Substrates

After it has finished draining, a pot of sand is 2.5 percent air.¹

The nursery trade measures four things about a growing medium, because it sells on numbers rather than on adjectives. One of them, the air left in the pot after free drainage, is the one that predicts whether the plant lives, and it is the one the bag in a garden centre leaves off.

The four numbers the trade actually runs on.

All four are measured on a saturated sample that has been allowed to drain freely. The standard American procedure is the North Carolina State University porometer; the field version is Handreck's bucket method², and for air-filled porosity the two answer the same question.⁴

Two media can hold identical amounts of water and behave nothing alike, because container capacity says nothing about how tightly that water is held. That is a fifth quantity and it has its own section below.

How much air is enough, on a ladder rather than a single figure.

Extension services give growers 10 to 20 percent air-filled porosity, with total porosity 50 to 70 percent³, and the wider American nursery range is 10 to 30 percent air space, container capacity 45 to 65 percent, bulk density 0.19 to 0.70 g per cm3. Handreck and Black did something more useful than publish a target. They wrote down what each level means in practice.

Two things fall out of that ladder for anyone keeping plants in a room rather than a nursery. The mix a household wants is more open than the professional standard, not less, because the failure mode at home is a pot that stays wet. And the chunky mixes the aroid keepers argue about are not eccentric: at 40 to 50 percent air they are sitting in the documented epiphyte band, which is the correct band for a plant that roots into bark and litter on the side of a tree. What comes with that band is the watering frequency, and that is the part usually left out.

What the materials measure, one dataset at a time.

Substrate numbers do not travel between studies. Method, pot height and particle grade all move them, so the two tables below are kept apart and each is quoted with its source. Do not average a row from one against a row from the other.

Classic components, drained in a container

United States Forest Service nursery manual, table 2.2.5, values as published.

The wet columns are worth as much as the air column. Peat goes from 104 to 694 kg per m3 when saturated, a 6.7 fold rise, and vermiculite does almost the same at 109 to 641. Perlite goes 96 to 394, a much smaller gain, because its cells are closed and water clings only to the outside of the particle. If the pot has to stay liftable when wet, perlite is the aggregate and vermiculite is not.

Mineral substrates, all sieved to the same grade

Eleven materials at 4 to 7 mm plus pumice at three grades, European pressure plate method. Available water is maximum water capacity minus the permanent wilting point at 15,000 hPa.⁵

Grade beats material. The same pumice delivers 21.5 percent available water at 1 to 3 mm and 9.5 percent at 4 to 7 mm. That gap is larger than the gap between pumice and perlite. Anyone selling pumice without stating a grade is selling an unknown.

Vermiculite is not a substitute for perlite. Same nominal job, and the measurements are not close: it holds 46.6 percent water against perlite's 24.2, it goes from 0.11 to 0.58 g per cm3 wet against perlite's 0.10 to 0.34, and it sits at pH 6.1 against perlite's 8.7. It is also the one mineral component with real cation exchange capacity, which makes it a chemically active ingredient pretending to be an inert one.

Water held is not water the plant can have.

The European convention defines the difference by suction. Air space is the water lost between 0 and 10 cm of suction, easily available water is what comes out between 10 and 50 cm, the reserve comes out between 50 and 100 cm, and everything still held beyond that is unavailable down to the wilting point at 15,000 cm.⁶ The American nursery convention arrives at the same uncomfortable place from the other direction: available water 25 to 35 percent and unavailable water 25 to 35 percent by volume, which is to say that in a good substrate roughly half the water in the pot is water the plant will never touch.

The consequences are not intuitive. Perlite has 91.5 percent porosity and gives up 11.8 percent available water. Zeolite has 56.5 percent porosity, nearly the lowest in the set, and gives up 18.5 percent. Vermiculite holds nearly half again as much water as fine pumice, 46.6 against 32.1 percent, and delivers the same amount to the plant, 21.4 against 21.5. The paper's own conclusion is that pore interconnection and pore size distribution decide hydraulic behaviour, not pore volume. Total water held predicts almost nothing.

