356 Professor Forses’s Researches on Heat. 
larization, but not to produce them. The numerical results which 
I gave (I. Series, 44), were intended chiefly to shew that, even 
under all these disadvantages, the effects which I observed were 
of an extremely important and obvious character, such as no slight 
or capricious anomaly could possibly have produced. No one 
who candidly reads the paper, or is aware of the labour of so ex- 
tensive an investigation in so new a field, can suppose that I in- 
tended to give these results as to the quantities of heat, of dif- 
ferent kinds, polarized by passing through mica bundles, as de/i- 
nitive numerical results. I certainly did suppose that the differ- 
ent kinds of heat were polarizable in different degrees under the 
same circumstances, a conclusion which I am now prepared to 
establish. 
18. The extent to which the former paper had swelled, like- 
wise prevented me from inserting the description of a multitude 
of precautionary measures, taken to shew that errors, of whose 
existence I was perfectly aware had no influence in producing 
the effects which I stated; nor am I now going to enter into 
details of manipulation, which the experimentalist must learn for 
himself, and which would be highly tedious to any other reader. 
[ will content myself with recalling two proofs (which I have else- 
where * stated), in justification of my experiments. Could an 
effect similar to polarization have been produced by the con- 
duction of heat, it must have been im the followimg way: My 
earliest experiments were made with bundles of thin mica, A 
and B, Fig. 2, cemented to soles or bases of wood, C and D, 
forming, with the mica, an angle of about 34°. Two of these 
being placed between the source of heat $, and the thermo-elec- 
tric pile P, the fact observed was, that when either of these bun- 
dles was set on edge, as in Fig. 3, so that the planes of refraction 
in the two bundles were perpendicular, /ess heat reached the 
pile than in the first position. If conduction produced this ef- 
fect, it must have been owing solely to the different quantities 
* Philosophical Magazine and Annals, for November 1835 and March 1836. 
