V 



■^ , 



66 MB. BENTHAM ON LOGANIACEJE. 



connected Detioles. As a eenus it differs from them essentially 



m 



length 



two cells separating in the first instance. On this account 



mtim 



the sBstivation supposed to he valvate. It proves, however, to he 

 as strongly imbricate as in Logania. The habit of the plant is 

 unlike that of anv of the allied genera, and reminds rather of a 



Minna/rtia. 



In this resnect it also forms 



Gilia amonff Polemoniacetje 



a loculicidal dehiscence of the capsule, abnormal in Loganiaceae^ 

 but characteristic in Polemoniacece, would add the latter to the 

 number of families with which Polvioremvm mierht be associated, 



lumi 



without 



in Pol&moniaeece. The genus contains but a single species, for 

 the P. Schlechtendahlii of Walpers appears to be merely an acci- 

 dental form of the common one. 



9. Logania, Br, 



I have already shown that the affinities of Logania are rather 

 with the Scrophularinecd than with the MuhiacecB or Apocyne<B. 

 From Gomphostigma, hitherto placed in the former family, there 

 is indeed little to distinguish it but the usually pentamerous, not 

 tetramerous flowers, and even this character fails in the Logania 

 micrantha mentioned below. The same circumstance also separates 

 Logania from the other Buddleiece^ except that in Buddleia itself 

 some species have often an admixture of pentamerous flowers. In 

 these cases the toothed leaves, indumentum, and general habit are 

 very different from those of Loga/nia. 



The species of Logania are now rather numerous. In addition 

 to the twelve Australian ones, enumerated by DeCandolle, four 

 from Swan River have been described by Nees, one from South 

 Australia by Schlechtendahl, a very distinct one from subtropical 

 Australia by Hooker, and a doubtful garden one by Kunth and 

 Bouche. There are also one or two in our herbaria which appear 

 to be undescribed, but, without more numerous specimens in all 

 states of some of the commoner species, it is very difficult as yet 

 to make out a good monograph. Some species are evidently very 

 variable. There appears to be a regular gradation among the 

 Eastern ones, from L.florihunda to L. angtistifolia^ L, revoluta, and 

 L. linifolia ; so also between i. latifolia, L. longifoUa, and their 



^:v- 



^ /<^ 







'-.*■- i. 



^<-' . -\ 



• --- 



■•:'. 



■■,-# 





