HAIGHT, Circuit Judge.
The defendant in error (the plaintiff below) sued to recover certain moneys which it claims were illegally assessed and exacted from it, by way of taxes for the year 1913, under the Income Tax Act of October 3, 1913 (38 Stat. L- 166, c. 16). The case was tried without a jury, pursuant to sections 649 and 700 of the Revised Statutes (Comp. St. §.§ 1587, 1668), and resulted in a judgment in favor of the plaintiff for the full amount claimed.
[1] The questions presented depend for their decision upon the construction to be given to certain of the provisions of section 2, G (b), of the before-mentioned act. The plaintiff is a mutual life insurance company without capital stock, conducting its business on what is known as the “level premium plan,” and annually declares “dividends” to its policy holders.- In order to relieve this opinion of the burden of a discussion of the method by which premiums are fixed under the level premium plan, and how and out of what funds the so-called “dividends” to policy holders are declared, we refer to the explanation on pages 202, 203, of the opinion in Mutual Benefit Life Insurance Co. v. Herold (D. C. N. J.) 198 Fed. 199. That aptly describes -the practice followed by the plaintiff in this case, except that a part of some of the dividends which enter into this controversy were derived from sources other than overpayment of premiums by the policy holders to whom such dividends were paid, as will be hereinafter set forth; but for present purposes the dividends in question [83]*83may be considered as derived from and as representing the excess of the actual premium or premiums previously paid by a policy holder over the subsequently ascertained cost of his insurance for any given year or years. Such excess is included in the premium as a margin of safety, and is in insurance parlance called, and will be fre'quently referred to, as a “redundancy.”
We deem it unnecessary to describe the different kinds of policies involved in this suit, or how and when and why the dividends accumulated and were payable thereon, respectively; it being, we think, sufficient merely to point out that during the tax year of 1913 (from March 1 to December 31, 1913) the plaintiff paid cash dividends— moneys actually paid, as distinguished from dividends used in abatement of subsequent premiums or to purchase paid-up additions to existing policies, etc. — aggregating in amount $683,729.03, to the holders of 10 different kinds of its policies, and that in many cases no premiums were received or due from the same policy holders that year, and in other cases the premiums received from such policy holders were less than the dividends paid to them, respectively. The company claimed the right to exclude the aggregate of those payments from its total income during that period, in order to ascertain the net income upon which, under the act in question, it was required to pay a tax. The government, on the other hand, took the position that the whole of this sum could not be so excluded. Out of these contrary views the present controversy arose.
The act imposed a normal tax “upon the entire net income arising or accruing from all sources” to every domestic insurance compan)-. Subdivision G (a)' of section 2. It also provided that the net income should be ascertained—
“by deducting from the gross amount of the income of such * * * insurance company, received within the year from all sources, * * * tlie net addition, if any, required by law to be made within the year to reserve funds and the sums other than dividends paid within the year on policy and annuity contracts.” Section 2, G (b).
It then, after making special provisions in separate provisos for mutual fire insurance companies and mutual marine insurance companies, contained, in the latter proviso, the following:
“And life insurance companies shall not include as income in any year such portion of any actual premium received from any individual policy holder as shall have been paid back or credited to such individual policy holder, or treated as an abatement of premium of such individual policy holder, within such year.”
The first of these clauses will, for convenience, be hereafter referred to as the “deduction” clause, and the latter as the “noninclusion” clause.
Leaving out of consideration for the moment, the grounds upon which the plaintiff contends that it is entitled to exclude those parts of the dividends paid to policy holders which do not represent the redundancies in the premiums previously paid by such policy holders, the plaintiff’s before-stated contention is based on the proposition that under the “noninclusion” clause-it may exclude from its total in[84]*84come (income derived from all sources) any moneys which it may have paid as dividends to policy holders during the tax year, provided that such payments represent redundancies in premiums previously paid by the same policy holders, irrespective of whether, during the tax year, it received anything by way of premium from them, respectively. The government’s contention proceeds on the theory that dividends paid ■ to a policy holder, or, for that matter, used in abatement of premiums or otherwise credited to a policy holder, may be excluded from the total income in any tax year only when the company has received a premium during the tax year from the individual policy holder to whom the dividend is paid, and then only to the extent of the amount of the premium so received, either actually or theoretically.
