These are protests against the collector’s assessment of duty on raw cotton imported from Mexico through the port of Brownsville, Tex., a subport of Laredo. The protests were consolidated at the trial. It appears from the consular invoices that the shipments consisted of 1,981 bales in entry No. 0136B (protest 124716-K) and 1,374 bales in entry No. 0137B (protest 124717-K). It also appears from the- record that samples of the merchandise were originally taken at Brownsville and forwarded to the appraiser at *169New York; that the appraiser at New York received 47 sacks containing 2,298 samples; that he reported that 1,722 of these were found to be cotton of one and one-eighth inches or more in staple length; that classification was made on the basis of that report. However, the collector, instead of assessing duty upon the weight of the specific bales found to contain long staple cotton, assessed duty at the rate of 3/ cents per pound under paragraph 783, Tariff Act of 1930, as amended, on 81 per centum of the total weight of the cotton imported.
The importer claims that the staple length of the imported raw cotton is less than one and one-eighth inches and is therefore entitled to free entry under paragraph 1662 of the Tariff Act of 1930 and further that the collector erred in assessing duty upon 81 per centum of the entire shipment instead of upon the actual weight of the bales found to be long. In its brief the Government conceded that the assessment of duty should have been upon the actual weight of the bales reported to contain long staple cotton and not upon a percentage basis.
The pertinent provisions of the tariff act are as follows:
Pab. 783 [as amended by the trade agreement with Peru, T. D. 50670]. Cotton having a staple of one and one-eighth inches or more in length_3)4$ per lb.
Par. 1662. Cotton, not specially provided for, and cotton waste. [Free]
The issue is confined to the question: Does the cotton contained in the 1,722 bales reported by the appraiser to be long staple have a staple of one and one-eighth inches or more in length?
The record in this case is voluminous, but a large part of it is concerned with certain preliminary motions which it is unnecessary to discuss here. As a result of these motions, plaintiff introduced into evidence the 47 sacks containing the samples of the merchandise which had been examined and manipulated by the appraiser at New York (plaintiff’s collective exhibit 1) and said samples were examined and manipulated, first, by plaintiff’s witnesses in a room of the United States Customs Laboratory in New York and, second, by defendant’s witnesses in a room of the New York Cotton Exchange.
At the trial Harold Fielding, textile analyst for the United States Customs Laboratory, called as a witness for the plaintiff, testified that he had been a customs analyst for 28 years; that the work of a customs analyst is to test all types of fibers and fabrics, to identify the fiber and its diameter, length, quality, and composition; that the work includes determining the staple length of raw cotton; that he worked for about 6 years with the Pacific Mills in Lawrence, Mass.,, before coming to the Customs Laboratory; that in the last 6 years he has done practically nothing but cotton stapling; that during said period he stapled 10,000 samples representing 10,000 bales of cotton; that long staple cotton is cotton having a staple of one and one-eighth inches or more in length; that cotton having a staple of less than one *170and one-eighth inches in length is considered short staple; that he did not buy or sell cotton commercially, but was just a technical analyst; that he stapled cotton for various customs ports including New York; that he had pursued a course of study in stapling cotton at the Department of Agriculture in Washington in 1945; that he stapled the cotton involved in the entries herein in the conditioning room of the Customs Laboratory about the 20th of September 1947; that that room is kept at a temperature of 70 degrees Fahrenheit and at 65 degrees relative humidity; that he stapled 2,245 samples; that he found 43 samples had a staple of one and one-eighth inches or more and the balance had a staple of less than one and one-eighth inches. Mr. Fielding’s report was received in evidence as plaintiff’s exhibit 6.
Mr. Fielding also testified that he was present when Mr. Neal and Mr. Neville worked in the laboratory; that during that time the room was kept in its standard condition; that the cotton came directly to him from the clerk of this court and was always kept under his jurisdiction; that he stapled the cotton after Mr.- Neal and Mr. Neville, except possibly the first day.
