United States v. Empire Gas Corporation

ROSS, Circuit Judge. This civil antitrust action against Empire Gas Corporation was brought by the United States under sections 1 and 2 of the Sherman Act, 15 U.S.C. §§ 1 and 2. The case was tried to the court, judgment was for Empire, United States v. Empire Gas Corp., 393 F.Supp. 903 (W.D.Mo.1975), and the government now appeals. We affirm the district court. Empire is a retailer and to a lesser extent a wholesaler of liquefied petroleum (LP). LP is a generic term for any of various gaseous fuels such as propane and butane, which are compressed into their liquid states for marketing. Retailers generally store their LP inventories at bulk plants, from which it is distributed by tank truck to consumers for heating, cooking and other uses. Because of the high cost of this method of transportation, a retailer’s sales are usually limited in area to approximately a 30 mile radius around his bulk plant. Frequently LP retailers are small, local, family-run businesses. Empire was founded in Missouri in 1963 and still maintains its home office in Wheaton. During the first 10 years of its life it greatly expanded its business in Missouri and added bulk plants in 24 other states. This expansion was largely through acquisition of 81 other LP gas retail companies with close to 400 bulk plants. The complaint alleged several violations of the antitrust laws, 1 but on appeal the United States has concentrated on their contentions that Empire violated section 2 of the Sherman Act by attempting to monopolize the retail sale of LP in areas surrounding Lebanon and Wheaton, Missouri, and that Empire restrained commerce in violation of section 1 by obtaining covenants not to compete from its employees and others. I. Attempt to Monopolize. In order to establish an “attempt to monopolize * * * any part of the trade or commerce among the several States * * * ” under 15 U.S.C. § 2, the government was required to show Empire’s specific intent to monopolize and a dangerous probability of success within a relevant *299 product and geographic market. Agrashell, Inc. v. Hammons Products Co., 479 F.2d 269, 284, 286 (8th Cir.), cert. denied, 414 U.S. 1022, 1032, 94 S.Ct. 445, 461, 38 L.Ed.2d 313, 323 (1973). The district court held that the plaintiff failed to prove any element of its case. A. Specific Intent. Specific intent need not be proved when it is alleged and proved that monopolization has been accomplished, United States v. Griffith, 334 U.S. 100, 105, 68 S.Ct. 941, 944, 92 L.Ed. 1236, 1242 (1948), but it is an element when, as here, the charge is an attempt to monopolize. Hiland Dairy v. Kroger Co., 402 F.2d 968, 971 (8th Cir. 1968), cert. denied, 395 U.S. 961, 89 S.Ct. 2096, 23 L.Ed.2d 748 (1969); United States v. Alcoa, 148 F.2d 416, 431-432 (2d Cir. 1945). The relevant geographic areas here are the Lebanon and Wheaton market areas; however, at oral argument the United States contended that Empire’s actions in other areas could support an inference of monopolistic intent in the relevant geographic areas. With this we agree. We have stressed the importance of viewing the evidence as a whole to give the antitrust plaintiff the full benefit of his proof, rather than tightly compartmentalizing the case and wiping the slate clean after considering each piece of evidence. Sanitary Milk Producers v. Bergjans Farm Dairy, Inc., 368 F.2d 679, 691 (8th Cir. 1966). Evidence of similar acts not charged is admissible in criminal actions when it is probative of the defendant’s intent to commit the crime for which he is under indictment, e. g., United States v. Calvert, 523 F.2d 895, 908 (8th Cir. 1975), and in a civil antitrust case where the burden of proof is less and there are fewer constitutional strictures, a more restrictive rule is not justified. See Kansas City Star Co. v. United States, 240 F.2d 643, 650-651 (8th Cir.), cert. denied, 354 U.S. 923, 77 S.Ct. 1381, 1 L.Ed.2d 1438 (1957). To show specific intent plaintiff introduced evidence of market allocation agreements, acquisitions of competitors and covenants not to compete, among other things. However, the greatest part of the evidence of specific intent, and that which we find persuasive, relates to pricing practices of the defendant. The evidence establishes a pattern followed by Empire in which it attempted to use price cuts or threats thereof to influence competitors’ prices or methods of competition. Empire’s Vice President of ■ Finance from 1969 to 1972 testified that around 1970 he was present at staff meetings when the company president, Robert W. Plaster, indicated that Empire’s competition should be encouraged to pass increased costs of LP supplies to the retailer on to the ultimate consumer, and sales department employees thereafter contacted Empire’s competitors for this purpose. Several of these competitors testified at trial. Exemplary of their testimony is that of W. L. Arthur, the owner of Arthur Gas and Appliances of Marshfield, Missouri. Arthur competed with Empire in the overlapping Niangua and Lebanon market areas. Mr. Rex Shaddox, one of the top officials of Empire Gas, frequently stopped in to visit with the witness or his father. During these conversations, Shaddox invariably tried to convince the Arthurs to sell their business to Empire and to set their LP prices higher. 