posed. It is true that the present scheme is, as you say, far more vicious in its tendencies, for the reason that the sums which may be raised for distribution may be large, as compared with the small prizes offered by the Cincinnati Enquirer, and, therefore, practically offer to the public the allurements of a lottery. Moreover, in the case of the Cincinnati Enquirer, the prizes were gratuitously given by that newspaper, while, in this case, the sums to be distributed are raised by the subscriptions of those who participate in the scheme, thus assimilating it to a lottery.

Nevertheless, the question must, in the present case, as in the case which was the subject of the opinion referred to, depend upon a reasonable construction of the statute which forbids the use of the mails for "any lottery, so-called gift concert, or other similar enterprise offering prizes dependent upon lot or chance." The prizes to be distributed in the present case are obviously not dependent upon lot, and the only question to be determined is whether they are dependent upon chance within the meaning of the statute. As Attorney-General Miller said in the opinion above referred to:

"In a certain sense and in a certain degree, perhaps, any prediction as to human action may be said to be dependent upon chance; that is to say, it is in some measure dependent upon circumstances, the happening of which can not be anticipated or foretold with any degree of certainty. But, at the same time, it can not be said that a prediction that a man who has lived a life of uprightness for fifty years will, during the remainder of his life, continue so to live, or that a man who has been a successful business man for fifty years will so continue, or that a man who has maintained certain opinions-religious, political, or economical-will continue in the same line, is dependent upon chance. It is, of course, quite possible that such man may utterly change his habits of life, business, or opinions, but such change will not be purely matter of chance. So with regard to the case in hand. A student of statistics might know approximately the number of Republican votes and the number of Democratic votes in the State of Ohio; he might approximate the ratio

in which one and the other might increase or decrease in a given year. It is quite likely that his estimates would often be wide of the mark, but it would not be by reason of chance, but by reason of causes in regard to which he had formed erroneous estimates. It would hardly do to say that a child or a schoolboy could form as correct an estimate in the matter as an experienced politician who had been giving weeks and months of steady attention to the consideration of the question."

The question therefore arises whether a similar contest, in which the prizes are dependent upon the accuracy with which the participant estimates the majority which the successful candidate will receive at the coming Presidential election, is distinguishable in principle from a scheme offering prizes for the nearest estimate to the majority to be given in an election for the office of secretary of state of the State of Ohio. As I have intimated, the method by which the money for the distribution of prizes is secured is unimportant, even though it make the present scheme both vicious and dangerous in its tendencies. Similarly, the amount of money to be distributed can not affect the question, though it is also true that the possible distribution, in a scheme of this character, of large sums of money makes the present scheme far more objectionable than the comparatively small prizes given by the Cincinnati Enquirer. Neither of these considerations in any way affects the question whether the distribution of prizes is "dependent upon lot or chance."

Ordinarily it would be more difficult to estimate the majority of a successful candidate for the Presidency than to estimate a similar majority at a State election for a State office. This, however, is not necessarily or invariably the case. There are many State elections in which the conditions change so rapidly from day to day that an estimate is quite as difficult as an estimate in a Presidential election, where the number of votes is much greater. In both cases the vote is so large that nothing is possible but an approximation. The difficulty of such approximation, as between the two cases, is one of degree only, and in both the elements of calculation, foresight, knowledge, inquiry, and 19395-VOL 23-02-14

information play no inconsiderable part. The trained observer of political events is far more likely to approximate the number than one who has not studied election statistics or past political struggles. Upon the other hand, it can be said that the exact number is so far beyond the power of human calculation that, as between competitors in such a contest of equal knowledge, the successful guess must necessarily be wholly a matter of chance. In other words, if two men, equally familiar with election statistics and the past and current movements of political parties, should attempt to make estimates, the question as to which would be the nearest to the exact number would no longer be a matter of calculation and reflection, but would be wholly dependent upon chance.

