The Verification Venue · the living litmus that isn't

The Flower That Reads the Soil Backwards

Everyone repeats it: the hydrangea is a living litmus test — blue in acid, pink in alkali. It isn't. A real anthocyanin indicator turns red in acid; the hydrangea turns blue. It runs the wrong way because the thing colouring the flower isn't acid at all. It's aluminium.

Soil chemistry doesn't paint the petal. It only decides whether one metal ion — Al³⁺ — can dissolve and climb the roots. Below is a Hydrangea macrophylla mophead. Drag the soil pH and watch it recolour. Then try to break it: keep the soil acid but lock the aluminium away, and the blue never comes.

soil pH 7.0pink

Aluminium reaching the sepal

— µg/g

band: —

The bloom reads

—

Soil pH — 7.0

Acid soil (below ~5.5) frees aluminium; neutral-to-limey soil (above ~6.5) locks it up as insoluble hydroxide. Watch the bloom cross from pink through purple to blue — the aluminium is what moves, not the acid.

High-phosphate soil Phosphate fertiliser fixes aluminium as insoluble AlPO₄. The pH stays low — but no aluminium reaches the roots. Co-pigment present The blue needs a third molecule — an acylquinic-acid co-pigment. Remove it and no stable blue can form, even with aluminium.

Cultivar

Bigleaf, blue-capable White (no anthocyanin)

The honest twist is in that first toggle. Lowering the pH does not guarantee blue. Flip on high-phosphate soil and the pH stays exactly where you left it, yet the flower snaps back to pink — because phosphate has locked the aluminium away as an insoluble salt the roots can't take up. The pH is the lever. The aluminium is the agent. Confuse the two and you have the myth.

sepal Al = —

The white cultivar makes the point from the other side: it carries no anthocyanin, so there is nothing for aluminium to complex with. You can acidify its soil all summer and it will not turn blue or pink — there is no pigment to move.

Now the part that unravels

Pull one thread and the tidy "living litmus" picture comes apart in your hands. Put a real anthocyanin indicator — the purple juice of red cabbage — next to the hydrangea and drive them with the same soil pH. They colour in opposite directions.

Hydrangea sepal

—

Red-cabbage litmus (true indicator)

—

Same pigment family — a delphinidin/cyanidin anthocyanin — and yet in acid the cabbage goes red while the hydrangea goes blue. The litmus responds to H⁺. The hydrangea responds to Al³⁺, which acid merely unlocks. Two flowers, two chemistries, pointed the wrong way from each other.

It gets stranger the closer you look. Push a proton-selective microelectrode into a single cell and the intuition inverts again: within one plant, the blue cells have a vacuolar pH of 4.1 and the red cells 3.3. The blue cells are less acidic inside — the exact opposite of what "acid makes blue" would predict, because this inner pH is a different quantity from the soil's and it points the other way.

And at the bottom of it all sits an open problem. The blue itself is a three-part complex — delphinidin-3-glucoside + Al³⁺ + a 5-O-acylquinic-acid co-pigment — that is so fragile it exists only in water, has never been crystallised, and gives no clean NMR. Chemists could only see it by spraying it through a mass spectrometer — which pinned its composition to a clean 1:1:1 (one pigment, one aluminium, one co-pigment). What the mass peak can't give is the geometry: the complex is too fragile to crystallise and yields no analysable NMR, so the three-dimensional structure of the blue is still unsolved. The garden trick is a century old; the molecule under it, chemists still cannot fully draw.

The check — every number recomputed in front of you

The colour bands below are measured aluminium content of the fresh sepal (Schreiber et al., BioMetals 2011). The green column is the published threshold — the cross-check, not a claim. The row you're on is highlighted.

colour	sepal Al (µg/g fresh)	published band

The gate, in numbers. Aluminium solubility follows the gibbsite equilibrium log[Al³⁺] = 8.5 − 3·pH: three protons per aluminium, so the free Al³⁺ your roots can reach multiplies about 1000× for every 1.0 the soil pH drops.

soil pH	[Al³⁺] in solution	vs pH one unit higher

For the pH you've set, the live colour engine, step by step:

—

The 10 and 40 µg/g thresholds and the 4.1 / 3.3 vacuolar-pH figures are measured; the pH → sepal-Al curve is a monotone model pinned to those two thresholds and the horticultural pH gate (named as a free choice below). Run every check yourself: node research/the-flower-that-tests-your-soil/verify.mjs.