2 At the time, Arthur was selling gas in the Lebanon area at a lower price than Empire. In late 1965, during one of these discussions of price, Shaddox told the Arthurs that Empire was too large to compete with, and could put Arthur out of *300 business. The following day Mr. Plaster, Empire’s president, called and asked that Arthur raise his LP prices. Arthur testified that when he refused Plaster said that he was going to put him out of business. A few days later Empire began doing business in Arthur’s home area of Niangua. Empire’s route salesman solicited door to door with an offer of LP gas at one and a half or two cents less than Arthur was selling for. Arthur testified he could not make a sufficient profit at this price. The Empire salesman also told those he solicited that Arthur Gas was going out of business, according to the witness. Arthur’s testimony concerning Empire’s retaliatory price cut was corroborated by Raymond E. Dore, Empire’s treasurer from June, 1963, to September, 1967. He testified that Shaddox and Plaster were upset about the competition that Arthur was giving Empire in the Lebanon area and therefore they planned to retaliate by selling low priced LP in the Niangua area where Arthur was located. Dore was instructed by Mr. Plaster to form Niangua Gas Company for this purpose, and he did so. Arthur testified that he and other companies in the Niangua area were affected by Empire’s price cut. Finally an intermediary worked out an agreement between Empire and the other gas companies which apparently included a temporary hike in the retail price of LP. In 1967 Empire acquired Arthur’s supplier. Because of the problems which Arthur had experienced with Empire, Arthur and his son met with Shaddox to see if there was a way out of the supply contract that Arthur had with Empire’s predecessor, the rights to which had been assigned to Empire. Arthur also offered to sell his company to Empire at this time. According to Arthur, Shaddox refused to cancel the contract or buy Arthur out, and told him that “we have you right where we want you.” Thereafter Empire raised the price at which it sold LP to Arthur so that he could not retail the product at a competitive price and still make a profit. Arthur also testified that Empire limited his credit to two loads of gas. This would not have been objectionable except that Empire would not mail Arthur invoices so he could pay his bills, and on this pretense they frequently refused to sell him the propane he needed. Finally, in the spring of 1967, Arthur broke his supply contract with the defendant. Other competitors testified to similar experiences with the defendant. Charles 0. Bridges operated a company in Adrian, Missouri, beginning in 1969. In the summer of 1970 an Empire district manager and plant manager called on him and said the price of LP gas in the area was going to be either 9.5 or 15.9. At the time Bridges was buying gas wholesale at 8.7 cents a gallon, and shortly thereafter his wholesale price rose to 9.7. When Bridges pointed out that a profit could not be made at a retail price of 9.5 cents the manager told him that Empire was large enough that some subsidiaries could carry others, and that Empire was selling at 9 or 9.5 in Springfield. Shortly thereafter Empire lowered its price. Bridges testified he came as close as he could to 9.5, but eventually set his price at 15.9 cents per gallon. Gary Jennings competed with Empire in the same area as Bridges, and was also contacted by the Empire district manager and plant manager in 1970. When he refused to raise his prices he was given the impression that Empire would cut their prices. At the time Jennings bought his gas from Empire at 8 or 9 cents. Shortly thereafter Empire dropped the price to its retail customers to 9 or 9.5 cents, according to Jennings’ testimony. A few weeks later he raised his price. Testimony was offered by Lloyd Geiger, who ran Gygrgas of Boonville, Missouri. In January, 1967, Mr. Shaddox called to express displeasure because Geiger was selling gas at two cents per gallon less than Empire in Jefferson City. Shaddox demanded that he come up two cents or Empire was going to play “burnout” with him. Geiger did not raise his price and two months later Empire set up a subsidiary in Boonville. The Empire company hired one of Gygrgas’ employees and solicited Geiger’s customers at a price of five cents per gallon. *301 Raymond Dore, Empire’s former treasurer, testified that Mr. Plaster, Empire’s president, had him incorporate the Boonville subsidiary to retaliate against Gygrgas for selling at a lower price in Jefferson City, just as the Niangua company had been organized to hurt Arthur Gas and Appliances. Rex Smith ran companies in Springfield and Weaubleau, Missouri. In the summer of 1970 he received three visits in the space of a month from an Empire- employee whom he believed to be defendant’s retail price coordinator. Smith was told that Mr. Plaster wanted the LP price in Springfield raised to 16 cents; at the time Smith’s price was about 14 cents and a competitor in Ash Grove was selling at 12. Because of the Ash Grove dealer, Smith did not feel he could raise his price. During the following heating season, in February, Empire brought 15 men into the Springfield area to solicit gas at 12 cents. The witness testified that when he dropped his price to 13 