92 Oliver. — 'On Physostoma elegans , Williamson , 
come under observation came from the collection of Mr. D. M. S. Watson — 
prepared by his own hands. To him I am indebted for the opportunity 
of examining the two longitudinal sections, one median, the other tan- 
gential. The former shows seventeen grains in the pollen-chamber, the 
latter thirty (see Text-figs. 6 and 7). These two sections alone thus retain 
close on fifty pollen-grains ; but the total number present must have been 
far in excess of this, when it is borne in mind how much of the pollen- 
chamber has been lost in the operations of cutting and grinding. 
The questions at once arise : How did the pollen-grains get there ? 
What was the agent of transport, and what were the arrangements at the 
orifice of the seed for the reception of pollen ? As to the last, perhaps we 
may accept the position of the tentacles shown in the accompanying 
illustration (Text-fig. 7), based on the median longitudinal section in 
Mr. Watson’s possession. In that case the tentacles behaved collectively 
Text-fig. 6. Tangential section 
through the pollen-chamber showing 
about thirty pollen-grains. Sketched 
from Mr. Watson’s preparation, x 50. 
Text-fig. 7. Sketch of Mr. Watson’s 
median section through the pollen-chamber 
of Physostoma , showing two arms approxi- 
mated above the orifice. The dotted lines 
below the sinus mark where the integument 
has broken away from the nucellus. x 20. 
as a closed tube, a micropyle in the making. No doubt a drop excreted 
from the pollen-chamber played an essential part in the mechanism, as in 
recent Gymnosperms. That the tentacles formed a close-fitting tube 
around the pollen-chamber and its orifice is also consistent with the rarity 
of pollen in the sinus at the base of the plinth. 1 Whether the ‘ micropyle ’ 
was occupied by a tubular extension of the pollen-chamber — which thus 
gained direct access to the exterior— remains uncertain. No traces of any 
such passage have been detected, but in view of the fact that the pollen- 
chambers of both L. Lomaxii and L . ovoides had direct access to the 
1 Notwithstanding the notable exception described on p. 94. 
