Download PDF - Oxford Academic - Oxford University Press

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Biogeography wikipedia , lookup

Transcript
250783
13/6/03
3:34 pm
Page 783
JOURNAL OF PLANKTON RESEARCH
VOLUME

NUMBER

PAGES
‒

Vertical microscale patchiness in
nano- and microplankton distributions
in a stratified estuary
LONE THYBO MOURITSEN AND KATHERINE RICHARDSON*
DEPARTMENT OF MARINE ECOLOGY, BIOLOGICAL INSTITUTE, AARHUS UNIVERSITY, FINLANDSGADE
, DK- AARHUS N, DENMARK
*CORRESPONDING AUTHOR: [email protected]
Microscale (decimetre) vertical heterogeneity in the distribution of nano- and microplankton was
studied on August 23, 1999 at two different sites (separated by ~40 m) in Aarhus Bay, Denmark.
At the time of sampling, the water column was stratified with respect to both temperature and salinity
and a subsurface fluorescence maximum (corresponding to ~10 µg chlorophyll a l–1) occurred
immediately below the primary pycnocline (9 m). Samples for species identification were taken at the
surface, near the bottom of the water column and at 20 15-cm intervals in and around the depth of
maximum fluorescence. The plankton communities recorded in the three different regions of the water
column differed dramatically from one another. In addition, significant differences were found in the
distribution patterns of species and functional groups in the region of sampling around the fluorescence peak. The same patterns in vertical species distributions were observed at the two stations. In
the region surrounding the subsurface fluorescence peak, diatoms were, generally, regularly distributed,
although the total diatom biomass decreased slightly with depth. Dinoflagellate species were mainly
non-regularly distributed and could be divided into two groups: (i) autotrophic or potentially
mixotrophic species (Dinophysis norvegica, Dinophysis acuminata, Dinophysis cf. dens, Prorocentrum micans, Gymnodinium chlorophorum and Ceratium macroceros) that mainly decreased in
numbers with depth or were aggregated in their distribution; and (ii) heterotrophic or potentially
mixotrophic species (Ceratium lineatum, Ceratium longipes, Ceratium furca, Ceratium tripos,
Ceratium fusus, Protoperidinium curtipes, Protoperidinium steinii, Diplopsalis spp. and Katodinium
glaucum). In the latter group, the species mainly increased in numbers with depth or were randomly
distributed. Most ciliates were uniformly distributed vertically in the water column. However, small
cells of the photosynthesizing ciliate Mesodinium rubrum were most abundant at the depth of
maximal fluorescence while large M. rubrum cells were equally abundant at all the depths sampled,
suggesting the two size groups of this organism may differ ecologically. Overall, the study demonstrates that traditional plankton sampling methods may lead to misinterpretations of species
co-existence and interactions.
I N T RO D U C T I O N
The patchy distribution in space of plankton organisms
under natural conditions is well documented. However,
most studies examining this phenomenon have dealt with
large scale (i.e. hundreds of metres to kilometres) patchiness in the horizontal plane. Horizontal heterogeneity in
plankton distributions is also known to occur at small (i.e.
millimetre–centimetre) scales but such patchiness is less
well studied and the potential causes of this patchiness less
well understood than that occurring at larger scales.
Plankton patchiness in the vertical plane is less well
studied than that in the horizontal. The most well documented examples of vertical heterogeneity in plankton
distributions are the occurrence of subsurface chlorophyll
peaks. Such peaks are, most often, identified through
fluorescence profiling and are generally considered over
scales of metres. Water sampling for the purpose of determining species abundances over the water column are
often taken at selected depths but such samples are seldom
taken at depth intervals of less than several (often up to
tens of ) metres. Thus, although vertical heterogeneity in
Journal of Plankton Research 25(7), © Oxford University Press; all rights reserved
250783
13/6/03
3:34 pm
Page 784
JOURNAL OF PLANKTON RESEARCH
VOLUME
phytoplankton distributions is recognized, knowledge
concerning the microscale vertical distribution of
plankton species is very limited.
Regardless of the scale of patchiness one is considering
in the ocean, it can come about through either physical
(giving rise to heterogeneity in the physical environment)
or biological [i.e. migration, sinking, grazing, predation or
other more subtle species interactions (Stoecker et al.,
1984; Tokarev et al., 1999)] processes or a combination of
these two types of process. The importance of biological
processes in leading to plankton patchiness is, however,
predicted to increase inversely with scale (Pinel-Alloul,
1995). Thus, it is at the microscale where one could
predict an important role for biological processes in
causing heterogeneity in plankton distributions. Documenting the occurrence of such interactions remains,
however, elusive.
A few field data sets have provided material on small
scale plankton distributions from which individual
species’ behaviour or species/community interactions
leading to plankton patchiness can be inferred. For
example, Bjørnsen and Nielsen (Bjørnsen and Nielsen,
1991) found that high densities of toxic dinoflagellates
were associated with a decreased abundance of microzooplankton and suggested a negative interaction
between these two plankton groups. In another study,
Stoecker et al. (Stoecker et al., 1984) found a positive correlation between the distributions of ciliates and their
prey and suggested positive species interactions between
these groups.
The purpose of the present study was to describe the
microscale (decimetre) vertical distributions of phyto- and
microzooplankton both at the species and community
levels in a stratified estuary. The observed patterns in
species distributions are confirmed by sampling at two
different stations in close proximity to one another. The
study is, in its nature, descriptive. However, we combine
our observations on species distributions with previously
published information concerning the autecology of the
species recorded in order to propose hypotheses to explain
how species behaviour or species/community interactions
may have influenced the observed patchiness in species
and community distributions.
METHOD
Sampling site
Sampling took place in Aarhus Bay, a relatively sheltered
estuary situated in the south-western part of the Kattegat,
at 56°0910N, 10°1920E, depth 16 m (Figure 1). The
hydrographic conditions in the Kattegat are mainly determined by the outflow of low salinity Baltic water at the


