HALL. — DEVIATIONS OF FALLING BODIES. 347 



the nearer side of the spout.* It is possible that in these trials, at some 

 points of the fall, the clear distance between the ball and the nearest 

 part of the spout was not more than 5 cm. On October 18 a set of 

 twelve balls was dropped with the spout 5 cm. north from its original 

 symmetriral position, and another set of twelve balls with the spout 

 5 cm. south from its symmetrical position. The first set fell about 

 0.004 cm. north and the second set the same distance south from the 

 plumb-line, a mean deflection of 0.004 cm. away from the nearer side 

 of the spout. On October 20 the experiments of October 18 were 

 repeated with the same displacements of the spout, but on this last 

 trial an average deflection of about 0.04 cm. away from the nearer side 

 of the spout was observed. We have, then, according to these not very 

 accurate tests : t 



Eccentricity in spout. Deflection due to eccentricity. 



10 cm. 0.32 cm. 



5 cm. 0.02 cm. 



During the general course of the experiments the eccentricity of the 

 line of fall was probably less than 1 cm., and, during the latter part of 

 the time at least, it was in such a direction as to produce a tendency 

 toward northerly deviations of the balls. This tendency must have been 

 very sliglit indeed, however ; too slight to deserve further consideration. 



The possible errors from inaccurate levelling fall naturally into two 

 classes: 1st, those which are made at the sending apparatus; 2d, those 

 made at the receiving apparatus. The errors of the first class are due 

 to the fact that, whereas the position of the plumb-line at its top is deter- 

 mined by the position of the " beak " (r in Fig. 2), the position of the 

 hall just before release is determined by its contact with the lower end 

 of the central hole in the release plate. The beak is about 4,5 cm, above 

 the centre of the ball in position, and any considerable error in leveiiing 

 would evidently put the tip of the beak a perceptible distance to one side 

 fi-oai the vertical through the centre of the ball. Much care was given 



* Newton, in suggesting to Hooke a study of tlie course of falling spheres, 

 warned him that " in a narrow well the bullet possibly may be apt to receive a ply 

 from the straitened air neare the sides of the well, if in its fall it come nearer to 

 one side than to anotlier." 



t In these experiments of October, 1902, lend was used instead of tallow as tlie 

 receiving medium. The individual variations of position of tlie fallen balls were 

 perhaps as great as when tallow was used ; but with lead twelve balls could be 

 received in one pan, and changes of temperature at the foot of the tower were no 

 longer troublesome. In further experiments I should use lead. 



