318 BACTERIAL POISONS. 



In closing the description of the preceding bodies it may 

 l>e well to present briefly our present knowledge as to thei'r 

 constitution. GATJTIER, starting out with the idea that 

 they are polymerization-products of hydrocyanic acid, has 

 deduced theoretically cyclic formulae, recalling the hexagon 

 of the benzole derivatives. These formulae, though formid- 

 able in appearance, are a complete failure so far as they are 

 expressions of chemical reactions. Thus, the formula of 

 guanine : 



N=CH 

 H CO 



HN NH 



fails to show either a urea or a guanidine residue, and yet 

 it is a well-known fact that guanine on oxidation yields 

 parabanic acid and guanidine (page 312). In a similar 

 manner, his xanthine formula fails to show up the urea 

 residues which we know to be present. 



HORBACZEWSKI'S synthesis of uric acid has thrown con- 

 siderable light upon the constitution of these bases. As a 

 consequence of his method of synthesis uric acid was shown 

 to possess the structural formula given below. E. FISCHER 

 has found, as a result of experimental work, the constitu- 

 tion of xanthine to be expressed by the subjoined formula. 

 We know that uric acid on treatment with nascent hydro- 

 gen is converted into xanthine, then into hypoxanthine. 

 It follows, therefore, that a relation exists between hypo- 

 xanthine and xanthine similar to that between xanthine 

 and uric acid. The formula of hypoxanthine, as deduced 

 from this relation, and given below, probably represents 

 its constitution quite closely. It is possible, however, that 

 the CH and CO groups will be found to occupy the 

 reverse position which they are given in this formula, in 

 which case corresponding changes must be made in the 

 formulae of guanine and adenine. The latter two arc b;isnl 

 upon the relation which. these bodies bear to xanthine and 



