April 29, 1909J 



NA TURE 



= 65 



Poldhu by day as well as by night, but no commercial use 

 of the station was made at that time, in consequence of 

 tfie fact that, although the signals came through by day 

 as well as by night, they were exceedingly weak and faint, 

 and also because the corresponding station on the same 

 plan had not yet been erected in Ireland. 



A further step in advance was the adoption at the Trans- 

 atlantic stations of the directional aerial shown in Fig. 12.' 

 The ordinary wireless telegraph aerials, which 1 have 

 already described, send out electric radiation equally in all 

 directions. This is, however, in many cases a disadvantage. 

 Many suggestions respecting methods for limiting the direc- 

 tion of radiation have been made by various workers, 

 notably by Messrs. Artom, Braun, and Bellini Tosi. 



In some of my earliest experiments, in 1S96, I used 

 copper mirrors, by the aid of which it was possible to 

 project a beam of electric radiation in a certain direction, 

 but I soon found that this method would only work over 

 short distances. 



About three years ago I again took up the subject, and 

 was able to determine that by means of horizontal aerials, 

 disposed in a particular manner, it was possible to confine 

 the effects of electric waves mainly to certain directions as 

 desired. True, the limitation of transmission to one direc- 

 tion is not very sharply defined, but it is nevertheless very 

 useful. The practical result of this method has been, so 

 far, that messages can be sent over considerable distances 

 in the desired directions, while they travel only over a 

 comparatively short distance in other directions, and that, 

 with aerials of moderate height, greater efficiency in a given 

 direction can be obtained than can be obtained all round 

 by means of the ordinary aerials. 



When this type of aerial was adopted at Glace Bay a 

 considerable strengthening of the received signals at Poldhu 



was noticed. It was therefore decided to adopt the direc- 

 tional aerial at all long-distance stations. 



A further improvement introduced at Clifden and Glace 

 Bay consisted in the adoption of air condensers, composed 

 of insulated metallic plates suspended in air at ordinary 

 pressure. In this manner it is possible to prevent the dissi- 

 pation of energy due to losses caused by the dielectric 

 hysteresis in the glass dielectric of the condensers previously 

 employed, and a very appreciable economy in working, 

 resulting from the absence of breakages of the dielectric, is 

 effected. These air condensers, which have been in use 

 since May, 1907, have been entirely satisfactory. After 

 very considerable delay and expense, the new station at 

 Clifden was got ready for tests by the end of May, 1907, 

 and experiments were then commenced with Glace Bay. 



The wave-length used during these tests was 12,000 feet, 

 the capacity employed i-6 microfarads, and the potential to 

 which the condenser was charged 80,000 volts. 



Good signals were obtained at Cape Breton from the 

 very commencement of the tests, but some difficulty was 

 encountered in consequence of the effects of atmospheric 

 electricity due to the prevalence of thunderstorms in the 

 eastern part of Canada during the first few days of the 

 tests. 



Simultaneously with these tests others were carried out 

 from Poldhu to Glace Bay with a new system of trans- 

 mitting apparatus, by means of which continuous or semi- 

 continuous oscillations could be produced. 



Proportionately to the energy employed the signals from 

 Poldhu were so much better than those from Clifden that 

 I decided at once to adopt this new method of transmission 

 at Glace Bay and Clifden. The apparatus which I have 

 been using for producing continuous or closely adjacent 

 trains of electric oscillations is as follows - : — \ metal disc 



' " On Methods whereby the Radiation of Electric \ 



Confined. &c. Proc. Rny. Soc, C. Marconi, A. Ixxii 



- Patent .Application No. 20,iig, September g, 1907. 



NO. 2061, VOL. 80] 



1 may be mainly 



A (Fig. 13), insulated from the earth, is caused to rotate 

 at a very high speed by means of a high-speed electric 

 motor or steam turbine. Adjacent to this disc, which I 

 shall call the middle disc, are placed two other discs, c,, 

 Cj, which may be called polar discs, and which also can 

 be rotated at a high rate of speed. These polar discs 

 should have their peripheries very close to the surface or 

 edges of the middle disc. 



If a small amount of energy is used, stationary knobs 

 or points may be used in place of the side discs. 



The two polar discs are connected respectively through 

 suitable brushes to the outer ends or terminals of two con- 

 densers K, joined in series, and these condensers are also 

 connected through suitable inductive resistances to the 

 terminals of a generator, which should be a high-tension 

 continuous-current dynamo. 



On the high-speed or middle disc a suitable brush or 

 rubbing contact is provided, and connected between this 

 contact and the middle point of the two condensers is in- 

 serted an oscillating circuit consisting of a condenser E 

 in series with the inductance, which last is connected 

 inductively or conductively to the aerial. 



If the necessary conditions are fulfilled, and a sufficient 

 E.M.F. is employed, a discharge will pass between the 

 outer discs and the middle disc, which discharge is neither 

 an oscillatory spark nor an ordinary arc, and powerful 



FlC l;v 



oscillations will be created in the signalling condenser K 

 and oscillatory circuit F. 



I have found that in order to obtain good effects a peri- 

 pheral speed of more than 100 metres per second is desir- 

 able ; therefore particular precautions have to be taken in 

 the construction of the discs. Electrical oscillations of a 

 frequency as high as 200,000 per second can be obtained. 



The apparatus works probably in the following manner : — 

 Let us imagine that the source of electricity is gradually 

 charging the double condenser and increasing the potential 

 at the discs, say c, positively and c, negatively ; at a 

 certain instant the potential will cause the charge to jump 

 across one of the gaps, say between c, and A. This will 

 charge the condenser E, which will then commence to 

 oscillate, and the chavge in swinging back will jump from 

 A to c,, which is charged to the opposite potential. The 

 charge of E will again reverse, picking up energy at each 

 reversal from the condensers v.. The same process will go 

 on indefinitely, the losses which occur in the oscillating 

 circuit Et^ being made good by the energy supplied from 

 the generator 11. If the disc is not rotated, or rotated 

 slowly, an ordinary arc is at once established across the 

 small gaps, and no oscillations take place. The efficient 

 cooling of the discharge by the rapidly revolving disc seems 

 to be one of the conditions necessary for the production of 

 the phenomena. 



By means of this apparatus tests were carried out, but 

 it was found, as was to be expected, that the oscillations 

 were too continuous and of too high a frequency to affect 

 a receiver, such as the magnetic detector, unless an inter- 

 rupter was inserted in one of the circuits of the receiver. 



