Tbe problem confronting us here is whether an importation of merchandise, described on the consular invoice as “Plus Adding Calculating Machines,” is within the provision for “Calculating machines specially constructed for multiplying and dividing” in paragraph 372 of the Tariff Act of 1930 (19 U. S. C. § 1001, *271par. 372), as modified by the Torquay Protocol to the General Agreement on Tariffs and Trade, 86 Treas. Dec. 121, T. D. 52739, supplemented by Presidential notification, 86 Treas. Dec. 265, T. D. 52763, and dutiable at 12% per centum ad valorem, as claimed by plaintiff; or whether it is subject to classification as “Other machines” within the provision of said paragraph 372, as modified by the General Agreement on Tariffs and Trade, 82 Treas. Dec. 305, T. D. 51802, and dutiable at 15 per centum ad valorem, as classified by the collector of customs.
The pertinent text of the competing provisions of paragraph 372, supra, is here set forth:
Paragraph 372, as modified by T. D. 52739 and T. D. 52763, supra—
Machines, finished or unfinished, not specially provided for:
Calculating machines specially constructed for multiplying and dividing- 12J4%adval.
Paragraph 372, as modified by T. D. 51802, supra—
Machines, finished or unfinished, not specially provided for:
Machines for packaging pipe tobacco; machines for wrapping cigarette packages; machines for wrapping candy; and combination candy cutting and wrapping machines * * *
* * * * * * *
Other * * *_15% ad val.
It is not disputed that the imported devices consist of key-drive calculating machines, and our sole inquiry is whether they are such calculating machines as are specially constructed for multiplying and dividing.
At the trial, the following exhibits were introduced by plaintiff and received in evidence:
Exhibit 1, representing imported model 909.
Exhibit 2, the same as exhibit 1, disassembled so as to expose its operating parts.
Illustrative exhibit 3, Monroe rotary-type calculating machine.
Illustrative exhibit 4, adding listing machine.
Illustrative exhibit 5, half keyboard machine.
Illustrative exhibit 6, photograph of Millionaire calculating machine.
Illustrative exhibit 7, photograph of “Mercedes” Euldid calculating machine.
Illustrative exhibits 8, 9, and 10, parts of imported machine.
Exhibit 13, page 6 of “U. S. Government Price List” for Monroe calculating, adding, and bookkeeping machines, insofar as it represents a photograph of the model L 200-X machine.
Collective exhibit 11, exhibit 12, and collective exhibit 14 were marked for identification only.
*272Defendant introduced exhibit A, pamphlet circulated by plaintiff corporation.
At the hearing, six witnesses were called, three of whom testified on behalf of plaintiff and three for the defendant.
Plaintiff’s first witness, Henry R. Mathieu, testified that he had been president of the plaintiff company since 1952, having been connected with it since its incorporation in 1949, being “responsible for the overall supervision of the marketing of the equipment in the United States; that is, sales, service, operator training, advertising, preparation of training material for the operators, and service training of salesmen; also, contacts with large corporations.”
Plaintiff’s second witness was Don R. Hogue, for 26 years promotion agent for Burroughs calculators. Prior to his association with Burroughs, he was, for 8 years, a salesman with the Felt & Tarrant Manufacturing Co., producer of computer adding and calculating machines.
Plaintiff’s third witness, Edgar Noll, testified that he was “known as an independent distributor of adding, and calculating machines of all types, covering the entire field as generally known in the industry as listing adding machines, rotary calculators, key driven calculators, and printing calculators,” having been in that business nearly 35 years; that he had handled every type of adding and calculating machine “that is known to anybody.”
The first witness for defendant was Ridgely D. Bryan, the educational director of the Monroe Calculating Machine Co. for approximately 25 years. His duties included the sales and promotion of the company’s equipment to schools and colleges; that the business of the Monroe Calculating Machine Co. was the manufacturing and selling of adding machines, accounting machines, and calculating machines. From 1918 to 1928, he was a salesman and branch manager in two different offices of the Monroe company, and, from 1912 to 1918 (except for a period in the Army), was sole New York representative for the Barrett Adding Machine Co.; that, while in the Army, he was in charge of a bureau of 30 comptometer operators in the Quartermaster Corps; and, prior to 1912, was in the accounting department of the Standard Oil Co., in charge of the Millionaire Calculating Machine Bureau, which Millionaire calculating machine he had operated in a commercial way.
The defendant’s second witness was Carl A. Zeller, New York manager of Felt & Tarrant Manufacturing Co., a concern engaged in the manufacture of the comptometer calculating machine. He had been associated with that company for 34% years, first as a service man, then as salesman, and finally as manager.
