of the Specula of Reflecting Telescopes, 261 



must also vary with each magnitude of speculum. I could 

 not describe a temperature for which there is no measure, nor 

 lay down a period of time which must depend partly on that 

 temperature, and partly on the size of the metal, and partly 

 also, I may add, on its form and thickness, and even on the 

 temperature of the atmosphere. But if each maker must 

 feel his way for himself and for each size that he may ca&t, let 

 him always remember, that the speculum metal is rather a 

 mixture than a true alloy, and that it has a tendency to sepa- 

 rate as to crystallize, when under liberty, by heat ; that the 

 best and most brilliant metal is that which cools most suddenly, 

 or which is soonest deprived of the power of crystallizing, and 

 therefore that his object must be to cool his speculum as 

 rapidly as it can possibly be cooled without breaking. And as 

 to the practical method of annealing, I may add, that I think 

 the ordinary mode bad, inasmuch as it is difficult to regulate. 

 Whether my own method may prove better in other hands or 

 not, it is not for me to pronounce, but it was to use sand for 

 that purpose. In this case, the mass of sand was placed over a 

 horizontal iron-plate flue, so as to be at a high red-heat in one 

 place and to cool gradually at another; and while the speculum 

 was placed in the hottest part after casting, it was slowly moved 

 to the cooler parts till the desired end was obtained. 



I have but one other suggestion to offer as to the effects 

 which have seemed to me to follow from polishing, for use, 

 specula, of which the texture had been thus crystallized. 

 Whether there should be two alloys, or an alloy and a metal, 

 or not, or whether the whole result is mechanical, depending 

 on the intersected surface of this confused crystallization, it is 

 certain that a polished surface of this nature is irregular in its 

 action on light, and that good images are scarcely attainable, 

 from optical causes that will be sufficiently obvious, to say 

 nothing of an absolute loss of light produced by this cause. And 

 I have further reason to believe from numerous experiments, 

 which it would be impossible as improper to detail here, that 

 such crystallized specula, undergo further changes of their crys- 

 talline texture from the ordinary vacillations of temperature 

 to which the uncrystallized ones are either less subject, or 

 from which they are exempt ; the consequence being a slow 



