NATURAL PHILOSOPHY. 125 



which the pressure or shock of the water might produce. Therefore, 

 the two tubes, and the reservoirs in which the motion of the water 

 was made, were sustained by supports independent of the apparatus, 

 and especially of the two lunettes ; it was, therefore, only the two tubes 

 which could suffer any accidental movement ; but both theory and 

 practice have proved that the motion or flexions of the tubes alone 

 were without influence upon the position of the fringes. The follow- 

 ing are the results obtained. 



When the water is set in motion the fringes are displaced, and, 

 according as the water moves in one direction or the other, the dis- 

 placement takes place towards the right or the left. The fringes are dis- 

 placed towards the right when the water is running from the observer, 

 in the tube situated to his right, and towards the observer in the 

 tube situated to his left. The fringes are displaced towards the left 

 when the direction of the current in each tube takes place in a direc- 

 tion opposed to that which has just been described. With a velocity 

 of water ecjual to two metres per second, the displacement is already 

 very sensible ; with a velocity of four to seven metres it is perfectly 

 measurable. 



After having demonstrated the existence of the phenomenon, I 

 endeavored to determine its numerical value with all the exactitude 

 which it was possible to obtain. By" calling that the simple displace- 

 ment which was produced when the water, at rest in the commence- 

 ment, was set in motion, and that the double displacement which was 

 produced when the motion was changed to a contrary direction, it was 

 found that the average deduced from nineteen observations sufficiently 

 concurring, was 0.23 for the simple displacement, which gives 0.4G for 

 the double displacement, the width of a fringe to be taken as unity. 

 The velocity of the water was 7.00 metres per second. 



The results were afterwards compared with those which have been 

 deduced by calculation from the different hypotheses relative to the 

 tether. According to the supposition that the gather is entirely free 

 and independent of the motion of bodies, the displacement ought to be 

 null. 



According to the hypothesis which considers the Eether united to 

 the molecules of matter in such a way as to participate in its motions, 

 calculation gives for the double displacement the value 0.92. Experi- 

 ment gave a number only half as great, or 0.46. 



According to the hypothesis by which the aether is partially carried 

 along, the hypothesis of Fresnel, calculation gives 0.40, that is to say, 

 a number very near to that which was found by experiment ; and the 

 difference between the two values would very probably be still less, if 

 it had been possible to introduce into the calculation of the velocity of 

 the water, a correction which had to be neglected from the want of 

 sufficiently precise data, and which refers to the unequal velocity of 

 the different threads of fluid. By estimating the value of that correc- 

 tion in the most.probable manner, it is seen that it tends" to augment 

 a little the theoretical value, and approach the value of the observed 

 result. 



An experiment similar to that which I have just described, had been 



