Aquaculture research, tập 41, số 12, 2010

Aquaculture Research, 2010, 41, 1727^1740

doi:10.1111/j.1365-2109.2009.02450.x

Natural zooplankton as larval feed in intensive
rearing systems for juvenile production of Atlantic cod
(Gadus morhua L.)
Kjersti Eline TÖnnessen Busch1, Inger-Britt Falk-Petersen1, Stefano Peruzzi1, Nora Arctander Rist2
& Kristin Hamre3
1

Department of Arctic and Marine Biology, University of TromsÖ,TromsÖ, Norway

2

Lo¢lab AS, Steine, Stamsund, Norway

3

National Institute of Nutrition and Seafood Research (NIFES), Bergen, Norway

Correspondence: K E T Busch, Department of Aquatic BioSciences, Norwegian College of Fishery Science, University of TromsÖ, N-9037
TromsÖ, Norway. E-mail:

Abstract

Introduction

The growth potential of cod larvae is not fully
achieved when rotifers (Brachionus spp.) are used

as live feed. In this experiment, we studied the effect of natural zooplankton (mainly copepods) on
the growth of cod (Gadus morhua L.) larvae reared
in intensive systems. Using a growth model developed for cod larvae, the growth rates observed
could be evaluated and compared with growth
rates reported previously . The cod larvae showed
optimal growth rates until age 19 days post hatch
(DPH) when they reached 9.77 Æ 0.25 mm standard length (SL). Early weaning (20^25 DPH) resulted in signi¢cantly longer larvae at age 30 DPH
compared with late weaning (25^32 DPH); however, in this period, the zooplankton concentrations were low. The experimental larvae showed
considerably higher growth rates compared with
rotifer (Brachionus spp.)-reared cod larvae in previous experiments. The nutritional composition of
cod larvae was analysed and compared with published results on rotifer-reared larvae. The levels of
iodine, manganese, selenium and n-3 PUFA were
considerably higher in larvae fed copepods compared with larvae fed rotifers. The di¡erences in
nutritional status may well explain the di¡erences
in growth observed between copepod and rotiferreared larvae.

Aquaculture of cod (Gadus morhua L.) is a growing industry in Canada, Faroe Islands, Iceland, Norway,
Scotland, United Kingdom and United States (Brown,
Minko¡ & Puvanendran 2003; Watson, Sales, Cumming, Fitzsimmons, Walden, Arthur, Saravanan &
McEvoy 2006; Bj˛rnsson, Steinarsson & Arnason
2007; Rosenlund & Halldorsson 2007). Norway is by
far the largest producer of farmed cod, and in 2006
the domestic production was 11087 tonnes whole
¢sh and10.5 million juveniles (Directorate of ¢sheries
2007). A reliable production of high-quality cod juveniles is essential to ensure an economically sustainable production. There are in principle two di¡erent
approaches to juvenile production of cod; the extensive and the intensive method. Extensive ¢sh farming
relies on an enclosed ecosystem where the food is
produced within the system, while an intensive rearing system requires a constant food supply due to its
small size and high larval densities (van der Meeren
& Naas 1997).

There is a long tradition for extensive production of
cod fry in Norway, the ¢rst attempt being carried out
at the FlÖdevigen research station in 1885 (Solemdal,
Dahl, Danielssen & Moksness 1984). After 1975, a
number of experiments have been carried out in
ponds and lagoons with the aim of studying ¢sh larval feeding, growth and survival (van der Meeren &
Naas 1997; SvÔsand, OtterÔ & Taranger 2004). In the
1980s, relatively high numbers of cod juveniles were
produced in ponds by commercial companies and

Keywords: Gadus morhua, copepod, rotifer, growthrate, nutrition, fatty acid
r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd

1727


Farmed cod larvae fed natural zooplankton K E T Busch et al.

government-¢nanced institutions for sea ranching
and aquaculture purposes. After some years of
scarce production due to low prices, production increased again after 2000 (Engelsen, Asche, Skjennum & Ado¡ 2004). At present, there is production
of cod juveniles in ponds with or without submerged
bags at three to four sites in Norway. For a description
of the bag-in-pond system, see van der Meeren and
Naas (1997). Among the main challenges that cod
farmers face relying on the extensive method are the
seasonality of plankton production and the low
stocking densities of cod larvae resulting in a limited
production of juveniles.

To avoid the seasonality and to increase production, development in cod culture has moved towards
an intensive production in indoor systems with a
year-round production of larvae fed rotifers (Brachionus spp.) (Rosenlund & Halldorsson 2007). At present,
there are 10 intensive cod juvenile producers in Norway. Production methods have, to a large extent,
been adopted from intensive production of other
marine species like European seabass (Diecentrarchus
labrax L.) and gilthead seabream (Sparus aurata L.)
with a reliance on rotifers and Artemia salina L. as
live feed during the ¢rst weeks of culture (SvÔsand
et al. 2004). A growing body of scienti¢c studies has
contributed to the intensive rearing methods of cod
larvae (Puvanendran & Brown 1999, 2002; Baskerville-Bridges & Kling 2000a, b, c; Brown et al. 2003;
Puvanendran, Burt & Brown 2006; Fletcher, Roy, Davie, Taylor, Robertson & Migaud 2007); however,
there are still challenges to be met in optimizing production and juvenile quality.
Although not well documented in the scienti¢c literature, there might be large di¡erences in the quality, in particular the growth potential, between
intensively and extensively reared cod juveniles.
Higher rates of normal pigmentation and a higher
stress resistance in marine larvae reared on natural
zooplankton compared with those reared on rotifers
or Artemia have been reported (reviewed in StÖttrup
2000). To our knowledge, only one study has compared long-term di¡erences in growth rates and deformities of extensively vs. intensively reared cod.
Imsland, Foss, Koedijk, Folkvord, Stefansson and Jonassen (2006) concluded that extensively reared cod
grew faster and had a lower rate of deformities compared with intensively reared cod, but the two experimental groups were initially raised at di¡erent
production sites, rendering a direct comparison di⁄cult (Imsland et al. 2006). Based on the assumption
that the fast growth rates of extensively reared cod

1728

Aquaculture Research, 2010, 41, 1727^1740

juveniles are due to the feed provided during the ¢rst
weeks of life, we hypothesized to obtain similar high
growth rates in an intensive system if natural zooplankton replaced rotifers as the start feed. For this
purpose, we combined the two technologies by stocking cod larvae at high densities in tanks where natural zooplankton were provided as feed during the ¢rst
weeks of life and where weaning occurred at an early
age. The nutritional composition of cod larvae reared
on natural zooplankton was analysed to serve as a
reference for rotifer-fed larvae.
A size- and temperature-dependent growth model
(STDG model) was developed for cod larvae by Folkvord (2005) and the use of this model allows comparison of growth rates of cod larvae reared at di¡erent
temperatures.We applied the STDG model to evaluate
the growth rates of cod larvae in this experiment.
Further, to test if the high growth rates achieved by
extensively reared cod larvae are likely due to the feed
(i.e. natural zooplankton) or to the rearing protocols
(low density of larvae, very large volumes), the
growth rates of cod larvae in a commercial semi-extensive production was compared with the intensively reared experimental larvae by the use of the
STDG model. Finally, the STDG model (Folkvord
2005) was applied to evaluate published growth rates
of rotifer-reared larvae.

Materials and methods
Cod larvae and natural zooplankton
This experiment was conducted at the cod hatchery
Lo¢lab AS in Lofoten, Northern Norway. Eggs were
collected from common spawning tanks of the ¢rstgeneration selected broodstock at the National
Breeding Programme, TromsÖ in 2 subsequent days
and incubated at 5.5 1C. At the age of 110^115 day
degrees post fertilization and 2 days post hatch
(DPH), the larvae were transported by airplane to

the experimental site. Here, the larvae were stocked
at 50 larvae L À 1 in three outdoor tanks of 3000 L, at
a total of 150 000 larvae per tank.
Natural zooplankton, primarily copepods, was
supplied from a 300.000 m3 enclosed seawater pond
regularly fertilized to enhance primary production.
The zooplankton was harvested from the pond by
three plankton concentrators where the pond water
was ¢ltered through a 120 mm Unik ¢lter (Unik Filtering Systems, Oslo, Norway). The concentrated plankton was then separated into the following size
fractions: small (80^150 mm), medium (150^290 mm)

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740


Aquaculture Research, 2010, 41, 1727^1740

and large (290^1000 mm) using a Unik ¢lter. The
small size fraction contained mostly copepod nauplii,
the medium fraction was dominated by copepodites,
while the large fraction contained mainly adult copepods. The small fraction of zooplankton was supplied
to the larval tanks from day 2 to 8 DPH, followed by a
mixture of the small and medium fraction from day 9
to 12 and a mixture of all three plankton fractions
from day 13 to the end of live feeding. Daily samples
of the zooplankton mixture fed to the experimental
tanks were collected and later analysed under a dissection microscope to determine the proportions of
nauplii, copepodites and copepods. Natural zooplankton was fed to the tanks at 08:00, 14:00 and
20:00 hours. The objective was to supply the tanks
with 1500 prey items L À 1 at each feeding. Unfortunately, at age 14 DPH the zooplankton concentration

in the pond dropped dramatically and the experimental tanks were thereafter supplied with the maximum amount of zooplankton available.

Farmed cod larvae fed natural zooplankton K E T Busch et al.

an eye-piece micrometer. At the day of sampling, 10
larvae from each tank were measured for SL (live
SL) before being freeze-stored at À 20 1C. The other
10 larvae were measured after freezing (frozen SL).
The shrinkage caused by freezing was estimated by
measuring 10 larvae from each sampling before and
after freezing and the following relationship was
found: live SL 5 (frozen SL À 0.6491)/0.8652,
R 2 50.98.
All lengths were converted to live SL (hereafter SL).
To obtain the dry mass (DM), all larvae were dried at
60 1C for 24 h and subsequently weighed using a precision balance (MX5 microbalance; Mettler-Toledo
AS, Oslo, Norway). Because of the large size of the
tanks, daily mortality could not be easily established.
At the day of termination, the biomass of larvae from
each tank was measured, and the number of juveniles was calculated by dividing the biomass by the
mean wet weight of the juveniles. The total mortality
was corrected for sampling mortality, where each
sampled larvae was counted as 0.5 survivors.

Weaning experiment
At age 19 DPH, larvae in each of the three original
tanks were split in two and placed into six new tanks
of 3000 L volume. In three tanks, the larvae were
weaned on formulated feed (Gemma micro 150 and
Gemma micro 300, Skretting, Norway) from day 20

to 25 DPH (Early weaning, EW) while in the other
three tanks larvae were weaned from age 27 to
32 DPH (late weaning, LW). The experiment was terminated at age 41DPH. Seawater was supplied from
the surface outside the pond and ¢ltered through a
60 mm Unik ¢lter. The water exchange was 16% h À 1
from the beginning of the experiment until age
32 DPH, and was then gradually increased to
36% h À 1 at age 41DPH. The salinity was 35 g L À 1
throughout the experiment. Oxygen was measured
daily and saturation never dropped below 90% in
any tank. The temperature was measured daily at
08:00, 14:00 and 20:00 hours.
Sampling
Upon arrival at the experimental site,10 larvae were
sampled for length and mass measurement. N 5 20
larvae from each tank were sampled for length and
mass measurements at age 6, 9, 16, 19, 23, 26, 30, 33,
37 and 41 DPH. All larvae were over-anaesthetized
in metacaine (MS-222) before the measurements
were undertaken. Standard length (SL) was measured under a dissecting microscope equipped with

Semi-extensive production
For the commercial production at Lo¢lab AS,
35 m 3 impermeable plastic bags were submerged
in the pond and ¢lled with ¢ltered pond water.
Newly hatched larvae were stocked at a density of
7 larvae L À 1. Filtered natural zooplankton of chosen size-fractions was constantly supplied to the
bags. In addition, rotifers were supplied to the
bags three times per day to ensure a su⁄cient
supply of food. The larvae were weaned on formulated feed from age 27 to 36 DPH. A growth series

of the commercial semi-extensive production at
Lo¢lab was obtained by weekly samples of 10 larvae from two rearing bags from age 2 to 44 DPH.
Standard length and DM were obtained as described above.
Growth rates of cod larvae reared on rotifers
In order to compare growth rates of cod larvae reared
on copepods vs. larvae reared on rotifers, the growth
rates of eight published experiments where rotifers
were used as the start feed were evaluated by the
use of an STDG model (described in ‘Growth model’)
(Baskerville-Bridges & Kling 2000b; Puvanendran &
Brown 2002; Callan, Jordaan & Kling 2003; Monk,
Puvanendran & Brown 2006; Park, Puvanendran,
Kellett, Parrish & Brown 2006; Fletcher et al. 2007;
Garcia, Parrish & Brown 2008). The protocols in the

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740

1729


Farmed cod larvae fed natural zooplankton K E T Busch et al.

di¡erent experiments varied regarding tank volume,
stocking densities, feeding densities, rotifer feeding
period and temperature (Table 1). However, using the
STDG model a comparison of growth rates across different rearing temperatures is facilitated.

