HIROTA: NATURAL HISTORY OF PLEUROBRACHIA BACHEI IN LA JOLLA BIGHT 



larvae were in very low abundance. The abun- 

 dance of eggs was generally the same as that for 

 larvae, except for the absence of eggs in April 1970 

 and the high abundances of eggs relative to larvae 

 in August-September. There were about 

 1,000-10,000 eggs/m2 in August-September, 

 which are presumed to be spawned by the high 

 abundance of postlarvae. The hatching time of 

 eggs is about 24 h at 15°C, so that the low densities 

 of larvae during this time were the result of large 

 seasonal changes in hatching success, increased 

 mortality rates of larvae, or both, assuming that 

 the observed abundances were not determined 

 mainly by physical processes. The data on sea- 

 sonal variation in length frequency distributions 

 of postlarvae considered below will provide some 

 information to support the interpretation of in- 

 creased larval mortality. From the abundance of 

 eggs, larvae, and postlarvae at station 5 it was 

 calculated that on 13 March, 1 May, 31 July, 21 

 August, 5 November, and 27 January the eggs and 

 larvae made up 89-99% of the numbers of indi- 

 viduals per square meter. On 13 August the eggs 

 and larvae constituted 69% of the total population. 

 The sample dates in which the eggs and larvae 

 made up a very low percentage of the population 

 are those in June 1970 and April- June 1971. 



During the seasonal study 18 pairs of replicate 

 samples were taken between 31 July 1970 and 2 

 June 1971 at both stations 5 and 6, 1.6 and 10 km 

 off Scripps Institution, respectively. The mean 

 abundance of postlarvae per cubic meter was cal- 

 culated at each respective station on each sam- 

 pling date, and a f/-test was performed on these 

 means to determine whether or not medians of 

 mean abundance over time were significantly dif- 

 ferent at stations 5 and 6. The null hypothesis is no 

 significant difference, with a one-tailed alternate 

 hypothesis that the median of station 5 is greater 

 than that of station 6. Results indicate sig- 

 nificantly greater median abundance at station 

 5 than at 6 (P<0.025). The median difference is a 

 factor of 4.2 and the mean difference is a factor of 

 4.8. This result supports the three offshore dis- 

 tribution studies which indicated decreasing 

 abundance with increasing distance from shore. 

 However, the observed decrease in abundance in 

 the first 10 km from shore was about tenfold for 

 the offshore distribution studies and about half 

 this for the seasonal study. The discrepancy of a 

 factor of two is probably real and may be caused by 

 sampling bias in relation to stage of the tide and to 

 seasonal changes in the patterns of currents. The 



TIDAL HEIGHT, cm 



Figure 15. — The relationship between abundance of postlarval 

 Pleurobrachia bachei and the tidal height. The abundances are 

 in number per square meter of sea surface (Y), and the heights 

 are in centimeters (X) for samples at all stations located 1.6 km 

 from shore. The data are for all stages of the tide. 



more accurate measure of variations in abun- 

 dance with distance offshore should be found in 

 the seasonal comparisons, but more carefully 

 planned sampling could now be carried out to bet- 

 ter sort out variations due to small scale tidal 

 motions, larger scale "true" spatial variations 

 offshore, and the effect of other types of motion and 

 the wind on patterns of abundance. 



When all of the 180 samples for the seasonal 

 study (15 samples on five sampling dates are ex- 

 cluded from the analysis as five or fewer 

 ctenophores occurred in all samples lumped per 

 date) are plotted against predicted tidal height in 

 centimeters for all tidal stages (Figure 15), the 

 resulting least squares regression is Y = 0.89X -I- 

 88.36. Y is the number of ctenophores per square 

 meter andX is the tidal height in centimeters. The 

 slope of the line is significantly different from zero 

 in a two-tailed ^-test (P<0.01). It is surprising to 

 find a significant positive regression coefficient. 

 The strength of the test is in the many degrees of 

 freedom and the removal of 15 samples which 

 might otherwise tend to pull the line down toward 

 a zero slope because of frequent absence data at 

 any tidal height. This result is unexpected be- 

 cause the tidal currents are probably not the same 

 at different locations along the coast. Variations 

 exist in depth, bottom topography, exposure to 

 wind, strike of the beach, etc. The pattern of circu- 

 lation will also be differentially affected by spatial 

 and temporal variations in the wind field. The 

 results suggest that over an annual average, 

 abundance at any one time and place of sampling 

 could be affected by as much as a factor of four due 

 to tidal variations alone at locations 1.6 km from 

 shore. This average range due to tidal effects is 

 about the same magnitude as the annual average 

 difference between mean abundances of stations 5 



311 



