ARTICLE IN PRESS
http://www.elsevier.de/protisPublished online date 2 April 2008
1
Correspondine-mail bollma
& 2008 Elsevdoi:10.1016/j
159, 369—381, July 2008
Protist, Vol.ORIGINAL PAPER
Morphological Variation of Gephyrocapsaoceanica Kamptner 1943 in Plankton Samples:Implications for Ecologic and TaxonomicInterpretations
Jorg Bollmanna,c,1, and Christine Klaasb,c
aDepartment of Geology, Earth Sciences Centre, University of Toronto, 22 Russell Street, Toronto, Ontario,Canada M5S 3B1
bAlfred Wegener Institute for Polar and Marine Research, Am Handelshafen 12,27570 Bremerhaven, Germany
cEarth Science Department ETH Zurich, Sonneggstrasse 5, CH-8092 Zurich, Switzerland
Submitted September 7, 2007; Accepted February 9, 2008Monitoring Editor: Michael Melkonian
Morphological analysis of Gephyrocapsa spp. in plankton samples confirms the existence of five outof six morphotypes that were previously reported from Holocene sediments. Our data suggest a muchhigher diversity within the genus Gephyrocapsa than the currently accepted species circumscriptions.Furthermore, we confirm the morphological species delineations made by Kamptner that allow theseparation of three morphological groups within the genus Gephyrocapsa: one group with largebridge angles (G. oceanica var. typica Kamptner 1943), a second group with small bridge angles(G. oceanica var. californiensis Kamptner 1956) and a third group of small coccoliths (G. apertaKamptner 1963). However, a seemingly continuous transition from small to large coccoliths withinG. oceanica var. typica along a temperature gradient points either to a high phenotypic plasticity ofG. oceanica var. typica or numerous sibling species highly adapted to specific environmentalconditions. Testing of these hypotheses is of utmost importance to understanding the diversity ofmarine plankton and its evolution, and to assessing the impact of future and past environmentalchange on primary producers such as coccolithophorids.& 2008 Elsevier GmbH. All rights reserved.
Key words: prymnesiophyte; coccolithophorids; morphospecies; phenotypic plasticity; marine protest;biodiversity.
Introduction
The oceans cover about 71% of the Earth’ssurface and thus marine micro-organisms mayplay a major role in shaping the global environ-ment. Coccolithophores represent a particularly
g author; fax +1 416 978 [email protected] (J. Bollmann).
ier GmbH. All rights reserved..protis.2008.02.001
important group of marine micro-organismsbecause they constitute a significant part of themarine food web as primary producers and areimportant players in global biogeochemicalcycles. Their production of dimethyl sulphide(DMS) and of calcite platelets, respectively, mighthave had a major impact on the global climate
ARTICLE IN PRESS
370 J. Bollmann and C. Klaas
since their first occurrence in the late Triassic(Westbroek et al. 1993). Therefore, knowledge ofthe factors driving their distribution in today’socean is essential to assessing the impact offuture and past environmental change on themarine ecosystem and on marine plankton evolu-tion. However, the biodiversity, biogeography,ecology and evolution of coccolithophores arestill not well understood (Thierstein and Young2004). One reason for our lack of understandingappears to be the morphology-based taxonomicspecies concept that currently does not accountfor small-scale morphological variability possiblyreflecting species-level biodiversity.
The traditional taxonomic perception of homo-geneous cosmopolitan species has becomequestionable in recent years because severalcosmopolitan species exhibit remarkable finescale morphological variations (Bollmann 1997;Bollmann and Herrle 2007; Hagino et al. 2000;Knappertsbusch et al. 1997; Quinn et al. 2005;Young and Westbroek 1991). These fine scalevariations allow the identification of morphotypesthat show a characteristic biogeographic patternor environmental adaptation (Brand 1981). Theseresults have led to questions regarding whethereco-phenotypic or genotypic variation within thecosmopolitan species such as Gephyrocapsaoceanica cause the fine scale morphologicalvariation (Bollmann 1997).
Species Definitions within the GenusGephyrocapsa
All extant species of the genus Gephyrocapsaform spherical to sub-spherical coccospheres ofelliptical coccoliths with a diagonal bridge cross-ing the central area of the coccoliths (Kamptner1943). Many variations of coccolith size, coccolithbridge angle and shape, the central collar, and thesize of the central pore exist. In his originaldefinition of the genus Gephyrocapsa, Kamptner(1943) included all coccoliths with a single bar(bridge) across the central area. Kamptner (1956)first used the angle between the bar crossing thecentral area and the short axis of the ellipticalcentral area (bridge angle) to distinguish betweenG. oceanica var. typica Kamptner, 1943 (largebridge angle) and G. oceanica var. californiensisKamptner, 1956 (small bridge angle), and thecoccolith size to identify G. aperta Kamptner, 1963(small size: 2—3 mm).
