(ii) Japan

20. Japan responds that there is a sufficient amount of literature and scientific data which indicates the possible presence of a statistically significant difference in the efficacy of known disinfestation measures across varieties of the same product, and that such a difference could require application of a different treatment.
21. First, Japan submits that in the specific case of MB fumigation the link between varietal differences and the divergent efficacy of a fumigation treatment may manifest itself by way of a difference in the CxT value for different varieties, i.e., a difference in the relationship between the fumigant gas concentration in the fumigation chamber and the time-period of fumigation. ²³⁵ Japan specifically refers to three studies which allegedly demonstrate a statistically significant difference in CxT values between tested variety samples: ²³⁶ (1) 1985 tests on three varieties of American walnuts where, according to Japan, CxT values were significantly different between the Franquett variety and the Payne variety²³⁷; (2) 1988 tests on three varieties of American nectarines where, according to Japan, CxT values of the May Diamond variety showed a statistically significant difference from the other two cultivars for most of the doses²³⁸; and (3) 1997 tests on three varieties of Japanese nectarines where there was, according to Japan, a statistically significant difference in CxT values between the Shuhou variety and the Fantasia variety. ²³⁹
22. According to Japan, differences in CxT values between varieties could be an indicator of differences in the efficacy of a fumigation treatment. Japan argues as follows: When MB gas is injected into the fumigation chamber to disinfest a particular load of fruit, it is absorbed by the surface or the pulp of the fruit. If the degree of sorption varied depending on the variety of the fruit, the amount of fumigant remaining in the chamber air will vary in an inverse relationship to the sorption. Then the CxT value, which is determined by the gas concentration remaining in the chamber and which is a known indicator to control the degree of efficacy of the treatment, will vary as well depending on the variety of the fruit. According to Japan, the physical and chemical properties of fruits are factors affecting sorption (e.g., differences in oil content or a rougher surface) and can be attributable to varietal characteristics. Therefore, Japan concludes, the reported differences in CxT value are due to varietal differences and are an indicator of differences in the efficacy of a fumigation treatment.
23. Second, Japan submits three studies, derived from dose-mortality tests²⁴⁰, which indicate a difference in LD 50 values (i.e., the level of dose required in the tests to kill 50 per cent of all codling moths²⁴¹) between tested variety samples: ²⁴² (1) 1987 tests on six varieties of American nectarines where, according to Japan, one of the varieties tested, namely Summer Grand, was found significantly more susceptible to MB fumigation (i.e., had a lower LD₅₀ value) than the others²⁴³; (2) 1987/1988 tests on five varieties of New Zealand cherries where, according to Japan, the LD₅₀ value for the Bing variety was significantly lower than that for two of the other varieties tested, namely Rainer and Sam²⁴⁴; and (3) 1983/1984 tests on two varieties of New Zealand nectarines where, according to Japan, the Fantasia variety showed a significantly lower LD₅₀ value than the Redgold variety. ²⁴⁵
24. Japan agrees with the United States that there may be a range of exogenous factors (e.g., differences in leakage of the fumigation chamber, fruit load, experimental errors, sorption by packaging material, natural variation of pest population and fruit-to-fruit variation such as different ripening times, seasonal variations and physical condition of the fruit) which may also account for the differential in CxT and LD₅₀ values reflected in the studies it refers to. However, Japan claims that most of these other variables can be controlled in such a manner as to minimize their effects and that such control is a normal practice for scientists. The Japanese hypothesis is that characteristics of a particular variety may affect fumigation efficacy and that there is not sufficient evidence to disprove this possibility. This is, according to Japan, a reasonable argument. For Japan, its policy of variety-by-variety testing is therefore based on a scientific hypothesis which, in turn, is supported by empirical data, in full conformity with the obligations contained in Article 2.2.
25. Japan acknowledges that existing treatment levels of host plants of codling moth have been found effective for additional varieties. However, for Japan, all this proves is the efficacy of the treatment on the tested varieties. According to Japan, this falls short of showing absence of varietal difference within a product altogether. Japan notes that only a limited number of varieties have been tested in full-scale trials. On the question of a buffer, Japan claims that in large-scale trials the buffer is not always added to the "highest minimum dose", as the United States argues. Japan submits that in some instances the amount of fumigant absorbed by, inter alia, bins or the interior of warehouses may exceed the 10 to 20 per cent buffer. Japan further refers to Dr. Ducom's response to Panel question 12 relating to the buffer ²⁴⁶ and highlights the uncertainty of the effect of the buffer. On these grounds, Japan concludes that the United States has not performed the required demonstration of efficacy of a treatment across all varieties.

