EPA-R1-73-005
December 1972
Environmental Health Effects Research Series
v.*... *^^v^£x'X*!%v* «ff^*!%*!v**.'K**v%%*»*!%.v *" *«*""*" v!*>%*?!. v%*«%"«R*i"X*i%*i.»!«!. v%.x v
-------
-------
EPA-R1-73-005
A SURVEY
OF AIR
AND POPULATION
LEAD LEVELS
IN
SELECTED AMERICAN COMMUNITIES
by
Lloyd B . Tepper and Linda S . Levin
Department of Environmental Health
Kettering Laboratory
University of Cincinnati
College of Medicine
Cincinnati, Ohio 45219
Contract No. PH 22-68-28
Program Element No. 1H1099
EPA Project Officer: Dr. Robert J. M. Horton
Office of the Director
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
December 1972
-------
This report has been reviewed by the Environmental Protection Agency and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does
mention of trade names or commercial products constitute endorsement
or recommendation for use.
The investigations reported here were supported jointly by the
Environmental Protection Agency (Contract No. PH 22-68-28), the
American Petroleum Institute, and the International Lead Zinc Research
Organization, Inc.
-------
TABLE OF CONTENTS
LIST OF FIGURES iv
LIST OF TABLES iv
SUMMARY 1
PREFACE 2
PURPOSE 2
PARTICIPATING ORGANIZATIONS AND ACKNOWLEDGEMENTS 3
METHODS AND PROCEDURES 3
Aerometric Sampling 4
Location of Sampling Sites 4
Sampling Equipment 4
Sampling Methods 4
Analytical Methods 4
Meteorological Considerations 11
Biological Studies 12
Population Selection 12
Population Lead Levels 21
Estimation of Alimentary Lead Intake 22
OBSERVATIONS AND DISCUSSION 23
Aerometric Data 23
Dlood Lead Concentrations 31
Urban-Suburban Blood Lead Levels 32
Male-Female Clood Lead Differentials 60
Fecal Lead Levels 61
Results of Analysis of Pennypack Data ^
Results of Analysis of Alpine County Data 64
Results of Analysis of Policemen Data 65
REFERENCES 67
APPENDIX 1 68
APPENDIX 2 69
APPENDIX 3 71
APPENDIX 4 72
APPENDIX 5 73
-------
LIST OF FIGURES
Figure
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Sampling Sites
Sampling Sites
Sampling Sites
Sampling Sites
Sampling Sites
Sampling Sites
Sampling Sites
Sampling Sites
Meteorological
Meteorological
Meteorological
Meteorological
Meteorological
Meteorological
Meteorological
Los Angeles ,
Philadelphia
Cincinnati
Los Alamos .
Houston . . ,
Chicago . . ,
Washington ,
New York . .
Factors
Factors
Factors
Factors
Factors
Factors
Factors
Cincinnati
Los Angeles .
Philadelphia
New York . ,
Chicago . . .
Houston . . ,
Washington
Lead-in-Air Data (1959-71) National Air Sampling Network
(Symbols Indicate 1961-62 and 1968-69 Population and
Air Lead Level Studies.)
Atmospheric Lead Levels by Month for Each Site
Distribution of Blood Lead Concentrations
Mean blood Lead Levels and n5% Confidence Limits by Age
Blood Lead Levels and Corresponding Mean Air Lead Levels
(5)
Page
6
6
7
7
8
8
9
9
13
14
15
16
17
18
19
31
33
43
52
60
LIST OF TABLES
Table
1 Comparison of Analytical Results from Duplicate Analyses
of Wedges from a Single Filter
2 Population Groups Sampled
3 Monthly Mean Concentrations of Lead in Los Angeles, Phila-
delphia, Cincinnati, and Los Alamos, 1968-69
4 Yearly Means and Confidence Intervals of Air Pb Concentra-
tions for Each Station
5 Annual Means of Monthly Average Concentrations of Lead and
Particulate Matter
6 Freouency Tables Showing Distribution of Clood Pb Levels in
Okeana, Philadelphia, and Pasadens
7 Geometric Means and Confidence Intervals of Dlood Lead
Levels Based on a Single Determination from Each
Subject
8 Analyses of Functional Relationship Between Blood Lead and
Age for Non-Smokers
9 Analysis of Variation Due to Smoking Status (Females) and
Location
lOa Relationship Between Mean Blood Lead Levels and Degree of
Urbanization in Paired Locations
Page
10
20
24
27
29
45
48
50
54
55
IV
-------
Table Page
lOb F-Ratios Testing the Difference Between Urban and Suburban
Lead Levels in the Total Sample (Including Previous
Smokers) 55
11 Geometric Means of Blood Lead Levels by Location and Smok-
ing Status 56
12 Further Analyses of Relationship Between Blood Lead and
Smoking Status 56
13 Geometric Means of Blood Lead Levels for Los Alamos Males
and Females by Smoking Status 57
14 Air and Blood Lead Concentrations (1968-71) 58
15 Air and Blood Lead Concentrations (1968-71) (lion-Smokers
Only) 59
16 Freouency Tables Showing Distribution of Hematocrit Values. 62
17 Analyses of Functional Relationship Between Blood Lead and
Hematocrit for Non-Smokers 63
-------
-------
1
SUMMARY
The concentration of lead in the ambient atmosphere was determined at 59
sampling sites in eight American communities during the period 1968-71. Nineteen
sampling sites had existed in a similar survey in 1961-62. At 14 of these sites
the lead-in-air value was found to be higher in the current study than in 1961-62.
The observed annual mean atmospheric concentration of lead varied from 0.14,ug/m3
(Los Alamos) to 4.55 jug/m^ (Downtown Los Angeles). Higher lead values were
associated with urbanization.
The concentration of lead in the blood of specific well-defined populations
was determined. Such populations lived in geographic proximity to specific air
sampling sites. In the three metropolitan areas from which both urban and sub-
urban population groups had been obtained, the mean blood lead levels were signifi-
cantly higher in the former. In two of the three areas the blood lead level was
higher in urban smokers and non-smokers than in the corresponding suburban pop-
ulations, classifi-d by smoking habits. At each location the concentration of
lead in the blood of smokers was greater than that in the blood of non-smokers.
The magnitude of the observed urban-suburban difference (for populations com-
parable in smoking habits) ranged from 0.9,ug/100 gms to 4.5 jug/100 gms. It is
probable that these observations partially reflect lead absorption from ambient
atmospheres differing in lead concentration. There was no significant concord-
ance between the ranking by site of mean air lead levels and that of the mean
blood lead levels prevalent in the related populations.
The observation that urban levels of blood lead are higher than suburban
levels, but that air concentrations of lead are not clearly reflected in blood
lead levels on a general national basis, suggests that factors other than the
atmospheric lead level are of relatively greater importance in determining the
blood lead levels in population groups.
No relationship was established between age of participant and the blood
lead level. In husband-wife pairs, presumably exposed to similar diets and
atmospheres, the males had significantly higher blood lead levels than did the
females. This difference could not be attributed to smoking habits or to
hematocrit levels. The possibility that the difference was due to dissimilar
quantities of foods in the respective diets of men and women was not examined.
Studies of dietary lead levels showed them to be generally lower than com-
monly reported in the literature, lOO^ug/day being a closer approximation in
the population studies than the widely quoted 300
-------
2
PREFACE
In 1958, the Surgeon General of the Public Health Service appointed a
committee to provide advice relative to the health implications of the use
of lead anti-knock compounds in gasoline formulations. The committee concluded
that an increase in the tetraethyl lead content of gasoline from 3.0 to 4.0
milliliters per gallon would not pose a hazard to the public health from air
pollution. It was recommended that studies be conducted to provide more defin-
itive data on levels and trends of atmospheric lead levels in selected urban
locations and on the blood levels of lead in selected population groups'.
In response to these recommendations the Surgeon General formed a Work-
ing Group on Lead Contamination to develop and execute a suitable program of
studies. This Group included representatives of the petroleum, automotive,
and fuel additive industries, the Public Health Service, the California State
Health Department, and the University of Cincinnati. With the guidance and
support of the Working Group air was sampled over a 12-month period at a num-
ber of locations in Cincinnati, Los Angeles, and Philadelphia. Blood and urine
samples were obtained from selected population groups in these cities.
The findings based upon these extensive observations were reported in
1965 in an official Public Health Service document: "Survey of Lead in the
Atmosphere of Three Urban Communities", Publication No. 999-AP-12.2 This
report presented average, maximum, seasonal, and diurnal atmospheric lead
data. These showed, among other things, that the higher lead levels were
associated with areas of high vehicular density, the fall and winter months,
and early morning peak traffic. The study of lead in the blood showed a
tendency toward higher concentrations in groups of persons as they varied from
rural to center-city areas in their places of residence and work. A number
of factors appeared to be related to these observations: geographical region,
tobacco smoking, occupational exposures. The Working Group also evaluated the
influence of inter!aboratory variation, since the required assays were performed
by several collaborating organizations.
It was a primary purpose of the investigation to establish a baseline for
the lead content of the atmosphere and in selected population groups. Those
responsible for the work anticipated a subsequent study, which would reflect
changes in these lead levels which might be attributed to various relevant
factors: vehicle density, lead content of fuels, mechanical condition of ve-
hicles, local climatology, and population patterns.
The work reported here summarizes studies conducted in 1968-72 to examine
those changes, if any, which had occurred during the interval since 1962. The
investigations were conducted by the Department of Environmental Health, Ket-
tering Laboratory, University of Cincinnati with the financial support of the
Environmental Protection Agency; the American Petroleum Institute; and the In-
ternational Lead Zinc Research Organization, Inc. The course of the study was
guided by the Subcommittee for the Surveillance of Air and Population Lead
Levels, which reflects the public health interests of government and industry.
The Subcommittee is a working unit of the Lead Liaison Committee, an advisory
group to the Environmental Protection Agency.
PURPOSE
The purpose of the investigative program reported here was to:
1) Determine whether or not a change in the concentration of airborne
lead had occurred after an interval of 8 years, at a series of refer-
ence locations designated in 1961-62.
-------
2) Determine the ambient atmospheric levels of lead at selected additional
sites in the cities studied in 1961-62 and in other communities.
3) Examine the extent to which levels of lead in the blood of selected
population qroups reflect exposure to lead at various levels in com-
munity atmospheres.
4) Determine whether or not a change in the blood levels of lead had oc-
curred after an interval of 8 years in selected population groups.
PARTICIPATING ORGANIZATIONS AND ACKNOWLEDGEMENTS
The Subcommittee for the Surveillance of Air and Population Lead Levels,
which provided guidance during the study, included representatives from the
American Petroleum Institute, Automobile Manufacturers Association, E.I. du Pont
de Nemours and Company, Ethyl Corporation, International Lead Zinc Research Or-
ganization, and the Environmental Protection Agency. Names of individual repre-
sentatives appear in the appendix. Environmental studies were conducted with
the cooperation of local organizations in each area: Chicago Department of
Air Pollution Control, District of Columbia Department of Public Health, Houston
Department of Public Health, Los Alamos Scientific Laboratory, Los Angeles
County Air Pollution Control District, New York City Department of Air Resources,
Philadelphia Health Department, and City of Vernon Health Department. Private
citizens, commercial establishments, and public institutions generously permitted
sampling stations to be installed and operated on their property.
The collection of biological materials was made possible through the coop-
eration of church and public service organizations.
The design and execution of this study depended upon scientific and tech-
nical competence in many areas. Names of principal participants appear in the
appendix. This study is based upon their individual and collective contribu-
tions.
Financial support for the investigations reported here was provided jointly
by the American Petroleum Institute, the International Lead Zinc Research Or-
ganization, and the Environmental Protection Agency (Contract PH 22-68-28).
METHODS AND PROCEDURES
The program of aerometric and biological studies was conducted in eight
regions reflecting a range of geographic and climatological characteristics.
The selection of Cincinnati, Philadelphia, and Los Angeles was based on the
fact that extensive sampling had been conducted in these areas during a 12-
month period in 1961-62, and that comparisons could now be made between cur-
rent observations and those obtained some 8-9 years earlier. Investigations
of a similar nature were included in the current study in the additional areas
of New York City, Metropolitan Washington (D.C.), Greater Chicago, Houston,
and Los Alamos, New Mexico. New York was selected because of its high vehicle
-------
and population density and Washington because of its lack of industrialization.
Chicago and Houston were added as examples of large "northern" and "southern"
cities, respectively. Los Alamos represented an area with a well defined popu-
lation of over 15,000 persons with minimal exposure to atmospheric metals.
