55 
has a fit companion in that other which interprets the leaves of Dischidia 
as protectors of the roots, but does not tell us what purpose roots serve 
in such a place. 28 As a matter of fact, these plants are also myrmecoph- 
ilous, the leaves furnishing shelter for ants, and the ants furnishing food 
which the roots absorb. Dischidia is rarely without ants and rarely 
without a considerable amount of debris about the roots inside each leaf 
brought by them. There are other Asclepiadacece, epiphytic without 
evident structural modifications, the roots of which are invariably in 
aerial ants’ nests. 
In all these cases it is likely enough that the plant derives some 
organic as well as mineral food from its tenants. 
REPRODUCTIVE STRUCTURES. 
The principles underlying the adaptations of the reproductive struct- 
ures of ferns (sporophytes) are very simple. The sporangia must be 
protected during their development against injury by desiccation or other- 
wise; the mature spores must dry thoroughly enoughly to be easily and 
well scattered ; and the drying of the spores must not involve too great a 
desiccation of the frond, for an insignificant number of Philippine ferns 
suffer an annual loss of their leaves. The structures found in ferns 
are a compromise between these rather antagonistic principles. 
Ferns almost always protect their sporangia, at the same time that 
they avoid interference with the illumination of the assimilating organs, 
by restricting the former to the nether surface; our physiological ex- 
ceptions are Psomiocarpa and Stenosemia, the vegetative and reproductive 
fronds of which are distinct, and Lecanopteris , which may not be entirely 
dependent on photosynthesis for its organic food. 
For the sake of facile nutrition and to preserve the normal exercise 
of its functions by the nether epidermis, the sporangia of practically all 
ferns, the vegetative and reproductive fronds (or pinnae) of which are 
alike, are collected into sori. Most ferns protect these sori by means 
of indusia. At San Eamon, 60 per cent of all Polypodiacece have indusia, 
the remaining 40 per cent including 13 members of the old genus Achros- 
tichum and a number formerly put into Gymnogramme, beside all those 
with well-defined nude sori. In the indusiate list are included the 
Pteridece (not including “Gymnogramme”), they having, bionomically, 
indusia as truly as any ferns do. Any full discussion of the forms and 
origin of indusia would be superflous here, in view of the attention they 
have received as most important structures in taxonomy, but it is per- 
tinent to the subject of this work to point out that their structure fits 
the local demands upon it. Thus, it is leathery in the two strongly 
28 Scott and Sargant, in Ann. of Bot. (1893), 7: 243, suggest that the roots are 
to absorb water, those in the inverted pitchers of D. rafflesiana condensing the 
water transpired by the interior of the leaves. 
