How Three Rivers Ended: the Closure Timeline Dispute

The popular account is that Ball Brothers bought the Three Rivers factory in 1937 and closed it in 1938. The state marker says so [@ev-0011; @ev-0041, p. 23], the TSHA Handbook of Texas says so [ev-0004], and Smith’s book says so [@ev-0038; @ev-0039].

It collapses at least four distinct events into one, and it gets the character of the central one wrong.

The correction comes from a source none of the compiled accounts used: the Transcript of Record in Hartford-Empire Co. v. United States, in which the United States prosecuted the glass-machinery combine and Ball Brothers was a defendant [ev-0059]. Three Rivers appears in it throughout — in sworn testimony, in the parties’ own correspondence subpoenaed from their files, and in the Government’s formal charge. It is the only held source in which the people who ended the company speak on the record about having done it.

The sale: a trustee’s sale, in a state court

Hartford-Empire’s secretary Safford was examined on the sale exhibits. Exhibit 5745 “states that the Court ordered notice to be given of the proposed sale on November 27th, 1936, of the Three Rivers Glass property to the Guardian Loan Company, Trustee” [@ev-0059, part 09, fol. 10219]. Exhibit 5746 is the notice itself, giving “notice that a hearing as to whether the sale should be confirmed will be held in the District Court at George West, Texas, on November 27, 1936, and calling on all parties at interest to appear and register any opposition they may have to the sale” [@ev-0059, part 09, fol. 10219]. Exhibit 5747 is “a letter of December 4, 1936, addressed to the witness by Mr. Searcy” [@ev-0059, part 09, fol. 10220].

Three things follow, and each corrects a held source.

It was not an auction. BRG reports that “the Three Rivers property was auctioned, and the Ball Brothers Corp. made the highest bid of $80,000” [ev-0042]. The trial record describes a trustee’s sale to a named trustee, with a confirmation hearing at which parties could object — a different procedure, and one in which there is no highest bidder. Nothing in the extract held supports the $80,000 figure.

It was a state court, not a federal one. Tips remembered that Ball “went into the Federal Court because the people that were working to put us out of business were from outside of the State,” and that “the Federal Court allowed them to foreclose on the first mortgage note and take over the factory” [@ev-0041, p. 44]. The confirmation hearing was in the District Court at George West, the Live Oak County seat [@ev-0059, part 09, fol. 10219].

A fuller reading of the record shows what he was conflating. There were three federal proceedings, and Hartford-Empire’s secretary enumerated them under oath: Hartford filed suit against Three Rivers “on October 21, 1932, in the United States District Court at Corpus Christi”, taking a consent decree for repossession of its machinery and judgment for royalties; a petition under 77-B put the company into federal reorganization under a second receiver until it was dismissed in March or April 1935; and a second 77-B petition moved the assets back to Corpus Christi in 1936 — “the third time”, as the court observed, that it had been in federal court [@ev-0059, part 09, fols. 10181, 10221-10224]. Tips was a defendant in a federal suit brought by the licensor over the machinery, and his company twice passed through federal reorganization. What was not federal was the sale itself, which the state court in his own county confirmed.

The creditors were bought down first. “Then in the Fall of 1936 the negotiations commenced between the Three Rivers Glass Company and the Ball Glass Company, or some of its representatives, and the creditors were asked to take forty percent of their claims. That was done. Hartford-Empire Company, as one of the creditors, received forty per cent of its claim, and the plant was sold, I believe, to Ball Glass Corporation” [@ev-0059, part 09]. No held source outside the trial record mentions the forty-percent composition.

