Study Site and CTD Casts
Data were collected on four cruises in the Sargasso Sea on board the R/V Atlantic Explorer. On each cruise, sampling was conducted at three stations: the center and edge of a mesoscale eddy and at one station outside of the eddy. Eddies were identified using satellite-derived sea level anomaly (SLA) data provided by Dr. Dennis McGillicuddy and Dr. Valery Kosnyrev of the Woods Hole Oceanographic Institution. Target eddies (one per cruise) were initially identified on the day of departure; the ship's position within the eddy (at the center or the edge, as appropriate) was confirmed by daily checks of SLA data.

At each station, high resolution CTD casts to ~2000 m were performed at noon to measure core physical, chemical and biological parameters of the water column. In addition to the core CTD casts, pre-dawn "Productivity" CTD casts were performed to collect water for measurements of size-fractionated biomass (as chl a) and size-fractionated primary productivity. Samples were obtained using the 24 bottle Niskin rosette from 3-4 depths (20 m, 40-50 m, deep fluorescence maximum (~80 m), and 100 m). Ten-liter polycarbonate collection bottles (covered with black tape) were pre-rinsed with sample water and were filled by draining the Niskin bottles through opaque tubing. All samples were pre-filtered through a 200 um Nitex screen. Further handling of the samples was done in the dark or under red light.

The 200 um pre-screened water from pre-dawn productivity casts was used for measurements of size-fractionated biomass (as chl a) and biomarker photopigments by HPLC and for measurements of size-fractionated primary productivity.  HPLC pigments were also used for taxonomic identification of total and size-fractionated phytoplankton groups using ChemTax analyses. Samples for microscopy were also taken from productivity casts as verification of the ChemTax results using methods described above for core CTD casts. 

Total phytoplankton biomass was measured directly by filtering triplicate aliquots of 1 to 2 liters of pre-screened water onto GF/F filters. This gave total chl a in the size fraction 0.7 to 200 um. The biomass of three size classes of phytoplankton was quantified by differential filtration: the picophytoplankton (0.7-2 um), the nanophytoplankton (2-20 um) and the microphytoplankton (20-200 um) as follows. Triplicate aliquots of 1 to 2 liters of pre-screened water were filtered through a 2 um Nuclepore filter then onto a GF/F filter (= picophytoplankton, 0.7-2 um). Triplicate aliquots of pre-screened water were also filtered through a 20 um Nitex mesh then onto a GF/F filter (= 0.7-20 um). Biomass of the nanophytoplankton size class was determined by subtracting the picophytoplankton biomass from the 0.7-20 um biomass. Microphytoplankton biomass was determined by subtracting the 0.7-20 um biomass from the total chl a value. Filters were folded and placed in 1.5 ml cryotubes and frozen at -80° C until later analysis at the University of South Carolina (USC) using the methods below. 

Primary Prodcutivity Measurements
For size-fractionated primary productivity measurements, 200 um pre-screened water collected from discrete depths were dispensed into Nalgene polycarbonate incubation bottles (7-8 clear bottles plus 1-2 dark bottles per depth; 800-1200 ml each). Bottles were spiked with 14C-labeled sodium bicarbonate (PerkinElmer Health Sciences Inc.) to a final activity 0.04-0.08 uCi ml-1 per bottle. An additional bottle per depth was used as a particulate blank (T0) (Barber et al., 1996). The T0 bottles were immediately filtered onto a GF/F, acidified with 500 ul 0.5 N HCl and left open to fume for 24 hours (Barber et al., 1996). Samples for total counts (Tc; 100 ul) were collected from one bottle per depth and combined with 200 ul of phenylethylamine (PEA) and 5 ml of scintillation cocktail (EcoLume, MPBiomedicals, Solon, Ohio). All bottles were incubated in situ at the depth of collection. Incubations were started before sunrise (usually between 05:00 and 06:00 h) and were terminated 24 h later. The productivity array was tracked using a Telonics, Inc. transponder platform subscribed to the ARGOS satellite tracking system.

Total phytoplankton primary productivity was measured directly by filtering triplicate incubation bottles onto GF/F filters. This gave total primary productivity in the size fraction 0.7 to 200 um. Dark bottle productivity was also measured directly by filtering dark bottles directly onto GF/F filters (= dark productivity; 0.7-200 um). Size-fractionated rates of primary productivity of the picophytoplankton, nanophytoplankton and microphytoplankton were made by differential filtration. Two 1 liter bottles were filtered through a 20 um Nitex mesh then onto a 2 um Nuclepore filter (= nanophytoplankton, 2-20 um). Two or three 1 liter bottles were filtered through a 20 um Nitex mesh then onto a GF/F filter (= 0.7-20 um).  Filters were treated with 500 ul of 0.5 N HCl and left under a fumehood for 24 hours, then combined with 10 ml scintillation cocktail.  Radioactivity was determined in disintegrations per minute (DPM) by the shipboard liquid scintillation analyzer (Packard Tri-Carb 2000CA). 

Rates of primary productivity (PP) were calculated in units of mg C m-3 d-1 using the methods of Barber et al. (1996) with the addition of dark bottles:

    PP = (DPM24 – DPM0 – DPMd)/(1.05)(25200 mg C m-2)(DPMtot * time)-1       

where DPM24 = activity on filter after 24 hour incubation; DPM0 = activity of (depth-specific) T0 particulate blank; DPMd = average of (depth-specific) dark bottles; DPMtot = total activity DPM of isotope added multiplied by volume of water filtered (DPM ml-1); 1.05 = constant that accounts for preferential uptake of the lighter isotope 12C over 14C; 25,200 = concentration (in mg m-2) of inorganic carbon in seawater.

The rate of primary productivity for the picophytoplankton size class was determined by subtracting the nanophytoplankton productivity from the 0.7-20 um productivity. Primary productivity for the microphytoplankton size class was determined by subtracting the 0.7-20 um productivity from the total primary productivity, 0.7-200 um. Total and size-fractionated rates of primary productivity were integrated to 100 meters using trapezoidal integration (mg C m-2 d-1).

Microscopy
Samples preserved in Lugol’s were settled overnight in a 100 ml sedimentation chamber and enumerated at 400x using a Nikon TS-100 Eclipse inverted microscope. A minimum of 400 cells per sample were counted to give a confidence interval of ±10 % (Guillard, 1973). Phytoplankton taxa were identified to the lowest taxonomic level possible, that is, in most cases, to genus.

HPLC and ChemTax
Samples for HPLC analysis were lyophilized for 24 h at -50° C, placed in 90% acetone (0.45-0.55 ml), and extracted at -20° C for 24 h.  Filtered extracts (350 µl) were injected into a Shimadzu HPLC equipped with a monomeric (Rainin Microsorb-MV, 0.46 x 10 cm, 3 µm) and a polymeric (Vydac 201TP54, 0.46 x 25 cm, 5 um) reverse-phase C18 column in series. A nonlinear binary gradient consisting of the solvents 80% methanol: 20% 0.50 M ammonium acetate and 80% methanol: 20% acetone was used for pigment separations (Pinckney et al. 1996). Absorption spectra and chromatograms (440 ± 4 nm) were acquired using a Shimadzu SPD-M10av photodiode array detector. Pigment peaks were identified by comparison of retention times and absorption spectra with pure standards (DHI, Denmark). The synthetic carotenoid ß-apo-8'-carotenal (Sigma) was used as an internal standard. Contributions of individual algal groups to total community composition and to each size class was determined by chemical taxonomy using the ChemTax matrix (Mackey et al., 1996).