¹ Grapes, , earning over one billion dollars, were the second leading crop in California. California Department of Finance, California Statistical Abstract, 1980 (Sacramento, 1980), pp 96–98.Google Scholar

² See, for example, Street, James H., The New Revolution in the Cotton Economy: Mechanization and Its Consequences, (Chapel Hill, 1957);Google ScholarFite, Gilbert C., “Recent Progress in the Mechanization of Cottom Production in the United States,” Agricultural History, 24 (01 1950), 19–28, and vol. 54 (January 1980), 190–207;Google ScholarDay, Richard H., “The Economics of Technological Change and the Demise of the Sharecropper,” American Economic Review, 57 (06 1967), 427–49;Google ScholarMusoke, Moses S., “Technical Change in Cotton Production in the United States, 1925–1960” (Ph.D. Dissertation, University of Wisconsin-Madison, 1976).Google Scholar

³ See, for instance, Street, The New Revolution, pp. 18–34;Google ScholarFulmer, John L., Agricultural Progress in the Cotton Belt Since 1920 (Chapel Hill, 1950), pp. 65–72;Google ScholarTang, Anthony, Economic Development in the Southern Piedmont, 1870–1950 (Chapel Hill, 1958);Google ScholarNicholls, William H., Southern Tradition and Regional Progress (Chapel Hill, 1960);Google Scholar and Parker, William N., “The South in the National- Economy, 1865–1970,” Southern Economic Journal, 46 (04 1980), 1019–48.CrossRefGoogle Scholar

⁴ Mumford, E. Philpott, “Early History of Cotton Cultivation in California: Research Among the Old Spanish Manuscripts,” California Historical Society Quarterly, 6 (06 1927), 156–66;Google ScholarHilgard, Eugene W., Report on the Physical and Agricultural Features of the State of California with a Discussion of the Present and Future of Cotton Production in the State, U.S. Census Report, (Washington, D.C., 1884), pp. 73–78;Google ScholarPacific Rural Press (11 November 1911), p. 393, (22 January 1916), p. 96.Google Scholar

⁵ U.S. Department of Agriculture, Statistics on Cotton and Related Data, Statistical Bulletin 99 (Washington, D.C., 1951), p. 6;Google ScholarCook, O. F., Expansion of Cotton Production in California, U.S. Department of Agriculture, Department Bulletin 533 (Washington, D.C., 1917), pp. 1–3;CrossRefGoogle ScholarPacific Rural Press (17 February 1917), p. 204.Google Scholar

⁶ Statistics on Cotton, 1951, p. 38.Google Scholar

⁷ For an excellent analysis of the San Joaquin valley environment, see Blick, James D., “An Analysis of Cotton Production in the San Joaquin valley, California” (Ph.D. Dissertation, University of California, Los Angeles, 1956).Google Scholar

⁸ This estimate rests upon an unweighted average of USDA estimates of yield losses. Statistics on Cotton, 1951, p. 67.Google Scholar

⁹ Ibid., p. 67.

¹⁰ Camp, Wofford B., Production of Acala Cotton in the San Joaquin Valley of California, U.S. Department of Agriculture, Department Circular 357 (Washington, D.C., 1925), pp. 2–4;Google ScholarCalifornia Cotton Journal (November 1925), p. 11;Google ScholarPacific Rural Press (6 May 1922), p. 523.Google Scholar

¹¹ Wofford B. Camp: Cotton, Irrigation and the AAA, an Interview conducted by Willa Klug Baum, University of California Regional History Office, Bancroft Library, Berkeley, 1971, pp. 72–78;Google ScholarSan Joaquin Power Magazine, 7 (09 1925), p. 4;Google ScholarCalifornia Cotton Journal (December 1925), pp. 11, 20; (April 1926), p. 11; Pacific Rural Press (9 Setember 1934), p. 248.Google Scholar

¹² For a discussion of the value of uniform quality, see Classification of Cotton, U.S. Department of Agriculture, Miscellaneous Publication 310 (Washington, D.C., 1938), p. 43.Google Scholar

¹³ Camp, Production of Acala, p. 2; Pacific Rural Press (21 November 1925), p. 552.Google Scholar