NUMBER

PAGES
‒

Fig. 1. Map of the study area. Arrow denotes location of Aarhus Bay,
where the sampling site is located.
surface and the inflow of high salinity North Sea and
Skagerrak water at the bottom, usually separated by a
stable halocline. The vertical position of the halocline is
strongly influenced by meteorological conditions such as
wind speed and direction as well as atmospheric pressure,
which cause the stability of the halocline to vary
(Rasmussen, 1997). Moreover, heating of the surface layer
by increased irradiance during the summer months reinforces stratification.
Sampling
Sampling was conducted on August 23, 1999 between
11.00 and 11.30 a.m. using a high resolution sampler
(HRS) as described in Bjørnsen and Nielsen (Bjørnsen
and Nielsen, 1991). The HRS was positioned so that the
middle was placed at the depth of maximal fluorescence
(Figure 2). Two HRS deployments were made in the same
depth interval at stations separated by a horizontal
distance of ~40 m (stations S1 and S2). Prior to HRS
sampling, continuous profiles of salinity, temperature,
density and fluorescence were recorded in 20-cm intervals
in order to determine the depth of the fluorescence peak.
CTD measurements were made using an Arop 2000
PowCom and fluorescence measurements of chlorophyll
were made with a BackScat fluorometer. Light attenuation in the water column was measured using a Li-Cor
quantameter.
The HRS collects 20 duplicate water samples (each
13/6/03
3:34 pm
Page 785
L. T. MOURITSEN AND K. RICHARDSON
VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES
microscope (Axiovert 135M, 100, 200, 400).
Encountered cells were identified to the lowest possible
taxonomic level. When appropriate, cells were assigned to
size classes. Nanoflagellates (2–20 µm) were only counted
in every second sample (separated by 30 cm). Cells smaller
than 2 µm were not counted due to the difficulties of
enumerating this size class using the Utermöhl method
(Montagnes et al., 1999). Abundance, biovolumes and cell
carbon content were calculated using the program
Algesys (Bio/consult A/S). Biovolume was calculated by
using geometric equations for the shape of the algae
(Olrik, 1991). Calculation of cell carbon content followed
Edler (Edler, 1979). Only healthy looking cells containing
chloroplasts (where appropriate) were enumerated.
0
2
Temperature
Salinity
4
Fluorescence
6
Depth (m)
250783
8
10
12
14
16
0
5
10
Temperature (¼C),
15
20
Salinity (ä),
25
30
35
Fluorescence (µg chl/l)
Fig. 2. Vertical distribution of temperature, salinity and chlorophyll
derived from fluorescence at station S1 on 23 August, 1999. The shaded
regions show sampling depths. Twenty depths were sampled (duplicates
collected at each depth) using the high resolution sampler (HRS)
between 8 and 11 m. Duplicate samples were taken at the surface and
in bottom waters.
150 ml) almost simultaneously (<1 s) at 15-cm intervals.
From each depth, the two water samples were pooled and
a subsample of 50 ml was filtered through a Whatman
GF/C filter and frozen for later analysis of nutrient
content. The filters were individually placed in 10 ml of
90% acetone for extraction of chlorophyll a. The remaining water was preserved in 1% acidic Lugol’s solution for
quantification of phyto- and microzooplankton. In
addition to the HRS samples, samples from the surface
water (depth = 1 m) and the bottom water (depth = 15 m)
were collected and treated as described above. At the time
of sampling, wind velocity was low (4–6 m s–1) and there
were scattered clouds.
Sample analysis
Chemical analysis
Chlorophyll a concentrations were measured fluorometrically on acetone extracts using a Turner fluorometer,
according to Yentsch and Menzel (Yentsch and Menzel,
1963). Analysis of silicate, phosphorus and nitrate
concentrations was carried out automatically, following
Grasshoff et al. (Grasshoff et al., 1983), at the Danish Institute for Fisheries Research.
Taxonomic identification, enumeration and calculation of cell
carbon content
All diatoms, flagellates and ciliates were quantified by
settling 25 or 50 ml of water from each sample for 48 h
in plankton sedimentation chambers (Hydrobios). Cells
larger than 2 µm were counted following the Utermöhl
technique (Lund et al., 1958) using an inverted Zeiss
Data analysis
Data reduction
Various methods have been applied to determine how
many cells to count in order to obtain satisfactory
precision on cell concentration estimates [e.g. (Allen,
1921; Venrick, 1978)]. In this respect, different authors
have reached rather different conclusions, probably as a
consequence of, for example, sampling design, the applied
counting method and the context in which the data have
been used, and/or different authors’ views of ‘satisfactory
precision’. Thus, in order to distinguish methodological
inaccuracy from real biological patchiness in this study, a
separate method, tailored to fit the obtained data and the
purpose of sampling, was developed.
Species-specific coefficients of variance (CV = variance
divided by mean) of cell concentrations from S1 (the 20
samples used as replicates) as a function of mean cell
count were used to identify the minimum number of cells
counted necessary to stabilize CV. Rejecting the possibility that rare species are inherently more patchily
distributed with respect to environmental variables than
abundant species, a relatively high CV among the rare
species suggests the influence of methodology (i.e. that a
few cells by chance appear in one sample whereas none
are found in the subsequent sample) rather than the
occurrence of real patchiness. This analysis showed that
the CV became relatively stable at mean cell counts per
sample of around five. One species present in very high
numbers in only a few samples (Kinetoplastida sp.) and
chain-forming species (Chaetoceros spp. and Skeletonema
costatum), which are aggregated in the sedimentation
chamber by nature (Holmes and Widrig, 1956), were
omitted from the analysis.
A comparison of samples from stations S1 and S2
further supported the conclusion that cell counts above
five can be considered reliable in the present context. The
samples taken in these two HRS deployments demonstrate a remarkable concordance regarding the vertical

250783
13/6/03
3:34 pm
Page 786
JOURNAL OF PLANKTON RESEARCH
VOLUME
distribution of the different species (see Results). Among
species with mean cell counts above five, 48% (n = 36)
(63% if chain-forming species are omitted) showed a
significant species-specific positive correlation (based on
Pearson’s correlation coefficients), while only 11% (n = 42)
of species with counts of less than five were significantly
correlated.
In addition, confidence limits, as described in Ricker
(Ricker, 1937), were applied to species concentrations
recorded in samples from S1 and S2 (upper limit = x +
2.42 + 1.96√[x + 1.5 ]; lower limit = x + 2.42 – 1.96√[x
+ 0.5]; x = concentration; 95% confidence limits). A
number between the concentration limits for each species
at each depth was chosen at random and Pearson’s correlation coefficients were calculated. This was done to
examine the reproducibility of the number of significant
positive correlations found between S1 and S2 applying
rather wide confidence limits to the concentration estimates. For species with a high mean number of cells
counted (n > 5), the ‘new’ concentrations gave rise to 33%
significant positive correlations, which is not statistically
different from the original 48% (2 = 0.58, d.f. = 1, P =
0.45). This result indicates that the correspondence of
species-specific distribution patterns between S1 and S2 is
preserved, even when broad confidence limits are applied
to cell concentrations.
Based on the above analyses, a mean number of five
cells counted was chosen as the reduction parameter,
leaving out 36 of the 80 species counted. The following
analyses are based on the reduced data set unless otherwise stated.
Multivariate analysis
Multivariate analyses were performed on biomass data
(square root transformed) using the program Primer 4.0
(Clarke, 1993). This was done in order to evaluate possible
clustering of samples within and between HRS samples.
Using non-metric multidimensional scaling (MDS), which
generates a geometric configuration of distances between
sample points, derived from a matrix of similarities
(Bray–Curtis) between points, two-dimensional scatter
plots of the samples were constructed. No fundamental
differences were found between MDS plots based on
reduced versus unreduced and transformed (fourth and
second root) versus non-transformed data, indicating that
observed differences between samples are mainly due to
the more abundant species.
Correlations
Spearman’s coefficient of rank correlation (rs) was calculated between species and between nutrients/
salinity/temperature and species in each of the HRS
samples. The non-parametric rs is used because a linear