McIntyre et al. (1970) used similar morphologi-cal characteristics, but provided for the first time
precise morphometric boundaries to distinguishspecies within Gephyrocapsa. McIntyre et al.(1970) distinguished three species: G. oceanica(warm water species with a bridge angle greaterthan 451), G. caribbeanica Boudreaux and Hay,1967 (cold water species with a bridge anglesmaller than 451) and the small (2.2—1.9mm) G.ericsonii McIntyre and Be, 1967. Later, Pujos-Lamy (1976), Breheret (1978) and Samtleben(1980) distinguished several morphotypes whithinthe genus Gephyrocapsa based on measurementsof size and bridge angle of Gephyrocapsa cocco-lith assemblages from sediment samples. Sincethen, various combinations of size, bridge angle,roundness (ratio of width/length), pore width andother descriptive features led to the descriptionof numerous morphospecies from the Mioceneto the Holocene. A detailed overview is givenby Perch-Nielsen (1985) and Bollmann (1997;Table 1).
Currently, the extant Gephyrocapsa speciesdelineations used closely resemble those pro-posed by Kamptner (1956, 1963) based on bridgeangle and coccolith size: G. oceanica var. typica(large bridge angle), G. oceanica var. californiensis(syn. G. muellerae Breheret, 1978; syn. G. car-ibbeanica Boudreaux and Hay, 1967; small bridgeangle) and G. aperta (syn. G. ericsonii McIntyreand Be, 1967; distinguishable by its minutecoccoliths). An additional distinction of Gephyro-capsa species with small coccolith size hasbeen used based on ornamentations of thecentral collar or the shape of the distal shield(i.e. Gephyrocapsa ornata Heimdall, 1973;Gephyrocapsa crassipons Okada and McIntyre,1977; Gephyrocapsa protohuxleyi McIntyre, 1970).In contrast to previous studies, the quantitativemorphological analysis of Gephyrocapsa assem-blages from globally distributed Holocene sedi-ment samples revealed six dominant morphologi-cal associations related to distinct environmentalconditions with respect to sea surface tempera-ture and productivity (Figure 1; Bollmann (1997)).Furthermore, morphological parameters such assize or bridge angle appear to vary with environ-mental conditions such as temperature, showingseemingly continuous transitions between allmorphological associations, suggestive of oneglobal species with high eco-phenotypic plasticity(Fig. 1).
Bollmann (1997) described the different mor-photypes informally because it is not evident fromhis sediment analysis whether the different mor-photypes correspond to discrete species, or toone species showing a high phenotypic plasticity.
ARTICLE IN PRESS
Figure 1. Morphological variation of Gephyrocapsa in Holocene sediments. (A) Morphological measure-ments determined from a single coccolith: BA, Bridge angle; L, Coccolith Length. (B) Six morphologicalassociations of Gephyrocapsa determined in Holocene sediment assemblages. (C) Biogeography of the sixdifferent Gephyrocapsa morphological associations determined in Holocene sediments. The lines over thesymbols indicate the mean orientation of the bridge within an assemblage (modified after Bollmann (1997)).
371Morphological Variation of Gephyrocapsa
Furthermore, taxonomic and ecological datainferred from sediment samples might be biasedby taphonomic effects such as selective dissolu-tion of morphotypes, lateral transport, sedimentmixing via bioturbation, or seasonal variations inthe production of different morphotypes.
Several feasiblility tests need to be conductedto verify whether the different Holocene morpho-types represent distinct species (Bollmann 1997).These tests include the following: (1) Studies ofplankton and trap materials from time-seriesstations would reveal whether the Holocenemorphotypes and morphological associations arepresent in the plankton, and might indicatewhether morphological changes in coccoliths onliving coccospheres along an environmental gra-dient are continuous (suggestive of phenotypicvariation) or discontinuous (suggestive of succes-sions of genetically distinct populations). (2)Culture experiments of monoclonal populationsof Gephyrocapsa isolates from various regionsgrown under various environmental conditionswould provide a test for phenotypic plasticity.A change in morphology of any cultured popula-tion into that of another morphotype wouldprovide positive evidence for phenotypic plasti-city. Consequently, populations of these differentmorphotypes would assume the taxonomic rankof subspecies. (3) Finally, genetic and morpho-
metric analysis of Gephyrocapsa from culturedstrains and in the field (Iglesias-Rodrıguez et al.2006; Saez et al. 2003) are needed.
Here, we analysed the morphology of Gephyro-capsa spp. in plankton samples from all oceanbasins covering a range of temperature andproductivity gradients (Fig. 2, Table 1), to testwhether (1) Holocene morphological associationsand morphotypes of the genus Gephyrocapsa canbe also identified in the plankton, and (2) differentmorphotypes represent discrete species or onespecies showing high eco-phenotypic plasticity.
