56 
The thread of “raw” or reeled silk is excessively strong, ductile, and 
elastic. As has been seen, it is- composed of several double filaments, 
drawn from as many cocoons. In common with other elastic threads, 
a given length of one of silk will resist a tendency to stretch to an ex- 
tent proportionate to its mean section. This is the un derlying princi- 
ple of the serigraph. The mode of determining the irregularities exist- 
ing in a thread of raw silk by means of this machine is as follows : The 
end of the thread is brought from the reel or bobbin on which it is wound 
around a drum, 8, (Plate I), thence over a pulley, R, and back around 
another drum, T, mounted on the same axis as S. From the drum T it 
is wound on a reel. The drum T is larger than 8, so that the former 
winds on the thread somewhat faster than it is paid off by the latter, 
and thus stretches it. In this manner we apply a constant force to the 
pulley _R, tending to draw it from its normal position. This pulley is 
attached to the base of a pendulum, U , which, under the action of the 
force mentioned, is drawn from the perpendicular. The weight of this 
pendulum overcoming the force thus applied to an extent inversely pro- 
portional to the mean section of the length of thread submitted to the 
test, the position of equilibrium taken by the pendulum depends upon 
and is an indication of that mean section. The portion thus tested is 
that between the two drums S and T , and as, through the constant 
action of the machine, successive lengths of thread occupy the position 
indicated, the pendulum oscillates through a course which depends upon 
the irregularities of the thread. These irregularities are graphically 
recorded by a pencil, attached to the pendulum, upon a band of paper, 
which moves constantly under its point. 
The serigraph, it will be seen, is an apparatus for continuously meas- 
uring the relative size of any thread passed over its drums and record- 
ing the irregularities in its size on a band of paper. 
From this machine to the automatic reeler was but a slight transition, 
easily accomplished. It has been in working out the details of the de- 
sired mechanism that the greatest difficulty has been met with. The 
result is attained in general by causing the pendulum TJ to close an 