These divergent views result from the different interpretations given to the words “portion of any actual premium.” The company would read these words as characterizing the source from which the dividend was derived; that is to say, it would permit the exclusion from the total income of dividends paid or credited to policy holders when they represent redundancies in previous premium payments, without regard to' whether any premiums were received that year from the policy holders to whom the dividends were paid. The government, on the contrary, would read the words as limiting the extent or amount of the dividend paid or credited to the policy holder, which may be excluded from total income, to the sum received by way of premium from that same policy holder during the year in which the dividend is paid or credited to him, without regard to the source from which the dividend was derived or when it became payable. It is also manifest that under the company’s interpretation it is necessary that the policy holders be treated as a whole, while under the government’s contention each-is dealt with separately.
Under the company’s theory, if iff received, for example, by way of premium, from all of its policy holders and from other sources, a total income, in any given tax year, amounting to the sum of $1,000-000, and paid to policy holders, from whom it received no premium whatever that year, dividends amounting to $100,000, and which represented redundancies in premiums previously paid by those same policy holders, it would report as its gross income $900,000, whereas in fact its actual gross income was $1,000,000. The result, of course, would be to produce an erroneous net income, because the latter would be based on a false gross income. Under the government’s theory, the company would not be entitled to deduct any dividends paid to individual policy holders from whom it had not in that year received some premium, or at least a premium equal to the amount sought to be excluded as paid to him; in other words, if A., a policy holder, in 1913 paid no premium, but received a cash dividend representing redundancies in premiums paid by him in previous years no account would be taken of the dividend paid to him in that year, so far "as the company’s taxable income is concerned. It would be left, out of the calculation altogether. On the other hand, if the company in that year had received premiums from A., B., C., D., etc., and had paid to each of them dividends representing redundancies in [85]*85premiums paid in a previous year or years, which as to each individual policy holder was less than, or did not exceed, the premium received from each, respectively, during the tax year, the company would be entitled, in ascertaining its gross income, to exclude the aggregate amount of the dividends so paid that year to those policy holders. The result would be that the amount thus reported as gross income would represent the total of the amount actually received by way of premiums during the tax year. The learned trial judge adopted the plaintiff’s theory.
Having thus ascertained the contentions of the respective parties, the theories upon which they are based, and the practical effect thereof, we now proceed to an analysis of the language of the statute, iu the light of existing statutory enactments and judicial interpretation thereof at ihe time the act in question was passed. It will be noted, primarily, that the act, so far as life insurance companies are concerned, provides two distinct steps in the ascertainment of the taxable net income, viz.: (a) The ascertainment of gross income; (b)
the deductions to be made from gross income to ascertain net income. In ascertaining gross income, it is specifically provided that certain parts of any actuad premium received from any individual policy holder shall not be included, and after the gross income has thus been ascertained, the net income upon which the tax is to- be assessed is arrived at by deducting various items, among which are the sums, other than dividends, paid within the year on policy and annuity contracts. This distinction was recognized by the company in making up its return for the tax year in question, because all of the moneys now in controversy were deducted from the company’s total income for the purpose of ascertaining gross income, and were not deducted from gross income for the purpose of ascertaining net income.
The original Corporation Excise Tax Act of 1909 (36 Stat. 112, c. 6, § 38), which, so far as it imposed a tax upon the “entire net income,” was, in all respects material to this case, the same as the act of 1913, authorized insurance companies to make the same deductions from their gross income, to ascertain their net income, as does the “deduction” clause of the act in question; hut it contained no such clause as the “noninclusion” clause of the act of 1913. Tor many years before the act of 1909 was passed, it had been the policy of life insurance companies, both mutual and stock (when the latter issued participating policies) doing business on the level premium plan, annually to declare “dividends” to policy holders, which represented, as before mentioned, the difference between the premium provided in the policy to be paid and the actual cost of the insurance. The companies permitted the insured to use these dividends in any one of several ways, viz. as an abatement of future premiums, in purchasing paid-up additions, withdrawal in cash, etc. When dividends were used in abatement of premiums, it was dear that: the company actually, as distinguished from theoretically, received, in the year in which the abatement was made, not the full amount of the premium charged and payable under the policy, but only the difference between the dividend and the premium; and so in respect to full-[86]*86paid additions, while the company received the full premium due and chargeable under the policy, it did not actually receive the dividend as premium on the paid-up addition to the policy.
