Document yk2vg05XGnvGqqbdjyr1OrBeD
A S B E S T O S : T H E P H O E N IX O F O C C U P A T IO N A L M E D IC IN E
Q <
<ri
43
Z
C
< D.
Ws
o
CQ u
>* 43 >t_
z o X X
Cd Ui
K
X3
p C .t: 3 X3 CO
-r: o 5
5 p >s U
CO
3
C3 w cd . _
c/3 00
3 to
<5
43
"cd 43 ' E g-o
4) T3 >
i> U -- N- --0r.-0
u S
8 *2 OO.C
o * 00 CA ra -- 3=^
#X CO
s
o to
^*5
O to
cu w
0 Ji jc H1 *^
> Ed
g
-5
o xs
. E-- CJ Ui
3
cd o t-- T3
00 4>
CL 3
to a>
43 xs
o 43 c
t/1
t>*
5 *
-* O
xs ' ^
<- T3 43
cd c
00 CO -iOg*i -Co-/3 ocl
oo
;
T3 CO xs
o 4>
XS CO
--* ,,
CL <j o> u,
CL 2 io-OO
.CrO. 2C y .ti w
T3 ^ to '5-1>
43 X3
C JZ cd u
^ 43 3 SZ
O X2
T3 x: co ft> r; b045 _13 >
>> O tt-f 43 o 45
J2D cLo. -3 43
.E ^E
.h: w *- m 43
>>
--
C o
43 c
uc
-
3
co oG*O.
o
to
l^]
-a to
5 jo .tS o
2 'So'H
|iE ic
a) o *5
4J box: to
`
L2
co
> o o
to 32 o
c o x:
C
o
QJ XT
cH
*--* -4-* W
^ O^
-- o .2"ci.
E
Kc
to
V)
> x>
C4> o
cd 3 ^
2 co,
x -a
CL.Js O 5x:
i>ns .o2. cO --C --ed t>o C3 ed
E xsT3 'O* V-
OQo
*3
* to
XS
1 .ti ced
, X3
X >^C ~ c
GO -T3
*; ed
c X3
c ed
^
GO C
"0--5 ui- --ct) --C3
'G 3
"O
E ^ao c c xi .2P w ^ M9 .2 t/j "O ed
X d-
n-O w o {* wi 4J*
^ ti
G C r* u cat;
co to
JZ JZ
i J4S) ij
o
.
C --
^E
to 3 .tS 43
cLd<5^5 o lS
i
O
c
-- 43
C to
CO
4) 43 to
Z
2 X . xc3: ed
(L> o
-- cwO 0)
4C3L'XCSO c
GO C
a,
t> ed
C CO 3 xws *4>
E-
.5 c
-o o
U3
4)
>, J
00 o c**
3 C
2 E
(J t-- <L> o
4>
c/3
Uo
n> --
o
U JJ 5
c
to
ed D-
co
ex *5 x to G
S o xc ^
u 01 C/3
43 X3
O to
euBd),
Z
4> --
c 3
<- L-1
? 4>
>sS C^
4-^ n. to 2,2 ed co to ti C 4>
a^! .is ed
-O ^ --
o2
ed o
'TU G
S 2 ed
to ^
5 > < X> X3
-C X) "i ed O
- 4)
-So s52
tUo. Uo ed X3
--u co o
to to >, ed
oc
0 u-*
c -- u v- O
to tsi
to X)
ed X2
--
t
1
8
2co
HS ^- -a
S|
x: --
c o -a
ed >*<lN .C5
43 T5 X2 to
O c
o3o
c -* G
.H
2
4-*
O 5 ed
^ a. E
E to . u. ; m cs o to
? -- *- SP
CO -5
o ed
-x
to
<2 raJ>
*T3 C
^ O
M3 *-> vi ed CL
1:3 O bo
to o -S
u ed L. CL
O 'r* h eSJx
to ed <u ttj
v> 3 <-
^ S. o
to J-*' i- co
to CL
SZ 5
to w
: c3-o
2 -- *tor +*~*t
S.s r*z-* G4
c
to
jz
"
t-
ed
Oto l>-*
to o
to tC;L
12 '3
^
>
X3 a
<D -
o
oo to
V> oo
-- v >
x: o XG
.S
o 1H
^ 1c
-a to
cf EC 2 - C
3* GO --
o<
oto --5
^ sto
.o ^ed
ed u. to c
rto~ T3 CL 4) CTM %t
8o
cl c to
^o
22 w
4> z: x:
X 43 __ C
T3
to 4>
T3 sz
G ed
3h aJ jo w o
T3
.
_ -
m ed
to tj
S 2?
2o
c2 u. i--4)
to r->
to c
jo to xn u
g.s
ed w
tzO; ^3
oj "q in O
5/5 H ed 3 a.
s > -5
to 22^ _ G
O
4U3r
u
^ 43
58
S 3to 4) G X3
to -O
> >
to
X CO
o -- u. *a
to *tJ
to CO
O
C>L* ced
*3 2
\
;
FMSI 05825
u
ASBESTOS: THE PHOENIX OF O CCUPATIO NAL MEDICINE
49 13 G lx C -
.
O Cu -* o
w ra
.2 'O u
LO.og-s|oa*! -3T.*
wDO --u [l---- 3^d 3r- ^IC 'jb3
c_^u22
x2 U3
^49
-5 ^ 49 2 -- JC 3
o 2 .SP!5
to 09
u 4>
c
2 -e a-5 e
=> -3 2 g
*149
E "f
___
~ T3
1/3 g
C2O
> -
49 o o~
*E
00 C U 49 ao.oj
T3 C
5> o rt = 2 >*.o PE^ |
3
LO 49
c
S' g
o
- -E
49
_Q JD '
c
o
.2>L 2^0` O'
5> O 3 *-*
u. CT 4>
d
^ .0"9 c3 u-3
_
JZ3. _Q
oo a C CO
&D' C 3_
3 G o
W>. -^T "O -g
c to 3
lo G
O^
*O-- wo-
o..n*W--e oO r* a -2 b
c2
3J3- jD
-3 CO
crO
oG
c 2P
o-ur 2 CO rrt
*r --
3
>w
Cl ?u>i
C3 fc60 c
>% cGa' a. "5O3 -2* .b o u a> E 53 oo g
>49 ? S co CL to ClG
2o
Lx -G J--
S rt 0.0-0 CL W 3 xO .m
to _r 49 --
oo cO a> CL >
~3
cO q> ,5 *g
E~
Lx T? 4> O *-
(/I
09 -- cO
s*
09
a>
c2: --g 4>
c 2 ^ *5
CD
Vw u < J3 O
O JZ
~
CO a
2 t ^ PSi 2o 0ra0
E
09 ^s-s
9
%m Q "T3
o
^2 c
j: ~ *
I- ^ 1 c
S-o
_ 49 ^ JZ
t> G d
3 4> *- O .? s -- -
u. bO O 9-S.S
J<-D ,Lc> 3b "O Sh S Ji
CO
u
E
'ra c cO
wa a> JO
72 ^
"o % CO v
.2
'o
s
--o
*5!
MS "O woc
--G*- --V) CO
W NJ
o-S
W Lx
to
3
3 3
g * a*
tT2 *5 c
49 3 J3
X
Ss <*- xO
gs 3 o
^ -o
o
^ u
JE
t/5
3 OLx
O Co
X) c:
u --
a> >
Cu a>
Sh .2 K '
lx Ul I
UO u C 5 <0^2 -S
8-
<L> t/i
^
c o
-o O
>, J3
l/J -*
W-Lx CCO V
D, C 3
*a J=> o 3
OC
O
tn
_OcO pto.
to
w C^O
a. g o
>>u S u
O 2 X l-
js^S
.
c9
=Pb3 j-cj;l;.
3 O
O. c c_
S)1*" C ^
>> a -2 .2 75
c ~ ow 3 v. 3 a> C
-- OXI u
0) o c
5:
C-'D
S o5.-"55 --^ S"*
5 g - ao T3
gCXuco. xw:
1> *x
09 4-*
to .. O
1
.5 x) co
O cw DO ^
J
cl-S x
a
o. Og
3 cr bOG c2
a> j=
cG wr DO ^ c - cO
> CL U 49 ou &rto aTM Cl o
^ E3 U u J2 S c *5 o JG
cw jti ww
Xu X rTMa u Gw
t= .
x
U . w cOo Wto Q> O rt ^O *a"
3 L3
G 8
Up I
8H
.u
to Ca> ^W o CL io J a> a>
49 49
* o 3
o -
w - ,C
c a 3 js o
-5 3 c:
.49 uZ
S o x" C
1g
^
cO _
49 -o O ^ U4
o o x -g .2 X C 09 49 G to ^
" X - -- 2 X Jr 3 x . 2 * 3
oh J aJ 2 E H x x x':
L H ts
c
CO
JZ a 4>
E CL C
*2 %Sb o o
13 49 tO cx*o
S'.S"
G
3
E
rHi**
GS
. 60
o to cO
=sc 8
v? to
O 49 0> > si
lx -O IG
C 'uj OL
-a w
S ^?-s cO co
G 4>
o
?_
a JS 3lx 4W9
S
2
<*xl to 13
Wc
O* 3 ^3 co
1
439
to
t">
3
Cl
tj
to
to
to
`G
u,
4b9o
sc 09 CS .
bo ^ JC G 52
s:s
.s
co H .S^ Ji
THE M E D IC A L BULLETIN
CO-2 .E-o^" _Q9
o . - >)_2 n:
49
cj ~ t -- -r* t-T --
to -C
gS
2o
?0 -
CO ^
i
tq4o9j
Si o 3 E ^ cS i JG
3
s
-2 3
-a
.2
<75
JZ w
49 >
J ^-s. T3 ^
X
- 3
3 x
*.
49 G
._c2___^L_-l?_-_r_5^_
-G
otfl
E . o 0 op s
1 2 oi1
g-S
4j 3
e
ao.. >- 5P ca .3 <-
'Lx.
G* .
_>^^49
O
OO ))) O c jTM -a ._ ^
o. G 49
u
to
oG.
2 _ 'E> c/i --
ax o i- J c i> U l- -* -o cs SxU5"
JD
.S *
.<"
<4x
o
aS
>-- . cO
to .G
1to cd T3 C3 Oc
--co _^G -_S^ `* 3 3x W
>49
tLox
5
49
3 -D ti
E u
-- CO-a --
?5 00 > P 'C 3 co ^G*-"^ ~<jj x
~ o u.
3Cao>
c 2^
3 l-r *-*
O Lx
r-- H
USSfh C3
O
o
O
a.
.-2
oo E
2 "a "o O S- CL
So ^ *riO"
CB" '5 SL.i -OP
c-- c '--0-*' --CO 44<990
. ^ -X 03 U Lx
u .2 o t-, _ f'E ^ g & x o -- ^ -S L) x
1-
g a> .S g X J E S Eco gf
,,E S
cu E
X x
E U. O 3
CL 3
o n
iZ CD o CO
49 to T3 4499
:>o
-a c
^3 3
fN
tO 49
o-
60'
C.
"3
49
09 JZ
T3
3 Lx
G
3 r-- C
CS Cl '
4tO> Or Lx t~-2 --S
3 49 Lx
to
09 -Q
3 cfl> ^ M <2 .^E
CO i>E- g E-S
< 2 g u IS K "o CO -xx^ ^ g o C *C " C
-- CD----c o o . P
2
c
to 09
3
.to
cO
O 3lx 3
JC
to33'\*-rr ^ ^ C49-G
t^/5 % .i3>5
u*D
o *--x
JEx
G 5-
2 CU'G^?Q-
roi-t<PLux P--a-< C3m
O 3 W -
2-G Gw g
3 O *>3s
ci-=G>2-- L^-2SGE jj^JUcOo^ o --G"o-
n u"*c;a^<? u.x3 .3ii^- " -,,o "Ou
v5QU.G.JrTj3>49
Lx .X C
JZ 3 P -3 2 g X - ^ i? c ^ > -
CL
x _ rt i=
Oto TiG3 -^-GLu4x9,_
n
49
X^>
3 j33 Ol- 6^0 ,^1-0^49 c J^,g
3 J~, 2 c
E
0-T)3CcL --o x1= X
E3O
>
3 >) P U 2
u2 '
o. 2
<g
> a 'H
c
*s w
49 g a
Lx -' G 60 09 J3
6g tflx
O33 C G J3
CL O W ,2 G
--
to w 60 w ^ c: g:
>> 49 49 co
uo
3^ GO
_E
-wg Jt3C
^
b3 -t-oi
o
3 49 iL
49 ^
C
o
--
3
,49
he 'o
--X >."
s8
B. 3 B
fi.2C^L*T6"0 W
49
co
C o
W c
tO rt 49 J
=O *--C<
3
o 3o. _C
.3
N . to
O
-- 60 tWo 3CTJ
"5 to
-G <o 3 .2
G --3
G ON
O C
SW TwS*
w CX
Lx
49
Lt-x
^
--
oo
^3 3 <
wo xIS> <32
L-,
2 jUr .y j"y4^ 3Lx
^3
13 g
CL O gG-t0o
-2G ^a> ~ J3 C to
3
3Lx 0O9 .52
*>
09 JG -
AO
CS
60 *
49
Lx 49 JZ
iS 4u9 tflo9 CO- .5
u u o<
. 49
3 60 to to
X
3' X)
to --.-XXX
G 3
09
3
G
O
49 JG to 60
<2 2 _S xG O w -a
0 co
d^
Lwx
.2 `S.-K w CD ,
c
Sg CL G
49 _ 3
co 2 3
3
C9
' O
-O
G
3
3
w
49 49
Uw
E
O49 oy > JD
G
3
`O JC o
oc
U^
4> 49
cC Ew O "O oc 49 3
JZ a
3d
o -
E u
" ca
JwC
--ns
xg
o ir2i
cl -G
a. t-ox, n2 49
H x -o
to CX to
o c t>
c
..
i= c E o
FMSI 05826
o
d CT3 RJ C
C t--o co
c wcd V)
3X
E 2
1 &
C O
o >
X u
Q. O m
o
go *
3 "3 X 'trt -- ^ >ii
-a
c cd
c 3
3
>
oc xft _c
a
ou
CO j-- CO
-
4) >
C 3
.2 x
t"a>ia?
3
cd
<0 Hog
3 cd
"5 to u. o
oE
. cd
(/] 3
3
+* CD X
Q- . cd
u. cd X c_>
3 E
U c
3
CD X
E C3 cd
>* <n 3
X DO
c
CL cd
cd CD co
co x
X
CO , 3 C
E ^tf)
s
g:
XO -
cc a.g< 3
3' t~ cd
CO *--* * ft_> oc**O-* 2 8- CO
CD
^ c8
E
*3 '>
XX .
HaC>O
. cd
3
=s
O
>% 3
<_ CO
-
3 -5
o
3
X
4> >
cd Ou CD
>U>> U
cd -G CD -*
CD cd D. cd CD
go
c
c u-o i; S-X 2? s
gS^
o 'a. o T3 4) cd CD X 3
0-7" u-
#o| j2
c o -> 3 e 3
ao
go.5:
G "3 3-i
4o-- CQ
cd Ih
`ELco
S
a uco cd
*5
3w 3
E CD
cn o
c o
UJ *3 to "3
U.
X
00 X
CD CD
3 -3
CD
&
to
CLD-
GO
Ou.
Cu
CD CD
3 *3
CD
DC
*o 3
cd GO
T3 O JD
X)
c
co
ID . oo 3e O3 cx
33
Ou E o
U- c
o CO
to to
to O
o
XZ
r-1 0> --> >% cd co
c _J <s rn rr
t/d o r-
*3 C cd
to GO
E 12 u
3
EE
U . bO
c3
3
3u* X
o.
3
*3
ca>
T3 O CCDO *_,
3 X
cco o xm
o
TJ a
bO c 3
to
. .23 "S
to "3
CD 3
C-'
cr
33
CD
a. o
to O CD
o *3
gCL ^>
3 >
Eo O* 3 c>
O X 3
CL 3
o~x 3^
wto 3
G .5
3 <D 3 t-- co CD X 3 0> C o S.S 3l--
=3 .2 EDC
CL 3
to
Zf
bO <
3 XX
X co
`to .G5 2 '= I -S -- <d
A
to bO
c
O
C
CD 3 C_Q
S
^ G
3 ^
3
Ou O
co co a>
3 rrt u
g> Mo
3 CD CL CL 3
X
3 -G
o .5 -a 2 o
Cd
S
3
s
>*> C1)
00 ^ OO CD 3 x
C T3 3_ O QQ - cd
as
CD >> <D C
o
CQ O C/3 H 3 o X
snd^^
OO
o-3 i|
5!= o
I 5
uu S
o-^ i
2.
fc* jG J2 5
H 15 o
53c"5Or.i=?<a31*-aoC<uKS5 S 25 Qto s*M
5.-u- SXT oU 2S c*; wo
sg S *" 3 -i5 S>u \co3 jg*e
S5= j: j=
2 - Sk ^*0 0.0 '1 " 2 E .2 .2 o
2 o "c
*-SSl S
00 C E3
V-.
3
CD >*
'. >
CD
e
3 O) o X
ooa 3 E
o<
Go sx 3 CD
Eo
o UJ
<0
>, .5 a>
CL 3 c
o
sX
o
cGo 3H --tO< 3X0
8 iE i
i> co co
a O ttZ? ^3
4D X
U-
s X 0 V3
C ttD o.2^
ECD
w O
CD ' t_ 3
r> Tt3>
`t:
:n
60-a CD
c X
3 ^
Cd
CD
> *c c -- 5 C
co
U TJ Cfl
CL 4C-D* 43- 3cE
CD * to t
cX
xt
nl *-* - 3 r? p;=--cT3tX-
3 O CD Lu O
X *-*
u. XTP -^c
^
3O
'*- co
3 O
See
8.S -a ^ > | .2 & Eu ca d
3
3
5rn 3 x <5
5 ?00 rr
ocg Sop %j-
-a w x S c d o ^ 3
E3 .
O
w X CD X
t' o!
"3
*-* 3
s CD O 3
_r E O
CD O
^
X
1 CD OD ' X CL
CD
to t-u
E3 c JS-2CD
CL
CD co
3O CL --* * ss
ECL X
3 CD CD to
*-S ^o < CD
CLD C>D
X
CD --
>>
EX 3
3
w CD 3 E
CD Id
*3CO
3
Cx
<3 E 3 SCD -- CL *3
WO
CD bO
X
$* g-s 3-|JSX .3.2 E j*2
3
oCL'
tip C
c
to o
y
1! >1 GO *"* coG o
M cd o tO G-
x 3 co x E-5 s3--> CD
>7
*3
O X
co O
co
u
XI co
0
3.
