L. HAND, Circuit Judge.
The defendant, Costello, appeals from a judgment entered upon the verdict of a jury, fi tiding him guilty upon three of four cour.ts in an indictment under § 145 (b) of the Internal Revenue Code, Title 26, U.S.Code: i. e., of wilful attempts “to evade: or defeat a large part of the income tax” for the years 1947, 1948 and 1949, “due and owing by him and his wife,” by understating their joint net income tas.. (The jury acquitted him on the first count — -1946—, which was for understating his separate net income tax.) Judge McGohey imposed a sentence of five years upon each of the three counts for 1947, 1948 and 1949 (to be served concurrently), and a fine of $10,000 on each count cumulatively, together with the costs of the prosecution —$4,111.38. On the appeal Costello raises six points, which we shall consider seriatim. First, he challenges the sufficiency of the evidence to prove that he and his wife had received more taxable income in the three years in question than he included in their joint tax returns for those years. Second, he asserts that it was error to refuse his offer in evidence of tax assessments against him for the years 1941-1945, which showed a higher net income received by him than the prosecution computed as his gross income for those years; and which were therefore relevant to show that he might have laid up a cash reserve on January 1, 1946. Third, that it was error to include his wife’s expenditures as part of his own. Fourth, that the court erred in the admission of various pieces of evidence offered by the prosecution. Fifth, that the judge’s charge to the jury was insufficient. Sixth, that the indictment should have been dismissed because no competent evidence was before the grand jury at the inquest.
The first question is the most important. The prosecution built up its case upon what has come to be known as the “net-worth method,” which the Supreme Court has very recently accepted as permissible, though it must be applied with the greatest caution.1 This method presupposes that the prosecution first proves what property the taxpayer had at the beginning of the year in question and what he had at the end of it. To the remainder obtained by subtracting the first from the second it adds whatever sums it can prove that he spent in the year in question. That is the putative gross income for the year; and the remainder, after deducting the amount of gross income reported, is by hypothesis the unreported gross income. However, this is not enough, for it does not follow that all that the taxpayer expended was necessarily taxable income, or indeed income of any kind. Conceding something for the difficulty of establishing by impreg[671]*671nable proof how much was income, the Court is satisfied with “proof of a likely source, from which the jury could reasonably find that the net worth increases sprang”. True, a “likely source” may not be the true source, so that it is necessary in addition to exclude the possibility that what he received did not come from gifts, inheritances or loans. Here too the Court does not exact precision, for, “where relevant leads are not forthcoming, the Government is not required to negate every possible source of nontaxable income, a matter peculiarly within the knowledge of the defendant.” Nevertheless, in the end the prosecution must “prove every element of the offense beyond a reasonable doubt though not to a mathematical certainty.” “The settled standards of the criminal law are applicable to net worth cases just as to prosecutions for other crimes. Once the Government has established its case, the defendant remains quiet at his peril. * * * The practical disadvantages to the taxpayer are lessened by the pressures on the Government to check and negate relevant leads.” Finally, it should be remembered that, as in all criminal prosecutions, the prosecution makes out a sufficient case to go to the jury, if the evidence would have been enough in a civil action; the only difference between the two is that in the end the evidence must satisfy the jury beyond any reasonable doubt.2
The prosecution’s proof started with a supposed “net worth” at the beginning of the year 1946, made up of four items which, less liabilities, aggregate $250,-000 3, and among which there is no item of cash on hand. Concededly the “net worth” at the beginning of each year would be falsified to the extent that any such sum was omitted; and with it would fall the computations for later years. However, we will disregard this possibility for the time being; and proceed on the assumption that the figure for “net worth” was accurate. Next, the prosecution proved that either Costello or his wife made large purchases in each of the four “indictment years,” varying from nearly $60,000 in 1948 to over $90,-000 in 1949. Then it proved the “net worth” at the beginning of 1947, and subtracted it from that at the beginning of 1946, which resulted in a minus quantity ($8500). This it deducted from the sum of the purchases; and the difference between what remained and the amount returned as gross income by Costello it asserted to have been fraudulently concealed. He raises a number of objections to this computation in addition to the omission of a concealed cash reserve. One of these is the failure to show a “likely source, from which the jury could reasonably find that the net worth increases sprang.” There was no difficulty in the case at bar in pointing to such a source. By his own admission Costello was a gambler, though he had no other occupation. He had substantial