at the York Meetivg, I84S. 
413 
applied at the third trial, all doubt vanished. jNIr. Cambrid^re's 
man believed that the power alone would be rejjarded in the third 
trial ; and accordinjrly fed so moderately that one man less was 
required to open sheaves for him than in the previous trials. The 
result was tli.it the machine now took Gniin. 4 sec. to thrash 
100 sheaves, though it had previously thrashed the same quantity 
in 4 miu. 12 see, so that at ilils last trial it was beaten in point of 
time by both Garrett's and Ilornsby's; yet, in spite of this moderate 
feedinnf, the power required to drive Cambridge's machine was 
represented by the number 33^, which was an increase of 17 per 
cent, upon that required by Garrett's, without the shaker, \\hich 
was onlv 28^. The only fair way of comparing machines of this 
description is by ascertaining the amount of work performed by 
each when driven bv the same amount of power for the same time ; 
and, when subjecled to this test, the three best machines stood in 
the following order : 1. Garrett's; 2. Hornsby's ; 3. Cambridge's. 
One defect was observable in most of the thrashing-machines 
exhibited at York, viz. the small diameter of the horse-walk. 
There are two reasons which make it important that the horse- 
walk should be large, the first of which is, that a horse cannot use 
his full strength when moving in a small circle, as he is placed in 
a constrained and unnatural position, which prevents his throwing 
his weight properly into his collar. The second reason is that in 
a small horse-walk the power is applied at a meclianical disad- 
vantage. The power would be most efficient if applied in the di- 
reciion of a tangent to the circle, and the smaller the circle is, the 
more does the line of draught fall within this line, and the greater 
therefore is the loss of power. Practically it is found that a 
horse never does his work with ease if he has to walk in a circle 
of less diameter than 24 feet. It is true that more time is required 
to make the circuit of a large horse-walk than of a small one, so 
that the horse-wheel makes fewer revolutions per minute, but 
this is exactly counterbalanced by the increased diameter of the 
circle, which makes the arms by which the horses draw propor- 
tionably longer. These arms act as levers, and by increasing 
their length an increase of power is obtained, which admits of the 
horse-wheel being so much enlarged that its additional number 
of cogs restores the speed lost by its making fewer revolutions: 
thus the balance between the power and speed of the machine is 
unaffected by the change, and its efficiency is increased both from 
the greater freedom of action given to the horses and the increased 
angle at which the power is applied. The only set off against 
these advantages is, that (as has been shown above) a larger horse- 
wheel is required. There is, however, another mode by which 
the speed may be regained, viz. by diminishing the size of the 
pinion which is driven by the horse-wheel, or by introducing an 
