delivered the opinion of the court:
This is an appeal from the judgment of the United States Customs Court, Second Division, awarding the importer recovery of the duties assessed upon certain cotton cloth by the Collector of Customs at the port of Philadelphia, Pa., under paragraph 924 of the Tariff Act of 1930.
The merchandise was classified basically under paragraph 904 (a) of the act. This basic classification is'agreed to have been correct and the duties levied in conformity with that paragraph are not in-*242dispute. The collector, however, seems to have been advised by the appraiser that the warp threads of the cloth were manufactured from cotton having a staple of more than 1% inches in length and that such warp threads (after allowing 2.1 per centum in weight for sizing) constituted 35.2 per centum in weight of the cloth. So, there was assessed and collected an additional duty of 10 cents per pound (the collector invoking the provision of paragraph 924) upon such 35.2 per centum. No additional duty'was assessed upon the weight (64.8 per centum) represented by the weft threads, apparently because the collector found that such threads had been manufactured from raw cotton having a staple length of less than 1 % inches.
As we understand it, the warp threads in a woven article such as the cloth here involved are those which, in the weaving process, are extended lengthwise in the loom to form the basis of the article, and the weft threads are those which (carried by a shuttle or other means) are woven across the warp threads by inserting them over and under the warp threads. The weft threads are frequently referred to in the record before us as filling threads.
The brief on behalf of appellee states:
We have not challenged the correctness of the weight of the warp threads (35.2%) or the weft threads (64.8%) as found by the Government, nor have we challenged the correctness of the weight of the sizing as found in the warp threads (2.1 %); the weight of the sizing being properly deducted.
Schedule 7 of the Tariff Act of 1930 provides duties for agricultural products, etc., and paragraph 783 of the act appearing in said schedule reads:
Pab. 783. Cotton having a staple of one and one-eighth inches or more in length, 7 cents per pound.
The act makes no provision for duty on cotton having a staple-length of less than 1 % inches.
Schedule 9 of the act makes provision for cotton manufactures, and embraces, among others, the following paragraphs:
Pab. 903. * * *. (b) In the ascertainment of the condition of the cloth or yarn upon which the duties imposed upon cotton cloth are made to depend, the-entire fabric and all parts thereof shall be included.
Pab. 904. (a) Cotton cloth, not bleached, printed, dyed, or colored, containing yarns the average number of which does not exceed number 90, 10 per centum ad valorem and, in addition thereto, for each number, thirty-five one-hundredths-of l’per centum ad valorem; exceeding number 90, 41K per centum ad valorem: Provided, That none of the foregoing shall be subject to a less duty than fifty-five one-hundredths of 1 cent per average number per pound;
Pab. 924. All the articles enumerated or described in this schedule (except in-paragraph 922) shall be subject to an additional duty of 10 cents per pound on the cotton contained therein having a staple of one and one-eighth inches or more in length.
*243Examination of the legislative history discloses that the duty fixed in paragraph 924 was provided as a compensatory duty for that levied upon the raw long-staple cotton defined in paragraph 783, supra.
During the trial of the case the importer introduced the testimony of six witnesses, and the Government that of two. Numerous exhibits were also placed in evidence.
In the brief on behalf of the Government before us it is said that determination of whether the duty of 10 cents per pound under paragraph 924 was properly assessed involves:
* * * (1) whether such duty is based upon the staple length of the cotton from which the article is made; (2) whether the collector considered the entire article and all parts thereof; (3) whether the procedure of ascertaining the staple of the cotton was correct.
The first ruling of the trial court related to the second of the above-quoted questions. Upon this the court said, in part:
It is conceded that the cotton cloth here involved consists of a warp and weft or filling yarn. Therefore the “entire fabric and all parts thereof” includes both the warp and the weft or filling yarn, and does not include only the warp yarn. In order for the collector tó comply with the clear mandate contained in paragraph 903 (b), in classifying the present cotton cloth, he was required to include “the entire fabric and all parts thereof,” and not having followed the congressional mandate his action was clearly erroneous. On the admission and concession that the collector in classifying the instant merchandise did not include “the entire fabric and all parts thereof,” as a matter of law his classification was wrong, and all presumption in favor of his classification is overcome and disappears.
