Prasinophyceae)1

Home , Micromonas, Prasinophyceae, Thalassiosira pseudonana

J. Phycol. 38, 1132–1142 (2002)

DIEL VARIATIONS IN OPTICAL PROPERTIES OF MICROMONAS PUSILLA (PRASINOPHYCEAE)1

Michele D. DuRand,2 Rebecca E. Green, Heidi M. Sosik,3 and Robert J. Olson Woods Hole Oceanographic Institution, Biology Department, MS #32, Woods Hole, Massachusetts 02543, USA

Micromonas pusilla (Butcher) Manton et Parke, a nitrogen ratio; Chli, intracellular chlorophyll concen- marine prasinophyte, was used to investigate how tration; CV, coefficient of variation; D , the diameter cell growth and division affect optical properties of of the mean cell; FLS, forward light scattering; G , phytoplankton over the light:dark cycle. Measurements geometric projected area of the mean cell; , spectral were made of cell size and concentration, attenua- slope; , growth rate; n, real refractive index; n, imag- tion and absorption coefficients, flow cytometric inary refractive index; a, absorption cross-section; forward and side light scattering and chl fluores- b, scattering cross-section; c, attenuation cross-sec- cence, and chl and carbon content. The refractive in- tion; Qa, efficiency factor for absorption; Qb, effi- dex was derived from observations and Mie scattering ciency factor for scattering; Qc, efficiency factor for theory. Diel variations occurred, with cells increasing attenuation; V , volume of the mean cell in size, light scattering, and carbon content during daytime photosynthesis and decreasing during nighttime division. Cells averaged 1.6 m in diameter and exhibited phased division, with 1.3 divisions per day. Scattering changes resulted primarily from changes introduction in cell size and not refractive index; absorption Diel variations in bulk optical properties such as changes were consistent with a negligible package ef- beam attenuation (Siegel et al. 1989, Hamilton et al. fect. Measurements over the diel cycle suggest that 1990, Cullen et al. 1992, Stramska and Dickey 1992, in M. pusilla carbon-specific attenuation varies with Gardner et al. 1993, 1995) and in single cell optical cell size, and this relationship appears to extend to properties such as phytoplankton forward light scat- other phytoplankton species. Because M. pusilla is tering (FLS) (Olson et al. 1990, DuRand and Olson one of the smallest eukaryotic phytoplankton and be- 1996, Vaulot and Marie 1999) have been observed in longs to a common marine genus, these results will numerous field studies in many of the world’s oceans. be useful for interpreting in situ light scattering vari- A number of researchers have documented diel varia- ation. The relationship between forward light scat- tions in cell light scattering during laboratory studies tering (FLS) and volume over the diel cycle for M. of various phytoplankton cultures (Stramski and Rey- pusilla was similar to that determined for a variety of nolds 1993, Stramski et al. 1995, DuRand and Olson phytoplankton species over a large size range. We 1998). The typical pattern exhibited is scattering min- propose a method to estimate cellular carbon con- ima at dawn and maxima near dusk. Diel variations in tent directly from FLS, which will improve our esti- beam attenuation have been related to primary pro- mates of the contribution of different phytoplank- duction (Siegel et al. 1989, Cullen et al. 1992, Walsh ton groups to productivity and total carbon content et al. 1995) and have been proportioned among in the oceans. groups of organisms in the ocean (DuRand 1995, Du- Key index words: absorption; attenuation; diel varia- Rand and Olson 1996, Chung et al. 1998, Binder and tion; flow cytometry; Micromonas pusilla; optical DuRand 2002). properties; phytoplankton; Prasinophyceae; refractive Light scattering by particles such as phytoplankton index; scattering is determined by their size and refractive index and, to a lesser extent, by their shape and internal struc- Abbreviations: a, absorption coefficient; a*chl, chl- tures ( Jerlov 1976). If diel changes in scattering cross- specific absorption coefficient; ac-9, absorption and sections can be related to changes in cell size and attenuation meter; b, scattering coefficient; b*chl, chl- carbon content, then phytoplankton productivity specific scattering coefficient; c, attenuation coeffi- (Cullen et al. 1992) and growth rates (DuRand 1995, cient; c*c , carbon-specific attenuation coefficient; Ci, Binder et al. 1996) can be estimated from measure- intracellular carbon concentration; C:N, carbon to ments over the diel cycle in the field. In laboratory studies, contrasting data have been obtained on the relative contributions of variations in particle size and in refractive index to diel changes in optical cross- 1Received 24 January 2002. Accepted 22 June 2002. sections. Changes in refractive index have been found 2Present address: Ocean Sciences Centre, Memorial University of Newfoundland, St. John’s, NL A1C 5S7, Canada. E-mail mdurand@ to be equal to or more important than changes in cell mun.ca. size for determining the attenuation cross-section of a 3Author for correspondence: e-mail [email protected]. diatom (Stramski and Reynolds 1993) and a prokary-

