-------
TabU 5.  (Continued)
Sj>sc!ss
Tissue
B loconcmtrat 1 on
Cheaica! Factor
Duration
iday»}
Haler«nc«

Other Arsenic Compounds
Cladocaran,
Oaphnla Magna
Cladoceran,
Oaphnla Maona
Scud,
GawMrus ps«udol tMnaeus
Scud,
GAMMBTUS pseudol iMnaeus
Snail,
He 1 1 sowa caMpanu 1 at a
Snail,
Heiisoma caMpanulata
Snaii,
Stagnicola eMarglnata
Snail,
Staynicoia uMTyinata
Stonaiiy,
Pteronarcys dor sat a
Stonetly,
Ptsroftarcys dcrsata
Ratabos trout.
Sal MO galrdner 1
Kalnbow trout.
Sat BO gaJrdner 1
Whole body
Whole body
Whole body
Whole body
Whole body
Whole body
Whoie body
Whole body
Whole body-
Whole body
Whole body
Whole body
Dl sod I UK M«thy 1
ar senate
SodluM dlMethyl
ar senate
DIsodliM Methyl
ar senate
£odluM dlMethyl
ar senate
DIsodluM Methyl
ar senate
SodluM dlMethyl
ar senate
DisodiuM Methyi
arsenatn
SodluM dlMethyl
arsenafv
DisOuiuM «etny t
ar senate
SodluM dlMethyl
ar senate
Oisodtua asthy!
ar senate
Sodlui* dlMethyl
ar senate
4
4
0
0
4
5
3
2
9
7
0
0
2!
21
28
28
28
28
28
28
28
28
28
28
Spehdr, et a!.
Spdhor, et ol.
bpahar . et sU
Spehar , et at.
Spohar, et al.
Sptthar, et 
-------
Tatt* S.  
-------
       Table 6.  Other data for arsenic
Chemical
Result
«   Reference
«B^U^_«B^B» •— «^»^™™.— «»^»W^B~i«»— • 4MB— M«^^ -»*«^M» 1 . ' 	 	 ' ~ -
FRESHWATER SPECIES
Trlvalent Inorganic Arsenic
C 1 adoceran,
(not specified)
Cl«doceran,
Papnnla Magna
Cladoceran,
(not specified)
Copepod 1 ado It),
(not specified)
Copopod,
(not spec Ml ad)
Rotifer,
(not specified)
Rotifer,
(not specified)
AwpMpod,
HyaUUtt knlcKwboctierl
A^blpod.
GaMurus psaudol lonaeus
Mayfly (ny^th),
Caen Is dlalnuta
Hayfly (ny^in),
Caen Is dlalnuta
Mayf ly.
Calllbaatls sp.
Toad (eabryo- larval) .
Gastruphryne carol Inensls
Sodium
arsenlte
Sodlun
arsenlta
SodluM
arsenlte
SodliM
arsenlte
Sodlun
arsenite
SodliM
arsenlte
Sodlu*
arsenl te
Arsenic
trlGKlde
Arsenic
trloxlde
Arsenic
trloxlde
Ar sen 1 c
trloxlde
Arsenic
trloxlde
Sodium
arsenlte
1 ok
26 nrs
16 «ks
16 wKs
1 Mk
t *k
16 Mks
5 days
7 days
5 days
5 days
5 days
7 days
Significant pop-
ulat Ion reduction
IC50 (Median
(Mobilization)
Reduced population
(one treatment)
Reduced population
(weekly treatments)
Significant popu-
lation reduction
Significant popu-
lat Ion reduction
Reduced population
(•onthly treatments)
70$ Mortality
80| Mortality
25> Mortality
62} Mortality
94$ Mortality
LC50
2,120
3,770
690
690"
2,320
2,320
690"»
4,469
961
2,234
5.958
4,469
40
Cowell, 1965
Crosby & Tucker. 1966
Glldernus, 1966
Gllderhus, t%6
Cowell, 1965
Cowell. 1965
Gllderhus, 1966
Surlier & Meehean,
1931
Spehar, et al. I960
Surber I Meehean,
1931
Surber & Meehean,
1931
Surber 1 Meuhuan,
1931
Utry». 1979
                         B-23

-------
Table 6.  (CoetlMUed)
                                                                        Result
Species
Rainbow trout
(e»bryo-larval).
Sales galrdnerl
Rainbow trout (Juvenile),
Sales galrdnerl
Brook trout,
Sal veil BUS tontlnalls
Goldfish (Juvenile),
Carasslus auratus
Goldfish (e*bryo- larval),
Carasslus auratus
Spot tall shiner,
Motropls hudson 1 us
Fathead *lnnow (juvenile),
PlMphalas promelas
Fathead •Innow,
PlneplMtles proaalas
Olueglll (Juvenile),
Lepoals *acrochlrus
Blueglll (adult),
Lepoals Mcrochlrus
Blueglll (Juveniles),
Lepoals *acrochlrus
Blueglll (flngerllng),
Lepoals *acrochlrus

Cheeiical
Sodlu*
arsenlte
Arsenic
trloxlde
Sodlu*
arsenlte
Sodlu*
arsenlte
Sodlu*
arsenlte
Sodlu*
arsenlte
Sodlu*
arsenlte
Arsenic
trlsulflde
Sodlu*
arsenlte
Sodlu*
arsenlte
Sodlu*
arsenlte
Sodlu*
arsenlte
(palletized)
Duration
28 days
21 days
262 hrs
336 hrs
7 days
72 hrs
336 hrs
96 tirs
16 wks
)6 wks
336 hrs
48 hrs
Effect
LC50
Decrease In fat
weight gain
LC50
LCM
LC50
LOO
LC50
LC50
Reduced survival
(one treatment)
Hlstoparhologlcal
alterations
(weekly treat*ents)
LCSO
LC50
MO
1,000
10,440
18,618
490
27,000
82,400
690
690"
18,328
290
Reference

Blrgo. 1979
Speyer, 1974;
Speyer & Leduc, I97i
Cardwel 1. et al. 1976
Cardwell, et al. 1976
tilrye, 1979
Boschettl 4
McLoughlln, 1957
Cardwell, et al . 1976
Curtis, et al. 1979
Gllderhus, 1966
GHderhus, 1966
Cdrdwell, et al. 1976
Huyhes &, Davis. 1967
                                                3-24

-------
Table 6.
                                                                            Result
Spaclas
Reference
Pentavalent Inorganic Arsenic
Cladocaran.
Daphnla aaana
Cladocaran.
Oaphala »agna
Green Minflsh C Juvenile).
Lepoals cy ana II us
Green sunflsh,
LepOMlS cyanallus
Green sunflsh,
Lepoals cyan* II us
Green sunflsh,
LefKMls cyanellus
Green sunflsh,
Leoomts Chanel lus
Green tuntlsh,
Lepceils cyanallus
Green sunflsh,
Leponls cyanallus

Sodlua
ar senate
SodliM
ar senate
Sodlu*
ar senate
SodliM
ar senate
SodliM
ar senate
SodliM
ar senate
ar senate
ar senate
SodliM
ar senate
3 MkS
3 «ks
39 hrs
2 Mks
678 hrs
210 hrs
124 hrs
527 hrs
209 hrs
SALTWATER
LC50
Chronic Malts
LT50
Ultrastructural
changes In liver
LT50 at 10 C
LT50 at 20 C
LT50 at 30 C
LT50 at 20 C
LT50 at 30 C
SPECIES
2.850
520-
1,400
40.000
31,700
60,000
60,000
60,000
30.000
30,000
Bleslnyer &
Chrlstensen, 1972
Bleslnger A
Chrlsten&en. 1972
Sorenson, I976a
Sorenson, I976b
Sorenson, I976c
Sorenson, I976c
Sorenson, I976c
SoreniOfi, 1976c
Sorenson, I976c
Trlvalent Inorganic Arsenic
Red alga,
Pluaarla a lagans
Polychaeta wor«,
Nereis dlverslcolor
SodliM
arsenlte
SodliM
arsenlte
16 hrs
192 hrs
7 day post expo-
sure - arrested
development o<
spore) Ings
LC50
577
>I4,500
booey, at ol. 1959
bryan, 1976
                                                  B-25

-------
Tabla 6.  (CoKtlawad)
Spacla»
Mud SMll.
NaSSarluS ObSOlatMS
Bay scallop (Juvanlla),
Argopactln Irradlans
Mhlta ihrlap
(Juvanlla).
Paftaauft sallfarus
Pink sal»on,
Oncorhynchus oprfauscha
Pink calaon,
Oncorhynchus ggrbuscha
Pink MlMon,
Oncorhynchu* oprbuscha
Chua sal aon,
Oncortiynchut kata

Cnwilcal
Sod) u>
ar$*nlta
SodlUM
arsanlt*
Arsanlc
trlMilflda
Arsanlc
trloKloa
Arsanlc
trlCKlda
Arsanlc
trlOKlda
Arsanlc
trlcKlda
Racult
Duration EMaet (na/l»* Rataranca
72 hrs 09 consuaption >2,OOU Haclnnas 1 Thurbarg
rtta 4*pratsad and 1973
abnormal bahawlor
4 days Btoconcantrat Ion - Nalson, at al. 1976
factor * 13
96 hrs LC50 24,700 CurtU, at at. 1979
96 hrs LCI 00 12,507 Holland, at at. I960
7 days LCIOO 7,195 Holland, at al. I960
10 days LC54 3,787 Holland, at al . I960
48 hrt LC50 6,330 Alderdlce t Bratt,
1957
•  Rasults ara axprassad as arsenic, not as tha co*pound«
*• Maasurad concantratlon attar 16 waaks was 9.040 |tg/l
••"Maasurad concantratlon attar 16 waaks «as 2,280 tig/1
                                                     B-26

-------
                                  REFERENCES

Alderdice, O.F.  and  J.R.  Brett.   1957,   Toxicity of sodium arsenlte to young
     salmon.   Prog.  Rep.  Pacific Coast  Stat. Fish. Res. 8d. Canada.  108: 27.
Anderson, A.C., et al.   1975.   The  acute  toxlclty of MSMA to black bass (Mj_-
cropterus dolomieu),  crayfish (Procambarus sp.), and channel catfish ( Ictal-
urus lacustris).  Bull. Environ. Contam. Toxicol.  14: 330.

Anderson, B.G.   1946.  The  toxicity  thresholds of  various  sodium salts de-
termined by the use of Daphnla maqna.  Sewage Works  Jour.  18: 82.

Anderson,  M.A.   1979.    Personal  communication.   University of  Wisconsin,
Water Chemistry Program, Madison, Wisconsin.

Biesinger, K.E.  and  G.M. Christensen.   1972.   Effects of  various metals on
survival,  growth,  reproduction,  and  metabolism  of  Daphnia maqna.    Jour.
Fish. Res. Board Can.  29: 1691.

Birge,   W.J.   1979.   Aauatic Toxicology  of Trace  Elements  in  Coal  and Fly
Ash.  In: Energy and Environmental  Stress 1n  Aouatlc Systems.   Thomas Hunt
Morgan  School  of Biological  Sciences, Lexington,  Kentucky.   Sel.  Water Res.
Abs. 12, W79-09248.
                                     B-27

-------
Boney, A.O., et  al.   1959.   The effects of various poisons on the growth and
vitality  of  soorelings of  the red  alga  Plumaria elegans.   (Bonnem.)  Schm.
Schm. Biochem. Pharmacol.  2:  37.

Boschetti, M.M.  and  T.F.  Mclouohlin.  1957.  Toxicity of  sodium arsenite to
minnows.  Sanitalk.  5: 14.

Bryan, G.W.  1976.   Heavy Metal  Contamination  in the Sea.   Ir±' Marine Pollu-
tion, Part 3.  Academic Press.

Calabrese, A., et  al.   1973.   The  toxicity of  heavy metals to embryos of the
American oyster, Crassostrea virginica.  Mar. Biol.  18:  162.

Cardwell, R.D.,  et al.  1976.   Acute  toxicity of selected toxicants to six
species of  fish.   Ecol. Res.  Series EPA 600/3-76-008.  U.S.  Environ.  Prot.
Agency,  o. 125.

Clemens,  H.P.  and K.E.  Sneed.   1959.   Lethal  doses  of  several  commercial
chemicals for fingerling channel catfish.   U.S.  Fish Wildl. Serv.  Sci.  Rept.
Fish.  No. 316, Washington, D.C., U.O. Oep. Inter,  p. 10.

Cowell, B.C.  1965.  The effects of  sodium  arsenite  and  silvex on  the plank-
ton populations in farm ponds.  Trans. Am. Fish. Soc.  94:  371.

Crosby,  O.G.  and  R.K. Tucker.   1966.   Toxicity of  aauatic herbicides  to
Oaohnia magna.  Science.  154: 289.
                                     B-28

-------
Curtis,  M.W.,  et  al.   1979.  Acute  toxicity of  12  industrial chemicals  to
freshwater and saltwater organisms.  Water Res.  13: 137.

Ferguson, J.F.  and J.  Gavis.  1972.   A  review of  the  arsenic cycle  in natu-
ral waters.  Water Res.  6:  1259.

Fish  Pesticide  Research   Laboratory.   1980.   Unpublished  laboratory  data.
Columbia, Missouri.

Fowler,  8.A.   1977.    International  conference on  environmental  arsenic: An
overview.  Environ. Health Perspect.   19: 239.

Gilderhus, P.A.   1966.  Some  effects  of sublethal concentrations  of sodium
arsenite  on  bluegills  and  the anuatic environment.   Trans. Am.  Fish.  Soc.
95: 289.

Hale, J.G.   1977.  Toxicity of metal  mining  wastes.   Bull.  Environ. Contam.
Toxicol.  17: 66.

Holland, A.A., et  al.   1960.  Toxic effects  of  organic  and inorganic pollu-
tants  on young  salmon and  trout.   State  of  Washington,  Dep.  Fish.  Res.
3ull.  No. 5.

Holm,  T.R.,  et  al.    1979.   Reprinted  from  ACS   Symposium  Series  No.  93.
Chemical Modeling in  Aoueous Systems,  E.A. Jenne, (ed.).   Am. Chem. Soc.
                                     B-29

-------
Huqhes, J.S. and  J.T.  Oavis.   1967.   Effects  of Selected Herbicides on Blue-
gill  Sunfish.   ]£:  Proc.   18th  Ann.  Conf.,  S.E.  Assoc.   Same  Fish  Comm.,
October  18-21,   1964.   Clearwater,  Florida.    S.E.  Assoc.  Game Fish  Comm.
Columbia, S.C.   D. 480.

Inglis, A.  and  E.L.  Davis.   1972.  Effects of  water  hardness  on the toxici-
ty  of  several  organic  and  inorganic  herbicides  to fish.   Bur. Sport  Fish
Wildl. Tech. Paper 67.  U.S. Deo. Inter,  p. 22.

Maclnnes, J.R.  and R.P.  Thurberg.   1973.   Effects of metals on  the behavior
and oxygen consumption of the mud snail.  Mar. Poll. Bull.   4:  185.

Nelson, O.A., et  al.   1976.   Biological effects of heavy  metals on juvenile
bay scallops, Arqooecten  irradians,  in short-term exposures.  Bull. Environ.
Contam. Toxicol.  16: 275.

Nelson, K.W.  1977.   Industrial  contributions of  arsenic to the environment.
Environ. Health Perspect.  19: 31.

Sanders, H.O. and O.B.  Cope.   1966.   Toxicitles of several pesticides to two
soedes of cladocerans.  Trans. Am. Fish. Soc.  95: 165.

Sanders, H.O. and O.B.  Cope.   1968.   The relative toxicities of several pes-
ticides  to  naiads  of  three  species  of  stoneflies.   Limnol.  Oceanogr.
13: 112.
                                     8-30

-------
Sorenson, E.M.8.   1976a.   Toxicity  and  accumulation of arsenic in qrsen



fish,  Lepomis  cyanellus,  exposed to arsenate  in  water.   Bull. Environ. Con-



tarn. Toxicol.  15: 756.







Soreson, E.M.8.   1976b.   Ultrastructura!  changes  in the hepatocytes of green



sunfish, Lepomis  cyanellus Rafinesoue,  exoosed to  solutions  of  sodium arse-



nate.  Jour. Fish 3iol.  8: 229.







Sorenson, E.M.3.   1976c.   Thermal  effects on  the  accumulation of arsenic  in



green  sunfish, Leoomis cyanellus.  Arch. Environ.  Contam. Toxicol.  4; 3.







Soehar,  R.L.,  et  al.   1980.   Comparative toxicity of  arsenic compounds and



their  accumulation  in  invertebrates  and  fish.   Arch.  Environ.  Contam.



Toxicol.  9: 55.







Soeyer,  M.R.   1974.   Some  effects  of  combined  chronic  arsenic  and  cyanide



ooisoning on the  physiology of  rainbow  trout.   M.S. Thesis,  Concordia Univ.,



Montreal, Canada.







Soeyer, M.R. and  G.  Leduc.  1975.   Effects  of Arsenic Trioxide on Growth of



Rainbow Trout.  In:  International Conference on Heavy Metals in the Environ-



ment.  Toronto, Ontario, Canada.  October, 1975.







Surber, E.W. and  O.L.  Meehean.   1931.  Lethal concentrations  of  arsenic for



certain aauatic organisms.  Trans.  Am. Fish. Soc.   61: 225.
                                     8-31

-------
U.S. EPA.   1978.   In-deoth  studies  on health  and environmental  impacts  of
selected water pollutants.  U.S. Environ. Prot. Agency.   Contract  No.  68-01-
4646.

U.S. EPA.   1980a.   Unpublished laboratory data. Environ. Res.  Lab.,  Duluth,
Minnesota.

U.S.  EPA.    1980b.    Unpublished   laboratory   data.    Environ.  Res.   Lab.,
Narragansett, Rhode Island.
                                     8-32

-------
 Mammalian  Toxico1ogy and  Human  Health  Effects
                                    EXPOSURE
 Inqestion  from Water
     In  a U.S.  Environmental  Protection  Agency national  study of residential
 tao water,  66.3  percent  of the one-time  grab samples  collected  from 3,834
 "esidences  had arsenic levels greater than  0.1  ug/1.   The average, minimum,
 and maximum arsenic  levels of  the  samples  were 2.37, 0.50,  and  213.6 ug/1,
 respectively  (Greathocise  and Oaun,  1978).   In 1975  it  was  reported  that 5
 out of  566  samples  collected from Interstate Carrier Water Supplies exceeded
 10  ug/1  and that  the maximum level  was 60 ug/1 (U.S. EPA, 1975).  Well water
 samples  collected during  1976 at 59  residences in a  Fairbanks.  Alaska sur-
 burban  community  had a mean  arsenic  content  of  224  ug/1 with a  range from
 1.0 to  2,450  "9/1  (u-s-  Public  Health  Service,  1977).   Valentine,  et  al.
 (1979)  reported that arsenic levels  in  the water  supply from five communi-
 ties  (Fairfax  and Edison  in Sakersfield.  California;  and Virginia Foothills,
 Hidden  Valley,  and  Fallon  in  Nevada)  to  be 6,  393, 51, 123. and 98 ug/7,  re
 spectively.
    There have  been  a  number  of  other reports  of isolated instances of high-
 er  than  usual  concentrations  of arsenic  in  well  waters.  Goldsmith,  et  al.
 (1972)  reported on  a study, in Lassen County,  California,  of the health  ef-
 fects associated  with  drinking  well waters  with  arsenic  levels  ranging from
 100 ug/1 or less  to  1,400  ug/1.    In  Perham,  Minnesota,  a newly  bored well
was associated with  illness in  13 people whose hair samples  contained  arsen-
 ic  at 37-1,680  ug/g.  The well  water  serving these patients contained  arsen-
 ic  from 11,800 to 21,000 ug/1; this was  later  determined  to  come  from  ground
 contamination  by  residual   arsenical  grasshopper  bait  (Feinglass,  1973).
                                     C-l

-------
forty-five out  °f  558 water  samples  collected from Lane  County,  Oregon  had
arsenic values greater than  50  ug/1.   The mean, maximum,  and  minimum  values
detected were 9.6,  2,150, and 0 ug/1,  respectively (Morton, et  al.  1976).
    ^ch information  nas  been collected concerning the  levels of  arsenic  in
fresh  surface  waters  (Table  1).   Arsenic  occurrence  is very  widespread  and
even occurs in  some rain  water.   Most  of the high values  reported  in  rivers
and lakes  are  probably due to industrial contamination  [National  Academy  of
Sciences (NAS), 1977al.   Angino, et al.  (1970)  have shown  that household  de-
tergents  (mostly  of  the  high-phosphate type)  widely  used   in  the  United
States contained arsenic  at  1-73  ug/g;  their use probably contributes  sig-
nificant amounts  of  arsenic  to surface  sources.   Sollins (1970),  however,
felt that, after dilution  during  use,  the concentration would be  well  below
the recommended maximum  and  constitute no particular  hazard.  It has  been
Generally  assumed that surface  waters, like the ocean,  are  "self-purifying"
with respect to arsenic  - i.e.,  arsenic is removed from solution  by  deposi-
tion with  sediments;  but  quantitative studies are  lacking.   Sediments  are
always  higher   in  arsenic  than  the waters  with  which  they   are  associated
(NAS, 1977a).
Inoestion from Food
    A  1966 food survey  found  arsenic in  3.2  percent  of  the samples at  a
range of 0.10 to 4.7  ug/g  (Cummings,  1966).   In a  1967 market-basket  survey,
arsenic was present in 10  percent  of  the composite samples (Ouggan and Lips-
comb,  1969).    In  1968,   arsenic  occurred  in  18  percent of the  samples.
Whether arsenic occurred naturally or as  a  result of  man's  activities  was
not known.
    Schroeder  and  Balassa  (1966)  sampled  foods and beverages from American
chain  stores  (Table  2).   Fish   and  seafoods contained  the most  arsenic
                                      C-2

-------
                                                TABLE 1

                                    Arsenic  in Fresh Surface Waters
                                       Arsenic Concentration
           Water
                                        Reference
United States

  Lakes:
    New York, Chautauqua
    Michigan
    Superior
    Wisconsin
    California, Searles
    Florida, Echols
    Florida, Magdelene

  Rivers:
    Hillsborough
    Withlacoochee
    Fox (polluted watershed)
    Yellowstone
    Narrow
    Providence
    Seekoink
    Sugar Creek
     (contaminated)
    Columbia
    Schuylkill

  Canals:
    Florida
    3.5-35.6
    0.5-2.4
    0.1-1.6
    4.0-117
198,000-243,000
    0.0-llOa
    0.0-2,000b
    3.58
    1.75
    0.25
    0.42
  100-6,000
    4.5
    0.90
    0.75-0.90
    2.48-3.45
   10-1,100

    1.6
   30-180
   10-20
Lis and Hopke, 1973
Seydel, 1972
Seydel, 1972
Chamberlain and Shapiro, 1969
White, et al. 1963
Livingston, 1963
Livingston, 1963
Braraan and Foreback, 1973
Braman and Foreback, 1973
Braman and Foreback, 1973
Braman and Foreback, 1973
Brown, et al. 1973
Ellis, 1934
Ray and Johnson, 1972
Ray and Johnson, 1972
Ray and Johnson, 1972
Durum et al. 1971:
 Wilder, 1972
Onishi, 1969
Kopp, 1967
Grantham and Sherwood, 1968
                                                 03

-------
                                             TABLE  1  (cont.)
           Water
Arsenic Concentration
        ng/1
          Reference
Puget Sound

  Rainwater:
    Rhode Island
    Washington, Seattle

Chile

Formosa, Mel) water
       1.5-1,200


       0.82
      17

     800

     800
Crecelius, et al. 1975;
Crecelius and Carpenter,  1974

Ray and Johnson 1972
Crecelius, et al. 1975

Borgono and Greiber, 1972

Fan and Yang, 1969
Dissolved solids, <2,000 mg/1
Dissolved solids, >2,000 mg/1
                                                 C-4

-------
                              TABLE  2

                         Arsenic  in  Foods*
                           (wet weight)
                                   Arsenic              Micrograms of
Food Sample                      Concentration            Arsenic
                                    (ug/D                  100
Fish and sea food
Haddock
Kingf ish
Oysters, fresh
Oysters, frozen
Scalloos, fresh
Shrimp, fresh frozen
Shrimo Shells
Clams, fresh frozen
Conch, fresh
Conch, dried, whole
Meats
Beef, stewing
Pork loin
Pork liver, No. 1
°ork liver, No. 2
Pork kidney
Lamt choo
Chicken breast
Gelatin
Egg lecithin
Vegetables and grains
Wheats, whole
Rye, seed
Corn
Corn meal
Corn oi 1
Corn oil lecithin
Rice, Madagascar
Rice, U.S.
Puffed rice
Kelloggs' Special K*
Cottonseed oil
Beets

2.17
8.86
2.9
2.7
1.67
1.50
15.3
2.52
3.1
5.63

1.3
0.06
1.07
1.4
0.0
0.35
0.0
0.19
0.0

0.17
0.16
0.11
0.78
0.0
0.0
0.48
0.13
1.6
0.66
0.0
0.0

305
886
580
540
160
132
—
525
311
311

58
21
75
98
0
19
0
6
0

5
5
3
22
0
0
13
3
46
19
0
0
                               05

-------
                               TABLE 2 (cont.
     Food Sample
   Arsenic
Concentration
   (ug/i)
*Source:  Schroeder and Balassa, 1966

aCalorie values from McCance and Widdowson, 1947
Micrograms of
 Arsenic per
   100 cala
Vegetables and grains (cont'd)
Seet greens
Swiss charci
Rhubarb
Red pepper
Garlic, fresh
Cherry tomatoes
Yellowpear tomatoes
Turnip
Mushrooms
Soy lecithin
Vegetables, St. Thomas, V.I.
Carrots
Peas, dried
Peas, fresh
Tomatoes, fresh
Egg plant
Ginger
Fruits
Apple
Orange
Pear
Grapes, wild
Miscel laneous
Cocoa, Hershey's®
Coffee
Tea
Salt, table
Salt, sea
Sugar, lump
Sugar, granulated
Milk, evaporated
Milk, dry skimmed
Butter, unsalted

0.24
0.56
0.48
0.06
0.24
0.37
0.10
0.0
2.9
0.0

0.0
0.09
0.0
0.0
0.82
0.0

0.0
0.0
0.0
0.17

0.59
0.0
0.89
2.71
2.83
0.10
0.0
0.17
0.0
0.23

240
215
800
„_
__
264
70
0
414
0

0
3
0
0
546


0
0
0
34

13
0
«.
_
_
3
0
11
0
3
                                   C-6

-------
 and  fruits  contained  the  least.   Generally,  the only foods which are high in
 arsenic are  seafoods.   Chapman  (1926)  found  that mussels,  oysters,  and scal-
 lops  contained  very  high  levels of  arsenic  (means of  up to 30 yg/g).   In
 comparison,  mixed freshwater  fish  contained  a mean  of  0.65 ug/g  arsenic.
 Zook, et  al.  (1976)  reported  that  in  a survey of selected seafoods  for metal
 content,  the overall arsenic  mean  content was 2.6 ug/g.   The   lowest  mean
 arsenic value  was found in wild catfish -0.1 ug/g.  Thus,  arsenic  appears
 to  be present  in  small amounts  in  nearly all  foods,  with marine  inverte-
 brates containing the highest arsenic levels (Table 3).
    A bioconcentration  factor  (BCF)  relates  the concentration of a  chemical
 in aauatic  animals to the  concentration  in  the water in  which they  live.   An
 appropriate  BCF can  be used  with  data  concerning  food  intake  to  calculate
 the  amount  of  arsenic  which might  be ingested from the consumption  of  fish
 and  shellfish.   Residue data  for a  variety of  inorganic  compounds  indicate
 that  bioconcentration factors for the  edible  portion of  most  aquatic  animals
 are   similar,  except  that  for  some  compounds  bivalve  molluscs   (clams,
oysters,  scallops, and  mussels)  should  be  considered  a  separate group.   An
 analysis  (U.S.  EPA,  1980a)  of data from a food survey was used  to  estimate
 that  the   per   capita  consumption  of   freshwater  and  estuarlne  fish  and
shellfish   is 6.5  g/day  (Stephan,  1980).   The  per  capita  consumption  of
bivalve molluscs is 0.8 g/day and that of  all  other freshwater and  estuarine
fish and shellfish is 5.7 g/day.
    Spehar, et al. (1980) obtained bioconcentration  factors of zero  for  four
different   arsenic compounds  in  rainbow  trout,  but  a  8CF  of  4  was  obtained
with  the  blueqill  (U.S. EPA,   1978).   Thus,  the BCF for arsenic  is  probably
about 1.0 for  many  aquatic  animals   a  BCF of  350 was  obtained for  sodium
arsenite  with  in oysters.   If  the values  of  350  and 1  are  used  with  the
                                     C-7

-------
                                   TABLE 3

                   Bioaccumulation Ratio Values for Arsenic
                            in Acuatic Organisms3*
Species
Haddock
Kingf ish
Crustacea and shellfish

Assorted fish
Assorted fish
Shrimp
Mackerel
Cod
Assorted freshwater fish

Arsenic in
Tissue
(ug/I)
2-10.8
8.86
1.5-3.1
0.018-1.06
0.076-2.27

-------
 consumption  data,  the  weighted  average bioconcentration  factor for arsenic
 and  the  edible  portion of  all  freshwater  and estuaHne  aauatic organisms
 consumed  by Americans  is calculated to be 44.
     Oecelius  (1977a)  has analyzed  19 samples  of  domestic table  wines for
 several  species  of  arsenic;  13 varieties  of white  and  red wines  were in-
 cluded.   The  Canoes  of concentrations were <1-420  ug/1,  <1-110  ug/1, and
 <1-530  ug/1  for arsenite,  arsenate,  and  total  arsenic,  respectively.   Mean
 levels  were  127,  32, and  153  ug/1,  respectively.   Clearly, the  majority of
 the  arsenic  was  present as arsenite.  Both  dimethylarsinic  acid  and methyl -
 arsonic were below the detection limit of 1 wg/1 in these wine samples.
     In  1966  Schroeder  and  Balassa (1966)  estimated  that the  average  daily
 diet contains  900 ug of arsenic.   This estimate was based on  the results of
 arsenic  determinations  for meats,  sea food,  and  vegetables  purchased  from
 Vermont chain  stores.   Arsenic  in  an institutional  diet was estimated  to be
 400  ug  per day.   One  reason  for  this  lower  level  is  that  the institutional
 diet did  not  contain  any  seafood.  The  World Health  Organization  reported
 that arsenic intakes vary  from 7 to  60 ug per day  (NAS, 1977a).  Jelinek and
 Corneluissen (1977) reported that  the  Food  and  Drug  Administration  (FOA) has
monitored  for  arsenic  in  its Total  Diet  Survey since  inception  of  the  pro-
 gram.  The data  from this program indicated that  the average daily  intake
for arsenic trioxide has decreased from about 130  wg/day in  1968 to about 20
ug/day in  1974.   It  is  likely  that  the differences  among  the estimates  of
total daily intake are  partially due to variations  in the species of arsenic
considered.
    Arsenic was known  as a therapeutic agent to the  ancient  Greeks  and Ro-
mans.  The introduction  of Salvarsan  (arsphenamine)  by Ehrlich  at  the  turn
                                     C-9

-------
of  the  century  gave  rise  to intense  activity on  the  part of  the  organic
chemists,  and  it is estimated that  more  than  32,000  arsenic  compounds were
synthesized  (NAS, 1977a).
    The  advent  of  penicillin disposed of  antiluetic arsenicals,  and  other
newer drugs  have nearly  eliminated  the use of  other  organic arsenicals.   In
current Human  therapeutics,  arsenicals are of  importance only  in  the treat-
ment of certain  tropical  diseases (Harvey, 1975).
Inhalation
    Suta  (1978)   has  evaluated  atmospheric  arsenic  concentration  data  for
1974  in  267  locations  representing  a  resident   population  of  more  than
58,000,000  people.   The  annual  average  concentrations for  all  sites ranged
from below  the detection limit  to  83  mg/m .  The mean of the  annual  average
concentrations for  all  locations was  3 mg/m  .   The average  concentration
for eiaht  locations near nonferrous smelters was  30 mg/m , and  the  average
concentration  for eight  locations in  remote  rural  areas was  0.4  mg/m  ,  as-
suming a concentration of  zero for  samples reported  as  below  detection lim-
it.  The  lower detection limit  for  an individual  arsenic sample  is  1  mg/m
(Suta, 1978).
    Suta (1978)  has estimated air arsenic concentrations and exposed  popula-
tion numbers associated  with  major  manmade sources of arsenic  in  the atmos-
phere (Table 4).   He states  that due to  the  paucity of relevant data  and  in-
formation,   the large number of required  assumptions,  and the  inherent  inac-
curacies of  the  modeling approach,  the accuracy of  these exposure  estimates
cannot  be   judged  quantitatively.   It should  be  noted that  the  exposure
concentrations shown  in  the  table  are annual  averages.  Exposures  for  se-
lected times may be much higher  or  lower than  the  annual  averages.  Popula-
tion exposures for  concentrations below  3 mg/nr are  not given  because they
                                     C-10

-------
                                                        TABU 4

                      Estimates of  Collation Exposures  to Arsenic  far Selected Emission Sources'
Emission Source
Average Annual
Concentration*
<«9/«5)
I.O-5.9
1.0-?. 9
0.60-0.99
0.30-0.59
0. 10-0.?9
O.t)6<>-0. 099
0.030-0.059
0.010-0.029
O.OOS-0.009
0.003-0.004
Copper
Smelters*
2.200
--
17.500
2H.OOO
92,000
2Btt,000
20,000
Mb ,000
23/.000
lead
Shelters':



JJOO
2.600
5,100
38.000
46.000
67,000
I\*c
Shelters'*



22.000
48.000
134.000
6V). 000
1.202,000
1.642.000
Cotton Pestle Me
Gin* Manufacturer'
5
100
200
700
2.000
J.OUO
5.900
20.000 60
Sb.OOO BOO
1)5,000 11.900
Gloss
Manufac tun ni)9


ISO
2«.,OOO
169.000
1. 251.000
1,534.000
6.0b<4,000
9,490.0m)
•Sroirce:  Sula. 1978
*Aver*oe QMiiillrect ion«l concentrations.  With the exception of cotton gin exposures, 24-hr worst-case
 c*n be estiMted by MMilliplyinq the <»muj*l averiqes of 12.5.  The 24-hr worst-case exposures for  cotton yini  may be
 obtained by «wltiplyinq the concentrations by 81.
leased on EPA's esti«ale of stack emissions.  Assumes 10 percent fugitive emissions.
cHased on an emission of 0.5 Ib of arsenic per each ton of lead produced.  Fugitive emissions are  estiHMieil lu be
 10 percent of stack Missions.
''Based on an emission of (.3 Ib of arsenic per ton of tine produced by pyrometal lurgical  ^Melters  and no sldik
 Missions at electrolytic shelters,  fugitive emissions assumed to be 10 percent  of the  I.) Ib/ton slack emtbiiitm
 for all spellers.
eAnnual averaqe exfHisure. assuming that ginnlitg exposures occur during 15 percent  of lite  year and  that there are  no
 exposures duftiM) the reminder of the year.
'ASSUMES that all large plant pesticide enisslons are well control led.
lAssuwes that 25 percent of pressed and blown glass is manufactured with arsenic and that  only certain «idiiufjitur-
 ers use arsenic in all of their pressed and blown glass production.


                                                               C-ll

-------
are assumed to be  equal  the  average  urban  background concentration.   Popula-
tion  exposures  are  not  given  for  concentrations  below  10  mg/rrr for  some
cocoer smelter  alternative  estimates, because  to  do so would  have  required
extrapolation of  modeling results  beyond  20  km for  the  source.  At  these
areater distances, the  accuracy  of the modeling results became increasingly
uncertain.
    It is  quite  apparent from these  qualified  estimates that  in  some  areas
of the country,  the  general  population is  exposed  to  high levels of  atmos-
oheric  arsenic  when  compared  to ambient  levels  in  uncontaminated  areas.
Klemmer, et al.  (1975)  analyzed  61  samples  of dusts collected  from  homes  in
Hawaii for arsenic content and found  that the  levels ranged  from 33  to 1,080
ug/g.  Since house dust has been  implicated as  a  significant  source  of human
oesticide burden,  it may  also be  a  significant  source  of arsenic exposure  in
some homes.
Dermal
    No information was  found concerning the levels  and/or  duration of  dermal
exoosure  to  arsenic.   Since arsenic  compounds  are  used  in  insecticides,
herbicides, fungicides,  algicides, sheepdips,  wood preservatives, and  dye-
stuffs and for the eradication of tapeworm in sheep and  cattle  (NAS,  1977a),
it seems  likely  that the dermal  route may be  a source of arsenic  exposure
for some segments of the population.
                               PHARMACOKINETICS
    Adequate understanding of  arsenic toxicology  is heavily dependent  upon
clear delineation  of differences  between various  arsenic forms  or compounds,
e.g., organic versus  inorganic  arsenic compounds or tHvalent  versus  penta-
valent  inorganic  arsenic species,  in terms  of various pharmacokinetic  as-
oects, e.g., adsorption, metabolism  (especially  in  vivo  biotransformations),
                                     C-12

-------
 distribution,  and excretion.  Important new information regarding  different-
 ial  characteristics  of various arsenic  forms  in regard to  such  aspects has
 emerged  in the  scientific  literature  during the past  5 years,  but has only
 recently begun  to be critically evaluated in regard  to  its  full  meaning and
 implications.   In addition,  newly emerging  evidence has  recently  been re-
 oorted  suggesting a  possible  essential role  for  arsenic  in  some  mammalian
 soecies,  carrying with  it  potential   implications  for analyses of  arsenic
 toxicology.   A  thorough  critical  assessment  of literature  bearing on  the
 above issues  appears  in  the  recently  prepared  EPA  Health Assessment Document
 for Arsenic (U.S. EPA, 1980b).
 Absorption
    The  major  routes  of  arsenic exposure of significance  for  general  public
 health are  inhalation and ingestion,  either via direct intake of  food and
 water or secondary to the  inhalation  of arsenic in  a  form  and  size whereby
 it undergoes  retrocilliary movement and is  eventually swallowed.   Inhalation
 is probably of  more  significance  in occupational settings, while  oral  intake
 is a more  widespread  exposure  route for the population  at  large.   Percutan-
 eous absorption of arsenic,  while  poorly studied, can occur  in man, based on
 isolated reports, but appears to be a  relatively minor route of exposure ex-
 cept under certain occupational exposure conditions.
    Confusing  the  picture of arsenical  absorption  is the  importance of the
 chemical  form  of the arsenical.   In  some studies,  this has  been  known with
more certainty  than  in  other studies  and it  is  difficult  at  times, to dis-
 cern clearly  relative uptake or  absorption characteristics  for  various  ar-
 senic forms under different exposure conditions.
    The  extent  of respiratory  absorption  of  arsenic adsorption depends  on
 chemical  species of arsenic  and the particulate  size, assuming that the air-
 borne arsenic  comoound  is in the  form of  an  aerosol.   Smaller-sized  parti
 cles (
-------
greater  subsequent absorption  likely via  the  alveolar parenchyma  than  for
larger-sized particles.   Larger  particles  tend  to  be deposited mainly in  the
upper  portion  of  the  respiratory tract, undergo  retrodliary  movement,  and
ultimately  are swallowed,  with  arsenic  absorption  then  determined by  the
characteristics  of gastrointestinal   uptake.   Precise  relative rates of  up-
take  and absorption of  airborne arsenic  compounds,  therefore, depend  upon
the  size of arsenic-associated  particles  generated  from particular emission
sources.  In the  case  of  emissions from  high-temperature combustion sources,
such as  smelters  and coal-fired  power plants,  emissions  of arsenic and  other
toxic trace metals  were found  by Natusch,  et al.   (1974) to be  mainly in  the
highly respirable size range of 
-------
    Animal data  have  also been reported on arsenic absorption via the respi -
 ratory  tract.   Bencko and Symon (1970) observed that hairless mice breathing
 a  solid  aerosol  of  fly  ash  containig  180 ug  As/nr  for several  weeks had
 increases  in  tissue arsenic values.   Since the  particle  size was determined
 to be only less  than  10 urn, part of this intake may have occurred via the GI
 tract.   Increases  in  tissue  arsenic  in two  exposure  groups  also  occurred
 when  rats were  exposed  to arsenic  trioxide  (condensation  aerosols:   1.0,
 3.7,  and  46 ug/m^) for  90 days (Rozenshtein,  1970).   Similarly,  relatively
 rapid absorption  of pentavalent arsenic was noted by Outkiewicz  (1977)  when
 rats  were exposed  intratracheally  (arsenate  solution  labeled  with  arsenic -
 74; 0.1  and 4.0 mg/kg).   Arsenic  tissue distribution  dynamics were  similar
 for the intratracheal  and a companion  intravenous  exposure study,  indicating
 that the  rate  of  arsenic  uptake intratracheally resembles  parenteral  admini-
 stration more than oral or percutaneous exposure.
    In man  and experimental animals,  factors  which govern  gastrointestinal
 absorption of  arsenic  include  both the chemical form of  the  element  and its
 physical characteristics.   It  can  be stated that soluble  arsenicals  will  be
 generally more extensively  absorbed than the insoluble forms.  On  the  other
hand,  one  should  be cautious  in extending  data for simple water  solubility
to the chemical milieu existing in  the GI tracts of various species.
    Taken  collectively,  the reports of  Coulson, et  al.  (1935),  Ray-3ettley
and O'Shea (1975), Oecelius (1977), and Mappes  (1977)  demonstrate  that  very
substantial  gastrointestinal   absorption  of   soluble   inorganic   tHvalent
 arsenic  occurs.   Greater  than   95  percent  of  inorganic  arsenic taken  orally
by man appears to be  absorbed,  with less than  5 percent of the  administered
 amount appearing  in feces  (Coulson,  et al.,  1935;  Ray-8ettley and  O'Shea,
 1975).
                                     C-15

-------
    Consistent with  this,  Mappes  (1977)  observed  that  daily intake orally of
an  aqueous  solution  of  -0.8  mg  tnvalent arsenic  led to a  daily  urinary ex-
cretion  rate  of 69  to  72 percent  of the  daily  intake by a  human  subject.
Also, Oecelius  (1977)  reported that  ingestion  of 50 ug trivalent and  13 ug
oentavalent  inorganic  arsenic  in  a  wine  sample  led  to 80  percent of  the
total  63  ug  of  arsenic  appearing   in  urine  within   61  hours.    Oecelius
(1977),  however,  reported  that ingestion  of  well  water  mainly  containing
identified  pentavalent  inorganic arsenic led  to urinary clearance  of half of
the intake  of -3 days.   Absence  of fecal   arsenic data preclude  determining
fecal loss or body retention of the remaining half.
    In contrast to the  relatively high  absorption rate  for  soluble inorganic
arsenic,   Mappes   (1977)   reported   that   insoluble   arsenic   triselenide
(AsjSe^),  when  taken orally,  passes through  the GI  tract with  negligible
absorption.
    While  available  data  for  human  GI  tract  absorption of  As£03  taken  UP
via inhalation are sparse, the  report of Smith, et  al.  (1977)  strongly sug-
gests that  levels of arsenic trioxide entering  the GI  tract of smelter  work-
ers strongly correlate with urinary arsenic levels.
    Analysis  of  arsenic  intake via  the  diet  of nonoccupationally  exposed
populations requires that one consider  the  issue  of  bioavailability  and dif-
ferences in the manner  in which arsenic  forms are incorporated into  the mat-
rix of various foodstuffs.   In  terms of concentration  levels  and  bioavaila-
bility factors,  the  arsenic content  of crustaceans  and other marine  foods
warrants special  comment.
    The  so  called  "shrimp" or  "seafood" arsenic  present in crustaceans  and
other fish  appears to represent a  complex  organic form of  the  element  which
has recently  prompted  considerable  study  (LeBlanc and  Jackson, 1973; Westoo
                                     C-16

