Evolution, development and complexity in Pere Alberch (1954–1998)


Andrés Moya Departament de Genètica, Dr Moliner 50. 46100 Burjassot, Valencia, Spain. Tel: 34 96 398 3180; fax: 34 96 398 3029; e-mail: andres.moya@uv.es

When Pere Alberch died on 13 May 1998, a fruitful career was truncated. Born 2 November 1954 in Barcelona, he was only 43 years old. Pere Alberch had reached a prominent place in the research on developmental biology, morphological evolution and theoretical biology. He studied biology and environmental science at Kansas University. He graduated in 1976 with the highest honours, he earned his PhD in zoology at the University of California in 1980. George Oster and David B. Wake were his co-supervisors. He lectured on biology at Harvard University from 1980 to 1989, where he was also Curator of Herpetology at the Museum of Comparative Zoology. Owing to his scientific prestige, he was also a member of the editorial or advisory boards of journals such as Trends in Ecology and Evolution (since 1993), Biodiversity Letters (since 1992), Journal of Theoretical Biology (since 1985) and Journal of Evolutionary Biology (1986–1991). In 1989, he came back to Spain, where he earned a position as research professor; he also carried out an important management role at the Museo Nacional de Ciencias Naturales (CSIC) in Madrid. When he gave up this last job, a new phase for his scientific activity began. In Valencia, the Instituto Cavanilles de Biodiversidad y Biología Evolutiva, a new research centre, became interested in incorporating Pere in its staff. He had new ideas and theories that could have been developed in this institution, and a great willingness to communicate his knowledge through advanced courses on evolution. Unfortunately, this was not possible due to his death.

We would like to remember Pere Alberch as an evolutionary biologist, his main activity. We say his main activity because Pere developed other facets during his short life, such as a love of music; he was also a keen connoisseur, collector and critic of modern art and, like D’Arcy Thompson, art gave him ideas that lead towards his important notion of developmental constraint on evolution. In order to understand his place in biological research, we have to look at the diversity of life. This shows two kinds of living entities. The first category consists of organisms displaying a variety of metabolic capabilities within a very narrow morphological range (mainly unicellular organisms), whereas the second category groups organisms with a limited genetic repertoire but displaying very different morphologies (mainly multicellular beings). Morphological evolution is mainly related to the latter category. In multicellular beings, there are genes that govern housework cell jobs and other genes that are involved in morphogenetic tasks. Morphological evolution is related to changes in developmental programmes and this was the field explored by Pere Alberch.

Pere was aware that the history of life was a relevant source of data for testing evolutionary hypothesis, palaeontology therefore playing a central role in evolutionary theory. Eldredge & Gould's punctuated equilibria (1972) involved two important features: the morphological stasis of lineages and evolutionary jumps from the ancestral to the descendant lineage. This introduced two important notions for modern evolutionary biology: stability and discontinuity; the gradualistic picture of evolution advocated by the modern synthesis was contested. On the other hand, organic change and development share a common feature: history. History at two levels: individual and biotic. During the nineteenth century, many scientists such as von Baer, Saint-Hilaire or Agassiz worked in the fields of both palaeontology and embryology. They conceived of organic change as ruled by very similar laws to developmental processes and, therefore, they tried to understand the history of life. Now there are other points of view, but interest remains in how evolution and development are related. Although this theme was not appreciated by Neodarwinians, some of them, such as De Beer in the 1940s, dealt with it; he criticised Haeckel's biogenetic law and revisited the very important concept of heterochrony. Waddington, in the 1950s, drew attention to development as a missing process in evolutionary theory. A new and important approach to the relation between development and evolution came from the palaeontologist Gould in his book Ontogeny and Phylogeny (1977). He decoupled growth, somatic development and the age of organisms through evolution. An immediate consequence of this was the well-known clock model, by which all De Beer's heterochronies were described. Moreover, Gould added two other kinds of heterochronic change: proportional gigantism and proportional dwarfism.

One of Pere Alberch's first relevant contributions was the quantitative approach of the clock model (Alberch et al., 1979) when he was 25 years old. The clock model was a qualitative description of heterochrony. Quantification of this model needs the concept of growth law for size and shape. Growth laws are formulated as differential equations y′ = f(y/α, β, k, S0), in which y is size or shape of a structure in the developing organism, α is the age of onset of growth, β is the age at which the structure terminates its differentiation and growth, k is related to growth rate and S0 is the initial size at age α of the system. By integrating these differential equations, an ontogenetic trajectory for the structure is obtained. During evolution, when a descendent is derived from an ancestor, its ontogenetic trajectories are modified; these modifications appear as changes in α, β or k and their developmental translation is heterochrony, because these three parameters involve developmental time. This was a seminal paper, because it stimulated a source of new ideas for research in palaeontology and related fields of evolutionary biology. An international meeting in Dijon in 1986 (Colloque International Ontogénèse et Évolution) showed the relevance of these new ways of tackling evolutionary problems. Pere Alberch talked there about monstrosities, one of the most exciting lectures presented during the Colloque. However, he saw the limits of heterochronic thinking when used in comparative biology, in phylogenetic analysis and cladistics above all (Alberch, 1985).

