44 Benedict : New hybrids in Drvopteris 



at all. This, of course, is likely to be especially marked if the 

 parent species differ markedly in habit. But even in such cases 

 one parent may predominate over the other to such an extent that 

 the hybrid may be identified as one of the parent species. 



B. The final tests depend then on an examination of the more 

 critical characters and the value of these characters seems to vary 

 inversely with the size of the part of the plant examined. For 

 example, D. Goldiana and D. Clintoniana, which are often very 

 similar in general leaf-form, differ in the shape and cutting of the 

 pinnae and o( the pinnulae, in the stipe-scales, and in the cell- 

 structure of the indusia. The characters of the indusia have been 

 very little used in fern classification, but they frequently furnish 

 valuable tests in determining the identity of a given specimen, and 

 are particularly useful in the class of ferns treated in this paper, 

 Dryoptcris Goldiana and D. marginalis have characteristic indusia, 

 as can often be determined from a gross examination, and in cell- 

 structure, these can easilv be distintruished from the indusia of D. 



D 



The 



last-mentioned have indusia of the same general structure as have 



cristata. but are much thinner, and in the 



D. Clintojiiana and Z? 



case of D: intermedia are invariably glandular, a character which 

 distinguishes this fern from all the other species of Dryoptcris 

 that occur in association with each other in this region. D. di- 

 iatata^ which Is occasionally glandular, is a fern of high altitudes 

 in this latitude and need not be considered under ordinary circum- 

 stances, although where it does occur with other species, hybrids 

 should be looked for. A hybrid between this fern and D. Filix- 

 mas is already known in Europe. But aside from this possible 

 exception, it seems safe to say that the presence of glandular in- 

 dusia on a suspected plant indicates D. intermedia as one of the 

 parent species. From what has been said, it will be seen that a 

 primary requisite in the study of fern hybrids is a thorough knowl- 

 edge of the parent species, not only of the general synoptical char- 

 acters, but also of the more minute differences. 



2. Sterility, greater vigor, and abnormality. The material of 

 D, cristata x uia^^glnalis and also of the other plants considered to 

 be hybrids, is almost uniformly sterile. This sterility manifests 

 it.qelf either in the form of abortive undersized sooran^jia. or. if the 



