456 Messrs. S. E. Sheppard and C. E. K. Mees. [Dec. 20, 



ratios demonstrated later. A conclusive experimental disproof of the 

 hypothesis is given by the fact that a plate exposed from the glass- 

 side develops normally, although in this case the layer containing 

 fewest reduced particles is reached by the developer first. 



Table III. 



Time. Density. K. 



11 -Grains. 0'477 0*0210 



15-0 0-612 0-0217 



20-0 0-681 0-0193 



33-0 0-897 0-0197 



75-0 M56 



120-0 1-156 For meaning of K, see later. 



The mass-time curve is quite normal. 



Another theory of development-velocity was proposed by Dr. Abegg,* 

 based on his " silver-germ " theory of the latent image. According to 

 this, the velocity of development, i.e., the number of silver germs 

 deposited in unit time is proportional to the number present at any 



time. This may be expressed by V = JT *=& (a + b + x), where a and & 



(M> 



are constant numbers of germs due to exposure and ripening respec- 

 tively, and x is the number due to development. It is evident that 

 as x increases the velocity should increase, and a steady acceleration 

 should be noticed till the whole film was developed through ; in this 

 unmodified form the expression is in obvious disagreement with the 

 facts of development. 



Theory of Development. The investigations of Boguski,t Noyes and 

 Whitney,| and especially of M. Wilderman, E. Bruner, and Nernst,|| 

 have led to a better understanding of the reaction-velocity in hetero- 

 geneous systems, among which the development of silver-bromide 

 naturally stands. When a solid dissolves, the rate is proportional to 

 its surface and to the difference between the saturation-concentration 

 and that at the given moment. At the boundary between the phases 

 saturation exists, so that the rate of solution depends on the diffusion 

 velocity. When a chemical action is superposed, Nernst assumes that 

 in most cases the equilibrium in the reaction-layer is adjusted with 

 practically infinite velocity compared with the diffusion process. If the 

 length of the diffusion path and the diffusion coefficients are known, 

 in many cases the velocity may be calculated in absolute measure. 



* R. Abegg, ' Archiv f. Wiss. Phot.,' 1899, vol. 1. 



t v. Ostwald, ' Lehrbuch,' 2te Auflage, 2te Bd., 2te Fl. ' Chem. Kinetik.' 

 J ' Zeit. f. Phys. Chem.,' 23. 



M. Wilder-man, ' Phil. Mag.,' October, 1902 ; < Zeits. Physik. Chem.,' vol. 30, 

 1899, p. 341. 



|| ' Zeit. f . Phys. Chem.,' 47, 1904, p. 56. 



