The main innovations in cotton processing are the saw gin, which made short-staple cotton cheap to separate from its seed (Eli Whitney’s spike-tooth gin, patented on 14 March 1794, and Hodgen Holmes’s saw gin, patented on 12 May 1796); roller drafting and machine spinning in British mills from 1769; dryers, seed-cotton cleaners and lint cleaners at the gin (early 1900s to the 1940s); the module builder (1971 to 1974); instrument (HVI) classing, required for every bale in the US loan programme from the 1991 crop; and rotor spinning (1967) and air-jet spinning (1978; vortex 1997). The result of the first of them was immediate. One worker on Whitney’s gin cleaned 50 pounds of lint a day, against about 5 pounds a day on a churka roller gin; US raw cotton output doubled every decade after 1800; by 1850 the United States grew three-quarters of the world’s cotton; and the gin greatly increased planters’ demand for enslaved labour to grow and pick it. Today a US saw gin plant handles 20 to more than 80 bales an hour.
Last verified: 12 October 2026. Every date and figure is cited to the US National Archives, a peer-reviewed paper, a museum record, the USDA classing handbook or the machine maker’s own page. Sources are listed at the end.
Innovations in cotton processing at a glance
| When | Innovation | Stage | What it changed |
|---|---|---|---|
| By the 5th century AD | Single-roller gin | Ginning | A roller pressed over seed cotton on a flat base pinches the seed out of the lint |
| Centuries old in Asia; in North America by 1750 | Churka (double-roller) gin | Ginning | About 2.3 kg (5 lb) of lint a day, five times the single roller |
| 1769 | Richard Arkwright’s water-driven spinning frame | Spinning | Roller drafting and water power; better yarn, made faster, and a far larger British demand for raw cotton |
| Patented 1770 | James Hargreaves’s spinning jenny | Spinning | One spinner works many threads at once instead of one on a wheel |
| 1772 to 1779 | Samuel Crompton’s mule | Spinning | Arkwright’s rollers plus the jenny’s moving carriage: fine yarn for weft |
| 14 March 1794 | Whitney’s spike-tooth gin | Ginning | A worker’s daily task of 50 lb of lint from green-seed (Upland) cotton |
| 12 May 1796 | Holmes’s saw gin | Ginning | Circular saws and continuous seed discharge; still the gin for Upland cotton |
| 1840 | Fones McCarthy’s reciprocating-knife roller gin | Ginning | 68 to 91 kg (150 to 200 lb) of lint a day per stand |
| Early 1900s; 1930s; 1940s | Seed-cotton cleaners; tower dryers (USDA-ARS); lint cleaners | Gin cleaning | Took out the leaf, sticks, burs and moisture that rougher hand harvesting, and later machines, brought to the gin |
| After World War II | Mechanical spindle pickers | Harvest | Shorter harvest season, so gins needed more capacity, drying and cleaning |
| 1963 | Rotary-knife roller gin | Ginning | Pima at one 218 kg bale an hour, five times the reciprocating-knife gin |
| 1967 | Rotor (open-end) spinning, BD 200 | Spinning | First rotor spinning mill, 10 machines, at Ústí nad Orlicí, Czechoslovakia |
| 1971 to 1974 | Cotton module builder | Harvest to gin | Seed cotton stacked in weatherproof modules in the field; no waiting trailers |
| 1978; 1997 | Murata Jet Spinner; Murata Vortex Spinner | Spinning | Air-jet spinning; vortex yarn spun directly from sliver |
| 1991 crop | HVI classing required for loan cotton | Classing | Every loan bale classed by instrument; today USDA classes practically all US cotton, with HVI measuring length, uniformity, strength, micronaire, colour and trash |
| 1999 | Batt-less saw lint cleaner (Lummus) | Gin cleaning | Lint cleaning with fewer neps and less short fibre than the controlled-batt type |
| 2005 | High-speed roller gin | Ginning | Pima ginned four times and Upland three times faster than on a conventional rotary-knife gin |
Each row is explained, with its source, in the sections below: the result of the new technology, what one gin could do, ginning, cleaning at the gin, harvest and modules, classing and selling raw cotton and the spinning mill.
What was the result of new technology in cotton manufacturing?
