In the summer of 1964, Dave Wilkinson and Peter Roll were on the brink of an epoch-making discovery. But as they busied themselves high on the roof of Princeton’s geology building, assembling the radio antenna with which they intended to take the temperature of the Universe, another antenna less than a hour’s drive east of Princeton was already registering a peculiar and persistent hiss of radio static that was coming from every direction in the sky.
For two deeply puzzled young radio astronomers at the Bell Telephone Laboratory in Holmdel the mysterious hiss marked the beginning of the most frustrating year of their lives – a year in which they were destined to spend more time removing pigeon droppings from their antenna than actually making observations of the Universe.
Arno Penzias was 31 years old, a dynamic New Yorker who had come to the US as a refugee from Nazi Germany. Robert Wilson was a taciturn 28-year-old who had moved east after completing his graduate work at Caltech in Pasadena. In 1963, the two had teamed up to work on a very unusual radio antenna which Bell Labs had built at its Holmdel site in northern New Jersey.
The antenna, designed for satellite communications, stood on top of Crawford Hill, a low wooded knoll which barely rose above the flat monotony of the surrounding New Jersey countryside. It looked nothing at all like a familiar radio dish; in fact, it was really rather hard to describe. Someone had referred to it as ‘an alpenhorn the size of a boxcar’, but a better description might be a giant ice-cream cone laid on its side.
A 20-foot-square opening had been cut in the side of the cone, just beneath where the ice cream should have gone. This opening collected microwaves from the sky. They were then funnelled down to a sensitive radio ‘receiver’ installed in a cramped wooden cabin at the cone’s tapered end. People could work in this cabin, fiddling with the receiver’s electronics while they watched a red pen trail across a chart recorder, depicting the radio signal the antenna was picking up. The entire antenna could be turned about two separate axes so that the 20-foot opening could be pointed anywhere at all in the sky.
Despite its unusual appearance, the antenna on top of Crawford Hill was simply a standard microwave horn. In fact, it was little more than a larger version of the one Dave Wilkinson had soldered together at Princeton. Bell Labs had built it in 1960 in order to bounce radio signals off the Echo 1 satellite, a sort of stone-age communications satellite which was the ancestor of all the satellites that make today’s world a smaller place. Echo 1 was like a silvered beach ball 100 feet in diameter. It hung up in space, a brilliantly bright artificial moon in the night sky.
The problem of picking up a weak radio signal reflected from a tiny satellite was a formidable one. The engineers at Bell Labs not only had to develop an extremely sensitive radio receiver capable of measuring differences in temperature of only a few tenths of a degree, but they also had to design a very special antenna.
The basic problem they faced was that the radio signal from a satellite – little more than a pinprick in the sky – would be utterly swamped by unwanted radio waves coming from nearby sources such as the ground. So the engineers at Bell Labs had to design an antenna that would keep out all radio waves except those coming from the direction of the satellite. By a coincidence, it was precisely the problem Wilkinson and Roll had to overcome in their search for the fireball radiation.
At Bell Labs, they solved it with the ice-cream-cone design. When the Holmdel antenna was pointed at a source in the sky, it was almost impossible for radio waves from the ground to bend their way into the 20-foot opening.
Echo 1 was superseded by a more sophisticated communications satellite called Telstar, and the ice-cream-cone antenna was adjusted so that it could transmit and receive microwave signals from Telstar instead. It was while the Telstar project was under way that Bell Labs hit upon the idea of hiring two radio astronomers to come and do some astronomy with their unique antenna. The company reasoned that since radio astronomers were also in the business of pushing the technology of detecting radio waves to its limits, Bell Labs might benefit from having some around.
Arno Penzias was recruited in 1962. He came straight from New York’s Columbia University, where he had been the student of Charles Townes. Townes was the inventor of the maser – the microwave predecessor of the now familiar laser.1 A year later, Bell Labs recruited Robert Wilson. At Caltech he had worked for John Bolton, an Australian who was one of the pioneers of radio astronomy.
While working on his thesis at Caltech, Wilson had got to know a Bell Labs man called Bill Jakes. Jakes showed up at Caltech at regular intervals to talk with the radio astronomers there and to ask if anyone might be interested in a job working with the 20-foot antenna at Holmdel.
