370 H. A. Hazen—Thermometer Exposure. 
come by its rapid motion through a large body of air. Obser- 
vations on clear nights in September and October have shown 
the ‘‘fronde” sometimes ‘2 or ‘8 degrees higher and some- 
times the same amount lower than a thermometer from which 
all radiation was cut off. Experiments are still needed in sum- 
mer and in different situations to fully settle the question, but 
it seems probable that the “ fronde,” if shielded from direct. 
sun heat during the day, will give at all times the most accu- 
rate temperature that can be obtained. 
The following brief description of some of the previous 
experiments in this field, will serve as an introduction to subse- 
quent work. Possibly the most complete results hitherto pub- 
lished are those from observations taken, under the auspices of 
the English Royal Society, upon a large open field at Strath- 
field Turgiss. The observations were taken at 9 A. M., 3 P. M, 
and Y p. M., from November, ’68 to April, 70 inclusive, January, 
70 only being omitted. The stands tested consisted of eight 
forms ranging between the open stand like Glaisher’s and the 
closed like Stevenson’s. Asa result of these tests it was deci- 
ded that the Stevenson was least faulty, though it was nob. 
claimed that even this was all that could be desired, and espe- 
35° (it would be a matter of much interest if the atmospheri¢ 
conditions giving such large differences between two shelters so 
near each other could be studied), These comparisons would 
seem to show a lack of ventilation in the Wild shelter ree 
was so much higher than the other that the wind should have — 
had freer access to it. lso a mere agreement between the 
two cannot be regarded as proving the accuracy of either, bat 
since there are manifest defects in the Stevenson we may COD 
clude that neither is satisfactory. Comparisons are also given 
