2. Materials and methods
2.1. Experimental design and data analysis Artemia franciscana cysts (Unibest™ 020730, marine strain) were hatched under standard conditions (Sorgeloos et al., 1986, 2001):
cysts were decapsulated and subsequently incubated for 24 h in 6 L
cylindrical tanks with water at 28 °C and salinity 28, strong bottom
aeration (near saturated oxygen levels, N5 ppm) and light (2000 lx)
at a maximum density of 2 g L−1. The number of newly hatched
A. franciscana nauplii obtained was estimated by sub-sampling (ten
replicates of 1 mL sub-samples).
A. franciscana naupliiwere then stocked at a density of 50 naupliimL−1
in 5 L cylindrical tanks with strong bottom aeration (near saturated
oxygen levels, N5 ppm) and light (2000 lx), and enrichedwith 0.2 g L−1
of AlgaMac 2000® as recommended by the manufacturer (Aquafauna
Bio-Marine Inc.). According to the manufacturer, this product has a 27
DHA:0.54 EPA ratio.
The effects of temperature (16, 20, 24 and 28 °C), salinity (3, 13,
23 and 33) and enrichment time/duration (6, 12, 18 and 24 h) on
A. franciscana nauplii survival, total length and fatty acid profile after
enrichment were tested using a factorial design; three replicates were
used for each combination (N=192). Temperature was controlled by
submersible heaters. The different salinities were generated by dilution
of natural seawater with freshwater. All water used was filtered (1 μm)
and sterilized through ultraviolet radiation.
For each replicate of a treatment at the conclusion of the respective
enrichment durations, the number of live nauplii in ten 1 mL subsamples
was counted; the average count was used to estimate enriched
A. franciscana percent survival of that replicate. In addition, 30 nauplii
per replicate were haphazardly chosen, fixed with a solution of iodine
(2%) for total length (TL) measurement under a stereomicroscope
(Olympus™, model SZ6045TR) with a calibrated micrometer eyepiece
to the nearest 0.001mm. The TL of newly hatched A. franciscana nauplii
was measured in triplicate (N=17,280).
To determine the fatty acid profile of A. franciscana nauplii enriched
under the different combinations of salinity, temperature and enrichment
time, enriched nauplii were rinsed in freshwater and sampled
(one sample of 1.06–1.75 g per replicate, 3 replicates for each treatment,
N=192). Newly hatched A. franciscana nauplii were also sampled in
triplicate. After samples were freeze-dried, samples were ground in a
Potter homogenizer with chloroform–methanol–water (2:2:1.8) (Bligh
and Dyer, 1959). An internal standard fatty acid (19:0) was added to
the extracts. After saponification and esterification of the lipid extracts
(Metcalfe and Schmitz, 1961), the fatty acid methyl esters (FAME) were
injected into capillary columns (30mfused silica, 0.32 I.D.) installed in a
Varian Star 3400CX gas–liquid chromatograph (GLC). Heliumwas used
as carrier gas at a flow rate of 1mLmin−1; oven temperaturewas 180 °C
for 7 min, then 200 °C (with a temperature gradient of 4 °C min−1) over
a period of 71 min. Both the injector and the FID detector were set at
250 °C. GLC data acquisition and handling were performed using a
Varian integrator 4290 connected to the GLC. Peak quantification was
carried out with a Star Chromatographyworkstation. Peak identification
was performed usingwell-characterized cod liver oil chromatograms as
a reference.
Knowing the average TL of the A. franciscana nauplii in each sample,
individual A. franciscana nauplius dry weight for each sample was
calculated using the equation:
Artemia nauplii dryweight = 1:7065 e0:7446 TL R2 = 0:98
where dry weight (DW, μg) and TL (mm), obtained by fitting a curve
to the data published by Reeve (1963) and Narciso (2000). Knowing
the fatty acid (FA) dry weight per unit of dry weight in each sample
and individual A. franciscana nauplius dry weight for all combinations
of the factorial design, the fatty acid content of each A. franciscana
nauplius was estimated. Using fatty acid composition per nauplius is
important since comparing DWof FA per DWof samples might lead to
erroneous conclusions (Rønnestad, 1995).
Response surface regression were used to analyze interactive,
polynomial and interactive by polynomial effects of the continuous
predictors (temperature, salinity and enrichment time) on the survival,
total length and fatty acid profile of the A. franciscana nauplii. A backward
stepwise method was used in these regressions to eliminate the
predictors that do not significantly affect survival, total length and fatty
acid profile of the A. franciscana nauplii (Kuehl, 1994). Analyses were
performed in Statistica 7.0 at a level of significance of 0.05.
