I am honored to introduce Professor David Ruelle's ``Poincare Prize'' lecture.

His scientific carrier is remarkable for his various contributions and for the conceptual continuity of the development through them. He has been among the first to realize the relevance of a rigorous derivation of the properties of equilibrium Statistical Mechanics as an essential step towards understanding the theory of phase transitions. His work has been an important guide to the scientists who in the sixties were attempting accurate measurements of thermodynamic quantities, like critical exponents, in various statistical mechanics models using the newly available electronic computational tools in conjunction with the use of rigorous results for assessing the correctness and reliability of the computations.

The treatise on Statistical Mechanics, 1969, has become a classic book and it is still the basis of the formation of the new generations of scientists interested in the basic aspects of the theory. He has written several other monographs which are widely known and used.

His critical work on the structure of Equilibrium Statistical Mechanics led him to undertake in 1969 the analysis of the theory of turbulence. The first publication on the subject was the epoch making paper ``On the nature of turbulence'' in collaboration with Takens. The paper criticized the theory of Landau, based on the increasing complexity of quasi periodicity arising from successive bifurcations in the Navier Stokes equations. The main idea that only ``generic'' behavior should be relevant was a strong innovation at the time: this is amply proved by the hundreds of papers that followed on the subject, theoretical, numerical and experimental.

The works making use of Ruelle's ideas stem also, and perhaps mainly, from the innovative papers Ruelle wrote (and continues to write) after the mentioned one. There he developed and strongly stressed the role that dynamical systems ideas would be relevant and important in understanding chaotic phenomena.

The impact on experimental works has been profound: one can say that after the first checks were performed, some by notoriously skeptical experimentalists, and produced the expected results we rapidly achieved, by the end of the seventies, a stage in which the ``onset of turbulence'' was so well understood that experiments dedicated to check the so called ``Ruelle--Takens'' ideas on the onset of turbulence were no longer worth being performed as one would know what the result would be.

The very fact that a study of the onset of turbulence was physically interesting was new at the time (the sixties). The ideas had been independently worked out by Lorenz, earlier (in a 1963 paper): this became clear almost immediately. However I think that Ruelle's view, besides reviving the interest in Lorenz' work, which had not been appreciated as it should have, were noticed by physicists and mathematicians alike, and perhaps had more impact, because they were more general and ambitious in scope and aimed at understanding from a fundamental viewpoint a fundamental problem.

In 1973 he proposed that the probability distributions that describe turbulence be what is now called the ``Sinai-Ruelle-Bowen'' distribution. This was developed in a sequence of many technical papers and written explicitly only later in 1978. In my view this is the most original contribution of Ruelle: it has not been well understood for years although it has been quoted in impressively many works on chaos. It had impact mostly on numerical works, but it proposes a fundamental solution to one of the most outstanding theoretical questions: what is the analog of the Boltzmann--Gibbs distribution in non equilibrium statistical mechanics? his answer is a general one valid for chaotic systems, be them gases of atoms described by Newton's laws or fluids described by Navier Stokes equations (or other fluid dynamics equations). Today the idea is still a continuous source of works both theoretical and experimental.

Since the beginning of his work he has studied also problems concerning other fields like operator theory and operator algebras obtaining results remarkable for originality and depth: I mention here only his results on the Lee--Yang theorem on the location of the zeros of polynomials (a subject to which he continued to add new results and applications) and the Haag--Ruelle theory of scattering in relativistic quantum fields, very widely studied and applied, which is still today virtually the only foundation for relativistic scattering theory, employed in mathematical Physics, high energy phenomenology and theoretical Physics.

In the last few years he has also provided important impulse to the development of nonequilibrium statistical mechanics: his work continues in this direction at the highest level. He has developed foundational papers for the thery of nonequilibrium Thermodynamics particularly with respect to the concept of entropy.

The work of Ruelle is of mathematical nature: but it is an example of how important a conceptually rigorous and uncompromising approach can be fruitful and lead to progress in very applied fields like experimental fluid mechanics or numerical molecular dynamics simulations. His work is in the tradition of the 1800's fundamental investigations in Physics.

His work, books and papers, is always very careful, clear and polished: every word, however, is important and requires attention.

The awarding of the prize recognizes the cultural influence that he has exercized in the last thirty years or so: and we are all here united in this recognition.

Giovanni Gallavotti