378  W.AH. Pickering—American Trotting Horse. 
Art. XLII.—Hvolution of the American Trotting Horse; by 
: . H. PIcKERING. 
THE two articles which have appeared on the above subject 
in the July and August numbers of this Journal have but just 
come to my knowledge, as I have been absent from the city 
for several weeks. Mr. Nipher has dealt with the question in 
a somewhat different manner from myself (see Science, May 
4th), and has arrived at quite different results. 
According to my calculations the value fo (s = seconds, 
T = years), instead of diminishing was nearly constant, with a 
slight tendency to increase. Or, in other words, horses were 
improving their speed slightly more rapidly now than they 
were a few years ‘ago. is of course might readily be ex- 
plained by the greater care bestowed in breeding and by greater 
attention being paid to the subject now than formerly. This 
could not last for many years, however, and the value of iT 
must finally diminish to zero. But in the mean time we have 
no means of saying how soon this will take place, or what will 
be the value of s. According to my work we could only pre 
dict what speed would be made in the course of the next few 
years, and it was concluded, for example, that two minutes 
would be reached about the year 1907. 
Instead of taking his dates directly from the races, Mr. 
Nipher prefers to interpolate them-from some curves which he 
has calculated. If all the observations contributed by Mr. 
Brewer could be included in this scheme, it would undoubtedly 
be the best method. As it is, it limits our author confessedly 
for accurate work, between the dates 1854 and 1873, whereas, 
the direct method is capable of including all observations from 
1824 up to the present time. It therefore seems to me that the 
greater number of observations more than compensates for the 
greater accuracy of the few. And indeed I think this is indl- 
cated by the fact that we shall be able to draw quite a different 
conclusion from our author, though using the results obtained 
by his own method, while, by using all the observations as 10 
my former article, I think but one conclusion can be drawn. 
At the top of page 22 (July number), Mr. Nipher gives 4 
table, the third column of which qT he calculates from alter- 
nate differences in the first two columns. Now where there 
are so few observations I think this method is hardly of suffi- 
cient accuracy, and that it is somewhat better to divide the 