What's exactly true, what's idealised, and what's still unsolved

Exactly true (measured)

The pigment. The colouring anthocyanin in Hydrangea macrophylla sepals is delphinidin-3-glucoside (delphinidin 3-O-glucoside). The blue is a three-component metal complex: that anthocyanin, Al³⁺, and a 5-O-acylquinic-acid co-pigment (5-O-caffeoylquinic and/or 5-O-p-coumaroylquinic acid). Remove the co-pigment and no stable blue forms (Kondo et al. 2005; Ito, Oyama & Yoshida 2018).

The colour bands. Fresh-sepal aluminium content sorts the colour: red 0–10, purple 10–40, blue above 40 µg Al per gram of fresh sepal (Schreiber et al., BioMetals 2011). Above ~40 µg/g the colour saturates — more aluminium does not make it bluer. That is why the model treats the colour as three bands with a cap, not a smooth gradient to infinity.

The gate. Soil pH does not colour the flower directly. It sets whether Al³⁺ is soluble and root-available: acidic soil (below ~5.5) frees it; neutral to alkaline soil (above ~6.5–7.5) locks it up as insoluble aluminium hydroxide. The correct one-liner is "soil chemistry gates aluminium; aluminium colours the flower" — not "pH has nothing to do with it." pH is the necessary lever.

The vacuolar-pH inversion. Within a single plant, the vacuolar pH of blue cells is 4.1 and of red cells 3.3 — blue cells are internally less acidic (Yoshida et al., Plant Cell Physiol. 2003). This intracellular pH is a different quantity from soil pH and can point the opposite way; keep them strictly distinct.

White cultivars. Anthocyanin-free white cultivars have no pigment to complex, so they cannot be turned blue or pink at any pH. (Some whites still pick up faint aging tints from unrelated causes, so "cannot be blued or pinked" is the honest claim, not "stay perfectly white forever.")

Idealised (free choices, named)

The pH → sepal-Al curve is a model. No field study has published a function mapping soil pH to sepal aluminium content — it depends on the existing soil aluminium, buffering, organic matter, cultivar and time. So the engine uses a monotone linear bridge pinned to two things that are published: the measured colour thresholds (10 and 40 µg/g) and the approximate horticultural pH gate. Concretely it puts the blue onset (40 µg/g) at pH 5.5 and the red boundary (10 µg/g) at pH 6.5, clamped to a 55 µg/g cap. The shape is illustrative; the endpoints are sourced.

The pH thresholds are approximate. Sources give the blue/pink gate as "blue below ~5.5, pink above ~6.5–7.5." These are horticultural rules of thumb, not sharp constants, and vary by source and soil.

The solubility constant is representative. The log*K = 8.5 in log[Al³⁺] = 8.5 − 3·pH is a representative gibbsite / soil-aluminium value (the literature spans roughly 8.1–9.4 depending on crystallinity; Lindsay 1979). The exact constant is a free choice and the absolute concentrations shift with it. What is exact is the slope of −3 — three protons consumed per aluminium dissolved — which fixes the 1000×-per-pH-unit behaviour regardless of the constant.

The bloom is not a dosing guide. Real colour shifts take months, depend on soil aluminium already present, and aluminium sulphate can be overdone to plant toxicity. The slider shows the mechanism, not a recipe or a timeline.

The cabbage litmus swatch uses the standard red-cabbage anthocyanin colour chart (red in strong acid → purple near neutral → blue/green in base) as a qualitative indicator, interpolated across the same pH range — it illustrates the opposite direction, not a calibrated titration.

Genuinely unsolved (and the page says so)

The three-dimensional structure of the blue is not known. The delphinidin-3-glucoside·Al·co-pigment complex is unstable, exists only in aqueous solution, and has resisted crystallisation and clean NMR. Chemists could observe it only by ESI mass spectrometry, which established its composition as a clean 1:1:1 — one anthocyanin, one Al³⁺, one co-pigment (a single molecular-ion peak at m/z 843 across 1:1:1, 1:2:1 and 1:3:1 mixing ratios; Ito, Oyama & Yoshida, Molecules 2018). What that measurement cannot fix is the geometry. As of 2026-07-09 the exact three-dimensional structure of the complex remains an open question — the composition is settled, the shape is not. The operable garden trick sits on top of a molecule chemists still cannot fully draw.

Species precision

Only bigleaf (H. macrophylla) and mountain (H. serrata) hydrangeas reliably change colour. Smooth (H. arborescens, e.g. 'Annabelle') and panicle (H. paniculata) do not turn blue at any pH — the "hydrangea" here means the two that can, not the genus.