A canal does not require nearly so much technology. So
the barrier to entry is less.
For example:
-- The Grand Canal in China was started circa 400 BC
and reached a length of 1800 km circa 500 AD.
-- The first Transcontinental Railroad was constructed
starting in 1863 and completed in 1869. That's quite
a bit later.
The interesting thing is that the Transcontinental RR used
/cast iron/ rails and /cast iron/ components on the engines,
because steel was not available at anything approaching a
reasonable price. There were precisely zero Bessemer
converters in the US before 1864, and the first inch of
steel rail in the US was produced in 1867. During the 1870s
the cost of steel dropped astonishingly, to the point where
it was economical to replace the rails on the Transcontinental
Railroad from end to end, replacing cast iron with steel,
which is dramatically more suitable for the purpose.
People take steel for granted ... but they really shouldn't.
It involves a treeeemendous amount of cleverness, not to
mention quite a lot of capital investment. It's cheap now,
but it wasn't always. Before Bessemer came along, a good
steel sword was worth a fortune.
The advent of cheap steel changed a lot of things.
Even once you have steel, building a locomotive is a
nontrivial undertaking.
This is relevant to this list as follows: Not everybody
with a BS in physics goes on to study hadron/hadron
collisions or gravitational waves. A lot of them do
materials science. And materials science is not just
silicon. There is a lot of steel in the world, and
there is a market for people who understand how it
works.
Tangentially related: Here's a classic piece of collaborative,
interdisciplinary detective work, solving the centuries-old
mystery of Damascus steel. A master swordsmith teamed up
with a materials scientist. I'm impressed, and I'm told by
experts that the more you know about it the more impressive
it is: http://www.tms.org/pubs/journals/JOM/9809/Verhoeven-9809.html