STURTION TRIBE. BALSAM TRIBE. 411 



spur will be found nearly or altogether deficient; whilst in 

 others it would be double or triple, the three being sometimes 

 united, sometimes divergent from each other. In one curious 

 specimen which came under the Author's notice, a second spur 

 grew from the first, not downwards, however, but upwards 

 into the flower, just as if we push back the finger of a glove, so 

 as to project into the part that receives the hand. The petals 

 are also more or less irregular, three being smaller than the rest, 

 and these being occasionally undeveloped. The stamens are 

 eight in number, and are perigynous or adherent to the calyx 

 (. 498). This, then, is an instance in which an exception 

 occurs to the character that has been taken as the chief guide in 

 classification ; for although the perigynous stamens, considered 

 in themselves, would cause the order to be removed into the next 

 class, yet its affinity with other Thalamiflorae, and especially 

 with Geraniacese, is so manifest, that it cannot be properly re- 

 moved from that group. The ovary is made up of three carpels, 

 adherent as in the last order, around a central elongated axis ; 

 these each contain a single seed. 



576. Another allied order is that of BALSAMINKE, the Balsam 

 tribe ; this has not, however, any of those medicinal virtues 

 which might be inferred from its name, but consists only of a 

 small number of plants, which are esteemed for the beauty of 

 their flowers and their elegance of aspect. Of these, one species 

 is a native of Britain, though it is rarely found wild ; and this 

 is remarkable for the curious manner in which its seed-vessel 

 opens (. 419). There is greater irregularity in the flower than 

 in either of the last two orders ; so that the nature of its parts 

 might not be understood at first sight. The sepals are five in 

 number, but unequal and irregular in form ; the two upper and 

 inner ones are adherent, while the lower one is spurred. There 

 are only four petals ; and these are adherent in pairs ; so that 

 the corolla appears as if it consisted of two only. The stamens 

 are five in number, and their position is regular ; so that, by 

 comparing their places with that of the petals, it becomes evident 

 that each of the apparently-single petals consists of two, and 

 that a fifth petal, which should occupy the space between them, 



EE2 