4> ^ X
to -Q>-*
to o
to X
2
co 3 CD o Xc to X 3 GO
r- 3
C -4> O^
* J> CD X--*
O T3 w CD co ^ *C 3 CD CD
CD -5^ 3_ CD
CD CL O
H CL CL 3
3 *3
co O
; ;
o_
to a u c -- o
3
E-B
.
s
o
-O
od
c ao -a '5--G
3 2 3 -- to -Q
GC
u "O CD
XX
CO
3 CD
o
CL CL
"P^w
'
3
L. H 3
CD X cd
co
J5
3 CO
3
*o
w
C to 3V
X 05 x .
> r>. co
3^._cBO'
lo- pPr --a>
r3
70^ c-3o U33 t-T. Jti o -a CD > o c X o
CD CD
r-- 3 >,
L-- X CiO
CD
go
w -3
c
t: 'S> o.
3 .
O. t c
CO
5o
i>
'LSLw/
g
--
CD X
'
3 CD
w t_
> O
bo c 3O O
CD .. CD
OO CL l^ CL C
co CD C
o >> CD
3 CD H> X
"0 3
oG
x
X co
'53
--
'S
3 O tn ^
o DO ^
o G T3
3 CD *3
CD
r-
.2
OD
o *-*
to CD
f%
C
co X c<-- v_ to O
U w
O. C 3
cd
>,5
--c
- g* =
> O x
a X>
Q
in o
CD jz 3 to
OX
3 60s 0 3?
"**' c 3 3 na
w XI x c
T3
G CD
3
OO c
?
bo X
c
X CL
O C
3 CD o
o X L-
co CD g
<
CD t7 u
>
C
O
Cil .tx
S
o
x
3
cr
--
CD C
c
^
w
u r*
o
to
CD ^ 3
wX
.c 3
o3 G CE xw J-?
o cd
CD ^ GO.3 HO
^ cO .
X co 3
C8 . W ^ O o
gc g -a c -- CD CD
SE
So" >.`u
CD
CD -->
Oc
'X
cx
C 3 CD
CD CD
CD X S
3 S tn
3 X
C
3
CO
?
>*o
5 ts
,,CD
3
o - gO
OX bO^ Z,
o
CD
s
X co
GO C
00 CD
3 ' cl <
c
co CD X3 ____2_
CD CD C4>
*3 C3
-s
^
"3
gco
"
.E
*n
CD
L_ CL
C *3
*^3 r,, CD od G
3 ^ GO^c
O
-3--rCD.
d^TDv -3
tO 33
3 *5 3
^3 L
J3 3 bg X 3
t= `t:
x 3 < Q..S ,, w>
co C to 3 3 3
3 -u X5
o
cC w
5
3 ^
eo 3
3 CD 1 r X u_ T3 _
co
3 ` u.
o
--' WJ
SX
^^x
CD ^ To
r3--I
w
31 H
3 o !2
Q-. G G*
3 3w
3w
C
GO . co m X
to
3 co E e: x -- ELu h rrtti
CD
3 CD 3 t: x
>> 5 > O
u. 3 O
^
3r-- **"tod r*"r*i CO
E mo 2 u
X ^ 3 o c -- -5
ylTxJ ^r
. Eh 3
c--3 -rj O S -0
y
c-o JEh 8
Id
^
cd oo.H *-
c 5t: o 3 CD
CD X -3
o UE
.So E
G 3
to CD
j
l-
3" U
C 3X
H. X qj CL -n GO eg ^ 1/5
3 CD .52
co 3
g^l'"
CD
to 3
>> 3
3
3
*3 O
oo C 3
0 3--; 3" . E
OO
*-* CD
C- - S
CD u
r- <J
x; CX.
3-2 J- ^
to
-- O
to
C CD
|-s-a
G
3
u rt
n: --
60
S
to
c 3
=5-
o3
CD
CD ^ -CD -T? 3 o *~ 2x g
e > ta
b g- ?2 C
X H
3 1/3 GO
< CO
~ tX CD ^
>3X3
CD H-E s E
FMSI 05827
2
o
cn -O' 4> cn G
C3 &> >
G Cl, S 3n
* o 41
eg wc
4~1
3 cn Cfl 4> "e
41 s*s
O _ DO WC O H2
Go
l/l ^ 3 bo
'X0 3 s
C a * o t/i c
2 oo
tV ji
>
-5
_
o u
H -J
.
E^
og c
*-- CQ
oS%.Sh-
41 g >
C/l *T3 O r- C/l *. 41 -- C 3*. O
cH *5 g tz o ai ,3
*T3 41
Gor5ri
Cd .2?
0 G CO O
3 So c .2 o
CU C t~\ u Au\ Q. W
<U C
81
^o
5 4= 5 s '5o >
G
i2 . M
*> V. cn 41 41 O 1
.y >; SZ'Z
-a O
Vi --
4) JO
4' O "
C/-1 60 >41 -cI |e -w ^-
-3 2 5
cl 3 c
-
ai v: 3 o >>5 E u2 c .*
2u
Ji T3 o 2' --= u
Vi
.2 S* Oft 3 .2
is
iH S
ex g a 3o G 41 -* > --ar T3 oo
cO
oO
E
_"
B
u_
X) cd
E 41
Ui 41 >
E 4!_1
4to1
uo
<U .y C3 C
3 vi 3
C
g> o~ ra 2 > 13
41 41 EG
>%
u
S
o'-S g N
6 s:l^
G 3
3
3 cn
12
-a
0 G
4) rM
cn cn
<-. cn O
E
CQ u 41
, *g --
-C- *^
0.0 -x TU
*^5
3 G
3 'll G
0ft
>- C/3
CLG
VO D ON
5 E.5?i
41 G
41 G
cx-*3 3
03 TJ 03
C CO
u
u u =3 V) O
a. x UJ
w cn
a3
(0
Q rM
U1 w r>
1 ^o * ca
z
IU
cc 2 E
C/3
VO
E a>
l/D >c
00 VO c r-
13 ^ C,,
o o c *3)
Om cu
o^ o
(JL
o Ui
>* U O
ov .2 ^
. u* U Um
vt r- coo c *--
>> ~o
3
cd OftfN u< -2 i*-*
'a. 0
C/3 a>
2 .5
CO u
.
4> OO 2 i
-G Jg -- CJ\
>.>
U uS
O-
m
^ai rc/l . .
O3 ^6,
cad> Co'. a. -
aW
vj
O
c/i \0 j> ov x -- Vi
<ri
* Ov >"T
^ 00 . 00 ^s
U
VO ro
x:
<-* 00
.e2 ^r<3
10 <N o rJ
C/l - T*--
-O vi
*7v3;
a> z: cf oo
w
c o v> 3 vO S2:
--* 725 0000 "3
0- ^rs c^l CM
s: s:
cT o
UJ
3 o --^ c/l o --^ O X)
-X3
c: s:
c o *W3 u_ UJ
3 4> -- CU
UJ 'ZZ 3 I u. o 2
-s'l
4/5 dH as
UJ
11ac
C/3 ^ >\
g*5
U-5 -
IS
o. W -5 BQ
||
30 *> m Co
DO *^3 .TE3 v^i
"O <S
--<: C3 Vj
u u> . 03
u b<i
c^ cd *
O-Ov rt --
i
43 X)
23UU Ov,. QI2_i. JU3
r
: s: r* U ^ <5
I? o|
;^ UJ
2U
ui >
UJ
.;
U -2 Jj s ^ UJ
C/3 . -o a ^y Ox
H^
^3 o s: 3- a cq oo
DO u* 4) uc JD E
10 r? orj ^Uh O .5
*
V)
C u -a 3o
O CQ
<Z
h .E 21-
Jn-Jj 0cn
C o-- ~c C3 3 'ZZ 4> CJ
u CU
UJ. uT t3 UJ 4/
^ -S
o
.o2
4-J
w4>-
o 00
C 4>
U3h1 Q^ *7 00
^c
.
C ^ UT o z>j |uo
{X o o > o Q
"O c 03
4>
Cj xc Q ^4
cd E o J3 4) -C
0- I ^
a 3 %-m
4>
to <N C/J U, a\
; u ^r ^ -X: r-
c O^
yi 00 ^
4>
H_ c.n .
qr o> **^v DO.
55
Dd1
Cd x
r?* -ii
00 Q* On
03 S
cn o r~ 2 a> -g ci
<^ V-J r^
2
1
4a3
G cd U
DO G 3
X
uSO^ a>
a .y ~ b j: ^ CQ &"=!
. mV_ C/3rTJ bDOft--3
9o S -o W
d o J? Crf -
T3 -o rt r u. CQ cd O
^
3 2^
s
>- 5 .zX -X 3 -3 NE
3g wE
o
4> X
< m r--c 0\ cd -- u*
o vn
o 00 UJ ^
o--
0C/3 T3
4)
1u
*0
o -G
O
cn E u Ui o .O c
43 - J3 C5O 2 -o o g 63 3 cn 3
4> -O 2 cd 3n 4> 4>
4> JD
G O a c SS
q
cd
o
o
cd O
!5 *5 wQ
V)
E
g
I.5
ui ui 4)
.. i> 3 2cc
s
= a2
I
00 cd T3 * u
cn
EE ic^
^-o Oi_
--2o --p
3 7JU3 x> g-a IP
o
g--n g-o
J* O cn c/3 u, ^
o **s
o Cd O
cn O uj co lo
C 4) 41
4> X>
G
.2
o .22 G in *Z3 O cn v-
5c
O
.22 7?
o1/5 5G u3 -2 3
rs j=
E. 41 cn
-G vi
-O ^0
C cd
"4) JO
"5 o.H
E W> c Pc o -E 00 s 2E u,
^G 4>
S'-H
1
O cn
3 rG- ^m .2 -
O2 *- n
41
GG O 4) *G - -- ^ Oft
J: 43" cn *r- ca bo 2 .2 O.Q
ra &0 ac 13
4>
z
a 4> c 52 Tw3 5O 4> Cu Ui cn o 4>
Jr; 4> V 24> pG cun >v O O
o 60-0 5r </i
o -- - cd
E
cn c O
O *- vO
->Si
4> c z2
Gm
E 3
co
Cv 3
G --
3 cn CM
1 sO
A > to
I^ [ cd 00 ic >-J
O ^5-
~ S
.b . n -> 4>
cn cd n *3 ca *rr o SQ
s
V G
JS s-g
cn
3 >> c y
G
ca co > E J2
2 C6
*a
>1 0
jc E w j-\
'u
u-
2 m<2
pto i a2-c
Ui cd
u c 2 S* s
E c o 4> a> o 4> ~ -25 c <J
2E ^ _ca oo ^ 11 uoc
Vi OO g jg ' u -
*c-* u2p^" H <*= G9
* .S -- CM
o c
o u2 C
-o
DO c
1) U S-S
3 2o
G E=
cn ^3
J-S ft> oo Jc cn vn
*
cd c3d ia)
3
4> cn
S E
60 <= cd
'5 5 5 .
tn >v
|-& 0 s
4> *_.
sESS
-O c %CO -43>
e cr
E 4> c^
4)
4> G
S .
G
o O
Ogv,
c cd
JS *-*
cn 4>
o
o
to
4) '
EG s
> cd cn
g
O
4) cn
DO 5 > cd*
E in Ja^ c 04
o
<a 2
> c
-o o cn
Gc
ov v* J41 .22 0 c 3 Cl .
41 OO-X GO
3E ~J cn
.2 g
C c 0 E ui
cn 2 " 'S u
*-* TJ c E
DO 3 e 5 -a *- 4>
G s>
0
m3n uZ2> rF- oo
Or E
6 cn
CO ^ U Cfl JD 0
^ o
4) G
O >U>i O cn
G
cd 4> > cd G cn cd E 2 U G
4>
E
c ' 4>
O u 41 CL >. tZ o> jb
cn GO DO O.
W5
cd o
4) a G 4) rs *> 41 >> cd
G cd
o >
o CL cn >> Dft . O DO ~X G
C cd
2E
3o
CL G CO `U
OO
O 4> `Z3 Cl C C3 GX crdj *3-. ccdn do o2 n`W 3C G -= cd *0 ^ ^
J! -ti
cd _ o2
to jn v'l
O ^ Ov U 4> -- GG 13 ** .
*3 C "O icd; cd *4-<)
l-.g.
E- 0rM5 ou.
3' cn a-- rt
G^ O--
-rt ?=?
u <n 41 G
cd
1= 2u -*153
41 G
C -T3 4) 0O w
--. G Jd
> cd ^G
G
cd O o
cd
CO Ol 00 10 o
CO IL
cd -o
5E
O On
C --
c 3
U .dj
60 ti rr C
Co
cd 43
c
cd no 44 O JL o *n
ti 53
s
'o5 <n C/3
V.
1) 3
a> X) S3
O. ON
< <N
c?ON
O co a^>
X On
W--
sf S3
a. o
z<D
-J
3>
CD
-I < o
o Ul 2 ia
3^ -J o'
oo o *r* 43 +- OO
o Ti <N
c os X 3 a, CuO.
4J < 4u4.
O
in <N
O --
t1o) Xt>n
NO
o
<
ofN
tn O
t1n>
43 tn <
U.
T3 C cd u 43
x
X1 "<3
43
tn O*-*
Id
O CO
E.l E oin a>
a
sui
c
CO 4c3o
<
n
fc: U4 O
i r***
* 2
On ` NO
^ to mO
o -* 44 D O X> O c/>
CJ <
s.
60
4o3
C
C
3 nJ
>>
O
cd
S= oX "5>
to
cd
lx o1>
CO Z
O3 OO `So rn
o
rN m
ON
o C*/>5 <nNj-
<N NO
FMSI 05829
iI
&5...22.S,, i . ^ s 53z
fr ill
3
E ss-1i ES>2 "I 3 2 J OQ
60 eO
-S -o
Cd > J3 42 cj
.52 w -a aj ----
w 3
25 cr
oj
oj
>
o2 o^
cst JD o dj on t)
(fl TM 4) O
CO CO
Uv 4J
d> --CJ
C c oon
2
aj CO a.
CO 3 _ dJ
Si-j: o
u on *3
CO
c -o 00
c O;
-Wwcaj
C *
dj o on c0
to r--- TO 05
O Q.
05
_
^ H co > 2 . Jj
XJ r-
<" -C J2
60 2 13 jj
M .2 "
U -O ^
c 3
--'* CO 3 O D3 cr c yo u
CO duj CO
To, wto ^Q 0J
< 3 3. 5
C 0)
2
3
tn U
t) J-- W-l -Q uO c
S
*a w u a_> 8^ 5-g
%
t) co e> to,. u 3 3> 3
05 CO 7-
co 5^3
wO *c3o QcoJ O U-
XJ ^ > 1> q_> -H tn r-| z: O C c^ a. u
o 3 h
60-a
.c3 C
c CO
"2 co co
.-- 3 3
to c *--
m u
^ VCJ
4-J
- cj o --a
.2 to
a
dj oj 3
34
u 3
P
c" O
'15
CO TJ CX*0
o> 3
cn ,, ~
2-f
3 c *1T3*4: E-s
-c3
o 1)
.2 u
L0 d5 u
co C .22,
a C
O j=
-- 60
c^ '5 o
-a r .>
-o
-o a> -a
<
-Q c0
CJ
o
co r c ^
CO
JD 3
W CO >
3 d: c
1> .
. xj ^o (= X3 a> Z3
S -- <u
Oo O..E
Si? ^
2oa
xo S E
w w a> O. in r; O UJ *TT c- | > So CJ VI O --*
r- O
jd *r 3 >o
S JC 3^
05 60 > 60
O ^ co
c^c 'CSS1 CX -C ftj u
=1 3
o
T3 3
60 O c
u>- cU r --j CO T3 T3
oO
co
,a.-o C3 JJ `C
o dJ -X
co dj
.2 co
cn " o -5 o co .2 2 - n>
d5 to 3
o
CL to
.c
C 2 co
wJ
o. g 2 jc
" .c
o
c_2 E u w,, 02
.C"l u O 3 Cl q
3 C2O g>
s
t- V) CL O
JtDn 60 cO C c 2
*3 u > -> JD cO 13 JC * -- >*?
c t_ QJ t O T3
ftj t- 1> 8 CO JD
60 O "O .5 S
*3 CO
> o 3 JC
C o CJ i
o o .2 t42>
c g O .3
.2 S3
-- JD CO
dj
jd
d> b
gji
cSf
ea 2 cr o -D u .
t/5 uo O
T3 u
CO N . " 60 i. 5
y.E <
2 ^ .
E "a S gU c VcoJ
? S1 2 CO 0) CO --* .fp > CL w
c c: JL
O 3 o2
to O
cl S3
w^ oo~
^
O. X
rES
52
O w
CO 0J
< a>
CL JD
fc- 5J .
SHco -->ov JD CL w <1 CO
> v
JZ w O L. .3 -c
i" .
3
-X 3
o ^
-a
O JtSo
w RO '3
X3 O
to c *- O
2 c co .ti " a.
1> > CO
o JCrL,
S- cj t
E
3 >. o u
dJ O 5
o -o a. y 3
3 =s 2
3 to o <D JS
pw-- .i JcJp w U T3 =2
o o >,
X
O
_on
x UJ
CL cC
E *n uj i>
o-CJ wa ES
t/5 on
. .b u
c t-
CU u3 >--
CO a)
3 5) O 3
^ 60
C c 4)]S
-o *3
C3
3 C-O* co ^ x: vj C s "
u- O
0u
--o1
oI
O .
3
u c
O CO
5/5 O
E
"3
J
b IU
O"
c o rW\
1C/O1
0
tn
g-E 1)0
CJ
l
cn
z
o
H <
a
z Ul
S
s
o
a Ul
z o
cc
-
<cc
o
o DC
Qcc.
H
Z
o o z
Oo
o
X X
(0 Ul
o
CO
Ul
oa
V)
<
z
o X X Ul
?5,:
i c c i U'o 0) c w vj
E o a> o " . 3 -- 3
oo ^
u is ^
rv
_r gcS
-
I!
