Just as Caltech makes no concessions to wealth, it won't sacrifice merit for diversity's sake. Caltech struggles to attract minority students, in part because the relatively few blacks or Hispanics who have the preparation in advanced math needed for Caltech likely have their pick of the Ivy League and other elite universities. Only one of 207 Caltech freshmen in 2004 was African American; overall, blacks make up 1 percent of the student body and Hispanics 7 percent. Women, an underrepresented group in the sciences, comprise about 30 percent of Caltech undergraduates, a lower proportion than at MIT (43 percent)—which is precisely why Vicki Loewer, with a 1560 SAT score including perfect scores on the math part and the math SAT II, chose Caltech over MIT.

  “I don't want to wonder why I got in,” Vicki, a chemical engineering major with a 3.88 average and a member of the freshman admissions committee, told me one morning as we sat at an outdoor cafe on campus. Although she was wearing casual clothes, she told me she had dressed up especially for the interview; during exam week, she usually wore blue pajamas everywhere, including to admissions committee meetings. Vicki, whose father is a Baptist minister and whose mother is chief of staff to a federal education official, graduated from a public magnet high school in suburban Washington, D.C.

  “I have excelled here. I found my stride,” she told me. “We have more freedoms than students at other schools. Caltech lets you embrace all your weird, freakish tendencies…. I intend to give lots of money to Caltech when I have it.”

  A Caltech senior in 2005, Vicki wanted to stay there for graduate school, but her department urged her to go elsewhere to broaden her experience. Wooed by Princeton, MIT, the University of Texas, and other premier schools, Vicki chose MIT, which she described to me as a “good runner-up” to Caltech. She was awarded a $30,000-a-year graduate fellowship from the National Science Foundation.

  At her first Caltech admissions committee meeting, Vicki told me, she asked if she was supposed to factor race and gender into her evaluations and was told no. “I would have had trouble taking them into account,” she told me. “It would be horrible if the school was full of stupid women. That's my nightmare. It doesn't make sense for science to be run that way. It wouldn't be logical to let someone in who wasn't that good.” Asked how the admissions committee would respond if asked to admit an applicant because the family might donate to Caltech, she said, “We would throw things.”

  The freshman admissions committee consists of admissions office staff, sixteen students, and sixteen faculty; only about half of the students and faculty show up at any given meeting. Of 2,300 applications from U.S. students in 2005, the admissions staff turned down the bottom one thousand applicants outright and forwarded each of the others to a faculty member and a student for review. If staff, professor, and student agreed that a candidate should be rejected, admitted, or placed on the waiting list, that was final. If they disagreed—as they did on 210 applicants in 2005— decisions were hashed out over three mornings and two afternoons in March by committee subgroups of half a dozen people, including at least one admissions staff member, professor, and student. The admissions office evaluated international applicants by itself, because it was most familiar with grading and testing policies overseas.

  Faculty participation in undergraduate admissions used to be more extensive. For decades, each professor on the committee would travel for two or three weeks each winter to a specified territory, such as the upper Midwest or (a more coveted assignment) Hawaii, and interview candidates there. Back at Caltech, they would advocate for their favorites before the full committee, and usually prevail. “It was a fantastic experience for the professors, and for the teachers at these various schools,” recalled Caltech archivist Judith Goodstein, a committee member in the 1980s. “The teachers understood if they had a shining star in their student body, somebody gifted in math or science, that kid, regardless of background, stood a chance of being accepted.” A number of such applicants interviewed by Caltech professors, Goodstein recalled, later joined the faculty themselves.

  Goodstein does remember one instance of pressure from Caltech fund-raisers—a call from the vice president of development complaining that she had not interviewed a donor's son. “I went back and looked at the kid's folder and it was just average,” she told me. “We did do a courtesy interview, but I don't think I did it. I bristled at the notion that somebody in development would tell me I had to go see a kid. I don't think the kid came to Caltech.”

