A side return in Dartmouth Park is the narrow strip of external ground running alongside the rear closet wing of a Victorian terrace, between the wing's flank wall and the boundary with next door. Infilling it is one of the few ways to widen a narrow terraced plan without pushing further into the garden. The first thing we draw is not a footprint but a measured survey with levels, because on this kind of ground a return cannot be assumed to behave like a flat one. Dartmouth Park's terraces sit within a hillside grain of sloping streets between Highgate, Kentish Town and Hampstead Heath, and where that fall runs across a plot it can pass straight through the return. Its true width, its length beside the wing, and the difference between external ground level and the internal kitchen floor are the three numbers that decide what the infill can actually be.
Once the levels are known the infill starts to design itself. Where the ground in the return sits above the internal floor, the new external wall and the reworked boundary edge take on a retaining function, and the design has to deal with lateral earth pressure, a reliable damp-proof line and drainage behind the retained face, rather than simply a wall on a strip footing. Where the ground falls away instead, the floor build-up has to be made good and the threshold to the garden becomes a step rather than a flush junction. Neither condition is a problem in itself, but both change foundation depth, waterproofing and the junction with the original solid brick wall — and all of that is far cheaper to resolve on paper than after the return has been broken out.
The return is also where a Victorian terrace keeps its plumbing. A rainwater downpipe from the main roof, the closet wing's own gutter and downpipe, a surface gully and, frequently, the soil stack from the bathroom above all live in that strip, and an infill builds over every one of them. There is no single answer: some runs can be re-routed clear of the new footprint, others have to be taken internally within a duct that retains access, and any drain connecting to a public sewer may require a build-over or build-near agreement with the water authority before work starts. On a cross-falling plot, the surface water that currently drains down the return has to be given somewhere new to go. We settle drainage before the plan is fixed, not afterwards.
Opening the flank wall of the closet wing is the structural heart of the job. That wall is loadbearing: it typically carries the wing's roof or the floor above, and often a chimney breast as well, so replacing it with a clear opening means a steel frame with beam and posts, padstones sized for bearing into soft lime-bedded Victorian brick, and a temporary propping sequence that keeps the wing supported while the wall comes out. The existing footings under a terrace of this age are usually shallow, so the new foundation for the infill has to be designed against what is actually there rather than assumed to match. Where a chimney breast is removed the flue above it must be supported. None of this is exotic, but it is what a purely spatial sketch tends to ignore.
A side return infill is a two-boundary problem, which is why the Party Wall etc. Act 1996 almost always engages. The strip is bounded on one side by the flank of your own wing, on the other by a boundary wall that may be a party fence wall or the neighbour's own structure, and at its far end by the neighbour's closet wing. Building on the line of junction is notifiable; so is excavating for foundations within three or six metres of an adjoining owner's structure, depending on depth. On a cross-falling plot, foundations are often deeper on the downhill side than a level-ground rule of thumb would suggest, which can bring the deeper six-metre test into play sooner than owners expect. Notices, awards and surveyor appointments take time, so we identify notifiable neighbours early.
The neighbour across the boundary matters more on a side return than on almost any other extension, because where a flank window serves their kitchen, utility or stair, their own return may be its only source of light. Building a full-height wall hard against the boundary for the entire length of the return can measurably reduce the daylight reaching it, and Camden will assess that using the established BRE tests. The usual moves are to keep the infill roof low, to step or set back the rearmost portion, and to stop short of the neighbour's own rear wall rather than running the full depth. Right to light is a separate civil matter that a planning consent does not resolve, so where an infill approaches a neighbour's window we flag it early rather than let it surface late.
Many owners want the return infill and a rear extension as one wrap-around move, and that is usually the right ambition — but the two parts are judged on different things, and the return is the part carrying the boundary, drainage and party wall risk. The junction between the two roofs, and the point at which the infill stops relative to the neighbour's rear wall, do more to determine whether the scheme is defensible than the rear projection does. It is also why a precedent borrowed from a flat London street, or even from a nearby street with a different fall, can mislead: a design that worked where the return was level can fail on a cross-falling plot, where the same footprint sits a course or two out of step with the ground it stands on.
Before any of this we confirm what is actually being extended. Dartmouth Park's stock includes maisonettes and converted flats as well as whole houses, and where a ground-floor flat has use of the rear garden the return may or may not be demised to it — a freeholder's licence to alter, and the leaseholders above, sit alongside planning rather than behind it. From there the route is ordinary and methodical: a measured survey with levels, a drainage investigation, a structural scheme, planning drawings for a householder application to Camden where one is required, and then the building regulations package covering structure, fire safety, insulation, ventilation, drainage and waterproofing. We do not promise approval. We set out the constraints honestly and design something that answers them.