interests in “slot machines” and “juke boxes” in Louisiana companies; and these were not his only ventures, for he gambled in “horses, cards and fights”; and there was evidence that he received $30,000 for keeping bookmakers away from a race tract for two years; which of itself showed him to have been a man of power[672]*672ful undisclosed influence. Gambling is an occupation with indeterminate possibilities that might well have brought in more than $100,000 in a single year — the highest with which he was charged in the four “indictment years.” If there were no accumulated reserve and no loans, gifts or inheritances, gambling and such payments as the $30,000 we have mentioned, were a “likely source” of the money that he spent during the year. To exclude the possibility that loans, gifts and inheritances might have contributed, the prosec.ution searched all likely records for inheri tances and for gifts and found none. Similarly, it found no credits on the returns of either Costello or his wife for the payment of interest on loans. The sam€ method was followed for each of the “indictment years”; and it fulfilled the tests laid down in Holland v. United Spates, supra. Upon any issue it is nearly always true that the party having the affirmative does not, and cannot, shut the door against every possible exception; and it would be idle to deny that there were no possible exceptions on the present occasion; but we cannot see how it can be doubted that a man, having no resources in loans, gifts or inheritances, could have spent what Costello did in these four years, unless it was out of his income or his wife’s, or unless he had a cash reserve accumulated from past years. As to his wife’s income, the evidence justified a finding that the money had not come out of it, except in 1946 whan she was credited with about $16,000 gross income, which she separately returned. In 1937 Costello had sworn that she had no income; in 1939 that, whatever money went into her bank acc.ount, he gave her; and in 1943 he told his lawyer that in 1941 she bought a motor car with his money, as well as “all the living expenses.” In applying for insu ranee in 1940 she stated that she was supported by her husband. Furthermore, the prosecution traced many cheques received by him into her bank account and in 1943 two payments of Costello’s estimated income tax were paid out of her account. From all this it was a permissible inference that in the four “indictment years” she had no separate income beyond what was credited to her in 1946.
Thus, the issue is narrowed to whether Costello had an accumulated cash reserve at the beginning of 1946, out of which the purchases might have come that were shown to have been made, and not declared. Since the aggregate of the net income omitted during the “indictment years,” when computed as we have described, came to more than $100,000, perhaps it might have been fair to start with an assumption that no such hypothetical reserve covered the unreported income for at least the year 1949. However, the prosecution did not rely on such an assumption; it undertook to establish that Costello had no reserve whatever on January 1, 1946. It started with a statement of his “net worth” on October 18, 1937, the day of a sworn statement made by him to an official of the Tax Bureau. On that day he said that he had a cash reserve of between $25,000 and $30,000 which he kept in currency. His bank deposits, investments and receivables, when added to this, made $45,600, from which was deducted loans of $6,000. Thus, he started with a “net worth” on that day of substantially $40,000. As we have said, on January 1, 1946, a little over eight years later, this figure had become $240,000, making a gain of $200,-000. To this the prosecution added purchases of $312,000, making $512,000; and, after deducting $6,400 of non-taxable receipts and $302,500 of gross income declared, it asserted an unreported difference of $203,000. In addition it proved, as it had for the four “indictment years,” that there was no record of any loans, gifts and inheritances.
In order to meet the possibility that Costello might have laid aside out of this a substantial cash reserve, it proved that he had allowed three judgments to remain unpaid, that he left a mortgage outstanding on his house; that he had allowed a loan on his life insurance to remain unpaid until 1946 when he paid less than half of it, and that he had borrowed [673]*673$25,000 of a friend in 1946 and the same amount again in 1947. Finally, it proved that on February 21, 1947, he paid over $21,000 as penalty and interest upon taxes that had remained unpaid since 1930 and earlier. It must be owned that the combination of these facts makes it unlikely that there was any reserve left on January 1, 1946; and we should not hesitate so to assume, were it proper to make any inference against an accused from his failure to answer. Nevertheless, a man in Costello’s position, whose affairs, as we can see, varied greatly up and down, might have allowed interest to run up, confident that in the end a turn would come; and have kept on hand a substantial cash reserve, prepared for a run of bad luck that would demand immediate calls for large payments. His situation was far from ordinary; and, as we have said, in the past he had always kept such reserves in large amounts. Considering the admonition with which the Supreme Court has hedged about the “net worth method,” we cannot therefore agree that the evidence, taken as a whole, justifies the inference that he had no reserve on January 1, 1946.