¹⁴ Herbert, F. W., quoted in the Pacific Rural Press (19 September 1925), p. 291.Google Scholar

¹⁵ Harrison, George J., “A History of Cotton in California,” 1955, typed manuscript in Giannini Library of Agricultural Economics, University of California, Berkeley, pp. 57–66;Google Scholar “Your Cotton Seed Came This Way,” California Farmer (18 June 1948), p. 624.Google Scholar

¹⁶ “California's Cotton Rush,” Fortune (May 1949), pp. 134, 136. The Acala P-18 strain had been adopted in 1944. Alter twenty years of continual improvements in the 4–42 strain, Acala SJ-l was introduced in 1967. See John H. Turner, History of Acala Cotton Varieties Bred for San Joaquin Valley, California, U.S. Department of Agriculture, Agricultural Research Service, ARS W-16 (February 1974); Harrison, “Cotton in California,” pp. 63–66.Google Scholar

¹⁷ U.S. Bureau of the Census, Sixteenth Census of the United States: 1940, Agriculture: Special Cotton Report, (Washington, D.C., 1943). We chose Texas, Mississippi, and South Carolina as representative states for comparison with California. All three were important cotton producing states, and they make up important parts of several distinct regions of production—the High Plains of the Southwest, the Delta, and the Piedmont and Coastal Plains. Throughout this paper, we concentrate on California, although major cotton growing areas in Arizona and New Mexico had similar characteristics and mechanization rates.Google Scholar

¹⁸ Data distinguishing between plantations and “farms” were first available in 1944. “Operating” units include both multiple-unit and single-unit operations. Given time for recontracting, the operators had considerable authority over production, as well as land utilization decisions. National Cotton Council of America, Special Report: Cotton Farms, Classified by Acreage Harvested. Prepared by the U.S. Bureau of the Census (Memphis, 1947), pp. 5, 13, 16–17, 44.Google Scholar

¹⁹ For changes in farm layout, see Pedersen, Harald A. and Raper, Arthur F., The Cotton Plantation in Transition: The Case Studies of a Mechanized and an Unmechanized Cotton Plantation in the Yazoo-Mississippi Delta, Mississippi State Agricultural Experiment Station Bulletin 508 (State College, 1954). The distinction between farms and operating units within South Carolina was not nearly so dramatic as in Mississippi. The smaller operating units in South Carolina may help account for the slower diffusion of mechanical pickers in that state.Google Scholar

²⁰ U.S. Bureau of the Census, Special Cotton Report (1943), pp. 2–15.Google Scholar

²¹ U.S. Bureau of the Census, U.S. Census of Agriculture: 1959, Vol. II, General Report: Statistics by Subjects (Washington, D.C., 1962), p. 214.Google Scholar

²² Pacific Rural Press (2 April 1927), p. 456. One of the more notable growers in Kern County was Herbert Hoover who regularly raised 400 acres of cotton on his 1,200 acre farm during the 1920s. See Los Angeles Times, Farm and Tractor Section (8 May 1921), p. 18; California Cotton Journal (April 1926), p. 16.Google Scholar

²³ California Committee to Survey the Agricultural Labor Resources, Agricultural Labor in the San Joaquin Valley: Final Report and Recommendations (Sacramento, 15 03 1951), pp. 177–196;Google ScholarFisher, Lloyd H., The Harvest Labor Market in California (Cambridge, Massachusetts, 1953).CrossRefGoogle Scholar

²⁴ U.S. Department of Labor, Bureau of Labor Statistics, Labor Unionism in American Agriculture, Bulletin No. 836 (Washington, D.C., 1945), pp. 30–42, 100–105;Google ScholarU.S. Congress, Senate, Subcommittee of the Committee on Education and Labor, Hearing on S. Res. 266, Violations of Free Speech and Rights of Labor, 74th Congress, 2nd sess. (Washington, D.C., 1939), Pt. 51, 54;Google ScholarDaniel, Cletus E., Bitter Harvest: A History of California Farmworkers, 1870–1941 (Ithaca, 1980), esp. pp. 167–221.Google Scholar

