222 



SCIENCE 



[N. S. Vol. XLIV. No. 1129 



exhibit symptoms, increasingly serious, of 

 linear differentiation. 6 



If we admit the permeability of the nu- 

 clear membrane for chromatin or its im- 

 mediate forerunners, we can with equal 

 justification attribute the exact character 

 of this membrane to the quantities and 

 qualities of the substances enclosed. Spe- 

 cific permeabilities at once suggest them- 

 selves and so, by selective exclusion, any 

 elements not true to one or the other of the 

 types already present within the nucleus 

 may, conceivably, be warded off (Macal- 

 lum). 



Having admitted only specific elements 

 to the nucleus, it becomes our duty to at- 

 tach them to particular places in specific 

 chromosomes. Here we are, necessarily, 

 thrown on our resources in analogies. 



Most suggestive is the behavior of opti- 

 cally active substances in various degrees 

 of dispersion. The common Japanese cam- 

 phor, dextro-rotatory in alcoholic solution, 

 is also dextral in gaseous as well as solid 

 form. A property therefore which in the 

 highest and intermediate states of dispersal 

 must be attributed to the configuration of 

 individual molecules, is preserved in ag- 

 gregates of these. This can only result 

 from specific orientation. 



Taken alone, this analogy is too simple. 

 It may enable us to form some notion of 

 the terms in which differentiation among 

 the chromosomes is conceivable; but each 

 chromosome, instead of being homogeneous, 

 is, if we can trust ourselves, a system of 

 heterogeneous complexes definitely ar- 

 ranged in space. 



Our starting point may again be a rela- 

 tively simple analog. The hexoses are also 

 systems of heterogeneous complexes defi- 



6 This evidence has been brought together con- 

 veniently by T. H. Morgan and others, in "The 

 Mechanism of Mendelian Inheritance. ' ' The Mae- 

 millan Co., 1915. 



nitely arranged in space. "While the actual 

 form of the hexose molecule is unknown, 

 the carbon atoms are distributed in a man- 

 ner conceivable as a linear series in which 

 aldehyde and ketone groups occupy the 

 only positions possible. 



Chromosomes, of course, are not large 

 molecules, but aggregates of complexes of 

 these. While the chemical forces deter- 

 mining the specific structure of the indi- 

 vidual molecules may be precisely analo- 

 gous to those which account for the nature 

 of the hexose molecule, aggregation into 

 linear series, in the case of the chromo- 

 somes, very likely involves elements not 

 strictly molecular. There is one sugges- 

 tion, however, that is bodily transferable 

 to the situation presented by the chromo- 

 some, namely: factors, in the Mendelian 

 sense, may occupy certain positions be- 

 cause these are the only loci possible. 



In this connection, the temporary unions 

 between enzymes and their specific sub- 

 strates are especially interesting because 

 they depend on the stereo-relations of large 

 complexes of molecules. Conditions, gen- 

 erically similar, may play a determining 

 role in the formation of more permanent 

 unions even though these are not chemical. 

 Stereometrically determinable fitness, de- 

 grees of fitness, or possibilities of fitness, 

 between various regions of persistent dif- 

 ferentiated chromosomes and the newly 

 synthetized elements by the lateral accre- 

 tion or incorporation of which, these re- 

 gions grow, enable us to visualize not only 

 the periodic restoration of chromosomes to 

 full size, but even the physical require- 

 ments for such phenomena as the single 

 and double cross-over. 



THE DIVERSITY OP DESCENDANTS 



We can hammer out, on the lines sug- 

 gested, a provisional interpretation of that 

 constancy in organisms which makes us 



