75
A great variety of HPLC equipment is now available, and the number of methods described
in the literature has been increasing steadily (Roy, 1987). With our own methods, which
resemble those of Abaychi and Riley (1979) and Mantoura and Llewellyn (1983), we obtain
chromatograms of very high quality (Fig. 9). Recent versions of our methods (one isocratic,
the other reversed-phase) have been described in Gieskes and Kraay (1983b), Gieskes and
Kraay (1986a) and Gieskes et al. (1988). A major difference with the method of Mantoura
and Llewellyn (1983) is that we did not use an ion-pairing reagent: buffering of solvents with
a phosphate buffer (ion strength 0.05, pH 7) and extraction of the sample in 2% ammonium
acetate in methanol was found to be sufficient to achieve good resolution of all
chloropigments, including ChI c1 +2, ChI c3, chlorophyllide a, etc. (see Fig. 9). Zeaxanthin
and lutein separation is notoriously difficult (Hager and Meyer-Bertenrath, 1967), but
co-occurrence in natural phytoplankton is in our experience exceptional. Ben-Amotz et al.
described a method for separation in 1982, in J. Phycol. 18:529-537. The two pigments have
a different absorption spectrum in all organic solvents (we prepare lutein standard from ferns,
zeaxanthin from maize), and this can be the test for which is present in a sample.
Our measurements of chI a after separation by HPLC are normally well-correlated with chI
a concentrations measured with the conventional fluorometric and absorption acidification
methods (see previous sections), both in cultures and in natural populations, in a wide variety
of water types (Fig. lOa). The regression equation is never far from Y = 0.67 X, so the
conventional methods give a mean overestimate of 30 %. This can normally be ascribed to the
presence of chI a allomers and isomers, chlorophyllide a, and unknown substances detected
as tiny but sometimes numerous peaks in the chromatograms. The large discrepancy between
HPLC and the conventional methods reported by Jacobsen (1978), Sartory (1985) and
Mantoura and Llewellyn (1983) has never been found by us, not even in the turbid Wadden
Sea, the polluted Rhine river, or Northwest European coastal waters (see Fig. lOa). On the
other hand, Trees et al. (1985) stated that HPLC measurements overestimate chI a
concentrations! Our own comparison between methods of TLC (Gieskes and Kraay, 1978),
HPLC (Gieskes and Kraay, 1983b, 1984, 1986, 1988) and conventional methods was always
quite satisfactory.
Nevertheless, it is clear that intercalibration remains a necessity. An improve intercalibration
exercise is presently being coordinated by R.R. Bidigare, now at Hawaii University,
Honolulu. The JGOFS program (Joint Global Ocean Flux Studies) for which this
intercalibration was designed, tries to define and estimate of the extent of oceanic carbon
transformations. Phytoplankton plays a vital role in the global carbon budget; the synoptic
mapping of these organisms by satellite which is considered essential for this program,
A great variety of HPLC equipment is now available, and the number of methods described
in the literature has been increasing steadily (Roy, 1987). With our own methods, which
resemble those of Abaychi and Riley (1979) and Mantoura and Llewellyn (1983), we obtain
chromatograms of very high quality (Fig. 9). Recent versions of our methods (one isocratic,
the other reversed-phase) have been described in Gieskes and Kraay (1983b), Gieskes and
Kraay (1986a) and Gieskes et al. (1988). A major difference with the method of Mantoura
and Llewellyn (1983) is that we did not use an ion-pairing reagent: buffering of solvents with
a phosphate buffer (ion strength 0.05, pH 7) and extraction of the sample in 2% ammonium
acetate in methanol was found to be sufficient to achieve good resolution of all
chloropigments, including ChI c1 +2, ChI c3, chlorophyllide a, etc. (see Fig. 9). Zeaxanthin
and lutein separation is notoriously difficult (Hager and Meyer-Bertenrath, 1967), but
co-occurrence in natural phytoplankton is in our experience exceptional. Ben-Amotz et al.
described a method for separation in 1982, in J. Phycol. 18:529-537. The two pigments have
a different absorption spectrum in all organic solvents (we prepare lutein standard from ferns,
zeaxanthin from maize), and this can be the test for which is present in a sample.
Our measurements of chI a after separation by HPLC are normally well-correlated with chI
a concentrations measured with the conventional fluorometric and absorption acidification
methods (see previous sections), both in cultures and in natural populations, in a wide variety
of water types (Fig. lOa). The regression equation is never far from Y = 0.67 X, so the
conventional methods give a mean overestimate of 30 %. This can normally be ascribed to the
presence of chI a allomers and isomers, chlorophyllide a, and unknown substances detected
as tiny but sometimes numerous peaks in the chromatograms. The large discrepancy between
HPLC and the conventional methods reported by Jacobsen (1978), Sartory (1985) and
Mantoura and Llewellyn (1983) has never been found by us, not even in the turbid Wadden
Sea, the polluted Rhine river, or Northwest European coastal waters (see Fig. lOa). On the
other hand, Trees et al. (1985) stated that HPLC measurements overestimate chI a
concentrations! Our own comparison between methods of TLC (Gieskes and Kraay, 1978),
HPLC (Gieskes and Kraay, 1983b, 1984, 1986, 1988) and conventional methods was always
quite satisfactory.
Nevertheless, it is clear that intercalibration remains a necessity. An improve intercalibration
exercise is presently being coordinated by R.R. Bidigare, now at Hawaii University,
Honolulu. The JGOFS program (Joint Global Ocean Flux Studies) for which this
intercalibration was designed, tries to define and estimate of the extent of oceanic carbon
transformations. Phytoplankton plays a vital role in the global carbon budget; the synoptic
mapping of these organisms by satellite which is considered essential for this program,