Limnol. Oceanogr, 39(5), 1994, 1044-I 060 0 1994, by the ...pc › ~pelagic › s › Montagnes 1994 Limn Ocean.… · Limnol. Oceanogr, 39(5), 1994, 1044-I 060 0 1994, by the American

Limnol. Oceanogr, 39(5), 1994, 1044-I 060 0 1994, by the American Society of Limnology and Oceanography, IIK.

Estimating carbon, nitrogen, protein, and chlorophyll a from volume in marine phytoplankton

David J. S. Montagnes’ and John A. Berges Department of Oceanography, University of British Columbia, Vancouver V6T 124

Paul J. Harrison and F. J. R. Taylor Departments of Oceanography and Botany, University of British Columbia, Vancouver V6T 124

Abstract

The size of 30 small (2-60 pm) phytoplankton species was examined with a microscope and a Coulter Counter before and after fixation. Acid Lugol’s iodine caused cells to shrink immediately. The shrinkage effect was constant for concentrations of l-10% Lugol’s iodine (in seawater). For optically measured cells fixed in 2% Lugol’s iodine, volume of live cells = 1.33 x (volume of fixed cells). Coulter Counter and optically measured volumes did not agree. For live cells, optical cell volume = 1.24-2.04 x (Coulter Counter determined volume); this difference is likely due to inaccurate volume measurements of nonspherical cells by the Coulter Counter and by inaccurate microscopy resulting from optical distortions (errors of ~0.5 pm in cell dimensions). Cell quota estimates were presented following the relation y = a.x?, where x = optically measured cell volume (pm3), y = any cell constituent (pg cell-‘), and a and b are constants. The constants a and b were 0.109 and 0.99 1 for carbon, 0.0172 and 1.023 for nitrogen, 0.043 and 1.058 for protein, and 0.00428 and 0.9 17 for Chl a. Our relation of carbon to volume differs from other literature values, in which there is no consensus. Our data can be used to determine carbon, nitrogen, protein, and Chl a estimates from field material that has been fixed with Lugol’s iodine, observed live, optically measured, or Coulter Counter measured; however, the variability in published data suggests that any of these estimates will have a large potential error.

Although automated methods (e.g. Steen 1990) and indices such as chlorophyll a can be used to estimate phytoplankton biomass, only microscopic analyses offer the ability to resolve species composition or morphological variation (Smayda 1978). Further, automated methods are often expensive and require elab- orate equipment, while microscopy is rela- tively inexpensive and technically simple. Consequently, the microscope is and will likely continue to be a mainstay for phytoplankton biomass evaluation.

Typically, when microscopy is used to de-

I Present address: School of Oceanography, University of Washington, Seattle 98195.

Acknowledgments We thank Cheryl Jerome for assistance with video mea-

surements and culturing, Jeanette Ramirez for providing phytoplankton cultures, Maureen Soon for CNS analyses, and an anonymous reviewer for constructive comments.

Financial support was provided by the Natural Sciences and Engineering Research Council of Canada (NSERC) through operating grants awarded to P.J.H. and F.J.R.T. D.J.S.M. received support from a University Graduate Fellowship, and J.A.B. was supported by a Killam Pre- doctoral Fellowship.

termine the biomass of phytoplankton from field samples, species abundances and cell volumes are determined from preserved cells. Us- ing published carbon : volume relationships, the total biomass is then calculated; analogous relationships may also be used to estimate other cellular constituents (e.g. nitrogen, protein, or Chl a). The final biomass estimates thus depend on the accuracy of the determination of abundances, cell volumes, and cell constituent : volume ratios.

Of these three factors, the determination of abundance has received most attention (e.g. Sournia 1978; Harris 1986; Li 1987). How- ever, estimates of cell volume and the equations for biomass conversion are equally important. We have investigated factors affecting these latter two parameters.

Several factors can influence estimates of cell volume. Fixation can alter cell volume in both phytoplankton and microzooplankton (Bor- sheim and Bratbak 1987; Breteler 1985; Choi and Stoecker 1989; Ohman and Snyder 199 1; Jerome et al. 1993), yet many estimates of C : volume consider live and fixed measurements as interchangeable and occasionally fail to mention whether live or fixed cells were ex-

1044

Estimating phytoplankton biomass 1045

amined. Although the effects of glutaraldehyde (Booth 1987; Verity et al. 1993) and formal- dehyde (Olivieri 198 5) fixation on phytoplankton volume have been examined, few studies have examined similar effects for Lugol’s iodine, which is the recommended (and likely most often used) general fixative and preser- vative for small flagellates (Throndsen 1978). Using linear dimensions of live and Lugol’s fixed phytoplankton (table 1 of Booth 1987), we discerned an effect of fixation on cell volume (Fig. 1). This finding agrees with studies that show a strong effect of Lugol’s iodine on reducing the volume of ciliates (Choi and Stoecker 1989; Ohman and Snyder 199 1; Je- rome et al. 1993). One of our goals was to determine whether Booth’s (1987) results were repeatable and, if possible, to improve on their resolution. By making such an improvement, we planned to provide a correction factor for estimating live volume from fixed material.

The use of electronic particle counters has improved our ability to measure cell numbers and volumes in natural populations (Sheldon and Parsons 1967; Sheldon 1978). Unfortu- nately, there is evidence that particle counters may give inaccurate volumes for nonspherical cells (Grover et al. 1982; Kubitschek 1987) that may not be comparable to microscopically determined volumes (Price et al. 1987). Al- ternatively, visual micrometry may also have biases associated with it; it is often difficult to measure all dimensions of a cell without manipulation, and there may also be optical distortions.

Our ability to convert from volume to carbon, nitrogen, protein, or Chl a is equally important for accurate estimates of biomass. The typically cited conversion factors for determining carbon from cell volume (Parsons et al. 196 1; Mullin et al. 1966; Strathmann 1967) were established prior to the awareness that small flagellates can contribute substantially to marine productivity (e.g. Parsons et al. 1979). For example, only 6 of the 30 species examined in the studies cited above were < 10 pm in diameter. In part, the neglect of small species was due to difficulties in culturing them and to their limited availability. Since the studies cited above were conducted, the number of species in culture has increased significantly. This increase can be partially attributed to im- provements in culturing methodology, pri-

El =r lo*

2

=: al

(: ZlO' x

l r(

I&

10' 10* lo3

Live-cell volume (pm’)

Fig. 1. Log-log relationship between Lugol’s iodine fixed- (unknown concentration) and live- (optically determined) cell volumes. Volumes were determined from linear dimensions presented by Booth (1987). The line reaching from corner to corner represents unity, the solid line through the data represents the least-squares regression, and the broken lines are the 95% confidence intervals. Note that the regression line and most of the data fall below the unity line.

marily seawater enrichment stocks (e.g. Har- rison et al. 1980; Keller et al. 1987), but also to the dedication of several researchers in es- tablishing phytoplankton collections (e.g. Guillard 19 8 8).

Further, the improved techniques which have allowed us to grow and maintain “difficult” phytoplankton species may also have improved our ability to maintain previously cultured species under conditions closer to those they would experience in situ. Because of these changes and advances, Verity et al. (1993) were recently able to examine the relationships of C and N to volume in several small phytoplankton species; they found differences from previous studies. However, they used a new method of image analysis that may not be comparable to the microscopical or electronic particle-sizing techniques. These latter techniques are still in use and need to be re-examined.

Finally, the extent to which culture condi-

1046 Montagnes et al.

tions affect cell constituent : volume ratios (particularly C : volume) has not been ade- quately addressed. Thompson et al. (199 1, 1992) summarized the many studies which have shown the influence of irradiance and temperature on cell volumes and carbon quotas. Despite this, the choices of temperature and irradiance in studies that focus on constituent to volume relationships are seldom rationalized and occasionally are not reported.

In this study, we have examined the possible biases associated with the determination of cell volume (specifically, the effects of fixation with Lugol’s iodine on cell volume and the problems inherent in both microscopic and electronic volume measurement) and biases associated with cell constituent to volume relationships (including the paucity of data for species between 2 and 10 pm in diameter and the effects of improved culturing methods on cell composition).

