The swimming crab，Portunus trituberculatus(Miers，1876)，naturally distributed in the coastal seasof Asia-Pacifi c countries，is an important farmedspecies in China. Commercial farming commenced inthe 1970s in China，and the annual production hasrecently reached 131 000 tons，valued at $4.3 billionUSD in China Fisheries Yearbook of 2012.

The industry，however，is impeded by the lowsurvivorship and unstable yield，because thecommercial characteristics(growth rate，diseaseresistance)of the cultured stocks have declined aftermany years of culturing(Qiao et al.，2012). Manystudies have been done both from China and elsewhereto improve the culturing models(Zhang et al.，2011;Xia et al.，2012)，genetic breeding(Gao et al.，2007;Cui et al.，2012; Liu et al.，2014)，growth and development(Romano and Zeng，2006; Liao et al.，2008)，food and nutrition(Wu et al.，2007; Baylon，2009)，and farming conditions(Bryars and Havenh and ，2006; Stickle et al.，2007; Liao et al.，2008; Wang et al.，2010).

Salinity is one of the most important factors thataffect marine crab’s geographical distribution(Bryars and Havenh and ，2006)，larval hatching(Hamasaki，2003)，growth and development(Anger，1991; Fisher，1999; Romano and Zeng，2006; Nurdiani and Zeng，2007; Liao et al.，2008)，gonadal development(Fisher，1999)，overwintering(Ye，2007; Baylon and Suzuki，2007; Wang et al.，2010) and feeding strategy(Wang et al.，2010). In all of these studies，salinity stress，combined with other environmental factors，wouldinfl uence the yield and output. The stress can berefl ected in the animal behavior，such as feedingbehavior(Ashley，2007). Furthermore，thecharacteristic of feeding behavior，feeding rhythm，and growth rate can be used to guide improvement offarming facilities，culture methods，and stockedquality for reducing the stress(Ashley，2007). In thisstudy，the effects of salinity on the feeding behavior，moulting，and survival of the juvenile swimmingcrabs were investigated.2 MATERIAL AND METHOD

Juveniles of the swimming crab used in the studywere selected from aquaculture stocks(DongchenFisheries Breeding Co.，Xiangshan，Ningbo，China)，with body weight of 0.89±0.14 g，carapace length of23.39±1.31 mm，carapace width of 11.07±0.82 mm and carapace height of 4.81±0.35 mm(mean±SD，n=423).

The experiment was carried out in June 2013 atNingbo University. The crab juveniles weretransported to the temporary cement pond，with 5 cmdepth of s and and 15 cm depth of seawater，a salinityof 20，a temperature of 26°C and was an air pump foraeration. Half of the seawater was replenished daily and the crabs were fed with frozen shrimp purchasedfrom the local market. Four days later，the crabs withall appendages intact were selected for the salinityexperiment.

The experiment was repeated 3 times with thesimilar size of juveniles. Each time，there were 5salinity levels of 5，10，20，30，and 40. Low(5 and 10) and high(30 and 40)salinity levels were achieved bymixing natural seawater from the local area(20)withappropriate amount of distilled water and sea salt(Qianglong Co. Ltd.，Weifang，Sh and ong Province，China)，respectively. The study was conducted inplastic containers(L60 cm×W45 cm×H40 cm)with5 cm depth of s and and 15 cm of seawater. Eachtreatment had 15 replicates(5 for enzyme analysis and 10 for growth investigation). The crabs werer and omly selected，weighed，and placed into thecontainers(one crab per container) and allowed toacclimate for 24 hours. At 9:00 AM every day，uneatenfood was removed，new food was provided，50% ofthe water was replaced，and the juveniles were givennew food. Every day，the new food and uneaten foodwere weighed for the measurement of food intake.Water temperature was maintained at 26±1°C. Theexperiment lasted for 5 days.

Each day of the 5 days duration of the study，onecrab from each treatment was sampled for enzyme(amylase，protease，lipase，superoxide dismutase(SOD)，peroxidase(PO)，catalase(CAT))activityanalysis with the kit from Nanjing JianchengBioengineering Corp.(Nanjing，Jiangsu Province，China). With the other 10 juveniles，food intake，moulting rate and survival rate(%)were measuredevery day，and the body weight gain was measured atthe end of the experiment. The food ration(g/g BW)was calculated by the food intake each day with unitbody weight of each crab. Moulting rate(%)was alsocalculated with the 10 juveniles. Every day for 5 days，the behavior of crab juveniles in every salinitytreatment was recorded by a SONY 2216E videocamera(Sony(China)Corp.，Beijing，China) and categorized according to the behavior description inTable 1.