The machinery was ordered out of the plant on the same day. On Hartford-Empire’s application of May 27, 1935, the court entered an order — “Entered and rendered in open court on November 27, 1936” — granting Hartford “the right of possession of said machines and equipment” and “permission to take possession of and remove said machines and equipment from the manufacturing plant of Three Rivers Glass Company at Three Rivers, Texas”, and directing the receiver to permit it [@ev-0059, part 18, line 66821]. Ball bought the plant on the day the feeders were ordered removed from it. Whether they were in fact removed is a separate question: in May 1938 Hartford was still writing to George A. Ball about “our equipment which the Ball Glass Corporation is now using at their Three Rivers plant” and pressing for a signed licence [@ev-0059, part 16, fol. 60665]. Hartford’s own licence history records the transaction in five words: “11/27/36 – Three Rivers – acquired by Ball” [@ev-0059, part 18, line 68242].

The concealment: Government Exhibit 1200

The step before the sale is documented in the buyer’s own counsel’s words. On September 14, 1936 — ten weeks before the confirmation hearing — Wm. C. Church of Church & Steger, San Antonio, wrote to S. S. Searcy, attorney at law at the Frost National Bank, enclosing in duplicate the letter they proposed sending to the creditors of the Receiver of Three Rivers Glass Company [@ev-0059, part 16, printed p. 13805].

“While we do not want it generally known, and ask that you and your clients keep the information confidential, the prospective purchaser that we represent is Mr. George A. Ball of Muncie, Indiana. He has asked us to get you and the Hartford Empire Company to assist him in purchasing the Three Rivers plant, and not doing anything about the new licensing agreement should he become the purchaser, until after he has obtained the plant and continues the operation.” [@ev-0059, part 16, printed p. 13805]

This is the documentary form of what Tips recounted from memory as a betrayal by “a crooked attorney” [@ev-0041, pp. 43–44]. The record shows something more specific than a single crooked lawyer: Ball’s identity as purchaser was deliberately withheld from the creditors being asked to accept forty cents on the dollar, and Hartford-Empire — itself one of those creditors — was asked to assist the purchase and to hold off on the licensing question until Ball had the plant.

Tips remembered the shape of it correctly without having seen the letter.

Production ended in December 1937, not 1938

Ball Brothers’ own correspondence with Hartford-Empire settles this. Writing to Hartford about feeder royalties, Ball states: “Three Rivers is a subsidiary of Ball Brothers Company, and therefore, the feeders in that plant are covered in our agreement. As you know, Three Rivers has not been operated since December—We are operating at Wichita Falls” [@ev-0059, part 16].

The 1938 and 1939 closure dates in the compiled accounts [@ev-0011; @ev-0004; @ev-0042] therefore describe something other than the end of production. The March 31, 1938 deed [@ev-0041, p. 62] is the corporate transfer, and Ball still had a factory manager at Three Rivers writing employment references in September 1938 [@ev-0041, p. 64] — a going administrative concern at a plant that had stopped making glass.

BRG judges Brantley’s 1939 “most likely correct” against Toulouse’s 1947 [ev-0042]. Brantley’s own p. 95 register, now read verbatim, says “closed in 1939 and dismantled, sold 1954”; his p. 30 gives 1937 for the acquisition, not 1936 — internally inconsistent on the sale year but unambiguous on 1939 for the closure [@ev-0062; @ev-0068]. Toulouse’s 1947 is the divestiture order date, not the production-cessation date [ev-0062]. On production, the trial record puts both later than the fact.

Why Ball closed it — in Ball’s own words

On July 16, 1938, Ball Brothers wrote to Hartford-Empire’s secretary:

“Thre Three Rivers plant and equipment was [acquired] from the receiver. It was in due time closed down for repairs and at the time those repairs were completed there was not enough business in that section to make possible the operation of both Three Rivers and Wichita Falls.” [@ev-0059, part 16, Government Exhibit 1210]

Ball goes on to explain that it had displaced Owens machines at Wichita Falls with feeder machines carrying higher royalties than “the two or three feeders formerly operated at the Three Rivers plant,” and that it therefore preferred “not to enter into any license agreement relating to the Three Rivers plant at this time” [@ev-0059, part 16].

This is not the story of a plant that failed. It is a statement that the acquiring company had two plants in one region and kept the other one.

