This appeal is from the judgment of the United States Customs Court, 79 Cust. Ct. 53, C.D. 4713, 440 F. Supp. 1237 (1977), dismissing appellant-importer’s protest to the .classification of an antibiotic drug, identified by the trademark Rimactane, imported from Italy. We reverse.
The structural formula of Rimactane is shown below with the portion of the molecule with which we are concerned within the dotted line:
*83The imported Rimactane is produced in a multi-step synthesis from a naturally-occurring mold extract known as Rifamycin B haying' the following structural formula:
The first intermediate formed from Rifamycin B in the course of producing Rimactane is known as Rifamycin S and has the following structural formula:
*84Comparison of the aboye formulae reveals that the singly-bonded, oxygen-containing substituents on the rightmost ring in Rifamycin B are initially oxidized to leave doubly-bonded oxygen substituents, with a concomitant alteration of the electron distribution in the ring, before ultimately returning to a singly-bonded status in the final product.
The Customs Service classified Rimactane in Schedule 4, Part 1, of the Tariff Schedules of the United States (TSUS), 19 USC 1202— Benzeniod Chemicals and Products; more specifically, under item 407.85, “Drugs: Other.” In protesting this classification, appellant-importer concedes that Rimactane is a “Benzenoid Chemical” of the sort comprehended by Schedule 4, Part 1, but contends that Rim-actane is excluded from classification therein by the operation of headnote 3 thereof, which generally excludes compounds containing naturally occurring “benzenoid” structures derived from non-coal-tar sources. Appellant contends, therefore, that Rimactane must be classified in Schedule 4, Part 3 — Drugs and Related Products; more specifically, under item 437.32, “Antibiotics: Other.”
The relevant statutory provisions read, in pertinent part:
Tariff Schedules of the United States (TSUS), 19 USC 1202: Schedule 4. — CHEMICALS AND RELATED PRODUCTS Part 1. — BENZENOID CHEMICALS AND PRODUCTS Part 1 headnotes:
1. Except as specifically set forth in the headnotes to other parts of this schedule, all products described in this part shall be classified hereunder even if more specifically described elsewhere in this schedule. Any product described in both subparts B and C of this part shall be classified in subpart C.
2. For the purposes of this part, the term “modified benzenoid” describes a molecular structure having at least one six-membered heterocyclic ring which contains at least four carbon atoms and having an arrangement of molecular bonds as in the benzene ring or in the quinone ring, but does not include any such molecular structure in which one or more pyrimidine rings are the only modified benzenoid rings present. [Emphasis in original.]
3. With the exception of the natural products provided for in subpart C, this part does not cover cyclic organic chemical products (such as, but not limited to, tannic, gallic and pyrogallic acids; estrone, estradiol, and corticosteroids; morphine, ergot, and cinchona alkaloids; rotenone; phenylalanine; tyrosine; epiner-phrine; and thymols) having a benzenoid, quinoid, or modified benzenoid structure, which are produced from animal or vegetable products in which such structure occurs naturally,' unless such cyclic organic chemical products were obtained, derived, or manufactured in part from any product provided for in subpart A, B or C of this part. [Emphasis ours.]
*85Subpart C.&emdash;Finished Organic Chemical Products
Subpart C headnotes:
1. The provisions of this subpart providing for products obtained; derived, or manufactured in whole or m part from products described in subparts A or B of this part shall also apply to products of lité chemical composition having a benzenoid, quin-oid, or modified benzenoid structure artificially produced by synthesis, whether or not obtained, derived, or manufactured in whole or in part from products described in said subpart A or B.
Classification:
Products suitable for medicinal use, and drugs: Obtained, derived, or manufactured in whole or in part from any product provided for in subpart A or B of this part:
* * * * * *
Drugs: * * * * * * *
Item 407.85 Other_ 1.7f5 per lb.+12.5% ad val. (as modified by T.D. 68-9).
❖ * , * ' ’* * * *
Part 3.&emdash;DRUGS AND RELATED PRODUCTS * * * * * * *
Claim:
.Antibiotics: • *******
Item 437.32 Other- 5% ad val. (as modified by T.D. . 68-9).
The Customs Court dismissed appellant's action, holding that Rimactane was not comprehended by the .headnote 3 exclusion because .the sequence of events involved in its production, as above described, did~ not show "the identifiable continuity of an original benzenoicl structure * * ~." 79 Cust. Ct. at 58, 440 F. Supp. at 1240. The trial court construed headnote 3, correctly we think, as requuiring "the benzenoid structure of the imported. product to be the very same structure found in the original animal or vegetable product from which it w~s produced," meaning that the original benzenoid structure "must remain identifiable at all distinct stages of the production process." Id. at 54, 440 F. Supp. at 1239. In the face of conflicting expert testimony on the point, the court found that the transformation of the right-most ring in Rifamycin B from a structure resembling benzene to a structure more closely resembling quinon.e in Rifamycin S constituted a modification of. the original "benzenoid" structure.
he Customs Court disposed of the importer’s contrary arguments, which generally correspond to its arguments before us, as follows:
*86The court cannot subscribe to plaintiff’s argument that the fundamental relevant structure is only the skeletal carbon ring and its associated electrons (the aromatic sextet) and that the bonding to oxygen which creates the quinoid resemblance has no more significance in principle than the addition of a peripheral hydroxyl or acetyl group. The headnote in question does not treat the subject in such an irreducible and abstract manner. It speaks of specific and distinct structures, characterized by their resemblance to the compounds benzene or quinone. When a strong resemblance to quinone supplants the resemblance to benzene in the course of production, the original benzenoid structure is at an end.