The result was that the companies sought, in making their tax returns under the act of 1909, to charge themselves only with the moneys actually received in the tax year, and thus to eliminate from their gross income those dividends which, as a matter of bookkeeping, they appeared to have received that year, but actually did not receive. _ The government, on the other hand, contended that those dividends were such “dividends” as, under the act, the companies were not permitted to claim deduction for, and that they must therefore be included , as part of their gross income. The question thus presented first came up for judicial decision in Mutual Benefit Life Insurance Co. v. Herold, supra, and in a well considered and reasoned opinion the late Judge Cross adopted the contention of the insurance companies. His decision was affirmed by this court on January 27, 1913 (Herold v. Mutual Benefit Life Ins. Co., 201 Fed. 918, 120 C. C. A. 256), and the Supreme Court later denied a certiorari to review it (231 U. S. 755, 34 Sup. Ct. 323, 58 L. Ed. 468). Thus by judicial construction one feature of the “noninclus'ion” clause of the act of 1913 had already been incorporated in the act of 1909 when the later act was passed; i. e., that insurance companies were not required to include in their gross income such part of the premiums payable by policy holders as they had not actually received, but which had been pro tanto abated by the application of dividends. It is proper to here observe incidentally that dividends could not be used in abatement of a premium of a policy holder unless a premium was due from that policy holder, greater or at least equal in amount to the dividend.
On October 3, 1913, the act in question was passed, and in it the “noninclusion” clause appeared for the first time. Reading the “non-inclusion” clause in the light of the before-mentioned contentions and the history of the litigation thereon, and bearing in mind that, so far as insurance companies are concerned, the deductions permitted to be made from gross income to ascertain net income were the same in the act of 1909 as in the act of 1913, it seems reasonable to conclude that the “noninclusion” clause was inserted for the purpose of setting at rest (which had not been authoritatively done when the act was passed, because the petition for the certiorari was not denied until December 15, 1913), the contention which the government was making in these matters, and to indicate clearly that it was not the intention of Congress to assess a tax in any given year on such parts of premiums as the company received only in theory during that year, but exclusively on that part of the premiums which they actually received — that which in fact, not in theory, came in to them. It will also be seen by reference to the decision of Judge Cross in the Mutual Benefit Case (198 Fed. 204) that .where policy holders withdrew their dividends in cash the total amount of the premium payable under the policy was included in the plaintiff’s statement of income received and the tax was imposed thereon by the government; [87]*87and the plaintiff in Conn. General Life Insurance Co. v. Eaton (D. C.) 218 Fed. 188, 203, appears to have done the same.
It would therefore appear that, when Congress used the words “paid back” in the noninclusion clause of the act of 1913, it made that act more liberal to the insurance companies, so far as “dividends” paid in cash were concerned, than the previous act had been construed even by the insurance companies. It took another step, and treated a “cash dividend” just as the courts had previously treated a dividend used as an abatement of a premium. It will be observed that the noninclusion clause provides that life insurance companies “shall not include as income in any year.” Congress did not say that life insurance companies might “deduct” from the total income in any year, etc. It used the words “shall not include.” The word “include” carries with it the implication of something received, either actually or theoretically. It is difficult to understand how one may be relieved of including in his receipts something which he never received. How could an insurance company not include in gross income certain moneys paid out to policy holders, if there were no premiums paid by those policy holders that year from which to exclude them? Furthermore, Congress did not say that a life insurance company should not include “such portion of any actual premiums received,” etc., but it used the word “premium” — the singular.
To adopt the plaintiff’s construction necessitates the reading of the word “premium” as “premiums,” in nearly every payment involved in this suit, unless it should be held that Congress intended to permit the deduction of such part or portion of only one previously paid premium as represented a redundancy in such premium. But no reason for such a limitation is conceivable. The words “that year,” which the plaintiff contends must be read into the clause after the word “received,” if the government’s construction is adopted, we think are there by implication, because of the use of the word “premium,” instead of “premiums” and the words “shall not include.” Furthermore, the act directs that life insurance companies shall not include such portion of the premium received “from any individual policy holder” as shall have been paid back or credited “to such individual policy holder or treated as an abatement of premium of such individual policy holder.” It seems quite clear, by the use, thus conspicuously, of the words “individual policy holder,” that Congress manifested an intention to treat each policy holder separately — his premium and his dividend — and not all of the policy holders as a whole; the aggregate premiums and the aggregate dividends. In this connection it may be observed that the company keeps a separate account with each policy holder.
