104 DR L. BECKER ON THE SOLAR SPECTRUM. 



full revolutions of the recording wheel are registered in a simple manner. A strong nail 

 was driven into the rim of the wheel, and filed away to a sharp edge, which leaves a dis- 

 tinct mark in the paper every time it passes beneath the roller. These marks served as 

 zero points in reading off the observations. We may mention that in the 3500 feet of 

 paper that contain the observations, not one of these marks is wanting ; and judging 

 from the intervals between them, the fillet has never once slipped. Apart from this 

 safeguard, the observer, when turning the roller, could always see in the viewing tele- 

 scope that the grating had moved ; and this could not possibly happen unless the fillet 

 had correspondingly advanced. If the recording wheel was intentionally held fast, it 

 was impossible to draw the fillet over it by turning the roller. 



As to the linear distance between two lines on the paper, it may easily be computed 

 from the figures given, that the D lines for instance are 19f inches apart, whilst the 

 whole region from A = 6024 to 4861 would require a strip 314 feet long. 



The apparatus works in the following manner : — The observer with his right hand 

 turns a toothed wheel on the same axis as the roller; this drives the recording wheel and 

 moves the paper along by friction. The long iron rod transmits this motion to the disk 

 of the connecting joint, and then by means of the cross to the other disk which is fixed 

 to the second pinion. This second pinion, acting through the wheels and endless screw, 

 slowly rotates the grating, thus causing the lines of the spectrum to move across the field 

 of view. When the line under observation coincides with the intersection of the wires, 

 the fingers of the left hand depress one or more of the needles according to the degree 

 of blackness of the line. If the lines of the spectrum are near together, they can be 

 registered as quickly as the eye can appreciate their individual characteristics. 



In spite of all the connections and the smallness of the worm-wheel, the probable 

 error of one observation of the relative position of the grating is but +0 ff '77 of arc as 

 computed from lines half-way between starfdard lines. This corresponds to -gu.^ju inch 

 in the circumference of the worm-wheel. 



For effecting a quick motion of the grating, the bracket to which the wheelwork is 

 fastened turns round a pivot at the upper end, and can be raised out of position by a 

 string. By a long wooden handle the observer can then rotate the tangent-screw 

 directly, without quitting his seat at the eye-end of the viewing telescope. 



The instrument could be much simplified. A small table moving easily round a 

 vertical axis from which a rigid arm projects as far as its rigidity permits, and of course 

 balanced, and a screw of low pitch acting on the arm similarly to the slow motion of a 

 Transit- Circle in Declination, would be a simple substitute for all our multiplying gear. 



When a great number of lines have to be determined by eye-observations, such an 

 instrument will always give accurate results in a comparatively short time, provided it is 

 possible to introduce a sufficient number of standard lines. 