Aquaculture Research, 2010, 41, 1727^1740

the treatment with the highest growth rates from each
of the published results. Modelled DM was converted
to SL by the following relationships given by Finn,
RÖnnestad, van der Meeren and Fyhn (2002):
Size range : 0:031 À 0:175 mg ln SL
¼ ðln DM þ 7:799Þ=3:109

Growth model
The growth rates of the cod larvae in this experiment,
in the semi-extensive production and in published reports were evaluated by the use of the STDG model
(Folkvord 2005). The model was developed based on
growth rates obtained by cod larvae reared on natural
zooplankton provided ad libitum and seems to represent the growth capacity of cod larvae (Folkvord
2005). The input of the model is the DM at start and
the temperatures observed throughout the experiment. For all comparisons, the modelled growth for
northeast Arctic cod was used. The modelled daily instantaneous growth rate [speci¢c growth rate (SGR)]
is given as: SGR 51.0811.79T À 0.074T Â ln DM
À 0.0965aT(ln DM)210.0112T(ln DM)3 (Folkvord
2005). As the growth in DM is exponential, published
graphs presenting DM might be di⁄cult to read. We
therefore chose to present the comparisons of modelled and observed growth as SL. An exception was
the results of Garcia et al. (2008), where the DM sampling series was longer than the SL series; hence, here
we used DM data converted to SL (see description of
conversion below). We chose to use the average SL of

Size range : 0:175 À 1:829 mg ln SL
¼ ðln DM þ 9:657Þ=4:129
Size range : 1:829 À 409 mg ln SL
¼ ðln DM þ 7:932Þ=3:406

Analysis of nutritional composition
At age 27 DPH, larvae from the LW treatment were
sampled for analyses of nutritional composition. As
a limited number of larvae were available from each
tank, pooled samples from the three tanks were used.
The larvae were ¢ltered and rinsed in tap water, dried
with a paper towel under the sieve and frozen in aliquots on dry ice for the di¡erent analyses. The larvae
were transported on dry ice to the lab and stored at
À 80 1C until analyses. Dry weight was determined
gravimetrically after drying the samples at 105 1C
overnight. Protein was calculated as N Â 6.25. Nitrogen was measured using a nitrogen analyzer (Leco
FP-528, St Joseph, MI, USA).Vitamin A and E were determined using normal-phase HPLC after saponi¢cation and extraction of the sample using hexane (Lie,

Table 1 Summary of rearing protocols in studies reporting growth rates of cod larvae

Experiment
Puvanendran and
Brown (2002)
Garcia et al. (2008)

Duration
(days)

Temperature
( 1C)

Tank
Stocking
volume density
(L)

(L À 1)

Rotifer
period
(days)

N per
feeding
(L À 1)

Number of
feedings Rotifer
(day À 1)
enrichment

42

8

30

40

3–42

4000

3

37

11–13

3000

50

0–37

4000

2

O’Brien-MacDonald et
al. (2006)
Park et al. (2006)

65

8

3000

50

0–65

4000

3

43

11

30

50

0–43

4000

4

Callan et al. (2003)

63

10

22

75

0–21

9000

6

Monk et al. (2006)

58

10.5

4500

50

0–45

4000

3–4

Baskerville-Bridges &
Kling 2000a–c
Fletcher et al. (2007)

71

22

55

0–35

11 000

4

80

75

0–54

8000–24 000

8

70

10–11
6.5–11

DHA protein selco
(INVE)
Pavlova and Algamac
s
2000
High Lipid Rotifer
Enrichment
Spray-dried
Crypthecodinium sp.
DHA-selco (INVE)
s
Algamac 2000

Isochrysis and Algamac
s
2000
DHA-selco (INVE)
Culture Selco 3000
(INVE)

Only the protocol of the treatment that resulted in the highest growth rates is hereby reported.

1730

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740


Statistical analyses
The e¡ect of early and late weaning on SL and DM
was tested using an analysis of variance (ANOVA),
where early/late weaning was the ¢xed e¡ect while
larvae nested in the tank and the original tank were
the random e¡ects. Separate analyses were run for SL
and DM, and each sampling day was tested separately. Values are represented as means Æ SE of
means. Level of signi¢cance was set to Po0.05. The
analyses were performed with the statistical software R (2.1.0) (R development core team 2008).

Results
Temperature and feeding
The average temperature from hatching to the age of
25 DPH was 10.2 1C; it then rose to 15.8 1C at day
30 DPH, and to 18.8 1C by the end of the experiment

(Fig.1a). On average, zooplankton was supplied at a density of 1400 prey items L À 1 between age 2 and 14 DPH.

In this period, the zooplankton consisted of a majority
of copepod nauplii, some copepodites and some adult
copepods. The copepod species observed were Pseduocalanus spp., Eurytemora longicornis, Centropages hamatus and some harpacticoids (Fig. 1b). At age 15 DPH, a
major drop in the zooplankton concentration in the
pond occurred, and an average feeding between age
15 and age 32 DPH was 200 prey items L À 1; in this period, copepodites dominated the zooplankton (Fig.1b).

Survival
The average survival from the start to the end (2^
41DPH) of the experiment was 7.2 Æ 3.98% in the
EW tanks and 7.2 Æ 0.42% in the LW tanks. An elevated mortality was observed on the last day of the
experiment.

Growth
The growth rates in all tanks were high from the beginning of the experiment and until age 19 DPH. At
this time, the average SL was 9.77 Æ 0.25 mm and
the DM was 0.54 Æ 0.07 mg (Fig. 2). From age 19 to
26 DPH, the growth rates declined in all tanks before
increasing again from age 26 to 41DPH. In all tanks,
(a)

20

Temperature (°C )

Sandvin & WaagbÖ 1994; N˛ll 1996; Moren, Naess &
Hamre 2004). Ascorbic acid was analysed using
HPLC with amperiometric electrochemical detector,

as described in M×land and WaagbÖ (1998). Thiamine analysis was carried out using HPLC according
to Comite' Europe'en de Normalisation (2003). For
trace element analyses, freeze-dried samples of diet
ingredients and rotifers were wet digested in nitric
acid with 30% hydrogen peroxide using a microwave
technique (Julshamn, Thorlasius & Lea 2000). The
samples were then analysed for Fe, Mn, Cu, Zn and
Se according to Julshamn, Lundebye, Heggstad,
Berntssen and BÖe (2004), and for P according to Liaseth, Julshamn and Espe (2003). Iodine was analysed
according to Julshamn, Dahl and Eckho¡ (2001).
Fatty acid composition of total lipids was analysed according to Lie and Lambertsen (1991), using 19:0 as
the internal standard. Brie£y, the lipid was extracted
with chlorform:methanol 2:1, with the internal standard added and methylated in methanol/NaOH with
BF3. The extract was separated by GLC and detected
by £ame ionization (FID). The columns used were a
head column (Silica 0.53 mm ID) and a CP-sil-88 column of 50 m WCOT, with an inner diameter of
0.32 mm. The fatty acids were quanti¢ed by an integration of the peak areas. The temperature programme of the GC was 50 1C for 1min, an increase
in 25 1C min À 1 to 155 1C, hold for 20 min, increase
of 3 1C min À 1 to 220 1C and hold for 10 min.

Farmed cod larvae fed natural zooplankton K E T Busch et al.

15
10
5
0

(b) 2000
Prey items (L–1)

Aquaculture Research, 2010, 41, 1727^1740

1600
1200
800

Others
Adult copepods
Copepodites
Nauplii

400
0
0

5

10

15 20 25 30 35
Age (days post hatch)

40

45

Figure 1 Average daily temperatures (a) and average
densities of natural zooplankton (b) supplied at each feeding to Gadus morhua L. larvae in a weaning experiment. In
the tanks where larvae received early weaning, zooplankton feeding was stopped after day 25 post hatch.

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740

1731


Farmed cod larvae fed natural zooplankton K E T Busch et al.

the growth rate was similar to the modelled growth
rate until age 19 DPH (Fig. 2a), and declined compared with the modelled growth from age 19 to
26 DPH. When the observed SL was $ 10 mm, the
observed and modelled SL were similar, while at SL
$ 20 mm the average observed SL was 30% lower
than the modelled growth (Table 2).

Aquaculture Research, 2010, 41, 1727^1740

vord 2005) until age 25 DPH. After this time, SL and
DM increased at rates close to the model in bag 1 and
somewhat slower in bag 2 (Fig. 3). The observed and
modelled SL were similar at $ 10 mm, while the
average observed SL was 14% lower compared with
the modelled SL at $ 20 mm (Table 2).
Growth of rotifer-reared larvae

Weaning experiment
The SL of larvae in the EW treatment was signi¢cantly higher than in the LW treatment at age
30 DPH (ANOVA, P 5 0.0015). No signi¢cant di¡erences between treatments were found on other sampling days (Fig. 2a). The DM was signi¢cantly higher
in the EW treatment at age 30 DPH (ANOVA,
P 5 0.0069), whereas no signi¢cant di¡erences between treatments were found on the other sampling

days (Fig. 2b). The e¡ect of tank of origin was not signi¢cant on either DM or SL on any sampling day.

Nutritional composition

Semi-extensive production

30

Standard length (mm)

In the two bags from the commercial production, SL
and DM increased as predicted by the model (Folk(a)

25
20
15
10
5
0

Dry mass (mg)

(b)

The increase in SL of cod larvae reared on rotifers
was lower than the modelled growth in the published
results (Fig. 4). The di¡erence between the observed
growth and modelled growth appears evident from
the start of the experiments (Fig. 4). When the observed SL was $ 10 mm, it was on average 44%
shorter than the modelled SL. In three out of the

eight published experiments, the larvae reached SL
longer than 20 mm. At SL of $ 20 mm, these larvae
were on average 45% shorter compared with the
modelled SL (Table 2).

10

1

Protein constituted 51% and nitrogen 8.1% of the larval DM (Table 3). Of the minerals found in copepodfed larvae, iodine, manganese and selenium were
considerably higher, copper and zinc were similar
and phosphorus was lower than in intensively reared
cod larvae fed rotifers.Vitamin A also appeared lower
in the larvae fed copepods than in those fed rotifers,
although the latter showed large variation. The levels
of thiamine, vitamin C and vitamin E are given for
the copepod-fed larvae, but were not analysed in the
larvae fed rotifers (Table 3). The fatty acid composition in the larvae fed copepods was characterized by
lower levels of monounsaturates and n-6 polyunsaturatted fattyacid (PUFA) and higher levels of n-3
PUFA than in larvae fed rotifers. This trend was also
seen in the levels of ARA, docosahexaenoic acid
(DHA) and eicosapentaenoic acid (EPA) and in the
EPA/ARA ratio (Table 4).

0.1

Discussion
0.01

0

7

14
21
28
35
Age (days post hatch)

42

Figure 2 Growth of Gadus morhua L. larvae measured as
standard length (mm) (a) and dry mass (mg) (b) in common
rearing tanks (ALL) and in early weaning (EW) and late
weaning (LW) treatments. The dashed line in the upper
graph shows the modelled growth at the temperatures observed. Data represent mean values Æ SE of means.

1732

Combining the use of intensive techniques
and natural zooplankton
The use of natural zooplankton as live feed for cod
larvae has traditionally been associated with extensive and semi-extensive rearing methods (van der
Meeren & Naas 1997). Some experiments on cod larvae have been conducted using indoor tanks and natural zooplankton (Otterlei, Nyhammer, Folkvord &

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740


Aquaculture Research, 2010, 41, 1727^1740

Farmed cod larvae fed natural zooplankton K E T Busch et al.

Table 2 A comparison between observed and estimated SL for Gadus morhua L. larvae of SL $ 10 and $ 20 mm
Standard length $ 10 mm

Standard length $ 20 mm

Experiment

Age
(DPH)

Obs.
SL (mm)

Mod.
SL (mm)

Diff.
(%)

Puvanendran & Brown (2002)
Garcia et al. (2008)
O’Brien-MacDonald et al. (2006)
Park et al. (2006)
Callan et al. (2003)
Monk et al. (2006)
Baskerville-Bridges & Kling (2000b)
Fletcher et al. (2007)

Weaning experiment (this study)
Commercial production (this study)

35
30
49
29
36
36
29
33
19
25

10.6
9.8
9.8
9.5
11
10.5
9.5
9
9.8
10.5

14.5
20.7
23
16.25
19.6

20.8
15.13
16.53
9.29
11

26.9
52.9
57.4
41.5
43.9
49.5
37.2
45.6
À 5.5
4.5

Age
(DPH)

Obs.
SL (mm)

Mod.
SL (mm)

Diff. (%)

64

23.5

40.5

42.0

64
62
41
38

20.9
22.5
20.3
19

42.1
40
28.1
22

50.4
43.8
27.8
13.6

(a) 30

Temperature and survival

Standard length (mm)

SL was estimated by a growth model for cod larvae (Folkvord 2005). The cod larvae in the eight published experiments were fed Brachionus (spp.) while the larvae in this study were fed natural zooplankton.
DPH, days post hatch; SL, standard length; Obs.,observed; Mod.,modelled; Di¡., di¡erence.

