When I went to the local library (well, Acton Memorial, not Citizens Free) in search of railroad wisdom, I found an Encyclopedia of something or other with a bibliography of 130 books about rail. From that, I made a selection available via the Minuteman system through ILL, including this one. It's been slow going, and this is the first one I've gotten through.
Short review: Interested in trains? This is worthwhile.
Vance's thesis is simple: rail in North America inherited little more than the idea of rail-guided transport from Britain. In Britain, markets and settlement patterns were in place (along with roads to serve them) for a good while before the railroad came along. Rail could be a drop-in replacement/supplement to existing transport. The market for that transport was thus quite predictable so capital could be invested in developing railroads with a high expectation of return. Because it is rail nature that the more you put into the road at the beginning, the cheaper it is to run trains on it forever after, British rail was thus heavily engineered and ran low-power locomotives very, very cheaply. By contrast, settlement patterns weren't even past the Appalachians when rail came to the US (from a non-indigenous perspective). Markets were barely a gleam. People had ideas about where other people would settle down and farm or cut down trees or dig coal out of the ground or transship goods or whatever, but not all of these theories would work out in reality. Lines were initially built as cheaply as possible and redeveloped if sufficient freight and/or passenger markets materialized to support them. More powerful locomotives were needed, among other things.
Also, we have taller mountains and more snow than can be found anywhere in Britain. Not really a trivial factor.
The ramifications of this basic thesis are extensive. Rail came early in and around Boston, but because of its location with respect to New York State and the rest of the country, Boston investors never successfully connected their city directly to the rest of the developing national network, but instead leapfrogged by investing in Michigan lines (among others). Rail also came early to the South, but for a variety of reasons (isolation of lines from each other by geography, politics, gauge, etc.; isolation of the south from the rest of the country due to economic structure, location of rivers, etc.; the B & O being physically located in the South but loyal to the North during the War) did not develop extensively until after the Civil War. Gauge choices bring to mind more modern standards wars. Pennsylvania's extensive and almost-successful canal system meant rail developed later there than it might otherwise have done; the Erie canal also distorted rail development from New York State. And that whole St. Louis v. Chicago thing? It might have been illuminating if _any_ of the analyses I have read in the past might have mentioned that St. Louis adopted the Maine-Canadian gauge for the railroads radiating from it.
Think long and hard before fighting standard gauge. Long and hard. And then don't do it.
Rail went through a long period of what I would call (Vance never does, as near as I can tell) thinking of itself as an artificial river or drainage. The idea seemed to be that you built a railroad and you got to own the markets it touched. There were two ways to make money off of a railroad: by building it (and getting whatever you could from the subscribers who put up capital, as individuals, corporations or governments) and then selling out, or by running it. People who wanted to make money by running a railroad weren't the kind to build the first, speculative line somewhere -- then tended to be the second through nth over the mountains or across the continent or whatever. The markets served by these monopolists often got over their gratitude, and turned with open arms to another line developer who brought competition and a reduction in rates to their town(s).
Huge chunks of the book are devoted to extremely detailed analyses of surveys done for the first transcontinental railroads in the US and Canada, and then followup on how the first one was decided upon, financed and built, and then still more on how the rest of the prospective routes were eventually also developed -- and precious few additional alternates were ever found.
Much of both the fun and the frustration of the book comes from the detailed description of terrain. It's fun, because I've ridden Amtrak up and down the West Coast, and I've ridden Via from Vancouver to Alberta, and I've ridden the Acela down to Newark, and I've ridden some of the commuter rail in and around Boston. Thus chunks of the book explain things I've seen and experienced -- and answered questions I had, or at least appreciated the answers to had I thought of asking them. I've driven around large chunks of the US, and hiked a bit here and there, and that gave me a little insight into a bit more of the geography. Stories from my Canadian relatives stretch that geographic knowledge a tiny bit further. But that leaves a vast amount of territory I am sadly lacking in experience of; those descriptions are just name after name with nothing to tie them to. The more geography you know, the more you're likely to get out of this book.
A lot of Vance's perspective as a historian is that other historians would do well to pay more attention to geography: physical geography, economic geography, etc. He makes some pointed remarks about people asserting nonsense they would never have emitted had they bothered to look at a map. I don't see a lot to disagree with him there.
There is only a limited amount in this book about technological developments in locomotives, bridge building, blasting and so forth. Vance is focused on the development and operation of rail networks. Remarks on signaling, bridge building developments, how unit trains work, catenary and so forth are strictly sidelights on the main show, which is where did the line go, what did it connect to, and (when) was it cost effective to run it.
_The North American Railroad_ is well-represented in library systems, but neither cheap nor easy to acquire (even used). I do recommend it. Its format is like a textbook (oversize, lots of maps, pictures, etc.), which makes the seemingly limited page count (323) deceptive. There's a lot here to digest. It would almost certainly reward rereading, altho I am unlikely to do so in the near future.