NEST GEARING. 389 
to remedy these defects in future designs. The pressure on the surfaces 
is such as allows one inch of breadth to transmit about 200 lbs. of force, 
but we do not yet know the wear under these conditions. The smallest 
of the rollers is 3 ins. in diameter. 
T am now inclined to think that the excentric method of tightening 
will prove less convenient than other plans, some of which will be 
presently described. The framing and system of levers for the excentric 
method are complex, and it is found necessary (in order to avoid useless 
friction) to slack back the tightening screw a little after each fresh 
adjustment of the tightening roller. This prevents the advantageous 
use of a spring to take up the wear by constantly forcing the tightening 
roller into the narrow part. I prefer the method of tightening in which 
two external reverse coned surfaces are pinched between two internal 
coned surfaces, as in fig. 5, but when tightening is effected in this way the 
surfaces of the two convex rollers, as at A and B, should be flat cylindrical 
surfaces rolling true together. 
Messrs. Ayrton and Perry have constructed a nest (not exhibited 
here) in which the pinching is effected by an ingenious arrangement of 
the outer nest ring, and in this nest the ring itself is stationary, while 
the nest of intermediate rollers, with their frame, revolves. These 
experiments have in certain cases given an apparent efficiency of 97 per 
cent. for this nest, multiplying the angular velocity nine times. The 
grinding action due to the coned surfaces is very much less than in the 
old forms of friction gear, and is indeed insignificant, as will be seen 
from a simple calculation of the difference of velocities between the mean 
circumference and the extreme on either side. When the breadth is 
say 4 in., the taper 1 in 5, and the diameter of the two rollers in gear 
say 18 ins. and 8 ins. respectively, the rubbing action of the two wheels 
rolling together at the mean diameter will, at the extreme edge of the path, 
be due to a gain or loss of ‘0625 in. in 18 ins., or about 0°348 per cent. 
Whatever be the coefficient of friction, if the pressure were only just 
sufficient to prevent gripping, this would entail a loss of only 0°348 per 
cent. in the power transmitted. It would actually be somewhat larger 
than this, but still insignificant where the breadth of bearing is small 
compared with the diameter, and where the intermediate rollers are 
large. I myself prefer to get the pinching by causing one or more 
intermediate rollers to expand lengthways. In that case only one inter- 
mediate roller need grind or bear on the cones. The two others may run 
on the flat surface between the cones.! 
In fig. 6 will be seen a modification which I believe to be wholly 
novel. In this arrangement we have a multiplication in a duplex ratio. 
The two parts D and D, of the nest ring may be joined or they may be 
wholly separate. The modes of tightening may be any of those described. 
The gear so far described has been suited for joining shafts which have 
their axes in one straight line, or in parallel lines which are not far apart. 
I have called it concentric nest gear. I next pass to right-angled 
nest gear, the first form of which was due to a suggestion of my son, 
C. Frewen Jenkin. In its simplest form this gear is shown in fig. 7, where 
the contact between the surfaces of D D,, the nest discs, with B and B, the 
- intermediate rollers, is confined to a single point. In this figure the 
1 Since reading this paper at Southport I have patented a plan in which, by inclining 
the two sides of this expanding roller, all the coned surfaces run true on each other. 
