Introduction

Molecular evolution is a scientific field where two distinct research traditions have come together. If we take a long-term perspective, it brings the eighteenth century organismic and mechanistic approaches to the study of living beings into a common set of explanatory models and representations of life’s diversity, but we do not need to go that far. For historians of science, the study of molecular evolution is a prime example of the revolutionary transformation of the life sciences along the twentieth century, that intensified after World War II thanks to increasing governmental funds for scientific research, the introduction of molecular technologies and techniques in the study of living phenomena and disease, and reliance on computer-based statistical analysis of large collections of sequencing data.

For at least two decades after the mid-1960s, molecular and organismic evolutionists faced a number of important scientific controversies that reflected the findings, theories, and scientific values of two distinct traditions. Most of these controversies have given way to a complex interplay between the two approaches, leading to new ways to think of evolutionary processes and patterns. Moreover, the creation of large protein and DNA databases, illustrated by the European Molecular Biology Laboratory database (funded in 1980), GenBank (created in 1982), as well as the DNA Databank of Japan (1986), and the development of computer science and networks, has revolutionized the practice of research on the molecular basis of evolution.Footnote 1

The history of this field has provided historians with insights into the mechanisms of discipline formation, the relevance of debates—and even of different scientific values—in the advancement of science, and the deep connection between areas of research relevant to what eventually became Molecular Evolution, and the pressing concerns that characterized postwar scientific inquiries.

Early Attempts

Since its beginnings in the first years of the 20th century, the use of molecular data was driven by a desire to eliminate subjective aspects of traditional taxonomy, where morphological and other organismic data were weighed according to personal judgment and authority-based claims on a given biological group (Hagen 1999, 2001, 2003; Suárez-Díaz and Anaya-Muñoz 2008; Strasser 2010a). The story of how the American–British physician George H. F. Nuttall (1962–1937) first developed the precipitin reaction to measure the proportion of immunological response between different species has been told before (Strasser 2010a; Suárez-Díaz 2014a). Nuttall had made several contributions to the study of malaria, immunology, and blood chemistry at the Molteno Institute Biology and Parasitology at Cambridge University, where he developed the precipitin reaction that measured the strength of the reaction between blood sera of different animal species.Footnote 2 Together with his two students, he tested thousands of samples from groups as diverse as the arthropods and mammals, and published his results on the relationships between species early on (Nuttal 1904). In 1901, Karl Landsteiner had discovered the first blood groups (A and B), and a decade later Ludwig Herszfeld and E. von Dungern demonstrated that they follow Mendelian inheritance. From that moment onward, immunological and serological data were taken as an indication of the amount of genetic relation between species, and thus as a reliable source to establish evolutionary relations. However, in the first part of the twentieth century, these data were seen only as complementary to the work of anatomists, taxonomists, and paleontologists, and in no way providing an independent source of knowledge about the history of living beings. This situation changed six decades later, as the early molecular evolutionists—properly speaking—struggled with organismic biologists and claimed that molecular data was preferable (and superior) to morphological traits (Zuckerkandl and Pauling 1965a; Suárez-Díaz 2007).

Inter-specific relations were not the only focus of attention for those working with serological data in the first part of the twentieth century. Until the mid-1950s, immunological and serological studies remained the only available source for data on the genetic variation within human populations. In the 1930s, data on blood groups frequencies were incorporated in the first mathematical analysis of population genetics, figuring prominently in the debate between the medical statistician Lancelot Hogben (1895–1975), who fiercely opposed eugenics, and population geneticist Ronald A. Fisher (1890–1962), one of the main proponents of this doctrine in the United Kingdom (Mazumdar 1992). Notwithstanding Hogben’s arguments on the interaction between inheritance and environment, as well as the difficulties in using blood groups to distinguish human “races” or “nations,” serologists claimed their data would provide an objective method for human group classification that would supersede the traits measured by physical anthropologists. As the historian and anthropologist Jonathan Marks has claimed, for a long time, serology sought to provide empirical data on human racial categories but with little success (Marks 1996, 2012).

