580 
Proceedings of the Royal Society 
dentations produced in this strip by a point attached to a vibrating 
disc stretched across a short brass cylinder 2|- in. diameter. 
The same vibrating disc was used to produce the indentations 
and to reproduce the sounds. 
The usual ferrotype plate had not been tried as a vibrator, for 
hard metal discs were found to give a disagreeable resonance to the 
voice and its reproduction. A slack tinfoil disc had been used with 
good results ; the marker was attached firmly to a small disc of 
stiff paper \ in. diameter, gummed to the tinfoil ; a short piece of 
watch spring was also attached to the marker, so as to give the disc 
a rapid period of vibration. 
A sentence reproduced by this disc was loud enough to be heard 
by many people standing round, and sentences had been heard by 
several persons who understood them without any previous idea of 
what they were. This result could not, however, always be secured. 
When the sentence was known to the hearers, it appeared to he 
given back with startling accuracy. The vowel sounds were more 
distinct than the consonants, hut the consonants also were dis- 
tinctly to be heard. The tinfoil vibrator gave more articulate 
sounds when slack and irregular than when it was neatly strained 
over the end of the tube. 
An oil-silk disc had also been tried with no spring, and a simple 
marker attached to a disc of mica, gummed to the oil-silk. This 
disc gave purer sounds than the tinfoil, but they were not nearly so 
loud. The indentations on the tinfoil were excellent to the eye, 
and quite as large as with the tinfoil. 
The oil-silk answered best when irregularly stretched. 
An india-rubber film, with a similar mica disc and rigidly attached 
marker, had also been tried, and gave beautiful records on the tin- 
foil strip, but this disc failed to reproduce the sounds accurately, 
having a note of its own. 
This latter disc was used to produce records which were subse- 
quently enlarged and shown on paper in the form of curves. The 
india-rubber was preferred for this purpose, because the gentle and 
uniform pressure which it gave did not tend to obliterate the records. 
The curves exhibited showed the form of the indentations magnified 
about five hundred times. This magnification was effected by two 
compound levers, of which the second was a glass siphon like that 
