Henry Floyd “Red” Box was convicted by a jury of violating 18 U.S.C. § 1955, the federal antigambling statute. On appeal, Box argues that the evidence was insufficient to support this verdict. We agree and therefore reverse the conviction.
Federal agents conducted an extensive investigation of several bookmaking operations in the Shreveport-Bossier City area during the 1973 football season, culminating in simultaneous raids on the last day of the season. A one-count indictment filed on April 25, 1974, charged appellant Box and ten other persons with the operation of an illegal gambling business in violation of 18 U.S.C. § 1955.1 The indictment named three unindicted principals as having been involved in the same illegal gambling business. One of the defendants was granted a continuance and severance, due to the death of his counsel. Six others entered pleas of nolo contendere or guilty prior to trial. Trial of the four remaining defendants began on September 30, 1974. The guilty plea of one of these was accepted on October 4, 1974. Later the same day the jury returned a verdict of guilty as to Box and the other two. Only Box has appealed.
Our review of the evidence and application of the law in this case require an understanding of the general nature of a bookmaking operation, and so we preface our consideration of the issues here with a very brief summary on that subject.
THE NATURE OF A BOOKMAKING OPERATION
This section might be subtitled, “How to Succeed in Gambling Without Really Gambling,” because a successful bookmaker makes his profit not from winning bets, but rather from collecting a certain percentage of the amount bet that losing bettors are required to pay for the privilege of betting.2 This percentage, 10% in the Shreveport area, is called “juice” or “viggerish,” and its effect is to require a bettor to risk $110 in *1261an attempt to win $100. So that betting odds can remain even on each game, a bookmaker normally has a “line” — on each game on which he is taking bets, one team will be favored by a certain number of points, called the “point spread.”3
In an ideal situation, a bookmaker would have bets from bettors exactly balanced on each contest, so that no matter which team “wins” (read: beats the point spread), the bookmaker is assured a definite percentage of the amount bet. That is, he would collect 110% of the amount he would be required to pay. With a multitude of bets each week, this ideal of perfectly balanced books cannot be achieved. When the bets placed with a bookmaker on a certain contest become very unbalanced on one side, however, there are certain measures the bookmaker might take to lessen the incumbent risk.4 He can refuse to take further bets on that side, hoping enough bets will be placed on the other side to effect some rough balance. Alternatively, he can adjust his “line” on the contest, thus making the underbet side more attractive.5
Another common solution to the bookmaker’s problem of grossly unbalanced bets on a game is the “lay off” bet. By this device, a bookmaker whose customers had bet $10,000 on Dallas + 6 and only $6000 on Pittsburg —6 would himself seek to make a $4000 bet on Dallas + 6 with another individual.6 This bet would have the effect of “laying off” $4000 of the $10,000 the bookmaker’s customers had bet on Dallas, leaving the bookmaker in the net position of having $6000 bet with him on each side. Normally, the bookmaker would look to another bookmaker to make this bet, and would be required to give up the same favorable 11 to 10 odds which he had received from the Dallas bettors. Indeed, several cases dealing with § 1955 have in dicta defined a lay off bet as a “bet between bookmakers.”7 It seems clear, however, that the individual accepting a lay off bet from a bookmaker need not be another bookmaker.8 That individual could be part of a professional “lay off” operation, an organization dealing only with bookmakers rather than with retail customers, and having sufficient capital so that risk-taking at 11 to *126210 odds posed little problem.. On the other hand, the individual could be a mere bettor who wanted to bet $4000 on Dallas + 6, but was told by his bookmaker that no more such bets were being taken and was invited by the bookmaker to accept instead a wager in which the bettor received 11 to 10 odds for agreeing to bet on Pittsburgh. The point of all this is that a “lay off” bet should be defined solely in relation to the occupation and the purpose of the person making the bet — the occupation and motives of the person accepting the bet are irrelevant to the definition.
We do not warrant the foregoing as constituting all the structural information a lay person (as distinct from a lay off person) would need to organize his or her own business, but we think it sufficient for our purposes, and we turn now to the case before us.
THE EVIDENCE RELATING TO BOX
During this five day trial, twenty-one witnesses testified and several kilograms of evidence were introduced. The testimony of the only four witnesses who had any knowledge concerning Box may be summarized as follows:
F.B.I. Agent Beinner testified that Lombardino, a bookmaker, visited the Guys & Dolls Billiard Parlor, an establishment owned by Box, on three separate Tuesdays during the 1973 football season. Beinner believed Tuesday to be “payoff day” in the bookmaking operations he had been investigating. Bein-ner had obtained and executed search warrants on the homes or places of business of eight of the defendants, but had been unsuccessful in his attempt to obtain a warrant on the home and place of business of Box. Beinner’s principal informant, whose information was the basis for the search warrant affidavit, described the other defendants who were named in the affidavit as “bookmakers” and described Box only as a “bettor.”9 It was through- the testimony of Agent Beinner that the government introduced the telephone toll records, discussed below.
