The defendant was tried before Judge Foley and a jury on two counts of wilfully evading personal income taxes in 1953 in violation of 145(b) of the Internal Revenue Code of 1939, 26 U.S.C.A. 145(b) and in 1954 in violation of the successor section, 7201 of the Internal Revenue Code of 1954, 26 U.S.C.A. 7201. The government used the net worth method to prove the amount of unreported income, and the jury convicted on both counts. Judge Foley sentenced the defendant to six months' imprisonment on each count to run concurrently and fined him $2,000 on each count. On appeal the defendant raises, inter alia, the sufficiency of the judge's charge to the jury, the adequacy of the evidence, and numerous alleged trial errors. We believe that the charge was not erroneous and that the evidence of guilt was adequate. We find, however, two specific errors as to the admission and use of evidence. Since these were prejudicial as to the indictment year 1953 but not as to 1954, we reverse the conviction as to 1953 and affirm as to 1954.

The defendant operated an industrial laundry business under the name of Star Overall Cleaning & Supply Company (Star) in Schenectady, New York, as a sole proprietorship. Star rented and laundered overalls and uniforms. In some cases the individual employee who wore the overalls or uniform paid Star's charges and in others his employer paid. Star also rented and laundered industrial wiping cloths. The government alleges that the defendant obtained unreported income by cashing some of the checks received from Star's customers without entering them on Star's books or reporting the income on his federal income tax return.

A second alleged source of unreported income is undeclared dividends. In the years 1948 to 1953 the defendant reported no dividends, and in 1954 he reported $255 of dividends.1 The government agent testified that according to financial reports the defendant should have received dividends in the years 1950, 1951, 1952, and 1953 in the amounts of $965, $1,674, $2,023, and $1,596, respectively.

The government submitted a summary statement of the defendant's assets, liabilities, and personal expenses,2 which showed net worth bulges, i.e., an excess of income computed on the net worth method over income reported on annual federal income tax returns, for the calendar years 1949 through 1954. The net worth bulges for the two indictment years, 1953 and 1954, were $13,922 and $20,475, respectively, computed as follows:

                            1953     1954
Taxable income computed
  on the net worth
  method ................. $18,148  $27,440
Taxable income reported
  on defendant's
  returns .................. 4,226    6,964
                           -------  -------
Unreported income ........ $13,922  $20,475
                           -------  -------

The bulges for the four previous years, 1949 through 1952, totaled $47,725 and were admitted only as proof of wilfulness. See, e.g., United States v. Ford, 237 F.2d 57, 60, 65, 67 (2 Cir. 1956).

The principal defense was that the defendant neither handled Star's books nor made out his own tax returns, and that unreported rental income was due to the default of Star's bookkeeper while unreported dividends were due to the default of the defendant's attorney, and that both were without defendant's knowledge. However, Benjamin Segal, the attorney who prepared defendant's 1948 through 1953 tax returns, testified that he had asked the defendant each year whether he had received any dividends, and the defendant had replied no. The trial judge correctly left the issue of wilfulness to the jury. See United States v. Schenck, 126 F.2d 702, 706-707 (2 Cir. 1942), cert. denied, Moskowitz v. U.S., 316 U.S. 705, 62 S. Ct. 1309, 86 L. Ed. 1773 (1942).

The defendant offered two reasons why he had nonfraudulently understated his taxable income on his 1953 and 1954 returns. He had taken a deduction for $2,700 for bad debts on his 1953 income tax return. Since he reported income only as it was received, i.e., was on the cash basis, he was not entitled to such a deduction. If the defendant took this deduction erroneously but not fraudulently, the 1953 bulge would have been explained consistent with lack of wilfulness to the extent of $2,700.

The second reason was explained by a certified public accountant whom the defendant had retained to examine his books in preparation for trial. He testified that certain income and expense items had been erroneously entered. The net effect of these errors was to reduce the 1953 and 1954 net incomes shown on the defendant's books approximately $5,000 and $6,800, respectively, below their actual levels. Since the defendant's tax returns were prepared from his books of account, these errors were reflected on his tax returns. If the errors were made by the defendant's bookkeeper without defendant's knowledge, the bulge would have been explained consistently with lack of wilful tax evasion to the extent of $5,000 in 1953 and $6,800 in 1954.

