RADIOTTCLEGRAPHY MARCONI. 127 



During these tests the receiving wire was supported by means of a 

 kite, as was done in my early transatlantic tests of 1901, the height 

 of the kite varying from about 1,000 to 3,000 feet. Signals and mes- 

 sages were obtained without difficulty, by day as well as by night, up 

 to a distance of 4,000 statute miles from Clifden. 



Beyond that distance reception could only be carried out during 

 nighttime. At Buenos Aires, over 6,000 miles from Clifden, the 

 night signals from both Clifden and Glace Bay were generally good, 

 but their strength suffered some variations. 



It is rather remarkable that the radiations from Clifden should 

 have been detected at Buenos Aires so clearly at nighttime and not 

 at all during the day, whilst in Canada the signals coming from Clif- 

 den (2,400 miles distant) are no stronger during the night than they 

 are by day. 



Further tests have been carried out recently for the Italian Gov- 

 ernment between a station situated at Massaua in East Africa and 

 Coltano in Italy. Considerable interest attached to these experi- 

 ments, in view of the fact that the line connecting the two stations 

 passes over exceedingly dry country and across vast stretches of 

 desert, including parts of Abyssinia, the Soudan, and the Libyan 

 Desert. The distance between the two stations is about 2,600 miles. 



The wave length of the sending station in Africa was too small to 

 allow of transmission being effected during daytime, but the results 

 obtained during the hours of darkness were exceedingly good, the 

 received signals being quite steady and readable. 



The improvements introduced at Clifden and Glace Bay have had 

 the result of greatly minimizing the interference to which wireless 

 transmission over long distances was particularly exposed in the 

 early days. 



The signals arriving at Clifden from Canada are as a rule easily 

 read through any ordinary electrical atmospheric disturbance. This 

 strengthening of the received signals has moreover made possible the 

 use of recording instruments, which will not only give a fixed record 

 of the received messages, but are also capable of being operated at a 

 much higher rate of speed than could ever be obtained by means of 

 an operator reading by sound or sight. The record of the signals is 

 obtained by means of photography in the following manner: A sen- 

 sitive Einthoven string galvanometer is connected to the magnetic 

 detector or valve receiver, and the deflections of its filament caused 

 by the incoming signals are projected and photographically fixed on 

 a sensitive strip, which is moved along at a suitable speed (pi. 1, fig. 2). 

 On some of these records, which I am able to show, it is interesting 

 to note the characteristic marks and signs produced amongst the 

 signals by natural electric waves or other electrical disturbances of 



