48 



Life atid Lettertt of Frane'iH (ru/fon 



aocompaiijiiig figure. Galton constructed his surface from a table of 400 

 values of the vapour-tension, 4(»0 iioles being l)()re(l into a solid rectangular 

 l)lock to these 400 values spaced pro|>erly apart, and then the remainder cut 

 away, Hletl anil smootluHl. The construction at that time did not cost more 

 than £6. Here agjvin it is easy to think of many purposes to which a machine 

 of this kind could Ik* put, but jis it luus never been made as a coiinnercial 

 article, it has never come into general use. Perhaps this brief notice may 

 remind investigatore of the existence of Galton's design. 



^ii*^ 



Galton's Trace Compnter— a machine for tracing a curve, whose ordinate in any arbitrary function 

 of two other variate vahies at the game absciRHa or time. 



A third instrument designed by Galton a little earlier (18G7) never came 

 in being, owing probably to a discouraging letter from Balfour Stewart at 

 that time at the Kew Ob-servatory, who laid great stress on comparison 

 of pairs of automatic jneteorological records at different intervals. Galton 

 was easily discouraged and was apt to treat the judgments of the really able 

 people whom he consulterl as sure to be better than his own. It certainly 

 was a pity that in this ca.se he was put off completing his model. It was 

 of the following nature: a map is mounted horizontally on, let us say, a 

 metal plate; then holes are drilled at each meteorological station, and a 

 rod of a convenient length is free to move vertically up and down in the 

 hole. Templates are now cut to the continuous automatic records of any 

 meteorological character for these stations and are fixed in vertical planes 