Wilson had already gained a good impression of Bell Labs because he had worked with some of their people up at Caltech’s Owens Valley radio observatory in northern California. The company had lent Caltech some experimental equipment for the radio telescopes, and Wilson had helped to install it.
Shortly after he finished his thesis, Wilson applied for the job with the 20-foot horn antenna. ‘At the time I wasn’t sure I wanted to continue in astronomy,’ he says. ‘But the Bell Labs job was an opportunity to continue doing astronomy, and I could also see a lot of other things going on at the company that were interesting to me.’
Wilson got the job and arrived at Bell Labs in March 1963. He soon met Arno Penzias, and they decided to join forces. ‘Arno and I were the only two radio astronomers at the place, so it was natural for us to team up,’ says Wilson.
It was destined to be a perfect partnership. Not only did the two radio astronomers have complementary technical skills, but their personalities complemented each other as well. Wilson was quiet and cautious, while Penzias was brash and outspoken. But though on the surface they were like chalk and cheese, they shared one important characteristic which would go a long way to ensuring their eventual success: when it came to doing science, both men were meticulous and painstakingly thorough.
Although Arno Penzias had been at Bell Labs for a year, he had not yet been able to get his hands on the 20-foot horn because it was still being used for Telstar. But that all changed shortly after he and Wilson teamed up. ‘The Telstar people agreed we could do some astronomy with it,’ says Wilson.
Immediately, Penzias and Wilson set about modifying the antenna for astronomy. ‘The 20-foot antenna was unique,’ says Wilson. ‘It produced very little in the way of unwanted radio signals, and there was the possibility of determining exactly how big those signals were and where they were coming from.’
This made the antenna good for making ‘absolute measurements’ – that is, measuring how bright a source at radio wavelengths really was rather than simply comparing it with the sky background. Of course, if Penzias and Wilson were to exploit it for this purpose, they would need an artificial source of radio waves with which to compare any astronomical source.
So it was that Penzias began building a cold load. The device he came to build was remarkably similar to the one Peter Roll was building down the road at Princeton. Both devices used a piece of wave guide and both were cooled by liquid helium to just 4.2 degrees above absolute zero. A crucial element of the design of the Bell Labs experiment was a switch which allowed the temperature of the sky and the cold load to be compared rapidly.
In 1964, it was probably fair to say that there were only two liquid-helium cold loads in existence in the world. It was rather a coincidence that they had been built independently by two groups of astronomers who were unaware of each other’s existence and who were only 30 miles apart.
Now equipped with its cold load, the Holmdel antenna was ideally suited for picking up a faint background signal from the sky. And that is precisely what Penzias and Wilson intended to do with it.
For his thesis with John Bolton at Caltech, Wilson had made a map of the Milky Way at radio wavelengths. He had suspected that surrounding the starry disc of the Milky Way was a vast ‘halo’ of gas glowing faintly at radio wavelengths, but he had been unable to prove it. The reason was that he had made his map by using the standard technique of comparing the brightness of the Milky Way with the background sky. His technique was therefore incapable of measuring the brightness of the Milky Way’s faintly glowing halo, as in effect this was the background sky.
The 20-foot horn, with its ability to measure the radio signals from faint background regions of the sky, was an ideal instrument for measuring the weak radiation from the Milky Way’s halo. Penzias and Wilson decided to look at a wavelength of 21 centimetres. If the halo glowed at all, it would glow at 21 centimetres. This was because it should be made of neutral hydrogen gas, which broadcasts a very distinct radio signature at this wavelength.
But the two radio astronomers knew that actually looking for the Galactic halo at 21 centimetres would be tough. The halo was likely to be very faint, and to register no more than a temperature of one degree at their antenna. Other, unwanted, signals from the antenna and receiver and the atmosphere would be much larger. So it was clear to Penzias and Wilson that before they attempted the halo measurement they would really need to understand their instrument and know where all the unwanted radio signals were coming from and just how big they were.
