| Dykes in The Genus Iris, 1913, under the name I. ensata; This widely distributed Asiatic species seems to stand so entirely alone as to demand separate treatment. Its peculiarities will be found in The Observations at the end of the account of the species. The most striking features are the long, narrow ovary and capsule, and the extremely short perianth tube.Description. Rootstock , a compact, close-tufted rhizome, covered with coarse remnants of old leaves, resembling those of I. spuria and I. longipetala. Leaves , linear, ¼-½ in. broad, length varying from a few inches at flowering time to 18-24 in. eventually, finely ribbed, of a glaucous grey. Stem , 2-12 in. long, flattened so as to be lozenge shape in section, bearing a single terminal head. Spathes , narrow, acuminate, green, usually 2-3 in. long, but sometimes much longer. Pedicels , of varying length even in the same spathe, ½-4 in. Ovary , 1-2 in. long, narrow, more or less trigonal, with two furrows on each face, usually twisted and passing gradually into the pedicel. Tube , very short and easily detached from the top of the)>Vary when the flower withers. Falls . The oblanceolate blade is separated by a slight and gradual constriction from the slightly panduriform haft. The colour varies considerably, from pure white with a few slight greenish veins to dark blue- or red-purple. In some cases the veining is very delicate and beautiful. On the haft the low central greenish-yellow ridge is often dotted with small purplish dots. Standards , oblanceolate unguiculate with canaliculate haft; the colour is usually darker than that of the falls and the back of the haft is greenish. Styles , narrow, rounded, not sharply keeled. Crests , triangular, overlapping. Stigma , a small, pendant triangular tongue. Filaments , white, short. Anthers , cream, longer than the filaments. Pollen , cream. Capsule , 2-3 in. long, narrowly oblong, with six ribs at equal intervals and a short sharp beak (see Fig. 11). Seeds , globose, dark brown, smooth-skmned; somewhat resemblmg those of the longipetala group.ObservationThis Iris extends from the Altai region to Shantung and Corea, and appears to have several local forms. Any attempt to arrive at a clear and complete classification of these forms is made the more difficult by several facts. In the first place, the slender rhizomes will not survive complete desiccation, and almost invariably succumb during the long journey to England from Central Asia. The surest method of introducing the various forms is therefore by means of seeds, which are freely produced and germinate readily. Unfortunately the plants do not begin to flower until, after three or four years, they have grown into clumps of some size.Another difficulty lies in the fact that the specimens found in herbarium collections have usually been gathered either at the flowering time or when the capsules of seed had developed. Between these two periods the foliage grows considerably, with the result that the plants appear quite different.The task of endeavouring to assign definite localities to the different forms of this Iris is probably further complicated by the fact that it has in Central Asia some importance as an economic plant. In Kashmir, for instance, the foliage is used as fodder (cf.' Gartenflora, 1898, p. 370), while in the neighbourhood of Kashgar it is woven into cord, and bunches of vegetables and of grapes are mostly tied with it. We are bound to admit therefore that the Iris may now have become naturalised in many places where it was not originally indigenous. In this connection it may be noted that Maximowicz, who endeavoured (vide I.e.) to distinguish three varieties, was obliged to state that all three were collected by Regel at Kuldscha.The economic value of I. ensata is doubtless increased by the fact that, owing to its deep rooting character, it is less affected by long periods of drought than any other Iris. This was very noticeable in September 1911, when the foliage of nearly all other Irises had withered away. The clumps of I. ensata remained quite green until they were cut down by frost. This peculiarity allows the plant to flourish where other vegetation fails, and I have seen photographs taken near Kashgar of a district where the ground is covered with great clumps of I. ensata many feet in circumference, and where the bare, deeply fissured hills afford abundant testimony to the aridity of the climate.Although it is impossible to come to any definite conclusion as to the number of really distinct forms of this species that may exist, there are at least three which should not be confused with one another, but which, at the same time, cannot be separated into distinct species. Of these three forms, by far the commonest is that which has flowers with narrow oblanceolate (alls and foliage which at length attains nearly, if not quite, to two feet in length. The flower stems are of variable length even on the same plant, and the relative proportions of the stem and of the leaves are not constant under different climatic conditions. Of this a striking illustration was afforded by the difference between Wilson's specimen from the Min valley in Szechuan (No. n81) and a plant raised from seed brought home from that locality