10 Nature of the Genetic Material 



an ideal gene molecule, except when we assume that the chromosome 

 is a single immensely long and thoroughly coiled molecule. 



There is no doubt that the centromere with the spindle-fiber 

 ) attachment plays a decisive role in the division of the chromosome, 

 for fragments without a centromere are doomed. This may be a 

 mechanical feature, meaning that the centromere alone is capable of 

 sprouting a spindle fiber or of making the connection with one, what- 

 ever theory of spindle formation turns out to be true, under the 

 assumption that only the centromere-spindle mechanism can separate 

 the two split halves. There is some reason to assume that centromeres 

 and centrioles are identical elements, since Pollister's ( 1943 ) work has 



(shown that in the atypical sperm of Paludina the centromeres of dis- 

 carded chromosomes behave subsequently like centrioles. If this is 

 true, the centromere is completely different from the other constituents 

 of the chromosome, though it remains a body endowed primarily with 

 the ability of self-duplication just like the centriole. This would mean 

 also that its location within the series of chromomeres could hardly be 

 used as evidence that it is a special kind of chromomere (e.g., a piece 

 of heterochromatin ) . It is more comparable to the cytoplasmic cen- 

 trioles and 'Tcineties." The position of the centromere as an apparent 

 member of the chromomeric series would be indicative of a mechanical 

 cause: the function of the centromere in division requires its being 

 anchored in the chromonema, in analogy but not in homology with 

 the chromomeres. 



The centromere, Hke the centriole, seems to have two main 

 properties : the ability to divide; and, after Pollister's ( 1943 ) work and 

 Carothers' (1936) observations on grasshopper spermatocytes, the 

 ability to sprout an axial fiber. Whether the latter is due to molecular 

 unfolding, as some work on flagella indicates, or to polymerization is 

 not known. Thus the centromere becomes a non-genic, non-chromatic 

 differentiation of the chromonema, which nevertheless has the property 

 of self-duplication. There is a strong suspicion that two centromeres 

 may unite, two sister centromeres in preleptotene and two homologous 

 centromeres in diakinesis. Though an origin of a centromere de novo 

 has never been proved, and though the acquisition of a centromere by 

 a separating chromosome arm may be due to translocation or ab- 

 normal duplication of the centromeric region, an origin de novo 

 cannot be called impossible; in parthenogenetic sea urchin eggs cen- 

 trioles may be formed de novo (if Wilson's old observations still 

 stand), which, in view of the interchangeability of the two organelles, 

 may likewise be a capacity of the centromere. There remains the 



