DRS. GUY BARLOW AND H. B. KEENE ON THE ANALYSIS OF SOUND. 145 
S3 
1 
3b 
■§ 
resonance with the microphone. The subharmonic series of 1052 was in consequence 
pronounced. Thus in the experiment a marked response 
was observed at 151, which is 1052 -f- 7 nearly, i.e. the 
seventh subharmonic of the fourth harmonic of the funda¬ 
mental vibration. In practice the presence of the sub¬ 
harmonics is generally of less inconvenience than this 
example might suggest, but it will be seen that analysis 
is particularly affected if the receiver possesses strong 
resonance-points very much above the region of frequency 
under analysis. 
(3) Complex Sound .—In order to test the reliability of 
the analysis, it was thought desirable to produce a complex 
sound containing known constituents each of which could 
be examined separately. Four electrically maintained forks 
were placed on the same table, and grouped in front of the 
microphone receiver suspended on rubber cords indepen¬ 
dently of the table. The four forks selected had the 
following frequencies:—A 71-3, B 89, C 100, D 261/sec. 
With all forks sounding together a careful analysis was 
first made, in which each response was examined separately 
in order to determine accurately its magnitude and 
frequency. The result of this analysis is shown in fig. 7. 
It will be observed that the following frequencies were 
present:—- 
(i.) The four fundamental vibrations A, B, C and D. 
(ii.) Weak octave of C. Those of A and B were not 
observed, while the octave of D was out of range. 
(iii.) All the important subharmonics of A, B, C and D, 
except in those cases where they were masked by 
other responses, e.g., AD was missed, as it was 
nearly coincident with the fundamental B. 
With the exception of uncertain responses at 158 and 
161 /sec., all frequencies found in the analysis are accounted 
for. The frequency 161 may be due to a combination- 
tone, either D — C — 161, or A -f- B = 160-3. A separate 
investigation showed that this response only occurred when 
the forks A and B were sounding together, and that it was 
not always present even under those conditions. 
The analysis was now repeated by making a rapid sweep 
wo—I 
■V'5 — 
"*s> SK — 
Sip - 
- 
S' 
c> 
o 
k 
■oo 
<5 
Tat 
if 
jo 
s 
<0 
ii 
S 
u 
U 
l 
o 
u- 
CM 
o 
d 
o 
m 
7 
LO 
w 
*5) 
cZ 
0) 
bt) 
G 
cZ 
PS 
3 -3 
f 
056 
a 
50 
s 
of the whole range by continuously increasing the speed of interruption without stopping 
