Aquaculture nutrition, tập 18, số 3, 2012

The literature reporting physical quality isdominated by 4–7 mm pellet diameter because most feeds Lupin Myallie Lupin Wodjil SBM Ingredient and inclusion level % Figure 4 A comparison o

1,2 1

Department of Animal and Aquacultural Sciences, Aquaculture Protein Centre, CoE, Norwegian University of Life Sciences,A˚s, Norway;2 Nofima, A˚s, Norway

Feed comprises the biggest cost in intensive fish farming

and the quality of feed is therefore important A vast body

of research has been carried out in order to investigate

nutritional quality of alternative ingredients Effects of

ingredients on physical quality are seldom included in these

investigations Physical quality of feed varies with

ingredi-ent composition and processing condition and may interfere

with feed intake, nutrient digestibility and therefore growth

performance of the fish In this review, physical quality of

extruded, high energy feed, and how ingredient composition

and processing conditions affect the quality will be

addressed Various pellet properties will be discussed and

methods used to evaluate physical quality will be reviewed

condi-tions, feed ingredients, feed quality, fish feed, physical quality

Received 16 May 2011; accepted 30 October 2011

Correspondence: Mette Sørensen, Aquaculture Protein Centre, CoE,

Department of Animal and Aquacultural Sciences, Norwegian University

of Life Sciences, A˚s, Norway, and Nofima, A˚s, Norway E-mail: mette.

sorensen@nofima.no

The demand for compound aquafeeds was estimated to be

29.3 million tonnes in 2008 and is expected to grow along

with increased global aquaculture production (FAO 2011)

Since 1995 the compound fish feed production has grown at

an average rate of 10.9 percent per year (FAO 2011) In

order to relieve the pressure on fish meal in the steadily

increasing aquaculture sector, alternative feed ingredients

from plant, microbial and other animal sources have been aprioritized field of research for many years Most studiesevaluate the value of new ingredients in terms of nutritionalquality with the main focus on digestibility, growth perfor-mance, health and feed intake Effects of ingredients onphysical quality of the feed are, however, most oftenneglected in feeding experiments with fish Ingredient com-position is the single most important variable affecting thephysical quality of steam pelleted feeds (Behnke 1996), and

is also important for the quality of extruded feeds (Refstie

et al.2006; Sørensen et al 2009, 2010, 2011; Glencross et al

2010, 2011a; Draganovic et al 2011; Kraugerud & Svihus2011; Kraugerud et al 2011) The purpose of this review is

to give an overview of some factors that influence physicalquality of extruded fish feed and how physical quality offeed may affect feed intake and nutritional quality Methodsused to evaluate extruded fish feed are also discussed

Over the course of the past 30 years, extrusion processinghas become the primary technique used for fish feed produc-tion, mainly because of the high physical and nutritionalquality of the feed (Hilton et al 1981) The extrusion systemconsists of a barrel housing with one or two rotating screws(single– or twin screw extruder) The system is also equippedwith a preconditioner as well as an accompanying machinecontrol system The preconditioner is a high speed mixingunit designed for the purpose of mixing water and steam intothe blend of dry ingredients The overall goal with precondi-tioning is to supply the extruder barrel with an evenly moist-ened and preheated mix Preconditioning allows moreefficient transfer of heat through friction in the extruder bar-rel, and also reduces the extruder barrel wear and energy

.

.

Aquaculture Nutrition

consumption during feed production Moisture is necessary

for gelatinization of starch and hydration of proteins

Dur-ing the extrusion cookDur-ing process, the blend of Dur-ingredients is

turned into a melt employing a combination of high

temper-ature (120–130 °C), high pressure (20–30 bar) and shear

forces This transformation of the ingredients is carried out

in the extruder barrel The extruder barrel consists of

extru-der screws, heads and shear locks The screw is a long

cylin-der with helical flight wrapped around it Usually screw

elements are mounted on a shaft with shearlocks or

knead-ing elements mounted between the elements A huge variety

of screw elements with different configurations exist, but

normal practice is to configure the screw as a series of

repeated conveying and mixing elements The conveying

screw(s) generate the pressure necessary for the material to

flow through the die, which causes the restriction in the

out-let of the extruder The final dimensions of the pelout-lets is

shaped in a die, but is also affected by the energy input The

two main sources of energy in an extrusion system are of

mechanical or thermal origin Mechanical energy, reported

as specific mechanical energy (SME), is generated by friction

caused by the dough in the extruder barrel as it is moved

for-ward by the rotating screw The SME is a product of torque

and RPM divided by mass flow rate (Mercier et al 1989)

Thus, heat generation is affected both by the choice of

hard-ware and processing variables during the feed production

Thermal energy, is for most expanded products added as

steam in the preconditioner, in which the mash is usually

heated to 80–90 °C in order to warm up and soften the

ingredients Thermal energy is thus contributing

approxi-mately 2/3 and mechanical energy 1/3 of the energy needed

to obtain an extrusion temperature of approximately 130°C

The time that the feed mash is exposed to heating during

preconditioning and extrusion, is normally less than five

minutes The extrudate is shaped into pellets using a die at

the outlet of the extruder in combination with rotating

kni-ves cutting the pellets at an appropriate length In order to

increase the shelf life, the feed pellets are dried by reducing

moisture content from approximately 300 g kg 1 to

80 g kg 1 For high energy pellets, additional oil is added

using a vacuum coating system before the pellet is cooled

and eventually bagged

Physical quality of pelleted feeds is often defined as the

ability of feed to be handled without creating an excessive

amount of fines All feeds used in intensive aquaculture

should be resistant to mechanical stress during transport,

handling and in pneumatic feeding devices (Aarseth 2004;

Aarseth et al 2006a) At the same time, the feed shouldhave a texture and size that facilitate high feed intake(Hardy 1989) and efficient digestion by the fish (Lovell1989; Baeverfjord et al 2006; Aas et al 2011b) Pellets thatare too hard may cause digestive disturbances in the fish(Pillay & Kutty 2005) The latter authors have reportedthat overfeeding with hard pellets may result in swellingand rupture of the stomach This condition is associatedwith fermentation and gas formation in the stomach Onthe other hand, soft pellets or pellets with low water stabil-ity may cause oil separation in the stomach that potentiallycause oil-belching in rainbow trout suffering from osmoreg-ulatory stress, also associated with abdominal distensionsyndrome (Baeverfjord et al 2006; Aas et al 2011b) Bulkdensity of the pellet is important and can be adjusted dur-ing the production to control sinking velocity and buoy-ancy control Water stability and sinking velocity also have

to be adjusted to the eating habits of the cultured fish cies Some species are fed at the surface with floating feed,others are fed in the water column with use of slow sinkingfeed and some aquatic animals are bottom feeders Sloweating aquatic species such as shrimp and sea urchin needfeeds that are water stable for hours without leaching nutri-ents Feed pellets used in the grow-out phase for Atlanticsalmon may contain up to 40% oil These feeds shouldtherefore be able to absorb and retain fat inside the pellet

spe-The different forces acting on pellets during conveying, dling and storage are usually defined as impact, compressionand shear (Winowiski 1995) Impact forces shatter the pelletsurface and any natural cleavage planes in the pellet Com-pression forces crush the pellet and also cause failure alongcleavage planes Shear forces cause abrasion of the edges andsurface of the pellet (Winowiski 1995) Several methods exist

han-to determine the physical quality of pellets, however, many ofthe methods used for terrestrial animal feeds are not appro-priate for fish feed Common methods used to analyze physi-cal quality of extruded fish feed will be presented herein,whereas in depth overview of methods used in the feed andbriquetting industry is given by other authors (Winowiski1995; Thomas & Van der Poel 1996; Kaliyan & Morey 2009)

Hardness or strength at rupture, defined as the maximumforce needed to crush a pellet, is commonly determined

.

using a texture analyser (Thomas & Van der Poel 1996;

Glencross et al 2010) This method assesses resistance to

breaking when pellets are exposed to external pressure, and

can be used to mimic the force on pellets during storage in

bins or silos, crushing of pellets in a screw conveyor, and

crushing of feed pellets between animal teeth (Kaliyan &

Morey 2009) Despite the reliability of the test, there is a

lack of standardization for analyzing the texture of feed

pellets Most reports are of analysis of laying pellets,

how-ever, if the test is carried out with standing pellets the force

needed to crush the pellet will increase (Sørensen 2011,

unpublished results) Results are therefore often not

comparable The result of the test may also vary with the

various probes and attachments used Hardness may be

reported as shear force if a knife is used (Adamidou et al

2009; Glencross et al 2010) or pellet hardness if a flat

ended probe is used (Aarseth et al 2006b; Øverland et al

2009; Sørensen et al 2009, 2010, 2011; Glencross et al

2010) Hardness of the pellet varies with degree of

expan-sion, ingredients and processing conditions (Table 1; Figs 1

& 2)

Durability is the amount of fines produced from a sample

of pellets after being subjected to mechanical or pneumatic

agitation (Thomas & Van der Poel 1996; Kaliyan & Morey

2009) Pellet durability simulates forces on pellets taking

place during filling of bins, during transportation from the

feed factory to the farm, and during distribution in the

feeding system at farms (Thomas & Van der Poel 1996;

Sørensen et al 2009) Pellets with high durability form

fewer small particles and fines during bagging and storage–

and finally, show low degradation in pneumatic feeding

devices when fed to fish (Aarseth et al 2006a; Sørensen

et al 2009; Aas et al 2011a) Different devices have been

developed to assess durability of pellets, however, most of

these cannot be used on oil-coated high energy extruded

feed

The tumbling box method is an accepted standard in the

feed industry in North America (ASAE 1997) and is used

to simulate formation of fines during mechanical handling

This method uses 500 g of sifted pellets The pellets are

placed in a box that revolves for a period of 10 min at a

speed of 100 rpm After testing, the pellets are screened on

a mechanical sieve shaker with a sieve size of about 0.8

times the pellet diameter The tumbling box pellet

durabil-ity index (PDI) is calculated as the mass of the pellets

retained on the screen divided by the total mass of pellets

The Holmen durability tester has been developed to testeffects of impact and shear forces during pneumatic convey-ing and is commonly used in Europe A sample size of 100 g

of sifted pellets is conveyed with high air velocity through atube with right angled bends for 30–120 s, simulating thefeed handling process of pellets subjected to impact andshear forces Fracture occurs when pellets strike the right-angle corners of the tester The Holmen PDI is calculatedusing the same procedure as for the tumbling box

The LignoTester is another testing device used to late degradation caused by shear forces during pneumaticconveying This procedure uses a sample of 100 g of siftedpellets and blows them around a perforated chamber for

simu-30 s Fines are removed as they are formed and pelletscome out at the end of the cycle The remaining pellets areused to calculate the Ligno PDI

The Norwegian fish feed industry has developed a newdevice, the DORIS Tester, to simulate the stresses thatpellets are exposed to in pneumatic feeding devices Thistesting device is made of an Archimedes screw that feedspellets into a vane and is simulating degradation by impactand shear (Aas et al 2011a) A sample size of 350 g is used

in the procedure and fines, small fractures and wholepellets are screened and separated on a mechanical sieveshaker with a rack of sieves (Aas et al 2011a)

The greatest limitation of the Holmen durability tester andLigno tester when used on high energy salmon diets is leak-ing of oil as the pellets are circulated with the air Because offat leaking, the collected feed samples will have an errone-ously low weight Besides, oil in the testing devices will ham-per proper collection of fine particles Holmen durabilitymeasurements on extruded fish feed are therefore most oftenreported for uncoated feed (Sørensen et al 2009, 2010,2011) The tumbling box durability reveals little differ-ences in durability on extruded feed, even with the adjustedprocedure adding hexagonal nuts to increase agitation(Table 1; Sørensen et al 2010) Durability of uncoated pel-lets measured with the Holmen test, and use of the compres-sion test to measure strength at rupture, are better methodsthan the tumbling box to evaluate differences in physicalquality in extruded fish feed (Sørensen et al 2010) TheDORIS has shown a high correlation with Holmen durabil-ity, and is therefore the most appropriate method for coatedhigh energy pellets (Sørensen et al 2011)

Water stability of feed is an important quality trait forslow eating aquatic animals when the feed has to be soaked

.

in water for hours with minimum leaching of nutrients.For fish that feed on a slow sinking pellet, water stabilitymay be important to mimic the degradation pattern of feed

in the gastrointestinal tract A procedure to determinewater stability over time was described by Baeverfjord

et al (2006) In brief the latter authors used 10 g samples

of pellets that were placed into circular wire netting basketswith 3 mm mesh size and a diameter of 8 cm The test iscarried out in triplicate Baskets with feed samples wereplaced in 600 mL beakers and 300 mL of tap water wasadded The beakers were then incubated in a water bath at

23°C and subjected to 100 shakings per min for 30, 60,

120 and 240 min, respectively When terminating the bation, the baskets were gently dried with paper tissues andweighed before the baskets were placed in a heating cabinet

incu-at 105°C for 18 h After drying, the baskets were againweighed to determine the residual dry matter in each bas-ket The Water stability was calculated as the difference in

DM weight before and after incubation in water divided by

DM weight of the feed before incubation

Sinking velocity is measured by dropping pellets one byone from a height of 5 cm above the water surface into thecentre of a transparent tube with a diameter of 300 mmand a height of 200 cm (Lekang et al 1991) The tube isfilled with tap water of drinking quality, or water with apredefined salinity Because temperature and salinity bothare factors that interfere with sinking velocity, they should

be constant during the course of the test Therefore, waterwith defined salinity should be left for 24 h to achieve aconstant temperature prior to the test The sinking speed ismeasured with a stop-watch over a distance of 150 cmbetween two fixed points 10 and 160 cm below the watersurface, respectively Single pellets are randomly selectedfor sinking rate measurements, and sinking velocity isrecorded as cm s 1

Bulk density is an important property that determines atability or sinking velocity of pellets (Chevanan et al

flo-2007, 2009), and is directly related to the degree of sion during extrusion (Glencross et al 2011b) A floatingpellet is more expanded and has a lower bulk density com-pared to a sinking pellet Bulk density of pellet needs to beadjusted according to feeding management practices andfeeding habits of the target species, and usually a bulk

density greater than 525 g L is needed for sinking pellets

in seawater 35 g L 1(Glencross et al 2011b) Bulk density

is analyzed by filling up a measuring cylinder of known

volume Pellets are carefully poured into a tared cylinder

until a pile of feed has developed on top A scraper is used

to remove the excess feed by pulling once gently over the

edge of the cylinder The content of the full cylinder is then

weighed on a balance In order to standardize the

proce-dure, pellets should be poured from a funnel, preloaded

with feed, and the cylinder should not be tapped prior to

weighing (Aarseth et al 2006b) Each measurement should

be carried out in triplicate, and bulk density for each

repli-cate is calculated as mass of the sample to the unit volume

of the sample (g L 1) Volumetric displacement methods

can also be used to measure specific density of the pellets

with improved accuracy (Draganovic et al 2011)

Fat absorption capacity of the pellets during top dressing of

oil in vacuum coating systems is another pellet property

that is highly correlated with expansion ratio (Sørensen

et al 2010) Expansion of the pellets is correlated to oil

absorption capacity, however, the bulk density needs to be

optimized in order to ensure sinking pellets in case of

salmonids Because energy content in feed is highly

corre-lated to growth performance and feed utilization in Atlantic

salmon (Hillestad & Johnsen 1994) the commercial grow

out diets used in Atlantic salmon production in Norway

contain up to 400 g kg 1fat A method used to analyze fat

absorption capacity is described in Sørensen et al (2011)

Oil in surplus is added to the vacuum coater When all the

oil is distributed onto the feed, the pressure is returned back

to atmospheric pressure, and the oil is retained inside the

pellets Surplus oil is removed by gently compressing the

pellets between of 4–5 layers of oil-absorptive linings Oil

absorption is determined as weight increase of sample (g)

divided by initial weight of sample (g) times 100

Oil leaking can be determined as the loss of oil from

pel-lets Leaking of oil from high energy diets is a problem

because of lowered energy content and different nutritional

profile Besides, an oil layer in the pipeline used for

pneu-matic conveying may act as a sticky layer leading to build-up

of small feed particles If a layer of oil and small particles is

allowed to build up, this will eventually result in blocked

pipes Oil leaking from the pellet should therefore be

avoided Research has shown that oil leaking is affected by

choice of ingredients and processing conditions (Øverland

et al 2007; Sørensen et al 2011) Sørensen et al (2010)

found that oil leaking was not associated with oil level in thefeeds, but was related to feeds with low absorption capacity

Neither was oil leaking correlated with expansion of thepellet or other physical quality parameters Most likely oilleaking is related to microstructure of the pellet Differentprocedures have been used to assess oil leaking Øverland

et al (2007) placed 1 kg coated feed in plastic bucketsequipped with an absorbing plastic coated lining in the bot-tom The bucket was closed with a lid and stored at roomtemperature (20–22 °C) After one week of storage the feedwas emptied from the bucket, and the fat runoff from thepellets was weighed in order to measure fat leaking Anotherand faster method is to incubate smaller amounts of coatedpellets, 100 g, at 40°C in a heating cabinet for 24 h (Søren-sen et al 2010) The pellets are placed on an absorptivelining in a plastic box After incubation, all pellets and dustare removed from the box, and the new weight of the boxand blotting paper is weighed on a balance No investigationhas been carried out to compare the two methods reported

by Øverland et al (2007) and Sørensen et al (2010)

Modern aqua feeds are made from a range of ingredientsthat are combined to meet the nutritional requirements

The feed mash undergoes significant changes during thecourse of the conditioning and extrusion process as it isheated, kneaded and sheared This process may affect thenutrient digestibility, growth rate and feed conversion effi-ciency of the feed A temperature higher than 100°C isneeded in order to flash off steam and expand the feed as

it leaves the die Effects of extrusion temperature on ibility and utilization of fish meal based diets by rainbowtrout were examined in two different experiments (Sørensen

digest-et al 2002, 2005) In the experiment of Sørensen et al

(2002) a fishmeal-based diet was extruded with a screw extruder at three temperatures (100, 125 and 150°C)

twin-In another experiment (Sørensen et al 2005), a based diet was extruded using a single screw extruder attwo temperatures (100 and 140°C) The two studiesshowed that extrusion processing with temperatures in therange from 100 to 150°C did not affect digestibility of pro-tein, individual amino acids or energy for fish meal baseddiets Neither was the feed conversion or net accumulationefficiency of protein and energy significantly different introut fed diets extruded at 100 and 140°C The growthrate was, however, improved at the highest extrusiontemperature in the latter study In line with these results,

fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal- fishmeal-.

Barrows et al (2007) found no significant effect of

extru-sion temperature on apparent digestibility of protein,

organic matter, lipid, energy or carbohydrate in diets

con-taining soybean meal for rainbow trout Increased retention

time in the extruder negatively affected feed intake and

weight gain, whereas high temperature (127°C) resulted in

improved feed conversion rate (FCR) compared to the

lower temperature of 93 °C (Barrows et al 2007) In

con-trast to Barrows et al (2007) improved digestibility of most

major nutrients and amino acids was reported in Atlantic

salmon fed diets with defatted soybean meal when the

extrusion temperature was increased from 110 to 150°C

during processing (Morken et al 2012) The latter

investi-gation was carried out using higher extrusion temperatures

compared to Barrows et al (2007) and this may explain

the differences in effects on nutrient digestibility These

studies with fish meal as protein source either alone

(Søren-sen et al 2002, 2005) or in combination with soy proteins

(Barrows et al 2007; Morken et al 2012) showed that feed

extruded at the highest temperature was better utilized

either due to higher availability of nutrients, higher

utiliza-tion of the nutrients, or a favourable feed structure that

stimulated feed intake Damage to proteins during heat

processing is a function of temperature, time, moisture and

the presence of reducing substances (Papadopoulos 1989)

The relatively high moisture content (250–300 g kg 1

)combined with short duration of exposure (0.5–2 min)

implies that extrusion is not detrimental to the nutritional

value of the feed as long as the temperature does not

exceed 150 °C

Moisture content is of critical importance in order to

main-tain nutritional quality during heating In order to prevent

losses of essential nutrients, a moisture content of 250–

300 g kg 1 during wet extrusion of diets for fish and pets

has been recommended (Rokey 1994) Improved growth

performance of shrimp fed diets extruded at high moisture

contents in comparison to ‘dry’ extrusion conditions

emphasizes the significance of moisture during processing

(Obaldo et al 2000) Low moisture content in combination

with heating of fish proteins to temperatures higher than

100°C, increased cross-linking between amino acids

(Ops-tvedt et al 1984; Finley 1989) causing reduced digestibility

of nearly all amino acids, especially cysteine (Andorsdottir

1985; Ljøkjel et al 2000) Reduced digestibility of cysteine

was also shown in rainbow trout when water addition to

the extruder was restricted, compared to when the diet wasproduced at elevated moisture contents (Sørensen et al.2002) Cysteine reacts readily during heat treatment toform disulphide bonds between cysteine units (Bender1978) The reduction in cysteine digestibility in heat treatedproteins has been explained by the formation of SS bonds,assumed to be resistant to proteolytic cleavage (Friedman1982) In contradiction, Aslaksen et al (2006) reported noeffect on protein digestibility or absorption of disulphideproducts in mink, despite an increased content of disul-phide bonds in soybean meal extruded diets These resultssuggest that the content of disulphide bonds and cross linkedamino acids probably need to exceed a threshold level beforenegative effects are noted on protein digestibility

Production methods used to produce high quality pelletsare challenging and investigation of the physical quality ofcommercial fish feed pellets, i.e pellet hardness, durability,sinking velocity and water absorption, has shown that thequality varies (Chen et al 1999) Although extrusion tech-nology is used with the purpose of producing feed withhigh physical quality, recent research has shown that physi-cal quality varies with diet formulation (Glencross et al

2010, 2011a; Draganovic et al 2011; Kraugerud & Svihus2011; Kraugerud et al 2011), extruder configuration(Sørensen et al 2009, 2010), and processing parameters(Kraugerud & Svihus 2011; Kraugerud et al 2011; Morken

et al 2012; Sørensen et al 2011) All variables includingfeed composition and processing conditions that interferewith expansion of the product influence pellet structure anddurability An inverse relationship between expansion ofthe pellet and physical quality reported as hardness(Fig 1a) or durability (Fig 1b) has been reported in anumber of studies (Aarseth et al 2006b; Hansen & Store-bakken 2007; Sørensen et al 2009, 2010, 2011; Glencross

et al 2011b; Kraugerud & Svihus 2011; Kraugerud et al.2011; Morken et al 2012)

Screw configuration directly influence the cooking andtransformation of the feed dough inside the extruder barrel(Barres et al 1990), due to changes in residence time (Olkku

et al 1980), degree of filling, energy input to the material(Erdemir et al 1992; Yam et al 1994; Sørensen et al.2011), and shear rate (Gogoi et al 1996) Changing theconfiguration of the screws can therefore be used to manip-ulate expansion, bulk density, hardness and durabilitytowards targeted values (Gogoi et al 1996; Sørensen et al

.