²⁵ Galarza, Ernesto, Farm Workers and Agri-business in California, 1947–1960 (Notre Dame, 1977), p. 121–27.Google Scholar

²⁶ Holley, William C. and Arnold, Lloyd E., Changes in Technology and Labor Requirements in Crop Production: Cotton, WPA, National Research Project Report No. A-7 (Philadelphia, 1938), p. 99;Google ScholarFree, Benjamin J., Seasonal Employment in Agriculture (Washington, D.C., 09 1938), pp. 21–23, 50–51. In this last report, all crops, not just cotton, are considered in the estimate. The peak employment in all crops during the chopping season exceeded that during picking in the deep South.Google Scholar

²⁷ Pedersen and Raper, The Cotton Plantation, p. 4.Google Scholar

²⁸ California State Relief Administration, Survey of Agricultural Labor Requirements in California, 1935 (December 1935), p. 239;Google ScholarAdams, R. L., Seasonal Labor Requirements for California Crops, California Agricultural Experiment Station Bulletin 623 (Berkeley, 07 1938), p. 16.Google Scholar

²⁹ “Flattening Labor Peaks Challenges Designers,” Implement Record (March 1950), p. 19; California Committee, Agricultural Labor in the San Joaquin Valley, pp. 135–55.Google Scholar

³⁰ Higher nominal wage rates, of course, do not necessarily mean that cotton pickers in California were better off than those in other regions, as San Joaquin Valley growers had asserted. One piece of evidence for a higher “real” wage in California cotton fields was the net migration of southern cotton pickers to the arid West. But this indicator is complicated by the possibility that cotton picking was just familiar, temporary employment for migrants seeking higher-paying jobs in the urban areas. While sorting out all of these conditions is beyond the scope of this paper, one welfare judgment seems justified: few cotton pickers in any region were really well off. On this point, there is general agreement.Google Scholar

³¹ Harrison, “Cotton in California,” p. 79.Google Scholar

³² For example, James Schooley built and operated a suction-type machine at Scaly in 1917. In 1918 a suction machine with five nozzles which purportedly picked 5000 pounds of seed cotton per day was demonstrated in the Imperial Valley. In 1921, several machines were tried and found wanting. For notes on these and other experiments, see: Scientific American (19 October 1918), pp. 315, 316, 326, 327; Los Angeles Times, Farm and Tractor Section (11 July 1920), p. 8, (16 January 1921), p. 14, (23 April 1922), p. 13, (21 May 1922), p. 11, (13 August 1922), p. 6, (11 November 1922), p. 8.Google Scholar

³³ The Harris Manufacturing Company claims to have used six machines at the Tagus Ranch in Tulare County. These machines picked cotton using two revolving brushes. The Gladiator Manufacturing Company, the Double-Creeper Tread Company, and the Smith-Haughton Company also fielded machines in 1927. Harrison, “Cotton in California,” pp. 79–80.Google Scholar

³⁴ Strippers proved ill-suited to the vegetative strains of cotton grown in California, and to the present have been given limited use.Google Scholar

³⁵ Harrison, “Cotton in California,” pp. 80–83; H. B. Walker, “Cotton Harvesting by Mechanical Methods,” Implement Record (March 1932), pp. 16–17;Google ScholarWalker, H. B., “Mechanical Cotton Harvesting Experience in California,” Agricultural Engineering (September 1938), p. 392.Google Scholar

³⁶ Harrison, “Cotton in California,” pp. 79–80.Google Scholar

³⁷ Street, The New Revolution, pp. 128–29. Mack Rust moved to Coalinga in about 1941, and worked out of a converted oil refinery building provided to him free of charge by a local farmer, M. J. Allen. Rust and his wife, with modest backing from local farmers, attempted to perfect his machine over a span of five or six seasons without much success. Street's reference to Mack Rust's activities as a “custom harvesting business” is a generous exaggeration. More accurately, it may be described as a hand-to-mouth research and development effort. Mack Rust regularly promised to harvest in the upcoming season, but invariably the machine was not ready when the harvest began, and would only make brief forays into the field before being returned to the shop for refinements. From a conversation with M. J. Allen, Coalinga, California, 1 December 1980.Google Scholar