2.2. Enrichment model
2.2.1. State variables, forcing functions and processes
The initial values of the state variables A. franciscana percent
survival, TL and fatty acid composition, were the ones presented by
the newly hatched A. franciscana nauplii (enrichment time=0). To
build a model to predict A. franciscana survival, length and fatty acid
profile (state variables), it is important to understand how the forcing
functions temperature, salinity and enrichment time affect the most
important processes occurring during enrichment (Table 1). The
forcing functions may accelerate, delay or have neutral affects on the
different processes. The equations of the processes were obtained
by differentiating the response surface regressions for survival, total
length and fatty acid composition (for the most important fatty acids
and groups of fatty acids) in order to enrichment time. Percent
survival decreases over enrichment time (ET) through the process of
mortality. A. franciscana nauplii total length increases because of the
growth process. The fatty acids composition of the nauplius changes
over enrichment time because of the processes of FA ingestion,
catabolism and synthesis. The ideal enrichment model would take
into account the concentration of each fatty acid in the enrichment
product and its oxidation rate (McEvoy et al., 1995), as well as the
filtration rate of the A. franciscana nauplius (which would probably
vary over development), and fatty acid conversions (synthesis and
catabolism) (Navarro et al., 1999). This approach is, however, too
complex to access with any form of precision since it is very difficult
to know what is happening in the nauplius and, even more difficult,
to distinguish between the processes of ingestion, excretion, synthesis
and catabolism. Therefore, the enrichment model developed used a
process that we named FA incorporation which entails all of these
processes without differentiating them. The differential equations
system that represents the change in each state variable over enrichment
time (due to the processes of FA incorporation, mortality and
growth) are:
dARA
dET
= + Incorporation of ARA ð1Þ
dEPA
dET
= + Incorporation of EPA ð2Þ
dDHA
dET
= + Incorporation of DHA ð3Þ
dLA
dET
= þIncorporation of LA ð4Þ
dALA
dET
= þIncorporation of ALA ð5Þ
dSFA
dET
= + Incorporation of SFA ð6Þ
dBFA
dET
= + Incorporation of BFA ð7Þ
dMUFA
dET
= þIncorporation of MUFA ð8Þ
dPUFA
dET
= + Incorporation of PUFA ð9Þ
dSurvivalk
dET
= − Mortality ð10Þ
dTL
dET
= + Growth ð11Þ
where LA is linoleic acid, ALA is linolenic acid, SFA is saturated fatty
acids, BFA is branched fatty acids and MUFA is monounsaturated fatty
acids.
The model was developed in Matlab 6.0® and modified and run in
STELLA 9.0.3®.
2.2.2. Sensitivity analysis
Sensitivity analyses were carried out to determine which inputs in
the model contributed most for the output variability. The analysis
was conducted by means of successive simulations, by varying each
parameter and forcing function by 10%, up and down, of their initial
baseline values (keeping the others at baseline) and recording the
corresponding change in the state variables. Thus, the sensitivity, S, of
a parameter, P, is defined as:
S =
δψ = ψ
δP = P
ð12Þ
where Ψ is the state variable under consideration (Jørgensen and
2.1. Experimental design and data analysis Artemia franciscana cysts (Unibest™ 020730, marine strain) were hatched under standard conditions (Sorgeloos et al., 1986, 2001):
cysts were decapsulated and subsequently incubated for 24 h in 6 L
cylindrical tanks with water at 28 °C and salinity 28, strong bottom
aeration (near saturated oxygen levels, N5 ppm) and light (2000 lx)
at a maximum density of 2 g L−1. The number of newly hatched
A. franciscana nauplii obtained was estimated by sub-sampling (ten
replicates of 1 mL sub-samples).
A. franciscana naupliiwere then stocked at a density of 50 naupliimL−1
in 5 L cylindrical tanks with strong bottom aeration (near saturated
oxygen levels, N5 ppm) and light (2000 lx), and enrichedwith 0.2 g L−1
of AlgaMac 2000® as recommended by the manufacturer (Aquafauna
Bio-Marine Inc.). According to the manufacturer, this product has a 27
DHA:0.54 EPA ratio.
The effects of temperature (16, 20, 24 and 28 °C), salinity (3, 13,
23 and 33) and enrichment time/duration (6, 12, 18 and 24 h) on
A. franciscana nauplii survival, total length and fatty acid profile after
enrichment were tested using a factorial design; three replicates were
used for each combination (N=192). Temperature was controlled by
submersible heaters. The different salinities were generated by dilution
of natural seawater with freshwater. All water used was filtered (1 μm)
and sterilized through ultraviolet radiation.
For each replicate of a treatment at the conclusion of the respective
enrichment durations, the number of live nauplii in ten 1 mL subsamples
was counted; the average count was used to estimate enriched
A. franciscana percent survival of that replicate. In addition, 30 nauplii
per replicate were haphazardly chosen, fixed with a solution of iodine
(2%) for total length (TL) measurement under a stereomicroscope
(Olympus™, model SZ6045TR) with a calibrated micrometer eyepiece
to the nearest 0.001mm. The TL of newly hatched A. franciscana nauplii
was measured in triplicate (N=17,280).
