32 Mutations and Evolution. 



2-6 un its, but since the chromosomes of the germ cells are shorter 

 and stouter than in the somatic cells this difference cannot be 

 directly compared with the length of the odd chromosome. Hance 

 also concludes that there is no evidence of pairing of the somatic 

 chromosomes. In this we cannot agree, as we believe that in 

 other (Enothera forms the pairing, though weak, is sufficiently 

 indicated in certain cases so that the extra or unpaired chromosome 

 may sometimes be directly determined by this method (see Gates 

 and Thomas 1914, pi. 35, Figs. 5, 6, 7, 10, 11, 15). Again, we have 

 not found that accurate drawings of the chromosomes will show 

 them to be of uniform width with even margin throughout, as 

 Hance's figures would indicate, although the departure from 

 uniformity in width is not. as a rule, great. 



Coming now to the question of chromosome number in somatic 

 tissues, it would seem at first impossible to reconcile the 

 apparently contradictory results of Hance and Miss Lutz. A 

 critical study has been made of the process of chromosome 

 fragmentation in somatic tissues, although only incidental 

 references to this work have been published. 1 We believe Hance 

 has over estimated the significance of these somatic segmentations. 

 We observed numerous " clear areas " and constricted portions of 

 chromosomes, but the segments were almost invariably in 

 alignment, showing clearly which belonged to the same 

 chromosome. Moreover, delicate linin connections are usually 

 visible connecting the two segments. Miss Lutz (1916) has made 

 similar observations. Constrictions or incomplete segmentations 

 have also been studied in Vicia Faba by Fraser and Snell (1911) and 

 Sakamura (1915). In nearly all such cases it is clear which portions 

 make up a given chromosome, and so it is much more sound to 

 treat them as a whole chromosome than to attempt to rearrange 

 these " fragments " on a hypothetical basis from measurements. 

 Formerly we observed no segmentation of chromosomes in 

 prophase nuclei, but more recently the writer has observed 

 several cases of fragmentation in prophase. Nevertheless, we are 

 inclined to look upon the phenomenon as a temporary one, related 

 only to the mitotic cycle, and the segmentations in the prophase of 

 one mitosis as probably bearing no relation to those of the next to 

 follow. The name " extra " or " supernumerary " chromosomes 

 applied by Hance to these segments is not a suitable one, for the 

 actual number of whole chromosomes can be determined without 

 difficulty by observation, as Hance himself has done in (E.scintillans 

 1 Gates and Thomas 1914. 