S
> 2 0 c c ^ "a S
s >O ^ .wr0 .3
c - - u Co
4If-*l r1,> oDor Cw C> u1> O d> 3
E to to 3 b o eo 3 *t-o .cb *oo
w oto .. w
t O 3
to 4>
O 4t?o to CO V 32
3' O
3 2 o r-
1/3 c u ^
c
w i: R-o-g 2
`g CL
pc -ua b 7-
CO
CX c0
O tn t-- UC o
wo
-S "a>
60 -
d^J >>31^ ^*-
60' cj .2 41 .2 'P 3 3 -- c0
5S-5'S5o
3 __
O JS CX dJ
--n_c_0
C *P3
4> fc L. u, 3MU0
a cj "O
1>
=> 6tt0i
c gO
j2
o -
s c
5 flj
'
.3 O C 3 CX*
2 -- *t3 4-* co ^ CD c c
X T3 C .2 O
dj C <u GO
C . 2 *E 3 .5
60
CJ o
ES -|
o 3 CO CO 3
o^
4CC~-O*
-M2 C
E U.
c o
l J n hfi -C t--i
c s * = SS
a: u_ duJ
E a>
E
JD
O 3
*o C c0
*-* * -- --, s C * CO dJ u_, CJ E 0 2 .2 dj qj V-- ti
8^ s
^
yco ^ o3
3C -O
60 QJ 3CN
VJ CO
QJ S
cu0 vs 0U CL 3
80
a>
dJ CJ
c
x QJ
dJ 3
CD
(rt , 'D w g-5 O 3
o !> h
_ca
ca
2 60 o.E
em
3 CJ CX
05 3-->
CL 3^ c
2E-^
d> 3
.E
E O JD U d>
o_
c W CO C3
c c ca
o tO V5 1
2c
^ 3 *3
1) o c co 2 5) |
a> c 34-* OVJ
8 3
.SJS E
X3 Jr S
ri '5 3 S O
S 4>
u lrn l. JO 00 ,, C3 M
T3 C 32-
O c0 cO S3
" O <n
3
tu' 3 JO -o
CC ---i i7S-^ 0 3 *3 ^3> > *o 53
Q3 cS 0
1g/5
3 -E X3 CO 4D- 2oj OL to S C CX c
1 ca 60 -- E. 2
u H *1 c
g2U. .W2C
2S.
3
d> U
as.c,
0b w O to 3
o
C C 5 >*
CO o
c o *a
OJ 0 -- fc- 4-
CO3
1 S.'! o Le -5 CL
dJ
a dj E ca-g
<r> co C
<S c X3 C C dj o 60 ES
h
uU> Wa> 60^
-E in CO
2--
&G e> O--
X3.2 -- _ C 4> CO --. *_ _ca --. -aca; 0O-
'= o 3 2s p
2 6> Ego
<- O .5 E -S ^
u E*2
C -u, vj ---r: rcof
'w _
5 tJ SE ^-- 28 .5b 2Cl t-Jo-J to
B to *""
ca 7to ]1_c_a_ c 5/5 o^ O C3
to 3 73
c ^ tc0 E o to
Xl tC_ cto0 P g. o 3
OC to o dj *__ 3o to 3 co w
CO N 2CO
^
CO CJ Q -- *u
TM -g 5
O CO -- og-o
3hi scl-s--
3 dJ 3 --T3 "3a> 3
d>J ^t3o 4O-*
w> vj Jt2o dj
u p
c3
o b
3--
u J=
5 8
~a c cd
2g
o J ^U uo S t; o cf
KC
O S
o
*0 O
O.
X
o II
o
3 to o
'ZZ 'ZZ D 2 S
W5 x
C --. r3
*- 1) dJ 3 Ui vj
w:h ? C0 CO
3 H
dJ
^c 35 Ec c 4-. a> c3
S
bcxt^o *5 S cr 3 c 0O
1g
l!
.
a. ^ j:
^>-s
ioi a 03 2 o . GP. J>5 ul "us > "o
u y? X3 sis
c E2
IE u o o2 S
uuo l> JD G. > ca x w -- i>
2>S o ca i> u-g cu x) ?
FMSI 05830
+--/ & _E E-* E .g2 ^`ES i a3, o r T> o uUl EDC ^3 ca o & S>i5Q
*5UJ o s t- a fcO
O
E u.DC
a
' 1^ uo s Ij g*z
-i
u S> o ma Ez
< x
t/>
o
3 2I-
to HI
*o "8co
to
-a _n .2 -o 1<
N CO JC >u>
G CO o
i T3 C
t-4 C .2
G
X
*13 a>
O cO m
mG
cO cO G
c0
G in C
l *T3
4> CO G 'ul G c Q->i
CO G in a> *-*
4> o m
*-
o
c3
>> S3
C "
3 "
1"o/n3
O Vfc-t G t- O- `
oO UCwO uV) to
^3
m C
in 4>
O 4>
m %>
X m
O X C0
u. m X
33
a> o
O CO G
X
-- G
<L>
a> CO t/3 --
3 S WJ V)
e3
3 O G uV3
*13 *ua">, 32
* 2EH| j2
W X "60*
g
u! --ag> E
Urt. X" Oo O. E
d> o
C3 c
JC eP
_ g-oio
m 5o D T3 cO O Gfi
^ ,/ U C W5 O "* _p- > G
- * - J* O
O c
.2 3 _4>
~o
i ob^
4> "O -S a.
U x .5
e .2
4) "O a>
cl-
s --
,
. ,
03
eO -* C
-O G 3
E
g: *
o
1
CO
(L>
o u
*0 o
c
33 S?
O "
5 3
in
<>
a> a>
Cl.5
i
G
Q
&> l. C
,,
t* c
U 3 T3
G
o CO >1 G O LO o -S -C -s .2 i> o o
bo 1< G *-* cn
cO
o .
CO
ts-s 3> *TS3 ci
o XIn 03
c
8,2 a'S
c Jt-
o
c .5P,
"CO
--V3
C
U 3 u
*Otr
60
u S
o "55 C >
4) `u5
QO T3
.wS
.6J-5-0r
co3 x:
`X c0 c
i> 4>
GE
c o T3 oC
co ^X
00 G
ox c --*
a S =5. E t> o r* w
4> C
t>: X,
c
--r a> co X *~` Ui o>
^3 .2 X Q-
,o X
n- w
2h
4> w -* Q
V-
Q. i/> 4> X3
^ G U uo
CO
*-* O .H
CO U
~3
4> (D T3 X *5 <*><*-
4) cn
T3 3 8.S
`5 5' li cO O O
SC02 oc/3 JCrL 13
xg
o
D O 1/3
>
CO X
U, 4) 3C
^o O ,ri CL io
4> c co
a> X
S *
O
s g T3 C
C w
J rt ^
=
Lo l*
.oX v0w> g' X1:
x V ,~
,_. CO
2 co cS g ~<oL>
in E
a> X
U 2 U -- a> tc m
9 y?
^ c/J
C0
x ^ x:
C /3
3 C/3 4> U.
03 cQ dl c/3 co x a>
8 60 2 a) c Ca/>3 S-5 X i/3 "O C0
3 O
u.
c0 C
i_
C3 ai c >X cO O > V,
u Ea> cfloj u. '2 T3 ^X -n x
^W
G C3
C/3 3
G_
E E o2
IX a> X
CO O
4> S E r! g C/3
602 c a>
cn o
.5 4>
*c<l
u 4>
8 -p
tn O
3
XUrin- XU-3
c Z0> in
c0 M cO
* c4o)
!r 2fX
X c/3 c0
c Jr
x c0
3 --
C0
x= E o
wX
_o_ >o
Q>
4) T3
CO G
2
S
=
g
wow "--w
C C0
w XE
Ew E
cE
G
G - 4> G *3 X
o O 1n/5\
CL
*G= X * C*-- m. wCrr
2CX U
*-"x
g S-S 1 s
fl 4) VI -- XC -C-O-
CtoXTft3>
<
T3 G
3 ra CL
> UU
ec
GO
Mg:"5
`E
t-+
"5
-
g*-g5-a urt
^3 -5 a " G W>S
R u *-
* O U ,-L
m ^3 u. X cO ,,v
|M
-E a.
Jr rv m
O c
2* *3
XO
mw
S 8*1 s g x ^
-s
in G T3
m- m<L> * O
-
8 '3 *" *X3 0>
C G u.
~i w
I C "0 X c
u 60T3
4> --O2 XG +-T3 C
r- 6) O
ir E 8
c cX
O
u cn
o tn
o c0 G c fll > O x 0_5Gu.lL> o " X 1> X
g-o jo 0T3
3 ft>
u.
So'G G X g * o X5 .52
< i-- JO O rO>
0> 3 w.C
U. --
G in CO bO"T3
FMSl 05831
4>* +-* X _Q O 4>
= a c y -- O'
o E; 01 s> *
oo
4-4
D to
to D
2 - Cl vi ui cd p Q,c~ 1> o
G u
S 3 S 03 3 **
w CL
.5 5*:
Ul cc
5 3P
o Ul
O 00 c ~CL. l- '.3. 0.-0 C
c O O Ou
a o u o Cl . ,rt O SH
oocc
o.
ic o oo
U CA u,
O u* cd
D 01
c C/3
a>
O.T_3, CL co c
a> X
cd (A -a -a
<0 --
cd
Cl 10
cd
~ 8 .E i=' -a
C o
0> 0)
0> CL O
c0
g|-e
o X 4--* E tfl
o S'0'r 'x) ^2
trt O 3 CL O
C_>
cd 4 0> tUOT> OU. r--
_cd ~CL c
0 00
SCcdLf-Do
G 60, >
.52 J5
'5 .s
O *a *"
Is!
cm cd
ox CC C
X E
CL.T3 *"* CA O O
OC
rk as 3
JJ c o--
pi
Mu 0>
9 *hZ
a 2 a> c
c
01
c o o 00 0 01 JD o <
O
c
o
a o
00
O
x co
3
JC 3 coa ian ajgs
0--0 = o 0> 2~ o
2E S| ^ 3 c --
^SSe
5 T3 tj *o -a
Cl 5 2 2 2 ti o o o o <5252
u fj oiX) -c 5 <ft , ~
2 E 2 = = E?J Q. <j ^
<5-0
JDo
X
a <0
* ?w
ii
2 !3 C3
| S|
Co 5o O* = O Qc. U
-- 'c= TJ*
C
oE
t "o '
5
,, ^CB ;2
1> 52.=3
QCO
c52 = a
E
OOO <0 <n O
n<rt x<0> xm> --2
< < < u.
?5^>2
s-.s; X CB 2
tm o OO
>
5 0a
3 O
a a o*.
a a o
a a ok>
o
0O
o
2 3
o
?: c 8|
cOo 2o --c 2
E
Q. 0
or
LU
c
o
o 0 o
.u>
0
>k_ o
w
a
0) 0
a:
01
n
OO
a > O u
zo
e >
03
&|3 =-
sS-5 i:
=
c =;
s 1 CA 2 2 -a -- "o t
P"3Jo Cl aS O c o O S>. oc ?co B3
5Q. 5 2.2. CO K
1 5. . SS E _ 0 EEP
ID
II<
o
Q Ul
2 ui
| i > ..E-s1/3 to *w4CDE0DcD>p--1r. C44(cOtOcCrO-o--dd*! MSuEtaEo*, E*Z
W L tf) O M W V3 o -o o .GC O t3i
^ -S
E"*2 0 3 *cL clu_
o
o,
S3 *T33
^3 c
cn
0u0 J3 00 i_
-- t: u - 1> Jx2 -|O
CL C
jz o 1> C 3 O *^3 i c
o 2 -O E u_
CL O .h: CL N *
3
4> O cd <
r* w
eg
j;
5 is *5 x: 3 E o *x
cS
xtn 3cd e
O -3 r o o E
* cd =* _c
'Z1 <->
:
_cd
^2 c
S.H
o O- 2 0
-- x: , cd < crx)
cn ^
1 .M
! 1
o X
T3 xi S3
.2
0 1/3
X) a> ^
V5 cd 6 ^ cd
m
cd u. l> O
t> w
V) 4J
XO
>
v> v
xo
cd
cd o
3E
> w u- 0C0
o
E-o a>
0^1 x0
c J= o .y
tm u, c
3 <0 d * -d
fN 1/) **--' 3 .S3 o
E "O x
33 o a* Xu
*0 to
cd
o G
co cd
00 o
3 O
J=
<- E _ i-
. 5 => | >,Je
0
cd E
-5 -O
.c5 -oa-
00 G
o >\
-a a. x cd
u O ,0 o
1c/d1 SCll
04 a. o o
2 a> U, 60 O x> *a C t_
Sxo 5; o
3 P17 ox^ r-- V3 .
S -a DtA. JOJ
V_3
^ J2
v) TO '
C 3 o S
,0 -a
o
~ u OX
3 SOi c6 X W
O 00
*"* cn X a> cd c W3 u1 cd cd O ft>
-o
PE OX X_ v) O
-4-J
i CL
fi w
__ P w
-- ` o x .2
H-
-- ii >,
cd
n yx ^w o4>
X U.
cd O
CL Ou
O 5:4 JC3/J -
-a ou. c^d ou X ^X cn * o-- cd X >
<A
Oc ^ no S'S *-* C^
5; Sw o 60
C CL. G o -.
O x' *r >x c ^o E-g >-
^ flJ U J2 c3 c
2 o UJ V3 G 1/3 1) O OO o. .S
60 Lco "o5 > " 3
c/5
oO
co U 1 <U u. G
w cd o _!>U O
"H4X t*" E^
00
2 5 .E
a O
S 00
G cd
-a .
> 60
o G L-
h: o O. _Q
o,, -o -S2 -0 "E 3
O -q X 00 V3 V3 c
00 2
G 1/3
'Si*
X c
2
S
p cd
CJ
WWC > i> --
4^ G .
O > TD
t> ^ 4>
_
CO cd
00
a .5
CL
cd x O cd
'Uo --C ^g CL
W.S LO
j3
1/3
cd G
E ca
< 8 X G cd u o
p o
X
U.. C '3
oS
*T3 3 0"0" D
Eo -a S
-Cl g
8 a> to
E'"5 E
c
0o
O
Eo u
O Q.
o cd o
CL
iC:L ato uX CL
X o
X op
cd ^ L, 00
E^ D
4i C cd
'5 ?
"e -*
*5 X oo
4> o X X
o i *
5 J=iAU^i *
o
c2 CO 3
E
< CL O u.
X
' !> U XX
3I.'3
cyo O
D x
8/5
^-i
_cd
G
63 .*S*3*
s x2.
TD 44->. O "
T3 *cdc
cd
CL
u Ul
a> cd
u. qj
to
On to r* --OC
CLX
O o
Q> X
X u X
2 c
P1
Ex
^ c
= p cr o
a> E
5S? tcod X "iO-> x E
u X ^ 03
o o
ID
oj
> cd -o
XD
w
cl y to H
DX
^ rthS 5 U. D
-a
c cd
Cd
c 1>
D E
^
o.l X ^
DX
g*2
3 u.
cr a>
Ul
CL
co
Cd
C D 0c
X a> ^03
O to
H
v_ cd
o
O CJ
*-*
u. co DD co _c
glE cd
X CL
c E
o
U,
DD
0c0
DE
X O
O co
,,v
o.-5
EE
| -3
G ZE xL E 0 o cd
O3 -"S 4-4 .
00 cd ^ c tr-->4 .1 00---
T3 D u -a
o.S
O O
c qj
E -a ^ E
<^
cD
X 2
o 8
E
>u,>4C>(N~O
O L- t^,
ti o o t--
gX
o
X bj- *-
CL pj G ^ -i e
v- cd
"cd
C o
c<
00 c
CL^1
cd D ^
XO .2 cd
cd A
^ "O X
XX D
cd D
X 4> L> D
M-S s 00
cd O
c 4-.
cUal D
cd
o
^
> cd
^XT3
D -1 ^
{/) O
X D*-<
X
D a>
OO
X
c X
4--*
'?.
m
to G O
u> D CL
8
O u
C o
D X
D X
D4 U X X
FMSI 05832
A S B E S tO S H A N D LIN G t llU U L D u n
*oo *X3 c .u
E
E
8.a
4j t-. lH <U
cd
53 E
c2
*3 .5
> cd
3C cr o &> o
t- c/j
1>
X c/5
4>
XI
o
t/5 CCS
C/3 ot-- X*-*
cd
-b ao
8" .5
GO o _C ? *S c=
o JI7>
<2 * t) <U C/3 X3 H w
o
J3
ooooo
73 r-ouimo N'T ooeo _ CO *- r- to to cc T3 O
Z
X (0 (0 o a <0
wm *O< as
a.
5c
5d = 5
3oc 2 CO S= < CO > t0
Oto C > CO
3
a> c (0
<
X 3 <0 3
OC *: 0*-oo*-ooi/)to<oooo JT CNCflT-r'OrtOI'OCVJ'~ Sg<r aracararggg-a:u.
o> > to w 03 <
C E
o m
CL
OC
X
1i
>o> h. a
oc
o> c
X*
3 a
m OCOOO*-OOlOCO-JOOO
Ja vn<D-*-or)aoocMO e ^'MftNNOOOJinO'-CO fc 0>a> 4-r-U3tDU3S-tTU.^CN
3 io m
CC <E
t- J o> CC
Z ^^
go
*5 + T3 ^ o
s
5 O ao
< m to co 55s = = qo*
50*2 >05 O Ui
S 55$$$<<^^000
<
d
X3 3
*o
Q
*a o
CC> to tO > ee|
? O
i g
s o
J O
5 O
5
o
5S
o~
z OJ
3 X3
3
tfj to 3^3
to 3
O OO o
3` "
C0 V) to 5 33 3 oo o
3 fc
I 5 3 m % c 2to --c c c c c
a
a
w C P m <T
CG
a o o!o o o O O
O CL 0. 0 o a o o o O O CL
a
U.
3 U.
A
*
t. o
s
s
fc.
a
oc
03 CM <X5O <0 03 <
<D tor- M O) I
CO I m<o f*o------
:5"
T- O O to
OT y~
toos
^ 5
co *T
<
<3
T3 <
O Co TSJ 0--5
Q. CL
>c _ op
Z c-o 5co cuo
<E"
* 3; aP S
5 o. 5 m < -i
3
(0
d
8
5
<
TH E M E D IC A L B U LLE TIN
c "O o ^0
3
O' a.
a:
O
c o
o
> o Z
9 c > jr
S
> T3 o
3 C *.* c o
a
03 O
<
00
o <
IS
a m^ 2o =< it.
CO mo
<
oo E w. 2o 3< U.
a a IwS 2o 5< u.