Defendant’s third witness, Harry A. Hicks, was vice president of the Remington-Rand Co., manufacturer and distributor of adding, *273calculating, accounting, and tabulating machines and of electronic equipment. He had been associated for 34 years with the Remington-Rand Co. and its predecessor, the Dalton Adding Machine Co., where he had put in a complete accounting system. From 1923 to 1927, he was assistant to the vice president in charge of sales. From 1927 to 1930, he was sales manager of the Dalton division, and, in 1933, became general sales manager of the “Adding, Calculating, and Adding and Accounting Division of Remington-Rand,” which position he held until 1943. From 1949 to the present, he had been employed as vice president of the Remington company.
The testimonial record before us is voluminous, and it would prolong this opinion unduly to attempt to set forth a detailed analysis of it. However, reference will be made to some of the salient features of the testimony, wherever deemed necessary.
We have been greatly aided by the carefully prepared briefs of adversary counsel and of amicus curiae in a rather complicated case.
As indicated above, the witnesses were men of wide experience with an intimate knowledge of the subject controversy. Several of the witnesses were asked to express their opinion as to whether exhibit 1 was specially constructed for multiplying and dividing, but, upon that question, there was a sharp divergence of view. It was not disputed by the witnesses that exhibit 1, representing the importation, and exhibit 3, the Monroe rotary type of machine, are calculating machines; that exhibit 3 is specially constructed for multiplying and dividing; and that both of those machines can be used to add and subtract.
In its brief, plaintiff sets up the “basis” of its claim herein this way:
We claim that a calculating machine is a mechanism that performs one or more of the functions of arithmetical calculations — adding, subtracting, multiplying or dividing — and that the President and the trade agreement negotiators had this kind of mechanism in mind when the duty was lowered only on those calculating machines that were constructed to multiply and divide. The higher duty (15%) remained in effect for such devices that were constructed primarily to add, or to add and subtract. Any calculator that is designed to multiply and divide must be capable of adding and subtracting because multiplication is repeated addition, and division is repeated subtraction. [Italics supplied.]
Plaintiff relies upon evidence to the effect that, since exhibit 1 is so constructed that it will multiply and divide efficiently and speedily, it is, ipso jacto, a calculating machine “specially constructed for multiplying and dividing.”
Reference is also made to evidence indicating that fewer motions are required to do a simple multiplication and division on exhibit 1 than on exhibit 3 and that all manually operated calculating machines, which are designed to be used commercially for multiplication and division, are specially designed to multiply and divide.
*274It will be observed that reference has been made to “simple” multiplication and, later, in its brief, plaintiff points out that the key-drive type of calculator, such as exhibit 1, has certain advantages over the rotary type, exhibit 3, for multiplying and dividing purposes in that “it is much faster for small computations.” [Italics supplied.]
Plaintiff admits, however, that “On the other hand the rotary type discloses all factors for checking purposes, and does not require the same degree of skill on the part of an operator.”
Plaintiff’s witnesses took exception to the statement contained in the New International Encyclopaedia, second edition, volume 4, at page 307, that:
* * * the multiplying and dividing machine will also add and subtract, but so slowly that it will never be used commercially for such work. * * *
But, as stated by the witness Noll:
* * *.We can never conceive in a general way of multiplying and dividing exclusively unless we get into a 'particular department where that is the primary object; therefore, all machines that are sold to do multiplication and division, every one of them must be practical for addition and subtraction, and every one that I know anything about, regardless of the dictionary, is practical for addition and subtraction; * * * [Italics supplied.]
Defendant contends that, whereas exhibit 1 and exhibit 3 both perform the four arithmetical functions of adding, subtracting, multiplying, and dividing, nevertheless, exhibit 1 basically adds, regardless of the function being performed; that division and subtraction are accomplished on exhibit 1 by a method of “complements,” which will be explained later.
Since it is admitted by the parties litigant that exhibit 3 is a calculating machine that is specially constructed to multiply and divide, it is interesting to note from the record that exhibit 3 has certain features not present in exhibit 1, namely:
1. A movable carriage.
2. Reverse operation of the mechanism itself.
3. Division is based upon completely mechanical principles, including a bell to indicate oversubtraction.
4. A keyboard which will hold one of the factors in it for checking.
Defendant grants arguendo that, since exhibit 1 may multiply efficiently and accurately, it might be regarded as having been constructed for multiplication, but points out, moreover, that the provision of law relied upon by plaintiff requires that a machine be specially constructed not only for multiplication, but for division as well.
Furthermore, defendant states in its brief that:
* * * It is clear from the record that the complementary system must be used to divide. This is a “system” and not a construction.
*275Tlie printing of large and small numbers on the keys could hardly be considered a construction. Nor are any of the other features used to facilitate division a construction in the true sense of the word as the term was used by the trade negotiators.
We pause Rere to consider what is meant by the term “complementary system,” which is not clearly defined in the record.
By reference to Webster’s New International Dictionary, second edition, 1948, page 546, we note the following:
complement, n. * * *
* * * % # * *
8. Math. * * * c Of a number, the numerical amount that must be added to the number to give the least number containing one more digit; as, the complement of 4 is 6, and that of 45 is 55.
