Journal of Experimental Marine Biology and Ecology,
234 (1999) 147–166
L
Nutrient control of phytoplanktonic biomass in atoll lagoons
and Pacific ocean waters: Studies with factorial enrichment
bioassays
P. Dufour a , *, B. Berland b
a
ORSTOM, Station Marine d’ Endoume, rue de la Batterie des Lions, 13007 Marseille, France
b
CNRS, Station Marine d’ Endoume, rue de la Batterie des Lions, 13007 Marseille, France
Received 29 July 1997; received in revised form 23 July 1998; accepted 30 July 1998
Abstract
Although the atolls of the Tuamotu archipelago (Central South Pacific) are located in an
oligotrophic oceanic area, some of their lagoons have experienced exceptionally harmful
phytoplanktonic blooms in the last 30 years. Twenty-four differential enrichment bioassays were
conducted on 10 atoll lagoons and 5 ocean sites at two different times of the year in order to
determine, among other factors, which nutrients may control phytoplanktonic crop. Complete
factorial (2 3 ) design with N, P and Si and fractional factorial (2 823 ) design with N, P, Si, chelator,
Fe, Mo, Mn and vitamins were performed. In vivo fluorescence (IVF) was used to follow the
growth of phytoplankton. Although this method is imperfect, we argue that the large increases in
fluorescence, observed in response to some spikes, indicate biomass shifts. Nitrogen, phosphorus
and sometimes silicate effects were significant. The nitrogen effect was greatest in 17 out of 24
samples. In the smallest lagoon, the phosphorus effect was higher than the nitrogen effect. In the
six other samples N and P effects were similar. Silicate spikes resulted in a significant effect for
only seven samples. Vitamins, Mo, Mn, iron and chelators had little or no effect. In 20 bioassays
there was also a synergistic effect when N and P were added simultaneously. This synergistic
effect was present in five bioassays when N, P, and Si were added simultaneously. The season or
origin (lagoon or ocean) had little influence on these effects. On average fluorescence attained by
samples supplemented with N alone was six times that of controls (unspiked). The highest
responses were observed with combined N 1 P or N 1 P 1 Si spikes. Combined N 1 P fertilization
produced an in vivo fluorescence ranging from 5 to 85 times the fluorescence attained by controls.
This synergy is consistent with the view that both N and P are in relatively short supply. Ocean
waters appeared to be nitrogen depleted. Phosphorus limitation increased in small lagoons with
low advective ocean waters and a large ratio of immersed surfaces to water volume. Atoll lagoons
and ocean waters from the Tuamotu archipelago appear to be highly susceptible to eutrophication
with the addition of both nitrogen and phosphorus.  1999 Elsevier Science B.V. All rights
reserved.
*Corresponding author. Tel.: 04 91 04 16 14; fax: 04 91 04 16 35; e-mail: [email protected]
0022-0981 / 99 / $ – see front matter  1999 Elsevier Science B.V. All rights reserved.
PII: S0022-0981( 98 )00135-X
148
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
Keywords: Atoll; Factorial bioassays; Nutrients control; Pacific; Phytoplankton
1. Introduction
There are 77 atolls in the Tuamotu archipelago. Several have experienced exceptional
phytoplanktonic blooms in the last 30 years. These were sometimes followed by massive
fish, coral and cultivated pearl oyster mortality (Harris and Fichez, 1995). Little is
known of the factors controlling the phytoplankton communities of these atoll lagoons.
Among many physical, chemical and biological factors, it is well documented that
nutrients regulate density, dominance and successions of phytoplanktonic species (e.g.
Tilman et al., 1982). Our objective was to identify which macro- or micro-nutrients may
control the phytoplanktonic standing crop of the Tuamotu atoll lagoons. To our
knowledge, phytoplankton control by nutrients has only been investigated in three of the
413 atoll lagoons in the world. In Tarawa lagoon (Republic of Kiribati, South Pacific
ocean), Kimmerer and Walsh (1981) suggested that N was limiting because of the low
dissolved inorganic N / P ratio. In the confined lagoons of Christmas Island and Canton
Atoll (North Pacific) reactive phosphorus appeared depleted well below ocean levels and
was assumed to limit primary producers (Smith, 1984).
In various studies on coral reef waters different nutrients were shown to control the
growth of primary producers: Mo (Howarth and Cole, 1985) or Fe (Entsch et al., 1983b)
or P (Atkinson and Smith, 1983; Entsch et al., 1983a; Smith, 1984; Littler et al., 1991)
or N (Patriquin, 1972; Laws and Redalje, 1979; Smith et al., 1984; Laws and Allen,
1996). These discrepancies may be related to different methodological approaches which
reflect different understandings of the concept of limitation (see below) but may also
reflect actual differences in community under investigation: pico-, nano-, or microphytoplankton or benthic algae. They may also be related to the spatial heterogeneity
within lagoons and other coral reef systems and to the diversity of their environment. In
any case, these systems are permanently exchanging waters with the surrounding ocean
waters of different nutrient richness. For instance, D’Elia and Wiebe (1990) reported a
wide range of dissolved inorganic phosphate levels (0.025–0.56 mmol ? l 21 ) in the
vicinity of coral reef. It would, therefore, be useful to know the nutrient content of the
surface ocean waters of Tuamotu. Only a few chemical indices argue for a limitation by
N (Rancher and Rougerie, 1994; Raimbault et al., 1998) or a colimitation by N and Fe
(Lindley et al. 1995). Due to this insufficient knowledge, we extended our research of
nutrient control to some samples of the surface ocean waters surrounding the atolls.
Tuamoto atolls have a high morphological diversity likely to influence the nutrient
regime of their lagoons. In order to avoid a dubious generalisation from some particular
cases, the study was processed on 10 atolls selected according to two morphological
parameters: lagoon area and proportion of submerged rim. Both parameters have a clear
influence on the degree of water exchange between lagoons and the surrounding ocean, a
hydrodynamic factor which was proved to play a decisive role on nutrient budget in
lagoons (Smith, 1984). Lagoon area of the 10 selected atolls ranges from 0.7 to 315
km 2 . The proportion of their submerged rim (submerged reef-top, spillways, passages...)
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
149
Table 1
Characteristics of the 10 prospected atolls (ordered by increasing surfaces)
Atoll
Slag
Sat.
Depth
Por.
Inh.