2010) Sørensen et al (2010) investigated the effect of three

different screw configurations that all generated unique

SME’s Screw configuration had a significant impact on the

strength and Holmen durability of the feed, however, there

were no correlation between SME and these quality

charac-teristics when different carbohydrate sources were used

Although variation was observed in Holmen durability and

strength, the same ranking was noted among carbohydrate

sources for three different screw configurations Screw

speed (RPM) in the extruder is directly correlated with

SME Greater expansion ratio is reported with increasing

RPM (Sørensen et al 2011) Optimizing pellet quality by

manipulation of RPM to change SME was also investigated

by Kraugerud et al (2011) and Kraugerud & Svihus (2011)

A total of 11 diets in which fish meal was partly replaced

with either plant protein ingredients or starch rich

ingredi-ents were produced at four different SMEs (Fig 2a,b)

Dif-ferent SMEs were obtained by adjusting moisture and

RPM of the extruder Feed containing different ingredients

showed great variation in physical quality, such as hardness

(Fig 2a,b) and responded differently to changes in RPM

Differences in torque were also recorded among the diets

suggesting that the diets formed different viscosities during

the extrusion process The recorded SME therefore varied

and the measured responses had moderate correlation to

SME Torque is associated with friction, or resistance to

flow, and is affected by RPM, fill and viscosity of the rial in the screw channel (Harper 1989) This study clearlydemonstrated the difficulty of producing pellets with pre-dictable physical quality The overall conclusion from theexperiment was that extrusion processing parametersneeded to obtain targeted physical qualities need to bebased on specific knowledge of each ingredient

mate-Temperature during extrusion has an impact on the tinization of starch and denaturing of proteins Elevated

Figure 1 Relationship between expansion ratio and hardness (a)

and expansion ratio and durability (b) (Hansen & Storebakken

2007).

0 20 40 60 80 100 120 140

Wheat starch

Peas Whole bean Dehulled

bean Wheat Oats Starch rich ingredients

(a)

(b)

0 20 40 60 80 100 120 140

Fish meal Corn gluten Soybean

meal Sunflower Lupin Rapeseed Protein rich ingredients

Standard Production Low High

Standard Production Low High

Figure 2 Differences in hardness (a, b) in pellets processed from either starch rich ingredients (a) or plant protein ingredients (b) at different SME’s achieved mainly by adjusting screw speed (RPM) and moisture level The study was divided into three feed produc- tion trials: (1) feed production at “standardized parameters’, (2) feed production at “commercial parameters’, (3) feed produced at two different SME levels The standardized processing conditions aimed to study how diets with different ingredients responded to identical treatment The flow rate of the extruder was kept con- stant and, the screw speed was set to 400 rpm and total moisture

to 308 g kg 1 For feed produced at ‘commercial parameters’ the extrusion process was optimized to give a bulk density of

460 g L 1 and a pellet diameter of 5 mm The targeted bulk sity was obtained by adjusting (releasing) steam pressure from sec- tion four in the extruder The High (45.4) and low (36.5) SME treatments were carried out by adjusting moisture and screw speed (Kraugerud et al 2011).

den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den- den-.

temperature in front of the die reduces the viscosity and

increases the steam pressure of the melt, which in turn

directly influence expansion of the extruded high energy

pellets (Aarseth et al 2006b; Morken et al 2012) The

lat-ter authors showed that pellets became more brittle due to

increased expansion and less resistant to mechanical

stres-ses with increasing extruder temperature

Functional properties of ingredients may be defined as the

ingredients’ ability to form pellets with defined physical

quality in terms of durability, hardness, density, oil

absorp-tion capacity and oil leaking Physical quality of pellets

depends on the bonding between particles and varies with

ingredient composition, particle size distribution of the

ingredients, conditioning, cooling and drying (Behnke

1996) Rumpf (1962) suggested that binding with or without

solid bridges is the main mechanism of bonding of particles

Bonds without a solid bridge may be important when

parti-cles are brought close together, whereas bonds with solid

bridge are reported to be important in pellets (Kaliyan &

Morey 2010) These bonds can, for example, be formed by

diffusion of molecules between particles at the contact

point, crystallization of particles from different ingredients,

chemical reactions, and solidification of melted components

(Kaliyan & Morey 2010) Water added as a liquid or steam

is activating natural binders such as soluble carbohydrates,

starch, proteins and minerals Moreover, moisture itself is a

film binder forming bonds between particles by weak forces

such as hydrogen bonds and van der Waals forces (Pietsch

2002) Inherent binding characteristics differ among

ingredi-ents depending on chemical constituingredi-ents and functional

properties (Behnke 1996; Cavalcanti & Behnke 2005a,b;

Lundblad et al 2009; Sørensen et al 2009, 2010, 2011;

Glencross et al 2011a,b; Kraugerud & Svihus 2011;

Kra-ugerud et al 2011), and is associated with water absorption

capacity of ingredients (Hemmingsen et al 2008) Small

particle size distribution in the mash also improves physical

pellet quality because smaller particles more readily absorb

moisture compared to large particles and are easier to

agglomerate (Hemmingsen et al 2008; Kaliyan & Morey

2009) Based on the inherent binding capability and effects

on pellet quality, different ingredients were assigned a

pelle-tability index (MacMahon & Payne 1991) Carbohydrates

and proteins are structure forming materials (Guy 2001)

that may improve the quality of pellets, whereas fat is a

lubricant reducing pellet quality (Guy 2001; Cavalcanti &Behnke 2005a,b; Morken et al 2012) Although extrudersystems are flexible and can produce pellets with highquality from ingredients with low pelletability index, feedconstituents such as starch, protein, fibre and fat affectphysical properties such as strength, durability and expan-sion ratio Because modern extruded fish feed is made from

a mixture of ingredients containing chemical componentswith different functionality, it is challenging to produce pel-lets with predictable physical quality Variation in physicalquality parameters of extruded feeds containing marineingredients, plant protein ingredients or different starchsources is shown in Table 1

Starch is added to feed for carnivorous fish primarily as

a partly digestible binder and to facilitate expansion of thefeed Kraugerud et al (2011) reported a moderate but posi-tive correlation between amount of added purified wheatstarch and expansion, durability and hardness No correla-tion was, however, found among total starch level (addedpurified starch and starch derived from the ingredients) andthese parameters These findings suggest that starch is notthe most important binder in extruded feed and are sup-ported by previous research with pelleted broiler feeds(Zimonja & Svihus 2009) Starch sources with good bindingand expansion properties are needed because starch is kept

to a minimum in diets for salmonids due to the low ity of carnivorous fish to digest and metabolize starch(Hemre et al 2002) The functional properties of starch areactivated by gelatinization, a process that is mediated byfree water and elevated temperature in the system Use ofpregelatinized starch improved pellet quality both in steampelleting (Wood 1987; Zimonja & Svihus 2009) and extru-sion (Sørensen et al 2010) compared to native starch.Therefore it has been generally accepted that high gelatini-zation of starch in extruded pellets improves pellet quality.However, a recent study comparing effects of different feedprocessing techniques (steam pelleting at low and high tem-perature, expander conditioning prior to pelleting andextrusion processing) on nutritional and physical quality offeed, Lundblad et al (2011) showed that the extruded feedhad the highest starch gelatinization and the lowest PDI.The authors concluded that this was mainly because theextruded diets were 25% more expanded compared to thepelleted diets The greatest improvements in extruded com-pared to steam pelleted diets are therefore higher waterstability and possibilities to adjust bulk density and fatabsorption capacity The functional properties differ amongstarches due to differences in amylose-amylopectin ratio,size and shape of the starch granule and other chemical

capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac- capac-.

components associated with the granule A high amylose

content increases the visco-elastic behaviour and flow

prop-erties of melted starch (Chinnaswamy 1993; Kokini 1993)

Wheat of food-grade quality is extensively used by the feed

industry due to high functional and nutritional properties

However, recent research has shown that replacing wheat

with potato starch improved durability and hardness of the

extruded fish feed, whereas pea starch improved durability

but not hardness of extruded fish meal based diets

(Table 1) Greater oil leaking was observed when potato

starch (Sørensen 2011, unpublished results) and pea starch

(Sørensen et al 2011) replaced wheat in fish meal based

diets Feed additives that affect pH in the diet may change

the binding properties of starch In a recent study Morken

et al (2012) demonstrated that formic acid added to

soy-based diets most likely reduced the hydrophilic properties

of starch resulting in greater expansion and reduced

bind-ing strength with increasbind-ing extrusion temperature (Fig 3)

Based on these findings the authors suggested that use of

acidic feed additives should be used with great care in diets

with poor inherent binding properties

Dietary fibre is made up by a mixture of soluble and

insoluble components that are indigestible for carnivore

fish and should be kept at a minimum in diets for these

species Dietary fiber can be classified as water soluble or

water insoluble, and the solubility is related to the

hydro-philic and hydrophobic properties of the repeating

mono-meric unit and bonds of the fibre The source and amount

of fibre affect quality of the pellets (Hsieh et al 1989; Lue

et al 1990; Hansen & Storebakken 2007; Kraugerud et al

2011) measured as viscosity, bulking effect, water holding

capacity and ability to form gel Water soluble fibres mayimprove the structural integrity of pellets by increasing theviscosity and thereby embedding coarser particles into anetwork (Thomas et al 1997) Recent research have shownincreased durability and pellet hardness with increasing cel-lulose inclusion in the diet (Hansen & Storebakken 2007;

Kraugerud et al 2011) The latter authors also suggestedthat the effect of fibre was an indirect effect associated withreduction in expansion mediated through either disruption

of cell walls, or through a reduction of die swell as theextrudate leave the extruder (Guy 2001) Water insolublefibres may prevent bonding between particles due to resil-ience characteristics (Thomas et al 1998) Water insolublefibres that entangle and fold between particles may result

in weak spots and more fragmentation of the pellet Refstie

et al.(2006) reported that the diets containing lupin kernelmeals had greater durability and were harder compared tolupin protein concentrates The differences between the ker-nel meals and the concentrates are mainly that concentratescontain less fibre than the kernel meal (Glencross et al

2005)

The functionality of protein as a binder is associatedwith the structure of the protein, which is mainly deter-mined by the amino acid composition and processing his-tory Proteins in native or undenatured state have betterbinding properties than heated or denatured proteins

Wood (1987) found that inclusion of raw and unprocessedprotein, rather than denatured protein, improved pellethardness and pellet durability Improved hardness was alsoobserved for extruded feed when untoasted rather thantoasted, soybean meal replaced fish meal in the diet (Søren-sen et al 2009) Native proteins have a higher solubilitythan denatured proteins, and solubility of main feed con-stituents appears to be an important factor for the quality

of the final product Heat treated proteins are alreadyunfolded and form less soluble aggregates Several investi-gations have reported improved hardness and durability ofpellets when fish meal was replaced with less heat treatedplant proteins such as soybean meal, pea protein concen-trate, sunflower meal, lupin, rapeseed meal and wheat glu-ten (Table 1; Fig 2; Draganovic et al 2011) However,Draganovic et al (2011) reported that although wheat glu-ten and soy protein concentrated was positively associatedwith hardness of the pellets, fish meal had unique func-tional properties not present in the plant ingredients used

in the experiment Fish meal contains mainly fibrous teins that has been through several steps of heating prior

pro-to processing inpro-to fish feed, whereas plant proteins aredominated by globular proteins When protein is exposed

Holmen durability index

Expansion ratio

Figure 3 Holmen durability and expansion ratio in diets with

soy-bean meal supplemented with ( +) or without ( ) sodium diformate

extruded at 100, 130 or 150 °C Increasing extrusion temperature

gave higher expansion and reduced Holmen durability, in

particu-lar for the diets added sodium diformate (Morken et al 2012).

.

to heat it denatures Unfolding of the protein starts at

45 °C and during cooling the protein re-associates and new

bonds are established between feed particles (Thomas et al

1997) Repeated heating of proteins reduces the gluing

effect (Wood 1987; Sørensen et al 2009)

Internal fat level in the feed mash to be processed has an

impact on pellet quality (Cavalcanti & Behnke 2005a,b;

Morken et al 2012) Use of oil rich ingredients such as full

fat soybean meal instead of defatted soybean meal

(Morken et al 2012), or fat levels above 12% during the

extruder process have a negative effect on pellet durability

(Rokey 1994) The lubricating effect of fat prevents efficient

conveying of the dough in single screw extruders and also

reduces the mechanical dissipation of energy Because of

less shearing force, the temperature needed for starch

gelati-nization should be increased with elevated internal fat levels

(Lin et al 1997) A reduction in shear stress is associated

with decreased plastification and cooking of the melt (Lam

& Flores 2003), resulting in lower quality of the pellet

Water has an important role for the integrity and physical

quality of the pellets (Lundblad et al 2009; Mathew et al

1999; Draganovic et al 2011) Kaliyan & Morey (2009)

sug-gested that free moisture between particles wets and spreads

in the interstices between particles causing cohesive forces

due to liquid bridges, holding particles together by capillary

and viscous forces Differences in chemical composition and

physical conditions on the surface of the ingredients, affects

the ability of different ingredients to absorb moisture as

water or vapour (Hemmingsen et al 2008) The latter

authors showed that greatest water absorption was observed

for finely ground starch rich ingredients hydrated in water

at 80 °C The protein rich ingredients showed a faster

hydration than the starch rich ingredients, and they were

more efficiently hydrated in moist air at 80°C The moisture

absorption capacity was less affected by particle size and

water temperature in the protein rich ingredients The

rela-tively limited addition of water into the preconditioner

(<30 W/W) suggests that there is a competition for water

and limited access for e.g full gelatinization of starch during

the extrusion process Gelatinization of starch is usually

defined as a phase transition in excess water Lund (1981)

suggested that a water :starch ratio of about 1.5 : 1 is

needed for complete gelatinization of starch Extrusion

sys-tems with water usually ranging from 15–40% are therefore

providing melting rather than gelatinization of the starch

(Qu & Wang 1994) Nevertheless, transformation of starch

by extrusion processing is referred to as gelatinization, and

is reported to range between 73–100% in extruded fish feed

(Hansen et al 2010; Kraugerud et al 2011)

A few studies have investigated the effects of physical ity on feed intake and nutrient utilization in fish (Hilton

qual-et al 1981; Obaldo et al 1999; Baeverfjord et al 2006;Barrows et al 2007; Venou et al 2009; Aas et al 2011b).Two of these studies showed that extruded diets gave pro-longed gastric emptying rate compared to steam pelleteddiets, and consequently reduced feed intake (Hilton et al.1981; Venou et al 2009) Hilton et al (1981) made a com-parison between extruded and steam pelleted diets andfound that extruded pellets were more durable, had higherwater stability and absorbed more water than steam pellets.Baeverfjord et al (2006) found no significant differences infeed intake and growth in rainbow trout fed diets with high

or low water stability The apparent digestibility of gen and lipid tended to be highest in feed with high waterstability in the latter study In line with this, Aas et al.(2011b) reported that apparent digestibility coefficients ofamino acids, starch, energy and dry matter was highest inthe feed with highest pellet hardness and water stability Incontrast to Baeverfjord et al (2006) a 23% higher feedintake was observed when rainbow trout were fed a dietwith low water stability compared to a diet with high waterstability (Aas et al 2011b) These studies suggest thathigher water stability of pellets result in longer gastricretention time and prolonged evacuation time of chymethrough the gastrointestinal tract In line with this hypothe-sis, Adamidou et al (2009) showed that inclusion of fababean and chickpea in diets for seabass increased the hard-ness of the extruded pellet, and these diets also had pro-longed gastrointestinal evacuation time

nitro-In contrast to the studies with rainbow trout ord et al 2006; Aas et al 2011b), reporting reduced feedintake with higher water stability and hardness of the pel-let, Glencross et al (2011b) showed that inclusion of lupin

(Baeverfj-in the diets gave higher hardness and feed (Baeverfj-intake (Fig 4).The contradictive findings suggest that hardness is not agood measure for pellet water stability and may give noindication of potential for feed intake in rainbow trout andAtlantic salmon

Extrusion results in higher gelatinization of starch in parison to steam pelleting due to higher water content,higher temperature and more mechanical energy used duringthe process (Lundblad et al 2011) The higher gelatinization

com-of starch in extruded feed also explains the higher waterabsorption capacity, expansion ratio and lower bulk densitycompared to steam pelleted diets reported by Honorato et al

.

(2010) A lower growth rate reported for silver perch and

pacu juveniles fed extruded diets compared to steam pelleted

diets may be explained by the lower bulk density of extruded

feed (Booth et al 2000, 2002; Honorato et al 2010) A lower

bulk density of extruded compared to steam pelleted diets

may result in a lower intake of energy because the juveniles

reach the point of satiation at a lower energy intake

Physicochemical properties vary among ingredients and

as a result of different processing and affects physical

prop-erties of feed such as water stability and hardness

Extruded pellets made with wheat inclusion were more

water stable than those made with field beans (Aas et al

2011b) and resulted in a lower feed intake Pellets with the

high water stability also need longer time to dissolve in the

stomach (Aas et al 2011b; Glencross et al 2011a)

Adami-dou et al (2009) observed longer gastrointestinal

evacua-tion time when faba bean and chickpea were included in

the diet compared to a wheat meal based control diet Feed

intake was, however, not recorded in the experiments

reported by Adamidou et al (2009) and Glencross et al

(2011a) Changes in gastrointestinal evacuation rate were

also reported in experiments with monogastrics as a result

of different ingredients (Fledderus et al 2007; Rosenfeld &

Austbø 2009) or processing of the diet (Amornthewaphat

& Attamangkune 2008; Rosenfeld & Austbø 2009) Such

changes in the gastrointestinal retention time are most

likely mediated through intestinal viscosity (Fledderus et al

2007), though changes in viscosity appear to be a less

sensi-tive method in comparison to faecal dry matter content in

salmonids (Refstie et al 1999; Aslaksen et al 2007)

Use of extrusion systems to produce aquafeeds enables duction of pellets with high physical quality for a range ofingredients Production of pellets with predictable quality

pro-is, however, challenging because physical quality is a result

of the interaction between processing conditions and dients Functional property is therefore an important qual-ity aspect that should be taken into consideration wheningredients are evaluated (Glencross et al 2007) Unfortu-nately, only a few publications report physical quality offeed, and the procedures differs among the labs A stan-dardization of methods is needed to build up a knowledgebase describing interactions between ingredients and pro-cessing on physical quality and possible implication forfeed intake and growth performance

ingre-This review has addressed variation in physical quality

of pellets and effects of ingredient composition as well asprocessing parameters In brief, different publications havereported that:

 Bulk density in extruded high energy diets needs to behigher than 525 g L 1 to avoid floating pellet in seawater(salinity∼35 g L 1

; Glencross et al 2011b)

 Hardness of pellets vary with ingredient composition,pellet diameter and expansion ratio Hardness rangedbetween 20 and 40 N in pellets with a diameter of 5–6 mm

Plant ingredient based pellets of this size usually showedhigher hardness ranging from 40 to 80 N

 Durability can be analyzed with a great variety of ing devices Durability of the same diet may range from44–99% depending on testing device (Sørensen et al 2010)

test-Using a Holmen durability testing device on uncoated fishmeal based diets, PDI values ranged between 40 and 80%,while durability of diets containing plant protein ingredi-ents showed higher values The DORIS tester is recom-mended as a standard for high energy (coated) extrudedpellets, because of the accuracy and the ‘easy to operate’

procedure

 Expansion ratio ranged between 20 and 30% for asinking pellet depending on the amount of fat targeted inthe final feed

 Starch gelatinization was reported to vary between 73and 100% in extruded fish feed (Hansen et al 2010; Kra-ugerud et al 2011)

This literature review has also revealed a scarcity of ature reporting physical quality of pellets smaller or largerthan 4–7 mm The literature reporting physical quality isdominated by 4–7 mm pellet diameter because most feeds

Lupin Myallie Lupin Wodjil SBM

Ingredient and inclusion level (%)

Figure 4 A comparison of hardness (g) and feed intake (g fish 1

)

in diets containing two genotypes of lupin or defatted soybean

meal at increasing inclusion level (Glencross et al 2011b).

