OF NATURAL HISTORY. 117 



otFier, to the top, contending which of them fhall fooneft receive 

 the benefit of the air *. In the fhallow parts of rivers, when fro- 

 zen, many fifhes are found dead. But, when parts of a river are 

 deep or rapid, the fi£hes fly from the ice, and by this means avoid 

 deftrudion, 



Thefe, and fimilar experiments, have been repeated by Mr Wil- 

 loughby, and many other modern authors ; and they have uniform- 

 ly been attended with the fame event. A carp, in a large vefTel full 

 of water, was placed in the receiver of an air-pump. In proportion 

 as the air was exhaufted by working the pump, the furface of the 

 animal's body was covered with a number of bubbles. The carp 

 foon breathed quicker, and with more difficulty : A little after, it 

 rofe to the furface in queft of air. The bubbles on its furface next 

 difappeared ; the belly, which before was greatly fwollen, fuddenly 

 collapfedj and the animal funk to the bottom, and expired ia con- 

 vulfions. 



Thus the refpiration of air is as necelTary to the exiftence of 

 fifhes as to that of land-animals; for none of them can live long 

 when deprived of this vivifying element. Filhes, indeed, feem to 

 require a fmaller quantity of air than animals who have a conftant 

 and free communication with the atmofphere. The bodies and 

 fluids of fifhes are colder than thofe of land-animals; and, of courfe, 

 if Do£tor Crawford's theory be well founded, fifhes require lefs air 

 to fupport the proportionally fmall quantity of heat they poiTefs. 



An analogy between fifhes and birds deferves here to be noticed. 

 Both of thefe clalTes of animals are rapid in their motions; and both 

 of them, befide refpiring by lungs or gills, have receptacles of air- 

 within their bodies, Fiflies tranfmit fmall quantities of air through 



their 



* Rondeletlus, lib. 4. cap. rj. 