If we understand the foregoing ruling aright it must have been the view of the trial court that in determining whether paragraph 924 is applicable to cotton cloth it is necessary to consider all the yarns or threads, both warp and weft, which enter into- the composition of the cloth, and that unless there be a preponderance of threads having a staple length of more than 1% inches the paragraph is not applicable to any portion of the cloth.
We agree that there must be a determination of the staple length of all the threads in the cloth (this, of course, to be done by an examination of representative samples thereof), but we do not agree that paragraph 903 (b) should be construed to mean that in order to render paragraph 924 applicable there must be in the manufactured article a preponderance of threads in which the staple length is 1% inches or more.
If that should be adopted as a'rule it would appear that where there is a preponderance of threads in the whole cloth in which the staple length is less than 1% inches, although there may be a large proportion of threads composed of the longer staple, no duty would be levied, and, on the other hand, should the threads containing a staple length of 1% inches or more predominate, a duty would be levied upon the whole. *244although, the cloth might contain a large proportion of threads made from short-staple cotton. It seems to us that a practice'of this character would not conform to the obvious purpose of the act.
No duty was levied in the Tariff Act of 1930 upon raw cotton having a staple length of less than 1% inches and there was no occasion to levy a compensatory duty upon articles made from cotton having a staple shorter than 1% inches. It appears to be a common practice, or at least not an unusual practice, to utilize yarns made from both short and long-staple cotton in the manufacture of cloth — that is both types are woven into the same piece of cloth — as was the case in the merchandise here involved, and it seems to us clear from the phraseology of paragraph 924, sufra-, that it was intended to provide the compensatory duty upon such portions of such imported articles as were made from the long-staple material, and upon those portions only. This of necessity requires segregation and, while the entire article must be examined to determine the staple length of its component threads, in our opinion, the act did not intend that the duty provided in paragraph 924, sufra., should be levied, or not levied, on the entire article upon the basis of the preponderance of the respective staple lengths in all the threads, both warp and weft, composing it.
For the reasons stated, we find ourselves in disagreement with what we understand to be the view of the trial court upon this phasenf the case, and, in this connection, it may be said that it is clear from the record in the instant case that the collector in applying paragraph 924, sufra, did endeavor to make segregation. Both the warp and weft threads, seemingly, were analyzed and no duty was levied under paragraph 924, sufra, upon the weight represented by the weft threads. The' question of whether he correctly analyzed and measured the warp threads will be discussed later.
With respect to the first question above quoted from the Government’s brief relative to the staple length of the cotton material in the cloth, it seems to have been the view of the trial court that the collector determined the staple length of the staple in the threads upon the basis of the staple length of the raw cotton from which the threads were spun rather than upon the-basis of the threads after they had been spun and woven into -the cloth,- and the court held this to be erroneous, saying, inter alia:
* * *. In paragraph 924 the additional duty is made to depend entirely upon the length of the staple of the cotton “contained therein,” and not upon the staple of the cotton of the kind provided for in paragraph 783. There is no raw cotton contained in the imported merchandise. The only cotton “contained therein” is manufactured cotton, and it is thé cotton “contained therein,” and not the cotton provided for in paragraph 783, upon which the Congress made the additional duty to depend. It is therefore clear that the additional duty provided for in paragraph 924 is made to depend upon the length of the staple of the cotton *245as it appears (therein) in the imported merchandise, and not upon the staple of the raw cotton from which the articles were manufactured, of the kind provided for in paragraph 783.
Moreover, another strong, if not conclusive, indication that Congress intended that the staple of the cotton should be determined by its length in the cloth and not the staple length of the raw cotton from which it was made, is the fact that in the enactment of paragraph 924, Congress did' not levy the same rate of duty on the staple length of the cotton contained therein having a staple of 1)4 inches or more in length as it did on the raw cotton in paragraph 783. It levied a duty of 10 cents per pound on the cotton in the cloth having a staple of 1)4 inches or more in length and only 7 cents per pound on the raw cotton in paragraph 783, which clearly shows that Congress knew that the staple length 'of the cotton in the cloth would necessarily be some shorter than the 'staple length of the raw cotton from which the cloth was made, and so put a higher rate of duty on the length of the staple in the cloth in order to procure a just and correct compensatory duty on the cotton cloth.
