By Duncan Steel 29/05/2019

[avatar user=”duncansteel” size=”thumbnail” align=”right” /]

Measurements of photographs obtained during the total solar eclipse of 29th May 1919 were pivotal in demonstrating the veracity of Einstein’s General Theory of Relativity, turning him into a household name. The centenary of that event is now upon us, and well worthy of being remembered. 

As I sit here typing on my keyboard, my favourite photo showing myself and my two sons aged 6 and 4 is beside me on the desk. I can tell you precisely where the photograph was taken: we were fossicking for crabs and shrimps in rockpools at St Michael’s Mount in Cornwall, England. I can also tell you the precise date: it was the tenth of August, 1999 – almost two decades ago.

The reason I can recall the date so easily is that the times of total solar eclipses are well-remembered by all that strange band of enthusiasts who chase them around the globe. If you’ve never experienced totality, it’s hard to explain… But if you’ve done it once, it can become an obsession.

The next day – the 11th of August – I was due to do the live commentary for BBC Radio 1 on the eclipse as it swept across parts of Cornwall and Devon in the far southwest of England, and then over France and regions of Europe further east. Millions upon millions went out to see it.

The writer awaiting totality over Cornwall, 11 August 1999. The screen shows the partial stage of the eclipse, obtained from an aircraft above the clouds. As can be seen, there is no blueness in the sky above, just a blanket of grey.

In the event the 10th was clear and sunny, as was the 12th, but in accord with Murphy’s Law the 11th was cloudy, even rainy. The best we could do was to feel the cold and dark as the Sun was covered by the Moon, and watch the live video feed from aircraft far above the clouds.

Published in 1999 (with an updated edition appearing in 2000), my book Eclipse was the best-seller when a total solar eclipse path crossed SW England twenty years ago.


The U.S. edition of my book Eclipse was published in 2001 by the National Academies Press. If you want to purchase any edition, get this one: it contains five additional chapters about historical eclipses. And I’ll sign it for free.

My boys were disappointed, but three years later I took them ‘home’ (they were both born in Adelaide) to see the eclipse of 4th December 2002 from just south of Roxby Downs in Outback South Australia. It was brilliant, the Sun being gradually hidden as it was setting low in the west (and so the eclipse path never reached NZ).

My sons Elliot (left) and Harry, the day before the total solar eclipse over South Australia in 2002. They have grown a bit since (so as to be far taller than me): Elliot (with a first in geophysics from U Queensland) now works for McKinsey and Company out of Sydney, whilst Harry took a first in Mechanical Engineering plus a BSc in mathematics and physics at the U of Sydney, and is just completing a DPhil/PhD in synthetic biology at St John’s College, Oxford. (Parents are allowed to be proud.)

Since then I have been to only one other: that of 13th November 2012, for which I drove north from my home in Canberra to far north Queensland, seeing the eclipse from the Atherton Tablelands (surrounded by coffee and mango plantations) inland from Cairns as the Sun rose over the Great Barrier Reef and the Coral Sea. Again, a wonderful experience, worth every minute of the drive, and every cent it cost to get there.

In the past there have been several notable total solar eclipses crossing New Zealand, but your next good chances from these islands are in 2028 (when the corridor of totality crosses Queenstown and Dunedin) and in 2038 (totality crossing Collingwood in Golden Bay and then the Kapiti Coast and Palmerston North; Nelson and Wellington are just too far south).

Too long to wait? (My answer: yes, definitely.) There is a total solar eclipse on 2nd July (the 3rd, NZ time) this year that starts out east of NZ, has the longest duration of totality somewhere near Tahiti (you’d need to be on a boat/ship to get into the track), the path then crossing Chile and Argentina. Typically the corridor of totality is only around 100 km wide, sometimes wider, sometimes narrower, but well-known in advance. You must be within that path to experience totality.

Typically there are two solar eclipses each year, but not all are total: some are annular, many are partial. To find out where and when you need to travel, the best website I know is that of Fred Espenak, who formerly worked at NASA’s Goddard Space Flight Center in Maryland. The name of the site: it has to be, because that’s Fred. There is a huge amount of vital information available there, such as a list of solar eclipses (and locations to head for) through to 2030.



