National Academies Press: OpenBook

Regulating Pesticides (1980)

Chapter: 7. Application to Chlorobenzilate

« Previous: 6. Evaluation of the Regulatory Options: Weighing the Risks and Benefits
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 154
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 155
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 156
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 157
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 158
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 159
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 160
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 161
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 162
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 163
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 164
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 165
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 166
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 167
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 168
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 169
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 170
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 171
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 172
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 173
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 174
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 175
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 176
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 177
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 178
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 179
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 180
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 181
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 182
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 183
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 184
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 185
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 186
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 187
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 188
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 189
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 190
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 191
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 192
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 193
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 194
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 195
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 196
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 197
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 198
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 199
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 200
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 201
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 202
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 203
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 204
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 205
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 206
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 207
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 208
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 209
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 210
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 211
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 212
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 213
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 214
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 215
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 216
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 217
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 218
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 219
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 220
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 221
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 222
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 223
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 224
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 225
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 226
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 227
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 228
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 229
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 230
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 231
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 232
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 233
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 234
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 235
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 236
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 237
Suggested Citation:"7. Application to Chlorobenzilate." National Research Council. 1980. Regulating Pesticides. Washington, DC: The National Academies Press. doi: 10.17226/54.
×
Page 238

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

INTRODUCTION 7 Application to ChIorobenzilate The preceding chapters have reviewed the methods currently used by the oPP in selecting pesticides for review and in analyzing the benefits and risks of alternative regulatory measures where they appear appropriate. A number of important recommendations for revising these procedures have also been made. But it is easier to recommend than to perform. Therefore, the Committee has felt responsible for applying its recom- mendations to an actual instance, to the extent that its resources permitted. This chapter reports on that application. It will also serve to help clarify the Con~mittee's recommendations by illustrating how they are implemented. Chlorobenzilate was selected as the pesticide for the test application. It was the first pesticide to complete the entire RPAR procedure and, consequently, all the data used in EPA'S evaluation were readily available for the Committee's use and appraisal. Because oPP has previously completed benefit and risk assessments for chlorobenzilate, including a comparison of benefits and risks associated with various regulatory options (U.S. EPA 1978a, 1979), the discussion in this chapter is to some extent framed in terms of a critique of the oPP analysis. The structure of the chapter basically follows the format of oPP's Chlorobenzilate: Position Document 3 (U.S. EPA 1978~. The first section reviews chemical and physical properties, registered uses, and environmental fate of the compound The second and third sections present the Committee's risk 154

Application to Chlorobenzilate 155 and benefit assessments, respectively. Finally, in the last section, benefits and risks are compared and presented in a manner that reveals not only the Committee's assessment of the trade-o~s for the major regulatory options considered by oPP, but also the uncertainty in the scientific base and the extent to which value judgments enter into the decision. This chapter focuses on recommended departures from oPP's analytical methodology; it is not intended to stand as an independent document on chlorobenzilate. BACKGROUND Chlorobenzilate (ethyl 4,4'-dichlorobenzilate), a chlorinated hydrocar- bon acaricide, is manufactured by esterification of dichlorobenzilic acid and is formulated principally as emulsifiable concentrates and wettable powders (Severe 1978~. The formulation marketed in the United States contains 45.5 percent technical-grade chlorobenzilate. Approximately 93 percent of this amount is pure chlorobenzilate; the remaining 7 percent consists of several unidentified intermediates and other impurities (U.S. EPA 1978b). Chlorobenzilate is registered for use on almonds, walnuts, apples, melons, cherries, citrus fruit, cotton, pears, ornamentals, and trees (U.S. EPA 1978a). Approximately 90 percent of the total amount used in the United States is applied to citrus to control the citrus rust mite (U.S. EPA 1978a); the principal crops on which chlorobenzilate is applied are oranges, grapefruits, and lemons (Luttner 1977a). Limited use also occurs on limes, tangerines, and tangelos (Luster 1977a). The predomi- nant method of applying chlorobenzilate to citrus groves is with a speed sprayer pulled by a tractor. The illustrative analysis in this chapter concentrates on the principal uses of chlorobenzilate, namely, mite control on oranges, grapefruits, and lemons. PROPERTIES OF CHLOROBENZILATE, ETHION, Ad DICOFOL As noted earlier, to assess the risks and benefits of adopting any regulation restricting the use of a pesticide, it is necessary to compare the risks and benefits of using that pesticide with comparable risks and benefits of alternative pesticides to which users are likely to resort. Ethion and dicofol are two important alternatives to chlorobenzilate for control of rust mites on citrus fruits (U.S. EPA 1978a). Their physical properties are compared with those of chlorobenzilate in Table 7.1. Examination of these properties and calculation of other molecular parameters indicates some common behavioral characteristics in the

156 REGULATING PESTICIDES environment. Each has a relatively low vapor pressure, which indicates that the rate of vapor loss from sprayed surfaces may not be high. The substantial polarity of ethion indicated by its molar refraction (100.1) probably indicates a strong adsorption on surfaces. Ethion may therefore persist as a surface residue allowing dermal exposure if reentry occurs before photochemical destruction or hydrolysis. Dicofol similarly shows low vapor loss, indicating ready adsorption. As with ethion, dicofol's physical and biological properties may afford exposure. Chlorobenzilate, while also having comparatively low vapor loss, is much more suscepti- ble to degradative chemical, biochemical, and photochemical reactions than dicofol. It should undergo metabolism more readily than dicofol and probably have less of a propensity for partitioning in lipid. Among the three compounds, ethion would probably be the least persistent and dicofol the most. In terms of biological activity, ethion is not as specific to rust mites as is chlorobenzilate, so that it is likely to have more extensive side ejects than chlorobenzilate on nontarget organisms. Dicofol is specific for mites, but is not as effective for rust mites as is chlorobenzilate. This cursory examination of physical and biological properties suggests that chlorobenzilate poses less of an environmental and human hazard in terms of persistence and the possibility of undesired exposure than its substitutes, dicofol or ethion. There are few actual data about the fate of chlorobenzilate in the environment. Several authors have studied its metabolism in plants and found that it is persistent in citrus and apple peels (Gunther et al. 1977, Severn 19781. When applied topically to soybean leaves, it translocates to the petioles unchanged after about 12 days (Hassan and Knowles 1969~. Miyazaki et al. (1970) found that chlorobenzilate can be metabolized by microorganisms, particularly yeast. A study of chlorobenzilate's persis- tence in Florida Lakeland and Leon fine sandy soils demonstrated a half-life of 1.5-5 weeks in Leon soil and 1.5-3 weeks in Lakeland soil (Wheeler et al. 1973~. This same study concluded that chlorobenzilate did not affect the microbiological activity in the soils. Finally, a study of environmental transport detected no chlorobenzilate in drainage water or in soil samples, following (1) the spraying of a citrus grove, (2) 39 hours of irrigation, and (3) a 2.41 cm rainfall a week later (U.S. EPA 1977). ANALYSIS AND ASSESSMENT OF THE RISKS The principal concern with the use of chlorobenzilate is the possibility that this chemical may increase the incidence of cancers in people exposed to it. The seriousness of this risk depends on three factors: (1)

Application to Chlorobenzilate TABLE 7. ~ Physical Properties of Chlorobenzilate, Dicofol, and Ethion 157 Solubility Vapor Molecular In Water Pressurea Refractive Compound Weighta (temperature ° C) (mm Hg) Index ,a nD20 Chloroben- 325.2 - 2.2 x 10-6 (20) 1.5727 zilate (tech. prod.) Dicofol 370.5 1320,ug/1 (25)b 1.20 ppm (20)C 1.606 ppm (3)C Ethion 384 2ppm (22)4 1.5 x 10-6 (25) 1.5490 2 ppm (?)e 1.530 to 1.542 (tech. grade) a Source: Martin (1971) . b Source: Well et al. (1974). c Source: F. Parveen, Environmental Health Sciences Center, Oregon State University, Cor vallis, personal communication, 1975. Source: von Rumker and Horay (1972). e Source: Gunther et al. (1968). the number of people exposed, (2) the dosages to which each of them is exposed, and (3) the probable health risks from receiving these dosages. The next subsections present estimates of the extent of human exposure, followed by an assessment of that the consequences of the exposure. In a final subsection, risks posed by chlorobenzilate substitutes are evaluated and comparison compounds selected. HUMAN EXPOSURE TO CHLOROBENZILATE The use of chlorobenzilate on citrus fruit exposes different segments of the U.S. population to widely differing doses. The largest doses are received by citrus spray applicators and citrus fruit pickers. Much lower doses are received by people who eat foods that contain residues of the pesticide. Some of the meat consumed by residents of Florida ~y contain residues transmitted in the citrus pulp that is used as animal feed in the Florida livestock industry (U.S. EPA 1978a). In addition, the U.S population in general, including the residents of Florida, receives small quantitites through the ingestion of citrus fruits, some minor use crops, and products made from them. The daily doses of chlorobenzilate to these three population groups-workers, the Florida population, and the general U.S. population are presented in Table 7.2; the derivation of these estimates and estimates of total exposure are discussed below.

158 I .o ·=U, o - D o o U. U. o - ·~ ~ D AX ·~ ~ - D D o I:: D · - X · Ct D 2 .° o o ) ·° ~ ~To {o 4= Lad ~ o ~ ^ A to - o to to to - to to o o o o o o rat o o - .O CL ~ o o ~ ^ ~ ~ x - ~ - o - ·o ~ cr~ r~ . . . o ~ ~ ~ oo ~ - - - . . . o o - Io o o o r~ - o - :^ - ~ _ o o ~ o o, o oo 1 o o o . ._ o o ~ . . ~ X _ _ ~ ~ ~ - ~ ~ a

159 o ~ ~ o _ ~ ~o~ ~o o o _ ~_ oo o - . _ ooo oo o o ~ o ~ ~o o o ~ ~o oo o ~ _ [_oo oo o o o ooo oo o _ _ o_o o . . ... . o o oo~ ~ C~ _ _ ~o :^ .a . ;~ _ _ r ~_ _ _ S o o o o V) ~o o o o o o o o o o C, C o ~ o . . . cr oo _ ~ 888 8 _ _ '_ _ _ o oo oooo ~X .. . . . . . o o _ o o ~ o o ~ - _ _ . . . . . o ~o o _ ~ ~ .~ ~ ~ ~ o ~ ~ ~ ~ ~ 3 ._ ~ _ · _ O~D cn O [L, O C~ CL X X _ ~ C°L ~ ~ UO, ~ ~ cn O 4., ~ U~ o _ o o tV V~ · a) D ~ ~=.O V, O 4-= ·_ C ,C ~O C O O u pC ~ D ~ ~ C) ~ ~a.' ~ ~ ~_ =0 ~ ~ Ct ~ ~ O-m S-~m V, ~ O g . - ~ · - ~ . u, o ~ ·m ~ ~ ~ . - D O U, S:: g O = U,~c-c', ~ ,C :m c ~ ;> ~ ~ ~ ·~ ~ C X C ~ a X c ~ ~ ~ ~> 0 CtS C .0 C cI C) E ~ E ~ E ' ~ ~ ~ - o ,, E ~ ~ ~ o

160 Occupational Exposures REGULATING PESTICIDES Citrus Spray Applicators The Committee basically accepts oPP's esti- mates of the worst-case exposure situation for spray applicators. Although these estimates are methodologically sound, they are only conjecture because there are no direct observations on the dosages received by chlorobenzilate applicators. The dosage received, therefore, had to be inferred from the experience of applicators of other pesticides believed to be analogous. Under the circumstances, the Committee sees no way to improve on the estimates. The estimated probable-case doses are similarly conjectural. A major modification of oPP's estimates, however, has to do with the number of years-a worker is expected to be exposed to chlorobenzilate. The Committee will use a value of 10 years as the additional expected economic life of chlorobenzilate for use on citrus, whereas oPP assumed that chlorobenzilate would continue to be used indefinitely (see Chapter 4~. If chlorobenzilate were only viable in the marketplace for an additional 10 years, the incremental exposure to citrus workers would occur only for those additional 10 years, not a full occupational lifetime. In a study of workers exposed to various pesticides, Wolfe et al. (1967) measured both dermal and respiratory (i.e., inhalation) exposure under a variety of ground spray application conditions (see Table 7.3~. All applications were to fruit orchards, using air-blast spray equipment. The technique for trapping residues of pesticides during application involved attaching absorbent pads to the body or clothing of the applicators to measure dermal exposure, and placing filter pads in respirators worn by the applicators to measure respiratory exposure. Trapped residues were extracted and chemical analysis of the various pesticides was carried out using a variety of analytical techniques. Measurements of residues derived from chemical analyses indicated that exposure from spray operations was greater for the dermal route than the inhalation route (see Table 7.3~. Data on dermal exposures were gathered under conditions in which applicators wore short-sleeved, open-necked shirts, with no hats or gloves (Wolfe et al. 19679. The investigators assumed that covered portions of the workers' bodies were completely protected. Since the data in Table 7.3 were obtained under conditions similar to those associated with the application of chloroben- zilate (i.e., similar type of crop, spray apparatus, and spray concentra- tion), oPP assumed that they provide a reasonable basis for estimating exposure of spray applicators to chlorobenzilate (Severe 19781. OPP'S estimates of the exposure of spray applicators are based on the

Application to Chlorobenzilate ·61 range of the mean values reported by Wolfe et al. (1967) for dermal exposures (rounded on to 15-50 mg/hour) and the maximum mean value reported for inhalation exposures (rounded off to 0.1 mg/hour) (see Table 7.3~. Assuming an 8-hour workday, oPP estimated the daily dermal dose to range from 120 to 400 mg and the daily inhalation dose to be approximately 1 mg (Severe 1978~. Data upon which to base an estimate of the rate of dermal absorption of chlorobenzilate were not available either (Severe 1978~. oPP therefore assumed that the chemical characteristics of chlorobenzilate are similar to those of DDT and other chlorinated hydrocarbon pesticides, and that chlorobenzilate would penetrate the skin at a rate comparable to that of DDT and other similar pesticides (Severe 1978~. Estimation of the amount of chlorobenzilate that penetrates the skin was based largely on a study by Feldmann and Maibach (1974), in which the authors studied the recovery from urine of radioactively labeled DDT, lindane, parathion, and malathion, following topical administration to the forearm in humans. They found that 5 percent of the applied dose was absorbed after the first day and that about 10.4 percent was absorbed after 5 days, although the subjects were allowed to wash the application site after the first day. The authors assumed that excretion and tissue distribution of the test compounds after dermal absorption are the same as after intravenous injection. Thus, oPP concluded that about 10 percent of the amount of chlorobenzilate that reaches the skin is absorbed (Severe 1978~. OPP estimated the daily dose of chlorobenzilate received by spray applicators based on the above data and assumptions. For dermal exposures without protective clothing or respirators, oPP considered 12 400 mg/day to be a reasonable estimate. Multiplying this range by a 10 percent absorption factor produced an estimate of 12-40 mg/day. oPP's daily inhalation exposure estimates assume 100 percent absorption by the lungs (Feldmann and Maibach 1974) and are estimated at 1 mg/day. Since protective clothing was not required, and climatic conditions where citrus is grown dictate against its use, oPP assumed, in the absence of other information, that spray applicators did not wear protective clothing (Severe 1978~. Thus, oPP estimated that daily occupational exposure per individual for spray applicators was 12 40 mg dermally and 1 mg by inhalation, or 13-41 mg total. Because the Committee has no new data with which to make better estimates, oPP's range is retained; the lower value is assumed to be a minimum-plausible and the higher value a maximum-plausible exposure estimate. To derive a probable-case estimate, as recommended in Chapter 4, the Committee has taken the

162 o~ .~ ~ o ~_ ·g c Ct C~ o o . ~ U. Ct a, os .o .e _ ._ ° ao O X ~: Ct ~ o oo ~ ~ ~ oo _ _ _ _ _ ~ _ - _ o _ o _ o . . . . . o o o o o oo ~ ~ ~ ~ o ~ o ~ o . . . . . Ol O Ol Ol Ol ~ C~ ~ C~ _ o o o o o . . . . . o o o o o oo _ _ ~ ;: ~ o ~ _ . _ _ o oo V) _ - , _ - ~ ~ _ ~ oo _ 1 1 ~ 1 1 _ ~ ~ ~ ~ . . . . . _ ~ ~ ~ _ V) 1 1 1 o oo ~ ~ ~ ~ oo o o . . ol ol o o o o o o o o o o a, o ~ . - o o ;! a a ~ ~ ~L ._ a~ .` Ct S L U' - .s ~ U' ._ U] .~ a' ~ ~ ~ O o ~ 3 ~ ~ ~o U, ~ ~ .= ~ .c _ ~ ~ o ~ .~- a =s - ~ ~Z O ~0 cr -

Application to Chlorobenzilate 163 mid-point between oPP's minimum- and maximum-plausible values (assuming that the dose-response curve is linear in this range) arriving at a probable exposure estimate for applicators of 27 mg/day (see Table 7.2~. Conversion of daily occupational exposures to total incremental exposure, were chlorobenzilate to continue in use, must take into account the duration of exposure. The USDA (1977c) estimates that the current use of chlorobenzilate in ground application to citrus is carried out by as few as 714 applicators for 30~0 days/year, or by as many as 1,375 applicators for 1~20 days/year (Severe 1978~. Again, oPP chose the worst-case exposure situation for an individual and assumed that the lesser number of applicators, 714, work for the greater number of days a year, 40, for approximately 40 years (U.S. EPA 1978a). Here the Con~n~ittee's estimates depart from oPP's by assuming that 10 years after a regulatory decision, chlorobenzilate will be gone from the marketplace. Instead of a 40-year exposure, then, the following calcula- tions assume 10 additional years of ground applicator exposure at 40 days/year to approximately 700 applicators. (These values 10 years, 40 daYs/vear. 700 applicators-are being treated as firm estimates for the , , , ~ ~ ~ . .. . .. .. ... . . ... .. . ... . sake of analytic simplicity, although In reality they should be presented as ranges.) Thus, the total incremental lifetime exposure of spray applicators to chlorobenzilate under assumed present conditions, i.e., no protective clothing, becomes: Probable case = 27 mg/dlay x 40 days/year x 10 years = 10,800 lag. Maximum-plausible case = 41 mg/day x 40 days/year x 10 years = 16,400 ma. OPP also considered occupational exposure of drivers of auxiliary vehicles and helicopter pilots. We adopt oPP's conclusions that (1) since the drivers only bring their trucks to the edges of the groves and are not generally in the immediate vicinity of the sprayer, the drivers' exposure is very much less than that of the applicators (Severe 1978), and (2) spraying by helicopters (in Florida) is sporadic and is unlikely to result in significant exposure to humans because of the small amount that is applied and because the pilot is protected by the enclosed helicopter cockpit (Luttner 1977b). The final regulatory action taken by oPP in concluding the chloroben

164 REGULATING PESTICIDES zilate RPAR stipulates that ground applicators have to use either protective clothing and a respirator, or a suitably equipped enclosed cab. oPP derived exposure estimates for spray applicators using protective clothing and respirators. oPP,s estimate of the reduction in dermal exposure afforded by protective clothing (coveralls, a cloth cap, and gloves) was based on the assumption that covered skin areas are completely protected. According to Hayes (1975), the body surface areas of hands, arms, face, and neck make up approximately 16 percent of total body surface area. If all these areas except the face are covered, the remaining exposed surface would be 3.5 percent of total body surface, resulting in a reduction of dermal exposure by a factor of approximately 4.5 (Severe 1978~. Using this factor of 4.5, oPP derived a dermal exposure of 3-9 mg/day (12~0 mg/day . 4.5) with protective clothing. oPP also concluded that respirators would electively eliminate exposure by inhalation (estimated at 1 mg/day) and further reduce dermal exposure to the face by 1-3 mg/day. Thus, for applicators wearing both protective clothing and respirators, daily exposure could be reduced to between 2 and 6 mg/day, minimum-plausible and maximum-plausible exposure, respectively (U.S. EPA 1979~. Total incremental lifetime exposure of an applicator in full compliance with the new chlorobenzilate regulations-that is, with protective clothing and respirators would be: Probable case = 4 mg/day x 40 days/year x 10 years = 1,600 ma. Maximum-plausible case = 6 mg/day x 40 days/year x 10 years = 2,400 ma. However, these estimates of the ejects of the regulation cannot be confirmed until the required applicator exposure data are submitted and evaluated. In fact, when calculating risk to citrus spray applicators associated with the regulatory option that requires protective clothing and respirators, the Committee assumes only 50 percent compliance as the probable case and 20 percent in the maximum-plausible exposure case. These assumptions are based on the difficulty of enforcement and the Committee's direct experience with citrus growers and researchers in Texas, which indicated a disbelief that chlorobenzilate is hazardous. One of the major uncertainties in the forgoing analysis is the validity of the dermal absorption rate (i.e., 10 percent). Chlorobenzilate is relatively polar compared to other compounds of its general type, such

