IF AN ART MUSEUM were burning and only one object could be saved, which would it be? A Raphael Madonna? An unforgettable Degas portrait? A 15th-century Flemish altarpiece? These were the most popular candidates for rescue at a museum where I once worked. But while favorites appeared over and over again, no one on the staff remained loyal to previous choices.

This absence of unanimity or consistency reflects a fact about art on which both Kenneth Clark in What Is a Masterpiece? and John Canaday in What Is Art? agree: There is no one work, no one formula, no one absolute that makes a work of art great. While these books do not offer any really new material or observations, they do provide some guidelines to understanding the criteria for, and the qualities of, great art.

Clark's book, based on a recent lecture of the same name delivered at the University of London, is a short text devoted to specific examples of Western European masterpeices. Canaday's, an expansion of a series of monographs he wrote for the Metropolitan Museum Seminars in Art in 1958 and 1959, is a lengthy volume touching on broad concepts of esthetics.

Both writers ascribe to the idea that a great work of art expresses a double relationship to its own era and to the past. This quality, which Clark refers to as "density," is rarely disputed as being fundamental to the so-called masterpieces of cultural history. The authors also share a conviction that the appreciation of great art requires retrospective evaluation. Canaday spells this out when he uses the new Georges Pompidou Art Center (Beaubourg) as an example of art about which final esthetic judgment should in his opinion, be reserved until the building's response to the goals of its own era and to its intended function can be assessed.

Clark, who has served in more notable positions in the museum and academic world than can be conveniently recounted and has produced many books, including the extraordinary popular text and television series, Civilisation, is to many of us the Alistair Cooke of art. In What Is a Masterpiece? he takes us on an esthetic journey from Giotto's frescoes at the Arena Chapel to Picasso's Guernica. With references to the stylistic chronology of Western European art from the 14th to the 20th century, he moves from one selected work to another while he elaborates on the checks and balances of subject, form and inspiration that, when in perfect equilibrium, support the concept of a masterpiece. His criteria for masterpiece status are as comfortable as Cooke's arm-chair. But they often reflect rather narrow and conservative standards of judgment.

For example, Clark discourses at lenght on the importance of a great theme to the creation of a masterpiece, empahsizing that the Christian story has been the most fertile soil for dramatic and inspirational works of art. While he allows that portraits can also be masterpieces and he calls Manet's Olympia (a painting of a nude prostitute being presented flowers from a male admirer) a masterpiece, he firmly believes that very few subjects can be the basis for a great work of art. Indeed, he seems to pay only lip service to what he considers less serious themes, observing that masterpieces must use the esthetic syntax of their day, "however degraded it may seem." While he is somewhat more liberal in his discussion of the effect of scale on the status of painting, Clark overlooks the fact that monumentality is often as influential a factor in contemporary art as it was in the acclaimed altarpieces, murals and history paintings of past centuries.

Clark steers clear of trouble by handing us a neat list of his "greatest hits" in art, almost all of which are undisputed in terms of quality and impact. But this isn't all bad, coming, as it does, from a true connoisseur. As conservative as it may be, his little book is a challenge for the uninitiated. It also brings out some satisfying lessons on the concept of a masterpiece, a topic on which few authorities like to comment. Indeed, the most recent issue of ARTnews, which polled 27 scholars, curators and museums directors on what makes a masterpiece, reported that only 14 among them would or could respond. Why? As some pointed out, masterpiece is an overblown, underrated and misunderstood word. Clark deserves some credit for trying to define it.

Unlike Clark, John Canaday, former art critic of The New York Times, takes a thematic rather than a chronological apprach to his subject, which includes but does not stop at what constitutes a masterpiece. In fact, his book amounts to a compendium of topics, ranging from the definition of art to its expressive, stylistic and technical methods. While this ambitious approach can be disconcerting, the persevering reader can benefit from some good things in it.

Canaday takes advantage of his thematic organization to compare specific examples of art in successive chapters, thereby building upon premises and observations to enlarge the reader's approach to art apreciation. Particularly helpful to the beginner is Canaday's sensitivity to visual strategies, which he explains in three chapters devoted to composition.

Canaday's chapters on techniques are also fresh and interesting. This may be because he flavors the potentially boring technical details of a medium with examples of how individual artists worked with it. For instance, he gives us an energetic view of Michelangelo as he created and then executed the frescoes for the ceiling of the Sistine Chapel, and a lively discussion of the woodcut. Throughout these chapters he also presents brief discussions of art conservation which, while superficial, are usually interesting to layman and art historian alike.

Canaday's book benefits -- and suffers -- most from his instinctive critical approach. To his credit, this approach leads him to point out the vagaries and periodic inertia of art criticism that have buried certain artists and styles, only to exhume them for plaudits at a later date. His critical insights are scattered throughout the book. But for all of their enrichment of the text, they also confuse the book's identity.

In the end What is Art? lies uncomfortably between a reference book, to be consulted for specific information, and a short general introduction to the visual and critical evaluation of art. Canaday's valuable analytic threads get lost in the scuffle.