Three `detectives' find Nijinsky designs. How did the great dancer actually stage the ballet `Le Sacre du Printemps'?

Nijinsky, undisputed greatest dancer of all time, choreographed four ballets. In only one of these, ``Le Sacre du Printemps,'' he did not dance. Ironically, this is also the ballet that was performed only eight times after its May 29, 1913, premi`ere in Paris. Last night, Nijinsky's ``Sacre'' was scheduled to be performed for the first time in 74 years by the Joffrey Ballet here at the Dorothy Chandler Pavilion. The reconstruction of this ballet, measure by painstaking measure, was the achievement of three individuals: Robert Joffrey, artistic director of the Joffrey Ballet; Millicent Hodson, dance historian; and Kenneth Archer, art historian.

The attempt to revive Nijinsky's lost choreography actually began in 1955, when Robert Joffrey was teaching his ballets to Marie Rambert's company in London. The late Rambert showed Joffrey steps from the original production, for which she was rehearsal assistant to Nijinsky, helping the dancers with Stravinsky's complicated rhythms. Handwritten notes in Russian on Rambert's score later helped Joffrey and Hodson decipher the puzzle of choreography, which was never written down.

Joffrey says he was further hampered in the detective work because ``Nijinsky never talked his choreography - he showed it.'' Only one section of the ballet was written, the solo for the Chosen One (the maiden who sacrifices herself). This was recorded by Nijinsky's sister, Branislava, who imitated the choreography as Nijinsky demonstrated it in their living room in November 1912. He started with her solo and worked backward on the score to add the other dancers. Unable to dance in the production because she became pregnant, Branislava made notes. The only other notations appear on Rambert's rehearsal score and on Stravinsky's score, which contains comments about entrances and exits of groups of dancers.

Connecting the gaps between known material was the job of Millicent Hodson, who first discussed the project with Robert Joffrey on the Berkeley campus in 1971, where she was studying for a PhD.

Hodson's search led her to interview Kenneth Archer, who was researching a biography of Nicholas Roerich, the scene and costume designer of ``Sacre.'' Archer and Hodson continued the research for more than a decade on three continents (marrying along the way), going through museums and archives, talking with Roerich's son in India, interviewing dancers, musicians, spectators, and critics who were present at one of ``Sacre's'' legendary performances.

Sadly, not one playbill remains, but the team found that critics' descriptions of movements and stage patterns were most helpful. Some mentioned that male dancers were percussive with their ``insistent shuffling of the feet.''

The 13-year-old daughter of the company physician remembers the atmosphere: ``It was like a volcano,'' when the performance caused a riot in the house. The young audience member said, ``I couldn't believe French aristocrats were acting that way - one woman stuck another with her hatpin.'' Diaghilev had to appear during the three-minute entr'acte and beseeched the audience to stop fighting among themselves.

Details of the entr'acte curtain caused much consternation for Archer, as he tried to reconstruct d'ecor for the Joffrey revival. Diaghilev's secretary told him, ``You'll never find the originals - they've been cut up to patch other d'ecors or scrub the stage floor.''

The costumes, which French critics in 1913 called ``a violent feast for the eyes,'' were scattered in an auction, but Archer went to Sotheby's and has been able to relocate half of them.

These costumes, with their symbolic geometric patterns, gave Hodson clues to the appropriate movements to be matched by the dancers.

Such interplay was typical of their discovery process: As Archer uncovered something in his office, he'd shout to Hodson in her nearby studio, ``I've found it!'' His little clue edged Hodson closer to bridging yet another choreographic gap.

Another source was the drawings of a Paris artist made while she watched several performances. These sketches were later translated into pastels and saved from a trash can at the time of her death by dance writer Richard Buckle. On view at the Victoria and Albert Museum today, they show formations of dancers and were useful in suggesting what the original Roerich d'ecor looked like.

While Hodson studied Stravinsky's notes and correspondence, she discovered that Nijinsky changed the composer's metronome tempos. Dancers and orchestra had only two rehearsals together before the premi`ere, when it was discovered that Stravinsky's tempo was much too fast for the dancers.

Those who remember working with Nijinsky say that he ``visualized the music literally,'' never used French terms in rehearsal, wanting to completely abandon ballet vocabulary by substituting turned-in feet and contained energy, the hands in fists or flat like ``spatulas.'' Piecing together these verbal descriptions with a few photos, Hodson choreographed the ballet scene by scene. When Joffrey watched a recent rehearsal, it triggered memories of something else taught to him by Marie Rambert, so there were 11th-hour changes and additions.

The result is believed by Robert Joffrey to be 85 percent Nijinsky, a degree of accuracy that owes a lot to Archer's nine dossiers and Hodson's 12 notebooks, all due to appear in their future book, ``The Search for Le Sacre du Printemps.''