16
C.C. Parrish
operating properly, the peak areas of each component in each of these standards
should be close in anyone injection. Some of the components of these quantitative
standards are also useful for identification purposes.
The reproducibility of repeated determinations of sample peak areas should be
regularly checked. For major peaks (>5% of the total fatty acids), Christie (1989)
indicates the coefficient of variation (CV) should be < 10%. For replicate injections of the same extract, a CV <5% for the major peaks is routinely expected in
my laboratory. However, it is important to replicate not just at the analysis step.
Samples are routinely extracted in triplicate for every lipid determination in my
laboratory and a CV < 10% is anticipated. Ideally, replication should occur at the
sampling stage in which significantly larger CVs are expected. CVs >20% have
been obtained in my laboratory for settling particulate matter from different
collection tubes on the same sediment trap frame deployed in the water column.
A new gas chromatographic method for profiling marine samples has recently
been developed at Memorial University of Newfoundland (Yang et aI., 1996). In
terms of the information provided, this method is located between detailed fatty
acid analyses and the class analyses provided by the Chromarod-Iatroscan system.
The carbon number profiling method groups compounds according to carbon
number within each class. By summing the groups of molecular species within
each class, total class amounts are obtained. This method uses a short (5.5 m)
nonpolar capillary column (DB-5: J. & w. Scientific, Folsom, CA) that is programmed to be heated through a wide temperature range (62-340°C). It has been
successfully applied to neutral lipids in a wide range of marine samples (Yang et
aI., 1996) but has yet to be applied to polar lipids. However, this approach is used
routinely for phospholipids in plasma samples after dephophorylation with phospholipase C (Myher and Kuksis, 1984).
1.3. Conclusion
There are probably nearly as many different variations on standard extraction
(e.g., Wainman and Lean, 1994; Ackman, 1993; Nelson, 1991; Shaikh, 1986) and
derivatization techniques (e.g., Liu, 1994) as there are laboratories that routinely
perform these procedures. The purpose of this chapter has been to indicate
procedures that have been used successfully over the past several years in my
laboratory for both freshwater and marine samples. These procedures have been
intercalibrated among themselves as well as with those used routinely in other
laboratories. For example, Chromarod-Iatroscan analyses have been compared
both with gravimetry (Parrish, 1987; Gardner et aI., 1985) and with GC analyses
(Yang et aI., 1996; Parrish et aI., 1995). Such intercalibration is to be recommended, especially for new users of any analytical approaches. A useful first step
is to repeat the chosen extraction or derivatization procedure on the sample and
then to compare the results of the original determination with those of the repeat
procedure. If the repeat analysis yields no more lipid than is present in a total
system blank, then one knows that one is at least close to the operational definition
C.C. Parrish
operating properly, the peak areas of each component in each of these standards
should be close in anyone injection. Some of the components of these quantitative
standards are also useful for identification purposes.
The reproducibility of repeated determinations of sample peak areas should be
regularly checked. For major peaks (>5% of the total fatty acids), Christie (1989)
indicates the coefficient of variation (CV) should be < 10%. For replicate injections of the same extract, a CV <5% for the major peaks is routinely expected in
my laboratory. However, it is important to replicate not just at the analysis step.
Samples are routinely extracted in triplicate for every lipid determination in my
laboratory and a CV < 10% is anticipated. Ideally, replication should occur at the
sampling stage in which significantly larger CVs are expected. CVs >20% have
been obtained in my laboratory for settling particulate matter from different
collection tubes on the same sediment trap frame deployed in the water column.
A new gas chromatographic method for profiling marine samples has recently
been developed at Memorial University of Newfoundland (Yang et aI., 1996). In
terms of the information provided, this method is located between detailed fatty
acid analyses and the class analyses provided by the Chromarod-Iatroscan system.
The carbon number profiling method groups compounds according to carbon
number within each class. By summing the groups of molecular species within
each class, total class amounts are obtained. This method uses a short (5.5 m)
nonpolar capillary column (DB-5: J. & w. Scientific, Folsom, CA) that is programmed to be heated through a wide temperature range (62-340°C). It has been
successfully applied to neutral lipids in a wide range of marine samples (Yang et
aI., 1996) but has yet to be applied to polar lipids. However, this approach is used
routinely for phospholipids in plasma samples after dephophorylation with phospholipase C (Myher and Kuksis, 1984).
1.3. Conclusion
There are probably nearly as many different variations on standard extraction
(e.g., Wainman and Lean, 1994; Ackman, 1993; Nelson, 1991; Shaikh, 1986) and
derivatization techniques (e.g., Liu, 1994) as there are laboratories that routinely
perform these procedures. The purpose of this chapter has been to indicate
procedures that have been used successfully over the past several years in my
laboratory for both freshwater and marine samples. These procedures have been
intercalibrated among themselves as well as with those used routinely in other
laboratories. For example, Chromarod-Iatroscan analyses have been compared
both with gravimetry (Parrish, 1987; Gardner et aI., 1985) and with GC analyses
(Yang et aI., 1996; Parrish et aI., 1995). Such intercalibration is to be recommended, especially for new users of any analytical approaches. A useful first step
is to repeat the chosen extraction or derivatization procedure on the sample and
then to compare the results of the original determination with those of the repeat
procedure. If the repeat analysis yields no more lipid than is present in a total
system blank, then one knows that one is at least close to the operational definition