284 Prof. W. H. Miller on a new Form of 



5° with the bright signal, and adopting the image of this in 

 the dark glass mirror W for the faint signal. The mirror, 

 about 40 millims. wide and 30 millims. high, is cemented to a 

 thin metal plate bent so as to make an angle of about 30°, 

 while by means of a screw tapped in the other part it can be 

 fastened to the upper end of a strip of brass 208 millims. long, 

 25 millims. wide and 3 millims. thick, having slits at each end 

 to receive screws, and bent at right angles at a point 40 millims. 

 from the lower end. 



By making the mirror revolve through a small angle round 

 the screw which fastens it to its support, and the support 

 round the screw by which it is fastened to the instrument, 

 the mirror is capable of suitable aclj ustment. A second mirror 

 and support will be found useful, the two mirrors being 

 attached on opposite sides of a vertical plane passing through 

 the centre of the instrument and the source of light. A mean 

 of the results will be free from the errors depending upon 

 the first powers of the eccentricity. For various reasons the 

 accuracy of the observation increases when the angle of inci- 

 dence of the light on the crystal is small. This angle can be 

 easily reduced to less than 15°. 



The accuracy of the observations is undoubtedly increased 

 by the use of collimators ; they also allow the observation to 

 be made when the space at the observer's disposal is very 

 limited. They may have object-glasses of from 20 millims. 

 to 30 millims. aperture and 500 millims. or more focal length. 

 The two collimators are mounted on supports having each 

 three nearly equidistant notches, so that one resting in the 

 middle notch of each support may have its axis passing 

 through the axis of the goniometer at a distance of 132 millims. 

 above the plane of the graduated circle : it has in its principal 

 focus a vertical adjustable slit, through which the light of the 

 sun is thrown from a heliostat-mirror in the direction of the 

 axis of the collimator. It is sometimes convenient to place a 

 lens in front of the slit in order to enlarge the emergent pencil 

 of light. The other collimator, resting in either pair of the cor- 

 responding notches, may have its axis passing nearly through 

 the middle of the dark glass mirror. The end opposite to the 

 object-glass is covered with a plate of glass ground rough on 

 the outer surface, and covered on the inner surface with indian 

 ink scratched away in a vertical diameter. 



When the crystal is large, or is implanted on a matrix of 

 some considerable weight, it is obviously impossible to secure 

 and adjust it on the ordinary Wollaston's branch. In such 

 cases it is fixed in a sort of vice having three parallel claws, 

 one of which is movable. By screwing up this claw the 