Main characteristics of the choice
Reka Reka
Tepoto Sud
Tekokota
Haraiki
Taiaro
Hiti
Nihiru
Hikueru
Marokau
Kauehi
0.7
1.6
5.1
10.4
12
15
80
82
217
315
5.2
6.2
7.3
24.6
17
25
100
107
256
343
1.5
5
3
12
13
10
21
28
28
45
2
15
59
19
1
18
25
18
17
22
0
0
0
20
3
0
20
300
50
200
small size, low porosity
small size, medium porosity
small size, high porosity
small to medium size, medium porosity
small to medium size, low porosity
small to medium size, medium porosityy
medium size, medium porosity
medium size, medium porosity, dystrophic crisis
big size, medium porosity
big size, medium porosity
Slag: surface area of lagoon (km 2 ); Sat: surface area of atoll (km 2 ); depth (m): estimated average depth;
porosity (Por): percentage of immersed atoll rim that allows water exchange between the lagoon and the ocean;
´ ¨ pers. com.
inhabitants (Inh). Surfaces and porosity from Andrefouet,
is between 1% and 59% of circumference (Table 1). These ranges include most values
´
of lagoon area and atoll rim porosity encountered among the Tuamotu atolls (Andref¨ pers. com.).
ouet,
The degree to which nutrient availability controls phytoplanktonic growth rates or
standing crop may be assessed by different approaches (Maestrini et al., 1984; Elser et
al., 1990). The biochemical approach relies on the cellular chemical composition or any
metabolic activity known to be affected by nutritive deficiencies. With the in situ
approach, nutrient concentrations in waters are compared with metabolic needs, for
instance cellular half-saturation constants for nutrient uptake. The experimental approach
uses the response of phytoplankton itself to the addition of nutrients. Each of these
approaches has its own advantages and drawbacks. Moreover, they lead to different
interpretations for nutrient limitations. Nutrient availability can regulate rate processes
according to the Blackman concept or can limit the phytoplanktonic standing crop
according to the Liebig concept (Cullen, 1991). In this study we used the experimental
approach with a protocol (batch cultures in small volumes, nutrient enrichments and in
vivo fluorescence as response parameter) in respect with algal growth as crop, not as
growth rate. Nutrient limitation was assessed with a large series (24) of enrichment
bioassays. Complete factorial and fractional factorial designs were used. These designs
are rarely used in oceanography in spite of their advantages as evidenced below.
2. Materials and methods
Samples were collected on two cruises (TYP3, TYP4), in November 1995 and March
1996, on the R.V. Alis, at atolls and ocean sites located between 158459S–188039S and
1418559W–1458099W (Fig. 1). The main characteristics of the studied atolls are shown
on Table 1.
Nutrient control was assessed from water samples collected in the center of every
lagoon. Due to the low intra-lagoonal variations in the Tuamotu Archipelago (Charpy et
150
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
Fig. 1. Location of the ten atoll lagoons and five ocean sites sampled in the Tuamotu archipelago (French
Polynesia).
al., 1997; Dufour and Harmelin–Vivien, 1997) these central sites were considered
representative of mean lagoon conditions.
Water was collected and handled with care in order to minimize contamination.
Subsurface water samples (0.5 m depth) were collected leeward from a plastic boat in
acid cleaned polyethylene jars and kept in an isotherm container, at in situ temperature,
in the dark. All samples were processed within 2 h of collection. We used polyethylene
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
151
or polycarbonate containers soaked overnight in 10% HCl and rinsed three times with
Nanopure water, then with sample water immediately before filling.
Limiting nutrients were identified by the effect of nutrient additions on the standing
crop of natural phytoplankton assemblages. In November 1995 (TYP3) we studied the
role of nitrogen, phosphorus and silicon. In March 1996 (TYP4) in addition to these
three macro nutrients we studied the impact of several trace compounds:
–a B12, Biotine and thiamine vitamin pool, as in environments with high UV they
can be photodestroyed (Carlucci et al., 1969).
–the trace metals Fe, Mo, and Mn which are generally related to respiration,
photosynthesis and nitrogen fixation and are known to be limiting in several natural
phytoplankton populations (Sournia and Citeau, 1972; Howarth and Cole, 1985;
Martin et al., 1994; Boyd et al., 1996).
–we also estimated the level of complexation of these waters from the response of the
populations to EDTA (Ethylene diamine tetracetic acid) addition.
Complete factorial and fractional factorial designs were used (Cochran and Cox,
1957; Vigier, 1988). Concentrations of nutrients were varied at two levels. At the lower
level (–) no nutrient was added resulting in concentrations identical to in situ
concentrations which varied largely according to the samples. Dissolved inorganic
nitrogen ranged from undetectable concentrations to 1.18 mM whereas dissolved
inorganic phosphorus ranged from undetectable concentrations to 0.37 mM (data not
shown). Dissolved organic nitrogen and dissolved organic phosphorus, measured
according to Pujo–Pay et al. (1997), reached 14.8 and 0.43 mM, respectively. At the
higher level ( 1 ) ultrapure chemicals (p.a. Merck quality) were added to the samples
(Table 2) to final concentrations below the concentrations suggested by Maestrini et al.
(1984) and Boyd et al. (1996) for bioassays in oligotrophic waters. Nevertheless nutrient
concentrations in spiked samples were high compared to those found in the surrounding
Pacific ocean waters of the area and in the most oligotrophic lagoons. The choice of
added concentrations depends on conflicting needs. (1) The nutrients supplied as
enrichments have to be in significantly greater concentration than in situ in order to
Table 2
Nutrients, chemical compounds and nutrient final concentrations (mole l 21 )
Nutrient
chemicals
final concentration
N
P
Si
Fe1Edta
Mn
Mo
Edta
Vitamin B12
Vitamin B1
Vitamin H
NaNO 3
NaH 2 PO 4 , H 2 0
Na 2 SiO 3 , 5H 2 O
FeSO 4 , 7H 2 01 Na 2 Edta
MnSO 4 , H 2 0
Na 2 MoO 4 , 2H 2 0
Na 2 Edta
Cyanocobalamine
Thiamine
Biotin
10 25 M
2310 26 M
10 25 M
10 28 M110 28 M
10 28 M
10 27 M
10 27 M
10 29 M
10 28 M
10 210 M
152
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
differentiate clearly between the effect of the added and native nutrient and to generate
rapid and significant responses. (2) Concentrations should be as low as possible so as
not to modify assimilation mechanisms nor introduce toxic effects associated with
excessive concentrations. In our case, nitrogen and phosphorus supplies were less than
an order of magnitude over the concentrations of DIN (dissolved inorganic nitrogen) and
PO 4 measured in the richest lagoons (8 and 5 times these concentrations, respectively).
