1. This concern is evident in the most influential models of state formation, including Charles Tilly's notion that the state is a form of protection racket. See Tilly, Charles, “War Making and State Making as Organized Crime,” in Bringing the State Back In, eds. Evans, Peter, Rueschemeyer, Dietrich, and Skocpol, Theda (Cambridge: Cambridge University Press, 1985), 169–187 CrossRefGoogle Scholar. Most scholars do not engage the question of the emergence of violence expertise directly (e.g., Olson, Mancur, “Dictatorship, Democracy, and Development,” The American Political Science Review 87, no. 3 (1993): 567–576 CrossRefGoogle Scholar). Indeed, even Tilly's more recent synthetic work deals with the ways in which types of violence interact with regimes rather than the emergence of expertise per se. See Tilly, Charles, The Politics of Collective Violence (Cambridge: Cambridge University Press, 2003)CrossRefGoogle Scholar. Important exceptions in the state formation literature include Gambetta, Diego, The Sicilian Mafia: The Business of Private Protection (Cambridge, MA: Harvard University Press, 1993)Google Scholar; Ikegami, Eiko, The Taming of the Samurai: Honorific Individualism and the Making of Modern Japan (Cambridge, MA: Harvard University Press, 1995)Google Scholar; Varese, Federico, The Russian Mafia: Private Protection in a New Market Economy (Oxford: Oxford University Press, 2001)CrossRefGoogle Scholar; Volkov, Vadim, Violent Entrepreneurs: The Use of Force in the Making of Russian Capitalism (Ithaca, NY: Cornell University Press, 2002)Google Scholar.

2. See Frymer, Paul, “Building an American Empire: Territorial Expansion in the Antebellum Era,” University of California-Irvine Law Review 1 (2011): 913–954 Google Scholar; Stefan Heumann, “The Tutelary Empire: State-and Nation-building in the 19th Century United States” (PhD diss., University of Pennsylvania, 2009); Thomas Ogorzalek, “Filibuster Vigilantly: Private Actors, the American State, and Territorial Expansion” (Unpublished Manuscript, Columbia University, 2008). On the role of the U.S. military as a source for bureaucratic and infrastructural capacity, see Wilson, Mark, The Business of Civil War: Military Mobilization and the State, 1861–1865 (Baltimore: Johns Hopkins University Press, 2006)Google Scholar; Adler, William D., “State Capacity and Bureaucratic Autonomy in the Early United States: The Case of the Army Corps of Topographical Engineers,” Studies in American Political Development 26, no. 2 (2012): 107–124 CrossRefGoogle Scholar; Adler, William D. and Polsky, Andrew J., “Building the New American Nation: Economic Development, Public Goods, and the Early U.S. Army,” Political Science Quarterly 125, no. 1 (2010): 87–110 CrossRefGoogle Scholar.

3. On the revisionist conception of early American state power, see Moore, Colin D., “State Building through Partnership: Delegation, Public-Private Partnerships, and the Political Development of American Imperialism, 1898–1916,” Studies in American Political Development 25, no. 1 (2011): 27–55 CrossRefGoogle Scholar; Baldwin, Peter, “Beyond Weak and Strong: Rethinking the State in Comparative Policy History,” The Journal of Policy History 17, no. 1 (2005): 12–33 CrossRefGoogle Scholar; Novak, William J., “The Myth of the ‘Weak’ American State,” The American Historical Review 113, no. 3 (2008): 752–772 CrossRefGoogle Scholar; King, Desmond and Lieberman, Robert C., “Ironies of State Building: A Comparative Perspective on the American State,” World Politics 61, no. 3 (2009): 547–588 CrossRefGoogle Scholar; Novak, William J., “The American Law of Association: The Legal–Political Construction of Civil Society,” Studies in American Political Development 15, no. 2 (2001): 163–188 CrossRefGoogle Scholar; Balogh, Brian, A Government Out of Sight: The Mystery of National Authority in Nineteenth-Century America (Cambridge: Cambridge University Press, 2009)CrossRefGoogle Scholar.

4. The notion of “violence expert” is similar to but distinct from that of “violence specialist” found in Tilly, Charles, The Politics of Collective Violence (Cambridge: Cambridge University Press, 2003), 34–41 CrossRefGoogle Scholar. Whereas Tilly essentially defines the violence specialist as a “political actor” who can “control means of inflicting damage on persons and objects” (p. 35), the violence expert is any actor (market or otherwise) who possesses expertise in violence. The distinction is important because we shouldn't presuppose that skill in violence is necessarily politically salient, a point that is somewhat confusing in Tilly's formulation. For example, he points out that “boxers, gladiators, bullfighters, and rugby players” all count as specialists “in doing damage” who “work outside of government”; yet, he also claims that “the category of political entrepreneur … overlaps with the category of violence specialist” (pp. 35–36). Maintaining the potentially non-political character of violence expertise, however, is crucial to problematizing the relationship between skill formation and state and market formation.

