Dixon — On the Chromosomes of Lilium longijlorwn. 709 



It also results fi'om these observations that the number of 

 "Chromosomes is not constant for the whole genus Liliujn ; for while 

 twenty-four is the most general number in L. Martagon {cf. Guignard, 

 loc. cit.) sixteen is more usual in Z. longiflorum, although twenty-four 

 also occurs. 



The meaning of the variation in number is difficult to explain. It 

 suggested itself that this difference among the nuclei might be due to 

 hybridisation, especially as Z. longiflorum is so highly cultivated a form, 

 but I could obtain no record of its being a hybrid, and when I compared 

 the nuclei of the variety "Harrisii" with the "type" I foujid that 

 the variation occurred in a similar manner in both. 



- Development of the Pollen. 



The nuclei of the young stamen may contain sixteen or twenty -four 

 ■chromosomes ; these two numbers are found even in the cells of the 

 archesporium up to the formation of the special mother-cells of the 

 pollen ; however, those with twenty-four are rarer than those with 

 sixteen. Preparatory to the first mitosis of the special mother-cells of 

 the pollen the nuclear thread breaks up into half the number of 

 ■chromosomes. In this stage — the prophase as well as in the diaster — ■ 

 I have foimd nuclei with eight chromosomes and others with twelve. 

 In these stages and in the subsequent divisions to fonn the nuclei of 

 the pollen grains I have counted twelve chromosomes in drawings made 

 from about forty nuclei as well as eight in similar drawings from about 

 forty more. Besides these nuclei there are also others present which 

 contain ten chromosomes, and I have seen one or two with thirteen 

 and fourteen, but these latter are very rare. Prom these numbers it 

 follows that each kind of asexual nucleus, i.e. those with twenty -four 

 and sixteen chromosomes and, perhaps, even those with intermediate 

 numbers have representatives in the mother-cells of the pollen, some 

 of the nuclei of the latter possessing twelve, some eight, and some ten. 

 Thus, not only is the number of chromosomes in the asexual nuclei 

 of this lily variable, but even the number found in the nuclei of the 

 pollen mother-cells also varies within the limits stated. 



With regard to the form of the nuclear figures of the divisions, 

 previous to the formation of the pollen grains, it may be noticed that 

 in the pollen mother-cells the nuclear thread, before being distiaguish- 

 able into individual chromosomes, is a complexly coiled delicate thread. 

 The granules of which it is composed, even with high powers, appear 

 to form a single row in the filament of linin ; at this time there are 



