Radiometer and Sunshine Recorder. 163 



The vapor chamber should be as small as possible, so the 

 cups are arranged to be lowered or raised bj screws in order 

 to just accommodate, and no more, the largest probable change 

 in temperature and pressure for the season, and to adjust for 

 elevation above sea level. Because, as the tension of saturated 

 vapor in any chamber is the tension of the coolest part ; so the 

 smaller the chamber the more sensitive it will be. The cir- 

 cumference of the cylinder O is 48 centimeters, which makes 

 one millimeter equal three minutes of time. A difference of 

 one millimeter in tension, by adjustment of the weight of car- 

 riage, causes a motion of two millimeters in the pen. . 



These dimensions seem to be large enough to nearly exclude 

 mechanical errors and not too large for convenience. If more 

 accurate values as regards time are required the motor should 

 be arranged for two different speeds by having, for example, 

 the first contact of the pointer made through a light spring, 

 which starts the motor as above, but upon increase of weight 

 it closes a second circuit which would raise a brake, or friction 

 wheel, from the motor, thereby increasing the speed. 



A motor that moves at a constant speed is either so slow 

 that the pen falls behind a sudden change in temperature, or 

 so fast that it causes oscillation on slight changes. 



The use of a saturated vapor in this way I believe to be 

 original. I have chosen ethyl alcohol as giving the best aver- 

 age results, though experience may prove that the use of ether 

 in winter and alcohol in summer, or some such combination, 

 may be better. 



The difference in vapor pressure in millimeters of mercury of 

 ethyl alcohol between 0° 0. and 5° C. is about 5 mm , and lower 

 down the difference is very much less, which would probably 

 give inaccurate results. 



To reduce the tracing on card there must be either a syn- 

 chronous tracing, preferably on the same card, by a thermo- 

 graph, or observations of a thermometer; for the curve repre- 

 sents the difference in temperature between a shaded and a 

 black bulb thermometer. 



The pen, when the sun is not shining, draws a straight line 

 around the card which is the base line of measurement, but 

 must have a different value for different temperatures. Sup- 

 pose that at twelve o'clock, noon, on a certain day the curve 

 measured 38*4 mm from the base, and at the same time a shel- 

 tered thermometer or the thermograph trace showed a tem- 

 perature of 25 C. ; then, as the tension of alcohol vapor at 25 

 C. is 58'86 mm , adding half of the 384 (for the machine multi- 

 plies by two) the sum equals 78'06 mm , equals tension at 30 C, 

 equals temperature of black bulb thermometer at that time. 

 This may be done by simple inspection by making a trans- 

 parent scale as follows : 