1132

M. PUSILLA DIEL OPTICAL PROPERTIES 1133 otic picoplankter (Stramski et al. 1995). In contrast, materials and methods DuRand and Olson (1998) reported that changes in cell size were more important than those in refractive Laboratory measurements. Micromonas pusilla (CCMP 489), a strain isolated from the Sargasso Sea, was grown in f/2 medium index in determining diel variations in beam attenua- without silica (Guillard 1975) in replicate batch cultures at 25 C. tion for a small chlorophyte. The filtered (0.22 m) seawater and nutrient solutions were au- Estimates of primary production in the ocean have toclaved separately and then added together after the seawater had cooled. To ensure particle-free medium, the solution was been made from diel variations in beam attenuation by converting changes in beam attenuation to carbon pro- then sterile filtered through a 0.22- m filter. The cultures were grown in 10-L carboys (containing 2 L of culture at the start of duction using a constant factor (Siegel et al. 1989, sampling) in an incubator on a 12:12-h light:dark cycle at 120 Cullen et al. 1992, Walsh et al. 1995). Ackleson et al. mol photonsm2s1 (cool-white fluorescent lights, measured (1993), however, reported carbon-independent changes with a QSL-100 4 quantum scalar irradiance sensor, Biospher- in cell scattering. In addition, several other laboratory ical Instruments, San Diego, CA, USA). The carboys were bub- bled with moisturized filtered air. Samples were forced out by studies on cultures have found that the carbon-specific air pressure through a sterile sampling port. The cultures were beam attenuation varies by 25%–30% over the diel cycle acclimated to the growth conditions for 7 days and diluted daily (usually increasing during the day) and mean values for 5 days before the day of sampling (with the last dilution 24 among the species studied vary by 21% (Stramski and h before the start of the sampling) to maintain the cells in ex- Reynolds 1993, Stramski et al. 1995, DuRand and Olson ponential growth. Samples were withdrawn every 2 h (starting at dawn, until 1998). It may be necessary to know the composition of the following dawn) and measurements made of cellular and the phytoplankton community to best interpret diel bulk properties (individual cell measurements of concentra- variations in beam attenuation as production (DuRand tion, forward and side light scattering, chl fluorescence, and and Olson 1996, Binder and DuRand 2002). cell size, and bulk measurements of absorption, attenuation, extracted chl content, and carbon content). Samples were pro- Stramski (1999) suggested that the strong correla- cessed in dim light during the nighttime sampling points. tion between intracellular carbon concentration and A modified Epics V flow cytometer (Beckman Coulter, Mi- real refractive index and between intracellular chl con- ami, FL, USA) with a Cicero acquisition interface (Cytomation, Fort Collins, CO, USA) was used to measure FLS (3–19 at 488 centration and the imaginary refractive index, observed for phytoplankton cultures, could be used to es- nm), side angle light scattering ( 54–126 at 488 nm), and chl fluorescence (660–700 nm) of the M. pusilla cells in undiluted timate carbon and chl content of individual cells in duplicate samples. Polystyrene microspheres of two sizes (0.66 natural water samples. Despite the fact that the deriva- m and 2.14 m, YG calibration-grade beads from Poly- tion of refractive index from single-particle analysis is sciences, Warrington, PA, USA) were added as internal stan- not currently routine, the results of such an approach dards and differentiated from cells by their orange fluorescence (555–595 nm). The samples were introduced into the are appealing because knowledge of the carbon-based flow cytometer using a peristaltic pump (Harvard Apparatus, size distribution of the plankton is crucial to an under- Holliston, MA, USA), and cell concentration was determined standing of upper ocean carbon cycling and knowl- from the pump flow rate and sample run time. Mean values of the measured parameters normalized to reference beads (2.14 edge of chl per cell is important for studies of phytoplankton productivity and for bio-optical modeling. m) were calculated using “CYTOWIN” software (D. Vaulot, www.sb-roscoff.fr/Phyto/cyto.html). Flow cytometric analyses of natural samples have The size distribution of the M. pusilla cells was determined us- shown that eukaryotic picophytoplankton are ubiqui- ing a Coulter Multisizer (Beckman Coulter, Miami, FL, USA) tous (e.g. DuRand and Olson 1996, Campbell et al. equipped with a 30-m aperture. The culture was diluted with fil- 1997, Zubkov et al. 1998, Shalapyonok et al. 2001), tered seawater (9- to 14-fold dilutions) to obtain coincidence rates of 6% or less. The size distributions of replicate samples were av- but detailed laboratory studies of diel variations in phy- eraged and normalized to the cell concentration from flow cyto- toplankton optical properties have focused on smaller metric analysis. These average 256-channel data of cell diameter and larger cell types, including prokaryotic picoplank- distributions were used to calculate the geometric projected area ton (Stramski et al. 1995) and eukaryotic nanoplank- of the mean cell (G , m2), the diameter of the mean cell (D , 3 ton (e.g. Stramski and Reynolds 1993, DuRand and m), and the volume of the mean cell (V , m ) to include the effects of polydispersion in subsequent calculations (equations Olson 1998). To fill this gap, we chose to examine a from Stramski and Reynolds 1993, Reynolds et al. 1997). small (approximately 1.6 m diameter) prasinophyte, Spectral absorption and attenuation at nine wavelengths Micromonas pusilla; recent studies on picoplankton di- (412, 440, 488, 510, 532, 555, 650, 676, and 715 nm) were mea- versity using molecular approaches suggest that pras- sured using an absorption and attenuation meter with a 25-cm pathlength (ac-9, WETLabs, Philomath, OR, USA). Two reser- inophytes, and the genus Micromonas in particular, voirs were attached with tubing to the inlet and outlet of the ac-9, are important in many different environments from samples (diluted 9- to 21-fold with filtered seawater) were gravity open ocean to coastal waters (Diez et al. 2001, Not et fed through the instrument, and data collection was closely mon- al. 2002, F. Not and D. Vaulot, personal communica- itored to ensure the absence of air bubbles. A dilution series was tion). Here we examine variations in physical, chemical, measured at the start of the experiment to ensure that operation was within the linear range of the instrument. Filtered seawater and optical properties over the diel cycle. In addition, we was run between consecutive samples and subtracted as a blank. evaluate the relative importance of cell size and refrac- A temperature correction was applied to account for the differ- tive index to changes in optical properties, examine vari- ence between water temperature of the samples and that during ability in the carbon-specific beam attenuation, and pro- instrument calibration (Pegau et al. 1997). The measured absorption and attenuation coefficients (a and c, m1) were nor- pose a method to estimate cellular carbon for different malized to the cell concentrations to obtain average cross-sections 2 phytoplankton groups directly from flow cytometric FLS ( a and c, m ). The measured efficiency factors for attenua- and an empirical calibration. tion (Qc) were determined as c/G.

