This work could not have been written without the assistance of the staff of the USNO Library, Morgan Black, Michael O’Connor and Rachel Poe. Daniel Belteki, Royal Museums, Greenwich provided guidance towards its initial scope. The suggestions of our anonymous reviewers were most helpful. We are particularly grateful to the former members of the USNO Transit Circle Divisions, Theodore J. Rafferty, F. Stephen Gauss, and the late Dr. Thomas Corbin, for their technical advice and manuscript review, as well as for the many nights that they spent in solitude under the stars.

1The U. S. Naval Observatory (USNO) in Washington, D. C. is the oldest scientific institution in the U. S. Navy, founded in 1830 as the “Depot of Charts and Instruments”, and tasked with the maintenance of the Navy’s navigational charts, instruments, and chronometers. From a modest shack on Capitol Hill, by 1844 it would rise to become the first American “national observatory”. Following the national observatories at Paris, Greenwich, Berlin and St. Petersburg, the USNO’s work was devoted to the needs of navigators: positional astronomy, timekeeping and celestial mechanics, becoming by the 20th century the pre-eminent force in American fundamental astronomy, demonstrating the Nation’s commitment to both basic and practical science (Dick, 1990). In 1901, Superintendent Charles Henry Davis II described USNO’s department of meridian observations as “the real raison d’être of the observatory itself” (Davis II, 1901, p. 4).

2Though nearly every major observatory had meridian circles at the close of the nineteenth century, by 1933 Britain’s Astronomer Royal Harold Spencer Jones noted:

• 1  Jones ­Harold Spencer (1933), to Robertson, James, 1933 May 3, cited in Hellweg, J. F. (1948), U.  (...)

The Observatories at Greenwich, Washington, and the Cape are the only observatories now which are making regular meridian observations of the Sun, Moon, and planets; the same observatories, with the addition of that at Pulkova are the only ones which are steadily continuing the fundamental observations of star places. These observations provide the basis upon which all astronomical knowledge is built.1

3In 1966, USNO’s Benny L. Klock wrote: “The U. S. Naval Observatory currently operates the only active transit circles in the United States, and consequently must assume most of the responsibility for fundamental positional astronomy in this country.” (Klock, 1967)

4In the twentieth century, as universities and private observatories generally left work in positional astronomy for more exciting topics in astrophysics and cosmology, national observatories continued work in the classical astronomy of position called astrometry (Dick, 1997). This was nominally in support of celestial navigation, but in a broader sense, it was in the pursuit of astronomical knowledge generally. The USNO transit circle divisions had as their objectives the establishment of a celestial reference system, and the observation of absolute positions of celestial bodies with respect to this reference system. Absolute or fundamental positions were those found independently of any previous measures, and referred to the dynamical reference frame, the celestial equator and equinox defined by observation of solar system bodies.

5The instruments of choice for fundamental astronomy in the 19th and 20th centuries were meridian circles. These functioned not as isolated instruments, but rather as ensembles of devices, techniques, and analyses in the hands of experienced observers. A glance at the annual counts of USNO meridian circle observations from 1846-1995 (Fig. 1) reveals the remarkable fact that, at least in terms of numbers of observations, its most productive era was in its final decade, the last of the 20th century.

6Five meridian circles have played a role in the history of the U. S. Naval Observatory (USNO) (Table 1). We have selected two of these, the Six-inch and Seven-inch Transit Circles, to illustrate the following aspects of meridian circle astrometry: the meridian circle as an assemblage of telescopes, calibrations and recording devices to enable precision observations; the tandem operations of globally dispersed instruments; and the contributions of meridian observations to science and society.

Table 1 - USNO Meridian Circles

Meridian Circle	Maker	Years in operation	Total observations	Locations
4.5-inch	Ertel	1845-1864	20,367	Washington, D. C.
4.5-inch	Pistor & Martins	1850-1852	27,304	Santiago, Chile
8.5/9-inch	Pistor & Martins	1865-1944	323,000	Washington, D. C.
6-inch	Warner & Swasey	1899-1999	973,908	Washington, D. C.
7-inch	USNO	1959-1996	640,253	El Leoncito, Argentina, Black Birch, New Zealand

Source: Compiled from annual reports and publications of the U. S. Naval Observatory

7The Six-inch Transit Circle would become the true workhorse of fundamental astrometry at the USNO in the 20th century, operating from 1899-1999 in the west transit building at the Massachusetts Avenue site in Washington, D. C. In that period, 974,000 visual observations were made with this instrument, 300,000 more than the Great Airy Transit Circle at Greenwich (1851-1954), perhaps more visual observations than any other transit circle in history.

8The Seven-inch Transit Circle was the only meridian circle built at the Naval Observatory. It was nearly indistinguishable in design from its 50-year-older cousin, the Six-inch Transit Circle. But while the Six-inch remained a visual instrument, the Seven-inch pioneered electronic imaging automation. These instruments were managed for most of their existence by separate USNO Divisions, yet each participated in the global campaign of the 1950s-1960s known as the International Reference Stars, and in the USNO Pole-to-Pole Program of 1985-1996.