In the United States the decisive machine was the gin. Planters inland could grow only a green-seed cotton whose sticky seeds were slow to separate; the long-staple cotton that separated easily grew only on the coast. Roller gins were already used on green-seed cotton (Hughs et al. 2020), but they were slow. Whitney’s gin, and the saw gins that followed it, removed that bottleneck; as Whitney wrote to his father, “One man and a horse will do more than fifty men with the old machines.” The US National Archives sums up the result: “the yield of raw cotton doubled each decade after 1800”, by 1850 America grew three-quarters of the world’s supply, most of it sent to New England or exported to England to be made into cloth, and by the mid-19th century the South provided three-fifths of America’s exports, most of it cotton.
The same source calls the growth of slavery the gin’s most significant effect. The gin cut the labour of removing seed, but not the labour of growing and picking cotton, and cotton became so profitable that planters demanded more land and more enslaved workers. There were six slave states in 1790 and fifteen in 1860; Southerners imported 80,000 Africans between 1790 and the 1808 ban on the trade; and by 1860 about one in three Southerners was enslaved. Whitney struggled to profit from the gin: a loophole in the 1793 patent act let others copy it, and the Archives notes that his idea drew on earlier gins and on ideas from others, including Catherine Greene and enslaved workers.
In Britain the new technology was in the mill. Arkwright’s water-driven spinning frame of 1769 “greatly improved yarn quality and production speed for British spinners”, and together with other textile improvements it raised Britain’s need for raw cotton just as the young United States needed trading partners (Hughs et al. 2020). The jenny (patented 1770) let one spinner work many threads; the Science Museum’s record of it says one spinner could spin 120 threads at a time, against one on a spinning wheel (object 1912-319). Crompton’s mule, built between 1772 and 1779, joined Arkwright’s rollers to the jenny’s moving carriage to spin fine weft (replica of Crompton’s mule). The Archives notes that demand for the South’s cotton was fuelled by these machines to spin and weave it, and by the steamboat to carry it. The Science Museum’s record of Arkwright’s carding machine names the other results: production moved from spinners’ and weavers’ homes into factories, where one power source drove a sequence of special-purpose machines, and “British machine spinning led to the destruction of the Indian cotton spinning industry, while demand for raw cotton sustained the slave economy in the USA” (object 1860-7). The wider industrial story is told on our page on industrial textile history, and the hand methods they replaced on spinning and weaving by hand.
How much one gin could do, then and now
The clearest measure of each innovation is output. The figures below come from the USDA Agricultural Research Service’s account of the gin’s development, published in the Journal of Cotton Science as a chapter of the Cotton Ginners Handbook (Hughs et al. 2020). A US bale in these statistics is 218 kg (480 lb) of lint unless stated.
| Gin | When | Output stated in the source |
|---|---|---|
| Churka double-roller gin, hand-cranked | Centuries old; in North America by 1750 | 2.3 kg (5 lb) of lint a day |
| Whitney’s spike-tooth gin, hand-turned | 1793 letter; patented 1794 | A worker’s set task of 50 lb of lint a day; paid per pound above that task, 60 to 80 lb a day for 10 to 15 days in a row (Whitney’s 1793 letter to Thomas Jefferson) |
| Double-roller gins of Sea Island planters | 1840 | About 11 kg (25 lb) of long-staple lint a day per stand |
| McCarthy reciprocating-knife roller gin | 1840 | 68 to 91 kg (150 to 200 lb) a day per stand |
| Line-shaft ginnery with several saw gin stands | 1920s | Four to six 227 kg (500 lb) bales an hour on dry, well-harvested cotton |
| Rotary-knife roller gin (Pima) | 1963 | One bale an hour per stand, five times the reciprocating-knife gin |
| High-speed rotary-knife roller gin | First commercial plant 2005 | Four times faster on Pima, three times on Upland, than a conventional rotary-knife gin; 73 stands in service by 2009 |
| Today’s saw ginnery | 2020 | 20 to more than 80 bales an hour |
The same chapter reprints the National Cotton Ginners Association’s count of US gins, which shows the result over a century: fewer, far larger plants.
| Year | US production (running bales, thousands) | Number of gins | Average bales ginned per gin |
|---|---|---|---|
| 1900 | 9,345 | 29,214 | 320 |
| 1930 | 13,932 | 14,494 | 961 |
| 1960 | 14,170 | 5,395 | 2,627 |
| 1990 | 15,065 | 1,533 | 9,827 |
| 2005 | 23,253 | 891 | 26,098 |
| 2018 | 17,909 | 532 | 33,664 |
Between 1900 and 2018 the number of gins fell by 98% while the cotton each one ginned rose about a hundredfold. The authors put it down to changes in production practice, rising labour costs, and technical and mechanical innovation.
