Every food forest has the same hidden ceiling: nitrogen. You can plant the most carefully chosen fruit trees in the world, build the perfect mulch layer, and chop-and-drop comfrey three times a season, and the system will still slow down if there is no perennial source of nitrogen feeding it from below. The answer that experienced food forest designers keep returning to is nitrogen-fixing trees: long-lived woody species that pull nitrogen straight out of the atmosphere and store it where surrounding production trees can use it.
This guide gives you the science, the species that actually work in US zones 2 through 10, the ones to absolutely avoid, and how to integrate them into a working food forest design.
Atmospheric nitrogen is 78 percent of the air, but plants cannot use it directly. The only way nitrogen enters the food forest naturally is through bacteria that strip it out of the air and convert it to ammonia in exchange for plant-supplied sugars. Trees that can host these bacteria fall into two distinct groups.
Legume trees plus Rhizobium. Members of the legume family (Fabaceae) including black locust, redbud, Kentucky coffee tree, Siberian pea shrub, mesquite, and acacia partner with Rhizobium and Bradyrhizobium bacteria. The bacteria enter root hairs and form discrete pink nodules in which they convert atmospheric N to ammonia. This is the most familiar nitrogen-fixing system and the one most cover-crop literature describes.
Actinorhizal trees plus Frankia. A second, much less-known partnership runs through Frankia, a filamentous actinobacterium that forms multicellular vesicles inside tree root nodules. Peer-reviewed work indexed at NIH documents the symbiosis with non-legume genera including alder (Alnus), sea buckthorn (Hippophae), Russian olive and autumn olive (Elaeagnus), buffaloberry (Shepherdia), Casuarina, Ceanothus, and Myrica. Critically, Frankia tolerates acidic and waterlogged soils where Rhizobium struggles, which makes actinorhizal trees the only practical option for many marginal sites.
Key Takeaway
If you only plant legume nitrogen-fixers, you will hit a ceiling on acidic, waterlogged, or nutrient-poor soils. Mix Rhizobium and Frankia hosts, and your food forest can build fertility on almost any site, including the rough ground most beginners start with.
| USDA Zone | Recommended N-Fixers | Partner |
| 2 to 5 (cold) | Siberian pea shrub (Caragana arborescens), sea buckthorn (Hippophae rhamnoides), speckled and grey alder (Alnus incana), goumi (Elaeagnus multiflora) | Rhizobium for Caragana; Frankia for the rest |
| 5 to 7 (temperate) | Black locust (Robinia pseudoacacia), Eastern redbud (Cercis canadensis), Kentucky coffee tree (Gymnocladus dioicus), red alder (Alnus rubra in PNW), goumi | Rhizobium for Robinia/Cercis; Frankia for Alnus/Elaeagnus |
| 7 to 10 (warm) | Honey mesquite (Prosopis glandulosa), velvet mesquite (P. velutina), Acacia spp., Tagasaste / tree lucerne (Cytisus proliferus) | Rhizobium |
Sources: USDA NRCS Plant Guide, Siberian pea shrub (PDF), USDA Forest Service Silvics, Red Alder, USDA Forest Service Silvics, Black Locust, Winrock International, Sea Buckthorn.
Common Mistake to Avoid
Honey locust (Gleditsia triacanthos) is in the legume family but does not form effective nitrogen-fixing root nodules. Plant it for its pods (12 to 13 percent protein for livestock and wildlife), its dappled shade, and its durable timber, but do not count it as your nitrogen source. (USDA NRCS Honey Locust Plant Guide PDF.)
If you can grow only one nitrogen-fixing tree, in most of the eastern half of the US, the answer is black locust. USDA Forest Service silvics data documents annual height growth of about 4 feet per year on good sites in the first decade, and peer-reviewed agroforestry literature places its nitrogen fixation at 100 to 200 kg per hectare per year. Cornell Small Farms describes black locust as "the most rot-resistant wood that can be grown in the Northeast climate" with flowers that supply Hungary's commercial honey industry. USDA Forest Products Laboratory documents heartwood durability sufficient for fence posts, mine timbers, insulator pins, and ship treenails without chemical treatment.
