188 BOTANICAL GAZETTE [SEPTEMBER 
be formed and the nuclear membrane may have disappeared on one 
side of the nucleus before any indication of spindle formation has 
appeared on the other side. 
Fig. 15 isa polar view of the two homotypic spindles in metaphase. 
There are 11 chromosomes on one spindle and 1o on the other. 
Fig. 25 is another at the same stage, showing 10 chromosomes. In 
fig. 24 the spindles are at right angles and all the chromosomes are 
not shown. In the side view of the spindle some of the chromosomes 
appear like tetrads, owing to their bivalent structure. On the homo- 
typic spindle, before the chromosomes divide, they are very regularly 
oriented in a single plane in the equatorial plate, as in jigs. 15; 25- 
This contrasts strikingly with the heterotypic spindle, in which the 
chromosomes are scattered for a considerable distance along the 
long axis of the spindle, so that there is usually no metaphase, strictly 
speaking. 
Fig. 26 is an early anaphase of the homotypic mitosis, showing 
one spindle in side view and one in polar view. In the polar view 
20 chromosomes are found by focusing through a short distance, and 
the remaining 2 are found on the next section. In the side view the 
chromosomes could not all be counted, but presumably there were 
20 after division. Two pale-staining nucleoli still persist in the 
cytoplasm. 
Fig. 16 shows three of the nuclei in the telophase of the homotyp!¢ 
mitosis. Many of the chromosomes have a characteristic two-lobed 
or dumb-bell shape. One chromosome is left behind in the cyto- 
plasm, leaving ro chromosomes each in two of the daughter nuclel. 
This two-lobed shape is a characteristic appearance of the chromo- 
somes in the telophase of somatic mitoses, but presumably bears no 
relation to the next mitosis, because if this were the beginning of a 
split for the next mitosis it would indicate that the division of the 
chromosomes is transverse. But metaphase and anaphase stag® of 
somatic mitoses show that the chromosomes divide longitudinally. 
Of course the possibility that the transverse axis of a chromosome 
in telophase should regularly become its longitudinal axis before the 
next metaphase, is not excluded, although this seems unlikely. 
The great difference in the size of the figures made it impossible 
to arrange them in order on the plates. The magnification is the same 