NUMBER

PAGES
‒

relationship between these parameters cannot be
assumed. Species-specific Pearson’s coefficients of correlation (rp) were also calculated between samples from S1
and S2. rp, which assumes a linear relationship between
species concentrations, was applied instead of rs in order
to emphasize the quantitative relationships between cell
concentrations in the spatially separated samples.
Assignment to functional groups
On the basis of taxonomic and morphological characteristics and a survey of available literature concerning their
autecology, the species included in the statistical analyses
were assigned to various groups according to (i) taxonomy
(diatoms, dinoflagellates, ciliates, other), (ii) size (small,
<20 µm; medium, 20–50 µm; large, >50 µm), (iii) mobility
(mobile or non-mobile) and (iv) nutritional form (autotrophy, heterotroph or mixotroph). All diatoms were
classified as immobile, regardless of reports of migrating
diatom mats (Villareal et al., 1993). The classification of
species into nutritional groups was the most difficult as the
trophic affiliations of several species are currently in
question [e.g. (Hansen, 1991a,b; Jacobson and Andersen,
1994)]. The assignments of species to the various groups
are summarized in Table I.
Distributional patterns of species and groups
The vertical distributions of species and groups within the
samples taken with the HRS were analysed in different
ways. Each species was assigned to a main distribution
category: ‘regular’, ‘random’ or ‘aggregated’. Note that
the statistical description ‘random’ does not necessarily
imply that the distribution of the species in question is
incidental or that it is unexplainable but, rather, that it is
irregular. The assignment was made by calculating the
variance to mean ratio, or index of dispersion (I ) of abundance data using each depth within the HRS as a replicate. I will approximate to unity if there is agreement with
a Poisson distribution (that is if data are randomly distributed). The departure of I from unity is assessed by reference to a 2 table, using the approximation 2 = I(n – 1),
where n = number of samples and d.f. = n – 1 [see (Elliott,
1977)]. Species that increased or decreased in number
with depth were identified by correlating cell concentration and depth using Spearman’s coefficient of correlation.
Differences and similarities between distribution
patterns of species and groups were analysed using
Kendall’s coefficient of concordance (W ) (Sokal and
Rohlf, 1995), measuring the degree of association among
a number of variables transformed to ranks. W varies
between 0 and 1, with W = 1, meaning that there is
perfect agreement between the distributions tested, resulting in a significant 2 value (P < 0.05). For each species,
250783
13/6/03
3:34 pm
Page 787
L. T. MOURITSEN AND K. RICHARDSON
VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES
Table I: Feeding form, size and mobility assigned to the different species
Dinoflagellates
Diatoms
Prorocentrum micans
(Mi, Me, Mo)
Chaetoceros compressus
(A, S, Im)
Dinophysis acuminata
(Mi, Me, Mo)
C. curvisetus
(A, S, Im)
D. cf. dens
(Mi, Me, Mo)
C. socialis/radians
(A, S, Im)
D. norvegica
(Mi, Me, Mo)
C. subtilis
(A, S, Im)
Ceratium furca
(Mi, L, Mo)
Guinardia flaccida
(A, L, Im)
C. fusus
(Mi, L, Mo)
Leptocylindrus danicus
(A, S, Im)
C. lineatum
(Mi, L, Mo)
Guinardia delicatula
(A, S, Im)
C. longipes
(Mi, L, Mo)
Rhizosolenia fragilissima
(A, S, Im)
C. macroceros
(Mi, L, Mo)
R. pungens
(A, L, Im)
C. tripos
(Mi, L, Mo)
Skeletonema costatum
(A, S, Im)
Gymnodinium chlorophorum
(Mi, S, Mo)
T. punctigera
(A, L, Im)
Katodinium glaucum
(H, S, Mo)
Pseudonitzschia sp.
(A, Me, Im)
Protoperidinium curtipes
(H, L, Mo)
Nitzschia closterium/longissima
(A, S, Im)
P. steinii
(H, Me, Mo)
Thalassionema nitzschioides
(A, S, Im)
Diplopsalis spp.
(H, Me, Mo)
Athecate dinoflagellate 2–10 µm
(Mi, S, Mo)
Athecate dinoflagellate 10–20 µm
(Mi, S, Mo)
Thecate dinoflagellate 2–10 µm
(Mi, S, Mo)
Thecate dinoflagellate 10–20 µm
(Mi, S, Mo)
Ciliates
Others
Salpingella sp.
(H, Me, Mo)
Dichtyocha speculum
(A, Me, Mo)
Lohmaniella oviformis
(H, Me, Mo)
Ebria triparthita
(H, Me, Mo)
Strombidium spp. 20–30 µm
(Mi, Me, Mo)
Kinetoplastida
(Mi, S, Mo)
Mesodinium rubrum <25 µm
(Mi, S, Mo)
Pyramimonas/Tetraselmis
(A, S, Mo)
M. rubrum >25 µm
(Mi, Me, Mo)
Prymnesiophyceae spp.
(A, S, Mo)
Cryptophyceae spp. 14–20 µm
(A, S, Mo)
A, autotrophic; H, heterotrophic; Mi, mixotrophic; S, small <20 µm; Me, medium 20–50 µm; L, large >50 µm; Mo, mobile; Im, immobile.
the relative abundance in each depth (relative frequency
of the total number of cells in the given HRS sample) was
calculated in order to standardize species distributions.
This made it possible to compare species independently
of differences in abundance/biomass.
Comparisons of vertical distributions of different
groups were made by choosing the median frequency of
the species in the group (i.e. mobile) in each depth and
comparing that with the median frequency of the species
in the other group (i.e. immobile). Median frequency was
chosen, instead of mean frequency, in order to give
weight to typical distributions, thereby minimizing the
impact of rare outliers. In the analysis of the agreement
of distribution of species with known or potential
predator–prey relationships, potential prey organisms
were identified from published literature [e.g. (Hansen,
1991; Nielsen and Hansen, 1999; Olli, 1999; Smalley
et al., 1999)].
The subsurface fluorescence peak sampled consisted of
several minor peaks, all of which were associated with
spikes in salinity and temperature. Differences in species
composition between the three largest peaks were tested
performing Kendall’s coefficient of concordance on the
sequence of the 10 most abundant species in terms of
biomass. The relative distributions of autotrophic and
heterotrophic species in each of these three peaks were
examined by comparing the mean of the depth of
maximal occurrence in the intervals where the peaks were
positioned. Differences were tested for significance using
a Student’s t-test.

13/6/03
3:34 pm
Page 788
JOURNAL OF PLANKTON RESEARCH

VOLUME
Statistical analysis
NUMBER

PAGES
‒

0
Whenever parametric tests were applied, assumptions
of normal distribution and homogeneity of variance
were shown to be met using appropriate tests
(Kolmogorov–Smirnov and Levenes). Statistical analyses
were carried out using the Statistical Package of Social
Science (SPSS) unless otherwise stated.
8
P1
Depth (m)
250783
P2
10
P3
R E S U LT S
Hydrographic conditions
12
In late August, a primary pycnocline was situated between
5 and 8 m depth. Salinity increased from 19‰ in the
surface water to 29‰ at the bottom, while temperatures
decreased from 17 to 12°C. Hydrographic conditions
were similar at the two stations. Data from station S1 are
shown in Figure 2. Salinity and temperature discontinuities usually coincided. Concentrations of nutrients in
the depth interval sampled by the HRS were relatively
low. The concentrations of phosphate and silicate
increased slightly with depth (P: rs = 0.54, n = 20, P =
0.013; Si: rs = 0.61, n = 20, P = 0.004), while the concentration of nitrate was characterized by minor peaks
throughout the water column (data not shown). Nutrient
concentrations were consistently low in the surface
water but higher in the bottom water (phosphate,
surface/bottom concentrations: not detectable/1 µmol
l–1; nitrate + nitrite: 0.85/0.71 µmol l–1; ammonium:
0.35/4.93 µmol l–1). A fluorescence maximum (corresponding to ~10 µg chlorophyll a 1–1) was measured
beneath the primary pycnocline at 9 m depth, corresponding to ~10% of surface photon flux density.
Heterotrophs
Mixotrophs
Autotrophs
16
0
50
100
150
200
250
300
350
µg C/l
Fig. 3. Vertical distributions of autotrophic, heterotrophic and
mixotrophic biomass (µg carbon 1–1) at the depths sampled at the surface
and bottom and with the HRS (station S1). P1, P2 and P3 indicate minor
peaks in the autotrophic plankton distribution.
Vertical heterogeneity
Biomass
The fluorescence profile (Figure 2) gives an indication of
the overall distribution of autotrophic biomass in the
water column and indicates the presence of several
distinct peaks in the biomass distribution. The profile of
biomass of auto- and mixotrophic phytoplankton estimated by microscopic examination (Figure 3) of the
samples in the region of the fluorescence maximum was
qualitatively similar to the fluorescence profile, although
it appeared that the peaks in biomass were displaced vertically upwards by ~0.5 m between the time of the fluorescence profiling and the water collection by the HRS.
Measurements of chlorophyll at the depth interval
sampled by the HRS (data not shown) agreed well with
the microscopically estimated total auto- and mixotrophic
biomass. However, the carbon:chlorophyll ratio decreased
significantly with depth (rp = –0.56, P = 0.01), indicating
that the phytoplankton in the deeper water layers, where

