the line wave-length 5535 readily reversed, while that with wavelength 5518 is less easily reversed; the line wave-length 4933 is comparatively easily reversed, whereas that with wave-length 4899 has not been reversed by us. On the other hand, the line wave-length 4553 has been reversed, but not the line wave-length 4524. In the case of strontium, the lines wave-length 4831 and 4812 have been reversed, but not the line wave-length 4784, and the two lines wavelength 4741 and 4721 remain both unreversed. In the group of five lines of calcium, wave-length 4318 to 4282, it is only the middle line wave-length 4302 which has been reversed. Of the potassium groups of lines wave-length 5831 and 5782, 5802, 5782 are reversed, the line wave-length 5811 has not been reversed, and of the others the line wave-length 5802 is the first to appear reversed. It is worthy of remark that the first of these lines is faint and the last is the brightest of the group. The group wave-length 5355, 5336, 5319 have been all reversed, but the last of the three (5319) was the most difficult to reverse: it is also the feeblest of the group. In the more refrangible group, wave-length 5112, 5095, 5081, the least refrangible is the only one reversed.

Making a general summation of our results respecting the alkaline earth metals, potassium, and sodium, and having regard only to the most characteristic rays, which for barium we reckon as 21, for strontium 34, for calcium 37, for potassium 31, and for sodium 12, the reversals in our experiments number respectively 6, 10, 11, 13, and 4. That is in the case of the alkaline earth metals about one-third, and these chiefly in the more refrangible third of the visible spectrum, the characteristic rays remaining unreversed in the more refrangible part of the spectrum being respectively 2, 5, and 4. In the case of potassium we reversed two in the upper third, all the rest in the least refrangible third. These experiments relate to mixtures of salts of these metals combined with the action of reducing agents. In a future communication we will contrast these results with those of the isolated metals, calcium, strontium, and barium.

IV. "Note on the unknown Chromospheric Substance of Young." By G. D. LIVEING, M.A., Professor of Chemistry, and J. DEWAR, M.A., F.R.S., Jacksonian Professor, University of Cambridge. Received March 27, 1879.

In the preliminary catalogue of the bright lines in the spectrum of the chromosphere published by Young in 1861, he calls special attention to the lines numbered 1 and 82 in the catalogue, remarking that "they are very persistently present, though faint, and can be distinctly seen in the spectroscope to belong to the chromosphere, as such, not being due, like most of the other lines, to the exceptional elevation

of matter to heights where it does not properly belong. It would seem very probable that both these lines are due to the same substance which causes the D, line."

Again, in a letter to "Nature," June, 1872, Young says, "I confess I am sorry that the spectrum of iron shows a bright line coincident with 1474 (K); for, all things considered, I cannot think that iron vapour has anything to do with this line in the spectrum of the corona, and the coincidence has only served to mislead. But there are in the spectrum many cases of lines belonging to the spectra of different metals coinciding, if not absolutely, yet so closely, that no existing spectroscope can separate them, and I am disposed to believe that the close coincidence is not accidental, but probably points to some physical relationship, some similarity of molecular constitution perhaps, between the metals concerned. So, in the case of the green coronal matter, is it not likely that though not iron it may turn out to bear some important relation to that metal ?" In 1876 he proves that the coronal line 1474 is not actually coincident with the line of iron.

In the catalogue of bright lines observed by Young at Sherman in the Rocky Mountains, to which we have directed special attention in one of our previous communications, it appears that the above-mentioned lines 1 and 82, along with D3, were as persistently present as hydrogen, the only other line approaching them in frequency of occurrence being the green coronal line 1474 of Kirchhoff, which was present on 90 occasions out of 100. It has occurred to us that these four lines may belong to the same substance. An analogy in the ratio of the wave-lengths of certain groups of lines occurring in different metals has been already pointed out by Stoney, Mascart, Salet, Boisbaudran, and Cornu; and without any special reductions, or claims to an exact ratio in whole numbers, the following analogies are worthy of note :

Hydrogen.

Wave-length :

(1) 6563.9

:

:

(2) 4862.1

(2) 6102

(2) 3837.8

(3) 4340

(3) 4970

(3) 3335

(4) 4102·4

(4) '4604

(5) 4130

*This wave-length is not so accurately known as the other rays belonging to the chromosphere.

The ratio of the wave-lengths of F to G of hydrogen ((2) to (3) in the table above) is nearly identical with the ratio of D, to the coronal green line ((2) to (3) in table above).

This near coincidence in the ratios of certain lines of hydrogen, lithium, and magnesium, substances belonging to the same type, combined with a similar ratio in the wave-lengths of the nearly equally persistent lines of the chromosphere, greatly strengthens the probability of the assumption that these lines belong to one substance.

The fact that the two less refrangible rays have no representative in the Fraunhofer lines, is by no means opposed to their belonging to one substance, since we know that aluminium behaves in a similar way in the atmosphere of the sun; and in the total eclipse of 1875 the hydrogen line h was not visible in the chromosphere, that is, we suppose, was on the limit between brightness and reversal; and during the late eclipse the two most refrangible rays of hydrogen were not detected from the same cause.

Until our knowledge of the order of reversibility of lines belonging to different types of metals has been extended, it would be rash to infer the group of metals to which it belongs, or its probable molecular weight.

V. "Contributions to Molecular Physics in High Vacua." By WILLIAM CROOKES, F.R.S. Received March 27, 1879.

(Abstract.)