The temperature rose to high levels at the end of the
experiment, but only at age 41DPH when the average
temperature reached 18.8 1C, high mortalities were
observed. Yin and Blaxter (1987) found that the temperature tolerance (where 50% of the larvae survived
24 h) was 18 1C for newly hatched cod larvae and
15.5 1C at yolk exhaustion for unfed cod larvae. Our
results indicate that the temperature tolerance of fed
cod larvae is higher than 17.8 1C and close to 18.8 1C.
The observed survival was in the lower range of reported survivorships in intensive and extensive systems. In intensive systems, survival of cod larvae of
3^40% have been reported for experiments terminated between age 43 and 72 DPH (BaskervilleBridges & Kling 2000b; Callan et al. 2003; Monk et al.
2006; O’Brien-MacDonald, Brown & Parrish 2006;
Park et al. 2006; Fletcher et al. 2007). Survival of cod
larvae from hatching until metamorphosis in marine
ponds have been reported to be on average 23%
(range 3^42%) (reviewed in Blom 1995). In the present experiment, the relatively high mortality may
have been caused by the low feeding densities from
age 14 DPH and by the high temperatures at the end
of the experiment.

25
20
15
10
5
0

Dry mass (mg)

(b) 100
10
1
0.1
0.01

0

7

14
21
28
35
Age (days post hatch)

42

49

Figure 3 Growth of Gadus morhua L. larvae measured as
standard length (mm) (a) and dry weight (mg) (b) in two
di¡erent rearing bags (1and 2) of the commercial semi-extensive production. The dashed line in the upper graph
shows the modelled growth at the observed temperatures.
Data represent mean values Æ SE of means.

Stefansson 1999). However, the larval densities have

been low ($ 3 larvae L À 1) compared with intensive
rearing systems. To our knowledge, the present experiment is the ¢rst attempt to combine the use of
large-scale intensive rearing methods with natural
zooplankton as live feed for cod.

Weaning
Until age 19 DPH (SL 9.77 mm), the growth rates were
optimal compared with the growth model (Folkvord
2005). Such high growth rates would probably allow
a very early weaning. Baskerville-Bridges and Kling
(2000b) managed to complete weaning on a micro-

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740

1733


Standard length (mm)

Farmed cod larvae fed natural zooplankton K E T Busch et al.

50

30

20

20

10

10

0

0

Standard length (mm)

10

20

30

40

50

60

70

50

0

10

20

30

40

50

60

70

30

40

50

60

70

40

50

60

70

20 30 40 50 60
Age (days post hatch)

70

50
O’Brien−MacDonald et al.2006
8 (°C)

40

Park et al. 2006
11 (°C)

40

30

30

20

20

10

10

0

0
0

Standard length (mm)

Garcia et al. 2008
11−13 (°C)

40

30

0

50

10

20

30

40

50

60

70

0
50

Callan et al. 2003
10 (°C)

40

30

20

20

10

10

10

20

Monk et al. 2006
10.5 (°C)

40

30

0

0
0

Standard lenght (mm)

50

Puvanendran & Brown 2002
8 (°C)

40

Aquaculture Research, 2010, 41, 1727^1740

10

20

30

40

50

60

70

50

0

10

20

30

50
Baskerville−Bridges & Kling 2000b
10−11 (°C)

40

Fletcher et al. 2007
6.5−11 (°C)

40

30

30

20

20

10

10

0

0
0

10

20 30 40 50 60
Age (days post hatch)

70

0

10

Figure 4 Growth in standard length (SL) (mm) from eight published results where Gadus morhua L. larvae were fed Brachionus spp. The points represent the average SL for the treatment with the highest growth rates from the reported experiments. In the graph from the experiment reported by (Garcia et al. 2008), dry mass has been converted to SL. The dashed
lines are the modelled growth at the temperatures observed.

particulate diet at a larval SL of 8.5 mm. In Norwegian commercial hatcheries, cod larvae are now commonly weaned directly from rotifers onto formulated
feed. Weaning may start when the larvae are SL of
$ 7.5 mm and 25 days old and ¢nish when the larvae
reach SL of $ 8.0 mm (T. A. Hangstad, pers. comm.).
With the high growth rates observed in our experiment, weaning could have been ¢nished well before
the age of 16 DPH, something which could substantially shorten the time depending on live feed using
current rearing protocols. Poor growth was observed
between age 19 and 26 DPH in both the EW and LW
treatment. These low growth rates may have been

caused by stress from the transfer of larvae to the ex-

1734

perimental tanks at age 19 DPH. However, the low
feed density in this period is a probable cause of the
poor growth rates. The prey concentrations needed
to support optimal growth rates are likely to vary
with larval densities, larval sizes and prey type. Puvanendran and Brown (1999) reported higher growth
rates for cod larvae that received 4000 rotifers L À 1
than for larvae that received 2000 rotifers L À 1.
Further, prey concentrations of 1000 rotifers L À 1 did
not support survival beyond age 32 DPH (Puvanendran & Brown 1999). Rajkumar and Kumaraguru vasagam (2006) fed Acartia clausi (Giesbrecht) at
densites of $ 1700 L À 1 with no signs of limited
growth rates in Lates calcarifer (Bloch).

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740


Aquaculture Research, 2010, 41, 1727^1740

Farmed cod larvae fed natural zooplankton K E T Busch et al.

Table 3 Protein, nitrogen, selected minerals and vitamins
on dry weight from Gadus morhua L. larvae aged 27 DPH
reared on natural zooplankton

Dry matter (% of wet wt)
Protein (% of dry wt)

Nitrogen (% of dry wt)
mg kg À 1 dry wt
Iodine
Manganese
Copper
Zinc
Selenium
Phosphorus
mg kg À 1 dry wt
Vitamin A (sum retinol)
Thiamine (B1)
Vitamin C
Vitamin E
(a-tokopherol)

Experimental
larvae

Larvae fed
rotifers

15.3
51
8.1

12.5 Æ 0.2
na
na

29

13
7
142
5.1
16 497

1.76 Æ 0.51
4.7 Æ 0.4
6.1 Æ 1.2
146 Æ 6
1.1 Æ 0.3
19 470 Æ 520

5.2
9.9
594
76

12.6 Æ 4.7
na
na
na

Each nutrient was analysed from a pooled sample of larvae from
three rearing tanks. Data from larvae reared on industrially produced Brachionus plicatilis (Mˇller) (26 DPH) are given for comparison (mean Æ SD, n 5 4; Hamre et al. 2008b).
na, not analysed; wt, weight.

Table 4 Fatty acid composition (% of total fatty acids) from
Gadus morhua L. larvae aged 27 DPH reared on natural
zooplankton

16:0
16:1 (n-7)
18:0
18:1 (n-9)
18:1 (n-7)
18:2 (n-6)
20:4 (n-6) ARA
20:4 (n-3)
20:5 (n-3) EPA
22:5 (n-3)
22:6 (n-3) DHA
Total saturates
Total monounsaturated
Total (n-3) PUFA
Total (n-6) PUFA
DHA/EPA
EPA/ARA

Experimental
larvae

Larvae fed
rotifers

16.5
1.4
6.2
4.9
2.5

0.8
3.3
0.8
11.3
1.6
37.6
24.8
10.3
53.6
5.1
3.33
3.42

14.4
3.0
8.6
9.6
2.7
3.4
2.8
1.3
7.2
2.7
25.4
25.1
19.9
38.2
7.1
3.5
2.5

Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ

0.1
0.2
0.1
0.1
0.0
0.1
0.0
0.0
0.1
0.0
0.7

0.2
0.4
0.6
0.1
0.1
0.1

Fatty acid composition was analysed from a pooled sample of larvae from three rearing tanks. Data from larvae reared on industrially produced Brachionus plicatilis (Mˇller) (26 DPH) are given
for comparison (mean Æ SD, n 5 4; Hamre et al. 2008b). The total sums include fatty acids not reported in this table.
PUFA, polyunsaturatted fattyacid; EPA, eicosapentaenoic acid;
DHA, docosahexaenoic acid.

In a pond experiment, prey densities in the range
of 1.5^32.3 prey L À 1 supported optimal growth of
cod larvae with an initial stocking density of
$ 30 larvae m À 3 (Busch, Folkvord, OtterÔ, Hutchinson & SvÔsand 2009). The present results indicate
that prey densities below 200 L À 1 do not support optimal growth of cod larvae when natural zooplankton is used as feed in intensive rearing systems.
After a period of slow growth, the larvae again
started to grow fast from age 26 and 30 DPH for the
EW and LW treatment respectively (Fig. 2). Larvae in
the EW tanks were signi¢cantly longer and heavier at
age 30 DPH, indicating that it was an advantage to
receive formulated feed at an early age in this experiment. The limited larval growth in the LW treatment
may have been caused by the low densities of prey
organisms. The weaning period was short compared
with what has been reported previously (Fletcher
et al. 2007; Baskerville-Bridges & Kling 2000b);
but with the high growth rates observed at the
end of the weaning periods, we consider weaning to
have been successful. To our knowledge, this is

the earliest weaning reported for cod larvae reared
on natural zooplankton. Such an early weaning
reduces the quantity of natural zooplankton
needed and would make it possible to increase
the number of larvae produced. Additionally, an
early weaning reduces the risks associated with the
use of live feed, e.g. sudden drops in zooplankton
availability.

Natural zooplankton vs. rotifers
The growth rates of the intensively reared experimental larvae and the commercial semi-extensive
production were both optimal until the larvae
reached $ 10 mm (Figs 2 and 3). As the larvae in
these two systems were produced by two entirely
di¡erent protocols, but with the same food, we
suggest that the high growth rates observed in
ponds and semi-extensive systems might be caused
rather by the natural zooplankton than by low
densities of larvae or large volumes. We evaluated
reported growth rates of cod larvae reared on rotifers
(Fig. 4 and Table 2) and found that despite a wide
range of protocols regarding light regimes, feed densities, rotifer enrichment and temperature (Table 1),
all of the reported growth rates were sub-optimal
and surprisingly similar across the studies. The modelled maximum SGR occurs at $ 10 mm (Folkvord
2005), being 19% at 10 1C. At this length, rotifer-

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740

1735

Farmed cod larvae fed natural zooplankton K E T Busch et al.

reared larvae were 30^50% shorter than the modelled SL (Table 2). Further, for the three experiments
where cod larvae were reared beyond an SL of
20 mm, the main di¡erences between modelled and
observed growth were established already at SL
10 mm. This indicates that rotifers are sub-optimal
as feed for cod larvae from the beginning of exogenous feeding.
Some experimenters have attempted to compare
growth rates of marine larvae fed rotifers vs. natural
zooplankton directly, all indicating the superiority of
copepods. In an experiment on southern £ounder
Paralichthys lethostigma (Jordan & Gilbert), larvae
that received a mixture of Acartia tonsa (Dana) and
rotifers were signi¢cantly heavier and had a signi¢cantly higher survival at 14 DPH than larvae that received rotifers only (Wilcox, Tracy & Marcus 2006).
Rajkumar and Kumaraguru vasagam (2006) followed seabass larvae until the age of 21DPH and
found that larvae fed Acartia clausi were signi¢cantly
heavier and had a signi¢cantly higher survival than
larvae fed rotifers or Artemia nauplii. These ¢ndings
support the idea that rotifers are sub-optimal as feed
for marine ¢sh larvae. The reason why rotifers do not
support optimal growth rates in cod larvae may be
numerous. The nutritional composition of natural
zooplankton and rotifers are di¡erent and may be
the main cause of growth di¡erences in cod larvae
(further discussed in ‘Nutrition’). However, cod
larvae, when o¡ered a choice, select prey items
of increasing sizes as they grow (Busch et al. 2009).

And while the size of rotifers is well suited for
newly hatched cod larvae, they are far below the
preferred size of larger cod larvae. To compensate
for the small size of rotifers, cod larvae will have to
ingest a vast number of prey items, and at some stage
this is likely to be a limiting factor for the growth of
cod larvae.

