- 9 8 - 
The complete double peristome is a device to prevent the too rapid escape 
of the spores in mosses with pendent or strongly cernuous capsules, but 
mosses growing on trees or cliffs can shed the spores much more freely and 
without waste as the much greater distance above the ground insures freer 
and more certain access to air currents. Hence imperfect peristomes would 
be no bar to reproduction, and natural selection would not keep the peristome 
up to its full perfection. In vertical erect capsules many spores must fail to 
escape at the most advantageous time or even at all until moisture or old age 
have rendered them valueless for reproduction. Hence a curved and cernuous 
or a drooping capsule with the mouth well guarded by a complete peristome 
would insure the greatest percentage of reproduction in mosses with a vertical 
seta, but if the substratum be vertical the erect capsule is horizontal or even 
pendent, as in Neckera , and the spores find easy and ready exit. 
I am not unaware of exceptions to this line of reasoning, Hypnum reptile 
with a tree habitat and curved capsules, or many species of Entodo7i with 
horizontal habitat and erect capsules with imperfect peristomes, but there 
must of necessity be some cases of progressing adaptation and of the sur- 
vival of a character after a change of habitat has rendered it more or less 
unsuited to new conditions. 
There are undoubtedly many mosses of a more primitive type of peri- 
stome in which the erect sporophyte'is the primitive condition and these as. a 
class may grow on almost any substratum, e. g. Georgia , Catharine a, many 
of the Tortulaceae . Dicranaceae and perhaps the Grimmiaceae. A second 
apparently similar case of habitat modification is found in the so-called 
Cleistocarpi and many gymnostomous species. These as a rule grow on 
moist soil, either bare from fresh disturbance or scantily covered with other 
vegetation. They are usually annual and develop their spores in the spring 
while the soil is moist, disappearing in many cases during the drier part of 
the season, though under favorable conditions it seems probable that the pro- 
tonema may persist for more than one season. The whole subject of the dur- 
ation of these forms is imperfectly known and more observation is needed. 
The leaves of these mosses are usually thin and soft although Astomum and 
Weisia still retain the leaf-structure] of the Tortulaceae and are probably 
not annuals. Some of these mosses may possibly be primitive types but 
most are degenerate members of families of a high degree of development. 
Such are Sphaerangium, Phascum , Pleuridium, Acaulon and Bruchia of 
the Cleistocarpi and Astomum , Physcomitrium, Aphanoregma and Pottia 
species, of the gymnostomous forms. Just how or to what extent the habi- 
tat of these degenerate forms has induced the common character is not clear 
but I believe there is a causal relation. 
Second. Modifications of the gametophyte. Mosses whose habitat is 
strongly xerophytic for any considerable portion of the time have small cells 
and very thick cell walls, e. g. Grimmia , Orthotrichum , many of the Tor- 
tulaceae and Leskeaceae. This condition obviously retards the escape of 
water. Manyof the larger mosses like some species of Poly trichum growing 
in places where seed plants have abundant moisture are at times thoroughly 