Both as to the bad debt deduction and as to the bookkeeping errors, the trial judge was correct in leaving the issue of wilfulness to the jury. And it was, therefore, proper for the government to resolve these issues against the defendant in preparing its summary of the defendant's net worth. The defendant could have introduced his own summary, resolving these factual questions in his favor. Scanlon v. United States, 223 F.2d 382, 391 (1 Cir. 1955).

On appeal the defendant claims that he used most of the money procured from cashing the unrecorded rental checks to pay business expenses which were not deducted on his income tax returns, and thus that the unreported income and expenses offset each other. The failure of the government agent to investigate this claim was, the defendant argues, reversible error. We disagree.

When a taxpayer furnishes leads which might reasonably explain his net worth bulge inconsistent with guilt, the government must investigate these leads or the trial judge should consider the taxpayer's explanations as true. Holland v. United States, 348 U.S. 121, 135-136, 75 S. Ct. 127, 99 L. Ed. 150 (1954). This leaves the burden of proof of tax evasion on the government, putting only the burden of production on the defendant. Id. at 137-139, 75 S. Ct. 127. However, if the defendant had actually expended the unreported income to pay unreported deductible expenses, he would not have had the cash on hand at the year end and his net worth would not have been increased. Thus, the defendant's claim does not tend to explain his net worth bulge.

The government had, in fact, introduced evidence of the unreported checks not as proof of the defendant's bulge, but only as proof of his intent, id. at 139, 75 S. Ct. 127, and as proof of a likely source of unreported taxable income, id. at 138-139, 75 S. Ct. 127. Rather than requiring the government to prove the negative proposition that the defendant had no other deductions, it is reasonable to require the defendant, if he wishes to disprove intent and likely source, to bear the burden of going forward when he alleges that he had additional deductions not claimed on his income tax return. See Clark v. United States, 211 F.2d 100, 103 (8 Cir. 1954), cert. denied, 348 U.S. 911, 75 S. Ct. 289, 99 L. Ed. 714 (1955); United States v. Link, 202 F.2d 592, 593-594 (3 Cir. 1953); United States v. Bender, 218 F.2d 869, 871-872 (7 Cir. 1955), cert. denied, 349 U.S. 920, 75 S. Ct. 660, 99 L. Ed. 1253 (1955); United States v. Lennon, 246 F.2d 24, 27 (2 Cir. 1957), cert. denied, 355 U.S. 836, 78 S. Ct. 60, 2 L. Ed. 2d 48 (1957).

The defendant also contends that the trial judge erred in refusing to allow proof of defendant's accounts payable for overalls and uniforms purchased which would have reduced the defendant's net worth. The trial court was correct. Taxpayers may elect various methods of accounting for use on their income tax returns. Although a taxpayer's total taxable income over the long run should be the same regardless of the method of accounting selected, in any given year a taxpayer on the accrual basis may have a substantially different amount of taxable income than if he had chosen the cash basis. The purpose of a net worth computation is to check the accuracy of the amount of income reported by the taxpayer. Thus, if the taxable income reported on his return is correct, and if no errors creep into the asset, liability, and expenditure figures used to ascertain successive years' net worth, the income computed on the net worth basis and that reported on his tax return should be the same. This is possible only if the figures used in computing net worth are keyed to the taxpayer's method of accounting. If a taxpayer disregards his accounts payable in reporting income on his annual tax return, i.e., reports as a cash basis taxpayer, then the government in computing his net worth in order to check the income reported on his tax return must also disregard these amounts. Whether this computation reflects net worth as accountants define it, i.e., the excess of the value of assets over liabilities, is irrelevant, United States v. O'Connor, 273 F.2d 358, 361 (2 Cir. 1959), since the government is not seeking to determine the taxpayer's true worth but rather to verify the accuracy of his income tax return for a particular year.