The Telstar people had left the 20-foot horn with a receiver set up for a wavelength of 7.35 centimetres or 4,080 megahertz. Penzias and Wilson therefore decided to take advantage of this and try to understand completely what was happening in their instrument at this wavelength before going to the trouble of building another receiver sensitive to 21 centimetres.
It turned out that observing the sky at 7.35 centimetres would be a particularly neat test of their ability to measure its temperature because at this wavelength the Milky Way’s halo should be essentially invisible. So when the antenna was turned on the sky and all the sources of unwanted static were accounted for, the 20-foot antenna should register only a signal from the antenna structure itself, and this should be almost zero. So if Penzias and Wilson pointed their antenna at the sky and had no signal left when they had done their accounting, then all would be okay.
Penzias and Wilson did just this in June 1964. They fully expected to measure a sky temperature of zero degrees. But it was immediately clear that something was very wrong. Their horn was generating more radio static than they expected. Even when they had accounted for every source of unwanted radio waves, the instrument was still registering a signal. It was precisely what would be produced by a body at a temperature of just 3.5 degrees above absolute zero.
‘When we made that measurement, Arno’s first reaction was, “Well, I made a good cold load,”’ says Wilson. If the cold load had been reflecting any radio waves back into the antenna, then it would have appeared to be hotter than 4.2 degrees and so screwed up Penzias and Wilson’s accounting job.
After satisfying themselves that the cold load was okay, the two astronomers wondered whether they might be picking up a man-made signal in the urban environment of northern New Jersey. ‘The best place to do radio astronomy is a completely isolated valley that’s shielded from all radio interference,’ says Wilson. ‘But the Holmdel antenna had been built on top of a hill so that it would get complete coverage of the sky for satellite communications.’
If the anomalous signal was man-made, then the obvious source was New York City, 30 miles to the north. But when Penzias and Wilson checked this out by pointing their antenna in the direction of the city, the signal on their chart recorder did not jump. In fact, the ghost signal at 4,080 megahertz stayed the same wherever they pointed the instrument around the horizon.
It turned out that Penzias and Wilson were not the first people to encounter the problem of the peculiar excess signal. As early as 1961, Ed Ohm, an engineer working on the 20-foot horn, had noticed that the instrument was registering more static than expected when it was pointed at the sky. With the horn receiving signals bounced off the Echo satellite, Ohm had added up all the sources of unwanted radio static. He had found that the antenna was picking up something like three degrees more than he could account for.
Ohm did not pay too much attention to this excess temperature because some of the contributions in his accounting sum were uncertain by more than three degrees. Without a cold load it was impossible to pin down where the excess static was coming from. Nonetheless, Ohm published this result in The Bell System Technical Journal.
Penzias and Wilson wondered whether their ‘amplifier’ circuits were producing the excess signal. Amplifiers are part of any radio receiver. They are needed because the electrical currents generated by radio waves in an antenna are so tiny that practical detectors usually cannot register them. Instead, the currents have to be magnified electronically by ‘amplifiers’ before they reach a detector.
The two astronomers compared the signals coming from the antenna when it was looking at the cold load and when it was not. Since the signal from the amplifier circuits had to be the same in both cases, it neatly cancelled out. What was left was the signal coming from the Holmdel antenna alone. They knew this was made up of contributions from the metal structure of the antenna, from the Earth’s atmosphere and from any astronomical sources of radio waves that happened to be in the direction the antenna was pointing.
The static from the atmosphere was easy to identify and subtract because of its distinctive characteristic: the hiss was strongest when the antenna was pointed at the horizon – the direction in which the atmosphere is thickest – and weakest when it was pointed straight up – where the atmosphere is at its thinnest.
Of course, the anomalous radio signal could have been real. But this seemed too ridiculous to contemplate. For a start, it could not be coming from the Sun or the Milky Way because neither covered the whole sky, and the signal quite definitely did. The only other possibility was that the signal was coming from the Universe as a whole. But the astronomers knew of no astronomical source that could be generating such a constant radio signal. Clearly, there must be a fault in the antenna causing it to generate more static than Penzias and Wilson had realised. They were confident that the antenna was generating very little static. This was one characteristic of the Holmdel instrument that had convinced them in the first place that it was uniquely suited for making the difficult Galactic halo measurement. But Penzias and Wilson were nothing if not meticulous. They decided to look at the antenna in detail.
