352 Olive . — Sexual Cell Fusions and Vegetative 
the other hand, concludes that chromosome-formation in these cases ‘ is 
in complete abeyance ; so that the division actually partakes of the nature 
of direct division 5 (’ 04 , p. 35b). In all the various types of Rust cells 
herein described — in those of the spermogonium and in the uninucleated 
cells of the gametophyte, as well as in the binucleated cells of the sporo- 
phyte — the mitoses are found to conform quite closely to the type figured 
by Hermann for the Salamander. All three of the species which I have 
carefully examined for the mitotic phenomena agree in this respect ; and, 
although it is true that the nuclei of various species of Rusts vary greatly 
in size as well as in favourableness for study of the details of the process, 
it is highly probable that the phenomena as observed in Triphragmium will 
be found to be characteristic for the group. 
As was described above for Triphragmium , the central spindle arises 
between the halves of the divided centrosome. Further, since the centro- 
some is located on the nuclear membrane, in a somewhat similar position 
to that occupied by the centre in Phyllactinia (Harper, ’ 05 ), the central 
spindle comes to be formed in a corresponding position, along the nuclear 
membrane (Figs. 1-4). Soon the nuclear membrane breaks down and the 
central spindle thus comes to lie as a dense, strongly-developed structure 
at one side of the irregular mass of chromatin (Figs. 13-15, 24). It 
will be remembered that Blackman thinks that the spindles in the case 
of the promycelial divisions in Gymnospora 7 igium are formed outside of the 
nucleus, free in the cytoplasm, between the two portions of a divided 
extra-nuclear centrosome. His figure illustrating the phenomenon resembles 
closely the nuclei of my Fig. 24. From the facts gained from a study 
of the earliest prophases in the spermogonium of Triphragmium , I am 
inclined to think that Blackman has missed the earliest stages of spindle- 
formation. It is quite probable, in my opinion, that Gy 7 H 7 io sporangium 
will be found to agree with Triphragmium in having the centre located 
on the nuclear membrane, during the earlier stages, instead of outside 
the nucleus, in the cytoplasm. On the breaking down of the nuclear 
membrane, an appearance similar to Blackman’s figure would then be 
presented. It is highly probable, further, that the description of the 
nuclear and cell division in the spermogonium of Phragmidium violaceum 
given by Blackman will be found to be based on poor preparations, since 
both processes in the case of the spermogonium of Triphragmium are 
perfectly normal and well differentiated. The division of the spermogonial 
nucleus in Triphragmium corresponds in all the essential features to the 
mitoses in other types of cells of this species ; and further, the cell division 
by means of which the spermatium is cut off presents no such peculiarities 
as Blackman describes. The latter figures a sort of thickening ring on 
the hyphal wall below the constriction which he thinks may somehow 
be connected with the disjunction of the spermatium. This ring does not 
