160 



THE JOUENAL OF BOTANY 



be considered essential in any study of the phenomena of 

 hybridity/'^ 



The sure method of identifying a naturally occurring plant sup- 

 posed to be a hybrid is the following. The supposed pure parental 

 species are cultivated in a garden and crossed artificially. The re- 

 sulting offspring (if any) are grown in the garden side by side with 

 the wild plant, and the comparison is made. The usual experimental 

 precautions must be taken ; and the plant should be observed over 

 a series of years, in order to allow for the differences in perennials 

 between the first year's growth and that of succeeding years. 



The experimental method is cumbrous and lengthy, but con- 

 clusive. Its adoption would obviate such a controversy as that 

 between Eev. E. S. Marshall and Mr. C. B. Clarke over the 

 existence and determination of wild hybrids {Journal of Botany, 

 xxix. 1891, pp. 225, 296 ; xxx. 1892, pp. 78, 106). Mr. Clarke 

 stated that " 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." Mr. Marshall termed this "caricature, 

 pure and simple" ; but, like all caricature, it contains an element 

 of truth. The fact on which it is founded is that systematists 

 have almost exclusively explained ignotium per ignotiis, and sub- 

 stituted questionable hypotheses for experimental proofs. Though 

 Mr. Marshall succeeded in proving his case so far as it went, the 

 lack of experimental evidence is manifest, and it was this lack 

 that made the controversy possible. 



The genus E^nlobium is remarkable for the frequency with 

 which interspecific hybrids are produced ; and as there also exist 

 a great number of forms and varieties, the genus is a difficult one 

 to study. The size and detail of Haussknecht's magnificent 

 monograph is witness to the complexity that exists in the genus. 

 Naturally occurring " hybrids " have been repeatedly observed, 

 and many such have been described with great care by Hauss- 

 knecht and others ; but it seems important for the better knowledge 

 of the genus that artificial hybrids should be made experimentally 

 between species and varieties of constant type (as tested by raising 

 from self-fertilized seed).f With this object I have produced 



* In the case of a supposed wild hybrid the speciesin whose company it 

 occurs may give vahiable hints as to its possible parentage ; and collateral 

 evidence is often furnished by a study of its geographical distribution with 

 respect to that of the supposed parent species. But such ex j^ost facto evidence 

 cannot have the cogency of direct experimental results. 



t The following are the Epilobium hybrids which are recorded in the 

 literature (so far as I know it) as having been made by design : — 



E.Lamyi x montanum: Haussknecht made the cross both ways, and found 

 the hybrids indistinguishable [Monograph, p. 27). 



E. montanum x parvijioriim : both reciprocal crosses were made by Hauss- 

 knecht (loc. cit.), and found almost identical. 



E. montanum, female x roseum, male: Focke [Pfianzenmischlinge, p. 528). 



E. montanum, female x obscurum, male: Focke (loc. cit.). 



E. montanum x '' tetragonum": Bell Salter (loc. cit.) made the reciprocal 

 crosses, and found the products identical. (See below.) 