03 tD 99 OO
O Ooca
CZ3
o
ffl
9 CD
a CD U _ 2o
X
H
o3 A --
a a. a. 33 3 ooo
*W-Q k_
aoio CB O)
< < < <<
03 03 (O
OOO <<<
ffi m a:
oOo <<<
ffi
o1 <
ffi
o <
ffi CO 03
ooo <<<
as ffi co
ooo <<<
cc oo zz
c o z
o
o Q aj O oo O 9O
u3.
mm
200^3 0 o5 = <s
,Ox
o
9
u
9
03 CO o om <<
oo
<<
:== = os >
cc
2o zo
cc
zo zo
cc oo zz
ffi o<
oo <<
cc
zo zo
> >
_01 (J
o 73
a o
CD
- T> C ^ CD C
5) ~c ^ 3E Oo f I
j: r 4) T -Q | O
*2 /\ > )
2 -->
i- 9 CO >
- T3 -C JSg Id rS .9 0 3 0.0 O
ca =*o a
o a
S a.
2I_ ~2
*. a> 3<d
o' c5, oo
-
d o
CB
c
c gc *-2 c s p>
-|oo|"|SId?p 31 "H os
d
-o S& w 40 2 ~
2o n O 3 cOO
2-- oM. jr 3 tT O . Oo o> c 2 -03 S OO
<2
Od
8
a a
CO
o T' O o c2 O3 oo
C t>:
ooto.
iuc
?E5 ?oo to
f3>C 9c BrS oS Q.O O
<
2
o B
OD
-o
*5
9 A
0)
u.
3co (f)
ga-cn> c. i- $
*w > aj i 3= > O'- --
Hc rs>g2
5 2-5 jd c x- Oa J_O oo 3<0Oo O! Q-J7 O 5
^a
* c. > Ot --> >- oa g ^ o'! jOffl3
k.
7O- d 0in---5 a> i
'^I
^
3 * CL
C 06 b o-H WXXJ'D
3 `C
Sg.S5H CL
.E O 3 c/3 <0.03
Mil
<moa
-- C
^O T3~
o -Q c ca S
t5 2
IS3g
r: o
t ty
* 15
O
a
--a 7c o c
*sx"z: ;a 7^ < 3
E |B 3 OJ
o
SS c
C 3C a
go <B JD 5
0o*5>
<c8 fc oo
t? Q 2
2 O i?i
tn ca 2 o co
'^3 5^33
QC >
3oC
>5
oI
Z r-
FMSI 05833
>
eo *0 a5
IS'-U
33 lo o
'c
cn (/)
2 w
.
1
6 5
Ou C
.2
-4o> 4) XI n tC!3 cQ
2 3
g E 60
-a.S
o S-2
Oo
*
o C_
<z O 00 c3 co 2
X g ti
CO .2 ^ 3
o
-- to w#v 3 o> GO
H u. v) C
<0 U1
o> o 'C
ta C 3
co H
* w to
4> a
5P*a a>
2gg
>u> 5 > O
Gp ,, .S s ' 5 '
jd *a <2
o <o
^o
S ts
cx-~ 3 Oto. CO o C *-
.2 o
4^1 tO 3O W 4= CH
4) . SV
0:2
uo
ss
o
4>
-Q O
h
1,1 " OS
<DC C '--' -~ __ u
a 8ES
-----------
b. b* ~ OO*-- js .
xa?> _ CQ > S, .
S u- l_ <->
D. 4> -
D-.ri <
2 e*s
(4 u (A
3 jD
`a.
3
O.
=
3to
srV 4) 43
>1 cd
3 a o o
0$ o cx > H
13 O
'O4-> 3 b* "EL to 4>
3 "O
eg eg 3
a
U
U* cd jd
0 a>
ex
33
to
CX
to
CX
lu -C
&
JS *-*
cd o
c
a>
cd u
c o
5
o a>
*0 o
c
e cd
cd
b O
3 b
*5.
o U
2
TT
to 4> b w to 3 T3
a> 00
rs c
o
Q *
eg u
JJ
a c
<
ra a
mX .
o -o
00 C
a> w
: 4- c
j a > uj
E to
eg w
---* to
3 U- "U
b O +- 3 b
*s.
to 4> b
to 3
T3
4> ^ 00 *=
j ^5
o u
^t
crt. --
S-a c _^ oH O ^ .E^l
eg ^ a3 C X3
1S
^v .2> Su Ku --0 o JZ 3 C 1/5 Q(J C/D UJ
2cj
2 u-
v~. U- O
*TD wO w
d
n U
4 I-
EE ^n 5oh
b O 3 u> 'Ex to 4> b 4-* 3 -o
X 4>
Ll.
cflQU(/)UJ
*d
d U t 0 Si J3
3 a
H
eo Sm H u2
5
<4 0- 02 . o
lz ?
15 "O
O 0 +- c:
b 0
3 b *SCO u b. 4- CO 3 *0
3 4) to 3 4) c
ex,
d J
6 V a & Si S.
E a -a c E
<si o. oe . cu
*a S2
vm H -J
4>
tX ii eg XS CQ C
_c o ea e h J4UJ
3
r-
'O m
I J0m Os (N
'g- --
ex ex
JH
z
-C 00
00 c
4 >
3 cd ^ <
0
T3 O
4= c
Os X
43 \rSo 200
otf
c
o
< S 1 I* t^5 Z*gtn tn unh
1U
Z3 m
O .2_j Sd<
o
a
u
S
u
E\oX
tn DC
O 1 3 ca
cn dc
UJ 1-
WQ
1U u
cc Q
QZ
s<
Z UJ < DC
01 3 111 H
eg o
3 5 CQ
a _
eg s o
c
u
o
eg ex
eg
c
ou
eg
-C
2 *-3
o 3
CO
5
to O 3 "3 O w
a,
>*
eg w
erx o
3 eg
CO 3
X3 >1 C
4> 00
3
`C
CO
Z
^Cg V45
o
00
w u. 4) T3 C 3 X 4>
w
3 X3
a
o o
a> > u Q
u cd ex, co >> cd
O
o m
cd
u
4> 4=
3 C
CO
3 43 44 O O b~ CQ JC
r
O z
*0
> CQ
4>
o c c
4> ex
o o tT
o U to o
4 u 3
ex, c
OO
o m
o O
c 4>
\o Tg" o^ c
3
3
4T u
_: c 4)
2 !H
4> f ^ c cX o /> r+s u 3 <N
rn
E3
3 C
o
u
3
ex,
u-v > to o
u
3 a,
0to Tf go c u. CW sM
eg 11
o o* >
v- o
HQ
cu S1 . -2
X o
<
. CN WO
C/3
O'
CQ o`
U
ex,
00 c 3 3 4> cd
<N <N NO X o CQ
O
OU
cj
c
CQ
UJ
4> *o to aj
ut a
:o
*2 ^
ba 3 5 c
o "O -- o
p
.^g-s T3 g
5 " ^ **
*c *- --C
'
co > C T3 ^ O UJ o
4-> X. 5
ex-
>E CO
g T3 c3
ex w ^Cfl 3 t$ ,,
3 co
OS
3
C 5: So ^ 4>
8^ *3 P > CX
-- X O ^
1
K
O Is.
5^/5 TD *3 Q 4) 3
^ X *U
O0c CO X) ^ C E
oc
S2 O
OC
t3c o1/5 w w
C 3
w O
'a. 3 O *>
o u 2 ^ rf <3
*" rt 0S.-0Sh<2
^bO 2 i> -S <A
TM 2 2 ?i
4> c
4> C 0
O 3
5 ix
1:1z r>
z
.2<
s < su. .2 U a <2 -- Eo
Ot <\DN
UO
eg 3 OO 4--1
CX u
Og
go
b.
EW CE
5
<c
sO rs
O U
w 0 h; -- 3 03
'C
a3 3o T3
o -S2
o U
c. >v
-
u
ow
o" Vi C 1> cO CJ ^
CO E
c3?
O U
3 (D
ex
O
3. "3 r^c a\ 4) r-
O
'*g*
a c
4> 00 CO*
ms 000
V <N i--l On
a 3 3
a.
ex
.
>t m 3^ u- On 3 1--1
V) .
gS
.5 c m 4>
2 ex,
*3 o or ex UJ
* cd . C/5 -g -m S^ to ^_ O
(Sz
o 3 *U O
Ou
g
4Q 4) t*_ 3 CO
>
c
o
.2 > < 4-* -- o O
C/3
o .2 c2
to 4--=
QU
- Cg .b c << >3 > --
uC P3 00 00
CD c
rOt
>. t2 .
C 3 ON CO
Q. cd CA p '3^
O fJ XT) W J2 c
a> 3 ^ > ro 4^
*3
O 3 n . O rn wO
a- so
c2
O to eg
rp c
>>
4) 3 _C 5
32 3 ctj -w5 "U. -o
CO 4> 4>
4) 2
,0. --4>=.(n
CO
4> ' J= 3
-O 4) -- *-
d
to to
*3
<g"S-
c ^5
, 4> 3 qj to .
t> o t- O - C/5
^ oX
C . CX &2
EU
rU; 0NO
FMSI 05834
A S B E S TO S H A N D LIN G PRO CEDURE
m
6 Z . t--
cd cd C/5
w 5
*o 4J
o- cd u
'S-
(O O)
*
D -L*_O4
3
T3
O O o
cu >*
-J
t 3,T -js
a
va V3 X
S'
C/3
t> >
as oo
ou, 4) a. c:
oo Zv> oL(J> 0~3 60 Ja JcJ
o 2:ls c
tt. 0 r- C/3 CD UJ
Tf
Tf o o Zu k* ctC c3 U.
0> JC
C/5 *5
--* jd
?
XJ 4>
T3 4>
O
0u3. '5u V) 2
D
3 *T3 O O O <N
a
eS
+0->* s
-J - 5 -J
- s<
-J
eo > wl5
C/3
i_ v- JD O O Ci r -ac
5 C CQ bo o o o _ QC Clir
HO--* cd w `H.
t/3
1> * o>
3 DO X) ~o O O CJ
a
at
O 4) rok-i u cu at
* D
1 Cwu 3
>> o -6<
3
"
w k. Xi
J
Tot toT cTS .fi ..CS *Ca
o o BO 2 J2
o o* J 5 ^
cu a. cn ns uj
cd o
0*3
Ou
o
rvi **
Xs
Tf m
. cu
o:
U
at
?i
-a *2 cd 5G CU V
o o>
i> x JZ -- *-* k-
>, 0>
o -O G
ct--d
XI h,r. *a->*
'cu cn
c
O
3a
0> E
3
TJ O o
>3 -S3-
c 0>
a
.2
w cd
o Jd vQ
cd J= I
*-* M, >< ^ 8 -
Ou* X3
--
* oO
Oh > z
St 3
" 5 >, i> > J3<
co Xo >^. O O- -OC
-j
C/3_ o T3
22:1 wc a, a. t- c/3 cd uj
O 00 5.S U -g
5E
cs -o
3 * X> i "o ,E-2
jc w
60
1 .S
a o
tj o c
u2$ 2 ,S
o l- S-.'O
`B.-0 ^ V5 __ . j> ^ o
. c: Z
va - , . -3a C4) cd
xj cu VO *" ^ VO k-. q
O
<T^22 .X2X cou
S g-s
o
a?
Cl CQ
rt >"0 W2OS S sf-S .5So Saow
TH E M E D IC A L B U LLE T IN .
o C0 cd h u.
*5,
CO co oj <U
33 -a -J
H .1 2
a> S cx >* S| h*
jij -s-I
cd g 5 c-g
X3 .5 -u o g
c S Sg-a Sz^
cd
s
33 X -J
.2 2
1> a a. | 2
H
CU c ^ JsJl
cd u
X 3 'S . C b
',w s2r. 2 -QJQ ucCj
cd
s
Ui 1>
c:
CO 3 XJ
CQ 0> o. >*
*u "*.c
wO _o_ cd
-5
cu tr co 0) w- -o ^ 1>
s*
X3 3 co
> Cu cd o
c5>u ca>
g
S 4= CO *-* T3 '5 * -c >
a w2
.2 Z 85
I "2 Si
-->z
g S C -g S "o 5
TM 42*
22
c O
CD C
n. cd o c 4> c cd
*5 ?
X o *H
u O
cu Jt>r
a>
x> -a
X8
!> G
>m H &> i> Q.
cd H
X C TO
cd 2
22 t 55
2z^i
<30
X o>
cd *3 co CU cd o c 0>
Jw=
XI o->
_
2 NO .x o 0.0
oo
o "2
*u Z
>
u cd
3 e 2
Z
a> Cu a.
>*
h
cuOo.
8 5? OQ o C *.5
r0 A-
cd E g g cz-O
"O
CQ Ssg-a
0) cu
5 z ^ i5
>*
H
"O *0-> cd 3 Cu cd o c a>
o
cd -C
XJ
*uoc-d* `CoI5n>O. --+&u1t -co ^ -
.-(NJ C0>U a >> Eb U c=d t3s in -a
u -J
E c/i UJ
c ^ je oo
B^-Z
2c q ra CQ
E >^ 2 X.-aC
au
l>
jS c r 5 eo a ea s c
U (A UJ c/j LU
3 *o
T3 w4J cd 3 to CU cd O
c o
_o
T3 >> cd
XJ u
o cd
CU co OJ Za>k
*2 s Zj o jC J
E 2> C/3 LU *o ^'O e= *>> jJa eoo
U-
3 -O
0S3'
01
CX,
3 CU
>. H
c *q> oa o*h.
Ea>u Iu _K = jc ! 5 so
rj <1 3 C
U CO W CO UJ
FMSI 05835
A S B E S TO S H A N D LIN G PRO CEDURE
SL o = o
22 5155 ~1 ST^
g^H o.ti c c--
o
to cd
JOx> ?
"O QJ X-- c: rn si
tdg Lx ^
C2 *3 o -- E CO
y ""to* 2
C/> Lx 3 1>
-O
Lx UO
2o
>s cd
E
*a
('XN_x. ^
La o
> s cd
o 3 Lx
Lx > -- o
0.00 to
O'X E
4LJx
Lx -E 2^ .io cd &s 2w
J0 4oj wto Lx Cd
*> xx oo *w-* CL.i2
--* o s >>
-x
2
a:
o ,o <X c
(9 2 TJ c
<3
|>
c o
00 e
-1 U4
JO o* IE .2 * c JD
o a,
3
~u" c^* ^ c s-'s y
SE
.CSL' *> p 3 -Q C rr_- JO -- 4> 2 U
^T3 > wO
Cj
cl. J>
*CL to a> a S
. JO
* "o
-* to T1
O_
WU?tx >W4
O S3 3
tx &0 y
CL V J3
c -2 4J Cd u^
c* *> r* o *,
JO CL
y .2 4>
<H
CC
> !-0c3 fu\
u.
3 ttOo
CL 5 C > y
: .o 3
*5 -w *4 O
..ti'
-io
w-jo
JO cd
,, bO
cl > a co O *G
uC a> -io
J=> -g *"r cd ;
> *a So
o ^ C C <2
^ 2'5 >OJ
oO
oc
; s-5
U S^
v* to cd cd *tD 4> 6 2
GO c
.a >
.s 53 S; g -
o C >, n, O ai >%
eo
c -H
-2 _`y cd
12 O .3
O CL
CO
3 E
*< j= H cU-
2* -t* g^SH Q O& Q erf '
>> 5 y O* CSgi^Ui>jW3fcUCJ0
ri c3 ca o
J ~< p Cd -Q
oto
.53
C >% cO
a 00 CO
> CJ 15 2>
> 1 & S-C
to S' o
jo = w &0 E Sa5
cH
CL
LL H O 2
Lx C3 co CJ 3 '
L. T3 yy
C s CO
3 GO C CL,
< DC 4~, o
n
"U O. 4>
X-- c
JO
3 IS
lx o
-J
<
cO L
T3 y
o "E.
co
QO 1U
XI
o
5
ui X
V) 3
O ^
s<3' < 2
co Lx *CL
fM CQ LO . SO -
I * <3 .. ^
<
-J
co 4J
S U. ^ p u
Q i- t> ra O ^
3 X3
3 ?-S 3 g O
3S Sbi*
g* g = CM <Om J3i5
fN 'O
c2 <^
o, S t*5
oo
T3 C cd
wV
C fvj
x> c cd
Lx o
C3 rsi
4> C CO
cd
JO
FMSI 05836
APPENDIX II
FMSI 05837
ASBESTOS INFORMATION ASSOCIATION/ North America '
1835 K Street, N. W. Suite 402
Washington, D. C, 20005 |
* This study develops a decision method for evaluating the social
acceptability of industrial controls on hazardous materials.
Decisions are based on a "multiple criteria approach" that jointly
considers measures such as risk-benefit tradeoff, minimum
.
reducible health risk, maximum acceptable cost and implicit value
of human life. Health risks are calculated by combining separate
l estimates of production and usage patterns, emissions to air and
water, effectiveness of controls, pollutant dispersion and human
susceptibility. Economicbenefits consider employment, trade and
consumer impacts, as well as direct costs of controls. The analysis
focuses on asbestos as an example hazard. Relative values of
hazard reduction alternatives are examined for asbestos
manufacturing exhaust filters and for asbestos substitutes in
brake linings. Preliminary calculations indicate risk reductions
of these alternatives cannot justify their social costs.
*
i
Risk-benefit analysis for industrial and social needs
KENDALL D. MOLL* and DENNIS P. TIHANSKY: Stanford Research Institute; George Washington University
The term risk-benefit analysis has become a example, value preservation of nature over costs
commonplace expression as both the public and of control; they might think it meaningless to
governmental agencies become increasingly compare ecological parameters with any
concerned with environmental protection and economic concept. On the other hand, many
the quality of life. Attendent with this concept is industrialists place greater weight on thecosts of
the notion that there is some sort of sheet used in preserving nature. U ncertainties or perhaps even
the decision-making process for regulating ignorance of environmental consequences, can
environmental quality.
significantly affect an individual's value system.
