207 



roller, with a narrow groove corresponding to the point of the reci- 

 procating stile or pen operating on the outside. This groove enables 

 the pen to indent and crease the paper, by pushing it downward with- 

 out resistance. These indentations, thus made, were to some extent 

 indefinite and rather illegible, unless when the light is received tan- 

 gentially; and were always liable to become obscure or obliterated 

 by pressure and long keeping. 



In registering upon a sheet of paper enveloping a revolving cylin- 

 der, this use of a groove is impracticable; and the paper lying on a 

 hard surface could with difficulty be marked by dot or indentation. 

 As a substitute for the groove, I have interposed a covering of velvet, 

 or the like substance, between the paper and the solid cylinder, and 

 I am thus enabled to mark by sharp needle punctures, in a manner, 

 as I think, scarcely needing further attempts at improvement. 



I enclose you a specimen cut from a sheet two feet square, and 

 containing the work of four hours. This specimen is a zone or belt, 

 including the circumference of the cylinder, which revolves once per 

 minute, and receives in that time sixty punctures, distant from each 

 other about ten millimetres. When the sheet is cut open and deve- 

 loped from the cylinder, it presents a table having readings by inter- 

 sections derived from a vertical column on the left, and a horizontal 

 column at the top. On the left the column of seconds reads from 

 to 60, and on the top the headings are hours and minutes. Thus 

 it appears that an observation, or a puncture intermediate between 

 any two seconds, can be read as easily in the developed sheet as on 

 the cylinder. The successive spirals or revolutions of minutes return 

 very close to each other (2|- millimetres or one-tenth of an inch), and 

 the uniformity of motion of the cylinder is most severely tested by 

 the lines which the returning punctures make more or less parallel to 

 the cylinder's axis. This sample was produced without the magnetic 

 regulator, the cylinder running with a free flying motion. It is re- 

 markable, that with this flowing motion we should be enabled to gene- 

 rate a helix 500 feet long, and arrive at a final predicted point within 

 five millimetres or two-tenths of an inch; or, in other words, after 

 four hours of work, in which 14,400 punctures had been made, the 

 lines of returning dots had at no point departed half of a second from 

 a line parallel to the axis of the cylinder. The several spaces repre- 

 senting the seconds, differ from the 60th part of the circumference of 

 the cylinder not more than 2s-k^^ P^^^' 



VOL. V. — 3 F 