In its discussion of tliis question the trial court referred to the decision of this court in the case of Morse Bros. (Inc.) v. United States, 13 Ct. Cust. Appls. 553, T. D. 41432, and stated its reasons for distinguishing that case from the instant case.
That case arose under the Emergency Tariff Act of 1921 which for the first time, so far as wo are informed, provided for the imposition of a duty on imported raw cotton upon the basis of staple length.2
Paragraph 16 of that act read:
16. Cotton having a staple of one and three-eights inches or more in length, 7 cents per pound.
Paragraph 17 provided a compensatory duty as follows:
17. Manufactures of which cotton of the kind provided for in paragraph 16 is the component material of .chief value, 7 cents per pound, in addition to the rates of duty imposed thereon by existing law.
In construing the foregoing paragraphs this court held, in effect, that the cotton provided for in paragraph 16 comprised cotton in which the fibers (the term “fiber” being there used interchangeably with the term" staple”) as long as 1% inches predominated, and that in determining whether or not the cotton cloth which was there involved was classifiable under paragraph 17 the inquiry was whether or not fibers shorter than 1% inches predominated in the raw cotton from which the cloth was made. In other words, it was there ruled that under the language of the statute the proper test was that of the length of the fibers in the raw cotton and not the length of the fibers in the threads after the latter had been woven into the cloth, it being *246said, as expressed in the syllabus, that “Paragraph 17 levies duty on manufactures of which such [long-staple] cotton, not such fibers, is the component material of chief value.”
It will be observed that paragraph 783 of the Tariff Act of 1930, except as to the matter of staple length, is identical with paragraph 16 of the Emergency Tariff Act of 1921, but that the language of paragraph 924 of the 1930 act differs somewhat from the language of paragraph 17 of the emergency act and that the rate is different. The change in language was emphasized in the trial court’s decision in the instant case and the conclusion was reached to the effect that under the existing law it is now required that the length of the fibers or staple in the threads woven into the cloth must be determined by measuring such fibers and not by the measurement of the fibers, or staple, of the raw cotton from which the threads were spun.
We feel constrained to disagree with the construction so made by the trial court, it being our view, based’ in part at least upon the legislative history of the pertinent paragraphs, that it is contrary to the intent of Congress.
It may be said that the Tariff Act of 1922 repealed the Emergency Tariff Act of 1921, and that no duty was provided for raw cotton of any character in the 1922 act, nor was any duty provided upon cotton cloth based on the length of the staple of cotton from which such cloth was manufactured.
Paragraph 783 of the Tariff Act of 1930 originated in the United States Senate while the bill which matured into the law was there being considered. It was offered as an amendment by Senator Shortridge of California (see Cong. Record, 71st. Congress, 2nd. Sess., Yol. 72, Part 4, page 4415) and adopted after spirited debate. The whole tenor of the debate shows clearly that the amendment was designed and intended to protect the growers of cotton of the kind described therein against such competition as might arise from the importation of similar raw cotton from foreign countries. During the discussion it was pointed out that if the amendment proposed by Senator Shortridge should be adopted, there would be required a further amendment to the bill so that compensatory duties would be provided respecting articles manufactured from the type of raw cotton involved in the Senator’s proposed amendment. The amendment so proposed having been adopted, there was subsequently offered by Senator Smoot of Utah (chairman of the Finance Committee of the Senate in charge of the bill on the Senate floor) an amendment which embodied the precise language of paragraph 924, supra. The amendment so offered by Senator Smoot was agreed to by the Senate and later by the conferees representing the respective legislative branches of the Congress and finally by the House. (For the *247Senate debate, etc., see Cong. Record, 71st. Congress, 2nd. Sess., Yol. 72, Part 5, page 4616, and subsequent pages. For the amendment offered by Senator Smoot see Cong. Record, 71st. Congress, 2nd. Sess., Vol. 72, Part 6, page 5833.)
