If your soil is sand, you already know the cycle: fertilize in spring, and by mid-summer the water has carried it away. This page answers a concrete question — how to build real fertility on poor, free-draining ground with plants instead of bags, and how to tell whether it's actually working.
The problem, plainly
Sandy soil holds neither water nor nutrients, so every fix has to be re-bought and re-applied. Hauling in compost helps, but it is a one-time payment; the soil food web is the only builder that keeps working after you leave the field. This is not a new idea. The oldest Chinese farming manuals — the Fan Shengzhi Shu from the first century BC and the Qimin Yaoshu from 544 AD — taught exactly this: fertility grown in place, and what the field gives, the field gets back.
What actually works
Sunn hemp (Crotalaria juncea) is the workhorse: a tropical legume that grows fast in summer heat, stands 1.5–2.5 m tall, and fixes significant nitrogen — in warm-season plantings it can add 100–150 kg N/ha to the soil as the residue breaks down. Its thick taproot and vigorous growth make it the closest thing to a free fertilizer plant for sandy, low-fertility ground, and the deep root system re-opens compacted layers — a biological broadfork. Cut at early flowering for the most palatable residue: the tops mulch, the roots feed. Timing is Heaven: sow after the last frost once the soil warms, and cut before it flowers to avoid hard stems and seed; in central Florida that means a May–June sowing cut 60–90 days later. Companions cover the rest of the year: cowpea or lablab for nitrogen and shade in summer; crimson clover or vetch for early nitrogen in fall–winter; daikon radish as a bio-drill through compacted layers in fall; rye or oats for biomass and erosion armor in winter; buckwheat for a fast weed-smother in summer gaps.
How to run it
- Read the bed. Spade test first — is the topsoil sandy, is there a hardpan layer, where does water sit?
- Sow into a real seedbed. Cover crops need contact, not depth — rake, broadcast, rake again, and roll lightly.
- Let it grow to its window, not to seed. Cut or crimp at flowering while the stems are still green and succulent.
- Leave the residue in place. As mulch it feeds the top; the roots rot below and open the structure. Let the bed rest 2–3 weeks before planting.
Where it fails
- Sowing into a crust or too deep. Contact, not depth — rake, broadcast, rake again, roll lightly.
- Letting it go to seed. The stems harden and the residue turns to bedding instead of food; the cut window is flowering.
- Cutting too early or too late. Too early and you lose biomass; too late and the stems won't break down.
- Sowing before the soil warms. Germination stalls; in central Florida the window is roughly May–June for a 60–90-day crop.
- Judging it in minutes. Soil is Earth phase; it changes in seasons and years, never in minutes. Two or three cover-crop cycles on poor sand will show a visible difference; five will change the land. Keep the field journal simple — same spot, same spade test, twice a year — and watch for darker crumb, worms where there were none, and water soaking instead of running.
The short version
Spade-test the bed; sow Sunn hemp, or the season's companion, into good contact after the soil warms; cut at flowering and leave the residue where it falls; rest the bed 2–3 weeks; measure the same spot twice a year. The residue feeds 100–150 kg of nitrogen per hectare into poor sand, the roots open the hardpan, and the fertility keeps compounding while the fertilizer bags stay in the shed.
From the Biological Soil Building Guide
Biological Soil Building — PDF ($7) →
printable · instant delivery