Materials and methods Culture conditions-Phytoplankton species

(Table 1) were obtained from the Northeast Pacific Culture Collection (NEPCC), Depart- ment of Oceanography, University of British Columbia. Cultures were grown in semicon- tinuous batch culture in enriched artificial seawater (ESAW) based on the recipe of Harrison et al. (1980), except we replaced sodium glyc- erophosphate with an equimolar concentration of sodium phosphate, ferrous ammonium sulfate with an equimolar concentration of fer- ric chloride, and added selenite, nickel, and molybdate to achieve 1 nM final concentration. The final salinity of this medium was 30.5o/oo. Temperature was maintained at 16 +_ 1 .O”C. Cultures were grown in l- or 2-liter glass flasks and stirred at 80 rpm with 5-cm Teflon-coated stir bars. Certain species that grew poorly when stirred were gently swirled once daily.

Cultures were grown on a 14 : 10 L/D cycle. Illumination was provided by Vita-lite fluorescent tubes at an irradiance of 20-60 pmol photon rnA2 s-l, measured in water inside culture vessels with a Biospherical Instruments QS- 100 quantum scalar irradiance meter. At these n-radiances, the growth of many species was below maximum. As Thompson et al. (1991) pointed out, light limitation occurs in nature; light absorption by particulate and dis-

solved materials means that photosynthesis may be limited in the upper parts of the water column, while in deeper waters, where nutrients are abundant, light is again limiting. Also, at higher latitudes during winter, light will be limiting. Previous studies have mainly considered cell composition at saturating irradiances.

Growth rates were followed by in vivo fluo- rescence, measured daily with a Turner De- signs model 10 fluorometer. Cells were usually grown for l-2 weeks, which was equivalent to 5-8 generations for most species but only l-2 generations for the slowest growing cultures (see Table I). Our growth conditions were similar to those of the stock cultures, as accli- mation was probably rapid.

Sampling protocol-AU sampling was conducted in mid- to late-logarithmic growth phase. Subsamples were immediately measured for live volume estimates and then fixed with 2% acid Lugol’s iodine (Throndsen 1978) to be measured -2 weeks later. Concurrently, duplicate samples for C, N, Chl a, and protein were taken from each flask.

For four species which showed evidence of a decrease in cell volume after preservation, we conducted more detailed experiments. Time-series experiments were conducted to determine when shrinkage occurred due to fixation with 2% Lugol’s iodine. Cultures of three flagellates and one diatom (Chlamydomonas sp., Chroomonas salina, Isochrysis galbana, and Thalassiosira weissflogii) were grown as described above. Live samples were taken and live volume was estimated. The samples were fixed, and volume was estimated immediately after fixation (time = 0) and again after 2, 7.5, 24,76,168,336,504, and 2,184 h. In addition, the effect of fixative concentration was examined on these four species. Samples were fixed with 0.2, 0.5, 1.0, 2.0, 5.0, and 10% Lu- gol’s iodine. After 2 weeks, the volumes of these cells were measured. In these time-course and fixative concentration experiments, only Coulter Counter measurements were made (see below).

Volume and abundance estimates-For each species, > 100 fixed and live cells were measured with an ocular micrometer in the ocular of a Zeiss IM inverted microscope equipped with a long distance (LD) condenser, a green filter, and a 12-V, 100-W halogen light source. Live and fixed samples were settled in a 3.5-

Tabl

e 1.

C

ell

para

met

er

mea

sure

men

ts

from

30

live

an

d tie

d (p

rese

rved

in

2%

Lug

ol’s

io

dine

) sp

ecie

s of

phy

topl

ankt

on

(n >

10

0 fo

r ea

ch m

easu

rem

ent).

Al

l cl

ones

wer

e ob

tain

ed

from

th

e N

orth

east

Pa

cific

C

ultu

re

Col

lect

ion

(NEP

CC

). Li

near

km

) an

d vo

lum

etric

&

m3)

m

easu

rem

ents

w

ere

dete

rmin

ed

mic

rosc

opic

ally

: D

-dia

met

er;

L-le

ngth

; W

-wid

th;

B-br

eadt

h (h

t);

Vol-c

ell

volu

me.

Sh

apes

(Sh

) us

ed t

o es

timat

e vo

lum

e fro

m

linea

r di

men

sion

s:

C-c

ylin

der;

S-sp

here

; PS

-pro

late

sp

hero

id;

PB-p

rola

te

sphe

roid

w

ith

brea

dth

as a

thi

rd

dim

ensi

on.

Cel

l vo

lum

es

(CC

Vo

l) an

d th

eir

equi

vale

nt

sphe

rical

di

amet

er

@SD

) w

ere

dete

rmin

ed

by C

oulte

r C

ount

er.

Cel

l qu

otas

of

C,

N,

Chl

a,

and

prot

ein

are

pres

ente

d as

pg

cell-

I.

p re

pres

ents

sp

ecifi

c gr

owth

ra

te (

d- *

) ca

lcul

ated

fro

m

incr

ease

s in

in

vivo

fluor

esce

nce.

(n

-d.

-not

de

term

ined

.)

spec

ies

N’E

PCC

Sh

C

lone

Li

ve

Fixe

d Li

ve

Fiie

d C

ell

quot

as

P

D

L W

B

Vol

D

L W

B

Vol

CC

Vol

ES

D

CC

Vol

ES

D

C

N

Chl

a

vrot

(d

-3

Baci

kiop

hyce

ae

Dtm

ulap

umila

Th

alas

sios

ira ps

eudo

nana

lh

alas

sios

ra w

e$ss

jk@

ch

iyso

phy~

4e

ndkZ

ka

5pin

iira

Pdag

ocvc

cus s

p. (I

) Pe

lugo

cac~

sp

. (2)

Ps

a&pe

dn&

py#o

tmis

C

hlor

ophy

ceae

c7

lhny

dom

onas

sp.

Dun

uhIla

te

dokr

a N

anno

chkw

vpsi

s ocu

lata

W

wM

-w==

C

?mxm

wna

s sah

a ch

roo~

nass

p.

Cry

pton

tona

spm

Jiul

da (I)

C

rypk

mw

n4as

p@nd

a (2

) c?

plzn

nona

s sp.

&w

nonw

nas

salh

a R

hodo

mon

askn

s D

inop

hyce

ae

q$w

vlod

ini.a

on s

angl

4iJl

am

Gj7

nnod

iniu

ntsi

mpl

u:

- .

cym

nodi

nwn

viago

G

y?EM

hiw

n au

Tedo

n G

yrod

iniu

mun

cars

rum

Pr

asin

ophy

ce!a

e M

a?lto

tiua

Squm

Mta

M

inom

onas

pusi

lla

(1)

Mirr

omoM

Fp&

(2)

prym

neai

ophy

- ch

ryso

chtv

~ he

rd-

Gxd

ithus

pc

lagi

clls

l3

nilia

nahu

rlcyi

Ls

oduy

ti ga

lban

a Pa

vkw

a lu

thet

i Ph

aeoc

yslis

pouc

hctii

Pq

ntnc

sium

parv

um

Rap

hido

phyc

eae

Het

ervs

igm

a ca

rtera

e

B 67

4 c

--

36.0

16

.2

-- 7,

395

-- 33

.7

15.5

--

6.36

0 4,

697

20.8

2,

278

16.3

35

5.20

61

.40

7.42

21

1.81

0.

20

B58

c --

4.0

3.6

-- 42

--

3.7

3.2

-- 30

20

3.

4 16

3.

1 5.

94

0.94

0.

20

3.49

0.

36

B636

c

-- 19

.1

6.1

-- 55

2 --

17.7

6.

1 --

522

286

8.2

155

6.7

64.4

4 11

.06

1.54

51

.10

0.29

CH

451

cH66

0

cH6a

cH49

S 9.

6 __

m

e __

S

4.8

__

-- m

e S

4.9

__

-- --

Ps

-- 9.

3 8.

7 --

4.64

59

63

36

5

7.8

-- --

-- 25

1 4.

2 __

-

__

38

4.1

-- --

-- 37

--

7.5

7.1

-- 20

1

222

7.5

127

6.2

47.6

0 9.

90

22

3.5

16

3.1

3.53

0.

68

18

3.3

15

3.0

2.77

0.

71

80

5.4

123

6.2

18.0

0 2.

93

2.28

19

.28

0.22

0.

05

2.59

0.

34

4,

h

0.06

1.

62

0.51

z

0.73

7.

86

0.27

g 2.

11

.31

245.

85

0.28

3

1.78

13

.22

0.32

0.

05

0.28

0.

29

2 c 0.

85

24.9

4 0.

41

1.74

76

.46

0.37

g

2.61

78

.94

0.60

s

2.53

74

.79

0.35

SF

3.

63

163.

17

0.10

1.

04

24.7

7 0.

29

$

0.69

39

.09

0.51

g

57.4

4 3,

509.

