PHYLOGENY 79 
found that, in Uredinopsis, the teleutospores are irregularly 
scattered throughout the spongy mesophyll of the host. But, 
in this position, their germination, or at least the liberation of 
their basidiospores, could not take place easily until the leaf 
had decayed, The transference to a place either (1) beneath the 
cuticle, (2) in the epidermal cells, or (3) just beneath the 
epidermis, and their aggregation into definite sori which by 
their upward pressure would burst through the overlying layers, 
would both be an advance in adaptation; so these various 
positions are found to be occupied in successive genera, and the 
most effective of all (the subepidermal sorus) is alone to be met 
with in the highest groups. 
In regard to a peridium, this can be supposed non-existent 
at first, (1) because there is no peridium in the Ustilaginales, 
(2) because a peridium could not exist so long as the trichogyne 
was functional. 
Here, it is true, there is a little deviation from what might 
be expected; a peridium is found in Uredinopsis, Milesina, 
Melampsorella and Melampsoridium (round the uredo-sori), but 
in Hyalopsora this is very rudimentary or completely absent, 
and it can scarcely be compared, in any case, with the peridium 
of the ecidial stage of Puccinia, being of a very different 
character. It must be. considered as a special development, 
separately originated for the protection of the uredo-sori in 
these lower groups. When one considers the cidiospores, one 
finds them in the intermediate types either without a peridium, 
or with encircling paraphyses, or with an irregularly shaped 
peridium, and it is only in the higher forms, such as Puccinia 
and Uromyces, that the beautifully outlined “Cluster-cup” arises. 
In respect to these higher groups it has been shown else- 
where (Grove, 1913) that the Endophyllaces constitute the 
starting-point from which the varied forms of the Pucciniacese 
have been evolved. A certain amount of advance went on, of 
course, simultaneously among the Impedicellate, though to 
nothing like the same extent. 
In Endophyllum the ecidiospore which is the product of the 
fusion-cell is also the teleutospore which germinates with a 
basidium: in accordance with theory it is accompanied by 
