204 THE SCENT OF FLOWERS AS A MEANS OF ATTRACTING ANIMALS. 



occurring in different species of Syringa, Tilia, and Samhucus can also be easily 

 recognized. In Eoses this phenomenon is even more remarkable. From their scents 

 Rosa alpina, pimpinelli folia, arvensis, Indica, moschata, canina, Gallica, cinna- 

 momea, Centifolia, and Thea can be at once distinguished with closed eyes by 

 anyone who has examined the numerous species of this genus even to a limited 

 extent. It is also remarkable that in closely-allied species the flowers of one will 

 smell while those of another will be scentless. Habenaria montana has no scent, 

 whilst Habenaria bifolia exhales a strong scent of cloves. Viola tricolor is scent- 

 less, Viola polychroma develops a strong violet scent. The flowers of Primula 

 Lehmanni have no scent, while those of Primula Auricula, which can hardly be 

 distinguished from the former, have a strong Auricula smell. These facts are not 

 without bearing in the theory of specific constitution of protoplasm, as will be 

 discussed later on in the chapter on the Origin of Species, and therefore should be 

 noted here in passing. 



We are liable to many erroneous inferences with regard to the perception of 

 flower scent by animals, since our judgment depends mainly on our owm sense of 

 smell, and it is very possible, even pi'obable, that the power of smell in flower- 

 visiting animals differs materially from ours. The olfactory sense of man is lodged 

 in a sharply-defined portion of mucous membrane in the upper part of the nasal 

 cavity. There the superficial cells of the mucous membrane join with the end 

 branches of the olfactory nerve in a peculiar net-work, and the scents must act 

 directly on this region if they are to produce the sensation of smell. But this is 

 again only possible if the odorous substances give off" fine particles into the air, 

 and if this impregnated air is wafted over the special part of the nasal mucous 

 membrane. It was formerly held that the substances passing thus over the 

 olfactory mucous membrane were dissolved in a fluid and were then distributed in 

 solution. Only in this way could they influence the nerve-endings. But this view 

 is contradicted by a series of facts of which the most important are the following: 

 it is well known that we can smell certain metals whose finely-divided particles 

 break away and enter the nose, although these metals are certainly not soluble in 

 the mucous membrane. We are also able to smell very different scents quickly, 

 one after the other, which would not be the case if the sense of smell were 

 dependent on a previous solution of the odorous substance in the fluid which 

 saturates the mucous membrane. Again it is a remarkable fact that the mucous 

 membrane is altogether absent from the olfactory organ of many animals. The 

 knobs and pegs on the surface of the feelers which form the olfactory organs of 

 insects are indeed connected on one side with gangliose nerve-endings, but they 

 have nothing resembling a mucous membrane which could contain or secrete a 

 fluid, and yet insects are characterized by their fine sense of smell. 



The stimulation of tlie nerve-endings in the olfactory organ cannot therefore 

 be the result of a previous solution of the odoriferous substance, but must be 

 considered as the transference of a movement. It seems as if the molecules of the 

 odorous substance which are present in the air undergo a rotatory, vibrating, or 