The calculation of subtraction can be performed on an adding machine by adding the number’s complement and striking off the added digit. We give the following as an illustration:
Problem
Subtract 3 from 9.
Solution
The complement of 3 is 7.
Add 7 to 9=16.
Remove the added digit 1, leaving 6.
Explanation
By definition, a complement is the amount that must be added to a number to give that number an added digit. In effect, it is what changes units to tens, or tens to hundreds, et cetera. Expressed differently, the complement is equal to 10 (or 100 or 1,000), minus the number itself.
By adding the complement (i. e., 10 minus the number) .and then removing the added digit 1, a person is really doing three things:
1. Adding 10,
2. Subtracting the number itself,
3. Canceling the 10.
The net result is equivalent to subtracting the number itself.
See also the New International Encyclopaedia, second edition, volume 4, at page 307.
We believe that the meaning of the term “specially constructed for multiplying and dividing” is so shrouded in doubt as to permit us to resort to extraneous aids for possible enlightenment.
The provision for “Calculating machines specially constructed for multiplying and dividing” does not appear in the basic Tariff Act of 1930. It made its first appearance in the trade agreement with Sweden, 68 Treas. Dec. 19, T. D. 47785, entered into in 1935.
The Digest of Trade Data relating specifically to that trade agreement contains the following, at page 133:
The calculated [sic] machines affected by reclassification are confined to those constructed essentially for multiplying and dividing. Such machines can be used *276to add or subtract, although these operations are usually more readily performed on adding machines.
* * * * * * *
Adding machines are not included under the new classifications for the reason that although many of them are capable of being used for multiplying and dividing, they are not constructed essentially for the performance of these operations. [Italics supplied.]
It will be observed that the matter above quoted from the Digest of Trade Data is quite in harmony with the views expressed in the New International Encyclopaedia, to which reference has been made, sufra.
From the foregoing considerations, it would appear that the calculating machine, represented by exhibit 1, was not the type of machine which the trade negotiators had in mind as a calculating machine specially constructed for multiplying and dividing.
The only other case wherein it was claimed before this court that an imported machine was a calculating machine specially constructed for multiplying and dividing is Addicalco Corporation of America v. United States, 31 Cust. Ct. 329, Abstract 57675. In that case, a machine, which performed the four arithmetical functions of adding, subtracting, multiplying, and dividing, was held not to be a calculating machine specially constructed for multiplying and dividing.
Carl A. Zeller, New York manager of Felt & Tarrant Manufacturing Co. of Chicago, who testified in that case, was also called as a witness in the present case. As indicated above, the record before us discloses that Zeller has been with Felt & Tarrant, manufacturer of the comp-tometer calculating machine, for 34% years, having first been employed as a service man, then as a salesman, and, finally, in his present capacity. He testified that the comptometer is basically the same type of machine as exhibit 1. Based upon his knowledge and experience, he stated that exhibit 1 is an adding calculating machine and that he used the term “adding calculating machine,” insisting that all the machine does is to add by a method, and, for that reason, he was of the opinion that exhibit 1 was not mechanically specially constructed to multiply and divide. He pointed out that there were certain features on exhibit 1 which enabled one to multiply and divide, but reiterated that, in so doing, the machine was merely adding.
Zeller also testified that he was familiar with the basic mechanics and operation of exhibit 3 which, he stated, was specially constructed for multiplying and dividing. Again referring to exhibit 1, Zeller stated that, in the act of division, it is the operator who performs the calculation rather than the machine, because of the mental requirements in applying the complementary method for the use of the machine, which is not required in the rotary type of machine, such as exhibit 3.
*277Zeller’s attention was drawn to his testimony in the Addiealco case, supra, in which reference was made to the Millionaire and the Mercedes calculating machines, illustrated by exhibits 6 and 7, respectively. He was familiar with their functioning, having operated them in the Solvay Process Co. plant in Syracuse, where he was employed in the thirties. Zeller regarded those two machines as constructed specially for multiplication and division, but as not being commercially adapted for the purpose of addition and subtraction.
Plaintiff’s witness Hogue also testified that the Millionaire and Mercedes machines were primarily used for multiplication and division, because their construction is especially adapted for that purpose.
Defendant’s witness Bryan testified that, prior to 1912, he was in the Accounting Department of the Standard Oil Co., in charge of the Millionaire Calculating Machine Bureau; that he is familiar with that machine, which is represented by exhibit 6, and had commercially operated it; that it multiplies and divides faster than any machine that was ever built; that he is also familiar with the Mercedes calculating machine, represented by exhibit 7, which, he stated, multiplies and divides rapidly, but that neither the Millionaire nor the Mercedes is used for addition and subtraction, because the process would be too slow.
Based upon the record before us, we are of the opinion that plaintiff has failed to show by a preponderance of evidence that the subject calculating machines are “specially constructed for multiplying and dividing,” as contemplated by the provision in paragraph 372, as modified, supra.
The protest of plaintiff is, therefore, overruled and judgment will issue accordingly.