AND CROSS-BREEDING. 159 



species are more variable than those from very distinct species, 

 and this shows that the difference in the degree of variability 

 graduates away. When mongrels and the more fertile hybrids 

 are propagated for several generations, an extreme amount of 

 variability in their offspring is notorious ; but some few cases, 

 both of hybrids and mongrels, long retaining uniformity of 

 character could be given. The variability, however, in the 

 successive generations of mongrels is, perhaps, greater than in 

 hybrids." So reservedly does Darwin deal with a subject on 

 which the opinions of others could be brought to bear ; but as 

 they are not all concurrent, and not unfrequently conflicting 

 (which they may well be from the various subjects experimented 

 on), he has said, with commendable moderation, all that can be 

 said on the subject. 



From my readers I respectfully claim the same kind indul- 

 gence which Darwin has shown to the testimony he has had to 

 deal with in judging of the views I have offered and am now 

 to offer on the experiments which I mean to lay before you. 

 But ere I enter upon them it is necessary to premise, especially 

 as regards that form of dimorphism which occurs among many 

 plants in the Linnaean classes from Pentandria (5-stamened) up 

 to Decandria (lo-stamened) in having very generally one if not 

 two pairs of stamens shorter than the other stamens in the 

 same flower, and the same dimorphic form often occurs in even 

 a more marked degree in many plants of the class Tetrandria 

 (4-stamened). It is also the distinctive character of the two 

 orders of Didynamia to have two long and two short stamens. 

 As described in my former paper, it is now seventeen years 

 since my attention was drawn to the long and short stamens, 

 but to the latter more particularly in muling operations I had 

 performed, where by using them I crossed that large species of 

 Rhododendron, R. rinnamomeum, on the pigmy Rhodothamnus 

 chamceristus. I refer to these short stamens again as the means 

 by which I succeeded in effecting some extraordinary crosses, 

 which I confidently believe but for their use and my improving 

 a propitious time would have been utterly impracticable. As I 

 have said, I at first worked only with short stamens ; these I use 

 in all cases where I wish to cross a large on a small species. 

 I have now found that the converse holds good, and use the 

 long stamens where I wish to cross a small on a large species. 

 In all extremes I use the longest or shortest pair of stamens as 

 the case demands. The short pair is generally well distanced 

 by the others the longest pair is often not just so much in 

 advance. There is often an intermediate pair of short stamens, 

 which in cases less extreme are exceedingly serviceable, but 