Well after World War II, serology and studies on the demography and geographical distribution of ABO groups and serum enzymes continued to provide data on human population genetics, as exemplified in the work of Boyd (1963) on blood group frequencies, and Harris (1966) on human blood serum components. By the mid-1950s, these human population studies were transformed by the introduction of molecular techniques, such as filter-paper and gel electrophoresis, by molecular anthropologists like Morris Goodman, John Buettner-Janusch, and later Vincent M. Sarich and Allen Wilson (Aronson 2002). However, the century long relation between interest in blood components, human classification, and evolutionary genetics is noticeable.

Searching for objective methods to classify other biological species (meaning the use of measurable, quantitative traits, and explicit classification criteria) remained a powerful drive in attempts to use serological data in taxonomy. From the end of the 1920s to the beginning of the 1960s, Nuttall’s precipitin reaction was refined by Alan A. Boyden, as part of his continued efforts to develop a “serological taxonomy” at Rutgers University in New Jersey (Strasser 2010a). To this end, Boyden created a Serological Museum which, along with other molecular collections, such as Hermann Lehman’s hemoglobin collection (De Chadarevian 1998), or the human sera collection of Arthur Mourant (Bangham 2014) provides historical clues on the origins of contemporary databases. In their exhaustive quest for samples, these pioneering collections constituted the first places where “analogical-molecular data” came to bear for evolutionary-comparative purposes.

A different route to the integration of molecular data and evolutionary biology began with the development of general or comparative biochemistry in the fourth decade of the last century. Ernst Baldwin’s An introduction to Comparative Biochemistry (1937) and Marcel Florkin’s L’evolution biochemie (1944, English trans. 1949) compared the products of secondary metabolism in plants and, perhaps more consequentially, the adaptability of organisms as reflected in the evolution of biochemical substances and metabolic “systems.”Footnote 3 Again, this project was driven, as stated in Florkin’s Prologue to the English edition, “to superimpose, upon the zoological classification, established by morphologists, a biochemical classification” (1949, p. iii). The two books covered a range of metabolic systems in animals, and argued for the unity of “chemical composition” in animals (and plants) as a result of their common evolutionary history, an affirmation that not everyone shared in their time.

Equally consequential was the work of British biochemist Marjorie Stephenson (1885–1948), who carried out the first comparative studies on carbohydrate intermediate metabolism in bacteria (Stephenson 1930/1949). Like Baldwin, Stephenson’s work was tremendously influenced by the “general” approach to biochemistry of Frederick Gowland Hopkins at the Dunn Institute in Cambridge University, a reminder of the important role played by a favorable research environment. In Stephenson’s work, as in the case of A. J. Kluyver’s, comparative biochemistry became a reliable, accepted tool for the classification of bacteria, though they both shared the contemporary uneasiness regarding the evolution of this group: lacking stable morphological and biochemical traits, and sexual reproduction, microorganisms were thought to evolve under Lamarckian principles. It was not until Max Delbrück and Salvador Luria’s fluctuation test in 1943 that bacteria slowly began their way into the realm of evolutionary biology, eventually benefitting from the DNA database revolution of the last two decades.

History, like evolution, does not follow a single path. It is a multidirectional trajectory redirected by contingent events and favored trends. Comparative biochemistry, for all that was worth, was in large part set aside by the early proponents of the field that would become molecular evolution.

Protein Chemistry and Biomedicine After World War II

As we know it today, the study of evolution at the molecular level owes its modern face to the rise of protein chemistry, biomedicine, and molecular biology after the Second World War. Fueled by large investments in new technologies, the mid-1950s saw the invention of gel-zone electrophoresis, further developments in crystallography and ultracentrifugation, and dropping prices and access to radioisotopes and scintillation counters, among other technologies that transformed the biological laboratory. The first major work in this direction, Christian Anfinsen’s The molecular basis of evolution (1959) was written in the aftermath of the elucidation of the DNA structure by Watson and Crick. In his argument for a new molecular approach to evolution, Anfinsen relied on his own work on protein structure and function, firmly rejecting Florkin’s biochemical comparative approach, and advocating for an exclusive emphasis on proteins and nucleic acids (informational molecules, or semantides, as Emilé Zuckerkandl called them a few years later) (Zuckerkandl and Pauling 1965a). Although the genetic code had not been elucidated at the time Anfinsen wrote his book (it was a few years later), recent advances in molecular genetics and biomedicine opened new venues for him and others to explore.

