Son of Cezanne.

The transparency and delicacy of the Analytical Cubist style were largely used by its founders to transform solid, compact, often rather massive objects: clusters of houses; chunky trees; sturdy tables bearing bulky instruments or containers; single human figures shown head-and-shoulders or truncated to half- or three-quarters-length, so that they are firmly grounded. That is to say, though late Cezanne was the source of their style, Braque and Picasso evaded many of late Cezanne's key motifs: stretches of land or water or sky; long tapering branches of trees; full-length figures; groups of bathers. Braque and Picasso declined to deal with things like spindly limbs, whether of trees or of human beings; they focused on things that were ample and stable as the pyramids.

Mondrian's Analytical Cubism began where theirs left off: with Gray Tree, usually said to have been painted in 1912 but tentatively ascribed by Joop Joosten and Angelica Zander Rudenstine in the present catalogue to September 1911; and The Sea, painted in the summer of 1912. The most substantial form in either picture is the slim, pliant trunk in Gray Tree. Otherwise, long, slightly curving lines traverse the canvases - in The Sea entirely from side to side, in Gray Tree curving a little toward the upper corners. Mondrian soon synthesized those contrasting movements in a somewhat more abstracted form in Flowering Appletree, perhaps the first work in his career that could be taken as an overture to what came after.

The Sea is a perfect crystallization of uninterrupted ebb and flow. The sense of movement outward from an axis in Gray Tree crystallizes a tree's life as a paradigm of annual ebb and flow. in both paintings, then, process rather than product is the subject - natural processes and, in particular, those that could have a metaphysical import. There is a metaphysical feel, of course, to all great Analytical Cubist paintings, because of their mysterious translucence, but with Mondrian's it is reinforced not only by their more rarefied, more ascetic quality but by their themes, which evoke the cycle of life and death and are the stuff of ritual and myth.

Earlier in 1911 Mondrian had completed his most overtly theosophical work, the triptych Evolution, with its three hieratic standing figures. Evolution had several attributes in common with Picasso's Les Demoiselles d'Avignon, painted four years earlier. One was its attempt to make a big statement. A second was the effort to subjugate its viewers by confronting them with towering idols in possession of compelling stares. A third was that the artist's friends found the result laughable. Where Evolution differed from Les Demoiselles was that the laughter never turned to amazement. (It will induce neither laughter nor amazement at the current exhibition, from which this space-intensive piece has sensibly been left out in line with the curators' emphasis on achievement rather than mere experiment.)

Yet turn from that work to The Tree A of early 1913, which comes uniquely close, it has always seemed to me, to conveying an atmosphere such as pervades Michelangelo's late drawings of the Crucifixion. The story of Mondrian is implicit in how, having failed to produce a religious art out of a study of religious texts, he almost immediately produced a religious art through a study of the art of two impeccably secular contemporaries; in total contrast to the innocence with which he got embroiled in the irrelevance of Evolution was the clearheadedness with which he perceived what he could and what he should not take over from Picasso and Braque. The moment of decision was, of course, the crisis of legibility that came to a head in the spring of 1912. Braque and Picasso fought their way out of it by starting to incorporate bits of the real world - actual bits or simulated bits - and gradually abandoning the whole apparatus of Analytical Cubism. Mondrian's decision amounted to a statement that painting, having established a certain distance from reality in the wake of Cezanne's achievement of -a harmony parallel to nature," ought not to be letting reality in again through the back door; that it ought to he insisting on its autonomy by becoming less and less referential, more and more musical.

The decision was expounded in 1913 in the series of pictures using an orthogonal grid and a palette of gray and rust and titled Tableau or Composition. The rejection of the path into Synthetic Cubism was dramatized in a related work from the same year, the oval picture in the Stedelijk titled Tableau No. 3. This is virtually a pastiche of Braque's Man with a Violin (in the E. G. Buhrle Foundation, Zurich), painted a year earlier, a classic example of the last days of Analytical Cubism. (By the time Mondrian painted his version, both Braque and Picasso had produced oval still lifes in the Synthetic Cubist style.) Mondrian's picture seems to be deliberately echoing its model in that, besides the close resemblances in morphology and composition, it has the same zone down the center that is more luminous than the rest. In Braque's composition, it is precisely in this luminous zone that the violinist and his instrument are faintly but unmistakably adumbrated; the corresponding area in the Mondrian is the area that is most immaculately nonfigurative.

So much for the negative aspect of Mondrian's choice of direction. The positive aspect is pithily summarized in Yve-Alain Bois' catalogue essay: "A comparison between the 1912 Flowering Appletree and the 1911 Gray Tree, whose motif it echoes, reveals a considerable difference between mere curves and curves as potential straight lines. From this new conception of the image in progress flows all of Mondrian's subsequent work. Most importantly, what emerges from it is the idea that a painting is the result of a kind of struggle, a precarious equilibrium that must remain suspended so as to sustain its intensity." I hope I am not putting words into Bois' mouth when I infer that Mondrian is thereby Cezanne's truest heir.

Of course the sense of the artist as a man of destiny is a very commonplace clement in the history and mythology of Modernist art. Even so, Mondrian is a singularly poignant and heroic case. A Mondrian retrospective is not just a procession of great pictures but a progression that is in itself an esthetic experience: the history of the man's art becomes as much a thing of beauty as the art. And because the exhibition charts that progression with the highest curatorial intelligence and rigor, it has been the most satisfying 20th-century one-man show I've seen. (I saw it at the Gemeentemuseum, with its partly natural top light and its handsome architecture and its civilized and not too numerous public; each venue has its comparative advantages and disadvantages.)

The variety of the work is amazing, given its unity and consistency of purpose. The development, for all its momentum and inevitability, is always able to surprise. The quality is sustained as it is by no other artist of this century. Periods one had not thought of as outstanding are suddenly seen to be high points: I had never before realized, for instance, that 1922 was a five-star vintage. There is very little letup in this exhibition. One can move from room to room as if in a state of shock. I use this shor-thand because I do not yet know how to describe what the particular state is that a Mondrian induces. (It remains much the same state, despite the variety of the corpus.) It is a sort of loss of self. The mind feels intent, but it also feels empty and unfocused. The body, though, does feel focused, as if it were channeling energy into the picture. It is not that the picture becomes in some way a reflection of oneself. It's that the picture has taken one over.

David Sylvester is an art historian living in London. His most recent book was entitled Looking at Giacometti (1994).