Notwithstanding this, I find it impossible to distinguish the case in principle from that which was the subject of Attorney-General Miller's opinion; and as the case is not free from doubt, and as questions of doubtful construction of criminal statutes should be resolved in favor of the liberty of the subject, I am constrained to adhere to the opinion of my distinguished predecessor. I do this with a full appreciation of the vicious and dangerous tendencies of the present scheme, to which you have with reason referred; and if this scheme should be used to an extent that would make it a public evil, I should not hesitate to advise such amendments to the lottery act as would bring similar schemes within the letter of the prohibitory statutes. I have the honor to remain, very respectfully, JOHN W. GRIGGS.

The POSTMASTER-GENERAL.

WAR-REVENUE ACT-TAX ON INSURANCE POLICIES.

The act of June 13, 1898 (30 Stat., 448), requires the stamp to be affixed to the policy of insurance and not to the preliminary application, although such application is expressly made a part of the contract of insurance.

The term "policy of insurance" as used in that act is a technical phrase, and ordinarily applies to the specific instrument by which the company agrees to pay a certain amount upon conditions therein stated.

Where the application for insurance expressly stipulates that the policy shall embrace four separate contracts, covering four consecutive periods aggregating one year, and shall remain in force after the first insurance period only as continued by further payments of premium, such policy is held to be a policy issued for one year, and the amount of tax- to be affixed when the policy is delivered is to be determined by the aggregate of the premiums for the entire year.

DEPARTMENT OF JUSTICE,

September 5, 1900.

SIR: I have the honor to acknowledge the receipt of your letter of December 2, 1898, transmitting a copy of a letter of the Commissioner of Internal Revenue dated November 30, 1898, in which you request my opinion as to the proper method of stamping certain applications and policies of installment accident insurance issued by the Travelers' Insurance Company. You state that the party desiring the insurance makes a written application to have the policy written for four consecutive periods, of two, two, three, and five months, respectively, each being covered by a distinct premium to be paid, respectively, from his wages. He simultaneously gives an order upon the paymaster of the corporation by which he is employed, whereby he assigns a sufficient sum from his wages to pay these four premiums. After receiving the application the Travelers' Insurance Company issues a single policy to the insured for the principal sum mentioned as the full amount of insurance in the application. The application expressly stipulates that "the policy shall embrace four separate insurance contracts, and shall remain in force after the first insurance period only as continued by payments of premium for the consecutive periods following;" and the assignment blank contains the statement that the payments named in this assignment are premiums for separate and independent contracts for consecutive periods of two, two, three, and five months; and each shall apply only to its corresponding insurance period.” The insurance company claims the right to affix the revenue stamps required by the act of June 13, 1898, upon the application instead of the policy, and it contends that the amount of the stamp shall be determined separately for each of the four successive periods during which the contract runs. The Commissioner of Internal Revenue, on the other hand,

has ruled that, although the payments are made in installments, the entire amount of the premium is and has been fully determined at the time of issuing the policy, and as the policy is for a stated period and a fixed premium, it should be stamped when delivered for the four terms of insurance included therein.

Two questions, therefore, are suggested by your inquiry: First. Should the stamp required by the act of June 13, 1898, be affixed to the policy of insurance, or to the application for such policy?

Second. Should the amount thereof be such as is required by the entire premiums to be paid during the period of one year for which the policy of insurance runs, in the aggregate, or may the stamp duty be divided into four parts and separately affixed with each successive payment during the periods in question?

As to the first question, I would advise you that the act of Congress requires a stamp to be affixed to the "policy of insurance, or other instrument, by whatever name the same shall be called." It is true that in the present case the application is expressly made a part of the policy, and it is therefore true that the application and the assignment of the wages do form a part of the entire contract of insurance. Nevertheless, the term "policy of insurance," as used in the act of Congress, is a technical word with a wellascertained meaning, and ordinarily applies to the specific instrument by which the company agrees to pay a certain amount upon conditions therein stated. The application does not of itself insure, and a policy of insurance may not be granted as applied for. To the extent that its statements, if false, may invalidate the policy, it forms a part of the contract of insurance, but it is not commonly regarded as the policy, but only an instrument of writing preceding and accompanying it. I am therefore of opinion that Congress intended that the stamp should be affixed to the policy of insurance and not to the preliminary application.

As to the second question, there can no longer be any doubt, in view of the case of Buckalew v. United States (102 Fed. Rep., 320), in which the circuit court of appeals for the fifth circuit decided upon facts which are precisely