Industries / Food, farming and forestry / Agriculture and farming / Nut and fruit orchards
Planting and rootstocks
Planting and rootstocks is choosing which variety to grow and which root to graft it onto, raising the young trees in a nursery and setting them out in the orchard.
6
Milestones achieved, from 1846 to 2019.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1846
Pecan grafting succeeds
Grafting joins a cut shoot (a scion) from a chosen tree onto a young seedling's roots, so the new tree bears the same nuts as the chosen one. After Dr. A. E. Colomb failed to graft a pecan tree with large nuts that grew on a neighboring plantation, Antoine, an enslaved gardener at Oak Alley plantation, grafted 16 trees with it in the winter of 1846 or 1847, and by 1865 the plantation had 126 grafted trees in bearing. Nuts from them won a diploma at the 1876 Centennial Exposition in Philadelphia, and the variety was later sold as 'Centennial'. In 1904 W. A. Taylor of the US Department of Agriculture called it the first pecan successfully propagated by grafting and the first planted as a commercial orchard, and dated "the beginning of modern pecan culture" to Antoine's grafts.
Credit
Antoine, gardener at Oak Alley plantation; Dr. A. E. Colomb; Telesphore J. Roman, the plantation's owner
1874
A seedless orange in Riverside
In 1870 William Saunders, superintendent of gardens and grounds at the US Department of Agriculture in Washington, received twelve budded navel orange trees from Bahia, Brazil, and propagated them for distribution. Trees sent to Florida did poorly, but Luther and Eliza Tibbets planted three in their yard in Riverside, California, in March 1874. The seedless, easy-peeling fruit won first prize at a Riverside citrus fair in January 1879, and within about a decade the "Washington" navel was the most widely planted variety in the area, by the account of the University of California, Riverside's Citrus Variety Collection. The fruit has no seeds, so the variety is multiplied by budding (grafting a single bud) onto rootstocks.
Credit
William Saunders (USDA); Luther C. and Eliza Tibbets; missionaries at Bahia who sent the trees
1930
Tristeza kills grafted citrus
After a rot caused by Phytophthora, a soil-borne water mould, destroyed Mediterranean citrus grown from seed, growers grafted oranges, mandarins and grapefruit onto sour orange roots, which resist it, and sour orange became almost the only rootstock in the Mediterranean and the Americas. Citrus tristeza virus, carried by aphids and in infected buds, kills trees on that combination. A 2008 review by Pedro Moreno and colleagues at Spain's IVIA research institute lists the worst epidemics in Argentina (1930), Brazil (1937), California (1939), Florida (1951), Spain (1957) and Israel (1970), with almost 100 million trees on sour orange lost.
LessonAn industry grafted onto one rootstock can be undone by one disease, and rebuilding took tolerant rootstocks such as Carrizo citrange together with quarantine and certified, virus-tested budwood.
Credit
a failure; reviewed by Pedro Moreno and colleagues, Instituto Valenciano de Investigaciones Agrarias
1970s
Rootstocks tested side by side
Until then, universities tested rootstocks separately, with different trees, varieties, designs and care, so their results were hard to compare. NC-140, a USDA-funded regional project formally created in the mid-1970s from groups working since the 1960s, plants the same rootstocks with the same scion variety and design at many sites, so that differences come only from soil and climate. It grew to about 22 cooperators in the United States, five in Canada and two in Mexico, and added peach, pear, plum and cherry trials.
Credit
NC-140 cooperators; described by Richard Marini (Penn State) and Wesley Autio (University of Massachusetts Amherst)
1991
Rootstocks that resist fire blight
Fire blight is a bacterial disease that infects apple flowers at bloom and spreads fast, and M.9, the standard dwarfing rootstock from England, is vulnerable to it. In 1968 James Cummins and Herb Aldwinckle began breeding rootstocks at Cornell University's Geneva station that keep the dwarfing but resist fire blight and soil diseases. The first fire-blight-resistant Geneva rootstock was released in 1991 and G.41 in 2005; by Cornell's figures G.41 had 5 million plants in production in 2017 and made up 20% of the US apple rootstock market by 2019. In Washington's 2018 fire blight epidemic, a company that had planted two orchards of 150,000 trees each lost about 20% of the trees on M.9 and none on G.41, by Cornell's account.
Credit
James Cummins and Herb Aldwinckle (Cornell); Gennaro Fazio (USDA Agricultural Research Service), who leads the joint USDA–Cornell program
2019
Eleven million trees, one apple
Bruce Barritt of Washington State University (WSU) crossed Enterprise with Honeycrisp in 1997, and the seedling became WA 38, sold under the trademark Cosmic Crisp. WSU holds the patent and released it as a managed variety: only Washington growers in the United States could plant it for 10 years from its launch. Nurseries started 2017 with 300,000 trees against grower requests for 4 million, so trees were allocated by drawing, and within three years the industry had planted more than 11.5 million trees and invested over half a billion dollars, by the WSU Foundation's count. The first 450,000 boxes of 40 lb (about 18 kg) went on sale in December 2019.