Mr. Fielding described the method of stapling he used as follows:
* * * I take about quarter of an ounce of cotton in my two hands, break it in half, discard one half, draw the fibers through the thumb and forefinger, into the other hand, discard the rest of the fibers, redraw them into the right hand, straighten off the ends, pull out the short, waste fibers, do the same on the other end, lay it on a piece of velvet, block it with the ends of the rule, or fingers, and measure it.
Q. All right. Now, when you did that what did you do, block off the fibers?— A. Yes.
Q. And then you measured? — A. That is right.
Q. Now, what do you call that tuft of cotton that you measure? — A. That is the stapled cotton.
Q. Yes, but what do you call that, a pull? — A. Yes, that is a pull.
Q. That is what you call a pull? — A. Yes.
Q. And then you measure that? — A. Yes.
Q. And you measured it with a rule? — A. Yes.
Q. Then you followed that procedure in each one of them? — A. Yes.
Q. And the results you gave were obtained in the manner you now just described? — A. Yes.
Joseph Hugh Neville, Jr., called as a witness by the plaintiff, testified that he was employed by Manget Brothers, a firm of cotton shippers; that he had been in the cotton business since 1933; that he had worked in the sample room to learn the business, had bought and sold cotton from 1934 to 1940, had graded and stapled cotton for the Clark Thread Co. from 1941 through July 1942; that in 1940, Manget Brothers received a contract from the Department of Agriculture to grade and staple stocks of cotton held by the United States Government amounting to a little over 7 million bales; that he had classed a *171little oyer 100,000 bales of that cotton between January and June of 1940.
Mr. Neville stated that he had stapled the cotton involved herein in the conditioning room of the Customs Laboratory in August 1947 with Mr. W. C. Neal; that Mr. Fielding was also present; that Mr. Fielding and his assistants gave him the cotton to staple; that it was returned to them; that after checking the conditions of the room with Mr. Fielding, he received from him a portion of a United States Government staple type of one and one-eighth inches; that that is a portion of cotton certified and issued under the seal of the Secretary of Agriculture to represent cotton of a staple of one and one-eighth inches in length; that he made several pulls of this type and checked them with a ruler to be sure he had the correct type; that he left these pulls on his little velvet board and stapled the cotton in question; that when there was any doubt in his mind whether the cotton was or was not one and one-eighth inches in length, he checked with the pull and with the rule; that he gave the results to Mr. Marvin Cooper who recorded them; that out of 2,288 samples submitted to Mr. Neal and himself they found 18 samples to be one and one-eighth inches or longer and the balance to be less than one and one-eighth inches in length.
On cross-examination Mr. Neville stated that he learned to staple cotton by actual experience in grading and stapling bales along with experienced classers; that in stapling the cotton herein he compared it with either the type or the rule in each instance; that when there was any question in his mind he placed the pull beside the Government type for a closer examination and checking; that there was a question in his mind about 200 times in stapling 2,288 samples; that in obtaining a typical pull, it is necessary to discard the long and short fibers to obtain those which are predominant in the bale, that is, the fibers of most frequency.
Mr. William Casper Neal testified on behalf of plaintiff that he is a spot cotton merchant and a broker; that he has been in the cotton business since 1905; that he had worked in the sample room of New-burger Cotton Co., learning to staple and grade cotton; that he after-wards engaged in buying, selling, and stapling cotton; that in 1918 he came to New York as first chairman of the Board of Cotton Examiners, a Government agency, which graded and stapled cotton for certification on the New York Cotton Exchange; that he has constantly graded and stapled cotton during all the time he has been in business; that he has been on boards of arbitration representing the Staple Cotton Cooperative Association; that he has been consulted by the United States Customs in New York in reference to cotton cases and has appeared as an expert witness on behalf of the United States.
