SELF-FERTILIZATION IN TROPICAL ORCHIDS. 545 



the ovarium. Now if we were to suppose the ridged rostellum of 

 Arundina to become adherent to the back of the stigma instead 

 of hanging down loose, we should have such a ridge as is seen in 

 Phaius Blumei ; so that it would seem that this marked ridge of 

 the Phaius is a remnant of a rostellum, which has become modi- 

 fied and adherent to the back of the stigma. 



Eria albido-tomentosa, Lindl., is an epiphytal Orchid, growing at 

 a height of about 2000 feet above the sea. It is singular in having 

 the flower-stalk and the external parts of the perianth densely 

 covered with a felty mass of white wool. It is rather dull and 

 unattractive as a flower. Its anther is situated above the stigma, 

 is separated from it by a narrow rim-like rostellum, and contains 

 two cells. The stigma is broad and rather shallow ; its viscid 

 fluid is gummy and not very liquid. 



The following are the general appearances in the bud : — The 

 pollinia are arranged in two cells. The viscid disk of the pollinia 

 (really composed of pollen-grains) occurs high up above the 

 level of the tip of the rostellum. In looking at the column 

 in front, only two of the masses in each cell can be seen ; the 

 other two, projecting behind, coalesce at the viscid disk of the 

 pollinia. The process of fertilization seems to be : — A swelling 

 of the viscid matter of the stigma takes place, the anther is more 

 or less inundated through the channels at the side of the rostellum. 

 The outermost pollen-mass in each cell is the first to be touched 

 by the stigmatic viscid matter, as they lie just above the channels 

 by which the stigmatic fluid passes into the anther. The higher 

 pollinia are generally untouched by the viscid matter of the stigma, 

 and after the fertilization of the flower, the lower masses come 

 forward and project over the other two. The pollen-masses them- 

 selves are short and somewhat spathulate, and lie against each 

 other by their flat sides. By the touch of a pencil at the spot 

 where the masses converge together, three pollinia, i.e. the two 

 hinder and the inner one of each cell, come away early ; but 

 the exterior one remains glued, but not firmly (I am describing 

 a bud), to the bottom of the anther, and is removable with 

 some manipulation. The outer pollinia, on becoming smeared 

 with the viscid mass, emit the pollen-tubes, which go down 

 through the stigma as two prominent ridges, which tend to 

 obliterate the rostellum. The flower-bud now expands ; for the 

 operations above described have generally taken place before the 

 opening of the bud. The viscid disk gradually loses its viscidity, 



