PEOF. H. M. WAED AND MISS E, DALE ON CEATEROSTIGMA PUMILTJM. 345 



Craterostigma, founded by Hochstetter (' Flora,' Jahrg. xxiv. p. 668) in 1841 on the two 

 AbyssiDian plants C. plantagineum and C. pumilum, and it is with the latter that our 

 plant agrees. 



Of the three or four species comprised in Cixitei'ostigma, only one has the flowers 

 solitary, viz. C. pumilum, Hochst., and it comes from Abyssinia or Somali-Land. 



On reference to the Kew Herbarium we found that Craterostigma pumilum, Hochst., 

 has inflorescences which are usually branched and bear about three to five flowers with 

 bracts — an umbellate cyme apparently ; but on the same sheet is a specimen with single 

 flowers marked Torenia auriculcpfolia, Dombr., and the closest examination shows no 

 other differences between it and the several-flowered Craterostigma pumilum^ Hochst., 

 and leads to the conviction that the Kew authorities are right in merging Dombrain's 

 plant in Hochstetter's species. Moreover, there can be no doubt that Dombrain's plant 

 is the same as ours, especially after consulting his figure in ' The Floral Magazine,' 1871, 

 p. 534 ; but as he gives no botanical description and no figure or mention of the red 

 roots or details, we tliiDk it worth while to off'er our results. 



On the same Kew sheet also there are some multi-flowered specimens collected by 

 Mrs. Lort Phillips, and all the specimens on this sheet show by the red stains that the 

 roots were coloured as in our specimen. The same is true of C. plantagineum, Hochst., 

 a much more hairy form, evidently very close to C. pumilum, and coming from the same 

 region. The dried herbarium specimens of C. pumilum also look more hairy than our 

 fresh plant. 



The question arises whether the single-flowered (auriculcBfoUum) form is a constant 

 variety of C. pumilum, or merely a more poorly-grown plant : some of the Kew 

 specimens with several flowers are as small as the one-flowered form. We have been 

 unable to detect any bracts on the inflorescence of our Cambridge plant, and Dombrain's 

 figure shows no traces of them. 



Internal Structure. 



Hoot. — A transverse section through the root (PI. XXXIV. fig. 15) shows the usual 

 structure of a central stele surrounded by cortex. The cortical cells are arranged 

 M'ith great regularity in radial rows, and between them are equally symmetrical inter- 

 cellular spaces, which, in section, are of quadrangular form, and contain a peculiar red 

 colouring-matter. The spaces are in many cases almost as large as the cells, and 

 generally occur throughout the cortex, but are absent from the piKferous layer and the 

 cells immediately internal to it. 



When highly magnified (PI. XXXV. fig. 16) the spaces are seen to be due to the 

 splitting of the middle lamella at the points where four cells would be in contact. Where 

 the walls of adjacent cells are united one to another they are flat, but where they Hmit a 

 space they project into it, so that, in transverse section, the cells approximate to the form 

 of a square with rounded angles, while the spaces are quadrangular with convex walls. 

 On account of the regularity of the arrangement of the cells, the spaces between them 

 are also very uniform in shape and size. 



3r2 



