314 
PROFESSOR TYNDALL ON THE ACTION OF FREE MOLECULES ON 
iodide of ethyl, while both of them were more diathermanous than benzol. In 1880, 
the case was precisely the reverse. Suspecting that the discrepancy might be due to 
impurity, I requested my friend Professor Dewar to subject the liquids to a further 
process of purification. Tested afterwards, they produced the following deflections :— 
ABC 
Bisulphide of Carbon. 5'0° 5’0° 4 , 0 C ' 
Chloroform. 17 - 0 15’0 6'0 
Iodide of Methyl. 39'4 33*0 8'0 
Iodide of Ethyl. 33'0 35'0 12'5 
Amylene. 42'0 4P0 16’0 
Sulphuric Ether. 44'3 43'5 18’2 
Acetic Ether. 46‘2 4-5'5 22'0 
Formic Ether. 47'5 46'9 22 - 2 
Under A and B respectively are the deflections produced by the liquids prior to and 
after purification. The iodide of methyl falls from 39*4° to 33°, while the iodide of 
ethyl rises from 33° to 35°. The relative positions occupied by the liquids in 1864 are 
thus restored. Benzol, however, remained permanently lower than before. Under C 
are the deflections produced by the vapours of the purified liquids. Here also the 
positions of the two iodides are reversed, vaporous absorption following the order of 
liquid absorption. I have frequently encountered cases of this character. The simul¬ 
taneous change of diathermic position of liquid and vapour indicates that the foreign 
ingredient, whatever it was, possessed approximately the same volatility as the sub¬ 
stance which it vitiated. 
§. 5. New Experiments. Hypothesis of Internal Films, 
a. The Experimental Tube. 
I wish now to come to closer quarters with the argument urged by Magnus and 
repeated by MM. Lecher and Pernter, namely that my results were due to “vapour - 
hesion ”—that is to say to liquid films condensed on my experimental tube and on my 
plates of rocksalt. The two investigators last named express unqualified surprise that 
I should have neglected the simple precaution of experimenting with blackened tubes. 
But this precaution was by no means neglected by me. I have repeatedly fortified 
myself by experiments of this character, with the result recorded in Table I. But I 
went further. A smooth coating of lampblack, however powerful as an absorber, 
might be held competent to reflect a certain portion of the incident heat. Hence my 
desire to get entirely rid of reflection, by avoiding all contact with the interior surface 
of the experimental tube. 
In fig. 3 an apparatus is sketched by means of which this has been accomplished. 
T U is a stout tube of brass, 36 inches long and of 6 inches internal diameter. Pro¬ 
jecting from its otherwise closed ends are screws 1 inch each in depth, surrounding 