1.  The Six-inch Transit Circle

9To introduce the Six-Inch Warner & Swasey Transit Circle we must travel back in time to December, 1887. More than a decade had passed since the Navy planned to move the Observatory from its miasmic location in Washington’s Foggy Bottom, and at last as funding flowed from the U. S. Congress, plans were underway to build a new Naval Observatory a few miles away to the northwest (Dick, 2003). USNO Superintendent Robert L. Pythian was consulting with Simon Newcomb, of the (then separate) Nautical Almanac Office, on instruments for the new site. Could American makers provide them? Newcomb wrote to Richard S. Floyd of the Lick Observatory, who assured Newcomb that if he were looking for another meridian circle, he would “without hesitation have it made in this country.” Floyd did not hold back, praising the Washington, D. C. firm of Fauth & Co., adding:

• 2  Floyd Richard S. (1887, Dec. 24), letter to Simon Newcomb, USNO Archives, “Administrative Files, A (...)

There are oceans of the finest mechanical skill developed in this country […] which if turned on a fine meridian circle would, I really believe, produce the very best one ever made […] I don’t think we should go abroad when we can do as well with strong chances of better at home—and especially in the case of our National government—which can well afford experiments of this kind to induce the ever abundant and ambitious skill of its citizens towards furnishing its scientists with instruments.2

13In this photo, from left to right, are seen the sidereal clock, a circle lamp, the Saegmuller chronograph, the observing couch, and above the objective, two of the magnitude screens. The vertical collimator towers above to the right. In front of the north doors is the tube of a horizontal collimator above the larger mark lens, with a telescope for reading the thermometer, and the reversing carriage for lifting and rotating the instrument.

14Six-inch Transit Circle observations of stars and the Sun, Moon, and planets commenced on June 5, 1899. These observations were strictly differential, with measurements relative to stars with well-known positions, in this case, the stars of the American Ephemeris, the Berliner Jahrbuch, and a zone of Bonner Durchmusterung stars (Davis II, 1900). They revealed a host of problems with this new instrument. The Six-inch’s marble piers were unstable, its circles were poorly engraved, and its pivots were soft. Its circle microscope objectives were defective, and the illumination of its micrometer wires, so critical to accurate bisections of stars, was uneven (Updegraff, 1903). The Six-inch was taken out of service from 1903-1908 to address these issues, during which time Warner & Swasey installed a new hand-driven “travelling-wire” or “impersonal” micrometer. This modification added a wire that moved by turning the right ascension micrometer screw. As the observer tracked the motion of a transiting star, ten contact points in the micrometer were triggered, signaling time electrically on a drum chronograph (Veeder, 1909). The travelling-wire micrometer was effective in reducing personal equation, (the variation in reaction time among observers to times of transits), particularly for Sun and bright-star observations (Cullen, 1921).

15Warner & Swasey also added a “self-registering apparatus” on the zenith distance micrometer. Designed by Joseph A. Rogers, this device had a series of engraved drum wheels and levers, each of which when lifted would lock one of the wheels at the indicated reading of the declination micrometer. Four or five readings could be taken during a transit, to be read off and recorded afterward (Gilliss, 1865).

17The man who brought the Six-inch Transit Circle from the 19th century into the 20th was Chester Burleigh Watts (1889-1971). Watts worked for nearly 50 years at the Naval Observatory, and directed the Six-inch Transit Circle Division from 1934-1959. In addition to leading the production of the most precise USNO star catalogues to date, Watts was adept at engineering, a skill that he applied to “improving every aspect of the transit circle that appeared to him to be deficient in accuracy or convenience […] The net result of his efforts is reflected in the overall increase in the accuracy of an observation and the ease with which an observer can guide on a star and record all instrumental readings without touching the telescope.” (Scott, 1972)

18Those contributions included the development in the 1930s of a motorized travelling-wire impersonal micrometer, the photographic registration of the micrometer screws, and the addition of film cameras to the six circle micrometers. Glass scales on the micrometer right ascension and zenith distance screws were photographed onto a roll of bromide paper. Watts’ circle cameras were triggered automatically at 4-second intervals by the sidereal clock, rather than by the observer (Hellweg, 1932; Watts, 1950). Watts then constructed the Automatic Measuring Engine (AME, pronounced “Amy” and referred to as “she”), a photoelectric measuring engine that scanned the photos of the zenith distance circle engravings (Fig. 3).

20The Naval Observatory catalogues W₂₅ and W₅₀, “Six-inch Transit Circle Observations: 1925-1941” were the last to use visual reading of the circle microscopes, comprising 100,128 visual observations by twelve different observers. Watts’ microscope cameras were first used for the 1941-1949 program known as the W2₅₀ (Watts & Adams, 1949). Whereas in 1888, observing the Sun had required three observers (one at the eyepiece, two reading the zenith distance micrometers and setting for the next limb) by 1945, with photographic recording, a single observer could make the same observations (Pythian, 1892, p. A.40; Graf, 1952).

21By the middle of the 20th century, the Six-inch programs were being conducted on a fundamental basis, with observations of standard stars, the Sun, Moon and planets. The next program, the W3₅₀, ran from 1949-1956 (Adams, Bestul & Scott, 1964). By 1950, decades of suburban development around Washington meant the loss of program stars fainter than about 9th magnitude, as USNO “gradually lost its enviable position at the northwest edge of the city and became engulfed in a huge metropolitan area extending in all directions.” (Scott, 1954).

22The evolution of the Six-inch entered a new era in 1959 when August Norwood Adams succeeded Watts as Director of the Six-inch Transit Circle Division (Fig. 4). Adams grasped the potential that electronic data processing held for the massive stream of data that Watts’ Division had previously been processing on desk calculators. Assisted by his data-reduction chief Sylvan M. Bestul, in the 1960s, Adams recast the Six-inch’s 1906 Warner & Swasey micrometer into a digital recording instrument with motor drives and synchro resolver transducers on the micrometers, while transforming the reduction process to modern punched-card operations.