An examination of the defendant's tax returns and books of account discloses that he reported sales and purchases on the cash method of accounting, disregarding accounts payable for purchased overalls and uniforms. Therefore, the trial court correctly excluded proof of such payables in reconstructing his taxable income on the net worth method. See Clark v. United States, 211 F.2d 100, 105 (8 Cir. 1954), cert. denied, 348 U.S. 911, 75 S. Ct. 289, 99 L. Ed. 714 (1955); Scanlon v. United States, 223 F.2d 382, 389 (1 Cir. 1955); Leeby v. United States, 192 F.2d 331, 334 (8 Cir. 1951).

The defendant raises five claimed factual mistakes in the government's net worth summary. However, on each the trial judge properly left the conflicting testimony to the jury and properly permitted the government to base its net worth summary on its version of the facts. See Smith v. United States, 236 F.2d 260, 263-264 (8 Cir. 1956), cert. denied, 352 U.S. 909, 77 S. Ct. 148, 1 L. Ed. 2d 118 (1956); United States v. O'Connor, 237 F.2d 466, 467-475 (2 Cir. 1956). Only three of these alleged errors require discussion.

The government listed the defendant's cash on hand at December 31, 1948, the starting point for the government's net worth summary, at $10,000 and the cash on hand at the end of each year from 1949 to 1954 at $300. The government based its figures on an oral admission the defendant had made to a revenue agent. The agent testified that the defendant had said during an interview that before 1948 his mother had given him approximately $10,000 which he retained in cash until 1949 when he deposited the remainder in a savings account, and thereafter kept only $200 to $300 on hand. Since the defendant testified at trial that these figures substantially understated his cash on hand, there must be corroborating evidence to make a case for the jury. Smith v. United States, 348 U.S. 147, 75 S. Ct. 194, 99 L. Ed. 192 (1954); United States v. Calderon, 348 U.S. 160, 75 S. Ct. 186, 99 L. Ed. 202 (1954). The record is replete with evidence corroborating the defendant's admission that he did not have large amounts of cash on hand after December 31, 1949. The defendant periodically borrowed money to meet payrolls and other indebtedness. Furthermore, there frequently were judgments outstanding against him during this period. These facts are sufficient to permit the question of cash on hand to go to the jury. See Friedberg v. United States, 348 U.S. 142, 75 S. Ct. 138, 99 L. Ed. 188 (1954); United States v. Sclafani, 265 F.2d 408, 411-412 (2 Cir. 1959), cert. denied, 360 U.S. 918, 79 S. Ct. 1436, 3 L. Ed. 2d 1534 (1959); United States v. Ford, 237 F.2d 57, 62-63 (2 Cir. 1956). See also Holland v. United States, 348 U.S. 121, 132-134, 75 S. Ct. 127, 99 L. Ed. 150 (1954).

The defendant testified that he owned $3,800 of securities purchased through Bache & Co. that were not shown on the summary sheet and that $3,000 of his 1952 cash inflow was due to a casualty loss and thus was non-taxable income. The government agent testified that he knew nothing of these two claims. As previously discussed, the government must investigate all of the taxpayer's reasonable explanations of his net worth bulge inconsistent with guilt or the trial judge may consider the explanations as true. Holland v. United States, 348 U.S. 121, 135-136, 75 S. Ct. 127, 99 L. Ed. 150 (1954). Thus if Vardine told the government agents about the securities and the casualty loss a sufficient period before the trial to permit investigation and they did not follow up these leads, the trial judge should have assumed their validity and required the government to reflect them on their summary statement. Scanlon v. United States, 223 F.2d 382, 388-389 (1 Cir. 1955). However, the defendant introduced no evidence that he gave these leads to the government before trial. Therefore, Judge Foley correctly left these issues to the jury.

The defendant presses two claims which have merit. As proof of the balance in defendant's checking account the government introduced the amount shown on the bank's records on December 31 of each year. The defendant then introduced evidence that he had checks outstanding at the end of the years 1952, 1953 and 1954. The trial judge left it up to the jury whether to deduct these outstanding checks from the defendant's year-end checking account balance, and the government, on their net worth summary, showed the balance as it stood on the bank's books. This was error. When a cash basis taxpayer delivers his check without any limitation on the payee's ability to cash it immediately, the taxpayer is entitled to a deduction at the time of delivery. Clark v. Commissioner, 253 F.2d 745 (3 Cir. 1958). Therefore, the trial court should have instructed the jury as a matter of law to reduce the defendant's year-end checking account balance by the amount of the outstanding checks and the government to follow suit on its net worth summary statement. See United States v. O'Connor, 237 F.2d 466, 475 (2 Cir. 1956); United States v. Altruda, 224 F.2d 935, 942 (2 Cir. 1955). The amount of outstanding checks were as follows:

December 31, 1952 .. $  417.
December 31, 1953 ... 1,824.
December 31, 1954 ..... 309.

