108 
The reference is undoubtedly incorrect. the plant being a narrow-leaved, 
rather glabrous form of the V. pubescens Ait. 
V. primulaefolia 1.., reported as rare in moist soil in Gibson and Posey 
counties, by Dr. Schneck, I am forced to regard as a form of V. blanda 
Willd. V. primulaefolia is an eastern plant, ranging from New England 
to Florida near the coast. A glance at the descriptions of blanda and 
primulaefolia will serve to show how, with slight foliar changes, it might 
be possible to mistake the two forms. I have examined for intervening 
stations so far as I was able, but have found none that indicate even the 
slightest western movement of the species. 
V. rostrata Pursh, reported from Jefferson County (“Clifty Ravine’), by 
C. R. Barnes, and from Noble County, by VanGorder is a rather rare 
northern form, extending southward along the Alleghanies. Of the two 
stations, that in Noble County would be the more probable. I have seen 
no specimen verifying either citation, but because of the known range of 
the form am inclined to refer it to a form of V. striata Ait. The most 
constant difference between rostrata and striata is in the spur. In the 
former it is slender and longer than the petals; in the latter it is thickish 
and shorter than the petals. It may, however, be a form of V. Labradorica 
Schrank (—V. canina var. Muhlenbergii Gray). I feel confident, however, 
that V. rostrata Pursh is not a member of the State flora. 
Viola rotundifolia Michx., reported from Dearborn County, by Dr. 
Collins, and from Jefferson County, by Professor Young, is another eastern 
mountain form, whose presence in our territory is scarcely possible and 
certainly is very improbable. The recorded range of the plant reads: 
“Cold woods; Maine to Minnesota and southward along the Alleghanies.” 
The form is so characteristic that it is difficult to understand with what 
speciés it may have been confused. The range probabilities, however, are 
so strongly against its presence in the State that in the absence of verify- 
ing specimens it must be excluded from the catalogue. 
The admitted forms of the genus are as follows: 
V. blanda Willd. 
V. Canadensis L. 
V. Labradorica Schrank (—V. canina var. Muhlenbergii Gray) a form 
not recorded north of Monroe County. 
V. lanceolata L. 
V. obliqua Hill (~V. palmata var. cucullata Gray). 
