190 DISSERTATION SECOND. [part j. 



deduced from it. It is indeed quite evident, that, indepen- 

 dently of experiment, Descartes himself could have put no 

 trust in it, and it is impossible not to feel, how much more 

 it would have been for the credit of that philosopher, to 

 have fairly confessed that the knowledge of the law was 

 from experiment, and that the business of theory was t» 

 deduce from thence some inferences, with respect to the 

 constitution of light and of transparent bodies. This I 

 conceive to be the true method of philosophizing, but it is 

 the reverse of that which Descartes pursued on all occa- 

 sions. 



The weakness of his reasoning was perceived and at- 

 tacked by Fermat, who, at the same time, was not very 

 fortunate in the theory which he proposed to substitute for 

 that of his rival. The latter had laid it down as certain, 

 that light, of which he supposed the velocity infinite, or 

 the propagation instantaneous, meets with less obstruction 

 in dense than in rare bodies, for which reason, it is refract- 

 ed toward the perpendicular, in passing from the latter 

 into the former. This seemed to Fermat a very improba- 

 ble supposition, and he conceived the contrary to be true, 

 viz. that light in rare bodies has less obstruction, and 

 moves with greater velocity than in dense bodies. On 

 this supposition, and appealing, not to physical, but to final 

 causes, Fermat imagined to himself that he could deduce 

 the true law of refraction. He conceived it to be a fact, 

 that light moves always between two points, so as to go 

 from the one to the other in the least time possible. Hence, 

 in order to pass from a given point in a rarer medium where 

 it moves faster, to a given point in a denser medium where 

 it moves slower, so that the time may be a minimum, it 

 must continue longer in the former medium than if it held 

 a rectilineal course, and the bending of its path, on entering 

 the latter, will therefore be toward the perpendicular. On 



r 

 * 