The situation changed a lot after the war. In 1943, Hopkins retired as head of the Dunn Institute at Cambridge and was replaced by Charles Chibnall, who instead of the general comparative approach to biochemistry favored a focus on proteins and their amino acid structure (De Chadarevian 1996, 2002). Among the Dunn PhD students, was Frederick Sanger, who found a supportive environment to his development of chemical methods to establish the amino acid sequence (a term he coined) of the two chains of insulin in 1951 and 1952. He moved to the Molecular Biology Laboratory at Cambridge, under the leadership of Max Perutz, in 1962.Footnote 4 It was there where Sanger later developed the first methods for DNA sequencing (Garcia-Sancho 2012).

Sanger’s structural attack to proteins became a powerful tool in the hands of evolutionary inclined biochemists. On the other side of the Atlantic, Anfinsen’s research provided the first clues as to the complex evolutionary relationship between protein structure and function (Creager 2007). Working with the enzyme bovine ribonuclease, he showed that the tertiary structure of proteins—and its function—depended, ultimately, on their primary structure. The sequence of amino acids in a given protein chain, which he thought must be related to the primary structure of the DNA molecule, provided crucial information as to the function—and lack thereof— and the evolutionary relations between proteins. At the same time, however, Anfinsen noticed that significant levels of structural variation (including deletion of several residues depending on each protein) could take place without “violating” its primary function (Anfinsen 1959, 1972). Based on his own research on ribonuclease, and on the large variety of hemoglobin anomalies that medical geneticists and biochemists had accumulated, he concluded that structure was tantamount, but not all amino acids carried the same weight in determining the function of a protein (Anfinsen 1959, 1989). It was one of the first formulations of the idea that functional and structural constraints were at the basis of evolution at the molecular level.

Extensive research on hemoglobin structural anomalies reflected not only broader social concerns, but also the intersection of medicine and basic biological research after the war. Concerned by the effects of radiation due to atomic testing and fallout, many of the most influential scientists of the time focused their attention on the genetic effects of radiation on human populations. Public health care programs (like the malaria eradication campaigns) were added to the interest in genetic variation around the globe (Suárez-Díaz 2014b). As illustrated in the work of Linus Pauling, no better molecule existed to link the study of variation and disease, than hemoglobin.

In 1949, Linus Pauling, Harvey Itano, John Singer, and Ibert Wells, using the so-called Tiselius apparatus (bound electrophoresis), had concluded that the difference between normal (Hb), and sickle cell hemoglobin (HbS) resided in its electrical charge. They inferred that the cause of the disease was a molecular alteration of the hemoglobin protein, for which Pauling coined the term “molecular disease.”Footnote 5 A few years later, at the Cavendish Laboratory in Cambridge, Vernon Ingram established that a single residue of glutamic acid had been substituted by a valine in HbS (Ingram 1956), using a technique that combined zone electrophoresis and paper chromatography, which he named protein fingerprinting. News of Ingram’s discovery arrived at Caltech, where Pauling convinced the recently recruited Austrian biochemist Emilé Zuckerkandl to redirect his career to focus on the evolution of hemoglobins. Protein fingerprinting was thus adopted for a first comparative study of primate hemoglobins carried out by Zuckerkandl and Pauling’s graduate student Richard T. Jones (Zuckerkandl et al. 1960).

A word, however, needs to be said as to why Pauling was interested in the comparative analysis of hemoglobins. The connection between molecular disease and evolution grew from Pauling’s involvement in the nuclear testing debate during the 1950s. As the controversy on the genetic implications of atomic fallout rose after the March 1, 1954 Bravo Test of an hydrogen atomic bomb at the Bikini Atoll of the Marshall islands, Pauling became more involved in pacifist activism, and increasingly interested on advances in mutation, evolution, and genetics research. The historian of science Gregor Morgan writes that “[b]y the time Zuckerkandl arrived [to Caltech] from France, in September 1959, Pauling was well versed in evolutionary theory and genetics” (Morgan 1998, pp. 159–160).Footnote 6

