GRAPHIC RECORDING OF SPEECH VIBRATIONS 
87 
ment of a compound lever.^’i^ He believes this inadequacy is 
due to the low degree of magnihcation obtainable (125 to 300 
times). But a comparison of his results with those of Hermann, 
Rosset, Marichelle and Miller, shows that the failure of Scrip- 
ture’s apparatus to record consonantal and overtone vibrations 
is due to its low limit of sensitiveness and not to its low degree 
of magnihcation. Thus Marichelle shows the occlusive and 
explosive phases of voiceless consonants with a magnihcation of 
less than thirty diameters.^^ Miller’s record of Lowell Institute 
as recorded in his Science of Musical Sounds^^ has a magnihca- 
tion only a little greater than that given by Scripture’s appa- 
ratus, but it plainly shows the vibrations of even the occluded 
phases of s and t. 
Muenzinger’s compound lever apparatus neutralizes the 
weight of all levers through counterbalancing, and practically 
eliminates play and friction of bearings and joints by point or 
V-bearings. His hrst results do not exceed those obtained with 
Scripture’s simple lever apparatus. Muenzinger’s apparatus 
ought to be further developed and perfected in order to deter- 
mine its full possibilities. Apparently any further progress in 
physical lever magnihcation of the sound groove must follow 
the lines that he has laid down. The future of the method must 
lie, not in seeking greater magnihcation as Scripture supposed, 
but in seeking greater sensitiveness, perhaps with lower magnih- 
cation if that is necessary to secure it. The absorption of 
minute vibrational phases (due in Scripture’s apparatus to play 
of bearings, inhnitesimal bending of levers, and friction of the 
recording stylus) must be eliminated. Muenzihger’s V-bearings 
will probably solve the hrst difficulty. The bending can perhaps 
be eliminated by using very light levers with delicate trussing 
or wire bracing, which are kept at a hxed temperature; the 
method probably cannot be perfected unless such rigid and 
delicate levers can be obtained. For the recording surface some 
■ Study of Speech Curves, p. 26. 
La parole d’apres le trace du phonographe, figs. 83-86, p. 80. 
Figure 184, p. 254. 