70

ad.

0.70

2.

31

53.1

7 0.

22

12.8

9 23

7.60

0.

21

37.7

3 1.

363.

20

0.05

CH

L73

s 18

.6

-- --

-- 3,

392

CH

L1

s 8.

2 __

_a

__

28

5 C

HL6

31

s 2.

6 __

m

m

__

9

19.6

--

- --

3,93

9 --

9.3

6.4

-- 20

1 23

m

a __

__

6

3,30

0 18

.5

1,72

2 14

.9

%9.

69

129.

57

1%

7.2

95

5.7

41.7

0 7.

90

4 2.

1 5

2.1

1.36

0.

23

CR

275

Ps

-- 11

.0

6.5

-- 24

5 -

10.1

6.

5 -

222

167

6.8

101

5.8

32.4

0 7.

92

CR

48

Es

-- 17

.8

8.9

7.4

621

-- 15

.4

7.4

6.9

424

714

11.1

18

9 7.

1 39

.30

10.0

0 cR

6s

Ps

-- 17

.4

9.7

-- 86

3 --

14.8

8.

5 --

560

765

11.4

31

1 8.

4 10

4.70

21

.00

CR

65

Ps

-- 16

.3

8.9

-- 67

0 --

15.2

8.

2 -

529

521

10.8

30

8 8.

4 10

1.40

17

.50

CR

160

Ps

-- 17

.9

9.6

8.8

766

-- 15

.5

8.1

7.9

543

807

11.6

24

6 7.

8 15

2.80

29

.80

CR

76

Ps

- 12

.4

7.6

- 37

9 --

11.0

6.

7 --

258

181

7.0

121

6.1

32.4

0 7.

05

CR

588

Ps

- 10

.5

7.0

-- 27

2 w

e 9.

9 6.

5 --

223

203

7.3

%

5.7

40.7

0 11

.40

D35

4 P

I3

-- 56

.0

37.7

31

.3

34,6

63

--

53.3

35

.4

30.5

30

,167

31

,761

39

.3

16,1

48

31.4

29

.13.

20

688.

40

D 1

19a

Ps

- 10

.6

8.3

-- 37

9 --

9.6

7.7

-- 30

0 22

4 7.

5 15

9 6.

7 38

.00

8.29

D

27

Ps

-- 14

.0

11.0

-

885

-- 14

.0

10.8

--

859

683

10.9

57

3 10

.3

113.

76

22.6

0 D

665

PB

-- 23

.9

19.3

13

.9

3362

--

23.9

20

.7

15.8

4,

128

3,39

9 18

.7

3,55

0 18

.9

513.

90

93.8

3 D

607

Ps

-- 31

.6

27.2

-

12,2

26

-- 30

.8

25.4

--

10,3

70

11,2

46

27.8

9,

217

26.0

2,

275.

U

441.

00

PS24

P29

P29

s 4.

1 __

--

-- 37

4.

0 --

-- --

33

25

3.7

12

2.8

3.56

0.

62

0.15

1.

64

0.23

Ps

-

1.6

1.3

-- 1

-- 1.

6 1.

4 --

2 2

1.6

1 1.

3 n-

d.

n-d.

0.

03

n.d.

0.

44

Ps

-- 1.

9 1.

5 --

2 --

1.8

1.8

2.9

3 2

1.5

1 1.

4 0.

80

0.08

0.

01

0.42

0.

37

Prvl

ls6

PM72

3 PM

659

PM 6

33

PM5

PM22

5

PM

673

s 7.

5 --

-- --

221

6.1

-- --

-- 11

7 74

5.

2 44

4.

4 8.

24

1.95

0.

22

3.38

0.

37

s 11

.5

- --

-- 79

0 9.

7 --

-- -

471

619

10.6

23

2 7.

6 65

.50

14.5

0 1.

92

35.3

7 0.

47

S 5.

1 -

-- --

69

4.7

- --

-- 53

25

3.

6 16

3.

1 4.

57

1.04

0.

06

3.88

0.

33

s 5.

5 _-

--

__

85

4.8

-- -

-- 60

38

4.

2 22

3.

5 6.

97

1.16

0.

17

2.63

0.

32

Ps

- 5.

5 5.

0 --

74

4.7

- --

-- 55

25

3.

6 43

4.

3 8.

55

1.24

0.

23

2.08

0.

14

Ps

- 5.

6 5.

2 --

80

-- 5.

4 5.

0 --

70

45

4.4

35

4.1

5.52

1.

26

0.13

6.

85

0.35

ps

;5

-

7.8

5.9

-- 14

1 --

6.3

5.2

-- 88

79

5.

3 46

4.

4 15

.11

2.03

0.

27

3.65

0.

28

5

Ps

- 15

.8

11.5

--

1096

--

12.3

9.

3 --

559

362

8.8

354

8.7

87.8

0 17

.70

3.75

62

.05

0.34

R

A560


ml settling chamber, 5 mm deep, with a 4-nm-r the phytoplankton cultures were diluted with clear glass cover. Samples were viewed with filtered 3% NaCl for counting. The dilution an F-Achromat 100/l .25 oil immersion ob- was consistent for measurements of a single jective and 12.5 x oculars. Cells typically ap- species. For most cultures, the dilution was 1 : peared to have a thin outer layer, a clear region 20 (cells and medium : 3% NaCl), but at times just inside the outer layer, and an opaque cen- it was as low as 1 : 4. A consistent 1 : 10 dilu- tral region. The linear measurements were tion was used during the time-course and fix- made from the edges of the thin outer layer. ative concentration experiments.

Several dinoflagellates moved too quickly to be accurately measured as described above. Gymnodinium sanguineum, Gymnodinium vitiligo, Gyrodinium aureolum, and Gyrodi- nium uncatenum were videotaped with a Sony Trinicon (DXC 1850) camera control unit, a Zeiss Photomicroscope 2, a Panasonic video cassette recorder (NV-8950), and a Sony Trin- itron monitor (model KX- 190 1). To videotape these dinoflagellates, we collected live cells with a fine-drawn Pasteur pipette, placed them on a coverslip rimmed with a Dow Corning vacuum grease, and covered them with a slide. Measurements were made from the monitor images.

Volume estimates were made from linear dimensions using appropriate geometric shapes (Table 1). Cell volume data were not always normally distributed. A log transformation normalized the data; mean dimensions and volumes were calculated as the antilog of the mean of the transformed data. In most cases where length and width differed, volume was estimated independently for each cell and then the mean volume was determined. When cell shape required that we measure a third dimension (breadth), we typically could not determine it from the same cell as length and width. In this case, mean volume was determined from the values of mean length, width, and breadth (calculated with transformed data, as above).

To calibrate the Coulter Counter, we measured microsphere standards under the microscope. The sizes were verified through the manufacturer (lot number, ASTM standard procedure) and by examination with an electron microscope. We determined whether there was any bias associated with the beads by measuring beads of 2.02-, 5.1 l-, 9.66-, and 43.27- pm diameters with the light microscope in 3% NaCl and, after desiccation, in immersion oil. Beads viewed with the light microscope in 3% saline or in oil were the same size, but in both cases there were light-diffraction rings around the beads, which caused halos. These halos changed with the focal position. Bead diameter was measured as the size of the sphere inside the halo when a sharp edge was seen. The beads were also examined under an electron microscope. Beads were desiccated at 50°C for 4 h, sputtered-coated with gold, and viewed with a Hitachi S-4 100 field emission SEM operated at an accelerated voltage of 5 kV.

Concurrently, volumes and cell densities were determined with a model TAII Coulter Counter equipped with a population accesso- ry. Either a 70- or 200-pm aperture was used, depending on the size of the cell, and the manufacturer’s recommended settings were followed (Coulter Electronics 1979).