Messina, a bookmaker who had been granted immunity by the government in return for his testimony, testified that he himself had never “laid off” bets to Box, but that he had personal knowledge that Cook had done so.
Cook, a bookmaker also given immunity, testified that he had occasionally “laid off” bets with Box and with several of the other defendants. Cook explained that when he lost such a bet to Box or one of the others, he would pay the winner an extra 10% in excess of the amount bet. Cook testified that Box, as a customer, also placed bets with Cook in which Cook received this 10% advantage. Cook did not consider Box a bookmaker and knew of no one who did. He related that Box had been free to take or reject bets offered by Cook, and he described Box only as a bettor.
Stewart, a bookmaker, testified that Box was one of his customers, i. e., a bettor. No one asked Stewart the direct question, “Is Box a bookmaker?”, but the prosecutor asked that question of Stewart concerning every other defendant remaining on trial when Stewart testified, and received an affirmative answer in each case. Stewart testified that Box placed bets with him, and that he (Stewart) placed bets for Box with other bookmakers. It was through Stewart that the betting slip testimony was introduced. Stewart testified that he made bets with two other bookmakers in which he gave the others 11 to 10 odds— some of these were “lay off” b'ets, and some were bets Stewart made because he liked the team. Stewart did not testify that he ever made such bets with Box.
The two items of documentary evidence which related to Box were as follows:
The Telephone Toll Records. No wiretaps or pen registers were used in this *1263case, but the Government introduced at trial several long distance telephone records, including those of the telephone at Box’s house and the telephone at Guys & Dolls, Box’s establishment. These records showed that during the period of the investigation (autumn, 1973), 20 calls were made from Box’s home and 223 calls were made from Guys & Dolls to one Price, a Baton Rouge bookmaker.
The Betting Slips. The simultaneous raids conducted on the last day of the 1973 football season yielded, inter alia, large numbers of betting slips which had been used in the Stewart operation. Most of these slips were marked in a similar simple manner, e. g., G.B. + 14 = 200 (translated, the bettor had wagered $200 that the score of Green Bay plus fourteen points would be greater than that of Green Bay’s opponent). In the lower right hand corner a name, a set of initials, or a number would appear, indicating the individual making the bet. Finally, an indication of the result was added, e. g., “ + 200” (the bettor won), or “-220” (the bettor lost and was required to pay the additional 10%).
A smaller number of these slips were marked in a second, distinct, fashion, e. g., G.B. + 14 = 330/300. On these, the results would be recorded as + 330 or -300. The testimony of Stewart on this point was quite confused, but it could be inferred that the slips marked in this second fashion represented bets in which he was giving 11 to 10 odds to the person with whom he was betting. Of the five individuals whose names or initials appeared on Stewart’s slips marked in this second fashion, four were clearly bookmakers. The fifth was Box. The seized slips represented about $230,000 of Stewart bets and approximately $3800 of this amount was comprised of slips labeled “Box” and marked “330/300”, “550/500”, or the like.
STANDARDS FOR SUFFICIENCY
In reviewing the evidence upon which the jury based its verdict of guilty, we of course examine the evidence in the light most favorable to the government. Glasser v. United States, 1942, 315 U.S. 60, 62 S.Ct. 457, 86 L.Ed. 680; United States v. Warner, 5 Cir. 1971, 441 F.2d 821. When the conviction is based upon circumstantial evidence, our question becomes whether the jury could reasonably conclude that the evidence excluded all reasonable hypotheses of innocence. United States v. Gomez-Rojas, 5 Cir. 1975, 507 F.2d 1213, 1221; United States v. Squella-Avendano, 5 Cir. 1973, 478 F.2d 433, 436.
WAS BOX A BOOKMAKER?
If we were to find that the jury could reasonably conclude that Box was a bookmaker (engaged in a business with the other defendants), our analytical task would be at an end, for the statute in express terms covers bookmakers.10 Even viewing the evidence most favorably to the Government, however, we are convinced that the jury could not reasonably reach such a conclusion. This evidence must be regarded as consistent with the hypothesis that Box was not a bookmaker.