light was less available than at the surface, responded by
increasing their cellular chlorophyll content [e.g. (Beardall
and Morris, 1976; Richardson et al., 1983; Rosen and
Lowe, 1984)].
Species composition
Dinoflagellates dominated the plankton biomass between
8 and 11 m. The contributions of ciliates and diatoms to
the plankton biomass were lower and relatively constant.
A minor increase in mixotrophic biomass was noted in the
deepest HRS samples. The dominant species in the
samples was Gymnodinium chlorophorum (Elbrächter and
Schnepf, 1996), an otherwise colourless dinoflagellate
containing an autotrophic endosymbiont. This species
was present as colonies embedded in a gelatinous matrix
(which is common for this species) and, together with the
dinoflagellates Ceratium tripos, Ceratium macroceros, Ceratium
fusus, Prorocentrum micans, Dinophysis norvegica and Protoperidinium curtipes, and the diatom Rhizosolenia pungens,
comprised ~90% of the total biomass. The dominant
species in both the surface and bottom waters were the
diatom R. pungens and the dinoflagellate C. tripos. Together,
these species comprised 50% of the biomass, while G.
chlorophorum (which comprised ~50% of the biomass in
the HRS samples) only comprised 14 and 3% in surface
and bottom waters, respectively.
A closer investigation of the three biomass peaks
(P1–P3) studied in the region of maximum fluorescence
sampled by the HRS (Figure 3) revealed a similar
sequence of the 10 most dominating species (comprising
250783
13/6/03
3:34 pm
Page 789
L. T. MOURITSEN AND K. RICHARDSON
VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES
90% of the total biomass) in the two uppermost peaks [P1
versus P2: W = 0.96, 29 = 17.2, P = 0.045; NB the
subscript on 2 here and in the following represents the
number of degrees of freedom in the Chi distribution, see
(Sokal and Rohlf, 1995)]. This species sequence was not
similar to that of the lowest peak (P1 versus P3: W = 0.81,
29 = 14.8, P = 0.10; P2 versus P3: W = 0.75, 29 = 13.5,
P = 0.14). All fluorescence and biomass peaks were associated with changes in salinity and temperature. A difference (although non-significant) was found between the
mean vertical positions of autotrophic and heterotrophic
species in each of the three peaks, where, in all cases, the
autotrophic species peaked above the heterotrophic
species.
The distances on the MDS plot between the data points
representing the plankton communities (unreduced data
set) found at the various depths at station S1 indicate a
large degree of similarity in the communities recorded in
the samples taken with the HRS compared with the
communities found at the surface and at the bottom.
Likewise, the distance between the points representing the
surface and bottom communities on this MDS plot indicates the presence of distinct communities at these depths
(Figure 4A). MDS analysis of samples taken with the
HRS at both stations (Figure 4B) shows that the shallowest samples cluster at the right of the plot while the
deepest are concentrated at the left of the plot. In all
cases, samples from the same depths at the two different
stations cluster closer together than samples from the top
and bottom of the HRS. This indicates a high degree of
resemblance for both species composition and biomass at
corresponding depths at the two sites, and that the similarity between communities sampled at a specific depth at
the two stations (~40 m horizontal distance between
stations) was greater than that between communities at
the top and bottom of the 3 m sampling range of the
HRS at an individual station.
Analysis of the species composition showed that a tiny
flagellate, Kinetoplastida sp., was present in very high
numbers (up to 5500 cells ml–1) but only in the deepest
sample taken by the HRS. This observation was made at
both stations. Furthermore, species-specific correlations
between S1 and S2 showed that 48% of the species were
significantly positively correlated (63% if chain-forming
species are left out; rp: P < 0.05). However, none of the
nanoplanktonic species (<20 µm) was significantly correlated between the two stations.
Distribution categories
Tables II and III show the results of the statistical distribution analyses. The departure of the index of dispersion
(I ) from a Poisson distribution divided the species into
three distributional groups (Table II): aggregated (18% of
22
a)
15
10
11
14
13
982456173
18
17
12
19
16
20
21
8
b)
20
16
18
19
38
40
39
36
14
12
6
24
32
26
34 2830
2911 7 25
17
31
35 15 33 9 27
5
3
13
10
23
21
4
22
2
1
Fig. 4. (a) MDS plot of samples taken from all depths at station S1.
Calculations are based on the unreduced, square root transformed data
set. Samples 1–20 were taken using the HRS in and around the fluorescence peak. Sample 21 is the surface sample and 22 the bottom
sample. The plot depicts the uniqueness of the plankton communities
occurring at the surface, bottom and middle of the water column. The
samples taken in the middle of the water column with the HRS clump
together in (a). This clump is expanded in (b): MDS plot of samples
taken using the HRS, i.e. only in and around the fluorescence peak at
station S1 and S2. Calculations are based on the unreduced, square root
transformed data set. Samples from station S1 are numbered 1–20.
Samples from corresponding depths at station S2 are numbered from 21
to 40. Points representing samples from corresponding depths at the two
stations are connected. Sample 37 was lost. The figure demonstrates
both the heterogeneity of the plankton communities sampled at the
different depths in the region of the fluorescence peak and the similarity
in plankton community structure at the two stations.

250783
13/6/03
3:34 pm
Page 790
JOURNAL OF PLANKTON RESEARCH

VOLUME
NUMBER

PAGES
‒

Table II: Distribution patterns of species in HRS samples from station S1
Species
Re
Prorocentrum micans
Ra
A
*
Species
Re
C. subtilis
*
Dinophysis acuminata
*
C. socialis /radians
*
D. cf. dens
*
Guinardia flaccida
D. norvegica
*
Leptocylindrus danicus
*
Ceratium furca
*
Guinardia delicatula
*
*
Rhizosolenia fragilissima
*
C. lineatum
*
R. pungens
C. longipes
*
Skeletonema costatum
*
*
C. macroceros
*
Thalassiosira punctigera
C. tripos
*
Pseudonitzschia sp.
*
*
*
Nitzschia closterium/longissima
*
Katodinium glaucum
*
Thalissionema nitzschioides
*
Protoperidinium steinii
*
Salpingella sp.
*
Lohmaniella oviformis
*
P. curtipes
*
Diplopsalis spp.
*
Athecate dinoflagellate 2–10 µm *
Strombidium spp. 20–30 µm
*
Mesodinium rubrum <25 µm
*
*
Athecate dinoflagellate 10–20 µm
*
Mesodinium rubrum >25 µm
Thecate dinoflagellate 2–10 µm
*
Kinetoplastida sp.
Thecate dinoflagellate 10–20 µm
*
Pyramimonas/Tetraselmis
A
*
C. fusus
Gymnodinium chlorophorum
Ra
*
*
Dictyocha speculum
*
Prymnesiophyceae spp.
*
Chaetoceros compressus
*
Choanoflagellida spp.
*
C. curvisetus
*
Cryptophyceae spp.
*
Re, regular; Ra, random; A, aggregated.
Table III: Species in HRS samples from station S1 whose distributions were found to decrease or
increase with depth
Species
I
D
Species
I
Dinophysis acuminata
*
Dictyocha speculum
*
D. cf. dens
*
C. curvisetus
*
D. norvegica
*
C. socialis/radians
*
R. pungens
*
Ceratium furca
*
C. fusus
*
C. macroceros
Thalassiosira punctigera
*
D
*
Nitzschia closterium/longissima
*
Protoperidinium steinii
*
Strombidium spp. 20–30 µm
P. curtipes
*
Mesodinium rubrum <25 µm
*
*
Diplopsalis spp.
*
Cryptophyceae spp.
*
I, increasing with depth; D, decreasing with depth.
the species), randomly distributed (26%) or regularly
distributed (56%). An aggregated distribution was typically ascribed to species with one or more abundance
peaks. For randomly distributed species, no pattern in