The Government’s charge

The United States put it plainly. Among its findings against Ball Brothers:

“(5) Ball Brothers acquired and shut down and is continuing to keep closed the plant of Three Rivers Glass Company, its principal competitor in the State of Texas.” [@ev-0059, part 18, fol. 64374]

The mission asks whether the end was business failure or anticompetitive destruction. The United States, in the proceeding where the question was litigated, charged the latter. That is not the same as a court having found it proved as to Three Rivers specifically, and this article does not claim it is — but it is a great deal more than folklore, and it is what Tips said had happened [@ev-0041, p. 44].

Disputed: did Three Rivers make fruit jars?

This matters because BRG’s account of the company’s undoing turns on it. BRG states that Three Rivers “inaugurated fruit jar production in 1928,” which “had not played the game according to the Hartford-Empire rules,” and that “the Ball Brothers repeatedly sent notices to Three Rivers to drop jar production” [ev-0042]. On that account the company provoked the competitor that destroyed it.

Three Rivers’ own chief chemist, examined under oath, said otherwise. Asked what the company manufactured, Coleman — later president of Knape-Coleman Glass — answered: “They made a general line of packers’ ware, and for several years manufactured milk bottles” [@ev-0059, part 04, fol. 3430]. Asked directly whether they made fruit jars:

“A. No, not so-called. They were fruit-packers ware, one trip container; they weren’t Mason type jars. Q. Did they sell their packers’ ware to be used as fruit jars? A. Well, I think Mr. Ball said they did.” [@ev-0059, part 04, fol. 3430]

The last answer is the crux and it should not be smoothed over. Coleman is not denying that the ware competed with fruit jars; he is saying the characterisation was Ball’s. Whether Three Rivers breached its Hartford-Empire licence or was said to have breached it is a different question with a different answer, and the held sources give opposite ones.

Note that BRG’s own product listings for the receivership years name “fruit jars” among what the plant made [ev-0042], and that the 1936 American Glass Review listing includes them [ev-0042].

Coleman was not the first Three Rivers man to give that answer. In January 1935 Hartford-Empire’s president wrote to Frank C. Ball, replying to a complaint about Mason jars reaching a grocery company, and recounted what Henderson Coquat had said when Hartford summoned him to explain whether the receivership was “violating his license and lease agreement… due to the fact that he was manufacturing a domestic fruit jar”:

“Coquat claimed that under no conditions had [he] manufactured fruit jars for the domestic trade—showed us letters in which he had declined to manufacture fruit jars for the domestic trade—but did admit that on orders received from one or two packers he had supplied a straight line jar with a screw top thread. His position then was that the packers had ordered these jars as far as he knew for packing purposes, and that if any of them had gone out and bought caps, and had then sold the jars and the caps as a domestic fruit jar, such action was beyond his control.” [@ev-0059, part 16, fol. 59030]

That is Coleman’s answer, given five years earlier by a different man, and recorded by the party he was answering to. And the industry’s own sales correspondence concedes the same facts from the other side: a memorandum on Three Rivers describes it “selling plain pint and quart jars, without caps, to be used by the wholesale grocers in connection with fruit jar closures bought from Bernardin or Crown Cork and Seal” [@ev-0059, part 16, fol. 60285].

So the dispute is narrower than it looked. Both sides agree on what left the factory: plain screw-thread packers’ jars, sold without closures. They disagree on what to call the result. BRG’s “inaugurated fruit jar production”, read against this, is a characterisation the company contested at the time and contested again under oath — not a fact it conceded. The conflict is real and is left open, but it is a conflict about a word.

Coleman also dates the receivership differently

Coleman testified that “The Three Rivers Glass Company went into receivership, I think, in the summer of 1932” [@ev-0059, part 04, fol. 3430]. BRG dates the bankruptcy proceedings to October 1932 [ev-0042]. Coleman’s “I think” is his own hedge and he was testifying years later; the variance is recorded, not resolved.

