Pates and Biological Effects of
   Polycyclic Aromatic Hydrocarbons in
   Aquatic Systems
   Savannah JUver Ecology Lab.,  Aiken,  SC
   Prepared for

   Environmental Research Lab.,  Athens,  GA
   Jul  83
                                                               PB83-250191
US.         tf

-------
                                   TECHNICAL REPORT DATA
                            (Ftteu read liusmdiont on the rmne I* tan coniplritngj
i  REPORT NO.
                                           T
                                            t
                                            i
                                                          3. P.F.CI,
4. TITLE ANOSUBTITLE
  Fates and  Biological Effects of Polycyclic Arotratlc
  Hydrocarbons in Aquatic Systems
                                                        6. REPORT DATE
                                                        6. PERFORMING ORGANIZATION CODE
TJS ACCESSION-NO.
~     250191
                                                         July 1QB3 ...„	
7. AUTHORlS)
  John  P.  Ciesy,  Steven M. Barteil, Peter F. Landrum,
  Gordon J.  Leversee, and John W. Bowling
                                                        ». P ERFORMINQ ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Savannah  River Ecology Laboratory
  U.S.  Department of Energy
  P.O.  Drawer E
  Aiken SC   29801
                                                        10. PROGRAM ELEMENT NO.
                                                           CCULIA
                                                           78-0-X0290
 12 SPONSORING AGEf.CY NAME AND ADDRESS
  Environmental  Research Laboratory—Athens GA
  Office of Research and Development
  U.S. Environmental Protection Agency
  Athens GA  30613
                                                        13. TYPE OF REPORT AND PERIOD COVERED
                                                           Final, 6/78-5/81
                                                        14. SPONSORING AGENCY CODE

                                                           EPA/600/01
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
        This  research project was conducted to test  the hypothesis  that  fates of poly-
  cyclic  aromatic hydrocarbons (PAH) in ecosystems can be predicted by mechanistic sim-
  ulation models  based on easily measured properties of the compounds  in this homolog-
  ous series.   To accomplish this goal our research efforts were  in four major areas:
   (I) development of a mechanistic, predictive simulation model based  ort kinetic rather
  than  thermodynamic considerations; (2) development of analytical  and quality assuranc
  protocols  forthe extraction and quantification of PAH associated  with  biological and
  geological  matrices;  (3) laboratory studies to determine the vectors  of and rate
  constants  for uptake, depuration and biotransformation of PAH by  aquatic organisms
  and sediments;  (These studies also examined the effect of PAH concentration, temper-
  ature and other exogenous factors on rata constants and determined whether rate
  constants were  first order.) and CO a microcosm study to compare the  results of
  simulation  and  laboratory scale studies to a larger scale ecosystem  study.
                               KEY WOHOS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                             b.lOENTIF I6RS/OPEN IKDfO TEBMS  C COSATI i icld/Gfcmp
         TIGN STATEMENT
                                              19, 61CUHITY CLASS
                                                 UNCLASSIFIED
                                                                         5. VI t'AOtS
                                                                             245
RELEASE TO PUBLIC
                                              JO. StCuntTY CLASS (nil (mgef
                                                 UNCLASSIFIl'D
                                                                        23. PRICE
EPA farm iJJO-l (»-7J)

-------
                                                                     PB83-250191
                                                                EPA-600/3-83-053
                                                                July  1983
                 FATES AND BIOLOGICAL EFFECTS OF POLYCYCLIC AROMATIC
                           HYDROCARBONS IN AQUATIC SYSTEMS


                                          by


John P. Giesy, Steven M. kartell, Peter P. Landruo, Gordon J. Leversee, John W. Bowling
                           Savannah River Ecology Laboratory
                                 University of Georgia
                                     P.O. Drawer E
                              Aaken, South Carolina  29801
                                 Inteiegency Agreement
                                       7B-D-X0290
                                        between
                               U.S. Department o£ Energy
                                          and
                               Environmental Protection Agency
                                    Project Oflicor
                                     Harvey W. Holes
                              Environiaental Systems Branch
                           Environmental Research Laboratory
                                 Athens, Georgia  30613
                                         RESEARCH LABORATORY
                           OFFICE OK RESEARCH AND DEVEI,OPMENT
                          U.S. EMVIRO5JMKNTAL PROTJX^TION AGENCY
                                 ATHENS, GEORGIA  30613

-------
                                 DISCLAIMER

     This report has been reviewed by the Savannah River  Ecology Laboratory,
U.S. Environmental  Protection Agency, and  approved for publication.   Ap-
proval does not signify that the contents necessarily reflect  the views  and
policies of  the  U.  S.  Environmental Protection Agency, nor does mention of
trade names or commercial products constitute endorsement or recommendation
for use.
                                    ii

-------
                                 FOREWORD

      Environmental protection efforts are increasingly directed towards
preventing adverse health and ecological effects associated with specific
compounds of natural or human origin.  As part of this Laboratory's research
on the occurrence, movement, transformation,  impact,  and control of envi-
ronmental contaminants, the Environmental Systems Branch studies complexes
of environmental processes that control the transport, transformation,  degra-
dation, and impact of pollutants or other materials in soil and water and
assesses environmental factors th?'. affect water quality.

      Polycyclic aromatic hydrocarbons (PAH)  from natural  and man-made  sources
are widely distributed in the environment and pose potential health problems
to aquatic animals and humans.  An important  phase in addressing these  pro-
blems is developing an understanding of the cycling and fate of these compounds.
This report presents the development and evaluation of a theoretical, predic-
tive nodel of the fate of PAH based on easily measured characteristics  of
the compounds and the environments to which they may  be released.

                                     William T.  Donaldson
                                     Deputy Director
                                     Environmental Research Laboratory
                                     Athens,  Georgia
                                    ill

-------
                                  ABSTRACT

     This research project  was  conducted to test ~he hypothesis that fates
of polycyclic aromatic hydrocarbons (PAH) in ecosystems can be predicted by
mechanistic  simulation  models based  on  easily measured properties  of  the
compounds in  this  homologous  series.   To accomplish this goal our research
efforts were in four major areas.  1) Development of  a mechanistic,  pre-
dictive  simulation model based  on kinetic  rather  than thermodynamic  con-
siderations;  2)  Development of analytical  and  quality assurance protocols
for the extraction and quantification of PAH associated with biological and
geological matrices;  3^  ^iborutory studies to determine the vectors of and
rate  constants  for  uptake,  depuration  and biotransformation  of PAH  by
aquatic organisms and sediments.  These studies also examined the effect of
PAH  concentration, temperature  and other  exogenous  factors on  rate  con-
stants and  determined if rate constants were  first  order;  and A) A micro-
cosm  study  to compare the  results  of  simulation  and  laboratory scale stu-
dies  to a larger scale ecosystem study.

      Laboratory  studies  indicated  that anthracene and benzo (a) pyrene are
rapidly  biotransfonned  by  fish  and dipteran larvae but not by periphyton
communities.   Biotransformation had  a  significant effect  on  the  steady
state  concentrations  of  parent compound  and  biotransformation products.
These  results demonstrated  that predictions of steady state concentrations
based  on    C-labeled  parent compound and  the  octanol-water partitioning
coefficient  of  the parent compound would be  in  error.   Thus, the octanol-
water  partitioning coefficient  would  not be  a  good predictor  of the  be-
havior of  PAH in aquatic organisms.   Uptake  and  depuration rate constants
were  first  order for  fish  but not  dipteran larvae.  Induction of biotrans-
formation  and  changes  of  this rate over  time  means  that  results  frou
short-term pharmacokinetic  studies, using radio-labeled compounds, will be
misleading for compounds which are  biotransforratd.

     Anthracene  (approximately  12 Jig*?.  )  was  acutely  t.oxic  to  bluegill
sunfish  dosed  in   outdoor  channels  microcosms.   This  mortality was  not
observed in  laboratory studies and  was shown to be a photo-toxic mechanism.
Therefore,  laboratory studies  must be  conducted  under the  same  lighting
conditions  if the  results  of these studies arc  to  be realistic represen-
tations of field conditions.

     Ihe modeled processes  that influenced PAH tjranfiport included losses to
volatilization,  photolytic  degradation,   sorption  to suspended particulate
taattcr  and  sediments  and  net  uptake  by biota.   The  biota  in  the raodel
included phytoplankton, periphyton, rooted enacrophytes, bacteria, zoopiank-
ton,  two  functionally   defined  benthic  invertebrate  components  and  two
functionally  defined  categories of fish.   Model  simulations were compared
to  results  of experiments  conducted   in artificial  streams.   A 0.6 ptnolar

                                    iv

-------
solution of Anthracene in ethanol was continuously added into the headwaters
for IS days  folio-red  by termination of anthtacene input.   The  Booel accu-
rately predicted  the  dissolved anthracene  concentration through  time  and
space.  Uptake by periphyton  was overestimated by the  model)  however,  the
rate of depuration  ot  anthracene  by peripnyton was  reasonably simulated.
Photolytic degradation  appealed  to be the  most important pathway  ot  flux
within the channels, both experimentally and in the simulations.

     This report was  submitted in  fulfillment  of Interagency Agreement  No.
EPA-7w-l>-X0230 by  the  Savannah  River  Ecology  Laboratory  with  the  U.S.
Environmental Protection  Agency.   This  report covers  the  period  June  1,
1978 to Hay  31, 1981, and work was completed as ot Hay 31, 1981.

-------
                                  CONTENTS

Foreword 	 .....  iii
Abstract	   iv
Figures	vii
Tables	xiii
Acknowledgments  	 xvii

     1.   Introduction	    1
     2.   Conclusions  	    5
     3.   Recommendations	    7
     4.   Laboratory Studies 	    8
     4.1  Introduction 	    8
     4.2  Anthracene Sorption by Organic Sediments 	    9
     4.3  Effect of Hufflic Acids on Bioavailability and Transport
             of PAH	   28
     4.4  Uptake, Depuration and Biotransformation Kinetics of Benzo
             (a) Pyrene and Anthracene by Periphyton Communities ...   40
     4.5  Uptake Depuration and Biotransformation of Benzo(a)pyrcne
             by the Midge (Chironomous riparius) 	   56
     4.6  Effects of Temperature and Anthracene Concentration
             on Uptake, Depuration and Biotransformation of
             Anthracene bv Chironomus riparius  	   70
     4.7  Uptake, Depuration and Biotransforraation of Anthracene
             By the Claea (Anodonta imbecillis)	   86
     4.8  Uptake, Depuration and Biotransfontiation of Anthracene
             and Benzo(a)pyrene by the Bluegill Sunfish
             (Lepoais raacrochirui.)	   96
     5.0  Simulation Model for Predicting the Fates of PAH  in
             Aquatic Systems 	  113
     5.1  Introduction	113
     5.2  Model Structure Description  	  115
     5.3  Sub«odel Descriptions	123
     5.4  Simulation of Anthracene in Channels Microcosms -
             Resi'itc and Discussion	152
     6.0  Channels Microcosms Studies  	  169
     6.1     Facility Description and Methods   	  169
     6.2     Water	182
     6.3     Sediments	190
     6.4     Periphyton	193
     6.5     Clams	200
     6.6     Fish	202

Appendix I	209

References	210

                                    vi

-------
                                  FIGURES

Figure 4.2.1   Flow diagram of procedures used to spike, extract
               and analyze anthracene in stream sediment  	    10

Figure A.2.2.  Relative ion intensities of anthracene.  OV101
               capillary column, 2 C min. increase in temperature  .  .    14

Figure 4.2.3.  Relative ion intensities of anthtaquinnne.  OV101 •
               capillary column 2 C min. increase in temperature  ...    15

Figure 4.2.4.  High pressure liquid chromatograms of anthracene
               and anthraquinone extracted from sediment with
               35/65 ml acetonitrile/benzene.  254 nm UV detec-
               tion.  I = interferring peak, AN = anthracene,
               Aq - antbraquinonc	    15

Figure 4.2.5.  Relative ioo intensities of 7 mass fragments
               of anthrace- e in benzene	    18

Figure 4.2.6.  Relation ion intensity of anthracene in sedi-
               ment extracts	    19

Figure 4.2.7.  Relative ion intensities of mass fragments of
               anthraquinone standard in benzene  	    20

Figure 4.2.8.  Relative ion intensities of mass fragments from
               anthraquinone in sediment extracts 	    21

Figure 4.2.9.  Relative ion intensities of mass fragments from
               anthraquinone in sediment extracts 	    22

Figure 4.3.1.  Accumulation of B(a)P <»nd anthracene by D. magna as
               a function of time	    33

Figure 4.3.2.  Salting-out of BaP at six salinities in the pres-
               ence of Aldrich huraics	    36

                               14
Figure 4.4.1.  Accumulation of   C B(a)P by periphyton commu-
               nities from UTRC and Castor Creek which had been
               colonized for 3 or 5 weeks.  Accumulation is nor-
               malized to a surface area (a and b) and dry weight
               basis (c and d).   Each point represents the mean
               of 3 replications with confidence interval = 1 SD  ...    45
                                    vii

-------
Figure 4.4.2.
Figure 4.4.3.
Figure 4.4.4.
Figure 4.6.5.
Figure 4.4.6.
Figure 4.4.7.
Figure 4.5.1.


