9 8 



NATURE 



[March 27, 1913 



case an interesting exception was observed. A piece 

 of iron which had not rusted on long exposure to the 

 action of air and water was placed in a strong- solution 

 of copper nitrate; after some time beautiful crystals 

 of copper were deposited on parts of the iron, whilst 

 other parts remained quite unaffected. 



T ] 



SOUTHERN HEMISPHERE SEASONAL 



CORRELATIONS. 

 'HE first of a proposed series of articles on this 

 important subject by Mr. R. C. Mossman, of 

 the Argentine Meteorological Office, appears in 

 Symons's Meteorological Magazine for February. Not- 

 withstanding the great labour involved in this kind 

 of research, it has received increasing attention from 

 leading meteorologists during recent years. Mr. 

 Mossman has collected a large mass of material re- 

 lating to the climate of South America, which is now 

 available for testing whether the sequences of weather 

 in that continent " show as pronounced resemblances 

 or contrasts, when compared with data from other 

 regions, as do those in the northern hemisphere." 



The inquiry now in question refers to the relation 

 between the Nile flood and the winter rainfall of 

 Santiago (Chile). The data used for the Nile floods 

 are the percentage values for the years 1869-1906, 

 published by Captain Lyons in " Rains of the Nile," 

 1906, and, for rainfall, the percentage values at San- 

 tiago for May-August of the same years. When 

 plotted^ on a diagram, it is seen that, on the whole, 

 there is a strongly pronounced opposition between 

 the two sets of values. The author points out that 

 the winter rainfall of Santiago, in common with 

 other stations between 32 and 39° S., varies with 

 the position of the South Pacific high-pressure area. 



The Chilean Meteorological Office has recently sup- 

 plied a complete set of instruments to Juan Fernandez, 

 and the^ island is in radiographic communication with 

 the mainland. This, with observations from a new 

 station on Easter Island (27 S., 109 W.), should, 

 Mr. Mossman thinks, afford useful information re- 

 garding the seasonal relations of the South Atlantic 

 and South Pacific anticyclonic belts, and later 

 on, when these data are compared with those at St. 

 Helena, there is little doubt that the chain linking 

 up the rainfall of Abyssinia with the Antarctic circula- 

 tion will be complete. Captain Lyons has shown 

 that the height of the Nile flood is dependent on the 

 June to September rainfall in Abyssinia. 



SOME METHODS OF MAGNIFYING FEEBLE 

 SIGNALLING CURRENTS. 1 



'"TELEGRAPHY over long submarine cables is con- 

 -*■ tinually on the increase, and I think it may be 

 brought forward as a fairly accurate statement that 

 the number of messages sent doubles itself every ten 

 years. It is therefore important that, besides the 

 increase in the number of the cables laid down each 

 year, means should be devised to increase the carrying 

 power. 



The instruments which I have invented and am 

 about to describe were designed primarily for cable 

 work, but they are equally applicable to recording 

 many other kinds of signalling impulses. 



For good reasons, recording by photographic means 

 is objected to by nearly every telegraphist. If the 

 photographic method were permissible, great advances 

 in speed would be available, but it is important that 



1 Discourse delivered at the Eighth E>hihition of Apparatus, held hv the 

 Physical Society on December 17, 1912, by : Mr. S. G. Brown. 



NO. ^265, VOL. 91] 



the record should be of a simple, cheap, and imme- 

 diate nature. 



Lord Kelvin invented the siphon recorder in 1867 

 — that is, about forty-five years ago ; he designed it 

 so carefully that no improvement in its sensitiveness 

 has been brought about until now. 



Short Siphon Recorder. 

 In siphon recorders of the moving-coil type what 

 has to be done consists of — 



(1) Overcoming the inertia of the coil and siphon. 



(2) Overcoming the back E.M.F. of the coil. 



(3) Overcoming the control of the suspensions. 

 ty Overcoming the friction of air, suspensions, and 



inking. 



As the siphon has to return to zero in a certain 

 time after the current in the coil ceases, it is neces- 

 sary for the coil and siphon to have a definite fre- 

 quency of oscillation depending on the speed of the 

 signals. For submarine telegraphy this frequency 

 lies between about 3 and 10 per second, and is ad- 

 justed by varying the control on the coil. As the 

 control necessary to give a certain natural period to 

 the moving system is proportional to its moment of 

 inertia, it follows that by reducing this inertia we 

 reduce the forces required both to accelerate the coil 

 and to overcome the control. 



The ordinary siphon recorder employed is a siphon 

 tube about 2!- in. long and from S to 12 mils in 



diameter. The moving coil consists of 500 turns of 

 2-mil wire at a mean radius of § in. The coil and 

 siphon are mounted on separate axes and are con- 

 nected by silk fibres so that the angular movement of 

 the siphon is about two to three times that of the coil. 

 By reducing the length of the siphon to * in. and 

 substituting a narrower coil it is possible greatly to 

 increase the sensitiveness of the recorder. 



In order to make the inertia effects of the moving 

 system a minimum, it is advisable to make them 

 equal for the coil and the siphon. Even a narrow 

 coil of 300 turns has about 100 times more inertia 

 than the siphon, so that it is necessary to move the 

 siphon through V 100 times the angle moved by the 

 coil. 



By reducing the number of turns on the coil and 

 increasing the field it is possible to reduce the natural 

 period for a given sensitiveness and back E.M.F., but 

 as the mass of the mountings and insulation of the 

 coil only decrease slightly as the turns are 

 reduced the gain is not very marked. In prac- 

 tice it is inadvisable to reduce the turns on 

 the coil below 50 or 100 turns, as with lower values 

 the power required to overcome the friction of the 





