302 Arber . — The Leaf Structure of the Iridaceae , 
expression. The thesis which I hope here to substantiate is that the leaf 
of the Iridaceae has no true lamina, but represents , in every case , either 
a petiole and leaf -base, or a leaf -base alone. 
A considerable body of work on the leaf anatomy of the Irids already 
exists , 1 that of the Italian writer, Hermann Ross, being particularly thorough 
and detailed. But I have found that, since none of the botanists who have 
hitherto dealt with the subject have even considered the possibility of the 
leaf in question being merely a modified leaf-base and petiole, their 
observations seldom include answers to the particular questions to which 
the phyllode theory gives rise. I have, therefore, found it necessary to 
attempt an independent examination of the principal leaf-types met with 
in the Iridaceae. In this connexion I must express my gratitude, for 
their kindness in supplying me with material, to the Director and to the 
Keeper of the Herbarium, the Royal Botanic Gardens, Kew, and to the 
Keeper of the Department of Botany, British Museum (Nat. Hist.), and 
also to Professor Sir Isaac Bayley Balfour, F.R.S., Professor Beguinot, of 
Padua, Mr. Joseph Benbow, of La Mortola, the late M. Augustin de 
Candolle, of Geneva, Mr. F. J. Chittenden, Miss Mabel Cobbe, Mr. W. R. 
Dykes, Mr. C. C. Lacaita, Miss C. E. Larter, Mrs. J. J. Lister, Mr. R. I. Lynch, 
Mr. J. H. Maiden, F.R.S., of Sidney, Dr. Schonland, of Grahamstown, and 
Professor A. C. Seward, F.R.S. I am much indebted to Mr. W. R. Dykes 
and to Professor J. Bretland Farmer, F\R.S., for suggestions and criticism, 
and to Miss E. R. Saunders for facilities for carrying out this work in the 
Balfour Laboratory, Cambridge. 
Regarding the classification of the Iridaceae there is considerable 
difference of opinion ; in the present paper I have provisionally adopted 
the scheme used in Engler’s ‘ Pflanzenfamilien \ 2 i 
2. Ensiform Petiolar Leaves. 
(i) The occrurrence of ensiform leaves in the Iridaceae. 
In a previous paper 3 I have briefly discussed the ensiform (or 
isobilateral equitant) leaf, and its relation to other phyllodic types. This 
form of leaf — the most widespread of those met with amongst Monocoty- 
ledons which show the type of anatomy that I have interpreted as 
phyllodic — is characteristic of the majority of the Iridaceae ; 4 it occurs in 
a large proportion of the Iridoideae and Ixioideae, though not in the 
Crocoideae. In examining the leaves of the family in the herbarium of 
the British Museum (Nat. Hist.) I have seen ensiform leaves in the following 
genera : Acidanthera, Antholyza , Aristea , Belamcanda , Bobartia , Crocosmia , 
1 Chodat, R., and Balicka-Iwanowska, G. (1892) ; Balicka-Iwanowska, G. (1892-3) ; Ross, H. 
(1892-3) ; Lindman, C. A. M. (1899). 
2 Pax, F. (1888). 3 Arber, A. (1918), p. 482 . 4 Ross, H. (1892-3). 
