24 Transactions. 



formation involves a loss to the flower which is usually made up 

 by suppression of other parts. In Cruciferff there are, for 

 instance, only six stamens, but a close study of the flowers 

 renders it probable that the six nectaries are remains of six 

 other stamens which are required to make up the symmetry of 

 the flower. Generally a direct connection between nectaries and 

 missing parts cannot be traced, but well-developed nectaries, as 

 in the higher Corolliflorcp, usually go along with great reduction 

 in stamens and carpels. One has only to compare the numerous 

 stamens and carpels of buttercups, roses, and mallows with the 

 four or two stamens and two carpels of most CoroUifhrcf to see 

 this clearly. 



Very often a regular nectary leads to a kind of bag being 

 formed to hold the lioney secreted. A rudimentary cup of this 

 kind occurs in the pouched sepals of some Cruciferre ; here it has 

 probably been formed by the impression of the nectaries which in 

 the buds occupy the part of the sepals afterwai-ds pouched, and 

 the cavity thus formed being useful as a honey receptacle, has 

 been maintained and improved by selection. In otiier cases one 

 is disposed to think that insects have directly started a vai'iation 

 of this kind. Thus some buttercups are very near the St. John's 

 Wort in being without any particular spot at which honey is 

 secreted. In the common butter cup, however, exudation of 

 lioney is confined to a particular shallow pit just at the base of 

 the petal. If one compares this shallow cavity with a Columbine 

 petal, one can scarcely resist the conclusion that the constant 

 pushing and probing of insects lias deepened and elongated this 

 shallow pit till it has come to form the long curved spur of 

 the Columbine. The lengthening of this pit may have gone on 

 for centuries, but as every little increase was of advantage both 

 to flower and insect it seems very possible that it has been pro- 

 duced in this way. Quite similar pits or spurs are found in the 

 Toadflax Butterwort and in violets, and similar structures are 

 extremely common in the orchid family. One orchid, Angru'cniii 

 sp.squipedale, has a spur of this kind eighteen inches long, which 

 is adapted to a kind of hawk moth which has a proboscis aljout 

 the same length. 



There would be no difiiculty in multiplying instances to show 

 the influence of insect visitors. I shall, however, forbear to tiy 

 the patience of members any longer, only pointing nut wjiat is 



