Mass Immigration and Race Classification in the United States Census

While Census Bureau officials rarely voiced openly racist views, they were, like most intellectuals in the early twentieth century, influenced by an ascendant belief in inherited racial distinctions that affected character and behavior.^{Footnote 2} They were also sensitive to the effects of immigrants from new origins. Francis Amasa Walker, director of the 1870 and 1880 censuses and the intellectual patriarch of the bureau, had, like Henry Cabot Lodge and other restrictionists, begun to voice a neo-Lamarckian view of hereditary inferiority in Southern and Eastern European immigrants. Joseph A. Hill, assistant director for the 1920 and 1930 censuses and a prominent figure in the bureau, epitomized Walker’s legacy. In the first decades of the twentieth century, he and other bureau staff engaged in a debate about the meaning of race that led to the creation of the Mexican race category.^{Footnote 3}

Joel Perlmann provides close analyses of deliberations within the bureau, a discussion influenced by the Dillingham Commission’s extensive and politicized assessment of immigration.^{Footnote 4} The debate did not result in a firm conclusion: among bureau staff, race remained an ambiguous term, at times a synonym for ethnicity subject to change, especially through assimilation, and at other times implying a more permanent, hereditary, and biological condition. Concern about high levels of immigration in the early twentieth century was manifest in bureau studies and in the private commentary of staff. Aware of the diverse populations within European territorial boundaries, bureau staff resisted expansion of the race variable to include categories among white Europeans, but they did introduce a mother-tongue question in 1910 designed to identify ethnic or “racial” origins more accurately than country of birth. In planning the 1920 census, Hill saw as “more important than ever before” questions on nativity and language, the latter providing “fairly accurate racial classification of our white population of foreign birth or foreign parentage.”^{Footnote 5} Still, until 1930 the census race variable remained reflective of the widely accepted “grand divisions” of white, Indian, African American, and Asian origin categories.

Nothing in documents before the 1920s reveals a particular concern about Mexicans or their racial makeup. Daniel Folkmar’s influential Dictionary of Races and Peoples, produced for the Dillingham Commission, casually referenced Mexicans as largely of “Indian or mixed origin.” Census analysis of Indian groups in the Southwest did not lead bureau staff to speculate on the racial makeup of persons of Mexican origin in the region.^{Footnote 6} Indeed, the problem of classifying Mexicans came to the bureau in the 1920s as an unwelcome surprise, sprung on them by the sudden arrival of large numbers of Mexican immigrants.

The Origins of the 1930 Racial Category

Mexican immigration in the nineteenth century had been modest. Individuals casually crossed an unmarked border, but more permanent entry was rare; census birthplace data indicate fewer than five thousand immigrants from Mexico per year before 1900. The only large concentration of native-born Hispanics, in northern New Mexico, clung somewhat successfully to an argument that they were of Spanish rather than Mexican origin. By the 1920s, however, immigration from Mexico had risen dramatically, exceeding eighty thousand per year in the data available to Census Bureau officials.^{Footnote 7} While not equivalent to the flow of Europeans before restriction, the arrival of Mexicans in places where they had not previously settled incited a sharp nativist reaction. Given the rising influence of racial theory among intellectuals, the racialized conceptualization of European immigrants already embedded in immigration law, and the racist view of Mexicans in many locales in the Southwest, it was axiomatic that xenophobia toward Mexicans would take on a racial tone. It did. In 1921, former Congressman James Slayden concluded that “in Texas the word Mexican is used to indicate the race, not a citizen or subject of the country.” Persons born in Texas of Mexican ancestry “are ‘Mexicans’ just as all blacks are Negroes though they may have five generations of American ancestors.” Eugenicists such as Charles Goethe warned of the demographic and social calamity in the mass arrival of the “Amer-ind (American Indian) peon.” Congressional attempts to extend restrictionist quotas to Mexicans often used racial inferiority as a justification.^{Footnote 8}

Despite the standing position of the 1897 In re Rodriguez decision that the Treaty of Guadalupe Hidalgo made Mexicans eligible for citizenship (and therefore considered white for naturalization requirements), restrictionists sought their prohibition on racial grounds. Glenn E. Hoover, writing in Foreign Affairs in 1929, argued that Mexicans were largely Indian in background and their immigration thereby violated the clause in the Immigration Act of 1924 excluding the entry of immigrants racially ineligible for citizenship, then available only to “‘free white persons and to persons of African nativity or descent.’”^{Footnote 9}

