Spectral fluorescence signatures in the characterization of phytoplankton community composition

Fluorescence spectral signatures from 28 algal cultures aredescribed.The cultures are split into four groups accordingto their accessory pigments. Phycocyanin and phycoerythrin,characteristic pigments of cyanobacteria, form groups I andII. The characteristic pigment found in group III is chlorophyllb (green and rasinophyte algae) and in group IV it is chlorophyllc (diatoms, dinophytes and some other algae).This preliminarycatalogue of spectral signatures was used to characterize fivenatural phytoplankton communities from brackish water environmentsas a comparison with phytoplankton species found in the samples.Accessory pigments such as phycocyanin and phycoerythrin, characterizinggroups I and II, can be used for identification without confusion.Distinguishing between groups III and IV is more complicated,because their accessory pigments do not have their own fluorescence.These groups can be characterized by increased fluorescenceof chlorophyll a induced by energy excited through chlorophyllb and c. The possibilities of applying the spectral fluorescencesignatures approach to the characterization of natural communitiesare discussed.     

Phytoplankton of an eutrophic tropical reservoir: comparison of biomass estimated from counts with chlorophyll-a biomass from HPLC measurements

The seasonal variation of phytoplankton in an eutrophic tropical reservoir was evaluated through photosynthetic pigments analyzed by HPLC. The contributions of algal classes to total chlorophyll a (TChl-a) were estimated by two procedures. The first one used fixed marker pigment/chlorophyll a ratio available from culture studies of the major species of each class. In the second procedure, a matrix factorization program (CHEMTAX) was used to analyze the pigment data. The pigment data were compared with carbon biomass estimated from microscope analysis. A significant correlation between total chlorophyll a (measured by HPLC) and total biomass was obtained, indicating only a slight variation in the content of algal chlorophyll a when compared to its fluctuations in carbon biomass. The interpretation of pigment data with CHEMTAX resulted in a good agreement with biomass. Although displaying some differences, the general pattern of the phytoplankton community dynamics and the major shifts in composition, biomass and the cyanobacterial bloom were evidenced. In contrast, Chl-a biomass estimates from fixed Xan/Chl-a ratios presented poor agreement with microscope data and did not register the principal changes in phytoplankton. Our results also highlighted the needs of better understanding of the relationships between marker pigments, chlorophyll-a and algal biomass.     

Algal pigment diversity in coastal sediments from Kerguelen (sub-Antarctic Islands) reflecting local dominance of green algae,euglenoids and diatoms

Summary High Performance Liquid Chromatography analysis of algal pigments from inter- and subtidal (deep and shallow) sediments from the Kerguelen Islands showed clear differences in the pigment composition at the different stations. High concentrations of chlorophyll c and fucoxanthin were present at all locations, indicating significant diatom densities, chlorophyll b was detected at all sites. At one station the other green algal pigments were also present; here green algae contributed more to chlorophyll a concentrations than diatoms, as estimated by using pigment ratios and microscopic observations. At another location chlorophyll b was associated with a high concentration of diadinoxanthin, indicating an abundance of euglenoids. This indicates that chemotaxonomy can be powerful tool in microphytobenthos studies since enumeration of living cells are difficult as many algae are attached to sediment particles (epipsammic algae). Ways of diagenesis, carotenoid degradation and the role of grazing are briefly mentioned. Phaeophorbide a-like pigments were the most significant chlorophyll a degradation products, with concentrations up to 110 g · g^–1 dry weight sediment, i.e. 10 times the chlorophyll a concentration. Some taxonomic estimations, based on pigments ratios, and their limits, are discussed.     

Phytoplankton Assemblages and Their Dominant Pigments in the Nervion River Estuary

