After Darwin’s Black Box:
Where Does a Laboratory Scientist Go to Contribute?

Ralph Seelke

Presented to the "Christian Scholarship Conference,"
Ohio State University October 22, 1999
Columbus, Ohio

Darwin’s Black Box (Behe, 1996) has made a strong and controversial statement about the inability of mutation and selection to produce irreducible complexity. A strength of Behe’s argument has been the depth of our knowledge about molecular machines. Numerous experimental scientists, working on such processes as motility, secretion, and the immune response, have unwittingly contributed to the argument for design. Are there areas of experimental science that need the deliberate contribution of those interested in Intelligent Design (ID)? Are those areas also of interest to those working within a neo–Darwinian framework? I believe that the answer to both questions is yes. My paper will address three questions that are important to those interested in ID and to those interested in understanding the world from a neo–Darwinian viewpoint. 1) What is the limit of mutation and selection to produce complexity? There are a number of experimental systems using microorganisms that would allow the researcher to see how far mutation and selection can go in producing complexity. This paper will discuss some of those experimental systems. 2) Can rearrangement of developmental genes produce major, advantageous changes in organisms? These highly conserved genes produce dramatically different effects in various phyla. There has been speculation that modification and/or rearrangement of these genes could produce major changes in body plans. This paper will discuss some of the strategies for accomplishing this type of sophisticated mutagenesis, and some of the expected results. 3) What are the genetic differences between organisms that have an unknown "common ancestor"? What would be needed to change one into another? We are approaching a time when we will have the entire DNA sequence for a variety of organisms. My paper will discuss how DNA sequence analysis will allow us to produce a more precise understanding of how organisms with different body plans differ at the DNA level, and the experimental uses that can be made of that understanding. Throughout my paper I will be emphasizing that, by working on areas that are important to evolutionary theory, experimental scientists can contribute to a more accurate understanding of the capabilities and limitations of evolutionary processes.

Darwin's Black Box (Behe, 1996) has invigorated the debate about the neo–Darwinian explanation for the origin of the diversity and complexity of life (a recent internet search revealed over 1 million matches for his name). Behe introduced the concept of irreducible complexity, and showed that irreducible complexity was pervasive at the molecular level. Behe has made the mousetrap famous as an explanation of irreducible complexity. Like the mousetrap, many molecular machines go from no function to complete functioning when all the parts are assembled, and gradual mutation–selection schemes are not up to the task of producing such complexity. He also showed how few molecular explanations were present in the literature for the evolution of complex systems such as sight, cellular motion, and the generation of antibody diversity.

One of the strengths of Behe's arguments is the wealth of our knowledge of molecular machines; neo–Darwinian explanations are suspect not because of gaps in our knowledge, but instead because of the way molecular biology has filled in those gaps. Because of what we know, intelligent design (ID) becomes more and more plausible.

Our understanding of molecular machines has come from men and women who, for the most part, accept the standard neo–Darwinian explanation. However, this explanation is not critical for them, since they are typically asking "how does it work?" questions rather than "where did it come from?" questions. However, by providing insight into the astonishing complexity of life, many researchers have provided unwitting support for design.

My paper is addressed to those engaged in experimental molecular biology, and who also find themselves favoring ID. Does intelligent design lend itself to a different set of research questions for experimental scientists? Are those questions also important to those favoring a neo–Darwinian explanation? I believe that the answer to both of these questions is "yes". In this paper I would like to explore some of the questions that arise from ID theory that should be of interest to experimentalists favoring either ID or Darwinian explanations.

What is the limit of mutation and selection to produce complexity?

ID predicts that mutation and selection are not capable of producing irreducible complexity. Yet what is the limit of what it is capable of producing? Answering this question will require mutation of literally trillions of organisms, and selection schemes that would allow us to detect the few mutants that would have new characteristics. With most organisms, this is not possible. However, bacteria and yeast offer the possibility of such experiments.

