This is a protest against the collector’s assessment of duty on fluorspar, imported from Mexico on June 20, 1955, in *10car No. L & A 9156, at $8.40 per long ton, under paragraph. 207 of the Tariff Act of 1930, as fluorspar, containing not more than 97 per centum of calcium fluoride. It is claimed that the merchandise is fluorspar, containing more than 97 per centum of calcium fluoride, and is subject to duty at $2.10 per long ton under paragraph 207 of said tariff act, as modified by the Torquay Protocol to the General Agreement on Tariffs and Trade, T. D. 52739.
The only issue in this case is whether or not the involved fluorspar contains more than 97 per centum of calcium fluoride.
It was stipulated at the trial that the official sample of the merchandise and that used by the importer were taken by the Eagle Pass Laboratory and that said samples are representative of the merchandise. The method of taking the samples was described by John E. Roche, owner of the laboratory, as follows: Samples in lump and gravel form are obtained by digging holes to the bottom of the car in all positions of the car. Such samples run from four to five hundred pounds on an ordinary sized car of 50 tons. [According to the entry, car No. L & A 9156 contained 63.50 long tons.] The samples' are crushed, mixed, and pulverized by various operations until particles small enough to pass through an 80-mesh screen are produced. After rolling on a rubber cloth 100 times to make the samples as homogenous as possible, they are ready for analysis.
One of the samples involved herein was analyzed for its calcium fluoride content at the Eagle Pass Laboratory under the supervision of Mr. Roche. This witness testified that he was a chemical engineer and that his experience included 23 years with the American Smelting Refining Co. in Mexico in all capacities from washing beakers to plant manager, consultation work on smelter operations, and the operation of his own laboratory for the sampling and assaying of ores. In the course of his career, he had made more than 20,000 complete analyses of fluorspar and had been employed by the Government to make analyses of fluorspar purchased by the General Services Administration for stockpiling purposes.
Mr. Roche testified that the method of analysis which he used was “a strictly wet method, part of which was developed by Dr. Schrenk at the Missouri School of Mines,” the products being determined through solution rather than through any baking action. In describing the process, the witness stated as to one operation: “The sample is taken down to heavy fumes on a hot plate, asbestos covered so that no bubbling action goes on and there is no chance of mechanical loss, one of the defects of other methods.” In the course of the process, the witness obtained what he called the total lime titration, which included calcium fluoride, calcium carbonate, calcium sulphate, and calcium phosphate. He explained the method by which the amount of calcium fluoride was found, as follows:
*11* * * The amount of fluoride that goes into solution under my conditions, and with my solutions, is six-tenths of a cc. * * *
:{: jjs }{{
* * * This amount must therefore be deducted from that titration. Assume I had a titration of 3 cc. I deduct 0.6 cc., giving me 2.4 cc., and since I used a 3 gram sample, and I must get it back to a half gram basis, I have to divide that by 6. Six into 2.4 gives 0.4. Then multiply that by the calcium carbonate factor of your solution, and this gives you the net soluble lime other than fluoride in that particular ore. Now you have the total titration, you have a titration against your permanganate solution of a half gram sample. Let’s take for example 46 cc. That includes your calcium carbonate, calcium sulphate, your calcium phosphate, calcium fluoride. That is everything in lime. From this you deduct the lime, the titration on a half gram sample which is 0.4, which corresponds to the lime part of that, and the gypsum, the sulphate and phosphate. That gives you 46 minus 0.4, gives you 45.6 cc. Multiply that by the calcium fluoride factor of your permanganate solution, and it gives you the calcium fluoride.
The witness testified that, in the instant case, he determined the calcium fluoride content to be 97.10 per centum, the silica 0.70 per centum, and the calcium carbonate 1.90 per centum.
Mr. Roche stated that he was familiar with the Bidtel method of analysis of fluorspar and the Bureau of Standards method, as well as with the Schrenk method, and that all three are accepted by chemists as methods for the analysis of fluorspar.
Manuel Davila, manager of a plant operated by Fluorita de Mexico, shipper of the instant merchandise, testified that there were three carloads of fluorspar involved in the shipment and that the fluorspar in car L & A 9156 was obtained from the stockpile at the plant. He said that the quality control of the fluorspar produced by his company is maintained by taldng a so-called grab sample every 15 minutes during processing. The samples are put in a box and every 8 hours the box is sent to the laboratory where the grab samples and other samples of the 8-hour pile of concentrates are analyzed by the Bidtel method. If the samples assay over 97 per centum, the material is placed in the stockpile from which shipments are made.
