Notes 1813

Limnol. Oceatzogr., 38(8). 1993, 1813-1818 (0 1993, by the Ameruzan Society of Limnology and Oceanography. Inc Chlorophyll a concentrations in the North Pacific: Does a latitudinal gradient exist?

Abstract -Chlorophyll a concentrations were tion properties of glass-fiber and membrane measured as a function of depth from 28 to 48”N filters have demonstrated that glass-fiber filters along 152”W in March 199 1 with Whatman GF/F and 0.2~pm Nuclepore filters. Surface Chl a concen- inadequately retain < 1-pm-diameter cells due trations measured with 0.2~ym Nuclepore filters were to their large nominal pore size (0.7 and 1.2 up to fourfold higher than those measured with pm for Whatman GF/F and GF/C filters). Low Whatman GF/F filters. The largest difference be- retention efficiencies of glass-fiber filters result tween the two filter types was found in subtropical waters, where picoplankton were a major constituent when Chl a concentrations are low (<0.5-l .O of the phytoplankton assemblage. Chl a concentra- pg Chl a liter-‘) and when picoplankton are a tions integrated from 0 to 175 m showed a threefold dominant fraction of the as- increase (9-26 mg Chl a m-I) between 28 and 48”N semblage (Phinney and Yentsch 1985; Taguchi when Whatman GF/F filters were used. However, integrated Chl a concentrations based on measure- and Laws 1988). Under such conditions the ments with 0.2~pm Nuclepore filters were nearly con- use of membrane filters with submicron pore stant (25-3 1 mg Chl a mpZ) over the transect. These sizes has been recommended (although not al- results lead us to question the existence of previously ways followed). Many studies have made ex- reported latitudinal gradients in integrated Chl a tensive use of glass-fiber filters due to their low concentrations in the North Pacific Ocean. cost and fast flow rates. Results from previous basinwide studies Attempts to understand the ocean’s role in have suggested the presence of a latitudinal global climate change require accurate esti- gradient in integrated Chl a concentrations be- mates of phytoplankton and produc- tween subtropical and subarctic waters. The tion. Recent work has shown that picoplank- purpose of this study was to compare Chl a ton constitute a major fraction of the concentrations in subtropical, transition, and phytoplankton biomass in tropical, subtropi- subarctic waters and to evaluate the existence cal, and temperate waters. As well, these small of a meridional gradient in integrated Chl a phytoplankters have been found to contribute concentrations across the North Pacific basin. significantly to the in open- Sampling was conducted from 20 to 55”N ocean ecosystems (Li et al. 1983; Platt et al. along 152”W in March 199 1 on the RV Dis- 1983). However, in oligotrophic tropical and coverer as part of the Climate and Global subtropical waters, where picoplankton dom- Change Program of NOAA. Water samples inate, Chl a concentrations and the level of were collected at -20-m intervals from the primary production may have been underes- surface to 100 m and then every 25 m to a timated due to the use of filters with inappro- depth of 175 m with lo-liter Niskin bottles priate pore sizes. attached to a rosette sampler. Duplicate 200- Numerous comparative studies of the reten- ml subsamples of seawater were filtered under low vacuum pressure (5 180 mm of Hg) onto either 25-mm-diameter combusted Whatman GF/F or 0.2~pm Nuclepore filters. Near the Acknowledgments We thank the captain, officers, and crew of the NOAA end of the filtration, 0.5 ml of saturated MgC03 ship Discoverer for their cooperation and assistance during was added to the samples. Chl a concentra- the cruise. We are grateful to Larry Small, Barry and Ev tions were measured with a Turner Designs 10 Sherr, and two anonymous reviewers for their helpful sug- fluorometer after a 24-h extraction in 90% ac- gestions and comments. etone at -20°C in the dark. Before the cruise, This research was supported by a NASA graduate stu- dent fellowship in global change research to M.-L.D. and the fluorometer was calibrated with a pure Chl a NOAA climate and global change grant (NA 16RC009- a standard purchased from Sigma Chemical 01) to P.A.W. Co. In addition, the fluorometer was also stan- 1814 Notes dardized before each use with a secondary Chl Table 1. Comparison of Chl a retained by combusted a standard, also from Sigma Chemical Co., and and noncombusted Whatman GF/F filters and by com- busted filters at vacuum pressures of 80 and 180 mm of with 90% acetone blanks and extracted filter Hg. Duplicate samples were compared for each treatment. blanks (Whatman GF/F and Nuclepore filters) The overall precision (SD) for these measurements was (Venrick 1987). Integrated Chl a concentra- 0.01, n = 6. tions were calculated with the trapezoid meth- GF/F filters od of integration. The vertical profile of the Depth (noncombusted/ Combusted filters Chl a concentration at 32”N, 152”W was not Cm) combusted) (80/ 180 mm of Hg) included in Fig. 1 because some of the samples 15 0.97 0.92 may have been mixed. However, there was 100 1.10 0.93 no significant effect on the integrated Chl a Mean + SD 1.04+0.09 0.93*0.01 concentrations. The mean precision * 1 SD for duplicate Chl a measurements was 0.025+0.017 for GF/Ffiltersand 0.015*0.004 GF/F filters were lowest in the subtropical wa- for 0.2~pm Nuclepore filters. ters (0.023 pg Chl a liter’) and highest in In September 199 1 during a cruise off north- subarctic waters (0.299 pg Chl a liter-l). A ern California, we used the same procedure deep, subsurface Chl a maximum layer (0.250- outlined above to test if a difference could be 0.300 pg Chl a liter-l) was located between detected in the amount of Chl a measured by 100 and 120 m in the subtropical and transi- combusted and noncombusted Whatman tion waters but was absent in the subarctic GF/F filters. We also checked whether lower water at 48”N. The 0.2~pm Nuclepore filters vacuum pressure (80 mm of Hg) increased the retained 3-4 times more Chl a than the What- amount of Chl a measured. The station se- man GF/F filters in subtropical waters and 2- lected for these tests was 650 km off northern 3 times more Chl a in the transition zone. Chl California (42”N, 132”W) in oligotrophic wa- a concentrations measured with 0.2-pm Nu- ters characterized by low ambient nitrate con- clepore filters were not different from those centrations (- 1 ,uM) and a low phytoplankton measured with Whatman GF/F filters below standing stock (-0.080-0.300 pg Chl a li- 100 m in the subarctic (48”N). However, in the ter- l). This site was chosen for its similarity upper 100 m the difference in the Chl a con- to stations in subtropical waters along the centrations between the two filter types was NOAA transect. significantly different from zero (Student’s Comparison of combusted with noncom- t-test, P < 0.001). The relative picoplankton busted Whatman GF/F filters showed no ap- in the Chl a maximum layer was preciable difference in the Chl a concentrations highest in subtropical water (85% of the total measured for two depths at the oligotrophic Chl a), intermediate in transition water (78% site off northern California (Table 1). Chl a of the total Chl a), and lowest in subarctic concentrations were found to be -7% lower water (35-48% of the total Chl a) (Table 2). when the vacuum pressure was 80 mm of Hg Chl a concentrations integrated from 0 to compared to 180 mm of Hg. Chl a concentra- 175 m (Fig. 2) showed a threefold increase (9- tions measured at 15 and 100 m were com- 26 mg Chl a m-*) between 28 and 48”N when parable to those encountered at the southern end of the NOAA transect. The results of these tests indicate that the low Chl a values for the Table 2. Percent abundance of picoplankton Chl a in Whatman GF/F filters during the NOAA tran- the Chl a maximum layer of subtropical, transition, and subarctic waters. The percentage of Chl a in the 0.2-3-pm sect were not due to increased porosity of the fraction was calculated by [( >0.2 pm - > 3 pm)/>0.2 pm] combusted Whatman GF/F filters, inappro- x 100%. ND-No data. priate vacuum pressures, or a combination of the two. Chl a (kg liter ‘) % Chl a (0.2-3-wrn Vertical profiles of Chl a concentrations at N lat Depth (m) >0.2 I.trn >3 I.rrn fraction) four stations extending from the central North 28” 100 0.244 0.037 84.8 Pacific gyre (28”N), through the transition zone 32” 20 0.211 0.047 77.7 (37”N), and into subarctic waters (42 and 48”N) 37” 100 0.293 ND ND showed very different patterns (Fig. 1). Surface 42” 40 0.350 0.228 34.9 Chl a concentrations measured with Whatman 48” 6 0.382 0.200 47.6 Notes 1815