Attention to the indigenous origins of Mexican immigrants appeared in both the English- and Spanish-language press. The New York Times applauded Hoover’s analysis, agreeing that Mexican immigration might well lead “to a new ‘race’ problem,” since “those who enter are largely Indian in blood, with only a veneer of Spanish culture.” La Prensa, the San Antonio paper with the greatest influence among the Spanish-reading population in the Southwest, judged nearly all persons of Mexican origin to be primarily Indian in ancestry (including New Mexicans). However, like most of the Spanish-language press, La Prensa argued that admixture yielded virtues rather than vices, publishing pieces by the Mexican intellectual José Vasconcelos, who celebrated the Mexican mestizo as a new and gifted race.^{Footnote 10} The noted anthropologist Manuel Gamio, a well-known scholar of Mexican immigration, had a less optimistic take on the Indian side of the equation, pointing to educational, cultural, and economic deficiencies. But he and other post-revolutionary intellectuals saw the integration of the Mexican population in mestizaje as essential in forjando patria, the construction of a nation founded on a unique and worthy race. Belief in the superiority of the “blended bronze and iron” of the mestizo, an inversion of the idea that hybrid races were degenerate, was firmly established in Mexico among intellectuals and celebrated by the state. Mexican American civil rights activists regularly voiced the same conviction.^{Footnote 11}

In the bureau, intense concern with recent immigration can be seen in the generational boundaries placed on the 1930 Mexican-race category: “In order to obtain separate figures for this racial group, it has been decided that all persons born in Mexico, or having parents born in Mexico, who are not definitely white, Negro, Indian, Chinese, or Japanese, should be returned as Mexican.” Instructions to enumerators noted the mixed racial background of “Mexican laborers,” emphasizing that those “not definitely” of another race be entered as Mexican.^{Footnote 12} The limitation to two generations recalled attempts to define European “races” in the early twentieth century, in which assimilation and loss of ethnicity was conceivable.^{Footnote 13} Still, a racial function appears manifest: census questions on birthplace and parental birthplace would have identified all those of recent Mexican immigrant origin. Officials knew most Mexican immigrants were mestizo in origin, but others appeared to be largely European: they wanted a measure of race within the population of Mexicans arriving in the United States.

The Bureau and the Category

The origins of the Mexican racial designation appear to lie within the Census Bureau itself, although scholars have argued otherwise. In an influential piece, Jennifer L. Hochschild and Brenna Marea Powell conclude that “political pressure” from Congress, abetted by the “deepening depression,” forced the Census Bureau to create the category. Since the idea emerged and the questionnaire was developed before the economic crisis, the latter could not have had an effect; moreover, Hochschild and Powell provide no direct evidence of Congress’s influence. Mark Reisler’s more careful account, based in governmental archival documents, also points to congressional pressure; he reports that the virulently restrictionist and racist Chair of the House Committee on Immigration and Naturalization (and a member of the House Committee on the Census), Albert Johnson, “convinced Secretary of Commerce, Robert P. Lamont, to have his census director classify Mexicans in a separate racial category in the 1930 census.” However, the correspondence cited refers to tabulation and publication of conventional data, not to the new racial category, nor is there documentary evidence of Johnson’s role. Margo Anderson’s authoritative history of the census states that the 1929 appropriation bill from the Congress “for the first time . . . did not specify in minute detail the questions to be asked.” Indeed Census Director Steuart, Lamont, and the Acting Secretary of Commerce, E. F. Morgan, successfully resisted the “elaborate plan” for racial schedules presented by Johnson and the eugenicists Alexander Graham Bell and Harry H. Laughlin. Commerce officials also opposed the restraints on Mexican immigration that eugenicists, racists, and other nativists demanded.^{Footnote 14}

The bureau’s role appeared first in February 1926, when Hill’s office presented a topic for discussion at a weekly brainstorming meeting held by Secretary of Commerce Herbert Hoover, under whom the bureau operated. The unsigned memorandum posed the following question: “The immigration law has greatly stimulated entrance into the U.S. of Canadians and Mexicans, to whom it does not apply. Is there any possibility of being able to shut off the Mexican whom many sociologists believe will be far more objectionable as a national problem than most of the Europeans who are being excluded?” Attached to the memorandum was a calculation of the difference between the number of Mexicans returning to Mexico and those staying, with the note that the sojourners were “Mostly Mexican race.”^{Footnote 15}

The first official step toward a separate racial category can be found in materials prepared for the December 1928 meeting of the (Joint) Census Advisory Committee. This committee, made up of leading members of the American Economic Association and the American Statistical Association, provided counsel about the enumeration schedule and the publication of census data. A staff memorandum dated November 23, 1928, listed eleven of the “Most important questions for addition,” to be taken up with the Advisory Committee, including the “The color classification of Mexicans, especially of recent immigrants from Mexico.”^{Footnote 16} The extant records do not show what action the Advisory Committee took, nor do subsequent agendas mention the Mexican race category. Reflecting in 1934 on the bureau’s decision, Hill recalled that “Mexicans were classified separately from the white population because of the feeling that they were not strictly white.”^{Footnote 17} Unless convincing evidence of outside pressure is discovered, it appears that the bureau was responsible for the creation of the category.

