292 THE SENSORIAL FUNCTIONS. 



those of the mammalia, although they do not, as in the latter 

 class, open into the the respiratory organs, is shown by the 

 curious transformation of the one into the other during the 

 development of the tadpole, both of the frog and of the sa- 

 lamander. During the first periods of their existence, these 

 animals are perfectly aquatic, breathing water by means of 

 gills, and having all their organs formed on the model of the 

 fish. Their nasal cavities are not employed for respiration 

 at this early period, nor even for some time after they have 

 begun to take in air, which they do by the mouth, swallow- 

 ing it in small portions at a time, and afterwards throwing 

 it out in bubbles by the same channel. But when they quit 

 the water, and become land animals with pulmonary respi- 

 ration, the nostrils are the channels through which the air 

 is received and expelled ; and it is here also that the sense of 

 smell continues to be exercised. 



We know very little respecting the seat of the sense of 

 smell in any of the invertebrated animals, though it is very 

 evident that insects, in particular, enjoy this faculty in a 

 very high degree. Analogy would suggest the' spiracles as 

 the most probable seat of this sense, being the entrances to 

 the respiratory passages. This office has, however, been as- 

 signed by many to the antennae; while other entomologists 

 have supposed that the palpi are the real organs of smell.* 

 Experiments on this subject are attended with great diffi- 

 culty, and their results must generally be vague and incon- 

 clusive. Those which Mr. P. Huber made on bees, seem, 

 however, to establish, with tolerable certainty, that the spira- 

 cles are insensible to strong odours, such as that of oil of 

 turpentine, which is exceedingly offensive to all insects. It 

 was only when a fine camel-hair pencil containing this pun- 

 gent fluid was presented near the cavity of the mouth, above 

 the insection of the proboscis, that any visible effect was pro- 

 duced upon the insect, which then gave decisive indications 



* On the subject of this sense in insects, See Kirby and Spence's Introduc- 

 tion to Entomology, vol. iv. p. 249. 