-------
 and  Rydalv,  1972; Munr0,  1976;  Edmonds,  et al.  1977;  Penrose,  et al. 1977;
 Oecelius,  1977;  Edmonds and  Francesconi,  1977).   In brief,  the  results of
 such  studies  indicate  that  the arsenic present  in shellfish and other marine
 foods  appears  to be extensively  absorbed and  rapidly  excreted intact  as  a
 complex  organoarsenical  by man and animals and,  as  such,  does not appear to
 oose  a oarticular health  threat to  man.   Thus,  it  is  not  appropriate to con-
 sider  high  human arsenic  intake  from diets heavy with  "seafood"  arsenic as
 representing relevant  exposure  inputs  for estimating  the  likely toxicity po-
 tential  associated with  overall exposure  of population  segments to inorganic
 arsenic  via multimedia routes.
    Studies of  the  oral  intake and absorption  of  arsenicals  in experimental
 animals  generally confirm  the  findings derived  from  the above human studies.
 More  specifically,  soluble inorganic  arsenic,  in either  trivalent  or  penta-
 valent solutions,  is  almost completely  absorbed  from the  GI tract of  rats
 (Coulson, et al.  1935),  with 88  percent  absorption was  observed for arsenic
 trioxide solution (Urakabo, et al.  1975;  Dutkiewicz,  1977)  and 70  to 90 per-
 cent  for arsenate  solution.   Similar  observations  have been  made in  pigs
 (Munro,  et  al.  1974), with 90 percent  of arsenic  trioxide  solution  being
 absorbed, and monkeys  (Charfaonneau, et  al.  1978a)  with  98 percent  of arsenic
 trioxide being  absorbed.  Charbonneau,  et al.  (1978a)  fed arsenic-containing
 fish (Atlantic  grey sole)  to  adult  female monkeys  as  a  homogenate  (1 mg fish
 arsenic/kg body weight)  and  noted  that  about  90 percent  was  absorbed,  of
which about 75  percent appeared in  urine  after  24 days.   In a related  study,
 swine and adolescent monkeys were seen  to absorb  approximately 70  to 50 per-
 cent, respectively.  On the other hand, arsenic trioxide  in suspension given
 orally to rabbits  and  rats was reported  to result in  only about  40  and  30
 oercent absorption, respectively (Ariyoshi and Ikeda,  1974).
                                     C-17

-------
    The  effect  of nutritional  status  or dietary factors  on  arsenic  absorp-
tion  has not been well  studied,  although  interactive  relationships  between
arsenic  and  elements  such  as selenium are  known.   Tamura  et al.  (1977)
showed  that  rats  given arsenic trioxide  in  either  milk  or cereal  diets  had
no differences  in  fecal  excretion  patterns  of  arsenic  over a  6-month  period.
Nozaki,  et  al.  (1975),  using  ligated  rabbit  intestine,  demonstrated  that
phosphate,  casein,  and a casein hydrolysate all  inhibited  trivalent  arsenic
uotake.   Tsutsumi and co-workers  (1976)  found  that  co-administration  of
metal  chelanting  agents,  such  as  dimercaprol  (BAL),  thioctic acid (TA),  or
                                                74
diisopropylaminodichloroacetate  (DAOA),  with    As  labeled   arsenate  into
the GI  tract  of the  rat  resulted  in markedly  retarded enteric  absorption  of
the arsenical, compared to controls receiving the labeled arsenate alone.
    Little  information exists on the extent  of cutaneous  absorption into  the
bloodstream  of inorganic   arsenic  by  human  subjects.    Evidence  for  skin
absorption sufficient  to induce clinical  manifestations of  arsenic  poisoning
stems  from  case  reports  of  either  individual accidents  with  arsenic  tri-
chloride  (Delepine,  1922,1923;  Buchanan, 1962), arsenic acid  solution  (Carb
and Hine, 1977),  or  (arsenical  paste)   (Robinson, 1975).   Patty  (1948)  notes
that arsenic  passage through  epidermal  lesions is more rapid  than  with  nor-
mal skin suggesting that, in  the case  of industrial  activity,  the skin burns
elicited by arsenic contact permit easier passage of arsenic  into the deeper
layers of the integument.
    Dutkiewicz (1977)  found that skin  absorption  of  arsenic  in  the  rat using
solutions of  arsenate, was  significant  and  the uptake  rate via  the tail  was
as  high  as   33.1   wg/cm  /hour  using  concentrations up  to 0.2  molar.   The
                                                 2
corresponding  absorption  in  man,   using 700  cm   as  the  surface  area  for
                                     C-18

-------
 both  hands,  was calculated to  be  as  much as 23.2 mg/hour; and tissue  levels
 of  arsenic from dermal  contact resembled the  distribution  dynamics of oral
 exoosure.
     Potential  fetal  exposure  to  toxic  elements via  transplacental passage
 from  the mother  is  of major importance given the potential sensitivity of uj_
 uterp  development to  diletarious impacts  of  exogenous toxic agents.
     In a  study of  maternal-newborn  tissue  sets  for arsenic, Kagey,  et  al.
 (1977)  reported  that cord  blood  levels  approximate those of mothers  in  101
 subject sets.   Tissue analysis (Kadowaki, 1960)  of fetus arsenic  in  a pre-
 sumably healthy  Japanese population  indicated  measurable arsenic  levels  at
 least  by the  fourth month  of  gestation and  increasing  to  the seventh  month.
 Of  importance  here  is  the  observation that brain levels, as  well  as those of
 bone,  liver,  and skin,  were  the  highest of all  tissue tested.    Since  the
 relative amounts of arsenic passing  the blood-brain barrier  in adult animals
 appears to  be small  relative  to  uptake  in  other  soft tissues,   these  data
 sugaest that  the  human fetal  nervous system may be particularly vulnerable
 to arsenic exposure early in development.
    Complicating the  issue  is  the  chemical  nature  of  the  tissue  arsenic  as-
 sayed  in either of  the two  studies noted  above, inasmuch as  precise chemical
 speciation was not  attempted.   Also,  the Japanese   study  presumably did  not
 select material in  a  manner such that dietary  histories could be discerned.
 Thus,  Questions can be  raised  regarding  full implications of these data  for
 toxicological  analyses.
    Transplacental   transfer of  arsenic  has also been demonstrated in experi-
mental animals,  For  example,  rapid  transplacental   transfer  has  been  demon-
 strated in  hamsters  given  arsenate   parenterally   (Perm, 1977;  Hanlon  and
                                     C-19

-------
Perm,  1977),  with  embryonic tissues showing levels close  to  those in mater-
nal blood 24  hours after dosing.   Tnvalent  arsenic,  given  as  such,  also re-
sults  in  transplacental passage  in pregnant  rats.   Arsenic  has been  detected
in newborn rats when the dams received arsenic trioxide in the diet.
Distribution
    Analysis  of  the -available  literature  dealing  with the tissue  distribu-
tion of inorganic arsenic must be  tempered by  current  awareness  that  i_n_ vivo
biotransformations of  arsenic  occur in many species  and  distribution dynam
ics  involve   the  transformation products  as  well  as  any  intermediates  or
original exposure forms.
    Blood  is  the main  vehicle  for transport  of arsenicals from  absorption
sites  to  various tissues,  with  the hemokinetic character  of arsenic  being
dependent  on  the animal species  studied.    It  is  readily apparent from the
literature that  the  rat constitutes an  anomalous  model  for  studies of the
fate of  inorganic  arsenicals  vn vivo and this  includes the clearance behav-
ior of  blood-borne  arsenic  in  the  rat (Hunter,  et  al.  1942;   Ducoff, et al.
1948;  Lanz,  et  al.  1950;   Ariyoshi  and  Ikeda,  1974;  Klaassen,  1974;  Tsut
sumi and Kato, 1975; Outkiewicz, 1977).   In  the case  of the rat,  arsenic  in
blood  is  only slowly  cleared  following  exposure,  with  about 80  percent  of
the  total  blood arsenic  content  localized  in  the  erythrocyte.  The  half-
times  of  blood  clearance   for  inorganic arsenic  in  the  rat  (trivalent  or
pentavalent)   is of  the order of 60 to 90  days (Lanz, et al.  1950;  Ariyoshi
and  Ikeda,  1974).   Given the  recent data of  Odanaka,  et al.  (1978),  cited
earlier, it 1s possible that erythrocyte arsenic is present as the dimethyl -
ated form.
    Arsenic in  the  blood  of other species  — man  (Ducoff,  1948;  Mealey,  et
al.  1959),  mice  (Lanz, et  al. 1950;  Oema,  1955),  rabbit  (Hunter,  et  al.
                                     C-20

-------
     ;  Ducoff,  1948; Klaassen, 1974), dog  (Lanz,  et al. 1950; Hunter, et  al.
 1942),  and  the  primate  (Hunter,  et  al.  1942;   Klaassen,   1974)  —whether
 given  as  the pentavalent form or  as trivalent form, is  rapidly cleared.
     In  some of  these  species,  a  three-compartment model for clearance is  ap-
 parent.   Qverby  and FredMckson  (1963)  calculated half-times  of ~6 hours  for
 the  raoid phase,  a slightly longer  time for  the second  phase  and  a  slow
 ohase  of  -60 hours.
     Clearance of arsenic in  dog  and  man was also  found to  fit  a three com
 oartment  model  by  Charbonneau,  et al.  (1978b)  with half-times of  1,  5,   and
 35  hours,  respectively.    When   contrasted  with   the   work  of Tarn,  et  al.
 (1978a),  which  reported the  time dependent  j_n vivo  methylation   of  arsenic
 and  excretion, the  various  components  presumably  relate to initial excretion
 of inorganic arsenic, followed by clearance of dimethylarsenic.
    Very  little  information is available  concerning the molecular character
 of binding in either  erythrocytes or plasma, and what little older  data  is
 available must  be  viewed in  the  light of what  is presently  known  about  i_n_
 vivo changes  in  arsenical forms.
    In  the rat  erythrocyte, arsenic appears  to be associated with  the  pro
tein moiety of hemoglobin (Hunter, et al.  1942; Lanz,  et  al.  1950).  Labeled
arsem'te  (As), when given  to  a human  subject, appeared to be associated
in plasma with *1-globulin  (Musi!  and Oejmal, 1957).
    Biliary transport of arsenic  has  been reported for a  number  of species.
8ile excreted arsenic is reabsorbed.  Cikrt  and Sencko  (1974)  noted that  the
rat  had  a  higher  biliary  excretion  rate  for  trivalent  than  for   the
oentavalent form (-10:1).   Klaassen  (1974) noted  that  the  biliary excretion
rate was  much greater for  the rat  than  for  either the rabbit  or  the  dog.
Biliary transport data for  man is  not available.
                                     C-21

-------
    The  tissue  partitioning of  arsenic  in man  has been studied  using  both
autopsy  and  dosing  data.  Kadowaki  (1960)  found that heart,  kidney,  liver,
and  lung contained  the  highest  levels  of arsenic  (0.04  to  0.05 ppm,  wet
weight)  of  the  soft tissues,  with  skin,  hair,  teeth,  bone, and  nails --
arsenic  storage  organs  -- housing  the  highest  absolute  amount.  Brain  tissue
(0.03  ppm wet  weight) had  an  arsenic  level slightly  lower than  other  soft
tissue.  Uebscher  and  Smith  (1968),  analyzing  tissue  samples from  nonex-
posed  sources  in Scotland,  observed  lung  to  have  the  highest  levels  (0.08
ppm  dry  weight), with  liver  and  kidney  levels  (0.03 ppm dry weight)  not
materially different from other  soft tissue.   Like  the  Kadowaki  study,  stor-
age  organs  such  as  bone,   hair,  nails,  and  teeth  had  the highest  overall
levels.
    In addition  to  the  autopsy studies by  Kadowaki  (1960)  and Uebscher and
Smith  (1968),  Larsen, et al.  (1979)  have  recently  reported  on  a  detailed
study  of the topographical  distribution of  arsenic in  normal  human  brain
tissue.  The study  results  (for 5 persons, 15 to 81  years  of  age)  revealed
widespread distribution  of   arsenic  throughout essentially all  of 24  brain
areas  sampled, with  markedly higher  concentrations  of arsenic  in  white  mat-
ter  (2.4-5.2  ng/g wet  tissue) than  in  grey matter  (1.2-2.6  ng/g wet  tis-
sue).  These arsenic concentrations  in  centra.l  nervous  system white  matter
are not  significantly different  from arsenic  concentrations reported  earlier
for peripheral  nerves  (Larsen, et al.  1972),  leading to the  interpretation
by  Larsen,  et  al.  that the metal  is  likely preferentially  accumulated  in
neural tissue  components (e.g.,  myelin)  high in  lipids,  phospholipids,  or
phosphatides.   This   interpretation  is  consistent  with  the  proposition  by
Schroeder and Balassa  (1966)  that arsenic  has a predilection  for accumula-
tion in fat.
                                     C-22

-------
     The  above  findings  appear  to be rather consistent in regard to the over-
 all  patterns of  tissue  distribution of arsenic  based  on  human autopsy mate-
 rial.   However,  since the above studies were  carried  out with little atten-
 tion to  dietary  histories,  particularly the  predominance  of seafood in diet,
 it  is  difficult  to compare the  study  results  in absolute quantitative terms
 or  to  draw  orecise conclusions from them  regarding trends  in  tissue  accumu-
 lation  with  age.  Kadowaki's  data  for infants and elderly  subjects,  never-
 theless,  suggest some age-dependent accumulation in skin and  kidney.   Also,
 the  above studies  do  not  provide a  basis  for assessing possible differential
 tissue distribution of  tri - or pentavalent-inorganic  arsenic.   Other  studies
 indicate, however, that when human  subjects  are  exposed  to  trivalent  arsenic
 parenterally,  highest  levels   of  arsenic  are   seen  in  liver  and  kidney
 (Hunter, et al.  1942; Ducoff, et al. 1948; Mealey, et  al.  1959).
     Exposure of  various  species to  either tri- or pentavalent  arsenic  leads
 to  the  initial  accumulation of  the element  in  liver, kidney,  lung,  spleen,
 aorta,  and  skin   (Hunter,  et  al. 1942;  Ducoff,   et  al.   1948;  Lanz,  et  al.
 1950; Peoples, 1964;  AriyosM  and Ikeda,   1974;  Cikrt  and Bencko,  1974;  Kla-
 assen,  1974;  Tsutsumi  and Kato,  1975;  Urakabo,  et  al.  1975;  Dutkiewicz,
 1977).   With  the  exception  of  the rat,   a  species in  which  metabolism  of
 arsenic is only a very limited model for study of this element (vide  supra),
clearance from  soft  tissue  is rather  rapid except for  the skin, where  the
high sulfhydryl group content  probably  promotes  tight  arsenical  binding.   As
also seen with human tissue,  arsenic  is  apparently  lodged  in the brain  of
experimental   animals  exposed   to  arsenic,   with  slow   clearance  reported
 (Crema, 1955).
    The more-or-less  similar tissue distribution profiles for  both  tri- and
pentavalent  arsenic  in  various  species probably  reflects   the  common  bio-
                                     C-23

-------
transformation oathways  for  inorganic arsenic that have  been  described  ear-
lier.   It  should  be  noted,   however,  that  since  presacnfice perfusion  of
animals  in  these  studies  was not  carried  out,  part  of  the  arsenic  tissue
burdens reported may  be  attributable  to trapped blood.   This  might,  for  ex-
ample,  account  for at  least  part  of  markedly elevated  tissue levels  noted
for spleen.
Metabolism
    The understanding of  assimilation  of inorganic arsenic by man  and  other
mammalian soecies  is  substantially complicated  by a  series  of  ne«1y-
-------
    Oecelius  (1977)  reported  the urinary excretion of form variable arsenic
when  a human  subject ingested  arsenic   in  known  oxidation  state  or  other
chemical  forms.    Ingestion  of a  wine sample  of known arsenic  content  and
fori  (50  ug  trivalent and  13 ug pentavalent) was followed  in  about  61  hours
by  major  clearing of the  63  ug  of  arsenic:  50 oercent  as  dimethylarsenic
acid,  14  aercent  as  monomethyl   arsenic,  and  8  percent  each  in  the  two
inorganic forms.
    Consumotion  of well  water containing  200  ug  arsenic  as  arsenate by  a
subject  in  the  same  study showed urinary  trivalent  arsenic  at  near  back-
ground  levels with an elevation  in pentavalent  form as well as increased  ex-
cretion of  dimethylarsenic.   Determining  exact percentages of  eacn  excreted
form was  complicated  by  recovery  of  but  half of the ingested  amount.  Arsen-
ic  as  contained  in canned  crab tissue was  also studied in this experiment.
It  would  appear  that arsenic  is  present  in marine  foods  in an  organic form
which is excreted  intact, but  from which  dimethylarsenic may be liberated by
chemical treatment.
    The study of  Smith,  et  al. (1977), using basically the same speciation/
analysis  techniaues  noted  in  the  previous  study and  involving  urinary pro-
files for  a group of copper  smelter  wokers,  also confirmed  transformation
processes i_n_ vivo  (Table 5).   In  controls  as well  as  in  three  study  groups
that varied  as  to intensity  of  airborne  trivalent arsenic oxide  exposure,
dimethylarsenic was the dominant  species  in  urine,  followed by methyl  arsen-
ic,  trivalent arsenic, and  pentavalent arsenic.
    Interestingly, the correlation of  dimethylarsenic  with  airborne  exposure
composed a  tighter fit than total arsenic.   Furthermore,  both  fine resp-ir.
able and  larger  (>5 urn)  fractions  of  arsenic trioxide  particulate correlated
with  all  four  forms  measured,  with  a  stronger relationship  seen  for  t"e
                                     C-25

-------
                                                 TABLE 5

                         Concentration of Arsenic In Urine and Airborne Samples
                                        Test and Control  Groups*
Urinary
Species
AS (III)a
As (V)a
Methyl arsonic adda
Dimethylarsinic acid*
Total urinary arsenic
Arsen1cb
Control
(n-41)
1.3(1.58)
1.31(1.59)
3.4(1.63)
11.5(1.47)
21.2(2.04)
3.6(1.56)

Low As
Exposure
(n-30)
2.2(2.19)
1.6(2.32)
4.9(2.13)
17.0(1.96)
24.7(2.01)
8.3(3.43)
As concn. (SD)a
Medium As
Exposure
(n-23)
4.8(2.08)
2.4(2.86)
9.7(1.90)
32.7(1.71)
51.8(1.61)
46.1(3.05)

High As
Exposure
(n-30)
8.6(2.62)
3.1(3.64)
20.8(2.55)
64.1(2.42)
66.1(2.14)
52.7(6.61)
*Source:  Smith, et al. 1977
aAll concentatlons are expressed as yg/1 of elemental arsenic,  geometric mean (standard deviation).
^All constituent concentrations are expressed as ng/m^ geometric mean (standard deviation).   Con-
 trols had 56.1 percent of samples less than detectable (<1.2 ug As/m3)  and the low group had 20
 percent less than detectable.
                                                  C-26

-------
        portion,  i.e., that  portion  wnich mainly enters  the  body via the GI
 tract.   It may  be  seen that  while  the  relative amount  of  dimethylarsenic
 acid,  (which  may be considered a  detoxification  form)  is invariant over the
 various  exposure groups,  the relative amount of  trivalent  arsenic (particu-
 larly  in comparison  to  pentavalent  arsenic) increases  almost  linearly  with
 increasing  exposure.   This  increase  in  trivalent  arsenic with increasing ex-
 posure  to  airborne  arsenic is  further  suggestive evidence  that it  is  the
 trivalent  form of  arsenic  j£ vivo  that  mainly  accounts  for toxic  effects
 seen  in  man and  is  consistent  with  dose-response relationships  for  various
 health effects found in epidemiologic studies.
    Reports in  the  literature dealing with  the  interconversion  of trivalent
 and  pentavalent  arsenic in  man  are  sketchy.   Mealey,   et  al.   (1959)  noted
 that  administration  of   As trivalent  arsenic  parenterally  to  human  clini-
 cal  subjects  resulted in excretion  of  levels of "pentavalent"  arsenic  that
 ranged from about 60  percent  1 day post-dosing  up to  80 percent  with  further
 time.  The  method employed  involved  the acidification of urine  samples  with
 hydrochloric acid followed  by benzene extraction.  Trivalent arsenic  is  ex-
 tracted by  benzene, but the  pentavalent form remains.   Since  this separation
 approach  differentiates  tri - and  pentavalent  inorganic  arsenic   {later  bio-
 analytically confirmed by both Mushak, et  al.  1977 and  Reinke,  et  al.  1975)
as well  as methyl  arsonous from  methylarsonic  and  dimethyl arsinous  from
cacodylate  (Mushak,  et  al.  1977),  it is  probable that  the "arsenate"  frac-
tion was  a  mixture  of cacodylic,  methyl  arsenic, and  inorganic pentavalent
arsenic.  Since  the amount  of  "arsenate"  determined was  greater proportion-
 ally  than   any  contaminating level  in  the  parenteral   dose,  conversion  to
arsenate and methylated forms had occurred.
                                     '-27

-------
    One  subtle  aspect  of this  Deport,  however,  is that mono- and,  more  im-
portantly, dimethylarsenic in these urine  samples  existed  as  dimethylarsenic
(cacodylic)  rather  than  dimethylarsinous  acid.   Were  the  case  otherwise,
i.e., methylated arsenic in  lower  oxidation  state,  then  extraction  into  ben-
zene  from hydrochloric  acid  solution of  the lower-state methyl  arsenicals
would  have  occurred,  as  indicated by  the  observations  of tfushak,  et  al.
^1977).   This  does not preclude the  possibility that  sufficient  oxygenation
of urine  samples occurs  in  the process of  collection  to allow oxidation  j_n
situ, but  the  work of both  Smith,  et al. (1977)  and  Oecelius  (1977)  indi-
cates that whatever  artifactive oxidation in  urine post-collection may  oc-
cur,   at  least  measurable inorganic trivalent  arsenic  remains  and  at  levels
proportional  to exposure to the trivalent  form.
    A number of animal studies  also provide  data regarding  the character  and
Quantitative aspects of  in vivo transformation processes of  inorganic  arsen-
ic.  Of necessity, the weight placed  on these studies  is  tied  to  the quality
of the methods of  analysis and  their  ability to  chemically  speciate the  var-
ious   forms.   This  also  necessitates  retrospective scrutiny of methods  used
in the earlier  literature  since the more reliable  speciation  techniques   -e
of recent origin.
    To  date,  transformation  processes  involving   arsenic  and  experimental
animals have  been  reported for the dog (Lakso  and  Peoples,  1975;  Tarn,  et  al.
1978,1979; Charbonneau, et al.  1979), cow  (Peoples, 1964;  Lakso  and Peoples,
1975), mouse (Bencko,  et al.  1976),  and  rat  (Winkler, 1962; Ginsburg,  1965;
Odanaka,  et al. 1978).
    Lakso and  Peoples  (1975) noted that  the oral  exposure of both  dogs  and
cows  to either arsenite or arsenate led to  conversion  of  either valency  form
                                     C-28

-------
 to  methylated arsenic  with  about 50  percent  conversion  to  methyl  arsenic,
 which may  formally be considered pentavalent.  The method  used  did  not per
 nit  distinction between valency forms of inorganic arsenic.
     Several more  recent studies  Have  provided  more  detailed data as  to arse-
 nic  transformation processes  in the dog.  Tarn, et  al.  (1978,1979) exposed a
 arouD  of  docs  to  radiolabeled  f  As)  arsenic   acid  given  intravenously.
 The  levels  of inorganic,  monomethyl -,  and  dimethylarsenic  were  then  timemon
 itored  in  urine and plasma using an  ion-exchange chromatographic tecnnigue.
 While inorganic  arsenic is the major species  in  plasma  up to about  2  hours
 post-dosing,  dimethyl  arsenic  formation can  be detected as early as  10 min-
 utes  affc?r administration.  By  6 hours,  virtually  all  (90  percent)  plasma
 arsenic  is  in the  dimethyl  form, with little monomethyl   species  detected.
 Oimethylarsem'c was the major form in the  urine from  days  1-6.   In a  compan-
 ion  study,  Tarn,  et  al.  (1978b)  used  thin-layer chromatography  to  further
 speciate the  inorganic  arsenic  fraction into both pentavalent  and trivalent
 arsenic.  Charbonneau, et  al.  (1978a,b)  noted  that when  labeled  arsenic acid
 (As)  was  given  to  dogs intravenously,  about  four-fifths  of  the  arsenic
 lodged in the  red  cells,  with  dimethylarsenic being  detected in  those  cells
 about 10 minutes after  dosing.  With  time, the arsenic  content  is partition-
ed between  cells   and  plasma,  total  conversion being seen by  6  hours.   At
about 1  hour,  dimethylarsenic is  detected  in  the  urine.  These  data  indicate
participation  of  the  erythrocytes and  liver  in   dimethylation and  transport
of the dimethylated species.
    Of interest, here are  the data of  Odanaka,  et al. (1978), who fed ferric
methanearsonate to adult male  rats  and  analyzed  the  blood, urine, and  feces
for  the  amount of various speciated  arsenical forms.  Dimethyl  arsenic  was
detected in  urine, feces, and  blood,  indicating methylation of  monomethyl
                                     C-29

-------
arsenic  j_n  vivo.  While  dimethylarsenic was  in  minor amounts  in  urine and
feces,  blood  arsenic was mainly  present as dimethylarsenic.   For  analysis,
these  workers  used  thin-layer  chromatography  for separation of  the  arseni
cals  and  gas-1iauid  chromatography in tandem with mass  spectrometry  to con
clusively determine  the structure  of the organoarsenicals  as  mono-  and di-
methylarsenic.
    The  Odanaka,  et  al. (1978)  oaper is of significance  on several  counts:
(1) it  demonstrates  that methyl  arsenic  can  be  methylated  to dimethylarsenic
and,  hence,  the monomethyl  form  can  be  an  intermediate in  the inorganicdi
methylarsenic transformation  (since these authors took  pains  to assure the
purity  of the methyl arsenic administered,  it  is  ^niikely that the dimethyl
form  arose from  contaminating  inorganic  arsenic);  (2)  mass spectral  analysis
confirms  the  presence of dimethylarsenic  in arsenic  transformations  in the
rat and,  by  inference,  other animal  models  reported;  (3)  the  minor  amounts
of dimethylarsenic in urine  or feces and  the major  amounts  in  blood  suggest
selective retention  of dimethylarsenic by  rat erythrocytes.   When contrasted
with  the  Charbonneau,  et  al.  (1978a,b)  data  noted  previously, it  appears
that  in  both  rat and dog the  erythrocyte  is at least the  transport  vehicle
for dimethylarsenic,  but,  unlike  the dog erythrocyte, release  of  dimethyl -
arsenic into rat plasma is much  slower.   This  is  consistent  with other known
facts  of arsenic distribution in the rat  as noted below.
    In  light of  the  preceding reports regarding methylation  processes, ear-
lier  reports dealing with  trivalent-to-pentavalent-arsenic conversion or the
reverse must be viewed carefully.
    _In  vivo  conversion  of   trivalent inorganic  arsenic  to  the  pentavalent
form  has  been reported  by several  authors.  Infusion of  arsenite (trivalent)
intravenously in  dogs led  to  the  detection of tri - and pentavalent  arsenic
                                     C-30

-------
 in  olasma,  urine,  and glomerular  filtrate  (Ginsberg,  1965).   tinkler  fi962!
 noted  that  livers  of rats fed arsenite contained mainly  arsenate.   *^e "lore
 Decent  study of Bercko,  et  a".  (1976)  is  particularly  significant  in that
 fivalent arsenic  conversion to  the  pentavalent  form  in  mice  was demonstrat-
 ed  using  oaper  chromatographic  techniaues  for separation  and removal  of
 urine  samples  through  the bladder to minimize artifactive  oxidation  of tri-
 valent  arsenic.  It  was  noted  that the  relative  amounts  of pentavalent arse-
 nic  formed   from  radio-isotopic  (   As)  arsenite hinged  on  the time  lapse
 after  pretreatment with  a  large  dietary level (250 mg/1)  in  drinking water.
 In animals  preexposed  for  13 days  prior  to  dosing with the labeled  arsenate,
 virtually no trivalent  arsenic  was  found  in  urine.   Since  dimethylarsenic
 acid was  not tested  in Bencko's chromatographic system,  it  is  possible that
 this was the form being identified as pentavalent inorganic arsenic.
    The case  for  HI vivo  reduction  of  pentavalent  arsenic to  the  trivalent
 fonn is  sketchier,  mainly due  to  analytical methods  employed.   The  aporoach
 of Lanz,  et  al.  (1950),  who  reported some  reductive  conversion of  arsenate,
 entailed  precipitation  as a  mixed  salt,  the residual  solubility  of  which
 could  have  been  enough to account  for  the  amount  labeled  as  soluble triva-
 lent arsenic  (MAS,  1977a).  The Ginsberg  (1965) report  employed an  extrac-
tion/chelation separation  method  involving  acidified  samples and removal  of
trivalent arsenic  wfth  chloroform containing ethyl  xanthate.   Since  ethyl
xanthate is  a  thioli'c  chelanting  agent  and  pentavalent arsenic  is very lab-
 ile in acid,  interaction  of  the  chelatinq agent with  pentavalent arsenic  to
form arsenic  (III)  and some disulfide (R-S-S-R)  cannot be  discounted.   Using
doos dosed with arsenate,  the  origin of the  trivalent portion  of the  inor-
 ganic  arsenic  fraction as isolated  and measured,  i.e.,  _i_n vitro versus  jjn
 vivo formation, cannot be definitely established  (Tarn, 1978).
                                     C-31

-------
    As can  best  be oresently determined, demethylation  of  methylated arsen-
ics formed jji vivo does not occur.  Support for this  is  chiefly  from data of
Stevens,   et  al.   (1977), who  saw no evidence  of  j_n_  vivo demethylation  when
animal were exposed to dimethylarsinic acid (cacodylic acid).
    Several  experimental  animal  studies  suggest  that  some  induction of  a
tolerance to arsenicals may arise in arsenic pretreated  animals  that are re-
exoosed.    Bencfco  and  co-workers  (Bencko  and  Symon,  1969,  1970; Sencko,  et
al. 1976)  have found  that  arsenic  pretreatment of  mice will markedly alter
the subseauent  tissue  distribution  and excretion  data  of  a  radio-isotopic
arsenic pulse.   The mechanism of this  process is  not understood,,  but  must
include the efficiency of the i_n_  vivo  methylation process(es)  described  ear-
lier.
    In summary;
    1.   Pentavalerit and trivalent  arsenic in  both man and animals undergo
         jm vivo  transformation  mainly to  dimethylarsinic  acid, which  pro-
         bably was  misidentified  as pentavalent  inorganic  arsenic  in early
         studies.
    2.   The i_n  vivo conversion of pentavalent  inorganic arsenic  to  the  tri -
         valent  forms  remains to be conclusively demonstrated, but cannot De
         ruled out based on  presently available information.
    3.   Methylation of  inorganic  arsenic can  be  considered as  detoxifica-
         tion in  that  cacodylic acid  is  much  less  toxic than the  inorganic
         forms,
    4,   As  a  detoxification process,  methylation  efficiency appears  con-
         stant as  a fraction  of total  speciable arsenic, although  the rela-
         tive amount of  trivalent arsenic will increase with  increasing  ex-
         posure.
                                     C-32

-------
     5.    However,  as  a  detoxification  process,  it can  eventually be  over-
          whelmed  or  Codified  as  is apparent from the extensive  literature on
          arsenic  toxicology in man  and  animals.
     5.    At  least  in  some species,  dimethylarsenic formation  involves the
          erythrocyte and the  liver  in biosynthesis  and transport.
 Excretion
     3enal  clearance  appears  to  be  the  major route  of  excretion of absorbed
 arsenic  in man  and  animals,  biliary  transport  of the  element  leading  to
 enteric reabsorption with  little carriage  in feces.
     In a  study assessing the  utility of urine  arsenic  measurement in occupa-
 tional  exposure   settings,  Nappes  (1977)   reported  excretion  data  for both
 single and multiple daily  dosing  for a  human subject ingesting arsenite sol-
 ution.  By 3  hours, renal excretion was  maximal,  with  about  one-auarter of
 the  single dose appearing  in  the  urine  by  day  1 post-exposure.  With succes-
 sive  arsenite  ingestion (0.8  rng As),  daily urinary clearance  after  5 days
was  two-thirds of daily intake.
     Crecelius  (1977)  noted that  arsenic  in wine  [50  ug  As (III),  13  ug As
 (V)l  after ingestion  led  to  a measured  level  in urine of  -80 percent  after
61  hours.   Oral   ingestion of arsenic  (V)  in  well water  (200 ug),  however,
led  to about  50 percent urinary  excretion  by 3  days after ingestion.  Mealey
(1959) measured urine  arsenic in patients  given  trivalent  arsenic by intra-
venous administration,  with -60  percent of the dose amount  appearing  in the
urine by 24 hours.  Hunter, et al.  (1942)  noted considerable variance,  30 to
80 percent after 4 to  5 days,  in urinary  clearance  of arsenic given parent-
erally in  a group of human  subjects.
    As might  be  predicted from  the j£ vivo behavior  of arsenicals  in  the
rat,  urinary  excretion of  arsenic  in   this  species  is  very  slow,  due  to
                                     C-33

-------
erythrocyte  retention  on the order  of  2  to 5 percent of  the  arsenic  intake
by  several  days  post-dosing  (Coulson,  et  al.  1935;  Ariyoshi  and  Ikeda,
1974).  Urakabo, et  al.  (1975)  calculated  a  half-time of  84 days  for arsenic
in the rat.
    Slow clearance of  arsenic  from the rat gave rise to  the widely  held  as-
sumption for many years  that arsenic is one of the elements that  accumulate
in the  body.   Other  species excrete arsenic rapidly.  Mice, rabbits,  swine,
dogs, and monkeys clear  the majority of injected trivalent  arsenic within  24
hours, with excretion  usually being  ^70 percent within that  time  period  (Ou-
coff, et al. 1948; Oema, 1955; Munro, et  al.  1974; Lakso  and  Peoples, 1975;
Tarn, et al.  1978a;  Charbonneau,  et al. 1978b).  Other studies  also  indicate
rapid urinary clearance  of  arsenic  given  in the pentavalent form to  species
other than  the  rat  (OuPont,  et al.  1942;  Ginsberg and Lotspeich, 1963;  Peo-
ples, 1964; Lakso and  Peoples,  1975).   Some calculated half-times for  either
tri- or pentavalent  arsenic urinary clearance are:   mice,  injected  trival-
ent, -4.5 hours; dogs and cows, oral tri- or pentavalent,  -36 hours.
    Deposition of arsenic in such organs  as hair and  nails  can be considered
an excretory mechanism for  arsenic.  Although hair  analysis has  had  a  long
history in  arsenic's chemical  and  forensic  literature,  for reasons of  both
analytical  convenience and the possibility of establishing  an  exposure  his-
tory from  sectional  analysis,  many  questions  remain unanswered.   The rela-
tionhsip between arsenic  deposition  in  hair and various  exposure  parameters
has not been well  defined on a Quantitative basis nor are  the physiological
mechanisms   well  understood.  The  chemical nature  of hair  arsenic  is  also
largely unknown.
    The long-held view of arsenic as an element that  accumulates  in  the  body
was mainly  based on  the behavior  of  arsenic  in  the rat,  an animal model
                                     C-34

-------
 which  in retrospect was  the  least helpful  in  understanding  the  fate of the
 toxicant i_n  vivo  for other mammalian  species and man.
     Based  on  current   arsenic  elimination  data  for  all  mammalian  species
 studied  other  than  the  rat  (vide supra),  one concludes that marked long-term
 accumulation  of  arsenic  generally does  not occur   in  physiologically  vital
 components of  the body.   This  is in contrast to, say,  marked long-term lead
 accumulation  in  bone  or cadmium accumulation in renal  cortex.  Autopsy tis-
 sue  data for human subjects  of different  ages  is   not  conclusive  regarding
 possible long-term  tissue accumulation.  Kadowaki   (1960)  did observe higher
mean levels of arsenic in skin  and  kidney samples of subjects about  50  years
of age  versus  infant  values,  but dietary histories  of  the subjects  were not
available to allow for differentiation  of increases  in  arsenic  levels due  to
current  versus past exposures for  the  older subjects.   Deposition  in  hair  is
really excretory  in nature,  not accumulative.
    Brune, et al. (1980)  have reported  that lung tissue from  retired smelter
workers, on  autopsy, had  median values for-  arsenic  which  were  approximately
8 times  higher than that  for a  control  group.   Kidney and liver values, how-
ever, were  not  significantly  different  between  smelter  worker  groups and
controls.  Arsenic  accumulation in the lung of smelter workers  even  after
several  years of  retirement and  removal from workplace  exposure (interval  of
2-19  years)  suggests that a very insoluble  form of  arsenic exists  in  smelter
ambient  air  and is inhaled by these workers.  That  this form may  be  arsenic
sulfide   is further suggested by the study of Smith, et  al. (1976) who  *ound
that  the  respirable air within  the confines of a  copper  smelter contained
arsenic  sulfide.  These two studies have  implications  for  the issue  of  occu-
pational  respiratory carcinogenesis associated  with  arsenic exposure.
                                     C-35

-------
                                    EFFECTS
Acute. Subacute, and Chronic Toxicity
    The multiplicity of  organ  systems  and  tissues  affected  in the manifesta-
tion of clinical  symptoms of acute poisoning  and  in the production  of  sys-
temic  health  effects  associated  with  subacute  or  chronic  exposure to  the
metal  reflect well  the widespread impact of arsenic  in  certain  subcellular/
biochemical processes  common to cellular components  of  many  different  types
of tissues.  At  the  same time, certain distinctive  features  of  arsenic  sys-
temic  toxicity,  e.g.,  its marked  effects on the skin,  are  better understood
in light  of  the intercession of  the  metal  into particular biochemical  pro-
cesses  most  intensely  characteristic  of   selected  cells  or tissue  types,
e.g.,  those comprising the  integumentary system.   The possibility of arsenic
playing a  beneficial  role,  at  very  low trace  levels, as hinted  at  by  newly
emerging  evidence  for its  essentiality 1n some mammalian  species,  is  also
best evaluated within  the  context  of  a discussion  of the impact  of the  metal
in subcellular/biochemical mechanisms.
    The following  discussion  focuses  on those  data dealing with  inorganic
arsenic interactions at  the  biochemical  and subcellular  level which have the
most relevance  for understanding  the  i_n  vivo  toxic effects of  the  agent  in
man and experimental animals.
    Several  reviews  (Peters,  1955;   Vallee,  et al.  1960;  Johnstone,  1963;
Webb,  1966) have described  the effects of  various arsenicals on enzymes and
enzyme-mediated  processes in  a number  of  species.    Many  of these  studies
have entailed  either purified  preparations, where question of  relevance  to
J_n vivo  conditions can  be raised, or heterogeneous, complex systems  where
the site of interaction(s) is left in doubt.
                                     C-36

-------
    The  literature  dealing with effects  of trivalent arsenic on  enzymes is
 rather  extensive,  while  that  for  pentavalent  arsenic  is  considerably more
 soarse.   This  stems in  large  measure from the widely accepted  fact  that it
 is  the  trivalent form  which can  chemically  interfere  directly with  enzyme
 action  via  formation of  arsenic-sulfur  bonds  with those thiol  groups  which
 oarticipate ^n either enzyme structure or function.
    Webb  (1966)  has  tabulated  no less than  78  enzymes from a wide  variety of
 soecies which have been  reported to be  inhibited  to  a varying  degree  by tri-
 valent  arsenic  (arsenite) at concentrations of 0.01  to  >10  millimolar.   Al-
 though  various  classes  of enzymes are  sensitive  to  arsenite, the  oxidizing
 enzyme  systems  appear  to be  particularly  vulnerable,  including:   pyruvate
 oxidase,  0-amino  acid  oxidase, 2-glutamic  acid oxidase, monoamine  oxidase,
 liver choline oxidase,  and glucose  oxidase.
    Evidence in  support  of thiol  binding as  the  biochemical site of  enzyme
 inhibition includes:  (1)  all  of  the oxidase systems noted  above can  be re-
 activated with glutathione,  a thiolic biochemical  factor  (Sarron and  Singer,
 1943); (2) lipoic acid  is  a  cofactor  in  a number  of these enzyme systems and
 possesses proximal  thiol  groups  expected  to  react  readily with  trivalent
 arsenic  to  form a  highly stable  five-membered heterocycle  (Vallee,  et  al.
 1960).  Such effects not only provide good  clues  as  to the  biochemical  basis
 for arsenic toxicity but also support  the  premise that  arsenicals are  gen-
eral metabolic poisons.
    One  particularly important oxidizing  enzyme systems  sensitive to  arsenic
 is  the  pyruvate  dehydrogenase  (POH)   complex  which plays  a  crucial  role  in
 cellular  energetics.  The pyruvate dehydrogenase  complex consists of  three
distinct  enzymes:  (1)  pyruvate dehydrogenase  (pyruvate   decarboxylase),  (2)
                                     C-37

-------
dehydrolipoate  transacetylase,  and (3) dihydrolipoated  hydrogenase.   Arsen-
ite could interfere with  the  latter two enzymes via binding  to  the  proximal
thiol groups of  lipoic  acid,  while  any effects  on  pynjvate decarboxylase are
likely to be associated with  the  inactivation/activation  reaction  controlled
by a phosphorylation/dephosphorylation process (Linn,  et  al.  1969).
    Decently, Schiller, et  al.  (1977) studied the pynjvate oxidation  system
using liver mitochondria  from rats  fed pentavalent arsenic (up  to 85  ppm  As
in drinking water)  in  order to pinpoint the  site  of  arsenic  interaction  in
the enzyme complex.  Pynjvate dehydrogenase (enzyme 1 of  the  complex)  activ-
ity was measured  before and  after  activation |£ vitro.   Basal  activity  be-
fore activation  was  reduced by 48  percent at 3 weeks in the  animals  fed  85
ppm As.   The  inhibition of pymvate  dehydrogenase activity both before  and
after activation suggests a direct  effect on  pyruvate utilization  not  invol-
ving lipoic acid.
    Since the  activation/deactivation  process  for POH  requires that  phos-
phate bind at some point  to both  POH  and  the phosphatase  and  kinase enzymes
involved in  activation/deactivation,  arsenate presumably  interferes by  com-
peting with  inorganic  phosphate.   Thus, in this particular system, effects
are imparted by both pentavalent  (Schiller, et  al.  1977)  and  trivalent arse-
nic (Webb,  1966).   Inhibition of the POH system  by  arsenic  influences  the
operation of the tricarboxylic acid cycle, with  decreased  acetyl-Co A  forma-
tion  and  subsequent decrease in  NADH generated  for ATP formation.   Fatty
acid  synthesis  and storage triglycerides are  also affected.
    Inorganic arsenic in the  trivalent  form has  also  been  known  to  interfere
with  active transport  processes and this  literature has been  critically  re
viewed by Webb   (1966).  Substance transport that is inhibited  includes:   po
tassium,   sodium,  rubidium,  hydrogen  Ion,  halide,  monohydrogen  phosphate,
                                     C-38

-------
 water,  proplonate,  glucose,  certain amino acids, marker dyes, and streptomy-
 cin.   According to  Webb (1966), it  is  difficult  to ascertain whether arsen-
 ate  has  a  specific effect  on  transport or  whether  the process  reflects  a
 general  lesioning of cellular  respiration  by the  arsenical.   Some  evidence
 suggests  that the chief mechanism  of  transport  inhibition  involves  pyruvate
 oxidation inhibition (Davenport, 1955).
    Arsenite  is  known  to  be  a potent  inhibitor  of chicken  liver  xanthine
 dehydrogenase  and  related  molybdoflavoproteins   (Rajagopalan  and  Handler,
 1964,  1967;  Peters  and  Sanadi, 1961;  Palmer,  1962)  and probably interacts
 with the  molybdenum  center in these enzymes  (Coughlan,  et  al.  1969).   John-
 son and Rajagopalan  (1978),  using  electron  paramagnetic resonance (EPR)  sig-
 nal modification  from  molybdenum,  found the  site  of  arsenite interaction  to
 be  a   reactive  group within the  molybdenum  complex  required  for  electron
 transfer  from  purine substrate to  the enzyme and is probably a  sulfhydryl
 unit binding the  metal  atom,  a  persulfide residue, or possibly  the metal  it-
 self.
    Although  inhibition  of enzymes due  to  arsenicals has been more  heavily
 studied,  enzyme  activation  by  arsenicals  is  also  known  to  occur  (Webb,
 1966).   This  includes  activation  of cytochrome  oxidase  of  rat brain at  an
 arsenate concentration  of  1.0 millimolar,  catalase malate  dehydrogenase  of
pig heart at  30 milUmolar  and,  apparently,  enzyme  systems  associated with
drug detoxification.  For example,  Ribeiro  (1971)  noted  that  trivalent  arse-
nic  oxide   reduced   hexobarbitone   anesthesia   time   in   mice   although
hexobarbitone oxidation  and  amlnopyrene demethylation rates  were  unaltered.
Also,  Wagstaff  (1978),  studying the effects  of  dietary arsenic trioxide  on
hexobarbitone sleeping  time, oxidation  cleavage  of  0-ethyl-0-p-nitrophenyl
                                     C-39

-------
phenylphosphonothioate  (EPN),  and  0-domethylation of p-nitroanisole  in  rats
at 100 to 5,000 ppm As, found  moderate  enzyme  induction  by trivalent  arsenic
but phenobarbital  induction of the enzyme system was  unaffected.
    Unlike arsenite,  pentavalent  arsenate appears to  exert biochemical  ef-
fects  via  interference with  phosphate  transport  and  phosphorylation  (NAS,
1977a; Fowler,  et  al.  1977),  through  uncoupling  of  mitochondria!  oxidative
phosphorylation, presumably  via competitive  replacement  of inorganic  phos-
phate  by  arsenate  to form  a  highly labile  arsenate ester that  decomposes.
Also,  arsenate  stimulates  succinate-controlled respiration of rat  liver  mi-
tochondria, an  effect retarded by  addition  of phosphate  (Crane  and  Lipman,
1953); and mitochondrial ATPase  is  stimulated by arsenate  (Azzone  and  Erns-
ter,  1961; Wadkins,  1961), the stimulation  being offset by inorganic  phos-
phate.  Arsenate inhibition  of mitochondrial respiration may occur via  com-
petition with  phosphate during oxidative  phosphorylation  and/or  inhibition
of NAD reduction by succinate  (Mitchell,  et  al.  1971).   Rats  chronically ex-
posed to arsenate show  decreased state  3 respiration and  respiratory  control
ratios in renal and liver  mitochondria  (Brown,  et al.  1976),  associated  with
swelling of the organelle in both  organs.
    From the  above,   it can  be seen that  the mitochondrion is one cellular
organelle particularly  vulnerable to the effects  of  inorganic  arsenic either
as arsenite or  arsenate (Webb, 1966; Fowler,  1977a; NAS,  1977a).   Mitochon-
dria readily take up  arsenic  and various i_n_ vivo and i_n_  vitro  studies  indi-
cate   that   biochemical    lesioning   includes   NAD-coupled   mitochondrial
respiration,  uncoupled  oxidative phosphorylation  and interference  with  steps
in  the  heme   blosynthetic  pathway which   are   intramitochondrial  (Fowler,
19775).

-------
    Arsenate  causes  mitochondria!  swelling both j_n  vitro  (Packer,  1961; De-
 Master  and Mitchell,  1970;  Mitchell,  et  al.  1971)  and in vivo (Fowler, 1974,
 1975;  Brown,  et  al.  1976).  Effects  on  liver mitochondria  after  prolonged
 oral  exposure of rats  to  arsenate  (20,  40,  85  ppm in  drinking  water)  were
 found  by  Fowler et al.  (1977)  to include  pronounced  UT_ situ mitochrondrial
 swelling  in the  40  and  85  ppm^As exposure group animals, as  well  as  lipidic
 vacuolization  and  fibrosis.  These structural changes  were  associated with:
 (1) decreased  state  3 respiration and respiratory  control ratios  for  pyruv-
 ate/malate but  not  succinate; and  (2)  marked increase  in specific  activity
 of  monoamine  oxidase and  cytochrome  oxidase, sited in  inner  mitochondria!
 membranes.  Effects  of  arsenate on  these  membrane marker enzymes  suggests
 direct  interaction  with membranes,  resulting in  increased   permeability  or
 conformational change.   The mechanism of these effects  is probably  arsenate
 interference in  phosphorylation  processes required  for  functioning of  pyruv-
 ate  dehydrogenase,   the first   enzyme  in  the  pyruvate  oxidation   complex
 (Schiller, et  al. 1977).
    In  a  related study, Woods and  Fowler (1977)  saw a  pronounced effect  of
 oral arsenate  administration  (1.2,  2.2,  and  3.5  mg As/kg;  6 weeks)   on  rat
mitochondria!  heme biosynthesis,  with heme synthetase  activity  decreased  to
63  percent of control levels at  the  highest  exposure  level, 3.5 mg  As/kg,
and a  resulting porphyrin urea.
    Incubation of  respiratory rat  liver mitochondria  with  arsenate  for  20
minutes at 2*C followed by  removal of  the agent  results in uncoupled  respir-
 ation with succinate  (Bhuvaneswaran, et al.  1972).   Since most  of the  arsen-
 ate in  this study was  removed prior to oxidative  phosphorylation  assay,  un-
 coupling of succinate oxidation  is not an  arsenolytic process.   Interesting-
 ly, the mitochondria! preparation was  capable  of  limited glutamate/malate  or
                                     C-41

-------
3-hydroxybutyrate oxidation  (AOP/0 values of 1.3 to  1.6).   Further study of
this system  (Bhuvaneswaran  and  Wadkins,  1978)  indicates  that arsenate treat-
ment preferentially  decreases  the  coupling  capacity  of  mitochondria at sites
2  and  3.   In  related work,  Bhuvaneswaran  and Wadkins  (1977)  found  that  a
small  fraction  of  arsenate added to the  mitochondrial  preparation  could  not
be  amoved  even with trichloroacetic acid treatment.  Since  arsenic binding
does not occur  with  cyanide,  oligomycin,  or  inorganic phosphate,  the binding
is  associated  with  an  electron  transport  chain  and energy-coupling  reac-
tions.   Dissociation  of the complex  could be  achieved after  partial restora-
tion of oxidative coupling at sites 2 and 3, i.e., ATP addition.
    Harris  and  Achenjang  (1977) found that  uptake  of arsenite by  rat liver
mitochondria to be energy-dependent  and  inhibited by  mersalyl or  N-ethymale-
imide.    Two modes of uptake were  kinetically discernible  and  may  involve
both membrane thiol  attachment  and  accumulation of free  or bound  arsenite in
matrix  space.
    Fowler et  al.  (1978)  studied microsomal and  mitochondrial oxidative  in-
teractions  in  preparations from  livers  of rats  exposed orally  to arsenate
(40  ppm  in  drinking water,  for  6  weeks).  Morphometric  studies   showed  a
doubling  of  the ratio between  rough endoplasmic  reticulum  surface  density
and mitochondria! volume density in  the  arsenic  treated  animals.   Microsomes
from arsenic-treated  animals contained 20 percent less  aminopyrine  demethy 1-
ase activity compared to  controls  after  mixing  with mitochondria  from  con-
trol animal  livers.   These  data point to  an  j_n_ vivo functional  interaction
between mitochondria  and microsomes  with  regard  to oxidative processes,  with
arsenate  disturbing mitochondria! NAD-linked oxidative capability and  reduc-
ing microsomal  mixed-function oxidative  capability.

-------
     Tn  addition  to characterization of mitochondria! effects, numerous stud-
 ies  have focused  on  the interaction  of  arsenic with  ONA as it  relates to
 chromosomal  effects.   Knowledge that  arsenate  can  compete with  phosphate in
 phosphorylation  processes,  as noted  earlier,  has  prompted  suggestions  that
 arsenate occasions chromosomal  abnormalities  by substituting far  phosphate
 in  the   DNA chain  (Petres  and Hundelker,  1968;  Petres,  at al=  1970),   This
 hypothesis  ignores the  fact  that  arsenate esters  are  so much more  labile
 thermodynamicany  than  the  phosphorus analogs  that it  is  questionable  if
 such esters have other than transitory existence.
    More likely is  interference  with DNA  repair  processes.   Jung  (1969.
 1971)  demonstrated decreased  ONA  repair  following  ultraviolet  irradiation
 and  incubation  of skin  grafts  in  arsenate  solution, concluding that  "dark
 repair"  of  DNA  in  these cells is  inhibited.   Results of studies by Rossman.
 et al. (1977) on effects of UV light  and arsenite on strains  of £._  coll  dif-
 fering  in  DNA  repair  functions  further  implicate   arsenite  as  interferring
with ONA repair processes.  Observations by  Grunicke. et al.  (1973) that  DNA
 removal   from tumor cells is retarded  by either arsenate or arsenite suggests
 that cross-linking of ONA and protein may be occurring.
    In regard to the possible role  of arsenic  as an essential element  at  low
 trace levels, early reports attempting to show  a nutritional  requirement  for
 the  element in  animals were inconclusive  (MAS.  1977a;  Underwood.  1977).
Part of  the problem was undoubtedly  technical  In  nature, i.e..  the  diffi-
culty of carrying  out  such  studies   in  an experimental  environment where
rigorous exclusion of a ubiquitous element from  the diet 1s  necessary.  More
 recently,  however,  several  carefully controlled  studies  appear  to  demon-
 strate nutritional essentiality for arsenic in some mammalian species.   For
example.  Nielsen,  et  al. (1978)  observed  that feeding  of arsenic-deficient
                                     C-43

-------
diets  to  pregnant rats  resulted in  greater  perinatal  mortality  among  pups
from  arsenic-deprived dams  and  post weaning  deficits  in  growth,  enlarged
spleens, and increased red cell osmotic fragility.
    Anke, et  al.  (1978) also  studied nutritional reouirements  for  arsenic,
using  goats  and mini-pigs  and a  semi-synthetic diet  containing  less  than  50
ppb arsenic.   Effects attributed to  arsenic  deficiency in both  species  oc-
curred  in  adult  animals and  their offspring.  Arsenic  deficiency increased
the mortality of  adult goats  as  well  as altering their mineral  profiles  for
copper  and  manganese.  Significant reproductive  effects for  both arsenicde-
ficient goats  and mini-pigs  included  reduction  of  normal  birth  percentages
and litter  sizes; and the  mortality  of kids and piglets from  the As-defic-
ient groups was significantly increased.   Manganese  levels were  elevated  in
As-deficient kids  and piglets, but no  perturbation of  hematological  indices
(hemoglobin, hetnatocrit, or mean corpuscular concentration)  was  noted.
    This  is  in contrast  to  observations  with rat  (Nielson, et  al.  1974),
where  decreased hematocrits,  elevated iron content  in spleen,  and increased
osmotic fragility of  cells  were seen.  Given that the rat is known to be  an
anomalous animal  model for  arsenic metabolism (see Metabolism  section)  how-
ever,   these  differences  are probably  peculiar  to that species.   Other  evi-
dence for the likely  essentiality of  arsenic in the  rat, includes the find-
ings of  Schwartz  (1977), who noted  enhanced  growth  effects of  arsenite  on
rats fed an  arsenic-supplemented diet.  An optimal  effect  was  seen  at  0.25
to 0.5 pom,  but pentavalent arsenic as sodium arsenate was  less  effective.
    Despite  the  above evidence  for  possible  arsenic  essentiality  in  some
mammalian species, any physiological  role  for arsenic, the existence of  any
specific carrier agent in the  body, or other  evidence of arsenic essentially
                                     C-44

-------
 in  man remains to be  independently  demonstrated.   Another factor complicat-
 ing  the  issue  is  the fact that one usual  feature  of  essential  element meta-
 bolism is  homeostatic  control  of  levels and movement  of a particular element
 j_n  vivo.   Based on  information  considered earlier,  there  appears to  be  no
 effective  absorption barrier for most  soluble  inorganic  arsenicals,  but ef-
 ficient  excretory  mechanisms (kidney,  hair) and biotransformation  appear  to
 provide  possible  control  over  an  absorption-excretion balance.   The Question
 of  arsenic essentially  in  man  is  made  even more interesting by  the  study  of
 Liebscher and  Smith  (1968)  (see Metabolism section),  who  showed  that  arsenic
 in  human  tissue appears  in a  log-normal  distribution, a commonly  observed
 biostatistical  characteristic  of  environmental  contaminants rather  than es-
 sential  elements.   Put in  terms  of physiology,  this  says  that  contaminant
 levels occur in tissues in  simple proportion  to the level  of exposure, i.e.,
 not  under  homeostatic  control.    However,  such a biostatistical  criterion
 added  to those of  Mertz (1970) is complicated  in  the  case of arsenic  if one
 does not know  the specific  partitioning of various  chemical  forms both  uj.
 vivo and in the human diet.
    The  more  physiologically   subjective   issue   of  arsenic beneficiality,
 particularly to  man, merits  some comment because the distinction  between
 beneficiality  and  essentiality  is not  always  made.   Given the  historical
 toxicological  character of arsenic  in  man and  animals,  beneficial  effects
from the past  (or present)  use of such "therapeutics" as Fowler's  solution
 (arsenite base),  and arsenic  pastes  have   not  always  been  considered  ;n  a
 framework of benefit-risk  balance.  The  beneficiality  of agents  such  as  Fow-
 ler's  solution  has  required that  the margin  of risk  perhaps  be  too  narrow
between  levels associated  with  both  beneficial   dose-effect  and  toxic'ty
                                     C-45

-------
dose-effect resoonses.  By contrast,  the  limited  data  on  arsenic  essentially
suggest a requirement for only  very small trace amounts,  leaving  a  huge  gap
between essentiality and toxicity.
    The Question of  possible  beneficial effects of  arsenic  also  substantial-
ly involves the  issue of  interactive effects between arsenic and  other  sub-
stances,  including  certain  important  protective  effects  exerted  by  arsenic
in relation to  reducing toxicity effects  associated with excess exposure  to
certain other trace metals.
Acute Toxicity
    The typical  systemic  manifestations  of  arsenic poisoning due to  inges-
tion  usually  include gastrointestinal  disturbances  (Dreisbach,  1971).   The
intensity and onset  of  symptoms  are  determined  to some  extent by  the physic-
al form of  the  arsenical,  quantity ingested, and whether or  not   a  meal  has
been  recently eaten.  Hemolysis  is the primary manifestation of arsine  poi-
soning.
    The first symptom  of  acute  poisoning  is often  a feeling of throat  con-
striction followed  by  difficulty  in  swallowing,  epigastric  discomfort,  and
violent abdominal  pain  accompanied  by vomiting  and  watery  diarrhea  (Buch-
mann,  1962).  Intense  thirst  is  usually  present  (Rentoul  and Smith,  1973).
Cramps may be present in muscles  of  the lower limbs.   Systemic collapse  with
severe hypotension probably reflects  widespread damage to the muscular  sys-
tem.    Death  which  is   generally  preceded by  restlessness,   convulsions,  or
coma, may result from cardiac failure.  In  subacute poisoning, symptoms  are
less  intense.   If  death is  not  immediate,  jaundice  and  oliguria or  anuria
occur after 1 to  3  days (Oreisbach,  1971).   The  toxic  action of  arsenic  on
the gut  lining  epithelium  is seen  microscopically  as  a cloudy swelling  and
                                     C-46

-------
 fatty  infiltration  (Buchmann,  1962).   In less severe cases of occasional oc-
 cupational  exposure,  recovery  often  occurs  and may  either be  complete or
 followed  by recurrent  manifestations  of symptoms  characteristic  of chronic
 DOT soni ng.
 Subacute  and Chronic Toxicity
     Several  reports  of  acute  arsenical  posioning by  ingestion  have  been
 cited  in  the  older  literature (Reynolds, 1901;  Mizuta,  et  al.  1956; Takane-
 hara,  et  al.  1956;  Yoshikawa,  et  al.  1960).  An acute poisoning episode also
 occurred  more  recently  in  two Indonesian  orphanages  from the  ingestion  of
 arsenic-contaminated rice which was  prepared independently in each orphanage
 (Tjaij  and Aziz,  1971).   Laboratory analyses  showed small  amounts  of arsenic
 in  the  urine  of  five  children  and in  the  vomit of one child.   Symptoms of
 vomiting, abdominal  pain, diarrhea,  lassitude,  dizziness,  and  headache  ap-
 peared  in 109 children  and  48 adults  1 to  2  hours after  ingestion of  the
 rice.   These symptoms  were  similar to these  typically seen  with  the earlier
 incidents and accidental poisonings.
    The  acute  oral  toxicity  of  arsenic  trioxide  using  commercial  grade
 (97.77  AsjO-j   with   1.18   percent   Sb20^)  and   highly   purified   arsenic
 trioxide  (99.99   percent  As^O^)  was  tested  in  mice and  rats by  Harrison,
 et  al.  (1958).   Test  solutions  were  administered intraesophageally.    For
 Webster Swiss  mice, acute  oral  LD^Q  was  estimated  as  39.9  mg As/kg  body
weight  for  purified As^  and 42.9 mg  As/kg body weight  for  the  commerc-
 ial  grade.  The  LD^Q  for  Sprague  Oawley albino rats was  15.1 mg As/kg  for
 the  pure  ^2^2  and  23.6 mg  As/kg  for  the crude  form.    While   the  IDjg
 for the purified arsenic was lower in both  species, the  purified  arsenic  was
 a  less  severe  gastrointestinal  irritant than the commercial  form  of arsenic
 trioxide.   Retching and  gastrointestinal  damage were attributed to  the  pre-
 sence of antimony in the unpurified preparation.
                                     C-47

-------
    Several points regarding acute  arsenic  poisoning  are  of  considerable  in-
terest.  For example, trivalent arsenic is  widely held  to  be  more  toxic  sys-
temically  than  the   pentavalent  form,  based  on  both  lethality  data   and
sublethal experimental studies.  Although it  is rarely  possible to  establish
precisely  the  exposure  level  in  acute  arsenic  poisoning,   Vallee,  et   al.
(1960)  have  estimated a  human  lethality dose  of  trivalent  arsenic  as   the
oxide  to be  on the order of 70 to  180  mg.   Individual  susceptibility may  be
much  less.   Holland  (1904)  described one  patient  showing subacute  symptoms
to -8 mg of arsenic in Fowler's solution (alkaline  arsenite).
    The matter of  reversibility or  nonreversibility of  acute  poisoning symp-
toms  is  also quite important.   Survivors  of  acute arsenic poisoning display
sequelae which  involve  the peripheral  nervous, hematopoietic, cardiovascu-
lar,  hepatic,  and  integumentary systems.  The peripheral neuropathy, with  an
onset  of several  weeks,  usually  involves the lower extremities and  histolo-
gically, is manifested by long axon Wallerian degeneration,   and can persist
for years  (Ohta,  1970).   Cardiac abnormalities range from certain  electro-
cardiographic disturbances, included  T-wave  abnormalities, to eventual  con-
gestive heart failure.
    The  anemia  and leukopenia of  acute arsenic poisoning appear  to be  re-
versible features.   These  are  in  contrast  to the   longer-lasting  symptoms
such  as skin lesions, Including erythematous  eruptions  followed by  pigmenta-
tion  and keratoses of the  extremities, which are  late-emerging sequelae  of
subacute or chronic arsenic exposure.   More detailed  discussion of  the organ
systems involved in subacute and  chronic arsenic toxlcity is  presented  below.
    Systemic exposure to  amounts of arsenic sufficient  to  cause symptoms  but
inadequate to  produce systemic collapse  is  of particular interest for  de-
velopment of human health criteria for arsenic exposure.   The exposed  pat-
ient may go for weeks or  months  with gradually increasing or variable signs
                                     C 48

-------
 and  symptoms  related  to  several  organ systems and giving the appearance of a
 progressive chronic  disease state.   If  death  occurs,  it may  appear to have
 been  the  conseauence of  the  inexorable  course of  an  obscure  "natural" dis-
 ease.   Information  bearing on the induction  of arsenic  toxicity  effects  by
 subacute  or  chronic  exposure  of  humans  has  been derived from  case reports
 and  epidemiologic  studies of  people exposed via use of  therapeutic arsenic-
 als,  homicidal  or accidental  poisonings, and  occupational  or environmental
 exposures.
    The  literature  describing the constellation  of health  effects  observed
 in  connection with  various  such   exposure  conditions  or  circumstances  is
 briefly  reviewed  next,  before  assessment of  salient  information,  including
 data  on  dose-effect  or dose-response relationships,  bearing  on  arsenic-in-
 duction of  different specific  classes  or types  of systemic  health  effects
 delineated by the organ systems affected.
    Health effects  associated  with medicinal  or  therapeutic uses  arsenicals
 have been best delineated  in  relation to the use of Fowler's  solution.   The
method of  arriving at a  therapeutic dose of  Fowler's solution is  based  on
 establishing  a  patient's  tolerance  to   increasingly  higher,  but  nontoxic
 doses of arsenic.  As described  by Holland  (1904),  the patient was typically
given 5 drops  (about  9 mg  of  arsenic trioxide, or  6.8  mg of  arsenic)  well
diluted,  after meals  (i.e., three  times  a day), increasing  the dose one drop
daily until the disease  is under  control or  until  the eyelids puff  and  the
bowels move too  freely.   The dose is then  reduced to a  safer  quantity,  and
continued until  the  warning  returns,  when  it  is again  reduced.    For  some
persons even  the  minimum  dose will  produce  unpleasant effects; one  case  of
erythroderma has  been reported  after a   patient  received 10  mg of  arsenic
 trioxide (7.6  mg of arsenic) taken  over a 2-day period.
                                     C-49

-------
     There  are other  case  reports  in  the literature of  subacute  to chronic
 arsenic  poisoning  due to the  use of Fowler's solution.   Silver  and Waiimian
 (1952) described a  patient  who ingested approximately 8,8 mg of arsenic tri-
 oxide as Fowler's solution daily  for a  total  period  of  23 months,  as a reme-
 dy  for  asthma.   Signs of arsenic poisoning,  manifested  as increased freckl-
 ing  and  as darkening  of  the  nipples, first  appeared  in  association with gas-
 trointestinal symptoms after  13 months;  redness  and  puffiness  about the eyes
 and  hyperkeratoses  developed at  approximately  1.5 years.   Neurologic  symp-
 toms in  the  form of paresthesias  and weakness were the last to be noted,  oc-
 curring  after 2  years.   When  the  arsenic intake was stopped,  the pigmenta-
 tion lightened, but the  hyperkeratoses remained, and the  asthma  became more
 difficult to  control.
    Also, Fierz  (1965)  examined  262 patients who had received  long courses
 of medicinal  arsenic  6-25 years previously  and  found  keratoses in  40 percent
 and  typical  skin  cancer in  8 percent.   There was  evidence of a  dose  rela
 tionship for  both keratoses  and skin cancer.   Patients  who had received more
 than 400 ml  of  Fowler's  solution (4 g  of arsenic trioxide) had  an incidence
 of hyperkeratoses of greater than 50 percent, but  as  little as 60  ml (600 mg
 of  arsenic  trioxide)  had resulted  in  keratotic changes  in  one  patient.   As
 little as 75  ml  (750 mg of  arsenic trioxide) had been consumed by  one pat-
 ient with skin  cancer.   The  shortest  time to cancerous  change was  6  years,
with an  average of  14 years, compared  with  Neubauer's  estimate of  18  years
 (Neubauer,  1947).   Fierz  (1965)  noted  that 1,450  invitations  for  a free  ex-
 amination had been   sent to patients who  had been given  the therapeutic  arse-
 nic.  Beside  the 262  who came for examination,  100 patients provided written
 reports, and  information was obtained  about  the  deaths  of 11.  Five of  the
 11 deaths were due  to systemic  cancer,  and  three to lung cancer.   Sixteen of
 the 21  patients with cancer had typical keratoses (Fierz, 1965).
                                     C-50

-------
     In  regard to health effects  observed  with homicidal  and accidental  poi-
 soning  cases, Holland  (1904)  used  the information  gained  from personal ob-
 servation  and reports  of  suicide and  criminal  cases which used  rat  or fly
 ooison,  as well  as  Fowler's  solution, to  describe the effects  of subacute
 and  chronic  arsenic  exposure.   Occasionally,  enthusiastic patients  would
 overdo  their  use  of  medicinal  arsenic,  but this  was uncommon,  because of the
 associated  discomfort.  Holland  described  subacute poisoning  as  producing
 loss  of appetite, fainting,  nausea  and some  vomiting, dry  throat, shooting
 pains,  diarrhea,  nervous weakness,  tingling of the hands  and feet, jaundice,
 and erythema.   Longer  exposure  resulted in dry,  falling hair;  brittle, loose
 nails;  eczema;  darker  skin;  exfoliation;  and a  horny  condition (hyperkera-
 toses) of the palms and soles.
    Mizuta, et  al.  (1956)  reported on  220  Japanese patients of all ages who
 had been  poisoned by contaminated soy  sauce,  with  an  average  estimated in-
 gest ion of  roughly  3 mg of arsenic  (probably  as  calcium  arsenate) daily for
 2-3 weeks.   In  this  group, 85 percent  had facial  edema and anorexia;  fewer
 than  10 percent  had  exanthemata, desquamation,  and hyperpigmentation;  and
 about 20 percent  had peripheral neuropathy.   Except for headaches and fever,
 the findings  in these  patients  appeared to be  very similar to  those reported
by  Reynolds  (1901).   Although the  majority  of  patients'  livers were  en-
 larged,  relatively few  abnormalities were  found  in liver-function tests; and
the histopathologic description of five liver  biopsies  did  not  reveal  severe
degenerative changes.   There were no findings  suggestive of congestive fail-
ure, but electrocardiograms were abnormal   in 16 of  20 patients,  and this co-
nfirmed the reports  of Josephson,  et   al.  (1951)  and Nagai, et  al.  (1956).
The symptoms tended to  diminish after  5 or 6  days,  despite  continued  intake
of  arsenic, and  neurologic symptoms  became  prominent  as  much  as 2  weeks
                                     C-51

-------
after arsenic  ingestion was  discontinued,  at  which  time  urinary  arsenic  con-
tent remained high.  Hair was found  to  contain  arsenic at  2.8-13.0 ug/g  near
the  root,  compared with  0-1.5  ug/g near  the end  and  0.4-2.8 ug/g  in  hair
from control patients.
    For a  few  months  in  1955, a  large  number of babies  in Japan  received  a
formula made  from  powdered  milk  contaminated  with  arsenic  (Masahika  and
Hideyasu,  1973; Okamura, et al.  1956a,b; Satake, 1955).
    The subacute  symptoms of poisoning in  these infants included  the  usual
coughing,  rhinorrhea, conjunctivitis, vomiting,  diarrhea,  and  melanosis,  but
the striking presenting features were fever and  abdominal  swelling secondary
to hepatomegaly.  Abnormal laboratory findings  included  anemia,  granulocyto-
penia, abnormal electrocardiograms,  and increased density at epiphyseal  ends
of long bones  similar to  the  familiar "lead  line".  Nagai,  et  al.  (1956)  re-
ported  on   a  group of  these children  who were followed  for more  than  6
months.  Except  for  a  measurable retardation  in  ulnar  growth,   they  found
that all other  features  of the syndrome had  disappeared,  including melanos-
is.  Follow-up is continuing, and  a  report  by the Japanese Pediatric Society
(1973) indicated that growth  was still  reduced  and  that  there  was  a probable
incidence  15-30 percent of leukomelanodemia  in the children (at  the  ages of
17-20 months).  The children  had  a 15 percent  incidence of keratosis  (vama-
shita, 1972).  Of greater concern, however, was  the observation  of increased
incidences  of mental retardation,  epilepsy, and  other findings that suggest-
ed brain damage in  the  arsenic-exposed  children.   Presumably,  future studies
in this population  (more  than 10,000 exposed  infants)  will help  to  resolve
some of the  standing questions regarding the  latent effects of arsenic  expo-
sure.
                                     C-52

-------
     In  regard to occupational  arsenic exposures, it Mas  noted  by Perry, et
 al.  (1948)  that  all  of a group of chemical workers handling inorganic arsen-
 ic  compounds had pigmentary changes and  that  one third  of them had "warts,"
 although  these were  not well   described.   They reported  that  the cutaneous
 "changes  were so evident that  (the  examiner)  could  readily tell whether the
 man  .  .  , was a  chemical  worker."   All these workers had increased  urinary
 arsenic  compatible   in  degree   with  the extent  of exposure;  this  indicates
 systemic  absorption  of  the  arsenic  from dust,  probably through the lungs and
 skin.   High-exposure areas of  the  plant had  arsenic concentrations  ranging
 from about 250 to 700 ug/m  ,
    Holmqvist  (1951)  also  reported  ezcematous and follicular dermatitis  in
 smelter workers,  primarily  on  exposed skin.   Patch tests  showed sensitivity
 to  both  trivalent and  pentavalent  arsenic.   Birmingham,  et al.  (1965)  re-
 ported similar lesions  that developed  within  a few months  of  the  startup  of
 a gold  smelter that  handled  ores  containing  large  amounts of  arsenic  sul-
 fide.  Dermatitis developed 1n  half the mill  workers and  1n 32 of 40  stud-
 ents in a nearby elementary school.
    Butzengeiger  (1940) reported  that,  of 180  vlnedressars  and  cellar-men
with symptoms of chronic arsenic poisoning, about  23  percent had evidence  of
 vascular disorders of  the  extremities.  Arsenical insecticides  were used  in
 the vineyards, and exposure occurred not  only  with spraying, but during  work
 in  the  vineyards  by  Inhalation of  contaminated dusts  and  plant debris.
 Sulfur and lime-sulfur were frequently applied  in  the same solution, or  as  a
 dust with  lead arsenate.   In   either  event,  arsenite is  formed as a  func-
 tion of  the contact  time  between  the  two materials.   Most of the workers
 consumed 1-2  liters of  wine per day,  especially that  made  from musk.   It has
 recently been  shown  by Crecelius (1977)  that wine  contains high  levels  of
                                     C-53

-------
arsenite from  reductioan  during  fermentation.   Wine made from  the  musk  con-
tains  even  higher levels of  arsenite  than normal.  Exposure,  therefore,  to
arsenite would  seem  to be  convincing  in this population, with  lesser  expo-
sures  of arsenate.   All  15  workers with vascular  disorders  had  hyperpigmen-
tation, and  all  but  two had palmer and  plantar keratosis; six of the 15  had
gangrene of the fingers and  toes.
    The same association  of vascular disorders, hyperpigmentation,  and  kera-
tosis  was  observed  in Taiwan  (veh,  1963).  Urinary arsenic content  average
0.324  mg/liter,  and  hair arsenic,  0.39 ug/g.  Butzengeiger  (1949)  reported
that  the  electrocardiograms of  36  of   192  vinegrowers  with chronic  arsenic
intoxication were definitely abnormal,  with no other evident cause.   The  ao-
normalities  included  prolongation  of   the  Q-T   interval  and  a  flattened
T-wave.   In treated  cases, these  abnormalities   diminished  with the  other
evidence of  toxicity.   Similar findings were  reported  by Barry  and  Herndon
(1962) and GTazener,  et al.  (1968).
    Turning to health  effects  induced by various  environmental  exposures,  in
the early  1960s,  physicians in Antofagasta, Chile, noted demiatologic  mani-
festations and some deaths,  particularly among children,  that were  traced  to
a water supply  containing 300 ug/1  of  arsenic.   This  water supply had  been
in operation only since  1958.   Borgono  and Graiber (1972)  have  reported  on
studies of  the  inhabitants  of this city.   They  compared 180  inhabitants  of
Antofagasta with  98  people  who lived  in a  city (iQuique, Chile)  with a  nor-
mal water  supply.  Most  of  the people  studied were  less than 10 years  old.
Among  the residents of Antofagasta  the  primary symptoms  reported  were abnor-
mal skin pigmentation  (80 percent); chronic coryza (60  percent);  hyperkera-
tosis  (36  percent);  various  cardiovascular manifestations, i.e.,  Raynaud's
                                     C-54

-------
 syndrome  (30 percent);  acrocyanosis (27  percent);  abdominal pain  (39 per-
 cent); chronic diarrhea  (7  percent);  and  lip  herpes  (13 percent).  The inci-
 dence of  these  symptoms  in  the control  population was substantially lower or
 nonexistent.
    Two  additional  reports on  the Antofagasta  studies  are worthy of  note.
 Zaldivar  (1974)  further  described a  study on a  total  of 457 patients  (208
 males,  249  females)  bearing  cutaneous   lesions  (leukoderma,  melanoderma,
 hyperkeratosis,  squamous-cell  carcinoma).   Children  (up to 15 years of  age)
 accounted  for  69.2  percent  of male  cases,  and  for  77.5 percent of  female
 cases.  These patients exhibited high arsenic  content  in  the  hair.  The mean
 concentration of  arsenic in drinking water  in the period  1968-1969 was  380
 ug/g versus 80 ug/g in 1971.  Such difference  was  attributed  to  a new  filter
 plant, which  started operation  in May, 1970.   The  average  incidence  rates
per 100,000  population  for cases  with  cutaneous  lesions in  1968-1969  were
 145.5 for males  and  160.0 for  females.   The  Incidence  rates  decreased  in
1971 to 9.1 for males and 10.0 for females.
    Among  the  337 registered  children,  5  died  showing thrombosis of  brain
arteries, thrombosis of  mesenteric artery, restriction of  lumen  of coronary
arteries, and/or myocardlal Infarction.   Of the  64 registered  adult males,  2
developed multiple skin  carcinoma with lymph node metastases.
    A number of  questions  are  raised regarding  this  report.   For example,
the decrease in cutaneous  lesions  seemed  to be  too rapid,  following  instal-
 lation of the water-treatment plant,  suggesting  other  factors  were  involved.
The 8- to  10-year-old  age  group  recovered in  three  years.   Adults  exposed
for more  than  15 years  also  had a decrease  1n  incidence rate  of  cutaneous
 lesions.
                                     C-55

-------
     In  a  follow-up study,  Borgano,  et al.  (1977)  investigated  clinical  and
epidemiologic  aspects  of the cases first  reported  in  1971.   Arsenic content
in  hair and  nail  clipping  samples  of the inhabitants of Antofagasta were de
termined  and  compared  to the  levels  measured  in the  initial  study reported
in  1971.   Similar measurements and comparisons  were performed for cultivated
vegetables  and carbonated  beverages.  Also, a clinical  study  was made  in
school  children,   looking  for  cutaneous  lesions  attributed to  arsenicism.
Six years  after the water  treatment  plant  started  to operate the problem had
diminished considerably.   Arsenic  determination of hair and  nails  of  child
ren  6  years  of age or  less,  born  since the water treatment  plant  went  into
operation, indicated no  cutaneous  lesions  in this  age  group.  However,  those
over  6 years  of   age  still  had  significant arsenic  residues  in  hair  and
nails.  Although  the  clinical manifestations have  improved,  arsenic  content
of water,  soft drinks, and in some  foods  are  still considerably above  safe
levels and reouire additional sanitary engineering improvements.
    Arguello,  et  al.  (1938)  reported on a  large group of patients  seen  for
arsenical   skin  cancers  in  the Cordoba region in Argentina,  which had  a  high
arsenic content in the  drinking water {Bergoglio,  1964)  and  found  keratoder-
ma in  100  percent  of the patients.   Most patients  also had associated  hyper -
hidrosis and  abnormalities  of pigmentation, whereas those reported by  Fierz
(1965)  did not.   Arguello,  et al.   (1938) noted  that the  pigmentation  ap-
peared early and was variable  among  the  patients.   It  was  described as  small
dark spots 1  to 10 mm in diameter, with  a  tendency to coalesce,  and  appear-
ing predominantly  on  the  trunk,  that is,  in  the  areas  not exposed to  the
sun.  These and other  authors  have  noted that atrophy  may be associated  with
telangiectasia and loss  of color,  or  leukoderma, between  the hyperpigmented
areas (the "raindrop"  appearance)  cited by Reynolds (1901).
                                     C-56

-------
     In  another  study of  large-scale  environmental  contamination exposure,
 Tseng,  et al. (1968) surveyed  a  group of 40,421  residents  of the southwest
 coast  of  Taiwan (from a  population  "at risk" of 103,154) exposed to arsenic
 via  well  water  and  found that they  suffered  from a  number  of dermatologic
 and  peripheral  vascular  problems.   The  overall  male  and  female prevalence
 rates  for the clinical  findings are as follows:  hyperpigmentation 18.4 19.2
 percent,  17.6 oercent;  keratotic  lesions  7.1 - 7.5 percent,  6.8 percent; and
 black  foot  disease  0.9  - 1.2 percent,  0.7 percent  for  males  and  females re-
 spectively.   The  reason  for  the  range  of values  for males  was  not  ex-
 plained.   Skin  cancer  prevalence  rates  corresponded  directly  to  age  and
 arsenic  exposure gradients  (Table 6)  (Tseng,  1977).  The  concentration  of
 arsenic in the wells ranged  from  17  to 1,097 1.3/1.   No cases  of melanosis  or
 keratosis were  found in  a group  of  2,552 people  living in  an area where the
wells contained  almost no arsenic.
    In considering arsenic health  effects in terms  of specific organ  systems
or tissues  effected, the effects  of arsenic on  skin  are clearly  among  the
more notable  and  striking  manifestations of  the  systemic  toxicity of  the
metal.   The characteristics  of  the  skin malignancies  found  in chronic arsen-
ism have  been reviewed by Yeh  (1963)  and  Yen,  et al.  (1968)  in their  reports
on the Taiwan cases.  A prominant,  even necessary,  clinical  feature of  arse-
nical   skin  cancer  is its  association  with  the characteristic keratoses  or
pigment irregularities on  the trunk.   Several  authors have  cited  a  similar
association in exposed  workers as  evidence  that  arsenic may  cause  internal
cancers,  especially  of  the  lung  (Braun,  1958;  Currie, 1947;  Htieper,  1951;
Osburn, 1957; Robson  and  JeHiffe,  1963;  Rosset, 1958; Roth,  1957).   In ad
dition, the skin lesions  are  characteristically multiple and  predominantly
on the  areas  of the body that are  protected  by  clothing.    Both  these  fea-
tures  are notable,  inasmuch  as "ordinary" skin cancers  tend to  be  single and
                                     C-57

-------
                          TABLE  6

           Prevalence of Skin Cancer for Males*
   mq/1           20-39 yr        40-59 yr         >60 yr
                   (30)            (50)            (70)
0 -0.29           0.0015          0.0065          0.0481

(0.15)

0.30-0.59         0.0043          0.0477          0.1634

(0.45)

>0.6              0.0224          0.0983          0.2553

(1.2)


*Source:  Tseng, 1977
                           c-59

-------
 have  b^en shown  to Kave  a  body  distribution  directly correlated  with the
 amount  of Sun exposure  (Birmingham,  1971;  Miescher, 1934).   Arsenical  les-
 ions  (both keratoses and  cancers)  also  appear  at an earlier average age than
 do  solar  (senile)  keratoses  and  related carcinomas  (NAS,  1977a).   The induc-
 tion  of skin lesions had  been  a characteristic result of  oral  ingestion  of
 arsenic under all exoosure situations discussed above.
    The histopathology  of the multiple  and  varied  lesions seen  in  arsenism
 has been  the  subject of considerable  interest  among dermatopathologists  (An-
 derson,  1932; Ayres and  Anderson,  1934;  Miescher, 1934;  Montgomery,  1935;
 Pinkus and Mehregan, 1969; veh,  1963; Veh, et  al. 1968).   Lesions  that  clin-
 ically  are  keratoses may  show  proliferation of  keratin  of a  verrucous  na-
 ture, may exhibit  precancerous  derangement  of  the  squamous portions of  the
 epithelium equivalent to  those  seen in  Bowen's disease and  solar  keratosis,
 or may even be frank sauamous cell  carcinomas.  Lesions that  are  less  kera-
 totic and more  erythematous  may contain  either squamous  cell  or  basal  cell
 carcinoma  or a mixture  of cell types.   Most  authors  seem  to agree  that  kera-
 totic lesions appear to be able  to  progress  to frank carcinoma, but  observa-
 tion of such an event is rare, and  most  cancers  appear  to  arise independent-
 ly of the  keratoses.
    The question  of the  association of  Bowen's disease  with arsenism  has
 stimulated considerable controversy.   Graham and Helwig  (1959) analyzed  36
autopsies  of  patients with Bowen's disease  in  whom arsenic intake  had  been
ruled out as  much  as possible.   It is  striking  that this group of  patients
 differed from patients  with  arsenism in  several  respects:  they  lacked  the
 typical  keratoses  and   pigmentation;  they had  a tendency  for the  "typical
 Bowenoid"  squamous  cell carcinoma  uj_  situ  to precede the other cutaneous
malignancies  by an  average of 6 years; there was  an  incidence of  approxi-
mately 80 percent  of associated internal  malignancies  (some  diagnosed  only
                                     C-59

-------
 at  autopsy);  and they  had  suggestive evidence of  a  familial  predisposition
 to  the condition.  Of  more than  100 living patients  with  the  dignosis  of
 Bowen's  disease  surveyed by the  same authors,  internal malignancy  had  been
 diagnosed  in  23.  These features  seem sufficient  to distinguish Bowen's  dis-
 ease  from  chronic  arsenism, despite the confusion  later introduced  by Grah-
 am, et al.  (1961).  If  Graham  and  Helwig's  cases  are  representative, the as-
 sociation  of  systemic  cancers is  much  higher in  Bowen's  disease  than  has
 ever been suggested for chronic arsenism (MAS, 1977a).
    The  effects  of  arsenic  exposure on skin may  occur  many  years after  ces-
 sation of  exposure  (NAS, 1977a).    For example,  Braun  (1958) reported on  16
 patients who  had been  exposed to arsenic  in  their  occupation  as  vintners
many years  before.   No known  exposure  to  arsenic  had  occurred  since.   All
 had keratoses,  nine had  leukomelanoderma  of the  trunk,  and seven  had  skin
cancer or intraepidermal carcinoma j£ situ.  Eight had lung cancer.
    Roth (1957) also studied 47 vintners whose  arsenic  exposure had  occurred
8-14 years  earlier.  His  population was  selected by  having  come  to  autopsy.
He found that  33 of the 47 had cancer.   A total  of  75 malignant  tumors  (40
of which were skin  cancers) of various  tissues were  observed:   18 cases  had
 lung cancer,  6 with hemangiosarcoma of the  liver,  5  with  esophageal carcin-
 ima, and 1 with bile duct carcinoma.
    Cardiovascular  effects  of  arsenic have been  demonstrated to  occur  with
acute and  subacute  exposure of humans to  inorganic arsenic and  may include
quite  severe  cardiovascular  involvement,  with  congestive  heart   failure
identified  as  a  cause  of death  in fatalities  encountered  in  one posioning
outbreak (Reynolds, 1901).  More  recent clinical  assessments of subacute and
chronic  arsenic poisoning of  large numbers  of  people  (Mizuta,  et al.  1956;
Hamamoto, 1955; Borgono and Greiber,  1972;  Tseng,  et  al.  1968)  indicate  that
                                     C-60

-------
 the  extent of  cardiovascular  injury with  the nature  of  exposure, subject,
 geographic area,  and  level  of arsenic intake.
     Hamamoto's  (1955) clinical  findings  of cardiovascular  injury  in  12,000
 infants   consuming   arsenic-contaminated  milk   included   elevation  of  the
 ST-wave  and extension of  the  QT-interval,  cardiographic  changes  which  were
 slow  to  revert to  normal  after  exposure and  ceased.   Similar cardiographic
 data  for  200  patients who  consumed arsenic-contaminated soy sauce were noted
 (Mizuta,  et al. 1956).
     In the clinical  survey by Sorgono  and  Greiber  (1972)  of  both  pediatric
 and  young adult  victims  of chronic  arsenic  exposure  via  a  drinking  water
 supply, cardiovascular symptoms  seen included  Raynaud's syndrome, acrocyano
 sis,  angina pectoris, hypertension, myocardial  infarction,  and  mesenteric
 thrombosis.
    Tseng  (1968,  1977) described  the incidence of  "black foot  disease"  in
 Taiwanese  consuming well water containing relatively high  levels  of arsenic.
 The disorder,  a peripheral vascular  derangement arising  from  arteriosclero
 sis and thromboangitis obliterans, results  in  gangrene  of  the  feet  and shows
 an increasing prevalence  with increasing arsenic content of drinking water.
    Chronic exposure  to  arsenic  in  occupational  settings  has  also  been  re
 ported to  be  associated with  various cardiovascular disorders.   Vine  dres-
 sers who  had  been exposed  to trivalent arsenic  showed  late onset  (30  years
post-exposure,  exposure  time of  20 years)  peripheral  vascular  sequelae  in
 the  form  of  endangiitis  obliterans  and  acrodermatitis atrophicans  (Grobe,
 1976).  The role  of arsenic in increased cardiovascular disease mortality  is
 suggested  by  the  epidemlological   investigations of  smelter workers by  both
 Lee and Fraumeni (1969) and Axelson,  et  al.  (1978).
                                     C-61

-------
    Little  information  is  available regarding  useful  animal models  of  the
 cardiovascular  effects  seen  in  arsenic-exposed  human  subjects.   However,
 oral  exposure  (1.5  mg  As/kg  body weight) of cats to either arsenate or arse
 nite  in feed  (Massmann  and Opitz,  1954)  was followed by  flattened T-wave and
 lengthened QT-time  in the electrocardiogram.
    Neurotoxic effects  of arsenic  have  long been recognized as being associ-
 ated  with  acute,  subacute,   and  chronic  exposures to relatively  high  levels
 of  inorganic  arsenic.   These  effects   include  both clinically  significant
 peripheral nervous  system  (PNS) and central nervous  system (CNS)  damage  re-
 ported  as  occurring in cases of accidental  or  homicidal  arsenic poisonings,
 prolonged  occupational   exposures,  and  certain  therapeutic  applications  of
 arsenical  compounds.   Such  marked neurotoxic effects have been  fairly  well
 characterized  in  terms of their major  pathophysiological  features,  clinical
 courses and  sequalae,  and associated  histopathology.  Much  less  well  char-
 acterized  are  quantitative   dose effect/dose response relationships  defining
 arsenic exposure  parameters  associated  with Induction of  neurotoxic effects
 in humans.
    Reynolds  (1901) provided  one  of the  earliest  detailed  descriptions  of
 arsenic-induced neurotoxic   effects  in  reporting  on clinical  findings  for
more  than  500  patients  that  had consumed  arsenic-contaminated beer.   As  de
 scribed elsewhere  (NAS,  1977a),  Reynolds (1901)  reported  that  neurologic
 signs  and  symptoms began before the appearance of  classical skin  lesions,
 but followed such an Insidious  course  of development so  as to have  gone  un-
 diagnosed  for  several  weeks.   Neurological  Involvement  started  with sensory
 changes, e.g.,  paresthesias, hypersthesias,  and neuralgias,  accompanied  3y
 considerable muscle tenderness.   Varying degrees of motor  weakness,  progres-
 sing  from distal  to proximal muscle groups, also  occurred  and  culminated  at
                                     C-62

-------
 times  in  paralysis of affected muscle  groups  or  extremities.   Certain indi-
 cations  of central  nervous  system  (CNS)  damage,  e.g.,  loss  of  memory  and
 general mental  confusion,  were  also observed but  were discounted by Reynolds
 (1901)  as being less likely due  to arsenic than chronic  alcoholism  or con-
 current excessive  selenium intake.
     Perpherial  nervous  system  (PNS) effects  similar  to those  described  by
 Reynolds  (1901)  have since been  observed  in numerous other cases  of  acute,
 subacute,  and chronic arsenic  exposures  (Silver and Wainman,  1952;  Mizuta,
 et  al.  1956; Heyman, et  al.  1956;  Jenkins,  1966;  Hara,  et al.  1968;  Chut-
 tani,  et  al.  1967;  Ishinishi,  et  al.  1973;  Nakamura, et  al.  1973; Nagamatsu
 and  Igata,  1975;  O'Shaughnessy  and Kraft, 1976; Frank, 1976;  Garb and Hine,
 1977;  LeQuesne  and Mcleod,  1977)   and are  now  recognized  as classic clinical
 symptoms  of  arsenic  poisoning.  Such symptoms  include peripheral  sensory  ef-
 fects  characterized  by the  appearance  of  numbness,  tingling, or  "pins  and
 needles"  sensations  in  the hands  and feet, as well  as  decreases  in  touch,
 pain,  and  temperature sensations  in a symmetrical  "stocking glove" distribu-
 tion.  These  symptoms are  often variously accompanied by  burning sensations,
 sharp  or  shooting pains,  and  marked muscle tenderness  in the  extremities.
 Peripheral  neuritis  symptoms  originate distally  and, over  the  course of  a
 few  weeks,  often  progressively  become more  widespread in both lower  and  up-
 per  extremities,   usually  appearing  first  in  the  feet  and  later  in  the
 hands.  Signs and  symptoms  of  peripheral  motor nerve effects  include:   sym-
metrical muscular  weakness  of  the  extremities,  predominantly  distal   but  at
 times extending to proximal  muscle groups and, rarely, the  shoulder  or pel-
 vic  girdle; evidence of foot and/or wrist drop; and, in some  cases,  rapidly
 developing  paralysis  and  atrophy  of  lower  leg muscles  and small  muscles  of
 the  hand.
                                     C-63

-------
    Collectively,  the  above  components  of  the  classical  clinical  syndrome
associated with excessive arsenic exposure  are highly  indicative  of  progres-
sive peripheral polyneuropathy,  involving both sensory and motor  nerves,  and
most  intensively  affecting  long-axon  neurons.   Several  studies  (Jenkins,
1966;  Nagamatsu  and  Igata,  1975; O'Shaughnessy  and Kraft,  1976;  Garb  and
Mine,  1977;  LeQuesne  and McLeod, 1977)  have provided quantitative  electro-
physiologic data,  in the form of  electromyographic  (EMG) or  nerve conduction
velocity (NCV) recordings, confirming arsenic induced peripheral  nerve  func-
tional  deficits  in  association with  the manifestations  of frank  clinical
signs and symptoms of the above  type.   In addition,  biopsy and  autopsy  stud-
ies have provided  histopathological  evidence verifying peripheral nerve dam-
age, especially Wallerian degeneration of long-axon  myelinated  nerve  fibers,
in cases of  human  arsenic  exposure  where frank neurological  signs and  symp-
toms were manifested  (Heyman,  et al. 1956;  Jenkins, 1966;  Chuttani, et  al.
1967; Ohta,  1970;  LeQuesne  and  McLeod,  1977).  Such degenerative changes  in
myelinated long-axon neurons are consistent  with  human autopsy  findings dis-
cussed earlier regarding the  uptake  of arsenic into peripheral nerves  (Lar-
sen, et  al.  1972)  and preferential  accumulation  of the metal  in CNS  white
matter with  high content  of "fatty"  components of  neural tissue, e.g., mye-
linated nerve fibers (Larsen,  et al.  1979).
    Several   additional  points  regarding  arsenic-induced  peripheral neuro-
pathies warrant  special  attention here  for  present  health   assessment pur-
poses,  including   consideration of  such  issues   as:    (1)  arsenical  forms
identified as  inducing clinical  neuropathies; (2)  pattern(s) of  development
of neuropathic effects;  (3) persistence of,  available  therapy for,   and  re-
covery from  such   effects;  and  (4)   effective  exposure  or dosage parameters
associated with their induction.
                                     C-64

-------
     In  regard  to such issues, clinically-manifest cases of peripheral neuro-
 oathies  have been demonstrated  to  occur as the  result of exposure  to many
 different  inorganic  arsenic  forms  under  a variety of circumstances.  For ex-
 ample,  peripheral  neuropathy effects have been documented  in  clinical  cases
 of  acute homicidal,  suicidal,  or accidental  poisonings  involving ingestion
 of  various  commercially  available  herbicides,  pesticides,  and  animal  poisons
 containing  inorganic  arsenic compounds  such  as:   lead arsenate;  sodium arse-
 nate;  calcium  tn'arseniate;  copper acetoarsenite  (Paris   Green);  arsenious
 oxide; and  arsenic  trioxide  (Heyman, et  al.  1956;  Jenkins,  1966;  Ohta, 1970;
 O'Shaughnessy  and  Kraft,  1976).   Similarly,  peripheral neuropathy have been
 observed  following  acute,   subacute,  or chronic  occupational  exposures  to
many of  the  same arsenic  compounds,  e.g., in the  course of agricultural  ap-
plications  of  calcium or  lead  arsenate  insecticide  sprays, or  occupational
exposure to  arsenicals  such  as:  arsenious acid and  other  tri- and pentaval-
ent  inorganic  arsenic  compounds  encountered  in  a coal gas  desulfurization
processing  facility (Hara,  et  al.  1968); and  arsenic trihydride  or arsine
 (Frank,  1976).    Thus, regardless of  the  particular inorganic  arsenic  form or
valence  state  involved,  it  appears  that excessive exposure to arsenic  from
any of the above substances can result  in severe peripheral  neuropathy.
    In regard  to the development  and  persistence of  peripheral  neuropathy
associated with  arsenic  exposure,  somewhat variable  patterns of  onset,  per-
sistence, and response to treatment have  been  observed, depending in  part on
the nature of specific exposure parameters.  LeQuesne and McLeod  (1977),  for
example, reported  fairly  rapid  onset  of peripheral  neuropathies  involving
both motor  dysfunctions  and paresthesias, which  appeared  in four  patients
within 10 days  to 3 weeks after ingestion of single  large doses of inorganic
arsenic  compounds  (e.g.,  sodium  arsenate and arsenious oxide).   Further  de
terioration occurred for  a  few  days in  3  of the  patients  and  progressively
                                     C-65

-------
worsened  for  5  weeks for the fourth,  as  indexed  by NCV recordings  and other
observations  of clinical  signs  and  symptoms.   All  improved slowly  there-
after,  but  after  6  to 8 years,  3  patients still  had  abnormal  neurological
signs and symptoms;  NCVs, too, were still  not  in  the  normal  range and  marked
atrophy of  affected  muscles was  evident  in some cases.   The pattern of onset
and persistence of neuropathy signs and  symptoms  observed  by others (Heyman,
et  al.  1956;  Jenkins,  1966; Nagamatsu  and  Igata,  1975;  O'Shaughnessy  and
Kraft,  1976;  Garb  and  Hine,  1977) for  acute arsenic poisoning are consistent
with  those  reported  by LeQuesne  and  McLeod (1977); that  is,  the neuropathy
typically become clinically manifest within  a week  or  two  after  exposure  and
slow  incomplete recovery  is  seen over a course of  years,  with some patients
continuing to require the aid of  leg braces  to  walk.   Delayed onset of symp-
toms, 1 to  6  months  after  acute  exposure to arsine has been reported (Frank,
1976) for six industrially exposed workers.
    Under more  chronic occupational exposure  conditions to  lower  levels  of
arsenic compounds, the  development  of  neuropathy symptoms can be more grad-
ual  and  insidious  and not  only bilateral  but unilateral  polyneuropathies
without motor paralysis have been  reported  (Ishinishi,  et  al.  1973;  Naka-
mura,  1973).   Again,  the  time  course for  recovery  from the  neuropathies,
once  induced, tends to be slow and on  the  order of  years.   Gradual  onsets  of
peripheral  neuropathies and  slow recoveries  have also been reported  with
subacute  or  chronic  exposures to arsenic  via ingestion of  contaminated  soy
sauce (Mlzuta, et  al.  1956)  or  anti-asthmatic  herbal  preparations containing
arsenic trloxlde or asenic sulfide (Tay and Shea,  1975).
    In  regard to  effective   dosage  parameters for induction of peripheral
neuropathies  by arsenic,  it  is  usually  not  possible  to determine  precise
doses involved or  periods of  exposure.   For most  acute poisonings,  it  is  us-
ually evident that high  level exposure (on the order of tens  or  hundreds  of
                                     C-66

-------
mg or more)  occurred,  frequently  involving  only  a  single dose.  For subacute
or  chronic  poisoning  situations,  information  exists  from  few  studies  by
which  effective  exposure  parameters  can   be  estimated.   Mizuta,  et  al.
(1956),  for example,  reported  that peripheral  neuropathies  occurred  in  20
percent  of 220 patients of  all  age groups  poisoned by  ingestion of  arsenic
contaminated  soy  sauce,  with approximately  3 mg  arsenic  (likely as  calcium
arsenate)  estimated  to be  ingested daily  for  2-3  weeks  resulting in  total
effective  doses  up to approximately  60 mg.  Also,  Tay  and Shea (1975)  re-
ported  polyneuropathies  in  approximately 50 percent of  74 patients  poisoned
by daily ingestion of  3.3 or  10.3  mg/day of  arsenic  trioxide  or arsenic sul-
fide  in  antiasthmatic  medicinal   pills.    Similarly,   Silver  and Wainman,
(1952) reported on a patient  that  had  ingested  approximately  8.8 mg of  arse-
nic trioxide daily for 28 months  as an asthma  treatment.  Signs of  peripher-
al neuropathy  appeared at  about  two  years, well  after  the  onset  of  other
arsenic-related effects,  e.g.,  skin  changes;  assuming  regular  ingestion  of
the arsenical each day for  two  years,  then, the neuropathy appear to be  as-
sociated with  gradual  exposure  to  a  maximum total  dose  of up to 650  mg  of
arsenic.  Comparison of this  estimate  (650 mg) with  that from the Mizuta,  et
al. (1956)  study  (60  mg) suggests  marked variation in  individual  suscepti-
bility to  neurotoxic  effects of  arsenic resulting  in frank  clinical  neuro-
pathies.
    The   above  studies characterizing  clinically-flianifest  peripheral  neuro
pathies   with relatively  high  acute,  subacute,  or  chonlc  exposure,   have
raised Questions as to whether  similar but  subtle neurotoxic  effects  are  in-
duced by chronic exposure to  lower  levels  of arsenic.   Takahashi  (1974),  for
example, reported that abnormal  electromyograms  (EMG)  were found in  the  ab-
sence of subjective  symptoms  among population  groups living  in  the  vicinity
of an arsenic mine and smelter in Japan.
                                     C-67

-------
     Few  other epidemiology studies have attempted to delineate more precise-
 ly  qualitative or quantitative  relationships  between chronic  arsenic expo-
 sure  and the  induction  of peripheral  neurotoxic effects indexed  by  EMG or
 NCV  recordings and neurologic examinations.  Landau, et  al.  (1977) reported
 relationships  between  length  and  intensity  of  occupational  arsenic exposure
 (mainly  to  arsenic trioxide  via  inhalation) of  smelter  workers  and altera-
 tions  in peripheral  nerve  functioning.  The manner  in  which  the  data  were
 reported,  however, precludes  precise characterization of  dose-effect/dose-
 response  relationships.   Similar  difficulties  exist  in  terms  of attempting
 to  characterize  such  relationships for arsenic-induced peripheral  nerve  de-
 ficits demonstrated by EMG  recording  in studies of two other chronically ex-
 posed populations:  (1)  an  Indian  population exposed  partly  via occupational
 contact with  arsenic in  a  gold  mining and smelting facility  in Yellow Knife,
 Canada, or via arsenic emissions from the facility into  the  ambient environ-
 ment (Canadian Public Health  Assoc.,  1978);  and (2)  a Nova  Scotia population
 exposed  via  geologically  natural  arsenic  contamination  of  wells   used  for
 drinking water (Hindmarch, et al.  1977).
    Several  of  the clinical  reports  dicussed  above not  only  document peri-
 pheral  nerve  damage induced  by exposure to arsenic,  but  also  contain  de
 scriptions of  arsenic-induced central  nervous  system (CNS)   disturbances  or
 encephalopathic  effects  ranging in severity from memory losses  and general
mental  confusion to convulsions, stupor,  coma,  and even  death (e.g., Heyman,
 et  al.   1956;  Jenkins,   1966;  Frank,  1976;   Nagamatsu  and  Igata,   1975;
 O'Shaughnessy and  Kraft, 1976;  Garb  and Hine,  1977).  The onset  and courses
 of  such  CNS  effects  have  not  been  well  defined,  but  appear  to  parallel
 rather closely  the development of peripheral  neuropathy; and cases of  pro-
 longed  encephalopathy  indexed  by  electroencephalogram  (EEG) recordings  of
 abnormal  brain wave patterns  up to a  year  after cessation  of  exposure  have
                                     C-68

-------
 been reported  (Freeman  and Couch, 1978;  BentaT,  et al. 1961).   Very  little
 information regarding dose-effect/dose-response relationships for  arsenicin-
 duction of CNS effects  can be derived from these studies, however, and  such
 effects appear to be a  much  less  constant feature of arsenic-induced neuro-
 toxic effects  in  adults  than  are peripheral  neuropathies.
     Certain studies  suggest,  in contrast, that children may be more suscept-
 ible to arsenic-induced  CNS  damage.   Severe  CNS  deficits were  observed in
 children  exposed for several  months  as   babies  to  arsenic-contaminated  pow-
 dered milk  formulas in  MoHnaga,  Japan (Hamamoto,  1955;  Okamura, et  al.
 1956b;  Vamashita,  et al. 1972; Masahiki   and Hideyasau, 1973;  Japanese Pedia-
 tric Society,  1973).  Follow-up studies  on the children exposed  to arsenic
 as  infants have  revealed:   (1) increased  incidence of severe  hearing  loss
 (>30 dB)   in  18 percent of 415 children  examined,  compared  to  less  than  1
 percent  incidence of hearing  loss  in corresponding age group  children;  (2)
 increased   incidence  of abnormal  electroencephalographic  (EEC)   brain  wave
 patterns  in 14 percent  of the exposed   children,  more  than  double the  ex-
 pected  rate for  comparable normal pediatric  populations;  and  (3)  observa-
 tions  of   increased  incidences of  persisting  mental  retardation,  epilepsy,
 and  other  indications of severe brain damage.   In  additon,  Ohira and  Aoyama
 (1972) reported not  only increased EEG abnormalities  but also  visual  system
damage, including pathological  eye  changes,  in  children  fed  the  arsenic  con-
taminated  powdered  milk  in  comparison   to nonexposed  breast-fed  infants.
Taking into account  known  information regarding length of exposure  and  dose
 levels, it  can  be calculated  that the above persisting  (probably permanent)
types  of  CNS  damage effects  resulted  from ingestion  of approximately  3.5
mg/day of arsenic resulting in a total intake of about 90-140 mg.
     In another  study (Bencko  and  Syman,   1970), hearing  losses  in  children
were  reported  to  be  associated with arsenic exposure  derived  from emissions
                                     C-69

-------
from  a  nearby power  plant  combusting high-arsenic  content coal.   Both  air
and bone  conduction  hearing  losses were observed, suggesting  inner  ear dam-
age.  Failure to find  analogous  hearing  losses  in  children  exposed  to atmos-
pheric  arsenic  emitted from  a  nearby copper  smelter in  the United  States
(Milham,  1977)  has  raised questions regrading  arsenic-induced damage  to  the
inner ear in children.   Evidence  supportive of  the possible occurrence  of
such  effects  has been  obtained  in an animal  toxicology study (Aly,  et  al.
1975) that  demonstrated hearing  losses  and histopathological  confirmations
of  destruction  of  the  organ  of  Corti  and other  inner ear damage in  guinea
pigs  exposed  to  arsenic over a  two-month  period  via  intraperitoneal  (i.p.)
injections of sodium  arsenate solutions  at a dosage  level  of  0.2 mg/kg body
weight.
    Very  few animal  toxicology  studies  have  focused  on  investigation  of
neurotoxic effects of arsenic on the CNS.   Rozenshstein  (1970),  for  example,
reported  evidence of  CNS functional deficits,  as indexed  by  altered  condi  -
tioned  reflexes, as well  as  histopathologic evidence  of  CNS  structural dam-
age,  e.g.,  pericellular edema and  neuronal  cytolysis in the  brain,  in rats
exposed  for  three months  to an  arsenic  trioxide aerosol  resulting  in  an
arsenic  concentration  of 46  ug/m .   Similar  but less  severe effects  were
also  obtained with  exposure  of  other  rats to  a  3.7 ug As/m3  aerosol.  CNS
deficits,  indexed  by   impaired  avoidance  conditioning   in the  absence  of
demonstrable  histopatholotic  changes  in  brain tissue,  were also  reported
(Qsata,  1977) for suckling rats administered 2  or  10 mg  arsenic  trioxide via
stomach intubation over a 40 day period.
Synergism and/or Antagonism
    Moxon (1938) first  demonstrated  the  protective effect of arsenic against
selenium  poisoning when he found  that  arsenic at  5,000  ug/1  as  sodium asen-
ite in  the  drinking  water largely prevented  liver damage in  rats whose diet
                                     C-70

-------
 contained  selenium  at  15  ug/g  as  seleniferous  wheat.   Moxon  and  Dubois
 (1939)  then  showed  that arsenic was unique in  its  ability to prevent selen-
 ium  toxicity;  all  other elements studied were  unable  to protect against all
 manifestations  of  chronic  selenosis.  Sodium  arsenite  and  sodium  arsenate
 were  equally effective against seleniferous grain,  but  the arsenic  sulfides
 were  ineffective  (Dubois,  et  al. 1940).  Arsanilic  acid  and 3-nitro-4-hydr-
 oxyphenylarsonic acid,  two  organic  arsenicals  used  as  "growth-promoters" for
 livestock, also  exhibited  a beneficial action  against selenium  poisoning  in
 rats  when  given  in the drinking water (Hendrick,   et  al. 1953).  There  is
 evidence that it would  be practical  to  use  these  two agents to  protect swine
 and  poultry  in high-selenium  regions  (Carlson, et  al.  1954; Wahlstrom,  et
 al.  1955).   Amor and Pringle  (1945)  even  suggested the  use of  an  arsenic-
 containing tonic  as a  prophylactic  agent  against  selenium poisoning  in ex-
 posed industrial workers.
    The  metabolic  basis for  the beneficial  effect of  arsenic   in  selenium
 poisoning  remained  confused  for some  time,  because  arsenic was known  to
 block the biosynthesis  of dimethylselenide,  a  detoxification  product  in  ani
mals  that  received  subacute doses  of selenium by  injection (Olson,  et al.
 1963).  Moreover, the protective  effect of   arsenic against dietary  selenium
was not  seen  if the arsenic was  given  in  the  diet, instead  of  the  drinking
water (Ganther and  Baumann, 1962a).   Frost  (1967) has  shown that the toxici-
ties of arsenic and selenium are  additive  if both elements are  given  in the
drinking water.  These  results  agree with those of  Obenneyer, et al.  (1971)
who recently  observed an additive  toxicity  between arsenite and trimethyl-
 selenonium chloride or dimethylselenide.
    Ganther and Saumann (1962)  studied  the  influence of  arsenic  on the meta-
 bolism  of  selenium when  both   elements are injected  in  subacute doses and
found that  the  excretion  of  selenium  into the  gastrointestinal tract was
                                     C-71

-------
markedly  stimulated  by  arsenic.   Levander  and  Baumann (1966a)  observed  an
inverse relationship  in  arsenic-treated rats between the amount  of selenium
retained  in  the liver and  the  amount  excreted  into  the gut; and  they  con-
cluded that  the bile  might be  the route  by which selenium was  appearing  in
the gastrointestinal tract.  This hypothesis proved correct when  it was  dis-
covered that  in three hours over 40  percent  of  the  selenium  injected  could
be recovered  in the  bile of rats that  also  received  arsenic,  whereas  only 4
percent of the  selenium was excreted  into the bile of rats not given arsenic
(Levander and Baumann, 1966b).  This  effect  of arsenic  on  the biliary excre
tion of  selenium was  not  confined  to  subacute  toxlcity experiments: a  re
sponse of  selenium to arsenic  was  seen  at  doses approaching a  rat's  daily
intake of  selenium when  fed  some crude  commercial  diets.   Sodium arsenite
was the most  effective form of  arsenic  in enhancing  the biliary excretion  of
selenium, but arsenate  and  3-nitro-4-hydroxylarsonate  were  also  active  to
some extent.    In  experiments  with  radioactive  arsenic, it  was found  that
selenium stimulated the biliary excretion of arsenic, Just as arsenic stimu-
lated the excretion of selenium.  Initial attempts to characterize  the  forms
of selenium  in  rat bile  suggested that the element  is probably present  in
several forms, including  some macromolecularly bound  selenium.
    Although  these studies provide  physiologic  information  concerning  the
interaction  of  arsenic and selenium,  the chemical  mechanism of  the process
is still  far from clear.   The most  logical  hypothesis to  account for  the
arsenic-selenium antagonism  from the  molecular   point  of  view  assumes  that
arsenic combines with  selenium--perhaps,  in  analogy  with sulfur  chemistry,
by reacting  with  selenol  (-SeH)  groups to  form a detoxification  conjugate
that passes  readily into  the bile (NAS,  1977a).
                                     C-72

-------
 Teratogenicity
     Although  few human epidemiologic studies have provided evidence of arse
 m'c-induced  reproductive  or teratogenic effects, several  studies  have shown
 that sodium  arsenate  induces  developmental  malformations  in  a  variety  of
 test animals:   embryo  chick, hamster,  rat,  and  mouse  (Ancel,  1946; Ridgeway
 and  Karnovsky,  1952;  Perm and Carpenter,  1968;  Hood  and Bishop,  1972; Beau
 doin, 1974).
     Pregnant  golden  hamsters injected  with  sodium  arsenate (15 to  25 mg/kg
 body weight)  produced  offspring with a  range of developmental  malformations
 including  anencephaly,  renal agenesis,  rib malformation,  cleft  lip and pal-
 ate,  and  anoohthalmia.  The percentages  of  living  embryos with various  se-
 lected malformations following maternal  treatment with  20 mg/kg  sodium arse
 nate  on  the 8th day of  gestation were  as  follows:   nearly 90 percent with
 all  malformations; over  80 percent with anencephaly; nearly  70  percent with
 rib malformations; and 30  percent  with  exencephaly.   The spectrum  of malfor-
 mations varied with  the  time of injection during critical  stages  of embryo
 genesis.   Malformations  induced  by  arsenate differed  from those  induced  by
 other teratogenic agents including certain heavy metals  (Ferm,  et  al. 1971).
     In another  study,  single intrapen'toneal  injections of sodium arsenate
 (45 mg/kg)  in Swiss-Webster  mice between the 6th and llth days  of gestation
consistently caused an increase  in fetal  resorptions, a  significant decrease
 (p<0.05) in  fetal  weights compared  to  controls, and a  number of  fetal mal-
formations, most  frequently the  following:  exencephaly,  shortening  of  the
jaws with  consequent protrusion  of  the tongue,  exophthalinos, missing  pinna,
cleft lip,  hydrocepnalus, umbilical  hernia, ectrodactyly,  micromelia,  and
shortened  or twisted tail or limb, or both.  Malformations were  dependent  on
the stage of embryogenesis.  Exencephaly  occurred  in 54 percent of  the fet-
uses when  the  injection was  administered on day 9  of  gestation;   fusion  of
                                     C-73

-------
the  ribs occurred  in  100 percent of the fetuses when the injection was given
on day 9; and fusion of  the  vertebrae  occurred  in  73 percent when the injec-
tion was given on day 10  (Hood and Bishop, 1972).
     In  a  later  report,  Perm (1977) demonstrated  that  administration of  20
mg/kg  of  sodium arsenate  intravenously or intraperitoneally to  golden  ham-
sters during days 8 to 9  of  gestation  induced  a specific  spectrum of malfor-
mations including  exencephaly,  encephaloceles,  skeletal  defects,  and malfor-
mations of  the  genitourinary system.   The  last effect,  which  appears to  be
uniaue to arsenate, occurred in both sexes and with high  frequency.
     Perm  (1977)  further  showed that  radioactive  arsenic   (  As)  injected
intravenously  into Golden  hamsters  on  day 8  of  gestation  was  transmitted
across the  placenta  during the critical stage  of  embryogenesis and appeared
in the fetal  tissues.   Perm (1977)  refers  to  a report  concerning  a  case  of
arsenic trioxide poisoning  during  human  pregnancy,  which  demonstrated  the
"ease with  which inorganic arsenic  crosses the human placenta at  term  with
extremely high  levels  in the fetal  liver,  brain,  and kidneys"  (Perm, 1977).
Introduction of  arsenic  into fertilized bird  eggs has led  to  malformations
of beak and brain (Peterkova and Puzanova, 1976).
    Hood,  et al. (1977) compared  the prenatal  effe  s of  oral and intraperi-
toneal administration  of  sodium  arsenate  in mice.   Intraperitoneal  admini  -
stration had a  considerably greater effect than oral administration  on  pre
natal mortality, reduction of fetal  weights,  and occurrence  of  fetal  malfor-
mations.  The dosages were 40 mg/kg  (intraperitoneal) and 120 mg/kg (oral).
    Hood,  et al. (1977)  further  noted  that although arsenite is considerably
more toxic  than  arsenate,  it has received  less attention  from teratologists.
Intraperitoneal  injection  of mice i_n  utero with 10  to- 12 mg/kg  arsenate  on
one  of  days 7 to  12  of pregnancy caused  significant increases  in prenatal
mortality  (o
-------
gross  and skeletal  malformations  similar  to  but  less  frequent  than  those
induced by comparably toxic levels of arsenate (Hood, et al. 1977).
    In another  study,  Tamura  (1978)  found no effects  on  growth and develop-
ment  of  rats  fed arsenic trioxide from  the  7th  to the 21st  day postnatal ly
at  a  dose level  of 1.5  mg/kg/day  in comparison  to  a 50  percent  mortality
rate at a 15 mg/kg dose.
Mutagenicity
    Most  mutagenesis  research  has  centered on chromosomal  reactions  to  sod-
ium arsenate.   There are  no  data based  on the  host-mediated  assay or  the
dominant  lethal technique (NAS, 1977a).
    One  of  the  earliest observations  that has  meaning  today  was made  by
Levan  (1945).   Root men'stem  cultures  of  A11 iurn  cepa  were  treated  for  4
hours with an  unspecified arsenic  salt at  10 concentrations,  from lethal  to
a no-effect.   Chromosomal  changes were  observed,  including  spindle  distur-
bances and metaphase arrests.   Similar effects,  with  minor variations,  were
observed  after  treatment with  salts  of  24  other metals (mostly  nitrates).
The changes resembled those caused by  colchicine, but  they  cannot  be consid-
ered serious  damage (NAS, 1977a).
    Petres and  Hundeiker  (1968)  and  Petres,  et  al.  (1970,  1972) have  re-
ported chromosomal breakage  in  human leukocyte cultures after  short-term  J_n
vitro  exposure  to  sodium arsenate and  in cultures obtained  after  long-term
exposure to arsenical compounds in  vivo.
    The cytotoxic  and  mutagenic effects  of sodium arsenate  were  tested  j_n
vitro on  phytohemagglutinin-stimulated lymphocyte cultures  at concentrations
of 0.05-30 ng/ml  of culture medium  (Petres,  et  al. 1970).   It  was reported
that 33 percent of metaphase plates  were pulverized at 0.1 ug/ml  and  80-100
Descent at concentrations of  2  ug/ml  or  greater.  The "mitosis  index"  and
the "( H)thymidine-labeling  index"  were  decreased.   Arsenate has  also
                                     C-75

-------
found  to  increase the total frequency of  exchange  chromosomes  in Drosophl la
melanoqaster  treated with  selenocystine,  (Walker  and  Bradley,  1969).   The
overall  significance of  these  chromosomal  studies  is  difficult  to  assess,
inasmuch  as  many unrelated  compounds  may cause  similar effects.  The  fact
that arsenic  compounds have  caused  chromosomal  damage  in  a  number of  biolog-
ic systems, however,  should  alert toxicologists to  a  possible  role of arsen-
ic in chemically-induced mutagenesis (NAS,  1977a).
    _In_  vivo studies  were  made on 34 patients  at  the University  of  Freiburg
Skin Clinic (Petres,  et  al. 1970).   Thirteen of  these  patients  had  recieved
intensive  therapy,  some  more than  20  years  before  the experiment;  most  of
these were psoriasis  patients.  The  control  group (21  patients)  consisted  of
14 psoriasis  patients and seven with  eczema,  none of  whom had  had  arsenic
treatment.  Phytohemagglutinin-stimulated  lymphocyte  cultures were  prepared
from each  patient for evaluation  of chromosomal  aberrations.   The incidence
of aberrations  was remarkedly  greater  in  the  cultures of patients  who  had
been treated with arsenic.   Expressed as the frequency  per  1,000 mitoses,  49
secondary constrictions occurred in  the arsenic group and  12  in  the  control;
gaps were  found  in  51  in  the  arsenic  group and seven in  the  control;  26
"other" lessions  occurred  in the arsenic  group and one in the  control;  and
broken chromosomes appeared at the rate of 65 per 1,000 mitoses  in the arse-
nic group and two in  the control.   Aneuploidy was found at  the expected  fre
auency  in  the  arsenic  group.   The extent  of abnormalities  attributed  to
treatment  with  arsenicals  is impressive;  it  is important that  this  study  be
repeated (NAS, 1977a).
    The occurrence of chromosome  aberrations  was  studied by Beckman,  et  al.
(1977)  in short-term cultured leukocytes from mine  workers  exposed to arsen-
ic at  the  Ronnskar smelter  in  northern  Sweden.  In the smelter  workers,  87
aberrations were found in  819 mitoses  (Table 7).   The  number  of aberrations
                                     C-76

-------
                              TABLE  7

       Oromosome Aberrations in Workers Exposed to Arsenic
                from Ronnskar,  Sweden  and  Controls*

No. of cells
No. of aberrant cells
No. of aberrations
Gaps
Chromatid aberrations
Chromosome aberrations
Total
No. of aberrations per cell
Freauency of aberrant cells, %
Arsenic
Workers
819
71

56
12
19
87
0.1062
8.7
Controls
1,012
13

9
3
1
13
0.0128
1.3
*Source:  Seckman,  et al.  1977
TABLE 8

   ancer,  by Site*
                               C-77

-------
 increased  from  0 to  25  aberrations per 100 cells.   In  the control  material
 10 aberrations were found  by  these investigators  in 1,012 mitoses.  Thus, it
 was  found  that  the  freauency  of  aberrations  was  significantly  higher
 (p<0.001)  among  the arsenic-exposed workers.   The  three  types  of aberrations
 observed  in  this study,, gaps  (p<0.001), chromatid  aberrations  (p<0.01),  and
 chromosome aberrations (p<0.001) were significantly increased.
    Paton  and Allison  (1972)  investigated  the effect  of  sodium  arsenate,
 sodium  arsenite,  and  acetylarsan  on chromosomes in cultures of  human  leuko-
 cytes and  diploid fibroblasts.  Subtoxic  dose  of  the arsenicals  were  added
 to leukocyte and  fibroblast cultures at  various times  between  2  and  48  hours
 before  fixation.   In  leukocyte  cultures  treated  with  sodium  arsenate  at
 0.29-1.8 x  10~^M for the  last 48  hours of the culture period,  60  percent
 of 148 metaphases examined were found to have  chromatid  breaks.   No  signifi-
 cant breaks  were found  in cultures treated  with  sodium arsenate at  0.58 x
 10~^M,  the highest  nontoxic  concentration.   However, treatment with  acety-
 larsan  at  6.0 x   10~^M  resulted  in 20 percent  chromatid breaks  in  50  meta-
 phases  examined.   Sodium   arsenate  caused  chromosomal  damage  in   diploid
 fibroblasts  to  which sodium  arsenite  {0.29-5.8 x  IQ^M)  was  added to  the
medium for the last  24  hours of culture;  chromatid  breaks  were found  in 20
 percent  of 459  metaphases  examined.   These results  supported  the j_n  vitro
observations of Petres,  et al. (1972)  and Petres and Hundeiker  (1968).
Carcinogenicity
    The  case  for the  association  of  inorganic arsenic  with  skin  and  lung
cancer,  as well  as  other visceral  carcinomas,  has  been  extensively  reviewed
 (IARC,   1973;  NIOSH,  1975;  Hernberg,  1977;  and others).   The most  salient
 points concerning pertinent literature and  reviews  are evaluated  below.
                                     C-78

-------
    The clinical  association  of skin cancer with  the  oral  administration of
arsenic compounds  began with  a report  by  Hutchinson  (1888).   He  described
six patients  in whom  skin  cancer  occurred  and  who had  suffered for very long
periods from  diseases  o^  the  skin (five with psoriasis,  one  with pemphigus)
typified  by  multiple  lesions.  Multiple lesions  occurred even  when sauamous
cancers arose in keratoses; there was an average of two lesions per case.
    The elapsed  time  from the  beginning of administration  of  the  arsenical
drug to the  beginning  of  the  epitheliomatous growth was  variable,  but  aver-
aged 13 years,  regardless of  the  type of  lesion.   In cases  with keratosis,
the latent period to the onset  of keratosis  was  about  half  the  latent period
to the onset  of  the epithelioma,  i.e.,  about 9  years.   In  spite of the long
induction   period,   arsenic-related  skin cancers  started when  the  patients
were relatively young,  33  percent when  they were  40 or  younger,  and  70 per-
cent when  they were 50 or younger.
    Of the  143  patients,  13  had or  developed miscellaneous cancers at other
sites,  but such cases  were not  reported systematically;  the reports commonly
presented  one or a  few case histories.   For example,  Regelson,  et al. (1968)
reported a case of  hemangioendothelial  sarcoma  of the  liver in  a 49-year-old
man who had  taken  Fowler's solution intermittently for  17 years to  control
osoriasis.
    There  have been numerous  reports of arsenic-induced  occupational  cancer,
such as those of the  excess  lung-cancer mortality among Southern  Rhodesian
miners  of gold-bearing  ores  containing large  amounts  of arsenic (Osburn,
1957),  and of the occurrence  of lung and liver  cancer and  clinical arsenism
among  German vineyard workers  exposed  to arsenic-containing   insecticides
(Braun, 1958; Roth,  1957,  1958).   The association of  cancer  with   a  high
degree of  arsenic  exposure has often  been based  on the existence  of oalmar
and plantar  keratoses  (Sommers  and  McManus, 1953).  However,  because of  the
                                     C-79

-------
 increased concentration of  arsenic  in  the lesions  of Bowen's disease, arsen-
 ic  has  been  considered as  a  possible  cause  of the  disease  and  accompanying
 visceral  tumors  (Graham,   et  al.  1961), without  overt  prior  exposure  to
 arsenicals.
    A  number of  relatively quantitative  studies  of cancer  attributable  to
 occupational exposure to arsenicals exists as discussed in this document.
    A death-record examination was made  of a  British plant that  manufactured
 sodium  arsenite sheep  dip  (Hill  and  Faning,  1948;  Perry,  et  al.  1948).   The
 factory was  in  a  small  country  town  within a specific birth and  death regis-
 tration subdistrict.   In  this and adjacent  subdistHcts,  death  certificates
 of  75 workers and  1,216 men (not factory workers)  in three other occupation-
 al  groups  were  obtained for  the period   1910-1943.   Of  the 75  deaths  among
 factory workers,  22  (29  percent) were  due   to  cancer;   of  the  other  1,216
 deaths, 157  (13  percent)  were due to  cancer.   The proportion of  deaths  due
 to  cancer was even higher  among  men  who  actually worked  with the manufacture
 and packaging of  the arsenic-containing  material:   16 of  the 31 deaths  of
men so  classified were due  to  cancer.   The  number  of deaths due to cancer
 according to  site for the  two  groups  is  shown  in  Table  8,  in which  those
 deaths  are  expressed as  a  ~>action of  cancer deaths and as a  fraction  of
 total  deaths.   The absolute  numbers  of  deaths  and  the  fraction of  cancer
deaths are from  the  author's paper;  the  fractions  of total  deaths were  cal-
culated for this  report.  The data  suggest a relative excess  in  the factory
workers of cancers of  the  respiratory  system and skin, whether calculated on
 the basis of cancer deaths  or of total deaths;  the corresponding deficits in
 cancers of the  digestive  organs  and peritoneum disappear  when calculated on
 the basis of total deaths.
                                     C-80

-------
                                                 TABLE 8

                                      Death Due to Cancer, by Site*
No. Cancer Deaths

Buccal cavity and
pharynx
Digestive organs and
peritoneum
Respiratory organs
Genitourinary
Skin
Other or unspecified
Total
Factory
Workers
2
5
7
2
3
_3
22
Other 3
Occupational
Groups
10
91
25
13
2
16
157
Fraction of
Cancer Deaths, Xa
Other 3
Factory Occupational
Workers Groups
9.1
22.7
31.8
9.1
13.6
13.6
99.9
6.4
58.0
15.9
8.3
1.3
10.2
100.1
Fraction of
Total Deaths, %«
Other 3
Factory Occupational
Workers Groups
2.7
6.7
9.3
2.7
4.0
4.0
29.4
0.8
7.5
2.1
1. 1
0.2
1.3
13.0
'Source:  Hill and Faninqt 1948
dThere were 75 deaths among the factory workers and 1,216 deaths in the other three  occupational
 groups (see text).
                                                 C-81

-------
    Although Hill  and  Faning  (19*8)  stated that the numbers of cancer deaths
 are  small,  they concluded  that  "there is a  suggestion  in the  figures  that
 the  factory workers have  been especially affected  in the lung  and skin."
 Hence,  there was  an  investigation  of the  environmental  conditions  at  the
 factory and  the  clinical condition  of the workers in question, compared with
 employees  in other branches of the  factory  who were not  exposed  to arsenic
 (Perry, et  al.  1948),   The  median  air arsenic content for the chemical work-
 ers  at  the  various  operations ranged  from  254  to  696  ug/mr.  As  an  upper
 limit,  this  was  stated to  represent  the  inhalation  of about  1 g  of arsenic
 per  year.   This amount  of  arsenic  is  roughly  equivalent to  the  amount  re-
 ceived by patients using arsenic medication for skin diseases.
    The excretion  of arsenic  in the urine of  127 current  employees  was  de-
 termined; the  scatter  of these values  was  very wide.  Some  exposed workers
 excreted  from  1  to  nearly  2  mg/day,  whereas  many  excreted  less  than  100
 ug/day.  A  few  of  the  persons in the control  group had  very  high excretion
 rates, for  which  the authors found  no  explanation.   It  is important to  note
 that 20 of the 31  factory workers had been exposed  to airborne sodium arsen-
 ite for more than  20 years, and five of  them for 40-50 years.  Furthermore,
 the median  age  of  the  31 exposed workers  was 52 years, and the  average  age
was  50.   None  of these  men's  lungs had pathologic  signs attributable  to
 their exposure to  sodium arsenite  (radiographs  were  made, and vital  capacity
 and exercise capacity were measured).
    The mortality  experience  of  8,047 white male smelter  workers  exposed  to
 arsenic tr1oxide  during 1938-1963  was compared  by  Lee and  Fraumani  (1969)
with  that  of  the white  male population  in the same state.   There was  a
 threefold excess total mortality from respiratory cancer  in smelter  workers,
 and  this  reached  an  eightfold  excess for  employees working  more   than  15
                                     C-82

-------
years  and heavily exposed  to  arsenic.   When respiratory  cancer  deaths were
grouped  according  to degree of arsenic  exposure,  the  observed mortality was
significantly  higher than expected in  all  three  groups:   approximately 6.7,
4.8, and  2.4 times the  expected mortality  in  the  heavy-,  medium-,  and light -
exposure  groups,  respectively.   In  addition to  arsenic tHoxide dust, smelt-
er  workers  were concurrently exposed to sulfur dioxide.   Exposure  to silica
and  ferromanganese and  lead dusts occurred in parts of  the refineries where
arsenic  concentrations  were  low.   Therefore, a  similar  classification  was
made for  relative  sulfur  dioxide  exposure.  Respiratory
-------
by  Milham  and Strong  (1974).   They criticized the methods  of  the  Pinto and
Bennett  study.   The  records  of  workers  from the  same  plant revealed  40
deaths from  lung cancer,  which  was  significantly  higher  than the 18 expected
on  the basis  of rates in  the  general  U.S.  population.   Recent  data  on  mor-
tality experience of  arsenic exposed workers  by Pinto, et  al.  (1977)  is  pre
sented in Table 9.
    Snegireff and Lombard (1951)  made  a statistical study of  cancer  mortal-
ity in a metallurgic plant (A)  in which  arsenic was  handled  and in  a  control
plant  (Z)  in which  "working conditions  approximate those  of Plant A except
that no arsenic is handled."  From  1922  to  1949,  there were  146 deaths among
the employees  of  Plant A  who  handled  large  quantities  of  arsenic  trioxide.
Of  these deaths,  18  were due to  cancer,  including  seven cases  of  cancer  of
the respiratory system.   In  the control  plant,  12 of 109 deaths between  1941
and 1949 were due to  cancer,  including six due to  lung  cancer.  The  authors
stated that  total  cancer mortality in the two plants was  not  significantly
different from the figures for  the  state as a whole,  and they  concluded  that
handling of arsenic trioxide in the industry  studied  does  not  produce a  sig-
nificant change  in  cancer  mortality  of the  plant  employees.   However,  as
pointed out  by the  National  Institute  for  Occupational  Safety and  Health
(NIOSH, 1975), there  are  a number  of  deficiencies  in the  report.   Specifi-
cally,  reanaJyses of  the  data  have revealed  that actually  there was  a large
excess  (approximately  fivefold) of  lung-cancer deaths relative  to  mortality
from all causes  among workers  in both plants.   Thus, the  data demonstrated
evidence of  a carcinogen  for  the  respiratory system among the workers  of
both the plant in which arsenic trioxide was handled and  the control plant.
    Findings of  increased risk of  lung  cancer among copper-smelter  workers
are not  limited  to  the United  States.  A  retrospective study  by  Kuratsune,
et al.  (1974) in Japan revealed that,   of 19 males who died  of  lung  cancer  in
                                     C-84

-------
                                                                       TABLE 9

                            Observed Ooatlts  and Standardised Mortality Ratios  (SHH) at Ages 66 and Over  fur the Period
                             January 1.1949. through December  31. 19/3. Mong 5JO Hen  Ketlrlng fro* the Tacoaa  Swelter
                                           by Cause of Death and Arsenic Exposure  Index  at  Retirement 4. >>
Arsenic Exposure lndexc

Cause of Death

All causes
Cancer (140-205)
Digestive (150-159)
Respiratory (160-164)
ty*phatic (200-203. 205)
Urinary (I80.IUI)
Oilier cancer
Stroke (330-334)
Heart disease (400-443)
foroiiary (420)
Other heart disease
Respiratory disease
(41« 493, 500-602)
AM oliv«r causes

Total
Under
3.000
3.000-5.999
6.000-8.999
9.000 11.999
I2.000*

Obs
324
69
20
32
2
3
12
43
144
120
24

II
57
SHR
.12.21
148.90
122.0
304. ft1*
95.2
90.9
84. 5
113.2
108.8
108.9
108.6

101. 8
92.2
Obs
87
15
6
5
1

3
18
33
25
8

4
"
SMR
98.1
107.9
121.2
165.6
166.1

56.3
150.3
81.4
74.9
111.5

116.8
89.2
Obs
124
28
9
11
1

7
12
63
52
II

3
18
SMR
110.3
156. Of1
140.4
279. 4*1
126.1

102.8
80.0
122.4
122.2
123.2

70.9
74.8
Obs
70
14
2
7
0

5
J
36
31
5

1
12
SHI
129.2J
151.6
62.7
306.90
0.0

150.4
104.5
I44.241
145.10
138.4

52.4
104.1
Obs
24
,
1
4
0

0
6
5
4
1

2
4
SMR
117.7
2IU.7
264.7
668. !>0
0.0

0.0
218.4
53.6
SI. 7
62.9

?!>O.U
91. 1
Obs
17
S
0
5
0

0
I
.,
5
0

0
6
SHH
130.0
217.2
0.0
810. 5J
0.0

0.0
64.7
BJ.2
95.5
0.0

0.0
212.4
4'.niiri .•:  Pinto,  et  al.
I'l •!•'•' ''•«! 'li->!<«< ct^uisuo' tiwU'i ilei ivfil I KM Hart jer.  el al.
•'•.I jl isl it .IMI'>).
                                                                    C-85

