124 Report S.A.A. Advancement of Science. 



But by using temperature and humidity results for half an hour 

 later in each case, so as to agree with the pressure and dew-pomt 

 hours, we get mean annual temperatures within o°'5, and humidity 

 percentages within °'T% 



4. For four observations i^r diem, equidistant at six hours 

 apart : 



Pressure Any four equidistant hours. 



Temperature II., VIII., XIV., XX., or 



v., XL, XVII., XXIII. 



Dew-point III., IX., XV., XXL, or 



VI., XIL, XVIIL, XXIV. 



Humidity As temperature. 



The s€t of hours, IL, VIIL, XIV., XX., is evidently of much 

 importance. If we make the dew-point set an hour earlier, for 

 agreement sake, the deduced annual mean would only then be 

 o°'2 in error — a quantity probably well within the limits for the 

 wet bulb. These hours for the barometer are of special con- 

 venience, because, although any four equidistant hours in twenty- 

 four would do, it happens that the mean at IL is nearly equal to 

 the annual mean, and consequently the mean of the obser\'ations 

 made at VIIL, XIV., and XX., must also be equal to the same 

 thing. It is only necessary, then, to read the barometer at VIIL, 

 XIV., and XX. These are the very convenient hours used at 

 Kenilworth for observation of the standard barometer, made for 

 the control of the photobarograph. As for the temperature, direct 

 readings of dry and wet bulbs may be made at VIIL, XIV., and 

 XX., and a registration of the values at IL easily effected by a pair 

 of reversing thermometers. * 



The annual normals of pressure in Table i are obtained from 

 the continuous record of a photobarograph. For the whole diurnal 

 period between VIIL and Midnight nearly all the readings of dry and 

 wet bulbs are made by eye observations of a standard. From I. 

 to VII., however, most of the readings are taken from the registra- 

 tions of reversing thermometers. And it becomes pretty clear, 

 in tabulating the results, and from such occasional tests as have been 

 practicable, that two at least of these reversing thermometers read a 

 trifle too low. The IL dry bulb, e.g., would seem to require a plus 

 correction of nearly o°'i5 throughout its scale, and the IV. wet bulb 

 almost exactly the same. The effect is to increase the ratio of 

 humidity at II. by about o"6%. and to decrease that at IV. bv nearlv 

 an equal amount. Observers acquainted with instnmients of this 

 class will understand their lapses in respect of the displacement of 



See also the same physicist's classical '' Handbook of Climat(jl(>gy," chap. 1. 



