space enclosed by the ring. Its position with reference to the central star is

It is evidently the star d of Holden's diagram. This star is very distinct on the negative made with 10m exposure, and is just visible on that exposed for 2 m. As it is at the very limit of vision with the 36-inch refractor, it must, like the central star, possess unusual photographic energy.

Holden's star f, at the preceding extremity of the major axis, also appears on the photographs. On some of the plates it is larger than other stars of the same magnitude, and slightly irregular, so that it is perhaps not a true star, but a very small bright patch of nebulosity. The other stars of the diagram seem to be bright patches due to the interlacing of the narrow rings, but in such cases as this the evidence of the visual telescope is perhaps to be preferred.

Barnard's small nebula.-The negative obtained with two hours' exposure shows that the small nebula discovered by Professor Barnard with the 36-inch refractor in 1893 is a lefthanded, two-branched spiral. The extreme diameter on the photograph is about 30.

The positive enlargements (11.1 diameters, Imm = 3.55′), sent with this, in the form of lantern slides, reproduced in Plates VII, VIII, and IX, show most of the details described in the foregoing article.

In all work with the Crossley reflector I have had the efficient assistance of Mr. H. K. Palmer, Fellow at the Lick Observatory.

LICK OBSERVATORY,

August 15, 1899.

Mon. Not. R. A. S., 48, 386.

2 A. N., 134, 130, 1893.

PRESSURE IN THE ELECTRIC SPARK.

By JOHN FRED MOHLER.

In an article published in the ASTROPHYSICAL JOURNAL for June 1899. Eduard Haschek and Heinrich Mache give a method by which they obtain the amount of pressure produced when a spark is passed through a gas. They give results for various media, for different electrodes, and for variation of capacity, spark length, and pressure of the surrounding media.

The work touches some done by Dr. Humphreys and myself1 in that we have found that when the arc is under pressure the period of vibration of the light emitted is a little greater than when the pressure is removed, and consequently the wave-length is increased a little when pressure is added. We have shown in the article referred to above that this change in wave-length or shift of the spectral lines is proportional to the pressure, and in some cases to wave-length, and that it varies with the element. producing the line.

The measured shift of the lines of a few of the elements under twelve atmospheres pressure, together with the shift reduced to wave-length 400, is given in the following table:

Using these data as a starting point it would seem possible to measure the pressure at the source of light if we knew the displacement of the lines. This assumes that a spectral line will be displaced by pressure whatever may be the source of light producing the line. The question of the amount of pressure becomes very interesting in the light of the recent experiments

of Professor J. Wilsing,' who obtained a spectrum very similar to that of the new stars by passing a spark between electrodes immersed in water.

My experiments described below were made to test the results obtained by Haschek and Mache. If their results are near the truth the shift of the lines due to pressure should be very considerable; indeed, as in many experiments they found a pressure of more than fifty atmospheres, the shift of the lines should be four or five times the largest shift we found in previous experiments with the arc, and the displacement due to one atmosphere is a measurable quantity with some elements.

I used in this investigation a four-inch concave Rowland grating, mounted in the usual way on piers of solid masonry, in a room of nearly constant temperature. Photographs of the spectrum under consideration were taken along with the spectrum of the arc, at atmospheric pressure, for comparison. The camera was very solidly mounted and the plate was exposed to the arc spectrum both before and after it was exposed to the spark. For varying the pressure around the spark and for investigating the effect of different gases the electrodes were inclosed in a heavy brass vessel similar to that used in the previous work. This vessel was mounted with the arc lamp on a swinging frame, so arranged that either spark or arc could be focused on the slit of the spectroscope without touching the mounting of the grating. For capacity I used a series of jars, some large and some small. Their capacity was not very accurately determined, but the error cannot be more than 4 or 5 per cent. The induction coil used was capable of giving an eight-inch spark when used without capacity. The spark gap was usually 3 mm or less. A filar micrometer in connection with a very low power microscope was used to measure the displacement of the lines. As the previous work had shown that cadmium gives a relatively large displacement under pressure I used that metal in most of my experiments.

1

Sitz. d. K. Akad. d. Wis. zu Berlin, No. 24, May 4, 1899; this JOURNAL, 10, 113, 1899.

2 This JOURNAL, 3, 116, 1896.

CAPACITY.

Messrs. Haschek and Mache found that under given conditions the pressure produced when the spark is passed through air varied with the capacity in a peculiar way. As the capacity increased the pressure increased to a maximum of fifty-one atmospheres and then decreased. Below is part of their table

giving the relation of capacity to pressure:

With cadmium electrodes and at atmospheric pressure I found the displacement of the green and blue lines with varying capacity to be as given in the following table. The measured shift is given in thousandths of an Ångström unit, the capacity is given in meters. Considering 0.008 as the measured shift of the cadmium lines per one atmosphere pressure, I give in the same table the pressure in atmospheres deduced by this method.

The above table shows that the pressure calculated from the shift of the lines is very much smaller than that given by Haschek and Mache. Considerable allowance must be made for errors in measurement of the shift, as the spark lines, with condensers in the circuit, become broader as the capacity is increased.

The last result with capacity of 70.2 meters gives but a slightly greater shift than a capacity of 48 meters. This seems to indicate that the pressure does not increase directly with capacity, but as indicated by Haschek and Mache, the pressure approaches a maximum value.

The effect of capacity on the position of the iron lines of the spark spectrum was also investigated. With small pieces of steel as electrodes several photographs were taken with a capacity of 22.2 meters. About twenty lines on these plates were measured. The shift was small, and the average measured displacement was 0.011 Ångström unit. The previous work indicated that the shift per atmosphere was about 0.002 Ångström unit for the iron lines. This would indicate a pressure in the spark of 5.5 atmospheres which is the average found for cadmium. for the same capacity.

PRESSURE.

The effect of the pressure of the surrounding medium is shown by comparing the shift produced by the spark with a definite capacity when the pressure about the electrodes is one atmosphere with the shift produced when the pressure is four times as much, the capacity remaining the same. The capacity used for this experiment was 22.2 meters. The pressure was four atmospheres and the measured shift was 0.160 Ångström unit, corresponding to a pressure in the spark of 20 atmospheres. This, compared with the shift at atmospheric pressure, shows that the pressure in the spark varies very nearly with the pressure of the surrounding medium. This is altogether different from the results given by Haschek and Mache, who find that the pressure in the spark increases very much faster than the pressure in the surrounding medium.

The effect of the kind of gas surrounding the electrodes as given by Haschek and Mache indicates that carbon dioxide had three times the effect of atmospheric air, and strangely enough, illuminating gas with a density of 0.47 had produced a greater pressure than air.