&EPA
             United States
             Environmental Protection
             Agency
             Environmental Research
             Laboratory
             Duluth MN 55804
EPA 600/3 78 071
  ;378
            Research and Development
Polychlorinated
Biphenyls
in Precipitation
in the Lake Michigan
Basin

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology  Elimination  of  traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1.   Environmental Health Effects Research
      2.   Environmental Protection Technology
      3.   Ecological Research
      4   Environmental Monitoring
      5.   Socioeconomic Environmental Studies
      6.   Scientific  and Technical Assessment Reports (STAR)
      7.   Interagency Energy-Environment Research and Development
      8.   "Special" Reports
      9.   Miscellaneous Reports

This report has been  assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials. Problems are assessed for their long- and short-term influ-
ences. Investigations include formation,  transport, and pathway studies to deter-
mine the fate of pollutants and their effects. This work provides the technical basis
for setting standards to minimize undesirable changes in living organisms in the
aquatic,  terrestrial, and atmospheric  environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                             EPA-600/3-78-071
                                             July 1978
POLYCHLORINATED BIPHENYLS IN PRECIPITATION IN THE
               LAKE MICHIGAN BASIN
                        by

                 Thomas J. Murphy
               Charles P. Rzeszutko
                DePaul University
             Chicago, Illinois 60614
                 Grant No.  803915
                 Project Officer

                Michael  D.  Mull in
           Large Lakes Research Station
    Environmental  Research  Laboratory - Duluth
           Grosse lie, Michigan 48138
        ENVIRONMENTAL RESEARCH LABORATORY
        OFFICE OF RESEARCH AND DEVELOPMENT
       U.S.  ENVIRONMENTAL PROTECTION AGENCY
             DULUTH, MINNESOTA 55804

-------
                                 DISCLAIMER


    This report has been reviewed by the Office of Research  and Development,
Environmental Protection Agency and approved  for  publication.  Approval does
not signify that the contents necessarily reflect the  views  and policies of
the Environmental Protection Agency, nor does mention  of  trade names  or com-
mercial products constitute endorsement or recommendation for use.
                                      11

-------
                                  FOREWORD
    The presence of hazardous materials in the environment is a topic of
major interest to the Environmental Protection Agency.  Elevated levels of
polychlorinated biphenyls (PCBs) in some species of Great Lakes fish have
necessitated the recommendation that persons limit, or eliminate, their die-
tary intake of these fish.  In order to develop an understanding of the
pathways by which PCBs move within a lake ecosystem, it is necessary to
determine first the routes and amounts of these pollutants entering the
lake.

    This report addresses the input of PCBs from the atmosphere to Lake
Michigan.  The differentiation between gaseous and particulate forms are
measured as well as the comparison between urban and rural input rates.


                                     Michael D. Mullin, Ph.D.
                                     Project Officer
                                     Large Lakes Research Station
                                     Environmental Protection Agency-Duluth
                                     Grosse lie, Michigan
                                    in

-------
                                   ABSTRACT


       Event rainfall samples were collected  in  Chicago,  Illinois  and on
Beaver Island, Michigan and analyzed for polychlorinated  biphenyls  (PCBs).
The volume weighted mean concentrationRof  3b  samples of rain was  111 ng/L
(111 parts in 10  ) of PCBs as AroclorK 1242  + 1254 +  1260.  This  would
result in the deposition of 4800 kg/yr of  PCBs to Lake Michigan from
precipitation.

       The concentrations of PCBs in rainfall were found  to be as  high on
Beaver Island as in Chicago.  This, and the finding of PCBs in the  air and
precipitation in remote areas, is interpreted as  indicating that  the sources
of PCBs to the atmosphere are diffuse and  that residence  times in  the
atmosphere are long.

       A comparison of the results obtained from air and  precipitation
sampling, including simultaneous sampling, indicates that the majority of
PCBs are present in the atmosphere as vapor with  the remainder being present
on particulates.  This result, along with  the high solubility of  PCBs in
non-polar organics, raises doubts as to the validity of measurements for the
dry deposition of PCBs obtained with the use  of  collectors covered  with
mineral oil or other non-polar liquids.

       Two gas samples obtained from a vented sanitary landfill had PCB
concentrations in the jug/m  (parts in 10  ) range.  This indicates  that
PCB containing materials incorporated into landfills may  be an important
source of PCBs to the atmosphere.

       With the decline in the use and discharge of PCBs  by major  industrial
facilities in the basin, PCB levels in Lake biota may  decline by  a factor of
four or more.  Precipitation is now the major known source of these
materials to the Lake, and the future problems in the  Lake with PCBs will be
determined mainly by the magnitude of the  atmospheric  inputs of PCBs to the
Lake.

-------
                                   CONTENTS
Forward 	       m
Abstract	        iy
Acknowledgements  	        vi

Section:

1. Introduction 	        1

2. Sampling	        4

3. Analysis	       10
     Calculations 	       12

4. Results	       13

5. Discussion	       18
     Atmospheric Inputs 	       18
     Distribution of PCBs in the Atmosphere Between Vapor and
       Particulates 	       20
     PCBs in Organic Particulates	       22
     PCBs in Gases from Landfills	       22
     Perch!orination  	       22

Bibliography  	       24

Appendix	       28

-------
                               ACKNOWLEDGEMENTS
       The authors would like to express their thanks and appreciation to
Dr. Matthew C. Hohn, director of the Central Michigan University Biological
Station on Beaver Island for his conscientious help  in collecting the
precipitation samples on Beaver Island.

       We would like to acknowledge the help of Cunera Buys, Edwardine
Nodzenski and Guiseppe Paolucci for their  able assistance with some of the
analytical work on this project.

       We wish to thank the Illinois Geological Survey for the use of a
recording rain gage for this project; Mr.  George Bottoms and the Forest
Preserves of DuPage County for permitting  the collection of the landfill gas
samples at Roy C. Blackwell Forest Preserve; Mr. Thomas Ward and the City of
Chicago Department of Waters and Sewers for their  permission and
co-operation in collecting the water samples from  the 68th Street (Dunne)
water  intake crib; and the National Park Service at  Mammoth Cave National
Park for their permission to sample and the Cave Research Foundation for
their  co-operation in the sampling in the  National Park.

       We would like to acknowledge the partial support of this project by
DePaul University.
                                       VI

-------
                                  SECTION  1

                                  INTRODUCTION


       In recent years, fish species  in Lake Michigan,  especially  the
salmonids, have contained high levels of DDT and  its degradation products,
and polychlorinated biphenyls (PCBs).  In  most cases, the  large fish are
above the Food and Drug Administration (FDA) limit for  these materials  of  5
mg/L (Willford 1975).  Among the  Great Lakes, the levels of DDT and  its
degradation products in the late  1960's and then of PCBs in the 1970's, have
been the highest in Lake Michigan (IJC 1974; 1976).

       PCBs are a serious health  hazard to many species of animals and  to
man.  It has been shown that the  cause of  the reproductive failures  in  ranch
mink fed Lake Michigan fish during the middle 1960's, was  due to PCBs in the
fish (Aulerich et at. 1973).  Fish are now practically  the sole source  of
PCBs in the human diet (Jelinek and Corneliussen  1976).  Breast fed  infants
throughout the country (Savage et _al_. 1973; Musial et >aJL  1974; Grant et a!.
1976) are now exposed to 10% of the dosage of PCBs which has been  shown to
cause serious health problems in  rhesus monkeys (Allen  and Norback 1976;
Barsotti et al. 1976).  A correlation between the number of Lake Michigan
fish consumecTand the blood levels of PCBs in humans has been reported
(Humphreys 1976).  The public press has reported the case  of one fisherman
who developed the symptoms of PCB poisoning, Yusho disease, after  having
eaten large quantities of Lake Michigan fish (Ingersoll 1977).

       The cause of the DDT problem was probably the use of this insecticide
in the agricultural regions around the Lake, and for Dutch Elm disease
control in the urban areas at the south end of the Lake.   It is believed
that an important route for the dissemination of DDT and its metabolites is
atmospheric transport as vapor, mist  and material adsorbed on particulate
matter (Woodwell et _al_. 1971; 1972).  This material could  then enter the
Lake by both wet and dry deposition.