Mr. Neal stated that he stapled the cotton involved in this case in a room provided by the Customs Laboratory; that Mr. Fielding *172or his assistants handed him the cotton and it was returned to them when he left; that before stapling the cotton, he checked the conditions of the room and found the temperature was 70 degrees and the humidity 65; that Mr. Fielding furnished him with a Government inch and an eighth type, which was put up by the Department of Agriculture; that he pulled the type to acquaint his eye with it, put it on a velvet pad, and measured the staple and length of the type to satisfy himself that it was in proper condition, and then pulled the samples; that in several instances at the beginning and then whenever there was a question in his mind he measured the pull against the Government type and verified it with the rule; that he did not measure every sample but only when there was a question; that there was a doubt in 150 to 200 instances; that he called out the results and Mr. Cooper recorded the numbers and he checked them; that he found 18 samples to be an inch and one-eighth or longer and the balance to be less than an inch and an eighth; that he and Mr. Neville worked together as a committee.
The record of the findings made by Mr. Neville and Mr. Neal as recorded by Mr. Cooper was admitted into evidence as plaintiff’s collective exhibit 7.
C. E. Killinsworth, called as a witness for the plaintiff, testified that he has been classing cotton since 1913; that from 1914 to 1919 he was employed by the Department of Agriculture principally in the preparation of grade standards, which had nothing to do with staple; that from 1919 to 1920 he was employed by McFadden Bros, at Atlanta, Ga., to class and staple cotton for purchase and shipment; that in 1928, he came back to the Department of Agriculture as a specialist in cotton; that he handles-75,000 to 100,000 bales a year; that his office (Board of Cotton Examiners) has been designated to examine cotton for the customs since 1945; that he stapled some of the cotton involved in this case; that the samples were compared with the Government inch and one-eighth type for length; that he stapled most of it in committee with another member of the Board; that they arrived at a unanimous conclusion; that he compared the pull with the Government inch and one-eighth type certified by the Secretary of Agriculture under the Cotton Futures Act of 1922; that he had a record made of his results; that he examined 953 samples and found 940 samples to be short staple or less than one and one-eighth inches in length and 13 to be equal or more than one and one-eighth inches; that the samples he received weighed about 12 ounces. His report was received in evidence as plaintiff’s collective exhibit 8.
The Government called Elwood B. Kern, Jr., who testified that he has been with George H. McFadden & Brother, cotton merchants, since 1923, as an apprentice, a cotton classer, and a salesman, succes*173sively, and for the last 8 years as manager of the Foreign Growth Department; that he has stapled about 50,000 or 60,000 bales of cotton annually for the last 8 years; that in December 1947, he classed the samples involved herein in a room of the New York Cotton Exchange, which was at a temperature of 72 degrees Fahrenheit and from 65 to 68 degrees humidity; that he classed the samples along with Mr. Coughlan and sometimes Mr. Murray; that they called off the class to the clerk; that each morning they would pull the Government type of one and one-eighth inches and lay it out on their felt-covered board; that they would class the cotton samples and lay the rule on them; that in each instance they compared the length with a rule; that he and Mr. Coughlan conferred on the samples and called out the results to Mr. Saul; that out of a total of 1,749 samples, 1,693 were one and one-eighth inches in length or longer, 20 were under one and one-eighth inches, and 36 were insufficient to class.
On cross-examination the witness stated that he bought foreign cotton either on type or description as to grade and staple length; that he has bought Mexican cotton since 1940; that it was bought on type, that is, sample; that cotton bought on type is accepted when it compares with the sample; that it makes no difference what the length is if it conforms to the sample; that in classing the samples herein, he used the rule in each instance; that in classing cotton for his firm, he seldom uses the rule, except on very long cotton; that he has used the rule when serving on arbitration boards for the Cotton Importers Association in the last 2 years; that he used the rule in doubtful cases; that he and Mr. Coughlan worked jointly on the samples, each taking a sample of his own; that some of the samples were very small, but they took a good enough piece to get a pull out of them. Counsel for the importer then showed the witness a sample from plaintiff's collective exhibit 1 and the witness stated that he would take at least half of that to make a pull. The witness then tore it and demonstrated the amount he needed. The piece was weighed by Mr. Fielding and found to weigh half an ounce. Mr. Kern then demonstrated the smallest amount of cotton that he considered sufficient to arrive at a proper staple. This was found by Mr. Fielding to weigh a quarter of an ounce. Mr. Kern testified that a sample may be handled by too many people to obtain a proper staple; that if the sample weighed one ounce and three people had already stapled it, there would be sufficient left for one other person to staple it, but not enough for three more people to do it; that no two cotton men in the world pull alike; that “some man pulls long, and he still knows it is inch and an eighth”; that “another man pulls short and he still knows it is inch and an eighth”; that actually on the ruler it would not be an inch and an eighth.