It did of course meet with approval from restrictionists in Congress, who regularly highlighted the racial threat Mexicans posed. Replying in November 1929 to a letter from Johnson, Secretary Lamont (who had testified against restriction of Mexican immigration) assured him that

The 1930 census enumeration classified 1,431,473 persons as Mexican race. Nearly a fifth were native born and had native-born parents and should not have been so designated under the census rules. Area of residence strongly affected classification. In parts of Texas, the enumerator was highly likely to write “Mex” in the race column. In contrast, racial identification was low in northern New Mexico and southern Colorado, home to a large, politically effective population of “Hispanos” fervently opposed to being identified as Mexican. In Torrance County, New Mexico, enumerator Sosteno Trujillo classified the entire Tapia family as white. While Mr. Tapia was born in the United States to U.S.-born parents, his wife was born in Mexico, so she and their six daughters met the census criteria for Mexican race.^{Footnote 19}

The Ambiguity of Race

Few protested the 1930 census classification. The only sharp reaction came from New Mexico, where the new immigrants rarely ventured and where Hispanic voters objected to any connection to Mexicans.^{Footnote 20} El Nuevo Mexicano, the leading newspaper in the state, strongly criticized the Mexican race category, advising its readers that they should inform the enumerator that they were “Americans and nothing less than Americans, the same as and equal to any other element of our citizenry,” and not to be classified as “Mexicans.” The paper’s editorials were reprinted in other Spanish-language newspapers in the state. On March 14, 1930, El Defensor del Pueblo of Socorro, New Mexico, urged its readers to remember that “we are American citizens of Hispanic” background, not Mexicans.^{Footnote 21}

Little to no opposition emerged from regions with large Mexican immigrant populations, made up of persons unlikely to be citizens, still oriented toward Mexico, many of whom expected to return to their homeland. They may have been at ease with a sense of themselves as distinct from the yanquis they encountered north of the border. Those familiar with racial classifications in Mexico would not have been startled by the category; that country’s census inquiry on raza distinguished between Indians, whites, and mestizos. The racial perspective of the mass of Mexican immigrants to the United States is ill understood, but Mexican American and Mexican intellectuals, like Vasconcelos, celebrated a separate, mestizo race.^{Footnote 22} Indeed, Spanish-language papers in the United States published explicit support for the 1930 census category from the Mexican American attorney, diplomat, and civil rights activist, Alonso S. Perales. Born in Texas and educated in Washington, D.C., and the chief architect in the creation of LULAC, Perales was by the early 1920s vigorously protesting discrimination against persons of Mexican origin. Yet in testimony in congressional hearings on immigration on September 3, 1929, he voiced pride in both the indigenous and Spanish contributions to the “mestizo” Mexican people, who descended from “Hildalgo and Cuahutémoc.” He and other Mexican American witnesses did not argue that Mexicans were white. In his testimony, LULAC co-founder J. T. Canales contrasted Mexican labor to “white labor.”^{Footnote 23} In 1930, Spanish-language newspapers published Perales’s stock explanation of the census in a column the bureau conventionally sought in foreign-language papers to encourage census participation. Perales emphasized but did not criticize the “classificación separada,” noting that this was “the first time the census will give to those of Mexican origin a separate classification.”^{Footnote 24}

Perales’s lack of censure and the inclusion of his column in leading Spanish-language newspapers demonstrate the absence of immediate opposition to a separate racial classification outside New Mexico.^{Footnote 25} Perales was, like most LULAC leaders, linked closely by family and culture to Mexico. In the mid-1930s, he wrote that the use of three categories in San Antonio’s vital statistics (“white, Mexican and colored” [i.e., African American]) had not previously prompted complaint “because we persons of Mexican descent, regardless of citizenship, are very proud of our racial extraction and do not wish to convey the impression that we are ashamed to be called Mexicans.” He and other LULAC founders were at pains to defend themselves against charges by Mexican critics that they had abandoned their racial identity, as had M. Flores Villar, a journalist who was a Mexican citizen. Flores attacked the LULAC founders as “renegades,” who denigrated “the Mexican race. . . . [I]nstead of saying with pride ‘We are Mexicans by race’ [they] state (without seeing for a moment in the mirror their bronze color and their totally Indian aspect) . . . ‘We are Americans.’”^{Footnote 26}

Practical concerns, not a lack of belief in a distinct Mexican race, had led Perales and others toward the assertion that Mexicans were white and “Caucasian.” Texas had a large African American population, a history of slavery, and a legal system that mandated discriminatory treatment of nonwhites. In that state (and in other parts of the Southwest), there were de jure and de facto separate levels of citizenship for whites and for persons of color. Despite their racial pride, LULAC’s leaders saw that racial classification in the census and other official statistics could mean a loss of rights. Perales and others therefore made what Neil Foley has called a “Faustian” bargain, insisting on a white identity in official statistics when the alternatives endangered equal treatment. In 1934, Perales elicited an opinion from James T. Allred, Texas attorney general, stating that Mexicans included “persons of Caucasian descent or of mixed Caucasian and American Indian blood”; Allred concluded that “‘Mexicans’ must be considered ‘white citizens.’”^{Footnote 27}

