delivered tbe opinion of tbe court:
Tbe merchandise in question was manufactured by a firm in Denmark and imported by appellant. Tbe collector at St. Louis classified tbe same as a nonenumerated article not otherwise provided for, under paragraph 385 of tbe tariff act of 1913, and assessed a duty at 15 per cent ad valorem. Tbe importer protested, claiming that tbe goods were specifically provided for in tbe free fist under paragraph 399 of tbe act of 1913 as:
Annatto, roucou, roeoa, or Orleans, and all extracts of.
Tbe commodity was invoiced as butter color and was intended to be sold to tbe creameries in tbe United States for that purpose. It is also used to color other food products. Commercial designation is not considered in tbe case, and tbe uses to which tbe commodity is put are not regarded as pertinent. Tbe questions for tbe decision of this court are, Is the importation annatto or extract of annatto? and also, Must its classification be governed by administrative practice regardless of what it may be found to be?
Tbe annatto preparation under consideration is manufactured as follows: Chemical manufacturers in Denmark import tbe seed or berry of tbe annatto plant, which is known as Bixa orellana. These seeds, somewhat resembling grape seeds, are coated with a rufous substance. They are first rinsed in water to remove tbe dirt particles and also that portion of tbe coloring substance which would readily dissolve in water. This rinsing water, which contains a large quantity of annatto, is not used in making tbe importation *190at band. Tbe record shows that the rinsing water in some instances is evaporated, and that the annatto residue has a commercial value as coloring matter, but that it is not used in making" the extract of annatto under consideration for the reason that, being soluble in water, it would color the buttermilk when placed in the churn. After the rinsing process the seeds are dried and then placed in vegetable fixed oils which had been previously mixed and heated to a temperature below boiling. The oils absorb or extract from the seed the coloring matter, and as the mixture is stirred from time to time it takes on a deep amber color, its shade depending on the amount of the seed substance which the oil will absorb. The amount of annatto absorbed by the oil varies from 2 to 7 per cent. The oil mixture is a trade secret, and the edible oils, other than cottonseed oil, which are used in small quantities are not named by the record. The mixture of oil and seeds is permitted to stand from six hours to several days, and then by a filtration process the liquid is drawn away from the seeds, and the seeds, together with all fibrous matter and impurities, are thrown away. The product is then known as extract of annatto. The evidence shows that it is then heated to a very high temperature. It is not explained why this is necessary, but the first heating is essential to the process of extraction for the reason that there will be more precipitation of the coloring matter into warm oil than into cold oil. When the process is so completed the annatto and oil are not free and can not be separated.
The commodity is imported in casks and is retailed and delivered in St. Louis for $1.95 per gallon. There is testimony in the record to the effect that the rinsing water and its annatto component is "water extract of annatto,” and also that the powder obtained by evaporating the rinsing water is also "extract of annatto.” The witnesses testify that there is a commercial article known as calce annatto, used in the pharmaceutical industry, and which is prepared, according to the New International Encyclopedia, Vol. II, p. 69, as follows :
It is made from the seed pellicle of an evergreen plant, the Bixa orellana, growing ,in Brazil, Cayenne, and several other places. To obtain it, the seeds of the fruit capsules are crushed and allowed to ferment in water; they are next rubbed upon a sieve, completely mashed, and the coloring matter washed away; after some time the water is decanted, and the coloring matter allowed to dry in the shade. It is then broken up into cakes and wrapped in leaves. Arnotto (or annatto) is insoluble in water; it dissolves with a red color in alcohol, in alkalies, and in fixed oils.
The witnesses testify that this commodity so made is not similar to the importation, that much of the cake annatto is soluble in water, and if used as butter color in the churn would color the buttermilk. *191It seems that there are but three extracts of annatto: Alkali extract, alcohol extract, and oil extract; to which might be added, although the proof is not satisfactory, the water extract. Webster’s New International Dictionary gives the derivation of the word “extract” as:
Extrahere — -ex, out of, plus trahere, to draw — to draw out or forth.
As a verb it is defined as follows:
To withdraw by * * * treatment with a solvent or other mechanical or chemical process * * *; also to treat with a solvent to remove soluble substance.
As a noun the same authority defines the word as follows:
Something prepared from a substance by expression, decoction, shaking with a solvent, or the like.
The word “decoction” is defined as:
The process of boiling anything in a watery fluid to extract its virtues.
One of the witnesses, a consulting chemist, defined an extract as follows:
An extract is a solution of one substance in another that is obtained by the process of extraction.
Which would seem to be an accurate definition.
Applying the above definitions to the statement of facts in this ease, the conclusion is that Exhibit 3, which is the subject of the importation, is an extract of annatto.
