[ Page 1 of 3 ]

by Charles Linn, FAIA

In 1890, Seth Low, Columbia's president, championed the relocation of that small college from Midtown Manhattan to Morningside Heights far to the north. When donors balked at the expense, Low decided to lead by example. He pledged $1 million to build a library to anchor the new location. The building would honor his father, a tea trader and owner of a fleet of clipper ships. Low Memorial Library is the centerpiece of McKim, Mead and White's 1894 master plan and is still regarded by many as the most prominent building on campus. Charles McKim designed the library to be constructed entirely of masonry, without a steel superstructure. It is Greek-cross shaped in plan, with a central rotunda covered by a pair of domes that nest one on top of the other. The inner dome is made of plaster on iron-frame-supported metal lath. The outer dome, whose drum is supported by monumental arches and pendentives, is sheathed with limestone panels. Until recently, the composition of its supporting structure had gone unstudied.

It is little wonder that the administration of Columbia University considers its collection of 24 McKim, Mead and White buildings to be one of its greatest assets. In recent years, more than $170 million has been spent to repair and preserve these architectural treasures. When attention turned to Low, a study of the stability of the dome was a priority. But the structure is complicated. University architect Irwin Lefkowitz, AIA, admits, "We didn't know where to start." When Helpern Architects was hired in 1999 to create plans for modernizing Low's infrastructure and renovating parts of the building, Helpern partner Margaret Castillo, AIA, and consulting engineer Ed Meade, of Robert Silman Associates, found themselves in the midst of an architectural-engineering detective story.

A photograph of Low Memorial Library taken not long after its completion gives an indication of the dome's overall size. Millions of measurements taken using a laser surveying instrument were transformed into AutoCAD models (above right) and construction drawings (below).

"We had no idea what was up there," says Castillo. "We had seen written accounts of it and drawings, but all of them turned out to be inaccurate." The only indication on Low's final ink-on-linen working drawings was a pair of concentric arcs from which the dome's location and proportions could be inferred. Construction photographs were not as helpful as the architects and engineers might have wished. One image shows construction on the stone inner portion of one of the arches when it was just finished, with the centering still in place. The next in the series shows the dome all but finished.

ADVERTISEMENT

While searching through boxes of correspondence written while the building was under construction, Castillo discovered that the decisions about the outer dome's construction were made in a fashion that would be called "fast-tracking" today. She found correspondence between the contractor, Norcross Brothers of Massachusetts, and the McKim, Mead and White office that revealed that even as the great arches and pendentives that would support the exterior dome were under construction, exactly how it would be built was still being discussed. Guastavino tile was actually shown on drawings discovered by Castillo, but according to a letter she located, Norcross declined to guarantee Guastavino's work. Another proposal drawing showed that thin-shell concrete was considered, but by the time the city building department had approved it, winter weather had set in, making it too cold to execute such work.

When Castillo and Meade began their initial visual survey of Low's upper reaches, university workers led them up a series of narrow circular staircases into a huge, pitch-black attic space between the inner and outer domes. There, the flashlights on their miner's helmets revealed that the outer dome was actually made of brick, not the steel framing that they had been led to expect by drawings that have appeared in reference materials about McKim, Mead and White buildings. Evidently, the masonry was laid up over wooden centering which had been covered with roofing felt—fragments of it were still stuck to the mortar inside the dome more than 100 years later.

Learning that the dome was brick encouraged speculation about a number of other issues. Was this a wholly masonry structure, or were metal reinforcing or other layers of material being concealed inside it? If metal reinforcing or anchors had been used to reinforce the dome, was it corroding and therefore prone to failure? How much did the dome move over the course of a year, and above all, was it safe? There had been cracks in the arches that support the dome for as long as anyone could remember, and the cast-iron window frames under the east arch had been deformed for decades. "What was alarming," says Castillo, "was that there were reports of sand coming down on the people who occupy two of the four balconies under the great arches. Why would that be happening after 100 years?"