Faustian perhaps, intentional surely, but not simply calculated. The claim of white identity rested as well on a bedrock of racism in the Mexican origin population that mirrored that among non-Hispanic whites in the United States. The Spanish term, la raza, is as ambiguous as its equivalent in English in the early twentieth century, meaning, at times, mutable ethnicity, and at other times, an innate, largely unchangeable nature. Las razas de color (African or Asian) were clearly seen as racially distinct by persons of Mexican origin in the United States. In 1928, Mauro Machado, an early associate of Perales, decried the “the cowardly way in which [Box Committee witnesses] try to make us possessors of negro blood.”^{Footnote 28} Hostility to African Americans and to Asians repeatedly erupted in local condemnations of any equating of Mexicans with these groups. In 1921, El Heraldo de México (Los Angeles) protested the seating of Mexicans in theaters with “la raza de color.” The New York newspaper Gráfico found it odd that American blacks saw Latin Americans as a colored race. In August 1925, the New York newspaper La Prensa noted the “indignant” reaction of Mexicans in Indiana Harbor, Indiana, to attempts to compel them to sit in areas of theaters given to “people of color.” In 1929, the Mexican community in the copper mining town of Miami, Arizona, objected to the “disrespectful and degrading” policy of a “Greek” theater owner who required that they sit with “people of color.” In the same year, the New York paper La Prensa reported the protest of children in Mexico against attempts in San Bernardino, California, to classify students of Mexican descent “with the black and oriental children” in that county. A similar segregation in the theaters in Lockhart, Texas, was protested in 1932, and one in Galveston, Texas, was successfully undone in 1935. In 1939, “La Unión Hispano Americana de Texas” objected to the placement of persons of Mexican origin “with negroes” in the hospital in Austin, Texas.^{Footnote 29}

Racially tinged views of African and Asian Americans guaranteed a quick reaction against any classification as “colored.” It led Machado and others to declare themselves “white,” and for registrars of vital statistics (many of Mexican origin) to place Mexicans in the white category when the only alternative was “colored.” Still, even these racially charged views among Mexicans pointed to the nefarious effects of discrimination: it was the practical penalty of a nonwhite classification that provided the impetus for a systematic and successful protest movement.

Civic Protest and the Elimination of the Mexican Racial Category

The Mexican” racial category did not appear in the 1940 census, but it had not died an easy death. While senior officials, including an aging Hill, clung to the category, it was not they who most effectively championed its continued use. New Dealers did bring to the bureau “an extraordinarily open, liberal view of the diverse ways of life,” yet it was New Deal staff and the experts who advised them who argued most strenuously to retain the Mexican category. During the Roosevelt administration, the bureau increasingly hired staff with Ph.D.s in statistics, sociology, or economics, among the first to identify themselves as “population scientists” or “demographers.” The “Class of 1940” sought to create the first scientific census.^{Footnote 30}

These experts favored a cohort component model to generate estimates of future population size and structure, an approach popularized in the late 1920s by Pascal Whelpton and Warren Thompson of the Scripps Foundation for Research on Population Problems. They proposed that subdivision of the population according to characteristics correlated with mortality and fertility—such as race and urban/rural status—would yield more accurate population projections. Mexicans and Mexican Americans had higher rates of fertility, infant mortality, and adult mortality than non-Hispanic whites and public health officials had by the 1920s begun to publicize the sharp differences between Mexicans and other groups in vital statistics. Well aware of this literature, demographers sought to identify them separately in order to improve the accuracy of projections. In 1936, Thompson explicitly called for separation of Mexicans and other groups from the catchall “colored” category in vital statistics, citing both socioeconomic and hereditary rationales.^{Footnote 31}

The initial lack of objection to the Mexican race category had led to its expanded use within the bureau. Census documents and staff publications employed it through the mid-1930s, as did such influential studies as President Hoover’s commissioned work, Recent Social Trends in the United States, 1929–1933.^{Footnote 32} But its most fateful application was in vital statistics. In its annual reports for 1930 through 1934, following the 1930 rule, the bureau’s Division of Vital Statistics stated that Mexicans “were not classified with the ‘white’ but tabulated with ‘other races,’” i.e., colored, though it was apparent that local registrars often did not follow this directive. Expressing concern as early as 1930, in 1932 Hill urged the bureau in his “Classification of Mexicans in Vital Statistics” to “instruct or educate the registrars” to enter Mexicans separately from whites.^{Footnote 33} In an Advisory Committee meeting in April 1934, Hill objected again to the inconsistency between vital statistics and the census, since the former did not effectively separate “Mexicans from the whites.” A rising figure in the bureau, Chief Statistician for Population Leon E. Truesdell, pointed to the problem in the field: in gathering vital statistics data, it was “not easy to segregate Mexicans as the Mexicans have a prejudice against returning themselves as other than white, and seventy-five percent of the local registrars in New Mexico and lower California are Mexicans who credit themselves with being white.” Census Director William Lane Austin “saw the same objection in the case of the population census, as in the same sections a large percentage of the enumerators and supervisors are Mexicans.”^{Footnote 34}

