BAROMETER. 13 



incide exactly with its image reflected below by the mercury. 

 This method may be very good when the surface of the mercury 

 is perfectly pure and brilliant ; but this is very rare. It is gene- 

 rally dimmed by a slight layer of oxide, which makes the coinci- 

 dence of the point with its image uncertain. It is safer to judge 

 of the contact in a different manner. From the moment when 

 the point does more than touch the surface, it forms around itself, 

 by capillary action, a small depression, which, breaking the direc- 

 tion of the reflected rays, becomes immediately very easy to dis- 

 cover. It is enough, then, to raise the mercury so as slightly to 

 immerse the point ; then to lower it gradually until the little 

 depression disappears. If care is taken to make a good light 

 fall on that portion of the mercury which is under the point, 

 and to use the aid of a magnifier, the adjustment of the point 

 thus made becomes not only easy, but very certain, and the 

 errors to which we are liable are almost insensible, for they do 

 not exceed two or three hundredths of a millimetre, or a thou- 

 sandth of an inch. 



d. The level being thus adjusted to the zero of the scale, we 

 proceed to observe the height of the summit of the column. 

 Take hold of the instrument with the left hand, above the at- 

 tached thermometer, without moving it from the vertical ; strike 

 several slight blows in the neighborhood of the top of the 

 column ; then, by means of the screw, lower the slide which 

 carries the vernier, until the plane passing through the two 

 lower opposite edges of it is exactly tangent to the summit of 

 the meniscus that is, the convexity which terminates the column. 

 We know that this is the case when, placing the eye exactly at 

 the height of the summit of the column, we still see the summit 

 of the column, without there being any trace of light between the 

 summit and the edge of the ring. To convince ourselves that 

 the barometer has remained quite vertical during its operation, 

 we leave it to itself, and when it is at rest, we look again to see 

 whether the ring has remained tangential to the summit of the 

 column. If it has not, the verticality has been disturbed ; it 

 must be adjusted anew. It is necessary, at the same time, to 

 examine if the adjustment of the surface of the mercury in the 

 cistern has remained the same. The attached thermometer will 

 also be read anew, and if it indicates a temperature noticeably 

 higher than at the commencement of the observation, a mean 



