53 
twice the size of R. hirsutus, their dimensions being respec- 
tively o and ah of an inch. 
I have not had the time to make similar careful measure- 
ments with the micrometer, but I have seen sufficient to be 
satisfied that while there is considerable variation in dimen- 
sions between the pollen of one species and that of another, 
they are tolerably constant in size in the same species. 
For some noticeable differences compare the smaller pollen 
of Epilobium brachycavpum with the larger pollen of E. 
Fleischeri or that of Senecio gallicus with S. incanus, the 
spines on the latter species being also much coarser. Again, 
the pollen of Silene acaulis is but half the size of that of 
8. alpina, the latter having some beautiful markings in 
addition; the pollen grains of this genus differ from the 
usual caryophyllaceous type in not having the pits or 
depressions common in the order, so that the grains become 
spherical rather than polyhedral. 
4. Colour. This is not so reliable a character for differ- 
entiation as the others noticed, since species differ amongst 
each other according to the soil, &c., of the place where they 
have grown. I remember gathering some years ago, near 
Ashbourne, Derbyshire, a variety of Stellar la Holostea 
having a dark purple pollen instead of the ordinary pale 
yellow. An example or two under this head will suffice. 
The pollen of Ajuga genevensis is yellow, but that of 
A. pyramidalis is usually white ; again, while the grains 
of Ornithogalum umbellatum are large and yellow, those 
of 0. nutans are small and white. 
Some objection may be raised to any reliance being 
placed upon the dry shrivelled-up grains of herbaria speci- 
mens — such specimens being in most cases the only ones 
obtainable for purposes of investigation ; but the structure 
of pollen is such as to bring into greater prominence the 
pores, folds, valves, and other markings which are met with 
on their surface after the grains have collapsed by the dis- 
charge of their contents. 