Results and Discussion
We analysed the morphology of about 500 cocco-liths from 16 globally distributed plankton samples(Figs 3 and 4) and applied the same criteria asBollmann (1997) to distinguish between unimodaland bi- or polymodal samples and to identifydifferent morphotypes. Our analysis revealed thatfour of the six Holocene morphotypes reported byBollmann (1997) were present in unimodallydistributed plankton samples (Figs 3A—F, M;4A—F, M, and 5A, B). These morphotypes areGephyrocapsa Equatorial, Gephyrocapsa Larger,Gephyrocapsa Cold, and Gephyrocapsa Minute.The morphotype Gephyrocapsa Transitional (GT)
ARTICLE IN PRESS
-60° -60°
-30° -30°
0° 0°
30° 30°
60° 60°-180° -150° -120° -90° -60° -30° 0° 30° 60° 90° 120° 150°
HOT52
Okada 6
JGOFS FH13
COD-X
MK-ZH21
P212/2 3509M42/1b 327
Meteor 32-X
Sonne 119-XFB160
Photo
2µm PC71
A
B
CD
E
Bridge Angle (°)
Freq
uenc
y
Freq
uenc
y
Length (µm) Length (µm)
GL
GC
Brid
ge A
ngle
(°)
Figure 2. (A) Locations and code names of sampling sites for plankton samples analysed in this study;names with an X represent locations where various samples were taken (Table 1). (B) Left panel: SEM pictureof a coccosphere of Gephyrocapsa oceanica. Right panel: dimensions measured from digitised SEM picturesof a single coccolith; bridge angle measured from the long axis of the central area (a), coccolilth length (b),coccolith width (c), length of the central area (d), width of the central area (e). (C) Example of a bimodalassemblage consisting of two morphological associations. Left panel: frequency distribution of the bridgeangle (interval size 101). Middle panel: frequency distribution of the length (interval size 0.4 mm). Right panel:scatter plot of length versus bridge angle showing separation of the assemblage into two distinctmorphological associations Gephyrocapsa Cold (GC) and Gephyrocapsa Larger (GL). The standard deviationof the assemblage is 19.911 for the bridge angle and 0.59 mm for length. After separating the two modes, thestandard deviations of the dominant mode (GC) is 7.51 for the bridge angle and 0.39 mm for length.
372 J. Bollmann and C. Klaas
could be identified only in one bimodal sample(Figs 3P, 4P, 5C) and the morphotype Gephyro-capsa Oligotrophic (GO) could not be identified inany plankton sample, although coccolith measure-ments on single coccospheres corresponded tothis morphotype in several of our samples (Fig. 3).The near-absence of Gephyrocapsa Transitionaland Gephyrocapsa Oligotrophic in the planktonsamples is possibly due to the poor geographicalcoverage of our samples in the open ocean ingeneral and in the oligotrophic gyres in particular.Hence, the existence of these morphotypes in theplankton cannot be ruled out.
The most common combination of morphotypesin polymodal samples was Gephyrocapsa Largerand Gephyrocapsa Cold (Fig. 3K, L, N, O and P),and to a lesser extent, Gephyrocapsa Minute
(Fig. 3H, K and O). A similar combination ofmorphotypes was found in Holocene sedimentsamples (Bollmann 1997). The co-occurrenceof the morphotypes Gephyrocapsa Larger,Gephyrocapsa Cold and Gephyrocapsa Minute inthe same plankton sample rules out the pre-sence of one single cosmopolitan species withhigh phenotypic plasticity dependent on environ-mental conditions. Furthermore, this observa-tion suggests that each of these morphotypesrepresents a separate biological species. There-fore, our data support the species delineation ofKamptner (1943, 1956, 1963), who distinguishedbetween G. oceanica var. typica (large bridgeangle) and G. oceanica var. californiensis (smallbridge angle) as well as G. aperta (small size:2—3 mm).
ARTICLE IN PRESS
Ta
ble
1.
Nam
es
and
locatio
ns
of
pla
nkto
nsam
ple
sused
inth
isstu
dy
tog
eth
er
with
enviro
nm
enta
lp
ara
mete
rsand
mo
rpho
metr
icd
ata
.(S
am
ple
)S
am
ple
nam
e.
(Date
)S
am
plin
gd
ate
.(D
ep
th)
Sam
plin
gd
ep
th.
(Tem
p)
In-s
itu
wate
rte
mp
era
ture
at
the
tim
eand
dep
tho
fco
llectio
n.
(Co
nc)
Gep
hyr
ocap
sacell
co
ncentr
atio
n.
