April 22, 1909] 



NA TURE 



235 



Prof. J. A. Fleming, of the University of London. The 

 transmitter at Poldhu was similar in principle to the one 

 I have already described, but it is obvious that the con- 

 siderable distance over which it was proposed to transmit 

 signals necessitated the employment of more powerful 

 electromagnetic waves than those ever previously used. 

 These were obtained by means of a generating plant con- 

 sisting of an alternator capable of an output of about 

 25 kilowatts, which, through suitable transformers, charged 

 a condenser having a glass dielectric of great strength. 



Time does not permit me to describe in detail all the 

 engineering difficulties which were encountered in control- 

 ling electrical oscillations of a power which at that time 

 was certainly unprecedented, and as the tests were made 

 possible by commercial organisation, the objects of which 

 do not consist solely in the advancement of science, you 

 will understand that a detailed description of the plant used 

 at the Transatlantic stations cannot, for the present at 

 least, be made public. 



My early tests on wireless transmission by means of the 

 elevated capacity method had convinced me that when 

 endeavouring to extend the distance of communication it 

 was of little utility merely to increase the power of the 

 electrical energy applied to the transmitting circuits, but 

 that it was also necessary to increase the area or height 

 of the transmitting and receiving elevated conductors. 



As it was economically impracticable to use vertical 

 wires of very great height, the only alternative was to 

 increase their size or capacity, which, in view of the facts 

 I had first noticed in 1895, seemed likely to make possible 



FIC. 5 



the efficient utilisation of large amounts of electrical 

 energy (Journ. Inst. Elec. Eng., xxviii., 1899, pp. 278-9). 



The form of aerial which I first proposed to employ 

 consisted of a conical arrangement of wires insulated at 

 the top and gathered together at a lower point in the form 

 of a funnel. This aerial was supported by a ring of twenty 

 masts each 200 feet high, arranged in a circle 200 feet 

 in diameter. 



During the first tests an arrangement of circuits (Fig. 5) 

 proposed by Dr. Fleming, and consisting of a modification 

 of the system shown in P'ig. 3, was employed. In this 

 arrangement, in place of one high-frequency oscillation 

 circuit, two are employed, and the constants of the two 

 circuits are so arranged that very high-tension discharges 

 can be obtained from one of the condensers — the one Vifhich 

 is inductively connected with the aerial — without danger of 

 damage to the circuits of the generator (" The Principles 

 of Electric Wave Telegraphy," 1906, p. 506). 



Simultaneously with the construction of the station at 

 Poldhu, the erection of another one on substantially the 

 same plan was undertaken at Cape Cod, in the United 

 States of America. 



The completion of the arrangements was delayed owing 

 to a storm, which wrecked the masts and aerial at Poldhu 

 on September 18, iqoi, but by the end of November the 

 aerial was sufliciently restored to enable me to complete 

 the preliminary tests which I considered necessary prior 

 to making the first experiment across the Atlantic. 



Another accident to the masts at Cape Cod seemed likely 

 to postpone the tests for several months more. I therefore 

 decided that in the meantime I would use a purely tem- 



KO. 2060, VOL. 80] 



porary receiving installation in Newfoundland for the pur- 

 pose of testing how far the arrangements in Cornwall had 

 been conducted on right lines. 



The transmitting elevated conductor employed at Poldhu 

 during the experiments with Newfoundland consisted of 

 fifty almost vertical copper wires supported at the top by 

 a horizontal wire stretched between two masts 48 metres 

 high and 60 metres apart. These wires converged together 

 at the lower end in the shape of a large fan, and were con- 

 nected to the transmitting instruments situated in a build- 

 ing (Fig. 6). 



The transmitting condenser used with this aerial had a 

 capacity of one-fiftieth of a microfarad, and was charged 

 to a potential sufficient to produce a suitable spark dis- 

 charge between spheres 3 inches in diameter, \\ inches 

 apart, the wave-length being 1200 feet. The actual power 

 employed for the production of the waves was about 

 15 kilowatts. 



I left for Newfoundland on November 27, 190 1, with two 

 assistants. As it was impossible at that time of the year 

 to set up a permanent insta,Ilation with poles, I decided 

 to carry out the experiments by means of receivers con- 

 nected to elevated wires supported by balloons or kites — a 

 system which I had previously used when conducting tests 

 across the Bristol Channel for the Post Office in 1897.' 



It will be understood, however, that when it came to 

 flying a kite on the coast of Newfoundland in the month 

 of December this method was neither an easy nor a com- 



FIC.G 



fortable one. When the kites were got up much difficulty 

 was caused by the variations of the wind producing con- 

 stant changes in the angle and altitude of the wire, thereby 

 causing corresponding variations in its electrical capacity 

 and period of electrical resonance. My assistants at Poldhu, 

 in Cornwall, had received instructions to send on and 

 after December 11, during certain hours every day, a 

 succession of S's followed by a short message, the whole 

 to be transmitted, at a certain pre-arranged speed, every 

 ten minutes, alternating with five minutes' rest. 



Owing to the constant variations in the capacity of the 

 aerial wire in Newfoundland, it was soon discovered .nat 

 an ordinary syntonic receiver was not suitable, although, 

 at one time, a number of doubtful signals were recorded. 

 1 therefore tried various microphonic self-restoring coherers 

 placed either directly in the aerial or included in the 

 secondary circuit of an oscillation transformer, the signals 

 being read on a telephone. 



On December 12 the signals transmitted from Cornwall 

 were clearly received, at the pre-arranged times, in many 

 cases a succession of S's being heard distinctly, although 

 probably in consequence of the weakness of the signals 

 and the' constant variations in the height of the receiving 

 aerial no actual message could be deciphered. The follow- 

 ing day we were able to confirm the result. The signals 

 were actually read by myself and by my assistant, Mr. 

 G. S. Kemp. 



1 ".Signalline throueVi Space without Wires." lecture by Sir William 

 Preece, Royal Institution, June 4, 1S97. Proc. R.I., xv., p. 467. 



