72 MITOSIS : THE VARIATION OF THE CHROMOSOMES 



degenerating (Fig. 12). It seems probable that these are ones 

 lacking a centromere. Others behave normally as new chromo- 

 somes. These are all the same size, about one-tenth the length of 

 a normal chromosome in this and other genera (e.g., Fritillaria and 

 Ranunculus). 



3. Except for such observations of the first post-meiotic division, 

 it is only possible to see those new structural types which have 

 survived many mitoses. These show a great range of form, but 

 certain other restrictions are evident. Thus fusion only occurs 



Fig. 17. — Metaphase of mitosis in the root-tip of a seedling of Tulipa 

 galatica, 2« — 24 + igff. Note the lack of synchronisation in 

 the orientation and division of fragments (black). x 2,^00 

 (Upcott, unpub). 



between new fragments and old chromosomes (as in translocation) 

 or between whole chromosomes which have a centromere close to 

 the end. Fusion then occurs near the centromere (e.g., in McClung's 

 observations on the Acrididae). Here again the fusions which occur 

 may be at random and those that are seen may be the selected 

 survivors. Such cases, as we have seen, are most easily explained 

 as the result of unequal interchange. 



4. Once the new structural type has survived the test of mitosis 

 it is potentially as permanent as the chromosomes of the normal 

 complement. But it often happens that in its size it is less well 

 adapted than these to the conditions of mitosis or meiosis. The 