Application to Chlorobenzilate 165 as DDT, SO that its relative dermal uptake should be less than that for DDT. Reliance on the general data developed by Wolfe et al. (1967) for estimating inhalation and dermal exposures is unfortunate, partly because so few compounds were tested but, more importantly, because the absorbent pads used in their tests (cotton gauze pads) lack all the significant characteristics of human skin. The result of these consider- ations is that errors of several orders of magnitude might be involved in these estimates and passed on to the estimated inhalation exposure calculated by the average dermal to inhalation ratios reported by Wolfe et al. Research on these topics would allow for better estimates through development of appropriate physicochemical correlates to relative dermal absorption of those pesticides already investigated. However, the data upon which the forgoing analysis relies for its estimates of dermal absorption of chlorobenzilate (i.e., Feldmann and Maibach 1974) are not helpful, as the correlation between the data and actual physicochemical activities appears small or non-existent. Consequently, use of an average dermal absorption rate for chlorobenzilate is necessary and unavoidable, though weak. Fruit Pickers The Florida citrus crop is harvested from about Novem- ber to May, using mostly migrant workers (Severe 1978~. Data cited by the Federal Working Group on Pest Management (1974) show that in January 1971, 25,431 migrant and contract workers were employed in Florida, whereas in July of that year only about 450 were employed. During harvesting season, as many as 25,000 30,000 citrus pickers may be occupationally exposed to chlorobenzilate (U.S. EPA 1978a). OPP'S analysis of the risks associated with the use of chlorobenzilate, however, does not evaluate the extent of exposure of citrus pickers during harvesting activities. The Committee agrees that lack of data on (1) the extent of dislodgeable residues on citrus fruit and foliage, and (2) the- extent of transfer of these residues to citrus pickers, either by dermal or vaporization routes, precludes quantitative assessment of exposure for this particular group (Severe 1978~. Since exposure of citrus pickers occurs after chlorobenzilate has been applied, oPP assumed that their exposure is less than that of spray applicators (U.S. EPA 1978a). However, chlorobenzilate residues on citrus fruit and foliage may be a primary source of contact (i.e., dermal) exposure. Residue data, together with factors indicating the degree of absorption by the skin and duration of exposure, would allow for estimation of exposure of citrus fruit pickers. The regulatory action taken

166 REGULATING PESTICIDES for chlorobenzilate calls for monitoring studies of the levels of chloro- benzilate residues at harvest time and of the actual dermal doses that the workers receive (U.S. EPA 1978a). Dietary Exposures As noted in Chapter 4, the Committee endorses the general procedure used by oPP to estimate dietary exposures. The general equation oPP uses to determine daily doses is: consumption (g/day) X extent of pesticide use on crop (percent) x maximum residue (ppm) = maximum ingestion (,ug/day). The Committee will, however, derive probable levels of exposure in addition to the worst-case (maximum-plausible) situation to which oPP's estimates are limited. Also, the Committee dissents from some of the detailed procedures used by oPP: in particular, those for determining assumed residue levels and the assumption that populations will receive chlorobenzilate through their diet for an entire 70-year lifetime. The estimates used by the Committee are derived in the following two subsections. Florida Population The population of Florida is assumed to undergo unique exposure to chlorobenzilate due to the use of treated citrus pulp in livestock feed. Chlorobenzilate residues are assumed to be ingested by Floridians via the meat and meat by-products of Florida beef and lamb, and via milk from the cattle. The procedure recommended and used by the Committee to estimate probable exposure involves using an observed residue accumulation ratio to estimate chlorobenzilate residue levels in beef, lamb, and milk, instead of analytical sensitivity levels. The advantage of this procedure is that it is pharmacodynamically sound, whereas the maximum-plausible residue estimates are not. The critical assumption in this procedure is that the accumulation ratio is constant over a wide range of dosage rates. An accumulation ratio for chlorobenzilate in cattle fat can be derived using data from a study by Mattson and Insler (~9664. Mattson and Insler studied levels of chlorobenzilate residues in cattle and sheep that had been fed a daily ration containing chlorobenzilate over a period of 28 days. Following analysis of several different tissues, residues of chlorobenzilate were found only in cattle fat at the highest feeding level, 340 mg/animal/day. Assuming that a cow eats approximately 20 kg of feed a day, the feed resulting in ingestion of 340 mg/animal/day would contain chlorobenzilate at a concentration of 17 ppm. The average residue level for three fat samples was 0.63 ppm. The observed

Application to Chlorobenzilate ·67 accumulation ratio, therefore, is ppm in fat . ppm in feed, or 0.63 . 17 = 0.037 _ 0.04. No residues were detected at an analytical limit of 0.3 ppm when the feeding level was 110 mg/animal/day. However, the probable residue in fat calculated by estimating from the accumulation ratio would be 0.22 ppm (a feeding level of 110 mg/day . 20 kg consumed = 5.5 ppm of chlorobenzilate in the diet; 5.5 ppm x 0.04 accumulation ratio in fat = 0.22 ppm residue in fat). Adopting oPP's estimate of 2 ppm as a reasonable upper limit of residue in citrus pulp (based on maximum measured residues in Florida citrus pulp, FY 1976 data; Severn 1978), the Committee estimates that the steady-state residue of chlorobenzilate in animal fat from using treated citrus pulp in livestock feed is the accumulation ratio (0.04) x the concentration of chlorobenzilate in the diet (2.0 ppm in pulp x 16 lb of pulp . 20 kg diet), or O.W x 0.73 ppm = 0.029 ppm. Beef cuts averaging 30 percent fat (a maximum value) would then contain 0.009 ppm residual chlorobenzilate. If Florida beef and lamb were consumed at a rate of 143.2 g/day (Schmitt 1977, based on USDA 1972), the probable estimate of ingestion of chlorobenzilate would be 1.29 ,Ag/day (based on all animals being fed pulp under the above conditions). However, we adopt oPP's estimate that only 10 percent of the animals are fed citrus pulp, reducing the probable ingestion estimate to 0.13 ,ug/day (see Table 7.2~. The method used by oPP, and adopted by the Committee, to derive maximum-plausible estimates of daily chlorobenzilate ingestion from beef and lamb is based on the detection level of chlorobenzilate in fat. This method may overestimate exposure by relying on maximum analytical sensitivity. In oPP's chlorobenzilate estimates, the key assump- tion fits the analytical detection limit to an experimental dosing rate that was observed not to produce residues exceeding that detection limit. Thus, the resulting estimate could be pharmacodynam~cally unsound. However, with chlorobenzilate this approach actually matched the pharmacodynamic estimate quite closely; 110 mg/day ingestion yielded an assumed 0.3 ppm concentration in fat (oPP) rather than 0.22 ppm as estimated in this report. The maximum practical amount of citrus pulp in animal feed is 16 lb/day/animal (Severe 1978~. The amount of chlorobenzilate ingested by an animal fed 16 lb/day of dried citrus pulp is 16 lb/day x 454 g/lb x 2 ppm residue in feed, or 15 mg/day. Using the 0.3 ppm detection level in fat as the maximum fat and meat residue to be expected from feeding 110 mg/animal/day of chlorobenzilate (Mattson and Insler 1966), by extrapolation oPP calculates the maximum residue level expected from

168 REGULATING PESTICIDES feeding 15 mg/day as 0.3 ppm x (15 mg . 110 mg), or 0.04 ppm. Multiplying this assumed residue by daily meat consumption and percent of animals receiving treated feed gives a maximum-plausible daily human ingestion estimate of 0.04 ppm x 143.2 g/day x 10~o = 0.57 ,ug/day (see Table 7.2~. To determine incremental lifetime doses for the Florida population, the daily doses are multiplied by the assumed duration of exposure, discounting past exposures. Using the assumption that chlorobenzilate will remain on the market for an additional 10 years if rereg~stered (see Chapter 4), the additional lifetime doses to Floridians from chlorobenzi- late in beef and lamb are: Probable case (beef, lamb) = 0.13 ,ug/day x 365 days/year x 10 years = 475 ,ug. Max~mum-plausible case (beef, lamb) = 0.57 ,ug/day x 365 days/year x 10 years = 2,081 lug. The accuracy and validity of the accumulation ratio procedure for estimating the daily dose of chlorobenzilate to the Florida population depend on the following assumptions: 1. That neither cumulative dosing, nor dosing at different rates, alters the metabolism of the chemical. Studies have shown the accumulation ratio to be linear over a 100- to 1,000-fold concentration range in several published studies of chemicals similar to chlorobenzilate (e.g., Quaife et al. 1967, WaLker et al. 1969~. 2. That steady-state equilibrium is assumed, thus continuous dosing to the time required for equilibrium is necessary. 3. That external modifiers affecting metabolism, distribution, and elimination of the chemical are constant. Most of these conditions also pertain to the worst-case (maximum- plausible estimate) procedure, although these latter estimates also include the analytical threshold as a major uncertainty. As noted earlier, if the accumulation ratio method were used in the case of chlorobenzi- late, the maximum dosing rate not producing an analytically significant residue (i.e., 110 mg/day in beef cattle) would be predicted to produce a 0.22 ppm residue in fat, rather than oPP's 0.3 ppm (minimum-sensitivity)

Application to Chlorobenzilate 169 value; this close agreement indicates the innocuousness of the analytical sensitivity approach in this instance, but oPP cannot expect such coincidence in general. Residues in milk are concentrated in mink fat and are essentially in equilibrium with body fat at concentrations similar to those in body fat. Therefore, probable estimates of residue concentrations in milk may be derived from the accumulation ratio approach described for beef. With chlorobenzilate residues in pulp at a level of 2 ppm, the resulting residue level in milk fat would be approximately 0.029 ppm (0.73 ppm in feed X the accumulation ratio, 0.04~. This establishes a residue estimate of 0.001 ppm for milk containing 3.5 percent fat (3.5% x 0.029 ppm). If 184.7 g of milk were consumed daily and all of it contained residues (adopting oPP's data, Schmitt 1977 based on USDA 1972), the probable-case estimate of chlorobenzilate ingestion for Floridians would be 0.18 ,ug/day (see Table 7.2~. This value is probably high, since animals that secrete milk eliminate a significant amount of daily fat and thus eliminate residual chlorobenzilate faster than in nonlactating animals; the result is in an overall lowering of the steady-state equilibrium level. oPP's maximum-plausible estimates of chlorobenzilate ingestion by Floridians via milk were derived from limited data based on analytical sensitivity indicating that cattle consumption of treated feed may result in chlorobenzilate residues in milk from 0.0024 to 0.04 ppm (U.S. EPA 1978a). The Committee used the higher value of this residue range, 0.04, to derive a single maximum-plausible exposure estimate of 7.39 ,ug/day (184.7 g/day x O.W ppm; see Table 7.2~. Again, incremental lifetime doses are calculated by multiplying the daily rate by 10 years: Probable case (milk) = 0.18 ,ug/day x 365 days/year x 10 years = 657 ,ug. Maximum-plausible case (milk) = 7.39 ,ug/day x 365 days/year x 10 years = 26,974 ,ug. General U.S. Population The Committee has prepared estimates of probable daily doses received by the general U.S. population through the diet and has adopted oPP's worst-case or maximum-plausible estimates. These estimates are presented in Table 7.2, and their derivation is sketched below. To convert these to lifetime exposures, one would

170 REGULATING PESTICIDES TABLE 7.4 Florida State Monitoring Data for Chlorobenzilate Residues in Fresh Citrus Fruit Percent of Samples Average Range of with Positive Tolerance Residue Residues Season Concentrations (ppm) (ppm) (ppm) 1974-1975 35 5 0.42 0.05-1.82 1975-1976 35 5 0.36 0.07-1.06 1976-1977 38 5 0.040 0.06-1.73 Source: Dennis (1977). multiply by the expected additional 10-year market life of chlorobenzi- late discussed earlier. The Committee's estimates of probable levels of daily chlorobenzilate ingestion from citrus fruits and products are 10 percent of the maximum- plausible estimates, except for flavored drinks (see Table 7.2~. Both the probable and maximum-plausible estimates rely on oPP's average consumption figures, taken from USDA surveys (Schmitt 1977, based on USDA 1972) and oPP's estimate of the percentage of the crop treated with chlorobenzilate (U.S. EPA 1978a). Both estimates also use the analytical sensitivity limit as the assumed residue level (except in the case of flavored drinks where residue data are used); however, the maximum- plausible estimates assume an analytical sensitivity of 0.1 ppm, whereas the probable case uses the sensitivity of new analytical methods, 0.01 ppm. Justification for the use of analytical detection limits comes from a review of actual residue data. Data on chlorobenzilate residues were provided by the Florida Department of Agricultural and Consumer Services (Dennis 1977) and the FDA (Severe 1978), and are presented in Tables 7.4 and 7.5, respectively. Both surveys indicate that residues of chlorobenzilate are well below the established tolerances in the samples tested. Thomas (1976) also carried out a limited monitoring study of chlorobenzilate residues in fresh, canned, and frozen citrus produce obtained in the Washington, D.C. area; residues above 0.1 ppm were found in 7 of 79 samples, all of which were fresh whole citrus. All of the citrus residue data were derived from analyses of whole citrus fruit. It is likely, however, that the residues originated principally from the peels, since detection of significant amounts of chlorobenzilate in the edible portion has not been reported (see, for example, Bartsch et

Application to Chlorobenzilate 171 al. 1971~. Even during juice-making procedures, transfer of chlorobenzi- late from the oil-rich peel to the expressed juice is minimal. According to Kesterson et al. (1971), there exists a misconception that oil is pressed from the peel or fruit during the extraction of citrus oil. They note rather that oil cells are ruptured by pressure or abrasion, and the oil washed away. Here, the partitioning behavior of chlorobenzilate between aqueous and oil phases would prevent even miminal residues in juice. Gunther (1969) presented experimental data showing that chloroben- zilate residues were undetectable in juice under conditions where the detection limit was 1/100 the residue concentration in lemon rind (20 ppm). Even after considering that the laboratory preparation of juice might be "cleaner" than commercial juicing operations, it remains unlikely that chlorobenzilate residues in juice would regularly approach 0.1 ppm concentration when the whole fruit residues established in the monitoring surveys were only 0.1{~.5 ppm, with these being detected in no more than 10 percent of the samples (see Table 7.5~. Whole juice monitoring has not produced detectable chlorobenzilate. Since the majority of citrus products consumed as food comes from the pulp and not the peel, it is assumed that chlorobenzilate residues ingested from consumption of citrus would not be detectable by current monitoring techniques. However, since residues can be present up to the limit of analytical sensitivity and still be undetectable, the complete absence of chlorobenzilate in citrus food products cannot be demon- strated. For maximum-plausible residual estimates, oPP concluded that the Bartsch et al. (1971) survey provided the best available data on chlorobenzilate residues in citrus. These workers found residues in pulp to be less than 0.1 ppm, using gas chromatography with a detection limit of 0.1 ppm. oPP therefore assumed that 0.1 ppm was a reasonable upper limit of actual chlorobenzilate residue levels that would pass undetected in edible portions of citrus (Severe 1978~. For the probable-case estimates, the Committee assumes that since new analytical methods are sensitive to 0.01 ppm, and the residue is virtually~all in the peel, residues in the edible portion of citrus milt be present up to a limit of 0.01 ppm. Estimation of the amount of chlorobenzilate in flavored fruit drinks was based on oPP's residue data for chlorobenzilate in citrus oils, which are used to flavor the drinks (U.S. EPA 19794. oPP's estimate was adopted for both the probable and maximum-plausible residues since that value seemed reasonable to the Committee, and the Committee had no basis for an alternative estimate. Probable and maximum-plausible estimates of chlorobenzilate ingestion

172 U. ._ .= o o Ct :: ._ 3 Ct ._ · V, ao C~ ._ o s~ o ~3 {4o- Ct o ._ ~. Ct U. .~ .= o o . V) m ;> ·3 ._ ~ o . - o Ct O ~ _ c o E <_ ~ ·3 0-- g ~ O ~ ~ . _ O Z 3o ~: .= E ~ o o o _ o. o _ ~ o ~, _ o _ ~ _ _ o o ~ _ o o o o _ _ o - oo _ oo _ _ _ o o o o o C~ C`- - - C` V) ~ o ct cO a, ~ ~. ~- U, V~ V~ { ~ ~ ~ ~ ~ ~ 80 ~ . _ _ ~ _ _ _ _ _ C~ a~ .= U, a' .H o - o - - 5 U, - C~ U, .~ ._ ~: E a~ U, ~ E ~ o ~_ ~ ~: -o ._ E o ~ ~ oo oo o o o ~ o . . . oo ~ ~ O r~ O o o o ~ _ - ~ - . - . ~_ .O .~ ~ _ ~ ~ ~° ~ ~ _~ ~ ~ C ~ ~ o O ~ ~ a~ cr ~ - - - - o ~ oo ~ ~ - o cr' ~ - Q> ~ c~ .. cO o-) o ~ gS cn

Application to Chlorobenzilate 173 via citrus are shown in Table 7.2 and are calculated using oPP's general equation: consumption x extent of use on crop x residue = ingestion. Because of a lack of data on residues in apples and pears, oPP used the 5.0 ppm tolerance value as a basis for estimating maximum-plausible doses. No large fruit sample analyzed, however, contained more than 1.0 ppm residual chlorobenzilate and only 1 percent of the samples analyzed had any detectable residue even though all the fruit had been treated (Table 7.5~. Therefore, not only is the tolerance level an excessive estimate for the concentration in whole fruit, but it also ignores the fact that these fruits are not primarily consumed whole, as assumed, but In several processed forms (e.g., washed, peeled, or cooked). The probable-case estimate, therefore, assumes that only 10 percent of the treated crop contains residues (i.e., residue frequency = 10 percent), and uses the average reported residue (0.50 ppm) to obtain a responsibly conservative on the side of maximizing risk value. For apples and pears, ingestion estimates are (see Table 7.2~: Probable case (apples) = 50.1 g/day x 0.065 (% crop treated) x 0.5 ppm residue x 10% residue frequency = 0.0016 ,ug/day. Maximum-plausible case (apples) = 50.1 g/day x 0.065 (% crop treated) x 5 ppm residue 0.16 ,ug/day. Probable case (pears) = = 5.1 g/day x 0.23 (% crop treated) X 0.5 ppm residue x 10% residue frequency = 0.0006 ,ug/day. Maximum-plausible case (pears) = 5.1 g/day x 0.23 (% crop treated) x 5 ppm residue = 0.06 ,ug/day. Calculations for chlorobenzilate ingestion via nuts are based on analytical detection limits and, for the probable case, a 10 percent residue frequency: Probable case (almonds) = 0.59 g/day x 6.8 (% crop treated) x 0.01 ppm residue x 10% residue frequency = 0.0(XK)4 ,ug/day.