Figure 4.5.2.
Autoradiographs illustrating the deposition of
14C B(a)P in A) Desmidium coarctatum, B) Spondylosimn
pulchrum, C) Netrium digitus, D) Neiilium iridis, and
E) Eunotia sp.  All autoradiographs are 500X magni-
fication 	

                     14
Percent of recovered   C as B(a)P, shaded, and
non-B(a)P, unshaded, extracted from Castor Creek
periphyton.  The upper figure represents periphyton
from three week colonization, while the lower
figure represents periphyton from six week coloni-
zation period.  Samples of both live and dead
periphyton were taken after 0.25, 4 and 24 h 	
                     14
Percent of recovered   C ar> B(a)P, shaded and
non-B(a)P, unshaded, extracted from Upper Three
Runs periphyton.  The upper figure represents
periphyton from a three week colonization while
the lower figure represents five weeks of coloniza-
tion.  Samples of both live and dead periphyton
were taken after 0.25, 4 and 24 h  	
Histograms of quantities of F>(a)P shaded and
non-B(a)P, unshaded, 14C in Upper Three Runs and
Castor Creek periphyton.  Samples of live and
dead periphyton were taken after 0.25, 4 and 24 h
                14
Accumulation of   C-anthracene by periphyton which
colonized glass slides in LTTRC.  Mean bioraass/slide
= 0.4 mg, dry weight-cm  .  Anthracene concentra-
tion in water = 22 pg • £  .  Concentrations are
normalized to an area basis but can be converted to
a biomass basis with the conversion factor given.
Each point represents the mean of 3 slides.  Confi-
dence intervals are i SD 	
Dcsorption of anthracene fro» UTRC periphyton which
had been exposed to 14C-anthracene for 24 h.  Each
point represents the mean of three slides.  Confi-
dence intervals are ± SD. ?Mean periphyton biomnss
= 0.4 mg, dry weight "  ~~
                                      cm
Distribution of B(a)P and metabolites in static
uptake test  	
              14
Depuration of   C by C. ripariua with and without sub-
strate.  Concentration is expressed os.pg B(a)P • g
(wet weight) chironomid, assuming all   C is 8(a)P.
Each point '.i> the mean of 3 determinations ± SD  .  .   .
                                                                          46
                                                                          48
                                                                          49
                                                                          50
                                                                          51
52


60




63
                                   viii

-------
                                  14
Figure 4.5.3.  Uptake and loss of   C in whole C. riparius, cxo-
               skeleton and viscera.  Each point represents the
               mean of 3 determinations t SO	   65

                               14
Figure 4.5.4.  Accumulation of   C B(a)P (A) as determined Ly
               TLC, and B(a]P plus netabolites (o) expressed as
               ng b(a)P • g   (vet weight) chironomid.  Each
               point represents the mean of .1 determinations
               ± SD	   67
                               14
Figure 4.6.1.  Accumulation of   C anthracene by C. riparius at
               25  and 22 yg • £  .  Each point represents the
               mean ± 2 SE of 3 replicates.  The line represents
               the least squares fit of equation 2	   72

Figure 4.6.2.  Log K  for C. riparius at 4 different concentra-
               tions of anthracene in water.  Estimate ± asymp-
               totic 951 CI	   76

                             14
Figure 4.6.3.  Depuration of   C anthracene by C. riparius in
               uncontaninated water and paper towel substrate at
               30°C.  X ± SE, n = 3	   79

Figure 4.6.4.  Biotransformation rate after 1 h exposure as a
               function of temperature.  X ± 2 SE, n = 4	   81

Figure 4.6.5.  Biotransformation rate as a_function of time of
               exposure and temperature.  X .+. 2 SE, n = 4	   82
                                    14
Figure 4.6.6.  Biotransformation of   C anthracene at 10, 25 and
               30°C.  X ± 2 SE, n = 4	   83

Figure 4.7.1.  Accumulation of anthracene by A. Hnbecillis shell
               as a function of time.  The fitted line is based on
               the first 5 h of uptake only because of Iocs of
               anthracene from the water, which did not go to
               the clams.  This results in a decrease in the
               amount of anthracene associated with the shells
               (points after 24 h)	   90

Figure 4.7.2.  Accumulation of anthracene by soft tissues of A.
               imbcci 11 is as a function of time	   91

Figure 4.7.3.  Depuration of anthracene from the clam A.
               imbcci His.  The data have been fit to two compo-
               nent depuration model (Equation 4.7.4) 	   93

                               14
Figure 4.8.1.  Accumulation of   C-anthracene and benzo(a)pyrene
               from well water with and without humic ncids.
               Humics present (solid circles), no humics (op^n
               circles)	101

                                    ix

-------
                               14
Figure 4.8.2.  Accumulation of   C-anthracene by fish and dep]--
               tion of   C--
-------
Figure 5.4.5.  Simulation of dynamics of anthracene in herbi-
               vorous fish	,	161

Figure 5.4.6.  Simulation of anthracene dynamics in carnivorous
               fish	162

Figure 5.4.7.  Relative importance of non-advective physical/
               chemical processes (solid dots, dashed line) and
               biological processes (open circles, solid line)
               in total transport of (a) anthracene, (b) naph-
               thalene, and (c) benzo(a)pyrene through artifi-
               cial  streams	164

Figure 6.1.1.  View  of channels microcosm facility  	  170

Figure 6.2.1.  Anthracene concentrations in reach 1 during the
               input period of the abiotic study.  The closed
               circles represent samples taken st dawn, while
               closed triangles represent samples taken at
               noon.  Each point represents the m°an of repli-
               cate  samples.  The confidence interval is 1 SD	186

Figure 6.2.2.  Histogram representing anthracene concentra-
               tions in water from reaches 1,3, end 5, during
               the biotic study	187
Figure 6.2.3.
Figure 6.3.1.
Concentrations of anthracene associated with
plastic liner during the two channel microcosm
studies.  The points which represent samples
taken during the abiotic study are labeled as
such.  The other points are from the biotic
study.  Earh point represents the mean of 6
samples.  The confidence interval is 1 SD  . .
                                                                         189
Uptake and depuration of anthracene by organic
sediments in petti dishes in the channel micro-
cosm.  Closed squares represent the concentra-
tion of anthracene on a dry weight basis while
closed circles represent the same data normal-
ized tc au areal basis 	
                                                                         193
Figure 6.4.1.
Figure 6.4.2.
Anthracene concentrations in peripLyton in reach
1 as a function of time.  Upper figure represents
               anthracene'cm
               anthracene-g
                while the lower figure represent*
              ,  dry weight, pcriphyton 	
Anthracene concentrations in periphyton in reach
3 as a functiot) of tin.e.  Upper figure represents
                while the lower figure represents
                dry weight, periphyton 	
               anthracene'cra
               anthracene'g
197
198
                                    xi

-------
Figure 6.4.3.
Figure 6.5.1.
Figure 6.6.1,
Anthracene concentrations in periphyton in reach
5 as a function of time.  Upper figure represents
               anthracene-cm
               antbracene*g
                           -1
                while the lover figure represents
                dry weight, periphyton 	
199
Uptake of anthracene by soft tissues of clams
during the period of anthracene input to the
channels.  Each point represents the mean of 6
clams.  Confidence intervals of 2 SE are present-
ed.  The least squares predicted regression curve
for the indicated model is given 	
Cumulative mortality of bluegill sunfisb in the
channels microcosm.  Mean drawn anthracene concen-
trations were 12 (JR*^  •   Two cages at each reach
with 8 fish per cage.  Time is given as cumulative
exposure to anthracene.  Periods of light and
dark are given.  *denotes time at which all fish
vere dead
                                                                         201
                                                                         203
Figure 6.6.2.
Cumulative mortality of bluegill sunfisb in shaded
and unshaded reaches of the channels microcosms.
The weighted averages (day-night) of anthracene
vere 13 pg-iv-l
               n =  13; noon = 7.4, n = 21) and 9.5 dawn = 12.5,
               n =  42: noon = 5.6, n = 33) for caches 1,3, and
               5 respectively.  Two cages with 8 fish per cage
               were placed in each reach.  *denotes time at which
               all  fish were dead
                                                          204
Figure 6.6.3.
Figure 6.6.
Cumulative mortality of bluegill sunfish after 72
hr of exposure to anthracene in shaded or unshaded
reaches.  After 72 hr of anthracene dosing, anthra-
cene inputs were stopped and fish exposed to light .  ,  .   206

Cumulative mortality of bluegill sunfish as a
function of time.  Fish vere exposed to 14.6 (Jg
anthracene•£   for 48 h in the dark.  Sets of 8
fish were allowed to depurate into clean water
for 24, 48, 72, 96 and 144 h then placed in
clean water in the light.  Each set of fish
were placed in the light 0800 hrs and observed
for 48 h	207

-------
                                   TABLES

                             14
Table 4.2.1.   Extraction of   C-anthracene from direct and slurry
               spiked UTRC sediment	   12

Table 4.2.2    Experimental design to determine effects of sediment
               type, equilibration time, moisture content and polarity
               of solvent on extraction efficiency of anthracene from
               sediment	   13

Table 4.2.3    Effect of time of equilibration, mositure content,
               sediment  type and solvent polarity on extraction of
               14C-anthracene from Steel Creek and UTRC sediment  ...   24

Table 4.2.4    Anova of  experimental results given in table 4.2.1 ...   25

Table 4.3.1    Radio-labeled PAH used in bioavailability studies  ...   29

Table 4.3.2    Inorganic salts added to test water and resulting
               water quality parameters 	   30
                                A
Table 4.3.3    Effect of Aldricb  humic acids on PAH accumulation by
               D. yagna   	,	   32

Table 4.3.4    Effect of Aldrich  htunic acids on PAH accumulation
               by D. oajjina  . . .	   35

Table 4.3.5    Effect of Skinface Pond and Upper Three Runs Creek or-
               ganics and particulates on B(a)P accumulation by
               Daphnia magna  	   37

Table 4.3.6    Salting - out of PAH in water with and without humics   .   38

                                                             14
Table 4.5.1    Bioaccuraulation parameter estimates for total   C in
               Chironomoua riparius using one and two • compartment.
               models	   62

                                    14
Table 4.5.2    Biotransformation of   C - B(a)P by Chironomous ripart-
               us	"	   66

Table 4.6.1    Uptake (Ku) and depuration (Kd) rate constants in
               C. riparius at four concentrations and three tem-
               peratures (estimate ± 95% CI)	   77
                                                  14
Table 4.6.2.   Bioconcentration factors (BCF) for   C-anthracene in
               the midge C. riparius.  BCF for total   C wa* calcu-
                                   xiii

-------
               lated from K -K "  after 10 and 30 h of uptake.  BCF
               for anthracene was calculated from concentrations of
               anthracene at 4 hr. X ± 95% CI .  '.	   84

Table 4.7.1.   First order rate constants for uptake and depuration
               of anthracene by A. imbecillis from 31 pg anthracene.
               Estimated by Marquardt iterative least squares
               procedure	   92

Table 4.8.1.   Effect of varying exposure concentration on the rate
               of uptake of 14C anthracene by bluegills	105

                                                  14
Table 4.8.2.   Rates of uptake and elimination of   C activity in
               bluegills exposed to anthracene or bcnzo(a)pyrene
               (B(a)P) in stream water  	  107

                               14
Table 4.8.3.   Distribution of   C activity in bluegills exposed to
               5.8 pg*£   anthracene (An) or 0.70 |Jg'£   benzo(a)pyrene
               (B(a)P) for 4 hours	108

Table 4.8.4.   Rates of biotransformfiion in bluegills exposed to
               8.9 pg*£   anthracene of 0.96 Mg'ji   benzo(a)pyrene
               (B(a)P)	109

Table 4.8.5.   Comparison of bioconcentration factors determined by
               three methods	109

Table 5.2.1.   Output from FOAM.  Each variable is simulated hourly .   .  121

Table 5.J-1.   Parameter estimates for consumer components of fates
               of aromatic model (FOAM).  See Table 5.3.3 for expla-
               nation of abbreviations	126

Table 5.3.2.   Parameter estimates for primary producer components of
               fates of aromatic model (FOAM).  See Table 5.3.3 for
               explanation of abbreviations 	  127

Table 5.3.3.   Key to parameter abbreviations used in FOAM for con-
               sumer and producer components  	  129

Table 5.3.4.   Contribution of light of 10 nm increments to total
               spectrum between 300 and 500 not	132

Table 5.3.5.   Comparison of observed quantum yield coefficients
               for reaction of PAH in air saturated water to pre-
               dicted quantum yields based upon regression with
               molecular weight 	  133

Table 5.3.6.   Feeding preference (W  ) for predator i feeding on
               prey (food) items j, and a  , fraction PAH assimi-
               lated by predator i per unit of PAH ingested in the
               form of prey (food) j	141