Chl a (1.19liter-‘) Chl a (pg liter-‘) 0 0.10 0.20 0.30 I I I I i

40 -

160 ISO/-\ / o-o GF-F o--o Nuclepore

Chl a (pg liter-‘) Chl a (pg liter-‘) 0.10 0.20

, -

120

I I I 42”N 152” W ;I’ o-o GF-F 160 I @ Nuclepore

Fig. 1. Vertical profiles of the Chl a concentrations at four locations obtained with Whatman GF/F and 0.2-hrn Nuclepore filters. 1816 Notes

Whatman GF/F filters were used. However, I integrated Chl a concentrations for phyto- 30- / T< plankton retained on the 0.2-pm Nuclepore c-7 filters were relatively constant along the tran- ‘E _ -s” // GF-F sect (25-31 mg Chl a m-2). This result was 5 20- / P due to a significant fraction of the Chl a in \ / n subtropical and transition waters passing w - // ‘\ / 5 / ‘0’ through the Whatman GF/F filters but being / g io- retained by the 0.2-pm Nuclepore filters. itt d Many investigators have been concerned about the loss of cells through glass-fiber filters t 01 I I I I I and the magnitude of the problem. Most com- 20 30 40 50 parative studies have concluded that mem- LATITUDE (“N) brane filters retain more Chl a than glass-fiber Fig. 2. Integrated Chl a concentrations (mg Chl a m ?) filters, especially when Chl a concentrations from 0 to 175 m between 28 and 48”N, 152”W calculated are ~0.5-1 .O pg Chl a liter-l (Venrick and from Chl a concentrations determined with Whatman GF/F Hayward 1984; Phinney and Yentsch 1985). and 0.2~pm Nuclepore filters. Munawar et al. (1982) found that 42% more Chl a was retained by 0.2-pm Nuclepore filters Past work has indicated the existence of a than Whatman GFK filters. Similarly, Ven- steep latitudinal gradient in phytoplankton rick et al. (1987) reported that 0.45-pm Mil- species composition, biomass (McGowan and lipore HA filters measured 27% more Chl a Williams 1973) and in situ fluorescence of Chl than Whatman GF/C filters at 24”N in the Pa- a (Pak et al. 1988) extending from the central cific. At 47”N the difference between the filter North Pacific subtropical gyre to subarctic wa- types was 10%. Phinney and Yentsch (1985) ters. Originally, this gradient was defined on compared Chl a retention by 0.45~pm Milli- the basis of differences in the surface Chl a pore HA membrane filters with various types concentration between subtropical and sub- of glass-fiber filters in laboratory and field ex- arctic waters. In our study, surface Chl a con- periments. In their studies with cultured phy- centrations also showed a latitudinal gradient toplankton from four oceanic regions, they between southern and northern ends of the found no significant difference in the Chl a transect, regardless of the type of filter used. concentration measured with either the 0.45- Data suggesting the presence of a latitudinal pm Millipore HA or Whatman GF/F filters. gradient in integrated Chl a concentrations were However, Whatman GF/F filters were found based on a series of basinwide cruises along to be inadequate and gave variable results when 155”W that took place from 1960 to 1980. used in a series of field experiments where Chl Results from these cruises have been reported a concentrations were in the range of O-5 pg elsewhere (Hayward and Venrick 1982; Hay- Chl a liter- l. Phinney and Yentsch concluded ward et al. 1983; Hayward and McGowan that under low biomass conditions the most 1985; Hayward 1987). During these cruises efficient glass-fiber filters retained only 60% of Chl a concentrations were measured with the Chl a. Whatman GF/C filters. The latitudinal pattern Our data confirm that membrane filters have that emerged from these data sets agrees with superior retentive properties compared to glass- our integrated values obtained with Whatman fiber filters but also indicate that the magnitude GF/F filters. However, our integrated >0.2- of the difference between the two filter types pm Chl a data indicate that a latitudinal gra- in subtropical waters may be greater than pre- dient between subtropical and subarctic waters viously thought. The 0.2-pm Nuclepore filters is weak at best when the picoplankton com- retained 67 and 74% more Chl a than the ponent of the phytoplankton assemblage is in- Whatman GF/F filters at 28”N in the chloro- cluded. Pak et al. (1988) reported meridional phyll maximum layer and in the surface water, variations in Chl a concentrations along 15 5”W respectively. However, this difference was in the North Pacific. Their fluorometric data smaller at higher latitudes. At 48”N there was show an increase in relative fluorescence be- only a 1O-27% difference between the two filter tween 23” and 57”N, however their extracted types. pigment results indicate no difference