23With the addition of astronomer and computer specialist F. Stephen Gauss to the Division in 1963, the following decades would bring about a continuous alignment of USNO transit circle data acquisition and processing with current computing technology (Hemenway, 1966; Dick, 2003, p. 431). Two more USNO fundamental catalogues followed, the W4₅₀ (Adams & Scott, 1968) and the W5₅₀ (Hughes & Scott, 1982). These catalogues continually improved astrometric positions that made up the fundamental reference frame.

24In 1967, a Farrand Inductosyn angular transducer was installed on the horizontal axis of the Six-inch (McDowell, 1968). It was intended to replace the engraved declination circles, but ultimately its inaccuracies and sensitivity to temperature relegated its use to pointing the telescope. To acquire data from the Inductosyn, in 1968 April a new IBM 1800 data acquisition and control system, developed by Gauss, was installed in a room adjacent to the transit circle. With its 8-kilobytes of memory and 1-megabyte of disk storage, this system collected environmental data, readings from the micrometer screw resolvers, circle scanners, and clock timing data. It also computed apparent places prior to the observation, determined the speed of the micrometer motor drive and the altitude to which the telescope was set manually, and performed preliminary reductions of observations. The instant feedback it provided to the observer took the form of messages on a terminal, tones and lights, and a voice synthesizer.

25In 1969 a new micrometer screw resolver-to-digital system was interfaced directly to the IBM 1800 data processor. Now the preliminary results of a night’s observing were available within the hour, and any problems could be detected before the next observing session.

26In 1973 the first of a series of experimental glass zenith circles by Teledyne-Gurley was installed on the Six-inch. The renewed interest in glass circles was prompted by testing of a new photoelectric circle scanning system by Watts Prototype, Inc. They found, as had Saegmuller, 85 years before, that circles of glass showed improved uniformity and definition of the division lines (Rafferty & Klock, 1982).

27Six new photoelectric scanning microscopes were built by the U.S. Naval Observatory’s Instrument Shop and installed on the Six-inch’s glass circle. Years would be spent testing and replacing imperfect Teledyne-Gurley circles. In 1975, USNO contracted with a second vendor, the Heidenhain Corp. of Traunreut, Federal Republic of Germany, for a new type of glass circle having a gold coating evaporated behind the division lines, to provide a uniform and highly reflective background. The Heidenhain circles were so successful that in 1981 a second glass circle was ordered for the Six-inch. The USNO’s photoelectric scanners read the 7,200 glass circle divisions to a precision of one part in 5,000. In the mid-1980s, CCD cameras would replace the photoelectric scanners on both the Six-inch and on the USNO Seven-inch, six on each of the glass circles.

28During the 1970s, the Six-inch objective was replaced by a new temperature-compensating lens from the Farrand Optical Corporation of New York. A newer Inductosyn system was introduced in 1979, and the IBM 1800 computer was replaced by a Hewlett-Packard HP1000-F floating point computer. Software by Gauss that had been developed on the Six-inch Transit Circle was ported to this system (Vohden, 1981).

29The USNO’s first star catalog to be referred to the International Celestial Reference Frame (ICRF) and equinox J2000 was the W1_J00, observed with the Six-inch Transit Circle in Washington, D. C., from 1977-1982. The W1_J00 provided both absolute positions reduced in the instrumental frame as well as positions differentially adjusted to the ICRF by way of the Hipparcos Star Catalog (Table 4, below) (Rafferty, Holdenried & Urban, 2016).

2.  The Seven-inch Transit Circle

30In 1944 the USNO retired its 80-year-old Nine-inch Transit Circle, and Watts began work on his crowning achievement, the design of its replacement, the Seven-inch Transit circle (Fig. 5).

33The Seven-inch made an additional 45,400 observations from 1959-1963. In December 1966, it was shipped to the Yale-Columbia Southern Observatory, El Leoncito, Argentina. There it participated in a world-wide effort to map the Southern Reference Stars (SRS) and stars of the Fourth Fundamental Catalogue (FK4) and its Supplement, under the direction of Francis P. Scott, John L. Schombert, and James A. Hughes. Observing from Argentina began in May 1967; by program completion in August 1973, 155,000 observations were made of 23,000 stars. Because no daytime solar observations could be made with the temperature-sensitive objective lens, night observations were taken of the minor planets Ceres, Pallas, Juno and Vesta for defining the equator and equinox (Branham, 1979). Film cameras recorded the wire and circle readings, with the film sent back to Washington to be processed on the AME. This program produced the Washington – Leoncito Fundamental Catalogue (WL₅₀) (Hughes, Smith & Branham, 1992).

