ERICACEAE. 



381 



impregnation would be more readily effected, were the whole disc 

 stigmatic, than it is, from the superposition of the cellular coat. 

 The application of the pollen sufficient to ensure fecundation, takes 

 place solely along the margin. In its effects it appears to be slow, 

 as, if we judge from the sphacelation sometime after the falling 

 of the perianth, the boyaux have not reached the foveola. The 

 perianth is deciduous, except the three outermost. 



The bases of the inflorescence are persistent, and each inflores- 

 cence is, folio floriferous, so that each axil of a sj^iniform leaf, is in- 

 definitely proliferous, a circumstance which does not usually obtain. 



Rhododendron. — Upwards of 12 species occur in Bootan between 

 Dewangiri and Tongsa, each has defined geographical limits, the 

 species appear in many cases to run one into the other, and I should 

 not be surprised, if, in a state of nature, many hybrids are produced. 

 Examination is particularly needful between the larger arboreous spe- 

 cies, and the lesser, which has not leaves argenteous underneath, al- 

 though a species with the flowers of this latter, has the under surface 

 of the leaves as in the larger. 



The inflorescence is compound, this is indicated by the upper flow- 

 ers opening^ first, and by the presence of two very narrow lateral 

 bractes on the pedicel. The parts of the flower vary, 5 being the 

 normal number; this, in the larger species is increased to 10. 



The corolla has as may sacs at its base, as there are petals. The 

 sacs contain honey. 



The anthers are very curious, not that the opening by a pore is 

 rare or singular, but the thickened margin of the pore would seem 

 to indicate that it is covered in its earlier stage by an operculum. 

 This is evidently not due to the almost total cohesion of the longi- 

 tudinal furrow. Other instances of porous opening must be ex- 

 amined. 



The pollen is compound, ternarily or quaternarily, and is rendered 

 adhesive by curious filaments. Each component part of the pollen is 

 furnished with one longitudinal furrow, furnished with a pore at its 

 centre. Of the nature of the intermixed filament, I much doubt ; 

 there is nothing in their appearance to lead one to suppose that they 

 are unabsorbed portions of the original poUiniferous mass. The ap- 

 pearance is precisely that of woody fibre. 



There is a tendency in the ovarium to a placentation, similar to 