Aerometric Sampling
Location of Sampling Sites
To permit an accurate appraisal of changes which may have occurred since
1961-62 in atmospheric levels of lead, stations which existed in the study con-
ducted at that time were duplicated in the current investigation. This proce-
dure was followed primarily for comparison purposes; the suitability or epi-
demiological significance of these individual sites was not reconsidered. A
comprehensive photographic record of the instrument installations in 1961-62
permitted accurate duplication of station position and orientation. Certain
limited changes were made in the distribution of stations in the current study
as opposed to their distribution in 1961-62:
Cincinnati: Two stations (#34,35) approximately 2 miles apart and 24 miles
northwest of Downtown Cincinnati were established on farms in the town of Okeana,
Ohio, near the Indiana line. The purpose of these stations was to provide a
rural "background" level for atmospheric lead in a locality with a population
available for sampling.
Philadelphia: The station at the Philadelphia Airport (#10) was omitted
because of small biological significance and major changes in the composition
of aviation fuels since 1962. Station #13 at 15th and Market was moved one
block east because of construction at the original site. An additional instru-
ment (#18) was installed on the 3rd floor balcony of an apartment on Ritten-
house Square so as to record representative exposure of a center-city population.
Two additional stations (#19,20) were installed in the suburban area of Ardmore-
Wynnewood, where a population was to be sampled.
Los Angeles: Station #2 was moved approximately 1/4 mile north because of
operational problems at the original location.
In the 5 regions where long-term sampling of this type had not been con-
ducted previously, sites were selected on the basis of relevance to population
exposure, accessibility, and the area to be covered. In communities in which
biological sampling was to be conducted, emphasis was placed on selecting an
instrument location which would reflect, insofar as possible, the usual human
exposure in that general neighborhood. In some locations, elevations of the
sampling instruments did not approximate those of the breathing zones of the
study population. For particulates of respirable size, however, the evidence
is that differences in elevation, such as existed in this study, influence air
lead concentrations only insofar as such elevations reflect distance from the
lead source.
It is essential to emphasize that all stations, although representing cer-
tain types of neighborhoods, elevations, and proximities to major roadways,
were selected arbitrarily. Data from a specific site have meaning for that
site only, and pooling or "averaging" for an entire city has limited validity.
-------
Alternative sites could have been selected to yield other "averages" for each
city. It would therefore not be judicious to make unqualified inference from
the findings of a number of stations in a city or other area to that city or
area; for example, the findings from eight selected locations in Los Angeles
cannot be generalized to the "situation in Los Angeles". Pooled data for a
city nay have validity in certain kinds of comparisons over a period of time
for the specific sites included in the pool. These data, however, cannot be
precisely compared with pools based upon other combinations of sites arbitrarily
selected.
The location of sampling sites in the various regions is shown in Figures
1-8.
Sampling Equipment
The equipment utilized in this study has been described'^). in essence
it consists of a filter holder for 106-mm Millipore filters (Type WS), a vane
pump and motor unit drawing a nominal free air volume of 5.9 cfm, a cooling fan,
and suitable electrical and air connections. The housing is constructed of
plywood and is designed to permit necessary ventilation of the equipment while
attenuating objectionable pump noise. Initial and final air flows through each
filter are determined from orifice meter measurements.
Sampling Methods
Sampling units were operated over a 12-month period at each site. Filters
were changed in most locations on Monday, Wednesday, and Friday. Because of
previous experience in Los Angeles with filter obstruction in the course of 48-
hour weekday exposures, tandem units with automatic switch-over permitted the
collection of 24-hour samples in parts of the City with relatively high traffic
density or industrialization. No problems of filter obstruction were encountered.
Filters were numbered, conditioned in a controlled atmosphere, and weighed
in Cincinnati prior to exposure. Shipments of filters, each in a plastic Detri
dish, were made to monitors in each city. Monitors recorded on each dish ori-
fice meter measurements before and after exposure, the duration of each exposure,
and notes (if any) about station operation. Exposed filters were returned by
mail to Cincinnati in their original dishes. Upon receipt, the filters were
again conditioned and weighed in the determination of total particulate weight.
Calibration curves for each orifice meter permitted the determination of volume
of air which had passed through each filter. A glass template was used to divide
each filter into ten equal wedges suitable for analysis.
Analytical Methods
Lead contained in particulate matter on the filter wedges was determined
in 1961-62 by the dithizone colorimetric technique.^) The three laboratories
cooperating in the study at that time modified the method slightly to suit the
needs of each. Interlaboratory comparisons were conducted to standardize the
analytical bases upon which inferences were drawn.
The developments in atomic absorption spectrometry (AAS) since 1961-62 in-
dicated that greater efficiency in filter analysis might be achieved in the
Text continued on page 10
-------
(N
OJ
a
H
fe
0)
a
ra
,1
H
fcfi
s
H
t(
o,
E
ra
&
a
a
a
c
m
o
c
ci
a
h
01
m
C
0
D
I
4J
H
a
a.
m
o
c
u
a
01
3
C
01
>
<
a
o
18
IK
0
.p
18
0
,Q
J
iH
O
C
o
0
I
01
fi
0
SB
O
18
a.
c
E:
01
0.
c
o
H
4J
18
O
o
M
<
H
o.
i
o
18
O
ci
f7
HI
fl
(Q
^
01
c
18
1
10
18
>
U
4J
C
18
&
4J
C
1
18
01
H
10
g-
in
01
to
3
g
p
a
OS
o
.-I
C
01
a
c
01
o
O
z
w
0)
rH
41
bo
c
2 '
3 tn
tuo w
CJ Ifl
s o
&
D
111
p:
O
f.
4-1
U
O
c
(U
rH
1
91
\ 01
4->
>,
QJ d)
n) c
H ^:
2^
D, O
&
(0 (U
(A
-------
10
.3
O.
CO
w
01
0
ra
O
8
o
I
01
f-l CO
-p
IT)
i-l
&
ID
&
o
a
H
O
C
H
u
PI
2 1
U) -P
H -H
fe co
bfi
a
I
£
I
8
o
o
CJ
!
I
1
-------
o
bJJ
OJ
O
<£> tn
0)
0) -P
H bfi
fa C
P.
TO
W
3
o
f
u
CO
id
P.
HI
8
a
8 §
g
CO
01
ndb
& 3
B
11
CO
&
H
H
0)
O*
X
C
g
co
a
K
s
id
g
10
ft
tn
I
3
m
a
Jj
B
11
O
H
0]
U
a
tj
o
X
r4
£
H
4J
B
U
O
^
m
K
"8
X
H
£
id -i
"4 J->
U B
Vl V
oi 10
§ ~4
i (0
0 V
0 *
H
4i
B
11
O
^-1
n
V
K
§
5
£
H
4J
S
1
«
o
VO
o
p
to
in i
0 in
H CQ
a
ca
ra
U
i
.u
h
id
a.
o
Q
11
U
H
&
1
0
D:
B
HI
H
rH
3
O
,B
4J
i-l
id
1)
S
1
X
^1
id
(X
C
c
£
4J
fi
+J
M
Q
tx
S
X
a
,13
V
n
H
b
*
o
i
o
4J
Q
-I
id
n
a
0)
f.
4J
id
U
E
3
O
U
4-1
CO
10 1J
id £
rp u
U Vi
4J id
co a
sidence
v
a
u
4J
id
>
H
U
O,
u
>
v4
VI
Q
O
4J
id
8
D
V
a
I
-------
00 I
a) ra
f-i a>
u> -H
H W
PN
tJO
c
a
E
co
w
rH
O
o
£
u
CO
O1
H
S
g
0
Ll
01
1
i
03
&
O
HJ
10
Ll
0
.£3
n
V
11
Ll
4J
Ul
rH
rH
1
g
11
rH
X
rH
__
U
Ul
rC
D>
H
E
C
O
W
o
^
1
01
c
(U
01
o
at
S1
rH
0
C
rH
to
1
C
rH
!
(0
c
Ll
O
Li
0
B
1
0
O>
n
H
H
* &
U ID
H 01
S C
U tt
11 Ul
O Q
rH
8
o
Ul
o
c
S
U)
1
I
H
III
H
C
O
V
id
U]
&
o
Ul
u
3
i!
O
c
11
H
rC
O
£
1
I
rH
II)
H
C
<0
p
n)
U)
Schoo.'
g
rH
id
u
,C
4J
g
ul
1
c
o
£
^
n
id
£
ft
|
CO
rH
n
Q
i
o>
2
111
rd
£
0,
8
X
o
X
H ^H
S £
U)
OJ
a
C
O -O
in X
O -rl
a £
c
(U
o
rl
U)
HI
a.
S
a
c
01
o
C
o
4-1
hD
C
H
x:
en
CO
Sin
o>
H +->
M
ft
E
CO
W
-p
c
i
4J
Ll
rt
a
o
Q
rH
ITJ
11
1
C
o
c
-H
Ll
<
rH
ID
a
H
O
H
C
3
£
1
11
3
C
11
CTi
a c
C -H
H rH
a -H
C 3
M m
4J
m
4>
Ul
a
5
i
V
3
C
1
to
u
id
u
I
X
?;
IT)
0)
rH
"8
0
s
c
id
rH
Ll
V
N
4-1
rH
w
id
lie
+J
+J
11
Ll
rH
Id
"71
o
3
O
K
id
H
c
H
Ll
TH
Q
"c
u
D
rH
rH
£
o
1
TJ
10
o
K
rH
rH
-H
X
Ll
H
I
fO
*o
CO
0)
p
m
c
n
a
C
1>
c
c
o
a
n
V
ce
21 ol
Si*1
-------
10
current study with this method. Consequently this procedure was adopted with
special attention to an examination of the relationship between analytical re-
sults obtained by these two methods: dithizone and AAS. Since a primary pur-
pose of the entire study was the determination of changes in air lead concen-
trations over a time interval, it was essential to assure that demonstrated
differences were not due to variations in technique.
Methodological checks were of two types: 1) assays of wedges from the
same filter by dithizone and AAS; and 2) assays of retained wedges from the
1961-62 study by current analytical techniques (AAS).^ In each type of evalu-
ation it is apparent that wedges from the same filter do not contain the same
amount of lead; i.e., there is inter-wedge variation. The magnitude of inter-
wedge variation was examined as a part of the consideration of variation between
methods. It was determined that the precision with which the quantity of lead
on a filter may be inferred by the analysis of one 10% wedge is inconsistent
from one filter to the next. Presumably this inconsistency reflects to some
extent particle size and influence on some filters of deposited heavy particles
which appear on only a few segments of the filter. In the estimation of lead
on a filter from one 10% wedge, the 95% confidence limits may be as narrow as
97-103% of the mean for the entire disc or as wide as 67-150%, depending upon
whether one uses the lower or upper estimate of interwedge variance. In a
general pool of analyses, however, it was determined that the 95% confidence
limits for the lead content of a filter disc as estimated from a 10% segment
is 85 to 120% of the mean. For most stations approximately 150 10% wedges
were collected over the course of 12 months; the site average is based upon
150 analyses.
The relative response of the atomic absorption method as compared with
that based upon dithizone colorimetry was found to vary between 0.92 and 0.98.
This observation approximates the 0.95 average value which had been obtained
from lead determinations on filters artificially spiked with lead (95% confi-
dence limits 0.93-0.97).
To construct a direct relationship between assays in 1962 and in 1969-71,
retained wedges of filters which had been exposed, cut, and assayed at the
earlier date were examined in the current study. There is of course variation
between wedges from the same filter. A variation between 1962 and 1969-71
methodologies exists in addition to the interwedge variation. Table 1 sum-
marizes a comparison of analyses of single 10% wedges in 1969 with single 10%
wedges from the same filters in 1962.
Table 1. Comparison of Analytical Results From Duplicate
Analyses of Wedges From a Single Filter.
City
Av. ratio (%)
1969/1962
No. of
filters
No. filters
with ratio
< 1
No filters
with ratio
> 1
z-Value
Los Angeles 103.47 39 16 19 1.28
Cincinnati 90.86 40 30 6 -6.29
Philadelphia 99.61 40 18 20 -0.21
-------
11
The magnitude of the z-value reflects the degree of difference between the num-
ber of filters with ratios less than and more than one. The 5% critical value
is _ 2.68. It is therefore concluded that the 1969 air lead concentration deter-
minations in Cincinnati are systematically significantly lower than those obtained
in 1962 (z = -6.29), the average ratio being 90.86%. Interpretation of the air
lead concentrations at the Cincinnati sites in 1969 must give this observation
appropriate weight. For Los Angeles and Philadelphia no significant differences
were found between the 1962 and 1969 measurements.
