At the outset, there was much uncertainty as to the effect of the curvature of the earth on the propagation of a Hertzian wave over a distance of many hundreds of miles. In the case of the Atlantic transmission between the station at Poldhu in Cornwall and that at Cape Cod in Massachusetts, U.S.A., we have two stations separated by about 45 degrees of longitude on a great circle, or one-eighth part of the circumference of the world. In this case, the versine of the arc or height of the sea at the half-way point above the straight line or chord joining the two places is 300 miles.
The question has recently attracted the attention of several eminent mathematical physicists. The extent to which a free wave propagated in a medium bends round any object or is diffracted depends on the relation between the length of the wave and the size of the object. Thus, for instance, an object the size of an orange held just in front of the mouth does not perceptibly interfere with the propagation of the waves produced by the speaking or singing voice, because these are from two to six feet in length: but if arrangements are made by means of a Galton whistle to produce air waves half an inch in length, then an obstacle the size of an orange causes a very distinct acoustic shadow. The same thing is true of waves in the ether. The amount of bending of light waves round material objects is exceedingly small, because the average length of light waves is about one-fifty-thousandth part of an inch. In the case of Hertzian wave telegraphy, we are, however, dealing with ether waves many hundreds of feet in length, and the waves sent out from Poldhu have a wave-length of a thousand feet or more, say, one-fifth to one-quarter of a mile. The distance, therefore, between Poldhu and Cape Cod is only at most about twelve thousand wave-lengths, and stands in the same relation to the length of the Hertzian wave used as does a body the diameter of a pea to the wave-length of yellow light. There is unquestionably a large amount of diffraction or bending of the electric wave round the earth, and, proportionately speaking, it is larger than in the case of light waves incident on objects of the same relative size.
Quite recently Mr. H. M. Macdonald (see Proc. Roy. Soc., London, Vol. LXXI., p. 251) has submitted the problem to calculation, and has shown that the power required to send given electric waves 3,000 miles along a meridian of the earth is greater than would be required to send them over the same distance if the sea surface were flat in the ratio of 10 to 3. Hence the rotundity of the earth does introduce a very important reduction factor, although it does not inhibit the transmission. Mr. Macdonald's mathematical argument has, however, been criticised by Lord Rayleigh and by M. H. Poincaré (see Proc. Roy. Soc., Vol. LXXII., p. 40, 1903).
The accomplishment of very long distances by Hertzian wave telegraphy is, however, not merely a question of power, it is also a question of wave-length. Having regard, however, to the possibility that the propagation which takes place in Hertzian wave telegraphy is not that simply of a free wave in space, but the transmission of a semi-loop of electric strain with its feet tethered to the earth, it is quite possible that if it were worth while to make the attempt, an ether disturbance could be made in England sufficiently powerful to be felt in New Zealand.
Leaving, however, these hypothetical questions and matters of pure conjecture, we may consider some of the facts which have resulted from Mr. Marconi's long-distance experiments. One of the most interesting of these is the effect of daylight upon the wave propagation. In one of his voyages across the Atlantic, when receiving signals from Poldhu on board the S.S. Philadelphia, he noticed that the signals were received by night when they could not be detected by day.[72] In these experiments Mr. Marconi instructed his assistants at Poldhu to send signals at a certain rate from 12 to 1 a.m., from 6 to 7 a.m., from 12 to 1 p.m., and from 6 to 7 p.m., Greenwich mean time, every day for a week. He has stated that on board the Philadelphia he did not notice any apparent difference between the signals received in the day and those received at night until after the vessel had reached a distance of 500 statute miles from Poldhu. At distances of over 700 miles, the signals transmitted during the day failed entirely, while those sent at night remained quite strong up to 1,551 miles, and were clearly decipherable up to a distance of 2,099 miles from Poldhu. Mr. Marconi also noted that at distances of over 700 miles, the signals at 6 a.m., in the week between February 23 and March 1, were quite clear and distinct, whereas by 7 a.m. they had become weak almost to total disappearance. This fact led him at first to conclude that the cause of the weakening was due to the action of the daylight upon the transmitting aerial, and that as the sun rose over Poldhu, so the wave energy radiated, diminished, and he suggested as an explanation the known fact of the dissipating action of light upon a negative charge.
