The Relations between Marine Animai and Vegetable Life. 



357 



3' 50" to a litre in IT 20", and stili less wlien cleeveased to a litre 

 in 28'. It sbould be mentìoned that, before collecting tbe samples 

 used for these analyses, tbe water was allowed to filter at tbe same 

 slow rate for several bours previously. With reference to fresb water 

 sand filtration , Koch bas sbown ' tbat to obtaiu water containing- 

 less tbau 100 germs })er ce., tbe rate of filtration must not be more 

 tban 100 mm. per bour. Calculated for tbe filter at present under 

 discussion, tbis would correspond to a rate of 1 litre in 34'. Tbe 

 New River Company, on tbe otber band, filter tbeir water 2 at a rate 

 corresponding to 1 litre in 22'. Tbese rates of filtration were bowever 

 determined witb reference to tbe removal of as many germs from the 

 water as possible, and not to tbe niost efficient removal of tbe 

 ammonia. 



Simultaneously witb tbe above series of observations, anotber 

 was made ou tbe filtering capacities of a specimen of sand whicb 

 bad been dug u}) from tbe sea sbore, at a deptb of a metre beneatb 

 tbe surface. It could obviously tberefore be uncontaminated witb 

 any vegetable growtb of a cbloropbyll containing nature. Tbe results 

 obtained with tbis sand are given in tbe accompanying table. As 

 before, tbe layer of sand was 16 cm. tbick, and of 177sq.cm. super- 

 ficial area. 



For tbe first few days we see tbat tbis sand bad practically no 

 ammonia-absorbing powers at ali, tbe sligbtly diminisbed amounts 

 present in tbe filtered water probably being due to tbe mecbauical 



1 ibid. 



^ NOTTER and Firth, Theory and Practice of Hygiene 1896 p. 4' 



Mittheilungen a. d. Zoolog. Station zn Neapel. Bd. 13. 24 