Furthermore, according to the witness, after the merchandise is loaded on the trucks to be taken to the railroad cars, a small sample is taken from each truck and is analyzed by the laboratory and a report issued. The report as to car L & A 9156 (plaintiff’s exhibit 2) shows a calcium fluoride content of 97.40 per centum. Reports on the other two cars involved in this shipment (defendant’s exhibits A and B) show a calcium fluoride content of 97.30 per centum for car L & A 9307 and a calcium fluoride content of 97.70 per centum for car TP 19289. The witness explained the differences in the percentages among the three carloads as being due to the fact that his company has a big stockpile and that all of it is not exactly the same. It is not mixed, as the only thing his company is interested in is that it contain over 97 per centum calcium fluoride.
*12O. M. McCombs, assistant chief chemist .in the United States Customs Laboratory in New Orleans, stated that he had received a bachelor’s degree in chemistry from East Texas State Commerce in 1948 and a master’s degree in 1949 and had done analytical work for two Army arsenals prior to his joining the Customs laboratory in 1950.
This witness testified that his office received two samples of the involved merchandise in the regular course of business, one on July 1, 1955, and a duplicate on August 22, 1955. They were analyzed under his supervision by two chemists in the laboratory who used the umpire method, which is prescribed by the Bureau of Standards (of the Department of Commerce). The analysis was carried out in duplicate on the samples and also on a Bureau of Standards reference sample, No. 79. Said sample No. 79 was received in evidence, together with a certificate of analysis thereof, as defendant’s collective exhibit D. The sample consists of fluorspar which has been analyzed by the Bureau of Standards and a number of industrial and research laboratories throughout the country. • According to the certificate of analysis prepared by the Bureau of Standards, the calcium fluoride content is 94.83 per centum. This amount appears to be the average of 14 determinations made by the Bureau of Standards method and 10 made by other methods. It is stated on the certificate that it is believed that the certified percentage of calcium fluoride is accurate to 0.25 per centum.
Mr. McCombs testified that, in the Customs laboratory, sample No. 79 is run through the entire proceedings just as are the other samples. If the results arrived at on sample No. 79 check with the certificate of the Bureau of Standards, then the results on the other samples are considered true and correct. In the instant case, a so-called short method was used with the first sample by means of which the approximate percentage of calcium fluoride was obtained, but certain impurities remained. When the duplicate sample was received, the analysis was done over and additional operations were performed. In all, 22 analyses were made of these samples, 18 of which were taken through the entire long method by which the exact percentage of calcium fluoride is obtained. The witness said that the lowest percentage of calcium fluoride found by these analyses was 96.0 per centum and the highest 96.74 per centum. The majority of the determinations corresponded very closely to the average, which was 96.64 per centum. In obtaining these results, a correction factor of two-tenths of 1 per centum was added to allow for the small amount of calcium fluoride which is dissolved in the initial step of the process. The witness explained that the extreme low of 96.0 per centum was for all practical purposes out of control and not conclusive of the average. He said that variances should not be in excess of two-tenths of 1 per centum, although, in sample No. 79, the findings of thedifferen't *13chemists and different laboratories using the Bureau of Standards method showed a greater variance, 0.28 per centum.
Mr. McCombs said that the umpire method differed from the Bidtel and the Schrenk methods in that those were lime determination methods. He explained that the Bidtel method assumes that the composition of fluorspar is basically of certain components, calcium carbonate, lime and aluminum oxide, silicon dioxide, and fluorspar; that the percentages of silica calcium carbonate, lime, and aluminum oxide are determined, and it is assumed that everything else is calcium fluoride. The Schrenk method is a wet assay method which determines the total calcium components; the soluble portion of calcium products other than calcium fluoride is determined and is subtracted from the total lime content, which purportedly includes salts of calcium only. However, barium, strontium, and, possibly, magnesium will precipitate at the same chemical points as calcium, unless they are very carefully controlled. In the opinion of the witness, the Schrenk method and the Bidtel method'do not definitely. assure that all the impurities other than calcium salts have been removed before the initial calcium analysis, and the final report under the Schrenk method might, and possibly does, include strontium fluoride, barium fluoride, magnesium compounds, and other salts. He added that, although the Bidtel method was at one time considered a standard method, many laboratories, including the Customs laboratory, have discarded it. This was done after known samples of fluorspar were analyzed by different laboratories throughout the country by various procedures and the deviations of the Bidtel method and the wet lime assay method, corresponding to the Schrenk method, were too high.