On the other hand, the jury was warranted in setting a limit upon its possible size; and it is quite likely that this is exactly what they did do, when they voted to acquit on the first count. Costello’s net income unreported for 1946 was only $1,541, if we allow no cash reserve; that for 1947 was $36,730; and, if the jury assumed a reserve of $30,000 (the figure he gave in 1937), it would still be true that he would have understated the joint income for 1947. On the other hand, since his understatement of the joint income for 1948 was $35,245, and the aggregate understatement for 1946, 1947 and 1948 was therefore $73,524, it would have taken a reserve on January 1, 1946 of that size to deprive the verdict on the third count of all support in the record. Nothing suggests any reason why on January 1, 1946, he should have nearly doubled the maximum of $40,000 which had been enough six years earlier, especially when we remember the heavy liabilities that were rolling up in interest and penalties. In deference to the limitations imposed upon any use of the “net worth” method, we feel obliged to say that the evidence did not justify a verdict based upon the assumption that on January 1,1946, there had not been a reserve of more than $30,000; or indeed of more than $40,000. On the other hand we also hold that, taking the evidence as a whole, it would be quixotic for a jury to assume that on that day Costello had had a reserve of over $73,000. It follows that while there was not enough evidence to support a verdict on the second count there was enough to do so on the third and fourth counts.
In view of what we have just said about the reserve, it is hardly necessary to discuss whether it was before or after October 18, that Costello received $66,-000 which he returned as received in 1937 from Kastel’s share in the Bayou Novelty Company; but in any event there was evidence enough to find that he did receive it before October 18. The Bayou company was a partnership between one, Kastel, and another man who died on June 19, 1937; and Costello had a half interest in Kastel’s share, which turned out to be about $132,000. Although Costello’s return does not show whether he received his half before or after October 18, all of Kastel’s share was in Kas-tel’s hands by June 26, because that was the balance of his account on that day after deducting that sum; and Costello testified that Kastel remitted to him “weekly, sometimes monthly.” This established, prima facie, that Costello had received before October 18 the share that he entered in his income tax return.
The second point raised is the judge’s refusal to allow Costello to put in evidence assessments by the Tax Bureau of deficiencies in his net income tax as reported for the years 1941-1945. These he offered to meet the prosecution’s evidence that he had no concealed cash reserve on January 1, 1946. They showed, he argues, that his “net income,” as “corrected,” had been larger by $292,000 than what he had reported, which was [674]*674$90,000 more of unreported net income in the five years, 1941-1945, than, according to the prosecution’s computation, he had had of unreported gross income in the eight years, 1938-1945. From this, he argues, the jury might have found that he had had, or at least might have had, a concealed cash reserve on January 1, 1946, which would have been enough to meet the purchases during the “indictment years.” Judge McGohey heard prolonged argument as to the relevancy of these assessments, and finally excluded them because he thought that, unaccompanied as they were by any explanation, they were more likely to mislead than to enlighten the jury. Costello did not suggest that he meant to supplement the assessments by showing how much of the alleged unreported net income had been spent for consumable purchases, or how much remained, or might have remained, as a concealed cash reserve on January 1, 1946. He stood upon their admissibility as they read.
