JohnsoN, Judge:
The merchandise involved in this case consists of fluorspar shipped from Spain on the S. S. Gloria and entered at the port of Wilmington, Del., on April 26, 1954. It was assessed with duty at $8.40 per ton under paragraph 207 of the Tariff Act of 1930, as fluorspar, containing not more than 97 percent of calcium fluoride. It is claimed to consist of fluorspar, containing over 97 percent of calcium fluoride, dutiable at $2.10 per ton under said paragraph 207, as modified by the Torquay Protocol to the General Agreement on Tariffs and Trade, T. D. 52739.
The pertinent provisions of the tariff act are as follows:
Par. 207. * * * fluorspar, * * * containing not more than 97 per centum of calcium fluoride, $8.40 per ton; * * *.
Par. 207 [as modified by the Torquay Protocol to the General Agreement on Tariffs and Trade, T. D. 52739], Fluorspar, containing over 97% of calcium fluoride, $2.10 per ton.
The issue before the court is whether or not the imported merchandise contained over 97 percent of calcium fluoride. Plaintiff claims that there are two ways to determine this, one, whether the merchandise was bought and sold as fluorspar, containing over 97 percent of calcium fluoride, and the other by chemical analysis. While, in the absence of chemical analysis, the fact that this merchandise was bought, sold, and accepted as fluorspar, containing over 97 percent of calcium fluoride, might be some evidence that the merchandise was such, we do not think it material where the fluorspar has actually been assayed. Since commercial designation has not been claimed or proved, the merchandise must be classified according to the common and usual meaning of the tariff term. Heylinger & Raubitschek v. United States, 11 Ct. Cust. Appls. 90, T. D. 38735; Armand Schwab & Co., Inc. v. United States, 32 C. C. P. A. (Customs) 129, C. A. D. 296. Obviously, that meaning is fluorspar which, in fact, contains more than 97 percent of calcium fluoride. That fact is most accurately established by chemical analysis.
[18]*18Therefore, plaintiff's offer of proof as to a contract providing for premium payments by consumers for percentages of calcium fluoride over 97 percent and the actual premiums paid was properly excluded by the trial judge. Plaintiff’s collective exhibit 9 for identification is also inadmissible, since the record does not show how the percentages of calcium fluoride given therein were determined.
The record indicates that calcium fluoride was not uniformly distributed in the imported merchandise and that the material had to thoroughly mixed in order to obtain samples representative of the calcium fluoride content of the entire shipment. It also appears that the calcium fluoride in fluorspar does not change over a period of time under normal conditions.
Before this merchandise was exported, samples were taken in Spain every time the excavating bucket loaded into the hold of the Gloria. Said samples were mixed daily, quartered, and samples taken from 2 quarters. The daily samples obtained during a period of 5 days were also mixed, reduced in size, and a general sample of 50 pounds sent to the Laboratorio Químico de Luchana, Bilbao, Spain.
When this merchandise arrived in the United States, it was sampled by Zenovy E. Graniak, a sampler for Booth, Garrett & Blair, analytical chemist, weigher, and sampler. As the merchandise was discharged from the ship into railroad cars, 3 conical-shaped piles were made in each car, and the sampler took about 10 scoops at random from each cone. The material was placed first in a can and then transferred to a drum. At the end of each day, the contents of the drums were emptied on a plate, one on top of the other, and shoveled from one pile to another, three times. The pile was then flattened into a “pie” and marked off in quarters. Two quarters were rejected and the remainder mixed and quartered again. Two quarters thereof were put into a drum and kept until the sampling of the cargo was completed. The samples from 3 days’ operations were then mixed and quartered down. From 2 quarters, 25 pounds were removed and placed in a friction top air-proof can. This constituted the moisture sample and was so identified by card. Then, the quartering-down process was repeated with the remaining material and from 2 quarters another 25-pound sample was drawn as an analysis sample and so identified by card. Those two samples were taken by the witness to the laboratory room of Booth, Garrett & Blair the next day.
Out of the final 2 quarters of the material there was also taken an amount which was placed in two small containers and given to the customs inspector. Mr. Graniak thought they were cans, but Francis J. Murphy, the customs inspector who supervised the unloading of the Gloria, said he received a can and a jar. He took them to the customhouse, and labeled the can and mailed it to the United States Customs Laboratory at Philadelphia (defendant’s exhibit G). He [19]*19kept the jar until September, when he mailed it to the Customs Laboratory also (plaintiff’s exhibit 12).
