



14 MB. G. BENTHA.M ON THE 



forniati genus (Githopsis) , the origin and connexions of which ar 

 as yet very problematical. It is a small annual of limited area, 

 with the capsule operculate, as in some of the above-mentioned 

 South-African Campanuleae, with which it cannot be supposed to 

 have any connexion. The character, however, is exemplified in 

 West America in some Lobeliese — a circumstance which, taken 

 together with the habit, might lead us to suppose that notwith- 

 standing its free anthers, GitTiopsis is rather derived from the 

 Lobelieae than from the Campanuleae, with which it is technically 

 classed. 



2. Oleaceae and Salvadoraceae. 



Following the suggestions of Endlicher and others, we have re- 

 united the Jasminese and Oleineae, so many of the supposed di- 

 stinctive characters having become invalidated by further observa- 

 tion ; and we long hesitated whether the Salvadoraceae should not 

 also have been brought in. The whole together form an isolated 

 and well-defined group, divisible into five, of nearly, though not 

 quite, equal value — Jasmineae, Syringeae, Fraxineae, Oleinese, and 

 Salvadoraceae. The connecting link of the whole, independently 

 of a number of minor less-constant characters, is the binary ar- 

 rangement of the stamens in continuation of that of the carpels, 

 and independent of that of the corolline lobes. This is strikingly 

 exemplified in the Jasminese and in Schrebera, where the corolline 

 lobes vary from the ordinary four to five, six, or more. Lindley ob- 

 serves, indeed (Veg. Kingd. 618), that the two stamens of Jasmi- 

 nese are probably connected with a quinary type ; but that does 

 not prove to be the case. In the Personate orders, where the 

 stamens are reduced to two, it is always by the abortion of the 

 remainder of the whorl, of which the rudiments may most fre- 

 quently be traced; and the two perfect ones are always alter- 

 nate with the lobes of the corolla. In Jasminese and in Schrebera 

 I have invariably found them, as in other Oleaceae, alternating 

 with the carpels and bearing no relation to the corolla-lobes. This 

 is further proved by two or three flowers in which I found 



car 



any 



[•responding difference in the corolla-lobes. In the Sal- 

 and in the very few species of Oleaceae (about half a 



280) 



true, alternate 



