Ti'iiHUf'tioii Stml'n'H 



M 



I 



Perhaps u simpler unungtMuent would be to replace the liiikiif^e \>\ j.ji.-. ai, 

 a and h running in a slotted Iwir turning about a pivot C We reprwluoe 

 (page 4(i) an illustration of the tinal apparatus; the pencil or style could be 

 replaced by a drill '. W^ith this instrument for twelve yeare the ronlinuoua 

 automatic weather records for seven stations (velocity and direction of 

 wind, drv and wet bulb thermometers, barometer, vapour-tension and ruin) 

 were reduced to manageable dimensions and published. Of this [>ublicatioti 

 Galton remarks: 



"It 8urpris«>H nie that meteorologists have not maile much more use than they have of these 

 comprehensive vuiiimeM. But thHre is no foretelling what asf>ect of meteorology will In? tnken 

 up by the very few earnest anil capable men who work at it. Kiu-h of them wnnt« voluminous 

 (InUi urriiiigwl in tlu; form most convenient for hi.n own |Mirticular inquiry'." 



Probably the use has not been made of these graphical charts that might 

 well have been made; but Galton's own results indicate that we need aimul- 

 taneous data for a far wider range than Great Brituin, and further modern 

 methods of multiple correlation, which seem likely to be most productive of 

 result in present day meteorology, demand numerical values, and these are 

 hard to obtain from the graphs; not only can they scarcely be read off with 

 the requisite accuracy, but to reconvert the graphs into any numbere whatever 

 is in itself a most arduous task. 



Galton's compound pantagraph has indeed a far wider field of u.sefulness 

 than reducing automatic weather returns. The difficulty is that it is not 

 made commercially and procurable at a moderate cost. 



A second instrument devLsed by Galton about this time will be found 

 described in the Report of the Mcteoroloyicul Committee, 187 1 (p. MO). It was 

 devised for obtaining mechanically the vapour-tension curve from the curves 

 of dry and wet bulb thermometers, but again it can be used to serve a much 

 more general purpose, namely to obtain the curve of a variate whose ordinate 

 is a given function of the ordinates of two other curves — all three curves 

 having the same abscissji. The machine depends upon the construction of a 

 surface corresponding to the function the variate is of the two other ordi- 

 nates (i.e. in CJalton's case the vapour-tension in terms of wet and dry bulb 

 thermometer readings). By fine screw adjustment the cross-hairs in two 

 microscopes are brought into accordance with the tops of the ordinates in the 

 two curves, but the screw which adjusts one microscope moves the surface 

 parallel to one axis, and the screw which adjusts the otner microscope moves 

 the surface perpendicular to this direction. Thus a vertiail style resting 

 on the surface raises to an ade(juate height a scriber which marks the 

 ordinate or function-value of the compound variate*. It would be out of 

 place here to give a more complete iiccount of the in.strument. but my 

 more mechanically minded readers will gnisp the general ideji from the 



' The theory is fully described in the Minuteni of th« MettorolofficfU CommitUe, 1869, p. 9. 

 It is alst) figured in the Katahuj vuUhfmntinclier AfodelU, Apparate und InttriimmU, of the 

 Deutsche -Mathonuitiker-Ven'inigung, 189l', p. 232. ' Memori't, p. 236. 



' In Galton's actual instrument (see our p. 48) the required curve w^as rec»)r»led on a xinc 

 plate (partly removed in figure to show scriber R). The scrib«'r received when adju^ted a 

 blow from the hammer H worked by the action of the operator's foot on a treadle. 