Their gaze settled on a pair of pigeons which were roosting deep inside the ice-cream-cone antenna, just at the point where it entered the wooden cabin. ‘It was a nice comfy place in there because the end of the antenna was up in our heated control room,’ says Wilson.
It might have been warm in there, but it was difficult to build a nest. Every few days, Penzias and Wilson turned their antenna, tipping the pigeons onto their heads.
The pigeons had left their distinctive mark on the inside of the great ice-cream cone. To Penzias, a radio engineer through and through, it was ‘a white dielectric material’. But to anyone else it was simply pigeon shit.
‘Until now, we’d been operating quite happily with the stuff in place,’ says Wilson. ‘There were no big heaps of the stuff because anything loose fell off whenever we turned the antenna around.’
So could the pigeon droppings that were coating the inside of the antenna be responsible for the mysterious static? Since everything above absolute zero gives out radio waves, the pigeon droppings would certainly be glowing at microwave wavelengths. By now, Penzias and Wilson were desperate enough to consider anything.
They decided to eject the pigeons. But this proved to be no easy task. From a local hardware store they bought a ‘Hav-A-Heart’ trap, which they put at the end of the antenna after removing some of their receiver. The Hav-A-Heart trap was a wire-mesh cylinder with a droppable gate at either end. You put food on a feeding tray at the centre of the cylinder, and in theory when an animal or bird walked in and disturbed the feeding tray the two gates were triggered to fall. It worked perfectly. ‘I think we got one pigeon one day and the other the next,’ says Wilson.
They put the pair of pigeons in a box and mailed them to Whippany, another New Jersey site of Bell Labs 40 miles to the north-west of Holmdel. ‘We sent them there because it was the most distant place we could send them in the company mail,’ says Wilson. At Whippany, a man had agreed to accept them and turn them loose.
Once the pigeons were on their way, Penzias and Wilson set about removing the pigeon shit. They climbed into the gloomy interior of the horn antenna armed with brooms. ‘It wasn’t a big job,’ says Wilson. ‘After an hour of sweeping, we’d removed everything.’
Penzias and Wilson thought they had seen the last of their pigeons, but they were wrong. ‘Two days later, the pigeons were back in the antenna,’ says Wilson. ‘By now we decided we had given them a good chance. There was a guy in the machine shop who was a pigeon fancier and he told us these were junk pigeons, and we were not to worry about them. One day, he brought in his shotgun and blew them to kingdom come.
‘The oddest thing’, says Wilson, ‘is that since our pair of pigeons, no others have ever nested up there in the 20-foot horn.’
With the pigeons gone for good, Penzias and Wilson thoroughly cleaned the interior of their antenna. The antenna was made of aluminium sheets which were riveted to aluminium beams. Thinking that the rivets might be causing the spurious signal, they put aluminium tape over them. So careful were they that they had even checked that the adhesive on the back of the tape generated negligible radio waves. Now, surely, they had thought of everything. At long last they would be able to do some radio astronomy.
Penzias and Wilson pointed the 20-foot antenna at the sky and looked at the reading on their chart recorder. To their dismay they saw that the spurious static had decreased only slightly. It had not gone away. The horn was still registering an anomalous temperature of 3.5 degrees above absolute zero.
By now the excess signal had persisted for almost a year. As far as Penzias and Wilson could tell, it was the same in all directions and it did not vary with the seasons. They were also able to rule out two additional sources of radio waves. It could not be in the Solar System because any source should have moved around the sky as the Earth orbited the Sun. Also, it could not be due to a nuclear test. In 1962, a high-altitude nuclear explosion had injected ionised particles into the Van Allen radiation belts, high above the Earth. But any radiation from this source should have reduced considerably within a year of the explosion.
Penzias and Wilson were at a loss for any further explanations. This tiny but persistent effect had sabotaged their plan to observe the halo of the Galaxy. But just when they were at their wits’ end, Penzias happened to make a phone call …
1. Townes was to win the Nobel Prize in 1964 for inventing the maser.