Edward McCabe Coughlan, called on behalf of the defendant, *174testified that lie has been in the cotton business for about 20 years; that after he left school he was employed by a group of mills, starting in the cotton department where he learned to staple and grade cotton; that he became successively assistant to the cotton buyer, and buyer; that for the past 9 years he has been employed by O’Donnell Brothers as buyer and seller of foreign cotton; that in the early part of December 1947, he was shown some samples in the conditioning room of the New York Cotton Exchange; that he and Mr. Kern stapled them; that he measured the pulls by rule; that out of 1,749 samples he found 1,693 to be one and one-eighth inches or more in length, 20 to be under one and one-eighth inches, and 36 to be insufficient to determine the staple length.
On cross-examination Mr. Coughlan stated that he first started working for the mills in 1923; that he learned to class as an apprentice at a couple of the mills after 2 or 3 years during the summer; that from 1931 to 1938 he bought and classed cotton for the Lisbon Spinning Co.; that since 1941 he has been a salesman and classer for O’Donnell Brothers; that he has sold cotton on description and on actual samples; that he did not personally sell Mexican cotton; that he has stapled Mexican cotton that his firm bought. The witness was then shown a portion of plaintiff’s collective exhibit 1 and asked if it was sufficient to arrive at an accurate and proper staple. The witness stated that it was and was about the minimum that could be used. Mr. Fielding weighed the piece and stated that it weighed one-tenth of an ounce. Mr. Coughlan stated that three men could easily staple a portion of cotton weighing an ounce; that six could do,it if they had to; that in stapling the cotton herein he was told to use the rule besides comparing with the Government type; that he had seldom used the ruler before in his work; that it is much easier to compare with the Government type, fiber against fiber, than to try to find that thirty-second of an inch on a ruler; that he would rather measure against type than use the ruler; that classing cotton is no fixed science but a matter of opinion; that among six classers, four out of six ought to agree; that the average variation might be 20 per centum; that good cotton men come within a thirty-second of an inch of each other.
The record of the findings made by Mr. Kern and Mr. Coughlan was admitted into evidence as defendant’s collective exhibit 10.
Plaintiff and defendant have each attached to their briefs tabulations showing the results found by the various witnesses insofar as the 1,722 bales involved herein are concerned. From these tabula-: tionsit appears that in ten instances the defendant’s witnesses agreed with the plaintiff’s witnesses that the cotton was less than one and *175one-eighth inches in staple length. The Government concedes in its brief that these bales should be free of duty.
Mr. Killinsworth and his Board of Cotton Examiners examined samples of 953 bales, of which only 689 are involved in this proceeding. Of these, they found 678 samples to have a staple of less than one and one-eighth inches in length and 11 to have a staple of one and one-eighth inches or more. The samples used by Mr. Killinsworth were taken from the original bales and were not part of plaintiff’s collective exhibit 1.
