* Position angle of slit 346° (about perpendicular to the general direction of the ellipse).

A brilliant elliptical ring with outlying lobes beyond the extremities of its minor axis. There are two straight threads of light running outward from the nebula in the direction of the major axis, which terminate in minute knobs, or condensations. The object suggests the appearance of Saturn when viewed with the rings "edge on."

The spectrum is shown to fair advantage in plates XLII and XLVIII.20 It will be seen by a glance at the photographs made with the slitless spectrograph that most of the light of wavelength 3727A is concentrated in two small areas at the extremities of the ellipse. This is just the place where the other lines are weak. The weakness is most pronounced in the 4686 image, from which the ends of the ellipse are practically gone. The small condensations mentioned as terminating the faint extensions of the major axis are relatively strong in the image at 3727A, though the extensions themselves hardly show. These extensions are, however, visible on the brighter images. The short diameter of 4686A is less than that of the other images, measuring 7'1 as compared with 85 for Hẞ and Hy.

In appearance the nebula strongly resembles N.G.C. 7662, and the spectra of the two objects are almost identical in the part of the photographic region covered by glass instruments. Helium (4471) is relatively stronger in No. 7009, while the Pickering series is more pronounced in No. 7662. Some of the ultra-violet lines are apparently missing in 7009, but allowance must be made for the fact that this object, on account of its southern declination, is observable only at a great zenith distance, and the lines in question are therefore much weakened by absorption. Another point of difference brought out by the slit spectrograms is that in N.G.C. 7009 the line 4363A terminates at the bright ring. In N.G.C. 7662 it extends beyond. Otherwise the spectra are remarkably alike. The similarity in form and spectrum of the three nebulae N.G.C. Nos. 3242, 7009, and 7662 is striking.

*The intensity measurements are on the side of the bright ellipse except for 3727A, in which case the two condensations at the ends are measured. These two condensations are of unequal brightness and measurements of both are given.

On a slit spectrogram which resolves these lines 3967A appears to be about 3 times as strong as He.

29 The faint hazy patch above Hy in fig. 5, pl. XLVIII, is probably a photographic defect.

The detached ultra-violet continuous spectrum is quite strong. This is rather a good spectrum in which to measure its lower limit, because the nebula itself is quite narrow and roughly serves the purpose of a slit. An estimate of the position, which is after all necessarily an approximate one, places the limit at about 3650A.

The spectrum of the nucleus is continuous, and has the usual uniform appearance characteristic of the spectra of nebular nuclei.

The nebula consists of two slightly elongated lobes lying parallel to each other and separated by a comparatively dark lane not quite equal to their width. Faint filaments of nebulosity surround and connect the lobes. One of the spectrograms is reproduced in plate XLVII, figure 2.

The image at 4686, although maintaining the general shape of the nebula, is very much smaller than the others, while 3727A is larger. There is no evidence of a difference in size or separation among the remaining images. The separation of the two lobes, measured center to center, is as follows:

The slight range in the measured separations of intermediate size is probably not significant. It has been suggested that the separating lane in this nebula is due to absorption. If that is true, it is evident that the absorption must be selective and within the nebula itself, and not beyond the luminous part, because the width of the lane varies with the size of the image, that is, it is different for different wave-lengths. It is exceedingly narrow for 4686A, and very wide for 3727. This fact is incompatible with the idea of its being caused by a generally opaque medium.

There is a very faint nucleus situated just to the east of the center of the lane. The light of this is concentrated in two broad bands, one extending from 3795 to 3835A, and the other from 4640-4690A. These estimates are only approximate. The second band is doubtless the Wolf-Rayet band 4650A. The nucleus of the nebula N.G.C. 6905 has an almost identical spectrum.

This is a rather small but exceedingly bright nebula, perhaps the brightest in the sky. It has two condensations or nucleii of unequal intensity. In any one of the slitless spectrograms the different monochromatic images appear to be of unequal size, but this is almost entirely an effect of image intensity. Properly timed images are nearly of the same dimensions, though the one at 4686 may be a little smaller than the others.30 There is not, however, the difference in this respect that one finds in many objects.

It is an excellent nebula for wave-length determinations, not only on account of its brilliancy but also because its spectrum includes a very great percentage of the known nebular lines. This is the reason for the large number of observations which have been made.

The spectrum shown in plate XLII was secured with the slit placed in the line of the two nuclei. Photographs taken with the slitless spectrograph are reproduced in plate XLVI.

The wave-lengths as measured in this and other nebulae are given in table 11. The mean results for all the measurements of a single spectrum are recorded there. It has seemed preferable not to publish the determinations from individual plates in extenso. The observations were made with instruments of many kinds and under a great variety of conditions, and a complete account of the reductions would be very long, and probably serve no useful purpose. There are some details of the measurement of particular lines, however, which seem to call for some explanation, and these are discussed here.

