M.-P. VOL. I.} CKAWFORD CONS7*ANT OF REFRACTION. 113 



for all three, and is o.3 (F). It is nearly always the case 

 (in this hemisphere) that the southern part of a room is a 

 trifle warmer than the northern. But this is not the case 

 on Mount Hamilton. The temperature of the air inside 

 is, on the average, very uniform and but very little (o.3) 

 warmer than the air outside. In his " Untersuchung liber 

 die Astronomische Refraction u. s. w.," Dr. Bauschinger 

 notes that the southern part of his observing room in 

 Munich was warmer than the northern, and that at night 

 the average difference between the inside and the outside 

 temperatures is i-3 (C). From his investigation, he con- 

 cludes that the temperature of the air within the observing 

 room should be taken into account. 



Because of these difficulties, many observers have seri- 

 ously considered the idea of mounting their instruments 

 under a movable house, so that when at work the instru- 

 ment will be entirely out of doors, and thus completely 

 obviate this difficulty. But this would needlessly endanger 

 the instrument. To accomplish the same purpose, the 

 Meridian Circle house being built at Kiel is to be con- 

 structed in the shape of a cylinder whose axis coincides 

 with the axis of the instrument. This is undoubtedly the 

 best form of construction. 



For the efficiency of the Meridian Circle house on Mount 

 Hamilton, the difference between the inside and the outside 

 thermometers can speak. As has been said, the average 

 difference (in the sense Inside-Outside) is -\- o.3 (F). 

 The maximum difference noted was one evening, a few 

 minutes before the sun had set, when the difference 

 was.+ i.i (F). The maximum difference noted here is 

 less than half the average at Munich. After this Meridian 

 Circle house has been completely opened for an hour and 

 a half, the temperature inside is practically the same as it is 

 outside. 



During the months October to December, inclusive, a 

 similar set of observations was secured. For these months 

 the average difference between the inside and the outside 

 temperatures is even less than for the summer months. 

 But the range of the difference is much greater for the 