We shall assume, arguendo, that the assessments were competent evidence on this trial that Costello had received the net income that they charged him wilh failing to report. We do not indeed agree that they were “judgments” which estopped the prosecution; for they were net.4 It has been recognized practice for centuries to levy on tax assessments 5; but the taxpayer has always been allowed to challenge their validity in one way or another, or to recover his money from the collector. Nevertheless, the ass assments rested upon findings of the Taa Bureau made under its statutory authority, and we will assume, again arguendo, that as such, they were competent as admissions against the prosecution in a criminal trial, although the point is not free from doubt. If they were competent, they were relevant to show that Costello had received $90,000 of unreported net income more than the prosecution had succeeded in proving that he had received of unreported gross income during the eight years before January 1, 1946. That could, indeed, have been one link in a chain of reasoning to prove that he might have had a cash reserve in his hands on January 1, 1946; but it would be only a link, and it proved nothing, unless accompanied by evidence as to what had been done with the income so received. So far as Costello or his wife might have spent it for consumable goods, or have otherwise disposed of it, it could not be part of a cash reserve; and there was no presumption that they had not so spent it. It is horn-book law that, although a party may offer evidence rationally relevant to the issue, that does not inevitably make it admissible. It may, just as Judge McGohey thought in this instance, tend rather to mislead than to enlighten the jury.6 When the prosecution showed the amount of his gross income during the eight years, it did so by proving in detail the purposes for which he spent his money, and thus supplied, just the element that was missing from the assessments. It was certainly within the judge’s discretion to decide whether to admit them until this omission was supplied; and we are in entire accord with his ruling.
The third point is that it was error to allow the jury to include Mrs. Costello’s purchases as having been made out of Costello's money. We have already stated the evidence that supports such an inference; and we think that it was sufficient, standing uncontradicted as it did. In this connection the question arises whether it was an added ground for this conclusion that she was not called as a witness for the defence, a wife now being a competent witness in her husband’s [675]*675favor.7 In a case where the wife was not herself concerned in the transaction that was being prosecuted as a crime, we held that it was permissible to infer from the husband’s failure to call her that her evidence would not be favorable ;8 but when, as here, she was a party to the transaction and might believe herself to be subject to prosecution, the inference becomes much weaker, if indeed it does not altogether disappear. In that situation she might wish to claim a personal privilege against self-incrimination, even though she were not in fact criminally involved. On the whole it appears to us therefore that no such inference should be drawn in the case at bar; and it is not necessary to decide the point, for there was evidence enough without it, as we have said; and the question was not raised at the trial.
The fourth point is that the judge admitted some evidence that was not adequately connected with Costello, and did not confine the prosecution to the same amounts of unreported gross income that it had alleged in a bill of particulars. The evidence whose admission is complained of consisted of five items appearing in the prosecution’s computation for the “indictment years”; their aggregate is about $29,000, about $20,000 of which was entered as spent in the year 1949. For that year the prosecution computed an unreported net income of $32,500, and as has already appeared, even a cash reserve of $73,000 on January 1, 1946, would have been exhausted by January 1, 1949. Thus, it would make no difference as to the fourth count even though these items of income had been omitted. We do not mean by this that they were not adequately proved; but it is not necessary to discuss them. The $9,000 entered in the earlier years was for about $1,500 for travel expenses, $4,400 for a Cadillac motor, and $3,000, the initial payment on a mausoleum. The travel expenses were proved by testimony of the railway fares to and from places that Costello had visited; and the objection is frivolous. Judge McGohey heard extended argument about the motor car, and, although the proof did not amount to demonstration, it is apparent that every reasonable inference pointed to its purchase for Mrs. Costello. As to the payment on the mausoleum, the testimony of Festa, who signed the contract, was that he called up some one whom he supposed to be Costello and said that the payment was necessary. His interlocutor agreed and soon thereafter he received $3,000 in $100 bills with which he made the payment. The mausoleum was for Costello, and the suggestion that some one else might have paid for it, does not deserve serious answer. Finally, although it is true that the prosecution proved about $44,000 more of unreported gross income than the total — $244,000 —which it had alleged in the bill of particulars, that was a variance that could not have prejudiced the defence. Indeed, that was not suggested at the trial. We have long since passed the time when a trial has to be conducted with the formality of the code duello; § 52(a) of the Rules of Criminal Procedure, 18 U.S.C., directs us to “disregard” any “variance” that “does not affect substantial rights”.9