       When the use of DDT was halted in Wisconsin and  Michigan, and then
natiowide in the late 1960's, the levels of DDT and its metabolites  in  the
fish dropped rapidly over the next several years.  For  instance, the level
in bloaters (chubs, Coregonus hoyi) dropped 87 percent  from 1970 to  1975,.
and the level in Lake Trout (Salvelinus namaycush) dropped 53 percent during
the same time period.  This indicates a relatively short residence time in
the Lake for these materials.

       The problem with PCBs became apparent during the time of the  problems
with DDT.  Since many of the chemical and  physical properties of the PCBs

-------
are similar to those of DDT, there was  a  lot of  accumulated  experience with
this type of problem to draw on.  Thus  it was expected  that  the  voluntary
ban on "open" uses of PCBs  in  1971 by the sole U.  S. manufacturer, would
lead to a decline in PCB levels  in the  fish similar to  that  which had
occurred with DDT (Buckley  1972).  Unfortunately,  this  did not happen.  PCB
levels in bloaters in 1974  were  unchanged from those in 1972 and the levels
in lake trout rose about 50% during the same time  period  (Willford 1975).

       In 1975 then, it was not  known why the PCB  levels  in  the  fish in Lake
Michigan were continuing to increase.   Several possible explanations for
this increase are:

             1. The chemistry of  the PCBs  in the  Lake is different from that
            of DDT.

            2. The slower degradation of many of the PCB  isomers increases
            the residence time of the PCBs  in the  Lake.

             3. One of the "closed" use  of PCBs was the  major source of the
             PCBs  getting into  the Lake.

             4. The  large reservoir of PCBs  manufactured before  1971, and
             already  used or still in use, continue to enter  the  environment.

       Since  the  measurement of  PCBs in tributaries and waste water
treatment plants  had shown  that  only a  small amount of  PCBs  were entering
the  Lake by those routes, since  the  atmosphere had been a major  source of
DDT  to the  Lake,  and since  PCBs  had  been  identified in  precipitation in
other  areas (Tarrant and Tatton  1968; Sodergren  1972;  1973), it  was thought
that the atmosphere  could also be transporting PCBs to  the Lake.  This
assumes that the  fourth explanation  above is the cause  of the PCB problem  in
the  Lake.

       This study was  initiated  to  determine the amount of PCBs  which were
being  transported to the Lake  by the precipitation scavenging of
PCB-containing  particulate  matter from  the  atmosphere,  and to make  some
measurements  on the  concentration of PCBs to the atmosphere. Information  on
the  distribution  of  PCBs between the soluble and particulate fraction  in
precipiation was  obtained  and  the  compostion of  the PCBs with respect  to the
different  isomers present  in  the samples  was determined.  Also,  the PCB
concentration and composition  in gases  vented  from a  sanitary landfill were
measured.

       While this project was  not  directed  to  the  determination  of  sources
of PCBs to  the  atmosphere,  a  convenient opportunity to  measure  such a  source
exists  in the Chicago  area  and such  measurements were  made.

       A potential  source of  PCBs  to the  atmosphere is  material  incorporated
into sanitary landfills.   If  PCBs  in paints,  plastics  or carbon-less paper,
or  in  containers  such  as PCB  filled  capacitors which  have broken open,  are

-------
present in a landfill, they can evaporate.   Since  gases  are  generated  by the
anerobic decomposition of organics within  landfills,  and  continually escape
to the atmosphere, PCBs which evaporate  in the  landfill would  be  carried out
along with these gases.

       An above ground landfill has been constructed  in a suburb  of  Chicago
and is used for recreational purposes.  The  landfill  has  been  carefully
sealed with clay, and several gas wells have been  put  in  place  to  vent gases
generated within the landfill to the atmosphere.   These gases  have been
sampled by the Illinois Geological Survey  and,  on  a volume basis,  consist
mostly of methane (69%), and carbon dioxide  (29%).

-------
                                  SECTION 2

                                   SAMPLING


       To meet the objectives of this project, air  and  event  precipitation
samples were collected in Chicago, Illinois  at DePaul University  (41
55.4'N; 87° 39.9'W) from July 1975 through January  1977,  and  rain  samples
were collected on Beaver Island (45  40'N; 85  31'W), 500 km  from
Chicago, from May through October 1976.  The sampler  in Chicago was  located
about 15 meters above the ground on the roof of a building, 2 km west  of
Lake Michigan, 4 km north of the central business district of Chicago  in a
densely populated urban area which also contained some  light  manufacturing.
The sampler on Beaver Island (Figure 1), 500 km north of  Chicago,  was  one
meter above the ground and about 30 meters from the shore on  the east  side
of the Island on the grounds of the Biological Station  of Central  Michigan
University.

       Air samples were collected in Chicago at the above location,  and
within the boundaries of Mammoth Cave National Park (37°  13'N; 86° 03'W).
In addition, several water samples from Lake Michigan were collected and
some samples obtained of gases vented from a sanitary landfill.

       When possible, samples were filtered  upon  collection,  and thus
separate samples of PCBs associated with the non-filterable as well  as the
filterable portions were obtained.  The samples were  cleaned-up and  analyzed
for the PCBs present.  Since the commercial  PCB products, the Aroclors  ,
contain mixtures of different PCB compounds, the  weight of the individual
PCB compounds found in the samples was determined,  summed, and Reported as
the Aroclor  in which they predominately occurred.  The Aroclor
mixtures reported here are 1242, 1254 and  1260.   In no  case did the  relative
amounts of the different PCBRcompounds found in the samples approximate the
ratios found in the Aroclors  .

       Polychlorinated biphenyls (PCBs) are  liquids or  solids at room
temperature.  They are non-polar, slightly soluble  in water (56x*
-------
                Waukegan^
                  Chicago
                                     Mammoth Cave NP
FIGURE 1.  MAP OF  SAMPLING LOCATIONS.

-------
analysis.  Also,  in order to  collect  remote  samples,  a means  of
concentrating or  processing the  sample  at  the  collection  site,  with  a
minimum of effort on the part  of the  sampling  personnel,  was  necessary.

       The polyurethane foam  collection method for PCBs of Uthe ejt al_.
(1973) was chosen.  This method  involves  passing a liquid or  gas stream  over
pieces of polyurethane foam coated  with a  non-polar liquid.  The PCBs are
soluble in the  liquid and the  foam, and are  very efficiently  scrubbed from
the water or gas  stream by the foam.  They are then extracted from the foam
in a Soxhlet extractor.

       Precipitation was collected  on a galvanized steel  surface and the
water ran-off to  the sampler  described  below.   The surface of the collector
was cleaned before each precipitation event  to minimize the amount of dry
fallout incorporated into the  sample.  The design of  the  sampler used for
the rain collections on Beaver Island is  shown in Figure  2.  The collector
used at DePaul  worked the same but  was  slightly different in  construction.
The sampler works as follows.  When rain  is  anticipated,  the  collector is
cleaned and the peristaltic pump is turned on.  The pump  creates a partial
vacuum in the apparatus.  The  water from  the collector is channeled  to the
funnel and drawn  through a tared glass  fiber filter (GF/A; Gelman
Instruments; Ann  Arbor, MI) by the  vacuum  in the apparatus.  This separates
the particulates  in the rain  from the "soluble" portion,  i.e.,  the
filtrate.  The  filtrate then  flows  under  gravity through  the  polyurethane
foam, coated with DC-200 silicone oil,  in  the  bottom  of the sampler.  The
water level  in  the sampler  is  maintained  above the foam by the  use of the
siphon tube  on  the discharge.  The  water  is  pumped from the sampler  into a
barrel.  After  a  precipitation event, the  volume of water in  the barrel  is
measured to  determine the amount of water  sampled.  The amount  of rainfall
in an event  was separately measured with  an  adjacent  rain gage.

       Besides  making the determination of the PCBs on particulates
possible, the filtration of the  rain  is necessary since it has  been  reported
that PCBs on particulate matter  pass  uncoilected through  polyurethane foam
(Bedford 1974).