Nutrition
As the larvae in this study were raised on natural
zooplankton, we may assume that the nutritional
composition of the diet was ideal for cod larvae (Sargent, McEvoy, Estevez, Bell, Bell, Henderson & Tocher
1999a), and hence these results may serve as a reference for future studies of live-feed enrichment.While
the nutritional composition of natural zooplankton
has been investigated previously (Van der Meeren, Olsen, Hamre & Fyhn 2008), a full-scale analysis of nutritional composition of cod larvae reared on natural

1736

Aquaculture Research, 2010, 41, 1727^1740

zooplankton has not been performed before. However, these analyses were performed on single pooled
samples of larvae from three tanks and the results
should therefore be treated with care.
The di¡erence in dry matter found between larvae
fed rotifers and copepods may be related to the di¡erence in larval size, as the dry matter increases as the
larvae grow (unpubl. results). Another cause for differences in dry matter may be di¡erence in nutritional status (Hamre, N×ss, Espe, Holm & Lie 2001).
We have no reference for the protein and nitrogen
content of cod larvae fed rotifers, but the data on larvae fed copepods may be used as a reference for
further studies. The protein content in dry matter is
inversely correlated to lipid content and may be used

as an indication of the energy status of the larvae.
Little information exists on the dietary requirements of micronutrients of marine ¢sh larvae, but
the levels of most of the trace elements are higher in
copepods than in rotifers (Hamre, Srivastava, RÖnnestad, Mangor-Jensen & Stoss 2008a). The data in
the present study, when compared with the results
of Hamre, Mollan, S×le and Erstad (2008b), show
that this is re£ected in the levels of iodine, manganese
and selenium in the larval bodies. Based on the idea
of using natural plankton as a reference for larval
nutrient requirements, iodine, selenium and manganese are thus de¢cient, while copper and phosphorus
appear to be su⁄cient in the control rotifers used in
Hamre et al. (2008b). The high zinc level found in rotifer-fed larvae by Hamre et al. (2008b), may have
been caused by high levels of zinc in the inlet water
originating from the zinc anodes of the pumps, and
hence we cannot conclude that zinc is su⁄cient in
standard rotifers. Hamre et al. (2008b) found that enriching rotifers with iodine and selenium to copepod
levels increased the survival of cod larvae at 26 DPH
by 32%, but the study was not conclusive with regard
to which one of the elements or if the combination of
them was the cause of the improved larval performance. The concentration of iodine in the larvae did
not increase signi¢cantly and the concentration of
selenium was only 3.5 mg kg À 1 dry weight. The difference in retention of iodine in the present study
compared with Hamre et al. (2008b), may indicate
that iodine exists in di¡erent forms in copepods
and enriched rotifers, and that these forms have different bio-availabilities. Studies on manganese have
not been performed with cod, but Tien, Satoh, Haga,
Fushimi and Kotani (2008) fed Artemia enriched
with zinc and manganese to red seabream larvae,
and found that manganese increased larval growth,

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740


Aquaculture Research, 2010, 41, 1727^1740

while both manganese and zinc reduced skeletal
deformities in this species. The data thus indicate
that some of the trace elements are de¢cient in
rotifers and Artemia, and in intensively reared cod
larvae.
We have no data on vitamin C, thiamine and vitamin E levels in larvae fed rotifers (Table 3), but we
know that vitamin E is most often lower in copepods
than in commercially produced rotifers (unpubl. results), supported by the relatively low level of vitamin
E found in the larvae in the present study.Vitamin E is
often used as an antioxidant in enrichment diets for
rotifers; in addition, vitamin E is added as a vitamin
supplement in large doses. It is not known if the high
vitamin E levels in commercial diets have a bene¢cial
e¡ect in terms of vitamin supplement.Vitamin C and
thiamine are also supplemented in commercial enrichment diets for rotifers (Hamre et al. 2008a; unpublished results) and the high levels of these
vitamins in larvae fed copepods in the present study,
justify this supplementation. Rotifers fed yeast-based
diets contain ample amounts of the other B vitamins
(Hamre et al. 2008a), and these vitamins were therefore not analysed. It is not known if algae-based rotifer diets, which have become more common in the
commercial market, have similar levels of B vitamins
as the yeast. Van der Meeren et al. (2008) found that
the levels of vitamin A in copepods were below the
detection limit of their analyses. The vitamin A found
in cod larvae in the present experiment is probably

caused by a conversion from astaxanthin to vitamin
A (Moren et al. 2004). Astaxanthin is present in high
levels in copepods (van der Meeren et al. 2008). Nevertheless, vitamin A was lower in larvae fed copepods,
than in larvae fed rotifers and the concentration of
vitamin A in the control rotifers in Hamre et al.
(2008b), was only half the vitamin A requirement in
¢sh (0.7 mg kg À 1, NRC 1993). This indicates that cod
larvae have a low requirement for vitamin A, or that
the active metabolites of vitamin A are synthesized
directly from the carotenoids.
The composition of fatty acids is of major concern
in start feeding of marine ¢sh (McEvoy, Naess, Bell &
Lie 1998; Este¤vez, McEvoy, Bell & Sargent 1999; Sargent et al.1999a; Sargent, Bell, McEvoy,Tocher & Estevez 1999b; Hamre 2006). In particular, the amount of
the essential fatty acids DHA, EPA and AA and their
relative proportions may be crucial for successful larval development (Sargent et al. 1999a, b). Not surprisingly, the fatty acid pro¢le in this experiment
resembles that found in wild-caught cod eggs and larvae caught around the Lofoten islands, with high le-

Farmed cod larvae fed natural zooplankton K E T Busch et al.

vels of DHA, EPA and AA (KlungsÖyr, Tilseth,
Wilhelmsen, Falk-Petersen & Sargent 1989) and is
characterized by higher levels of DHA, EPA and n-3
fatty acids than in larvae fed rotifers.

Long-term e¡ect of di¡erent live feeds
The long-term e¡ects of di¡erences in start feed during the larval stage have not been thoroughly investigated. Cod farmers report that juveniles originating
from extensive rearing systems grow faster than juveniles originating from intensive systems. The difference in weight at slaughter at an equivalent age
might be as much as 1000 g (K.-P. Myklebust, pers.
comm.). Imsland et al. (2006) found signi¢cant di¡erences in growth rates of cod juveniles originating
from two di¡erent rearing regimes, one using natural

zooplankton and the other rotifers. The experiment
started when the cod juveniles were 8.5 Æ 0.3 g and
was terminated 3 months later. At termination, the
zooplankton group was 26^36% heavier than the rotifer group (Imsland et al. 2006).We observed similar
di¡erences between rotifer and zooplankton-reared
juveniles in a pilot trial. The juveniles were
28.8 Æ 11.6 g at the start of the trial. The wet weight
33 days later in 9 1C water were 64.4 Æ 19.1g for the
zooplankton group and 50.6 Æ 14.2 g for the rotifer
group (Students t-test, Po0.01) (unpubl. results).
These results and observations indicate that zooplankton-reared cod will continue to outgrow rotifer-reared ¢sh throughout life. This might be due to
qualitative di¡erences between the two groups,
where zooplankton-reared juveniles are better developed for providing high growth rates than rotifer
reared cod. An alternative explanation is that rotifer
reared larvae fail to utilize the very high growth potential at $ 20 DPH, and that no (or very little) compensatory growth occurs after this. It follows from
the last hypothesis that growth rates must, to a large
extent, be age dependent rather than size dependent.
However, more experimental data are needed to investigate the long-term e¡ects of improved early
growth rates.

Conclusions and future research
In this experiment, we have shown that natural zooplankton may be used as a start feed in intensive
rearing systems. The main advantage of using natural zooplankton as a start feed is the high initial
growth rates of cod larvae, which must be closely

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740

1737

Farmed cod larvae fed natural zooplankton K E T Busch et al.

linked to the nutritional composition of the copepods.
Our results indicate that the di¡erences in nutritional status between larvae fed natural plankton
and those fed rotifers are large in some respects and
may well explain the di¡erences in growth between
these larvae.
Early growth may signi¢cantly increase long-term
growth of farmed cod, but the e¡ect of early growth
on later performance needs to be further investigated.We suggest that the focus should be placed on
optimizing the growth of cod larvae. This could be
achieved by the use of natural zooplankton, by
further improvement of rotifer enrichment and rearing methods or by development of other live prey organisms, such as the copepod Acartia spp. (StÖttrup,
Richardson, Kirkegaard & Pihl1986; Rajkumar & Kumaraguru vasagam 2006; Wilcox et al. 2006). The
use of natural zooplankton in intensive systems
may have great potential and optimizing larval
stocking densities and prey densities are likely to increase the larval production from such systems considerably.

Acknowledgements
We thank the sta¡ at Lo¢lab AS for letting us use their
experimental facilities, for the supply of live feeds and
for technical support. Thanks are due to Fride Tonning for great help during the experiment. We acknowledge the National Breeding Programme,
TromsÖ for the supply of cod larvae free of charge.
This experiment was supported by the Norwegian
Research Council.

References
Baskerville-Bridges B. & Kling L.J. (2000a) Development and
evaluation of microparticulate diets for early weaning of

Atlantic cod Gadus morhua larvae. Aquaculture Nutrition
6,171^182.
Baskerville-Bridges B. & Kling L.J. (2000b) Early weaning of
Atlantic cod (Gadus morhua) larvae onto a microparticulate diet. Aquaculture 189, 109^117.
Baskerville-Bridges B. & Kling L.J. (2000c) Larval culture of
Atlantic cod (Gadus morhua) at high stocking densities.
Aquaculture 181, 61^69.
Bj˛rnsson B., Steinarsson A. & Arnason T. (2007) Growth
model for Atlantic cod (Gadus morhua): e¡ects of temperature and body weight on growth rate. Aquaculture 271,
216^226.
Blom G. (1995) Production of juvenile Atlantic cod (Gadus morhua L.) in marine ponds with emphasis on the recruitment

1738

Aquaculture Research, 2010, 41, 1727^1740

process. Dr. scient. thesis. Departement of Fisheries and
Marine Biology, Bergen, Norway.
Brown J.A., Minko¡ G. & Puvanendran V. (2003) Larviculture of Atlantic cod (Gadus morhua): progress, protocols
and problems. Aquaculture 227, 357^372.
Busch K.E.T., Folkvord A., OtterÔ H., HutchinsonW.F. & SvÔsand T. (2009) E¡ects of female spawning experience and
larval size on feeding and growth of cod larvae Gadus
morhua L. reared in mesocosms. Marine Biology Research
5, 286^296.
Callan C., Jordaan A. & Kling L.J. (2003) Reducing Artemia
use in the culture of Atlantic cod (Gadus morhua). Aquaculture 219, 585^595.
Comite¤ Europe¤en de Normalisation,TC 275 EN 14122 (2003)
Foodstu¡s^ Determination of Vitamin B1 by HPLC. European Committee for Standardization, Brussels, Belgium.
Directorate of ¢sheries (2007) Key ¢gures from the Norwegian aquaculture industry. Available at: www.¢skeridir.no
(accessed 23 November 2009).

Engelsen R., Asche F., Skjennum F. & Ado¡ G. (2004) New
species in aquaculture: some basic economic aspects. In:
Culture of Cold-Water Marine Fish (ed. by E. Moksness, E.
KjÖrsvik & Y. Olsen), pp. 487^515. Blackwell publishing,
Oxford, UK.
Este¤vez A., McEvoy L.A., Bell J.G. & Sargent J.R. (1999)
Growth, survival, lipid composition and pigmentation of
turbot (Scopthalmus maximus) larvae fed live-prey enriched in Arachidonic and Eicosapentaenoic acids. Aquaculture 180, 321^343.
Finn R.N., RÖnnestad I., van der Meeren T. & Fyhn H.J.
(2002) Fuel and metabolic scaling during the early life
stages of Atlantic cod Gadus morhua. Marine Ecology-Progress Series 243, 217^234.
Fletcher R.C., Roy W., Davie A.,Taylor J., Robertson D. & Migaud H. (2007) Evaluation of new microparticulate diets
for early weaning of Atlantic cod (Gadus morhua): implications on larval performances and tank hygiene. Aquaculture 263, 35^51.
Folkvord A. (2005) Comparison of size-at-age of larval
Atlantic cod (Gadus morhua) from di¡erent populations
based on size- and temperature-dependent growth models. Canadian Journal of Fisheries and Aquatic Sciences 62,
1037^1052.
Garcia A.S., Parrish C.C. & Brown J.A. (2008) A comparison
among di¡erently enriched rotifers (Brachionus plicatilis)
and their e¡ect on Atlantic cod (Gadus morhua) larvae
early growth, survival and lipid composition. Aquaculture
Nutrition 14, 14^30.
Hamre K. (2006) Nutrition in cod (Gadus morhua) larvae
and juveniles. ICES Journal of Marine Science 63, 267^274.
Hamre K., N×ss T., Espe M., Holm J.C. & Lie Ò. (2001) A formulated diet for Atlantic halibut (Hippoglossus hippoglossus, L.) larvae. Aquaculture Nutrition 7, 123^132.
Hamre K., Srivastava A., RÖnnestad I., Mangor-Jensen A. &
Stoss J. (2008a) Several micronutrients in the rotifer Brachionus plicatilis may be limiting for growth, survival and

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740

Aquaculture Research, 2010, 41, 1727^1740

normal development of cod larvae. Aquaculture Nutrition
14, 51^60.
Hamre K., Mollan T.A., S×le Ò. & Erstad B. (2008b) Rotifers
enriched with iodine and selenium increase survival in
Atlantic cod (Gadus morhua) larvae. Aquaculture 284,
190^195.
Imsland A.K., Foss A., Koedijk R., Folkvord A., Stefansson
S.O. & Jonassen T.M. (2006) Short- and long-term di¡erences in growth, feed conversion e⁄ciency and deformities in juvenile Atlantic cod (Gadus morhua) startfed on
rotifers or zooplankton. Aquaculture Research 37, 1015^
1027.
Julshamn K.,Thorlasius A. & Lea P. (2000) Determination of
arsenic in seafood by electrothermal atomic absorption
spectrometry after microwave digestion: NMKL collaborative study. Journal of AOAC International 83, 1423^
1428.
Julshamn K., Dahl L. & Eckho¡ K. (2001) Determination of
iodine in seafood by inductively coupled plasma/mass
spectrometry. Journal of AOAC International 84, 1976^
1982.
Julshamn K., Lundebye A.K., Heggstad K., Berntssen M.H.G.
& BÖe B. (2004) Norwegian monitoring programme on inorganic and organic contaminants in ¢sh caught in the
Barents Sea, Norwegian Sea and North Sea, 1994^2001.
Food Additives and Contaminants 21, 365^376.
KlungsÖyr J., Tilseth S., Wilhelmsen S., Falk-Petersen S. &
Sargent J.R. (1989) Fatty acid composition as an indicator
of food intake in cod larvae Gadus morhua from Lofoten,
Northern Norway. Marine Biology 102,183^188.

