146 BELL SYSTEM TECHNICAL JOURNAL 



was decided to construct a model of the proposed antenna system on 

 this plot and to operate small transmitting sets to determine the 

 cross-talk and other important conditions. The antenna system 

 decided upon consisted of three poles arranged in the form of an 

 equilateral triangle supporting three antennas — one from the middle 

 of each span to the transmitting shack at the center of the triangle. 

 The dimensions of this model system were 50 meters (164 ft.) by 10 

 meters (33 ft.) high. Three experimental transmitting sets of small 

 power were set up under the antenna system and studies were made 

 of the interference produced between channels when all three channels 

 were in use. Three receiving sets of the general form proposed were 

 built and taken to a location near Elberon, New Jersey, about 16 miles 

 (26 km.) from ClifTwood, at which place it was decided to locate the 

 three-channel receiving station to co-operate with the New Jersey 

 transmitting station a mile (1.6 km.) away. 



In November, 1919, the first test of a three-channel system was 

 held between Cliffwood and Elberon with the result that the receiving 

 sets resolved conversations on carriers of frequencies of 725, 750, and 

 775 kilocycles without any cross-talk altho the received volume was 

 so large as to be audible all over the room. This is a frequency 

 difference between channels of approximately three per cent. A 

 change in frequency to 747, 759, and 777 kilocycles resulted in a 

 barely perceptible cross-talk on the middle channel, with no cross- 

 talk on the others. These results indicated that the loop receivers 

 which had been developed were sufficiently sensitive and selective 

 to carry out the proposed three-channel work; and, altho a great deal 

 of development work was done later on the receiving sets, the general 

 principles were retained. It was found that some reliance must be 

 placed upon the directional properties of the loop antennas, and con- 

 siderable care was used to secure very sharp directional selectivity. 

 This was done by compensating for the vertical antenna effect of the 

 loop by a balanced connection to ground. 



During the whole course of this ship-to-shore work very little 

 trouble was experienced thru interference by continuous wave sta- 

 tions, even when their frequencies came within two or three per cent, 

 of those to be received. We did, however, have much difficulty due 

 to interference from spark stations, since they inherently occupy a 

 wide frequency range. 



Provision of Stations and Development of Apparatus 



During the time the model system was being constructed at Cliff- 

 wood, land had been purchased at West Deal, Monmouth County,. 