Meteorological Considerations
The previous study (1961-62) demonstrated that storm conditions with strong
winds were associated with low atmospheric lead levels; precipitation alone pro-
duced no significant effect nor did flooding, snow cover, or high space heating
demands on cold days.
To determine whether or not unusual meteorological conditions might have
influenced the observations reported here on the atmospheric levels of lead, a
brief examination was made of the representativeness of the sampling period.
Significant deviations from conditions recorded in 1961-62 might have important
influences upon the conclusions to be drawn from the current study.
Meteorological data obtained at the Greater Cincinnati Airport are summar-
ized in Figure 9. Precipitation during the sampling period was somewhat below
normal, especially in late winter. Heating requirements were slightly less than
normal. Average wind speeds were near normal and higher in the winter than in
the summer months. The overall situation appears to approximate that which ex-
isted in 1961-62 and is not inconsistent with the 30-year average.
Meteorological data obtained at the Los Angeles International Airport are
summarized in Figure 10. January precipitation totalled 9.6 inches, a signifi-
cant departure from normal. Average monthly wind speed for the year was con-
sistently higher than the 30-year average. Average wind speed in 1968-69 was
higher than in 1961-62 except in the month of May. Advisory statements of high
air pollution potential for Los Angeles were issued for May 10-11 and November
3-5, 1969. An increase in total particulate matter and atmospheric lead levels
at these times was observed at most sites. In 1961-62 the relatively high con-
centrations of atmospheric lead in the fall and winter were attributed in part
to relatively low wind speeds during these seasons. For 1968-69, atmospheric
lead levels during these months was again relatively high as compared to those
recorded in spring and summer months, the differential being somewhat greater
than observed in the previous study. The average monthly wind speed in 1968-
69, however, varied less than in the previous study.
With respect to Philadelphia, precipitation was generally below normal and
the 1961-62 level except for the very wet summer in 1969. Heat requirements
for the two test periods and the 30-year average were consistent. Wind speed
in the winter of 1968-69 was significantly greater than the 30-year average; in '
1961-62 wind speed had been less than average. Inspection of atmospheric lead
concentrations as recorded 1961-62 and 1968-69 shows that lower levels or rela-
tively small increases were recorded during these relatively windy winter months
of 1968-69, as compared with the corresponding still months of 1961-62. An .
-------
12
advisory of high air pollution potential in the Philadelphia area for September
23-24, 1969, was not associated with unusually high levels of atmospheric lead,
although total particulate matter increased at all sites.
Corresponding meteorological data for Los Alamos are not published; however
observations during the test year do not indicate major departures from normal.
Meteorological data recorded at Central Park in New York City indicate that
precipitation was approximately half of normal during the summer of 1970. Heat-
ing requirements approximated normal levels except for November and December,
which were cold. The average wind speed for the test period was slightly less
than normal except in December, 1970.
Precipitation recorded in Chicago during 1970-71 was generally well in ex-
cess of normal. Heating requirements were somewhat below normal for the year.
Wind speeds were slightly above the 30-year average in March and October, 1970.
Corresponding data for Houston show that 14.39 inches of rain fell in May,
1970, more than three times normal; positive and negative departures from normal
occurred throughout the year. Heating requirements approximated the 30-year
average. Throughout the year the average monthly wind speed was less than nor-
mal by approximately 20-30%.
Meteorological data obtained at Washington National Airport are summarized
in Figure 15. Precipitation for the year approximated the normal rainfall.
High air pollution potential advisories on April 29-30, June 9-12, July 28, and
August 15-16, 1970, were not associated with significant elevations in the air
concentrations of lead. Heating requirements and wind speeds approximated the
30-year mean.
Biological Studies
Population Selection
A principal objective of this investigation was an evaluation of the rela-
tionship between atmospheric lead levels and the concentrations of lead in the
blood of persons exposed to the atmospheres in question. While a number of dif-
ferent population groups have been studied in attempts to establish this rela-
tionship, multiple uncontrolled variables associated with these groups have cor-
fused the appraisal of the actual exposure and the construction of the correct
inference. In the absence of personal monitoring devices one cannot establish
the total integrated exposure over months or years of police officers, garage
mechanics, "drivers", "commuters", aircraft workers, or other similar groups.
Occupational and wide-ranging geographical factors are not reflected in air
sampling at fixed locations not clearly related to the population in question.
For these reasons particular attention was focused in this study on the
definition of populations with a specific consistent relationship to known air
levels of lead. Populations sampled in this study were derived from women volun-
teers living within a prescribed region surrounding an air sampling instrument.
Women, in general, currently tend to spend a greater proportion of their time
at home or in the neighborhood of their home than do men. Furthermore, women
who are employed at a distance from their homes rarely work in lead-using trades.
For inclusion in the study a continuous residence time of not less than 5 years
Text continued on page 20
-------
13
W
SH
O
P
O
II CO
9> ra d
<-i o d
3 -H -H
& M> o
H 0 d
fe rH -H
O O
fn
O
01
ro
S3HONI
Q
LU
LJ
Q.
CO
Q L
LJ
CD
<
cr
LU
yduu 'O33dS QNIM 39VU3AV
to
h-
z
LU
S
LU
O
LU
tr
<
LU
O O O O O O O
o o o o o o o
o o> oo r- (£> in
o o o o
o o o o
-------
o
LU
a:
o_
1 ' '
ill
O
S3HONI
14
ta
^
o
+j
u
OJ W
O (it
*T-xi 1 1 r
il
V)
cP
ONIM 39W3AV
-------
15
o
LU
o:
Q_
w
fn
O
CO 03
i-H fe -H
rH J3
r-H C,
0) CO i-l
?H O 0)
3 -H T3
M be oj
H O rH
6n i-H -H
O JZ
h &
O
ooi^-tDio^t rooj
S3HONI
UJ f-
^
UJ
£
o
UJ
o:
<
LJ
V
iV
o o o o o o
o o o o o o
OJ O O) tD ^- OJ
SXH7 33U93O
-------
16
o
M
O
cj w o
fc CJ ><
SB ho s
H O 0)
O
0)
if) co
S3H3NI
Q
LU
UJ
CL
CO
Q
z
?
UJ
(T
UJ
I
^
*
CO
if CJ
VOfc/ '
-------
17
~ii r
O
UJ
a:
CL
Q
2
O
tn
SH
O
+J
y
Ct!
<1> ca aj
5n O y
S-H -H
bfi rj
H O o
O
0)
-p
I
rocvj
H<
LU
^
Z)
o
UJ
a:
CD
ui
T.
39VU3AV
o o o o o o o
o o o o o o o
ro CD h- 10 ro
SAVQ
-------
o
LJ
or
Q.
18
1 - 1 - 1
_J i .X _\_
O
CO
to
i-(
o
JJ
o
0!
0)
S3H3NI
Q
LJ
LJ
CL
CO
LJ
O
<
(T
LJ
LJ
2
LJ
cr
o
LJ
(£.
LJ
I
'O33dS ONIM 39VU3AV
o o o o o
o o o o o
if) *fr ro CVJ
SAW 33U93O
-------
19
O
O
LU
o:
CL
~ii
^ i
I I I I
00 h-
CO W)
10
CO
o
a;
Q
LJ
LU
Q_
(O
Q
UJ
CC
LJ
(f)
LfdUU '
CVJ O 00 CD ^ CJ
39VU3AV
i-
UJ
5
ID
O
Ld
cc
h-
<
LJ
X
33U93CI
-------
20
within this area was required. The area was defined either by political bound-
aries ("Pasadena", "Port Washington") or by the circumference of a circle one
mile or less in radius drawn from the reference station. Homogeneity of land
use within each area was reasonably consistent. This study was concerned ex-
clusively with absorption of ambient community atmospheric lead via the respira-
tory system and did not involve pediatric exposures nor exposures of men with a
variety of possible occupational opportunities for additional lead absorption.
The extent to which blood lead levels in women reflect levels in the general
population was not examined.
An attempt was made to sample "suburban" and "urban" populations in several
of the regions under study. In some areas this objective was confounded by the
fact that highly mobile populations and rapidly changing neighborhoods made it
unlikely that a group of acceptable volunteers could be recruited without ex-
tensive personal solicitation. In the same region a comparison of "suburban"
and "urban" populations would tend to minimize the influence of regional factors
which might influence regional biological lead levels. Climate and sources of
food and water would tend to affect equally populations living in the same gen-
eral metropolitan area. Populations sampled in the current study are presented
in Table 2.
Table 2. Population Groups Sampled
Community
Philadelphi
Cincinnati
Los Angeles
Los Alamos
Washington
New York
Chicago
Houston
Urban
Station
a Rittenhouse
Square
Pasadena
fWoodly Park \
^Cleveland Pk.j
Greenwich Vil .
Bridgeport
Nw. Houston
18
4, 5
74
85
63
56
N
137
209
224
140
148
204
Suburban
N* Station N
136 Ardmore-
Wynnewood
Okeana
(rural )
193
Los Alamos
219
140 Port Wash.
147 Lombard
191
19, 20 156
34, 35 166
41 , 42 204
88, 89 203
66 208
N*
150
162
191
198
208
*A11 statistical computations performed on populations corrected to exclude those
persons with possible unusual lead exposure and those who were less than 20 or
more than 79 years of age. See p. 31.
In addition to the groups of women which participated as described, smaller
studies were conducted in four other populations.
1) Husbands of 100 women participants in Los Alamos were sampled to see
whether or not there is a male-female differential in the blood level of lead.
It was possible to define the occupational exposures of these men, most of whom
were employed by the Los Alamos Scientific Laboratory. The atmosphere at the
Loboratory does not vary from that in the residential areas. Husband-wife pairs
-------
21
were considered to consume approximately the same foods, although the quantity
consumed by the husband will, of course, generally exceed that-consumed by the
wife. In the absence of occupational lead exposure and under these circumstances
it was believed that the degree of difference in male and female blood lead levels
could be determined.
2) In 1962, blood levels of lead were determined in 120 Cincinnati police
officers. In 1971, it was possible to identify 65 of these men still on the
police force. Of these 48 volunteered to provide blood samples. The purpose
of this study was to see whether or not blood lead levels had changed in this
population over the course of 10 years. Such changes as might be observed could
be attributed to the effect of aging or to actual differences in absorption over
the intervening period.
3) A group of 87 men and women residents of the Pennypack Woods section of
north-eastern Philadelphia was studied in April, 1961, as representative of a
"suburban" population. Of these 44 had moved from the neighborhood by April,
1971; 5 more had died. Of the remaining 38, it was possible to obtain blood
samples from 23 persons in 1971.
4) The report issued in 1965 included blood lead data from assays performed
on 37 men and women who had participated in a multiphasic screening program in
Alpine County, a rural area in the Sierra Nevada. Most of these persons were
of American Indian ancestry and had lived in the County for most of their lives.
Blood samples were obtained in October, 1971, from 39 persons who had lived in
Alpine County for at least 5 years. This group did not represent to a signifi-
cant extent those individuals who had been sampled previously; however they were
considered to be representative of persons who had been long-term residents of
Alpine County.
Population Lead Levels
Data Collection: In order to examine the correlation between air and blood lead
levels, it is necessary to control those variables which might be expect d to in
fluence the concentrations of lead in the blood. The selection of populations
permitted residential specifications as to location and duration. Through the
use of an interview, other variables were evaluated: occupation, hobbies, smok-
ing habits, sources of food and water, medical history. The principal study was
based exclusively on women, a fact which eliminated differences in blood lead
due to difference in sex, and to a significant degree, in occupation and exposure.
Determination of BJood Lead Levels: Persons in the several populations studied
who met the residential criteria for inclusion and who had been interviewed con-
tributed blood samples for analysis. Blood was drawn from the ante-cubital vein
into lead-free evacuated tubes of 20 ml capacity. Tubes were stored under re-
frigeration until the time of analysis. Assays.were performed according to the
standard dithizone method of Bambach and Burkey'^' as modified by Cholak and
Burkey.*
There is no argument with the position that an epidemiological investigation
of this type would be improved by the determination of the total body burden of
lead in exposed populations. It is apparent, however, that these determinations
or active clinical procedures such as provocative lead excretion tests with
*Method available upon request to J. Cholak, Kettering Laboratory, Cincinnati,
Ohio 45219.