In the Nervion River estuary surface samples were taken from March to September 2003 at six sites covering most of the salinity range with the aim to know the biomass and taxonomic composition of phytoplankton assemblages in the different segments. Nine groups of algae including cyanobacteria, diatoms, dinoflagellates, chlorophytes, prasinophytes, euglenophytes, chrysophytes, haptophytes, raphidophytes and cryptophytes were identified by means of a combination of pigment analysis by high-performance liquid chromatography (HPLC) and microscopic observations of live and preserved cells. Diatoms, chlorophytes and cryptophytes were the most abundant algae in terms of cells number, whereas fucoxanthin, peridinin, chlorophyll b (Chl b) and alloxanthin were the most abundant auxiliary pigments. Based on multiple regression analysis, in the outer estuary (stations 0, 1, 2 and 3) about 93% of the chlorophyll a (Chl a) could be explained by algae containing fucoxanthin and by algae containing Chl b, whereas in the rest of the estuary most of the Chl a (about 98%) was accounted for by fucoxanthin, Chl b and alloxanthin containing algae. The study period coincided with that of most active phytoplankton growth in the estuary and fucoxanthin was by far the dominant among those signature pigments. Several diatoms, chrysophytes, haptophytes and raphydophytes were responsible for fucoxanthin among identified species. Besides, dinoflagellates with a pigment pattern corresponding to chrysophytes and type 4 haptophytes were identified among fucoxanthin-bearing algae. Cryptophytes were the most abundant species among those containing alloxanthin. The maximum of Chl b registered at the seaward end in April coincided with a bloom of the prasinophytes Cymbomonas tetramitiformis, whereas the Chl b maxima in late spring and summer were accounted for by prasinophytes in the middle and outer estuary and by several species of chlorophytes in the middle and inner estuary. Other Chl b containing algae were euglenophytes and the dinoflagellate Peridinium chlorophorum. Dinoflagellates constituted generally a minor component of the phytoplankton.     

Determination of chlorophylls and carotenoids of marine phytoplankton: separation of chlorophyll a from divinylchlorophyll a and zeaxanthin from lutein

A rapid reverse-phase HPLC method is presented for the identificationand quantification of most of the phytoplankton pigments. Thismethod yields the resolution of divinyl-chlorophyll a and chlorophylla, as well as the partial resolution of lutein and zeaxanthin,and of divinyl-chlorophyll b and chlorophyll b. In addition,chlorophylls c_1,2 and c₃ are well resolved. The analysis timefor one sample is 20 mm, which makes this method particularlysuited when large numbers of samples have to be processed.     

HPLC analysis of phytoplankton pigments from Lake Kinneret with special reference to the bloom-forming dinoflagellate Peridinium gatunense (Dinophyceae) and chlorophyll degradatio products

Photosynthetic pigments extracted from the paniculate materialof the water column of Lake Kinneret were studied throughoutthe periods of May 1988-June 1989, and November 1993-November1994, by means of HPLC. The temporal and vertical variationof the pigment suite found agreed with the microscopically determinedphytoplankton record. The regression calculations of taxon-specificbiomass with the corresponding signature pigments suggest thatpigment analysis may be a useful tool for the monitoring ofbloom-forming species, e.g. the dinoflagellate Peridinium gatunenseNygaard. The HPLC pigment analysis permitted the identificationand quantification of chlorophyll degradation products, providingfor the first time information about their composition in LakeKinneret. Chlorophyllide a was the major detectable degradationproduct of chlorophyll a, varying between 1 and 9% of the chlorophylla concentration. Other chlorophyll a derivatives appeared mostlyin minor quantities. Pheophytin a was virtually lacking in allthe samples. Removal rates of pigments, measured by sedimentationtraps, indicated that the degradation of chlorophyll a via chlorophyllidea is a dynamic process that continues during the sedimentationof the phytoplankton particles.     

Photosynthetic pigments of plankton in some karst lakes of central Russia

Phytoplankton pigments were studied in seven soft water, weakly mineralized karst lakes in central Russia (Vladimir oblast). The lakes Kshchara, Sankhar, Yukhor, Bol’shoye Poridovo, Svetlen’koye, Bol’shie, and Malye Garavy were investigated. The seasonal dynamics, vertical distribution of chlorophyll a, and its content in the phytoplankton biomass were considered. The relationship between chlorophyll a content and abiotic factors was analyzed. In addition to algae pigments, high concentrations of bacterial chlorophyll d were recorded in meta- and hypolimnetic layers of lakes Yukhor, Kshchara, Sankhar, and Svetlen’koye.     

