Puzzled looks and hesitant steps are the norm for first-time visitors arriving on the Mall by way of the Smithsonian Metro station. These reactions are understandable. Indeed, they are predictable.

As the visitors emerge from the ground, they are provided with little ceremony and less information. The transition from the vaulted subterranean cavern to one of the world's great urban open spaces is, to put it mildly, low key. The dinky information kiosk looks like what it is -- an insufficient afterthought. The Metro pylon gets lost in the sea of space. The trampled ground seems a symbol of past meanderings on the part of bewildered guests.

Surely something should be done to improve this sorry introduction to the nation's number one public park, but the question of what to do is by no means an easy one. As part of the research for this column, I helped to judge a competition called "A Metro Portal for the Mall," and I came away from the experience with an enhanced appreciation of just what a difficult, delicate mission it is to build anything at all on or near the Mall.

The competition was open to students from seven local architecture schools, and each student or student team had just one week to produce a solution from the time they picked up the competition guidelines -- a daunting task. Not surprisingly, no individual or team came up with the perfect, buildable idea. Architect Joseph Passonneau, a fellow juror, consoled the contestants in a tough-minded way by telling them that even after many years spent studying the urban design complexities of the Mall and the city around it, he doubted his own ability to come up with an apposite solution in a week's time.

But he also said, "In architecture there is a right way and a wrong way." In many respects the hours we jurors spent poring over the student boards duplicated the efforts the students themselves had made to decide what would be architecturally right, and what wrong, in this crucial location.

The idea originated with Profs. Jaan Holt and Gregory Hunt of the Washington-Alexandria Center of the College of Architecture and Urban Studies of the Virginia Polytechnic Institute, charged this year with organizing the competition. "After the King Street Metro station opened, we began riding the trains," Hunt explained, "and it prompted us to think about the ways the underground transportation network connects to the city." They selected the Smithsonian station, he said, because both the need and the challenge were great, but the notion was to stimulate students and perhaps even Metro's management to think about improving other stations in the system.

In truth, despite the beauty and monumental presence of Harry Weese's coffered vaults underground, the points are few and far between where Metro connects with the city in visually pleasing or functionally interesting ways. There are the busy activity centers at Rosslyn, International Square and the Connecticut Connection; the stunning escalator exit at Judiciary Square (beautifully aligned on the north with the central gable of the National Building Museum); and the exhilarating passage toward the oval-framed sky at Q Street north of Dupont Circle. But basically the system was designed as an efficient, pleasing people-mover. Making an event out of the transition from underground to the city and its buildings clearly was not high on the list of its designers.

Of all the stations, excepting the fiasco of the non-connection at National Airport, the Smithsonian is the biggest missed opportunity. Functionally, it fails to provide what many of its users need most: information and orientation. Physically, it is simply an unprepossessing hole in the ground in a location that obviously calls for more.

Competitors were given great leeway -- more, probably, than designers in the real world would get. They were not permitted to fool with the alignment or the form of the existing underground vault, but beyond this there was almost total liberty. They did not have to consider the impressive overlapping powers of the review boards and institutions governing changes on the Mall, they didn't need to think about costs, and, most important, they were invited, "if they consider it appropriate," to "disregard any prohibitions against the building of structures on the Mall."

Predictably, the proposed solutions divided into two principal categories: those that took this bait and those that did not.

Among those that did -- that is, among those foreseeing some kind of building or landscaped element in the central greensward -- there were great ingenuity and variety. Two contestants could not resist the almost irresistible temptation to bring people out of the ground in the exact center of the greensward, so that either the Capitol or the Washington Monument would be their first astonishing sight. Another entry proposed a sunken court marked by pristine geometrical forms (shades of Gordon Bunshaft's original proposal for a cross-Mall sculpture garden for the Hirshhorn Museum). One tour de force suggestion was to summarize, more or less in miniature, the landscape history of the city, from Pierre L'Enfant to Andrew Jackson Downing to the McMillan Commission plan. Several proposed vast underground programs (for transportation museums, electronic orientation galleries and so on) linking the north and south sides of the Mall. Quite a few proposed some kind of vertical marking in the center of the greensward, from the skinniest needle to a mighty obelisk to a "Parabolic Fresnel Lens Transducer/Diffuser" looking much like a radar antenna in the middle of the Mall. None of these projects won prizes.

First prize (submitted by Stephen Wilczynski and Lila Snider of the University of Maryland) went to a scheme that kept the Metro precisely where it is, and quite simply improved it by allowing for the differing needs of commuters and tourists. Second prize (to Deborah Auten of the University of Maryland) was awarded to a proposal that, likewise, kept the station where it is, and improved it with straightforward gateways and a slight realignment and redesign of the escalator corridor. Third prize (to Norris McLeod of the Cornell Center in Washington) was given to an elegant building, beautifully drawn and based upon Beaux Arts precedents.

The jurors -- Passonneau, myself, Kent Cooper of the Cooper-Lecky Partnership, Robert W. Nordstrom of Harry Weese Associates and E. Michael Vergason of Edaw Inc. -- did disagree at the beginning about what was permissible and what was not. By day's end a consensus was clear, at least to me: Although something does need doing at the Smithsonian station, the simpler, the more elegant and the more respectful it is, the better. Chalk up another victory for the magnificent Mall.