The trial record favours October on the narrow question of when the receiver was appointed. Safford’s sworn narrative runs: “In July, 1932, was the first time which we sent cancelation notice to the Three Rivers Glass Company. And then we filed suit against them on October 21, 1932, in the United States District Court at Corpus Christi. Subsequent to our filing suit the first mortgage bondholders had a receiver appointed in the State court” [@ev-0059, part 09, fol. 10221]. The appointment therefore followed October 21. Coleman’s “summer of 1932” may be remembering when the trouble became terminal — Hartford’s cancellation notice went out in July — rather than when a court acted.

The trade association’s count: “several times”

One more voice in the record counts the receiverships, and it is neither the company’s nor a creditor’s. Among the exhibits is a letter from the Glass Container Association, signed by its assistant business manager, answering somebody who was thinking of building or buying a glass plant. The whole letter is an argument against doing it. It sends the correspondent to look at idle capacity nearby — Hazel-Atlas at Blackwell, Oklahoma, “down for the last year due to lack of volume”; Obear-Nester’s Kansas City plant “down for several years”; Liberty Glass at Sapulpa, which “manufa[ct]ures milk bottles entirely for the Southwestern distric[t]” and has been handicapped because the district could not support it. Then it widens the circle:

“Stretching out a little further we have the Dixie Glass Bottle Manufacturing Company at Jackson, Mississippi, which was organized about two years ago, and which has recently gone into the hands of the receiver because of unprofitable operations. I might also mention the Three Rivers Glass Company of Three Rivers, Texas, which has been in the hands of the receivers several times during the last few years.”

“My earnest plea, however, is that yo[u] do make an investigation before investing a considerable amount of money in a plant which may have slight possibilities of a successful performance.” [@ev-0059, part 16, before fol. 59554]

Three things follow, and one of them is a caution against the other two.

It is an outside count, and it is in the plural. The trade body for the industry, writing to a stranger with no reason to flatter or to blame, treated repeated receivership at Three Rivers as a known fact of the trade — the sort of thing one cites without needing to prove. That is independent of Safford, of Coleman, and of Tips, all of whom had a position.

It also puts the ending in its industry. This plant is offered as an example of a general condition, alongside a Mississippi company two years old and already gone, and modern plants standing idle for want of volume. Whatever Three Rivers did wrong, it was doing it inside a market the association was telling people to stay out of.

The caution is that the letter is undated in the window this expert holds and its recipient is not named, so “the last few years” cannot be anchored. It cannot be used to fix how many receiverships there were or when. What it can be used for is the shape of the thing: by the time this was written, the plant’s instability was general knowledge in the trade, and repetition rather than a single collapse was what the trade remembered about it.

What the sequence looks like once separated

Four different endings, four different dates, and the popular account collapses them.

What ended	When	On what
The company’s independence	September–November 1936	Concealed purchase plan, September 14 [@ev-0059, part 16]; creditors composed at forty percent; trustee’s sale confirmed at George West, November 27 [@ev-0059, part 09]
Production	December 1937	Ball to Hartford: “not been operated since December” [@ev-0059, part 16]
The corporate transfer	March 31, 1938, filed April 11	Live Oak County deed, Book 71 p. 507 [@ev-0041, p. 62]
Ball’s administrative presence	after September 1938	Ball Glass Corporation employment reference, September 9, 1938 [@ev-0041, p. 64]

What is still open

The removal of the machinery — three questions, and only one of them open. An earlier version of this section said that nothing in the extract held records when the plant was dismantled, and rested that on Government Exhibit 1209. Both halves were wrong, and separating the questions dissolves most of the gap.

