334 Arber . — The Leaf Structure of the Iridaceae , 
phyllodes, but even the minor detail of the association of ensiform leaves 
and winged axes, sometimes met with in the Iridaceae and other Mono- 
cotyledons, can be paralleled in the genus Acacia (pp. 308-10). The nature of 
the ‘ radial leaves J of Hermodactylus , certain species of Iris , &c., is discussed, 
and it is shown that these leaves are merely variants upon the ensiform type 
(pp. 3 10- 1 2), and that they lead on to the leaves of certain Irises and Moraeas, 
which are mainly dorsiventral, but have a monofacial apex ; such leaves 
are regarded as leaf-bases terminating in a more or less vestigial petiole 
(pp. 3 1 2-1 6). The leaves of the Juno Irises are interpreted as being exclusively 
of leaf-base nature (p. 314). The leaves of the Ixioideae are next con- 
sidered, and the origin of the cruciform leaf of Gladiolus tristis is traced 
in detail. It is shown that all such leaves are again mere modifications of 
the ensiform type, and that the leaves of various species of Gladiolus can 
be paralleled within the genus Acacia (pp. 317-19). Attention is drawn to 
the resemblance which many Irids and Acacias bear to one another in their 
tendency to excessive foliar fibrosis, and the significance of this is considered 
(pp. 319-20). The peculiar foliated leaf-types met with in Babiana , Cypella , 
&c., are examined, and it is shown that they arise from a simple petiolar 
structure, through invaginations, sometimes associated with the development 
of keels or wings (pp. 320-3). The leaves of the Crocoideae then come 
under consideration, and it is demonstrated that the gulf hitherto supposed 
to separate those of Crocus and Romulea has no real existence, but that 
the transition from sheath to limb shows that both leaves are petiolar 
phyllodes of an essentially similar type and that their divergent mature 
forms are merely the results of differing types of invagination (pp. 323-8). 
It is pointed out that the tendency to reduction or loss of the median 
bundle, which seems inherent in the leaves of the Iridaceae, can be paralleled 
in certain rare cases among Dicotyledonous petioles (pp. 328-9). 
Finally, the evolutionary history of the Irid leaf is discussed, and it is 
concluded that the ensiform petiolar phyllode is probably primary for the 
family, and that the various leaf-types met with at the present day are to 
be interpreted either as reduced from the ensiform type by the more or less 
complete loss of the petiolar region, or else as elaborated from this petiolar 
region by invagination and winging (pp. 329-33). 
Balfour Laboratory, Cambridge. 
