270 Mr. F. Galton on Meteorological Statistics [Apr. 24, 



shown in the first column of Table III., and the four different settings 

 correspond to the four lines in the body of that Table. 



The slides n, e, s, and w move to and fro parallel to the tramways, 

 and are each furnished with bars that can be pushed vertically up and 

 down, and a rod projects horizontally from each of the bars towards the 

 zinc plate opposite to it. Graduations referring to the force of the wind 

 are placed on the bars, which are thereby adjusted to come in contact at 

 the proper levels with the zinc plates, and to be pushed back through a 

 distance equal to the length of the abscissae at those levels, when A A is 

 run forwards on its tramway. 



Thus far we have obtained the result that n, e, s, and w shall be seve- 

 rally pushed back to the distances which would be sailed over in one day 

 if the wind blew on 100 occasions with specified forces from each of those 

 quarters, and if the ship were sailing to that quarter whose initial is 

 opposite to the zinc plate marked 0. As A A is adjusted in the figure, 

 that quarter would be N. ; if it were moved one step to the left it would 

 be W., if 2 steps it would be S., and if 3 steps it would be E. 



"We have now to diminish the movements impressed upon n, e, s, and 

 w in the ratio of the percentage of occurrence of the several winds. 

 This is effected by linking them to another series of slides, N, E, S, and 

 ~W, as shown in the figure, and by attaching adjustable centres to the arms, 

 there shown in a vertical position. The standards to which those centres 

 would be clamped are almost wholly removed in the figure, but the top and 

 bottom of them can be seen. The reduction will be correct within such 

 limits as we need, when the links are somewhat longer than in the figure 

 and the arm does not swing through more than 40°. I therefore do not 

 care to propose in this case the somewhat more complicated, but per- 

 fectly accurate arrangement which I contrived for the parallel slides of 

 the pantagraphs now in use at the Meteorological Office, of which a de- 

 scription is given in the Report of the Meteorological Committee for 

 1870, p. 30. 



I place the adjustable centres between the links in the cases of N 

 and S, and above them in those of E and W ; consequently, when n, e, s, 

 and iv are all pushed back, N and S will advance, and E and "W will 

 retreat : the reason for doing this will be seen presently. In order to 

 graduate the arms for the adjustable centres, we must take the part 

 of the distance between the pivots of the links as the unit of measure- 

 ment, and measure in both cases from the pivot of the upper link as the 

 zero-point. Then if p=the percentage to which the movements are to 

 be reduced, the graduations for p between the links are determined by the 



formula -jff^ and when above the links by -^-P . 

 100 + p J 100 — p 



The results we have now attained are that, when A A is pushed forward, 



N, E, S, W shall move alternately forwards and backwards, each through 



a distance corresponding to that which a ship would sail towards the 



