HM Revenue & Customs v GE ION Track Ltd

Judgment

MR JUSTICE BRIGGS:

1. This is an appeal by the Commissioners for Revenue and Customs against a decision of the VAT and Duties Tribunal, given on 27^th October 2005. The Tribunal held, allowing an appeal by GE Ion Track Limited, against the decision contained in a review letter dated 18^th February 2005, that the correct classification of three products imported by GE Ion Track, namely; Entry Scan, Vapour Tracer, and Itemiser, to which I will refer as “the products” under the combined nomenclature, to which I will refer as “CN,” was under CN heading 9027, rather than, as the Commissioners had asserted in a binding tariff information on 14^th July 2004, and the review process had determined, CN heading 9022.

2. The relevant facts about the products are to be found summarised in paragraphs 2 to 3 of the decision of the Tribunal:

“2. The products are designed for the detection of explosives and narcotics, for example at airports. Entryscan looks like the type of machine one walks through at airport security. VaporTracer is a hand-held device and Itemiser is a desktop device. All three Products take a sample of air and analyse it using the Appellant’s patented Ion Trap Mobility Spectrometer (“ITMS”) technology, which is a type of Ion Mobility Spectrometry. These are described in the Appellant’s brochure as follows:

‘IMS separates ionised molecular compounds on the basis their transit times (sometimes called ‘time of flight’ or ‘drift time’) when subjected to an electric field in a tube. This time is then compared to stored transit times of known compounds making it possible to distinguish the target material (explosives or narcotics) from other molecules. ...[G]aseous samples enter an ionisation chamber where an ionisation source emits low-energy beta particles resulting in ion formation in the gaseous phase. A gating mechanism allows the ions of the correct polarity to pass through the shutter grid and enter the ion drift region where an applied electric field mobilises the ions... The rate at which these ions traverse the ion drift region is inversely proportional to the size of the ion. This correlation allows for the identification of the analyte of the interest...’

ITMS is described in the brochure as follows:

‘ITMS, like IMS, separates ionised vapours and then measures the mobility of the ions in an electric field. In the typical implementation of ITMS, the gaseous sample passes through a semi-permeable membrane prior to ionisation. Also like IMS, the gaseous samples then enter an ionisation chamber where an ionisation source emits low-energy beta particles resulting in ion formation in the gaseous phase. Unlike IMS, however, the ionisation in ITMS is allowed to reach equilibrium in a field-free region and then pulsed into the drift tube where an electric field accelerates the ions to the collector. Note that in the ITMS detector, the shutter grid does not exist, resulting in a much greater portion of the ions entering the drift tube.’

“3. We heard expert evidence in the form of three reports and oral evidence from Mr Bevan John Clues, Chartered Electrical Engineer and a Member of the Institution of Electrical Engineers, who also commented on the Appellant’s expert’s report by Mr Ching Wu, who did not give evidence, but there was little disagreement between them. We make the following findings of fact from his evidence:

(1) Spectroscopy is the study of the interaction between electromagnetic radiation and matter. The Products are mass spectrometers. By way of comparison an optical spectrometer resolves a beam of light into components according to their wavelength i.e. colour component, whereas a mass spectrometer resolves a beam of positive ions according to their mass/charge ratio, or, if all have a single elementary charge, according to their masses. There are different types of mass spectrometers; the distinguishing feature between them is in the manner in which separation is achieved. Essentially ions of the material being analysed have to be produced.

(2) An ion is an atom with either extra electrons or missing electrons. Thus an ion can be positively or negatively charged. Methods of producing the ions include sparks, heated filament and acceleration by the use of an electrical potential, electric field at a sharp point or edge, electrospray ionisation, radiation source, e.g. alpha, beta or gamma rays. The Products use beta radiation because it requires zero power consumption, and has long-term stability and universal ionisation ability, which makes it the best choice for portable instruments. The use of beta radiation generally is being replaced by other means, such as electrospray.

(3) The source of beta radiation is an integral and inseparable part of the Products. Beta rays are part of the electromagnetic spectrum, covering radio, infra-red light, visible light, ultra-violet rays, X-rays, gamma rays, beta rays and cosmic rays. Beta rays are streams of very fast electrons with speeds of up to within a few tenths of a per cent of the speed of light.

