4 BOTANY OF SOCOTRA. 



known only from Socotra and Somali Land. The genus is very closely allied 

 to Robert Brown's Notoceras from the Canary Islands, Mediterranean region 

 and Western Asia, from which indeed it is but doubtfully separated by th 

 more clothed habit and the colour of its flowers. 



D. incana, Balf. fil. in Proc. Roy. Soc. Edin. ix. (1882), 500. Tab. II. 



Notoceras sinuata, Franch. Sert. Somal. in Miss. BeVoil. 9. 



Herba incaua ; foliis oblongis v. ovatis, obtusis, repandis ; racemis laxis elongatis ; floribus 

 niagnis ; siliquis tetragonis. 



Erecta basi suffruticosa incana. Folia oblonga v. ovata |-1| poll, longa, ^— | poll, lata, basi 

 cuneata inpetiolumbrevemattenuata,apice obtusa v. subacuta, repanda dentata v. dentato- 

 serrata, subfloccosa. Bacemi laxi elongati ssepe pedales. Alabastri oblongi. Flores magni % 

 poll, longi. Sepala margine scariosa, lateralia basi saccata. Petala erythrina dimidio sepala 

 superantia. Siliqua tetragona vix torulosa |-| poll, longa floccosa, cornuis stylo dimidio 

 lorgioribus. 



Socotra. On sandy spots of the plains about Galonsir. B.C.S. nn. 136, 

 161. 



Distrib. Somali Land. 



A very beautiful new species of this genus, of which D. floccosa, Boiss. (Flor. 

 Orient, i. 315), andZ>. canescens, Boiss. (loc. cit.) are the other members. These 

 were brought from Persia by Aucher Eloy. 



The Socotran plant is very distinct, resembling most D. floccosa, from which, 

 however, its larger leaves, long lax racemes, large flowers with saccate lateral 

 sepals, as also its tetragonous siliqua separate it. These last two characters 

 are points in which the species varies from the generic type as given by 

 Bentham and Hooker {loc. cit.), where I find "sepala basi sequalia" "siliqua 

 teretiuscula torulosa." The specimens of Boissier's species to which I have 

 had access are somewhat fragmentary, and do not afford flowers for making a 

 satisfactory comparison, but the fruit of I), canescens, which is described by 

 him as subtetragonous, shows an approach to the marked tetragonous con- 

 dition which is so conspicuous a feature in the Socotran plant. 



Franchet, ignorant of my diagnosis of the species as found in Socotra, 

 described (loc. cit.) ReVoil's Somali Land specimens under the genus Notoceras. 

 I have in Paris, through the kindness of M. Franchet, examined ReVoil's plant, 

 and find it is the same as the Socotran one. 



2. FARSETIA. 



Farsetia, Desv. Journ. Bot. iii. (1814), 173 ; Benth. et Hook. Gen. PI. i. 72. 



A genus of about twenty-five species of whitish woolly herbs or half-shrubby 

 plants, characteristic of the dry plains of the circum- Mediterranean region, 

 Arabia, Persia, and north-west India. One species appears endemic in 

 Socotra. 