1134 MICHELE D. DURAND ET AL.

Absorption was measured in a 1-cm cuvette using a Lambda 18 UV/VIS dual-beam spectrophotometer with a 60-mm inte- grating sphere (Perkin-Elmer, Shelton, CT, USA). Syringe- filtered (0.22 m) culture was used in the reference cuvette and as the blank. A dilution series was measured at the start of the experiment to ensure that in subsequent measurements multiple scattering effects were negligible. The value at 750 nm, where absorption by phytoplankton is negligible, was subtracted from the absorption values across the spectra to account for any scattering, which is assumed to be spectrally flat. The measured absorption coefficients (a, m1) were then normalized to the cell concentrations to obtain the average cellular 2 absorption cross-sections ( a, m ). The measured efficiency factors for absorption (Q a) were determined as a/G. The absorption data presented here are from the spectrophotometer measurements, which were strongly correlated with those from the ac-9 (r 2 0.97, n 208). On average the ac-9 measurements were 90% of those from the spectrophotometer, but the bias was somewhat wavelength dependent (with ac-9 values lower by as much as 12% at 412 nm or higher by as much as 16% at 650 nm). For the measurements of chl, triplicate 1-mL samples were collected on GF/F filters and extracted overnight in cold 90% acetone. Samples were analyzed on a 10-AU fluorometer (with optical kit 10-040R, Turner, Sunnyvale, CA, USA) using the non-acidi- fication technique (Welschmeyer 1994). The fluorometer was calibrated with pure chl a (Sigma-Aldrich, St. Louis, MO, USA). For analysis of carbon and nitrogen content, duplicate 25-mL samples were collected on pre-combusted GF/F filters, frozen, and later dried overnight at 60 C. The samples were analyzed on a Perkin-Elmer 2400 CHN analyzer with acetanilide as the standard, and dry pre-combusted filters were analyzed as blanks and subtracted. Calculations of refractive index. Mie scattering theory, which Fig. 1. Time series of (A) cell concentration and cell vol- assumes homogenous and spherical particles, was used to calcu- ume (V ) and (B) flow cytometric forward light scattering (FLS, late both the real (n) and the imaginary (n ) parts of the refrac- bead units) and carbon per cell for replicate carboys of Mi- tive index (algorithm from Bohren and Huffman 1983). First, cromonas pusilla. The black bar denotes when the lights were off preliminary n and then n were calculated using the anomalous in the incubator (12–24 h past dawn). diffraction approximation through iterations to closely match the theoretically calculated Q a (or Q c) to the experimentally measured Q a (or Q c) (Van de Hulst 1957, following equations creased at night upon cell division (Fig. 1A). V at the in Morel and Bricaud 1986, Bricaud and Morel 1986). Then, end of the experiment (dawn) was half that at its max- the final n and n were derived from Mie theory using a multi- dimensional unconstrained nonlinear minimization (“fmin- imum (dusk), which corresponded to an approximate search” from MATLAB, The MathWorks, Natick, MA, USA) of doubling in cell numbers. This result is expected if we the absolute value of the difference between measured and the- assume that a cell must double in volume before it oretical Q a added to the absolute value of the difference be- can divide into two cells and is relevant to the estima- tween measured and theoretical Q c. The measured values used in this estimation of n and n are the absorption and attenuation of division rates from cell volume (DuRand and tion coefficients, cell concentrations, and cell diameter distri- Olson 1998). Mean cell diameter, D , averaged 1.60 butions. Calculations were made at eight wavelengths (412, m with a range of 1.37 to 1.89 m over the diel cycle. 440, 488, 510, 532, 555, 650, and 676 nm). Flow cytometric analysis showed similar patterns in FLS (Fig. 1B) and side light scattering (data not results and discussion shown). Carbon per cell was also minimum at dawn Cell growth and division. Micromonas pusilla cell divi- and maximum at dusk (Fig. 1B). Linear regressions sion was phased to the light:dark cycle with most of between V , FLS, and cell carbon all showed strong the division occurring during the first 8 h of darkness correlations (between V and FLS, r 2 0.91; between (Fig. 1A). The cultures in this experiment were in ul- carbon and FLS, r 2 0.92; between V and carbon, tradian growth, with a growth rate over the 24-h ex- r 2 0.93). The strong relationships between these periment of 0.9 d1 (1.3 divisions per day). Approxi- measurements can be used to interpret diel variations mately one doubling per day took place between dusk in natural assemblages, for which individual cell carbon and dawn, and the remainder occurred during the content, for example, cannot be measured but FLS of light period. Nighttime division is common among recognizable cell types can. eukaryotic phytoplankton in the ocean (Chisholm Cell constituents and refractive index. For both the real 1981). A recent study found that cell division in five part of the refractive index, n, and the imaginary part eukaryotic picoplankton species, including M. pusilla, of the refractive index, n, there was no discernible started just before dusk and continued for several diel pattern that was consistent between the replicate hours into the night (Jacquet et al. 2001). cultures (Fig. 2A). n (650 nm) varied from 1.049 to The mean cell volume, V , increased during the day 1.066 (mean 1.058, SD 0.003). n (676 nm) varied as the cells were photosynthesizing and then de- from 0.0057 to 0.0082 (mean 0.0066, SD 0.00052). M. PUSILLA DIEL OPTICAL PROPERTIES 1135

surements will be propagated and affect the result of the calculations. Previous studies have shown variability in n with a tendency to increase during the day and decrease at night, though with irregularities, for cultures of Thalassiosira pseudonana (Stramski and Rey- nolds 1993), Synechococcus (Stramski et al. 1995), and Nannochloris sp. at some growth conditions (DuRand and Olson 1998). The variation in intracellular carbon concentration, Ci, over time was similar to that of n (Fig. 2B), with no discernible diel pattern. The C:N ratio, however, did not appear to follow trends in n over the diel cycle (Fig. 2B). The daytime increase in C:N ratio, com- bined with examination of the patterns in particulate organic carbon and particulate organic nitrogen over the day (data not shown), indicates that cells accumu- lated relatively carbon-rich compounds (carbohydrates) during the day faster than they did nitrogen-rich compounds (proteins). At night, the C:N ratio decreased due to loss of carbon (20% of total carbon), presumably from respiration, whereas the total nitrogen remained constant. The mean C:N ratio we observed for M. pusilla was 6.7, close to the Redfield ratio of 6.6. The common trend found in Ci and n but not in the C:N ratio indicates that changes in the chemical constituents of the cell were less important than the ratio of water to cellular material in determining the refractive index. A similar conclusion was reported by Aas (1996), who calculated n from the composition of phytoplankton cells. The concentration of chl a per cell increased during the day, peaked 2 h before the dark period, and then decreased (Fig. 2C). Flow cytometric chl fluorescence per cell followed a similar diel pattern (data not shown). The ratio of FCM chl fluorescence to cellular chl a concentration exhibited a coefficient of variation (CV) of Fig. 2. (A) Time series of the refractive index, n (650 nm), less than 6% over the diel cycle (Fig. 2C), with some in- and the imaginary part of the refractive index, n (676 nm), for dication of a systematic increase in the second half of the replicate carboys of Micromonas pusilla. Values were derived us- light period. The small variations seen here are consistent ing measurements and Mie scattering theory. (B) Time series with low packaging and only minor changes in pigment 3 of intracellular carbon concentration, Ci (kg m ), and C:N ra- composition over the course of the day. tio and (C) cellular chl concentration (pgcell 1) and ratio of flow cytometric chl fluorescence to cellular chl a concentration Single-cell optical cross-sections and efficiencies. The mean for replicate carboys of M. pusilla. cellular cross-sections for absorption, a, increased during the day and decreased during the night (Fig. 3A). The ratio of FCM chl fluorescence to absorption cross-section ( a) showed only small variations, with a CV of 6% over the diel cycle (data not shown). These Multiple solutions for n were not found, presumably results suggest that a for these cells can be estimated due to the small size of the cells (see discussion in with reasonable accuracy from measurements of flow Stramski and Mobley 1997). Most of the variability in cytometric chl fluorescence. This is important for nat- both n and n occurred between the first sampling ural samples that include assemblages of mixed species point (“dawn”) and the next sampling point (2 h each of which may have very different a values. Dif- later). The variability in both n and n over the 24-h ferent a values cannot be separately assessed by bulk sampling period was small, however, and the calcu- spectrophotometry, but each cell can be indepen- lated values were too noisy to detect a diel pattern. dently characterized by flow cytometry. Some of the variability in our estimates of refractive The mean cellular cross-sections for attenuation, c, index may be associated with how it was derived. Be- increased during the day and decreased during the night cause n and n were calculated from the measured val- (Fig. 3A). The scattering cross-section, b (data not ues of absorption and attenuation coefficients, cell shown), was the major contributor to c (78%–84%). size, and cell concentration, any errors in these mea- The b obtained from bulk methods includes scattering 1136 MICHELE D. DURAND ET AL.