(4) The use of a drift tube is an established technique for mass spectrometers.

(5) Because the Products are based on an ionisation radiation source and a drift tube and that a beta source has been designed into the Products, in Mr Clues’ opinion the Products are “based on” beta radiation.’”

On one view the finding number 5 in paragraph 3 might be said to be a mixed conclusion of fact and law, but in any event it is common ground.

3. The relevant legal framework is also largely common ground, and I can summarise it for present purposes as follows: The European Union is a contracting party to the International Convention on the Harmonised Commodity Description Coding System, known as the Harmonised System, or HS. The Convention required that contracting states’ tariffs and statistical nomenclatures conform to the Harmonised System. That system is administered by the Customs Cooperation Council, and formerly known as the World Customs Organisation, or WCO. The WCO produces explanatory notes to the Harmonised System. All contracting states use the headings and sub-headings of the Harmonised System.

4. The official interpretation of the Harmonised System is given in four volumes of explanatory notes published by the WCO, and these are known as HSENs. The EU’s tariff and statistical nomenclature, “the Tariff,” is defined in Article 20.3 of Council Regulation 2913/92/EEC setting up the community customs code. Customs duties which are legally due on the importation of goods are determined by reference to the tariff. The tariff comprises, among other things, the combined nomenclature of goods of a CN. The CN is composed of three elements: (1) the nomenclature of the Harmonised System, or HS. (2) the community sub-divisions to the HS, and (3) the preliminary provisions, additional section or chapter notes and footnotes relating to the CN sub headings.

5. This case involves interpreting two headings of the CN which are, as they are required to be, identical to those contained in the HS. The CN is produced at Annex 1 of Council Regulation 2658/82/EEC; the tariff regulation. It is updated annually by the Commission. Part 1 contains, at section 1A, the General Rules for the Interpretation of the CN, or GIR. Part 2 contains the headings themselves, which amount to a comprehensive goods nomenclature, designed to classify all goods imported into the EU in an appropriate heading.

6. The European Court of Justice has repeatedly set out the requirement for objectivity in classifications, and the basic principle governing the court’s approach is set out, for example, in Holz Geenen v. Oberfinanzdirektion MünchenCase 309/98 at paragraph 14:

“It is settled case-law that, in the interests of legal certainty and for ease of verification, the decisive criterion for the classification of goods for customs purposes is in general to be sought in their objective characteristics and properties as defined in the wording of the relevant heading of the CN. The explanatory notes drawn up, as regards the CN, by the Commission and, as regards the HS, by the Customs Cooperation Council (‘the HSENs’), may be an important aid to the interpretation of the scope of the various tariff headings but do not have legally binding force (see Case C-405/97Mövenpick Deutschland v. Hauptzollamt Bremen [1999] ECR I-2397, paragraph 18).”

7. It is apparent from the ECJ decisions that the court pays regard regularly to the HSENs in interpreting the scope of the headings of the CN. Nonetheless, it is equally apparent that the HSENs are not to be applied where the effect would be to alter the meaning of the heading at issue, and I turn now to the particular headings in issue in this case. The first is heading 9022 and it provides as follows:

“Apparatus based on the use of X-rays or of alpha, beta or gamma radiations, whether or not for medical, surgical, dental or veterinary uses, including radiography or radiotherapy apparatus, X-ray tubes and other X-ray generators, high tension generators, control panels and desks, screens, examination or treatment tables, chairs and the like.”

The second is heading 9027, which provides as follows:

“Instruments and apparatus for physical or chemical analysis (for example, polarimeters, refractometers, spectrometers, gas or smoke analysis apparatus); instruments and apparatus for measuring or checking viscosity, porosity, expansion, surface tension or the like; instruments and apparatus for measuring or checking quantities of heat, sound or light (including exposure meters); mocrotomes.”

8. The appropriate classification of goods in the combined nomenclatures is governed by the principles of interpretation, which as I have said, I shall summarise as the GIR. Rule 1 of the GIR provides as follows:

“The titles of sections, chapters and sub-chapters are provided for ease of reference only; for legal purposes, classification shall be determined according to the terms of the headings and any relative section or chapter notes and, provided such headings or notes do not otherwise require, according to the following provisions.”

It is common ground that there are no section notes or chapter notes, which are relevant for the purposes of assisting in the classification of the products the subject matter of this appeal.

