The Life Seam · a live disagreement, handled honestly

The Fraction That Reaches the Tree

A tree fixes carbon in its leaves. Weeks later, a labelled atom of it turns up beside a neighbour of another species. That much is real, measured, and repeatable. The famous wood-wide web argument is not about whether carbon moves — it does. It is a fight over one number nobody has measured: of the carbon a scientist detects at the neighbour, how much is inside the neighbour plant, and how much is sitting in fungal tissue running its own economy? Partition the real fluxes yourself. The mixing arithmetic is exact. The fractions are the open question — and this page picks no winner.

In 1997 Suzanne Simard and colleagues sealed paper-birch and Douglas-fir seedlings in the ground in British Columbia, fed one species ¹³C and the other ¹⁴C as carbon dioxide, waited, and harvested. Carbon labelled in the birch showed up in the fir, and vice versa. The transfer was bidirectional, and it netted out toward the fir: the fir's net gain averaged 6% of the carbon isotope it took up through its own photosynthesis, rising when the fir was shaded. A non-mycorrhizal western-red-cedar control barely picked up any label — evidence the carbon moved mainly through the fungal threads, not loose through the soil.¹

Two numbers set the stakes. Simard's 6% net, fir-ward is a modest, careful field figure — and note the wording exactly: 6% of the isotope the fir itself fixed, not 6% of the fir's whole carbon budget, and a net figure hiding a larger two-way gross flow. Nineteen years later, Tamir Klein and colleagues labelled the canopy of a 40-metre spruce for five years with heavy-carbon-depleted CO₂ and estimated that tree-to-tree transfer amounted to ~40% of fine-root carbon — about 280 kg of carbon per hectare per year. A huge number. But whether that carbon truly crossed through the fungal network, rather than through overlapping roots and soil, is itself contested.²

Follow a hundred labelled atoms into the ground

Here is the honest core. A field isotope experiment does not watch carbon travel. It labels a donor, waits, harvests tissue near a receiver, and measures the isotope excess in that sample. To read that excess as "the neighbour tree was fed," you have to assume where the labelled carbon ended up. So make the assumption explicit. Below, the donor pushes 100 units of labelled carbon below ground. You decide how it splits between four destinations. The bar always sums to 100 — that part is bookkeeping, and it is exact.

That last sentence is the entire controversy in miniature. In 1999 David Robinson and Alastair Fitter made the point sharply: much of the label recovered "in" a receiver may sit in fungal tissue, not the plant — the mycorrhizal fungus is a large living thing running its own carbon economy, and a plant-to-plant reading quietly credits the plant with the fungus's biomass. Their argument reframed the naive question. Not did carbon move? — yes. But what fraction reached the neighbour plant, versus stayed in the fungus? That is a measurable quantity. It has not been measured in the field.³

One measurement. Two stories.

So freeze the measurement. Whatever the partition above, the isotope reading collapses to one number: the label detected at the neighbour. Call it 100% of the signal. Now slide the one dial nobody has pinned — the share of that signal that is really fungal tissue, not plant — and watch the same, unchanged datum tell opposite stories.

The dial is not a rhetorical trick. It is a real, physical quantity — the fraction of detected isotope that resides in fungal rather than plant tissue — and it is genuinely unknown for field networks. You cannot dissect a metre of intermingled hyphae and fine root at the resolution the question demands, and bulk digestion and mass spectrometry only ever hand back the sum. Push the dial left and Simard's 6% becomes a real subsidy between trees. Push it right and Robinson & Fitter are correct that the plant barely sees it. The published data are consistent with both. That is not a failure of anyone's experiment; it is where the evidence actually stops.

One sentence, three different amounts of evidence

The popular claim — "trees talk and share food through an underground fungal web, and mother trees nurture their young" — is really three claims welded together. Karst, Jones and Hoeksema separated them in 2023 and asked what the peer-reviewed literature actually supports. The three answers are not the same.⁴

They also documented a positive citation bias: papers were increasingly cited as if they had shown benefit or communication even when their own results were neutral or negative — the story outrunning the data. (A minor 2023 author correction to their paper removed four stray references and renumbered; it changed none of this.⁴)

The argument, with no winner picked

This is a live scientific disagreement between serious researchers, and it would be dishonest to resolve it for you. Here are the positions in their own register.

One boundary worth keeping straight. Everything above is about ectomycorrhizal (EM) symbioses — the fungi of conifers, birches, oaks — which is what Simard and Klein studied. The other great guild, arbuscular mycorrhizal (AM) fungi of most herbs and grasses, is a different system with different carbon rules; some work (e.g. Pfeffer et al. 2004) argues the AM fungus does not pass its carbon on to plant tissue at all. Don't let a headline blur the two.