AND ANATOMY OF THE GENUS RESTIO. 223 
The fibro-vascular bundles traverse the cellular tissue at’ inter- 
vals, in a straight direction from the base of the sheath up- 
wards. Their constitution is exactly like that of the bundles of 
the stem. 
The epidermis on the inner side of the sheath consists of smaller, 
thinner cells than on the outside; ; they are, moreover, compressed 
from back to front, and are not ium T by stomata. 
The bracts and glumes are similar in structure to the sheaths. 
The hairs that invest the carina of the glumes are long, flattened, 
tortuous, irregularly branched or compound hairs. The stigmatic 
hairs are in tufts, in appearance like those just po but 
smaller and unbranched. 
Arrangement of Species. —With reference to the grouping of the 
species, great difficulty and uncertainty must exist for the home 
observer, in the due matching of the sexes. In many cases one 
sex only is known, while in others the male and female plants are 
so different in appearance that they have been described as con- 
stituting different species. For these reasons I have, in the fol- 
lowing arrangement, deviated but little from that published by 
Kunth, which is, I believe, on the whole the best that has been 
proposed. When the male and female plants of all the species are 
known with certainty, a more natural arrangement can doubtless 
be made by grouping the species according to the form and 
arrangement of the male and female spikelets. 
I have endeavoured to group the species in a manner that may 
be practically useful to botanists. No one knows so well as 
myself the defects and shortcomings that pervade this arrange- 
ment, but after numerous trials the one adopted has seemed the 
best. I first of all take Kunth’s first section, or true Restios, in 
which there are two connate styles. The male flowers of this 
section are readily recognizable by the absence of a rudimentary 
pistil or pistillodium. The presence or absence of staminodia in 
the female flowers, or of pistillodia in the male flowers, I find to 
afford valuable characters from the constancy of the distinguishing 
marks so afforded. The next subdivision is according to the pre- 
sence of one, two, or of more florets in the ripe female spikelets. 
This group is again subdivided into those which have loose sheaths 
and those in which those organs closely wrap around the culm. In 
practice I find this to afford valuable characters, though apparently 
of little physiological importance. The succeeding groups are 
established according to the arrangement of the inflorescence in 
the male plants. I have experienced greater difficulty in subdi. 