It would seem obvious that to be a true compensatory duty the length of the staple of the raw cotton from which the thread was spun must be taken into account. A manufacturer in this country is required to pay a duty of 7 cents per pound on any raw cotton having a staple of 1% inches or more in length which he imports. If, in the process of manufacturing an article from such imported cotton, there be a loss in staple length (which the testimony in this case indicates often happens) the manufacturer receives no rebate of duty on that account. If, on the other hand, a foreign manufacturer uses long-staple cotton in the making of an article which is imported into this country and there be a loss in staple length in the process of manufacture, so that the staple in the article falls below l}g inches in length, the domestic manufacturer obviously would be placed at a disadvantage should the reduced length of the staple of the cotton in the thread be made the basis of duty rather than the staple length of the raw cotton from which the thread was made. As has been said, the duty provided in paragraph 924 was intended to be a compensatory duty in order to protect domestic manufacturers against the effects of paragraph 783. Except for the enactment of paragraph 783 there would have been no occasion for the enactment of paragraph 924 and there is no reason to suppose that it would have been enacted.
We have carefully considered the suggestion of the trial court, quoted above, respecting the conclusion which should be drawn from the fact that in the 1930 act the compensatory rate of duty on the manufactured article was fixed at 10 cents per pound while in the emergency act the rate was fixed at 7 cents per pound (the rate on the raw cotton — 7 cents per pound— being the same in both acts), but we are unable to agree that this has any particular significance in construing paragraph 924 in connection with paragraph 783 of the 1930 act. Both paragraph 17 of the emergency act and paragraph 924 of the 1930 act were intended and adopted for the single purpose of providing duties of a compensatory character, and the determination of the rate was, of course, a matter wholly within the legislative discretion and authority.
Furthermore, it appears as a part of the legislative history that after the respective amendments constituting paragraphs 783 and 924 of the act had been adopted by the Senate, the United States Tariff Commission in a “Supplement to Tariff Information” on items in the bill subject to conference, page 387, discussed those amendments as follows:
*248Comment on phraseology and rates. — In Schedule 7 (Agricultural Products and Provisions) the Senate added a new paragraph rea'ding:
Par. 783. Cotton having a staple of 1)4 inches or more in length, 7 cents per pound.
In Schedule 9 (Cotton Manufactures) the Senate added a new paragraph, No. 924, to provide a compensatory duty on long-staple cotton, such as is made dutiable under paragraph 783, when imported in the manufactured state.
Long-staple cotton, of 1)4 inches or more in length, is usually combed before spinning and'the waste averages about 30 per cent, that is, from 100 pounds of raw cotton there is obtained approximately 70 pounds of yarn. A duty of 7 cents a pound means $7 per 100 pounds and this divided by 70 gives a compensatory duty of 10 cents per pound. * * *.
When the fundamental purpose of paragraph 924 of the 1930 act is considered in connection with the fundamental purpose of paragraph 17 of the emergency act, we do not think the differences in language are such as to justify the conclusion that, upon the point immediately under discussion, paragraph 924 should be given a construction different to that given paragraph 17 of the emergency act in the decision of the Morse Bros. case, supra. See also the case of Wilkes-Barre Lace Mfg. Co. v. United States, 11 Ct. Cust. Appls. 519, T. D. 39664, decided May 24, 1923.
It may be said that this court’s decision in the Morse Bros. case, supra, was rendered February 25, 1926, and it and the decision in the Wilkes-Barre Lace Mfg. Co. case, supra, were known to Congress at the time it was considering the bill which became the Tariff Act of 1930. It would seem that if it had been the intention of Congress to provide a construction different from that made in those cases it would have used language differing more widely from paragraph 17 of the emergency act than does paragraph 924 of the 1930 act.
So, upon the questions of law thus far discussed we find ourselves in disagreement with the trial court.
Our holding in that regard, however, is not conclusive of the case.
There remains to be considered the question whether the staple length of the cotton in the warp threads of the imported cloth was l}& inches or more in length when tested upon the basis of the staple length of raw eotton such as that from which the threads were spun.