³⁸ Rust, John, “The Origin and Development of the Cotton Picker,” Western Tennessee Historical Society Papers, 7 (1953), 52–53.Google Scholar

³⁹ Conversations with Trimble Hedges, an agricultural economist at the University of California-Davis. Hedges conducted extensive interviews with machine owners in 1949 and 1950. Hedges interviewed about 15 operators of Rust machines. Of these, only two were satisfied; the remainder complained about continual breakdowns and an assortment of mechanical problems. Hedges noted that the two satisfied users were widely regarded as superior mechanics. Owners of international Harvester machines found their pickers far more reliable.Google Scholar

⁴⁰ Hedges, Trimble R. and Bailey, Warren R., Economics of Mechanical Cotton Harvesting: A Report of Studies Made in the San Joaquin Valley of California, California Agricultural Experiment Station Bulletin 743 (Berkeley, 1954);Google ScholarBailey, Warren R. and Hedges, Trimble R., Economics of Mechanical Cotton Harvesting: Study Made in San Joaquin Valley, 1949–50, California Agricultural Extension Service Unnumbered Publication (1951);Google ScholarBurlingame, Burt B. and Bailey, Warren R., Cost of Harvesting Cotton with Mechanical Pickers: California, 1948, California Agricultural Experiment Station Mimeo (Berkeley, 1950);Google ScholarVenstrom, Cruz, Experiences in 1945 With Mechanical Cotton Pickers in California, U. S. Bureau of Agricultural Economics Mimeo (October 1946).Google Scholar

⁴¹ These cost figures differ from the adjusted results (based on the assumption that each machine picked 300 bales) given prominence in Hedges's and Bailey's study. Their adjusted average savings is $19.25 per bale. We should note that both figures are within the range of savings claimed by Harold and Waldo Weeth, cotton farmers in the Coalinga area. Weeth's calculations of hand picking costs included money spent on maintaining worker camps. Hedges and Bailey do not include such non-field, labor expenses. Conversations with Weeth, Harold, 11 1980, and California Farmer (5 November 1949), p. 387, (17 June 1950), p. 577.Google Scholar

⁴² Again we emphasize that these figures are based on Hedges's and Bailey's actual survey results. Using their adjusted savings based on a 300-bale utilization rate yields savings of $9.81 in the North and $21.50 in the West Side.Google Scholar

⁴³ Grade loss represented a major cost in mechanical harvesting. The value of the loss in grade between machine and hand picked cotton found by Hedges and Bailey averaged $10.32 a bale. This, combined with an average field waste of $3.15 a bale, represented almost one half of the actual average cost of machine picking. if the machines were used to near capacity—that is, to pick 300 bales a year—grade loss and field waste would have accounted for more than one half of the valley-wide average cost. Moreover, differences in grade loss were a major component in regional cost differences found within California. Other important components of Hedges's and Bailey's cost estimates deserve mention. They use a 4 percent interest rate and assume that a picker, if used to capacity, lasted five years. They employ straight line depreciation techniques. They charge 54 percent of the tractor's fixed cost to the picking operation.Google Scholar

⁴⁴ Estimated from U. S. Bureau of the Census, Census of Agriculture, 1950, II, p. 630.Google Scholar The David-Olmstead exchange on the diffusion of the reaper raises the question to what extent cotton pickers were shared or rented. Unlike the reaper, it appears that early cotton pickers in California were not frequently shared or jointly purchased. Hedges could not recall any sharing arrangements. But an active custom market with itinerant harvesters analogous to those found in midwestern wheat fields did emerge early. Hedges and Bailey (p. 9) estimate, for example, that over 10 percent of the machines in California in 1951 had been brought in to do custom work exclusively. David, Paul A., “The Mechanization of Reaping in the Ante-Bellum Midwest,” in Rosovsky, Henry, ed., Industrialization in Two Systems: Essays in Honor of Alexander Gerschenkron (New York, 1966), pp. 3–39;Google ScholarOlmstead, Alan L., “The Mechanization of Reaping and Mowing in American Agriculture, 1833–1870,” this JOURNAL, 35 (06 1975), 327–52.Google Scholar

