364 BOTANICAL GAZETTE [may 
logical study. In any event, it suggested a species for study, 
and the allied 7. purpurascens, abundant in the vacant lots i in 
Chicago, was selected. 2 
The work was begun early in the summer of 1900 and carried _ 
on at the Hull Botanical Laboratory of the University of Chicago. © 
Acknowledgments are due to Professor John M. Coulter for — 
much suggestive advice during the prosecution of the work, and — 
also to Dr. Charles J. Chamberlain and Dr. Burton E. Livingston 
for assistance in collection, technique, and interpretation. Mr. 
Andrew C. Moore, now of the University of South Carolina, also 
gave much assistance in collecting material. 
METHODS. 
- A compound microscope was taken into the field in order to 
determine whether the flowers were pistillate or staminate. Only 
such plants as were found by this means to be pistillate were 
oo used. In fact, the flowers were all too young to be determined 
in any other way. A dozen such pistillate plants were 10's" 
in the greenhouse of the laboratory, nine ~ of which surviv 
< The plants were numbered, and watered with solutions of various 
= fertilization would have taken place normally and u | 
ceased to be receptive. Each week some of ae 
ng flowers. were killed in a one per cent. solution of chrom 
acid and kept i in 70 per cent. alcohol for further study. - 
oe autumn these plants were dried and their rootstocks preserv 
oS pots over winter. These were forced about the first of April, 
_ all produced abundant pistillate flowers long before 
— doors had blossomed. This made it certain that pollina , 
Se not occur, as these flowers were mature and their re 
i. ceased to be receptive long before those out of ste”. ha 
oo begun to bloom, much less to produce pollen. * 7 
oe Fifteen or sixteen other pistillate plants were chosen He 
same time by the same means and transplanted into” the 
| of the laboratory. The inflorescences of these were 
covered or _ epee with paper bags so as to see < 
