samples at the four wavelengths was not con- 

 stant, the departure from linearity is not 

 sufficiently great to introduce a large error 

 in the chlorophyll a concentration value. Any 

 nonlinear absorption due to the sample ma- 

 terial itself has not, of course, been taken into 

 consideration. 



Additional evidence is now available which 

 shows that the sea salt retained on the filter 

 is the major cause of the high blank values 

 at 750 m^ (J. D. H. Strickland, personal com- 

 nnunication). 



Accuracy and Precision of Chlorophyll a 

 Determinations and Sampling Errors 



Richards with Thompson (1952) gave an 

 estimate of total error of 9 percent for 

 chlorophyll a. This estimate was obtained with 

 identical samples and presumably reflects the 

 reproducibility of the method under ideal 

 laboratory conditions when quantities of chlo- 

 rophyll are large. Humphrey (1960), who ana- 

 lyzed the chlorophyll a concentration in six 

 aliquots of a well-mixed surface sea water 

 sample, found the standard deviation (0.053 mg, 

 per cubic meter) to equal about 6.6 percent of 

 the mean value (mean 0.81 mg. per cubic meter) 

 of the six determinations. 



In 1956 and 1957 a number of simple field 

 studies were carried out which gave a rough 

 estimate of the variability in chlorophyll a in 

 repeated sampling. The results have been 



tabulated in table 12 together with the mean 

 value, standard deviation, and coefficient of 

 variation for each series at each sample 

 depth. Unfortunately, the interpretation of 

 these data is not as straightforward as X 

 might wish. The difficulty arises from the fact 

 that not all of the data in each series could be 

 collected sinnultaneously, with the result that 

 the effect of the diurnal rhythm of chlorophyll 

 a (Doty and Oguri, 1957; Holmes and Haxo in 

 Holmes et al., 1958; McAllister, 1963; Shima- 

 da, 1958; Yentsch and Ryther, 1957) almost 

 certainly introduces another source of vari- 

 ability distinct from that associated with spatial 

 heterogeneity. Furthermore, in the statistical 

 analysis I have considered the samples tohave 

 been collected at random and the values 

 obtained at a given depth to be normally dis- 

 tributed. 



Of the data presented in table 12, series B 

 and C best reflect the variability encountered 

 in synoptic sampling and yield coefficients of 

 variation of 14 and 27 percent, respectively. 

 The remaining surface sample series (A, D, 

 and E) yield values below 27 percent even 

 though the sampling series required several 

 hours for completion. A similar analysis of 

 Shimada's (1958) data, which covered a time 

 interval of 44 hours, gives a coefficient of 

 variation of 16.5 percent (mean 0.112; stand- 

 ard deviation 0.0184) in spite of the clear 

 evidence of diurnal periodicity. Thus at the 



Table 12. --Variability in field determinations of ohlorophyll a, eastern tropical Pacific, 1957 



, nine surface 



1958). 



surface samples 



; six samples 



^ Series A taken by SCOPE, 22 November 1956: 0915-1202, lat. 9°25' N. , long. 89°31' W. 

 samples collected at and around a surface parachute drogue at 4.8-laii. intervals (Holmes, 



Series B taken by Orea, 28 February 1957: 1145, lat. 33° 07.5' N. , long. 117° 50' W.; 

 collected in rapid succession with a plastic bucket at six locations around vessel. 



Series C taken by Orca , 26 February 1957: 0848, lat. 32° 39.3' N., long. 117° 12.7' W. 

 taken at 1-m. depth in rapid succession from the same position on the ship--inshore samples. 



Series D taken by Orea. 27 February 1957: 0254-0700, lat. 33° 6.5' N., long. 118° 10.5' W. to lat. 

 33° 05' N. , long. 118° 08.3' W. Samples at 0, 20, and 40 m. collected at five stations at the following 

 times: 0254, 0331, 0440, 0550, and 0700. Ship drifting during entire sampling period. 



Series E taken by Orea, 27 February 1957. Samples collected at four depths at about 6.4-km. intervals: 

 E-1 at 1422, lat. 33° 10.4' N. , long. 117° 28.4' W. ; E-2 at 1518, lat. 33° 10.2' N. , long. 117° 32.2' W. ; 

 E-3 at 1551, lat. 33° 09.8' N., long 117° 37' W. ; E-4 at 1900, lat. 33° 09.7' N., long. 117° 41.8' W. ; 

 E-5 at 1950, lat. 33° 09.3' N. , long 117° 46.5' W. 



All times are Pacific standard time. 



19 



