March. 3, 1910J 



NATURE 



to flicker, and presently the flitkerings shaped themselves 

 into letters and words. The cable had awakened to life. 

 A few days more and it too was complete. 



Throughout the operations Thomson was in the ship ; 

 Varley remained at Valencia. Thanks to their labours, 

 and to those of Mr. Willoughby Smith, the contractors' 

 electrician, the appliances for testing on board ship had 

 been brought to a degree of perfection that left nothing 

 to be desired. By this time it was generally recognised 

 that the credit for Atlantic telegraphy, regarded as an 

 electrical achievement, belonged to Thomson, though in his 

 characteristic manner he would, when speaking of the 

 subject, dwell on the parts played by others. Along with 

 Mr. Canning, the engineer of the expedition, and Captain 

 Anderson, who commanded the Great Eastern, he received 

 the honour of knighthood. 



For a time his mirror galvanometer remained the only 

 instrument by which conversation could be carried on. 

 He now proceeded to design a substitute for it which 

 should give a record of the successive electric impulses 

 instead of merely exhibiting them to the watchful eye of a 

 skilled clerk. To secure greater power in the movement 

 of the indicator he inverted the function of magnet and 

 coil, making the coil the movable piece and the magnet 

 the fixed piece. The coil was, therefore, made very light ; 

 the magnet, which being stationarj' might now be very 

 heavy, was made exceedingly strong, and was arranged so 

 that the coil lay in an intense field between its poles. 

 The movement of the coil actuated a very light pointer, or 

 rather pen, in the form of a siphon-shaped tube of fine 

 drawn glass, from which ink was deposited on a running 

 paper band. Here we find the earliest example of the 

 moving coil type of galvanometer, often called the 

 D'Arsonval t>pe by those who do not recognise its real 

 origin. It is a type now familiar in many practical instru- 

 ments for the measurement of direct-current amperes and 

 volts ; but an important element in the invention is still 

 to be named. It was essential that the glass pen should 

 write without friction, and Thomson efifected this by the 

 happy device of electrifying the ink so that the ink and 

 the paper attracted one another, with the result that the 

 siphon was maintained in a constant state of rapid vibra- 

 tion, alternately advancing to the paper to deposit a minute 

 drop of ink and then springing back, but all the time free 

 to follow, without friction, the movements of the coil in 

 obedience to the electric impulses arriving through the 

 cable. Dynamically the siphon recorder has to satisfy the 

 same conditions as those that determined the design of the 

 mirror galvanometer. It draws on the moving strip of 

 paper a curve of arrival for every one of the successive 

 currents of which the signals are composed. 



To this day the recorder remains in universal use as 

 the standard instrument in submarine telegraphy. It has 

 been simplified by the substitution of permanent field 

 magnets for electromagnets, and by the use of an electro- 

 magnetic vibrator for the siphon instead of electrification 

 — changes which were made in later years by Thomson 

 himself. 



It is time now to turn to Lord Kelvin's work in naviga- 

 tion. Taking the two oldest aids to navigation, the com- 

 pass and the sounding-line, he revolutionised them both. 

 \\ here most men would have thought there was nothing 

 left for invention to do he found much. He has earned 

 profound gratitude for appliances which add immeasurably 

 to the security of all who go to sea. He has been called 

 the best friend the sailor ever had ; and it is said that a 

 bluejacket was once overheard to remark. " I don't know 

 who this Thomson may be, but every sailor ought to pray 

 for him every night." 



It was about 1873 that he began to study the compass 

 Seriously, partly because he had undertaken to write an 

 article on it for Good Words, and partly because he had 

 occasion to prepare, for the Royal Society, a biographical 

 sketch of his friend .\rchibald Smith, containing an account 

 of Smith's work on the theory of the perturbation of the 

 compass caused by the magnetism of iron ships. Kelvin's 

 first patent for an improved compass was taken out in 

 1876. 



