In general, the new roads followed long-established “desire lines” between cities, serving the most populous towns in between. Those corridors, of course, were often constrained by the need to traverse mountain ranges using particular passes, or to cross rivers at narrow or shallow spots. Thus, it wasn’t unusual for the new Interstate to follow the same path used by the predecessor US highway—which in turn might have followed a path used by a predecessor railroad or wagon road. Those, in turn, had often found the easiest way through the landscape by following Native American trails or animal paths.

Many big cities and state highway departments were planning and even building superhighways in the 1930s, but World War II slowed that quite a bit, except for projects like the Willow Run Expressway outside Detroit that were useful for new defense plants. But after the war, the states and big cities were anxious to get the planned superhighways under way. Cities like New York, Chicago, and LA cobbled together state and local funding, while many states turned to toll financing for corridors where traffic projections were high enough. By 1957, about 1.5 percent of "federal-aid" and toll highways were already superhighways; today—though the statistics are not strictly comparable—about 19 percent are. It took several years for Congressional leaders to hammer together the financing scheme for the full national network that we call the Interstate System, and the 1955 addition of urban routes to the network was a way to attract interest from big-city mayors and congressmen, who wanted to prevent further abandonment of central cities and portside industrial areas.

Because large volumes of intercity traffic had been moving on the nation’s highways since the early 1930s, it was easy for state highway departments to see the intercity corridors that had the most demand. Based on this state data, the 1955 "Yellow Book," General Location of National System of Interstate Highways Including All Additional Routes at Urban Areas, showed the Interstate network’s general routings almost exactly as built between cities, and this book is what was given to all members of Congress. Of course, exact alignments weren't shown and the state highway departments (which determined actual routings) faced many decisions about geographic barriers, costly takings, and the safest alignment of the roads—both during and after construction. In much of the West, the previous two-lane US highway became one half of the finished four-lane superhighway, though kinks and grades that didn’t meet the Interstate geometric criteria often had to be bypassed or modified. Especially delicate alignments, such as I-280 on the San Francisco Peninsula, or I-70 through Colorado’s Glenwood Canyon, were often the most costly and last to be opened.

In urban areas, freeway alignment proved to be a much more contentious issue. Urban superhighways are the most expensive parts of the network to build, and—before electronic tolling—were very difficult to design as toll facilities. The pre-Interstate "turnpikes" built in America—Pennsylvania, New Jersey, Illinois, Oklahoma, Maine, Florida, etc.—were cross-country facilities that didn't even attempt to enter big cities, with a few exceptions like the Massachusetts Turnpike.

The Yellow Book alignments were very general, and states sometimes folded in urban freeways already planned or under construction. Finding an alignment through dense urban neighborhoods is inherently problematic, as seen in Robert Moses’s quote that “You can draw any kind of picture you want on a clean slate . . . but when you operate in an overbuilt metropolis, you have to hack your way with a meat ax.” In many cities, the urban superhighway era coincided with slum-clearance and urban renewal efforts. Though that meant the highway wasn’t technically the proximate cause of residential displacement, the result was the same. In other places, the cheapest and most direct route was alongside an existing railroad or watercourse—which for other reasons was often the location of lower-class housing. Some cities sought to avoid condemnation of homes by routing the new highways through parks, as in St. Louis, Philadelphia, San Antonio, or Memphis. The “freeway revolt” of the late 1960s began with concerns about harm to specific inner-city neighborhoods or parks, and eventually became part of more general 1970s concerns about the environment and about who makes decisions regarding disruptive public works projects. The result was more federal restrictions on routing, but also a “trade-in” program allowing big cities to use earmarked Interstate funding instead for transit improvements. Some urban segments, such as I-40 through Memphis or I-95 through Boston, were simply never built.

There’s a widespread popular conception that the network was largely designed for military or civil defense purposes. However, throughout the years of congressional debate, military strategists repeatedly testified that they didn't need any particular routes or geometric specifications, always saying that highways built to promote commerce would also serve their needs. The words "and Defense" were added to the name of the "National System of Interstate Highways" in conference committee, almost as an afterthought, and played no role in congressional voting. See Congressional Record 102, Part 8, pp. 10991-10997. FHWA has a page on “Interstate Myths.”

I like Earl Swift’s The Big Roads as the most readable history of the Interstates, but Mark Rose's Interstate: Express Highway Politics 1939-1989 is the more scholarly source on the congressional machinations. America’s Highways 1776-1976, a Bicentennial project of the Federal Highway Administration, offers some good history on highways generally, and how the Interstates got built. A more recent book by Joseph DiMento and Cliff Ellis, Changing Lanes: Visions and Histories of Urban Freeways, recounts the history of how traffic engineers and city planners wrangled with this new force reshaping the city.