In 1935, in a major bureau reorganization, Halbert L. Dunn became chief statistician in the Division of Vital Statistics, part of the shift toward more highly trained and demographically oriented staff. Dunn, a biostatistician credited with the establishment of a national vital statistics system in the United States, became a central figure in the struggle to retain a separate classification. Dunn and other advocates were well aware of the distinct characteristics of vital statistics in the Mexican origin population. As his office was to state in 1947, “The mortality rates for white infants” were powerfully affected by “the presence, in the white population of [Southwestern] States, of a large number of Mexican agricultural workers in low income groups, among whom the rate of infant mortality is extremely high.” At times, this distortion led the Division to remove these states when calculating comparisons in mortality between jurisdictions. Separate categories were crucial to the accuracy of population models the new staff wished to build. Including persons of Mexican origin distorted the rates for “white” populations. The instruction manuals for coding vital statistics from Dunn’s office listed Mexicans as a distinct race.^{Footnote 35}

The attempt to classify Mexicans as nonwhite in vital statistics would ultimately spark broad protest against the Mexican race category in the census, but reaction sprang first from a different source, and in the most sensitive state, New Mexico. On February 18, 1935, Austin responded to protests sent to him on the 15th and 16th by New Mexican Republican Senator Bronson Cutting, who complained about the classification of farm operators as “Mexican” rather than white in the 1935 Census of Agriculture.^{Footnote 36} Revealing the link the bureau had forged between Mexicans and Indians, he advised the Senator that “there has been a good deal of discussion . . . concerning the proper classification by the Bureau of ‘Mexican’ population and ‘Indian’ population. There is quite a difference of opinion among statisticians and ‘race experts’ concerning the proper classification of the two ‘races’ mentioned, and it is my intention to bring these questions up for discussion and final settlement previous to our next census of population.”^{Footnote 37}

After renewed criticism, remarking that he was subject to “pressure from various learned individuals and organizations,” Austin promised Cutting that the agricultural census enumerators would not classify farm operators as Mexican unless they had been born in Mexico. He had directed all “supervisors and field representatives in New Mexico” to ensure that those of “Spanish” descent “be reported as ‘white.’”^{Footnote 38}

The next, and much more powerful reaction, came out of Texas as a result of the removal of Mexicans from the white category in vital statistics. Beginning with sheer indignation that Mexicans would be classified with African Americans, protests soon revealed, as Paul Schor argues, the driving, practical concerns in Texas that made emerging Mexican American civic organizations hostile to any designation other than white in a state where segregation of Mexican children in schools was a core issue. Shortly before the election of 1936, a set of complaints came to Austin’s direct supervisor, Secretary of Commerce Daniel C. Roper, a stalwart Democratic appointee. The epicenter of the public debate, examined closely by the historian Mario García, lay in the border city of El Paso, Texas.^{Footnote 39}

The controversy had its roots in the extraordinarily high rates of infant mortality in that city, which often placed El Paso in the unenviable position of first in the nation, a notoriety due entirely to high mortality among infants of Mexican origin. As the city’s health officers remarked, “The infant mortality rate among American families . . . is way below the national average. The numerous cases which occur among Mexican American infants of poorly educated families, living in unhealthy and squalid conditions, are the ones that increase our rate so rapidly.” In October 1936, City Registrar Alex K. Powell announced that his office would join four other Texas cities “in classifying Spanish-speaking residents as ‘colored,’” a category he claimed the Census Bureau had approved. El Paso’s Health Department chief, Dr. T. J. McCamant, stated that the State Registrar had permitted this classification and that L. P. Bishop, his counterpart in San Antonio, “had been authorized by H. L. Dunn, chief of statistics for the Department of Commerce in Washington to classify Mexicans as ‘colored.’” The city officers argued for the need for more accurate accounting of the sources of differential infant mortality so that it might be addressed, emphasizing the socioeconomic sources of the problem, rather than racial ones. García concludes that the main intent of the new category was to make El Paso’s mortality rate more acceptable by identifying Mexicans as the source of the city’s poor performance.^{Footnote 40}

The decision ignited a firestorm of protest in El Paso. Most important, it centered the attention of Mexican American leaders, especially in Texas LULAC chapters, on racial classification in official statistics. On October 8, 1936, Frank J. Galvan, the president general of LULAC, sent out circulars to all chapters reporting the attempt to classify “the members of the Spanish speaking race . . . as colored people.” He identified the “Census Bureau of the Department of Commerce” as the source of the grave affront, calling on each council to ask their congressmen to insist that “our classification” be “white.” Adolpho de la Garza, president of LULAC Council No. 5 in Mission, Texas, reported, “Councils all over the state of Texas, and some in New Mexico, protested to Washington over this error.”^{Footnote 41} Complaints about the El Paso decision filled the pages of the Spanish-language press.

