Day and Shepherd — Lime-Silica Series of Minerals. 269 



time factor is always very important in dealing with a Seger 

 cone. Bondonard himself states (p. 343) : U A very small differ- 

 ence in temperature, or a few minutes additional heating* 

 often suffices for the softening stage to pass into one of complete 

 fusion." If this statement was made understandingly, the 

 method merited rejection by Boudouard himself. Furthermore, 

 the use of Seger cones for exact work will always be unsatis- 

 factory because it depends upon the judgment of the operator 

 to say when a cone has "fallen" sufficiently to be considered 

 melted, and different observers almost always obtain widely 

 different results under like conditions. As has been pointed 

 out by Day and Allen, f any method of measurement which is 

 not based upon some reasonably sharp physical change must be 

 expected to give different results in different hands,;}; Suppose 



* Italics are ours. 



f Arthur L. Day and E. T. Allen, "The Isomorphism and Thermal Properties 

 of the Feldspars";'' this Journal (4), xix, p. 93, 1905. Zeitschrift t Phys. 

 Chem. liv, p. 1, 1905. Publication of the Carnegie Institution of Washing- 

 ton, No. 31. 



\ Since the text of this paper was written, a very recent paper by Doelter 

 (C. Doelter, "Die Untersuchungsmethoden bei Silikatschmelzen," Sitzungs- 

 ber. d. Wien. Akad. cxv, 1, May, 1906) has come to our attention in which he 

 reaffirms his confidence in and preference for subjective methods for the 

 investigation of silicate solutions, — more particularly the viscous silicates. 

 This question of methods of attack in problems of wide scope and considera- 

 ble difficulty is not an academic one ; it is a matter of the very first importance, 

 particularly in view of the increased attention which is coming to be paid 

 to the' minerals as solutions. We have preferred to avoid subjective methods 

 wherever possible on the general ground that no observation so made is 

 exactly reproducible. Subjective observations are therefore always much 

 more satisfactory to the observer than to any one else. Prof. Doelter has 

 probably had a greater and more varied experience in the observation of 

 mineral melting points than anyone now living, and he is therefore able to 

 form consistent judgments upon the changes which he observes. But even 

 under his exceptionally, competent hand we have seen the "melting points" 

 of the feldspars rise a little higher each year in his successive publications 

 upon the subject, and the feldspars are very viscous minerals of the type to 

 which he finds the optical methods especially adapted. 



Doelter then criticises the thermoelectric methods in use in this laboratory 

 on the ground of inexactness, i. e., because the recorded time-temperature 

 curves contain no period of absolutely constant temperature, although he 

 appears to be very familiar with the fact that the phenomena themselves are 

 not sharp. This seems to be unfortunate and unproductive criticism. Under 

 fair conditions a phenomenon is obviously the same, whether observed by 

 looking at the charge or by exposing a thermoelement in it. If the supposed 

 melting "point" does not occur at a point, it cannot be recorded as such. 

 So far from failing in its purpose, therefore, the thermoelement has revealed 

 to us a hitherto unfamiliar phenomenon with great fidelity. Our principle 

 in the choice of methods is therefore diametrically opposed to Doelter's. If 

 the change of state were sharp and well marked, it would really matter very 

 little how it was determined. If, on the other hand, it is a slow change, we 

 should greatly prefer the unprejudiced record of a thermoelement if it could 

 be obtained. Nothing is so difficult of observation or gives rise to so much 

 difference of opinion between observers as a slow-moving phenomenon. An 

 interesting example of this is to be found in this (Doelter's) paper itself (p. 

 12). Nearly all observers of the constants of silicates are now completely 

 agreed that the glasses are merely undercooled liquids which of course have 

 no melting point but change continuously from a hard amorphous to a soft 



