SEA-FISHERIES LABORATORY. 163 



permanently in the form of a jelly, or a viscous liquid, according 

 to the concentration. The interstices of the mesh work are, 

 nevertheless, so fine as to prevent the transport of even such 

 small particles as the bacilli. How these spread through 

 the jelly is a subject for microscopical investigation by sections 

 of the hardened substance — -I do not know if this has been 

 attempted. The agar jelly we may imagine to be like a sponge 

 containing liquid, and the liquid consists of the dilute agar 

 solution, and of the solutions of the nutritive and other 

 substances entering into the composition of the medium. 

 In a very concentrated agar jelly, the phase containing much 

 water would consist of droplets enclosed by the phase containing 

 little water, so that diffusion of electrolytes through the 

 jelly would be very slow. But in the dilute agar jelly there 

 would be little more resistance to diffusion than if the whole 

 were a true molecular solution. 



The surface of the jelly is always moist, since it is contracting 

 and liquid is being forced out. A colony on the surface will, 

 therefore, grow more rapidly than one in the depth of the 

 plate, since it will take up the dissolved nutritive substance 

 round itself, and the concentration of the latter will be renewed 

 by diffusion from surrounding areas of the surface as rapidly 

 as it is lowered. Therefore, these superficial colonies will be 

 larger, and slightly different in appearance. 



Diffusion will be rather slower in the depth of the plate, 

 and adjacent colonies will act as centres of absorption for the 

 dissolved nutritive substances. If the colonies are numerous 

 they ought to be smaller, since they are growing at the expense 

 of the liquid food substances added to the jelly rather than 

 the latter itself. And this is what we actually find. The 

 colonies in the crowded plates are always smaller than those 

 in plates which have been inoculated by only a few bacteria. 



Further, and this is a matter of considerable practical 

 importance in methods, a crowded plate ought to indicate 



