If a man should stand in room 1 at A, he will be able to see a friend standing in room 4 at B, but the two friends will not be able to converse. When A speaks, the sound waves that he produces will spread out and will fill room 1. A very small percentage of them will strike the doorway or opening into room 2. In their turn these sound waves will be diffused all through room 2, and again but a small percentage of them will find access into room 3. The sound waves will by this time be so much attenuated that the voice of the man standing in room 1 will be lost. Any little tone, however, that may remain will become dissipated in room 3, and it will not be possible for a person standing in room 4 to hear the voice. [Illustration: Fig. 12. The Principle of the Sound Trap] This plan illustrates the principle of the sound trap joint. Figure 13 shows in section the joint between two Swell shutters. A small proportion of the sound waves from inside the Swell box striking the sound trap joint, as indicated by the arrow, will pass through the nick between the two shutters, but these sound waves will become greatly weakened in charging the groove A. Such of the sound waves is pass through the second nick will become attenuated in charging the chamber B. They will be further lost in the chamber C, and practically none will remain by the time the chamber D is reached. It is Hope-Jones' habit to place the shutters immediately above the pipes themselves, so that when they are opened the Swell box is left practically without any top. It is in such cases not his custom to fit any shutters in the side or front of the Swell box. [Illustration: Fig. 13. Sound Trap Joint] To relieve the compression of the air caused by playing for any length of time with the shutters closed, he provides escape valves, opening outside the auditorium. He also provides fans for driving all the cold air out of the box before using the organ, thus equalizing the temperature with the air outside--or he accomplishes this result through the medium of gas, electric or steam heaters, governed by thermostats. The Hope-Jones Vacuum Swell Shutters, with sound-trap joints, are shown in Figures 14 and 15. It is well known that sound requires some medium to carry it. Readers will doubtless be familiar with the well-known experiment illustrating this point. An electric bell is placed under a glass dome. So long as the dome is filled with air the sound of the bell can be heard, but directly the air is pumped out silence results, even though it can be seen that the bell is continuously ringing. As there is no air surrounding the bell there is nothing to convey its vibrations to the ear. That is why the hollow swell shutter, from the interior of which the air has been pumped out, is such a wonderful non-conductor of sound. The shutters shown in Figures 14 and 15 are aluminum castings.