.

were made for feeding experiments with small fish

Com-mercial fish feed producers in Norway have reported that

this pellet size is the easiest to produce and usually have

the least complaints from the customers Density of the

pel-let to ensure sinking is often the major challenge in feed

smaller than 3 mm in diameter, while feed particles larger

than 10 mm often cause problems with physical integrity

Big feed pellets are usually the most expanded because they

need to absorb the largest amount of oil Consequently, fat

leaking and small particles are common problems in high

energy extruded fish feed

Most experiments evaluating alternative ingredients or

new diets are reported without a detailed description of the

equipment and processing conditions used Variable

pro-cessing conditions may explain different digestibility of

energy and nutrients (Hilton et al 1981; Venou et al 2009;

Glencross et al 2011c) A more systematic reporting of

processing parameters and ingredient functionality is also

important to gain more knowledge about the main factors

affecting physical quality of extruded fish feed In order to

describe the feed manufacturing process, information about

the equipment and procedure is needed These include

grinding, mixing, conditioning, type of extruder (or other

processing equipment) and drying Processing condition

parameters such as screw configuration, temperature,

mois-ture added, temperamois-ture profile in the extruder, RPM,

pres-sure (in front of the die), SME and torque are also of

pivotal importance in order to understand the behaviour of

diets during processing

Research has shown that extrusion processing have no

negative effect, at least when dry conditions are avoided,

and even improves nutrient digestibility and feed utilization

when plant protein ingredients are included in the diets for

salmonids However, more research is needed to investigate

the possible interactions between physical quality of feed,

feed intake and utilization of the diet Differences in

physi-cal quality may affect gastric emptying rate and feed

intake, and conclusions about palatability of ingredients

may easily be confounded by different physical quality of

the feed Whether these effects are ingredient specific, or

directly caused by the physical properties, still needs to be

resolved In vitro water stability test (Baeverfjord et al

2006) appears to correlate well with loss of integrity in

in vivopellet integrity analysis (Aas et al 2011b; Glencross

et al 2011a) There are indications that feed with high

water stability result in lower feed intake (Aas et al

2010b)

In conclusion, analysis of physical quality of

experimen-tal feeds should be carried out routinely, in particular when

new ingredients are investigated or if processing conditionsdiffers among the feeds It is concluded that standardizedand suitable methods are needed to evaluate the physicalquality of extruded fish feed

This work was financially supported by Aquaculture tein Centre, CoE (Project No 14949 from The ResearchCouncil of Norway) Dr Karl D Shearer and Professor

Pro-Dr Anders Skrede are highly appreciated for their valuablecomments to the manuscript

Aarseth, K.A (2004) Attrition of feed pellets during pneumatic conveying: the influence of velocity and bend radius Biosyst Eng., 89, 197–213.

Aarseth, K.A., Perez, V., Bøe, J.K & Jeksrud, W.K (2006a) Reliable pneumatic conveying of fish feed Aquacult Eng., 35,

14 –25.

Aarseth, K.A., Sørensen, M & Storebakken, T (2006b) Effects of red yeast inclusions in diets for salmonids and extrusion temper- ature on pellet tensile strength: Weibull analysis Anim Feed Sci Techn., 126, 75–91.

Aas, T.S., Oehme, M., Sørensen, M., He, G., Lygren, I & A˚sga˚rd,

T (2011a) Analysis of pellet degradation of extruded high energy fish feeds with different physical quality in a pneumatic feeding system Aquacult Eng., 44, 25–34.

Aas, T.S., Terjesen, B.F., Sigholt, T et al (2011b) Nutritional responses in rainbow trout (Oncorhynchus mykiss) fed diets with different physical qualities at stable or variable environmental conditions Aquacult Nutr., 17, 657–670.

Adamidou, S., Nengas, J., Alexis, M., Foundoulaki, E., oulou, D., Campell, P., Karacostas, J., Rigos, G., Bell, G.J & Jauncey, K (2009) Apparent nutrient digestibility and gastroin- testinal evacuation time in European seabass (Dicentrarchus lab- rax) fed diets containing different levels of legumes Aquaculture,

Nikolop-289, 106–112.

Amornthewaphat, N & Attamangkune, A (2008) Extrusion and animal performance effects of extruded maize quality on digest- ibility and growth performance in rats and nursery pigs Anim Feed Sci Technol., 144, 292–305.

Andorsdottir, G (1985) Protein quality, methionine tion and fat levels in starter diets for salmon fry Cand Scient thesis, University of Oslo, pp 109.

supplementa-ASAE (1997) Cubes, pellets, and crumbles: definitions and ods for determining density, durability, and moisture ASAE Standard S269.4 Agricultural Engineers Yearbook of Standards American Society of Agricultural Engineering, St Joseph, MI, USA.

meth-Aslaksen, M.A., Romarheim, O.H., Storebakken, T & Skrede, A (2006) Evaluation of content and digestibility of disulfide bonds and free thiols in unextruded and extruded diets containing fish meal and soybean protein sources Anim Feed Sci Technol.,

128, 320–330.

Aslaksen, M.A., Kraugerud, O.F., Penn, M., Svihus, B., Denstadli, V., Jørgensen, H.Y., Hillestad, M., Krogdahl, A˚ & Storebak- ken, T (2007) Screening of nutrient digestibilities and intestinal

.

pathologies in Atlantic salmon, Salmo salar, fed diets with

legumes, oilseeds, or cereals Aquaculture, 272, 541–555.

Baeverfjord, G., Refstie, S., Krogedal, P & A˚sga˚rd, T (2006) Low

feed pellet water stability and fluctuating water salinity cause

separation and accumulation of dietary oil in the stomach of

rainbow trout (Oncorhynchus mykiss) Aquaculture, 261, 1335–

1345.

Barres, C., Vergnes, B., Tayeb, J & Della Valle, G (1990)

Trans-formation of wheat-flour by extrusion cooking: influence of

screw configuration and operating-conditions Cer Chem., 67,

427 –433.

Barrows, F.T., Stone, D.A.J & Hardy, R.W (2007) The effects of

extrusion conditions on the nutritional value of soybean meal

for rainbow trout (Oncorhynchus mykiss) Aquaculture, 265, 244–

252.

Behnke, K.C (1996) Feed Manufacturing technology: current

issues and challenges Anim Feed Sci Technol., 62, 49–57.

Bender, A.E (1978) Food Processing and Nutrition Academic

Press, London, pp 243.

Booth, M.A., Allan, G.L & Warner-Smith, R (2000) Effects of

grinding, steam conditioning and extrusion of a practical diet on

digestibility and weight gain of silver perch, Bidyanus bidyanus.

Aquaculture, 182, 287–299.

Booth, M.A., Allan, G.L., Evans, A.J & Gleeson, V.P (2002)

Effects of steam pelleting or extrusion on digestibility and

per-formance of silver perch Bidyanus bidyanus Aquac Res., 33,

1163 –1173.

Cavalcanti, B & Behnke, K.C (2005a) Effect of composition of

feed model systems on pellet quality: a mixture experimental

approach I Cereal Chem., 82, 455–461.

Cavalcanti, B & Behnke, K.C (2005b) Effect of composition of

feed model systems on pellet quality: a mixture experimental

approach II Cereal Chem., 82, 462–467.

Chen, Y.-S., Beveridge, M.C.M & Telfer, T.C (1999) Physical

characteristics of commercial pelleted Atlantic salmon feeds and

consideration of implications for modeling of waste dispersion

through sedimentation Aquac Int., 7, 89–100.

Chevanan, N., Muthukumarappan, K., Rosentrater, K.A &

Jul-son, J.L (2007) Effect of die dimensions on extrusion processing

parameters and properties of DDGS-based aquaculture feeds.

Cereal Chem., 84, 389–398.

Chevanan, N., Muthukumarappan, K & Rosentrater, K.A (2009)

Extrusion studies of aquaculture feed using distillers dried grains

with solubles and whey Food Bioprocess Technol., 2, 177–185.

Chinnaswamy, R (1993) Basis of cereal starch expansion

Carbo-hydr Polym., 21, 157–167.

Draganovic, V., Van der Goot, A.J., Boom, R & Jonkers, J.

(2011) Assessment of the effects of fish meal, wheat gluten, soy

protein concentrate and feed moisture on extruder system

parameters and the technical quality of fish feed Anim Feed Sci.

Technol., 165, 238–250.

Erdemir, M.M., Edwards, R.H & McCarthy, K.L (1992) Effect of

screw configuration on mechanical energy-transfer in twin-screw

extrusion of rice flour Lebensm Wiss Technol., 25, 502–508.

FAO (2011) Aquaculture development 5 Use of wild fish as feed

in aquaculture FAO Technical Guidelines for Responsible

Fish-eries No 5, Suppl 5 FAO, Rome, p 79.

Finley, J.W (1989) Effects of processing on proteins: an overview.

In: Protein Quality and the Effects of Processing (Phillips, R.D.

& Finley, J.W eds), pp 1 –8 Dekker, New York.

Fledderus, J., Bikker, P & Kluess, J.W (2007) Increasing diet

vis-cosity using carboxymethylcellulose in weaned piglets stimulates

protein digestibility Livest Sci., 109, 89–92.

Friedman, M., Grosjean, O.K.K & Zahnley, J.C (1982) tion of soya bean trypsin inhibitors by thiols J Sci Food Agric.,

Inactiva-33, 165–172.

Glencross, B., Evans, D., Dods, K., McCafferty, P., Hawkins, W., Maas, R & Sipsas, S (2005) Evaluation of the digestible value

of lupin and soybean protein concentrates and isolates when fed

to rainbow trout, Oncorhynchus mykiss, using either stripping or settlement faecal collection methods Aquaculture, 245, 211–220.

Glencross, B.D., Booth, M & Allan, G.L (2007) A feed is only as good as its ingredients: a review of ingredient evaluation for aquaculture feeds Aquaculture Nutr., 13, 17–34.

Glencross, B., Hawkins, W., Maas, R., Karopoulos, M & Hauler,

R (2010) Evaluation of the influence of different species and tivars of lupin kernel meal on the extrusion process, pellet prop- erties and viscosity parameters of salmonid feeds Aquacult.

cul-Nutr., 16, 13–24.

Glencross, B., Rutherford, N & Jones, B (2011a) Evaluating options for fishmeal replacement in diets for juvenile barramundi (Lates calcarifer) Aquacult Nutr., 17, e722–e732.

Glencross, B., Rutherford, N & Hawkins, W (2011b) A son of the growth performance of rainbow trout (Oncorhynchus mykiss) when fed soybean, narrow-leaf or yellow lupin meals in extruded diets Aquacult Nutr., 17, e317–e325.

compari-Glencross, B., Hawkins, W., Evans, D., Rutherford, N.,

McCaffer-ty, P., Dods, K & Hauler, R (2011c) A comparison of the effect

of diet extrusion or screw-press pelleting on the digestibility of grain protein product when fed to rainbow trout (Oncorhynchus mykiss) Aquaculture, 312, 154–161.

Gogoi, B.K., Choudhury, G.S & Oswalt, A.J (1996) Effects of location and spacing of reverse screw and kneading element combination during twin-screw extrusion of starchy and protein- aceous blends Food Res Int., 29, 505–512.

Guy, R.C.E (2001) Raw materials for extrusion cooking In:

Extrusion Cooking Technologies and Applications (Guy, R.C.E.

ed.), pp 5 –28, Woodhead Publishing Limited, Cambridge.

Hansen, J.Ø & Storebakken, T (2007) Effects of dietary cellulose level on pellet quality and nutrient digestibilities in rainbow trout (Oncorhynchus mykiss) Aquaculture, 272, 458–465.

Hansen, J.Ø., Penn, M., Øverland, M., Shearer, K.D., Krogdahl, A˚., Mydland, L.T & Storebakken, T (2010) High inclusion of partially deshelled and whole krill meals in diets for Atlantic salmon (Salmo salar) Aquaculture, 310, 164–172.

Hardy, R.W (1989) Practical feeding: salmon and trout In: tion and Feeding of Fish (Lovell, T ed.), pp 185 –203 Van Nostrand Reinhold, New York.

Nutri-Harper, J.M (1989) Food extruders and their applications In:

Extrusion Cooking (Mercier, C., Linko, P & Harper, J.M eds),

pp 1 –15 American Association of Cereal Chemistry, St Paul, MN.

Hemmingsen, A.K.T., Stevik, A.M., Claussen, I.C., Lundblad, K.

K., Prestløkken, E., Sørensen, M & Eikevik, T.M (2008) Water adsorption in feed ingredients at different temperatures, particle size and ingredient combinations Dry Technol., 26, 738–748.

Hemre, G.-I., Mommsen, T.P & Krogdahl, A˚ (2002) drates in fish nutrition: effects on growth, glucose metabolism and hepatic enzymes Aquacult Nutr., 8, 175–194.

Carbohy-Hillestad, M & Johnsen, F (1994) High-energy/low-protein diets for Atlantic salmon: effects on growth, nutrient retention and slaughter quality Aquaculture, 124, 109–116.

Hilton, J.W., Cho, C.Y & Slinger, S.J (1981) Effect of extrusion processing and steam pelleting diets on pellet durability, pellet water absorption and the physiological response of rainbow trout (Salmo gairdneri) Aquaculture, 25, 185–194.

.

Honorato, C.A., Almeida, L.C., Da Silva Nunes, C., Carneiro, D.

J & Moraes, G (2010) Effects of diets with distinct levels of

carbohydrates and lipids: the outcomes on growth of pacu

(Pia-ractus mesopotamicus) Aquacult Nutr., 19, 91–99.

Hsieh, F., Mulvaney, S.J., Huff, H.E., Lue, S & Brent, J (1989)

Effect of dietary fiber and screw speed on some extrusion

pro-cessing and product variables Lebensm Wiss Technol., 22, 204–

207.

Kaliyan, N & Morey, R.V (2009) Factors affecting strength and

durability of densified products Biomass Bioenerg., 33, 337–359.

Kaliyan, N & Morey, R.V (2010) Natural binders and solid

bridge type binding mechanisms in briquettes and pellets made

from corn stover and switchgrass Biores Technol., 101, 1082–

1090.

Kokini, J.L (1993) The effect of processing history on chemical

changes in single- and twin-screw extruders Trends Food Sci.

Technol., 4, 324–329.

Kraugerud, O.F & Svihus, B (2011) Effects of online

pretreat-ment of plant ingredients on processing responses and physical

properties in extruded fish feed Anim Feed Sci Technol., 168,

250 –256.

Kraugerud, O.F., Jørgensen, H.Y & Svihus, B (2011) Physical

properties of extruded fish feed with inclusion of different plant

(legumes, oilseeds, or cereals) meals Anim Feed Sci Technol.,

163, 244–254.

Lam, C.M.D & Flores, R.A (2003) Effect of particle size

and moisture content on viscosity of fish feed Cereal Chem., 80,

20 –24.

Lekang, O.I., Andersen, J., Bøe, J.K & Berre, B (1991) Devices

and methods for measuring physical properties of particulate

material (Utstyr og metoder for ma˚ling av fysiske egenskaper

ved partikkelmateriale) Technical note 106/91 In Norwegian.

Department of Mathematical Sciences and Technology, The

Norwegian University of Life Sciences, A˚s, Norway.

Lin, S., Hsieh, F & Huff, H.E (1998) Effects of lipids and

pro-cessing conditions on lipid oxidation of extruded dry pet food

during storage Anim Feed Sci Technol., 71, 283–294.

Ljøkjel, K., Harstad, O.M & Skrede, A (2000) Effect of heat

treat-ment of soybean meal and fish meal on amino acid digestibility in

milk and dairy cows Anim Feed Sci Technol., 84, 83–95.

Lovell, T (1989) Feed formulation and processing In: Nutrition

and Feeding of Fish (Lovell, T ed.), pp 107 –127 Van Nostrand

Reinhold, New York.

Lue, S., Hsich, F., Peng, I.C & Huff, H.E (1990) Expansion of

corn extrudates containing dietary fiber: a microstructure study.

Food Sci Technol., 23, 165–173.

Lund, D.B (1981) Influence of time, temperature, moisture,

ingre-dients and processing conditions on starch gelatinization CRC

Crit Rev Food Sci Nutr., 20, 249–273.

Lundblad, K.K., Hancock, J.D., Behnke, K.C., Prestløkken, E.,

McKinney, L.J & Sørensen, M (2009) Adding water into the

mixer improves pellet quality and pelleting efficiency in diets for

finishing pigs with and without use of expander Anim Feed Sci.

Technol., 150, 295–302.

Lundblad, K.K., Issa, S., Hancock, J.D., Behnke, K.C.,

McKin-ney, L.J., Alavi, S., Prestløkken, E., Fledderus, J & Sørensen,

M (2011) Effects of steam conditioning at low and high

temper-ature, expander conditioning and extruder processing prior to

pelleting on growth performance and nutrient digestibility in

nursery pigs and broiler chickens Anim Feed Sci Technol., 169,

208 –217.

MacMahon, M.J & Payne, J.D (1991) The Pelleting Handbook.

pp 72, Borregaard Lignotech, Norway.

Mathew, J.M., Hoseney, R.C & Faubion, J.M (1999) Effects of corn sample, mill type, and particle size on corn curl and pet food extrudates Cereal Chem., 76, 621–624.

Mercier, C., Linko, P & Harper, J.M (1989) Extrusion ing, pp 471 American Association of Cereal Chemists, St Paul, MN.

Cook-Morken, T., Kraugerud, O.F., Sørensen, M., Storebakken, T., Hillestad, M., Christiansen, R & Øverland, M (2012) Effects of feed processing conditions and acid salts on nutrient digestibility and physical quality of soy-based diets for Atlantic salmon (Salmo salar) Aquacult Nutr., 18, 21–34.

Obaldo, L.O., Warren, D.G & Gihyung, R.H (1999) Growth, survival and diet utilization of Pacific White Shrimp, Penaeus vannamei, fed dry extruded diets with different levels of shear and moisture J Appl Aquacult., 9, 47–57.

Obaldo, L.G., Dominy, W.G & Ryu, G.H (2000) Extrusion cessing and its effect on aquaculture diet quality and shrimp growth J Appl Aquacult., 10, 41–54.

pro-Olkku, J., Antilla, J., Heikkinen, J & Linko, P (1980) Residence time distribution in a twin-screw extruder In: Food Process Engineering (Linko, P., Malkki, Y., Olkku, J & Larinkari,

J eds), pp 791 –794 Elsevier Applied Science Publishers, London.

Opstvedt, J., Miller, R., Hardy, R.W & Spinelli, J (1984) induced changes in sulfhydryl groups and disulfide bonds in fish protein and their effect on protein and amino acid digestibility

Heat-in raHeat-inbow trout (Salmo gairdneri) J Agric Food Chem., 32,

929 –935.

Øverland, M., Romarheim, O.H., Ahlstrøm, Ø., Storebakken, T.

& Skrede, A (2007) Technical quality of dog food and salmon feed containing different bacterial protein sources and processed

by different extrusion conditions Anim Feed Sci Technol., 134,

124 –139.

Øverland, M., Sørensen, M., Storebakken, T., Penn, M., dahl, A˚ & Skrede, A (2009) Pea protein concentrate substitut- ing fish meal or soybean meal in diets for Atlantic salmon (Salmo salar): effect on growth performance, nutrient digestibil- ity, carcass composition, gut health, and physical feed quality Aquaculture, 288, 305–311.

Krog-Papadopoulos, M.C (1989) Effect of processing on high-protein feedstuffs: a review Biol Wastes, 29, 123–138.

Pietsch, W (2002) Agglomeration Processes – Phenomena, nologies, Equipment, pp 622 Wiley-VCH, Weinheim.