9. GIR 3 provides as follows:

“When, by application of rule 2(b) or for any other reason, goods are prima facie classifiable under two or more headings, classification shall be effected as follows:

(a) the heading which provides the most specific description shall be preferred to headings providing a more general description. However, when two or more headings each refer to part only of the materials or substances contained in mixed or composite goods or to part only of the items in a set put up for retail sale, those headings are to be regarded as equally specific in relation to those goods, even if one of them gives a more complete or precise description of the goods;

(b) mixtures, composite goods consisting of different materials or made up of different components, and goods put up in sets for retail sale, which cannot be classified by reference to 3(a), shall be classified as if they consisted of the material or component which gives them their essential character, in so far as this criterion is applicable;

(c) when goods cannot be classified by reference to 3(a) or (b), they shall be classified under the heading which occurs last in numerical order among those which equally merit consideration.”

10. I turn now to the relevant parts of the HSENs for each of headings 9022 and 9027. Under the general heading, “Apparatus Based on the Use of X-rays,” this HSEN provides under paragraph C as follows:

“X-ray apparatus for industrial use. There are many industrial applications of X-rays. They are used,, for example, in metallurgy to locate blisters or to check the uniformity of alloys; in engineering to check the accuracy of assemblies; in the electrical industry to check heavy cables or frosted glass lamps; in the rubber industry to check the reactions of the inner casings of tyres (e.g., stretching of canvases); in various industries for checking or measuring the thickness of materials. For all these various applications the apparatus generally resembles that used for diagnostic purposes described above, except that it may be equipped with adaptors and ancillary equipment for particular purposes.

The heading also covers:

(1) Special apparatus (X-ray diffraction and X-ray spectrometry equipment) used for the examination of the crystalline structure as well as the chemical composition of materials; the X-rays are diffracted by crystals and then made to fall on a photographic film or an electronic counter.”

And then under the heading: “Apparatus Based on the Use of Alpha, Beta or Gamma Radiations”:

“Alpha, beta or gamma radiations emanate from a radioactive substance with the property of emitting radiations by spontaneous transformation of its atoms. This radioactive substance is placed in a container, normally of steel coated with lead (bomb), which has an aperture designed to let the radiations pass in one direction only. Gamma radiations can be used for much the same purpose as X-rays.

The following types may be distinguished, according to the radiations employed and the use for which they are designed:

(1) Therapy apparatus, in which the radioactive source is a charge of radium, radio-cobalt or some other radioactive isotope.

(2) Apparatus for radiological examinations, used mainly in industry for checking metal parts, etc., without damaging their structure.

(3) Apparatus having a measuring instrument such as beta and gamma ray thickness gauges for measuring the thickness of materials (sheets, linings and the like), apparatus for monitoring the contents of packages containing any product (pharmaceutical products, foodstuffs, sporting gun cartridges, perfumes, etc.) or ionisation anemometers. In these apparatus, the required information is generally obtained by measuring the change in the amount of radiation applied to the factor under examination.

(4) Fire alarms incorporating smoke detectors containing a radioactive substance.

“The heading does not cover instruments and apparatus which are not designed to incorporate a radioactive source and which merely measure or detect radiation even when such instruments are calibrated in arbitrary terms (heading 90.30).”

11. The HSEN for heading 9027 includes the following relevant parts. Firstly: under the general title, “This heading includes,” sub-paragraph 5 is as follows:

“5) Spectrometers. These instruments are used to measure the wave-lengths of emission and absorption spectra. They consist essentially of an adjustable slit collimator (through which the beam of light to be analysed passes), one or more adjustable prisms, a telescope and a prism table. Some spectrometers (particularly those used for infra-red or ultra-violet rays) are fitted with prisms or with diffraction gratings.

This group includes: spectroscopes for the observation of spectra; spectrographs for recording the spectrum on a photographic plate or film (spectrograms); monochromators, instruments for isolating a particular line in a line of spectrum or for isolating certain parts of a continuous spectrum.

But the heading excludes spectroheliographs and spectrohelioscopes, used for solar observation (heading 90.05); spectrum projectors, for examining an enlarged spectrogram projected on to a screen (heading 90.08); micrometric microscopes and spectrocomparators incorporating microscopes (for comparative examination of spectrograms by optical observation) (heading 90.11) and spectrum analysers for measuring or checking electrical quantities (heading 90.30).”