174 Maximum-plausible case (almonds) REGULATING PESTICIDES = 0.59 g/day x 6.8 (% crop treated) x 0.1 ppm residue = 0.004 ,ug/day. Probable case (walnuts) = 0.59 g/day x 0.46 (% crop treated) x 0.01 ppm residue x 10% residue frequency = 0.000003 ,ug/day. Maximum-plausible case (walnuts) = 0.59 ~g/day x 0.46 (% crop treated) x 0.1 ppm residue = 0.0003 ,ug/day. The sum total of daily doses of chlorobenzilate received by the general U.S. population in the diet, shown in Table 7.2, ranges from 1.12 to 4.56 ~g/day, or 4,088 to 16,644 ,ug/lifetime total, assuming chlorobenzilate is present for 10 more years. PATHOLOGICAL ACTIVITY OF CHLOROBENZILATE On May 26, 1976, oPP issued an aPAR for chlorobenzilate based on oncogenic ejects (U.S. EPA 1976b). Although the studies on which the RPAR was based included data indicating that chlorobenzilate has adverse ejects on the testes of rats, oPP's RPAR notice neither mentioned the issue of testicular atrophy nor invited registrants to comment on this issue in their rebuttals. Concern for this potential adverse eject was not raised until the later stages of the chlorobenzilate RPAR proceeding, and then oPP concluded that the data were insufficient to establish the biological significance of adverse testicular ejects. Thus, oncogenic ejects were the main issue in the chlorobenzilate RPAR proceedings. The following assessment of the pathological activity of chlorobenzilate is restricted to oncogenic ejects. OPP'S decision to issue an RPAR against chlorobenzilate was based on two studies in which tumors developed in rats (Horn et al. 1955, Woodard Research Corporation 1966) and one study in which tumors developed in mice (Innes et al. 19691. Subsequently, NC! submitted additional chlorobenzilate carcinogenesis bioassay data on both rats and mice. oPP's final risk estimates were based on the Innes et al. (1969) study, which showed that chlorobenzilate produced statistically sig

Application to Chlorobenzilate 175 nificant increases in the incidence of tumors in male mice (U.S. EPA 1979~. After additional review and validation procedures, oPP determined that the Horn et al. (1955) and Woodard Research Corporation (1966) studies were unreliable. Upon further consideration, the EPA'S CAG chose to base the chlorobenzilate risk estimates on the Innes rather than the NC] study because the oncogenic response per unit of dose of cWoroben- zilate in the Innes study was 5 times greater (U.S. EPA 1978a). Also, animals in the Innes study were fed the compound begs ng at a younger, more susceptible age. Thus, the CAG concluded- that the Innes data were more appropriate for risk calculations under the conservative assumption that human response is similar to the most sensitive animal species and because of the possibility that people will be exposed to chlorobenzilate as infants (U.S. EPA 1978a). The Committee's estimate of the carcinogenicity of chlorobenzilate is detailed in Appendix C. The Committee endorses oPP's decision to use the Innes data as a starting point and based its own calculations on those data (see Appendix C), using the procedure recommended in Chapter 4. The CAT calculated for chlorobenzilate from the Innes data lies in the range 0.15~.38 (using a 90 percent confidence interval) with a most- probable estimate of 0.27. The use of data from the laboratory experiment that showed the highest incidence of tumors, disregarding several other experiments that also were available, naturally imparts a bias toward a worst-case estimate. Because the relationship between tumor incidence in laboratory animals and cancer incidence in humans is so obscure, the Committee elected to follow the policy of the CAG in this regard. This bias, however, should be kept in mind in interpreting the following analysis. ESTIMATION OF RISK UNDER VARIOUS REGULATORY OPTIONS By risk we mean the combined eject of the number of people exposed to pathogenic compounds, the levels of dosage received by those people, and the pathological activity of the compounds. All three components of risk have been discussed, and it remains to put them together. Separate estimates of risk are required for each regulatory option considered, so that they can be compared. The Committee evaluated the risks of the following five options, which it believes to be the salient ones (see the final major section of this chapter for a discussion of why these options are chosen):

176 REGULATING PESTICIDES A. Continue registration of all uses. B. Cancel all noncitrus uses. C. Continue registration of chlorobenzilate use on citrus and amend the terms and conditions of registration to require protective clothing and respirators; cancel all other uses. D. Cancel chlorobenzilate use on citrus to take eject after 5 years, and in the interim apply Option C. E. Cancel all uses. The first step in estimating the risks associated with an option is to estimate the lifetime doses of chlorobenzilate to which members of different population groups will be exposed if that option is adopted. This step is the subject of the next subsection. But one consequence of any option other than the status quo may be to increase the use of alternative pesticides (see Chapter 6), some of which may be carcinogen- ic (or otherwise hazardous). Thus, the second step is to estimate the doses of those alternative pesticides to which members of the population will be exposed if chlorobenzilate is regulated in accordance with that option. This step will be taken in the second subsection. These two estimates must then be combined to determine the "equivalent chloro- benzilate lifetime exposure" for each population. This combining, however, is not one of simple addition; the doses of chlorobenzilate and each of its alternatives have to be weighted in proportion to their CAPS (or other appropriate pathogenic activity indicator) before being summed (see Chapter 4~. This step is taken in the third subsection. Finally, a conception of the virulence of chlorobenzilate must be introduced. This is accomplished by comparing the experimentally observed CAl for chlorobenzilate to those of appropriate reference compounds. In this way the relative virulence of chlorobenzilate becomes known. This step is taken in the last subsection before the economic evaluation of benefits is undertaken. Lifetime Doses of Chlorobenzilate Both the length of exposure to chlorobenzilate and the intensity will vary with the various regulatory options. Exposure will be of maximum duration and dosage for all populations for Option A. For Option B. exposure to the Florida and general U.S. populations will be reduced by the noncitrus contribution. As it happens, the doses received by ingestion of noncitrus products are too small to affect the numerical estimates (which are truncated after two or three significant figures) that will be derived below except in the maximum-plausible cases. Option C diners from Option B in that the dosage received by citrus ground applicators is

Application to Chlorobenzilate 177 also reduced. For Option C, 50 percent compliance with the protective clothing and respirator requirements is assumed as the probable case and 20 percent as the maximum-plausible exposure case, as explained earlier (see this chapter, section on Human Exposure to Chlorobenzilate). This assumption affects the estimated lifetime exposure values. For Option D, citrus exposures will continue for only 5 additional years, noncitrus exposures will be eliminated immediately, and the ground applicator exposures during the 5-year period will be reduced as in Option C. For Option E, there will of course be no chlorobenzilate exposure. The resulting estimates of exposure to chlorobenzilate under the five options are presented in Table 7.6. The first column displays the daily doses, in millimoles per kilogram of body weight, that the Committee estimates to be most probable for members of three population groups under each of the five options. These figures are derived from the estimated daily doses in micrograms, whose calculation was described in the preceding section, by dividing them by the gram molecular weight of chlorobenzilate (325.2) and by the average weight of an adult man (70 kg). For example, the probable daily dose received by a ground applicator under Options A or B was computed by: 27,000 ,ug/day 325.2 g/mole x 70 kg = 12 x 10-4 m mole/kg. The second column is similar except that it records the maximum daily doses that the Committee judges to be at all possible. In view of the many inadequacies of the data and the uncertainties that have been discussed above, this judgment is necessarily subjective. In each case where such judgments had to be made, the Committee endeavored to arrive at a figure that corresponded to the upper limit of a formal 90 percent confidence interval, that is, a figure that we felt would be exceeded by actual results if the option were adopted in about 5 percent of the cases in which such judgments were made. All the estimates of maximum-plausible exposure in this analysis are to be interpreted in that sense. The third column computes the average number of days of exposure of members of the population to the estimated dosages. The genesis of these estimates has already been described. The figures shown are most- probable estimates; in fact, the remaining economic life of chlorobenzi- late may turn out to be greater or less than 10 years, and the average number of days that an applicator is exposed may differ from the 40 days a year that we believe most likely. In the interest of simplicity, no attempt was made to define a range of estimating errors. We believe that

178 ~D - C~ ao ~o U. o au ._ ._ o .~ o o Ct a' D C~ ·k D ,00 V o a, o .~ - - o - U. o - .H ~ w _ o ~: .° - a ~ D ._ . _ U ,] ~ _ o _ W ~ O . ._ ~ O ~ ~ O O . . . . O O O O O O O O O _ _ _ _ X X X X _ O mm . . . . ~ _ _ O O O O O _ _ _ _ o X X X X oo oo ~ ~ - , ~ _ ~ ~ ~ ~ ~ ~ _ 00 000 0000 0 ~ ~ ~ $- O O O ~ _ ~4 ~ X X X X O O O O ~ ~ d. oo oo _ _ ~4 _ 8888 o o o o o 00 OG o ~ ~ ~ ~ o _ ~ o o oo oo o o o o s" o o o o o o .e - § o o o _ _ ~ X X X X :^ ~ :^ U~ o 1 1 1 1 o o o o _ _ _ _ X X X X ~ ~ r> ~ ooo oo ~ ~ ~D ~ o o o o _ ~ _ ~ X X X X o oo o ° o o o o x oo - .S - ~s ~ ~ ~ V ~ ~ o

179 o o o o _ _ t_ _ X X X X cr . . . . o o o o 1 1 1 1 0 o o o _ _ _ _ ~ ~ ^ X X A =~ oo oo oo oo _ _ _ o o o o o o o o o ~ _ _ _ ~ X X X X :^ V~ ~o ~o o 1 1 1 1 o o o o _ _ _ _ X X X X _ o ~ ~ _ _ _ d_ . . . . ~] o o o o o x ~ r~ - 1 1 1 1 o o o o _ _ _ _ X X X X o~ ~ o~ o o o o . . . . o o o o o ~48 9-; E st) 8.~ ~y q ~Y ~ E j ~E ~ o o a ~y D ~ a'a~ ~ ~ x X X ~ ~ a ~ ~ ~ s~ ~ ~ I ~, C ~ ~ ~ . ; a ~a ~ a i 3; ~, ~ ~ a .= - c~ ao u, O ,., c .= u, .s: .O ~ ·~0 08 ~ d- 8 q) oo ~ o~ ~, t.> ~a .> O a, o C, ct~ Ct :^- ~ .~ ~_ a, ~ Do ~ ·S O == C~o ~ 3 0 D _ ~ ~ ~ :^ ~ ~ _ _ O ~ ~ - ~ ~'

180 REGULATING PESTICIDES if the attempt were made, it would not substantially affect evaluations of the regulatory alternatives. The final two columns are estimates of lifetime doses, obtained by multiplying either the first or second columns by the third one. Risks from Alternative Pesticides The considerations discussed in Chapter 3 determine how much information about and assessment of the alternatives to chlorobenzilate use on citrus are necessary (see Chapter 3, section on Modification to the Preliminary Ranking: The Role of Alternative Pesticides). For example, if it were assumed that no Class A (i.e., apparently a potential toxic hazard) or B (i.e., insufficient data to permit a reasonable judgment) substitutes continue to be available, and the RPAR evaluation of chlorobenzilate alone showed risks to exceed benefits, then no consider- ation of alternatives would be necessary; chlorobenzilate use on citrus should be severely restricted or cancelled. In fact, however, if chlorobenzilate were the only miticide registered for use on citrus, its withdrawal would result in a significant deteriora- tion in the cosmetic quality of the fruit (especially on Florida and Texas citrus). In addition, some have argued that the failure to treat for rust mites could lead to significant losses in yield per acre (see Luttner 1977b). It is conceivable that benefits of chlorobenzilate use would outweigh risks if no other miticides were available. Consequently, it is necessary to investigate benefits and risks of chlorobenzilate use under the assumption that some or all of the currently registered alternatives will remain available. oPP has identified at least seven apparently viable alternatives to chlorobenzilate for use on citrus: fenbutatin-oxide, dicofol, ethion, sulfur, oil, propargite, and carbophenothion. Because pesticide formula- tions have not yet been classified into Classes A, B. and C, as recommended in Chapter 3, it is unclear to which class each chlorobenzi- late substitute belongs. On the basis of preliminary toxicity profiles prepared by HED, concern might be raised about each of the substitutes for reasons of either acute or chronic toxicity (Burnam 1977, Bushong 1977~. In particular, ethion and carbophenothion appear to have high acute toxicities, and dicofol has been shown to be carcinogenic in an NC! rodent bioassay study (NC! 1978~. For the following discussion, then, it is prudent to assume that all of the substitutes would be either Class A or Class B compounds. If it is assumed that all of the substitutes will remain available, the benefit-risk estimates for chlorobenzilate should be calculated taking

Application to Chlorobenzilate 181 into account risks and benefits of the substitutes if chlorobenzilate were regulated. (oPP has done this in a rough, largely qualitative fashion in PD'S 3 and 4.) If this assessment clearly shows that benefits of continuing to use chlorobenzilate exceed risks' then it should be reregistered for citrus. If this is not clear, however, it then becomes important to identify the substitute pesticide or pesticides causing the ambiguity and to conduct a "qualified" RPAR on it or them (see Chapter 3~. To illustrate how alternatives would be taken into consideration using the risk assessment procedures recommended in Chapter 4, the Comm~t- tee has chosen dicofol. This selection is based on oPP's information that, from NC! animal bioassay data, dicofol is about 12 times more potent as a carcinogen than chlorobenzilate (U.S. EPA 1978a). Thus, an assessment of the risks from increased use of dicofol if chlorobenzilate were to be cancelled appears crucial to making a sound decision on chlorobenzilate. The following discussion is divided into two subsections addressing, first, exposure, then pathological activity. Dicofol Exposure Estimates of the additional daily dose of dicofol to each exposed population if chlorobenzilate were to be cancelled are presented in Table 7.7. Their derivation is sketched below. Estimates of total lifetime incremental exposure to dicofol for the various regulatory options are then presented in Table 7.8. To calculate the increased exposure to dicofol if chlorobenzilate were cancelled (regulatory Option E) or restricted (Options B to D), it is necessary to have an estimate of the extent to which dicofol will replace chlorobenzilate in citrus and noncitrus uses. From the benefit analysis in the next major section of this chapter, the Committee estimates that dicofol will replace from 13,500 to 67,500 chlorobenzilate acre-treat- ments, or 2-10 percent of the total citrus use.i For simplicity, the Committee assumed that the replacement rate for dicofol on noncitrus (other fruits and nuts) would be the same as for citrus. For applicators it was assumed that daily dicofol exposure would be the same as for chlorobenzilate multiplied by the 2.4 times greater application rate (Bushong 1977) of dicofol's active ingredient. Implicit in this assumption is the expectation that the mode of application and subsequent behavior of dicofol, including metabolism, will be similar to that of chlorobenzilate. The basis for these assumptions has not been explored, and better exposure estimates might be calculable; but the ones used here will suffice for the sake of illustration. Estimates of the incremental daily doses of dicofol from dermal and inhalation routes are shown in Table 7.7. To take account of the fact that under Options D and E when

182 E 0, i_ o ~ .a Cat .d D ~ _ v, | D ~C, o 8 o o ·C~ o .~ A o Cd .e LL an D o US 8 ~ o ~ Pi o ~ U. ~ ~ V ° ~ U. ~ o ~ x o o - C~ ~ a, mo cat 8 8 o ~ ~Go g 8 8 ~Do To ~ To .a o .g ILL ~ .o ~ o Ct o ~ ~ D ~ ~ o C=,, S ~ ~ ~ ~ ~ X ~ g O. ~0 3 ~ D o COL °O ._ _ ~ oo ~ . . . O O ~ O ~_~ ~ _ _ _ . O O O O O O O ~O. O O ~It- .. ~oo __ ~ D ~ ~ ~ ~ - ~ ~ Xi a ~

183 o ~_ o (5 ~ _o o _o oo o .. ... .. . ooo ooo o o _ ~ o ~ ~o o o ~ ~_ oo o ~ _ ~oo oo o o o o oo oo o . . . .. .. . ooo oo ooo -. _ o o j ~ U~ s o o oo ~ ~ ~;~ o o _o . . .. oooo o ~o o o~ o o\~, o ~ ~o _ ~. o o~o o o_ o . .. . o oo o o ~ o ~_ _ . . ... .. . oo ~o ~o o _ ~_~ ._ .e ° ~ ~ o ~ o ·c _ . .o ~ _ X X _ t: Q a' ~ ~ =_` ~ z ao ° 0 ~·= .V, O CO .k C 0 0 =O ~ X C ~Cd =-3 a' Ko ~ C~ o o D ~a ~ V] _ ~ 0> ~ O ~ ' O ~ .O - a) C ~ Q- ~ a,,_% O X C~ _ .U ~Ct ~q ~ _ i~ D ,~, c> D C O o ~=.~. C~ ~C) ~ ~ E C D ~^ O ~ ~ ~ ~ =2 s3 ~3 2 ~ ~ ~ . ~ ~ , . ·_~ Ct ·- - >D O C'0 OI - ~_o U, = - OLL, ·- O. O `:~ ~ U] ._C . C ~ ~- 0 ~ ~ ._ S ~ ~ O ~ ~ r ~° .Ta ~ ·~ ~ .~ ~ ~ O V, ~· ;> a~ a' ~ ~ ~t aO0~ D i~ ~O C 5)0 C _ ~-0 C.~-~C) - ·3 c~cd :^ CL C~ o ~o o O ~ ~ a, ~- ~ ~ ° ~ ~ ~ X ~ ~ , 2 X e A D ~ y,c ~ E Y ~ ~ E E 3 · u, .e ~ ~ ~ · V, - ~ 32 ·3 O ~D o - ' a, D ~° a' ~ u, ~ ~D ·3 ~ g g --3= ·33 _ ~ O - ~s~ ~ o ~ ~O ~t,_ O O ~o ~c O ~ O C ~g - . _ ~ O ~O 3= o . 3 ~ C g g C D 8 al C D 8 E ~ 8 e - 8, ~ ~ ao 3 ·~ ~ ~ ao O O ~ _ ~o.= _ O.0 ao ~ ~ ~ _ ~ X ~ ~ C C 8 ~ 3 ~ E ~ ec ~ O C ~

184 ~ g E ~ o ~ ._ D Hi._ -0 3 O o .0 o ~0 A: 65 o .C) o ·> ·= ~0 . ~ .~ Cd D o C) ~ D 60 ~ ~ At o D o ._ 00 ED - o a' in o .a ~ D e4 _ D D o 1 ~ O a, *O o t-~ A O O OO O ~)~ OO .. 000 00 0 V)O O_ O O O OO -c~ oo .. o o o oo o :^ ~o ~ - o - - o x o x o - - - 1 ~ 1 - ~ - ~ ~ x ~ x o o o o o ~ ~ oo oo 8 8 ooo o o - .e :S ~ ~ ~ V ~ ~ v) o. o o o o o o o o o _ _ X X (.~, oo oo oo . . . ~ ~ o 1 1 o o _ _ x - ~D 1 1 1 1 0 0 o o _ _ _ _ ~x, ~ ~ ^ ~ A A oo c`4 ~ ~ O _ . . . . 00 0 0 1 O 1 1 - O O _ _ X X O ~ V~ _ _ X X ~_ O O ~o 1 1 1 1 0 0 O O _ _ X X X X \d ~_ _ _ ~V) ~ ~ O O O O O O I_ ~_ I_ O O ~O _ _ ') _ X X X X L. ~o 1 1 1 1 1 O O O O O _ _ _ _' _ X X X X X ~ ~ ~ _ _ . . . . . O ~ ~ ~ O O ~ 0\ /= ~0 \0 O O O O O _____ 00 X X X X X _ ~t ~ ~t ~ ~ ~00 00 00 _ _ ·_ O O O O O _9 O O O O O O . O oO blo 00

185 o ~o 1 1 1 0 1 o o _ o _ _ _ X X X X o o r~ 1 1 1 0 o o _ - - X ^ ^ d o . . . o o o ~o 1 1 1 0 o o _ - - X ~ ~ o oo oo _4 . . . ~ ~ o ~o 1 1 O 0 1 _ _ o X X - _ _ X . . . o oo o o _ _ ~ X X X :^ ~o o - x o. o o - x a~ ~4 o 1 o - x - o o x r_ .~ o . ~5 ~o CO - ~o ox ~ 1 1 1 1 1 0 0 o o o _ _ _ ~ _ ~ `, X X X ^ ^ _ _ _ o o . . . . . o ~ ~ ~ o o Cr~ 1 1 1 o~ O O 0 1 1 _ _ ~4 o o X X X - - ~ ~ ~ ^ oo oo oo o o o . . . . . o oo ~ ~ .S o C~ g ._ - o s~ . - ,o. - - o g2 x <= 3 ~o (,, o C)~ .= ~·O X Ce .=o ·tD · ~_ .o "o~ ·- C: o ~o ~8 ~ a, ~ ·E '2 ~ _ s ~ ~a;, ~ o 5 ~ ~o ~c x ~ 8 2 ~o ~ ~5 8 , i ~ =5 ~ ~ ~ i . ~ -5 8 =5 ~ 8 8 ~ y -g 5 ~ ~ ~ ~ ~ oo