(b) Is the varietal testing requirement maintained without "sufficient scientific evidence" in the sense of Article 2.2?

26. We first examine that part of Article 2.2 requiring Japan to "ensure that [the varietal testing requirement] � is not maintained without sufficient scientific evidence".
27. We recall that Article 2.2 provides for an alternative to the requirement not to maintain phytosanitary measures without sufficient scientific evidence, namely to adopt provisional measures in accordance with Article 5.7. ²⁴⁷ Whether Japan can validly invoke Article 5.7 in this dispute, is addressed below in paragraphs 8.48 and following.

(i) The meaning of a measure "maintained without sufficient scientific evidence"

28. As referred to above²⁴⁸, the general obligations in Article 2.2 have to be read together with the more specific obligation imposed on Japan in Article 5.1, namely the obligation to ensure that the varietal testing requirement is "based on" a risk assessment. The Appellate Body, in its report on EC � Hormones, elaborated on the meaning of the term "based on" as used in Article 5.1 as follows:

"We believe that "based on" is appropriately taken to refer to a certain objective relationship between two elements, that is to say, to an objective situation that persists and is observable between an SPS measure and a risk assessment". ²⁴⁹

"We believe that Article 5.1, when contextually read as it should be, in conjunction with and as informed by Article 2.2 of the SPS Agreement, requires that the results of the risk assessment must sufficiently warrant � that is to say, reasonably support � the SPS measure at stake. The requirement that an SPS measure be "based on" a risk assessment is a substantive requirement that there be a rational relationship between the measure and the risk assessment". ²⁵⁰

29. We consider this statement (with respect to Article 5.1) to provide guidance also for our examination as to whether the varietal testing requirement is "maintained without" sufficient scientific evidence (in the sense of Article 2.2). In our view, for a phytosanitary measure to be "maintained without" sufficient scientific evidence, there needs to be a lack of an objective or rational relationship between, on the one hand, the phytosanitary measure at stake (in casu, the varietal testing requirement) and, on the other hand, the scientific evidence submitted before the Panel (in casu, in particular the six studies referred to by Japan²⁵¹).
30. When conducting this examination, we consider it to be important to make a clear distinction between (1) the Japanese requirement that it is for the exporting country (in casu, the United States) to demonstrate the efficacy of the quarantine treatment it proposes in order to gain access to the Japanese market for certain products and (2) the Japanese requirement that the exporting country (in casu, the United States) needs to make such demonstration for each variety of a given product. The United States does not contest the first requirement. ²⁵² It accepts that it needs to demonstrate quarantine efficacy. Only the fact that it needs to do so for each variety, i.e., only the second requirement (the varietal testing requirement), is at issue in this dispute. Under Article 2.2, Japan has the obligation not to maintain this requirement without sufficient scientific evidence. ²⁵³
31. Our task in this dispute is to determine whether or not Japan, to date, is in breach of this obligation²⁵⁴; not whether in the future scientific evidence could be produced which would allow Japan to comply with its obligation. ²⁵⁵ If to date there were not sufficient scientific evidence in support of the varietal testing requirement, Japan would be in breach of its obligations under the SPS Agreement. ²⁵⁶

(ii) The opinions of the scientific experts advising the Panel

32. To determine whether or not the varietal testing requirement is maintained without sufficient scientific evidence (i.e., whether there is a lack of an objective or rational relationship between the measure at issue and the scientific evidence before the Panel), we need to refer to the opinions we received from the experts advising the Panel. ²⁵⁷ We recall that these expert opinions are opinions on the evidence submitted by the parties. We are not empowered, nor are the experts advising the Panel, to conduct our own risk assessment. ²⁵⁸
33. At the end of our meeting with the experts advising the Panel, we requested them to confirm a number of understandings we had drawn from their answers and statements. The experts unanimously confirmed the following understandings:

