Sir Peter Medawar has in recent years become nearly as well known for his essays on the nature of scientific activity as he had earlier for his experimental investigations in immunology and tissue transplantation. The popularity of his writing owes as much to his incisive, pithy style as to his trenchant insights, such as his definition of science as "the art of the soluble."

"The Limits of Science," which returns to many of the themes Medawar has treated in earlier books, actually consists of three separate essays. The first, "An Essay on Scians," is a collection of short, aphoristic discussions of disparate aspects of science and its relations to the broader world such as "Science and Cricket," "Science and Culture," "Women in Science" and "On Being too Punctilious in the Execution of Experiments." The second and much shorter essay poses the question, "Can Scientific Discovery Be Premeditated?" and features three capsule accounts of discoveries that Medawar claims could never have been premeditated. He goes on to argue that government support for scientific investigation should not be based on "customer/contractor treaties" that specify what the researcher should find. The third essay, from which the book receives its title, asserts that there is no foreseeable limit to the progress of science, within the realm of questions that science is suited to answer, but that science has nothing to say about "first and last things," such as the origin of the universe and the purpose of human existence.

As in his earlier writings, Medawar sums up general characteristics of scientific activity in felicitous, illuminating passages. It would be hard to improve on the following:

"The truth is that there is no such thing as 'scientific inference.' A scientist commands a dozen different stratagems of inquiry in his approximation to the truth, and of course he has his way of going about things and more or less of the quality often described as 'professionalism' -- an address that includes an ability to get on with things, abetted by a sanguine expectation of success and that ability to imagine what the truth might be that Shelley believed to be cognate with a poet's imagination."

In other places, however, Medawar's succinct style has less fortunate effects. He moves so quickly across issues of great complexity that the reader often cannot tell how he has reached his conclusions. He is fond of making pronouncements that are to him "so obviously right as not to be worth the exertion of justifying them." Sometimes he seems to be settling old scores -- against politicians who dislike scientists, against humanists who regard scientists as uncultured, against a public that blames science for the ills of the world. The abruptness with which he dismisses opinions of his "brethren in humanistic studies" borders on arrogance. Among the harshest of his dismissals is that of the "IQ psychologists," whose acceptance of the fraudulent data of Sir Cyril Burt he ascribes to their "sheer stupidity." As if that were not bad enough, he adds, "They are almost unteachable, too."

Some of the topics to which Medawar addresses himself in this book he has discussed more fully and coherently in his previous essays. But to search here for carefully reasoned arguments in support of Medawar's positions is to misunderstand the nature of "The Limits of Science." This book is not an argument for Medawar's views but a profession of his unbounded faith in the glorious adventure of science to which he has devoted his life. Only in the last section of the book, when he confesses that he is not proud of his inability to share in the religious faith that sustains others, do we glimpse the moving vulnerability of a man committed to rationalism in an age in which he fears that rationalism is going out of fashion. As a statement of the convictions that have guided Medawar through his rich professional life, "The Limits of Science" has an attraction strong enough to overcome its shortcomings.