Out of the 25-pound analysis sample received by Booth, Garrett & Blair, two 1-ounce samples were forwarded to Arnold H. Miller, the American representative of the Spanish producer, and one was sent by him to Andrew S. McCreath & Son and the other to C. O. Parker & Co. Another 1-ounce sample (defendant’s exhibit E) was sent to . the Customs Laboratory on May 5, 1954. Still another sample was Dent to the Ozark-Mahoning Co. and was forwarded to the firm’s chemical laboratory at Rosiclare, Ill. A portion of this sample (defendant’s exhibit J) was subsequently sent through intermediaries to the Customs Laboratory.
Out of the moisture sample, a 1-pound portion (defendant’s exhibit I) was sent to Arnold Miller on May 19, 1954, and was forwarded through intermediaries to the Customs Laboratory in September. Likewise, another 1-pound portion (defendant’s exhibit E) was forwarded through intermediaries to the Customs Laboratory in November or December 1954.
The entire shipment of the fluorspar from the Gloria was received at the Wilmington plant, where it was dried and loaded onto trucks by means of a hose. About 5 or 6 samples were taken from each truckload as the material flowed into the trucks. The samples were identified with the truck number, cargo, and date and sent to the chemical laboratory at Rosiclare.
All of these samples were analyzed for their calcium fluoride content by different persons, by various methods, and with differing results.
The first analysis was made from the samples obtained from the excavating buckets in Spain by Ignacio Barrenengoa, proprietor and chemical analyst of the Laboratoria Químico de Luchana. He holds an engineering degree and has had experience in analyzing and testing fluorspar since 1925. He is the official chemist of the Mining Board and was the official analyst for the Allied Control Commission for several years. In making the test of this merchandise, he first prepared the sample in accordance with international norms of mixture, splitting, and separation and then followed the instructions received from the Ozark-Mahoning Co., a copy of which was attached to his deposition. He considered the procedures standard and “the most reasonable which exist in technical material and concise in procedure.” He made two analyses of the sample and found the calcium fluoride content to be 97.25 percent in the first test and 97.27 percent in the second.
A number of tests of different samples were made in the United States Customs Laboratory in Philadelphia by Joseph Lamb, assistant chief chemist. Mr. Lamb received his degree in chemistry from the University of New Hampshire in 1930, had been employed as a [20]*20chemist by the Brown Co. at Berlin, N. H., the U. S. Gypsum Co. on Staten Island, and the Bemus Industries, Portsmouth, N. H., and has been with the United States Customs Laboratory since 1937.
Mr. Lamb testified that when defendant’s exhibit G was received in the laboratory, it was dried, prepared with screening and grinding to pass through an 80-mesh screen, and thoroughly mixed. Thereafter, he analyzed it for the calcium fluoride content, using the United States Customs Laboratory method, described in defendant’s exhibit H, which is a modification of the National Bureau of Standards method (defendant's exhibit B). The witness made four tests of the sample and subsequently made tests of other samples, using the same procedures. The results he obtained varied from 96.73 percent in one analysis of defendant’s exhibit E to 97.73 percent in an analysis of defendant’s exhibit J.
Two samples were analyzed by Bernard E. Kelley, an analytical chemist holding the degree of bachelor of chemical engineering from Villanova College, who has been employed by Booth, Garrett & Blair since 1946. In making the analysis of the samples from the Gloria shipment, he used a modification of the National Bureau of Standards method, described in defendant’s exhibit D. He found the calcium fluoride content of one sample to be 96.91 percent, and of a resampling, made on May 7, 1954, to be 96.93 percent.
According to the testimony of William Smith, assayist for the Ozark-Mahoning Co. at Rosiclare, the sample which was received from Booth, Garrett & Blair was analyzed by Walter Millhouse, deceased, with the assistance of Mr. Smith. The procedure used was the Bidtel method, described in defendant’s exhibit A, with some modifications. The amount of calcium fluoride found was 97.01 percent.
Mr. Millhouse also analyzed the samples taken from the various truckloads shipped out of the Wilmington plant. A schedule prepared from the official records of the company shows the calcium fluoride content of each sample as found by Mr. Millhouse (plaintiff’s exhibit 2). The percentages range from 95.52 percent bo 97.69 percent, the average found by Mr. Smith being 97.07 percent.
