28 THE GENERAL CHARACTERS OF THE PROTEINS 



colour is developed. A solution thus prepared can be kept for a long 

 time without changing. 



Determination of the "Gold Number". Small quantities of the 

 colloid solution under investigation are introduced into a series of 50 

 c.c. beakers. The colloid should be measured out from a 0*2 c.c. 

 pipette graduated in thousandths of a cubic centimetre. The quan- 

 tities generally used are 0005, croi, 0*02, 0*05, up to 0*5 c.c. Larger 

 quantities of solution are to be avoided. Five c.c. of the gold solution 

 are then introduced into each beaker, and the mixture is then rapidly 

 stirred. After three to five minutes 0*5 c.c. of sodium chloride solu- 

 tion (100 grams NaCl to 900 c.c. water) is introduced into each 

 beaker. By this method a higher limit can be observed in which 

 no change takes place, and a lower limit in which the red solution 

 is converted into violet. In this way the limits are determined. The 

 number of milligrams of colloid in each of these limits, multiplied 

 by 2, gives the interval which is designated the " gold number". This 

 factor is the one generally determined (see table above). 



SECTION XI. THE FRACTIONAL FILTRATION OF PROTEINS. 



In 1896 C. J. Martin described a gelatin filter, through which, 

 under pressure of 40-50 atmospheres, water and simpler substances, 

 such as sugar and salts, could be made to pass, whereas more complex 

 colloidal bodies, such as the proteins, were held back. The apparatus 

 employed consisted essentially of a Pasteur-Cham berland filter candle, 

 which acted as a support for a membrane of gelatin or silicic acid. 

 This was mounted in a gun-metal filter case, which was connected 

 with a steel gas cylinder containing air under the requisite pressure. 

 The liquid to be filtered was introduced into the filter case, which 

 was then connected with the compressed air cylinder. The water and 

 simple bodies commenced to filter through as soon as the requisite 

 plus-pressure was attained. This process of filtering served as a 

 means of concentrating protein solutions and at the same time freeing 

 them from contamination with simpler substances. 



It has since been employed in a limited number of cases for 

 obtaining some information as to the relative sizes of molecular 

 aggregates in protein solutions. Thus, for example, W. A. Osborne 

 has shown that the sodium salt of caseinogen will pass through a 

 Martin filter, whereas the salt of the dibasic calcium, with approxi- 

 mately double the molecular weight, is held back. Craw has also 

 employed this filter for the separation of toxins and antitoxins. 



The principle involved in the Martin filter has recently been 

 extended by Bechnold, who has devised a method of fractional 

 filtration. Instead of employing a filter candle for impregnation 

 with the filtering membrane he uses filter paper. This is soaked in 

 the impregnating membrane (either acetic acid solution of collodium 

 or an aqueous solution of gelatin) in vacuo in a specially devised 

 apparatus. In this way a relatively thin filter with a large surface 

 can be obtained. This is introduced into a suitably constructed 

 filter funnel, in which it is supported on a nickel gauze ; the funnel 

 is inserted in a metallic cylinder, which can be connected either with 

 compressed air or a force pump, so as to produce a plus-pressure on 

 one side of the filter ; a stirrer can also, if necessary, be inserted so 