distributions could be discerned and regularly distributed
species were more or less uniformly distributed in the
region of HRS sampling.
The distribution category assigned in Table II describes
3:34 pm
Page 791
L. T. MOURITSEN AND K. RICHARDSON
VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES
the general distribution pattern of the species in the water
column. Within all of these categories of distribution,
there is the possibility that numbers of an individual
species may increase or decrease with depth. Those
species for which a significant pattern of distribution with
respect to depth in the HRS samples was observed are
identified in Table III. The concentrations of 18% of the
species decreased significantly with depth, while 24% of
the species increased in number with depth (Spearman’s
coefficient of correlation, P < 0.05). The remaining
species showed no systematic trend with depth.
0
Depth (m)
8
10
Dinophysis norvegica
12
Distribution patterns within taxonomic groups. Within the
different taxonomic groups, some patterns in distribution
were noted. Almost all of the diatoms encountered in the
HRS samples were found to be regularly distributed,
although the total diatom biomass decreased slightly with
depth. In contrast, the majority of dinoflagellate species
were either randomly or aggregatedly distributed in the
water column. The dinoflagellates could be divided into
two groups with respect to distribution. One group, which
consisted of potentially mixotrophic species (G. chlorophorum, D. norvegica, Dinophysis acuminata, Dinophysis cf. dens,
P. micans and C. macroceros) showed the same overall
distribution pattern as the autotrophic species. Examples
of this distribution type are shown in Figures 5 and 6. The
other group consisted of heterotrophic species and the
remaining mixotrophic species (Ceratium lineatum, Ceratium
longipes, Ceratium furca, C. tripos, C. fusus, Protoperidinium
curtipes, Protoperidinium steinii, Diplopsalis spp. and
Katodinium glaucum). These species mainly increased in
16
0
5
10
15
20
25
30
35
µg C/l
Fig. 6. Vertical distribution of the dinoflagellate D. norvegica
(µg carbon 1–1) between 8 and 11 m. This distribution pattern is
described as aggregated and the cell numbers decrease with depth.
Closed circles represent samples taken at station S1, open circles at S2.
0
8
Depth (m)
13/6/03
10
0
8
Protoperidinium steinii
12
16
0
Depth (m)
250783
1
2
3
µg C/l
10
Fig. 7. Vertical distribution of the dinoflagellate P. steinii (µg carbon 1–1)
between 8 and 11 m at stations S1 (closed circles) and S2 (open circles).
This distribution pattern is classified as regular and increasing with
depth.
12
Gymnodinium chlorophorum
abundance with depth or were randomly distributed. The
distribution of P. steinii is shown in Figure 7 as an example
of a typical distribution for this group. Most ciliates were
regularly distributed in the water column.
16
0
50
100
150
200
µg C/l
Fig. 5. Vertical distribution of the dinoflagellate G. chlorophorum
(µg carbon 1–1) between 8 and 11 m at stations S1 (closed circles) and
S2 (open circles). This was the most abundant plankter in the HRS
samples and its distribution pattern is classified as aggregated.
Distribution of groups. No statistically significant similarities
in the vertical distribution of various groups of species
(grouped according to taxonomy, nutritional form,

250783
13/6/03
3:34 pm
Page 792
JOURNAL OF PLANKTON RESEARCH
VOLUME
mobility and size), were found (Kendall’s coefficient of
concordance, dinoflagellates/diatoms/ciliates: W = 0.38,
218 = 20.7, P = 0.29; auto-/mixo-/heterotrophs: W =
0.38, 218 = 15.1, P = 0.66; mobile/immobile: W = 0.66,
218 = 23.8, P = 0.16; small/medium/large cells: W =
0.48, 218 = 25.9, P = 0.10). These tests were performed
on the median of the depth-specific frequencies of species
comprising the various groups. In contrast, a significant
similarity in distribution patterns was found among
heterotrophic species (W = 0.34, 219 = 38.3, P = 0.005)
and autotrophic species (W = 0.17, 218 = 45.9, P =
0.001), when the concordance of the distribution of the
species within these groups was tested. No overall concordance was found among the vertical distributions of
species of the other groups. Thus, there appeared to be
common factors determining the distribution of species
in the autotrophic group and factors influencing the distribution of the species comprising the heterotrophic group.
Species–species relationships
Of all possible relationships between species, 13% were
significantly positive and 7% were significantly negative
(Pearson’s coefficient of correlation, P < 0.05). Some of
the negative relationships could be explained by major
differences in distribution patterns (e.g. increasing or
decreasing with depth), suggesting a possible response to
a third external factor, for example irradiance or nutrient
availability. However, others might be interpreted as being
the result of real negative interactions between species.
Dinophysis acuminata and D. acuta were, for example, negatively correlated with 7 and 19% of the remaining species,
respectively. Of the species negatively correlated with
both D. acuminata and D. acuta, no obvious explanation
(such as increasing as opposed to decreasing numbers
with depth) for the negative associations were found for
C. longipes, Rhizosolenia fragilissima, Chaetoceros curvisetus and
Leptocylindrus danicus.
Nutritional form
The distributions of some of the species for which nutritional form is in question or has been demonstrated to
vary [e.g. (Hansen, 1991a,b; Jacobson and Andersen,
1994) were examined by calculating Kendall’s coefficient
of concordance with respect to the vertical distribution of
these species and species representing nutritional groups
they could either belong to or be associated with. A high
degree of concordance was found between D. norvegica,
D. cf. dens and D. acuminata (W = 0.70, 219 = 39.9, P =
0.003). These Dinophysis spp., when tested together with
the group ‘heterotrophic dinoflagellates’, gave no concordance (W = 0.08, 219 = 10.4, P = 0.94). However, a
significant concordance was found between these Dinophysis spp. and the autotrophic species that decreased with


NUMBER

PAGES
‒

depth (W = 0.41, 219 = 46.1, P < 0.001). Autotrophs that
decreased with depth were tested separately from total
autotrophs as the cells that decreased with depth might be
expected to be actively photosynthesizing and growing. In
contrast, autotrophs that increased in abundance with
depth could be sinking or dying cells. Nevertheless, testing
Dinophysis spp. together with all autotrophs also gave a
significant, although low, concordance (W = 0.15, 219 =
50.5, P < 0.001).
Likewise, the distributions of P. micans and C. macroceros
were found to resemble the above-mentioned Dinophysis
spp. (W = 0.51, 219 = 48.1, P < 0.001) and autotrophic
species that decreased with depth (W = 0.31, 219 = 29.1,
P < 0.06). No concordance was found between the other
Ceratium species (C. furca, C. fusus, C. lineatum, C. longipes, C.
tripos) and Dinophysis spp. However, a low but statistically
significant concordance (W = 0.21, 219 = 42.7, P =
0.001) existed between the distribution of these Ceratium
species and the heterotrophic species (Protoperidinium spp.,
Diplopsalis spp., K. glaucum, Salpingella sp., Lohmaniella
oviformis), suggesting that these species may share a
common feeding form. The distribution of G. chlorophorum
(Figure 5), which harbours an autotrophic endosymbiont
(Elbrächter and Schnepf, 1996), showed a statistically
significant similarity with the distribution of Mesodinium
rubrum (both size classes combined), which utilizes chloroplasts from ingested cryptomonad prey for photosynthesis
(Gustafson et al., 2000), and Strombidium spp. (20–30 µm),
which is also possibly mixotrophic (Nielsen and Hansen,
1999) (W = 0.41, 219 = 31.3, P = 0.038). A more detailed
analysis of the distribution of M. rubrum cells revealed that
while large M. rubrum cells were more or less uniformly
distributed, small M. rubrum cells increased with depth in
the depth interval surrounding the fluorescence peak
(Figure 8). Furthermore, the distribution of the potentially mixotrophic ciliates resembled that of the
mixotrophic dinoflagellate G. chlorophorum, which may
indicate that feeding mode can be a significant factor
influencing the distribution of ciliates.
Predator–prey relationships
A statistically significant concordance between the distributions of the dinoflagellates Protoperidinium spp. and
various potential diatom prey (Chaetoceros spp., Rhizosolenia
spp., L. danicus) may suggest a possible predator–prey
relationship between Protoperidinium spp. and these species
(W = 0.28, 219 = 47.5, P < 0.001). No agreement was
found between the distribution of Protoperidinium spp. and
the distribution of potential dinoflagellate prey (Dinophysis
spp., P. micans, G. chlorophorum, K. glaucum) (W = 0.09, 219
= 15.6, P = 0.68).
A similarity in the distribution of the heterotrophic
ciliates and potential nanoflagellate prey was found.
13/6/03
3:34 pm
Page 793
L. T. MOURITSEN AND K. RICHARDSON
VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES
indicating that the patterns we observed do not simply
reflect a chance distribution of the species recorded at
the time of our sampling. A striking demonstration of the
similarity in the vertical distribution of species was the
fact that one flagellate, Kinetoplastida sp., was found in
only one sample at both stations. It was very abundant in
both samplings but found only in the deepest sample
taken with the HRS. In contrast to the findings for microplankton, the vertical distribution patterns observed for
nanoplankton were not found to be similar at both
stations. This may suggest that the smaller plankton are
not able to control their vertical positions in the water
column to the same extent as microzooplankton.
0
< 25 µm
8
> 25 µm
Depth (m)
250783
10
Mesodinium rubrum
12
Distribution of diatoms
16
0
1
2
3
4
µg C/l
Fig. 8. Vertical distribution in the water column of M. rubrum (cells
smaller than 25 µm represented as closed circles; cells > 25 µm as open
circles). Both cell sizes classify as being regular in their distribution and
the numbers of the smaller cells increase with depth.
Significant concordances were found between the ciliate
species, Salpingella spp. (W = 0.39, 29 = 28.6, P = 0.001),
L. oviformis (W = 0.25, 29 = 17.2, P = 0.038) and Strombidium spp.[(20–30 µm), W = 0.33, 29 = 23.7, P = 0.005]
and their potential prey organisms [Cryptophyceae spp.
(6–14 µm, 15–20 µm), athecate dinoflagellates (2–10 µm),
thecate dinoflagellates (2–10 µm), Pyramimonas/Tetraselmis
spp., Choanoflagellida spp. and Chrysochromulina spp.]. A
rather high, but non-significant, concordance was found
between C. furca and the potential ciliate prey (Strombidium
spp. 20–30 µm and Salpingella sp.) (W = 0.45, 219 = 25.8,
P = 0.14).
DISCUSSION
Interpretation of studies of the vertical distribution of
plankton species is usually hampered by the time interval
occurring between collection of samples at various
depths. Here, we have used a novel sampling device that
allowed nearly simultaneous sampling at 20 different
depths over a 3-m depth range. This allowed us to
overcome the problems of time intervals between
sampling and turbulence (which potentially disturbs the
microscale distributions of plankton species relative to
one another) caused by raising and lowering a sampling
device in the water column. The study clearly demonstrates microscale vertical heterogeneity in, especially, the
distributions of the microplankton species. Furthermore,
the major patterns observed in this vertical heterogeneity
of microplankton were similar at the two stations studied,
Statistical analysis of the distributions of different species
and groups of species indicated different patterns of
distribution in the region of the water column sampled by
the HRS: regular, aggregated or random distribution.
Most of the diatoms encountered in the HRS samples
were regularly distributed. Exceptions were Guinardia
flaccida (randomly distributed), Rhizosolenia pungens
(randomly distributed) and Thalassiosira punctigera (aggregated). We have no explanation for why these species
might exhibit different general distribution patterns than
the rest of the diatoms encountered but note that
remnants of an earlier bloom, where cells are now sinking
out of the water column, might not be expected to show
a regular distribution. Several diatom species were found
to decrease in number with depth (Table III), which might
be expected to be a response to light availability. Two
diatoms, including the aggregatedly distributed T. punctigera, increased in cell number with depth. This, again,
might suggest that T. punctigera was sinking out of the
water column. Sinking of living and, in particular, dying
cells will always be a significant process influencing the
distribution of diatoms, especially under conditions of
nutrient depletion [(Bienfang et al., 1981) and references
therein; (Passow, 1991)].
Some diatom species, (i.e. Chaetoceros spp. and Nitzschia
spp.), were mainly found in the upper and lower part of
the depth interval sampled by the HRS, i.e. at complementary depths to those where the dominating dinoflagellate G. chlorophorum was most abundant. A possible
explanation for this pattern is that the gelatinous matrix
of G. chlorophorum may interfere with the spiny structures
of Chaetoceros spp. and Nitzschia spp. and, thereby, impede
the growth of these species. This hypothesis is supported
by the observation that also C. macroceros, a dinoflagellate
with very long and slender horns, was also distributed
complementarily to G. chlorophorum. Evidence of possible
negative interactions between G. chlorophorum and other
plankton species has not previously been reported.

