(c) CxT Values

109. Japan claimed that in the specific case of MB fumigation, the link between varietal differences and divergent efficacy of the fumigation treatment could manifest itself by way of the difference in the CxT value. ⁹¹ The process could take the following sequence: When MB gas was injected into the fumigation chamber, it would be absorbed by the surface or the pulp of the fruits. If the sorption varied depending on the variety of the fruits, the amount of fumigant remaining in the chamber air would vary in an inverse relationship to the sorption. Then the CxT value, a known indicator to control the degree of efficacy of the treatment ⁹², would vary as well depending on the variety of the fruit. Japan claimed that it could reasonably be assumed, by ways of the CxT values, that there was a route by which the characteristics of a fruit of particular variety could affect the outcome of disinfestation efficacy.
110. Japan claimed that there were three empirical cases which demonstrated a statistically significant difference in the CxT value between tested variety samples, including the case of three cultivars of walnuts the United States referred to in their submission⁹³:

1. 1985 tests on three varieties (Hartley, Payne and Franquett) of American walnuts (1985 Walnut Test Report): CxT values had been significantly different between Franquett (75.7 gh/m³ at the first replicate; 71.5 gh/ m³ at the second replicate) and Payne (109.4 gh/m³, and 99.9 gh/m³, respectively). ⁹⁴

Table 8 Fumigant Concentration and CxT Values for 3 Varieties of American Walnuts
Variety	Replicates	Fumigant Concentration (g/m3)		CxT Value (gh/m3)
		Immediately after Injection of Fumigant	4 hours after Injection
Hartley	1 2	54.0 54.1	16.2 15.0	101.0 87.4
Payne	1 2	51.2 50.8	18.9 17.5	109.4 99.9
Franquett	1 2	51.3 33.7	12.7 12.9	75.7 71.5

2. 1988 tests on three varieties (May Glo, Mayfire and May Diamond) of American nectarines: CxT values of May Diamond had showed a statistically significant difference with the other two cultivars for most of the doses. ⁹⁵

Table 9 CxT Values of American Nectarine Varieties and Doses (2 hours, 21oC, 50% Load Factor)
Variety	CxT Values (gh/m3�SD) for Fumigant Doses
	48 g/m3	64 g/m3	96 g/m3	128 g/m3
May Glo	64.8�0.8	86.6�1.8	132.7�5.1	173.8�7.3
Mayfire	63.8�1.2	86.4�3.4	135.0�4.9	182.4�10.2
May Diamond	72.4�4.4	98.2�8.4	158.2�7.5	208.2�9.7

3. 1997 tests on three varieties (Fantasia, Flavortop and Shuhou) of Japanese nectarines: There had been a statistically significant difference in CxT values between Shuhou and Fantasia. ⁹⁶

Table 10 Residual Gas Rate of Fumigant, CxT Value and Fumigant Sorption of Fantasia, Shuhou and Flavortop Nectarine Varieties in MB Fumigation (48g/m3, 20oC, 2 hours, 0.143t/m3 Load Factor)
Variety	Replicates	Weight (kg)	No. of Fruits (Unit)	Residual Gas Rate (%)	CxT values		Sorption (mg/kg)
					gh/m3	Mean � SD
Fantasia	1 2 3	4.218 4.213 4.218	32 32 32	68.2 67.1 68.4	71.1 69.9 72.0	71.0 �1.1	133.1 139.8 134.2
Shuhou	1 2	4.217 4.212	30 30	69.7 70.3	66.1 65.9	66.0 �0.2	146.7 144.9
Flavortop	1 2	4.213 4.216	26 26	69.7 68.4	69.8 67.8	68.8 �1.4	136.2 141.9

111. Japan noted that these significant differences were attributable to varieties as the CxT value was affected by physical and chemical properties of the fruits, which were attributable to varietal characteristics. For example, in the case of walnuts, Japan suspected that the differences in the oil content of the walnuts, a qualitative feature of the particular variety, affected the sorption of fumigant and the resulting CxT values. Similarly, the CxT value variation in the 1997 nectarine experiment (third table under paragraph 4.110) was considered attributable to Shuhou�s rougher surface, another feature characteristic of the variety, compared to other varieties.
112. Japan noted that despite test-to-test variation, statistical analysis using Tukey's multiple range test had found a statistically significant difference in the CxT values between Payne and Franquett varieties of walnuts. Statistically speaking, Franquette's CxT value was lower than that of Payne at the confidence level of 95 per cent. Similarly, the same statistical tool had found a significant difference among the nectarine varieties in the second and third experiments set out in paragraph 4.110 encima.
113. Japan claimed that the study on Satsuma mandarins and pumpkins which the United States had cited (paragraph 4.95) was in fact fully consistent with the present Japanese practice. Japan recalled that the hypothesis underlying the Japanese policy was that varietal differences could be such as would affect sorption of fumigant by the fruits, resulting in different CxT values and hence efficacy of MB fumigation treatment. In the cited experiment, faced with the difficulty of rearing and placing mealybugs on Satsuma mandarins⁹⁷, scientists had chosen to use pumpkins as a proxy for Satsuma mandarins because the CxT values between the two plants were remarkably close. It was incorrect to state that "Japanese scientists were willing to accept efficacy of quarantine treatment for a pest as it related to two radically different products". Quite the contrary, these products were almost identical as far as the crucial factor � the CxT value � was concerned. This experiment by no means undermined the hypothesis relating to the varietal differences.