Tech-Pillay, T.V.R & Kutty, M.N (2005) Feed and nutrition In: Aquaculture: Principles and Practices, 2nd edn pp 105 –173 Blackwell publishing, Oxford.

Qu, D & Wang, S.S (1994) Kinetics of the formations of nized and melted starch at extrusion cooking conditions Starch- Sta¨rke, 46, 225 –229.

gelati-Refstie, S., Svihus, B., Shearer, K.D & Storebakken, T (1999) Nutrient digestibility in Atlantic salmon and broiler chickens related to viscosity and non-starch polysaccharide content in different soyabean products Anim Feed Sci Technol., 79, 331– 345.

Refstie, S., Glencross, B., Landsverk, T., Sørensen, M., Lilleeng, E., Hawkins, W & Krogdahl, A˚ (2006) Digestive function and intestinal integrity in Atlantic salmon (Salmo salar) fed kernel meals and protein concentrates made from yellow or narrow- leafed lupins Aquaculture, 261, 1382–1395.

Rokey, G & Huber, G (1994) Extrusion processing of culture feeds In: Feed Manufacturing Technology IV (McEllhiney, R.R ed.), pp 509 –515 AFIA, Inc., Arlington, VA.

aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua- aqua-.

Rosenfeld, I & Austbø, D (2009) Effect of type of grain and feed

processing on gastrointestinal retention times in horses J Anim.

Sci., 87, 3991–3996.

Rumpf, H (1962) The strength of granules and agglomerates In:

Agglomeration (Knepper, W.A ed.), pp 379 –418 John Wiley

and Sons, New York.

Sørensen, M., Ljøkjel, K., Storebakken, T., Shearer, K.D &

Skre-de, A (2002) Apparent digestibility of protein, amino acids and

energy in rainbow trout (Oncorhynchus mykiss) fed a fish meal

based diet extruded at different temperatures Aquaculture, 211,

215 –225.

Sørensen, M., Storebakken, T & Shearer, K.D (2005)

Digestibil-ity, growth and nutrient retention in rainbow trout

(Oncorhyn-chus mykiss) fed diets extruded at two different temperatures.

Aquacult.Nutr., 11, 251–256.

Sørensen, M., Stjepanovic, N., Romarheim, O.H., Krekling, T &

Storebakken, T (2009) Soybean meal improves the physical

quality of extruded fish feed Anim Feed Sci Technol., 149, 149–

161.

Sørensen, M., Luyen, G.Q.N., Storebakken, T & Øverland, M.

(2010) Starch source, screw configuration and injection of steam

into the barrel affect physical quality of extruded fish feed.

Aquacult Res., 41, 419–432.

Sørensen, M., Morken, T., Kosanovic, M & Øverland, M (2011)

Pea and wheat starch possess different processing characteristics

and affect physical quality and viscosity of extruded feed for

Atlantic salmon Aquacult Nutr., 17, e326–e336.

Thomas, M & Van der Poel, A.F.B (1996) Physical quality of leted animal feed 1 Criteria for pellet quality Anim Feed Sci.

pel-Technol., 61, 89–112.

Thomas, M., Van Zuilichem, D.J & Van der Poel, A.F.B.

(1997) Physical quality of pelleted animal feed 2 Contribution

of process and its conditions Anim Feed Sci Technol., 64,

173 –192.

Thomas, M., Van Vliet, T & Van der Poel, A.F.B (1998) Physical quality of pelleted animal feed 3 Contribution of feedstuff com- ponents Anim Feed Sci Technol., 70, 59–78.

Venou, B., Alexis, M.N., Fountoulaki, E & Haralabous, J (2009) Performance factors, body composition and digestion character- istics of gilthead sea bream (Sparus aurata) fed pelleted or extruded diets Aquacult Nutr., 15, 390–401.

Winowiski, T.S (1995) Pellet quality in animal feeds, pp 1 –5.

Technical Bulletin, American Soybean Association, FT21-1995, http://www.asasea.com/technical/ft21-1995.html).

Wood, J.F (1987) The functional properties of feed raw materials and their effect on production and quality of feed pellets Anim.

Feed Sci Technol., 18, 1–17.

Yam, K.L., Gogoi, B.K., Karwe, M.V & Wang, S.S (1994) Shear conversion of corn meal by reverse screw elements dur- ing twin-screw extrusion at low temperatures J Food Sci., 59,

113 –114.

Zimonja, O & Svihus, B (2009) Effects of processing of wheat or oats starch on physical pellet quality and nutritional value for broilers Anim Feed Sci Technol., 149, 287–297.

.

1 2 1 1 3 1

Institut National des Sciences et Technologies de la Mer, Salammbo, Tunisia; 2 University of Namur (FUNDP), Unit ofResearch in Organismal Biology (URBO), Rue de Bruxelles, Namur, Belgium;3 Ifremer, UMR 1067, Plouzane, France

The aim of this study was to evaluate the effects of dietary

phospholipids (PL) sources (fish gonad G-PL and soybean

lecithin S-PL) and levels (50 and 90 g kg)1dry matter) on the

performances and fatty acid (FA) composition of pikeperch

larvae From day 10 to day 34 posthatching (p.h.), larvae

were fed with three isoproteic and isolipidic microdiets The

best results of growth and skeletal development were related

to a high phospholipid level regardless of their origin and FA

profile Jaw deformities seemed associated with high dietary

highly unsaturated FA (HUFA) level The optimal level of

eicosapentaenoic acid and docosahexaenoic acid (EPA +

DHA) for pikeperch larvae appeared to be around 12 g kg)1

(dry matter) associated with a PL level around 90 g kg)1 FA

composition of diets and larvae revealed a better

incorpo-ration of arachidonic acid, EPA and DHA into PL fraction

especially in larvae fed with soybean PL Moreover,

34-day-old pikeperch larvae may have capability of converting 18

carbon n-3 FA into the n-3 HUFA Hence, for pikeperch

larvae, PL from plant origin were as efficient as those from

marine fish origin

key words: deformities, HUFA, larval development,

nutrition, phospholipids, pikeperch

Received 25 February 2011, accepted 10 June 2011

Correspondence: Neila Hamza, Institut National des Sciences et

Technol-ogies de la Mer, 28, Rue 2 Mars 1934, 2025 Salammbo, Tunisia E-mail:

hamza.neila@gmail.com

Several studies evidenced the beneficial role of phospholipids

(PL) on the survival, growth and development of marine fish

larvae (Salhi et al 1999; Sargent et al 1999; Cahu et al.2003a; Tocher 2003; Morais et al 2004; Gisbert et al 2005;Villeneuve et al 2005) as well as freshwater fishes (Geurden

et al.1997a, 1998a, 2008; Fontagne´ et al 1998, 2000; Olsen

et al.1999, 2003) Recent studies concerning the nutritionalrequirements of percid fishes like Eurasian perch (Perca flu-viatilis) or pikeperch (Sander lucioperca) focused on juvenilestage (Kestemont et al 2001; Xu et al 2001; Zakes et al.2001; Xu & Kestemont 2002; Abi-Ayad et al 2004; Nyina-Wamwiza et al 2005; Schulz et al 2005, 2006, 2007, 2008;Molnar et al 2006; Kowalska et al 2010) In these studies,requirements in lipids for pikeperch juvenile were estimatedbetween 100–160 g kg)1 (Nyina-Wamwiza et al 2005) and170–180 g kg)1(Molnar et al 2006; Schulz et al 2008) Inthe first study concerning pikeperch dietary requirements atthe first developmental stages, Abi-Ayad et al (2004) showedthat starved pikeperch larvae utilize preferentially polyun-saturated fatty acids (FAs) especially eicosapentaenoic acid(EPA) and docosahexaenoic acid (DHA), but also mono-unsaturated and saturated FAs as metabolic substrates Theoptimal dietary PL level was estimated to be around 90 g kg)1dry diet by Hamza et al (2008) Another study showed thatgrowth was similar for pikeperch larvae fed with Artemiacontaining different levels of essential FAs [arachidonic acid(ARA), EPA and DHA] (Lund & Steenfeldt 2009)

It is well known that PL play a major role in maintainingthe structure and function of cellular membranes (Kanazawa1985; Tocher 2003) They are reported to increase the effi-ciency of absorption and transport of dietary lipids (Teshima

et al.1986; Kanazawa 1991; Fontagne´ et al 2000; Izquierdo

et al.2000; Hadas et al 2003; Morais et al 2004) They alsoaffect the skeletal development (Geurden et al 1998b; Cahu

et al.2003b; Villeneuve et al 2005) Moreover, recent studiesdemonstrated an earlier maturation of the digestive tract infish larvae fed with PL-enriched diets (Cahu et al 2003a; .

2012 18; 249–257

Aquaculture Nutrition

Gisbert et al 2005; Hamza et al 2008) Recent reviews

(Cahu et al 2003b, 2009; Tocher et al 2008) have

synthe-sized data about their role and their mode of action (lipid

transport, lipoprotein synthesis, etc.)

A previous study showed that PL supplementation

signif-icantly improves growth and the activities of digestive

en-zymes in pikeperch larvae (Hamza et al 2008) Indeed, the

increase in dietary PL from 14 g kg)1 to 95 g kg)1 led to

50% enhancement in weight in 34-day-old larvae and the

earlier maturation of digestive structures This study was

conducted using vegetable PL and fish PL, the main

differ-ence between these two PL sources being FA profile, mainly

highly unsaturated FAs (HUFA) content Certain studies

have evidenced the role of these HUFA on morphogenesis

and notably on the skeletal development of marine fish larvae

(Cahu et al 2003a; Villeneuve et al 2006)

Owing to the increasing world demand in fish oils and the

plateau reached by the production, there is a great interest to

optimize their utilization and to use other sources of oils in

fish feed formulation This explained the interest in plant oils

for replacing fish oils in larval (Geurden et al 1997a) and

grow-out feeds (Olsen et al 2003; Menoyo et al 2004, 2010;

Kowalska et al 2010) It is particularly important to

deter-mine the essential requirements in PL levels and FA for fish

at the larval stage because their requirements are higher than

those of later stages (Koven et al 1993)

Considering that freshwater and marine fish do not have

the same FA requirements and utilization capacities, it

seemed interesting to evaluate the possibilities of replacing

fish oil by plant oil in larvae diet formulation of a valuable

species like pikeperch

The aim of this study was to determine the effects of

die-tary PL sources (plant and fish origin) and levels (50 and

90 g kg)1dry weight) on the pikeperch larvae performances

(survival, growth and deformities) and FA composition

Pikeperch larvae were obtained from the private hatchery

(Viskweekcentrum Valkenswaard, Leende, The

Nether-lands) and transferred to INSTM (Institut National des

Sciences et Technologies de la Mer) Tunisia on day 2

posthatching (p.h.) Upon arrival, they were acclimated into

two 500-L tanks supplied with UV-sterilized recycled fresh

water (20–22C) From mouth opening, on day 4 (p.h.),

they were fed ad libitum each hour from 8AM to 8PM, on

newly hatched small size Artemia nauplii (AF; INVE,

Dendermonde, Belgium) On day 10 (p.h.), the larvae weretransferred to the experimental unit consisting in a 12 cyl-indroconical tanks of 60 L each, supplied with UV-steril-ized recycled fresh water Initial stocking density was

20 larvae L)1 Temperature and dissolved O2, measureddaily, were maintained at 21–23C and above 6 mg L)1,respectively, with a water exchange of up to 100% h)1(flow rate: 1 L min)1) Nitrites and ammonia were mea-sured twice a week and maintained at levels lower than 0.1and 1 mg L)1, respectively The larvae were kept under lowlight intensity (30 lux maximum) except for cleaning andfeeding time Tanks were cleaned twice a day by siphoningthe faeces and dead larvae The number of dead larvae wasregistered daily, and the number of sampled larvae wastaken into account for survival calculation

From day 10 to day 34 posthatching (p.h.), larvae of eachtreatment were fed with one of three isoproteic and isolipidicmicrodiets (Table 1) formulated at IFREMER according tothe patent WO0064273 (November 2000) and containing PLfrom two origins fish gonad (G-PL) or soybean lecithin(S-PL) at two levels (50 or 90 g kg)1dry matter) Microdietswere processed as follows: dietary ingredients were mechan-ically mixed with water, pelleted and dried at 50C for

20 min The pellets were sieved to obtain two sizes of cles: 200–400 lm used between days 10 and 16 (p.h.), and400–700 lm used from day 17 (p.h.) until the end of theexperiment Food was distributed manually each 30 minfrom 8AMto 8PM Three tanks were randomly assigned toeach experimental diet

parti-The feeding levels were fixed at 250 g kg)1, 200 g kg)1,

150 g kg)1and 100 g kg)1of larval wet weight during the first,second, third and fourth week, respectively, corresponding to1–5 g tank)1day)1 A period (days 10–15 p.h.) of co-feedingwith Artemia was applied to habituate the larvae to the dry diet

Growth was monitored by sampling 30 larvae per tank ondays 4 and 10 (p.h.), and 10 larvae per tank on days 16, 22, 28and 34 (p.h.) Larvae were collected in the morning, beforefeeding, in order to have empty gut The larvae were weighedimmediately per group from day 0 to day 22 and individually

on days 28 and 34 At the end of the experiment, the weighedlarvae were also examined to detect malformation (lordosis,scoliosis and jaw malformation) The dead larvae werecollected and counted during tank cleaning

.

Growth, survival and deformity rates were calculated as

follows:

Specific growth rateðSGR;% day1Þ ¼ 100 ðlnWflnWiÞ DT1

Survival rateðS; %Þ ¼ 100 Nf=ðNi  NsÞDeformity rateðD; %Þ ¼ 100 Nd=Nswhere Wf and Wi = final and initial weight of larvae (mg),

T= time (days); Nf and Ni = final and initial number of

larvae; Nd = number of larvae displaying deformity and

Ns = number of sampled larvae

On day 34 p.h., 20 larvae per tank were collected for

analysis of total lipids and lipid classes content and FA

profiles Samples were taken before food distribution andimmediately stored at)80 C until analysis

Samples were homogenized by Ultra Turax (Yellow LineDI25 basic, Staufen, Germany) The lipid content in diets andlarvae was analysed after extraction with chloroform/metha-nol (2 : 1 v/v) according to the method of Folch et al (1957).Whole crude lipids were subsequently separated into polar andnon-polar fractions using Sep-Pak Silica cartridges (Waters,Milford, MA, USA) Chloroform and methanol were used asthe mobile phases for neutral and phospholipids, respectively(Juaneda & Rocquelin 1985) The FA methyl esters were pre-pared by transesterification with borontrifluoride in methanol

as described by Metcalfe et al (1966) and then separated by gaschromatography using an Agilent Technologies chromato-graph 6890N (Agilent Technologies, Palo Alto, CA, USA)equipped with a flame ionization detector, a splitless injectorand a polar INNOWAX 30M silica capillary column (0.25 mmi.d and 0.25 lm film thickness) The temperatures of theinjector and detector were 220 and 275C, respectively He-lium was used as a carrier gas at a flow rate of 1.5 mL min)1

Results were given as mean values and standard deviations.Survival, malformation, cannibalism rates and FA percent-ages were arcsin transformed and larval weights were log10transformed before statistical analysis Data were compared

by a two-wayANOVAfollowed by a least significant difference(LSD) test when significant differences were found at

P <0.05 The two analysed factors were phospholipids gin and level The homogeneity of variances was first checked

ori-by Box Ms test

On day 34, survival rate varied between 28% and 38%, and nosignificant differences appeared between treatments A sig-nificant effect of dietary PL level on larval growth was detectedfrom day 28 onwards At the end of the experiment, meanweights of larvae fed the highest level of PL (90 g kg)1) weremore than twice those of larvae fed 50 g kg)1PL, regardless

of PL origin (Fig 1, Table 2) According to statisticalanalysis, growth differences between these two groups wereexclusively attributed to dietary PL level (P = 3.6· 10)6)

Table 1 Experimental formulation and proximate composition of

the diets with varying phospholipids (PL) level and origin

Diet ingredients 1 (g kg)1) S-PL5 S-PL9 G-PL5 G-PL9

1 Dietary ingredients were commercially obtained Fish meal,

hydrolysed fish meal (CPSP G, Concentre´ de Prote´ines Solubles de

Poissons) and cod liver oil from La Lorientaise (Lorient, France) The

soy lecithin was from Ets Louis Franc¸ois (St Maur des Fosse´s, France).

2 Soybean lecithin contains: phospholipids 950 g kg)1 including

260 g kg)1phosphatidylcholine, 200 g kg)1

phosphatidylethanol-amine and 140 g kg)1phosphatidylinositol (PI).

3 Fish gonad lecithin contains: 600 g kg)1phospholipids including

450 g kg)1phosphatidylcholine, 200 g kg)1

phosphatidylethanol-amine, 160 g kg)1PI), 5% TAG, 15% cholesterol.

4 Per kg of vitamin mix : choline chloride : 200 g, retinyl acetate

0.34 g; all-rac-a-tocopherol 10 g; cholecalciferol 0.5 g; niacin 1 g;

D-calcium pantothenate 2 g; thiamin 100 mg; riboflavine 400 mg;

pyridoxin 0.3 g; ascorbic acid 20 g; folic acid 0.1 g; cyanocobalamin

1 g; biotin 1 g; menadione 1 g; meso-inositol 30 g.

5 Per kg of mineral mix : KCl 90 g, KL 4 O 40 mg, CaHPO 4 Æ2H 2 O

500 g, NaCl 40 g, CuSO 4 Æ5H 2 O 3 g, ZnSO 4 Æ7H 2 O 4 g, CoSO 4 Æ7H 2 O

20 mg, FeSO 4 Æ7H 2 O 20 g, MnSO 4 ÆH 2 O 3 g, CaCO 3 215 g, MgSO 4 Æ

7H 2 O 124 g and NaF 1 g.

6 Calculated as: lipid · 37.7 J kg)1; protein · 16.7 J kg)1.

.

Vertebral column deformities were almost exclusively

limited to the lordosis Less than 10% of the larvae fed high

PL level (S-PL9 and G-PL9) exhibited lordosis, while these

deformities affected almost one-third of the larvae of S-PL5

group and two-third of the larvae fed G-PL5 diet These

differences were both owing to PL level (P < 0.001) and to

PL source (P = 0.021)

The occurrence of scoliosis in the larvae fed with high

PL level was very low (<1%) and significantly lower

(P = 0.042) than in the larvae fed with 50 g kg)1PL whatever

the source Concerning the deformities affecting the larval jaw,

they were limited to the retracted upper jaw The incidence of

this deformity was significantly higher in larvae fed S-PL diets

(33–50%) than in larvae fed G-PL diets (16–27%), and

differ-ences were exclusively attributed to the PL source (P = 0.04)

Diets The FA composition of the four diets S-PL5, S-PL9,

G-PL5 and G-PL9 are detailed in Table 3

Differences between G-PL and S-PL diets were mainlyobserved in PL fraction and were essentially observed at thelevel of n-6 and n-3 FA series The n-3/n-6 FA ratios variedfrom 0.2 (S-PL9) to 17.2 (G-PL9) Diets formulated with PLfrom soybean origin were richest in short n-6 FA essentiallyrepresented by C18:2n-6 (almost 50% of total FAs) and inshort and long n-3 FA essentially represented by C18:3n-3and C22:6n-3 (3% and 4–6% of total FAs, respectively) Onthe contrary, diets containing PL from fish origin wererichest in long n-6 FA (C20:4n-6) and in long n-3 FA (par-ticularly HUFA, C20:5 and C22:6) Diets formulated withsoybean lecithin contained 1.9% and 1.2% of EPA + DHArelated to their dry weight On the opposite, G-PL dietsincluded 3.2% and 4.3% EPA + DHA related to dryweight

Larvae The FA composition of the larvae globally reflected

FA composition of the diets (differences in n-6 and n-3series) especially in the neutral lipid fraction (Table 4)

In PL fraction of G-PL9 larvae, the level in total n-3 FAand in particular the level in DHA were not significantlydifferent from n-3 and DHA level in S-PL5 or S-PL9 larvaedespite the important differences in HUFA composition ofthe diets ARA, EPA and DHA concentrations were three- tosixfold higher in S-PL larvae than observed in the corre-sponding S-PL diets Inversely, they were in lower concen-tration in G-PL larvae compared with G-PL diets (especiallyDHA)

Consistent with our previous study (Hamza et al 2008), thesurvival of pikeperch larvae was not affected by the dietary

PL level This study showed in addition that PL source didnot have effect on larval survival

Results of this study were in agreement with other feedingstudies indicating a growth and development promotingeffect of dietary PL on several marine and freshwater fish

a

a c

b a ab

Figure 1 Growth of pikeperch larvae fed different dietary

phos-pholipid levels and origins (S-PL5, S-PL9 = PL from soybean at

50 g kg)1and 90 g kg)1; G-PL5 and G-PL9 = PL from fish gonad at

50 g kg)1 and 90 g kg)1) Age is expressed in days posthatching

(dph) PL, phospholipids Values with different letters are

sig-nificantly different (P < 0.05).

