Sert.-Oct., 1918. } THE ORCHID REVIEW. 195 
accompanied by the perfectly gratuitous assumption that they arose directly 
ata single operation. The question is only one out of over a dozen others, 
which are grouped under the following general enquiry: ‘‘ How did the 
following commence, and what advantage was gained by the rudimentary 
beginning, sufficient to insure the completion of the organ in question? .. . 
where does Natural Selection come in as formative or explanatory?’ The 
pollinia of Orchids are not mentioned, but they are well adapted to answer 
the enquiry, because the stages of progress are so well marked that one may 
almost trace their development step by step in the different groups. 
The pollinia of Orchids are very diversified in character, and in many 
cases exceedingly complex, but the character that they possess in common 
is that they are formed of compound pollen grains, grains that are united 
together in fours. In the highly specialised tribe Vandez the pollinary 
apparatus consists of a combination of pollinia, stipes and gland, the two 
latter derived from the rostellum, but in Epidendrez, Neottiew and Ophrydeze 
the rostellum does not produce a separable stipes, though bodies somewhat 
similar in appearance, known as caudicles, are in many cases produced as an 
outgrowth from the pollinia. The significance of the difference will be seen 
Presently. In proceding back to the more primitive Cypripediez we find that 
neither rostellum nor pollinia are developed, the pollen grains being simple, 
though cohering in a sticky mass through their own viscidity. Finally, in 
the ancestral Apostasiez the absence of a rostellum is accompanied by the 
pollen grains being both simple and not cohering together, so that the 
Pollinary condition of an ordinary Monocotyledon is reached. 
ome authors regard the Apostasiz as forming a distinct Natural Order, 
but this view would also exclude the Cypripediew, which are very similar 
lM general structure. In fact Bentham unites the two, but it is better to 
regard them as distinct tribes of the sub-order Diandre, which is distinguished 
ftom Monandre by the different staminal arrangement. The point, however, 
'S immaterial for the purpose of this enquiry, for their structure and 
telation to each other remain, whatever we may call them. It may make 
the point clearer if we say that there is a large natural group of 
Monocotyledons, called _Microsperme, from the organisation of its 
humerous minute seeds, and that this is divided into two Natural Orders, 
Burmanniacez, with regular flowers, and Orchidacez, in which they are 
highly irregular. The latter is again divided into two sub-orders, Diandre, 
two or occasionally three stamens, simple pollen grains, and no 
vellum, and Monandre, with a single stamen, compound pollen grains, 
Cohering into pollinia, and a rostellum. These groups are perfectly natural, 
and the characters given indicate their evolutionary progress. 
We May attribute the origin of the Orchidacez to the modification of 
the flowers as an adaption to fertilisation by insects, two stamens of the 