Moreover N and P addition were in the same order of magnitude as the total dissolved
nitrogen and phosphorus. Silicon additions were no more than three times the
concentrations in the richest lagoon.
The full factorial design used in TYP3 included 8 (2 3 ) bioassays with 3 replicates
(Table 3). To decrease the number of bioassays in TYP4, a fractional factorial design
without replication with 32 (2 823 ) bioassays was used (Table 4). Test tubes, filled with
10 ml samples, were incubated in shadowed (10%) natural light on the ship’s deck. They
were cooled to in situ temperature by a circulation of surface water and the tubes were
inverted twice daily to mix.
Growth of phytoplanktonic biomass was approximated daily by measuring in vivo
chlorophyll fluorescence (IVF). The tubes were in the dark for 30 min every evening
before their fluorescence was measured according to Lorenzen’s method (Lorenzen,
1966) with a Turner 111 fluorometer fitted with a high intensity F4T5 blue lamp and a
red sensitive photomultiplier, a blue Corning filter CS.5–60 for excitation light and a red
filter CS.2–64 for emitted light. Incubation was prolonged until maximum IVF was
reached (4–6 days).
Only maximum IVF was used in calculations. The kind of orthogonal design used
enables regression analysis of single and combined effects (named also contrast in
fractional design) of nutrients. The effect of each nutrient or combination of nutrient (C)
was calculated using either the 27 (3 3 2 3 for TYP3) or 32 (2 823 for TYP4) fluorescence
bioassays regardless of the other nutrients levels, 1 or 2 , (Cochran and Cox, 1957;
Vigier, 1988) with the formula:
C 5 1 / 2(IVF at level 1 ) 2 (IVF at level 2 )
(1)
In each atoll, significant effects were determined by a graphic method developed for
the analysis of two-level factorial experiments (Henry’s line: Daniel, 1959). The
calculated effects are sorted by increasing order and plotted as a function of their order
number: the values of the effects along the x-axis on a linear scale, their order number
along the y-axis on a Gaussian scale (Fig. 2). In this representation, a normal distribution
becomes linear. Assuming that the non-significant effects follow a normal distribution,
they are aligned. Significant effects corresponding to nutrients or interactions of nutrients
affecting in vivo fluorescence depart from this regression line. The level of significance
was assessed using multiway ANOVA (Vigier, 1988). Significant effects are reported on
Tables 5 and 6.
To calculate the modelled response (modelled IVF) of the phytoplankton community
in each lagoon (Figs. 3 and 4), the significant effects were used in a polynomial
regression model:
Assay
nutrient
N8
1
2
3
4
5
6
7
8
N
2
1
2
2
1
1
2
1
Highest response (fluorescence)
P
2
2
1
2
1
2
1
1
Si
2
2
2
1
2
1
1
1
Rek
11
12
8
12
33
13
12
65
Tep
4
31
8
0
122
31
3
164
Tek
2
20
6
4
55
66
0
124
Har
0
10
1
1
68
6
1
55
Tai
0
23
3
5
78
27
4
122
Tep
4
31
8
0
122
31
3
164
Hiti
4
6
4
3
55
7
5
132
Har
0
10
1
1
68
6
1
55
Tai
0
23
3
5
78
27
4
122
Nih
14
17
16
15
220
24
18
250
Hik
2
0
2
3
43
9
6
36
Mar
2
30
3
4
64
55
9
81
Kau
5
10
9
9
74
18
13
143
Oc. 1
5
10
1
1
39
2
0
81
Oc. 4
5
7
6
0
55
5
4
64
Oc. 6
8
14
11
9
123
15
6
162
Rek (Reka Reka); Tek (Tekokota); Tep (Tepoto Sud); Har (Haraiki); Tai (Taiaro); Nih (Nihiru); Hik (Hikueru); Mar (Marokau); Kau (Kauehi); Oc. (Ocean).
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
Table 3
Matrix of 2 3 factorial design for 13 lagoons and ocean samples in November 1995 (TYP3); (1) when nutrient was added and (2) when nutrient was not added in an
assay. Highest fluorescence response for each assay are shown
153
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
154
Table 4
Matrix of 2 823 fractional factorial design in March 1996 (TYP4); (1) when nutrient was added and (2) when
nutrient was not added in an assay
Assay
Added nutrient
N8
N
P
Si
Vit.
Fe
Mn
Mo
EDTA
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
2
1
1
2
2
1
1
2
2
1
1
2
2
1
1
2
2
1
1
2
2
1
1
2
2
1
1
2
2
1
1
2
2
2
2
1
1
1
1
2
2
2
2
1
1
1
1
2
2
2
2
1
1
1
1
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
1
2
1
1
2
2
1
1
2
1
2
2
1
1
2
2
1
2
1
1
2
2
1
1
2
1
2
2
1
1
2
2
1
2
1
1
2
1
2
2
1
2
1
1
2
2
1
1
2
1
2
2
1
2
1
1
2
1
2
2
1
1
2
2
1
1
2
2
1
2
1
1
2
2
1
1
2
2
1
1
2
2
1
1
2
1
2
2
1
1
2
2
1
Fluorescence responses are not shown.
Y 5 Ym 1 Xi ? Ci 1 Xi Xj ? Cij
(2)
where Y stands for the calculated response (IVF) for each sample, Ym 5mean IVF of all
assays for each sample; Ci 5effect of nutrient i; Cij 5effect of the interaction of the
nutrients i with j; Xi or Xj 5 11 or 21, according to the presence or absence of the
nutrient i or j in the assay.
Modelled responses were validated by testing, curve fitting, the normality of the
residues on a normal plot of the residues vs. their probability (Daniel, 1959).
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
155
Fig. 2. Three examples of normal plot for determination of significant effects. Calculated effects are sorted by
increasing order and plotted as a function of their order number: on the horizontal axis, effect values on a
linear scale (IVF units) and on the vertical axis, order number on a gaussian scale. The points that are aligned
(s) follow a normal distribution and represent the unsignificant effects. Inversely the points that depart from
the line (d) represent the effects which are significant.
