This is the first of eight posts on evolutionary research to celebrate Darwin's bicentennial.
If you were a designer tasked with creating a machine for collecting and processing light, the last thing you would come up with is the human eye. Darwin marvelled at the eye as an "organ of extreme perfection", but in this, he was wrong. Aside from the many illusions that can fool them, our eyes have a major structural flaw. In humans and other back-boned animals, the light-sensing cells of the eye - the "photoreceptors" lie at the back of the retina.
In front of these sensors lie several layers of nerve cells that carry their signals, and blood vessels that supply them with nutrients. The nerves join to the main optic nerve which passes through a hole in the centre of the retina and connects to the brain. It's a back-to-front design. Light has to pass through several layers of nerves, not to mention blood vessels, before it hits the retina itself. It's a bit like designing a camera, and sticking the wiring in front of the lens.
Octopuses and squid have a very similar eye to ours, but theirs' are much more sensibly structured. Their nerves and blood vessels connect to the light sensors from behind so that light can hit the photoreceptors without having to negotiate an obstacle course. And because their retina doesn't need a hole to accommodate the optic nerve, they have no blind spot.
In our own retinas, nerves and vessels are random in their spacing and irregular in their shape. The light that shines past them is reflected, scattered and refracted. It's amazing that our eye can see at all. But even though there is clearly no designer, evolution has done a pretty good job instead, with its remarkable capacity for making the best of a bad lot. In the case of our eye, some of the obscuring cells act as living optic fibres, to funnel light onto the sensors they cover.
Kristian Franke and colleagues from the Paul Flechsig Institute for Brain Research first noticed these fibres by shining light onto the retinas of guinea pigs. They looked at a cross-section near where the photoreceptor lie and saw a very regular pattern of bright spots. Clearly, some parts of the retina were transmitting light far better than others.
As they looked at further cross-sections throughout the retina, they realised that the bright spots were the endpoints of long tubes that stretched throughout the retina. Near the top, the tubes widened into funnels. Franze identified these tubes as Muller cells. These brain cells aren't nerves themselves, but are part of their supporting cast. They are long cylinders arranged in columns across the entire retina, and provide a route for light to pass through the tangled morass of nerves and blood vessels.
The Muller cells gather light at the top of the retina and channel it to the light sensors as a tight beam. Along the way, the light is barely reflected or scattered and little is lost when it finally reaches the photoreceptors, just like modern optic fibres.
Light enters the Muller cells at a shallow angle and is slowed down considerably by the cells' high refractive index. When it hits the cells' boundaries, it is almost completely reflected back along the tube. Their funnel shape allows the Muller cells to gather and transmit as much light as possible. But as they narrow in the middle, they take up a very small amount of space and leave plenty of room for the blood vessels and nerves that the retina needs.
On average, each Muller cell serves a single cone cell and several rod cells. This one-to-one system ensures that the images that eventually hit the light sensors keep strong contrast, and are not distorted.
Evolution has given the vertebrate eye a remarkably ingenious solution to its ludicrous inverted retina. The eye may not be the perfect organ that Darwin thought, but new insights into its evolution still provide us with awe-inspiring surprises.
Reference: Franze, Grosche, Skatchkov, Schinkinger, Foja, Schild, Uckermann, Travis, Reichenbach & Guck. 2007. Muller cells are living optical fibers in the vertebrate retina. PNAS 104: 8287-8292.