The ancient paper, made from sliced reeds that grew in abundance in the Nile River, presents documents and records -- even some literature -- from as far back as 3000 b.c. Scholars use the papyrus to get clues about the economy, politics, and literature of ancient life.

Relatively few papyri have been published, though. Papyrus collections usually are only frequented by specialists who find the time and money to travel to the originals and who can translate the texts. The originals are mostly in Greek, though some are in Latin, Persian, Aramaic, Arabic or one of four different Egyptian scripts. The leaders of this digital project think students at all levels, and even the general public, will find interesting nuggets in the papyrus papers if they can get to them easily and read them in English. Duke and Michigan, which already have parts of their collections on the Web, have gotten thousands of hits from outside their universities, including some from the elementary school level.

Papyrus is rarely on the market now, and when it is it goes for exorbitant prices. Columbia got its collection going at the beginning of the century with a few thousand dollars per year approved by President Nicholas Murray Butler. Now the collection is stored in the Rare Book and Manuscript section of his namesake library, where pieces lay sandwiched between panes of glass or preserved in acid-free folders.

The idea for the digital project, called the Advanced Papyrological Information System (APIS), came from Roger Bagnall, chairman of the Department of Classics and one of 100 to 200 papyrologists in the world. He had the idea to digitize and integrate collections back in 1992.

"Everyone was focused on separate projects," he said. "But in real life, you follow leads, look something up, stumble across something. With this, you'll be able to weave in and out of images, bibliographies, original text."

With a grant from the National Endowment for the Humanities, six universities which own the most significant American papyrus collections are forming the core of the library: California, Columbia, Duke, Michigan, Princeton and Yale.

The process involves taking digital photos of each piece of papyrus, feeding images and text into a computer, then linking all the collections together. Each institution will maintain its own Web-based collection, and APIS will provide an interface to allow users to jump around in what is planned to be a relatively seamless way.

From a main index, which will be on Columbia's server, users can search all of the collections at once, then click to go to the image, text, translation or commentary.

After the original six members have contributed, other institutions are expected to join in, such as the Ashmolean Museum at Oxford, which houses the most important collection in Britain.

Columbia's Academic Information Services (AcIS) is working on one of the biggest challenges facing the project, which is technical compatibility. "Using digital information is a moving target," Bagnall said. "Every six months there is a different answer."

Another major problem with large digital efforts is obsolescence, the fear that computers will be speaking a different language in the future. "If papyrus had been digital in antiquity we wouldn't be able to read them at all now," Bagnall said.

Once the $600,000 NEH grant (which was divided among the six member institutions) runs out next year, the
project is expected to be up and running. The library will need to be a low-maintenance operation where material can be added easily with no full-time administrators required. Bagnall expects the digital collection to double every few years.

Is there a possible downside to the project, in that once it becomes easy to leaf through papyrus on a computer, people might not bother to seek out the originals?

"It would be a disadvantage if you only had the digital form," Bagnall said. "It doesn't give you a feeling for the dimensionality of the papyrus."

For that, you still need to tour the reeds.