BRENNEN

Industries / Food, farming and forestry / Agriculture and farming / Cotton ginning

Ginning

Separating the fiber (lint) from the seed: saw gins for most cotton, roller gins for long-staple cotton.

7

Milestones achieved, from 1796 to 2025.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1796

    Saws make ginning continuous

    Eli Whitney's 1794 gin pulled fiber off the seed with wire spikes on a wooden drum, but the spikes worked loose and the ginner had to stop to dig out the cleaned seed. Hodgen Holmes's patent, signed on 12 May 1796, replaced the spikes with circular saws, the slotted bar with iron ribs, and opened the bottom of the box so seed fell out while the gin ran. USDA-ARS historians write that Holmes's continuous saw design still dominates ginning of Upland cotton, the short-staple kind that is most of the world's crop.

    cotton.org

    Credit

    Hodgen Holmes

  2. 1840

    A knife for long fibers

    Saw gins were fast but broke long fibers. Fones McCarthy of Demopolis, Alabama, patented a gin in which a leather-covered roller drags fiber under a fixed knife while a second, moving knife pushes the seed off. One stand ginned 68 to 91 kg of fiber a day, where the older two-roller gins managed about 11 kg of long-staple fiber. USDA-ARS authors write that outside the United States, countries that gin long-staple cotton still use McCarthy-type gins.

    cotton.org

    Credit

    Fones McCarthy

  3. 1923

    Weevils end Sea Island ginning

    The boll weevil, a beetle that eats cotton buds, cut Sea Island cotton on the US southeast coast from 50,000 bales in 1918 to no commercial crop after 1923, and roller ginning there ended with it. Idle roller gins moved west to gin the new long-staple Pima cotton of Arizona and California.

    Lessona gin type lasts only as long as the crop it was built for, and roller ginning survived by following long-staple cotton west.

    cotton.org

    Credit

    recorded by USDA-ARS

  4. 1963

    Roller gins five times faster

    The USDA-ARS Southwestern Cotton Ginning Research Laboratory at Mesilla Park, New Mexico, gin makers and private gins developed the rotary-knife roller gin, which ginned Pima cotton at one 218-kg bale an hour, five times the rate of the McCarthy-type gin it replaced. The last commercial McCarthy-type plant in the United States closed by 1980.

    cotton.org

    Credit

    USDA-ARS Southwestern Cotton Ginning Research Laboratory and partners

  5. 2005

    High-speed roller gins

    The same laboratory designed a high-speed rotary-knife roller gin. The first, installed in a commercial gin in 2005, ginned Pima four times and Upland three times faster than a conventional roller gin, and 73 stands were in use by 2009, mostly in California. By 2011 California gins roller-ginned 188,000 bales of Upland for a mill premium of 13 to 26 US cents per kg. A 2024 study by New Mexico State University economists with USDA-ARS, using cost data from Arizona and California gins, found high-speed roller ginning cost 14.18% more than saw ginning and conventional roller ginning 24.42% more.

    cotton.orgdoi.org

    Credit

    Carlos Armijo and M. N. Gillum, USDA-ARS

  6. 2006

    India pays for upgrades

    Under India's Technology Mission on Cotton, the Ministry of Textiles paid 25% of the cost, about Rs 20 lakh a factory, for ginning and pressing factories to modernize, aiming at 1,000 by 2009. By November 2006 it had sanctioned 777 and 467 were finished, but 84% of approved factories were in two states, Gujarat and Maharashtra.

    Lessonan optional matching grant reached few gins outside the two states where owners applied.

    business-standard.com

    Credit

    Ministry of Textiles and Cotton Corporation of India

  7. 2025

    Fewer, bigger gins

    USDA's National Agricultural Statistics Service (NASS) counted 419 active US gins in the 2025 season, down from 446 in 2024, ginning 13.5 million bales; 60% ginned more than 20,000 bales each. In 1900 there were 29,214 US gins averaging 320 bales each.

    esmis.nal.usda.govcotton.org

    Credit

    USDA NASS

Watch for

  1. Variety sets ginning speed

    Using 2021 records from eight commercial gins, USDA-ARS, SAS Institute, Cotton Incorporated and the National Cotton Council modeled how cotton variety, fiber quality and moisture change bales per hour. Of 73 varieties, 40 ginned slower than the average and 11 faster; ginning rate drops sharply above a threshold of seed-cotton moisture; and the first three weeks of the season run slower. They released a calculator for gins, while writing that their models predict rate "to a low level of precision".

    Achieved whenthe calculator's predictions are checked against measured rates at gins outside the original eight over another season, and the results are published.

    cotton.org

    Claimed by

    Jaya Shankar Tumuluru and colleagues, June 2025

Open

  1. 01

    Map the world's gins

    An open map of active cotton gins worldwide, giving each gin's name, location, country, gin type (saw or roller) where known, and the public source for every entry.

    The brief
    Credit · unclaimed
  2. 02

    Gin records on one clock

    Open-source software that puts a gin's machine logs (gin-stand load and feed rate, dryer temperature, lint moisture, press cycles) on one time standard, Coordinated Universal Time (UTC), and stamps each bale's number onto the minute it was pressed, so ginning rate can be studied bale by bale against the settings that produced it.

    The brief
    Credit · unclaimed

Open 01 · the brief

Map the world's gins

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An open map of active cotton gins worldwide, giving each gin's name, location, country, gin type (saw or roller) where known, and the public source for every entry.

Why it is still open

OpenStreetMap held 36 features tagged as cotton gins worldwide in November 2026, and USDA NASS publishes only counts of active gins by county and state, 419 in the United States in 2025. Queensland's government publishes an open map layer of 11 gins in Queensland and nearby New South Wales.

data.gov.autaginfo.openstreetmap.orgesmis.nal.usda.gov

Built with what exists

NASS county tables as a completeness check; public directories of ginner associations and government programs; satellite imagery, where gins show as module yards and bale stacks; QGIS; a low-cost AI model that reads directory pages and fills in fields (see the Note "The jobs agents repeat").

  1. First step

    Map Texas's 150 active gins and compare the result with the NASS count for each county.

  2. Done looks like

    At least 3,000 gins in the eight largest cotton-growing countries, each with a source; US counts per county within 5% of the NASS table; 200 randomly chosen entries checked on imagery, with the error rate published.

Open 02 · the brief

Gin records on one clock

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

Open-source software that puts a gin's machine logs (gin-stand load and feed rate, dryer temperature, lint moisture, press cycles) on one time standard, Coordinated Universal Time (UTC), and stamps each bale's number onto the minute it was pressed, so ginning rate can be studied bale by bale against the settings that produced it.

Why it is still open

The 2025 USDA-ARS study of eight gins found the time stamps of different gin systems were not standardized, predicted ginning rate only "to a low level of precision", and asked the industry to adopt UTC and network time. An open plugin for gin data from AgGateway's ADAPT project exists, but it only imports records.

cotton.orggithub.com

Built with what exists

Python; network time protocol (NTP) on a small computer such as a Raspberry Pi; AgGateway's ADAPT framework and cotton gin data plugin; the gin's existing control-system exports.

  1. First step

    Write the shared record format and test the merge on one made-up day of gin-stand, dryer and press records.

  2. Done looks like

    Records from at least three gins for one season, shared with permission and anonymized; each machine log's clock offset measured and stated; and the published ginning-rate model rerun on the merged data, with its prediction error compared to the original.