March, 1922] BENEDICT — NEW VARIETIES OF NEPHROLEPIS 
originate in association with the leaves, one stolon being paired with each 
leaf in branching from the crown-stem axis. It should not be difficult to 
determine whether there is any association between leaf variations on the 
original crown and runner variation on the paired stolon. For example, 
do the stolons paired with reverting leaves tend to produce new plants 
showing a reduction in stability of leaf division in a manner analogous to 
the behavior in vai legated Pelargonium? 
The problem is not as simple as might at first appear from the above 
statement, because the relation of stolon to old and new plants is not en- 
tirely simple. The stolon from an original stock plant may bear along its 
several feet of length a considerable number of plants arising as lateral buds, 
generally on short spur branches. When the new bud plants are taken up 
for potting, each one may retain some portion of the parental stolon capable 
of continued bud reproduction. The progeny of such a plant would then 
consist of new plants from the original parental stolon together with others 
from the new stolons of the bud plant, i.e., "sisters" and "daughters." 
With the possibility of such complications, the doubts expressed with 
reference to the relationships of the series of four connected plants figured 
by Boshnakian will be readily appreciated. 
A not infrequent anomaly in the behavior of the stolons is what appears 
as a dichotomy of a stolon in which one half becomes a leaf without any 
evident associated stem axis while the other half continues as a normal 
stolon. I have seen three successive leaves produced in this manner on 
one stolon. Sometimes the stolon growing point is lost with the formation 
of a single leaf, and the appearance is given of the transformation of a stolon 
into a leaf. I have made no determination of the stelar behavior in these 
cases. 
Reversions oj other secondary sports. 
Only one other secondary sport of bostoniensis has been recorded as 
having given rise to a reversion of the mutation type, another Pierson 
variety, elegantissima. This is a variety much like elegantissima-compacta in 
leaf division, but lacking the dwarf character of that form. Robert Craig 
of Philadelphia has reported and introduced as a new variety, named robusta, 
a form which has almost complete agreement in characteristics with plain 
Piersoni (PI. VI, fig. 5). If its origin is correctly given, it is to be grouped 
with the other Piersoni-Yike reversions which have already been discussed. 
Summarizing the facts recorded for the reversions of the secondary 
sports, we find the reversionary behavior similar to that described for the 
primary forms. 
I. Reduction in the stability of the original progressive change as shown 
in the degree of division and in the proportion of divided leaves was ex- 
perimentally demonstrated. Furthermore, this reduction was shown in 
elegantissima-compacta to pass through an almost completely reverted 