The Coulter Counter was calibrated with 5.1 l- (70-pm aperture) or 43.27- (200~pm aperture) pm microspheres. Mean cell volumes were calculated following Thompson et al. (199 1). Beads were also measured in 2% Lu- gol’s iodine, and it was verified that the fixative did not alter the calibration. When necessary,

Cell composition - In all experiments, cell C and N quotas were determined by filtering samples at low vacuum (< loo-mm Hg) onto precombusted 13-mm Gelman type AE glass- fiber filters (except for Micromonas pusilla, for which 25-mm Whatman GF/F filters were used); these filters were processed with a Carlo Erba model 1500 CNS analyzer. Samples for protein and Chl a determination were collected on precombusted 25-mm Whatman GF/F filters. Homogenates for protein analyses were prepared as described by Berges et al. ( 1993) based on the method of Dortch et al. (1984). The samples were ground with 3% trichloro- acetic acid (TCA) and solubilized in 1 N NaOH. Protein was determined by the method of Bradford (1976), using the microassay procedure of the Bio-Rad protein assay kit (Bio-Rad Laboratories, 500-000 1) with bovine serum al- bumin (BSA, Sigma Chemical Co., Cat. NO. A 7638) as a standard. Chl a was determined by the in vitro fluorometric method of Parsons et al. (1984). The coefficients of variation (C.V.) for duplicate C, N, and Chl a analyses were


o” IO' lo2 lo3 IO4 IO5

Fixed-cell volume (pm3) Fig. 2. A. Log-log relationship between optically measured fixed- and live-cell volumes of 30 phytoplankton species

(Table 1). B. Same as panel A, except for Coulter Counter-measured fixed- and live-cell volumes. For both panels: diatoms-o; chrysophytes-O; chlorophytes-W, cryptophytes-V; dinoflagellates-m; prasinophytes-0, prymne- siophytcs- a; raphidophytes- 0. The line reaching from corner to corner represents unity, the solid line through the data represents the least-squares regression, and the broken lines are the 95% confidence intervals. The equations of the lines (as described in the text) and their associated r2 values are shown.

always < 10%. The C.V. for protein analyses was always < 15%.

Data presentation and treatment -To nor- malize variance and facilitate data presentation over four orders of magnitude of cell volume, we logarithmically transformed (base 10) and analyzed the data with linear regression techniques. The relation between regressions of this type and linear plots is not necessarily intuitive. The linear regression equation

log y = b log x + a

would be of the form

Y = axb

on normal axes. If the relation is linear on normal axes, the constant b = 1. In this case, the constant a will give the proportionality between y and X. If b f 1, the relationship is nonlinear; i.e. y increases allometrically at a slower (b < 1) or greater (b > 1) rate than X. It follows then that when b # 1, a does not represent a proportionality over the entire range

of the data. We have used the symbols a, b, x, and y throughout the following discussion and in our figures.

When there is variation in both the x and y components of a regression and the data are obtained from normal or near-normal distri- butions, a geometric mean (model 2) regression should be used to describe the central trend. However, for a regression in which the object is to predict y from X, the regression of y on x (model 1) is most appropriate (Ricker 1984). Since our goal was to determine predictive equations, we have focused on model 1 analyses, but we have also examined model 2 regression (following the procedures of Clarke 1980) to assess whether the central trends differed from those described by model 1 analyses.

Results and discussion Fixation efects on cell volume--In general,

phytoplankton cells shrank due to fixation with 2% Lugol’s iodine, whether they were mea-

Montagnes et al.

l ThalassioSiT~ w eissJlog~~

0 Chlamydomona= v ChTOOmOnaS salha v Isochtysis galbana

Time (h) Fig. 3. Change in relative cell volume (fixed volume/live volume) over time for four species of phytoplankton fixed

with 2% Lugol’s iodine. The thick solid line represents the mean shrinkage rate (of the four species) from 72 to 2,184 h [slope = 1.3 x 10-3% (of live volume) d-l].

sured by microscope or by Coulter Counter (Table 1, Fig. 2; note the 95% confidence limits do not overlap the 1 : 1 line). For optical measurement, the slopes (b) of the log fixed-cell volume to log live-cell volume were not statistically different from 1 (P > 0.5, models 1 and 2), strongly indicating a linear relation between the untransformed parameters. This finding also held for Coulter Counter measurements (P > 0.05, models 1 and 2); it suggests that there is no size-dependent effect on shrinkage due to fixation for optical measurements and that parameter a can be used to calculate average shrinkage. Thus, due to 2% Lugol’s iodine fixation, for optically measured

, , , , , , , , , l Thalassiosira

t-

w eissy1ogii 0 Chlamydomonas v ChToomonas salin V Isoohrysis galban

0.3 012345

X Concentration of 678910

Lugol’s iodine

Fig. 4. Change in relative cell volume (fixed volume/ live volume) of four species of phytoplankton with chang- ing concentration of fixative.

cells, volume of live cells = 1.33 x (fixed volume).

For the Coulter Counter, if b = 1, live volume = 1.39 x (fixed volume). However, in this case, the exponent b does influence the predicted value (although b is not statistically different from 1 at a = 0.05). For example, if a value near the modal fixed-cell volume (N 300 pm”) is used to suggest a “typical” shrinkage effect, live volume = 1.6 x (fixed volume).

Although there was scatter in the data around the regression lines (Fig. 2), the variation could not be attributed to individual taxa (e.g. dinoflagellates did not shrink more or less than cryptophytes). Surprisingly, the cell volume of diatoms also decreased due to fixation. This decrease was not anticipated, since diatoms have a silica frustule that we expected would be unaffected by preservation. We emphasize that this observation is based on only three species.

The effect of Lugol’s iodine was species-specific; certain species appeared to increase in size after fixation, although microscopic and electronic methods sometimes disagreed. For example, after fixation, the volume of the smallest cell examined (M. pusilla) increased when it was measured optically but decreased when it was measured with the Coulter Coun- ter. M. pusilla is l-2 ,um in diameter, which is near the limit of our ability to make optical measurements (see volume measurements). Since the potential error in volume estimates increases exponentially with decreasing linear measurements, our estimates of optically measured volume at this size may be less precise


0 1 2 3 4

Coul:eDr Count’et-measiied live-Yell voltAFe (pm’) lo5

0 1 2

AttiE CounteY-rneasuitd 3 4

fixedc0ell volu’k lo5

(pm’)

loo lo1 lo2 lo9 10’ Coulter Counter-measured fixed-cell volume (pm’)

lo6

Fig. 5. Log-log relationships between Coulter Counter- and optically measured volumes of live cells, Coulter Counter- and optically measured volumes of fixed cells, and Coulter Counter-measured volumes of fixed cells and optically measured volumes of live cells. For each relationship, measurements were made on 30 phytoplankton species. Symbols and lines as in Fig. 2 and the equations of the lines (as described in the text) and their associated r2 values are shown.

than those made by the Coulter Counter for cells shrank at a nearly constant rate (Fig. 3). certain cells (see volume measurements). Thus, our measurements, made after 2 weeks,

Cells shrank quickly in Lugol’s iodine. All four species examined in time-course experi-

are near the maximum extent of shrinkage due to fixation. But our data may not be valid for

ments shrank substantially over the first 24 h samples preserved for long periods. Although and then shrank at a much reduced rate for some cells are reported to swell immediately the next 2,160 h (Fig. 3). From 72 to 2,184 h, after fixation (Ohman and Snyder 199 1; Bre-


teler 1985) we did not observe this; all four species began to shrink immediately (Fig. 3).

The concentration of Lugol’s iodine used for fixation appeared to have little affect on cell shrinkage based on the four species examined (Fig. 4); the differences between fixed- and live- cell volumes were far greater than the differences between volumes fixed at different concentrations of Lugol’s iodine. When l-l 0% Lugol’s iodine was used, the cells shrank to the maximum extent. Cells fixed with 0.2-0.5% Lugol’s iodine tended to be larger than those fixed at higher concentrations, although this was variable between species. These data indicate that measurements made on cells fixed with 2% Lugol’s iodine (the concentration we selected) are representative of volume changes of cells fixed with l-l 0% Lugol’s iodine.

There is variation in the recipes for making Lugol’s iodine [e.g. the ratio of I2 (g) : KI (g) : glacial acetic acid (ml) : H,O (ml) varies from Throndsen 1978 (1: 2 : 2: 20) to Kimor 1976 (1 : 2 : 20 : 200) to Sigma Chemical Co., Cat. No.L6146(1:2:0:270)]andawiderangein concentrations of Lugol’s iodine used in the literature. To our knowledge, the effects of these different recipes have not been examined; we chose a recipe recommended for phytoplankton (Throndsen 1978). Further, many investigators do not document the concentration used. In fact, many judge concentration only qualitatively (e.g. adding Lugol’s iodine until a “weak tea” color is achieved). We surveyed several phytoplanktologists in the Oceanog- raphy Department at the University of British Columbia and found that the Lugol’s iodine concentrations associated with the colors they picked were O-8-2.0% (mean = 1.2%). Some workers recommend concentrations of 0.2- 0.4% (Throndsen 1978) while others use up to 20% (Gifford 1988). It has also been sug- gested that higher concentrations (15%) re- duce the loss of microzooplankton by lysis (D. K. Stoecker pers. comm.). Although our choice of 2% is perhaps slightly above that normally used, based on the minimal effect of concentration on shrinkage, our conversion factors should be applicable to most studies.

