EGG-SECRETIONS OF ARBACIA AND ASTERIAS. 379 



The observations were so made that intervals of one minute 

 elapsed between the readings. Thus if A, B, C, D represent 

 four culture dishes, B was read one minute after A, C, one minute 

 after B, D one minute after C, and the second reading of A was 

 taken four minutes after the first. Under such conditions one 

 could introduce quite a large error from the mere arrangement of 

 the dishes. This was guarded against both by reversing the 

 order of reading from time to time, as well as by changing the 

 order of the dishes in the several experiments. As a matter of 

 fact the significant differences in my tables are always greater 

 than the maximum error which might have resulted from the 

 order of observation. 



In the presentation of the results I have not compared the 

 intervals for each of the three stages considered, separately 

 (although the data for such comparison are given) , but the aver- 

 age of the intervals that elapsed between insemination and the 

 2, 4, and 8-cell stages respectively. This procedure further 

 eliminates errors attaching to any specific observation, besides 

 reducing the number of comparisons necessary. 



In the right hand division of the following table are contained 

 comparable observations made in sea-water whose C OH had 

 been raised by the addition of 1.75 c.c. N/io NaOH to 100 c.c. 

 sea water. These experiments are included because they serve 

 as checks on those in normal sea water. The theoretical con- 

 siderations which prompted them were based on the fact brought 

 forward by Loeb ('98) that the development of Arbacia is ac- 

 celerated in alkaline sea-water, and depressed in acid. This is 

 easily verified if normally fertilized eggs are allowed to develop 

 to the blastula stage and are then divided into three lots, one 

 for control, a second to which NaOH is added in the proportion 

 of 1.75 c.c. N/io per 100 c.c. of sea water, whereas the third is 

 acidulated with HC1 in the proportion of 1.75 c.c. N/io per 

 100 c.c. of sea water. In such an experiment gastrulae, with 

 only here and there a short armed pluteus, predominate in the 

 HC1 culture at a given hour; the control at the same time contains 

 a large number of plutei in various stages and some late gastrulae, 

 whereas the NaOH dish holds practically nothing but complete 

 plutei. Loeb, in the paper referred to attributed this result to 



