The Mortar

In 1996, Sara Wright, a soil scientist at the USDA Agricultural Research Service, extracted a protein from soil that no one had detected before. The protein was a glycoprotein produced by arbuscular mycorrhizal fungi — the symbiotic fungi that colonize the roots of roughly eighty percent of terrestrial plant species. Wright named it glomalin, after the fungal order Glomales. Over the next decade, her group showed that glomalin constitutes between fifteen and twenty-seven percent of soil carbon in temperate soils, binds soil particles into water-stable aggregates, and is present in virtually every soil sample ever tested.

The protein had been invisible. Not because it was rare — it was among the most abundant proteins in soil — but because the standard method for extracting soil proteins could not dislodge it. The Bradford assay uses mild detergent extraction. Glomalin resists mild extraction because it is the agent that holds soil aggregates together. The gentleness of the method assumes the soil is already coherent. Glomalin is the reason it is coherent. To extract it, Wright had to autoclave the soil — 121 degrees Celsius, sixty minutes, citrate buffer. The extraction that reveals the binding agent destroys the binding.

In 2018, Neil Theise and colleagues at Mount Sinai published a paper in Scientific Reports describing a previously unrecognized anatomical feature: the interstitium. It is a network of fluid-filled compartments supported by collagen bundles, distributed throughout the body — beneath the skin, lining the digestive tract, encasing blood vessels and airways. Theise proposed it functions in fluid transport, mechanical buffering, and possibly the spread of metastatic cancer cells.

Anatomists had examined every layer of this tissue for over two hundred years. The interstitium was not hidden behind anything. But the standard method for examining tissue — histological preparation — requires fixing the sample in formaldehyde, dehydrating it, and slicing it into thin sections for mounting on glass slides. The process drains fluid. Fluid-filled compartments collapse into dense, solid-appearing tissue indistinguishable from the surrounding connective stroma. The interstitium was systematically destroyed by the method designed to reveal anatomical structure. Theise saw it because he used confocal laser endomicroscopy, imaging living tissue in situ without fixing or dehydrating. The structure no one had seen was the structure that the seeing had been destroying.

In 1933, Fritz Zwicky calculated the mass of the Coma Cluster of galaxies by two independent methods. He estimated visible mass from luminosity and total mass from the velocities of the galaxies using the virial theorem. The two numbers disagreed by a factor of more than a hundred. There was far more mass than anything that emitted light could account for. Zwicky called the discrepancy dunkle Materie.

Four decades later, Vera Rubin and Kent Ford measured the rotation curves of spiral galaxies and found the same problem at a different scale. Stars at the outer edges orbit at roughly the same velocity as stars near the center — flat rotation curves where Newtonian gravity predicts declining velocities. Something is holding them in. That something constitutes approximately twenty-seven percent of the mass-energy of the universe, exceeds visible matter by a factor of five, and does not interact with electromagnetic radiation at any wavelength. It cannot be seen, photographed, or spectroscopically analyzed. Every telescope ever built — radio, optical, infrared, X-ray, gamma — is an electromagnetic instrument. The dominant binding force of the galaxy is invisible to every instrument designed to observe the galaxy.

In 1856, Rudolf Virchow named the non-neuronal cells that fill the spaces between neurons: Neuroglia — nerve glue. Three decades later, Santiago Ramón y Cajal established the neuron doctrine: the nervous system is composed of discrete cells, neurons, that communicate across synaptic gaps. The doctrine earned him a Nobel Prize and created a category for everything else. Astrocytes maintained ion balance. Oligodendrocytes provided insulation. Microglia handled immune functions. None were considered part of the computational machinery. They were the mortar between the bricks.

In 1999, Alfonso Araque and colleagues proposed the tripartite synapse: astrocytes are not passive bystanders but active participants in synaptic transmission. They release neurotransmitters, propagate calcium waves independently of neuronal activity, modulate synaptic strength, and participate in memory consolidation during sleep. The cells that Virchow named "glue" turned out to be computational elements. But the neuron doctrine had defined computation as what neurons do. The glue was not silent. It was speaking in a channel the framework had classified as scenery.

In each case — soil, tissue, cosmos, brain — the binding agent was present, abundant, structurally essential, and invisible. Not because it was hidden behind something. Because the observation method presupposed the structure the binding provides. The soil assay assumes coherent aggregates. The histological preparation assumes solid tissue. The telescope assumes luminous matter. The neuron doctrine assumes neuronal computation. And the binding is not rare. Glomalin constitutes a quarter of all soil carbon. The interstitium lines every organ. Dark matter outweighs visible matter five to one. Astrocytes are as numerous as neurons. The binding was always the majority. What made it invisible was not its absence but the fact that every question addressed to the structure was a question that assumed the structure was already there. You do not see the mortar when you are counting bricks. You do not see it because counting bricks is only possible on a wall that is already standing.