Traces of paramagnetic oxygen, for example, would completely invalidate the results. The non homogeneous field method may be employed, the force exerted on bulbs filled with the gas being measured; or the gas may be used as the surrounding medium with a suitable test body. In the Gouy method, as applied by T. Sone, a cylinder is divided into two compartments by a horizontal partition. The lower half is filled with a liquid or gas of known susceptibility, or is evacuated, the upper half with the gas under investigation. The cylinder is suspended from the arm of a special balance, the partition being midway between the poles of a magnet. The force depends on the relative susceptibility of the substances in the two compartments. The Quincke method, elaborately modified, has been used by E. Bauer and A. Piccard for the determination of the susceptibility of oxygen. In another arrangement the susceptibility was deter mined from the difference in the pressure of the gas at two points one of which is in a field H, the other in zero field, this difference being equal to -1-K A neat modification of the Quincke method has been devised by A. P. Wills and G. Hector. An 0 tube is em ployed, the liquid being a solution of nickel chloride. Water being diamagnetic, and the salt paramagnetic, solutions of any desired susceptibility at a particular temperature may be made up; the variation with temperature is known from other experi ments. The volume susceptibility of the gas is proportional to the pressure ; if the gas and the solution have the same suscepti bility there will be no movement of the meniscus when the field is switched on. Magnetic balance, indicated by absence of move
ment of the meniscus, could be brought about by varying the pressure of the gas at a particular temperature, or varying the temperature at a particular pressure, the susceptibility of the gas being then equal to that of the solution. A diagram of the ap paratus is shown in fig. 25.
The field is applied at the meniscus, but the meniscus itself was not observed; instead, the microscope was focussed on gum mastic particles in the solution at the constricted portion of the tube. This "indicator" was very sensitive to any movement of the meniscus on application of the field.
For comparative measurements at low pressures a method sug gested by W. Wien has been used by A. Glaser. A small para magnetic test rod is suspended by a delicate fibre in a non homogeneous field. In general, when the field is switched on, the test rod will be subject to a couple depending on the difference of the susceptibility of the test rod and that of the surrounding medium it will be deflected, but may be restored to its zero position by turning the torsion head through an angle a,.; and Let be the angle corresponding to a vacuum, a, that to a gas of known susceptibility to the gas under investigation of susceptibility K2.
method is sensitive, but at very low pressures, unless special precautions are taken, spurious effects may occur owing to the presence of absorbed layers on the test rod.
Up to about 1920 measurements of most gases were unreliable, but since then investigations have been undertaken with an appre ciation of the precautions necessary; the experimental difficulties have been surmounted and for a number of gases satisfactory data are available.