REMINISCENCES. Ixxxv 



The colours are apparently somewhat variable, but less so than we 

 have been led to suppose ; but their proportions and direction in 

 the several stages in the course of reproduction seem to be very- 

 constant. It has been said that the relative length of the stamens 

 and styles is the result of dimorphism. It is doubtless so in many 

 plants, but observation has not led rubologists to the conclusion 

 that such is the case amongst Ruhi. It would appear that the 

 dehiscence of the outer ring of anthers at the time when the 

 stigmas are ripe aflFords a sufficient security for cross-fertilization ; 

 the fertilization has usually taken place before the inner rows of 

 stamens have produced any pollen ; but insects continue to frequent 

 the flowers, and convey the pollen of these later stamens to another 

 flower, having already done this with the product of the first 

 ripened anthers. 



It is very much to be wished that collectors would make a note 

 of the characters afforded by the flowers, as well as record the 

 direction of the growing stem ; as the want of such information 

 renders their specimens of very much less value. I have been as 

 much at fault as others in former years, and thus a considerable part 

 of my collection consists of specimens scarcely determinable. 



The points which seem to require especial attention are the 

 direction of growth of the barren stem of the year, the form of its 

 transverse section, and its armature ; also when leaves are men- 

 tioned without any distinction, those found on that stem are 

 intended. In the description of those leaves attention should be 

 paid to the stalked or sessile state of the leaflets, especially the 

 lower or outer pair; the form of the terminal leaflet, all parts 

 of it being considered ; and the relative length of it and its partial 

 petiole ; and the character of its toothing. The form and structure 

 of the panicle is also very important; the form and direction at 

 different stages of the sepals and their armature ; the length and 

 direction of the stamens relatively to the pistils and their colour, 

 and that of the petals. Colour is usually considered by botanists 

 to be of very little value, but it seems to be important and often 

 quite permanent in many Ruhi. 



There is also another point concerning which I know very little, 

 which our great masters in this study consider of value ; I mean 

 the presence or absence of hairs on the young germens. It will be 

 seen that many of these things can only be observed on the living 

 plant ; it is therefore most important that they should be noted at 

 the time when the specimen is collected. The want of this care on 

 the part of collectors has caused exceeding difficulty in correctly 

 naming many of their specimens which may be in most other 

 respects well preserved. 



It has been well remarked by Weddell {Ann. Sci. Nat. s6r. 6, 

 ii. 356) that, "Except in a very few cases, it is impossible to 

 distinguish exactly one species from its neighbours by one single 

 character alone." This is the case in all groups where the species 



