No College official would reveal any information about the actual structure of the three curves, but Glimp disclosed that "It's better to rank first in a small class than a little way down in a larger one. Apparently being first in one's class-no matter what the size-indicates some quality that enables a student to do well at Harvard."

One curiosity of the formula is the omission of the College Board mathematics aptitude score, considered very important by most high school advisors. Whitla said that the score is left out because experiments have shown that including it does not increase the PRL formula's ability to predict.

Glimp explained that math aptitude shows up significantly in achievement scores and in rank in class. Hence the mathematical capabilities of the Harvard applicants are indirectly included in the PRL formula.

Nevertheless, the omission of math aptitude is a reversal in emphasis from the testing policies around 1959, at the height of the satellite race. That year the testing office drew up a completely different PRL formula for scientists because, Whitla said, "We felt the present PRL formula might be unfair to them."

For a few years thereafter, the College computed two PRL's for some students, but when the science formula proved less accurate than the one already in use it was junked.

"The figures have shown that it takes a high verbal aptitude to be a good scientist at Harvard," Glimp said.

The Office of Admissions sees all the computed PRL's, and the ratings for the applicants selected are given the following year to freshman advisors. Senior tutors get the student's PRL after he has chosen a House.

Advisors are interested in PRL largely out of curiosity. Many say it's a fantastic game, though most also contend that the figures are in some vague way useful for advisory purposes. One freshman advisor said he would dissuade a student with an unusually low PRL from a taking a fifth course.

Standish Meacham, Jr., former Winthrop House senior tutor, said, "If a student's doing poorly. I want to find what's wrong, and PRL is meaningless for that." Grinning, he added, "But I do look at it now and then just to see if a student is above or below what they thought he'd be."

PRL is more meaningful for admissions purposes and Glimp too has a weakness for playing with the number. "Artists have PRL's lower than the Harvard average, while musicians are way above the mean. I think that's sort of an interesting fact." he said.

Glimp warned that although PRL is a "great aid" to the admissions process, "We must remember that it's only one of many things we're looking for. There are a lot of guys more interesting and more desirable than the potential summa. But if you get somebody with a 1.6 who looks like he'll hang together for four years, you've got to take him.

Glimp added that the use of PRL in admissions is limited by its degree of inaccuracy. But this error is smaller than that of most similar indices, according to Whitla. "Given the narrowness of the range it's pretty good, but you can see how easy it is for the prediction to be wrong." Glimp said. "PRL can't consider any sort of family problems or emotional pressures or what type of courses the guy's going to take at Harvard."

"So what makes our job fun is deciding which PRL's to believe. For instance the boy in the class of 1964 with the lowest PRL graduated magna cum laude. We didn't believe his PRL and so we let him in."

"We expect a case like that every now and then," Whitla said. "After all, if PRL was right on the nose every time, there wouldn't be any sense in grading people after they got here, now would there?"