Following the Load Path from Roof to Ground

The members of a pitched roof lean, so the load on them does not travel straight down. Part of it appears at the feet as an outward push against the wall head, and how that push is dealt with is the thing that separates one roof carcass from another. A tie at ceiling level, a collar, or a beam at the ridge are the usual answers, and each hands load to a different place.

This matters at the very first step of the trace. A roof carried on a ridge beam takes load up to that beam and down through whatever supports it at each end, which may be a position with nothing beneath it at the storey below.

Rafters and joists do not usually bear on masonry directly. A plate bedded along the top of the wall spreads their concentrated loads along it instead of letting them arrive at points, and gives them something to fix to. Where that plate is unevenly bedded it does not spread bearing as intended, and where it is not restrained the roof's outward push has nothing holding it.

The relationship runs both ways: the roof restrains the top of the wall, so a roof taken off for repair leaves a wall standing in a condition it was not designed to be in.

From here the trace splits. In load-bearing construction the walls themselves accumulate the load of each floor and of the roof and pass it down through their own section, so the plan of the walls and the plan of the structure are one drawing. In a framed building the load arrives at points instead, which is what allows the walls to be open and carry nothing.

Knowing which you are in changes every later question, because the foundations beneath a line of load and beneath a point of load are different in kind.

A floor delivers what is on it into the walls or beams supporting it, and the whole of a member's load passes through the small area where it bears. How much of the joist actually sits on the wall, and whether the masonry beneath can accept that concentrated load, determine whether the support is real rather than apparent.

The floors are also holding the walls in line. Cut a floor away to form a void, or take out a returning wall to open a plan, and the wall loses what was keeping it upright even though nothing about its vertical load has changed.

Where the supports above are not in the same positions as the supports below, something has to redistribute the load, and that element exists precisely because of the offset. It collects load from columns or walls at one arrangement and sets it down on a different one, and everything above it then depends on it.

Those positions are worth finding because the consequences of anything going wrong there are assessed differently. A transfer element is not simply a larger beam: what arrives on it is concentrated rather than distributed, and the structure it supports is sensitive to its deflection in a way a floor finish is not.

Older buildings rarely carry a record of what was altered and when, and the published masonry entry treats establishing which walls are carrying, what has already been removed and what is currently propping what as a survey exercise rather than something read from appearance.

The useful move is to look for discontinuity. A wall above that does not continue below, a beam end resting where nothing obvious receives it, a chimney breast removed on one storey but not another, a floor opening that arrived later than the floor, a wall built across a room at an angle nothing else follows.

At the bottom the load leaves the structure and enters the ground, and the element doing the spreading is chosen for the shape in which that load arrives. A continuous wall delivers along a line; a column delivers at a point and also needs holding in position. The concrete outside is only half the answer, because what the element can do is governed by what it is standing on.

This is the step where the trace stops being something you can do by looking. Founding level, width and reinforcement all follow from ground that has been investigated on that plot.

A trace is not a conclusion. It is a set of positions you can point at and ask about, which turns a vague question about whether a wall can go into a specific one about what happens at a named place. It also tells the engineer where to look first, which is worth more than any opinion you could offer.

Write it down, ideally with photographs at each level and at each point where the path seemed to change. That record survives the visit and is what a later alteration will be judged against.