Reirospective Criticism. 619 



certain number of degrees ; then mark the nxmiber of seconds that elapse, 

 till it fall exactly half the number of degrees raised. Raise the alcohol 

 again the same number of degrees, and expose the ball to the full impression 

 of the wind, and mark the number of seconds that it takes to fall half the 

 number of degrees it rose. Divide the number of seconds elapsed in still 

 air by the number of seconds elapsed in the full play of the wind; throw 

 oft' 1 from the quotient, and multiply by 4^ ; the product expresses the 

 velocity of the vvjud in miles per hour." Now I think that one glance of 

 the mind will satisfy any person tiiat this formula, or rather the principle of 

 the fornuda, is incorrect. Even granting that the cooling power oi a stream 

 of air is exactly in proportion to its velocity, which, I consider, remains to 

 be proved, yet from the slow heat-conducting power of the glass cover of 

 the alcohol, and the not innnediately regularly throughout cooling of the 

 alcohol itself as the velocity continues to increase, the cooling must pro- 

 ceed in a slower ratio than the velocity ; and this disproportion will be con- 

 siderable where the glass is thick, the' alcohol in considerable volume, and 

 the velocity very great. For example, take extreme cases : suppose the 

 alcohol covered with so bad a heat-conducting envelope that it shall require 

 one hour (.3600 seconds) in still air to fall from 60° to 30°, the air itself 

 being at 40"; can we believe, that, in the case of the current of air being 

 85i miles per hour, the alcohol would cool twenty times faster (in 180 

 seconds) within this almost hnpervious to heat envelope than in the still 

 air ? In this last case, the outside of the cover would immediately be 

 cooled down almost to the temperature of the atmosphere, and the rapid 

 current of wind would extract very little more heat than the calm air 

 extracted. T am aware that the difference of temperature between the 

 surface of the glass and the interior fluitl comes into account in both cases, 

 as well when the air is still (that is, when there is only a slight current 

 upward, b}' reason of the surrounding air being diminished in density by 

 the expansion caused by the heat given out by the glass), as in the full 

 play of the wind; but in the former case, from the slower cooling, the 

 difference of temperature between the surface and interior is very small, 

 and in the latter, especially should the cooling be rapid, the difference is 

 considerable. Different instruments will also reciuire different allowances for 

 radiation, and the loss of heat by radiation with the same instrument will 

 also be variable, affecting the results. Granting Professor Leslie's premises, 

 this anemometer could "only be correct should the outside of the glass 

 maintain equal temperature with the whole interior of the niass ; and 

 should the very small loss of heat by radiation go on increasing in the 

 same proportion as the loss by contact, which is impossible. Let it be 

 understood, that I have no knowledge whatever of Professor Leslie's 

 anemometer but through your Magazine, (p. 231.) I think it impossible 

 but that the above-stated causes of error must have been obvious to him. 



As distinguishing epithets for different velocities of air, the following terms 

 are pretty generally understood, perha[)s definitely enough for general pur- 

 poses : — 



Gentle breeze - - - = 5 miles per hoiu*. 



Breeze - - - - = 10 



Smart breeze - - - = 20 



Gale =30 



Hard gale - - - - = 43 



Tempest or storm - - = 60 



Violent tempest or storm . = 80 



We ilismiss hurricane to lower latitudes. The sense may be cultivated 



to apjiropriate these terms pretty correctly to determinate velocities of air, 



hv Irequently taking a connnanding situation, overlooking a country, the 



distance between whose remarkable lines or points is known to you, and 