• First, referring to the evidence before the Panel, there may be differences between varieties of the products in dispute which may, in turn, be relevant for quarantine purposes, i.e., which may affect the efficacy of an MB treatment approved for one variety of a product if applied to another variety of the same product. ²⁵⁹
• Second, the question whether varietal differences, if any, are significant for quarantine purposes cannot be determined on the basis of the evidence before the Panel ²⁶⁰
• Third, if, and to the extent that, differences between varieties are significant for quarantine purposes, they are mainly or even exclusively related to different levels of sorption of the fruit. ²⁶¹

As noted earlier²⁶², the scientific evidence before the Panel (i.e., the evidence evaluated by the experts advising the Panel) relates to either apples, cherries, nectarines or walnuts. However, in the view of the experts advising the Panel the understandings referred to above equally apply to the other products at issue (apricot, pear, plum and quince). ²⁶³

34. Replying to a US question (posed at the meeting with the experts) as to whether the experts are aware of any situation where differences in variety have resulted in a different treatment level for the products at issue in this case, Mr. Taylor stated: "No, I have no information or have seen any published data so the answer I have to give is no"; Dr. Heather replied: "In my experience there have been no differences of this kind. In fact it's been to the contrary. Most of my experience has been with insecticide dips with the material dimethoate and here we find that the same treatment not only goes across varieties but across commodities but I can see that this sorption problem with methyl bromide is something very special and that's why I wish to defer to my colleagues with experience as fumigation experts".²⁶⁴
35. When we asked the experts advising the Panel whether, in their expert opinion, there is an objective or rational relationship between, on the one hand, the varietal testing requirement imposed by Japan for MB treatment and, on the other hand, any of the evidence submitted by the parties, they stated unanimously that � even though in theory there may be relevant varietal differences � to date there is not sufficient evidence in support of the varietal testing requirement. ²⁶⁵
36. In his written answer to Panel question 16, Dr. Ducom states the following:

"The arguments put forth by Japan for requiring varietal trials are not based on scientific data. They are supported by a few experimental data in which varietal difference exists, in terms of LD₅₀, among a lot of other data in which it does not. These observations lead them to suspect all existing varieties and even more so those of the future, in which, in their eyes, genetic engineering and biotechnology might well create even greater differences. This is not based on any scientific data".²⁶⁶

37. According to the experts advising the Panel, there is scientific evidence before us � in the form of small-scale dose-mortality tests ²⁶⁷ carried out on different varieties of the same product � which indicate different test results (either a different CxT value ²⁶⁸ or a different LD₅₀ value²⁶⁹) for different varieties. ²⁷⁰ This cannot be disputed. The statistical and biological relevance of the differences in these test results and, especially, the factors causing them are less clear.
38. First, the experts advising the Panel question the value to be attached to the test results for purposes of checking quarantine efficacy. ²⁷¹ They express doubts as to whether LD₅₀ values derived from dose-mortality tests can be used to compare the efficacy of quarantine treatment between varieties. ²⁷² They also note that even though the test results � both with respect to CxT and LD₅₀ values � show statistical differences between varieties, biologically speaking these differences are not pronounced. ²⁷³
39. Second, and more importantly, the experts advising the Panel are of the view that even if confidence is given to these differences in the test results there is no evidence before the Panel that these differences � both in CxT and LD₅₀ values � are due to varietal differences.
40. In the studies referred to by Japan, the same tests were carried out on different varieties and in some instances the results differed. However, according to the experts advising the Panel, these differences could have been caused by a series of factors which are not related to varietal differences, such as differences in leakage of the fumigation chamber, fruit load, experimental errors, sorption by packaging material, natural variation of the pest population and fruit-to-fruit variation such as different ripening times, seasonal variations and physical condition of the fruit. Japan contends that experimental factors can be and were controlled to the extent possible. However, Japan does not contest that other factors (not related to varietal differences) could also explain the differences in the test results. The differences might also be linked to varietal differences. However, on this Dr. Ducom states the following (confirmed by the other two experts advising the Panel²⁷⁴):

"It is impossible by a simple DMT [dose-mortality test] to find out the relevant impact of the factors playing a role in the varietal differences � mainly because varieties ripen at different times � The DMT presented by the parties are designed to give information on insect sensitivity. The search for possible causes of varietal variations cannot be determined with precision by them, but only with a specific research program".²⁷⁵