One 1-ounce sample from the original sampling was analyzed by Bruce D. Krecker, director of the laboratories of Andrew S. McCreath & Son. Mr. Krecker received a degree of bachelor of science in chemistry from Albright College and has been connected with Andrew S. McCreath & Son since 1949. At the time of the trial, he was completing his thesis for a master of science degree from Franklin and Marshall College. He had tested many samples of fluorspar during the 6 years he worked in the laboratory. The method he used was a modification of the Bidtel method, described in plaintiff’s exhibit 10. The principal modification was the use of spectrographic [21]*21analysis to determine the purity of the final product, by which, according to the witness, a trained man can arrive at definite conclusions in terms of contaminants. In the instant case, the witness completed two determinations of the sample, finding 97.07 percent of calcium fluoride in one and 97.03 percent in the other, making an average of 97.05 percent.
Another 1-ounce sample from the original sampling was analyzed by Charles O. Parker, founder and manager of Charles O. Parker & Co. Mr. Parker spent his early years in the laboratory of Walter E. Burlingame. Upon return from service in World War I, he worked in the same laboratory while attending the Colorado School of Mines, graduating second in his class in 1923. Thereafter, he did research testing for Anaconda Copper Co., served as chief chemist for the American Smelting and Refining Co., then took over the business of Walter E. Burlingame, which, subsequently, became the Charles 0. Parker & Co. He has served as chemist for the Internal Revenue Department in the prohibition branch and the alcohol tax unit, as chief chemist for the Denver Grain Exchange Association, as chief assayer for the United States Mint, and has performed analytical work for the General Services Administration in its emergency procurement and stockpiling program of calcium fluoride concentrates. He has made possibly 100,000 determinations of the calcium fluoride content of fluorspar and has served as umpire in a number of cases. The testing of fluorspar is one of the daily occurrences in his office.
He personally analyzed the sample of fluorspar from the Gloria shipment received about May 22, 1954. He made four tests of the sample and, at the same time, ran identical tests on four samples of a standard sample obtained from the United States Bureau of Standards, containing a stated content of calcium fluoride, predetermined by many analysts. This was done to remove the personal equation, check minor discrepancies, and for use as a factor in titration in the oxalate method. He described in detail the method he used, stating that he followed the Bidtel method with certain modifications. He was familiar with the Bureau of Standards method and stated that its end method was identical to his, that is, the calcium oxalate titration. His finding of 97.17 percent of calcium fluoride was the result of very close figuring, being the average of four determinations that checked within five one-hundredths of a percent.
In view of the number of tests made of samples of the imported merchandise and the varying results obtained, it is necessary to examine each test carefully in order to find which is most accurate. At the outset, it is clear that some of the tests should be disregarded, including two tests made by Mr. Lamb of the moisture sample (defendant’s exhibits I and F), since the evidence indicates that the [22]*22moisture sample was not representative of the chemical composition of the merchandise. The tests made of the truckload samples should also be disregarded because the samples were grab samples, rather than carefully mixed material representative of the merchandise. There is such a wide discrepancy between the percentage of calcium fluoride in the various samples that the average determined by Mr. Smith cannot be taken as representing the correct percentage of calcium fluoride in the shipment. Furthermore, although the actual analyses were made by Mr. Millhouse, who was apparently employed by the Ozark-M&honing Co. for that purpose, there is nothing in the record to indicate what his qualifications were. For the latter reason, the test made by him of the sample sent by Booth, Garrett & Blah-should not be considered.
In addition, we think the test made by Mr. Kelley of the resampling of May 7, 1954, should not be taken into account, because there is nothing in the record to indicate how or where this resampling was done or whether it was a resampling of the entire shipment or of the samples in the laboratory of Booth, Garrett & Blair.
The Government claims that Mr. Lamb’s test of plaintiff’s exhibit 12 should also be disregarded on the ground that the sample was contained in a jar, whereas Mr. Graniak stated that he gave the customs inspector two cans of the merchandise. The witnesses disagreed as to when the practice of giving two cans to the inspector was changed to a can and a jar or bottle, but it evidently occurred prior to the sampling herein, since the inspector testified that he received a jar, and the jar was produced at the trial. The inspector had kept it on his reserve shelf, had later sent it to the Customs Laboratory, and it was accepted by customs officials as the customs reserve sample of the merchandise. Consequently, tests of the material contained therein may properly be considered.
