NOTES ON EPILOBIUM HYBEIDS 159 



of hybrids occur in this genus, therefore many plants will yield 

 mixed characters," though we may accept the statement without 

 demur, it must not blind us to the fact that the historical process 

 would be more fitly summed up in the inverse form of words, 

 thus : — " A large number of plants yielding mixed characters 

 have been found in this genus, therefore extensive hybridization 

 has been assumed." 



Now it is quite clear that this method, though in the hands of 

 skilled observers it may often yield accurate results, is essentially 

 faulty. Eecent experimental w^ork has demonstrated that a cross- 

 bred is by no means necessarily intermediate between its parents ; 

 it may even be indistinguishable from one of them, and show no 

 trace of the other. The phenomena of dominance and the Men- 

 delian segregation of unit-characters are of the greatest impor- 

 tance to students of naturally occurring hybrids ; but the general 

 acceptance of these facts by systematists has been tardy, and in 

 only a few cases has a sense of their importance been expressed." 

 Whatever may be the exact applicability of the newly discovered 

 phenomena to the mutual relationships of the highly complex 

 congeries of characters which separate naturally occurring species, 

 the experimental method is now seen to be indispensable for the 

 solution of problems of hybridity. The mode of work should be 

 changed, from the almost mathematical search for extremes to 

 suit a given mean, to the emasculation and artificial-pollination 

 processes of the experimental garden. 



To take one or two examples. Writing of Epilobium roseum, 

 Sir James Smith remarked in 1800 : "Is it possible to have arisen 

 from seeds of E. tetragonum impregnated by the pollen of E. mon- 

 tanum ? " But it was not till 1842 that Bell Salter made the 

 suggested cross, t and found that the result was not identical with 

 E. roseuvi. (See below.) 



Again, I grew plants in my garden which were absolutely 

 indistinguishable from typical Epilobnnn hirsutum. These had, 

 however, been grown from seed produced by fertilising the 

 occasionally occurring white (or rather, very pale pink-tinged) 

 variety with pollen taken from the typical red-flowered plant. 

 They were thus true cross-breds ; and when artificially self- 

 fertilized they set seed which produced, out of fifteen plants, 

 eleven having the full red flower of the type hirsutum, and four 

 having the pale pink flower of the variety, the result approximating 

 to the well-known Mendelian ratio 3 : 1. 



There are thus two tests for hybridity : (1) The sowing 

 of artificially self-pollinated seeds and the observation of the 

 off-spring ; (2) The reconstruction of the hybrid by the union 

 of selected parents. The first test is often inapplicable, may- 

 be because of sterility, or the complexity of the characters 

 involved ; but the second is usually practicable, and should 



* See, for instance, Moss, "The Pimpeinels," Journ. But. I'Jll, p. 44. 

 t FInjtologist, 1852, p. 737. 