Factors influencing the short-term variation in phytoplankton composition and biomass in coral reef waters

The short-term temporal dynamics of phytoplankton composition was compared among coral reef waters, the adjacent ocean and polluted harbour water from July until October along the south-western coast of Curaçao, southern Caribbean. Temporal variations in phytoplankton pigment 'fingerprints' (zeaxanthin, chlorophyll b, 19'-hexanoyloxyfucoxanthin, fucoxanthin, 19'-butanoyloxyfucoxanthin, chlorophyll c₂ and c₃ relative to chlorophyll a) in the ocean were also observed in waters overlying the reef. However, with respect to specific pigments and algal-size distribution, the algal composition in reef waters was usually slightly different from that in the oceanic water. Phytoplankton biomass (chlorophyll a) was either higher or lower than in the oceanic water. The relative amount of fucoxanthin and peridinin was usually higher, and the relative and absolute amount of zeaxanthin was significantly lower than in oceanic water. Zeaxanthin-containing Synechococci were significantly reduced in reef water. Average algal cell size increased from the open water to the reef and the harbour entrance. Large centric diatoms (>20 m Ø) were better represented in reef than in oceanic water. In reef-overlying waters, the nitrate and nitrite concentrations were higher than in oceanic water. In front of the town, anthropogenic eutrophication (sewage discharge and ground water seepage) resulted in higher NH₄, NO₃ and PO₄ concentrations than at other reef stations. This concurred with significantly enhanced phytoplankton biomass (chlorophyll a), chlorophyll c₂ and peridinin amounts at Town Reef compared with the other reef stations. Polluted harbour water usually showed the highest phytoplankton biomass of all stations, dominated by diatoms and dinoflagellates. Conditions in reef waters and harbour water promoted the occurrence and the relative abundance of diatoms and dinoflagellates. Harbour water did not influence the phytoplankton composition and biomass at reef stations situated >5 km away from the harbour entrance. We conclude that phytoplankton undergoes a shift in algal composition during transit over the reef. The dominant processes appear to be selective removal of zeaxanthin-containing Synechococcus (by the reef benthos) and (relative) increase in diatoms and dinoflagellates. The difference in the phytoplankton composition between reef and oceanic waters tends to increase with decreasing dilution of reef water with ocean water.     

Diel activity of Gammarus pulex (Crustacea) in a South Swedish stream: comparison of drift catches vs baited traps

The benthos of a perennially ice-covered Antarctic lake, Lake Hoare, contained three distinct signatures of lipophilic pigments. Cyanobacterial mats found in the moat at the periphery of the lake were dominated by the carotenoid myxoxanthophyll; carotenoids: chlorophyll a ratios in this high light environment ranged from 3 to 6.8. Chlorophyll c and fucoxanthin, pigments typical of golden-brown algae, were found at 10 to 20 m depths where the benthos is aerobic. Anaerobic benthic sediments at 20 to 30 m depths were characterized by a third pigment signature dominated by a carotenoid, tentatively identified as alloxanthin from planktonic cryptomonads, and by phaeophytin b from senescent green algae. Pigments were not found associated with alternating organic and sediment layers. As microzooplankton grazers are absent from this closed system and transformation rates are reduced at low temperatures, the benthos beneath the lake ice appears to contain a record of past phytoplankton blooms undergoing decay.     

Does the chlorophyll a content of phytoplankton vary with trophic status in lakes on the New Zealand central volcanic plateau?

Recently, it has been shown that ratios of chlorophyll a toparticulate phosphorus (Chl a/PP) and chlorophyll a to particulatenitrogen (Chl a/PN) were significantly higher in eutrophic thanoligo/mesotrophic waters in 17 lakes on the central volcanicplateau, North Island, New Zealand. This difference was thoughtto be due to an increase in the chlorophyll a content of phytoplanktonin these eutrophic lakes. Corresponding measurements of chlorophylla and phytoplankton cell volume made during this study do notsupport this hypothesis. However, ratios of chlorophyll a toadenosine triphosphate and estimates of percentage phytoplanktonbiomass were significantly higher (P<0.05) in our eutrophicthan oligo/mesotrophic samples, suggesting that Chl a/PP andChl a/PN may be high in eutrophic waters simply because phytoplanktoncomprise more of the total microbial biomass. This hypothesisis supported by a strong linear relationship (r=0.88, P<0.001)between Chl a/PP and percentage phytoplankton biomass in sixof our study lakes where corresponding measurements were made.     