Table 11 CxT Values for Satsuma Mandarins and Pumpkins98
Subject	Replicates	Load (kg)	CxT Value (mgh/l)
Satsuma Mandarins	1 2 3	1.45 1.50 4.10	92 92 98
Pumpkins	1 2 3	1.45 1.50 4.10	94 93 95

114. The United States claimed that minor differences in CxT values between varieties did not indicate differences in varieties of a single product. While CxT values could assist in developing a commercial treatment to kill codling moth, they did not in and of themselves indicate differences in varieties. The difference in CxT could be just as pronounced within the same variety. The experts advising the Panel had not accepted the notion that these differences could be attributed to varietal differences and had furthermore pointed out that, whatever the cause of CxT variation between varieties, these differences had not been observed to affect treatment efficacy.
115. Reasons for CxT value variations included (i) minor differences in leakiness of fumigation chambers; (ii) the amount of toxicant taken up by the product or the packaging material; (iii) the amount of product being fumigated (the load); (iv) the accuracy of the various measurements used to calculate the CxT; and, (v) the times at which the concentrations in the chambers were sampled. The United States stressed that CxT values were used by researchers only as a control mechanism for fumigation trials. They were not useful for concluding that there were differences in varieties. In addition, the United States pointed out that the experts advising the Panel had confirmed that test-to-test variation was inevitable in small-scale tests because of natural variations in pest populations, testing equipment and conditions, and product samples. Also, variations within a product were likewise inevitable from season to season, tree to tree and fruit-to-fruit within a variety. The experts had, according to the United States, in this regard noted that fruit-to-fruit variation could greatly exceed variety-to-variety variation, if any. Variation could also occur because of the difficulty of having different varieties in the same physiological state because of ripening times.
116. Importantly, the United States stressed that the experts advising the Panel had made very clear that it was not possible to conclude that that test-to-test variation in CxT values were attributable to varietal differences rather than any of the other sources of variation described.
117. In the 1992 New Zealand study on cherries, the CxT values for the Sam variety of cherry varied from one season to the next (paragraph 4.106). Moreover, from the Summer Grand and May Grand varieties of nectarines tested in the Vail et al., 1997 nectarine re-test (paragraph 4.93), it was possible to see that CxT values differed within the same varieties compared to the 1987 nectarine study results. In large-scale US tests of nectarines, it was again evident that there could be as much difference in CxT values within a single variety as among varieties. The difference in CxT value ranges within May Grand nectarines, in 1991, was comparable to the difference between May Grand and Royal Giant nectarines in 1989 and 1991 ⁹⁹ (Table 12 below). The United States noted that in respect of walnuts, the Hartley variety referred to by Japan exhibited a difference in CxT values from one test to the next; this difference was greater than the difference in CxT values between the varieties Hartley and Payne. ¹⁰⁰ Hence, in the view of the United States, these results illustrated that there could be any number of reasons why CxT values varied and their variations were not indicative of any varietal characteristics.

Table 12 CxT Values From Large-scale Tests on U.S. Nectarine Varieties in Carton Containers
	1989	1991	1992
Royal Giant	69.6 72.1		67.8
May Grand		70.8 79.4
Fantasia		73