                                    xiv

-------
Table 5.3.7.   Initial bioroass conditions for simulations reported
               here g • dry weight-m	

Table 5.3.8.   Biological state variables for simulation of anthra-
               cene dynamics.  See Table 5.3.3 for explanation of
               abbreviations   	  149

Table 5.3.9.   Biological state variables for anthracene dynamics
               simulation	150

Table 5.3.10.  Parameters for  the simulation of volatilization
               of anthracene	151

Table 5.3.11.  Parameters for  photolytic degradation.of anthracene.
               Quantum yield coefficient = 7.5 X 10~	  151

Table 5.4.1.   Predicted and observed concentrations of dissolved
               anthracene in channels microcosms  	  152

Table 5.4.2.   Comparison of predicted and observed concentrations
               of anthracene in stream sediments  	  156

Table 6.1.1.   Mean water quality of treated well water 	  171

Table 6.1.2.   Recovery and precision of anthracene and anthra-
               quinone from various matrices  	  177

Table 6.1.3.   Background and  limits of detection for analysis of
               Anthracene And  Anthraquinoae from various matrices . .  .  178

Table 6.2.1.   Percent recovery of anthracene and anthraquinonc
               from control water samples in abiotic channels
               experiment	  182

Table 6.2.2.   Calculated input of anthracene to channel and
               measured values in reach one	183

Table 6.2.3.   Concentration (Jg*£   of anthracene and anthraquinone
               in water	•	184

Table 6.2.4.   Percent accountability nM anthracene + nM anthra-
               quinone/actual  anthracene input (nM) 	  185

Table 6.2.5.   Input and mass  balance of anthraccne/anthraquinone
               in reach one	188

Table 6.3.1.   First order rate constants for uptake and release
               of anthracene from organic sediments during the
               channels microcosm experiment based on dry weight
               of pediment.  Data was fit by n = 3, PF < 0.001 for
               all regressions	192

-------
Table 6.3.2.   First order rate constants for uptake and release
               of anthracene from organic sediments, based on area
               of sediment.  Data was fit by the marquardt itera-
               tive least squares procedure.  X ± SE, n = 3,  PF
               < 0.0001 for all regressions	194
                                    xvi

-------
                              ACKNOWLEDGEMENTS

     The cooperation and support of the staff of the University of Georgia's
Savannah River Ecology Laboratory (SREL) is much appreciated.   The research
and report preparation weve supported, in part, by interagency agreement
No. DE-AC09-76SR00819 between the United States Department of  Energy and
SREL.  We wish to acknowledge the technical support ot James Cheatham and
Susan Giddings.  We wish to thank Hs. Jean Coleman, Tonya Willingham and
Debbie Perks for their help in preparing this report.  Karen Brown assisted
with sampling and statistical analyses of experiments on sediment studies.
                                   xvii

-------
                                  SECTION 1

                                INTRODUCTION

     Polycylic  (polynuclear)  aromatic  hydrocarbons  (PAH  or  PNA)  are  a
homologous  series  of  compounds composed  of 2  or more  condensed  benzene
rings with  occasional  incorporations  of cyclopentene rings, such as in the
fluorenes  or  "hetero-octanes"  (N,  0  or  S).   PAH  derived from  natural
(Andelm&n and  Snodgrass,  1974) and man-made sources are widely, distributed
in the  environment.   PAH occur as natural  products  in  plants and microbes
and natural  pyrolytic  processes,  such as forest fires and volcanic activi-
ty, and  human  activities such as manufacturing  and  fossil  fuel conversion
(Braunstein  et  al. ,  1977;  Harrison et a±., 1975; and Suess, 1976).  Inputs
to  aquatic  systems  are  due to  human activities, as  found in  east  coast
near-shore  marine  sediments  (Kites,  et al., 1977) or of natural origin in
sediments of pristine  lakes  (Brown and Starnes, 1978) and marine sediments
(LaFlamme and Kites, 1979).  LaFlamme and Kites  (1978) reviewed the sources
of PAH in the environment and concluded that while some of the PAH could be
from natural sources,  most of the PAH were from pyrolytic processes due to
human activities.   PAH released to the environment by human activities are
due  to  oil  spills  and fossil fuel conversion  (Neff,  1979).   Over 230,000
metric  tons of PAH  enter  the  oceans  and surface waters  of  the world  each
year (Neff,  1979).   The  American Petroleum Institute (1978) has compiled a
lengthy  review of  the  sources and environmental  concentrations  of  PAH in
the  environment.   Inputs of  PAH into aquatic systems  are  expected  to in-
crease with  the commencement of the synthetic fuels program and emphasis on
coal combustion as  a  primary energy source  (Gehrs  e_t  al.,  1976).   After
reviewing  the  available  literature  on projected  effluent  concentrations,
solubility,  degradation  waste treatment  efficiency  and  acute  and chronic
effects  of  coal conversion effluents, Herbes et  al.  (1976)  concluded  that
higher molecular weight polycyclic and heteroaromatic compounds possess the
greatest potential  for bioaccumulation,  carcinogenic and mutagenic effects
in aquatic animals and man, and acute toxicity to aquatic organisms.  Thus.
the  importance of  developing  predictive  models  of PAH  fate  in  aquatic
systems  must be emphasized (Baughman and Lassiter, 1978).  For this reason
we undertook two  lines of research:   1) to develop an understanding of the
cycling  and fate of  these compounds including  bioaccumulation, sorption,
degradation  and transformation  in a quasi-natural  environment and 2) to
determine effects of PAH compounds on aquatic ecosystems.

     Toxic,  mutagenic,  and carcinogenic  properties  of  PAH  have been  much
studied  (LaVoic et aj.., 1979; Norden, ejt aj.. , 1979).   Many PAH arc carcino-
genic or have  carcinogenic biotransformation products  (Christensen ej. al. ,
1975; Neff,  1978)  and  may be hazardous  to  man  through bioraagnification or
be directly mutagenic to aquatic organisms  (Neff, 1979).

-------
     Mechanisms of  PAN  biotransformation, mode of action,  and  analyses  in
biological and geological  matrices  have been much studied and this litera-
ture has  beec reviewed  (Braunstein e_t al.,  1977;  Jones  and  Leber,  1979;
Neff, 1979; Bjorseth and Dennis, 1980). However, few studies have attempted
to integrate laboratory, field and simulation studies.

     Arsessment of  health  risks, associated with introduction  of  PAH into
the  environment,   depends   in  part  upon  quantification  of  environmental
transport  and subsequent  dose  concentrations  experienced by  inhabitants
(Crawford  and Leggett,   1980).   Estimates of transport  and dose  are  inde-
p*ndent of the specific nature  of  the threat  to human  health,  either di-
rectly  through contamination of potable water,  for  example,  or indirectly
through damage to ecological life support systems.

     We have  the recent experiences, for example, of worldwide distribution
of DDT and PCB in aquatic  life of many rivers and bays, mercury and cadmiua
poisonings of the  populations  in certain industrial areas, and vinyl  chlo-
ride  exposure of  industrial workers. These  experiences  arose  from  acci-
dental  release,  uncontrolled industrial  releases, and unsafe  overuse.  In
order  to  avoid these types of experiences in  future years with  the  thou-
sands  of  new chemicals  being released to  the  environment, a  toxic effects
testing strategy is needed which will "flag" potential problem contaminants
prior  to  their widespread dispersal.   This testing strategy  should  cover
both  plant  and  animal  species  and  be  concerned  with  acute  and chronic
exposures.   Because of  restrictions on time, resources and tiuraan cognitive
capacity,  a  number of techniques have been established  to assess probable
effects of trace  contaminants.   These techniques include  toxicity tests on
individual  or standard  organisms,  individual   bioaccumulation  tests,  mea-
surement  of  physical  properties of the chemical, environmental monitoring,
studies  in simple systems and  investigation of transport and degradation.

     To depend only on retrospective environmental studies to elucidate the
knowledge  required  to avoid adverse effects is to court disaster.  Rather,
what  is  urgently  required  is  a systematic  program of  investigation the
scale  of  which time  and  resources permit  expeditious  acquisition of the
essential  data and understanding.   While  the  tests listed  above provide
useful  information,  they  do  not  consider  complex  interactions  between
biotic  and  abiotic components  of  the  environment   and  do not  allow for
evaluation  of  complex  biological  interactions.   Over  time,   very  subtle
effects  may  have  serious  impacts  on  populations  or communities  as they
aodify complex interdependences.

     The  so-called "benchmark11  techniques  have been developed on the as-
sumption that  the chemical and physical characteristics of a trace contami-
nant  will deteimine  its effects on  the biocenose  and  how it  fluxes and
cycles  through the environment and that the fate of a compound can be pre-
dicted by  simple first or  second principles.  While  this assumption has not
been tested  and  may,  in fact, be untestable,  correlative studies indicate
that  the   assumption  is somewhat valid.   However,   low  predictability and
high variability  indicate  that  the correct parameters  are not being mea-
sured  for  each compound or  that the  interactions  between the compounds cf
interest  with biotic  and  abiotic  components  are non-additive.   That is,

-------
there  are  significant  interaction terms  which  can  not  be  evaluated  by
staple tests  or knowledge  of  the first principles of  trace  contaminants.

     Because  of the  vast  difference between  simple  laboratory tests  and
complex field situations, scientists have formed conceptual and operational
bridges between  the  two systems.   These include models of theoretical con-
ceptualization  of  how complex  systems operate.   Computer  science  has pro-
vided  the  bookkeeping .capacity  to  describe  these  systems  with  complex
linear and non-linear functions.  The science of modeling and  environmental
predictability  suffers  from a  lack  of steady state  conditions, rate con-
stants and an Mnderslanding of the most important parameters for increasing
the power  of  predictability.   If  the correct  physical  and cheaical  param-
eters  of  both an organic compound of interest and the environment  to which
it is  being released are known, a model could be constructed to predict the
gross  movement  of  the compound of interest within  that environment.   This
model  could also include degradation and transformation due to hydrolysis,
photolysis  and  by biota.   Another useful type of model  of ecosystems are
the  simplified  subsets  of ecosystems called  microcosms  (Metcalf et al.,
1971;  Lee and Takahashi, 1*77; Lu et al.., 1977; Giesy, 1980).   We feel that
a microcosm approach (artificial  streams) allowed us to evaluate a predic-
tive  model under  more  natural, complex conditions with  control of  inputs
and replication  of the experimental conditions.

     For example,  most information on plutonium  indicated  that  it was not
ecologically  hazardous  since  it was not readily taken up by plant  roots or
assimilated through  the intestinal wall of animals.  Hence, its presence in
the environment  was  more a problem of physical transport in air and  coils,
end  its  biological  significance,  with  regard to effects, was  limited  to
inhalation or movement into wounds.  Later, field data showed  quantities of
plutonium  in  trophic  levels and tissues where it was not expected.   Studies
were  conducted  which  showed  that in the environment plutonium  can  be or-
ganically  complexed  by microorganisms  in the soil,  and  that  it  was then
much  more available  for  transfer through  food  chains all the way  up  to
vertebrate  consumers of the vegetation.   Effects within  the  system could
then be predicted based on new modes of entry within the intact system, and
a  change  in  rate constants  for  transfers  between components  within the
system.

     Thousands of  different PAH compounds are chemically possible.  There-
fore,  application  of elaborate screening protocols (Duthie, 1977)  to indi-
vidual PAH in order to estimate major processes of transport,  accumulation,
and .degradation  is  impractical for purposes of  rick  analysis.  However,  it
IB important  to  answer questions  about these  compounds  such  as:  Can human
exposure  be  limited  by more  rigorous drinking  water treatment  aimed  at
removing  organics?  Do  the various  compounds  become  concentrated  in and
magnified  by  trophic levels culminating in human exposure?  Is concentra-
tion  in  various trophic components proportional  to  time  and  concentration
of  exposure?   What  are the  rates of  accumulation  and  elimination from
various components?   What  are  the rates of uptake and elimination relative
to  immobilization  rates  by sorption  processes  and  photo-  and roicrobial

-------
degradation  rates.  Because of  the large number  of  compounds  and environ-
ments to  be  evaluated a certain level  of  predictability must be achieved.

     This  study tested the hypothesis  that  a  theoretical predictive model
of  the  fate of PAH  compounds  or classes of compounds  in a given environ-
ment, or class of environments, can be constructed based on characteristics
of  the  compound of  interest  and the environment to which it  is released.
This  report  describes a test of the hypothesis that the transport of three
PAH,  anthracene,   naphthalene,   and   bcnzo(a)pyrene,   in  lotic  systems
(streams,  rivers,  reservoirs)  could   be  predicted  from  the  fundamental
chemistry  of  individual  PAH   which  embodies  sufficient  information  for
prediction.   While  we  do  not   feel  that a  simple  characteristic,  such as
nolecular  weight  is  necessarily the best parameter, from which to predict
environmental  fates,  we  feel that  the law of parsimony dictated t'at we try
a  simple parameter,  such  as molecular weight, to  try  to make predictions
from  and determine where  this  simplification  fails  to  predict accurately.
Many  modeling efforts  rely  on measured properties, such as octanol-water
partitioning  coefficients, which  are  a function of first principles of a
compound.  These  measured  parameters often  integrate several properties of
compounds, which  is  an advantage  because more information can be included
in  a single empirically measured  value.   However,  Ihese coefficients must
be  measured, which  is time-consuming  and  they may be  less accurate than
molecular  weight measureoients.

