482 Scientific Intelligence. 
ultra red rays, or in other words to the thermocrosis of the several 
ie OSE employed. 
might be éxpected, the differences of qualities just described 
are es out most strikingly when coatings of opposite classes 
are balanced against each other on the vanes of the movable fl 
of a radiometer, and this instrument can be constructed with a 
movable glass cap made tight by cement, so that the fly may be 
removed and the vanes shifted at pleasure. We quote from Mr. 
Crookes’s papes the following description of the singular results 
he thus obtai 
“ Disks pace on alternate sides with chromic oxide and pre- 
cipitated selenium move in one direction to the naked flame of a 
candle and in the other direction when a water screen is inter- 
posed, ith saffranin and hydrated zinc oxide, the instrument 
does not move at all when exposed to the naked flame, but re- 
volves when a water screen is interposed. With thallic oxide and 
Magnus’s green platinum salt, the instrument moves strongly 
when no screen is interposed, but is stopped te a water-screen. 
These results are all in conformity with the figur 
“A pith radiometer coated with sates pa tae and 
chromic — was exposed to the radiation from a colorless gas 
e of a Bunsen burner, and also to the same colored intensely 
the repulsion of the chromic a was however gy wagens as 
strong as when the non-luminous flame was used. This experi- 
to a ray of iene polarized in one plane and white to aray ecient 
action does not exist to ay appreciable degree. He experimen- 
= on a plate of tourmaline suspended in vacuo on a torsion bal- 
nee, and measured the amount of repulsion produced by a beam 
of polarized light when in different planes, and he explains the 
negative results by sayin ge while the repulsion resulting from 
radiation is almost entirely a surface action, the a of a tour- 
maline on a polarizer is le in ger thickness i is necessary. 
The analysis which Mr. Crookes next gives of ae effect of the 
shape of the vanes of a radiometer in influencing the amount and 
direction of che eae = very ingenious, and highly ngrregaaec 
but the discussion cannot be made intelligible without the nume- 
rous figures with which ais papers are illustrated. It is silico 
to rad that the effects are shown to be in accordance with the 