The use of spectral fluorescence methods to detect changes in the phytoplankton community

In vivo fluorescence methods are efficient toolsfor studying the seasonal and spatial dynamics ofphytoplankton. Traditionally the measurements are madeusing single excitation-emission wavelengthcombination. During a cruise in the Gulf of Riga(Baltic Sea) we supplemented this technique bymeasuring the spectral fluorescence signal (SFS) andfixed wavelength fluorescence intensities at theexcitation maxima of main accessory pigments. Thesemethods allowed the rapid collection of quantitativefluorescence data and chemotaxonomic diagnostics ofthe phytoplankton community. The chlorophylla-specific fluorescence intensities (R) and thespectral fluorescence fingerprints were analysedtogether with concentrations of chlorophyll a indifferent algal size-groups, phytoplankton biomass andtaxonomic position. The lower level of R in thesouthern gulf was related to the higher proportion ofcyanobacteria relative to total biomass and the lowerabundance of small algae. The phycoerythrinfluorescence signal was obviously due to the largecyanobacteria. The basin-wide shift in the shape ofchlorophyll a excitation spectra was caused bythe variable proportions of differently pigmentedcyanobacteria, diatoms and cryptomonads. This revised version was published online in July 2006 with corrections to the Cover Date.     

Phytoplankton and zooplankton development in a lowland, temperate river

The longitudinal and seasonal patterns of plankton developmentwere examined over 2 years in a lowland, temperate river: theRideau River (Ontario, Canada). Following an initial decreasein phytoplankton and zooplankton biomass as water flowed fromthe headwaters into the Rideau River proper, there was an increasein chlorophyll a (chl a) and zooplankton biomass with downstreamtravel. At approximately river km 60, both phytoplankton andzooplankton reached their maximum biomass of 27 µg l^–1(chl a) and 470 µg l^–1 (dry mass), respectively.Downstream of river km 60, the biomass of both planktonic communitiesdeclined significantly despite increasing nutrient concentrationsand favorable light conditions. These downstream declines maybe due to the feeding activity of the exotic zebra mussel (Dreissenapolymorpha) which was at high density in downstream reaches(>1000 individuals m^–2). There was no evidence forlongitudinal phasing of phytoplankton and zooplankton, as increasesand decreases in chl a and zooplankton biomass appeared to coincide.Overall, chl a was best predicted by total phosphorus (R²=0.43),whereas zooplankton biomass was best predicted by chl a (R²=0.20).There was no evidence for significant grazing effects of zooplanktonon phytoplankton biomass.     

In vivo fluorescence excitation and absorption spectra of marine phytoplankton: I. Taxonomic characteristics and responses to photoadaptation

Absorption and fluorescence excitation spectra were measuredfor batch cultures of five species of marine phytoplankton grownunder high and low light. These spectra were examined for propertiescharacteristic of taxonomic position and of photoadaptive response.While regions of absorption and excitation of chlorophyll afluorescence diagnostic of pigment composition were identifiable,photoadaptive response had greater influence on spectral variability.Although reduced growth irradiance caused changes in both theabsorption and fluorescence excitation spectra, the fluorescenceexcitation spectrum appears to be more sensitive to alterationsin the ambient light field for growth than does the absorptionspectrum. For a single species. the fluorescence excitationspectrum for a sample grown at low irradiance showed greaterstructure than that for the sample grown at a high irradiance.Under low light conditions, the excitation of chlorophyll afluorescence by accessory pigments increased relative to theexcitation by chlorophyll a itself The highest fluorescenceyields occur in the blue-green region of the spectrum, correspondingto bands of peak absorption by the accessory pigments. Changesin absorption spectra are less marked, but two features recur.First. in the blue-green region of the spectrum from -500–560nm. absorption is enhanced in the low-light cells relative tothat of the high-light cells. Second, the ratio of absorptionat 435 nm to that at 676 nm was greater for the high-light cells.Correlating changes in pigment concentrations were observed.The influence of photoadaptation on the properties of fluorescenceexcitation spectra is as great or greater than the influenceof pigment complements characteristic of specific algal taxa.     

Bias in satellite-derived pigment measurements due to coccolithophores and dinoflagellates

Satellite-derived estimates of phytoplankton pigments are thoughtto be affected by the phytoplankton species composition. Measurementsof surface algal chlorophyll and satellite-derived pigment werecompared for waters containing coccolithophores and dinoflagellates.Satellite-derived chlorophyll concentration was underestimatedby a factor of 2–3 in a patch of the large coccolithophore,Umbilicosphaera sibogae, and also in a bloom of the dinoflagellate,Gonyaulax polyedra. Overall abundance and species-specific propertiessuch as light scatter and vertical migration probably causedthese results.     