The following tabulation lists the analyses which remain for examination:
[23]*23
In the instant case, the Government contends that there is nothing in the record to show that the National Burean of Standards method of analysis, as used by the United States Customs Laboratory, is less accurate than any other method; that the method used by the Government analyst was proper; that the analyses of representative samples (defendant’s exhibits E and G), as run by the Government chemist, should prevail; that there is nothing in the record to show that the finding of the Customs Laboratory was erroneous. In making the claim that the record shows that the merchandise did not contain more than 97 percent of calcium fluoride, the Government is disregarding the analysis made in the Customs Laboratory of plaintiff’s exhibit 12, for the reason stated above, and the analysis of defendant’s exhibit J, for no given reason.
It is well settled, of course, that the methods of testing and weighing merchandise used by customs officials are presumptively correct and that the burden is upon the importer to show that such methods or the results obtained are erroneous. United States v. Gage Bros., 1 Ct. Cust. Appls. 439, T. D. 31503; United States v. Lozano, Son & Co., 6 Ct. Cust. Appls. 281, T. D. 35506; Draper & Co., Inc. v. United States, 28 Cust. Ct. 136, C. D. 1400.
In a recent case involving the analysis of fluorspar, T. H. Gonzalez v. United States, 40 Cust. Ct. 9, C. D. 1949, rehearing granted 40 Cust. Ct. 453, Abstract 61560, we held that the Bureau of Standards method had not been discredited or proved inaccurate by the evidence presented. It was recognized as a standard method by the plaintiff’s witness, who did not criticize it, except to say as to his own method, “there is no chance of mechanical loss, one of the defects of other methods.” However, it did not appear that his method was more accurate or eliminated all correction factors. In that case, analyses were also made by the Bidtel method, but the court disregarded them, because the samples used came from a stockpile which was not uniform.
[24]*24In the instant case, we have a different situation. As stated, there is an appreciable variation in the results obtained in the Customs Laboratory, from as little as 90.73 percent to as much as 97.73 percent. For this reason, a further examination of the methods used and the results found by the various chemists is necessary.
In discussing the methods of analyzing fluorspar, the chemist Parker stated:
Well, if you are referring to analytical procedure, an analytical procedure is only a procedure devised to remove from that element which you wish to-determine all of the interfering elements or substances which might be contaminated, and you end up with either the element itself, the mineral rather, or a compound, or some form representing that element from which you can calculate back to the element or the mineral which you wish to determine. In this instance, we are determining calcium fluoride, chemically CaFi. The mineral from which calcium fluoride is derived, and which constitutes that mineral, is fluorite, or fluorspar, known to the trade. We have the duty to determine calcium fluoride, and we can either attack it by determining the calcium contained as the calcium fluoride, or the fluorine contained as calcium fluoride, and general practice has [been] to attack it from the standpoint of determining the calcium content after all other calcium minerals, compounds, have been effectively removed, and that calcium determined as calcium fluoride.
It appears from tbe descriptions of the Bidtel method, that calcium carbonate is first removed, then silica, then other impurities. The residue remaining plus a correction representing losses in the operations is taken as the weight of calcium fluoride. The material is then tested for purity by converting to calcium sulfate, which is calculated to calcium fluoride by the factor 0.5735. If barium sulfate is present, it must also be deducted.
The initial step in the Customs Laboratory method appears to be the same as that in the Bidtel method. Thereafter, the methods differ but, in the final calculations in the Customs Laboratory method, two-tenths of 1 percent is added to the percent of calcium fluoride found to correct for the fluoride dissolved by acetic acid in the first step.
Mr. Krecker stated that he did not believe fluorspar analysis was as accurate as many other analyses, because of the mineral form of the material; that the degree of accuracy that can be obtained is not as great, no matter what method of analysis is used. He said that, at the present time in the McCreath laboratory, every fluorspar that comes in is analyzed by four different methods. The results obtained in the strict Bidtel method, as well as most of the modified Bidtels, compare favorably with the results produced by the different methods, and the results reached by the National Bureau of Standards method are in reasonable agreement with results produced by the Bidtel methods. The Bidtel method is an accepted method in industry and with the Government, the General Services Administration.
