r bracts
with no trace of floral structure. Fig. 79 shows this in a species of
_Willdenovia_, and a very good example is figured in a bamboo,
_Pseudostachyum polymorphum_, by General Munro.[156]

"Rose willows" (fig. 80) owe their peculiar appearance to a similar
cause, the scales of the catkin being here replaced by closely crowded
leaves. These aggregations of scales or leaves are not confined to the
inflorescence, but may be found in other parts of the plant, and may be
frequently met with in the willow, birch, oak, &c., generally as the
result of insect puncture. On the other hand, the production of leaves
or leaf-buds in place of flowers is, as is well known, generally the
consequence of an excess of nutrition, and of the continuance rather
than of the arrest of vegetative development.[157] It has even been
asserted that a flower-bud may be transformed into a leaf-bud by
removing the pistil at a very early stage of development, but this
statement requires further confirmation.[158]

=Viviparous plants.=--The spikelets of certain grasses are frequently
found with some of their constituent parts completely replaced by
leaves, like those of the stem, while the true flowers are usually
entirely absent. A shoot, in fact, is formed in place of a series of
flowers. In these cases it generally happens that the outermost glumes
are changed, sometimes, however, even the outer and inner paleae are
wholly unchanged, while there is no trace of squamulae or of stamens and
pistils within them, but in their place is a small shoot with miniature
leaves arranged in the ordinary manner.

The grasses most commonly affected in this manner are _Dactylis
glomerata!_, _Poa bulbosa!_, _Poa annua!_, _P. trivialis!_,
_pratensis!_, _alpina!_, _angustifolia_, and _laxa_, _Cynosurus
cristatus_, _Festuca nemoralis_, _F. ovina!_, _Glyceria fluitans!_, _Gl.
aquatica_, _Aira alpina!_, _caespitosa!_, _Phleum phalaroides_, _Lolium
perenne!_, _Alopecurus pratensis!_, _Agrostis alba_, _Holcus mollis!_

[Illustration: FIG. 81.--Portion of panicle of _Aira vivipara_ and
separate floret.]

From an examination of the structure of viviparous grasses Von Mohl was
led to the conclusion that the lower palea is to be considered as a
bract, and not a perianthial leaf, because the base of the palea
surrounds the stem or axis of the spikelet entirely, and both its
margins cohere towards its lower extremity.[159]

A similar condition occurs not infrequently