coplankton-sized cells (Stramski et al. 1995). The mean 2 1 a*chl at 676 nm was 0.0201 m (mg chl) (range over time was 0.0180 to 0.0229), which is extremely close to the value for pure chl a dissolved in acetone (0.0207 at 663 nm; Morel and Bricaud 1986). This is also consistent with virtually no package effect in these cells. The chl-specific scattering coefficient (b*chl), though relatively constant during most of the light period, increased shortly before dusk, peaked near dusk, and then declined during the dark period (Fig. 4A). The peak near dusk was likely due to the slight time lag between the peak in bulk chl concentration (10 h) and the peak in bulk carbon concentration (12 h); this lag suggests that the cells stopped synthesizing chl before they stopped producing other cell material. The diel variations we observed in b*chl (21% increase from dawn

Fig. 3. Time series of (A) optical cross-sections (m2 at 488 nm) for absorption ( a) and attenuation ( c) for replicate carboys of Micromonas pusilla. Note that the y-axes are scaled differently. (B) Time series of efficiency factors for absorption, Q a, and scattering, Qb, both at 440 nm for replicate carboys of M. pusilla. from all angles; nonetheless it was strongly correlated with FLS from the flow cytometer (r 2 0.94). It is important for interpreting measurements on natural samples that FLS values for the angles measured by the flow cytometer strongly contribute to total scattering. The efficiency factor for absorption, Q a (440 nm), tended to increase during the day, reached a peak 2 h before dusk, and then decreased, though the pattern was not striking (Fig. 3B). As with n, the diel pattern in Q a was not dramatic, because the variations over the course of the diel cycle were small and the data were noisy. The Q a values were low, which is consistent with little packaging. The efficiency factor for scattering, Qb (440 nm), clearly increased during the day and decreased during the night (Fig. 3B). These variations were driven primarily by changes in cell size (Fig. 1A). The observed variations in n (which were similar at 440 nm to those at 650 nm in Fig. 2A) were not large enough to noticeably influence the diel pattern in Q b. Bulk optical properties. The chl-specific absorption coefficient (a*chl) did not exhibit a diel pattern (Fig. 4A) even at 440 nm, the wavelength at which the most 2 Fig. 4. Time series of (A) chl-specific absorption (a*chl, m variability is expected due to potential changes in the 1 2 1 (mg chl) ) and chl-specific scattering (b*chl, m (mg chl) ); (B) package effect and in pigment composition. This indi- absorption coefficient, a (m1), and scattering coefficient, b (m1); cates that the package effect is negligible for M. pu- and (C) scattering-to-absorption ratio, b:a, all at 440 nm for repli- silla, which is consistent with the findings for other pi- cate carboys of Micromonas pusilla. M. PUSILLA DIEL OPTICAL PROPERTIES 1137 to dusk) were small compared with the variations among b :a may be affected by the planktonic community species, which have been found to be 8- to 10-fold composition, with higher values expected for a com- (Bricaud et al. 1983, Morel 1987). This implies that munity with a significant heterotrophic bacterial con- chl-based models of scattering (such as Morel 1987) tribution, for example. This example for pure phy- should be relatively insensitive to the time of day the toplankton and the variations observed over the diel measurements were taken, even for phytoplankton cycle should be useful for interpreting absorption and that exhibit phased division patterns. scattering coefficients from in situ instruments such as The bulk absorption coefficient a (440 nm) (Fig. 4B) ac-9 meters (e.g. Sosik et al. 2001). increased during the day and then was relatively con- Spectral variability. The spectral shape of a did not stant during the night. The daytime increase was prima- vary with time of day, but the spectral shape of b did (Fig. rily caused by the increase in chl per cell (Fig. 2C) and 5A). As an indicator of the shape of the b spectrum, a associated increase in a (Fig. 3A) and not to cell con- spectral slope ( ) was estimated for each observation centration, which was relatively constant (Fig. 1A). In based on least-squares regression to a power function: contrast, the constant values of a at night reflect a bal- b() , where and are fit parameters. Values ance between increasing cell concentration (Fig. 1A) of varied systematically with cell size, with the greatest and decreasing a (Fig. 3A). These results are also con- values occurring when the cells were smallest (Fig. sistent with a negligible package effect because a did not 5B). In situ instruments capable of quantifying the spec- increase during the night when the cells were getting tral shape of both a and b are now in common use, and smaller. As a result, the pattern in a reflects that of bulk this kind of information about potential sources and am- chl in the culture, which also increased during the day plitudes of variation will be critical for adequate inter- and was constant at night (data not shown). pretation of observations from natural waters. The bulk scattering coefficient b (440 nm) exhib- Both parts of the refractive index, n and n, also ex- ited a daytime increase similar to a (440 nm), but in hibited spectral variability, although with no discern- contrast to a, b decreased during the night (Fig. 4B). ible change over the course of the day (Fig. 6). The The daytime increase was primarily caused by the in- observed spectral shape of n is similar to that previ- crease in diameter and associated increase in b per cell and not by changes in n or cell concentration, which were both relatively constant (Figs. 1A and 2A). The nighttime decrease in b is predicted theoretically for cells of this size undergoing division. Even though the redistribution of cell material from one cell (diameter 1.8 m pre-division) to two smaller cells (assuming conservation of volume) results in a 26% increase in the sum of the geometric cross-sections, theory predicts a 10% decrease in b (440 nm) (using the anomalous diffraction approximation with n and n set to the average values, 0.0095 and 1.057, respec- tively). The observed scattering coefficients for these M. pusilla cultures showed a decrease of approximately 20% during the nighttime division period. There are several possible explanations for the larger decrease than that predicted by theory. It could be due, in part, to losses of cell carbon through respiration; bulk carbon concentration in the culture decreased by approximately 20% during the night hours. In addition, the effects of intracellular heterogeneity and changes in cell shape upon division that cannot be accounted for by Mie theory could also contribute. Changes in the refractive index may also be a factor; however, the changes observed were not consistent (n increased in one culture and decreased slightly in the replicate), suggesting that this is not a significant factor determining the observed variability in b (which decreased in both replicate cultures during the night). The ratio of scattering to absorption, b :a, at 440 nm tended to increase during the day and decrease during the night (Fig. 4C). The average value for b :a Fig. 5. A) Spectral absorption (a) and scattering (b) coeffi- for M. pusilla was 3.14 (CV 11% over diel cycle), cient normalized to 676 nm for two time points, dawn and dusk, which is in the range observed for other types of phy- for Micromonas pusilla carboy 1. (B) Time series of the spectral toplankton (Stramski and Mobley 1997). In the ocean slope of b, , for the replicate carboys. 1138 MICHELE D. DURAND ET AL. ously observed for another picophytoplankton, Syn- echococcus (“CYA” in Stramski and Mobley 1997). Both spectra exhibited a maximum near 490 nm and also increased in the red part of the spectrum. The spectral shape of n, the part of the refractive index attrib- utable to absorption, was similar to that of a (Fig. 5A). Size and refractive index effects on attenuation. To evaluate the relative effects of variations in cell size and variations in refractive index on the attenuation cross- section ( c), a sensitivity analysis was performed. We used Mie scattering theory to calculate Qc and c (at 650 nm) from estimates of n, n, and measured D . A series of calculations were made with either n and n or D held constant at the observed mean values, whereas the other input(s) were allowed to vary with the observed diel pattern. The theoretical c values derived with constant n and n and with varying D Fig. 7. Results of sensitivity analysis for effects of varying closely matched the measured c values, whereas diameter (D, with n and n constant) or varying refractive index those calculated with constant D and varying n and n (n and n, with D constant) on the predicted attenuation cross- were nearly constant over the diel cycle (Fig. 7). section, c. For comparison, the measured c for Micromonas pu- These calculations emphasize that variation in cell silla carboy 1 are also shown (results for the replicate carboy were similar). size for M. pusilla is the primary factor determining the diel variability in c. This is similar to the results found in a diel study of Nannochloris sp. (DuRand and parable with the variation of 25%–30% between dawn Olson 1998), where variations in cell size were a more and dusk previously reported for T. pseudonana (Stramski important source of change in attenuation than varia- and Reynolds 1993), Synechococcus sp. (Stramski et al. tions in refractive index. For the diatom Thalassiosira 1995), and Nannochloris sp. (DuRand and Olson 1998). pseudonana (Stramski and Reynolds 1993) and for the The mean value of c*c we observed for M. pusilla was prokaryotic Synechococcus (Stramski et al. 1995), how- 2.9 m2(g C)1, within the range of 2.5 to 3.8 m2(g C)1 ever, variations in n were found to be more important reported for the other species. Among these four spe- or equal in importance to variations in cell size. This cies of phytoplankton that have been investigated in difference may arise because Synechococcus cells do not laboratory studies over the diel cycle, as well as Syn- exhibit highly synchronized cell division (Binder and echocystis that has been studied over a range of differ- Chisholm 1995) and because T. pseudonana is a dia- ent light levels (Stramski and Morel 1990), c*c varies tom and thus has constraints on its size set by its silica systematically with cell size (Fig. 9). The range of val- frustule and, additionally, may not be expected to have ues for c*c among species is greater than the variability tightly synchronized cell division (Chisholm et al. 1980). over the diel cycle for a single species. Nonetheless, Interspecific variability in carbon-specific attenuation. the intraspecific variability (for M. pusilla and Nan- The carbon-specific attenuation (c*c ) at 650 nm (Fig. 8) nochloris sp., at least) did show the same pattern: in- increased during the day and decreased at night for M. pusilla. The variation over the diel cycle was approximately 30% of the value at dawn, which is com-