Ginning: how the cotton gin worked, and why roller gins came back
Whitney’s gin pulled lint off the seed with rows of sharpened iron spikes set in a wooden cylinder 15 to 23 cm across. The spikes dragged the fibre through slots in a metal bar too narrow for the seed, and a brush cylinder behind the bar cleared the spikes. It had two weaknesses: the spikes worked loose, and it ran in batches, so the ginner had to stop and dig out the cleaned seed. Holmes’s design replaced the spikes with circular saws, the slotted bar with flat-iron ribs, and opened the bottom of the roll box so that seed fell out continuously. Buchele and Mayfield (2016), cited by Hughs et al., argue that Holmes may have held a caveat of invention from 1789, five years before Whitney’s patent; whoever first had the idea, it is Holmes’s continuous saw gin that the industry adopted (Hughs et al. 2020).
Saw gins are fast but rough. On Upland cotton, whose fibre clings to a fuzzy seed, the speed was worth it, and the saw gin still gins Upland. On long-staple Sea Island and later Pima cotton, saw ginning broke fibres and added neps, so roller gins carried on: McCarthy’s reciprocating knife from 1840, the rotary-knife gin by 1963, and the high-speed roller gin developed at the USDA-ARS Southwestern Cotton Ginning Research Laboratory, first installed commercially in 2005. Roller-ginned Upland then found a market of its own: by 2011 California ginned 188,000 bales of it, for a reported mill premium of 13 to 26 US cents a kilogram (Hughs et al. 2020). The steps inside today’s gin plant are described on our pages on cotton ginning and ginning for textile manufacturing.
Cleaning at the gin: dryers, seed-cotton cleaners and lint cleaners
Early cotton was picked by hand and arrived at the gin fairly clean. As crops grew, harvesting became faster and rougher, then mechanical: strippers before the 1940s and spindle pickers after World War II. Seed cotton reached the gin with more leaf, stick, bur and moisture. Gins answered in stages: seed-cotton cleaners from the early 1900s, tower dryers developed by USDA-ARS and adopted in the 1930s, and lint cleaners after the gin stand from the 1940s. By the 1960s most US gins had at least two drying stages, several seed-cotton cleaning stages and at least two stages of lint cleaning (Hughs et al. 2020; van der Sluijs 2020). A modern plant for machine-picked cotton runs in this order: module feeder, dryer, hot-air cleaner, stick machine, inclined cleaner, extractor feeder, saw gin stand, air and saw lint cleaners, battery condenser and bale press.
How much each stage removes was measured at two commercial gins in Moree, New South Wales, in the 2015 season. Each ginned 18 modules of spindle-picked Upland cotton from the same field (van der Sluijs 2020). The seed cotton arrived with 7.3% and 8.2% trash, calculated from the lint weight and an assumed seed content of 47.2%. The seed-cotton cleaners removed 22.7% of it at Gin A and 32.8% at Gin B, and the lint cleaners took most of the rest. Lint turnout, the share of the seed cotton’s weight that leaves as lint, was 42.4% at Gin A and 41.7% at Gin B.
Worked example: Gin B’s seed-cotton cleaning. Cleaning efficiency = (trash in − trash out) ÷ trash in × 100 = (8.2 − 5.5) ÷ 8.2 × 100 = 32.9%. Per tonne of seed cotton that is 27 kg of trash taken out, leaving 973 kg. The trash and cleaning efficiency calculator, opened with these figures, shows 32.93% and 27.00 kg. The paper reports 32.8%, presumably because it works from unrounded values. Throughput is entered as 1,000 kg so that the panel reads per tonne. Lint lost is set to zero because the study did not measure it at these machines, so the “clean fibre” line here is cleaned seed cotton.