If the purpose of inserting the “noninclusion” clause in the act of 1913 was as before stated, then it seems apparent that Congress used the language that it did, advisedly, for that language quite as explicitly expresses the thought that a life insurance company was to be taxed only on the income which it actually received, as distinguished from theoretically, in any tax year, as it is inappropriate to convey the meaning which plaintiff seeks to gather from it. If Congress [88]*88had' intended that the aggregate of all dividends, representing redundancies in previous premium payments, paid to policy holders in a tax year, should be deducted from the aggregate premiums received from all policy holders in that year, without regard to whether premiums were received that year from all of the policy; holders to whom dividends were paid that year, it would not, we think, have used the words “shall not include,” but rather the word “deduct”; nor the word “individual,” qualifying the words “policy holder,” and would have used the word “premiums” instead of “premium.” The manifest reason why a difference is made in respect to dividends paid or credited to policy holders which do not exceed the premium received from the same policy holders during the tax year, and those which are in excess thereof, or in respect to those cases where no premium is received from the policy holder to whom a dividend is paid, and those cases where a premium is received, is that Congress desired to tax only the moneys actually received during the tax year, and not those theoretically received, as the government had been contending up to that time.
The plaintiff seeks to support its argument on the proposition that as the moneys paid back to policy holders represent previous excess premium payments upon which, in fact or in theory, the company has theretofore paid a tax, it is unreasonable, in the absence of clear and explicit lánguage to that effect, to presume that Congress intended that they should be taxed again. But this argument loses sight of the fact that no tax whatever is levied on a dividend, other than the initial one imposed when the premium of which it is a part is paid. No tax was imposed for 1913, for instance, on the moneys paid back to policy holders as cash dividends in that year. The tax would be assessed only on such part of the moneys received by the company during that year as premiums from policy holders as was in excess of the dividends paid to the same policy holders that year. On the other hand, to permit the cash dividends, in excess of moneys received, either actually or theoretically, as premiums from the policy holders to whom they are paid, to be deducted in 1913, would relieve the company from paying any tax whatever upon them, because by deducting them from gross income in that year it would nullify or neutralize to that extent the tax which had been imposed upon the premiums, of which they were a part, in previous years. It is thus apparent that what plaintiff really seeks to accomplish is, not to avoid double taxation, but to be relieved from paying any tax whatever upon such parts of premiums, no matter when paid, as represent redundancies, and which are subsequently returned to the policy holders who paid the premiums.
It may be entirely equitable and just that Congress should have adopted stjch a course, but we do not think that we would be justified in finding such intention in a statute which expressly provides that the tax shall be levied “upon the entire net income arising or accruing from all sources,” in the absence of a clear expression to that effect, and especially when to do so would necessitate a strained and unnatural construction of the language used by Congress. Moreover, [89]*89immediately preceding the clause in question are those dealing with, mutual fire insurance companies and mutual marine insurance companies. The provision in respect to the latter is as follows:
“Provided, furl her, that mutual marine insurance companies shall include in their rei urn of gross income gross premiums collected and received by them less amounts paid for reinstxrance, but shall be entitled to include in deductions from gross income amounts repaid to policy holders on account of prpmimns previously paid by them, and interest paid upon such amounts between the ascertainment thereof and the payment thereof.”
The language there used is quite different from that used in respect to life insurance companies, and the meaning to be gathered from it is clear. It specifically permits those companies to make the very deduction, as deductions, which the plaintiff claims that life insurance companies may make by way of noninclusion. The fact that there is such a marked difference in the language used in respect to the two kinds of companies, and each respectively conveys a different meaning, tends to strengthen the conclusion that so far as life insurance companies were concerned, they were, in ascertaining gross income, ’ to exclude payments or credits to policy holders in the tax year only to the extent of the amounts received, actually or theoretically, from the same policy holders, respectively, during the same year. The same deduction and noninclusion clauses were carried into the Income Tax Act of 1916 (Act Sept. 8, 1916, c. 463, 39 Stat. 756), and later into the act of 1919 (Act ffeb. 24, 1919, c. 18, dO Stat. 1057). The distinction between the two clauses is accentuated in the latter act, as they are placed in separate and distinct sections. Section 233 and section 234. Section 233, which defines gross income, specially provides that in the case of life insurance companies “there shall not be included in gross income such portion,” etc., using the same language as the act of 1913; that mutual marine insurance companies must include in gross iucome the same items as the act of 1913 required, but it makes no mention of what either kind of company may deduct. Section 234 deals with deductions exclusively, and permits the same deductions for life insurance companies (paragraph 11) and for mutual marine insurance companies (paragraph 12), as did the act of 1913.
