212 BENEDICT: NEW VARIETIES OF NEPHROLEPIS 
is of course much larger still. Some deduction from the total 
thus indicated must be made in connection with smaller growers 
because of the fact that many buy part or all of their stock of 
ferns each year as small plants to raise to larger sizes. A single 
plant may thus be sold twice at wholesale, and once at retail, 
and the non-growing jobber may come in to make a third wholesale 
transfer. Taking everything into consideration, however, it seems 
safe to put the total number of plants raised every twelve months 
at not less than one million, and probably nearer two. As will 
readily be seen, this immense number offers large possibilities 
for the occurrence of variation. 
A third factor connected with the selection and preservation of 
new varieties of these ferns is found in the fact that some of these 
varieties have been the means of large profits to their producers. 
The first good two-pinnate sport of var. bostoniensis created a furor 
in the florist trade, and was awarded gold medals and other insignia 
which no other new plant has ever received. The regular price 
for a small plant of a standard variety from a two and a half inch 
pot is five cents in hundred lots. New forms of merit now sell 
at introduction for twenty-five cents apiece, and some new forms 
have sold as high as one dollar per plant. In the past this has 
led to the exercise of great watchfulness for new forms of possible 
merit, and has thus aided in the discovery and development of 
new varieties. So many forms have now been developed, however: 
that there appears to be a reaction from the interest in new forms. 
The figure as to the number of plants grown per year suggests 
the possibility of determining an approximate coefficient of varia- 
tion, but this seems to be an almost impossible problem. For one 
variety about which I have made inquiry, the number indicated is 
at least one in three thousand. For another variety which 's 
appearing rather frequently in different establishments at present, 
the coefficient is undoubtedly higher, perhaps one in one thousand. 
For yet another variety a figure of one in a million is probably too 
high an estimate. It should be added that these figures are for 
progressive variations only. Reversions or regressive variations 
appear much more frequently as a rule. 
Two conclusions are evident from the foregoing: (1) the ©o- 
efficient of variation differs with different varieties; (2) the co 
efficient of variation is never high for progressive variations. 
