282 OBSERVATIONS ON THE NATURAL FAMILY
this genus is referable to Saliva, whose principal characters would consist in the want of corolla or perhaps its accretion with the persistent style in the female florets ; in the peri- carpia being more or less winged, and presenting their disk instead of their margins to the centre of the capitulum. 102] Sir James Smith has already pointed out the error M. de Jussieu has been led into in referring Hippia minuta Linn, to his Gymnostyles nasturtiifolia, a plant much more nearly related to Hippia stolonifera of Brotero ; which, from repeated examination, I can with confidence refer to the same genus.
Gymnostyles anthemifolia is stated by M. de Jussieu to be a native of New South Wales : but as I have observed it only in cultivated ground in the neighbourhood of Sydney, and as it has certainly been found in South America, of which four other species of the genus are un- questionably natives, it has probably been imported into New South Wales, perhaps from Brazil; nor is it alto- gether improbable that Hippia stolonifera of Brotero may have been introduced into Portugal from the same quarter.
Grindelia,
described by Willdenow in the Transactions of the Natural History Society of Berlin for 1807, and subsequently in his Enumeratio Plantarum Horti Berolinensis, flowered in Kew Gardens for the first time in 1815, when I had an opportunity of examining it, and of determining its very near affinity with Donia, a genus proposed in the second edition of Hortus Kewensis, and adopted by Mr. Pursh in his Flora of North America : the principal distinction between these two genera consisting in a difference in the number of radii of the pappus, which in Grindelia is described by Willdenow as of two rays, and according to my observations has more frequently one only. But as even in Donia the number of rays, though indefinite, is variable, and the structure of the pappus is very nearly similar in both genera, which in all other respects agree, it
�� �
