PB80-213887
Chlorbenzilate
Position Document 3
(U.S.) Environmental Protection Agency
Arlington, VA
11  Jul 78
                                                                       J

-------
som-ui
 REPORT DOCUMENTATION
   	PAGE
 4. Tf.it end SubtHI.
                       1. REPORT NO.
               EPA/SPRD  80/J35	[	


Chlorobenzilate:   Position  Document  3
 7. Authord)
             J.B.  Boyd

 9. Performing Ortani, Mion Name and Address
    Special Pesticide Review  Division
    Environmental Protection  Agency
    Crystal Mall 2
    Arlington,  VA
 12. Sponsoring Organization Nam* and Addretl
    Environmental Protection  AGency
    401 M  St.  S.VJ.
    Washington,  D.C.   20460
ilptont't Accession Mo.
    80
                                                                    *. Report Data
                                                                    . Performinat Organization Rapl. No.
                                                       10. ProtKt/Tatk/wark Unit No.


                                                       1 1. Contract(C) or Grant(3) No.

                                                       (C)

                                                       (G)


                                                       13. Typa of Raport 1 Period Covered
 1$. Supplementary Note*
 If. Abttract (Limit: 200 words)
     Risk/benefit analysis:   qualitative &  quantitative  risks  of  a pesticide,
     value  of  crop  uses,  availability  of alternative  pesticides,  exposure to
     man  and environment.    Identificiation  of  risk  reducing  regulatory  options
     and  proposed Agency  action.
 17. Document Analyst*  a. Descriptor*

             0504,0606,0703




   b. Identtfiers/Opan-ErKlea' Teum
   C. COSATI Field/Uroup
 If. Availability Statement

           '  Release Unlimited
                                         19. Security Cl».« (Tnlm Report)
                                           UnclassifieQle
                                                     2O. Security Clan (This Put*)

                                                         •»••»  rtT O *2! C? 4 "fleO/4
       21. '-to. al Pagn

           138
                                                                              22. Prict
(See ANSI-ZM.ll)
                                            a
                                                                OPTIONAL FOMW 272 (4-77)
                                                                (Formerly NTIS-35)
                                                                Department of Commerce

-------
CHDORCBENZILATE:  POSITION DOCUMENT 3
                                               - 20/35"
             J.B. BCXD
          PROJECT MANAGER
  SPQCZAL PESTICIOE REVIEW DIVISION
    OFFICE OF PESTICIDE PBO3BM6
  O.S, EN/OOWNIAL PSCC&CTIQN BGEMOf
                   NAT10NAI TECHWCAl
                   ^FORMATION SWV1CE

-------
                               CHLDBDBENZIIATE

Working Group
J. Boyd, Project Manager, OPP
D. Fatten, Attorney, OGC
D. Barnes, Chemist, OTS
D. Kuroda, Physical Scientist, OBD
D. Dal ton, Entomologist, PTSED

Technical Support Team
J. Boyd, Project Manager, OPP
M. Luttner, Economist, CED, OPP
D. Severn, Chemist, CED, OPP
R. Potrepka, Pharmacologist, CED, OPP
M. McHhorter, Entomologist, CED, OPP
D. Coppage, Aquatic Biologist, CED, OPP
D, Patton, Attorney, OGC
K. Keaney, Economist, OPP
E. Siegler, Attorney, OGC
D. Reed, Chemist, RD, OPP
G. Cmara, Economist, CED, OPP
F. Kutz, Chemist, TSD, OPP
T. Dixon, Chemist, TSD, OPP

Pesticide Chemical Review Camnittee (PCRC)

W. Wells, Chairman, SPRD, OPP
E. Kuroda, Physical Scientist, OBD
J. Moore, GAG Liason, SPRD, OPP
A. Jennings, 0PM
E. Anderson, CMS
M. Menotti, OGC
J. Neyian, PTSED

-------
                         TABU: OP OONTENTS
                                                                            Page
I.    Introduction	  1
        A.  Background	  1
             1.  Chemical and Physical Properties.....	  1
             2.  Registered Uses	,	  2
             3.  Environmental Fate	  2
        B.  Regulatory History	  4
II.   Analysis and Assesanent cf Risk	  8
        A.  Analysis of Rebuttal Submissions	  8
             1.  Successful Rebuttal Arguments	  8
                  a.  Horn et al. (1955)	  8
                  b.  Wbodard (1966)	  9
             2.  Unsuccessful Rebuttal Arguments	 10
                  a.  Innes et al. (1969)	 10
                       i.   Criticisms Relating to Experimental
                              Design and Methods	 10
                             - Size of Test Groups	 10
                             - Choice of Strains	 11
                             - Assignment of Littennates to the
                                 Same Group	,	 12
                             - Treatment of Newborns and Route
                                 of Administration	 12
                             - Pathogen-Free Animals	 13
                       ii.  Criticisms Relating to Pathological
                              Diagnoses	 14
                  b.  NCI (1977)	 16
                       i.   Nd Appraisal of Results	 16
                       ii.  Arguments Relating to Control Animals	17
                       iii. Variable Dose Levels	 19

                                  (iff)

-------
                                                                            Page
                  c.   Other Arguments Against the Innes and NCI Studies	  19
                  d.   Other Arguments	  20
                       i.   Other Cancer Tests.	  20
                       ii. Negative Mutagenicity Testing	  21
                       iii. Lade of Exposure Necessary to Cause Adverse
                            Effects	  22
                       iv.  Epideroiological Data	  22
             3.  Adverse Testicular Effects in Rats	  22
        B.  Exposure  Analysis	  25
             1.  Dietary Exposure	,	  25
             2.  Occupational Exposure	  28
             3.  Cther Potential Exposure	  29
        C.  Risk Assessment	  31
            Risk:  Oncorienic Effects	  31
            Risk:  Testicular Effects	  34
                  a.   Dietary Exposure......	  34
                  b.   Occupational Exposure	  34
III.  Benefit Analysis	  39
        A.  Intvoduction	  39
        B.  Uses	  39
             1.  Citrus Uses	  39
                  a.   Florida	  40
                  b.   Texas	  41
                  c.  California	  42
                  d.  Arizona	  42
             2.  Non-Citrus Uses	  43

                                         (ii)

-------
                                                                           Page
       C.   Alternatives  to Chlorobcnzilate	  44
             1.   Citrus Alternatives	  44
                  a.   No  Miticide	  44
                  b.   Alternative Miticides:   Florida  and Texas	  46
                       i.  Selective Miticides	  46
                       ii.  Non-Selective Miticides	  50
                  c.   Alternative Miticides:   California	  51
                  d.   Alternative Miticides:   Arizona	  55
             2.   Non-Citrus Use Alternatives	  55
        D.   Grower Inpacts	  57
             1.   Citrus Uses	  57
                  a.  Florida	  57
                  b.  Texas	  62
                  c.  California	  63
                  d.  Arizona...	  64
             2.  Non-Citrus Uses	  64
        E.   Costs to the  Citrus Pulp  Feed  Industry	  64
        F.   Costs to Consumers	 65
             1.  Citrus Uses	 65
             2.  Non-Citrus Uses	 66
IV.  Risk-Benefit Analysis of Alternative  Courses of Action	 68
        A.   Continue Registration of  All Uses	 74
        B.   Cancel All Uses	 74
        C.   Continue Registration at  Chlorobenzilate Use on Citrus and
            Mend the Terms and Conditions of Registration; Cancel All
            Other Uses	  76

                                       (iii)

-------
                                                                            Page

             1.  Economic and  Environmental Considerations ..................  76

             2.  Proposed Restrictions ......................................  78

        D.  Cancel Chlorobenzilate Use  on Citrus to Take Effect After
           Fiv Years,  and in the Interim Amend the Terns, and  Con-
           ditions of Registration; Cancel All  Other Uses ..................  81

        E.  Continue Registration of Chlorobenzilate Use on Citrus;
           Amend  the Terms and Conditions of Registration, Require
           That Exposure Data be Submitted to EPA in 18 Months;
           Beevaluate the Use on Citrus After Additional Exposure
           Data Becomes Available; Cancel All Other Uses ...................  Sj
            1,  Citrus Fractionation Studies ..................................  84

            2.  Feeding Citrus Byproducts to Cattle Study ....................  85

            ".  Citrus Pickers Exposure and Reentry Studies ................. ...  85

            't.  Aerial Application Exposure Study .......................... ...  85

            5.  Ground Applicator Exposure Study ..............................  86

            6.  Residue Monitoring of Milk From Pulp-Fad Cattle and
               Residue Monitoring of By-Products of Citrus Processing .......   86

        F.  Continue Registration of Chlorobenzilate Use en Citrus in
            Florida, Texas, and California, Amend the Terms and Conditions
            of Registration, Require that Identified Exposure Data be Sub-
            mitted to EPA in 18 Months; Reevaluate the Use on Citrus After
            Additional Exposure Data Becomes Available; Cancel Use on Cit-
            rus in Arizona and All Other Uses ...............................   87

        G.  Continue Registration of Chlorobenzilate Use on Citrus, Amend
            the Terms and Conditions of Registration; Prohibit the Use
            of Pulp from Chlorobenzilate-Treated Citrus as Cattle Feed;
            Establish Complementary Tolerances; Cancel All Other Uses ......   87

V.   Recommended Regulatory Action.... .....................................   88

        A,  Introduction [[[   88

        B.  Comparison of Options ..........................................   90

        C.  Reconnenaaticn of Option F ............. , .......................   94

Appendix A.  Summary of Other Cancer Stuc'ies
Appendix B.  Other Toxicity Data
Appendix C.  Risk Considerations Relating to Pesticide Substitutes for
             Chlorobenzilate

-------
                                                                          •   PAGE
1.  Agricultural Use of Chlorobenzilate, 1975	    3
2.  Survival Rates for Animals Examined Histopathologically, NCI Study	   18
3.  Reasonable tipper Limit of Dietary Exposure to Chlorobenzilate	   27
4.  Chlorobesizilate Occupational Exposure	   30
5.  Innes Data Used To Estimate Risk	   32
6.  Chlorobenzilate Potential Occupational Cancer Risk	   35
7.  Chlorobenzilate Potential Cancer Risk Through Dietary Exposure;
    Florida Population	   36
8.  Chlorobenzilate Potential Cancer Risk Through Dietary Exposure;
    U.S. Population (Except Florida)	   37
9.  Estimate of the Maximum Economic Value Lost as a Result of Uncontrolled
    Mite Tr-festations in Citrus	   45
10. Per-Acre Treatment Costs in Citrus with Chlorobenzilate; Alternative
    Hiticides and Sealicides	   48
11. Estimated Extent of Use of Chlorobenzilate and Selected Substitutes to
    Replace Chlorobenzilate for Control of the Citrus Mite Complex and Scale
    Insects in Florida	   52
12. Estimated Extent of Use of Chlorobenzilate and Substitutes to Replace
    Chlorobenzilate for Control of the Citrus Mite Complex in Texas	   53
13. Estimated Extent of Use of Chlorobenzilate Compared with Amount of Oil
    Treatment to Replace Chlorobenzilate for Control of the Citrus Bud Mite
    on California Lemons	,	   54
14. Chlorobenzilat* and Alternatives for Non-Citrus Uses	   56
15. Impacts Projected to Grower? Resulting from Cancellation of
    Chlorobenzilate	   58
16. Projected Cost of Scale Insect Control in Florida Citrus During Initial
    Five Year Period Following Cancellation of Chlorotenzilate	   59
17. Economic Impact of the Loss cc* Chlorobenzilate During Initial Five
    year Period Following Cancellation	   61
18. Regulatory Options and Maxinum Risk Incidence from ChJ.orobenzilate Use..  69
19. Economic Impacts Resulting from Chlorobenzilate Regulavory Options	  71
                                  FIGURES
1.  Metabolic Routes of Chlorobenzilate	   5
                                     (y)

-------
Page Intentionally Left Blank

-------
I.   Introduction

       Hie Federal Insecticide, Fungicide, and Rodenticide Act (FIFPA) and
the related regulations require the Environmental Protection Agency to re-
view the risks and benefits of the uses of registered pesticides. On May 26,
1976, the Agency initiated this review for chlorobenzilate with the issuance
of a notice of rebuttable presumption against registration and continued
registration (RPAR) of pesticide products containing chlorobenzilate  (41 PR
21517, May 26, 1976).  Based on information developed through the RPAR re-
view, this position document presents the Agency's analyses of the risks
and benefits of chlorobenzilate uses and recomendations regarding regula-.
tory alternatives.
                   y
     A.  Background

          1.  Chemical and Physical Properties
       Chlorobenzilate (ethyl 4,4'-dichlorobenzilate) is a chlorinated hy-
drocarbon which is also known by its trade names:  Acarfaan, Akar 338,
Rospin, Geigy 338, Benzilan, Folbex, and Kop-Mite.  In pure form it is a
yellowish-brown viscous liquid with a melting point of 35° to 37°C at
0.06 mm Bg and a vapor pressure of 2.2 x 10   ran at 20°C.  It is highly
soluble in most organic solvents and petroleum oils but Is insoluble  in
I/ The chlorobenzilate SPAR was one of the first RPARs  issued by the Agency.
   At the tine it was issued, Agency RFAR procedures were still in a for-
   mative stage, and a detailed position document did not accompany the
   chlorobenzilate RPAR.  For this reason, the Agency has included in  this
   position document information which under  current procedures would  appear
   in Position Document 1 and acocnpany the RPAR.

                                (1)

-------
water.  The molecular weight  is 325.2 «nd its structural formula  is:
          2.  Registered Uses
       Chlorobenzilate, an acaricide,  is registered for use on  almonds,
apples, melons, cherries, citrus fruit, cotton, pears, walnuts, ornamen-
tals, trees, and in certain outdoor areas.   It  is  also registered  to con-
trol spiders on boats and docks.  Ninety percent of the current usage
applied  is on citrus crops (Table 1).  there are 18 Federally-registered
chlorobenzilate products; six applications  for  Federal registration are
pending, and there are eight State-registered products for which notices
of application for Federal registration were filed, pursuant to 40 CFR
162.17.  All pending applications are  for cites already registered.
          3.  Environmental Fate
       Little  is known about the metabolism of  chlorobenzilate  in  nan or
other organises or about  its degradation  in soil and  water or by light.
The primary reactions which chlorobenzilate is  likely to  undergo after
field application may include hydrolysis, decarboxylation, conjugation,
and oxidation.  Chlorobenzilate soil persistence studies  found  that its
half-life was 1.5 to 5 weeks, however, the  degradation products were
not  identified or measured (Boyd, H.,  1978).
       Several studies show that chlorobenzilate  is not metabolized by
plants.  In studies on apples  (Murphy  et  al. 1966) and citrus fruit

                                 (2)

-------
                                         Table 1
Registered
  Uses

Citrus:
  Oranges, Lemons,  Fl
  Grapefruit
  Other (Limes,
  Tangerines,
      elos, etc.)

  All Citrus

Cotton
Fruits, Nuts, &
Miscellaneous
Crops

All Uses
V
Agricultural Use of Chi orobenzi late, 1975
Site
Fl
TX
CA
AZ
US
US
US
US
US
Active Ingredients
Used (Pounds)
805,000
101,500
7,500
6,000
920,000
75,700
995,700
39,000
81 ,000
Al 1~ c/ c/
Agricultural Farms Acreage
Uses (%)
72.1
9.1
0.7
0.5
82.4
6.8
89.2
3.5
7.3
No. % No.
8,314 80.1 523,000
2,584 50.8 43,000
315 4.1 4,000
207 6.7 3.000
11,421 43.4 573,000
N/A - 47,000
11,421+ - 620,000+
39,000
24,000
%
67.3
56.6
1.6
.2.8
47.0
45.9
50.9
0.41
1
US
1,116,000
100.0
715,000
i/  Source:  Preliminary Benefit Analysis (Luttner, 1977).

b/  Percent of total U.S. chlorobenzilate used on commodity noted.

£/  Percent is of total farms and total acreage producing each commodity.
                                              (3)

-------
(Gunther et al. 1955), chlorobenzilate was found only in the peel.  Hassan
and Knowles (1969) found that chlorobenzilate rapidly penetrated soybean
leaves and was translocated unchjnged into the leaf stalks.
       Miyazaki et al. (1970) found that chlorobenzilate was metabolized
to dichlorobenzophenone (DBF) by micro-organisms, especially yeasts.  Horn
et al. (1955) found that chlorobenzilate was hydrolyzed to the free acid
(DBA) by dogs.  Knowles and Ahmad  (1971) found that chlorobenzilate is
metabolized by rat hepatic enzymes to at least four and perhaps as many
as seven metabolites.  These results indicate that chlorobenzilate can
be metabolized by microorganisms and animals.  Figure 1 illustrates these
routes and products.

     B.  Regulatory History
       Based on a study by Innes et al.  (1969), an Advisory Committee to
the Secretary of Health, Education and Welfare  (the Mrak Ccranission),
recommended that human exposure to chlorobenzilate be minimized and that
use of this pesticide bt restricted to those purposes for which there are
benefits to human health which outweigh  the potential hazard of carcino-
genicity (DHEW, 1969).
       On May 26, 1976  [pursuant to 40 CFR 162.11(a)(3)]  the Agency  issued
a notice of rebuttable presumption against registration  (SPAR) of pesticide
products containing chlorobenzilate  (41  FR 21517, May 26, 1976) based on
studies in which  tuners developed  in rats  (Horn et al. 1955; Wbodard, 1966)
and mice (Innes et al. 1969) which had been orally exposed to  the pesticide.
After the notice  of. rebuttable presumption was  issued, the National Cancer
                                 (4)

-------
                             FIGURE 1
               METABOLIC BOOTES Of CHLOROBEN2ILATE
                               CH-
                             or H+
                                                    000-
                                      ->  Cl
       CHLOROBENZILATE
                        -CO
 Cl-
 H


•-C

 CH
4,4' -DICHLORGBENZHYDROL
PRODUCT POUND: IN VITRO
RAT HEPATIC ENZYMES
                                4,4'-DICHDORCBENZIUC ACID
                                PRCOUCT FOUND: WMM&L URINE;
                                YEAST
                                         VftTER SOLUBLE CONJUGATES
                                                (UNVERIFIED)
               -OH
                                         4,4'-DICHLORCBENZOPHCNONE
                                         PRODUCT FOUND: IN VITRO
                                         RAT HEPATIC ENZYMES; YEAST
P-CHLOROBEN2OIC ACID
PRODUCT FOUND: IN VITRO
RAT HEPATIC ENZYMES
                                 (5)

-------
Institute (NCI} completed a chlorobenzilate carcinogenesis bloassay which
showed a statistically significant increase of tumors in mice.  The Agency
provided copies of this study to registrants who requested it; these data
and EPA analyses of the data are summarized in Appendix A.  Other toxidty
data are summarized in Appendix B.
       Registrants and other interested parties were offered an opportunity
to review the data upon which the presumption was based and to submit in-
formation to rebut the presumption.  Respondents could rebut the presump-
tion by showing that the Agency's initial determination of risk was in
error, or by showing that consideration of use patterns and exposure indi-
cates that use of the pesticide is not likely to result in any significant
chronic adverse effects [40 CFR 162.11(a)(4}].  Also, registrants and other
interested persons were offered the opportunity to submit evidence as to
whether the economic, social, and environmental benefits or the use of the
pesticide outweigh the risk of  its use [162.11(a)(5))iii)].  Although the
presumption was based on three  studies, the preliminary results of the
NCI carcinogenesis bloassay were available, and comments were received
on all four studies.  The Agency received 35 submissions, 12 from chloro-
benzilate registrants and 23 from other  interested parties.
        As summarized in Section II of this position document, the Agency
has concluded  that Information  submitted in rebuttal to the Horn and
Woodard studies raises serious  questions about the reliability of these
data  for assessing the oncongenicity of  chlorobenzilate, and that the re-
spondents have therefore successfully rebutted these data.  The Agency has
also  concluded that  respondents failed to rebut data in the  Innes study
                                 (6)

-------
and the NCI study.  As a result, the Agency has used the Innes and NCI data
to assess the risks associated with the uses of chlorobenzilate.  In Sec-
tion III, the Agency has analyzed information on the benefits of chloro-
benzilate uses and the probable costs of regulatory action to cancel or
otherwise restrict uses of this pesticide.  An analysis of the risks and
the social, economic, and environmental benefits which would result from
each of six different regulatory options is presented in Section IV.
Finally, Section V presents the Agency's reccnmended option and an expla-
nation of why this option achieves a sound balance between the risks and
benefits of the uses of chlorobenzilate considered in this analysis.
                                 (7)

-------
II.  Analysis and Assessment of Risk
                                         y
     A.  Analysis of Rebuttal Submissions
       Persons submitting rebuttals contended  that data in the four studies
were defective in several critical respects  and  that  exposure to chloro-
benzilate would not lead to significant adverse  effects.   The Agency has
reviewed the carcinogenicity data again in light of  the rebuttal ooranents,
and has concluded that  information and  arguments submitted in rebuttal to
the Horn and Wcodard studies  indicate that these data may not be reliable
for assessing chlorobenzilate's onccgenic effects.   Accordingly, the Agency
accepts the rebuttal arguments against  use of  these  data for assessing the
cancer risk of chlorobenzilate.  The  Agency  also has concluded that argu-
ments and data submitted regarding the  Innes and NCI  studies did not rebut
or otherwise  invalidate the presumption that chlorobenzilate is onccgenic.
          1.  Successful Rebuttal Arguments
               a.  Horn et al.  (1955)
       Horn et al. administered  chlorobenzilate  in the  diet to rats fron
weaning until 104 weeks of age.  Tumors were observed in some animals.
2/ The rebuttals and connents were reviewed by the Criteria and Evalua-
   tion Division (CEE)  of the Office of Pesticide Programs (OPP), the
   EPA Cancer Assessment Group (CAG), and/or two consulting firms.  These
   reviewers extracted  items from the rebuttal submissions which specifi-
   cally and authoritatively addressed the risk data upon which the pre-
   sumption against chlorobenzilate was based.  The reviewers did not
   analyze testimonials and other conments not supported by data or
   references.  The reviewers evaluated rebuttal items as they related
   to the risk information and submitted their contents and conclusions
   to the Agency Working Group on chlorobenzilate.
                                    (8)

-------
Respondents argued, among other things, that too few animals were ex-
amined histopathologically for proper statistical analysis.  For ex-
ample, although 80% of the male control animals survived, tissues were
studied from only 25% of the survivors.  Agency consultants reevaluated
the use of this data to oanonstrate oncogenic effects and reported that
too few animals were examined histopathologicalJy, that tumor incidence
in the control groups was often greater than in the experimental groups,
and that examination of only representative animals and tissues may have
biased the results  (Preudenthal and Leber, 1977; Savage and Hayts, 1977).
The Agency agrees that these factors indicate that the data may not be
reliable indicators of onoogenicity.

               b.   Moodard  (1966)

       The Ciba-Geigy Corporation, a chlorobenzilate registrant, had the
Woodard Research Corporation conduct a study in which rats were given
chlorobenzilate in  the diet.  Registrants, including Ciba-Geigy, con-
tended that the tumor incidence  in the control animals was cften higher
than  the treated animals, that the investigators failed to conduct ne-
cropsies of animals that died before the  scheduled end of the study,
and that they failed to examine  histopathologically an adequate number
of animals.   For example, no more than 20% of the control and test ani-
mals  from each group were histopathologically examined.  Agency consul-
tants confirmed that the studies were defective  in these respects and
that  the data were unreliable for assessing  the onoogenicity of chloro-
benzilate  (Freudenthal and  Leber, 1977).
                                 (9)

-------
       these consultants also reported that there was an increased inci-
dence of liver timers at the highest dose level which, although not sta-
tistically significant, indicated the need for further study.  Since the
Znnes and NCI studies provided reliable data en chlorobenzilate-induced
liver tumors in rodents, data in this study are consistent.  However, the
Agency agreed with the consultants' conclusion that these data alone do
not indicate that chlorobenzilate is an oncogen.
          2.  Unsuccessful Rebuttal Arguments
               a.  Innes et al (1969)
       Innes et al. tested the tumorigenicity of 120 pesticides and  in-
dustrial compounds by continous oral administration to two hybrid strains
of mice.  Tuners of the liver were observed in 52.9%  (9/17) of the male
mice ingesting 603 ppm chlorobenzilate compared with  10.1%  (8/79) of the
control animals.  Liver tumors were not observed in the females.  These
data are detailed in Table 7, Subsection II,C.  This section summarizes
the rebuttals of the Innes study and the Agency's response.

                    i.  Criticisms Relating to Experimental Design
                        and Methods'
       Size of Test Groups;  Registrants argued that only 18 male and
18 female mice were fed chlorobenzilate, and that this was too f"w mice
to permit a valid statistical analysis  [Alikonis, 1976a, 1976b, 1976c
                 I/
(23-25:30000/3)].   The Agency acknowledges that the  number of animals
 3/ The  names  and numbers in parentheses identify the source of each
    rebuttal in terms of the signer,  date,  and EPA identification number.
                                 (10)

-------
in a study does limit the sensitivity of the test, but notes that  in the
Innes study an appropriate proportion of the animals were autopsied and
that both chlorobenzilate-treated groups of F hybrid male mice demonstrated
a statistically significant increase (p £ 0.02) in animals with  tumors
compared to controls (Freudenthal and Leber, 1977).  Toe Agency  also notes
that the Mrak Report concluded that the number of animals per group in the
Innes study was sufficiently large to provide a sound basis for  statistical
analysis of the results.
       Some registrants noted that the Innes study was a mass screening
study which involved irany chemical canpounds snd few animals and that, in
the introduction, the authors urged against drawing conclusions  from the
study without confirmatory research  (Jovanovich, 1976a  [1:30000/3]; Alikonis,
1976a, 1976b, 1976c  [23-25:30000/3]).  These considerations do not negate
the statistically significant (p < 0.001}  increase of hepatctnas  in chloro-
benzilate-treated male mice (Freudenthal and Leber, 1977).   In computing
significance, sample size  is necessarily considered.  Further, the Mrak
Commission referred to the Innes study as  a "fine-mesh screen designed
tr>  identify as many as possible of the carcinogens submitted to  it," and
concluded that  it "performed this task with considerable  success." More-
over, an analysis of the results of  the  recent NCI study  confirm that
chlorobenzilate  is oncogenic  (Barton, 1977).
       Choice of Strains:  Registrants argued  that  the  two first genera-
tion hybrid strains of mice used by  Innes  et al. did  not  have  the genetic
heterogeneity of normal animals and,  therefore, should not be  used in can-
cer research.  They noted  that randomly  bred animals  are  recommended by
                                 (ID

-------
the Mrak Commission, the Food and Drug Administration Panel on Carcirogene-
sia in 1971, and the World Health Organization in 1961  [Balser, Jovanovich,
1976b (20:30000/3); Jovanovich, 1976 (31:30000/3)].  However, the Agency
has concluded that the hybrid mice used in this study wre suitable test
animals.  A low incidence of tumors appeared  in control groups, and the
strain was highly susceptible to induction of tumors by positive control
chemicals (HEW, 19699.  While randomly bred animals are preferred by
some scientists, true random breeding is difficult to achieve in practice
and results in broad, ill-defined gene pools.  Moreover,  inbred strains
have been described as the best means for investigating chemical carcino-
genesis (Goldberg, 1974).
       Assignment of Littermates to the Same Group;  Registrants argued
that the assignment of littermates to the same experimented groups was
inappropriate  [Jovanovich, 1976a (1:30000/3); Murphy, 1976a  (22:30000/3)].
The Mrak Conoission reported that when the Innes data were reanalyzed
under a nore rigorous statistical procedure to take this  possible bias
into account, the differences reported as statistically significant  in
the original study remained significant  (HEN, 1969).
       Treatment of Newborns and Route of Administration! Registrants
stated that the FDA criteria for cancer testing recommend that newboms
not be  included in cancer tests  [Balser, Jovanovich, 1976b  [20:30000/3);
Alticonis, 1976a, 1976b, 1976c  (23-25:30000/3)].  Specifically, the Innes
study VJBS faulted for beginning to administer chlorobenzilate when the
mice were 7 days old. Registrants argued that the  innune  system  in these
very young animals  is  incompletely developed, rendering them more sus-
                                 (12)

-------
oeptible to the development oi' tumors,  the registrant cited the FDA
guidelines .jhich state, This might be a satisfactory screening proce-
dure under limited conditions but cannot be reeonsmendad as a routine
test procedure" (FDA, 19/1).
       Ihe Agency has rejected this argument because young and more sus-
cpetible animals are used in oncogenicity studies to optimize detection
of oncogenic activity.  In addition, when a relatively small number of
animals is ussd in feeding studies, it is appropriate to  increase the
sensitivity of the experiment with this technique.  This  approach is ex-
plained in the NCI Guidelines for Carcinogenic Bioassays  in Small Rodents
(Sontag et al. 1976) which state that chronic study animals should be
weanlings, if possible, since a poorly developed  immune system may sen-
sitize these animals to carcinogens and make them more susceptible.  Hu-
man  infants also exhibit a weak immune response.
       One registrant contended that the administration of chlcrobcnzilate
to infant mice by gavage flawed the study  [Jovanovich, 1976a  (1:30000/3)].
Gavage is an accepted means of administering a test material  to animals,
especially to young animals or when an exact oral dose  is required  (Wilson,
J.G., 1973; Goldberg, 1974).
       Pathogen-Free Animals;  With reference to  the FDA  guidelines  (FDA,
1971), one registrant stated,  "Specific pathogen-free animals should be
avoided because of possible  increased susceptibility to Infection"  [Balser,
Jovanovich, 1976b  (20:30000/3)].  The FDA's recotmendation  is intended as
an experimental design measure  to safeguard against early deaths  from  in-
fection; reductions  in  the number of  animals  surviving  a  long-term  study
                                 (13)

-------
would result in a less valid study.  Since there was a high survival rate
among the animals in the Innes study and no data indicating that there was
significant infection, the FDA precaution is no1-  relevant  in this case.

                    ii.  Criticisms Relating to Pathological Diagnoses

       One registrant argued that, since most of the animals in the Innes
study did not die of cancer but were sacrificed at the end of  the study,
chlorobenzilate is not a potent carcinogen  [Balser, Jovanovich, 1976b  (20:
30000/3}].  Agency regulations which authorize the rebuttable  presumption
review provide that a presumption will arise if a chemical is  oncogenic,
or tumor-producing.  Such tumors need not cause the death  of the animals.
       One registrant stated, "Many of the cancers are detected only by
microscopic er-amination" in the Innes study  [Balser, Jovanovich, 1976b
 (20:30000/3)1 and concluded that chlorobenzilate should  not be considered
an oncogenic risk.  Diagnosis of tunors by microscopic examination of  his-
tological preparations  is essential to a complete review of the effect of
a chemical on tissues.  The risk criterion provides that a rebuttable
                              »
presunption shall arise  it a ccnpound  is oncogenic; there  is no require-
ment that tumors be macroscopically detectable.
       One registrant argued that  chlorobenzilate should not be considered
an oncogenic risk wince  tumors found  in the  Innes study  had not generally
metastasized  [Balset, Jovanovich,  1976b  (20:30000/3)].   The EPA "Interim
Guidelines for Carcinogen Risk Assessment"  (41 FR 21404, Hay 25,  1976)
 specify that:
                                 (14)

-------
          Substantial evidence [that an agent is a human car-
          cinogen]  is provided by animal tests that demonstrate
          the induction of malignant tumors in one or mare
          species including benign tumors that are generally
          recognized as early stages of malignancies.

       Because diagnoses of cancer can be made only by expert pathologists,
several registrants have concluded that the evaluations of the Innes study
are subjective [Balser, Jovanovich, 1976b (20:30000/3); Alikonis, 1976a,
1976b, 1976c (23-25:30000/3}].  The Agency acknowledges that pathologic
diagnosis is a special skill and relies on the known ability of  the pa-
thologists who reviewed the Innes study slides.  The independent diagnoses
of the patholocj ists who reevaluated the slides add considerably  to the
Agency's confidence in this judgment.
       One registrant cited FDA's 1971 Criteria  for Cancer Testing (FDA,
1971) which recommended, "It is desirable that both gross and microscopic
examinations be conducted without knowledge of the treatment of  specific
animals."  The registrant argued that the pathologists who reevaluated the
Innes study slides would know the treatments prior to examining  the slides
[Balser, Jovanovich, 1976b (20:30000/3)].  The registrants have  not submit-
ted any specific information that would cause the Agency to doubt the va-
lidity of the observations of the pathologists who reevaluated these slides.
The pathological evaluations were performed through a  "blinded"  evaluation
of the tissue preparations of unknown origin.  This procedure precludes  the
possibility of the suggested bias, regardless of whether or not  the patholo-
gists knew the experimental protocol.
       One registrant contended that EPA demonstrated  its  lack of confi-
dence  in the Innes study when the Agency's Toxicology  Branch, Registration

                                 (15)

-------
Division (RD), refused to accept data fron this study  as  evidence  that cer-
tain other chemicals were not oncogenic  [Murphy, 1976a (22:30000/3)].   The
refusal of RD to accept the fact that sane Innes study data demonstrated
that sane chemicals are not care:icgenic does not  conflict with the accep-
tance of data relating to chlorobenzilate.   Specifically, since chemicals
tested in this study were not administered throughout  the entire lifetimes
of the animals and since a small number  of animals was tested,  conditions
were not optimum for tumor detection, and therefore there are bases for
questioning negative findings.

               b.  NCI (1977)

       A summary of the final  individual animal pathology data was pro-
vided by NCI on January 27, 1977  (NCI, 1977).  ICI tested the oncogenicity
of chlorobenzilate  in B6C3F1 mice  and Osborne-Mendel rats.  Neither sex
of rats showed statistically significant responses.  The incidence of
hepatocellular carcinonas  in male  mice was 68%  (32/47) at 4,000 ppm and
49%  (22/45) at 6,000 ppm,  compared to 24%  (4/17)  in 'die  on'rol group.
These tumors were observed in  23%  (11/49) of the  female trice ingesting
3,200 pan and 26% at 6,400 (13/50), compared to 0% (0/20) in the control
group.  An EPA analysis  indicates  that  the chlorobenzilate-treated mice
snowed statistically significant increases  in the number of total tumors
and  heptatocellular carcinomas at p - 0.01V  and p - 0.0001, respectively
 (Earton, 1977a).
                     i.  NCI Appraisal of Results

       Registrants  submitted  a letter from the Chief of  the Carcinogen

                                  (16)

-------
Bioassay and Program Resources Branch, NCI, which stated that  "the com-
pound appears not to be carcinogenic to mice and rats of both  sexes"
[Jovanovich, 1976a  (1:30000/3); Murphy, 1976a  (22:30000/2)].   The letter
refers only to a preliminary inspection of unverified data as  of May 13,
1976, and does not  necessarily reflect the final NCI interpretation of the
data.  EPA's analysis of the verified histopathology report of January 27,
1977, indicates a statistically significant  (p » 0.017) increase  in total
tunors for mice of  both sexes fed chlorobenzilatd (Barton, 1977a).  Id
has not yet completed their analysis of this final  data.

                    ii.  Arguments Relating  to Control Animals

       The occurrence of tumors  in the control group was not the  sane as
the overall incidence usually observed in untreated mice of this  strain.
One registrant argued that  if the  incidence  of tumors  in the treated mice
were compared to the historical baseline  incidence  of  tumors  in this strain,
the difference in tumor occurrence would  not be statistically  significant
(Murphy, 1976a  [22:30000/3]).  tCI reports that the historical incidences
of spontaneous primary liver tunors  in untreated male  and  female  B6C3F1
mice are 15.6% and  2.5%, respectively; the comparable  rates for concurrent
controls in the study were  23.5% and 0%.  The  Agency's Cancer  Assessment
Group  (CM) advised that because this difference  is negligible for a sanr
ple size of 20 control animals,  the  historical control data does  not appre-
ciably affect the significance of  the chlorobenzilate  findings (Albert,
1978).
       Citing the World Health Organization's  (VHO) recommendations, one
                                 (17)

-------
registrant claimed that the control group should have  contained  70 animals

[Balser, Jovanovich, 1976b (20:30000/3)].  The ocntrol group contained  only

20 mice although the treatment groups contained 50 mice each.  C&G evaluated

this rebuttal and the WHO guidelines, and pointed out  that there should have

been 35 animals in each test group to provide optimal  statistical efficiency

(Anderson, 1977).  It was emphasized/ however/ that  even  though  the  test was

not as sensitive as it. could have been, a statistically significant  ele-

vation in tumor incidence was found  (Barton, 1977a).

       One registrant argued that, since only 60% of the  male mice  in the

control group survived to the end of the study, meaningful statistical

evaluation of the study was not possible  [Balser, Jovanovich,  1976b  (20:

30000/3)].  The statistical analysis was based on the  animals that survived

and were examined histopathologically.  The  survival rate in the controls

was better or the sane as that of the  low dose group to which it was com-

pared.  This comparison  is shown  in Table 2.  As a  result the significant

"p" value  is not biased by better test group survival. Using the Fishers


TABLE 2.  Survival Rates for Animals Examined Histopathologically (tCI, 1977)

         Week of Study       Control Group       Low Dose Group

             48                 1.000                 0.979
             50                 0.941                 0.915
             SI                 0.941                 0.872
             36                 0.941                 0.851
             63                 0.941                 0.830
             68                 0.941                 0.809
             69                 0.941                 0.787
             72                 0.882                 0.787
             73                 0.824                 0.787
             74                 0.765                 0.787
             77                 0.765                 0.766
             86                 0.765                 0.745
                                 (18)

-------
exact 2x2 test, the "p" value is less than 0.002, which  indicates  that
the differences in tumor incidence observed between treated and  control
animals were highly significant (Anderson, 1977).

                    iii.  Variable Dose Levels
       Che registrant pointed out that dose levels varied  accoring to the
age and sex of the animals  [Balser, Jovanovich,  1976b  (20:30000/3)].   The
dosing regimen was selected on the basis of preliminary  tests to determine
maximum tolerated doses.  Standard procedure for selecting dosage levels
permits variation according to sex.   In addition, food intake relative to
body weight can vary with age; thus,  an adjusted concentration of the chemi-
cal in the diet to accommodate for this is solid experimental procedure
(Mishra, 1977; Edwards, 1977).

               c.  Other Arguments Against the  Innes and ICI  Studies
       Cne registrant argued that neither Innes nor NCI  adhered  to all FDA
criteria for carcinogenic it? testing  (FDA, 1971), but  did  not note specific
inconsistencies  in the NCI  study  (See Section II,A,2,a,  for  Innes inconsis-
tencies).  Although the FDA criteria  were developed to test methods  and
to provide a sensitive screen for cancer-causing agents, deviation from
these guidelines does not alone  invalidate a study.  In this case, since
both the NCI and Innes studies showed statistically significant  increases
in tumors  in mice, the studies were clearly sensitive  enough to  detect on-
oogenic activity.
       Another argument was that  the  secondary  toxic effects noted  in these
studies made pathological evaluation  difficult  [Balser,  Jovanovich,  1976b
                                 (19)

-------
(20:30000/3); Alikonis, 1976a, 1976b, 1976c  (23-25:30000/3)].  Vhile  sec-
ondary toxic effects may complicate the diagnosis of tuners, the regis-
trants have offered no evidence which casts  doubt on the validity of  the
patholegists' conclusions  in these studies.
       Several registrants contended that  the Agency's presumption  of
chlorobenzHate's oncogenic risk was based on false positive results  from
both Innes aid NCI studies [Jovancvich, 1976c (31:30000/3)].   No data were
offered to support this argument, and based  on  both studies, the probability
of false positive results  for  the  increase of tumors  in chlorobenzilate-
treated mice was very slight (Barton, 19T7).
       Several registrants argued that data  fron the  Innes and NCI  studies
were unreliable because  inbred strains may carry tumor-causing viruses.
These registrants maintained that  the tuners found  in these studies were
similar to those produced  by viruses  [Balser, Jovanovich,  1976b  (20:30000/
3); Alikonis,  1976a, 1976b, 1976c  (23-25:30000/3); Jovanovich, 1976c  (31:
30000/3)].  This argument  was  rejected  for two  reasons  (Mishra,  1977).
First, although the  association of viruses (or  C-type particles)  and  the
development of tumors  in test  animals has been  demonstrated,  no  causal
relationship  has been established.   Second,  even if viruses in mice were
oncogenic, the control  animals used  in  both tests  showed markedly fewer
tumors  than the chlorobenzilate-tested  animal.
               d.  Other Arguments

                     i.   Other Cancer Tests
        One registrant  [Murphy, 1976a (22:30000/3)]  cited a 1965  Bazelton
                                (20)

-------
Laboratories' study in which chlorobenzilate was fed at levels of 0, 100,
500, and 3,000 ppm active ingredient to six beagle dogs.  No evidence of
oncogenic activity was observed.  However, the animals were exposed to
chlorobenzilate for only two years, which is generally too short a portion
of the dog's life span for the development and detection of tumors.  It is
generally accepted that in an oncogenic bioassay, a test agent should be
administered continuously for the larger part of an animal's life span to
achieve "greatest confidence" in a negative result.  This translates to a
test duration of 7-10 years for carcinogenicity tests in dogs (Page, 1977).

                    ii.  Negative Mutagenicity Testing

       Two registrants referred to the frequently observed correlation be-
tween nutagenic and oncogenic activity and cited data which suggested that
chlorobenzilate is not nutagenic and, therefore, not oncogenic  [Murphy,
1976a (22:30000/3); AlDeonis, 1976d  (33:30000/3)].  This argument can be
rejected for two reasons.  First, mechanisms other than mutation may cause
cancer.  Second, although there is a high correlation for some classes of
chemicals between carcinogenicity  in nannalian test systems and nutage-
nicity  in certain microbial systems, this correlation is not perfect, and
false positive and false negative  indications do occur.  For instance, the
reversion assay in Salmonella with metabolic activation  , i.e. the Ames
test, has a high "false negative"  correlation for chemical classes such
as cyclodienea, chlorinated hydrocarbons, and certain metals,  foe these
classes of compounds  the Ames test is often negative, although there are
positive results in mamnalian bioassays  for carcinogenicity.  Chloro-
                                (21)

-------
benzilate falls into the potentially  "false negative" group of  chemicals
(Pertel, 1977).  For these reasons  the negative results  in the  in vitro
test battery submitted in rebuttal  of the carcinogenicity  of  chloroben-
zilate are not convincing (Pertel,  1977;1978).

                    iii.  Lack of Exposure Necessary to  Cause Adverse
                          Effects
       Two registrants concluded that even  if chlorobenzilate were  carci-
nogenic, the human exposure would not be high enough to  pose  any risk
[Murphy, 1976a (22:30000/3); Murphy 1976b  (22:30000/3)].  The exposure es-
timates  in Section II,B, of this document and the  risk estimate in  Section
II,C, in the Agency's judgment, indicate that humans may be exposed to
amounts  of chlorobenzilate which may cause  sufficient adverse effects to
require  the Agency to consider whether uses  of  chlorobenzilate  offer off-
setting  social, econonic, or environmental  benefits.
                    iv.  Epidemiological Data
       A preliminary company report on chlorobenzilate manufacturing epi-
demiology  [Murphy, 1976a  (22:30000/31] was  submitted to  show that no un-
usual health problems had been detected  among employees  of a  chloroben-
zilate production facility  in Albania. The study was determined to be
inconclusive  (Rossi, 1976), and was not  further considered.   No other
reliable epidemiological data  is currently available.
           3.   Adverse Testicular Effects in Rats

       The studies on which the presumption against chlorobenzila'ce was
                                 (22)

-------
based include data indicating that chlorobenzilate has  adverse  effects
en the testes of rats.  The Agency did not review these data  in its RPAR
notice and the registrants were not  invited  to  comment  in  their rebuttals.
However, in reevaluating chlorobenzilate studies  in  connection  with the
RPAR review, the Agency concluded that the testicular effects'  data also
required consideration.  Re-examination of these studies disclosed that
testicular toxicity  in rats has been reported  in five of the  studies
examined.
       Horn et al. (1955)  in a two-year study  using  50  and 500  ppm,
found a dose-related increase in the number  of  small and/or soft testes
among the survivors  of the study.  Conpared  with 25% (4/16) of  the con-
trols, at 50 ppm chlorobenzilate, 69% (9/13} of the  male rats evidenced
testicular effects,  and at 500 ppro chlorobenzilate,  100%  (14/14) showed
              i/
these effects.  In a 2-year study using 40,  125 and  400 ppm,  Woodard
      */
(1966)  found, among the animals examined, more frequent  testicular
changes at 125 ppm and 400 ppm than  in the  controls  or  at  40  ppm.  At
125 ppm, 33%  (2/6) of the  animals evidenced  change and  at  400 ppm, 60%
(3/5) cccpared with  0%  (0/5)  in the  controls and with  20%  (1/5) at 40
ppm.  A dietary level of 40 ppm was  considered the no-effect level in
this study.
 4/ The rebuttable presumption against chlorobenzilate was based in
    part on orcogenicity data presented in this study, and the Agency
    has concluded that respondents sucessfully rebutted these data as
    Co the oncogenic presumption.   However, these rebuttal arguments
    do not apply for all purposes and independent analyses of the tes-
    ticular effects' data is appropriate.
                                 (23)

-------
       A two-year study by NCI (1977), in which rats were  fed 1,600 or
3,200 ppm for 78 weeks, reported an  increased  incidence of abnormal tes-
ticular pathology among treated males.  Sixty-six percent  (33/50)  evi-
denced adverse effects at both the high and  low doses  compared with 18%
(9/49) of the controls.  Testicular  atrophy  was the most conronly  observed
effect.  Further, in a 99-day subacute study using 20,100, 500,  and 2,500
ppm, spenniogenetic injury and atrophy of the  gonads were  found  in 25%
(5/20) of the rats at the highest dose level.  In this study, 500  ppm was
judged to be the no-effect level (Fotrepka,  1978a).  In a  3-generation  study
in which F  rats were fed 50 ppm and subsequent generations fed  either  25
or 50 ppm, Wbodard (1966) found decreased testicular weights in  F  males.
                                                                  Ib
At 25 ppm, mean testicular weight was reduced  to 3.04  g and at 50  ppm,
significantly reduced  (p » 0.05) to  2.75 g compared with 3.24 g  in the
controls  (Quaife, 1966).
       Chlorobenzilate studies using other species  are limited,  and  those
that are  available generally do not  contain  conments on  testicular changes.
In a review of slides  from the Innes study on  mice, one pathologist  indicated
that there might be a  treatment-related  increase  in testicular atrophy.
These data could only be considered  suggestive because of  the limited num-
ber of samples studied  (Frith, 1976).  Testicular  effects  were not re-
ported for mice  in the NCI study.
       Die biological  significance of the adverse  testicular effects has
not been  established.  Although  the  reproduction study examined  did not dis-
close significant differences  in fertility between treated and untreated
animals,  the  regular occurrence  of testicular atrophy in animals exposed
                                 (24)

-------
to chlorcbenzilate warrants concern.  Since none of the other studies
were designed to measure reproductive capacity or testicular function,
the reproductive and physiological significance is unknown.  Consistent
with its obligation to protect human health, the Agency must assume that
the data (shewing chlorobenzilate-induced injury to the tastes) indicates
that chlorobenzilate may also interfere with the endocrine or spennato-
genic function of this organ, in the absence of data establishing that no
such interference occurs.