-------
22
chelating compounds require experimentation upon humans in a way which cannot
be currently justified in epidemiological studies. Blood lead levels do, how-
ever, reflect recent major changes in exposure and the body burden of lead,
insofar as there has been long-term exposure at a relatively constant level so
that a steady-state relationship between blood and the body burden exists. In
the situations under examination the likelihood of recent major changes in ex-
posure is small. If the exposure to lead has been constant over a period of
time, it is reasonable to assume that the levels of lead in the blood of these
populations are meaningful indices of body burden for these adult populations.
Estimation of Alimentary Lead Intake
Since lead levels in man reflect intakes from both the respiratory and gas-
trointestinal systems, it is not prudent to make quantitative inferences about
absorption from the lungs with knowledge reflecting only total lead absorption.
Without standardized alimentary intakes of lead or knowledge of its level in
each population, a situation may exist in which differences in blood lead levels
may be incorrectly attributed to variability in atmospheric lead levels. It
is also true that a lack of difference in mean blood lead levels between popula-
tions may be due to compensatory variations between sources of absorption for
those populations. In other words, high alimentary lead and low respiratory
lead levels in one population may produce the same blood lead level which is ob-
served in a population exposed to high atmospheric and low food lead levels. To
infer that different atmospheric lead concentrations have no influence on blood
lead levels is obviously incorrect in this situation which omits consideration
of lead intakes in food and water.
In carefully controlled studies of lead metabolism, it has been possible to
collect accurate duplicate diets of the experimental subjects, for whom alimentary
lead intakes could then be determined. This approach did not lend itself to the
study reported here, however, because of the cost of duplicate diets for large
numbers of people and the difficulty in obtaining accurate duplicates from in-
experienced people. It was determined that information of approximately the same
utility might be derived from measurement of lead excreted in the feces, wh;^ii
contain some 90-92% of lead ingested in the diet. Accordingly 20 volunteers in
each region were recruited to participate in a ten-day excretory collection.
Urine and feces were collected separately in low-lead plastic containers. A
measured aliquot of urine and all fecal samples were shipped to Cincinnati for
assay. To assure total collections only intelligent and motivated persons were
invited to participate. A detailed instruction period and daily collections of
materials by a nurse contributed to the completeness of the collections.
The manner in which, the metabolic study was designed reflects certain as-
sumptions which affect its validity. It has been assumed that persons who live
in a general metropolitan region derive their food and water from essentially
equivalent sources, so that measurements of alimentary lead in Port Washington
on Long Island also describe the situation in New York City. Similarly, measure-
ments in Chicago or Philadelphia suburbs are applicable as well to the respective
urban area. There are obviously differences between individuals in their choice.
of foods, but we believe that the sampling method gave a reliable index for the
region. We further believe that the dietary habits of 20 people over the course
of ten days are representative of the general habits of the parent population
of 200 people.
-------
23
It is also apparent that extrapolation of a ten-day sample to an expression
of annual lead ingestion is not fully satisfactory. It is impractical, however,
to sample throughout a year in the absence of compelling scientific reasons and
participants with special motivations and attitudes. Since the samples used in
this study were collected during times of the year when an abundance of fresh
foods was not normally available, however, we believe that the ten-day sample
does reflect the general situation for most of the year. Only during the sum-
mer, when fresh fruits and vegetables are more readily obtained, might there be
major changes in the food sources of some individuals. We have permitted the
assumption, therefore,that the sampling as described does reflect general ali-
mentary lead intake.
Analyses of excreta for lead were conducted by the atomic absorption tech-
nique; 10% of the samples were also analyzed by duplicate dithizone procedures.
In the case of the fecal samples, which were received and weighed in plastic
bags, the entire sample with the bag was ashed for assay. Lead content of con-
trol bags had been determined and found to be only several micrograms per bag.
OBSERVATIONS AND DISCUSSION
The investigative program described in this report was designed to:
1) Determine whether or not a change in the concentration of airborne
lead had occurred after an interval of 8 years at a series of refer-
ence locations designated in 1961-62.
2) Determine the ambient atmospheric levels of lead at selected additional
sites in the cities studied in 1961-62 and in other communities.
3) Examine the extent to which levels of lead in the blood of selected
population groups reflect exposure to lead at various levels in com-
munity atmospheres.
4) Determine whether or not a change in the blood levels of lead had oc-
curred after an interval of 8 years in selected population groups.
Except in the case of the fourth objective, data were developed which are useful
in answering the questions presented.
Aerometric Data
Aerometric data are presented in Tables 3 and 4. With respect to changes
in atmospheric levels of lead during the interval 1961-62 to 1968-69, the reported
data (Table 5) demonstrate higher lead levels during the latter period at 14 of
the 19 sites which were re-established. A careful examination of experimental
methodology tends to exclude the possibility that this change is an artifact re-
flecting changes in technique. Current assays of 1961-62 filter wedges from Los
Angeles and Philadelphia have validated the lead assay techniques used during
the period 1968-71. Because of the observation that filter wedges from the 1961-
62 Cincinnati study, when re-assayed in 1969, yielded results approximating 91%
of those obtained in the earlier period, it is possible that increases in at-
mospheric lead levels at the re-established Cincinnati sites are greater than
may be apparent. Consequently, increases in the current study can hardly be
Text continued on page 30
-------
24
f»
g
ia , Cincinnati, and Los Alamos, 1968-6;
.c
a
Q)
o
to
rH
H
«
10
a>
rH
0)
hn
UU
c
cn
^
£
H
a
ca
3
OCMOO
otoeoooincnincom tocMrHOtjit~cocorHtocM coooco
o
mto^cococo^in 'tf CMCMincorHrHrHCMCMi i CM CM I-H co
tofocM'
to ^
ra -H 3 "O
tp Q)^^ rH c" ^ ? CO
P bJD "^ ^O cl) CD O CU cj
Q O
w OL a. -H occC2a
Q J3 -H
rH PXl U
CM
i-H
0
rH
r-t
rH
CM
rH
O
00
0
O)
o
N
O
J5
0
^
o
m
o
to
0
CO
CO
CO
o
in
o
CO
o
a
CO
o
CO
o
CO
o
CO
o
m
o
TT
O
CO
O
rH
O
*
CO
(O
c
(U
J^*"
o
CO
o
in
0
0
T»<
0
CO
0
CM
o
CO
0
co
0
m
o
rf
o
CO
o
CM
O
*
m
CO
fl3
e
a>
^
0
rH
rH
CO
00
O
00
O
oo
0
,_,
1-1
00
0
to
o
to
o
4-
tn
-------
25
d
0)
^
C
rH
P
C
O
tJ
«
fjQ
m
f$
nd Houston, 1970-'
CO
X
0
i»
^
cu
CD
O
rl
rj
C
lflj
c
CO
3
p
H
O
rH CD
0 O
CM CM
TJ< in
CO CO
O CM
rH CO
CO CO
m CD
m CD
rH rH
CD CO
rH rH
rH rH
co m
rH rH
CD 00
coin
o'oo
coco
00 CO
, f
Chicago 60 ]
61 ]
in co
rH CO
O O
m o
o oo
Tf CO
in rH
CD 00
rH CM
CM CM
CD rH
iH CM
O iH
O 00
CM iH
O CD
oco
C0£
*s
CM £
s-
o
Q.
cr
-a
'£.
GO
CD
m
m
CD
CD
o
CM
CM
CM
CM
00
iH
CN
0
0
CO
s
CM
S
CD
rH
O
oo
CD
CM
O>
co
CM
CN
CD f-
rH CM
CM CO
00 CN
rH rH
r> o
SS
rH 0
rH 0
rH rH
lO CO
co co
"O
to
o
_l
00 00
CO CM
co o
CO N
, ,
CM CO
MI CN
00 T}<
t*« 00
O CM
CD m
rH CO
rH CO
CM CM
t- 0
r- o
rH (N
CM CD
rH CM
OJrH
rH rH
m CM
5S
CM
00
CD
in
CO
iH
CD
rH
CN
rH
CM
rH
00
rH
CM
rH
rH
rH
un
+-
CQ
0)
P
rH
CM CO
O CO
CO M1
CD 00
1 rH
i m
CO CM
CO ^
T CO
M1 CD
CO CO
rH rf' CM
CD CD 00
t~ ao o
CD 00 CM
rH O CD
O 00 CD
CN rH CN
CO CO CD
00 00 CD
rH rH CM
CO CD CD
O O ^
N CM CN
t> CM 00
CD t~- rH
rH rH CN
CO CD rH
CD 00 Tji
rH rH CM
CD rH i 1
oo CD m
rH rH CM
m CD CD
M* CN CD
*
co rf in
00 00 00
o
1
3
0)
CD
i.
C3
CO f
O O)
I-H d
O N t-
O 00 CD
rH O O
£* ^^ C^
lO CO CD
CD CO Tj*
tO C^J O
00 C(OrHO
CMrHCOTjCQOOCDoi
SCD O CM rH 00 in CD
co CM ir r CD oo CD
rHOrHrHrHOOO
t>rHt>CMrHCn^l
-------
26
d
0)
4J
O
U
CO
3
pa
IS
O
be h
c -H
H CO
M
u
H
ja
p
u fi
a o
c a
<0 -H
P
c
co
iI 4N H r-l f-l 4N r-4 1-4
O Wi-tOOl-tOi-l
Sr-i oo o in co TJ< to
oo co o o eft 01 c^
^ .......
l-l rH |H rH rH r-4 O ^
(o to m cn oo
-------
27
Table 4
Yearly Means and Confidence Intervals of Air Pb Concentrations For Each Station
Area
Station Geometric Mean 95% Confidence Limits
(ug/nP)
Cincinnati
Dec 1968 -
Nov 1969
Philadelphia Suburban
Dec 1968 -
Nov 1969
Philadelphia Urban
Dec 1968 -
Nov 1969
Okeana
Okeana
Ardmore
Wynnewood
Los Anqeles
Dec 1968 -
Nov 1969
Rittenhouse Sq.
Pasadena
Pasadena
Los Alamos
Dec 1968 -
Nov 1969
Washington, D.C.
Oct 1970 -
Sept 1971
New York Suburban
Apr 1970 -
Mar 1971
Wood ley Park ~l
Cleveland Park]
Port Washington
Port Washington
30
31
32
33
34*
35*
19*
20*
11
12
13
14
15
16
17
18*
1
2
3
4*
5*
6
7
41*
42*
70
71
72
73
74*
75
76
77
88*
89*
1.92
1.45
2.13
0.85
0.31
0.33
1.28
1.01
1.89
1 .74
3.75
2.18
1.39
1.09
1.08
1.67
4.03
4.34
3.46
3.72
06
36
39
55
0.14
0.20
1.10
1.64 (7 mo.)
2.21
1.78
1.19
1.13
2.34
1.08
1.14
1.10
(1.63,
(1.16,
(1.75,
(0.71,
(0.24,
(1.63,
(1.52,
(3.30,
1.90.
1.19,
(0.90,
(0.92,
(0.95,
(1.36,
(1.91,
(1.53,
(1.07,
(0.99,
(2.05,
2.27)
1.83)
2.60)
1.01)
0.41)
(0.20, 0.40)
(1.06, 1.55)
(0.86, 1.19)
20)
00)
26)
52)
65)
33)
27)
(1.42, 1.93)
(3.40,
(3.58,
(2.96,
(3.20,
(2.62,
(1.92,
(2.62,
4.76)
5.26)
4.05)
4.32)
3.57)
2.90)
4.39)
(3.71, 5.60)
(0.11 t 0.18)
(0.16, 0.25)
1.27)
2.00)
2.55)
2.06)
1.32)
1.28)
2.67)
(0.94, 1.24)
(1.03, 1.27)
(0.98, 1.24)
* blood sampled
-------
28
Table 4 (Continued)
Area Station Geometric Mean 95% Confidence Limits
(uq/mj)(ug/m3)
New York Urban 81 1.76 (1.44, 2.15)
Apr 1970 - 82 2.04 (1.74, 2.38)
Mar 1971 83 1.73 (1.57, 1.90
84 1.71 (1.54, 1.89
Greenwich Vil. 85* 2.08 (1.79, 2.43)
May 1970 86 1.34 (1.13, 1.58)
Apr 1971 87 1.22 (1.06, 1.39)
Chicago Urban 60 1.30 (1.17,1.45)
Mar 1970 - 61 1.35 (1.23, 1.49)
Feb 1971 62 1.53 (1.33, 1.76)
Bridgeport 63* 1.76 (1.60,1.93)
64 1.83 (1.67, 2.02)
65 1.57 (1.39, 1.78)
67 1.55 (1.37, 1.76)
68 1.87 (1.51, 2.31)
Chicago Suburban Lombard 66* 1.18 (7 mo.) (0.83, 1.68)
March -
Sept. 1970
Houston 50 1.15 (0.92, 1.43)
Mar 1970 - 51 1.02 (0.86, 1.21)
1971 52 2.13 (1.83, 2.47)
53 2.26 (1.87, 2.73)
54 1.32 (1.16, 1.49)
55 0.87 (0.73, 1.03)
NW Houston 56* 0.85 (0.75, 0.96)
* blood sampled
-------
in
0)
3
LU to
«* rl
R -°
H? E
r^ E
uu q,
i o
r- (D
^B
o §
81
i
i§
O 01
oo 5
H CO
!H E
-P X.
