CENTROMERES 537 



somewhat variable, the whole adjoinmg region of the chromosome 

 may share its properties, and in some cells in Tradescantia, following 

 acid vapour pre-treatment, it has appeared as a double particle 

 transversely orientated at first metaphase (D., unpub.). This 

 doubleness, however, being inconsistent, is probably a non-charac- 

 teristic artefact. 



Since the centromeres that have been seen are located at charac- 

 teristic and permanent positions in the chromosome we may expect 

 them to have the permanence of other chromosome parts. This has 

 been shown most simply by the results of crossing-over in inversions, 

 and of the breakages of chromosomes by X-rays already described 

 (Ch. X). They make it clear that each centromere continues to 

 function even when more than one lies in a particular chromatid, 

 and that when a chromatid arises with no centromere it ceases to 

 make any movement in response to the spindle. Centromeres do not 

 arise de novo (Navashin, 1932 ; Mather and Stone, 1933). Wherever 

 they are found they control the spindle movements of the chromo- 

 somes and the special prophase movement of terminalisation. We 

 may therefore infer that bodies analogous to the centromeres are 

 present wherever their functions are fulfilled and the centric con- 

 strictions that they determine are observed. The failure to see 

 them in the smallest chromosomes is to be expected if they are pro- 

 portionate in size to the bodies of the chromosomes, for in the largest 

 they are near to the limit of visibility. 



When we apply this conclusion generally we find that the chromo- 

 somes of some organisms have a less clearly defined relation to the 

 spindle, which may imply an anomalous character in their centro- 

 meres. In Ascaris, as we have seen, the apparently diffuse 

 character of the anaphase repulsion is due to the chromosome being 

 polycentric. The possibility of doubleness in the centromere sug- 

 gested by direct observation (D., 1933) and inferred from the results 

 of X-ray breakage (McClintock, 1933 h), seems to be incompatible 

 with its mechanical individuality. In other organisms the anomaly 

 is not clear (e.g., in the Coccidae, F. Schrader, 1932). 



The question of the genetic relationship of centromeres is impor- 

 tant in considering their mechanical functions. It must not be 

 supposed that the centromeres, since they belong to individual 



