_ 
46 PHILOSOPHICAL SOCIETY OF WASHINGTON. 
special case of the triaxial ellipsoid in which they are straight. 
In all other cases they are curved. 
The full text of this paper may be found in the Archives of 
Ophthalmology, Vol. XII, Nos. 3 and 4. 
Mr. H. A. Hazen followed with a communication on 
THERMOMETER EXPOSURE. 
[ Abstract. ] 
Without entering upon the question, Where in any locality shall 
the air temperature be observed, it is proposed to discuss the even 
more important question, What shall be the environment of a 
thermometer that it may give the true temperature. ‘The practice 
has been very various: in England the Stevenson shelter is re- 
garded as a standard: this is a double-louvred frame, wholly of 
wood, 18 x 10 x 18 inches, and placed about 4 feet above grass. In 
Russia we find a large wooden outside shelter of single louvres 
open to the north, inside of which is placed a metallic screen, the 
whole being exposed 12 or 13 feet above grass. In any exposure we 
should seek, first, to allow the freest possible access of the outer air, 
and second, to screen the thermometer from direct sun heat, from 
precipitation, and from radiation, whether (a) from surrounding 
objects by day or (6) to the sky at night. 
It is important that we adopt some ready means of accurately 
determining the air temperature which may answer as a standard 
of comparison. This we have in the swung thermometer, which, 
by its free motion through a large body of air shaded from direct 
sunlight in the daytime, is calculated to give good results. 
Experiments have been tried with a so-called “ Pattern” shelter 
constructed of wood,.of single louvres, inclined 30° to the hor- 
izontal, thus giving a good air circulation. The size is 4x3x3 
feet, and it is erected at a height of 13 feet above a tin roof. In 
order to determine the Jeast admissible size for a shelter, thermom- 
eters were placed in the Pattern 5 inches apart and running in an 
east and west direction, and these were observed morning and after- 
noon. It has been found that with a hot sun and still air the heat 
from the louvres rapidly diminishes with distance and becomes in- 
sensible at 15 inches. Comparisons have also been made for several 
weeks between the: Russian and Pattern shelters; and the means 
of 100 sets of continuous observations on a still day, and again on 
a windy day, are shown n the following table: 