186 REGULATING PESTICIDES chlorobenzilate is cancelled dicofol is assumed to replace chlorobenzilate for 2-10 percent of the current citrus acre-treatments, the product of the daily dose and number of days exposed is multiplied by 2-10 percent (see Table 7.8~. This calculation assumes that the entire citrus applicator population will spray dicofol 2-10 percent of the time they would have sprayed chlorobenzilate, rather than a decreased applicator population doing all the dicofol spraying. For dietary exposure estimates, the 2-10 percent replacement rate of dicofol for chlorobenzilate would be factored in by multiplying the extent of chlorobenzilate use on specific crops by 2-10 percent in the equation: Ingestion (pa/day) = consumption (g/day) x extent of pesticide use on crop (~O) x assumed residue (ppm). In this equation, food consumption will remain the same for dicofol as for the chlorobenzilate calculations. The range of assumed residues will diner, however. For the Florida population, dicofol residue levels in meat and meat by-products from Florida beef and lamb and milk from the cattle are calculated using the accumulation ratio procedure recommended for chlorobenzilate. Only one estimate is calculated representing both the probable and maximum-plausible residue levels because the Committee found no basis for making an alternative estimate. The final daily dosage estimate is a range because the estimated replacement rate of dicofol for chlorobenzilate is a range. It was necessary to derive an estimated accumulation ratio for dicofol in beef by indirect means, using various published data from beef and poultry experiments. Accumulation ratios for dicofol and DDT in poultry fat can be calculated using data from Fries (1969) and McCaskey et al. (1968~. These values are 2 and 6, respectively. Similarly, an accumulation ratio of approximately 1 for DDT in beef fat can be calculated using data from Cummings et al. (1967~. An estimated accumulation ratio for dicofol in beef fat can then be derived using the relationship: Dicofol accumulation ratio in poultry fat DDT accumulation ratio in poultry fat = Dicofol accumulation ratio in beef fat . DDT accumulation ratio in beef fat

Application to Chlorobenzilate ·87 Inserting the calculated accumulation ratios into the above equation and solving for dicofol in beef fat gives: Dicofol accumulation ratio in beef fat = 2/6 x 1 = 0.33. An estimate of the maximum-plausible dicofol residue in citrus pulp can be derived by multiplying oPP's estimate of chlorobenzilate residue in pulp (2 ppm) times the greater application rate of dicofol (2.4x) (Bushong 1977) times the greater persistence of dicofol residues over chlorobenzilate residues (2-3x) (Carmen et al. 1972, Gunther 1969~. Dicofol residue in pulp is estimated to be approximately 10 ppm. The steady-state residue of dicofol in animal fat, then, is the accumulation ratio (0.33) x the concentration of dicofol in the diet (10 ppm in pulp x 16 lb of pulp - 20 kg diet), or 0.33 x 3.6 = 1.19 ppm. Beef cuts averaging 30 percent fat (a maximum value) would contain 0.36 ppm residual dicofol. If Florida beef and lamb were consumed at a rate of 143.2 g/day (oPP's value), the estimate of daily dicofol ingestion by Floridians from this source would be 143.2 g/day x 2-lO~o dicofol replacement rate x 10% of chlorobenzilate crop treated x 0.36 ppm residue = 0.1~0.52 payday (see Table 7.7~. As with chlorobenzilate, estimates of dicofol residue concentrations in milk may be derived from the accumulation ratio approach described for beef. With dicofol residues in pulp at a level of 10 ppm, the resulting residue level in milk fat would be 1.19 ppm. This establishes a residue estimate of 0.042 ppm for mild containing 3.5 percent fat. If 184.7 g of milk were consumed daily (oPP's value), the estimate of dicofol ingestion would be 0.1~0.78 ,ug/day (see Table 7.7~. For the general U.S. population, the Committee's estimates of probable and maximum-plausible incremental daily doses of dicofol (resulting from replacement for chlorobenzilate) are based on the assumed residue levels of chlorobenzilate in citrus and noncitrus products multiplied by a factor of 5. This factor is rationalized as was done previously when estimating dicofol residues in citrus pulp, i.e., the greater application rate of dicofol (2.4 x ~ times the greater persistence of dicofol residues (2-3 x). Estimates of additional daily doses of dicofol for the U.S. population are shown in Table 7.7. For the Florida and general U.S. populations, total (lifetime) incremental exposure can be estimated by multiplying the daily dietary dose by the number of years dicofol would be expected to be used under the various regulatory options (see Table 7.8~. Again, the Committee assumes an economic lifetime of 10 additional years for dicofol (see Chapter 4~.

188 REGULATING PESTICIDES Pathological Activity of Dicofol The following assessment of the patho- logical activity of dicofol is restricted to its carcinogenic potential. The data for the calculations come from the 1978 NC! Bioassay of Dicofolfor Possible Carcinogenicity (NC! 1978~. The procedure recommended in Chapter 4 and illustrated in Appendix C was used to calculate the CAT for dicofol. Both high- and low-dose male mice showed statistically significant increases in hepatocellular carcinomas, although no statisti- cally significant association between dose and mortality was observed. For the low dose, the CAT lies in the range 0.25-1.16 (using a 90 percent confidence interval) with a most-probable estimate of 0.86. For the high dose the range is 0.3~0.66 (using a 90 percent confidence interval) with a most-probable estimate of 0.55 (see Table 4.3~. Since the low-dose test group developed a tumor yield of approximately 55 percent- which is comparable to the tumor yield in the Innes et al. (1969) chloroben~ilate study the low-dose CAT estimate will be used in subsequent calculations. Translating Dicofol Lifetime Exposures into Chlorobenzilate Exposure Equivalents As noted earlier, to present the most useful information to a decision maker about the pesticide exposure for each affected population attendant to the various regulatory options, it is necessary not only to estimate exposure outcomes for the pesticide in question-in this case chlorobenzilate but also to incorporate exposure to alternatives into the estimate. This is accomplished here by translating the lifetime exposure from dicofol (Table 7.8) into chlorobenzilate exposure equivalents using the method described in Chapter 4 (see the section on Combining Exposure and Pathological Activity). Chlorobenzilate exposure equivalents for dicofol are shown in the last column of Table 7.8. The uncertainty associated with the CAT values for dicofol and chlorobenzilate was taken into account as described in a footnote to the table. The chlorobenzilate exposure equivalents for dicofol will be combined with the chlorobenzilate exposure values in Table 7.6 in the final major section of this chapter, to produce a single exposure estimate for each population segment under the various regulatory options (see Table 7.26~.

Application to Chlorobenzilate Reference Compounds The implications of lifetime exposures expressed in millimoles per kilogram of body weight are hard to perceive. To gain an appreciation of the ejects on public health of using chlorobenzilate it is, therefore, virtually essential to compare its carcinogenic potency and anticipated dose levels with those of other pesticides that have been considered extensively and have been subjected to regulatory decisions. The chlordane-heptachlor complex is well suited to this purpose. Their chemical structures and metabolic behavior in mammals are roughly similar to those of chlorobenzilate. Laboratory experiments under comparable test conditions and within comparable parameters are available for chlordane, heptachlor, and chlorobenzilate (see Table 4.3~. The uses of all three compounds in agriculture and the resultant patterns of exposure are roughly similar. The nature of the principal risk to public health that chlordane and heptachlor present- cancer is identical with that of chlorobenzilate. The registrations of both chlordane and heptaclor for most uses were suspended in 1975 on the grounds that "the human cancer hazard posed by these pesticides and the lack of benefits to outweigh this risk were the bases for their suspension" (CEQ 1976:31~. Their registrations were cancelled after further investigation in March 1978. Some determinations of the CAYS for both chlordane and heptachlor are included in Table 4.3. Since the CAT adopted for chlorobenzilate is derived from experiments with male mice (females appear to be less sensitive), the determinations for chlordane and heptachlor derived from experiments with males are the relevant ones for comparison. There is only one such determination for heptachlor, giving a CAT of about 19. There are five determinations for chlordane using males at five different dose levels. At the lowest concentration, 5 ppm, the indicated CA' was 3.1, but the observed rate off excess tumor response was so low that the estimate is not reliably greater than zero. At higher dosages, the indicated percent of animals in which tumors were induced by administration of chlordane were 80 percent at 25 ppm and slightly less, but not enough to be statistically significant, at the higher doses. We infer that a substantial increase in tumorgenesis is induced by the administration of 25 ppm of chlordane, but that thereafter any increase in responses is too gradual to be detected by samples of the size used in the experiments reported. Accordingly, we have adopted a CAT of 14.5, corresponding to the 25 ppm dose level of chlordane. These estimates of carcinogenic activity are displayed graphically in 189

190 REGULATING PESTICIDES Figure 7.1, which shows the accompanying uncertainty, and they are used in the evaluation in the last section of this chapter. AN ECONOMIC EVALUATION OF THE BENEFITS This section applies the principles and recommendations set forth in Chapter 5 to assess the benefits of using chlorobenzilate. Since oPP has previously completed a benefit assessment for chlorobenzilate (Luttner 1977a, by, the discussion in this section will employ many of the data compiled for the oPP analysis. The preliminary benefit analysis by oPP (Luttner 1977a) estimated that 1.1 million lb of chlorobenzilate were applied in 1975 (the base year for the analysis). Most of this chlorobenzilate, approximately 920,000 lb, was applied to control the citrus rust mite on Florida and Texas oranges and grapefruit and the citrus bud mite on Arizona and California lemons. Another 76,000 lb were used on other citrus crops (e.g., limes and tangelos). Cotton farmers made slight use of chlorobenzilate, applying around 39,000 lb. Finally, the remaining 81,000 lb were sprayed on a wide variety of fruits, nuts, and other miscellaneous crops. The analysis in this section will be restricted to the major uses of chlorobenzilate, namely, mite control on oranges, grapefruits, and lemons.2 The presentation is divided into three parts. The effects of chlorobenzi- late use on yield and quality of citrus are discussed in the first part. Estimates of the eject that chlorobenzilate use has on pest control costs in the citrus industry are presented in the next part. Finally, the estimated yield, quality, and cost ejects are translated with the aid of conventional benefit-cost analysis into overall estimates of the benefits of chlorobenzilate use (or, alternatively, of the benefits forgone because of withdrawal of chlorobenzilate from the market). EFFECTS ON YIELD AND QUALITY The oPP preliminary benefit analysis (hereafter, PBA) adopts a number of important assumptions concerning effects on yield and quality likely to be associated with the loss of chlorobenzilate. Briefly, the oPP analysis assumes that the available alternative miticides will " . . . provide yields and product quality comparable to chlorobenzilate . . . " (Lutt- ner 1977a:45a). However, it also assumes that failure to treat for mites would have significant adverse effects on fruit size, appearance, crop yield, and tree-stock stamina (Luttner 1977a:44~. This section considers each of these key assumptions.

Application to Chlorobenzilate 25 15 23 10 5 -to Heptachlor Chlordane 191 ...................................................... ....................................................... ............................................... ............................................ ................................................... ....:.................................... ................. ; - ·................................ ...... 2 .. ............ ... : .. ............................................ ........................... 2. , : 2 : ... ...................... : ............................... .. .. ................... .. I......... . ........... I................ ~ .... I... E .... . .....i ~I:;. :;:;:;--: :-:-.-:-:-.-:-:;::;:-.;:-:-:;:;:-:-:-' Dicofol Chlorobenzilate FIGURE 7.1 CAN'T used in Chapter 7 calculations (see section on "Estimation of Risk Under Various Regulatory Options" in this chapter) with confidence intervals for chlorobenzilate, dicofol, and the selected comparison compounds. Source: Table 4.3.

192 REGULATING PESTICIDES No Miticide Treatments If chlorobenzilate were not available, citrus growers would have the choice of either treating with an alternative miticide or not treating at all. Consideration of the no-treatment option led oPP to the following conclusion (U.S. EPA 1978a:44 46~: . . . uncontrolled mites cause reductions In fruit size of 12% for oranges and 17% for grapefruit. Fruit size declines also occur in lemons, but these ejects have not been fully quantified. Also, overall yield can be reduced by mite infestations due to fruit drop (Allen 1978~. It is estimated that such reductions In fruit size and overall yield would reduce grower gross revenues by about $159 million per year (approximately 16 percent of total industry revenue). These estimates are supported by data developed by Allen (1978) and brought to oPP's attention by USDA'S assessment team for chlorobenzi- late.3 A subsequent search of the literature revealed that the USDA assessment team failed to bring to oPP's attention eight additional studies of the eject of chemical miticides on citrus yields. None of these additional studies which are summarized in Table 7.9 found a statistically significant difference in yields between sprayed and un- sprayed plots that could be attributed to rust mites. The largest percentage decrease in yield (23 percent) was measured by Reinking (1967), but it was not statistically significant because of the great variability in yield between trees. Simanton (1962) and Griffiths and Thompson (1953) found 20 percent decreases; Griffiths and Thompson's finding was not statistically significant, however, and Simanton did not report a significance test. In no case did the unsprayed trees have greater yields than the sprayed ones, so that the possibility of some decrease cannot be ruled out. If there is a decrease, however, it is too small to be detected by any of the experiments found in the Committee's search of the literature. Allen and Stamper (1979) examined the frequency distribution of citrus rust mite damage on citrus fruit and concluded that the effect of russetting was not serious until 50-75 percent of the surface was russetted. Generally, less than 5 percent of the oranges and 40 percent of the grapefruit in unsprayed groves were heavily scarred by rust mite. Allen (1978) found that the drop rate of fruit scarred by rust mite did not increase above that of unscarred fruit unless 75-80 percent of the surface skin was russetted. Allen (1979) also found that weight of citrus fruit decreased with 50 percent or greater scarring, while the percentage of total soluble solids increased in proportion to the amount of rust mite

Application to Chlorobenzilate 193 damage. It was also found that grapefruit achieved a smaller diameter when more than 87.5 percent of the surface was scarred. All of the above effects were attributed to increased water loss from scarred fruit. Irrigation of citrus groves with citrus rust mite infestation would alleviate some of the problems of water loss caused by severe russetting. In Florida, 78 percent and 72 percent, respectively, of the orange and grapefruit acreage is irrigated, while all of the bearing citrus acreage in Texas and California is irrigated (U.S. PEA 1976~. In fact, oranges with rough lemon rootstock, with its extensive root system and greater water- gathering capacity, are less affected by rust mite damage than orange trees with less extensive root systems (Allen 1979~. McCoy et al. (1976b) found that 2~30 percent of the outside surface of oranges could be damaged by rust mite with no effect on yield of fruit. They also found that if extensive damage (50 percent of all the fruit in a grove with extensive surface bronzing) occurred, the fruit could either be harvested early or irrigated to prevent excessive moisture loss and peel shrinkage. McCoy (1976) analyzed the effects of rust mites on leaves in Valencia orange groves. He concluded that uncontrolled mite populations could cause a 3 percent increase in leaf drop, but this would not be severe enough to affect the vigor of the tree or its subsequent yield. Again, he suggested irrigation, a common practice in Florida, as a means of controlling this problem. Van Brussel (1975) analyzed the interrelation of greasy spot and rust mite in citrus groves in Surinam. He found that trees with heavy rust- mite and greasy-spot infestations suffered defoliation, while trees with higher levels of rust mites that were sprayed to control greasy spot did not show any signs of defoliation. He concluded that greasy soot was the agent responsible for leaf drop in citrus. Another study by McCoy (1977) found that unsprayed orange groves had 16-38 percent of the fruit damaged by rust mites, while groves sprayed with chlorobenzilate had 1-8 percent of the fruit damaged. McCoy's previous work (1976a, b) and work by Allen (1978, 1979) indicated that this level of damage to an unsprayed grove would probably not cause a decrease in fruit or juice yield. Yothers (1918) presented data indicating that russetted oranges and grapefruit were 1~17 percent smaller than clear fruit. However, his data did not reveal the yield or russetting differences between sprayed and unsprayed plots. Moreover, Sinclair (1972) found that, on the basis of fruit volume, smaller grapefruit yielded significantly more juice than larger fruit. Apparently, the peel affects the percentages by volume more than the percentages by weight. v , .

194 o cn . ~ ·' ~ o Cal Ct in V, ._ ·o o U' - ._ a: .zO =O ~ V, o Cal Cal ~ ^ ~ ~ Cal _ ~ in, At, \o \4, in, _ 8 8 y a +~l =8 =8 +8 ~ ~ ~ ~ ~ ~ ~ Cal z Z Z Z Z it- Z V, .8 ~ ~ _ _ D ~ .8 .8 g =\ t- ~ ° ~ ~ O ~ ° try ~of D ~D 5 ~ D D ~ _. =^ ~i _ - _ ~ a. ~ O ~ ~ ~ ~ ~ ~ ~ ~ Ce ~ C~ ~ C. C. ~ ~ cq .c ·c ~ ~ ·c ·c ·c ·c ~X X ~ ~ ~ ~0 .= ao ~_ .g D - o w o .3 :- 08 - U' :^ :^ D :- g U) .` U. - ._ :^ - :~ U. d · C ~._ 9 D~ .Q ~ ·s ·~ ~ s~ ~ ·- ~ ~ o a~ ~a u~ - ~ ~ ~ C~

Application to Chlorobenzilate 195 The weight of all this evidence supports the conclusion that the economic value of chlorobenzilate treatments of citrus fruits is far less than the estimate of $159 million a year adopted by oPP. (oPP's quantitative estimate is based only on yield erects; it does not consider the impact on consumer demand of cosmetic erects from not treating.) However, for purposes of estimating this value, it is misleading to compare the yields and qualities of orchards treated with chlorobenzilate with those of orchards where no protective measures have been taken. Citrus rust and bud mites are perceived by most growers and citrus entomologists as major pests. Relatively economical alternatives to chlorobenzilate are available. Consequently, growers will almost certain- ly continue to treat for mites, either with chlorobenzilate or with an alternative. Alternative Miticide Treatments The oPP assumption that the alternatives will provide yields and product quality comparable to that provided by chlorobenzilate was both plausible and acceptable to the USDA assessment team (J. Knapp, University of Florida Extension Service, Gainesville, personal communi- cation, December 19781. A number of relevant studies lend support to the assumption. For instance, in a study of Texas orange and grapefruit groves, Reinking (1967) analyzed the effect of oil on citrus pests. Although he stated that the correct grade of oil controls Texas mites, during the periods of his study neither the experimental nor the control plots contained significant mite populations. More importantly, however, he found that the correct grade of oil did not injure the trees or affect the yield and quality of fruit. Jeppson et al. (1955) analyzed the erects of chlorobenzilate and oil in the control of the citrus bud mite and the citrus red mite in California. They found that when chlorobenzilate was used, 20 percent of the buds became infested with bud mite, while 32 percent of the buds were infested when petroleum was used. However, petroleum sprays electively controlled the red mite, while chlorobenzilate was relatively ineffective. Townsend (1976) analyzed 2 years of integrated pest management (IPM) programs in Florida orange groves. In the IPM groves, oil was used to suppress mite populations, but in some instances, large mite populations were treated with chlorobenzilate. In the fall only a few groves along the ridge were sprayed for rust-mite control, while none of

196 REGULATING PESTICIDES the groves along the east coast of Lake Okeechobee needed mite control for late russetting or bronzing. Oil was also found to suppress greasy-spot fungus at low fungal densities. The IPM demonstration blocks, using little chlorobenzilate, had rust-mite densities 2 times greater than the conventional plots with no reduction in yield or soluble solids. McCoy et al. (1976a) also examined a reduced pesticide program for pest control in Florida orange groves. Three spray programs were analyzed over a 4-year period: (1) conventional; (2) IPM; and (3) no spray. The IPM blocks received, on average, 1.2 fewer sprays for rust mites per year than the conventional blocks and showed no more rust- mite activity than the control. The unsprayed blocks also showed no increase in late rust-mite damage over the 4-year period. Yields in the unsprayed plot were reduced by greasy-spot fungus and dry weather. Under humid growing conditions, the fungus Hirsutella thompsonii controls citrus rust-mite populations. McCoy et al. (1976a) found that in the unsprayed plots this fungus effectively controlled citrus rust mite over a 4-year period. Indeed, the fungus thrived in the unsprayed plots compared to sprayed plots. Townsend (1976) stated that H. thompsonii controls the citrus rust mites when mite populations reach high densities in July and August. McCoy et al. (1971) reported suppression of citrus rust-mite populations, equal to that achieved through the use of chlorobenzilate, with a 5 percent spray of the fragmented mycilia of H. thompsonii. The investigators stated that their methods worked well at high levels of mite infestation and under humid conditions. They did not determine whether this process was economical or whether it would control subeconomic mite populations (i.e., populations for which the cost of treatment exceeds the losses from not treating). Van Brussel (1975) also achieved excellent rust-mite control in Surinam using H. thompsonii as a biological control agent. Chlorobenzilate has been found to cause a 50 60 percent reduction in entomopathogenic fungi (Olmert and Kenneth 1974), and thus it can interfere with control of citrus rust mite by H. thompsonii. In summary, the evidence indicates that the yield and quality of citrus crops will not be diminished appreciably, if at all, if farmers are required to replace chlorobenzilate treatments with some alternative. CHANGES IN PEST CONTROL COSTS it' To estimate the change in pest control costs (ARC) that cancellation of chlorobenzilate's registration would occasion, it is necessary to develop measures for several variables. The requisite information is indicated by

