76 PRELIMINARY EXPERIMENTS ON 



size worked from the same motor. All the jars were filled with water 

 from the offing beyond the Breakwater, but in one (A) the water had 

 been standing in the Laboratory for some weeks and had previously 

 contained Echinoderm larva*, while in the other two the water was fresh 

 from the sea. One of the latter jars (C) was immersed in a tank of 

 circulating water downstairs in order to maintain a fairly constant 

 temperature ; the other (B), together with the jar of stale water, was 

 placed in front of a window facing west, and was fully exposed to the 

 changes of temperature in the air of the building. The same number 

 of larvae (twenty-five) was placed in each two-gallon jar (August 4th). 

 The larvic in the stale water were all dead or moribund in three days. 

 Half of the larvas in each of the remaining jars had died within a week, 

 and all were dead on the tenth day, except one which survived until the 

 fourteenth. The rapid mortality in experiment A was obviously attri- 

 butable to some impurity in the stale water ; that in B and C was 

 traceable partly to the excessive movements of the plunger, as described 

 above, partly to insufficient food in the first week, and partly to the 

 over-accumulation of dead plankton in the limited quantity of water. 



Tahlc showing, for each experiment, the initial numher of larvcc, and the 

 number surviviny on every fourth day. 



Aug. 23rd. 

 J K L 

 10 10 10 



4 7 



all moribund. 



• Deatlis caused by the water becoming impure and turbid. 



t Deaths caused by meciianical accidents. 



X Deaths caused chiefly by transference of larvre to a tank of circulating water in which, 

 by an accident, tiiey were exposed to the attacks of rapacious Copepods. 



§ Deaths probably caused by mechanical accidents due to a defect in the jar subsequently 

 discovered (August 26th), and rectified same day. 



I 



