that it is known how singular are the anomalies which occasionally break out in the flowers of some Orchideæ, where hybridity is quite out of the question. The wild hybrids described in Dianthus, Galium, Hieracium, and Stachys, appear to me to be exceedingly doubtful; and in the single alleged instance among Gramineæ, that of the hybrid between Ægilops and Triticum, one of the parents at least is in a cultivated state. We must also bear in mind the observation of C. F. Gaertner, how numerous are the genera, where several nearly-allied species grow together in the greatest abundance all over Europe, and are never known to hybridise. Such are Ranunculus acris, repens, and bulbosus; Brassica Sinapistrum and nigrum; Stellaria Holostea and graminea; Geranium molle, pusillum, and rotundifolium; Potentilla argentea, verna, reptans, and anserina, &c., &c.
Admitting, however, that in the extensive and diversified Flora of Europe, wild hybrids have been observed in some twenty to twenty- five genera, if we consider that the species in those genera which will hybridise are but few; that the individuals raised are always very few, and often isolated; that they are either not reproduced in a second generation, or their offspring is a further approach to the parent species; and that even two individuals sprung directly from the same two parent species generally differ quite as much from each other as from one of their parents; we shall find it very difficult to believe in the permanent establishment of wild hybrid intermediate races, distinguished by positive characters; and we cannot but reprobate the modern practice of introducing into Floras and systematic works so-called hybrid species, races, or varieties, with a pretended diagnosis, which are, in fact, nothing but descriptions of individuals. The reader is thus misled; for the chances are that the diagnosis will not apply to any fresh individual he may find of the same hybrid. A mere indication in the Flora or other work, under each parent species, of the existence or suspected existence of hybrids with such and such other species, is always sufficient for all legitimate purposes.
None of the above observations apply to artificial hybrids, the subject of so much careful experiment on the part of W. Herbert, C. F. Gaertner, A. Braun, Naudin, and others, whose labours have done much towards elucidating the physiology of hybrids in general. But the plants thus experimented upon were placed in exceptional circumstances; and the results obtained bear but indirectly on the evidences of wild hybridity, or are often indeed calculated in some measure to mislead. The fact that artificial impregnation between certain species can be effected with great facility, is no proof that these species, or others allied to them, are the more apt to produce hybrids in a wild state. It is well known, for instance, how numerous are our garden hybrids in the genus Erica. When I worked up that genus for the Prodromus, I had before me wild specimens from various collectors of almost every Cape species, and often in considerable numbers, including the original specimens of Masson, Niven, and others, from whom were obtained the majority of our garden forms; I examined them all with great care, as
