50 



NA TURE 



[May 21, 1874 



But in the case of photography the low altitude of the 

 sun introduces a much more serious dii^culty. The light 

 has in this case to pass through a great length of the 

 earth's atmosphere, in its lowest and densest regions. 

 Much of the light is absorbed by the atmosphere, as is 

 shown by the fact that the rising or setting sun may be 

 gazed at with impunity. But further, it is found that of 

 all the colours composing the sun's light, those which 

 affect most powerfully a photographic plate arc the most 

 greedily absorbed. Hence it has been found at St. 

 Petersburg that at m.id-winter a photographic plate must 

 be exposed to the sun 360 times as long as at the equi- 

 noxes, when the altitude of the sun is about 6° or 7°. 

 This is a difficulty which cannot be surmounted except 

 by exposing the plate a longer time than is desirable. 



It has been already stated that considerable dis- 

 crepancies in determining the times of contact might 

 arise from observers noting different phenomena. The 

 employment of the Model Transit of Venus ensures con- 

 cordance among the observers of each nation ; but all 

 European observers will be much indebted to M. Struve, 

 who has actually compared his own observations with 

 those of the Russian, German, English, and French ob- 

 servers, so that comparisons will be possible between the 

 observations of these different nations. 



Everything being now prepared for observing as suc- 

 cessfully as possible the actual phenomenon of contact, it 

 remains to describe the means by which the time can be 

 determined accurately. All clocks and watches are set 

 and regulated by observations of the stars, or by com- 

 parison with other clocks so regulated. An astronomical 

 clock counts the hours up to 24h. The clock is set to oh. 

 at the instant when a particular star passes the meridian. 

 If then we have a means of determining the time when 

 this happens, we can set our clock accurately to local 

 time. But a star does not pass the meridian of Greenwich 

 at the same time as it passes the meridian of a place 

 having any other longitude. By the aid of a transit in- 

 strument the local time can be determined ; but to 

 determine actual Greenwich time at another place we 

 must, as before stated, know accurately the longitude of 

 that station. T/wse iwi} things, the absolute time ami the 

 longitude, are so connected, that if we Icnorj the one, the 

 otlicr can be immediately deduced. 



The longitude may be determined in a variety of 

 different ways. If the two places whose difference of 

 longitude is to be determined be not very distant, a 

 simple method may be employed. A rocket is sent up 

 from some point between the two stations. An observer 

 at each station notes the local time at which the rocket is 

 seen to burst. The difference between these times gives 

 the difference of longitude. A flash from a lamp or re- 

 flected sunlight may be similarly employed. 



The absolute time (and consequently the longitude) 

 can also be found by transporting chronometers from one 

 station where it is known to another where it is not 

 known. First-rate chronometers must be used, and a 

 large number to check one another's errors. The main 

 error of a chronometer is due to the influence of tempera- 

 ture on the momentum of the balance vi'heel and the 

 strength of its spring. The Russians have of late years 

 introduced with great success a method of secondary 

 correction for this error. Along with the compensated 

 chronometer at least one is sent without any compen- 

 sation. The difference between this chronometer and 

 others is a measure of the sum total of the tempera- 

 tures to which they have been exposed ; and by the aid of 

 a table carefully drawn up from a number of observa- 

 tions, the amount of secondary correction necessary can 

 be fairly estimated. It is said that the employment of 

 this device is of the very greatest service. Ten well-tried 

 chronometer?, accompanied by a single uncompensated 

 one, if carried between stations ten days apart (V.^-. St. 

 Petersburg and Cazan) will, in one journey, give the longi- 



tude of an intermediate station (such as Moscow) cor- 

 rectly within j^,7 of a second of time. By the aid of this 

 contrivance chronometers may be employed, even for 

 very long journeys, to determine the longitude. This 

 method is quite new, and has not been tested by any 

 nations except the Russians. The results obtained by 

 them are, however, perfectly satisfactory. Theoretically 

 the idea is almost perfect ; the outstanding temperature 

 error being the main fault of chronometers, and the em- 

 ployment of an additional chronometer uncompensated 

 giving us a means of determining the amount of this 

 error, the time deduced by this means ought to give very 

 satisfactory results. There is but one objection to the 

 method, which is only a partial one. After a series of 

 alternately very hot days and very cold nights, the differ- 

 ence between the compensated and uncompensated chro- 

 nometers might be the same as after the same period, 

 with a tolerably uniform temperature ; but the correction 

 necessary in these two cases might be very different in- 

 deed. It is easy, however, to keep chronometers at a 

 temperature which does not vary rapidly, and the experi- 

 ments made by the Russians warrant us in saying that by 

 the aid of this method longitudes may be determined, 

 with very great accuracy indeed, in voyages of such length 

 that the ordinary chronometric method would be unavail- 

 ing, and that in every case where longitudes are required 

 by the use of chronometers this method should be 

 employed. 



A third way of determining the absolute time is by the 

 use of telegraphic signals. An operator at Greenwich 

 may arrange to telegraph a signal to another at 

 Alexandria at a certain definite time of day. If the trans- 

 mission ot the current from Greenwich to Alexandria 

 were instantaneous the person at Alexandria would at 

 that instant receive the exact time. But a current through 

 a submarine cable is retarded. Suppose it to be retarded 

 two seconds ; the time received at Alexandria will be too 

 late by two seconds. If now an operator at Alexandria 

 telegraphs to Greenwich he will dispatch the signal two 

 seconds before it reaches Greenwich. The longitudes de- 

 termined by the two currents in opposite directions will 

 therefore differ by four seconds. The mean of these 

 values gives the true longitude, and half the difference 

 between the two determinations is the time of transit of 

 the currents. It is found, however, both from theory and 

 experiment, that if there be a leak in the cable nearer to 

 Greenwich than to Alexandria the current will pass more 

 slowly in going to Alexandria than in the reverse direction. 

 This difference, however, can never be very great. 



Considerable differences have been found by the 

 Americans to exist between comparative observations of 

 longitude by the telegraphic method and by the lunar 

 method, which will presently be described. The Americans 

 rushed to the conclusion that the error existed in the lunar 

 method. This is not necessarily so. The American 

 system of telegraphing over long distances consists in 

 using a relay. A relay is an arrangement to overcome 

 the difficulty of sending a current through a long 

 line. It is placed at an intermediate station. It 

 consists essentially of an electro-magnet which attracts 

 a piece of iron when a current which has originally 

 been sent through the primary station passes through its 

 coils. This attraction of a piece of iron makes contact 

 with a new electric circuit with a separate battery, and so 

 the current is passed on to the final station, or to a second 

 relay. The piece of iron must move through a sensible 

 distance before the second circuit is completed. Ithas 

 hitherto been supposed that the time lost in employing a 

 number of relays could be eliminated by sending the 

 current in alternate directions as above described. This 

 is certainly not the case. The time elapsing before con- 

 tact is made by a relay depends upon tlie strength of the 

 current. The strength of the current depends upon the 

 length of the wire through which it is passing, and also 



