468 



SCIENCE. 



[N. S. Vol. XX. No. 510. 



variation of personal equation. The observ- 

 ers were so placed that personal equation is 

 eliminated in the final longitude of Guam and 

 enters only once in the final longitudes of 

 Midway and Manila. It does not seem that 

 these longitudes can be seriously in error on 

 account of personal eqtiation. 

 , A method of automatically recording time 

 signals over long cables had been successfully 

 used by the Canadian and English observers in 

 the determination of the difference of longi- 

 tude between Greenwich and Montreal in 1892, 

 but no description of this method was avail- 

 able. A visit was made to the office of the 

 Commercial Cable Co., at ISTew York, where 

 information was obtained which led to the 

 designing and construction at the Coast and 

 Geodetic Survey office, of an ajsparatus for 

 recording automatically time signals over the 

 cables. 



The instrumental outfits of the five stations 

 of the Commercial Pacific Cable Co. are 

 similar. The Muirhead Syphon Recorders are 

 exclusively used, and a description of them 

 can be found in ' Submarine Telegraphs ' by 

 Charles Bright. The record is made upon a 

 slip of paper by a syphon pen attached by two 

 fibers to the coil of the recorder. The slip of 

 paper is made to move by an electric motor 

 at the rate of about two centimeters per sec- 

 ond. In order that the paper and pen may be 

 adjusted to the proper relation the rollers over 

 which the paper passes are attached to a stand 

 capable of a vertical and two horizontal ad- 

 justments. It was necessary that all cable 

 signals should be recorded on this slip of paper 

 at the cable office. The problem was to refer 

 the record of a signal received over the cable 

 to the time as recorded by a break-circuit 

 chronometer. It was solved in the following 

 manner. To the armature of a twenty-ohm 

 Morse relay was attached an arm of aluminum, 

 which carried a syphon pen exactly like the one 

 on the cable receiver. This relay was mounted 

 on a stand capable of a vertical and two hori- 

 zontal adjustments. This was called the 

 chronometer cable recorder. It could be placed 

 on the table in front of the cable receiver and 

 so adjusted that its syphon pen would record 

 ,on the slip of paper parallel to the record 



made by the pen of the cable recorder. The 

 chronometer and battery being put in circuit 

 with the chronometer cable recorder, a record 

 of the chronometer seconds was made on the 

 slip and any signal coming over the cable was 

 recorded by the pen of the cable recorder also 

 on the slip. This cable signal could be brought 

 vertically down to the record of the chronom- 

 eter and its time read off by the usual scale. 

 On account of the impracticability of always 

 adjusting the pens exactly opposite each other, 

 the cable signal has a correction, the deter- 

 mination of which will presently be explained. 

 Signals were sent by a key such as is used in 

 correspondence over the cable. It is a double 

 key by which a positive or negative current 

 can be sent to the cable. An attachment was 

 made to these keys at the Coast and Geodetic 

 Survey office by which the circuit through the 

 chronometer recorder would be broken the 

 instant the current was put on the cable and 

 thus record the chronometer time of the signal 

 sent. When sending signals the local cable 

 receiver is generally disconnected, but by so 

 arranging a shunt that a small portion of the 

 sending current would pass through the coil 

 of the local receiver a sharp record was also 

 made by the pen of the cable receiver and thus 

 the relation of the two pens was obtained. 



At the stations San Francisco, Honolulu 

 and Manila the observatories were so far from 

 the cable office that the chronometer cable 

 recorder at the cable office could not be placed 

 in circuit with the chronometer at the ob- 

 servatory. Another chronometer was, there- 

 fore, placed at the cable office to be used in 

 the exchange of the cable signals. At these 

 stations the cable offices and observatories 

 were connected by land lines by which the two 

 chronometers were compared before and after 

 the exchange of cable signals. At Midway 

 and Guam the cable offices and observatories 

 were only a few hundred feet apart, so that 

 the chronometer cable recorder at the cable 

 office could be placed directly in circuit with 

 the chronometer at the observatory and no 

 comparisons of local chronometers was neces- 

 sary. The results indicate that time signals 

 over long cables can now be exchanged with 

 as great accuracy as over the best land lines. 



