of Edinburgh, Session 1877-78. 
737 
low notes than with high — because, though the rotating apparatus 
always produces a siren effect, the intensity of this effect is very 
small at low speeds. A very curious case occurs when m = n (i.e., 
the siren giving the same note as the pipe), for then the first har- 
monic comes in very marked. 
8. Note on a Mode of Producing Sounds of very great 
intensity. By Professor Tait. 
Two years ago I had an opportunity of making from the deck of the 
steamer “ Pharos” some observations on the performance of the fog- 
siren at Sanda, off the Mull of Cantire. The instrument is worked 
by air at about 1| atmospheres pressure; and, though driven by a 
powerful air-engine, sounds for 7 seconds only per minute. One 
obvious defect of such an arrangement I saw to be the waste of 
energy in producing a current of air through the trumpet of the 
siren along with the oscillations. It then occurred to me that a 
regular alternation of puffing and sucking — exactly analogous to the 
air-disturbance produced by a drum — must be a much less costly 
source of sound. I have since constructed a siren on this double 
action principle, the air in the trumpet, which acts as a resonator, 
being put alternately in connection with reservoirs of compressed 
and rarefied air. The small model has given very good results, and 
a larger one is in progress. The only defect which my model 
showed was a waste of energy in the form of pulsations in the 
tubes leading to the exhausted receiver and to that containing 
compressed air. This can be very greatly reduced, but I do not 
yet see how to get rid of it entirely, unless it be possible to make 
both receivers so exactly as to act as additional resonators to the 
siren. If this can be carried out in practice there will be no 
energy spent except in sound. It is obvious that the principle just 
described is approximated to in practice whenever steam is employed 
in a siren : — the vacuum being produced by the condensation of the 
steam. 
Another device of a somewhat different character was suggested 
to me by the experiments described in the preceding paper. After 
trying, without much success, to reduce the intensity of the siren 
notes by filing the edges of the apertures, it occurred to me that I 