The same result turns up on the organic side. Thirteen coir dusts from Asia, the Americas and Africa all measured above 94 percent total porosity, with air content ranging from 24 to 89 percent by volume, and they held less easily available water and less total water than sphagnum peat.⁷ That is the opposite of how coir is sold, and the 24 to 89 percent spread is not a specification, it is a lottery. Coir is worth using. It is not one material.

A mix that feels damp in the hand and holds all of its water below the wilting point is, from the plant's point of view, a dry mix. Chase easily available water and air space together, and treat container capacity as a distraction.

A handful of grit in a bag of compost does less than nothing.

This is the least intuitive result in the subject and it has been measured twice, forty years apart. Mix particles of two sizes and the resulting volume is always less than the sum of the parts, because the small ones fill the gaps between the large ones. Spomer worked out where the turning point sits. Below it, added grit occupies volume without creating drainable pores between grains, so total porosity falls. At the point where the grains just touch and the fine material exactly fills the gaps, porosity is at its minimum, which he called the worst possible mixture to grow plants in. Only above that point does aeration start to rise.⁸

For sand, perlite, calcined clay and similar amendments, the threshold is about 10 volumes of amendment to two and a half or three volumes of soil, which is roughly 77 to 80 percent amendment by volume. Nobody adds 80 percent grit. Everybody adds 20 or 30 percent, which is on the wrong side of the curve, which is exactly why it makes the mix denser rather than looser.

It is not a nineteen seventies curiosity either. The measurement was repeated on a modern nursery substrate:

Stated properly: a coarse aggregate improves aeration only when you add enough of it that the coarse particles carry the load and the fine material rattles around between them. Below that, you are filling pores rather than making them. Organic openers such as bark and coarse coir chips reach their threshold at a lower proportion than sand does, which is the honest reason to prefer them.

Every pot has a drowned layer at the bottom. Its height belongs to the medium.

Water a pot until it runs out of the hole, wait until it stops dripping, then cut the pot open. There is a layer at the very bottom in which every pore is full of water and there is no air at all, with water content falling as you move up. That is the perched water table, and it is in every container, whatever you do about drainage holes.

The reason is that the suction available at any point comes from its height above the free drainage surface. At the drainage hole that height is zero, so there is nothing to pull the water out of the pores and they stay full. A centimetre up there is a centimetre of water column pulling, ten centimetres up there is ten. In a field soil the free drainage surface is the water table, metres down. In a pot it is the hole. That single difference is why an excellent garden soil can be a lethal potting medium: a fine soil in a container will probably remain saturated after watering and drainage.¹⁰

Pore size sets the thickness. Capillary rise is roughly 15 divided by the pore radius in millimetres, so a 1 mm pore is beaten by about 15 mm of water column and a 0.01 mm pore holds against about 1.5 metres.¹¹ Fine media therefore have thick saturated layers and coarse media have thin ones, and that is the whole physics behind chunky mixes.

The consequence people miss is what happens when you change the pot rather than the mix. The depth of the saturated layer is a constant for a given medium and is unaffected by the size and shape of the container, so the same soil in a shallower pot has a higher percentage water holding capacity than in a deeper one.¹²

The drainage layer correction.

Almost every source you will find, including the ones worth trusting on other subjects, says that a layer of gravel in the bottom of a pot raises the perched water table and makes matters worse. We have said it ourselves. The evidence does not support it as stated, and this is the most consequential correction on this site.

How much weight that deserves is a fair question, and the paper answers it honestly. It is one study. It characterised no pore sizes or water potentials, used a single container geometry, saturated the pots fully rather than watering them the way a person does, and ran no plant growth trial at all. It tells you where the water went, not whether anything grew better.

So the useful version is neither slogan. A drainage layer does not remove the drowned zone from the system, it relocates it: in the loam trials the grit layer itself sat saturated after draining. If what you care about is roots, moving the drowned zone into a bed of stones is an improvement. What it costs you is certain while the benefit is contested, because a 60 mm layer in a 175 mm pot is about a third of the container given over to stones, and the fifty year old alternative is to mix the coarse material through the whole medium instead, which buys the same aeration without losing the volume. The claim to retire is not that layers help or that layers hurt. It is that a layer of stones substitutes for a suitable medium. It does not.

