ON THE CAULOTAXIS OF BRITISH GERANIUMS. 299 



branch in its axil. There is little difference, however, between 

 these branches as regards vigour of growth, and as they diverge at 

 nearly equal angles from the central axis, especially at the lower 

 nodes, the caulotaxis is practically a dichasium. In some cases 

 the lower nodes are destitute of a central prolongation of the 

 internode below, but the middle and higher ones usually have 

 a two-flowered peduncle in the fork between the lateral branches. 

 When the central peduncle is absent, a reminiscence of it usually 

 exists in an undeveloped central bud. 



G. sylvaticum, G. p?/?'e7«rucii»?., and G. rotundifolium have a 

 caulotaxis which fluctuates a little, but in the main agrees with 

 that of G. jyratense, and approaches more or less closely to a true 

 dichasium. The departure from this, which these species some- 

 times present, alwaj^s shows itself in a tendency towards the 

 formation of a pseud-axis, owing to a slight accentuation of the 

 differences between the lateral shoots produced at each node. 



G. dissectum conforms to the type of G. Rohertianum, The 

 central peduncle is usually present at every node, both lateral 

 branches are developed, and a distinct pseud-axis is organised. 

 Of the same type is the caulotaxis of G. hicidum. 



In G. sanguineum and G. columhinum the caulotaxis exhibits a 

 higher degree of divergence from the dichasial type than even 

 G. Roberiianum. The axillary branches at the lower nodes are 

 both developed, — one stouter than the other and taking part in the 

 formation of a well-marked pseud-axis, — but at the upper nodes one 

 branch often remains undeveloped, so that the peduncles at these 

 points often appear to be axillary. In most cases hitherto 

 examined, however, a bud has been found to occupy the position 

 of the absent branch. 



But the most peculiar form assumed by the caulotaxis of the 

 Geraniums is that presented by G. molle. In this species there is 

 only one leaf at the majority of the nodes, and the two-flowered 

 peduncles are thus generally leaf-opposed — an arrangement which 

 would be very perplexing did this species stand alone. If, however, 

 its caulotaxis be interpreted by the light of the facts brought 

 forward in what has preceded, it will hardly seem an unwarrantable 

 hypothesis that in G. molle the peduncles are terminal as in all the 

 other species, and that at each node one leaf, with its axillary 

 branch, has been completely suppressed ; in other words, that the 

 caulotaxis is of that form which has been designated a cicinal 

 cyme. 



G. imsillum appears to be rather more variable than the other 

 species, sometimes approaching to the type of G. columhinum, and 

 at others to that of G. molle, especially at the higher nodes. 



Taking then the whole genus of British Geraniums, it will be 

 seen that four types of caulotaxis are distinguishable, viz., those 

 of G. xjratense, G. Rohertianum, G. columhinum or sanguineum, and 

 G. molle respectively, No doubt these types pass more or less 

 gradually hito one another, and individual species do not in- 

 variably and entirely conform to one type. But in the great 

 majority of specimens examined during the preparation of this 