(AvN
)M
ean
annual
nitra
teco
ncentr
atio
navera
ged
over
the
up
per
100
mo
fth
ew
ate
rco
lum
n.
(Pho
sp
)P
ho
sp
hate
co
ncentr
atio
nat
the
tim
eand
dep
tho
fco
llectio
n.
(AvP
)M
ean
annualp
ho
sp
hate
co
ncentr
atio
navera
ged
over
the
up
per
100
mo
fth
ew
ate
rco
lum
n.
(Chla
)M
ean
annualsurf
ace
Chlo
rop
hyll-
aco
ncentr
atio
n.
(Ang
le)
Sam
ple
avera
ge
brid
ge
ang
le.
(Leng
th)
Sam
ple
avera
ge
co
cco
lith
leng
th.
(Po
re)
Sam
ple
avera
ge
po
resiz
e.
(N)
Num
ber
of
co
cco
liths
measure
din
the
sam
ple
.(S
TD
EV
)S
tand
ard
devia
tio
n.
(NN
)N
ot
measure
d.
Sam
ple
Latitu
de
(1)
Lo
ng
itud
e(1
)D
ep
th(m
)D
ate
Tem
p.
(1C
)S
alin
ity
Co
nc.
(cells
ml�
1)
Nitra
te(m
mo
ll�
1)
AvN
a
(mm
oll�
1)
Pho
sp
.(m
mo
ll�
1)
AvP
a
(mm
oll�
1)
Chla
a
(mg
l�1)
Ang
le(1
)S
TD
EV
(1)
Leng
th(m
m)
ST
DE
V(m
m)
Po
re(%
)S
TD
EV
(%)
N
CO
D1-3
381
23.50 N
91
14.90 W
43
20.0
4.9
913.8
NN
NN
NN
3.5
6N
N0.2
10.2
753.8
417.6
14.4
90.8
743
3.9
42
CO
D1-1
381
09.5
80 N
91
29.4
50 W
50
20.0
4.9
914.4
NN
NN
NN
3.4
7N
N0.1
90.2
838.0
918.7
73.8
90.9
841
4.3
43
CO
D2-5
381
440 N
91
320 W
815.0
6.9
914.5
35.8
5N
N3.5
3.5
30.4
90.2
10.2
556.0
120.2
14.5
70.8
346
4.1
30
JG
OF
SF
H13
491
40.90 N
201
47.70 W
309.0
6.9
014.7
235.5
4N
N1.7
5.6
50.2
0.2
70.5
523.1
66.7
33.3
80.2
545
3.9
28
FB
160
ST-
5371
29.00 N
071
22.30 E
25
21.0
2.8
815
37.1
632.7
6N
N1.9
9N
N0.0
70.4
146.7
818.1
14.3
20.6
745
4.9
50
CO
D2-1
381
400
N91
560 W
915.0
6.9
917.5
35.8
5N
N0.5
3.4
60.3
0.2
00.2
442.0
720.3
04.2
20.9
452
4.6
38
Mete
or
32-5
-438
181
31.80 N
571
20.50 E
40
04.0
8.9
519.7
535.7
43.6
021.4
810.9
61.6
91.0
70.2
165.6
710.4
44.2
70.4
947
4.9
30
MK
-ZH
21
181
510 N
191
36.50 W
5N
N19.9
NN
NN
NN
9.8
3N
N0.6
20.6
762.4
111.9
45.1
20.5
744
3.8
33
M42/1
b_327-7
2291
10.20 N
151
300 W
25
01.0
7.9
821.1
736.7
90.9
00.0
91.8
10
0.1
20.2
628.7
012.6
82.2
21.1
058
7.8
21
HO
TS
52
221
450 N
15810 W
50
01.0
2.9
422.6
35.1
710.0
00.0
10.2
90.0
70.1
80.0
969.6
99.0
44.8
30.5
156
5.8
29
P212/2
3509-1
291
05.3
50 N
15103.4
880 W
25
24.0
9.9
523.1
36.8
0.0
40.0
01
1.7
40.0
20.1
20.2
926.8
210.6
11.9
90.5
257
3.3
34
So
nne
119-3
7201
33.90 N
601
03.10 E
20
03.0
6.9
725.9
936.2
55.5
0N
N11.8
5N
N1.0
90.5
664.8
010.7
74.5
00.5
649
3.3
30
Mete
or
32-5
-404
091
58.80 N
651
01.10 E
100
24.0
7.9
527.7
836.5
542.3
00.7
42.9
10.2
80.4
10.1
071.0
88.7
03.9
80.4
052
4.9
30
Mete
or
32-5
-404
091
58.80 N
651
01.10 E
20
24.0
7.9
527.9
36.5
548.8
00
2.9
10.1
90.4
10.1
071.5
97.8
73.9
40.5
157
4.9
30
OK
AD
A-6
101
30 N
1541
55.00 W
30
NN
28.5
NN
NN
NN
1.8
7N
N0.4
80.0
767.6
313.0
63.9
10.3
745
5.1
58
So
nne
119-7
161
12.10 N
601
18.50 E
20
24.0
5.9
729.7
736.5
14.0
0N
N6.7
7N
N0.7
20.2
471.6
09.1
43.4
60.3
346
4.4
30
aF
rom
Levitus
et
al.