Lint cleaners are where cleaning starts to cost fibre. The same study sampled the lint after the gin stand and after the second of two saw lint cleaners. Gin A ran batt-less saw lint cleaners, the type Lummus introduced in 1999. Gin B ran the controlled-batt type, based on cleaning principles developed in the 1940s.
| Measured on the lint (HVI and AFIS) | Gin A, two batt-less saw lint cleaners | Gin B, two controlled-batt saw lint cleaners |
|---|---|---|
| HVI trash, % of surface area | 0.47 → 0.18 | 0.43 → 0.19 |
| HVI colour grade | 41-1 (Strict Low Middling) → 31-1 (Middling) | 31-2 (Middling) → 21-2 (Strict Middling) |
| Upper-half mean length, mm | 31.18 → 30.58 | 30.88 → 30.05 |
| Length uniformity, % | 82 → 82 | 83 → 81 |
| Total neps per gram (AFIS) | 248 → 324 | 212 → 334 |
| Lint turnout of the gin | 42.4% | 41.7% |
Both types bought a better colour grade with shorter fibre and more neps. The author puts Gin B’s lower turnout down to its more aggressive controlled-batt cleaners, which removed usable lint as well as trash. A larger US study of 2005 and 2006, cited in the same paper, found that each controlled-batt saw lint cleaner cut fibre length by 0.38 mm and length uniformity by 0.7%, and added up to 45 neps per gram. More than two saw lint cleaners is generally discouraged for that reason. The paper also reports that gin process controls measuring moisture and leaf grade on line were being installed in many gins, to help ginners limit fibre damage and keep turnout up. The site’s records for ASTM D2812 (non-lint content) and ASTM D5866 (neps, the AFIS method used here) cover how these quantities are tested.
Harvest to gin: the module builder
Seed cotton used to travel from field to gin in wagons, trucks or trailers, and harvests stalled while haulers waited at the gin to unload. Professor Lambert Wilkes of Texas A&M University developed the cotton module builder between 1971 and 1974, with support from Cotton Incorporated. It pressed seed cotton into compact stacks in the field that stood up to weather and could be fed straight into the gin later. When the American Society of Agricultural and Biological Engineers made it a historic landmark in 2002, it was used for more than 90% of the US crop (ASABE). Pickers now form round modules on board, and that has pushed gins towards still higher capacity (Hughs et al. 2020). The Moree trial’s field was harvested with one, a John Deere 7760.
Classing: how raw cotton is graded and sold
Raw cotton is traded bale by bale on its measured quality, and that system is itself an innovation. US grade standards grew from a 1907 meeting of the international trade in Atlanta. Under USDA classification today, the gin presses lint into bales of about 500 pounds and tags each one with a 12-digit Permanent Bale Identification number. A licensed agent cuts a sample of at least 4 ounces (115 g) from each side of the bale. The sample goes to one of ten USDA classing offices, where high volume instruments (HVI) measure fibre length, length uniformity, strength, micronaire, colour and trash in seconds; leaf grade is now calculated from the HVI trash reading, and only extraneous matter and special conditions are still judged by a classer. At the peak of the season USDA classes up to two million bales a week (The Classification of Cotton, USDA Agricultural Handbook 566, Cotton Incorporated edition, 2018).
The switch to instruments was made compulsory through the loan programme: beginning with the 1991 crop, all cotton needed an HVI classification to be eligible for the price support loan (Mayfield, 1990 Beltwide Cotton Conferences). Module averaging, a voluntary programme offered since 1991, assigns every bale in a module the average of its bales’ HVI readings for strength, length, uniformity and micronaire. Classing data sit in USDA’s National Database, open to the bale’s current owner. That is how merchants buy and sell raw cotton by its numbers, and how a mill knows what it holds. Cotton Incorporated’s Engineered Fiber Selection (EFS) software then lets a mill plan its bale laydowns from those HVI values, so each mix spins the same yarn (AH 566; see our page on fibre blending).
The tests behind the numbers are standardised. HVI testing is ASTM D5867, micronaire is ASTM D1448 and ISO 2403, bundle strength is ASTM D1445/D1445M, and the lighting of classing rooms is ASTM D1684. The site’s micronaire converter turns a classing micronaire into fineness, and the HVI replicate and outlier checker flags readings that disagree. What the HVI values mean for the fibre itself is set out on our cotton fibre page.