Chasing eclipses is by no means a new thing. The event in 1927 over the north of England prompted the largest peacetime movement of people ever recorded in the UK, with authors such as Virginia Woolf writing about their experiences.

Tourism poster promoting travel to watch the total solar eclipse over northern England in 1927. It was a clear day, and millions of people did indeed get the thrill of their lifetimes.

I don’t want to push my Eclipse books too hard (hey, I’ll never make another cent in royalties), but one of the things that fascinated me in writing about eclipses is their historical importance, and how they have affected societies in one way or another. Eclipses have been the subject of myths, legends, superstitions… and early movies:



So, why am I writing about solar eclipses now? Well, May 29th is the centenary of the total solar eclipse of 1919, and in terms of science that is one worth remembering. Shown below is the pre-determined path it took.

The narrow path of totality of the total solar eclipse of 1919, which was observed by major expeditions to Brazil and the islands of São Tomé and Príncipe in the Gulf of Guinea.

Looking at that chart you might wonder how it was calculated, given that it was before the development of electronic computers (with which we are now able to compute eclipse paths with utmost precision). There are two answers to that question: one is that eclipses were of huge importance, scientifically, and so great effort was put into determining the paths using pencil, paper, tables of logarithms, and hand-cranked calculating machines. The other answer is that eclipses follow a whole set of nested patterns, one of which is the Saros cycle which had been recognised in ancient times: eclipses recur after 18 years plus 10/11 days (depending on the leap-year phase) with offsets across the Earth’s surface. Here is a diagram showing six solar eclipse paths of totality, of which 1919 is one member:

Six eclipses from one Saros series: 1901, 1919, 1937, 1955, 1973, and 1991. These paths were known well ahead of time. The eclipse of 1919 has its centenary on 29th May. The eclipse of 1973 is notable because it set a record for the duration of observed totality, although there was some cheating involved: the eclipse was chased across Africa on board the supersonic Concorde airliner.

Let’s get back to 1919. Better still, let’s get back to late 1915. In November of that year Albert Einstein published his General Theory of Relativity (GTR). Fundamentally, this is a gravitational theory that shows where Newton’s theory from two centuries earlier is not quite correct. In 1905 Einstein had published his Special Theory of Relativity, and I distinctly recall reading about it at age 14 in the book Thirty Years that Shook Physics by George Gamow. (I was shaken to the core. I still am now. It was this book that diverted me into studying physics. Ten years after I read it I found myself occupying an office in the Gamow Tower at the University of Colorado.)

One of the predictions of GTR – falsifiable hypotheses are vital ingredients in the progress of science – was that the gravitational attraction of the Sun on light from distant stars is twice that which the Newtonian theory would produce. There are a few physicists now grinding their teeth at me writing things like that, but the idea I am wanting to convey is valid.

The point here is that photons of light, having an equivalent mass (via E=mc²), would be expected to be deviated by the Sun’s gravitational attraction. But not by much, even if those photons were skimming just above the top of the solar atmosphere. So how could they be measured?

The answer comes from a total solar eclipse. With the Sun’s phenomenally-bright disk being blocked by the Moon, stars that happened to be close to the limb of the Sun would appear to be in slightly different positions than when they might be measured six months before or after such an eclipse, when they would be in a dark night sky. Newton’s theory said that they would have their apparent positions shifted by a certain amount (which is actually tiny, and so difficult to determine with precision), while Einstein’s theory said that the shift would be twice as much (but still only 1.75 arcseconds, about the same amount that a star is blurred by atmospheric turbulence at most locations).

The concept of the deviation of light from distant stars by the Sun’s gravitational attraction. The dashed lines show the apparent positions of stars photographed during the total solar eclipse. The solid lines show the actual positions of the stars, which can be determined from photographs when they are far from the solar direction (e.g. six months before or after the eclipse). The Newtonian theory of gravity says that the angular shift of the stars’ images should be a certain amount, whilst Einstein’s theory says that it should be twice that. (Diagram from my books entitled ECLIPSE, natch.)

