, M., Johnsson, J. I. & Devlin, R. H. (2007a). Dispersal potentialis affected by growth-hormone transgenesis in coho salmon (Oncorhynchus kisutch).Ethology 113, 403–410.

Sundstrom, L. F., Lohmus, M., Tymchuk, W. E. & Devlin, R. H. (2007b). Gene-environmentinteractions influence ecological consequences of transgenic animals. Proceedings ofthe National Academy of Sciences of the United States of America 104, 3889–3894.

Tindo, M. & Dejean, A. (2000). Dominance hierarchy in colonies of Belonogaster junceajuncea. Insectes Sociaux 47, 158–163.

Townsend, C. R. & Winfield, I. J. (1985). The application of optimal foraging theory to feed-ing behaviour in fish. In Fish Energetics. New Perspectives (Tytler, P. & Calow, P.,eds), pp. 67–98. Sydney: Croom Helm.

Tymchuk, W. E., Abrahams, M. V. & Devlin, R. H. (2005). Competitive ability and mortalityof growth-enhanced transgenic coho salmon fry and parr when foraging for food.Transactions of the American Fisheries Society 134, 381–389.

Utne-Palm, A. C. & Hart, P. J. B. (2000). The effects of familiarity on competitive interac-tions between threespined sticklebacks. Oikos 91, 225–232.

Vehanen, T. (2003). Adaptive flexibility in the behaviour of juvenile Atlantic salmon: short-term responses to food availability and threat from predation. Journal of Fish Biology63, 1034–1045.

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472

1472 M . D UA N E T A L .

Venugopal, T., Anathy, V., Kirankumar, S. & Pandian, T. J. (2004). Growth enhancementand food conversion efficiency of transgenic fish Labeo rohita. Journal of ExperimentalZoology 301A, 477–490.

Wang, Y., Hu, W., Wu, G., Sun, Y., Chen, S., Zhang, F., Zhu, Z., Feng, J. & Zhang, X.(2001). Genetic analysis of “all-fish” growth hormone gene transferred carp (Cyprinuscarpio L.) and its F1 generation. Chinese Science Bulletin 46, 1174–1177.

Werner, E. E. & Gilliam, J. F. (1984). The ontogenetic niche and species interactions in size-structured populations. Annual Review of Ecology and Systematics 15, 393–425.

Westerberg, M., Staffan, F. & Magnhagen, C. (2004). Influence of predation risk on individualcompetitive ability and growth in Eurasian perch, Perca fluviatilis. Animal Behaviour67, 273–279.

Zbikowska, H. M. (2003). Fish can be first–advances in fish transgenesis for commercialapplications. Transgenic Research 12, 379–389.

Zhu, Z., Li, G., He, L. & Chen, S. (1985). Novel gene transfer into the fertilized eggs of gold-fish (Carassius auratus L. 1758). Zeitschrift fur Angewandte Ichthyologie 1, 31–34.

© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 1459–1472