HORARY VARIATIONS OF THE BAROMETER. 419 



It is readily understood that, under the tropics, wliere tht- duration of the days 

 varies so little, the barometrical heights should be taken at moments always 

 equally distant from mid-day, and therefore always equally distant from the 

 rising and setting of the sun, or at least very nearly so. Such are the observa- 

 tions made by Humboldt, M. Doussingault, &c But, in France, how are we 

 authorized to collate and compare the effect of the warming of the earth's sur- 

 face by the solar rays, m sumvier, at nine o'clock in the morning when the sun 

 is already five hours high, with that which has been produced at tiie same hour 

 of the morning, in winter, when the sun is scarcely above the horizon and its 

 rays reach the surface of the earth only through thick clouds or mists 1 There 

 is evidently no comparison possible between observations made under conditions 

 so different. It would seem that if the maximum of the morning corresponds, 

 under the torrid zone, to eight or nine o'clock, as Humboldt has verified, this 

 maximum will, with us, occur in summer, when our days are long, at about six 

 o'clock. Further north it would be at five o'clock, while near the; pule there 

 would be no maximum either morning or evening. In the winter, on the con- 

 trary, the maximum would, in our latitude, occur at about ten or eleven in the 

 morning and be necessarily not very obvious. In high latitudes, where the sun 

 scarcely shows itself, all would be confounded, and the diurnal variation, if not 

 wholly absent, escape the notice of observer.-? by reason of its diminutive pro- 

 portions. Continuous observations, directed to a verification of the suggestions 

 thus thrown out, could not fail to be of much interest, especially if it were prac- 

 ticable to make them with a registering instrument aud a barometer filled with a 

 liquid much lighter than mercury. 



The inquiry has sometimes occurred to me whether there might not be found 

 in the experience of our every-day life some simple facts which would serve to 

 illustrate and confirm the above theory on the diurnal period, and it has seemed 

 to me that what takes place in our ordinary chimneys has much analogy with i 

 the oscillations of the barometric column. Let us suppose the fire to have been , 

 extinct for some time, and the air of the chamber, like that of the chimney, 

 nearly motionless. Let fire be now supplied on the hearth ; the lower air will 

 become heated and have a tendency to rise, but it is necessary that it should 

 make for itself a way in the column which occupies the body of the chimney, 

 that, so to say, it should displace this whole column and set it in motion. The 

 struggle is of more or less duration ; sometimes the warm air, not being able to 

 rise quickly enough, diffuses itself in the chamber; the cMmwcy sinokes. As 

 long as the struggle continues there is a i epression of the air. and consequently 

 an augmentation of pressure in the body of the chimney. Here we have the 

 first phase of the diurnal period; it con-esponds to ihe warming of the surface 

 of the earth by the newly risen sun. By and by the moment arrives when the 

 pressure of the aif in the body of the chimney attains its greatest intensity, 

 corresponding to the maximum of eight or nine o'clock in the morning. After 

 this, all the air being finally put in motion and the force of inertia overcome, 

 the tension diminishes, and through the velocity acquired, this tension in the 

 tunnel of the chimney must become even inferior to what it was before the 

 kindling of the fire on the hearth, i-epresenting the period from nine or ten 

 o'clocTi. to midday and to even three o'clock in the afternoon. 



Suppose now that after having kindled a brisk fire and established a strong 

 draught, we cease to supply fuel; the draught will diminish, the air will ascend 

 less rapidly, the pressure will augment in the body of the chimney, and if there 

 were a barometer therein, the mercury would ascend in the tube. .A.t length, 

 when the fireplace has become altogether cooled, the entire column of air will 

 acquire a movement from above downward, and will exert a pressure on the 

 hearth ; it is what often occurs in our dwellings during summer, when great heat 

 prevails. The external air, wh.ch is warmer than that of the chamber, becomes 

 cooled in penetrating from above, descends along the tunnel and throws down 



