574 Transactions of the Society. 



turbid or colloidal solutions, fixed or fluid, are similarly suitable, 

 provided that the average distance of the single particles is no 

 smaller than half a wave-length. 



Let us suppose that the dimension in every direction of these 

 small particles is less than half a wave-length. In that case 

 it is clear that their microscopical images will only be diffraction 

 discs. Now such, for simplicity's sake, will be called " ultra- 

 microscopic " particles ; for the expression will at the same time 

 indicate that the resolution of detail in the structure of these 

 particles lies beyond the resolving power of the Microscope. 



It might be suggested that mere evidence afforded by such 

 diffraction discs does not sufficiently differentiate the respective 

 particles, and therefore such investigations as the present ones 

 might be treated as superfluous. But I believe that the experi- 

 ments with gold ruby glasses which I was able to make, at the 

 instigation of, and together with Dr. Zsigmondy, have afforded an 

 optical proof that distribution of gold in these glasses is discon- 

 tinuous, and have also demonstrated that there are a number of 

 phenomena characteristic of diffraction discs such as colour, order 

 of position, condition of polarisation and brightness, and in fluids 

 also kinds of movement. So many properties seem to warrant a 

 careful diagnosis. 



Now, microscopic investigations relating to ultra-microscopic 

 particles cannot be effected by the usual methods. The coloured 

 ruby glasses, in which the distribution of the various particles was 

 demonstrated, showed no sign of their existence when examined in 

 the ordinary way, or even when examined by dark-ground illu- 

 mination of the usual kind. Under such conditions the glasses 

 appeared perfectly homogeneous. One might almost have expected 

 that these gold ruby glasses in thin slices would have given some 

 indication of heterogeneity, because they might be supposed to be 

 somewhat analogous to stained bacterial preparations. 



It was therefore necessary to devise a new method which would 

 permit these small particles to become visible by direct observa- 

 tion as far as possible. The main feature of this method depends 

 upon the regulation and arrangement of the illumination, which, 

 as will be observed, differs materially from that hitherto em- 

 ployed. As in general particles to be optically imaged are not self- 

 luminous in themselves (or where they might be slightly self- 

 luminous the light would be so weak as not to be of any service), 

 we are from the outset compelled to rely upon an artificial light- 

 source of great specific intensity, such as the electric arc or 

 brilliant sunlight. 



When this is made to impinge upon the particles they become 

 visible by the cone of rays which they diffract. But the intensity 

 of the illuminating rays is naturally very much higher than that 

 of the rays diffracted by the particles. In order to make smaller 



