386 



n. M. Vernon 



Here we see that filtration of the water through a layer of sand 

 16 cm. iu depth in each case diminished the baeteria present, but 

 not veiy markedly. Of the three specimens of sand used, that dug 

 up a metre below the surface of the shore coiitained uo appreciable 

 vegetable growth, and so must bave acted merely as a mechanical 

 filter, The other two specimens ou the other band coutaiued small 

 quantities of vegetable matter. After two days' growth of the larva?, 

 the specimens of filtered water were found to be bacterially ver}^ 

 much purer than the normal water, though in every case the 

 number of baeteria had largely increased. After eight days' growth 

 the number had diminished again, but much more for the normal 

 water than for the samples of filtered water, so that these latter 

 now contained more baeteria than tlie former. 



After a current of water had been allowed to run continuously 

 through two of the specimens of sand for 29 days, fresh bacteriological 

 determinations were made. 



We see that now, after the sand has become impregnated with 

 vegetable growth, the filtration enormously increases the number of 

 baeteria in the water, instead of diminishing it. This increase was 



