March, 1922] BENEDICT NEW VARIETIES OF NEPHROLEPIS I4I. 
appear in the division series and in the ruffling series, and division may ap- 
pear in the dwarfing series (cf. chart, page 144). 
5. The varieties, even in a single series, are almost invariably charac- 
terized by several associated differences. Thus Giatrasi, which is half the 
size of hostoniensis, differs also in the possession of red, wiry, somewhat 
sinuous petioles and rachides, and by differently shaped pinnae. In only 
a few cases is there but one distinguishing character obvious. 
6. The new varieties are stable, i.e., they reproduce themselves true to 
type in a high percentage of cases. Occasionally in the division-series forms,, 
reverting leaves may appear; that is, leaves with less division than is 
typical; and furthermore, new runner plants may develop which show a 
return in characters toward the parent form. The latter are uncommon, 
however, and, as will be indicated later, are actually new saltations or 
mutations. Some varieties have not as yet been known to show any vari- 
ations from type. 
7. The progressive varieties have been infrequent. If reversions are 
uncommon, as noted in the preceding paragraph, new progressive forms 
have been much more so. Of the millions of plants of the original hosto- 
niensis variety grown for sale, only six or eight sports have been reported. - 
The large number of one hundred progressive sports in twenty years has- 
been due to two factors: to the fact that millions of these plants are grown- 
each year and that florists are on the watch for new forms, and to the fact;, 
also, that the coefficient of mutation has been higher in some of the derived 
varieties than in bostoniensis itself. 
As a basis for the description and comparison of the reversionary forms, 
reference is here made to the revised chart of the relationships of the pro- 
gressive varieties considered in the earlier paper. The revision consists 
mainly in the omission of a few forms of some uncertainty of origin and of no- 
particular interest in connection with the present paper. Schilleri and 
SchuUheisi, before attributed to bostoniensis as primary sports, are not in 
circulation in the trade or likely to be. Wittboldi, included doubtfully in 
the bostoniensis chart of the earlier paper, may almost certainly be eliminated 
from any such relationship. Although no plants have been obtainable from 
the original producer, specimens have been received from a number of 
sources, French, English, and American, which by reason of their exact 
agreement in characteristics may be considered authentic. Judged from 
this material, Wittboldi is a variety of some species other than exaltata, to 
which it was originally attributed. It probably belongs with biserrata. 
At any rate, it is of no interest in the present paper. 
The position of the three- to four-pinnate Amerpohli, originally placed 
as a primary sport of bostoniensis, has been changed to bring it opposite 
varieties of a similar amount of division. No more evidence of its exact 
origin is available, but since this is in doubt in any event, it seems best to 
place it with similar forms. Smithi and Craigi, originally attributed in 