34The WL₅₀ was the final visual program by the Seven-inch, for the decade of the 1980s would see major changes transforming it into USNO’s first fully automated meridian circle. In 1849 USNO’s Lt. Matthew Fountain Maury wrote, “[…] you have only to invent a contrivance, by which the star itself, as it makes the transit, shall touch the key; and astronomers may then, for the first time, boast of one perfect instrument.” (Maury, 1846) That dream became something of a reality in 1981 with tests of the prototype Image Dissector Micrometer (Hughes & al., 1986). The IDM featured vertical wires and a collimating lens mounted onto a ball slide which was tracked by two precision position encoders. A photocathode with filters matching the spectral response of the human eye, operated in photon-counting mode. Digital scans of the stellar and planetary images were processed in real time with centroid analysis. With a temperature compensating objective built by the Farrand Optical Corp. of New York, the Seven-inch Transit Circle with the IDM was able to observe fainter stars over the full range of zenith distance, including all stars to magnitude 4.5 by daytime, all year long. This enabled absolute astrometric observations of the Sun, Mercury and Venus for deriving corrections to the equator and equinox. The greater number and better distribution of the fainter daytime star observations provided a more homogeneous system of clock, refraction, and latitude corrections (Rafferty & Loader, 1993). The 1980s saw also the addition of graduated glass circles fabricated by Heidenhain Corp. of Germany to be scanned by a new photodiode system circle.

35As Hughes reported to the International Astronomical Union in 1984, “A simultaneous, coordinated program of pole-to-pole absolute observations will take place for the first time in astrometric history.” (Hughes, Robinson, Gauss & Stone, 1986) The “Pole-to-Pole Program” was envisioned as a means of improving the reference frame of the Astronomisches Rechen-Institut’s Fifth Fundamental Catalog (FK5) (Fricke, Schwan & Lederle, 1988).

36It would involve the Six-inch in Washington, observing from April 1985 to February 1996, in coordination with the Seven-inch at Black Birch, near Blenheim, New Zealand. The USNO “Pole-to-Pole Program” was carried out under the leadership of Hughes, Clayton A. Smith, and F. Stephen Gauss, directors of the Astrometry Department. The observing program was overseen by Thomas E. Corbin and Theodore J. Rafferty.

37Following 3½ years setting up operations at Black Birch, New Zealand, in 1987 the Seven-inch commenced observations of the southern portion of the USNO “Pole-to-Pole Program” (Loader, 1997). The Black Birch station was directed by M. Dennis Robinson, Ellis R. Holdenried, Theodore J. Rafferty, and Carl S. Cole. Nearly 32,000 observations with the Image Dissector Micrometer were obtained in the first year alone (Anawalt, 1988). The final product of the “Pole-to-Pole Program” was the issuance of the W2_J00 catalog of 44,395 stars on the system of the ICRS (Rafferty, Holdenried & Urban, 2016).

38At the conclusion of the “Pole-to-Pole Program” in 1996, in the wake of the success of the European HIPPARCOS astrometric satellite, the decision was made to end transit circle observing at USNO. In 1999, a final set of visual observations on the Six-inch was made by Rafferty to close out 100 years of service by that venerable instrument, which is now on display at the Naval Observatory in Washington, D. C. The Seven-inch was shipped back to Washington, where it has remained in storage for the past quarter-century.

3.  An Instrumental Ensemble

40The ideal meridian circle would be a completely inflexible telescope, impervious to change of temperature, with a perfectly level horizontal axis which precisely constrains its line of sight to the meridian. It would operate at a precisely known latitude and longitude, on a planet without atmospheric refraction nor polar motion. Its zenith distance circle would be divided with unerring accuracy, its micrometers having translation screws of perfect pitch, its wires exactly aligned with its optical axis, and its observer, human or machine, able to measure transit times with invariant consistency, regardless of any star’s magnitude or declination.

41Of course, “it is impossible to make a geometrically ideal instrument or to set it up exactly. Therefore, an investigation of the instrument for determining its errors is absolutely imperative to the performance of any astrometric work.” (Podobed, 1962) Thus, a meridian circle functioned as an instrumental ensemble of devices aimed at quantifying deviations from the ideal instrument, along with others designed to increase observing efficiency and recording of data. In Table 2 we list some of the devices that have at times made up a meridian circle device ensemble.

Table 2 - USNO Instrument ensemble devices

Device	Purpose	Type of device	Frequency of use	Accuracy
Sidereal clock	Right ascensions	Pendulum/quartz/radio/atomic synthesized	Every star	10-2–10-6 sec.
Chronograph	Transit timing	Drum recorder	Every star	±0”.01
Horizontal collimators	Collimation	Telescopes with micrometers on piers in the transit circle pavillion	Every 2-3 hrs.	±0”.25
Vertical collimator	Measure vertical flexure	Telescope with micrometers, mercury basin artificial horizon	Every 2-3 hrs.	±0”.25
Marks (mires)	Azimuth	Very long focus (100 m) lens and artificial star on mark house pier	Every 1-3 hrs.	±0”.30
Striding level	Level of pivots	Striding (bubble) level	Every 2-3 hrs.	±0”.15
Mercury trough	Nadir	Artificial horizon (mercury basin)	Every 2-3 hrs.	±0”.35
Bonsdorff flexure tubes	Telescope flexure	Steel tubes attached to center cube supporting a flat mirror	Experimental (1930)	-
Reversing carriage	Pivot errors	Mechanical lift	Monthly	<0”.005
Roger’s self-registering apparatus	Marking zenith distance observations	Engraved drums with locking levers on the declination micrometer	Every bisection	-
Travelling-wire micrometer	Reduce personal equation	Electro-mechanical right ascension slide, manual or motorized	Every tic	--
Dove prism	Reduce personal equation	Optical East-West field reversing prism	Each object	-
Personal Equation Machine	Measure differences of P. E.	Motorized artificial star with pushbutton keypad	Monthly	-
Circle cameras	Photo registration	Film or digital imaging camera	Every star	-
Micrometer cameras	Photo registration	Film or digital imaging camera	Every tic, bisection	-
Measuring engine	Film measuring	Manual or automated	Each frame	±0”.06
Screens	Reducing brightness of stars	Wire gauze or louvered slats in the optical path	Used on brightest stars	-
Sympiesiometer	refraction	Combination of thermometers and barometer for density calculation	Varied	-
Thermometers	refraction	Mercury standard	Varied	-
Barometers	refraction	Mercury standard	Varied	-
Dew cells	Refraction	Hygrometer used to measure dew point	Varied	-

