ASTRONOMY i8i 



of this effect depends upon the developer, being greatest with 

 pyro-metol. Examination of sections of images developed 

 with different developers shows that these effects are produced 

 by strains in the gelatine resulting from unequal drying. The 

 developed image has a much smaller water content than that 

 of the surrounding gelatine, and therefore dries more rapidly. 

 This gives rise to stresses, parallel to the supporting plate, 

 which are greatest near the edge of the image. The stresses 

 cause the edge to move inwards, dragging along the surrounding 

 gelatine by an amount which decreases approximately ex- 

 ponentially with the distance from the edge. This accounts 

 for the translatory motion of the star images. The amount 

 of the contraction depends upon the size of the image, its 

 density, and upon the developer. 



This phenomenon is of importance in photographic measures 

 of close double stars. The contraction will cause the measured 

 distance to be too small. The same is true of measurements 

 of close spectral lines, and of the relative positions of stars in 

 dense clusters. For such purposes, therefore, developers such 

 as caustic hydroquinone and pyro should not be used, but 

 preferably metol- or chlor-hydroquinone, with which very little 

 contraction results. 



A third paper deals with " Photographic Sharpness and 

 Resolving Power" {ibid., 52, 201, 1920), and investigates the 

 conditions required to obtain sharp images and high resolving 

 power. The two important quantities for determining the 

 sharpness are 7 and F, defined above. P is the rate of increase 

 of diameter of image with logarithmic exposure, and may be 

 used as a measure of the " turbidity " of the emulsion. If the 

 turbidity is high, the photographic image will rapidly extend 

 beyond the limits of the geometrical image ; 7 provides a 

 measure of contrast, defining the rate of increase of density. 

 It is found that the decrease of intensity at the edge of an 

 image follows an exponential law, with a coefficient which is 

 inversely proportional to the turbidity, whilst the " sharpness," 

 which may be defined as the rate of change of density with 

 distance from the geometrical edge of the image, is equal to the 

 ratio of the contrast to the turbidity. Practical investigation, 

 based upon the study of enlargements of knife-edge contact 

 prints exposed to light of varying wave-length, confirms this 

 conclusion in a general way, although actually the phenomena 

 are complicated somewhat by secondary actions. The result 

 indicates, however, that in order to secure sharp images, 

 emulsions should be chosen which have a high contrast and a 

 low turbidity. It is also important that the characteristic 

 curve of the emulsion should have a minimum amount of 

 " toe/' so that the slope (which defines the contrast 7) should 