-------
 a  particular  town,  11  had  been employed as smelter workers in a local copper
 refinery, and in all cases  the  disease  Had become  manifest after the men had
 stopped working  at  the  refinery.   The author's  conclusion  was that prolonged
 sxoosure to arsenic, and oossibly  also  other  compounds,  seemed  to  be associ-
 ated with cancer of the  lings.   Additional groups  exposed  to  inorganic arse
 nic such as gold miners  in Rhodesia  (Osburn,  1969), hard-rock miners  in the
 United  States (Wagoner, et al.  1963),  and  nickel  refinery  workers  (Rock-
 stroh, 1959)  have shown an  increased  mortality  from  lung cancer, but evalua-
 tion  of  the role of  arsenic  is difficult  because  of  the  presence  of other
 carcinogens in the working atmosphere.
    A  study  at  the Dow Chemical  Company  examined the incidence of  respira
 tory cancer among  173  descendents  who were exposed primarily to lead arsen
 ate and calcium  arsenate  «>nd  1,809 descendents who worked in  the  same plant
 and were not  exposed to those  compounds  (Ott, et al. 1974).  Data were pre-
 sented on the relationship  between  cumulative arsenic exposure  and the ratio
of observed  to  expected deaths  from  lung  cancer.   The  average exposure  of
each worker  was  calculated on the basis  of  records  of  job  assignments and
data  on  the  arsenic content  of  the air  in  various  parts  of the  plant.
Deaths from  respiratory malignancy were  seven  times  greater than  expected
for total  inhaled quantitites  of  29.8 g and  2-4 times greater  for 0.13-6.56
g.   There  was no  association  between  the extent  of exposure  and  the  time
from beginning of exposure  to  death;  most of the  respiratory cancers occur-
red 20-40 years  after initial  exposure,  regardless  of total exposure.
    The ratio of observed  to  expected  deaths  was even  higher (3.85:1)  in
another category,   malignant  neoplasms  of the  lymphatic  and  hematopoietic
tissues except leukemia, than  it  was  in malignant neoplasms of  the  respira-
                                     C-86

-------
 tory system (3.45:1).  Six lymphomas were reported, with the following  diag-
 noses  on the  death  certificates:  four  cases of  Hodgkin's disease,  one of
 lymphoblastoma,  and one  of reticulum cell  sarcoma.
     By contrast with  the  Dow Chemical Company workers,  orchard  workers who
 sprayed  lead  arsenate were reported as showing no evidence of increased can-
 cer  (Nelson,  et al.  1973).   A mortality  study  involving a cohort  of  1,231
 morbidity  survey of  the  effects  of  exposure  to  lead  arsenate  insecticide
 scray  was  conducted   in  1968-1969.   Air  concentrations  of  arsenic  during
 spraying  averaged  0.14 mg/m  .  The population was grouped according  to ex-
 posure  in  three categories  and compared  in terms of  standardized  mortality
 ratios with  the mortality experience  of the  state  of  Washington.   There was
 no evidence  of increased  mortality from  cancer,  heart  disease,  or vascular
 lesions.
     In  1974,   the  mortality  experience  of  retired  employees  of   an  Allied
 Chemical Company pesticide plant  in Baltimore was analyzed  (Baetjer,  et al.
 1975; NAS, 1977a).  The  employees  had  been exposed to  a number of industrial
 chemicals, including  arsenicals;  there were no data on the extent  of  expo-
 sure to  the various  chemicals.   Incidence of death  among the retirees  was
 3.5  times that  among  the  general  Baltimore population.   The excess  mortality
was  concentrated   in   cancer-caused   deaths   (14   times  the   expected),
particularly respiratory  cancer and lymphatic cancer.    The  noncancer  deaths
were at the expected  rates.   These calculations were based on  a  total  of 22
deaths in men  from all causes during the period 1960-1972.
    Several  human studies  not generally available  were reviewed  in the  Na-
tional  Institute of  Occupational  Safety and Health document on  occupational
exposure to  inorganic arsenic  (NIOSH, 1975), including  unpublished reports
to Kenncott  Copper  Corporation  in  1971  and  1974;  unpublished papers  pre-
sented  at the  Conference  on  Occupational  Carcinogenesis  in New York City on
                                     C-87

-------
March 24=27,  1975;  and  an  evaluation by NIOSH of the study by Nelson, et al.
(1973).   In the  latter  case,  independent sources of information investigated
by  NIOSH  contradicted,  ratisr  than  confirmed,  the rgpcrt by  Nelson,  et al.
(1973).   The  conclusion  drawn  was that  the report  apparently  did  not accur-
ately depict  the cancer incidence of  persons exposed  to  lead  arsenate spray
in the Wenatchee Valley  (NIQSH, 1975),
    High  incidences  of  skin  cancer have been reported  in  several  population
groups exoosed  to  high  concentrations of arsenic  in  drinking  water,  inriijcj-
ing people  in the district of  Reichenstein in  Silesia.  (Geyer.  1898),  Cor-
doba  Province in  Argentina,  (Bergoglio,  1964), and  Taiwan  (Tseng,   et  al.
1968).
    Chronic arsenical poisoning,  including  skin  cancer and a  gangrenous  con-
dition of the hands  and feet  called Blackfoot's  disease,  has occurred  in
several  communities  exposed  to  arsenic in drinking  water.   The  best docu-
mented instance  of such  arsenical  poisoning is   in Taiwan  (veh,  et al. 1968;
veh, 1963, 1973; Tseng,  et al.  1968; Tseng, 1977).  In a house to-house  sur-
vey of  40,421 people  in 37  villages along the  southwest coast  of  Taiwan,
Tseng, et al.  (1968) found that  the  prevalence  of skin cancer, hyperpigmen-
tation,  and keratosis each correlated with  the  arsenic content of  the water.
The highest range of concentrations  was greater  than  0.6 ppm,  and the lowest
range was less  than  0.29 ppm.  From  a general   survey  of  these inhabitants,
the prevalence  rate  for arsenical  skin cancer  was 10.6  per  1,000 residents
in 1965.  The skin cancer  --ate  for well water containing  <0.60  ppm was  21.4
per 1,000 people,  while at. <0.29  ppm it was  2.6  (Tables  10 and 11;  Figures
1, 2, and 3).
    The sources  of  drinking  water for  these  areas  were deep  artesian wells,
used from around 1910  until  1966,  when a tap  water supply  was  installed.
                                     C-88

-------
                                   TABLE 10

                 Age-specific  and  Sex-soecific  Drevalence  3ate
                               for Skin Cancer*
Age
0-19
20-29
30-39
40-49
50-59
80-60
70*
Total
Male
Per 1,000
_
1.0
9.7
25.9
80.8
124.8
209.6
16.1
Female
Number
0
2
20
40
99
92
__57
310
Per 1,000
__
1.1
1.5
8.0
28.9
57.0
53^8
5.6
Number
0
3
4
16
38
40
_17_
118
Total
Per 1,000
._
1.1
5.0
15.7
53.7
91.9
12L1
10.6
Number
0
5
24
56
137
132
74
428
*Source:  Tseng, 1977
                                     c-89

-------
                                   TABLE 11

               Causes of Death in Patients wit" Skin Cancer and
                       Patients with Blackfoot Disease*
Skin Cancer
Patients
Cause of Death
Cancer
Lung
Skin
Bladder
L i ver
Colon
Kidney
Stomach
Nasal cavity
3one
Uterus
Esoohagus
Miscellaneous
Cardiovascular disease
Gangrene
Cerebrovascular disease
Respiratory disease
Pulmonary tuberculosis
Pneumonia
Others
No.
68
15
15
10
6
5
5
3
2
2
1
—
4
30
7
32
46
10
17
19
*
27.9
6.1
6.1
4.1
2.5
2.0
2.0
1.2
0.8
0.8
0.4
0
1.6
12.3
2.9
13.1
18.9
4.1
7.0
7.8
Blackfoot
Disease
Patients
No.
99
21
12
17
21
3
—
4
5
4
2
4
6
83
70
63
100
41
28
31
%
18.8
4.0
2.3
l'.2
4.0
0.6
0
0.8
0.9
0.8
0.4
0.8
1.1
15.7
13.3
12.0
18.9
7.8
5.3
5.9
General
Pooulation in
Endemic Area
No.
125
21
3
16
17
12
—
13
16
2
6
2
17
87
—
91
231
55
117
67
*
13.1
2.2
0.3
1.7
1.3
1.3
0
1.4
1.7
0.2
0.6
0.2
1.8
9.1
0
9.5
25.1
5.8
12.3
7.0
Disease of
  alimentary tract

Senility
13
12
5.3
4.9
34
22
6.4
4.2
118
 50
12.4
 5.3
Renal disease
Miscellaneous
Unknown
Total
7
13
16
244
2.9
5.3
6.6

21
30
6
528
4.0
5.7
1.1

34
207
—
951
3.6
21.3
0

*Source:  Tseng, 1977
                                     c-90

-------
            200
                                             Molt
                                             Tata I
            100-
                            40   90   «0  TO   «0
                             FIGURE  1

      Age and Sex-soecific - Prevalence Rate for Skin Cancer
                Source:  Adapted from Tseng, 1977.
              Both Sexes
  200
   ISO
   too
 I
looo
         H>  M M
        < — IM T
           10. »
                         M
                 40-
                                             in
                              H  H^n  QjM ft Ov«r
                              M  M*4  030 - 0-59
                              I  L«w  OAO - 0.2*
                              V
Total
                             FIGURE 2

Age-specific Prevalence Rate  (1/1,000)  for Skin Cancer by Arsenic
                   Concentration  in Well  Water

                       Source:  Tsenq, 1977
                               C-9L

-------
><  20 1—
 «
o


u

e
-*

w
                                                              0.3 -  0.6 mg/1
     8  —
     4





     0
                                     FIGURE 3



                        GracMca! Representation of Figure 2
                                      c-92

-------
 The  concentrations of  arsenic  in  the water  ranged  from  0.01  to  1.82  ppm
 (median range was  0.4 + 0.6  ppm).   The chemical  form  of arsenic  in the water
 was  not  clearly determined,  but  it may have  been  either  tn'valent (because
 the  well  water is  probably  anaerobic) or  a  methylated arsine  (because  the
 authors observed a combustible gas, perhaps methane,  bubbling from the water
 storage tanks).   Initial  attempts to measure  the arsenate  to arsenite ratio
 of  the  water have  been confined to measurements in  a  United States  labora-
 tory of a sample shipped  from Taiwan  with  no  special  precautions to preserve
 the  speciation  occurring  at  the  collection  point   (Table  12)  (Irgolic,
 1979).  Therefore,  from  the  best  available information,  people in  that  re-
 gion  of  Taiwan  could  have  been  exposed  to  both  tHvalent  and pentavalent
 arsenic compounds.
    Assessing the  Taiwan  situation  is  more complex  than  simply identifying
 the  two  oxidation   states  of arsenic,  as  suggested recently  by Lu,   et  al.
 (1975, 1977a,b).   These  workers have  observed nonarsenical  fluorescent  com-
 pounds in water  samples  from  the areas where Blackfoot  disease  is  endemic
 and have identified one of the fluorescent components  as  an alkaline  hydro-
 lysate of  ergotamine,  lysergic  acid,  or  a   related  compound  (Lu,  et  al.
 1977b).   They  have also  shown that one  of the fluorescent  components  pro
duced abnormalities in developing chick embryos (Lu,  et  al.  1977a).   It  is
not known whether  ergotamine was the compound that produced  these abnormal-
 ities.
    The evidence of arsenical waters  in  an eastern area of  the province  of
Cordoba,  Argentina, has been known for many  decades  and is  associated  with
 the  occurrence  of   hyperpigmentation,  keratosis, and  skin  cancer.  A  study
made in 1949-1959 indicated  a  higher proportion  of  deaths  from cancer  in  the
 arsenical  region  than in  the  rest  of the  province  — 23.8  percent  versus
                                     C-93

-------
                                   TABLE 12

                Analysis Results for the Taiwan Water Samples*
                                                 Geographic Location
    Element                                  Pei  Men              Pu Tai
Arsenite^, ppm As
Arsenate^, ppm As
Arsenite and Arsenate
total As, ppm (AAS)2
total As, ppm (NAA)3
0.05
0.52
0.57
0.72
0.76
0.09
0.63
0.72
0.76
Sodium, ppm                                 282                    223
Cooper, pom                                  <0.1                   <0,
Manganese, ppm                               <0.1                   <0.
Zinc, ppm                                    <0.1                   <0,
Iron, ppm                                    <0.1                   <0,


•Source:  Irgolic, 1979.
^Determined by GC-MES
2F1ameless Atomic Absorption Spectrometry
^Neutron Activation Analysis
                                       C-94

-------
 15.3  percent  (Bergoglio,  1964),  The excess was  due  mainly to cancer of the
 respiratory and  digestive tracts in  both  men  and women.   The excess cancer
 was unrelated to socioeconomic  differences.
    In  contrast  to the above epidemiology studies yielding  evidence  for in-
 creased  cancer  rates  in  populations exposed  to  arsenic via  drinking  water
 supplies,  a study  conducted  by  Morton,  et  al. (1976) failed  to  demonstrate
 any increased  incidence of cancer  in Lane County, Oregon,  the only  area  in
 the United States  where  the  drinking  water  supply  has  elevated  levels  of
 arsenic.   Several  possible  explanations  can  be  offered  for the   lack  of
 effects seen in  Lane  County  in comparison to  positive observations  in  other
 areas, e.g., Taiwan.   These include the following:
 (1) arsenic concentrations  in  the  Lane  County  drinking  water   supply  were
    distinctly lower than those measured in Taiwan;
 (2) the predominant  form of  arsenic  in  the  Taiwanese  water was  trivalent
    arsenic which  tends  to be more toxic than the pentavalent form found  in
    Lane County;
 (3) an insufficient number of subjects to  reveal  small  increases  in cancer
    rates  may have been  studied in the Lane County area,  an  area much  less
    densely populated than Taiwan and having fewer total numbers  of  subjects
    available  for study;
(4) differences  in  racial characteristics and  nutritional  status  between the
    two  experimental  populations may have affected the results;
(5) the  presence of  other carcinogenic contaminants  in  the Taiwanese  water
    but  not in the  Lane County water  may  have  increased  the cancer rate in-
    dependent  of  the  presence  of arsenic.
                                     C-95

-------
While  none  of the  above explanations can  be  ruled out as  crucial  factors,
neither can  the  observed differences in cancer  rates  be  conclusively attH -
buted to any one of them.
    In  general,  animal  studies  have not  shown  carcinogenicity  for  arsenic
compounds, even when  administered  at  near  the  maximally tolerated dosage for
long  periods.   Certain  notable  exceptions are  described  first,  and  then
several of the negative studies.
    Askanazy  (1927)   noted  benign  and  malignant  teratomas  in   rat  embryos
transplanted into  the peritoneal  cavity  of rats  whose drinking water  con-
tained  arsenic.   Embryonal  cells  are especially sensitive  to arsenic  which
provoked in  them  signs  of degeneration even  in  concentrations of 0.25  ug/1
of cultivation medium (Goeckerman and Wilheim,  1940).
    In 1962, Halver reported the occurrence of hepatomas in  trout fed  a syn-
thetic  diet  containing  carbarsone  at 4.8  mg/g  of  diet  (the data were  re-
viewed by Kraybill  and  Shimkin  (1964);  the original  report is  not  readily
available).   Of  50  trout  exposed  to carbarsone,  five developed hepatomas.
There were  no hepatomas  in a  large control  group fed  the synthetic  diet
without carbarsone.   However,  aflatoxin contamination  of  the diet may  have
been a confounding variable.
    More recently,  Osswald  and  Goerttler  (1971) reported  that  subcutaneous
injections  of  sodium arsenate  in  pregnant  Swiss mice  caused  a  considerable
increase in  the  incidence  of  leukemia in  both  the  mothers and their  off-
spring.  A 0.005 percent aqueous  sodium arsenate solution  was injected  daily
during pregnancy  for a  total  of  20  injections, each  containing  arsenic  at
0.5 mg/kg.   Some  groups of offspring from the  arsenic-treated  females  were
                                     C-96

-------
 given  an  additional  20 subcutaneous  injections of  arsenic  (0.5 Tig/kg)  at
 weekly  intervals.   Leukemia  occurred  in  11 of 24 mothers  (46  percent),  7 of
 34  male offspring (21 percent), 6 of  37  female  offspring  (16  percent),  and,
 in  the  offspring given  the additional  20  injections,  17  of 41  males  (41  per-
 cent) and  24  of  50 females  (48 percent).   Leukemia  developed  in only 3 of 35
 males  (9 oercent) and  in  none of 20  female  offspring of  untreated  control
 mice.   Furthermore,  11  of 19  mice  (58 percent) developed  lymphoma  after 20
 weekly  intravenous injections of 0.5 mg each of arsenic as  sodium arsenate.
    Long-term  studies  of effects of  arsanilic  acid  on  chickens, pigs,  and
 rats were  reoorted by Frost, et  al.  (1962).   No  adverse  effects were seen in
 the chickens  and  pigs after  4  years  of feeding,  nor in pigs fed 0.01 percent
 arsanilic  acid for three  generations.   Male  and  female weanling rats  from
 the F£  generation  of a  six-generation  breeding  study in  which 0.01  percent
 and 0.05 percent arsanilic acid was fed were  held on  the 0.01  percent arsan-
 ilic acid  diet or on  the control diet for 116 weeks.  The overall  tumor in-
 cidence was the same in  all  groups and  resembled the  historical incidence of
 tumors  in  the  colony, 35-45  percent.   The significance of  these data lies in
 the fact  that transplacental exposure to a carcinogen  followed  by  lifetime
 exposure  to  the same carcinogen is  often the most sensitive  technique  for
detecting  carcinogenicity of a  substance   (Tomatis and Mohr, 1973),  but  this
 test was negative.
    Soutwetl   (1963) used female mice  (Rockland  and a specially  bred  strain
highly susceptible to skin tumors) in  a test  for cocarcinogenicity  of potas-
 sium arsenite.   It  was  tested  as an  initiator,  both  orally by stomach  tube
 (a total of 2.4 mg in  five days) and  locally  (a  total  of 1.2 mg in  eight ap-
plications during  five  days).   This  initiating treatment was  followed  by
 topical  application of croton oil twice a  week for  18  weeks.   He also tested
                                     C-97

-------
potassium  arsenite  as  a  promoter  by  daily  applications  (a  total  of  2.3
mg/week)  after a  single  75 ug  dose of dimethylbenzanthracene  (OMBA).   The
prolonged  skin applications  of  potassium  arsenite  were hyperkeratotic  and
ulcerogenic.   Other  experiments  were done  to  determine  whether  arsenic  would
increase  the  yield  of  skin cancers  caused by  a suboptimal regimen  of  DMBA
plus  croton  oil  given  either  at  the  time  OM8A  initiation  or during  the
24-week  period of  croton  oil promotion.   Under the  latter condition,  the
mice were  fed  potassium arsenite at 169 ug/g of food.   This  dietary  concen-
tration of 169 wg/g {as potassium  arsenite)  is  very high,  compared with  the
0.5 ug/g usually found in the human  diet.   In no case was there an  effect of
arsenite on skin  carcinogenesis  in  these  experiments.   Many tumors  developed
in the  positive  control mice,  beginning as early as  six weeks  after treat-
ment began.
    Baroni, et  al. (1963)  carried  out  a  similar study with male and female
Swiss mice,  testing  the  oral effects  of  potassium  arsenite  (100  mg/1  in
drinking water)  as  an initiator  with croton  oil promotion  and  as a promoter
with OMBA  and  urethane  initiation.   Local  skin  applications of  sodium arsen -
ate were  tested  as  a promoter  after initation  with  OMBA  or urethane.   The
arsenicals had no  effect on  tumorigenesis:  and only a very  slight degree of
keratosis was observed.
    Milner (1969) used three strains of mice  that differed  in susceptibility
to the  induction of  skin  tumors by  the application to the skin of  methyl -
cholanthrene-impregnated paraffin disks for 2-3  weeks.   The treated site  was
transplanted syngenetlcally  and  observed  for eight weeks  for  tumor forma-
tion.  Arsenic trioxid« (100 mg/1  in drinking water) was administered either
during  methylcholanthrene  exposure,  to  animals with  transplanted  skin,  or
                                     C-98

-------
 both.  Arsenic  exposure  produced  a  small  increase in the yield of papillomas
 in  the low-susceptibility strain, a small decrease  in  the high-susceptibil-
 ity  strain, and no effect in the  intermediate susceptibility strain.
     Byron,  et al. (1967) fed  either sodium  arsenite  or sodium  arsenate  to
 Osborne-Mendel  rats  in  a  2-year  study at  dietary concentrations  of  15-250
 ug/g  for arsenite  and  30-400  ug/g  arsenate.   No carcinogenic  activity  of
 either material  was found.   These  investigators  also  did a  2-year  arsenic
 feeding  experiment on  dogs,  with  negative results; however, this  was  an  in-
 adequate observation period for studying carcinogenic responses in dogs.
    Hueper  and  Payne  (1962)  incorporated arsenic trioxide  in  the  drinking
water  (either plain  or with 12 percent ethanol)  of  groups  of  rats and mice.
The  initial concentration of 4  mg/1  was  increased by 2 mg/1 each  month to a
maximum of 34 mg/1 at  15 months.  Thus,  the daily intake of arsenic trioxide
ranged from 0.1  to  0.8  mg/rat.   The  administration  of  arsenic  trioxide  was
continued until  24 months.  Neither  the rats  nor the mice developed any can-
cers in suspected target organs — skin,  lung, and liver.
    Kanisawa and Schroeder  (1969) and Schroeder,  et  al.  (1968) found  no car-
cinogenic effects on mice exposed to  potassium arsenite at 5  mg/1  in  drink-
ing water  from  weaning  to  senescence (Kanisawa  and Schroeder, 1969)  or  on
rats on the same regimen (Schroeder,  et  al.  1968).
    Kroes,  et al.  (1974) studied the carcinogeniclty  of lead  arsenate  and
sodium arsenate with SPF-W1star-derived male  and female  rats.   In addition,
some groups were intubated  with a subcarcinogenic dose  of diethylnitrosamine
to investigate  a  possible synergistic action  leading to  lung  tumors.  Food
intake and  body weights were  recorded,  and  complete gross and  microscopic
examinations were made on all  animals.   Lead arsenate  that was incorporated
in the diet at  1,350 ug/g was  toxic and  caused  increased mortality;  an ade-
noma of the renal  cortex and a  bile duct carcinoma were  found  in this  group,
                                     C-99

-------
but  no  significance can be attached  to  one OF two tumors in  any  group.   No
cancer was  associated with  the  feeding  of lead arsenate at 463  ug/1  or sod-
ium  arsenate  at  416 ug/1.   No  synergism  with  the nitrosamines was observed.
There was  a high  spontaneous-tumor  incidence  in this experiment.   The test
diets were  fed to  female rats  from  the time of  parturition until  the  young
were  weaned,  and  these  young were  the  test  animals.   Surviving  rats  were
killed after 29 months of feeding.
    As  Fraumeni  (1975)  has pointed  out,  it  1s  largely because  laboratory
studies have  not  succeeded  in  producing  tumors  in animals that  arsenic  has
not been accepted  universally as  a carcinogen.   There is evidence from  clin-
ical  observations   and  occupational   and  population  studies  that  inorganic
arsenic is  a  skin  carcinogen  in man.  There  1s  a characteristic sequence of
skin  effects  of  chronic  exposure to  arsenic that  Involves  hyperpigmentation
initially,  then hyperkaratosls  (keratosis), and  finally  skin cancer (Yeh, et
als 1968),  This sequence has been  oserved under  a variety of circumstances
involving chronic  exposure:   potassium  arsenlte  (Fowler's  solution) was used
medicinally  (Neubauer.  1947),  vineyard  workers  used sprays   and/or  dusting
powders  containing  arsenic compounds  and drank  arsenic-contaminated  wine
(8raun4 1958: Roth.  1957, 1958).  chemical  workers manufactured sodium arsen-
ite for use as  i sheep dip (Perry,  et  al. 1948), and residents of a south-
west area of Taiwan  had. as their only  source of drinking water for  over 45
years, artesian wells eontaminattd by arsenic  from geologic  deposits (Tseng.
et aU 1968),  The  similarity of  responses under these diverse circumstances
is important*  because  studies in human  populations always  involve variables
that cannot be controlled as  in laboratory t.xn*rimtnts;  htncts the credibil-
ity of  information derived from  human  studies depends  on  the demonstration
of comparable effects under different conditions.   This  requirement has been
amply met regarding arsenic as a cause of skin cancer (NAS, I977i)s
                                     C-100

-------
     The  earliest skin effect  of chronic arsenic exposure, hyperpigmentat ion
 (melanosis),  occurs in  a  dappled pattern predominantly  in  unexposed areas.
 After  the onset of melanosis, the  skin  begins  to  atrophy in  a patchy way in
 hyperpigmented areas,  with the formation of  keratoses that are the pathogon-
 omom'c  lesions of  chronic  arsenic exposure (Yen, et al. 1968).  Only a small
 proportion  of the  keratoses  evolve into  skin  cancer,  and this  takes  place
 only after  many years.   The sequence is  illustrated by the Taiwan data   the
 prevalence  of melanosis,  keratosis,  and  skin  cancer reached 10  percent  in
 the  male population roughly  at ages of 18,  30, and 60  years,  respectively
 (Tseng,  et  al. 1968).   Chronic  exposure to  inorganic arsenic thus  causes  a
 slowly  progressive form of  patchy skin  damage involving the epidermis  and
 adnexal  structures,  as  well as  the underlying  dermis,  with  the  precancerous
 keratoses and  cancers  forming in the areas of  chronic  atrophy.   The chronic
 damage and  tumorigenesis resulting from arsenic are similar  to  the  effects
 of ionizing and ultraviolet radiation on the skin (HAS, 1977a).
    Arsenical  skin  cancer  is  readily distinguished from  skin  cancer  induced
 by sunlight,  in that it occurs  predominantly on surfaces that  are  shielded
 from sunlight  and  multiple lesions  are  much  more  common  in  arsenic-induced
cases;  for  example, in 428 of the  429  cases  of skin cancer  studied  in  Tai-
wan,  there was more than one cancer (veh, et al. 1968).
    Substantial doses of inorganic  arsenic are  required to produce an appre-
ciable  incidence of skin cancer.  The average  intake  of persons  treated  with
Fowler's solution who developed  skin cancer was around  20-30  g.   The preval-
ence of  skin  cancer 1n  Taiwanese men exposed  to  drinking water  containing
arsenic at 300-600 mg/1 was about  15  percent  at age 60 and over.  The normal
 incidence is  2-3  percent.   On the  basis of a  2 I/day water  intake  for  the
period  over which the artesian wells were  used  (45  years), the total  arsenic
                                    C-101

-------
 intake  must have  been  about 15  g,  which  is  roughly  in the same  domain as
 that  in clinical cases  of  the use of Fowler's  solution.   Thus,  the Taiwanese
 data  that  demonstrated  the requirement for large doses  of  arsenic to obtain
 even  a  modest yield of  skin cancer are  consistent  with the relatively  low
 frequency  of skin cancer  in patients  treated  with Fowler's solution.   The
 low potency of  inorganic arsenic may explain why  no skin  effects  have  been
 reported in  people treated  for  syphilis with  organic arsenicals,  inasmuch as
 the total  doses  amounted to only a few grams.  However,  It  is  also possible
 that  the metabolism  of the  organic  arsenicals  is  sufficiently different to
 preclude the occurrence of skin  cancer and other forms  of  arsenical  damage
 even  at higher doses (MAS,  1977a).
    The  relative  frequency of  melanosis, keratosis,  and  skin  cancer  was
 roughly  similar  in  the Taiwanese  population  and  the  chemical  workers  who
manufactured  sheep dip.   On direct examination, the latter  showed  a 90  per-
 cent  prevalence  of melanosis and  30  percent  prevalence of  keratosis, for  a
 ratio of melanosis to keratosis  of  3:1.  At comparable  ages,  the Taiwanese
 showed  a  ratio  of about 4:1.   Two  of  the nine keratosis  patients  in  the
 sheep dip  factory  had  already been treated for  skin cancer, and  the propor-
 tionality  between  keratosis and  skin  cancer  was about  the same  in Taiwan.
As in the  Taiwan experience,  the sheep  dip chemical workers had been exposed
to large doses of  Inorganic arsenic  (up  to 1  g/year),  but much of  this  was
by inhalation (MAS, 1977a).
    It  1s  possible  that  the trivalent and pentavalent  forms  of  inorganic
 arsenic produce  the same effects  on  skin.   This  is  of  interest,  particularly
 in view of  the different metabolic patterns of  trivalent and pentavalent in-
organic  arsenic  with the  former by  interaction  with  sulfhydryl  groups  and
the  latter  by  substituting  for  phosphate.  The  clinical  use  of  Fowler's
                                     C-102

-------
 solution  and the manufacture of sodium arsenite as a sheep  dip both  involved
 exposure  to trivalent  inorganic  arsenic.   The two categories  of people de-
 veloped  similar skin responses.  The Rhodesian gold  miners, in whom the in-
 cidence of  typical  arsenical  keratoses  was very high, were exposed to arsen-
 opyrlte,  1n which the arsenic becomes trivalent on weathering; the reactions
 of  arsanopyrite in  the  body are unknown (NAS,  1977a).   The  chemical  form of
 arsenic  in  the Taiwanese  artesian-well water is  still  being investigated,
 but,  the  reported occurrence of methane gas 1n the water could preclude the
 existence  of arsenic 1n  the  pentavalent form  (MAS,  1977a)  and certain pre
 Umlnary  results  by Irgollc (1979)  suggest that the trivalent form of arsen
 1c  predominates.  The failure to find  Increased  incidence of cancer in Lane
 County, Oregon,  where pentavalent,  inorganic  arsenic  tends  to predominate in
 water  supplies  lends some support to the  possibility that  trivalent arsenic
 has the greatest  carcinogenic potential and 1s of the greatest concern.
    Of the  published reports on mortality from respiratory  cancer in copper
 smelters, the most  Impressive is  that of Lee  and  Fraumenl (1969).  The study
 involved  a  population of 8,047 white male smelter  workers who were followed
 for 26 years;  for each  employee,  Information  was available  on  time, place,
 and  duration of  employment, maximal  arsenic  and  sulfur dioxide exposures
 (descriptive,  rather than  numerical),  and cause  of  death.    The life-table
method was  used  to evaluate age-specific mortality rates  for  the  various
causes of  death,  and the rates were compared with  those of the  states  in
which  the  smtlters  were.   The  number  of  deaths  available  for  analysis  was
very  substantial; 1,877.   The study demonstrated  a systematic  gradient  for
respiratory  cancer  according to  the  magnitude and duration of  exposure  to
both  arsenic  and sulfur  dioxide.   These  agents,  however,   were  inseparably
linked, because of  the  nature of the smelter  operations. The amount of ex-
cess cancer  was impressive, with  an  eightfold increase  in   the  workers  who
                                    C-103

-------
had  the  heaviest  arsenic exposure for  the  longest  duration,  i.e., more than
15  years.   The  latent  period -- the interval  between  first  employment  and
death  from respiratory cancer --was extraordinarily long  and  was inversely
related to the magnitude of  exposure:   34,  39,  and  41 years for the categor-
ies  of heavy, medium, and  light arsenic exposure.  There  were deficiencies
in  the study, some of  which were unavoidable.   For example,  no  indication
was  given  of whether  the  study  population  was representative of  the  total
workers population; the  exposure rankings were based on  the  maximal  arsenic
concentrations,  rather  than weighted  averages  derived  from  work  histories.
No  quantitative  data  were  available on  exposure.   No  attempt  was made  to
validate the  stated causes  of  death.   No  smoking histories were  obtained.
However, none of  these deficiencies  could be seriously  regarded as invalida-
ting the conclusions of the study (MAS, 1977a).
    The Kuratsune, et al. (1974)  report  dealt with  a smaller  study that com-
pared  lung-cancer mortality  rates calculated from the  22 deaths  that  occur-
red in a 30-year period in a smelter  town with  the  lung cancer experience  in
the  same period  in a  neighboring city and  in  Japan  as a  whole.   The  stan-
dardized mortality rate for males in  the  smelter  towns  was  four times  higher
than that  for the rest  of  the  country,  but equal  to   that for  women.   This
4-fold excess is comparable with  the  3.3-fold excess observed in  the  Lee and
Fraumeni  (1969)  study.   Although many  of the men 1n the town  worked  in  the
refinery,   a much  higher proportion  of  the  lung-cancer cases,  compared with
controls,   occurred  in men  who  were  heavily exposed to  arsenic  as  smelter
operators.   As 1n the case  of the Lee and Fraumeni  (1969)  study,  the  latent
period from  first exposure  to  the diagnosis of  lung  cancer was  very  long,
ranging from 26 to 48 years.  The duration  of employment  was  also very long,
with a median of about  30  years, although  two cases occurred  in  people who
worked for only 2-3 years.
                                     C-104

-------
     Two  lung cancer  studies  of the  American  Smelting and  Refining Company
 smelter  have produced conflicting  results.  The  1963 Pinto  and  Bennett re-
 port examined  the proportional  mortality from lung  cancer  in a  total  of 229
 deaths  in  the  period 1946-1960.  This  study dealt only with  pensioners and
 workers  who  died during their employment and did  not include people who had
 left the plant.   The reported data  showed that the  18  lung-cancer  deaths in
 the  plant  population as  a whole was  higher than  the  rate  in the  state of
 Washington.  However,  the excess lung  cancer  for the plant  as  a  whole was
 due  to the high  occurrence in controls,  i.e.,  in  workers  who were considered
 not  to have  arsenic exposure.  Milham  and  Strong (1974)  by  contrast,  found
 in the years 1950-1971,  that  there  were records  of 39 deaths due to respira
 tory cancer  in Pierce County  (the  smelter locale) in people  who  were  stated
 to have  worked  at  the  smelter.   Application of U.S. mortality rates  to the
 published  figures  for  the smelter  population  at  risk  yielded an  expected
 number of 18 respiratory  cancer deaths,  compared  with the  39 deaths observed
 (MAS, 1977a).
     Pinto,  et al.  (1977)  recently resolved  the  discrepancy  between  the Pinto
 and  Bennett  (1963) and Milham  and  Strong  (1974)  papers  in  a study  of the
same smelter that  Devaluated the exposure categories used  in the  Pinto and
Bennett  paper  (1963)  (which were apparently in error) and also included  a
longer observation  period and  therefore more  deaths.  The data included  a
total of 32 respiratory  cancer cases  and  show  a  progressive  increase  in
standardized mortality ratio  with increasing arsenic exposure.  The arsenic-
exposure index  was  calculated as a  weighted  average  based on urinary arsenic
concentration and duration of employment.   It  is  of interest that the eight-
fold excess in respiratory cancer for workers with  the  highest exposures and
                                     C-105

-------
 the  threefold excess  for  all  the  smelter  workers  reported by Pinto  et  al.
 (1977)  were very close  to  the figures reported  by  Lee and  Fraumeni  (1969)
 and Kuratsune, et al.  (1974).
    The  studies  described  here indicated that  excess  respiratory  cancer  oc-
 curs  in  copper-smelter workers as  a  function  of the  magnitude and  duration
 of  exposure to  arsenic, with  latent  periods  of  three to  four decades  from
 the time of  initial  exposure.  However,  the studies  do not permit  a  conclus-
 ive resolution of  the issue of whether concomitant  exposure  to  sulfur diox-
 ide and other  smelter  dusts  is necessary  for the carciogenic  response.  Evi-
 dence  from studies  involving  entirely different  circumstances  of  exposure
 including workers  in three  pesticide manufacturing plants  (Hill  and  Faning,
 1948; Ott,  et  al.  1974), vintners who applied  pesticides  (Braun,  1958),  and
 Rhodesian gold miners  (Osburn, 1969), however, suggests that  sulfur  dioxide
 and  other   unspecified  smelter dusts  are  not essential  cofactors  for  the
 respiratory  carcinogenicity of  arsenic.   All  the  nonsmelter  studies  had
obvious limitations, but the lung  cancer  excess 1n  each study was  relatively
 large and,  taken as  a  group, they provide  significant  evidence that  arsenic
 is a lung carcinogen (MAS,  1977a).
    The Hill  and Faning (1948)  study of  75 deaths in a  sheep dip  factory
used  the  indirect method  of  proportional  mortality  to  evaluate  the  small
group of  22 deaths  from cancer;  seven of them were cancers  of the  respira
tory tract, compared with  an  expected 2.4  deaths.   The Dow  arsenic  workers
 (Ott, et  al.  1974)  were evaluated  in  two ways: (1) by an  analysis  of death
 records of  those who  died  from  lung cancer  (28,  or  16.2 percent,   of  173
 chemical-worker deaths, compared with  104,  or  5.7  percent,  of the  1,809 con-
 trol-case deaths), and  (2)  then,  as  a retrospective  cohort study,  a  compari-
 son of the  mortality from respiratory cancer (obtained from the  records used
                                    C-106

-------
 in  the  first approach) among 603  persons  identified  as having worked in the
 arsenic  plant  from 1940 to  1973  with the mortality  among  the corresponding
 U.S.  white  population.   The two  approaches  gave  essentially the  same  re-
 sults  -  a  threefold  to fourfold  excess.   However,  the puzzling aspect of the
 data  is  that almost  60 percent  of the respiratory-cancer deaths were in peo
 pie who  had  worked with arsenic for  less  than  a year, three decades earlier.
 Most  of  the arsenic workers  were unskilled short-term  employees,  of whom  a
 large proportion  left  the company after  a brief period of  employment.   The
 follow-up  study,  however,  dealt  only with  the people  who remained  in  the
 company.  A  confirmation of  the  excess  lung cancer in  a follow-up  of short-
 term  arsenic workers who  left  the company would be  very  useful.   Neverthe
 less, there were about a dozen cases  in people  who  worked  longer  than a year
 and who  were in the  highest  dose  categories,  where  the excess risk  was maxi-
 mal (fourfold to  sixfold).   It  is possible that the  apparent  twofold excess
 in  lung  cancer  in the lower exposure categories,  including  those who worked
 with  arsenic for  less  than a year, would  not  be ascribable to arsenic,  be-
 cause there  was  no change  in  cancer risk over a  wide  range  of  total  doses
 (0.04-1.56 g).  Furthermore,  these  low  dose categories consisted  predomin-
 antly of short-term  unskilled workers who as  a group might have had higher
exposures to other hazardous chemicals than the controls (MAS, 1977a).
    The  Allied  Chemical  Company  pesticide manufacturing operations produced
 a  range  of  products,   including   some  arsenical compounds.   A  preliminary
study of the proportional  mortality among retired  employees  showed a seven-
fold  excess  of  lung  cancer  that  accounted  for   about  40 percent of  all
deaths.  Both  the Dow and Allied  studies also showed  a  few  excess deaths
from  lymphoma and  Hodgkin's  disease.  The  results  of a more  detailed study
of the Allied  Chemical Company that  is  now in progress will  be  very useful
 (MAS,  1977a).
                                     C-107

-------
     Arsenic  sprays and  dusts  were widely  used in Germany  between  1925 and
 1942,  at which  time  they were  banned  (Braun, 1958; Roth, 1957, 1958).  Vine-
 yard workers also drank wine containing  arsenic.   Hundreds  of  workers  de-
 veloped  acute and chronic arsenic poisoning.   In  the  1950's,  vineyard work-
 ers  with  lung  cancer  began  to appear in hospitals serving  the vineyard re-
 gions.  An  association between arsenic and  lung cancer  is  further suggested
 by  the high  proportion of  vineyard  workers  with lung  cancer who  had  the
 characteristic  hyperpigmentation  and  keratoses  associated with chronic arse-
 nic  exposure  (MAS, 1977a).
     The same  high  degree of  association  of  skin arsenism and  lung cancer oc-
 curred  in  Rhodesian  gold miners  who were  heavily  exposed to arsenopyrite
 dust (Osburn, 1969).   In the period  1957-1963,  the occurrence  of 37 cases of
 lung cancer  in  gold  miners  represented an  incidence of  206/100,000 compared
 with 34/100,000 for  adult males  in the Gwanda  region  of Rhodesia.  This re-
 presents a sixfold difference in  lung cancer in miners  (MAS, 1977a).
    The  probability  of  death  from  lung  cancer in  persons with  keratosis,
 ranges from 32  to  56  percent,  which  is roughly 5-10 times  higher  than might
 be expected.  The  data  suggest  that  there  is a very high risk  of lung cancer
when the exposure to  inorganic arsenic dust  is  high enough  to  cause keratos-
es (MAS, 1977a).
    The only  evidence that  arsenic  is  a  liver carcinogen  comes  from German
vintners.  Thirteen of  the 47  persons whose autopsies were  reported  by Roth
 (1957,  1958)  had cirrhosis,  and  six  had  angiosarcoma, a rare form  of liver
cancer associated  with  exposure  to vinyl  chloride and  Thorotrasr*,  and  one
with a duct carcinoma.   Only two  cases of angiosarcoma have  been reported in
people treated  with  Fowler's solution (Regelson,  et al.  1968).   There is no
evidence of either cirrhosis or  liver damage in any of  the  other  studies on
                                     C-108

-------
arsenic.   It  is  possible that the  combined  effect  of a high alcohol  intake
and arsenic  is responsible for the unusual forms of cirrhosis and  liver can-
cer observed  in  vintners.   It should also  be  pointed out that the chemical
form of arsenic in wine is  unknown (NAS,  1977a).
                                    C-109

-------
                             CRITERION FORMULATION
 Existing Guidelines and Standards
     In  1942,  the  U.S.  Public Health Service set a maximum allowable level of
 50  ug/Hter  for  arsenic  in  drinking water  supplied  by  interstate  Carrier
 Water Supplies.   The  arsenic standard remained  at that level  in the 1962 re-
 vision  of  the Drinking Water  Standards  and has  been  continued in the  U.S.
 Environmental  Protection  Agency Drinking  Water  Standards  which became effec-
 tive in June  of 1977.
    The  American   Conference   of  Governmental  and  Industrial   Hygienists
 (ACGIH, 1977) has set 0.5  mg/m3 as  the  Threshold  Limit  Value-Time Weighted
 Average (TLV-TWA)  for airborne  arsenic.   This  means  that  the time-weighted
 average concentration  of airborne  arsenic  for a  normal  8-hour  workday  or
 40-hour workweek  should  not exceed  0.5  mg/m3.   The Conference has  issued a
 Notice  of  Intended Change  (for  1977} which to reduce the TLV-TWA from 0.5
 mg/m3 to 0.05 mg/m3 (ACGIH,   1977).
    The National  Institute  of  Occupational  Safety and  Health  has  recommended
 a ceiling  level  of 2  vg/m3 for  airborne  inorganic arsenic for any  15  min-
 ute period of the workday.
    The new  (August,  1978)  Occupational   Safety  and  Health  Administration
 (OSHA) standard for airborne inorganic arsenic  is 10 wg/m3 TWA.
 Current Levels of Exposure
    A broad  range of  arsenic  levels  have been found  in  drinking  water  sam-
 ples.   In a  U.S.  Environmenal  Protection Agency  national  study of resident-
 ial   tap  water, 66.8  percent  of the  one  time grab samples   collected  from
 3,834 residences  had  arsenic  levels  greater   than  0.1  ug/1.  The  average,
minimum,  and  maximum  levels  of the  samples   with  detectable arsenic  were
 2.37, 0.05,  and  213.6 ug/1,  respectively  (Greathouse  and  Craun,  1978).   In
                                     C-110

