426 FETCH : 



face of the latter streaked with, or almost uniformly coloured, 

 purplish-red. 



There are three bracts to each flower. The largest one, 

 beneath the flower, is triangular, and measures about 14 mm. 

 in length and 6 mm. in breadth ; it falls off before the flower 

 opens. The other two are strap-shaped, tapering towards the 

 tip, about 9 mm. long and 4 mm, broad, and are situated 

 laterally, closely overlapping the bud and overlying the line 

 of separation of the two sepals ; they fall off when the flower 

 opens. 



On page 161 Miers states that the sepals of Bertholletia are 

 notched at thek apex by three small teeth. In his descriptions 

 he writes that excelsa has the sepals tridentate, while in nohilis 

 they are obsolutely crenulate, and he further states that 

 nohilis differs from excelsa in its rounder and more entire 

 calycine lobes ; but in his figures of B. nohilis (plate 33, fig. 3) 

 he shows the calyx with three fairly strongly developed teeth, 

 and describes it, in his explanation of plates, as the calyx 

 which splits into two semiglobular segments, each tridentate 

 at the apex. Exammation of the fresh specimens shows that 

 this feature is a variable and accidental one. The globular 

 calyx spHts into two hemispherical sepals, whose margin is 

 usually quite entire, but sometimes obscurely three-toothed. 

 The teeth, however, are formed by the splitting of the sepal as 

 the flower opens, and this spHtting is accentuated as the sepal 

 dries. Even sepals which are entire when fresh may become 

 obscurely tridentate on drying. I have not, however, found 

 sepals so markedly dentate as in Miers' figure, though I have 

 seen them split with a single fissure halfway down to the 

 base. The degree of fission at the apex of the sepal 

 would probably depend on the degree of expansion of 

 the flower. The flowers of the Ceylon trees do not expand 

 wo widely as shown in Miers' figure ; indeed, the petals, 

 tliougli recurved at the apex, are so closely applied to the 

 andro])horum that it would appear that only self-fertilization 

 can occur. It would seem probable, however, from Miers' 

 description, that his figure of the calyx is that of B. excelsa. 

 Tliere is no difference between the calyces from tlie two 

 Ceylon trees. 



