Given the problems Michelson was having getting reliable measurements with the San Jacinto and Santiago Peak stations, it’s no surprise he began looking around for other options, and had been doing so for a couple of years. His clear favorite was some sort of pipe, either evacuated or filled with Hydrogen gas. At one point he even considered a very long wooden box, but it’s not clear if his intent was to fill it with Hydrogen, or somehow treat it to make it resistant to air and partially evacuate it.
By early 1929, Michelson and Mount Wilson staff were actively looking to “borrow” some sort of existing pipeline for their experiments. They approached a number of gas companies, as well as water districts, seeing if there were any pipelines available for their use. Not surprisingly, the answer they received was “NO”, in varying levels of loudness. San Diego County was approached, and they reminded Michelson of their obligation to bondholders, which meant delaying operation of their for a scientific experiment of uncertain duration.
Although Michelson was still lacking a specific location for a long, evacuated tube, he moved ahead in testing his apparatus, minus the beam tube. As described in this article in the Los Angeles Times of June 5, 1929, Michelson set up his mirrors 1,100 feet apart on an old railroad line in Arcadia adjacent to the Ross Field airport and made preliminary tests.
While looking for locations in Orange County, Mount Wilson staff happened to speak at length with the Engineer for the Irvine Ranch, C.R. Browning. He told them that while they didn’t have a pipeline Michelson could use, his boss, James Irvine, might be interested in letting Michelson temporarily install a surface pipeline on Irvine Ranch. Since Michelson had at his disposal $67,500 (Rockefeller funding of $30,000, a Carnegie Grant of $27,500 and University of Chicago money of $10,000), at that point it appeared the only way he would have his pipeline was if he built it himself. So on July 20, 1929, Adams wrote James Irvine formally asking permission to use part of the Irvine Ranch for the next round of Michelson’s speed of light experiments. On July 24th, Irvine replied in the affirmative, and work began.
Prior to this, Mount Wilson staff had been looking around for a pipe manufacturer capable of supplying a three foot diameter pipe a mile in length. On top of that, they also wanted it welded and leakproof. They were turned down repeatedly. Finally the California Corrugated Pipe Company, in Los Angeles, said they could provide pipe, but Michelson would be on his own when it came to sealing for vacuum.
In the Fall of 1929 construction the facility began on the Irvine Ranch. Due to manufacturing issues with the company supplying the pipe only an initial 1,100′ of the 36″ diameter pipe was available. The remaining 4,000′ wouldn’t be available until about April of 1930. Because of this, it was decided to construct most of the facility (observing room, pumphouse, generator facility) and begin testing with the shorter length of pipe. All of the fixed structures remained where they were constructed at the southerly end of the pipe. When the additional pipe became available, the pipe would added on to towards the north. By the start of Summer, 1930, testing had been completed and the full mile of pipe was in place.
Method of operation
The main reflecting mirrors were 22″ diameter flat mirrors at each end of the evacuated tube. Between these two mirrors was reflected a collimated beam from an arc lamp. This was very similar to what Michelson had done with his early tests, but there was a new twist here. In order to get the maximum length possible of the light path, the light made multiple trips up and down the length of the pipe, for a total of either 8 or 10 miles. There was the usual rotating, faceted mirror, but this time it had 32 facets and was spun by compressed air at 1,000 revolutions per second (60,000 RPM!).
Proper sealing of such a multi-component, lengthy system was, not surprisingly, problematic. After a lot of trial and error, a multiple layer approach using canvas, inner tubes, friction tape and Glyptal (sort of a paint/sealant) was used. Each section of pipe was 60′ in length, so many sealings were required across the full length of the tube. Staff spent much of their time chasing after vacuum leaks.
From Los Angeles Times, June 23, 1930
The next ten photographs I discovered in the files of the Los Angeles Public Library and they generously granted me permission to include them here:
In the course of collecting everything I could in terms of things Michelson, I came across an image for sale on a site selling historical photos. I immediately recognized the setting and grabbed it at a reasonable price (i.e., cheap!). When the photo arrived, I could see that it was some sort of press photo, dated June 21, 1930. The back side contained the notation “SF TO CLEVELAND: NY ACME Another view of Dr. Albert Michelson”. I assume its intent was to be used in a newspaper, since there was considerable touching up of the photo’s glossy surface with dark paints to reduce the background detail and make the image of Michelson more pronounced. After scanning it as it originally was, I used assorted solvents to remove the old school Photoshopping and uncovered the background details. Cool pic!
