From these circumstances this plant seemed to me worth studying, more especially from the great contrast between the small, dull, elongated, irregular flowers of the wild pansy, and the beautiful, flat, symmetrical, circular, velvet-like flowers, more than two inches in diameter, magnificently and variously coloured, which are exhibited at our shows. But when I came to enquire more closely, I found that, though the varieties were so modern, yet that much confusion and doubt prevailed about their parentage. Florists believe that the varieties (10/184. Loudon's 'Gardener's Magazine' volume 8 page 575: volume 9 page 689.) are descended from several wild stocks, namely, V. tricolor, lutea, grandiflora, amoena, and altaica, more or less intercrossed. And when I looked to botanical works to ascertain whether these forms ought to be ranked as species, I found equal doubt and confusion. Viola altaica seems to be a distinct form, but what part it has played in the origin of our varieties I know not; it is said to have been crossed with V. lutea. Viola amoena (10/185. Sir J.E. Smith 'English Flora' volume 1 page 306. H.C. Watson 'Cybele Britannica' volume 1 1847 page 181.) is now looked at by all botanists as a natural variety of V. grandiflora; and this and V. sudetica have been proved to be identical with V. lutea. The latter and V. tricolor (including its admitted variety V. arvensis) are ranked as distinct species by Babington, and likewise by M. Gay (10/186. Quoted from 'Annales des Sciences' in the Companion to the 'Bot. Mag.' volume 1 1835 page 159.) who has paid particular attention to the genus; but the specific distinction between V. lutea and tricolor is chiefly grounded on the one being strictly and the other not strictly perennial, as well as on some other slight and unimportant differences in the form of the stem and stipules. Bentham unites these two forms; and a high authority on such matters, Mr. H.C. Watson (10/187. 'Cybele Britannica' volume 1 page 173. See also Dr. Herbert on the changes of colour in transplanted specimens, and on the natural variations of V. grandiflora, in 'Transact. Hort. Soc.' volume 4 page 19.) says that, "while V. tricolor passes into V. arvensis on the one side, it approximates so much towards V. lutea and V. Curtisii on the other side, that a distinction becomes scarcely more easy between them."
Hence, after having carefully compared numerous varieties, I gave up the attempt as too difficult for any one except a professed botanist. Most of the varieties present such inconstant characters, that when grown in poor soil, or when flowering out of their proper season, they produced differently coloured and much smaller flowers. Cultivators speak of this or that kind as being remarkably constant or true; but by this they do not mean, as in other cases, that the kind transmits its character by seed, but that the individual plant does not change much under culture. The principle of inheritance, however, does hold good to a certain extent even with the fleeting varieties of the Heartsease, for to gain good sorts it is indispensable to sow the seed of good sorts. Nevertheless, in almost every large seed-bed a few, almost wild seedlings reappear through reversion. On comparing the choicest varieties with the nearest allied wild forms, besides the difference in the size, outline, and colour of the flowers, the leaves sometimes differ in shape, as does the calyx occasionally in the length and breadth of the sepals. The differences in the form of the nectary more especially deserve notice; because characters derived from this organ have been much used in the discrimination of most of the species of Viola. In a large number of flowers compared in 1842 I found that in the greater number the nectary was straight; in others the extremity was a little turned upwards, or downwards, or inwards, so as to be completely hooked; in others, instead of being hooked, it was first turned rectangularly downwards, and then backwards and upwards; in others, the extremity was considerably enlarged; and lastly, in some the basal part was depressed, becoming, as usual, laterally compressed towards the extremity. In a large number of flowers, on the other hand, examined by me in 1856 from a nursery-garden in a different part of England, the nectary hardly varied at all.