What ideas dominated his thinking? Zoology was his early interest and his work on herpethology was his zoological grounding. He was interested in the development of vertebrate limbs and a large amount of his empirical work was devoted to this theme. Today his paper with his brother Jordi on quantitative studies on heterochronic processes in salamanders (Alberch & Alberch, 1981) has been recognized as a classical model for evolutionary thinking, but he wrote many other papers dealing with developmental features of vertebrates (e.g. Alberch, 1983a; Shubin & Alberch, 1986; Müller & Alberch, 1990; Blanco & Alberch, 1992).

However, an important part of the theoretical background underlying these quantitative concepts came from Waddington's (1957) book The Strategy of the Genes. Waddington highlighted the buffered character of development. He coined concepts such as canalization and creode (canalized developmental pathways), which refer to the stability of developmental processes. Waddington conceived of developmental systems as reaching stable steady states. Thus, steady states were defined by quite broad boundary conditions and a gradual transition between two different steady states would not be possible. For Alberch et al. (1979), ontogenetic trajectories were quantified expressions of creodes. In Alberch (1980), his main ideas about morphological evolution related to development were outlined. The role of genes would not be as relevant. Genes could not be considered outside development; development is the effect and the cause of genic expression (Alberch, 1991a,b).

He had a diametrically opposed perception of evolution based on his knowledge of development. Neodarwinian theory deals with isolated genes; genome is seen as an atomized entity. Since genic expression takes place through development and is regulated by it (Alberch, 1982a, 1983b), phenotypic consequences of genes are strongly constrained and an atomized conception of the genome cannot be sustained. This is the concept of developmental constraint on evolution (cf. Alberch, 1982b), one of his main contributions to evolutionary thinking. This new way of seeing evolution collides with another proposition of the Modern Synthesis by which random mutation, mainly sorted by natural selection, results in adaptive diversity. According to Alberch, there are only a finite number of phenotypic solutions and they consist of stable steady states. Thus, a more internalist character of evolution has to be reconsidered (Alberch, 1989a). The play between the environment and phenotypes produces a selection of the latter, but only a previous subset of morphologies is developmentally possible. Selection again constrains the field of possible morphologies. Adaptation is a new constraint, but developmental constraints act beforehand (Oster & Alberch, 1982). Therefore, Neodarwinism predicts well which phenotypes survive in a given environment, but it does not tell us which will most likely be produced by development (Alberch, 1982b).

There is a limited set of phenotypes, but each one has a different probability of being produced by developmental processes. Since the developing organism can be described as a nonlinear dynamic system, thresholds, bifurcations and inequiprobability between transitions are essential features of them and, thus, of developing organisms (Alberch, 1980; 1982b; Oster & Alberch, 1982). Thresholds and bifurcations challenged the gradualist metaphor of evolution because a continuous genetic variation would be driven into phenotypical discontinuities through development. Development would be the main reason for morphological stasis and the punctualistic character of evolution shown by the fossil record, a topic that Eldredge and Gould had evidenced in their seminal paper in Models in Palaeobiology, edited by T. J. M. Schopf in 1972. Alberch also wanted to attack the question of developmental constraints on evolution by means of another argument: the nature of monstrosities, which he had already outlined previously (Alberch, 1980). Around this theme, he gave his presentation as invited lecturer in Dijon (Alberch, 1989b). If morphological diversity exhibits empty zones as shown in the morphospace of possible forms, these discontinuities are interpreted by Neodarwinism as a consequence of the work of natural selection. Therefore, teratological morphologies should show infinite variability before they are removed by selection, but they have patterns of variation as restricted as those of the ‘normal’ forms. Thus, limited variation is not a product of selection but a result of developmental constraints for monsters; the same line of reasoning may be used for explaining these empty regions in the morphospace of normal forms. Another consequence of this was that convergence and parallelism would be ubiquitous along the evolutionary process because only a small set of developmental pathways could be followed. Finally, this limits chance. Natural selection could not sort an infinite and continuous variety of shapes, but only a limited and discontinuous one, a position very close to that of D’Arcy Thompson or Waddington. This is the most important evolutionary consequence of this review of teratologies from the point of view of evolutionary biology.

Morphogenetic processes have been considered to be governed by physical forces since the times of D’Arcy Wentworth Thompson, and Waddington's proposals took into account this concept. Now, new conceptual tools have arisen in order to deal with complex systems. The study of dynamical systems is a new conceptual field with important applications in different sciences such as biology, chemistry or economics. On the other hand, nonlinear dynamical systems are ubiquitous in nature. Pere had always argued from this perspective since he published his early papers. Dynamical systems theory was an adequate tool for him because he was not a reductionist biologist, another unorthodox but more fruitful position. Current developments in dynamical systems involve chaos theory, self-organization and complexity. He saw in these topics a new line for advancing towards understanding of life. His last project, a book entitled ‘An Introduction to Chaos Theory and Complexity with a Special Emphasis on Biological Sciences’ was on this branch of theoretical biology. This would have been a great contribution to biological thought but unfortunately, it will never be published. We agree with David Wake that his best was yet to come and Pere's early death has cost the scientific community dearly.