"There is a lack of precise studies on this subject. A priori, one could cite almost anything, the size of the fruit, the nature of the epidermis, the average sugar content, the ripeness of the fruit, its physiological condition, the time between harvest and fumigation, etc. However, in this, there are hypotheses that merit studying". ²⁷⁶

41. The experts advising the Panel point out that so far no attempt has been made to determine whether varietal differences actually constitute a factor which causes the differences in the test results. ²⁷⁷ And this even though, according to the experts advising the Panel, technically speaking such determination can relatively easily be made, for example, by conducting sorption tests on different varieties of a product. ²⁷⁸ Japan did not further test or try to confirm its so-called hypothesis according to which varietal differences affect quarantine efficacy. However, as Japan itself notes: "The task of scientific demonstration begins, not ends, with a discovery of variables". ²⁷⁹

To continue with Evaluation by the Panel

²³⁵ The concept of a CxT value is further explained in paragraph 2.11.

²³⁶ These studies are described in more detail in the descriptive part of our report, paragraphs 4.109 - 4.135.

²³⁷ Vail, P. V. et al., Walnut On-Site Operational (Demonstration) Test Report to Japanese MAFF, USDA/ARS, Horticultural Crops Research Laboratory, Fresno, California, 2-14 December 1985.

²³⁸ Vail, P. V. et al., Report on Efficacy of Methyl Bromide for Codling Moth on Nectarines: Consideration of Nectarines as a Product Group, Prepared for Approval by the Japanese MAFF, Horticultural Crops Research Laboratory, Fresno, California, December 1988.

²³⁹ Research Division, Yokohama Plant Protection Station, MAFF, 1997, Unpublished.

²⁴⁰ The meaning of a "dose mortality test" is explained in paragraph 2.12.

²⁴¹ The notion of LD values is further explained in paragraph 2.14.

²⁴² These studies are dealt with in more detail in paragraphs 4.81 - 4.108.

²⁴³ Yokoyama V.Y. et al., Methyl Bromide Fumigation for Quarantine Control of Codling Moth (Lepidoptera: Tortricidae) on Nectarines, Journal of Economic Entomology 80, 1987, pp. 840-842.

²⁴⁴ Waddel, B.C. et al., Disinfestation of New Zealand Cherries, Cultivar Comparison Test 1987/1988, Department of Scientific and Industrial Research, Auckland, June 1988.

²⁴⁵ Batchelor, T.A. et al., Disinfestation of New Zealand Nectarines 1983/1984, Department of Scientific and Industrial Research, July 1984.

²⁴⁶ See paragraphs 6.87 and following.

²⁴⁷ Article 2.2 provides that "Members shall ensure that any � phytosanitary measure � is not maintained without sufficient scientific evidence, except as provided for in paragraph 7 of Article 5" (emphasis added).

²⁴⁸ See paragraph 8.15.

²⁴⁹ Op. cit., paragraph 189, pp. 76-77, italics in original, underlining added.

²⁵⁰ Op. cit., paragraph 193, underlining added.

²⁵¹ See paragraphs 8.21 and 8.23.

²⁵² See paragraph 8.10.

²⁵³ Unless the Japanese measure is imposed in accordance with Article 5.7, a provision to which Article 2.2 explicitly refers. As noted earlier, this question is dealt with in paragraphs 8.48 and following.

²⁵⁴ In this respect, we recall the rules on burden of proof we set out earlier in paragraph 8.13.

²⁵⁵ Article 11 of the DSU directs us to "make an objective assessment of the matter before [the Panel], including an objective assessment of the facts of the case".

²⁵⁶ Assuming the measure cannot be considered to be a provisional measure in accordance with Article 5.7, an issue we examine below in paragraphs 8.48 and following.

²⁵⁷ For the procedures we followed to appoint these experts and to obtain their views, see paragraphs 6.1 and following.

²⁵⁸ See also panel report on Australia � Salmon, op. cit., paragraphs 8.41, 8.126 and 8.172.

