The Relations between Marine Animai and Vegetable Life. 



387 



stili maintained after the larvse liad grown foiir days in the water, 

 but after eight days the differenee was much less marked. The 

 water fìltered through Aquarium sand contained, at any rate after 

 four days' growth, a very inueh larger mimber of germs than the 

 water filtered through deep shore sand, and this may have been the 

 cause of the larvse grown in it being only .1% larger than the normal, 

 as compared with the iucrease of 8.5^^ produced by the other water. 

 After the Aquarium tank sand filter had been kept in darkness 

 for 24 days, and was therefore no longer impreguated with Chloro- 

 phyll containing growth, another series of determinations was made. 



From this table it may be seen that in the two determinations 

 in which the water passed very slowly through the filter, the uumber 

 of bacteria in the outflowing water was considerably increased. On 

 the other band, when the rate of flow was moderately quick, the 

 number of germs was dimiuished to about half. This is rather an 

 unexpected result, for sand filters as a rule remove a greater number 

 of bacteria, the slower the current of water. In this case, however, 

 it is evident that the sand must have contained numbers of multiplying 

 bacteria, so that whilst on the one band many of the bacteria already 

 in the water would be removed by mechanical filtration, others in 

 larger or smaller numbers would be added to it from the sand itself. 

 It should be mentioned that when determining the number of bacteria 

 in the unfiltered water, the water was drawn off from the tap at 

 approximately the same rate as that at which it had been running 

 through the filter. 



The layers of sand iised in these experiments were there- 

 fore much too thin to act as efficient bacterial filters. Thus, as 

 previously mentioned, they were only 16cm. thick, instead of the 

 100 cm. or so usually employed for filtration at water works. 



As the following results show, the introduction of algse into the 

 water seemed to favour bacterial growth. 