Mr. Fielding’s results, as shown on plaintiff’s exhibit 6, indicate a finding of 43 samples as long staple. Of these, one was a duplicate and five were not involved in this proceeding. Of the 37 remaining samples found by Mr. Fielding to be long staple, three were also found long by Mr. Neville and Mr. Neal, and one of these three was found short by Mr. Killinsworth. The other two were not examined by him. Ten samples found by Mr. Fielding to be long were found to be short by Mr. Killinsworth as well as by Mr. Neville and Mr. Neal. Three samples which Mr. Neville and Mr. Neal found to he long were found by Mr. Fielding and Mr. Killinsworth to be short. All the samples found by Mr. Killinsworth to be long were found by Mr. Fielding and Mr. Neville and Mr. Neal to be short. Thus, while all of plaintiff’s witnesses found some samples to be long, there is no agreement among them as to which ones were long.
In large measure, therefore, plaintiff’s witnesses found the samples to be short staple cotton and defendant’s witnesses found them to be long. Since all of the witnesses were well-qualified experts with long experience in stapling cotton, either commercially or in Government service, it will he necessary to examine closely the method of procedure in order to determine which results are more accurate.
The leading case on this subject is United States v. General Rubber Co., 22 C. C. P. A. 308, T. D. 47350. The controversy in that case turned on the question of whether the staple length of cotton should be determined by comparing the pull with the' official cotton type by the use of the unaided eye or by actual measurement of the imported cotton regardless of the actual measurement of the official type. The court stated that the importer had the burden not only of showing that the method employed in classifying the cotton was wrong and brought about inaccuracies but also of proving that the methods employed by its witnesses were proper and produced a more accurate result. The court found that the importer had not sustained its burden, stating (pp. 316, 317):
Obviously, whether the rule is used to measure or the eye is used to estimate, the length of the cotton between the blocked-off ends must be ascertained. It would seem clear that if the witnesses on both sides of this case are correct in their *176determination of the portion of the pull to disregard (the blocked-off ends), the distance between the blocked-off ends is the length of the staple. Since both sides agree that the uneven ends should be disregarded, it is difficult to understand how it may be logically claimed that applying the rule in measuring the distance between the blocked-off ends can produce an inaccurate result. It is our view that any estimation or comparison of the portion which lies between the blocked-off ends which produces a result different from actual measurement would be an incorrect result.
The Government does not urge that the rule be used in cases where it is obvious that the proper measurable portion of the pull of the cotton to be stapled is shorter than the standard in question, but does contend that in close or doubtful cases the rule should be used in order to insure accuracy. It is fair to conclude from the testimony of all the witnesses that the imported cotton is close to the border line and that such a condition calls for accurate measurement.
It appears from the record in the instant case that plaintiff’s witness Fielding and Government’s witnesses Kern and Coughlan measured the pull in each instance with a rule, while plaintiff’s witnesses Neville, Neal, and Killinsworth used the rule only in cases which they considered doubtful. In view of the conflicting findings of the witnesses herein, it is obvious that the staple length of all of the cotton was close to the border line and that measurement with the rule was called for, as provided in United States v. General Rubber Co., supra. Therefore, the results obtained by plaintiff’s witnesses Neville, Neal, and Killinsworth by the comparison method must be held less accurate than those obtained by the other witnesses who used the rule in each instance.
On the other hand, plaintiff questions the accuracy of the results obtained by the witnesses Kern and Coughlan on the ground that they had little experience with the rule method and on the further ground that the samples were insufficient by the time these witnesses manipulated them. As to the first ground, the witnesses stated that they had used the rule on other occasions, particularly in doubtful cases. Although they had more experience with the comparison method and expressed a preference for it, they themselves considered the rule method more accurate, since they used it in all cases where they had any doubt.
