EXPERIMENTS ON KAOLINIZATION. 291 



the unit mass of feldspathic rock in the unit of time. T. E. K. may, there- 

 fore, a priori, be either positive, zero, or negative. It must be regarded, 

 moreover, as a function of the time during which the action has been going 

 on, of the temperature and of the quantity of feldspar contained in a 

 given sample of rock. 



The problem pi-esented is none other than thie measurement of very 

 small increments of temperature with all the accuracy attainable. For such 

 a purpose either thermometric or electrical means are applicable. The for- 

 mer requiring specially constructed apparatus, had at once to be discarded. 

 It is a question, moreover, whether the thermometric method of research 

 will not, under all circumstances, offer obstacles of a very serious character. 

 In the measurement of small increments at the boiling point it becomes a 

 matter of great importance to keep the mercury column throughout at a 

 temperature as nearly as possible equal to that of the bulb — a condition 

 which can be realized only with great difficulty, when a division of the stem 

 into very small fractions of a degree is also required.' Electrically, there 

 are two methods applicable The first, however, based on the relation 

 between temperature and resistance, would have necessitated the measure- 

 ment of increments of the latter quantity amounting t(> scarcely 0.0005 per 

 cent, of the whole, in order to arrive at the accuracy desired. Though 

 even this is feasible in the laboratory, I despaired of being able to reacli 

 such nicety with the means at my disposal. In view of these facts, it was 

 finally determined to try how far a thermo-electric method of research 

 might be successful in answering the question. 



Processes of this kind, in which the effect observed is due to chemical 

 action, are usually accelerated by the application of heat. In other words, 

 the assumption is warranted that the thermal effect of the action of aqueous 

 vapor on feldspar (T. E. K.) will increase, and will therefore be more easily 

 detected as the temperature of the vapor increases; provided, of course, that 

 this temperature is not chosen so high as to dissociate the products of de- 



' A greater diflSculty still •would probably be encoantered from the fact that the reservoir of a 

 thermometer subjected to large differences of temperature is by no means constant in vnliiuie, but sub- 

 ject to variations dependent upon the glass chosen. (Phenomena of "after-action.") 