between Notes 1817 vertically integrated pigment concentrations at Our results suggest that historical data sets 28” and 42”N. It is not clear at present whether may have underestimated Chl a concentra- in situ fluorescence can be used as a proxy for tions in subtropical and transition waters by a extracted Chl a. factor of two to four whenever glass-fiber filters Unfortunately there are no other basinwide were used. Primary production measurements data sets that include submicron Chl a with also may have been underestimated due to the which we can compare our results with Nu- loss of small cells through glass-fiber filters. clepore filters. Nonetheless, the integrated Chl The magnitude of the underestimation may be a concentration measured over O-200 m at one significant when one considers that there is a station (28”N, 155”W) with 0.45~pm Millipore linear relationship between integrated Chl a HA filters during the PRPOOS experiment was concentrations and integrated primary pro- -26 mg Chl a m-2 (Venrick et al. 1987). This duction (Hayward and Venrick 1982) and value closely corresponds to our estimate of -60% of the primary production in tropical 25 mg Chl a rnA2 at 28”N, 152”W from 0 to and subtropical waters can be attributed to pi- 175 m obtained with 0.2-pm Nuclepore filters, coplankton (Li et al. 1983; Platt et al. 1983; and both values exceeded the Chl a trapped Iturriaga and Mitchell 1986). Obviously, there on Whatman GF/F filters by almost a factor is a serious need to use filters with submicron of three. pore sizes in low biomass waters to ensure ac- We can only presume that the component curate measurements of phytoplankton stand- of the picoplankton assemblage retained by the ing stocks and primary production. 0.2~pm Nuclepore and not the Whatman GF/F Mary-Lynn Dickson’ filters consisted of a marine prochlorophyte, Patricia A. Wheeler Prochlorococcus. The largest discrepancy in the Chl a concentrations between filter types was College of Oceanic and Atmospheric Sciences found at the lower latitudes along the transect, Oregon State University in nutrient-poor water. This pattern is consis- Corvallis 9733 l-5503 tent with other studies reporting this taxonom- ic group to be numerically abundant in oli- gotrophic waters (Olson et al. 1990) and to References contain a large fraction of the Chl a below the CHISHOLM, S. W., AND OTHERS. 1988. A novel free-living mixed layer (Campbell et al. unpubl.). As well, prochlorophyte abundant in the oceanic euphotic zone. Nature 334: 340-343. the small size of prochlorophyte cells (Chis- HAYWARD, T. L. 1987. The nutrient distribution and holm et al. 1988 report a mean diam of primary production in the central North Pacific. Deep- 0.7 pm) makes it likely that a significant por- Sea Res. 34: 1593-1627. tion of the prochlorophyte population passed - AND J. A. MCGOWAN. 1985. Spatial patterns of chlorophyll, primary production, macrozooplankton through the glass-fiber filters. If prochloro- biomass, and physical structure in the central North phytes were primarily responsible for the ob- Pacific Ocean. J. Res. 7: 147-167. served differences in Chl a concentrations, it -, AND E. L. VENRICK. 1982. Relation between sur- raises the possibility that the difference be- face chlorophyll, integrated chlorophyll and integrat- tween filter types might actually be larger than ed primary production. Mar. Biol. 69: 247-252. - -, AND J. A. MCGOWAN. 1983. Environ- that observed, due to the presence of Chl b mental heterogeneity and plankton community struc- (Chisholm et al. 1988). Significant quantities ture in the central North Pacific. J. Mar. Res. 41: 7 1 l- of Chl b can cause overestimation in pheopig- 729. ment concentrations and result in underesti- ITURRIAGA, R., AND B. G. MITCHELL. 1986. Chromo- coccoid : A significant component in the mation of the Chl a concentration (Loftus and food web dynamics of the open ocean. Mar. Ecol. Carpenter 197 1). However, LeBouteiller et al. Prog. Ser. 28: 291-297. (1992) found no bias in the Chl a estimation LEBOUTEILLER, A.,J. BLANCHOT,AND M. RODIER. 1992. arising from the presence of Chl b as long as Size distribution patterns of phytoplankton in the the ratio of Chl b : Chl a is < 1. They further western Pacific: Towards a generalization for the trop- ical open ocean. Deep-Sea Res. 39: 805-823. report a maximum Chl b concentration in deep subtropical water where the Chl b : Chl a is -0.2. Thus, it seems unlikely that our Nucle- pore results were influenced by the presence I Present address: Graduate School of Oceanography, of Chl b. University of Rhode Island, Narragansett 02882-l 197. 1818 Notes