Hitherto Einstein had been largely ignored by the public, and many physicists refused to believe that GTR was correct. Newton wrong? Surely not!

The eclipse of 1919 rendered an opportunity to validate Einstein’s relativity theory, though. British and Irish expeditions were mounted to Brazil, and the islands in the Gulf of Guinea.



Later accounts make it sound like a pleasant voyage for a few days in the tropics, but it was far from that. As noted earlier, the locations of the reference stars were needed six months away from the eclipse. Just as during what I like to think of the Heroic Age of Eclipse Expeditions in the nineteenth century (one voyage to the Pacific to observe an eclipse resulted in giant land crabs stealing the astronomers’ leather boots), things were tough for people like the Cambridge dons who set out for São Tomé and Príncipe off the coast of mainland Africa.

Their leader was Arthur Stanley Eddington. Many people have heard of Eddington (although they may not know him by name) due to various peculiar prognostications he is reputed to have written or uttered. Perhaps the most famous is the idea that he started a question he set for the Cambridge University Tripos examinations with the statement: “Consider a perfectly spherical elephant, whose mass may be neglected.”  Of course those of us who have studied both physics and Latin know that an elephant is a pachyderm, and so cannot be perfectly spherical. (My comment there may explain why I am rarely invited to parties.)

Anyhow, here is Eddington:

Sir Arthur Stanley Eddington (1882-1944), who may have been more fun that he looks.

…and here is one of the photographs obtained during totality in 1919, by Eddington and his colleague Edwin Cottingham:

Total solar eclipse photograph from 1919, obtained by Arthur Stanley Eddington and Edwin Cottingham. (Courtesy the Royal Astronomical Society.)

Obviously enough, no stars are readily apparent (bar the Sun, of course). But with suitable magnification and image processing – not simple in those pre-computer days – the background stars could be detected, and their positions measured.

Eddington, clearly, was not the only astronomer along on the trip. A similar team was sent to Sobral in Brazil. The Astronomer Royal, Sir Frank Watson Dyson, who was reputed to be a disbeliever of Einstein’s GTR, was one of the main motivators behind the expeditions. An attempt to observe a total solar eclipse the previous year, in 1918, had not been successful. In 1919 the heavyweights of British astronomy were dispatched.



Although the eclipse occurred as anticipated on May 29th (under skies that were mostly cloudy, though the eclipse could be seen occasionally through the gaps), in these days of instant global communications it is hard to understand why it took until November of that year for news of the results to become public.

Actually, a large part of the reason is that (as described above) various photographs of the background stars (in the constellation Taurus) needed to be obtained whilst they were in the dark, night sky. Another reason is that, contrary to the stories told to the media at the time (and often repeated nowadays), the results were not clear-cut. The headlines (such as that below) bellowed that “EINSTEIN THEORY TRIUMPHS” — but in reality it was not quite that way, for the scientists directly involved.

Headlines in the New York Times dated 10th November 1919.

Rather than bore you all here with the details, I will merely refer you (dear reader) to an excellent article that appeared (like the above clip) in the New York Times, but in this case less than two years ago. Denis Overbye was the writer in question. Therein he relates how three sets of measurements of the apparent angular shift of the stars due to the solar gravitational attraction were obtained from the expeditions to Sobral in Brazil, and the island of Príncipe off the coast of Africa.

The thing is this… If one took all three sets of data, then one might conclude that the answer is closer to the prediction of Newtonian theory, than Einsteinian. However, clearly one set of data (that from an astrograph in Brazil) was unreliable, because the stars were out of focus. Throwing that set out, Eddington concluded that Einstein’s GTR was correct.

Did Eddington and Dyson make the right call? The answer is that subsequent measurements of the positions of stars confirmed that the shifts due to the Sun’s gravitational pull are in accord with Einstein’s GTR, and therefore the Newtonian theory of gravity, although a very good approximation for most applications, is not a complete theory.

Nowadays astronomers do not need to wait for eclipses to measure the shifts of stars, or travel to extreme locations to be within a solar eclipse track. Radio telescopes working at suitable wavelengths can observe stars as they pass behind the Sun without needing to block out solar emissions.