12. Then at the end of the HSEN there are a number of exclusions lettered (a) to (ij) under the general heading, “The heading also excludes”: I need only read (e):

“X-ray, etc., apparatus (heading 9022).”

13. I turn now to a summary of the reasoning of the Tribunal -- and there is no point of distinction between each of the three products for the purposes of the present appeal -- which was as follows:

“The products prima facie fall within heading 9022, since they are, to a sufficient extent based on the use of beta radiation. (2) The products also fell prima facie within the terms of heading 9027 as being instruments for physical or chemical analysis, and as mass spectrometers were expressly included in the heading by reason of the reference in the heading to spectrometers. (3) To the extent that the HSEN for heading 9027 purports to restrict the meaning of spectrometers to optical spectrometers only, it went beyond a mere explanation of the wording. That is the terms of heading 9027. (4) The reference by way of exclusion of x-ray etc, apparatus in HSEN heading 9027 did not exclude any item merely because it had an x-ray, or as here, beta ray element within it. It merely pointed the reader to heading 9022 as a separate and possibly more appropriate classification for such equipment than heading 9027. Alternatively, if that exclusionary sub-paragraph did purport automatically to exclude such apparatus from heading 9027, again, it went impermissibly beyond a mere explanation of the terms of the heading. (5) Therefore, the products fell prima facie within heading 9027 as well as heading 9022. (6) Applying the tie-breaker rule in GIR 3A, heading 9027 was a more specific description of the products as instruments for chemical analysis, since whereas heading 9022 referred only to one aspect of the mechanism of the products, heading 9027 referred to the function of the products as a whole.”

14. The Commissioners challenged that analysis as wrong in law. Mr Owain Thomas, who appeared for the Commissioners, submitted in summary as follows:

(1) The Tribunal was wrong to treat exclusion (e) in the HSEN for heading 9027 as merely a pointer that there was a separate heading for x-ray etc apparatus, which might be more appropriate.

(2) The Tribunal was also wrong to treat the HSEN for heading 9027 as in any way contrary to the wording of the heading, either in limiting the types of spectrometer to which it applied, or in excluding x-ray apparatus properly described as a spectrometer.

(3) Even if both headings 9022 and 9027 were prima facie applicable, the Tribunal ought to have construed heading 9022 as containing the most specific description of the products within the meaning of GIR 3A.”

15. The Commissioners’ case disclosed a large and potentially far-reaching difference of approach to the purpose and legal effect of HSENs, and in particular, the frequently found exclusions in them, from that adopted either by the Tribunal, or by the Respondent in its submissions before me, and I will deal with that issue first. Mr Thomas said, in relation to this issue, as follows:

(1) The HSENs are a vital tool for identifying, when applying GIR 1, the single heading applicable to a particular product.

(2) Wherever an HSEN for a particular heading excluded a type of product on the basis that some other heading applied to it, that meant that wherever that other heading prima facie applied, the instant heading is dis-applied, in other words excluded, so that the decision-maker, having regard to the terms of the HSENs, never in those circumstances, has to apply GIR 3, and he submitted that this was so even if, ignoring the exclusion in the HSEN, the application of GIR 3 would, as the Tribunal thought in this case, lead to the opposite conclusion.

(3) He submitted that this was how HSEN exclusions were habitually used, and that the result produced in relative terms simplicity, international uniformity, and predictability, none of which would be realised by a constant need for recourse to the judgments called for in applying GIR 3. The HSENs were, he submitted, there to avoid that.”

Mr Thomas gained some assistance in support of his submission from the decision of Lawrence Collins J, in VTech Electronics (UK) Ltd v. Commissioners of Customs and Excise[2003] EWHC Ch 59. That case was about whether a particular product modelled like a computer, but which contained electronic functions designed for use by children was a toy under heading 9503, or a game, under heading 9504. The Tribunal had decided, applying GIR 1, that it was a toy, despite having some characteristics of a game, and was assisted in reaching that conclusion by reference to the relevant HSENs. The passage of the judgment relied upon by Thomas is at paragraphs 105 to 107:

“105. Both the Regulation and the Tribunal (in its primary finding on this aspect) cannot be criticised for stopping at GIR rule 1 (and the corresponding rule 6) and not going on to apply rule 3. I accept the submission for the Commissioners that goods are not considered to be "prima facie classifiable under two or more headings", with the result that rule 3 must be invoked, purely by virtue of the fact that they could technically be brought under two or more headings, and that the purpose of rule 3 is to arbitrate between headings whose application remains finely-balanced.