The fifth point is that the judge’s charge was inadequate. In Holland v. United States [348 U.S. 121, 75 S.Ct. 132] the Supreme Court said that in “net worth” prosecutions “charges should be especially clear, including, in addition to the formal instructions, a summary of the nature of the net worth method, the assumptions on which it rests, and the inferences available both for and against the accused.” In the margin we have put those portions of Judge McGohey’s charge that cover this [676]*676feature of the case,10 and we can find nothing omitted that would have added to the enlightenment of the jury. It is unnecessary to discuss the refusals to charge. We understand the defence also to raise the point that the acquittal on the first count was inconsistent with the conviction on the other three. That is not true in i'act, as has already appeared; but the point is ill taken in any event, for rational consistency between the verdicts of a jury is never necessary.11
The last point is that the grand jury had no competent evidence at the inquest cn which to find the indictment; and that for this reason it should have been dismissed. During the trial Costello's counsel asked each of the witnesses for the prosecution, whether he had appeared before the grand jury, and only three had done so. Those who had appeared had had no personal acquaintance with Costello or his affairs; they were accountants or experts and we find it impossible to escape the conclusion that their computations must have been based on hearsay. On this showing at the close of the prosecution’s evidence, and again at the close of all the evidence, Costello moved to dismiss the indictment. He had pleaded not guilty in April 1953, and in February, 1954, he obtained an order to show cause from Judge Weinfeld for an inspection of the minutes of the grand jury, and to dismiss the indictment because “there was no lawful evidence” to [677]*677support it. In support of this order he filed an affidavit asserting his innocence, and that he was “firmly convinced that there could be no legal or competent evidence before the Grand Jury,” because he had always reported his full income. On return of the order Judge Weinfeld denied the motion, and we do not understand that the defence asserts that in so doing he was wrong; at any rate he was plainly right. Eule 12(b) (3) of the Eules of Criminal Procedure provides that such a “motion shall be made before the plea is entered,” but that “the court may permit it to be made within a reasonable time thereafter”; and subdivision 4 of Eule 12(b) provides that, although it “shall be detei'mined before trial”, the hearing upon it may “be deferred for determination at the trial of the general issue.” Perhaps, in the light of this, it would be permissible to dispose of the denials of the motions made at the close of the prosecution’s case and of the whole ease, by saying that they should not be reviewed because Costello had failed to ask Judge Weinfeld to defer their determination until “the trial of the general issue.” It seems to us that, especially in a prosecution so jealously circumscribed as this, we ought not to dispose of a point going to the very heart of it upon such a formal and harsh ground. Costello had done all he could to raise the point; and, since he was without that not uncommon help that the names of witnesses called at the inquest must be endorsed upon the indictment,12 he was stripped of all means of showing that the whole proceeding had no lawful beginning, unless the course he adopted was left open. We cannot accept it that such a trivial omission or slip in procedure should have such a decisive consequence; and we hold that we may not escape deciding whether it is ground for dismissing an indictment that all the evidence adduced at the inquest was hearsay.
It is indeed well settled that the admission of incompetent evidence at the inquest is not a ground for dismissing the indictment;13 but at times the courts have assumed that it is otherwise, if all the evidence is incompetent. In Nanfito v. United States, 8 Cir., 20 F.2d 376, 378, and Brady v. United States, 8 Cir., 24 F.2d 405, 407, 59 A.L.R. 563, that was made the basis of the decision, and Judge Phillips declared in the second of these cases that it was “the settled law” in the Eighth Circuit; but we have found no other federal appellate decisions which can be said to have actually turned upon the point. If “incompetent” is to cover all evidence, however rationally persuasive it may be, that would be excluded at a trial with great deference we cannot agree. Legal rules presuppose that the occasions to which they apply, shall be decided under the ordinary postulates of reasoning; and the exclusory rules are an exception, for they apply to evidence that is relevant rationally, but that courts will not accept, not because it does not prove the issue, but it is thought unjust to the opposite party to use it against him, or because it is within some privilege to suppress the truth. We should be the first to agree that, if it appeared that no evidence had been offered that rationally established the facts, the indictment ought to be quashed; because then the grand jury would have in substance abdicated. But that is in no sense true of hearsay, as is shown both by the numerous exceptions, based upon the inaccessibility of any better evidence, and by decisions that have again and again held that it may be as dependable a reliance [678]*678at a trial as any other evidence; the only condition upon its use being that the opposite party shall not object to it.14 Indeed, we conduct our most serious affairs upon the strength of it; it would be impossible to carry on a day’s business without it.
