34 NATIVE BRITISH ORCHIDACE.E 



successfully raised.' Incidentally it may be remarked that species between which natural 

 hybrids practically never occur are remarkably faitlaful to type, whilst species prone 

 to hybridisation, such as Orchis nmculata, present a perfect riot of variation (PL 50). 



When the existence of natural hybrids was at length admitted, it was still mamtamed 

 that only crosses between races or varieties were fertile, that those between species 

 were sterile, and that none could arise between genera. Experiment has shown that 

 not only can hybrids occur between genera (of which indeed Gymnigritelk smveokns 

 mentioned above was an example), but that various hybrids have been raised involving 

 three distinct genera. A list of thirteen and coloured plate of nine European bi-generic 

 natural hybrids appeared in Genetica, ix, 24 (1927). The behef that the existence of 

 bi-generic hybrids amounted to a proof of the identity of the parent genera has no 



foundation in fact. 



European orchids, except about ten wholly or partly self-fertilised species, are 

 mainly visited by Hymenoptera. They alight at the base of the spike and work 

 upwards, thus assuming the right position for the removal of poUinia, and nearly 

 always confine themselves to the species which they first begin to visit on startmg 

 their round. This selective instinct plays a great part in keeping species true to type, 

 and the fidelity with which they carry out this principle is remarkable. A natural 

 hybrid is an accident, and as far as my experience goes, such accidents are rare, often 

 extremely rare (with the partial exception of the marsh orchids). Bees do not often 

 appear to change over in the course of their round from one species to another, 

 unless the species with which they began proves too few and far between. Hence 

 the conditions most favourable for the production of hybrids are not when both 

 parents are abundant, but when one is plentiful and the other scarce, or when few 

 plants of the one are in flower whilst the other is in full bloom. Of some hybrids 

 only a single specimen has ever been found, whilst others, e.g. Caloglossum viride 

 X Gymnadetm conopsea, have only occurred in one or two localities, though the parents 

 often grow together. This may be due to the presence of an insect in the localities 

 in question which is absent from continental stations, or to pure accident. 



Transmission oj characters. All the characters of each parent are rarely traceable in 

 a hybrid. One or more characters must be perceptible, or they could not, if growmg 

 wild, be recognised as hybrids. Even conspicuous characters may disappear in crosses 

 between different genera, e.g. in all hybrids between Orchis and the spurless S. European 

 Serapias, the spur of Orchis appears to be entirely suppressed. There was once an 

 idea that the leaves of a hybrid resembled those of one parent, and the flowers those 

 of the other. An example of this is seen in PI. 50, B, Orchis incarnata x 0. maculata, 

 in which the leaves were those of incarnata, and the flowers were much like those of 

 maculata. This, however, is quite exceptional. One plant of Orchis papilionacea x 



' Genetica, ix, 26 (1927). ^ 



