119 
Although in some forms no stolons of any kind are present, the majority 
possess stolons more or less well developed. In many cases they are 
wholly underground, scaly, and without fully developed leaves, or finally 
arching upward, breaking through the surface of the soil and developing 
green leaves at their tips. In other forms the stolons are wholly over- 
ground, trailing on the surface of the soil, and rooting at the nodes whenever 
the character of the ground permits the development of roots, or, as is 
sometimes the case in rich situations, winding through the bottom vegeta- 
tion, in some places as much as 6 inches or more above the ground. In 
still other cases the stolons may be developed in water, growing in long 
strands generally more or less permanently submerged. In either case the 
stolons may be simple or branched in various degrees. 
With this wide latitude in stolon development, the question is whether 
the presence of surface stolons should be deemed of sufficient taxonomic 
importance to serve as a species character. In considering this question 
it should be recalled that there are grass species outside of the genus 
Agrostis either with or without stolons, perhaps the most conspicuous being 
Phragmites communis Trin. This species as a rule has no stolons, but in 
some cases it is found with surface runners after the fashion of those occur- 
ring in many forms of the Agrostis stolonifera group. These runners, 
when well developed, are trailing on the ground, more or less freely rooting 
at the nodes, and may reach a length of 20 to 30 feet. When well devel- 
oped, the stolons, as the writer has had the opportunity to observe in 
southern Sweden on the shore of the Baltic sea, constitute practically the 
whole foliaceous part of the plant, the upright stems being suppressed in a 
seemingly direct ratio to the increase in stolon development. The stolon- 
iferous form of Phragmites communis is most strikingly different from the 
typical one, but, so far as far as the writer is aware, no modern systematic 
botanist regards it as anything but a mere variety of the species. 
As far as the genus Agrostis is concerned, it should be remembered 
that the stolonifera group is not the only one in which forms either with 
or without surface stolons are found. Thus, there are stoloniferous forms 
both in Agrostis tenuis Sibth. and A. canina L. These are universally 
regarded as mere varieties or forms of the species in question. 
In view of the above it seems logical to hold the opinion that, as long 
as the stoloniferous character in such species as Phragmites communis , 
Agrostis tenuis, and Agrostis canina is regarded at most as a varietal 
character only, the stoloniferous habit of corresponding forms within the 
Agrostis stolonifera group cannot in itself be considered but a varietal 
character either. 
Concerning the presence or absence of awn on the lemma there are 
very few modern students of Agrostis who consider this character of 
sufficient value to be regarded as a specific one. The awn in the stolonifera 
group, when present, varies considerably in length and also in respect to 
insertion. When best developed it is inserted below the middle of the 
lemma, geniculate, and protruding far beyond the apex of the spikelets. 
In other cases it is short, straight, and inserted near the tip of the lemma. 
In many instances, however, plants are found in which only a certain 
number of the spikelets are awned, whereas others in the same panicle 
are perfectly awnless. These latter forms are on a par with similar forms 
of A. tenuis Sibth., of which the writer has examined numerous specimens 
having only certain spikelets awned, the awned spikelets in very many 