Complete amino acid sequences were laborious and time consuming to obtain at the time, but the limitations of protein fingerprinting convinced Zuckerkandl to shift to protein sequences analysis. In 1961, he and Walter A. Schroeder published a paper on the determination of the amino acid structure of gorilla hemoglobin (Zuckerkandl and Schroeder 1961). Their results showed that, regarding their amino acid composition, the α- and β-chains of gorilla hemoglobin were almost undistinguishable from the corresponding chains of human hemoglobin. The chance to explore the full implications of protein structural comparative analysis (which Zuckerkandl dubbed “chemical paleogenetics”) came with a paper he co-authored with Pauling (1962), as part of a Festschrift in honor of Albert Szent-Györyi. Tellingly, the paper’s aim was not molecular evolution per se, but the connection between disease and evolution at the molecular level.Footnote 7

In that paper, Zuckerkandl and Pauling advanced the idea that the α- and β-chains of human hemoglobin were homologous (something that had been hypothesized by Itano in 1957 and by Ingram in 1961), and for the first time they postulated the concept of the molecular clock (without using the term).Footnote 8 They again emphasized the close relationship between man and gorilla: “since gorillas get along well with their hemoglobin, as they prove by existing, it is not likely that the gorilla β-chain, if it were present in human, would cause molecular disease…thus, if the gorilla β-chain occurred in a human family the physician’s attention would probably not be attracted to it” (Zuckerkandl and Pauling 1962, p. 200). By bringing together the study of human variation and inter-specific variation, the new technical procedures allowed a whole new conceptualization of the possibilities of the molecular approach.

The new protein sequence analysis approach was carried on by others, notably by biochemist Emanuel Margoliash, then at Abbot Laboratories in Illinois. In contrast to the biomedical origins of hemoglobin research, Margoliash concentrated his efforts on an evolutionary relevant molecule, cytochrome C. Reflecting the optimism of early molecular evolutionists, he aimed to establish the whole lineage of life using this single molecule (Margoliash 1963). His comparative analysis of amino acid sequences led to results similar to Anfinsen’s and Zuckerkandl and Pauling’s: the recognition of variable regions in the protein that did not affect its function, while simultaneously revealing a high degree of conservation in some of the amino acid residues. This kind of analysis was complemented by the crystallographic analysis of cytochrome C`s tertiary structure, carried out by Richard Dickerson at Cambridge. In sum, protein chemistry led to the establishment of hypotheses concerning the function of conserved residues, and the forces or constraints retaining them in their place (Buettner-Janusch 1962; Zuckerkandl and Pauling 1962, 1965b).

In principle, the analysis of protein sequences could provide a quantitative method to assess similarities and differences between homologous proteins. By the end the 1960s, sequencing a complete protein was still a very slow and also a costly process; sequences were so scarce that they could only provide the basis for structural comparative analyses, but nothing more. Only when the increased efficiency, and later the automation of sequencing, produced enough data, a sophisticated statistical quantitative analysis was possible. This was made possible, thanks to developments in chromatography and electrophoresis, and later by the Edman degradation reaction (Edman and Geoffrey 1967; Garcia-Sancho 2010, 2012). At the end of the decade, Pehr Edman built a fully automated sequencing machine (the sequenator) that used the degradation reaction he had developed some years before.Footnote 9 Almost simultaneously, Stanford Moore and William H. Stein, of the Rockefeller Institute in New York, sequenced the 174 amino acids of ribonuclease in half the time it had taken Sanger to sequence the insulin molecule, for which they were awarded the Nobel Prize in 1972 (they shared it with Anfinsen).

More Than One Way

The study of the molecular basis of evolution was not restricted to advances in protein chemistry. At the end of the 1950s, and well into the following decade, at least two other experimental methods competed with sequence analysis in that realm: gel-zone electrophoresis and nucleic acid hybridization.

The introduction of zone electrophoresis provided a cheap, easy, portable, and reliable method to separate biological molecules according to its molecular weight and electrical charge. In 1955, Oliver Smithies published a method which included starch gel as support material, and four years later Leonard Ornstein and Baruch Davies of Mount Sinai Hospital in New York proposed the use of a synthetic gel of polyacrylamide, with the advantage of the adjustability of the pore size of a synthetic gel (Chiang 2009, p. 513). In contrast to its predecessor, developed by Arne Tiselius, zone electrophoresis separated molecular species in a solution, but the practical advantages of the new methods were equally important. Gel electrophoresis could be carried out in almost every laboratory in the world, and in contrast to the Tiselius apparatus, it was not an expensive technology, nor required costly facilities.