This involves the third query embraced in the quotation above made from the Government’s brief, viz., “whether the procedure [followed by the collector] of ascertaining the staple [length] of the cotton was correct.”
Elsewhere in the Government’s brief it is said:
We respectfully assert that the fundamental error in the lower court’s decision ■is that it failed to find the one fact upon which the issue here turned. The question in issue is: “What is the staple length of cotton contained in this cloth?” That question cannot be answered until an understanding is reached as to the meaning of the term “length of staple” as applied to cotton in cloth.
It is proper to say that Congress has never stated a legislative definition of the term “staple” as applied to cotton, nor has it ever *249stated a rule for determining the staple length of cotton for customs purposes, either with respect to raw cotton or articles made from raw cotton.
During the discussion of the amendment proposed by Senator Smoot, which became paragraph 924 of the act, the following colloquy occurred:
. Mr. Walsh of Montana: Mr. President, I should like to inquire of the Senator whether practically it is possible to determine in any fabric the amount of long-staple cotton as distinguished from the ordinary standard grade?
Mr. Smoot: Of Course it is, in a piece of cloth. It is very easily determined.
The Senator did not describe any method for such determination, however, nor is any defined in the law.
By authority given in different acts of Congress relating to domestic trade in raw cotton the Bureau of Markets of the Department of Agriculture has formulated cotton standards and adopted a method for determining the staple length of raw cotton and, so far as we are advised, the standards established and the method suggested have been generally accepted and followed in the domestic trade and, to a certain extent at least, they have been utilized by the customs officials in applying the provision for duty upon long-staple raw cotton. The Department of Agriculture, however, has had no occasion officially to deal with the matter of staple length respecting manufactures of cotton.
With respect to a definition of the terms “staple” and “staple length” as applied to raw cotton, we recite the following:
One of the witnesses called by the Government in the instant case was Mr. Harold C. Slade, who stated that he was with the Agricultural Marketing Service in the Department of Agriculture, being in charge of the preparation and distribution of official cotton standards, chairman of the Final Appeal Board of Review Examiners who supervise the classification of cotton, and on the Board of Cotton Examiners; that he had been with the Government 25 years; that for 8 years prior to entering the Government service he was engaged in different types of business relating to “raw” cotton, in which he classified “raw cotton,” stapling it to determine the length of the staple, and that in the course of his experience he had determined the staple length of more than a million bales of cotton.
He was asked, “Will you state what a staple of cotton is?” He replied, “It is a typical portion of the fibers of any kind.”
Concerning the definition of “staple length” the following appears in his testimony:
Q. Vou stated what a staple of cotton is. Will you state what the staple length of cotton is? — A. The staple length of cotton is the length' of a typical portion of the fibers.
Q. In other words, the length is the length of the staple? — A. That’s right.
*250There is no controversy here respecting the definitions of “staple” and “staple length” as applied to raw cotton thus stated.
As to the method, Mr. Slade testified, in substance, that in arriving at the staple length of raw cotton the fibers are not counted, nor is the average length of the fibers determined; that a staple of cotton can be arrayed by separating the fibers upon a piece of cloth with the long fibers at one end and the short fibers at the other; that such an array shows the “modal length,” or the length of greatest frequency of the fibers, and that this (not the average length) determines the staple length. A demonstration of stapling was made by the witness before the trial court in conformity with the plan outlined.
From the testimony of Mr. Slade, taken in connection with the official publications of the Department of Agriculture placed in evidence, we deduce that in stapling a bale of cotton in order to determine staple length, tufts in sufficient number to assure that they are representative of the bale content are pulled and analyzed, and that the staple length of a tuft does not mean that every individual filament shall be of a given length, but that it is determined by the length of those of greatest frequency in the tuft. For example, if filaments 1% inches or more in length are of greater frequency in the representative tufts or “pulls,” even though some filaments may be shorter, the bale will be classified as long-staple, but if there be a greater frequency of filaments shorter than 1 % inches the bale will be classified as short staple, even though some filaments may be longer.
