122 A TEXT-BOOK OF GRASSES 
The student should take careful note of the theoretical relations 
of the parts of the spikelet, since it is a knowledge of these relations 
that enables one to assign a morphological status to an absent organ. 
The glumes and lemmas are morphologically equivalent, namely 
bracts. But in the great majority of species of grasses the lower 2 
bracts of the spikelet are empty and the others above contain flowers. 
By definition the lower pair are called glumes and those above are 
called lemmas. The glumes are nearly always differentiated structu- 
rally from the lemmas. 
The theory of the evolution of organisms teaches us to trace 
the development, progressively or retrogressively, of organs through 
groups of allied species. Such an examination will usually enable us 
to interpret correctly the morphology of the organs. For example, we 
wish to know the morphology of the spikelet of Reimarochloa and 
Homalocenchrus. In the former we have a spikelet consisting of 1 
empty bract and 1 flowering bract. How is this to be interpreted? 
In the first place we are confident that the genus belongs to the large 
tribe Panicez and that it is closely allied to Paspalum and Panicum. 
The typical spikelet of the Panicez consists of 4 bracts, the upper- 
most of which contains a perfect flower. This bract, by definition 
a fertile lemma, is distinctly different from those below. The first 
and second bracts are empty and by definition are glumes. The third 
is by definition also a lemma even though it contains no flower. An 
examination of the spikelets of various genera shows that there are 
all gradations between species in which the lower lemma, usually 
called the sterile lemma, contains a perfect flower (Isachne) to those 
which contain stamens, or only a palea, and finally to those which 
are empty. This, of course, confirms the statement that the third 
bract is a lemma. No transitions are found between the glumes and 
the lemma. But we do find a tendency on the part of the glumes to 
retrogress in size. The first glume is usually smaller than the second, 
and the retrogression can easily be traced through its slight develop- 
ment in Syntherisma and Panicum to its disappearance in Paspalum. 
Similarly the second glume shows a tendency to disappear, cul- 
minating in its absence in Reimarochloa. Furthermore, there is 
no tendency for the second glume to disappear before the first. From 
the above we conclude that the single empty bract below the fertile 
lemma in Reimarochloa is the sterile lemma, that is, it is homologous 
with the third bract or sterile lemma of the typical spikelet of the 
tribe. 