To determine the fatty acid profile of A. franciscana nauplii enriched
under the different combinations of salinity, temperature and enrichment
time, enriched nauplii were rinsed in freshwater and sampled
(one sample of 1.06–1.75 g per replicate, 3 replicates for each treatment,
N=192). Newly hatched A. franciscana nauplii were also sampled in
triplicate. After samples were freeze-dried, samples were ground in a
Potter homogenizer with chloroform–methanol–water (2:2:1.8) (Bligh
and Dyer, 1959). An internal standard fatty acid (19:0) was added to
the extracts. After saponification and esterification of the lipid extracts
(Metcalfe and Schmitz, 1961), the fatty acid methyl esters (FAME) were
injected into capillary columns (30mfused silica, 0.32 I.D.) installed in a
Varian Star 3400CX gas–liquid chromatograph (GLC). Heliumwas used
as carrier gas at a flow rate of 1mLmin−1; oven temperaturewas 180 °C
for 7 min, then 200 °C (with a temperature gradient of 4 °C min−1) over
a period of 71 min. Both the injector and the FID detector were set at
250 °C. GLC data acquisition and handling were performed using a
Varian integrator 4290 connected to the GLC. Peak quantification was
carried out with a Star Chromatographyworkstation. Peak identification
was performed usingwell-characterized cod liver oil chromatograms as
a reference.
Knowing the average TL of the A. franciscana nauplii in each sample,
individual A. franciscana nauplius dry weight for each sample was
calculated using the equation:
Artemia nauplii dryweight = 1:7065 e0:7446 TL R2 = 0:98
where dry weight (DW, μg) and TL (mm), obtained by fitting a curve
to the data published by Reeve (1963) and Narciso (2000). Knowing
the fatty acid (FA) dry weight per unit of dry weight in each sample
and individual A. franciscana nauplius dry weight for all combinations
of the factorial design, the fatty acid content of each A. franciscana
nauplius was estimated. Using fatty acid composition per nauplius is
important since comparing DWof FA per DWof samples might lead to
erroneous conclusions (Rønnestad, 1995).
Response surface regression were used to analyze interactive,
polynomial and interactive by polynomial effects of the continuous
predictors (temperature, salinity and enrichment time) on the survival,
total length and fatty acid profile of the A. franciscana nauplii. A backward
stepwise method was used in these regressions to eliminate the
predictors that do not significantly affect survival, total length and fatty
acid profile of the A. franciscana nauplii (Kuehl, 1994). Analyses were
performed in Statistica 7.0 at a level of significance of 0.05.
2.2. Enrichment model
2.2.1. State variables, forcing functions and processes
The initial values of the state variables A. franciscana percent
survival, TL and fatty acid composition, were the ones presented by
the newly hatched A. franciscana nauplii (enrichment time=0). To
build a model to predict A. franciscana survival, length and fatty acid
profile (state variables), it is important to understand how the forcing
functions temperature, salinity and enrichment time affect the most
important processes occurring during enrichment (Table 1). The
forcing functions may accelerate, delay or have neutral affects on the
different processes. The equations of the processes were obtained
by differentiating the response surface regressions for survival, total
length and fatty acid composition (for the most important fatty acids
and groups of fatty acids) in order to enrichment time. Percent
survival decreases over enrichment time (ET) through the process of
mortality. A. franciscana nauplii total length increases because of the
growth process. The fatty acids composition of the nauplius changes
over enrichment time because of the processes of FA ingestion,
catabolism and synthesis. The ideal enrichment model would take
into account the concentration of each fatty acid in the enrichment
product and its oxidation rate (McEvoy et al., 1995), as well as the
filtration rate of the A. franciscana nauplius (which would probably
vary over development), and fatty acid conversions (synthesis and
catabolism) (Navarro et al., 1999). This approach is, however, too
complex to access with any form of precision since it is very difficult
to know what is happening in the nauplius and, even more difficult,
to distinguish between the processes of ingestion, excretion, synthesis
and catabolism. Therefore, the enrichment model developed used a
process that we named FA incorporation which entails all of these
processes without differentiating them. The differential equations
system that represents the change in each state variable over enrichment
time (due to the processes of FA incorporation, mortality and
growth) are:
dARA
dET
= + Incorporation of ARA ð1Þ
dEPA
dET
= + Incorporation of EPA ð2Þ
dDHA
dET
= + Incorporation of DHA ð3Þ
dLA
dET
= þIncorporation of LA ð4Þ
dALA
dET
= þIncorporation of ALA ð5Þ
dSFA
dET
= + Incorporation of SFA ð6Þ
dBFA
dET
= + Incorporation of BFA ð7Þ
dMUFA
dET
= þIncorporation of MUFA ð8Þ
dPUFA
dET
= + Incorporation of PUFA ð9Þ
dSurvivalk
dET
= − Mortality ð10Þ
dTL
dET
= + Growth ð11Þ
where LA is linoleic acid, ALA is linolenic acid, SFA is saturated fatty
acids, BFA is branched fatty acids and MUFA is monounsaturated fatty
acids.
The model was developed in Matlab 6.0® and modified and run in
STELLA 9.0.3®.
2.2.2. Sensitivity analysis
Sensitivity analyses were carried out to determine which inputs in
the model contributed most for the output variability. The analysis
was conducted by means of successive simulations, by varying each
parameter and forcing function by 10%, up and down, of their initial
baseline values (keeping the others at baseline) and recording the
corresponding change in the state variables. Thus, the sensitivity, S, of
a parameter, P, is defined as:
S =
δψ = ψ
δP = P
ð12Þ
where Ψ is the state variable under consideration (Jørgensen and