3. Results
Observed maximum IVF within each of the 24 assays of TYP3 and TYP4 varies up to
3 orders of magnitude (Table 3 and data not shown). Significant effects deduced from
the normal plots (e.g. Fig. 2) and ANOVA analysis on maximum IVF are shown on
Tables 5 and 6. The effect and interaction of N and P fertilisation in ocean and lagoon
samples was significant in all cases except in Marokau in November where only N was
significant (TYP3 cruise). In Reka Reka, Nihiru, Kauehi lagoons and one ocean sample
(TYP3) and in Hiti lagoon (TYP3 and TYP4), N and P effects were identical. The effect
of P was highest in Reka Reka (TYP4) while the effect of N was higher than the effect
of P in the 17 other samples. The NP interaction term was as high or nearly as high as
the main effect of N or P. During the TYP3 cruise the effects of Si and of the interaction
SiN were significant once and the interaction PSi was never significant. During TYP4
the effects of Si and the interaction SiN were significant 6 times and the interaction PSi
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
156
Table 5
Nutrient effects and interaction C (calculated according to Eq. (1)) in 13 lagoons and ocean samples in
November 1995 (TYP3); significant effect ( p,0.05) in bold type; last column: the significant nutrients and
interactions are classified according to their importance
Site
N
P
Si
NP
NSi
PSi
NPSi
Reka Reka
Tepoto
Tekokota
Haraiki
Taiaro
Hiti
Nihiru
Hikueru
Marokau
Kauehi
Ocean 1
Ocean 4
Ocean 6
10
42
32
17
30
23
60
9
27
26
16
14
35
9
29
12
14
19
22
58
9
8
25
13
14
32
4
4
14
22
7
10
9
1
6
8
4
0
5
10
27
12
13
19
22
57
8
7
25
14
13
32
3
6
15
22
5
10
8
0
4
9
5
2
6
4
5
2
21
5
10
7
22
21
8
7
2
4
4
5
4
21
6
9
6
22
22
10
6
1
5
N5P5NP
N. .P.NP
N. .NSi 5Si 5P5NP
N.P5NP
N. .P5NP
N5P5NP
N.P5NP
N5P5NP
N
N5P5NP
N.P5NP
N5P5NP
N.P5NP
was significant 5 times. None of the trace elements added (TYP4) had significant effect
on IVF with the exception of iron which had a slight and negative effect in 2 lagoons.
IVF responses modelled according to Eq. (2) were used to assess the importance of
various nutrients in terms of magnitude (Figs. 3 and 4). With the exception of Reka
Reka, the addition of N induced higher values of fluorescence than in the control. The
response was sometimes weak but on average, fluorescence attained by samples
supplemented with N alone was 6 times the fluorescence of controls (from 1 in Reka
Reka to 40 in Tepoto sud in March). Cruise (TYP3 or TYP4) or origin (lagoon or ocean)
did not have significant effect on the IVF response to N. There was a slight response
above control levels with the addition of P in 6 out of 24 samples. On average
fluorescence attained by samples supplemented with P alone was 1.7 times the
fluorescence of controls. There was no significant IVF response when other nutrients,
metals, vitamins, Si and chelator were added. In all cases the highest responses were
observed in assays enriched with N1P or N1P1Si. Combined N1P fertilisation
produced an in vivo fluorescence ranging from 5 to 85 times the fluorescence attained by
controls (mean527, SD522, n524). These responses were stronger than responses to
fertilisation by N or P alone with the exception of Marokau in November where the IVF
responses to N1P and N enrichments were equal. In seven samples the response to
N1P enrichment was lower than the response to N1P1Si enrichment which ranged
from 12 to 261 times the control (mean548, SD569, n524). The frequency and
magnitude of these synergistic responses emphasise the interplay between N, P and Si
availability in constraining the algal standing crop in the Tuamotu lagoons and the
surrounding ocean.
Site
N
P
Reka Reka
Tepoto
Tekokota
Haraiki
Taiaro
Hiti
Nihiru
Hikueru
Marokau
Ocean 7
Ocean 9
35
107
24
34
8
47
69
49
38
36
25
38
64
16
13
4
47
66
40
34
23
22
Si
48
10
10
Fe
27
214
45
18
11
NP
36
64
16
13
3
47
66
40
34
22
22
NSi
46
11
7
NFe
PSi
26
41
6
PFe
NPSi
NPFe
41
6
25
2
214
215
213
47
46
47
19
12
18
8
17
9
P.NP5N
N.P.NP.Si5NSi.PSi5NPSi
N.P5NP.NSi5Si.Fe5PSi5NPSi5Fe
N.P5NP.Si.NSi.PFe
N.P5NP5PSi
N5P5NP..PFe5Fe5NFe5NPFe
N.P5NP.NSi5NPSi5PSI5Si
N.P5NP
N.P5NP.NSi5Si5PSi5NPSi
N.P5NP.NSi5Si5NPSi5PSi
N.P5NP
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
Table 6
Significant ( p,0.05) nutrients effects and interactions C (calculated according Eq. (1)) in 11 lagoons and ocean samples in March 1995 (TYP4); last column: the
significant nutrients and interactions are classified according to their importance
157
158
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
Fig. 3. November 1995 (TYP3); modelled fluorescence responses, calculated according to Eq. (2), for 13
lagoons and ocean samples fertilized with different nutrients and combinations of nutrients. Note the different
fluorescence scales on the two diagrams.
4. Discussion
4.1. Advantages and limits of the factorial bioassays
There were several advantages to the experimental enrichment method and the
factorial design that was applied: i) The response of the algae themselves was used
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
159
Fig. 4. March 1996 (TYP4); modelled fluorescence responses, calculated according to Eq. (2), for 11 lagoons
and ocean samples amended with nutrients and combinations of nutrients. Note that the fluorescence scale is
different on the two diagrams.
therefore indicating nutrient availability for phytoplankton. ii) Natural phytoplanktonic
communities were studied and the results therefore related directly to organisms
originating from the water mass under study. iii) Factorial designs used all the assays in
the nutrient effect calculation allowing greater accuracy of the nutrient additions effect.
iv) Factorial design allowed the evaluation of the effects of interactions between
160
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
nutrients. v) The response of the algal community to nutrient additions could be
modelled using regression equations (see methods). vi) Moreover, the use of fractional
factorial designs allowed a decrease in the number of assays. vii) Finally, it was possible
to follow simultaneously a large number of assays as the measurement of IVF is rapid.