      We  feel  that  a  coordinated  synthesis  of  information  on ecosystem
structure  and function is  required to  assess  environmental impacts  of any
technology which must ultimately be interfaced vith the environment.   While
this  research  was on  PAH compounds (anthracene and benzo(a)pyrene specific-
ally)  it  tested  the  hypotheses that  valid generalizations about  the be-
havior  of general classes  of organic compounds in complex  environments can
be  made from  knowledge about  the compounds obtained  in relatively simple
laboratory studies  and  the  environments to which  these compounds p.re re-
leased.   Knowledge  of  this  type can be transferred to other orgar.ic com-
pounds  and has  broad  applicability.

      Anthracene was  chosen as  a model  PAH  for laboratory, field and simu-
lation  studies  because  it  is a  commercially  important PAH which  is produced
in  large quantities  and used extensively as a  reagent  in organic synthesis
(Archer  et al., 1979).  Anthracene has  also  been  used frequently  as a model
PAH  for  studies  of  environmental fate and transport  in  aquatic syotetrs
(Ausmus  et al., 1980)  or physiological  disposition in aquatic biota (Roubal
cj.  a1.,  1977).  Anthracene is  non-carcinogenic  (MAS,  1972) and  relatively
uon-toxic  (HAS, 1972;  Dcickman  and Gerarde,  1969).

-------
                                  SECTION 2

                                 CONCLUSIONS

     The results of the laboratory studies demonstrated that PAH are rapid-
ly biotransformed by bluegill sunfish and chironomid larvae but not unionid
clams or periphyton  assemblages.   Biotransformation rates were affected by
temperature  and time  of exposure.   The  rates  of biotransfonnation  also
affected the bioconcentration factors predicted  from  short-term pharmaco-
kinetic studies.  Temperature and food ration also affected rates of uptake
and  depuration.  Uptake  and depuration were  first-order with  respect  to
forcing function concentrations.   Depuration was generally multiphasic  with
some  of  the biotransformation products  bound to  tissues,  such  that  they
were very slowly depurated.

     Recovery  of  PAH  from  geological  materials  such  as  sediments,  was
inversely  proportional  to time  of exposure.  Internal  standards  added at
the  time of  extraction  did not allow  accurate determination  of extraction
efficiencies.   Drying of  sediments resulted in reduced recovery of anthra-
cene  from  sediment.   Capillary  gas-liquid chromatography coupled  to  mass
spectrooetry  (GC/MS)  showed  that anthracene could be  extracted from sedi-
ments and separated from interfering peaks but anthraquinone (a transforma-
tion product of anthracene) could not.

     Dissolved, naturally occurring organic acids (humic acids) reduced the
availability of anthracene, benzo(a)pyrene and dirccthylbenzanthracene to D.
magna but  increased  the availability of 3-roethylcholanthracene and dibenz-
anthracene.  The presence  of particulates also reduced the availability of
PAH to D. magna.

     The simulation model, Fate of Aromatics Model (FOAM), was developed to
predict the  fates  of PAH from molecular weight  of  the PAH of interest and
parameters of the ecosystem to which it is to be released.  FOAM accurately
simulated  and  predicted the  flux along physical-chemical  pathways.  How-
ever, the  model  was  less accurate in  predicting  the  accumulation and bio-
tmnsformation  of  PAH  by aquatic  biota.   Future  versions  of predictive
simulation models  shoulu  be kinetically  based, os  FOAM is,  if  accurate
information  on  cycling  is to be  predicted.   However, this  type  of model
requires a great amount of information.   Because of the scale of the input
parameters, models such as FOAM  will be useful  in  describing overall pro-
cesses and fluxes  along different pathways but will  not  be very n  ful in
simulating the  concentrations of PAH and  PAH transformation  [not.jcts in
individual  biotic components.  Also,  some of the physical processes, which
are  important   in  determining the  fates  of PAH  in   natural  syntens,  are
discontinuous functions or catastrophic  events,  such as  stonas,  which do
not lend themselves  to  mathematical  simulations.  Therefore, models of the

-------
type developed here will always be limited to relatively gross predictions.
Accurate predictions or concentrations in individual types of organisms may
not be attainable.

     The  channels microcosms  were useful  for  testing the  reliability  of
FOAM to  simulate  the behavior of PAH in a complex ecosystem.  However, the
greatest  utility  of the-channels microcosm was in allowing  studies  to  be
conducted under more natural conditions than allowed by laboratory systems.
From these  studies  we  learned that exposure to sunlight during exposure to
anthracene  was  acutely toxic.  Thus, laboratory studies  do  not accurately
predict  toxicity  observed  under more natural conditions.   While microcosms
may not  be  used as a central monitoring tool, they are useful in studies of
classes  of  compounds and to verify  laboratory  and sinulation conclusions.
Because  of  their  scale, however, microcosms will  never be  useful in vali-
dation of long-term or  global  simulation models.  The test of the benchmark
hypotheses  was  only partly  supported by our  studies.   Future simulations
will  need  to consider  more   complex measures  of chemical  behavior.   The
proposed use  of octanol/ water partitioning coefficients shows promise for
predicting  uptake by organisms but not biotransformation.   Thus, the bench-
mark approach will  probably not  give adequate  predictions  of the dynamics
of organic  compounds in complex environments.

-------
                                  SECTION 3

                               RECOMMENDATIONS

     Future studies  of the fates and effects of  trace  contaminants  should
be conducted  in laboratory, field and  siraulat.lcn modes  concurrently.   All
three  modes of  investigation  added  to  the  overall understanding of  the
dynamics of PAH in aquatic ecosystems.

     Simulation models should be developed to predict  general  behavior of
PAH  in aquatic systems.   That  is,  to determine when the  greatest mass of
PAH  accumulates and where  the most  sensitive  ecosystem  components  are.
Where  physical  and  chemical processes dominate, simulation models will be
able to  accurately predict the overall dynamics of PAH, but tare events of
large  nagnitude will  reduce  the accuracy of  predictions  on a  short-tt;rm
basis.   Optimization of time step duration relative to  sy:.teu>  level  vari-
ability  needs  t.o  be further investigated to  increase the  accuracy  of pre-
dictive  models.

     The current  state of predictive simulation models  will  not allow  the
accurate prediction of concentrations of single P^H in  individual  species
but  can  be useful  in  simulating  overall procc< . es  in  a  gross  manner.
Future simulations  should be kinetic in  nature but  the  relationships used
to predict  rate constants ne~d to be based on structure-activity relation-
ships.

     The channels  nicrocosms used  in this study  were   sufficient to test
some simulation processes and we recommend tne use of this type of a  system
in  future   studies.   However,  microcosms of  this  magnitude  will  not  he-
useful as  screening tools.   The utility of the realism of such systems  was
demonstrated by the  photo-toxir effect observed in the  microcosm, which was
not observed under laboratory conditions.  We suggest  this type of  system
needs  to be used  to validate processes  which are  indicated  by laboratory
and/or simulations  because  they  allow testing in a more  complex  and more
natural  system than that  of  the  laboratory,  while  not   releasing  trace
contaminants to the  biosphere.

-------
                                 SECTION 4.0

                             LABORATORY STUDIES

                                 SECTION A.I

                                INTRODUCTION

     A series  of laboratory  studies  was conducted to determine  rate  con-
stants for uptake,  depuration and biotransformation of PAH in several  iso-
lated  components  of aquatic  systems.   The components studied  were  water,
sand, organic  sediments,  peripbyton,  dapnnia,  chironomids, fish and  clams.
Beside determining  rate constants for  uptake,  depuration  and biotransfor-
nation, these  studies  investigated  the effects of factors  such as tempera-
ture,  presence or absence  of substrate, and concentration of  PAH on  rate
constants.  The  goal  of these studies was  to  supply  independent  estimates
of  rate  constants  and to  determine  the precision with which these  rate
constauts could be measured in the laboratory.   The rate  constants could be
used to parameterize the simulation model as well as to develop statistical
relationships  between  rate  constants  and first principles  of the  compounds
of interest.

     Subsequent  comparison  of rate constants and  state  variables measured
under  controlled  laboratory conditions could  be compared to those observed
in  microcosm  studier  and  those  predicted  from simulation  models.   Such
comparisons can  provide estimates of  relative variability  as well as rela-
tive  importance  of  pathways and interactions to be used  in formulating and
modifying the predictive simulation model.

-------
                                 SECTION 4.2

           ANTHRACENE SORPTION AND DESORPTION BY ORGANIC SEDIMENTS
INTRODUCTION

     When PAH  are  added to aquatic systems  they  rapidly  become associated
with suspended  and bottom sediments (Dunn,  1976; Lee  and Takahashi,  1977;
Huller  and  Bohnke,  1977;  Lee et al.,  1978; Ciddings  et  aJL.,  1978;  Herbes
and Schwa 11, 1978; Teal e_t al_.,  1978; Prahl and Carpenter, 1979; Gearing et
al.,  1980;  Hinga  et al. ,  1980).   A number of  techniques  for extracting
organic  compounds  from solid matrices arc  available  and  different extrac-
tion efficiencies  have been reported for each (Kookc e_t al., 1981).  There-
fore  laboratory studies  were  designed  to  determine the  best extraction
conditions  and quantification  techniques.   Rate  constants  for sorption of
anthracene by organic sediments were also determined.
MATERIALS AND METHODS

Extraction Procedures

     The  first  set of experiments was conducted  to  determine  the best ex-
traction  protocol  (Figure A.2.1).  The methods which  were  common to these
two  experiments are  given  first, with the methods  which were  specific to
each experiment afterward.

     Two  different sediments were used in these studies.  Upper Three Runs
Creek  (UTRC)  sediment is a  fine,  silty sediment  high in organic matter (•>•
20%  by weight).   Steel Creek sediment is  a  coarse,  sand-clay mixture with
little  organic  matter.   Sediment from the two streams was collected with a
glass  beaker  and  sieved through  5 ma opening stainless steel screen into a
stainless steel bucket.     C-anthi_acene  used to spike  the  sediments had a
                                 -1
specific activity of 3.3 mCi'mmol  .  All soxhlet extractions were done for
18 h.  Liquid scintillation counting was done using 12 mis of Research Pro-
ducts International 3d70b  counting cocktail.  Counting time was 10 minutes
per  sample witti  a  Beckman  Model  LS 8100  liquid scintillation  counter.

     High  pressure  liquid  chromatography analyses were  don? on  a  Varian
Model 5000  LC,  which was equipped with a V» -ian 30 cm HCH-10 reverse phase
column operated at  30°C,  and a flow rate of 1 ffil-min  .  Solvents were 35%
acetonitrile  in water  initially, programmed  to  100% acetonitrile  in 40
minutes, held at  100% for 10 minutes, with  a  return to initial conditions
for 5 minutes before the next  injection.   Injections  were done automatic-

-------
                                ILVftKT ]*OO »f 0<«*
-------
ally by a 25 pi sample loop.  The detector was a 254 OJD fixed-wavelength uv
detector, connected  to a  Varian CDS  111-L  integrator, and  a  strip chart
recorder.   A fluorescence  detector with  360 ma  excitation  and >  460 nm
emission wavelength filters, connected to a second strip chart recorder was
operated downstream from the UV detector.  Quantitation was by the internal
standards method with chrysene as the internal standard.

                                               14
     Combustion  of sediment, for  unextracted   C  was done  on  a  Packard
Tricarb  sample  oxidizer.     C-CG   was trapped  in 7  ml cf  a CO  trapping
agent  and  added to  12-15 ml  of counting cocktail.   Prior  to combustion,
sediment  samples,  ^  250  mg each,  were mixed  with  cellulose  powder  and
Combustaid   to improve -combustion efficiency.

     The  first  experiment  was conducted  to  determine the  solvent  system
which  allowed the greatest  recovery  of  anthracene  from  sediment, while
minimizing  the extraction of interfering compounds.  In addition, we wanted
to  determine if  direct spiking and more natural incorporation gave similar
estimates  of extraction efficiency.  The  flow  diagram for this experiment
is  given by  the solid line  in Figure 4.2.1  Three  solvents were  used to
extract  UTRC sediment, which  had been  slurry  spiked.  Three aliquants of
lOg, wet filtered  weight were soxhlet extracted with  each of three solvent
systems  (Table 4.2.1).