The only direct testimony on this matter clearly categorizes Box as a bettor rather than a bookmaker. Of course, the jury might not have credited this testimony, although we note that Cook and Stewart had no hesitation in labeling the other defendants as “bookmakers”. The fact remains that there is no evidence in this record upon which an opposite conclusion, i. e., that Box was a bookmaker, could be based. Bookmakers have customers. The names of over 150 bettors were seized during the raids, numerous bettors were interviewed by the FBI, and bettors who were customers of each of the other defendants on trial testified, but no evidence was introduced relating to any “customers” Box might have.
The testimony of Cook and the betting slips of Stewart indicate that *1264Box on occasion accepted “lay off” bets from two bookmakers.11 The Government argues that since a lay off bet must be defined as a bet between two bookmakers, Box was a bookmaker simply because he accepted lay off bets. As explained above, we reject the premise of this argument — a lay off bet is one placed by a bookmaker, but the individual accepting the bet need not be a bookmaker.
An additional characteristic of a bookmaker is that she distributes a “line.” There is no testimony that Box ever distributed a line, either to customers or to bookmakers. Finally, we note the calls made from Box’s telephones to Price. Assuming the jury could conclude that Box himself made all 223 calls to Price from the Guys & Dolls phone, it cannot be said that this number of calls in that direction is inconsistent with the hypothesis that Box was merely a heavy bettor, placing bets with Price.
§ 1955 AND NONBOOKMAKERS
Having established that Box cannot be labeled a bookmaker, we have not yet shown him to be within an unassailable hypothesis of innocence, because § 1955 clearly was meant to proscribe some bookmaking-related activities of individuals who were not themselves bookmakers. The legislative history indicates that § 1955
applies generally to persons who participate in the ownership, management, or conduct of an illegal gambling business. The term “conducts” refers both to high level bosses and street level employees.12
This reflects an intent to reach employees of large bookmaking operations, and that intent has been followed in cases affirming § 1955 convictions of runners, telephone clerks, salesmen, and watchmen.13 On the other hand, individuals who are only bettors or customers of bookmakers clearly are not within the scope of the statute.14 The case before us cannot be fit easily into either of these two categories. No evidence supports the theory that Box was an employee of other bookmakers; yet, Box’s acceptance of lay off bets arguably makes him more important to the operation of a bookmaking business than would be a mere customer. Our question, then, is in what circumstances can an individual who accepted lay off bets from bookmakers be convicted under § 1955? The language of the statute does not resolve this, so we turn again to the legislative history.
Clearly, the dominant concern motivating Congress to enact § 1955 was that large-scale gambling operations in this country have been closely intertwined with large-scale organized crime, and indeed may have provided the bulk of the capital needed to finance the operations of organized crime.15 The target of the statute was large-scale gambling operations — local “mom and pop” bookmaking *1265operations were to be left to state law.16 In this connection, the requirements of dollar volume ($2000 gross on any day) or duration (30 days or more), and number of participants (5), were drafted into the legislation.17 These requirements are such that relative small-fry can conceivably be ensnared in the statutory strictures,18 but apparently Congress was of the opinion that the size of gambling operations was often much larger than could be proved, and that law enforcement officials needed some flexibility in order effectively to combat the large-scale operations.19
There are indications in the legislative history of a concern that one way in which large-scale organized crime profited from bookmaking operations was to act as a regular market for lay off bets from local bookmakers.20 Remarks of supporters of the bill demonstrate that the Congress was aware of the general function of lay off betting. For example, Senator McClellan stated:21
[describing a lottery operation] The gambler thus seldom gambles. In addition he hedges his bet by a complicated layoff system. .
