drop with vigorous shaking in order tliat any insulin thrown out will 

 at once pass into solution again, and this addition of the acid is con- 

 tinued until a slight amount of turbidity persists permanently. The 

 flask is then carefully set aside overnight. On the following morning 

 the bottom of the flask as well as its sides are coated with an abundance 

 of colorless, highly refractive crystals. The small amount of amor- 

 phous insulin that falls out in very minute granules or floccules— an 

 expression of the above-mentioned turbidity— is readily separated by 

 means of the pipette during the washing with cold water, as the com- 

 pact crystals obtained by this method sink quickly to the bottom of the 

 tube like so much sea sand. This process of crystallization may be re- 

 peated as often as is thought to be necessary. Only a few experiments 

 have as yet been made with the object of crystallizing the compound 

 from hot solvents, and to prepare crystalline derivatives. 



Professor Swartz of the Department of Geology of the Johns Hop- 

 kins University has kindly examined samples of the insulin crystals 

 and reports them as being very uniform in character suggesting a 

 high degree of purity, as doubly refracting and as belonging to the 

 rhombohedral division of the hexagonal system. [See illustration be- 

 tween pages 16 and 17.] The crystals which are allowed to form more 

 slowly out of acidulated disoditim hydrogen phosphate solution are 

 very much larger than those precipitated quickly by pyridine out of 

 an acetic acid solution, such crystals being readily perceptible to the 

 naked eye as individual crystals. Irrespective of the manner of their 

 preparation the crystals melt sharply at 233 °C. with slight browning 

 occurring at 215°C. The melting point has remained constant on re- 

 crystallization. 



Chemical Reactions.— A large fraction of the sulphur of the mole- 

 cule is in the labile state previously described as characteristic of all 

 insulin fractions of high rabbit unitage. Quantitative analyses of both 

 the labile and more firmly bound sulphur are in progress. The com- 

 pound gives a beautiful biuret reaction passing from a clear pink to a 

 fine purple on further addition of copper sulphate to the alkaline 

 solution. Millon's reaction is unequivocal and positive. The Pauly 

 and the ninhydrin reactions are also positive. This pure insulin is 

 very sensitive to alkali, as shown by the fact that boiling for 1 5 minutes 

 with iV/10 sodium carbonate deprives it entirely of its hypoglycaemic 

 activity. It will be seen that the chemical reactions here described are 

 of considerable significance as shedding light on the chemical struc- 

 ture of the compound. It is now being examined in respect to its other 



78 