106. There are several decisions in the European Court where products were capable of being classified under more than one heading, and which were decided without reference to rule 3. Thus in Case C-177/91 Bioforce [1993] ECR I-45 the product (extract of hawthorn with added alcohol, taken as a heart tonic) was capable of being classified both as a medicament and as a spirituous beverage. Both Advocate General Gulmann and the European Court (who came to opposite conclusions) arrived at a classification without reference to rule 3. In Case C-338/95 Wiener SI GmbH v. Hauptzollamt Emmerich [1997] ECR I-6495 the product was capable of being worn both as a " nightdress" and as a "dress", but the classification was resolved without recourse to rule 3, after consideration of the HSENs and on the basis of the objective characteristics of the product as reflected particularly in its intended use.

107. I also accept the Commissioners' contention that these cases confirm that the purpose of examining the "objective characteristics" of a product, and of having regard in doing so to such constructional aids as the HSENs and the intended use of the product, is to find the category in which it should be placed, and not to assemble a list of theoretically possible but increasingly implausible categories between which the provisions of rule 3 must be used to decide.”

16. Ms Marie Demetriou, who appeared for the Respondent, submitted on this issue as follows:

(1) The purpose and legal effect of HSENs is governed by their third ranking in the hierarchy of applicable legal norms relevant to the classification process.

(2) The terms of the headings themselves, and any relevant section, or chapter notes come first in that hierarchy as required by GIR 1. (3) Unless those terms and relevant section, or chapter notes otherwise require, the other GIRs, General Rules for Interpretation, must also be applied. They also have the force of law. Again, she relied upon the language of GIR 1. That necessarily, she submitted, imports GIR 3, the tie-breaker provisions, in any appropriate case.

(4) That HSENs are not of legal force, and never more than an aid to construction.

(5) There are numerous exclusions to be found expressly stated, both in headings, and in section, or chapter notes.”

She referred me in particular to the chapter notes at the beginning of chapter 90 itself, which contained a series of express exclusions of particular types of articles from the whole of chapter 90, including, for example, paragraph 5, which excluded:

“Measuring or checking optical instruments, appliances or machines which, but for this note, could be classified both in heading 9013 and in heading 9031 are to be classified in heading 9031.”

She also referred me to a series of express exclusions in headings 9006, 9010, 9013, 9017, 9026, 9029, 9030, 9031, and for convenience, to give examples of the two types of exclusion which appear there, I refer to 9010, which reads as follows:

“Apparatus or equipment for photographic (including cinematographic) laboratories (including apparatus for the projection or drawing of circuit patterns on sensitised semi-conductor materials), not specified or included elsewhere in this chapter; negatascopes; projection screens.”

That is an example of a residual heading, which applies only if there is no reference to such equipment elsewhere in chapter 90, and I can refer to heading 9026 to illustrate the other main type of express exclusion. That heading reading as follows:

“Instruments and apparatus for measuring or checking the flow, level, pressure or other variables of liquids or gases (for example, flow meters, level gauges, manometers, heat meters), excluding instruments and apparatus of heading 9014, 9015, 9028 or 9032.”

17. Miss Demetriou submitted that where those express exclusions apply, then whenever a particular product is capable of falling within another named heading it is automatically excluded from the instant heading with no need for recourse to GIR 3. But next she submitted that the purpose of the exclusionary sections frequently found in HSENs, and in particular, exclusion (e) in the HSEN for heading 9027, is to point the reader by way of shorthand to another heading, or headings, which in relation to particular types of product will usually, but not invariably be either the only appropriate heading, or the more appropriate of two or more prima facie applicable headings, applying the terms of the headings, the section chapter notes, and the GIR as a whole, including of course, Rule 3.

