Measurement of Small Pressures. 457 



still open at the bottom. Knowing the inclination and the 

 distance a b, we know the difference of level between the two 

 points, and therefore the difference of pressures. One arm of 

 the gauge a b is permanently connected with a McLeod gauge 

 and a mercurial pump, and care is taken that this side shall be 

 tilled with a gas for which the McLeod gauge may be trusted. 



With this apparatus, a measurement can scarcely be made 

 in less than three minutes ; even after much practice. We 

 have accordingly constructed two instruments in which we 

 still utilize the principle of measuring the tilting of the 

 apparatus needed to bring mercury levels to the fiducial points, 

 but in which Mr. E. has applied a very different method of 

 reading. At a b, figure 2, are the two arms of the gauge, seen 

 in horizontal section through the fiducial points. At c is a 

 pair of mirrors so placed that, in a microscope at d, the two 

 points a b are seen side by side. These two mirrors are 

 wrought on the end of a glass rod some six millimeters in 

 diameter. They are mounted on a spherical base, so that they 

 can be rotated in' any desired way about a point in their 

 intersection. They can also be moved in the direction c d. 



The two arms of the gauge, the mirrors, and the microscope, 

 are all rigidly fixed to a support whose inclination is deter- 

 mined by a micrometer screw. Since the points, mirrors, and 

 microscope are all moved together, vision of the points always 

 takes place under precisely the same conditions. 



The amount of mercury in the gauge is capable of adjust- 

 ment, somewhat as in the first form, but the adjustment does 

 not require to be so delicate. 



When, with suitable illumination, we look in the micro- 

 scope, we can not see the surface of the mercury, but we see 

 two images of each point. One image of each point is seen 

 directly, and one image is seen by reflection from the surface 

 of mercury. It is important that the ends of the points shall 

 be small hemispheres. Suitable adjustment causes the four 

 images to appear as in figure 4. When a difference of pressure 

 is to be read, the micrometer screw tilts the whole apparatus, 

 and the two tangent lines a b and e d are quickly made parallel, 

 when the reading is taken. Ten seconds may well suffice, 

 especially if the eye is assisted by an eyepiece micrometer with 

 parallel lines. It will of course be noticed that the reflection 

 from the surface doubles the distance to be measured. 



In the instruments which we have used, which were 

 intended only for the measurement of small pressures, we 

 made our U-tube less than an inch long. To do this with 

 tubes six centimeters in diameter required us to connect the 

 two vertical arms by a horizontal cross-piece, so that our 

 apparatus looks more like the letter H than U. To secure 