cents Empire lowered its LP price to 10.7 cents per gallon the next day and offered to lease customers’ tanks for nothing. At this time Smith’s LP cost was 12 cents so he could not afford to go lower than a retail price of 13 or to offer free tank leases. Mr. Smith stated that Empire’s 10.7 retail price continued from February until November, at which time Empire and he both went back to 14 cents. Several other dealers testified to threats and price cuts of the defendant which were designed to make them raise prices or stop soliciting Empire customers. The district court found that the United States had failed to show that Empire had ever threatened to reduce prices with the intent of inducing competitors to increase prices or refrain from soliciting Empire’s customers, or that it ever reduced its prices for this purpose. The court also found that Empire never attempted to enter into any price fixing arrangement. The trial judge concluded that the government had not established Empire’s specific intent by the evidence of pricing actions or otherwise. We find the district court’s conclusion in this regard to be clearly erroneous. The evidence of price fixing attempts, threats, and price reductions for the purpose of disciplining low price competitors and those who solicited defendant’s customers was not given by one witness alone. A number of witnesses testified to the existence of some or all of these practices. Two former Empire officers corroborated competitors’ testimony in certain respects. From our reading of the record these government witnesses were neither impeached nor was their testimony rebutted in any material respect. Some of the evidence is circumstantial. However, we need not rely on instances where the defendant cut its price after merely discussing prices or solicitation with a competitor. In the majority of these instances Empire’s drastic price cuts followed express threats by its officers and employees that such price cuts would result unless there was cooperation with defendant’s plans. In the face of these express declarations of intent, related by so many of the government’s witnesses, we cannot accept the district court’s conclusion that Empire’s price cutting was done for innocent competitive reasons. Nor can we agree with the district court that the instances related are not probative because the Empire officers or employees who delivered threats to its competitors were not shown to have any real or apparent authority with respect to pricing. Some of the statements relied on to show specific intent were made by Mr. Plaster himself, the president of Empire Gas Corporation. Some threats were made by Mr. Shaddox, the defendant’s senior vice president, or other vice presidents who were shown to have authority in retail pricing. It may be true, as Empire contends, that other employees, such as its plant managers or even its district managers, did not actually have price setting authority. However, in the instances where these individuals represented their employer’s wishes with regard to price or competition, or delivered ultimatums affecting these areas, the circumstantial evidence indicates they had authority to do so. This determination is obvious because in a great many instances the *302 company delivered on these men’s promises. Whether Empire’s managers had authority to set prices or were acting as mere conduits between competitors and Empire officials who did have such authority makes no difference. The circumstantial evidence establishes that lower level employees, in these instances at least, had authority to convey Empire’s wishes and threats, and the company itself followed up with retaliatory price cuts when in its judgment they were called for. Specific intent to monopolize or to obtain monopoly power may be defined for purposes of section 2 of the Sherman Act as an intent to control prices or to restrict competition unreasonably. United States v. Du Pont & Co., 351 U.S. 377, 389, 76 S.Ct. 994, 1003, 100 L.Ed. 1264, 1277 (1956). In Schine Chain Theatres, Inc. v. United States, 334 U.S. 110, 120-121, 68 S.Ct. 947, 953, 92 L.Ed. 1245, 1254 (1948), the Court recognized that price cutting in certain circumstances could violate the Sherman Act: [Pjrice cutting without more is not a violation of the Sherman Act. It is indeed a competitive practice which this record shows to have been common in the industry. It may be used in violation of the Act. Thus it may be the instrument of monopoly power to eliminate competitors or to bring them to their knees. But since it is not unlawful per se, facts and circumstances must be adduced to show that it was in purpose or effect employed as an instrument of monopoly power. (Emphasis added.) The specific intent which makes price cutting a monopolistic tool may take various forms. The record before us establishes that the defendant attempted to use price cuts or threats of price cuts to prevent competitors from soliciting Empire customers. Actions such as these, designed to prevent competitors from increasing their share of the market at the expense of the antitrust defendant, clearly show intent to monopolize. American Tobacco Co. v. United States, 328 U.S. 781, 806-807, 66 S.Ct. 1125, 1137, 90 L.Ed. 1575, 1592 (1946); see United States v. Grinned Corp., 384 U.S. 563, 570, 86 S.Ct. 1698, 1703,16 L.Ed.2d 778, 785 (1966). In addition to the above actions designed to reduce competition from other retailers, the evidence establishes that many of the defendant’s price cuts were designed to give Empire control over the retail price of LP gas. The defendant, through its agents, represented to competitors that because of its size other subsidiaries could carry those which had to resort to price cuts for some time in order to discipline uncooperative competitors. See Bergjans Farm Dairy Co. v. Sanitary Milk Producers, 241 F.Supp. 476, 484 (E.D.Mo.1965), aff'd, 368 F.2d 679 (8th Cir. 1966). Whether this representation was true or not, it seems obvious that the defendant wished to intimidate its competitors into raising their retail prices to the level that Empire thought would insure it a sufficient profit. Liberal price cuts reenforced the message. These actions taken with the object of manipulating the price of LP gas to the consumer show specific intent to monopolize, as do Empire’s attempts to control competitors’ solicitation of its customers. Price and competition are so intimately entwined that any discussion of theory must treat them as one. It is inconceivable that price could be controlled without power over competition or vice versa. United States v. Du Pont & Co., supra, 351 U.S. at 392, 76 S.Ct. at 1005, 100 L.Ed. at 1279. Accordingly, an attempt to control price, competition or both demonstrates specific intent to monopolize. We find that the government has proved both and' has thus met the burden of showing that intent here. B. Dangerous Probability of Success. This circuit follows the majority view that proof of dangerous probability of success of an attempt to monopolize must include a showing of the relevant market within which that probability occurred. Morton Buildings of Nebraska, Inc. v. Morton Buildings, Inc., 531 F.2d 910, at 918 (8th Cir. 1976); George R. Whitten, Jr., Inc. v. Paddock Pool Builders, Inc., 508 F.2d 547, 550 (1st Cir. 1974), cert. denied, 421 U.S. *303 1004, 95 S.Ct. 2407, 44 L.Ed.2d 673 (1975). There are two components in the concept of relevant market: relevant product market and relevant geographic market. The district court held that neither of these components was shown by the government at trial. 1. The Relevant Product Market. We disagree with the district court’s conclusion that there was insufficient proof to establish the relevant product market. The district court based its conclusion that the sale of LP did not constitute a distinct product market on its finding that the evidence “generally established that LP gas is functionally interchangeable in virtually every one of its common uses with such other fuels and energy sources as natural gas.” Since the government does not (nor could it) contend that Empire had a dangerous probability of success in monopolizing the sale of all fuels or energy sources, this finding alone would defeat their cause. Morton Buildings of Nebraska, Inc., supra; Acme Precision Products, Inc. v. American Alloys Corp., 484 F.2d 1237, 1244 (8th Cir. 1973). The trial judge applied an incorrect legal standard in order to determine whether LP gas sales was a relevant product market or submarket for purposes of section 2 of the Sherman Act. It is true, as he found, that other fuels will serve the same function as LP. However, the inquiry does not end there. Whether a particular product’s sales constitute a relevant market or submarket depends on the cross-elasticity of demand for that product; in other words, the readiness and ability of consumers to turn to reasonable alternatives to the product in question. United States v. Du Pont & Co., supra, 351 U.S. at 394, 76 S.Ct. at 1006, 100 L.Ed. at 1280; United States v. Grinnell Corp., supra, 384 U.S. at 571, 86 S.Ct. at 1704, 16 L.Ed.2d at 786. The functional interchangeability of other products with LP is thus germane to the inquiry. United States v. Du Pont & Co., supra, 351 U.S. at 399, 404, 76 S.Ct. at 1009, 100 L.Ed. at 1282. However, Du Pont also makes clear that such factors as price and qualities of the product must be considered. Id., 351 U.S. at 396—404, 76 S.Ct. at 1008, 100 L.Ed. at 1281. The cross-elasticity of supply would seem to be as important as the demand factor in determining relevant product market. 3 Brown Shoe Co. v United States, 370 U.S. 294, 325 n. 42, 336, 82 S.Ct. 1502, 1523, 8 L.Ed.2d 510, 535 (1962); Twin City Sportservice, Inc. v. Charles O. Finley & Co., 512 F.2d 1264, 1271 (9th Cir. 1975); Note, Telex v. IBM [510 F.2d 894 (10th Cir. 1975)]: Defining the Relevant Market, 61 Iowa L.Rev. 184, 189, 219-221 (1975). In defining the relevant part of commerce for any product the reality of the marketplace must serve as the lodestar. Brown Shoe Co. v. United States, supra, 370 U.S. at 325, 82 S.Ct. at 1502, 8 L.Ed.2d at 535; Kansas City Star Co. v. United States, 240 F.2d 643, 659-660 (8th Cir.), cert. denied, 354 U.S. 923, 77 S.Ct. 1381, 1 L.Ed.2d 1438 (1957). We agree that wood, coal, fuel oil, natural gas and electricity are all functionally interchangeable to a considerable degree with LP in some or all of its major uses. Because of the inferior qualities of wood, coal and fuel oil, however, there does not appear to be a high degree of cross-elasticity of demand between these products and LP. Although there was some testimony, mostly conjectural, that