Liaseth B., Julshamn K. & Espe M. (2003) Chemical composition and theoretical nutritional evaluation of the produced
fractions from enzymatic hydrolysis of salmon frames
with Protamex. Process Biochemistry 38,1747^1759.
Lie Ò. & Lambertsen G. (1991) Fatty acid composition of glycerophospholipids in seven tissues of cod (Gadus morhua)
determined by a combined HPLC/GC method. Journal of
Chromatography 565, 119^129.
Lie Ò., Sandvin A. & WaagbÖ R. (1994) Transport of a-tocopherol in Atlantic salmon (Salmo salar) during vitellogenesis. Fish Physiology and Biochemistry 13, 241^247.
M×land A. & WaagbÖ R. (1998) Examination of the qualitative ability of some cold water marine teleosts to synthesise ascorbic acid. Comparative Biochemistry and
PhysiologyA 121, 249^255.
McEvoy L.A., Naess T., Bell J.G. & Lie Ò. (1998) Lipid and fatty
acid composition of normal and malpigmented Atlantic
halibut (Hippoglossus hippoglossus) fed enriched Artemia:
a comparison with fry fed wild copepods. Aquaculture
163, 237^250.
Monk J., Puvanendran V. & Brown J.A. (2006) Do di¡erent
light regimes a¡ect the foraging behaviour, growth and
survival of larval cod (Gadus morhua L.)? Aquaculture
257, 287^293.
Moren M., Naess T. & Hamre K. (2004) Conversion of betacarotene, canthaxanthin and astaxanthin to vitamin A

Farmed cod larvae fed natural zooplankton K E T Busch et al.

in Atlantic halibut (Hippoglossus hippoglossus L.) juveniles.
Fish Physiology and Biochemistry 27,71^80.
N˛ll G.N. (1996) High performance liquid chromatography
analysis of retinal and retinal isomers. Journal of Chromatography 721, 247^259.
NRC (1993) Nutrient Requirements of Fish. National Research
Council, National Academy Press, Washington, DC,
USA.
O’Brien-MacDonald K., Brown J.A. & Parrish C.C. (2006)

Growth, behaviour, and digestive enzyme activity in larval Atlantic cod (Gadus morhua) in relation to rotifer lipid.
ICES Journal of Marine Science 63, 275^284.
Otterlei E., Nyhammer G., Folkvord A. & Stefansson S.O.
(1999) Temperature- and size-dependent growth of larval
and early juvenile Atlantic cod (Gadus morhua): a comparative study of Norwegian coastal cod and northeast
Arctic cod. Canadian Journal of Fisheries and Aquatic
Sciences 56, 2099^2111.
Park H.G., Puvanendran V., Kellett A., Parrish C.C. & Brown
J.A. (2006) E¡ect of enriched rotifers on growth, survival,
and composition of larval Atlantic cod (Gadus morhua).
ICES Journal of Marine Science 63, 285^295.
Puvanendran V. & Brown J.A. (1999) Foraging, growth and
survival of Atlantic cod larvae reared in di¡erent prey
concentrations. Aquaculture 175,77^92.
Puvanendran V. & Brown J.A. (2002) Foraging, growth
and survival of Atlantic cod larvae reared in di¡erent
light intensities and photoperiods. Aquaculture 214,
131^151.
PuvanendranV., Burt A.L. & Brown J.A. (2006) Can Atlantic
cod (Gadus morhua) larvae be weaned faster onto dry feed
at higher temperatures? Aquaculture 255, 334^340.
Rajkumar M. & Kumaraguru vasagam K.P. (2006) Suitability of the copepod, Acartia clausi as a live feed for Seabass
larvae (Lates calcarifer Bloch): compared to traditional
live-food organisms with special emphasis on the nutritional value. Aquaculture 261, 649^658.
R development core team (2008) A Language and Environment for Statistical Computing. R Foundation for Statistical
Computing,Vienna, Austria.
Rosenlund G. & Halldorsson O. (2007) Cod juvenile production: research and commercial developments. Aquaculture
268, 188^194.
Sargent J., McEvoy L., Estevez A., Bell G., Bell M., Henderson
J. & Tocher D. (1999a) Lipid nutrition of marine ¢sh during

early development: current status and future directions.
Aquaculture 197, 217^229.
Sargent J., Bell G., McEvoy L.,Tocher D. & Estevez A. (1999b)
Recent developments in the essential fatty acid nutrition
of ¢sh. Aquaculture 177, 191^199.
Solemdal P., Dahl E., Danielssen D.S. & Moksness E. (1984)
The cod hatchery in FlÖdevigen ^ background and
realities. In: The Propagation of Cod Gadus morhua L
(ed. by E. Dahl, D.S. Danielssen, E. Moksness & P. Solemdal), pp. 17^48. Institute of Marine Research, Arendal,
Norway.

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740

1739


Farmed cod larvae fed natural zooplankton K E T Busch et al.

StÖttrup J.G. (2000) The elusive copepods: their production
and suitability in marine aquaculture. Aquaculture Research 31,703^711.
StÖttrup J.G., Richardson K., Kirkegaard E. & Pihl N.J. (1986)
The cultivation of Acartia tonsa Dana for use as a live food
source for marine ¢sh larvae. Aquaculture 52, 87^96.
SvÔsand T., OtterÔ H. & Taranger G.L. (2004) The status and
perspectives for the species. In: The Propagation of Cod Gadus morhua L (ed. by E. Dahl, D.S. Danielssen, E. Moksness
& P. Solemdal), pp. 433^474. Institute of Marine Research,
Arendal, Norway.
Tien N.V., Satoh S., Haga Y., Fushimi H. & Kotani T. (2008)
E¡ect of Zn and Mn supplementation to Artemia on

growth and vertebral deformity in red sea bream (Pagus
major) larvae. XIII International Symposium on Fish Nutrition and Feeding. June 1^5, Florianopolis, Brazil
van der Meeren T. & Naas K.E. (1997) Development of rearing
techniques using large enclosed ecosystems in the mass
production of marine ¢sh fry. Reviews in ¢sheries science
5, 367^390.

1740

Aquaculture Research, 2010, 41, 1727^1740

van der Meeren T., Olsen R.E., Hamre K. & Fyhn H.J. (2008)
Biochemical composition of copepods for evaluation of
feed quality in production of juvenile marine ¢sh. Aquaculture 274, 375^397.
Watson J.T., Sales S., Cumming G., Fitzsimmons S.D.,Walden
J., Arthur G., Saravanan S. & McEvoy L.A. (2006) Comparison of performance of two size groups of farmed cod (Gadus morhua L.) juveniles following transfer to sea cages.
ICES Journal of Marine Science 63, 340^345.
Wilcox J.A., Tracy P.L. & Marcus N.H. (2006) Improving live
feeds: e¡ect of a mixed diet of Copepod Nauplii (Acartia
tonsa) and Rotifers on the survival and growth of ¢rstfeeding larvae of the Southern Flounder, Paralichthys
lethostigma. Journal of the World Aquaculture Society 37,
113^120.
Yin M.C. & Blaxter J.H.S. (1987) Temperature, salinity tolerance, and buoyancy during early development and starvation of clyde and North-Sea Herring, Cod, and
Flounder Larvae. Journal of Experimental Marine Biology
and Ecology 107, 279^290.

r 2009 The Authors
Aquaculture Research r 2009 Blackwell Publishing Ltd, Aquaculture Research, 41, 1727^1740

Aquaculture Research, 2010, 41, 1741^1747

doi:10.1111/j.1365-2109.2009.02460.x

Effects of temperature on fertilized eggs and larvae of
tawny puffer Takifugu flavidus
Yong-Hai Shi1,2, Gen-Yu Zhang1,Ya-Zhu Zhu1, Jian-Zhong Liu1 & Wei-Ling Zang2
1

Shanghai Fisheries Research Institute, Shanghai, China
Shanghai Ocean University, Shanghai, China

2

Correspondence: Y-H Shi, Shanghai Fisheries Research Institute, Shanghai 200433, China. E-mail:

Abstract

Introduction

Tawny pu¡er Takifugu £avidus is a species found
in China considered to have potential for aquaculture. Experiments were conducted to determine the
optimal temperature for its incubation and larval
culture. Fertilized eggs collected from cultured
broodstocks that were induced to ovulate with a
[D-Ala6 -Pro9-Net]-luteinizing
hormone-releasing
hormone analogue were inseminated. The e¡ect of
temperature (19, 20, 23, 26 and 29 1C) on the hatch
rate, incubation period, viability of 24 h post-hatch

larvae and total mortality rate was assessed. The effect of temperature (20, 23, 26 and 29 1C) on the
growth and survival of larvae from 3 to 19 days after
hatching (dah) was also assessed. The results
showed that the optimal temperature for successful
development of fertilized eggs ranged from 23 to
26 1C, and the highest hatch rate, the optimal viability of 24 h post-hatch larvae and the lowest total mortality rate were all predicted using quadratic
equations. The relationship between temperature
and the incubation period of tawny pu¡er eggs was
determined using the e¡ective degree-day model.
The temperature at developmental zero (t0) was
11.34 1C, and the sum of e¡ective degree-days (k)
was 52.356. The survival rate of tawny pu¡er larvae
at 20 1C was signi¢cantly lower than among 23, 26
and 29 1C, whereas the survival rate was not signi¢cantly di¡erent from that at 23, 26 and 29 1C. The larval growth rate increased rapidly as the temperature
increased, showing a linear relationship in the range
of temperatures investigated. The optimal temperature for larval culture ranged from 23 to 29 1C.

Pu¡er ¢sh are widely distributed, with approximately
100 di¡erent species in the world. Some species are
important ¢shery resources and appear to be promising for aquaculture (Yang & Chen 2005). Tawny
pu¡er Takifugu £avidus (Tetraodontiforms,Tetraodontidae) is mainly sourced from the East China Sea,Yellow Sea and Bohai Bay. It is a coastal temperate
bottom ¢sh and is not migratory (Yang, Zhang &
Kuang 1991). Tawny pu¡er has a potential for aquaculture in China because of its commercial importance and high market value. However, because of
environmental degradation, over ¢shing and other
factors, tawny pu¡er may become endangered (Shi,
Zhang, Zhu, Yan, Liu & Zhu 2009). This is occurring
despite arti¢cial breeding to compensate for its declining number and to continue meeting consumer
demands (Shi et al. 2009). Relatively little is known
about the environmental requirements to support
the early life stages of tawny pu¡er, and no research

conducted on the e¡ects of temperature on the eggs
and larvae of this species.
Temperature is one of the most critical external
factors of ontogeny in the early life stages of ¢sh
(Kamler 2002). The e¡ects of temperature on ontogeny are determined by the rate of enzymatic reactions (Blaxter 1969). There are two ways in which
temperature a¡ects ontogeny. First, temperature, if
within a viable range, strongly a¡ects the rate of ontogeny (Yang & Chen 2005). A temperature beyond
this range is lethal for the species (Brett 1979). Second, temperature a¡ects the hatch rate (Hart &
Purser 1995; Gracia-Lo¤pez, Kiewek-Mart|¤ nez & Maldonado-Garc|¤ a 2004; Cook, Guthriel, Rust & Plesha
2005; Uehara & Mitani 2009), incubation period
(Hamel, Mangan, East, Lapointe & Laurendeau 1997;

Keywords: temperature, eggs, incubation, larvae,
tawny pu¡er,Takifugu £avidus

r 2010 The Authors
Aquaculture Research r 2010 Blackwell Publishing Ltd

1741


Temperature e¡ects on tawny pu¡er eggs and larvae Y-H Shi et al.

Kamler 2002; Yang & Chen 2005; Petereit, Haslob,
Kraus & Clemmesen 2008), the size of the newly
hatched larvae (Scoppettone, Buetter & Rissler 1993;
Bermudes & Ritar 1999), larval yolk absorption and
utilization (Fukuhara 1990; Hart & Purser 1995;
Hardy & Litvak 2004), larval feeding behaviour
(Johnston & Mathias 1994), larval survival (Gadomski & Cadell 1991; Bidwell & Howell 2001; Hansen &

Falk-Petersen 2001; Berlinsky,Taylor, Howell, Bradley
& Smith 2004) and larval growth (Hansen & Falk-Petersen 2001; Hardy & Litvak 2004; Fielder, Bardsley,
Allan & Pankhurst 2005; Petereit et al. 2008).
The objective of this study is to investigate the
e¡ects of temperature on the eggs and larvae of
tawny pu¡er to determine the optimal temperature
for incubation and larval culture of the species. The
results of this study will be useful in increasing the
production of this species through incubation and
larval culture.

Materials and methods
Fertilized eggs and larvae collection
Fertilized eggs of tawny pu¡er were obtained from
cultured broodstocks maintained in 20 m3 concrete
tanks with a salinity and temperature of 13.5 g L À 1
and 21.0 Æ 1.0 1C respectively. Twenty-eight 4-yearold females (with body weights of 0.50^0.75 kg)
were injected with a [D-Ala6 -Pro9-Net]-luteinizing
hormone-releasing hormone analogue at a dose
of 35 mg kg À 1 body weight. After the 30th hour following the hormonal treatment, abdominal palpation was performed every 1^2 h to determine the
completion of ¢nal oocyte maturation. The majority
of females ovulated about 35^40 h after hormonal
treatment. The eggs were manually stripped and arti¢cially fertilized (Shi et al. 2009). At 24 h post fertilization, when fertilized eggs were at the stage of late
gastrula, dead and physically damaged eggs were removed using a wide-mouth pipette. Only developing
fertilized eggs from a single female were placed into
units of Experiment 1.
The fertilized eggs were incubated in 700 L conical
tanks, which were ¢lled with ¢ltered brackish water
(13.5 g L À 1) and provided with constant aeration at
21.7 Æ 0.5 1C. The density of the eggs was 1.5^

2.0 Â 105 m À 3. Tank water was kept static, and
about 70% of the content of the tank was drained
and replaced with ¢ltered brackish water daily (Shi
et al. 2009). Most of the hatching began about 100 h
and ¢nished about 120 h after fertilization. Larvae

1742

Aquaculture Research, 2010, 41, 1741–1747

were kept in the incubation tanks for 3 days after
hatching (dah). Afterwards, the larvae were used in
Experiment 2.