Electrophoresis was foremost used to separate human blood proteins and was rapidly associated with immunological techniques in the study of Primate evolution, as in the work of Morris Goodman (Goodman 1962, 1964; Hagen 2010). It provided a relatively easy way to assess genetic variation in human and animal populations (Powell 1994; Dietrich 2006). By the mid 1960s, an army of medical geneticists and physical anthropologists alike were sampling all types of human populations around the world (Suárez-Díaz 2014a; De Souza and Santos 2014), and electrophoresis was used to quantify relations between species and within populations. It is the latter type of research, particularly the studies on the genetic variation of Drosophila pseudoobscura, by Hubby and Lewontin (1966), which frequently gets cited as given way to the neutral theory of molecular evolution by Motoo Kimura (1968; see Lewontin 1974). Despite its enormous impact on population genetics, by now it may be clear that theorizing evolution at the molecular level required much more than the contributions of population geneticists (Dietrich 1994).

Besides the many contributions of protein chemistry to the idea that molecular evolution might differ from processes taking place at the organismic level (see above), advances in nucleic acid hybridization revealed the complexity of the eukaryotic genome, and the limitations of the pan-adaptationist view that was so prevalent in those years. Nucleic acid hybridization originated in research on the physicochemical properties of DNA, by Julius Marmur and Paul Doty at the Department of Chemistry at Harvard University (Marmur and Doty 1959). In 1960, they reported that DNA in solution lost its double helix structure when heated, and then recovered it when the solution was slowly cooled (Doty et al. 1960). Very soon, Ellis T. Bolton and Roy J. Britten, from the Laboratory of Terrestrial Magnetism at the Carnegie Institution of Washington, started to use hybridization to quantify the degree of affinity between DNA from two different species, as measured by their proportion of hybridization. The quantification of the hybridization reaction was interpreted as a direct measurement of similarity between two DNA species (Suárez 2001). Here was a quantitative method that offered a direct overall measure of the “degree of affinity” between two different genomes. In these features relied the specific difference of hybridization techniques and, as claimed by its supporters, its main advantage. However, in terms of molecular evolution, its most important outcome was the characterization of repetitive sequences in the eukaryotic genome (Britten and Kohne 1968), a biological phenomenon that soon was incorporated in the non-darwinian theory of evolution of King and Jukes (1969), as well as in the idea of junk DNA (Ohno 1972).

As with other techniques, hybridization required calibration with paleontological data and, often, the assumption of the molecular clock. However, the method became plagued by more serious problems, like the fact that it did not yield reliable data in the case of plants, and it was difficult to assign statistical relevance to data from different experimental settings (Felsenstein 1987; Templeton 1988). Moreover, hybridization became embroiled in debates concerning the evolutionary relation between human, gorillas, and chimpanzees (Sibley and Ahlquist 1984; Sibley et al. 1990; Schmid and Marks 1990; Suárez-Díaz 2014a).

The Rise of Statistical Phylogenetics

By the time hybridization was practically abandoned, advances in protein and DNA sequencing, in combination with the introduction of computers and the mathematization in biological research, had changed the face of comparative biology. Computers had become more efficient and less expensive, and according to historian of biology Joel Hagen, in the early 1960s, “15% of colleges and universities in the United States had at least one computer on campus, and many major research universities were purchasing so-called ‘second generation’ computers based on transistors rather than vacuum tubes” (cited in Hagen 2001, p. 293). Some of its early applications in biology took place in interaction with developments in (organismic) systematics. The field was entering a new era, with pheneticists introducing numerical criteria in taxonomy (Sneath and Sokal 1973; Hagen 2003) organismic systematists introducing parsimony (cladism), and all of them contesting the—until then—dominant view of evolutionary systematics, in what has become to be known as “the systematics wars” (Hull 1988; Felsenstein 2001). The mathematization of phylogenetics was heir to previous attempts (by serologists, as we saw before) to develop objective explicit criteria in classification and phylogenetic analysis. This approach required statistical analysis, made possible by increased availability of larger numbers of molecular data.

