OBSERVATIONS UPON MULING AMONG PLANTS. 167 



sitce, but artificial impregnation has failed to produce a true 

 hybrid, and the so-called hybrids are probably mere crosses of 

 varieties. Amongst cellular plants, as the sexual organs are for 

 the most part unknown, or their nature very imperfectly ascer- 

 tained, no direct experiments have been made ; and it is doubtful 

 whether the numerous forms which arise in the genus Gymno- 

 gramma, and which have veiy much the air of hybrids, are really 

 so, or mere varieties. Since, however, even dichogamous plants 

 are far less apt for hybridization than those which are her- 

 maphrodite, it seems scarcely probable that true hybrids should 

 exist in Ferns. 



Hybrids are certainly more common in polyspermous genera, 

 but this probably depends upon the greater chance of fecunda- 

 tion with strange pollen in capsules abounding with seeds than 

 those which contain but a few. It might be supposed, too, that 

 dichogamous plants would present a greater fitness, from the 

 very circumstance of the access of the native pollen being more 

 difficult ; but precisely the contrary state of things obtains 

 amongst these vegetables. Indeed, were they peculiarly suscep- 

 tible, the original races must long since have perished. 



With respect to genera, it is observable that nearly allied 

 genera show a great difference in their aptness for producing 

 hybrids. Most species, for instance, of Aquilegia admit of 

 union, as far as opportunity of trial has occurred ; whereas 

 Aconitum and Nigella have shown no such mutual attraction ; 

 while in Delphinium a union took place only between D. con- 

 solida and Ajacis. Pelargonium is notoriously apt for pro- 

 ducing hybrids, but not so Geranium, though it presents no 

 appreciable impediments. And many equally strong examples 

 might be adduced. 



Much fallacy has arisen with regard to the supposed fre- 

 quency of the union of different genera of the same family, in 

 consequence of concluding that union had taken place because 

 apparently perfect seeds had been produced, without waiting to 

 examine whether they were really fertile, and in case of their 

 germination, observing the produce. The author had himself 

 been led into error in this respect during the earlier part of his 

 studies, and a long series of experiments undertaken in conse- 

 quence of the discovery of his error produced but a single 

 successful result, namely, the union of Lychnis diurna with 

 Cucubalus viscosus. He had, however, during the course of 

 other experiments, effected a union between Lychnis diurna <j> 

 and Silene noctijlora % , as also with Agrostemma coronaria. 

 The union of Lychnis vespertina $ with Cucubalus viscosus £ 

 is much more difficult than that of Lychnis diurna $ , and the 

 hybrid type is entirely different, which gives the clearest proof 