In the spinning mill: the line and what changed
From the bale, cotton for ring-spun yarn passes through the usual sequence of machines of a short-staple mill. Each stage takes something out or evens something up:
| Stage | What it does | What it takes out | A dated innovation |
|---|---|---|---|
| Bale opening and blowroom | Opens bales into small tufts, blends many bales, cleans | Trash and dust | Laydowns planned by software from each bale’s HVI data |
| Carding | Separates the tufts into single fibres on wire-clothed cylinders and forms a sliver | Fragments of leaf and seed; the card also straightens the fibres | Arkwright’s carding machine of about 1775, close to Daniel Bourn’s cylinder card of 1748 |
| Drawing | Combines several slivers and draws them out between pairs of rollers | Unevenness (nothing is removed) | Roller drafting, as on Arkwright’s spinning frame of 1769 |
| Combing, for fine yarn | Combs out the short fibres from laps of carded cotton | Comber noil: fibres shorter than half an inch | Automatic lap change and piecing at the comber (Rieter ROBOlap) |
| Roving and ring spinning | Draws the sliver finer and inserts twist | Nothing; this is where yarn is made | Mule (1772 to 1779) and jenny (patented 1770) before the ring |
| Rotor or air-jet spinning, instead of roving and ring | Spins drawn sliver directly into yarn on a package | Rotor: impurities, at a trash-removal device before the rotor (BD 200-RC, 1976) | First rotor mill 1967; air-jet 1978; vortex 1997, up to 550 m/min on the 870 EX |
Sources for the table: carding machine, Science Museum Group; comber noil length, Cotton Incorporated (comber noils under 0.5 in); laydown software, AH 566; rotor spinning, Rieter, link 72 (2017); air-jet and vortex spinning, Murata Machinery history and VORTEX page. The machines themselves are explained on our pages on the blowroom and carding, and their production arithmetic in spinning formulas.
Rotor spinning shows how a mill innovation happens. Ring spinning had reached its technical limits, and Czechoslovakia set out after the war to raise yarn output. Staff of the Cotton Industry Research Institute (VÚB) in Ústí nad Orlicí began spinning trials in 1958. A first complete machine ran in May 1962. The decisive step was an opening roller that fed single fibres into the spinning rotor. In August 1967 the world’s first rotor spinning mill opened at Ústí with 10 BD 200 machines. At ITMA 1971 in Paris, 11 makers showed 15 rotor spinning exhibits (Rieter, link 72). Murata developed its air-jet Murata Jet Spinner in November 1978 and showed the Murata Vortex Spinner No. 851 at OTEMAS in October 1997. Its VORTEX III 870, released in 2011, reached a maximum speed of 500 m/min, and the 870 EX 550 m/min (Murata history; VORTEX page).
Questions people ask
What was the result of new technology in cotton manufacturing and production?
The cotton gin made short-staple cotton profitable to grow inland, so US raw cotton output doubled each decade after 1800, and by 1850 the United States grew three-quarters of the world’s cotton. Because the gin cut only the labour of removing seed, not of growing and picking, it greatly increased the demand for enslaved labour: slave states rose from six in 1790 to fifteen in 1860. Most of that cotton was made into cloth in New England or English mills, on the new spinning and weaving machines.
What are some innovations and advancements made in the cotton industry?
In order: the churka roller gin; machine spinning (the water frame of 1769, the jenny patented in 1770, the mule of 1772 to 1779); Whitney’s gin (1794) and Holmes’s saw gin (1796); the McCarthy roller gin (1840); gin dryers and cleaners (1900s to 1940s); mechanical pickers after World War II; the rotary-knife roller gin (1963); rotor spinning (1967); the module builder (1971 to 1974); air-jet spinning (1978, vortex 1997); HVI classing of every loan bale (1991 crop); and the high-speed roller gin (2005).
What is raw cotton merchandising?
It is the trade in ginned lint between the grower or gin and the spinning mill. In the United States each bale carries a permanent identification number and its own USDA classing record, measured by HVI, and the record passes to each later owner of the bale, mainly merchants and mills, so bales are bought and sold on those numbers. Mills then use the same data to plan their laydowns.
