*

The lantern slides are a little enlarged from the original negatives. On the slides from the reflector-plates 1 degree equals 86 mm., and on the slide of the 1-in. lens I degree equals 4 mm. It is interesting to compare photographs 7 and 8, which were taken at the same time, but with very different optical means— 28-in. and 1-in.

The tail on No. 8 is somewhat over 12 degrees long, while the small field of the reflector gives the head and its neighbourhood only. The bright star near the head of the comet on No. 8 is y Geminorum.

* In the reproduction on Plate 6 one degree on the reflector plate=79 mm., and one degree on the 11-in, lens plate = 3.7 mm.

Astrophysical Observatory,
Heidelberg (Königstuhl),
1907 December 10.

On an improved method of illuminating the field in a Transit Instrument, and its effect on the discordance in reversed positions of the instrument. By Sir W. H. M. Christie, K.C.B., F.R.S., and H. A. H. Christie, B.A.

(Plates 7, 8.)

It is known that with a reversible transit instrument observations made in the two positions of the instrument generally show discordant results for clock error for stars reduced with the same collimation error for the two positions. This discordance has been attributed to what is called lateral flexure, and a correction depending on cos Z.D. has sometimes been applied, but at Greenwich we never felt satisfied that there was any real ground for this assumption. We have suspected sometimes that it was due to looseness in the mounting of the object-glass, and in the Paris-Greenwich longitude determination of 1888 this was found to be the case with both the portable transits. But though this was remedied by an improvement in the mounting of the objectglass, this puzzling discordance still remained in subsequent longitude determinations. Whilst there is in each series of observations a general tendency in one direction, there are considerable variations in the discordance between the mean clock errors for each night in the two positions of the instrument, as will be seen from Tables I., II., III., IV., V., VII., and VIII.

Mr. Hollis, on thinking over the matter in connection with the observations for longitude of Killorglin in 1898 and of Paris in 1902, had come to the conclusion that to obtain the correct clock-error from the mean of results in reversed positions it was necessary to keep the focus of the eyepiece absolutely fixed, and that to secure this it was expedient to keep the diagonal eye-tube which was used for the observations, and was liable to sag, rigidly attached to the tube of the instrument. In the Paris longitude

observations in 1902 special attention was paid to this, the eyepiece being adjusted to focus at the beginning of the evening, and not altered afterwards. The discordance, however, was not altogether got rid of, and on some nights it exceeded oo 2 or oo 3. There was the further difficulty that, as will presently be explained,, the observer's eye is liable to change of focus when the bright star comes into the field.

In 1906 and 1907 Mr. Harold Christie was working with two of the portable transits (B and C) previously used in the Paris longitude determinations of 1888 and 1902, with a view to other longitude determinations, and the large discordances in his results for the two positions again called attention to the question. He noticed that if the eyepiece was adjusted for distinct vision of the wires in an illuminated field a fresh adjustment was required when the star appeared, the eyepiece having to be pulled out further from the object-glass owing to change in the focus of the eye, the wires being, by an unconscious mental process, naturally referred to a finite distance (say 2 or 3 feet), and the star, when it came in, to an infinite distance. This change of focus still took place when the other eye was covered up. He further noticed (on March 1, 1907) that while adjusting the eyepiece for focus on a slow-moving polar star the star seemed to cross the wire from one side to the other as the eyepiece was moved in or out, the apparent movement of the wire being as much as 20". A similar effect had been previously noticed on the meridian mark (a bright point) seen in an illuminated field. It was at first thought that this might be due to the wires not being exactly in the focus of the object-glass, but readjustment of the focus failed to get rid of the effect. On trying the same experiment on Polaris by daylight next day it was found that no apparent movement of the wires could be got by movement of the eyepiece in or out. This indicated that the effect was due to illumination of the field. It was further noted that the movement was less apparent with a strong illumination than with a reduced light. It may here be explained that the illumination of the field is given by a gilt annular reflector in the transit axis, the inclination of which can be varied from a minimum of 45° to a greater angle with the axis so as to reduce the light.