This (although expressed in somewhat different 'language) was stated in our decision .in the case of United States v. General Rubber Co., 22 C. C. P. A. (Customs) 308, T. D. 47350, where an importation of raw cotton was involved.
Mr. Slade testified, in effect, that he had had no experience with respect to determining staple length in manufactures of cotton, and his testimony related solely to raw cotton.
The necessity for defining the terms “staple” and “staple length,” of raw cotton, and describing the method of determining such staple length in the' instant case is due to our view, as already set forth, that it is necessary to make comparison, in some manner, of the staple length of the fibers constituting parts of the threads woven into the cloth with the staple length of raw cotton.
No samples of the foreign raw cotton from which the threads were spun were available to the customs officials for examination and, in the absence of such, they, of necessity, had to resort to a comparison with samples of domestic raw cotton.
That was done in respect to the cloth which was involved in the Morse Bros. case, supra, but there was no clear statement there of the plan or method which was followed by the customs officials, and we felt *251constrained to hold that the importers had produced testimony which overcame the presumption of correctness attaching to the collector’s classifications.
The principal witness for the importers in that case was Dr. Hugh Byron Gordon, a chemist with the United States Testing Co., who had prepared and published in the Textile World of October 13, 1923, an article relating to cotton fiber length and stapling length as applied to cotton in fabrics. In our decision there we quoted much of his testimony respecting the method which he pursued. 'It accorded with that stated in his article. There were other witnesses for the importers, the substance of whose testimony was stated, and, after a discussion of all the testimony in the case, we said:
While there is no recognized method of determining the staple of cotton after it is manufactured into cloth, Congress must have contemplated that some means could be devised for determining this fact. The testimony of the witnesses for the importers discloses a method used by them, which seems to be a reasonably fair one, and, while not capable of producing absolute accuracy, it would seem to be reasonably calculated to produce the result desired by Congress. The method pursued by the Government cannot be regarded as a fair effort to determine the staple in the cotton before it is manufactured, much less is it to be relied upon in determining staple from the cloth.
So, to the extent indicated and upon the facts there appearing, this court there approved the so-called “ Gordon method” which is described in our decision and need not be repeated in detail here. It does not follow, however, that the decision there with respect to the method of determining staple length is controlling here. It was not our intention to attempt to establish the “Gordon method” as an arbitrary standard, nor was it meant to suggest that there might not be some better method for determining staple length in cotton manufactures.
The importer here seems to have proceeded upon the theory that the customs officials followed the “Gordon method” with respect to the instant merchandise, and introduced testimony for the purpose of showing (1) that such method is not adequate to secure an accurate determination, and (2) to show what is claimed to be a more accurate method.
Whether the “Gordon method” (as same was described in our decision in the Morse Bros. (Inc.) case, supra,) was used by the customs officials in this case is not determinable from the record before us.
■ During the making of preliminary statements at the beginning of the trial there was introduced in evidence a sample of the imported merchandise which is identified as Exhibit 1.
The first witness called by the importer was Mr. Thomas J. Desmond an examiner of merchandise at the port of Philadelphia, who testified, in substance, that he submitted a sample (Exhibit 1) of the cloth to Mr. William 0. Beard, a chemist in the Division of Laboratories of the customs service at that port, for analysis, and that he advised *252classification upon the basis of Mr. Beard’s report, the sample itself not being returned to him. Mr. Beard was then called by the importer. He identified Exhibit 1, as the sample which he analyzed, and stated the conclusions which he reached relative to “yarn number,’’ “sizing,” “weight,” etc., but he was not asked any questions as to the method he used and gave no testimony in that regard.