It is argued by the Government that the same commodity has been under consideration by the Board of General Appraisers on at least three prior occasions, and that they in the Minnesota Creamery Supply Co. case, Abstract 19944 (T. D. 29339); the Berlin Aniline Works case, Abstract 21248 (T. D. 29763); the American Express Co. case, Abstract 25732 (T. D. 31654) classified it as a nonenu-merated manufactured article not otherwise provided for, and that we are therefore bound by administrative practice. .The records in the above cases are very incomplete, and we know little of the facts, but if the commodity under consideration was identical with the one at hand, the decisions would be erroneous, and precedents clearly erroneous will not be followed. —Lloyd Co. v. United States (9 Ct. Cust. Appls. 280; T. D. 38217); Pacific Creosoting Co. v. United States (1 Ct. Cust. Appls. 312; T. D. 31407); Pierce v. United States (1 Ct. Cust. Appls. 171; T. D. 31215). Furthermore where a ruling of this court is sought upon the ground of long-continued administrative practice, it certainly is incumbent upon the one advancing the argument to show in a case of classification *192that the goods under consideration were substantially identical with the merchandise in the cases relied upon as precedents. —Ulmann & Co. v. United States (4 Ct. Cust. Appls. 77; T. D. 33363); United States v. Schumacher & Co. (3 Ct. Cust. Appls. 301; T. D. 32586).
It might be well to consider briefly the three cases cited above:
In the Minnesota creamery case, supra, decided in 1908, construing the act of 1897, the opinion of the board discloses that the sample was an oil containing 1.5 to 2 per cent of annatto coloring, and that the record contained, a letter from the manufacturer of the commodity in Copenhagen, Denmark, stating that the substance was made from “pure annatto seeds and oil.” The opinion then quotes the New International Encyclopedia process, given heretofore in this opinion, and states that no testimony was introduced in the case. It is furthermore stated in the opinion that—
From the wording of paragraph 475 (which is the same as the free-list paragraph under consideration) [parenthesis ours], we do not think it was intended to admit free of duty thereunder a commodity such as this, largely made up of a dutiable article (oil), when the ingredient (annatto color) entitled to free entry forms but an insignificant part of the product imported.
If the last quoted statement can be said to be the reason for the classification, then the board was wrong, if the importation was an extract. If Congress sees fit to place upon the free list a concoction or extract which contains ingredients which if separated would be dutiable, it clearly has the right to do so, and may have done so in this case, although it is not shown that any of the ingredients which entered into the composition of the extract are dutiable. But the Minnesota creamery case may be distinguished from the case at bar in so far as the opinion discloses that the importation then under consideration was made, according to the New International Encyclopedia’s statement, from the crushed seeds and through a process of fermentation. We do not mean to hold that a commodity so made would not be an extract, but it would not be the same kind of a commodity in its consistency or use as the merchandise at hand.
In the Berlin Aniline Works case, supra, the official chemist reported the importation to be “coloring matter extracted from annatto, and containing alkali carbonate.” The board stated that:
Extract of annatto is shown to be soluble in oil and in alkali, but practically insoluble in water, whereas the article under consideration, by reason of the treatment to which it has been subjected is actually soluble in water.
In the case at bar, the commodity is shown not to be soluble in water. The meager facts set out in the opinion would seem to disclose that the article under consideration had in part been extracted from annatto, and we are unable to understand how the board arrived at the conclusion that it was not an extract of annatto, although the *193facts in that case, which are not fully disclosed, might have justified the conclusion reached by the board.
The opinion in the American Express Co. case, supra, is in five lines, and states that the merchandise is “similar to the extract of annatto which was the subject of the board's decision” in the Berlin Aniline Works case, and follows that ruling.
It is indisputably proved in this case that the annatto coloring is by a well-known process extracted from the .annatto seeds into a solvent. The fact that more than 90 per cent of the product is oil and not annatto does not prevent the whole concoction from being an extract. ■ If due weight is given to the evidence and the definitions of the pertinent words used, it is difficult to see how any other reasonable conclusion can be reached than that the importation is extract of annatto. The fact that a “new and different commodity” is produced, and that it is manufactured, and that it takes labor and skill to produce it, would seem to be a poor reasoning for the conclusion that it was not an extract. We are not justified in saying that Congress meant only that such annatto extracts as could be produced without labor or skill would be free of duty when it placed in the free list “annatto * * * and all extracts of”; nor can we say that Congress intended that no “new extracts” of annatto should be admitted free.
In the tariff act of 1922, we find substantially the same provision:
Annatto, and all extracts of * * * all of the foregoing not containing alcohol.
In the light of recent legislation and events, the reason for excluding extracts containing alcohol is self-evident, but to our minds it also indicates that Congress was cognizant of the fact that extracts contained menstruums such as alcohol, oil, and alkali, and that they would be imported in that condition, and that the presence of the menstruum would not preclude its classification as an extract.
The judgment of the of Board General Appraisers is reversed.