Although the facts seem to support this view, another explanation may be suggested. It has been shown by Professor J. J. Thomson that gaseous ions or electrons can absorb the energy of an electric wave, if present in a space through which waves are being transmitted.[73] If it be a fact, as suggested by Professor J. J. Thomson, that the sun is projecting into space streams of electrons, and if these are continually falling in a shower upon the earth, in accordance with the fascinating hypothesis of Professor Arrhenius, then that portion of the earth's atmosphere which is facing the sun will have present in it more electrons or gaseous ions than that portion which is turned towards the dark space, and it will therefore be less transparent to long Hertzian waves.[74] In other words, clear sunlit air, though extremely transparent to light waves, acts as if it were a slightly turbid medium for long Hertzian waves. The dividing line between that portion of the earth's atmosphere which is impregnated with gaseous ions or electrons is not sharply delimited from the part not so illuminated, and there may be, therefore, a considerable penetration of these ions into the regions which I may call the twilight areas. Accordingly, as the earth rotates, a district in which Hertzian waves are being propagated is brought, towards the time of sunrise, into a position in which the atmosphere begins to be ionised, although far from as freely as is the case during the hours of bright sunshine.
Mr. Marconi states that he has found a similar effect between inland stations, signals having been received by him during the night between Poldhu and Poole with an aerial the height of which was not sufficient to receive them by day. It has been found, however, that the effect simply amounts to this, that rather more power is required by day than by night to send signals by Hertzian waves over long distances.
Some interesting observations have also been made by Captain H. B. Jackson, R.N.,[75] on the influence of various states of the atmosphere upon Hertzian wave telegraphy. These experiments were all made between ships of the British Royal Navy, furnished with Hertzian wave telegraphy apparatus on the Marconi system. Some of his observations concerned the [effect of the interpositon of land] between two ships. He found that the interposition of land containing iron ores reduced the signalling distances, compared with the maximum distance at open sea, to about 30 per cent. of the latter; whilst hard limestone reduced it to nearly 60 per cent. and soft sandstone or shale to 70 per cent. These results show that there is a considerable absorption effect when waves of certain wave-length pass through or over hard rocks containing iron ores. It would be interesting to know, however, whether this reduction was in any degree proportional to the dryness or moisture of the soil. Earth conductivity is far more dependent upon the presence or absence of moisture than upon the particular nature of the material which composes it other than water.
The observations of Captain Jackson, however, only confirm the already well-known fact that Hertzian waves, as employed in the Marconi system of wireless telegraphy, within a certain range of wave-length, are considerably weakened by their passage through land, over land or round land. In some cases he noticed that quite sharp electric shadows were produced by rocky promontories projecting into the line of transmission. His attention was also directed (loc. cit.) to the more important matter of the effect of atmospheric electrical conditions upon the transmission. The effect of all lightning discharges, whether visible or invisible, is to make a record on the telegraphic receiver. On the approach of an atmospheric electrical disturbance towards the receiving station on a ship, the first visible indications generally are the recording of dots at intervals from a few minutes to a few seconds on the telegraphic tape. Captain Jackson states that the most frequent record is that of three dots, the first being separated from the other two by a slight interval like the letters E I on the Morse code, and this is the sign most frequently recorded by distant lightning. But in addition to this, dashes are recorded and irregular signs, which, however, sometimes spell out words in the Morse code. He noted that these disturbances are more frequent in summer and autumn than in winter and spring, and in the neighbourhood of high mountains more than in the open sea. In settled weather, if present, they reach their maximum between 8 p.m. and 10 p.m., and frequently last during the whole of the night, with a minimum of disturbance between 9 a.m. and 1 p.m. Another important matter noted by Captain Jackson is the shorter distance at which signals can usually be received when any electrical disturbances are present in the atmosphere, compared with the distance at which they can be received when none are present. This reduction in signalling distance may vary from 20 to 70 per cent, of that obtainable in fine weather. It does not in any way decrease with the number of lightning flashes, but rather the reverse, the loss in signalling distance generally preceding the first indications on the instrument of the approaching electrical disturbance. It is clear that these observations fit in very well with the theory outlined above, viz., that the atmosphere when impregnated with free electrons or negatively-charged gaseous ions is more opaque to Hertzian waves than when they are absent. Captain Jackson gives an instance of ships whose normal signalling distance was 65 miles, failing to communicate at 22 miles when in the neighbourhood of a region of electrical disturbance. These effects in the case of wireless telegraphy have their parallel in the disturbances caused to telegraphy with wires by earth currents and magnetic storms.
Another effect which he states reduces [the usual maximum signaling] distance is the presence of material particles held in suspension by the water spherules in moist atmosphere. The effect has been noticed in the Mediterranean Sea when the sirocco wind is blowing. This is a moist wind conveying dust and salt particles from the African coast. A considerable reduction in signalling distance is produced by its advent.