Statist leal fit vestigatiinm 



:«M» 



for measurement'. One of these papera deals with here<Hty and will be con- 

 sidered in the following chapter. The first of the remaining three belongs 

 to (Jalton's work on coni[)08ite photogrrtphy. It is entitle<l : "Conventional 

 Representation of the Horse in Motion'.' In this paper (Jalton endeavours 

 to construct the conventional attitude of sculptun; or of painting from com- 

 posites of Mtiyhridge's photographs of the "Horse in Motion." The final 

 result ought to represent a nude, for it is certaiidy a very hybrid structure. 



"The first composite sIiowh the hind logs di.stiuctly, thn s«h;oih1 fthows the fore logs distinctly; 

 And if duplicates of the first and second woodcuts are each divided into two halves and the b^t 

 defined halves of each art< uiiitt^i (in a way that might have occurred t4) Itaron Munchiius«'n if a 

 second rider's iiorse had sutl'ered as his own, and there had l)een a mist^ike in piecing them) a 

 result is protlucinl that shows a very fair corn'sjxindence with a not unconinion representation 

 in scul|)t.ure." 



The second paper is entitled: "The Ameriam Trotting Horse'." The 

 interest of this paper lies in its great value lus indicating the uitluence of long 

 continued selection on a character. Galton, dealing witli the statistics of the 

 speed of American trottere, shows that every three yeitrs from 1871 to 1880 

 the speed of the best horse increjised about two seconds, or in the nine years 

 there was an improvement Irom about a mile in 2 mins. 17 sees, to a mde in 

 2 mins. 1 1 sees. Perhaps the luost noteworthy point i.s, however, that not only 

 the speed of the fastest horee thus imiiroved, but the first hundred horses 

 maintained their relative speeds, or all increased by about the same two 

 seconds. (Jalton's final table is as follows: 



Number of Seco7ids and Tenths of Seconds in Excess of Two Minutes that arc 

 required for Rxmniwj One Mile lyy the Horses, whose order in the Rate 

 of Runniny in each Year is given at the Toj) of the Column. 



Mem. The first horse runs the mile ill about five or six seconds less than the tenth horse. 



Clalton's anticipation for 1890 is obtiiined by noticing that the values in 

 the vertical columns are nearly linear and then exti-apolating. 



"Supjiosing the coiulitions to he umintained, T should anticipate that in 1890 there will Im* 

 alM>ut If) horses that will run a mile in 2 minutes 15 seconds or less, .'inil tlint tin- fastest horse 

 of that year will run a mile in about 2 minutes 8 seconds'." 



' See our p. .317 H seq. » Nature, Vol. xxvi, pp. 228-9, July 6, 1882. 



' Xalure, Vol. xxvui, p. 29, May 10, 1883. 



 Actual best values for 1891, 1900, 1910, and 1922 were: 2' 8"-25, 2'3"-25, 1' 58"-75 

 and r 56"-75 respectively, so that Gallon's prediction was exact, if we take 1891 for 1890. 



