REPORT OF THE SECRETARY. 71 



As completed, the constants of this galvanometer are as follows : 



Resistance of coils, each ohms . . 30 



Diameter of coils : 



Exterior millimeters.. 34 



Interior do 2 



Distance apart of coils do 1.5 



Weight of needle: 



First milligrams.. 2.5 



Second do 6. 5 



Length of quartz fiber centimeters.. 30 



Diameter of quartz fiber millimeter.. . 0015 



Current in amperes giving l mn » deflection on scale at l ni at time of 

 single swing of needle 10 seconds when coils are connected in 

 series parallel (total resistance hence 30 ohms) : 



With first needle 000000000005 



With second needle 000000000020 



Constant under similar conditions for galvanometer used last year. .000000000100 



Notwithstanding the greater sensitiveness of the first needle, the second has been 

 employed in taking holographs for two reasons: First, because its mirror (of 2 mil- 

 ligrams weight, by Brashear) gave better definition ; and second, because the needle 

 was steadier on account of its greater weight. It is probable that means could be 

 devised, though not without considerable time and expense, to use the first needle 

 with as much satisfaction as the second, and thus to gain four times in sensitiveness. 



The achievement of dhis very considerable advance in sensitiveness made the use 

 of both a narrower bolometer and a narrower slit to the spectrobolometer possible. 

 The former was already at hand. To reduce tlie linear width of the slit would have 

 resulted in a waste of radiations because of diffraction, a danger to which some 

 reference was made in last year's report. To reduce the angular width of the slit 

 by a collimatiug system of spherical mirrors of longer equivalent focus would have 

 resulted in a waste of radiations, because of the vertical spreading out of the beam. 

 In these circumstances an arrangement of cylindrical collimatiug mirrors was 

 designed, with the aim to avoid both horns of this dilemma. These mirrors, one 

 convex of 57 centimeters focus and one concave of 544 centimeters focus, were exe- 

 cuted by Brashear and give equally as good definition as the spherical concave mir- 

 rors before employed, while reducing the angular width corresponding to a given 

 linear aperture of the slit to about one-seventh of its former magnitude. The 

 angular widths of slit and bolometer strip have now each been reduced to about 1.3 

 seconds of arc. 



In the use of the new arrangements much trouble was at first experienced from 

 "drift" and accidental disturbances of the galvanometer. The "drift" wasreduced 

 to nearly its former harmless magnitude by added precautions to avoid temperature 

 changes. But the accidental disturbances, especially with the very narrow bolom- 

 eter (0.03 millimeter), were very serious, and on days when there was the slightest 

 breeze absolutely prohibitive to observation. It was found at length that by mak- 

 ing the chamber occupied by the bolometer air-tight to a difference of pressure of 

 one-third of an atmosphere all prejudicial effects of the wind, except such slight 

 ones as were due to mechanical jarring, were avoided. Bolographs may now be 

 taken with good results on the very windiest days. 



Mechanical jarring of the galvanometer has been reduced by floating it in a pan 

 of mercury, which is itself supported upon the table of the Julius suspension intro- 

 duced here in 1895. 



Bearing in mind the advantage derived from inclosing the bolometer in an air- 

 tight compartment, an air-tight galvanometer case was designed, and having been 

 constructed was made use of for some of the latest bolographs. The advantage 

 derived from its use, however, proved slight. 