5. The best (though dated) bibliography of this vast literature is Adams, Ramon F., Six-Guns and Saddle Leather (Norman, OK: University of Oklahoma Press, 1969)Google Scholar. Other general works consulted in preparing this article include Brown, Richard Maxwell, No Duty to Retreat: Violence and Values in American History and Society (Oxford: Oxford University Press, 1992)Google Scholar; Brown, Richard Maxwell, Strain of Violence: Historical Studies of American Violence and Vigilantism (Oxford: Oxford University Press, 1975)Google Scholar; Rosa, Joseph G., The Gunfighter: Man or Myth? (Norman, OK: University of Oklahoma Press, 1969)Google Scholar; Metz, Leon C., The Encyclopedia of Lawmen, Outlaws, and Gunfighters (New York: Facts on File, 2003)Google Scholar; O'Neal, Bill, Encyclopedia of Western Gunfighters (Norman, OK: University of Oklahoma Press, 1979)Google Scholar; Sonnischen, C. L., I'll Die Before I'll Run: The Story of the Great Feuds of Texas (New York: Harper & Bros., 1951)Google Scholar; DeArment, Robert K., Deadly Dozen: Twelve Forgotten Gunfighters of the Old West (Norman, OK: University of Oklahoma Press, 2003)Google Scholar; DeArment, Robert K., Deadly Dozen: Forgotten Gunfighters of the Old West, Volume 2 (Norman, OK: University of Oklahoma Press, 2007)Google Scholar; DeArment, Robert K., Deadly Dozen: Forgotten Gunfighters of the Old West, Volume 3 (Norman, OK: University of Oklahoma Press, 2010)Google Scholar; DeArment, Robert K., Ballots and Bullets: The Bloody County Seat Wars of Kansas (Norman, OK: University of Oklahoma Press, 2006)Google Scholar; DeArment, Robert K., Assault on the Deadwood Stage: Road Agents and Shotgun Messengers (Norman, OK: University of Oklahoma Press, 2011)Google Scholar.

6. For a scholarly account of gunfighters as social bandits, see White, Richard, “Outlaw Gangs of the Middle Border: American Social Bandits,” The Western Historical Quarterly 12, no. 4 (1981): 387–408 CrossRefGoogle Scholar; as reckless deinstitutionalized men, see Courtwright, David T., Violent Land: Single Men and Social Disorder from the Frontier to the Inner City (Cambridge, MA: Harvard University Press, 1998)Google Scholar; as carriers of the honor code of the West, see Herman, Daniel J., Hell on the Range: A Story of Honor, Conscience, and the American West (New Haven, CT: Yale University Press, 2010)Google Scholar.

7. Brown, Richard Maxwell, “Western Violence: Structure, Values, Myth,” The Western Historical Quarterly 24, no. 1 (1993): 5–20 CrossRefGoogle Scholar; Brown, No Duty to Retreat, 43–46.

8. Heumann, “The Tutelary Empire”; Frymer, “Building an American Empire.”

9. I thank Anne Holthoefer for this formulation.

10. Katznelson, Ira, “Flexible Capacity: The Military and Early American Statebuilding,” in Shaped by War and Trade: International Influences on American Political Development, eds. Katznelson, Ira and Shefter, Martin (Princeton, NJ: Princeton University Press, 2002), 82–110 Google Scholar. Also, see Frymer, “Building an American Empire” and Ogorzalek, “Filibuster Vigilantly.”

11. Frydl, Kathleen J., “Kidnapping and State Development in the United States,” Studies in American Political Development 20, no. 1 (2006): 18–44 CrossRefGoogle Scholar.

12. See Rao, Gautham, “The Federal Posse Comitatus Doctrine: Slavery, Compulsion, and Statecraft in Mid-Nineteenth-Century America,” Law and History Review 26, no. 1 (2008): 1–56 CrossRefGoogle Scholar; Kaufman, Jason, “‘Americans and Their Guns’: Civilian Military Organizations and the Destabilization of American National Security,” Studies in American Political Development 15, no. 1 (2001): 88–102 CrossRefGoogle Scholar; Isaac, Larry, “To Counter ‘the Very Devil’ And More: The Making of Independent Capitalist Militia in the Gilded Age,” American Journal of Sociology 108, no. 2 (2002): 353–405 CrossRefGoogle Scholar; Ilia Murtazashvili, “The Political Economy of Claim Clubs: Squatters, Presumptive Rights, and the Origins of Legal Title on the American Frontier” (PhD diss., University of Wisconsin, 2009).

13. Parrillo, Nicholas, “The De-Privatization of American Warfare: How the U.S. Government Used, Regulated, and Ultimately Abandoned Privateering in the Nineteenth Century,” Yale Journal of Law & the Humanities 19 (2007): 1–96 Google Scholar.

14. Thomas Ogorzalek has made a somewhat similar argument in relation to filibusters. See Ogorzalek, “Filibuster Vigilantly.”

15. See Prassel, Frank Richard, The Western Peace Officer: A Legacy of Law and Order (Norman, OK: University of Oklahoma Press, 1972), 44–73 Google Scholar, 94–101; Ball, Larry D., Desert Lawmen: The High Sheriffs of New Mexico and Arizona Territories, 1846–1912 (Albuquerque, NM: University of New Mexico Press, 1996), 38–50 Google Scholar.

16. Calhoun points out that, in unorganized territories, federal marshals were the key law enforcement officials, although even in territories most legislatures allowed local counties to incorporate and create their own shrievalties. Calhoun, Frederick S., The Lawmen: United States Marshals and Their Deputies, 1789–1989 (Washington, DC: Smithsonian Institution Press, 1990)Google Scholar, 150. Moreover, sheriffs in organized counties would often take responsibility for patrolling unorganized areas; e.g., Ellis, Mark R., Law and Order in Buffalo Bill's Country: Legal Culture and Community on the Great Plains, 1867–1910 (Lincoln, NE: University of Nebraska Press, 2007)Google Scholar, 21,71.

17. Walter, John, The Guns that Won the West: Firearms on the American Frontier, 1848–1898 (London: Greenhill Books, 2006), 90–112 Google Scholar.

18. Abbott, Edith, “The Civil War and the Crime Wave of 1865–70,” Social Service Review 1, no. 2 (June 1927): 212–234 CrossRefGoogle Scholar; Courtwright, Violent Land, 45–46.

19. Collins, Randall, Violence: A Micro-sociological Theory (Princeton, NJ: Princeton University Press, 2009), 399–409 Google Scholar.

20. Faust, Drew Gilpin, This Republic of Suffering: Death and the American Civil War (New York: Knopf, 2008), 3–60 Google Scholar.

21. See Fellman, Michael, Inside War: The Guerrilla Conflict in Missouri During the American Civil War (Oxford: Oxford University Press, 1990)CrossRefGoogle Scholar.

22. A corollary to this point is that one can quickly gain a reputation for skill in violence by challenging and defeating an actor whose reputation is already well-known.

23. McGrath, Roger D., Gunfighters, Highwaymen, and Vigilantes: Violence on the Frontier (Berkeley, CA: University of California Press, 1987), 76–85 Google Scholar.

24. Brown, No Duty to Retreat, 3–20.

25. Stiles, T. J, Jesse James: Last Rebel of the Civil War (New York: Knopf, 2002), 210–211 Google Scholar, 247–248; Brown, No Duty to Retreat, 84–85; Rosa, Joseph G., They Called Him Wild Bill: The Life and Adventures of James Butler Hickok (Norman, OK: University of Oklahoma Press, 1974), 105–110 Google Scholar.

26. Courtwright, Violent Land.

27. White, Richard, Railroaded: The Transcontinentals and the Making of Modern America (New York: W.W. Norton, 2011), 466–482 Google Scholar.

28. Moore, Jacqueline M., Cow Boys and Cattle Men: Class and Masculinities on the Texas Frontier, 1865–1900 (New York: NYU Press, 2009), 37–38 CrossRefGoogle Scholar.

29. Ibid., 92–101.

30. U.S. Congress. House. Depredations on the Frontiers of Texas. 42nd Cong., 3rd sess., 1872. Ex.Doc.39, 3,8,18. To put the threat into perspective, the same commission reported that approximately 300,000 cattle were grazing in 1872 in 11 of the Texas counties bordering Mexico.

31. Cool, Paul, Salt Warriors: Insurgency on the Rio Grande (College Station, TX: Texas A&M University Press, 2008), 96–98 Google Scholar; Hine, Robert V. and Faragher, John Mack, The American West: A New Interpretive History (New Haven, CT: Yale University Press, 2000), 308–310 Google Scholar.

32. White, Richard, It's Your Misfortune and None of My Own: A New History of the American West (Norman, OK: University of Oklahoma Press, 1993), 344–346 Google Scholar; Graybill, Andrew R., “Rural Police and the Defense of the Cattleman's Empire in Texas and Alberta, 1875–1900,” Agricultural History 79, no. 3 (2005): 253–280 CrossRefGoogle Scholar.

33. Hagan, William T., Charles Goodnight: Father of the Texas Panhandle (Norman, OK: University of Oklahoma Press, 2012), 72–78 Google Scholar.

34. Smallwood, James M., The Feud That Wasn't: The Taylor Ring, Bill Sutton, John Wesley Hardin, and Violence in Texas (College Station, TX: Texas A & M University Press, 2008), 101–135 Google Scholar; Smallwood, James M., Crouch, Barry A., and Peacock, Larry, Murder and Mayhem: The War of Reconstruction in Texas (College Station, TX: Texas A & M University Press, 2003), 73–82 Google Scholar, 92–129; Pickering, David and Falls, Judy, Brush Men & Vigilantes: Civil War Dissent in Texas (College Station, TX: Texas A & M University Press, 2000), 129–136 Google Scholar.

35. Crouch, Barry A. and Brice, Donaly E., The Governor's Hounds: The Texas State Police, 1870–1873 (Austin, TX: University of Texas Press, 2011), 176–177 Google Scholar.

36. See Burton, Arthur T., Black, Red, and Deadly: Black and Indian Gunfighters of the Indian Territory, 1870–1907 (Austin, TX: Eakin Press, 1991)Google Scholar. This article, unfortunately, cannot do justice to the manifold ways gunfighting—traditionally seen as a paragon of white masculinity—was shaped by racial conflict and alliance. In particular, much more work is needed on unpacking how the marketization of violence affected the relationship between the use of force and racial categories in the Southwest, as well as the ways in which such categories were used on behalf of large-scale economic incorporation. See Benton-Cohen, Katherine, Borderline Americans: Racial Division and Labor War in the Arizona Borderlands (Cambridge, MA: Harvard University Press, 2009), 48–79 CrossRefGoogle Scholar.

37. Pinkerton, William A., “Highwaymen of the Railroad,” The North American Review 157, no. 444 (November 1893): 530–540 Google Scholar.

38. White, “Outlaw Gangs of the Middle Border”; Stiles, Jesse James.

39. Wooster, Robert, The American Military Frontiers: The United States Army in the West, 1783–1900 (Albuquerque, NM: University of New Mexico Press, 2009)Google Scholar.

40. Graybill, “Rural Police and the Defense of the Cattleman's Empire in Texas and Alberta, 1875–1900.”

41. Graybill, Andrew, “Texas Rangers, Canadian Mounties, and the Policing of the Transnational Industrial Frontier, 1885–1910,” The Western Historical Quarterly 35, no. 2 (2004): 167–191 CrossRefGoogle Scholar.

42. Continual complaints about the military's effectiveness managing the rustling threat (which was neither fully a native nor a Mexican problem) led to a number of congressional inquiries. E.g., U.S. House, 42nd Cong., 3rd sess., Ex.Doc.39; U.S. Congress. House. El Paso Troubles in Texas. 45th Cong., 2nd sess., 1878. Ex.Doc.93; U.S. Congress. House. Lawlessness in Parts of Arizona. 47th Cong., 1st sess., 1882, Ex.Doc.58.

43. This was true, for example, in the Lincoln County War of New Mexico (the event in which Billy the Kid cut his teeth), a war between commercial factions for control over local trade. When the event reached a dangerous standoff in Lincoln in the summer of 1878, the appearance of a unit commanded by Colonel Nathan Dudley ostensibly to “keep the peace” actually helped precipitate the shootout at Alexander McSween's house by dramatically shifting the balance of power, leading to a collapse in local order and a rapid increase in rustling and raiding by the independent groups of gunfighters brought in to work for the factions. Utley, Robert M., High Noon in Lincoln: Violence on the Western Frontier (Albuquerque, NM: University of New Mexico Press, 1990), 110–117 Google Scholar.

44. Ball, Desert Lawmen, 235–237.

45. Clayton Laurie has argued that, prior to 1878, the military was a fairly active and willing participant in local law enforcement on the frontier, acting as a kind of para-posse for local forces. Laurie, Clayton D., “Filling the Breach: Military Aid to the Civil Power in the Trans-Mississippi West,” The Western Historical Quarterly 25, no. 2 (1994): 149–162 CrossRefGoogle Scholar. At the same time, General Sherman and other central elites, concerned with the possibility of lawsuits for wrongful arrest and preoccupied with ending Reconstruction, continually expressed their unease with this kind of activity. The military's growing professional ethos also meant that, on the whole, they did not necessarily regret institutional occlusion from local posse service.

46. The U.S. Army, for example, had been very active in suppressing the 1877 national strike, creating tense relationships with both industrial laborers and their antimonopoly homesteader allies in the West, who suffered from the price-gouging of the railways and the complicity and occasional corruption of military officers. White, Railroaded, 291–292.

47. “Dangerous Dan” Tucker, for example, marshal of Silver City, New Mexico, organized a posse of local toughs (including the notorious John Kinney) to “patrol” the border during the conflicts of 1877 and 1878 on the Rio Grande. Ironically, this adventure also led some of the participants to engage in rustling of their own. See U.S. House, 45th Cong., 2nd sess., Ex.Doc.93, 4, 17, 64, 79.

48. Ball, Desert Lawmen, 43, 209–211.

49. This was often the source of intense consternation by large corporations, and was frequently invoked to justify the hiring of private guards. See U.S. Congress. 52nd Cong.., 2nd sess. Congressional Record, 1280 (1893), 84, 129–130, 161–162, 242.

50. See Hagan, Charles Goodnight, 74, 78.

51. Horan, James David, The Pinkertons: The Detective Dynasty that Made History (New York: Crown Publishers, 1968), 189–202 Google Scholar, 360–394; Morn, Franklin, The Eye that Never Sleeps: A History of the Pinkerton Detective Agency (Bloomington, IN: University of Indiana Press, 1982), 78–79 Google Scholar, 91–109.

52. Prassel, The Western Peace Officer, 132–133, 138–142.

53. See U.S. Congress, 52nd Cong., 2nd sess., Congressional Record, 1280 (1893), 136–138.

54. Horan, The Pinkertons, 194–202.

55. Charlie Siringo, for example, was hired by the firm after presenting references from Pat Garrett, whom he had known in his cattle driving days. Lamar, Howard R., Charlie Siringo's West: An Interpretive Biography (Albuquerque, NM: University of New Mexico Press, 2005), 135–136 Google Scholar.

56. Calhoun, The Lawmen, 25–119.

57. Prassel, The Western Peace Officer, 223.

58. Rao, “The Federal Posse Comitatus Doctrine,” 46–53; Traub, Stuart H., “Rewards, Bounty Hunting, and Criminal Justice in the West: 1865–1900,” The Western Historical Quarterly 19, no. 3 (1988): 287–301 CrossRefGoogle Scholar. The U.S. Marshalcy was considered an important patronage role not only for the Marshal himself, but also deputies. See Ball, Larry D., The United States Marshals of New Mexico and Arizona Territories, 1846–1912 (Albuquerque, NM: University of New Mexico Press, 1978), 6–17 Google Scholar.

59. Ball, Desert Lawmen, 34–35.

60. Dykstra, Robert R., The Cattle Towns (New York: Knopf, 1968), 1–73 Google Scholar; Hine and Faragher, The American West, 307–308.

61. I wish to thank an anonymous reviewer for suggesting this point.

62. Einstadter, Werner J., “Crime News In The Old West,” in Media, Process, and the Social Construction of Crime: Studies in Newsmaking Criminology, ed. Barak, Gregg (London: Routledge, 1995), 49–68 Google Scholar.

63. For instance, I located 108 articles published in the New York Times and Chicago Tribune between 1865 and 1885 with the keywords “Dodge City,” “Abilene,” or “Wichita” in the title. Proquest Historical Newspapers Database, accessed July 4, 2013. These articles frequently dealt with violent events, though they occasionally covered economic developments in the cattle industry as well.

64. “A Desperado's Declaration,” The National Police Gazette August 16 (1879): 3; “Dodge City's Sensation,” The National Police Gazette July 21 (1883): 5; Nichols, George Ward, “Wild Bill,” Harper's New Monthly Magazine 34, no. 201 (1867): 273–286 Google Scholar.

65. Dykstra, The Cattle Towns, 144. Dykstra identified at least 15 killings between 1876 and 1885 in Dodge.

66. Roberts, Gary L., Doc Holliday: The Life and Legend (Hoboken, NJ: Wiley, 2006), 104–108 Google Scholar.

67. DeArment, Ballots and Bullets, 38–40.

68. Traub, “Rewards, Bounty Hunting, and Criminal Justice in the West,” 297–298, Newspaper Clippings, Stephen Hart Library, David J. Cook Papers MSS 725, Prassel, The Western Peace Officer, 135–136.

69. See, for example, Miscellaneous Papers, Denver Public Library (WHC) Cyrus Shores Papers (1883–1901), B1 FF3. This file includes letters from those requesting Shores’ help in finding work as deputies.

70. Ellis, Law and Order in Buffalo Bill's Country, 66–68.

71. DeArment, Robert K. and DeMattos, Jack, A Rough Ride to Redemption: The Ben Daniels Story (Norman, OK: University of Oklahoma Press, 2010), 10–21 Google Scholar, 52–55.

72. Ibid., 57–65, 70–71.

73. Ball, The United States Marshals, 95–97.

74. Moore, Cow Boys and Cattle Men, 175–204.

75. Ibid., 176–177, 186–199.

76. Courtwright, Violent Land, 66–108.

77. Isenberg, Andrew C., Wyatt Earp: A Vigilante Life (New York: Hill & Wang, 2013)Google Scholar.

78. See Swidler, Ann, “Culture in Action: Symbols and Strategies,” American Sociological Review (1986): 273–286 Google Scholar.

79. DeArment, Robert K., Knights of the Green Cloth: The Saga of the Frontier Gamblers (Norman, OK: University of Oklahoma Press, 1982), 80–99 Google Scholar.

80. 41 of the 255 participants in gunfighting in my sample worked as saloonkeepers/owners, dance hall proprietors, or gamblers (mostly in boomtowns) at some point in their careers.

81. Siringo, Charles A., Two Evil Isms, Pinkertonism and Anarchism (Chicago: C. A. Siringo, 1915), 43–57 Google Scholar.

82. O'Neal, Encyclopedia of Western Gunfighters. I supplement O'Neal's work with data from DeArment, Assault on the Deadwood Stage, DeArment, Deadly Dozen Vol. 1; DeArment, Deadly Dozen Vol. 2; DeArment, Deadly Dozen Vol. 3. I code the period between first and last gunfight as participation in the gunfighting system for a given gunfighter.

83. Though likely incorrect, this ratio still may not be that far off from the actual figure; for example, of the 1,786 employees of Fort Smith between 1865 and 1922 for whom we have information (Fort Smith was the U.S. District Court to which the Deputy Marshals in the Indian Territory and Oklahoma in the 1870s through 1910s were attached), only a dozen or so are listed as African-Americans. See the Fort Smith National Historic Site Database, National Park Service, accessed March 15, 2013, http://fosmcourtdatabase.nps.gov/.

84. See Burton, Arthur T., Black Gun, Silver Star: The Life and Legend of Frontier Marshal Bass Reeves (Lincoln, NE: University of Nebraska Press, 2006)Google Scholar.

85. Ed Bartholomew has claimed that there were over 25,000 individuals involved in the gunfighting system as a whole, though this number has never, to my knowledge, been independently verified. See Bartholomew, Ed, “Review of Encyclopedia of Western Gunfighters by Bill O'Neal,” Arizona and the West 22, no. 1 (1980): 77–78 Google Scholar.

86. See Figure 5. The online appendix includes information about the aggregate numbers of fights in various states.

87. See, for example, McKanna, Clare V., Homicide, Race, and Justice in the American West, 1880–1920 (Tucson, AZ: University of Arizona Press, 1997)Google Scholar. Also see the table in Journal of the Reconstruction Convention which met at Austin, Texas, June 1, A.D., 1868, vol. 1 (Austin, TX: Tracy, Siemering & Co., 1870)Google Scholar, 194, which lists 939 fatalities associated with Reconstruction violence in Texas alone from 1865 to 1868. These were precisely the kinds of events in which many future gunfighters participated. In making this argument, of course, I am aware that I am directly contradicting a strain of historical revisionism that contends that the post-bellum U.S. frontier was not an unusually violent place. However, recent scholarship has conclusively demonstrated the high frequency of firearm use and violence in the West, which is the general view I have adopted in this article. See Roth, Randolph, “Guns, Murder, and Probability: How Can We Decide Which Figures to Trust?,” Reviews in American History 35, no. 2 (2007): 165–175 CrossRefGoogle Scholar; McKanna, Clare V., Race And Homicide In Nineteenth-Century California (Reno, NV: University of Nevada Press, 2002)Google Scholar; Roger D. McGrath, Gunfighters, Highwaymen, and Vigilantes and especially Roth, Randolph, American Homicide (Cambridge, MA: Harvard University Press, 2009), 354–85Google Scholar. For older revisionist views, see Dykstra, Robert R., “Body Counts and Murder Rates: The Contested Statistics of Western Violence,” Reviews in American History 31, no. 4 (2003): 554–563 CrossRefGoogle ScholarPubMed; Hollon, W. Eugene, Frontier Violence: Another Look (Oxford: Oxford University Press, 1974)Google Scholar; Dykstra, Robert R., “Overdosing on Dodge City,” The Western Historical Quarterly 27, no. 4 (1996): 505–514 CrossRefGoogle Scholar; Prassel, Frank Richard, The Western Peace Officer: A Legacy of Law and Order (Norman, OK: University of Oklahoma Press, 1972)Google Scholar.

88. This decline is not simply a matter of mortality or professional disengagement; over 78 percent of gunfighting Civil War veterans were still alive in 1875, and more than 60 percent of veterans were still working as professional violence experts in the later period. Indeed, as late as 1895, 40 percent of known veterans in the gunfighting system were still alive.

89. I exclude service in guerilla militaries, as well as in militia, feuding, and vigilante groups, even if the experience involved some economic benefit, since these were primarily partisan commitments. I also exclude conventional military participation as professional work because such violence was not uniquely a matter of individual or small group effort. Instead, I try to identify those for whom individual competence in gunplay was remunerated.

90. There were 91 professionals and 52 non-professionals active exclusively in non-entrepôt settings during this period. Fisher's Exact Test for this difference in outcomes (two-tailed) is significant at p < .05.

91. If anything, entrepôt fighters were more likely to die violently than other kinds of gunfighters. For example, 22 entrepôt fighters died violently, while only 8 did not (27 percent). Conversely, 95 of non-entrepôt fighters (40 percent) appear to have died non-violently, while 139 or so did not. This difference, however, is not statistically significant.

92. Of fighters active between 1865 and 1885 who were participants in the gunfighting system for more than one year, 46 in non-entrepôt and 10 in entrepôt towns became professionals in later years, while 34 of the non-entrepôt and 1 of the entrepôt fighters remained non-professionals. This difference in outcomes is significant at p < .05 (Fisher's exact two-tailed test).

93. Difference of means tests assuming unequal variances allow us to reject the null that the value of a one-tailed difference of means statistic is greater than zero with p < .05; t = –0.89 with 47 observations of number of moves by non-entrepôt professionals (mean = 2.45, s.d. = .23) and 23 of number of moves by entrepôt professionals (mean = 2.83, s.d. = .36).

94. If we examine those who had their first fights in the six most “important” gunfighting towns from 1885 to 1900 (defined as the areas in which the largest number of fighters congregated), a statistically insignificant number became professionals in later years. 18 individuals fought in these key gunfighting areas; 14 (or 78 percent) became professional fighters, while a similar proportion (33 of 47, or 70 percent) of fighters in other areas did the same (Fisher's Exact Statistic [two-tailed]). The six areas (followed by number of gunfighters active during the period in question in parentheses) are: Tombstone, AZ (28); Ingalls, OK (14); Johnson County, WY (12); Gila County, AZ (11); Las Cruces, NM (10); El Paso, TX (10).

95. Bob Meldrum, for example, was good friends with Tom Horn and Cyrus Shores, as well as with A.E. Carver, Pinkerton's Denver Assistant Superintendent, who was key in helping connect him to reward and employment opportunities. See Dalton Brothers Gang, correspondence (1891, 1909–1915) 1/2; Pinkerton's National Detective Agency Part B: Criminal Case Files. Series 2, reel 1, microfilm collection (Indiana University). Jim Miller married into the Hardin family and was fairly close to John Wesley Hardin during his career as a bounty killer and “lawman” in southwest Texas. See Metz, Leon C., John Wesley Hardin: Dark Angel of Texas (Norman, OK: University of Oklahoma Press, 1998), 217–222 Google Scholar.

96. A few were “Unknown” and hence excluded from the analysis. The “commitments” of non-professional gunfighters were also coded, though they play little role in the following analysis.

97. See Tilly, “War Making and State Making as Organized Crime”; Olson, “Dictatorship, Democracy, and Development.” Of the 188 members of the subsample, 89 or 47 percent worked for a large-scale incorporation interest like a private detective firm, a cattle association, or for the state or federal government at some point in their careers, a figure that included well over half of the total number of mixed fighters (58 percent).

98. Fisher's Exact Test (two-tailed) is significant at the p < .05 level for 90 non-entrepôt (63 non-mixed; 27 mixed) and 28 entrepôt (12 non-mixed; 16 mixed) professional gunfighters.

99. Of the 23 professionals who had career trajectory sequences of more than three moves and who were active in entrepôts between 1865 and 1885, 13 shifted from incorporation to non-incorporation roles or vice versa after their time in such towns; of the 17 who had careers prior to their time in entrepôts, only 3 had pursued such opportunistic activity. This difference is significant at p < .05 (Fisher's Exact Test [two-tailed]).

100. Fisher's Exact Test (two-tailed) of differences in career trajectories of pre-entrepôt fighters and non-entrepôt fighters with careers of longer than length 3 who were also active between 1865 and 1885 is not significant at p > .10.

101. For simplicity, I simply define groups as weak components; components are the maximal subgraphs from a network in which a non-directional path exists from every node and every other node.

102. To be clear, clusters are not the same thing as groups as I have defined them. Clusters are sets of actors involved in some form of coherent activity that I have defined exogenously, while groups are endogenously defined to include all those among whom we can trace some connection in a network.

103. Component complexity should be distinguished from a well-known approach to evaluating signed graphs: structural balance. Specifically, complex components can either be highly balanced or not (that is, they can be partitioned into two distinct sets of actors that share positive ties “internally” and only have negative ties with actors in the other group). Most of these complex components were, indeed, quite balanced, indicating that actors were unlikely to fight against those with whom they had been allies. This might be interpreted as consistent with either a pattern of partisanship or one of market-based organization. Arguably, neither balance nor complexity alone tells us whether or not these actors fought according to a logic of partisanship/dueling or a market, however, since the distinction between actors fighting in factions according to economic or partisan interests requires having information about the specific careers of participants rather than merely topological information about their fighting patterns. Complex components do tell us that actors were involved in systems of interaction that allowed information about fighting others also known within the system to diffuse to both enemies and allies, which is consistent with a market approach to organizing force.

104. The data included in the network analysis involve observations for all fighters active during the period in question, whether or not they had ties to other fighters in the set. Multiple ties within each relation are dichotomized for simplicity.