This paper is a continuation of one "On the Illumination of Lines of Molecular Pressure, and the Trajectory of Molecules," which was read before the Royal Society on the 5th of December last. The author has further examined the action of the molecular rays electrically projected from the negative pole in very highly exhausted tubes, and finds that the green phosphorescence of the glass (by means of which the presence of the molecular rays is manifested) does not take place close to the negative pole. Within the dark space there is absolutely no phosphorescence; at very high exhaustions the luminous boundary of the dark space disappears, and now the phosphorescence extends all over the sensitive surface. Assuming that the phosphorescence is due either directly or indirectly to the impact of the molecules on the phosphorescent surface, it is reasonable to suppose that a certain velocity is required to produce the effect. The author adduces arguments to show that within the dark space, at a moderate exhaustion, the velocity does not accumulate to a sufficient extent to produce phosphorescence, but at higher exhaustions the mean free path is long enough to allow the molecules to get up sufficient speed 2 N

VOL. XXVIII.

to excite phosphorescence. At a very high exhaustion there are fewer collisions, and the initial speed of the molecules close to the negative pole not being thereby reduced, phosphorescence takes place close to the pole.

Experiments are described in which a pole folded into corrugations is used at one end of a tube, the pole at the other end being flat set obliquely to the axis of the tube, and having a plate of mica in front pierced with a hole opposite the centre of the pole. The questions which this apparatus was designed to answer are:-(1.) Will there be two sets of molecular projections from the corrugated pole when made negative, one perpendicular to each facet, or will the projection be perpendicular to the electrode as a whole, i.e., along the axis of the tube? (2.) Will the molecular rays from the oblique flat pole, when this is made negative, issue through the aperture of the screen along the axis of the tube, i.e., direct to the positive pole, or will they leave the pole normal to the surface and strike the glass on its side? With the corrugated pole experiment shows that at high exhaustions molecular rays are projected from each facet to the inner surface of the tube, where they excite phosphorescence, and form portions of ellipses by the intersection of the planes of molecular rays with the cylindrical tube. When the oblique flat pole is made negative, a stream of molecules shoots from it nearly normal to its surface, and those which pass through the hole in the plate of mica strike the side of the tube, forming an oval patch of a green colour.

The oval patch in this apparatus happens to fall on a portion of the glass which has previously had its phosphorescence excited by the molecular discharge from the other corrugated pole. The phosphorescence from this pole is always more intense than that from the flat pole, and the glass, after having been excited by the energetic bombardment, ceases to respond readily to the more feeble excitement from the flat pole. The effect, therefore, is, that when the oval spot appears, it has a dark band across it where the phosphorescence from the other pole had been taking place. The glass recovers its phosphorescent power to some extent after rest.

In this apparatus a shifting of the line of molecular discharge is noticed. If the coil is stopped and then set going repeatedly, always keeping the oblique pole negative, the spot of green light occurs on the glass at the spot where it should come supposing the discharge were normal to the surface of the pole. But if once the flat pole is made positive, the next time it is made negative the spot of light appears nearer the axis of the tube, and instantly shifts to its normal position, where it remains so long as its pole is made negative. There seems no limit to the number of times this experiment can be repeated.

A suggestion having been made by Professor Stokes that a third, idle, pole should be introduced between the negative and positive elec

trodes, experiments are described with an apparatus constructed accordingly. The potential of the idle poles (of which there are two) at low exhaustions is very feebly positive; as the exhaustion gets better the positive potential increases, and at a vacuum so good as to be almost non-conducting, the positive potential of the idle poles is at its greatest. The result is that an idle pole in the direct line of fire between the positive and negative poles, and consequently receiving the full impact of the molecules driven from the negative pole, has a strong positive potential.

It is found that when the shadow of an idle pole is projected on a phosphorescent screen, the trajectory of the molecules suffers deflection when the idle pole is suddenly uninsulated by connecting it with earth. The same result is produced by connecting the idle pole with the negative wire through a very high resistance, such as a piece of wet string, instead of connecting it with earth. A tube, which has already been described in a paper read before the Royal Society on December 5th last, is used to illustrate this deflection. The shadow of an aluminium star is projected on a phosphorescent screen. So long as the metal star is insulated the shadow remains sharp, but on uninsulating the star by connecting it with an earth wire the shadow widens out, forming a tolerably well-defined penumbra outside the original shadow, which can still be seen unchanged in size and intensity. On removing the earth connexion the penumbra disappears, the umbra remaining as before.

It is also found that the shadow of the star is sharply projected when it is made the positive pole, the negative pole remaining unchanged.

These experiments are explained by the results just mentioned, that the idle pole, the shadow of which is cast by the negative pole, has strong positive potential. The stream of molecules must be assumed to have negative potential; when they actually strike the idle pole they are arrested, but those which graze the edge are attracted inwards by the positive potential and form the umbra. When the idle pole is connected with earth, its potential would become zero were the discharge to cease; but inasmuch as a constant supply of positive electricity is kept up from the passage of the current, we must assume that the potential of the idle pole is still sufficient to more than neutralize the negative charge which the impinging molecules would give it. The effect, therefore, of alternately uninsulating and insulating the idle pole is to vary its positive potential between considerable limits, and consequently its attractive action on the negative molecules which graze its edge. The result is a wide or a narrow shadow, according to circumstances.

After a definite shadow is produced, it is found that increasing the exhaustion makes very little change in the umbra, but it causes the