⁴⁵ The new practices included improved bed preparation, seed delinting, plant spacing, pesticide application, weed control, fertilizer use, thinning, irrigation, and defoliation.Google Scholar

⁴⁶ See Fairbank, J. P. and Smith, K. O., “Cotton Mechanization in California,” Agricultural Engineering (May 1950), 219–22;Google ScholarHarrison, George J., “Breeding and Adapting Cotton to Mechanization,” Agricultural Engineering (September 1951), 489–92;Google ScholarTavernetti, James R. and Miller, H. F. Jr, Studies on Mechanization of Cotton Farming in California, California Agricultural Experiment Station Bulletin 747 (Berkeley, 11 1954);Google ScholarTavernetti, James R. and Carter, Lyle M., Mechanization of Cotton Production, California Agricultural Experiment Station Bulletin 804 (Berkeley, 08 1964);Google ScholarBainer, Roy, The Engineering of Abundance, Interviews conducted by A. I. Dickman, Oral History Center, University of California-Davis (1975), p. 253.Google Scholar

⁴⁷ Harrison, “Breeding and Adapting Cotton,” pp. 486–88; Fairbank and Smith, “Cotton Mechanization in California,” p. 219.Google Scholar

⁴⁸ The Shafter station during the late 1940s and early 1950s annually offered Mechanization Field Days for cotton farmers that demonstrated the latest equipment and explained new cultural techniques. To top off these events, the National Cotton Council held its Sixth Annual Cotton mechanization Conference in the San Joaquin Valley in 1952. See Cotton Mechanization, Proceedings, Sixth Annual Cotton Mechanization Conference (Memphis, 1952). For the field days, see California Farmer, (24 September 1949), p. 256, (31 December 1949), p. 540, (5 September 1953), p. 206, (13 October 1956), p. 321, (9 May 1959), p. 520; Implement Record (April 1949), pp. 23, 55, (December 1950), pp. 23–24.Google Scholar

⁴⁹ Maier, Frank H., “An Economic Analysis of Adoption of the Mechanical Cotton Picker” (Ph.D. Dissertation, University of Chicago, 1969), p. 7.Google Scholar

⁵⁰ The rapid increase in the percentage of cotton mechanically harvested throughout the South stems in part from marginal lands being transferred to other uses. Federal acreage allotments along with increased competition from the western states and from newly mechanized southern farms led to sharp declines in southern cotton acreage. For example, between 1950 and 1960, cotton acreage in Mississippi declined 25 percent. U. S. Department of Agriculture, Satistics on Cotton (1974), p. 71.Google Scholar Another important trend was the evolution of southern cotton farms towards the structure of California farms. Between 1949 and 1969 the acreage in cotton per cotton farm in California increased 31 percent. During this same period the cotton acreage per Mississippi cotton farm increased 195 percent, and in South Carolina it increased 130 percent. U. S. Department of Commerce, Bureau of the Census, U. S. Census of Agriculture: 1959, Vol. II, General Report: Statistics by Subjects, (Washington, D. C., 1962) pp. 829, 835,Google Scholar and U. S. Census of Agriculture: 1969 Vol. 11, General Report, Chapter 6: Crops, Nursery and Greenhouse Products, Forest Products (Washington, D. C., 1973), p. 31.Google Scholar

⁵¹ The latter, of course, would also depend on past misuse resulting from institutional shortcomings.Google Scholar

⁵² Crowe, Grady B., Mechanical Cotton Picker Operation in the Yazoo-Mississippi Delta, Mississippi Agricultural Experiment Station Bulletin 465 (State College, 07 1949);Google ScholarBrunner, Warner L. Jr, and Stanley, Felix E., “Cost and Utilization of Mechanical Cotton Picker in the Delta Cotton Areas of Louisiana, 1948 and 1949.” Louisiana State University Department of Agricultural Economics Mimeograph Circular No. 117 (August 1950).Google Scholar

