From this arrangement, it will be seen that every division on the vernier will be the tenth part of 11 tenths of an inch, or every division is equal to 11 hundredths of an inch, and, consequently, every division on the vernier exceeds, by one hundredth of an inch, every division on the principal scale. Suppose the vernier placed, as in the diagram, so that its upper edge d may be exactly even with the surface e of the mercury in the tube. If we examine the principal scale, we shall perceive that the mercury stands somewhat higher than 29 inches and 4 tenths. If we now look at the vernier, we shall find that the division 4, on its surface, coincides with a particular division on the scale. Now, as we have seen that each division on the vernier is one hundredth of an inch greater than one division on the scale, it follows that the space from 29 up to the level of the mercury, is 4 tenths and 4 hundredths of an inch, and, consequently, the true altitude of the column is 29 inches, 44 hundredths.

In using the vernier, considerable precision must be employed in making its upper edge a tangent to the curved surface of the mercury in the tube, and, consequently, its accuracy will still depend upon the skilful manipulation of the observer.

To obviate this, and render the barometer self-regulating, Mr. Christie, Secretary to the London Mechanics' Institution, has invented and constructed a barometer, in which, by means of a float, the vernier is set at its proper height by the rising or falling of the mercury itself. The annexed sketch will show in what the improvement consists. Here a b c represents a glass barometer tube about 35 inches long, exclusive of the part b c. It is not less than a quarter of an inch diameter inside, and enlarged at the top a to about two inches. On the surface of the mercury, shown by the shading, rests a small glass ball, or float d, supporting a slender steel wire, d e with its attached vernier f. This wire passes loosely through a guide hole in the projection g to keep it close to the scale of inches h i. In other respects, the scale and vernier are similar to those of the common barometer. The upper part of the enlarged tube being vacuous, the mercury is prevented from descending by the pressure of the atmosphere on the surface at d; and as the pressure increases, it will force down the mercury in the part c b of the tube, and, consequently, cause it to rise in the leg a b.

As the surface at d falls, the float carrying the vernier falls with it, and thus the edge of the vernier being always kept at a given distance from the surface of the mercury, will indicate the precise amount of any changes that may occur in atmospheric pressure.

Fig. 1.

Fig.2.

In order to increase the extent of the barometric changes, a contrivance is sometimes adopted, called the diagonal barometer, which is represented in Fig. 4. b c d is the glass tube bent at c, the altitude of which is less than 28 inches; hence c b includes the whole barometric range in the present form, while a c is the range it would have were the whole of the tube vertical. Now it is manifest, that by decreasing the angle at c, so as to bring b c nearer the horizontal position, we can make its proportionate length to a c as great as we please. Suppose b c so inclined that its length shall be three times greater than a c, then every rise or fall of 1 inch in a c would be equivalent to a rise or fall of 8 inches in b c. The difficulty, however, of observing the precise height of the mercury in this arrangement, more than counterbalances the advantage resulting from the extended range, and this form is, therefore, seldom adopted.

The wheel barometer is another contrivance for enlarging the scale, and rendering minute changes more easily observed. This, which is the common domestic barometer, is represented in Fig. 5, in which d b is the longer leg, a b the shorter, in which the changes of altitude occur: for as the diameter of the bulb d at the top is very large, compared with the diameter of the tube at c, a small fall in d will be equivalent to a considerable rise in the tube a b. Thus the surface at c may rise 3 inches, and thereby shorten to that amount the distance between the two surfaces of the column, which is the true height sustained by atmospheric pressure, without sensibly affecting the level at d. At c there is an iron ball, floating on the surface of the mercury, and partly supported by means of a thread passing over the pulleys, and carrying a small counterpoise at w. On the axis of the pulley, an index i is fastened, which moving with it, branches the circumference of the circular plate, on which a scale is drawn, to represent the rise or fall in inches; and also the terms fair, change, rain, etc, which certain altitudes have been improperly considered to indicate.

In this arrangement, it is manifest that as the column is supported by the air's pressure on the surface c, any diminution of that pressure will cause the mercury in the longer leg to fall, and that in the shorter to rise. On the other hand, if the atmospheric pressure increases, the mercury will rise in the longer, and fall in the shorter, leg; but, as before observed, the change of level will be scarcely perceptible in the former, on account of the enlarged diameter of the upper part. The changes, then, that occur in the shorter leg, may, without material error, be considered the representatives of the changes that are continually occurring in the atmosphere. Now it will be evident, on inspection, that as the ball c is partly supported by the mercury, it will partake of its motion. If the pressure of the air increase, the surface at c will be depressed, and as the iron ball must sink with it, the thread to which it is attached will, at the same time, communicate its motion to the pulleys, and through it, to the index, which will, consequently, move from the right towards the left of the graduated circle.