73 



fications in the germ ])lasm must take place very frequently. 

 The problem of describing all these multifarious types is most 

 extensive and calls at first for descriptive treatment before 

 any thoroughgoing attempt is made to work out their pos- 

 sible mode of origin through experiment or cell study. While 

 the general run of papers on variation in most plants have to 

 deal with one or two new and undescribed types, the investigator 

 of Nephrolepis meets the problem of differentiating and describing 

 scores, even hundreds of new sports. Some of these new types 

 receive the recognition of a florist's name and are likely to be 

 preserved but many do not have the qualities required for horti- 

 culture and their only chance of preservation lies in a scientific 

 experimental collection. So many have been accumulated at 

 the Brooklyn Botanic Garden that in the space available it is 

 scarcely possible to do more than grow single plants of many 

 of the most important and interesting varieties. As a matter 

 of convenience, all Nephrolepis forms may be classified into four 

 groups. 



1. Systematic species. All the wild forms are once pinnate, 

 tropical ferns, found in varying habitats and differing in details 

 of habit, color, size, form, scaliness, shape of pinnae, shape of 

 indusium, and other minutiae. The American species appear to 

 number eight or nine, well distinguished on the bases of habitat 

 and form characteristics. When Old World forms are also con- 

 sidered the classification is more complicated as several of the 

 American forms occur also in Africa and Asia. Study of these 

 wild types recognized as species is of special interest as a basis 

 for comparison with the new forms continuallv appearing under 

 conditions of cultivation. 



2. Bud variation in the Boston fern. From a single variety 

 discovered in cultivation twenty-five or thirty years ago, the 

 Boston fern {Nephrolepis exaltata hostoniensis) , there have arisen 

 at least one hundred horticulturally named varieties by bud 

 sporting. As many more have appeared which have not re- 

 ceived any name. The majority of named types occurred as the 

 result of progressive variation away from the Boston fern along 

 four lines of sporting; viz., increased leaf division, dwarfing, 

 ruffling, and cresting. Six primary sports embodying these four 

 types of variation in different manifestation were followed by 

 a group of secondary sports in which the same four lines of 