Fig. 6. Spectra of the real part of the refractive index (n) Fig. 8. Time series of carbon-specific attenuation coefficient 2 1 and the imaginary part of the refractive index (n ) both at mid- (,c*c m(g C) at 650 nm) for replicate carboys of Micromonas pu- day (6 h after dawn) for replicate carboys of Micromonas pusilla. silla. M. PUSILLA DIEL OPTICAL PROPERTIES 1139 creasing c*c with increasing size over the diel cycle (data been suggested that diel variations in phytoplankton not shown). Several investigators used a single value of light scattering in natural samples can be used to esti- c*c to relate in situ measurements of beam attenuation to mate phytoplankton growth rates (DuRand 1995, carbon concentration for estimates of production from Binder et al. 1996, DuRand and Olson 1998, Vaulot and diel studies (Siegel et al. 1989, Cullen et al. 1992, Walsh Marie 1999), with some caveats, such as potential com- et al. 1995). These results indicate that it is important to plication by changes in refractive index. This application consider not only diel variations in c*c but also variability is supported by our finding that a large-scale calibration with community structure (DuRand and Olson 1996) between FLS and cell size based on a wide range of dif- because c*c appears to vary systematically with phy- ferent sized phytoplankton (equivalent spherical diame- toplankton cell size. ter from 0.87 to 39 m) also applies over the smaller size Relating FLS to cell size. An important outgrowth of range (approximately doubling) that occurs in a single laboratory-based experiments on phytoplankton optical species over the diel cycle (Fig. 10). The data presented properties is their application to studies in natural waters here do not reflect a comprehensive study of different with mixed assemblages of cell types. This is especially species and growth conditions, and it is possible that fac- relevant for flow cytometric studies where several differ- tors such as variations in the pigments that absorb at 488 ent groups of phytoplankton can be distinguished in nm (the wavelength of the FCM laser), differences in natural samples. In a number of studies, investigators cell shape and internal structures, and changes in scat- have used cell concentrations from flow cytometry of dif- tering efficiency with cell size may add additional unex- ferent phytoplankton groups to calculate their contribu- plained variance to the relationship shown in Figure 10. tion to microbial carbon using an empirical factor or cal- Even considering these limitations, using FLS to esti- ibration to convert from assumed or measured cell size mate cell volume has advantages over assuming a mean to carbon (Li et al. 1992, Campbell et al. 1994, 1997). size for all cells in a population or assemblage because it Some recent studies of natural assemblages have been provides a means to examine diel and other high fre- based on a similar approach, but in these cases cell vol- quency variations. ume was determined from FLS measurements and an Relating optical properties to carbon and chl. For many empirical relationship for phytoplankton cultures and ecological questions, determining cell volume is an in- then volume was converted to carbon content via a sec- termediate step toward estimating biomass (e.g. carbon ond empirical relationship (Gin et al. 1999, DuRand et al. 2001, Shalapyonok et al. 2001). In addition, it has