Informal comparisons of risks and benefits do indeed occur in setting most regulations. The underlying decision structure, however, does not always incorporate three crucial considerations: the necessity of making tradeoffs; the likelihood that tradeoff impacts are noncommensurate (that is not expressed in the same units, as dollar benefits vs. loss of life); and uncertainty about
Fear of unknown risks often instills greater conservatism in action than is warranted by the objective situation. For instance, a minute but well publicized probability' of death from exposure to a hazardous material might make an unknow ing person think that he will be affected. Aggregated over all individuals, this risk could thus be exaggerated over its actual level.
s impacts of alternative decisions. Formal risk- Publicity about risks and benefits could also
!
benefit analysis is useful here because it be a poor indicator of individual values. Surveys systematically applies economic theory and are thus important in assessing variations of
decision analysis to help the policymaker public opinion, patterns of risk-taking among
understand the tradeoffs.
members of the general population and
Risk-benefit analysis has different compari perceptions of risks and benefits.
sons for different persons. Some ecologists, for
A large spread of personal evaluations should
be expected given data deficiencies on risks and
Present address: Castle and Cooke. Inc.. San Francisco. CA.
benefits. Information about ambient concentra-
&
American Industrial Hygiene AssociaNcn JOURNAL (38) 4/77
I S3
FMSI 05838
i*
1
v
(1
M'.
lions of hazardous materials and the degree of human exposure is not generally available. More fundamental is the lack of information on inherent biological effects of these chemicals on human health. Epidemiological studies have been outstripped by the technology to produce and distribute an ever growing number of hazardous materials. For some new chemicals, the regulatory agency relies on the manufacturer and developer of derived products to furnish health information. But without a character or checklist of needs for conducting risk assessments, the manufacturer cannot really be expected to assume this additional respon sibility.
Insights on benefits of product use are typi cally as deficient as judgements on hazard levels. Conditions for optimality in regulation depend on the "welfare function" to be maximized. One criterion frequently employed is maximum value of total output, which in principle should include such non-monetary aspects as safety, aesthetic qualities and reductions in physic tension or anxiety. Economists generally contend that these benefits of controlling hazards can be assessed collectively as the sum of the "willingness-to-pay" of all individuals affected by the action. Analytically, this requires the formulation of a consumer demand curve for feasible controls. Benefit assessment then involves integration of the area under this curve.
Determination of the overall curve relating benefits to emission control policies requires a multi-step analysis, in which the economics input occurs only after exposures and health effects have been determined. Consequently, the lack of benefit estimates for regulations can be only partly blamed on economists. The lack of meaningful exposure and dose-response relationships falls in the domain of the life and physical sciences. Without accurate knowledge of these relationships, attachment of economic values is a misleading and perhaps futile exercise.
Most regulatory agencies as well as researchers lack an appreciation of the complexities underlying benefit estimation. Asa result, they tend to establish controls which minimize risks without explicitly considering product benefits. "Zero-tolerance laws," which
Figure 1 -Risk/benefit methodology: Analytical Steps.
prohibit any detectable amount of a hazardous material, are inflexible and can quickly become too costly. They may obligate large budgets; they may remain in effect over an extended time span; and they may adversely affect the lives of many individuals.
Although risk-benefit analysis should make decision-making more comprehensive, there are a number of limitations to its current usefulness. First, it represents an abstraction of reality and is usually simplistic. It may not be translatable into practice if it does not adequately account for all impacts. Second, the magnitude of uncertainity about both risks and benefits is typically much greater than the magnitude of their most likely values. This difficulty emerges as the most severe constraint of any formal analysis. Proxy measures are often used to measure impacts, and these proxies may be quite crude. Some authors, for example, represent risk in terms of death rates.' While vital, this measure fails to incorporate non-fatal results that are likely to be far more prevalent and in some cases more damaging in total than the number of deaths.
Another limitation is that risks and benefits of various actions do not usually come in single pairs. Rather, each control level implies a diversity of beneficial and costly impacts. The analyst, with limited resources and time, must
154 Am. Ind Hyg. Assoc. J (38) April. 1977
select the those o-' more, i another; risks. Ini evaluation needs and
There; limitatioi Because 1 risk-benefi vehicular: for auto investigati of hazard: struck h person-' and a i individual ratio.1 Thi as the mo willing to
Anothe represer
conn:
risk^ corn; additioi. claims, be public refi the existt acceptanc was not d
In the r new appri incorporal
American Indus'
FMSI 05839
STANDARDS:
reps.
dous ome they pan; .iany
nake e are ness, nd is into or all linity tuich ikely eve re
roxv
>, and hors, death !s to to be more laths.
fits of single ies a . The must
till. I3T7
select the most important impacts and exclude those of hopefully less importance. As he learns more, he may change the list of impacts. Still another limitation concerns weights assigned to risks. Individuals seem to waver in their evaluation of generally low-probability events as needs and perspectives fluctuate.
There have been few attempts to consider such limitations in formal risk-benefit analyses. Because highway data are readily available, risk-benefit studies are most numerous for vehicular accidents and choices of safety features for automobiles, (e.g.).2'5 One extensive investigation of risk-benefit ratios for a number of hazardous events, such as driving and being struck by lightning selected "fatalities per person-hour of exposure" as the measure of risk and a benefit index defined as "value to the individual" in the denominator of a risk-benefit ratio.' The benefit in most cases was quantified as the monetary investment that consumers are willing to support for a risky venture.
Another more formal, theoretical model represents benefits of an item as the price a consumer is willing to pay for it.4 As hazards or risks increase with item use, its value declines correspondingly. However, there is an additional "non-pecuniary" part of risk, he claims, beyond which the consumer or genera! public refuses to buy or use the item. This implies the existence of a threshold level of risk acceptance. Whether such a level actually exists was not demonstrated.
In the remainder of this paper, we present a new approach toward risk-benefit analysis. It incorporates risk and benefit comparisons, as in
the above referenced s' idies. But it goes beyond their scope by including uncertainty factors and by showing how various decision-making approaches can result in different priorities. The conceptual model to be presented was applied to a selection of feasible controls on hazardous materials, namely, cadmium and asbestos usage in the United States.5
The methodology follows very closely that recommended in the recent National Academy of Sciences report on Decision Making for Regulating Chemicals in the Environment.6 The most significant aspect of this methodology is its reliance on a "multiple criteria" approach. A multiple criteria approach involves the consideration of many factors bearing on the decision in addition to those of "risk" and "benefit," such as "minimum reducible risk," "maximum socially acceptable risk," "maximum acceptable cost," and "value of human life."
As shown in Figure 1, the analysis starts with a review of present standards and proceeds to parallel examinations of the existing pollution system and alternative controls that might be applied to it. The parallel approach continues with second stage investigations of economic benefits on one branch and of population exposures and resultant health hazards on the other. Finally, the economic benefits and the health risks are combined with other relevant parameters in a graphical presentation of the multiple criteria affecting the decision. This presentation is applied under a variety of possible decision rules to derive acceptable standards and develop priority research needs for future improvement of the decisions.
American Industrial Hygiene Association JOURNAL {38) 4/77
155
Ti y .j * C
?
i
t
3
FMSI 05840
CONTR: 0 R 1 G!:
CONTR: EFF
NATIC
Figure 3-U. S. Asbestos How (metric tons per yearj.
When we examine the top box, Present Standards, in more detail, we discover the kind of control system illustrated in Figure 2. This figure shows how various types of standards impinge on the hazardous waste system and on the monitoring and control mechanisms that accompany it. Standards must be compared with the actual amount of hazardous waste present at a given point in the hazardous waste flow system. For example, production standards must be compared with some measure of the actual production. The difference between the standard and actual production is used as a signal for the monitor and its control system to implement corrective measures if necessary. The same kind of feedback control must operate for usage standards, emission standards, ambient air or water standards, exposure standards and ingestion standards. Note from the figure however, that controls do not have to be applied at the same point that the monitoring signal is observed. Excessive exposures, for example, could be corrected by more restrictive controls over emissions concentrations. These multiple feedback possibilities allow for very complex control systems.
The multiple feedback loops do impose two requirements on the monitoring and control
systems. Whatever standards are developed must be measureable and they must be consistent with other standards that they may be applied at other points in the control process. For example, the OSHA asbestos standard was reduced from 5 visible fibers to 2 fibers per milliliter on July 1, 1976. However, the visual methods currently used to count asbestos concentrations account for less than 5% of total fibers and no one knows whether the visual identifiable fibers affect health more than the smaller fibers.7 Standards cannot be made very precise until this conversion problem is solved.
When we decide that some measure of pollution, say weight, is most meaningful, then the second step is to analyze the industrial pollution of various media (air, water, and land) that occurs at different stages of the system, from extraction to final disposal. These amounts are shown in Figure 3. We have used a materials balance approach to estimate the amounts of asbestos going to each use and each disposal media and reconcile their totals with our best estimates of overall use and disposal. The quantities (not all of which are shown in the figure) therefore are additive both vertically and horizonatally along the different flow paths. This requires some very difficult data search and
156 Am. Ind Hyg Assoc J (33) April, 1977
Figur
recor creati How> miss ; the cl plausl systc high't be rei base ;
In t dispo^ pollut tion i fabric exami also h: the se emissii brake effectir polluti steps, manuf fabric contro 96% oi nation, about million
Sub: brake I
American I
FMSI 05841
EMISSION SOURCE:
MANUFACTURING FACILITIES
AUTOMOBILE BRAKES
CONTROL METHOD:
FABRIC FILTERS
SUBSTITUTE MATERIAL
ORIGINAL EMISSIONS:
547
METRIC TONS
129 METRIC TONS
CONTROL EFFECTIVENESS :
96%
100%
NATIONAL COST: LOW
MEAN HIGH
$2.5 MILLION $5
$3.6
y
CO
$52 MILLION $65
Figure 4-Asbestos control costs and effectiveness.
reconciliation work, almost to the point of being creative when the data is particularly sparse. However, we feel it is necessary in order not to miss large parts of the problem simply because the data is difficult to obtain. It gives us a plausible base from which to make an overall systems analysis and, at the same time, highlights areas where existing data may need to be refined. It also provides a first-order priority base for examining alternative controls.
In our analysis, we observe that although land disposal accounts for the great bulk of asbestos pollution, the largest source of air contamina tion is the 547 metric tons emitted from fabrication operations. We therefore have examined controls for asbestos fabricators. We also have examined the possibility of eliminating the second most significant airborne source: emissions from users of friction products (i.e., brake linings). Economic costs and operational effectiveness of control alternatives for these two pollution sources were examined in the next two steps, as shown in Figure 4. For asbestos manufacturing processes, we ascertained that fabric filters are the most effective method to control emissions.8 They can eliminate about 96% of stack emissions. We estimated the total national cost for such a program would run about S2 million per year, plus another SI million for monitoring and enforcement.
Substitution of other materials for asbestos brake linings is speculative since no satisfactory
AREAS EXPOSED " SQUARE KILOMETERS
Figure 5-Asbestos air concentrations.
material has yet been found. But if one assumes that a materia! can be found for a 50% cost increase, the total extra cost would run to about S65 million per year.
These economic costs have to be balanced against the control program's reduction in public exposure and consequent hazard, which we also examined. Figure 5 shows the total areas now exposed in the U.S. to various concentration levels from industrial sources and from brake lining emissions, according to our simplified model of emission sources and air dispersion. Note that industrial sources are more concentrated over a smaller area than are brakeshoe sources. Total exposure to the population is about SO mg/m5 from each source.
The big hurdle in our analysis, as in most such analyses, came in trying to develop a doseresponse function to evaluate the effects of these exposures. To do this, we looked at death rate estimates for the three most important causes of death from asbestos: mesothelioma, other cancers and non-cancer respiratory diseases.
Arencan Industrial Hygiene Association JOURNAL (38) 4-'T7
157
vt'
FMSI 05842
Figure 6-Asbestos dose-response: respiratory system disease
Figure 6 shows excess deaths versus accumulated asbestos dosage for non-cancer respiratory diseases.9 As with most of our data, we had some problems here in converting from one set of measures and dosage assumptions to another. For Figure 6, uncertainties in these conversion factors were the major uncertainties in determining the confidence limits that are shown.
Figure 7 illustrates excess death rates of people exposed to asbestos from all cancers except mesothelioma.9 The Mae West shape is simply a matter of curve fitting and has no known physiological basis.
Figure 7- Asbestos dose-response: cancer !except mesothelioma).
Excess deaths from mesothelioma only are shown in Figure 8. This diagram presents the widest uncertainty of all because its two references10'" both attributed their original data to a common source but differed by a factor of five in their resultant calculations.
By summing these three causes of excess death and combining their uncertainties as independent random variables, we arrived at the overall dose-response curves of Figure 9. We show these for analytical purposes as straight lines emanating from the zero intercept.
158 Am. !nd. Hyg. Assoc J (33) April. 1977
FMSI 05843
although in fact we have no strong basis for assuming that the mean value is a straight line or that the uncertainties are fixed fractions of the dosage.
Mesothelioma appears to be the most significant contributor to the total excess death rale. Asbestosis produces a significant number of illnesses in addition to the excess deaths, but morbidity effects were not included in our analysis.
Hazards from ingestion of food and water were likewise not considered, since most studies of relative hazards indicate that the principal mode of er ' of retained asbestos is via the lungs.K,M
Finally, we did not consider the effect of
population mobility on the hazard, even though
people with high exposures are statistically very
unlikely to live next to an asbestos factory for
their entire lives.
;
But these neglected effects are all relatively insignificant compared to the order-ofmagnitude uncertainties and the other difficulties of reaching a tradeoff between risks and benefits. An overall method of evaluating and presenting risk/benefit tradeoffs is illustrated in Figure 10. Here, the two main parameters are measured in the separate dimension of the chart. The vertical scale shows risk in terms of lives saved by a control alternative relative to the existing "status quo" situation and the horizontal shows negative economic benefit (or cost) involved. Each alternative, surrounded by an ellipse representing the confidence limits, can be shown in terms of this tradeoff of lives versus non health economic benefits. The dollar tradeoff between health and economic effects is not charted, but any particular valuation of human life can be represented by a diagonal line originating at the status quo position. Alternatives lying abov e this tradeoff line would be cost-effective in terms of that particular life valuation, whereas alternatives below the line would presumably not be.
Other constraints can also be shown on the chart; these act to restrict the feasible domain with which alternative solutions may besought. At the top a "minimum reducible risk" line represents a limit in the number of lives that
K.'-LIGRAK - TEARS PER CUBIC NETER IN AIR
Figure 9-Asbestos dose-response: total.
cost (un-health dollar benefits)
Figure 10-Multiple criteria comparison method.
might be saved by any feasible alternative. The minimum reducible risk might beconsidered as a background level of contamination below which further reductions are extremely difficult. At the opposite side of the feasible domain, the "maximum acceptable social risk" represents a
At-;r,c.vi Irst-aiial Hygiene Association JOURNAL <3Si 4/77
159
t
i \i :r
jl"
[i'
'r
I
T %
A
:
%
!
?
FMSl 05844
LIVES SAVED PER YEAR
limits of about a factor of two in each direction. But implementation costs ofthetwoalternatives differ by more than 20:1.
If one wishes to assign a value to saving lives, the chart shows that the factory filter alternative costs less than SIO million per life saved whereas brake substitution costs about S100 million per life. Neither alternative comes close to the S300.000 valuation that workers in hazardous occupations implicity give to their own lives'^'16 (see the upper diagonal line).
Figure 11 -Asbestos pollution control alternatives.
number of lives lost that will not be readily accepted by society. The maximum risk might be some vague social limit such as the prevailing rate of disease, or it may represent a "here and now" risk limit as defined by existing standards and regulations. Another constraint, not usually mentioned together with risk limits but nevertheless logically comparable, is that of "maximum acceptable social cost." This represents the maximum expenditure that society is willing to obligate at a particular time to solve a particular pollution problem. Equity considerations among the population are also now generally accepted as valid constraints. Equity considerations can be considered by making comparable charts for the analysis of each ethnic, income, geographic, generational, or other identifiable interest group. Together, the risk, benefit, and other constraints make up the "multiple criteria" problem.
For this reason, both alternatives appear inefficient as measures for protecting the general public, if economic tradeoffs are considered. As always in such studies, however, we must qualify our conclusions. First, we have not considered potential effects of these measures on the health of industrial workers, which very likely would be more significant than those to the general public. Inhalation hazards to brake shoe installers, for example, would be completely eliminated by substitution of some other material for asbestos in brakes. Second, we have only examined two alternatives. Calculations based on our exposure model indicate that about 37 people per year could potentially be saved by eliminating asbestos from our ambient air (see top horizontal line in Figure 11). This potential life saving is 50 times as great as we get from either of the two alternatives considered, so additional protection possibilities certainly are worth investigating.
These and other factors that go beyond the assumptions of a particular analysis almost need to be considered by decision-makers in real life. Therefore, no single chart can give a complete answer. In addition to the formal trade-off charts of the type shown in Figures 10 and 11. a full presentation should include the listing of many supplementary criteria by which decision makers or other interested parties can derive their own values. Such a list is shown in Figure 12.6
When we apply this methodology to the asbestos control alternatives we obtain Figure 11, which has been derived specially for this paper. The first feature one notices in this logarithmic scaled chart is that lives saved by the two alternatives are about the same: namely, 0.6 person per year each, give or take confidence
Choices among the criteria presented can be made on the basis of many decision procedures, including those of expected value and ordinary old fashioned biases such as optimism, pessimism and probability. So many selection methods arc available, in fact, that decision makers in some ways will have greater freedom
160 dm. Ind Hyg. dssoc J (33i April. \S7;
FMSI 05845
I6 ;s
Figure 12-Display of benefits and costs.
of action than they ever had before. The choice among alternatives may remain ambiguous. But at least decision makers will have explicit, quantitative means for weighing the practical tradeoffs that in the long run are going to have to be considered.
references
1. Staff. C.: Social Benefit Versus Technological Risk. Sci 765:1232 (1969)
2. Lave. L.B. and W.E. Wever: A 8enefit-Cost Analysis of Auto Safety Features. Appl. Econ. 2:1 (1970).
3. Calibresi, G.: The Cost of Accidents. Yale University Press, New Haven. Conn. (1970).
4. Muehlhouse, C.O.: Risk-Benefit Analysis in Decision-making. National Bureau of Standards. Washington, D C., unpublished manuscript {1972).
5. Moll, K.D., S. Baum, E. Capener, F.S. Dresch and R.M. Wright: Hazardous Wastes: A Risk-Benefit Framework Applied to Cadmium and Asbestos. Stanford Research Institute for Environmental Protection Agency (September 1975).6
6 Davies. J.C. ed: Decision Making for Regulating Chemicals in the Environment. Chapter 5 and Appendix H. National Academy of Sciences. Washington. D C. (1975).
7. Background Information on the Development of National Emission Standards for Hazardous Air Pollutants: Asbestos. Beryllium, and Mercury. APTD1503, Office of Air and Water Programs. U. S. Environmental Protection Agency, p. 34 (March 1973).