250783
13/6/03
3:34 pm
Page 794
JOURNAL OF PLANKTON RESEARCH
VOLUME
In contrast, negative effects of the DSP toxin producing species, D. acuminata and D. acuta, on other microalgae
have been demonstrated (Windust et al., 1996). In the
present study, D. acuminata and D. acuta were negatively
correlated with a number of other species, in particular
diatoms, and this can be envisaged to be a result of toxic
or inhibitory effects of exudates released from Dinophysis
cells.
Distribution of dinoflagellates
The majority of dinoflagellate species were either
randomly or aggregatedly distributed in the region of
the HRS sampling. Dinoflagellate distributions could be
divided into two groups. One group, which consisted of
autotrophic and potentially mixotrophic species, showed
the same overall distribution pattern as other
autotrophic species (mainly diatoms). The other group
consisted of the heterotrophic species and the remaining mixotrophic species. The species in the latter group
mainly increased in abundance with depth or were
randomly distributed.
Examining the first group more closely, we find the
most dominant of the organisms sampled by the HRS to
be G. chlorophorum. This species was aggregatedly distributed within this depth region and showed greater abundance here than in either the surface or bottom samples.
This distribution pattern is unlike that observed by
Elbrächter and Schnepf (Elbrächter and Schnepf, 1996),
who found G. chlorophorum to accumulate at the surface.
However, the bloom in Aarhus Bay, which was first
observed at the pycnocline during this study, expanded to
cover the upper part of the water column in September
and October (P. Andersen, personal communication).
Dinophysis norvegica, D. cf. dens and D. acuminata were
predominantly found in the upper part of the depth
interval, at the level of 10% of surface light, but not in
the surface water. A similar observation concerning light
preference was made by Carpenter et al. (Carpenter et al.,
1995), who found D. norvegica to be most abundant deeper
in the water column at low photon flux densities. This
may indicate a relatively narrow range of photon flux
densities preferred by Dinophysis spp. in contrast to
another member of this first group, P. micans, which was
uniformly distributed in the HRS samples, abundant in
the surface water and rare at the bottom. The wide
vertical distribution of P. micans (i.e. the upper two-thirds
of the water column) during mid-day has also been
demonstrated by other authors (Staker and Bruno, 1980;
Eppley et al., 1984; Edler and Olsson, 1985) and suggests
a tolerance of a broad span of photon flux densities by
this organism.
The second group of dinoflagellates, which consisted of
the heterotrophic species and the remaining mixotrophic