Table 2 Survival, growth and per cent deformities in the four experimental groups S-PL of 34-day-old pikeperch larvae fed different levels and origin of dietary phospholipids

Values are the mean ± SD of triplicate determinations (n = 3).

Values with different letters in the same line are significantly different (P < 0.05).

SGR, specific growth rate; PL, phospholipids.

.

larvae and juveniles (Geurden et al 1998a,b, 2008; Olsen

et al 1999; Fontagne´ et al 2000; Cahu et al 2003a,b;

Gisbert et al 2005; Hamza et al 2008; Kjørsvik et al 2009)

The results observed at the end of the experiment revealedthat the differences in growth between S-PL5 and G-PL5from one part and S-PL9 and G-PL9 from another part were

Table 3 Fatty acid (FA) composition

(% of total FAs) of the phospholipids

(PL) and neutral lipid (NL) fractions of

the experimental diets S-PL5, S-PL9,

Total EPA + DHA (%

diet dry weight)

Values with different letters in the same line are significantly different (P < 0.05).

HUFA, highly unsaturated FA; MUFA, monounsaturated FA; SFA, saturated FA.

.

because of the dietary PL level whatever their origin Our

previous study already showed that pikeperch larvae fed

with a diet containing a high PL level (90 g kg)1) had

better growth and digestive development compared with

larvae fed with diets containing 14 g kg)1or 50 g kg)1 PL

(Hamza et al 2008)

The observed results concerning deformities occurrence

showed a similar trend compared with growth results

Indeed, deformity rate represented by lordosis and scoliosis

was lower for larvae fed with the highest PL level regardless

of their origin Several authors demonstrated the beneficial

role of PL on the skeletal formation process in sea bass

Dicentrarchus labrax(Cahu et al 2003a,b; Villeneuve et al

2005) Geurden et al (1998b) showed that

phosphatidylino-sitol (PI) fraction prevents particularly skeletal deformities in

common carp, Cyprinus carpio According to the same

authors, animal PL contains less PI than vegetable PL

(Geurden et al 1997b) In our study, G-PL5 and S-PL5

larvae that received the lowest dietary PL level showed the

highest levels of lordosis (65% and 32%, respectively) and

scoliosis (3.3% and 6.7%, respectively) These skeletal

deformities may be related not only to the PL level but also

to the PI content in the diet Indeed, the used fish lecithin

contains 60% PL in which PI represents 16%, while the used

soybean lecithin contains 95% PL in which PI represents

14% Correspondingly, G-PL5 and S-PL5 larvae received

7.7 g kg)1PI diet and 10.6 g kg)1PI This may explain that

the occurrence of lordosis was higher in larvae fed with

G-PL5 than in the larvae fed with S-PL5

Several studies (Coutteau et al 1997; Cahu et al 2003b;

Villeneuve et al 2005, 2006; Kjørsvik et al 2009) suggested

that HUFA can also affect the skeletal formation during

larval development According to Villeneuve et al (2005), a

moderate level of EPA + DHA (2.3%) in the PL fraction of

the diet induced better growth and lower deformities rates,

while an excess of HUFA (4.8% EPA + DHA dry matter in

PL fraction) in the diet led to lower growth and higher

incidence of deformities in sea bass larvae Our results

showed that larvae fed with G-PL5 showed similar growth

but much more deformities (lordosis and scoliosis) than

S-PL5 larvae We may suppose that this difference was

be-cause of the dietary HUFA profile in the PL fraction (1.3%

and 0.3% EPA + DHA dry matter, of G-PL5 and S-PL5

diets, respectively) However, this hypothesis was not

sustained by the fact that the larvae fed with the richest diet

in n-3 HUFA (G-PL9: 2.9% EPA + DHA dry matter)

showed the best growth and lowest deformity rates as well as

those fed with the poorest diet in n-3 HUFA (S-PL9: 0.3%

EPA + DHA dry matter) In the limit of these experimental

values, it seems that PL was determinant for the growth andskeletal development independently of their HUFA profile

In the same way, Kjørsvik et al (2009) demonstrated that, inthe case of cod larvae, the ossification process depends more

on lipid source than on the FA quantity in the diet

On the other hand, retracted upper jaw rate was cantly lower in larvae fed with fish PL but reached 33–50%

signifi-with S-PL diets These differences were attributed to PLsource and may be related to HUFA profile of the diets

According to Villeneuve et al (2006), inadequate dietarylevels of HUFA led to different skeletal malformations in seabass larvae, dietary n-3 HUFA being involved in the mod-ulation of the genes implicated in bone development

In this study, G-PL9 and S-PL9 larvae had similar growthdespite the differences in FA composition and particularlyHUFA of corresponding diets As already showed in ourprevious study (Hamza et al 2008), the best larval growthand development were obtained with the highest dietary PLlevel independently of the FA composition Analogousresults were obtained by Lund & Steenfeldt (2009) whoreported similar growth with pikeperch larvae fed withArtemiasupplemented with ARA, EPA or DHA

Compared with other freshwater species, we can considerthat pikeperch larvae have relatively high PL requirements(around 90 g kg)1dry matter), near to those of marine spe-cies like sea bass 120 g kg)1(Cahu et al 2003a) However, itappears that its essential fatty acids (EFA) requirements arelower than those reported for marine fish Indeed, according

to the results of previous and present studies, the optimallevel of EPA + DHA for pikeperch larvae could be situatedaround 1.2%, on a dry matter basis), thus lower than thosesuggested for marine species (around 3%) (Sargent et al

1999; Cahu & Zambonino Infante 2001) This assumptionwas sustained by Abi-Ayad et al (2000) who showed thatpikeperch have low physiological requirements for EFA,even when compared with a close species like the Eurasianperch Nevertheless, a high level of EPA + DHA (4.3% dietdry weight) in the diet did not induce high incidence ofdeformities contrary to the sea bass larvae (Villeneuve et al

2005) Thus, it seems that pikeperch larvae tolerate highlevels of HUFA in the diet

However, these relatively low requirements for pikeperchlarvae have to be associated with a relatively high dietary PLlevel, around 90 g kg)1, from soybean origin A previousstudy showed that a diet formulated with FA profile similar

to the PL9 diet but with 14 g kg)1 PL from cod liver oil,which is triglycerides, gave poor results in terms of growthand development (Hamza et al 2008) In the same way, 1.5%

EPA + DHA was shown to be sufficient for a good .

development of sea bass larvae when associated with

120 g kg)1PL (Cahu et al 2003a)

In this experiment, ARA, EPA and DHA levels (% total

FA) were three- to sixfold higher in PL fraction of the

larvae fed with soybean PL compared with the

corre-sponding diets (S-PL5 and S-PL9) On the contrary, larvae

fed with fish PL contained lower levels of these HUFA than

the diets This may be explained by a better incorporation

of ARA, EPA and DHA into PL fraction in larvae fed with

soybean PL although these diets contained much lower n-3

HUFA levels than G-PL diets Same results were observed

by Xu et al (2001) with Eurasian perch This better

incor-poration of HUFA with S-PL9 diet may be attributed to

the stimulatory effect of soybean PL on intestinal

lipopro-tein secretion (Fontagne´ et al 1998; Geurden et al 1998a)

The transport of dietary lipids was thus improved (Teshima

et al 1986; Fontagne´ et al 2000; Hadas et al 2003) as

well as FA absorption in intestine (Geurden et al 1998a,

2008)

Moreover, pikeperch larvae probably have the capability

to elongate and desaturate FA Indeed, the increase in EPA

and DHA contents in PL fraction (three- to fivefold higher in

the S-PL5 and S-PL9 larval tissues than in S-PL5 and S-PL9

diets) and the marked decrease in C18:3n-3 (in the larvae

compared with the diets) may indicate a high capacity of

converting 18 carbon n-3 FA to the n-3 HUFA This

capa-bility has already been demonstrated for Eurasian perch (Xu

et al.2001; Xu & Kestemont 2002) and pikeperch juveniles

(Schulz et al 2005) fed with different lipid levels or origins

Similar conclusions could also be drawn from our previous

study (Hamza et al 2008) but should be confirmed by

labelled FAs

In conclusion, it is the first study that showed the effect of

the origin and level of dietary PL on the performances on the

first stage of pikeperch development with inert diets This

study evidenced that plant PL were as efficient as fish PL to

promote pikeperch larval growth and development This

seems to be due to the plant PL properties but also to the

capacity of pikeperch larvae to elongate and desaturate FAs

Fish oil can be totally replaced by soybean oil with high level

of PL in diet for pikeperch larvae These results showed the

importance to determine the requirements specific to a

species and to a stage of life to optimize and equilibrate diet

formulation, PL/HUFA in this case

This study was initiated by a cooperative project between

INSTM (Tunisia) and FUNDP (Belgium) and supported by

a CGRI-DRI grant, French-speaking community of Belgiumand Ministry of the Walloon Region Authors thank

P Quazuguel and H le Delliou (IFREMER, France) fortheir technical assistance They also express their gratitude

to Saloua Sadok (INSTM, Tunisia) for access to herlaboratory

Abi-Ayad, S.-M.E.-A., Boutiba, Z., Me´lard, C & Kestemont, P (2004) Dynamics of total body fatty acids during early ontogeny of pikeperch (Sander lucioperca) larvae Fish Physiol Biochem., 30, 129–136.

Abi-Ayad, S.-M.E.-A., Kestemont, P & Me´lard, C (2000) Dynamics of total lipids and fatty acids during embryogenesis and larval development of Eurasian perch (Perca fluviatilis) Fish Physiol Biochem., 23, 233–243.

Cahu, C.L & Zambonino Infante, J.L (2001) Substitution of live food by formulated diets in marine fish larvae Aquaculture, 200, 161–180.

Cahu, C.L., Zambonino Infante, J.L & Barbosa, V (2003a) Effect of dietary phospholipid level and phospholipid:neutral lipid value on the development of sea bass (Dicentrarchus labrax) larvae fed a compound diet Br J Nutr., 90, 21–28.

Cahu, C., Zambonino Infante, J & Takeuchi, T (2003b) Nutritional components affecting skeletal development in fish larvae Aqua- culture, 227, 245–258.

Cahu, C.L., Gisbert, E., Villeneuve, L.A.N., Morais, S., Hamza, N., Wold, P.-A & Zambonino Infante, J.L (2009) Influence of dietary phospholipids on early ontogenesis of fish Aquac Res., 40, 989–999 Coutteau, P., Geurden, I., Camara, M.R., Bergot, P & Sorgeloos, P (1997) Review on the dietary effects of phospholipids in fish and crustacean larviculture Aquaculture, 155, 149–164.

Folch, J., Lees, M & Stanley, G.H.S (1957) A simple method for the isolation and purification of total lipids from animal tissues.

J Biol Chem., 226, 497–509.

Fontagne´, S., Geurden, I., Escaffre, A.M & Bergot, P (1998) tological changes induced by dietary phospholipids in intestine and liver of common carp (Cyprinus carpio L.) larvae Aquaculture,

His-161, 213–223.

Fontagne´, S., Burtaire, L., Corraze, G & Bergot, P (2000) Effects of dietary medium-chain triacylglycerols (tricaprylin and tricaproin) and phospholipid supply on survival, growth and lipid metabolism

in common carp (Cyprinus carpio L.) larvae Aquaculture, 190, 289–303.

Geurden, I., Coutteau, P & Sorgeloos, P (1997a) Effect of dietary phospholipid supplementation on growth and fatty acid compo- sition of European sea bass (Dicentrarchus labrax L.) and turbot (Scophtalmus maximus L.) juveniles from weaning onwards Fish Physiol Biochem., 16, 259–272.

Geurden, I., Coutteau, P & Sorgeloos, P (1997b) Increased sahexaenoic acid levels in total and polar lipid of European sea bass (Dicentrarchus labrax) postlarvae fed vegetable or animal phospholipids Mar Biol., 129, 689–698.

doco-Geurden, I., Bergot, P., Schwarz, L & Sorgeloos, P (1998a) tionship between dietary phospholipid classes and neutral lipid absorption in newly-weaned turbot, Scophtalmus maximus Fish Physiol Biochem., 19, 217–228.

Rela-Geurden, I., Marion, D., Charlon, N., Coutteau, P & Bergot, P (1998b) Comparison of different soybean phospholipidic fractions

as dietary supplements for common carp, Cyprinus carpio, larvae.

Aquaculture, 161, 225–235.

Geurden, I., Kaushik, S & Corraze, G (2008) Dietary

phosphati-dylcholine affects postprandial plasma levels and digestibility

of lipid in common carp (Cyprinus carpio) Br J Nutr., 100, 512–

517.

Gisbert, E., Villeneuve, L., Zambonino Infante, J.L., Quazuguel, P.

& Cahu, C.L (2005) Dietary phospholipids are more efficient than

neutral lipids for long chain polyunsaturated fatty acid supply in

european sea bass Dicentrarchus labrax larval development Lipids,

40, 1–10.

Hadas, E., Koven, W., Sklan, D & Tandler, A (2003) The effect of

dietary phsophatidylcholine on the assimilation and distribution of

ingested free oleic acid (18:1n-9) in gilthead seabream (Sparus

aurata) larvae Aquaculture, 217, 577–588.

Hamza, N., Mhetli, M., Ben Khemis, I., Cahu, C & Kestemont, P.

(2008) Effect of dietary phospholipid levels on performance,

enzyme activities and fatty acid composition of pikeperch (Sander

lucioperca) larvae Aquaculture, 275, 274–282.

Izquierdo, M.S., Socorro, J., Arantzamendi, L & Hernandez-Cruz,

C.M (2000) Recent advances in lipid nutrition in fish larvae Fish

Physiol Biochem., 22, 97–107.

Juaneda, P & Rocquelin, G (1985) Rapid and convenient separation

of phospholipids and non phosphorus lipids from rat heart using

silica cartridges Lipids, 30, 40–41.

Kanazawa, A (1985) Essential fatty acid and lipid requirement of

fish Nutrition and Feeding in Fish Academic Press, London.

Kanazawa, A (1991) Essential phospholipids of fish and

crusta-ceans In: Fish Nutrition in Practice, Biarritz (France) June 24–27,

1991 (INRA ed.), pp 519–530 INRA, Paris 1993 (Les colloques

N61).

Kestemont, P., Vandeloise, E., Me´lard, C., Fontaine, P & Brown,

P.B (2001) Growth and nutritional status of Eurasian perch Perca

fluviatilis fed graded levels of dietary lipids with or without

eth-oxyquin Aquaculture, 203, 85–99.

Kjørsvik, E., Olsen, C., Wold, P.-A., Hoehne-Reitan, K., Cahu,

C.L., Rainuzzo, J., Olsen, A.I., Oie, G & Olsen, Y (2009)

Com-parison of dietary phospholipids and neutral lipids on skeletal

development and fatty acid composition in Atlantic cod (Gadus

morhua) Aquaculture, 294, 246–255.

Koven, W.M., Kolkovski, S., Tandler, A., Kissil, G.Wm & Sklan,

D (1993) The effect of dietary lecithin and lipase, as a

function of age, on n-9 fatty acid incorporation in the tissue

lipids of Sparus aurata larvae Fish Physiol Biochem., 10, 357–

364.

Kowalska, A., Zakes, Z., Jankowska, B & Siwicki, A (2010)

Substituting vegetable oil for fish oil in pikeperch diets: impact

on growth, internal organ histology, blood biochemical

parameters, and proximate composition Aquacult Nutr., 17,

e148–e163.

Lund, I & Steenfeldt, S.J (2009) The effects of dietary long-chain

essential fatty acids on growth and stress tolerance in

pikeperch larvae (Sander lucioperca L.) Aquacult Nutr., 17, 191–

199.

Menoyo, D., Izquierdo, M.S., Robaina, L., Gines, R., Lopez-Bote,

C.J & Bautista, J.M (2004) Adaptation of lipid metabolism, tissue

composition and flesh quality in gilthead sea bream (Sparus

aurata) to the replacement of dietary fish oil by linseed and

soyabean oils Br J Nutr., 92, 41–52.

Menoyo, D., Lopez-Bote, C.J., Obach, A & Bautista, J.M (2010)

Effect of dietary fish oil substitution with linseed oil on the

performance, tissue fatty acid profile, metabolism, and oxidative

stability of Atlantic salmon J Anim Sci., 83, 2853–2862.

Metcalfe, L.D., Schmitz, A.A & Pelka, J.R (1966) Rapid tion of fatty acid esters from lipids for gas chromatographic analysis Anal Chem., 38, 524–535.

prepara-Molnar, T., Szabo´, A., Szabo´, G., Szabo´, C & Hancz, C (2006) Effect of different dietary fat content and fat type on the growth and body composition of intensively reared pikeperch Sander lucioperca (L.) Aquacult Nutr., 12, 173–182.

Morais, S., Cahu, C., Zambonino Infante, J.L., Robin, J., tad, I., Dinis, M.T & Conceic¸a˜o, L.E.C (2004) Dietary TAG source and level affect performance and lipase expression in larval seabass (Dicentrarchus labrax) Lipids, 39, 1–10.

Rønnes-Nyina-Wamwiza, L., Xu, X.L., Blanchard, G & Kestemont, P.

(2005) Effect of dietary protein, lipid and carbohydrate ratio on growth, feed efficiency and body composition of pikeperch Sander lucioperca fingerlings Aquacult Res., 36, 486–492.

Olsen, R.E., Myklebust, R & Ringø, E (1999) Lipid digestibility and ultrastructural changes in the enterocytes of Arctic char (Salvelinus alpinus L.) fed linseed oil and soybean lecithin Fish Physiol Bio- chem., 21, 35–44.

Olsen, R.E., Dragnes, B.T., Myklebust, R & Ringø, E (2003) Effect of soybean oil and soybean lecithin on intestinal lipid composition and lipid droplet accumulation of rainbow trout, Oncorhynchus mykiss Walbaum Fish Physiol Biochem., 29, 181–192.

Salhi, M., Hernander-Cruz, C.M., Bessonart, M., Izquierdo, M.S &

Fernandez-Palacios, H (1999) Effect of different dietary polar lipid levels and different n-3 HUFA content in polar lipids on gut and liver histological structure of gilthead seabream (Sparus aurata) larvae Aquaculture, 179, 253–263.

Sargent, J., Bell, G., McEvoy, L., Tocher, D & Estevez, A (1999) Recent developments in the essential fatty acid nutrition of fish.

Aquaculture, 177, 191–199.

Schulz, C., Knaus, U., Wirth, M & Rennert, B (2005) Effects of varying dietary fatty acid profile on growth performance, fatty acid, body and tissue composition of juvenile pikeperch (Sander lucioperca) Aquacult Nutr., 11, 403–413.

Schulz, C., Gu¨nther, S., Wirth, M & Rennert, B (2006) Growth performance and body composition of pikeperch (Sander luciop- erca) fed varying formulated and natural diets Aquacult Int., 14, 577–586.

Schulz, C., Bo¨hm, M., Wirth, M & Rennert, B (2007) Effect of dietary protein on growth, feed conversion, body composition and survival of pikeperch fingerlings (Sander lucioperca) Aquacult.

Nutr., 13, 373–380.

Schulz, C., Huber, M., Ogunji, J & Rennert, B (2008) Effects of varying dietary protein to lipid ratios on growth performance and body composition of juvenile pikeperch (Sander lucioperca).

Aquacult Nutr., 14, 166–173.

Teshima, S.-I., Kanazawa, A & Kakuta, Y (1986) Effects of dietary phospholipids on growth and body composition of the juvenile prawn Bull Jap Soc Sc Fish., 52, 155–158.

Tocher, D.R (2003) Metabolism and functions of lipids and fatty acids in teleost fish Rev Fish Sci., 11, 107–184.

Tocher, D.R., Bendiksen, E.A., Campbell, P.J & Bell, G.J (2008) The role of phospholipids in nutrition and metabolism of teleost fish Aquaculture, 280, 21–34.

Villeneuve, L., Gisbert, E., Zambonino-Infante, J.L., Quazuguel, P.

& Cahu, C.L (2005) Effect of nature of dietary lipids on European sea bass morphogenesis: implication of retinoid receptors Br J.

Nutr., 94, 877–884.

Villeneuve, L.A.N., Gisbert, E., Moriceau, J., Cahu, C.L & bonino Infante, J.L (2006) Intake of high levels of vitamin A and polyunsaturated fatty acids during different developmental periods

Zam-modifies the expression of morphogenesis genes in European sea

bass (Dicentrarchus labrax) Br J Nutr., 95, 677–687.

Xu, X.L & Kestemont, P (2002) Lipid metabolism and FA

com-position in tissues of Eurasian Perch Perca fluviatilis as influenced

by dietary fats Lipids, 37, 297–304.

Xu, X.L., Fontaine, P., Me´lard, C & Kestemont, P (2001) Effects of

dietary fat levels on growth, feed efficiency and biochemical

compositions of Eurasian perch Perca fluviatilis Aquac Int., 9, 437–449.