     The second  experiment was conducted to compare variation in extraction
efficiency  due to  sediment type, time of equilibration between sediment and
anthracene,  the  effect of sediment moisture on the extraction of anthracene
from  wet or  dry sediment, and  solvent  type  (p'.larity).   The flow diagram
for this  experiment  is  given by  the dashed  liiy? in Figure  4.2.1.   The
experimental design  was a completely randomized 2  factorial with 3 repli-
cations  per treatment combination (Table 4.2.2).  Spiked sediment slurries
were mixed  for 24 h  or  1  wk before extraction.  Wet  sediment samples were
soxhlet  extracted  immediately  after filtration.  Dry sediment was air-dried
in  a  hood for 24  h.  Extraction  time was 18 h with either 100 ml of benzene
or  35  ml of acetonitrile +  65 ml  of benzene.  Crude  extracts were sampled
for quantification of  C activity.

     Anthracene  and  anthraquinone peaks were separated  on an OV101 capil-
lary column and  mass spectra  recorded  to  determine if the compounds which
were being  spearated  and identified were anthracene and anthraquinone. The
relative ion intensities of  the peaks wnich we had  identified as anthracene
and anthraquinone  are  given in Figures 4.2.2  and 4.2.3 respectively.  The
retention  indexes  (RI)  vene  calculated  with  equation  4.2.1.  The  HPLC
protocol  separated   anthracene,  anthraquinone  and   chrysene  well  (Fig.
4.2.4).  However,  sediment extracts often contained a  compound which eluted
very near  Anthraquinone (Fig.  4.2.4).  This  peak was  almost  always present
but often  obscured   the  anthraquinone  in  spiked  samples.   The retention
indices  of  the  anthracene standard and an anthracene  sample  extracted  from
sediment were 1717 and  1712  respectively.  The retention indices of  anthra-
quinone  standard and  raasple were  1807 and 1789  respectively. The retention
indices  indicated  that the peaks we had chosen as  anthracene  and anthraqui-
none in  our chromatograms of  sediment  extracts  were accurate.  However, to
                                     11

-------
TABLE 4.2.1.  EXTRACTION OF 14C-ANTHRACENE FROM DIRECT AND SLURRY SPIKED UTRC SEDIMENT
Fraction Recovered
Treatment
1 Direct Spike
2 Slurry Spike
3 Slurry Spike
4 Slurry Spike
N
3
3
3
3
Solvent System" x
35 ml ACN * 65 ml $H .976
35 ml ACN + 65 ml H .808
35 ml ACN + 65 ml Chex .771
16 ml MEOH + 42 ml ATN + ml CHEX .898
SD CV
.020 2.05
.030 0.37
.015 1.95
.027 3.01
 ACN = Acetonitrile




  $H = Benzene




 ATN = Acetone




CHEX = Cyclohexane




MEOH = Methanol

-------
TABLE 4.2.2    EXPERIMENTAL DESIGN TO DETERMINE EFFECTS OF SEDIMENT TYPE,
               EQUILIBRATION TIME, MOISTURE CONTENT AND POLARITY OF SOLVENT
               ON EXTRACTION EFFICIENCY OF ANTHRACENE FROM SEDIMENT
Main Effects
Sediment Type Spike Time Moisture
Steel Creek 24 hr Vet

Dry

1 wk Vet
Dry

UTRC 24 hr Vet
Dry

1 wk Vet

Dry

Solvent
System
«H
4>H/ACN
4H
$H/ACN
JS/ACN
«H
4>H/ACN
«>H/ACN
.„
(|iH/ACN
<>H
(J>H/ACN
<»H
$H/ACN
     * Benzene, ACN = Acetonitrile.
                                    13

-------
lOO-
ts 80-
0s
*" 60-
H 40-
co
Z 20-
UJ
h- 0-
- 100-
g 80-
i- eo-
-» 40-
UJ
K 20-
n-
ANTHRACENE





1
2



., ,J
T I i i T i i ; ; r \ i i i •
!0 40 60 80 100 120 140 160



!,
             180    200    220    240    260    280

                           MASS  OF  FRAGMENT
300    320
Figure 4.2.2.   Relative  ion  intensities  of  anthracene.  OV101 capillary
              column, 2°C min. increase  in tcmprr.iturc.

-------
   too-
5  80

>  60
t  40
CO
2  2G-
LU
 ~ 100
 UJ
 >  80
             20
                         ANTHRAQUINONE
                    40
60
80
100
120    140
160
            180    200    220    240    260    280   300   320
                        MASS OF  FRAGMENT
Figure 4.2.3.  Relative ion  intrnsitirs of anlhraquinonc.
             column 2°C min.  increase in temperature.
                                                   OV101 capillary
                               15

-------
                 0  4  8 12 16    24   32   40    48   56
                                TIME  (min)
Figure 4.2.4.  Hij;h pressure  lirjrji!  chrcmaloRrams of anthracene nti'l .inthr.i-
               quino/ic  exlr.irtcd from scdimont  with 35/65 ml .icelonitrilc/
               bcnr.cnc.   254  nm UV  dctrctioti.   1 = inLt-rfcrrinR pe.ik, AN =
               anthracene-, A(| =  anthr.iqiiinono.
                                     16

-------
check  the accuracy  of our  determinations  further  we  examined  the  mass
spectra of these peaks.

     The  retention  indices  and matching coefficients were calculated using
equations A.2.1 end 4.2.2.

                      SU - SRI    „„„                     4.2.1
                 RI = SR2 - SRI X 2°°

       where:  RI = Retention Index

            SU = Scan Number of unknown peak
           SRI = Scan Number of first hydrocarbon standard peak
           SR2 = Scan Number of second hydrocarbon standard peak

The matching coefficient of our sample peaks and standard peaks were calcu-
lated  from  the mass spectra of anthracene ana anthraquinone (Figs.  4.2.5 -
4.2.9)  and  equation 4.2.2.   The  matching  coefficients  for anthracene and
anthraquinone were 81.4 and 99.9% respectively.
                              - RISA)/J x 100%               4-2-2
            MC =
                       n
                       Z(RIST + RISA)
                       1=1
       where:  MC = Matching Coefficient
            RIST  = Relative intensity of each mass fragment in the standard
                    peak
            RISA  =  Relative  intensity of each mass fragment in the sample
                    peak
                l - ith ion fragment
                n = Number of confirming ions

     A  laboratory study was  performed to determine the  rate constants for
uptake  and depuration  from  a  quiescent sediment.  Upper  Three Runs Creek
sediments  were .placed in crucibles with a volume of 4.8 ml and top surface
area of  4.2 cm .   Crucibles  were placed  in  a flow through exposure system
and  exposed  to  35 pg'£   of    C-anthracene.  The  flow  rate  through the
system  was adjusted  so  that  the concentration of anthracene  in  the water
was not  significantly depleted  due to uptake  by  the  sediment.   The uptake
rate constant  was  calculated  by assuming first order  kinetics  and initial
rates conditions.
                                    17

-------
       ANTHRACENE  STANDARD IN BENZENE
 CO
UJ 100

Z  50^
—   0
Z
o

UJ
 UJ
 o:
           100% = 10376
           100% = 18320
           100% =7172
           100% =7948
           100% = 13160
A
           100% =1143803
                                               177
           100% = 17664
           100% =119616
                                               179
                                               178
                                               176
                                               152
                                               151
                                               89
/\
T"
      i	1	]	1	1	—,	,	,	
     950    1000   1050    1100   1150   1200   1250   1300   1350
                            SCAN NUMBER
Figure 4.2.5.  Rcl.'it ivo ion  intensities of 7 mass  fr.i>;mcnts of anthracene
             in  benzene.
                               18

-------
        ANTHRACENE IN COMBINED SEDIMENT EXTRACTS
 cn
 z:
 LU
 \-
o

LJ
 UJ
 ce
100
 50 ^
  0
       100% =13600
          f
       100% = 81088
       100%= 7724
     100% = 13288
       100% =
     100% = 6768
       y\
          100% =444928
                               s^ ^
     100% =6283264
                                                     179

                                                     178
                                                     177
                                                    176
                                                     152
                                                     151
                                                    Til
     900     9oO    1000   1050    1100    1150    1200   1250
                          SCAN NUMBER
Figure 4.2.6.  Relation ion  intensity of anthracene  in sediment  extracts,
                              19

-------
        ANTHRAQUINONE STANDARD  IN BENZENE
1250
1300    1350
1400    1450     1500
SCAN  NUMBER
1550
                                                                 1600
Figure 4.2.7.  RolaLivc  ion  intriisi tirs of ni.iss  fragments  of ant Iir.'ujui nono
             staiid.iril  in brnzcuc.
                                20

-------
       ANTHRAOUINONE  IN COMBINED SEDIMENT EXTRACTS
LU
o:
__— 	 \_ 	 --v _ _s*~ \^_ . 	 , 	
100%
= 6340508 Til ^
                      1300    1350    1400
                            SCAN NUMBER
1450     1500    1550
Figure ft. 2. R   RrKitivo  io:i  inLcn.siLicr  of .nnss  f r.ij-.int'iiL:;  from •intlir.T
                     ii) sediment extract:;.
                                21

-------
       ANTHRAQUINONE  IN  COMBINED  SEDIMENT EXTRACTS
   100
 CO
 z
 UJ
 2
 o

 LJ
 UJ
 a:
50-
              ;	1	1	1	1	1	1	r
     1050    1100   1150   1200   1250   1300   1350   1400   1450
                           SCAN NUMBER
Figure 4.2.9.  Relative  ion  intensities  of  mass  fragments  fiom  antlirj-
             quiuonc in sediment extracts.
                                22

-------
RESULTS AND DISCUSSION

Solvent Systems
                                                                       *
                     1A
     The recovery of   C-anthracene added to the samples immediately before
extraction did  not  give the same extraction efficiency from samples, which
had been spiked as a slurry (Table A.2.1). Thus, spikes added just prior to
sediment  extraction are  not  useful  in  determining  the efficiency  of  an
extraction procedure for  extracting anthracene from a particular sediment.

     The greatest extraction efficiency of "naturally incorporated" anthra-
cene was  by  the most polar solvent system, which was methanol, acetone and
cyclohexane (Table A.2.1).  However, this solvent mixture extracted much of
the  polyphenolic humic  substances  from  the  sediment  which could  not  be
easily  separated from anthracene before  analysis.  The  mixture of aceton-
trile and  cyclohexane  yielded anthracene in two solvent phases and was not
practical  when  all  of  the relative  merits  of  the solvent  systems were
considered.   We adjudged  the acetontrile-benzene solvent mixture to be the
most  practical extraction  system.   The  recovery of  anthracene,  using the
•cetonitrile-benzene mixture was only slightly less than that with the more
polar mixture but the selectivity for anthracene was greater.

     The  results of the  factorially designed  experiment  are  presented  as
mean  recovery and standard deviations of the  factorial main effects (Table
4.2.3)  and analysis of variance (Table  A.2.A).   All of  the  main effects
were  statistically  significant.   This indicates that  all of  the factors
examined  significantly affect  the  efficiency of extraction  of anthracene
from  sediment.  The two-way  interaction terms  between  equilibration time
and  moisture  content and equilibration time and solvent polarity were also
significant so that  the exact magnitude of the main effects of these treat-
ments  can not be interpreted directly.  The three-way interaction terra was
also  statistically  significant. Thus,  any discussion of the two-way inter-
actions must  be guarded.   The  two-way interaction between moisture content
and  solvent  polarity  was  highly significant  for Steel  Creek  sediment but
not  statistically  significant  for the  more organic  sediment  from UTRC
(Table  A.2.A).

     Similar  generalizations  could  be made  for both   the  sediment types
studied.   The  longer  anthracene  is  equilibrated  with  sediments,  in  an
aqueous slurry,  the  lower the recovery (Table  A.2.3).  This is probably due
to  partitioning  into  clay IntXice  spaces and  sediment organic matter.