All of the witnesses herein testified that calcium fluoride is a stable component which does not change under ordinary circumstances. It was stipulated at the conclusion of the trial that if James P. Peterson, chief chemist of the Customs laboratory at New Orleans, were called to testify, he would testify to the same facts that Mr. McCombs did.
Thus, it appears that three laboratories each using a different method of analysis found different percentages of calcium fluoride in the within merchandise. The Customs laboratory, using the umpire or Bureau of Standards method, determined that the calcium fluoride content was 96.64 per centum. Mr. Roche, using the same samples but a different method (a modification of the Schrenk method), found that it was 97.10 per centum. The laboratory of Fluorita de Mexico, using different samples and a different method (the Bidtel method), found that it was 97.40 per centum.
Plaintiff claims that the presumption of correctness of the collector’s classification of the merchandise on the basis of the Customs *14laboratory analysis has been overcome. This contention is made on the ground of the testimony of Mr. Davila that the merchandise was shipped by Fluorita de Mexico from its stockpile of fluorspar, containing over 97 per centum calcium fluoride, the analyses made by Fluorita de Mexico, and the analysis made by the Eagle Pass Laboratory. It is further claimed that Mr. McCombs’ criticism of the Schrenk method was not the result of his personal experience and that the method he used was a modification of the Bureau of Standards method and involved a correction factor.
The Government’s contentions are summarized in its brief as follows:
1. It is not sufficient to show that two analysts found over 97% calcium fluoride while the Government analyst found under 97 %. The plaintiff must also show that the method used by his analysts was a standard, approved method, for the results obtained by an unapproved method obviously are of no consequence.
2. The Bureau of Standards method followed by the Customs Laboratory
has been approved by this Court, and nothing in the instant record discredits that method. • .
3. The Government analyst has not been challenged in any way. Therefore, the analysis, resulting in a determination of under 97% calcium fluoride by an approved method, supports the collector's classification.
It is well settled 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; Resolute Paper Products Corp. v. United States, 31 Cust. Ct. 285, Abstract 57595.
In the instant case, the Customs laboratory followed the method of analysis prescribed by the Bureau of Standards, the only modifications being the use of the short method on the first sample and the limitation of the determinations to the calcium fluoride content. The Bureau of Standards method was approved by this court in Wm. A. Bird v. United States, 63 Treas. Dec. 980, T. D. 46437, wherein the court found the analysis presented by the plaintiff more accurate, stating (p. 985):
In analyzing the fluorspar the commercial chemists used the method which had been determined by the Bureau of Standards as the most exact method to obtain consistent results. This .method was not used by either of the chemists who testified for the Government and who obtained a lower content of calcium fluoride. Samples should have been obtained by the Government in accordance with instructions contained in T. D. 43095. We are of the opinion that the testimony of the Government chemist with respect to his analysis of improperly identified samples by some method other than the one recommended by the Bureau of Standards is not sufficiently competent evidence to rebut the accuracy of analyses made by plaintiff’s witnesses, commercial chemists, based upon properly identified samples in accordance with the method approved by the Bureau of Standards.
*15Here, the Government chemists analyzed representative samples in accordance with the Bureau of Standards method. That method has not been discredited or proved inaccurate by the evidence presented. Mr. Roche recognized it as a standard method of analysis and did not criticize it, except to say as to one operation in his method, “there is no chance of mechanical loss, one of the defects of other methods.” While it was admitted that a loss does occur in the first step of the method, a correction factor was used to offset this. Since Mr. Roche also used a correction factor in his calculations, it does not appear that his method was more accurate or eliminated all correction factors.
While all of the witnesses recognized the Bidtel method to some extent, Mr. McCombs pointed out that it had been discarded by many laboratories, including the Customs laboratory. In any event, we do not think the analyses made by the Bidtel method by Fluorita de Mexico have been shown to be more accurate than the determinations made by the Customs laboratory. While plaintiff refers to three analyses made by the Mexican firm, only one specifically covers the merchandise in question. The others involve grab samples and other samples generally taken from 8-hour productions of fluorspar, and analyses of merchandise in other carloads. We are concerned only with the merchandise in car No. L & A 915.6. The analysis of said merchandise was made from small samples taken from the truckloads making up the carload. The evidence does not show how these samples were obtained, and, since Mr. Davila stated that the stockpile was not uniform, it cannot be presumed that such samples were representative of the merchandise.
On the record presented, we hold that plaintiff has not met his burden of overcoming the presumption of correctness attaching to the collector’s classification of the merchandise and has not established that it was properly dutiable as fluorspar, containing more than 97 per centum calcium fluoride. The protest is overruled and judgment will be rendered for the defendant.