I ,",,s Gates, Tctrapluid Mutants and Chromosome Mechanisms. 



Certain plants of the Fj were examined and found to have fourteen 

 chromosomes. Of course there is the bare possibility that the 

 wrong capsule was collected from the Lamarcldana plant in the 

 original cross. But this seems improbable. Another explanation 

 is that all the functional pollen grains of the giyas parent, owing 

 to irregularities in meiosis. contained only seven chromosomes, or 

 again as happens in certain Echinoderm hybrids, the extra chromo- 

 somes may have been extruded and lost from the nuclei in the 

 early mitoses of the fertilized egg. Geerts (1911) found that in 

 Lamarckiana ^ ( gigas in the F 2 the number of chromosomes returned 

 to fourteen, but contends that the F 2 hybrids were still identical 

 with those of the F x , an observation which is open to grave doubt. 

 In my paper on meiosis in O. luta X gigas (Gates, 1909b), 

 I demonstrated clearly that in my material the twenty-one chromo- 

 somes on the heterotypic spindle regularly segregated into groups 

 of ten and eleven chromosomes, with only occasional cases of a 

 912 distribution. Scores of nuclei were counted in interkinesis, 

 and in every case the result was as above stated. No cases were 

 observed of greater irregularity in the heterotypic distribution, and 

 none were found in \vhich chromosomes were left out in the cyto- 

 plasm during interkinesis. These results were established beyond 

 the slightest doubt in my paper above mentioned, yet Geerts 

 (1911) attempted to throw doubt upon them because, as he thought, 

 his own results were incompatible with them. Another point, which 

 was referred to by Strasburger (1910) and subsequently by Geerts 

 (1911) was with regard to the possible paired arrangement of the twenty- 

 one heterotypic chromosomes. Strasburger reproduced my figures 9 

 and 10 (plate XII) and 11 (plate XIII) as giving some evidence of 

 such a paired arrangement, which they probably do. But I never 

 found the pairing in the hybrid evident enough to be quite con- 

 vinced of its significance, although I studied this point before my 

 paper was published. As I first showed several years ago, the 

 chromosomes are very loosely arranged on the heterotypic spindle, 

 so that even in pure races of Lamarcldana forms and of O. bicnnis, 

 the evidence of pairing at this stage is often very doubtful. Davis 

 has since confirmed these results for (1910) O. biennis and (1911) 

 O. LfimarcJfiana. Even assuming, as is not improbably the case 

 (although I did not obtain thoroughly convincing evidence of it), 

 that in the twenty-one-chromosome hybrid seven chromosome pairs 

 are regularly formed, still the fact that the remaining seven chro- 

 mosomes were almost invariably distributed in groups of 4 and 8, 

 remains to be accounted for, as I pointed out (1 909 b, p. 194). The 

 fact remains that in my material the heterotypic mitosis was passed 

 through with great regularity and uniformity. The homotypic 

 mitosis also was completed with very few irregularities, and the 



