496 Mr. B. V. Hill on Accidental 



attribute the phenomena of diffusion directly to a fall of 

 pressure, scarcely any doubt can remain that there exists no 

 essential difference between colloid and crystalloid substances. 

 For the extreme slowness of diffusion indicates, upon the 

 one hand, a minute driving-force, i. e. a small osmotic pres- 

 sure, and, upon the other hand, a large viscosity which the 

 molecules possess in their translation through the water. 

 Both the phenomena can be explained upon the assumption 

 that the colloids possess an extremely large molecular weight." 

 As a corroboration of this theory, the molecular weight of 

 egg-albumen in water, determined by the freezing-point 

 method, is 14,000. This is a value nine times as large as the 

 simplest formula? would indicate. Such a condition would 

 occasion no surprise. In the case of silicic acid the case is a 

 little different. Here the osmotic methods give it a molecular 

 weight of " not less than " 49000. If silicic acid were repre- 

 sented by H 4 Si0 4 , it would have a molecular weight of 96 - 4. 

 That is, in this case we must assume a molecule 500 times as 

 large as had been supposed. 



According to Lord Kelvin*, the probable average dis- 

 tance between the centres of contiguous molecules is about 

 •00000005 cm If we imagine 500 of these strung together, 

 we shall have a line '000025 cm. long, which is less than 

 half the wave-length of sodium light. While this arrange- 

 ment may be imagined, it is not likely that such combinations 

 would be frequent. 



There are several reasons to doubt that we are dealing with 

 true solutions. With crystalloid solutions we may mix equal 

 parts of water and the solution, and obtain one which is one- 

 half the original concentration. With gelatine this is not 

 the case. We find that if 100 grams of gelatine be dissolved 

 in a given amount of water, boiled and cooled, it gives, when 

 examined under a strain, a certain rotation of the plane of 

 polarization of light. If now this fluid be immediately taken 

 out of the apparatus, diluted to one half its former concen- 

 tration, boiled, cooled, and examined as before, it gives one 

 half the effect. If, however, it were replaced in the chest 

 and examined without boiling, the result would be not only 

 different, but to some extent uncertain, as is shown by the 

 results given in Table V. This cannot be due to the 

 slowness of the diffusion, as the gelatine must be thoroughly 

 mixed through the water by the rapid motion of the 

 cylinders. 



If the solution be boiled and cooled a number of times, the 

 amount of double refraction is increased. This is not due 

 * ' Popular Lectures and Addresses,' vol. i. p. 224. 