The trade-offs are real. Bark, leaves, and seeds contain robin and phasin, both toxic to livestock. Cornell's Plants Poisonous to Livestock database lists horses as most susceptible, with symptoms including depression, anorexia, abdominal pain, diarrhoea, laminitis, and paralysis. Black locust also has stipular spines, and outside its native Appalachian range it can spread aggressively, which has put it on invasive species watchlists in western states and parts of the Midwest. In its native range east of the Mississippi it is unambiguously a gift; in the West, plant only if you are prepared to manage suckers actively.
Red alder (Alnus rubra) is the most productive nitrogen-fixer in the Pacific Northwest, and the only commercial timber species west of the Rockies that fixes atmospheric nitrogen at all. USDA Forest Service Silvics documents soil nitrogen accretion in red alder stands at 40 to 300 lb per acre per year (45 to 355 kg/ha/year). At the high end of that range, a single mature alder stand exceeds what most production fertilisers contribute over a decade.
Speckled alder (A. incana) and grey alder extend Frankia-based nitrogen fixation into zones 2 and 3, particularly on wet sites. Because Frankia tolerates acidic soils that defeat most Rhizobium strains, alders are the right choice for waterlogged riparian zones, podzolised forest soils, and acidic glacial till common across the Northeast and Upper Midwest. They live 60 to 70 years, which makes them a successional species: plant them early to build fertility, let longer-lived production trees gradually take over the canopy as the alders age out.
For zones 2 and 3, where black locust and most legume trees struggle, two species do most of the work. Sea buckthorn (Hippophae rhamnoides) forms Frankia nodules and has been measured fixing up to 179 kg N/ha/year in a UK stand. Its orange berries are among the highest natural sources of vitamin C and omega-7 fatty acids, which has built a commercial juice and supplement market. Plant a male for every six to eight females for pollination.
Siberian pea shrub (Caragana arborescens) remains the most reliable Rhizobium-host woody nitrogen-fixer at zone 2. USDA NRCS documents its use across the northern Great Plains for windbreaks and shelterbelts. Its small peas feed chickens and wildlife, and its dense thorny structure traps snow on priority beds and blocks deer. Coppice it on a 2 to 3 year cycle for chop-and-drop biomass.
Why This Works: Biological Nitrogen Banking
A legume cover crop fixes nitrogen for one season and then dies. A nitrogen-fixing tree fixes nitrogen for 30, 50, or 100 years and stores it in growing biomass that you can release on demand through pruning. Over a 20-year horizon, a single mature black locust contributes more biologically available nitrogen than several decades of annual cover crops on the same land, with none of the tillage and replanting cost. This is the deep logic behind permaculture's preference for perennial systems over annual ones.
Russian olive (Elaeagnus angustifolia) and autumn olive (E. umbellata) are both Frankia-hosts and both prolific nitrogen-fixers. They are also among the most damaging woody invasives in North America. USDA Forest Service Fire Effects Information System documents Russian olive on 46 state noxious weed lists, with western US riparian areas forming self-replacing thickets that exclude native willows and cottonwoods. Penn State Extension reports autumn olive aggressively invading from Maine to Virginia and west to Wisconsin and Nebraska, with a single mature plant producing up to 80 lb of fruit that birds spread widely.
Both species fix nitrogen impressively, and that is exactly why they are invasive: they out-compete native plants on the marginal sites where alternatives cannot establish. Substitute goumi (Elaeagnus multiflora), a related actinorhizal shrub with lower invasive risk, or one of the alders above.
Martin Crawford at the Agroforestry Research Trust recommends planting roughly one nitrogen-fixer for every three to five production trees during the establishment years, then reducing the ratio as the system matures. This works because nitrogen-fixers are fastest-growing in years 1 through 8 and contribute the most biomass during the establishment period when production trees most need it.
Pick zone-appropriate species, not Pinterest favourites
Use the zone table above. Black locust does not work in zone 3. Sea buckthorn struggles in zone 9. Honey mesquite does not survive Vermont winters. Match the plant to the climate before anything else.
Plant 1 nitrogen-fixer per 3 to 5 production trees
Interplant them rather than blocking them. A line of black locust will leave gaps in the production rows; a black locust between every third apple tree feeds the apples directly through root contact and leaf drop.
Inoculate when planting in disturbed or new soil
Rhizobium and Frankia strains are usually present in long-established soils, but disturbed sites, formerly tilled land, and new fill often lack them. Commercial Rhizobium inoculants are species-specific, so order one matched to your tree genus. Frankia inoculation is harder to source commercially; a scoop of soil from an established stand of the same genus works as a field inoculant.