By the end of the 1960s, the rapid accumulation of partial protein sequences, accelerated by advances in automation, led to the first collection by Margaret O. Dayhoff and her colleagues, published as the Atlas of Protein Sequence and Structure (1965 1st edn.) as a result of her research at the National Biomedical Research Foundation. Dayhoff’s pioneering collection of protein sequences constituted an important precedent to the creation of the first databases (Smith 1990; Strasser 2010b, 2011). She not only collected the sequence data published every year, but developed computational tools for its mathematical analyses, including parsimony and substitution matrices to asses the probabilities of substitution between amino acids (Dayhoff and Schwartz 1978, see below). In her work (Dayhoff 1969; 2010b), and in the collaboration of Walter Fitch and Emanuel Margoliash (Fitch and Margoliash 1967, 1968), computers were first introduced for the comparative analysis of amino acid sequences. Using Margoliash’s data, consisting of the 20 known amino acid sequences of cytochrome C, Fitch developed an algorithm to construct a molecular phylogeny based on a measure of similarity that he defined as minimum (mutational) distance between two molecules. Other algorithms were being applied at the time to other types of molecular data.Footnote 10 Dayhoff’s and Fitch’s use of computers was possible by the introduction of a first generation of these technologies in university laboratories (mostly mainframe IBMs of the 360 Series, and the common DEC PDP-8). This was part of a broader project launched by the National Science Foundation and the National Institutes of Health.Footnote 11 The project gained impetus at the intersection of important developments in computer technology that took place within giant companies, such as IBM and DEC in the United States.

The mathematization of systematics, however, remains as one of the most contentious arenas in the history of 20th century biology, one that intersected with the rise in the use of molecular over morphological data (Felsenstein 2001; Hagen 2003). Extending from the early 1960s to the late 1980s, the clash between a predominantly organismic cladism, and those supporting the need for numerical methods in phylogenetic inference, resulted first in the dismissal of classic evolutionary systematics. Meanwhile, ever more sophisticated algorithms to organize and compare data were developed (for instance, Needleman and Wunsch 1970). By the end of the 1980s and the early 1990s, a more pragmatic attitude prevailed, increasingly leaving behind philosophical commitments to parsimony, and enhancing the use of statistical criteria to discern between ancestral and derived characters, as well as many other decisions concerning the construction of evolutionary trees (Fitch 2000). Statistical tools such as maximum likelihood, the bootstrap, parametric bootstrapping, the KHT test, and many others, were introduced into software packages, such as PAUP, MacClade, and PHYLIP (Felsenstein 2001; Suárez-Díaz and Anaya-Muñoz 2008; Suárez-Díaz 2010).

The development of algorithms and the mathematization of comparative analysis of sequences also intersected with a rapid transformation in the way data collections were managed at the end of the 1980s. Databases were targeted first as services, and then as crucial research resources, and managed via the intervention of major research institutions and governmental agencies. The rationale behind was not evolutionary biology, but once again biomedical research; still, as some of us have claimed, comparative evolutionary also played a role in the emergence of genomics in the years to come (Suárez-Díaz 2010; Garcia-Sancho 2012). The creation of large protein and DNA databases, such the European Molecular Biology Laboratory database (funded in 1980, see Garcia-Sancho 2012, Chap. 4) and GenBank (created in 1982 by initiative of the NIH in conjunction with Los Alamos National Laboratory and the private firm Bolt, Berenek and Newman, see Temple 1990, Strasser 2011), opened the possibilities for new ways to practicing evolutionary research. They represent the transition from the “analogue” molecular collections of Boyden, Lehmann, and others into the digital format (Suárez-Díaz 2014a), and illustrate the integration of the two longstanding traditions implicit in molecular evolution: natural history and experimental biology (Strasser 2011).

Time to Debate

Although many organismic and molecular evolutionists had favored the view that both approaches provided complementary insights into biological evolution, by the mid-1960s, there were enough differences between the two fields to provoke a number of interesting scientific debates. Instead of a conciliatory tone, molecular evolutionists now defended the superiority of molecular over morphological data in the construction of phylogenetic trees, and for the elucidation of evolutionary mechanisms (Margoliash 1963; Zuckerkandl and Pauling 1965a), they also hypothesized on the constancy of substitution rates in proteins (the molecular evolutionary clock) and the possibility that mechanisms other than natural selection could be prevalent at the molecular level (Kimura 1968, 1969; King and Jukes 1969; Crow 2008). Their research had led to the establishment of unexpected phenomena, such as the presence of highly repetitive sequences in the eukaryotic genome, and the realization that some—or many—amino acid residues could be substituted in a given protein without “violating” its function. The molecular clock not only meant a tool to calculate divergence times, as when Sarich and Wilson (1967, 1969) used it to lower the time of divergence between man and large primates to 4–5 million years), but rapidly became a main argument in the dispute between selectionism and neutralism, and between defendants of the evolutionary synthesis, and the new molecular evolutionists.