8. Paddock, R.E. et a/.Comprehensive Study of Specified Air Pollution Sources to Assess the Economic Impact of Air Quality Standards. Vol. II. Asbestos. Beryllium. Mercury. PB-222 858. prepared for U.S. Environmental Protection Agency by Research Triangle Institute (August 1972).
9 Enterline. P., P. DeCoufle and V. Henderson: Mortality in Relation to Occupational Exposure in the Asbestos Industry J Occup. Med. 14:897 (1972).
10. Bruckman. L. and R.A. Rubino: Rationale Behind a Proposed Asbestos Air Quality Standard No. 74-222. presented at the 67th Annual Meeting of the Air Pollution Control Association. Denver, Colorado (9-13 June 1974)
11. Selikoff. I.J.: Asbestos Criteria Document Highlights. ASSE J. (3):26 (1974).
12. A Study of the Problem of Asbestos in Water, by the American Water Works Association Research Foundation. AM. Water Works Assoc. Vol 66, No. 9. Part 2, p. 1 (September 1974).
13. Merliss. R.R.: Talc-Treated Rice and Japanese Stomach Cancer. Sci. 173.1141 (1971).
14. Masson, T.J., F.W. McKay and R.W Miller: Asbestos-Like Fibers in Duluth Water SupplyRelation to Cancer Mortality J. Am Med Assoc. 228:1019 i 1374).
15. Thaler, R. and S. Rosen: The Value of Saving a Life: Evidence from the Labor Market, paper presented 30 November 1973. published by University of Rochester.
16. Melinek, S. J.: A Method ofEvaluatingHuman Life for Economic Purposes. Fire Research Note No. 950, Herts , England (November 1972).
Accepted
23 1976
hJcstnal Hygiene Association JOURNAL (33) 4/77
161
a i.
1 r:
fi ll
>- -5*. ST
EMS! 05846
A method is described tor calculating confidence intervals lor particle or fiber concentration, and for dust collector penetration. The span of the interval depends upon the value of fiber concentration or collector penetration reported and upon the number of particles or fibres counted.
Uncertainty in particle counting and sizing procedures
DAVID LEITH and MELVIN W. FIRST Harvard School of Public Health, Department of Environmental Health Sciences, Boston, Massachusetts 02115
ron
Introduction
The concentration of particles in a gas can be determined by passing a known gas volume through a filter and counting particles on representative filter portions. Particle concen trations are valuable to determine compliance with legal standards, as for asbestos fibers in workroom air, or to determine the particle size collection efficiency of a dust, collection device by making counts of simultaneous upstream and downstream samples. For both applica tions, it is important to estimate the count reliability. Although enough particles must be counted to establish the validity of the result within acceptable limits, it is wasteful to insist upon excessive counting to obtain needlessly high reliability.
Particles for microscopic counting are conveniently collected on membrane filters1-2 or electron miscroscope grids1-2,1. The "stratifi cation" particle counting method2'7 illustrated in Table I is often used to reduce counting time. An initial traverse is made by examining
a representative number of fields under the microscope. All particles seen are segregated into convenient, continuous size categories. Subsequent traverses note entries for only those size ranges in which particles are present in relatively small numbers. The average num ber of particles in each size range per traverse, x, is then calculated as shown in Table I. Stratified counting is a way to emphasize those particles whose concentrations are most diffi cult to assess with statistical reliability because of their relative rarity.
Nomenclature
A -- inverse of the fraction of total filter area examined per traverse
G -- inverse of the total volume of gas passed through a filter
m -- number of an equal area, counting outward from filter center
M -- total number of equal areas into which a filter is divided
David Leith, Assistant Profes sor at the Harvard Uni versity School of Public Health, holds Bachelors and Masters degrees in chem ical engineering from the University of Cincinnati, and a Doctorate in Environ mental Health Sciences from Harvard. His interests lie in industrial hygiene and air pollution control.
American Industrial Hygiene Association Journal
Melvin W. First, Professor of Environmental Health Engineering at the Harvard University School of Public Health, is a diptomate of the American Academy of Environmental Engineers and a Director of the Amer ican Board of Industrial Hygiene. Dr. First was a Di rector of AIHA from 1964 1967.
103
FMSI 05847
;- *
* . *
TRAVERSE
1 2 3 4 5 6 Total. N Mean Count per Traverse,
Std. Deviation of Mean, u-
95% Confidence Interval
TABLE I Data Tabulation for Particle Sizing by Stratified Counting!
< 0.44
57
PARTICLE SIZE RANGE, MICROMETERS
0.45 0.63- 0.89- 1.30* 1.80- 2.50 3.50
0.62 0.68 1.29
1.79 2.49 3.49
4.99
87 54 36 21
24 6 12 12 6
57 87 54 36 21
24 18 18
5.00 7.09
0 3 3 2 2 1 11
7.10- TOTAL PER 10.00 TRAVERSE
3 300 0 21 03 13 13 23 7 333
57 87 54 36 21
24
9
9
1.8 1.2
300
7.55 9.33
72 105 to to 42 69
7.35
68 to 40
6.00
48 to 24
4.58
30 tn 12
4.90
34 to 15
2.12
13 to
4.8
2.12
13 to
4.8
0.55
2.9 to 0.7
0.45
2.1 to 0.3
a N Nilown NUP P Pdown Pup Pt Tin
Tf
X
V
a CTx
CTv t
-- number of traverses performed in a stratified counting procedure
-- total number of particles of a certain size counted through all traverses
-- N for filter downstream of a particle collector
-- N for filter upstream of a particle collector
-- number of particles per volume of gas
-- P for gas downstream of a particle collector
-- P for gas upstream of a particle collector
-- dust collector penetration (1 -- ef ficiency), Pdowr/Pup
-- distance from center of filter to point where microscope is to be focused for equal area m
-- radius of filter -- number of particles in a certain size
range present in one traverse -- mean number of particles in a cer
tain size range found per traverse
-- standard deviation of x
-- standard deviation of x
-- standard deviation of Pt
Microscopic field selection
When a filter holder has a small diameter in let relative to the filter diameter, the largest particles may concentrate on that part of the
filter directly downstream of the gas inlet. Dennis and co-workers8 found that such radial concentration gradients did not occur for particles smaller than 20 micrometers diameter when filter holders were used which had a ten degree included angle between gas inlet and filter surface. However, filter holders of this design are not always practical because of their large size.
To avoid bias when using a conventional holder with small inlet, fields are usually selected at random from a pie shaped piece of the circular filter paper. About twenty fields must be examined2 to complete one unbiased estimate of the particle size distribution. Totally random field selection gives an un biased estimate of particle concentration when a sufficiently large number of fields is ex amined. However, the same result can be obtained with fewer fields by an ordered ap proach. Because the pressure drop across a membrane filter is sufficient to assure uniform gas velocity normal to the filter surface, the volume of gas flowing through each equal, concentric area on the filter will be the same, unless the central areas become plugged be cause of excessive particle deposition there. The overall dust concentration for the gas sampled will be the average of the concentra tions found in the gas passing through each of the equal areas. To utilize an ordered approach, the microscope should be focused at the center of each equal area ring present in the sector, and a single field examined. After one field in each equal area has been inspected, all data can be combined to make an entry for one traverse, as shown in Table I.
104
February, 1976
l
-"S V.
FMSI 05848
> -
l-\ i'
i
I
I
r
Additional traverses are made along different radii of the filter sector.
This equal, area traverse method for locating counting fields is analogous to the method used for positioning a pilot-static tube when determining the average gas velocity in a round duct. Average gas velocity could be determined by measuring the velocity at many random points within the duct cross section and averaging the results found. However, the number of measurements needed to reach the same statistical reliability using this approach is greater than for the equal area method, and the random approach is not used. By analogy, it is as logical to use an ordered approach for locating counting fields on a membrane filter as it is to use it for locating pitot-static meas uring points in a duct.
The distance, r,,,, from the center of a filter of radius rf to the midpoint of each of M equal areas can be found from
Iei -- i2tn 1 r, \ 2M
(1)
Here, m is the number assigned to an equal area, starting from the filter center and count ing outwards. Alternatively, values can be found in a reference giving the relative distance from the wall at which a pitot probe should be placed in order to have an unbiased estimate of average gas velocity in a circular duct.9'10
95% confidence intervals
After stratified counting procedures have been employed and mean concentrations for particles in each size range calculated, it is important to determine confidence intervals for these values. Systematic sources of error such as anisokinetic sampling, inaccuracies of flow measuring de vices, improper microscope calibration and the like can be minimized through careful experi mental technique.11 Assignment of particles to improper size categories is not a significant problem when trained observers use a standard Porton graticule for determining particle diameter.12
A source of non-systematic sampling error that cannot be eliminated by control of experi mental procedures is associated with random variations in the number, x, of particles in a certain size range which are present in each traverse of a stratified counting procedure
comprising in traverses. A Poisson distribution describes the variation in these x values 12 H. The standard deviation of these values, <r, therefore equals the square root of the mean number of particles in that size range per
traverse, x, i.e.
v=^ x
(2)
The standard deviation of the mean, crx , is the standard deviation, cr, divided by the square root of n, the number of traverses.
From Equations 2 and 3,
where N is the total sum of all particles in a certain size range counted through all traverses. Equation 4 is an application to stratified counting of the expression given by Chapman and Ruhf for the relative error associated with repeated counts of particles in liquid suspen sion.15 Appropriate values for the standard de
viation of the mean, 0-7 , are given for the data shown in Table I. The number concentration of particles in a certain size range, P, is proportional to x, the mean number of particles of this size per traverse, to the inverse, A, of the fraction of total filter area examined per traverse, and to the inverse, G, of the total gas volume passed through the filter.
P = IT A G
(5)
The standard deviation of a product can be found from18
With careful technique, the standard error associated with A and G can be made small compared to that for x. The substitution of Equation 4 into Equation 6 gives the relative standard error associated with a measurement of concentration, P.
p _ /_L. -P~ ~ \N
/ (
American Industrial Hygiene Association Journal
105
FMSI 05849
4
Figure 1 -Number of paFtides counted versus relative error of particle concentration.
<5
id
cr SE
0s1
z NUMBER OF PARTICLES ON DOWNSTREAM FILTER, Ndow,,
Figure 2 -Number of particles of one size counted on downstream filter versus number of particles of that size counted on upstream filter, with relative error of penetration as parameter.
until the standard is no longer contained within the confidence interval, as one can then state with 95 % certainty whether or not the standard has been met. When the fiber con centration is close to the standard, it will be necessary to count a larger number of particles to establish with 95% confidence whether or not the standard is met, than when the con centration is clearly well above or below the standard.
For example, after counting 50 asbestos fibers on a membrane filter, one might find that the mean concentration in the air passed through the filter was 1.5 fibers/cc. Equation 8 and Figure 1 show that one can state with 95% confidence that the true fiber concentra tion was 1.5 1.96 X 1.5 X (1/50)'/1, or from 1.08 to 1.92 fibers/cc. The upper bound of the confidence interval for this example is below the 1976 OSHA standard of 2.0 fibers/ cc. Counting additional fibers would make the 95 % confidence interval smaller, but would be unnecessary if 95% confidence that the standard is met is sufficient.
Equation 4 can also be used to determine confidence intervals about an experimentally determined value for dust collector penetration or efficiency. Penetration for particles of a certain size is the ratio of particle concentration in the downstream gas to the analogous con centration in the upstream gas.
The standard deviation of this quotient is16
A 95% confidence interval about P will ex tend plus and minus 1.96 times the standard deviation for P.
P !.96 P v'l/N
()
Figure 1 is a plot of the relative standard error of particle concentration, cri>/P, against the number of particles counted, N.
Applications
When asbestos fibers are counted to determine compliance with an applicable standard, it is prudent to determine periodically the mean fiber concentration and associated confidence interval. The count should be continued only
The substitution of Equation 7 into Equation 10 yields
gpt Pt
UD
Therefore, a 95% confidence interval about
Pt will extend plus and minus 1.96 times the
standard deviation of Pt, as shown in Equation
12.
Pt 1.96 Pt VI/N-, + l/H*.*
(12)
106
February, 1976
I
FMSI 05850
Equation 11 indicates that the relative error in penetration for particles of a certain size is only a function of the total number of particles of that size counted upstream, N,,n, and downstream, N,i,,Wn, of the collector. This relationship is plotted in Figure 2.
Equation 11 and Figure 2 show, for ex ample, that to be 95% confident that pentration of particles in a certain size range is between 40 and 60% (1.96 tm = 10% with mean penetration of 50%) it will be necessary to count 200 particles in this size range on the upstream filter and 200 particles in the same size range on the downstream filter. Alter natively, 500 particles counted upstream and 125 downstream would give the same result. However, the fewest total particles that must be counted to achieve a given relative error in penetration will always be found when the number of particles counted on the upstream and downstream filters is equal. This can be proven by differentiating Equation 11 with respect to Nup, setting the derivative equal to zero, and proceeding in the usual manner. When the particle deposit is less dense on the downstream filter, it becomes necessary to make more traverses for that filter in order to count about the same number of particles as are counted on the upstream. Or, a larger field size could be used for the less dense filter. When the particles observed on a filter are separated into many size categories to de termine collector particle size efficiency, it is ..ecessary to observe a large total number of particles to generate adequate confidence in the penetration or efficiency results for each size range considered.
When it is desirable to maintain a 95% confidence interval of constant size, i.e. a constant value of ere, for all size ranges. Equation 12 shows that fewer particles need be counted in each size range as penetration decreases. The stratified counting technique can be used with good effect to concentrate the microscopist's efforts on the size ranges where penetration is high. Therefore, it is worthwhile to identify these size ranges as soon as possible by making a preliminary estimate of penetration based on an initial traverse of the upstream and downstream fillers.
This method of calculating confidence intervals applies whenever concentrations are
determined by counting. The techniques out lined above can be used for data from auto matic counting devices such as optical instru ments working on light scattering principles, as well as for data from other automatic counting devices.
Summary
An equal area traverse method is described for selecting microscopic fields when counting particles or fibers on a membrane filter. This method is analogous to that used to position a pitot-static tube in a duct when determining average gas velocity. The equal area traverse approach is an aid in avoiding inadvertent counting bias due to nonrandom field selection.
Although confidence intervals are im portant to establish the significance of particle concentration or collector efficiency data, they are seldom calculated or reported. When particle size data are generated by a stratified counting procedure, the method described can be used to establish with 95% confidence whether or not mean concentrations are below or above a fixed value.
Confidence intervals about values of pene tration or efficiency can be calculated in a similar manner. The stratified counting ap proach allows the microscopist's efforts to be concentrated onto those particle size ranges where small confidence intervals for penetra tion are the most difficult to achieve. Charts have been prepared that make it possible to determine easily the number of particles or fibers which must be counted to assure desired confidence intervals.
References
1. SILVERMAN, L,, C. F.. BILLINGS and M. W. FIRST:
Particle Size Analysis in Industrial Hygiene. Academic Press, New York (1971). 2. edwards, G. ii. and j. R. lynch: The Method Used by the U. S. Public Health Service for Enumeration of Asbestos Dust on Membrane Filters. Ann. Occitp. Hyg. 11:1 (1968). 3. morrow, p. e. and T. T. aji rcer: A Point to Plane Electrostatic Precipitator for Particle Size Sampling. Am. Inil. Hyg. Assoc. J. 25:8 (196-t). 4. billings, c. e. and l. Silverman: Aerosol Sam pling for Electron Microscopy. /. Air Pollut. Control Assoc. 12:586 (1962).
American Industrial Hygiene Association Journal
107
5. stcm t. ii. s.: On (lie Size Distribution of Airborne Mine Dust. J. S. Afr. Inst. Min. Met. 58:171 (1957).
6. hoel.p. o.: Introduction to Mathematical Statistics. Wiley, New York (1949).
7. wiiitby. K. T.: Determination of Particle Size Distribution--Apparatus and Techniques for Flour Mill Dust. Univ. of Minn. Eng. Expt. Station Bull. No. 32 (January, 1950).
8. DENNIS, R,, L. SILVERMAN. C. E. BILLINGS, E. KRISTAL. D. M. ANDERSON AND P. DRINKER: Air Cleaning Studies Progress Report for July I, 1955 to June 30, 1956. A. E. C. Contract No, AT(30-1)84) (March 16, 1959).
9. American Conference of Governmental Industrial Hygenists: Industrial Ventilation. 13th ed. P. O. Box 453, Lansing, Michigan (1974).
10. hemlon, w. C.: Plant and Process Ventilation. 2nd ed. Industrial Press, New York (1963).
I 1. HAWKSLKY. P. G. W,, S. BADZIOCIl and J. It. hLackett: Measurement of Solids in Flue Ciayes. British Coal Utilization Research Assn., Leatherhead, Surrey, England (1961).
12. fairs, g. l.: Xll--Developments in the Technique
of Particle-size Analysis by Microscopical Ex
amination. J. Row Microscop. Soc. 71:209
(1951).
'
13. corn, m.: Statistical Reliability of Particle Size Distributions Determined by Microscopic Tech niques. Am. Ind. Hyg. Assoc. J. 26:8 (1965).
14. ill rdan, G.: Small Particle Statistics. 2nd ed. Academic Press, New York (1960).
15. chapman, H. M. and R. c. ruhf: Dust Counting Reliability. Am. Ind. Hvg. Assoc. Quart. 16:201 (1955).
16. auxin, h. and R. R. colton: Statistical Methods. 5th ed. Barnes and Noble, New York (1970).
Accepted October 15, 1975
t
submission Of new
manuscripts ^
The American Industrial Hygiene Association JOURNAL, as the
publication of the Association, provides a medium for timely publication of scientific articles and technical reports covering the broad fields of industrial hygiene and occupational health. These relate to the detection, evaluation and control of problems con cerned with occupational, en vironmental and radiological health, as well as air "pollution, human and animal toxicology, product safety and related sub jects.
Worthy contributions to the literature are welcomed. Manu scripts will be acknowledged and reviewed for acceptance promptly. When approved, they will be scheduled for publication at the earliest possible date.
Only manuscripts submitted exclusively to the AIHA JOUR NAL, not published elsewhere and not being considered by another publisher will be reviewed. Manu scripts not meeting these stipula tions should not be offered.
Detailed guidelines covering manuscript preparation to suit the new format requirements of the JOURNAL will be published frequently. Reprints cf guidelines also are available to authors. These current guidelines should be re viewed carefully and observed, before the final draft of a manu script is prepared. Requests for reprints of guidelines for authors and submission of manuscripts should be made to the editor, American Industrial Hygiene Asso ciation JOURNAL, 66 S. Miller, Rd., Akron, OH 44313.
