THE PHYSICAL PROPERTIES OF THE PROTEINS 29 



that the surface of the filtering liquid in contact with the membrane 

 can be continually changed. With this apparatus comparatively 

 small plus-pressures only are necessary (from \ to 5 atmospheres). 



By employing impregnating membranes of different concentra- 

 tions Bechhold has succeeded in devising a method of differential 

 filtration. Thus, for example, a filter paper impregnated with 3 per 

 cent, gelatin, when tested with a I per cent, haemoglobin, allowed 

 appreciable quantities of the solute to pass ; a similar filter impreg- 

 nated with 4 per cent, gelatin was not permeable to haemoglobin. 



The more concentrated the impregnating solution the higher the 

 plus-pressure necessary for filtration. 



The filters made from more concentrated solutions have neces- 

 sarily smaller pores ; such filters will allow the passage only of 

 substances in a state of comparatively simple aggregation ; they can 

 in this way be separated from the substances in which the state of 

 aggregation is more complex. An interesting example of the method 

 of employment of the filter is afforded by Bechhold in his description 

 of the attempt to separate the constituents of Witte's peptone. Pick 

 (see p. 14) has described the following fractions obtained with am- 

 monium sulphate : Protalbumose (24-42 per cent, saturation), deutero- 

 albumose A (54-62 per cent), deutero-albumose B (70-95 per cent), 

 deutero-albumose C (100 per cent. + acid). A clear 5 per cent, 

 solution of Witte's peptone solution, which gave a precipitate on 23 

 per cent, saturation, was submitted to filtration under I atmosphere 

 plus-pressure through a 3 per cent, gelatin filter ; a similar filter 

 impregnated with 2*5 per cent, gelatin was just not permeable to 

 haemoglobin. 1 The residue on the filter was twice diluted with water 

 and twice subjected to filtration with a similar filter. Precipitation 

 in the filtrate commenced only when the saturation with ammonium 

 sulphate reached 34 per cent., whereas the residue on the filter com- 

 menced to precipitate when the liquid was only 23 per cent saturated ; 

 hence a separation into two fractions, viz., one precipitable at 34 per 

 cent, saturation and one at 23 per cent. Other experiments of 

 similar nature gave analogous results. They are summarised below : 



Fraction and Precipitation Limits, 



(NH 4 ) 2 SO 4 . Filter Used. Result. 



Protalbumose (24-42 per cent.) 3 per cent. Residue commenced to precipitate at 



34 per cent, saturation. 

 Deutero-albumoses A + B Residue precipitated between 34 and 



(54-95 P er cent.) 4 ,, 95 per cent, saturation. 



Deutero-albumose C (100 per Filtrate precipitated between 95 and 



cent. + acid) 4 ,, 100 per cent. 



Fractional filtration gives results, therefore, which are analogous 

 to those obtained by fractional precipitation with salts and may serve 

 as a valuable additional adjunct to this process. Here, again, how- 

 ever, complete separations are not to be expected ; for neither the 

 size of the aggregates in the solution to be filtered, nor the size of the 

 filter pores, are of uniform size, as Bechhold has experimentally 

 demonstrated. The method is still new, and promises to be a 

 valuable addition to the technique for dealing with proteins. It 

 might prove of great utility in conjunction with a tedious fractional 

 precipitation method, such as that employed by Haslam. 



1 One per cent, haemoglobin was used as a test solution for filter membranes ; the 

 minimum concentration necessary to produce a filter not permeable to haemoglobin in 

 this solution was determined. To express this the formula 3 per cent. (H 2-5 per cent.) 

 is employed. 