(1998).
373Morphological Variation of Gephyrocapsa
ARTICLE IN PRESS
Figure 3.
374 J. Bollmann and C. Klaas
ARTICLE IN PRESS
Figure 3. (Continued)
375Morphological Variation of Gephyrocapsa
ARTICLE IN PRESS
Figure 3. Bivariate plots (left panels) and frequency histograms of coccolith length (middle panels) andbridge angle (right panels) of coccoliths measured during this study in individual plankton samples (A-P).Sample name and temperature at the time of collection are given for each sample in the middle panel. Linesin the bivariate plots indicate morphological boundaries reported by Bollmann (1997).
376 J. Bollmann and C. Klaas
The seemingly continuous change in morpholo-cical charateristics along a temperature gradientas reported from Holocene sediments (Bollmann1997) was observed in the coccolith size ofplankton samples (Fig. 5D) but not in the bridgeangle. This is a reflection of the fact that only oneGT assemblage and no clearly defined GOassemblages were found in plankton samples.However, in contrast to observations from Holo-cene sediments, the variation in coccolith lengthwith temperature (or other associated environ-mental variables) is due to a continuous transitionbetween morphotypes with a large bridge angle(Gephyrocapsa Larger and Gephyrocapsa Equator-
ial; Fig. 5B, D). Coccolith length of GephyrocapsaLarger/Gephyrocapsa Equatorial assemblagesincrease with decreasing temperature, reachingmaximum values at 20 1C. Below 20 1C no changein average coccolith length is observed (Fig. 5D).This can be explained by: (a) mixing between thetwo morphotypes Gephyrocapsa Larger andGephyrocapsa Equatorial along the temperaturegradient from 20 to 29.6 1C, where GephyrocapsaLarger represents the ‘‘Cold’’ end member andGephyrocapsa Equatorial the ‘‘Warm’’ end mem-ber; (b) the presence of numerous species highlyadapted to a specific temperature; or (c) thepresence of phenotypic plasticity within a single
ARTICLE IN PRESS
Figure 4. Box plots of plankton samples analysed in this study. Error bars represent the range betweensmallest and largest value of corresponding morphological parameter in each sample. Lower quartile (lowerboundary of open rectangle), median (line inside open rectangle), upper quartile (upper boundary of openrectangle) and average (solid squares) are shown. Note — Sample names in bold indicate bimodal orpolymodal samples. In addition, plots of each morphological association of the bimodal or polymodalsamples are shown. The separation of morphotypes within polymodal/bimodal assemblages is done basedon the bivariate plots. GE ¼ Gephyrocapsa Equatorial, GO ¼ Gephyrocapsa Oligotrophic, GT ¼ Gephyr-ocapsa Transitional, GC ¼ Gephyrocapsa Cold, GL ¼ Gephyrocapsa Larger, GM ¼ Gephyrocapsa Minute.
377Morphological Variation of Gephyrocapsa
species in response to changes in temperature orco-varying environmental parameters.
The assumption of a single species is contra-dicted by the variability in growth rates of variousstrains of G. oceanica determined by Brand (1982)in culture experiments. Additional evidence for thepresence of more than one species withinGephyrocapsa oceanica was reported by Haginoet al. (2000). Based on the proportion of thecentral area as compared to total coccolith area,Hagino et al. (2000) differentiated between twodifferent morphotypes of Gephyrocapsa oceanicawith morphotype 1 restricted to eutrophic areasand/or deeper layers in the water column andmorphotype 2 occuring in oligotrophic regions.Although information on in-situ nutrient concentra-tions and primary productivity were not available
for several sampling sites in this study, ourdata essentially confirm the finding of Haginoet al. (2000) as G. oceanica coccoliths with a largecentral area seem to occur exclusively in areas withhigh SST (420 1C), low nutrient concentrations,and low Chlorophyll-a concentration (Figs 5E, 6).Based on the available data, however, we cannotrule out that the transition from a large to a smallproportion of the central area is due to eco-phenotypic response to environmental conditions.
