A cluster if observations poses puzzles.

If the life of the universe were a book, the later parts would be clearly legible. It's the early chapters that remain fuzzy.

Using the world's largest optical telescope, the 10-meter W.M. Keck atop Hawaii's Mauna Kea, several teams of astronomers have recently taken a leap back in time, sketching in some of the details about the cosmos shortly after its birth some 10 to 20 billion years ago.

In May, Judith G. Cohen of the California Institute of Technology in Pasadena and her colleagues reported that distant galaxies, viewed as they appeared when the cosmos was half its current age, were clumped together rather than distributed evenly across the sky. The finding suggested that this lumpiness in the cosmos may have arisen earlier than some theories can easily account for (SN: 4/27/96, p. 260). Now, another Keck observation hints that such clustering might have occurred earlier still, when the cosmos was less than one-fifth its current age.

The new findings focus on the region surrounding the distant quasar BR 2237-0607. Last year, Richard G. McMahon of the University of Cambridge in England and Esther M. Hu of the University of Hawaii in Honolulu found what appears to be a young, ordinary galaxy in the neighborhood of the quasar (SN: 9/30/95, p. 212). The galaxy's measured redshift of 4.5 means that the light now reaching Earth left the galaxy when the cosmos was just a few billion years old. McMahon and Hu have spied a second galaxy in the vicinity, they reported June 10 at a meeting of the American Astronomical Society in Madison, Wis.

A grouping of two ordinary galaxies and a quasar does not necessarily a cluster make, McMahon emphasizes. Quasars, dazzling powerhouses that may represent an unusual type of galaxy, seem more likely than other objects to reside in groups. Yet McMahon says that the discovery, together with evidence of clustering later in the history of the cosmos, suggests that astronomers may find larger groupings in the early universe as telescopes like Keck deepen their view.

"Seeing even one object this far away is difficult, and so finding even another one is suggestive of clustering," says Mark A. Dickinson of the Space Telescope Science Institute in Baltimore.

"It's dicey; they really don't have enough data," says Cohen. But, she adds, the finding could be the "beginning of a great discovery."

Cohen says she and other astronomers, including Lennox L. Cowie of the University of Hawaii, have confirmed her team's evidence of clustering later in cosmic history by analyzing the Hubble Deep Field, the most detailed deep-sky images ever made. At a workshop in Cambridge next week, Dickinson plans to present observations of a cluster of galaxies dated slightly earlier than that of Cohen's team.

At the Madison meeting, Matthew A. Malkan and his colleagues at the University of California, Los Angeles reported finding two clusters of infant galaxies dating from about one-third the universe's current age. This places them even earlier in cosmic history than Dickinson's finds but later than those reported by Hu and McMahon.