Zakes, Z., Szkudlarek, M., Wozniak, M., Karpinski, A & Demska-Zakes, K (2001) Effect of dietary protein:fat ratios on metabolism, body composition and growth of juvenile pikeperch, Stizostedion lucioperca (L.) Czech J Anim Sci., 46, 27–33.

.

1 1 1 2 2 1 1

1

Centro de Investigaciones Biolo´gicas del Noroeste, S.C., Mar Bermejo No 195, Col Playa Palo de Santa Rita, La Paz, BCS,

Me´xico;2 Laboratory of Aquaculture and Artemia Reference Center, University of Gent, Ghent, Belgium

Stocking shrimp at high densities increases yield during

cul-ture, but growth is generally compromised and weakened

immune response associated with poor water quality has also

been reported Therefore, we tested if supplying more

ara-chidonic acid (ARA) in the diet, a precursor of eicosanoids

such as prostaglandins E from the series II (PGE2) that

enhance immune response can counteract the negative effect

of stocking shrimp at high densities The effect of physical

crowding was separated from the effect of water quality, both

a result of high density, by using tanks divided by a hard

plastic net that allowed water flow between two density

conditions Crowding reduced weight gain by 8.3%,

although the effect was more evident with deteriorated water

quality from combined effects of high total ammonia and low

dissolved oxygen levels (18.4%), but no effect on survival was

found A clear food imprinting of ARA levels in hemocytes

was observed, but ARA did not clearly counteract the

neg-ative effects of high density on overall performance

How-ever, ARA could minimize stress response of sampling and

enhance some effectors of the immune system, such as

clot-ting and respiratory burst The increase in PGE2metabolite

in shrimp fed with the high-ARA diet was not consistent, and

thus, the effects of ARA were not necessarily mediated by

these eicosanoids

key words: arachidonic acid, density, immune response,

polyunsaturated fatty acids, prostaglandins, shrimp

Received 13 December 2010, accepted 7 July 2011

Correspondence: Elena Palacios, Centro de Investigaciones Biolo´gicas del

Noroeste, S.C., Mar Bermejo No 195, Col Playa Palo de Santa Rita, La

Paz, BCS 23090, Me´xico E-mail: epalacio@cibnor.mx

Stocking shrimp at high densities is desirable to increase yieldduring shrimp culture However, increasing stocking densitiesaffect growth and sometimes survival of shrimp (Allan &

Maguire 1992; Martin et al 1998; Li et al 2006; Herrera et al 2006) Lemonnier et al (2006) attributeddecreased survival of Litopenaeus stylirostris during ÔSummersyndromeÕ episodes to high densities (30–40 shrimp m)2) Wu

Mena-et al.(2001) concluded that Penaeus japonicus reared at higherdensities (260 shrimp m)2compared to 73 shrimp m)2) weremore prone to horizontal transmission of viruses throughcannibalism and waterborne route Li et al (2006) found thatincreasing stocking density produced an increase in pheno-loxidase activity and a decrease in haemolysin activity inFenneropenaeus chinensisshrimp, as indicators of the capacity

of immune response Reduced survival in shrimp stocked athigh densities is possibly a result of weakened immuneresponse and overall energy unbalance, associated with poorwater quality, such as hypoxia and high ammonia concen-trations that depress immune response in shrimp (Le Moullac

& Haffner 2000) In fact, the effect of stocking density onshrimp can be separated into physical (crowding) and chem-ical (water quality) stresses, although very few studies havetried to separate between both effects Nga et al (2005) con-cluded that the well-known effect of high density on Penaeusmonodon shrimp growth seems to be caused by chemicalrather than physical interference

Several studies have investigated the use of diets rich inhighly unsaturated fatty acids (HUFA) to counteract thecompromised immune response capacity caused by stressfulconditions Chim et al (2001) showed that a diet rich in x3HUFA increased agglutination titre of plasma and respira-tory burst of hemocytes in Litopenaeus stylirotris In

2012 18; 258–271 .doi: 10.1111/j.1365-2095.2011.00892.x

.Aquaculture Nutrition

addition, Litopenaeus vannamei shrimp fed HUFA-rich diets

had significantly higher levels of lipoprotein-beta glucan

binding protein (Mercier et al 2009) A greater resistance to

environmental short-term stressors, such as a decrease in

salinity and temperature, was also observed after having fed

shrimp on diets containing high levels of HUFA (Tackaert

et al.1989; Rees et al 1994; Mourente & Rodriguez 1997;

Palacios et al 2004) However, to date, the majority of

studies focus on x3 but not on x6 HUFA, namely

arachi-donic acid (20:4n-6 or ARA) It has been observed that

dietary ARA attenuated the negative effects of handling

stress in fish larvae, although it increased mortality when fish

were exposed to repetitive salinity change (Koven et al

2003) ARA is a precursor of eicosanoids, such as

prosta-glandins E from the series II (PGE2) and of other

eicosa-noids, which affect the immune system in vertebrates and

insects (Stanley & Miller 2006), and for this reason could

enhance the immune response capacity, thereby

counteract-ing the effects of stress induced by high density

The purpose of the present study was to test the influence

of two dietary levels of ARA and two culture densities on

growth, survival, levels of a metabolite of prostaglandin E2

metabolite (PGEM), immune response and biochemical

composition of shrimp Two experiments were conducted to

separate the effect of physical crowding (high densities) from

the chemical effect (water quality) by using tanks divided by a

hard plastic net

Two diets with high- and low-ARA levels were formulated

using the software MIXIT-WIN (Agricultural Software

Consultants, San Diego, CA, USA) Diets were formulated

to be isolipidic and isonitrogenous, according to L vannamei

nutritional requirements (Akiyama & Dominy 1989; Camba

et al.1993) Prior to preparing the diets, the dry ingredients

were pulverized and sieved through a 250-lm mesh sieve and

mixed in a food mixer Wet ingredients, namely soybean

lecithin and different lipid emulsions were mixed, added to

the dry ingredients and homogenized Water was blended

into the mixture to attain a consistency appropriate for

pel-leting The resulting mixture was pelleted in a meat grinder

through a die with 3-mm diameter holes Pellets were dried in

a forced-air oven at 35C until 80–100 g kg)1 moisture

content was reached, and they were stored in plastic bags at

)20 C Different dietary ARA content was achieved by

including different ICES emulsions (ICES, 1997) The

prox-imate composition of the diets was determined using themethods of AOAC (2005) Gross energy was determinedusing a bomb calorimeter (Model 1261; Parr Instrument,Moline, IL, USA) The diet formulations and proximatecompositions are presented in Table 1, and fatty acid com-position for both diets is shown in Table 2 For convenience,experimental diets were termed low- and high-ARA dietswith final ARA levels of 0.8% (86.2 ± 5.4 mg 100 g)1) and3.2% (323.4 ± 27.7 mg 100 g)1) of total fatty acids

Two different experiments were conducted to analyse theinfluence of crowding only (Experiment C) or crowd-ing + water quality changes produced by high densities(Experiment C + W) In Experiment C, eight outdoor con-crete tanks (1.38· 1.11 m side by side, ·0.6 m, depth) wereused These outdoor concrete tanks were shaded by a sheetmetal roof suspended above, and a black plastic was placeddirectly on top of each tank covering it partially to reducelight incidence, phytoplankton proliferation and avoidshrimp jumping out of the tanks Tanks were divided in two

by a mesh (10 mm), with one half stocked with a low density

of shrimp (12 organisms per half tank = 16 shrimp m)2)and the other half with a high density (50 organisms per halftank = 68 shrimp m)2) The water was allowed to flowbetween both halves to obtain the same water quality forboth densities Each dietary treatment was randomlyassigned to four tanks (eight half tanks for each dietarytreatment, i.e., four half tanks for each diet–density combi-nation) In Experiment C + W, the same mesh-separatedtanks were used but in this case, both halves of each tankwere stocked with the same density of shrimp: low(16 shrimp m)2) or high (68 shrimp m)2) As for Experiment

C, each diet was provided to four tanks, corresponding tofour half tanks for each diet–density combination

Juvenile Pacific whiteleg shrimp L vannamei were lected from culture ponds at the Centro de InvestigacionesBiolo´gicas del Noroeste (CIBNOR) and only apparenthealthy organisms of 3.5 ± 0.1 g (mean ± standard error)were selected for the experiments A 40% daily exchange ofwater was used for each tank and they were equipped withone air stone A strong reduction in dissolved oxygen and anincrease in total ammonia by the end of the 2nd week promptfor a 100% exchange of water for a day in all the systems(day 12–13) and then for 60% of daily exchange of water.Temperature was maintained at 26 ± 0.8C with salinity at

col-38 ± 1 g L)1 Photoperiod and illumination were natural(approximately 12 : 12) Daily values of dissolved oxygen .

and weekly values of pH and ammonium were measured

before water exchange To allow density maintenance, dead

shrimp were replaced by similar-sized shrimp that were

tagged using a plastic ring around the eye, and which werenot used for final sampling

Shrimp were first acclimated for 5 days to the experimentaltanks, while they were fed a commercial diet (350 g kg)1protein, PIASA, La Paz, B.C.S., Mexico) twice daily Low-

or high-ARA diets were then provided The total daily feedwas initially set at 5% of the biomass in each tank, distrib-uted manually in two rations (50% at 10:00 h and 50% at18:00 h) Rations were adjusted daily, based on apparentconsumption, where the added feed closely matched themaximum feed ingested, with little or no feed remaining

After 30 days, 12 h – fasted shrimp in intermolt and earlypremolt were sampled for analyses

A total of 24 shrimp was sampled for each experiment (n = 6for each group of diet–density combination) Approximately

250 lL haemolymph was sampled from the base of thepleopods of the first abdominal segment, using 5% sodiumoxalate as anticoagulant Haemolymph was centrifuged at

800 g for 5 min at 4C and stored at)70 C PGEM wasassayed by the Prostaglandin screening EIA Kit (CaymanChemical No 514531, Ann Arbor, MI, USA) As indicated

in the Cayman Kit, PGE2 is rapidly converted in vivo toPGEM and so measurement of this metabolite is more

Table 1 Ingredients and proximate composition of low- and

Gent crude protein 351.9 ± 2.1 353.2 ± 1.2

Gent free nitrogen extract 467.2 468.8

Gent energy (kJ g)1) 18.6 ± 0.1 18.5 ± 0.1

ARA, arachidonic acid; BHT, butylated hydroxytolune.

1 Harinera Parayas, Jalisco, Mexico.

2 Monterrey sardine, Conservera San Carlos, Puerto San Carlos,

B.C.S., Mexico.

3 Solvent extracted, AGYDSA, Jalisco, Mexico.

4 Sigma-Aldrich, A-7128 St Louis, MO, USA.

5

Vitamin composition (g kg)1 premix): vitamin A acetate

(20 000 IU g)1), 5; vitamin D3 (850 000 IU g)1), 0.001;

DL-a-toc-opheryl acetate (250 IU g)1), 8; menadione, 2; thiamin–HCl, 0.5;

riboflavin, 3; pyridoxine–HCl, 1; DL-Ca-pantothenate, 5; nicotinic

acid, 5; biotin, 0.05; inositol, 5; cyanocobalamin, 0.002; folic acid,

0.18; filler a-cellulose, 865.266.

6

Sigma-Aldrich, S-0876 St Louis, MO, USA.

7 Mineral premix (g kg)1 premix): CoCl 2, 0.04; CuSO 4 5H 2 O, 2.5;

FeSO 4 7H 2 O, 40; MgSO 4 7H 2 O, 283.98; MnSO 4 H 2 O, 6.5; KI, 0.67;

Na 2 SeO 3, 0.1; ZnSO 4 7H 2 O, 131.93; filler a-cellulose, 534.28.

8 Sigma-Aldrich, C-8503 St Louis, MO, USA.

9

ICN Biomedicals Inc., 101386 Aurora, OH, USA.

10 Stay C 35% active agent, Roche, D.F., Mexico.

11 ODONAJI, Distribuidora de Alimentos Naturales y Nutricionales,

La Paz, BCS, Mexico.

12

Laboratory of Aquaculture and Artemia Reference Center,

Department of Animal Production, Gent University, Gent, Belgium.

13 Based on Vevodar oil, from the Laboratory of Aquaculture and

Artemia Reference Center, Department of Animal Production, Gent

University, Gent, Belgium.

14 Butylated Hydroxytoluene, ICN Biomedicals Inc., 101162 Aurora,

OH, USA.

15

Results are presented as mean ± SD, except for the results of the

free nitrogen extract that are given as means only.

Table 2 Fatty acid composition (% of total fatty acids) of the and high-ARA diets (means ± SD)

low-Low-ARA diet (n = 3)

High-ARA diet (n = 3)

reliable for estimation of actual PGE2 production in

biological samples According to the recommendations of the

manufacturer a purification step was performed on C-18 SPE

cartridges (Waters, Milford, MA, USA)

A different set of organisms (n = 6 shrimp for each diet–

density combination for the Experiment C or Experiment

C + W) was used for the analyses of immune variables, for

which haemolymph was collected using EDTA as

anticoag-ulant (Vargas-Albores et al 1993), with a haemolymph:

EDTA dilution of 1 : 2.5 Total hemocyte count (THC) and

superoxide anion production (as a measure of respiratory

burst) were analysed from haemolymph samples as

previ-ously described (Mercier et al 2006)

A different set of organisms was used to determine osmotic

pressure; 10 lL haemolymph was obtained from the same

ventral sinus without using anticoagulant and with a pipette

with a tip previously cooled at 4C for osmotic pressure

determination using a vapour pressure osmometer (Model

5520; Wescor, Logan, UT, USA) from different shrimp For

each experiment, n = 12 shrimp (3 shrimp for each half

tank) were sampled for each diet–density combination

For clotting time, haemolymph was obtained without using

anticoagulant from the same shrimp used for osmotic

pres-sure Haemolymph samples were withdrawn from the

peri-cardial cavity through the intersegmental membrane between

the cephalothorax and the abdominal segment Clotting time

was determined as described by Jussila et al (2001) with

minor modifications A capillary test tube (inner diameter:

1.1–1.2 mm; length: 75 mm; Corning, Monterrey, NL,

Mex-ico) was filled with haemolymph to 30% of its capacity The

tube was then turned slowly several times to allow

haemol-ymph to flow through the tube until haemolhaemol-ymph set Time

required for haemolymph to clot in the tube was registered as

the clotting time When haemolymph did not clot within

100 s, the result was not taken in consideration

Haemolymph (200 lL) was obtained from the ventral sinus

at the base of the pleopods of the first abdominal segment

using a 3-mL syringe rinsed with a 5% sodium oxalate in

isotonic saline cooled anticoagulant solution (Mercier et al

2006) For each experiment, n = 12 shrimp were sampled foreach diet–density combination Unfortunately, samples fromExperiment C + W were lost because of technical problems.Haemolymph was centrifuged at 800 g for 5 min at 4C andtotal protein, and haemocyanin was analysed immediately inplasma (Mercier et al 2006) The rest of the plasma wasstored at)70 C for lactate, glucose, triglyceride, cholesteroland total lipid analyses, as described by Palacios et al (2000).Precipitated hemocytes were carefully transferred to aglass vial using a sterile Pasteur pipette, and plasma wastransferred to Eppendorf tubes that were immediately frozen at)70 C For fatty acid determination, pools of hemocytes fromthree organisms were made for each treatment (n = 4 pooledsamples for each group of density and diet treatment), lipidswere extracted using chloroform:methanol (2 : 1 v/v, approx-imately 100 mg of sample to 6 mL of solvent) and adding 1%butylated hydroxytolune (BHT) as antioxidant and C 23 : 0 asinternal standard Fatty acids from neutral and polar lipidfractions were separated using a 6% hydrated silica gel-packedPasteur pipette as micro-column and collected in vials afterelution of neutral fraction with 10 mL chloroform : methanol(98 : 2 v/v) and of polar fraction with 15 mL of methanol,additional 23 : 0 and BHT were added to the vials containingthe polar fraction Each fraction was transesterified with bor-on-trifluoride methanol (BF3 14% methanol, Supelco), andmethyl esters were analysed in a gas chromatograph (GCModel 6890M, Agilent Technologies, Santa Clara, CA, USA)equipped with DB-23 Silica column (30 m· 0.25 mm

ID· 0.25 lm film thickness), flame ionization detector withhelium as the carrier gas (0.7 mL min)1) and a temperatureramp from 110 to 220C Fatty acids were identified by com-paring their retention times with those of standards (Sigma,Bellefonte, PA, USA) with the concentration of each fatty acidcorrected by correlation with the response of the correspond-ing standard Data were analysed using GC ChemStation Rev

A 10.02 (1757, Agilent Technologies, 2003)

After haemolymph sampling, shrimp were weighed andfrozen in liquid N2and stored at)70 C for further analyses.Frozen hepatopancreas (n = 12 shrimp) and muscle (n = 12shrimp) were dissected using a cold plate, and total protein,total lipid and carbohydrate concentrations were measured inhepatopancreas, and lactate, triglyceride, total protein, totallipid and carbohydrate concentrations were measured inmuscle, as described by Palacios et al (2000)

One-way ANCOVA was applied to compare water qualityvariables in each tank (both halves) obtained from both .

experiments, using the three density combinations (16/16, 16/

68 and 68/68 shrimp m)2) as independent variable, and time

as a covariable to eliminate daily variation throughout the

duration of the experiment

Two-way nested ANOVA was applied to analyse all

bio-logical variables within each experiment, using low- and

high-ARA diets as the first independent variable and density

(16 and 68 shrimp m)2) as the second independent variable

Differences between means for each group (individual means)

were determined by a post hoc Tukey test only when the

interaction between both independent variables was

signifi-cant Otherwise significant differences between global means

(pooled means of either dietary treatments or stress

condi-tion) are indicated in the text of the results section

Analyses were performed using STATISTICA version 5.5

(StatSoft, Tulsa, OK, USA), and differences were reported

as significant if P£ 0.05

The water physicochemical variables are reported in Fig 1;

low- (16/16, i.e., average 16 shrimp m)2) and high-density

treatment (68/68, i.e., average 68 shrimp m)2) from

Experi-ment C + W, and intermediate density from the ExperiExperi-ment

C, which correspond to a single density combination (16/68,

i.e average of 42 shrimp m)2) (see Material and methods

section for further details on grouping water quality ables) Average values of total ammonia–nitrogen (TAN) weresignificantly higher at the highest density The average TANvalues during the experiment were 0.6, 1.3 and 2.4 mg L)1forthe 16/16, 16/68 and 68/68 densities, respectively (P < 0.01) Areduction in TAN was observed for the highest density, whenwater flow was increased (1 June 2007) Dissolved oxygen wasaffected (P < 0.01) by the three different densities, with higheraverage values (5.6 mg L)1) for the low-density treatment(16/16), average values of 4.6 mg L)1 for the intermediatedensity (16/68), and lower average values (3.7 mg L)1) for thehigh-density treatment (68/68) However, the lowest value(1.8 mg L)1) was registered once during the first week for thehighest density, while values in the intermediate and lowdensity were above 3 mg L)1; an increase in dissolved oxygenwas observed for the highest density when water flow wasincreased (1 June 2007) pH decreased with increasing density(8.04, 7.95 and 7.86 for 16/16, 16/68 and 68/68, respectively;

vari-P <0.01), and no significant differences were found fortemperature between density treatments (25.4–26.0C)

Shrimp weight increased by 2.4 times in the high-densitytreatment and by 3 times in the low-density treatment duringthe Experiment C, but dietary ARA level did not affectgrowth (Table 3) The effect of density on final weight wasmore pronounced in Experiment C + W (global means from

Figure 1 Water quality in relation to the different densities: 16/16 or 68/68 shrimp m)2for the Crowding + water quality Experi- ment (C + W) and 16/68 to the Crowding Experiment (C) in which Litopenaeus vannamei juveniles were grown One-way

quality parameters, using the three densities

as the independent factor and time as a variable to eliminate daily variation throughout the duration of the experiment, and means not sharing the same letter are significantly different (P < 0.05) and shown

co-as tables inserted in each figure.

