FERTILISATION AND PROPAGATION. 



125 



the most appropriate for making such an experiment. There were sown 

 together the crested form of our common Royal Fern (Osmimda reqalis 

 cristata) with the North American 0. cinnamomea, as also with 0. Claytoniana 

 {interrupta) . These, again, were sown together with 0. javanica, different 

 from all other known Osmundas through its evergreen nature and the 

 coriaceous texture of its pinnate fronds ; with the slender North American form 

 of 0. regalis, called gracilis or spectabilis ; and with the crested Japanese form 

 of 0. regalis, called japonica corymbifera. These, we thought, might have 

 produced, even if only a small percentage, offspring partaking to a greater or 

 lesser degree of the character of both parents ; but, to our great disappoint- 

 ment, although several thousands of young plants were raised from these 

 sowings, they simply reproduced one or the other of the parents, and not the 

 slightest trace of variation was observable amongst them. The same thing 

 happened in the cases of other genera, though, on account of the proceedings 

 being less carefully booked, the trials were not of such importance as the one 

 referred to above. 



It is difficult, if not even impossible, to trace the parentage of such crested 

 forms of Maidenhairs as Adiantwm cuneatum grandiceps and A. c. Ludclcmann- 

 ianum, or the deeply-laciniated A. Capillus- Veneris Jissum, or, again, the drooping 

 A. Bausei and A. cuneatum defiexum ; for there were existent no parents already 

 possessing either crested, laciniated, or drooping characters, which could be trans- 

 mitted to the offspring through fertilisation. The same remarks apply equally to 

 the Davallias, two crested forms of which were raised about the same time as the 

 above-named Maidenhairs, viz., D. elegans polydactyla and D. Mariesii cristata, 

 one of which, at any rate, if not both, must have originated from parents 

 with plain foliage. These, we should say, are natural sports or accidental 

 variations in seedlings, which may possibly later on produce numerous varieties 

 more or less crested, as is now the case with the almost endless forms of 

 Pteris serrulata. It may be taken as a general rule that when the original 

 form has once varied, it has a greater tendency to vary again. In that respect 

 we have, as conclusive examples, the numerous forms of the Lady Fern, the 

 Male Fern, the Hartstongue, &c, in which the tendency to variation is so 

 developed that many distinct forms are generally found in any large batch of 

 seedlings, while a small portion only reproduce the variety from which the 

 spores have been gathered. 