with recent rises in LP prices a few customers may have turned to these sources, most LP dealers testified that these fuels were not competing products even with the change in price. See United States v. Grinnell Corp., supra, 384 U.S. at 573-575, 86 S.Ct. at 1705, 16 L.Ed.2d at 787; International Boxing Club v. United States, 358 U.S. 242, 251-252, 79 S.Ct. 245, 250, 3 L.Ed.2d 270, 277 (1959); American Crystal Sugar Co. v. Cuban-American Sugar Co., 259 F.2d 524, 530 (2d Cir. 1958). The qualities of natural gas and electricity, on the other hand, are comparable and perhaps even superior to those of LP gas. *304 Because of the cost of constructing new natural gas pipelines and of converting LP-using facilities to electricity, however, the cross-elasticity of supply between these energy sources and LP remains low, at least for the present. See United States v. Columbia Steel Co., 334 U.S. 495, 510-511, 68 S.Ct. 1107, 1115, 92 L.Ed. 1533, 1544 (1948); Twin City Sportservice, Inc. v. Charles O. Finley & Co., supra, 512 F.2d at 1271. The evidence reveals that natural gas is slowly moving into the rural areas where LP is presently used, as is electricity. However, even with the rise in LP prices these products are not yet serious competitors with LP in these areas. Cf. United States v. Alcoa, 148 F.2d 416, 425-426 (2d Cir. 1945). Our conclusion, from examining all of the pertinent factors which make up cross-elasticity of supply and demand, is that LP retail sales is a relevant product market, or perhaps more accurately a relevant sub-market within the energy market, within the meaning of section two. Brown Shoe Co. v. United States, supra, 370 U.S. at 325, 82 S.Ct. at 1523, 8 L.Ed.2d at 535; International Boxing Club v. United States, supra, 358 U.S. at 251-252, 79 S.Ct. at 250, 3 L.Ed.2d at 277. 2. The Geographic Areas. On appeal the government has relied on evidence of dangerous probability of success in only the Lebanon and Wheaton areas. At trial the United States attempted to show that there was a dangerous probability that Empire would be successful in obtaining monopoly power in 13 different geographic areas. Seven of these, including the Wheaton and Lebanon areas, were areas designated in a report prepared by Empire which estimated where its sales were for certain bulk plants, and these seven areas were more or less contiguous. Two of the seven contiguous areas, the Lebanon and Niangua areas, overlapped each other. In their market report empire also named the companies which it thought retailed LP in each of Empire’s sales areas, and estimated the approximate percentage of the market each of these various competitors held. The report was not intended to be an admission of the relevant geographic market areas. The government altered some of the sales areas submitted by Empire and designated them as relevant geographic markets. The seven contiguous areas were designated as geographic markets without change. No reasons were given for acceptance or rejection of any of the areas, nor were the criteria used in selecting areas entered into evidence, despite repeated requests therefor by the judge. In fact, the government was unable to say who had drawn any of the lines which delineated its 13 areas. The plaintiff did attempt to learn whether there were other competitors not mentioned in Empire’s report, but did not find any in the Lebanon and Wheaton areas, at least. The sales area of a defendant is not necessarily the same as the relevant geographic market for antitrust analysis. The relevant geographic area is that in which the seller competes for distribution of his product and to which the buyer can turn for alternative sources of supply. United States v. Phillipsburg National Bank, 399 U.S. 350, 362, 90 S.Ct. 2035, 2042, 26 L.Ed.2d 658, 671 (1970); see Morton Buildings of Nebraska, Inc. v. Morton Buildings, Inc., 531 F.2d 910, at 918 (8th Cir. 1976). The parties agree that generally a retailer markets LP in an area within a 20 to 30 mile radius of his bulk plant, but some retailers testified that they sell as far as 50-60 miles away. Both the Lebanon and Wheaton areas are considerably smaller in area than a circle with a 20 mile radius. It has sometimes been held that a national company such as Empire should be considered as competing within a national market area although local outlets carried on its day to day operations. United States v. Grinnell Corp., supra, 384 U.S. at 575, 86 S.Ct. at 1706, 16 L.Ed.2d at 788; American Football League v. National Football League, 323 F.2d 124, 130 (4th Cir. 1963). We are aware that a certain amount of fuzziness is often inherent in the task of defining a relevant geographic market, and the final decision must often be a compromise. United States v. Philadelphia Na *305 tional Bank, 374 U.S. 321, 360-361, 83 S.Ct. 1715, 1739, 10 L.Ed.2d 915, 943 (1963). Nevertheless we have misgivings about the government’s method, or lack of method, of designating Wheaton and Lebanon as relevant geographic markets or submarkets. While we might not accept them in another case, we do so for purposes of this appeal in view of the result we reach on the question of dangerous probability of success. 