CO CM N CM
m o o to to CM cn
rH t^~ in ^ CO 00 00 I I I
CM rH "J' CM 00 tD r>
rH rH rH
tO (f> If CM Tj< f~ 00 1 1 1
in rH co CM in Tf ^*
N to CM in CM to i i i
rH rH rH
1 1 1
* * * *
* * * *
to >-H in CM to in to
CO rH CM CM CM 1 1 1
+ 1 + + + + +
o^^*inooo^CiOOh*oorH
OOt^NrHCOOOtOCMO
CJ> in to co o Is- to
CO N CM rH rH 00 00
rH rH CO CM rH O O
* * *
rH W CO TP lO tO t^ 00 d ^5
a
H
X!
P.
rH
0)
a
CO
rH
H
£
O 'J* N O N O
00 to 00 00 O rH
oo TC m to to m
to t** t** m ^ ^j*
rf rH CM in
m to N oo i i
co to <* CM
m to t- to i i
CM
i i
* *
* *
co co in CM
rH CO CM II
+ + + 1
CM m co m I-H co
Ol TP rH 00 CO CO
o en o t>
t^ o t~ oo
ii
rH rH i-H O
* *
o rH CM oo ^ m
H
P
03
C
H
O
a
H
o
m
P
a en
a>
i-H -P
a c
E a)
in u
en -H
=H
o &
i-H -H
CQ CO
* *
*
-------
30
attributed to the lead assay technique. Air flow measurements have been based
upon identical procedures during the two test periods. Insofar as can be deter-
mined, the sampling and assay procedures during the two periods are directly
comparable.
One must consider the possibility that the atmospheric lead levels reported
in 1961-62 were inaccurate and too low, a possibility which would tend to show
that current levels erroneously reflect a non-existent increase. Evidence for
this might rest in the fact that in 1961-62 a number of Los Angeles samples had
to be excluded from the site averages. If the excluded samples represented
periods of relatively high atmospheric lead levels, the apparent monthly or
annual mean lead concentrations at the affected sites would have been too low.
Most of the excluded samples had been collected during the photochemical! smog
season, from June through November. Organic constituents in the air altered
the porosity of the filter media in a way which reduced air flow and made
volume measurements unreliable. There were times in 1961-62 when only one or
two out of eight stations were providing usable data. Special tandem pump units
and a reduction in smog levels in 1968-69 eliminated this problem of filter
obstruction.
Most of the high-reading samples in 1961-62 in Los Angeles occurred during
periods of surface inversions associated with low temperatures rather than during
the smog season. The number of lost samples in Los Angeles in 1961-62 approxi-
mated 3% of the total filters exposed. While it is possible that the operational
situation in Los Angeles in 1961-62 resulted in loss of certain high-reading
samples, it is quite unlikely that the losses were of sufficient magnitude to
result in the apparent change which is observed in comparison with 1968-69 sam-
ples. The problem of filter obstruction never constituted an important problem
in other cities.
An examination of meteorological conditions is relevant to the comparison
of data collected during the two test periods. Since no meteorological instru-
ments were located at the actual sampling sites, inferences with respect to
climatological conditions must be drawn from the data collected by the National
Oceanic and Atmospheric Administration (NOAA) at its reference location in each
city. These data are valid for precipitation and heating requirements; the
data are somewhat less satisfactory for wind velocity and direction but have
general validity for the area. That is, wind conditions at the Los Angeles In-
ternational Airport do not invariably reflect the situation in Pasadena. How-
ever, over long periods, such as those used in these studies, local deviations
between sampling sites and the NOAA reference station would not be expected to
differ to an important degree from year to year. An examination of climatologi-
cal data during the sampling periods and over a 30-year period does not suggest
that either sampling period is non-representative.
That air levels of lead measured in 1961-62 and 68-69 are sufficient to
demonstrate a trend at the sampling sites (based upon two points in time) has
been questioned. Especially in Los Angeles, the National Air Sampling Network
lead data at the reference station show wide annual deviations, upward and down-
ward, over the course of more than ten years (Figure 16). While the results
of the study reported here show relatively higher levels of lead in 1968-69
than in 1961-62 at 2 of the 4 Cincinnati sites, at 2 of the 7 Philadelphia sites,
and at all of the 8 Los Anqel^ sites, the NASN data cannot be ignored and de-
serve further consideration
-------
31
4
0
Los Angeles
' Cincinnati'
60 62 64 66 68 70 72
ANNUAL MEAN
Figure 16. Lead-in-Air Data (1959-71) National Air Sampling Network^)
(Symbols Indicate 1961-62 and 1968-69 population and Air Lead Level Studies.)
Note: The Los Angeles NASN site is Station # 1 in this study.
The Philadelphia NASN site is Station # 14 in this study.
The Cincinnati NASN site is Station # 30 in this study.
It is relevant that the NASN system was designed for a less specific purpose, that
different collection media and analytical techniques are used, and that the annual
mean at each site is based on approximately 26 samples.
Information on atmospheric levels of lead at sites newly established m 1968-71
serves to confirm previous observations which associate higher levels with urban pop-
ulation centers and lower levels with areas of low population densities. Data de-
rived from Okeana (Stations #34,35), Los Alamos (Stations #41,42), and Alpine County
reflect an almost "background" situation with respect to atmospheric lead levels.
Month-to-month variations in atmospheric lead levels were generally observed.
For the most part highest levels occurred during July-October while the lowest levels
occurred in February-April. The variations were most apparent at the Los Angeles
sites, where the highest lead levels tended to occur in October-December and the
lowest in April-July. Examples of the monthly variations are presented in Figure 17.
Blood Lead Concentrations
Histograms showing the distributions of blood lead concentrations in populations
meeting specific epidemiological criteria and living in areas of dissimilar character
and atmospheric lead levels are presented in Figure 18. Corresponding frequency
distributions are given in Table 6. Mean blood lead values and their 95% confidence
limits are presented in Table 7. For epidemiological purposes subjects were excluded
from tables and statistical consideration when they:
-------
32
1) were under 20 or more than 79 years of age;
2) when there appeared to be significant occupational or avocationa'l
exposure to lead;
3) when there was a significant amount of wild game in the diet
The inclusion of these subjects in the analyses does not change significantly
the computed data means for the several population groups. Appendix 3 presents
the blood lead concentrations with no exclusions from the study population.
The age distribution was not consistent for subjects at each area under
consideration (Appendix 4). Consequently an age effect might have been apparent
in the mean blood lead levels and in their distributions in the various groups.
The relationship between age and blood lead levels was examined by means of
regression analyses (Table 8). These were carried out for each location separ-
ately since patterns of mean levels differed among locations (Figure 19). When
a third order model was assumed, the regression of lead on age was shown to
be not significant in each location at an overall 5% level. Therefore no ad-
justment for age has been made in the analyses of blood lead data presented in
this report. Mean blood lead levels by age for each location, given in Figure
19, show that there is considerable overlapping of the confidence intervals,
supporting the conclusions of non-significant regressions of lead on age.
Since the distribution of persons by cigarette smoking status was not consis-
tent in each location (Appendix 5), and since cigarette smoking was believed to have
a possible influence on blood lead levels, tests of the differences in mean blood
lead values among locations were carried out after partitioning the sample into three
cigarette smoking categories (Tables 9,11,12). The F-ratio testing the differences
between the 11 locations is 31,47, which is significant at the 1% level.
Table 9 also shows the influence of smoking upon blood lead levels to be clear-
ly significant (F = 8.27). The blood lead level of smokers exceeded that of non-
smokers and previous smokers (Tables 11,12). The relationship held true for males
as well as females in the husband-wife pairs in Los Alamos (Table 13).
Since the interaction term in Table 9 is not significant (F = 0.81), one mav
conclude that the significant differences among mean blood lead levels in the
various locations exist in each smoking category. That is, the average blood
lead values of smokers is significantly different among these eleven locations;
the same holds true for previous smokers and non-smokers.
Urban-Suburban Blood Lead Levels
The investigation of whether mean blood lead levels reflect degree of ur-
banization was carried out by applying a binomial test to the data from three
metropolitan areas where an urban and suburban population had been sampled. In
each location a sample of smokers and non-smokers was used for this test, giving
a total of six comparisons between urban and suburban blood lead levels (Table lOa).
In both smoking categories the Philadelphia urban groups ("Rittenhouse") were
higher than the suburban ("Ardmore"); the Chicago urban groups ("Bridgeport") were
higher that the suburban ("Lombard"); and in New York the urban groups ("Green-
wich Village") were higher than the suburban ("Port Washington"). The proba-
bility of obtaining 6 relationships in this direction (urban > suburban) out of 6
comparisons is approximately equal to 1/2^ = 1/64 = 0.015, if one assumes either
direction is to be equally likely. Consequently it is concluded that urban mean
Text continued on page 49
-------
n
bj) Q)
3-P
-H
w
Wl /-N
rH .d W
N a) o aj
r-l > CO rH
0) W 0)
(DM U)
^ h C
3 T3 O <
H qj oi
O C
H O
0)
Q.^1
O
E
-p
c
o
o
t-
Oi
00
CD
Ol
I
r-f
CD
-------
34
in
O 00 CD
I
O
00
<£>
Cl
o
0)
3
c
o
o
3
hO
H
fcl
a
a
as
-------
35
O
CM
O
hw
O 00 CO
CNJ
00 CO
d 6
i
O
OS
CO
oo
<£
OS
o
d>
c
H
-P
ti
O
o
0)
T3
CO
CD
CM
CD
I
((
CD
O CO CO
-------
36
eo
o
I-CO
o oo co
CO CD
d CD
Q
N
I
O
CD
I
00
c
0)
c
o
o
5
CO
§
rl
U
H
O
O CO CO <* CJ
CO
C\J
,- 00 CO
d> o
* co
ci ci
L°
-co
p
c
,0
O CO CO t
CO
eo
O CO CD
CM
to
I
O)
II
,m/Slf qd
-------
m
in
°
O 00 CD
00 CO
0 O
37
3
C
H
O
O
c
O
+-!
CQ
3
a
O 00 CD
X!
-p
-------
38
3
C
H
4->
C
O
O
N
r-t
O
bfi
eo
u
H
(0
O 00 CD
O
(O
o oo CD
z
-o
-co
(O
i
I
O
I
QO
CD
O 00 CD
TI 1 1
00 CD
O O
CO
<0
_ ^
<
O CO CD
00 CD
C) CD
-------
39
(O
(0
o
0)
3
C
d
o
o
£
3
- Z
O
CO
O 00 CD
C\J
U
H
I
I
o
I
oo
CO
O 00 CD
-------
40
O 00 CD
O 00 CD
Q
T3
0)
a
O
O
C
o
-!->
M
C
H
O 00 CO
(O
O 00 CD
o
-to
00 CD
d a
I
O
ft) 5t
^ <>
a
H
-o
O 00 CD
CO
r^.
