220 THE JOURNAL OF BOTANY 



Kew. is one labelled " The Highlands," a dwarf form, with a stem 

 barely 35 mm., which exactly agrees with the Spanish specimens 

 on another sheet, and is quite distinct from either of the other two 

 British varieties. On the evidence of this specimen, I think that 

 var. congesta may be added to the Scottish flora, and possibly it 

 may be found in lofty stations suitable for its growth, if diligently 

 looked for. If this can be satisfactorily settled, it will be seen that 

 all the three European varieties of Anteiuiarla dioica elvc to be found 

 in the Scottish Highlands. 



The late Dr. F. B. White, in Scottish Xaturalist, 1886, p. 323, 

 and in Fl. of Perthshire, p. 180, has described a var. peclicellata, 

 found at Glen Tilt and near Strowan Station, in which the calathia 

 are not so close together, on pedicels from 12-28 mm., and, as he 

 says, "probably only an extreme state, as intermediate forms occur." 

 It seems indeed to be only a form sliglitly different from the type 

 of the species, scarcely to rank as a variety. Among the specimens 

 in the British Herbarium, the description seems to fit some speci- 

 mens gathered by A. Croall in 1854 at Little Craigandhal, and 

 labelled, " Plants of Braemar no. 105." In these, the calathia are 

 spread out in a semicircle, owing to the greater length of the pedicels. 



In the plate which accompanies this paper, the single sheet in the 

 Linnean Herbarium, which contains eight specimens of Antennaria 

 dioica, is reproduced by photography. At the bottom of the sheet 

 is the word " dioicum " in Linne's handwriting. In three of the 

 specimens rose-coloured scales are more noticeable ; in three others 

 no rose-coloured scales are to be seen. One specimen has two 

 flowering stems. 



SOME BEITISH VIOLETS.— II. 



By Edmund G. Baker, F.L.S. 



The Pansies growing in hilly or mountainous districts in this 

 country can, with trifling exception, be arranged in two of the 

 groups referred to in my previous paper (pp. 9-12). The repre- 

 sentative species for these groups are V. lutea Hudson and V. saxa- 

 tilis Schmidt* (F. alpestris Jordan). These groups have many 

 points in common — the plants are perennials or subperennials, 

 never annuals, as is generally the case with the groups of V. ar- 

 vensis Murray and V. tricolor L. sensu stricto ; the flowers are 

 nearly always showy, the petals being always distinctly longer 

 than the sepals, except in the case of T'. lutea Huds. var. hamulata. 

 The two groups differ in the stipules, which in the group of V. lutea 

 are digitately multipartite or digitately pinnatifid, while in the 

 saxatilis group they are pinnatipartite. Particular attention must 

 also be paid to the character of the rootstock. In mountainous 

 or hilly country in close proximity to the sea — as, for instance. 



* This species has not been recorded as growing in Britain, but is widely 

 spread on the Continent. 