All of the above points to a synoptic view of the role of several nutrients.
On the other hand, in vitro incubations may be misleading as they isolate the algal
assemblage from external sources and sinks of nutrient and alter the physical, chemical
and biological environment (e.g. Maestrini et al., 1984; Elser and Kimmel, 1986). In
addition, other environmental factors (e.g. light, temperature, grazing, mortality, sinking,
vertical mixing, water dilution) may control phytoplankton standing crop in situ.
Methodological drawbacks were decreased wherever possible. The fluorescence measurements were conducted through the culture vials which limits handling and culture
contamination. Temperature and light settings during incubation reproduced in situ
surface conditions. It was of course impossible to simulate natural turbulence and its
associated patterns of irradiance. We did not observe any initial variation of IVF which
would have likely been a response to incubation stress. The recorded IVF increase
lagged the nutrient addition by 1–3 days which may indicate changes in species
population. Species with greater growth rates but less affinity for nutrients at low
concentrations may be favoured. Picophytoplankton accounts for most of the chlorophyll
biomass in the lagoon and ocean waters of the Tuamotu (Charpy et al., 1997; Charpy
and Blanchot, 1998). Generally large eukaryotic phytoplankton, particularly diatoms,
respond better than picoplankton to nutrient increases (Landry et al., 1996). This drift in
species population may affect results when nutrients that regulate natural phytoplankton
are studied but is of special interest if this drift reproduces population changes as is the
case in artificial or accidental enrichment in situ.
In our samples, grazing was present. However, zooplankton larger than 35 mm was
rare (2700 / m 3 on average, Leborgne, pers. com.) and, in most cases likely excluded by
sampling and incubation in small volume tubes. On the other hand protozoans could not
be eliminated. Artefacts due to protozoan grazing and the cascade effect caused by their
excretion of nutrients would be expected in our experiments. They may favour the large
phytoplankton cells that are not grazed by micropredators. As with protozoans, bacteria
cannot be eliminated from any bioassays on natural populations. Therefore, in our
experiments the response of phytoplankton was tested in the presence of nano and
picoplankton communities.
In vivo fluorescence was chosen to measure responses to nutrient additions. It requires
small water volumes and measurement is rapid. Other methods would not have permitted
the monitoring of 456 cultures. However, IVF is an imperfect phytoplanktonic biomass
indicator with variable responses according to species composition and physiological
state of the cells, itself mainly dependent on temperature, light and nutrient history (e.g.
Jeffrey, 1997). We avoided fluorescence variations linked to light stress and day light
cycles by keeping the phytoplanktonic cells in the dark for half an hour before
measurement and by processing the measurement every day at the same hours.
Incubation conditions and measurements were identical for all test tubes of the same
sample. We are aware that the ratio of IVF to chlorophyll a concentration decreases with
nutrient satiation (e.g. Kiefer, 1973). As a consequence, the increase of IVF in spiked
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
161
tubes underestimates the increase of phytoplankton biomass, and the actual differences
should be even more significant than those evidenced here. On the other hand, the
‘‘chlorophyll a / phytoplankton carbon’’ ratio generally increases under nutrient replete
conditions (e.g. Geider et al., 1997). In absence of accurate biomass measurements we
cannot know the importance of the variations of ‘‘IVF / phytoplankton biomass’’ ratios in
the test tubes. Despite these limitations, the large differences in IVF with the different
experimental treatments (Figs. 3 and 4) indicate undoubtedly differences in phytoplanktonic biomass.
4.2. Nutrient control in the atoll lagoons and ocean waters of the Tuamotu
archipelago
The nitrogen effect was greatest in 17 out of 24 samples (Tables 5 and 6). This
suggests that nitrogen is most frequently the nutrient that limits the standing crop of
phytoplankton in the atoll lagoons and ocean waters of the Tuamotu archipelago. Only in
the two Reka Reka samples was the phosphorus effect higher than (TYP4), or equal to
(TYP3), the nitrogen effect. In the last five samples the N effect could not be
distinguished from P effect.
A dominant N limitation may seem paradoxical on the basis of chemical and
biological considerations. The highly oxygenated upper layer of the carbonate rich
sediment of atoll lagoons (Entsch et al., 1983a; Charpy–Roubaud et al., 1996) suggests
that it should adsorb P rather strongly. In Tikehau lagoon Charpy–Roubaud et al. (1996)
demonstrated that diffusional nutrient fluxes from sediment to water column were weak
in comparison to the requirements of primary production and were rather phosphorus
deficient with a mean N:P atomic ratio of 30 (SE52.5). Sorokin (1995) reported a net
flux of PO 4 –P from water column to different bottom biotopes of coral systems.
Moreover N fixation by cyanobacteria is known to occur in coral environments (Smith,
1984; Larkum et al., 1988) as is the case in the Tuamotu atolls (Charpy–Roubaud et al.,
1997).
Despite the above observations other arguments may explain why P might in fact be
less rapidly exhausted from the environment than N: (1) It is generally admitted that the
turnover time of P is faster than that of N (Sorokin, 1995) in coral reef environments.
(2) In the surrounding ocean, DIN / P–PO 4 ratios were all below 1 at. / at. while DIN
concentrations were on average 0.03 mM (data not shown). Such low values indicate a
potential limitation by N rather than by P. (3) The DIN concentrations in lagoons were
generally higher than in the ocean whereas the reverse was observed for PO 4 (Dufour
and Harmelin–Vivien, 1997). Consequently the ratios DIN / P–PO 4 were higher in
lagoons than in the surrounding ocean. In Reka Reka lagoon the DIN / PO 4 –P ratios
were 107 at TYP3 and 24 at. / at. at TYP4 which strongly suggests a limitation by P
rather than N and confirms the bioassays results in this lagoon. In all the other lagoons
the DIN / PO 4 –P ratios were less than 4: values well below the Redfield ratio which
suggest a potential limitation by N rather than P (Redfield et al., 1963; Karl et al., 1993).
´
(4) Torreton
et al. (1998) have tested the effect of N–NH 4 , PO 4 and glucose additions
on the growth of bacterioplankton assemblages issued from the same samples as this
study. They concluded that N–NH 4 was the nutrient most frequently stimulating
162
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
bacterioplankton growth in the 10 lagoons tested. This correspondence between bacterial
and algal nitrogen limitation suggests increased nitrogen competition to the advantage of
the bacterioplankton (e.g., Cotner and Wetzel, 1992).