     B.  Exposure Analysis

       Registrants, tne U.S. Department of Agriculture, and other sources
provided data on patterns of chlorobenzilate use which the Agency has used
to  identify populations which may b~- exposed to chlorobenzilate and to es-
timate the extent of exposure.
          1.  Dietary Exposure
       The Agency's estimates of human dietary exposure are based on resi-
due data and the extent to which chlorobenzilate is used on each of the
food crops for which it is registered.  A reasonable upper limit of the
dietary exposure of the general U.S. population teas calculated, based on
the average individual's consumption of oamodities,  including orange
juice, produced fron crops treated  with chlorobenzilate  (Severn, 1978).
Because there were no detectable residues  in most of  the edible portions
of  these foods, these sources were  assumed  to contain 0.1 ppra, which is
the limit of detection  in the analytical method used  to measure represen-
tative samples  (FDA, 1971).  Accordingly,  the calculations presented in
                                 (25)

-------
Table 3 nay be regarded as reasonable upper  limit estimates.    In  addition,
since apples and pears are eaten whole, residues in  these crops were  esti-
mated at 5 ppm, the established tolerance level.  Finally, although chloro-
benzilate-treated citrus  is processed into other products, e.g. citrus oil,
which nay also be sources of dietary exposure, data  for estimating expo-
sure from these sources were not available (See Section TV, Option E).
       Florida residents may ingest additional significant amounts of
chlorobenzilate because pulp from chlorobenzilate-treated citrus fruit
is fed to dairy and beef  cattle which are raised and marketed in Florida,
although a limited EPA survey of Florida milk "sarples  detected no  resi-
dues at the 20 ppb detection limit  (TSD, 1978).
       Formica, et al.  (1975) reported  that  on days  1  thorough 42  during
which cows were fed pulp  to which 20 ppm chlorobenzilate had  been  added,
chlorobenzilate levels  in milk ranged fron 0.03 ppm  to 0.04 ppm.   EPA also
completed a limited survey of Florida citrus pulp-mixed feed  and found that
the chlorobenzilate content averaged 0.16 ppm  (TSD,  1978).  Other  pulp data
FDA, 1976 and proprietary data  (Reed, 1978d) indicate that chlorobenzilate
can occur at 2 ppm  in citrus pulp fed to cattle.  Based on these data, the
Agency has estimated  that chlorobenzilate may be present  in milk at 1 to
3 ppb (Reed, 1978c).  This level could  not be detected by the current FDA
 4/ The Agency's estimates of exposure and risk are based on the data,
     information and assumptions cited for each estimate.  In many cases,
    a range of values or several reasonable assumptions, tested or un-
    tested, are appropriate for the analysis.  The Agency generally se-
    lects values and assumptions which permit a conservative (from the
    standpoint of protecting tne public health) risk estimate rather
    than  using average values or generalized assuiptions.
                                (26)

-------
                                 Table 3
       REASONABLE UPPER LIMIT OF DIETWOr EXPOSURE TO CHLOHOBENZILATE
U.S. Population Exposure
Ccmnodity
Citrus:
Oranges (inc.
Grapefruit
Other Citrus
Other Fruit:
Nuts:
Total U.S.
S/
Consunption
(g/dayT"
juice) 42.00
19.30
12.70
55.20
1.18

Extent
of Use by
Crop (%)

47.80
60.90
31.00
0.08
3.60

St
Assumed
Maximum
Residue (pptn)

0.1
0.1
0.1
5.0
0.1
<
Maximum
Ingestion
(ug/day)
< 2.01
< 1.18
< 0.40
< 0.21
< 0.004:
3.8 (0.002
Beef and Lamb

Milk
                             [Percent with
                              Potential  ~.
                              Occurrence]^

                      143.2      10            0.04           < 0.57

                      184.7     100      0.0024 - 0.04    < 0.44 - < 7.39

                        Total Florida Additional < 1.01 - < 7.96

                            Grand Total Florida  < 4.81 - < 11.76
                                              (< 0.0025 - < 0.0061 ppn)
£/
Ssvern, 1978
Doane, 1976
Detection level in the nose representative sampling
Tolerance level
Feeding by-products of citrus processing (pulp and molasses) to cattle
in Florida is viewed as an indirect dietary source of chlorobenzilate.
It results in additional dietary exposure for the Florida population.
Based on limited EPA survey (Luttner and McWhorter, 1978)
                                    127)

-------
monitoring method.  However, because  these data  are  inconclusive,  a range
has bean  expressed (Table 3) up to 40 ppb which would  be  the maximum theo-
retic?! occurrence (Reed/ 1978a).  In addition,  based on the Mattson and
Insler study  (cited in Severn, 1978), the Agency estimated that chloroben-
zilate may be present at 0.04 ppm  in  beef (Table 3).
          2.  Occupational Exposure
       Ground applicators and citrus  pickers are exposed to chlorobenzilate
through  its use  on citrus crops.   The USDA estimates  that  the current use
of chlorobenzilate in ground application on  citrus is carried-out by as
few as 714 applicators for 30 to 40 days per year or by as many as 1375
applicators for  10 to 20 days per  year  (Severn,  1978).   The worst case,
which  is represented by 714 applicators  for  40 days  per year, is shown in
Table 4A.  No data is available to indicate  the  actual  amounts of chloro-
benzilate exposure to these applicators; however data on exposure to ground
applicators during application of  other  pesticides (Wolfe, et al.1967) was
used to  estimate these amounts at  between 120 mg and 440 mg dennally and
1 rag by  inhalation during each day of application.  Since  similar pesticides
are known  to  be  absorbed through  the skin at a rate of  only about 10%
 (Feldtnann  and Maibach, 1974), the  daily dermal dose range was estimated to
be 12  to 40 mg;  Feldian and Maibach  assumed  that pesticides which are  in-
haled  are  absorbed 100%.  Thus, the  total daily dose for ground applicators
was estimated at 13 mg to 41 mg.   Assuming up to 40 years of occupational
exposure to  chlorober.zilate,  the  average daily dose ranged from 0.81 mg to
2.57 mg.  For purposes of relating this estimated exposure to the animal
dose-response data,  the  average daily dose  is expressed as 0.39 ppn to
                                 (28)

-------
1.3 ppm dietary equivalents  (Table 4).
       There are approximately 25,000 to 30,000  citrus pickers who may be
occupationally exposed to chlorobenzilate, but there  is  too little data
available at this tune to permit a estimate of exposure  and potential  risk.
Since citrus pickers work in the groves after the pesticide is applied,
the Agency has assumed that  their exposure  is less  than  the applicators'
exposure.  The Agency has also assumed that surface ohlorobenzilate  resi-
dues on fruit and foliage surfaces may be a primary source of contact  ex-
posure.  If citrus pickers are exposed to the pesticide  in this manner,
an exposure estimate could be based en residue data together with the  10%
dermal absorption and 40-year duration figures used in estimating appli-
cator exposure (See Section  IV, Option E).

          3.  Other Potential Exposure

       Chlorobenzilate  is registered for aerial  application to citrus
for the control of citrus rust mites.  This manner  of application could
result ir drift, depending on the speed of  the wind and  the size  of  the
5/  The applicator  exposure estimate  is based on amounts of chloro-
    benzilate to which applicators may be exposed through the skin
    and through lung  absorption during pesticide application.  In this
    case, the values  range front 0.81  to 2.57 mg/day.  However, since the
    risk estimates  (See Section II,C) are based in part on animal test
    data derived from dietary exposure, reported as ppm/day, the hu-
    man exposure and  the animals' exposure are not expressed in the
    sane terms.  Therefore, for purposes of relating the animal dose-
    response data which is expressed  in ppm to the human exposure
    data which  is calculated initially as mg/day, the latter has beer.
    converted to dietary equivalents  (Albert, 1978).
                                 (29)

-------
                                  Table 4


                               CHLORQBENZILATE

                           OCCUPATIONAL EXPOSURE
                GROUND APPLICATORS (AGRICULTURAL-CITRUS ONLY)

                                                    SJ
                Maximum Extent of Exposure (Absorbed /day)
                   Inhalation               - 1 ing/day
                   Dermal                   - 12-40 mg/da>

                Duration of Exposure
                   Assumption - 40 days/year for 40 years
                                                                           k./
Reasonable Upper Limit of Exposure - (0.39 to 1.3 ppm)(Dietary Equivalents)

Maximum Number of Ground Applicators at This Level of Exposure - 714
a/  Adapted from Wolfe, et al.  (1967) data on similar pesticides (Severn, 1978)

b/  Dally amount, time-weighted on duration of exposure (Thorsland, 1978)
                                   (30)

-------
spray droplets.  In this way there could be exposure to people near the
vicinity of application. However, there is no data on which to base an
estimate of the magnitude of the potential exposure in this situation.
     C.  Risk Assessment
          1.  Risk: Oncoqenic Effects
       The chlorobenzilate cancer risk assessment is based on the princi-
ples and procedures outlined in the EPA cancer risk assessment guidelines
(41 PR 21402, May 25, 1976).  These guidelines specify that a substance
will be considered a "presumptive cancer risk when it causes a statis-
tically significant excess incidence of benign or malignant tumors  in hu-
mans or animals," that current and anticipated exposure levels are appro-
priate considerations, and that cancer risk estimates may be derived  from
a variety of risk extrapolation models such as the log-prohibit  and linear
non-threshold models.
       In accordance with these principles, the  EPA Cancer Assessment Group
(CAG)  (Albert,1978), and Agency consultants  (Felkner and  Lcrobardini,  1978)
developed risk estimates using several different models and a range of ex-
posure estimates.  CAG has concluded that  "...the weight  of evidence  indi-
cates  that chlorobenzilate is a possible human carcinogen"  (Albert, 1978).
After  reviewing  the data sources and the preliminary risk estimates,  CAG
concurred in recommendations that  the final risk estimates be based on data
                                                                       */
from the Innes study using the one-hit model  (Table 5)  (Albert,  1978).
CAG and the consultants recommended using  the Innes rather  than  the Nd
data because the onccgenic response per unit  of  dose of chlorobenzilate
 6/ CAG has reviewed risk according to the NCI data (Albert, 1978b).

                                 (31)

-------
                                 Table 5
                    INNES DMA  USED TO ESTIMME RISK
         Strain
          "X"
Dose (ppm)
0
603
C
603
Incidence
Hepatcraas
Hale Female
h/
3/79 (10.1%)
9/17 (52.9%)
5/90 (5.6%)
7/17 (41.2%)
0/87 (0%)
0/18 (0%)
1/82 (1.2%)
0/18 (0%)
a/ This was an eighteen-month feeding study on two hybrid strains of mice.
   "Strain X" - (C57BL/6 x C3H/Arf)F,;  "Strain Y" * (C57BL/6 x AKR)F,,
   Innes et al., Journal of the National Cancer Institute, 42:1101-1114,
   1969

b/ Used to estimate risk
                                 (32)

-------
in the Innes study was 5 times greater  (Albert, 1978).  CAG explained

that animals in the Innes study were fed the conpound beginning at a

younger, more susceptible age.  In addition, GAG concluded that the Innes

study data were more appropriate  for the risk calculations because the

human response is more likely to  be similar to  the most sensitive animal

species and because of the possibility  that people will be exposed to

chlorobenzilate as infants.

       Human risk is defined mathematically as  the probability that an

individual exposed to chlorobenzilate will develop a tumor due to that

exposure during his or her lifetime.  To develop a risk estimate, CBG

and ftgency consultants evaluated  the animal test data and the human ex-

posure data using several different models.  They selected the one-hit

model as providing the most conservative estimate.   This model relates

the probability of the development of tumors  in hunans to the exposure

and animal test data as  follows:

                                   -(Bx)
                         Risk - 1-e

                            B >

          ftere:
                 B
                 P
                 PC
                 Pt
slope coefficient of the one-hit model
(Pt-Pc)/(l-Pc)
Incidence of hepatonas in control animals
Incidence of hepatcrnas in test animals
Test animal exposure (ppm)
Potential human exposure
2/  The linear and one-hit models were both used to calculate risk po-
     tential (Felkner,  1978).   However, the one-hit model projected the
     most risk and was  therefore chosen as the more conservative basis
     of projecting potential risk from the Innes data.   The Log-Probit
     model was considered inappropriate for estimating  risk rron the
     results of the Innes study (Felkner, 1978).
8/  The slope B can be derived from the general expression for the
     one-hit model which is: risk - 0 + 
-------
The risk estimates are summarized in Tables 6, 7, and 8.
          2.  Risk;  Adverse Testicular Effects

       The primary routes of exposure to chlorobenzilate are  in the diet and
during spray application.  The following estimates are based  upon analyses
by the Criteria and Evaluation Division.
               a.  Dietary Exposure

       As previously determined, the average  human exposure to chlorobenzi-
late from the diet  is 0.0038 rag/day for the general population and 0.0095
mg/day for the Florida population (Severn, 1978).  For a 70 kg male,  this
converts to daily dietary equivalents of 0.002 ppm for the general popula-
                                           s/
tion and < 0.006 ppm for the Florida group.   Because human exposure  from
the diet is a lifetime possibility, it  is appropriate to use  a whole  life
feeding model for risk calculations.  Comparing  the estimated exposures to
a no-observable effects level  (NOEL) of 40 ppm  (See Section II,A,3),  indi-
cates that the margin of safety  is approximately 15,000 for men exposed to
chlorobenzilate in  the diet  (Potrepka, 1978a).

               b.   Occupational  Exposure

       Maximum exposure to unprotected spray applicators has  been estimated
to be 13 to 41 mg/day  (Table 4).  A direct  comparison between applicator ex-
posure and a NOEL based on daily exposure over  an entire life span may not
be appropriate because of the  non-continuous nature of applicator exposure;
 9/  See Footnote 5.

                                   (34)

-------
                                  Table 6


                              CHLOROBENZIIATE

                        POTENTIAL OCCUPATIONAL CANCER RISK
                                                              S/
               GROUND APPLICATORS (AGRICULTURAL - CITRUS ONLY)
                                                                         S/
                                              by                  Maximum
                                       Maximum            Mathematical Expectation
                                Lifetime Probability        of Numbers of Tumors
                                 of Tumor Formation          Djring a Lifetime
                               	(Less Than)	    	(Less Than)	

One-hit Model (MCI Data)

  Observed                      80 to 300 in 1 million            0.1 to 0.2


One-hit Model (Innes Data)


  Observed                     400 to 1400  in 1 million          0.3 to 1.0
a/ There  is  insufficient data  to  allow estimates  of  risk to aerial  ap-
   plicators, non-citrus applicators or harvesters of the treated crops.

b/ Assumes that ground applicator's daily dietary exposure to chloro-
   benzilate is 0.39  to 1.3 ppm.  The one-hit model  is a conservative
   technique for projecting risk  from laboratory  animals to man.

c/ Since  lifetime  animal studies  were used  to make risk estimates,  the
   probability of  cancer in humans  is estimated as a lifetime proba-
   bility, and should, therefore, be  interpreted  as  an index or "mathe-
   matical"  expectation rather than a "clinical"  expectation.

d/ In addition to  normal spontaneous rate;  estimate  based on NCI  study
   male mice with  hepatocellular  carcinomas [32 cut  of 48 (treated); 4
   out of 19 (controls)] (Albert, 1978b).

e/ In addition to  normal spontaneous rate;  estimate  based on Innes
   Study  "Strain X" male mice  with hepatcmas [9 cut  of 17 (treated);
   8 out  of  79 (controls)]  (Albert, 1978).

f/ Estimate  used.

                                  (35)

-------
                                      Table 7
                            CANCER RISK THROUGH DIETARY. EXPOSUPE


                          FLORIDA POPULATION  (8,000,000)

                                                                       t/
                                            a/                  Maximum
                                     Maximum            Mathematical Expectation
                              Lifetime Probability        of Numbers of Tumors
                               of Timor Formation           During a Lifetime
                                  (Less Than) _    _ (Less Than) _
                        <•/
One-hit Model  (NCI Data)

  Observed                    0.5 to 1.2  in 1 million            4 to 10

                          3/
One-hit Model  (Innes Data)


  Observed                    2.7 to  6.5 in  1 million          22 to 52
a/ Assumes that dietary exposure occurs  at  the  level of exposure  expressed
   as reasonable upper limit  (0.0025  to  0.0061  ppra daily  throughout life-
   time) , and model projects  conservative expression of risk.

b/ Since the  animal study was conducted  throughout lifetime  exposure,  the
   chance of  cancer occurrence  is extrapolated  as the potential of a cancer
   event during a  lifetime, and should,  therefore, be  interpreted as an in-
   dex or "mathematical" expectation  rather than a "clinical"  expectation.

c/ In addition to  normal spontaneous  rate;  estimate based on NCI  study male
   mice with  hepatocellular carcinomas [32  out  of 48  (treated); 4 out  of 19
   (controls)] (Albert, 1978b).

d/ In addition to  normal spontaneous  rate;  estimate based on Innes study
   "Strain X" male mice with  hepatomas [9 cut of 17  (treated); 8  out of 79
   (controls)} (Albert, 1978).

e/ Estimate used.
                                   (36)

-------
                                    Table 8
                  POTENTIAL CANCER RISK THRCUGH DIETMCf EXPOSURE
                   O.S.  POPUIATION (EXCEPT EDORUA) (212,000,000)
                        s/
One-hit Model (NCI Data)

   Observed
                          *
One-hit Model (Innes Data)
                                           a/               Maximum
                                    Maximum          Mathematical Expectation
                              Lifetime Probability    of Numbers of Tutors
                               of Tumor Formation      During a Lifetime
                                  (Less Hum)        _ (Less Than) _
                                               ~
0.4 in 1 million
2.1 in 1 million
                                                               86
                                                              445
a/ Assumes that dietary exposure occurs at the level of exposure expressed
   as reasonable upper limit (0.0025 to 0.0061 ppra daily throughout life-
   time) , and node! projects conservative expression of risk.

b/ Since the animal study was conducted throughout lifetime exposure, the
   chance of cancer occurrence is extrapolated as the potential of a cancer
   event during a lifetime, and should, therefore, be interpreted as an in-
   dex or "mathematical" expectation rather than a "clinical" expectation.

c/ In addition to normal spontaneous rate; estimate based on MCI study male
   •ice with hepatocellular carcinomas [32 cut of 48 (treated); 4 out of 13
   (controls)] (Albert, 1978b).

d/ In addition to normal spontaneous rate; estimate based on Innes study
   "Strain X" male mice with hepatomas [9 out of 17 (treated); 8 out of 79
   (controls)] (Albert, 1978).

e/ Estimate used.
                                (37)

-------
similarly, use of a NOEL from a subacute study does not consider the fact
that applicator exposure may be repetitive over most of the life span.
Assuming the analysis employed in the estimation of onoogenic risk ade-
quately adjusts for the difference between continuous and repeated ex-
posure, a dietary equivalent of 0.39 to 1.3 ppm would be derived for a
70 kg man (Albert, 1978).
       Based upon a NOEL of 40 ppm, the margin of safety for unprotected
spray applicators would range from 55-169.  The values given for margin
of safety were calculated based upon a comparison of approximate dose
levels (mg/kg) rather than dietary concentration (ppm) (Potrepka, 1978).
Use of either figure would result in a margin of safety within the same
order of magnitude.
       If the assumption is made that the subacute model is more analo-
gous to applicator exposure, no correction would be made for the tine
span of the exposure, and the applicable NOEL would be 500 ppm.  Calcu-
lations based on this assumption would yield an estimated maximum dietary
equivalent of 6.7 - 2.1 ppm.  Using the subacute NOEL, 500 ppm, as a
basis, the margin of safety would range from 43 -135.
                                    (38)

-------
                      10/
III.  Benefit Analysis
     A.  Introduction

       This section summarizes the benefits of the principal  uses  of
chlorobenzilate.  The sunnary identifies the uses of  the pesticide; es-
timates the quantities used; identifies and evaluates the  registered al-
ternatives, their availability and their costs; and evaluates the  consa-
quences of cancelling chlorobenzilate for these uses,  including  the pro-
jected in^acts on crop production costs and retail food prices,  this
information is derived in part from  rebuttal submissions.

     B.  Uses

       For purposes of discussion, the uses of chlorobenzilate may be
grouped into two categories, citrus  and non-citrus uses.   Table  1  presents
the complete usage pattern  for cnlorobenzilate in the United  States.
       -  1.  Citrus Uses
       The roost extensive use of chlorobenzilate  is  to control mites on
citrus crops, principally oranges, grapef."  ;, and lanons. The  major
target pests are the citrus rust mite on oranges and grapefruit  and  the
citrus bud mite on lemons.  In 1975, chlorobenzilate use on these  three
crops accounted for approximately 920,000 pounds of  active ingredient;
other citrus uses (limes, tangelos,  tangerines, other specialty  citrus
10/ This  section  is  based  on analyses prepared by M.  Luttner and
     M. McWhorter, Criteria and Evaluation Division, OFF,  EPA.
     reference1:  ire from Luttner unless otherwise noted.
                                 (39)

-------
fruit) accounted for 76,000 pounds.  In total, citrus usas accounted for
89.2% of all chlorobenzilate used in the United States.
     About 50% of U.S. citrus acreage, over 620,000 acres and 11,400 farms,
is treated with chlorobenzilate.  Florida citrus growers use chloroben-
zilafe most extensively.  Two-thirds (523,000 acres) of the Florida acreage
used to grow oranges, grapefruit, and lemons is treated with chloroben-
zilate, accounting for 72% of the total chlorobenzilate used in  the United
States.  Approximately half (43,000 acres) of the Texas acreage  is treated
with chlorobenzilate, which accounts for 9% of the total chlorobenzilate
used  in the United States.  Only 1.6% of the California citrus acreage  is
treated with chlorobenzilate, which accounts for under 1% of the total
chlorobenzilate used  in  the United States.  Another 7% of chlorobenzilate
used  in the United States  is applied to limes, tangerines, tangelos, and
other specialty citrus crops.
               a.  Florida
      In Florida, chlorobenzilate plays an  integral part  in established  cit-
rus  integrated pest management  programs.  At present  these programs  are
directed at controlling  the principal pests of Florida citrus, the citrus
rust mite and the citrus snow scale  (Brogdon, 1976).  Chlorobenzilate  is
recortmended for use  in  these programs because  it controls citrus rust
mites without harming the  natural predators and parasites of  the scale
insects and because  it  is  cost-effective.  Florida  citrus IPM programs
have reduced previously important pests, such  as puple  scale  and Florida
red  scale, to relative  insignificance  through  the  introduction and estab-
lishment of parasites on virtually all of  the Florida citrus acreage.
                                  (40)

-------
Through the IPM program, another parasite, the Hong Kong wasp,  is being
introduced to control snow scale.
               b.  Texas
       The principal pest in Texas citrus production  is the  citrus  rust
mite (French et al., 1978).  As in Florida, chlorobenzilate  is  effective
in controlling the citrus rust mite while preserving  beneficial insects
inportant in the control of chaff, California red, Florida red, purple,
and brown soft scales.  Seventeen species of beneficial insects were  re-
leased in Texas citrus production areas during the period  1952-1960 (Cocke
et al. 1978); as a result, beneficial  insects have provided  significant
control of purple and Florida red scale (Dean, 1955,  1975).
       The citrus mealybug is also regarded as a potentially major  pest
that is currently being controlled in  Texas by species of  lady  beetles and
brown lacewing.  Field experimentation indicates that the  lady  beetle can
effectively control citrus mealybug  in orchards  treated with chlorobenzilate.
       Extension education programs  backed by citrus  IPM research at the
Texas Agricultural Experiment Station, Texas Ail University, and the USDA-
Subtropical Texas Area Citrus Insects  Laboratory are  helping Texas  growers
to become aware of and to adopt citrus IPM control strategy. Because of
the growing acceptance of IPM, a Texas Citrus Pest Management Program that
will include  insect, mite, and disease control is being developed  (Cocke
et al. 1978}.
 ll/ Citrus  IPM in Texas is currently on a less formal basis than citrus
    IPM is  in Florida.   However,  beneficial insects established through
    releases  provide  biological control cf scale pests.   In Texas as m
    Florida,  the use  of selective irdticides like chlorobenzilate protects
    these beneficial  insects,  requiring less use of broad-spectrum scali-
    cides than would  otherwise be required.
                                 (41)

-------
               c.  California
       In California, chlorobenzilate  is used on an  "as needed" basis
(alone and in combination with oil} to control citrus bud mite on lemons
and citrus rust mite on oranges.  About one eighth (approximately 5,000
                                                     12/
acres) of the lemon acreage  in the southern counties  is treated with
chlorobenzilate in any given year; approximately one tenth  (3,300 acres)
of the orange acreage in the sane area is  treated annually  (USDA, 1977).
Integrated pest management programs in the southern  counties  are  u&ed to
varying degrees for control of California  red, purple, and  black  scales,
and brown soft scales, aphids, and mealybugs.  Chlorobenzilate  is used
in these programs because it is compatible with the  use of  the Aphytis
parasites to control California red scale, and has no adverse effect on
the natural predators of mites and other pests  (Jeppson, 1959).
               d.  Arizona
       Three major mite problems occur annually on citrus crops grown  in
Arizona - citrus red rrlte, citrus flat mite and Yuna spider mite.  A fourtli
species, the Texas citrus mite, is an  occasional problem in local areas
(Luttner, 1977a).  The citrus red mite is  principally a pest  specific  to
lemons in Arizona, while the other mite species affect all  of the citrus
crops.  Chlorobenzilate is used by Arizona growers for the  same  reasons
it is used elsewhere, i.e.,  its selectivity for mites and negligible
effects upon beneficial insects.  Approximately 5% (3,000 acres)  of the
12/  The  southern California counties in question include the following:
     Imperial,  Los Angeles,  Orange,  Riverside,  San Diego, Santa Barbara,
     and  Ventura.
                                 (42)

-------
Arizona acreage used to grow oranges, lemons, and grapefruit (principally
lemons) is treated with chlorobenzilate  (Table 1).
          2.  Non-Citrus Uses
       Cotton use accounts for 39,000 pounds or  5% of the total chloroben-
zilate used in the United States (1975).  This ancunt was used on 39,000
acres of cotton or 0.4).% of the total U.S. cotton acreage.  Non-citrus
fruits and nuts (apples, pears, cherries, aironds, and walnuts) account for
81,000 pounds or 7.3% of the total chlorober.zilate used  in  the United States
(1975).  This amount was used on 24,000  acres of fruits  and nuts or approxi-
mately 1% of the total U.S. fruit and nut acreage.
       There are other registered uses of chlorobenzilate,  including melons,
                              13/
ornamentals, boats, and docks.    Little chlorobenzilate is applied for
these uses  (OSDA, 1977).
13/ The registered miticide uses of chlorobenzilate other than the  citrus
    use,  the cotton  use,  an?  the uses on fruits and nuts are:
 agricultural crops      - melons  (c&saba,  cantaloupes,  crenshaw,  honeydew,
                          Persian);
 ornamentals             - (lawns and turf) - grass;
                         - (herbaceous plants and bulbs) - aster,  carnations,
                          chrysanthemums,  gladioli, iris, marigold, phlox,
                          snapdragon, zinnia;
                         - (woody shrubs, trees and vines) - arbovitae,
                          azaleas, birch,  boxwood, camellia, Douglas  fir,
                          elm,  hawthorn, hemlock, holly, juniper, lilac,
                          locust, maple, oak, ornamental shrubs,  ornamen-
                          tal trees, pine, poplar, rhododendron,  roses,
                          spruce, willow yew;
 domestic dwellings;    - (areas other  than edible-product areas) -
 medical  facilities  &    outdoor areas, boats,and docks.
 schools; commercial
 establishments
    Source:  EPA Compendium of Registered Pesticides (U.S. Environmental
    Protection  Agency,  1973).
                                 (43)

-------
     C.  Alternatives to Chlorobenzilate

          1.  Citrus Alternatives
       If chlorobenzilate were not available for use  as  a miticide,  citrus
growers would face a choice of using no miticides or  using miticide  alter-
natives to chlorobenzilate.  If growers in Florida, Texas, and  Arizona  de-
cided to use an alternative niticide, they would face the decision whether
to use one of the two selective alternatives or one of the non-selective
alternatives.  California growers would not have the  option  of  choosing a
selective alternative.
               a.  Ho Miticide
       Uncontrolled, mites affect fruit size, appearance, crop  yield, and
tree stock stamina (Table 9).  Fruit size and appearance are important  for
the fresh-fruit market because of consumer preference.  Approximately 57%
of the citrus fruit grown  in Texas goes to the  fresh  fruit market.  In
                                                                            W
California,  52% of the lemons and 65% of the oranges  go  to  the  fresh market.
While only 5% of the Florida citrus crop goes to the  fresh  fruit market,
growers cannot identify the  fruit which goes to that  market  until the end
of the season.  Since the  fresh fruit market is more  lucrative  than the
process market, growers try  to produce  for this market by protecting the
appearance of their fruit.
       Studies have shown  (Allen, 1978) that uncontrolled mites cause re-
ductions  in  fruit size of  12% for oranges and 17%  for grapefruit.  Fruit-
size declines also occur  in  lemons, but these effects have  not been fully
 14/ State-wide average  (Luttner,  1973a).

                                 (44)

-------
                                     Table 9
ESTIMATE OF THE MAXIMUM ECONOMIC VALUE LOST 1
AS A RESULT OF UNCONTROLLED MITE INFESTATIONS IN
y
State Crop
CA
FL
TX
Total
lemons
oranges
g-fruit
oranges
g-fruit
Per Year
Average
Production
(tons)
735,000
8,119,000
2,057,000
248,000
405,000
Loss Due
To Drop
(tons)
183,750
811,900
205,700
24,800
40,500
Loss Due To Total
Size Reduction Loss
(tons) (tons)
	 183,750
906,080 1,717,980
314,721 520,421
27,677 52,477
61,965 102,465
CITRUS
Value
($7t5n~)
116
58
57
52
51
/
Total Value
3/ of Production
Loss
$ 21,315,000
$ 99,643,000
$ 29,664,000
$129,307,000
$ 2,729,000
$ 5,226,000
$ 7,955,000
$158,577,000
I/ Luttner, (1978a)



2/ No estimate of inpact oould be derived for Arizona



3/ 3-Year average based on USDA statistics
                                      (45)

-------
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 (Table 9).  Total grower gross revenues fron sales of cit-
rus crops approximate $1 billion per year.  Finally, failure to control mites
reduces the life span of citrus trees, causing further economic impacts.
               b.  Alternative Miticides;  Florida and Texas
       Chlorcbenzilate is used widely because its selectivity makes it
compatible with integrated pest management programs using predators and
parasites of pests other than mites, principally scale insects.  Such IPM
programs provide inexpensive control of a number of major pests.  Accord-
ingly, assessment of pesticide alternatives to Chlorobenzilate must focus
on the economic consequences of replacing chlorebenzilate with a selective
miticide, compatible with these integrated pest management programs, or
with a non-selective miticide, which would disrupt those programs.
                     i.  Selective Miticides
       Only two registered  selective miticides    are potential  chloroben-
zilate alternatives  in Florida  and Texas:  dicofol  and  fenbuta tin-oxide.
IS/ Several pesticides  unregistered  for  citrus  uses may be useful  as
    Chlorobenzilate  alternatives.  Diflubenzuran (under pre-HPAR review
    by the Agency) has  performed adequately in  pre-development  testing.
    Hirsutella, a naturally-occurring  fungal disease  of mites,  has
    shown effectiveness under some conditions (McWhorter,  1978).   The
    Agency is  not evaluating  unregistered potential alternatives in
    this document.
16/ The Agency has conducted  a preliminary  risk assessment of dicofol
    and fenbutatin-oxide (see Appendix C).   In  broad  summary, dicofol
    appears  to be an oncogen  based upon  preliminary results of  an NCI
    study, while fenbutatin-oxide appears to cause reproductive effects
    and may  pose other  problems.
                                 (46)

-------
Both pesticides are included in the respective State recoimendations  for
mite control (Luttner, 19T7a, 1978b).  However, each poses  substantial
problems as an acceptable substitute  in IPM programs.
       Total per-acre treatment costs with dicofol are  approximately  33%
higher than total per-acre treatment  costs with chlorobenzilate (Table  10).
If dicofol is used in groves infested with snow scale  (approximately  75%
of Florida groves), the snow scale populations increase, causing serious
infestation and necessitating the use of  scalicides  (Florida Cooperative
                         12/
Extension Service, 1977).
       Growers using dicofol in place of  chlorobenzilate  in groves infested
with snow scale would need to supplement  dicofol  applications with scalicide
applications.  However, since snow scale  is not a citrus pest in Texas, it
is unlikely that scale problems  in Texas  would be aggravated by the use of
dicofol.  The per-acre cost of treatment  with dicofol  is approximately
$32.67, compared to $24.62 per-acre using chlorobenzilate.
       Fenbutatin-oxide  (marketed under the trade name Vendex) is a selec-
tive miticide which does not cause proliferation  of  snow scale or other
pests  (Codec et al. 1978).  However,  feiVoutatin-oxide  is expensive in com-
parison with chlorobenzilate and other alternative miticJdes.  In Florida,
the per-acre cost of treatment with fenbutatin-oxide ranges from $63.31 to
IT/ The observed  field sex ratio of snow scale is normally 5 males to 1
    female.  Applications of dicofol have increased average populations
    approximately three-fold.  Additionally,  the sex ratio of the F.
    progeny  is approximately 3 females to 1 male, thus greatly expanding
    the population's potential for increase (Brooks, 1973; Huffaker,  1978).
                                (47)

-------
                                    Table 10
          PER-ACRE TREATMENT POSTS IN FDORIDA CITRUS WITH CHL3ROBENZILATE;
                                                                 !7
                    SEI£CTED ALTERNATIVE MITICIDES AND SCALICIDES
             Pesticide

Mites    Chlorobenzilate 4E
         Dioofol 4MF
         Ethion 4E
         Sulfur 95%
         Ethion 4F & Oil 97%

         Oil 97%
         Fenbutatin-Oxide 50WP

Scales   Ethion 4E
         Parathion 8E
         Carbophenothion 4E
         Oil 97%
  Cost ($)

 16.00/gal
 17.40/gal
 13.50/gal
120.00/ton
 13.50/gal
    $l/gal
  1.00/gal
 11.65/lb

 13.50/gal
 16.00/gal
 13.50/gal
  1.00/gal
                                           y
  Material
Cost/f-rre ($)

   5.00
  13.05
  12.66
   3.00
  12.66 + 6.00
  18.66
   8.00
  43.69

  12.66
   5.00
  12.66
  10.00
Total Cost/Acre
 Treatment ($)

    24.62
    32.67
    32.28
    22.62
    38.28

    27.62-,
    32.28
    24.62
    32.28
    29.62
I/ Selected on basis of use or potential  for use.

2/ Material costs as reported for Florida by the USDA Chlorobenzilate
   Assessment Team.

3/ Material costs per acre based on Florida costs  and application rates
   specified in the Florida Citrus Spray  and Dust  Schedule.

4/ Total costs include application costs  per acre  of $19.62,  which repre-
   sents typical Florida costs for a dilute (1,000 gallons spray/acre)
   spray treatment with tractor-pulled  air-blast equipment.

5/ The total cost/acre treatment would  increase to $69.87  if  6 pints of a
   surfactant were added per acre.
                                      (48)

-------
                                                       IB/
$69.87 (depending upon whether a surfactant  is added);    in Texas,  the per-

acre cost of treatment with fenbutatin-oxide  ranges  from $40.01  to  $46.57.

The per-acre cost of treatment with chlorobenzilate  is $24.e>2 (Table  10).

       Another problem associated with  the use of  fenbutatin-oxide  in

Florida (but not Texas)  is its incompatibility with  oil.  Fenbutatin-oxide

and oil are phytotoxic when applied together or within 30 days of one
        19/
another.    Oil is the treatment of choice for greasy spot, a fungal  di-
                                                                 20/
sease which is one of the major pest problems of Florida citrus.   Mite

problems and greasy spot problems  frequently occur at the sane time.   If

it is passible to delay  treatment  of these pests for 30 days, then  fen-

butatin-oxide and oil can both be  used  for control of greasy  spot disease

and mites; however, the  use of fenbutatin-oxide  in the spray  program  would

increase grower costs, since  two separate treatments would be required in

place of a single chlorobenzilate  plus  oil treatment.  This increase  would

amount to the $19.62 per-acre application cost.   It  is frequently not pru-

dent, however, to defer  treatment  of greasy  spot or  mites for 30 days.  In

such situations, fenbutatin-oxide  could only be  used if seme  other  pesti-

cide could be substituted for oil  to control greasy  spot disease.  If
 IS/ Because fenbutatin-oxide does not act as rapidly as chlorobenzilate,
     the use of  fenbutatin-oxide may require the addition of a surfactant
     to produce  results equivalent to those obtained with chlorobenzilate.

 19/ The phytotoxicity problem ocoirs primarily with immature fruit and
     foliage. However, since all trees in a grove may contain both mature
     and immature fruit and/or foliage at any one time, phytotoxicity is
     appropriately treated as a generic problem.

 20/ Other pesticides registered and recommended by the State of Florida for
     control of  greasy spot disease are:  di Cola tan, benonyl, copper, oil.


                                 (49)

-------
       21/
benonyl   were substituted for oil, the additional cost  to growers would

be $8.00, the difference between the material cost of bencrryl and the ma-

terial cost of oil.

       Further, while fenbutatin-oxide can substitute for chlorobenzilate on

oranges and most grapefruit varieties, it is not registered  for  the remain-
                                22/
ing chlorobenzilate citrus uses.    Fenbutatin-oxide has not been used widely

in the past, and it nay not be immediately available in  sufficient quantity

to be used as a substitute for chlorobenzilate.  There is no reason to be-

lieve, however, that the supply of fenbutatin-oxide would not  increase to

meet demand if that demand were created by cancellation  of chlorobenzilate.

                                                23/
                    ii.  Non-Selective Miticides

       Certain of the non-selective alternatives to chlorobenzilate —

ethion and ethion plus oil — can control mites as effectively as chloro-

benzilate.  Another non-selective alternative, sulfur, does  not  provide

the level of mite control achieved with chlorobenz'late. Were these alter-
                                                    24/
natives repeatedly used  in place of chlorobenzilate,   the populations of
21/ Currently  under RPAR review.

22/ Fenbutatin-oxide  is  currently registered for all citrus fruit except
    tangerines,  tangelos, Heed grapefruit,  or Webb Red Blush grapefruit
    (Luttner,  1978b).
23/ The non-selective miticides for citrus  mite control are carbophe-
    nothion, ethion, prcpargite,  sulfur,  ethion plus oil,  carfoopheno-
    thicc; plus oil, ethion plus sulfur, oil, dicofol plus oil.   Only
    these materials judged to be major chlorobenzilate alternatives
    were evaluated  for their impact upon  beneficial insects (USDA, 1977).

24/ Non-selective miticides  are currently in use;  however, the  level of
    use is compatible with maintenance of beneficial insect populations.
                                 (50)

-------
the predators and parasites of the scale insects and other pests would be
reduced to levels incapable of providing economic control, necessitating
the use of combinations of chemical pesticides in large volumes (Table 11
and 12}.  In Florida, for example, the impact on existing IPM programs
would be severe, and it is expected that at the end of the fifth year after
cancellation of chlorobenzilate, growers using a non-selective alternative
would need to treat 100% of their acreage with one application of a miticide
and two applications of a scalicide (Table 11).  The per-acre cost of such
                                             25/
treatment would range from $72.00 to $103.00.    A similar impact could be
experienced by Texas citrus growers, many of when ?re using pest management
techniques similar to those used in Florida.
       The use of non-selective chlorobenzilate alternatives and widespread
scalicide treatments in Florida may have adverse effects upon fruit yield
or grade and tree vitality.  Adverse fruit quality effects may also occur
in Texas (over one-half of all citrus produced in Texas is utilized in the
fresh market).  However, no valid estimates on the amount of fruit that
would be damaged or unusable or the extent of tree injury are available to
evaluate this potential impact.
               c.  Alternative Miticides;  California
       There is only one chlorobenzilate alternative (oil) for control of
the citrus bud mite or the southern California lemon crop.  Further, only
one chlorobenzilate alternative  [wettable sulfur)  is recommended by the
State of California for control of the citrus ruse mite on southern
 25/ The cost range is  based on  the following  spray regimes:  sulfur,
    parathicn, parathion ($72.00)  and ethion  plus oil,  ethion,  ethion
     ($103.00).
                                 (51)

-------
                                       Table 11

     ESTIMMED EXTENT OF USE OF CHD3ROBENZILAIE ftND SEI£CTED SUBSTITUTES TO REPIACE
                                                                                     I/
  CHLOROBENZILATE FOR CONTROL OF THE CITRUS MITE COMPLEX AND SCALE INSECTS IN FLORIDA
                          Miticide Use          Scalicide Uses            Total Use
                       Year 1     Year 5       Year 1     Year 5     Year 1      Year 5


Chlorobenzilate       805,000     805,000       	       	       805,000     805,000



Fenbutatin-Oxide    2,415,000   2,415,OOC       —       	     2,415,000   2,415,000
Dicofol               418,000     418,000       	       	       418,000     418,000

Ethicn                979,000     979,000     319,000  1,594,000  1,298,000   2,573,000

Carbophenothion         	         	       319,000  1,594,000    319,000   1,594,000

^arathion               	         	       213,000  1.063,000    213,000   1,063,000

Sulfur              7,834,000   7,834,000       	        	    7,834,000   7,834,000

Oil (gals.)         1,949,000   1,949,000     850,000  4,250,000  2,799,000     6,199,000
I/  Note that all of the  listed materials  (except  fenbutatin-oxide) would  have  to
    be used  in  the quantities  indicated to replace chlorobenzilate.

2/  Quantity indicates potential use  level (see discussion at Section III,C,b,i).
    Fenbutatin-oxide poses  oil conpatibility problems (see discussion at Section
    III,C,b,ii).  In addition, there  are problems  concerning the availability of
    fenbutatin-cxide in the event  of  cancellation  of  chlocobenzilate  (see  dis-
    cussion  at  Section III,C,b,i).  Finally, fenbutatin-oxide is not  registered
    for sane Texas citrus uses (see discussion at  Section Til, C,b,i).
                                         (52)

-------
                                    Table 12



        ESTIMATED EXTENT OF USE OF CHIQROBENZILATE AND SELECTED SUBSTITUTES
                                                                               I/
     TO REPIACE CHLOBOBENZILKTE FOR CONTROL OF THE CITRUS MITE ODMPIEX IN TEXAS





                                    	Quantity Required
                                     Year 1             Year 5
Chlorobenzilate                     102,000             102,000
Fenbutatin-Oxide                     51,000              51,000
Dicofol
jofol
Ethion
Carbophenoth ion
Oil (gals.)
174,000
31,000
79,000
39,000
199,000
174,000
31,000
79,000
39,000
199,000
I/  Note that all of  the listed materials  (except  fenbutatin-oxide) would have to
    be used  in the quantities  indicated to replace chlorobenzilate.

2/  Quantity indicates potential use  level (fee discussion at Section  III,C,b,i).
    There are problems concerning the availability of  fenbutatin-oxide in the
    event of cancellation of chlorobenzilate  (see  discussion at Section III,C,b,i).
    Finally, fenbutatin-oxide  is not  registered for  seme Texas citrus  uses  (see
    discussion at Section III, C,b,i).
                                      (53)

-------
                                        Table 13






 ESTIMATED EXl«ff OF USE OP CHLOBUBENZIIATE COMPARED WITH AMOUNT OF OIL TOEAIMEMT



TO REPLACE CHLOHDBENZIIATE FOR CONTROL OF TOE CITRUS BUD MITE ON CALIFORNIA LEMONS










                                    	Quantity Required
                                     Year 1             Year 5
Chlorobenzilate                       7,500               7,500
Oil (gals.)                         493,000           2,465,000
                                          (54)

-------
                                26/
California oranges (USDA, 1977).    Oil and sulfur  are  not  completely

satisfactory substitutes for chlorobenzilate since  they are not  as effec-

tive and their use damages fruit quality and tree vitality.   However,  the

use of sulfur has an adverse effect on soil chemistry requiring  conpen-

sating soil treatments with line.  The per-acre  treatment cost of oil

and sulfur would be approximately $76.00 and $23.00, respectively, con-

pared with $61.00 for chlorobenzilate  (USDA, 1977).

               d.  Alternative Miticides;  Arizona

       Although chlorobenzilate use is quite limited  in Arizona  citrus

groves, the adoption of non-selective alternatives  on those acres where

chlorobenzilate is currently used would adversely effect the  endemic para-

site and predator populations  in a manner  similar to that described  for

Florida and Texas.

          2.  Non-Citrus Use Alternatives

       There are numerous alternative pesticides registered to control

mites on cotton and on fruits  and nuts  (Table  14).  Should  chlorobenzilate

be cancelled as a cotton miticide, 10 of  the 14  State-recaimended alter-

natives would have a lower pesticide cost  per  acre.  Even if  efficacious

control of the cotton spider mite could only be  achieved with alternatives

more expensive than chlorobenzilate, the  increased  pesticide  cost  to con-

trol mites on domestic cotton  would be minimal - at most approximately

$125,000 per year.  Under a similar  "worst case" assumption,  the increased
 26/ Registered  alternative treatments in use in Florida and Texas have
    been evaluated in California,  but because of phytotoxicity problems,
    conparitive ineffectiveness, or other considerations, they have not
    been included for grower use.
                                 (55)

-------
                              Table 14
          CHIOROBENZIIArE AND ALTERNATIVES FOR NDN-CITRUS USES
Conned ity
Pesticide
Pesticide Cost/Acre ($)
Cotton
Walnuts
Apples
                    Chlorobenzilate
                    Aldicarb
                    Carbophenothion
                    Dane ton
                    Dicofol
                    Dicrotophos
                    Disulfoton
                    Ethion
                    Methidathion
                    Methyl Parathion
                    Monocrotophos
                    Parathion
                    Phorate
                    Propargite
                    Sulfur
                    Chlorobenzilate
                    Carbophenothion
                    Dicofol
                    Dioxathion
                    Ethion
                    Oil
                    Phosalone
                    Propargite
                     Chlorobenz ilate
                     Carbophenothion
                     Cyhexatin
                     Dicofol
                     Ethion
                     Propargite
                     Tetradifon
                                  5.45
                                 12.64
                                  2.20
                                  3.80
                                  5.99
                                  C.71
                                  2.40
                                  3.44
                                 10.00
                                  0.99
                                  3.16
                                  1.12
                                  4.20
                                  6.04
                                  5.25
                                  3.18
                                  7.00
                                  6.22
                                  9.76
                                 11.68
                                  7.75
                                 14.22
                                  8.68
                                   2.C4
                                   3.50
                                   3.00
                                   2.92
                                   1.46
                                   2.17
                                   1.77
                               (56)

-------
U.S. pesticide cost to control mites on fruits and nuts would also be mini-

mal - at most approximately $69,000 per year.


     D.  Grower Impacts


          1.  Citrus Uses

               a.  Florida

       The economic impacts on growers associated with use  of selected

chlorobenzilate alternatives  are elaborated  in Tables 11, 16, and 19.

Based upon the information previously presented,  it  has been determined
                                                   27/
that the non-selective chlorobenzilate alternatives   would be  the pre-

dominant replacement materials  if chlorobenzilate were no longer available.

       It was also previously explained that the use of dicofol for mite

control would not have an adverse effect upon the beneficial  insects  that

control purple and Florida red  scale but would require additional use of

sealicides to control  increased populations  of snow  scale.

       In the short run, fenbutatin-oxide  is more expensive ($63.31/acre)

than the dilorobenzilate miticide alternatives  ($22.62 to $38.28 per  acre).

