60 Mr Lavvijon on the Temperature and Atmosphere 



As no observations were made from 11 p.m. to 4 a.m., it was necessary 

 that this gap should be filled up, before the mean temperature could be 

 found. Sir David Brewster had long ago shewn, in the case of the thermo- 

 metrical observations at Leith Fort, that the curve of daily temperature 

 might be represented, with considerable accuracy, by parabolic arcs. 

 On projecting the above observations, their resemblance to parabolic 

 curves was sufficiently obvious, though differing in their arrangement 

 considerably from those at Leith. In the present case, the 10 p.m. 

 observation gives a temperature less by 0^.08 only than that obtained by 

 calculation, and the 5 a.m. observation one greater by 0°.05; conse- 

 quently, the actual curve of temperature must have crossed that obtained 

 by calculation, between 10 p.m. and 5 a.m. ; and it is not likely, that, 

 had the observations been continued at the intermediate hours, the 

 resulting temperatures would have differed materially from those af- 

 forded by calculation ; and, as any difference that may exist between 

 the latter and former will affect the mean temperature, derived from 

 the whole observations, by only one-fourth of its amount, I have not 

 hesitated to insert them in the above table, in place of actual observa- 

 tions, in order to derive from the whole the mean temperature for the 

 three months. 



The mean temperature of June, July, and August 1841, in Barbadoes, 

 is thus found to be 80".384 Fah. ; and the diurnal curve of temperature 

 crosses it at 7 h. 46 m. a.m., and again at 6 h, 34 m. p.m. The diurnal 

 changes of temperature will be readily comprehended by inspecting the 

 accompanying drawing, in which the broad horizontal line indicates the 

 mean temperature, 80''.364, while the continuous curved line represents 

 the curve, as derived from observation, and the dotted line, the results 

 obtained by calculation. The actual curve of temperature presents a 

 remarkable depression between mid-day and 2 p.m. This is not acci- 

 dental, but is owing to the air, which is in immediate contact with the 

 ground, then becoming much heated, which causes it to ascend, when 

 its place is supplied by colder air from above. About mid-day, espe- 

 cially when afresh trade wind is blowing, the thermometer is in a constant 

 state of oscillation from this cause, and it is not unfrequently seen to 

 alter even a whole degree in 15 or 20 seconds. 



As the observations on each side of 1 p.m. seem to indicate that as very 

 nearly the hour of maximum temperature (supposing the depression in 

 the curve above noticed had not existed) of the day, I have in the cal- 

 culation assumed it to be so, and have farther supposed the temperature 

 then to be such as would best represent the afternoon portion of the 

 curve according to observation, that being the part of the day when the 

 danger of error from radiation was least, and the breeze generally 

 strongest. The hour which gives the lowest temperature is 6 a.m., 

 though the minimum of the 24 hours, during July and August, actually 

 occurred about oh. 45 m. a.m., nearly the time of sunrise. By inspecting 

 the drawing, it is obvious that those portions of the diurnal curve, ex- 