Volume measurements-While both techniques indicated cell shrinkage by fixation (Fig. 2) optically measured cell volumes were almost always greater than those measured by the Coulter Counter (Fig. 5, Table 1). The relation between live optically and live Coulter

Counter-measured cells (Fig. 5A) was size-dependent (i.e. a test of parameter b = 1 gave P < 0.05 for model 1 and P < 0.001 for model 2). For the relations between fixed optically and fixed Coulter Counter-measured cells (Fig. 5B) and live optically and fixed Coulter Coun- ter-measured cells (Fig. 5C), model 1 regressions were size-independent (for both P > 0.2) but both were size-dependent (P < 0.02) for model 2 regressions. If we assume parameter b to be unity for the latter two relations (i.e. using model 1 for predictive purposes), fixed optical volumes are 2.27 x (fixed Coulter Counter volumes), and live optical volumes are 3.03 x (fixed Coulter Counter volumes). However, the relation between live optically and Coulter Counter-measured cells (Fig. 5A) is not so simple. Over the cell volume range of 1 O-l O4 pm3, live optical volume would range from 2.04 to 1.24 x (live Coulter Counter volume).

These three regressions (Fig. 5) indicate a notable difference between optical and Coulter Counter volumes, but they do not distinguish which method is more accurate. One reason for the difference between methods may be a bias in optical measurements. The Coulter Counter was regularly calibrated with 5.1 l- and 43.27~pm-diameter beads. Although the reported bead diameters (Coulter Corp.) agreed with those observed by electron microscopy, they did not agree with those determined by light microscopy (Table 2). Light microscopy overestimated the diameter of the smaller beads by -0.2-0.5 pm. The reason for this is unclear, but we suspect it was due to our inability to distinguish the edge of the beads, due to the halo effect described earlier. The halo effect also seemed to vary somewhat depending on the microscope used; 5.11 -pm beads were observed with a Zeiss Photomicroscope 2 and were determined to be - 5.50 pm in diameter (cf. Table 2). Presumably,the effect seen for beads can be extended to our measurements of phytoplankton, but we believe that the outer layer, used as an outer boundary of most cells, was not the same as the halo. However, we have examined the effect of such an error on estimating cell carbon (see cell composition- volume relationships).

It appears that the common assumption that optically observed values are accurate may be erroneous. This optical distortion could arti- ficially increase the volume of small cells (2

Estimating phytoplankton biomass

Table 2. Comparison of reported diameters (RD, Coulter Corp.) and optically measured diameters (OMD, this study) + 1 SD of four sizes of microspheres from Coul- ter Corp. For each size, 30 beads were measured with the light microscope used to measure phytoplankton cells (see text fir details). Diff = (reported - measured diameter).

RD OMD Microspheres

bm) Diff Lot No. (date)

2.02 2.19kO.16 0.17 11014 (Jul 79) 5.11 5.37f0.10 0.62 5177 (Sep 91) 9.66 10.48kO.56 0.82 5685 (Ott 91)

43.27 43.03kO.19 -0.24 5690 (Mar 91)

pm) by almost a factor of two, while larger cells (20 pm) would be increased by only - 10%. Obviously, the effect of this bias would change over the range of cells we measured; in theory, the slopes of the regressions in Fig. 5 would be significantly < 1. All of the model 2 and one of the model 1 regressions indicated that this was the case (P < 0.02). However, for larger cells (e.g. ~20 pm), presumably more accurately estimated under the microscope, this optical distortion does not fully explain the difference between Coulter Counter and optically determined volumes; other factors must also have influenced volume estimates.

Another source of error is associated with the Coulter Counter volume estimates. Theory indicates that not only is the volume of nonspherical cells underestimated by the algo- rithms used by electronic particle counters, but other factors, including cell compressibility and cell conductivity, can also alter Coulter Coun- ter measurements (Grover et al. 1982). Un- derestimation of actual cell size has also been demonstrated empirically (e.g. the Coulter Counter volume of the ciliate Tetrahymena was 0.72-0.63 of the cell size measured visually, Nilsson 1990).

For our data, Coulter Counter measurements of elongate cells could have underestimated the true cell volume by a factor -0.7 (using a length : width ratio of 3 : 1 and the relation provided by Grover et al. 1982). Our data also support this “shape-factor” effect: the equivalent spherical diameter calculated from Coulter Counter-determined volume was compared with optically determined cell dimensions. There appeared to be a 1 : 1 relationship between the short axis of the cell and the diameter determined by Coulter Counter, but this was not the case for the long axis (Fig. 6). The Coulter Counter TAII measures vol-

7 60

-3

10 20 30 40 50 60

Optically measured linear dimension (pm)

Fig. 6. Relationships between the equivalent spherical diameter determined by Coulter Counter and optically determined linear dimensions of 30 phytoplankton species. Cell diameters-o; widths-O; lengths-Cl. The line reaching from corner to corner represents unity. Note that most of the circles lie near the line but all the squares are below the line.

ume from the peak amplitude of an electronic pulse generated as the cell passes through the aperture (Kubitschek 1987). Our data support the notion that cells were drawn into the aperture with their short axis perpendicular to it, so that only this axis determined the volume. Thus, the volume estimate would be a function of only the short axis length; therefore, the true volume would likely be underestimated.

The combination of bias due to geometry and the previously mentioned optical-size dis- crepancy was likely the cause of the difference between the two methods. Because the optical measurements of beads were variable and biases due to particle counters may vary between instruments, models, and makes, we have not offered a universal correction factor. Instead, we support other studies that caution against using electronic particle counters to provide accurate volume estimates (e.g. Grover et al. 1982; Kubitschek 1987; Nilsson 1990). We believe that optical measurements best estimate volume and therefore use them in the following analyses.


Table 3. Comparison of the parameters of the formula y = axb from several sources, where y is C, N, protein or Chl a (pg cell-l); x is cell volume (pm’); and a and b are constants. The taxa of cells used in each study are represented as marine or freshwater and as diatoms (D), flagellates (F), and cyanobacteria (C).

Material a b SOWCe

y = carbon

Marine: D, F Marine: D Marine: F, D Marine: F, C Marine: F, D Freshwater: D, F, C Freshwater: D, F, C

y = nitrogen

Marine: F, C Marine: F, D Marine: F, D

y = protein

Marine: F, C Marine: F, D

y=Chla

Marine: F, C Marine: F, D

0.513 0.75 0.485 0.712 0.396 0.829 0.436 0.863 0.109 0.991 0.120 1.051 0.210 1.041

0.0652 0.838 0.0824 0.837 0.0172 1.023

0.489 0.802 0.043 1.058

0.00892 0.722 0.00429 0.917

Mullin et al. 1966 Strathmann 1967 Moal et al. 1987 Verity et al. 1993 This study Roacha and Duncan 1985 Chalk 198 1 cited IIJJ Rocha

and Duncan 1985

Moal et al. 1987 Verity et al. 1993 This study

Moal et al. 1987 This study

Moal et al. 1987 This study

Cell composition and culture conditions- For many of the species we used, there are no detailed measurements with which to compare our volume and cell composition values. How- ever, two common indicators of culture condition (C : Chl and C : N) suggest our cultures were healthy. The mean C : Chl ratio was 40.7, which is well within the range of values found for healthy cells growing with excess nutrients (Banse 1977). Also, on an atom : atom basis, the mean C : N ratio was 6.25, which is close to the Redfield ratio of 6.6 and is well within the typical range of healthy phytoplankton (Parsons et al. 1979).

We grew our cultures on a light-dark cycle; many other studies used 24 h of light. Further, we used subsaturating light intensities, which may better mimic conditions often experi- enced in situ. These two factors may make our cell quota estimates more like those found in many natural situations, and our data may be more applicable to field studies where light is limiting and nutrients are not. However, carbon and volume may be lower for phytoplankton grown on subsaturating light (see Thomp- son et al. 199 l), and this may partially explain why our values are different than others.

Cell composition-volume relationships-For composition-volume relationships, we selected the optically measured cell volumes, which

we believe best represent the true volume. Log- log regressions between optically measured live-cell volume and C, N, protein, or Chl a cell-l were all linear, with slopes equal to 1 (i.e. b was not significantly different from 1, P > 0.1 for model 1 regressions; Fig. 7, Table 1). Once again, there was scatter in the data for these estimates, but the variation did not appear to be taxa related. Surprisingly, the two smaller diatom quotas (Fig. 7, solid circles) did not appear to be different from those predicted for “soft-bodied” cells (note that the dinoflagellates we examined were athecate). This finding is contrary to the dogma that diatoms are to a large extent composed of an “empty” vac- uole (Strathmann 1967).

