Travel Reference
In-Depth Information
workstations had made his computing center the rival of small software cor-
porations. Even more impressive, Tom actually knew how to use them.
But how to test the strength of an ammonite? How to test the hypoth-
esis that increasing septal complexity made the shell stronger? One could
make exact molds. But that would require many weeks of skilled artisan
work, and even then it would be impossible to determine whether the artifi-
cial ammonites were responding to pressure in way faithful to the original—
whether the artificial shell was actually of even strength and was failing be-
cause of material rather antique design properties. Test an actual specimen?
As beautiful as many ammonite fossil are, it is doubtful that any have sur-
vived the minimum of 65 million years entombed in sediment since their
last demise without undergoing mineral changes that would compromise the
delicate balance of shell strength. These intractable problems had stymied
all past investigators. Tom Daniel proposed taking a new tack: Build the am-
monite on the computer, and then subject it to simulated pressure.
Just before our chance arrival (a visit rued by all at times over the next
4 years), Daniel had bought the site license for a new generation of engi-
neering software program. One of the most powerful of all engineering pro-
cedures is called finite element analysis, or FEA. FEA works by subdividing
complex structures into small elements and then monitoring forces that act
at the junctures, or nodes, of each of these elements. It thus embodies the
time-honored method of attacking a large and difficult problem one small
step a time. Any structure is broken up into thousands of tiny regions, and
the stress acting on each region is examined separately. A map can then be
drawn to show how forces acting on the entire structure are encountered.
FEA had long been the province of those blessed with large mainframe
computers. But the computer revolution of the past decade had packed com-
puting power into very small boxes. The Silicon Graphics workstation had the
processor and enough memory to construct a charging
Tyrannosaurus rex
—or
to build an ammonite shell, node by node, and then subject it to pressure.
And so we began building ammonites by computer. Tom's first efforts
were crude, but just as evolution starts simply and builds up, so too did the
virtual ammonites become more and more elegant. And even our early ef-
