MULLIN ET AL.: VERTICAL STRUCTURE OF PLANKTON OFF CALIFORNIA 



rence of the median value before vs. after the 

 storm. This tendency was also shown by isotherms 

 (see above). The range of concentrations of 

 chlorophyll in the water column tended to in- 

 crease, meaning that chlorophyll maxima were 

 accentuated after the storm, though hypothesis 4, 

 based on the median profiles, was not rejected 

 (0.01 < P < 0.025). 



No phytoplanktonic or protozoan taxa we 

 examined migrated dielly into and out of the upper 

 50 m nor did any taxon migrate dielly within the 

 upper 40 m. The two-way ANOVA of four profiles 

 detected significant decreases in poststorm abun- 

 dances of the diatoms Nitzschia spp., Bacterias- 

 trum spp., Rh. alata, and Rh. fragilissima; all but 

 the last of these decreases were also significant by 

 nonparametric Test 3 applied to the full 13- 

 member set of integrated profiles. This latter test 

 also revealed a significant decrease in poststorm 

 abundance of another diatom, S. costatum. Only 

 the dinoflagellate, Prorocentrum, was more abun- 

 dant after the storm by the ANOVA test. When the 

 data set of 13 integrated profiles was examined by 

 nonparametric Test 3, significant increases were 

 also detected in the poststorm abundance of 

 Lohmanniella (a potential larval fish food) and 

 Ceratium spp. Neither Gymnodinium splendens 

 nor Cochlodinium catenatum (two potential food 

 items for larval anchovy) changed significantly in 

 abundance in samples taken at the depth of the 

 fluorescence maximum layer from the second ves- 

 sel. The large diatom category, Chaetoceros spp., 

 did not change in total abundance, but the species 

 comprising this category changed at the time of 

 the storm; in particular, Ch. constrictus was the 

 dominant member of the genus after the storm, 

 but was not encountered in the prestorm samples. 



From the ANOVA, no phytoplanktonic taxa had 

 poststorm vertical distributions different from 

 their prestorm ones, when the criterion of P "^ 0.01 

 was used for significance, and only Rh. alata and 

 Prorocentrum had significant changes as defined 

 by P < 0.05. Hence, as far as we could tell from the 

 four profiles which were usable in the ANOVA, the 

 storm had much less effect in changing the verti- 

 cal distributions of specific phytoplankters (and 

 protozoans) than it did for zooplankton. This con- 

 clusion is, however, suspect (see below). 



Nonparametric Test 4, for which eight profiles 

 were usable, indicated that the poststorm range of 

 abundances in the upper 40 m was greater than 

 the prestorm range for five of the dinoflagellate 

 taxa, Mesodinium rubrum, and Lohmanniella, 

 while Nitzschia, Rh. fragilissima , and S. costatum 



had significantly smaller poststorm ranges. These 

 eight profiles strongly suggested poststorm shoal- 

 ing of the vertical distributions of the potential 

 food species, C. catenatum, G. splendens, and 

 Laboea, but the data sets were too small to estab- 

 lish statistical significance at P ^ 0.01. 



The general changes associated with the storm 

 were therefore decreases in the abundances and in 

 the degree of stratification of some diatoms, and 

 increases in abundances and degree of stratifica- 

 tion of some dinoflagellates and protozoans. How- 

 ever, significant changes in the pattern of stratifi- 

 cation with depth were more difficult to detect 

 because of the reduced data sets, except for the 

 shoaling of the distribution of chlorophyll. 



The floral composition of the profiles permitted a 

 clear separation into prestorm and poststorm as- 

 semblages, with the exception of the last prestorm 

 profile, which was quite different from the others 

 (Fig. 3B). This result was different from the analy- 

 sis of zooplankton (Fig. 3A), where the first post- 

 storm profile was unexpectedly grouped with pre- 

 storm profiles. Both results, however, indicate that 

 the compositional changes associated with the 

 storm were gradual rather than abrupt. Unlike 

 the faunal assemblages, the floral grouping 

 showed no tendency to separate day from night. 

 The difference in correlation coefficients between 

 dendrograms A and B probably reflects the fact 

 that quite different numbers of taxa were counted, 

 and that samples were counted by different 

 techniques, rather than any fundamental distinc- 

 tion between phytoplanktonic and zooplanktonic 

 assemblages. 



C Relations Between Zooplankton and 

 Phytoplanktonic Biomass 



If positive correlations between the abundances 

 of particle-grazing zooplanktonic taxa and 

 chlorophyll existed before the storm, it is reason- 

 able to hypothesize that such correlations would 

 be weaker or nonexistent after the storm due to 

 turbulent disruption of associations. 



We examined the following taxa of zooplankton 

 in this regard, sometimes combining categories 

 from the Appendix: Naupliar Acartia, naupliar 

 Calanus, naupliar "Paracalanus" , copepodid and 

 adult Acartia, CI-CIV Calanus, CV and female 

 Calanus (nocturnal only), copepodid and adult 

 "Paracalanus" , adult Metridia (nocturnal only), 

 adult Pleuromamma (nocturnal only), and the 

 appendicularians. We grouped data into four sets 

 of profiles: three diurnal, prestorm; three noc- 



159 