Source: (Corbin, 1985; Watts, 1950)

4.  An Assemblage of Reductions

43The precision of meridian circle astrometry resulted from a rigorous system of observation developed by generations of observers, who understood that a meridian circle was by no means a static device. Rather, it was ever an animate object, constantly in motion, ever changing its dimensions, its spatial orientation and its optical response. Instrumental changes on small scales had to be managed. As Dr. Thomas Corbin explained, “very small mechanical shifts in the mounting of the instrument or its optics can have serious effects on the observations if not corrected. A shift of only 10 microns in one of the mounting cages of the Washington Six-inch Transit Circle can produce a shift of 2.3 arcsec in the azimuth.” (Corbin, 1985) We now look at some of the corrections that were made to meridian circle observations.

Clock Star Corrections

44It was usual for USNO to observe in each tour from a list of a few hundred “clock stars”, a set of well-observed fundamental reference stars, evenly spaced around the sky within 30° of the equator. The positions of these stars were so well-determined that any observed differences in their right ascensions were attributed to the observing sidereal clock. They provided a framework for solving for systematic errors in the right ascensions of the program stars. A brief overview of the determination of such systematic corrections to star positions can be found in (Newcomb, 1906, p. 354-359). The clock star corrections for the Six-inch Transit Circle, 1963-1971, varied by right ascension over a range of ±0.07 arcseconds.

Personal Equation

45By 1823 Bessel came to realize that personal equation, the varied reaction time among meridian circle observers, meant that the observer himself was an instrument in need of calibration: “…one can no longer doubt, that an involuntary constant difference can take place between two observers, which clearly exceeds the limits of the accidental uncertainty.” (Bessel, 1823) The “impersonal” or “travelling wire” micrometer, along with use of a reversing prism, did much to reduce personal equation. At Greenwich in 1855, George Airy found that personal equation (using the chronograph) among his dozen observers varied over a range of 0^s.19 of time (Airy, 1855). During the USNO W1_J00 Program, 1977-1982, this range was only 0^s.03. (Rafferty, Holdenried & Urban, 2016).

Pivot irregularities

46The horizontal axis of a meridian circle consists of hardened steel cylinders (pivots) which rest on two pieces of softer material, usually brass (wyes). Even in the late 20th century it was “not yet possible to make the pivots so perfectly cylindrical that the axis remains parallel to itself when the instrument is rotated” (Eichhorn, 1974). Wear on the pivots of meridian circles was an ongoing concern. USNO examined the figure of the pivots at the start and end of each observing program.

47To compensate for pivot error, once a month the telescope was lifted and reversed on its pivots, and observations made of the same stars in both positions. Pivot errors for the Six-inch were on the order of up to 0.2 arcseconds, depending on the rotation of the circle.

Day Corrections

49Daytime stellar observations at USNO were made up to 4.5 hours before and after noontime. Hughes and Scott in 1982 solved for the systematic differences in the positions of stars taken both day and night, finding position differences of nearly 0.1 arcsecond. The differences had been attributed to a variation in room refraction resulting from a rise in temperature during the daylight hours (Watts, 1949).

Variation of Latitude

50Meridian circle observations of zenith distances must be corrected for the variation of astronomical latitude caused by motion of the Earth’s crust, a phenomenon discovered by Karl Friedrich Küstner (1856-1936) in 1884 (Küstner, 1890). This variation has two prominent periods, an annual period attributed to meteorology and a period of 427 days, the so-called “Chandler wobble” defined by oceanic patterns and the elasticity of the Earth’s crust (Rafferty, 1982). From 1915 to 1997 the observations for variation of latitude were provided by the USNO Photographic Zenith Tubes (PZTs) operated by the Time Service Department. Variation of latitude is a large effect, up to 0.5 arcseconds.

Refraction

51In meridian circle astrometry atmospheric refraction played a large role, its value is a full arc minute at altitude 45°. The refraction correction was computed using a formulation by Danjon as a function of local atmospheric pressure, temperature and dew point (Danjon, 1952).

Diurnal Aberration

52The Earth’s rotation is continually carrying the meridian circle toward the east with a speed of nearly 500 meters/sec times the cosine of its latitude. The effect at the latitude of Washington is a displacement of observed objects eastward by up to 0.24 arcseconds.

Collimation

53In a meridian circle, a straight line drawn from the middle wire to a star, when that star is on the wire, was called the “line of collimation”. If that line was not perpendicular to the axis of the telescope, the telescope was said to be “out of collimation”. The “collimation error” was the time that a star on the equator would take to transit from the line of collimation to the true perpendicular to the axis. The collimation error was measured using a pair of horizontal telescopes with micrometers (“collimators”) mounted on masonry piers within the transit building to the north and south of the meridian circle. The meridian circle was pointed to each collimator in turn, from which the observer would make sightings on the meridian circle wires (Hughes, Smith & Branham, 1992, p. 167).