Hitchcock (1982) showed nonlinear relationships between diatom chlorophyll and protein vs. cell volume and clear differences between diatoms and dinoflagellates, while Moal et al. (1987) did not. Our data agree with Moal et al., but we were examining smaller species than Hitchcock, possibly there are more pronounced effects in larger species. For in- stance, the large diatom Detonulla pumila did appear to fall below predicted values in all four relationships (Fig. 7, solid circle of volume - lo4 pm3); this suggests a low amount of cy- toplasm per cell. Because we used a small number of diatom species, it is also possible that


I = lo3 : ha lo2 PC G 10' & $1, loo

15

r = 0.954 1

2 1o-l

- lo4 rl I i=i :

lo3

2 lo2 V

.Ef 10' Q

4 : loo

PI w

‘+ + ,.-n 1.058 7

r( I = JO2

al o 10'

F n m-51 F JJ

Y R 10-l

A

0" h 1o-2 GjF I I III

w - 0.917 1 y=o.o0429x

I I111111 1 I I ,,,*,I 1 , , ,111, l < -A d

8 1 O0 10' lo2 lo3 lo4 lo5

Live (optically measured) cell volume (pm3) Fig. 7. Log-log relationships between four cellular constituents and live-cell volume. Symbols and lines as in Fig.

2 and the equations of the lines (as described in the text) and their associated r2 values are shown.

size-dependent relationships were eclipsed by the larger number of other species included in the regression. More work is obviously required in this area.

Many field studies on phytoplankton require an accurate carbon-volume relationship. The relationship presented here differs from earlier ones. In fact, there is little agreement among

workers (Fig. 8A, Table 3). Our relation is close to that obtained by Rocha and Duncan (1985) for freshwater species, but their slope (b) is greater than ours; thus, their carbon estimates for large cells are substantially greater than ours (Fig. 8A, Table 3). Rocha and Duncan (1985) fixed some of their cells with Lugol’s iodine but measured others live. They also used data


Fig. 8. A. Comparison of our carbon-volume relationships with others (see Table 3 for the associated equations). The lines are from the following sources: 1 - Rocha and Duncan 1985; 2-Verity et al. 1993; 3-this study; 4-Moal et al. 1987; 5-Mullin et al. 1966; and 6-Strath- mann 1967 for diatoms. B. Comparison of our data to that of Verity et al. (1993) after some manipulation of both data sets: 1 -equation of Verity et al.; 2-equation of Verity et al. assuming cells shrank by 28%; 3 -equation of Verity et al. assuming cells shrank by 28% and were 0.5 wrn larger than measured; 4-the equation from this study; 5 -the equation from this study assuming cells were 0.5 pm smaller than measured. (See text for details.)

from other sources and did not report whether the cells were live or fixed. Lugol’s iodine shrinks cells, so this may have caused the differences between our predictions and theirs.

Measurements by Parsons et al. 196 1 (used by Strathmann 1967) were made on live cells (T. R. Parsons pers. comm.); those made by Mullin et al. (1966) (also used by Strathmann 1967) were likely made on cells fixed with either Lugol’s or Rodhe’s iodine (M. M. Mullin pers. comm.). This mixing of live- and fixed- cell volumes means that estimates of carbon from cell volume presented by Mullin et al. (1966) and Strathmann (1967) may be biased.

Our carbon to cell volume relationship also differs from one established by Verity et al. (1993) (Fig. 8B, line 4 vs. 1). They used a num-

ber of small marine species; thus, we might have anticipated similar results. This discrep- ancy could come from two sources: differences in volume estimates and (or) differences in carbon quotas.

For volume estimates, we have examined two reasons for the difference between our relationship and that of Verity et al. (1993): shrinkage due to fixation and a potential error of 0.5 pm in measuring linear dimensions (see above). The phytoplankton measured by Ver- ity et al. were fixed with 0.5% glutaraldehyde which shrank the cells by 28%; the equation they present for C : cell volume is based on fixed cells and is therefore not strictly comparable with others who used live cell volumes. We have recalculated their equation for live cells (Fig. 8B, line 2). It should be noted that their light microscope image system was de- signed to analyze fluorescing cells and was calibrated with fluorescent beads. It is possible that their system did not “see” the clear outer region of many cells (see above), and is therefore not comparable to our bright field-based analysis. We have also shown that light microscopy can give inaccurate estimates of diameter of up to 0.5 pm. We have recalculated the equation of Verity et al. for live cells that are 0.5 pm larger in diameter, based on spherical volumes (Fig. 8B, line 3). Alternatively, it may be that we have overestimated linear dimensions by 0.5 pm; we have also recalculated our equation, assuming that we overestimated linear dimensions by 0.5 pm, based on data in Table 1 (Fig. 8B, line 5). After correcting for these potential biases, our relationship be- comes more similar to that of Verity et al. (see Fig. 8B).

Carbon estimates are another source of variation in the relationship. Our data set and that of Verity et al. (1993) have four species in common: Dunaliella tertiolecta, Pavlova luth- eri, Heterosigma carterae [=Olithodiscus lu- teus (clone OLNB) = Heterosigma akashiwo], and Emiliania huxleyi. Despite the numerous potential errors in volume estimation dis- cussed above, the volume data for these species are remarkably close in the two studies (Table 4). However, our carbon cell quotas are sometimes two to four times lower than the data that Verity et al. present. It is unlikely that errors in methodology account for these differences because our techniques for carbon

Tabl

e 4.

C

ompa

rison

of

vo

lum

e an

d ca

rbon

es

timat

es

avai

labl

e in

the

lite

ratu

re

for

sele

cted

ph

ytop

lank

ton

spec

ies.

,G

row

th

cond

ition

s of

irr

adia

nce

(irra

d;

iLm

oI p

hoto

n m

-*

s-l),

da

y le

ngth

(L

: D

), an

d te

mpe

ratu

re

are

pres

ente

d.

As w

ell,

met

hods

of

cel

l vo

lum

e m

easu

rem

ent

[V

met

hod:

m

icro

scop

e (M

), C

ouhe

r C

ount

er

(CC

), or

im

age

anal

ysis

sy

stem

(IA

)] an

d ca

rbon

co

nten

t [C

met

hod:

w

et o

xida

tion

(WO

) or

CH

N

anal

yzer

(C

HN

)] ar

e do

cum

ente

d.

(n.d

. -n

ot

dete

rmin

ed.)

Snec

ies

Sour

ce

Tem

n “C

Irr

ad

L:D

Vo

l ur

n3

V M

etho

d C

DE

cell-

l C

Met

hod

Not

es

Dun

alie

llu

terti

olec

ta

Pavb

va

Iuth

eri

Het

eros

igm

a ca

rtera

e

Emiliu

nia

huxl

eyi

Eppl

ey

and

Sloa

n 19

65

21

210

24:0

n.

d.

Mul

lin

et a

l. 19

66

10 o

r 21

va

rious

24

:0

300-

316

Law

s an

d W

ong

1978

?

270

12:ll

n.

d.

Moa

l et

al.

1987

15

10

0 24

:0

230

Bro

wn

1991

20

70

-80

12:ll

29

5 Th

omps

on

et a

l. 19

9 1

17.5

10

-200

24

:0

173-

227

Thom

pson

et

al.

1992

1 O

-25

220

24:0

18

0-22

2 Ve

rity

et a

l. 19

93

21

200

14: 1

0 49

1 pr

esen

t st

udy

16

25-5

0 14

: 10

285

pres

ent

stud

y 16

25

-50

14:lO

19

5.9

Pars

ons

et a

l. 19

61

18

120

24:0

28

La

ws

and

Won

g 19

78

? 27

0 12

:ll

n.d.

Br

own

1991

20

70

-80

12:ll

91

Th

omps

on

et a

l. 19

91

17.5

10

-200

24

:0

32-6

5 Th

omps

on

et a

l. 19

92

lo-2

5 22

0 24

:0

33-4

8 Ve

rity

et a

l. 19

93

21

200

14:lO

73

pr

esen

t st

udy

16

25-5

0 14

:lO

73.6

pr

esen

t st

udy

16

25-5

0 14

: 10

24.5

Verit

y et

al.

