Mr. A. P. Trotter on a New Photometer. 85 



the front opening or diaphragm carries a sight or pointer, 

 and the back screen is provided also with a pointer, prefer- 

 ably at its upper edge. The photometer may be examined 

 from a distance, say six or eight feet, and moved by cords. 

 The two sights should be brought in a line with each other, 

 and the carriage is to be moved until the band of uniform 

 tone is bisected by the pointers. If the carriage be moved 

 slowly it will be seen that the band of uniform tone remains 

 fixed relatively to the scale on the photometer-bar. Readings 

 might be taken by noting the position of the band over the 

 scale, but I prefer to follow the usual custom in photometry, 

 and to keep the scale hidden until an adjustment has been 

 made, to avoid any bias. The operation of estimating the 

 bisection of the band by the pointers is, I think, easier and 

 therefore more conducive to precision than the estimation of 

 the similarity of two images as in the Bunsen, or of two 

 tones as in most other photometers. 



I will now give details of the development of this photo- 

 meter for the guidance of those who may wish to construct 

 them. 



Perforated zinc of the ordinary pattern having holes about 

 0*08 inch diameter and ^ inch pitch, although countersunk 

 on the back and painted dead white on the front, was found 

 to be of no use, probably because the area of the holes was 

 considerably less than that of the remaining metal. Half-inch 

 holes punched in thin card at 0*6 inch pitch gave on the whole 

 a better result than ^-ineh or ^-inch wooden slips, bevelled, 

 glued to a frame at a distance apart equal to their width, 

 and painted dead white. I then tried a number of different 

 patterns of perforated zinc and other screens, 30 centim. long 

 by 6 centim. high. Circular perforations of 4 millim. and 

 6 millim. diameter, about half the metal being removed, were 

 less satisfactory than zinc with holes 13 millim. diameter, 

 rather more than half the metal being removed. A fancy 

 pattern was tried, but without good results. The suitability 

 of the screens was compared in some cases by comparison of 

 the " mean error of a single observation " calculated in the 

 ordinary way from ten measurements, two similar glow-lamps 

 being used, sometimes 2 metres apart, and sometimes 3 metres. 



Finding that thinness of edge and perfect flatness were 

 very important, I made several screens of card and of paper 

 stretched while wet on wooden frames. The Coarsest of these 

 consisted of strips of two-sheet Bristol board 8 millim. wide 

 and 8 millim apart, 18 such strips, with an 8 millim. space 

 at each end, making 30 centim. The strips were cut with 

 sharp scissors to avoid any burr on the edge. Another 