       The partial vacuum  in  the sampler  created an additional  problem.
Since the water being filtered is saturated  with air, it  is rapidly degassed
after passing through the filter.  A  bypass  was provided  for  this air to
prevent  it from interfering with the  foam extraction.  Some of  the PCBs  in
the water could vaporize at this time.  To collect these  PCBs,  and also  any
that evaporate  from the particulates  when  air  passes  through  the filter,
polyurethane foam was placed  in  the by-pass.

       Finally, to ease the remote  operation of the sampler,  the portions of
the sampler  containing the foam  are made  detachable.   Thus, the person
handling the sampler  at Beaver Island was  supplied a number of  filters and
several sets of the interchangeable parts  of the sampler, containing clean
foam.  After a  rain event, the detachable  portions of the sampler containing
the foam were wrapped in aluminum foil  and mailed to  DePaul University for
extraction and  analysis.  A new  filter  and containers with clean foam were

-------
                           Filter
                                     24/40
                         24/40 j
                        Foam
                  Pump
                                    \


                                 18/9
55/50

Foam
       FIGURE 2.  PCB SAMPLER FOR WATER SAMPLES.
Gas
Well
^iri
     Pump
                    Foam
 FIGURE 3.  SAMPLER SET-UP FOR SAMPLING GASES FROM A LANDFILL.

                        7

-------
placed on the sampler.  About  200 ml of  70%  ethanol  were  added  to  cover the
foam in he bottom of the sampler in order  to force the  air  out  of  the  foam
and to permit the water to flow freely thru  it.

       A final complication with this sampler is  that the flow  rate  is fixed
by the pumping rate, one liter/minute for  the pumps  used  in this  project.
Thus, during a rain storm when the rainfall  rate  is  high, not  all  of the
water will be sampled.  Also,  as the filter  becomes  clogged with
particulates, the filtering rate decreases.   Fortunately, much  of  the
particulates are in the early  part of the  rain when  the filtering  efficiency
of the sampler is high.  Loss  of sample  due  to overflow of  the  filter  will
affect the results only if the concentration of PCBs in the portion  which
overflows  is different from that which  is  sampled.
                                                          2
     2 The interception area of the collectors was  1.29 m -at  DePaul and
1.2 m  at  Beaver Island.  A rain rate of 0.8 mm/min  (1.9  in/hr) then
corresponds to about one liter/minute of water.   The average rainfall
intensity  for most rain events is much  less  than  this,  but  many contain
periods of high rainfall intensity.

       To  determine the severity of the  problem,  a  recording rain  gage
(Belfort  Instrument Co.) was used  in Chicago in addition  to the standard
accumulation type gage.  Inspection of  the records  for  20 rain  events  showed
that while a significant amount of rain  from some events  occurred  during
periods of high rainfall rates, the rainfall rate for most  of  the  water
which fell was below 0.8 mm/min.   With  some  variability,  the average volume
of  sample  per  unit of  precipitation was  1.06 L/mm in Chicago and  0.98  L/mm
on  Beaver  Island.

       This  problem could  be minimized  further or corrected for in several
ways:  the area of the collector could  be  decreased; the  area  of  the funnel
could be  increased to  lower  its flow resistance and  spread  the  particulates
over a larger  area; the volume of  the funnel could  be  increased for  more
storage capacity;  and  finally, an  overflow could be incorporated  on  the
funnel to  collect excess water in  an  adjacent container which  would  then be
manually  poured through the  sampler  after  the rain  event.

       Air samples were collected  with  a standard high  volume  air  sampler
(General  Metal Works;  Cleves,  OH)  modified to permit plugs  of polyurethane
foam  (Bidleman and Olney  1974) to  be  inserted in the air stream after  the
filter paper.  Glass fiber  filters (GF/A;  Gelman Inst.; Ann Arbor, MI) were
used to collect the particulates.   Five to 7.5 cm of foam was  used.  The
foam was  coated with DC-200  silicone  fluid  (Uthe et ji]_. 1972).

       The gases discharged  from the  landfill vent  were sampled at the exit
port of the  vent pipe. The  apparatus  used  is shown on  Figure 3.   It
permitted  different sample  tubes  containing polyurethane foam to  be
connected  to the vent. The  extra  piping on the sampler permitted excess gas
to  be vented,  and the  pressure in  the  vent monitored to insure that there
was always positive pressure  in the  vent pipe so that outside air was  not
drawn into the sampler.

-------
       The flow rate of the gases which passed through  the  sampler  was
measured by passing this gas through three rotometers connected  in
parallel.  The rotometers had been calibrated by connecting  the  intact
sampling train to a closed 216 L (55 gal) drum which had a manometer
connected to it.  The sampler was operated for short periods,  and volume  of
air pumped from the drum was calculated from the known  volume  of the  drum
and the pressure change caused by the pumping.  This permitted the  flow rate
to be determined.  The system was calibrated at several different flow rates.

-------
                                  SECTION 3

                                   ANALYSIS


       Solvents used were pesticide grade (Burdick  and Jackson Laboratories;
Muskegon, MI) or were distilled from glass through  a packed  column.  Drying
agents, cotton, and chromatographic packing materials wece extracted with
hexane/acetone and dried before use.  Activated Florisil  was dried at
130 C before use and stored  in a closed bottle at room temperature  .
Aroclor  analytical reference standards (U. S. EPA, Health Effects
Research Laboratory, Research Triangle Park, NC) and individual  PCB isomers
(Analabs Inc.; North Haven,  CNKwere used as received.   Polyurethane foam
(density = 0.024 or 0.036 gm/cm ) was extracted with water,  acetone, and
acetone/hexane before use.

       The polyurethane foam plugs from the samplers were extracted with
hexane/acetone in a Soxhlet  extractor.  The foam from the air bypass was
extracted along with the foam from the filtered ("soluble")  water fraction.
The glass fiber filter containing the particulates, the  non-filterable
fraction, was separately extracted in a Soxhlet by  placing it in a  clean,
tared cellulose extraction thimble.  The extracts from the particulates were
evaporated to dryness and their weight determined.  This weight  is  reported
as the organic portion of the particulate fraction.  The extracts from the
foam contained the silicone  coating and thus the weight  of the organic
fraction from the filtered portion could not be obtained.

       The extracts from the filtered and non-filterable fractions  in
hexane, were placed on a hexane-washed Florisil column.  The hexane
fraction was collected and concentrated in a Kuderna-Danish  evaporator with
a three-ball Snyder column.

       The weight of the non-extractable particulates was determined by
weighing the extraction thimbles containing the filter paper after
extraction and subtracting the weight of the thimble and filter. Since the
extraction thimbles were hygroscopic, the following procedure was followed
to determine their weight.   The thimbles, with or without contents, are
heated to 110 C for 30-45 minutes  in  an oven, wrapped quickly in a  piece
of aluminum foil, and allowed to cool.  The foil-wrapped thimble is weighed,
emptied of its contents, reheated, and reweighed.   The weight of the
particulates is obtained by  subtracting the weight  of the filter and thimble
from the total weight.  The  tare weight of the thimbles  was  reproducable to
±2 milligrams.
                                       10

-------
       The PCBs in the samples were quantified by the method  of Webb  and
McCall (1973).  This involves the use of AroclorR standards and the
percent contribution to these standards of eight or more of their major
components to determine the absolute GC detector response factors  (area/ng)
for each of the 23 major peaks which are separated by the GC.  These
response factors are then used to determine the amount  (ng) of each of the
23 different peaks.  Rather than reporting the amount of each of these 23
peaks, they were grouped in a manner very similar to the scheme proposed by
Webb and McCall (1973).  The peaks which comprised the  di-, tri- and most of
the tetrachlorobiphenys, were summed and are reported as AroclorR 1242.
The peaks which comprised the remainder of the tetrachlorobiphenyls,  along
the penta-, and most of the hexachlorobiphenyls were summed and reported as
AroclorR 1254.  The peaks which comprised the remainder of the
hexachlorobiphenyls, along with the hepta-, and octachlorobiphenyls were
summed and reported as AroclorR 1260.  Although the PCBs found in the
samples are reported as AroclorsR, in no case did the pattern of PCB peaks
in a sample closely resemble that which occurs in a commercial AroclorR
sample.