1993

21

20

0 14

: 10

903

Thom

pson

et

al.

1991

17

.5

lo-2

00

24:0

49

1-1,

763

pres

ent

stud

y 16

25

-50

14: 1

0 1,

096

pres

ent

stud

y 16

25

-50

14: 1

0 55

8.91

Eppl

ey

and

Sloa

n 19

65

21

210

Mul

lin

et a

l. 19

66

10 o

r 21

va

rious

Fl

ynn

1990

18

80

m

yM

1990

18

80

D

. M

uggl

i un

publ

. 16

50

Ve

rity

et a

l. 19

93

21

200

pres

ent

stud

y 16

25

-50

pres

ent

stud

y 16

25

-50

24:o

24

:0

14: 1

0 14

: 10

14: 1

0 14

: 10

n.d.

14

.2-2

0.6

24.8

9.

14

n.d.

70

69

.3

25

- M

orC

C

- cc

M

cc

cc

IA

M

cc

cc

- M

cc

cc

IA

M

cc

IA

cc

M

cc - cc

M

M - IA

M

CC

39.4

wo

33-5

2 wo

19.7

-34.

1 C

HN

31

.3

CH

N

n.d.

-

40.2

C

HN

45

-55

CH

N

63.2

C

HN

41

.7

CH

N

41.7

C

HN

18.4

9.

3-13

.1

n.d.

15

.1

11-1

6.8

16.5

8.

55

8.55

wo

CH

N

CH

N

CH

N

CH

N

CH

N

CH

N

192

CH

N

119-

429

CH

N

87.8

C

HN

87

.8

CH

N

4.3

3.7-

4.1

n.d.

n.

d.

13-2

0 21

.7

4.57

4.

57

wo

wo - -

CH

N

CH

N

CH

N

CH

N

pres

erve

d ce

lls

varie

d N

lim

itatio

n

varie

d lig

ht

varie

d te

mpe

ratu

re

2 2.

3 a (g

‘a

varie

d N

lim

itatio

n s-

s b va

ried

light

F

varie

d te

mpe

ratu

re

5 s SY

s 5

varie

d lig

ht

pres

erve

d ce

lls

with

mcc

oli~

s na

ked

with

coc

colit

hs

nake

d K

nake

d 4


were very similar. Verity et al. were more rig- orous in correcting for background carbon and bacterial contamination, but if these biases occurred in our procedures, we would expect to find higher carbon quotas than they did. In Table 4, we have compared our estimates and those of Verity et al. to other estimates for these species found in the literature.

For the chlorophyte D. tertiolecta, there is reasonable agreement for both volume and carbon data despite a huge range of growth conditions and different volume and carbon methodologies (Table 4). The values of Verity et al. (1993) tend to be higher than average. D. tertiolecta is noted for its ability to grow well under extreme conditions; this may be reflected in the constancy of size and cell composition under a variety of conditions (Thompson et al. 199 1).

Data for P. Zutheri are much less consistent. The volume of P. Zutheri found in our study and that of Verity et al. (1993) are nearly iden- tical, yet our carbon value is half of theirs. Parsons et al. (196 1) grew this species on continuous saturating irradiance and found cells with carbon contents near those observed by Verity et al. but with a much smaller volume (28 pm3). Parsons et al. (196 1) used a Coulter Counter to determine cell volume; their results are so similar to our Coulter Counter volume (24.5 pm3) that we believe their volume to be underestimates of optically determined volumes. Laws and Wong (1978) found that nitrogen limitation changed the carbon content markedly; the lower values of carbon quota they reported are similar to ours. Thompson et al. (199 1) showed that cell volume and carbon quota in this species generally increased with irradiance. Other workers whose carbon values are higher than ours also used much higher light levels. Similarly, Thompson et al. (1992) showed increases in cell carbon with temperature increases from 15 to 20°C. We note that most investigators have selected higher temperatures as well (e.g. 2O”C, Verity et al. 1993). We believe these differences in condition may explain the lower carbon quotas obtained in this study.

The arguments for P. Zutheri apply to H. carterae as well. We and Verity et al. (1993) both found a similar cell volume, but our carbon quota for this species is less than half of theirs. Thompson et al. (199 1) also found wide

variability in cell volumes and cell quotas for this species and related these with irradiance.

For the fourth species in common, E. hux- Zeyi, there are unique considerations. This species forms coccoliths in nature but frequently ceases to produce them in culture. Coccoliths are composed of CaCO,, so they can significantly affect the C : volume ratio. Workers frequently neglect to mention whether their cells are coccolith bearing. In Table 4, it is clear that there is wide variation in both carbon and volume for this species. The studies of Eppley and Sloan (1965), Mullin et al. (1966), Verity et al. (1993), and ours all used naked cells. Our volumes and those of Verity et al. are similar, but our carbon quotas are much lower. Our carbon quotas are very similar to those given by Mul- lin et al. and Eppley and Sloan. The cells used by Mullin et al. appear to have a smaller volume (- 14-20 E.cm3), but this is probably a bias due to the Coulter Counter used; our Coulter Counter volume was 25 pm3 for this species. Typically, coccolith-bearing species in culture at the NEPCC have a substantially higher carbon content (D. Muggli perscomm.), and ac- cording to Flynn (1990), who used microscope measurements, these coccolith-bearing cells are nearly three times larger than naked cells. Such differences are particularly important because E. huxleyi has been one of the smaller species measured in several previous studies and will be particularly influential in these carbon-volume regressions. To further complicate the is- sue, coccoliths dissolve in acid Lugol’s preser- vative. We are uncertain what effect glutaraldehyde may have. Further work should be done with coccolithophorids to resolve these difficulties in C : volume relationships.

Thus, fixation, measurement methodologies, and culture conditions apparently affect estimates of C : cell volume. Note that all the equations predict a similar C : volume ratio for cells -20-l ,200 pm3 (Fig. 8A). Our data, those of Verity et al. (1993), and those of Rocha and Duncan (1985) were primarily from cells 1 02- 1 O3 pm3 in volume; those of Mullin et al. (1966) and Strathmann (1967) considered cells primarily 1 03-lo6 pm3 in volume. Thus, part of the difference between regressions may be due to a nonlinear relation between log cell volume and log carbon cell-l, which is seen only over a large range of cell volume. Again, more work needs to be done in this area.

Recommendations

Our data appear to give reasonable estimates of C, N, protein, and Chl a from volume for cells measured with a bright-field light microscope and ocular micrometer (Table 3). We suggest that for small cells and for cells growing under lower light (20-60 pmol photons m-2 s-l) conditions these are better estimates than those previously established. However, we caution that there are large variations in light, temperature, and nutrient limitation in nature which we believe will change the carbon-volume relationship. Attention should be focused on determining how these factors affect the carbon to volume relationships. There is a large uncertainty associated with carbon estimates from volume if all published relationships are considered (Fig. 8A), but if only our data and those of Verity et al. (1993) are considered, the variation is less (Fig. 8B). This potential error should be considered when applying carbon- volume conversions to field studies.

Phytoplankton shrink due to fixation with Lugol’s iodine. For field studies with mixed populations, we recommend that the conversion factors presented above be used to correct for shrinkage. However, the shrinkage effect can be species-specific. For laboratory studies on cultured species, we recommend that researchers determine the effect of fixation on the particular organism used.

Model TAII Coulter Counters (and other machines that use the same principle) apparently underestimate phytoplankton cell volume. Whether to use an electronic particle counter or measure cells visually is not a simple choice. Optical measurements are un- doubtedly more accurate (if sufficient data are collected), but they are time consuming. We suggest that electronic particle counters are not good tools for estimating absolute volume, as cell shape likely alters volume measurements. Some researchers may continue to use electronic particle counters to determine phytoplankton biomass from field samples. We draw their attention to the potential underestima- tion of cell volume we have demonstrated, but we suggest that a single correction factor is not useful, since it will be applicable only to a particular set of conditions.

We suggest that five areas require attention: determining whether all diatoms shrink due to


fixation; re-examining the belief that diatoms are distinct from other species in terms of carbon-volume relationships; further examining carbon-volume relationships for coccolithophorids; investigating the nonlinearity of cell quota-volume relationships over a large range of volumes; and finally, systematically examining the effect of growth conditions (light and temperature) on cell quota-volume relationships.

References BANSE, K. 1977. Determining the carbon to chlorophyll

ratio of natural phytoplankton. Mar. Biol. 41: 199- 212.