3. Dangerous Probability of Success. The record shows that the LP gas business is highly competitive in the Lebanon and Wheaton markets as elsewhere. The defendant has many competitors wherever it does business, and new ones spring up frequently. The barriers to entry in this industry are minimal; all that are needed are a supply of LP, a truck, and perhaps a storage tank. The United States conducted a survey which purported to show that Empire had about 50% of the market in dollars’ sales in the Lebanon area and 47% in Wheaton. We agree with the district court that the survey’s reliability was extremely doubtful. 4 The questionnaire upon which it was based did not specify any records that should be used in submitting responses and the government’s witness did not know what records were used. One of the government’s witnesses who responded to the questionnaire testified that the figures he submitted came “off the top of my head.” The government made no real attempt to verify any of the figures in the responses. Data were collected for only one year, 1972. Cf. American Tobacco Co. v. United States, 328 U.S. 781, 790, 66 S.Ct. 1125, 1129, 90 L.Ed. 1575, 1583 (1946). But even if we accept the government’s survey as proof that Empire had 50% of the Lebanon market and 47% of the Wheaton market, that alone is not sufficient to show a dangerous probability of success. Hiland Dairy, Inc. v. Kroger Co., 402 F.2d 968, 974 (8th Cir. 1968), cert. denied, 395 U.S. 961, 89 S.Ct. 2096, 23 L.Ed.2d 748 (1969); United States v. Alcoa, supra, 148 F.2d at 424. It is the theory of the United States, however, that Empire’s large market shares in conjunction with its anti-competitive conduct, discussed supra, gave rise to a dangerous probability that it would be able to control prices in the Lebanon and Wheaton markets. The record, however, reveals no instance in which competitors in these areas were susceptible to Empire’s intimidation. There is insufficient proof that the competitors of Empire: (a) raised or fixed prices because of Empire’s threats; (b) stopped soliciting Empire customers, or (c) decided not to enter the LP business or decided to leave the business on account of defendant’s actions. The government relies heavily on the testimony of its expert economist to show dangerous probability. The economist testified that a greater likelihood of monopoly power existed because of Empire’s size and anti-competitive practices. He also felt that the price trends in the Lebanon and Wheaton markets were indicative of monopolistic power. In the former area he found prices to be steadily rising and at a high level and in the latter area prices were extremely stable: both patterns characteristic of monopoly, according to the economist. Finally, he found Empire’s profitability to be significantly higher than other LP businesses or American industry as a whole, which indicated monopoly power. The expert testified that by looking at industry structure, industry conduct and industry performance he could conclude that there was a dangerous probability that Empire would attain or had attained monopoly power, but he would not attempt to reach a conclusion without looking at all three factors. We have several problems with the expert’s testimony, as did the district court. *306 It was brought out that the LP companies whose performance he compared with defendant’s were not really comparable, since Empire is an undiversified company and the other two companies derived from 29 to 53 percent of their sales from industries unrelated to LP. While Empire’s profits from 1963 to 1973 averaged 11.1 percent, according to the economist, we do not regard these profits as being so extremely high that we must necessarily conclude that Empire was successful in manipulating prices or competition in the relevant market areas. The government attempts to contrast pricés and the trend of prices between Lebanon and Wheaton and other areas, such as Niangua and Springfield. We do not find these comparisons persuasive. First the government states that prices in Lebanon averaged three cents higher than those in the nearby Niangua area for the period from May 16, 1966, to mid-August 1970. A government witness testified that he found the average price for Empire’s Lebanon subsidiary over this period was 16.7 cents, while the price for its Niangua subsidiary averaged 13.7 cents. The witness testified that he reached this result by adding all of the authorized price changes made during this period and dividing that sum by the number of such prices. On cross examination it was brought out that this failed to account for the gallons sold at each price. In other words, a price of 17 cents was given the same weight as a price of 15, even though one gallon may have been sold at the former price and a thousand gallons could have been sold at the latter. We also find that no consideration was given to the length of time a given price was in effect. Therefore a price of 17.9 cents, which was in effect for two days in the Lebanon area, is given the same weight as a price of 16.9 cents which was effective for over seven months. These same errors infect graphs prepared by the government’s expert economist. These graphs purportedly show that prices in the Lebanon area were high and trending steadily upward, and that prices in the Wheaton area tended to be extremely stable. Other graphs, for the Niangua and Springfield areas, supposedly show the norm in price trends where there are no barriers to