The previous analysis also negates plaintiff’s additional argument that benzenoid, quinoid and modified benzenoid structures are to be considered as a single class and presumably, a change from one to the other in the course of production would not affect the “natural” occurrence of the final structure. Though they are linked together in the statute and though they may have a strong “family resemblance” and even occasionally describe the same compound, their separate enumeration requires that they he considered as distinct structures when it is reasonable to. do so. In this instance the description of the structure of rifamycin S as quinoid surpasses in accuracy and alternative description and cannot be ignored. [Id. at 55, 440 F. Supp. at 1240, emphasis ours.]
Opinion
We hold that the Customs Court erred, first, in its characterization of Rifamycin S as more accurately described as “quinoid” and, second, in its conclusion that the separate enumeration of the words “benze-noid” and “quinoid” in the statute required compounds so characterized to be considered legally distinct.
We note initially that the compounds here involved are complex, fused-ring structures which are neither easily depicted nor easily named. The structural formulae reproduced above are merely models which represent, with varying degrees of precision, the complex relationship between the nuclei and electrons of atoms which have become associated as.a larger molecular entity through the action of forces which, if understandable at all, are understood only with difficulty. One cannot blindly rely on visual similarities or dissimilarities in structural formulae; the limitations of these models must recognized. In particular, the double bonds illustrated in the fused-ring structure of Rifamycin B do not illustrate fixed double bonds, such as those in ethylene. Instead, due to the unique nature of the six-member carbon ring, they can be regarded only as indicating a degree of endocyclic unsaturation which is distributed throughout the ring. By the same token, the double bonds illustrated in the rightmost ring of Rifamycin S, while unquestionably “different” from those in Rifamycin B, are, nonetheless, similarly “delocalized” to *87some extent so that the impression conveyed by the structural formula is, in many ways, misleading.
The issue presented by this appeal, as we see it, is whether the benzenoid, quinoid, or modified benzenoid structure of Rifamycin S, conventionally illustrated above, is properly characterized for tariff classification purposes as naturally occurring or as synthetic. If naturally occurring, the government apparently concedes that all further modifications in the preparation of Rimactane do nothing to alter the basic benzenoid, quinoid, or modified benzenoid structure, such that Rimactane would be excluded from Schedule 4, Part 1, by headnote 3. If synthetic, all further derivatives of Rifamycin S, including Rimactane, must be classified in Schedule 4, Part 1, in .view of the last clause of headnote 3.
In resolving the issue, we must reject the approach taken by the Customs Court in finding that Rifamycin S is more accurately described as “quinoid.” The Fifth Supplemental Report of the Tariff Classification Study, United States Tariff Commission (1963), states, in pertinent part, at 22:
The terms “benzenoid”, “quinoid”, and “modified benzenoid” are not mutually exclusive. A single chemical substance may have a molecular structure which may be aptly described by any one or combination of these terms. For example, anthraquinone is both benzenoid and quinoid, quinoline is both benzenoid and modified, benzenoid, and pyridine is modified benzenoid only.
The discussion of anthraquinone 1 is directly applicable to Rifamycin S and indicates that characterization as “benzenoid” or “quinoid” loses its exclusivity in a complex, fused-ring system, such as the one at bar, which is equally accurately described as either “benzenoid” or “quinoid.”
With respect to the significance to be attributed to the wording of headnote 3, we agree with appellant that the statutory language “cyclic organic chemicals having a benzenoid, quinoid, or modified benzenoid structure” was intended to describe a single class of compounds. The Tariff Classification Study, U.S. Tariff Commission *88(1960), Schedule 4; Part 1, Explanatory Notes, states at 17 (emphasis ours):
The obsolete, misleading, and often ambiguous term “coal-tar products” used in tariff paragraphs 27, 28, and 1651 of the Tariff Act of 1930, as amended, is replaced by the more specific and scientifically accurate term “cyclic organic chemicals having a benzenoid, quinoid, or modified benzenoid structure.”
We would, therefore, discount the Customs Court’s reliance on the separate enumeration of the words “benzenoid” and “quinoid” as indications of a legislative intention that compounds so described must be regarded as legally distinct for purposes of tariff classification.