There are numerous bacteria and yeast that can produce a new copy of themselves in an hour or less. The time required for a microbe to reproduce itself is called it’s generation time, and a cell generation is considered one organism reproducing itself, thus making two organisms (Weaver and Hedrick, 1995). In microbial terms, a generation is usually the doubling of a bacterial population. Thus, a generation could consist of billions of cell generations. A thousand bacteria can produce a billion bacteria within a day; this would be considered 20 generations, and the equivalent of almost a billion (1 billion –1000, or 999,999,000) cell generations. This amount of bacteria, along with its growth medium, can easily fit into a milliliter of fluid. If these same thousand bacteria are allowed to grow in a liter of medium then a trillion cell generations can be produced in two to three days. Alternately, several billion cell generations can be obtained each day if the cells are grown and a portion of them transferred to new growth medium daily. A few days of unrestricted microbial growth can produce as many organisms as a more complex organism might produce in thousands or even millions of years. For example, consider a population of small rodents. A stable population of 1 million mice could easily require 100 million new mice per year in order to maintain its population. That population would thus require 10,000 years to produce as many mice as would be produced in two days by a 1000 bacteria allowed unrestricted growth. Trillions of cell generations per hour are not inconceivable using more sophisticated culture techniques.

Such techniques are being currently used by a few researchers to examine microevolution in bacteria (Appenzeller, 1999). Richard Lenski has conducted experiments of this type using Escherichia coli, producing over 12,000 generations of bacteria, and trillions of cell generations. In all of these experiments, the bacteria are subjected to selective pressure. So far only the small changes expected from microevolution have been sought and found. To my knowledge, none of these researchers have sought to produce new, irreducibly complex characteristics in these bacteria. However, this type of system would lend itself to the search for such characteristics. Below is a list of potential characteristics that, if they arose, would require the production of irreducibly complex molecular machines:

Acquisition of any of these characteristics would provide a significant selective advantage; mutants acquiring these characteristics would also be easily identified. ID predicts that no amount of blind mutation and selection will produce irreducible complexity. Darwinian explanations would predict that simple, primitive adaptations would eventually arise that would have selective advantage. Carefully performed experiments could determine if such primitive adaptations do arise and if not, give an indication of the number of generations that pass without their development. Parallel experiments, selecting for traits that are NOT irreducibly complex, would provide valuable comparisons between those changes that can be produced by mutation and selection and those that cannot. While not providing proof, such experiments would provide estimates on the limits of mutation and selection. Again, both proponents of evolution and ID would be keenly interested in the results of these types of experiments.

Can rearrangement of developmental genes produce major, advantageous changes in organisms?

One of the major breakthroughs in developmental biology has been the discovery of developmental genes such as the hox genes. These genes activate developmental pathways, causing hundreds or thousands of other genes to be activated. When defective, they produce bizarre, usually lethal mutations: fruit fly larvae that have no heads or no tails, or reduced numbers of body segments, or legs that grow from the heads of flies where antennae should grow. These genes are found in all multicellular organisms. Interestingly they produce vastly different effects in different organisms. The same gene that acts to control spine development in sea urchin will be involved in leg development in insects. The conservation of these genes is unexpected; classical Darwinian theory would have divergence in structure and function correspond to divergences at the molecular level, not the same parts used in different ways. A controversial evolutionary interpretation of developmental genes (Schwartz, 1999) proposes that, by changing the "downstream" genes affected by developmental control genes, major changes in body plans can result. Recently, a similar proposal was made by Knoll and Carroll (1999). They have developed a proposal for the role of developmental genes in the evolution of arthropods (such as insects and crustaceans) from a hypothetical ancestor for all animals with bilateral symmetry. They proposed that in the evolution of arthropods, certain key developmental genes were duplicated, modified, and relocated. The ancestor had seven developmental genes; two duplications and modifications then result in the nine developmental genes present in arthropods. Other changes in developmental genes are invoked to explain the further evolution an ancestor arthropod into insects, crustacea, and other arthropod classes. These proposals provide testable models for evolution. They invite attempts at experimental evolution in which developmental genes are added or removed from organisms. The proposals of Knoll and Carroll suggest that these manipulations would result in dramatic changes in an organism, changes that would mimic key evolutionary steps. Of course, these types of developmental mutations, resulting in major changes in an organism, have never been observed. In fact, Jonathan Wells (1999) has proposed that development requires such rigid adherence to each step along the way that disruptions in developmental plan genes are invariably deleterious. However, our knowledge of development had reached the point where the role of developmental genes in producing evolution can be tested. Sophisticated mutagenesis experiments involving developmental genes would be a key component of such testing. With currently available technology, developmental genes can be altered singly, or in various combinations. Blocks of developmental genes can be moved to different locations; and new mixtures of developmental genes can be produced. ID theory would predict negative results in terms of generation of drastic changes in body plan. However, the phenotypes of the resulting mutants would be likely to generate a wealth of information about how these genes behave normally during development. They would also provide an understanding of the limits of random rearrangement in producing basic changes in body plans. These experiments would both add to our knowledge of development, and add experimental limits to the speculations that can be entertained about the power of developmental genes to be agents of macroevolution. Again, they are the types of experiments that proponents of both ID and Darwinian evolution would support.

