February 9, 1922] 



NATURE 



181 



itinue to collect the facts and trust to cautious 



iductive reasoning and observations to give us the 



rue interpretation of them. All the phenomena 



;em, however, to point to the existence of three 



iperimposed layers in the atmosphere : one, the 



ligher, beginning perhaps above 80-100 km., 



irhich is permanently ionised with negative 



IS. The other, the middle, which has in part 



iriable ionisation, depending on the position of 



lat part with regard to the sun. The third or 



lower level has a relatively small ionisation, but 



jlectromagnetic waves travelling in it may have their 



lergy considerably affected and reduced by the 



iture of the eartli's surface over which they are 



loving. Powerful absorption is caused by some 



)ils and by vegetation for certain wave-lengths. 



From the earliest days of long-distance wireless 

 jlegraphy the difficulties in reception due to vagrant 

 natural electric waves and atmospheric electric 

 lischarges passing down the receiving aerial have 

 ;en the bane of the wireless telegraphist. These 

 raves create sounds in the telephone in aural recep- 

 tion which often drown completely the signal sounds 

 and make false records in the case of printing or 

 photographic reception. In the case of telephone 

 reception, these noises have been classified into 

 (i) rattling or grinding, (2) hissing, (3) clicking or 

 snapping, and (4) crashing noises. These last two 

 seem to be associated with thunderstorm conditions. 

 Ha\ing regard to the fact that the positive atmo- 

 spheric electric potential gradient of the earth in- 

 creases at the rate of abou"- 100 volts per metre of 

 ascent, it is not surprising that aerials several 

 hundred feet high may be traversed by quite 

 large currents, due to this cause alone, which 

 may utterly swamp the feeble signal currents. The 

 strength of a signal or noise in the telephone is 

 generally estimated by its " audibility," and this 

 is measured by ascertaining the resistance S of the 

 shunt which must be put across the telephone of 

 resistance R just to render the sound inaudible to a 

 normal ear. The audibility A is given by the ex- 

 pression A = (R-f S)/S. Hence the audibility is 

 unity for a just audible sound. We can in this 

 way measure the audibility of a signal on a back- 

 ground of disturbing noise, and a readable signal 

 is generally obtained if the ratio of signal audi- 

 bility to stray audibility is more than 25 per cent. 



Owing to the serious extent to which these strays 

 hinder regular reception, especially at certain times 

 of the day and year, an enormous amount of atten- 

 tion has been given to their study and to the problem 

 of eliminating them. They are most troublesome in 

 the summer and during the night, and more severe 

 in tropical than in temperate climes. Even in our 

 latitudes they hinder reception at times immensely. 

 Dr. L. W. Austin has stated that receiving at Wash- 

 ington, U.S.A., with a simple loop aerial from high- 

 power radio stations in Europe with aerial sending 

 currents up to 300 amperes, signals were unreadable 

 for about 2000 hours a year. Tn tropical countries 

 over long-distance circuits the power required to get 

 a signal through may be often from six to eight 

 times that which must be used at favourable times, 

 NO. 2728, VOL. 109] 



and there are short periods when signalling is abso- 

 lutely impossible. Having regard to the effect such 

 interruptions have upon the earning power of a com- 

 mercial station or upon certainty of communication 

 in time of war or other urgent occasions, the problem 

 of elimination of strays is perhaps the most impor- 

 tant of all the practical questions connected with 

 long-distance wireless telegraphy. It has been the 

 subject of countless patents already. Early attempts 

 went on the supposition that the strays w^ere highly 

 damped vagrant waves or had particular frequencies 

 and could be eliminated by giving the receiving 

 system a very pronounced resonance and making it 

 a socalled stiffly-tuned circuit. These methods had 

 a very limited application, for the reason that any 

 impulse given to the receiving aerial sets it in electric 

 vibration with its own natural period. Then, again, 

 a number of inventions depend upon the peculiar 

 properties of certain detectors, such as crystals and 

 thermionic valves, in limiting the current which they 

 pass or rectify. One most practically useful dis- 

 covery was that by giving to the spark or wave train 

 in the case of spark systems, or to the beats in the 

 case of C.W. heterodyne reception, a regular fre- 

 quency of 500 or 600, thus imparting a rather shrill 

 musical sound to the signal, the ear could much 

 more readily fasten attention on it even against a 

 background of irregular but louder noise due to 

 atmospherics. 



Dr. de Groot made an immense number of ob- 

 servations on stray strengths at various hours of the 

 day and months of the year about 1916 in the Dutch 

 East Indies, and prepared diagrams showing the 

 mean stray strength for various hours for each 

 month of the year. From these he prepared a 

 diagram giving the hourly stray strength during the 

 day averaged throughout a year. The results were 

 that in general the strays were more numerous 

 and stronger during the night than during the 

 day. 



We have seen that there must be a certain inter- 

 mediate but high-level region in the atmosphere in 

 which the gases are ionised by the ultra-violet sun- 

 light during the day, but re-combine again during 

 the night. This region lies beneath the permanently 

 ionised layer. In this permanently ionised layer 

 there are drifting collections or masses of positively 

 electrified and negatively electrified solar dust. If 

 these masses are drawn together by their electric 

 attractions or commingled, it is highly probable that 

 electric recombinations will occur, which would 

 generate electric waves. Suppose, then, that we 

 assume the origin of a certain part of the strays to 

 be in the upper permanently conductive layer of 

 the atmosphere, these natural waves would find a 

 certain obstacle to their downward transmission in 

 the conductivity produced by the ionisation of the 

 middle layer of the atmosphere by day. But at 

 night-time this middle layer ionisation largely dis- 

 appears and the natural electric disturbances in the 

 upper layer would more easily find their wav down 

 to the earth. Tn other words, there would be a 

 more unhindered access for the strays to descend. 

 , Hence in the night-time they would be more numerous 



