The length of a novel, more surely even than any of its other qualities, needs to be determined by the subject. The novelist should not concern himself beforehand with the abstract question of length, should not decide in advance whether he is going to write a long or a short novel; but in the act of composition he must never cease to bear in mind that one should always be able to say of a novel: “It might have been longer,” never: “It need not have been so long.”

Length, naturally, is not so much a matter of pages as of the mass and quality of what they contain. It is obvious that a mediocre book is always too long, and that a great one usually seems too short. But beyond this question of quality and weightiness lies the more closely relevant one of the development which this or that subject requires, the amount of sail it will carry. The great novelists have always felt this, and, within an inch or two, have cut their cloth accordingly.

Mr. A. C. Bradley, in his book on Shakespeare’s tragedies, threw a new and striking light on the question of length. In analyzing “Macbeth,” which is so much shorter than Shakespeare’s other tragedies that previous commentators had always assumed the text to be incomplete, he puts the following questions: If the text is incomplete, at what points are the supposed lacunas to be found? Does any one, on first reading “Macbeth,” feel it to be too short, or even notice that it is appreciably less long than the other tragedies? And if not, is it not probable that we have virtually the whole play before us, and that Shakespeare knew he had made it as long as the subject warranted and the nerves of his audience could stand? Whether or not the argument be thought convincing in the given case, it is an admirable example of the spirit in which works of art should be judged, and of the only system of weights and measures applicable to them.

Tolstoy gave to “Ivan Ilyitch” just enough development to make a parable of universal application out of the story of an insignificant man’s death. A little more, and he would have dropped into the fussy and meticulous, and smothered his meaning under unnecessary detail. Maupassant was another writer who had an unerring sense for the amount of sail his subjects could carry; and his work contains no better proof of it than the tale of “Yvette” — that harrowing little record of one of the ways in which the bloom may be brushed from a butterfly.

Henry James, in “The Turn of the Screw,” showed the same perfect sense of proportion. He had ventured to expand into a short novel the kind of tale usually imposed on the imagination in a single flash of horror; but his instinct told him that to go farther was impossible. The posthumous fragment, “The Sense of the Past,” shows that he was again experimenting with the supernatural as a subject for a long novel; and in this instance one feels that he was about to risk over-burdening his theme. When I read M. Maeterlinck’s book on the bee (which had just made a flight into fame as high as that of the insect it celebrates) I was first dazzled, then oppressed, by the number and the choice of his adjectives and analogies. Every touch was effective, every comparison striking; but when I had assimilated them all, and remade out of them the ideal BEE, that animal had become a winged elephant. The lesson was salutary for a novelist.

The great writers of fiction — Balzac, Tolstoy, Thackeray, George Eliot (how one has to return to them!) — all had a sense for the proportion of their subjects, and knew that the great argument requires space. There are few things more exquisite in minor English verse than Ben Jonson’s epitaph on Salathiel Pavy; but “Paradise Lost” needs more room, and the fact that it does is one of the elements of its greatness. The point is to know at the start if one has in hand a Salathiel Pavy theme or a “Paradise Lost” one.

In no novelist was this instinct more unerring than in the impeccable Jane Austen. Never is there any danger of finding any of her characters out of proportion or rattling around in their setting. The same may be said of Tolstoy, at the opposite end of the scale. His epic gift — the power of immediately establishing the right proportion between his characters and the scope of their adventure — seems never to have failed him. “War and Peace” and Flaubert’s “Education Sentimentale” are two of the longest of modern novels. Flaubert too was endowed with the rare instinct of scale; but there are moments when even his most ardent admirers feel that “L’Education Sentimentale” is too long for its carrying- power: whereas in the very first pages of “War and Peace” Tolstoy manages to establish the right relation between subject and length. But there is another difference between the great novel and the merely long one. Even the longest and most seemingly desultory novels of such writers as Balzac, Flaubert and Tolstoy follow a prescribed orbit; they are true to the eternal effort of art to complete what in life seems incoherent and fragmentary. This sense of the great theme sweeping around on its allotted track in the “most ancient heavens” is communicated on the first page of such novels as “War and Peace” and “L’Education Sentimentale”; it is the lack of this intrinsic form that marks the other kind of long novel as merely long.

M. Romain Rolland’s “Jean-Christophe” might be cited as a case in point. In a succession of volumes, planned at the outset as parts of a great whole, he tells a series of consecutive soul-adventures, none without interest; but such hint of scale as there is in the first volume seems to warrant no more than that one, and the reader feels that if there are more there is no reason why there should not be any number. This impression is produced not by the lack of a plan, but of that subtler kind of composition which, inspired by the sense of form, and deducing the length of a book from the importance of its argument, creates figures proportioned to their setting, and launches them with a sure hand on their destined path.

The question of the length of a novel naturally leads to the considering of its end; but of this there is little to be said that has not already been implied by the way, since no conclusion can be right which is not latent in the first page. About no part of a novel should there be a clearer sense of inevitability than about its end; any hesitation, any failure to gather up all the threads, shows that the author has not let his subject mature in his mind. A novelist who does not know when his story is finished, but goes on stringing episode to episode after it is over, not only weakens the effect of the conclusion, but robs of significance all that has gone before.

But if the form of the end is inevitably determined by the subject, its style — using the term, in the sense already defined, to describe the way in which the episodes of the narrative “are grasped and coloured by the author’s mind” — necessarily depends on his sense of selection. At every stage in the progress of his tale the novelist must rely on what may be called the illuminating incident to reveal and emphasize the inner meaning of each situation. Illuminating incidents are the magic casements of fiction, its vistas on infinity. They are also the most personal element in any narrative, the author’s most direct contribution; and nothing gives such immediate proof of the quality of his imagination — and therefore of the richness of his temperament — as his choice of such episodes.