[A bookmaker] has at least the virtue of exploiting primarily those who can afford it. Yet he seldom gambles either. He gives track odds or less without track expenses, pays no taxes, is invariably better capitalized or “lays off” a certain percentage of his bets with other gamblers
Nothing in the legislative history, however, deals with the question of whether the recipient of a lay off bet, on that basis alone, should be convicted under the statute.22
The phenomenon of lay off betting has been a factor in a large number of cases which have construed § 1955.23 In almost every case, the question has been whether the exchange of lay off bets, usually in addition to the exchange of line information, could be enough to link two separate bookmaking operations into one business for the purposes of meeting the § 1955 jurisdictional requirement of five participants in one *1266business.24 The answer has in every case been affirmative — the regular direct exchange of lay off bets and line information can connect otherwise independent gambling operations, which alone would be illegal under state but not federal law (because less than five participants were involved), into one business. Further, the case law supported by legislative history establishes that an individual who is in the business of providing a regular market for a large volume of lay off bets should also be considered to be part of the gambling operation he services.25 Finally, it seems clear that, at least in this circuit, a professional gambler who accepts bets in the nature of lay off bets and, additionally, provides line information to the same bookmaking operation can be convicted as part of that operation under § 1955.26
The cases establish, then, that one who accepts lay off bets can be eon-victed if any of the following factors is also present: evidence that the individual provided a regular market for a high volume of such bets, or held himself out to be available for such bets whenever bookmakers needed to make them; evidence that the individual performed any other substantial service for the bookmaker’s operation, as, for example, in the supply of line information; or evidence that the individual was conducting his own illegal gambling operation and was regularly exchanging lay off bets with the other bookmakers. Our review of the legislative history, and our adherence to the doctrine that statutes mandating penal sanctions are to be strictly construed, convinces us that one of the listed factors, or other evidence that the defendant was an integral part of the bookmaking business, is necessary before an individual who accepts lay off bets can be convicted under the statute.27 *1267Evidence establishing only that a person received occasional lay off bets from bookmakers cannot be considered inconsistent with the possibility that the individual was for all practical purposes only a bettor.28
In these circumstances, we do not feel that the cases finding “lay off bettors” within the scope of § 1955 are disposi-tive.29 If dicta in these cases can be read to indicate that a “lay off bettor,” as the recipient of a lay off bet, is on that basis alone a part of an illegal gambling operation, we reject such dicta as being based on an erroneous assumption regarding the nature of lay off betting. We stress again that the recipient of a lay off bet need not be a bookmaker, but rather might be any individual willing to accept a single bet. § 1955 was directed at the professionals — the persons who avoided gambling themselves, but profited from the gambling of others.30 Although a heavy bettor might be a crucial source of revenue for a bookmaking operation, the statute was meant to exclude bettors. Gambling becomes a federal case only when a person is charged with more than betting, and evidence that a person accepted lay off bets, without more, is insufficient to expel that person from § 1955’s sanctuary of bet-tordom.
[H] The question remaining, then, is whether the evidence relating to Box, viewed most favorably to the Government, could sustain a jury finding that one of the additional factors noted above was present in this case. Such a jury finding would in effect be a conclusion that the evidence was inconsistent with any hypothesis of innocence. In reaching this conclusion, of course, the jury is limited to evidence in the record and supportable inferences therefrom. If a conclusion that all hypotheses of innocence have been excluded by the evidence could be reached only as a result of speculation or assumptions about matters not in evidence, then the jury verdict must be overturned.
The evidence against Box shows that he accepted lay off bets of undetermined amounts from Cook on a number of occasions, and that in one week he may have accepted $3800 in lay off bets from Stewart. These are the only two pieces of evidence which distinguish Box in any way from the “mere bettor” so clearly excluded from the statute’s scope. We do not find any reasonable basis in the evidence upon which the jury could conclude that Box was an integral part of these bookmaking operations. While the volume of bets with Stewart was substantial, no evidence indicates that Box regularly accepted lay off bets from Stewart. There is no evidence on amounts from Cook, and while Cook’s testimony could support a conclusion that Box accepted lay off bets on several occasions, that testimony flatly contradicts any suggestion that Box held himself out to be a regular market for such bets upon which local bookmakers could *1268depend.31 As we have already noted, no evidence supports the suggestion that Box was himself a bookmaker, or that he provided line or other gambling information to bookmakers.
Box may have gamboled with the gamblers, but he has not been shown to be a gaming entrepreneur. Nothing indicates that he solicited the lay off bets that he accepted. Box was a customer of bookmakers and was perhaps a bargain-seeking bettor, but the record does not permit him to be cast in a role as a necessary or integral part of a gambling operation. The testimony of admitted bookmakers, the multiplicity of phone calls and the shower of betting slips suggest only that Box bet with continuity and in magnitude, and on occasion received a discount when the professionals with whom he dealt needed to lay off a bet. We conclude, then, that the jury could not reasonably find the evidence inconsistent with the hypothesis that Box was simply a heavy bettor who on occasion received favorable odds in bets with bookmakers.32 For purposes of § 1955, this hypothesis is one of innocence.
Since we thus have concluded that the evidence in this case was insufficient to support a verdict of guilty, we need not reach any of the other eleven points argued by appellant.33 The conviction of Box is reversed, the sentence is vacated, and the ease is remanded to the district court for entry of a judgment of acquittal.
REVERSED.