| Dykes in The Genus Iris; 1913;THE PSEUDOREGELIA SECTIONIn substituting the name Pseudoregelia for Baker's name Pseudevansia, I wish to draw attention to the far greater affinity of the plants contained in the group to the Regelia Irises than to the Evansias. Of Baker's seven species only four were said to show traces of the rudimentary crest, from which the hairs of the beard spring, and which was the character on which the group was based. Of these I. Clarkei was clearly so described by a mistake (see p. 29); /. Alberti is so utterly unlike the other members of the group and so obviously a Pogoniris that it is difficult to see how it ever came to be included here and, if I. kumaonensis and /. Hookeriana have any crest at all, it is so extremely rudimentary and obscure that it can hardly be looked upon as of sufficient importance to make it the basis of a group.As a matter of fact, the examination of a number of flowers of the commonest Pogoniris will certainly provide examples of beards that "spring from a rudimentary crest'," or which end in a distinct crest. The amount of crest will be found to vary in the different flowers on the same plant and even in the various segments of the same flower. Foster's opinion of the taxonomic value of this character is worth quoting in full. "As I have elsewhere (Gard. Chron. 1887, i. p. 611) urged, I cannot attach any importance to the ridge on the fall. Indeed, in general I may say that the prominence, or want of prominence, of a median ridge on the fall, and whether it is a mere even ridge, or toothed, laciniated, or cut up into a beard, is about the least trustworthy character that can be appealed to in determining the affinities, and so the classification, of Irises•."What then are the characters in which /. goniocarpa, I. kumaonensis, I. Hookeriana and I. sikkimensis agree and which separate them from all other Irises? They are found in the rhizome, in the capsules, in the seeds and in the shape and colour of the flowers.First, they all possess a very compact rhizome, closely set with nipple-shaped growing points and presenting a gnarled appearance. In some ways there is a resemblance to the Oncocyclus and Regelia rhizomes, but the growth is much more compact and I have never yet seen any sign of a stolon.Secondly, the capsules taper to a pointed apex, which remains undivided even when the seeds are ripe and the capsule dehisces below the top. This also is a character shared by the Oncocyclus and Regelia groups, but in the case of the Pseudoregelia species the capsule is much broader relatively to the length and never attains the size of those of the other groups.Thirdly, the seeds show distinctly a creamy white aril, though here it is a mere flat ring and not a thick wax-like appendage often nearly as big as the seed proper.Lastly, the flowers all agree in having conspicuously oblong and consequently blunt-ended standards and are curiously mottled with two shades of purple.These characters, taken together, seem to define the group and certainly point to a far greater affinity to the Regelias than to the Evansias in which the capsules, seeds, rhizomes and flowers are entirely different.The group is confined to Northern India and Western China and its members seem to be the collateral relatives on the south side of the great Hindukush ranges to the Regelias on the north. The plants are somewhat difficult to cultivate because they come from regions where the year is divided into a dry and a rainy season. From March until October growth is rapid and moisture required and then the plants should lie dormant for six months in a comparatively dry soil. Since conditions in English gardens tend to be the very reverse of those that prevail in their native homes, the difficulties can be foreseen. Water must be supplied freely during the growing season and then during our winter it will be advisable to provide some sort of shelter in order to keep the plants comparatively dry. Failing this the root fibres are apt to rot and the plants thus weakened fail to flower properly in the following season.The rule that Iris seeds that possess a conspicuous aril are slow to germinate seems to hold good here as it certainly does with the Oncocyclus and Regelia species. All the Pseudoregelia seeds that I have obtained and sown, even when they were homegrown and sown at once, have gertninated very poorly, very irregularly and at long intervals. I even once went so far as to build a heap of snow three feet high over some of these seeds, but the reward was nil, although some of these identical seeds have now germinated after an interval of several years. I have not yet tried whether the solution of chloride of lime recommended for Oncocyclus seeds is equally effective with the Pseudoregelias, but it is certainly desirable to discover some means of hastening germination. Failing this, these Irises are never likely to become common in our gardens, since, owing to the character of the rhizomes, it is seldom that they lend themselves to multiplication by division.