       The PCBs were separated on a 1.8 m x 2 mm, 3% OV-101 GC column
temperature programmed from 180 to 22QOC.  The injector was kept at
260°C and the detector at 350°C.  A Ni-63 electron capture detector was
used with a pulsed voltage, constant current source.  The areas of the
different peaks were obtained with the use of a digital electronic
integrator (Supergrator-1; Columbia Scientific Ind. Corp.; Austin, TX).

       The efficiency of the extraction and clean-up of the PCB samples was
checked on several occasions.  The efficiency of the recovery of PCBs from
the foam and their clean-up was checked twice by adding 2,3,4,5,6-penta-
chlorobiphenyl (740 & 880 ng) to the foam and extracting with a mixture of
hexane/acetone in a Soxhlet for seven hours.  The extracts were cleaned-up
and analyzed in the normal manner.  The overall recovery was 78 and 84
percent.  Further extraction of the foam yielded less than one percent
additional material.

       The efficiency of the clean-up procedure was independently checked on
four occasions by adding known amounts of PCB isomers (296 to 840 ng) to the
chromatographic clean-up column and proceeding in the normal manner.
Recoveries of the added PCBs averaged 95+11 percent.

       On two occasions the three foam plugs from the water sampler were
separately extracted.  The bottom foam contained no detectable PCBs on both
occasions.

       One one occasion, the two pieces of foam from the air  sampler were
separately extracted and analyzed.  One fourth of the total PCBs were found
on the second piece of foam indicating an overall sampling efficiency of
about 88%.

       In order to be consistent with the reporting procedures followed by
other workers, and to permit easy comparison with other reports, the  results
here are uncorrected for analytical efficiency.

                                      11

-------
CALCULATIONS

       The object of this project is to determine the inputs of PCBs into
Lake Michigan.  This is not directly accomplished by determining the
concentration of PCBs in precipitation because the concentration of PCBs in
a precipitation sample is usually inversely proportional to the amount of
precipitation (Hicks 1966; Gatz and Dingle 1971).

       To take into account the concentration of PCBs as well  as the amount
of precipitation, the volume weighted mean concentration of PCBs was
determined for the samples.  This mean is calculated by multiplying the
concentration of PCBs found in the sample (ng/L) by the amount of
precipitation (mm) to give an input of PCBs in ng/m2.  The volume weighted
mean for the  samples in then calculated  by summing the inputs  for the
samples and dividing by the total amount of precipiation the samples
represent.

   Volume weighted mean PCB concentration = £ Concentration X  Precipitation
                                                       Precipitation
                                       12

-------
                                  SECTION 4

                                   RESULTS
      The volume weighted mean concentration for all of the samples from
Chicago and Beaver Island which gave useful results are shown in Table 1.

      Four of the samples from Chicago collected early in the project, were
not filtered and, based on the results of Bedford (1974), it was presumed
that the particulate matter in these samples and their proportion of the
PCBs, were inefficiently collected.  The concentrations of PCBs found for
these samples was then assumed to come only from the filtered portion.  From
the other 18 samples, it was determined that 34 + 23 percent of the PCBs
were on the particulates.  These four samples were included in the results
shown above by increasing their PCB concentration by 51 percent (1/(1-0.34))
to approximate the inclusion of the particulates.

      A number of different types of samples were collected and analyzed for
PCBs during the course of this project.  Besides the rain samples, these
include: snow samples, Lake Michigan water samples, air samples and gas
samples collected at the sanitary landfill.  The results obtained from the
analyses of these samples are shown in Table 2 along with the percent of the
PCBs which was associated with the filtered, ("soluble") part of the sample.

      The composition of the samples with respect to the different PCB
compounds varied considerably from sample to sample.  In general, however,
there were different patterns from the different types of samples as shown
in Figure 4 and in Table 3.  The compositions shown were calculated from the
sum of the different AroclorsR for aT! of the samples of each type.  For
       TABLE 1.  VOLUME WEIGH I ED MEAN CONCENTRATION OF PCBs IN RAINFALL
                                                  Chicago   Beaver Island
        Total amount of precipitation (mm)         386.5        110.1

        Number of samples                            22           13

        Volume weighted mean concentration (ng/L)   104          138


                                      13

-------
                   TABLE  2.   THE  MEAN  CONCENTRATION  OF  PCBs
                                               Arithmetic    Percent  PCBs  in
      Sample  Type                  Number  of      Mean  PCB       Filtered
                                   Samples    Concentration      Portion

   Rain  (ng/L)  Chicago               22           160             66
               Beaver  Island         13           215             53

   Snow  (ng/L)  Chicago                4           212             36

   Lake  Water (ng/L)  Chicago          2            41             30
                     Beaver Island    1            30             43

   Air (ng/m3)  Chicago                4           7.6             97
               Mammoth Cave NP        1           6.7             95

   Landfill Gases (ng/m3)             2           3240
the rain samples, a volume weighted mean distribution of the PCBs was also
calculated.  The results showed little difference from the arithmetic mean.
To be consistent, the arithmetic mean of the rain samples is included in
Table 3.

      For a number of the rain samples, the amount of the hexane/acetone
extractable material, the organic fraction, of the particulate portion of
the sample was derermined.  If it is presumed (see Discussion) that all of
the PCBs in the rain samples came from particulates scavenged by the
precipitation, then the concentration of PCBs on the particulates and in the
organic fraction of the particulates can be determined.

      Shown in Table 4 are the results of calculations which show the
concentration of PCBs in the total particulates and organic particulate
fractions if it is assumed that 1) either all of the PCBs present in the
sample were associated with the particulate organics; or 2) that only the
particulate PCBs were associated with the particulate organics.  For
instance, the mean amount of PCBs found on particulates in the rain samples
in Chicago was 697 nanograms/sample.  If these PCBs are assumed to be
associated only with the organics in the particulate matter (8 milligrams/
sample), the PCB concentration  in the organics is 88 mg/kg.  Also, the PCB
concentration in the total particulates (260 mil Iigrams/sample) then is 2.6
mg/kg.

      Since the organic fraction  is  small, the errors  in their determination
are large.  Therefore, the weights of this fraction and the calculations
based on them are not very precise and  should be treated only as
approximations.


                                       14

-------
                                           Percent of Total Sample
RAIN-Chicago



    -Beaver Island



SNOW-Chicago



LAKE WATER



AIR


LANDFILL
  Filterable
Participates
       Total

  Filterable
Participates
       Total

  Filterable
Participates
       Total

  Filterable
Particulates
       Total

  Filterable
Particulates
       Total

       Total
0

25
1
50
1
1
=|
75
1
1
100
66

                                   i
                         1  C
J
                 53
                 47
                 36
                 64
                                   Aroclor 1242  Aroclor 1254  Aroclor  1260
        FIGURE 4.  THE PCB COMPOSITION OF THE DIFFERENT TYPES OF SAMPLES

-------
       TABLE 3.   THE AROCLORR COMPOSITION OF THE DIFFERENT SAMPLE TYPES
                                  Filtered           Non-Filterable
                                  Aroclors              Aroclors
          Sample Type         1242  1254  1260      1242  1254  1260
Rain -
Snow -
Chicago
Beaver Island
Chicago
Lake Water
Air
Landfi

11
54
34
54
16
85
98
33
43
31
73
14
2
13
23
15
10
1
0
29
31
24
42
19
-
56
46
57
54
34
-
15
23
18
4
47
-
      There is quite a bit of variability between different samples which is
not reflected in the results shown in Tables 1-4.  The arithmetic mean and
standard deviation of the concentration of PCBs in the rain samples from
Chicago and Beaver Island are 160 + 125.4 ng/L and 215 ± 161 ng/L
respectively.  A t-ratio test indicates that the difference between these
means is not significant (t=1.128; N=35).  Therefore the PCB concentration
in precipitation in Chicago and Beaver Island are not different.  The
geometric means for these sample sets are 109 mg/L and 164 mg/L.