[25]*25Mr. Parker stated that the Bidtel method and the National Bureau of Standards method are accepted by various chemists, although they involve the use of empirical corrections. He, personally, does not like such corrections, because different empirical figures are quoted by various people. Therefore, in his method, he tests standard samples along with the samples he is analyzing and eliminates the necessity of an empirical figure. His analysis of the standard invariably checks very closely with the stated content thereof. He has been using the method he described since 1917 and said it was a standard method. He did not use the Bureau of Standards method, because he did not care for the use of hydrogen sulfide, as it involved an additional manipulation.
Mr. Lamb stated that the Customs Laboratory method was approved on March 15, 1949, and that, prior to that, the National Bureau of Standards method was used in customs laboratories.
Mr. Kelley stated that his firm stopped using the Bidtel method in 1936 and was currently using the method described in defendant’s exhibit D, which had been developed by his firm.
Mr. Barrenengoa stated that the method of analysis prescribed by the Ozark-Mahoning Co. did not require the use of an empirical correction. According to the English translation of his answer to the 13th direct interrogatory, the procedures he used “appear in the Bureau of Standards Fluorspar No. 79 from Booth, Garrett & Blair.” However, the description of the method attached to the interrogatories does not appear to correspond with either the Bidtel or the National Bureau of Standards method.
According to the record presented, both the Bidtel method and modifications thereof and the Bureau of Standards method and modifications thereof are currently in use by analytical chemists and are accepted by other chemists in industry. The evidence does not establish which of the methods, if any, is per se the most accurate.
In our view, the controlling factor herein is the accuracy of the particular analyses presented for consideration. Mr. Parker, Mr. Krecker, and Mr. Kelley stated that the results obtained by two chemists should check each other within a quarter of 1 percent (0.25%). Mr. Parker added that this percentage is an arbitrary figure used by the United States Bureau of Standards and that his experience has been, in many instances, far closer than that. Mr. Krecker said that, while the accepted tolerance is about a quarter of 1 percent, he thought that, with close work, fifteen one-hundredths would be expected. Mr. Parker stated, further, that while two chemists should check each other within a quarter of a percent, the individual should check much closer than that on his individual work and that, on an umpire determination, no report is made unless the determinations are within five one-hundredths of a percent.
[26]*26In the instant case, it appears that the percentages of calcium fluoride found by Mr. Lamb in the same sample, defendant’s exhibit G, using the same method, ranged from 96.84 percent to 96.95 percent, a spread of 0.11 percent. In his analyses of plaintiff’s exhibit 12, there was a variance of 0.21 percent, from 97.08 percent to 97.29 percent; in his analyses of defendant’s exhibit E, there was a variance of 0.25 percent, from 96.73 percent to 96.99 percent; and in his analyses of defendant’s exhibit J, there was a variance of 0.24 percent, from 97.49 percent to 97.73 percent. Mr. Kelley submitted only one figure for his analysis of the original sample; therefore, we cannot determine to what degree his own results checked with each other.
Mr. Barrenengoa’s results checked within 0.02 percent of each other, and those of Mr. Parker and Mr. Krecker within 0.05 percent. The results obtained by each of them checked with those obtained by the others within 0.25 percent. All of them found the material to contain more than 97 percent of calcium fluoride. The record indicates that the methods used by Mr. Krecker and Mr. Parker had been developed with a view toward greater accuracy than the original Bidtel method. Mr. Barrenengoa and Mr. Parker had had many years of experience in analyzing fluorspar. Mr. Parker’s experience was longer than that of any of the chemists who testified herein. He had grown up running fluorspars and had made possibly 100,000 determinations. He had made analyses in the commercial field and for the Government. In making the analysis herein, he made four tests of the sample and, at the same time, ran four tests of the standard sample. His four determinations checked within five one-hundredths of a percent. We conclude that the analyses of these three chemists are the most accurate of those presented for consideration.
On the record herein, we find that the weight of the evidence establishes that the imported merchandise contained over 97 percent of calcium fluoride. It is, therefore, properly dutiable at $2.10 per ton under paragraph 207 of the Tariff Act of 1930, as modified, as fluor-spar, containing over 97 percent of calcium fluoride. The protest is sustained, and judgment will be rendered for the plaintiff.