Flexure

54Flexure is the result of unequal yielding of the two ends of the telescope tube to gravitational force. Horizontal collimators (and a vertical collimator with mercury basin) were used to measure flexure. Observing stars in the artificial horizon, with the telescope inverted 180° was another method used for a time.

55During the W₅₀ program in 1930, an apparatus for measuring flexure was installed along the telescope tube of the Six-inch Transit Circle. Based on a device invented by Ilmari Bonsdorff at the Pulkovo Observatory in 1914 (Bonsdorff, 1915), it consisted of four steel tubes attached to the cube and supporting a plane mirror in front of the objective. Reflected images of the wires at various zenith distances mapped the flexure of the tube and optical system (Hellweg, 1930, p. 13). In the 1970s, Klock affixed strain gauges to the tube of the Six-inch, but the readings showed wide variation and were inconclusive. For the USNO Seven-inch Transit Circle, 1967-1973, the adopted flexure correction, for zenith distance z, was Δz = 0.335 sin z arcseconds (Hughes, Smith & Branham, 1992).

Magnitude Equation

56In 1898 Newcomb wrote “Among all the systematic errors affecting the observed Right Ascensions of stars, the most serious at the present time is that arising from the now well-known fact that the personal equation of an observer is generally different, according as a faint or a bright star is observed.” (Newcomb, 1898) Bright stars, having a larger disk, in transit will appear to be closer to the vertical wires than does the smaller disk of a fainter star. As described previously, screens became important additions to the Six-inch and Seven-inch Transit Circles in reducing the apparent brightness of the bright stars.

57The effort to determine the above corrections was prodigious. In Table 3 we compare two USNO observing programs, the W₂₀, (Morgan, 1933), observed with the USNO Nine-Inch Transit Circle, and the W2_J00, (Holdenried & Rafferty, 2016), observed with the USNO Six-inch and Seven-inch Transit Circles. Note the vast numbers of clock stars, solar system, azimuth, day, and refraction stars that were observed to refine the positions of the program stars! The improvements of 75 years are seen in the final three rows, with the standard error of right ascension observations much decreased, and the efficiency (ratio of program to calibration observations) and number of observations per day increased.

Table 3 - Observations of program stars and calibration and reference objects

Meridian Circle	9-inch Washington	6-inch Washington	7-inch New Zealand
Program (Years)	W20 (1913-1926)	W2J00 (1985-1995)	W2J00 (1987-1996)
Program stars	42,980	242,518	279,715
Clock stars	16,800	37,520	29,728
Solar system bodies	2,626	10,292	8,312
Azimuth stars	3,000	14,817	98
Day stars	3,600	9,241	36,343
Refraction stars	1,000	5,748	8,086
Totals	73,330	319,316	386,291
Std. error single observation R.A./dec. (arcseconds)	0.49/0.22	0.19/0.20	0.20/0.23
Ratio of program to calibration observations	1.43	3.22	2.62
Observations per day	15	87	118

Source: USNO Annual Report, 1926; Rafferty, Holdenried & Urban, 2016

5.  An Assemblage of Observations

58The transformation of the results of a USNO meridian circle program into an observation catalogue, reduced to a standard astronomical frame of reference, involved man-years of labor, often by the same observers who worked nights.  In the 1850s Maury repeatedly bemoaned the need for more Observatory assistants, “not only eyes to observe, but power to treat observations after the instruments have performed their office.” (Maury, 1860) The results of the USNO transit circles observations were published in the form of USNO fundamental catalogues of positions, reduced to standard epochs (Table 4). A star catalogue is “a list of spherical coordinates of stars […] obtained by the evaluation and reduction of observations made expressly for the accurate determination of stellar positions. Its principal function is in defining an inertial system, i.e., a coordinate system which allows the comparison of Newtonian mechanics with observation.” (Eichhorn, 1974) USNO observations contributed to the formation of the most important compiled fundamental catalogues, the Fundamental Catalogues of the Astronomisches Rechen-Institut, most recently, the FK6 (Wielen, et al., 1999). 

Table 4 - Star catalogues of the Six-inch and Seven-inch Transit Circles

Catalogue	Contributed to	Years observed	Transit Circle	Number stars	Epoch	Pub. date
W10	FK3-FK5	1909-1918	Six-inch	2,714	1910	1927
WZ25		1928-1930	Six-inch	3,520	1925	1938
W25	FK3	1936-1941	Six-inch	2,383	1925	1949
W50	FK3-FK5	1936-1941	Six-inch	1,536	1950	1949
W250	FK5-FK6	1941-1949	Six-inch	5,216	1950	1952
W350	FK5-FK6	1949-1956	Six-inch	5,965	1950	1964
W450	FK5-FK6	1956-1962	Six-inch	2,554	1950	1968
W550	FK5-FK6	1963-1971	Six-inch	14,916	1950	1982
WL50	FK5-FK6	1967-1973	Seven-inch	23,001	1950	1992
W1J00		1977-1982	Six-inch	7,476	2000	1998
W2J00		1985-1996	Six-inch	26,204	2000	1998
W2J00		1985-1996	Seven-inch	28,186	2000	1998

Source: compiled from Annual Reports and Publications of the U. S. Naval Observatory

6.  An Assemblage of Meridian Circles

59 It is important to emphasize that meridian circles worldwide have also been used as assemblages, in the sense of coordinated observations, at least since the mid-19th century. This was also true at the USNO. An early case in point is the quest to determine the value of the solar parallax, one of the most important astronomical constants, which defined the Astronomical Unit, the Earth-Sun mean distance. One way to measure it made use of meridian circles at about the same longitude and separated greatly in latitude, giving a north-south baseline to compare declinations. The Sun itself need not be observed; nearer solar system bodies would suffice.

