XIV FUNDAMENTAL PROBLEMS ^ 433 



and with our very limited number of trees and shrubs we 

 have about eighteen spiny or prickly species, more, apparently, 

 than in the whole endemic floras of the Mauritius, Sandwich 

 Islands, and Galapagos, though these are all especially rich 

 in shrubby and arboreal species. In New Zealand the prickly 

 Eubus is a leafless trailing plant, and its prickles are probably 

 a protection against the large snails of the country, several of 

 which have shells from two to three and a half inches long.^ 

 The " ■wild Spaniards " are very spiny herbaceous Umbelliferre, 

 and may have gained their spines to preserve them from being 

 trodden do'wn or eaten by the Moas, which, for countless ages, 

 took the place of mammals in New Zealand. The exact use 

 or meaning of the spines in palms is more doubtful, though 

 they are, no doubt, protective against some animals ; but it is 

 certainly an extraordinary fact that in the entire flora of the 

 Mauritius, so largely consisting of trees and shrubs, not a 

 single endemic species should be thorny or spiny. 



If now we consider that every continental flora produces 

 a considerable proportion of spiny and thorny species, and that 

 these rise to a maximimi in South Africa, where herbivorous 

 mammalia were (before the settlement of the country), perhaps, 

 more abundant and varied than in any other part of the 

 world ; while another district, remarkable for well-armed 

 vegetation, is Chile, where the camel-like vicugnas, llamas, and 

 alpacas, and an abundance of large rodents wage perpetual 

 war against shrubby vegetation, we shall see the full signifi- 

 cance of the almost total absence of thorny and spiny plants in 

 the chief oceanic islands ; and so far from " excluding the 

 hyjDothesis of mammalian selection altogether," Ave shall find 

 in this hypothesis the only satisfactory explanation of the 

 facts. 



From the brief consideration of Professor Geddes's theory 

 now given, we conclude that, although the antagonism between 

 vegetative and re])roductive growth is a real agency, and must 

 be taken account of in our endeavour to explain many of the 

 fundamental facts in the structure and form of plants, yet it 

 is so overpowered and directed at every step by the natural 

 selection of favourable variations, that the results of its 



^ Placostylis bovinus, Zh inches long ; Paryplianta Busbyi, 3 in. diani. ; 

 P. Hochstetteri, 2| in. diam. 



2 F 



