O'CONNELL: XARIABILITY OF NEAR-SURFACE PLANKTON 



Consecutive 1.6-km samples were separated by 

 shifting the incoming stream to a new filter every 

 6.5 min. Water volume, recorded for each sample 

 from a meter in the incoming line, averaged 636.7 

 liters (standard deviation 47.8) . Water temper- 

 ature of the incoming stream was recorded at 

 approximately the midpoint of each series of 

 block samples. Continuous thermograph records 

 indicated that surface water temperature did 

 not change appreciably during the sampling of 

 individual blocks. 



SAMPLE PROCESSING 



Four of the nine samples from each block were 

 randomly selected for the estimation of numbers 

 of organisms. Of the remaining five, one was 

 chosen at random to be reserved for special 

 purposes such as length and dry weight measure- 

 ments of species groups, and the other four were 

 pooled to obtain a dry weight value for the block. 

 Estimates were standardized as quantities per 

 m^ on the basis of the actual volumes of water 

 filtered in the samples. 



Estimates for two size categories of copepods, 

 0.2 to 0.9 mm in length and all over 1.0 mm long, 

 and for euphausiids and chaetognaths, were 

 made by volumetric subsampling with replace- 

 ment, i.e., each subsample was returned to the 

 sample before the next subsample was drawn. 

 The volumetric subsampling technique yields 

 estimates of satisfactory precision only if or- 

 ganisms are randomly distributed in the sample 

 container prior to the removal of each sub- 

 sample. Simple stirring accomplished this for 

 all species groups except the small copepods, 

 probably because they were entangled in ph>i;o- 

 plankton present in the samples. Random distri- 

 bution was assumed to exist where the value sVx 

 did not exceed x'/^-l for a series of subsamples 

 (Holmes and Widrig, 1956). A random distri- 

 bution was achieved for the small copepods by 

 subjecting the sample to a few 1-sec bursts of 

 rapid stirring in a Waring Blender." However, 

 because this treatment fragmented some larger 

 organisms a two-step jn-ocedure was emiiloyed: 



" The use of trade names is merely to facilitate de- 

 scriptions; no endorsement is implied. 



all organisms except the small copepods were 

 estimated by subsampling following gentle stir- 

 ring; the sample was then agitated in the War- 

 ing Blender, after which the small copepods were 

 estimated by subsampling. 



Estimates for the small copepods were always 

 based on subsamjjle counts totaling 200 to 300 

 from the sample. With the assumption of ran- 

 dom distribution, the number in the sample 

 should in all ca-ses be within 15'^ r of the estimate 

 for j) = 0.05 (Holmes and Widrig, 1956). More 

 than half of the sample estimates for the other 

 three species groups were based on counts of 

 30 or more, for which the number in the sample 

 should have been within 40'^r of the estimate. 

 For the remainder, where numbers counted 

 were low, examination was not extended beyond 

 subsamples totaling one-third the volume of the 

 sam|)le container, 2,000 ml. 



In addition to the four species groups counted, 

 the samples contained larvaceans and small in- 

 vertebrate eggs (0.15-0.35 mm diameter), some- 

 times in moderately high numbers. Larvacean 

 tails and heads were separated, however, and in- 

 vertebrate eggs were not readily distinguishable 

 from the latter. Cladocerans and polychaetes 

 were generally absent or low in number, though 

 each occurred in high numbers in a few samples. 

 Fish eggs occurred rarely and in low numbers. 



SIZE OF ORGANISMS 



Length measurements for 10 day samples 

 (0600-1800) and 10 night samples are summa- 

 rized in Table 1 and Figure 2. Measurements 

 were total length except for euphausiids, which 

 were measured from the carapace behind the 

 eye to the junction of the abdomen and telson. 

 Data from day and night samples were pooled 

 for small copepods and chaetognaths but not for 

 large copepods and eujihausiids, which showed 

 appreciable size frequency differences for the 

 two periods. 



The length-frequency distribution for the small 

 copepod group, composed largely of naupliar 

 and copepodite stages, is nearly symmetrical, 

 with the mean and the median close to the mid- 

 ])oint of the ]3redetermined size range (0.20-0.99 

 mm long). Almost one-third of the organisms 



683 



