photographic plate is fixed and the measuring microscope is moved.
The chief drawback to type A is that the errors of the screw are liable to change by wear, otherwise the apparatus, as made and used at Potsdam, is, on the whole, a convenient and accurate one. In determining the errors of the screw of the Potsdam form of machine it is necessary to have regard to the fact that the screw is placed at one side of the slide, as in fig. 20.
Fig. 20.
The result is that, if the screw is bent—if, for example, the end of the frame next the screw-head is raised and that next the end p is lowered in the diagram—a twist will be given to the web-frame, and the centre of the web will be moved nearer to the micrometer drum than it should be, whilst the reverse effect will follow when the head has been turned 180º. This would, of course, create a periodic error, which would be determinable for the motion of any particular point (say the middle) of the web, but which would be smaller for a point near the axis of the screw and greater for a point farther from that axis. In the Potsdam form of this apparatus the micrometer is, for convenience, provided with a motion at right angles to the axis of the screw, and it has been found at the Cape Observatory that the periodic errors in this apparatus do vary very sensibly according as the microscope is directed to a point more or less distant from the measuring screw. Since the discovery of this fact all measurements have been made in that fixed position of the microscope with respect to the axis of the screw for which the errors of the screw have been determined.
In the apparatus of type B as made by Zeiss there are two microscopes attached to a base-plate, one of which views the spectrum-plate (or other object) to be measured, while the other views a scale that moves with the slide on which the spectrum plate is mounted. In this way the scale can be viewed by a microscope of much higher magnifying power than can be employed for the photographed spectrum. Indeed, if the scale were subdivided to 110 mm. the power employed might only be limited by the sharpness of the division-lines. But for refined work this would imply the investigation of too many divisions of the scale; it is therefore more usual to divide the scale into single millimetres or half-millimetres and to provide a micrometer which subdivides the millimetre into 1000 or, by estimation, into 10,000 parts. For very accurate work it is desirable that the base-plate, the slide and the scale should be of nickel steel, having the same thermal coefficient of expansion as glass.
The forms of measuring machines of type C, often seen in physical laboratories, should be at once rejected for refined measurements, because it is impossible to construct slides of such perfection that the axis of the microscope will remain absolutely normal to the surface of the plate (assumed to be a plane) throughout the range of measurement. Even if the slide itself is mechanically perfect, the irregularity in the thickness of the lubricating oil between the bearing surfaces of the slide is apt to produce a variable error.
Bakhuyzen (Bulletin de Com. perm. congres. astrog, i. 164) described a measuring-machine by Repsolds, in which the micrometer microscope tilts in the bearings of the chariot on which it moves, so that it can view either a graduated scale or the photographic plate. We have, in fact, in this instrument a combination of types B and C. Even in this apparatus if the slide on which the chariot moves is not perfect (and no slide is perfect), the azimuth of the axis of the microscope is liable to change in the course of movement of the slide, and thus equal spaces on the scale will not be represented by equal spaces on the plate under measurement. The remedy proposed by Repsold for this proved fault is to cause the whole slide to tilt instead of the microscope only; this should prove a complete remedy.
The Travelling Wire Micrometer.—An important modern application of the micrometer, which is not dealt with in the article Transit Circle, is that which is now called “the travelling wire micrometer.”
In the Astronomische Nachrichten, No. 2940, Dr Repsold proposed a method of meridian observing which consists in causing a web to follow the image of a star in transit by motions communicated by the observer’s hands alone, whilst electrical contacts on the drum of the micrometer screw register on the chronograph the instants corresponding to known intervals from the line of collimation. The purpose of his paper was to show that if the axis, by which the observer imparts motion to the slide on which the travelling web is mounted, is provided with two disks at its extremities, so that the observer can use the thumb and finger of both hands in rotating it, there is no difficulty, after a little practice, in keeping the web constantly bisecting the star in transit, and that with a little practice the mean of the absolute errors in following the star becomes nearly zero.