  According to Professor Michael Hoffmann, who chaired the admissions committee at the time, these faculty tours stopped in the late 1980s because fewer professors were willing to spare time from research. “It was just harder to get somebody to commit as the pressure in the academic world became higher,” he told me.

  Several prominent faculty remain extraordinarily dedicated to undergraduate admissions, none more so than Politzer. When his Nobel Prize was announced in October 2004, Politzer skipped Caltech's press conference heralding the award. Yet the next morning, he was the first member to show up at the freshman admissions committee meeting. Politzer says faculty participation safeguards the quality of Caltech's student body. “Again and again, I find I have a very different perspective than a staff member as to the attractiveness of a particular student,” he told me. “I make a scientific judgment of the merit or aptitude or intensity shown by their work. The staff people, who don't have backgrounds in science, don't have a way to distinguish between endeavors of young people that to my mind might have enormously different levels of intensity, aptitude, and commitment.”

  In one case, he said, the admissions staff wanted to turn down an applicant with a math SAT score below 700. But the youngster's research, which to the admissions reader “looked like a science fair project,” would actually have made “a fine senior thesis at Caltech, except he did it on his own, not with a faculty member,” Politzer said. “I liked him, and I recommended him.”

  IN 1926 Pasadena was one of the richest municipalities in the United States, a mecca for wealthy midwestern retirees. On March 9, in the home of railroad tycoon Henry Huntington, one hundred local businessmen each pledged $ 1,000 a year for ten years to the fledgling California Institute of Technology, which had changed its name just six years earlier from Throop College of Technology. The Associates, as the group was known, helped sustain Caltech through the Great Depression; over the years, members of the group have contributed to more than fifty Caltech buildings, including several dormitories named after them.

  Today, the Associates continue to be a valuable and primarily local Caltech resource. Of its 1,448 members—who now pay either $35,000 all at once or $3,000 a year for twenty years—only 380 are alumni. Like the group's founders, most current Associates didn't attend Caltech, nor did their children. They're southern California businessmen, motivated by the civic pride that Caltech cultivates. In return for their gifts, they can travel overseas with Caltech faculty, are invited to public and private lectures, and enjoy meals, meetings, and celebrations at the Athenaeum, the elegant inn on campus where Albert Einstein and other visiting scientists have stayed.

  “Caltech is so small that it doesn't overrun Pasadena,” said Caltech trustee and Associates member Stephen Onderdonk, explaining the affinity between the institute and local business circles. His wife's grandfather, Fred Alberston, was a founding Associate. “A lot of the local community are members of the Associates,” the Pasadena businessman said. “People use the Athenaeum for wedding and other social engagements. Caltech welcomes community participation.”

  Caltech archivist Judith Goodstein said, “It was so important for the institute to have the names of prominent members of the community be listed as founding Associates. It became quite a successful organization with a glittering array of names well known in LA. There are a serious number of significantly wealthy individuals, leaders of the business or legal world, who seem to really get a charge out of rubbing shoulders with Nobel Prize winners.”

  The Associates and other nonalumni boosters are one key to Cal-tech's prosperity. As at Cooper Union and Berea, Caltech fund-raisers have to overcome several handicaps—a tiny pool of twenty thousand graduates, one-fifth the alumni base of other elite universities; no admissions break for donors; and a student body heavily drawn from public high schools and immigrant and lower-income families. One-third of alumni donate to Caltech, far below Princeton (61 percent), Harvard (48 percent), Dartmouth (47 percent), and Yale (45 percent).

  “From a fund-raising point of view, all those kids are still paying off loans,” said Gary Dicovitsky, the vice president for development. “They don't have a buffer for their own children. They have to save, save, save and can't give until later in life. That's not a dream for a fund-raiser, but I wouldn't have it any other way.”

