Jan., lUl/] A miotic Parthowgenesis in Taraxacum 99 



size, formed in tetrads, as by means of ordinary reduction 

 division. In other cases tetrads are found consisting of two 

 large and two small cells. Again, in both species but partic- 

 ularly in the examples of T. Icevigatum which were studied, the 

 pollen grains are formed in groups of irregular numbers, at times 

 as many as six or seven being found in a cluster. In such 

 cases, wide disparity of size is the rule, the larger grains fre- 

 quently being devoid of stainable nuclear material. 



Finally, in the red-fruited species, T. Icevigatum, bodies 

 exactly like mature pollen grains in outward form are found in 

 pairs. This type of pollen development corresponds exactly to 

 the diads found by Osawa^ to be so frequent in the amiotically 

 parthenogenetic Japanese species, T. albidum (?). The presence 

 of such diads would superficially suggest the complete loss of 

 reduction division, already true of the embryo-sac-mother-cell. 



Detailed studies of pollen genesis indicate that, as a matter 

 of fact, the tetrads of normal appearance are formed by ordinary 

 reduction division, just as they are in the sexual species of 

 dandelion that have been investigated. In many cases observed, 

 however, the divisions were seen to be quite irregular, in that 

 the spindle mechanism was unable to effect a simultaneous 

 transfer of all the chromosomes entering into the first division. 

 This means that a varying number of chromosomes lag behind 

 and are either (a) included in one of the two daughter nucleii, 

 augmenting its size at the expense of the other, or (b) formed 

 into separate supernumerary nucleii. In the former case of 

 course a second division would result in the formation of two 

 cells of large size, and two of reduced size, a situation which, 

 as indicated above, is often found. In the latter of the two 

 cases mentioned subsequent divisions might or might not occur, 

 in either case giving rise to an irregular number of pollen 

 grains of widely varying size and nuclear content. 



Again, a number of clearly defined cases were observed in 

 which nucleii were dividing amitotically, as indicated by 

 dumb-bell forms and other criteria laid down by careful workers 

 in the past. These amitotic divisions were found at times to 

 replace the first division of the mother cell, as well as subsequent 

 divisions. This finding also coincides with the condition obtain- 

 ing in T. albidum, as worked out by Osawa.^ 



Finally, clear explanation of the large number of anucleate 

 pollen grains, generally found in association with smaller grains, 

 was found in the frequent occurrence of well-defined nuclear 