60The particularly favorable oppositions of Mars in December 1849 and January 1852 offered an opportunity to refine its value, and in 1848 the Secretary of the Navy was directed by special act of the US Congress to establish an astronomical expedition to Chile (Gilliss, 1856; Hermosilla, 2017; Dick, 2020). Lt. James M. Gilliss organized a three-year long observing campaign enlisting the USNO, Harvard, Greenwich, and other northern observatories to make precise measurements of the positions of Venus, Mars, and their surrounding stars.

62At the opposition of Mars in October 1862, Gilliss, then heading the USNO, repeated the solar parallax experiment. This time, observations with the meridian circles at Pulkovo, Helsingfors, Leiden, Greenwich, Albany and Washington were combined with southern observations from Williamstown, Cape of Good Hope, and Santiago, resulting in a measure of the Earth-Sun distance only 0.6% too small (Davis, 1867). Subsequent international solar parallax campaigns in 1900-1901 and 1930 were organized around the close approaches of minor planet 433 Eros, in which the USNO 26-inch Equatorial and the 9-Inch and Six-inch Transit Circles participated (Davis II, 1901; Hellweg, 1930). Internationally, fifty-eight observatories participated in the 1930 Eros campaign.

63International assemblages of meridian circles were often enlisted in stellar reference catalogue efforts, with northern and southern hemisphere meridian circles dividing up the sky by declination zones, each taking regions appropriate to their latitude. In 1887, as part of a program organized by Dr. Arthur Auwers, President of the Astronomische Gesellschaft, the USNO agreed to make observations of Bonner Durchmusterung stars in a four-degree zone of south declinations (Skinner, 1908).

64In 1953, the International Astronomical Union (IAU) called for “observations of a sufficient number of reference stars […] by as many meridian circles as possible” to provide positions and proper motions in creating the AGK3R catalogue of reference stars, coordinated by a committee appointed by the IAU Commission 8 (Positional Astronomy) at its 1955 general assembly in Dublin. Meridian circles at twelve northern observatories participated: Babelsberg, Bergedorf, Bordeaux, Heidelberg, Herstmonceaux, Lund, Nicolaiev, Ottowa, Paris, Pulkovo, Strasbourg, and Washington. Together, this international consortium made 85,000 observations of the AGK3R from 1956 to 1963 (Corbin, 1977). Each observatory sent its observations to the USNO, which assumed the role of reducing and combining the observations on the system of the Astronomisches Rechen-Institut’s Fourth Fundamental Catalog (FK4) (Eichhorn, 1974, p. 217).

65The IAU in 1959 formulated the AGK3R extension, the Southern Reference Stars (SRS), to address an urgent need for fundamental star positions in the southern hemisphere. The USNO took a leading role in both observing and reductions, using the Six-inch Transit Circle in Washington and sending its Seven-inch Transit Circle to Argentina in 1966, to observe zones unreachable from the northern hemisphere (Dick, 2003, p. 435-437). The combined catalogues, AGK3R and SRS, came to be known as the International Reference Stars (IRS), a reference frame so important that it made up almost 40% of the HIPPARCOS astrometry program (Corbin, 1993). In 1991 USNO’s Thomas E. Corbin and Sean Urban published the “Astrographic Catalog Reference Stars (ACRS)”, an international compilation catalog of 325,416 star positions and proper motions computed from 1.6 million observations. The authors described the ACRS as “the most accurate high density reference star catalog ever available…derived from 165 photographic and meridian circle catalogs…” (Corbin, Urban & Warren, 1991).

66In the final example of observing by coordinated USNO instruments, as described previously, the USNO Six-inch and Seven-inch transit circles worked in tandem from Washington, D. C. and Black Birch, New Zealand to map the heavens in the Pole-to-Pole Program, 1985-1996, to create the W2_J00 Catalogue. As the catalogue authors noted, “The Pole-to-Pole project is the latest and largest of a long series of transit circle catalogs produced by the U. S. Naval Observatory. It is also, because of advancing technologies, certainly the last.” (Holdenried & Rafferty, 2016).

67These examples show that meridian circles could multiply their effectiveness, efficiencies, and accuracies by combining observations worldwide, an activity that not only affected the science but also involved a great deal of international coordination and social interaction.

7.  An assemblage of professionals

68A meridian circle is a complex instrument requiring a carefully monitored environment. Many professionals of varied skills made up its support team. At USNO these included observers, instrument makers, computers, record keepers, electricians, carpenters, painters, finance and procurement agents, and program managers. The Transit Circle divisions of the USNO have historically been particularly advantaged in the support of such an assemblage, as part of a relatively large and continuously funded government agency. In each of the voluminous Publications of the U. S. Naval Observatory associated staff are memorialized. For example, in their discussion of the Six-Inch Transit Circle observations for the W₁₀ Catalogue, 1909-1918, Hammond and Watts acknowledged observers Milton Updegraff, Capt. F. B. Littell, Matt Frederickson, Edgar D. Tillyer, Roscoe M. Packard, David Rines, Charles C. Wylie, and Ralph Aston (Hammond and Watts, 1927).

70The adhesive that bonded together this team of professionals was the U. S. Navy, represented by the Observatory Superintendent, whose rank was Captain or above. The Navy’s interest in, and dependence upon, the work of the USNO was felt by every member of the team. The authors can attest to the pride that they themselves felt in contributing to the mission of the USNO.

