How do I lay out an Inga agroforestry plot?

Dimensions, recommended densitites, plot architecture and other deep-nerd practice guidelines.

Note: this article is still under construction, and information is being corroborated. Especially the matrix diagrams. This is a bricolage protocol; which means we are building it in public! Some pages are under construction — but check back soon its updating fast!

There is more than one way to lay out an Inga or Guamo plot. The right plot architecture should be chosen for the crop.

• Alley cropping supports annual grain crops between hedgerows of Inga that are pruned sporadically, leading to long-term mulch and weed suppression.
• Matrix planting is for perennial crops grown alongside Inga that's lopped sporadically for shade management. The 4 m spacing is the same, but they are fundamentally different systems.

Jhony and Paulita on the Selva team planting Inga in hedgerow conformation.

Karen on the Selva team planting Inga in matrix conformation

Alley or hedgerow architecture

Alley architecture leads to the classic alley view. It's best used for permaculture crops.

TL:DR = 5k trees/ha = 4m alleys × 0.5m within-row spacing.

In plain language, the optimal recommended spacing/planting density for our sites at this time is four-meter alleys with a half-meter plant spacing within rows, which calculates to five thousand Inga trees per hectare.

This assumes 100% survivability, and in practice, most sites plant more densely, then weed to achieve the desired spacing (see Operational calendar below).

Optimization of alley spacing

There has been some debate in the field as to the optimal spacing, with competing farmer groups in Honduras and Ecuador, and iterative real-world testing over time. We included the historical variance and who promoted it below (see Table A), but all our sites use the density from Hands 2021 published from the multi-decade research series.

We know Inga alley cropping's weed-control mechanism depends on canopy closure and permanent mulch cover (Hands 2021). Hands' trials tested this between 5,000–10,000 trees/ha, and the Inga Foundation recommends 5,000 (Hands 1995, 1998, 2021, pers. com. 2026).

Table A. Historical variance in Inga alley cropping density, with Savimbo's current recommended alley or hedgerow density of 5k/ha

Density	Source	Status
10,000/ha	Comparison protocol Hands 1995, 1998 papers	Experimental variation, not for farmers
5,000/ha	Hands' 2021 paper, 2026 pers. comm. recommendation	What Savimbo uses
2,500/ha	Valle 2010 and Rainforest Saver, Ecuador	See debate below

We cannot find a published controlled trial to shed light on the debate about lower densities. Rainforest Saver reports good results at ~2,500/ha (1m) across 150+ Ecuadorian sites, based on field observation rather than controlled trials.

In our assessment, whether 2,500/ha reliably achieves the optimal canopy closure and permanent cover has not been adequately tested head-to-head. Because the Honduras/Costa Rica and Ecuador programs differ in both ecology and planting density, the two variables are confounded — part of why our team flags the field reporting as hard to compare directly. However, we have chosen to go with the higher (5,000/ha) density for our plots, given that we also use direct sowing rather than seedling cultivation, so the additional planting costs are minimal.

Matrix or nurse species architecture

Matrix architecture leads to a grid pattern. It's best used for establishing crops or interplanted crops. This is for helping native species establish themselves when planted into invasive grassland.

TL:DR = 3 options, 4m × 4m grid (nurse), 2:2 (cacao), tresbollillo lattics (hardwoods)

Many of our sites report significant repeated machete clearings to get trees to establish themselves. Inga is tough and competitive. and needs some machete work to establish, but far less so than many native trees. Using Inga decreases manual labor and makes it easier for embattled farmers to reforest their sites. Native species are added after establishment, but before canopy closure (Hands 2021).

When Inga is used as a matrix species there are three different conformations we are seeing in the literature. We've done our best to make diagrams of what we've read, but admit these aren't as good as on-site training with authors of original material.

Grid matrix (mixed species)

Plant Inga in 4m grids, intercropping with production crops like (Fig Aa). Reportedly this works well for chontaduro, banano, plátano, possibly pepper, and sinchi inchi.

Figure Aa: Standard planting configuration for Inga matrix nurse species

Banding matrix (cacao)

Plant Inga at 4m in staggered bands with cacao trees (Fig Ab).

Figure Ab: Banded planting configuration for cacao plots using Inga matrix nurse species

Tresbollilo lattice (hardwoods)

Plant Inga in tresbolillo at 4 m. But leave space for native hardwoods in a staggered configuration (Fig Ac). Geometrically, this creates a grid where every hardwood is bounded by 6 Inga — a hexagon. After 6–12 months, when the Inga is established and providing shade, but before the canopy closes, plant the hardwood saplings into those hexagonal gaps. (Hands 2021).

Figure Ac: Tresbollilo lattice planting configuration for hardwoods using Inga matrix nurse species

Hands reports in the 2000 Cam project: 2 ha planted this way, 15 species of rainforest trees in the gaps (including Terminalia and Swietenia as eventual emergents). Twenty years on, the hardwoods now tower over the Inga nurses, and the strip functions as a biological corridor — toucans and brown jays have brought in additional understory species like Xylopia frutescens that nobody planted.

On semantics

There are lots of different words for "things grown together". (And a lot of infighting, defining, and redefining as a result, exacerbated by endless indexing, ontology creation, and search terms).

We confess that we found this confusing, and suspect that smallfarmers do too.

• Permaculture uses vocabulary terms like guilds, polyculture, food forests, stacking, zones, and sectors.
• Silviculture, forest ecology, and restoration ecology use the word 'matrix species' to describe a dominant overstory or nurse species that defines the structural background into which target species get interplanted.

At the end of the day we're interested in action not talk so don't want to waste a lot of time on terms, just clarity of the right actions. We'll use the word matrix in this guide as its clear on actions.

Table B. List of terms for "things grown together" and what they describe.

Term	Field of origin	What it describes
Alley cropping	Tropical agronomy (IITA, 1970s)	Hedgerow trees + intercrop in alleys
Matrix planting	Forest restoration / silviculture	Nurse/canopy species + interplanted target species
Guild	Permaculture	Mutually beneficial species cluster
Food forest	Permaculture	Multi-strata edible polyculture
Polyculture	Agronomy (broad)	More than one species in a plot

Things we are unsure about

Precise information on which type of conformation is best for which crop is conflicting and sometimes paywalled. We defer to Inga Foundation papers where possible, and peer-learning from like ecosystems (altitude and soil matching) or site-experimentation when that's not possible.

Notes:

• Hands 1998 reports the 2.5 m alleys at La Conquista produced "the highest bean yields" in his trials, but were "too narrow for maize". Perhaps because beans tolerate shade better than maize, but we were unable to obtain updated information.
• Farmers in Ecuador reported growing cassava (yuca) and peas at 3m, but Hands didn't test these. No data either way.
• Cacao, chontaduro, banano, and plátano have been planted at 4 m alleys in some of our sites, but Hands does not use alley cropping for these crops at all. Same 4m spacing but matrix conformation. We think this is protocol drift, not a viable variant.