Experiment 1: E¡ect of temperature on the
hatch rate, incubation period, viability of 24 h
post-hatch larvae and total mortality rate
The temperatures were set to 17, 20, 23, 26 and 29 1C
for the treatments. Each treatment was replicated
thrice, which was conducted in water baths equipped
with thermoregulators and immersion heaters. Incubation units were comprised of 300 mL glass beakers
¢lled with 250 mL brackish water. Sixty embryos
were placed in each beaker. Eggs were incubated statically in the beakers under natural light and photoperiod. All temperature gradients were adjusted
gradually at a rate of 1 1C h À 1 (Fielder et al. 2005).
Fifty per cent of the water in each beaker was replaced daily with fresh brackish water. For all replicates, mortalities were removed and counted each
day until all the larvae had hatched. During the incubation period, salinity was 13.5 g L À 1, dissolved oxygen was 5.0^6.5 mg L À 1 and pH was 8.0^8.5.
The e¡ects of temperature on eggs were evaluated
using the following criteria (Yang & Chen 2005): (1)
total hatch rate, which is the percentage of stocked
embryos that hatched regardless of viability; (2) incubation period, which is the time interval between the

period of egg activation and the period when 50% of
the fertilized eggs have hatched (Kamler 2002; Gracia-Lo¤pez et al. 2004). To ensure accurate measurement of time interval, eggs were monitored every
few hours to see how many had already hatched,
and the times when the observations were made
were recorded once hatching began; (3) viability of
24 h post-hatch larvae, which is the percentage of
live larvae 24 h posthatch. To determine the viability,
all newly hatched larvae were moved from the incubation beaker and placed in a 1L glass beaker containing the same temperature of water for 24 h. At
24 h, deformed and moribund larvae were counted
and (4) total mortality rate, which represents the
combined percentage of dead eggs, abnormal fry and
dead and moribund larvae.

Experiment 2: E¡ect of temperature on the
growth and survival of larvae from 3 to 19 dah
The temperatures of the experimental treatments
were 20, 23, 26 and 29 1C. About 2400 larvae (3 dah;

r 2010 The Authors
Aquaculture Research r 2010 Blackwell Publishing Ltd, Aquaculture Research, 41, 1741^1747


Aquaculture Research, 2010, 41, 1741^1747

Temperature e¡ects on tawny pu¡er eggs and larvae Y-H Shi et al.

3.26 Æ 0.02 mm total length, mean Æ SEM, n 5 20)
were haphazardly harvested from the 700 L tank
and divided into ¢fteen10 L plastic buckets ¢lled with
brackish water (13.5 g L À 1; 22.0 1C). The initial stocking densities were 20 larvae L À 1. Three haphazardly

selected replicate buckets were used per temperature
treatment. All temperature gradients were adjusted
as described above. Larvae in all buckets were fed
with enriched rotifers Brachionus plicatilis until they
reached a total length of about 5.0^6.0 mm and a
density of about 5^7 rotifers mL À 1. When the larval
total length reached about 5 mm, Artemia nauplii larvae were also o¡ered, and its density increased from1
to10 mL À 1 at the end of the experiment. Larvae were
fed at 09:00 and 15:00 hours each day. Dead larvae,
faeces and other debris were siphoned out, and the
number of dead larvae was registered every day. Forty
percent of the water in each bucket was replaced daily with fresh brackish water. Each bucket was provided with gentle constant aeration. The buckets
were placed in a water bath under natural light and
photoperiod (14 L:10 D). The experiment lasted for 16
days.Water quality variables were as follows: salinity
13.5^14.5 g L À 1, dissolved oxygen 5.0^6.5 mg L À 1
and pH 8.0^8.5.
A haphazardly selected sample of 10 larvae was
collected from each bucket and placed in a beaker at
approximately 14:00 hours every 4 days. Larvae
anaesthetized with MS-222 were observed under a
stereoscopic microscope equipped with an eyepiece
micrometer (the nearest 0.01mm) and their total
lengths were measured. Larvae were not returned to
their original buckets to avoid handling-related mortality, and the number of larvae sampled through out
the experiment was reduced from the initial number
to calculate the survival rate.

Daily water measurement
Temperature was measured four times daily using

the YSI Model (Yellow Springs, OH, USA): 30-10 FT
(0.1 1C). Salinity, pH and dissolved oxygen were measured daily using theYSI model:30-10 FT salinity meter (0.1g L À 1),YSI Model: No. pH 100 (0.1pH unit) and
YSI Model: 58 dissolved oxygen meter (0.1mg L À 1)
respectively.

Statistical analyses
All data were presented as mean Æ SEM. The e¡ects
of temperature on the indices were analysed using

the one-way ANOVA after arcsin transformation of data
in case of percentages, followed by the Student^Newman^Keuls test (SNK). All analyses were performed
with a signi¢cance level of Po0.05. The e¡ective degree-day model was used to describe the relationship
between the incubation period and temperature in
Experiment 1 (Hamel et al. 1997; Kamler 2002; Yang
& Chen 2005): y 5 k/(T À t0), where y is the incubation period in days, T is the temperature in degrees
Celsius, k is the sum of degree-days and t0 is the temperature at developmental zero. The growth rate in
Experiment 2 was calculated using the formula (Hart
& Purser 1995; Gracia-Lo¤pez et al. 2004): growth rate
(mm day À 1) 5 (¢nal length (mm) À initial length
(mm))/time (days). Statistical analyses were conducted using the SPSS 13.0 software.

Results
Experiment 1: E¡ect of temperature on the
hatch rate, incubation period, viability of 24 h
post-hatch larvae and total mortality rate
Hatch rate at 26 1C was signi¢cantly higher
(Po0.05) than that at 17 1C, whereas hatch rates at
20, 23 and 29 1C were not signi¢cantly di¡erent
(P40.05) from either 26 or 17 1C (Table 1). The viability of 24 h post-hatch larvae at 23, 26 and 29 1C was
signi¢cantly higher (Po0.05) than that at 17 and

20 1C, and that at 20 1C was signi¢cantly higher
(Po0.05) than that at 17 1C. However, there was no
signi¢cant di¡erence (P40.05) among 23, 26 and
29 1C (Table 1). The total mortality rate at 17 1C was
signi¢cantly higher (Po0.05) than that at a higher
temperature, whereas there was no signi¢cant di¡erence (P40.05) among 20, 23, 26 and 29 1C (Table 1).
The relationships between temperature and hatch
rate, temperature and viability of 24 h post-hatch larvae, and temperature and total mortality rate were
all modelled bya quadratic equation (r240.90) (Fig.1).
From these equations, the highest hatch rate, the
optimal viability of 24 h post-hatch larvae and the
lowest total mortality rate were all predicted to occur
at 23^26 1C.
Incubation periods decreased with an increase in
the water temperature (17^29 1C). There were signi¢cant di¡erences (Po0.05) in the incubation period at
all temperature treatments. The relationship between
temperature and incubation period was determined
using the e¡ective degree-day model (r2 50.9872;
F-test, Po0.001) (Fig. 2).

r 2010 The Authors
Aquaculture Research r 2010 Blackwell Publishing Ltd, Aquaculture Research, 41, 1741^1747

1743


Aquaculture Research, 2010, 41, 1741–1747

Temperature e¡ects on tawny pu¡er eggs and larvae Y-H Shi et al.

Table 1 Hatch rate, incubation period, viability of 24 h post-hatch larvae and total mortality rate of tawny pu¡er eggs at
di¡erent temperatures (mean Æ SEM, n 5 3)
Treatment
( 1C)

Temperature
( 1C)

17
20
23
26
29

17.47
20.07
23.12
26.07
28.97

Æ
Æ
Æ
Æ
Æ

0.15
0.05
0.04
0.04

0.09

Hatch rate
(%)
73.33
83.33
83.89
87.22
83.33

Æ
Æ
Æ
Æ
Æ

Viability of 24 h post-hatch
larvae (%)
Ã

1.67b,
0.96ab
3.89ab
3.38a
3.47ab

74.81
83.90
97.39
97.40

96.64

Æ
Æ
Æ
Æ
Æ

4.51c
4.32b
0.55a
0.74a
0.72a

Total mortality
rate (%)
45.00
30.00
18.33
15.00
19.44

Æ
Æ
Æ
Æ
Æ

4.41a
4.41b

3.47b
3.85b
3.64b

Incubation period
(days)
8.49
5.78
4.74
3.40
3.01

Æ
Æ
Æ
Æ
Æ

0.08a
0.01b
0.01c
0.00d
0.17e

100

10

80

80

8

60

60

40

40

20

20

0
14

17

20 23 26 29
Temperature (°C)

0
32

Figure 1 Regression analysis de¢ned the relationship between temperature and hatch rate, viability of 24 h posthatch larvae and total mortality rate (~ hatch rate:
y 5 À 52.872111.152x À 0.2227x2, r2 50.9152; 4 viability of 24 h post-hatch larvae: y 5 À 116.05116.241x À
0.3075x2, r2 50.9709; & total mortality rate: y 5

303.04 À 22.271x10.4308x2, r2 50.9992; data presented
as mean, n 5 3).

Experiment 2: E¡ect of temperature on the
growth and survival of larvae from 3 to 19 dah
The survival rate at 20 1C was signi¢cantly lower
(Po0.05) than that at 23, 26 and 29 1C, whereas there
was no signi¢cant di¡erence (P40.05) among 23, 26
and 29 1C (Table 2).
Temperature strongly a¡ected larval growth. Signi¢cant e¡ects were observed at 11dah. The total
lengths at 26 and 29 1C were signi¢cantly longer
(Po0.05) than those at 20 and 23 1C, and the total
length at 23 1C was signi¢cantly longer (Po0.05)
than that at 20 1C. There was no signi¢cant di¡erence
(P40.05) between 26 and 29 1C. From 15 dah, larval
growth was signi¢cantly faster (Po0.05) at higher
temperatures (Fig. 3). The growth rate increased with

1744

Incubation period (day)

100

Total mortality rate (%)

Hatch rate (%) and Viability
of 24h post-hatch larvae (%)

ÃMean values within a column followed by di¡erent letters were signi¢cantly (Po0.05) di¡erent.

6

4

2

0
14

17

20
23
26
Temperature (°C)

29

32

Figure 2 Regression analysis de¢ned the relationship
between temperature and incubation period. The ¢tted
line is based on the e¡ective degree-day model
[y 5 52.356/(T À 11.340), r2 50.9872; ~ data presented as
mean, n 5 3].

an increase in the temperature from 20 to 29 1C
(Table 2). The linear equation between growth rate
and temperature was established (r2 50.9855) (Fig.4).

Discussion
The optimal temperature of incubation coincided
with the natural spawning temperature of this species (Mihelakakis & Yoshimatsu 1998). In the present
study, based on quadratic equations, the highest
hatch rate, the optimal viability of 24 h post-hatch
larvae and the lowest total mortality rate were all
predicted to occur at 23^26 1C. Therefore, we conclude that the optimal temperature for tawny pu¡er
egg incubation was from 23 to 26 1C. This optimal
temperature range was similar to the water temperature observed from May to June in Bohai Bay, China

r 2010 The Authors
Aquaculture Research r 2010 Blackwell Publishing Ltd, Aquaculture Research, 41, 1741^1747


Aquaculture Research, 2010, 41, 1741^1747

Temperature e¡ects on tawny pu¡er eggs and larvae Y-H Shi et al.

Table 2 Survival rate and growth rate of tawny pu¡er larvae at di¡erent temperatures from 3 to19 dah (mean Æ SEM, n 5 3)
Treatment ( 1C)

Temperature ( 1C)

20
23
26
29

20.12

23.00
26.00
28.95

Æ
Æ
Æ
Æ

Growth rate (mm day À 1)

Survival rate (%)

0.02
0.01
0.01
0.03

24.86
47.93
37.48
40.00

Æ
Æ
Æ
Æ

Ã

4.22b,
3.08a
1.48a
0.62a

0.102
0.292
0.564
0.700

Æ
Æ
Æ
Æ

0.007d
0.012c
0.025b
0.022a

ÃMean values within a column followed by di¡erent letters were signi¢cantly (Po0.05) di¡erent.
dah, days after hatching.

16

29°C

12

a

26°C
b

10
b

8
6

a
ab
b

a
a

b

c

4
2
0

3

7

c

23°C

c
b
d

11
15
Days after hatching

d

20°C

Growth rate (mm d–1)

Total length (mm)

0.8

a

14

0.6

0.4

0.2

19

Figure 3 Total length (mean Æ SEM, n 5 3) of tawny
pu¡er larvae grown from 3 to 19 days after hatching
(dah) at di¡erent temperatures. Di¡erent letters within
the same age indicate signi¢cant di¡erences (Po0.05)
among treatments.

