A 1.724Mb PDF of this article is available by clicking HERE
In the modern era of getting fast measurements with GPS, probably few surveyors today realize the painstaking efforts made by the United States Coast & Geodetic Survey (C&GS) to measure their triangulation baselines in the late 1800s. The distances measured by C&GS during this era were measured in increments of microns, or one-thousandths of a millimeter, which could only be ascertained by the use of powerful microscopes.
Obtaining a precisely measured baseline was essential if the triangulation network could be continued with high certainty. In simple terms, a baseline is a measured side of a triangle that is needed to mathematically determine the lengths of the other two sides when used together with the measured angles. Once the other sides are determined, the next set of triangles in the triangulation network depends upon the accuracy of the previously calculated line lengths. This process of only measuring angles was carried on for many miles, so the precision of the baseline was extremely important.
C&GS became a world leader in developing new base apparatuses for measuring distances. A base apparatus is a device used to precisely measure the ground distances of the baselines in the triangulation networks. One of the main obstacles to overcome was the effect of temperature change upon a calibrated measuring bar that would naturally cause contraction or expansion. It was felt that if a measuring bar could be kept at the same constant temperature throughout the entire process of measuring, the results would therefore be substantially more accurate.
Robert Simpson Woodward, a C&GS Assistant and former USGS employee, developed a theory for a base apparatus that counteracted this effect of temperature change. The main component of his measuring device would be a small rectangular-shaped steel bar measuring 8mm wide, 32mm tall, and 5 meters long. This highly-calibrated bar was to be carried along in a trough of melting ice while microscopes would be used to mark its successive positions.
The scientific study to create the Ice Bar apparatus began in the autumn of 1890 and continued through the early winter months of 1891 with the plans and specifications being finished and approved in the spring of 1891. The most critical aspect of the entire device, the 5-meter measuring bar, was manufactured at the steel works in Lancaster, Pennsylvania to exact specifications under the supervision of C&GS personnel. At the upper ends of the bar the first 2cm were cut away to receive graduated plugs of platinum-iridium, inserted so that the upper surfaces resided in the neutral surface of the bar. Three lines were ruled on each plug, two in the direction of and one transverse to the length of the bar. To secure the alignment of the bar, eleven German-silver plugs 5mm in diameter were inserted at intervals of 495mm along the bar so that they projected 1mm above the top surface of the bar. The remainder of the apparatus was constructed in Washington, D.C. by machinists from E.N. Gray & Company and D. Ballauf. The instrument division of C&GS also provided other essential parts. When completed, Woodward’s Ice Bar officially became known as Bar No. 17, or B17.
A steel trough formed in the shape of a "Y" provided the support and alignment for the bar and also contained the melting ice that surrounded it. The trough was riveted together with uplifted sides at the bottom, forming an angle of 60º. At intervals of every half-meter, saddles were secured to the sides of the trough to support the bar. Each saddle contained two lateral and one vertical adjusting screw which provided the alignment of the bar both vertically and horizontally while resting inside the trough. These screws, except at the ends, were offset alternately high and low to prevent pinching and to provide a means of rotating the bar if necessary. The vertical adjusting screws came up from the bottom of the trough through slots which also served as drainage passageways to expel water from the melting ice. These slots were stuffed with cotton to prevent air from circulating up into the trough, but allowed the water to percolate through the cotton to escape. The ends of the trough were enclosed by wooden "V"-shaped blocks, and the entire trough was covered with a heavy white cotton felt which protected it and the ice from direct radiation. The entire trough without the measuring bar and ice weighed 82kg (180.78 lbs).
Once the bar was in place, the inside of the trough was completely filled with 40kg (88.185 lbs) of pulverized ice that surrounded the bar and was mounded above the top. The "Y"-shaped design of the trough allowed the ice to always stay in close proximity to the bar from its weight on the sloping sides. A jack plane was used to shave ice to the consistency of snow, which was then packed around the ends of the bar. This finely shaved ice could be easily moved and permitted a viewing hole to the graduated plugs to take the measurements.
The trough was mounted onto two cars with saddles. Each car had three wheels that moved along on rails spaced 30cm (11.81 inches) apart. Three sections of track each five meters long were used during operation and were staggered so that the ends of the ice bar would always be located in the middle of a section. As the apparatus moved along the track, the trailing sections would be moved forward, thus eliminating the need for the bar and trough to ever be lifted or carried forward. The entire assembly was further shielded by a portable shed that protected it from direct rays of the sun.
Repsold micrometer microscopes spaced 5 meters (16.40 feet) apart were mounted upon firmly set wooden posts which provided the means of examining and recording the finely-made graduation lines on the ends of the bar. Each microscope had 2cm of lateral movement in order to make the measurement. One revolution of the adjusting screw corresponded to 0.1mm of movement. Four microscopes were used with two always being across from the bar for observation and two others being carried forward and mounted upon the posts to be ready for the next measurement. The microscopes were shielded by individual umbrellas.
Once the bar was surrounded by ice, it reached its stable position within ten minutes, and 90% of the contraction occurred within the first minute. To completely rule out the possibility of further contraction, the bar was allowed to be surrounded by ice for up to 60 minutes before taking measurements.
The precautions were so extreme that C&GS scientists even considered the effect that temperature might have on the small exposed areas on the bar where the measurements were being taken with the microscopes. It was concluded that the temperature difference at these locations was only a few thousandths of a degree different and therefore not considered a measurable concern.
