Who this guide is for
- Clients coordinating more than one contractor or work package on an envelope
- Self-builders acting as their own coordinator between trades
- Renovators whose extension meets an existing wall built by someone else
- Anyone reviewing a set of details before work on site begins
- Owners trying to understand why a junction was never built as drawn
The failure is contractual before it is physical
The control layer transition record puts the case most directly. Transitions sit on package boundaries by definition, so the exposure is contractual as well as physical, and the recurring failure is an unowned junction rather than a defective one. The planes on either side may both be complete and both perform perfectly across their own area while the assembly as a whole has an open path straight through it.
The damp-proof barrier record calls that the clearest case in the library of a complete component producing an incomplete system. The floor barrier and the wall course only become a barrier through the lap between them, and that lap is formed at the moment when neither trade regards the work as theirs. Both components remain intact; the line does not exist.
- Each side's scope can be fully discharged and the junction still be missing
- A complete component is not the same as a complete system
- The boundary is drawn on a contract, not on a section
- Nobody has to make a decision for this to happen
The places the records name
The list is short and it repeats. A cavity tray at an abutment belongs to the masonry trade while the flashing outside it belongs to the roofer, and the junction between their two pieces of work is inside the wall where neither can see it. The band across a slab or beam edge on an infill wall is described as belonging to nobody by default. The perimeter around an opening is where the window package, the wall package and the finishes package all stop.
The rooflight kerb adds a variation worth recognising, because it is a handover to a manufacturer rather than to another trade. The kerb belongs to the roof, its top face belongs to a manufactured unit, and responsibility changes at a line that appears on nobody's drawing by default. Recording where one scope ends and the next begins is what lets either party see the gap before it is built.
At the wall head, the floor edge and the base, the same pattern holds. The wall records ask which trade joins the wall's layers to the floor's and the roof's, and the answer is rarely written down anywhere that both trades will read.
- The tray inside a wall at a roof abutment
- The band across a floor or slab edge on a framed or infill wall
- The perimeter around every window and door opening
- The lap between a floor barrier and a wall course at ground level
- The line where a kerb stops and a manufactured unit begins
- The wall head, where wall layers have to meet roof layers
Timing is half the problem
Several of these junctions cannot be built by the party who is technically responsible, because that party is no longer on site. The abutment record states it plainly: provision for the concealed tray has to be made while the wall is being built, by a trade that will have left before the roof is weathered. Where it is missed, the remedy involves taking part of the wall apart again.
The control layer record generalises the same point into a question worth asking of every junction. At which point in the construction programme does this junction become permanently concealed? A transition drawn carefully and installed after the trade that buries it has left is a transition nobody has confirmed exists, and the window in which it could have been verified was set by the programme rather than by the design.
Whoever arrives last closes it
When a junction has no owner, it still gets closed, because a building has to be finished. The window opening record describes the outcome exactly: a perimeter with no single owner tends to be closed by whoever arrives last. The same record notes that whichever layer is installed last is the only one anyone will ever see, and that a perimeter is normally judged on the appearance of that layer alone.
That is why a finished junction is such poor evidence about an unfinished one. The visible closure was made by a trade doing reasonable work within its own scope, using what was to hand, in a position where the layers it was meant to connect had already been covered. Nothing about the result looks improvised.
What to ask, and when to ask it
The published design questions across these records converge on a small set. Which package does the junction itself fall inside, as distinct from the planes on either side of it? Who is responsible for the receiving substrate's condition at the moment the work arrives? Who inspects the junction, at what stage, and what record is kept before it is covered? Those are questions for the design team and the contract administrator rather than for the operative on the day.
The right moment for them is early, because the abutment record's own conclusion is that recording who provides the tray is more useful than specifying the flashing metal. A specification can be faultless about the material and silent about the party, and the party is what decides whether anything is installed at all.