      For both the Chicago and Beaver Island rain samples, there was an
inverse correlation, significant at the 5 percent level, between the amount
of precipitation and the concentration of the PCBs in the samples.  This is
not an unusual finding and can be explained by a depletion of the
particulates being scavenged by either rainout (Hicks 1966) or washout (Huff
and Stout 1964).  See also Gatz and Dingle  (1971).

      Two precipitation samples were collected simultaneously with air
samples.  This permitted the scavenging ratio (ng/kg PCBs in precipitation)/
(ng/kg PCBs in air), and the types of PCBs  to'be simultaneouly determined in
each of these samples.  The scavenging ratios found for the two samples were
13 and 37 for a mean of 25.

      From the results listed above  in Table  1, an input of PCBs from
precipitation to Lake Michigan can be calculated.  Since the difference
between the means from the samples from Chicago and Beaver Island was not
statistically significant, the samples from both locations were combined and
a volume weighted mean concentration of 111 ng/L was calculated for all of
the samples.  This concentration, when multiplied by the mean precipitation
on the Lake, 740 mm/yr (Jones and Meredith  1973), the area of the Lake
(58,000 km2), and the units corrected, results in an input of 4800 kg/yr
of PCBs to Lake Michigan from precipitation.

                                      16

-------
            TABLE 4.  THE CONCENTRATION OF PCBs IN ORGANIC MATTER
                                                  Sample Type

                                            Rain          Snow  Lake   Air
                                      Chicago  Beaver Is.       Water
Number of Samples
Particulates per Sample (mg)
Organ ics
Total
PCB Concentration in Particulates
Total PCBs per Sample (mg/kg)
Organic Particulates
Total Particulates
Non-Filterable PCBs per Sample
Organic Particulates
Total Particulates
17

8 + 8.4
260 ± 195


340
8
(mg/kg)
88
2.5
11

5 ± 5.3
76 + 56


240
16

120
7.5
2

9.8
140


220
11

150
5.5
3

3.8
82


140
5.5

100
4
4

10 ± 4
58


600*
200*

18*
3*
 Biased by the fact that most of the PCBs on air particulates are
collected as vapor.
                                     17

-------
                                  SECTION 5

                                  DISCUSSION

ATMOSPHERIC INPUTS

Extent of the Problem

      An unexpected result,  and  a significant finding, of this study was the
finding that the concentration of PCBs in precipitation was as high on
Beaver Island as in Chicago.  This means that either there is a local source
of PCBs to the atmosphere  in the vicinity of Beaver Island, which is
unlikely, or that PCBs  are being very widely distributed by the atmosphere.

      There is other evidence to support the wide distribution of PCBs by
the atmosphere.  Concentrations  of 230 ng/L of PCBs have been found in snow
samples collected on Isle  Royale National Park in Lake Superior (Swain
1977), 800 km north of  Chicago.  Also, the air sample collected in this
project in Mammoth Cave National Park, which is  located in a rural area
remote from centers of  population and industry,  and 650 km south of Chicago,
had concentrations of PCBs similar to those found in Chicago (Table 2).
Strachan et al.  (1977)  report lower  concentrations of PCBs found in
precipitation  in the Lake  Ontario basin.

Effects on_ Lake Michigan

      The results of the measurements made  in this project are that 4800
kilograms of PCBs are coming into Lake Michigan  in precipitation per year.
To gage the significance of these inputs, the other known and estimated
sources of PCBs  are listed in Table  5.   It can be seen that precipitation  is
now the major  source of PCBs to  Lake Michigan.

      The industrial use of PCBs for such uses as heat transfer fluids,
hydraulic fluids, in die casting operations, etc., has greatly decreased
since 1971 when  they ceased to be sold for those purposes.  As there was
little or no control over  the use or discharge of PCBs prior to 1971, it is
not known how much of these materials were discharged and reached Lake
Michigan, or where these discharges  occurred.  An illustration of the
possible magnitude of these sources  to the Lake, is a manufacturer in
Waukegan, Illinois, who used a commercial hydraulic fluid containing PCBs.
The water effluents from this plant  go mostly to Waukegan harbor, but some
go to a ditch leading directly to the Lake.  The manufacturer estimated that
of the 3.8 X 106 kilograms of PCBs they  purchased over a twenty year
period, as much as 15%  to  2056 could  have been discharged with their water
effluents (Thomas 1976).  Sediment samples taken in the harbor in 1976 show
concentrations of PCBs  as  high as 4  mg/gm,  and sediments from the ditch were
as high as 0.6 mg/gm (IJC  1976b).
                                      18

-------
              TABLE b.  INPUTS OF PCBs (KG/YR) 10 LAKE MICHIGAN
                                              Prior to
           Sources                              1975       1977
           Major Industrial Discharges*       (25,000)

           Precipitation*                       4,800     4,800

           Streams and Waste Water Effluents1-"   1,650     1,650

           Others5                             (2,500)   (2,500)
           Total                               33,950     8,950

 This estimate is a lower limit.  The actual amount may be several times
larger.

*This work.

^Hesse, J. L. (1976).  Conference Proceedings.  National Conference on
Polychlorinated Biphenyls, Nov. 19-21, 1975, Chicago, IL., EPA 560/6-75/004;
NTIS, PB-253 248/9PSX; p. 127.

^Kleinert, S. (1976).  Conference Proceedings.  National Conference on
Polychlorinated Biphenyls, Nov. 19-21, 1975, Chicago, IL., EPA 560/6-75-004;
NTIS, PB-253 248/9PSX; p. 124; Personal communication.

LdSchacht, R. A.  Pesticides in the Illinois Waters of Lake Michigan.  EPA
660/3-74-002, January 1974.

$An estimate to attempt to account for dry deposition of PCBs on the Lake
surface and to allow for small industrial discharges not yet identified.
      To gage the significance of the precipitation inputs of PCBs to the
Lake, the other known and estimated sources of PCBs to the Lake are listed
in Table 5.  These include the effluents from waste water treatment plants,
those industries known to have PCBs in their discharges, and tributary
inputs.

      If Table b is reasonably accurate, and industrial discharges have
greatly decreased, then when the Lake comes to equilibrium with the lower
inputs, the concentration of PCBs in the water and biota should be much less
than the present values.  The data here indicate that the atmosphere will be
the major contributor of PCBs and therefore, the levels of PCBs in the Lake
and biota should reflect the magnitude of the inputs from this source.

                                      19

-------
      The most recent information on the concentration of PCBs in Lake
Michigan fish indicates that levels in chubs  and  coho salmon have declined
by a statistically significant amount below the maximum  values of recent
years (Seelye 1977).  Thus, it seems as  if the concentration of  PCBs  in Lake
biota are now beginning to respond to the decreased  inputs of PCBs  to the
Lake.

      The data in Table 3 and Figure 4 indicate that the different
compartments associated with the  Lake contain very different compositions of
PCBs.  For  instance,  the PCBs in  air are 84 percent  AroclorR 1242 while
the Lake water is only 34 percent; and 66 percent of the PCBs in rain are
filterable  while only 33 percent  of the  PCBs  in the  Lake are.  There  are too
few samples  in most  cases to give reliable information,  and many of the
compartments  in the  Lake are not  represented, but it seems clear from these
different compositions that,  in the past, sources other  than the air  and
precipitation have contributed PCBs to Lake Michigan.

DISTRIBUTION  OF PCBs IN THE ATMOSPHERE BETWEEN VAPOR AND PARTICIPATES

      At the  present, it has not  possible to  determine  the extent to  which
PCBs  are present  as  vapor  and on  particulates in  the atmosphere. This  is
because  PCBs  on  particulates, captured on the filter paper of high  volume
air  particulate  samplers,  evaporate  into the  air  stream  and  are  collected  as
vapor.   Thus,  the  total  amount of PCBs present  in the  atmosphere can  be
readily  determined,  but  their distribution between  vapor and particulate
forms  can  not be.   The  information  obtained  in this  project, however, leads
to a tentative resolution  of  this problem.