How did Eli Whitney’s cotton gin work?
A hand-turned wooden cylinder studded with iron spikes pulled the lint through slots in a metal bar. The slots were too narrow for the seed, so the seed stayed behind, and a brush cylinder stripped the lint off the spikes. Two years later (patent of 12 May 1796), Holmes’s gin replaced the spikes with circular saws and let the seed fall out continuously, and that saw gin is still used for Upland cotton.
What role does technology play in turning cotton into yarn?
Every step after the gin is a machine: bale openers and blowroom cleaners, the card, the drawframe, the comber for fine yarn, then ring, rotor or air-jet spinning. The quality data from HVI classing decide which bales go into each mix.
Which machines are used for cotton processing?
At the gin: module feeder, dryers, seed-cotton cleaners (hot-air cleaner, stick machine, inclined cleaner, extractor feeder), the saw or roller gin stand, lint cleaners, condenser and bale press. In the mill: bale opener, blowroom cleaners and mixers, card, drawframe, comber, roving frame and ring frame, or a rotor or air-jet machine in place of the last two. Long-staple cotton such as Pima is roller-ginned and Upland is saw-ginned, because the saw gin damages long staple.
Sources
- US National Archives, Eli Whitney’s Patent for the Cotton Gin (educator resource, reviewed 16 December 2021): output, exports and slavery figures, and Whitney’s letter to his father.
- S. E. Hughs, G. A. Holt, C. B. Armijo, D. P. Whitelock and T. D. Valco (USDA-ARS), “Development of the Cotton Gin”, Journal of Cotton Science 24 (2020): 34–43, a chapter of the Cotton Ginners Handbook: gin history, the patent dates, the output figures and the table of US gins (courtesy of the National Cotton Ginners Association).
- M. H. J. van der Sluijs, “The Effect of Various Processing Stages During Ginning on Fiber Quality”, Journal of Cotton Science 24 (2020): 44–59: the Moree gin trial, trash, turnout and lint-cleaner results, and the dates of seed-cotton cleaners, dryers and lint cleaners.
- Science Museum Group collection: spinning jenny (1912-319), Crompton’s mule and Arkwright’s carding machine (1860-7).
- American Society of Agricultural and Biological Engineers, Cotton Module Builder, historic landmark (2002).
- USDA Agricultural Marketing Service, The Classification of Cotton, Agricultural Handbook 566, edition produced by Cotton Incorporated with USDA (2018).
- W. Mayfield, “Improving Quality under a Modern Grading System: Changes to Expect, 1991 and Beyond”, 1990 Proceedings, Beltwide Cotton Production Research Conferences, p. 514 (National Cotton Council).
- Rieter, “The Ball is Rolling: a journey through 50 years of rotor spinning”, link 72 (2017); Rieter, ROBOlap product page.
- Murata Machinery, textile machinery history and VORTEX spinning machine.
- Cotton Incorporated, Cotton for Nonwovens Technical Guide: Cotton Preparation (fibre lengths of comber noils, gin motes and linters).
How this page is kept
Last verified 12 October 2026. History and output figures come from the US National Archives, the USDA-ARS gin history in the Journal of Cotton Science, and Science Museum Group records. Gin trial data come from the peer-reviewed study cited. Classing comes from the USDA handbook and a 1990 Beltwide paper, the module builder from ASABE, and spinning dates from Rieter and Murata, the makers concerned. The worked example is arithmetic on the study’s published figures, and the calculator link opens with them, checked on this date. Maker claims, such as the ROBOlap efficiency gain, are attributed to the maker.
Change log. 12 October 2026: rewritten in full. The earlier version gave no dates, no machines by name and no sources. It described unnamed “case studies” of a mill that cut labour costs by 30%, another that raised production by 35%, and plants that halved water use; none could be traced, and all have been removed. It also said the cotton gin improved fibre quality. The saw gin raised capacity, not quality: it damages long staple, which is why long-staple cotton is still roller-ginned. An image of fabric abrasion testing captioned as fibre testing, a watermarked loom photograph and a garment-factory photograph have been removed.
Found an error? Please tell us through the contact page, naming the section and the source that disagrees.