LI, W. K. W., AND OTHERS. 1983. Autotrophic pico- phytoplankton chlorophyll technique: Toward auto- plankton in the tropical ocean. Science 219: 292-295. mated analysis. J. Plankton Res. 7: 633-642. LOFKJS, M. E., AND J. H. CARPENTER. 197 1. A fluoro- PLATT, T., D. V. SUBBA RAO, AND B. IRWIN. 1983. Pho- metric method for determining chlorophylls a, b and tosynthesis of picoplankton in the oligotrophic ocean. c. J. Mar. Res. 29: 319-338. Nature 301: 702-704. MCGOWAN, J. A., AND P. M. WILLIAMS. 1973. Oceanic TAGUCHI, S., AND E. A. LAWS. 1988. On the micropar- habitat differences in the North Pacific. J. Exp. Mar. titles which pass through glass-fiber filter type GF/F Biol. Ecol. 12: 187-217. in coastal and open waters. J. Plankton Res. 10: 999- MUNAWAR, M., I. F. MUNAWAR, P. E. Ross, AND A. DAGE- 1008. NAB. 1982. Microscopic evidence of phytoplankton VENRICK, E. L. 1987. On fluorometric determinations of passing through glass-fibre filters and its implication filter-retained pigments. Limnol. Oceanogr. 32: 492- for chlorophyll analysis. Arch. Hydrobiol. 94: 520- 493. 528. -, S. L. CUMMINGS, AND C. A. KEMPER. 1987. Pi- OLSON, R. J., S. W. CHISHOLM, E. R. ZETTLER, M. A. coplankton and the resulting bias in chlorophyll re- ALTABET, AND J. A. DUSENBERRY. 1990. Spatial and tained by traditional glass-fiber filters. Deep-Sea Res. temporal distributions of prochlorophyte picoplank- 34: 195 1-1956. ton in the North Atlantic Ocean. Deep-Sea Res. 37: -, AND T. L. HAYWARD. 1984. Determining chlo- 1033-1051. rophyll on the 1984 CalCOFI surveys. Calif. Coop. PAK, H., D. A. KIEFER, AND J. C. KITCHEN. 1988. Me- Oceanic Fish. Invest. Rep. 25: 74-79. ridional variations in the concentration of chlorophyll and microparticles in the North Pacific Ocean. Deep- Submitted: 25 June 1992 Sea Res. 35: 1151-l 171. Accepted: 17 March 1993 PHINNEY, D. A., AND C. S. YENTSCH. 1985. A novel Revised: 28 April 1993