Ecological Range
The morphotypic composition of the planktonsamples indicates that the patterns observed inHolocene sediment samples are not obscured or
ARTICLE IN PRESS
Figure 5. (A) Scatter plots of mean length versus mean bridge angle of morphotypes defined in Holocenesediment samples. Gephyrocapsa Equatorial (GE): mean bridge angle larger than 561 and mean lengthbetween 3.1 and 3.9 mm. Gephyrocapsa Oligotrophic (GO): mean bridge angle between 271 and 561 andmean length larger than 3.1 mm (characteristics of the subtropical central gyres). Gephyrocapsa Transitional(GT): mean bridge angle between 271 and 561 and mean length between 2.4 and 3.1 mm. Gephyrocapsa Cold(GC): mean bridge angle less than 271 and mean length larger than 2.4 mm. Gephyrocapsa Larger (GL): meanbridge angle larger than 561 and mean length larger than 3.9 mm. Gephyrocapsa Minute (GM): mean bridgeangle between 201 and 501 and mean length less than 2.4 mm. (B) Mean bridge angle versus mean length ofall unimodal plankton populations. (C) Mean bridge angle versus mean length of morphotypes within thepolymodal plankton populations. (D) Mean length of all Gephyrocapsa Larger and Gephyrocapsa Equatorialmorphotypes in plankton samples versus in situ temperature. (E) Mean relative pore size versus annual meanchlorophyll concentration. Lines indicate the mean values reported by Hagino et al. (2000). Note — B—E:error bars ¼ 95% confidence limit of the mean.
378 J. Bollmann and C. Klaas
ARTICLE IN PRESS
Figure 6. Scatter plots of pore size versus coccolith length of all unimodal GE and GL associations. Linesindicate the mean values reported by Hagino et al. (2000). Each panel represents values for one planktonsample. In situ water temperatures (1C), annual mean chlorophyll concentrations (mg Chl-a l-1), sample codeand sampling depth (m) are given in the panels.
Table 2. Comparison of temperature range of the six Gephyrocapsa morphotypes in Holocene sediments andplankton samples, respectively. (Name) Morphotype name. (SST range) Temperature range from Holocenesamples based on mean sea surface temperatures. (In-situ temp.) In-situ temperatures at the time and depthof collection of plankton samples.
Name SST range (1C) In-situ temp. (1C)
Gephyrocapsa Equatorial 25—29 27.8—29.8Gephyrocapsa Larger 18—23 13.8—25.8Gephyrocapsa Cold o21 14.8Gephyrocapsa Transitional 19—20 13.8Gephyrocapsa Oligotrophic 22—25 22.6Gephyrocapsa Minute NN 23.1
379Morphological Variation of Gephyrocapsa
biased by non-biological processes in the sedi-ments. The comparison of ecological ranges ofthe different morphotypes as reported fromHolocene sediments and our plankton data revealno major differences except for the GephyrocapsaTransitional morphotype (COD1-3, Table 2).
Furthermore, our data support the findings ofBollmann (1997) that Gephyrocapsa Cold/Transi-tional and Gephyrocapsa Larger morphotypes arefound within similar temperature ranges but indifferent ecological provinces. The occurrence ofunimodal Gephyrocapsa Cold association inplankton samples from the open North Atlanticand the combination of Gephyrocapsa Cold and
Gephyrocapsa Larger morphotypes in the Iberiancoasts (Fig. 3L, N—P) suggests different habitatslinked to nutrient conditions or other parametersassociated with the transition from coastal to openocean systems.
Conclusion
Our results confirm the morphological speciesconcept introduced by Kamptner (1943, 1956,1963) within the genus Gephyrocapsa comprisingone group with large bridge angles (G. oceanicavar. typica), a group with small bridge angles
ARTICLE IN PRESS
380 J. Bollmann and C. Klaas
(G. oceanica var. californiensis) and a group withtiny coccoliths (G. aperta). Furthermore, wedemonstrate the presence, in plankton samples,of five of six morphotypes reported by Bollmann(1997) from his analysis of Holocene sedimentsamples (Gephyrocapsa Larger, GephyrocapsaEquatorial, Gephyrocapsa Cold, GephyrocapsaTransitional and Gephyrocapsa Minute).
In addition, we show a gradual change inmorphological characteristics between Gephyro-capsa Larger and Gephyrocapsa Equatorial alonga temperature gradient which could result from theexistence of several species with narrow environ-mental range, eco-phenotypic plasticity within asingle species, mixing of the two morphotypes, ora combination of these factors. Testing of thesehypotheses requires genetic and culture studiesunder varying environmental conditions usingmonoclonal strains from different ocean basins,and is of utmost importance to understanding thediversity of marine plankton and its evolution, andto assessing the impact of future and pastenvironmental change on coccolithophorids.
Finally, our study demonstrates that quantitativeanalysis of coccolith morphology in the genusGephyrocapsa and possibly in most extant coc-colithophore genera is necessary for addressingkey taxonomic, ecological and evolutionaryissues.