-------
 1975  it  was  reported that  5 out  of  566 samples  collected  from  Interstate
 Carrier Water Supplies exceeded  10 ug/1 and  that  the maximum  level  was 60
 ug/liter  (U.S. EPA,  1975).   Well  water samples  collected  during 1976 at 59
 residences  in  a  Fairbanks,  Alaska  suburban  community had a mean  arsenic con-
 tent  of 224 yg/liter with, a range from  1.0 to 2,450 ug/1  (U.S. Public Health
 Service,  1977).   Moderately elevated levels of arsenic,  10 to 330 ug/1, are
 present in  potable  waters  of some  smaller communities  in  Nevada  and Cali-
 fornia  (Valentine,  1979).   There have  been  a  number of other reports of iso-
 lated  instances  of  higher than usual concentration  of arsenic in well waters
 (Goldsmith, et al.  1972;  Feinglass,  1973;  Morton, et al. 1976).  The highest
 value  reported in  these  studies  was 21,000 ug/1 in  well water  contaminated
 by arsenical grasshopper  bait.
    There is  a wide diversity in  the estimates of daily  intake of arsenic in
 foods.   Schroeder  (1968)  has estimated  that  the  average diet  provides  an
 arsenic  intake of  about  1,000  ug/day.  Arsenic in  a  sample institutional
 diet amounted  to  about 400  ug/day  (Schroeder  and  8alassa, 1966).  This lower
 level  is  attributed, at  least  partially,  to the absence  of  seafood,  a pri-
 mary  source of  arsenic,  in  the  institutional  diet.   In contrast  to  these
 levels,  the World  Health Organization  (WHO)  reported  that  average arsenic
 intakes for Canada,  the United Kingdom, the United  States, and France varied
 from 25 to 33 ug/day; specific values ranged from 7 to 60 ug/day  (WHO, 1973).
    According to  Suta  (1978),  the  levels of atmospheric arsenic  in locations
where major arsenic emitting sources are absent  range  from below the detec-
 tion  limit  of 1  ng/nr to  83  ng/rrr with  an  average  of 3 ng/rrr.   The  an-
 nual  average  near major  emission  sources  (copper,  lead,  and  zinc smelters,
 cotton  gins, pesticide manufacturers,  and  glass  manufacturers) ranged from 3
    •j                •}                            •>
 ng/'nij  to 5,900   ng/m-  with  most  below  290 ng/mj.   Assuming  normal  daily
                                      -111

-------
 inhaled volumes  of 21.2 and 11.1 cubic meters for men and women, respective-
 ly,  the ranges  of  daily airborne arsenic  exposures  in  uncontaminated areas
 are  760 ng  and  11-921  ng  for men  and  women respectively.   In  areas where
 arsenic emitting sources are  located daily, inhaled exposure  levels may be
 as  high as  6,148  to  125,080  ng  and 3,219  to  65,490 ng for  men  and women,
 respectively.
     No  quantifiable  information  was  found concerning present levels of expo-
 sure from drugs  or dermal contact.
 Special Groups at Risk
    Adverse  effects  have been  demonstrated  in all  age  groups  of both sexes.
 Children may have  an increased susceptibility to  arsenic-induced  CNS damage
 (Hamamoto, 1955; Okamura, et al., 1956; Yamashita, et al., 1972).
 Basis and Derivation of  Criterion
    As  described in the Carcinogenicity Section,  a  number of  studies  have
 shown that arsenic is  important  in the  etiology  of human cancers.   Clinical,
 occupational,  and  population  studies  have demonstrated  that  both  ingestion
 and  inhalation  exposures to arsenic compounds  increase the risk  of cancer
 induction  in the  tissues  of  the  lung  and skin  and  possibly other  sites.
 There appears  to be general agreement  that  arsenic  is  a  human  carcinogen,
despite of the fact  that there has  been  general  failure to demonstrate  this
effect in any animal  model.  Hence,  it is  necessary  to  rely totally on human
data  rather  than supplement  it with  appropriate  animal  toxicity and carcino-
genic data.   This  limitation  causes serious problems  since  animal  studies
are  the only practical  means  to  effectively evaluate  relative  toxicities,
 absorption rates,  etc.  for  different compounds  and  routes of  administra-
 tion.   Instead,  these  types of questions  must  be answered  based  on  effects
and  observation  of  exposed populations  recognizing the numerous  unknowns
                                     C-112

-------
 (levels  of  arsenic  and  other  environmental  exposures,   dietary  patterns,
 genetic differences, etc.) and different routes of exposure.
     The  only study that  relates  levels of arsenic  ingestion  to  skin cancer
 is  the one conducted by  Tseng (1968)  in  southwest  Taiwan.   He found  a  con-
 sistent  dose response  relationship  between  the exposure variable  levels  of
 arsenic  in  drinking  water and  age  and  stein  cancer  prevalence.   Questions
 concerning  comparability between  the  U.S. and  Chinese populations  must  be
 raised since some  areas in  the U.S.  have  similar arsenic  levels  without the
 reported dermatological  manifestations.   It is  very  possible that major  dif-
 ferences in  dietary patterns  (the Chinese diet is  low  in  protein and  fat)
 (Veh,  1973), other environmental and/or  occupational  coexposures,  socioeco-
 nomic  status,  etc.  may account for  the differences.   However,  since  similar
 health responses have been observed  in Antofagasta,  Chile  (Borgono  and Grei -
 ber,  1972),  Cordoba,  Argentina  (Bergoglio,  1964),  German  vineyard  workers
 (Oenk,  et   al.  1969),  and  those  who  ingest  Fowler's  Solution  (Neubauer,
 1947), it must be  assumed that arsenic is at  least one component of  the en-
 vironmental  exposures  responsible for  the observed effects.   Secondly,  the
clear  dose  response relationships both by length  of  exposure, as  indicated
 by age, and  by level  of waterborne  arsenic provide  additional  evidence  that
 arsenic is at  least one of  the agents responsible for the  observed effects.
 It  seems  auite unlikely that other environmental,  occupational,  or socio
economic factors which might  be  responsible for  variations  in  skin  tumor in
cidence would  have a  similar gradient to the  waterborne  arsenic  gradient.
 Hence it appears reasonable to use the  Taiwan data as  a  basis  for estimating
 a  level  which will not increase the  lifetime  risk of  cancer by  more  than
 1/100,000.    It is  recognized  the calculated level may be  quite conservative
 since  the  Taiwan experience  may  represent a worst ease  situation due to ex-
 posures and other agents, possibly dietary deficiencies.
                                     C-113

-------
     The  EPA Cancer Assessment  Group  has  developed  a mathematical prediction
 model  for  estimating  an acceptable level based  on  the  published Taiwan data
 (Tseng,  1977).   Their material  is  included  in Appendix  I  to  explain  the
 model  and  the  calculated estimates.
     Under  the  Consent  Decree  in NRDC  v. Train, criteria are to state "recom-
 mended  maximum   permissible  concentrations   (including  where  appropriate,
 zero)  consistent  with  the  protection  of aquatic organisms,  human health,  and
 recreational activities."   Arsenic  1s suspected  of  being a  human carcinogen.
 Because  there  is  no  recognized safe  concentration  for a  human carcinogen,
 the  recommended  concentration of arsenic in water for  maximum protection of
 human health is zero.
     Because  attaining  a zero concentration  level  may be infeasible  in some
 cases  and  in order to  assist the  Agency and  states in the  possible  future
 development  of water quality  regulations, the  concentrations  of arsenic cor-
 responding  to  several  incremental   lifetime  cancer  risk  levels  have  been
 estimated.  A  cancer risk  level  provides  an  estimate of the additional  inci-
 dence  of  cancer  that may  be  expected  in an exposed population.  A risk  of
 lO"^ for  example, indicates  a  probability  of  one additional  case  of  cancer
 for  every  100,000 people  exposed,  a  risk  of  10"6  indicates  one  additional
case for every million people exposed, and so forth.
     In the  Federal Register  notice of availability  of draft  ambient  water
quality criteria,  EPA  stated  that  it  is  considering setting  criteria  at an
 interim  target risk  level of   10"5,  10"6,  or  10"7 as shown  in  the  fol-
 lowing table.
     Exposure Assumptions       Risk Levels  and Corresponding Criteria  (1)
        (per aay;ng/i
                                        10-7          10-6       10-5
 2 liters of drinking water            0772          772         ~72
 and consumption of 6.5 g
 grams fish and shellfish. (2)
 Consumption of fish and               1.75           17.5         175
 shelIfish only.
                                    C-114

-------
     (1)   Calculated  by applying a relative risk model for epidemiology  stud-
          ies,  as  discussed in the Human Health Methodology Appendices to the
          October  1980 Federal Register notice which announced the availabil-
          ity of this  document and  as  discussed in Appendix I.  Since the ex-
          trapolation  model is  linear  at low  doses,  the  additional  lifetime
          risk  is  directly proportional  to  the water  concentration.   There-
          fore,  water concentrations corresponding  to other  risk  levels can
          be derived  by multiplying or  dividing one  of  the  risk  levels and
          corresponding  water  concentrations  shown  in the table  by  factors
          such  as  10,  100,  1,000, and so forth.
     (2)   Thirteen  percent  of  the arsenic exposure  results  from  the  consump-
          tion  of  aquatic organisms  which  exhibit  an average bioconcentration
          potential of 44-fold.   The remaining 87  percent  of arsenic  exposure
          results from drinking water.
    Concentration  levels  were derived  assuming a  lifetime exposure  to  var-
ious amounts of arsenic, (1)  occurring  from the consumption of both  drinking
water and  aquatic  life grown  in waters  containing  the corresponding  arsenic
concentrations and,  (2) occurring  solely from consumption  of  aquatic  life
grown in the waters containing the corresponding arsenic  concentrations.
    Although  total  exposure  information for  arsenic  is  discussed  and  an
estimate  of  the contributions  from other  sources  of exposure  can be  made,
this data will not be factored into ambient  water  quality criteria  formula-
tion until additional  analysis  can be made.   The criteria presented,  there-
fore, assumed  an Incremental  risk from ambient water exposure only.
    The  criterion  as  estimated  by the  methodology may  appear  unreasonably
low.  However, Inorganic arsenic is clearly established  as  a human  carcino-
gen  including  ingestion in  drinking  water.   Further, negative findings  in
                                    C-115

-------
large  population  groups  (Harrington,  st  al.  1980;  norton,  et  al.  1976;
Southwick, st al.  1980) have been carefully evaluated  by  the Agency  to  check
if the  criteron predicts an  incidence above what  was found  in these  stud=
ies.   The Agency concludes  that  the Taiwan experience is not  invalidated  by
the lack of skin cancer  incidence in  areas of the United  States  where people
are exposed to arsenic through drinking water.
                                    C-11S

-------
                                  REFERENCES

Aly,  S.M.,  et  al.   1975.  Toxic deafness.  Histological  study  of the effect
of  arsenic, salicylates, and  auinine  on the organ of  Corti  of guinea pigs.
Jour, of the Egyptian Med. Assoc.  58: 144.

American   Conference   of  Governmental  and  Industrial  Hygienists.   1977.
Threshold  limit  values  for chemical  substances and physical  agents  in work-
room  environment with intended changes for  1977.

Amor, A.J.  and  P.  Pringle.    1945.  A review of  selenium as  an industrial
hazard.  Bull. Hyg.  20: 239.

Ancel,  P.   1946.   Recherche Experimentale  Sur Le  Spina Bifida.  Arch. Anat.
Mier. Morph. Exp.  36: 45.

Anderson,  N.P.   1932.   Bowen's  precancerous  dermatosis  and  multiple benign
superficial epithelioma.  Arch. Derm.  Syphilol.  26: 1052.

Angino, E.E.,  et  al.   1970.  Arsenic  in  detergent:  Possible  danger and pol-
lution hazard.  Science.  168: 389.

Anke,  M.,   et  al.   1978.   Essentiality and  function  of  arsenic.    .In:  M.
Kirchgessner  (ed.),  Trace Element Metabolism  in  Man  and  Animals.   Arbeits-
kreses dur Tierernohrungs forschung, Werhenstephen, Germany.  3:  248.
                                     C-117

-------
Arguello,  R.A.,  et  al.   1938.   Cancer  y  arsenicismo  regional  endemico  en
Cordoba.  Rev. Argentina Oermatosifilol.  22: 461.

Ariyoshi, T.  and T.  Ikeda.   1974.   On  the tissue  distribution  and  the excre-
tion of  arsenic  in  rats  and rabbits  of  administration  with  arsenical  com-
pounds.  Jour. Hyg. Chem.  20: 290.

Askanazy, M.   1927.   Das  experimentelle  karcinom.   Schweiz.  Med. Wochenschr.
57: 1209.

Axelson, 0.,  et  al.   1978.   Arsenic exposure and mortality: A case  referent
study from a Swedish copper smelter.  Br. Jour.  Ind. Med.   35:  8.

Ayres,  S.,  Jr. and N.P.  Anderson.  1934.   Cutaneous manifestations  of arse-
nic poisoning.  Arch. Derm. Syphilol.  30: 33.

Azzone, S.F.  and L.  Ernster.   1961.   Compartmcntation  of  mitochondrial phos-
phorylation  as  disclosed  by  studies  with  arsenate.    Jour.  Biol.  Chem.
236: 1510.

Baetjer, A.M., et  al.   1975.  Cancer  and Occupational Exposure to Inorganic
Arsenic.  |n_:  Abstracts.   18th Int. Cong.  Occup. Health Brighton,  England,
Septemer 14-19.  p. 393.

Baronl, C., et al.   1963.  Cardnogenesis tests of  two inorganic arsenicals.
Arch. Environ. Health.  7: 668.
                                    C-113

-------
 Barren,  E.S.  and  T.P.  Singer.   1943.   Enzyme systems containing active sulf-
 hydryl  groups.  The  role of  glutathione.  Science.  97: 356,

 Barry.  K,G.  and E.G. Herndon.  jr.  1962,  Electrocardiographic changes asso-
 ciated  with acute  arsenic poisoning,  Med* Annu,s Washington, 0»CS  31: 25.

 Bartier.  A.,  et  al.  Analysis  of mortality  experience  of Allied  Chemica!
 plant,   (Unpubl.)

 Beaudoin,  A,RS   1974,   Teratogenicity  of sodium arse.nate  in  rats.   Terato-
 logy,   10: 153.

 Seckman.  G,~  it  al,  1977,   Chromosome aberrations  in  workers exposed  to
 arsenic.  Environ, Htilth Perspect,  19: 1455

 Bencko,  V. and  K,  Symon,   1969.   Dynamics of  arsenic  cumulation in  hairless
mice after peroral  administration.  Jour,  Hyge  Epidemic!,  Microbiol.  Immu-
 nol,  13: 248.

Bencko,  V. and  K.  Symon.   1970.   The  cumulation dynamics  in  some tissue  of
hairless mice inhaling arsenic.  Atmos. Environ.   4: 157.

Bencko, V., et al.   1976.   Blotransfonnation of  As  (III)  to As (V) and arse-
nic tolerance.  Arch. Toxlcol.   36: 159.

Bental,  £.,  et  al.  1961.   Electroencephalographic changes due  to  arsenic,
thallium, and strychnine poisonings.  Confin. Neurol.  21: 233.
                                     C-119

-------
 Bergoglio,  P.M.   1964.   Mortalidad por cancer en  zonas  de  aguas arsenicales
 de   la   Provincia  de  Cordoba,  Republica  Argentina.   Presa  Med.  Argent.
 51:  994.

 Bhuvaneswaran,  C.  and  C.L.  Wadkins.   1977.  Detection  of  a  tightly-oound
 arsenic  containing component(s)  in rat  liver  mitochondria: Conditions  for
 its  formation and dissociation.  Fed. Proc.  36:  727  {Abst. i239S).

 Shuvaneswaran,  C.  and  C.L.  Wadkins.   1978.  The influence of  NAD*-!inked
 substrates  on energy  conservation  at sites 2 and  3  in  mitrochondria  treated
 with inorganic arsenate.  Biochem.  Biophys. Res.  Commiin.  82: 648.

 Bhuvaneswaran,  C.,  et  al.   1972.   Inactivation   of  coupled respiration  of
 mitochrondria by inorganic  arsenate and  partial  restoration by ATP.   Bio-
 chem. Biophys. Res. Commun. 49: 690.

 Birmingham, O.J.  1971.  Arsenic.   _l£:  T.  Fitzpatnck,  et  al. (eds.),  Derma-
 tology in General Medicine.  McGraw-Hill Book Co., New York.

 Birmingham, O.J.,  et al.   1965.   An outbreak of arsenical  dermatosis  in  a
mining community.  Arch. Derm.  91: 457.

 Borgono, J.M. and  R.  Greiber.  1972.   Epidemlological  Study  of  Arsencism  in
 the City of Antogagasta.  ^n_:  O.D.  Hemphill  (ed.), Trace Substances in  Envi-
 ronmental  Health,  V.   Prov.  of  University  of  Missouri's  5th  Annu.  Conf.
 Trace  Subst.  Environ.  Health.  June 29-Ouly 1,  1971.   University of  Mis-
 souri, Columbia,
                                      C-120

-------
 Borgono,  J.M.,  et al.   1977.   Arsenic in the drinking water  of  the  city of
 Antofagasta: Epidemiological  and  clinical  study  before and  after  the  instal-
 lation  of a treatment  plant.   Environ.  Health Perspect.,  U.S. Oept.  Health
 Edu.  Welfare.   19: 103.

 Boutwell,  R.K.    1963.   A carcinogenlcity  evaluation  of potassium  arsenite
 and arsanilic acid.  Jour. Agric. Food Chem.  11: 381.

 Braman,  R.S.  and C.C.  Foreback.   1973.   Methylated forms  of  arsenic  in  the
 environment.  Science.  182: 1247.

 Braun,  W.   1958.  Carcinoma  of the skin  and  the internal  organs caused  by
 arsenic:  Delayed  occupational  lesions  due to arsenic.  German  Med. Monthly
 3: 321.

 Brown,  E.R.,  et  al.   1973.   Frequency  of fish  tumors  found  in   a polluted
watershed as compared to nonpolluted Canadian waters.  Cancer Res.  33:  189.

 Brown,  M.,  et  al.  1976.   Intracellular effects of chronic arsenic  admini-
stration  on  renal proximal  tubule  cells.   Jour.  Toxicol,  Environ.  Health.
1: 505.

Brune,  0.,  et  al.  1980.  Distribution  of 23 elements in  kidney,  liver  and
 lung  of a control group  in  northern  Sweden and of  exposed  workers  from  a
 smeltery and refinery.  Sci. Total Environ.  (In press)
                                     C-12I

-------
 Buchanan,  W.D.   1962.   Toxicity  of Arsenic  Compounds.   Elsevier  Publishing
 Co.,  New  York.

 Butzengelger, K.H.   1940.   liber periphere Zinkulationsstorungen bei  chronis-
 cher  Arsenvergfiftung.   Klin. Hochenschr.  19: 523.

 Butzengeiger,  K.H.    1949.   Chronic arsenic  poisoning.   I.  £KG  alterations
 and other  symptoms observed  on  the heart and vascular system.  Mucons Dtsch.
 Arch. din. Med.  194: 1.

 Byron, W.R., et  al.   1967.   Pathologic  changes in rats and dogs from twoyear
 feeding of sodium arsenite  and  sodium  arsenate.  Toxicol.  Appl.  Pharmacol.
 10: 132.

 Canadian  Public  Health  Association  Task Force on Arsenic.   1978.   Electro-
myography,  Final  Report, Yellowknife and  Hay River,  Northwest Territories.
 Canadian Public Health Association, Ottawa, Canada, November 1978.

 Carlson,  C.W.,  et  al.  1954.   Some effects of  selenium,  arsenicals,  and
 vitamin 8^ on chick growth.  Poult. Sc1.  33: 768.

 Chamberlin. W.  and  J.   Shapiro.   1969.   On  the biological  significance  of
 phosphate  analysis:  comparison  of  standard and new methods  with  a  bioassay.
 Limn. Oceangr.  14: 921.

 Chapman,  A.C.   1926.   On   the  presence  of  compounds of  arsenic in  marine
 crustaceans and shellfish.   Analyst.  51: 548  (London)
                                     C-122

-------
 Charbonneau,  S.M.,  et al.   1978a.   Arsenic  excretion by monkeys  dosed with
 arsenic-containing  fish  or with  inorganic  arsenic.  Bull.  Environ.  Contam.
 Toxicol.  20: 470.

 Charbonneau,  S.M.,  et al.  1978b.   Phannacokinetics  and metabolism of  inor-
 ganic arsenic in the dog.  Trace Subst. Environ. Health.   12: 276.

 Charbonneau,  S.M.,  et  al.   1979.   Metabolism  of orally administered  inorgan-
 ic arsenic in the dog.  Toxicol. Lett.  3: 107.

 Chuttani, P., et al.  1967.  Arsenical neuropathy.  Neurology.   17:  269.

 Cikrt, M.  and V. Bencko.   1974.   Fate of  arsenic after parenteral  admini-
 stration to  rats,  with particular references  to excretion via bile.   Jour.
 Hyg. Epidemic!.  M1crob1ol. Immunol.  18: 129.

 Coughlan, M.P.,  et al.   1969.   The  role  of  molybdenum  in  xanthine  oxidase
 and related enzymes.  Jour. B1ol,  Chem.  244:  2658.

Coulson,  E.J., et  al.   1935.   Metabolism in the  rat  of  the  naturally occur-
ring  arsenic of  shrimp  as compared with  arsenic  trioxide.   Jour.  Nutr.
10: 255.

Crane, R.K.  and  F.  lipmann.  1953.  The effect  of arsenate  on  aerobic  phos-
phorylation.   Jour. 81ol.  Chem.   201: 235.
                                     C-123

-------
 Crecelius,  E.A.   1977.  Arsenite  and  arsenate  levels  in wine.   Bull.  Envi -
 ron. Contam. Toxicol.  18: 227.

 Oema, A.   1955.   Distribution  et  elimination  de  1'arsenic  76  chez la souris
 noimiale et cancereuse.  Arch.  Int. Pharmacodyn.  103:  57.

 Cummings, J.G.  1966.  Pesticides  in the total  diet.  Residue Rev.  16:  30.

 Currie, A.M.   1947.   The  role of  arsenic in carcinogenesis.   8r.  Med.  Bull.
 4: 402.

 Davenport,  H.W.   1955.   Locus of  action of  sulfhydryl  inhibitors  on  gastric
 acid secretion.  Fed. Proc.  14: 35.

 Oelepine, S.   1922.  Observations  upon  the effects of exposure  to  arsenic
 trichloride upon health.  Jour. Ind. Hyg.  4: 346.

 Oelepine,  S.   1923.   Observationsupon  the  effects of  exposure  to  arsenic
 trichloride upon health.  Jour. Ind. Hyg.  4: 410.

 DeMaster, E.G.  and  R.A.  Mitchell.   1970.   The  insensitivity of  mitochon-
dria! -catalyzd arsenate-water exchange reaction to  dinitrophenol  and  to  ali-
gomycin.   Biochem. Biophys. Res. Commun.   39: 199.

 Oenk,   R.t  et  al.   1969.  Late disorders  caused by  arsenic  in  postmortem
 examined  winegrowers of the Mosel region.  Med.  Welt.   11: 557.
                                    C-124

-------
 Doll,  R.   1971-  ^he  age  distribution of  cancer:  Implications  for model of
 carcinogenesis.   Jour. Roy.  stat. Soc.  A134: 133.

 Dreisbach,  R.H.   1971.  Metallic Poisons.  jln_:  Handbook  of Poisoning,  Diag-
 nosis  and Treatment.   7th ed., Lange Med. Pub., Los AHos, California.

 Dubois, K.P.,  et  al.   1940.   Further  studies  on the effectiveness of arsenic
 in preventing  selenium poisoning.  Jour. Nutr.  19: 477.

 Ducoff, H.S.,  et  al.   1948.   Biological  studies with  arsenic.   II. Excretion
 and tissue localization.  Proc. Soc. Exp. Biol. Med.  69: 548.

 Duggan, J.G.  and  G.Q. Lipscomb.   1969.   Dietary intake  of  pesticide chemi-
 cals  In the  United  States  (II),  June  1966-April  1968.   Pestic.  Monitor.
 Jour.  2: 153.

 DuPont, 0.,  et al.   1942.   The  distribution  of radioactive arsenic  in  the
 normal and tumor-bearing (Brown-Pearce) rabbit.   Am.  Jour.  Syph.  Gon. Bener.
 Dis.   26: 96.

Durum,  W.H.,   et  ah   1971,  Reconaisance of  Selected Minor  Elements  in
Surface  Waters of   the  United   States,  October  1970=   Geological  Survey
Circular 643,  Washington,  O.C.:  U.S.  Dept.  Intr,  p.  49,

 Dutkiewicz,  T.  1977,  Experimental  studies on arsenic  absorption  routes in
rats.  Environ. Health Perspect.  19:  173.
                                    C-125

-------
 Edmonds,  J.S.  and K.A.  Francesconi.   1977.  Methylated arsenic  from  marine
 fauna.  Nature.  265: 436.

 Edmonds,  J.S.,  et  al.   1977.  Isolation, crystal  structure  and  synthesis  of
 arsenobetaine,  the  arsenical constituent of  the  western rock  lobster  Panu-
 Urus longiopes cygnus George.  Tetrahedron Lett.  18: 1543.
Ellis, M.M.  1934.  Arsenic storage in game fish.  Copeia.  1934(2):  97.

Feinglass,  E.J.   1973.   Arsenic intoxication  from  well  water  in  the  United
States.  New Eng. Jour. Med.  188:  828.

Perm,  V.H.   1977.  Arsenic  as  a  teratogenlc  agent.   Environ. Health  Per-
spect.  19: 215.

Perm,  V.H.  and  S.J.  Carpenter.   1968.  Malformation induced  by sodium arse-
nate.  Jour. Reprod. Fertil.  17: 199.

Perm,  V.H., et al.  1971.  The  teratogenic  profile  of  sodium arsenate  in the
golden hamster.  Arch. Environ.  Health.  22: 557.
Pierz, U.   1965.   Katamnestische Untersuchungen uber  die  Nebenwirkungen  der
Therapie  m1t   anorganischem  Arsen  bei   Hautkrankheiten.    Oermatologica.
131: 141.
                                     C-126

-------
 Fowler  8.A.      .1.   The  morphological  effects- of  mercury,  cadmium,  lead,
 and arsenic  on  the kidney  in  trace  metals  in water  supplies:  Occurrence,
 significance, and control.   Proc.  16th  Water Qual. Conf.  Champaign-Urbana,
 Illinois,   p. 65.

 Fowler,  B.A.  1975.  The Ultrastructural  and Biochemical Effects of Arsenate
 on  the Kidney.  |n_:  Proceedings of the XVIII International Congress on Occu-
 pational Health,  Brighton.   (Abst.)

 Fowler,  B.A.  1977a.   Toxicology of  Environmental Arsenic,   hi:  R.A. Gayer
 and M.A.  Mehlman  (ed.),  Toxicology  of Trace Elements.   Halstead  Press, New
 York.   p.  79.

 Fowler,  B.A.  1977b.   International  conference  on  environmental arsenic: An
 overview.  Environ. Health  Perspect.   19: 239.

 Fowler, 8.A.,  et  al.   1977.  Ultrastructural  and biochemical  effects of pro-
 longed arsenic exposure on  liver mitochondria of rats.   Environ. Health Per-
 spect.  19:  197.

 Fowler, B.A.,  et  al.   1978.  Effects  of  arsenic on mitochondrial  and micro-
 somal  oxidative   interactions.   Toxicol.  Appl.  Pharmacol.  45: 308  (Abst.
a 206).

 Frank,  G.   1976.   Neurologische  und  psychiatrische Folgesymptome  bei akuter
 Arsen-Wasserstoff-Vergiftung.   Jour.  Neurol.  213:  59.
                                     C-127

-------
Fraumeni,  J.F.,  J-, .   1975.  Respiratory  carcinogenesis.  An  epidemiologic
appraisal.  Jour. Natl. Cancer Inst.  55: 1039.

Freeman,  J.W.  and  J.R. Couch.  1978.   Prolonged  encephalopathy  with  arsenic
poisoning.  Neurology.  28: 853.

Frost,  O.V.    1967.   Recent  Advances  in Trace  Elements:  li±:  Proc.  of  the
Cornell Nutrition Conference for Feed Manfactures.  p. 31

Frost,  O.V., et  al.  1962.  Further Considerations on  the  Safety of  Arsani -
lie  Acid for  Feed  Use.   |n:  Proc.  12th  World's Poultry Congr.   Sydney,
Australia.  Section Pap.  p. 234.

Ganther,  H.E.  and  C.A. Baumann,  1962.  Selenium  metabolism.   I.  Effects  of
diet, arsenic,  and cadmium.  Jour. Nutr.  77: 210.

Garb,  L.G.  and  C.H.  H1ne.   1977.  Arsenical  neuropathy:  Residual  effects
following acute industrial exposure.  Jour. Occup. Med.  19: 567.

Geyer,  L.   1898.   Uber  die chronischen Hautveranderungen  beim Arsenicismus
und Betrachtungen  uber die Massenerfcrankungen 1n  Reichenstein  in Schlesien.
Arch. Derm. Syphilol.  43: 221.

Ginsberg,  J.M.   1965.  Renal mechanism for excretion  and  transformation  of
arsenic in the dog.  Am. Jour. Physiol.  208: 832.
                                     C-128

-------
 Glazener,  F.S., et  al.    1968.   Electrocardiograph^  findings  with  arsenic
 poisoning.  Calif. Med.  109: 158.

 Goeckerman,  L.H.  and  L.F. Wilheim.   1940   Arsenic as  the  cause of  cancer
 mucous membrane.  Arch. Oermat. Syph.  42: 261.

 Goldsmith, J.R., et  al.   1972.   Evaluation  of  health  implications  of  elevat-
 ed arsenic in well water.  Water Res.  6: 1133.

 Graham,  J.H.  and  E.B. Helwlg.   1959.   Bowen's disease and  its  relationship
 to systemic cancer.  AMA Arch. Derm.  80: 133.

 Graham,  J.H.,  et  al.  1961.   Chemistry  of  Bowen's disease:  Relationship  to
 arsenic.  Jour. Invest. Derm.  37: 317.

Grantham,  R.G.  and  C.B.   Sherwood.   1968.   Chemical  Quality of  Waters  of
 Broward  County, Florida.   Florida Geological  Survey,   report of  Investiga-
 tions No. 51.  Tallahassee: State of Florida, Board of Conservation,   p.  52.

Greathouse, D.G. and G.F.  Craun.   1978.   Cardiovascular disease  study  -  oc-
currence of  inorganics  in household tap  water and relationships  to  cardio-
vascular mortality rates.   Trace Subst. Environ. Health - XII.

Grobe, J.W.   1976.  Perophere  Durchblutungs  storunger  und  Akrocyanose bie
 arsmje schadlgten  Mosel win zern.  Bemfs-Oemiatosen.   24:  78.
                                   C-129

-------
Grunicke,  H.t  et  al.   1973.   Effect  of  alkyTating antitumor agents  on  the
binding of ONA to  protein.  Cancer Res.  33: 1048.

Hamamoto,  E.  1955.   Infant  arsenic  poisoning by  powdered  milk.  Jap.  Med.
Jour.  1649: 2.

Hanlon, D.P.  and V.H. Perm.   1977.   Placenta!  permeability of  arsenate  ion
during early embryogenesis in the hamster.  Experientia.  33:  1221.

Hara,  I.,  et  al.  1968.   Case studies of  polyneuritis  believed  to  have been
caused by  arsenical  poisoning.  TR-79-0524, Labor  Health Department  of  the
Osaka  Public  Health  Research  Center,   Osaka   University.    Sangyo  Igaku.
11: 84.

Harrington, J.M.,  et al.   1978.   A  survey of  a  population exposed to  high
concentrations of  arsenic  in  well water  in Fairbanks,  Alaska.  Am.  Jour.
Epid.  108: 377.

Harris, E.J. and P.M. Achenjang.   1977.   Energy-dependent uptake of  arsenite
by rat liver mitochondria.  Biochem.  Jour.  168:  129.

Harrison,  J.W.E., et al.  1958.  Acute oral  toxicity and chemical  and  physi-
cal properties of arsenic trioxides.   AHA Arch.  Ind. Health.   17: 118.

Harvey, S.C.   1975.   Heavy  Metals.  The Pharmacological  Basics  of  Therapeu-
tics, 5th ed.   A.6. MacMlllan Publishing  Co., Inc.
                                    C-130

-------
 Hendrick,  C., et  al.   1953.  Effect  of 3-oitro-4-hydroxyphenylarsonic acid
 and  arsanilic acid  on  selenium poisoning  in the  rat.  Jour. Nutr.  51:  131.

 Hernberg,  S.   1977.    Incidence  of  Cancer  in  Population With  Exceptional
 Exposure   to  Metals.    Jji:  Cold   Spring   Harbor   Conferences  on   Cell
 Proliferation.   Vol.   4.   Origins  of  Human Cancer.  Book  A.   Incidence  of
 Cancer  in  Humans.   H.H.  Hiatt,  et  al.  (eds.), Cold Spring  Harbor Lab., Cold
 Spring  Harbor, New  York.  p. 147.

 Heyman, A.,  et al.   1956.   Peripheral  neuropathy caused  by arsenical  intoxi-
 cation.  A  study of 41 cases with  observations  on the  effects of  8AL  (2,3-
 dimercapto-propanol).  New Eng. Jour. Med.  254:  401.

 Hill, A.8.  and E.L.  Faning.   1948.   Studies in the incidence of  cancer in a
 factory handling  inorganic compounds  of arsenic.   I.  Mortality experience  in
 the factory.  Br. Jour. Ind. Med.  5: 1.

 Hindmarsh,  J.T.,  et al.   1977.   Electromyographic abnormalities  in  chronic
environmental arsenical ism.  Jour.  Anal. Toxicol.  1:  270.

Holland, J.W.   1904.   Arsenic.  _In_:  F.  Peterson  and W.S. Haines  (eds.),  A
Textbook of  Legal Medicine and Toxicology.  W.S.  Saunders and Co., Philadel-
phia.  2:  404.

Holland,  R.H.,  et  al.   1959.   A  study of  inhaled  arsenic   -  74 in  man.
Cancer Res.  19:  1154.
                                    0131

-------
 Holmqvist,  I.   1951.    Occupational  arsenical  dermatitis.   A  study  among
 employees  at a  copper ore  smelting  worlc  including  investigations of  skin
 reactions to  contact with arsenic compounds.  Acta Derm. Venereol.   31:  1.

 Hood,  R.O.  and  S.L. Bishop.   1972.   Teratogenic effects of  sodium arsenate
 in mice.  Arch.  Environ. Health.  24:  62.

 Hood,  R.D., et al.  1977.  Effects in the mouse  and rat of  prenatal exposure
 to arsenic.   Environ. Health Perspect.  19:  219.

 Hueper, W.C.  1951.  Environmental lung cancer.  Ind.  Med. Surg.   20:  49.

 Hueper, W.C.  and  W.W.  Payne.  1962.   Experimental studies in metal  carcino-
 genesis.  Chromium, nicxel, iron, arsenic.  Arch. Environ. Health.   5:  445.

Hunter, F.T., et  al.   1942.   Radioactive tracer studies  on arsenic  injected
 as potassium  arsenite.   Jour. Phartnacol. Exp.  76: 207.

Hutchinson, J.  1888.  On some examples  of  arsenic-keratoses  of  the skin and
of arsenic-cancer.  Trans. Path. Soc.   39: 352.

 International Agency for  Research  on  Cancer.  1973.  IARC Monographs on the
evaluation  of carcinogenic  risk of  chemicals  to  man:  Some inorganic and
organometalUc compounds.  2: 48.

 Irgolic, K.   1979.   Spedation  of Arsenic in Water Supplies.  0.  Prog. Rep.
 Nov. 1, 1978-Jan.  31, 1979.  EPA Grant R-804774.
                                    C-132

-------
 Ishinishi,  M.   1973.   Nihon Rinsho. 31: 1991.  (Jap.)

 Japanese  Pediatric Society.   1973.   Morinaga arsenic-tainted  powdered  milk
 poisoning  Investigation  special  committee.   Summary  of  Report  of Activities
 of  the Morinaga Arsenic-tainted Powdered Milk  Investigation.  May 26, 1973.

 Jelinek, C.F.  and  P.E. Corneluissen.   1977,   Levels  of arsenic  in the United
 States food supply.  Environ. Health Perspect.  19: 83.

 Jenkins,  R.B.   1966.    Inorganic  arsenic and the  nervous system.   Brain.
 89: 479.

 Johnson, J.L.  and  K.V. Rajagopalan.   1978.   The  interaction of  arsenite  with
 the molybdenum center of  chicken  liver  xanthine dehydrogenase.   Bioinorg.
 Chem.   8: 439.

 Johnstone,  R.M.    1963.    Sulfhydryl   Agents:  Arsem'cels.  J_n:  Metabolic
 inhibitors.   A Comprehensive  Treatis.   Vol.  II.   R.H. Hochster  and  J.H.
Quastel. (eds.), New York:  Academic Press,  p. 99,

Josephson,  C.J.,  et al.   1951.    Arsine:  Electrocardiographic  changes  pro-
duced  in acute human poisoning.  AMA Arch. Ind. Hyg.  Occup.  Med.  4: 43.

Jung,  E.  1969.  Arsenic as an  inhibitor  of the  enzymes concerned  in cellu-
 lar recovery (dark repair).  Ger,  Med. Mon.  14:  614.
                                    C-133

-------
 Jung,  E.   1971.  Molekularbiologische  Untersuchungen  zur  chronischen Arsen-
 vergiftung.  Z. Haupt. Geschlechtske.   46: 35.

 Kadowaki, K.   1960.   Studies  on  the arsenic  contents  in organ tissues of the
 normal Japanese.  Osaka  City  Med.  Jour.  9:  2083.   (In Japanese with English
 summary.)

 Kagey,  B.T.,  et al.   1977.   Arsenic  levels  in maternal-fetal  tissue  sets.
 Trace Subst. Environ. Health.  11: 252.

 Kanisawa, M. and  H.A.  Schroeder.  1969.  Life  term studies  on the effect of
 trace elements on spontaneous tumors in mice and rats.  Cancer Res.  29:  892.

 Klaassen, C.O.   1974.   Biliary  excretion of arsenic  in rats,  rabbits,  and
 dogs.  Toxicol. Appl. Pharmacol.  19: 447.

 Klemmer,  H.W.,  et  al.    1975.   Arsenic  content  of  house  dusts  in  Hawaii.
 Bull. Environ.  Contain. Toxicol. 4.

 Kopp, J.F. and R.C. Kroner.   1967.   Tracing  water  pollution  with an electron
spectrograph.   Jour. Water Pollut. Control Fed.   39(1): 1659.

Kraybill,  H.F.  and  M.B. Shimkin.   1964.   Carcinogenesis  related  to  foods
 contaminated by processing and fungal metabolites.   Adv. Cancer Res.  8:  191.
                                    C-134

-------
Kroes, R.,  et  al.   1974.  Study on  the  carcinogenicity  of  lead arsariate and
sodium  arsenate  and on  the  possible  synergistic  effect of  diethylnitrosa-
mine.  Pood Cosmet. Toxicol.   12: 671.

Kuratsune,  M.,  et  al.    1974.   Occupational  lung cancer among  copper  smelt-
ers.  Int.  Jour. Cancer.  13: 552.

Lakso,  J.U.  and S.A.  Peoples.   1975.  Methylation  of inorganic arsenic  by
mammals.  Jour. Agric.  Food Ghent.  23: 674.

Landau,  E.,  et al.  1977.   Selected non-carcinogenic effects  of  industrial
exposure to inorganic  arsenic,  EPA  569/6-77-018, U.S.  Environ.  Prot.  Agency,
Washington, DC.

Lanz, H., Jr.  et al.   1950.   the metabolism of arsenic in  laboratory animals
        74
using As   as a tracer.  Univ. Calif. Publs. Pharmacol.  2:  263.

Larsen, N.A.,  et al.   1972.   Neutron Activation Analysis of  Arsenic,  Manga-
nese  and  Selenium  Concentrations  in Organs of  Uraemic  and  Normal  Persons.
_In_:  Nuclear  Activation  Techniques  in  the   Life  Sciences,  IAEA,  Vienna.
p. 561.

Larsen, N.A.,  et  al.   1979.   Topographical distribution of  arsenic,  manga-
nese, and selenium in the normal human brain.  Jour. Neurolog. Sci.   42: 407.

LeBlanc, P.J.  and  A.L.  Jackson.  1973.  Arsenic in  marine  fish and inverte-
brates.  Mar. Pollut. Bull.  4: 88.
                                    C-135

-------
 Lee,  A.M. and  J.F.  Fraumeni,  Jr.   1969.  Arsenic  and  respiratory cancer in
 man:  An  occupational  study.  Jour. Natl. Cancer Inst.  42: 1045.

 Levander,  O.A. and  C.A.  Baumann.   1966a.   Selenium metabolism.   V.  Studies
 on  the distribution  of  selenium  in  rats given arsenic.   Toxicol.  Appl.  Phar-
 macol.   9: 98.

 Levander,  O.A. and  C.A.  Baumann.   1966b.   Selenium metabolism.   VI.  Effect
 of  arsenic on  the  excretion  of selenium in  the  bile.   Toxicol.  Appl.  Pharma-
 col.  9:  106.

 Liebscher, K.  and H.  Smith.  1968.   Essential  and nonessential   trace  ele-
 ments.  A method of  determining whether an  element  is  essential or nonessen-
 tial in human  tissue.  Arch. Environ.  Health.   17:  881.

 Linn, T.C.,  et al.  1969.  A-keto acid  dehydrogenase complexes.   X.  Regula-
 tion  of   the  activity  of  pymvate dehydrogenase  complex  from  beef kidney
mitochondria  by phosphorylation   and  dephosphorylation.   Proc.  Natl. Acad.
 Sci.  64: 227.

Lis,  S.A. and P.K.  Hopke.  1973.    Anomalens  arsenic  concentrations  in
Chantaugua Lake.  Environ. Lett.   5:  45

Livingston, D.A.   1963.   Data of Geochemistry (6th  ed.)  Chapter 6.  Chemical
Composition of  Rivers and  Lakes.   U.S. Geological   Survey  Professional Paper
440-6.  Washington, O.C.:  U.S.  Government Printing  Office,  p. 64.
                                    c-136

-------
 Lu,  P.J.,  et al.   1975.   Physico-chemical  characteristics  of drinking water
 in Blackfoot  endemic  areas of Chia-T and Tainan Hsiens.  Jour. Formosan Med.
 Assoc.   74: 596.

 Lu,  F.J.,  et  al.   1977a.   Studies  on  fluorescent  compound  in drinking water
 of Blackfoot  endemic  areas.  I. The toxic  effect  of  fluorescent  compound on
 the  chick embryos.  Jour.  Fonnosan Med. Assoc.  76: 58.

 LeOuesne,  P.M.  and J.G.  McLeod.   1977.   Peripheral  neuropathy  following  a
 single exposure to  arsenic.   Jour. Neurol.  Sci.  32: 437.

 Lu,  F.J., et  al.   1977b.   Studies  on fluorescent  compounds  in drinking water
 of Blackfoot  endemic  areas.  II. Isolation and identification of fluorescent
 compounds.  Jour. Formosan  Med. Assoc.  76: 209.

 Mappes,  R.   1977.   Experiments  on  excretion  of arsenic  in  urine.   Versuche
 zur  Ausscheidung  von  Arsen  in  Urin.   Int.  Arch.  Occup.  Environ.  Health.
 40: 267.

Masahiki, 0.  and  A. Hideyasu.   1973.   Epidemiological  studies on  the Mori-
naga  powdered milk poisoning incident:  Final  report  of the joint  project
team  from Hiroshima and Okayama Universities  for  survey of  the  Senoi area.
 Jap. Jour.  Hyg.  27: 500.

Massmann, W.  and  H.  Opitz.  1954.  Experimentalle Untersuchungen  uber  Ekg-
 Vernderunger bei  chronischer Arsenvergiftung.  Z.  Kreislaufforsch.  43: 704.
                                    C-137

-------
Mealey,  J.,  Jr.,   et  al.   1959,   Radloarsenic   in  plasma,  urine,  normal
tissues, and intracranial neoplasms.  Arch. Neurol. Psychiatr.   81:  310.

Mertz, H.   1970.  Aspects of nutritional  trace  element  research.   Fed.  Proc.
29: 1482.

Miescher, G.   1934.  Statische  agnaben  aus  der krebstatisik  der  dertnatolo-
gischem klinik Zurich.  Dermatol. Wochrnschr.  98:  420.

Milham,  S.,  Jr.   1977.   Studies  of  morbidity near a copper smelter.   Envi-
ron. Health Perspect.  19: 131.

Milham, S., Jr. and  T. Strong.   1974.   Human arsenic  exposure  in  relation to
a copper smelter,  Environ.  Res.  7: 176.

Milner,  J.E.   1969.  The effect of ingested arsenic on  methylcholanthrene-
induced skin tumors in mice.  Arch. Environ.  Health.  18:  7.

Mitchell,  R.A.,  et  al.   1971.   Inhibition  of mltochondrial   energy-linked
functions of  arsenate.   Evidence  for  a  non-hydrolytic  mode of  inhibitor
action.  Biochem.  19: 1049.

Mizuta, N., et  al.   1956.   An outbreak of acute  arsenic  poisoning  caused by
arsenic  contaminated soy  sauce  (shoyu):  A  clinical  report  of  220  cases.
Bull. Yamaguchl Med. Sen.  4:  131.
                                    c-138

-------
Montgomery,  H.    1935.   Arsenic as  an etiologic  agent  in  certain  types  of
epithelioma.   Differential  diagnosis  from,  and  further  studies  regarding
superficial  epitheliomatosis  and  Bowen's  disease.  Arch.  Derm.  Syphilol.
3?:  218.

