92 Mr, J. Li nd ley's Observations on 



coriaceous ; so that the thinner the lining of the cells is, the 

 thicker becomes the coat of the seeds ; as if some sort of pow- 

 erful covering were indispensable for the protection of the em- 

 bryo, and therefore supplied by the testa when the pericarpium 

 is insufficient. 



The chalaza is generally conspicuous, in the form of a some- 

 what depressed areola, situated at that end of the seed which is 

 next the hilum. Its presence proves the coriaceous envelope of 

 the abortive seeds of Raphiolepis to be testa and not endocarp. 



The embryo has the same form as the seed, in consequence of 

 the almost absolute absence of albumen, which only exists in the 

 form of a very thin scale adhering to the testa in certain species 

 of Pyrin. The cotyledons are flat, and parallel with the pla- 

 centa ; the radicula small and conical, obliquely turned towards 

 the hilum ; somewhat longer in the pinnated Pyri than in the 

 rest of that genus. 



Three-fourths of the species are found in the temperate regions 

 of Europe, North America, and Asia ; a few are peculiar to the 

 north of India, and one species comes from the Sandwich islands. 

 They would therefore have nearly the same geographical distri- 

 bution as Roses. But two species have been found in Peru by 

 Ruiz and Pavon ; and a Pyrus from Mexico, sent to this coun- 

 try by M. Pavon, exists in the herbarium of Mr. Lambert. It is 

 much to be regretted that we have no information of the altitude 

 at which these southern species were observed. 



We have only now to consider whether the foregoing genera 

 should be retained as a distinct natural order, as has been pro- 

 posed by M. Richard (see Analyse du Fruit, Eng. edit. p. 23), 

 or be understood only as a section of Rosacea, according to the 

 decision of M. de Jussieu. 



The principal peculiarity by which M. Richard proposes to 

 characterize Pomacece appears to be the ascending direction of 



their 