Quite clearly, all of the structures involved in the preparation of Rimactane fall within the statutory class of cyclic compounds referred to in headnote 3, but, contrary to appellant’s implicit assumption, such a conclusion does not dispose of the question of whether the cyclic structure of Rifamycin S can be characterized as naturally occurring. Resolution of this question entails examination of the essential characteristics of the original benzenoid structure and analysis of the nature of the structural changes that have occurred, as far as thet are known.
We find in the Fifth Supplemental Report, supra, an indication that the drafters regarded the simple presence of endocyclic 'unsaturation as the essential characteristic of a benzenoid, quinoid, or modified benzenoid structure. It is stated:
On the other hand, methylcyclohexanone (item 403.78) or cyclo-hexane (item 403.80), both being of a saturated molecular structure without double bonds in the ring, do not in themselves have a benze-noid, quinoid, or modified benzenoid molecular structure, and therefore, are in Part 2 unless the particular shipment involved is .derived from a product having a benzenoid, quinoid, or modified benzenoid structure, [Id. at 22, emphasis ours.]
There is no expressed concern for the quantum-mechanical niceties of exactly where the available electrons are spending most of their time. As a matter of technical fact, one of the few things the opposing experts agreed upon at trial was that one really could not say where the available electrons in these cyclic structures were but must speak, instead, in terms of varying degrees of “delocalization” and “electron density.”
Focusing now on the nature of the transformation in issue, examination of the right-most ring in Rifamycin B reveals a structure resem*89bling a partially-etberified hydroquinone.2 Examination of the rightmost ring in Bifamycin S reveals a structure resembling the quinone-which corresponds to the partially-etherified-hydroquinone portion of Bifamycin B.
According to Defendant’s Exhibit L, a U.S. patent3 describing,, inter aha, the production of Bifamycin S from Bifamycin B, conversion of the partially-etherified hydroquinonic form to the quinonic-form occurs when Bifamycin B is simply stored in water in the presence-of air.4 The reverse reaction, whereby quiñones are converted by a mild reducing agent, such as ascorbic acid, to the corresponding-hydroquinones, was described at trial by the Government’s expert witness, Dr. Soloway, as “a well established reaction- in organic-chemistry” and is,- in fact, used in the further modification of'Bifamy-cin S to form the hydroquinonic Bimactane. Whatever the electronic configurations in these molecules are, they would appear to be- relatively closely energetically balanced in view of the ready interconversion of the two.5
Furthermore, Defendant’s Exhibit N, another U.S. patent,6 describes a different, albeit related, cyclic intermediate product bearing-either a pair of hydroxyl groups or a pair of doubly-bonded oxygen atoms on the right-most ring as “hydroquinonic” and “quinonic”' “form[s]” (emphasis ours), respectively, of the same intermediate compound. This ante litem motem indication of the perception in the-trade of the relationship between the ring structures of compounds similar to Bifamycin B and Bifamycin S is particularly revealing.
To the extent that Bifamycin S possesses the endocyclic un-saturation characteristic of benzenoid, quinoid, or modified benze-noid structures, that unsaturation is naturally occurring. The 10-member carbon ring skeleton and its associated electrons are, similarly, *90naturally occurring. The exocyclic oxygen atoms are also naturally occurring. It is thus seen that the unsaturated ring structures of the two compounds vary only in the distribution of available ring-carbon electrons. The position of electrons in such complicated molecules seems to be a matter of statistical “uncertainty” 7 and expert dispute far beyond the probable interest or expertise of Congress. With respect to Rifamycin B and Rifamycin S, this variation seems to be at most a matter of degree. In view of this gradational variation in electronic configuration, and in view of the fact that hydroquinonic and quinonic versions of these molecules seem to be susceptible, due to their ready interconversion, of characterization as “forms” of the same ring structure, we are of the opinion that the Customs Court erred in failing to conclude that the unsaturated ring structure of Rifamycin S can be characterized for tariff classification purposes as naturally occurring. The Customs Court, therefore, erroneously dismissed the importer’s action seeking classification of Rimactane under TSUS item 437.32.
The judgment of the Customs Court is reversed.
The structural formula for anthraquinone is:
Hydroquinone has the following structural formula:
U.S. Patent No. 3,301,753 issued January 31, 1967, to Piero Sensi et al. for “Antibiotic.”
As a matter of fact, this reaction is expedited in the formation of Rimactane by the addition of an oxidiz-ng agent, e.g., sodium persulfate, in acidic media.- '
Accord, R. Morrison and R. Boyd, Organic Chemistry (3d ed. 1973) at 878.
U.S. Patent No. 3,542,762 issued November 24, 1970, to Anacleto Gianantonio et ai. for “Process Por Rifamyeins.”
Appellee’s expert witness, Dr. Soloway, referred, in this regard, during his testimony at trial to the ■“Heisenberg Uncertainty Principle” which postulates that simultaneous measurements of an electron’s ¡position and momentum are intrinsically “uncertain.”