154 
AMERICAN JOURNAL OF BOTANY 
[Vol. 9. 
On more than one occasion and with diverse varieties, attempts to reproduce 
some reverting leaf type by setting out the new form in bench for vegetative 
reproduction have resulted in the continuance in the new crown ot its new 
characteristics but with the reappearance in the runners derived from it 
either of the original form of division or of that form with some reduced 
stability. The larger size of the reverted crown is typical generally of 
reversions, both temporary and permanent sorts. 
With these facts in mind, exception must be taken to the conclusion 
adopted by Boshnakian (Jour. Hered. 7: 233. 1916) and repeated by 
Babcock and Clausen {Genetics in Relation to Agriculture, fig. 130 and ex- 
planatory description), interpreting a series of four connected Nephrolepis 
runners which show differences in form as four different mutations. The 
series shows respectively (i) a four-pinnate variety, magnifica; (2) a derived 
once pinnate plant; (3) a third generation (?) plant with unstable division, 
and (4) another third generation (?) plant like magnifica. Similar series 
of runners have occurred in my cultures (PI. IX) which on further cultiva- 
tion have resolved themselves into the three types of plants cited above, 
viz., (a) second generation reverted crowns, usually only one; (b) third 
generation crowns, like the original variety in division and in stability of 
division ; (c) other third generation crowns like the original variety but with 
more reverted leaves per plant. Unless the plants figured by Boshnakian 
were afterwards planted and found to continue distinct, it is not safe to 
cite them as so many mutations; rather, they seem to belong in the category 
of fluctuations. It has been my experience that while a plant of a divided 
variety may often produce occasional reverted leaves and sometimes a 
new runner plant with only reverted leaves, the production of a stable self- 
reproducing reversion, in other words a mutation, is most uncommon. 
In general, any given division variety may be expected to produce in 
cultivation one of three types of plants. The most numerous — an over- 
whelming majority — will be of the form typical for the parent stock, the 
only modification being in occasional single reverting leaves such as occur 
in the stock plants themselves. Second, an occasional plant will be de- 
veloped which will appear almost entirely reverted but which, on further 
reproduction, will develop new plants like those of the original form, or 
at most with some reduction in stability of division. Such plants would be 
classed as fluctuations. It is obviously important for growers in selecting 
crowns for propagation to avoid such undesirable stock plants. Very 
rarely, a third type of plant will appear, distinct not only in its own charac- 
teristics but also in its runner progeny, a real mutation. These are some- 
what more common in the regressive direction than in the progressive. 
There is a very interesting morphological problem involved in this 
matter of the production of new forms, whether of regressive or of progres- 
sive type. The problem can only be indicated here, but the main facts 
are worthy of note. It is a well known fact that the stolons of Nephrolepis 