       Where do the PCBs  found  in  the  precipitation  come from?   Either from
PCBs present as  vapor which dissolved  in the  precipitation,  or  as  PCBs
adsorbed on particulates  which  were  scavenged by precipitation.   The
scavenging  of PCB vapor  by rain  can  be ruled  out  on  theoretical  and
experimental  grounds.  On  theoretical  grounds,  the  distribution  constant  for
the partition of PCBs between  air and  water,  the Henry's Law Constant (HLC),
 is about 20 for  AroclorR 1242  and 1254 (Junge 1977;  Mackay and  Leinonen
 1975;  Mackay and Wolkoff 1973).   It  should  be similar  for  Aroclor  1260.
Since the  data in Table  2 indicate  that  the  the  concentration  in
precipitation is  about  400 times  that  calculated  using the mean
concentration in air and the HLC, PCB  vapor   is  not  an  important  source of
PCBs to  precipitation.   McClure   (1976)  has  arrived  at  similar  conclusions
for DDT.

       Experimentally, the  scavenging of  PCB  vapor by the precipitation also
can not  explain  the amounts of  Aroclor 1260 found in the precipitation
samples.   From the summaries of the compositions of the air  and
precipitation samples in Table 2 and Figure   4,  the  air samples  contain about
 3 percent  AroclorR 1260, while the precipitation samples contain more than
 20 percent-.  The lower water solubility of the AroclorR 1260 compared to
 the other AroclorsR studied here, would preclude the factor  of six
 enhancement in the precipitation as compared to the air.  The greater
 percentages of the AroclorR 1260 in the precipitation samples can only be
 explained by the air particulates being higher in AroclorR 1260, and the

-------
PCBs in precipitation coming chiefly from the scavenging of these
participates.  The greater proportions of the lower molecular weight
AroclorsR found in the air samples must then be due to material present  in
the atmosphere as vapor.

      Precipitation then is providing the solution to the sampling dilemma
which exists for PCBs in the atmosphere.  If all of the PCBs in the
non-filterable fraction of the precipiation samples, and most, if not all,
of the PCBs in the filtered portion are coming from air particulates, then
precipitation is serving as a selective sampler for air particulates.  And,
while a rain droplet may have a different scavenging efficiency for
different sized air particulates, this would make a difference only if the
PCB composition of different sized particulates were different.

      A comparison of the air and rain samples collected permits an
approximate resolution of the PCBs present in the atmosphere into vapor  and
particulate forms.  As mentioned before, the air samples show a mean
Aroclor^ 1260 concentration of about 3 percent, while the rain samples
show a mean AroclorR 1260 composition of 20 percent.  Thus, a mass balance
based on Aroclor* 1260 indicates that only about 15 percent of the PCBs  in
the atmosphere are on particulates.  These results are in accord with the
predictions of Junge (1977) that compounds with vapor pressures of about
10^ mm of Hg, should be present in the atmosphere chiefly as vapor, even
in urban atmospheres.

      The data in Table 4, page 17, support this conclusion.  The values
shown for the total PCB concentration in air particulates are clearly high
by a factor of 2 to 4, if it is assumed that all of the PCBs found in the
vapor and particulate phases were on the particulates.  The simple explana-
tion for this is that not all of the PCBs collected were on particulates,
but rather were present in the atmosphere as vapor.  The low value for the
non-filterable PCBs/sample for the air samples, is due to the evaporation of
the PCBs from the filter and their collection as filtered PCBs.

      The evidence that much of the PCBs in the atmosphere in the Lake
Michigan region are present as vapor has important ramifications for the
validity of methods recently reported for the collection of the dry fallout
of PCBs (Sodergren 1973; McClure iy76) and DDT (Young et al. 1976).  These
methods all utilize collectors coated with a non-polar liquid to capture and
hold particulate matter-  But, since PCBs are quite soluble in non-polar
liquids, the partition coefficient into the liquid would be such that the
equilibrium concentration on PCBs would be in the milligrams per liter
range.  Thus, PCBs present in the atmosphere as vapor would be expected  to
be efficiently scavenged by the liquid coating.  The deposition results
reported from the use of these sampling techniques would include not only
PCBs from dry deposition, but also from vapor scavenging.  In fact, the
scavenging of PCBs in the vapor state by non-polar liquids is so efficient
that it is the basis for a method for the collection of PCB vapor  (Harvey
and Steinhauer 1974).

      Further evidence that the PCBs in precipitation are coming from the
scavenging of air particulates comes from measurements of the size of the

                                      21

-------
organic fraction and the scavenging efficiency one would expect for such
particulates.  Gatz (1977) has reported a correlation between scavenging
ratios and the mass median diameter (MMD), of the particulates being
scavenged.  The smallest scavenging ratio he reported was 76 for lead, with
a MMD of about 0.5 micrometers.  Andren (1976) and Strand (1977) have
reported that organic particulates in the atmosphere above Lake Michigan are
less than 2 micrometers in MMD.  The scavenging ratio of 25 found  in this
project is, when corrected for the fact that much of the PCBs are  present  as
vapor, is in good agreement with the sizes found by Andren and Strand.
Although other factors undoubtedly contribute to precipitation scavenging,
the low scavenging ratio found for these samples indicate that the
particulates being scavenged  are quite small.


PCBs  IN ORGANIC PARTICULATES

       In order to be  able  to  account for the distribution of PCBs  in the
environment  and their movement through the environment, the concentration  of
the PCBs  in  the different  compartments of the environment must be  known.
Due to their non-polar  nature and  large n-octanol/water partition
coefficient, PCBs not in  the  vapor phase are always found associated with
non-polar organic matter.  Therefore,  if the concentrations of PCBs  in this
organic matter, and the movement of the organic matter through the
environment  is known, the  movement of the PCBs through the environment will
also  be known.

       Data  on  the mean  amount of hexane/acetone soluble  (organic)  and
insoluble particulates  present  in  the  non-filterable fraction of the
different types of  samples collected  in the  project are in Table 4.  Also
shown are the  results of  calculations  based  on different assumptions on the
source of the  PCBs  in the  samples.  Probably most of the PCBs in
precipitation  samples which  are filterable are associated with small
particulates.  Thus the concentration  of PCBs  in organic material  in the
atmosphere  may be best  represented by  the concentration of PCBs  in the
organic portion of  the  particulate fraction  of the precipitaion
samples.These  results are  all close to 100 mg/kg.  The similarity  of these
values to those found for  the concentrations of PCBs in the fat  of fish from
Lake  Ontario (Heile et  al. 1975),  may  be more than a coincidence.

PCBs  IN GASES  FROM  LANDFILLS

       PCBs  in  landfills have  heretofore been considered only as  a  source of
water pollution.  While samples from  one day, from a single landfill, allow
no  conclusions to  be  drawn,  the  high  concentration of PCBs found in  these
samples  (Table 2, p.  13)  suggests  that landfills may be a source of  PCBs to
the atmosphere and  that further  studies are  in order.

PERCHLORINATION

       The quantification  of  PCBs  is  complicated  by the large number  of
separate  chlorinated  biphenyl compounds present  in a sample  and  by the
different sensitivity of  the  electron  capture  (EC) detector tot  he different

                                       22

-------
compounds.  It has been reported (Armour 1973) that, upon reaction with
antimony pentachloride, PCBs are quantitatively perchlorinated to the fully
chlorinated biphenyl, decachlorobiphenyl (DCB).  Thus all of the different
PCBs present are converted into one compound, and of all of the PCBs, the EC
detector has the greatest sensitivity to DCB.  It was hoped that the use of
this reaction would greatly simplify the quantification of the PCBs.

      While some initial control experiments using AroclorR 1254 were
satisfactory, further experiments with AroclorR 1242 gave only about 25
percent conversion to the DCB.  Careful attention to the evaporation of the
samples raised this to 38 percent.

      To check on the possibility that the perchlorination of the pure PCB
standards was not different from the perchlorination of samples which
contained large amounts of other organics, the following standard addition
experiment was carried out.  A portion of a precipitation sample was
perch!orinated and the DCB produced was determined.  A known amount of
AroclorR 1242 was then added and the sample again perchlorinated.  Only 40
percent of the theoretical amount of DCB was formed.