(Yang et al. 1991). In addition, the optimal temperature range for this species was higher than that for
the same genera of ¢sh (obscure pu¡er Takifugu obscurus) eggs (from 19 to 23 1C) (Yang & Chen 2005).
The tolerance limits of eggs to incubation temperature are commonly determined in relation to their
mortality (Bermudes & Ritar 1999). In this study, the
highest total mortality rate and the lowest viability of
24 h post-hatch larvae occurred at 17 1C. This suggests that tawny pu¡er eggs incubated at or below
this temperature are not suitable for this species. This
result is similar to that for the obscure pu¡er T. obscurus (Yang & Chen 2005).
The incubation period of tawny pu¡er decreased
(from 8.49 to 3.01 days) with an increase in the incubation temperature (from 17 to 29 1C). Many studies
show that temperature strongly a¡ects the rate of ontogeny, and the ontogenetic rate of ¢sh is slow at low
temperatures and fast at higher temperatures within
a viable range (Hart & Purser 1995; Hamel et al. 1997;
Mihelakakis & Yoshimatsu 1998; Bermudes & Ritar

0
17

20

23

26
Temperature (°C)

29

32

Figure 4 Regression analysis de¢ned the relationship
between temperature and growth rate (y 5 À 1.30681
0.0702x, r2 50.9855; ~ data presented as mean, n 5 3).

1999; Hansen & Falk-Petersen 2001; Kamler 2002;
Gracia-Lo¤pez et al. 2004; Yang & Chen 2005; Uehara
& Mitani 2009).
Many models can be used to describe the relationship between water temperature and the incubation
period of ¢sh eggs, such as the power-law model, the
quadratic equation, the exponential equation and the
e¡ective degree-day model (Kamler 2002; Yang &
Chen 2005). More than one model can be used for
the same species of ¢sh eggs (Yang & Chen 2005).
However, some studies that compared these models
reveal that no theoretical basis has been found from
the use of the power-law model, the quadratic equation and the exponential equation, while the derived
parameters (k and t0) of the e¡ective degree-day model have signi¢cant biological meaning (Hamel et al.
1997; Kamler 2002;Yang & Chen 2005). Based on this
¢nding, many scholars prefer to use the e¡ective degree-day model to describe the relationship between
temperature and incubation period (Hamel et al.1997;
Kamler 2002; Yang & Chen 2005; Uehara & Mitani

r 2010 The Authors

Aquaculture Research r 2010 Blackwell Publishing Ltd, Aquaculture Research, 41, 1741^1747

1745


Temperature e¡ects on tawny pu¡er eggs and larvae Y-H Shi et al.

2009), although this model has often been criticized
because of its lack of accuracy at extreme temperatures (Wagner, Wu, Sharpe, School¢eld & Coulson
1984; Highley, Pedigo & Ostlie1986). Therefore, in this
study, although the relationship between temperature and incubation period of tawny pu¡er can be determined using the power-law model, the quadratic
equation, the exponential equation and the e¡ective
degree-day model, we chose to use only the e¡ective
degree-day model because of its biological meaning,
ease of computing and the good ¢t to our data. In addition, this model has been shown to be useful for interspeci¢c and intraspeci¢c comparisons from
di¡erent latitudes or altitudes (Kamler 2002).
Based on the e¡ective degree-day model for tawny
pu¡er eggs, the temperature at developmental zero
(t0) was 11.34 1C, and the sum of the e¡ective degreedays (k) was 52.356. Yang and Chen (2005) showed
that t0 at 7.6033 1C and k at 78.905 were the e¡ective
degree-days of obscure pu¡er (T. obscurus) eggs. Kamler (2002) showed that in coldwater species, a low t0
was accompanied by a high value of k, whereas in
warm-water species, the opposite was true. Based on
these results, we also concluded that the optimal
temperature for tawny pu¡er eggs was higher than
that for obscure pu¡er. This also agreed with the initial conclusions in this study.
Temperature a¡ected larval survival probabilities.
The survival rate of tawny pu¡er larvae at 20 1C was
signi¢cantly lower than that at other temperatures.
Reduced survival at suboptimal rearing temperatures had been reported for a number of cultured ¢sh,

including California halibut Paralichthys californicus
(Gadomski & Cadell 1991), witch £ounder Glyptocephalus cynoglossus (Bidwell & Howell 2001) and black
sea bass Centropristis striata (Berlinsky et al. 2004).
However, the survival rate was not signi¢cantly different among 23, 26 and 29 1C. That temperature has
no in£uence on the larval survival in optimal rearing
temperature range had also been reported for some
species, such as the leopard grouper (Mycteroperca rosacea) (Gracia-Lo¤pez et al. 2004) and the Australian
snapper (Pagrus auratus) (Fielder et al. 2005). In addition, the optimal temperature range of early larval
culture was very similar to the natural spawning
conditions (Fielder et al. 2005). In the present study,
the optimal temperature for tawny pu¡er larvae was
similar to the natural spawning conditions in Bohai
Bay, China (Yang et al. 1991).
Larvae grew in direct relation to temperature. The
larval growth rate increased rapidly with an increase
in the temperature, showing a linear relationship in

1746

Aquaculture Research, 2010, 41, 1741–1747

the range of temperatures investigated in this study.
This agreed with many studies, such as those on the
greenback £ounder (Rhombosolea tapirina Gˇnther,
1982) (Hart & Purser 1995), spotted wol⁄sh (Anarhichas minor Olafsen) (Hansen & Falk-Petersen 2001),
leopard grouper (Mycteroperca rosacea) (Gracia-Lo¤pez
et al. 2004) and Australian snapper (Pagrus auratus)
(Fielder et al. 2005).
In summary, the optimal conditions of temperature for incubation and larval culture of tawny pu¡er
ranged from 23 to 26 1C and from 23 to 29 1C respectively. The e¡ective degree-day model for tawny puffer eggs yielded satisfactory results, with the

temperature at developmental zero (t0) at 11.34 1C
and the sum of e¡ective degree-days (k) reaching
52.356. The larval growth rate increased rapidly with
an increase in the temperature, showing a linear relationship from 20 to 29 1C.

Acknowledgments
We would like to thank Zhiwen Zhang, Jiabo Xu,
Yongde Xie and Xiaodong Zhu for their advice and
assistance throughout the course of this study. We
are also grateful to the two anonymous reviewers for
their helpful suggestions and English corrections,
which considerably improved the manuscript. This
study was funded by the Fisheries O⁄ce of Shanghai.

References
Berlinsky D.L.,Taylor J.C., Howell R.A., Bradley T.M. & Smith
T.I.J. (2004) The e¡ects of temperature and salinity on
early life stages of black sea bass Centropristis striata. Journal of theWorld Aquaculture Society 35, 335^344.
Bermudes M. & Ritar A.J. (1999) E¡ects of temperature on
the embryonic development of the striped trumpeter (Latris lineate Bloch and Schneider, 1801). Aquaculture 176,
245^255.
Bidwell D.A. & Howell W.H. (2001) The e¡ect of temperature
on ¢rst feeding, growth, and survival of larval witch
£ounder Glyptocephalus cynoglossus. Journal of the World
Aquaculture Society 32, 373^384.
Blaxter J.H.S. (1969) Development: eggs and larvae. In: Fish
Physiology,Vol. 3 (ed. by W.S. Hoar & D.J. Randall), pp.177^
252. Academic Press, NewYork, NY, USA.
Brett J.R. (1979) Environmental factors and growth. In: Fish
Physiology, Vol. 8 (ed. by W.S. Hoar, D.J. Randall & J.R.

Brett), pp. 599^675. Academic Press, NewYork, NY, USA.
Cook M.A., Guthriel K.M., Rust M.B. & Plesha P.D. (2005) Effects of salinity and temperature during incubation on
hatching and development of lingcod Ophiodon elongatus
Girard, embryos. Aquaculture Research 36, 1298^1303.

r 2010 The Authors
Aquaculture Research r 2010 Blackwell Publishing Ltd, Aquaculture Research, 41, 1741^1747


Aquaculture Research, 2010, 41, 1741^1747

Temperature e¡ects on tawny pu¡er eggs and larvae Y-H Shi et al.

Fielder D.S., Bardsley W.J., Allan D.L. & Pankhurst P.M.
(2005) The e¡ects of salinity and temperature on growth
and survival of Australian snapper, Pagrus auratus larvae.
Aquaculture 250, 201^214.
Fukuhara O. (1990) E¡ects of temperature on yolk utilization, initial growth, and behaviour of unfed marine ¢shlarvae. Marine Biology 106, 169^174.
Gadomski D.M. & Cadell S.M. (1991) E¡ects of temperature
on early-life stages of California halibut Paralichthys californicus. Fishery Bulletin 89, 567^576.
Gracia-Lo¤pez V., Kiewek-Mart|¤ nez M. & Maldonado-Garc|¤ a
M. (2004) E¡ects of temperature and salinity on arti¢cially reproduced eggs and larvae of the leopard grouper
Mycteroperca rosacea. Aquaculture 237, 485^498.
Hamel P., Mangan P., East P., Lapointe M. & Laurendeau P.
(1997) Comparison of di¡erent models to predict the in
situ embryonic development rate of ¢sh with special reference to white sucker (Catostomus commersoni). Canadian
Journal of Fisheries and Aquatic Sciences 54, 190^197.
Hansen T.K. & Falk-Petersen I.B. (2001) E¡ects of egg
disinfection and incubation temperature on early life
stages of spotted wol¢sh. Aquaculture International 9,

333^344.
Hardy R.S. & Litvak M.K. (2004) E¡ects of temperature on
the early development, growth, and survival of shortnose
sturgeon, Acipenser brevirostrum, and Atlantic sturgeon,
Acipenser oxyrhynchus, yolk-sac larvae. Environmental
Biology of Fishes 70, 145^154.
Hart P.R. & Purser G.J. (1995) E¡ects of salinity and temperature on eggs and yolk sac larvae of the greenback £ounder
(Rhombosolea tapirina Gˇnther, 1982). Aquaculture 136,
221^230.
Highley L.G., Pedigo L.P. & Ostlie K.R. (1986) DEGDAY: a
program for calculating degree-days, and assumptions
behind the degree-day approach. Environmental Entomology 15, 999^1016.
Johnston T.A. & Mathias J.A. (1994) The e¡ects of temperature on feeding in zooplanktivorous walleye, Stizostedion

vitreum, larvae. Environmental Biology of Fishes 40,
189^198.
Kamler E. (2002) Ontogeny of yolk-feeding ¢sh: an ecological perspective. Reviews in Fish Biology and Fisheries 12,
79^103.
Mihelakakis A. & Yoshimatsu T. (1998) E¡ects of salinity and
temperature on incubation period, hatching rate and
morphogenesis of the red sea bream. Aquaculture International 6, 171^177.
Petereit C., Haslob H., Kraus G. & Clemmesen C. (2008) The
in£uence of temperature on the development of Baltic Sea
sprat (Sprattus sprattus) eggs and yolk sac larvae. Marine
Biology 154, 295^306.
Scoppettone G.G., Buetter M.E. & Rissler P.H. (1993) E¡ect of
four £uctuating temperature regimes on cui-ui, Chasmistes cujus, survival from egg fertilization to swim-up,
and size of larvae produced. Environmental Biology of
Fishes 38, 373^378.
Shi Y.H., Zhang G.Y., Zhu Y.Z., Yan Y.L., Liu J.Z. & Zhu J.M.

(2009) Study on arti¢cial breeding techniques of Takifugu
£avidus in Estuary area of Hangzhou Bay. In: Modern Fisheries and Sustainable Development,Vol. 1 (ed. by X.L. Zhang
& J. Liu), pp. 165^172. Ocean Press, Beijing, China.
Uehara S. & Mitani T. (2009) E¡ect of temperature on the development of eggs and the daily pattern of spawning of
round herring Etrumeus teres. Fisheries Science 75, 159^
165.
Wagner T.L.,Wu H.-I., Sharpe P.J.H., School¢eld R.M. & Coulson R.N. (1984) Modeling insect development rates: a literature review and application of a biophysical model.
Annals of the Entomological Society of America 77, 208^
225.
Yang Z. & Chen Y.F. (2005) E¡ect of temperature on incubation period and hatching success of obscure pu¡erTakifugu obscurus (Abe) eggs. Aquaculture 246,173^179.
Yang Z.F., Zhang H.Q. & KuangY.H. (1991) Studies on biology
of fugu £avidus in Bohai bay. Marine Sciences Bulletin 10,
44^47 (in Chinese with English abstract).

r 2010 The Authors
Aquaculture Research r 2010 Blackwell Publishing Ltd, Aquaculture Research, 41, 1741^1747

1747


Aquaculture Research, 2010, 41, 1748^1758

doi:10.1111/j.1365-2109.2009.02475.x

Rearing cuttings of the soft coral Sarcophyton glaucum
(Octocorallia, Alcyonacea): towards mass production
in a closed seawater system
Ido Sella & Yehuda Benayahu
Department of Zoology, George S.Wise Faculty of Life Sciences,Tel-Aviv University,Tel-Aviv, Israel
Correspondence: I Sella, Department of Zoology, George S.Wise Faculty of Life Sciences,Tel-Aviv University,Tel-Aviv 69978, Israel. E-mail:

Abstract
The octcoral Sarcophyton glaucum has a wide IndoPaci¢c distribution and is known for its diverse content of natural products. The aim of the current study
was to establish a protocol for rearing miniature cuttings of S. glaucum in a closed seawater system. In order to determine the optimal conditions for rearing,
the survival, average dry weight, percentage of organic weight and development of the cuttings were
monitored under di¡erent temperature, light, salinity
and feeding regimes. At 26 1C, the highest dry weight
was obtained, and at 20 1C, the highest percentage of
organic weight. The dry weight of the cuttings increased with the light intensity, while under 35^
130 mE m À 2 s À 1, survival was high. Salinity did not
a¡ect any of the colonies’ features. Feeding intervals
of 7 and 30 days yielded a better result than of 2 days.
A comparison of the colonies derived from the closed
system with the colonies reared in a £ow-through
system, those reared in the sea and with ¢eld-collected colonies revealed the importance of environmental conditions in determining the features of the
colonies. The study emphasizes the advantages of a
closed seawater system in controlling the conditions
needed for rearing cuttings of S. glaucum for targeted
farming.