     Incomplete extraction of  PAH  fro.n  sediments  has  been  a continuing
problem and has  affected the qualitative as well as quantitative analytical
results (Gearing et al. ,  1978).  Lake £t  al.  (1980)  reported that because
of  varying  extraction efficiencies   clos
-------
TABLE 4.2.3.   EFFECT OF TIME OF EQUILIBRATION, MOISTURE CONTENT, SEDIMENT
               TYPE AND SOLVENT POLARITY ON EXTRACTION OF   C-ANTHRACENE
               FROM STEEL CREEK AND UTRC SEDIMENT
Treatment >C
24hr,
24hr,
24hr,
24hr,
Iwk,
Iwk,
Iwk,
Iwk,
Main
Wet,
Wet,
Dry,
Dry,
Wet,
Wet,
Dry,
Dry,
*H
(>H/ACN
4>H
(|iH/ACN
H
4>H/ACN
_a
X
.9?!
.879
.454
.707
.315
.547
.455
.533
Steel Creek
SD
.037
.024
.047
.026
.017
.032
.003
.010
3
2
10
3
5
5
0
1
CV
.83
.72
.4
.73
.23
.77
.58
.82
_a
X
.796
.783
.819
.860
.253
.470
.719
.741
UTRC
SD
.040
.061
.026
.012
.130
.035
.038
.004
CV
4.
7.
3.
1.
51.
7.
5.
0.
97
81
12
34
5
49
31
59
Effects
Spike time-24hr


Moisture-


Solvent-
System
Iwk
Wet
Dry
H
H  =  Benzene, ACN  = Acetonitrile
                                    24

-------
             TABLE A.2.4.   ANOVA OF EXPERIMENTAL RESULTS  GIVEN  IN TABLE 4.2.1
ro
Source
ACSpike Time)
B(Moisture)
C(Solvent System)
A x B
A x C
B x C
AxBxC
Error
Total
DF
1
1
1
1
1
1
1
16
23
SS
.5049
.1192
.0832
.2452
.0075
.0138
.0980
.0110
1.0830
Steel Creek
MS F
.5049 695. 45a
.1192 164. 19a
.0832 114.60*
.2452 337. 74a
.0075 10.33a
.0138 19.018
.0980 134.99°
.0007

SS
.4334
.2623
.0269
.1525
.0166
.0074
.0235
.0520
.9746
UTRC
MS F
.4334 133. 75a
.2623 81.003
.0269 8.30a
.1525 47.07a
.0166 5.12b
.0074 2.29 NS
.0235 7.25b
.0032

              Significant at P < 0.01




             Significant at P < O.fj




             NS = Not significant at P < .05

-------
     PAH  are generally  extracted from wet  sediments (Windsor  and Kites,
1979) because  it has  been observed that drying  sediments  reduces  the. re-
covery.   This effect  may  be  due  to volitalization  or  to  other  physical
processes which  cause  the PAH to bind more  tightly to sediments upon dry-
ing.  In this  study  we  observed greater extraction  efficiercies  from wet
Steel Creek sediments and dry UTKC sediments.  This nay be due, in part, to
the  different  organic contents  of the two  sediment  types.   This  accounts
for  the equilibration time-moisture  interaction term.  After  being equi-
librated  for a  longer  period  of  time  the  extraction  efficiency  was. no
longei  affected by  sediment  moisture because  of migration  of  anthracene
into organic and donon spaces.  The benzene-acetonitrile solvent system was
significantly  better  at recovering  anthracene  from both  sediment types.

     From the mass fragment spectrum for anthracene (Fig. 4.2.3) we select-
ed  masses 179,  178,  177,  176,  152, 151  and 89  and  monitored the  relative
ion  intensity of a standard in benzene (Fig. 4.2.5) and determined chat our
standard was pure.   We  then monitored  the relative  ion  intensities  of
masses  179,  178, 177,  176, 152,  151 and 57 in a sediment extract.  The fact
that the onset of all of  these peaks was in the same scan number indicates
that all of the material which eluted  in  this  peak was  anthracene.   An
analysis of  our anthraquiaone standard in benzene  was done for mass frag-
ments  209,  208,  188,  152, 151,  150 and  76 (Fig. 4.2.7).   This  analysis
showed  that  our anthraquinone standard was  also  pure.   A similar  analysis
of  mass fragments 209, 180, 152 and 57 (Fig. 4.2.8) and mass fragments 209,
208 and  160  (Fig.  4.2.9) in sediment extracts  in benzene demonstrated an
impurity in  the peak  in  our  samples which we had tentatively identified as
anthraquinone.   The  209  mass  fragment peak in scan 1390 is not in the same
scan as the  208  aau  160  mass  fragments, which have peaks in scan 1375 (Fig.
4.2.9).   This  information indicates that there is a unknown compound which
also has 208  and  160 mass fragments, as  does  anthraquinone,  which elutes
slightly sooner  than those from anthraquinone.  This compound does not seem
to  have a very intense  209  mass fragment.  Therefore, we did  not report
anthraquinone  recoveries from our sediment.  Maher e_t a\. (1978) also found
colored organicE were  extracted  from sediments which interfered with subse-
quent  quantification of  PAH  compounds.  In a study of extracts free marine
sediments, Overton et  al.  (1977)  found that capillary column chromatography
was  able  to distinguish  between indigenous hydrocarbons  of contemporary
origin  and  those known  to be associated with  fossil hydrocarbon pollution
of  marine sediments.
First-Order Rate Constants

     The  first order  rate constant  for sorption was reported on both a mass
and  area  of  sediment basis.   Based cm mass of  sediment  the rate constant
for  uptake was 0.04  min  ,  assuming a density  of  approximately  1  for the
cilty  sediment  used in this study.  While the sediment density is slightly
greater than  1,  this  will not significantly affect the rate constant.  The
first  order  rate  constant for sjorptionj. based on surface area of sediment
was  found to  be 0.009 ml  •  cm   •  min  .   These values  relate to the rate
                                    26

-------
of sorption  as well as  diffusion  into the quiescent sediment and are  not
the rate  constants  which would be derived for sorption  in completely mixed
systems.
                                    27

-------
                                 SECTION 4.3

        EFFECT  OF  HUHIC  ACIDS  ON  BIOAVAILABILITY  AND  TRANSPORT  OF  PAH
INTRODUCTION

     Evidence  suggests  that the most  important pathway  of bioaccumulation
of  PAH is  direct uptake  from  water  by  simple  partitioning  into  lipids
(Neeley, et al., 1974).  It has been demonstrated  that humics in freshwater
bystems may  form  stable complexes affecting the transport of trace organic
compounds  including  pesticides  (Wershaw,  et  al.,  1969) and  cholesterol
(Hassett and  Anderson,  1979).   In marine systems, dissolved organic matter
(DOM)  reportedly  increased the  solubility of n-alkanes  (Boehm and Quinn,
1974) and decreased the uptake ot aromatic and other petroleum hydrocarbons
by the vivalve Mercenaria mercenaria (Boehm and Quinn, 1976).  It is gener-
ally  believed that DOM in rivers is precipitated on  exposure  to  seawater
and  deposited  in  estuarine and nearshore  marine  sediments  (Gardner  and
Kenzel,  1974; Hedges and  Parker,  1976).  The potential  for deposition of
trace contaminants in estuarine  sediments is largely unknown.

     The effects of humics on direct uptake of PAH by Daphnia magna and  the
potential  for salting-out or  co-precipitation of PAH-humics  at estuarine
salinities  were determined  in  laboratory  studies.   Studies on humics  as
they  affect  PAH accumulation  in D^ magna were conducted to detinaint ]) if
a  "standard"  htunic  acid  (Aldrich  )  at  one  concentration  (2  mg'£   DOC)
affected the  bioaccumulation of  several PAH, 2) if observed effects in some
PAH  varied  predictably with  humic  concentration,  3) if  observed  effects
changed over a range of PAH concentrations, some  of  which exceeded limits
of water  solubility,  and 4) if  humics in natural  waters behaved similarly.
In addition,  the potential for salting-out of PAH-humics was determined  for
six  PAH at  one  concentration   of  Aldrich    humics   in  20  /oo artificial
seawater.
METHODS AND MATERIALS

     Benzo(a) pyrene,  napthalene,  anthracene,  1,  2, 5, 6  -  dibenz anthra-
cene,  dimethyl  benz(a)anthracene  and  3-raethylcholanthrene   (Table  4.3.1)
were studied.   Stocks  of PAH were diluted in solvents as shipped and added
to water  in volumes  of 1-30  pl'£   .   The  H  -  dibenzanthracene specific
activity was adjusted  to 1027 mCi'mmol   by addition  of  uiilabeled dibenz-
anthracene  (Aldrich, 97% pure) to  water at the time of labeling. Water was
labeled in bulk (1-2 £) and thoroughly shaken for one to two  rain.
                                    28

-------
           Table 4.3.1.  Radio-labeled PAH used  in bioavailability  studies.
to
vO
Compound Label Specific
Activity
benzo(a)pyrpne 7,10-14C 21.7
naptbalene 1-'*C 8.8
14
anthracene 9- C 3.3
1,2,5,6-
dibenzantbracene H(G) 30.3x10
7,12-diusethyl dimethyl 14C 97.4
benzo (a) anthracene
3-methylcholanthrene 6- C 57.0
Solvent Radiopurity Source
toluene > 96 Amersham
Batch 29
ether > 98 California
bionuclear
lot 2545
acetone > 98 California
bionuclear
lot 770824
benzene > 97 New England
Nuclear
lot 1253-021
benzene > 98 New England
Nuclear
lot 1267-176
benzene > 98 New England
ethanol 9:1 Nuclear
lot 1231-171

-------
     Test water was  prepared  by adding inorganic salts  to  deionized water
(Milli-Q ) (Table A.3.2).
      Table 4.3.2.  Inorganic salts added to test water and
                    resulting water quality parameters.
Added Salt
NAHCO,
3
CaSO • 2H 0
MgS04
KC1
mg-£
48

30
30
2.0
                     Water Quality Parameters

                     pH                   7

                     hardness             20 mg-£   as CaCO_

                     Alkalinity          120 mg'£~  as CaCO
                                                     _2
                     Conductivity        130 (Jmhos-cm
     A humic stock solution was prepared from Aldrich humic acid (H-1675-2,
Lot  #082091)  as  follows:   1)   humics dissolved in 0.1 N  NaOH;  2)  centri-
fuged  at 13,000  x  g  for  30  min;  3)  pH  adjusted  to 2 with 6.0 N  HC1 and
allowed  to stand for  18  h;  4)  humic acid  precipitate  was  collected  by
centrifugation  (4000  x g for 20 min).  The above steps were repeated three
times, the  final  humic solution adjusted  to  pH 7,  and dialyzed repeatedly
against  deionized water until  conductivity was <  10 (jmho.   Stock humics
were added  to  labeled test water- to  a final, nominal concentration of 2.0
mg  • £    DOC.   Total carbon  and  inorganic  carbon  were determined  by a
Beckman  Model  915B  Carbon  Analyzer and DOC was calculated by  differences.

     Daphnia magna  from laboratory  cultures  (Athens Environmental Research
Laboratory, U.S.  EPA,  Athens,  GA.)  were exposed to PAH in paired one liter
beakers  with  and without   humics   (60-70  Daphnia  £   );  for six h.   Five
sanples  of  ten  animals v/ere removed by pipet from each test water,  collec-
ted on nylon  netting,  rinsed quickly  in  50 ml deionitr.ed  water, and poured
through  a Millipore  filter apparatus.  Animals were  collected  on  tared 25
IMI Type  HA  cellulose  acetate filters, dried over desiccant for 24-48 h and
sample dry weights determined on a Cahn Model 4700 Electrobalance.   Animals
                                    30

-------
                                                                       14
and  filters  were combusted  in a Packard Model  306 sample oxidizer,   CO.
and   H.O collected  in  suitable trapping  agents,  and  counted  by liquid
scintillation counting.

     To  evaluate the  importance of natural  particulates  and DOC on  B(a)P
bioavailability,  similar experiments were conducted using water  from  Upper
Three  Runs  Cree1'  (UTRC) and  Skiuface  Pond on  the Department of Energy's
Savannah River Plant,  near Aiken, SC.   B(a)P  uptake was determined in  water
directly from the creek,  in water filtered  through  Whatman GFC precombustcd
(550   C,  8  h) glass fiber filters  (0.45 pm)  to  remove particulates, and in
filtered  water UV photo-oxidized  (f.aJolla Scientific  Co.)  to remove DOC.
Salting-out of PAH  - Humic Complexes:

     Each  of  the six PAH was  dissolved  in 0.5 L of water in paired beakers
containing Milli-Q water with humics  (8.5  mg-£    DOC) and without huwics
(0.2  rag   •  £    DOC).    Initial  PAH  concentrations  were  determined  from
triplicate assays of activity  in  1 ml  water  samples.  A nominal salinity of
20  o/oo was obtained by  adding sea  salts  (Instant Ocean, Aquarium Systems,
East  Lake, OH.)  to each  beaker  and PAH in  solution  after  24  h was deter-
mined as above.
 RESULTS  AND DISCUSSION

 Bioconcentration in Daphnia  magna

      Concentrations of five PAH >fere  f*.eterminea  in  Daphnia jnagna after 6 h
 exposures  in water with Aldrich   humics  (DOC =  2.0 tng'£  )  and in water
 without  hiusics  (DOC £ 0.2  mg-£   ).   Daphnia  in  huroic waters had signifi-
 cantly lower concentrations of  benzo(a)pyrene (BaP) (-24.5%) than compar-
 able animals in non-humic waters  (Table 4.3.3).   On the  other hand, humics
 increased   daphnia  concentrations  of  methylcholanthrene   (MC)  (+210%).
 Huraics  had  little  effect   on  dimethylbenzanthracene  (-13.1%)  or  dibenz-
 anthracene (-t-19.1%), and while  the effect on anthracene  (-46.7%) appeared
 large, it was  not  significant due to large sample variance.  Results re-
 ported assume  that all radioactivity  was parent compound.  These results
 are  significant  for two reasons:  1)  they  demonstrate  that  hiunics  at low
 concentrations  affect  bioaccumulation  of some  PAH;  2)  they demonstrate that
 huaiics may either  increase or  decrease bioaccumulation.

