262 SCIENCE PROGRESS 



The writer of this article has recently examined nearly 

 eighty instruments of different types, testing them not merely 

 subjectively, but by two physical methods of extreme delicacy, 

 the instrument being connected to an extremely sensitive 

 manometric flame, and also to a drum membrane pendulum 

 detector. Nearly all the instruments of classes B, C, and D 

 were found to be virtually useless ; it was only in class F that 

 some slight improvement was obtained. 



Class F constructionally does not differ from D. See fig. 5. 

 There is usually an air cavity A, short and wide, and resounding 

 especially to one note of fixed but generally high pitch, con- 

 nected by a flexible tube B to a monaural or binaural ear- 

 piece at the further end. The metal portion A is intended to 

 rest on a table. This class of instruments depends principally 

 on solid conduction, and requires that the table on which the 



A 



Fig. 5. 



receiver is placed should be of wood, not covered by a cloth. 

 So far as resonance of the air chamber of the receiver plays 

 any part in the mechanism, the criticism urged against the 

 preceding classes applies, viz. that the resonant air-chamber 

 is too small, and that its fundamental tone lies above the tones 

 usually employed in speech. 



On account of the complex shape the fundamental tone of 

 the air-chamber cannot be computed, though it may easily be 

 found experimentally ; but that the dimension is in general too 

 small may be judged from this, that to resound to the note 

 g' = 192, a spherical Konig receiver would have to be 15 cm. 

 diameter, or nearly 2 litres capacity. 



With regard to class G, we have here a class of instrument 

 dependent on the principle of the microphone, in which the 

 energy of the incident waves may, by supplementing it by the 

 energy of an electric battery, be increased any desired number 



