68 'Wisconsin Academy of Sciences, Arts■, and Letters. 
spread by artificial infection. Later it was only with the great¬ 
est care that new leaves conld be infected. Infection was more 
often secured by clamping together with a light spring the un¬ 
dersides of two leaves, one of which contained ripe spores and 
then covering the plant with a bell jar. 
Our observations on the process of spore formation agree in 
general with those of the authors cited. Rosen (12) says that 
the wedge-shaped sterile cells in the aecidiospore rows are cut 
off alternately on the opposite sides. In our material they 
seem to be cut off quite irregularly. Cases were also observed 
in which apparently no sterigmata were formed. The divisions 
of the promyceliumi simply rounded themselves off and became 
spores as frequently happens in Gymnpsporangium. 
Our preparations of the nuclear phenomena in the develop¬ 
ment of the uredospore agree with those of Poirault and Faci- 
borski (3). It is perfectly certain that the nuclei lie at first 
in the long axis of the cell and that later they lie side by side 
in the transverse axis where they divide simultaneously. The 
cell division always results in the cutting off in the daughter 
cells of a pair of nuclei, each member of wflich is formed from 
one of the pair in the mother cell. From this it is seen that 
the pair of nuclei in the daughter cell are never sister nuclei. 
By reference to Fig. 21, it will be seen that the two masses 
which are to form the two nuclei in each new cell are still con¬ 
nected, not with each other, but with the corresponding mass in 
what is to be the other daughter cell. As to whether a. single 
spindle is formed for the two nuclei as claimed by Poirault and 
Faeiborski, it is very difficult to say. However, their figures 
do not seem to be conclusive on this point. Also as to the num¬ 
ber chromosomes the evidence given by Poirault and Pacibor¬ 
ski and Sappin-Tr ouffy seems hardly conclusive. Our figures 
show the chromatin as solid masses in all stages of division, but 
it is quite likely that these appearances are due to poor fixation 
and are not to be trusted as showing details either as to the 
number of chromosomes or the spindle. Indeed for rea¬ 
sons which will be given later it seems highly probable that 
there are more than two chromosomes in each division figure. 
It is perfectly clear that the nuclei lie side by side to divide 
