NEW CHROMOSOMES 71 



later. Chromosomes with two centromeres and with none (dicentric 

 and acentric) are found only immediately after they have arisen in 

 special X-ray experiments, for they are unable to survive mitosis 

 indefinitely. They must result from interchange that is asym- 

 metrical with respect to the centromere (Ch. XII). 



Thus the common observation of mitosis enables us to infer the 

 occurrence of several different kinds of changes, but it does not 

 alone enable us to define the nature of these changes. The observa- 

 tions of meiosis and of special types of mitosis is necessary for this. 

 They reveal many changes that are necessarily invisible at mitosis 

 such as equal interchanges, inversions not including the centromere 

 or symmetrical with regard to it, and very small translocations. 

 These show that all the conceivable structural transformations 

 of chromosomes occur in nature and are responsible for the changes 

 in number and size that are found in chromosome complements. 



The relative importance of the different kinds of visible change 

 occurring spontaneously has been shown in two species of Crepis 

 from analysis of 6,000 plants by Navashin (1926) as in Table 8. 



(ii) Time and Place of Structural Changes. From these observa- 

 tions several definite conclusions can be reached in regard to when, 

 where, and how structural changes arise, as follows : — 



1. Structural changes occur in chromosomes both in the mitotic 

 and in the meiotic phase. They probably occur during the resting 

 stage, or, at meiosis, during the prophase (v. Ch. VII, secondary 

 changes and crossing over). The occurrence in mitotic nuclei is 

 mostly clearly shown by Navashin's observations (1931 a) of 

 chimaeras in Crepis roots consisting of altered and unaltered cells 

 and by four pollen mother-cells in Secale having an interchange 

 while all their neighbours were normal, the interchange therefore 

 having taken place two mitoses earlier. The occurrence at meiosis 

 is most clearly shown by observations of new fragments both at the 

 metaphase of the first meiotic division and at the first post-meiotic 

 division in Tradescantia. 



2. New structural types may fail to survive. Obviously a new 

 acentric fragment will not survive mitosis unless it develops a new 

 one, which it apparently cannot. Fragments of various sizes are 

 found at the pollen grain mitosis in Tradescantia virginiana, 



