1900 ] DEVELOPMENT OF THE POLLEN GRAIN 97 
very readily lost; for in such stages clove oil must be used very 
cautiously. In plants whose cytoplasmic structure is very 
loosely netted, these precautions are unnecessary. In general, I 
have secured better differentiation of the chromosomes, or rather 
more distinct outlines, by using strong gentian. On the other 
hand, spindle effects are usually superior from the use of dilute 
stains. Considerable experimentation is necessary in order to 
determine what is the best method of procedure for each plant 
studied; but for this purpose time may well be afforded. 
Iron hematoxylin gave some brilliant results where the 
Flemming combination was least effective, as in the second 
division of the pollen mother-cells of Symplocarpus, and espe- 
cially in stages of the maturing pollen grain and of the dividing 
microspore nucleus. It was found well to keep the sections in 
the iron alum for more than an hour, and then to stain in hema- 
toxylin from twenty-four to thirty-six hours, from which a better 
differentiation results. Otherwise, it is difficult to get such a 
distinct effect in the chromosomes when the cytoplasm is prop- 
erly decolorized. 
BOTANICAL DEPARTMENT, 
CORNELL UNIVERSITY. 
EXPLANATION OF PLATES I AND II. 
PLATE I. Symplocarpus fetidus, 
All figures drawn with the aid of the camera lucida, at 12 ins. projection, tube 
length 155™™; nos, I-12, 14-19, and 21-24 with comp. ocular 8 and hom. imm.; 
no. 13, ocular 3 and objective as before; no. 20, comp. ocular 18 and objective as 
efore. 
Fig. 1. Nucleus in synapsis; peculiar orientation of kinoplasm in the 
cytoplasm. 
FG. 2. Spirem stage in which the chromatin disks are prominent on the 
linin framework, 
Fig. 3. Early development of the chromosomes; chromatin appearing in 
tetrad-like groups; fusiform nucleolus. 
Fic. 4. A segmentation stage in which loops and X and Y forms are 
abundant. 
Fig. 5. Forms of nucleoli not infrequently observed. 
Fic. 6. Multipolar spindle with only a few prominent poles. 