Galley proofs will be provided for final reading, but not for re writing or revision, just prior to publication. Reprint orders are made available to authors with galley proofs.
Inquiry regarding current status of a manuscript previously submitted for consideration should be made in writing. Be sure to mention the manuscript's JOUR NAL number, if known, plus the first author's name and the manu script's full title.
.1 .
i
1 108
i
February, 197S
I
FMSI 05852
V..
r. 1975 :nt in
V. P. Health : Anal. d. Ily'ygiene SonnaMark:
Counting Asbestos Fibers by the Most Probable Number Method
PARKER C. REIST, Sc.D.
Department of Environmental Sciences and Engineering, University of North Carolina, Chapel Hill, North Carolina 27514
A procedure for evaluating asbestos fiber counts is described wliicli uses the most probable number method of bacteria counting. This technique is faster than conven tional counting methods, with approximately comparable accuracies, although it suffers from a lack of rigor and requires the observer to estimate fiber concentrations to within an order of magnitude before counting. For the routine assessment of a large number of asbestos samples this procedure would seem to be more desirable than conventional counting because of the economy of time as well as being easier on the observer.
Ini rod fiction
Recent findings on the toxi city of asbestos have led to increased interest in sampling and analytical proced ures for determining the concentration of asbestos fibers in air. The Occupational Safety and Health Administration has estab lished an interim eight-hour time weighed average airborne allowable concentration of five asbestos fibers greater than 5 microns length per cubic centimeter of air, and this standard will be lowered to 2 fibcrs/cm3 on July 1, 1976.
Evaluation of asbestos fiber concentra tions in air is carried out using samples col lected on membrane fillers and viewed with phase contrast illumination at 400X-450X.1-2 This method is the standard field sampling method adopted by the Public Health Service.
Fibers arc assumed to be distributed ran domly over the filter surface and at least 20 but no more than 100 fields are to be viewed. At least 100 fibers arc counted which gives a 95% confidence limit of 20%. Because of the relatively small amount of sample collected, more than 100 fields would> have to be viewed if one were to find 100 fibers on a 10 minute sample collected
at a rate of 2 liters per minute from air containing 5 fibers (5 microns length) per milliliter, and if the concentration were only 2 fibers per milliter, 435 fields would have to be assessed.3 Of course, the number of fields necessary could be decreased fjy in creasing the sampling lime--ini;the latter case a 90 minute sample would yield 100 fibers in 50 fields--but the flexibility of short-term samples is then lost.
Counting fibers is a tedious and time consuming business fraught with a number of subjective decisions for the microscopist to make. For example, in the case of two fibers lying side by side he must decide when a fiber is a fiber or when it is only a large particle. (It is considered that any par ticle having an aspect ratio of three or greater is a fiber).
If the sample is relatively light, a great deal of information is lost which actually can be used to determine fiber density. Be sides actually counting the number of fibers, there is another quite distinct statistical method which could be used for estimating the number of fibers randomly distributed over a surface, the so-called most probable number method. In this paper the method will be applied to asbestos fiber counting
FMSI 05853
380 May, 1975
anil the advantages and limitations oC its use will be discussed.
Theory
The concept of the most probable number method for csitmating randomly distributed number densities was first described by McCrady4 and more recently by Cochran5 and principally was applied to the problem of estimating bacterial densities. Chapman6 applied the method to dust counting but it was not received with much enthusiasm for reasons which will be discussed later. Most recently the technique has fallen into dis favor even for estimating bacterial concen trations in milk and water samples, mainly because of the advent of more direct membrane filter techniques. For fiber count ing however, the method appears promising because it eliminates the need to resolve individual fibers and with the low densities normally found, gives the same counting accuracy in a much shorter period of time and with much less eyestrain.
Consider a filter of area A which is broken
up into a fields. If there arc m fibers dis tributed randomly over the filter surface, then the average number of fibers per field, f, is
The probability, P, than n fibers lie in a cer tain field can be expressed using the Poisson relationship
P(n) -- - (2)
provided that a is a large number. The
probability of a field being void of fibers is,
from above
P(o) = e->
(3)
If the void fields are distributed randomly
across the filter and L fields are sampled,
the probability that exactly / of these fields
will be void is
Pi(> = /| \l~-D'}P(o)]'[l ~ P(o>] <L
' (4)
Equation 4 gives a distribution of probabili ties which is very small for small / and rises
TABLE I 95% Confidence Interval for a Given Average Number of Particles Per Field
Particles/Field True Average
L = 25
95% Confidence Interval Expressed a6 a Percentage of the True Average (left column, minus; right column, plus)
L = 50
L = 100
L = 200
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 . 1.4 1.5 1.6 1.7 1.8 1.9 2
100 -- 229.1 76.6 -- 156 69.6 -- 111.3 54.8 -- 90.3 56.3 -- 79.4 52.7 -- 76.2 47.5 -- 55.2 45.6 -- 73.8 44.3 -- 75.6 43.2 -- 57.5
75.8 -- 138 58 -- 79.7 49.9 -- 58.8 42.1 -- 59.8 41.1 -- 52 37.4 -- 40.8 35.6 -- 47.4 33.6 -- 38.8 32.8 -- 38.9 32.2 -- 39.7
' -f
56.7 -- 84.9 41.3 -- 51.7 34.1 -- 43.1 31.5 -- 36 27.8 -- 36.2
27.9 -- 33.6 23.9 -- 32.2 25.1 -- 31.5 22.3 -- 31.4 21.9 -- 31.7 23.9 -- 28.7 21.8 -- 29.6 22 -- 26.8 23.3 -- 28.4 22.6 -- 29.9 20.9 -- 26.8 21.4 -- 28.7 22.1 -- 25.6 20.2 -- 33.1 23.5 -- 30.3
43.3 -- 51.1 29.8 -- 35 25.9 -- 28.1 23.1 -- 25.5 21.8 -- 22.6 18.7 -- 22.8 18.8 -- 21.7 18.1 -- 21.2 17.6 -- 19.4 17.4 -- 19.5 17.4 -- 18.1 16.3 -- 18.5 16.5 -- 19.2 15.6 -- 18.1 , 15.9 -- 19 , 16.3 -- 17.9 16.8 -- 19 16 -- 20 15.2 -- 21.4 15.7 -- 20.2
4
A mericai
to somi small a that has Turning concent
The ; pic prex tion. A dctcrmii present the natu fields Si fields cc most pi can be
As tl creased, creases, range fa at vario Equatio 100 fie 30% rcgardle providet density
Figure microsco
FMSI 05854
American Inilttxlrial Hygiene Association Journal
to some maximum value before becoming
small again. The most likely probability,
that having the largest value, occurs at
l = Lc-i
(5)
Turning this around, the most probable fiber
concentration is thus
/ = In (-j-)
(6)
The above development results in a sim ple procedure for assessing fiber concentra tion. A number of fields arc scanned to determine only whether fibers arc or are not present in any given field. Then by taking the natural log of the ratio of the number of fields scanned divided by the number of fields containing no fibers (Equation 6) the most probable number of fibers per field can be estimated.
As the number of fields scanned is in creased, the accuracy of the estimate in creases. Table I shows the 95% probability range for scans of 20, 50, 100 and 200 fields at various fiber densities, as calculated from Equation 6. Thus, for example, assaying 100 fields will yield results within about 30% of the true value 95% of the time, regardless of the total number of fibers seen, provided, of course, that the average fiber density is somewhere around two fibers per
Figure 1. Various orientations of fibers in the microscope field.
381
field or less. Assaying 200 fields will in crease the accuracy of the 95% confidence interval to something less than 20% of the true count.
For the ease of L = 200, the accuracy ap pears to increase with an increasing average number of particles per field. This--will continue until the point is reached where there is a good chance that every field con tains at least one fiber. This occurs when the average number of fibers per field is slightly in excess of three. Thus for this method it is necessary to estimate ahead of time the fiber concentration to within one order of magnitude so that there will be about 0.2 to 2 fibers per field, or, if there appear to be plenty of fibers in evidence, to then use the direct count method. Of course, the number of fibers per field can be easily varied by changing the field size.
Definition of a Field Containing a Fiber
Thus far, for the purpose of development of the theory, fibers have been considered as if they were particles. But they are not. A fiber has length and as such may starrin one field, extend through another or several others and finally terminate in yet another field. Figure 1 illustrates such a situation. The fiber originates in the upper left hand corner, continues through the lower left hand corner and then ends in the lower right hand corner. The fiber in the upper right corner represents no problem.
There arc several ways in which the fiber that passes through several fields can be treated. First, only the lower (or upper) end of the fiber can be considered, and the field that it lies in then is a field not devoid of fibers. In the rare case where the fiber is perfectly horizontal, some convention, such as choosing the left hand side of the fiber would be appropriate. A field would be considered. blank if the lower end of a fiber were not in it. Thus, in Figure I, only the fields on the upper right hand side and lower right hand side would be considered, to contain fibers. Since each fiber is asso-
rp*< m
---
FMS1 05855
382
ciaied with one "lower" end, the estimated number ends would equal the estimated num ber of fibers.
A second approach would be to call a field void only if it contained no fiber ends at all. In Figure I only the lower left hand field would be considered to be void. The most probable number of ends would then be estimated and since each fiber has two ends, the most probable number of fibers would be the estimated number of ends di vided by two. An obvious disadvantage of this approach is that the upper limit of density which could be used is half of what otherwise would be used.
The most reasonable approach is to assess only one end of the fiber. If the fiber is a bundle with a rough end lying on the edge of a field so that there is some question as to whether it is in or outside of the field, then the same rules as those used in particle counting could be applied to determine
May, 1975
whether the fiber is in the field or not.
Fxperimcntal
In order to determine the efficacy of the proposed counting procedure a number of asbestos samples were counted using a di rect counting method and a record was kept of the number of void fields observed. Most probable number data as determined from
Figure 3. Plot of most probable number versus direct count for data from computer simulation for 20 fields.
DIBECT COUNT, FIBERS PER FIELD
Figure 2. Most probable number data as deter mined from Fquulion fi versus direct count.
FIBERS PER FIELD
Figure 4. Plot of most probable number versus direct count for data from computer simulation for 100 fields.
FMSt 05856
1
American hulttslrial llyniene Association Journal
383
Equation 6 arc plotted as a function of (lie direct count which was observed for the samples and the results are shown on Fig ure 2. Also shown on this figure arc the 95% confidence limits for both the MPN method and direct counting. Although there is some spread in the data, the general repro ducibility is apparent. However, it appears that the most probable number method con sistently gives results which arc lower than those determined by direct count. This ob servation is consistent with a similar otic of Chapman's who surmised that the differ ence could be due to failure to see a single particle in an otherwise void field.
For more extensive work a computer sim ulation was developed in which fields of 10,000 bits were assigned particles randomly corresponding to some preset particle den sity. Then fields of various sizes were ran domly chosen and the average number of particles per field determined using the two methods. In addition, the absolute number of particles per field was determined. From this simulation it was possible to carry out non-biased counts using both conventional counting and the most probable number method for samples of 20, 100 and 200
Figure J. Plot of most probable number versus
direct count for data from computer simulation
for 200 fields.
.
fields. These data are shown in Figures 3, 4 tind 5. Unlike the actual experimental data, however, there appears to be no bias toward the direct count information, indicating that the higher count averages noted on the actual direct counts results from a bias inlroduced by the observer rather than by the technique. Similar to Figure 2, error limits for the 95% confidence interval arc shown as dotted lines.
Advantages and Disadvantages
The advantages of the most probable number method arc threefold. Sampling times are shorter, lighter samples with less chance of overlap can be used, and counting times arc shorter. Using fairly light samples for asbestos concentration assessment means that shorter sampling times arc needed in the field, often an advantage to the industrial hygienist. For a given number of fibers ob served, the MPN method implies greater ac curacy if this total number is relatively small. There is less eyestrain for the microscopist since he only has to determine whether there is or is not something there, and not re solve a specific number of fibers. If prob lems of fiber clumping have occurred they will be more evident because of the lighter sample density. Finally, since fields arc be ing counted instead of particles, the counting should proceed at a faster pace. For exam ple, in discussing particle counting by the most probable number method, Chapman pointed out that one observer could deter mine particle concentrations about twice as fast using the most probable number method compared to standard counting methods, while another observer was three times as first using the MPN method. We have not studied counting times objectively, but sub jectively the people in this laboratory who have compared the two methods for count ing asbestos fibers also feel that the most probable number method is much faster.
The principal disadvantage of the most probable number method is that it is not
wr**m~-t *,**,rmsr
FMSI 05857
1
384
rigorous. An observer could,' in theory, ac curately count the total number of fibers deposited on a filter whereas using the most probable number method, even if the whole filter were assessed, the observer would in the end still only have an estimate of the number of fibers present. In addition, for a given number of particles per field, the 95% confidence intervals for the direct counting method arc slightly narrower than for the most probable number method, the effect becoming increasingly pronounced when the average number of fibers per field exceeds two. The accuracy is sufficient, however, for routine asbestos counting.
Summary
A procedure for evaluating asbestos fiber counts is described which uses the most probable number method of bacteria count ing. This technique is faster than conven tional counting methods, with approximately compariblc accuracies, although it suffers from a lack of rigor and requires the ob
May, 1975
server to estimate fibers to within an order of magnitude before counting. However, for the routine assessment of a large number of asbestos samples this procedure would seem to be more desirable because of the economy of time as well as being easier on the observer.
References
1. Edwards G. H., anil J. R. Lynch: The Method
Used by the Public Health Service for Enu
meration of Asbestos Dust on Membrane
Fillers. Ann. Occnp. Ilya. II: (1968).
3.Joint AIHA-ACGIII Aerosol Hazards Evalu
ation Committee: Recommended Procedures
for Sampling and Counting Fibers. Amer. bid.
Ilvg. Assoc. J. J6.83 (1975).
3. Annon.: Occupational Exposure to Asbestos,
p. viii-5, HSM 72-10267. U.S. Dept. H.E.W.,
NIOSH. Washington (1972).
4. McCrady, M. H.: The Numerical Interpreta
tion of Fermentation Tube Results. J. Infec.
Pis. 77:183 (1915).
. ;1 .
5. Cochran, W. G.: Estimation of Bacterial Den
sities by Means of the Most Probable Num
ber. Biometrics 6.105 (1950).
6. Chapman, H. M.: Dust Counting by the Most
Probable Number Method. A.M.A. Arch. In-
Justr. Hyg. 8:234 (1953).
Int
E
dot V/ pot WI tivi cm cot cm die tox
cid hui res icil I tre j stu
me for the
ph.
bet ac am me tai
grit
FMSI 05858
i- ,
JOURNAL OF PAINT TECHNOLOGY
Handling Asbestos
Chrysotile Asbestos in Plastics
John L. Myers
Union Carbide Corporation*
Asbestos.Jiaa-xeceived_a great deal of attention and publicity in recent years, especially after it was designated a "target health hazard" by OSHA and a "hazardous air pollutant" by the EPA. Many of the articles on asbestos by the press have been emotionally oriented or distorted and, in some cases, stories have been sensationalized, based on obvious misinterpretation of facts. The use of half-truths or unsubstantiated statements has led to general con fusion and the unfair castigation of asbestos and products containing asbestos. The purpose of this paper is to put the matter of asbestos use and asbestos hazards into a logical and practical perspective. In this paper, the different types of asbestos and their many uses are discussed, along with government regulations controlling the use of asbestos. The health hazards associated with asbestos, both occupational and environmental, and some industrial ex perience with air sampling and dust control measures are also covered.
KEY WORDS: Asbestos; Plastics; Air pollution; Toxicology.
What is Asbestos?
Asbestos is a commercial or generic term used to describe six naturally occurring "asbestiform" minerals that are fibrous, hydrated metal silicates. The six varieties are divided into two classes--serpentine and amphibole-- based on their crystal structure. Chrysotile is the only member of the serpentine class, while the amphiboles include crocidolite, amosite, anthophyllite, tremolite, and actinolite. Chrysotile is by far the most-used variety and accounts for over 9555 of U.S. consumption, as noted in Table 1.
Crocidolite, also known as blue as bestos, is imported from South Africa. Because of its high mechanical strength and good resistance to acids and alkalis, it is used to reinforce a limited variety of plastics where its pronounced color is not objectionable. Amosite, also imported from South Africa, is used primarily in thermal
Presented at the Golden Gate Society's Man agement Seminar held in San Francisco, Calif., June 16. 1975.
* Metals Div., Niagara Falls, N.Y. 14302.
Vol. 47, No. 611, December 1975
insulation. Although there are some deposits of anthophyllite in the U.S., most of it is imported from Finland. It is used primarily as a filler for polypropylene and in insulating ma terials. A comparison of the four varieties of asbestos which are of com mercial importance is presented in Table 2. It should be noted that there are significant differences between chrysotile and amphiboles with regard to chemical composition and certain physical properties.
Where is Asbestos Used and Why?
Asbestos has served mankind for more than 100 years in a broad vari ety of applications. The general areas in which asbestos is used in the U.S. are shown in Table 3. Based on infor mation from asbestos producers and consumption surveys, it is estimated that the plastics industry uses about 33% of the 800,000 tons consumed an nually, which makes it the largest single user of asbestos fiber.
The largest uses of asbestos by the plastics industry are in vinyl-asbestos
floor tile and in phenolic molding compounds. It is also used in other plastics such as polypropylene, poly ester, nylon, melamine, epoxy, sili cone, and vinyl. Asbestos provides a valuable function in such products as brake linings, clutch facings, electrical components, automotive parts, furni ture, boats, sealants, coatings, adhe sives and mastics. The most important functions of asbestos in plastics are re inforcement, dimensional stability, heat resistance, flow control, and gen eral-purpose filling. Most of the func tions are supplied by short-fiber chrysotile asbestos fiber, although longer chrysotile fibers and other as bestos varieties are sometimes re quired for particular properties.
Why Use Chrysotile?
Among the several advantages of chrysotile, which set it apart from the amphibole minerals and aocount for its widespread and increasing us age, are world-wide availability, me chanical strength, flexibility, positive surface charge, low iron content, soft ness, and low refractive index. It is conservatively estimated that chryso tile asbestos is used in over 3,000 ap plications and, in most of these appli cations, it is an essential ingredient for which no replacement is readily available.