The resulting situation: that is, that hearsay will serve, only when supplemented by some modicum of first-hand evidence, has nothing to commend it. The reason why evidence, incompetent for any reason, will ordinarily upset the judgment is that, except in rare cases, it is impossible to know how far it may have determined the judgment; yet it is obviously just as impossible at an inquest as at a trial to know that it was not the hearsay alone that convinced the jurors. We make no effort, and can make none, to ascertain what part it may have played in the result; all we can do, if hearsay alone is not to be enough, is to insist that there must be some first-hand evidence, no matter how feeble and untrustworthy the jury may have thought it. We can see no justification for such an amorphous compromise; and it is particularly unsatisfactory in a unilateral investigation like an inquest. Apparently the notion persisted down into the 18th Century that the objection to hearsay was that it allowed disputes to be “tried” otherwise than by oath;15 but already in the 17th Century it had been held that the reason against accepting depositions was that “the defendant not being present when they were taken before the mayor, and so had lost the benefit of cross-examination.” 16 “As our system changed in character from investigative to adversary the rule rejecting hearsay and the rule making opportunity for cross-examination a requisite of admissibility developed side by side. * * * Consequently, not the jury system, but rather the adversary theory of litigation, coupled with the then currently accepted notions as to the value, of the oath, accounts for the hearsay rule as it was at the opening of the nineteenth century.”17 So far as the right of cross-examination is its sanction, first-hand evidence would be as inadmissible at an inquest as hearsay; and a refusal to recognize it as sufficient proves too much; for the accused is normally not present at an inquest and cannot cross-examine. True, there is no doubt some warrant of reliability in an oath, and more in the opportunity to observe the witness; and both are absent in the case of hearsay; but so far as these may be deemed to count, they should have excluded it altogether, which, as we have seen, they do not do. The excuse no doubt is that the stake is not the accused’s guilt but only the burden of the risk and expense of a trial; but why should it be thought that that burden is perceptibly lightened by the requirement that any bit of added first-hand evidence will be enough, be it ever so little persuasive? Surely that is in practice an utterly illusory protection; and often would lead, as it would have here, to a monumental obstacle and expense to the prosecution of crime. We will not add one more count to Professor Morgan’s indictment of the whole subject: “The fact is, then that the law governing hearsay to-day is a conglomeration of inconsistencies, developed as a result of conflicting theories. Refinements and qualifications within the exceptions only add to its irrationality. [679]*679The courts by multiplying exceptions reveal their conviction that relevant hearsay evidence has real probative value, and is capable of valuation by a jury as well as by other triers of fact.”18 Consider also what another acknowledged master of the subject says, Dean Wig-more. “But suppose that there is a limitation of the grand jury’s sources of investigation in the shape of a rule that they may receive only such kinds of evidence as would be receivable on a trial before a petit jury. Such a rule is a plain obstruction of justice, reprehensible in policy. * * * In the grand jury’s proceedings, if the rule is to be enforced at all, as it is for petit juries, the fact must be allowed to be shown by the grand jurors or others present. If, then, any community is willing to accept so deleterious a rule of criminal procedure, its enforcement in the only feasible way must be permitted by showing the facts.”19
The question is still undecided by the Supreme Court, and in this circuit we are uncommitted except for an old decision of Wallace, J., in a ruling as district judge, which he confined to “extreme cases, when the court can see that the finding of a grand jury is based upon such utterly insufficient evidence, or such palpably incompetent evidence, as to indicate that the indictment resulted from prejudice, or was found in wilful disregard of the rights of the accused”.20 We hold that it is immaterial that only hearsay was adduced at the inquest in support of the allegations of the indictment and it follows that although the conviction on the second count will be reversed, that on the third and fourth counts will be affirmed.
Judgment on the second count reversed; judgment on the third and fourth counts affirmed.