ON THE IMPROVEMENT OF PLANTS. 449 



minated ! Three of the plants were curious cross-breeds, 

 of no floral value, and having little in common with either 

 parent ; and one, in leaf, habit, and flower was very 

 similar to the wild Dog Rose. As the seeds were sown in 

 pots and placed under glass I do not think any stray seed 

 could have found place among them. I was here, no 

 doubt, unfortunate in the choice of my parent or parents, 

 and regard this issue as an instance of the well-known 

 tendency which the offspring of some cultivated plants 

 have of reverting to the normal form. The same year I 

 crossed about forty other flowers, but the crop of seed was 

 indifferent, and the result nothing worthy of record. This 

 non-success led me to submit to microscopical examination 

 the flowers of a number of varieties of the Rose with the 

 view of ascertaining which were likely to prove the best 

 seed-bearers. The result of that examination is given in 

 the " Rose Garden," ninth edition, p. 112 et seq. The 

 conclusions I drew were : 



First. That certain varieties are sterile, and incapable 

 of forming perfect seeds under any circumstances. Of 

 these I found such kinds predominate as roll the petals of 

 the flowers inwards, the centre of the flower being 

 quartered in the manner of a crown. In others the pistils 

 were weak or imperfect. Secondly. That many kinds 

 where the pistils are perfect, which in their natural state 

 form seed-pods which wither before arriving at maturity, 

 may be induced to perfect their seeds by artificial fertilisa- 

 tion. This class of Roses is the best for him who intends 

 raising seedlings to choose his female parents from, because 

 there is little here to interfere with, mar, or counteract his 

 plans. Be it remarked, however, that there are certain 

 kinds which must not be confounded with the above 

 kinds which, owing to the length of time the seed-vessels 

 are in arriving at maturity, never perfect their seeds in this 

 country. Thirdly. That those kinds which we find 

 seeding abundantly in their natural state are self-fertilised, 

 and that their abundant production of seed is due to this 



2C 