by Wilson himself. In China the flowers appear when the leaves are very short', and this actually occurred in some plants of I. ensata, which came into flower soon after Christmas 1911, during a spell of mild weather following on the thorough ripening that the summer's drought had given the plants. In England, the plants usually do not flower until May or June, when the foliage has already grown to 12-18 inches, and when the stems are consequently much overtopped by the leaves. This variety of I. ensata, which I take to be the typical form, has grown in my garden in the form either of imported plants or of seedlings raised from seeds from the following localities :-Japan, Szechuan (Wilson no. 1181, 1908), Shantung, Kashmir.The second form, which I propose to call var. grandijlora, has much larger flowers, the falls being obovate rather than oblanceolate. The ground colour is a pale creamy primrose, delicately veined with pale violet. At the edge the veins become diffuse and spread over the whole surface. The standards are also broad and of the same pale violet colour. For some reason this Iris is often known in catalogues and in gardens as the variety oxypetala, to which name it has, of course, less claim than any other known form of I. ensata. The only specimens of this Iris that I have obtained from a known locality are a batch of plants raised from seeds sent from Gyantze in Thibet and handed on to me by the generosity of the late W. E. Gumbleton.The third form of this Iris is one which I have raised from seeds sent to me by Mons. H. Correvon, but without any guarantee of their exact provenance. All the plants that I raised agree in having short, stiff leaves, with a distinct corkscrew twist, which differentiates them at once from the leaves of the other forms, when all are grown side by side. The flowers of this third form closely resemble those of the typical plant, except that the veining extends further over the blade of the falls, which expand rather more suddenly and then narrows to a pointed tip. I am inclined to think that this may be the variety chinensis, which grows abundantly on the plains round Pekin. As herbarium specimens it is impossible to separate this variety from the typical plant, and it is only occasionally that the variety grandijlora displays its flowers sufficiently well in the dried state to allow of its being separated from the other varieties. Pallas' specimens (BM) appear to belong to this variety.Specimens of I. ensata are very common in herbarium collections, often erroneously identified. The stems are so variable in length, both absolutely and in proportion to the leaves, and the appearance of the plant at flowering time is so totally different to that which it has assumed by the time the capsules of seed are mature, that such confusions are almost unavoidable until we, are acquainted with the peculiarities of the living plant.In the dry state, the distinguishing features are the narrow finely-ribbed leaves, often flushed at the base with purple, which becomes a deep brown in the course of time, the sl~er rhizome closely covered with coarse, but not fibrous, remains of old leaves and not unlike the rhizome of I. spuria, the long, narrow, obscurely-ribbed and usually slightly twisted ovary, and the long, narrow, six-ribbed capsule, supported on a long pedicel (see Fig. 11, p. 87). These characters are common to all the known forms, and serve to distinguish them from any other species.In cultivation the plant presents no great difficulty. I. ensata will grow in any soil, either in sand or m heavy clay, and is moreover unlike some other apogons, in that it has no dislike to lime in the soil. Not only is this true in England, but I have found that the soil still remaining attached to the roots of plants that have reached me from Central Asia is often strongly impregnated with lime.When growth begins in spring, the young leaves appear of a pale yellow, as though they had been blanched and only gradually turn to the somewhat grey shade of green that they subsequently assume.After flowering, the leaves still grow considerably. In some cases they are stiff and upright and grow with a curious spiral twist ; in others the length is greater and the leaves droop until their tips almost touch the ground. They remain green until quite late in the autumn, in fact until they are cut down by sharp frosts, when they turn to a peculiarly dirty, brownish black, wholly -unlike the various shades of yellow or red-brown found among the withered leaves of other species.The only species to which I. ensata seems to show any affinity, are the Western American I. longipetala and its relatives I. missouriensis and I. montana. With these it agrees in the characters of the rhizome and, to some extent, of the foliage and capsule (the latter being, however, much narrower in I. ensata ). The seeds, also, are all of the same type, although those of I. ensata are so much smaller and more spherical as to be readily distinguishable from those of the American species.It is possible that the Chinese I. Grijsi is also allied to I. ensata, but our ignorance of the seeds and capsules of the former makes it impossible to ascertain whether there is any real relationship between the two species.