578 THE FERTILISATION OF FLOWERS. [PART lv. 
only injurious so long as more specialised visitors are not abundant. 
enough to accomplish all the work of fertilisation ; but as soon as 
the development of special structures, to contain and to protect 
the honey, permits a larger accumulation thereof, the number of 
long-tongued visitors increases so much as to more than com- 
pensate for :the falling off in short-lpped visitors. Delay in 
the work of fertilisation, owing to concealment of the honey, is 
diminished by a great variety of contrivances, and sometimes 
entirely removed; the disadvantage that honey sheltered from 
rain is not directly visible, is alleviated by pathfinders (coloured 
spots or lines), which point towards the honey and enable the 
more intelligent visitors to find it in a moment; delay in obtain- 
ing deeply placed honey is lessened by the development of 
convenient standing-places, of apertures specially fitted for the 
insect’s head or proboscis (Labiates, Echiwm, etc.), and by the 
grouping of many flowers in close association, so as to permit 
rapid passage from flower to flower (Thymus, Mentha, Jasione). 
Finally, in Scabiose and Composite, we see how, by means of 
exserted reproductive organs and dichogamy, fertilisation en masse 
is rendered not incompatible with concealment of the honey. 
In the account that I have just given of the evolution of 
flowers with more or less deeply placed honey, I have stated that, 
in the first place, the advantages which led to such development 
were the protection of the honey from ‘rain, and the possibility of 
accumulating a large store; and that the exclusion of short-lipped 
insects was an advantage obtained indirectly. It may be sug- 
gested that possibly this statement should be reversed. But since 
the long proboscides of anthophilous insects and the long tubes of 
entomophilous flowers have been developed in mutual adaptation, 
and have advanced together to greater and greater perfection, we 
must suppose that the first honey-yielding flowers exposed their 
honey on flat surfaces, and that the first flower-visiting insects 
were only furnished with organs capable of licking up fully ex- 
posed honey. Under these circumstances elongation of the 
proboscis would be of no advantage to any insect so long as 
there was only freely exposed honey to lick; but shelter from rain 
and increased room for accumulating honey would be beneficial 
to the plant even before insects became divided into short-tongued 
and long-tongued. Therefore it is certain that the commencing 
evolution of structures to store and protect honey preceded the 
elongation of insects’ tongues. It was not until honey-receptacles 
and a more copious secretion of honey had been attained in certain 
"os ae 