The most important adjustment prior to making the initial measurement was the alignment of the bar in the "Y"-trough. The bar had to be precisely centered so that it was parallel with the trough, perfectly aligned with its sides vertical, and the ends completely level with each other. The alignment plugs in the top of the bar that projected 1mm were each examined and then the lateral adjustment screws were used to bring each one into one successive straight line between the end points. Originally a small sharp-pointed plumb bob was suspended from a fine brass wire that was stretched between the ends of the trough to align each plug. This method was found to have brought the plugs to within 0.1mm when the trough was fully loaded with ice. Subsequent experiments showed that by stretching a fine thread closely over the top of the plugs while the trough was only four-fifths full o
f ice yielded equal results. One fear that C&GS discussed before using the apparatus was the effect of the daily temperature range and its possible measurable effect upon the length of the bar due to the change in curvature of the trough. Tests were conducted with a strident level that could be placed upon the trough and found that the change in the curvature of the trough resulted in an insignificant change of only a few tenths of a micron upon the length of the bar.
Moving the Bar
The actual measurement of a line with the Ice Bar required the work of eight men three microscope observers, one recorder, one to move and adjust the microscopes, and to move the car tracks, the microscope shades, and the ice and ice-crusher. A clamp secured the trough into place onto the track so it could not move while measurements were being taken. Those observing and making the measurements through the microscopes called out their readings to the recorder. Each observer then purposely moved his microscope at least a half turn out of position, and they exchanged positions and brought them back into position. Each observer again called out his readings in microns with special attention given if a reading was significantly different than previously stated. This process eliminated personal blunders through the check of measurements. A total of four readings were taken at each location with each of the four readings expected to be within a few microns. While the bar was still under the microscopes, a third observer measured the distance at the front end and realigned the bar to an even graduated line. This process ensured the graduated lines on the bar at the next reading would be within the observing tolerance and had not progressively moved further away. The bar, trailing microscope, and track were then moved forward, leaving the former front end microscope in place for the rear measurement on the next section. While taking measurements, the observers stood upon platforms away from the microscope posts and were extremely careful not to come in contact with the posts, trough, or the track it rested upon.
At time intervals of twenty to forty minutes the apparatus was purposely moved away from the microscopes and fresh ice was supplied. This involved stirring the entire contents and replacing that which had melted. Then the trough was moved back into place and the measurements resumed.
The rate in which a designated portion of a baseline was measured with the Ice Bar varied upon the circumstances at each location, but a goal was to measure at least 100 meters per hour, and measuring a kilometer per day was not found to be excessive. The speed in which these precise measurements were made was the direct result of men having been fully trained to not only perform their duties well, but to work in complete unison with each other.
Woodward’s Ice Bar had its debut on the Holton Baseline at Holton, Ripley County, Indiana. The location of this baseline was determined by reconnaissance in October and November of 1890 during the time when Woodward was finalizing his theory. Between the towns of Holton and New Marion, the baseline had a total length of 5,500 meters (3.418 miles) in a nearly north and south direction. By June of 1891 C&GS crews were busy making preparations along the baseline, such as setting the surface and subsurface marks, clearing trees, fencing the camp ground, pitching twenty-two tents, setting the beechwood posts for the microscopes, and determining the height of the north base point by means of running levels from the nearest bench mark of the transcontinental line of the geodetic leveling.
As with all baselines, several other types of measuring devices including long bronze and metallic tapes were also used to each provide a check upon the other. The Ice Bar was delegated to measure only one kilometer of the baseline where special stones with copper bolts were placed to mark the end points of the Ice Bar section at 3,900 and 4,900 meters. Stones were also placed at various other intervals along the baseline where measurements began and ended. In charge of the measurements at Holton were Otto H. Tittman and Robert S. Woodward. The first measurement upon the kilometer designated for the Ice Bar was finished on September 10, 1891, by Woodward. Measurements were made on six other days during the following three weeks with the working hours usually being at night to avoid extreme temperature changes. The initial use of the device as a precision base apparatus was deemed a complete success.
In order to determine the true length of Ice Bar (B17) it had to be compared to a device of a standard length in terms of the International Meter known to that era. Prototype Meter No. 21 (M21) was used as a standard. The extreme probable error for M21 was determined to be 0.2 micron in length. Its comparison to B17 was made at the temperature of melting ice. The results determined that Woodward’s Ice Bar when used at the temperature of melting ice was 5 meters, 16.2 microns or expressed as 5.0000162 meters.
During the C&GS centennial celebration on April 6, 1916, the Ice Bar was lauded as the only bar in the United States, and probably the entire world that gave entire satisfaction so far as complete accuracy is concerned. The Ice Bar was replaced by the use of steel tapes for measuring baselines, but it was retained for standardizing other apparatus, and for that purpose it remained unexcelled. The whereabouts of the Ice Bar today is a mystery, according to Chief Geodetic Surveyor David Doyle with the National Geodetic Survey, who has diligently searched for this important historical measuring device.
Jerry Penry is a Nebraska licensed land surveyor. He is a frequent contributor to The American Surveyor.
Woodward authored numerous geodesy-related publications. The source for much of the information from this article is The U.S. Coast and Geodetic Survey Report for 1892, Part 11, No. 8. Below is a time line that highlights Woodward’s life and career.
• 1849 Born July 21 in Rochester, Michigan
• 1868-1872 Engineering student at University of Michigan; graduated C.E.
• 1872-1882 Assisted with triangulation network of Great Lakes region for U.S. Lake Survey
• 1882-1884 Assistant astronomer U.S. Geological Survey
• 1884-1890 Astronomer/geographer U.S. Geological Survey
• 1890-1893 Worked for C&GS; developed triangulation methods of surveying (such as Ice Bar)
• 1897-1898 Vice-president, American Mathematical Society
• 1899-1900 President, American Mathematical Society
• 1899-1904 Professor of Mechanics and Mathematical Physics at Columbia University
• 1904 Moved to Washington, D.C. where he became president of the Carnegie Institution
• 1924 Died on June 29 in Washington, D.C.
A 1.724Mb PDF of this article is available by clicking HERE