Lucien de Rubempe (in “Les Illusions Perdues”) writing drinking songs to pay for the funeral of his mistress, who lies dying in the next room; Henry Esmond watching Beatrix come down the stairs in her scarlet stockings with silver clocks; Stephen Guest suddenly dazzled by the curve of Maggie Tulliver’s arm as she lifts it to pick a flower for him in the conservatory; Arabella flinging the offal across the hedge at Jude; Emma losing her temper with Miss Bates at the picnic; the midnight arrival of Harry Richmond’s father, in the first chapter of that glorious tale: all these scenes shed a circle of light far beyond the incident recorded.

At the conclusion of a novel the illuminating incident need only send its ray backward; but it should send a long enough shaft to meet the light cast forward from the first page, as in that poignant passage at the end of “L’Education Sentimentale” where Mme. Arnoux comes back to see Frederic Moreau after long years of separation.

“He put her endless questions about herself and her husband. She told him that, in order to economize and pay their debts, they had settled down in a lost corner of Brittany. Arnoux, almost always ailing, seemed like an old man.Their daughter was married, at Bordeaux; their son was in the colonial army, at Mostaganem. She lifted her head: ‘But at last I see you again! I’m happy’ . . She asks him to take her for a walk, and wanders with him through the Paris streets. She is the only woman he has ever loved, and he knows it now. The intervening years have vanished, and they walk on, “absorbed in each other, hearing nothing, as if they were walking in the country on a bed of dead leaves.” Then they return to the young man’s rooms, and Mme. Arnoux, sitting down, takes off her hat.

“The lamp, placed on a console, lit up her white hair. The sight was like a blow on his chest.” He tries to keep up a pretense of sentimentalizing; but “she watched the clock, and he continued to walk up and down, smoking. Neither could find anything to say to the other. In all separations there comes a moment when the beloved is no longer with us.” This is all; but every page that has gone before is lit up by the tragic gleam of Mme. Arnoux’s white hair.

The same note is sounded in the chapter of “The Golden Bowl” where the deeply, the doubly betrayed Maggie, walking up and down in the summer evening on the terrace of Fawns, looks in at the window of the smoking-room, where her father, her husband and her stepmother (who is her husband’s mistress) are playing bridge together, unconscious of her scrutiny. As she looks she knows that she has them at her mercy, and that they all (even her father) know it; and in the same instant the sight of them tells her that “to feel about them in any of the immediate, inevitable, assuaging ways, the ways usually open to innocence outraged and generosity betrayed, would have been to give them up, and that giving them up was, marvelously, not to be thought of.”

The illuminating incident is not only the proof of the novelist’s imaginative sensibility; it is also the best means of giving presentness, immediacy, to his tale. Far more than on dialogue does the effect of immediacy depend on the apt use of the illuminating incident; and the more threads of significance are gathered up into each one, the more pages of explanatory narrative are spared to writer and reader. There is a matchless instance of this in “Le Rouge et le Noir.” The young Julien Sorel, the tutor of the Reynal children, believes a love-affair with their mother to be the best way of advancing his ambitions, and decides to test his audacity by taking Mme. Reynal’s hand as they sit in the garden in the summer dusk. He has a long struggle with his natural timidity and her commanding grace before he can make even this shy advance; and that struggle tells, in half a page, more of his fatuities and meannesses, and the boyish simplicity still underlying them — and more too of the poor proud woman at his side — than a whole chapter of analysis and retrospection. This power to seize his characters in their habit as they live is always the surest proof of a novelist’s mastery.

But the choice of the illuminating incident, though so much, is not all. As the French say, there is the wanner. In Stendhal’s plain and straightforward report of the scene in the garden every word, every stroke, tells. And this question of manner — of the particular manner adapted to each scene — brings one to another point at which the novelist’s vigilance must never flag. As every tale contains its own dimension, so it implies its own manner, the particular shade of style most fitted to convey its full meaning.

Most novelists who have a certain number of volumes to their credit, and have sought, as the subject required, to vary their manner, have been taken to task alike by readers and reviewers, and either accused of attempting to pass off earlier works on a confiding public, or pitied for a too-evident decline in power. Any change disturbs the intellectual indolence of the average reader; and nothing, for instance, has done more to deprive Stevenson of his proper rank among English novelists than his deplorable habit of not conceiving a boy’s tale in the same spirit as a romantic novel or a burlesque detective story, of not even confining himself to fiction, but attempting travels, criticism and verse, and doing them all so well that there must obviously be something wrong about it. The very critics who extol the versatility of the artists of the Renaissance rebuke the same quality in their own contemporaries; and their eagerness to stake out each novelist’s territory, and to confine him to it for life, recalls the story of the verger in an English cathedral, who, finding a stranger kneeling in the sacred edifice between services, tapped him on the shoulder with the indulgent admonition: “Sorry, sir, but we can’t have any praying here at this hour.”

This habit of the reader of wanting each author to give only what he has given before exercises the same subtly suggestive influence as all other popular demands. It is one of the most insidious temptations to the young artist to go on doing what he already knows how to do, and knows he will be praised for doing. But the mere fact that so many people want him to write in a certain way ought to fill him with distrust of that way. It would be a good thing for letters if the perilous appeal of popularity were oftener met in the spirit of the New England shop-keeper who, finding a certain penknife in great demand, did not stock that kind the following year because, as he said, too many people came bothering him about it.