80 ON EPILOBIUM DURIiEl. 



Haussknecht's book." I am not disposed to disparage mono* 

 graphers, and I had the pleasure of meeting Haussknecht here 

 when he was working through the Kew Epilobiums ; I have the 

 greatest respect for him botanically, and for his deep knowledge of 

 Epilobiwn ; but I am not disposed to spare his book. It bestows 

 quarto pages of descriptions upon hybrids between E. montanum, 

 collinum, Duriai, &c., when it appears difficult, even for himself, to 

 sort these species themselves apart ; and when there is no evidence 

 whatever adduced to show that the plants in question are hybrids 

 •far less what their parents were. 



I have not, as Mr. Marshall suggests, dismissed hybrids off- 

 hand ; where I have had time to come to close quarters with them, 

 however, they have invariably broken down. They largely arise in 

 the following way : — The species (A) is diagnosed as having hairy 

 leaves, acute sepals, and the closely allied species (B) as having 

 glabrous leaves, obtuse sepals. Some hybrid-monger gets an 

 example that has glabrous leaves, but acute sepals, and at once 

 describes the new " hybrid." There is of course the obvious 

 probability that (A) and (B) themselves are only the ends of a 

 series — that the two species "sensim transeunt" into each other, 

 and that there is no hybridity in the matter. But there is another 

 case which I meet with; the text-book diagnoses (A) and (B) as 

 above by " obvious " and usually sufficient characters ; but it often 

 happens that (A), though much resembling (B), may be found on 

 more careful examination to differ from (B) by many other and 

 more essential, though less patent, characters ; and it may turn 

 out that the supposed hybrid between (A) and (B) is altogether (A), 

 ?'. e., the acute sepals may carry with them several important 

 (perhaps small) structural points, while the hairiness of the leaves 

 in the species (A) may prove in a wide geographic series a variable, 

 a trivial, or finally a merely racial local character. Scirpus sylvaticus 

 Linn, has clustered spikelets, S. radicans Schkuhr has solitary 

 spikelets ; the favourite hybrid, S. sylvatico -radicans Baenitz!, has 

 some spikelets pedicelled, some sessile clustered. But the species 

 S. sylvaticus and S. radicans may be much more fully diagnosed, 

 and I am satisfied that the supposed hybrid is an accidental form 

 (not worth even mention, but should be covered by the diagnosis) 

 of S. sylvaticus, and no hybrid whatever. The makers of hybrids 

 often go no further than the diagnostic characters of systematists ; 

 their hybrids are not hybrids between any two plants that ever 

 lived, either species, crosses, or individuals, but hybrids between 

 two of the hybrid-monger's own diagnoses. 



I would illustrate my views of species -making by a concrete 

 case. Close round Andover are 3000 acres of M primeval forest " 

 — an underwood of oak and hazel exclusively. There is one frequent 

 form of hazel — var. pracox I will call it — that ripens its nut early ; 

 the nut is oblong, with a thin shell, much protruded from the 

 involucre; the plant appears of thin foliage, perhaps mainly 

 because the leaves wither early, and is at once recognised by its 

 general aspect by the children who pick nuts. There is another 



form of nut, "var. serotina" which has short ovoid thick-shelled 



