390 Scott Elliot —On the Fertilisation of South 
a form being early visited by insects might begin to form its 
seeds before this inner circle was complete. 
It would have in this way an advantage, and the inner 
perianth-segments would become permanently more and more 
checked in development till they disappeared altogether. 
The second great distinction is the enormous development 
of the style-branches, and here one may get a little nearer to 
the solution of the question. In almost all Iridaceae the 
stamens are fully formed before the style has finished grow- 
ing ; and this late growth of the style is one of the most 
striking features of the order. In Sparaxis , e. g., the young 
style lies feebly on the sides of the perianth, but as it becomes 
mature rises up, and stands eventually quite upright. Hence 
in flowers with a perfectly vertical axis we see how, if this 
excessive growth of the style late in life is any advantage, 
there would always be favourable variations arising. More- 
over the separation or splitting of the style-branches down to 
their very base which characterises some Moraeas, is not entirely 
inexplicable, for their weight must tend to produce this result. 
In the zj^gomorphic or Gladiolus-line of development there 
are also transitional forms. Geissorhiza secunda , Ker, tends 
in that direction and in a (possibly abnormal) form of Babiana 
(B. stricta var. v-illosa fide Prof. Macowan) I found the stamens 
quite straight and the style-branches appearing between the 
anthers just as in Romulea ! The greater development of the 
lower perianth-parts is of course a common feature of zygo- 
morphic flowers, but I thought I found that in those where 
this was most marked the upper perianth-segments were in 
the direction of the incident light and the lower transversely 
inclined to it. Of course the action of gravity would equally 
afford an explanation. At any rate in Babiana ringens , where 
the upper perianth-segments are much larger than the lower, the 
inclination of the flower is such that the lower are in the direc- 
tion of the incident light and the upper transverse to it. In 
Aristea spiralis there is the same zygomorphy of the perianth, 
but the parts are morphologically different. The twisting of 
the filaments in Gladiolus , by which the anthers are made to 