However, it would be too simplistic to account for the organismic biologists’ reaction as a mere refusal to scientific change. For the paleontologist George G. Simpson, biology was a profoundly historical science, and the idea of a constant rate of evolution oversimplified the particularities of each species, and the contingencies each faced in evolutionary time. Nothing would be more upsetting to Simpson than what he perceived as the molecular evolutionists’ arrogance, regarding the relation between the primates and men, which obviated the anatomical and behavioral differences between the two species. For Simpson, and the scientific tradition he represented, the more morphological, behavioral, and paleontological data one could accumulate, the better it was for a proper classification (Aronson 2002; Suárez-Díaz 2007; Simpson 1964). On the other side, it was clear from Simpson’s arguments that human evolution and primate taxonomy was a highly sensitive field, and those molecular evolutionists, in their quest for quantitative classification criteria, did not share the values of traditional taxonomists (Suárez-Díaz and Anaya-Muñoz 2008).

Ernst Mayr and Theodosius Dobzhansky’s defense of organismic biology and natural selection was more professionally inclined than Simpson’s. The philosopher and historian of biology John Beatty has forcefully argued that, for the architects of the evolutionary synthesis, the advent of molecular biology and evolution threatened the already scarce financial support, the number of bright students, and of university jobs, that went into the fields traditionally linked to natural history and evolution They translated their professional concerns into a conceptual defense of the complementary nature of organismic and molecular approaches, including the well-known distinction, by Mayr, of ultimate and proximate explanations (Beatty 1990, 1994). However, while Dobzhansky was more open to the new molecular biology, Mayr and Simpson felt a major threat to their views on the nature of biological research.Footnote 12

However, in the decades of 1970 and 1980, the neutral theory was challenged on many fronts. Not all tests supported the neutral theory. Francisco Ayala and his colleagues, for instance, used electrophoretic data on the genetic variability of Drosophila populations, and noticed that the frequency distribution of heterogenous loci did not support the distribution predicted by the neutral theory. Neutralists responded with several refinements of the original theory, in particular Ohta (1992), who developed the theory of nearly neutral alleles, and Masatoshi Nei, who theorized that population dynamics could explain some phenomena complicating Kimura’s neutral theory (Crow 2008; Dietrich and Suárez-Díaz 2016).

The molecular clock also remained controversial. From a tool to calculate divergence times, it had become the central evidence in the neutralist–selectionist debate. The struggle to define how much variability was reasonable to expect from the molecular clock remained polemic. Some argued that the variability of the clock was the result of a few mutations favored by natural selections (Kimura and Ohta 1971); others (like Goodman 1983) argued that a clock that changed its rate was not a clock anymore. In the following decades, as DNA sequences and statistical tests became readily available, the debate started to depolarize. In the work of Kreitman (1983), a student of Richard Lewontin, and his students and statistical tests were devised that compared synonymous and non-synonymous mutations in DNA, within a species and between two species (Hudson et al. 1987). Monte Carlo tests, in particular, were applied to molecular and population genetics data, to discriminate between alternative hypotheses on the mechanisms of evolution at the molecular level (Dietrich 1996). These tests, and others, eventually led to the common contemporary view that the neutral theory, as a useful null hypothesis, and as a reliable model of substitutions, is an integral part of the understanding of molecular evolution (Kreitman 1996, 2000; Dietrich 2006).

Still, the same DNA databases and statistical tests that have facilitated the dilution of the neutralist–selectionist debate have provided the means for new and enriching debates on the patterns of the early history of life. Molecular evolutionists debate if a tree is still an adequate representation of life’s history (Doolittle 1999, 2005; Sapp 2005; O’Malley and Boucher 2005; O’Malley et al. 2010), historians of science observe, with fascination, the uncertainties and the complex development of a new scientific field.