Application to Chlorobenzilate 197 an examination of the operational definition of APC relevant to a cancellation. For any specific region and type of citrus, ~ K APC = ~ ~[AjTl(MCj + ACj)] - ~ [AkTk(MCk + ACk)] j c1 k~1 where j denotes one of J (nonchlorobenzilate) pest control methods used on citrus; k denotes one of K spray mixtures containing chlorobenzilate (e.g., chlorobenzilate-sulfur); Aj and Ak denote acres treated per year by method j or k, respectively; Tj is the average number of treatments per year by method j, and Tk is defined similarly; MCj and MCk are material costs per acre-treatment; ACj and ACk are application costs per acre- treatment; and lt denotes the change in the variable or expression that follows it that would be induced by cancelling the registration of chlorobenzilate for the specific use being analyzed. In implementing this measure, it is assumed that total citrus acreage remains constant. The assumption is not crucial to estimating the change in production costs per unit of output (the ultimate objective of this section), and it will be relaxed in the next section. The expression AfAj Tj (MCj + ACj )] is the change in expenditures on control method j (which might be a nonchemical method) arising from the cancellation of chlorobenzilate. Usually (but not always) this term will equal zero (when total acreage is held constant), unless the pest control method itself provides an alternative to the use of chlorobenzi- late. Presumably, for the alternatives this term will be positive or zero. In order to estimate /`PC occasioned by the denial of chlorobenzilate, it is necessary to have information on each of the variables in the equation. In the next section we discuss oPP's approach to quantifying the expected APC in the event that chlorobenzilate's registration for use on citrus is cancelled. We then discuss and implement some alternative measures for some of the components of APC. Data and Assumptions The main source of the data used by oPP for measuring APC is the Doane Specialty Crops Study for the years 1972, 1973, and 1974 (reported in Luttner 1977a). The Doane report provided estimates by state and type of citrus for acres treated with various miticides, average number of treatments per year, application rate per acre-treatment, and the unit cost of the miticides (e.g., dollars per gallon). These statewide estimates are based upon samples of citrus growers in the various states (approximately 300 in Arizona-California, 375 in Florida, and 200 in

198 REGULATING PESTICIDES Texas: E. Dixon, Doane Agricultural Service, Inc., St. Louis, Mo., personal communication, November 1978~. oPP supplemented these data from Doane with information from published and unpublished papers, personal communications with citrus experts, and the report submitted by the USDA assessment team for chlorobenzilate (USDA 1977b).4 In the following discussion, the specific sources of the data used by oPP for each variable are identified. 1. The K currently used miticide spray combinations containing chlorobenzilate are obtained from the Doane survey. The PBA considers four chlorobenzilate combinations for oranges, two for lemons, and seven for grapefruit (Luttner 1977a:5~52~. 2. The J pest control methods that would be more widely used following a chlorobenzilate cancellation were identified in several ways. Identification of the relevant alternatives to chlorobenzilate was based upon three criteria. To be classified as an alternative for use on a specific type of citrus in a state, actual use of the chemical had to be reported in the Doane survey, the chemical had to be listed among recommended miticides in the state's citrus guide, and it had to be identified by state entomologists as an alternative likely to be used in place of chlorobenzi- late (Luttner 1977a:48~. The number of alternatives, according to state and type of citrus is (Luttner 1977a:5~52~: Oranges Arizona NA California NA Florida Texas Grapefruit Lemons NA NA 7 3 5 s NA NA = Not applicable, because no chlorobenzilate use was reported for this site. In addition, the Supplement to the PBA (hereafter, S-PBA) concluded that loss of chlorobenzilate would eventually occasion increased use of scalicides in Florida because the alternative miticides most likely to be used are lethal to insects parasitic on scale (Luttner 1977b). 3. Data on the average number of applications (~ per year for miticides are available from the Doane survey. However, with one exception, the PBA does not use these estimates of T; rather, the PBA generally assumes that T for each of the alternatives is the same as the

Application to Chlorobenzilate 199 average T for the various chlorobenzilate mixtures relevant to a use site. The one exception is the use of oil on California lemons; in this case, T is set equal to 2 (a value consistent with the Doane survey) for two thirds of the California lemon acreage (Luttner 1977b: 9), rather than 1 as is assumed for chlorobenzilate mixtures used on California lemons. The assumption that T is the same for virtually all the miticide mixtures (chlorobenzilate mixtures and substitutes) considered is not supported by the data in the Doane survey. oPP's estimates of T for the Florida scalicides are based upon information provided in the USDA assessment team report (USDA 1977b), hereafter abbreviated USDA-ATR. 4. oPP's estimates of base acres treated (Ak) with any one of the chlorobenzilate mixtures are taken from the Doane survey (Luttner 1977a). 5. For the most part, oPP assumes that the total number of acres treated with miticides would not be affected by a regulatory action against chlorobenzilate. Specifically, for miticide treatments in Arizona, Florida, and Texas, it is assumed that I:j ~ Aj = Ilk Ak . The exception occurs in connection with California lemons: to account for the claim in the USDA-ATR that cancellation of chlorobenzilate would lead to an increase in the number of acres requiring treatment, oPP inferred that base acres treated would grow to 41,000 (after 5 years) from the current 5,000 (Luttner 1977b:9~. In addition, oPP inferred (also from relatively general statements in the USDA-ATR) that the loss of chlorobenzilate would so hamper the IPM program in Florida that an the Florida citrus acreage (around 850,000 acres) would require two additional scalicide treatments (Luttner 1977b:6~. 6. The oPP estimates of acre-treatments for the various chlorobenzi- late mixtures (Ak Tk ~ are easily constructed from the data provided by Doane on base acres and average number of treatments (Luttner 1977a:51-53, 114-128~. . - 7. With the exception of California lemons, the change in total acre- treatments for the alternatives is taken as equal to the total acre- treatments for the chlorobenzilate mixtures. That is, oPP generally assumes that for the alternatives, I:j 1` Aj Tj = ~ k Ak Tk (see, e.g., Luttner 1977a:5~524. For California lemons, acre-treatments are assumed to expand (after 5 years) to 68,470 from 5,000 with chlorobenzilate an estimate dependent upon the assumption that 41,000 acres would require at least one additional treatment of oil (as a replacement for chlorobenzi- late) and two thirds of those 41,000 acres would require two additional oil treatments per year (Luttner 19?7b:9~. The additional acre-treatments for the scalicides in Florida are estimated by oPP to be 1.7 million after 5

200 REGULATING PESTICIDES years (850,000 acres x 2 additional treatments per year [Luttner 1977b: 8~. 8. oPP estimated the anticipated change in acre-treatments for any specific alternative (with the exception of oil on California lemons) by simply dividing the estimate of the change in total acre-treatments by the number of viable alternatives (Luttner 1977a:50~: ~AAjTj AA jTj = ~ . The estimation of additional acre-treatments for oil on California lemons and for scalicides in Florida was discussed above. 9. oPP estimated material costs per acre-treatment (MC!J for the chlorobenzilate mixtures and the alternatives in two different ways. Estimates for MC for all of the various chlorobenzilate mixtures and alternatives are available from the Doane report. However, in some cases " . . . the expenditure data in the Doane material were perceived (by oPP) as being either excessively low or high . . . " (Luttner 1977a:48~. In such instances, the expenditure per acre-treatment was derived by using prices from current pesticide price lists and the recommended application rates reported in the state citrus guides (Luttner 1977a: 5 52~. oPP's estimate of MC for the additional scalicide treatments in Florida is based upon information provided by the assessment team (Luttner 1977b:8). 10. The PBA generally assumes that the application costs per acre- treatment are the same for all of the chlorobenzilate mixtures and the alternatives (Luttner 1977a:5~52~. Consequently, this cost element is ignored (except in the case of California lemons and scalicide treatments in Florida) in estimating the change in treatment costs that cancellation of chlorobenzilate would occasion. That is, the estimate of ARC costs is collapsed to: APC= ~ AAjT,fMCj) ~ A kTk~ M Ck) k The oPP estimates of application costs per acre-treatment for the oil sprays on California lemons and scalicide treatments on Florida citrus are based on information provided in the USDA-ATR. The oPP estimates for the change in annual pest control costs that are likely to occur if chlorobenzilate is withdrawn are shown in Table 7.10. The total is $57.6 million.

Application to Chlorobenzilate 201 TABLE 7.10 OPP's Estimated Change In Annual Pest Control Costs on Citrus Following Cancellation of Chlorobenzilatea fin thousands of 1975 dollars) Type of Citrus State Grapefruit Lemons Oranges Total Arizona $ O $ 0 California $4,821b $ 4,821 Florida Non-IPM $ 468 $ 34 $ 1,541 $ 2,043 IPM $8,570b $ 398b $41,523b $50,491 Texas $ 44 $ 230 $ 274 Total $9,082 $5,253 $43,294 $57,629 a These estimates are from Luttner (1977a; pp 50-52; 1977b, pp. 9, 12). b These estimates apply to the fifth year following withdrawal of chlorobenzilate. The extent to which oPP's overall benefit estimate is highly dependent upon two key assumptions is demonstrated through the sensitivity analysis in Table 7.1 1. Clearly, these crucial assumptions by oPP concerning Florida scalicide treatments (accounting for 87.6 percent of the total benefits) and bud-mite control on California lemons (account- ing for 8.4 percent of the total) require scrutiny. We shall review, in order, the impacts of disallowing the use of chlorobenzilate on (1) the Florida lPM program, (2) bud-mite control on California lemons, and (3) mite control on Arizona, Florida (non-'PM effects), and Texas citrus. Florida IPM Effects Currently, the most important Florida citrus pests are citrus rust mite, greasy-spot disease, and citrus snow scale. Before 1960, purple scale and Florida red scale were also major pests of citrus. However, the introduction of two hymenopterous parasites, Aphytis lepidosaphes and A. holoxanthus, relegated these two scale insects to the role of minor pests on all the Florida citrus acreage (Brooks 1977:31~. In addition, Florida entomologists are currently attempting to establish a third parasite, the Hong Kong wasp (A. Iingnanensis), to provide control over citrus snow scale (Luttner 1977b:2~. An important advantage of chlorobenzilate is its specificity to mites: it has little deleterious eject upon the parasites used to control these scale insects. The oPP analysis assumes that the alternatives to chlorobenzilate

202 REGULATING PESTICIDES TABLE 7. l ~ Contribution of Key Assumptions to OPP's Estimate of Annual Benefits from Chlorobenzilate Use I. OPP's estimate of annual benefits It. Key assumptions underlying benefit estimate A. Cancellation would disrupt the Florida IPM program leading to two additional scalicide treatments on all Flora trus acreage (850,000 acres) at a per acre cost of $10.08 for materials and $19.62 for application: B. Cancellation would result in acre-treatments for bud- m~te control in California expanding to 68,470 from the present 5,000. Per acre-treatment costs with chlorobenzilate average $76.60 (including application costs), whereas the alternative (oil) averages $76 (including a $40 application cost). Thus, the net cost InCIeaSe IS: Overall contribution of the two assumptions: $57,629,000 $50,491,000 (87.6 percent of total) $ 4,821,000 (8.4 percent of total) $55,312,000 (96.0 percent of total) Source: DenvedfromLuttner(1977a,b). would destroy the hymenopterous parasites of the scale insects. "This situation would result in a return to the pre-'PM scale insect control practices . . . in which two dilute scalicide sprays were applied to all of the commercial Florida citrus acreage each year" (Luttner 1977b:3~. However, this assumption appears to be unrealistic. The PBA identifies the major alternatives to chlorobenzilate in Florida as dicofol, ethion, oil, and sulfur. Dicofol and oil have little deleterious eject on the beneficial insects, so use of these two alternatives would certainly not be disruptive of the IPM program. He use of dicofol in groves infested with snow scale can create complications, however; this point is discussed below.) Both ethion and sulfur are toxic to the scale parasites. However, ejects of these two pesticides are relatively short-lived (USDA 1977b: Attachment V). Consequently, while an ethion or sulfur treatment may destroy most of the adult scale parasites in a grove, it will not prevent the progeny from reestablishing the predator population. In fact, the Florida Citrus Spray and Dust Schedule 1977 (University of Florida 1977) recommends a summer application of ethion and oil, and expresses no concern that

Application to Chlorober~zilate 203 such an application will interfere with the natural control of purple and Florida red scale. Of course, frequent or excessive applications (especial- ly of sulfur) would eventually destroy the Aphytis populations (J. Knapp, University of Florida Extension Service, Gainesville, personal communi- cation, December 1978~. The key to continued natural control of the scale insects lies in the avoidance of frequent applications of materials toxic to the parasite populations. Since most citrus groves in Florida receive only two to three spray or dust applications a year (USDA 1977b: Attachment XII, 1), the maintenance of the predator populations would appear to be a feasible goal, contrary to the conclusion reached by oPP. As noted above, the use of dicofol is consistent with natural control of purple and Florida red scale. However, dicofol treatments can produce abnormally high populations of citrus snow scale (Brooks and Whitney 1977:431~. Consequently, the Florida Citrus Spray and Dust Schedule 1977 recommends that dicofol be combined with a scalicide for use in groves infested with citrus snow scale. This evidence clearly supports oPP's view that " . . . the projected overall increase in use of dicofol following a cancellation of chlorobenzilate would also increase the use of scalicides for snow scale control" (Luttner 1977b:3~. The relevant issue, of course, is: how costly would these additional scalicide treatments be? Approximately 45 percent of Florida citrus groves (perhaps 350,000 acres) harbor snow scale, and about 25 percent of the groves super economic infestations requiring treatment (Brooks and Whitney 1977:427~. Thus, dicofol treatments could aggravate the snow-scale problem on about 350,000 acres. However, approximately 200,000 of these acres are already being treated for snow scale. These treatments are assumed in the Committee's analysis to be adequate to handle any problems created by the use of dicofol. Only those 150,000 acres with noneconomic infestations would require additional scalicide treatments as the result of using dicofol. The Committee assumes that the scalicide (material) cost per acre- treatment is $10.00 (the S-PBA reports scalicide costs ranging from $5.00 to $12.66 per acre-treatment). According to the Florida Citrus Spray and Dust Schedule 1977, the scalicides can be applied along with dicofol. However, the application of a scalicide complete coverage of all wood is essential will increase the amount of water used per acre and increase the time required to spray each acre (J. Knapp, University of Florida Extension Service, Gainesville, personal communication, December 1978~. The additional application cost is assumed to be $10.00 per acre- treatment (the S-PBA estimates the application cost for an entire additional treatment to be $19.62 per acre).

204 REGULATING PESTICIDES If we assume that loss of chlorobenzilate would result in one additional dicofol treatment on all acreage currently harboring noneco- nomic infestations of snow scale (an assumption that would produce a maximum-plausible estimate of the benefits of chlorobenzilate use), then the additional scalicide costs would be about $3 million per year. For a minimum-plausible estimate of chlorobenzilate benefits, we adopt the assumption that chlorobenzilate's role in the IPM program can be assumed by an alternative (such as oil), thereby mitigating the snow- scale complication. Under this assumption there would be no cost effects peculiar to the IPM program, so assessment of any cost increases related to the use of higher cost alternatives can appropriately be included in a subsequent section dealing with non- cost ejects in Florida (and other states). In summary, the Committee's appraisal of the effect of disallowing chlorobenzilate on the cost of combatting snow-scale infestation is in the range $~$3 million a year, far lower than oPP's estimate of $50 million. Bud-Mite Control on California Lemons According to the USDA-ATR (USDA 1977b), there are about 41,000 lemon acres in California that could require treatment for bud mites. However, in any given year only about 5,000 acres require treatment with chlorobenzilate, with the acres being treated possibly changing each year. Loss of chlorobenzilate would leave oil as the only registered alternative for treatment of bud mites on California lemons. Dr. Glen Carman, a University of California entomologist, reports in the USDA-ATR (USDA 1977b:2, California section) that " . . . much of the lemon acreage in the citrus bud mite areas will receive one petroleum oil treatment a year for other pest control purposes." He further notes that "dependence upon petroleum oil sprays for the control of citrus bud mite would be expected to result in the mandatory use of one petroleum oil treatment each year in all bud-mite-infested properties and both a spring and a fall treatment in some localities or during some seasons." As a maximum-plausible case assumption, we accept oPP's assessment that all 41,000 acres would require one mandatory treatment with oil for bud-mite control and that 27,500 of those acres would require two treatments with oil (USDA 1977b). However, we also allow for the fact that virtually all of these acres already receive at least one treatment of oil for other purposes. Consequently, loss of chlorobenzilate (under these assumptions) would lead to 27,500 additional acre-treatments with oil. On the basis of information provided by Carman in the USDA-ATR, material costs are taken as being $33.02 per acre-treatment (26 gal/treatment at $1.27/gal). We assume that this oil can be included in

Application to Chlorobenzilate 205 spray mixes that would be applied in any event; consequently, application costs for the additional oil treatments are assumed to be zero. In fact, Carman notes that this additional oil treatment would probably supplant an existing spring (nonchlorobenzilate) treatment for red mites, since oil provides some control for this type of mite (Dr. Glen Carman, University of California, Riverside, personal communication, November 1978). This set of maximum-plausible case assumptions implies that loss of chlorobenzilate for bud-m~te control on California lemons would occasion increases in pest control costs equal to $870,000 ($910,000 for oil less the $40,000 currently spent on chlorobenzilate). Of course, the estimate may overstate the costs, because it fails to account for any cost savings occasioned by the substitution of oil for chlorobenzilate in treating red mites. To make m~nimum-plausible estimates of the benefits of chlorobenzi- late use, we assume that each chlorobenzilate treatment can be replaced with a single treatment of oil so that total acre-treatments remain at 5,000. In this instance, the loss of chlorobenzilate would occasion additional treatment costs of $125,000. oPP' using somewhat different assumptions, estimated the increase in pest control costs at $4.8 million a year. Other Cost Effects In this section we estimate the changes that loss of chlorobenzilate would occasion in the cost of controlling citrus pests in Arizona, Florida, and Texas. The discussion concerning Florida citrus focuses on those cost ejects that would be independent of the IPM program. We proceed in this section by considering the Committee's assumptions and the types of information needed to implement the measure of APC. 1. The list of K spray mixes containing chlorobenzilate that are currently used on citrus is taken from the Doane survey (Luttner 1977a). 2. The pest control methods (in this instance, they are all alternatives) that would probably be more widely used following a loss of chloroben- zilate are selected by merely adopting the list of alternatives (and the selection criteria) employed in oPP's analysis. The various alternatives to chlorobenzilate are listed in Table 7.12. 3. Information concerning the average number of treatments per year for chlorobenzilate and the alternatives is also presented in Table 7.12. The data in the table reveal that contrary to oPP's assumption the average number of treatments with chlorobenzilate mixtures may diner from the treatment rate for the alternatives. Consequently, in estimating

206 Cal .~ Ct o . ~ .G ~ :~ o Ad a, Ct Cal 3 ~ I:: an - U' a, :O . o ~ To _ ~ ~ . . . ~ _ _ ~ o ~ ~ ~ ~ ~ _ _ _ _ _ ~ ~ o . . . - %~% 1_ I_ . . . _ _ _ ~ __ _% K ma, - US US O ° E _ ~ s ° =° ~ ° .= ~ o ,,, `:, ~ In, - ~0 ~ ~

207 ~ So rid ~ ~ ~ I~ ° °° ~ ~ ° 1 1 _ ~ ~ ~ ~, , ~4 ~ oo I '') ° ~ ° ~ ° 1 1 Hi ~ Hi ~ ~ 1 1 -. %~0 ~ >i U. an x .§ C C ~ ~ ~_ ~ ~ ~ ~ ~ C .~ ~ ~X ~ ~ U. . am - a, U. .O : - o ~ _4 CL ~ ~ Cat o ~ i_ C ~ _ . . d) a, ~ ° o

208 TABLE 7.13 Acres Grown and Base Acres Treated With Chlorobenzilate, by State and Type of Citrus REGULATING PESTICIDES Type of Citrus Grapefruit Lemons Oranges AcresAcres Acres Acres AcresAcres State GrownTreated Grown Treated GrownTreated Arizona 9,7000 19,200 2,000 22,0000 California 20,7000 52,800 5,000 180,7000 Florida 132,000107,000 7,600 6,000 640,000416,000 Texas 42,00022,000 100 0 34,00015,000 Source: Doane Specialty Crop Studies for 1972, 1973, and 1974 (reported In Luttner (1977a)) and U.S. Department of Commerce (1978, Vol. II, Part 6, pp. 118-119). Data on lemon acreage in all states and grapefruit and orange acreage in Arizona and California are from U.S. Department of Commerce (1978). cost erects of a chlorobenzilate cancellation, we examine the implica- tions of replacing oPP's measure of treatment rates (no difference between the chlorobenzilate mixes and the alternatives) with the measures provided in Table 7.12. 4. Following oPP's lead, our estimate of base acres treated (Ak) with any one of the chlorobenzilate mixes is taken from the Doane report. Table 7.13 presents information on base acres treated annually with chlorobenzilate, by state and type of citrus. To provide some perspective on the relative extent to which chlorobenzilate is used, Table 7.13 also reports acres grown for each type of citrus. 5. In this part of the analysis, we assume as did oPP that total acres requiring miticide treatments would not be affected by loss of chloroben- zilate; i.e., I:j AAj-Ik Ak . The assumption is implicitly relaxed in the ~. , subsequent analysis that allows for price and quantity responses to whatever higher production costs a chlorobenzilate cancellation would occasion. 6. Our estimates of the acre-treatments for the chlorobenzilate mixtures are identical to those employed by oPP. 7. For part of the following analysis we accept oPP's assumption that the change in total acre-treatments for the alternatives will equal the present total acre-treatments with chlorobenzilate (i.e., /~:5jAj Tj = I:kAk Tk). However, we also investigate the implications of allowing differences in the treatment rates (Tj and Tk) to affect the acre-treatment estimates.