When Hartford became entitled to take its equipment back is answered exactly: November 27, 1936, by the order set out above [@ev-0059, part 18, line 66821]. Whether it took it is answered in the other direction. In May 1938 Hartford was still writing about “our equipment which the Ball Glass Corporation is now using at their Three Rivers plant” [@ev-0059, part 16, fol. 60665]; Ball wrote back that “the two or three feeders formerly operated at the Three Rivers plant” were covered by the general Ball licence and declined a separate agreement [@ev-0059, part 16, fol. 60668]; and on cross-examination Safford was put the point directly — that although he had obtained an order, he left the machinery where it stood [@ev-0059, part 09, fol. 10268]. A wider sweep of the transcript confirms this: the royalty stipulation shows Ball Glass Corporation paying Hartford $800 a year for machines in a plant making no glass, and George A. Ball’s own testimony in Part 5 agrees that the feeders were never removed [ev-0069]. The feeders did not leave in 1936.

The itemisation this section previously credited to Exhibit 1209 is in a different document and points the other way. It is in Church’s own September 1936 circular to the creditors, and it is a statement about what the receiver did not own: “One of the lehrs in the plant is owned outright by the Simplex Engineering Company and the other two lehrs and all of the feeders are owned outright by the Hartford Empire Company, and the Receiver has no interest therein” [@ev-0059, part 18, fol. 63510]. That is worth stating plainly, because it changes what the November sale conveyed: the forming equipment and two of the three lehrs were never the receivership’s to sell.

What remains genuinely open is the third question — when the plant was finally stripped. Smith has Ball beginning to dismantle in the last days of 1937 [ev-0038] and the marker research records Ball keeping a night watchman for several years [ev-0041]; the four parts held carry nothing that dates the dismantling, and the last thing they show is equipment still in place and still being argued over in the summer of 1938.

That question has since been narrowed, and the narrowing corrects how it was posed rather than answering it. The plant was not stripped on one occasion. The melting tank outlived the buildings: Ball remodelled it, and when Bryan Boyd and Harry J. Schulz bought the remaining property in September 1954 it was “in perfect working condition” [@ev-0041, pp. 62, 65]. Whatever came down in 1937 or 1938 was the sheds, the machinery and the railroad loading dock — not the furnace, which stood for another sixteen years, was offered to glass companies that all refused it, and was still being sold for its refractory brick in 1960. See concepts/three-rivers-glass-factory-site-1938-1973.md.

The $80,000 figure — now disputed. BRG gives $80,000 for the sale [ev-0042], and the extract held does not contain a sale price at all — the only capitalisation figure in it, $72,000, is the capital of Knape-Coleman and has nothing to do with this sale [@ev-0059, part 04, fol. 3431]. But the San Antonio Evening News of December 5, 1936, reporting under the headline “Three Rivers Glass Plant Reorganized,” gives a different figure: $130,000, naming Church and L. L. Bracken [ev-0066]. Whether BRG’s $80,000 comes from Roller or from another source, it is now contradicted by a period newspaper.

Fifty-four parts unread. The admitted evidence is an OCR extract of parts 4, 9, 16 and 18 of a fifty-eight-part item [ev-0059]. The pleadings — Ball’s Answer, and George A. Ball’s own sworn testimony — are not in those four parts and have not been read. A research note has searched all fifty-eight parts and located material this extract does not carry, including George A. Ball’s testimony in Part 5 and the Government’s brief in Part 54 [ev-0069]; those readings should be checked against the primary pages when the extract is extended.

The 1937–1941 trade directories. The American Glass Review glass factory yearbook for 1937–1941 would settle the closure year outright but those editions are not digitised; physical holdings are at the Rakow Research Library, Corning Museum of Glass [ev-0064]. The 1939 Census of Manufactures counted one glass-container plant in Texas [ev-0064], consistent with the Three Rivers plant being closed and Ball’s Wichita Falls plant remaining open. The trade journal Glass Industry does not mention Three Rivers in any of its 110 issues from January 1935 through February 1943 [ev-0070].

The exact wording. The extract is machine OCR with no page images matched [ev-0059]. Every quotation above should be treated as reliable in substance and provisional in wording until checked against the page image.