Sometimes a major discovery - like finding evidence to support the theory of dark matter - just requires a bit of creative thinking over a curry, as Dr Karl explains.
When professional astronomers design new telescopes, it takes forests of paperwork, big buckeroonies (tens of millions of dollars minimum), and at least a decade.
But hoorah for lateral thinking and hobbies. An astronomer's interest in nature photography led him to a radical new telescope very quickly and cheaply - and also got us one step closer to solving the mystery of how galaxies spring into existence.
One top theory of how galaxies form involves dark matter. Way, way back, over 13 billion years ago, just a few thousand years after the big bang, practically all the mass in the universe was this mysterious dark matter.
The dark matter began to clump together, thanks to gravity, and began to shape itself into roughly spherical objects - which began to collapse inwards.
Various gases (such as hydrogen and helium) collected at the centres of these spheres, turning into the first stars - and the first galaxies.
After a few billion years, the small galaxies merged with each other, eventually evolving into giant galaxies, like our own Milky Way.
But there's a catch with this dark matter theory. If galaxies formed that way, there should be (around each big galaxy) vast messy debris fields left over from the creation process. We would expect to see random ejected stars, partially eaten halos of gas, bulges and streams of matter, and lots and lots of very faint dwarf galaxies. But we can't find them. What cosmic dust bin have they been swept into? Or is the theory wrong?
You need to know a bit about telescopes to understand the next bit.
Most telescopes are inherently not very good at seeing this left-over debris - because it's soft and fuzzy and faint. And this is because the overwhelming majority of professional telescopes catch the incoming light with curved mirrors, not curved lenses. Curved mirrors are really good are seeing small bright objects - which is the exact opposite of what we are looking for.
One good thing about curved mirrors is that you can make them really big. Over the last half-century, there have been tremendous advances and improvements in the performance of mirror-type telescopes. In the specific field of gathering light from small bright objects, they're 100 times better than before.
But when it comes to gathering faint light from large diffuse objects, mirror-type telescopes have had no real improvement over the last half-century. This is for various technical reasons. First, the mirrors themselves cause scattering of the light, due to micro-roughness and dust. Also the secondary mirror in mirror-telescopes creates a large obstruction in the light path, and furthermore, the supports that hold this secondary mirror cause diffraction and bending of the incoming light.
So the ideal telescope for looking at large faint objects would have no mirrors and no obstructions to the incoming light- in other words, it would use lenses, not mirrors. But apart from solar telescopes, professional astronomers haven't used lens-type telescopes for a century.
Now back in 2011, two professional astronomers, Roberto Abraham and Pieter van Dokkum, were in a Nepalese restaurant. After curry and rice and lots of beer, they were shooting the breeze (as you do) about how to find these large faint bits of debris that they believed should be there - left over from the creation of galaxies.
Pieter can Dokkum was a keen nature photographer, and he suddenly realized that a recently released camera lens with a wonder coating on the front might just be perfect for their needs. It was a 400 mm f2.8 SuperTelephoto lens, costing around $10-15,000. The lens coating was called Subwavelength Structure Coating - actually, countless tiny cones or pyramids, on the front of the lens, all pointing outwards. These cones are microscopic - smaller than the wavelengths of visible light. The Physics is complicated, but the end result is that less light is scattered inside the lens - so there's less of what the photographers call "ghosts" or "flares".
And though it's not what this lens is designed for, it can pick up very faint objects.
By March 2012, they had spent about $15,000, done some testing, and found that their single off-the-shelf lens had captured what other astronomers had previously only got hints of. They saw a clear, but very faint, halo of diffuse matter surrounding the galaxy called M51.
Well, if one lens is good, surely three must better. So they swiped the credit card, got another two lenses, and built a special structure so that all three lenses were perfectly lined up. Very shortly afterwards, in September 2012, they got some results - yah! it all worked. By 2013, they were running eight lenses in parallel. In 2014, they published their findings that the galaxy called M101, or the Pinwheel Galaxy, had three previously undiscovered very faint dwarf galaxies orbiting around it.
The astronomers (who admit they can't leave well enough alone) have since upgraded their telescope to 50 lenses. They can now get images in hours, not weeks. Their credit card bill has run up to about half-a-million dollars - but it's still a lot less than tens of millions of dollars, and it was much quicker because the camera company had done all the expensive design work.
They call their array of professional lenses the Dragonfly Telephoto Array for two reasons. First, with 50 commercial telephoto lenses, it looks like the eye of a dragonfly - not a mirror to the soul, but a lens to previously hidden galaxies. And second, Pieter van Dokkum really likes taking photos of dragonflies …