Application to Chlorobenzilate 209 8. The anticipated change in acre-treatments for any one alternative is estimated through a two-step procedure. First, the acres treated with chlorobenzilate are divided among the alternatives in accordance with their current relative importance. That is, the increase in base acres treated with thejth alternative is estimated by l\Aj = pj teak where pi is the ratio of (1) base acres currently treated with the jth alternative to (2) base acres currently treated with all the relevant chlorobenzilate alternatives. We also investigate the implications of a maximum-plausible benefits assumption that loss of chlorobenzilate would lead to adoption of only the more expensive alternatives. (We explain this assumption more fully below.) The second step in the procedure is to translate these acreage estimates into estimates of acre-treatments. For this task, we use the alternative assumptions mentioned above concerning the average number of treatments per year. This procedure diners greatly from the approach employed by oPP (involving an equal division of acre-treatments among the alternatives). The implications of these different approaches to estimating the change in acre-treatments for the alternatives is illustrated in Table 7.14 for the case of Florida oranges. (See Appendix E for data on other use sites.) 9. The estimates of per acre material costs (MC) in the Doane miticide report appear to be unreliable (Luttner 1977a:48~. Consequently, for chlorobenzilate and the major alternatives, we derive independent estimates of MC using information on product prices and application rates recommended in the state citrus guides.5 For the various spray mixtures and a few of the alternatives, we accept the MC estimates used by oPP. Table 7.15 presents the estimates of MC per acre that are used in the subsequent analysis. 10. Since growers usually spray their citrus groves a number of times in a season, it seems reasonable to assume that any regulatory decision concerning chlorobenzilate would not affect the number of times growers sprayed their groves. That is, in connection with the chlorobenzilate analysis, application costs can be taken as fixed an assumption generally followed by oPP. The Committee's estimates of increased pest control costs yielded by the various assumptions discussed above are presented in Table 7.16, together with oPP's estimates of the same quantities. Depending upon the

210 REGULATING PESTICIDES TABLE 7.14 Estimates of Acre-Treatments of Chlorobenzilate Alternatives: Florida Oranges Allocations Based on Current Use Patterns (2) (3) (4) (1) OPP's Estimate Using OPP's Estimate Using Doane's Alternative Estimate Treatment Rates Treatment Rates Dicofol Ethion Ethion/oil Sulfur Oil 103,000 103,000 103,000 103,000 103,000 10,300 32,400 127,200 319,300 25,800 (2.0%) (6 3%) (24.7%) (62.0%) (5.0%) 8,300 28,000 102,800 309,500 20,800 (1.8%) (6.1%) (21.9%) (65.8%) (4.4%) Total 515,000 515,000 (100 %) 470,200 (100 %) Note: Column (2) assumes an equal allocation of chlorobenzilate's base acres among the alternatives; columns (3) and (4) allocate the base acres in accordance with the current rela- tive importance of the various alternatives. Source: Derived from Luttner (1977a). set of assumptions selected, we estimate the increase in pest control costs following loss of chlorobenzilate to be somewhere between $2.4 million and $9.2 million (the range represents a 90 percent confidence interval; see Appendix F). Even under the maximum-plausible benefits assump- tions, our estimate of the control cost savings from chlorobenzilate falls far short of oPP's estimate of $57,629,000. The fundamental reason for the divergent estimates lies in the assumptions concerning scalicide treatments in Florida and bud-mite treatments in California (see Table 7.11). ECONOMIC EVALUATION OF THE BENEFITS OF CHLOROBENZILATE This section presents an appraisal of the combined effect of yield, quality, and pest control cost changes within the context of conventional benefit-cost analysis. That is, it attempts both to measure the extent to which the use of chlorobenzilate occasions "real" gains i.e., enhances society's opportunities for production or consumption and to assess the distribution of gains and losses (including those of a purely pecuniary nature). The availability of effective substitutes (in addition to other reasons outlined in preceding sections) argues against attributing significant yield and quality ejects to the use of chlorobenzilate.

Application to Chlorobenzilate 211 TABLE 7.15 Unit and Per-Acre Material Costs for Chlorobenzilate and Alternatives, by State Per-Acre Material Costs, by State Unit Cali Miticide Cost Arizona fornia Florida Texas Chlorobenzilate 4E Chlorobenzilate mixtures Carbophenothion 4E Dicofol 4MF Ethion 4E Oil 97 percent Sulfur 95 percent Carbophenothion/ oil Dicofol/oil Dicofol/sulfur Ethion/oil Ethion/sulfur $ 15.60/gal $ 7.80 $ 7.80 $ 4.88 $ 4.18b $ 13.50/gal $ 18.24/gal $11.40 $ 11.44/gal $ 1.02/gal $113/ton $33.02e $ 4.95 $ 6.83 $ 1 l.soc $ 8.78d $13.68 $ 8.93 $11.44 $ 7.65 $ 2.83 $ 14.78d $18.24 $16.50f $15.05g $15.81d $ 1o.ood a Average of the material cost for all of the mixtures. b This cost applies to Florida oranges (170,000 acre-treatments at an average of $4.18 per treatment); the cost for Florida grapefruit is $6.71 per acre (32,000 acre-treatments). C This estimate applies to Texas oranges (10,000 acre-treatments); the cost for Texas grapefruit is $17.50 (16,000 acre-treatments). Doane reports estimates of acre-treatment costs for Chlorobenzilate mixtures on Texas oranges that average to $4.81, an unbelievably small figure since Chlorobenzilate alone is estimated to cost $6.83 per acre-treatment (Luttner 1977a) . The Doane information on grapefruit suggests that these mixtures exceed the cost of Chlorobenzilate alone by as much as 70 percent; thus, we estimate the cost of these mixtures on Texas oranges to about $11.50 ($6.83 x 1.7 = $11.61). d Estimates taken from EPA (Luttner 1977a, pp. 50-52). e The price used for narrow-range type oil in California is $ 1.27/gal (USDA 1977b) . f Estimated by summing the costs of dicofol and sulfur. g Based on OPP's assumption that ethion will be applied with 6 gallons of oil ($8.93 + $6.10 = $15.03). Consequently, we need consider only the treatment-cost implications of Chlorobenzilate use. As suggested in Chapter 5, conventional benefit-cost analysis defines the benefit of continued Chlorobenzilate use as the sum of (1) the value of productive resources saved due to its use and (2) the value of any additional output it allows. In technical terms, the correct measure of benefits is the sum of the changes in consumers' and producers' surpluses

212 ~ U. o o ._ Ct - Cal Ct on ._ o - o V, o - o so o u U. ._ U. Cal Ct o Cal a' Ct ._ _ ~ _ ._ o Cal ~ ~ '5' _ ~ o . ~ ~ o m ~ [_ _' :3 ._ Cal G sit m .o a ~ . V, ~ ~ ~ U. 0 ~ U. ~ a Ce ~ _ Cal ~ I:: ~ 0 ~ Ct ~ ~ _ 0 ~ =i ~ ~ c, ca U. ~ ~ O a LLl ~ ~ a' Cal a _ 00 ~9 69 <5~ ~ ~00 ~ - ~C ~ ~ ~ r', V ~r 00 ~ _ 69 69 6 ~6 - 69 O O O O t- O - 1 ~ 00 ~ - ^ 6~ 6 ~ 6~3 6~ - o 00 6e o ~4 6~3 ~ 1 o 8 0 - ^ 69 6~ e9 O O O 69 6f} o 1_ g O r OC) ~ C-l C- o9 ~ 6~ 6~ O O O O e9 e9 e9 O O O O l_ r~ _ 69 ~ ~9 O O O 6~ e9 69 1 O O O a~ ~ O r - ~ 6~ e9 1 ~ ~ ~ ~ ~ ~ ~ O ~ ~ ~ c E E E ~ ~ 3 em ~ ~ E ~=

213 o 6~ o o _ so He o o _ ~ o o He 69 - ._ ~ Ho fig o o cr - o To o - 6~ o 6~ o 69 lo o 69 o - 6~ 1 ~ ~ ~ 8 ~ ·= =~] .^ ~ ~ _ ~ ~ ~ o 8 ~% $ - Co; I o ~ R.o .^ ~ --hi ~ §80._ - 8~? ] O ~ ~ G 3 ~ T ~ ~ 3 3~_ a-y ~ ~ ,- 3 5 ~ ~ ~ ~ ~ ~ ~ ~ 8 ~ A JO C 3 o i Pa AL by To ~- 9 8 ~ E s s ~ ~ ~ ~-~ ~ ~ ~ ~ ~ ~6. Cat ~q, ~

214 REGULATING PESTICIDES occasioned by the use of chlorobenzilate. In principle, measurement of these benefit components requires reliable estimates of (or reasonable assumptions about) market demand and supply functions (although, as we note below, the estimates of treatment-cost ejects developed in the preceding section probably provide a close approximation to the theoretically correct measure of the benefits of chlorobenzilate). Like- wise, an analysis of the distributional ejects of chlorobenzilate use (e.g., price increases to consumers) requires some information about market supply and demand conditions. Demand and Supply Elasticities Numerous estimates of demand elasticities for different types of citrus and citrus products are available (George and King 1971, Riggan 1965, Tilley 1977, U.S. EPA/USDA 1978, Ward and Kilmer 1978~. Unfortunate- ly, the estimates vary considerably across studies. The demand elasticity estimates range from 0.68 to 4.98 for fresh Florida oranges, and from .71 to -3.06 for processed products made from Florida oranges (e.g., frozen concentrate), although most of these estimates fall between 0.7 and -2.0. The one available estimated demand elasticity for fresh Texas oranges exceeds -31.0. Estimates for fresh Florida and Texas grapefruits center around 4.0, although demand elasticities for all grapefruit (fresh and processed) center around -1.0. We were unable to obtain any demand elasticity estimates for lemons. As a result of the obvious uncertainty about the actual magnitude of the price elasticities of demand for the various types of citrus and citrus products, the subsequent analysis evaluates (with two exceptions) the efficiency and distributional implications of chlorobenzilate use under the alternative assumptions that the price elasticities of demand equal either 0.7 or -2.5. The exceptions to this procedure are associated with evaluating impacts on Texas oranges (both fresh and processed) and Florida lemons. Texas orange growers account for about 2 percent of the total U.S. production, and Florida lemon growers provide for no more than 6 percent of the total lemon market. Producers supplying such relatively small portions of a market almost surely confront highly price- elastic demands for their product (recall that the estimated price elasticity of demand for fresh Texas oranges was around -31.0~. Consequently, the analysis centering on Texas oranges and Florida lemons assumes that the price for those products is electively demand- determined (that is, the price elasticity of demand is assumed to be infinite or, at least, very large). Estimates of the price elasticity of supply for the various types of citrus

Application to Chlorobenzilate are not, to our knowledge, available. However, previous studies of short- run supply elasticities for certain agricultural commodities (e.g., see Tomek and Robinson 1972) suggest that the short-run supply elasticity for citrus probably lies around 0.2 and almost surely does not exceed 0.5. The subsequent analysis uses two alternative values for the price elasticity of supply for each of the various types of citrus, namely, zero (the usual oPP assumption) and 0.5. 215 Analysis In addition to the assumptions concerning demand and supply elastici- ties, a number of other important assumptions will be invoked in the benefit assessment in this section. First, the demand and supply functions are assumed to be linear (strictly speaking, they need be linear only in the "neighborhood" of the equilibrium values for price and quantity). The assumption facilitates conversion of the elasticity measures into slope and intercept "estimates" for the demand and supply curves. Second, it is assumed that the average of prices and quantities for the 197 ~ 1975 and the 1975-1976 seasons are short-run equilibrium values. The actual price-quantity data used in the analysis are presented in Table 7.17. Third, the increase in variable costs per unit of output for the industry (that is, the upward shift in the industry supply curve) is assumed to equal (l`PC- initial industry output). For instance, in the case of fresh Florida oranges, unit costs are estimated to rise initially under the maximum-plausible benefits assumptions-by about 4 cents per box following the loss of chlorobenzilate (i.e., $554,000 in higher pest control costs divided by 12,564,000 boxes of fresh oranges). The shift in the supply curve is assumed to leave the slope of the curve unchanged. The results of the benefit assessment by type of citrus are reported in Tables 7.18-7.23. Table 7.24 presents estimates of the aggregate ejects. These estimates are based upon the maximum-plausible benefits assump- tions discussed in the preceding section. The corresponding oPP estimates are also reported to facilitate comparisons. Note that our estimates are substantially lower than oPP's; virtually the entire difference arises from our differing appraisals of the eject of disallowing chlorobenzilate on the cost of IPM. The estimates in these tables that correspond to the zero-elasticity-of- supply case are identical to the estimated maximum-plausible increases in pest control costs following a loss of chlorobenzilate. A perfectly inelastic supply curve means that growers would (as indicated) absorb all

216 REGULATING PESTICIDES TABLE 7.17 Average Price and Quantity Data for Citrus over 1974-1975 and 1975-1976 Seasons Fresh Processed Price Quantity Price Quantity State (per box) (boxes) (per box) (boxes) Grapefruit F10r~dla $2.77 19,583,000 $0.76 27,267,000 Texas $2.14 5,960,000 $0.75 3,040,000 Other 4,818,000 - 4,908,000 Total U.S. 30,361,000 35,215,000 Lemons Arizona $4.39a 4,810,000 California $4.39a 18,800,000 Other 1,790,000 Total U.S. - -5,400,000 Oranges Florida $2.18 12,564,000 $1.66 164,689,000 Texas $1.74 2,745,000 $1.18 2,575,000 Other - 36,590,000 21,135,000 Total U.S. 51,899,000 188,399,000 a The pace data are not specific to a state; it is assumed that Arizona and California prices are the sane. Source: For oranges and grapefruit, price data are from Ward and Kilmer (1978); quantity data are from Growers Administrative Committee (1977). For lemons, price data are from U.S. EPA/USDA (1978); quantity data are based on information in USDA (1977a) and the U.S. Department of Commerce (1978). cost increases and continue supplying the same quantity that was provided before the cost increase. It is notable that allowing for some price responsiveness in supply hardly affects the estimate of net benefits from the use of chlorobenzilate. Thus, oPP's neglect of supply responses was a justifiable approximation for the purpose of estimating net aggregate benefits. However, Tables 7.18-7.24 also reveal that demand and supply conditions have to be taken into account in estimating the distributional implications of a chlorobenzilate cancellation (or other action). If short- run citrus supply is responsive to changes in price and variable production costs, then disallowing chlorobenzilate would lead to slightly higher prices (around 1 - cents per box in most cases) for citrus and a

Application to Chlorobenzilate TABLE 7. IS Alternative Estimates of Forgone Benefits Under Maximum-Plausible-Case Assumptions: Arizona and California Lemons OPP Estimate: -$4,821,000 Committee Estimates Alternative Demand Conditions Ed = -0.7 Users Nonusers Consumers Alternative Supply Conditions Es =o -$875,000 o o ES = 0 5 -$761,000 + 241,000 - 354,000 Net effect-$875,000 -$874,000 En = -2.5 Users- $ 875,000 - $ 784,000 Nonusers0 + 52,000 Consumers0 - 141,000 Net effect -$875,000 -$873,000 Note: Ed is demand elasticity; Es is supply elasticity. 217 slight decline in the quantity supplied (not only by users of chlorobenzi- late, but by the entire industry). These changes have the eject of transferring some of the increased pest control costs from users of chlorobenzilate to consumers. In addition, the higher prices confer large windfall gains on the nonusers of chlorobenzilate, again at the consum- ers' expense. This transfer between consumers and nonusing growers is the predominant reason why the total losses (to consumers and users) reported in Tables 7.18-7.24 exceed the net eject of a chlorobenzilate cancellation. Some additional suggestive evidence on the probable distributional ejects of a chlorobenzilate cancellation is provided in Table 7.25, which presents information about size distribution of orange farms in Califor- nia and Florida. If we assume that these data are roughly representative of citrus farms generally, it appears that much of the burden imposed on chlorobenzilate users would fall on the relatively large operations (i.e., those with more than 100 acres-a category for which average annual sales per farm exceed $300,000~. Nearly three fourths of the burden on users is expected to fall on Florida orange growers, and the data in Table

218 REGULATING PESTICIDES TABLE 7.19 Alternative Estimates of Forgone Benefits Under Maximum-Plausible-Case Assumptions: Fresh Florida Oranges OPP Estimate: -$3,057,500 Committee Estimates Alternative Demand Conditions Ej = -0.7 Users Nonusers Consumers Net effect Ed = -2.5 Users Nonusers Consumers Net effect Es =o -$554,000 o o -$554,000 -$554,000 o o -$554~000 Alternative Supply Conditions ES = 0 5 -$406,000 + 79,000 - 226,000 -$553,000 -$496,000 + 31,000 - 88,000 -$553,000 Note: Ed is demand elasticity; ES is supply elasticity. 7.25 indicate that around 77 percent of those costs will be borne by the larger operations. COMPARISON OF REGULATORY OPIIONS The preceding sections of this chapter have presented detailed analyses of the risks and benefits associated with uses of chlorobenzilate. This section will summarize the principal findings of those analyses and will apply them to the assessment of the costs (forgone benefits) and the reduction in risk that can be anticipated from a number of regulatory options. As mentioned in the risk section of this chapter, five options are being considered. They are: A. Continue registration of all uses. B. Cancel all noncitrus uses. C. Continue registration of chlorobenzilate use on citrus and amend the terms and conditions of registration to require protective clothing and respirators; cancel all other uses. D. Cancel chlorobenzilate use on citrus to take eject after 5 years, and in the interim apply option C. E. Cancel all uses.

