REPRODUCTION BY ATTRACTION 501 



The nature of this attraction is indicated, as we saw earlier 

 (Ch. VIII), by the differential condensation of the parts of chromo- 

 somes concerned. It seems to be an attraction between nucleolar 

 materials such as leads to the fusion of nucleoli at any stage in the 

 life of the nucleus. It has nothing to do with the specific attraction 

 of genes and chromomeres. 



(c) Reproduction. We have seen earlier from direct observation 

 that the chromosomes are undivided at the earliest prophase of 

 meiosis and divided throughout the prophase of mitosis. The 

 chromosomes must therefore divide during the resting stage in 

 mitosis, while in meiosis the prophase must be supposed to anticipate 

 this division. This conclusion is borne out by the evidence of 

 abnormal precocious prophases in polymitotic and variable sterile 

 maize (Beadle, 1932, 1933), where the chromosomes are still un- 

 divided as in meiosis. It has been confirmed by the results of X-ray 

 analysis in showing that the genes and the thread divide at a 

 particular period of the resting stage between meiosis and the 

 following mitosis (Ch. X). The evidence of X-ray effects on the 

 smallest visible objects must always have a higher validity than 

 that of visible or ultra-violet light on account of the shorter wave- 

 length, and it seems that these results must overbear any contra- 

 diction from direct observation. It is therefore unnecessary to 

 account for the many records of a spUt in the metaphase and 

 anaphase chromatids that have appeared in the past (Nebel, 1932 ; 

 Huskins and Smith, 1935, cf. Lorbeer, 1934, for list). Comparison, 

 in fact, shows that these records are internally as well as mutually 

 inconsistent. The divided " leptotene " chromosomes have not 

 been shown to pair. The supposed cleavages in the anaphase 

 chromatids have cut across either the major spiral or the minor 

 spiral, of whose existence the observers have been unaware. Those 

 observations which have taken account of the internal structure have 

 failed to reveal any split before the resting stage (Geitler, 1935 ; 

 D., 1935 a). There are, however, two exceptions, observations made 

 with a knowledge of structure, which suggest a split in the second 

 anaphase chromatids in Tradescantia (Kuwada and Nakamura, 

 1935 ; Sax, 1935). They depend, however, on the detection of 

 longitudinal doubleness in cylinders of less than one-half the wave- 