²⁵⁹ See Transcript, paragraphs 10.268 - 10.273. Dr. Heather stated the following (confirmed by the other two experts):

²⁶⁰ See Transcript, paragraphs 10.274 - 10.279. As noted in the previous footnote, Dr. Heather stated unambiguously:

See also the introductory statement by Dr. Ducom:

See also Mr. Taylor:

²⁶¹ Transcript, paragraphs 10.280 - 10.285. See also, for example, the introductory statement by Mr. Taylor:

See also Mr. Taylor, Transcript, paragraph 10.140:

²⁶² See paragraph 8.6.

²⁶³ Transcript, paragraphs 10.223 - 10.225.

²⁶⁴ Transcript, paragraphs 10.155 - 10.158. The question was not directed at Dr. Ducom who did not answer it.

²⁶⁵ Transcript, paragraphs 10.167 - 10.174:

²⁶⁶ See paragraphs 6.104 - 6.105.

²⁶⁷ The concept of a dose-mortality test is explained in paragraph 2.12.

²⁶⁸ The concept of a CxT value is explained in paragraphs 2.11 and 8.21.

²⁶⁹ The concept of an LD50 value is explained in paragraphs 2.14 and 8.23.

²⁷⁰ See paragraphs 8.21 - 8.23.

²⁷¹ In this respect, it should, from the outset, be recalled that all the studies referred to by Japan were designed and carried out in order to comply with the varietal testing requirement. None of the studies before us specifically examines the appropriateness of the requirement itself.

²⁷² See answers to Panel question 1 by Dr. Ducom: "In practice, the LD50 test constitutes a fairly unreliable method to compare the efficacy of quarantine" and Mr. Taylor: "LD50 values are extremely useful in comparing the toxicity of different chemicals and in the measurement of resistance. However, these values are less useful in investigations of much higher levels of toxic response such as are necessary in relation to quarantine treatments, where LD values of 99 or 99.9 are more appropriate and useful". In his answer to Panel question 12, at para. 6.8, however, Dr. Ducom noted that "while the dose-mortality test (LD₅₀) did not give any confidence in respect of the varietal factor, it did give an indication of the relative sensitivity of the products tested". See also Dr. Heather's statement at the meeting with the experts:

²⁷³ At the meeting with the experts Dr. Ducom stated:

See also the statement by Dr. Heather at the meeting with the experts:

²⁷⁴ Addressing, for example, a study conducted on three varieties of nectarines and 13 apple varieties (Kawakami, F., et. al., Methyl Bromide Sorption in Fruit Varieties, Research Division, Yokohama Plant Protection Station, Japan Exhibit 36), Dr. Heather, speaking on behalf of all three experts, stated:

With respect to the same study, Dr. Ducom noted the following:

See also Dr. Heather's answers to Panel question 1: "� where differences between varieties are small, fruit to fruit variation could greatly exceed variety to variety variation. Such variation is an inherent characteristic and is usually overcome by ensuring adequate robustness of the treatment"; to Panel question 3: "Although statistical differences are evident between some varietally based experiments, this does not provide an assurance that the origin of the difference lies predominantly in varietal characteristics"; and to Panel question 8: "If the criterion used is statistically significant differences between experimental samples of different varieties there are differences as identified by Japan, but there is no certainty that they are attributable to unique varietal characteristics. In subsequent research on additional varietal samples these differences were too small to cause the efficacy of a treatment based on varieties used in initial trials, to fail in further testing".

²⁷⁵ Dr. Ducom's answer to Panel question 3.

²⁷⁶ Dr. Ducom's answer to Panel question 9. Answering a US question at the meeting with the experts � as to whether varietal differences are widely known to result in significant differences in efficacy of treatment as opposed to a number of sources of fruit variation which include temperature, moisture, daylight, rainfall, cultivation conditions and other natural conditions of the harvest year, which according to Japan are not widely known to result in significant differences in efficacy of treatment � Dr. Ducom stated as follows:

²⁷⁷ See footnote 260.

²⁷⁸ Since, according to the experts advising the Panel, varietal differences, if and to the extent they exist, would mainly or exclusively be due to differences in sorption, tests could, for example, be conducted on different varieties of a product to check whether there are any such differences in sorption. See Dr. Ducom's statement at the meeting with experts:

See also the following statement of Mr. Taylor at the meeting with the experts:

See also statements by Dr. Ducom and Mr. Taylor, Transcript, paragraphs 10.187 - 10.196.

²⁷⁹ Second submission of Japan, p. 15.