MULLIN ET AL.: VERTICAL STRUCTURE OF PLANKTON OFF CALIFORNIA 



at 250 X, usually in one or two rows across the 

 settling chamber (1.13 or 2.3 ml). A precipitate 

 developed in certain samples after several months 

 storage, so profiles 1, 2, 8, and 12 could not be 

 included in the analysis based on discrete depths; 

 profile 13 was also excluded in order to balance the 

 data. 



To measure the amount of plant pigments in the 

 guts of selected zooplanktonic taxa, we used an 

 approach similar to that of Mackas and Bohrer 

 (1976). In a darkened room, the frozen contents of 

 each fiberglass filter were washed onto a circle of 

 Nitex mesh (180 /xm) and then sucked dry. The 

 mesh disk was transferred to a Petri dish, wetted, 

 and then examined visually using low magnifica- 

 tion and low-intensity green light. Organisms 

 were removed singly from each mesh, dipped in 

 filtered seawater, and then sorted into scintilla- 

 tion vials sitting in an ice bath and containing 

 small amounts of 90% reagent grade acetone. 



After obtaining enough organisms, we in- 

 spected the contents of each vial visually to insure 

 that they were taxonomically homogeneous and to 

 record the number of individuals present. The con- 

 tents were then homogenized with a motor-driven 

 teflon pestle in a glass grinding vessel to which 

 MgCOg and acetone were added. The homogenate 

 was transferred by several rinses to a 15 ml 

 screw-cap test tube and the volume was adjusted 

 to 10 ml. All test tubes were stored in a light-tight 

 container in a refrigerator for —1 h, after which 

 the homogenates were given an additional half 

 hour to extract and to warm to room temperature. 



The homogenate from each tube was first mixed 

 and then filtered through a fiberglass filter to re- 

 move the MgCOs and animal tissue/exoskeleton. 

 The amounts of chlorophyll a and phaeopigments 

 in the filtrate were determined fluorometrically 

 (Holm-Hansen et al. 1965) using a Turner Model 

 111 fluorometer equipped with a high-sensitivity 

 door. 



In order to evaluate the method, we collected 

 copepods by oblique net hauls over the Scripps 

 Canyon (—2 km from shore), sorted them, and 

 placed them in filtered seawater to starve for 18-24 

 h. On other occasions copepods were similarly col- 

 lected, starved to void their guts, and then allowed 

 to become satiated on mixtures of cultured phyto- 

 plankton. All animals were frozen before pigment 

 extraction. 



To assess (ex post facto) whether preservation of 

 pigments by freezing was complete, we took 

 oblique net tows (total duration ~2 h) over Scripps 

 Canyon. Each net haul was immediately strained 



through pieces of Nitex (<100 ixm) netting and 

 then quick-frozen using dry ice. Twelve samples 

 thus obtained were stored in the same freezer as 

 the cruise samples and processed in a similar 

 manner. One sample (To) was processed the same 

 day, the other samples at various times thereafter 

 up to 700 d. We were unable to detect a decrease in 

 total pigments over this time period by linear re- 

 gression, and therefore believe the freezing to be 

 adequate. 



The first group of hypotheses we wished to test 

 concerned temporal changes in patterns of verti- 

 cal distribution. One general procedure was to 

 treat several samples of one kind (e.g., all diurnal 

 samples from a particular depth before the storm) 

 as replicates accounting for variability due to 

 technique and to real patchiness, and then to look 

 for significant differences through an analysis of 

 variance (ANOVA) on log-transformed abun- 

 dances. Details are in Table 1. This was done for 

 those taxa for which the variances (of log- 

 transformed data) were homogeneous by 

 Bartlett's and/or Cochran's tests (Dixon and Mas- 

 sey 1957). Where the variances were heterogene- 

 ous (i.e., P < 0.01 of homogeneity), we tested 

 analogous hypotheses through nonparametric 

 tests, as indicated in Table 2. Taxa for which it was 

 necessary to employ the battery of nonparametric 

 tests are indicated by asterisks in the Appendix. 



A second group of hypotheses concerned correla- 

 tions between measured properties, such as the 

 concentration of chlorophyll and the abundance of 

 a particular taxon. These hypotheses were tested 

 by nonparametric correlation or concordance 

 tests; details are in Section C below We also tested 

 for changes in overall community composition by 

 constructing dendrograms based on rank differ- 

 ence correlation coefficients. All nonparametric 

 tests are from Tate and Clelland (1957). 



RESULTS 



The overall abundances and vertical distribu- 

 tions of 43 zooplanktonic and 18 phytoplanktonic 

 and protozoan taxa in the upper 50 m are shown in 

 the Appendix, based on median abundances for 

 diurnal and nocturnal profiles, before and after 

 the storm, together with the distributions of 

 chlorophyll. Depending on dietary preferences of 

 the visually feeding larval anchovy (e.g., Arthur 

 1976), some combination of the diurnal distribu- 

 tions of several taxa represents the "typical" verti- 

 cal distribution of larval fish food (see Section D 

 below). We will discuss results in the following 



153 