The information in Table 1 shows that the use of chrysotile asbestos and its share of the total market are stead ily increasing. This is due partly to technical advances permitting the broader use of chrysotile in plastics and the general decline in the use of asbestos in certain fireproofing and in sulating materials. In addition, there is increasing evidence that crocidolite and amosite are more hazardous to
FMSI 05859
4
Year
1967 1968 1969 1970 1971 1972
Table 1--Apparent U.S. Consumption of Asbestos, Tons3
Total
Chrysotile
Amosite
Crocidolite
720,583 817,363 784,321 728,131 758,571 808,554
686,044 (95) 6 775,711 (95) 749,708 (96) 695,770 (96) 729,272 (96) 791,020(98)
12,558(1.7) 20,467 (2.5) 14,618(1.9) 14,261 (2.0) 14,580(1.9)
7,125(0.9)
14,917 (2.1) 13,965(1.7) 10,558(1.3) 8,936(1.2) 6,953 (0.9) 5,374 (0.7)
(a) Information based on import and production data from U. S. Bureau of Mines' Minerals Yearbooks.
(b) Percent of total shown in parentheses.
Table 3--Apparent U.S. Consumption of Asbestos
By General Use Areas
Area of Use Percent of Consumption
Construction Floor tile Felt paper Friction and packing Insulation Textiles Other
40 15 15 14 3 2 11
human health than chrysotile.' Since 1970, the use of crocidolite in Britain has been restricted after ai panel of experts "concluded there was sufficient evidence to suggest other types of fibre should be substituted for crocidolite wherever possible."'
What is the Asbestos Hazard?
It is readily accepted that asbestos, like many other foreign bodies, can cause disabling lung damage (pul monary fibrosis), commonly referred to as asbestosis. This disease and bronchogenic carcinoma (lung can cer) are the two most common as bestos-related diseases. It is important to note that, based on epidemiological data, these diseases have occurred primarily in workers with high, long term exposures to asbestos dust. It is of further interest that one noted re searcher has reported that neither of these diseases is peculiarly related to
or caused solely by the inhalation of asbestos fiber.3 Another important consideration is the relation between cigarette smoking and lung cancer, as reported by Dr. E. C. Hammond and Dr. I. J. Sehkoff.* In this Study, they reported that: "It seems clear, then, that lung cancer is uncommon among asbestos insulation workers who have no history of cigarette smoking and that if the risk is increased, such in crease is not great."
A third disease, mesothelioma, has more recently been associated with persons exposed to asbestos. Meso thelioma is an extremely rare cancer of the lining of the chest (pleura) or the abdominal cavity (peritoneum). In contrast to the lung diseases, there is some evidence that mesothelioma can occur after brief exposures to relatively high fiber levels.
According to the 33-memlber Advisory Committee on Asbestos Can-
Table 2--Comparative Data for Asbestos Minerals3
Formula
Chrysotile
Crocidolite
Amosite
Anthophyllite
3Mg02Si02 2H-0 NauO Fe2Oa 1.5MgO 5.5FeO 7MgO 8SiOa H-O
8SiO, H.O 8SiOs HjO
Composition, % SiO, MgO FeO FcsOa AGO,
H.jO CaO Na-O CaO-fNa.O
37-44 39-44 0--6
0--5 0-2 12-15 0--5
_
-
49--53 0-3 13--20 17--20
--
2-5
_
4-8
-
49--53 1-7
34--44
_
2-9
2-5
_ _
0--3
56-58 28-34
3-12
__
0-2 1-6
_ _
-
Crystals Color Texture Flexibility Hardness, mohs Fiber diameter, A Tensile, mpsi Surface charge Resistance to acid Resistance to alkali
Fine fibers Gray green Soft silky Very good 2.5-4 180-300 800 Positive Poor Good
Brittle fibers Blue Harsh Good 4 600-900 600 Negative Good Good
Prismatic Gray/brown Harsh Good 5.5-6 600-900 200 Negative Good Fair
Prismatic Gray Harsh Poor 5.5-6 600-900
<4 Negative Very good Good
(a) Sources: Modern Plastics Encyclopedia (1973), and Encyclopedia of Chem. Tech., Vol. 2.
oers of the International Agency for Research on Cancer (a division of the World Health Organization): "There is evidence of an association of mesothelial tumors with air pollution in the neighborhood of crocidolite mines and of factories using mixtures of as bestos fiber types. The evidence re lates to conditions of many years ago. There is evidence of no excess risk of mesotheliomas from asbestos air pollution which has existed in the neighborhood of chrysotile and amosite mines. There are reported differ ences on incidence of mesothelioma between urban and rural areas, the causes of which have not been estab lished. There is no evidence of a risk to the general public at present."3
The same body quoted above has also concluded that there is, at pres ent, no evidence of lung damage by asbdstos to the general public; and such evidence as there is does not in dicate any risk of cancer resulting from asbestos fibers present in water, beverages, food, or in the fluids used for the administration of drugs.
While there seems to be general agreement that the public is not in any present danger from asbestos, it is also recognized that excessive, long term occupational exposure can cause serious health problems. Also, if man made emissions aTe not controlled, then environmental contamination could could approach harmful levels.
During the past two years, significant
legislation has been enacted by the Federal government to reduce and control occupational exposure to as bestos fibers and to minimize fiber emissions to the environment. Addi tional standards or regulations have been proposed or enacted by many states and local governments.
Summary of OSHA Regulations
The William-Steiger Occupational Safety and Health Act of 1970 be came effective on April 28, 1971, with
Journal of Paint Technology
FMSI 05860
prudent to permit additional con tamination of the public environment with asbestos. Continued use at min imal risk to the public requires that the major sources of man-made as bestos emission into the atmosphere be defined and controlled.' "7
What is Industry Doing?
The Asbestos Information Associa tion/North America reports that, dur ing the past 30 years, the asbestos in dustry has spent millions of dollars to improve mining, milling, and manufac turing methods.8 The establishment of safer working conditions has been a prime target and this work continues unabated and in close association with government agencies and independent medical researchers.
The ultimate goals of the asbestos industry are: Reduction of work-area dust to minimum levels; Protection of workers from asbestos-related dis eases; Maintenance of environmental emissions at levels low enough to pre clude public endangerment.
Air Sampling
In order to comply with OSHA standards and to determine the need for dust control measures, air monitor ing should be conducted in areas where asbestos is regularly handled or used. OSHA standards require that "all determinations of airborne eoncentrat'ons of asbestos fibers shall be made by the membrane filter method at 400-450X (magnification) (4 milli meter objective) with phase contrast illumination."10 The equipment for collecting air samples costs less than $400 and is readily available. A phase contrast microscope can be obtained for as little as $600, or an existing microscope can be modified for "counting" the asbestos fibers in com pliance with NIOSH criteria.11
Air samples have been collected and analyzed on a regular basis by
the asbestos industry for years. Ex cept for a few applications where dust control is an engineering problem, in dustry is finding that dust levels are already within acceptable standards or that minimum changes are nec essary to achieve compliance. Al though data on many asbestos/plastic applications are not available, the summary in Table 4 is tvpical of our measurements of dust levels during asbestos handling in various types of plants and operations. The dust levels reported are ceiling concentrations, and it should be noted that the allow-
Table 4--Typical Air Sampling
Results"
Ceiling
Type Plant
Concentration,
Or Operation
Asbestos Fibers/cc
Floor tile Polyester Phenolic compounding Handling phenolic Caulks and sealants Gypsum compounds
1-3 1-3 2-5 3-14 0-8 2-9
(a) .Source: Union Carbide Corp., from air monitoring reports.
able OSHA level is 10 fibers/cc. In most cases, the TWA exposure level would be well below OSHA stan dards. Most of the data were collected before the installation of any special dust-control measures. Monitoring of ten shows that obviously dusty con ditions are caused by materials other than asbestos. This does not preclude the need for controls, but it could change their scope and facilitate com pliance with government regulations. Because of "bad press," asbestos is frequently ordered out of use without regard to whether or not a hazard actually exists due to air contamina tion. If acceptable dust levels are feasible, there is no need to replace asbestos at the expense of product quality or economic penalty. Gordon Everett of EPA points out that in formation on the biological effects of asbestos is very limited and that the effects of many substitutes have not been investigated at all. Before as bestos is replaced, it should be cer tain that a safer alternative is avail able."
Obviously there are more people exposed to products containing asbes tos than there are to raw asbestos fibers. As noted previously, more than 90% of the ashes tos used in this coun try is in products in which the asbes tos is "locked in" or bound with ce
ment, plastics, or other binders, so that there is no release, or at least no significant release, of fibers in work areas or to the environment. Materials or products with locked-in fibers would include: floor tile, polyester resins, phenolics, sealants, coatings, brake linings, friction materials, rub ber, roofing compounds, and rein forced plastics. Since an abrading ac tion on some of these products could release asbestos fibers, appropriate monitoring and/or control measures should be instituted if it is thought that such action would release fibers.
Dust-Control Measures
Asbestos producers and users are spending a considerable amount of time and money on various dust-con trol measures. Conventional means to achieve minimum dust levels include: filtered ventilation systems on process equipment, local ventilation for saws and similar tools, conversion to a "wetted" operation, leak-proof pack aging, vacuum clean-up, more care in bag disposal and other asbestos waste handling, and automatic bag openers. Unusual innovations include: pel letized asbestos, special packaging, and treated products.
Only short-fiber chrysotile is avaib able as pellets, but this product serves a fair portion of the asbestos market. Pellets not only reduce dust during conventional handling but they are also available in bulk hopper oars and can be transferred and used in totally enclosed systems. Barring leaks in the system, dust in work areas is virtually eliminated. Used in bulk, asbestos pellets also reduce shipping costs, eliminate warehouse storage and handling, facilitate automation, reduce clean-up, and eliminate bag handling and disposal. The pellets contain no binder and are friable enough to be dispersed in dry form or in aqueous or resinous systems with conventional high-shear grinding equipment.1*
Several types of special packaging are currently available, and suppliers consider customer requests for unus ual requirements. The floor tile indus try can obtain asbestos in plastic bags which can be added directly to the compounding operation. Asbestos in bleached paper bags assembled with water-soluble glue and printed with water-dispersible ink can be added directly to paper-making finishes or acoustical tile formulations.
Although "wetted" asbestos is not generally available, most suppliers are working with customers to pro vide "dustless" products. When justi fied by market demand, asbestos can
be treated with water, mineral spirits, glycol, or other materials compatible with the application or system.
Conclusion
Asbestos is one of industry's many raw materials which involves a po tential hazard when not used with reasonable respect and care. Although all forms of asbestos are recognized as hazardous to health when inhaled excessively, there is growing evidence that crocidolite and amosite are more hazardous than chrysotile. Fortunate-
Journal of Paint Technology
FMSI 05861
the following Congressional purpose: "To assure so far as possible every working man and woman in the na tion safe and healthful working con ditions and to preserve our human resources." The Act established the Occupational Safety and Health Ad ministration (OSHA) within the De partment of Labor, which has re sponsibility for administration and enforcement. Research and related functions are handled by the Depart ment of Health, Education and Wel fare (HEW) through the National Institute of Occupational Safety and Health (NIOSH). Five million em ployers and 60 million of the nation's 80 million workers are covered by OSHA. Specifically excluded from coverage are government employees and operations which are protected under other federal health and safety laws. In a news release issued January 4, 1972, OSHA announced a Target Health Hazards Program aimed at improving health factors associated with working conditions. The follow ing five substances were designated to be the focus of initial and con certed efforts by OSHA and NIOSH: asbestos, cotton dust, silica, lead, and carbon monoxide.
At the present time, new standards have been established only for asbes tos, although, of the 8,000 toxic sub stances on the NIOSH list, only 500 are covered by standards and many of those need updating. The new Standard for Exposure to Asbestos Dust was published in the Federal Register, Vol. 37, No. 110, on June 7, 1972. The basic exposure standard is an 8-hr time weighted average (TWA) of five fibers, longer than 5 micrometers, per cubic centimeter of air. The TWA limit is to be reduced to two fibers per cubic centimeter on July 1, 1976. A peak concentration of 10 fibers per cubic centimeter is not to be exceeded at any time. All of the fiber concentrations are those to which an employee may be ex posed without protective clothing and equipment. The first basic require ment of the new standard is monitor ing to determine whether or not fiber concentrations are in excess of the exposure limits. Some asbestos sup pliers provide a monitoring service to customers, and a similar service may be obtained from state health depart ment officials, insurance carriers, or private consultants. The law requires that monitoring be repeated as nec essary to ensure that employees are not exposed to levels in excess of the exposure limits.
Improper interpretation of the reg ulations has created many miscon ceptions about equipment and pro cedures needed to properly use asbestos. If exposure limits are not exceeded, there are no further com pliance requirements except for med ical examinations. Medical examina tions are required for all employees in any occupation exposed to airborne concentrations of asbestos fibers. The examinations are relatively simple and should cost no more than $50 per year, per employee.
Respirators and special clothing are required in the construction trade for the spray application of insulation and fireproofing materials, and for the removal of such materials. This spe cial protection is not required for any other use of asbestos unless exposure l'mits are exceeded. This is also true for other items such as specially equipped tools, change rooms, clothes laundering, and waste disposal. Res pirators are not a substitute for en gineering controls, but the law allows their use while controls are being implemented, in special situations where controls are not feasible or adequate, in emergencies, and for in frequent short-term, job assignments. Caution labels are required on prod ucts containing asbestos except where the fibers have been modified by a bonding agent or other material to prevent dusting during any normal subsequent use or handling. Besides raw asbestos fiber, products which require package labeling could in clude: dry acoustical spray products and joint cements, unsaturated roof ing felt and textiles, and some in sulating products made without ade quate binders. The labeling of a prod uct does not prohibit its use. It should be noted here that in at least 90% of the products containing as bestos. the fibers are solidly locked into the product thereby presenting little danger of dust generation dur ing normal use and handling of the product.'
EPA Standards
Whereas OSHA is responsible for the protection of the worker, the En vironmental Protection Agency (EPA) is charged with improving the en vironment to which the general public is exposed. On March 31, 1971, as bestos, along with beryllium and mercury, was identified as a "hazard ous air pollutant" by the Administra tor of the EPA. National Emission Standards for asbestos were then pub lished by EPA in the Federal Regis ter, Vol. 38, No. 66, April 6, 1973. Although no numerical emission stan dards were established, operating cri teria are prescribed to prevent or limit asbestos emissions to the out side air from asbestos mills, roadways, certain manufacturing operations, building demolition, and the spray-on application of materials used to in sulate or fireproof equipment and machinery. The law further requires that spray-on materials used to in sulate or fireproof buildings, struc tures, pipes, and conduits shall con tain less than \% asbestos on a dryweight basis. This should significantly reduce emissions to which the gen eral public may be exposed, especially in large urban areas.
". . . the Administrator (of EPA) has determined that, in order to pro vide an ample margin of safety to protect the public health from as bestos, it is necessary to control emis sions from major man-made sources of asbestos emissions into the atmos phere, but that it is not necessary to prohibit all emissions.
"In this determination, the Admin istrator has relied on the National Academy of Sciences' report on as bestos, which concludes: "Asbestos is too important in our technology and economy for its essential use to be stopped. But, because of the known serious effects of uncontrolled inha lation of asbestos minerals in indus try, and uncertainty as to the shape and character of the d ...e-response curve in man, it would hr highly im-
JOHN L. MYERS, Marketing Manager for the Calidria Asbestos Group in Union Carbide's Metals Div., received his B.S. Degree in Chemical Engineering from Purdue University in 1951. Joining Union Carbide that same year, he served in the Nuclear Division until 1966, when he be came a Research Engineer in the asbestos group. He was promoted to his present position in 1970.
Vol. 47, No. 611, December 1975
FMSI 05862
ly, the plastics industry uses primarily ehrysotile asbestos and, in most prod ucts, the fibers are locked-in to pre vent airborne contamination. Al though asbestos dust levels are gen erally lower than expected, industry continues to expend large amounts of time and money to further improve the quality of the workplace. Al though the general public is not cur rently in danger, occupational controls are required to prevent future en vironmental contamination.
Chrysotile asbestos is an important and necessary raw material, vital to the nation's safety and economy; and, with proper control, it can be used safely and in compliance with govern ment regulations. Medical, scientific, government, and industrial personnel must continue to work closely to gether to establish reasonable expo sure limits, provide safe work areas, and eliminate any possibility of public endangerment. [~~|
References
(1) Enterline, P. E. and Henderson, V..
Arch. Environ. Health, 27, 312
(Nov. 1973).
(2) Wagner, J. C., Ann. Occup. Hyg.,
15, 61 (1972).
(3) Wright, G. W,, statement before
U.S. Dept, of Labor, Occupational
Safety and Health hearing on pro
posed occupational asbestos stan
dard, March 16, 1972, p. 3.
(4) Hammond E. C. and Selikoff, I. J.,
"Relation Of Cigarette Smoking To
Risk of Death of Asbestos-Associ
ated Disease among Insulation
Workers in the U.S.A.," presented
at meeting of the Working Group
to Assess Biological Effects of As
bestos, International Agency for Re
search on Cancer. Lyon, France
(Oct. 4, 1972).
(5) `Report of the Advisory Committee
on Asbestos Cancers," Brit. J. In-
dustr. Med., 50, 180 (1973).
(6) "Asbestos," National Safety News
(Oct. 9, 1973).
(7) "National Emission Standards for
Hazardous Air Pollutants," Federal
Register, 58, No. 66, 8820 (April 6,
1973).
(8) "Protecting the Asbestos Worker,"
Booklet No. 101D37, The Asbestos
Information
Association/North
America, p 5.
(9) Selikoff, I. J., Industr. Medicine, 59,
No. 4, 21 (April 1970) .
(10) "Standard for Exposure to Asbestos
Dust," Federal Register, 57, No. 110,
11320 (June 7, 1972).
(11) Bayer, S. G., et al, "Equipment and
Procedures for Mounting Millipore
Filters and Counting Asbestos Fibres
by Phase-Contrast Microscopy,"
Bureau of Occup. Safety and
Health, U.S. Dept, of Health, Edu
cation, 8c Welfare (Feb. 1969).
(12) "Asbestos Health Question Per
plexes Experts," Chem. Eng. News,
18 (Dec. 10, 1973).
(13) Myers, J. L,, "Calidria Asbestos Pel
lets," ASBESTOS (Oct. 1971).
Vol. 47, No. 611, December 1975
Asbestos Information Association North America
1660 L Street, N. W. Washington, D. C. 20036
FMSI 05863