BoEDDiCKEE — On Lunar Radiant Heat. 501 



6. Explanation of the preceding Table. — Columns I. and II. After the remarks 

 sect. 4 (page 486), no more need be said in explanation of these columns. Each 

 value G is, unless anything to the contrary is mentioned in the last column, the 

 mean of ten differences of eleven consecutive readings of the galvanometer. 



Column III. — The probable errors in this column are based on the deviation of 

 the single readings from the mean ; they give, therefore, a tolerably clear idea of 

 the steadiness of the apparatus. As was to be expected, the very considerable 

 increase in sensitiveness of the thermocouples since 1884 has also increased the 

 probable error of each heat-value G. Besides this, a very great increase is 

 naturally to be perceived during the periods of the most rapid change of the 

 lunar radiation. The mean probable errors, the absolute values as well as when 

 expressed in per cents, of G, according to the phases of the eclipse stand as 

 follows : — 



Before totality. 



Average P.E. 

 ^ — _ ^ , 



in Divisions of 

 the Galvanometer in °/„ of G. 



Scale. 



1. Before the first contact mth the penumbra (18 observations), . . ± 8-ll_ + 1-47. 



2. During the progress of the penumbra (81 ohs.), ±19-11 ± 4'14. 



3. During the progress of the shadow (52 obs.) ± 12-75 + 19-88. 



After totality . 



4. During the progress of the shadow (76 obs.), ± 14-86 ± 22-64. 



5. During the progress of the penumbra (88 obs.), ± 13-60 ± 3-68. 



6. After the last contact with the penumbra (132 obs.), . . . . ± 14-55 + 2-89. 



The increase in the last column during the progress of the shadow is here, of 

 course, chiefly due to the smallness of the values of G, the atmospheric and other 

 disturbances (as, for instance, the errors of observation) being the same as during 

 the other periods. The mean probable error of the whole series of 447 observa- 

 tions amounts to 14-55 divisions of the galvanometer-scale, or to 8*55 per cent, of 

 the reading G. It appears doubtful whether the sensitiveness of the apparatus 

 can be pushed much further if the mode of observing entirely in the open air be 

 adhered to. 



Columns IV. and V. — s is the Moon's true zenith-distance, and G^ the lunar 

 radiation reduced to zenith by means of Dr. Copeland's table for the extinction of 

 the lunar heat in our atmosphere, given in the Philosophical Transactions of the 

 Royal Society for 1873, p. 598. 