.

data pooled for both dietary treatments: 10.25 versus 8.45 g

for low and high densities, respectively, i.e., 18.4% lower

values) than in Experiment C (10.4 versus 9.4 g for low and

high densities, respectively; i.e 8.3% lower values)

Survival was not affected by density in Experiment C, and

a slight but significant decrease was observed at the high

density in Experiment C + W (Table 3) Dietary ARA level

did not affect survival in either experiment

Opposite effects of density on feed intake were observed

depending on the particular experiment, with an increased feed

intake at high density in Experiment C (global means = 0.36

versus 0.44 g tank)1per day) and a decrease in Experiment

C + W with differences in water quality in addition to

crowding (global means = 0.45 versus 0.36 g tank)1per day)

Prostglandin E2metabolite levels in plasma and hemocytesincreased with high dietary ARA level in shrimp fromExperiment C, but were not significantly affected in Experi-ment C + W (Fig 2) No differences were found for PGEM

in hemocytes or plasma in relation to density in eitherexperiment

The hemocyte fatty acid composition is shown in Table 4.The global means for main saturated fatty acids 16 : 0 were13.6–18.7% and for 18 : 0 they ranged from 11.2% to 16.6%

Figure 2 Prostaglandin E 2 metabolite levels

(pg mL)1) in hemocytes (a) and plasma (b)

of the Experiment C, and in hemocytes (c)

and plasma (d) of the Experiment C + W

.

of total fatty acids (results not shown), whereas 18:1n-9

(8.3–12%), 18:1n-7 (3.3–7%) and 20:1n-9 (1–5%) were the

main monounsaturated fatty acids (MUFA) The main

polyunsaturated fatty acids (PUFA) are shown in Table 5,

and were 22:6n-3, 20:5n-3, 18:2n-6, followed by 20:4n-6

which was significantly affected by the treatments and will be

discussed in more detail below

In Experiment C, only the level of ARA was affected; in

the neutral lipids, it was significantly higher by 64% in

shrimp fed the high-ARA diet (global means from data

pooled for density: 8% versus 13% of total fatty acids for

low- and high-ARA diets, respectively) Although there

was no significant interaction, this effect was less

pro-nounced for the high-density condition Increase in ARA

was at the expense of the decrease, although not cant, in the proportion of other fatty acids such as 20:5n-3(EPA from 12.1% to 10.6%, global means), saturated fattyacids and MUFA A similar increase in ARA (from 5.9%

signifi-to 11.2%, global means) in the polar lipids of Experiment

C was observed as a result of dietary ARA level, but nosignificant effect on EPA was found as a result of diet(Table 4)

In Experiment C + W, ARA levels in the neutral lipidssignificantly increased by 113% in response to dietaryARA level (global means 5.8 versus 12.4% of total fattyacids for low- and high-ARA diets, respectively) regardless

of the density A significant decrease in EPA from 15.2%

to 9.3% (global means, Table 4) was observed as a result

Table 4 Selected fatty acids (%, means ± SE) in hemocytes of Litopenaeus vannamei fed low- (0.8% of total fatty acids) and high- (3.2% of

total fatty acids) ARA diets, and grown at different densities (low: 16 shrimp m)2and high: 68 shrimp m)2)

ARA, arachidonic acid; MUFA, monounsaturated fatty acids.

Only the sum of main categories and selected individual fatty acids were included: 18:2n-6 as precursor of ARA (20:4n-6), and the three

essential fatty acids ARA, EPA (20:5n-3) and DHA (22:6n-3) See Table 3 for statistical analyses.

.

of dietary ARA level The n-3/n-6 ratio was significantly

higher in the low-ARA diet In the polar lipids in

Exper-iment C + W, the increase in ARA was less pronounced

than in the neutral lipids (6.2–10% of total fatty acids; i.e.,

increased by 61%) In this case, the increase in the content

of ARA in tissues resulted in a significant decrease in EPA

from 13.2–10.9% (global means, Table 4) and a

non-significant decrease in saturated fatty acids The n-3/n-6

ratio was significantly higher in the low-ARA diet

Sig-nificant interactions were found for MUFA and 18:2n-6

(Table 4)

Total hemocyte count, haemolymph osmolarity (Table 5)

and baseline superoxide anion production (not shown) were

not affected by density or dietary ARA level in either

experiment In Experiment C, a non-significant increase in

superoxide anion production was observed in shrimp fed the

high-ARA diet but only when grown at the highest density

(interaction ARA· density, P = 0.09) The clotting time

was higher for shrimp grown at low densities when fed the

low-ARA diet and higher for the low-ARA diet fed shrimp

at low density, as shown by the significant interaction

In Experiment C + W, high dietary ARA level decreased

laminarin-stimulated superoxide anion production only in

shrimp grown at the lowest density, whereas at high density,

a non-significant opposite effect was obtained, as indicated

by a significant interaction Clotting time was not affected

by density, but it was lower in shrimp fed the high-ARA

diet

For Experiment C, haemolymph glucose was significantlyhigher in shrimp fed the low-ARA diet (global means 16.5versus 11.0 mg dL)1for low- and high-ARA diets, respec-tively) (Table 6) Although no significant interaction betweendiet and density was found (P = 0.1), the effect of dietaryARA levels was greater for shrimp grown at high density.Lactate in haemolymph was also higher for the low-ARAdiet (27 versus 17.3 mg dL)1for low- and high-ARA diets,respectively) and was also significantly lower in shrimpgrown at high density (24.8 versus 19.4 for low- and highdensities ARA diets, respectively) A similar density effectwas observed for the levels of lactate in the hepatopancreaswith lower values at high density (0.6 mg g)1) than at lowdensity (0.9 mg g)1; Table 6) As shown by a significantinteraction, glycogen in muscle was differentially affected bydensity and dietary ARA level with a decrease at high densitybut only for the low-ARA diet (Table 6) No other differ-ences in biochemical composition were observed in Experi-ment C

In Experiment C + W, analysis of haemolymph was notcarried out (see Material and methods) Several biochemicalcomponents of hepatopancreas and muscle were differen-tially affected by density depending on ARA levels in the diet

as indicated by significant interactions (Table 6) Glycogenlevels in hepatopancreas increased at the high density butonly in shrimp fed the high-ARA diet, whereas the inversepattern was observed for lipids Finally, glycogen in muscleincreased in shrimp at high density but only in those fed thelow-ARA diet

Table 5 Immunological variables and haemolymph osmolarity (means ± SE) of Litopenaeus vannamei fed low- (0.8% of total fatty acids) and high- (3.2% of total fatty acids) ARA diets, and grown at different densities (low: 16 shrimp m)2and high: 68 shrimp m)2)

THC, total hemocyte count (determined on n = 6 shrimp); SO, superoxide (determined on n = 6 shrimp); AU, arbitrary units (optical density · dilution factor); ARA, arachidonic acid.

1 Laminarin-induced superoxide anion production Clotting time was determined on n = 12 shrimp See Table 3 for statistical analyses See text for comparison between global means (pooled means of either dietary treatments or stress conditions) Means not sharing the same superscript are significantly different (P < 0.05).

.

Survival was not affected by density, in contrast to previous

studies (Ray & Chien 1992; Martin et al 1998; Nga et al

2005; Li et al 2006) although the relation between survival

and growth in relation to density is not always consistent

(Moss & Moss 2004; Mena-Herrera et al 2006) Growth rate

is reduced at higher stocking densities, a well-known

influ-ence observed for L vannamei at postlarval (Moss & Moss

2004) and juvenile stages (Mena-Herrera et al 2006) The

present study shows that although crowding per se reduces

growth, the effect is more evident when there are differences

in water quality In accordance, Nga et al (2005) reported

that differences in water quality alone without crowding

(chemical stress mainly because of ammonia) were enough to

explain some growth depression in 25 mg (dw) shrimp In the

present study, maximum levels of TAN attained (2.4 mg L)1)

were far below the TAN lethal levels reported for

L vannamei in juvenile stages, which are usually above

20 mg L)1when exposed for 120-h, and closer to those ported as ÔsafeÕ that were estimated between 3.5 and6.5 mg L)1 (Frı´as-Espericueta et al 2000) Values of pHdecreased slightly with increasing densities, as observed inprevious studies (Ray & Chien 1992; Mena-Herrera et al

re-2006) which would decrease the proportion of unionizedammonia and thus would not explain a higher relativetoxicity of TAN at increasing densities

DO levels can reduce growth of L vannamei shrimpexposed to 1 mg L)1 for 16 days (Seidman & Lawrence1985) Here, the average value obtained at the highest densitywas 3.7 mg L)1, but DO levels oscillated around 2 mg L)1during 1 week at the beginning of the experiment However,high ammonia levels increase oxygen consumption (Chen &

Lin 1995; Racotta & Herna´ndez-Herrera 2000), and thephysiological effects of high ammonia levels are potentiated

by low levels of DO (Mugnier et al 2008), and ammonia

Table 6 Biochemical composition (means ± SE) of haemolymph, hepatopancreas and muscle of Litopenaeus vannamei fed low- (0.8% of total

fatty acids) and high- (3.2% of total fatty acids) ARA diets, and grown at different densities (low: 16 shrimp m)2and high: 68 shrimp m)2)

ARA, arachidonic acid.

See Table 3 for statistical analyses Means not sharing the same superscript are significantly different (P < 0.05).

.

toxicity increases at low DO (Allan et al 1990) Thus, the

interactive influence of crowding, high ammonia, and low

DO levels could explain a higher growth depression in

Experiment C + W compared to Experiment C

Crowding alone increased feed intake (Experiment C),

whereas a decrease was observed when combining crowding

with lower water quality (Experiment C + W) In

accor-dance, feed intake increased with increasing stocking

densi-ties in Penaeus chinensis, when similar TAN levels were kept

between the different densities (Li et al 2006)

It has been suggested that animals in their natural settings

produce chemicals that alert co-specifics on limiting

re-sources; this chemicals could prove troublesome when shrimp

are stocked at high densities in intensive aquaculture if they

affect growth or impair the immune system (Kamps & Neill

1999) Thus, overcrowding without a deterioration of water

quality (Experiment C) might increase competition for

available food, while crowding accompanied with decreased

water quality (Experiment C + W) might be perceived as

increased stress which can result in reduced food ingestion

and thus limited growth

Even if it was not the aim of this study, we observed that the

low-ARA diet (0.8% of total fatty acids or 86 mg 100 g)1

diet) seemed to satisfy the requirement of shrimp for this fatty

acid, as no differences on growth and survival were obtained

when compared with the high-ARA diet (3.2% of total fatty

acids or 323 mg 100 g)1 diet) Glencross & Smith (2001)

reported that growth of P monodon juveniles was not

affected by ARA levels below 4.7% in the diet, but decreased

when ARA levels increased above 10%, whereas

Gonza´lez-Fe´lix et al (2003) found better growth and survival of

L vannameijuveniles with diets containing 8% ARA, which

increased ARA to 18% in the neutral fraction of muscle after

6 weeks In diets devoid of ARA fed to P chinensis juveniles

for 2 months, ARA levels in tissues were maintained at

1–3%, in contrast to diets enriched with ARA where tissues

had up to 20% ARA (Xu et al 1994) Bottino et al (1980)

reported that 1 month was enough time to show the time

dependence for the adaptation of laboratory-reared P

se-tiferusto PUFA dietary lipids, but ARA had a slower

turn-over rate, because it was maintained at initial values (10%)

for 3 months in spite of the ARA content (0.7%) in the diet

As a comparison, wild P setiferus shrimp sampled in the Gulf

of Mexico have 8% of ARA (Bottino et al 1980) and wild

L vannameihave 5–6% of ARA (Browdy et al 2006), which

are similar to the values found in hemocytes in this study

The relationship between membrane fatty acid tion and function of immune cells is well known invertebrates (Calder 2007) and more recently for invertebrates(Le Grand et al 2011) ARA content in the diet was directlyreflected on the levels of ARA in the hemocytes; an effectmore evident in neutral lipids (e.g triglycerides) reserves than

composi-in polar lipids (i.e phospholipids), as previously observed forx3 HUFA-food imprinting in shrimp hemocytes (Mercier

et al.2009) The increase in the content of ARA in hemocytes

in shrimp fed the high-ARA diet was counterbalanced by adecrease in the proportion of EPA; similarly to what hasbeen described in fish, where EPA, DHA and ARA competefor the incorporation into phospholipids, resulting in aninverse relationship between these fatty acids in the tissues(Bell et al 1995) The DHA/ARA ratios (2.3 for the low and1.0 for the high-ARA diet) and EPA/ARA (1.6 for the lowand 0.7 for the high-ARA diet) in the membrane hemocyteswere lower compared to same ratios in the diets (17.3 forDHA/ARA and 10.6 for EPA/ARA in the low-ARA diet,and 4.6 for DHA/ARA and 2.6 for EPA/ARA in the low-ARA diet), suggesting a selective incorporation of ARA intohemocytes phospholipids, as has been observed for othertissues in tilapia fed high-ARA diet (Van Anholt et al 2004).Arachidonic acid is a precursor of several prostaglandins,and higher levels of PGE2or PGEM have been reported infish (Bell et al 1995; Lund et al 2008) and oysters (Hurtado

et al.2009) fed enriched ARA diets In Atlantic salmon thatcan elongate PUFAs to HUFAs, diets rich in 18:2n-6 pro-duced higher ARA levels in phospholipids of leucocytes andincreased plasma concentrations of certain eicosanoids (Bell

et al 1992) Here, we found that the levels of PGEMincreased threefold in hemocytes and plasma of shrimp fedthe high-ARA diet in the Experiment C It is interesting thatPGEM levels were affected in the Experiment C and not inshrimp of Experiment C + W, although the fatty acidcomposition was more affected in the later (Table 4) Thesynthesis and secretion of PG depends on the amount ofARA that can be released from phospholipids in response tothe stressor (Van Anholt et al 2004) In hemocytes, thestressor is commonly an immune challenge, during which PGare secreted from blood cells to plasma in mammals (Levy2006) and insects (Stanley & Miller 2006) Apparently, thestress induced by high densities was not enough to modifyARA or PGEM levels in shrimp PG have a short half-life,ranging from seconds to minutes; changes in levels of PGmight not be appreciated using a long-term stressor ascrowding On the other hand, the enzymes that utilize ARA

as a substrate to synthesize PG of the series II can also useEPA to synthesize PG of the series III, so these fatty acids .

compete with one another It has been suggested that an

increase in n-3/n-6 ratio increase the eicosanoid metabolic

products from ARA, such as PG, thromboxanes,

leukotri-enes, hydroxy fatty acids and lipoxins (Simopoulos 2006)

Here, the n-3/n-6 ratio of the diet was similar (0.84 for the

low-ARA diet and 0.91 for the high-ARA diet, Table 2) and

within the range established by Glencross et al (2002) for an

adequate growth range for shrimp However, the n-3/n-6

ratio was affected by the diet in the Experiment C + W, but

not in the Experiment C (Table 4) A negative correlation

was found between PGEM in hemocytes and n-3/n-6

(R =)0.89; P < 0.01) or with the EPA/ARA ratio (R =

)0.75; P < 0.01), and a positive with PGEM in hemocytes

and ARA levels (R = 0.83; P < 0.05), all in the polar

frac-tion of hemocytes Interestingly, none of the above were

significantly correlated to PGEM levels in plasma, but DHA

in polar lipids were negatively correlated (R =)0.65;

P <0.05) As with EPA, DHA competes with ARA but for

release from the n-2 position of phospholipids by

phospho-lipases, and diets rich in DHA can reduce turnover of ARA

to eicosanoids and prevent inflammation (Farooqui et al

2007; Tassoni et al 2008) Thus, a lack of diet-induced effect

on PGEM levels in shrimp in Experiment C + W compared

to Experiment C could be associated with changes in other

hemocyte fatty acids in the former, and these changes cannot

be explained in the light of the present results but could

indicate that their accumulation is differentially affected by

water quality

Clotting time is generally perceived as an indicator of

immune capability, and shrimp with shorter clotting time

are considered as better prepared to confront a stressful

condition Smith et al (2002) found that immune capability

and clotting were adversely affected in shrimp exposed to

contaminated water Shrimp under short-term stress

condi-tions such as handling (Mercier et al 2009) or tail flipping

(Jussila et al 2001) had higher clotting time Here, we found

that clotting time was lower in shrimp fed the high-ARA

diet for both experiments (C and C + W), which could

indicate that they were better prepared to respond to a

stressful condition PGE2 is not generally associated with

clotting in vertebrates, although dietary supply of linoleic

acid in mammals has been shown to increase levels of ARA,

PGF2a and of thromboxane B2 (TBX2), in platelet

phos-pholipids (Hwang et al 1982) Thromboxanes are strong

clotting agents in vertebrates (Wu et al 2005) Thus, the

increase in PGEM in experiment C is probably parallel to

the decrease in clotting time, and not directly responsiblefor it

Several studies have clearly shown the influence of TANand DO levels on shrimp immune response capacity thatcould be similar to the conditions found in the Experiment

C + W of the present study Exposure of L vannamei tohigh levels of TAN (5.2 mg L)1or more) (Liu & Chen 2004)

or nitrites (Tseng & Chen 2004) increased respiratory bursttogether with other effects that were considered as indicators

of a decreased immuno competence In contrast, hypoxia hasbeen shown to decrease respiratory burst in L stylirostris (LeMoullac et al 1998) and Machrobrachium rosenbegii (Cheng

et al 2002), and a decreased immunocompetence was alsorecognized Considering this with the present results, the in-crease in respiratory burst (superoxide anion production) inshrimp stocked at high density could be a result of highammonia and nitrite levels, rather than to DO level orcrowding Nevertheless, a clear depression of immuneresponse capacity was not observed because THC was notaffected and, as discussed above, an increase in respiratoryburst does not directly pointed out an immune depression Inaddition to the effect of TAN and DO, we can speculate thatdifferences in water quality in Experiment C + W alsoinclude a differential microbial load between both conditionsthat in turn could increase respiratory burst, in accordance toinfluence of microbial substances or live bacteria on super-oxide anion production (Mun˜oz et al 2000) In any case, thehigh-density-induced increase in respiratory burst wasobserved together with high-ARA levels in hemocytes Anincrease in respiratory burst associated with high-ARA levels

in hemocyte phospholipids was also observed in oysters withARA supplementation in the water (Delaporte et al 2006)

Contrary to the original hypothesis of the present work, PGwere apparently not involved because no changes of PGEMlevels were observed in Experiment C + W It is possiblethat instead, other eicosanoids such as leukotriene LTB4 areinvolved, as they were shown to enhance production ofreactive oxygen species in mammals (Calder 2001) In ver-tebrates, it has been shown that LTB4 increase in parallelwith PGE2as a direct result of increased ARA proportion inimmune cell membrane phospholipids and that their levelsare modulated by dietary supply of ARA (Kelley 2001;

Calder 2007)

Lactate levels in haemolymph obtained in Experiment C(17.3–27 mg dL)1) were relatively higher than baseline valuesreported in other studies for the same species, which range .

between 4 and 10 mg dL)1 (Racotta & Palacios 1998;

Racotta & Herna´ndez-Herrera 2000; Mercier et al 2006) In

L vannamei, lactate levels have been reported to increase

over 15 mg dL)1 during stressful conditions following

manipulation stress (Mercier et al 2009; Aparicio-Simo´n

et al.2010) Thus, values obtained in the present work are

more similar to values attained by stressed shrimp; stress

because of sampling itself (Mercier et al 2006, 2009;

Aparicio-Simo´n et al 2010) probably occurred in the present

work, because no special precautions to avoid stress were

taken If so, lactate levels would reflect the magnitude of

stress response that seems to be differential depending on the

treatment Hence, ARA dietary supply apparently blunted

the increase in lactate induced by sampling stress, because its

levels were lower (i.e the increase was less pronounced in

response to stress) with the high-ARA diet On the other

hand, lactate levels in haemolymph and hepatopancreas were

lower in shrimp grown at higher densities; it is possible that

continuous crowding increased the stimulus threshold to

elicit the response of shrimp to sampling stress

Shrimp fed the low-ARA diet also had higher glucose

levels but only when grown at high density Hyperglycemia

is a well-known short-term stress response in Crustacea

(Hall & Van Ham 1998; Racotta & Palacios 1998; Lorenzon

et al.2008) that can also be observed under some chronic

conditions of stress (Mercier et al 2009) In the present

work, the slight increase in glucose levels in crowded

con-ditions was not observed when shrimp were fed a high-ARA

diet In parallel, this effect was also observed for glycogen

levels in muscle: apparently, this reserve was used under the

long-term stress condition of crowding but its use was

prevented by a high content of ARA in the diet Indeed,

muscle glycogen mobilization could be the main source of

haemolymph glucose because it has not been described if

mobilization for glucose release preferentially occurs from

hepatopancreas or muscle glycogen reserves Several

crusta-cean hyperglycemic agents such as Crustacrusta-cean Hyperglycemic

Hormone (Sedlmeier 1985) and biogenic amines (Hsieh et al

2006) act both on hepatopancreas and muscle The

mecha-nisms by which ARA interferes with high

density-stress-induced glycogen mobilization, associated hyperglycemia

and anaerobic glycolisis with lactate production remain to

be elucidated

The metabolic response in haemolymph could not be

evaluated in the C + W experiment However, in tissues,

some responses were quite different than those obtained with

high density consisting of crowding only (Experiment C)

The decrease in muscle glycogen at high density was no

longer observed; on the contrary, an increase in glycogen in

muscle (low-ARA diet) or hepatopancreas (high-ARA diet)was observed in shrimp at high densities, but in the light ofthe present results, it is difficult to explain these differences

Crowding itself was enough to reduce weight gain, althoughthe effect was more pronounced when a deterioration ofwater quality exist as a result of high stocking density, whichcould be in turn attributed to the combined effects of hightotal ammonia levels and low dissolved oxygen levels

A possible beneficial role of ARA in the diet counteractingthe negative effects of high density was not clearly observed

on overall performance However, some results suggest thatARA could minimize stress response because of samplingand enhance some effectors of the immune system such asclotting and respiratory burst The increase in PGE2metabolite in shrimp fed the high-ARA diet was not con-sistent in both experiments, and thus, the effects of ARAwere not necessarily mediated by these eicosanoids

This research was funded by SEP-CONACYT (Grants 43249and 49191) The authors thank O Arjona, L Mercier, M.A.Hurtado, A Santamarı´a, A Carren˜o, M Reza, R Floresand R Herna´ndez-Herrera for their help during samplecollecting and analyses Vero´nica Aguilar was a recipient of aCONACYT fellowship

Akiyama, D.M & Dominy, W.G (1989) Penaeid shrimp nutrition for the commercial feed industry Texas Agricultural Extension Service and Texas A&M University, Sea Grant College Program,

50 p.