Silicate enrichment resulted in a significant effect for only 7 out of 24 samples (Tables
5 and 6). When it occurred Si effect was always less than N or P and NP effects. This
may be due to the low abundance of siliceous diatoms in this central Pacific region and
in Tuamotu lagoons dominated by nonsiliceous taxa of picoprocaryotes and
nanoeucaryotes (Landry et al., 1996; Charpy and Blanchot, 1998).
Iron appeared to have little or no effect on the maximum IVF. We don’t explain the
slight negative effect observed in Tekokota and Hiti (Table 6) other than by a decline of
the ratio of IVF / biomass due to a shift in phytoplanktonic species and / or a change in
their physiological state. A lack of positive effect contradicts the N and Fe simultaneous
limitation speculated in the oligotrophic South Pacific at 128S, 1408W by Lindley et al.
1995, and the Fe limitation in carbonate environments of the Great Australian Barrier
Reef (Entsch et al., 1983b). It appears that iron enrichment stimulates the net growth of
micro or nanophytoplankton but not of picophytoplankton like cyanobacterium Synechococcus (Wells et al., 1994). Cyanobacteria can compensate for growth limiting levels of
iron by producing extracellular siderophore which facilitate acquisition of iron that was
previously biologically unavailable (Wilhelm et al., 1996).
In all samples, additions of Mo, Mn, vitamins and EDTA showed no effect on IVF.
As in situ concentration levels were not measured, it is unknown whether this absence of
effect is related to an adaptation of the phytoplankton community to low concentrations
of these micronutrients and chelators or whether the surrounding corals provide their
regular supply.
Measurements were conducted at the end of the dry season (November) and at the end
of the rainy season (March). There were no changes in the most significant limiting
nutrient with the season: in both series of experiments, N was the primary limiting
nutrient, either alone or together with P, with the notable exception of Reka Reka in
March (Tables 5 and 6).
Reka Reka, the only P limited lagoon, is the smallest and the shallowest of the ten
lagoons studied (Table 1). With the highest ratio of immersed surface area (soft sediment
and hard substrates) to water volume, this lagoon has probably the highest N 2 fixation
and export from immersed surfaces to lagoon water (Larkum et al., 1988; D’Elia and
Wiebe, 1990; Charpy–Roubaud et al., 1997). Moreover, phosphorus limitation is also
expected to increase in lagoons with low advective ocean waters which appear rather
nitrogen depleted as seen above. Such a gradient in phosphorus concentration was
reported in profiles from ocean to lagoon in Canton lagoon (Smith and Jokiel, 1978).
The two other small lagoons of our selection, Tekokota and Tepoto sud (Table 1), also
have a high ratio of immersed surface area to water volume. Accordingly, we should
expect high N fluxes from the surface to the water column. However, these two atolls
have medium or large lagoon openings (Table 1). Therefore, dilution of lagoon water by
the influx of low N ocean water through these openings is likely the reason that these
lagoonal waters maintained N deficiency of phytoplankton, as do most of the atoll
lagoons of the Tuamotu. A shift to P limitation needs both low oceanic fluxes and large
immersed surfaces with actively N 2 fixing organisms. Such shifts on both sides from the
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
163
N / P optimum ratio may explain why some lagoons and other coral reef systems are N
limited (e.g. Kimmerer and Walsh, 1981; Smith et al., 1984; Laws and Allen, 1996;
Larned, 1997) while others are P limited (e.g. Smith, 1984; Entsch et al., 1983a; Littler
et al., 1991).
Multiple nutrient effects were the rule in our experiments. Nitrogen, phosphorus, and
sometimes silica effects were significant. In addition, in all samples but one the NP
effect was significant and in seven samples, the NSi, PSi or NPSi effects were significant
(Tables 5 and 6). Simultaneous nutrient control of phytoplankton has been shown in the
subtropical gyre of the Central North Pacific (Eppley et al., 1973) and in other systems
(Dufour et al., 1981; Elser et al., 1990; Kivi et al., 1993). Simultaneous limitations
appear to contradict Liebig’s law of a single limiting nutrient. It is possible that the
pooled response of a community consisting of algal species with different nutritional
requirements showed combined limitations. The response of the community should
therefore be additive, i.e. equal to the sum of the responses to single nutrient addition. In
contrast, we noticed a synergistic response to N1P additions and less frequently to
N1Si, P1Si and N1P1Si additions. In the case of a synergistic response with Si, it is
likely that Si enrichment channelled N and P enrichment into diatoms and therefore
changed phytoplanktonic species composition. These observations are consistent with a
view that both N and P are in relatively short supply, such that enrichment by one
nutrient without the other produces only a brief growth enhancement until depletion of
the other nutrient occurs.
5. Conclusion
As pointed out by Elser and Kimmel (1986), ‘‘the results of enrichment bioassays
reflect algal growth responses to increased nutrient availability under the specific
conditions of the experimental incubations and therefore, can only indicate a potential
for in situ nutrient limitation of algal growth in the absence of other limiting factors’’.
From the enrichment bioassays processed in this study, it may be concluded that both
N and P are in relatively short supply in the lagoons and surrounding ocean waters of
Tuamotu archipelago. However nitrogen appears to be the main limiting nutrient in
ocean water and most atoll lagoons and we predict an increasing role for phosphorus in
the smallest and most confined lagoons, as seen in Reka Reka. Si addition may increase
phytoplankton standing crop when N and P requirements are initially satisfied. Natural
concentrations of vitamins, Mo, Mn, Fe and chelators seem to be sufficient to support
increases of phytoplankton standing crop in the lagoons and surface ocean waters of
Tuamotu.
Phytoplankton communities in the lagoons of Tuamotu appear to respond quickly, and
for some intensely, to increases in the levels of combined N and P. The phytoplanktonic
blooms that have been observed on several occasions in the last 30 years are probably
related to these increases. Si could be responsible for shaping the phytoplanktonic crop
away from the picophytoplankton which is scarcely retained by shells and fishes. In the
management of the atoll lagoons of the Tuamotu N, P and Si input should be considered.