Morton,  W.,  et  al.   1976.  Skin  cancer  and water  arsenic  in  Lane  County,
Oregon.  Cancer.  37: 2523.

Moxon,  A.L.   1938.  The  effect of  arsenic  on  the  toxicity of  seleniferous
grains.  Science.  88: 81.

Moxon, A.L. and  K.P. DuBois.   1939.   The  influence of arsenic  and other ele-
ments on the toxicity of seleniferous grains.  Jour. Nutr.   8:  447.

Munro, I.C.  1976.  Naturally  occurring toxicants  in foods  and their signif-
icance.  Clin. Toxicol.  9: 647.

Munro,  I.C.,  et  al.   1974.   Biological  availability  of arsenic  from  fish.
Toxicol. Appl. Pharmacol.  29:  111  (Abst. #92).

Mushak, P.  1977.   Advances  in the  analysis  of toxic heavy elements  having
variable chemical forms.  Jour. Anal. Toxicol.  1: 286.

Mushak, P., et  al.   1977.  Flameless  atomic  absorption  (FAA)  and gas-liquid
chromatographic  studies  in  arsenic  bioanalysis.  Environ.  Health Perspect.
19: 5.
                                    0139

-------
 Mu$1l,  J.  and V,  Dejmal.  1957.  Experimental and clinical administration of
 radio-arsenic (76As).   Cas.  Lek.  Cesk.   96: 1543.  (In  Czech,  with English
 summary).   (Chem. Abst.  52:  14008, 1958)

 Nagal,  H., et  al.    1956.   Subacute-chronic "arsenic" poisoning  in infant-
 subseouent  clinical observations.  Ann. Pediat.  2: 124.

 Nagamatsu,  K.  and  A.   Igata.   1975.   Dominant  and  non-dominant  arsenical
 neuropathy,   TR-79-0523, Kagoshim University.  Rinsho Shinkei.  15: 1.

 Nakamura,  I.,  e al.   1973.   Study on  the  effect  of  arsenic  on human bodies,
 Part  I.  Japanese Public Health  Association  (Translated for  EPA  by SCITRAN,
 Santa Barbara, Calif., EPA translation No. TR-120).

 National Academy of  Science.  1977a.  Arsenic.   National  Academy  of  Sci-
 ences, Washington, O.C.

 National Academy  of   Science.   1977b.   Drinking water  and health.   National
Academy of Sciences,  Washington, D.C.

National Institue of  Occupational  Safety  and Health.   1975.   Criteria  for  a
recommended  standard...Occupational   exposure  to  inorganic  arsenic.   U.S.
Dept. Health,  Edu.,  Welfare.  DHEW  Publ.  No. (NIOSH)  75-149.   U.S.  Govern-
ment Printing Office, Washington, D.C.

Natusch,   O.F.S.,  et   al.    1974.    Toxic  trace  elements:   Preferential
concentration in respirable particles.  Science.  183:  202.
                                    C-140

-------
 Nelson,  W.C.,  et al.   1973.  Mortality among orchard workers exposed to  lead
 arsenate spray:  A  cohort  study.   Jour.  Chron. Dis.  26: 105.

 Neubauer,  0.   1947.   Arsenical cancer:  A review.  Br. Jour. Cancer.  1: 192.

 Nielson,  F.H., et  al.   1978.   Newer Trace  Elements  -- Vanadium (V) and Arse-
 nic  (As) Deficiency Signs  and  Possible Metabolic Roles.   Ln:  M.  Kirchgess-
 ner  (ed.), Trace  Element  Metabolism in Man and  Animals.   Arbeitskreses  fur
 Tierernahrungsforschung, Weikenstephen, Germany.  3: 244.

 Nielsen,  I.H., et  al.   1974.   Evidence  of  a possible requirement for arsenic
 by the rat.  Fed.  Proc.  34: 923.

 Nozaki,  $., et al.  1975.   Effect of casein on entera1  absorption of arsenic
 trioxide.  Jap.  Jour. Pharmacol.  25 Suppl:122P.  (Abst.  #171)

 Obermeyer, B.D., et al.   1971.   Toxicity  of trimethyselenonium  chloride in
 the rat with and without arsenite.  Toxicol. Appl. Pharmacol.  30:  135.

 Odanaka,  v.,  et  al.   1978.  Identification of  dimethylated arsenic  by  gas
 chromatographywnass spectrometry  in blood, urine, and feces  of  rats treated
with ferric m«thanearsonate.  Jour. Agric.  Food Chem.  26: 505.

 Ohira, M.  and H.  Aoyama.   1972.   Epidemiological   studies  on  the  Morinaga
 powdered  milk  poisoning   incident.   (Translated  for  Information  Sciences
 Division,  EPA,  by  Leo  Kanner Associates,  Redwood  City,  California 94363).
 Jap. Jour. Hyg.  27: 500.

-------
Ohta,  M.   1970.   Ultra-structure  of  sural  nerve  in  a  case of  arsenical
neuropathy.  Acta Neuropathol.  16: 233.

Okamura, *•» et al.   1956a.   Symposium  on  arsenic  poisoning  by powdered milk
(2).  Oiagnos. Ther.  9: 155.  (Jap.)

Okamura, K., et al.   1956b.   Symposium  on  arsenic  poisoning  by powdered milk
(2).  Oiagnos. Ther.  9: 240.  (Jap.)

Olson, O.E.,  et  al.   1963.   Selenium toxicity.   Effect of  arsenic  on  sele-
nium metabolism in rats.  Jour. Agr-fc. Food Chem.  11: 531.

Onishi, H.  1969.  Aresnic Chapter  33.   lr±:  K.H. Wedepohl  (ed.),  Handbook of
Geochemistry.  Berlin: Sponger-Verlay.

Osata, K.   1977.   Effects  of oral  administration of  arsenic trioxide  during
the suckling stage of rats.  Fukuoka Acta Med.  68: 464.

Osburn,  H.S.   1957.   Cancer  of  the  lung  in Gwanda.   Central  African  Jour.
Med.  3:  215.

Osburn, H.S.  1969.   Lung  cancer  in a mining district in Rhodesia.  S.  Afri-
can Med. Jour.  43: 1307.

O'Shaughnessy,  E.  and   G.H.  Kraft.    1976.   Arsenic  poisoning:  Long-term
follow-up of a nonfatal  case.  Arch. Phys. Med. Renabil.  57: 403.
                                     C-142

-------
Osswald,  H.  and  Kl.  Goerttler.  1971.   Laukosen bei  der  Maus nach  diapla-
centarer und postnataler Arsenik-Application.  Otsch. Gesmte Path.   55:  289.

Ott,  M.G.,  et al.   1974.   Respiratory cancer  and  occupational  exposure  to
arsem'cals.  Arch. Environ. Health.  29: 250.

Overby,   L.R.   and   R.L.   Fredrickson.    1963.    Metabolic  stability   of
radioactive arsenilic acid in chickens.  Jour. Agric. Food  Chem.   21:  378.

Packer,  L.   1961.   Metabolic  and  structural  states  of mitochondria.   II.
Regulation by phosphate.  Jour. Biol. Chem.  236: 214.

Palmer, G.  1962.   Some kinetic studies on  aldehyde oxidase.  Biochem.  Bio-
phys. Acta.  64:  135.

Paton, G.R. and  A.C. Allison.  1972.   Chromosome damage in human  cell  cul-
tures induced  by metal salts.  Mutat. Res.  16: 332.

Patty, F.A.   1948.  The  Mode of  Entry and  Action  of Toxic Materials.   J>K
F.A. Patty (ed.),  Industrial  Hygiene and  Toxicology.   Interscience  Publish-
ers, New York.  1: 175.

Penrose,  W.R., et  al.   1977.   Implications of inorganic/organic  interconver-
sion on  fluxes  of arsenic  in marine  food webs.  Environ.  Health  Perspect.
19: 53.
                                     C-143

-------
 Peoples,  S.A.   1964.   Arsenic  toxicity in  cattle.   Ann.  N.Y.  Acad.  Sci.
 Ill:  644.

 Perry,  K.,  et al.   1948.   Studies in the  incidence  of cancer in  a  factory
 handling  inorganic  compounds  of  arsenic.   II.  Clinical  and  environmental
 investigations.  8r. Jour. Ind. Med.  5: 6.

 Peterkova,  R. and  L.  Puzanova.   1976.   An effect  of  tri- and  pentavalent
 arsenic on the early development of chicken embryos.   Folia Morphol.  24:  5.

 Peters, J.M. and D.R. Sanadi,   1961.   Effect of  arsenite and cadmium  ions  on
 xanthine oxidase.  Arch. Biochem. Siophys.  93: 312.

 Peters, R.A.  1955.  Biochemistry  of  some  toxic  agents.   Bull.  Johns  Hopkins
 Hosp.  97: 1.

 Petres, J. and M.  Hundeiker.   1968.   "Chrotnosomenpulverisation" nach  Arsene-
 inwirkung auf Zelljulturen in vitro.  Arch. Klin. Exp. Dermatol.   231: 366.

Petres, J., et al.   1970.  Chromosomenaberrationen an menschlichen  Lymphozy-
ten bei chronischen Arsenchaden.  Dtsh. Med. Wochenschr.  95:  79.

Petres, J.t  et  al.  1972.   Zum Einfluss  anorganischen  Arsens auf die  ONS-
 Synthese nunschlicher Lymphocyten in vitr.  Arch. Derm. Forsch.  242:  343.

 Pinkus, H. and A. Mehregan.  1969.  Decrease and Increase of  Spidennal  Mela-
 nin,  and  Bowen's Precancerous  Oermatosis and Erythroplasia of Queyrat.   hK
A Guide to Dermatohistopathology.  Meredith Corp., New York.  p.  324.
                                    c-144

-------
Pinto, S.S. and  B.M.  Bennett.   1963.   Effect of arsenic trioxide exposure on
mortality.  Arch. Environ. Health.  7: 583.

Pinto,  S.S.,  et  al.   1976.   Arsenic  trioxide absorption  and  excretion  in
industry.  Jour. Occup. Med.  18: 677.

Pinto, S.S.,  et  al.   1977.  Mortality  experience  in relation to  a  measured
arsenic trioxide exposure.  Environ. Health Perspect.  19:  127.

Rajagopalan,  K.V.  and P.  Handler.   1964.   Hepatic  aldehyde oxidase.   III.
The substrate-binding site.  Jour. Biol. Chem.  239:  2027.

Rajagopalan,  K.V.  and  P.  Handler.   1967.   Purification  and properties  of
chicken liver xanthine dehydrogenase.   Jour. Biol.  Chem.  242: 4097.

Ray,  B.J.  and  D.L.   Johnson.    1972.   A  method  for  neution   alteration
analyses of natural  water for arsenic.  Anal. Chem. Acta.  63: 196.

Ray-8ettley,  F.  and  J.A. O'Shea.   1975.   The absorption of  arsenic  and  its
relation to carcinoma.  8r. Jour. Dermatol.  92: 563.

Regelson,  W., et  al.   1968.    Hemangioendothellal   sarcoma  of   liver  from
chronic arsenic intoxication by Fowler's solution.   Cancer.   21:  514.

Reinke,  J.,  et  al.   1975.  The determination of  arsenite and   arsenate  in
fish and shellfish by  selective  extraction  and  polarography.   Environ.  Lett.
8: 371.
                                    C-145

-------
Rentoul,  E.  and  H.  Smith,  (eds.)   1973.  Arsenic  and  Its Compounds.   Jn.:
Glaister's Medical  Jurisprudence and Toxicology.   13th  ed.  Churchill  Liv-
ingstone, London,  p. 537.

Reynolds, E.S.   1901.  An  account of  the  epidemic  outbreak  of  arsenical  poi-
soning occurring  in  beer-drinkers  in  the north of England  and  Midland  Coun-
tries in 1900.  Lancet.  1: 166.

Riberio, H.A.   1971.  Effects  of  certain metal  ions  on hepatic  microsomal
enzymes.  Proc. Western Pharmacol.  Soc.   13:  13.

Ridgway, L.P. and O.A. Karnovsky.   1952.   The  effects  of metals on the  chick
embryo: Toxicity  and  production of  abnormalities in development.  Ann.  N.Y.
Acad. Sci.  55: 203.

Robinson,  T.J.    1975.   Arsenical   polyneuropathy  due to  caustic  arsenical
paste.  Br. Med. Jour.  3: 139.

Robson, A.O. and A.M.  Jelliffe.  1963.   Medicinal  arsenic poisoning  and  lung
cancer.  Br.  Med. Jour.  2: 207.

Rockstroh, H.  1959.   Zur  Atiologie  des  Bronchialkrebses  in  arsenverarbeiten-
den  Nickelhutten.   Bletrag  zur Syncarcinogenese  des  Serufskrebses.   Arch.
Geschwulsforsch.  14: 151.

Rosset,  M.   1958.   Arsenical   keratoses  associated with  carcinomas of  the
internal organs.  Can. Med. Assoc.  Jour.  78: 416.
                                     C-146

-------
Rossman, T.G.,  et  al.   1977.  Effects of  arsenite  on  ONA  repair in Eschen .
chia coli.  Environ. Health Perspect.  19: 229.

Roth,  F.   1957.   The seauelae of chronic  arsenic poisoning  in  Moselle  vint-
ners.  German Med. Monthly.  2: 172.

Rozenshtein,  I.S.   1970.   Sanitary  toxicological  assessment of  low  concen-
trations of arsenic trioxide in the atmosphere.  Hyg. Sanit.   34: 16.

Satake, S.   1955.   Concerning the cases of arsenic  poisoning caused  by pre-
pared powdered milk.  Jap. Jour. Pub. Health.   2: 22.

Schiller,  C.M., et  al.   1977.  Effects of arsenic  on  pyruvate  dehydrogenase
activation.  Environ. Health Perspect.  19: 205.

Schroeder,  H.A.  and J.J.  Balassa.   1966.  Abnormal  trace elements  in  man:
Arsenic.  Jour. Chron.  Ois.  19: 85.

Schroeder, H.A., et al.   1968.   Germanium, tin,  and arsenic  in  rats:  Effects
on  growth,   survival,   pathological   lesions   and   life  span.   Jour.  Nutr.
96: 37.

Schwartz,  K.  1977.  Essentiality versus  Toxicity of Metals.  Ln:  S.S.  Brown
fed.),  Clinical Chemistry and  Chemical   Toxicology  of  Metals.   Elsevier,
Amsterdam.  1: 3.

Seydel, I.S.   1972.   Distribution and circulation  of  arsenic  through water,
organisms  and sediments of Lake Michigan.   Arch. Hydrobiol.  71: 17.
                                    C-147

-------
 Silver,  A.S.  and  P.I.  Wainmen.   1952.  Chronic  arsenic  poisoning following
 use  of  an  asthma  remedy.  Jour. Am. Med. Assoc.  150: 584.

 Smith,  T.J.,  et al.  1976.  The chemistry  of  sulfur and  arsenic in airborne
 copper  smelter  participates.  Bull. Environ. Contam. Toxicol.  15: 651.

 Smith,  T.J.,   et   al.   1977.   Airborne arsenic  exposure  and   excretion  of
methylated arsenic compounds.  Environ. Health Persp.  19: 89.

 Snegireff, L.S. and O.M.  Lombard.   1951.   Arsenic  and cancer.   Observation
in the metallurgical industry.  AMA Arch. Ind. Hyg.  4:  199.

Sonins,   L.V.    1970.    Arsenic  and  water   pollution   hazard.    Science.
170: 871.

SomffierS, S.C.  and R.G. "enanuS.   1953.  Multiple  arsenical  CanCerS  of  the
skin and internal  organs..  Cancer.   6: 347.

Southwick,  J.W.,  et  al.   1980.   Community health associated  with  arsenic  in
drinking water in  Hi Hard  Co.,  Utah.   Utah  Stats department of  Health,
Division of  Environmental  Health  Statistics,  and  State   Health  Laboratory.
Grant No.  R804 617-01.   Prepared  for Health  Effects  P*esearch  Laboratory,
U.S. Environ.  Prot. Agency.

Spehar,   R.L.  et  al.  1980.  Comparative  toxicity  of  arsenic compounds  and
their  accumulation  in  invertebrates  and   fish.   Arch.  Environ.  Contam.
Toxicol.  9:  55.
                                    0148

-------
 Stevens,  J.T.,  et  al.    1977.   Disposition  of  14C and/or  74As-cacodylic
 acid  in rats  after intravenous,  intratracheal,  or  peroral  administration.
 Environ. Health Perspect.  19: 151.
 Stephan, C.E.  1980.  Memorandum to J. Stara.  U.S. EPA.  July 3.

 Stockinger,  M.E.  and R.L.  Woodward.   1958.   Toxicologic methods  for  estab-
 lishing drinking water standards.  Jour. Am.  Water Works Assoc.  50:  517.

 Suta, B.E.   1978.  Human  exposures to atmospheric  arsenic.   SRI  Project  EGU-
 5794.  Cress Report No. 50.  U.S. Environ. Prot.  Agency.  (Unpubl.)

Takahashi,  K.    1974.   Investigation  of  chronic  arsenic  poisoning.    Jap.
Publ. Health Assoc.  (Jap.)

Tarn, K.H.,  et  al.   1978.   Separation of  arsenic  metabolites in  dog  plasma
and  urine   following   intravenous   injection   of    As.    Anal.   Biochem.
86: 505.

Tarn, K.H.,  et  al.   1979.   Confirmation of  inorganic  arsenic and  dimethyl -
arsinic acid in urine  and plasma  of dog  by ion-exchange  and  TLC.   Bull.
Environ. Contam.  Toxicol.   21:  371.

Tamura, S.   1978.   Study  of  arsenic metabolism  (Report 20)  - Influence  of
arsenic  agents   on   the   brain   of  rats   in   the   developmental   stage.
TR-79-0356.  Folia Pharmacol.  Jap.   74: 1.
                                    C-149

-------
Tamura, S., et al.   1977.   Studies  on  arsenic  metabolism.   XX.  Arsenic  accu-
mulation  in  organs  and excretion into  the feces and  urine of rats  chroni-
cally poisoned with arsenic.  Folia  Pharmacol.  Jap.   73:  877.

Tay, C.H. and C.S.  Shea.   1975.   Arsenic  poisoning  from  anti-asthmatic  herb-
al preparations.   Med. Jour. Aust.  2:  424.

Tjaijt J.K. and  0.  Aziz.   1971.  A mass  acute poisoning from  rice  contami-
nated with arsenic in two orphanges.  Paediatr. Indon.   11:  91.

Tokanehara, S., et  al.  1956.   Blood findings  in infantile  arsenic  toxicoses
caused by powdered milk.   Shonika Ninsho.   9:  1078.   (EPA translation No.  TR
110-74)

Tomatis,  L.   and  U.   Mohr,   (eds.)   1971.   Transplacental  Carcinogenesis.
Proceedings  of  a  meeting  held  at  the  Medizlnische  Hochschule,  Hannover,
Federal Republic  of Germany,  6-7  October.  Int. Agency Res.  Cancer,  1973.
Lyon, France.

Tomatis,  L.  and  U.  Mohr.   1973.  Transplacetal  carcinogenesis.   Mod. Trends
Oncol.  1: 99.

Tseng, W.  1977.  Effects  and  dose-response relationships of  skin cancer  and
Blackfoot disease with arsenic.  Environ. Health Perspect.   19:  109.

Tseng, W.P.,  et  al.   1968.  Prevalence of  skin  cancer  in an  endemic area  of
chronic arsenicism in Taiwan.  Jour. Natl. Cancer Inst.  40: 453.
                                     C-150

-------
 Tsutsumi,  S.  and  K. Kato.  1975.   Effects  of  dimercapol  or thioctic acid
on
 the  distribution  and  excretion of  74As  in  rats.   Bull. Tokyo  Dent.  Coll.
 16:  69.
Tsutsumi,  S.t  et al.  1976.  Studies on  the  arsenic metabolism.   Report 18.
On  effects  of various  antidotes  on  the enteral  absorption of  arsenical.
Bull. Tokyo  Dent. Coll.  L7: 73.

Underwood,  E.J.   1977.  Trace  Elements in Human  and -Animal  Nutrition,  4th
ed.  Academic Press, New York.

Urakabo,  G.,  et  al.   1975.   Studies  on the fate of  poisonous metals  in ex-
perimental animal (V).   Body  retention  and  excretion of arsenic.   Jour.  Food
Hyg. Soc. Jap.  16:  334.  (Jap.)

U.S. EPA.   1975.  Chemical  analysis of interstate carrier water  supply  sys-
tems.  EPA 430/9-75-005.  U.S. Environ. Prot. Agency.

U.S. EPA.   1978.   In-depth  studies on health  and  environmental  impacts  of
selected  water pollutants.   Contract  No.  68-01-4646.   U.S.  Environ.  Prot.
Agency.

U.S. EPA.   1980a.   Seafood  consumption  data  analysis.  Stadford  Research
Institute  International,   Menlo   Park,  Calif.,  Final  report,   Task   II.
Contract No. 68-01-3887.
                                    C-ISl

-------
U.S.  EPA.   1980b.   Health  assessment  document  for  Arsenic U.S.  Environ.
Prot. Agency.  External Review Draft.

U.S. Public  Health  Service.   1977.   High  environmental  exposure  to arsenic -
Fairbanks,  Alaska.   Center  for  Disease  Control,  Atlanta.   EPI  76-111-1.
(Unpubl.)

Valentine,  J.L.   1979.  Arsenic  levels  in  human  blood,  urine,  and  hair  in
response to exposure via drinking water.  Environ. Res.  20: 24.

Vallee,  B.I.   1973.   Arsenic  (Air  Quality  Monograph  #73-18).   American
Petroleum Institute, Washington, O.C.

Vallee, B.L., et al.   1960.   Arsenic toxicology  and  biochemistry.   AMA Arch.
Ind. Health.  21: 132.

Wadkins,  C.L.    1961.   Stimulation  of adenosine  triphosphates  activity  of
mitochondria and sub-mitochondrial particles by  arsenate.   Jour.  Biol. Chem.
238: 3300.

Wagoner,  J.K.,  et  al.  1963.   Unusual  cancer  mortality  among  a group  of
underground metal miners.  New Eng. Jour. Med.  269:  284,

Wagstaff, O.J.   1978.   Alteration  of hepatic detoxication  enzyme activity by
dietary arsenic trioxlde.  Food Cosmet. Toxicol.   16: 423.
                                     C-152

-------
 Wahistrom,  R.D.,  et al.   1955.   The  effect  of arsanilic acid and 3-nitro-4-
 hydroxyphenyarsonic  acid  on  selenium poisoning in the pig.  Jour. Anim. Sci.
 14:  105.

 Walker, G.W.S.  and  A.M.  Bradley.   1S6S.   Interacting effects of sodium mono-
 hydrogenarsenate  and  selenocystine  on  crossing  over  in  Qrosphi la  nieiario-
 gaster.  Can. Jour.  Genet. Cytol.  11: 677.

 Webb,  J.L.   1966*   Enzyme  and  Metabolic  Inhibition.   Malonate,  Analogs,
 Dehydroacetate.   Sulfhydryl  Reagent,  s-Iodosobenzoate,  Mercurials.  Academic
 Press, New  york,  2: 1237,
Westoo,  G.  and  M.  Rydalv.   1972.   Arsenic   levels  in  foods,   Var  Foda,
24: 21.  (In Swedish with English summary)

White,  O.E.,  et  al.   1963.   Data  of Geochemistry  (6th  ed.)  Chapter  F.
Chemical Composition  of Subsurface  Waters.   Geological  Survey  Professional
Paper 440-F.  Washington, O.C.: U.S. Government Printing Office,   p.  63.

Wilder,  H.8.   1972.    Investigation  of  the   Occurrence  and  Transport  of
Arsenic  on   Upper  Sugar   Creek   Watershed,  Charlotte,   North   Carolina,
p. 0205-0210.    hi:  Geological  Survey  Research.   Chapter  0.    Geological
Survey  professional   paper  800-0.    Scientific  notes   and   summaries  of
investigations  in  geology,  hydrology and  related  fields.   Washington,  O.C.:
U.S. Government Printing Office.
                                    C-153

-------
Winkler,  W.O.   1962.   Identification  and  estimation  of the  arsenic  residue
in  livers  of  rats  ingesting  arsenicals.   Jour.  Assoc.  Off.  Anal.  Chem.
45: 80.

Woods, J.S.  and  B.A.  Fowler.   1977.  Effects of  chronic arsenic  exposure  on
hematopoietic function  in  adult  mammalian  liver.  Environ. Health  Perspect.
19: 209.

Woolson,  E.A.   1975.   Sioacculation of Arsenical.  lr±:  Arsenical  pesticides
E.A. Woolson, (ed.), Am. Chem. Soc. Washington, O.C.  p. 97.

World  Health  Qrganizaton.   1973.   Trace  elements in  human nutrition.   Rep.
WHO Expert Committee.   WHO  Tech.  Rep.  Ser. No. 532.  World Health  Organiza-
tion, Geneva.

Yamashita,  N.,  et al.   1972.   Current state  of Kyoto children  poisoned  by
arsenic tainted Morinaga dry milk.  Jap. Jour.  Hyg.  27: 364.   (Jap.)
   , S.  1963.  Relative incidence of  skin  cancer  in  Chinese in  Taiwan:  With
special reference to arsenical cancer.   Natl. Cancer Inst.  10:  81.

Yeh, S.  1973.  Skin cancer in chronic  arsenicism.   Human Pathol .  4:  469.

Yen, S.,  et al.   1968.   Arsenical  cancer  of  skin.   Hlstologic study  with
special reference to Bowen's disease.   Cancer.  21: 312.

Yoshikawa,  T., et al.  1960.  Concerning the mass outbreak  of  chronic arse-
nic toxicosis in Niigata Prefecture.   Chiryo.  42:  1739.
                                    c-154

-------
Zaidivar, R.   1974.   Arsenic  contamination of drinking water and  foodstuffs
causing endemic chronic poisoning.  Beitr.  Pathol.  6d.   151;  334.

Zook,  E.G., ct  ai.   1976.   Preliminary survey of selected seafoods  for  mer-
cury,  lead, cadmium, chromium, and arsenic content.  Jour. Agric.  Food Chem.
24.
                                    C-15S

-------
                                   APPENDIX
Mathematical Prediction Model
    Due  to  the stable  copulation  in  a rural area along  the  southwest coast
of  Taiwan,  the data collected  by  Tseng,  et  al.  (1968) may  be  viewed as  a
lifetime  feeding  study where measured  amounts  of arsenic  in well  water  are
consumed  by  a  study population of  40,421  individuals.   Thus, this  data  may
be  used   to  predict the  lifetime  probability of skin  cancer caused  by  the
ingestion of arsenic.
    A model  estimating  the  cancer  rate as  a function of drinking water arse-
nic concentration was generated using  the  information  in  its  published form,
which is  a  summary  of data collected  by the investigators.   If  the original
data had  been  available,  a  more exact mathematical  analysis  would  have been
possible.
    Doll   (1971) has  suggested that  the relationship  between the  incidence of
some site specific  cancers,  age,  and exposure  level  of a population  may be
expressed as:
                             (1)  I(x,t)  - vBxV'1
where x  is  the exposure level of  a carcinogen,  t is the age of  the popula-
tion, and B, m, v are unknown parameters.
    However, the data collected by  Tseng,  et al. (1968) was  obtained  at  one
point in  time, and  since  skin cancer  has only a marginal  effect  on the death
rate, the obtained rates may be viewed  more  accurately  as the probability of
having contracted   skin  cancer  by  time t.   The  relationship  between  this
probability, often  referred  to  as the cummulatfve  probability density  or
prevalence,  and  the  incidence  or  age specific  or  hazard   rate  may  be  ex-
pressed as:
                     (2) F(x,t)  - 1 - exp  [  ("^(x.s) dsl
                                           o'
                                    C-156

-------
     Utilizing  the suggestion  of  Doll  (1971)  for the form  of  the incidence
 rate,  the  prevalence may  be  expressed as:
                         (3)  F(x.t)  » 1  - exp (-8xmtv)
 which  is a  Welbull distribution.
     In  Table  1,  based on  information in  Tseng,  et  al.  (1968),  we have esti
 mates of F(x.t) for different  age and exposure groupings for males.
     To  use  this data,  specific values for x and t had to be obtained for the
 intervals.  Where the  intervals were closed the  midpoint  was  utilized.  For
 the  greater than  0.6  ppm group, the midpoint between 0.6  and  the greatest
 recorded value 1.8 was  taken,  resulting in 1.2 ppm.  For  age  60 or greater,
 a  value of  70 was  utilized  somewhat  arbitrarily,  being the  same increase
 over the lower level as that in the  other two  age intervals.  The values for
 (x,t)  to  relate  to  the  prevalence  estimates are  shown  in  parentheses  in
 Table 1.
    From eauation  (3) it follows that:
               (4) ln(-ln[l-F(x,t)])  - ln(B) *  m ln(x)  *  v  ln(t)
which  is  multiple linear  in  form.    Estimating  the parameters  by  the usual
 least square techniques, we obtained the  relationship:
        (5) ln( -  ln[l - F(x.t)]) » 17.548  * 1.192 ln(x) * 3.881 ln(t)
which  is  an  excellent fit    having  a  multiple  correlation coefficient  of
0.986.
    Eauation (5)  may be expressed as:
             (6)  F(x,t) - 1  -  expC-10"7 x 0.2429x 1>192 t3-881] .
                           - 1 -expC-H  (t)  x 1.192]
 If the  coefficient m  • 1.192 was in fact  equal  to  1,  then for a given value
of t equation (6)  would be "one-hit" in form.
                                     C-157

-------
                         TABLE  1

       Age - Exposure  - Specific  Prevalence  Rates*
ppma
0 - ,29
(0.15)
0.30 - 0.59
(0.450)
>0.6
(1.2)
20-39
(30)
0.0013

0.0043

0.0224

40-59
(50)
0.0065

0.0477

0.0983

>60
770)
0.0481

0.1634

0.2553

*Source:  Tseng, et al. 1968
aRange given.  Midpoint is In parenthesis,
                             c-158

-------
     To  test  this Hypothesis (i.e., Ho: in  «  1)  the  student  "t"  test is used,
 giving  the result:
                                       -1
which  is  r»ot  significant  at  the 0.1  level.  The value 0.138  is  the  standard
e^or  of  m.  Thus  there  is  insufficient  evidence to reject  the  hypothesis
that the  dose response  relationship  is "one-hit" even at the  0.1  level  even
though the standard error of the regression coefficient is quite small.
    Fixing m » 1 we have the relationship:
                         (7)  Ffx.t) . 1 - exp[-g(t)x]
Transforming this equation to its  linear form  and  obtaining  the  least  square
estimates of B and v, we find that:
       g  (t) » exp(-17.5393)  t3'853, where  8 -  2.41423  x  LO"8, v - 3.853
In this case,  the fit is still  quite good  as  represented  by  a correlation  of
0.971.   The data used to obtain  the  estimates  is  shown  in  Table 2 and  the
goodness of fit is illustrated  in Figure 1.
    The  function  F(x,t)  .  1 -exp[ -2. 41423x1 O"8  x t3*853],   is  the  proba-
bility of  contracting skin   cancer  by age  t  given that  a  individual  had  a
life-time exposure to x ppm  in his drinking water (and lived  until  age  t).
    In  Appendix  ! of  the  CAG (1978)  coke oven document,  the  lifetime  proba-
bility of cancer in the presence  of competing  mortality  was  derived  from the
age-specific incidence rate.   For  the  case where the  cancer rate in the  ab-
sence of exposure  is  near zero  (as in this case where the skin  cancer  is  of
3  rats form that was  virtually  unknown in  other  parts of Taiwan) the  life-
time probability may be expressed as:
                              Q(*) * Bx/(8xV)
                                    C-159

-------
                        TABLE 2



Data Utilized to Obtain Predictor Equation and Figure 1
pp»
Arsenic

X

0.15


0.45


1.20


Age at Medical
Examination

t

30
50
70
30
50
70
30
50
70
Skin Cancer Prevalence


Observed
Rate
0.0013
0.0065
0.0481
0.0043
0.0477
0.1634
0.0224
0.0983
0.2553
Rate
F(x,t)
Expected
Rate
0.0031
0.0127
0.0455
0.0053
0.0375
0.1304
0.0141
0.0969
0.3110
Transformed
Skin Cancer
Prevalence Rate
-ln( -ln[l-
-17.5393 + 3

Observed
6.64474
5.03269
3.00993
5.44699
3.01849
1.72368
3.78739
2.26844
1.22155
.8531nt+lnx

Expected
6.33160
4.36341
3.06695
5.23299
3.26480
1.96834
4.25216
2.28397
0.98751
                    C-160

-------
 F(x,t)  -

 0.0009  7.0
                   t-30
 0.0025  6.0
0.0067 5.0
0.0181 4.0
0.0486 3.0
0.1266 2.0
0.3078 1.0
                   U50
0.6321 0.0
         -2.     -1.6     -1.2     -.8     -.4      .0     .4    Ine(ppm)

          0.135   0.202    0.301   0.449   0.670   1.000  1.492  ppm
                                   FIGURE  1

                       Relationship between  Transfonned
               Prevalence  and  log  ppm Arsenic  in  Water,  log  age
                                    C-161

-------
whe»-e  p   »   In   t^,  (where  tffl  is  the  median   lifetime  of  tne  popu-
 lation).   Assuming tm » 68  and v *  3.853, is the same  for  total  mortality
 as  the appearance  of  skin cancer, we have:
               0 t  \        2.41423 x
               2   > "  2.41423  x * 6.02793
The  level  of  x that  results  in a  lifetime probability of  skin  cancer  equal
to  1C"5  is   found by solving  Q (» )  .   10"5  for x  giving  X  «  2.4969  x
10"5 mg/liter  or 0.025 ug/liter.
    Under  the  assumption that  the  average consumption  of  water is  two liters
in both the U.S. and Taiwan we  estimate a water criteria concentration of:
                        2(.025)  - C(2  * 0.0065  x 44) or
                        C -    —1^5	 - 0.0219 ug/1
                                2.2860
Where 0.0065 is the average  fish consumption  in kilograms and  44 is  the bio-
accumulation factor for fish (supplied by Don Mount of U.S. EPA).
    A standard  for waterborne  arsenic  of 22 ng/1  would  thus  insure  a  life-
time risk of cancer of less than 10  .
    It is  recognized  that inorganic and organic compounds differ in  terms  of
toxicity and likely  in terms of carcinogenic  potential.   However, since  the
recommended level  is  to be  based  on  carcinogenic potential and no  informa-
tion is available  concerning the relationship(s) of specific arsenic  species
and cancer a single  all  inclusive limit  must  be set.  Even if the data were
available to permit separate  standards,  the level  of development of  the  re-
quired analytical  methodology  is not  sufficient  to  permit reliable  and  re-
oeatable speciation measurements, a necessity  before  setting a standard (Or.
Irgolic,  Texas AAM University, personal communication).
                                     C-162

-------
     For  comparative purposes, the  Stockinger  and Woodward (1958) method  was
 applied  to the  present and  proposed  airborne arsenic  standards to  compute
 comparable waterborne  arsenic  levels.

     American  Conference of  Governmental  Industrial Hygienists:
 1.   Existing  threshold  limit value  - time-weighted average  - 500  ug/m
     500  ug x  10 m3  x 5  work days x  20X  absorption  m 5 000  g/week
      m3   work day    week
     5,000  ug  x 1 week  x     1     x      Allowed    » 445 ug/i
      week     7 days    Z liters    8055  absorption
     Applying  the recommended safety factor of  100 the comparable  drinking
water limit is 4.46 ug/1.
2.   Proposed  threshold  limit value  - time-weighted average  - 50 ug/m^
     50 ug  x 10 m3 x 5 work days x 2056 absorption ^ ^n  Q/wk
     m3    work day    week
     500 ug x  1 week x     1    x 	 Allowed      . 44.5 vg/]
     week     7 days   2  liters   8056 absorption
     Applying  the recommended safety fator of 100 the comparable drinking
water limit is 0.45 ug/1.
     Occupational  Safety  and Health  Administration:
1.   Eight-hour average  - 10 ug/m3
     10 ug  x 10 m3 x 5 work days x 2055 absorption m 10Q ug/week
    week   work days     week
     100 ug x  1 week x     1    x    Allowed     « 3.93 ug/l
     week     7 days   2  liters   so* aosorption

    Applying  the recommended safety factor of  100, the comparable drinking
water limit is 0.09 ug/1.
                                    C-163

-------
    Assuming  that  the absorption factors  (air.-20  percent,  water-80 percent)
and methods  recommended  by  Stockinger  and  Woodward (1958)  are reasonable and
that  the safety  of  100  is appropriate,  it  is dear  that  the  recommended
water standard  is  even more restrictive  than  the air standards.   The differ-
ences are likely due  at  least  partially  to variations in extrapolation meth-
ods and  levels  of  acceptable risk.
    It  is  of interest to  see  what cancer  risk  would be associated  with  an
air exposure  equivalent  to  the recommended water  standard of  0.02  wg/1.   If
we make  the following assumptions:
    (1)  Total  daily  average absorbed arsenic from water is:
         0.8  x  0.02 (2  + 0.0065  x  44) > 0.0366 ug, where 80  percent is  the
         absorption rate.
    (2)  The  breathing  rate is   1  m  /hr and  20 percent of  the arsenic  is
         absorbed.
Then, the air concentration,  X,   required to  obtain  the  same  absorbed amount
of arsenic is:
                          0.2 x 24 x X  -  0.0366  ug  or
                               X  -  0.008 ug/m3
    From the  1978  CAG report  on  the  risk  associated with airborne  arsenic
the  lifetime  cancer risk associated  with  X  ug/M3  of  arsenic in the  air  is
estimated to be:
                               P » 3.418  x  10"3X

If instead of basing  our risk on the most  sensitive study we use  the geo-
metric mean of  the three studies, the lifetime cancer risk would be:
                               P  .  1.95  x  10'3  X
                                    C-164

-------
The  risks  associated  with   X  .  0.008  are  thus  2.73 x  10"5  and  1.56  x
10"^.  Thus,  if  our water standard  was  based on  the  geometric mean  of  the
human epidemiological  air  studies,  it would  be  0.013  ug/1  instead of  0.02
ug/1, which 1s a  remarkably consistent result.
                                    C-165

-------

                United States
                environmental Protection
                Agency
                Office of Water
                Regulations and Standards
                Criteria and Standards Division
                Washington DC 20460
EPA 440/5-30-021
Octooer 1980
vvEPA
Ambient
Water  Quality
Criteria for
Arsenic

-------
      AMBIENT WATER QUALITY CRITERIA FOR

                 ARSENIC
                 Prepared By
    U.S.  ENVIRONMENTAL PROTECTION AGENCY

  Office of Water Regulations and Standards
       Criteria and Standards Division
              Washington, O.C.

    Office of Research and Development
Environmental Criteria and Assessment Office
              Cincinnati, Ohio

        Carcinogen Assessment Group
             Washington, D.C.

    Environmental Research Laboratories
             Corvalis, Oregon
             Duluth, Minnesota
           Gulf Breeze, Florida
        Narragansett, Rhode Island

-------
                              DISCLAIMER
      This  report  has  been  reviewed by the  Environmental  Criteria  and
Assessment Office, U.S.  Environmental  Protection  Agency,  and approved
for publication.   Mention of trade names or commercial products does  not
constitute endorsement or recommendation for use.
                         AVAILABILITY  NOTICE
      This  document  is available  to  the public through  the  Nationa'
Technical Information Service, (NTIS), Springfield, Virginia  22161.
                                   11

-------
                                FOREWORD

     Section  304  (a)(l) of the  Clean  Water  Act of 1977 (P.L. 95-217),
 requires  the Administrator of  the  Environmental  Protection  Agency to
 publish  criteria for  water  quality  accurately  reflecting the  latest
 scientific knowledge on the kind and extent  of  all  identifiable effects
 on  health and  welfare  which  may  be expected  from the  presence of
 pollutants in any body of water, including ground water.  Proposed water
 quality criteria  for  the  65  toxic pollutants listed under section 307
 (a)(l)  of the Clean  Water Act were  developed  and a notice  of their
 availability was  published for  public comment on March 15, 1979  (44 FR
 15926), July 25,  1979 (44 FR  43660), and  October  1, 1979 (44 FR 56628).
 This  document  is a revision  of those proposed  criteria  based  upon a
 consideration of  comments  received  from  other  Federal  Agencies, State
 agencies,  special interest  groups,  and  individual  scientists.   The
 criteria contained in  this  document replace any  previously published EPA
 criteria  for the 65  pollutants.    This criterion  document  is also
 published in satisifaction of paragraph  11 of the  Settlement Agreement
 in  Natural Resources  Defense  Council, et.  a],  vs.  Train,  8 ERC 2120
 (D.O.C. 1976), modified, 12 ERC 1833  (O.O.C. 1979).

    The term "water  quality  criteria" is used  in  two  sections of the
 Clean Water Act, section 304 (a)(l) and section 303  (c)(2).   The term has
 a different  program impact in  each  section.   In section 304, the term
 represents a  non-regulatory,  scientific  assessment  of  ecological  ef-
 fects. The criteria presented  in  this publication are such scientific
 assessments.    Such water  quality  criteria associated with  specific
 stream uses when  adopted as State water quality standards under section
 303  become enforceable maximum acceptable  levels  of  a  pollutant in
 ambient waters.   The water quality criteria adopted in the State water
quality standards could have the same numerical  limits as the criteria
developed under section 304.  However, in  many situations States may want
 to adjust water quality criteria developed under  section 304 to reflect
 local  environmental   conditions and  human  exposure patterns  before
 incorporation into  water  quality standards.    It  is not  until  their
adoption as part  of the State  water  quality  standards that the criteria
become regulatory.

    Guidelines to  assist  the  States  in  the modification  of criteria
presented  in this  document,   in  the development  of  water  quality
standards, and in  other water-related programs of this Agency, are being
developed by  EPA.
                                    STEVEN SCHATZOW
                                    Deputy Assistant Administrator
                                    Office of Water Regulations and Standards

-------
                             ACKNOWLEDGEMENTS
Aquatic Life Toxicology

   Charles E. Steohan, ERL-Ouluth
   U.S. environmental Protection Agency
John H. Gentile  ERL-Narragansett
U.S. Environmental  Protection Agency
Mammalian Toxicology and Human Health Effects

   Dan Greathouse (author)
   U.S. Environmental Protection Agency
   Oebdas Mukerjee (doc. mgr.), ECAO-Cin
   U.S. Environmental Protection Agency

   Jerry F. Stara (doc. mgr.), ECAO-Cin
   U.S. Environmental Protection Agency

   Jeff Gaba, OGC
   U.S. Environmental Protection Agency

   Paul Hammond
   University of Cincinnati

   Steven 0. Lutkenhoff
   U.S. Environmental Protection Agency

   Robert McGaughy, CAG
   U.S. Environmental Protection Agency

   Ed Wool son
   U.S. Department of Agriculture
Roy E. Albert*
Carcinogen Assessment Group
U.S. Environmental Protection Agency

Thomas Clarkson
University of Rochester

Patrick Durkin
Syracuse Research Corporation

Lester Grant, ECAO-RTP
U.S. Environmental Protection Agency

Terri Laird, ECAO-Cin
U.S. Environmental Protection Agency

Bill Marcus, OOW
U.S. Environmental Protection Agency

Harry Ska!sky
Reynolds Metal Company
Technical Support Services Staff:  D.J. Reisman, M.A. Garlough, B.L. Zwayer,
P.A. Daunt, K.S. Edwards, T.A. Scandura, A.T. Pressley, C.A. Cooper,
M.M. Denessen.