competition. There are six of these graphs in all, reflecting periods ranging from slightly more than a year for some subsidiaries to over six years for others. As previously mentioned, there is no weight attached according to the time a particular price was in effect. In order to rectify this and to compare prices in various markets during the same time periods, this court recharted the prices on a single graph with the horizontal axis reflecting the time periods during which the prices were in effect and the vertical axis depicting price per gallon. We believe that this more accurately reflects the trends in price and permits a more meaningful comparison of the Lebanon and Wheaton markets with the others. This graph reveals the following: 1. The price trend in the Wheaton area, rather than being extremely stable in comparison to the Niangua area as argued by the government, does not appear to differ appreciably as far as stability is concerned from that in Niangua. 2. The prices in the Springfield area are not reflected over a long enough period during which Lebanon and Wheaton prices are also available to be meaningfully comparable. 3. The prices of Empire’s Lebanon subsidiary did appear to be higher than in some other areas, but certainly did not average three cents over those in Niangua. In fact there was rarely three cents’ difference between the two. We do not consider the price trends reflected to be of much significance, especially since gallons sold are not considered. This would seem to be of some importance, since the testimony established that LP sales are extremely seasonal, being, as one would expect, heavy in winter and light in summer. Taking the evidence of dangerous probability of success as a whole, as is proper in an antitrust case, Sanitary Milk Producers v. Bergjans Farm Dairy, Inc., 368 F.2d 679, 691 (8th Cir. 1966), we cannot find *307 the requisite proof. The government established that: (a) Empire had a large share of the market in Lebanon and Wheaton in 1972: about 50%; (b) it has engaged in anticompetitive conduct in various areas, which has not been shown to have been effective in the Lebanon or Wheaton areas; (c) the price of LP in Lebanon may be slightly higher than in some other areas, but the reasons for this are a matter of speculation; (d) the defendant has had a fairly high rate of profit for the past ten years. We cannot conclude from this inconclusive evidence that there is a dangerous probability that the defendant will be able to monopolize the LP gas business, or exert control over prices or competition in the Lebanon or the Wheaton market areas. Therefore, we hold that the district court was not clearly erroneous when it found that the government failed to prove dangerous probability of success in its attempt to control competition or its attempt to raise prices by price intimidation. See American Tobacco Co. v. United States, supra, 328 U.S. at 784-787, 66 S.Ct. at 1126, 90 L.Ed. at 1580. II. Unreasonable Restraint of Trade. It is the United States’ position on appeal that the defendant violated section 1 of the Sherman Act, 15 U.S.C. § 1, by obtaining numerous covenants not to compete from those whose companies it acquired and from its own employees. According to the government there were 3,239 such covenants signed by employees of Empire between August 1963 and November 1971. There were also 81 different acquisition agreements which contained noncompetition agreements between August 1963 and August 1973. 5 Both the employee and the acquisition agreements vary greatly as to duration and geographic scope of the covenants not to compete. Covenants not to compete executed in conjunction with the purchase of a business allow the purchaser to obtain the value of the good will for which he has paid. Generally such covenants are valid if reasonably limited in time and geographically in order to serve this valid business objective. E. g. Goldberg v. Tri-States Theatre Corp., 126 F.2d 26, 29-32 (8th Cir. 1942); see generally Annot., 46 ALR 2d 119 (1956); Annot., 45 ALR 2d 77 (1956). Employee covenants not to compete prohibit the employee from competing after his employment is terminated. Such agreements must also be reasonably limited in area and duration, see generally Annot., 43 ALR 2d 94 (1955); Annot., 41 ALR 2d 15 (1955) and moreover, it must usually be shown that the employee was in a position to obtain valuable personal contacts or trade secrets because of his employment which might be lost to the employer if the employee entered into competition. Blake, Employee Agreements Not to Compete, 73 Harv.L.Rev. 625, 653 (1960). Covenants of these two types have not generally been considered violative of the antitrust laws. United States v. Addyston Pipe & Steel Co., 85 F. 271, 281 (6th Cir. 1898), modified, 175 U.S. 211, 20 S.Ct. 96, 44 L.Ed. 136 (1899). However, in Schine Chain Theatres, Inc. v. United States, 334 U.S. 110, 119, 68 S.Ct. 947, 952, 92 L.Ed. 1245, 1253 (1948), the Court recognized that such agreements, even if legal under local law, could be unlawfully used in restraint of trade. The record in this case reflects that the successful retail sale of LP depends to a considerable extent on the route salesman who delivers the product by truck. In some instances, according to the testimony, a company has been able to obtain a considerable number of its rival’s customers by the simple expedient of hiring its rival’s route man, who knows where and who the customers are and has developed a personal relationship with them. Most of the employee covenants not to compete about which there is testimony in the record appear to concern employees of this nature or *308 those with access to customer lists. 