118. In respect of walnuts, the United States noted that Japan had pointed to the results of the 1985 Walnut Research Report ¹⁰¹ to assert that a difference in CxT values in dose-mortality tests applied to varieties of fruits supported the need for varietal testing. Again, the United States claimed that test-to-test variation of one variety had been as great as that found among all three tested varieties of walnuts. ¹⁰² CxT values were an indication of how much toxicant was available to the target insect. Regardless of the observed differences in the CxT values, there was no difference in the efficacy of the quarantine treatment for the three varieties of walnuts. Japan approved the methods and data generated during these studies and had not requested different quarantine schedules for the three varieties. ¹⁰³ In this respect, the United States noted that in bilateral talks in March 1997, the parties had discussed the Japanese decision not to allow import of the walnut variety Eureka due to perceived differences in sorption of methyl bromide. In fact, Japan had refused importation because, although its de-sorption rate had not been significantly different, Eureka had consistently higher residues, likely attributable to higher natural oil content (that would retain the methyl bromide), as compared with other varieties of walnuts. ¹⁰⁴ Concerns over methyl bromide residues were exclusively related to food safety and had nothing to do with the efficacy of the quarantine treatment. The United States insisted that perceived differences in de-sorption rates (rate at which fumigant leaves a product) did not indicate varietal differences with respect to quarantine protection. Although Eureka had consistently higher residues than the other three varieties tested, these differences had not been statistically significant and had not affected treatment efficacy. In fact, preliminary tests showed that there was no survival of test larvae in any of the varieties treated. The authors of the 1991 report on the 1985 tests on walnuts, had stated that "no significant differences were found in mortality of larvae among the four walnut cultivars tested, nor was variation in the size of walnuts of each cultivar a significant factor". ¹⁰⁵ At the second substantive meeting of the Panel, the United States submitted an article which clarified that oil content in walnuts did not vary by variety. ¹⁰⁶ Hence, even if sorption was affected by oil content, there was no basis for concluding that this varied by variety.
119. The United States pointed at the Japanese admission that for apples, cherries, nectarines and walnuts, there was no disagreement as to the efficacy of the treatment as applied to the approved varieties (paragraph 4.6). The United States claimed that implicit in this acknowledgement was a recognition that there would be variations in the preliminary dose-mortality testing and these differences did not alter the final quarantine treatment. These variations were evident in the test results of these approved varieties. Japan�s recognition acknowledged that CxT values could vary, but that they were not relevant in reaching efficacy (in killing the required amount of codling moth). To suggest that the variations of data in dose-mortality testing on its own represented differences in varieties was inconsistent with this concession on the part of Japan.
120. Moreover, the United States claimed that CxT was not an indicator of likeness among varieties or products. In reference to Satsuma mandarins and pumpkins Japan had asserted that it was within reason to accept treatment for a pest on two different products so long as the two products had identical CxT values (paragraph 4.113). This ignored all other physical aspects of the product such as size, as well as what biological stage of the pest would interact with a product. The United States recalled that the CxT value measured the amount of fumigant outside of a product in a fumigant chamber. Yet codling moth was not only on the outside of nectarines. The pest actually burrowed into apples, cherries and walnuts. CxT was silent on the way the fumigant interacted with the pest inside the product. If CxT were the standard of efficacy, then treatments would be developed that were ineffective in addressing the stage of pest, where the pest was located on the product, and the obvious differences among products that necessitate different treatments. Irrespective of what accounted for differences in CxT values, it had to be remembered that these values were merely tools in estimating and establishing a treatment level that resulted in mortality. Treatment efficacy did not rely on CxT values. It relied on the ability to kill the required level of pests.
121. Finally, the United States claimed that Japan had, in attempting to justify its theory, mischaracterized the meaning of the article by E.J. Bond that it cited (footnote to paragraph 4.109). While Japan seemed to be suggesting that Bond linked CxT values to variety of the product and differences among those varieties, he was in fact discussing CxT values as they related to the pest, and specifically the role of CxT values as a tool to observe the interaction of the fumigant and the pest. The Bond article did not suggest that differences in CxT indicated differences in variety. ¹⁰⁷
122. In addition, the United States noted that Japan had stated that the method it used to test the statistical significance of walnut data was "Tukey's multiple range test". However, tests such as Tukey's were based on an analysis of variance. In order to be used correctly, several assumptions had to be met, among them the requirement that samples used in replicates had to be randomly and independently selected and the requirement that the variances of the different samples were homogeneous. Tukey's test was not relevant to an analysis of the walnut data Japan examined ¹⁰⁸ because this data failed to meet both assumptions. For each of the varieties listed in Table 8 (page *), the walnuts used in each "replicate" were taken from the same batch of walnuts. The samples were thus not independently selected, and it was inaccurate to describe the test on each variety as "replicates". Second, the variances of the data for each variety were not homogeneous. ¹⁰⁹ Finally, the United States pointed to the observation of the experts advising the panel that even if Japan had demonstrated "statistically significant" differences in CxT, such differences would not necessarily have been a reflection of significant biological differences which could affect efficacy.
123. In sum, CxT values did not indicate differences among varieties. Those differences were just as pronounced within varieties. The United States claimed that Japan implicitly had to have understood this conclusion since it did not require different quarantine treatments or testing for the same variety despite differences in CxT values within varieties.