18. Next, she submitted that like the HSENs generally, the exclusion in the HSENs must not usurp the true meaning of the binding provisions, and finally, she submitted, to the extent that Lawrence Collins J in the VTech case suggested that GIR 3 was only a rule of last resort where, after using the HSENs, the choice between two or more headings remained finely balanced, that decision was wrong.

19. I have not found the choice between those two very different submissions about the nature and effect of HSENs, and in particular their exclusionary provisions, easy. But in my judgment, Ms Demetriou’s submissions are broadly to be preferred to those of Mr Thomas. My reasons are as follows; (1) The unanimous jurisprudence of the European Court of Justice is that the HSENs are not of legal force, but only a guide to construction to the terms of the headings, the section and chapter notes, and the GIRs, all of which are the legally binding structure for classification purposes.

(2) Nothing in the VTech case could or even purports to require a contrary conclusion.

(3) It cannot be right, as the Commissioners seek to do, to treat the exclusionary notes in HSENs as a separate self-standing code for the resolution of apparent ties between headings, independent of and to be used before any reference is made to GIR 3, so that GIR 3 is excluded in any case where an HSEN exclusion breaks the tie.

20. My reasons for that particular conclusion are as follows: (a) GIR 1 calls for the use of Rule 3, save where the heading in question, or any relevant section, or chapter notes otherwise requires. (b) Where the heading, and the section, or chapter notes do require a rigid exclusion in favour of another heading, they expressly so provide. Mr Thomas had no convincing explanation for splitting two sets of rigid exclusionary provisions between, on the one hand, the binding structure of the headings and the section and chapter notes, and on the other hand the explanatory notes and the HSENs, except for the avoidance of prolixity. But in fact, when one comes to read them, the exclusionary provisions of the HSENs are very short and not prolix at all. (c) Far from being a last resort the first part of GIR 3A is simple common sense: the specific excludes the general. The HSEN exclusions do point to other headings, which will usually be more specific, and more appropriate. For an example: an x-ray machine expressly referred to in exclusion (e) in the HSEN for heading 9027, is an instrument for physical or chemical analysis, but the 9027 heading is in that sense very general, whereas the 9022 heading, which as its own HSEN explains, generally applies to machines which use the x, or gamma, or beta, or other rays as the measuring device, is more specific. In such cases heading 9022 will be the more specific description. (d) The treatment of HSEN exclusions as a rigid code will produce counter intuitive results, for example, where a product is technically within another heading, let us say, heading B, but more specifically described in the instant heading, that is, heading A. If the effect of the exclusion of the instant heading in the HSEN for heading A in favour of heading B is rigidly applied, it will produce a worse mischief than was castigated in VTech, because the product would fall outside the heading where it naturally belonged and into a heading where it only gained a technical foothold. (e) The present is a case in point. The products are only based on beta rays, within the meaning of heading 9022, because the rays are used to ionise the subject matter for analysis. The rays are not used as the means of analysis.

21. The Tribunal thought, correctly in my judgment as I will explain in due course, that heading 9027 was a more specific description of these products than heading 9022. In reply, Mr Thomas modified his submission by saying that the HSEN exclusion applies in favour of another heading only if the product really does fall within the terms of that other heading, rather than merely coming technically within its confines. But that imports a judgmental analysis, which is either the same as the analysis required by GIR 3A, or if not, a different tie-break analysis than that which GIR 3 bindingly requires. If it is the latter, then the HSEN must give way - see, for example, the decision of the European Court of Justice in Develop Dr Eisbein -v- Hautpzollamt Stuttgart-West, Case C35/94, in particular at paragraphs 13 to 22.

22. (4) I do not read the VTech case as pointing to any different conclusion from that which I have set out above. There, the Appellant’s case was that GIR 3C applied, which, it will be recalled, reads as follows:

“When goods cannot be classified by reference to 3A or B, they should be classified under the heading which occurs last in numerical order among those which equally merit consideration.”

That rule is truly a last resort.

23. No doubt Lawrence Collins J was correct to say that a careful reading of the relevant terms of the relevant headings of any relevant chapter and section headings, assisted, but not governed, by the HSENs, while using them as a guide to interpretation, will often identify only one appropriate heading, rather than calling for a Rule 3 tie-breaker as it did in that case. Often the positive rather than the exclusionary provisions of the relevant HSENs will shed brighter light on the appropriate heading to be adopted.