However, in  the long run, the use of chlorobenzilate alternatives and scali-

cides would  lead to per-acre  costs ranging from $72  to $103.  Thus,  in

strictly economic terms, fenbutatin-oxide  would appear to be  an attractive
 27/ While non-selective materials would predominate,  the economic assessment
     of  the  likely consequences of chlorobenzilate cancellation performed
     by  the  Agency (Luttner,  1977b) assumed that materials identified as
     likely  replacements by the USDA Assessment Team would be used in equal
     distribution.  In the case of Florida, on* such material was dicofol,
     which  is  evaluated separately in this discussion  front the standpoint
     of  its  suitability as a  total cnlcrobenzilate replacement.  Similarly,
      in  the  case of Texas, the analysis assumed that fenbutatin-oxide would
     be  utilized to sane extent.

                                 (57)

-------
                                               Table 15
                                                                                     y
          lapactn Projected to Growers Resulting frca Canoellation of Chlorobenzilate
         Present Value          Present Value            4/   % Change            5/ % Change
         of Change in            of Change in     Current    in Current    Current   in Current
    2/   Control Cost  Affected  Control Cost/ Pest Control  Pest Control Production Production
Site	($000)	Acres   Affected Acre   Cost/Acre    Cost/Acre    Cost/Acre  Cost/Acre


PL-citrus   40,079     850,000       $47           $108        443.5        $346      +13.6


CA-lenons    3,678      41,000       $90           $220        +40.9      $1,264       +7.1
I/  Source:  Luttner, 1977a.  Assumes non-selective alternatives are used to
    replace chlorobenzilate.

2/  Comparative data for Florida citrus is based on oranges, which comprise
    approximately 75% of the Florida acreage.

3/  Cost data represents fifth-year inpacts calculated to present values
    using a 7% rate of discount.

4/  Represents an average for Florida oranges and California lemons, including
    pest, disease, and weed control.  Based on budgets developed by Muraro and
    Abbitt (Florida) and Gustafson and Rock (California).

5/  Represents averages for Florida oranges and California lemons using budgets
    cited in footnote 13.  Includes cultural costs and management only.
                                     (58)

-------
                                    Table 16
           Prcjectec Cost of 5;alt Ir-sect Control  ir.  Fiorisa  Citrj; During
         Tnitia- Fivt "tar ?«iec Felixine Car'cellatior  of Chlora.-»enzUate
V
Scalicioe
Scalicide +
Arolication 3/
Cost/'Xcre- Acre-Trssr^ents/Tfear
Trsaaaentt?) Year 1 Year 2 Year 3 Y=ar 4 Year 5
Cariopherothion      32.2B        85,000     17C,000     255,000    340,000    425,000

EthiOl               32.28        85,000     17C.OOC     7.55,000    340,000    425,000

Oil (97%)            29.62        85,000     170,000     255,000    340,000    425,000

Para tt ion 8E         24.62        85,000     170,000     25;,OOP    34C.OOO    425,000

Tbtils               	         340,000     680,000   1,020,000  1,36C,000  : .700,000
Scalicide
Carbopherath ion
Ethion
Oil (971)
Parathion BE
Ibtali

Y*ar 1
2.744
2,744
2, .518
2,093
10,399
Cost of
Year 2
5.488
5,488
5,035
4,185
20,196
Cortroi.'Tsar
fear 3
8,231
8,231
7,553
_6,278
X.293
ISOOO)
Year 4
10,975
10,975
10,071
8,371
40,392

Year 5
13,719
13,719
12.589
1C, 464
50,491
I/  The  s;aiicides listed appear in tie 1*77 Florida Citr.s SiTay- and Dust
    Sc.VA.le.   TtK aszxiption U.at  t^°se material: would be widely used for
    scale control appears to be rusor^ble based on ejecting registrations,
    effectiveness, ard lou cant relative to other available scalicides.

2/  Costs baaed on tte following:   application rates as soecified ui the 1977
    Florida Spray arei Sust Schedule;  pefticide prices sufplied fcy the Assessment
    T»m; ar^>lication OD«t of SIS. 62 per *cr» represents us* of dilute spray
     (two 500 gallon tanks/acre ) as  developed in a current production budget for
    Florida (Kuraro a,-x! Aobitt, 1977 citaJ in Luttner, 1977b).
 3/  Adciticnal scale control treatnents are assured to be required on 100% of
     the Florida acreage Capprox.  850,000 acres) over a five year period.  Q»
     annual incremntal increase in acres rsguirirr truitirent is assured bo be
     equal.  Tra cate of increase  in affected acres represents the projected rate
     of spread of econor.ically darjcing scalg fooula: icus throughout the State.
     Tse of scalici3es on an equal tasis represents an assjiptior. by the analyst.
     Uthcugh chlorobenzilate is not used on lOCt of Florida citrus each year,
     the prelected IP* inlets will irvolve all of the State's acreage.  QUoro-
     benzilau's use pattern lb?»  of Florida acrtsce treatwl'year) indi^tas
     that it is used two out of every three years on the average acre for nut*
     CDTitrol, vitn the run»rous alt.err.at lues used in rotation djring the third
     year.  This oc=«si£nal rotation with alterratives (sane of which do not
     cause serious adverse effects upon beDeficial tis«cts) pearits the oon-
     tin;>ed success of the Tr"?! prcgrsn.  Bowever, oontLiaous use of the al-
     ternatives wxild eventually lead to development of scale control problem*
     on all of the Florida acreage.

 4/  Prcd'jct of scalicide + application ctst/acre-treaUrent tires acre— treatneuts
     for U* respective years.
                                     (59)

-------
alternative.  However, the profitability of the Florida citrus industry is
highly variable and in many instances has been only marginally profitable
or unprofitable as measured by return on investment (Brooke, 1973).  There-
fore, growers are likely to take a short-run view when selecting alternative
pesticides, even though the most economical long run alternative would be a
selective miticide.
       Moreover, there are additional reasons why fenbutatin-oxide is un-
likely to be adopted as an alternative to chlorobenzilate in Florida.  These
other reasons (both discussed earlier) are the oil incompatibility with
fenbutatin-oxide and the fact that fenbutatin-oxide is not registered for
some Florida citrus crops.
       In Florida, using the non-selective regime to control rust mites
would also require the use of two additional dilute scalicide treatments
per year to counterbalance the reduction of currently established scale
insect parasites to levels of population incapable of providing economic
control.  This phenomenon would occur on 100% of the commercial Florida
citrus acreage over a five year period.
       Using the non-selective regime would increase grower treatment costs
by $2,043,000 per year or $3.17 per  acre-treatment.  The subsequent use of
pesticides for scale  insect control  in Florida would increase grower costs
by $50,491,000  in the fifth year aftei cancellation, i.e., when all of the
Florida acreage would be receiving scale control treatments  (Table 17).
The  costs  could be expected to continue beyond the fifth year at the sane
relative level unless alternative scale control measures were developed.
                                 (60)

-------
                                    Table 17

               Economic Inpact of the Loss of Chlorobenzilate During
                 Initial Five Year Period Following Cancellation .
Year After
Cancellation
Area
Cost of Mite
Control with
Alternatives
   ($000)
             y
Economic Cost
of IPM iirpacts
   (SOOO)
Present Value of
Total Economic
Impact ($000)
1




2




3




4




5




AZ
CA
FL
TX
US
AZ
CA
FL
IX
US
AZ
CA
FL
TX
US
AZ
CA
FL
TX
US
AZ
CA
FL
TX
US
0
658
2,043
274
2757?
0
1,698
2,04.3
274
4,015
0
2,739
2,043
274
37015"
0
3,780
2,043
274
"STol?
0
4,821
2,043
274
77T3F
0
—
10,099

10,099
0
	
20,196

20,196
0
	
30,293

30 ,293
0
	
40,392

40,392
0
	
50,491

50,491
0
658
12,142
274
13,074
0
1,587
20,785
256
22,628
0
2,392
28,242
239
30,873
0
3,086
34,640
224
37,950
0
3,678
40,079
209
43,9&
 I/ Costs for Arizona,  Florida,  and Texas derived in the Preliminary
   Benefit Analysis  of Chlorobenzilate.   Assumes that selected miticide
   alternatives (Luttner,  1977a;  Tables  22,  23,  and 24) would be utilized
    in an equal  distribution on  all acreage currently treated with Chlorobenzi-
   late.  For a detailed discussion of the methodology utilized to derive the
   list of selected  alternatives* (Luttrier,  1977).

 2/ Assumes 2 additional  scale control treatments on 100% of Florida
   citrus acreage at year  5 and thereafter.

 2/ Sum of cost  of mite control  with alternatives plus cost of IPM im-
   pacts; present values calculated using a  7 percent rate of discount.
                                    (61)

-------
                  28/
The present value   of this change  in grower costs ($38,520,000) combined
with the present value o* the increased cost of chiorobenzilate alternatives
for mite control in Florida ($1,559,000)  is approximately $40,079,000 per
year after five years.  Given the annual  cost of the current Florida
citrus pest control program ($72.5  million), the impact  ($40.08 million)
projected to occur in the fifth year following a cancellation of chioro-
benzilate would represent a 55.3% increase in the annual cost of the
state's total citrus pest control program.
       The loss of chiorobenzilate  would  increase pest control costs to
Florida citrus growers by about $47 or 44% per acre after 5 years  assuming
widespread adaption of the non-selective  regime.  Average per-acre pro-
duction costs for Florida citrus are projected to increase after 5 years
by 13.6% (Table 15).
               b.  Texas
       In Texas, the use of chiorobenzilate alternatives for control of
citrus mites  is projected to  increase grower treatment costs by $274,000
per year or $4.72 per acre-treatment  (Table 17).
       Since  snow scale  is not a pest of  major importance on Texas citrus,
diccfol can be considered a satisfactory  replacement for chiorobenzilate
in existing pest management programs.  If dicofol were the sole alternative
used on those acres currently treated with chiorobenzilate, Texas  citrus
growers would incur additional mite control costs of approximately $432,000
per year or $7.32 per acre treatment.
     "Present value"  is an accounting  concept used to represent future
     monetary impacts at a cannon point in time.
                                 (62)

-------
        The per-acre cost of treatment with fenbutatin-oxide would in-
 crease from $24.62 to $40.01 (without surfactant) or to $46.57 (with sur-
 factant).   If all citrus acreage currently treated with chlorobenzilate
 were treated with fenbutatin-oxide,  the total annual cost increase to
 growers would range from $908,000 (without surfactant) to $1,295,000 (with
 surfactant).  Approximately 51,000 pounds of fenbatatin-oxide active in-
 gredient would be required to replace the 101,500 pounds chlorobenzilate
 currently used in Texas citrus (Table 12).
                c. California
         In California, the use of petroleum oil sprays to control citrus
 bud mites on lemons would require a 3 spray on all of the infested acreage
 and two sprays on two-thirds of the infested lemon acreage.  This would
 occur over a five-year period on all of the lemon acreage in the southern
 counties.  After five years, approximately 2,465,000 gallons of oil would
 be required to replace the 7,500 pounds chlorobenzilate active ingredient
 currently used on California lemons (Table 13).
        In California the cost of controlling citrus bud mites on lemons
»
 is projected to increase pest control costs to growers by an additional
 $4,621,00 per year in the fifth year after cancellation (Table 17).  This
 cost could also be expected to remain at the sane level unless alternative
 citrus bud mite control measures were developed.  The present value of the
 annual Inpact in the fifth year is approximately $3,678,000.  Given an es-
 timated cost of pest, disease, and weed control on lemons in the southern
 California counties of $9,020,000 per year, cancelling ohlorobenzilate
                                 (63)

-------
zilate would increase the cost of pest control on the affected lemon
acreage by about 40.8%.
       The loss of chlorobenzilate would increase pest control costs for
California lemon growers by about $90 or 41% per acre after 5 years.  Aver-
age per acre production costs would increase after 5 years by about 7.1%
(Table 15).
               d.  Arizona
       The loss of chlorobenzilate and adoption of alternative miticides
is projected to have no net cost to Arizona citrus growers.  Using alter-
natives may disrupt IPM strategies in Arizona, but the extent of such dis-
ruption has not been identified nor the resulting costs quantified.
          2.  Non-Citrus Uses
       The cost impacts to growers for non-citrus uses of chlorobenzilate
are discussed in Section C,2.

     E.  Costs to the Citrus Pulp Feed Industry

       Citrus pulp is a by-product of citrus processing; during the 1960's
citrus pulp became a principal feed ration ingredient for dairy cattle  in
Florida.  The majority (approximately 90%) of  the domestic pulp feed mar-
ket consists of sales to Florida dairymen  with sane sales to Florida beef
ranchers.  Approximately half of the 1,000,000 tons produced annually is
exported to European markets where it  is sold  for use as dairy cattle feed.
The use of citrus pulp as cattle feed produces gross revenues  of approxi-
mately $70 to $90 million to the processors; however, due to the large
amount of energy required to dry the pulp, net profits  to tbe  processors
                                 (64)

-------
ace not significant.  However, the use of citrus pulp for cattle feed has
solved a serious waste disposal problem for the citrus processors.  Prior
to the development of the citrus pulp-feed industry, the processing wastes
were routinely dunped and left to rot in citrus groves.  The utilization
of pulp as an animal feed ingredient has thus turned a net cost segment
of citrus processing into an outlet which provides sane (though limited)
net return to processors (Luttner, 1978g).
       Since there is currently no alternative use for citrus pulp, a more
costly disposal procedure (e.g., incineration or land-filling) would pro-
bably be necessary if citrus pulp were to be disallowed as cattle feed
(Luttner, 1978g).  Therefore, in order to protect the cattle feed ingre-
dient business as an outlet for the by-products of citrus processing, the
citrus processors would probably refuse to purchase chlorobenzilate-treated
fruit if a restriction is enacted which prohibits chlorobenzilate residues
in pulp.  The residue restriction would result in a de_ facto cancella-
tion of chlorobenzilate, leading to the citrus IPM impacts previously
discussed.

     P.  Costs to Consumers

          1.  Citrus Uses

       Consumers would be adversely affected due to higher prices end/or
fruit quality considerations only  if  the loss of chlorobenzilate leads  to
significant reductions  in yield and/or fruit grade.  However,  since yield
                                 (65)

-------
and/or quality changes are not  inroedlately projected  to occur,  there would
be little change in the quantity of citrus supplied.   Also,  the citrus in-
dustry is relatively stable since growers cannot elect alternative land
uses in the short run; therefore, a substantial supply response would  not
be expected in response to changes  in  cost.   Because  potential  yield/
quality effects are not indicated, and given the history  of  excess pro-
duction in the domestic citrus  industry  (which leads  to a relatively weak
                            29/
market position for growers)    the projected costs would  be  absorbed by
growers, at least in the short  term, with little or no consumer impacts
anticipated.
          2.  Non-Citrus Uses
       The analysis indicates that  the potential cancellation of chloro-
benzilate for use on non-citrus crops  would  have  insignificant  effects
upon growers, marketers, and consumers of these crops. The  reasons are:

          1)  Only small quantities of chlorobenzilate are used to
              control mites on  cotton  and a  wide variety  of  fruit
              and nut crops.
          2)  Numerous chlorobenzilate alternatives are both regis-
              tered and recommended for  use  on cotton, fruits,  nuts,
              and other crops.
          3)  Certain of the alternatives can be used at  a lower pesti-
              cide cost per acre.
       Should the cancellation  of chlorobenzilate  result  in  the use of
higher-cost alternatives on non-citrus crops, the  total estimated increase
 29/ This phenomenon is reflected by citrus cold storage figures, which
    relect the large stocks of citrus products carried over from one
    year to the next (OSDA, 1977).
                                 (66)

-------
in pesticide cost is $194,000 per year.  Of this total, cotton accounts
for $125,000; fruits, nuts, and other crops for approximately $69,000.
If the total non-citrus production cost-increase were ccopletely passed on
to final domestic consumers under the assumption of no reduction in yields,
the consumer impact would be minimal.
                                 (67)

-------
IV.  Risk-Benefit Analysis of Alternative Courses of Action



       Evaluation of the risk and benefit data suggests seven principal

regulatory options:



     A.  Continue Registration of All Uses.

     B.  Cancel All Uses.

     C.  Continue Registration of Chlorobenzilate Use on Citrus and Amend
         th3 Terms and Conditions of Registration; Cancel All Other Uses.

     D.  Cancel Chlorobenzilate Use on Citrus To Take Effect After Five
         Years, and in the Interim Amend the Terms and Conditions of
         Registration; Cancel All Other Uses.

     E.  Continue Registration of Chlorobenzilate Use on Citrus, Amend
         the Terms arid Conditions of Registration, Require That Identified
         Exposure Data Be Submitted to EPA in 18 Months; Reevaluate the
         Use on Citrus After Additional Exposure Data Becomes Available;
         Cancel All Other Uses.

     F.  Continue Registration of Chlorobenzilate Use on Citrus in Florida,
         Texas, and California, Amend the Terms and Conditions of Registra-
         tion, Require that Identified Exposure Data be Submitted to EPA in
         18 Months; Reevaluate the Use on Citrus After Additional Exposure
         Data Becomes Available; Cancel Use on Citrus in Arizona and All
         Other Uses.

     G.  Continue Reg.stration of Chlorobenzilate Use on Citrus, Amend
         the Terms and Conditions of Registration, Prohibit the Use of
         Pulp from Chlorobenzilate-Treated Citrus .is Cattle Teed; Estab-
         lish Complementary Tolerances; Cancel All Other Uses.
       Tables  18 and  19 summarize  the risks and benefits of each option.

The specific risks  and benefits pertaining to each option are described

below.
                                 (68)

-------
OPTIUN
                      none A err mil
                        COMSUHCI8
                                                                        T.bl. II

                                           UCULATOIY OPTIONS AMU NAXIHUU NISK INCIDENCE FROM CLOIOUiiZILATE USI
          or us
CITRUS CONSUMER
CITHUS PtSUClOe
  APPLICATOR
                                                                                        FLOIIUH ClTRUg
                                                                                          PICKKRS
NOM C1TMI8
USE CONSUMERS
    MOM-CITKUS           MOM-CITniS «
PESTICIDE APPLICATION       PICKEIS
A.  Continue  Re|iit ra-
    tion of All Ui«i

1.  Unc*l All U«e.
                          1.6 to *.»
                          MI nil lion
                                                  2.0
C.  Cunlinui  Keg lit ret ion   2.t to 6.4
    of Chlorvlienitlitf       ptl •illion
    lUei on Cltiui  and
          III* TCIHI  *i .a II
    Munllii; ««sy..l.i»i«  the Ul*
    on Cilru-. AH«i  A'MUlonal

    Av.iUhle; C.nc.l All
    Otber Dm*
                                             l*r
                                                   2.0
                                             per Billion
                                                  0.1
                                             ftr •illioa
                     400  to UOO
                     per  Million
                     4   to ?7S
                     per Billion
                     0.1  to  20
                     per  Billion
                                                                4 to lit
                                                                ptr  Billion
                        Mo Dill


                           0

                        No B«tl
                        Iteduod
                        Elpoiur*
                                                                                                             0.1
                                                                                                            pet Billion

                                                                                                               0

                                                                                                               0
                      No  0«t«


                         0

                         0
                            No D
-------
                                                                       (Table  1J> Cent 'o)
c.
Continue K«|iatration   2.6 to *>
of Chloroben.ilete       per •illion
Dee on Cicrua  in
Florida, Teiae, and
California; ARMIM!
TerMa ami I:OIH! i t i one  at
keg-iiic r«t ion,  M««|uire
ili II. Indentifieil Eiuoiure
llel* he •unfitted to  EPA
in 18 Hontha;  Keevalu>ta
tlie Uae on Citrua After
AJdilinnal F.«poaure uii Citrua  in Aricoita
•n.l Al I lltlier U.e«
                                W             W
I'xktitinue Megiatrat ion  of      2   per        2
CtiloiobanEilate II.«  on CU-   eilllicn    per Hillion
ro«; Aaienil tlie Teraia and Condi-
tion! ot Neiiiatralion;  Prohibit
the Uae o» |Milp  !•>•«• Chloroben^
xilate-Treateil Citrua  A< Cattle
feed; Initiate Action  for EPA
to Kateuliali a Toteraete on Chloro-
b«n
-------
                                        T.bl. 19
            ICONOM1C  IMPACTS  IESULT1NC  UGH CHLOROBEMZILATE KECULATOIY OPTIONS
Opt Ion
Cammotlti	Icgnomic l»|>«ct
                                                                                  II
A.  Continue ••glfttttloi  at  All  Iliti
Citrui
Cut'on
Fruit >/Nut•
                                                                   None
                                                                   Non*
                                                                   Nona
    C
-------
Cancel Cklorobaniila la  Uaa  on Citrua to
Taka (ffaet After  fi«e  Tear,  and
in tha IntariB  Aaand  the  Taraa and
Uondltioita  for  ka| la t rat ion;  Cancel All
Othar Uae*
(TAB1£ 19 Cbnf d)
Citrua Ho Inpict Year*
Are* IconoBic
AI
CA
?L
TX
Ub
' 0
400
1,700
200
9.300
1-S
lB)>act i
0
1.100
14.100
100
16.100
n Year Aftr Cancellation (SOOO)
0
l.«00
20.000
:oo
22.00U
0
1.100
24.600
100
17.000
0
2,500
18.500
300
31.300
Continv* lagiatratloa  of  Chlorobaniilate
Uae on Citrua, Aaand tha  Terae  and  Condi-
tion* for lagiitratlon, Require That  Iden-

tified tipuaura Data be aubaitted  to  EfA
in IS Montha; devaluate  tha  Uaa on Citrua
After Additional Knuoaura  Data  lecoaea
Available; Cancel All  Othar Uiea
Continua Ragiat rat ion
of Ch1orubans11 ata
Uaa on Citrua in

Florida, Taaaa, and
California; Araand
Tarna and Cor.ditioni of
Hafiat rat ion , Hai|uir«
that I iidant 1 f i«d Cmpoawra
l>al a b« aub»ittad  to EPA
in 14 Moulin, la«valua[a
tha (la* on Citrua Aftar
Additional Cipoaura Data
•acuaaa Avaliablaj Caiicrl
Uaa uii Citrua in Ariaona
and All Olliar Uaaa
                                                Cotton.
                                                fruica/Nuta

                                                Citrua
                                                Cotton
                                                Iruita/Muta

                                                Citrua
  and thereafter    -
94.4 to $ 19 . 1 Billion
  for aaended uaa   ..
  direction         —
»m.OOO/Yr
$69.000/Yr

 |4.4 to $19.1 Billion
   far ••enilcd u• •
   d ir«ct ion       3/

 PoCcncit.! for ttddl-
   tion«» i.p.ct,
   Jttpcndcnt  on teat
   reaulla

 JUS.OOO/Yr
 JfcV,000/Yr

 ?4.4 to 19.1 Billion
 for a«en«t«*d  uaa
 direction 3/

 Aricona - no econoaiic
 iBpact  Potential for
 addtlonal iaitact*
 dependent on teat
 reaulta
                                                  (72)

-------
                                                      (TAtlLE 19 COnt'd)
C.   Continue  lit'>t rat Ion of Ch lorobeni i 1 a t *
     Uae  on  Cicrue. Ae^nd the Ter«a  and  Cor-
     dltionl of la|tit rat ion, Ftohlbit th«  Ui«
     of Pulp froM Chloruhenti1*ta Treated  Citrue
     aa C»tcl« Feed; Initiate Action  tor  E»A  to
     Katiblleh a Tolerance on Ch lorobenai I • te
     Nariduu In Citrve Pulp; Cancel All  Other Uaea
                                                     Cotton
                                                     fruita/Muti

                                                     Hit rue
M 2*,OOO/Vr
9 (9,000/Tr

 Bcoaoaiie Inpact  In Year  After Cancellation ($000)
                         0
                    2,400
                   28,300
                      200
                   30,900
      0        0
    600     1.600
 12,200   10,«00
    300      200
 13,ODD   22,600

  and thereafter -'•-'
I*.* to tit.I
  mi 1 I ion for
  • •Mitded uee
  diraction
     0
 1.000
36,600
   200
38.000
     0
 3,700
40,000
   200
44,000
                                                                                       J/
I/  All  future  doll«F impact* arc (Ivan  in  preaant  valuei.  The citrua  can-
    cullatlon  laipa*-t  in year 1 < $ 13 f 000,000)  ia  of  tha aaaie na|aitude  aa  th<
    firat  y«ar  i«|>a>.c if it ia delayed  until  the aiitli year ($9,32 2, OO'j I .
    However,  th«  diicauntinc factuv Kaducaa  praaaiit value i>f tha aint'i-yaar
    Impact  ralatlva  to the firat-year  impact.

