274 Mr. J. R. Cotter on an 
imagine that a cylinder of hot lime, if it could be made air- 
tight, would still be permeable to carbon dioxide. As a rule 
the disintegration produced by chemical action would prevent 
such partitions from being effective for any length of time, 
but still cases of this kind are possible and must therefore be 
considered. | 
The diffusion, which takes place in these cases by the gas 
particles being handed from one molecule to the other, follows 
in general the same kind of laws as those which have first 
been considered. ‘There is, however, one very important 
difference conditioned by the fact that, the reaction being 
reversible, there is a definite dissociation pressure for each 
temperature. When the external pressure is > the dissocia- 
tion pressure, the whole of the superficial layer is turned 
into the compound, so that the solid cannot transmit a 
pressure greater than the dissociation pressure. Hence, if 
we start with a very high pressure on one side of the dia- 
phragm and zero pressure on the other side, the pressure on 
the low pressure side will rise until it is equal to the dis- 
sociation pressure, when no further transference will take 
place. On the other hand, if the pressure on one side is 
always kept at zero, whilst that on the other side is 
capable of taking all values, then the rate of flow through 
will be a uniform function of the pressure up to the dissocia- 
tion pressure, at which there will be a discontinuity, and the 
rate of flow will be independent of the pressure for all 
higher pressures. 



XXVI. AnJnstrument for Drawing Conics. By J. R.CorrEer, 
M.A., Assistant to the Professor of Experimental Physics, 
Trinity College, Dublin*. 
APF OPOS of Prof. Karl Pearson’s article in the ‘ Philo- 
sophical Magazine’ for February, I should like to 
mention that in the year 1894 I designed an instrument for 
drawing conics which has the advantage of always keeping 
the drawing-pen parallel to the direction of the curve. The 
compasses will draw any kind of conic, given the foci and a 
point on the curve. The accompanying figure is drawn from 
the actual instrument, but as it is only a rough home-made 
model it is faulty in construction. 
AGBF' is a rhombus formed of four equal and freely 
jointed flat brass rods. The corner A slides freely along 
the slot of the bar BC. [FH is another flat brass bar pivoted 
* Communicated by Prof. John Joly, F.R.S. 