FIG. 10. Relationship between cell volume (m3) and flow cytometric forward light scattering (FLS, bead units). “Mix1” indicates measurements on different phytoplankton species at Fig. 9. Relationship between the carbon-specific beam attenu- one time of day (25 batch phytoplankton cultures consisting of 2 1 ation (c*c , m (g C) ) and diameter for five different phytoplank- 10 dinoflagellates, 5 prymnesiophytes, 4 Synechococcus strains, 3 ton: Synechococcus (“Syn1,” Stramski et al. 1995), Synechocystis prasinophytes, 2 chlorophytes, and 1 diatom; n 33 because (“Syn2,” Stramski and Morel 1990), Micromonas pusilla (“Micro,” some cultures were measured on more than one occasion). this study), Nannochloris sp. (“Nanno,” DuRand and Olson 1998), “Mix2” designates measurements on a number of different phy- and Thalassiosira pseudonana (“Tp,” Stramski and Reynolds 1993). toplankton species (Green 2002, Green et al. in press). “Diel” For each data point the mean SD over the diel cycle is shown, ex- indicates data from the M. pusilla diel sampling experiment. cept for Syn2 for which the mean SD for seven different irradi- The line shows a fourth-order polynomial fit, r 2 0.98, for ances is shown. The c*c is at 650 nm for all except Syn2 at 660 nm. both “Mix” data sets. 1140 MICHELE D. DURAND ET AL. or chl). Because relationships between cell volume and light scattering can be complicated, more direct approaches to estimating cell biomass from optical properties have been examined. Significant positive relationships between Ci and n and between intracellular chl concentration and n have been established from laboratory experiments on phytoplankton cultures (Stramski and Reynolds 1993, Stramski et al. 1995, DuRand and Olson 1998, Stramski 1999, Stramski et al. 2002). The diel variations we observed in M. pusilla are consistent with general relationships that describe results from a variety of phytoplankton cultures (Fig. 11). Stramski (1999) suggested that individual particle analysis of natural samples could yield information about the refractive index of different types of microorganisms that could then be converted to Ci and intracellular chl concentration that, together with data on cell size, could be used to estimate contributions of different particle types to carbon and chl concentration in the ocean. This approach is based on theoretical considerations and empirical data but does not require the potentially prob- lematic conversion from cell volume to carbon content. The most difficult component of this approach is determining refractive index from individual particle optical measurements in natural samples; although some limited estimates of real refractive index have been made (Ackleson and Robins 1990), it is not a routine measure- ment with currently available instrumentation. (Green 2002, Green et al. in press) investigated this extensively with flow cytometry and concluded that large real refractive index differences, such as between phytoplankton and mineral particles, can be resolved. It is clear, however, that differences among phytoplankton are difficult to quantify accurately. Estimates of refractive index with this approach are limited in accuracy by required assumptions about particle homogeneity and shape. We propose instead that cellular carbon content be directly estimated from flow cytometric measurements of FLS. Cellular carbon content and FLS are well correlated (Fig. 12, r 2 0.98) over a large range of phytoplankton species and cell sizes; though it is not a comprehensive study, this data set does include variability over the diel cycle and over different irradiance Fig. 11. (A) Relationship between real refractive index, levels and growth conditions. The exact details of the n(650), and intracellular carbon concentration, Ci, for Synechococcus relationship we present are specific to our flow cytom- (“Syn,” Stramski et al. 1995), Micromonas pusilla (“Micro,” this study), Nannochloris sp. (“Nanno,” DuRand and Olson 1998), eter and its optical configuration, but a similar cali- Thalassiosira pseudonana (“Tp,” Stramski and Reynolds 1993, bration could be obtained with relative ease for other Stramski 1999 and “Tp2,” Stramski et al. 2002), and a number instruments. Using our laboratory-based calibration of different phytoplankton species (“Mix,” Green et al. in press, to convert FLS to cellular carbon content will allow us Green 2002). Data for n are at 660 nm for Syn, Nanno, Tp, and to estimate the contribution of different plankton Tp2. Linear regression and 95% confidence intervals are plotted. (B) Relationship between imaginary refractive index, n(676), groups to total carbon content in the oceans with sim- and intracellular chl concentration, Chli, for the same experi- ple flow cytometric counts and measurements of FLS. ments as in A, excluding Nanno. Data for n are at 675 nm for This “FLS-carbon method” should be more reliable Syn and Tp and at 674 nm for Tp2. Linear regression and 95% than empirical conversion of cell volume to carbon confidence intervals are plotted. content (reviewed by Stramski 1999). Much of the variability in cell volume to carbon relationships is associated with differences in cell constituent concen- tween FLS and cellular carbon is subject to less vari- trations (i.e. carbon and water content) that also af- ability than the relationships between FLS and cell fect refractive index. Because FLS is affected by both volume or between cell volume and carbon content. cell volume and refractive index, the relationship be- Even though this is not apparent in our data for M. M. PUSILLA DIEL OPTICAL PROPERTIES 1141