Implications of rapid changes in chlorophyll-<Emphasis Type="Italic">a</Emphasis> of plankton,epipelon, and epiphyton in a Pampean shallow lake: an interpretation in terms of a conceptual model

Biomass assessments of algae in wetlands usually include only the phytoplankton community without considering the contribution of other algal associations to total algal biomass. This omission prevents an accurate evaluation of the phytoplankton community as an integral part of the total ecosystem. In the present work, the biomass contributions (expressed as chlorophyll-a content per m² of lake) of phytoplankton, epiphyton on both submerged and emergent macrophytes, and epipelon were measured in Lacombe Lake, Argentina, for the purpose of (1) establishing the relative importance of the phytoplankton and (2) evaluating the entire contribution of algal biomass within the context of the Goldsborough & Robinson conceptual model. Our sampling was carried out monthly for a year in sites representative of different conditions with respect to water depth and type of macrophytes. Physicochemical analyses of water were performed following standard methods. Plankton was collected in a five-level profile at deeper stations and in subsurface samples at the shallow one. Samples of sediment obtained with corers were collected for epipelon sampling and segments of plants were cut at different levels, so as to obtain the epiphytes by scraping. Pigment was extracted with aqueous acetone and calculations were made by means of the Lorenzen equation. According to the Goldsborough & Robinson model, a Lake State developed here during the winter (phytoplankton maxima: 150 mg chlorophyll-a per m²). Then, through the subsequent growth of the submerged macrophytes, an Open State was observed, characterized by a maximum epiphyton biomass (at 3,502 mg chlorophyll-a per m²) along with lower levels of phytoplankton biomass. The epiphytic algae on the emergent macrophytes were always present but attained only relatively low biomass values (maximum: 120 mg of chlorophyll-a per m² in February). The epipelon biomass varied between 50 and 252 mg chlorophyll-a per m², registering a considerable contribution of settled algae from the water column (phytoplankton). This study contributes to our knowledge of wetland dynamics through its assessment of the rapid changes in the relative contributions of both planktonic and attached algae to the total algal biomass within the context of specific environmental factors. Guest editors: U. M. Azeiteiro, I. Jenkinson & M. J. Pereira Plankton Studies     

Detection and Quantification of Snow Algae with an Airborne Imaging Spectrometer

We describe spectral reflectance measurements of snow containing the snow alga Chlamydomonas nivalis and a model to retrieve snow algal concentrations from airborne imaging spectrometer data. Because cells of C. nivalis absorb at specific wavelengths in regions indicative of carotenoids (astaxanthin esters, lutein, β-carotene) and chlorophylls a and b, the spectral signature of snow containing C. nivalis is distinct from that of snow without algae. The spectral reflectance of snow containing C. nivalis is separable from that of snow without algae due to carotenoid absorption in the wavelength range from 0.4 to 0.58 μm and chlorophyll a and b absorption in the wavelength range from 0.6 to 0.7 μm. The integral of the scaled chlorophyll a and b absorption feature (I_0.68) varies with algal concentration (C_a). Using the relationship C_a = 81019.2 I_0.68 + 845.2, we inverted Airborne Visible Infrared Imaging Spectrometer reflectance data collected in the Tioga Pass region of the Sierra Nevada in California to determine algal concentration. For the 5.5-km² region imaged, the mean algal concentration was 1,306 cells ml⁻¹, the standard deviation was 1,740 cells ml⁻¹, and the coefficient of variation was 1.33. The retrieved spatial distribution was consistent with observations made in the field. From the spatial estimates of algal concentration, we calculated a total imaged algal biomass of 16.55 kg for the 0.495-km² snow-covered area, which gave an areal biomass concentration of 0.033 g/m².     