>
i 1 1 r i
0 00 CD *
*~
C\J i- 00
C)
LL
- 3
i \
CD O 00
d> ^
CD
O 00 CD
Q
I
oo
<£>
ff)
-------
41
CO
00
O CO (O
T- op CO
o 6
o
co
0)
3
C
H
P
C
O
o
(1)
o
>
(D
SB
o oo co
i
h-
I
o
CD
I
CD
O CO CD
-------
42
T3
0>
3
0
f*
O>
00
o oo co
I
o
I
oo
'-D
CD
00
CO
O 00 (D
c\i
-------
43
TJ
C3 >->
(b hn
3 o
0
TD r-H
O hi
oo .-i a
rH CQ O
0)
-------
44
TJ
(U
3
fi
H
P
C
o
o
oo
I
o
£
2
LU
LU
CC
0
1 I
O
CVJ
LU
O
<
_J
_l
>
1
0
Oo>
2 1
1
0
o
CO 1
rO
||
X
^
1 |
c
o»
O
O
\
o>
LU
_J
O
O
3
DO
o ° o o
AON3IT03UJ 3AU.V13U%
-------
45
0)
r-l
8
E-t
CU
CU
XI
O CU
O tJ
rH (0
CQ cn
o
c c
o
id
cu
id
to
CU
in
O
0)
-P
-P
rH
CU
iH
O
&
4
IU
id C
> -rl
1 I /"|
M U
cu a
-P tJ
C -rH
H S
-Q rH
di id
OH
rHCU
tjl-p
O(-l
t-l
r-5 H
S££££S2coSc£,HrHo'
o o o o o r-
iH
"a-
o o o o o «*
r~ r~- c r~- r^ co
VO VD 0 V£) VD CO
o o CM o CM in
o o rH ro co CM
rH
u
o
o oo co co criTr
r~~coo^ii'a ro v£> vc CMOO-slCMrH
rHiHCMCMCO^t^J'in A
OOOOrHrHrHrHiHrHiHHrH
co
co
m
ro VD
o o
CM O
m in
CM OJ
ro us
C X
-H (d
g S
-------
46
CO
rH
0)
> c
Q) O
r4 -P
X) C
PJ *H
(Q
£
(J M
rH rH CM r** C*)
O O O VO CM
O O rH H CM VO CO
rH CM
oo in
CM O
O O O O O CM O
rH
CM l> CM «* ^J1 VO
in in vo rH o r
O O >H CM CO CM rH
rH CM
0 0
m in
o o o o o r- r>-
rH
"Si
o
O OOOOOQ in r-
CM oo m ro in
^ vo o vo »* o o
CM i-H CM
«* vo in CM
en vo oo in
o ro r» o o o o
CM
o in o o in
O CM O O CM
o " rH o in o 1-1 o
CO rH rH
oo *3* *^ o\ OA in in
r^ o co vo o in in
A
oo ^r o vo CM oo
CM co ^r ^r in m
rH rH rH rH rH rH in
1 1 1 1 1 1
CM 00 *» O VO CN rH
CM CM CO ^'J1 in A
rH rH rH rH i-H rH
o
00
V0
O\
CM
o
»
CO
o
co
o
en
rH
CN
CTi
(N
CO
co
o
CO
C X
H fti
s a
-------
47
c
o
+J
10
3
O
a
J^
U
H
*J;
c
0)
(U
VH
U
C
0)
3
cr
M
0)
^
rH
p
rH
(U
rJP
x;
o
H
s
0)
0
o
rH
'O
t ,
H
0)
-p
c
H
JJ
E
Ti
H
m
a
£
o
(U
ft
rH
H
£>
4)
a
rH
a
o
H
S
vooomcM^oo^'eroooo
f
rH CM CM rH rH
COOO-'J'CM'^'iHOOOOOOOCM'^OO
ooi-HrHcorrinr~r-CMrHrHooo
CM CM rH i-H
rH VOrHnOOOOrOr-l>£)i-HrH
i^ cor-^J-incM^r^cor-r^
OOOOCMinrHcOCnrHOCMOOO
CM CM i-H i 1 rH
^^
tn
0
CU GO OO OO ^^ C^ P**1 LO LO 00 ^J* ^5* 0^ CF\ LO ^O
O^ V i~H i~H *~H <i rH f^l CN O4 CO rO fO
3 A
o o
o o
o c»
ro o
(M
o o
CD
i-H
rH
00
CM
rH
1
CM
CM
rH
CO
i-H
1
CO
CM
rH
0
rH
1
CO
rH
S
rH
1
0
rH
CM
m
i-H
I
rH
00
rH
1
CM
in
rH
oo
in
-------
Table 7
Geometric Means and Confidence Intervals of
Blood Lead Levels Based on a
Single Determination from Each Subject
(Rank)
5
9
11
7
3
10
4
6
2
8
1
Area
Ok can a
Ardmoro
Rittenhouse
Pasadena
Los Alamos (~)
Washington, D.C.
Port Washington
Greenwich Village
Lombard
Bridgeport
Houston
N
162
150
136
193
191
219
198
140
208
147
191
Geometric M<
(mg/lOOg)
0.0157
0.0180
0.0205
0.0175
0.0149
0.0192
0.0153
0.0166
0.0139
0.0176
0.0125
(mg/lOOg)
(0.0150, 0.0165)
(0.0171, 0.0190)
(0.0196, 0.0215)
(0.0169, 0.0182)
(0.0144, 0.0155)
(0.0186, 0.0198)
(0.0148, 0.0159)
(0.0159, 0.0173)
(0.0135, 0.0144)
(0.0169. 0.0183)
(0.0120, 0.0130)
All Females
1935
0.0162
(0.0160, 0.0164)
Los Alamos (M)
80
0.0172
(0.0163, 0.0182]
-------
49
blood lead levels are significantly higher than those from related suburban
areas (at the 1.5% level), based on data from three paired locations in
Philadelphia, New York, and Chicago.
Tests of the differences in blood lead levels in urban and suburban pop-
ulations of smokers only and of non-smokers only for each location separately
(i.e., Philadelphia, New York, and Chicago) were carried out by calculating
F-ratios for the comparisons. The significance tests are made by comparing
these F-ratios.to the critical values specified by the appropriate statistical
test. On the basis of the judgment that the Tukey statistical test is most
appropriate to the situation under examination, it is inferred that a significant.
urban-suburban gradient for both smokers and non-smokers exists in the Philadelphia
and Chicago areas. The differential in the New York area was not shown to be clear-
ly significant.
The relationship between mean blood lead levels and degree of urbanization
was also investigated by testing the significance of F-ratios obtained when the
urban and suburban mean blood lead levels were compared in each of three loca-
tions. Based on the total sample (smokers, non-smokers, previous smokers), the
urban mean lead value is higher than the suburban in each location and all loca-
tions combined on the basis of the Tukey test (Table lOb).
Table 14 and Figure 20 present mean blood lead levels and corresponding mean
air lead levels at the sites where both were measured. The association between
these mean values was measured by means of the Pearson Product Moment Correlation
Coefficient and Kendall's Rank Correlation Coefficient. The values obtained were
0,412 and 0.354, respectively, for all data. Corresponding values for non-smokers
only were 0.400 and 0.345. These values were not significant at the 5% level.
It was therefore concluded that the data do not contradict the hypothesis of no
association between average blood and air lead concentrations under the total en-
vironmental circumstances which exist at the study locations.
The observations that urban levels of blood lead are significantly higher
than suburban levels, but that air concentration of lead are not clearly reflected
in blood lead levels generally, suggest that other variables are more important
than ambient air lead-levels in determining concentrations of lead in the blood.
The precise nature of the variables, which evidently differ in the several regions
studied, remains undefined. Presumably alimentary lead intake plays a significant
role; however this was not thoroughly evaluated in the short-term metabolic stuny.
Climatic factors may also be relevant.
In a given metropolitan area, urban-suburban comparisons tend to minimize
the influence of diet and climate. That airborne lead contributes to the rela-
tively higher blood lead concentrations in center-city populations would seem
to be the most probable interpretation of this consistent observation.
Mention should be made of the possibility that persons who live primarily
indoors, as do the population groups in this study, are exposed to air levels of
lead which differ from those which exist in the ambient outside neighborhood at-
mosphere. While this is very probably the case, it is likely that the ratio
between indoor and outdoor air lead concentrations is similar in different com-
munities. Consequently indoor exposures are directly related to lead levels
Text continued on page 60
-------
50
Table 8
Analyses of Functional Relationship Between Blood
Lead and Age for Non-Smokers
Okeana
N = 131
Ardmore
N = 99
Rittenhouse
N = 76
Pasadena
N = 134
Los Alamos (F)
N = 106
Washington, D C.
N = 105
Port Washington
N = 87
Greenwich Village
N = 56
Source of Variation
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
df
3
127
125
2
3
95
93
2
3
72
70
2
3
130
128
2
3
102
102
_
3
101
99
2
3
83
81
2
3
52
50
2
SS
0.1346
2.6019
2.5953
0.0066
0.1653
1.7854
1.7795
0.0059
0.0413
0.9678
0.9487
0.0191
0.1072
2.0087
1 . 9656
0.0431
0. 0907
1.2900
1.2900
0.0210
0.9962
0.9909
0.0053
0.0454
1.1190
1.0573
0.0617
0.0004
0.6132
0.6088
0.0044
MS
0.0448
0.0205
0.0210
0.0033
0.0551
0.0188
0.0190
0.0029
0.0138
0.0134
0.0140
0.0096
0.0357
0.0155
0.0150
0.0215
0.0302
0.0126
0.0126
*"
0.0070
0 0099
0.0100
0. 0027
0.0151
0.0135
0.0130
0.0309
0.0001
0.0118
0.0120
0.0022
F
2.19 (NS)
0.16
2.93 (NS)
0. 15
1.03 (NS)
0.69
2.30 (NS)
1.43
2.40 (NS)
0.71 (NS)
0.27
1.12 (NS)
2.38
0.01 (NS)
0.18
(NS) Setting a 5% significance level for testing the significance of regression on age
for all locations, we must use a 0.5% significance level in each individual loca-
tion. The critical valuefor this test is equal to 4.28. (df=3 ,*) .
-------
51
Lombard
N = 125
Bridgeport
N = 85
Houston
N = 123
Los Alamos (M)
N = 32
Source of Variation
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
Due to regression
About regression
Error
Lack of fit
df
3
121
119
2
3
81
79
2
3
119
117
2
3
28
28
"
SS
0.0029
1.3071
1.2793
0.0278
0.0353
0.6980
0.6557
0.0423
0.0286
1.4867
1.4756
0.0111
0.0443
0.4348
0.4348
~
MS
0.0009
0.0108
0.0110
0.0139
0.0117
0.0086
0.0080
0.0211
0.0095
0.0125
0.0130
0.0056
0.0147
0.0155
0.0155
"~
F
0.08 (NS)
1.26
1.36 (NS)
2.64
0.76 (NS)
0.43
0.95 (NS)
(NS) Setting a 5% significance level for testing the significance of regression on age
for all locations, we must use a 0.5% significance level in each individual loca-
tion. The critical vaJue for this test is equal to 4.28. (df=3 ,^4 .
-------
§
25
20
15
25
20
10
25
20
15
10
52
Figure 19
OKEANA
RITTENHOUSE
LOS ALAMOS (F)
ARDMORE
LOS ANGELES
LOS ALAMOS (M)
I I I I I I
20- 30- 40- 50- 60- 70- 20- 30- 4O- 50- 60~ 70-
29 39 49 59 69 79 29 39 49 59 69 79
AGE (Years)
Mean Blood Lead Levels (ug/lOOg)
and 95% Confidence Limits by Age
-------
53
I
51
25
20
15
10
S 25
s1
^ 20
15
10
25
20
15
10
WASHINGTON, D.C.
I I I I I
GREENWICH VILLAGE
I I I I I I
BRIDGEPORT
1 1 1 1
PORT WASHINGTON
I I I I I I
1 i
HOUSTON
20- 30- 40-50- 60-70- 20- 30" 40-50-60-70-
29 39 49 59 69 79 29 39 49 59 69 79
AGE (Years)
Figure 19 (Continued)
-------
54
Table 9
Analysis of Variation Due to Smoking Status
(Females) and Location
Source df SS MS
Smoking
Location
Interaction 20 0.2120 0.0106 0.81
Error 1893 24.8060 0.0131
** Significant at the 1% level.
2
10
0.2167
4.1244
0.1083
0.4124
8.27**
31.47**
-------
55
Table lOa
Relationship Between Mean Blood Lead Levels
and Degree of Urbanization in Paired Locations
Smokers
Comparison
Rittenhouse vs Ardmore
(N=40) (N=29)
Greenwich Village vs Port Washington
(N=47) (N=69)
Bridgeport vs Lombard
(N=54 (N=99)
Rittenhouse vs Ardmore
(N=77) (N=99)
Greenwich Village vs Port Washington
(N=56) (N=87)
Bridgeport vs Lombard
(N=85) (N=125)
F-ratio
6.65*
on 1.67
26.09*
Non-Smokers
8.29*
on 1.31
31.77*
Direction
Urban >
Urban >
Urban >
Urban >
Urban -
Urban ->
Suburban
Suburban
Suburban
Suburban
Suburban
Suburban
Table lOb
F-Ratios Testing the Difference Between Urban and Suburban
Lead Levels in the Total Sample (Including Previous Smokers)
Rittenhouse vs Ardmore
(N=133) (N=150)
Greenwich Village vs Port Washington
(N=139) (N=197)
Bridgeport vs Lombard
(N=146) (N=207)
Total: Urban vs Suburban
(based on Philadelphia, New York, and Chicago)
*Significant at the 5% level (Tukey Test)
Note:
F-ratio
18.18*
5.98*
65.43*
71.56*
Urban
Urban
Urban
Suburban
Suburban
Suburban
Urban > Suburban
Critical Values (5%) for Judging Significance
of Comparisons in Tables lOa and lOb
"Standard" F
3.84
Tukey Critical Value
4.55
Scheffe Critical Value
18.30
-------
56
Table 11
GEOMETRIC MEANS OF BLOOD LEAD LEVELS
BY LOCATION AND SMOKING STATUS
(mg/lOOg)
Non-Smokers Previous Smokers
Okeana
Ardmore
Rittenhouse
Pasadena
Los Alamos (F)
Washington D. C.
