222 BENEDICT: NEW VARIETIES OF NEPHROLEPIS 
This variety is called mucosa (PLATE 12, FIG. 6) because of the 
dense, finely divided condition of the leaf. It seems to represent 
the same amount of division of leaf as is seen in the forms elegan- 
tissima and elegantissima-compacta, that is, it is thrice-pinnate. 
Tracing the ancestry of muscosa, we find the four forms, bostoni- 
ensts—Piersoni—superbissima—muscosa, which are, respectively, 
one-pinnate, two-pinnate, dwarf, and three-pinnate. 
Two generalizations are suggested by the facts shown in the 
preceding series, and borne out by others to be shown later, which, 
if thoroughly understood, will simplify the description and expla- 
nation of the remaining forms. They rest not only on these few 
examples, but, as will appear, are true in the case of nearly all 
the variations considered in this paper. 
I. Progressive variation takes place naturally along the three 
lines already described, viz., increase in leaf division, increase in 
ruffling or crisping, and dwarfing. f 
2. Any form which has not reached the limits of possibility in 
variation along the first and last mentioned lines, may be expected 
to give rise to new forms showing further progressive variation 1m 
one or both of them. 
To particularize, we may say that any once-pinnate form 0g 
be expected to give rise to a twice-pinnate variety. A  twice- 
pinnate variety may be expected to give rise to a thrice-pinnate 
form. The most highly divided form which has so far appear 
is five times pinnate (see PLATE 10, FIG. 5; PLATE II, FIG. 6), and 
there appear to be constitutional physiological reasons why further. 
division is unlikely. As is the case in the progressive division 
in the leaf, so also with dwarfing or pumilism. Any large form 
whether once- or twice-pinnate, may be expected to give rise to 
a dwarf variety, as superbissima was derived from Piersoni, and 
as dwarf forms have occurred in the other series to be describ 
below. With respect to progressive increase in ruffling, it is un- 
certain whether the second generalization applies to this line of 
variation or not. It may be that it does not apply in this line at 
all, or it may be that in the division series, a change in the degree 
of waviness would be too relative to admit of exact determuna- 
tion. It is certain that dwarfing occurs in all three lines, and that 
increase in division occurs also in the dwarfing series. Whether 