      Since the yield of DCB was a function of the composition of the PCBs
in the sample, and the percent of AroclorR 1242 in the different types of
samples varied from 9 percent to 98 percent, it was decided that this method
was not suitable for the quantification of the PCBs in the samples collected
in this project.
                                      23

-------
                                 BIBLIOGRAPHY


Allen, J. R. and D. H. Norback.  1976.  Pathobiological Responses of
Primates to Polychlorinated Biphenyl Exposure.  National Conference on
Polychlorinated Biphenyls.  Conference Proceedings, EPA-560/6-75-004, U. S.
Environmental Protection Agency, Washington, D. C., Chicago, It.  p. 43.

Andren, A. A.  1976.  10th Great Lakes Regional Conference, American
Chemical Society,  June, 19/6,  Evanston, IL.  Quoted by J. 0. Delfino, Envir.
Sci. and Tech., 10, 986.

Armour, J. A.  1973.  Quantitative  perchlorination of PCBs as a method for
confirmatory residue measurement and analysis.  J of the Assoc. of Off. Anal
Chem., 56_, 987.

Aulerich, R. J., R. K. Ringer  and S. Iwamoto.   1973.  Reproductive failure
and mortality in mink fed on Great  Lakes fish.  J. Reprod. Pert. Suppl., 19,
365.                                                                     ~~

Barsotti, D. A., R.  J. Marlar  and J. R. Allen.  1976.  Reproductive
dysfunction  in rhesus monkeys  exposed to low  levels of polychlorinated
biphenyls  (Aroclor 1248).  Fd.  Cosmet. Toxicol. ]_4, 99.

Bedford,  J.  W.   1974.  The use of polyurethane  foam plugs for extraction of
PCBs  from natural  waters.  Bull. Envir. Contam. and Tox., 12, 622.

Bidleman, T. F-  and  C. E. Olney.  1974.  High volume collection of
atmospheric  PCBs.   Bull.  Envir.  Cont. and  Tox., 11, 44.

Buckley,  J.  L.   1972.  Polychlorinated biphenyls; Still prevalent but less
of  a  problem.  Quoted  by  T.  H.  Maugh, II,  Science 178, 388.

Gatz,  D.  F.  and  A. N.  Dingle.   1971.  Trace substances in rain water:
concentration  variations  during convective rains, and their interpretation.
Tell us 23,  14.

Gatz,  D.  J.   1977.  Scavenging ratio measurements in METROMEX.
Precipitation  Scavenging  1974.   Champaign, IL,  Symposium Series f41, ERDA
(CONF-741003), p.  71

Grant.,  D.  L., J.  Mes, and R.  Frank.  1976.   PCB residues in human adipose
tissue and  milk.   Conference Proceedings;  National Conference on
Polychlorinated  Biphenyls, Chicago, IL.  U. S.  Environmental Protection
Agency,  EPA-560/6-757-004; p.  144.

Greenfield,  S. M.   1957.  Rain scavenging  of  radioactive particulate matter
from  the  atmosphere.   J.  of  Meteor.  1_4, 115.

Hague, R.,  D. W.  Schmedding  and V.  H. Freed.  1974.  Aqueous solubility,
adsorption  and vapor behavior  of AroclorR  1254.  Envir. Sci. and Tech. 8,
139.

                                      24

-------
Harvey, G. R. and W. G. Steinhauer.  1974.  Atmospheric transport of PCBs to
the North Atlantic.  Atm. Envir. 8, 777.

Halle, C. L., G. F. Veith, G. F. Lee and W. C. Boyle.  1975.  Chlorinated
hydrocarbons in the Lake Ontario ecosystem.  U. S. Environmental Protection
Agency, EPA-660/3-75-022.  Grosse He, Mich.

Hicks, B. B.  1966.  Nucleation and the wet removal of fallout.  J. of
Applied Meteor. 5_,  169.

Huff, F. A. and G.  E. Stout.  1964.  Distribution of radioactive rainout in
convective rainfall.  J. Appl. Meteor-  3^,  707.

Humphrey, H. E. B.  1976.  Evaluation of changes of the level of
polychlorinated biphenyls in  human tissue.  Mich. Dept. of Health; Final
report on U. S. FDA contract.

Ingersoll, B. 1977.  The miracle that got away.  Chicago, WFMT  Inc., 500 N.
Michigan Ave., Chicago, IL  60611.  August, p. 10U"!

International Joint Commission.  1974.  Great Lakes Water Quality 1974,
Appendix B, Surveillance Subcommittee Report,  p. 161.

International Joint Commission.  I976a.  Great Lakes Water Quality 1976.
Appendix B, Surveillance Subcommittee Report, p. 79.

International Joint Commission.  1976b.  Great Lakes Water Quality 1976.
Append-ix B, Surveillance Subcommittee Report, p. 36.

Jelinek, C. F. and  Corneliussen.  1976.  Levels of PCBs in the  U. S. food
supply.  Conference Proceedings.  National  Conference on Polychlorinated
Biphenyls.  Chicago, IL.  U.  S. Environmental Protection Agency,
EPA-560/6-75-004.   p. 147

Jones, D. M. A. and D.  D. Meredith.  1972.  Great Lakes hydrology by
months.  1946-65.   Proc. 15th Conf. of Great Lakes Research,  p. 477.

Junge, C. E.  1977.  Basic considerations about trace constituents in the
atmosphere as related to the  fate of global pollutants.  Proc.  ACS Symposium
on the Fate of Pollutants in  the Air and Water Environment.
Wiley-Interscience.  I. H. Sut'fet, Ed., p.  7.

Mackay, D. and A. W. Wolkoff.   1973.  Rate  of evaporation of  low-solubility
contaminants from water bodies  to atmosphere.  Envir. Sci. and  Tech. 7_f 611.

Mackay, D. and P. J. Leinonen.  1975.  Rate of evaporation of low-solubility
contaminants from water bodies  to atmosphere.  Envir. Sci. and  Tech. 9_,  11/8.

McClure, V.  E.  1976.   Transport of heavy chlorinated hydrocarbons in the
atmosphere.  Envir. Sci. and  Tech. 10, 1233.
                                       25

-------
Musial, C. J., 0. Hutzinger and V. Zitko.  1974.  Presence of PCB, DDE and
DDT in human milk in the provinces of New Brunswick and Nova Scotia,
Canada.  Bull, of Envir. Contain, and Toxic. 12_, 258.

Neely, W. B., D. R. Branson and 6. E. Blau  1974.  Predicting
bioconcentration potentials.  Envir. Sci . and Tech., £$,
Reed, N. P-  1976.  I). S. Federal Agency Roles and Actions: The Department
of the Interior.  National Conference on Polychlorinated Biphenyls.
Conference Proceedings,  EPA-560/6- 75-004, U. S. Environmental Protection
Agency, Washington, D. C.  p.  434.

Savage, E. P.,  J. D. Tessari,  J. W. Malberg, H. W. Wheeler and J. R. Bagby.
1973.  A search for polychlorinated biphenyls in human milk in rural
Colorado.  Bull, of Envir. Contam. and Toxic.  9_, 222.

Seeyle, J.   1977.  Results of  Fish Contaminant Monitoring.  Personal
Communication.  Great Lakes  Fishery Laboratory, 1451 Green Road, Ann Arbor,
MI   481 U5.

Sbdergren, A.   1972.  Chlorinated hydrocarbon residues in airborne fallout.
Nature 236,  395.

Sodergren, A.   1973.  Transportation, distribution and degradation of DDT
and  PCBs  in  a  south Swedish  Lake ecosystem.  Vatten 2^, 90.

Strachan, W. M. J., F. C.  Elder, H. Huneault and W. M. Schertzer.  1977.
Presented at the 20th Conference on Great Lakes Research; Ann Arbor. MI;
May.  1977.
                                                                    «
Strand, 0. W.   1977.  Polyaromatic hydrocarbons and total organic carbon
aerosol content above Lake Michigan.  20th  Conference on Great Lakes
Research.  May, 1977, Ann  Arbor, MI.