  Still, its $1.4 billion in assets in 2005 ranked Caltech 18th among private institutions in per-pupil endowment, ahead of MIT, and it has raised more than $1 billion toward a $1.4 billion campaign goal. Besides its nonalumni support, Caltech has also benefited enormously from its ability to groom successful entrepreneurs. These self-made multimillionaires who amply repaid the institute for its role in their fortunes include the late Arnold Beckman, who earned a doctorate in chemistry in 1928 and invented the electronic pH meter and other scientific instruments; 1954 alumnus Benjamin Rosen, chairman emeritus of Compaq Computer Corp.; and billionaire Intel cofounder Gordon Moore, who earned his Ph.D. at Caltech in 1954, one of America's wealthiest men.

  Moore, whose father (a deputy sheriff in California) never finished grammar school, told me that Caltech's $1.4 billion campaign is “pretty tough sledding,” even with his $600 million gift. “We don't have the alumni base to draw on that bigger schools do. We have to appeal broadly to people who don't have a connection to Caltech.”

  Caltech also appeals to the federal government. Its faculty generated $227 million in fiscal 2004 in federal research grants, of which about one-third covers overhead for facilities and administration. In addition, NASA paid Caltech $1.6 billion that year to manage the Jet Propulsion Laboratory in Pasadena, including an “award fee” of $21 million pegged to performance. Compared with research-based funding in its operating budget, Caltech's tuition revenue is so small that the institute has on occasion considered eliminating tuition altogether.

  According to Politzer, Caltech trustees once debated whether to use a major gift to replace tuition. “I think we could have done it, “he told me, “but I've heard there was a trustee who said, ‘This is America: people believe you get what you pay for. Do it, and you will denigrate your value in the marketplace.’”

  Toward the end of my Caltech visit, I saw disturbing signals of what might be called “legacy creep”—impending deference toward alumni and donors in admissions. Dicovitsky, the vice president for development, told me he planned to ratchet up Caltech's fund-raising from parents of current students; such programs, as we have seen at Duke, can lead to favoring applicants whose parents are expected to donate. Bischoff, who was hired as admissions dean in 2004 after holding the same job at the University of Chicago, confessed he is considering changing Caltech's practice of sending the same letter to all rejected applicants, substituting a gentler one for spurned legacies.

  I asked Baltimore whether it would be easier to achieve Caltech's $1.4 billion campaign goal if it offered admissions preferences to the wealthy. The campaign “has stretched us enormously to find new donors,” acknowledged Baltimore, who later in 2005 would announce his retirement. “We have to sell the vision of Caltech, the overriding excellence of everything we do, the participation in great intellectual adventures, the importance of Caltech to the economy. We're doing a much, much harder thing [than other universities], hewing to our idealism while still raising money.”

  Yet he assured me that Caltech would never compromise its standards. “People should be judged not by their parentage and wealth but by their skills and ability to contribute to society,” he said. “Any school I'm associated with, I want it to be a meritocracy.”

  On May 23, 2004, giving his first commencement address as president of Amherst College, Anthony Marx called for America's elite colleges to reclaim their mission as engines of upward mobility.

  “America today—and our great colleges and universities—have hit a wall of blocked opportunity,” President Marx told more than one thousand graduates, family members, and faculty gathered in Amherst's verdant quadrangle. “At our top colleges, only one-tenth of our students are drawn from the poorer half of the population, only 3 percent from the bottom quarter…. If we believe in diversity of class, ethnicity, origin, and interest among our students, then we must also embrace economic diversity.”

  Nearly six months later, a little-noticed session on the same campus demonstrated why Amherst and other premier colleges have so few low-income students—and how the impulse for change crumbles before the imperative to mollify alumni and donors.

  On the Saturday morning of homecoming weekend, before the football game with archrival Williams College, Amherst hosted what it billed as an “admissions workshop” for alumni and their children—a perk denied other applicants. Braving the season's first snowstorm, about forty-five parents and students gathered in Converse Hall to hear a presentation by Katharine Fretwell, Amherst's director of admissions. Fretwell walked the audience through Amherst's admissions process and statistics—the college accepts one-fifth of its 5,500 applicants, expecting 400 to enroll—and asked for questions. A gray-haired graduate in a white sweater raised his hand. “I think this is a question everyone in the room wants to ask,” he said. “I feel a little bad about asking it, but is there an advantage to a family relationship?”

