in the rays corresponded with increasing brightness of the second spectrum.

Since hydrogen in the normal state does not absorb the second spectrum, the latter cannot be due to collisions of neutral molecules which do not result in ionization; but beyond this there is at present no sufficient evidence to determine the state of charge of the molecules before or after the collisions which lead to its emission.

4. THERESTING" SPECTRUM.

It is clear from an inspection of Pl. IV. No. 1, that when atoms only are present in the positive rays, the second spectrum occurs only very faintly, if at all. It is probable, however, that even in this case it is not entirely absent (a very faint blackening can be detected on the negative of No. 1 between Hg and Hy, though there is no trace of any definite lines), and is presumably due to the rays exciting the spectrum of the gas through which they pass (the "resting" spectrum of Stark). When molecules are present the second spectrum occurs in considerable quantities, and one would naturally expect it to show the Döppler effect if examined along the direction of motion of the rays. Stark and Wilsar (loc. cit.) have found that the second spectrum shows no such effect. To account for this, one of two suppositions must be made. Either, which seems rather improbable, the second spectrum is only emitted as the result of collisions between molecules moving comparatively slowly with respect to each other-i. e., rays near the end of their path; or Stark and Wilsar never got molecules in their positive rays at all. The second hypothesis seems the most likely. They used a narrow tube which is not favourable to the development of molecules, and in Stark's account of his experiments with the second spectrum the potential differ ence used in the discharge tube never seems to have exceeded 7500 volts. In all the work done at Cambridge on positive rays fairly powerful induction coils have been used.

When a discharge is produced by these means in a tube. at the low pressures found favourable for the production of molecular rays, the potential difference is much greater than 7000 volts. Stark used a wide range of pressures, but it seems not improbable that his comparatively low-potential batteries are less effective in producing molecular rays than the coil discharges used in the Cambridge experiments.

5. EFFECT OF MAGNETIC FIELD ON THE DISCHARGES.

A curious phenomenon was noticed with the discharge tube illustrated in fig. 2. This tube gave only atomic rays under ordinary conditions except at very low pressures, and sometimes even then the molecule was barely visible. When, however, the discharge was influenced by an electromagnet held near it in a certain position, the molecular parabola would suddenly appear. This only happened over a certain range of pressure, and at the pressure at which Pl. IV. No. 1, was taken it had no effect. In addition to bringing out the molecular parabola the magnet always increased the deviation of the atomic parabola, and as this depends chiefly on the cathode fall of potential, it presumably made the tube go easier; while the atomic parabola was generally long and frequently beaded, indicating a wide range of velocities, the molecular parabola produced by the magnet was always short, being little more than a patch of light at the place corresponding to the greater velocity. The magnet had the effect of spreading out the negative glow, which was originally concentrated near the mouth of the anode tube, and bringing it forward towards the cathode. In addition to this effect the magnet would occasionally brighten the atom as well, but careful inspection showed that this was always due to a better centring of the beam of positive rays on to the fine tube, owing to slight deviations of the rays themselves by the magnet.

6. INTENSITY OF SECOND SPECTRUM LINES.

The brighter lines in the second spectrum generally occur in groups of two or three, too close together to be separated with the dispersion used. It may, however, be of interest to note that lines (e. g., the group near 4680) which Merton* found abnormally weak in a mixture of helium and hydrogen occur quite strongly in the positive rays. Unfortunately, the region examined does not contain any of the lines which Merton found intensified under his conditions.

* Proc. Roy. Soc. A, vol. xcvi. p. 382.

SUMMARY.

1. The Balmer series is produced when positively charged hydrogen atoms pass through hydrogen as positive rays.

2. The second spectrum is produced when positively charged hydrogen molecules pass through hydrogen as positive rays.

3. Stark and Wilson's failure to find the Döppler effect for the second spectrum of hydrogen is probably to be explained by there being no molecules in the positive rays they used.

4. In certain circumstances a great change can be made in the nature of the positive rays given by a discharge tube by subjecting it to a magnetic field.

In conclusion, I wish to express my sincere thanks to my father, Sir J. J. Thomson, for his interest and advice during the progress of the work described above.

XXVIII. Proceedings of Learned Societies.

GEOLOGICAL SOCIETY.

[Continued from vol. xxxix. p. 699.j

December 17th, 1919.-Mr. G. W. Lamplugh, F.R.S., President, in the Chair.

THE following communication was read:-

'A Rift-Valley in Western Persia.' By Prof. S. James Shand, D.Sc., F.G.S.

Asmari Mountain, near the oilfields of Maidan-i-Naftun, in the Bakhtiari country of Western Persia, is an inlier of Oligocene limestone among the beds of the Fars System (Miocene), the latter consisting, in the lower part, of bedded gypsum with intercalated shales and a few thin limestones. The mountain is a whale-back, 16 miles long and 3 miles wide at the middle, formed by a simple symmetrical anticline plunging at both ends. The north-western end plunges rather steeply, and shows no abnormal structures; but at the south-eastern end the fold has collapsed along its length for a distance of 3 miles, letting the gypsum-beds down into a trough in the limestone.

This trough is bounded by two main faults hading north-eastwards and south-westwards respectively, with an average hade of 20°, and marked by steep escarpments. The northern scarp, which lies practically along the axis of the anticline, is at one point 500 feet high; but the southern one, being low down on the flank of the

anticline, is much less conspicuous. Besides these main faults, there are at least three other big faults parallel to them, which produce minor scarps within the valley; the valley-floor thus descends in terraces towards the south-west, besides having a general south-eastward inclination of some 10°. Uphill, towards the crest of the mountain, the downthrow of the limiting faults diminishes gradually to zero, and the valley dies out on the broad top of the anticline. Downhill, towards the plunging nose of the anticline, the trough is closed abruptly by a cross-fault nearly at right angles to the anticlinal axis. The length of the whole trough is 24 miles, and its width half a mile.

The northern boundary-fault at its maximum has a downthrow of about 500 feet, the parallel faults range from 150 to 200 feet, and the cross-fault throws about 100 feet.

The gypsiferous beds which once completely filled the trough have been partly removed by erosion, clearly revealing the faultwalls in the lower part of the valley. These fault-scarps in their lower portions are remarkably fresh, and show the smoothed and fluted surfaces produced by the friction of the sliding faces.

The drainage of the faulted region is curious in several respects. The only perennial stream that traverses the valley cuts sheer across it from side to side. Rising in the gypsum-beds on the north-eastern flank of the anticline, the stream turns southwestwards, and cuts right through the limestone in a deep cañon ; the latter breaches the northern fault-scarp where the downthrow is greatest, and from here the bed of the stream lies in the gypsum of the valley-bottom, except for a short stretch between two faults, where the limestone-floor is again exposed to view. The stream finally breaks through the southern limestone-wall, and so makes its escape from the valley.

These conditions imply that the river assumed approximately its present direction when there was a cover of gypsiferous beds over the whole area, and that, although younger than the initiation of the faulting, it is older than the sculpturing of the inside of the fault-trough. The eastern end of the trough is again tapped by another and smaller stream with a precisely similar development.

It might be surmised from the excellent preservation of the fluted surfaces on the fault-scarps that the last movement took place at a very recent date, but the behaviour of the river in crossing the scarps does not favour this supposition. The succession of events appears to have been as follows:

(1) Formation of the trough with gypsum-infilling by partial collapse of the anticline.

(2) Levelling of the gypsum surface and development of a stream across the position of the buried trough.

(3) The stream, cutting down through gypsum, discovers the faults, etches them out, develops subsequents along them, and thus gradually eats out the gypsum filling the rift, until the present topography results.