Port Washington
Greenwich Village
Lombard
Bridgeport
Houston
All Females
Los Alamos (M)
0.0156
0.0180
0.0202
0.0172
0.0144
0.0181
0.0149
0.0158
0.0136
0.0167
0.0126
0.0158
(1127)
0.0165
Table 12
0.0139
0.0177
0.0200
0.0179
0.0154
0.0205
0.0150
0.0167
0.0146
0.0166
0.0120
0.0162
(304)
0.0168
Smokers
0.0171
0.0181
0.0214
0.0183
0.0160
0.0201
0.0162
0.0173
0.0146
0.0191
0.0124
0.0173
(495)
0.0193
Further Analyses of Relationship Between
Blood Lead and Smoking Status
Comparison
F
Direction
Smokers vs. Non-Smokers 40.01** Smokers > Non-Smo"kers
Smokers vs. Prev. Smokers 10.37** Smokers > Prev. Smokers
Prev. Smokers vs. Non-Smokers 2.57 Prev. Smokers > Non-Smokers
** Significant at the 1% level.
-------
57
Los Alamos (F)
Los Alamos (M)
Table 13
Geometric Means of Blood Lead Levels
for Los Alamos Males and Females by
Smoking Status
Non-Smokers
0.0144
0.0165
Previous Smokers Smokers
0.0154 0.0160
0.0168 0.0193
All
0.0149
(191)
0.0172
(80)
Analysis of Variation due to Sex (Los Alamos data)
Source df SS MS F
Sex
Smoking Status
Interaction
Error
1
2
2
263
0.1822
0.1093
0.0167
2.9402
0.1822
0.0546
0.0084
0.0112
16.30**
4.89**
0.75
Correlation Coefficients between Los Alamos Couples
Smoking Status
Same
Differing
No. Correlation Coefficient
Couples (95% Confidence Interval)
36 0.353*
(0.027, 0.610)
43 0.265 (NS)
(-0.038, 0.581)
* Significant at 5% level.
**Significant at 1% level.
-------
58
Table 14
Air and Blood Lead Concentrations (1968-71)
(All Subjects)
Geometric Means
Air
Rank
1
2
3
4
5
6
7
8
9
10
11
Area
Los Alamos
Okeana
Houston
Port Washington
Ardmore
Lombard
Washington , D. C .
Ri ttenhouse
Bridgeport
Greenwich Village
Pasadena
0.
0.
0.
1.
1.
1.
1 .
1.
1.
2 _
3.
Air
Gug/m3)
17
32
85
13
15
J8 (7 mo. )
19
67
76
08
39
Blood
(mg/lOOg. )
0.0149 (F)
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0157
0125
0153
0180
0139
0192
0205
0176
0166
0175
Blood
Rank
3
5
1
4
9
2
10
11
8
6
7
Relationship Between Means of Air and Blood Lead (Females)
Based on Pearson Product Moment Correlation Coefficient (r)
and Kendall's Rank Correlation Coefficient (T)
r = 0.412 (NS)
T -= 0.354 (NS)
(F) Females
(NS) Not significantly different from 0 at the 5% level.
-------
59
Table 15
Air and Blood Lead Concentrations (1968-71)
(Non-Smokers Only)
Geometric Means
Air
Rank
1
2
3
4
5
6
7
8
9
10
11
Area
Los Alamos
Okeana
Houston
Port Washington
Ardmore
Lombard
Washington, D.C.
Rittenhouse
Bridgeport
Greenwich Village
Pasadena
Air
(ug/m3)
0.17
0.32
0.85
1.13
1. 15
1.18 (7 mo. )
1.19
1.67
1.76
2.08
3.39
Blood
(mg/lOOg. )
0.0144 (F)
0.0156
0.0126
0.0149
0.0180
0.0136
0.0181
0.0202
0.0167
0.0158
0.0172
Blood
Rank
3
5
1
4
9
2
10
11
7
6
8
Relationship Between Means of Air and Blood Lead (Females)
Based on Pearson Product Moment Correlation Coefficient (r)
and Kendall's Rank Correlation Coefficient (T)
r = 0.400 (NS)
T = 0.345 (NS)
(F) Females
(NS) Not significantly different from 0 at the 5% level.
-------
60
measured outdoors. It is not immediately apparent that certain groups spend
relatively greater amounts of time outdoors. If this point were significant,
presumably the Pasadena population would experience relatively greater outdoor
exposure. Accordingly we do not feel that a consideration of indoor-outdoor
time or air conditioning alters the principal inferences drawn from this study.
Figure 20
Blood Lead Levels and Corresponding
Mean Air Lead Levels
0.03
0.02
i
O.Gl
0
?if)
c:
c;
^
.o
C;
BLOOD LEAD LEVELS
AS A FUNCTION OF
AIR LEAD LEVELS
() - number of people
;
"S
1 J
-t 1 ~^
1.0
<£
!
2.0
3.0
- 1
4.0
LEAD IN AIR, jjg/m3
Male-Female Blood Lead Differentials
The examination of blood lead levels in husband-wife pairs in Los Alamos
revealed a significant difference between males and females (Table 13). The
basis for this difference remains to be established. We believe that it cannot
be attributed to occupational factors, for the work of each individual was char-
acterized and conducted under comprehensive and detailed industrial hygiene
supervision. It is generally true that men smoke more than do women; however,
the blood lead levels in males remained higher than in female smokers and non-
smokers .
Since men have generally higher hematocrit levels than do women and since
lead is transported in association with red blood cells, the possibility exists
that higher blood lead concentrations in men reflect a higher level of circu-
lating red cells. An associated question related to whether or not relatively
higher hematocrits in women of Los Alamos (elevation: 7,300 feet) might be re-
sponsible for their blood lead levels being higher than would be the case at sea
-------
61
level. The anticipated gradient in hematocrits, Los Alamos males > Los Alamos
females > Pasadena females, was observed (Table 16). The analysis of variance
testing the significance of the difference among the hematocrit values of these
three groups was significant at the 1% level (F = 80.26). The comparison be-
tween Pasadena and Los Alamos females also was significant at the 1% level,
(F = 11.54), the mean value of Los Alamos females being higher than that of
Pasadena females.
The functional relationship between blood lead and hematocrit level was
examined within each of the three groups separately (non-smokers only) to see
if an "adjustment" could be made in the blood lead values which would make it
possible to obtain better comparisons among the three groups (Table 16). In
Los Alamos, the regressions were not significant; therefore it was concluded
that no significant association between blood lead determinations and hemato-
crit values could be found among Los Alamos males or females. In Pasadena a
significant second-order regression was found which is difficult to interpret
as far as the effect on blood lead values is concerned. For this reason no
"adjustment" for hematocrit level was made on the blood lead values of this
area.
Since the mean blood lead value for Los Alamos is lower than that of Pasa-
dena and since there was no significant (linear) correlation between blood lead
and hematocrit level, the relative size of the mean blood lead levels in Pasadena
and Los Alamos(F). would not be effected by an "adjustment" due to the average
hematocrit levels of these two populations.
It is also worth mentioning that although both men ^and women of the studied
Los Alamos couples consumed diets presumably similar in composition, the men
consumed a generally greater quantity of food. It is therefore likely that the
alimentary exposure of men to lead may be greater than that of women.
Fecal Lead Levels
The analysis of fecal lead levels, (Table 17) conducted to appraise one of
these variables, viz., alimentary lead intake, did not demonstrate high or low
levels compensating for respiratory exposures so as to obscure effects of the
latter. That is, the failure of the Pasadena group to show a "high" blood lead
level was not due to a low alimentary lead intake. Conversely the failure to
observe "low" blood lead levels in Okeana or Los Alamos was not due to a com-
pensatory high food lead level. It is of interest that the Houston fecal level
of lead was relatively high while the blood lead level was relatively low. This
observation was not further investigated in this study.
In each area the alimentary level of lead appeared to be much lower than
the 300 jug/day commonly quoted as the mean lead intake. This higher number is
based upon studies of men, who very probably ingested larger amounts of food
than did the women in this study. Nevertheless, the somewhat lower food intakes
by women cannot account for the discrepancy. It is possible that current im-
proved food manufacturing practices are associated with reduced lead levels in
foods.
Text continued on page 64
-------
62
Table 16
Frequency Tables Showing Distribution of Hematocrit Values
% Relative Frequency
Hematocrit
Interval
<34
34-35
36-37
38-39
40-41
42-43
44-45
46-47
48-49
S50
Hematocrit
Location
Pasadena (F)
Los Alamos (F)
Los Alamos (M)
Pasadena
(F)
0
4.9
4.4
21.2
29.1
18.7
15.8
2.5
3.0
0.5
Mean Values and
Los Alamos
(F)
1.0
1.0
4.5
11.9
17.8
38.1
16.3
6.4
2.0
1.0
Los Alamos
(M)
0
0
0
2.0
1.0
17.3
34.7
24.5
15.3
5.1
Their 95% Conference Limits
N Arithmetic Mean
203
202
98
40.97
41.97
45.55
95% Confidence Limits
(40.54, 41
(41.55, 42
(45.03, 46
.40)
.39)
.07)
Analysis of Variance Testing Differences in Hematocrit Levels
Source df SS_ MS F
Groups 2 1418.62 709.31 80.26**
Error 500 4418.94 8.84
Further Comparison Between Hematocrit Levels
Comparison F_ Direction
Pasadena (F) 11.54** Los Alamos >Pasadena
vs
Los Alamos (F)
** Significant at the -1% level
-------
63
Table 17
Analyses of Functional Relationship Between
Blood Lead and Hematocrit for Non-Smokers
Pasadena (F)
N = 131
R - 0.26
Los
N
R
Los
N
R
Alamos
= 104
= 0.18
Alamos
= 32
= 0.44
(F)
(M)
*Significant at 5%
**Significant at 1%
Due to regression 3
Cubic 1
Quadratic 1
Linear 1
About regression 127
Due to regression 3
About regression 100
Due to regression 3
About regression 28
level
level
0.1393
0.0008
0.1381
0.0004
9346
1
0.0435
1.3027
0.0933
0.3858
0.0464
0.0008
0.1381
0.0004
0.0152
0.0145
0.0130
0.0311
0.0138
3.05*
0.05
9.08**
0.02
1.11 (NS)
2.25 (NS)
Table 18
Geometric Averages of Fecal Lead Ten Day Totals
Area
Okeana*
Ardmore
Los Angeles
Los Alamos (F)
Washington, D.C.
Port Washington
Lombard
Houston
20
20
20
20
20
20
20
20
Fecal Lead
(mg/10d.)'
0.852
139
147
0.973
1.168
0.967
0.880
1 .505
*Total for donor #0112 adjusted to account for two
unusually high values.
Analysis of Variation of Fecal Lead
(Ten Day Total) Due to Location
Source df SS^ MS_ £
Location 7 0.9133 0.1304 2.00 (P = 0.08)
Error 152 9.9089 0.0651
-------
64
Results of Analysis of Pennypack Data
In April 1961 and'April 1971 blood samples from 23 persons in Pennypack
were obtained. (Each person was sampled twice: once in 1961 and once in 1971).
The geometric mean of 1961 blood lead values is 0.0124 mg/lOOg; the geometric
mean of 1971 values is 0.0138. The ratio of 1971 to 1961 mean values is 0.0138/
0.0124 = 111/7.. The result of a paired T-test performed to test the significance
of the difference between mean blood lead values in 1961 and 1971 yielded T =
1.764, df = 22. This value would occur by chance with probability P = 0.09
if in fact the mean lead values in 1961 and 1971 were equal. It was therefore
concluded that there may be a real difference between them; more evidence is
needed to make definite conclusions. Interpretations must take into account
the fact that the sampled participants were ten years older in the latter study
than they were in the former.