Limnol Oceanogr., 38(8), 1993, 1818-1822 0 1993, by the American Society of Limnology and Oceanography, Inc effects on epibiont communities: Epibiont pigmentation effects

Abstract-We observed the development of epi- Epibionts on zooplankton are thought to in- biont communities on freshwater crustacean zoo- crease the visibility and vulnerability of their plankton from 12 July-3 1 August 1991 in three ponds: one with planktivorous fish, a second with substrate organisms to visually orienting planktivorous larval salamanders, and a third with (Green 1974; Evans et al. 1979; planktivorous fish added midway through the sam- Willey et al. 1990; Chiavelli et al. 1993). The pling period. Prevalence, on zooplankton, of pig- effect of epibionts on zooplankton visibility mented euglenoid epibionts (Colucium vesiculosum), seems obvious (Fig. l), has been long suggested alone and with unpigmented peritrich (Vor- ticella campanula and Epistylis sp.), was significantly (Green 1953, 1974), and is supported by re- reduced by the addition of planktivorous fish in the lated work which shows that eyespots and oth- third pond, while the prevalence of peritrich ciliates er forms of cladoceran pigmentation affect sus- alone was not. Neither of the ponds with unaltered ceptibility to predators (Zaret 1972; Mellors planktivore regimes showed a similar pattern of change in prevalence of pigmented and unpigmented 1975; Konecny et al. 1982). Experimental work epibionts on zooplankton. The relative loss of pig- demonstrating that epibiont-enhanced visibil- mented euglenoid epibionts, after planktivore ad- ity is a mechanism of increased vulnerability dition, suggests that epibiont color or contrast may to planktivores is still lacking (Threlkeld et al. increase the susceptibility of their substrate organ- 1993). isms to planktivores. In a series of subalpine ponds in Colorado, we observed that many epibiotic taxa, includ- ing some pigmented photosynthetic (e.g. Co- Acknowledgments lacium vesiculosum) and some unpigmented We thank Phil Mason and Terry Schneider for assis- heterotrophic taxa (e.g. Vorticella campanula), tance, and D. A. Chiavelli for editorial comments. This study was funded, in part, by a grant from the often coexist on the same substrate organisms Colorado Division of Wildlife for monitoring salamander (Threlkeld and Willey 1993; Willey and populations in western Colorado. Threlkeld 1993). Here we present evidence that