Methods
Plankton samples: Sixteen plankton samples were analysedfrom the Indian, Atlantic and Pacific Oceans (Fig. 2A, Table 1).Water samples were obtained from discrete depths (listed inTable 1) using Niskin and immediately filtered on individualmembrane filters. After filtration, samples were dried overnightat approximately 50 1C and stored in separate containers forfurther analysis (for additional details see Bollmann et al.(2002)). Information on environmental parameters at the timeand depth of collection are provided in Table 1, whereavailable, together with climatic estimates for nutrient con-centrations and Chlorophyll-a (Levitus et al. 1998).
Morphometric measurements: Morphometric measure-ments of plankton samples were carried out according toBollmann (1997) (Fig. 1 B). A minimum of 30 coccoliths weremeasured per sample. Isolated coccoliths as well ascoccoliths on coccospheres were analysed, when possible.All measurements were collected using a Hitachi S2300 and aPhilips XL30 Scanning Electron Microscope (SEM) at amagnification of 8000� .
Because of instrument-related uncertainties when a SEM isused for geometric measurements (for details see ASTMCommittee E-4 (1993)), the geometry and accuracy of sizemeasurements in this study were controlled with measure-ments of �30 calibration spheres (width and length) of 2, 5, 7and 10 mm nominal diameter, respectively, before and aftereach plankton sample measurement series. Correction factors
were applied when average sizes for the calibration spheresmeasured before and after each plankton sample measure-ment series differed from the nominal sphere sizes. Allmeasurements with an apparent size offset were corrected.
The comparison between measurements of identicalcalibration spheres (nominal diameter of 1.9870.1 mm stan-dard error) conducted by Bollmann (1997) and this study,respectively, revealed that all measurements reported byBollmann (1997) are about 6.5% too small. The size ofcalibration spheres measured by Bollmann (1997) was about1.85mm rather than 1.98mm (Bollmann, unpublished data).This small size offset was assumed to be negligible, as themeasurements varied within the given statistical standarddeviation of 70.1 mm for this type of calibration sphere.Therefore, the measurements were not corrected for theapparent size offset in Bollmann (1997). However, in thisstudy, all measurements and the resulting morphologicalboundary values published by Bollmann (1997) were adjustedby a factor of 1.065 in order to avoid a biased sample set.All data are available at http://www.ngdc.noaa.gov/mgg/geology.
Data treatment: The data treatment was conductedaccording to Bollmann (1997) using the most importantindependent morphological characteristics of Gephyrocapsacoccoliths, namely, bridge angle and coccolith length (othermeasured morphological characteristics where not usedbecause they tend to be correlated with coccolith length).Briefly, the mean value and variance of all morphometricmeasurements in each assemblage were calculated (Table 1).Box plots were used to compare the variance of coccolith sizeand bridge angle between single samples (Fig. 4).
Samples were separated into two subsets: one samplesubset with low variance in length and bridge angle (unimodalassemblages), and a second sample subset with highvariance in length and bridge angle (polymodal assemblages,see Fig. 2C, left and middle panel). Assemblages withstandard deviations of less than 161 for the bridge angle orassemblages with standard deviations of less than 0.7 mm forcoccolith length were considered to be unimodal. These limitswere chosen after visual inspection of frequency histogramsof bridge angle and length measurements for each individualplankton sample (example in Fig. 2C, left and middle panels).The remaining assemblages with standard deviations 4161for bridge angle or standard deviations 40.7 mm for coccolithlength were considered bi- or polymodal. The polymodalassemblages were further subdivided into associations on thebasis of the frequency distributions of bridge angle and length.Scatter plots of length versus bridge angle were also used inall cases to support the separations (example in Fig. 2C, rightpanel). The width of size classes in frequency histograms were101 for the bridge angle and 0.4 mm for the coccolith length. Inaddition, a Kruskal—Wallis non parametric variance test (at a99.9 probability level) of samples (Gephyrocapsa EquatorialGephyrocapsa Larger ) along a temperature gradient from 20to 29 1C revealed that median lengths are separable amongthis sample set.
Acknowledgements
Michael Knappertsbusch, Hisatake Okada, HansR. Thierstein and Ralf Schiebel provided planktonsamples. Mara Y. Cortes, Rebecca R. Ghent andJens O. Herrle made very useful comments on an
ARTICLE IN PRESS
381Morphological Variation of Gephyrocapsa
earlier version of the manuscript. The commentsof three anonymous reviewers greatly helped toimprove the manuscript. This research wasfunded by the Swiss National Fund Projects: No.2053-053676: Plankton Ecology and Taxonomy,and JB’s Natural Sciences and EngineeringResearch Council Canada Discovery Grant.