To continue with CxT Values

⁹¹ Japan noted that in fact, the mode of action of MB treatment on insects was not fully known, and there could be other mechanisms by which varietal differences could affect the efficacy and referred to by Bell et al., (Price and Chakrabarti, "The Methyl Bromide Issue", 1996. (Japan, Exhibit 11))

⁹² Japan noted that E. J. Bond had stated that "[t]he use of integrated c x t products is particularly useful in routine fumigations when the reaction of a particular species or groups of species has been carefully worked out under the range of conditions likely to be encountered. It has been used successfully in large-scale eradication campaign" (Bond, E.J., "Manual of Fumigation for Insect Control", 1984. (Japan, Exhibit 12)).

⁹³ Japan noted that in addition, in a test on New Zealand cherries, it had been found that "[g]as chromatograph readings during the fumigations suggested that �Bing� cherries absorbed less methyl bromide than 'Dawson' cherries", Wearing, Batcholor, Maindonald, "Disinfestation of New Zealand Cherries", Department of Scientific and Industrial Research, 1981. (Japan, Exhibit 13).

⁹⁴ Vail, Hartsell and Tebbets, "Walnut On-Site Operational (Demonstration) Test Report to Japanese Ministry of Agriculture, Forestry and Fisheries", USDA/ARS, 1985. (Japan, Exhibit 14)

⁹⁵ Vail, "Efficacy of Methyl Bromide for Codling Moth on Nectarines � Consideration of Nectarines as a Product Group", USDA, 1988. (Japan, Exhibit 15)

⁹⁶ Research Division, Yokoyama Plant Protection Station, "Methyl Bromide Sorption in Nectarine Varieties", unpublished, 1997. (Japan, Exhibit 16)

⁹⁷ Japan noted that it was difficult to rear mealybugs on Satsuma mandarins because, under conditions suitable to the insects, the fruits easily decayed and/or the peel hardened. If one tried to place mealybugs on Satsuma mandarins after rearing them on other fruits, on the other hand, their natural mortality would be very high because of possible damage the insects would suffer upon such placement.

⁹⁸ Supra note , pp.57-68. (US Exhibit 13).

⁹⁹ Yokoyama, Miller, and Hartsell, "Methyl bromide efficacy and residues in large-scale quarantine tests to control Codling moth (Lepidoptera: Tortricidae) on nectarines in field bins and shipping containers for export to Japan". 87 J. Econ. Entomol. 730-735, 1994. (US Exhibit 36)

¹⁰⁰ US Exhibit 31.

¹⁰¹ Vail, Nelson, Hartsell and Tebbetts. "Development of a Quarantine Treatment for the Codling Moth in Walnuts for Export to Japan. (Not published in a scientific journal). Walnut Research Report 1985. Walnut Marketing Board, Sacramento, CA., pp.149-154. (US Exhibit 17)

¹⁰² Ibid., p.150. (US Exhibit 17)

¹⁰³ The United States recalled that the uniform treatment for walnuts was 56g/m³ methyl bromide fumigation for 4 hours at 15.6� C at 100mm HG, and a load factor of less than 50 per cent.

¹⁰⁴ Hartsell, Tebbets, and Vail, "Methyl Bromide Residues and Desorption Rates from Unshelled Walnuts Fumigated with a Quarantine Treatment for Codling Moth (Lepidoptera: Tortricidae)" 84 J. Econ. Entomol. pp. 1294-1297, 1991. (US Exhibit 18)

¹⁰⁵ Hartsell, Vail, Tebbets and Nelson. "Methyl Bromide Quarantine Treatment for Codling Moth (Lepidoptera: Tortricidae) in Unshelled Walnuts", 84 J. Econ. Entomol. pp. 1289-1293, 1991. (US Exhibit 19)

¹⁰⁶ Greve, McGranahan, Hasey, Soder, Kelly, Glodhammer and Labavitch, "Variation in Polyunsaturated Fatty Acids Composition of Persian Walnut", Department of Pomology, University of California, USA , J. Amer. Soc. Hort. Sci 117(3):518-522, 1992. (US Exbibit 40)

¹⁰⁷ Bond 1994, p.25. (Japan, Exhibit 12)

¹⁰⁸ In paragraph 81 of Japan's first submission.

¹⁰⁹ US Exhibit 39.