BERGES, J. A., A. E. FISHER, AND P. J. HARRISON. 1993. A comparison of Lowry, Bradford, and Smith protein assays using different protein standards and protein isolated from the marine diatom Thalassiosira pseudonana. Mar. Biol. 115: 187-193.

BOOTH, B. C. 1987. The use of autofluorescence for an- alyzing oceanic phytoplankton communities. Bot. Mar. 30: 101-108.

BORSHEIM,K.Y.,ANDG.BRATBAK. 1987. Cellvolume to carbon conversion factors for a bacterivorous Mo- nas sp. enriched from seawater. Mar. Ecol. Prog. Ser. 36: 171-175.

BRADFORD, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein using the principle of protein-dye binding. Anal. Biochem. 72: 248-254.

BRETELER, W. C. M. K. 1985. Fixation artifacts of phytoplankton in zooplankton grazing experiments. Hy- drobiol. Bull. 19: 13-l 9.

BROWN, M. R. 199 1. The amino-acid and sugar composition of 16 species of microalgae used in maricul- ture. J. Exp. Mar. Biol. Ecol. 145: 79-99.

CHOI,J. W., AND D. IL STOECKER. 1989. Effects offix- ation on cell volume of marine planktonic protozoa. Appl. Environ. Microbial. 55: 1761-1765.

CLARKE, M. R. B. 1980. The reduced major axis of a bivariate sample. Biometrika 67: 44 l-446.

COULTER ELE~RONICS. 1979. Coulter counter model TAII operator’s manual 420 10440/May 1979. Coul- ter Counter Inc.

D~RTCH, Q.,J. R. CLAYTON, JR., S. S. THORESEN, AND S. I. AHMED. 1984. Species differences in accumulation of nitrogen pools in phytoplankton. Mar. Biol. 81: 237-250.

EPPLEY, R. W., AND P. R. SLOAN. 1965. Carbon balance experiments with marine phytoplankton. J. Fish. Res. Bd. Can. 22: 1083-1097.

FLYNN, K. J. 1990. Composition of intracellular and extracellular pools of amino acids, and amino acid utilization of microalgae of different sizes. J. Exp. Mar. Biol. Ecol. 139: 15 1-166.

GIFFORD, D. J. 1988. Impact of grazing by microzooplankton in the Northwest Arm of Halifax Harbour, Nova Scotia. Mar. Ecol. Prog. Ser. 47: 249-258. -

GROVER,N.B.,S.A. BEN-SASSON,ANDJ.NAAMAN. 1982.


Electrical sizing of cells in suspension, p. 93-109. In N. Catsimpoolis [ed.], Cell analysis. Plenum.

GUILLARD, R. R. L. 1988. The Centre for culture of marine phytoplankton: History, structure, function and future. J. Protozool. 35: 255-256.

HARRIS, G. P. 1986. Phytoplankton ecology. Structure, function and fluctuation. Chapman and Hall.

HARRISON, P. J., R. E. WATERS, AND F. J. R. TAYLOR. 1980. A broad spectrum artificial seawater medium for coastal and open ocean phytoplankton. J. Phycol. 16: 28-35.

HITCHCOCK, G. L. 1982. A comparative study of the size- dependent organic composition of marine diatoms and dinoflagellates. J. Plankton Res. 4: 363-377.

JEROME, C. A., D. J. S. MONTAGNES, AND F. J. R. TAYLOR. 1993. The effect of the quantitative protargol stain (QPS) and Lugol’s and Bouin’s fixatives on cell size: A more accurate estimate of ciliate biomass. J. Euk. Microbial. 40: 254-259.

KELLER, M. D., R. C. SELVIN, W. CLAUS, AND R. R. L. GUILLARD. 1987. Media for the culture of oceanic ultraphytoplankton. J. Phycol. 23: 633-638.

KIMOR, B. 1976. The fixation and preservation of marine protozoa: Some problems and general considerations, p. 231-235. In H. F. Steedman [ed.], Zooplankton fixation and preservation. UNESCO.

KUBITSCHEK, H. E. 1987. Buoyant density variation during the cell cycle in microorganisms. CRC Crit. Rev. Microbial. 13: 78 l-799.

LAWS, E. A., AND D. C. L. WONG. 1978. Studies of carbon and nitrogen metabolism by three marine phytoplankton species in nitrate-limited continuous culture. J. Phycol. 14: 406-4 16.

LI, W. K. W. 1987. Experimental approaches to field measurements: Methods and interpretation. Can. Bull. Fish. Aquat. Sci. 214: 251-286.

MOAL, J., V. MARTIN-JEZEQUAL, R. P. HARRIS, J.-F. SA- MAIN, AND S. A. POULET. 1987. Interspecific and intra-specific variability of the chemical composition ofmarine phytoplankton. Oceanol. Acta 10: 339-346.

MULLIN, M. M., P. R. SLOAN, AND R. W. EPPLEY. 1966. Relationship between carbon content, cell volume, and area in phytoplankton. Limnol. Oceanogr. 11: 307-311.

NILSSON, J. R. 1990. On the correlation of the volume of a prolate-spheroidal cell, Tetrahymena, as determined by electronic particle counters and morphom- etry. J. Protozool. 37: 500-504.

OHMAN, M. D., AND R. A. SNYDER. 1991. Growth ki- netics of the omnivorous oligotrich ciliate Strombidi- um sp. Limnol. Oceanogr. 36: 922-935.

OLMERI, E. T. 1985. Feasibility of estimating phytoplankton size and biomass in fresh and preserved samples from the Benguela Current with a Coulter Counter. S. Afr. J. Mar. Sci. 3: 99-l 10.

PARSONS, T. R., Y. MAITA, AND C. M. LALLI. 1984. A manual of chemical and biological methods for seawater analysis. Pergamon.

-, K. STEPHENS, AND J. D. H. STRICKLAND. 196 1. On the chemical composition of eleven species of marine phytoplankters. J. Fish. Res. Bd. Can. 18: 100 l- 1016.

-, M. TAKAHASHI, AND B. HARGRAVE. 1979. Bio- logical oceanographic processes, 2nd ed. Pergamon.

PRXE, N. M., P. A. THOMPSON, AND P. J. HARRISON. 1987. Selenium: An essential element for growth of the coastal marine diatom Thalassiosira pseudonana (Ba- cillariophyceae). J. Phycol. 23: l-9.

RICKER, W. E. 1984. Computation and use of central trend lines. Can. J. Zool. 62: 1897-1905.

ROCHA, O., e A. DUNCAN. 1985. The relationship between cell carbon and cell volume in freshwater algal species used in zooplankton studies. J. Plankton Res. 7: 279-294.

SHELDON, R. W. 1978. Sensing-zone counters in the laboratory, p. 202-2 14. Zn A. Sournia [ed.], Phytoplank- ton manual. Page.

-, AND T. R. PARSONS. 1967. A practical manual on the use of the Coulter Counter in marine science. Coulter Electronics Sales Co.

&N-DA, T. J. 1978. What to count, p. 165-166. Zn A. Sournia [ed.], Phytoplankton manual. Page.

SOURNIA, A. 1978. Phytoplankton manual. Page. STEEN, H. B. 1990. Flow cytometry instrumentation, p.

3-30. Zn S. D. Demers [ed.], Particle analysis in oceanography. Springer.

STRATHMANN, R. R. 1967. Estimating the organic carbon content of phytoplankton from cell volume or plasma volume. Limnol. Oceanogr. 12: 411-418.

THOMPSON, P. A., M. Guo, AND P. J. HARRISON. 1992. Effects of variation in temperature. 1. On the bio- chemical composition of eight species of marine phytoplankton. J. Phycol. 28: 481-488.

-, P. J. HARRISON, AND J. S. PARSLOW. 1991. In- fluence of irradiance on cell volume and carbon quota for ten species of marine phytoplankton. J. Phycol. 27: 35 l-360.

THRONDSEN, J. 1978. Preservation and storage, p. 69- 74. Zn A. Sournia [ed.], Phytoplankton manual. Page.

VERITY, P. G., ANI> OTHERS. 1993. Relationships between cell volume and carbon and nitrogen of marine pho- tosynthetic nanoplankton. Limnol. Oceanogr. 37: 1434-1446.

Submitted: 29 January 1993 Accepted: 13 December 1993 Amended: 18 January 1994

Documents

Limnol. Oceanogr, 39(5), 1994, 1044-I 060 0 1994, by the ...pc › ~pelagic › s › Montagnes 1994 Limn Ocean.… · Limnol. Oceanogr, 39(5), 1994, 1044-I 060 0 1994, by the American