The factorial enrichment bioassays using the above protocol allowed us to test the
164
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
effect of 10 macro and micronutrient additions and their interactions on potential
phytoplankton standing crop. To study such a variety would not have been possible with
another method. However, due to the limitations of this enrichment approach, conclusions of a further article in prep., using biochemical and in situ approaches, will be
necessary for a more comprehensive understanding of phytoplankton control in the
Tuamotu atoll lagoons and the surrounding ocean.
Acknowledgements
This research was supported by the Institut francais de recherches
´
ment en cooperation
(ORSTOM) and the Programme National
(PNRCO). We would like to thank the crew of the R.V. Alis for
inhabitants of the Tuamotu atolls for their hospitality. We thank
providing data on atoll morphology (Table 1).
pour le developpe´
Recifs
Coralliens
their help and the
´ ¨ for
S. Andrefouet
References
Atkinson, M.J., Smith, S.V., 1983. C:N:P ratios of benthic marine plants. Limnol. Oceanogr. 28 (3), 568–574.
Boyd, P.W., Muggli, D.L., Varela, D.E., Golblatt, R.H., Chretien, R., Orians, K.J., Harrison, P.J., 1996. In vitro
iron enrichments in the NE subartic Pacific. Mar. Ecol. Prog. Ser. 136, 179–193.
Carlucci, A.F., Silbernagel, S.B., MacNally, P.M., 1969. Influence of temperature and other radiation on
persistance of B 12 , thiamine and biotin on seawater J. Phycol. 5, 302–305.
Charpy, L., Blanchot, J., 1998. Photosynthetic picoplankton in French Polynesia atoll lagoons: estimation of
taxa contribution to biomass and production by flow cytometry. Mar. Ecol. Prog. Ser. 162, 57–70.
Charpy, L., Dufour, P., Garcia, N., 1997. Particulate organic matter in sixteen Tuamotu atoll lagoons (French
Polynesia). Mar. Ecol. Prog. Ser. 151, 55–65.
Charpy–Roubaud, C., Charpy, L., Larkum, A.W., 1997. Contribution to N fixation to N primary productivity of
the lagoon of Tikehau. In: Lessios, H.A. (Ed.), Proceed. 8th Int. Coral Reefs Symp., Panama. Allen Press,
New York, pp. 803–808.
Charpy–Roubaud, C., Charpy, L., Sarazin, G., 1996. Diffusional nutrient fluxes at the sediment–water
interface and organic matter mineralization in an atoll lagoon (Tikehau, Tuamotu archipelago French
Polynesia). . Mar. Ecol. Prog. Ser. 132, 181–190.
Cochran, W.G., Cox, G., 1957. Experimental design. Ed. J. Wiley and Sons, New York, 617 p.
Cotner, J.B., Wetzel, R.G., 1992. Uptake of dissolved inorganic and organic phosphorus compounds by
phytoplankton and bacterioplankton. Limnol. Oceanog. 37 (2), 232–243.
Cullen, J.J., 1991. Hypotheses to explain high-nutrient conditions in the open sea. Limnol. Oceanogr. 36 (8),
1578–1599.
Daniel, C., 1959. Use of half-normal plots in interpreting factorial two-level experiments. Technometrics 1 (4),
311–341.
D’Elia, C.F., Wiebe, W.J., 1990. Biochemical nutrient cycles in coral-reef ecosystems. In: Dubinsky, Z. (Ed.),
Coral Reefs. Elsevier, pp. 49–74.
ˆ nutritif de la biomasse du seston dans une lagune
Dufour, P., Lemasson, L., Cremoux, J.-L., 1981. Controle
ˆ d’Ivoire. II. Variations geographiques
´
`
tropicale de la Cote
et saisonnieres.
J. Exp. Mar. Biol. Ecol. 51,
269–284.
Dufour, P., Harmelin–Vivien, M., 1997. A research program for a typology of atoll lagoons: strategy and first
results. In: Lessios, H.A. (Ed.), Proceed. 8th Int. Coral Reefs Symp., Panama. Allen Press, New York, Vol.
1, pp. 843–848.
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
165
Elser, J.J., Kimmel, B.L., 1986. Alteration of phytoplankton phosphorus status during enrichment experiments:
implications for interpreting nutrient enrichment bioassay results. Hydrobiologia 133, 217–222.
Elser, J.J., Marzolf, E.R., Goldman, C.R., 1990. Phosphorus and nitrogen limitation of phytoplankton growth
in the freshwaters of North America: a review and critique of experimental enrichments. Can. J. Fish.
Aquat. Sci. 47, 1468–1477.
Entsch, B., Boto, K.G., Sim, R.G., Wellington, J.T., 1983. Phosphorus and nitrogen in coral reef sediments.
Limnol. Oceanogr. 28 (3), 465–476.
Entsch, B., Sim, R.G., Hatcher, B.G., 1983. Indications from photosynthetic components that iron is a limiting
nutrient in primary producers on coral reefs. Marine Biology 73, 1–5.
Eppley, R.W., Venrick, E.L., Mullin, M.M., 1973. A study of plankton dynamics and nutrient cycling in the
central gyre of the North Pacific Ocean. Limnol. Oceanogr. 18, 534–551.
Geider, R.J., MacIntyre, H.L., Kana, T.M., 1997. Dynamic model of phytoplankton growth and acclimation:
responses of the balanced growth rate and the chlorophyll a:carbon ratio to light, nutrient-limitation and
temperature. Mar. Ecol. Prog. Ser. 148, 187–200.
´
Harris, P., Fichez, R., 1995. Observations et mecanismes
de la crise dystrophique de 1994 dans le lagon de
´ Française). Oceanographie,
´
l’atoll d’Hikueru (Archipel des Tuamotu, Polynesie
Notes et documents, Vol. 45,
ORSTOM Tahiti ed., 25 pp.
Howarth, R.W., Cole, J.J., 1985. Molybdenum availability, nitrogen limitation, and phytoplankton growth in
natural waters. Science 229, 653–655.
Jeffrey, S.W., 1997. Application of pigment methods to oceanography. In: Jeffrey, S.W., Mantoura, R.F.C.,
Wright, S.W. (Eds.), Phytoplankton pigments in oceanography: guidelines to modern methods. Unesco, pp.
127–166.
Karl, D.M., Tien, G., Dore, J., Winn, C.D., 1993. Total dissolved nitrogen and phosphorus concentrations at
US-JGOFS Station ALOHA: Redfield reconciliation. Marine Chemistry 41, 203–208.