24. I have not ignored the practical advantages of treating HSEN exclusions as a rigid code, rather than as a guide to the answer to be found by applying the terms of the headings, the section and chapter notes, and the GIR as a whole. Nor have I ignored the respect obviously due to the HSEN as the considered product of an international committee, highly experienced in a difficult field. But the combined nomenclature must be applied and interpreted in accordance with law, and the HSENs must be treated as a servant rather than as a master in that process.

25. I recognise also that as Mr Thomas submitted, looking at the exclusionary sections of the HSENs on their own, they are, in their note form, worded more as rigid rules than as pointers, but that is to ignore the status of the HSENs as a whole, which is as guidelines, or aids to interpretation, rather than rules.

26. It follows, in my judgment, that the Tribunal was correct to treat the question whether heading 9027 as well as 9022 were prima facie applicable as a question turning on the terms of the headings, there being no applicable section or chapter notes, and to treat exclusion (e) in the HSEN for heading 9027, as a pointer rather than as a rule. They were also correct on that basis to conclude that the products prima facie fell within both heading 9027 and 9022.

27. In my judgment, subject to one point to which I will come, that part of the decision -- that is paragraphs 16 to 20 -- cannot be faulted in law, and the Tribunal were therefore called upon to apply GIR 3 to resolve the tie.

28. The point to which I have just referred, is that I would not construe the earlier part of the text of the HSEN to heading 9027 as stating that heading 9027 applies only to optical spectrometers. There may be other spectrometers, which are neither optical nor based on x-rays or other relevant rays. But that is a small point and reinforces rather than undermines the Tribunal’s conclusion that heading 9027 prima facie applied.

29. I turn to the second question; namely whether the Tribunal erred in law in their application of GIR 3. They expressed their decision in paragraph 21 of the decision, as follows:

“In order to resolve the dual headings we apply GIR 3(a) that the “heading which provides the most specific description shall be preferred to heading providing a more general description.” It is difficult to weigh up the specificity of two headings which are different in nature: heading 90.22 looks at the means by which the apparatus operates regardless of its use (“based on the use of...beta....radiation, whether or not for medical [etc] use”), and heading 90.27 looks at the function it performs (“for physical or chemical analysis”). The existence of this approach makes it likely that items will fall within two headings. The question is not which heading is wider in general, but which heading is a more specific description of the Products. In our view it is a more specific description to say that the apparatus is an instrument for chemical analysis, than to say that it is based on the use of beta radiation because the latter is merely the means by which one part of it operates. Analysis is a description of what the apparatus does as a whole, whereas being based on the use of beta radiation is a description of how only part of it operates, which is not even the part that carries out the analysis, but a necessary preliminary to the analysis by producing the ions which can be analysed. If the apparatus actually measured the beta radiation the question would have been more finely balanced but here the radiation performs a preliminary (but essential) function only. Mr Clues’ description of the apparatus as a mass spectrometer is essentially a description that it performs analysis. Accordingly we conclude that the heading providing the more specific description is 90.27.”

30. Ms Demetriou submitted that there is a necessary distinction to be drawn between, (a) construing GIR 3 and then (b), applying that construction to the products. Step (a), she submitted, gives rise to a question of law, whereas step (b) only gives rise to factual questions, as to which the Tribunal have the last word. For that purpose she relied upon the opinion of Advocate General Jacobs in Veiner v. Hautpzollamt Emericht (Case 338/95), and in particular, paragraphs 23 to 30 of his opinion, and the decision of Dyson Jin Commissioners of Customs and Excise and General Instrument UK Ltd[1999] EWHC Admin 255, and in particular at paragraph 28. She submitted that if the Tribunal’s answer at step (a) is correct then it is irrelevant that the court might agree or disagree with its conclusion at step (b).

31. Mr Thomas submitted that paragraph 21 of the Tribunal’s decision demonstrated a failure to appreciate, as a matter of interpretation, that heading 9022 is inherently more specific than heading 9027.

32. In my judgment the Tribunal’s reasoning cannot be faulted as a matter of law on this issue either. I have not found it necessary to make the distinction submitted as appropriate by Ms Demetriou. In my judgment paragraph 21 is plainly correct, both as a matter of law, and if necessary, fact.

33. It follows that I must dismiss this appeal.