Lastly, I’m including an image of Michelson’s wiring schematic for the Irvine Ranch beam tube. This was found in the archives at Mount Wilson and is too fascinating not to share, as is. Clicking on it will pop it out larger
In the photos above, a few show the placement of the “B” tank, a welded steel housing that connected segments of the sealed corrugated pipe, and contained the viewing port through which measurements were made. Throughout the length of the beam tube there were a total of four of these tanks, each with a slightly different design. I have been told, but have not seen it with my own eyes, that at least one of these tanks still sits today on the grounds of Mount Wilson in the residential area.
During the operational life of the apparatus, from February 1931 until February 1933, a total of 233 measurements were made, roughly divided into fourths. Michelson died on May 9, 1931 when the first 36 measurements had been completed. The balance of the measurements were taken by Francis Pease and Fred Pearson.
The results were not as satisfactory as the San Antonio experiments. Much of it was due to the vacuum, or lack thereof. Michelson originally thought a vacuum of 2″ (50 mm of Mercury) would be adequate for good seeing. However it was found that to get sharp images in the tube, vacuum levels had to be brought down to 1 to 2 mm of Mercury. Given the many pipe joints, method of sealing and power of the pumps, this was almost impossible to achieve.
Beyond the vacuum, there were other problems. Heating of the residual air in the beam by the Sun caused the image to vanish, thus most measurements had to be taken after dark. Best measuring seemed to be when the tube was enveloped in the seasonal thick fog that occurs in the area. And there were strange, unexplained cyclic drifts in the measurements. It was thought for a while it could be a tidal influence, but correlation was weak. Some of it was put down to the clay like soil the tube was anchored to, which shifted due to changes in moisture content. The tube was also placed parallel to a drainage ditch, in which varying flows of water occurred.
Considering all 233 measurements, a mean velocity of light in vacuum was measured at 299,774 Kilometers/second, reasonably well grouped. This caused a bit of consternation, as the older San Antonio measurements resulted in a very well grouped 299,796 Kilometers/second, a 22 Kilometer/second discrepancy.
This resulted in quite an exchange with the US Coast and Geodetic Survey, who did the survey measurements for both experiments. They were certain the San Antonio measurements contained an error of no more than 1 part in 500,000. They felt the relatively simple Irvine tube measurement was performed to at least 1 part in 500,000. However the implied discrepancy between the results was 1 part in 12,000 for which no one had a good explanation.
For a while it seemed as if it could be a simple measurement error, as the Mount Wilson staff repeatedly measured the beam tube control points and obtained a markedly different value from the US Coast and Geodetic Survey. Upon hearing that, the Geodetic Survey folks loaned Mount Wilson some of their temperature calibrated Invar measuring tapes, and the inconsistencies vanished.
Thus the discrepancy between the Irvine Ranch measurements and the San Antonio measurements was never resolved. It was generally felt the San Antonio measurements were the more accurate, owing to their tighter grouping, and that value became the accepted value for the speed of light for many years. It was also widely believed that had Michelson lived a few years longer, his genius for experimental technique would have prevailed and the Irvine Ranch beam tube would have produced the superior results.
Einstein visits the Irvine Ranch
In researching this project I came upon an interesting fact: Albert Einstein had visited the Irvine Ranch in March of 1931. Specifically he came with members of the Mount Wilson staff (including Walter Adams the observatory Director) to meet with Michelson at the project site late one evening. In the Walter S. Adams papers held at the Huntington Library, there is a letter dated April 1, 1931 from Adams to Dr. Max Mason of the Rockefeller Foundation. In it, Adams writes:
We were somewhat afraid that you might have lost your way somewhere in the wilds of Santa Ana on the evening when we visited the Michelson pipeline. I know that Professor Einstein cast an anxious glance up the road many times to see if the lights of your car were in sight. I am very glad to learn, however, that you did not undertake the trip after all, although I know you would have enjoyed seeing Einstein and Michelson sitting side by side on an old army cot discussing the details of the apparatus.
