50 Transactions British Mycological Society. 
families Burmanniaceae, Orchidaceae, Gentianaceae and Erica- 
ceae as examples of this. Regarding the Pyrolaceae as a sapro- 
phytic sub-family of the Ericaceae we can trace a relation 
between increasing saprophytism and a more intensive develop- 
ment of mycorrhiza. In the root-tip region we get an ascending 
series in the amount of fungus present from Chimaphila um- 
bellata where the epidermal cells of some roots are without hyphae 
and other roots with hyphae, but not in every cell, to C. macu- 
lata with a greater number of the epidermal cells filled with 
hyphae; in Pyrola rotundifolia and P. elliptica all the cells are 
infected, and there is the beginning of intertwined hyphae round 
the root tip; then in Monotroba Hypopitys an increase in the 
width and extent of the sheaths and a division into two zones— 
an outer loosely woven mass of hyphae and an inner more 
compact one—and finally in M. uniflora a still greater width of 
the fungal sheath. In the least saprophytic species the epidermis 
soon dies off, carrying with it the fungal hyphae as in Chima- 
phila and Pyrola, whereas in Monotropa, especially M. uniflora, 
the epidermis is still living and filled with hyphae when the 
root is quite old. 
Corresponding with this increase in saprophytism there is an 
increase in the number of seeds produced and a reduction in 
their size and structure. “‘The endosperm in the Pyrolaceae 
consists of relatively few large cells—the embryo of about 
twenty-five to thirty cells with no trace of cotyledons. In the 
Monotropaceae the number of endosperm cells is still less and 
the cells are larger, the embryo also is very small, composed of 
only nine or five cells.’’ As these seeds also have their integu- 
ment in the form of a net-work there is an exceedingly close 
superficial resemblance to those of orchids. 
Comparing the members of the Pyrolaceae as a whole with 
the Ericaceae it would seem exceedingly probable that their 
seeds are even more dependent upon infection by the mycorrhizal 
fungus than are those of their chlorophyllous relatives. It will 
be interesting to learn at what stage infection takes place and 
whether or not a close approximation to the more advanced 
orchid type obtains. It is probable that the fungus will be found 
to be generally distributed in these plants as in Calluna. 
BURMANNIACEAE AND GENTIANACEAE. 
The other two families in which mycorrhizas are typically 
developed are the Burmanniaceae and the Gentianaceae*—in 
* “Most of the Orchideae are humus-plants, and it is noteworthy that 
dicotylous saprophytes, such as the Pyrolaceae, the gentianaceous Voyvia, 
and others, show a reduction of the embryo like that of the Orchideae. In 
Monotropa the embryo has but nine cells. The germination of the seeds of 
these dicotylous saprophytes is unknown. It takes place only in the presence 