A core feature in the agitation, reflecting a new phenomenon outside New Mexico, was the capacity of middle-class Mexican American organizations to persuade politicians to support their causes. The first to act, by his own account, was C. K. Quin, mayor of San Antonio. Quin claimed that he had immediately complained to both Texas senators about the classification of Mexicans among “other non-white races,” unjustly categorizing “some of our best citizens as ‘colored’ when they are not in fact ‘colored’ as that term is commonly used.” Cutting’s Democratic successor in the Senate, Dennis Chávez, enthusiastically joined in the El Paso dispute, despite its being “outside my official jurisdiction.” In the October 17 issue of La Prensa, the senator promised that he “would do all in my power to make sure [the category] is changed as soon as possible.”^{Footnote 42} Roper’s acting director, Ernest G. Draper, replied to criticism from Chávez, betraying his ignorance of census classifications, or a willful decision to ignore them:

On October 19, Secretary Roper himself responded to a similar protest from Senator Tom Connally of Texas, as he did to other politicians, making the same promise. He maintained again, and quite falsely, that the vital statistics “classification referred to is not in accordance with the established classifications of the Census Bureau in its report on population, agriculture, etc.”^{Footnote 43}

Texas politicians were especially sensitive to the political implications. Maury Maverick, a House member from the San Antonio area, had also sent Austin a telegram and a letter protesting the classification. Like Quin and a select few Texas politicians (including Vice President John Garner), Maverick had close ties with a Mexican American constituency upon which he depended for reelection.^{Footnote 44} Maverick described the “deadly insult” implied when “citizens in Texas of Spanish and Mexican extraction or descent are classified as ‘colored.’” In his view, Mexicans were like “Italians in New York City.” Maverick suggested officials provide a category “Other Whites—Mexican.” Austin reiterated the new bureau position in a letter to Maverick: “It was unfortunate, of course,” he wrote, “that an error was made by our Division of Vital Statistics in classifying Mexicans along with the ‘colored.’” He assured Maverick that such would “not occur again in any public information given out by the Bureau of Census. Mexicans will be classified as ‘white.’”^{Footnote 45}

Allies south of the border also reacted to implied linkages to African Americans. The El Paso newspaper El Continental reported that Masons in Ciudad Juárez had sent a protest to the Grand Lodge, denouncing the classification of Mexicans as a “colored race, that is to say, black,” and petitioning their Grand Master to lodge a protest with the Mexican Ambassador in Washington. Representatives to Mexico’s Chamber of Deputies, joined by the Mexican Chamber of Commerce, called for official protests that Mexicans would be classified as “negroes” and “an inferior race.” La Prensa reported that the Mexican ambassador had contacted the U.S. Department of State urging that Mexican elements in the U.S. population “be included in the white race in any official Government document.”^{Footnote 46}

The controversy prompted persons of Mexican origin who viewed Mexicans as a race distinct from whites to shift from that view. In letters to public officials in November 1936 that record his pride in his “racial extraction,” Perales judged the threat of being classified as colored sufficient to justify changing San Antonio’s three-part classification to two categories, “White, including Mexicans,” and colored. In order to “avoid confusion or incidents like that which has just occurred in El Paso,” there should be no “intermediate status of ‘mexicans.’” Such “is in accord with the Laws of Texas, which consider residents of Mexican origin to be of the white race.” Perales’s unhappy transition from one racial identity to another can be followed in his drafts of an article prepared in 1939. Not mentioning his own endorsement, he argues that the 1930 census classification “did us great damage . . . , achieving very evident[ly] what the intention was—that we would not be considered as ‘Whites,’” with grave social consequences.^{Footnote 47}

Dunn’s effort to get local registrars to classify Mexicans separately, in accord with the 1930 census category and essential to projection models, came to an end. Schor reports that, on October 15, 1936, Austin sent a sharp memorandum to Dunn announcing a strict policy:

In less than two weeks, wide protests and the mobilization of political power by middle-class Mexican American groups had resulted in a seeming capitulation by the bureau. Schor concludes that the success of the Mexican American protest in 1936 was both “complete” and “remarkable,” since theirs was “the only case in which a group obtained a modification of their classification.”^{Footnote 49}

It was, however, not yet complete. Bureau staff greatly regretted Austin’s concession to political rather than supposedly scientific standards; experts inside and outside the bureau remained convinced of the value of a separate classification. They were to mount a campaign to bring the category back for the 1940 census.

