ity and it will fall from any leaf which is not absolutely horizontal and 
quiet ; this is the case in respect to any leaf the surface of which is 
unwettable. 
We have just seen that the surface tension of a drop must prevent its 
passing around a sharp edge ; now, if drops run to a toothed margin, 
they must, in part run over the sinuses. If the sinus were a curve with 
a radius equal to or greater than that of the drop, or if the radius were 
not very much less, the drop would have an area of eontact greater than 
it would have in passing an entire margin, and so it would pass more 
readily. Such sinuses are found in Dipteris , but they are so placed that 
water, to reach the majority of them, must run uphill. I have found such 
sinuses in no other fern. If a drop runs to a sharp sinus, or to one 
much narrower than the diameter of the drop, then the sides of the sinus 
will hold it up and make it run outward until the sinus widens or the 
ends of the teeth are reached. If the sinus widens sufficiently to allow 
the drop to run through, then even though the frond is quiet enough 
so as not to cause it to be shaken • off when its contact is limited, and 
although it might return to the body of the frond while still touching 
both sides of the sinus, without - its having to run uphill, nevertheless 
it is more than likely to leave the frond because of its impetus in falling 
through the sinus; this condition is easily demonstrated. If a drop 
runs out onto a tooth, it loses its opportunity to pass to the nether surface 
by diminishing its possible contact, just as it does if it runs on a caudate 
tip. In general water must run to the teeth, rather than over a' sinus, 
because its surface tension prevents its starting over an edge if it can 
run along it. . Thus "a toothed or cut margin in all parts of its periphery 
is provided with a water-removing structure' such as caudate leaves have 
at the apex. I have demonstrated the inability of water to pass around 
a narrowly or sharply cut margin by experiments on various ferns. 
Among ferns with serrate margins are Nephr odium syrmaticum, 
Davallia decurrens, Diplazium pallidum ) D. 1667, D. polypodioides, 
Asplenium vulcanicum, A. macro phyllum, Athyrium silvaticum, Ble- 
chnum egregium , sterile parts of Pteris ensifolia (a serrate fertile margin 
of Pteris is impossible), and the fertile frond of Drynaria rigidula. 
Beside being serrate (rather obtusely), Davallia decurrens has the seg- 
ments so close together that a drop will run on the upper surface from 
one to another. In practice I have watched a drop run across eight 
segments without a particle passing through between them. Ferns with 
more deeply incised margin are Polystichum, Nephrodium 1712, N. in- 
termedium, N-. Bordenii, Aspidium difform'e , ' Davallia solida, Micro- 
lepia hirsuta, LindSaya Havicei, Diplazium Williamsi, D. dolichosorum, 
D. meyenianum, D. fructuosum, Asplenium caudatum, A. cuneatum, 
Adiantum diaphanum, A. mvndanaoense, Scliizostege calocarpa and 8. 
pachysora. A very much larger proportion of terrestrial than of- epi- 
phytic ferns have such margins; as epiphytes by virtue of their position 