Application to Chlorobenzilate TABLE 7.20 Alternative Estimates of Forgone Benefits Under Maximum-Plausible-Case Assumptions: Processed Florida Oranges OPP Estimate: -$40,006,500 Committee Estimates Alternative Demand Conditions Ed = -0.7 Users Nonusers Consumers Net effect E,' = -2.5 Users Nonusers Consumers Alternative Supply Conditions Es JO -$7,246,000 o o -$7,246,000 -$7,246,000 o o ES = 0 5 -$5,296,000 + 1,030,000 - 2,953,000 -$7,219,000 -$6,563,000 + 339,000 983,000 Net effect -$7,246,000 -$7,207,000 Note: E`, is demand elasticity; ES is supply elasticity. TABLE 7.21 Alternative Estimates of Forgone Benefits. Under Maximum-Plausible-Case Assumptions: Florida Lemons OPP Estimate: -$432,000 Committee Estimates Alternative Demand Conditions Alternative Supply Conditions E&= - ~Es=0 Es=O.S Users -$95,000 -$94,000 Nonusers O O Consumers O O Net effect -$95,000 -$94,000 Note: En is demand elasticity; ES is supply elasticity. 219

220 REGULATING PESTICIDES TABLE 7.22 Alternative Estimates of Forgone Benefits Under Maximum-Plausible-Case Assumptions: Florida and Texas Grapefruit OPP Estimate: -$9,081,000 Committee Estimates Alternative Demand Conditions Ed - -07 Users Nonusers Consumers Net effect Ed = -2.5 Users Nonusers Consumers Es =o -$1,942,000 o To -$ 1,942,000 -$1,942,000 o o Alternative Supply Conditions ES = 0 5 -$ 1,401,000 + 300,000 835,000 -$ 1,936,000 -$1,724,000 + 128,000 334,000 Net effect -$ 1,942,000 -$ 1,930,000 Note: Ed is demand elasticity; Es is supply elasticity. TABLE 7.23 Alternative Estimates of Forgone Benefits Under Maximum-Plausible-Case Assumptions: Texas Oranges OPP Estimate: -$230,000 Committee Estimates Alternative Demand Conditions Ed= _ ~ Users Nonusers Consumers Net effect Es =o -$ 160,000 o o -$ 160,000 Alternative Supply Conditions ES = 0 5 -$159,000 o o -$ 159,000 Note: Ed is demand elasticity; ES is supply elasticity.

Application to Chlorobenzilate TABLE 7.24 Alternative Estimates of Forgone Benefits Under Maximum-Plausible-Case Assumptions: All Citrus OPP Estimate: -$57,629,000 Committee Estimates Alternative Demand Conditions Alternative Supply Conditions Ed= -0.7Or- ~Es=0 Es=0.5 Users-$9,200,000 -$6,869,000 Nonusers0 + 1,397,000 Consumers0 - 3,696,000 Net effect-$9,200,000 -$9,168,000 Ed= -2.5Or- ~ Users-$9,200,000 -$8,310,000 Nonusers0 + 465,000 Consumers0 - 1,308,000 Net effect-$9,200,000 -$9,153,000 Note: Ed is demand elasticity; ES is supply elasticity. TABLE 7.25 Farm Size Distribution: California and Florida Oranges 221 Size Category 0.1 -24.9 25-49.9 Acres Acres 50-99.9 100 or Acres More Acres Percentage of total state orange production attributable to each category California 15.4 16.7 21.2 46.7 Florida 6.7 7.1 9.0 77.2 Average annual sales per farm California $12,000 $40,500 $82,200 $305,200 Florida - $ 7,600 $20,700 $43,300 $336,000 Source: All estimates based on data from U.S. Department of Commerce (1978, p. 118).

222 REGULATING PESTICIDES These options are basically the same as those considered by oPP (see Chapter 6), although the Committee lists two fewer. (It should be noted, however, that the Comm~ttee's labeling of the options diners somewhat from oPP's. Thus, the Committee's Option B. for example, does not correspond to oPP's Option B.) The PD 3 for chlorobenzilate (U.S. EPA 1978a) discusses the possibility of cancelling chlorobenzilate use only in Arizona (oPP's Option F). However, the option was eventually dropped from consideration by oPP. Thus, the Committee has chosen not to treat this potential option separately from the option that would cancel the uses of chlorobenzilate in all states. The PD 3 also lists as a regulatory option the prohibition of chlorobenzilate-treated citrus pulp as cattle feed (oPP's Option G). The selection of this option would amount to a de facto cancellation of chlorobenzilate use on citrus. Apparently, the profits from the continued sale of citrus pulp for cattle feed would more than offset the losses that would be incurred from (voluntary) withdrawal of chlorobenzilate (U.S. EPA 1978a). In addition, in Options C through G. the PD 3 discusses allowing the use of enclosed cabs in lieu of protective clothing and respirators when applying chlorobenzilate. Growers would find this option to be expensive since it would require substantial modifications to the existing fleet of tractors. The most likely response of growers would be to substitute other chemicals for chlorobenzilate, rather than incur the capital costs associated with this regulatory option (U.S. EPA 1978a). Accordingly, the Committee has not considered the prohibition of citrus pulp as cattle feed or the use of enclosed cabs during application separately from the option to cancel all uses of chlorobenzilate (the Committee's Option E). The major consequences to be expected from these options are summarized in Table 7.26 and portrayed graphically in Figures 7.2 through 7.5. Because of the substantial uncertainties surrounding all the estimates, more than one value is presented in each cell of the table (see Appendix F). In the cost row the three numbers for each option (except Option B) show the range from the lowest cost that the Committee believes to be at all conceivable to the highest one. (Recall that for Option B cancel noncitrus uses oPP's estimate is used and therefore no range is shown.) The cost measures presented in Table 7.26 and Figures 7.2 through 7.5 represent estimates of the discounted present value of the future benefits from chlorobenzilate that would be forgone as a result of adopting a particular option. These present values were calculated under the assumptions that the remaining economic lifetime of chlorbenzilate is 10 years and that the appropriate discount rate is 7 percent (see Chapter 5 for further discussion).6

Application to Chlorobenzilate 223 In the dose rows, the values are to be interpreted differently. The first number indicates the Committee's best judgment of the lifetime dose likely to be received by the relevant population group in the event of the indicated regulatory option. The second number is a worst-case or maximum-plausible exposure estimate, that is, the dose that the Committee believes might conceivably be attained over a lifetime but is highly unlikely to be exceeded. In every case, the recorded dose is the sum of the lifetime dose of chlorobenzilate plus the equivalent dose of dicofol- one of the most risky alternatives to chlorobenzilate and one that is expected to replace 2-10 percent of chlorobenzilate under certain regulations that could be anticipated under the regulatory option shown. As one moves across Table 7.26 from Option A to Option E, the costs, or value of forgone benefits, increase. The probable lifetime doses decrease for all population groups, or at least do not increase, as one moves from continuing registration to increasingly stringent regulations. But the maximum-plausible doses, it should be noted, actually increase for the Florida and U.S. populations as one moves from Option D to Option E. This averse behavior results from the increased use of dicofol as the use of chlorobenzilate is more and more closely restricted. Thus, while we do not expect that the lifetime dose received by the Florida population will be any greater under Option E than under Option D, we think that, because of the larger amount of dicofol whose use will be induced by Option E, the dose might be as much as 23 percent greater. The same data that are in Table 7.26 are presented in Figures 7.2 through 7.5 in a form that is easier to assimilate. In addition, the figures show dashed horizontal lines introducing a relevant comparison com- pound, heptachlor. These lines are labeled "chlorobenzilate-equivalent of heptachlor" and are drawn at the mean lifetime dose of heptachlor estimated for the total U.S. population and (where relevant) the statistically calculated dose that the most highly exposed 1 percent of the U.S. population was receiving before heptachlor was suspended, trans- lated into chlorobenzilate-equivalent doses using the CAN'S of heptachlor and chlorobenzilate as described in Chapter 4. It should be noted that in order to keep the graphs reasonably simple, the uncertainty in the CAI'S (see Figure 7.1) has not been represented. Ideally, they should be, as described in Chapter 6. The Committee found two reports that estimated the intake of heptachlor by the average member of the U.S. population: one by Nisbet (1976) and the other by CAG (19774. (In fact, Nisbet's report was prepared specifically for CAG.) Hence, there are two sets of heptachlor values in the figures. Nisbet estimated that at the time of the heptachlor

224 ·4= e~ oo o .~ .> (4o - U, o o U. .o Ct a C~ ·4 - .~ ~_ .> ~5 o ._ o o Ct a~ U' o0 00 00 0 r~ ~ o ~o o _ _ ~ _ . . ~ oo ~ oo _ ~o o _ ~ 1 0 1 _ o x Ir} X _ ~D o _ 1 0 1 _ o X X _ r~ o _ 1 0 1 _ o x x o ~ ~ ~ ~ ~ o _ ~i r ~O O O 1 r~ 0 1 o x oo C ~(~) X ~ ~ o _ o o o 1 0 1 _ o oo ~- , X ~ ~ o . . . . o o o o o o ~ - ~ ^ - o U' o _` ~o ~o ~ ~_ CQ o ~ X ~ D~ ~ ~ ·8 =~ ~D~ e ~3, .ii ~ ~ ~ E ~ ~, E O ~ ~ ~ 3 ~ ~; v ~

225 1 0 1 _ o x ^ ~ o o ~ o o 1 1 o o _ ~ X X _ o ~4 ~ o o 1 1 o o _ _ X X oo _ ~ o o 1 1 o o _ _ X X oo _ . . o o 1 1 o o _ _ X X oo _ o o ^ . - ~o a,, X X ~ 8 - ~ O O D ·~§ cn ~ ~ 5: o C) ao .> C) a~ o ·3 o 1 U' o o ·= ~o au ~ ,D,, . _ , _ :, ·6X D ._ _ .~ 8 D so C~ C) ~ .c D ~ C~ ·cq V, a,, ~-° -~ ,D 75 ~ ~ i ~ b e ~a O U, O ~ ~ ~-~ ~ ~ ~ O ~ a ~ a ~. a ~ a ~ ~ ~k ~ o o V, C~ o . · U-~ . O .= ao . ~. ~ O O.0 ~C~ =00 .b CL ~4_ ~Q ~.c ~ .5 ~ O 4_ ~ ao.G ~ _, .= CO U, . ~. ~ o ~·V, :- ~ ~o U) ·C) U,

226 0.20 ye ~O. ~ O o E - LL In o C) LU ~0.0003 U] 0.0002 ,0.0001 U] 1 _ o 1 1 1 1 1 _ REGULATING PESTICIDES Chlorobenzilate Equivalent of Heptachlor, Mean U.S. Lifetime Dose (Nisbet 1976) Chlorobenzilate Equivalent of Heptachlor, Mean U.S. - Lifetime Dose (CAG 1977) · · Florida Population (8 X 106) ). .( General U.S. Population (212 X 106) A B D A B D ' )~- E E OK 0 10 20 30 40 50 60 70 DISCOUNTED COST ($ million) FIGURE 7.2 Equivalent probable lifetime doses for the Florida and U.S. populations and ranges of discounted costs under five options for regulating chlorobenzilate; heptachlor comparison shown (see text for discussion). Source: Table 7.26 and text. suspension in 1975, the average member of the U.S. population was ingesting about 2.4 ,ug/day of heptachlor. At the time of the suspension, the limited economic lives of pesticides were not taken into account in estimating risk. Therefore, the Committee infers that the suspension was imposed in order to prevent the continuation of heptachlor intake over a typical lifetime of 70 years, an intake that would result in ingesting 0.00238 m moles heptachlor/kg of body weight/lifetime based on Nisbet's data. (The Administrator's order to suspend most heptachlor and chlordane uses was based on findings of widespread human exposure and the judgment, based on animal data, that the pesticides were carcinogenic. The imminent hazard posed by these findings was judged to outweigh the benefits of continued use.) The CAI for heptachlor is 19.4 (see the section on Reference Compounds in this chapter), or 72 times that of chlorobenzilate (0.27, based on Innes et al. 1969), so that on the basis of Nisbet's data the dose of heptachlor that was being received at the time of its suspension is

Application to Chlorobenzilate 1 .80 1.50 1.20 0.60 0.30 227 _ o . . , . ,, ,,,,,, ,,, > ,< ho: I... Chlorobenzilate Equivalent of Heptachlor, Highly Exposed 1%s ~ of U.S. Population (Nisbet 1976) Chlorobenzilate Equivalent of Heptachlor, Highly Exposed 1% of U.S. Population (CAG 1977) A B - C Chlorobenzilate Equivalent of Heptachlor, Mean U.S. n Lifetime Dose ~(Nisbet 1976) 1 . 1 1 1 1 , , - 0 10 20 30 40 50 60 70 DISCOUNTED COST ($ million) FIGURE 7.3 Equivalent probable lifetime doses for citrus ground applicators (700) and ranges of discounted costs under five options for regulating chlorobenzilate; Heptachlor comparison shown (see text for discussion). Source: Table 7.26 and text. equivalent to about 0.17 m moles chlorobenzilate/kg body weight as a probable lifetime exposure. Similarly CAG'S value of 1.2 ,ug/day as the estimate of Heptachlor ingested by the average member of the U.S. population is equivalent to about 0.084 m moles of chlorobenzi- late/kg/lifetime. The most highly exposed 1 percent of the population was estimated by Nisbet to receive ~10 times the mean daily exposure level. Therefore the Nisbet and CAG estimates are each multiplied by 6 and 10 to produce a range of estimates, shown in Figures 7.3 and 7.5, for the highly exposed groups. According to Nisbet, the highly exposed groups included children and breast-fed infants, freshwater fishermen and their families, and persons living near treated fields and in treated buildings. The mean cut-off equivalent dose of chlordane, the other comparison compound, is not shown; it is more than 3 times as great as the mean Heptachlor dose being received by the general U.S. population

228 - 0. 10 o E Al o C) LU ~ 0.003 he' - ~ 0.002 J ~ 0.001 a o . l _ i REGULATING PESTICIDES 0 20- Chlorobenzilate Equivalent of Hoptachlor, Mean U.S. Lifetime Dose (Nisbet 1976) _ _ , Chlorobenzilate Equivalent of Hoptachlor, M - n U.S. Lifetime Dose (CAG 1977) ~ - Florida Population (8 X 1 o6) 9. lC General U.S. Population (212 X 106) A \.,.~. E A \ C X>~ B K E OK 1 ' I 1 1 20 30 40 50 60 70 I I I ~ 0 10 DISCOUNTED COST {$ million) FIGURE 7.4 Equivalent maximum-plausible lifetime doses for the Florida and U.S. populations and ranges of discounted costs under five options for regulating chloroben~- late; heptachlor comparison shown (see text for discussion). Source: Table 7.26 and text. (based on an estimated lifetime exposure of 0.0081 m moles/kg [Nisbet 1976 and CAG 1977~. Recall that chlordane was also suspended and its major uses ultimately cancelled. Figures 7.2 and 7.3 present probable lifetime doses of Chlorobenzilate for each option, while Figures 7.4 and 7.5 show the maximum-plausible dose estimates. In each case, the first figure of the pair shows the general U.S. and the Florida populations while citrus applicators appear separately in the second. For the U.S. and Florida populations only the Chlorobenzilate equivalents of the mean heptachlor dose are graphed. This is because Chlorobenzilate doses under all five options are already below these values, making it unnecessary to add the higher chlorobenzi- late-equivalent of heptachlor lines. For the citrus applicators, both the mean and high-exposure level chlorobenzilate-equivalent doses of hetptachlor are shown. In the following discussion, the Committee has not attempted to

Application to Chlorobenzilate 1 a8or en ~ 1.50 o ~.~ 1.20 oh o 0.90 It '~ 0.60 of us > 0.30 a Lo 229 ~ _ ~tit:~.,~t,)~t~,It.~t $. 'I .... , Chlorobenzilate Equivalent of Heptachlor, Highly Exposed 1% ... ... .~; of U.S. Population (Nisbet 1g76) ................... ....... ., .... ~ . ,, ,, ,. ; ......... .................................................................................................... ..... , , , , i, . .... ............................... ' " "a"'.' "" " " "' '' '1 '' "''"' "' "' '' "'X X" "' " ~ '''X "' '''X ' "''' ' " ' ' "X" '' "' ' " """" "'''" . ~ X '2.' ~ ' ~ ~ go ' 5'' 5' ~ ~ ~ X X '' ' ' . ~ · ~ ~ ~ ~ ~ ~ ~ . ~ . ;~; ~,~, ..~; ;; ;~; j,` j,; A ;_, - ,,, . ~;~; ;,.;;;; . me, ..;~j;; ;; I, ^,,,~ ; I; i If,, A ~ ~.~... ~ ~- A Chlorobenzilate Equivalent of Heptachlor, Highly Exposed 1% Gil B ..~. of U.S. Population (CAG 1977) ..~. . _ ~2~ D , Chlorobenzilate Equivalent of Heptachlor, Mean U.S. Lifetime Dose (N'sbet 1976) Chlorobenzilate Equivalent of Heptachlor Mean U.S. Lifetime Dose (CAG 1977) , ~ , ~ O ~ 1 1 1 1 1 1 1 0 10 20 30 40 50 60 70 DISCOUNTED COST ($ million) FIGURE 7.5 Equivalent maximum-plausible lifetime doses for citrus ground applicators (700) and ranges of discounted costs under five options for regulating Chlorobenzilate Heptachlor comparison shown (see text for discussion). Source: Table 7.26 and text. analyze similarities or differences in the size or compositions of the populations exposed to Chlorobenzilate with those exposed to hepta- chlor. Such considerations may well be important. For example, is it more useful to compare a population of about 700 Chlorobenzilate applicators to a population of 212 x 106 average Americans exposed to Heptachlor or to a smaller population of breast-fed babies and young children receiving very high doses (from milk) for a short period of their life? This is a matter of judgment left to the Administrator. Furthermore, the economic impacts of the Heptachlor suspension are not discussed. Although EPA did conduct an economic impact assessment (U.S. EPA 1976a), the decisions to suspend and ultimately cancel most Heptachlor uses appear to have been based more on the risks involved than on the economic impacts, because alternative pesticides were predicted to be available and economic impacts were predicted to "be relatively minor in general and . . . (to) have no significant eject on production and

230 REGULATING PESTICIDES prices of agricultural commodities, retail food prices, and otherwise on the agricultural economy" (U.S. EPA 1978c: 12375~. It is clear from Figure 7.2 that even without regulation, the intake of chlorobenzilate to which the U.S. population and Florida residents are exposed is orders of magnitude less than the equivalent intake of heptachlor corresponding to the mean levels at which its use was forbidden. The Committee recognizes that it is possible that heptachlor would have been suspended even if its dosage rate had been substantially smaller than the estimated levels. It is even possible that heptachlor would have been suspended at a dose level one thousandth as great as the estimated mean level being experienced at the time of the suspension in 1975, in which case its intake would be about equal to the intake of chlorobenzilate without regulation in terms of carcinogenic activity. That is a matter for the Administrator's judgment. The chart does make clear that the probable doses of chlorobenzilate received by food consumers are some 3 orders of magnitude below equivalent doses of heptachlor to which they had been exposed. The contrast is even greater for chlordane (not shown). On the other hand, if we turn to Figure 7.3, it can be seen that citrus ground applicators are expected to receive nearly 4 times the mean chlorobenzilate equivalent of heptachlor cutoff for the general U.S. population under our Options A and B. about twice as much under Option C, a comparable amount under Option D, and about half as much under Option E. If, however, the probable chlorobenzilate doses to applicators are compared to the chlorobenzilate equivalent of heptachlor for the 1 percent of the U.S. population most highly exposed to heptachlor, all the options are expected to expose applicators to less than the high-level chlorobenzilate equivalent of heptachlor, although Options A and B are so close to the high-level cut-off based on CAG data, that they may be questionable. The main implications of Figures 7.2 and 7.3 appear to be that Options A and B expose ground applicators to levels of risk that are nearly comparable to those being experienced by the populations that were highly exposed to heptachlor. Options D and E expose chlorobenzi- late applicators to levels of risk roughly comparable to those that were experienced by the average member of the U.S. population in the heptachlor situation. Exposure of applicators under Option C falls in between. At the same time, the risks to the U.S. and Florida populations associated with the probable dietary exposures to chlorobenzilate appear to be too small to be at issue. It would obviously be useful to compare the doses of chlorobenzilate being received by the U.S. and Florida populations to chlorobenzilate equivalents of a previously regulated

