ON FILTRATION OF SERUM COMPLEMENT 91 



for rabbit's corpuscles passed through the filter. Vedder 1 , as 

 a result of filtering fresh serum through a porcelain filter, found 

 that some bacteriolytic complements were retained, e.g. those for 

 the Bacillus coli and Staphylococcus aureus ; whilst others, e.g. that 

 for the typhoid bacillus, passed through. Muir and Ferguson 2 

 found that when rabbit's corpuscles were lysed with the minimum 

 dose of hsemolytic serum, the surplus of the receptors of the red 

 corpuscles were retained by a porcelain filter, whilst in the case of 

 lysis by water, .a very small proportion of the receptors passed 

 through. They also mentioned in connexion with these experiments 

 that immune-body passes through a filter practically unchanged, 

 whilst complement is to a large extent removed from the serum. 



The following investigation was undertaken in order to 

 determine certain factors influencing the filtration of serum 

 complement. The experiments were performed with small 

 Berkefeld niters 5-5 cm. in length, and all the results stated 

 below refer to filters of this kind. Some observations were 

 made with Maassen filters, but these were found to be more 

 permeable to complement, although the fluid passed through 

 more slowly than in the case of the Berkefeld filters. Their 

 use was accordingly discontinued. As pointed out by Levy, 

 Berkefeld filters have often a distinctly alkaline reaction, 

 and, accordingly, in all cases before using a filter we have 

 passed water through it till no trace of alkalinity was given 

 by the filtrate. The filters were then dried at 57 C. for 

 twenty-four hours. Fresh serum of normal guinea-pigs was 

 used in all the experiments, and was diluted with an equal 

 volume of 0-85 per cent, solution of sodium chloride before 

 filtration. Comparative experiments were always performed 

 at the same time, and, as nearly as possible, the rate of 

 filtration was uniform ; as a rule about 5 c.c. passing through 

 in about two minutes after the filter was saturated with the 

 fluid. As we obtained at first very discordant results when 

 the same filter was used more than once, the explanation 



1 Vedder, Journ. Med. Research, 1903, vol. ix, p. 475. 



2 Muir and Ferguson, Journ. Path, and Bacter., 1906, xi, p. 84. 