In the Astron. Nach., No. 3377, Repsold gives a detailed description of two forms of eye-ends of transit circles, fitted with means of observing in this manner, to which he gives the name of “the impersonal micrometer.” This method of observation was very successfully employed, under Seeliger at Munich, in an extensive series of meridian observations, and, under the auspices of the Geodetic Institute at Potsdam, in telegraphic longitude operations. Still more recently the method has been largely employed at the Cape of Good Hope and elsewhere.
Under the date March 1901 Dr H. Struve published an account of the application of clockwork as an aid in Repsold’s method; and, later, Dr Cohn published a more elaborate paper on the same subject in the Astron. Nach., 3767. The method consisted in having motion transmitted to the micrometer screw from an axis on which is mounted a disk that presses with friction-contact upon a cone that revolves uniformly by clockwork. The velocity of rotation of the micrometer-screw could therefore be varied for stars of different declination by varying the distance from the apex at which the revolving disk presses upon the revolving cone. In the Königsberg transit instrument used by Struve and Cohn, the clockwork was attached to the eye-end of the instrument—a condition which is obviously undesirable both from the necessarily unsymmetrical position of the clockwork with respect to the optical axis, and from the impossibility of securing the uniform going of the clock in different positions of the instrument. In more recent instruments at the observatories of the Cape of Good Hope and Paris the motion is transmitted from a separately mounted cone and clock by a light rod passing through a perforation in the pivot of the transit instrument and thence through bevel-wheels in the cube of the axis to a second rod leading to the eyepiece. This rod turns a worm-screw which acts on a worm-wheel fitted “spring tight” upon the axis of the micrometer-screw.
It should be mentioned that an essential feature of the travelling wire micrometer is that the eyepiece as well as the wire shall be moved by the micrometer-screw. Thus, if the star’s image is kept in bisection by the wire, both star and wire will appear at rest in the field of view.
The distinction between the old and new method of observation may thus, in one sense, be described as the difference between shooting at a moving object and in shooting at one at rest; In the case of the original Repsold plan without clockwork the description is not quite exact, because both the process of following the object and correcting the aim are simultaneously performed; whilst, if the clockwork runs uniformly and the friction-disk is set to the proper distance from the apex of the cone, the star will appear almost perfectly at rest, and the observer has only to apply delicate corrections by differential gear—a condition which is exactly analogous to that of training a modern gun-sight upon a fixed object. It is impossible in this article to give a detailed description of the apparatus, but the reader is referred to Astron. Nach., 3377, for an illustrated account of the original Repsolds instrument and to the History and Description of the Cape Observatory for a complete description of the most modern form of its application to the Cape transit circle, with and without clockwork.
The Hartmann Spectrocomparator.—For accurate measurement of the displacements of lines of stellar spectra which are produced by the relative motion of star and observer in the line of sight, a very beautiful instrument has been devised by Dr J. Hartmann of Potsdam, and is described by him in the Publicationen des astrophysikalischen Observatoriums zu Potsdam, Bd. 18, Stück 53 (1906). An English translation of this paper is given in the Astrophysical Journal, xxiv. 285–302. The method originally used by Huggins, who first conceived and proved the possibility of measuring stellar velocities in the line of sight, was to measure with a filar micrometer the displacement of some well-known line in a stellar spectrum relative to the corresponding line of a terrestrial spectrum. Vogel of Potsdam introduced the method of photographing stellar and terrestrial spectra on the same plate, and in this way obtained an immense advance in the ease and precision of observation. Vogel and his successors employed one or other form of measuring machine, provided with a microscope having single or close parallel webs which could be successively pointed on the photographed lines of the star spectrum and the lines of the terrestrial spectrum. To derive the stellar velocity in the line of sight relative to the observer it was then necessary to assume that the normal wave-lengths of the stellar and terrestrial spectra are accurately known. But in the
