The subdivisions of the genus and its geographical distribution 1 3 
These six segments surround the central style column, which likewise branches into three. Each 
style branch bears on its under-surface the stigmatic lip, usually near the upper extremity, and above 
that point splits into two stigmatic crests. The style branches are usually arched and concave on 
their under-side and so afford protection to the anthers which lie close beneath them. When the style 
branches curve closely down on to the haft of the falls as in the Oncocyclus species, a kind of tunnel 
is formed with the anther lying along the roof and the stigma projecting near the mouth. In theory 
the visiting insect is supposed to dust its back with pollen as it makes its way down the tunnel 
towards the nectar in the tube at the base, and then to pollinate the next flower it visits. In practice, 
however, we find that very few Pogoniris 1 become fertilised in this way while in large groups of 
Apogon Irises self-fertilisation is the rule, without the intervention of insects*. Bees fail to fertilise the 
larger Pogoniris, partly because the pollen clings so much more firmly to the anthers than does that 
of the Apogons, partly because bees visit flowers of different genera indiscriminately, partly because 
they often take the shortest path to the nectar by entering the tunnel near its base from the side and 
not by the main entrance, and partly also because the straight stigma of the Pogoniris is so much less 
prominent than the tongue-like point or points of the Apogons. 
The ovary of an Iris is divided into three longitudinal cells, corresponding to the three style 
branches, and the placenta forms a central column to which the seeds are attached. 
The ovary is usually supported on a pedicel which, however, is sometimes extremely short. To- 
gether usually with the tube it is generally encased in the spathes, the valves of which, however, are 
in some cases so divergent as to expose both tube and ovary. 
The subdivisions of the genus and its geographical distribution. 
Paradoxical as it may seem, it is hardly possible to take a comprehensive view of the whole Iris 
genus without first dividing it into groups or sections. 
It would doubtless be most convenient if one single character could be selected as the basis for 
this classification, such for instance as the length of the perianth tube or the shape of the seeds, or 
even that of the pollen, but unfortunately the first of these would separate many plants that are 
obviously closely allied, and the other two, though they are of great help in dividing the genus into 
groups, are of little assistance in separating the species within the groups. 
We are forced, therefore, to divide the genus into groups in accordance with the more obvious 
external characters and then to arrange the species within the groups on the basis of other and some- 
times less obvious characters in which they differ among themselves. 
The first and most obvious division of the genus is into bulbous and non-bulbous species, although 
the essential difference between a bulb and a rhizome seems merely to be that in the former the 
growth of one year is entirely absorbed and disappears in the process of preparation for the next 
year's growth, while in the latter the new growths are formed without the absorption of the material 
of the old. 
Whether a bulb developed originally from a rhizome or a rhizome from a bulb, or both inde- 
pendently from a common ancestor are questions not easy to answer, and the answers to them are 
hardly relevant to our present purpose. It may, however, be noticed here that seedlings of I. Grant 
Dujffii (or I. Aschersoni), if lifted during the resting period at the end of their first or second year’s 
growth, appear to have formed bulbs with netted coats, which can only with difficulty be distinguished 
from those of I. reticulata. It is only in subsequent years that those remnants of former growth 
appear, which constitute the rhizome. See also p. 210. 
When this primary division into bulbous and non-bulbous species has been made, we cannot 
progress any further unless we take more than one character into consideration. For an obvious dis- 
tinction, such as the presence or absence of a beard, will not allow us to distinguish between several 
well defined subgenera, such as Pogoniris and Regelia. Moreover, it is somewhat difficult to define 
what we mean by a beard in an Iris, for many of the so-called beardless species have on the blade of 
the fall a distinct pubescence which under a microscope has all the appearance of a beard. However, 
the hair-like processes have a structural difference and are unicellular in the Apogons, whereas in the 
Pogoniris the hairs are multicellular. 
The first subdivision among the rhizomatous Irises will therefore contain those species in which 
the centre line of the falls remains smooth or is, at least, only covered with a minute pubescence of 
unicellular processes. It is to these Irises that the name Apogon' is given. They occur over the 
whole area in which Irises are found at all, that is to say in the Northern Temperate Zone from 
the Pacific Coast of North America in the west to China and Japan in the east, and from Alaska and 
Labrador, Siberia and Kamskatka in the north to Florida and Hongkong in the south. 
Besides their wide distribution and the fact that they far outnumber all the other subdivisions, 
there is another noticeable characteristic of the Apogon species, and that is that practically every one 
can be recognised by its capsule or seeds, so distinct are these in each case. There are within the 
subdivision certain well defined groups, such as those of which I. sibirica and I. spuria may be taken 
1 See p. 156. * See p. 19. 
* From nvytov a beard, and the privative prefix a. 
