174 Evans: TAXONOMIC STUDY OF DUMORTIERA 
occur over extensive areas, according to the information at hand. 
In Europe and Africa, for example, no forms with crowded papilli- 
form cells have as yet been reported, although some of the plants 
have vestigial chambers while others have none. In such areas, 
therefore, the difficulties of distinguishing between forms with 
crowded papillae and smooth forms are eliminated, although the 
difficulties of distinguishing between forms with air chambers and 
those without them still remain. In the writer’s opinion the latter 
distinction is less significant than the former, and it seems inex- 
pedient at the present time to attempt to use it in the delimitation 
of species. 
The receptacles of Dumortiera, which represent stalked branch- 
systems, have been repeatedly described (see, for example, Ernst, 
4, Pp. 173-178). The stalk of the male receptacle is extremely 
short but shows two rhizoid-furrows, agreeing in this respect with 
the much longer stalk of the female receptacle. The disc of the 
male receptacle is not clearly lobed and the antheridia are not 
arranged in radiating rows, although they clearly arise in acropetal 
succession. The disc of the female receptacle develops normally 
from eight to sixteen groups of archegonia on its lower surface. 
hen the receptacle is young there are no distinct marginal 
lobes, but these become evident later on in case fertilization 
has taken place, the groups of archegonia being situated beneath 
the lobes. The involucre is thick and fleshy and shows a small 
apical opening; on its surface it bears scattered bristles. Similar 
bristles, which have been interpreted as modified rhizoids, occur 
on the upper surface of the disc, sometimes abundantly and some- 
times sparingly, sometimes restricted to the marginal portions 
and sometimes more evenly distributed. Vestigial air chambers 
are not developed, and the surface-cells, although sometimes more 
or less convex, do not form papilliform outgrowths, even when 
the vegetative thallus forms them abundantly. 
The features used in separating species have been drawn mainly 
from the female receptacle and relate more especially to the mar- 
ginal lobes and the number and distribution of the dorsal bristles. | 
According to Stephani (25) the receptacle of D. trichocephala is 
strongly convex and very bristly, while that of D. hirsuta is less 
convex, with the bristles confined to the marginal portions. He 
