A REVISION OF THE GENUS BERTIERA 159 



In nearly all the specimens seen the branchlets are covered 

 with a relatively dense light-brown strigose indumentum ; but 

 the presence of this, unfortunately, does not appear to be a safe 

 criterion — at least, not in the dried state —as the specimens from 

 Peru and Panama are almost perfectly glabrous. Part of the 

 latter, moreover, is glabrous and part clothed similarly to the 

 typical B. cjuianensis, and both parts were gathered in all proba- 

 bility from the same plant. In these circumstances it becomes 

 necessary to distinguish the present species from B. gonzaleoides ; 

 this has been attempted under the head of that species (q.v.). 



It will be seen that B. guianensis is unrecorded from the West 

 Indian area. There is one sheet in the Kew herbarium bearing a 

 queried label: — "Jamaica, Dr. Dancer, Herb. Forsyth"; but this 

 is apparently an error, and Mr. Fawcett, late Director of Public 

 Gardens and Plantations, Jamaica, assures me that the genus 

 Bertiera is not, so far as he knows, included in the flora of that 

 island. 



13. B. LONGiTHYRSA Baker in Journ. Linn. Soc. xxv. 322 

 (1890). 



Hab. Madagascar : Baron, next 5789 ! Hildehrandt, 3002 ! 

 Hbb. Mus. Brit, and Kew ; Perville, 759 ! probably should be 

 assigned to this species. 



Notable for its conspicuously long, subfoliaceous bracteoles. 



14. B. FERA A. Kich. in Mem. Soc. Hist. Nat. Paris, v. 254 

 (1829). B. rufa DC. Prod. iv. 392 ; De Cordemoy, Fl. de la 

 E6un. 504. 



Hab. He de la Reunion : de I'Isle, 7676 ! Balfour ! Hb. Kew. 



The name rufa in the Prodromus appears to be the result of a 

 clerical error, as the authority there cited is " A. Rich, in mem. 

 soc. h. n. Par. v. 5, p. 254," from which de Candolle's description 

 is, essentially, transcribed ; and in Richard's memoir the name of 

 the same species appears as B. fera. This error persists even in 

 de Cordem.oy's flora (loc. cit.). 



The species is readily distinguishable by the dense ferruginous 

 indumentum, which recalls that of B. africana, and by the rela- 

 tively long (2 mm. or more) calyx-lobes, erect in the fruit. 



15. B. BORBONiCA A. Rich, in Mem. Soc. Hist. Nat. Paris, v. 

 254 (1829) ; DC. Prod. iv. 392. 



Hab. La Reunion : M. Richard ! Carmichael, 271 ! de I'Isle, 

 909 ! Hbb. Mus. Brit., Kew, and Univ. Cambridge. Mauritius : 

 Boutonl Hb. Univ. Cambridge. 



De Cordemoy (Fl. de la Reun. 504) regards this as identical 

 with the Mauritian B. Zaluzania, but it appears, from the material 

 available in the English herbaria, to be quite distinct, in the 

 character of the erect fruiting calyx-lobes, and in its much more 

 diffuse and widely branched inflorescence and smaller flowers 

 very acuminate in the bud. The anthers, moreover, are linear in 

 this species ; in B. Zaluzania they are shorter and broader, 

 almost ovate. 



16. B. BiSTiPULATA Boj. RamuHs foliisque glabratis, stipulis 