Allan, G.L & Maguire, G.B (1992) Effects of stocking density on production of Penaeus monodon Fabricius in model farming ponds Aquaculture, 107, 49–66.

Allan, G.L., Maguire, G.B & Hopkins, S.J (1990) Acute and chronic toxicity of ammonia to juvenile Metapenaeus macleayi and Penaeus monodon and the influence of low dissolved-oxygen levels Aquaculture, 91, 265–280.

AOAC (2005) Official Methods of Analysis, 18th edn Association of Official Analytical Chemists, Gaithersburg, MD, USA.

Aparicio-Simo´n, B., Pin˜o´n, M., Racotta, R & Racotta, I.S (2010) Neuroendocrine and metabolic responses of Pacific whiteleg shrimp Litopenaeus vannamei exposed to acute handling stress Aquaculture, 298, 308–314.

Bell, J.G., Dick, J.R., Sargent, J.R & McVicar, A.H (1992) Dietary linoleic acid affects phospholipid fatty acid composition in heart and eicosanoid production by cardiomyocytes from Atlantic salmon Salmo salar Comp Biochem Physiol., 103A, 337–342.

.

Bell, J.G., Castell, J.D., Tocher, D.R., MacDonald, F.M & Sargent,

J.R (1995) Effects of different dietary arachidonic acid:

docosa-hexaenoic acid ratios on phospholipid fatty acid compositions and

prostaglandin production in juvenile turbot (Scophthalmus

maxi-mus) Fish Physiol Biochem., 14, 139–151.

Bottino, N.R., Gennity, J., Lilly, L., Simmons, E & Finne, G (1980)

Seasonal and nutritional effects on the fatty acids of three species

of shrimp, Penaeus setiferus, P aztecus and P duorarum

Aqua-culture, 19, 139–148.

Browdy, C., Seaborn, G., Atwood, H., Davis, A.D., Bullis, R.A.,

Samocha, T.M., Wirth, E & Leffler, J.F (2006) Comparison of

pond production efficiency, fatty acid profiles, and contaminants

in Litopenaeus vannamei fed organic plant-based and

fish-meal-based diets J World Aquac Soc., 37, 437–451.

Calder, P.C (2001) N-3 polyunsaturated fatty acids, inflammation

and immunity: pouring oil on troubled waters or another fishy

tale? Nutr Res., 21, 309–341.

Calder, P.C (2007) Immunomodulation by omega-3 fatty acids.

Prostag Leukotr Ess., 77, 327–335.

Camba, E., Pedrazzoli, A., Yaguachi, M & Akiyama, D.M (1993)

Requerimiento de proteı´nas en dietas artificiales para juveniles de

Penaeus vannamei Acuicultura Trop., 1, 7–12.

Chen, J.C & Lin, C.Y (1995) Responses of oxygen consumption,

Ammonia-N excretion and Urea-N excretion of Penaeus chinensis

exposed to ambient ammonia at different salinity and pH levels.

Aquaculture, 136, 243–255.

Cheng, W., Liu, C.H., Hsu, J.P & Chen, J.C (2002) Effect of

hypoxia on the immune response of giant freshwater prawn

Machrobrachium rosenbergii and its susceptibility to pathogen

Enterococus Fish Shellfish Immunol., 13, 351–365.

Chim, L., Lemaire, P., Delaporte, M., Le Moullac, G., Galois, R &

Martin, J.L.M (2001) Could a diet enriched with n-3 highly

unsaturated fatty acids be considered a promising way to enhance

the immune defenses and the resistance of Penaeid prawns to

environmental stress? Aquac Res., 32, 91–94.

Delaporte, M., Soudant, P., Moal, J., Giudicelli, E., Lambert, C.,

Seguineau, C & Samain, J.F (2006) Impact of 20:4n-6

supple-mentation on the fatty acid composition and hemocyte parameters

of the Pacific Oyster Crassostrea gigas Lipids, 41, 567–576.

Farooqui, A.A., Horrocks, L.A & Farooqui, T (2007) Modulation

of inflammation in brain: a matter of fat J Neurochem., 101, 577–

599.

Frı´as-Espericueta, M.G., Harfush-Melendez, M & Pa´ez-Osuna, F.

(2000) Effects of ammonia on mortality and feeding of postlarvae

shrimp Litopenaeus vannamei Bull Environ Contam Toxicol., 65,

98–103.

Glencross, B.D & Smith, D.M (2001) A study of the arachidonic

acid requirements of the giant tiger prawn, Penaues monodon.

Aquac Nutr., 7, 59–69.

Glencross, B.D., Smith, D.M., Thomas, M.R & Williams, K.C.

(2002) The effect of dietary n-3 and n-6 fatty acid balance on the

growth of the prawn Penaeus monodon Aquac Nutr., 8, 43–51.

Gonza´lez-Fe´lix, M., Gatlin, D.M III, Lawrence, A.L &

Perez-Velazquez, M (2003) Nutritional evaluation of fatty acids for the

open thelycum shrimp, Litopenaeus vannamei: II Effect of dietary

n-3 and n-6 polyunsaturated and highly unsaturated fatty acids on

juvenile shrimp growth, survival, and fatty acid composition.

Aquac Nutr., 9, 115–122.

Hall, M.R & Van Ham, E.H (1998) The effects of different types of

stress on blood glucose in the giant tiger prawn Penaeus monodon.

J World Aquac Soc., 29, 290–299.

Hsieh, S.L., Chen, S.M., Yang, Y.H & Kuo, C.M (2006)

Involve-ment of norepinephrine in the hyperglycemic responses of the

freshwater giant prawn, Macrobrachium rosenbergii, under cold shock Comp Biochem Physiol., 143A, 54–263.

Hurtado, M.A., Reza, M., Ibarra, A.M., Wille, M., Sorgeloos, P., Soudant, P & Palacios, E (2009) Arachidonic acid (20:4n )6) effect on reproduction, immunology, and prostaglandin E2 levels

in Crassostrea corteziensis (Hertlein, 1951) Aquaculture, 294, 300–

305.

Hwang, D.H., Chanmugam, P & Anding, R (1982) Effects of dietary 9-trans,12-trans linoleate on arachidonic acid metabolism

in rat platelets Lipids, 17, 307–313.

ICES (1997) Report of the Working Group on Marine Fish Culture for the International Council for the Exploration of the Sea, Murcia, Spain, 23–26 June 1997, ICES CM1997/F,7.

Jussila, J., McBride, S., Jago, J & Evans, L.H (2001) Hemolymph clotting time as an indicator of stress in western rock lobster (Panulirus cygnus George) Aquaculture, 199, 185–193.

Kamps, R.H & Neill, W.H (1999) Aquacultural effluents: directive signals to the system downstream? J Chem Ecol., 25, 2041–2050.

Kelley, D.S (2001) Modulation of human immune and inflammatory responses by dietary fatty acids Nutrition, 17, 669–673.

Koven, W., Van Anholt, R., Lutzky, S., Atia, I.B., Nixon, O., Ron,

B & Tandler, A (2003) The effect of dietary arachidonic acid on growth, survival, and cortisol levels in different-age gilthead seabream larvae (Sparus auratus) exposed to handling or daily salinity change Aquaculture, 228, 307–320.

Le Grand, F., Kraffe, E., Marty, Y., Donaghy, L & Soudant, P.

(2011) Membrane phospholipid composition of hemocytes in the Pacific oyster Crassostrea gigas and the Manila clam Ruditapes philippinarum Comp Biochem Physiol., 159A, 383–391.

Le Moullac, G & Haffner, P (2000) Environmental factors affecting immune responses in Crustacea Aquaculture, 191, 121–131.

Le Moullac, G., Soyez, C., Saulnier, D., Ansquer, D., Avarre, J.C &

Levy, P (1998) Effect of hypoxic stress on the immune response and the resistance to vibriosis of the shrimp Penaeus stylirostris.

Fish Shellfish Immunol., 8, 621–629.

Lemonnier, H., Herbland, A., Salery, L & Soulard, B (2006)

‘‘Summer syndrome’’ in Litopenaeus stylirostris grow out ponds in New Caledonia: zootechnical and environmental factors Aqua- culture, 261, 1039–1047.

Levy, R (2006) The role of cytosolic phospholipase A2-alfa in ulation of phagocytic functions Biochim Biophys Acta, 1761, 1323–1334.

reg-Li, Y., reg-Li, J & Wang, Q (2006) The effects of dissolved oxygen concentration and stocking density on growth and non-specific immunity factors in Chinese shrimp, Fenneropenaeus chinensis.

Aquaculture, 256, 608–616.

Liu, C.H & Chen, J.C (2004) Effect of ammonia on the immune response of white shrimp Litopenaeus vannamei and its susceptibility to Vibrio alginolyticus Fish Shellfish Immunol., 16, 321–334.

Lorenzon, S., Giulianini, P.G., Libralato, S., Martinis, M & rero, E.A (2008) Stress effect of two different transport systems on the physiological profiles of the crab Cancer pagurus Aquaculture,

Fer-278, 156–163.

Lund, I., Steenfeldt, S.J., Banta, G & Hansen, B.W (2008) The influence of dietary concentrations of arachidonic acid and eico- sapentaenoic acid at various stages of larval ontogeny on eye migration, pigmentation and prostaglandin content of common sole larvae (Solea solea L.) Aquaculture, 276, 143–153.

Martin, J.L.M., Veran, Y., Guerlorget, O & Pham, D (1998) Shrimp rearing: stocking density, growth, impact on sediment, waste output and their relationships studied through the nitrogen budget in rearing ponds Aquaculture, 164, 135–149.

.

Mena-Herrera, A., Gutierrez-Corona, C., Linan-Cabello, M &

Sumano-Lopez, H (2006) Effects of stocking densities on growth

of the Pacific white shrimp (Litopenaeus vannamei) in earthen

ponds Isr J Aquac Bamid., 58, 205–213.

Mercier, L., Palacios, E., Campa-Co´rdova, A.I., Tovar-Ramı´rez, D.,

Herna´ndez-Herrera, R & Racotta, I.S (2006) Metabolic and

immune responses in Pacific whiteleg shrimp Litopenaeus vannamei

exposed to a repeated handling stress Aquaculture, 258, 633–640.

Mercier, L., Racotta, I.S., Yepiz-Plascencia, G., Muhlia-Almazan,

A., Civera, R., Quin˜ones-Arreola, M., Wille, M., Sorgeloos, P &

Palacios, E (2009) Effect of diets containing different levels of

highly unsaturated fatty acids on physiological and immune

responses in Pacific whiteleg shrimp Litopenaeus vannamei (Boone)

exposed to handling stress Aquacult Res., 40, 1849–1863.

Moss, K.R.K & Moss, S.M (2004) Effects of artificial substrate and

stocking density on the nursery production of Pacific white shrimp

Litopenaeus vannamei J World Aquac Soc., 35, 536–542.

Mourente, G & Rodriguez, A (1997) Effects of salinity and dietary

DHA (22:6n-3) content on lipid composition and performance of

Penaeus kerathurus postlarvae Mar Biol., 128, 289–298.

Mugnier, C., Zipper, E., Goarant, C & Lemonnier, H (2008)

Combined effect of exposure to ammonia and hypoxia on the blue

shrimp Litopenaeus stylirostris survival and physiological response

in relation to molt stage Aquaculture, 274, 398–407.

Mun˜oz, M., Ceden˜o, R., Rodriguez, J., Van der Knaap, W.P.W.,

Mialhe, E & Bache`re, E (2000) Measurement of reactive oxygen

intermediate production in haemocytes of the penaeid shrimp

Penaeus vannamei Aquaculture, 191, 89–107.

Nga, B.T., Lu¨rling, M., Peeters, E.T.H.M., Roijackers, R., Scheffer,

M & Nghia, T.T (2005) Chemical and physical effects of

crowding on growth and survival of Penaeus monodon Fabricius

post-larvae Aquaculture, 246, 455–465.

Palacios, E., Ibarra, A.M & Racotta, I.S (2000) Tissue biochemical

composition in relation to multiple spawning in wild and

pond-reared Penaeus vannamei broodstock Aquaculture, 185, 353–

371.

Palacios, E., Bonilla, A., Pe´rez, A., Racotta, I.S & Civera, R (2004)

Influence of highly unsaturated fatty acids on the responses of

white shrimp Litopenaeus vannamei (Boone) postlarvae to low

salinity J Exp Mar Biol Ecol., 299, 201–215.

Racotta, I.S & Herna´ndez-Herrera, R (2000) Metabolic responses

of the white shrimp, Penaeus vannamei, to ambient ammonia.

Comp Biochem Physiol., 125A, 437–443.

Racotta, I.S & Palacios, E (1998) Hemolymph metabolic variables

in response to experimental manipulation stress and serotonin

injection in Penaeus vannamei J World Aquac Soc., 29, 351–356.

Ray, W.M & Chien, Y.H (1992) Effects of stocking density and

aged sediment on tiger prawn, Penaeus monodon, nursery system.

Aquaculture, 104, 231–248.

Rees, J.F., Cure, K., Piyatiratitivorakul, S., Sorgeloos, P & Menasveta, P (1994) Highly unsaturated fatty acid requirements

of Penaeus monodon postlarvae: an experimental approach based

on Artemia enrichment Aquaculture, 122, 193–207.

Sedlmeier, D (1985) Mode of action of the crustacean hyperglycemic hormone Am Zool., 25, 223–232.

Seidman, E.R & Lawrence, A.L (1985) Growth, feed digestibility, and proximate body composition of juvenile Penaeus vannamei and P monodon grown at different dissolved oxygen levels.

J World Maricul Soc., 16, 333–346.

Simopoulos, A.P (2006) Evolutionary aspects of diet, the omega-6/ omega-3 ratio and genetic variation: nutritional implications for chronic diseases Biomed Pharmacother., 60, 502–507.

Smith, D.M., Burford, M.A., Tabrett, S.J., Irvin, S.J & Ward, L (2002) The effect of feeding frequency on water quality and growth

of the black tiger shrimp (Penaeus monodon) Aquaculture, 207, 125–136.

Stanley, D.W & Miller, J.S (2006) Eicosanoid actions in insect cellular immune functions Entomol Exp Appl., 119, 1–13 Tackaert, W., Abelin, P., Dhert, P., Le´ger, P., Grymonpre´, D., Bombeo, R & Sorgeloos, P (1989) Stress resistance in postlarval penaeid shrimp reared under different feeding procedures J World Aquac Soc., 20, 74A.

Tassoni, D., Kaur, G., Weisinger, R.S & Sinclair, A.J (2008) The role of eicosanoids in the brain Asia Pac J Clin Nutr., 17, 220– 228.

Tseng, I.T & Chen, J.C (2004) The immune response of white shrimp Litopenaeus vannamei and its susceptibility to Vibrio alginolyticus under nitrite stress Fish Shellfish Immunol., 17, 325– 333.

Van Anholt, R.D., Spanings, F.A.T., Koven, W.M & Bonga, S.E (2004) Dietary supplementation with arachidonic acid

Wendelaar-in tilapia (Oreochromis mossambicus) reveals physiological effects not mediated by prostaglandins Gen Comp Endocrinol., 139, 215–226.

Vargas-Albores, F., Guzman, M.A & Ochoa, J.L (1993) An coagulant solution for haemolymph collection and prophenoloxi- dase studies of penaeid shrimp (Penaeus californiensis) Comp Biochem Physiol., 106A, 299–303.

anti-Wu, J.L., Namikoshi, A., Nishizawa, T., Mushiake, K., Teruya, K.

& Muroga, K (2001) Effects of shrimp density on transmission of penaeid acute viremia in Penaeus japonicus by cannibalism and the waterborne route Dis Aquat Org., 47, 129–135.

Wu, K., Matijevic-Aleksic, N & Dahlback, B (2005) Molecular aspects of thrombosis and antithrombotic drugs Crit Rev Clin Lab Sci., 42, 249–277.

Xu, X., Wenjuan, J., Castell, J.D & ÕDor, R (1994) Essential fatty acid requirement of the Chinese prawn, Penaeus chinensis Aqua- culture, 127, 29–40.

.

1,2 2 1 3 2 1 2

2

1

Key Laboratory of Ecology and Environmental Science in Guangdong Higher Education, Guangdong Provincial Key Laboratory

for Healthy and Safe Aquaculture, College of Life Science, South China Normal University, Guangzhou, China;2 Institute of

Animal Science, Guangdong Academy of Agricultural Sciences, Guangzhou, China; 3 National Institute of Nutrition and

Seafood Research, Bergen, Norway

b-1,3-Glucan with 1.0 g kg)1 was supplemented to a basal

diet [control (C)] with different feeding schedules:

perma-nently b-1,3-glucan diet (BG) and 14 days BG–14 days

con-trol diet (C + BG) Growth and immunological responses

[respiratory burst (RB), superoxide dismutase (SOD),

cata-lase (CAT), lysozyme and total protein] in haemolymph,

hepatopancreas and muscle of Litopenaeus vannamei were

recorded after 84-day feeding and exposure to nitrite-N

(20 mg L)1) for 120 h b-1,3-Glucan administration did not

affect growth performance However, as compared with

control, elevated CAT and lysozyme activities were seen in

haemolymph of both BG groups, while significantly higher

activities of SOD, lysozyme and RB in hepatopancreas, and

higher activities of CAT and lysozyme in muscle were only

seen in C + BG group After nitrite-N stress, significantly

higher haemolymph protein, and hepatopancreas activities of

lysozyme and RB were observed in both BG groups, but

significantly higher activities of haemolymph SOD was only

seen in C + BG group The mortality in groups BG and

C + BG was significantly lower than that in group C, but

C + BG group showed a trend of higher nitrite-N resistance

compared with BG group Considering dietary cost and

immunostimulatory effects, the feeding schedule with 14 days

BG–14 days control diet is more recommended

key words: b-1,3-glucan, immune response, Litopenaeus

vannamei, nitrite-N stress

Received 5 May 2010, accepted 5 July 2011

Correspondence: Jun-Ming Cao, Institute of Animal Science, Guangdong

Academy of Agricultural Sciences, Guangzhou 510640, China E-mail:

junmcao@163.com and An-li Wang, Key Laboratory of Ecology and

Environmental Science in Guangdong Higher Education, Guangdong Provincial Key Laboratory for Healthy and Safe Aquaculture, College of Life Science, South China Normal University, Guangzhou 510631, China.

E-mail: wanganl@scnu.edu.cn

In aquaculture, immunostimulants are substances that ulate immune responses and increase the ability of the im-mune system to fight against infections and diseases Among

stim-a number of commercistim-al immunostimulstim-ants, b-glucstim-ans,which consist of glucose units linked through b-1,3 or b-1,6glycosidic linkage, are regarded as an effective immuno-stimulant and have been successfully used to enhance resis-tance of crustaceans against parasitic, bacterial and viralinfections (Robertsen et al 1990; Sung et al 1994; Scholz

et al.1999; Chang et al 2003; Misra et al 2004) However,some researchers demonstrated that the immunostimulatoryeffect of b-glucans may be a short-term response (Sakai1999) Sung et al (1996) observed that the apparently bene-ficial effects of glucan on Penaeus monodon only lasted for amaximum of 24 h before declining to control levels Chang

et al (2000) also reported that the immunostimulatoryenhancement in P monodon adults peaked at day 24 afterdietary b-1,3-glucan (2.0 g kg)1) exposure and subsequentlydecreased to the prefeeding level at the end of the 40-dayfeeding trial In our previous study, a significant increase wasobserved in immune parameters after 14 days in Litopenaeusvannamei, when a diet with 1.0 g kg)1of b-1-3-glucan wasadministered permanently during 35 days (Wang 2007)

Therefore, taking into consideration that long-term uous b-1,3-glucan administration may not maintain highimmunostimulatory effects in crustaceans, but increase thedietary cost, it is necessary to investigate the effect of

contin-2012 18; 272–280 .doi: 10.1111/j.1365-2095.2011.00893.x

.Aquaculture Nutrition