182 Profs. Percy Frankland and Marshall Ward. 



paper. This was done, as it was quite conceivable that the presence 

 of comparatively large suspended particles should exercise a marked 

 influence on the behaviour of the anthrax introduced, whilst this 

 filtration would not remove more than a portion of the bacteria 

 already present in the water in its natural state. It was, moreover, 

 especially desirable to ascertain whether the removal of these coarser 

 suspended particles would influence the result, as in the follow- 

 ing sub-series (3) all suspended particles, including the water 

 bacteria themselves, were removed prior to the introduction of an- 

 thrax. 



(3.) Experiments with the Thames water after removing all 

 suspended particles, including bacteria, by filtration through porous 

 porcelain (Chamberland filter), or, in other words, Thames water 

 sterilised without the agency of heat. 



(4.) Experiments with the Thames water after filtration through 

 Swedish paper as in sub- series (2), and subsequent sterilisation with 

 steam. 



Chemical Composition of the Thames Water employed. — The water 

 was submitted to analysis, (a) in its natural condition, (b) after filtra- 

 tion through Swedish paper, and (c) after filtration through porous por- 

 celain, with the following results : — 



Results of Analysis expressed in Tarts 'per 100,000. 







(5). 



(c.) 



Total solid residue (dried at 100° C.) 



35-20 



33-60 



33-60 





0-207 



0-212 



0-189 





0-023 



0-039 



0-021 





0-004 



0-003 



0-007 





0-016 



o-oio 



0-014 



Oxygen consumed by organic matter, 









as measured by the reduction of 









permanganate acting for three hours 











0-076 



0054 



0-064 





0-230 



0-272 



0-229 





0-256 



0313 



0-256 





1-6 



1-65 



1-7 





17:3 



171 



16-5 





5-1 



5-3 



5-9 



Total „ 



22-4 



22-4 



22-4 





very 



clear 



clear 





turbid 







These results show that the sample contained only a moderate 

 amount of organic matter, and was representative of the river when 

 in its purest condition. They also show that neither the filtration 



