1907] GATES—POLLEN DEVELOPMENT AND MUTATION 97 
they do not take the chromatin stain. In an iron-haematoxylin- 
orange stain they take only the orange; thus it is difficult to see how 
the view that the nucleoli of the daughter nuclei are reconstructed 
from fragments of the former nucleolus (SCHAFFNER 26) can apply 
here, for fragmentation of the nucleoli before their disappearance has 
never been observed, and their disappearance is preceded by loss of 
staining power, as if the chromatic-staining substance were first 
dissolved away, leaving the pale ground-substance. The method of 
final disappearance of the latter has not been observed. As already 
mentioned, the reverse series of events occurs in the reappearance of 
the nucleoli in the daughter nuclei. When first observed in the daugh- 
ter nuclei, they have their characteristic size and shape, but stain 
yellow and only acquire the chromatic stain later. They were never 
observed until there was a considerable accumulation of nuclear sap, 
and are certainly not visible until some time after the membranes*of 
the daughter nuclei have been formed. & These changes in the staining 
power of the nucleoli might be attributed to chemical changes in the 
nuclear sap, the medium in which they float, such as would alter the 
Staining reaction ‘of bodies contained in it. ¢ But while the nucleolus 
undergoes these changes before its disappearance, the chromosomes 
remain constant in their staining properties, which makes it appear 
more probable that the change in staining is due to actual loss of 
material from the nucleolus. 
Several of the figures (figs. 37, 38, 39) show that the number of . 
chromosomes in the O. Lamarckiana hybrid is greater than in O. lata. 
A count has not been made in the prophase, but from numerous 
counts in the metaphases and anaphases, the sporophyte number 
is found to be at least twenty and probably very near that number 
(fig. 37). It will be remembered that the sporophyte number in O. 
lata is fourteen. The bearing of these facts will be discussed later. 
The study of several other interesting features of the telophase 
of the heterotypic as well as of the homotypic division has not yet 
been completed. 
The pollen tetrads 
The pollen tetrads in O. Jata show numerous interesting features. 
As already stated, development to this stage frequently takes place, 