Afterwards observations of stars were made with the eyepiece inside the focus, at the best focus, and outside focus. The results were that the discordance W-E was positive and large (+ 1o5) with the eyepiece far in, and negative (about 06 or 0·7) with the eyepiece far out.

It is to be observed that the mean of the results W and E is sensibly unaffected by the discordance, as will be seen from the last columu.

Transit C.-Annular Illumination Experiments with Eyepiece focus.

By Mr. H. Christie.

N.B.--The "good focus" position was the ordinary one for star observations; the eyepiece was displaced from this position until observation was only just possible.

These experiments seemed to clearly establish the connection. between the discordance and eyepiece focus, and they further suggested that the annular reflector was in fault.

It is to be noted that the wires at night are not seen by their own light, but as dark shadows cutting off the light of the illuminated field, and thus the illuminating surface, which casts the shadow of the wire, has an important influence on its appearance with the annular reflector.

When the wire is out of focus the

shadow of each point of the wire is an annulus, and the superposition of these will give two dark lines parallel to the wire with a comparatively slight haze between, the two sides of the annulus being more effective in forming the shadow images than the top and bottom, as will be seen from the accompanying figure, which shows the portions of the annulus at the sides and at the top and bottom respectively, which are effective in forming the corresponding shadow images of a vertical wire. Thus two fairly clear and separated images are formed one on each side of the true position of the wire. Now if one side of the annulus is more distinctly

On examination of the wires by daylight illumination from the objectglass, it was found that when the eyepiece was out of focus, instead of two images a single image was seen, with bands on each side.

lighted than the other, and the whole illumination of the field is not very bright, the eye will lose one of the two images of the wire, and only see a single image, which will appear fairly clear, and will be displaced by an amount depending on the diameter of the out-of-focus annular ring.

On investigation this proved to be the case. When the annular reflector was at an angle of 45° to the axis (giving a maximum of illumination) it was fairly evenly illuminated, but when the inclination was increased, in order to reduce the light to a convenient amount for star - observing, one side of the gilt reflector became considerably brighter than the other. This was the actual position always used in practice, as the full illumination was too strong for observing stars. Also, it was found that the side of the reflector which gave most light was the one which would, by the above theory, cause the clock error micr. W to be greater than the clock error micr. E, as the star-observations showed to be the case. In order further to test this, Mr. H. Christie made observations on April 5, 6, and 10 with the reflector kept at 45° to the axis (a rheostat being used to reduce the light from the electric lamp), and obtained better results.

It had long been felt that illumination by means of an annular reflector in the transit axis was not satisfactory, as the image formed at the eye-ring from the illumination of the field was an annulus outside the circle formed by a star, and there was risk of the eye failing to receive the whole of the rays on both sides coming from the reflector.

Illumination by rays coming centrally within the cone of rays from the object-glass would be preferable in this respect, but there are objections to the way in which this has usually been carried out. In 1870, Sir G. B. Airy successfully applied central illumination to the Water Telescope which he planned, mounting a piece of looking-glass at an angle of 45° in front of the object-glass to receive the light from a gas flame, without any condensing lens. In this way the light was well diffused over the field, the gas flame as well as the mirror being well out of focus at the plane of the wires.

But in other forms which have been used since, a condensing lens with a small electric lamp in its focus has been introduced, which, in combination with a small reflecting prism or other specular reflector fixed in front of the object-glass, forms a magnified image of the source of light in the plane of the wires, exaggerating its defects, and making it difficult to get tolerable uniformity of illumination of the field and sharp definition of the wires.

In planning the New Altazimuth, a sketch arrangement was proposed in 1896 January for central illumination by light reflected from the axis lamp inside the telescope tube to a matt surface gilt reflector attached centrally to the inside surface of the objectglass; but, owing to practical difficulties, this plan was not carried out, and the ordinary annular reflector in the axis was at first