As to the second ground, it appears from the record that the clerk of this court weighed each of the samples in plaintiff’s collective exhibit 1 and found that each of 1,894 samples weighed less than one ounce and that each of 394 samples weighed more than one ounce. As shown by the record, these samples were stapled by the witnesses Neville, Neal, and Fielding in the order named and thereafter by the witnesses Kern and Coughlan (and in some cases by a Mr. Murray, who was not called as a witness). Subsequent to the trial, it was *177stipulated by the parties herein that if the witnesses were recalled, each would testify that the respective minimum weight sample used by him to determine the staple would be discarded and could not be used by another classer. Mr. Fielding and Mr. Neal each testified that he used a one-quarter-ounce portion of the sample to determine the staple. Thus, these witnesses and Mr. Neville together used more than one-half ounce of the sample. Since most of the samples weighed less than one ounce to begin with, there was little more than one-quarter ounce left for the Government witnesses to use. Mr. Kern' testified that the least he could use to obtain a proper staple was one-quarter ounce, but Mr. Coughlan testified that he could obtain a staple with one-tenth of an ounce. Mr. Kern stated that where a one-ounce sample had been stapled by three people, there would be enough left for orily one more person to obtain a proper staple, but Mr. Coughlan said six men could staple a one-ounce sample if necessary. Although the samples used by the Government’s witnesses must have been small, they found them sufficient except in 37 instances where no staple was found.
The court in United States v. General Rubber Co., supra, quotes from a bulletin issued by the Bureau of Markets of the Department of Agriculture on the method of pulling cotton, wherein it is indicated that the proper size of the tuft of cotton to be used is about one-quarter ounce.
Plaintiff’s reply brief herein contains the following quotations from publications of the’ Department of Agriculture:
(c) Samples should be handled with care so that their representativeness will not be impaired. Classers should not attempt to class samples that have been excessively handled. [Handbook for Cotton Classers.]
The useful life of a sample will depend to some extent upon its size, the care given it, and the frequency with which it is handled. After a sample has been classed more than three times, its representativeness is ordinarily subject to question.
* * * Some of the errors in classing may be traceable to an attempt to class samples that are too small to represent accurately the cotton in the bale. The class of the sample may be correct, but it may not be the true class of the bale from which the sample was drawn. Undoubtedly other differences are due to improperly conditioned, poorly handled, or excessively handled samples which, because of such conditions, fail to be representative of the bale. [Classification of Cotton, Miscellaneous Publication 310.]
In view of the foregoing, there is some doubt as to the accuracy of the results reached by the Government’s witnesses.
There remains the testimony of Mr. Fielding and the results reached by him. While Mr. Fielding had no experience in stapling cotton commercially, he had years of experience as a textile analyst for the Government and has done practically nothing but staple cotton for *178the last 6 years. Since he testified that he used the rule in each instance and since the samples came to him when they were still sufficiently large to be manipulated, his results appear to be the most accurate.
On the record herein we find that the cotton in the following bales out of the 1,722 in dispute is one and one-eighth inches or more in staple length:
Entry 0136B ESK 14 24730
0136B EJJ 39 26031
0136B FSG 3 17705
0136B FSG 3 17738
0136B FSG 3 17764
0136B FSB 42 12711
Entry 0137B FSSA 27 28028
0137B FSSA 27 28074
Entry 0136B ESI 26 26302
0136B FSD 15 27014
0136B FSH 16 26255
0136B FSA 7 12962
0136B ESH 34 26245
0136B EJA 5 11756
Entry 0137B FSX 4 25959
Entry 0136B ESF 17 29517
0136B ESF 17 29649
Entry 0137B FSS 19 18343
Entry 0136B ESF 43 29645
0136B EJH 36 27320
0136B EJH 36 27338
0136B EJH 36 27339
0136B EJH 36 27313
0136B FSH 35 17830
0136B FSH 35 17837
0136B FSI 20 24609
0136B FSI 20 24596
0136B FSI 20 24679
0136B FSFB 41 29192
0136B ESK 2 24526
0136B FSC 8 27016
0136B FSD 30 27001
0136B FSD 30 27247
0136B FSD 30 27216
0136B FSD 30 27076
0136B EJA 6 11892
0136B EJA 6 11923
Such cotton is held to be dutiable at 3K cents per pound under paragraph 783 of the Tariff Act of 1930, as amended by the trade agreement with Peru, T. D. 50670. The balance of the cotton involved herein is held to be free of duty under paragraph 1662. To that extent the protests are sustained and judgment will be rendered accordingly.