14^2 



quired before satisfactory results can be attained. These remarks 

 are therefore to be regarded as strictly preliminary. 



The importance of good material is so great that a few sug- 

 gestions as to collecting, it is hoped, will be pardonable. A small 

 flowering or fruiting branch makes a beautiful specimen, but for 

 systematic study it is' often quite useless, in that it fails to furnish 

 some of the more essential characters. Specimens should be col- 

 lected either in flower or in fully matured fruit — ^just before the 

 falling of the leaves; the latter is to be preferred. A few of the 

 younger, sterile growths of the season should be included. Since 

 stoutness or depauperateness modifies very markedly the char- 

 acters, it is desirable to collect from both. When possible, the 

 whole bush should be pressed, dividing it in sections if necessary. 

 At all events, enouMi of the stem should be secured to show the 

 normal arrangement of the spines ; tins not infrequently varies 



in different parts of the bush. 



Few plants are more strikingly modified by differences In en- 

 vironment than roses. Even the younger growths appear quite 

 different from the older, so much so indeed as to cause them to 

 be taken for different species. A knowledge, therefore, of tlie 

 value of characters is desirable. Quite contrary to what was 

 once diouglit, the varying degrees of pubescence, glaucousness, 

 glandulosity, and, to some extent, of prickles, possess little diag- 

 nostic value ; and are to be considered most frequendy as acci- 

 dents of growth depending on peculiarities of soil and location 

 for their development. Not that they are wholly devoid of val- 

 ue, but are so only when taken in connection with characters of 



the first order, 



r 



The larger prickles, commonly but erroneously called spines, 

 furnish important indications, not so much by their shape and 

 size, as by their arrangement on the stem and branches. Valu- 

 able as this character is, there is no other that taxes more the 

 experience and judgment of the rhodologist In rank bushes the 

 spines may be stout and curved ; hi depauperate slender and 

 straight, yet belong to the same species. They are frequently 

 absent from bushes to which they normally belong, and this from 

 no known cause. The friction of high water often removes them 

 from such as grow on the banks of streams or low places subject 