As time goes by while I am typing this, it is looking like I might not get it completed before 29th May begins, at least on New Zealand time. So I need to bring this post to a close.

I did want to finish up by noting that, whilst the centenary of the 1919 eclipse seems to have come unnoticed by many, others have been planning to celebrate the event. At Sundy in Príncipe they are clearly determined to remember when their little town hit the global headlines:

Eddington’s expedition observed the total solar eclipse of 1919 from the small town of Sundy in Príncipe.

Here is their rather nice website in which a host of events commemorating the eclipse in 1919 are listed and, most important, it is clear that people in São Tomé and Príncipe are using this centenary to encourage interest in science amongst their younger citizens. Wonderful!

There’s another centenary this year: that of the International Astronomical Union (of which I am a member). Noting the conjunction (a good astronomical term) with the 1919 eclipse, the IAU has instigated an educational programme entitled Einstein Schools.



There are many educational resources freely available via the Einstein Schools website. In view of that, it seems a pity that in Australia and Oceania only three schools appear to have joined the project so far, and all of those are in New South Wales. Becoming a member has a wide variety of benefits, including opportunities to interact with schools spread globally.


Leaving that aside, for more information about the centenary of the IAU, see the dedicated website.

This year also marks the centenary of the International Astronomical Union.


Midnight on 28th May is rapidly approaching, at a rate of one minute per minute, and so I need to finish this long post up and publish it. Let me conclude with mention of the Royal Astronomical Society, which part-sponsored the eclipse expeditions of 1919. As you might anticipate, the RAS has been a major player in commemorations of the significance of this centenary, and many of the original photographs obtained in 1919 can be viewed in the lovely post that appears here. The RAS, along with other UK institutions and universities, is also involved in a website dedicated to the 1919 eclipse and its significance in the history of science. I much recommend a perusal of the pages and images available there.

But the clicking clock has beaten me. It’s now fourteen minutes into 29th May, the centenary date (at least on NZST), and it’s time for me to put this blog post out and head for bed.



Update 30 May (NZST; still 29 May UTC): The IAU has put out a newsletter detailing some of the commemorative events taking place. Here is a webpage describing celebrations at Sobral in Brazil, one of the main places where eclipse photographs were collected in 1919. A three-day conference just concluding in Paris (where the IAU has its headquarters) is entitled Arthur S. Eddington: From Physics to Philosophy and Back Again. A short planetarium show called Checking up on Einstein is available, covering why light is deflected in a gravitational field and how the 1919 eclipse was used to verify that Einstein’s General Theory of Relativity is (so far as we know, at this stage in our scientific development) correct. As noted above inter alia, there are several other websites dedicated to information concerning the significance of the 1919 eclipse observations, such as this one.



Update 31 May: The first-ever movie of a total solar eclipse (in 1900) has just been made available. Take a look here. What bemuses me is that this article in the Guardian does not mention the fact that the photographer (a professional magician) was the namesake of the fifth Astronomer Royal (Nevil Maskelyne, 1732-1811). He (the photographer-magician, not the Astronomer Royal) also invented the pay toilet, so it is to him that we owe the phrase “spending a penny”.

Readers might also find interesting this recent article about the eclipse of 1919 in the Observer (the Sunday edition of the Guardian, in essence).

Addendum 6 June: I just came across this article from The Economist issue of 28 November 2015 – the centenary month of Einstein first presenting GTR to the world – and it begins with these words:
“ALFRED, it’s spinning.” Roy Kerr, a New Zealand-born physicist in his late 20s, had, for half an hour, been chain-smoking his way through some fiendish mathematics. Alfred Schild, his boss at the newly built Centre for Relativity at the University of Texas, had sat and watched. Now, having broken the silence, Kerr put down his pencil. He had been searching for a new solution to Albert Einstein’s equations of general relativity, and at last he could see in his numbers and symbols a precise description of how space-time—the four-dimensional universal fabric those equations describe—could be wrapped into a spinning ball. He had found what he was looking for.

The article was headed by the image below: