Ordinarily, a full-scale test would involve samples taken from large pits spaced systematically across the field, as Fuller had proposed in 1908. Material from each pit would be taken carload by carload to the plant for careful processing. This way the engineers would know precisely where diamonds came from and could judge the general quantity and quality at the spot. In previous testing, however, Reyburn’s operation had generally followed the standard procedure, and had consumed time and money for very little return except from the natural concentration of diamonds on the eroded surface. Covering the thirty-five acre slope satisfactorily with a grid and pits might be conducive to good recordkeeping, but was it necessarily the best method for gathering and processing the huge tonnage required for a full-scale test of this particular property?
To excavate the peridotite rock, Sam Reyburn and Stanley Zimmerman chose to use trenches instead of pits and a dragline bucket scraper in place of tramcars. Four wide trench lines were staked out, beginning at a single point near the plant and radiating upslope some 1,200 feet. Three branched northward across the main part of the diamond field; the other angled more westward, providing a sampling of the two harder types of peridotite as well as the volcanic breccia (the plant finally included a crushing unit). After the surface layer was removed, the rock matrix was prepared by drilling holes and inserting “30 per cent ammonia dynamite fired by electricity.” Then the big dragline scraper, pulled with the steam engine at the plant, eventually cut the trenches to a depth of some twenty feet.
In the big plant, however, the surface material still proved frustratingly hard to break down. The black gumbo dissolved to a certain extent, but then tended to roll into little balls, usually with solids concealed inside. Run over the grease table, these simply passed through. Only simple “hand washing,” with a set of screens in a tub, could break the clay down completely: surface soil was manually stirred and forced through each screen, “sizing” the material progressively smaller. If the mesh of the bottom screen was fine enough, virtually no clay balls survived as tailings. But hand washing, alone, could hardly suffice in a full-scale test; Zimmerman and his plant manager, M. J. (“Scotty”) Eunson, had to come up with a faster, yet reasonably efficient solution. Without the diamonds from that rich surface layer, the overall operation would seem anemic, to say the least.
In late December 1920, the plant shut down for over two months so the equipment could be adjusted. About sixty workers, mostly from Murfreesboro, were sent home. After restarting, the operation lasted only about two months—just long enough to produce rumors of large diamonds and again raise hopes that the development of local resources might “become a reality instead of a dream.” Then Zimmerman closed the plant indefinitely and went to England and South Africa to study methods and machinery.
New machinery began arriving at the mine in July 1921, including a new crusher to improve treatment of the harder rock matrix on the west side of the test area. Awaiting Zimmerman’s return, as well as remaining shipments of surface-mining equipment, field crews used “small hand washing operations” to work a selected area of the slope above the plant. They focused on sand-and-gravel deposits where natural erosion had left a concentration of heavy minerals, and hit what must have seemed a small mother lode. In one two-week period, they washed 48.0 carats of diamonds from 100 loads; in another, 33.0 carats from 100.
That 48.0-carat batch proved doubly exciting when it produced a record diamond. Described as “a fairly good 20.1+ carat white stone,” the find was hastily evaluated at $10,000, uncut.
In October 1921, Zimmerman and crew finished setting up a new surface-mining operation. Switching to the small plant nearby, the one John Fuller built in 1909, they modified the screens, added a grease table, installed a water pump and engine, and then ran a flume from the plant out into the field. Instead of merely washing the surface layer through a sluice box, powerful water hoses flushed it down the long chute to a holding pond. The saturated, broken down material was scooped up and sent through the simple screen-and-jig processing plant. In a general sense, the ADC had gone back to the method Lee Wagner used previously.
Sam Reyburn visited the mine in early December and found it “running at full capacity.” And, for the first time, the hydraulic mining operation was treating the gumbo thoroughly. Understandably, the Pike County Courier lauded the man whose “faith and influence” had created the foundation of a promising future. The paper again projected images of a much larger mining operation that would employ “hundreds of employees and would insure Murfreesboro of a rapid growth.” The company, its success now “an assured probability,” seemed on the verge of paying large dividends on its stock. There were rumors that the company, if profitable, intended to build homes near the mine for its employees.
As the eighteen-month test period ended in late spring of 1922, work halted except for the “placer mining” on the surface. “Many men were thrown out of employment by the stopping of part of the machinery,” the Courier reported in May, “but there still remains a large force at work and from rumors we understand that operations may open again in full blast within a few weeks.” After that, the rumors continued, but in a few months it was clear the test of the matrix was over and the big plant had closed.
Only one record of the test results is available, John Fuller’s comprehensive mining report of 1931. As expected, the surface yield ran much higher than that of the matrix, even though the problems with the gumbo reduced the overall average considerably. The yield of the basic run of matrix—0.38 carat per 100 loads—to some extent reflected the effect of unproductive rock out in the west trench, beyond the breccia. If Fuller’s figures were correct, however, the abnormally low yield mostly indicated the loss of small diamonds: only 253 diamonds weighing 176.07 carats were recovered from 45,961 loads of pure peridotite, for an average size of 0.70 carat, well above the ADC’s historical average of almost 0.50 carat. This either meant a loss from faulty processing or, more likely, reflected the mesh of the smallest screens used in the plant to size the rock.
Later, John Fuller and others insisted the low yield of the matrix resulted from inefficient management and faulty processing—Zimmerman was simply inexperienced and inept. “When I watched the plant a few times,” Fuller said, “there was a tendency to rush the material through the plant. . . . The way the equipment was put together and the haste with which the material was run through the plant led to the loss of diamonds.” Some critics suspected the test had been designed to fail: DeBeers and the international diamond cartel supposedly had managed to suppress the Arkansas mine. Yet, the fundamental problem lay not in management or machinery, but in the consistently poor performance of the matrix. Whatever losses occurred during processing derived overwhelmingly from the surface material; and as Reyburn and Zimmerman acknowledged at the beginning, that thin layer had a very short commercial life span.
 Reyburn and Zimmerman understood the difficulties of extracting diamonds distributed at random in the expansive matrix. See their long discussion in “Diamonds in Arkansas,” Engineering and Mining Journal, 984-985 (section headed “Ratio of Concentration in Diamond Washing 14,500,000 to 1”), which was preceded by essentially the same treatise in “Prepare for Test in Diamond Field.”
 “Diamonds in Arkansas,” 985; Fuller, map, “Prairie Creek Peridotite Occurrence”; photos, 23-1, 23-24, 23-26, and 23-87, VIII, Crater archive. Number 23-1 is a three-shot panorama of the four cuts, with the plant’s water tower and outbuildings nearby. A few photographs are also in the Mauney Records, Murfreesboro.
 Sources in the following endnotes made it clear the surface layer was the basic problem, but the balling was not mentioned specifically until almost a year after the big plant closed permanently: “15,000 Gallon Tank Falls at Mine,” Pike County Courier, August 17, 1923, p. 1. The piece referred to the current mining of the surface layer, “which they were unable to mine properly in the big plant on account of the dirt balling up and passing over the grease board.”
 “Hand washing” appeared noticeably in the sources, after 1920 and in previous years as well. The photographic record best clarifies the equipment. Experience washing the clayey material during recreational digging at the Crater leaves no doubt the balling would have been an insurmountable problem in the automated equipment of the big plant, especially in the finer-mesh screens (where clay bits would be retained and keep rolling around, gathering additional solid particles).
 “Diamond Plant Ceases Work,” Courier, January 7, 1921, p. 1 (at first, said “probably 65 or 75” laid off); “Diamond Mines May Open Soon,” ibid., March 4, 1921, p. 1 (corrects: “The mine ceased operating in December and about 60 men were thrown out of employment. Most of the employees are residents of Murfreesboro . . .”)
 “Diamond Mines Attracting Attention,” ibid., March 18, 1921, p. 1. The article cited State Representative Ebel’s statement about diamonds weighing “80 carats” and “18 carats,” reportedly found on the ADC’s property. An informed source denied the first. At varying times, other published items also referred to diamonds of eighteen carats (statement of Sam Reyburn, in “Prepare for Test in Diamond Field,” Arkansas Gazette, November 14, 1919, p. 5: without mentioning the color, Reyburn said the diamond weighed 18¾ carats and sold in the rough for $100 a carat, which was much less than the value of the canary yellow found in 1917); “Big Stone Found at Diamond Mine,” Gazette, October 9, 1921, p. 1, reprinted as “Rich Find at Diamond Mine,” Courier, October 14, 1921, p. 1: while reporting a 20¼-carat just found at the mine, the article referred to the previous record diamond, “a perfect white stone weighing 18.3 carats, [which] was sold in the rough for $6,800 to Tiffany of New York”).
 Arkansas Oil and Mineral News, reprinted as “Drop in Diamonds,” Courier, May 27, 1921, p. 1.
 “New Diamond Machinery is Arriving,” Courier, July 22, 1921, p. 1
 “Big Stone Found at Diamond Mine,” Gazette; “Rich Find at Diamond Mine,” Courier. Zimmerman described the finds. Fuller’s composite mining map of 1931 outlined the surface workings of 1920-1922, north of the big plant and looping around the smaller area where Lee Wagner washed out the big canary yellow gem in 1917 (“The Prairie Creek Peridotite Occurrence, Showing the Arkansas, Ozark and Mauney Mines, with Surface Workings, Topography and Approximate Locations of Intrusive Rocks,” December 1931, large folded map, VI.3, Crater archive; microfilm).
 “Big Stone Found at Diamond Mine”; “Rich Find.”
 Ibid.; cf. O. L. Brace, Petroleum and Mining Geologist, Camden, Arkansas, report to Gordon Ingalls, McWilliams Bldg., El Dorado, Arkansas, October 17, 1923, “Misc.” box, Crater.
 “Prominent Visitor Here,” Courier, December 9, 1921, p. 1.
 Brief untitled item, Courier, May 19, 1922, p. 2.
 The composite mining map accompanying Fuller’s report of 1931 used October 1922 as the break between the test and the continuing surface operation with the small plant (Prairie Creek Peridotite Occurrence, supra).
Shortly before his death, Sam Reyburn commented publicly about the test: “I decided to send Zimmerman to South Africa to observe and, through my Morgan contacts, got letters of introduction to the De Beers people. He came back and we kept going until 1922, when we gave up. . . . I offered to quit as president, but the stockholders wouldn’t let me” (Gross, “Diamond Mine Mystery,” 99). Reportedly, Reyburn testified during a court case in 1925 that he also put his stock up for sale, and that his wife, the only bidder at public auction in New York, got them for 5¢ a share (“Diamond Operators ‘Inefficient,’ Avers State, Resting Case,” El Dorado Daily News, October 16, 1925, pp. 1-2).
In late 1923, an engineer visiting the mine reported the “manager” told him the company planned “to remodel the large plant and resume operations in the near future” ((O. L. Brace, Petroleum and Mining Geologist, Camden, Arkansas, to Gordon Ingalls, McWilliams Bldg., El Dorado, Arkansas, October 17, 1923, “Misc.” box, Crater archive).
 Fuller, “Estimate of Unit Value—2. Tests of Diamond Contents in Arkansas Mines,” in “Reports and Information,” 36. The report used only one entry for the period of testing (October 1919-October 1922) and the following surface mining (October 1922-July 1925): “Oct. 1919-July 1925.” The composite mining map accompanying the report divided the two.
Although Fuller’s two entries for surface mining were labeled “Surface and Peridotite,” the latter probably indicated that some of the softened green matrix immediately underlying the humus-enriched “black ground” had been taken along with that layer and its heavy minerals, particularly during hydraulic mining. Recreational digging in buried drainage channels at the Crater often have shown a thick black layer of wash overlain by a thin green band. The black layer consisted of light clay soil and very small gravel, and was virtually barren of diamonds or other heavy minerals. The green held larger, heavier gravel and rock, and consistently produced diamonds. This indicates the lighter humus material washed ahead, forming a layer in the drainage cut, while the heavies lagged behind.
 In his speech in Little Rock, April 7, 1928, Fuller said in his opinion the results were negative “because the work was improperly done. The property today is just as valuable and is in exactly the same position so far as its prospective value is concerned as it was when I first saw it in 1909. All of the money that has been spent on it, in my opinion, might just as well be thrown into the Arkansas River because results were not obtained in the right way” (“Public Statement,” in “Reports and Information,” 33; also see “Historical Summary,” ibid., 2: “This work, as supervised by Zimmerman, proved commercially unsuccessful, due to the failure to apply known methods of testing diamond bearing areas”).
As some critics said, Zimmerman’s experience before 1920 was in other kinds of mining, and he himself acknowledged he had been a mining engineer only seven years (Affidavit, ibid.). Yet, he was hardly ignorant of diamond-processing methods, as the article in the Engineering and Mining Journal demonstrated (Reyburn and Zimmerman, “Diamonds in Arkansas“). In May 1931, Zimmerman wrote a letter of recommendation for his long-time foreman who had joined him in Murfreesboro in 1919, M. J. (“Scotty”) Eunson. Zimmerman mentioned both his and Eunson’s mining experience in South America and elsewhere. He indicated he was a manager in those jobs, with Eunson as foreman (copy of letter, S. H. Zimmerman, Murfreesboro, to [name blacked out], Denver, May 6, 1931, Conway papers, Pike County Historical Society archive, Murfreesboro). Zimmerman’s affidavit of June 11, 1920, stated he had received his education at the Michigan College of Mines.
 Howard Millar dominated this school of thought. His memoir was especially blunt toward Zimmerman, whose operation “was not designed to find many diamonds.” To Millar the reason seemed clear: “Looking back, I can see why he did not avail himself of any help I offered. I also understand why he had so few security arrangements around his recovery equipment. I strongly believe that he was sent to Murfreesboro with instructions to ‘prove’ the field to be potentially non-productive of profit. Thus the diamond syndicate would face less possibility that others would finance operations by my father and myself, operations that might prove competitive to the syndicate” (Finders-Keepers, 60-6l).
Martin Gross’s equally imaginative article reflected the broader suspicion reemerging after World War II (“Incredible Mine Mystery,” 99). Gross cited Howard Branch, a New York diamond merchant who had cut some of the ADC’s diamonds: Branch “intimates that Zimmerman’s giant flop may not have been accidental. ‘What better way to discourage future mining than make a big show and have it fail?’ he asks. Zimmerman, despite his failure, reportedly turned up in an important New York executive position.”