What are the genetic differences between organisms that have an unknown "common ancestor"? What would be needed to change one into another?

We are obtaining the complete nucleotide sequence for the genomes of more and more organisms. This information will allow us to analyze and experiment in unprecedented ways. Darwinian theory holds that major phyla, such as annelids (worms) and arthropods, must have had a common ancestor. We will be in a position to determine, at the genetic level, just exactly how a primitive arthropod differs from, say, a primitive annelid. This analysis will go beyond the level of sequence comparisons for similar genes such as are now commonly done. It will include the relationships of genes to each other (i.e., the "neighborhood" for genes or groups of genes), new genes produced, and possibly information on when important genes are expressed in development. Analysis of this type has already be done with bacteria. A recent article used nucleotide sequence information to compare Hemophilus influenzae with Escherichia coli (Tamames et al., 1997). H. influenzae has only 40% of the DNA of E. coli. However, these organisms share a number of similarities. These authors found similar genes clusters in both, and they were able to identify a number of the genes present in E. coli but missing in H. influenzae. When applied to more complex organisms, analyses will allow us to define with much more precision the genetic differences between phyla. It will also mean we will have some grasp of the genome of a putative common ancestor, and the number and types of changes needed to turn an annelid into an arthropod, or turn both in to a common, primitive ancestor. This type of research can and should be pursued by proponents of ID. By illuminating the complexity of even simple organisms, and the genetic differences between major phyla, such research will highlight the difficulty of achieving these ends by mutation and selection.


Much has been written about whether ID can result in a viable research program (see, for example, Moreland, 1994). I believe the time has come for ID proponents to be actively contributing to important research areas. The examples I have given are meant to stimulate thinking about areas of research for ID proponents; they are by no means exhaustive. The ID interpretation of the results of this type of research will clearly be different from they typical neo–Darwinian explanation. But in time, the weight of the evidence would make the design inference more and more attractive.

Literature Cited

Appenzeller, Tim. 1999. Test tube evolution catches time in a bottle. Science 224:2108–2110

Behe, Michael. 1996. Darwin’s Black Box. Free Press.

Knoll, A, and S. Carroll. 1999. Early animal evolution: emerging views from comparative biology and geology. Science 224:2129–2137.

Moreland, J.P. 1994. "Theistic Science and Methodological Naturalism" in The Creation Hypothesis, Chapter 1, pp 41–66. Inter–Varsity Press.

Schwartz, Jeffrey H. 1999. Sudden Origins: Fossils, Genes, and the Emergence of Species. John Wiley & Sons.

Tamames, J., G. Casari, C. Ouzounix, A. Valencia. 1997. Conserved Clusters of Functionally Related Genes in Two Bacterial Genomes. Journal of Molecular Evolution 44:66–73.

Weaver, R., and P. Hedrick. 1995. Basic Genetics, 2nd Edition, (William C. Brown) p. 279

Wells, Jonathan, and P. Nelson. 1999. Re–conceptualizing evolutionary developmental biology. Poster presented at The Developmental Basis of Evolutionary Change, 5/13/99 – 5/15/99, The University of Chicago


Ralph Seelke is a faculty member of the Dept. of Biology, University of Wisconsin Superior.