8.  Conclusion

71As the 20th century neared its conclusion, the USNO quietly ended meridian circle operations, as did most other observatories. Very Long Baseline Radio Interferometry, USNO’s new generations of CCD astrographic instruments, and especially, the astrometric satellites of the European Space Agency had by then assumed the mantle of delivering astrometric accuracies down to the 100 micro-arcsecond level. Yet, the remarkable precision of astrometry from space in no way diminishes the countless hours spent over the centuries by the pioneer meridian circle observers. As Daniel Michalik pointed out, “astrometric measurements obtained in the past, even of moderate accuracy by modern standards, have lasting value as they represent a state of the Universe that is never repeated” (Michalik, Lindegren, Hobbs & Lammert, 2014). Maury recognized this in 1858, when he proposed to the Secretary of the Navy “[…] an inventory of the starry heavens […] this work being carried out, we should enable posterity in the remotest future to reproduce the skies as they appeared to us, and to say of a certainty how far the great law of change to us extends to the skies and among the stars” (Maury, 1858). Thus, it is not surprising to find Gilliss’s Catalogue of 1,248 stars of epoch 1840 included as a contribution to the 20th century Fundamental Catalogues of the Astronomisches Rechen-Institut. Roland Wielen explained that even the satellite astrometry of HIPPARCOS is improved by combination with meridian circle catalogues.

The measuring accuracy of these ground-based positions is such that they allow, especially in combination with the HIPPARCOS positions, to derive proper motions which have significantly smaller measuring errors than the HIPPARCOS proper motions […] Due to undetected astrometric binaries, the HIPPARCOS proper motions can deviate significantly from the more typical proper motions of the center-of-mass. HIPPARCOS proper motions are derived essentially as ‘instantaneous’ proper motions from data collected within about 3 years only. In contrast, the ground-based observations cover often more than two centuries (Wielen & al., 1999).

72In the use of ground-based data for the improvement of HIPPARCOS proper motions, George Gontcharov of the Pulkovo Observatory found the long series of observations made with the Washington Six-inch Transit Circle “of exceptional importance” for their low mean errors of position (Gontcharov, Andronova, Titov & Kornilov, 2001).

73Stellar proper motions include the combined effect of random motion and a systematic component caused by Galactic rotation. Therefore, a stellar reference frame like the FK5 is not truly inertial. In the 1990’s, the International Astronomical Union (IAU) set about building a new International Celestial Reference Frame (ICRF), based on positions of distant extragalactic radio-optical sources known as quasars. The current ICRF3 (2020) includes positions of 6,000 quasars, replacing the dynamical reference frames of the meridian circle era (Charlot & al., 2020).

74USNO meridian circle observations have played an essential role in improving planetary ephemerides. The series of Developmental Ephemerides (DE) of NASA’s Jet Propulsion Laboratory provide precise predictions of the positions of the Moon and planets, generated by least-squares fitting numerically integrated orbits of the Moon and planets to observations (Standish, 1982). They are used by JPL for navigation of spacecraft and have been the basis of planetary data in the Astronomical Almanac since 1984. From ephemerides DE96-DE118 (1975-1981), the only transit circle data used by JPL were the USNO Six-inch and Nine-inch observations made after 1911, the date when travelling-wire micrometers came into use (Standish, 1990). The recent JPL ephemerides, DE421 (2009) and DE430 (2014), included USNO, Bordeaux, and La Palma transit observations of Jupiter, Saturn, Uranus, and Neptune, but noted that “since the measurement noise is relatively large for the transit measurements, they do not contribute significantly to the ephemeris solution. The transit measurements are included mainly for historical comparison.” (Folkner, Williams, Boggs, Parks & Kuchynka, 2014). The latest JPL ephemeris, DE440 (2021), includes USNO transit observations only of Uranus and Neptune from 1926-1993, with the same disclaiming quote as above (Park, Folkner, Williams, & Boggs, 2021).

75The astrometric contributions of the USNO and other meridian circle observers have laid a foundation into many areas of astrophysical research, particularly in improving stellar proper motions (Kopeikin & Makarov, 2020). These include:

• Stellar kinematics and dynamics of the Milky Way, globular clusters and dwarf galaxies of the Local Group. Refinement of the Oort Constants calibrate the actual Galactic rotation speeds, velocity of the Sun with respect to the Galactic Center, mass density distribution of the Galactic disk.

• Secular aberration and secular parallax. Just as the Earth’s rotation produces through diurnal aberration a universal eastward shift of celestial objects, so, on a larger scale, secular aberration results from the motion of the Local Standard of Rest (LSR) about the Galactic center. Secular parallax results from the motion of the Milky Way Galaxy with respect to the isotropic Cosmic Microwave Background. Each induces a component of the proper motion of quasars and galaxies.

• Tests of general relativity, gravitational waves, and the density of dark matter. For example, plane gravitational waves will produce characteristic periodic fluctuations in the apparent positions of distant astronomical sources (i. e., proper motions of quasars and galaxies).

76Finally, in providing an ever-improving celestial framework for the U. S. Navy’s Nautical and Air Almanacs for celestial navigation, the meridian circles truly fulfilled the motto of the USNO as taken from the Astronomica of Marcus Manilius: Pervenit in astra, et pontum caelo conjunxit. The stars are scaled, and sky with ocean joined.