6 The government asserts, however, that many of the employees who signed noncompetitive covenants did not occupy a position which gave them customer contact or access to confidential information, and thus those covenants were unreasonable restraints of trade. The government further argues that the employee covenants and the acquisition covenants are unnecessarily restrictive in time and area to protect Empire’s legitimate business interests. Finally they contend that the practice of obtaining agreements from such a large number of potential competitors constitutes a restraint of trade in itself regardless of the validity of individual contracts. The difficulty with the government’s position is that there is no showing that any restraint was unreasonable so as to violate the Sherman Act. Counsel for the United States apparently believes that the burden is on Empire to establish the reasonableness of each of the more than 3,000 contracts with their varying terms. However, the burden of showing unreasonableness of a restraint of trade, except where there is a per se violation of the Act, is on the plaintiff. Alders v. AFA Corp., 353 F.Supp. 654, 657-658 (S.D.Fla.1973), aff’d mem., 490 F.2d 990 (5th Cir. 1974); see United States v. Arnold, Schwinn & Co., 388 U.S. 365, 374 n. 5, 87 S.Ct. 1856, 1863, 18 L.Ed.2d 1249, 1257 (1967). In addition the record is barren as to the actual effect of these covenants on competition in the LP retail market. White Motor Co. v. United States, 372 U.S. 253, 261-263, 83 S.Ct. 696, 700, 9 L.Ed.2d 738, 745 (1963); Board of Trade v. United States, 246 U.S. 231, 238, 38 S.Ct. 242, 243, 62 L.Ed. 683, 687 (1918). We agree with the district court’s conclusion that the mere existence of a large number of covenants not to compete does not establish a 15 U.S.C. § 1 violation. III. Conclusion. We conclude that the district court was not clearly erroneous in holding that the government failed to prove a dangerous probability of success in an attempt to monopolize under section 2 of the Sherman Act and that it correctly determined that there was insufficient proof to show an unreasonable restraint of trade under section 1. Accordingly the judgment of the district court is affirmed. Footnotes [1] Allegations of threats and attempts to destroy the business or property of competitors, price fixing, and reciprocal purchasing arrangements have been abandoned on appeal. The destruction of property allegations were the subject of an earlier criminal trial in which Empire, its president and the president of two private security firms were acquitted by a jury verdict. [2] Mr. Shaddox died before trial. Empire contended that testimony concerning conversations with him were inadmissible under the Missouri dead man’s statute, V.A.M.S. § 491.-010. This statute did not disqualify the witnesses who testified to statements of Shaddox because none of them were parties to this action. Hunter v. Norton, 412 S.W.2d 163, 165 (Mo. 1967); Galemore v. Haley, 471 S.W.2d 518, 521, 522 n. 2 (Mo.Ct.App.1971). Since this case was governed by former Fed.R.Civ.P. 43(a), competency of the witnesses under Missouri law qualified them to testify in federal court. 5 Moore’s Federal Practice 43-70-43-71 (2d ed. 1976). [3] This is especially relevant in comparing LP gas to natural gas. Natural gas is in greater demand but in the relevant geographic market areas is not readily available. [4] The government disregarded the court’s order that the questionnaire to be used be submitted for its approval and that the court’s subpoena power would be used in securing responses, and sent out its own questionnaire without the court’s knowledge. The reason for following this procedure is far from clear, but it appears to us that the government would have benefited in many respects if it had been more cooperative with the court. It is regrettable that this dispositive question must be determined on an inadequate and incomplete record. [5] Many of the acquisition covenants required the seller to obtain other covenants from stockholders, employees, or family members of the seller. Therefore we do not know how many covenants pursuant to Empire’s purchases of other businesses are involved. [6] In addition to route salesmen and managerial personnel the government introduced evidence concerning Mrs. Sarah Ann Guy, a clerical employee who was sued for enforcement of a covenant after she left Empire. Her testimony revealed that she had worked with defendant’s accounts receivable during her employment and afterwards she went to work for an accountant who did work for various LP businesses. The lawsuit was settled when Mrs. Guy stipulated that she would not solicit Empire business or customers as of the date she was terminated. It is reasonable to assume that by the nature of her work with accounts receivable Mrs. Guy would have access to customer lists. Many of the acquisition covenants explicitly required that the seller not reveal such lists to Empire competitors.