It therefore appears that Congress has in each revision of the income tax law which it has made since 1913, retained the same language, as respects both life insurance companies and mutual marine insurance companies, as it did in the original act. It is a reasonable conclusion, therefore, that the language used in the act of 1913 was advisedly used, and, as it is different in respect to the two kinds of companies, that Congress intended that there should be a difference in the sums which each could, respectively, deduct on account of previous excess premium payments. That the construction for which we. have indicated a preference is the correct one is, we think made more manifest when the action of Congress in using the same language in the act of 1919, as it did in the acts of 1913 and 1916, is considered in the light of the regulations issued by the Treasury Department under the act of 1916. Regulations No. 33, issued January 2, 1918. Those regulations were as follows:
[90]*90“Premium Income Paid Bade. — Life insurance companies are authorized to-omit from gross income such portion of any actual premium received from any individual policy holder as shall have been paid back or credited to the policy holder or treated as an abatement of his premium.
“The amount authorized by this provision to be excluded from gross premium income on account of any premium refunded to any individual policy holder is explicitly limited to an amount not in excess of the actual premium paid by the individual policy holder within the tax year.
“Gash Dividends. — Life insurance companies are entitled under the foregoing holding to exclude from gross income any part of the premium received which is paid back to the individual policy holder within the same return year. Where the dividend is in excess of the premium received, there can be excluded from gross income only the amount of the premium received from such individual policy holder within the same return year.”
It must be presumed that Congress, in passing the act of 1919, had in mind the construction which the Treasury Department, in its regulations and instructions to internal revenue collectors, had adopted,, and that, if such construction did not properly express the intention of Congress, a change in the language, to express clearly a contrary meaning, would have been made in the act of 1919. Indeed, the Senate committee which had the act of 1919 in charge recommended a new basis for the taxation of life insurance companies, but in conference the Senate receded therefrom, and the basis upon which life insurance companies had been theretofore taxed was continued. See Montgomery’s Income Tax Procedure 1919, p. 669.
This situation is then presented: If the construction sought by the plaintiff is followed, although an equitable result would, in one-aspect, be therefore accomplished, it would be done only by a strained construction of the statute, and one, indeed, which would not be in harmony with what we think Congress, by the language which it used, manifested as its intention and confirmed by its subsequent action. If, on the other hand, the construction urged by the government is adopted, a reasonable result is reached, which gives effect to-what, it may be properly assumed, was the intention of Congress,, namely, to tax as the income of life insurance companies in any year only so much of any premium as may have been actually received, as distinguished from theoretically, from any policy holder during that year. In this contingency, there is, of course, no question as to-which construction should be adopted. We conclude, therefore, that the plaintiff was not entitled to exclude from its total income during the tax year in question, for the purpose of ascertaining its gross income, any dividends paid or credited to policy holders from whom it did not receive any premium during that year, and as to such policy holders as it did receive premiums from that year it was entitled to-exclude only such part of the dividends paid to those policy holders as did not exceed the amount received from them, respectively, by way of premiums, during that year.
[2] Although plaintiff has not contended, except as to items of interest paid to policy holders (which feature of the case will be hereafter referred to), that it is entitled, in ascertaining its net taxable-income, to credit for the dividends paid to policy holders, on the theory that such dividends may be “deducted,” as distinguished from-[91]*91“not included,” under the “deduction” clause, we have, nevertheless, considered that question. It might readily be conceived that dividends paid to policy holders, representing as they do excess premiums. are paid “on policy * * * contracts.” It may also be conceded that such dividends are not “dividends” in the commercial sense, as representing profits. Whether the word “dividends” is used in the “deduction” clause in the commercial sense, or in the sense in which insurance, companies ordinarily use it, we do not feel called upon to decide. If it be assumed, however, that the dividends in question are “sums * * * paid * * * on policy * * * contracts,” and that they are not dividends in the commercial sense, and that the statute in the clause in question refers to only commercial dividends, it might be argued with force that insurance companies are entitled to deduct them, under the “deduction” clause, from gross income, to ascertain net income, were it not for the noninclusion clause. At the threshold of any such argument, however, we would be met with the fact that dividends paid or credited to policy holders are unquestionably covered by the “noninclusion” clause. Manifestly, if it had been the intention of Congress that these dividends could have been deducted under the original “deduction” clause, there rvould have been no reason for adding the “noninclusion” clause. Indeed, by the latter, Congress explicitly defined what moneys returned, either in cash or by credit, to policy holders might be excluded by the companies, in ascertaining their taxable net income. Consequently we do not think that any of the cash dividends paid by the plaintiff to its policy holders in the tax year in question, which represented redundancies in previous premium payments, were deductible as “sums other than dividends paid within the year on policy * * * contracts.”
