The Relations between Marine Animai and Vegetable Life. 415 



tank, were found to be in a perfectly healthy condition six weeks 

 later, whilst such weeds when 'introduced into brightly illuminated 

 tanks with white marble floors began to decay within a week 

 or two. 



Probably a more potent factor than the macroscopie alga3, both 

 for good and fov e vii, is to be found in the Dìatomace« and the 

 microscopie algie. Thus we bave seen how rapidly sand impregnated 

 with this vegetable growth can remove the ammonia frora the water, 

 bat we also saw that if the circulation of the water were diminished 

 to at ali a slow rate, it immediately became contaminated with more 

 ammonia than it originally contained. This lowly organised vegetable 

 growth appears to be exceedingly sensitive to the least defect of the 

 water which brings it its food material, and hence it follows that 

 the vegetable slinie which becomes deposited in the tanks of aquaria 

 rapidly becomes a contaminating agency. A layer of slime so thin 

 that ali strafa of it get suffìcient circulation of water, is a liighly 

 eftìcient purifying agent, but once the lowest Stratum is deprived of 

 this water, it begins to decompose. In order to effectually make 

 use of the purifying action of diatoms and algie, it would seem to 

 be necessary to filter the water, at one part of its circuit, in a con- 

 tinuous stream through a hiyer of sand impregnated with this vege- 

 table matter. In the Naples Aquarium the water is only pumped 

 for every alternate two hours, and hence, if it were required to purify 

 this water after it had run out of the tanks, it would bave to be 

 coUected in a subsidiary reservoir, from which it could filter in a 

 continuous stream through a layer of sand. This layer need not be 

 of more than a few centimetres in depth, though it would probably 

 bave to be renewed every few weeks owing to the multiplication in 

 its substance of bacteVia which are carried away by the water. This 

 increase of bacteria in the water could of course be obviated by sub- 

 sequently filtering it through a deep sand filter, such as is used in 

 water works. As to the best rate of filtration, the data obtained on 

 the small scale seemed to show the faster the rate the better, and 

 that with a layer of sand about 16cm. in depth the greatest possible 

 rate, i. e., at 20 to 30 litres per superficial Square metre per minute, 

 is the best. Needless to say, a filter of this kind would bave to be 

 exposed to daylight. 



Should it, on the otlier band, be diftìcult to arrange for the 

 filtration of the water in a continuous stream, or for the sand itself 

 to be exposed to light, a bacterial sand filter could be employed 



