The insect eye: compound benefits

One of the topics you researched at the Gatty was the compound eye. What does the compound eye look like, and how did it evolve?

A compound eye is an eye of many facets. It is like an array of detectors, each with its own lens. The array itself is like a thistledown spray, with the axes pointing in all directions, and each unit acts as a miniature camera with one silver grain. In effect, it is one axis of detector. And you can have a great many of these. Some species of insects have very few facets, but some have tens of thousands.

Insects can have eyes like knobs on sticks. A nocturnal dragonfly, which catches mosquitoes at night, has its eye almost completely enveloping the whole head. And a mantid shrimp has three visual axes in each eye, so it has six visual axes looking at an object at the same time.

Compound eyes occur in the very first animals. Down in the Burgess Shale at four or five hundred million years ago there are animals with compound eyes. And this has evolved because it is extremely efficient for panoramic vision – 360° around, and over the top and underneath. It is impossible to make a panoramic eye with a single lens. Even if you make a wide-angle lens, the aberrations build up very quickly as you increase the field size.

Light guides in insect eyes

At about the same time there is the first mention of light guides in insect eyes.

That too came out of teaching, and also I had done the vision of the compound eye in the book with Bullock, covering all the literature on compound eyes. In 1962, two scientists in Britain published a paper saying that light comes in through several facets at the same time, and interferes behind the facets within the eye. Like one or two other scientists around the world, I just did not believe that eyes of things like bees, flies and grasshoppers work in this way. So, with microelectrodes, I recorded from locust eyes and showed basically that each facet has its own field, the field is generated by the interaction between the lens and the end of the light guide, and the rod-shaped structure containing the visual pigment acts as a light guide.

It is exactly the same in our eyes. Our rods are light guides. But in 1962 this was not a popular subject. There were perhaps two or three people in the world interested in the optics of rods and cones; nobody was interested in the optics of insect light guides.

I had a couple of good students at the time. One of them, John Scholes, used to get up so late that he had to work at night. And so he discovered that the locust eye sensitivity increases about a thousand-fold at night, and you can then record single photon arrivals. So one day he comes in and says, 'Look at these things I've recorded. You turn down the light to nothing, and this is the response to my white shirt!' This was in the dark room, with nothing but a glow from an oscilloscope indicator light to provide a bit of light. The facet of the locust eye is, say, 25 to 30 microns, so the area to catch photons is pretty small. The photon flux cuts down to about 10 a second with a very dim light.

Scholes said, 'What are all these little bumps?' They were random in height, and randomly spaced. I suppose someone said, 'They probably are photons.' Someone else probably said, either then or within a day or two, 'Are they Poisson distributed?' and of course they were. Then we did very careful measurements and showed that it was first-order Poisson system – there was no summation between photons required in order to get a response, and if two photons are required, then it is a second-order Poisson. And so on. We were onto that in 1962. You could do exactly the same thing with a photo-multiplier tube, of course, but we had an insect eye.

Focus questions

• How do compound eyes differ from human eyes?
• Horridge says that it is impossible to make a panoramic eye with a single lens. Why do you think this is impossible?