REE AE” E A Vat Pee tk ae Ragin ne AN UN Poem. POONA Sa SEE OTe EWS MR SI gt ST a re ee SS e oe a ale ee ee ee tS Gr on eee oe 
TERATOLOGY AND VARIATION 47 
shows structures of this type to be only superficially confluent. 
TEXT FIG. 57 shows a vertical section of the ovaries shown in 
TEXT FIG. 56. It seems that the crowding of the young ovaries 
prevents the usual closing of the stylar opening, and the sub- 
sequent development of hairs on the increased stigmatic margin, 
forms a seemingly continuous brush. A better example of true 
dedoublement is shown by the two-celled ovary figured in TEXT 
FIGS. 58, 59. 
A third line of divergence leads towards bisexual flowers. As 
might be expected, many such flowers are found on spikes bearing 
both staminate and pistillate flowers. The specimens here de- 
scribed, with many other similar forms, were collected in the field, 
where they had grown under usual conditions. TEXT FIGs. 60, 
61 show two views of a flower with a partially developed ovary 
wall, three normal ovules, and two functional anther cells. TEXT 
FIGS. 62, 63 show an ovary with normal stigma and ovules, and 
with two functional anther cells on a lobe of the ovary wall. TEXT 
FIG. 64 shows a flower with the ovary wall almost entirely sup- 
pressed, four functional anther cells and an unusually large ovule. 
TEXT FIG. 65 shows a flower with two functional anther cells, and 
instead of ovules, two sterile bracts similar to those in TEXT FIGs. 
54, 55. TEXT FIGS. 66, 67 show a flower with an unusual 
ovary wall suggestive of that in TEXT FIG. 62, and with functional 
ovules, but without any anther. In TExT FIGS. 68, 69 is shown a 
cleft ovary wall with normal stigma and two sterile anther cells, 
but with ovules entirely suppressed. It should be noted that in 
all bisexual flowers examined the anther formation has been 
connected with the ovary wall. In no case has there been found a 
suggested transition from ovule to anther, even when the ovule 
shows degeneracy, as in TEXT FIG. 65, or suppression as in TEXT 
FIGS. 68, 69. 
Quite distinct from the abnormalities just described are the 
following. A group of ten plants, probably arising from one corm 
and its offshoots, was found with leaves having four leaflets as 
shown in TEXT FIG. 53. The plants were of medium size with 
normal inflorescence and leaves normal as to size, surface, and 
texture. All the plants showed a severe attack of Uromyces Caladit 
and died before another growing season made possible farther 
