3()0 H- M. Vernon 



filtered through it was largely increased. Nevertheless the larvse were 

 increased A.l % in size. In Exp. 101 sand dug up a metre below 

 the surface of the sea shore was used. This sand contained no vege- 

 table growth, and must have acted on the water as a mechanical 

 filter pure and simple. The larvse grown in the water filtered through 

 it were 1.8^ larger than the normal, and henee it may reasonably 

 be concluded that when in the other experiments the larvse were 

 increased in size to a greater extent than this, the additional growth 

 was due to an actual chemical purification to the water, as distinguished 

 from a merely mechanical one. In Expts. 123 and 124 the water 

 was filtered through the apparatus, described above, which contained 

 sand taken from the Aquarium, whilst in Expts. 125 and 126, as also 

 in the above mentioned Exp. 101, it was filtered through a similar 

 apparatus containing sand from a metre below the surface. The 

 samples of water used were collected on the 29*'' day of the above 

 described experiments. In Exp. 123, in spite of the fact that 86^ 

 of the free ammonia was removed, the larvse were only very slightly 

 increased in size; not so much in fact as in Exp. 124, in which, 

 owing to the slowness of the filtration, the organic ammonia was 

 slightly increased in amount, and the free ammonia more than 

 quadrupled. These apparently contradictory results, as the evidence 

 adduced later on would suggest, may perhaps be ascribed to bacterial 

 action. In Expts. 125 and 126 on the other band, the larvse are 

 considerably increased in size. Thus in Exp. 1 25, in which the water 

 filtered through rather slowly, owing to the formation of a vegetable 

 crust on the top of the sand, the larvse were increased 8.5^ in size. 

 In Exp. 126 the water filtered through much more rapidly, as this 

 top layer of sand, to a depth of half a centimetre, was scraped 

 away. The apparent purifying efifect on the water was however even 

 greater than before, bat the larvse were increased by only 4.2^, or 

 less than half as much as before. It should be noted that in all 

 these four last experiments the number of larvse developing from the 

 ova was much greater than the normal, it being on an average %! % . 

 As a general re sul t therefore it may be concluded that by 

 filtration of the water through sand impregnated with vegetable 

 growth, it is not only rendered chemically much more pure, but, as 

 far as can be judged by the growth of these Plutei, physiologically 

 purer as well. At the same time it must be admitted that as far 

 as the results given in the above table can show, there is no general 

 and Constant relation between the degree of chemical purification 



