TERATOLOGY AND VARIATION 45 
the flower spike, and another in which the sterile portion of the 
spadix formed an irregular monstrous form. Plants have been 
observed by the writer in which the spathes were contorted and 
reduced in size as the result of evident injury; and the fact that 
the spathe begins its development before the differentiation of the 
spadix would make possible an injury which would entirely check 
the spathe’s development at a time when no harm would come 
to the spadix. As noted in connection with the experiments on 
change in sex, some plants with mutilated corms produced in- 
florescence only. One such flower cluster showed a normal spathe 
with normal staminate spike of flowers, but with a mere knob to 
represent the sterile spadix. Later two similar specimens were 
collected in the field. 
The only definite report of observations of abnormalities in 
the form of individual flowers is that of the confluence of the 
stigmas of two separate ovaries, reported by Rennert (21, p. 248). 
The occurrence of mixed spikes has been discussed in the section 
on sex distribution. It was there stated that the staminate flowers 
found on spikes chiefly pistillate were usually normal and function- 
ally perfect. In some cases, however, the stamens are borne on long 
pedicels (PLATE 4, FIG. 59) and show either small, sterile anthers 
or bract-like sterile growths. Many of the stamens formed entirely 
above a pistillate spike are undersized (PLATE I, FIG. 22), and some 
do not mature pollen. Close observation of a large number of 
flower spikes shows that there are three lines along which the 
flowers may vary from the normal form. The most common is 
that just mentioned and represented in PLATE 4, FIG. 59, 1. e., the 
more or less complete transformation of floral into vegetative 
structures. This is found not only with staminate flowers but is 
very frequent on otherwise purely pistillate spikes, where the 
transformed parts appear as more or less convolute bracts (TEXT 
FIGS. 54,55). A second line of abnormal development is that rep- 
resented by the confluence of parts. The confluence of the short 
filaments of -stamens and of anthers was mentioned in the section 
dealing with the staminate flower. The confluence of stigmas as 
reported by Rennert has been noted. An extreme case of the last 
named peculiarity is shown in TEXT FIG. 56, where four distinct 
= ovaries have a common stigmatic brush. Close examination 