Kiefer, D.A., 1973. Chlorophyll a fluorescence in marine centric diatoms: response of chloroplasts to light and
nutrient stress. Mar. Biol. 23, 39–43.
Kimmerer, W.J., Walsh, T.W., 1981. Tarawa atoll lagoon: circulation, nutrient fluxes, and the impact of human
waste. Micronesia 17, 161–179.
Kivi, K., Kaitala, S., Kuosa, H., Kuparinen, J., Leskinen, E., Lignell, R., Marcussen, B., Tamminen, T., 1993.
Nutrient limitation and grazing control of the Baltic plankton community during annual succession. Limnol.
Oceanogr. 38 (5), 893–905.
Landry, M.R., Kirshtein, J., Constantinou, J., 1996. Abundances and distributions of picoplankton populations
in the central equatorial Pacific from 128N to 128S, 140W. Deep-Sea Res. II 43, 871–890.
Larkum, A.W.D., Kennedy, I.R., Muller, W.J., 1988. Nitrogen fixation on a coral reef. Mar. Biol. 98, 143–155.
Larned, S.T., 1997. Nitrogen versus phosphorus limited growth of coral reef macroalgae from Kane’Ohe Bay,
O’Ahu. In: Lessios, H.A. (Ed.), Proceed. 8th Int. Coral Reefs Symp., abstracts. Allen Press, New York, p.
114.
Laws, E.A., Allen, C.B., 1996. Water quality in a subtropical embayment more than a decade after diversion of
sewage discharges. Pac. Sci. 50 (2), 194–210.
Laws, E.A., Redalje, D., 1979. Effects of sewage enrichment on the phytoplankton population of a subtropical
estuary. Pac. Sci. 33, 129–144.
Lindley, S.T., Bidigare, R.R., Barber, R.T., 1995. Phytoplankton photosynthesis parameters along 1408W in
the equatorial Pacific. Deep-Sea Res. II 42, 441–463.
Littler, M.M., Littler, D.S., Titlyanov, E.A., 1991. Comparisons of N- and P-limited productivity between high
granitic islands versus low carbonate atolls in the Seychelles archipelago: a test of relative-dominance
paradigm. Coral Reefs 10, 199–209.
Lorenzen, C.J., 1966. A method for the continuous measurement of in vivo chlorophyll concentration. Deep
Sea Res. 13, 223–227.
Maestrini, S.Y., Bonin, D.J., Droop, M.R., 1984. Phytoplankton as indicators of seawater quality: bioassay
approaches and protocols. In: Shubert, L.E. (Ed.), Algae as ecological indicators. Academic Press Inc.,
London, pp. 71–132.
Martin, J.H., Coale, K.H., Johnson, K.S., Fitzwater, S.E., Gordon, R.M., Tanner, S.J., Hunter, C.N., Elrod,
V.A., Nowicki, J.L., Coley, T.L., Barber, R.T., Lindley, S., Watson, A.J., Van–Scoy, K., Law, C.S., 1994.
Testing the iron hypothesis in ecosystems of the equatorial Pacific ocean. Nature 371 (6493), 123–129.
166
P. Dufour, B. Berland / J. Exp. Mar. Biol. Ecol. 234 (1999) 147 – 166
Patriquin, D.G., 1972. The origin of nitrogen and phosphorus for the growth of marine angiosperms. Mar. Biol.
15, 35–46.
Pujo–Pay, M., Conan, P., Raimbault, P., 1997. Excretion of dissolved organic nitrogen by phytoplankton
assessed by wet oxidation and 15 N tracer procedures. Mar. Ecol. Prog. Ser. 153, 93–111.
Raimbault, P., Slawyk, G., Boudjellal, B., Coatanoan, C., Conan, P., Coste, B., Garcia, N., Moutin, T., Pujo–
Pay, M., 1998. Biomass, new production and export in the equatorial Pacific at 1508 W: evidence for
intense nitrogen recycling. J. Geophys. Res. (In press).
´
´ Française).
Rancher, J., Rougerie, F., 1994. L’environnement oceanique
de l’archipel des Tuamotu (Polynesie
Oceanologica Acta 18 (1), 43–60.
Redfield, A.C., Ketchum, B.H., Richards, F.A., 1963. The influence of organisms on the composition of sea
water. In: Hill, M.N. (Ed.), The sea. Wiley, Vol. 2, pp. 26–77.
Smith, S.V., 1984. Phosphorus versus nitrogen limitation in the marine environment. Limnol. Oceanogr. 29 (6),
1149–1160.
Smith, S.V., Jokiel, P.L., 1978. Water composition and biochemical gradients in the Canton atoll lagoon. Atoll
Res. Bull. 221, 15–53.
Smith, S.V., Chandra, S., Kwitko, L., Schneider, R.C., Schoonmaker, J., Seeto, J., Tebano, T., Tribble, G.W.,
1984. Chemical stoichiometry of lagoonal metabolism. Preliminary report on an environmental chemistry
survey of Christmas Island, Kiribati. Sea Grant Cooperative Report, UNIHI-SEAGRANT-CR-84-02, 30 pp.
Sorokin, Y.I., 1995. Coral Reef Ecology. Springer, 2nd printing, 465 pp.
` en mer et ses relations avec la production
Sournia, A., Citeau, J., 1972. Sur la distribution du molybdene
primaire. C.R. Acad. Sci. Paris 275, 1299–1302.
Tilman, D., Kilham, S.S., Kilham, P., 1982. Phytoplankton community ecology, the role of limiting resources.
Ann. Rev. Ecol. Syst. 13, 349–372.
´
Torreton,
J.-P., Talbot, V., Garcia, N., 1998. Nutrient stimulation of bacterioplankton growth in Tuamotu atoll
lagoons. Supplement to Eos, Transactions, American Geophysical Union 79 (1), 154.
´
Vigier, M., 1988. Pratique des plans d’experiences.
Les Editions d’Organisation, Paris, 191 p.
Wells, M.L., Price, N.M., Bruland, K.W., 1994. Iron limitation and the cyanobacterium Synechococcus in
equatorial Pacific waters. Limnol. Oceanogr. 39, 1481–1486.
Wilhelm, S.W., Maxwell, D.P., Trick, C.G., 1996. Growth, iron requirements and siderophore production in
iron-limited Synechococcus PCC 7002. Limnol. Oceanogr. 41, 89–97.