  The director paused to choose her words, recognizing that her answer would be of critical importance to the alumni whose gifts filled Amherst's coffers—and who likely wondered whether the president's ambition for economic diversity would leave their children in the cold.

  “There can be,” she assured the audience, noting that the application form asks candidates about relatives who attended Amherst. “We do communicate with the alumni office as we go through the admissions cycle.”

  Then the admissions director recited a litany of statistics that starkly documented the edge Amherst College gives to legacies. Over the last fifteen years, she said, Amherst had admitted 50 percent of alumni children, as against 20 percent of all applicants. Amherst, she continued, rates applicants on an academic scale from 1 (outstanding) to 7 (unqualified). The college accepts 85 percent of applicants given a 1 rating, with most of the exceptions being impoverished foreign students to whom it can't afford to give financial aid. Legacy status comes into play with candidates given a 2 rating—strong students who have taken challenging courses and scored in the 1400s on their SATs. Amherst admits 40 percent of such candidates overall—but 100 percent of alumni children with a 2 rating.

  As the workshop ended, Fretwell revealed an exclusive entree for the privileged few. She had already told the audience that in the interest of fairness Amherst does not interview applicants; the college lacks the staff to visit all 5,500 candidates in their hometowns, and holding optional meetings on campus would discriminate against low-income students who can't afford to travel. Now Fretwell confided to the departing crowd that, despite the ban on official interviews, she and admissions dean Thomas Parker were always available for informal “conversations” with graduates and their children. All they had to do, she told them, was call the admissions office and identify themselves as alums.

  IT HAS become fashionable for college presidents to grab headlines by promising to admit more low-income applicants. In recent years, Harvard, Yale, Princeton, and the University of Virginia have all expanded financial aid for such students in hopes of boosting their enrollment. But such initiatives should be eyed with skepticism until Marx and his brethren answer the vital question “Who will lose out?”

  Admissions is a zero-sum game. Elite colleges are reluctant to increase their enrollment because their inaccessibility enhances their prestige. Hence, pledges to make room for poor students run up against the reality that a large proportion of slots at these institutions are reserved for the rich. At Amherst College, for instance, alumni children make up 10 percent of the student body. Another 15 percent are recruited athletes who tend to be “wealthier and less ethnically diverse than other students,” according to a 2005 report; the college fields men's and women's varsity teams in upper-class sports such as squash and golf. Thus it's no wonder that President Marx's policies have not matched his inspiring rhetoric. He told me in January 2006 that Amherst planned to enroll an extra twenty-five low-income students a year at most. When I asked if he would consider abolishing legacy preference to make space for more, he responded, “This college is what it is because of close to two hundred years of alumni investing in the place. We value that historical connection.”

  Like Amherst, few colleges are eager to alienate alumni and donors by curtailing favoritism toward their children. Colleges have boxed themselves into a corner by degrading admissions into a fund-raising tool; even if they want to be more democratic in admissions, they're afraid to jeopardize the flow of checks. Nor are colleges eager to reduce affirmative action for minority applicants. Hence, their flexibility is limited; any extra slots for low-income candidates have to come largely at the expense of students who don't qualify for any preferences and already face the highest odds against admission. That prospect would leave the colleges favoring the poor and the rich—and locking out middle-class candidates with outstanding credentials.

  This dilemma appears to be playing out at Harvard. In February 2004, former president Lawrence Summers announced that he was eliminating tuition for students from families earning under $40,000 and reducing it for students from families with incomes between $40,000 and $60,000. (In 2006, Harvard said parents earning under $60,000 would no longer have to make any payment.) Nevertheless, even Summers—whose resignation in 2006 deprived Harvard of its most powerful advocate for socioeconomic diversity—clung to the preferences of privilege, insisting that the legacy boost is “integral to the kind of community that any private educational institution is.”