APPLICATION OP THE ELECTKIC TELEGRAPH. 327 



accurately measured to tenths, and- may be estimated to 

 hundredths of a second. On some ac- 

 counts it is more convenient to employ 

 a scale consisting of diverging lines, as 

 represented in the annexed cut, so that 

 the breadth of the scale may always 

 exactly comprehend the interval be- 

 tween the second dots, which intervals 

 must necessarily vary somewhat in length. 



It is important that the paper upon which the observa- 

 tions are recorded should be reduced to a convenient 

 form for preservation. The long fillet of paper employed 

 in ordinary telegraphing, is very inconvenient for astro- 

 nomical purposes. . If the paper is allowed to run off at 

 the rate of one inch per second, the length of fillet re- 

 quired for one hour's observations would be 3600 inches, or 

 300 feet; and for a single night's work of an observatory, 

 a length of nearly half a mile would be required. The 

 inconvenience of managing such a strjp of paper detracts 

 materially from the value of the method. 



In the summer of 1849, Mr. Saxton completed a re- 

 gister which is well adapted to the regular work of an 

 observatory. It consists of a cylinder which may be 

 made of any convenient dimensions, say six inches in 

 diameter and two feet long, enveloped with paper which 

 may be removed at pleasure. This cylinder is made to 

 revolve with a uniform motion, while the registering pen 

 moves forward at the rate of one tenth of an inch to 

 every revolution of the cylinder. Thus the pen is made 



