GENERAL REMARKS, 375 



the peristome, and above all the columella. The first how- 

 ever presents a structure not much more complicated than 

 the changes undergone by the teguments of some ovules in 

 their progress to maturity. 



Also the want of correspondence in relation between the 

 embryos and the apex of the nucleus. 



Also their germinating separately from the seed, but Santa- 

 lum is analogous in some respects to this. Strong arguments 

 in favour of this view are deducible from Hepaticae, in which 

 all the anomalies are reduced, and in some of which as Ric- 

 cia, that alone involving the extreme plurality of embryos re- 

 mains. 



But so long as we view subjects in an isolated manner, 

 however ingenious the hypotheses we may frame, proofs of 

 their truth must always be wanting. We must instead of par- 

 tial views, take a comprehensive view of the vegetable king- 

 dom ; as well of its corresponding divisions in the animal 

 world, and it is perhaps one reason why botany is so back- 

 ward, that no botanist has hitherto ever looked for illustra- 

 tions of his views out of his own particular science. 



Question, are not such mosses as Schistostega, Fissidens, 

 etc., really frondose, in as much as their leaves are oblique in 

 their attachment ? At any rate, whether they are leaves or not, 

 this obliquity affords a passage to Hepaticae, in which the 

 frondose insertion is so rigidly adhered to, that scales formed 

 by the growth of the radicles obey the law. 



It is a singular fact, that of the two most reasonable hypo- 

 theses, that of viewing the sporula as pollen grains, and mine 

 as sketched above, the analogies are in a precisely reversed 

 ratio. For the first steps in the developernent of a theca 

 are fatal to its being considered as an anther ; and the last 

 steps in its developernent, fatal to the idea of its being a seed. 



The antheriform hypothesis gives no explanation of the 

 complication of the apparatus, although it fairly explains the 

 columella, yet it does not do so with regard to the inner mem- 

 brane, unless absolute reliance is placed on Brongniarts re- 