He found the compass full of serious defects. For one 

 thing it was very unsteadv— that is to say, it was liable 

 to be set swinging through a large angle when the ship 

 NO. 2105, VOL. 83] 



rolled. Sometimes an attempt was made to reduce this 

 unsteadiness by introducing friction at the pivot, which, 

 in a way, made matters worse by causing the compass 

 to stick, pointing in a wrong direction. Under a mistaken 

 idea of what would lead to steadiness, the card was made 

 heavy and the needles long, and the long needles made it 

 impossible to correct the compass properly for the 

 magnetism of the ship. This was the most serious defect 

 of all. In iron ships, and especially in ironclads, the 

 compass is at the mercy of disturbing influences, which do 

 much to mask the true directive force of the earth's mag- 

 netic field. To neutralise these is indispensable ; the way 

 to do it, as a matter of theory', had been pointed out, but 

 i". was only through the radical change in construction 

 which we owe to Kelvin that it became possible to carry 

 the process into efifect. 



He recognised that for this purpose the needles mtist be 

 short. Further, that for steadiness what was wanted was 

 a long period of horizontal oscillation — in other words, 

 small magnetic moment relatively to the moment of inertia 

 of the card ; but, to keep the frictional error down, the 

 weight of the card, including the needles, should be small. 

 So he made the card as light as he could get it — a mere 

 aluminium rim tied by silk threads to a small central 

 boss, just as the rim of a bicycle wheel is tied to the nave 

 by wire spokes, and from the silk-thread spokes he hung 

 short pieces of magnetised knitting-needle to serve as the 

 magnets. The result was that not only was the total 

 weight very small, but it was nearly all in the rim, where 

 it is most useful for giving moment of inertia and con- 

 sequent slowness of period. Magnets and all, the card 

 only weighs 180 grains for a lo-inch size, and yet its 

 period of oscillation is much longer than that of the old 

 standard compass, while its friction error is less. 



Another admirable feature of Kelvin's invention was his 

 method of keeping the compass always level and free from 

 pendulum-like oscillation. He hung the bowl, as usual, 

 from gimballs, but with knife-edges instead of the usual 

 round spindles at the trunnions, and under the card he 

 provided a chamber at the bottom of the bowl partly filled 

 with castor-oil. You see this in the glass bowl now on the 

 table. There is a glass partition to separate the place 

 where the compass card stands from the lower part of the 

 bowl, and in the lower part is the castor-oil. Its function 

 is to damp out any oscillation of the bowl that may tend 

 to be set up by the rolling or pitching of the ship, and it 

 does so by dissipating the energy of such swings. At the 

 same time the knife-edge gimballs leave the compass per- 

 fectly free to take up a true level. 



Another feature is that the bowl and gimballs as a 

 whole is hung from springs to withstand vibration caused 

 by the action of the screw, or in warships by gun-fire. 



Now as to the correction for the magnetism of the ship. 

 Let me indicate very briefly the nature of that problem, and 

 how it is solved. 



An iron ship is a great magnet, or rather a great aggre- 

 gate of many magnets. Her magnetism at any instant 

 springs from two causes. First, there is the more or less 

 permanent part, which she takes up first when she is built ; 

 it depends to a great extent on how her head lay while 

 she was on the stocks. Then there is the induced part, 

 which changes with every change of course — a transient 

 effect due to the induction of the earth's magnetic field. 

 Strictly speaking, the induced magnetism is not entirely 

 transient, nor is the other by any means entirely per- 

 manent : but the ideal division into transient and permanent 

 is a highly useful one provided we understand the limita- 

 tion within which it is to be accepted. Now think of 

 what happens when the ship is " swung," that is, turned 

 so that she heads successively on all points. The per- 

 manent magnetism will cause an error of the compass 

 which will be of the same nature as you would find if you 

 placed a compass needle on a fixed pivot and disturbed it 

 bv turning a bar magnet slowly round a vertical axis. 

 This error will reach a maximum twice in the revolution, 

 once to one side and once to the other side — in other words, 

 once in each semicircle. Hence it is called the semi- 

 circular error. The permanent magnetism of the ship has 

 a vertical component, and this causes not only semicircular 

 error, but- also a heeling error, namely, a deflection of the 

 compass when the ship inclines to either side. 