Credit
Bruce Barritt and Kate Evans (WSU apple breeding program); Washington growers and nurseries
Watch for
Gene-edited citrus rootstock
Soilcea, a Florida company, and University of Florida scientists used CRISPR gene editing to change genes in a citrus rootstock, CarriCea T1, so that the bacterium behind citrus greening (huanglongbing, which has cut Florida's citrus production since it was found there in 2005) interacts less with the tree; they say it tolerates the disease. The US Environmental Protection Agency approved its registration in April 2026, and Soilcea's chief executive, Yianni Lagos, said 200,000 trees on it were going through nurseries for planting in the 2026–27 season.
Achieved whengrove trials run by someone other than Soilcea publish several seasons of infection rates, yield and fruit quality for varieties grafted on CarriCea T1 against the same varieties on standard rootstocks.
Claimed by
Soilcea and the University of Florida, April 2026 (reported by FreshFruitPortal)
Open
01
Grade nursery trees by camera
A camera station or phone app that grades bare-root fruit trees as they leave the nursery: it measures trunk caliper (diameter at a set height), counts the feathers (side shoots) and records each feather's height above the ground and length, then assigns the tree its grade.
The briefCredit · unclaimed02
Forecast rootstocks from pooled trials
An open table of NC-140 rootstock results, site by site and year by year (trunk cross-sectional area as a measure of tree size, cumulative yield, yield efficiency, survival, root suckers), with each site's soil and weather attached, and a model that forecasts how a rootstock and variety will perform at a new site.
The briefCredit · unclaimed
Open 01 · the brief
Grade nursery trees by camera
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A camera station or phone app that grades bare-root fruit trees as they leave the nursery: it measures trunk caliper (diameter at a set height), counts the feathers (side shoots) and records each feather's height above the ground and length, then assigns the tree its grade.
Why it is still open
High-density apple systems such as the Tall Spindle use highly feathered trees, with 10 to 15 or more feathers, to crop early, and Cornell's Terence Robinson and colleagues found tree price a large factor in which planting density pays (New York Fruit Quarterly, 2007). Grading by hand is inconsistent: when a University of Guelph researcher and a nursery manager measured the same 25 trees with the same caliper, only 64% landed in the same size class. Guelph's truck-mounted camera system measures caliper at 1,000 trees an hour, within one size class of a person's reading in 88.5% of cases, and the authors write that nurseries want it to determine grade next; a search found no tool that counts or measures feathers on fruit trees.
Built with what exists
A depth camera such as the Intel RealSense D435 or a phone with a LiDAR sensor; open image models such as Ultralytics YOLO or Meta's SAM 2 to find the trunk and each shoot; Open3D to measure in 3D; a backdrop and a turntable or conveyor; a digital caliper and tape for the hand-graded reference set.
First step
Ask a fruit tree nursery that is digging and grading trees this autumn to let a researcher photograph 100 trees on the grading table, each next to its grader's tag, and measure them by hand.
Done looks like
On 500 bare-root apple trees from a commercial nursery, the tool's caliper is within 1 mm of the mean of two hand measurements on at least 90% of trees, its feather count matches the count two graders agree on for at least 90%, and its grade agrees with the graders at least as often as they agree with each other; the photos and labels are published.
Open 02 · the brief
Forecast rootstocks from pooled trials
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open table of NC-140 rootstock results, site by site and year by year (trunk cross-sectional area as a measure of tree size, cumulative yield, yield efficiency, survival, root suckers), with each site's soil and weather attached, and a model that forecasts how a rootstock and variety will perform at a new site.
Why it is still open
Rootstock results change with the site and the variety. Richard Marini's 2020 analysis of the 1994 and 2003 NC-140 apple trials found that differences in rootstock vigor showed more at high-vigor sites than at low-vigor ones, and a 17-year New York trial of 19 rootstocks with five varieties found significant variety-by-rootstock interactions for tree deaths, growth, cumulative yield and yield efficiency (Lawrence, Fazio and colleagues, 2025). Yet NC-140 results are published trial by trial, and growers' guides such as Cornell's Geneva comparison chart rank rootstocks on fixed scales with no site conditions. A search found no public tool that pools the multi-site data.
journal.americanpomological.orgfrontiersin.orgctl.cornell.edu
Built with what exists
Published NC-140 papers in the Journal of the American Pomological Society and state reports; an AI model that reads PDF tables, with every number checked against the page by a second reader; PRISM or Daymet weather and USDA SSURGO soil data for each site; mixed models in R (lme4) or Python (statsmodels), and GGE biplots as Marini used.
First step
Extract the tables from the papers on the 1994 and 2003 NC-140 apple plantings, check every number by hand against the PDFs, and write to the trial coordinators to ask whether the raw data can be shared.
Done looks like
Results from at least ten NC-140 apple plantings in one public table, each number linked to the page it came from, and a model fitted on the older plantings that predicts tree size and cumulative yield per rootstock and site for the most recent finished planting, with its error reported against the simple baseline of each rootstock's average.