Michelson, who at 78 had been in severely declining health due to a stoke and other maladies, died about 6 weeks after the visit.
The final blow to Michelson’s tube came on March 10, 1933, delivered by the Long Beach earthquake. With a magnitude of 6.4, the earthquake not only caused damage to the pipe itself, it changed some of the ground distances. So by the end of March, the tube was dismantled and sold off to the County of Orange Highway Department for $4,687.20. The pieces of this historic experiment would then finish out their lives as normal drain pipes in a variety of unknown locations in Orange County, its days of glory over. All that remained on the Irvine Ranch site were six piers and 32 posts for baseline measurement.
For Michelson’s very detailed full 1935 “Measurement of the Velocity of Light in a Partial Vacuum” paper, click here.
Contemporary press coverage of Irvine Ranch work
(Unless otherwise specified, LAT = Los Angeles Times and SAR = Santa Ana Register newspapers)
May 18, 1930 (LAT) Irvine Ranch tube has now been extended to its full mile length.
June 6, 1930 (LAT) “Speed of light tests tedious”.
June 9, 1930 (SAR) Final tube segments have been installed and Michelson will arrive next week.
June 17, 1930 (SAR) Michelson enroute to test site.
June 25, 1930 (Santa Cruz Evening News) Pictures of beam tube facilities.
July 24, 1930 (SAR) Testing delayed due to “accident” (vacuum pump break).
Jan 28, 1931 (LAT) Tube evacuation about to start.
Feb 20, 1931 (SAR) Einstein visits the Irvine Ranch site.
Mar 22, 1931 (Oakland Tribune) Full page spread of Michelson’s work with pictures.
The site today
Ummmm….Things are just a little bit different in Irvine from when Michelson was there. How to find this damn thing?
The plan of the experiment, Figure 2 in Michelson, Pease and Pearson’s “Measurement of the Velocity of Light in a Partial Vacuum” put me in the ball park. Then the Huntington Library had a 1932 aerial of a portion of the Irvine Ranch actually showing the tube (if you looked real closely).
With that in hand, I paid a visit to Historicaerials.com who carries a variety of old aerial photos. The oldest they had of this area was 1946, by which time Michelson’s installation was long removed. BUT….Most of the other landform features shown in the 1932 image were still on the 1946 image so I knew exactly where the installation was on the 1946 aerial. The final step was to turn on the layer showing today’s roads (It’s a cool website, worth playing around on). That gave me a very precise location of the tube and each end. Final tweaking was possible on the Historicaerials.com website by turning on layers to look at modern aerials from 2000 to present. I was able to locate each end of the pipeline within probably 150′.
Note in the old images of Michelson’s installation there’s a drainage ditch on the easterly side. That became an Orange County Flood Control District channel known as F08S01. Its southwesterly portion was eventually realigned to swing west, but its northeasterly portion is at the same location as when Michelson’s tube was sitting next to it on the west. Thus it’s an excellent reference point.
I should point out that while I was doing research for this, I did find on the Innnerwebz an exhibit purporting to show the location of Michelson’s tube. I’m not sure who originally created it but from the drawing and font it looks like something that may have appeared in the Orange County Register newspaper. It is incorrect. It shows the tube much too far to the northeast, seemingly centered on Alton Parkway. In fact, the tube began approximately at Main Street (at Mercantile) and ended at Alton Parkway (at Armstrong).
With this incorrect location, I did come across a mention that there was a plaque of some sort marking the site. I have walked the length of both the incorrect and correct locations of the beam tube and failed to uncover any plaques. Following that, I contacted the City of Irvine to see if they knew of a plaque’s existence and they responded they had no info on such a thing. The damn thing needs a plaque!
Alignment of beam tube with current land use (T. Mahood and Google Earth)For those of you who geek out at this sort of stuff, here’s a link to the KML file for the tube and its end points. It may be used with Google Earth or other GPS software.