present study of M. pusilla. Because phasing to the light:dark cycle is a widespread occurrence, we investigated diel variations in physical, chemical, and optical properties and found that M. pusilla cells increased in cell size, carbon content, single-cell scattering, attenuation, and absorption during the light period; all these properties then decreased with cell division during the dark period. Neither the real nor the imaginary refractive indices showed any significant trend over the light:dark cycle. The relationships between optical properties and physical or chemical cell characteristics observed for M. pusilla over the diel cycle were similar to trends observed across a range of species. With regard to single- cell light scattering, this is consistent with our finding that changes in refractive index of M. pusilla over the diel cycle were small (i.e. that light scattering was affected by cell size primarily). More generally, these results suggest that empirical approaches for estimating cell properties from light scattering measurements are applicable to a wide variety of species. In particular, we propose that the relationship observed between flow cytometric FLS and cellular carbon will allow esti- Fig. 12. Relationship between FLS (b.u.) and cellular carbon mation of the carbon content of different phytoplank- content (pgcell 1) for Micromonas pusilla (“Micro,” this study), Nannochloris sp. (“Nanno,” DuRand and Olson 1998), and a ton groups in natural assemblages. For cells similar to M. number of different phytoplankton species (“Mix,” Green pusilla, the relationship can also be used to resolve diel 2002, Green et al. in press). A second-order polynomial fit to variations related to cell growth and division. the log data is shown (r 2 0.98). In many natural communities, eukaryotic phytoplankton are numerically dominated by cells similar to M. pusilla. The ability to estimate their cell size and pusilla (because variations in n were small), it was evi- carbon content, population growth rates and primary dent in a study of Nannochloris grown under different productivity, and contributions to bulk optical prop- conditions (DuRand 1995, DuRand and Olson 1998). erties such as beam attenuation will expand our un- In addition, the FLS-carbon method is not susceptible derstanding of the role these cells play in planktonic to errors associated with deriving n and cell size using ecosystems. Mie scattering theory and the inherent assumptions of sphericity and homogeneity. Even though scattering theory indicates FLS is not a monotonic function We are grateful to Anne Canaday and Alexi Shalapyonok for as- sistance in the laboratory. We also thank Rick Reynolds and of cell size and refractive index, we have found a strong Dariusz Stramski for providing their data. Supported by NASA correlation between cellular carbon content and FLS. grants NAGW-5217, NAG5-7538, and NAG5-8868 (to H. M. S.) Further work is required to determine how sensitive and ONR grant N000149510333 (to H. M. S. and R. J. O.). This this relationship is to other factors that would be ex- is WHOI contribution 10608. pected to add variability, such as inclusion of cells that contain coccoliths or gas vacuoles or that have highly Aas, E. 1996. Refractive index of phytoplankton derived from its elongated shapes. Despite these caveats we suggest metabolite composition. J. Plankton Res. 18:2223–49. that routine FLS measurements, coupled with an em- Ackleson, S. G. and Robins, D. B. 1990. Flow cytometric determina- pirical relationship between FLS and cellular carbon tions of North Sea phytoplankton optical properties. Neth. J. such as that presented here, could provide simple esti- Sea Res. 25:11–8. Ackleson, S. G., Cullen, J. J., Brown, J. & Lesser, M. 1993. Irradi- mates of the carbon content of different planktonic ance-induced variability in light scatter from marine phyto- microorganisms and thus yield important information plankton in culture. J. Plankton Res. 15:737–59. on the partitioning and size distribution of carbon in Binder, B. J. & Chisholm, S. W. 1995. Cell cycle regulation in marine the oceans. In addition, because variations over the Synechococcus sp. strains. Appl. Environ. Microbiol. 61:708–17. diel cycle in this relationship fall on the same line as Binder, B. J., Chisholm, S. W., Olson, R. J., Frankel, S. L. & Worden, A. Z. 1996. Dynamics of picophytoplankton, ultraphytoplank- the general relationship, diel changes in FLS could be ton, and bacteria in the central equatorial Pacific. Deep-Sea Res. converted to diel changes in carbon content, thus al- II 43:907–31. lowing carbon productivity to be determined for groups Binder, B. J. & DuRand, M. D. 2002. Diel cycles in surface waters of of phytoplankton similar to M. pusilla. the equatorial Pacific. Deep-Sea Res. II 49:2601–17. Bohren, C. F. & Huffman, D. R. 1983. Absorption and Scattering of Conclusions. Even though they are ubiquitous in Light by Small Particles. John Wiley & Sons, New York, 530 pp. the oceans, eukaryotic picophytoplankton have not Bricaud, A., Morel, A. & Prieur, L. 1983. Optical efficiency factors been subject to detailed optical analysis until the of some phytoplankters. Limnol. Oceanogr. 28:816–32. 1142 MICHELE D. DURAND ET AL.