The line 3426.4A was measured twice with the quartz spectrograph (f). This is the only one of the ultra-violet group 3313 to 3445A which I have measured with a slit spectrograph in any nebula, the others being too faint for the present instrumental equipment. It is one of the ultra-violet lines discovered by Palmer, the other being 3346A. Dr. Palmer observed with a very small slitless quartz spectrograph, and was unable to more than indicate the positions of his lines, since they are far to the violet of any other nebular lines known at that time, and he was obliged to extrapolate about 300 units beyond his nearest standard, 3727A. The wave-lengths

30 In a previous paper I stated that the nebula appeared to be perfectly homogeneous. That view was based on observations with a slit spectroscope. It may be necessary to modify the assertion upon further investigation with more efficient apparatus.

given by Palmer differ considerably from my determinations. They are 337 and 345μp. The second of these should not be confused with 3445A of the present system. That is a line which happens to fall near his estimated position, but it was not observed by him.

The other lines in the ultra-violet group just mentioned were measured in this and other nebulae on slitless spectrograms, using the line 3426.4 and other known nebular lines as a basis for inter- and extrapolation. They are probably not more than a unit or two in error.

The line 4686A was measured on four plates, secured with a battery of three prisms. The single determinations are as follows:

4686.00
85.90

85.93

85.92

Mean 4685.94

4363A is recorded on one of these plates. The wave-length as measured is:

4363.39

2

A discussion of the wave-lengths of N, and N, is given in another paper and need not be repeated here.

A word may not be out of place concerning the suggestion by Fowler that a search be made in the spectrum of this nebula for the line at 3203A, observed by him in the laboratory and designated as the first line in the second principal series of hydrogen. The so-called principal series is comparatively strong in this nebula, the leading line, at 4686A, being among the brighter lines of the spectrum, and this fact led Professor Fowler to suggest that the least refrangible line of the second series, which he found associated in laboratory spectra with the principal one, be searched for in the spectrum of this nebula. The line is not recorded on any of my plates, but that fact furnishes little evidence that it is not present in the spectrum. A silvered mirror is transparent for light of that wave-length, and its reflectivity is therefore, unfortunately for this purpose, very low, so that a line of considerable strength at this particular point might well fail to record. The observation should be attempted with a metallic or nickel-plated mirror.

There is a very faint continuous spectrum due to the whole of the brighter part of the nebula. Following it in the direction red to violet, it fades out to vanishing intensity at about 4000A. Another continuous spectrum begins abruptly at about 3700 and continues far into the ultraviolet.

This is a stellar nebula. Monochromatic images of proper exposure are about 21⁄2" in diameter, which is about as small as can be expected for any star. N1 is intense even on the shorter exposure and gives a 4" disc. It is probable that the increased diameter is due to circumstances of observation. As is the case with all stellar nebulae, the intensity measurements are rather uncertain.

A small elliptical nebula with the major axis lying north and south. The definition is not. of the best on either of the spectrograms, but as far as one can judge the intensity of the images. falls off from the center outwards. The strongest discs therefore appear to be the largest. Most of the apparent differences in size can be ascribed to this cause, but 4686 is quite certainly confined closely to the nucleus, and the image at 4471 impresses one as being small. The images given by the lines 4740, 4712, and 4640, which usually conform to the behavior of 4686A, are too faint to permit of an estimate of their sizes. 3727A gives a faint image of the same general size as the others, but more nebulous.

There is a faint continuous spectrum due to the center of the nebula which, as noted above, appears to include 4686A, and possibly other lines, as localized radiations.

This is a beautiful nebula, consisting of an oval ring lying on a more or less circular disc of nebulosity. Distributed around the edge of this disc are a number of bright lobes which give the impression of detached portions of an outer ring system.

A number of spectrograms are reproduced in plates XLII and XLVI. They show at a glance the variety in the forms of the monochromatic images. The most striking image is that at 3727A, rather inadequately depicted. Here the light is almost all in the outer broken ring, and is distributed in a string of stellar points, most of which are resolved on the original photograph but are run together in the reproduction. The outstanding point shown in plate XLVI, figure 5, above and to the right of this image lies on a faint thread of nebulosity following the general form of the outer ring. Little more than a trace of the inner ring is shown.

The absence of exterior nebulosity from 4686A is very noticeable. The ring itself is rounder than the others, due to the fact that its minor axis is slightly greater and its major axis slightly less than those of, say, the hydrogen rings. The measured diameters of the principal rings are