314 Arber. — The Leaf Structure of the Iridaceae , 
accepting Mr. Dykes’s view, since Irises of the Spuria group are found over 
the whole range of I. Xiphium. 
In Iris filifolia, Boiss., another of the Xiphium Section, the only leaf 
which I was able to cut had no cylindrical apical region, but it was probably 
exceptional in this respect. 1 Here, as in the other members of the Section 
which I have examined, the median bundle is less conspicuous than the 
main laterals (Fig. 33, p. 313). 
The rare species, I. Kolpakowskiana , Regel, 2 whose affinities are 
somewhat uncertain, is generally associated with the Xiphium Section. Its 
leaf is described as Crocus- like in external appearance, but the similarity 
is in reality quite superficial (contrast Figs. 34 A and B with Fig. 56 I, 
p. 324, representing the limb of Crocus Tomasinianus) . The leaf is grooved 
on the adaxial side to the very apex ; it may be regarded as consisting 
primarily of leaf-sheath, but as including towards the apex the region of 
transition to the petiolar limb which has itself been lost. It bears a certain 
resemblance to the dorsiventral region of the leaf of Iris filifolia. This 
resemblance is not, however, very close, and the median bundle is more 
developed than is usual in the Xiphium Section. In this point the leaf rather 
recalls the Irises of the Juno Section. As /. Kolpakowskiana is confined 
to Turkestan, it would seem more probable that its affinities would be with 
the Juno Irises, which range from the Mediterranean region to the North- 
west Indian frontier, than with the Xiphiums, which are restricted to Spain 
and North-west Africa; but it is evidently a somewhat isolated form. 
The Juno Irises differ from the general Xiphium type in being 
dorsiventral to the extreme apex ; that is to say, the leaf is, on my view, 
more reduced than that of the Xiphium Section, and consists merely of 
leaf-sheath, having lost all trace of the petiole in which it once terminated. 
The leaves of the Juno Irises may be somewhat sharply folded along the 
midrib, which is, as a rule, at least as important as the main laterals. 
Iris orchioides , Carr. (Figs. 35 A and B) is a typical example. It shows 
very distinctly the difference between the lignified fibres (JP associated 
with the phloem, and the non-lignified hypodermal fibres (f 2 ) which lie 
between the (f) group and the epidermis. Chodat and Balicka-Iwanowska 3 
have shown that the presence or absence of these hypodermal fibres is 
a character of considerable systematic importance. 
In the closely related genus Moraea we find parallels both for the 
ensiform and bifacial leaves of Iris. The Section Eumoraea ( Dietes ) 
contains ensiform leaves, such as those of Moraea Macleai , Hort. (Fig. 12, 
p. 304), and M. Robinsoniana , C. Moore et F. Muell. (Fig. 11, p. 304), which 
recall those of the Irises of the Apogon Section, &c. But we also meet with 
a type of leaf which resembles that of the Xiphium Section in being 
1 Goebel, K. (1913), p. 285 . 2 Hooker, J. D. ( 1880 ). 
3 Chodat, R., and Balicka-Iwanowska, G. (1892). 
