Open a trench along an ordinary kerb and the wall of it is a cross-section of the city. Within a corridor a metre or two wide sit gas, water, sewer, electricity, telecommunications and, in some cities, heat. Six services, six owners, six sets of records — and no single organisation responsible for the volume they share.

Nora Whitfield reads that wall as a document. The order of the layers is the order in which the city acquired its services: drainage first, because water only goes downhill and cannot be persuaded otherwise; then gas in cast iron, much of which is still down there; then electricity, telephone and fibre, each arriving to find the convenient positions already occupied. Published guidance shows a tidy arrangement with everything at its proper depth and in its proper colour. That drawing describes a street somebody might build next year. It does not describe this one.

What makes the subject difficult, the book argues, is not the physics but the knowledge. A utility drawing records a claim about where something probably is, made by whoever had time to record it, and civil engineers have found it necessary to grade such claims according to how they were obtained — a records search at one end, physically uncovering the pipe at the other. Detection has its own asymmetry: electromagnetic methods need something conductive, so the shift from metal pipes to plastic improved the network and made large parts of it undetectable by the cheap method. Certainty still requires a hole.

Each service then gets a chapter on its own terms. Cast iron gas mains fail through the combination of an ageing metal and a joint never meant to be rigid, which is why ground movement can finish, years later, what a nearby excavation began. Water mains carry pressure, so their failures do not seep but dig, wrecking the street and often the services beside them. Sewers cannot be diverted around anything, which is why they lie deepest and everything else is arranged around them. A buried power circuit is limited by heat rather than voltage, making the surrounding soil part of the design and any neighbouring circuit a rival for the same cooling — and turning a cable rating, quoted alone, into a meaningless figure.

The closing chapters cover the methods that avoid opening the ground at all: curing a resin-soaked liner inside a failing pipe so the old one serves as formwork, fracturing a pipe outward while drawing its replacement in behind, steering a bore beneath an obstacle. All were invented for the same reason — the hole, not the pipe, is what costs money.

Two dozen diagrams were prepared specifically for this book, alongside forty-seven references, chiefly standards, patents, regulator filings and agency reports, listed after the last chapter. Where a frequently repeated statistic could be traced only to somebody's summary of a document, it has been left out and the omission explained.