Bricaud, A. & Morel, A. 1986. Light attenuation and scattering by Morel, A. 1987. Chlorophyll-specific scattering coefficient of phy- phytoplanktonic cells: a theoretical modeling. Appl. Opt. 25: toplankton. A simplified theoretical approach. Deep-Sea Res. 34: 571–80. 1093–105. Campbell, L. Nolla, H. A. & Vaulot, D. 1994. The importance of Morel, A. & Bricaud, A. 1986. Inherent optical properties of algal Prochlorococcus to community structure in the central North Pa- cells including picoplankton: theoretical and experimental re- cific Ocean. Limnol. Oceanogr. 39:954–61. sults. Can. Bull. Fish. Aquat. Sci. 214:521–59. Campbell, L., Liu, H., Nolla, H. A. & Vaulot, D. 1997. Annual vari- Not, F., Simon, N., Biegala, I. C. & Vaulot, D. 2002. Application of flu- ability of phytoplankton and bacteria in the subtropical North orescent in situ hybridization coupled with tyramide signal ampli- Pacific Ocean at Station ALOHA during the 1991–1994 ENSO fication (FISH-TSA) to assess eukaryotic picoplankton composition. event. Deep-Sea Res. I 44:167–92. Aquat. Microb. Ecol. 28:157–66. Chisholm, S. W. 1981. Temporal patterns of cell division in unicel- Olson, R. J., Chisholm, S. W., Zettler, E. R. & Armbrust, E. V. 1990. lular algae. Can. Bull. Fish. Aquat. Sci. 210:150–81. Pigments, size, and distribution of Synechococcus in the North Chisholm, S. W., Morel, F. M. M. & Slocum, W. S. 1980. The phas- Atlantic and Pacific Oceans. Limnol. Oceanogr. 35:45–58. ing and distribution of cell division cycles in marine diatoms. Pegau, W. S., Gray, D. & Zaneveld, J. R. V. 1997. Absorption and at- Environ. Sci. Res. 19:281–300. tenuation of visible and near-infrared light in water: depen- Chung, S. P., Gardner, W. D., Landry, M. R., Richardson, M. J. & dence on temperature and salinity. Appl. Opt. 36:6035–46. Walsh, I. D. 1998. Beam attenuation by microorganisms and detri- Reynolds, R. A., Stramski, D. & Kiefer, D. A. 1997. The effect of ni- tal particles in the equatorial Pacific. J. Geophys. Res. 103:12669–81. trogen limitation on the absorption and scattering properties Cullen, J. J., Lewis, M. R., Davis, C. O. & Barber, R. T. 1992. Photo- of the marine diatom Thalassiosira pseudonana. Limnol. Ocean- synthetic characteristics and estimated growth rates indicate ogr. 42:881–92. grazing is the proximate control of primary production in the Shalapyonok, A., Olson, R. J. & Shalapyonok, L. S. 2001. Arabian equatorial Pacific. J. Geophys. Res. 97:639–54. Sea phytoplankton during Southwest and Northeast Monsoons Diez, B., Pedros-Alio, C. & Massana, R. 2001. Study of genetic diver- 1995: composition, size structure and biomass from individual sity of eukaryotic picoplankton in different oceanic regions by cell properties measured by flow cytometry. Deep-Sea Res. II 48: small-subunit rRNA gene cloning and sequencing. Appl. Envi- 1231–61. ron. Microbiol. 67:2932–41. Siegel, D. A., Dickey, T. D., Washburn, L., Hamilton, M. K. & Mitch- DuRand, M. D. 1995. Phytoplankton Growth and Diel Variations in ell, B. G. 1989. Optical determination of particulate abun- Beam Attenuation through Individual Cell Analysis. Ph.D. thesis. dance and production variations in the oligotrophic ocean. Massachusetts Institute of Technology/Woods Hole Oceano- Deep-Sea Res. 36:211–22. graphic Institution, Woods Hole, MA, 263 pp. Sosik, H. M., Green, R. E., Pegau, W. S. & Roesler, C. S. 2001. Tem- DuRand, M. D. & Olson, R. J. 1996. Contributions of phytoplank- poral and vertical variability in optical properties of New En- ton light scattering and cell concentration changes to diel vari- gland shelf waters during late summer and spring. J. Geophys. ations in beam attenuation in the equatorial Pacific from flow Res. 106:9455–72. cytometric measurements of pico-, ultra-, and nanoplankton. Stramska, M. & Dickey, T. D. 1992. Variability of bio-optical proper- Deep-Sea Res. II 43:891–906. ties of the upper ocean associated with diel cycles in phyto- DuRand, M. D. & Olson, R. J. 1998. Diel patterns in optical proper- plankton populations. J. Geophys. Res. 97:17873–87. ties of the chlorophyte Nannochloris sp.: relating individual-cell Stramski, D. 1999. Refractive index of planktonic cells as a measure to bulk measurements. Limnol. Oceanogr. 43:1107–18. of cellular carbon and chlorophyll a content. Deep-Sea Res. I 46: DuRand, M. D., Olson, R. J. & Chisholm, S. W. 2001. Phytoplankton 335–51. population dynamics at the Bermuda Atlantic Time-series sta- Stramski, D. & Morel, A. 1990. Optical properties of photosynthetic tion in the Sargasso Sea. Deep-Sea Res. II 48:1983–2003. picoplankton in different physiological states as affected by Gardner, W. D., Walsh, I. D. & Richardson, M. J. 1993. Biophysical growth irradiance. Deep-Sea Res. 37:245–66. forcing of particle production and distribution during a spring Stramski, D. & Reynolds, R. A. 1993. Diel variations in the optical bloom in the North Atlantic. Deep-Sea Res. II 40:171–95. properties of a marine diatom. Limnol. Oceanogr. 38:1347–64. Gardner, W. D., Chung, S. P., Richardson, M. J. & Walsh, I. D. 1995. Stramski, D., Shalapyonok, A. & Reynolds, R. A. 1995. Optical char- The oceanic mixed-layer pump. Deep-Sea Res. II 42:757–75. acterization of the oceanic unicellular cyanobacterium Synecho- Gin, K. Y. H., Chisholm, S. W. & Olson, R. J. 1999. Seasonal and coccus grown under a day-night cycle in natural irradiance. J. depth variation in microbial size spectra at the Bermuda Atlan- Geophys. Res. 100:13295–307. tic time series station. Deep-Sea Res. I 46:1221–45. Stramski, D. & Mobley, C. D. 1997. Effects of microbial particles on Green, R. E. 2002. Scale Closure in Upper Ocean Optical Properties: From ocean optics: a database of single-particle optical properties. Single Particles to Ocean Color. Ph.D. thesis. Massachusetts Insti- Limnol. Oceanogr. 42:538–49. tute of Technology/Woods Hole Oceanographic Institution, Stramski, D., Sciandra, A. & Claustre, H. 2002. Effects of tempera- Woods Hole, MA, 168 pp. ture, nitrogen, and light limitation on the optical properties of Green, R. E., Sosik, H. M., Olson, R. J. & DuRand, M. D. Flow cyto- the marine diatom Thalassiosira pseudonana. Limnol. Oceanogr. metric determination of size and complex refractive index for 37:392–403. marine particles: Comparison with independent bulk esti- Van de Hulst, H. C. 1957. Light Scattering by Small Particles. John mates. Appl. Opt. (in press). Wiley & Sons, New York, 470 pp. Guillard, R. R. L. 1975. Culture of phytoplankton for feeding ma- Vaulot, D. & Marie, D. 1999. Diel variability of photosynthetic pico- rine invertebrates. In Smith, W. & Chaney, M. [Eds.] Culture of plankton in the equatorial Pacific. J. Geophys. Res. 104:3297–310. Marine Invertebrate Animals. Plenum, New York, pp. 29–60. Walsh, I. D., Chung, S. P., Richardson, M. J. & Gardner, W. D. 1995. Hamilton, M., Granata, T. C., Dickey, T. D., Wiggert, J. D., Siegel, D. The diel cycle in the integrated particle load in the equatorial A., Marra, J. & Langdon, C. 1990. Diel variations of bio-optical Pacific: A comparison with primary production. Deep-Sea Res. II properties in the Sargasso Sea. SPIE: Ocean Optics X 1302:214–24. 42:465–77. Jacquet, S., Partensky, F., Lennon, J. & Vaulot, D. 2001. Diel pat- Welschmeyer, N. A. 1994. Fluorometric analysis of chlorophyll a in terns of growth and division in marine picoplankton in cul- the presence of chlorophyll b and pheopigments. Limnol. ture. J. Phycol. 37:357–69. Oceanogr. 39:1985–92. Jerlov, N. G. 1976. Marine Optics. Elsevier, Amsterdam, 231 pp. Zubkov, M., Sleigh, M., Tarran, G., Burkill, P. & Leakey, R. 1998. Li, W. K. W., Dickie, P. M., Irwin, B. D. & Wood, A. M. 1992. Biomass Picoplanktonic community structure on an Atlantic transect of bacteria, cyanobacteria, prochlorophytes and photosynthetic from 50 degree N to 50 degree S. Deep-Sea Res. I 45:1339–55. eukaryotes in the Sargasso Sea. Deep-Sea Res. 39:501–19.