Eye Evolution: A Closer Look
In a previous article I described how theories of innovation provide insight into the limits of natural selection. I will now apply those concepts to hypotheses regarding the evolution of the vertebrate eye, a subject that, since the time of Charles Darwin, has been near center of the debate over the creative power of natural selection. As Darwin himself stated in the Origin of Species:
To suppose that the eye with all its inimitable contrivances for adjusting the focus to different distances, for admitting different amounts of light, and for the correction of spherical and chromatic aberration, could have been formed by natural selection, seems, I freely confess, absurd in the highest degree.
He did, however, still believe it could evolve over numerous gradual increments.
Today, evolutionists propose several of the stages in what they believe to be a plausible evolutionary path. Science writer Carl Zimmer has outlined the standard story:
In 2007, Trevor Lamb and his colleagues at Australian National University synthesized these studies and many others to produce a detailed hypothesis about the evolution of the vertebrate eye. The forerunners of vertebrates produced light-sensitive eyespots on their brains that were packed with photoreceptors carrying c-opsins. These light-sensitive regions ballooned out to either side of the head, and later evolved an inward folding to form a cup. Early vertebrates could then do more than merely detect light: they could get clues about where the light was coming from...a thin patch of tissue evolved on the surface of the eye. Light could pass through the patch, and crystallins were recruited into it, leading to the evolution of a lens. At first the lens probably only focused light crudely...Mutations that improved the focusing power of the lens were favored by natural selection, leading to the evolution of a spherical eye that could produce a crisp image.
See Wikipedia for a chart illustrating "Major stages in the evolution of the eye."
To add weight to this narrative, two biologists created a computer simulation, demonstrating, in their view, the incremental evolution of an eye in fewer than 400,000 generations.
This often-repeated tale sounds impressive at first, but it is not unlike most supposed explanations of the evolution of complex features. It scores high on imagination and flare but low on empirical evidence and thoughtful analysis. It most certainly does not represent a "detailed hypothesis." Likewise, the simulation does an admirable job of describing how a mechanical eye could develop incrementally, but it is completely disconnected from biological reality. In particular, it ignores the details of how a real eye functions and how it forms developmentally. When these issues are examined, the story completely collapses.
To fully appreciate why that is so requires a basic understanding of developmental biology. During development, cells divide, migrate, and differentiate into a wide variety of types. Throughout this process, the cells send chemical signals to their neighbors, and these signals cause proteins known as transcription factors (TF) to bind to genes in regulatory regions, which control the corresponding genes' activity. The TFs bind to what are called transcription factor binding sites (TFBS), and the correct binding enables the genes to produce their proteins in the right cells at the right time in the right amount.
The evolution of additional components in the vertebrate eye requires that this network of intercellular signals, TFs, TFBS, chromatin remodeling, as well as many other details be dramatically altered, so that each developmental stage can progress correctly. For instance, the seemingly simple addition of a marginally focusing lens -- that is to say, a lens that directs slightly more light onto a retina -- requires a host of alterations:
Ectodermic tissue folds into a lens placode, which then forms a lens vesicle.
Cells in the lens vesicle differentiate into lens fibers, which elongate to produce the proper lens shape.
The lens fibers then undergo several key modifications, including tightly binding together, filling almost entirely with special refractive proteins called crystallins, developing special channels to receive nutrients, and destroying their organelles.
All of these steps must proceed with great precision to ensure the end product focuses light in an improved manner. The development of the lens in all vertebrates is very similar, and it even resembles that in other phyla. Therefore, the development of the first lens should have closely followed the steps outlined above with only minor differences, inconsequential to the basic argument.
The challenge to evolution is that, short of completion, most of these changes are disadvantageous. A lens that has not fully evolved through the third step noted above would either scatter light away from the retina or completely block it. Any initial mutations would then be lost, and the process would have to start again from scratch. In the context of fitness terrains, an organism lacking a lens resides near the top of a local peak. The steps required to gain a functional lens correspond to traveling downhill, crossing a vast canyon of visually impaired or blind intermediates, until eventually climbing back up a new peak corresponding to lens-enhanced vision.
Once an organism has a functional lens, natural selection could then potentially make gradual improvements. However, moving from a reasonably functional lens to one that produces a high-resolution image is rather complex. In particular, the refractive index (i.e., crystalline concentration) has to be adjusted throughout the lens to vary according to a precise mathematical relationship. A gradual decrease from the inside to the outside is needed to prevent spherical aberrations blurring the image.
Even more steps are required for the improved image to be properly interpreted:
Feedback circuitry must be added to allow the lens to automatically refocus on images at different distances.
The neural networks in the brain have to be rewired to properly interpret the pre-processed high-resolution images from the retina.
Higher-level brain functions must be enabled to identify different objects, i.e., dangerous ones such as a shark, and properly respond to them.
Until steps 2 through 4 are completed, a high-resolution image would likely prove disadvantageous, since most of the light would be focused on fewer photoreceptors. In insolation, the alterations of perfecting the lens and those involved in step 1 would hinder the analysis of large-scale changes to the field of view, such as identifying the shadow of a predator. Natural selection would thus remove most of the initial mutations, and evolution of the eye would come to a halt.
The difference between blurry and high-resolution vision is well illustrated by the box jellyfish. It has several eyes around its body. Two have lenses, which can produce highly focused images. However, the focal point is past the retina, so the retinal images are blurry. An ability to focus more clearly than is actually useful seems to be an example of gratuitous design. Zoologist Dan Nilsson comments:
For such a minute eye it is surprising to find well-corrected, aberration-free imaging, otherwise known only from the much larger eyes of vertebrates and cephalopods. The gradient in the upper-eye lenses comes very close to the ideal solution...The sharp image falls well below the retina and it would seem that the sharp focus of the lenses is wasted by inappropriate eye geometry.
However, for the box jellyfish a high-resolution image would be disadvantageous, since its neurology is engineered to respond to such bulky features as the edge of a mangrove. Is this blurry vision the result of the jellyfish not having yet evolved high-resolution vision? No: its neural organization is radically different from that needed for the latter. As Nilsson comments, "Another, more likely, interpretation is that the eyes are 'purposely' under-focused."
"Purposeful"? Yes, it would seem so. The example illustrates that low-resolution vision is not at an inferior point on the same fitness peak as high-resolution vision. Instead, both systems reside near the peaks of separate mountains. For any species, upgrading to high-resolution vision requires massive reengineering in a single step. Such radical innovation, coordinated to achieve a distant goal, is only possible with intelligent design.