      Evidence from this1, and  other  studies  suggest  that  results reported for
 the  6h uptake tests were not  transient differences  but were  representative
 of steady  state  differences  for  all PAH tested.   In  preliminary 24 h uptake
 studies,  steady state  radioactivity in D. magna was reached within 6 h for
 naphthalene,  anthracene: and  Bap.   Approximately 60-80% of steady state was
 reached  after  6 h  with  DMBA, DBA and MC.    In  no  case  did  the relative
 difference between daphnia PAH in  humic and  non-humic waters  change after 6
 h.   In Daphnia  pulex,  Southworth,  e_t  a 1. ,  (1978) reported that 6-h expo-
 sures  produced   steady  state concentrations  of naph uaiene and anthracene,
 while  up to 24  h  were  required  for benzanthracene,  results consistent with
                                     31

-------
                               fit
Table A.3.3.  Effect of Aldrich  humic acids on PAH accumulation by D.  magna.
PAH
Anthracene
row 178.2
Benzo(a)pyrene
aw 252.3
Dimethylbenz-
anthracene
mw 256.3
Methyl
cholanthrene
mw 268.3
Dibenzanthracene
mw 278.4
Bioaccumulation:
Bioconcentration
a b
PAH Concentration Humic Concentration Bioaccumulation Bioconcentration
(Mg-A~ ) (DOC in mg-£ ) (6 h) Factor
1.96 0.2
2.0
1.15 0.2
2.0
0.33 0.2
2.0
1.29 0.2
2.0
0.73 0.2
2.0
(nmoles PAH • g dry wt. D. ma;
factor: ng PAH • g" wet wt D.
82.3 (19.7)
A3. 9 (4.0)
131.6 (7.3)
99.4 (7.7)
64.2 (23.2)
55.8 (30.0)
34.0 (7.4)
105.5 (10.8)
20.3 (0.8)
24.2 (2.4)
;na; Dry wt/Wet wt Ratio = 0.06)
magna
607
319
2745
2158
968
666
667
2064
652
773


                                    ng PAH  • ml
                                               -1
C X (SE) n - 5.

-------

10
0
X
3
w.
CT
O>
c
X
D.

50
40

30

20

10
°c
a-Benzo(o)pyrene BCF = I7I6
O-Anthrocene BCF = 389
o
o
o
-
o
o
0

'ee ° ° ° °
! 1 1 1 1 1 1 1 1
) 1 23456
HOURS
Figure 4.3.1.
Accumulation  of  Ii(.i)P and  antliracene by D. m.igna as a  func-
tion of time.
                                      33

-------
our observations.  Previous studies have shown that daphnia have relatively
slow PAH biotransformation rates relative to accumulation rates (Herbes and
Risi, 1978;  Leversee,  et aK , 1981).   In  this  study with   C-BaP, after 6
h, parent  compound  represented 91.3 + 2.1% of test water activity and 92.8
+  0.8% of  D.  magna  activity  respectively.   Active  biotransfonnation and
excretion of PAH reported for some invertebrates and fish are probably not
important factors in D. magna PAH accumulation.
                           '          £\
     The  effect  of varying  Aldrich   humic  acid  concentrations  (0.3-5.7
mg*£   DOC) on daphnia accumulation of BaP, anthracene, and naphthalene was
determined.   Humics  significantly  reduced  the  accumulation  of BaP and
anthracene,  but  did not affect naphthalene (Table  4.3.4).   In this  simple
unsubstitutcd  PAH  series,  the  effect  of  hiunics  was  greatest for  BaP  >
anthracene  > naphthalene, which  is consistent with  the  order of reported
octanol-water  partition coefficients.  This suggests  that  hydrophobic PAH
may  be partitioned  competitively  between daphnia  lipids  and  humics.  Al-
though  the  effect of humics on  BaP and anthracene was greatest at highest
humic  concentrations,  there  was not a simple linear relation between huaic
concentration and daphnia PAH concentration.  The. greatest effect was found
between no  humic (0.2 mg-2   DOC)  and  1-2  mg'2~   DOC.  Additional data at
several concentrations  of PAH and  humics will be necessary before dismiss-
ing  the possibility  that predictive sorption isotherms can be developed for
PAH  like BaP and anthracene, but these data suggest that such a possibility
is unlikely.
                                    34

-------
^_                              wV
Table 4.3.4.  Effect of Aldrich  hukic acids on PAH accumulation by D.
     PAH
Huinic Concentration
                      (DOC in

Bioaccumulation
      (6h)
Bioconcerit ration
     Factor
Benzo(a)pyrene
mw 252



Anthracene
mw 178



Naphthalene
mw 128


0.3
1.5
5.7


0.3
1.5
5.7


0.3
1.5
5.7

139.2 (14.2)
90.9 (7.6)
72.6 (5.7)
F ,,=12.19 (P<.005)
2,12
37.2 (2.0)
34.9 (2.0)
31.0 (1.1)
F .,=3.16 (P<0.1)
2,12
1.7 (0.2)
1.6 (0.?.)
1.8 (0.1)
F, ,=0.95 (NS)
2,6
1716
579
8?8
,

389
362
340


61
74
57

 a                                -1                   -
 Bioaccumulation:   n moles  PAH-g  dry wt D. magna.  X (SE), n=5.  Naphthalene
 n=3.

 Bioconcentration  Factor:   ng PAH'g   wet wt D. magna
                                         -1
                                ng PAH'inl   water

  Initial  B(a)P water  concentration was  1.5 Mg'£   (6.0 mM); Anthracene was
  7.2 aM;  Naphthalene  was 6.6 mM.
 Salting - out

      An  initial  study  was  conducted  with BaP  to determine the  range of
 salinities  producing a  salting-out  effect and  the  time  to  steady state
 water  concentrations.   Initial  concentration  of Aldrich   huraic  acid  was
 8-10 mg-£~  DOC in all studies.  At 17-?5°/oo salinity, 70% of BaP was lost
 from the water column aftei  24 h (Fig. 4.3.2).  Below 5 /oo, no salting out
 was  apparent.   In beakers  where significant salting  out occurred,  a fine
 organic  floe  was  evident  on  bottom  and  sides  of  the  beaker.   PAH mass
 balance was  greater  than  95% in all  cases,  indicating that unaccountable
 losses were minimal.

      Co-precipitation  of  all  six  PAH  studied  was  determined  at  20°/oo
 salinity after 24 h.  In the presence of humic acids.  PAH concentrations at
 24 h  were reduced  significantly (5-79%) below initial values  for all PAH
                                     35

-------
o

h-

cc
r-
2

o
Z
O
o

a.
 o
m
u
o
a:
LJ
a.
   120
    100
    80
    60
r-  40
    20
                 SALTING  OUT OF BaP HUMICS
Initial Humic = 8.5mg/L  DOC

Initial BaP  = 1.2/J.g/L
\2         18

  HOURS
                                    SALINITY (ppt)


                                            0
                                                    25.5
                                       24
                                                             48
Figure A.3.2.  Salting-out of B.iP  at  six  salinities  in  thn presence of

            Aldrirh humics.
                              36

-------
 (Table 4.3.6).   la  water  without humics, salting out was found only for MC
 (-12.5%) and DBA (-15.2%).  Mass balances were all in excels of 93%, and no
 losses were observed  from control beakers (PAH in Milli-Q  water).  DOC at
 24 h was not determined.
Table 4.3.5.  Effect of Skinface Pond and Upper Three Runs Creek organics
              and particulates on B(a)P accumulation by Daphnia magna.


Water Source and     Organic Carbon    Bioaccumulation     Bioconcentration
    Treatment          (mg • 8,' )          (6 hour)              Factor
Upper Three Runs
Creek
Untreated
Filtered (O-ASpm)
Filtered-oxidized
Skinface Pond
Untreated
Filtered (O.ASpm)
Filtered-oxidized


10.0
5.5
0.2

12.2
—
0.2


227.4
304.4
367.2

70.4
156.2
206.7


(12.0)C
(7.1)
(13.9)

(2.6)d
(5.6)
(13.1)


2571
3581
4656

903
2312
3292
 Mean  (SE),  n=5.  Values = runole  • g   dry weight

 mg  BoP-g    wet weight Daphnia

       ng BaP-ml   water

 °ANOVA:  F     =  37.87 p < .001.
          i, \.i

 dANOVA:  F.  .  =  67.84 p < .001.
                                     37

-------
Table 4.3.6.  Salting - out of PAH in water with and without hunics  .
                                             PAH (Hg'*')      PAH
       PAH                  Time (H)       Without Humics      With HumicsC
Naphthalene
mw 128.2
Anthracene
nw 178.2
Benzo(a)pyrene
mw 252.3
Dimethylbenzanthracene
mw 256.3
Methylcholanthrene
mw 268.3
Dibenzanthracene
mw 278.4
T =
T =
T =
T =
T =
T =
T =
T =
T =
T =
T =
T =
0
24
0
24
0
24
0
24
0
24
0
24
2
o
*»
2
2
1
1
0
0
1
0
0
0
.10
.13
.81
.90
.21
.18
.91
.96
.04
.91
.72
.61
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
(0
.02)d
.16)
.01)
.05)
.04)
.05)
.01)
.03)
.04)
.10)
.01)
.01)
2
1
2
2
1
0
0
0
1
0
0
0
.21
.56
.81
.55
.26
.39
.78
.39
.02
.25
.72
.15
(0
(0
.07)
.04)
(0.10)
CO. 04)
(0
(0
(0
(0
(0
(0
(0
(0
.01)
.03)
.05)
.01)
.04)
.00)
.02)
.00)
  Salinity in all cases  20o/oo,  Instant Ocean Salts.

  Water without huuics:   Dissolved  Organic Carbon = 0.2 mg-£   at T = 0.

 CWater with Aldrich  humic acids:   DOC = 10.0 mg-2"1 at T = 0.

      values = X (SE), N = 3.
      While the importance of PAH  sorption  to suspended mineral and en ganic
 particulates  has been  recognized  (Karickhoff,  ejt  al. ,  1979),  the role of
 dissolved  organic matter  has  received  little attention.  Hassett and Ander-
 son   (1979)  reported that  DOM  in river  water  reduced the  efficiency for
 solvent  extraction  of  cholesterol.  Boehm  and Quinn  (1976)  found that
 dissolved  organic matter  (DOM)  in  sea  water  increased the  solubility of
 n-alkanes  and isoprenoid hydrocarbons, but had no  effect on the solubility
 of  the PAH phenanthrene and  anthracene.  Landrum and Giesy (1981) reported
 that  sorption of B(a)P  to  XAD-4 resins is reduced by  the  presence of hu-
 mics.
     The  importance of humics on  the environmental  fate of hydrophobic PAH
like B(a)P is clear from  this study.  Our  results show that B(a)P bioaccu-
taulation  is  significantly reduced by  huraic concentrations of  2.0 mg'£
DOC.   Downstream transport  of  hydrophobic  PAH  may  be increased by humics


                                    38

-------
depending on particular*"  load,  character of humics, and partitioning among
these  compartments.   Precipitation  of  humic  PAH complexes  by  increasing
salinity, as reported  here for B(a)P, may  represent  a  significant pathvay
for  accumulation of PAH  in estuarine sediments.   Thus,  this work  demon-
strates that fresh water humics may significantly affect results  of studies
on PAH bioavailability and environmental transport.
                                    39

-------
                                 SECTION 4.4

            UPTAKE, DEPURATION AND BIOTRANSFORMATION KINETICS OF

           BENZO(A)PYRENE AND ANTHRACENE BY PERIPHYTON COMMUNITIES
INTRODUCTION

     Algae are  the  foundation of most aquatic food webs,  thus,  changes in
the size,  structure  or metabolic activity of the  algal  community may have
important  effects on  the  structure  and  function of  the entire  aquatic
community.   While many stream systems are not autotropbic,  the  periphyton
community  may  be important  to  critical  species  or  at  critical  periods
during  the growing  season.   Also,  the periphyton community can  be  an im-
portant accumulator  of organic compounds.  Thus,  periphyton communities may
be  an important  component  of simulation models by removing  organics from
solution,  biotransforming trace organics or acting, directly or indirectly,
as  a  source  of these compounds to foraging insects and other components of
the food web, directly or indirectly.

     The effects  of  hydrocarbons on algae have been  investigated (Soto et
a_l.,  1975;  Schindler  et  al.,  1975;  Soto et  al.,   1979a  and  1979b;  and
Giddings,  1979} but few  studies have been made of the  uptake,  depuration
and  biotransfonnation,  which are  needed  for  dynamic  simulation  models
(Payer and Soeder, 1975; Walsh et al., 1977).

     Evidence  suggests  that  the most  important pathway  of bioaccumulation
by  periphyton  is  direct uptake from water by simple  partitioning of B(a)P
into  lipids  (Neeley, et al., 1974).  Elimination  from biota  includes par-
titioning  of parent  compound into clean water and active biotransformation
and  excretion  of metabolites   (Neff,  1979).   Reported  biotransfonnation
potential  is great in midge larvae (Leversee, et. al.,  1981a) and fish (Lee,
et  al.,  1972b;  Bend  et «!.,  1979), and is ninor in Daphnia (Southworth, et
al.,  1978).   The potential  for food chain  bioaccumulation is  still not
resolved.   It  has  been  shown  that  there  is significant  accumulation of
naphthalene  by Chlamydomonas angulosa  with no apparent bictransformation
(Soto,  et  al., 1975).   In marine microcosms, Lee, ejt al.  (1978) reported
accumulation, and no apparent biotransformation,  of B(a)P and other PAH by
phytoplankton.  Accumulation of  B(a)P from algal  food  by  hardshell clam
larvae is  reported  to  equal  that of direct uptake from water (Dobroski and
Epifanio,  1980).   In  freshwater  lotic  systems,  periphyton  represents an
important  primary producer  for  food chains and may serve as the functional
organic substrate for PAH accumulation.