Application to Chlorobenzilate comparison compound that were lower than the current chlorobenzilate exposure levels. In the case of chlorobenzilate, our bank of reference compounds (Table 4.3) was insufficient to supply such a comparison. We turn now to a consideration of the maximum-plausible dose levels of chlorobenzilate shown in Figures 7.4 and 7.5. Under all options, the highest doses that members of the U.S. population and Florida residents can plausibly be expected to receive by way of ingestion are of the order of 1 percent of the estimated chlorobenzilate-equivalent mean dose of heptachlor at the time that that pesticide was suspended (Figure 7.4~. Citrus ground applicators, on the other hand, may plausibly be exposed to doses greater than the mean chlorobenzilate equivalent of heptachlor under any of the options (Figure 7.5~. The excess is smallest under Option E, which requires immediate cancellation of chlorobenzilate use on citrus. Under Options A, B. and C, applicators may plausibly be exposed to doses comparable to the chlorobenzilate equivalent of heptachlor that the highly exposed members of the U.S. population were experiencing when heptachlor was cancelled. Options D and E may plausibly result in chlorobenzilate exposures to applicators that lie in between the mean chlorobenzilate-equivalent of heptachlor Put-on and the highly exposed. Ultimately, the choice among options depends upon a judgment as to whether the indicated reductions in the doses received by ground applicators of chlorobenzilate are sufficient to justify the extra costs associated with adopting the more stringent regulations. The critical figures to review are Figures 7.3 and 7.5 pertaining to the probable and maximum-plausible chlorobenzilate exposures to citrus ground applica- tors together with the cost information along the x-axes. For example, looking at the move from Option C to Option D, it can be seen from Figure 7.3 that the cost of reducing the lifetime exposure of about 700 applicators by about 0.1 m moles/kg/person is expected to be about $16 million. Recall that this would place applicators at a risk roughly comparable to that being experienced by the average member of the U.S. population when heptachlor was cancelled. Similar comparisons of incremental costs with incremental reductions in exposure and the concomitant risk implications as one moves form one regulatory option to the next will provide the scientific basis for making decisions about regulating pesticides. (It is interesting to note that had dicofol been submitted to RPAR together with chlorobenzilate, as would arise if the procedures recommended in Chapter 3 were adopted, the regulatory options would diner. For example, it is likely that they would include the option to cancel both chlorobenzilate and dicofol. This, in turn, might easily result in applicators' risks being reduced below those associated 231

232 REGULATING PESTICIDES with the mean chlorobenzilate-equivalent of heptachlor cotton. Of course, without having considered that option explicitly, it is difficult to predict what risks or costs from other substitute pesticides might enter the picture.) NOTES 1. oPP assumes that about 19 percent of the Chlorobenzilate acre-treatments would be replaced by dicofol (Luttner 1977a). However, as the benefit analysis in a subsequent portion of this chapter demonstrates, this assumption seems to have no factual basis and is not rigorously defended in the oPP benefit analysis. The range (2-10 percent) adopted by the Committee is admittedly arbitrary, but nevertheless is consistent with the treatment levels implied by current usage levels (see the Doane Specialty Crops Survey included in Luttner 1977a). Presently, dicofol accounts for only about 5 percent of the treatments on lemons and grapefruits and about 2 percent of the treatments on orange trees. In contrast, Chlorobenzilate presently accounts for about 13 percent of the acre-treatments on lemons and about one third of the acre-treatments on oranges and grapefruits. 2. oPP estimates that cancellation of Chlorobenzilate would increase annual pesticide costs on noncitrus crops by $194,000, with cotton accounting for $125,000 and fruits, nuts, and other crops accounting for the remaining $69,000 (U.S. EPA 1978a:67). The Committee does not question these estimates, largely because their magnitudes are so small relative to the estimated impacts on citrus growers and consumers. 3. The benefit assessment for Chlorobenzilate was performed before the joint USDA/EPA assessment procedure was initiated. Therefore, the Preliminary Benefit Analysis of Chlorobenzilate (Luttner 1977a) was an EPA product to which the USDA reacted by producing a USDA assessment team report (USDA 1977b). EPA'S supplement to the PBA (Luttner 1977b) responded to the USDA report. 4. The USDA assessment team was formed after oPP had completed the PBA (see note 2); consequently, information from the assessment team is used only in the Supplement to the PBA of Chlorobenzilate (Luttner 1977b). 5. Prices for chlorobenzilate, oil, sulfur, ethion, and dicofol (and various combina- tions of these miticides) were obtained from the USDA-ATR (USDA 1977b) and the Anderson- Muraro production budgets (Luttner 1977a). Whenever these two sources reported different prices for the same product, we used the average of the reported prices in our analysis. 6. The documented costs in Table 7.26 and Figures 7.2 through 7.5 are based upon the following estimates of Anne forgone benefits (over a Midyear period). The noncitrus uses of chlorbenzilate are estimated to yield annual benefits of $194,000 (see note 2 to this chapter). The benefits from citrus uses range from $2.4 to $9.2 million annually, with the probable-case estimate being $5.8 million (see section on Changes in Pest Control Costs of this chapter for derivation of these estimates). Finally, the cost of providing applicators with protective clothing and respirators is assumed to be $ 10~$500 per applicator per year for a total annual cost of $70,000 $350,000. Neither oPP (U.S. EPA 1978a) nor the Committee developed actual empirical measures of the costs of protective clothing and respirators.

Application to Chlorobenzilate REFERENCES 233 Allen, J.C. (1978) The Effect of Citrus Rust Mite Damage on Citrus Fruit Drop. University of Florida, Lake Alfred, Florida. (Unpublished) Allen, J.C. (1979) The effects of citrus rust mite damage on citrus fruit growth. Journal of Economic Entomology 72(V: 195-201. Allen, J.C. and J.H. Stamper (1979) The frequency distribution of citrus rust mite damage on citrus fruit. Journal of Economic Entomology 72(3):327-330. Bartsch, E., D. Eberle, K. Ramsteiner, A. Tomann, and M. Spindler (1971) The carbinole acaricides: chlorobenzilate and chloropropylate. Residue Reviews 39: 1-88. Brooks, R.F. (1977) Integrated control of Florida citrus pests. Citrus Industry 58(4):31, 34- 36. Brooks, R.F. and J.D. Whitney (1977) Citrus Snow Scale Control in Florida. Vol. 2, pages 427~31, Proceedings: I Congreso Mundial de Citricultura 1973. Murcia-Valencia, April 29-May 10, 1973. Murcia, Spain: Ministerio de Sciencia Consejo Superior des Investigaciones Scientifica, Centro de Adafologia y Biologia Aplicada del Segura. Burnam, W.L. (1977) First Draft of Chlorobenzilate Substitutes. Transmitted to J.B. Boyd, 6-22-79, Special Pesticide Review Division, Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C. (Unpublished) Bushong, C. (1977) Chlorobenzilate risk analysis, fish and wildlife. Memorandum to J.B. Boyd, oPP. August 8, 1977, Special Pesticide Review Division, Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C. (Unpublished) Carcinogen Assessment Group (1977) Risk Assessment of Chlordane and Heptachlor. Roy Albert, CAG Chairman, U.S. Environmental Protection Agency, Washington, D.C. 20460. (Unpublished) Carman, G.E., W.E. Westlake, and F.A. Gunther (1972) Potential residue problem associated with low volume sprays on citrus in California. Bulletin of Environmental Contamination and Toxicology 8:3845. Council on Environmental Quality (1976) Environmental Quality: The Seventh Annual Report of the Council on Environmental Quality. Stock no. 041-010-00031-2. Washing- ton, D.C.: U.S. Government Printing Offlce. C''mmings, J.G., M. Eidelman, V. Turner, D. Reed, K.T. Zee, and R.E. Cook (1967) Residues in poultry tissues from low level feeding of five chlorinated hydrocarbon insecticides to hens. Journal of the Association of Official Analytical Chemists 50(2):418-425. Dennis, J.D. (1977) Untitled, unpublished letter to F. Maxwell, University of Florida. August 15, 1977. (Available from Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C.) Federal Working Group on Pest Management (1974) Occupational Exposure to Pesticides. Report of the Task Group on Occupational Exposure to Pesticides, T.H. Milby, Chairman, to Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C. (Unpublished) Feldmann, R.J. and H.I. Maibach (1974) Percutaneous penetration of some pesticides and herbicides in man. Toxicology and Applied Pharmacology 28: 126-132. Fries, G.F. (1969) Comparative excretion and retention of DDT analogs by dai~y cows. Journal of Diary Science 52(11): 180~1805. George, P.S. and G.A. King (1971) Consumer Demand for Food Commodities in the United States With Projections for 1980. Monograph Series Number 26. Berkeley, Calif.: Giannini Foundation.

234 REGULATING PESTICIDES Griffiths, J.T. ana W.L. Thompson (1953) Reduced spray program for citrus for Canning plants in Florida. Journal of Economic Entomology 46:930-936. Growers Administrative Committee (1977) 1976-77 Season Annual Statistical Record. Lal~eland, Florida. Gunther, F.A. (1969) Insecticide residues in California citrus fruits and products. Residue Reviews 28: 1-119. Gunther, F.A., W.E. Westlake, and P.S. Jaglan (1968) Reported solubilities of 738 chemicals in water. Residue Reviews 20: 1-148. Gunther, F.A., Y. Inata, G.E. Carman, and C.A. Smith (1977) The citrus reentry problem. Residue Reviews 67: 1-139. Hassan, T.K. and C.O. Knowles (1969) Behavior of three Ci4-labeled benzilate acaricides when applied topically to soybean leaves. Journal of Economic Entomology 62(3):618- 619. Hayes, W.J., Jr. (1975) Toxicology of Pesticides. Baltimore: Williams and Wilkins Company. Horn, H.J., R.B. Bruce, and O.K. Paynter (1955) Toxicology of chlorobenzilate. Jounce of Agricultural and Food Chemistry 3(9):752-756. Innes, J.R.M., B.M. Ulland, M.G. Valerio, L. Petrucelli, L. Fishbein, E.R. Hart, A.J. Pallotta, R.R. Bates, H.L. Falk, J.J. Gart, M. Klein, I. Mitchell, and J. Peters (1969) Bioassay of pesticides and industrial chemicals for tumorigenicity in mice: A preliminary note. Journal of the National Cancer Institute 42: 1101-1114. Jeppson, L.R., M.J. Jesser, and J.O. Complin (1955) Control of mites on citrus with chlorobenzilate. Journal of Economic Entomology 48: 37~377. Kesterson, J.W., R. Hendrickson, and R.J. Braddock (1971) Florida Citrus Oils. Bulletin 749 (technical), Agricultural Experiment Stations, Institute of Food and Agricultural Sciences. Gainesville, F1.: University of Florida. Luttner, M.A. (1977a) Preliminary Benefit Analysis of Chlorobenzilate. Criteria and Evaluation Division, Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C. (Unpublished) Luttner, M.A. (1977b) Supplement to the Preliminary Benefit Analysis of Chlorobenzilate. Criteria and Evaluation Division, Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C. (Unpublished) Martin, H., ed. (1971) Pesticide Manual. 2nd ed. British Crop Protection Council. (Copies can be obtained from Mr. A.W. Billitt, Clacks Farm, Boreley, Ombersley, Droitwich, Worcester, England.) Mattson, A.M. and M. Insler (1966) Chlorobenzilate Residues in Sheep and Cattle Tissues. Report submitted by Geigy Research Analytical Department to Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C. (Unpublished) McCaskey, T.A., A.R. Stemp, B.J. Liska, and W.J. Stadelman (1968) Residues in egg yolks and raw and cooked tissues from laying hens administered selected chlorinated hydrocarbon insecticides. Poultry Science 47:564-569. McCoy, C.W. (1976) Leaf injury and defoliation caused by the virus rust mite, Phyllocoptruta oleivora. Florida Entomologist 59:40~410. McCoy, C.W. (1977) Resurgence of citrus rust mite populations following applications of methidathion. Journal of Economic Entomology 70:74~752. McCoy, C.W., A.G. Selhime, R.F. Kanavel, and A.J. Hill (1971) Supression of citrus rust mite populations with application of fragmented mycelia of Hirsutella thom~sonii. Journal of Invertebrate Pathology 17:27~276.

Application to Chlorobenzilate 235 McCoy, C.W., R.F. Brooks, M.C. Allen, A.G. Selhime, and W.F. Wardowski (1976a) Effect of reduced pest control programs on yield and quality of 'Valencia' orange. Proceedings of the Florida State Horticulture Society 89:74-77. McCoy, C.W., P.L. Davis, and K.A. Munroe (1976b) Effect of late season fruit injury by the citrus rust mite, Phyllocoptruta oleivora (Prostigmata: Eriophyoidea) on the internal quality of Valencia orange. Florida Entomology 59:335-341. Miya7~ki, S., G.M. Boush, and F. Matsumura (1970) Microbial degradation of chloroben- zilate (ethyl 4,4'-dichlorobenzylate) and chloropropylate (isopropyl 4,4'-dichloroba~zy- late). Journal of Agricultural and Food Chemistry 18(1):87-91. National Cancer Institute (1978) Bioassay of Dicofol for Possible Carcinogenicity. cats No. 115-32-2, NCI-CG-TR-90. DHEW Publication No. (NOSH) 78-1340. Washington, D.C.: U.S. Government Printing Office. Nisbet, I.C.T. (1976) Human Exposure to Chlordane, Heptachlor, and Their Metabolites. Prepared for the Cancer Assessment Group, U.S. Environmental Protection Agency, under contract no. WA-7-1319-A, by Clement Associates, Inc., 1055 Thomas Jefferson Street, N.W., Washington, D.C. 20007. (Unpublished) Ol~nert, I. and R.G. Kenneth (1974) Sensitivity of entomopathogenic fungi, Beauveria bassiana, Verticillium lecor~ii and Verticillium sp. to fungicides and insecticides. Environmental Entomology 3:33-38. Quaife, M.L., J.S. Winbush, and O.G. Fitzhugh (1967) Survey of quantitative relationships between ingestion and storage of aldrin and dieldrin in animals and man. Food and Cosmetic Toxicology 5:39. Reinking, R.B. (1967) Evolution of some spray oils used on citrus in Texas. Proceedings of Rio Grande Horticulture Society 21 :28-34. Riggan, W.B. (1965) Demand for Florida Oranges. North Carolina State University at Raleigh. (Unpublished Ph.D. dissertation) Schmitt, R.D. (1977) Food Factors. Registration Division, Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C. (Unpublished) Severn, D.J. (1978) Exposure Analysis for Chlorobenzilate. H~rd Evaluation Division, Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C. (Unpublished) Simanton, W.A. (1962) Losses and production costs attributable to insects and related arthropods attacking citrus in Florida. U.S. Department of Agriculture Cooperative Economic Insect Report 12: 1182. Sinclair, W.B. (1972) The Grapefruit, Its Composition, Physiology, and Products. Division of Agricultural Sciences. University of California. Thomas, R.F. (1976) Chlorobenzilate Residues in Selected Citrus Products in the Washington Metropolitan Area, Fall 1966. Technical Services Division, Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, D.C. (Un- published) Tilley, D.S. (1977) A Retrospective Look at the 1976-1977 Season: Has a Structural Shift Occurred? Florida Department of Citrus, Tallahassee, Florida. (Unpublished) Tomek, W.G. and K.L. Robinson (1972) Agricultural Product Prices. Ithaca, N.Y.: Cornell University Press. Townsend, K.G. (1976) Two year summary of extension integrated pest management program. Proceedings of Florida State Horticulture Society 89:59-62. U.S. Department of Agriculture (1972) Household Food Con~sumption Survey, 196~1966, Report No. 12. Food Consumption of Households in the United States, Seasons and Years, 196~1966. USDA Agricultural Research Service. Washington, D.C.: U.S. Department of Agriculture.

236 REGULATING PESTICIDES U.S. Department of Agriculture (1977a) Agricultural Statistics, 1977. Washington, D.C.: U.S. Department of Agriculture. U.S. Department of Agriculture (1977b) An Economic and Biotic Evaluation of Chlorobenzilate and Its Alternatives. (Note: This unpublished Final Report is available from Office of Pesticide Programs, U.S. Protection Agency, Washington, D.C.) U.S. Department of Agriculture (1977c) Comments on EPA Report on Estimates of Human Exposure to Chlorobenzilate Prepared by D.J. Severn, January 25, 1977. Submitted to U.S. Environmental Protection Agency on September 12, 1977. (Unpublished, available from Office of Pesticide Programs, U.S. EPA, Washington, D.C.) U.S. Department of Commerce (1978) 1974 Census of Agriculture-Statistics by Subject. Volume II, Part 6. Washington, D.C.: U.S. Department of Commerce. U.S. Environmental Protection Agency (1976a) EPA Actions to Cancel and Suspend Uses of Chlordane and Heptachlor as Pesticides: Economic and Social Implications. EPA- 540/4-76-004. Washington, D.C.: U.S. Environmental Protection Agency. U.S. Environmental Protection Agency (1976b) Pesticide Programs: Notice of Presumption Against Registration and Continued Registration of Pesticide Products Containing Chlorobenzilate. 41 Federal Register 21517-21519. U.S. Environmental Protection Agency (1977) Fertilizer and Pesticide Movement from Citrus Groves in Florida Flatwood Soils. Environmental Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Athens, Ga. EPA- 600/2-77-177. August 1977. U.S. Environmental Protection Agency (1978a) Chlorobenzilate: Position Document 3. Special Pesticide Review Division, Office of Pesticide Programs, U.S. ~PA, Washington, D.C. (Unpublished. Prepared under the general supervision of J.B. Boyd, oPP.) U.S. Environmental Protection Agency (1978b) Environmental Fate Profile: Chlorobenzi- late. Chemistry Branch, Criteria and Evaluation Division, Office of Pesticide Programs, U.S. BPA, Washington, D.C. (Unpublished) U.5. Environmental Protection Agency (1978c) Velsicol Chemical Co., et al., Consolidated Heptachlor/Chlordane Cancellation Proceedings. 43 Federal Register (58)12372-12375. U.S. Environmental Protection Agency (1979) Chlorobenzilate: Position Document 4. Special Pesticide Review Division, Office of Pesticide Programs, U.S. SPY, Washington, D.C. (Unpublished. Prepared under the general supervision of J.B. Boyd, oPP.) U.S. Environmental Protection Agency and U.S. Department of Agriculture (1978) Economic and Social Impacts of Cancelling Use of DBCP as a Pesticide for all Registered Use Sites with Known Current Usage. (Note: This unpublished Final Report is available from oPP' U.S. EPA, Washington, D.C.) U.S. Federal Energy Administration (1976) Energy and U.S. Agriculture: 1974 Data Base. Federal Energy Administration, Office of Energy Conservation and Environment. Vol. 1. FEA/D-76/459. Washington, D.C.: U.S. Federal Energy Administration. University of Florida (1977) Florida Citrus Spray and Dust Schedule 1977. Circular 393-C. Florida Cooperative Extension Service, Gainesville, Florida. van Brussel, E.W. (1975) Interrelations between citrus rust mite, Hirsutella thom~sonii and greasy spot on citrus in Surinam. Landbou~vproefstation Suliname/Agricultural Experiment Station Surinam. Bulletin 98. von Rumker, R. and F. Horay (1972) Pesticide Manual, Part II: Basic Information on Thirty-Five Pesticide Chem~cals. U.S. Agency for International Development. Berkeley, Calif.: University of California Press. Walker, A.I.T., D.E. Stevenson, J. Robinson, E. Thorpe, and M. Roberts (1969) The toxicology and pharmacodynamics of dieldrin (HBOD): Two year oral exposures of rats and dogs. Toxicology and Applied Pharmacology 15:345-373.

Application to Chlorobenzilate Ward, R.W. and R.L. Kilmer (1978) The United States Citrus Subsector: Organization, Behavior and Performance. Department of Food and Resource Economics, University of Florida, Lake Alfred, Florida. (Unpublished) Well, Von L., G. Dure, and K.-E. Quentin (1974) Wasserloslichkeit van insektiziden chlorierten kohlenwasserstoffen und polychlorierten biphenylen im hinblick auf eine gewasserbelastung mit diesen stoffen. Zeitschrift fuer Wasser und Abwasser Forschung 7: 169-175. Wheeler, W.B., D.F. Rothwell, and D.H. Hubbell (1973) Persistence and microbiological efl-ects of Acarol~ and chlorobenzilate in two Florida soils. Journal of Environmental Quality 2(1):115-1 18. Wolfe, H.R., W.F. Durham, and J.F. Armstrong (1967) Exposure of workers to pesticides. Archives of Environmental Health 14:622-633. Woodard Research Corporation (1966) Chlorobenzilate Safety Evaluation by Dietary Feeding to Rats for 104 Weeks: Final Report. Geigy Chemical Corporation, Yonkers, New York. (Unpublished) Yothers, W.W. (1918) Some reasons for spraying to control insect and mite enemies of citrus trees in Florida. U.S. Department of Agriculture Bulletin 645. Washington, D.C.: U.S. Department of Agriculture. 237

Next: Appendix A: Scientific Limitations to Extrapolating Data on Cancer Risk From Animals to Humans »
Regulating Pesticides Get This Book
×
Buy Paperback | $85.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF
  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!