These figures would have reduced the defendant's 1953 net worth bulge by $1,407 and increased his 1954 bulge by $1,515. The government contends that since the net effect of these errors favors the defendant, there has been no harm. But since there are two separate indictment years, each must be considered separately.

The other error concerns machinery and trucks. During 1953 and 1954 the taxpayer purchased new machinery and trucks for his business. As part payment of the purchase price he traded in used machinery and trucks. In four instances the trade-in allowance was larger than his adjusted basis in the assets traded-in. The Internal Revenue Code provides that such paper gains on a trade-in of old for new equipment shall not be recognized. Internal Revenue Code of 1954, 1031, 26 U.S.C.A. 1031; Internal Revenue Code of 1939, 112(b)(1), 26 U.S.C.A. 112(b)(1). Gain on the traded-in asset is postponed by taking as the basis of the new asset the basis of the old asset plus any additional cash paid. 1954 Code 1031(d); 1939 Code 113(a)(6). The trial judge left it to the jury whether to accept the new assets at their full selling price or at the lower price that would be computed by excluding the gain on the trade-in and permitted the government to show these assets at the higher value on their net worth summary. This was error. The government argues that the defendant 'waived' his rights under 1031 of the 1954 Code and its predecessor by erroneously taking depreciation deductions on the higher basis. But 1031 and its predecessor are mandatory, not optional; a taxpayer cannot elect not to use them. Therefore, the trial judge should have instructed the jury as a matter of law to accept the lower basis for these assets and directed the government to alter its net worth computation accordingly. See United States v. O'Connor, 237 F.2d 466, 475 (2 Cir. 1956); United States v. Altruda, 224 F.2d 935, 942 (2 Cir. 1955). These errors caused the defendant's net worth bulge to be overstated by $592 in 1953 and $761 in 1954.

The total effect of the two errors can be summarized as follows:

                                  1953              1954
Outstanding checks ............. $1,407       $(1,515) *
Trade-ins ......................   592            761
                            ----------------  ----------------
Total errors against the
  defendant .................... $1,999       $ (754) *
                            ----------------  ----------------
* Figures in parenthesis are favorable to the defendant.

While the 1953 error of $1,999 is small in comparison to the 1953 bulge of $13,922 shown on the government's net worth summary, it is impossibly to tell how many of the factual questions the jury decided in the defendant's favor. In view of the other items in dispute,3 the $1,999 error might, therefore, have accounted for the entire 1953 bulge as it was found by the jury or it might have been sufficiently material in relation to the bulge found by the jury so that we should not second guess whether they would have convicted on this count. See Flemister v. United States, 260 F.2d 513, 517 (5 Cir. 1958).

The two errors did not harm the defendant as to 1954, but on the contrary, reduced the net worth bulge shown on the government's summary below what it should have been. Although the jury might have been relying on the apparent 1953 bulge, which we have found to be at least partially erroneous, as proof of wilfulness in 1954, the other proof of defendant's intent was sufficient to preclude reversal as to 1954. The fact that defendant consistently failed to report his dividend income and cashed Star's checks without recording them on the books or reporting them on his return coupled with the consistent net worth bulges for the four years preceding the prosecution years make the $1,999 error in 1953 insignificant in proving wilfulness as to 1954. Cf. United States v. Sclafani, 265 F.2d 408, 412-413 (2 Cir. 1959), cert. denied, 360 U.S. 918, 79 S. Ct. 1436, 3 L. Ed. 2d 1534 (1959); Scanlon v. United States, 223 F.2d 382, 388-389 (1 Cir. 1955).

Affirmed as to 1954 and reversed as to 1953.