The sun-filled formal dining room features floor-to-ceiling windows and an elaborate chandelier.

Once Lombardi gained possession of the Octagon House, the most urgent task was to stabilize and raise its sagging dome. Consulting with engineers, he devised a way to squeeze it using high-tension steel cables and turnbuckles while jacking it up. Miraculously, the technique worked: Over two years, the dome gradually moved back into place.

Another crucial step was to find out what the house looked like before time and various owners took their toll. By digging through old public records, Lombardi learned that a more modest eight-sided home had been built on the site in 1860 for its first owner, New York City banker Paul J. Armour. That structure rose only two stories, did not have a verandah, and was topped with a flat roof. Few traces of that house remained; in 1872 it had been almost completely transformed for tea merchant Joseph H. Stiner into the exotic structure Lombardi wanted to save.

Hoping to get more information about the Stiner house, Lombardi attempted to locate his descendants but found few. At least five people—including an insane woman, Jessie Clarke; a Columbia biochemistry professor, Dr. Erwin Brand; an adman, John P. Cunningham; and the writer Carl Carmer—lived in the Octagon House after Stiner moved out, but their stories revealed nothing about the 1872 house, not even who conceived and built it. "That was a big frustration," Lombardi says.

Lombardi recovered the original Renaissance revival bedstead and dressers.

The best clue Lombardi turned up was an 1882 photo of the home's exterior that had been stashed away at the New York Historical Society. By blowing up the negative, he was able to get a clear view of the home's missing pieces, like dormers with curlicued scrollwork and a pair of curved staircases flanking the entrance. (The helpful photo, which Lombardi calls "the Bible," now hangs in the Octagon House library.) Lombardi hired a French carpenter living in Michigan, Serge Dupin, to re-create the lost pieces based on the photo and surviving fragments.

An analysis of the paint then revealed the exterior's rainbow of colors—a "wonderful surprise," Lombardi says. He found no similar photos of the interior, so he again used paint analysis to understand what that part of the house looked like.

A view from the dome?s fourth floor.

The first round of paint analysis, conducted at least 20 years ago, picked up some of the colors from 1872. But an analysis begun in 2003 by Jablonski Berkowitz Conservation, Inc. of Manhattan, using powerful microscopes previously employed only by museums to study and restore paintings, detected more complex colors and designs. "The technology is so much better than 20 years ago," Lombardi enthuses.