A mix does not fail because the plant outgrew it. It fails because it ran out of air.

Four things close the pores. Organic particles decompose. Fines wash down and lodge between the larger particles. Vermiculite plates fracture during mixing and handling. And roots grow into the large air-filled pores and fill them, which means the mix tightens fastest when the plant is doing well.

Coarse survives and fine does not, in every comparison in that table. The coarse pine substrate shrank half as much as the fine one from the same material. Mineral aggregates genuinely do not compact. What follows from that is an estimate rather than a measurement, and it is worth having as long as it is labelled as one: a mix has failed when its air-filled porosity falls below roughly 10 percent, so its service life is the starting figure minus ten, divided by the rate of loss.

Four things the hobby is sure about that the measurements do not support.

LECA holds no water. False. Commercial expanded clay measured 16.6 percent of its dry weight in water and laboratory fired material 14.62 percent¹³. What is true is that it gives almost none of it back: of the 11.0 percent by volume it held, 2.7 percent was plant available, the lowest of eleven mineral substrates tested. The pores are a reservoir for the ceramic, not for the plant.

LECA is inert and pH neutral. Not as supplied. It measured pH 8.40 and an electrical conductivity of 135.6 microsiemens per cm, the second highest in that set, attributed to production residues. It is also not the light option: saturated it is 0.78 g per cm3 against 0.69 for saturated sphagnum peat, before the standing water in any reservoir is counted. Expanded clay runs without a product standard behind it, too. Firing temperature alone moved water absorption up to a maximum of 14.62 percent in one clay and down in another as the glassy phase sealed the surface pores.

Horticultural charcoal is inert and sweetens the mix. Reviewed values put biochar total porosity at 74 to 85 percent, air space at 3 to 34 percent, bulk density 0.09 to 0.17 g per cm3, and pH from 5.4 for sugarcane bagasse to 10.1 for mixed hardwood.¹⁴ An air space that can be 3 percent is not an aeration amendment, and a pH that can be 10.1 is not inert.

Sphagnum moss is a stable material. Its water holding comes from large dead storage cells, and it is spectacular: across 21 species, saturated absorption ran from about 19 to about 44 times dry weight, in proportion to the volume of those cells in the leaves.¹⁵ A 2.3 fold spread between species means any figure published without a species name is meaningless, and commercial moss is not sold by species. As a wrap for aerial roots it is excellent. Packed into the body of a pot it compresses towards peat within a season.

Peat, said plainly.

Physically it is the best material in the table. Sphagnum peat measures 84.2 percent total porosity, 58.8 percent water and 25.4 percent air, at 104 kg per m3 dry. Nothing else combines that much water with that much air at that weight. The bag is also not one material: the darker, more decomposed hypnum peat gives 12.4 percent air from the same test, half of what the young fibrous sphagnum gives.

The case against it is not physical. Peatlands store roughly twice as much carbon as all the world's vegetation combined while covering about a tenth of the area of forest; ten centimetres of peat takes about a century to form and extraction can remove twenty two metres of depth; at least 80 percent of the peatlands in the United Kingdom are already damaged.¹⁶

The replacement question is genuinely open. A cradle to gate life cycle assessment of a cubic metre of growing medium delivered to Spain found coir best overall despite the sea voyage, because it is a by product, and expanded perlite worst in seven of eight impact categories at 620 kg CO2 equivalent per cubic metre.¹⁷ The same study found coir had the highest water consumption before regional scarcity weighting. If a mix is sold as peat free and leans on perlite, the awkward ingredient is the perlite.

How this feeds the vessels.

Sources.

Where these numbers come from. The component values here are measured ones, and the figure for a mix is worked out from its parts — which is how every published ratio for an aroid or semi-hydro recipe is arrived at. The compaction figures come from nursery and greenhouse growing, over weeks to about a year, so they hold best over that span; on a windowsill across five years, read them as the direction of travel.