Sources for Spectroscopic Observation

There is also a peculiar slight disturbance in the positions of some of the lines near the poles of an arc, called the "pole effect," which makes it necessary to establish a standard source when the iron arc is used to provide standards of wave-length. In the "Pfund arc," which has been recommended by the Inter national Astronomical Union, the anode is below and consists of a bead of iron oxide supported on a massive rod of iron, while the cathode above is a rod of iron 6-7 mm. in diameter, having a massive cooling cylinder of copper or brass close to the end of the rod. The current may be five amperes or less at r ro-25o volts, and the arc 12-14 mm. in length. Only the central zone at right angles to the axis of the arc, not exceeding z i mm. in width is to be used.

The "vacuum arc" is an important modification of the ordinary arc in which the arc is enclosed in a vessel which can be ex hausted. Under these conditions the spectrum tends to approach that given by the spark, and the lines are very sharply defined. The same arrangement may also be used for passing the arc in any desired gas, such as hydrogen.

The Electric Spark.

The luminosity for observations of spark spectra is usually produced by discharges from an induction , coil between terminals made ofthe substance to be examined, or between terminals of platinum on which the substance is sup ported. Transformers suitable for the purpose are also available. As shown in fig. 6, a Leyden jar or other condenser, J, is con nected in parallel with the leads A and B from the coil to the terminals, S, which are separated by only a few millimetres. Without this condenser, the luminosity is very feeble and the spectrum consists almost entirely of bands of nitrogen.

The intensity of the discharge can be greatly varied by making use of an auxiliary spark gap, shown at C in fig. 6; the longer this gap the greater the potential to which the condenser must be charged before the discharge will pass.

The condensed spark yields lines special to the substance under examination, and, in addition, lines due to atmospheric nitrogen and oxygen. The air lines are common to all the spectra observed in this way and can therefore readily be identified is such. Air lines can also be eliminated by passing the spark in I closed vessel containing hydrogen, the spectrum of which is much simpler than that of air. The spectra of gases surrounding

he spark can be abolished altogether by introducing a self Mduction coil in series with the spark, as between C and S in 6, in accordance with the investigations of Schuster and Elemsalech.

Vacuum Tubes.

The spark discharge as described above can :learly be used to obtain the spectra of gases. Under these con litions, however, the gas lines are mostly too broad for accurate measurement, and observations are accordingly more generally made with the gases at reduced pressures in so-called "vacuum tubes." Such tubes are made in great variety according to the object of the experiments undertaken. Fig. 7a represents a form called the Geissler or Plikker tube, which can be purchased ready prepared for spectroscopic examination. It consists of two glass tubes of about three inches in length and half an inch in diameter, connected by a short length of capillary, and having a platinum wire sealed in at each end. The tube is exhausted and the desired gas admitted through a side tube, which is sealed off with a blow pipe when suitable conditions are established. When the electrodes are connected with an induction coil, a bright discharge passes through the tube, varying in colour and spectrum with the gas which has been introduced. The discharge is especially brilliant in the capillary tube, where the current density is greatest, and the spectrum shown by this is usually different from that given by the wider parts of the tube. The spectrum of the same gas also varies according as the discharge is passed with or without a condenser in the secondary circuit, the changes being com parable with those observed in passing from arc to spark.

Another common form of tube is shown in fig. 7b ; this is designed to give an "end on" view along the capillary, and since glass is not transparent to ultra-violet light, a side tube with a quartz window is often added as shown in the diagram. With this form of tube the spectrum appears with greater intensity. Some times the whole vacuum tube is made of fused silica glass. In most experimental work, the vacuum tubes are not sealed, but remain in connection with the exhaustion pump and the source of the gas under investigation, so that the pressure and other con ditions can be conveniently controlled. Much care is necessary to obtain pure gases in the tubes.