]_/  Aaaueiaa  no  ea«ab 1 1 ehaiant of viable  euualituta  coupatlbl« -j-ft- 1PH  «nj
    at coat  appraxinate1y equivalent to  ch1orob«nziI at«'a.

j[/  Protective  clothing coata for applicacora  uaa  aamaied to be negligible.
    (Kan  .Mure  ill  each).  Kucloeed  cab coata  rente fro* $4.230 (cab  pur-
    rliaiaf  t.) SH.aSO (ne« tractur with  cab).  Total  $4.4 li $lreaeura air filtration  ayatene.

4_/  Aaauaiaa  luaa  of  ftedetuff Market will cauaa  fr<'Wera to caaae uaini
    chlcrobeiK i I itu |  if not then laiae  laipac:  ai  Option C.
                                                          (73)

-------
     A.  Continue Registration of All Uses
       Adopting option A would indicate the Agency's conclusion that the
risks associated with each use are outweighed by the respective benefits
ana therefore, that none of the uses of chlorobenzilate cause unreasonable
adverse effects.  This option would return pesticide products which con-
tain chlorobenzilate to the registration process for reregistration.
       This option would not reduce the risk of cancer nor the risk of
testicular effects associated with the use of chlorobenzilate.  The po-
tential lifetime risk of cancer fron all uses would remain at 2.1 in one
million for the U.S. population; at 2.7 to 6.5 in one million for the
Florida copulation; and at 400 to 1400 in one million for ground appli-
cators of chlorobenzilete (Table 18).  The ground applicators' margin of
safety from testicular effects would remain in the range from 43 to 169.
       There would be no adverse economic impacts associated with this
option.  This option would retain the usefulness of chlorobenzilate as a
cost-effective tool for control of the mite complex, as well as the exist-
ing benefits from its use in citrus integrated pest management programs.
       The choice of this option would indicate the Agency's willingness
to tolerate a level of risk greater than the levels of risk estimated for
the other options (Table 18) because there are no adverse economic effects
(Table 19) or loss of other benefits.
     B.  Cancel All Uses

         Adopting option B would  indicate the Agency's conclusion that the
risks associated with each use outweigh the respective benefits and thereby
                                 (74)

-------
result in unreasonable adverse effects. This option would eliminate all uses
of chlorobenzilate.
       Cancellation would eliminate all the risks associated with the use
of chlorobenzilate (Table 18) but at great cost to citrus growers.  These
costs are based on the assumption that no alternative miticides as effective
as chlorobenzilate nor as compatible with citrus integrated pest management
programs would be developed and marketed at competitive oast.  In the fifth
year after the cancellation of chlorobenzilate, the citrus industry's addi-
tional pest management control costs would stabilize at $44 million per
year.  Florida growers would  incur 90% of the  increased cost generated by
using chlorobenzilate substitutes (Table 19).
       In addition, the switch to alternatives may  involve undesired risX
consequences.  To obtain the  degree of pest control currently obtained in
citrus integrated pest management programs using chlorobenzilate, increased
amounts of these substitute chemicals would be required (Table 9).  At the
present time, the safety data on several of these substitutes  (carbopheno-
thion, ethion, sulfur, and propargite) is incomplete.
       The annual economic impacts of cancelling chlorobenzilate's non-
citrus uses are relatively small, ranging from $£9,000 on fruits and nuts
to $125,000 on cotton.
       The choice of this option would indicate the Agency's unwilling-
ness to tolerate the level of risk associated with  all uses, based on a
conclusion that the benefits  do not outweigh those  risks  (Tables 18 and
19).
                                 (75)

-------
     C.  Continue Registration of Chlorobenzilate Use on Citrus and Amend
         the Terms and Conditions of Registration; Cancel AIlTfeher Uses"


       Option C would indicate the Agency's conclusion that the benefits

arising fron the use of Chlorobenzilate on citrus in Florida, Texas, Cali-

fornia, and Arizona exceed the risks (reduced by amending the terms and

conditions of registration) estimated to exist from the use of chloroben-

zilate on citrus in each of these four States.  Option C would indicate,

moreover, that the risks associated with the non-citrus uses of chlorcben-

zilate are not outweighed by these benefits.

          1.  Economic and Environmental Considerations

       The economic and other inpacts of cancellation of Chlorobenzilate

for use on Florida, Texas, Arizona, and California citrus are discussed

in detail in Part III of this document.

       In general, it is likely that cancellation of the uses of  Chloroben-

zilate on citrus in Florida would result in extensive use of non-selective

mitic-des - principally organcphosphates - and consequent destruction  of

existing IPH programs.  Per—acre treatment costs would  increase sharply -

from $24.62 to between $72 and $103  in the fifth year after cancellation

and thereafter.  The aggregate cost  impact of the expected use of non-

selective pesticides could range from $40.3 million to  $66.6 million annu-

ally in the fifth year after cancellation and thereafter.  In addition to

inposing a severe cost  inpact, the use of non-selective organcphosphates

would  involve a substantial  increase in  the volume of pesticides  used

which  could cause untiesired adverse  effects,  including  adverse health  im-

pacts.  ?enbutatin-oxide, a selective raiticide  compatible with IPM pro-
                                 (76)

-------
grans, could be used to sane extent but fenbutatin-oxide would be unlikely

to gain acceptance because of its high cost and inappropriateness in seme

situations.

       In Texas, dicofol probably would gain increased usage, since its

cost would be only 33% higher than ohlorobenzilate's and it could be used

in IPM programs.  However, recent preliminary NCI data indicate that di-
                                                                 30/
cofol may pose a more sertous risk of cancer than chlorobenzilate.   Fen-

butatin-oxide could substitute for chlorobenzilate  in many  instances and

would be conpatit.e with IPM programs.  Fenbutatin-oxide's per-acre treat-

ment cost  increase over chlorobenzilate is lower  in Texas than  in Florida,

but treatment costs with fenbutatin-oxide would be  twice as high as treat-

ment with chlorobenzilate.  Moreoever, fenbutatin-oxide may pcse a more
                                                          30/
serious risk of reproductive effects than chlorobenzilate.    In Texas,

as in Florida, growers may choose non-selective pesticides  to replace

chlorobenzilate; the IPM impacts of such a choice would be  similar to

those predicted for Florida.  As for Florida, a conclusion  that the bene-

fits of the citrus uses of chlorobenzilate in Texas exceed  the  risks would
                                                                       *
be based upon both economic and health concerns but the health  concerns

would weigh more heavily and the economic concerns  less heavily than for

Florida.

       In Arizona, the economic consequences of cancellation would be

insignificant.
30/  See Appendix C.
                                 (77)

-------
       In California, a conclusion that chlorobenzilate's  citrus  uses  should
be continued would reflect a determination that the risks  are snail when
weighed against the absence of satisfactory alternatives for the  control of
the citrus bud mite and the citrus rust mite.
          2.  Proposed Restrictions
       The data show that the populations at risk with  respect to chloro-
benz ilate use are the U.S. population at large, Florida residents, pesticide
applicators, and citrus pickers.  Under this option, pesticide applicator
exposure would be reduced by changes to the terms and conditions  of regis-
tration.  Directions for use would be modified so that  citrus growers  would
                                                               2i/
be allowed to select one of the following application methods:

          i)    Ground application with applicators using additional
               protective clothing and respirators.
          ii)   Ground application with applicators using suitably-
               equipped enclosed  cabs.
          iii)  Aerial application

       In Addition, the following registration amendment would be required
for the continued use of chlorobenzilate on citrus:

          iv)   Classify as a restricted use pesticide,  for use by
               or under the direct  supervision of certified ap-
               plicators.
       Based on data and experience  frcm other pesticides, the Agency has
concluded that the measures described below would reduce exposure to chloro-
 31/ See Appendix D for specific label amendments.
                                 (78)

-------
benzilate applicators and therefore would reduce a substantial portion of
the risk from chlorobenzilate use.
       Protective clothing and respirators could reduce  the ground appli-
cators' exposure and potential risks.  The exposure estimates for ground
applicators  (Severn, 1978) are based on exposure to arms, hands  and face
(15.8% of the total body surface); covering  the arms and hands would re-
duce the dermal exposure from between 12-40  ing/day to  between 2.4-8.0 rag/
day.  Therefore, to reduce the exposure the  Agency would require applica-
tors to wear heavy-duty work gloves and full-length, long-sleeved, one-
piece jumpsuits made of fine weave fabric  (jersey)  (Griffiths, 1978).  Both
would have to be laundered after each day of use.  Applicators would also
be required  to wear a broad-brimmed hat.  Further, face-piece respirators
would effectively eliminate exposure by  inhalation, estimated at 1 rog/day
without protection.  Therefore,  the Agency would require applicators to
wear suitable respirators which  fit over the mouth and nose  and  have a
filtering cartridge  (NIOSH approved respirators  for pesticide spray ap-
plicators referred to  in Appendix D).
       Protective clothing and respirators would reduce citrus pesticide
ground applicators'  estimated lifetime  cancer risk  (Table  18) by a factor
of five  (Severn 1978).  There would also be  a greater  margin of  safety
 (215 to  845) from testicular effects.   These risk reductions would out-
weigh  the minimal oast  for  the protective  clothing and the respirators.
       Use of suitably-equipped  enclosed cabs by applicators would also
reduce ground applicator  exposure and potential  risks.  The Agency would
require  that these rarn be  completely enclosed and  employ  systems for
                                (79)

-------
delivering filtered air to the operator, as described  in a  recent study
(Taschenberg and Bourke, 1975). In this study, the cab substantially  re-
duced (99%) the amounts of pesticide spray which could cone into contact
with the applicator.  Applying the results of this study to chlorobenzilate
exposure estimates would reduce exposure from between  13 and 41  rug/day to
between 0.13 and 0.41 ing/day when an enclosed cab  is used.   This would re-
duce the lifetime cancer risk to ground applicators from between 400  and
1400 to between 4 and 278 in one million.  The margin  of safety  from  tes-
ticular effects would be 4300 to 16,900.  The costs of this approach  can-
not be fully assessed.  However, the capital cost of air-conditioned  cabs
(which are anticipated to be somewhat  less expensive than filtered  air
cabs) would run between $4.4 and $19.1 million  if all  growers selected
this measure, depending upon the extent to which existing equipment could
be retrofitted (Luttner, 1978).  While growers have  indicated interest in
equipment of this type, it  is probable that requiring  enclosed cabs for
all ground applicators would drive citrus growers  to use other chemicals,
rather than incur the capital cost of  new application  equipment.
       Aerial application of pesticides  is favored by  seme  large growers,
and current information indicates  it would continue  to be an accepted
method of application where suitable.  When chlorobenzilate is applied
aerially, usually by helicopter,  there is a minimum exposure to  the ap-
plicator, but exposure from drift could potentially  increase for people
in the vicinity of  citrus groves.
       Classification to use only by certified  applicators  would also re-
sult  in reductions  in applicator  exposure.   .The key  concept behind classi-
                                 (80)

-------
ficaticn to use by certified applicators  is that certification can gene-
rally upgrade applicator skills and that  with more skill and knowledge
applicators are more likely to use pesticides carefully and efficiently.
It would be reasonable to conclude that a general upgrading of the skills
of chlorcbenzilate applicators would, for these reasons, result  in re-
duced exposures.
       Cancelling the non-citrus uses of  chlorobenzilate would have  the
impacts discussed in Option B.
       Choice of this option would indicate acceptance of  the level  of
risk to consumers frctn chlorobenzilate citrus use*; and would reduce  the
level of ris/c  (lowered by amended terns and conditions of  registration)
posed to applicators.  However, this risk level is based on somewhat un-
certain assumptions due  to the lack of definitive data on  exposure  to
chlorobenzilate.  While  the Agency would  impose additional safeguards with
regard to  the  level of risk posed to applicators,  it has not recommended
protective measures for  citrus pickers because  it has no data base  upon
which to evaluate risk to them.
     D.  Cancel Chlorobenziiate Use on Citrus to Take Effect After
         Five Years  and In the Interim Amend the Terms and Conditions
         of Registration;  Cancel All Other Uses.
        Option D would indicate the Agency's unwillingness to accept the
 risk levels  of Option C indefinitely,  but would demonstrate acceptance
 of  the Option C level cf risk for a five-year period in order to reduce
 the initial  economic irapact and encourage technological innovation.
        This  option would reduce by approximately 14-fold the potential
                                 (81)

-------
lifetime cancer risk and risk from tes'cicular effects, since the period of



exposure would be reduced fron a lifetime to five years  (Table 18).  The



combined effect of the amended use directions ard the limited tine span



would reduce the risk of cancer for citrus pesticide applicators from 400



to 1400 per one million to between 0.3 and 20 per one million  (Table 18).



The testicular effects margin of safety for citrus pesticide applicators



would be increased from between 43 and 169 to between 3,010 and 236,000.



       This option would allow time to develop  and register an alternative



to chlorobenzilate (e.g. Zaroex and Hirsutella)  for use  in citrus IPM pro-



grams.  The five-year phase out would lessen the otherwise substantial  im-



pacts fron the loss of chlorobenzilate.



       Choice of this option would reflect a conclusion  that die risks  as-



sociated with the citrus uses are acceptable for the period of time  neces-



sary to develop and register an alternative compatible with IPM programs,  in



order to avoid the econonic  irapacts of  immediate cancellation  (See Option B).



However, choice of this option also would reflect a conclusion that  indefi-



nite, future use of chlorobenzilate  involves risks which are unacceptable  in



view of the benefits, and that  indefinite continued registration of  chloro-



benzilate creates unacceptable barriers  to the  development of  alternatives.



       This option would require the Agency to  deal with two  important  areas



of uncertainty:  the period of time  that would  be necessary to develop  and



regist3r an alternative(s), and the  fact that the economic and environmental



 impacts of the alternatives which may be developed necessarily cannot  be



assessed at this time.  Both areas of uncertainty would  be addressed by tte



selection of a five-year phase out  for  chlorobenzilate.   Generally  speaking,
                                   (82)

-------
3 years sre necessary to carry a pesticide all the way through the develop-
ment and registration process, using the date that a compound is selected
as a candidate for development in early screening tests.  Since Zardex and
Hirsutella are already well beyond this stage, a five year-period would seem
adequate to accomplish development and registration of alternatives.  More-
over, seler.ion of a five-year phase out period would also  reflect an Agency
decision that sufficient time would be available after the  registration  (or
failure to register) of an altemative(s) to permit applications for renewed
registration of chlorobenzilate to be considered and acted  upon.
       Finally, it should be noted that selection of this option would re-
flect an Agency decision that  incentives are necessary to stimulate de-
velopment and registration of alternatives,  and that the creation of  in-
centives justifies the uncertainties and burdens associated with the op-
tions.  Selection of this option would further reflect a decision that
the probability that an alternative would be developed and  registered  is
unacceptably low  without the creation of  incentives by the  Agency.

     E.  Continue Registration of Chlorobenzilate Use on Citrus, Amend
         the Terms and Conditions of Registration, Require  That Identified
         Exposure Data Be Submitted  to EPA  in  18 jtonths; Reevaluate  the
         Use on Citrus After Additional Exposure Data Becomes Available;
         Cancel All Other Uses.

       Option E would reflect  an Agency conclusion  that  the risks of the
non-citrus uses of chlorobenzilate outweigh the benefits and that  the
benefits of  the citrus  uses of chlorobenzilate outweigh  vie risks,  as  re-
duced  by modifications  to  the  terms  and  conditions of  registration (see
Option C).  Option E would further  reflect  the Agency's  conclusion that
                                 (63)

-------
additional exposure data is necessary, and a determination to reevaluate
the citrus uses of chlorobenzilate after these data are available.
       •RiLs option would eliminate all risks associated with chlorobenzi-
late's non-citrus uses.  It would require the same amended use directions
as those explained in Option C for the continued use of chlorobenzilate pro-
ducts on citrus.  It would also preserve the usefulness of chlorobenzilate
in integrated pest management, at least until the additional data has been
submitted to and evaluated by the Agency.
       Option E would allow risks to continue at levels comparable to that
of Option C but would require new data to substitute or refute current risk/
benefit estimates or to indicate the need for a revised evaluation of risks
and benefits.  Should new data indicate higher risks or lower benefits than
have been found in the present analysis, the Agency would reassess its regu-
latory decision.  If the data confirms the present assessment, this option
would be equivalent to Option C.  The isroediate economic inpact of this op-
tion would be ccoparable to that of Option C.
       If this oprion  is adopted, registrants would be required to submit
data derived from the studies described belcw; specific protocols have to
be submitted by the registrants for approval within six months.
          1.  Citrus Fractionation Studies
     Very little information is available on chlorobenzilate residues  in
citrus by-products.  Limited data indicates that residues are present  in
citrus pulp and suggests the potential for residues  in other by-products
such as citrus oil.  A fracticoation study  is necessary to measure chloro-
benzilate residues  in  the by-products of citrus processing and, using  these
                                   (84)

-------
measurements, to estimate exposure fron the products which have commercial
utility.  Since each residue analysis step  (1 per by-product)  is estimated
to cost $100 to $200, and fractionation of citrus fruit requires approximately
ten steps, the study would cost an estimated $1000 to $2000 for each citrus
fruit type.
          2.  Feeding Citrus By-products to Cattle Study
       Existing data indicates that chlorobenzilate residues are present
in the citrus pulp used to feed cattle in Florida, and that milk and beef
from cattle that were fed chlorobenzilate-contaminated pulp may contain
chlorobenzilate residues.  The proposed feeding study is needed to measure
chlorobenzilate residues  in milk and meat from cows that were  fed  pulp frcra
citrus fruit treated with chlorobenzilate.  Data from this study would be
used to estimate exposure to Florida consumers from these dietary  sources.
The study would cost approximately $100,000.
          3.  Citrus Pickers Exposure and Re-entry Studies
       The Agency has no data to determine  whether chlorobenzilate residues
on fruit surfaces and tree foliage create an exposure source  to citrus
pickers.  Procedures which determine whether dislcdgable  chlorobenzilate
residues adhere to pickers and the degradation rate of chlorobenzilate in
field conditions will permit estimates of the occupational exposure and
risk levels  for pickers.  These studies would cost between $100 and $200
per sample.  Because sixty samples are estimated to be required,  the total
study would  cost between  $6,000 and $12,000.
          4.  Aerial Application Exposure Study
       The Agency lacks data on exposure  levels  which  result  from the aerial
                                 (85]

-------
application of chlorobenzilate.  This stud' wwould provide relevant data
concerning exposure to the drift which results from aerial application of
pesticides in order to estimate ohlorobenziiate exposure  levels  and eval-
uate the risk to nearby inhabitants.
       The Agency estimates that the number of samples  required  rnay range
from 500-1000 and at $100-5200 pur sample, the study would cos.  between
$50,000-$200,000.
          5.  Ground Applicator Exposure Study
       The current exposure estimates for  citrus ground applicators are
based on studies with other pesticide sprays.  An exposure study of chloro-
benizilate ground applicators would allow  better analysis and  consequently
better assessment of their potential risk  from spray applications  of  chloro-
benzilate.  The study should cover mixing/ loading and  application exposure.
The Agency estimates that 50 samples would be required  at a  cost of be-
tween $100 and $200 per sample; the entire study would  cost  $5,000-510,000.
          6.  Residue Monitoring of Milk from Pulp-Fed  Cattle  and
              Residue Monitoring of By-products of Citrus Process'i.ig
       Vhile a preliminary analysis of milk samples  from  Florida does not
indicate detectable residues of chlcrobenzilate  (Kutz,  1978),  an earlier
study  (Formica et al., 1975) demonstrated  chlorobenzilate residues in milk
from cattle that were fed pulp to which chlorobenzilate had  been added.
In addition, there are studies by Mattson  and Insler (cited  in Severn,
1978} which indicate that chlorobenzilate  residues could  be  expected  to
occur  in beef.  Therefore, a need exists  to establish methods  for  detect-
ing chlorobenzilate residues  in milk and beef from cattle that have been
fed pulp from chlcrobenzilate-treated  citrus fruit.

                                 (86)

-------
application of chlorobenzilate.  This study would provide relevant data
concerning exposure to the drift which results from aerial application of
pesticides in order to estimate chlorcbenzilate exposure levels and eval-
uate the risk to nearby inhabitants.
       The Agency estimates that the number of samples required nay range
frcm 500-1000 and at $100-5200 per sample, the study would cost between
$50,000-$200,000.
          5.  Ground Applicator Exposure Study
       The current exposure estimates for citrus ground applicators are
based on studies with other pesticide sprays.  An exposure study of chloro-
benizilate ground applicators would allow better analysis and  consequently
better assessment of their potential risk frcm spray applications of chloro-
benzilate.  The study should cover mixing, loading and application exposure.
The Agency estimates that 50 sanples would be required at a  cost of be-
tween $100 and $200 per sample; the entire study would cost  $5,000-$10,000.
          6.  Residue Monitoring of Milk from Pulp-Fed Cattle  and
              Residue Monitoring of By-products of Citrus Processing
       Wiile a preliminary analysis of milk samples from Florida does not
indicate detectable residues of chlorobenzilate  (Kutz, 1978),  an earlier
study  (Formica et al., 1975) demonstrated chlorobenzilate residues  in milk
from cattle that were fed pulp to which chlorobenzilate had  been added.
In addition, there are studies by Mattson and Insler  (ciced  in Severn,
1978) which indicate that chlorobenzilate residues could be  expected to
occur  in beef.  Therefore, a need exists to establish methods  for detect-
ing chlorobenzilate residues  in milk and beef frcm cattle that have been
fed pulp frcm chlorobenzilate-treated citrus fruit.

                                 (86)

-------
       Refinement of the analytical detection method is estimated to cost
approximately $55,000.  This assumes a 10% overhead cost.  The estimated
cost of monitoring the samples would be $100-200 per sample or an esti-
mated $3,000-$6,000 for the entire sample monitoring study.

     F.  Continue Registration of Chlorobenzilate Use on Citrus in
         Florida, Texas, and California, Amend the Terms a->d Con-
         ditions of Registration, Require that Identified Exposure
         Data be Submitted to EPA in 18 Months; Reevaluate the Use
         on Citrus After Additional Exposure Data Becomes Available;
         Cancel Use on Citrus in Arizona and All Other Uses!
       Option F is the same is Option E, except that Option F would -.ndi-
cate that the Agency recognizes a different risk/benefit argument for
citrus uses between regions.  In Arizona, the loss of Chlorobenzilate and
adoption of alternative miticides is projected to have no economic impact.
Cancellation of citrus use in Arizona would yield a marginal reduction
in risk to consumers; however, it would eliminate the risk to Arizona
applicators.

     G.  Continue Registration of Chlorobenzilate Use on Citrus, Amend
         the Terms and Conditions of Registration; Prohibit the Use of
         Pulp  from Chlorobenzilate-Treated Citrus  as Cattle  Feed;  Estab-
         lish Complementary Tolerances; Cancel All Other Uses.
       Option G  is  the same  as Option C,  except that Option G would  in-
dicate that the Agency is  unwilling to accept the potential risk posed to
the Florida population through consumption of milk and  beef from cattle
fed pulp fron chlorcbenzilate-treatcd citrus.  Accordingly, this option
would prohibit the  feeding of citrus pulp to cattle and would propose the
establishment of  complimentary tolerances under the Federal Food,  Drug,
and Cosmetic Act.
                                 (87)

-------
V.  Recommended Regulatory Action
         OPTION F:
         Continue Registration of Chlorobenzilate Use on Citrus  in Florida,
         Texas, and California, Amend the Terms and Conditions of Registra-
         tion. Require that Identified Exposure Data be Submitted to EPA
         in 18 Months; Reevaluate the Use on Citrus After Additional Expo-
         Data Becones Available; Cancel Use on Citrus in Arizona and All
         Other Uses.
     A.  Introduction


       The foregoing review sumnarizes  and analyzes  information on the

risks and benefits of  the uses of Chlorobenzilate  and evaluates a series

of regulatory options.  Several particularly significant  factors stand

out  in the analysis.