The Last Campaign

In 1937, as Patrick Lukens notes, Dunn complained to Assistant Secretary of State Wilbur J. Carr, that “carrying into effect the decision with respect to Mexicans means the virtual destruction of the census of vital statistics insofar as concerns their scientific use in determining certain facts in regard to health, length of life, birth rate and other important matters.” Dunn argued that there were “fundamental biological differences between the average American and the average Mexican in the way in which they react to disease.” He cited a set of federal agencies that classified Mexicans separately, adding that insurance companies did the same. According to Lukens, Dunn thought a designation engineered by State Department staff of “White,” “White—Mexico,” with similar white categories for other Latin Americans, would function.^{Footnote 50}

It was planning for the 1940 population census, however, that prompted a coordinated campaign. A “General Memorandum on Changes Under Consideration for the 1940 census of Population,” dated February 6, 1939, revealed the staff view. The memorandum lamented the imminent removal by “accidental circumstance” (i.e., political interference) of a classification that had “produced statistics of considerable value, since the Mexicans form a distinct social and economic class in those areas where they are numerous.”^{Footnote 51} In June 1939, in the “Recommendations of the Central Statistical Board” for the 1940 Population Schedule, the first of eighteen recommendations from the staff proposed a solution close to the one that Dunn had found reasonable: “1. Mexicans. There is general agreement that the information regarding Mexicans obtained in the 1930 census was of great value. A desire for similar information from the 1940 census is wide-spread. It is urged, therefore, that steps be taken in some way to meet this need, perhaps by subheads under the category ‘white’ in column 13, named ‘white except Mexican’ and ‘Mexican.’”^{Footnote 52}

In seeking to reestablish the category, the bureau faced a new and formidable opponent. Austin’s 1936 dictum had used intriguing language, noting that racial classification was “a very delicate matter to the United States Government and our classification must always be in accordance with the policy of the Federal Government.” In 1939, Under Secretary of State Noble confirmed what the Spanish-language press had reported in the controversy over vital statistics: the Mexican government had complained to the Department of State.^{Footnote 53}

Opposed in the 1920s to quotas on Mexican immigrants, and in the 1930s to any other action that imperiled good relations with a neighbor still teetering on its revolutionary axis, the State Department had become wary of any potential insult. Lukens’s analysis of the little-known Andrade court case of 1935, which threatened to make Mexican immigrants ineligible for naturalization based on their Indian ancestry, details the State Department’s efforts to thwart such racial classification. Privately conceding that most Mexicans were Indian, the department’s preemptive legal strategy was to set a wide precedent by pressuring all federal government offices to classify Mexicans as white.^{Footnote 54}

Census staff were aware of the State Department position. Truesdell and Austin sought the support of the Advisory Committee to win over that department: “1. Mexicans. Dr. Truesdell expressed the opinion, with which the Director agreed, that the Bureau’s policy with regard to the Mexican classification must be guided by the attitude of the State Department, although he felt it desirable to have an expression from the Committee as to whether it was believed the retention of the 1930 classification was urgently desirable, or slightly desirable, or undesirable. . . . After further discussion, Dr. Thorp moved that the 1930 classification be retained.”^{Footnote 55}

Thorp’s view was not easily ignored, since it represented that of the Commerce Department itself. Austin asked George Wythe of that department to facilitate an appeal to the State Department for the 1930 classification, admitting but regretting that he had “issued an order that all Mexican statistics were to be classified as white.” In September 1939, Under Secretary Noble appealed directly to Secretary of State Cordell Hull: “The Bureau of the Census has requested me to enlist your cooperation in working out a racial classification of the population data relating to Mexicans . . . which will satisfy the needs of the various Government agencies and other users of the population data without giving offense to the Mexican Government.”^{Footnote 56}

It was “natural that protests should have arisen from Mexican groups against such a classification since the word ‘Colored’ was misinterpreted to signify Negro.” Nonetheless, Austin’s decision “to classify all Mexicans as ‘White’” was “hasty and unfortunate and . . . the policy of classifying Mexicans as ‘White’ should be reversed for the . . . 16th decennial census, which will be taken in April of 1940.” He justified separate classification on the far higher birth and death rates of Mexicans living in the same areas as whites; public health agencies would be “handicapped if they cannot have separate data for Mexicans.” Moreover, “From sociological, cultural and economic standpoints, . . . the Mexicans are different from the other racial groups. The solution of many labor problems throughout the Southwest thus depends on a knowledge of their numbers and geographic distribution. . . . The Public Health Service, the Department of Labor, and the Social Security Board are particularly concerned, and, to a lesser degree, the War Department and the Department of Agriculture.”

Noble maintained that the State Department itself, as well as the Mexican government, might find reason to want to know about “the Mexicans residing in this country.” He claimed that visiting staff from Mexico’s census bureau, including Emilio Alanís Patiño, Mexico’s director general of statistics, had a positive view of the category.^{Footnote 57} Noble concluded with a proposal that hardly solved the domestic and international problems of a racial identification, since it extended its range, as a racial measure ought, beyond the 1930 limitation of two generations. Moreover, it discarded the solution Carr and Dunn brought forward to classify Mexicans as a category within whites. Instructions to enumerators would be: “Mexicans. Persons of Mexican birth or parentage who are of unmixed White blood are to be reported as ‘White.’ It will be found, however, that many persons of Mexican origin are of a racial mixture usually well recognized and known as ‘Mexicans’ or ‘Spanish Americans’ in the localities where they are found. Such persons, including both those born in Mexico and those whose parents or earlier ancestors were born in Mexico, are to be returned as ‘Mexican.’”