105. This is the component ratio score. If Cs the number of components (including dyads and isolates) in a network, and ns the number of individual nodes, then the component ratio is $\displaystyle{{C - 1} \over {n - 1}}$ . See Borgatti, Stephen P., “Identifying Sets of Key Players in a Social Network,” Computational & Mathematical Organization Theory 12, no. 1 (2006): 21–34 CrossRefGoogle Scholar.

106. This measure is simply Blau's well-known heterogeneity index. If H is a set of relations among actors in a network and Q the set of components with more than 2 nodes in the network (where p _h is the proportion of edges of relation h and n _q is the number of nodes in component q), then component tie heterogeneity (F) is 1 − ∑ _h p _h ² for all edges in q. The average weighted heterogeneity across the network for all elements of Q is $\displaystyle{{\sum_q {f_q}{n_q}} \over {\sum_q {n_q}}}$ .

107. Crucially, these numbers should not be confused with the number of gunfights between dyads themselves, since a given dyad may have been involved in multiple gunfighting events.

108. To perform this analysis, I used Stephen P. Borgatti, Martin G. Everett, and Linton C. Freeman, “UCINET 6,” Analytic Technologies, 2002. (See the online appendix for a graphical depiction of this multi-dimensional scaling.)

109. Because it only assumes weak monotonic functional transformations between observations, the non-metric MDS usually provides a more robust representation of similarities or dissimilarities. Stress scores indicate how well the scaling fits the data; scores below .1 on two dimensions with 30 observations are considered acceptable. Though somewhat unconventional, I prefer this measure of distance to raw Euclidean distance scores as a means of calibrating social spatial relations because the coordinate measures allow for direct graphical comparison to physical location. Moreover, a powerful advantage of MDS is that it depicts dimensional relations among the data that distance measures alone cannot capture. Although the interpretation of the MDS along these lines has not been pursued in this article, future work will unpack the underlying dimensionalization of social space in violence markets in depth.

110. Mantel test (Monte Carlo): simulated p < .1; based on 1500 replications; n = 30. Based on Euclidean distance among observations in both social and geographic space (defined as latitude/longitude centroid). Euclidean distance between two objects i and j (where (x, y) are Cartesian coordinates) is simply $\sqrt {{{\lpar {x_i} - {x_j}\rpar }^2} + {{\lpar {y_i} - {y_j}\rpar }^2}} $ . To explore this relationship, I also performed a Quadratic Assignment Procedure correlation (a revised version of the Mantel test) in the UCINET software for social networks (5,000 permutations) correlating raw Euclidean distance scores between locales (rather than Non-metric Multidimensional Scaling [NMDS] coordinates) in terms of flow of gunfighters with the circle distances of the latitude and longitude centroids of the gunfighting locales and actually found a negative correlation of –.19 (p < .05). Like Mantel correlation, quadratic assignment procedures (QAP) use simulation to manage interdependency among the observations, though the Mantel is suitable for hypotheses explicitly framed in terms of distances. Regardless of whether it is positive or negative, both tests agree that the strength of the underlying association between social and physical space is minimal.

111. This result applies to multiple measures of geographic proximity. Taking the inverse of Euclidean distances of latitude/longitude centroids of locales with gunfights (rather than simply those sharing gunfighters) as a weighted distance measure (n = 30), the Moran's I expected value of –0.03 differs from the observed value of 0.03 (s.d. = .03) with a p-value of .02, allowing us to reject the null hypothesis that there is no spatial autocorrelation among gunfights at α of .05. Alternatively, using an adjacency matrix of state borders for western states only (n = 22) to weigh proximity, the level of clustering among fights again indicates spatial autocorrelation (Moran's I expected value of –.0476 differs from an observed value of .213 (s.d. = .113) with a p-value of .02).

112. Heumann, “The Tutelary Empire.”

113. For the identity of participants in the Adobe Walls conflict, see Baker, T. Lindsay and Harrison, Billy R., Adobe Walls: The History and Archaeology of the 1874 Trading Post (College Station, TX: Texas A & M University Press, 2001), 75–92 Google Scholar.

114. The online appendix includes a graphical depiction of this network.

115. Including both conflict and alliance ties, only 8 percent or so of the possible variation in the degrees of the 196 actors in the network was actually present in the network.

116. It is worth noting that the pattern of alliance and neutral ties within the main component of the direct inter-gunfighter network is highly clustered, implying that actors interacted mainly with those who also knew each other (the average weighted clustering coefficient for alliance/neutral ties within this component is .620, while the network density—the proportion of possible alliance/neutral ties present across the entire main component—was a mere .022, indicating strong neighborhoods among actors). This also suggests, however, that conflict was critical in forming the market, since over 64 percent of the nodes in the main component were unable to reach each other through alliance/neutral relations alone.