      With Respect to  Risks
          Several  studies  provide  a reliable basis for concluding
          that Chlorobenzilate induces oncogenic effects in mice.
          Under the Agency's Interim Cancer Assessment Guidelines,
          these laboratory studies provide substantial evidence that
          Chlorobenzilate  poses a  cancer risk to man.   In view of
          the human exposure which may result from its uses, Chloro-
          benzilate poses  a cancer risk to man of sufficient magni-
          tude to  require  the Agency to determine whether these
          use; offer offsetting social, economic, or environmental
          berefits.  The key populations at risk with respect to
          di'.orobenzilate  are the  U.S. population at large, Florida
          residents, pesticide applicators, and citrus pickers.
          Chlorobenzilate causes adverse effects to the testes of
          male rats,  and may cause adverse effects to the testes of
          applicators.   Exposure levels of male pesticide applicators
          are high enough, in comparison to the "no observable effect"
          levels for  adverse testicular effects in rats/ to warrant
          a conclusion that Chlorobenzilate may pose a risk of ad-
          verse effects to humans of sufficient magnitude to require
          the Agency  to determine whether offsetting social, envi-
          ronmental or economic benefits result from the uses of the
          pesticide.
                                 (88)

-------
With Respect to Benefits
  Chlorobenzilate  is used on citrus crops in Florida, Texas,
  Arizona, and California to control mites.   Most use occurs
  in Florida  (72.1%).  Significant adverse economic effects,
  including production losses, would occur if mite pests are
  not controlled.  Chlorobenzilate  is  utilized  in citrus in-
  tegrated pest management because  it  is  selective to mites
  and does not kill natural predators  and parasites used to
  control citrus scale pests.  Such  integrated  pest .oanage-
  ment approaches  are used extensively in Florida, and  to a
  lesser extent  in the other citrus growing  regions. There
  are several other selective miticides registered for  use
  on citrus crops, and a number  of non-selectiv> miticides
  are also registered for use on these crops.
  Cancellation of  chlorobenzilate would significantly in-
  crease pest control  costs in Florida.  Non-selective mi-
  ticides would  be the predcr.-L-.ant replacements for chloro-
  benzilate,  for economic and other reasons developed in
  detail in Section III.   This would result in abandoning
  BW control of scale, because populations of beneficial
  insects would  b> reduced, and large volumes of chemical
  pesticides  would have to te used to control scale insects.
  Relatively small amounts of chlorobenzilate are used in
  California and only on a few citrus crops in one area.
  However,  cancellation of chlorobenzilate would have sig-
  nificant  impacts, because there are no registered alter-
  natives that are regarded as suitable chlorobanzilate
  replacements.
  The loss of chlorobenzilate and the adoption of alterna-
  tive miticides is projected to have no net cost to Arizona
  citrus growers.   Using alternatives may disrupt IPM strate-
  gies in Arizona, but the extent of any such disruption has
  not been identified, nor the resulting cost quantified.
   Registered efficacious alternatives are available for each
   of the other uses of chlorobenzilate; in some cases these
   alternatives are less expensive and achieve comparable
   levels of control.
                          (89)

-------
     B.  Comparison of Options
       In selecting a regulatory option, the Agency must decide which of
the proposed options achieves the most appropriate balance between risks
and benefits.  This decision turns in part on the key factual elements
suranarized above, and in part on the relative merits of each proposed
option.
       Option A, which would continue registration of all uses, and Option
B, which would cancel all uses, represent an all-or-nothing approach to
regulation.  Under Option A, the Agency would do nothing whatsoever to re-
duce the known risks of chlorobenzilate nor would it otherwise recognize
that the RPAR review confirmed the presunption of oncogenicity.  By con-
trast. Option B would succeed in eliminating risk, but only with substan-
tial increased costs for substitute pesticides and possibly serious en-
vironmental and agricultural consequences as the result of disrupting
established IPM programs or using unsatisfactory subsitutes.  The  latter
result would be particularly unfortunate because of the Agency's avowed
interest in promoting the use of IPM as an alternative to the  indis-
criminate use of pesticides.
       Further, Options A and B are even less  tenable  in view of the range
of available moderately restrictive measures described  in Options  C through
G, which would reduce risk to sane extent without significant benefit  im-
pacts  and would avoid the harsh consequences of across-the-board cancella-
tion.  These considerations  indicate that Option A would be  reasonable only
if the benefits clearly outweighed the  risks,  and  if risk reductions could
not be achieved without unacceptable benefit consequences.   Such ccnsidera-

                                 (90)

-------
tions further indicate that Option B would be reasonaole only  if the
risks clearly outweigh the benefits, and significant reductions in risks
cannot be achieved by measures short of cancellation, without  unaccep-
table benefit impacts.  The facts indicate that neither situation per-
tains, and that litese options are not reasonable regulatory measures  in
this case.
       The analysis of risks and benefits of the uses of chlorobenzilate
indicates that risks and benefits from citrus uses  in Florida, Texas,
and California nay be. close to equilibrium; however, in each situation
significant risk reductions can be achieved, without significant  inpacts
on the benefits of the use.  With respect to the citrus use  in Arizona
and the non-citrus uses on the other hand, the analysis suggests  that risks
exceed benefits, primarily because the projected impacts of cancellation
are insignificant while the risks of the uses  (particularly  to applicators)
are not insignificant.  Option C and the options which  follow  it  all  rep-
resent possible regulatory responses to this general assessment of  the
risks and benefits of chlorobenzilate uses and the  balance that should
properly be struck between them.
       Option C has  three distinct  components, each of  which is designed
to reduce the risks  of cancer and adverse testicular effects associated
with  the uses of chlorobenzilate without simultaneously creating  the ad-
verse economic, social, or environmental  impacts associated  with  cancel-
lation.  In proposing measures  to reduce risks by cancelling some uses
and restricting the  conditions  of use  for the  registrations  which re-
main  in effect, this option  is  distinguishable from Option A which would
                                 (91)

-------
allow the currently projected risks to continue indefinitely.  And by
preserving the uses of chlorobenzilate in the major IPM programs, this
option is distinguishable fron Option B which Mould eliminate IPM use
and produce substantial adverse economic and agricultural inpacts.
       Options D and G are based on Option C and, like Option C, each
would preserve sane chlorobenzilate uses under restricted use conditions,
thereby reducing riskr and also avoiding substantial economic impacts.
Each also has a special feature designed to reduce risks which would not
be affected by the terms of Option C.  The special feature of Option D,
cancellation in five years, would reflect a judgment that lifetime ex-
posure to the risks of chlorobenzilate is unacceptable.  To ameliorate
the adverse economic impact of iranediate cancellation as described in
Option B, Option D would propose that chlorobenzilate be phased out over
a period of five years, durirg which time satisfactory alternatives may
be developed.  There is, of course, no certainty that these alternatives
would be developed in this time.  However, since alternatives are cur-
rently under development, cancellation would not be necessary to bring
this about.
       The special features of Option G, measures to reduce exposure to
Florida residents frcm milk and beef, address the risk which the popu-
lation in this State may experience.  Since these measures would elimi-
nate the market for citrus pulp, this option would reduce growers net
income and create a costly disposal problem for processors.  Tc avoid
these consequences processors may not purchase chlorabenzilate-treated
fruit, and growers may not use chlorobenzilate.  In effect, then, elimi-
                                 (92)

-------
nation of this market oxxd constitute de facto cancellation.
       Although Option C and, derivatively, Options D and G would achieve
risk reduction without serious adverse economic impacts, these options
would not provide measures to validate or expand the information base and
thereby make the bases for the proposed regulatory action more certain.
       Although the Agency believes that it has made a sound and prudent
assessment of the available data, any analysis of this sort requires that
a number of uncertainties be dealt with by assumptions drawn from the
Agency's expedience and expertise.  The Agency has strived to use conser-
vative assunptions, consistent with its responsibilities to protect public
health.  However, it is possible that the Agency has underestimated po-
tential human exposure, and therefore, underestimated the risks of chloro-
benzilate use.  if so, Option C (which would permit continued use of
chlorobenzilate based upon these estimates) would allow a potentially un-
acceptable risk to continue  indefinitely.
       Option D presents the reverse problem.  If the uncertain data base
has resulted in over-estimates of probable human exposure, Option D would
propose more stringent regulatory action than would be required to reduce
actual risk, and would do so on the speculative assunption that more satis-
factory alternatives will be developed.
       Options C, D, and E would allow continued use of chlorobenzilate on
all citrus with the provisions discussed above.  However, there is no  in-
formation in the record to indicate that substantial benefits derive from
the use of chlorobenzilate in Arizona.  Acceptance of this option by the
Agency would ignore the apparent imbalance of risks and benefits in Arizona,
particularly with respect to Arizona applicators.

                                 (93)

-------
     C.  Recommendaticn of Option F
       Option F eliminates the risk from the citrus use  in Arizona as
well as the risk from all non-citrus uses by cancellation of  those uses.
This option also reduces the risk to spray applicators,  the population
at greatest risk from citrus uses of chlorobenzilate, by amending the
terns and conditions of registration so that chlorobenzilate  may be ap-
plied only by certified applicators and only in accord with label direc-
tions which reduce exposure.
       Specifically, this option requires two amendments to the terms
and conditions of chlorobenzilate registrations for citrus use. The
first  is a requirement that this pesticide be applied only by certified
pesticide applicators to ensure, to the extent possible, that pesticides
will be applied only ty persons skilled and knowledgeable  in  handling
pesticides.  The second requirement  is that chlorobenzilate be applied
only if applicators are protected by protective clothing (hat, gloves,
coverall-type suit), respirators, or by working in an  enclosed air-
filtered cab.  The clothing cost would be minimal. The cabs  would add
a substantial cost to the application process, but if  the Agency requires
cabs for other pesticides that may be harmful to  applicators, this cost
would  not be fully attributable to the chlorobenzilate regulatory program.
       The cancellation and use restriction elements  in this  option  are
basad  on the conclusion that when used  in accordance with the modified
terms  and conditions of registration, the risks associated with the major
citrus ores are not unreasonable,  in view of  the  benefits of  those uses
                                 (94)

-------
and the related probable adverse economic impacts of cancellation.  In
addition, this option also reduces risks appreciably without unacceptable
benefit consequences.
       Since the continuation of citrus registrations without any other
restrictions on use does not reduce the risks to consumers, a decision
to continue these uses reflects a determination that these uses do not
cause unreasonable adverse effects with respect to this group.
       Option F imposes new costs to replace chlorobenzilate with al-
ternatives for non-citrus uses, and other costs in connection with new
protection for spray applicators and testing to develop additional data
for exposure estimates, but avoids substantial adverse economic effects.
Specifically, the cost of cancelling non-citrus uses is minimal ccnpared
to the cost of cancelling citrus uses.  Further, since cancelling non-
citrus uses does not affect IPM programs as would cancelling citrus uses,
this option avoids disruption of these programs.
       Finally, Option F is preferable to the other proposed options be-
cause it provides a regulatory mechanism to clarify and enlarge the data
base for the exposure estimates which underlie the risk assessments and
risk/benefit analyses.  1t> accomplish this objective, Option F requires
chlorobenzilate registrants to develop additional data to confirm or
evaluate the Agency's chlorobenzilate risk assessments and the related
regulatory decisions.
                                 (95)

-------
1.  Albert, Roy E. (CM), 1978.   EPA memo of June 8, 1978 to J.B.  Boyd
    (SPRD).  Comments on the chlorobenzilate risk assessment.

2.  Albert, Roy E. (GAG), 1978a.  EPA memo of June 23, 1978 to J.B. Boyd
    (SPRD).  Chlorobenzilate risk assessment issues.

3.  Albert, Hoy E. (CAS), 1978b.  Carcinogen Assessment Group's Surmary
    and Conclusions for Assessment of Carcinogenic risk of Chlorobenzi-
    late.  Unpublished EPA report dated June 16, 1978.

4.  Allen, Jon C., 1978 "The Effect of Citrus Rust Mite Damage on Citrus
    Fruit Drop", Unpublished report of the University of Florida,  IPAS,
    Agricultural Research and Education Center.

5.  Anderson, Elizabeth (CAG), EPA memo of May 29, 1977, to J.B. Boyd
    (SPRD).  CAG assessment of rebuttal points on chlorcbenzilate.

6.  Barton, Anne (OTS), EPA memo of December 7, 1977, to J.B. Boyd (SPRD).
    Chlorobenzilate:  overall false positive rate.

7.  Barton, Anne (OTS), 1977a.  EPA memo of February 15, 1977 to J.B. Boyd
    (SPRD).  Statistical analysis of td chlorobenzilate study.

8.  Boyd, Hudson (CED), 1978.  Environmental Fate Prof ile:  Chlorobenzilate.
    Unpublished report.

9.  Boyd, J.B.  (SPRD).  EPA memo of December 6, 1977 to Edwin L. Johnson
    (DAA for Pesticide Programs).  Chlorobenzilate RPAR Rebuttal Decision.

10. Boyd, J.B.  (SPRD).  EPA roeno of April 1, 1977 to Clayton Bushong (CED).
    Determination of ecological risks associated with the use of chloroben-
    zilate alternatives.

11. Boyd, J.B.  (SPRD).  EPA memo of April 1, 1977 to Lamar Dale (CED).
    Determination of health risks associated with the use of chloroben-
    zilate alternatives.

12. Boyd, J.B.  (SPRD).  EPA memo of March 6, 1978 to SPRD Record.  Rebut-
    tals and contents on the chlorobenzilate RPAR.

13. Brogdan, James E., 1976. Extension entomologist. University of Florida,
    Gainesville, Florida, personal communication of August 6, 1976 to
    Mark Luttner  (cited  in Preliminary Benefit Analysis of Chlorobenzilate).

14. Brogdan, Janes E. and Fred P. Lawrence, 1974. Insects and Mites of
    Florida Citrus,  University of Florida, Institute of Food and Agricul-
    tural Sciences, Gainesville, Florida, (cited  in Preliminary Benefit
    Analysis of Chlorobenzilate).

-------
15, Brooke, Oonale L., 1573.  Citrus production costs and returns in
    Florida season 1971-1972 with comparisons.   Cooperation Extension
    Service, Institute of Food and Agricultural Sciences.  University
    of Florida.

16. Brooks, R.F., 1977.  Integrated Control of Florida Citrus Pests,
    the Citrus Industry.

17. Brooks, R.F. and J.D. Whitney, 19  .  Citrus snow scale control in
    Florida.  Florida Agricultural Experiment Stations Journal,  Series
    Mo. 4890.

18. Bullock, R.C., 1977.  Aerial Application to Florida Citrus (Unpolished
    Report)*

19. Burnam, W.L..  EPA memo of May 4, 1978 to J.B. Boyd (SPED).   Chloro-
    benzilate Alternatives Review.

20. Burnam, W.L.  EPA nemo of June 20, 1978 to J.B. Bcyd (SPED).  Vendex
    Review.

21. Bushong, Clayton (CED).  EPA memo of August 8, 1977 to J.B.  Boyd (SPRD).
    Chlorobenzilate risk anlaysis, Fish and Wildlife.

22. Codce, J., G.C. Lyons, and A. Amader, 1978.  Texas Citrus Pest
    Management Proposal.  Texas Agricultural Extension Service.   The
    Texas A & M University System.

23. California Dept. of Agriculture, 1972.  Pesticide Use Report 1971,
    Sacramento, California.

24. California Department of Food and Agriculture, 1974 Pesticide Use
    Report 1973, Sacramento, California, (cited in Preliminary Benefit
    Analysis.

25. Dale, Lamar  (CED).  EPA mano of May 4, 1978 to J. B. Boyd (SPRD).
    Limitation of the Chlorobenzilate alternatives review.

26. Dean, H.A.,  1975.  "Complete biological control of Lepidosaphes beckii
    on Texas citrus with Aphytis Lepidosaphes."  Environ. Entrool. 4:110-114.

27. Dean, H.A.,  1955.  "Factors affecting biological control of scale in-
    sects on Texas citrus."  J. Econ. Entnol. 48:444-447.

28. Dean, H.A. and J.C. Bailey, 1964.  "Certain post blccm treatments
    for control  of Texas citrus mites and theic effect on chaff scale
    parasites."  J. Rio Grande Valley Hort.  Soc. 18:29-32.

29. Dean, H.A.,  W.G. Hart, and S.J. Ingle, 1977.  "Pest management
    considerations of  the effects of pesticides on Texas citrus pests
    and certain  parasites."  J. Rio Grande Valley Hort. Soc. 31:37-44.

-------
30. Doane Agricultural Service Inc., 1976.  "Pesticide Use Data on Se-
    lected Specialty Crops," Current Pesticide Use and User Profiles for
    Selected Pesticide Intensive Crops; Report No. 4, EPA Contract 68-
    01-1928, St. Louis, Missouri.

31. Edwards, Gordon S.  More of August 2, 1977 to Elizabeth Anderson (CftG).
    Comments on criticisms of the Innes study raised by pesticide regis-
    trants and within EPA.

32. Federal Register, May 26, 1976.  Notice of presumption against regis-
    tration and continued registraion of pesticide products containing
    chlorobenzilate. 41(103):21517-21519.

33. Federal Register, June 11, 1976.  Notice of presunption against regis-
    tration and continued registraion of pesticide products containing
    chlorobenzilate. 41(114):23753.

34. Federal Register, July 14, 1976.  Hot ice of presumption against regis-
    tration and continued registraion of pesticide products containing
    chlorobenzilate. 41(136):29015.

35. Federal Register, July 14, 1976.  Notice of extension of period for
    submission of rebuttal evidence and comnents with regard to presunp-
    tion against registration and continued registration of pesticide
    products containing chlorobenzilate.  41(136):29015.

36. Federal Register, May 25, 1976.  Notice on "Interim Procedures and
    Guidelines for Health Risk and Economic Inpact Assessments of Sus-
    pected Carcinogens." 41(102):21402-21405.

37. Feldman, R. J. and H. I. Maibach, 1974. "Percutaneous Penetration
    of Some Pesticides and Herbicides  in Man*.  Tax. Appl. Pharm.
    28:126-132.

38. FeUcner, I.C., 1978.  Commentary on cancer risk  from exposure to
    chlorobenzilate.  Unpublished EPA report, dated  April 23, 1978.

39. Florida Cooperative Extension Service, 1977.  Florida Citrus Spray
    and Oust Schedule, 1977; Circular  393-C.  University of Florida,
    Gainesville.

40. Food and Drug Administration (FDA), 1971.  Food  and Drug Adminis-
    tration Advisory Ccnrcittee on protocols for safety evaluation;
    Panel on carcinogenesis  report on  cancer testing in the safety
    evaluation of food additives and pesticides.  Toxicol. Appl.
    Pharmacol. 20:419-438.

41. Forjnica, G., W. Blass, and D. Eberle, 1975.   "Gas chronatographic
    residue determination and residues of chlorobenzilate and two major
    metabolites  in plant materials and milk."  Meded. Fac. Landbouwwet.,
    Rijks Univ. Gent.  40(2, Pt.2): 1135-1148.

-------
42  French, V., P.J. Rathwell and Reed J. Reeve, 1978.  Management of
    Texas citrus.  Citrograph. 63:77-79.

43. Freudenthal, R.I. and P. Leber, 1977.  Evaluation of chlorobenzi-
    late toxicology data to determine its potential health hazards to
    nan or domestic animals.  Battelle Columbus Laboratories, Columbus,
    Ohio.  (Unpublished).

44. Frith, Charles, 1976.  National Center for Toxicological Research.
    Analysis of Innes Slides and Animal Diagnosis. (Unpublished).

45. Golberg, L., ed., 1974. Carcinogenesis testing of chemicals.  CRC
    Press, Cleveland, Ohio.

46. Griffith, Jade  (HEMB).  EPA memo of June 20, 1978 to J.B. Boyd (SPRD).
    Permeability of protective clothing.

47. Gustafson, C.D. and Robert C. Rock, 1977.  Lemon Production Costs.
    Cooperative Extension Service, University of California, San Diego.

48. Hufiaker, G., 1978.  (Editor). Accomplishments Torward Improving
    Integrated Pest Management for Citrus.  Principles, Strategies,
    and Tactics of Pest Population Regulation and Control in Major
    Crop Ecosystems.  In press, Wiley and Sons.

49. Hassan, T.K. and C.O. Knowles, 1969.  "Behavior of Three (T -labeled
    Benzilate Acaricides When Applied Topically to Soybean Leaves."
    J. Boon. Entonal. 62(3):618-619.

50. Horn, H.J., R.B. Bruce, and O.E. Paynter, 1955.  Toxicology of chloro-
    benzilate.  J. Agric. Food Chem. 3(9):752-756.

51. Innes, J.R.M., B.M. Ulland, M.G. Valeric, L. Petrucelli, L. Fishbein,
    E.R. Hart, &.J. Pallotta, R.R. Bates, H.L. Falk, J.J. Gart, M. Klein,
    I. Mitchell  and J. Peters, 1969.  Bioassay of pesticides and  indus-
    trial chemicals  for  tumorigenicity  in mice:  a preliminary note.  J.
    Nat'l. Cancer Inst.  42:1101-1114.

52. Jeppson, Lee R.  (Entomologist), 1976.  University of California,
    Riverside, California, personal ocmrounication of September 12, 1976
    to Mark Luttner.   (cited  in Preliminary Benefit Analysis).

53. Knowles, C.O. and S. Ahmad, 1971. "Comparative metabolism of  chloro-
    benzilate,  chloropropylate and brcmo-propylate.  Can. J. Physiol.
    Phannaool.  49(6):590-597.

54. Luttner, Mark A., 1978a.  EPA memo  of April 27, 1978 to J. B. Boyd (SPRD).
    Estimate of the maxim an economic value lost rs a result of uncon-
    trolled mite infestations in  citrus.

-------
55. Luttner,  Mark A.,  1978b.  EPA memo of June 13,  1978  to J.B. Boyd  (SPRD).
    Vendex as a replacement for chlorobenzilate.

56. Luttner,  Hark A.,  1978c.  EPA msno of June 13,  1978  to J.B.Boyd  (SPRD).
    Vendex.

57. Luttner,  Mark A.,  1978d.  EPA memo of March 21, 1978 to J.B. Boyd (SPRD).
    Chlorobenzilate exposure analysis.

58. Luttnar,  Mark A.,  1978e.  EPA memo of April 14, 1978 to J.B. Boyd (SPED).
    Chlorobenzilate: estimated cost of equipment modification requirement.

59. Luttner,  Mark A.,  1978f.  EPA memo of March 28, 1978 to J.B. Boyd (SPRD).
    Considerations re:  restriction  of chlorobenzilate  use to aerial
    applications only.

60.  Luttner, Mark A., 1978g.  SPA mono of  May 2, 1978 to J.B. Boyd  (SPRD).
     Chlorobenzilate (Impact of prohibiting use of citrus pulp as  cattle
     £e«d).

61.  Luttner, Mark A., 1977a.  Preliminary  Benefit Analysis of Chloroben-
     zilate.   Unpublished EPA report.

62.  Luttner, Mark A., 1977b. Supplement to the Preliminary Benefit  Analysis
     of Chlorobenzilate.  Unpublished EPA report.

63.  Maxey, Robert.  Pesticides Monitoring  Laboratory Memo of June 6, 1978
     to Han Tai.  Chlorobenzilate.

64.  McWhorter, Mike (CED), 1978a. EPA mare of April 5,  1978  to J. B. Boyd
     (SPRD).   Utilization of Dimilin and Hirstuella fungus for control of the
     citrus mite complex.