[3] Up to this point we have dealt with only those parts of the cash dividends as represented redundancies in previous premiums paid by the same policy holder to whom the dividends were paid. It appears, however, that some of these dividends represent, in part, interest which had accumulated on dividends which had been left with the company under certain forms of policies, and, in part, also on some policies, the share of the policy holders to whom the dividends were paid in funds accruing to holders of the same kind of policies, which for one reason or another had been forfeited, and which, under the terms of the policies, were divided among those policy holders of the same class who had not thus forfeited such benefits. The company claims that it is entitled (although it did not follow that course in making up its tax return) to deduct from its total income, to ascertain the taxable net income, such parts of the dividends as represent interest due to the policy holders, under the “deduction clause” of the statute, as payments made on policy contracts. It also claims the right to exclude from its total income the parts which represent shares of forfeitures, upon the grourid that the sums forfeited, and which are divided among the persisting policy holders, represent redundancies in the premiums paid by those policy holders who forfeited their right to the return of such redundancies. Hence it is argued that, under its construction of the words “portion of any actual premium,” such [92]*92sums are portions of actual premiums theretofore received, and deductible as such, irrespective of whether they originally came from the policy holder to whom the dividend is paid.
Under the construction which we have given*to the statute, as long as the dividends paid or credited to any individual policy holder within the tax year do not exceed the amount paid by way of premium by that policy holder during that year, it is quite immaterial from what source or sources the dividend paid or credited arose. Hence the independent question regarding the right to exclude such parts of the dividends as represent shares in the forfeitures of other policies is of no moment. However, as it is sought to procure credit, through the “deduction clause,” for the parts which represent interest, it is necessary to ascertain whether such interest payments fall within that clause. Under all of the policies involved in this suit, with one exception, the rate of interest allowed to policy holders on redundancies which were left with the company, and not withdrawn by the policy holder every year, was, in reality, dependent upon the earnings of the company upon its investments. Without attempting to decide, as before stated, because we find it unnecessary to do so, whether the word “dividend” is used in the deduction clause in its commercial sense, as distinguished from the insurance sense, we think that such interest payments are dividends in the commercial sense, as they clearly are in the insurance sense. They represent the policy holders’ pro rata share in the earnings on the company’s investments over a given period of time, and, as respects most of the policies involved in this suit, they are the result of compound interest. Hence they may not be deducted under the deduction clause.
[4] The one exception above referred to is the interest paid on what have been termed in the testimony “trust certificates.” These moneys are not paid to policy holders, but to the beneficiaries under installment policies, which provide for the payment of the face of the policy, not in bulk on the death of, the assured, but in installments extending over a period of. years. The fund thus left with the company, which, of course,' diminishes from installment period to installment period, draws interest, and the interest is then paid to the person entitled to the installments. It is accordingly, in neither an insurance nor a commercial sense, a dividend, but is a sum paid on a policy contract, and as such we think deductible under the “deduction clause” of the act. It appears from the record that, although all 'of the dividends in controversy were paid in cash to the 'policy holders, the company, during the tax year in question, received premiums that year from some of the policy holders to whom such dividend payments were made. The government concedes the right of the company to recover the tax paid on such part of the dividends paid to policy holders during the tax year as does not exceed the premiums received that year from the same policy holders, respectively, as it does the tax on an item of $2,774.35, which was inadvertently not included in the original return of the company. The actual amount which the government concedes that the plaintiff is entitled to recover is, however, we think, excessive, because it includes the sums received from holders of “life rate endowment policies” and holders of “ac[93]*93celerated endowment policies,” which were considerably more than the sums paid to them, respectively, during that year by way of dividends. Manifestly the company is only entitled in those cases to credit for the amount paid.
As the aggregate amount received by way of premiums from policy holders of the policies in question during the tax year was much less than the amount paid to them by way of dividends, and as judgment was awarded in the court below for the tax assessed on the total amount of dividends paid, it follows that the judgment must be reversed, and a new trial granted.