The Department of State instructed Ambassador Josephus Daniels to ask the Mexican foreign office for its view. Daniels simply copied Noble’s appeal to Hull, which described the 1930 category, the 1936 protests against it, and the desire of the Census Bureau to reinstate it, primarily for its utility in vital statistics. Mexican officials then asked their leading demographers for advice.^{Footnote 58} Noble had been correct about their view. Manuel Gamio, directing the “Departamento Demográfico de la Secretaria de Gobernación,” approved of a separate category. A firm believer in the distinct status of the mestizo, he thought it “not only to the interests of American statistics but also to the majority of Mexican workers resident in the United States that it be indicated which of them are white and which mestizo.” He opposed placing Mexicans in a category with African Americans, but approved of distinguishing between Mexicans of “European origin” who had a “higher standard of life and cultural index” and those of “Indo-European origin,” with lower levels of both. Alanís Patiño agreed: “I am of the opinion that the classification that the [U.S. Census Bureau] intends to make in relation to the Mexicans resident in the country ought to be accepted.” Patiño (whose views likely came too late to have any influence) thought that even those Mexicans classified as white were usually mestizo. Prompted perhaps by a renewed request from the Embassy, the Mexican foreign office sent a memorandum to Daniels, quoting Gamio’s approving opinion word for word, without further elaboration.^{Footnote 59}

Whether Daniels had this view in hand or not, it was to no avail. On November 16, 1939, Noble wrote Sumner Welles, under secretary of state, acknowledging Welles’s letter of November 9:

On November 22, the Mexican ambassador to the United States, Castillo Najera, informed his government that there would be no separate classification. No record of the official position of the Mexican government has yet been produced, but the final decision fit well their earlier stance in the Andrade case.^{Footnote 61}

The Penultimate Step: The Hispanic Identifier

In February 1940, Spanish-language newspapers printed copies of a letter that Welles had written to E. D. Salinas, the general president of LULAC. Welles assured Salinas that “there would be no separate classification of Mexicans in the 1940 Census.” Secretary Hopkins also denied the rumor that Mexicans were not to be classified as white, stating that Mexican American leaders “who had expressed this fear were completely misinformed.”^{Footnote 62}

The decision was not well received by the experts who advised the bureau. Advisory Committee member Dr. Murray R. Benedict, professor of agricultural economics at Berkeley, found it inexplicable, given the prominent role of Mexicans in farm labor. Truesdell suggested to the members that sample line inquiries on mother tongue and country of birth offered a partial solution, and the bureau continued to seek other ways to identify Mexican Americans as a distinct population group.^{Footnote 63} In 1950, it inaugurated an analysis of persons of Spanish surname in five states in the Southwest, using the regular census schedule, an assessment expanded in 1960 and 1970. In 1970, the census introduced a new question in the 5 percent sample, asking a person’s “origin or descent,” listing the categories Mexican, Puerto Rican, Cuban, Central or South American, “Other Spanish,” or none of these. This self-identification as Hispanic (that term first appearing in 1980) has now become a regular part of the full enumeration schedule, next to but separate from the race variable.

The new Hispanic identifier was made possible, indeed mandated, not by bureau ambitions but by pressure from the Mexican American community. Federal legislation and policy attentive to minorities changed the ground rules. As Harvey Choldin demonstrates, Mexican American leaders, like those of other minority groups, became aware in the 1960s that remedying discriminatory conditions and gaining access to federal resources depended on census counts. Disappointed to find that vital statistics data did not list persons of Mexican origin separately despite patent mortality differentials, these leaders called for a new category, some arguing that it be racial. Mexican American organizations pressured the bureau for separate identification “in order to take advantage of opportunities resulting from the new federal legislation.”^{Footnote 64}

The bureau resisted this political interference, objecting especially to self-reported data they thought unscientific, but to no avail. Even though the full census forms had already been printed, the Nixon White House insisted on compliance with Mexican American leaders’ demands and the bureau placed the identification question on the form for the 5 percent sample. The item devised was placed separately from the race variable, but like that variable, it was not bound by generational limits. It suggested a permanent condition.^{Footnote 65} Continued political pressure led to the appointment to special census advisory committees of leaders of Hispanic organizations with no social science or statistical expertise (a practice evident for other interest groups). The eventual result was the placement of the Hispanic identifier on the main enumeration form, and to a variety of procedures in government and census statistics designed to facilitate the count of Hispanics.^{Footnote 66}