Growth of grana from "primary" thylakoids in Phaseolus vulgaris

The plastids of young dark-grown bean leaves, exposed to periodiclight are agranal, devoid of chlorophyll b and contain primarythylakoids and chlorophyll a. Transfer of these plants to continuousillumination results in synthesis of new chlorophyll a, chlorophyllb and grana. This study was done in order to study whether andhow the grana are formed from preexisting primary thylakoids.¹⁴C--aminolevulinic acid was used to label the chlorophyll aof the primary thylakoids, and its fate was studied after transferof the plants to continuous light. It was found that chlorophyll b and grana become ¹⁴C-labelled.The total radioactivity of chlorophyll b per bean increasedwith the parallel decrease of that of chlorophyll a. All subchloroplastfractions, obtained after digitonin disruption of chloroplasts,contained chlorophyll a of equal specific radioactivity. Thespecific radioactivity of chlorophyll b was lower than thatof chlorophyll a, and, in addition, it was lower in the granathan in the stroma lamellae fraction. The data suggest that chlorophyll b is formed from chlorophylla; the grana are formed by stacking of preexisting primary thylakoids;chlorophyll b is synthesized faster in the grana than the stromalamellae; the newly formed chlorophyll a molecules are distributedat random throughout the developing photosynthetic membraneand not on specific growing sites. (Received April 24, 1976; )     

Studies on chlorophyllase of Chlorella protothecoides II. Formation of atypical chlorophyllide a

When chlorophyll a was incubated with a preparation of chlorophyllaseextracted with Triton X-100 from methanol-acetone powder ofChlorella protothecoides, the substrate was changed to chlorophyllidea, and subsequently to an atypical form of chlorophyllide a.The formation of an atypical form of chlorophyllide was notdetected in the reaction with chlorophyll b as substrate, norin the reaction with another preparation of chlorophyllase extractedfrom the algal cells. (Received August 18, 1969; )     

Assessment of grazing by the freshwater copepod Diaptomus minutus using carotenoid pigments: a caution

In order to assess feeding selectivity in freshwater zooplankton, we conducted feeding trials using Diaptomus minutus isolated from two Wisconsin lakes. Copepods were fed an algal assemblage comprised of an equal biomass of a centric diatom, a cryptomonad and a coccal green alga. The total amounts of photopigments were tracked using high-performance liquid chromatography. The removals of carotenoids and a-type phorbins (chlorophyll a and phaeopigments) from feeding suspensions were compared with their presence in the guts of animals, in fecal pellets, and in the final suspension. Diaptomus minutus generally removed either the diatoms primarily or all three algal cells equally. These removals were not reflected in gut extracts of the animals, however, where alloxanthin (marker of cryptomonads) was always present, but where fucoxanthin and diadinoxanthin (markers of diatoms) were never observed. Pigment disappearance was variable for total a-type phorbins, but frequently >90% for carotenoids, particularly for fucoxanthin and diadinoxanthin. Phaeophytin a was the major a-type phorbin detected after gut passage. Our results indicate that evaluations of zooplankton grazing which assume that algal carotenoids remain detectable throughout gut passage must be made with substantial caution and that differences in pigment processing are likely to occur among zooplankton species. Furthermore, grazing experiments designed to evaluate decreases of specific pigments in feeding suspension, rather than their appearance in animals' guts, may prove a more valuable approach to understanding the feeding selectivity of copepods.      

The presence of chlorophyll b and the estimation of phaeopigments in marine phytoplankton

A reverse-phase h.p.l.c. technique was used to estimate theconcentration of chlorophyll b in phytoplankton cultures, fecalpellets of Calanus pacificus, and suspended paniculate matterfrom the Central North Pacific, Oregon coastal waters, and DabobBay (a temperate fjord in Puget Sound, WA, USA). The purposewas to assess the distribution of this pigment in the euphoticzone and its effect on the fluorometnc estimation of phaeopigments.Analyses of natural waters confirm high chlorophyll b concentrations(median mass ratio of b:a > 0.3) at the depth of the chlorophylla maximum in tropical waters while values for temperate planktonare relatively low (median mass ratio of chl b:a = 0.05) andpatchy. Zooplankton fecal pellets showed a significant enrichmentin chlorophyll b, suggesting grazing as a mechanism to explainhigh concentrations of this pigment at the bottom of the euphoticzone. It is estimated that the presence of chlorophyll b couldcause an average overestimation of phaeopigment concentrationby the fluorometnc technique of 38% between 0 and 200 m in theCentral North Pacific. This effect is more pronounced at thelayer of chlorophyll b maximum (120–140 m). ¹Present address: Marine Biology Research Division, A-002, ScrippsInstitution of Oceanography, La Jolla, CA 92093, USA