References
ASTM Committee E-4 (1993) Standard Practice for Calibrat-ing the Magnification of a Scanning Electron MicroscopeASTM Designation E 766. ASTM International, West Con-shohocken, PA, USA pp 614—617
Bollmann J (1997) Morphology and biogeography of Gephyro-capsa coccoliths in Holocene sediments. Mar Micropaleontol29: 319—350
Bollmann J, Cortes MY, Haidar AT, Brabec B, Close A,Hofmann R, Palma S, Tupas L, Thierstein HR (2002)Techniques for quantitative analyses of calcareous marinephytoplankton. Mar Micropaleontol 44: 163—185
Bollmann J, Herrle HO (2007) Morphological variation ofEmiliania huxleyi and sea surface salinity. Earth and Planet SciLett 255: 273—288
Brand LE (1981) Genetic variability in reproduction ratesin marine phytoplankton populations. Evolution 35:1117—1127
Brand LE (1982) Genetic variability and spatial patterns ofgenetic differentiation in the reproductive rates of the marinecoccolithophores Emiliania huxleyi and Gephyrocapsa ocea-nica. Limnol Oceanogr 27: 236—245
Breheret JG (1978) Stratigraphie: Formes nouvelles quater-naires et actuelles de la famille des Gephyrocapsaceaem(Coccolithophorides). C R Acad Sc Paris 287: 447—449
Hagino K, Okada H, Matsuoka H (2000) Spatial dynamics ofcoccolithophore assemblages in the Equatorial Western-Central Pacific Ocean. Mar Micropaleontol 39: 53—72
Iglesias-Rodrıguez MD, Schofield OM, Batley J, MedlinLK, Hayes PK (2006) Intraspecific genetic diversity in themarine coccolithophore Emiliania huxleyi: the use of micro-satellite analysis in marine phytoplankton population studies.J Phycol 42: 526—536
Kamptner E (1943) Zur Revision der Coccolithen-SpeciesPontosphaera huxelyi Lohmann. Anz Akad Wiss Wien, Math -Naturw Kl 80: 43—49
Kamptner E (1956) Das Kalkskelett von Coccolithus huxleyi(Lohmann) Kamptner und Gephyrocapsa oceanica Kamptner(Coccolithineae). Arch Protistenkd 101: 99—202
Kamptner E (1963) Coccolithineen-Skelettreste aus Tiefseea-blagerungen des Pazifischen Ozeans. Ann Naturhistor MusWien 66: 139—204
Knappertsbusch M, Cortes MY, Thierstein HR (1997)Morphologic variability of the coccolithophorid Calcidiscusleptoporus in the plankton, surface sediments and from theEarly Pleistocene. Mar Micropaleontol 30: 293—317
Levitus S, Boyer T, Burgett R, Conkright M (1998) WorldOcean Atlas 1998 (WOA98). Ocean Climate Laboratory,National Oceanographic Data Center Data available at/http://ferret.wrc.noaa.gov/las/main.htmlS
McIntyre A, Be AWH, Roche MB (1970) Modern Pacificcoccolithophorida: a paleontological thermometer. N Y AcadSci Trans 32: 720—730
Perch-Nielsen K (1985) Cenozoic Calcareous Nannofossils.In Bolli HM, Saunders JB, Perch-Nielsen K (eds) PlanktonStratigraphy. Cambridge University Press, New York,pp 427—554
Pujos-Lamy A (1976) Emiliania et Gephyrocapsa (Nanno-plancton calcaire); Biometrie et interet biostratigraphique dansle Pleistocene superieur marin des Acores. Rev EspanMicropaleontol 9: 69—84
Quinn PS, Cortes MY, Bollmann J (2005) Morphologicalvariation in the deep ocean-dwelling coccolithophoreFlorisphaera profunda (Haptophyta). Eur J Phycol 40:123—133
Saez AG, Probert I, Geisen M, Quinn P, Young JR, MedlinLK (2003) Pseudo-cryptic speciation in coccolithophores.Proc Natl Acad Sci USA 100: 7163—7168
Samtleben C (1980) Die Evolution der Coccolithophoriden-Gattung Gephyrocapsa nach Befunden im Atlantik. PalaeontolZ 54: 91—127
Thierstein HR, Young J (2004) Coccolithophores fromCellular Process to Global Impact. Springer, Berlin, p 565
Westbroek P, Brown-Christopher W, van Bleijswijk J,Brownlee C, Brummer GJ, Conte M, Egge J, FernandezE, Jordan R, Knappertsbusch M, Stefels J, Veldhuis M, vander Wal P, Young J (1993) A model system approach tobiological climate forcing; the example of Emiliania huxleyi.Global Planet Change 8: 27—46
Young JR, Westbroek P (1991) Genotypic variation in thecoccolithophorid species Emiliania huxleyi. Mar Micropaleon-tol 18: 5—23