By agreement of counsel a piece, marked “Exhibit 1-A,” was cut. from the official sample and turned over to a representative of the .firm of customs brokers who handled the clearance of the merchandise. The piece so taken was later analyzed, at the instance of the importer, by Mr. Alfred E. Davieau and Mr. Benjamin A. Holgate who were connected with the United States Testing Co., a company which engages extensively in testing textile, fibers, yarns, and fabrics for jobbers, retailers, manufacturers, and, in some instances, for consumers. After being analyzed by them the exhibit (at the instance of the importer) was sent to and analyzed by Mr. George Sherman Monroe, director in charge of laboratory operations of the Better Fabrics Testing Bureau which is engaged in work similar to that of the United States Testing Co. Also, at some time during the course of the trial, another piece (identified in the record as Exhibit 1-X) was cut from the official sample, at the instance of Government counsel, and forwarded to Mr. Henry A. Gassmann, Government analyst in the Division of Laboratories, located in the appraiser’s stores,- New York, for his analysis.
Before discussing the evidence further we deem it proper to advert to our decision in the General Rubber Co. case, supra, in which there was involved a controversy respecting the measuring method necessary to be followed properly to determine the staple length of raw cotton.
Our holding upon that phase of the controversy was summarized in the second headnote of the decision, which reads:
In determining tito length of staple of cotton for tariff purposes, it is proper to use a rule to measure the “blocked off” portion of a “pull” of the cotton, and any estimation or comparison (by the use of the unaided eye) of said “pull” with a “pull” made from a sample of “official type cotton” of 1)4 inches staple length, put out by the Department of Agriculture, which produces a result different from actual measurement is incorrect. The case of Morse Bros., Inc. v. United States, 13 Ct. Cust. Appls. 553, held not controlling of the issues at bar.
The principle so announced in that case, we think, is applicable here under our view that the statute requires a test of the staple in cotton cloth by some proper comparison with the staple of raw cotton such as that from which the cloth is manufactured.
If we understand aright the. trial court’s view upon the method followed by the customs officials in the instant case, it is that they determined the staple length by visual examination, or some com-*253parativo method, without making any measurement of the kind held to be necessary in the General Rubber Co. case, supra.
So far as the testimony of Messrs. Desmond and Beard, the officials primarily responsible for the application by the collector of paragraph 924, sufra, is concerned, it throws no light upon the method used. Mr. Gassmann’s testimony was given long subsequent to the collector’s classification and he personally had nothing to do with that classification, oven in an advisory capacity, but he did analyze Exhibit 1-X, taken from the official sample and his testimony was to the effect that the collector’s finding that the warp threads which constituted 35.2 per centum by weight of the whole cloth were manufactured from cotton having a staple of 1X inches or more in length. It is claimed in the Government’s brief that in making his analysis Mr. Gassmann followed the applicable definition of staple length, as stated in the General Rubber Co. case, sufra, and properly applied it to the imported cotton cloth; that is, he took a typical portion of the fibers in the threads (particularly the warp threads, there being no dispute with regard to the weft threads) and properly determined their staple length, upon the basis of greatest frequency.
The brief on behalf of the Government states:
He took a portion of the merchandise, unraveled the cloth, separated the warp and weft threads, discarded the weft as short-staple cotton, and untwisted the warp threads into individual fibers * * *. It is undisputed that the cloth used was representative and typical. It follows then, that these fibers constituted a typical portion of the cotton contained in the cloth because they were all the fibers of a typical portion of the cloth. The witness arrayed these in graduated sizes, measured them and determined that the length of the fiber which occurred with greatest frequency was V/z inches or over. This would seem to establish without question that such cotton had a staple of 1/% inches in length. The evidence is squarely within the purport of the General Rubber case, supra.
The testimony of Mr. Gassmann with respect to his method of analysis is properly summarized in the Government brief as follows (page references being omitted):
He stated that he had taken a part of Exhibit 1, marked Exhibit 1-X, and had withdrawn some of the warp threads; that he had disregarded both ends of said threads because they contained cut fibers. The threads were then desized and untwisted and the fibers therefrom were arrayed on a piece of board covered with black satin. Two parallel lines were drawn on the black satin V/z inches apart * * *. The fibers were arrayed within the lines according to length. From this array he determined that the staple of the cotton contained in the warp threads was over 1/z inches. Such an array was made twice by this witness; once before desizing the yarn and once after desizing the yarn. * * *.