Clerical Staff:   C.A. Haynes, S.J. Faehr, L.A. Wade. D. Jones,  3.J.  Bordicks,
B.J. Quesnell. P. Gray, B. Gardiner.
*CAG Participating Members:
   Elizabeth L.  Anderson, Larry Anderson, Dolph Arnicar, Steven Bayard,
   David L.  Bayliss, Chao w. Chen, John R. Fowle III, Bernard Haberman,
   Charalingayya Hiremath, Chang S.  Lao, Robert McGaughy, Jeffrey Rosen-
   blatt, Oharm  V. Singh, and Todd W.  Thorslund.
                                    w

-------
                                 TABLE OF CONTENTS

                                                                     Page

Criteria Summary

Introduction                                                         A-l

Aquatic Life Toxicology                                              8-1
      Introduction                                                    B-l
      Effects                                                         8-4
          Acute Toxicity                                             B-4
          Chronic Toxicity                                           B-6
          Plant Effects                                              B-7
          Residues                                                   B-7
          Miscellaneous                                              B-8
          Summary                                                    B-10
      Criteria                                                        B-12
      References                                                      B-26

Mammalian Toxicology and Human Health Effects                        C-l
      Exposure                                                        C-l
          Ingestion from Water                                       C-l
          Ingestion from Food                                        C-2
          Inhalation                                                 C-10
          Dermal                                                     C-12
     Pharmacokinetics                                                C-12
          Absorption                                                 C-13
          Distribution                                               C-20
          Metabolism                                                 C-24
          Excretion                                                  C-33
     Effects                                                         C-36
          Acute, Subacute, and Chronic Toxicity                      C-36
          Subacute and Chronic Toxicity                              C-47
          Synergism and/or Antagonism                                C-70
          Teratogenicity                                             C-73
          Mutagenicity                                               C-75
          Carcinogenicity                                            C-78
     Criterion Formulation                                           C-l10
          Existing Guidelines and Standards                          C-110
          Current Levels of Exposure                                 C-110
          Special Groups at Risk                                     C-112
          Basis and Derivation of Criterion                          C-112
     References                                                      C-117
Appendix                                                             C-156

-------
                               CRITERIA DOCUMENT
                                    ARSENIC

                                 Aquatic Life
         ^reshwatef  aauatic  life the concentration of total  recoverable
valent inorganic  arsenic  should  not  exceed  440 ug/1  at  any time.  Short-term
effects  on embryos  and  larvae of aquatic vertebrate  species  have  been  shown
to occur at concentrations as low as 40 ug/1.
    The  available data  for  total  recoverable  trivalent  inorganic arsenic  in-
dicate that  acute  toxicity  to  saltwater aquatic  life  occurs  at  concentra-
tions as low  as  508 ug/1  and would occur at  lower concentrations  among  spe-
cies that  are  more  sensitive than those tested.   No  data  are available  con-
cerning  the  chronic  toxicity of  trivalent  inorganic  arsenic  to  sensitive
saltwater aauatic life.

                                 Human Health
    For  the  maximum protection of  human  health  from the  potential  carcino-
genic effects  due to exposure  of  arsenic  through ingestion  of contaminated
water and  contaminated  aquatic  organisms,  the  ambient water  concentrations
should  be  zero  based  on  the  non-threshold  assumption for  this  chemical.
However,  zero  level  may not  be  attainable  at the present  time.  Therefore,
the levels which  may result  in incremental  increase  of cancer risk  over  the
lifetime   are   estimated   at  10  ,   10  ,    and   10  .    The  corresponding
recommended criteria are 22  ng/1,  2.2  ng/1,  and  0.22  ng/1,  respectively.   If
the above estimates  are made for consumption of  aquatic organisms  only,  ex-
cluding  consumption  of  water, the levels are 175  ng/1,  17.5  ng/1,  and  1.75
na/1, respectively.
                                     vi

-------
                                  INTRODUCTION

     Arsenic  is a naturally occurring  element  often referred to  as  a metal,
 although  chemically classified  as a  metalloid.   Arsenic and  its compounds
 are  used in  the manufacturing of  glass,  cloth,  and  electrical  semiconduc-
 tor,  as  fungicides and wood preservatives, as  growth  stimulants  for plants
 and  animals,  as well as  in  veterinary applications (U.S. EPA,  19765).   The
 United  States  consumes  half  of  the  world  production  of arsenic,  or  about
 37,500  tons  per year,  and produces about  18,000 tons  per year  itself.   The
 principal emission  source for arsenic in  the  United States  is  thought  to be
 coal-^uel power plants, which emit  approximately  3,000 tons of  arsenic  per
 year (Nelson,  1977).
    Environmental concentrations  of arsenic have  been  reported  at 5 mg  per
 kg in the earth's crust  (U.S. EPA,  1976a).   Arsenic is  found also  in  air  and
 in all  living  organisms.   Analysis  of  1,577  U.S.  surface waters  showed  arse
 nic to  be present in 87  samples, with  concentrations  ranging from 5 to  336
 uq/1 and  a  mean  level  of 64 ug/1  (Kopp,  1969).  Bowen  (1966)  reported  3.0
 ud/1 in sea  water.
    A member  of Group  VB  of  the  periodic table,  arsenic  has  five electrons
 in its  outer  shell, giving rise  to the oxidation  states  of *5, +3, 0,  and
 -3.  Arsenic  as  a free  element (0) is  rarely  encountered  in natural  waters.
 Soluble inorganic arsenate (+5)   predominates  under normal  conditions  since
 it  is  thennodynamical ly more  stable  in water  than arsenite  (+3)  (Ferguson
 and Gavis, 1972).   Elemental  arsenic  is  a gray, crystalline material with  a
molecular weight of 74.92, a  density of  5.727,  a melting point  (at 28 atmos-
 pheres) of  817°C,  and   a  boiling  point  (sublime) of  613°C (Weast, 1975).
 The low toxicity  of elemental arsenic is  attributed  to its  virtual  insolu-
 bility in water or in the body fluids  (U.S. EPA, 1976b).
                                     A-l

-------
    A  distinction   should  be  made  between  different  means of  classifying
arsenic compounds.   The  compounds  of arsenic may be  classified  according to
the  oxidation  state  of  arsenic  (As  ",  As  ,   and  As  )  and  according  to
whether or  not  arsenic  is  in the  organic  form  (i.e.,  the arsenic  atom is
covalently attached to at least one carbon atom).
    Conditions  of  low pH,  low  Eh  (standard oxidation-reduction  potential)
and  TOW  dissolved  oxygen  in  water  favor the  formation  of lower  oxidation
state arsenicals such  as arsenite   (+3)  and  arsine  (-3) whereas more  basic,
oxygenated waters result in  an increase  in  the  percentage  of arsenic present
in the oentavalent state.  The reducing  action  of certain  organisms may also
cause arsenite  to  be the predominate  form.   In waters of  high  organic con-
tent, a considerable amount  of arsenic  may  be bound  to colloidal  humic mat-
ter (Ferguson and Gavis,  1972).
    Both  arsenate  and  arsenite can  be  removed  from the water column  by co
precipitation or  adsorption  onto  iron  oxides   (LaPeintre,  1954;   Gupta  and
Ghosh, 1953).  Arsenate  species  can  also  be  removed  by adsorption  onto alum-
inum hydroxide and clays, while  arsenite  is  readily  adsorbed onto  metal sul -
fides (Ferguson and Gavis,  1972).
    Oxidation of arsenite to  arsenate occurs  slowly  at neutral  pH  (faster in
strongly acid or alkaline solutions), while  methylation of arsenic to methyl
and dimethylarsine  by  methanogenic bacteria  is  known to  occur  (McBride and
Wolfe, 1971).
    Arsine  (AsH-j)  and  its  methyl  derivatives   are   the  most  acutely  toxic
compounds of  arsenic.   However,  they do  not occur  in  drinking  water  or in
ambient water.  Human exposure has  occurred  only through  generation of these
compounds in  occupational settings.   Thus,  arsine compounds are not further
considerated in this document.
                                      A-2

-------
    The  organic'   '"hich are not  naturally  occurring,  are the  largest group
of arsenic compounds.   The  two  most  common  organic  arsenic compounds are the
arsonic  adds,  R-AsO-(OH)2,  and  the  arsinic  acids,   R,R'-AsO-OH,  where  R
and R1 refer  to  a  variety  of  organic (alkyl) groups (U.S. EPA,  1976b).   The
organic  arsenic  compounds   considered to  be of environmental  importance  are
those containing methyl  groups, the  aromatic arsenic derivatives  employed as
feed  additives and in veterinary  medicine,  and others which  may  have  impor-
tance in biological systems (U.S.  EPA, 1976a).
    Arsenic forms  a  complete  series  of trihalides,  while  arsenic  (V)  fluor-
ide is  the only simple  pentahalide  known.   All of  the  arsenic  halides  are
covalent compounds that  hydrolyze in the presence of water  (Standen,  1967).
Additional information on inorganic arsenic  compounds is given in  Table 1.
                                     A-3

-------
                                                   TABLE  1

                               Properties  of  Some Inorganic Arsenic Compounds*
     Compound
Formula
  Hater Solubility
 Specific Properties
Arsenic trloxlde
Arsenic pentoxide
Arsenic hydride
   (arsine)
Arsenic (HI) sulfide
Arsenic sulfide
Arsenic (V) sulfide
As203
As?05
12 x 10* wg/1 • O'C
21 x 106 wg/l • 25'C
2,300 x 10* Mg/l • 20"C
                   20 Ml/100 g
                   cold water
                   520 »g/l • 18*C
                   1.400 i.g/1 O O'C
Dissolves in water to
form arsenious acid

(H3As03:
K . 8 xlO-10H 25"C)

Dissolves in water to
form arsenic acid
                                                            2.5  x  10~4
                                                            5.6  x  10
                                                            3  x  10-13)
                                                                                 *3
                                    This  compound and  Its
                                    methyl  derivatives are
                                    considered  to be the
                                    most  toxic.

                                    Burns in  air forming
                                    arsenic trioxide and
                                    sulfur  dioxide; occurs
                                    naturally as orpiment.

                                    Occurs  naturally as
                                    Realgar.
*Source:  Standen, 1967; U.S. EPA, 1976a,b
                                          A-4

-------
                                  REFERENCES

Alderdice, D.F.  and  F.FL  Brett.   1957.   Toxicity of sodium arsenite to young
chum  salmon.  Prog.  Sep. Pacific Coast Stat. Fish. Res. Board Can.  108: 27.

Ancel,  P.   1946.  Recherche experimental e . sur  le  spina  bifida.   Arch. Anat.
vier. M0rph. Exo.  36: 45.

Anderson,  3.G.    1946.    The  toxicity  thresholds  of  various  sodium  salts
determined by the use of Daphnia magna.  Sewage Works Jour.  18:  82.
Beaudoin,  A.R.    1974.   Teratogenicity  of  sodium  arsenate  in  rats.
toloay.  10: 153.
Biesinger, K.E.  and  C-.M. Christensen.   1972.   Effects of various  metals  on
survival,  growth,  reproduction,  and  metabolism  of   Oaphm'a  magna.   Jour.
Fish. Res. Board Can.  29: 1691.

Bowen,  H.J.M.   1966.   Trace  Elements   in  Biochemistry.   Academic  Press,
London -New York.

Browning, E.   1961.  Toxicity of Industrial Metals.  Buttersworth, London.

Calabrese, A., et al.  1973.  The toxicity of  heavy metals  to  embryos of t.he
American oyster, Crassostrea virginica.  Mar. Biol.  18: 162.
                                     A-5

-------
Car-dwell,  3-D.,  et al.   1976.   Acute toxicity of  selected  to-      :  to six
soecies  of *ish.   Ecol.  Res.  Ser.   EPA  600/3-76-008.   U.S.  •      1.  Prot.
Aqency, Washington, D.C.

Clemens,  H.P.  and  K.E.  Sneed.   1959.   Lethal doses  of  several  commercial
chemicals  for  fingering  channel catfish.  U.S. Fish Wildl.  Ser.  Spec.  Sci.
Sep. Fish. No. 316.  U.S. Dep. Inter., Washington,  O.C.

Ferguson,  J.F. and  J.  Gavis.   1972.   A  review of the arsenic  cycle  in  natu-
ral waters.  Water Res.  6:  1259.

Fenn, V.H. and  S.J.  Carpenter.  1968.  Malformation induced  by  sodium  arse-
nate.  Jour.  Reprod. Fertil.  17: 199.

Gilderhus, P.A.   1966.  Some  effects of sublethal concentrations of  sodium
arsenite  on  bluegills  and  the aquatic environment.  Trans.  Am.  Fish.  Soc.
95: 289.

Guota,  S.R. and  S.  Ghosh.   1953.  Precipitation of brown  and yellow hydrous
iron oxide.  III. Adsorption of arsenious acids.  Kolloid-Z.  132:  141.

Hood, R.D. and  S.L. Bishop.   1972.   Teratogenic  effects  of  sodium  arsenate
in mice.  Arch. Environ. Health.   24: 62.

Hughes,  J.S.  and J.T.  Davis.   1967.   Effects  of Selected Herbicides  on  Blue-
gill Sunfish.   In: Proc.  18th  Annu.  Conf.,   S.E.  Assoc.  Game  Fish  Comm.,
October   13-21,  1964.   Clearwater,   Florida.   S.E. Assoc.  Game  Fish  Comm.,
Columbia, South Carolina,   p.  480.
                                     A-6

-------
 KODP,  J«p«   1969,   The  Occurrence of  Trace Elements  in Water,   In: o.D.
 Hemphill  fed.}.  Proc,   3rd  Annu.  Conf. Trace  Substances  in  Environ,  Health.
 University of Missourif Columbia,

 LaPeintre,  M.    1954.   Solubilization  oar  les eaux natjrelles  de 1'arsenic
 lie' au fer dans les roches sedimentaires.  C.R. Acad.  Sci.  239: 359.

 McBride,  B.C.   and  R.C.  Wolfe.   1971.   Biosynthesis  of dimethylarsine  by
 Vethanobacterium.  Biochem. Jour.   10: 4312.

 Nelson, D.A., et al.   1976.  Biological effects of  heavy metals on juvenile
 bay scallops, Argooecten  irradians,  in short-term  exposures.   Bull. Environ.
 Contam. Toxicol.  16:  275.

 Nelson, K.W.  1977.  Industrial  contributions  of arsenic to  the environment.
 Environ. Health Persoect.   19:  31.

 Ridgeway,   L.P.   and  O.A.  Karnovsky.   1952.   The  effects of  metals  on  the
 chick  embryo: Toxicity  and  production  of  abnormalities  in development.  Ann.
N.v. Acad.  Sci.   5:  203.

Sanders, H.O.  and  O.B.  Cope.   1968.   The relative  toxicities of  several
oesticides  to  naiads  of  three  species  of  stoneflies.   limnol.  Oceanogr.
13: 112.

Sorenson,  E.^.B.  1976.   Toxicity  and  accumulation of arsenic  in  green sun-
fish,  Lepomis cyanellus,  exposed  to arsenate  in water.   Bull.  Environ. Con-
tarn. Toxicol.   15:  756.
                                     A-7

-------
Soehar, R.L.   Comparative  toxicity of arsenic  compounds  and  their accumula-
tion in invertebrates and fish.  Manuscript.

Standen, A.  fed.)    1967.   Kirk-Othmer  Encyclopedia of Chemical  Technology.
Interscience Pub.,  New York.

Tseng, W.P., et  al.   1968.   Prevalence of  skin cancer  in  an  endemic  area  of
chronic arsenicism in Taiwan.  Jour. Natl. Cancer  Inst.   40:  453.

U.S. EPA.  1976a.  Arsenic.  Subcommittee on Arsenic, Com. on Med.  and Biol.
Effects of Environ.  Pollut.  NRC/NAS, EPA 600/1-76-036.   U.S.  Environ.  Prot.
Agency, Washington, O.C.

U.S. EPA.  1976b.  Arsenic  and  its  compounds.   EPA  560/6-76-016.   U.S.  Envi-
ron. Prot.  Agency,  Washington,  O.C.

Weast, P.C.  (ed.)   1975.   Handbook of Chemistry and Physics.   56th  ed.   CRC
Press, Cleveland, Ohio.

-------
Aquatic Life Toxicology*
                                  INTRODUCTION
    Arsenical  compounds  are  found in all  living organisms including those  in
aquatic  systems.   Althougn  important  sources  of arsenic  in  the environment
are  industrial  (Nelson,  1977; Fowler,  1977),  such  as  smelters of nonferrous
ores  and  coal-fired  power  plants  using arsenic-rich coal, suostantial  arsen-
ic contamination  of  water can also  occur  from the improper use of arsenical
pesticides  such  as   sodium   arsenite  which  is  often  used   as  an  aquatic
herbicide.
    The chemistry of arsenic is  quite  complex,  consisting of chemical, bio-
chemical,  and  geochemical  reactions which  together  control  the  amount  of
dissolved arsenic concentrations  in  aquatic  systems.  A cycle for arsenic  in
natural waters  has been diagrammed  in an extensive review  by  Ferguson and
Gavis  (1972).
    Arsenic  is  stable in water  in  four oxidation  states  (+5,  +3,  0,  -3)  as
both  inorganic and   organometallic  species  and  in  dissolved  and  gaseous
states.   Common  arsenic  species  are  arsenate,  arsenite, methanearsonic acid
and dimethyl arsenic  acid  (cacodylic  acid).   Arsenic as the free element (0)
is rarely encountered in water  but  appears to be thermodynamically stable at
lower  Eh  (standard oxidation reduction  potential)  values.  At  very  low Eh,
AsH-j  (arsine,  -3) may be formed which  is  only slightly soluble.   Arsenic
sulfides have low  solubilities and occur  as  stable  solids at pH values below
5.5 and lower Eh  conditions.  Arsenite (+3)  may also be present if the Eh is
*The reader  is  referred to the Guidelines  for  Deriving Water Quality Crite-
ria for  the  Protection of Aquatic Life  and Its Uses in order  to  better un-
derstand  the  following discussion and recommendation.   The following tables
contain  the  appropriate  data  that were  found  in  the literature,  and  at the
bottom of  the appropriate table  are  calculations for  deriving various  mea-
sures of toxicity as described in the Guidelines.
                                      8-1

-------
 less  than  0.1V.  Arsenic  (+3)  has a strong  affinity for sulfur and  reaaily
 adsorbs or coprecipitates with metal sulfides.
     In  aerobic  water,  reduced forms of  arsenic  tend to be oxidized to arse-
 nate  (*5),  the  predominant form in these  waters.   The rate  of  oxidation of
 arsenite to  arsenate by  oxygen  is  slow  at neutral pH, but proceeds measurab-
 ly  in  several  days  in  strongly  alkaline or  acidic  conditions  (Ferguson  and
 Gavis,  1972).   This oxidation,  however, probably never  proceeds  to  comple-
 tion.   Arsenate can coprecipitate  with  or adsorb on  nydrous  iron  oxides  and
 form  insoluble  precipitates with calcium,  sulfur, aluminum,  and oarium com-
 pounds  (Holm, et al. 1979).   Arsenate  is chemically  similar to pnosphate  and
 may  be  enriched in  phosphate minerals,   although  arsenic  affinity  to  iron is
 predominant.  The  adsorption of arsenate  by  metal  oxides  and  the  formation
 of  arsenic  sulfide  appears to remove arsenic from  solution to the sediments
 and  prevent nigh  arsenic  concentrations   from  being  present   in  solution.
 Studies by  Holm,  et al.  (1979)  and others  on the heterogeneous interactions
of  arsenic  in  aquatic  systems  indicate  that  arsenate  is more  strongly  ab-
 sorbed  to  sediments  than  are  other arsenic forms.   Generally,  adsorption
 processes are very  dependent  on  arsenic  concentration,  sediment characteris-
 tics,  pH,  and  ionic concentration of other  compounds.  Arsenic can  be  re-
moved from the sediments by volatilization and recycled in the water.
    Inorganic arsenic  can  be  converted  to organic  alkyl arsenic  acids  (+3
and +5) and  to methylated  arsines  (-3) under  anaerobic conditions  by  fungus,
yeasts  and  bacteria,  although biomethylation  may occur  under aerobic condi-
tions as well.
    Little  is known about  the mechanism of  arsenic  toxicity to aquatic  or-
                                      8-2

-------
ganisms;  however,  arsenic  readily  forms kinetically stable  bonds  to sulfur
and  carbon in  organic compounds.   Like mercury,  arsenic  (+3)  reacts  with
sulfhydryl  groups  of  proteins;  enzyme  inhibition  by  tnis  mechanism  may be
the  primary  mode  of  arsenic  toxicity.   Arsenate  does   not  react  with
sulfhydryl  groups  as  readily  but  may  uncouple  oxidative  pnosphorylation
(Anderson,  1979).
    Although  considerable  information has  been  published on the effects of
arsenic on  freshwater  organisms, knowledge  of  its  toxicity  is  less  than  com-
plete since much of  the work  has  been devoted to monitoring or field assess-
ment studies  or  were  studies  that contained information  that was  not  useful
for  deriving  a  water  quality criterion.   Virtually  no data on  chronic  ef-
fects of  arsenic on fish  species  exist, and  only one  invertebrate  chronic
test was  found acceptable.  Only  two  references  dealing with  arsenic biocon-
centration by freshwater species accurately reported  exposure  concentrations
or calculated useful bioconcentration factors.
    The arsenic  data  base  for  saltwater organisms  is  inadequate  to  assess
comparative sensitivity among a variety  of organisms  and their  life  stages
or to assess  the importance of  water quality parameters  sucn as  salinity to
arsenic toxicity.   In  addition,  these  data do  not  distinguish differences,
if any,  among various oxidation states.
    The present  data  base  for arsenic is separated  into  trivalent  inorganic
arsenic,  pentavalent   inorganic arsenic,  and  other  arsenic compounds  since
the majority  of  toxicity  tests were conducted with  the trivalent form,  par-
ticularly  sodium arsenite,  and  because toxicity  may  be related to  the  form
of arsenic  present  in  solution.  All  results are expressed in terms  of  ar-
senic,  not as the compound.
                                     B-3

-------
                                    EFFECTS
Acute Toxicity
    Seven acute  tests with  freshwater  invertebrate species have been report-
ed with trivalent  inorganic  arsenic  ana  all  were testea with sodium arsenite
(Table 1).   Only one {U.S.   EPA, 1980) was a flow-through test with measured
concentrations;  the  others were  static tests with unmeasured concentrations.
Crustaceans, comprised of three  cladoceran and one scud species,  showed some
variation among  species but  were more  than four  times  more  sensitive  than a
stonefly species,  an aquatic insect.  Tne range  of  acute values  for crusta-
ceans was 812 to 5,278 ug/1.  Daphnia magna  appeared to be the most tolerant
cladoceran  although  it was  difficult to  compare sensitivities  due  to  the
small data  base.   All  crustacean species  were more  than  twice as  sensitive
to  trivalent  inorganic  arsenic  as  were the fish species  tested.   Stonefly
sensitivity was  within  the   range of sensitivity of  fish  based on  12  acute
tests with fish.
    The acute  toxicity  of  trivalent inorganic arsenic  to freshwater  fishes
is also summarized  in  Table 1.   One-half  of the tests were  static  with  un-
measured concentrations and  the  others were flow-through tests with measured
concentrations.  Seven fish  species  are  represented and  sodium arsenite  was
used  in  all tests.   Rainbow trout  and  brook trout were  the most  sensitive
species  and bluegills  were  the most  tolerant.  The  total  range of  LCcQ
values was narrow for the seven  species  (13,340  to  41,760 ug/1).   The three
values reported  for  bluegills by Inglis  and Oavis (1972) were for tests con-
ducted in  soft,  medium and  hard water  (50,  200» and  370  mg/1  as  CaCO^,  re-
spectively).   No significant difference was  demonstrated to  indicate  that
hardness had any effect on arsenic toxicity.
                                      3-4

-------
    Values  reported for  two  invertebrate and  four  fish species  exposed to
pentavalent  inorganic  arsenic  and  other arsenic  comoounds  are  listed  in
Table  1.   All  tests  were static  with  unmeasured concentrations  except for
one  flow-through,   measured  test  with  rainbow  trout  and  sodium  arsenate.
Values  reported  for  Daphnia  magna  and  rainbow  trout exposed  to  sodium
arsenate  (*5)  are  comparable  to those  for  exposures with these  species and
sodium  arseoite (+3K  Although this data  base is  limited,  the  two  valence
States appear to be  similarly toxic.  The extremely  high values  reaorted for
crayfish,  channel   catfish,  and   smallmouth  bass   exposed   to  monosodium
methanearsonate  indicate  that  organic  arsenic  may  be  much  less  toxic  than
both  trivalent  and  pentavalent  inorganic arsenic.   The 96-hour  LC^«  value
of 82.400  ug/1  shown for  fathead  minnows and  arsenic  trisulfide  (Table  6)
was aooroximately  5  times higher than the value  for this  species exposed to
sodium arsenite  (Table  1).  This  is  probably because  arsenic  trisulfide  is
less soluble than sodium arsenite.
    Based on  the  above  data base,  the Freshwater  Final Acute  Value for tri-
valent inorganic arsenic,  based on calculations described  in  the Guidelines,
is 440 ug/1 (Table 3).
    Acute  toxicity data  representative  of  trivalent   inorganic  arsenic  and
saltwater aauatic  life  are limited to two  fish and three  invertebrate  spe-
cies (Table 1).   Nelson,  et al. (1976)  employed  a renewal  test  to determine
the toxicity  of  sodium  arsenite to juvenile bay  scallop, and  Calabrese,  et
al. (1973) evaluated  toxicity to American oyster  embryos using  sodium  arse-
nite in static  tests.   Toxicity  tests with  unmeasured  concentrations  defined
                                      B-5

-------
a  96-hour  LC^  Of  3,490  ug/1  for  bay  scallops  and  a  48-hour  LC5Q  of
7,500  ug/l  for  American  oyster.   The  lowest  arsenic  acute value  reported
(508  ug/1)  was  for  a  cooepod  tested  in static  toxicity tests with  sodium
arsenite.
    Alderdice and Brett  (1957)  assessed the toxicity of  arsenic  trioxide  to
chum  salmon using  a  renewal  test with unmeasured  concentration  to determine
the  48-hour LCgg  of 8,330  ug/1  (Table  6).   The 96-hour  LC5Q  values  for
arsenic  for  the fourspine stickleback  and  Atlantic  silverside were  deter-
mined to be 15,000 and  16,000 ug/1, respectively (Table  1).
    Toxicity of  arsenic  trisulfide  to  juvenile white  shrimp  was tes  ?d  by
Curtis,  et  al.  (1979).   A 96-hour  LCgQ  of 24,700  ug/1  was determined  for
this  less soluble form of the element  in  static  tests (Table 6).   No compar-
able  data  are available with  this species for  any other form  of trivalent
arsenic, but the highest of the five available values is 16,033  wg/1.
Chronic Toxidty
    Only  one  chronic test was  reported  that  could be  used to calculate  a
chronic  value  for arsenic and  freshwater  aauatic organisms.   A  life-cycle
test  with  Daphnia magna (U.S.  EPA,  1980) (Table  2) exposed to  sodium arse-
nite  resulted in  a chronic value of 912  ug/1  based on chronic  limits  of  633
and  1,315  ug/1.   A  life-cycle  test with the same  species  (Table  6) exposed
to sodium arsenate could  not be used  in  the  calculation of a  chronic value
because the test concentrations  were not  measured  as  specified  in the Guide-
lines.  However,  the upper and  lower chronic  limits  in this test,  based  on
reproduction  growth, and  enzyme  Inhibition  were  nearly identical   (520  to
1,400 ug/1) to  that  reported  above for Daphnia magna.   Both  tests were con-
                                      8-6

-------
ducted  in  Lake  Superior water.   The similar toxicity reported in these tests
using  trivalent  and oentavalent inorganic arsenic  suggests  that  these forms
are  similarly toxic as  was noted previously with acute tests.
     Because  less  than  the  required  number of  chronic tests were reported for
arsenic according  to  the  Guidelines,  a Freshwater  Final  Chronic  Value cannot
he calculated.
     No  data  are  available  on the  chronic  toxicity of  arsenic  to  saltwater
fish or invertebrate species.
Plant Effects
     The effect  of trivalent  inorganic arsenic on  three  species  of  algae and
one  submerged plant are reported  in Table 4.   All  tests were  conducted  with
sodium  arsenite  (Cowell,  1965).  The  sensitivity  of aouatic  plants  is  com-
parable to that for sensitive invertebrate species exposed  in acute  tests.
     No data are available for saltwater algae or vascular plants.
Residues
     Bioconcentration factors  for  freshwater organisms and  arsenic  are shown
in Table 5.  Values were obtained for  four  invertebrate  and  two  fish species
for  trivalent  inorganic  arsenic  compounds.   Six  species  were tested  with
other arsenic compounds.   Numerous  other studies  reporting  bioconcentration
factors for  aouatic organisms  were  not  used since they did not   meet  the
requirements  described  in the Guidelines.
     In the study  by Spehar,  et  al.  (1980),  arsenic was tested to compare tne
bioconcentration  of  four arsenic compounds  after  approximately  28  days  of
exposure.   Results  indicated that Oaphnia magna  and one  snail, Helisoma  cam-
panulata,  had the highest residues when  exposed to trivalent  inorganic  ar-
senic.  Another snail  species,  Stagnicola  emarginata,  and  stoneflies exposed
to trivalent  inorganic arsenic had  values similar  to organisms exposed  :?
                                     3-7

-------
the  other  compounds.  Sioconcentration  factors  for rainbow  trout  and scuds
were reported as zero because  residues  in  exposed  animals were not different
from those  in the controls.   No  value was reported for  scuds and trivalent
inorganic  arsenic  (arsenic  trioxide)  because the  concentration  tested  was
lethal  after 2 weeks of exposure.
    A  bioconcentration  factor of  4  was obtained  for  bluegills  and  arsenic
trioxide in another  study  (U.S.  EPA,  1978).  The half-life  in bluegill  tis-
sue was one day.   The  low  bioconcentration and  short half-life of arsenic in
fish tissue  suggest that  no  residue  problem will  occur at concentrations
that are not directly toxic.
    A  bioconcentration  factor  of  350 was  obtained  for the  oyster,  Crassos-
trea virqinica, after 112 days of exposure (U.S.  EPA.,  1980b)
    No  Residue  Limited  Toxicant  Concentration  (RLTC)  for arsenic could be
determined since no  maximum permissible tissue concentration  for  arsenic is
available.
Miscellaneous
    Data on  other  toxicological  effects  show that  there  is  a wide  range of
sensitivity of  freshwater   invertebrate  and fish species  to  arsenic  (Table
6).  Comparison of these data for fish  with  the fish  acute  values (Table 1)
indicates that 1n  almost all  cases,  arsenic  toxicity was  increased  with  in-
creased duration  of exposure.   One  value  for  bluegills  (Hughes  and  Davis,
1967)  was an  exception  resulting  in   a  48-hour  LCg* of  290  ug/1.   This
value was Included in the document because it was  verified by the author and
because there  was  no reason  to  exclude the  data.   A  specialized pelletized
form of sodium arsenite was used  which  may have accounted for its high tox-
icity.  The  invertebrate data were too  variable to indicate a trend in tox-
icity  in regard to duration of exposure.
                                      B-3

-------
     Not  enough  data were obtained to compare the toxicity of  trivalent  inor-
 ganic  arsenic  to that of other arsenic compounds.  As  with  the acute tests,
 me  trivalent  form,  particularly  sodium  arsenite,  was  the  compound  most
 extensively  tested.
     T-emperature  was the only  variable  tested to determine  effects  of envi-
 ronmental  factors  on  arsenic  toxicity to  freshwater   organisms.   Sorenson
 (1976c)  found  that  increased  water temperature decreased the median letnal
 time of  green  sunfish after exposure  to  two concentrations of  sodium arse-
 nate (Table  6).
    Generally,  the  lowest  freshwater   toxicity  values  for  arsenic  were OD-
 tained in  exposures witn early life stages  of  fish.    Values  for  early  life
 stage  exposures  with  rainbow  trout  and  goldfish  embryos and  larvae were sev-
 eral  times  lower  than  those  for older  juvenile  stages  of  these, species.
 Data for  bluegills showed  that fingerling stages exposed  to sodium arsenite
 were more  sensitive than juveniles and adults  of this  species.   Acute  data
 {Table 1)  also  showed that channel catfish fingerlings were  slightly  more
 sensitive  than juvenile  stages  exposed  to  sodium arsenite.   The lowest value
 obtained  for all  of  the arsenic  data  was  for  an early  life  stage exposure
with the toad which resulted in a 7-day LC50 of 40 ug/1   (Birge, 1979).
    Values obtained  for  early  life  stages  of fish  species  were  lower  than
 those  obtained   for  the  most  sensitive invertebrate species  (Table  1)  and
were below the  limits obtained  for  a life cycle test with  Daphnia  magna
 (Table 2).
    Bryan  (1976)  exposed the  saltwater  polychaete  worm, Nereis diversicolor,
 to  sodium  arsenite  and estimated  the  192-hour  LCcQ  to  be greater  than
 14,500 ug/l  (Table 6).
                                      B-9

-------
    Sodium  arsenite  caused  other  effects  which include depressea oxygen con-
sumption rate  and  behavioral  changes  in  mud  snails exposed to  sodium arse-
nite at concentrations  of arsenic  greater than 2,000 ug/1 for 72 hours (Mac-
Innes  and  Thurberg,  1973)  and  arrested  development  of red  alga  sporelings
following  exposure  to  577  ug/l  for 18 hours  and a post-exposure  period  of
seven days  (Soney, et al. 1959).
    Holland, et al.  (1960)  determined  tolerance  levels  of pink  salmon to ar-
senic  trioxide  and  determined  a  96-hour LC1QQ of  12,307 ug/1;   a  7-day
LC100 of 7'195 u9/1;  and a 10~day LC54 of 3»787 u9/]<
    The Dioconcentration factor of  15,  calculated from  Nelson,  et  al. (1976)
for the  bay scallop after  only  a  4-day exposure, has  been  included  for in-
formational value.
Summary
    The  chemistry  of arsenic  in water  is complex  and the form  present  in
solution  is dependent   on such  environmental   conditions  as Eh, pH,  organic
content, presence  of suspended solids, and sediment characteristics.  Based
on  freshwater  data,  trivalent inorganic  arsenic  (with  the  exception  of ar-
senic  trisulfide)  and  the pentavalent  form  appear to  be   imilarly toxic  to
aquatic  organisms.   Organic  arsenic  compounds and  arser :  trisulfide  were
much  less   toxic  but additional  data  are needed  to  adequately  determined
their effect on aquatic life.
    Acute  data for  14  freshwater  species show that differences  in toxicity
were  not   related  to the  type  of exposure   (i.e.,  static or  flow-through
tests).   Acute values  for  trivalent  inorganic   arsenic  ranged  from  312  to
41,760 ug/1.   A life cycle  test  was conducted with  Daphnia  magna  which gave
a chronic  value  of  912 ug/1.  No  chronic tests  with freshwater fish species
were reported.
                                     3-10

-------
    The  freshwater  residue data indicate tnat arsenic is not  bioconcentrated
 to  a  high  degree  and that lower foms of  aouatic  life may accumulate Mgher
 arsenic  '•esidues  than  fishes.   Arsenic  accumulation in  freshwater aiuatic
 oraam'sms does not  apoear  to  be  qreat'y  affected  by the  form of arsenic ore-
 sent,  although  the highest residues  were  seen in  exposures  with  the triva-
 
-------
                                   CRITERIA
    For freshwater aquatic  life  the concentration of total  recoverable  tri-
valent inorganic arsenic should  not  exceed  440  ug/1  at  any time.   Short-tern
effects on embryos and  larvae  of anuatic vertebrate soecies nave  been  shown
to occur at concentrations  as low as 40 ug/1
    The available data for total  recoverable  trivalent  inorganic  arsenic  in-
dicate that  acute  toxicity  to  saltwater aquatic  life  occurs   at  concentra-
tions as low as 508 ug/1 and would  occur at  lower concentrations  among  soe-
cies that  are  rcore  sensitive than those tested.   No data  are  available  con-
cerrnnq the  chronic   toxicity  of trivalent  inorganic   arsenic  to  sensitive
saltwater aouatic life.
                                     3-12

-------
                                    Table I.  Acute values for arsealc
Species
Method*
Cheatcat
LCM/EC90
Specie* Neaa
Acute Value
Reference

FRESHWATER SPECIES
Cladoceran.
Oaphnla aagaa
Cladoceran,
Oaphnla pulax
Cladoceran,
Oaphala pulax
Cladoceran,
SlMocephalus serrulatus
Scud.
GeaMrus pseudol lanaeus
Stonefly.
Ptaroaarcys call torn lea
Stonefly.
Ptaronarcys call torn lea
Rainbow trout,
Satan galrdnari
Brook trout,
Salvellnus tontlnalls
Goldfish (Juvenile),
Car ass 1 us auratus
Fathead Minnow
(Juvenile),
Channel catfish
(Juvenile).
S. U
s. u
s. u
s. u
FT, M
s. u
S, U
s. u
FT, M
FT, N
FT, M
FT, M
Trlvalent
Sodlua
arsanlte
SodliM
arsanlte
Sodliw
arsanlte
Sodlu*
arsenlte
SodliM
arsanlta
SodliM
arsanlte
SodliM
arsanlta
SodliM
arsanlta
Sod lim
arsanlta
SodliM
arsanlta
Sodlua
ar sen! ta
SodliM
arsanlte
Inorganic Arsenic
5,278
1,044
1,740
812
879
22.040
22.040
13,340
14,964
26.042
15,660
18.096
5,278
1.348
812
879
22,040
13,340
14,964
26,042
15.660
Anderson, 1946
Sanders A Cope,
FPRL, I960
Sanders & Cope,
U.S. EPA, 1980a
Sanders & Cope.
FPRL. I960
FPRL, 1980
Car dwell, et al.
Car dwell, et al.
Cardwell, et al.
Cardwell. et al.
1966
1966
1968
1976
1976
1976
1976
Ictalurus punctatus
                                                        B-13

-------
Table  I.  (Continued)
Species Hethod*
Channel catfish S, U
(tlngerl Ing),
Ictalurus punctatus
FlagtUh (fry). FT. H
Jordanella I lor Ida*
eiuegill (juvenile),, FT, H
Lepoals mecrochlrus
Bluoglll. S, U
LepoMls wacrochlrus
Blueglll. S, U
lepoals Mcrochlrus
Blueglll, S, U
L«po*ls Mcrochlrus
Blueglll, S, U
LepoMlt Mcrochlrus
Cladoceran, S, U
Oaphnla maq/na
Rainbow trout (Juvenile), FT, M
Sat ao ^alrdaerl
Crayfish. S. U
Proca«barufc sp.
Channel cattish, S, U
Ictalurus punctatus
Chemical
SodliM
Sodlu«
arsenlt*
SodliM
ar sen 1 te
SodliM
arsenl te
SodliM
arsanlte
SodliM
arsenl te
SodliM
arsenite
Pentavalent
LC50AC50
(»fl/l)M
15.022
28.130
41.760
15,370
16,240
15,486
17,400
Inorganic Arsenic
SodliM 7,400
ar senate
Sodlua 10.800
ar senate
Other Arsenic Compounds
HooosodliM
•at hanear sooa t e
HooosodliM
•etnanearsonat*
506,000
1.403.000
Species Mean
Acute Value
(ug/l)"*
18,096
28.1)0
41,760
7.400
tO. 800
506,000
1,403,000
Reference

Clttftuits & Sneed, 1959
CardMul 1. et at. 1976
Card*«l 1, et al. 1976
Inylls 1 Udvls, 1972
Inglls 1 Udvls. 1972
Inglls A Davis. 19/2
FPRL, I960
Bleslnger &
Clrlstensun. 1972
Hale. 1977
Anderson, et al. 1975
Anderson, et al. 1975
                                                   B-14

-------
Table I.   (Continual)
Species Method8
Sinai inouth bass S. U
( Hngarl Ing),
Micropterus doioaiMj
Bay scallop (Juvenile), R, U
Argopecten I r radians
Aserlcsn oyster, S, U
Crassostrea virgin lea
Copepod, S, U
Acartla clausl
Foursplne stickleback, S, U
Ape It as guadracus
Atlantic silver si da S. U
(Juvenile),
Menldla eanldla

LC50AC50 Acute Value
Cheat cai <|ig/)i" CM9/i)BB Reference
Monoiodiu- 4i4,000 414,000 Andar&on, or a\ . 1975
•at hanear sona te
SALTWATER SPECIES
Trlvalont Inorganic Arsenic
Sodltm 3,490 3,490 Nelson, et al. 1976
arseftite
Sod Jus 7,300 7,500 Calebruse, el -!,
arsenlte 1973
SodluM 506 506 U.S. LPA, I980b
arsenlte
SodluM 14,953 15,000 U.S. EHA, I960b
arsenlte
Sodlu* 16,033 16,000 U.S. LPA, I980b
arsenl te
* S • static,  R -  reneMal, FT • flow-through, U » unmeasured, H  *> Measured



•"KesulTs are expressed  as arsenic, not as the compound.
                                                     3-15

-------
                                 TabU 2.  Chronic values for arsenic
Specie*
Test1 Chemical
Limit*
(Hfl/l)11
Chronic Value
(HO/1 )" Reference
FRESHWATER SPECIES
Cladoceran,
Daphnla Mgna

Trlvalent Inorganic
LC Sodium
arsenlte
Arsenic
633-1.315
912 U.S. EPA. I960a
• LC » IIU cycle or  partial  life cycle

"Results are expressed as  arsenic, not  as  the compound
              Cladoceran,
              Daphnla aagna
                                          Acute-Chronic Ratio
                                               Chemical
                                                                 Acute
                                                                 Value
Trlvalent Inorganic Arsenic

          Sodium          5,270
         arsenlte
                                     Chronic
                                      Value
                                      (U9/I)     Ratio
                                                                               912
5.8
                                                  B-16

-------
TabU 3.  Spacla* Bean acut« valuas and aoita-chronlc ratio*  tor arsmlc
Spaclcs Maan Spael*s Maa*
Ae»ita Vain* Acuta-Chronlc
nk» SpMlas (Mfl/D Ratio
FRESHWATER SPECIES
12
II
10
9
8
7
6
5
4
3
2
1
Trlvalont Inorganic
Bluoglll,
Lapo*ls Mcrochlrus
Flagtlsh.
Jordaiwlia tlorldae
Goldtlsti,
Car ass 1 us aura t us
Stonef ly,
Ptaronarcys call torn lea
Chaonal catfish,
Ictalurus punctatus
Fathead Minnow,
Plaaphal«s proa* las
Brook trout,
Salvallnus fontlnalls
Rainbow trout,
Sal«o ifalrdnwl
Cladocaran,
Daphnla aagna
Cladocaran,
Daphnla pulax
Scud,
Ga««arus psaudollanawis
Cladoceran,
SlMOcaphalus s«rrulatus
Arsenic
41,760
28,130
26,042
22,040
18,096
15,660
14,964
13,340
5,278 5.8
1,348
879
812
                                   B-17

-------
    3.  
Baolt*
Saaclaft
Aorta Valva
(M/l>
AcMta-Caroalc
Ratio
SAtTMATER SPECIES
Triwalaat laorgjalc Arsaalc
5
4
2
1
Atlantic •llvarslda.
Fowrsalaa stick 1 aback,
Apaltas ^uadracus
Aaarlcaa oyctar,
Cra»«o«traa vlrglalca
Bay scallop.
Copapod.
Acartla clausl

16.033
14,953
7,500
3.490
500
-
Ranked fro* l«a»t sansltlva to «o»t tancltlv* ttascd on spacias aaan
acuta valua.

Frashnatar Fl«al Acuta Valua for trlvalant  Inorganic  arsanlc - 440 ug/l
                                    3-18

-------
                                 Tobl* 4*  Plant valves for arsenic
Specie*
Chealcal
Effect
(pa/!)* Reference
FRESHMATER SPECIES
Trivalent inorganic Arsenic
Aiga.
Cladophora sp.
Alga,
Spirogyra sp.
Aiga,
Zygnema sp.
Submerged plant,
PotaaogstoB sp.
SodiiM
arsenlte
Sodium
orsenite
Sod Jus
arsenlte
SOdllM
iOO> kiii in
2 «ks
1001 kill In
2 wks
!OC$ ki i ! ir.
2 wks
95| kill In
! so
2,520 Coweii, 1965
2.320 Cowell, 1965
2,320 Cows!!, 5965
2.320 Cowel 1, 1965
* Results  are expressed as arsenic,  not as the coapound.
                                                   B-19

-------
Spaclas
Tissue
Tabla 5.  Raslduas  tor arsanlc


    Chaalcal
Bloconcantrat Ion
     Factor
Duration
 (days)      Rafaranca
FRESHWATER SPECIES
Trlvalant Inorganic Arsanlc
Ctadocaran,
Daphnla tugna
Snail.
HallsoM cMpanulata
Snail,
Stagnlcola aaarglnata
Stonatly,
Pteronarcys dor sat a
Rainbow trout,
Sal«o galrdnarl
Bluaglll,
Lapoals Mcrochlrus
Cladocaran,
Daphnla «agna
Scud,
Gaawarus psaudollamaaus
Sna 1 1 ,
HallsoM cjMNMWulata
Snail,
Stagnlcola eaarglnata
S tonal ly,
Ptaroaarcys dor sat a
Rainbow trout,
Salao galrdnerl
Whola body
Whola body
Whola body
Whola body
Whola body
Whola body
Whola body
Whola body
Whola body
Whola body
Whola body
Whola body
Arsenic
trloxlda
Arsanlc
trloxlda
Arsanlc
trloxlda
Arsanlc
trloxlda
Arsanlc
trloxlda
Arsanlc
trloxlda
Pantavalant Inorganic
Arsanlc
pantox 1 da
Arsanlc
pantoxlda
Arsenic
pantoxlda
Arsanlc
pent oxide
Arsenic
pantoxlda
Arsanlc
pantoxlda
10
17
3
9
0
4
Arsanlc
4
0
6
3
7
0
21
28
28
28
28
28
21
28
28
28
28
28
Speh