804 



SCIENCE 



[N. S. Vol. XXXVII. No. 960 



supported by wires from the extremities of 

 the cross rod B. The ends of the rod A were 

 made adjustable perpendicularly to the plane 

 of the ring, so that the ring might be made to 

 swing on an axis parallel to its plane. The 

 ends of the rod were swung in solid supports, 

 adjustable to make the axis horizontal. In 

 order that the motion of the water might be 

 detected, a mixture of linseed oil and oil of 

 cloves of the same density as water was pre- 

 pared, and a few drops of the mixture were 

 shaken up with the water with which the tube 

 was to be filled. The globules of oil were 

 observed at a point C, between the ends of the 

 axis, through a micrometer microscope. 

 Difficulties from the astigmatic refraction of 

 the light by the water in the cylindrical glass 

 tube were overcome by sealing a tubular par- 

 affine cap, closed with a cover-glass and filled 

 with water, on the part of the glass tube under 

 the microscope, thus presenting a plane sur- 

 face through which to make the observation. 

 One side of the ring was weighted, so that on 

 releasing a catch at the side of the observer 

 the tube swung around through 180 degrees 

 in a definite time, and was held again by the 

 catch just under the microscope. 



In taking a reading, the microscope was 

 focused as nearly as possible on the center of 

 the tube, and the ring was left in position 

 until the oil globules had no appreciable mo- 

 tion. As soon as the catch which held the 

 ring in position was released, the time was 

 counted, with the aid of a metronome ticking 

 half-seconds, until the tube had turned and 

 an oil globule had been fixed upon to follow. 

 The globule was followed through a measured 

 length of time by turning the micrometer 

 screw, and the distance through which it 

 moved was recorded. Examples of these ob- 

 servations are given in the first three columns 

 of Table I. 



Variations in the readings arose from the 

 fact that the part of the ring toward the east 

 was near a cold wall, so that convection cur- 

 rents were produced as soon as the tube left 

 the horizontal position in making a turn. 

 This effect was made as small as possible by 



stirring the air with an electric fan. Other 

 variations came from the fact that it was 

 found impossible to adjust the horizontal axis 

 so nearly parallel to the plane of the ring as 

 to prevent a slight effect from turning the 



Time from 

 Releasing 

 Catch to 

 Following 

 Water's 

 Motion 



Time of I Distance 

 Fol- I Through 

 lowing' which 

 Water's' Water is 

 Motion, I Followed, 

 Sec. ! Mm. 



Time from 

 Completion 

 of Turn to 

 Following 

 Water's 

 Motion, Sec. 



Time on 

 Curves 



of Com- 

 pletion 



of Turn 



Initial 

 Veloc- 

 ity, '', 



Averaj 

 Case III. 



e V: Case I. =.0434; Case II. = .1580; 

 = - .0633 ; Case IV. = .0671. 



tube. Errors from the first cause were cor- 

 rected by reversing the direction of turning 

 in alternate readings. Those from the latter 

 cause were nullified by taking readings with 

 one side of the ring weighted and then shift- 

 ing the weight to the other side. In this 

 manner ten readings of each of four different 

 kinds were taken (Cases I., II., III. and IV.), 

 and the fact that the predominant motion is 

 positive, or toward the west as observed on the 

 south side, shows that the earth is turning 

 from the west to the east. 



Calculation of the Initial Velocity 

 In order to make an accurate estimate of 

 the velocity corresponding to any given read- 

 ing, the rate of decrease of velocity of the 

 water in the ring must be determined. If the 



