Rhodora 50(594): 145-147 (June 1948)
M. L. Fernald

In working over the species of Rosa in eastern North America I find myself looking upon some plants, which have been published as true species or varieties, as minor forms. Those which demand new combinations are the following:

ROSA SETIGERA Michx., var. TOMENTOSA Torr. & Gray, forma serena (Palmer & Steyermark), stat. nov. Var. serena Palmer & Steyermark in Ann. Mo. Bot. Gard. xxii. 569 (1935).

On the same page Palmer & Steyermark treated the unarmed glabrous-leaved Rosa setigera as a form, forma inermis Palmer & Steyermark, l. c., while the unarmed shrub with dull leaflets tomentose beneath was called a variety because "It appears to be more distinct and constant in its distinguishing characters, as well as more isolated geographically, than the variety based solely on the more or less pubescent character of the leaves''. Since, however, unarmed or essentially unarmed shrubs with the foliage of var. tomentosa occur far outside "the Ozark region'', in Ohio, Indiana, Alabama, etc., the geographic isolation seems less pronounced and the rose which is serene in having no prickles seems to be a form parallel with forma inermis, which in the Gray Herbarium is represented chiefly from Ozark Co., Missouri.

R. VIRGINIANA Mill., forma nanella ( Rydb.), stat. nov. R. nanella Rydb. in N. Am. Fl. xxii6. 497 (1918).

Surely Rosa nanella is only the most stunted and rather xerophytic extreme of R. virginiana. It occurs on wind-swept crests, barrens, talus and sand-dunes from eastern Newfoundland south in the coastwise area to New Jersey, but I find no morphological character to separate it from the taller and larger R. virginiana.

R. CAROLINA L., forma glandulosa (Crépin), stat. nov. R. parviflora Ehrh., var. β. glandulosa Crépin in Bull. Soc. Bot. Belg. xv. 68 (1876).

Plants with more or less glandular teeth or with glands on the leaf-rachis occur occasionally throughout the broad area of glandless typical Rosa carolina, from New England to southern Ontario and southward. They seem to have no distinctive range, and the number of glands on the foliage is thoroughly inconstant. There seems no justification for treating it as a species, as is done by Rydberg in N. Am. Fl. l. c. 500 (1918) and none for his calling this species R. serrulata Raf. As characterized by Rydberg R. serrulata is said to have ''Leaflets glandular-dentate and usually glandular on the rachis. Branches not bristly or rarely slightly so; teeth of the leaflets ovate" (Key, p. 484), "stipules ... strongly glandular-ciliate . . . ; petioles and rachis glandular-hispid ... leaflets . . . lance-elliptic or rarely oval ..., with gland-tipped teeth''. How very different was the account of R. serrulata Raf. in Ann. Gén. Sci. Phys. v. 218 (1820), repr. as Prodr. Monog. Rosiers, 9 (1820):

23. Rosa serrulata. Raf. Tige et pétioles aiguillonés et hispidules, aiguillons stipulaires droits, stipules ciliées; folioles 5-7, obovées, serretées et serrulées, glabres, pâles en dessous; fleurs 1-3, calices hispides, sépales simples serrulés: fruits globuleux hispidules. Var. rotundifolia. Aiguillons menus, droits et nombreux; folioles ovales, arrondies, hase entière.

Obs. Arbuste d'un pied; dans les bois avec le précédent, à fleurs roses peu odorantes, médiocres, pétales à peine échancrés. Les dents des feuilles sont serrulées. La variété croît en Kentuky; c'est peut-être une espèce distincte.

Surely Rafinesque would have detected the glands and he would not have called the leaflets ''obovées'' if they were "lance-elliptic or rarely oval''. As to Crépin's R. parviflora, var. glandulosa, his description was clear: "à dents plus ou moins composées-glanduleuses". Crépin cited no type but several characteristic sheets in the Gray herbarium bear his annotations made in 1896, ''Rosa humilis Dents com posées-glanduleuses", he then recognizing it as an unarmed form of R. humilis Marsh. (1785) = R. carolina L. (1753).

R. BLANDA Ait., forma alba (Schuette), stat. nov. Var. alba Schuette ex Erlanson in Papers Mich. Acad. Sci. v. 88 (1926).

A frequent albino, but certainly a mere color-form, not a geographic variety.

* extended forward.

One of the most remarkable of roses only recently described is Rosa Rousseauiorum Boivin in Naturaliste Canadien, lxxii. 225 (1945), the third very characteristic species endemic to the area centering on the lower River and the Gulf of St. Lawrence, the new one strongly marked by its very large and dilated stipules (2-3.5 cm. long) bordered by crowded red stipitate glands so that the teeth often appear glandular-pectinate, the sepals 1.8-2.5 cm. long. The earliest collections cited for this endemic of the lower St. Lawrence were made in 1927. It is, therefore, worth noting that among the accumulation of "unidentified" roses in the Gray Herbarium there is a very characteristic specimen from "Canada, Herb. Shepard'', with Crépin's note ''R. blanda Ait. var. à dents composées-glanduleuses''. What Crépin could not have known from this very old specimen (just in bud) is the fact that in maturity the sepals would have become reflexed against the fruit, the very striking character which distinguishes this species, R. Williamsii Fernald and R. johannensis Fernald from R. blanda, in which the sepals form a porrect* beak at summit of the fruit. Since this specimen came from "Herb. Shepard'' it is probable that John Shepard received it from Frederic Pursh, who explored the lower St. Lawrence.