368 ERWIN CHARGAFF 



with the first part of the expression representing the portion of apurinic 

 acid sensitive to alkaU and Pu, Thy, Cy denoting purine nucleotides and 

 thymidyhc and cytidyhc acids respectively. Whether the structure proposed 

 for calf thymus nucleic acid, namely, that of a chain in which tracts 

 consisting principally of pyrimidine nucleotides are followed by stretches 

 in which purine nucleotides predominate, will be borne out by further 

 work, remains to be seen. As regards the nondiffusible residue produced 

 by the action of alkali on apurinic acid, it may be of interest to note that 

 about one-half remains nondialyzable even against 2 M NaCl solution and 

 that an almost quantitative yield of the corresponding nucleosides is af- 

 forded by the action of phosphodiesterase and phosphomonoesterase.^^^ 



X. Correlations and Concluding Remarks 

 1. Simplifying Generalizations 



Even that master of inductive reasoning, Macaulay's infant who "is led 

 by induction to expect milk from his mother or nurse, and none from his 

 father," would have rejected as spurious some of the generalizations about 

 nucleic acids that have been made in the past and are still found in some 

 textbooks. The effort to force nature into a strait-jacket of puerile approxi- 

 mations has yielded many short-term successes in the natural sciences. As 

 is true of soap sculpture, they were pretty, but easily washed away. Whether 

 the attempt to teach the "gesta Dei per mathematicos" in three easy lessons 

 was more harmful than useful, may remain a matter of controversy. Nor 

 is plausibility a criterion of durability; quite the contrary: the gospels of 

 the future often are the heresies of the present. But the writers of textbooks 

 thrive on premature explanations. 



The "tetranucleotide theory" continues, for this reason, to lead a stub- 

 born existence. There never were any but psychological reasons for its 

 formulation, as has been pointed out before;^"'* but even in a very recent 

 and massive treatise there will be found the statement that thymus nucleic 

 acid is a large chain consisting of 500 to 1000 tetranucleotide units and that 

 each tetranucleotide is formed by the combination of four nucleotides con- 

 taining adenine, cytosine, guanine, and thymine, respectively."" Actually, 

 a glance at the tables in which the information on the composition of many 

 different deoxypentose nucleic acids is assembled (Tables V-XIII) will 

 show that out of almost 50 different species so far investigated only E. coli 

 (Nos. 5-9 in Table XII) has yielded preparations that could be called "sta- 

 tistical tetranucleotides," though not much good will come of it. One must 



289 M. E. Hodes and E. Chargaff, in preparation. 



"" B. Flaschentrager and E. Lehnartz (eds.), "Physiologische Chemie," Vol. I, p. 

 768. Springer, Berlin, Gottingen, Heidelberg, 1951. 



