amount in a solution of other ampholytes whose solubility and pre- 

 cipitation coefficients lie very close to one another. It seemed worth 

 while to the writer therefore to employ only weak bases and acids, 

 more especially such as can act as buffers, in the hope that there might 

 result from their use such an alteration of solubilities, that impurities 

 would either fall out of or remain in solution while the hormone in 

 each case took the opposite course. The successful outcome of the 

 experiments proved the essential correctness of the idea. 



For a description of our earlier methods of purification the reader 

 is referred to the first paper of this series. The employment of the 

 simple method there described, led uniformly to the preparation of 

 an insulin fraction. Fraction IV, with a rabbit unitage of 40 or more 

 units to the milligram, a unitage that has lately been increased by a 

 more skilful use of these earlier methods. The next steps in the process 

 leading directly to the crystallization of insulin are as follows. 



One gram, approximately, of the so-called Fraction IV, is dissolved 

 in a little more than the required volume of N/6 acetic acid, enough 

 water is added to bring the volume up to 60 cc. or thereabout, and the 

 contaminating substances (together with some insulin) are then pre- 

 cipitated by the addition of an acidulated solution of brucine contain- 

 ing 6 gms. of brucine in 95 cc. of N/6 acetic acid. The resulting clear 

 supernatant fluid which contains nearly pure insulin is separated 

 from the precipitate by centrifugalization. Insulin remaining in the 

 precipitate may be removed by dissolving in N/6 acetic acid and 

 precipitating with the brucine solution as before. How profitable it 

 may be to repeat the process has not been determined at the moment. 

 The clear colorless centrifugalate containing the insulin is then pre- 

 cipitated ^vith iV/6 pyridine and the precipitate and fluid are immedi- 

 ately centrifuged. The precipitate which settles out is largely crystal- 

 line in character, the sides of the tube are found to be coated with 

 glistening highly refractive crystals and the topmost layer of the pre- 

 cipitate consists of similar crystals. It is not a difficult matter to remove 

 this topmost layer of crystals by means of a rubber-mounted pipette 

 and to free them from adherent pyridine and acetic acid by frequently 

 washing them with distilled water at room temperature in which 

 medium pure crystalline insulin is quite insoluble. The crystals may 

 be recrystallized by dissolving them in N/6 acetic acid and reprecipitat- 

 ing with N'/6 pyridine. It has been my practice, however, of late to 

 dissolve this crystalline insulin together with whatever fine granular 

 material there happens to be mixed with it in M/15 disodium hydro- 

 gen phosphate. To the clear solution xV/6 acetic acid is added drop by 



77 



