174 Mr J. D. FORBES on the Horary Oscillations 



Arctic Regions in winter, we have precisely the reverse. It is 

 hardly necessary to remark, how totally opposed this is to the 

 theory of M. BOUVARD*. 



12. I would remark, fourthly, upon M. BOUVARD'S formula, 

 that, by fixing the mean temperature from which he calculates 

 his numerical results at the mean of the period from 9 A. M. to 

 3 P. M., he has given himself certain limits of which it is almost 

 impossible for a speculator not to avail himself, and which must 

 be at all times unsatisfactory on account of the great want of the 

 thermometrical observations for fixing the mean temperature of 

 any particular portion of the day in so many latitudes. I have 

 to observe, in the last place, that the results are far from satis- 

 factory even when the difficulties of the subject are considered. 

 By consulting the last column of the table already given, we 

 find the deduction of the formula for the oscillation at the Equa- 

 tor range from 2.72 to 4.82 miUimetres ; and, what is worse, that 

 the + and errors from the mean are not well balanced 

 throughout the quadrant of latitude, but that there is a general 

 rise in their value as we proceed farther from the Equator. As 

 it is undeniable that the tropical observations, from the small- 

 ness of the reductions required, afford the most correct results, 

 I am strongly of opinion that the mean of his numbers would 

 give us an equatorial oscillation greatly above the truth. By 

 looking over M. BOUVARD'S table, we shall find not a single 

 authentic series of observations gives an oscillation of three mil- 

 limetres (we except the observations at St Louis Maranhan, which, 

 wanting the results of the attached thermometer, should be dis- 

 carded), the maximum being 2.93, which occurred at the very 

 Equator. HUMBOLDT considers 3.3 millimetres the extreme limit 



* In 1829, the mean temperature of the seasons, distributed as above, was 

 1.5 R. + 4.8 R. + 0.5 R., and 4.6 R. In 1830, it was colder, as follows: 

 Q.l R. + 4.5 R., 0.7 R., and 7. 7 R. respectively. 