Results of Analysis of Alpine County Data
In 1960 and 1971 blood samples from two independent samples of males and
females from Alpine County, California were measured and recorded.(2,6) yne
geometric means of each group are as follows:
Mai es Fema 1 es Sex Unknown l£ifll
1960 0.012 0.009 0.013 0.011
(16) (11) (10) (37)
1971 0.0201 0.0159 - 0.0177
(18) (21) (39)
The standard errors of the 1971 mean values are 0.0205 for males, and
0.0223 for females. No data on the variation of 1960 values are available.
To determine if there is a difference in the 1971 blood lead values of
males and females in Alpine County an unpaired T-test was performed. The re-
sultant T-value was T = 3.393, df = 37. This value would occur by chance wit'
probability P = 0.002 if in fact the mean lead values of males and females wp'-r
equal. It was therefore concluded that the difference between them is high1/
significant, that of males being higher. The ratio of 1971 male to female me*
values is 0.0201/0.0159 = 126.42%.
Note: The F-statistic testing the equality of variances yielded F = i.397
(df = 20, 17), P = 0.245.
To determine if there is a statistically significant difference between
the average 1960 and 1971 blood lead values, 95% confidence intervals about the
1971 mean values were found. These are, for males and females, respectively,
(0.0182, 0.0222) and (0.0143, 0.0176). Since the 1960 mean values, 0.012 and
0.009 respectively, do not lie within these intervals, we conclude that the
1960 and 1971 mean blood lead values are significantly different at the 5%
level; for both males and females, the 1971 mean values are higher than those
in 1960.
-------
65
Continuous air sampling over a 6-month period at a representative site in
Alpine County permitted the mean atmospheric lead level to be determined:
0.03 jLiq/m^. There is no reason to believe that a higher or lower level pre-
vailed in 1960.
In each of these 2 situations the higher 1971 blood lead determinations may
reflect an actual increase or an apparent increase due to differences in method-
ology used in the two study periods. Standards are not available which would
permit a longitudinal comparison of methods. For the Alpine group, the differ-
ence is very likely a reflection of method, for there has been no apparent change
in the region which would account for the observation. Accordingly the 1960
levels were too low, or the 1971 levels were too high. Insofar as we can deter-
mine, from interlaboratory comparisons and from an inspection of data generated
in association with the original Alpine County survey" and in this study, the
1971 values are of relatively greater validity. There is no evidence that the
regional sources of food have changed during the interval between the studies.
The possibility that cultural and dietary habits of the 2 sampled populations
are dissimilar cannot be excluded from consideration.
Results of Analysis of Policemen Data
In March 1962 and March 1971 blood samples from 45 policemen in Cincinnati
were obtained. (Each policeman was sampled twice: once in 1962 and once in
1971.) Blood lead values were measured and recorded. The geometric mean of
1962 blood lead values is 0.0235 mg/lOOg; the geometric mean of 1971 values is
0.0179 mg/lOOg. The ratio of 1971 to 1962 mean values is 0.0179/0.0235 - 76%.
In order to determine whether the average blood lead value in 1971 is signifi-
cantly different from that in 1962, a paired T-test was performed. The re-
sultant T-value was T = 8.579, df = 44. This value would occur by chance with
probability P = 0.00 (to two decimal places) if in fact the mean lead values
in 1962 and 1971 were equal. It was therefore concluded that the difference
between them is highly significant. Here again an interpretation requires con-
sideration of the fact that the policemen were older in 1971 than in 1962. The
job assignments of a number of policemen had changed over the years; si e men
had changed from traffic duty to assignments in police cars.
The significance of data derived from the studies of Cincinnati police,
Pennypack residents, and Alpine County residents is unclear. There are no direct
standards which would permit comparison of 1961-62 and the later blood lead as-
says. Any interpretation of the observed differentials must take into consid-
eration the fact that the determination of lead concentrations in biological
materials is a difficult analysis and that experienced laboratories report vari-
able results when presumably identical samples are assayed at different times
or under different circumstances. Accordingly, longitudinal comparisons of
blood lead levels may contain random and systematic errors which complicate in-
terpretations. While it would seem desirable to run daily standards, it is an
unfortunate fact that blood "standards" are often of dubious uniformity and
that standards which are introduced into only the final step of the determina-
tion fail to reflect the ashing or sample preparation steps.
Experience suggests that day-to-day systematic variability may reduce the
validity of comparison studies when samples from different groups are assayed
at different times. Consequently it would seem appropriate to intermix samples
-------
66
from groups to be compared so that samples from all groups are assayed under
the same conditions. Such a procedure would dictate that all samples be col-
lected prior to their intermixing and assay. The impracticality of this approach
was apparent in this study. Insofar as possible, however, urban and suburban
samples from the same metropolitan area were intermixed and assayed simultane-
ously.
-------
67
REFERENCES
1. Public health aspects of increasing tetraethyl lead content in motor fuel,
U.S. Dept. Health, Education, & Welfare, PHS Pub. No. 712, Washington, 1959.
2. Survey of lead in the atmosphere of three urban communities, U.S. Dept. Health,
Education, & Welfare, PHS Pub. No. 999-AP-12, Washington, 1965.
3. Pierce, J.O., and Meyer, J.H., Technical note. Sampling and analysis consid-
erations in evaluating levels of atmospheric lead, Atmos. Environ. 5:811,
1971.
4. Bambach, K. and Burkey, R., Microdetermination of lead by dithizone, Ind.
Eng. Chem. Anal., 14:904, 42.
5. Shearer, S.D., Akland, G.G., Fair, D.H., McMullen, T.B. and Tabor, E.G.,
Concentrations of particulate lead in the ambient air of the United States,
Division of Atmospheric Surveillance, Environmental Protection Agency, Re-
search Triangle Park, 1972.
6. Lead in the environment and its effect on humans, Department of Public Health,
State of California, Berkeley, 1967.
-------
68
APPENDIX 1
Project Group - University of Cincinnati
Analytical Procedures
Jacob Cholak, principal
Roland E. Burkey
Bernard G. Meiners
James 0. Pierce
Bernard E. Saltzman
Lawrence J. Schafer
David W. Yeager
Atmospheric Studies
Lloyd B. Tepper, principal
Jacob Cholak
Bernard G. Meiners
Jozef Svetlik
Data Management and Statistical Evaluation
Joseph H. Meyer, principal (1968-69)
Linda S. Levin, principal (1970-72)
C. Ralph Buncher, consultant
Katherine Hayes
Kathleen Poskus
Lawrence J. Schafer
Population Studies
Lloyd B. Tepper, principal
Eva C. Petering
Subcommittee for the Surveillance of Air and Population Lead Levels
Robert J.M. Horton, Chairman, Office of Air Programs, Environ-
mental Protection Agency
Vincent J. Castrop (General Motors Corporation), Automobile
Manufacturers Association
Jerome F. Cole, International Lead Zinc Research Organization
(from 1969)
Robert E. Eckardt (Esso Research and Engineering Company),
American Petroleum Institute (through 1969)
Don G. Fowler, International Lead Zinc Research Organization
(through 1968)
Harold H. Golz, American Petroleum Institute (from 1970)
Howard E. Hesselberg, Ethyl Corporation
Alden J. Pahnke, E.I. du Pont de Nemours and Company
Elbert C. Tabor, Office of Air Programs, Environmental Pro-
tection Agency
-------
69
APPENDIX 2
Statistical Notes
1) Blood lead concentrations (mg/100 g) were measured to three
decimal places. Tables in this study show blood lead measure-
ments to four decimal places. This has been done to prevent
rounding to the third decimal place and hence losing accuracy
in comparing results.
2) In the analyses of this report, the data are assumed to be
normally distributed. Confidence limits and F-tests obtained
from analysis of variance techniques depend to varyina dearees
on the validity of this assumption. It was sugaested by pre-
vious studies that using the log transformation on blood lead
values successfully normalizes the data. To test this assump-
tion the Kolmogorov-Smirnov statistic was calculated for each
location separately. (Actually a Z-transfornation of this
statistic valid for large samples was used.) This statistic
was used to test the normality of the logarithm of the blood
lead values in each area. The mean values and standard de-
viations of the normal curves against which the observed data
were tested were obtained from the observed data in the usual
manner; that is:
X =
EX.
si
(X.-X)?
~ N-l --J
where X. = logarithm of blood lead.
The overall a level in these tests has been set at 0.20; each
individual a level, therefore, is equal to 0.01. The table
below shows resultant Z-values and their probabilities P(Z).
-------
70
Table of Z-Values Obtained From Kolmogorov-Smirnov
Tests of Normality of Log^o Lead Concentrations
Area Z_ P (Z) t
Okeana 1.055 0.22
Ardmore 1.142 0.15
Rittenhouse 0.695 0.72
Pasadena 0.948 0.33
Los Alamos (F) 1.613 0.01*
Washington, D.C. 1.384 0.04
Port Washington 1.180 0.12
Greenwich Village 1.244 0.09
Lombard 1.425 0.03
Bridgeport 1.290 0.07
Houston 1.282 0.07
Los Alamos (M) 0.696 0.72
The conclusion from these tests of normality is that there
is one location where the distribution of the logarithm of blood
leads differs significantly from normal; however, since this test
is conservative, one must question normality in those locations
where the Z values have probabilities close to 0.01, namely,
Lombard and Washington, D.C. In spite of this, there is enough
justification in the table above to assume log normality of the
blood lead data, and for this reason all tables of mean values
present the geometric mean rather than the arithmetic mean, the
geometric mean being the antilog of the arithmetic mean of the
logs. In addition, all the analyses of variance are done using
log values.
*Significant at 1% level.
tSince two parameters are estimated from the data, it is suggested
by Siegel, Non-Parametric Statistics, McGraw Hill, 1956, P. 60,
that this test is a conservative one.
-------
71
APPENDIX 3
(Rank)
Area
5 Okeana
9 Ardmore
11 Rittenhouse
7 Pasadena
3 Los Alamos (F)
10 Washington, B.C.
4 Port Washington
6 Greenwich Village
2 Lombard
8 Bridgeport
1 Houston
All Females
Los Alamos (M)
and Confidence Intervals of
'lOOg) Levels Based on a
nation from Each Subject
L Sample Before Exclusions)
N Geometric Mean 95% Confidence Limits
166
156
137
209
204
224
203
140*
208*
148
204
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0158
0180
0206
0176
0149
0192
0154
0166
0139
0176
0124
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
0150,
0171,
0194,
0168,
0144,
0186,
0149,
0159,
0135,
0169,
0120,
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0166)
0190)
0218)
0184)
0154)
0198)
0159)
0173)
0144)
0183)
0129)
1999
101
0.0162
0.0172
(0.0160, 0.0164)
(0.0164, 0.0180)
*No change in N.
-------
72
APPENDIX 4
0)
o
CO
CM
CO
lO
Ol rH
(0
o
CO
c
-------
73
APPENDIX 5
00 t» CM
CM T}<
0
CO
3
O
CM
CO
o
OJ
CM
o
a
CO
a
CO
o
H
.a
CJ
CO
CM
CO
CM
O rH
00
O
CM
r<
o
a
o
Q
~
a
o
P
b£>
a
H
A
(0
CO
e
a
H
w
0)
rH
CO
O
CK
oo m T)< ooi o
CM CM r-( i-l ^
1 r^
c
0
4^
M
C
H
J3
tn
ra
^:
P
t,
£
f-H
CO rH (^ CM CM irtl 00
CM CM CO f-H ^
1 ,_^
cj
P
fc
e
o
+j
bO
e
H
j-t
*XH
CQ
ca
=E
r-l t^ rH ^f W r-l Oi
CM CM CM CM i-l
CM
&
CM
I
O
CM
O5
CO
I
O
CO
cn
m
o
m
05
to
o
CO
o>
CO
P
£
-------
3 S
O> 2_
- -TO O 2
5 5~-c
10 TO a) i/> a.
3 3 CD O
<" CT °" S =
~ sSs
0.
o z "H. o>
3 D TO w
"O - 0) -,
^ -ou
fo =r -r n
3 TO mo
O.SH
af^i
,-t, I ro
o - m -
3 ;
CD
CD
-i _^ lu -,
° O ^ ~
T3 ° " tn
J ^ O _
CD _ TO
n &'~^
3 ? =
^ "> o
a>
z
<
-H 39
2 O
c
-2
-n
r~
O
-< o o 5
*
z S « =!
°!5i
M 3 J>
n
<
30
O
p 3
> >
CO i
1/5 u>
2 Si
00
o
rr>
2
n
------- |