METHODS 39 



organoleptic qualities ; it is easily understood that such 

 qualities concern the relations of chemical bodies with man 

 and not their relations with this or that definite compound. 



Accordingly such properties are useful only in man's 

 direct knowledge of substances. We ought to remember 

 that all our means of knowledge are, to a certain degree, 

 organoleptic. Even when we measure rigorously by means 

 of the most exact apparatus, when in other words we use 

 science in its strictest methods, we are always decomposing 

 the descriptions of things into elements which strike our senses. 

 Length and thickness measured in centimetres, temperature 

 measured by the thermometer and the rest are so many 

 descriptive elements that concern our sight or sense of touch ; 

 and, possibly, they may not be directly concerned in the rela- 

 tions between the body which we study and some other body 

 than ourselves. 



On this account we must distrust what is called the 

 simplicity of things. A phenomenon which is simple to us 

 may, on the contrary, be complicated in the extreme in 

 relation to some other phenomenon with which it struggles 

 in nature. 



Reciprocally, and this especially is important to us here, 

 a phenomenon very complicated in human analysis may 

 have very simple relations with some other phenomenon of 

 nature. For example, we shall see that, in a struggle between 

 two different colloid states, phenomena appear which can 

 be summed up in a simple formula, whereas the colloidal 

 state itself appears to us men something so complicated 

 that, in the present state of science, we are not yet able to 

 give a direct definition of it. 



So, when we study a new phenomenon, it is always best 

 to abandon for the time being our human knowledge of the 

 phenomenon and to seek for the other phenomena of nature 