Table 1 Ethogram of Portunus trituberculatus activity patterns

Food ration，moulting rate，survival rate and body weight gain among the treatments were compared using the two way repeated measures ANOVA. Allstatistical analyses were performed at the signifi cance level of 0.05. The behavior categories among the different salinity treatments were analyzed with theDunn tests(Xu et al.，2011).

Table options

The juveniles at the low salinity treatments of 5 and 10 largely buried into the s and (Rh)at about1.0 cm depth with an angle of 30–45°，with only theeyes and antenna exposed. They only w and ered(As)outside once or twice(1.2±0.45)per day for feeding(for 8.2±1.23 minutes each time and average totalfeeding time of 12.6±5.44 minutes per day). Therewas no signifi cant difference in feeding behavior(Q≤3.36)，feeding times and total feeding periodbetween the two treatments.

At high salinity of 40，the juveniles also hid in s and (Rh)，but at a shallower depth(about 0.5 cm)，withonly the carapace covered. They also only fed once ortwice(1.6±0.55)per day，but each time lasted about30 minutes(30.2±5.34)(Table 2).

Table 2 Comparison of feeding characteristics among the different treatments

Note: a，b，c indicate signifi cant differences among the treatments in each column.

Table options

At the salinity of 20 and 30，the juveniles showedthe rest behavior of Rh or Rs，but w and ered outside(As)frequently in addition to feeding. In these twotreatments the juveniles were present around thefeeding sites in advance for feeding after the secondor third day. When the food was introduced，the crabrapidly approached the food，grabbed it with its one ortwo big chelipeds(the feeding behavior was Fe or Fh). In 24 hours，there was 2–3 feeding times(2.6±0.55)，each time lasted about 25 minutes(25.4±7.16)，and total fed period was 73.2±22.65minutes. There was no signifi cant difference in thefeeding behavior between them，but signifi cantlydifferent with that of high salinity(40)in feedingbehavior(Q≥3.76) and duration of each feeding(P<0.05)，and signifi cantly different with that of lowsalinity(5 and 10)in feeding behavior(Q≥5.39)，feeding times and total feeding period(Table 2). 3.2 Effect on food intake，growth and survival

The effect of salinity on food intake is shown inFig. 1. The average food intake was 0.036 9±0.007 2 g/gBW，0.081 5±0.007 7 g/g BW，0.190 8±0.011 3 g/gBW，0.165 0±0.010 7 g/g BW，0.109 9±0.008 6 g/gBW for the crabs of 5，10，20，30 and 40，respectively(Fig. 1). There was no signifi cant differences betweenthose of 20 and 30，which were signifi cantly higherthan those in the other treatments(

Fig. 1 The average food intake of the Portunus trituberculatusjuveniles reared for 5 days in various salinity levels

Figure options

In all the 3 duplicates of the experiment，no juvenilemoulting was observed at salinity 5，about 16.7%，50%，and 66.7% of the juveniles moulted at 10，40，and 30，respectively，and all the juveniles moulted at20(Fig. 2). The moulting rate was signifi cantlydifferent among the treatments(P<0.05). Themortality rate showed the reverse pattern from that of the moulting rate. All the juveniles died at 5 and allsurvival at 20，about 60%，40%，and 10% of thejuveniles died at 10，40，and 30，respectively(Fig. 2).There was no signifi cant difference in mortality ratebetween that of 20 and 30，which was signifi cantlyhigher than those in the other treatments(P<0.05).

Fig. 2 The moulting and mortality rates of Portunustrituberculatus juveniles reared in various salinitylevels

Figure options

The juveniles at the salinity 20 had a highest valueof gained body weight of 0.796±0.128 g，and the gainrate was 87%–96%，signifi cantly higher than those at10，30 and 40(P<0.05)(Fig. 3). The feed conversionratios(FCR)of every treatment were 21.69±4.61，1.20±0.09，1.43±0.62，and 3.25±0.97 for 10，20，30 and 40，respectively. There was no signifi cant differencebetween those of 20 and 30，which were signifi cantlylower than those of the other treatments(P<0.05).