NUMBER

PAGES
‒

species, comprised species that mainly increased in abundance with depth or were randomly distributed. Ceratium
lineatum and C. longipes were both randomly distributed
within the HRS samples and very rare at the surface and
the bottom samples. An overall distributional similarity
between C. longipes and C. lineatum and the strictly heterotrophic species was demonstrated. No concordance was
found in the distribution of C. longipes and potential ciliate
prey species.
In contrast, the distributions of the other potentially
mixotrophic Ceratium species in this group (C. furca, C. tripos
and C. fusus) were very wide, with equal concentrations in
samples from the surface, the HRS and the bottom,
although all three species tended to increase with depth
within the HRS samples. The vertical distribution of
C. furca around noon has been demonstrated to be wide
but also to vary between successive samplings in a number
of studies (Blasco, 1978; Eppely et al., 1984; Edler and
Olsson, 1985). The distribution of C. fusus has also been
found to be wide but dependent on the nutritional state
of the cells (Mikaelyan and Zavyalova, 1999).
The variable distribution pattern characteristic of
C. furca suggests heterotrophy to be a more likely feeding
mechanism in this species than autotrophy. Less regularity would be expected in the distribution of heterotrophic
cells dispersed in accordance with the dynamic distribution of various prey species than in the distribution of
autotrophic cells determined by the less dynamic environmental gradients controlling photosynthesis. A similarity
between the distribution of C. furca and of potential ciliate
prey was found. Together with a high abundance of
ciliates and C. furca, C. fusus and C. tripos at the surface, this
suggests the distribution of Ceratium spp. to be strongly
influenced by the distribution of prey in the present study.
The heterotrophic dinoflagellates P. curtipes and P. steinii
were predominantly found in the lower part of the depth
interval sampled by the HRS and only rarely in the
surface or bottom waters. In contrast, Protoperidinium divergens was more abundant in the surface layer than in the
HRS samples. These observations agree well with those
of Eppley et al. (Eppley et al., 1984) who demonstrated that
P. divergens migrate vertically in the water column and are
most abundant in the surface layer at noon. In contrast,
Blasco (Blasco, 1978) found another species of the
Protoperidinium genus, Protoperidinium depressum, to be distributed predominantly in the subsurface layers, probably as
a result of a negative phototactic response. In the present
study, the distributions of P. curtipes and P. steinii suggest
that the distributions may be influenced by the
distribution of potential diatom prey species, indicated by
a distributional similarity.
The heterotrophic species Diplopsalis spp. and K. glaucum
were found mainly in the deepest HRS samples and the
250783
13/6/03
3:34 pm
Page 795
L. T. MOURITSEN AND K. RICHARDSON
VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES
‘bottom’ sample. Negative phototaxis may explain this
distribution but prey availability may also be important.
Diplopsalis spp. feed on other large (>20 µm) dinoflagellates (Hansen, 1991b), which were shown to increase with
depth in the present study. We might expect the small K.
glaucum to feed on nanoflagellates. The concentration of
heterotrophic nanoflagellates, which are known to feed
on, for example, bacteria associated with sinking,
decaying algal cells, can be predicted to increase with
depth. In this study, the group Cryptophyceae spp. (14–20
µm), which may include heterotrophic species, was found
to increase with depth and the bacteriovore, Kinetoplastida sp., was only found in the deepest samples. This could
potentially explain the observed distribution of
K. glaucum.
Cells of the photosynthesizing ciliate M. rubrum were
divided into two distinct size classes, as is customary
(Montagnes et al., 1999). Mesodinium rubrum may represent
a ‘species complex’ [(Gustafson et al., 2000) and references
therein]. Morphological variation is large (Crawford,
1993), as is the variation of cell-specific carbon:chlorophyll ratio and photosynthetic rate (Stoecker et al., 1991).
Furthermore, several authors have found the vertical
distribution of M. rubrum to vary considerably both on a
diel and a seasonal basis (Lindholm and Moerk, 1990;
Williams, 1996; Crawford and Lindholm, 1997; Olli,
1999). In the present study, a significant difference was
found between the distribution of small (∼15 12 µm)
and large (∼25 22 µm) cells of M. rubrum, suggesting the
two size classes to be different not only morphologically
but possibly also ecologically. The small cells were most
abundant in the lower part of the depth interval sampled
by the HRS and rare at the surface and bottom, while
large cells were uniformly distributed in the HRS samples
and abundant in the bottom sample (Figure 8). In a recent
study, Pérez et al. (Pérez et al., 2000) also found small M.
rubrum cells (length 15 µm) to accumulate around the
depth of a subsurface chlorophyll maximum, thus
avoiding the surface and bottom layers.
Both small and large cells of M. rubrum are believed to
engulf Cryptophyceae spp. (Owen et al., 1992; Gustafson
et al., 2000). The smallest cryptomonad cells were
predominantly found in the lower part of the depth
interval sampled by the HRS, coinciding with the small
M. rubrum cells, while the larger cryptomonad cells were
more randomly distributed, suggesting the distribution of
prey cells influences the distribution of M. rubrum.
However, nutrient availability (Nielsen and Kiørboe,
1994; Gustafson et al., 2000) and irradiance can also be
expected to influence the distribution of photosynthesizing M. rubrum cells.
Mobility
Two-thirds of the immobile species but only about onequarter of the mobile species were regularly dispersed in
the HRS samples (Tables I and II). This pattern, also
found by Owen (Owen, 1989) and Olli (Olli, 1999),
suggests that the immobile species are subject to dispersal
through turbulent water motion, whereas the mobile
species are, to a larger extent, able to overcome external
physical forcing and exert control over their position
within the water column.
Size
The percentage of species demonstrating a non-regular
distribution (i.e. random or aggregated) in the region of
the HRS sampling increased with increasing size group.
The groups of medium and large species were dominated
by mobile hetero- and mixotrophic species, while the
group of small species included many immobile
autotrophic species but also a number of mobile species.
Although the relative swimming velocity of small cells is
higher than that of larger cells (Kiørboe, 1993), their
ability to actively position themselves in the water column
in the face of opposing water motion is smaller. The
susceptibility to dispersion by water movements (including those caused by our sampling procedure) may, thus,
be expected to be greater in small than in large cells.
The influence of hydrography on microscale
plankton distributions
A comparison of the profiles of fluorescence, nutrient
concentrations, salinity and temperature in the depth
interval sampled by the HRS revealed a layered structure
common to all parameters, suggesting that physical
mechanisms also influence vertical microscale plankton
patchiness. A multi-layered structure, with discontinuities
in salinity and temperature associated with peaks in fluorescence, may result from various hydrographic phenomena, e.g. intrusions of intermediate salinity water
transporting nutrients and/or plankton. Differences in
the species composition between the lowest and the two
uppermost biomass peaks sampled by the HRS also
indicate a layered structure consisting of different
plankton communities. The overall tendency of auto- and
mixotrophic species to peak above the heterotrophic
species within the depth interval sampled by the HRS is
repeated in each of the minor biomass peaks. The basis
for the vertical microscale patchiness observed in the
present study may, thus, originate from a hydrodynamically established structure.
However, on the basis of existing knowledge concerning the autecology of specific plankton organisms and the
statistical analyses reported here, we hypothesize that
biological processes (i.e. behaviour of individual species

250783
13/6/03
3:34 pm
Page 796
JOURNAL OF PLANKTON RESEARCH
VOLUME
and/or interactions between species or communities) may
influence the fine-scale vertical distribution of individual
plankton species.
This study indicates that a mid-water column fluorescence peak may, in fact, contain a complex pattern of
species distributions that can only be revealed through
microscale sampling. We have shown that individual species
exhibit distinct and reproducible patterns in their
microscale vertical distributions. The presence of vertical
microscale heterogeneity as demonstrated here may have
important implications for sampling strategies and
interpretation of results. Food web structure, species interactions and the relative importance of the species comprising a heterogeneous fluorescence peak may be difficult to
infer from a single sample. Furthermore, sampling by vertically positioned water bottles that integrate water from a
column the length of the bottle may lead to misinterpretations of species co-existence and species interactions.
AC K N O W L E D G E M E N T S

