The signs on the bulletin board across from the fourth-floor elevator in the biological laboratories usually change from week to week.

But one bulletin appears to have been posted there for nearly four years. Dated October 10, 1989, it is called, simply, "Gilbert Laboratory Notebook Practice."

The two-page sheet, with its eight instructions, has a timely warning for workers in the lab of Loeb University Professor Walter Gilbert '53 that belies its yellowing color. Write down everything, keep neat notebooks, take nothing for granted, the sheet says.

"The notebook will help protect you from the greatest danger that faces a scientist: self-deception," reads the sheet. "It helps prevent wishful memory [from] improving one's recollection of past experiments."

This piece of paper constitutes the bulk of efforts by Gilbert, who serves as chair of the Department of Cellular and Developmental Biology, to prevent errors or fraud in research. Some professors say the two-page sheet represents a greater effort to head off fraud than many of their colleagues ever attempt.

But if an allegation of research fraud by a graduate student in Gilbert's lab--a charge currently being investigated by the University and, perhaps, the federal government--is true, it's possible that the professor's two-page warning may not be enough to stop deception.

What's more troubling than that prospect, 10 professors interviewed by The Crimson say, is that there's little more they can do to stop it.

"I think that's a real problem," says one Harvard biology professor, who spoke only no condition of anonymity. "In some sense, it's the professor's responsibility to check. But how do you do it, unless you're going to repeat every experiment in the lab?"

Professors soya vigilance is largely a question of time and energy. And because of that, higher-profile professors like Gilbert, who frequently have larger labs and less free time to police them, may be more susceptible to research fraud by their lab workers.

"There are very different styles that depend on how large a group it is," says Assistant Professor of Biology Markus Meister. "If you have a larger group on the order of 20 people, you have to adopt a different style."

But many professors, according to some researchers, can be lulled to sleep by the infrequency of allegations of fraud, which happen as rarely as once per decade. Reproducing the experiments of researchers also takes costly time.

"It's a question of your time horizon," says one biology professor, who spoke on condition of anonymity. "So when do you reproduce [experiments]--a week from now, a month from now, a year from now? A year later, it's an effort to go back. Some people don't want to make the effort."

Fear of research fraud, however, has led some professors to establish a more hierarchical structures inside their laboratories. This way, researchers with Ph.D's check on graduate students seeking Ph.D's, who check on undergraduate researchers.

In addition, by working on complimentary projects that build on one another, results from other experiments are automatically checked as the research progresses.

"The way you feel secure is that you have people working on complimentary projects so they're checking on each other," says Raymond L. Erikson. American Cancer Society professor of cellular and developmental biology. "If each experiment builds on others, there's a built-in check mechanism."

But some professors say checking can go too far, and ruin the atmosphere of a laboratory. A laboratory notebook should be a science aid, Meister says, not a tool to prove one's innocence.

"I don't think I would recommend to someone that they maintain a notebook to withstand scrutiny by the Secret Service or forensics experts," Meister says.

"Any lab has to be run on the basis of trust," says Vleck Professor of Pure and Applied Physics Paul C. Martin '52, a member of the faculty committee on professional conduct.

But trust and hierarchy do not satisfy some. Four professors interviewed by The Crimson said they believed that every professor should have personally supervise and check every experiment. Each, however, admitted that they themselves could not always live up to that standard.

"A professor sponsoring research is obligated to have hands-on contact with the data," says one biology professor who spoke on condition of anonymity. "But this contact is not common."

In the end, professors say, the possibility of research fraud is just another reality of scientific research, much like test tubes, gloves or laboratory notebooks. Researchers bent on deceiving their boss usually can.

Says another Harvard biology professor: "A real professional who wants to cheat will pull wool over anybody's eyes.