Keywords: soft corals, closed seawater system,
Sarcophyton glaucum, coral farming, Red-Sea

Introduction
A large variety of reef invertebrates, including soft
corals (Octocorallia), have long been used as a source

1748

for diverse natural products with pharmaceutical or
cosmetic value (e.g., Blunt, Copp, Munro, Northcote &
Prinsep 2005; Slattery, Gochfeld & Kamel 2005; Sipkema, Osinga, Schatton, Mendola,Tramper & Wij¡els
2005), as well as for the reef-aquarium trade (Wabnitz, Taylor, Grenn & Razak 2003). The increased demand for these organisms has led to their massive
harvesting (Castanaro & Lasker 2003) and has raised
the need for e⁄cient farming methodologies (Ellis &
Ellis 2002; Mendola 2003).
Coral propagation has been commonly used for the
production of daughter colonies, rather than harvesting naturally grown ones (e.g., Soong & Chen 2003;
Fox, Mous, Pet, Muljadi & Caldwell 2005). This practice has been based on the ability of corals to reproduce asexually (e.g, Delbeek 2001; Ellis & Ellis 2002).
Captive-grown colonies, or ¢eld-collected ones, are
used for the production of cuttings or fragments. The
latter are usually glued or attached by various means
to bases or stubs, until natural attachment is
achieved (Delbeek 2001).
Over the last decade, coral propagation in closed
arti¢cial seawater systems has been commonly used,
making the process more simple and cost e¡ective
(Borneman & Lowrie 2001; fkeeping.
com). The advantage of these systems lies in their
ability to control abiotic and biotic parameters
(Wheeler 1996; Borneman & Lowrie 2001; Abramovitch-Gottlib, Katoshevski & Vago 2002). Such systems have been proposed as a source by which to
obtain corals and other invertebrates with commercial value (e.g., Wabnitz et al. 2003; Sipkema et al.
2005), as well as for restoration and conservation
purposes (Becker & Mueller 1999; Petersen, Laterveer, Van Bergen, Hatta, Hebbinghaus, Janse, Jones,

r 2010 The Authors
Aquaculture Research r 2010 Blackwell Publishing Ltd

Aquaculture Research, 2010, 41, 1748^1758

Rearing cuttings of the soft coral S. glaucum I Sella & Y Benayahu

Richter, Ziegler,Visser & Schuhmacher 2006; Okuzawa, Maliao, Quinitio, Buen-ursua, Lebata, Gallardo,
Garcia & Primavera 2008).
Species of the soft coral genus Sarcophyton (Octocorallia) have a wide Indo-Paci¢c distribution and are
known for the diverse content of their natural products (e.g., Look, Fenical, Matsumoto & Clardy 1986;
Tanaka, Yoshida & Benayahu 2005; Nguyen, Tran,
Phan, Chau, Eun & Young 2008). S. glaucum (Quoy &
Gaimard, 1833) is the most common species of the
genus, also on the northern Red Sea reefs, where it
was found to be a dioecious broadcaster with the onset of reproduction at the age of 6^10 years (Benayahu & Loya 1986). S. glaucum is zooxanthellate and
extremely rich in natural products, which have been
studied extensively (e.g., Fridkovsky, Rudi, Benayahu,
Kashman & Schleyer 1996; Badria, Guirguis, Perovic,
Ste¡en, Muller & Schroder 1998; Tanaka et al. 2005).
The aim of the current study was to establish the
scienti¢c ground for a protocol suitable for rearing
miniature cuttings of S. glaucum in a closed seawater
system. Our objectives were to study the e¡ect of temperature, light intensity, salinity and feeding regimes
on the survival, average dry weight, percentage of organic weight and development of the cuttings. In addition, colonies reared in a closed seawater system were
compared with those reared in a £ow-through seawater system in the sea and with ¢eld-collected ones.

Materials and methods
In a preliminary work, we examined attachment of
the S. glaucum cuttings to stubs made of ceramic and
plastic materials, using a cyanoacrylate adhesive
(3M) and rubber bands. We also examined the minimal size of cuttings yielding the highest survivorship
(10^400 mm2) (Sella 2007). Consequently, in the current study, we decided to use cuttings measuring

36 mm2 (6 Â 6 mm) with an average dry weight of
0.0077 Æ 0.0024 g and 8.864 Æ 2.461 average percentage of organic weight (n 5 30 random cuttings
from colonies used in the experiments; see‘Collection
of colonies, preparation of cuttings and experimental
setup’). The cuttings were glued using a cyanoacrylate adhesive to cylindrical ceramic (Fig. 1a) stubs
that we produced ourselves (clay WBB Fuchfs GmbH
& Co R2502, burnt at 1200 1C for 12 h) at a cost of
$ 4 US cents (without labour). The stubs measured
5 cm in length and 1cm in diameter, with a hexagonal indentation (ca. 3 mm deep and 7^8 mm wide) at
one end, into which the cuttings were attached.

Experimental closed seawater system
We constructed a closed seawater system at Tel Aviv
University (TAU). The system comprised three 1m3
PVC tanks, each ¢lled with arti¢cial seawater (Red
Sea saltr) and 200 kg of live rocks obtained from Eilat (Gulf of Aqaba, northern Red Sea). Each tank was
provided with 40 Trochos dentatus snails for algal
grazing. Every 10 days, their faeces were removed
from the bottom of the tanks, together with ca. 5% of
the water volume, which was then replaced. A set of
24 glass aquaria (30 L each) was connected to the
tanks by 26-mm-diameter pipes. The aquaria were illuminated by neon bulbs (22 W, 400^800 nm, JEBO
daylight, 12 L:12 D, 35 mE m À 2 s À 1 on the colony surface). The water circulated from the aquaria to the
tanks, which were equipped with protein skimmers
(JEBO 3500), a chiller (JEBO 2000, Zhongshan Zhenhua Aquarium Accessories, Zhongshan, China) and
a 5000 L per hour circulation pump (JEBO). The
water temperature in each aquarium was controlled
using an individual heater (60 W). Salinity, temperature and pH were monitored daily, and the nutrient
levels weekly (nitrite o0.05 ppm, nitrate o10 ppm,
ammonia 0 ppm and 8.1^8.3 pH) (see Delbeek 2001).

Collection of colonies, preparation of cuttings
and experimental set-up
Colonies of S. glaucum with a polypary diameter of 5^
7 cm (the polyp-bearing part of the colony) were collected by SCUBA from three reef sites in Eilat (5^7 m,
August 2004^January 2007). For each experiment,
we collected 10^15 colonies, growing at least 5 m
apart in order to avoid the use of asexually produced
colonies (Benayahu & Loya 1986). In order to prevent
possible bias in the study, only colonies that did not
contain gonads were used for formation of the cuttings (see Okubo, Motokawa & Omori 2007). At the
Interuniversity Institute for Marine Science (IUI),
the colonies were placed in a £ow-through seawater
system for 3 days and all epibiotic organisms were removed. The colonies were £own to TAU in plastic containers ¢lled with seawater and placed in thermally
isolated boxes. They were then maintained in the experimental system for a 1-month quarantine period,
under conditions resembling the average ambient Eilat ones: salinity 40 ppt, pH 8.2 and 26 1C (IUI Data
Base, />For preparation of the cuttings, the colonies of
S. glaucum were handled with latex gloves. The perimeter of the polypary was removed using a scalpel

r 2010 The Authors
Aquaculture Research r 2010 Blackwell Publishing Ltd, Aquaculture Research, 41, 1748^1758

1749


Rearing cuttings of the soft coral S. glaucum I Sella & Y Benayahu

(see section ‘General observations’ ahead), and the
remaining part was then kept for future regeneration
(Sella 2007). The perimeter of the polypary was thoroughly rinsed with arti¢cial seawater for removal of

mucus and cellular debris, and 36 mm2 (6 Â 6 mm)
cuttings were prepared using autoclaved scissors,
scalpels and forceps and paper blotted. An adhesive
was applied to the cuttings’ lower epidermal surface,
and they were mounted on the stubs and kept in the
air for 30 s to dry. Notably, the cuttings retrieved for
the di¡erent treatments of each experiment were
randomly derived from the same parent colonies and
therefore represented similar genotypes. The stubs
were then inserted into test tube stands (Fig. 1a), and
kept in arti¢cial seawater for 20 min in order to wash
away any adhesive residue before introduction into
the aquaria. During the following week, if mortality
did not exceed15%, dead cuttings were replaced with
new ones and only then the 60-day experiments
were started; otherwise, all cuttings were discarded
and a new experiment was set up as it occurred once.
In all the experiments, each treatment comprised
three aquaria and a control, which was maintained
under Eilat average environmental conditions. Unless stated otherwise, the aquaria were supplied
every 25 days with 5000 nauplii of Artemia salina
(Sorgeloos & Persoone 1975; Sella 2007).
Development and transformation of the cuttings
into small colonies with a stalk and a polypary was
monitored weekly by digital photography and their
survival was calculated at the end of the experiments
(day 60). The cuttings were then removed with a scalpel from the stubs and placed individually in aluminium dishes (5 cm in diameter). Their dry weight was
determined by drying (Binder oven,60 1C) for 24 h and
then weighing to an accuracy of four decimal points
(Mettler balance, Toledo AB54-S, Mettler-Toledo,

Columbus, OH, USA). Subsequently, they were burnt
(Carbolite furnace,450 1C) for 6 h and their inorganic
weight was determined. The organic weight of each
cutting was calculated by subtracting the inorganic
from the dry weight, and the result was used to calculate the percentage of organic weight of each cutting. The stubs were reused after burning (800 1C,
4 h) for removal of fouled material.
E¡ect of environmental parameters
In order to determine the e¡ect of temperature on the
cuttings of S. glaucum, they were reared at 20, 23, 26
and 29 1C, falling within the range representing Eilat
annual seawater temperatures Æ 1 1C (IUI Data Base,

1750

Aquaculture Research, 2010, 41, 1748–1758

) (10^15 cuttings per aquaria:
March^May 2005). Seawater at 20 1C was circulated
from the tanks to all the aquaria. The 23, 26 and 29 1C
seawater temperatures were obtained using 60^300 W
heaters with thermostats. The temperature was routinely monitored using an alcohol thermometer.
To determine the e¡ect of illumination on the cuttings, they were reared at light intensities of 20, 35,
130 and 250 mE (mE m À 2 s À 1), as measured on their
surface, falling within the prevailing Eilat values
along the depth gradient (Winters, Loya & Beer
2006) (20^30 cuttings per aquaria: December 2005^
January 2006). In order to supply 20 mE, a PVC ¢lter
(PLASON UV1) was placed between the light source
and the surface of the respective aquaria, while for
130 and 250 mE, one or two additional neon bulbs

were added. Light intensity was monitored daily
using an underwater-light photometer (Bio-sciences,
Light 250L Li-CO).
To determine the e¡ect of salinity on the cuttings,
the aquaria were disconnected from the tanks. The
cuttings were reared under salinities of 30, 34, 37 and
40 ppt, representing the salinity within the geographic
distribution of the species (Levitus 1982) (15^25 cuttings per aquaria: February^March, 2006). Each aquarium had a 60 W heater with a thermostat, a ¢lter
(JEBO Eco) and an air stone.Water from the 1m3 tanks
(40 ppt) was diluted with distilled water in order to obtain the lower salinities. Ten litres of seawater was
changed daily in each aquarium and the salinity was
monitored using a light refractometer and an electronic conductivity meter (MRC Salinity 1, Lutron YK315A, Lutron Electronic Enterprise,Taipei,Taiwan).
To determine the e¡ect of feeding, cuttings were fed
at intervals of 2, 7 and 30 days, where a 2-day interval is commonly used by aquarists (f
keeping.com) (12^20 cuttings per aquaria: April^
May 2006). Five thousand nauplii were supplied to
each aquarium for each feeding episode, which lasted
3 h, and was terminated by replacing the water.
Organic weight in ¢eld-collected colonies
To examine the percentage of organic weight in
S. glaucum colonies of di¡erent sizes, samples were
removed from colonies of three polypary diameters:
5^7, 10^15 and 20^30 cm (n 5 6 for each size group:
August 2006). Each sample constituted a longitudinal section comprising ca. 10% of each colony, and
included respective portions of the polypary and the
stalk. Their dry weight, inorganic weight and percentage of organic weight were determined.

r 2010 The Authors
Aquaculture Research r 2010 Blackwell Publishing Ltd, Aquaculture Research, 41, 1748^1758