This witness stated that the modal length is the length of greatest frequency * * * and that the modal length of the fibers was greater than V/z inches * * *. He also made a test of cotton having a staple length of 1/z inches to show the individual lengths and the percentage of fibers having such lengths contained therein. * * *.
*254We may add that there were placed in evidence as illustrative exhibits (identified by Mr. Gassmann as having been prepared by him or under his direction) arrays of fibers, taken from the warp threads of Exhibit 1, arranged on pieces of board covered by black satin. One of the exhibits (Exhibit N) was prepared from the unsized threads and ■one (Exhibit O) from the sized threads. Parallel lines were drawn 1 % inches apart on the covering and the fibers are shown so placed that their length can be seen with reference to the lines. As so shown, it appears (from an inspection of the exhibits) that fibers l}i inches or more in length greatly preponderate in number — that is, there is a greater frequency of fibers of such length.
The importer called as witnesses Messrs. Davieau, Holgate, and Monroe, who analyzed Exhibit 1-A, and, each of them testified at considerable length.
We deduce from the testimony of Mr. Monroe that he followed the Gordon plan. His testimony was to the effect that following that plan he found the staple length of the fibers in the warp threads •to be approximately 1)32 inches in length.
Mr. Davieau stated that in 1923 he assisted Dr. Gordon “in conducting a research to establish the staple length in fabrics”; that “Prior to that time, there had been no method, recognized or established, for accomplishing this,” and described the “research” then made in detail. He testified that subsequently after receiving the ■sample (Exhibit 1-A) he did other “research” work which resulted in his adding steps by which certain “factors,” not obtainable by the Gordon plan, could be obtained.
The test, or analysis, of the sample seems actually to have been made by Mr. Holgate. He testified that he made it in conformity with the plan which Mr. Davieau developed and described the method pursued as follows:
Samples of the warp yarns were taken from the exhibit and treated with an agent to remove any sizing or finishing materials from the yarns, and these yarns were untwisted and wisps of fibers were taken from the individual yarns, and each fiber and each wisp were measured up to a total of five hundred fibers. These measurements were taken and put into what is known as cells, and the •cells were put in increments of sixteenths of an inch. From these cells the total •■number of fibers in the cell was multiplied by the length which would give yoii •a figure from which a cumulative curve could be plotted. That is the total or the product of the number in the cell, which is multiplied by the length in that ■•cell, would give a certain figure and the total of those figures for each cell would •give you the total length in sixteenths of an inch of the fibers measured. That figure divided by five hundred would give you the average length. Or, inversely you can take each one and find the percentage of the total.
The witness then testified, in. substance, that by applying the method so described it was found that the average staple length of the ■“cotton in the warp of Exhibit 1-A” was % of an inch, and there *255was introduced in evidence as Illustrative Exhibit A a photographic copy of a «document which he had prepared showing a “curve” supposed to conform to the “cumulative curve” mentioned in his description of the method above quoted.
No physical exhibits, such as those introduced by the Government above described, showing actual measurements were introduced on behalf of the importer. Its case in that regard seems to rest more upon theory than upon an actual showing. The testimony of its ■witnesses indicate that in determining the staple length involved they based their calculations on average length of fibers rather than upon frequency as to length of the fibers. Obviously the average length of the fibers does not necessarily represent the fiber length of greatest frequency.
The trial court obviously considered the factual testimony from the standpoint of its view representing the interpretation which it felt should be made concerning the law. If we were able to agree with its interpretation of the law, we would not feel at liberty to reverse its decision upon the question of fact, but, since we feel compelled to disagree on the questions of law we think it must be held, for the reasons stated, that the importer has failed to overcome the presumption of correctness which attached to the collector’s action.
The judgment appealed from is, therefore, reversed.
It is interesting to note (although of no importance in this case) that a customs duty on cotton was first provided in an act of August 10, 1790, and that thereafter duties on cotton were continued (the rates being varied in different acts) up to, but not including, the act of March 3, 1883. The duties so provided, however, were not based upon staple length, but upon “raw” cotton irrespective of its staple length, color, or other condition. See Government publication issued in 1896. entitled “United States Gustoms Laws Imnosing Duties on Imports From 1789 to 1895.”