of the laser made experiments easier. More important in the
long run, the laser opened a new horizon for holography that
mercury lamps couldn’t reach—recording fully three-dimensional objects.
Strictly speaking, photographic transparencies are three-dimensional because they have a finite thickness. The reference
and object beams in the first Leith-Upatnieks laser holograms
traveled nearly identical paths, so, with sufficient care, filtered
mercury lamps were coherent enough to record holograms. In
principle, lamps might even have been used to record holograms of thicker transparent three-dimensional objects.
However, most three-dimensional objects are opaque, and
opaque objects required more coherent light and posed a much
bigger challenge. Leith and Upatnieks spent a couple of days
in July 1963 trying to make laser holograms of reflective 3-D
objects, but they failed. Busy with other work, they didn’t
return to 3-D holography for months.
That return was largely in response to reporters who called
Leith after their first report on laser holography appeared in the
December JOSA, asking what might come next. “He offhand
mentioned that 3-D objects could be recorded and they would
be three-dimensional, and no one believed it,” Upatnieks recalls.
“Since Emmett said it would be done, we had to show it,” and
they went back to the lab, determined to succeed.
Officially, they were still working on the radar project.
Transparent holography offered insight into ways to improve
optical processing of radar data, but 3-D holography did not.
Fortunately, it was a time before military contracts tightly
directed research. “The subject was interesting, and the
sponsors weren’t picky if you didn’t stay exactly on the task,”
Upatnieks says. The university also provided some support for
the 3-D project.
They faced tough technical problems. Early lasers often
drifted into multimode operation, with coherence length too
short for 3-D holography. Stability requirements were much
more stringent because vibrations of reflective objects changed
path lengths much more than those of transparent objects,
making it hard to record holograms. They tried to record
holograms of simple objects laying around the lab, such as an
aluminum block, a microscope objective, and wire figure, but
the results were poor. They needed to isolate their optics from
wavelength-scale vibrations that blurred interference patterns.
Fortunately, they already had a massive granite optical
bench, which they had been using to build a Mach-Zender
interferometer. Recording holograms on the granite bench
greatly improved image clarity, but they were disappointed
that they didn’t see the more dramatic difference they expected
from a truly three-dimensional image. Then they realized the
problem: They were recording images of objects only about an
inch across, too small to give enough parallax for the eye to see
the image as three-dimensional.
Recording holograms of larger objects on 4- by 5-in. plates
showed what they wanted—something completely new. The
reconstructed image looked exactly as the object appeared
when illuminated by laser light. “It was incredible, just totally
incredible, the one thing that excited us most,” Leith recalled.
Their first image was a pile of loose objects they had collected in the laboratory; it looked like a pile of junk, interesting
only because it was a hologram. But as they refined their technique, they came upon an iconic object that made a striking
hologram—an HO-gauge toy train that they borrowed from
a technician. Eventually, the lab paid him for it, and Leith
and Upatnieks filled it with epoxy and glued it to the tracks to
stabilize it. They recorded more holograms, refining their technique and testing other off-axis configurations for combining
the reference and object beams. They found they could record
two holograms on the same photographic plate mounted at different angles, and then reconstruct the two holographic images
separately without crosstalk by illuminating the plate at the
Visitors streamed through the lab to see the holograms, but
the floodgates opened in April at OSA’s 1964 spring meeting.
Upatnieks presented a 15-min. paper on Friday afternoon, the