32 
BOTANY: W. TRELEASE 
its relatives the sepals are blood-red even before an thesis. Some species 
are known to be apogamous and apogamy is to be expected in many 
others, a circumstance very probably connected, as in Taraxacum and 
Hieracium, with polymorphism or close affinity in species as now 
understood. 
In all of our own species the plants are strictly dioecious and, as a rule, 
staminate spikes are longer than pistillate and bear more flowers. This 
is known to be true also of a number of tropical species, such as P. Wattii: 
in others, prevailingly if not exclusively staminate and pistillate spikes 
showing something of the same dimorphism occur monoeciously on the 
same plant. Though usually not too closely appKcable as between 
related species, the number and arrangement of the flowers on a given 
spike present equally characteristic differences, but with the quaHfi- 
cation that flowers of the uppermost joints may be fewer in number and 
simpler in grouping than below, while one or two of the lower joints may 
be partly or entirely without flowers, the lowermost almost universally 
being reduced to a sterile peduncle. The greater number of tropical 
species differ from those of the north in being androgynous through the 
occurrence of a number of staminate flowers on spike-joints that are 
otherwise pistillate, or, less commonly and sometimes differentiated by 
the term 'gynandrous, ' through the occurrence of a few pistillate flowers 
on otherwise staminate joints, — as many of Eichler's accurately drawn 
plates show very beautifully. Except in a broad way, these differences 
do not appear to be practically applicable in contrasting species though 
representing in part morphological differences of fundamental taxonomic 
value. The prevaihng grouping of the flowers is in 2, 4, or 6 series on 
each joint of the spike; i.e., in 1, 2 or 3 ranks over each of the two scales 
by which it is subtended. Examples of the first and last are given by 
P. laxiflorum (2), and P.flavescens (6), and where the joints are unisexual 
these numbers commonly prevail, though four series may be found by 
reduction and as many as ten by increase when the number is typically 
six. When the joints are androgynous, the staminate flowers often 
occur at top between the normal two ranks over each scale, and this 
condition is usually accentuated on luxuriant spikes and sometimes on 
all by the downward intrusion of a partial or complete third series over 
each scale. For the separation of the groups into which tropical species 
fall, I have found it most convenient to use the prevalence of 2 or 6 series 
of flowers on the joint as a differential, providing as an intermediate the 
prevalence of the interjected two series under the designation 4+2. 
A glance at P. domingense (2), P. trinermum (4 or 4 + 2), P. hexastichum 
(6) and P. Lindavianum (6-10) will make these distinctions evident, — 