⁵³ For an analysis of these reports, see footnote 97 in Musoke, Moses S. and Olmstead, Alan L., “A History of Cotton in California,” Agricultural History Center, working Paper No. 6 (University of California, Davis, 07 1981).Google Scholar

⁵⁴ Hedges and Bailey, Economics of Mechanical Cotton Harvesting, pp. 18–20; Brunner and Stanley, “Cost and Utilization,” p. 9.Google Scholar

⁵⁵ Crowe, Mechanical Cotton Picker Operation, pp. 6–7.Google Scholar

⁵⁶ Hedges and Bailey, Economics of Mechanical Cotton Harvesting, pp. 4, 46–48.Google Scholar

⁵⁷ Brunner and Stanley, “Cost and Utilization,” p. 7. For Crowe's study, yields are never explicitly stated, but may be inferred from related statements. Crowe, Mechanical Cotton Picker Operation, p. 10.Google Scholar Besides reducing yields, southern climatic conditions also induced a far greater variance in yields than existed in California, creating another problem for mechanization. As a rough approximation, between 1928 and 1950, yield variability (as measured by the coefficient of variation) at any given location in the South was about twice that found in the San Joaquin Valley. This had two effects. First, with capital market imperfections there was an increased probability of being unable to make the payments on the larger fixed costs associated with mechanization. Second, the expected value of the costs of mechanically harvesting a given acreage increases as the variance of yield increases. The overall impact is to increase the expected costs of mechanization. This conclusion requires only that the cost function is convex from the origin; empirical evidence indicates that it was. Botts, Ralph R., Variability of Cotton Yields by Counties in the United States, U. S. Department of Agriculture, Bureau of Agricultural Economics (Washington, D. C., 08 1952), mimeo.Google Scholar

⁵⁸ The variable costs associated with field waste and grade loss were typically higher in the South than in California. In the Delta, the reported field waste ranged between 5 and 8 percent during the 1947–1949 period. In the San Joaquin Valley, only 3 percent was lost. Reports for all regions claim that hand picking deteriorated everywhere in the late 1940s and 1950s, which largely explains the drop in field waste attributed to the machine over time. But southern environmental conditions tied to the varieties of cotton grown and climatic conditions reduced machine picker efficiency in the Delta relative to California. A comparison of California and Arizona offers evidence of the importance of environmental factors. The institutional structures were very similar but field waste in Arizona exceeded that of California by a factor of 3 to 4 times. This difference was largely a result of the different varieties grown. The reported value of the grade loss due to mechanical picking was about the same in the two regions. But the reported figures probably understate the true interregional differences. The quality loss could be reduced by extra attention to weeding and general field preparation. Descriptive evidence indicates that farmers who mechanized had to maintain cleaner fields and that Delta farmers needed to devote more effort to obtain acceptable field conditions. These added expenditures are not included in the reported costs of harvesting. In addition, the use of defoliants can reduce grade loss. In the late 1940s, California cotton growers defoliated both hand and machine harvested fields. In the Delta, defoliants were much more likely to have been used only in conjunction with machine harvesting. Inclusion of defoliation expenditures would add about $2.00 a bale to the cost of machine picking in the South. See Brunner and Stanley, “Cost and Utilization,” pp. 22–23; and Crowe, Mechanical Cotton Picker Operation, pp. 10–11.Google Scholar

⁵⁹ Since California represents our institutional basis of comparison, its practices at any point in time are “advanced” by definition. The fact that California farmers were well educated, operated on a large scale, were not shackled by long-term labor contracts (and were thus relatively free to substitute labor and capital to meet changing market conditions), and were tied into a well-developed, organized research network, provides ample justification for our definitional choice.Google Scholar

⁶⁰ Leavitt, F. Haven, “Agricultural Ammonia Equipment: Development and History,” in Agricultural Anhydrous Ammonia, Technology and Use, McVickar, Malcom, ed. (Memphis, 1966), pp. 127–28;Google ScholarStreet, The New Revolution, pp. 137–38.Google Scholar

⁶¹ It is also important to understand that some of the increase in yields resulted from mechanization. The closer association of costs and yields with the mechanical mode of production created an incentive to use more fertilizer at the margin.Google Scholar