The structural literature settles on five pylon families — H, A, inverted-Y, diamond and vase — plus the single central mast. Below, each one in the three views that matter, with a real bridge you can look up.
In a strict longitudinal elevation — the bridge seen from the side — H, A, inverted-Y, diamond and vase all look like the same single tower. You are sighting along the line of the two legs, so one hides the other. The shape only becomes visible across the bridge or from above.
So BMS makes a deliberate choice: draw each pylon's transverse silhouette at its position in the elevation band. Not strict projection — but the BMS diagram was never strict projection. It already stacks an elevation, plan bands and component tables in one canvas. The pay-off is that a site engineer recognises the structure instantly, and the client sees their bridge rather than a generic post.
The middle column below is what the diagram actually renders. The left column shows what a strict elevation would give you — identical for every row, which is exactly why we don't use it.
Shape is one key in the per-support JSON. Everything else is derived from it — no extra fields.
| pylon_shape | Deck anchor lines | Cross beam component? | Elevation renderer | Plan renderer |
|---|
count × planes) and how many lines the plan view draws. Single Mast anchors in the median — one line. Every other shape anchors at both deck edges — two lines.
The first draft of pylon_shape listed Delta. The structural literature's five canonical families are H, A, inverted-Y, diamond and vase — delta isn't one of them. Page 1 has been updated.
The GAD doesn't show a transverse section, so we can't read the shape directly. What it does show is pile clusters on both deck edges at P2–P5 and a legend entry for "RCC Lower Pylon" — implying a distinct upper pylon above deck level.
Two legs, therefore two deck anchor lines. That rules out Single Mast and nothing else.
H-frame — the most common choice for a highway bridge of this span range, and the safest default to draw.
Request the transverse section sheet. One drawing settles it, and it also gives us the real pylon height and cable count.
Because shape only affects the plan-view band and the cross-beam count, getting this wrong is cheap to correct later — unlike the cable numbering, which is not.