pritchard's wedge photometer. 323 



constant of the wedge if used to measure a star emitting rays composed entirely 

 of these wave-lengths. As might be expected, when the invisible rays are included, 

 the effect is still more marked. The portion of the total energy transmitted by the 

 thick end of the wedge, as determined by the bolometer, is fifty times as great as 

 the corresponding fraction of the visible light. 



A series of photometric measurements were made of the wedge at the Harvard 

 College Observatory. A photometer was used in which two portions of the same 

 beam of light could be compared by a Nicol and double-image prism. The measures 

 were made after one of these portions had passed through the wedge at a given 

 distance from its end. If we employ the term ahsorjAion to denote the numerical 

 inci'ease of stellar magnitude effected by the action of the wedge, these observations 

 could be represented by the formula ?« — 0.6 + 1-3;/, in which vi denotes the 

 absorption, and n the readiug of the scale corresponding to the part of the wedge 

 employed in the observation. This gives, for the absorption of the jioint marked 

 zero, 0.6, or the quantity whose logarithm is 0.2-i. In like manner the absorption 

 for each inch in length is 1.3 raagn., corresponding to the logarithm 0.52. This 

 agrees closely with 0.51, the value found by Professor Langley for the yellow ray, 

 X = 0.6. The photometric measures described above failed to show the gradual 

 diminution in the coefficient of absorption as the thickness increased. In fact, the 

 deviation appears to be in the opposite direction. Perhaps this is due to light 

 reflected from the rear surface of the glass, or other sources of error. As the light 

 transmitted by the thick end of the wedge is only about one five-thousandth of 

 the total light, a slight error in its measurement would be sufficient to produce this 

 effect. 



Another determination of the absorption of the wedge was made by photograph- 

 ing the solar spectrum through it. An exposure of 61 minutes gave a good 

 photograph of the spectrum of the light of the sky when passing through the part 

 of the wedge 3.5 inches from its end. Several exposures of from 5 to 30 seconds 

 were made on the same plate after removing the wedge. From about X = 0.5 where 

 the photographic image began, to X = 0.43, the spectrum with the v/edge had nearly 

 the same intensity as the spectrum without the wedge, having an exposure of 

 10 seconds. Photographic action appears to be nearly independent of the time 

 during which a given amount of energy is expended upon a plate. In other words, 

 the light required to produce a given image is nearly inversely proportional to the 

 exposure. The logarithm of the proportion of the light transmitted by the wedge 

 would therefore be 7.44. Allowing 0.24 for the absorption of the wedge at the 