Fig. 3 The average and the rate of body weight gain(g)of the Portunus trituberculatus juveniles reared invarious salinity levels

Figure options

The effect of salinity on enzyme activities is shownin Fig. 4. The activity variations of protective enzymes(SOD，PO，and CAT)were similar in different salinitytreatments. At the salinity 20，the activities of SOD，PO，and CAT were minimum，and increased withincreasing or decreasing salinity(Fig. 4). Only PObetween 20 and 30 and CAT between 10 and 20 werenot signifi cantly different. The activity variations ofdigestive enzymes(Amylase，Protease，Lipase)alsowere similar among the different salinity treatments，but showed opposite pattern with that of protectiveenzymes(Fig. 4). However，there was no signifi cantdifference among the 20，30 and 40 treatments，whichwas signifi cantly higher than those at salinity of 5 and 10(P<0.05).

Fig. 4 Variations in enzyme activities of the Portunus trituberculatus juveniles reared in various salinity levels

Figure options

Salinity is one of the most important factors thataffect the growth and survival of crustacean juveniles(Anger，1991; Zhang and Li，1992). Many studiesshowed the effect of fl uctuating salinity onphysiological status of crustaceans，such as on feedingintake and consequently growth of juveniles ofFenneropenaeus chinensis(Ding et al.，2008); on thegrowth and development of the larvae of Panulirusstimpsoni(Chen et al.，2003). In this study，weinvestigated the effect of salinity on the feedingbehavior，as well as on the growth and survival ratesof the swimming crab P . trituberculatus(Figs.2 and 3). At a salinity of 5，the juvenile’s behavior and activity showed dramatic changes，such as less frequency of feeding and reduced feeding duration，which resulted in decreased food intake. Thisphenomenon indicated that the stress of low salinitystrongly affect the crab’s feeding behavior，and consequently its growth. They also had a low activityof digestive enzymes and high activity of protectiveenzymes(Fig. 4)，which was similar with the results of Ding et al.’s(2010)study on the activities of enzymesof swimming crab fed with different food，where thelow digestive enzymes and high protective enzymessuggest that the crabs dislike the food. With theincreasing salinity，the juveniles’ feeding behaviorwas restored to normal status，and the food intake alsoincreased correspondingly. But，when the salinityexceeded 20，the salinity stress slightly affected thefood intake. However，the high activities of digestiveenzymes suggested the juvenile crab's increasedpropensity to eat，but the stress of high salinity gaverise to high activities of protective enzymes，whichconsumed more energy and affected the juveniles’growth and survival(Fig. 2). This present study iscomparable with that of Wang et al.’s(2010)reportthat swimming crab juveniles had a slower growth at30 than at 20.4.2 Relationship between feeding behavior and growth characteristics

Crustacean growth performance was expressed asbody weight gain and moulting. Moulting is a complexphysiological process，which consumes large amountsof energy(Anger，1991; Ding et al.，2008). However，a successful moulting can lead to the animal’s rapidgrowth(weight gain). In the present study，all thejuveniles had moulted once during the fi ve days whenreared in a salinity of 20，and gained body weight of 87%–96%. This process mainly uses the stored energymaterial in its body to maintain normal metabolism(Chang，1995; Ding et al.，2008). Therefore，the crabhad to increase food intake during the moultinginterval，and deposit energy material enough for itsmoulting and growth. At low salinity treatments，stress resulted in the change of feeding behavior，which reduced feeding frequency and subsequently，reduced food intake. Furthermore，low salinity stressalso required the animals to allocate more energy forresisting the stress(Chen et al.，2003; Ruscoe et al.，2004). 4.3 Study of feeding behavior for improvingculture technology

If the feeding behavior was normal，the juvenile had a good growth rate. However，the juveniles’growth was not the same even in the treatments withnormal behavior，such as 20 and 30 treatments. Thefeeding behavior，the food intake，and activities ofdigestive enzymes were similar，but the moulting and body weight gain were different，Because the stresspresented at 30 consumed most of the stored energyof the crabs. This is probably the reason why activitiesof protective enzymes were higher and unstable. It istherefore necessary to maintain optimum rearingconditions in the culture of the juveniles of thisspecies of crabs. Farming conditions，aquaculturefacilities，breeding methods and techniques must beenhanced. To promote and improve hatcheryprotocols，future research on behavior research isstrongly recommended.