vapor pressure of mercury is only 0.000016. At ordinary temperatures the green line remained plainly visible at hydrogen pressures above 10cm, when only an exceedingly faint, apparently continuous, hydrogen spectrum was visible. The luminosity of mercury vapor, mixed with hydrogen at a given pressure, was found to be nearly proportional to the density of the mercury vapor. The yellow and blue lines appeared only at temperatures

above 10°.

3. The addition of small quantities of mercury vapor to pure hydrogen materially reduced the luminosity of both the elementary and the compound spectrum of the latter. The relative luminosities of the hydrogen and the mercury vapor seemed to be proportional to their relative densities below 6mm hydrogen pressure. At higher pressures of hydrogen the mercury possesses relatively greater luminosity. The addition to hydrogen, free or nearly free from mercury vapor, of 4 per cent. of its own mass (or one molecule of mercury to every 2500 molecules of hydrogen) reduced the luminosity of the entire hydrogen spectrum by more than one half.

4. In tubes with external electrodes pure hydrogen showed a maximum luminosity at a pressure of about 3 mm, the conditions of current supply remaining constant. Tubes with internal electrodes under the same conditions gave no maximum down to pressures of about 0.6mm. The position of this maximum is probably a function of the current strength and the size of the tube.

5. The addition of small quantities of oxygen to hydrogen caused considerable changes in the luminosity of the latter. The luminosity was increased at pressures under I mm, and diminished at higher pressures. The maximum luminosity of the hydrogen was shifted to lower pressures with increasing quantities of oxygen present.

6. Water vapor produces effects quite similar to those of oxygen. Water vapor is probably formed when the discharge passes through the mixture of hydrogen and oxygen.

7. The so-called "compound" spectrum of hydrogen is

really due to hydrogen, not to impurities, as has been often

claimed.

8. A very small quantity of hydrogen added to oxygen will excite luminosity in mercury vapor present. The cause of this is not clear.

My thanks are due to Professor Warburg for his continuous advice and help during this investigation.

PHYSICAL INSTITUTE OF THE UNIVERSITY OF BERLIN,

July 1899.

ON PROFESSOR KEELER'S PHOTOMETRIC RESEARCHES, BY PHOTOGRAPHIC METHODS, ON THE NEBULA IN ORION.

By J. SCHEINER.

IN the March number of the ASTROPHYSICAL JOURNAL, Professor Keeler has given an account of some investigations, in which a photographic method is used to show that the quality of the light emitted by the Orion Nebula is different for different parts of the nebula.

The method is briefly as follows: A negative of the nebula was made on an ordinary plate, with short exposure, by means of a reflecting telescope; and another negative was made, with a longer exposure, on an orthochromatic plate, which was protected from the action of the more refrangible rays by a yellowishgreen plate of glass. The exposure times were so chosen that the bright part of the Huyghenian region appeared equally strong on both plates. The result was, that on the color-sensitive plate the fainter details in the outlying parts of the nebula were much less conspicuous than on the ordinary plate. From this fact the following conclusion was drawn: "We infer, therefore, that in the remote parts of the nebula the two lowest nebular lines are weak, or the hydrogen lines strong, as compared with the Huyghenian region. Thus the results of spectroscopic researches are confirmed, and are extended to parts of the nebula which are too faint for visual observation."

This conclusion involves the theorem, which is also directly stated in another place, that equally dense parts of the negative correspond to equally bright parts of the object. This theorem. is, however, incorrect, and can only be justified under two limitations equal times of exposure and similar plates. These two limitations are necessary, because, in the first place, the product of time and intensity is not a constant, and because in the second

place, the complicated law which is to be substituted for it involves constants which have very different numerical values for different kinds of plates. In the research under consideration both these limits are very greatly exceeded-first, in the variation of the time of exposure from six minutes to two hours and twenty minutes, and second, in that an ordinary plate was brought into comparison with one which had been sensitized with erythrosin. The fact that, even with equal times of exposure, the different degrees of darkening by different grades of brightness may vary widely for different kinds of plates is so generally known, that professional photographers give special orders to factories for "hard" or "soft" plates, each to be used for its appropriate. purpose. A further consideration of this point might therefore be deemed superfluous; but I wish very briefly to sketch an example of such investigations as should have preceded the conclusion which I have cited.

An ordinary Schleussner plate a was exposed in a sensitometer for thirty seconds, and an orthochromatic plate b, protected by a yellow color-screen, for five minutes. The darkest square I of plate b corresponded with the squares 3-4 of a. The faintest perceptible impressions were, for b the squares 1516; for a, 14-15. The interval on b was therefore 15.5 squares. and on a 11 squares; the corresponding intensity intervals are 33 and 11.3. If these plates had been used on the Orion Nebula, the result would, on the assumption of uniform quality of light, have been the reverse of that found by Professor Keeler. On the orthochromatic plate the outlying parts of the nebula could have been traced farther than on the ordinary plate. But we should have to guard ourselves against drawing the conclusion that the hydrogen lines are relatively weaker in the outlying parts than they are in the Huyghenian region; we should rather, on the basis of the experiments with the sensitometer, just mentioned, explain the observed appearance as being in harmony with the fact that in this case the orthochromatic plate was the "softer." With equal intensity of the brighter regions, it had rendered visible portions which were three times fainter than

those which appeared on the ordinary plate; and it is known that even greater differences may occur than those given in this example.

A particular interest, on the other hand, attaches to Professor Keeler's statement that on the orthochromatic plate the Proboscis Major is materially stronger than the streamer which is parallel to it, and which on an ordinary plate nearly equals it in intensity. If these two objects were in fact exactly equal in brightness this experiment would prove the correctness of the views held by Professor Keeler; but if there exists even a small difference of brightness, the conclusion is again inadmissible, since, according to the example of differences of plates, which I have given, small contrasts are vastly exaggerated photographically.

With respect to the method used by Professor Keeler, of weakening the brighter parts of the nebula, I have merely to say that the application of such methods, in the comparison of brightnesses which are already near the limit of visibility, seems to me a doubtful proceeding.

ROYAL OBSERVATORY, POTSDAM.

June 1899.