- Which package contains the junction, not the planes either side
- Which trade is on site when it has to be formed
- What condition the receiving surface has to be in, and who delivers that
- Who inspects it, and before which following operation
- What the fallback is if a substrate arrives different from the one assumed
Why a test at the end will not find it
It is tempting to assume that a whole-building check settles the question. The control layer record answers that directly: an air leakage test reads the envelope as a whole and cannot say which junction is leaking or whether a bond will last. It is evidence about the building, not about the detail, and it arrives long after the junction has been covered.
The same is true of the visual inspection an owner can make. Almost everything on this list is concealed by the time the building is handed over, which is why the records place so much weight on a record made while the junction was open, and on knowing who made it.
The record that makes it findable later
Every one of these junctions is eventually buried, and the question then becomes what a later investigation can start from. The records ask for the same handful of things: which junctions were considered high risk, what was assumed about each, who inspected them and what was photographed before the covering went on.
That record is also what makes a later alteration possible without repeating the whole exercise. Someone cutting a new opening, adding a support or bringing a roof up against a wall is working into junctions that already exist, and whether they can be judged rather than guessed at depends entirely on what was written down while they were visible.
Questions to settle before the packages are let
- 1List every junction where two work packages meet on this envelope
- 2For each one, ask which package the junction itself sits inside
- 3Ask which trade is on site at the moment each junction has to be formed
- 4Identify the junctions that must be built by a trade who will have left before the next arrives
- 5Ask who provides the concealed tray at every abutment, in writing
- 6Establish who carries the control layers across each floor edge and wall head
- 7Ask who owns the perimeter around each opening, as distinct from the wall and the window
- 8Ask where the roof designer's scope ends and a manufactured unit's begins
- 9Confirm what condition each receiving substrate has to be in, and who delivers it
- 10Fix a stage at which each junction is inspected before it is covered
- 11Agree what will be photographed, and who holds the photographs
- 12Ask what happens if a substrate arrives in a different material from the one assumed
- 13Record which junctions were treated as high risk, and why
Common mistakes to avoid
- Assuming a junction is covered because both sides of it are in someone's scope
- Specifying the material at a junction without naming the party who installs it
- Leaving a concealed tray to a trade that will not be on site when the wall is built
- Treating an air test at completion as evidence that individual junctions were built
- Letting the last trade on site close a perimeter nobody had been made responsible for
- Reviewing details in isolation rather than against the programme that builds them
- Handing over without a record of who inspected each junction and when
When to involve a professional
- Ask the design team to mark, on the details, which package each junction sits in
- Ask the contract administrator to confirm those boundaries against the appointments
- Ask for an inspection stage to be fixed before each junction is concealed
- Ask who provides the tray at every abutment and who forms the chase that retains the flashing
- Ask who carries the control layers across the floor edges and the wall head
- Agree what record of each junction will be handed over, and to whom
Frequently asked questions
Questions readers ask about this topic
Why does an unowned junction get missed rather than built badly?
Because nobody decides to leave it out. Each party completes its own scope to the boundary and stops, and the work between the two boundaries was never instructed to anybody. The result is a gap rather than a defect, which is why reviewing the quality of either side tells you nothing about whether the junction exists.
Which junctions do the records identify most often?
The tray inside a wall at a roof abutment, the band across a floor or slab edge, the perimeter around openings, the lap between a floor barrier and a wall course at ground level, the wall head where wall and roof layers meet, and the line where a kerb hands over to a manufactured unit.
Can an air pressure test tell me whether these junctions were built?
The control layer record states that such a test reads the envelope as a whole and cannot say which junction is leaking or whether a bond will last. It is useful evidence about the building, but it arrives after every junction on this list has been covered and it identifies no individual detail.
When is the right time to raise this?
Before the packages are let, because the question is about scope rather than about workmanship. Several of these junctions have to be formed by a trade who will have left the site before the following trade arrives, so the programme itself decides whether the work can be done at all.
What should be recorded about a junction before it is covered?
The records ask for which junctions were treated as high risk and why, what was assumed about each, who inspected them and at what stage, and photographs keyed to their positions. That is what allows a later damp investigation or alteration to begin from evidence rather than from guesswork.
Keep reading