NUMBER

PAGES
‒

Edler, L. (1979) Recommendations for marine biological studies in the
Baltic Sea. The Baltic Marine Biologists Publications, No 5, pp. 1–38.
Edler, L. and Olsson, P. (1985) Observations on diel migration of
Ceratium furca and Prorocentrum micans in a stratified bay on the Swedish
west coast. In Anderson, D. M., White, A. W. and Baden, D. G. (eds),
Toxic Dinoflagellates. Elsevier Science Publishing Co., Inc., New York.
Elbrächter, M. and Schnepf, E. (1996) Gymnodinium chlorophorum, a
new, green, bloom-forming dinoflagellate (Gymnodiniales, Dinophyceae) with a vestigial prasinophyte endosymbiont. Phycologia, 35,
381–393.
Elliott, J. M. (1977) Some Methods for the Statistical Analysis of Samples of
Benthic Invertebrates, 2nd edn. Scientific Publication No. 25. Freshwater
Biological Association, Ambleside.
Eppley, R. W., Reid, F. M. H., Cullen, J. J., Winant, C. D. and Steward,
E. (1984) Sub-surface patch of a dinoflagellate (Ceratium tripos) off
Southern California: patch length, growth rate, associated vertically
migrating species. Mar. Biol., 80, 207–214.
Grasshoff, K., Ehrhardt, M. and Kremling, K. (1983) Methods of Seawater
Analysis. Verlag Chemie, Weinheim, Germany.
Gustafson, D. E., Stoecker, D. K., Johnson, M. D., Van Heukelem, W.
F. and Sneider, K. (2000) Cryptophyceae algae are robbed of their
organelles by the marine ciliate Mesodinium rubrum. Nature, 405,
1049–1052.
We would like to thank T. G. Nielsen, National Environmental Research Institute (Denmark) for lending us the
HRS. Aarhus County (Denmark) is thanked for allowing
us to carry out sampling with their monitoring ship,
Thyrfing. This study was partially supported with a grant
from the Danish Natural Science Research Council.
Hansen, P. J. (1991a) Dinophysis—a planktonic dinoflagellate genus which
can act both as a prey and a predator. Mar. Ecol. Prog. Ser., 69,
201–204.
REFERENCES
Jacobson, D. M. and Andersen, R. A. (1994) The discovery of mixotrophy in photosynthetic species of Dinophysis (Dinophyceae): light
and electron microscopial observations of food vacuoles in Dinophysis
acuminata, D. norvegica and two heterotrophic dinophysoid dinoflagellates. Phycologia, 33, 97–110.
Allen, W. E. (1921) A brief study of the range of error in microenumeration. Trans. Am. Microsc. Soc., 40, 14–25.
Beardall, J. and Morris, I. (1976) The concept of light intensity adaptation in marine phytoplankton: some experiments with Phaeodactylum
tricornutum. Mar. Biol., 37, 377–387.
Hansen, P. J. (1991b) Quantitative importance and trophic role of
heterotrophic dinoflagellates in a coastal pelagial food web. Mar. Ecol.
Prog. Ser., 73, 253–261.
Holmes, R. W. and Widrig, T. M. (1956) The enumeration and collection of marine phytoplankton. J. Cons. Int. Explor., 22, 21–32.
Kiørboe, T. (1993) Turbulence, phytoplankton cell size, and the structure of pelagic food webs. Adv. Mar. Biol., 29, 2–72.
Bienfang, P., Szyper, J. and Laws, E. (1981) Sinking rate and pigment
response to light limitation of a marine diatom: implications to
dynamics of chlorophyll maximum layers. Oceanol. Acta, 6, 55–62.
Lindholm, T. and Moerk, A. C. (1990) Depth maxima of Mesodinium
rubrum (Lohman) Hamburger and Buddenbrock—examples from a
stratified Baltic Sea inlet. Sarsia, 75, 53–64.
Bjørnsen, P. K. and Nielsen, T. G. (1991) Decimeter scale heterogeneity
in the plankton during a pycnocline bloom of Gyrodinium aureolum. Mar.
Ecol. Prog. Ser., 73, 263–267.
Lund, J. W. G., Kipling, C., Le Cren, E. D. (1958) The inverted microscope method of estimating algal numbers and the statistical basis of
estimations by country. Hydrobiologia, 11, 143–171.
Blasco, D. (1978) Observations on the diel migration of marine dinoflagellates off the Baja California Coast. Mar. Biol., 46, 41–47.
Mikaelyan, A. S. and Zavyalova, T. A. (1999) Vertical distribution of
heterotrophic phytoplankton of the Black Sea in summer.
Okeanologiya, 39, 895–902 (in Russian).
Carpenter, E. J., Janson, S., Boje, R., Pollehne, F. and Chang, J. (1995)
The dinoflagellate Dinophysis norvegica: biological and ecological observations in the Baltic Sea. Eur. J. Phycol., 30, 1–9.
Clarke, K. R. (1993) Non-parametric multivariate analyses of changes
in community structure. Aust. J. Ecol., 18, 117–143.
Montagnes, D. J. S., Poulton, A. J. and Shammon, T. M. (1999)
Mesoscale, finescale and microscale distribution of micro- and
nanoplankton in the Irish Sea with emphasis on ciliates and their prey.
Mar. Biol., 134, 167–179.
Crawford, D. W. (1993) Some observations on morphological variation
in the red-water ciliate Mesodinium rubrum. J. Mar. Biol. Assoc. UK, 73,
975–978.
Nielsen, T. G. and Hansen, P. J. (1999) Dyreplankton i danske farvande.
(In Danish). Miljø- og Energi- ministeriet. Danmarks Miljøundersøgelser,
1999.
Crawford, D. W. and Lindholm, T. (1997) Some observations on vertical
distribution and migration of the phototrophic ciliate Mesodinium
rubrum (= Myrionecta rubra) in a stratified brackish inlet. Aquat. Microb.
Ecol., 13, 267–274.
Nielsen, T. G. and Kiørboe, T. (1994) Regulation of zooplankton
biomass and production in a temperate, coastal ecosystem. 2. Ciliates.
Limnol. Oceanogr., 39, 508–519.

Olli, K. (1999) Diel vertical migration of phytoplankton and
250783
13/6/03
3:34 pm
Page 797
L. T. MOURITSEN AND K. RICHARDSON
VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES
Statistics in Biological Research. W.H. Freeman and Co., New York, pp.
593–608.
heterotrophic flagellates in the Gulf of Riga. J. Mar. Syst., 23,
145–163.
Olrik, K. (1991) Planteplankton metoder. Miljøministeriet, Miljøstyrelsen.
Owen, R. W. (1989) Microscale and finescale variations of small
plankton in coastal and pelagic environments. J. Mar. Res., 47,
197–240.
Owen, R. W., Gianesella-Galvão, S. F. and Kutner, M. B. B. (1992)
Discrete, sub-surface layers of the autotrophic ciliate Mesodinium
rubrum off Brazil. J. Plankton Res., 14, 97–105.
Passow, U. (1991) Species-specific sedimentation and sinking velocities
of diatoms. Mar. Biol., 108, 449–455.
Pérez, M. T., Dolan, J. R., Vidussi, F. and Fukai, E. (2000) Diel vertical
distribution of planktonic ciliates within the surface layer of the NW
Mediterranean (May 1995). Deep-Sea Res., 47, 479–503.
Pinel-Alloul, B. (1995) Spatial heterogeneity as a multiscale characteristic of zooplankton community. Hydrobiologia, 300/301, 17–42.
Rasmussen, B. (1997) The near-surface horizontal buoyancy flux in a
highly stratified region, Kattegat. Estuarine Coastal Shelf Sci., 45,
405–414.
Richardson, K., Beardall, J. and Raven, J. A. (1983) Adaptation of
unicellular algae to irradiance: an analysis of strategies. New Phytol.,
93, 157–191.
Staker, R. D. and Bruno, S. F. (1980) Diurnal vertical migration in
marine plankton. Bot. Mar., 23, 167–172.
Stoecker, D. K., Davis, L. H. and Anderson, D. M. (1984) Fine scale
spatial correlations between planktonic ciliates and dinoflagellates.
J. Plankton Res., 6, 829–842.
Stoecker, D. K., Putt, M., Davis, L. H. and Michaels, A. E. (1991) Photosynthesis in Mesodinium rubrum: species-specific measurements and
comparison to community rates. Mar. Ecol. Prog. Ser., 73, 245–252.
Tiselius, P., Nielsen, G. and Nielsen, T. G. (1994) Microscale patchiness
of plankton within a sharp pycnocline. J. Plankton Res., 16, 543–554.
Tokarev, Yu. N., Williams, R. and Piontkovski, S. A. (1999) Identification
of micro-scale structure of plankton communities of the Black and
Ionian Seas by their bioluminescence characteristics. Hydrobiologia,
393, 163–167.
Venrick, E. L. (1978) How many cells to count? In Sournia, A. (ed.),
Phytoplankton Manual. UNESCO, Paris, pp. 167–180.
Villareal, T. A., Altabet, M. A. and Culver-Rymsza, K. (1993) Nitrogen
transport by vertically migrating diatom mats in the North Pacific
Ocean. Nature, 363, 709–712.
Ricker, W. E. (1937) Statistical treatment of sampling processes useful in
enumeration of plankton organisms. Arch. Hydrobiol., 31, 68–84.
Williams, J. A. (1996) Blooms of Mesodinium rubrum in Southampton
Water—do they shape mesozooplankton distribution? J. Plankton Res.,
18, 1685–1697.
Rosen, B. H. and Lowe, R. L. (1984) Physiological and ultrastructural
responses of Cyclotella meneghiniana (Bacillariophyta) to light intensity
and nutrient limitation. J. Phycol., 20, 173–183.
Windust, A. J., Wright, J. L. C. and McLachlan, J. L. (1996) The effects
of diarrhetic shellfish poisoning toxins, okadaic acid and dinophysistoxin-1, on the growth of microalgae. Mar. Biol., 126, 19–25.
Smalley, G. W., Coats, D. W. and Adam, E. J. (1999) A new method using
fluorescent microspheres to determine grazing on ciliates by the
mixotrophic dinoflagellate Ceratium furca. Aquat. Microb. Ecol., 17,
167–179.
Yentsch, C. S. and Menzel, D. W. (1963) A method for the determination
of phytoplankton chlorophyll and phaeophytin by fluorescence. DeepSea Res., 10, 221–231.
Sokal, R. R. and Rohlf, F. J. (1995) Biometry: The Principles and Practice of
Received on June 5, 2002; accepted on March 28, 2003