65.  McWhorter, Mike (CED), 1978b.  EPA mono of April 14, 1978, to J. B. Boyd
     (SPUD).  Efficacy of Zardez.

66.  Mishra, Lakshni C.  (CED).  EPA memo of February 3,  1977  to J.B. Boyd
     (SPBD).  Evaluation of the documents submitted by registrants in re-
     buttal of chlorobenzilate FPAR.

67.  Mishra, L.C.,  and P. Hebborn, 1971.  Ultraoentrifugation of mouse
     plasma containing Rausc^r virus.  Life Sci. 10 (Part II): 1375-1380.

68.  Miyazaki, S.,  G.M. Boush, and P. Matsumura, 1970.  "Microbial Degra-
     dation of Chlorobenzilate  (Ethyl 4,4'-dichlorobenzylate) and  chloro-
     propylate  (Isopropyl 4,4'-dicnlorobenzylate).  J. Agric. Food Chen.
 69.  Murphy, R., R. Kahrs, and A. M. Mattson, 1966.  Dnpublished Report,
     Geigy Cwnical Corporation.
                                                                             100-

-------
70.  National Cancer Institute (NCI), 1977.  Final individual an:onal pa-
     thology table for chlorobenzilate.  U.S. Oept. of Health, Elation,
     and Welfare, Bethesda, Maryland.  (Unpublished).

71.  Page, N.P., 1977.  Chronic Toxieity .^nd Carcinogenicity Guidelines.
     J. Envtl. ?dth. and Tex.. 1:161-182'.

72.  Parrish, Dale W. (HEMB-TSD).  Meno of April 6, j.978 to Director, Texas
     Epidemiologic Studies Program.  Assessment of carcinogenesis data
     (chlorobenzilate).

73.  Pertel, Ruth (SPRD).  Memo dated January 27, 1978 to J.B. Boyd (SPRD)
     and Pat Miller (SPRD).  Mut^enicity rebuttal on chlorobenzilate:
     oonments.

74.  Pertel, Ruth (SPRD).  EPA memo of January 31, 1977 to J.B. Boyd (SPRD).
     Mutagenicity rebuttal on chlorobenzilate: carments.

75.  Potrepka, Robert F.  (CED), 1978a.  EPA meno of June 5, 1978 to J.B. Boyd
     (SPRD).  Testicular atrophy in rats from chlorobenzilate.

76.  Potrepka, Rocert F.  (CED), 1978b.  EPA memo of June 9, 1978 to J.B. Boyd
     (SPRD).  Testicular atrophy in rats from chlorobenzilate, addendum to
     June 5, 1978 memo.

77.  Reed, Donald (KD), 1978a.  EPA memo of March 17, 1978 to J.B. Boyd (SPRD).
     Chlorobenzilate residues in milk and meat.

78.  Reed, Donald (RD), 1978b.  EPA memo of May 3, 19"8 to J.B. Boyd (SPRD).
     Chlorobenzilate residues in milk.

79.  Reed, Donald (RD), 1978c.  EPA memo of June 2, 1978 to J. B. Boyd  (SPRD).
     EPA Chlorobenzilate residue analysis of citrus pulp.

80   Reed, Donald (RD), 1978d.  EPA memo of May 16, 1978 to J.B. Boyd (SPRD).
     Chlorobenzilate residues in citrus pulp; re: discussion of 5/19/78.

81.  Rossi, Lois  (SPRD).  EPA memo c: April 4, 1978 to J. B. Boyd (SPRD).
     Observed deaths among workers in Chlorobenzilate area of Ciba-Geigy
     Mclntosh plant.

82.  Savage, E.P., and R.O. Hayes, 1977. Chlorobenzilate.  Colorado State
     University, Fort Collins, Colorado.   (Unpublised).

83.  Severn, David J., 1978.  Exposure Analysis for Chlorobenzilate.  Un-
     published EPA report dated March 14, 1978.  CONFIDENTIAL.

84.  Sontag, J.M., N.P. Page, and U. Saffiotti, 1976.  Guidelines for
     carcinogen bioassay  in small rodents.  National Cancer Institute,
     Bethesda, Maryland.

-------
85.  Sutton, Harvey J. 1977.  Economics of Helicopter Application for Pest
     Control at Waverley Growers Cooperative (Unpublished Report).

86   Taschenberg, E.F. and John B. Bourke, 1975.  "An Air Filter-Pressuriza-
     tion Unit to Protect the Tractor Operator Applying Pesticides,*  BuU.
     of Environmental Contamination and Toxicology 13(3):263-268.

87.  Technical Services Division (TSD). "Chlorcbenzilate Analysis - citrus
     pulp and feed." Unpublished EPA paper, dated May 22, 1978.

88.  Three Month Feeding Study of Chlorobenzilate, 1965.  Bonn, Germany.
     Submitted to EPA on February 1, 1965 in support of Food Additive
     Petition No.6Hl980.

89.  U.S. Dept. of Agriculture (USDA), 1970.  PR Notice 70-8. Loose-
     leaf pub.

90.  USDA Cooperative State Research Service, USDA/State Chlorobenzilate
     Assessment Team.  An economic and biotic evaluation of Chlorobenzilate
     and its alternatives for rust mite control on citrus in Florida; Re-
     camended Chlorobenzilate treatments and evaluation of alternative
     treatments on citrus in California; the state of Chlorcbenzilate use
     in Texas citrus	1977.  1977, Washington, D.C.

91.  USDA, 1977b.  Fruit Situation Economic Research Service, U.S. Dept.
     of Agriculture.  TFS-205.

92.  USDA Economic Research Service, 1974.  Farmers use of pesticides
     in 1971.... Quantities.  Washington, D.C.

93.  U.S. Dept. of Health, Education, and Welfare (HEW), 1969.  Report
     of the Secretary's Commission on Pesticides and Their Relationship
     to Environmental Health.  Washington, D.C.

94.  Wilson, 1973.  Environment and Birth Defects.  Academic Press, New
     York.

95.  Hblfe, H.R., H.F. Durham, and J.F. Armstrong, 1967.  "Exposure of
     Markers to Pesticides."  Arch. Environ. Health. 14:622-633.  (Cited
     in Severn, 1978).

96.  Wbodard Research Corporation, 1966.  Chlorobenzilate safety eval-
     uation by dietary feeding to rats for 104 weeks: final report.  Geigy
     Chemical Corporation, Yorkers, New York.   (Unpublished).  CONFIDENTIAL.

-------
                           REBUTTAL SUBMISSIONS


1.  Alikonis, Robert J., June 25,  1976;  Crystal Manufacturing Corpora-
    tion; [19:30000/3].

2.  Alikonis, Robert J., August 27, 1976;  Crystal Manufacturing Corpora-
    tion; [19:30000/3],

3.  Alikonis, Robert J., August 24, 1976;  Crystal Manufacturing Corpora-
    tion; [19:30000/3],

4.  Alikonis, Robert J., June 25,  1976;  Tower Chemical Co.;  [11:30000/3].

5.  Alikonis, Robert J., August 25, 1976;  Tower Chemical Co.  [17:30000/3].

6.  Alikonis, Robert J., August 27, 1976;  Tower Chemical Co.;  [23:30000/3A],

7.  Alikonis, Robert J., Sept. 9,  1976;  Tower Chemical Co.;  [33:30000/3].

8.  Alikonis, Robert J., June 25,  1976;  Trans. Chemical Ind.  Inc.;  [10:
    30000/3].

9.  Alikonis, Robert J., June 25,  1976;  Trans. Chemical Ind.  Inc.;  [18:
    30000/3].

10. Alikonis, Robert J., August 27, 1976;  Trans. Chemical Ind. Inc.;  [25:
    30000/3C].

11. Altraan, Melvin D., June 25, 2976; Shanna; [15:30000/3].

12. Baldi, A., August 11, 1976,; Aceto Agricultural Chemicals Corp.;  [34:
    30000/3].

13. Baldi, A., August 26, 1976; Aceto Agricultural chemicals Corp.;  [28:
    30000/3].

14. Balser, Richard L., Andrew P.  Jovanovich, August 27, 1976; Agrico
    Chemical Co. and Western United Resources, Inc.; [20:30000/38],

15. Davis, Kent J., DVM, June 7, 1976; [2:30000/3].

16. Goldberg, Melvin, June 10, 1976; Solchem Inc.; [4:30000/3].

17. Goldberg, Melvin, August 26, 1976; Solchem inc.; [29:30000/3],

18. Griffiths, J.T., June 22, 1976; Florida Citrus Mutual; [8:30000/3],

19. Hall, Graydon B. by Ross A. Pinbauch,  June 2'-, 1976; County  of  Santa
    Barbara, CA, Office of Agricultural Ccranissi  er;  [13:30000/3].
                                                                              103-

-------
20.  Hinkle, Maureen K., August 27,  1976;  Environmental Defense Fund;
     [21:30000/31.
21.  Hunkapiller, Paul,  June 25, 1976;  Helena  Chemical Co.;  [9:30000/3].
22.  Hunkapiller, Paul,  August 27, 1976;  Helena Chemical Co.;  [26:30000/31.
23.  Jeppson, L.R.,  June 1, 1977; University of California,  Riverside
     Dept. of Entomology; [35:30000/3].
24.  Jovanovich, Andrew P., May 17,  1976;  Agrico  Chemical  Co.;  [1:30000/3].
25.  Jovanovich, Andrew P., August 30,  1976; Agrico Chemical Co.;  [31:30000
     /3J.
.26.  Jovanovich, Andrew P., October 25, 1976;  Western United Resources Inc.;
     [20A: 30000/3].
27.  Kerdrick, Janes B. Dr., L.T. Wallace, August 24, 1976;  State  of Califor-
     nia, Dept. of Food & Agriculture;  [27:30000/31.
28.  Lichty, Ralph W., June 15, 1976; Industry Committee on Citrus Additives
     & Pesticides, Inc.; [5:30000/3],
29.  Lichty,. Ralph W., Oct. 5, 1976; Industry Committee on Citrus  Additives
     & Pesticides, Inc.; [5A:30000/3].
30.  Little, Kenneth K., June 23, 1976; County of San Diego; [14:30000/3].
31.  McCog, Clayton W., Sept. 24, 1976; University of Florida, Institute
     of Food b Agricultural Sciences;  [30:30000/3].
32.  McKown, Bobby F., June 18, 1976; Florida Agricultural Research  Insti-
     tute;  [7:30000/3].
33.  Murphy, R.T., June 7, 1976; Ciba-Geigy; [3:30000/3].
34.  Murphy, R.T., Sept. 7, 1976; Ciba-Geigy;  [32:30000/3],
35.  Murphy, R.T., Nov. 4, 1976; Ciba-Geigy; [22A.-30000/3B].
36.  Vice,  Bob L., June 17, 1976; San Diego County Farm Bureau;  (6:30000/3).
37.  Weidenfeld, Edward L., June 28, 1976; Attorney foe Agriro Co.;  [16:
     30000/3].
                                                                              104-

-------
          APPEMJDC A
SUMMARY OF OTHER CANCER STUDIES
                         Preceding page Wank



                                                      106

-------
                        HonyHazelton Study, 1955*
Dose
(PP»)
0

50
500

Nur.ter
Sex
M
F
M
M
F
Initial
20
20
20
20
20
Survived
16
12
13
14
14
Ccnplete
Necropsy
4
10
5
6
6
Rats with
Tunors
2
1
2
2
5
*  Horn,  H.J.,  et  al., J. flgri^. Food Chan., 3s752-756, 1955.

-------
                                             Vtoodard Research Corporation
                                               19S6 TVo-year Rat Study
                       Dose                      Number of Animals             Animals With
                        (ppm)       Sex    Initial   Survived     NecropsieJ    Liver Sarcomas


                         0         H       30          8             5               1
                                   F       30         1.1             5

                         40         M       30          5             5
                                   f       30         14             5

                       125         M       30         10             6               4
                                   F       30         14             5

                       400         M       30         10             5
                                   F       30         15             5
5
/\

-------
          SUMAKY* OF RECENTLY COMPLETED NCI  BIOASSAY ON FEMME OSBORNE MQCBL RATS

Dose (ppm)
0




1175




2350




Site
Mamnary Gland
Pituitary
Thyriod
Uterus Nos
Subcut Tissue/Flank
Maimary Gland
Pituitary
Adrenal
Pancreatic Islets
•Hiyriod
Manraary Gland
Pituitary
Adrenal
Uterus Nos
Thyroid
Tumors
Fibcoadencra Nos
Chranophobe Adenoma
C-oell Adenoma
Endometrial Stromal Polyp
Hemang ioearcona
Pibroadenona Nos
Chranophobe Adenana
Cortical Adenoma
Islet-cell Adenoma
Follicular-cell Adenona
Fibroadenona Nos
Chronophobe Adenana
Cortical Adenoma
Endcnetrial Stromal Polyp
Follicular-cell Carcinoma
Mnber of
Incidence Animals
15 /50
14
5
4
2
14 /4U
11
2
2
2
16 /48
U
5
4
3
*  Provided by NCI,  1977
                                    (iii)

-------
                        SUMMARY* OP RECENTLY COMPLETED NCI BIQASSAY ON MAI£ 06BORNB MENDEL RATS
Dose (ppm)
                             Site
                                       Tumors
                                                    Incidence
                                      Hunter of
                                      Animals
               Thyroid
               Spleen Nos
               Pituitary
               Thyroid
               Urinary Bladder Nos
                         Follicular-cell carcinoma     4
                         Hemangiosarcuma               4
                         Chranophobe Adenoma           4
                         Follicular-cell Adenoma       3
                         Papillona Nos                 3
                                                                                              /49
1600
Pituitary
Adrenal
Multiple Organ Nos
Thyroid
Subcut Tissue/Back
Chranophobe Adenoma           7
Cortical Adenoma              6
Malig.Lynphoma Histiocyti     2
Follicular-cell Adenoma       2
Bemangiosarooma               2
                                                                                              /50
3200
Pituitary
Thyroid
Adrenal
Subcut Tissue/Axilla
Thyroid
Chranophobe Adenoma           4
Follicular-cell Adenoma       4
Cortical Adenoma              3
Fibruna Nos                   2
Follicular-cell Carcinoma     2
                                                                                              /50
              *  Provided by NCI, 1977
s
/I

-------
     APPENDIX B
OTHER TOXICrrc DMA

-------
                       BODENT SUBACOTE TCKICm
Bats
44 weeks at 40 ppm
                       44 weeks at 800 ppm
                       99-day at 2,500 ppm

                                    500 ppm
                                    100 ppm
                                     20 ppm
histologic damage to
adrenals and pancrease

retanV'd growth;
incres.^d htsncpoietic
activity

atrcphic testes;

no effect
Sensitivity -
                        90-day rat  feeding  study     no effect
                        of  2,500 dichlorobnezilate
                             Miscellaneous
         r'bbit:  eye  irritation - severe primary skin irritation -
                  slight to moderate

         hunanr.:  repeated skin patch test - no priroary irritation or
                  sensitizati'jn
Neurotoxicity -

         Not tested

-------
                          MCJIftSENIC lESTS


Negative in these systems:


          E. ooli

          S. typhimurium
             Host mediated with S. typhimurium

          Bacillus  subtil is
                                                    Preceding page blank

                                                                         114

-------
                       FISH ACUTE roxicrrc
Rainbow Trout





Sheephead Minnow
Bobwhite Quail
Mallard Dude
48-hour DC
          50
48-hour 1C
                                     50
                       AVIAN ACUTE TOXICITY
  7-day
                                     50
  5-day DC
                                     50
                      RODENT ACOTE
Rats

Mice

Rat

Rabbit

Rabbit
Oral 1C
50
Oral IX)
50
Dermal ID
50
Dermal 1C
50
Inhalation 1C
  0.71 mgA





  1.0  iog/1
  3,375 ppn
> 8,000 ppm
                                     50
                                                      702 ng/feg





                                                      729 mg/kg





                                                   > 4 g/hg/clay





                                                    > 10.2  g/kg





                                               > 21? < rog/1 air

-------
                                                   CHRONIC TOXICm
              Rats                           2 years at 50 ppn     Slight decreased  size of
                                                                   testes;  slight growth effect

                                                    at 500 ppn     Decreased size of testes;
                                                                   retarded growth


              Rats                          52 weeks at 40 ppn     No effect

                                                    at 125 ppn     Testicular asymetry

                                                    at 400 ppn     Testicular asymetry:  markedly
                                                                   reduced  Henoglobin

              Rats                        2 years at 1,600 ppn     Testicular atrophy

                                                     3,200 ppn     Testicular atrophy


              Dogs         2 years;  1 to 14 weeks at 5,000 ppn     Moderate to severe anemia
                                                                   within 14 weeks

                                  20 to 104 weeks at 3,000 ppn     Organ weight changes;
                                                                   Effects  on liver  and  spleen;
                                                                   Hematcpoiesis

                                                       500 ppn     to effect level
K
H
05

-------
               APPENDIX C




RISK CONSIDERATIONS REIATIN3 TO PESTICIDE




     SUBSTITUTES FOR CHLOFOBENZILA.TE

-------
                               APPENDIX C
                RISK CONSIDERATIONS BELKTIUS TO PESTICIDE
                     SUBSTITUTES FOR QgflROBENZILME
I.  FEMBUTftTIN-OXIDE

       Preliminary inspection of Registration Division file data indi-
cates that toxic effects have been reported for test animals exposed to
fenbutatin-oxide in several different studies (Burnara, 1978).  Decreased
liver, brain, spleen, and kidney weights and decreased body weight gain
was reported for animals ingesting 100, 300 and SCO ppm fenbutatin-oxide
during the first three months of a two-year chronic feeding study, and for
animals ingesting 500 and 100C ppm during a 28-day subacute study (Shell,
Proprietary).  Serum alkaline phosphatase was elevated at 300 and 600 ppm,
indicating tissue injury at these doses.  The no-effect level is 100 ppm
for the 3-month exposure and 250 ppm for the subacute exposure (Shell,
Proprietary).  These changes are indicators of general toxicity.
       In the chronic study the test is weights and the testis-to-body-
weight ratios were increased at 300 ppm and 600 ppn fenbutatin-oxide, but
these changes were not accompanied by hypertrophy or other fenbutatin-
oxide related changes.  Later  in the two-year study, animal growth rates
were normal and no fenbutatin-oxide related tumors or lesions were reported.
       Adverse reproductive effects were reported in a three-generation
reproduction study in rats, with two litters  in each generation.  The
viability index was seriously  reduced in the first litter of the first
generation at 300 ppm and moderately reduced in both third generation 1it-
                                                                              118-

-------
ters at this dose level.  At 300 ppm, the parents and pups were smaller
in size, were reported to be somewhat hyperactive and irritable at tines,
and the lactation index was reduced in the second litter.  Weaning body
weights were reduced in five of the six litters produced during the study.
At 100 and 300 ppm, there was a statistically significant decrease (p •
0.05) in the testis to body weight ratio, but at 50 ppn the test and con-
trol data were ccoparable.  The no-effect level for all parameters includ-
ing testis weight was 50 ppm (Shell, Proprietary, 1974).
       Because fenbutatin-oxide and chlcrobenzilate are applied in the
sane manner, the chlorobenzilate exposure estimates were applied to make
the preliminary fenbutatin-oxide risk assessments.  Assuming that spray
applicators may experience dermal exposure of 0.57 rag/kg and based on a
no-effect level of 100 ppm, the margin of safety for subacute exposure
                                            I/
to an applicator would be approximately 8.7.
I/  The absorption  factor  for  fenbutatin-oxide  is unknown  and probably
    less than 10%.  If 10% is  used, the exposure would be  0.57 mg/kg.
    Based on a no-effect level of 100 ppm,  the  margin of safety  for
    subacute exposure to an applicator would be:  100 ppn  » 5 mg/kg,
    then 5 mg/kg divided by 0.57 mgAg * approximately 8.7.
    The margin of safety of 8.7 should be put in perspective.  Since expo-
    sure is estimated at 0.57  mg/kg/day tines 100 * 57 mg/kg, and  1 mg/kg
    - 20 ppra in the rat's  diet.  Total dietary  exposure is (57 x 20) 1140
    ppm.  Very few  pesticides  can be fed at levels  of 1140 ppm and those
    having a no-effect level at 1140 are very,  very few in number.
    The only conclusion is that any pesticide applied in a manner  similar
    to chlorobenzilate would have a margin  of safety or safety factor of
    considerably less than 100 (Burnham, 1978).
                                 tii)

-------
17.  DIOOPOL

       In comparing dicofol with chlorobenzilate ths following facts are
pertinent:  a) both compounds induce bepatocellular carcinomas in male
mice; b) chlorobenzilate induces the same kind of lesion in female mice
but at a lower rate, and dicofol has no carcinogenic effect on female
mice; c) neither compound induces a significant tumorigenic response in
maleor female rats; and d) the corapcunds have similar chemical structure,
implying that their mechanism of tumor induction may be tl » same.  In
view of these facts, it is legitimate to compare the potency of the cam-
pounds.  This is done by taking the ratio of the one-hit slope parameters
from the NCI experiments.  The results are B(dicofol)/B (chlorobenzilate)
« 2.40 x 10~3/2.02 x NT4 - 11.9.  Under similar test conditions, di-
oofol is about 12 times more potent than chlorobenzilate (Albert, 1978b).
III.  NDM-SEI£CTIVE SUBSTl'lUIES

     A.  Wildlife

       Ethion, sulfur, propargite, and carbophenothicn do not appear to
present significant acute toxicity risks to wildlife from use on citrus
(Boshong, 1977).

     B.  Aquatic Organisms

       Acreages involved in citrus uses of chlorobenzilate are large
enough to be contiguous with biologically significant todies of water
                                 (iii)
                                                                               120-

-------
and the alternatives, ethion and carbophenothion, appear to present more
risk from acute effects than chlorobenzilate (fish kills likely to occur
in contiguous waters).  Propargite appears to present about the sane acute
risk as chlorobenzilate while sulfur probably presents no acute risk.
     C.  Human Effects

        Use of ethion and carbophenothion present greater risk due to acute
toxicity than chlorobenzilate as evidenced by tests of laboratory animals
and pesticide episode data (Burnam, 1978).  It appears that there would he
a greater potential hazard in association with immediate injury to those
occupationally involved during or subsequent to use of these substitutes
on citrus.
                                 (iv)

-------
    APPENDIX D




LABEL REQUIREMENTS

-------
                              APPPPIX D
            THE FOLLOWItC STATEMENTS MUST APPEAR ON THE IABELS
             OF PESTICIDE PBODUCTS CONTAINING CHDOBOBENZILftTE

Restricted Use Pesticide
    For retail sale to and use only by certified applicators or persons
    under their direct supervision and only for those uses covered by
    the certified applicators certification.
General Precautions
A.  Take special care to avoid getting chlorobenzilate in eyes, on skin,
     or on clothing.
B.  Avoid breathing vapors or spray mist.
C.  In case of contact with skin, wash as soon as possible with soap
    and plenty of water.
D.  If chlorobenzilate gets on clothing, remove contaminated clothing
    and wash affected parts of body with soap and water.  If the extent
    of contamination is unknown, bathe entire body thoroughly.  Change
    to clean clothing.
E.  Nash hands with soap and water each tine before eating, drinking,
    or socking.
F.  At the end of the work day, bathe entire body with soap and plenty
    of water.
G.  Wear clean clothes each day and launder before reusing.

Required Clothing and Equipment for Application
A.  Fine weave cotton fabric  (Jersey), one-piece junpsuit, long sleeves,
    long pants.
B.  Wide-briimed hat.
C.  Heavy-duty fabric work gloves.
D.  Any article which has become contaminated must be replaced.
                                 (i)
                                                                                 123-

-------
B.  Pace-piece respirator of the type approved f-r pesticide spray applica-
    tion* by the National Institute for Occupational Safety and Health.

F.  Instead of the above specified clothing and equipment, the applicator
    can use an enclosed tractor cab which provides positive pressure and
    a filtered air supply.  Aerial application nay be conducted without
    the aeove specified clothing and equipment.


Handling Precautions


A.  Heavy-duty rubber or neqprene gloves and apron roust be worn during
    loading, unloading, and equipment clean-up.
                                 (ii)

-------