292 PROFESSOR C. PIAZZI SMYTH ON 



But for the present we may now go forward to consider the observations actually 

 made in reference to the march of cosmical super-annual variations; and need not be 

 delayed by any difficulty in obtaining new values of the minute corrections applied to 

 the former set of instruments; for they seldom amounted to more than from ^ to ^ of 

 the super-annual effects of the natural changes from without the earth, which it is now 

 our proper business in an astronomical Observatory to inquire into. Without further 

 detail therefore I introduce three varieties of numerical tables. 



Table I. Constants of the new Earth Thermometers, 1879-87. 



Table II. Original observations of the new Earth Thermometers, and two smaller- 

 auxiliary thermometers, weekly for the years 1879 to 1888. 



Table III. Monthly, Quarterly, and Annual Means taken quarterly, of the new Earth 

 Thermometers from 1879-1888; the last column having annual, or 

 seasonal, i.e. summer and winter effects entirely eliminated. 



Table IV. Parallel observations of Scottish Meteorological Society in Shaded Air- 

 temperature, and Rain-fall depth, each expressed in single annual 

 means. 



Also one Plate, representing all the observations, together with the Sun-spot-numbers, 

 as kindly furnished by Mr Whipple, Director of the Kew Observatory, for the times 

 concerned. 



Now on contemplating said Plate and its several contents, we can hardly but be struck 

 with the remarkable testimony which each of the four earth-thermometer curves bears to 

 the others, both in their resemblances and graduated differences ; the effect of any change 

 of temperature always decreasing with depth, but increasing in retardation therewith ; 

 testifying therefore to its production by waves of super-annual heat from outside the 

 earth striking on the surface and being slowly transmitted into its substance. Also that 

 such waves are of a very intense and powerful character, though short-lived, seldom lasting 

 more than two years ; and indicate radiations from the Sun, observing similar periods. 



These conclusions are borne out abundantly by the four earth thermometers, when 

 their annual means are taken every quarter ; but not very satisfactorily by single annual 

 means, as with the Scottish Air temperature, the Scottish rain-fall, and the Sun-spots 

 themselves. Wherefore I would beg to call attention among those Observatories which 

 are concerned therein, to the importance of their present single annual means of various 

 data, being replaced by quadruple annual means ; or, better still, by duodecimal annual 

 means — whenever the search for super-annual cycles includes those of sub-annual char- 

 acter as well. 



But can that necessity be distinctly shown in this case ? 



I am happy to say that it can be ; and has just transpired in a new and very 

 unexpected manner. 



Look for instance along each earth-thermometer curve, and note the length of line 

 from quarter to quarter of a year, when the external temperature is increasing ; and see 