Swain, W. R.   1977.  Persistant organic compounds in the vicinity of Isle
Royale.   J.  of Great Lakes Research,  in the press.

Tarrant,  K.  R. and  J. O'G. Tatton.  1968.   Organochlorine pesticides in
rainwater in the British Isles.  Nature 219, 725.

Thomas, H. T.   1976.  Letter to Mr. James McDonald; Chief, Enforcement
Division, Region V, U. S.  EPA;  March 15.

Uthe, 0.  F., J. Reinke and H.  Gesser.   1972.  Extraction of organochlorine
pesticides from water by porous polyurethane coated with selective
adsorbants.  Envir. Letters  3, 117.
         »
Willford, W. A., R.  J. Hesselburg  and L. W. Nicholson.   1976.  Trends of
polychlorinated biphenyls  in three Lake Michigan fishes.  Conference
Proceedings, National Conference  on Polychlorinated Biphenyls, Chicago, IL.
U.  S. Environmental  Protection Agency,  EPA-560/6-75-004, p.  177.


                                       26

-------
Webb, R. G. and A. C. McCall.  1973.  Quantitative PCB standards for
electron capture gas chromatography.  J. of Chrom. Sci.   11,  366.

Woodwell, G. W., P. P. Craig and H. A. Johnson.  1971.  DDT  in the
Biosphere: Where does it go?  Science  174. 11U1.

Woodwell, G. W., P- P. Craig and H. A. Johnson.  1972.  DDT  in the
Biosphere: Where does it go?  Science  177, 725.

Young, D. R., D. J. McDermott and T. E. Heeson.  1976.  Aerial fallout of
DDT in southern California.  Bull of Envir. Cont. and Toxic.   16, 604.
                                      27

-------
                                   APPENDIX


      Data from the different samples collected and analyzed in this project
are included in the appendix.  They include:

            1. Rain samples collected in Chicago.

            2. Rain samples collected on Beaver Island.

            3. Lake Michigan water samples.  The May, 1976 sample was
            collected in about 0.5 m of waber at Beaver Island.  The other
            samples were collected from a water intake crib of the city of
            Chicago in about 10 m of water.  All of the samples were
            collected from beneath the surface.

            4. Snow samples collected in Chicago.

            5. Air samples.  The Nov., 1975 sample was collected in Mammoth
            Cave National Park.

      The percent composition of the AroclorsR in the non-filterable
portion of the samples shown in the appendix was calculated from a data base
which included the measured percent composition for the filtered and total
portions only.  In those cases where the non-filterable portion of a
particular AroclorR constitutes only a small portion of the total amount,
there may be  significant errors in the percentage shown in the appendix
tables.  The  results  in the report for the  non-filterable portions of the
samples were  based on the measured values and thus should be free from these
errors.
                                      28

-------
CHICAGO
                          SAMF>L
SAMPLING
DATE
BEGUN DURATION
YY MM DD 
-------
SAMPLING
DATE
BEGUN DURATION
YY MM DO (HR)
76 05 09
76 06 13
76 06 15
76 O6 18
76 O6 24
76 07 29
76 08 00
76 08 27
76 08 31
76 09 14
76 09 29
76 10 04
76 10 14
5.7
9.5
3.7
2.3
13.0
0.0.
0.2
0.0
16.0
7.0
5.0
13.O
3.0
RAIN
(MM)
12.2
5.0
2.3
5.6
0.9
5.8
14.7
12.2
13.2
4.5
5.2
26.7
1.8
VOLUME
COLL EC
15.8
6.1
2.1
6.5
1.3
7.2
7.5
4.8
16. 5
5.2
6.5
28.3
2.1
FILTERED
X AROCLOR TOTAL
1242 1254 1260 
37.4 52.7 9.8
42.0 46.1 11.6
19.2 69.9 10.8
20.7 43.3 35.9
43.3 40.0 16.5
51.3 33.4 15.2
19.6 46.3 33.9
40.0 55.3 4.6
24.1 73.5 2.3
58.0 32.6 9.2
22.5 56.5 20.8
37.0 34.1 28.8
29.8 44.5 25.6
392
879
285
711
342
1429
337
670
570
220
39ft
1034
506
NON-FILTERABLE
X AROCLOR TOTAL
124S 1254 J26O 
-------
CO
        SAMPLING
      DATE             VOLUME
      BEGUN  DURATION  COLLEC
    YY MM DO   (HR)       
-------
  SAMPLING
DATE             VOLUME
BEGUN  DURATION  COLLEC

CO
ro



YY MM DD
75 11 27
75 12 05
76 02 20
76 03 29
76 04 19

32.0 20. 7 21.4 20.0 28.9 (M3) 1788 11O6 1O72 1138 954 FILTERED x AROCLGR "TOTAL 1242 1254 1260 (NG) 80.3 19.1 0.5 31469 93.2 6.2 0.6 7517 86.4 10.2 3.4 3646 75.3 24.4 0.2 11275 94.7 4.7 0.5 1O385 NON-FILTE.RABLE X AROCLOR ' TOTAL 1242 1254 126O
-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
1. REPORT NO.
  EPA-600/3-78-071
                              2.
                                                           3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
  Polychlorinated  Biphenyls in Precipitation in the
  Lake Michigan  Basin.
             5. REPORT DATE

               July  1978  issuing date
             6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
  Thomas J. Murphy
  Charles  P.  Rzeszutko
                                                           8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
                                                           10. PROGRAM ELEMENT NO.
  DePaul University
  1036 W. Belden  Avenue
  Chicago,  Illinois  60614
                1BA769
             11. CONTRACT/GRANT NO.

                Grant 803 915
12. SPONSORING AGENCY NAME AND ADDRESS
                                                           13. TYPE OF REPORT AND PERIOD COVERED
  Environmental  Research Laboratory  -  Duluth
  Office of Research and Development
  U.S. Environmental Protection Agency
  Duluth, Minnesota  55804
                Final  Report
             14. SPONSORING AGENCY CODE
                EPA 6QO/U3
15. SUPPLEMENTARY NOTES

   Project Officer:   Michael D. Mullin,ERL  -  Duluth, Grosse He, MI  48138
16. ABSTRACT
       Rainfall  samples were collected  in  Chicago, Illinois,  and  on Beaver Island,
  Michigan, and  analyzed for polychlorinated biphenyls  (PCBs).  The precipitation
  weighted mean  concentration of 35  samples  of rain was 111 ng/1.   (Ill  parts per
  trillion).   This  would result in the  deposition of 4800  kg/yr of PCBs  to the Lake
  from precipitation.   Presently available evidence on  other  sources of  PCBs to the
  Lake indicates that  precipitation  is  now the major source of PCBs to the Lake.
  The future  PCB problems in the Lakes  will  then be determined mainly by the magni-
  tude of atmospheric  inputs to the  Lake.
       The concentrations of PCBs in  rainfall  were found to be as  high on Beaver
  Island as in Chicago.
       Results obtained from the simultaneous  sampling  of  air and  precipitation in-
  dicate that PCBs  are present in the atmosphere as vapor  as  well  as being present
  on particulates.   This result raises  doubts  as to the validity  of results for the
  dry deposition of PCBs obtained from  the use of collectors  covered with mineral
  oil or other non-polar liquid.
       PCB concentrations in the parts  per billion range obtained  from gas samples
  from a vented  sanitary landfill, indicate  that PCB containing materials incor-
  porated into landfills may be an important source of  PCBs to the atmosphere.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.lDENTIFIERS/OPEN ENDED TERMS
                             COS AT I Field/Group
    Precipitation
    Atmospheres
    Air
    Rain
Lake Michigan
Polychlorinated biphenyls
Particulates
Landfills
PCB vapor
                                                                            08/H
18. DISTRIBUTION STATEMENT
  Release to  the  Public
19. SECURITY CLASS (This Report)
  Unclassified
21. NO OF PAGES

       39
                                              20. SECURITY CLASS (This page)
                                                Unclassified
                                                                         22. psice
EPA Form 2220-1 (9-73)
                                             33  £• U. S. GOVERNMENT PRINTING OFFICE: 1978-757-140/1421 Region No. 5HI

-------