Coppice or pollard to release the nitrogen
Nitrogen sits in the wood and leaves of the living tree until you cut it. Coppicing (cutting near ground level) or pollarding (cutting above browse height) on a 2 to 5 year cycle drops a flush of nitrogen-rich biomass for chop-and-drop mulching, and the root die-back that follows releases another pulse of nitrogen directly into surrounding soil.
Plan the exit
Most nitrogen-fixers are pioneer species and shorter-lived than the production trees they support. Plan to selectively remove or coppice the lowest-performing nitrogen-fixers as the canopy of the production trees closes, year 8 onward. Combine this with the broader seasonal food forest maintenance cycle.
| Mistake | Why It Fails |
| Counting honey locust as a nitrogen-fixer | Honey locust is a legume but does not form effective nodules. Plant it for pods, shade, and timber, not fertility. |
| Planting Russian or autumn olive anywhere in the US | Both are listed as noxious weeds in dozens of states. Substitute goumi (E. multiflora) or sea buckthorn. |
| Using black locust as pasture shade | Bark, leaves, and seeds are toxic to horses, cattle, sheep, goats, and poultry. Fence livestock out of the seed zone. |
| Skipping inoculation on new sites | Disturbed, eroded, or formerly tilled soils often lack the matching Rhizobium or Frankia strain. Without it, nodulation fails and nitrogen fixation does not happen. |
| Planting too few or too many fixers | Below 1:5, growth stalls. Above 1:3, the fixers shade out the fruit and nut crop. Match the ratio to soil fertility and adjust over time. |
| Forgetting to coppice or prune | Without cutting, nitrogen stays locked in the standing tree. The biomass and nitrogen pulse only happens when you cut. |
Sources: USDA Forest Service FEIS, Russian olive, Penn State Extension, Invasive Olives, Cornell, Plants Poisonous to Livestock.
Get Our Free Companion Planting Chart
Join 10,000+ gardeners getting weekly tips on what to plant together, soil health, and permaculture techniques.
Send Me the ChartFor temperate North America (zones 5 to 7), black locust is the most-recommended workhorse, delivering 100 to 200 kg N/ha/year along with rot-resistant timber and excellent pollinator forage. For cold zones (2 to 4), Siberian pea shrub and sea buckthorn are the most reliable choices. For Pacific Northwest sites, red alder fixes 45 to 355 kg N/ha/year. Eastern redbud and Kentucky coffee tree fill smaller niches in the same zones.
Peer-reviewed research documents 100 to 200 kg N/ha/year for established black locust stands in temperate zones. A typical 0.4 hectare food forest with five mature black locust trees receives roughly 40 to 80 kg of biologically fixed nitrogen per year, equivalent to substantial conventional fertiliser input.
No. Despite being in the legume family, honey locust does not form effective root nodules and does not fix significant nitrogen. It is still valuable in food forests for its high-protein pods, dappled shade, and durable timber, but plant it for those benefits, not for fertility.
Rhizobium is the genus of bacteria that partners with legume trees like black locust, redbud, mesquite, and Siberian pea shrub. Frankia is a filamentous actinobacterium that partners with non-legume "actinorhizal" trees like alder, sea buckthorn, Russian olive, autumn olive, buffaloberry, and casuarina. Frankia tolerates acidic and waterlogged soils better than most Rhizobium strains, making actinorhizal trees the right choice for marginal sites.
No. Both species are listed as noxious weeds in dozens of US states and are among the most damaging woody invasives in North American riparian zones and forest edges. Substitute goumi (Elaeagnus multiflora) or sea buckthorn for the same Frankia-based nitrogen fixation without the invasive risk.
Dig carefully around a few roots and look for nodules. Legume nodules are typically pink, red, or brown spheres a few millimetres across, attached to root hairs. Frankia nodules look more like dark coral-coloured clusters. If you see no nodules at all on a year-2 or year-3 tree, the soil may lack the matching bacterial strain and warrant inoculation.
Martin Crawford and most permaculture practitioners recommend one nitrogen-fixer for every three to five production trees during establishment (years 0 to 8), then reducing the ratio as the canopy closes and surface soil fertility builds. On poor soils start at 1:3; on already-fertile soils 1:5 is enough.
Ready to Grow Smarter?
Get our free beginner's guide to permaculture gardening, 20 pages of practical tips you can use this weekend.
Read the Free Guide