-------
     A  series of  studies was  conducted,  under  laboratory  conditions, to
determine the  rates  of uptake, depuration and biotransformation of the PAH
compounds, benzo(a)pyrene and anthracene by periphyton communities on glass
slides.   These studies were  designed to  provide information not  only on
rate constants and biotransformation but also on  community seasonal differ-
ences, as well as length of colonization time.
MATERIALS AND METHODS

Periphyton Communities

     Natural periphyton communities were collected from two streams on the
U.  S.  Department of Energy's  Savannah  River  Plant near Aiken, South Caro-
lina.  Upper Three  Runs Creek  (UTRC)  is a deep,  fast-flowing,  blackwater
stream  in  which diatoms comprised  approximately 90% of the winter periphy-
ton community.  Eunotia  incisa  and Eunotia sudetica were the dominant forms.
The water  quality of  UTRC  is  given  in  Giesy and Briese,  1978.   Castor
Creek,  by comparison, is a  shallow, partially beaverdam-impounded system in
which  the periphyton community was more  diverse.   Species of filamentous
and non-filamentous desmids  from  the  genera  (Hyalotheca,  Desmidium, Spon-
dylosium,  Netrium, Euastrum and Pleurotaenium) accounted for approximately
50% of the Castor  Creek attached  flora.   Various  diatom  species, particu-
larly  Anomoeoneis  serians  var.  Apiculata and Frustulia  rhomboides,  were
representative  of  the remainder  of  the  community.


Exposure

     Naturally  colonized glass microscope slides (7.62 x 2.54 cm) in plexi-
glass  slide holders  were  harvested after  3  or 5  weeks  colonization from
UTRC  and after 3 or 6 weeks  colonization  frora Castor Creek,  beginning in
January of  1980.  .These  slides,  .with  their  attached  communities,  were
exposed  to 7,10     C;Benzo(a)pyre^'i   (B(a)P)  (Amersham-Searle,  specific
activity  21.7 mCi-mmol   )  in  a labci».-•story flow-through system.  The anthra-
cene  used  in  these studies  was !5,    C  labelled  (Cal.  Biochem., specific
activity  J.3 mCi-mmol"  , Lot  #770624'-.   Millipore  filtered  (0.45 \m) stream
water  was  labelled  in bulk  witb    C-EaP approximately  one hour  before
exposure  was  begun.  Labelled water was  gravity  fed  from a 40£ glass head
tank,  via  stainless steel  tubing  and micrometer valves, tp each of six 250
ml  glass  staining dishes.  Flow rates  averaged 100 ml-h    and were suffi-
cient  to maintain  a constant  B(a)P  water concentration of 1.0 pg'£   .  They
did not simulate stream flow  conditions.   Six  slides,  supported in stain-
less  steel  wire mesh baskets,  were placed  in each staining dish.  Experi-
ments  were  performed under gold fluorescent  light  (X 5^ 500 nm) to minimize
photolysis  of   B(a)P  anthracene.  Uptake,  biotransformation and autoradio-
graphic studies were  conducted simultaneously.

-------
Uptake and Depuration Rate Constants

     Slides used in the determination of uptake rates were removed from the
flow-through system after 0.25, 1, 2, A, 8 and 24 h of exposure.  One slide
from  each of  three  reservoirs was  harvested at each of  the six sampling
times. After  removal  from the dosing system, each slide was allowed to air
dry  and  the  periphyton  were  then  scraped  into a  tared  paper thimble.
Samples were  dried in a dessicator  for 24 h  and weighed on a Mettler Model
HL52  Balance. , They  were  then combusted  in a  Packard Model  306  sample
oxidizer  and    CQ& collected in a scintillation  cocktail  consisting of IS
ml  of Permafluor  and  5 ml of Carbosorb  .    C  activity was subsequently
measured  in  a  Beckman Model  LS-8100  liquid  scintillation  counter  with
quench correction  using the  sample channels ratio method.

      Uptake  data  were fit to  linear least squares regressions on a concen-
tration   (ng  PAH-g   dry weight)  and a  slide  surface  area  (ng PAH-cm  )
basis.  Data  were  analyzed using Procedure GLMr Statistical Analysis System
(Barr,  et al.,  1979).  First  order  uptake rate  constants  were  calculated
from  the  flux of  the  initial uptake  rate and  depuration  rate constants from
the slope of  the  log-linearized depuration regression.
 Autoradiography

      Dipping emulsion  autoradiographic  techniques were  used  to visualize
 sites of   C-B(a)P accumulation.  Two slides were  removed at each sampling
 time, dipped in 2% Lugol's  solution for  ca. 30  seconds and then allowed to
 air dry overnight.  Standard  autoradiographic techniques using Kodak KTB-2
 dipping emulsion were used  in processing the samples  (Gude, 1968).  Slides
 were stored in light-tight  boxes  for 5 days, then  developed and mounted in
 Permount.   Resultant autoradiographs were examined and  photographed  on a
 Zeiss Model IM-35  inverted microscope.
 Biotransformation

      The  biotransfonsation  potential . of  periphyton communities  was  de-
 termined by analyzing^ periphyton  for    C as  B(a)P and B(a)P-transformation
 products  (non-B(a)P    C).   Periphyton  were  assayed  by  a  combination of
 solvent extraction,  thin  layer chromatography (TLC) and liquid scintilla-
 tion counting (LSC).  More  specific methods  are  given in section  4.3.  In
 order to demonstrate that  any  observed breakdown  of  B(a)P was the result of
 •some living cellular process and not of  phenomena such ax photooxidation, a
 "dead cell control" was  used in the study of  biotransformation.  Two slides
 each of living and  dead periphyton (fixed for 0.5 h in 4% Lugol's  solution
 and rinsed prior  to exposure)  were sampled after  0.25, 4 and 24 h exposure.
 Living  periphyton  were  immediately   placed   in  Lugol's  solution  (4%) to
 insure cessation of  snv .ongoing biotransformation  of B(a)P.   Since subse-
 quent measurements  of   C activity of  this  Lugol's solution revealed   C
 concentrations at or below background, it was felt that no appreciable  loss
 of   activity  occurred   during  this   step.    Following  overnight  storage
 (-20°C),  periphyton  were  scraped from  slides and  sample wet weights  were
                                     42

-------
measured.  Samples were  stored in glass scintillation vials at -20°C under
benzene:  ethyl  acetate  (1:1, V/V).  Samples were  warmed  to room tempera-
ture  and extracted  by  grinding in  a Ten  Broeck glass  homogenizer with
benzene:  ethyl-acetate  (1:1,  V/V).   Horaogenates were  centrifuged  at 1000
rpm  for  0.5  Jy and (IEC model  HN-S centrifuge) supernatant volumes adjusted
to  10 ml.    C  activity in  pellets  WAS determined as  a  measure of bound
compound.

     For  TLC analysis,  the  supernatant was  sampled  (0.5  ml)  to determine
activity, and  the volume was  reduced  to  100 jjl by N_ evaporation.  It was
ihen  co-chromatographed  with  B(a)P standards on  Merck Silicanized Silica
Gel  60 channel  plates.   Plates were  developed in 9:1  hexane:benzene and
visualized using ultraviolet  light.  Each sample channel was divided into a
B(a)P  section (the standard B(a)P spot  + 1 cm), R,  > B(a)P (less polar) and
R   <  B(a)P  (more  polar).    Each  section was  scraped  from the  plate and
transferred  to^a scintillation vial with cocktail (Research Products Inter-
              MB       1 £*
national  3a70B ) for   C activity measurement.

      In  addition to  determining levels of  these  compounds  in periphyton,
the  breakdown of  B(a)P  dissolved  in  experimental  water was  monitored  as
well.  Water  samples  (3 x 100  ml) were collected at  time "0" and subsequent-
ly  at 0.25,  4 and 24 h  and frozen  (-20°C).   Samples were thawed, acidified
to  pH 4, and extracted sequentially  with  benzene (3  x 50 ml)  and ethyl
acetate  (3  x 50 ml).  Extracts  were combined, dried over anhydrous Nfc-SO.,
reduced  in volume and run on  TLC plates as described above.


RESULTS

Periphyton Communities

      The Castor  Creek and UTRC  periphyton communities differed markedly in
both species composition and  amount of  biomass accumulation after 3 or 6 wk
colonization of  glass slides.  Approximately 50%  (relative abundance)  of
the  Castor Creek attached flora  consisted of  desmids, including such ornate
species  as  Desmidiura coarctaturo (6%),  Hyalotheca sp. (14.5%), Spondylosium
pulchrum (16%)  and Euastrum  pjnnatum  (4%).   Diatoms, particularly Anomoeo-
nesis  serians var. apiculata  (32%) and Frustulia  rhomboides (12%), repre-
sented the  remainder of the  community.  Mucilaginous secretions are common
in  members  of the Desmidiaceae, and extensive  sheaths surrounded D. coarc-
tatum and S. pulchrum.   In comparison,  the diator-dominated UTRC community,
consisted primarily  of  Eunotia incisa (40%) and  Eunotia  sudetica (21%).
This  diatom  conanunity appeared  to  provide  less  in the way of both adsorp-
tive  surface area and mucilage secretion.   Periphyton biomass  on Castor
Creek  slides was greater than UTRC after  both 3 wk (X  =  2.99 vs.  1.72 mg
dry  weight/slide) and  6 wk  of colonization  (X  =  6.88  vn. 5.36  mg dry
weight/slide).
                                    43

-------
Uptake

     B(a)P was rapidly accumulated from water by both UTRC and Castor Creek
periphyton communities.  Uptake of   C-B(a)P was linear during the first 24
h  of accumulation  when  expressed  as  a  function  of concentration  (r  =
.59-.93)  and slide  surface  area  (r   =  .85-.93)  (Fig.  4.A.I).   Biphasic
uptake, consisting of  a rapid surface  sorption  followed  by slower accumu-
lation during the  first 24 h was suggested by the positive y-intercepts in
all  cases (Fig. 4.4.1).   Approximately 37% of  the  total B(a)P  present in
the  UTRC  samples  after 4 h was  accumulated  within the first 15 minutes of
exposure.   For  Castor  Creek  periphyton,  this value was 26%.   The  uptake
rates  determined  on a  surface area  basis from regression  plots  differed
significantly between the two streams after similar  colonization periods (P
<  .01),  and between colonization periods  within streams  (P < .01).  Uptake
rates, on a concentration basis, were  more  variable than those normalized
to  surface area so significant effects were more difficult to demonstrate.
Accumulation by periphyton which  had  been  colonizing for 6 wk in  Castor
Creek was significantly greater  (P < 0.01) than by that which had only been
colonized for 3 wk.  The  slides which were colonized in UTRC had the lowest
biomass and  were the only ones which approached  steady state concentrations
of  B(a)P  (Fig.  4.4.1).   The comparatively  low  regression  r   (C.59)  of
accumulation of B(a)P  by the UTRC  peripbyton after  3  wk of colonization
suggests  a departure from linearity. A semi-log plot of  the same data only
slightly  improved the  fit (r  = 0.66) (Fig. 4.4.1C).

     The  uptake flux can be  determined as the initial slope of the linear
regressions.   The first-order  rate  constants can  then  be calculated with
equation  4.4.1.
           where
                J. =  initial  uptake  flux
                C  =  water  concentration
                K  =  uptake rate  constant


Autoradiography

     Autoradiographs  suggest surface  sorption and s'-.catn  complexation as
the  principal mechanisms  ol  B(a)P accumulation by algae  of both communi-
ties.  Although it  is not possible to differentiate  conclusively between
intra  vs.  extracellular  accumulation without  the use of  ultre-thin sec-
tioning  techniques (R. Knoechel, per.s. comm.),  label did appear to consist-
ently  acjumulate  on  cell   surfaces.  There is virtually no label associated
with the shrunken p.-«.toplast of Netrium  digit.us  (Fig.  4.4.2).   The exten-
sive mucilaginous sheaths  surrounding Desmidimn coarctatum nud Spondylosimn
pulchrum readily  complexed  B(a)P  (Fig.  4.4.2).   Dense labelling  of the
diatom  Neidium viridis var.  amphigorophus was  unique  (Fig.  4.4.2).  Unla-
belled  controls  did  not  show   concentrations  of  silver grains  within or
around algae,  although a background level was apparent.
                                    44

-------
      ? ».ooo
      ^ 4.0OO
      1
      o
      21
      t> 2.0OO
      z
                  UPPER THREE RUNS CREEK
               UD'e*r Oo'o ot Function of tm^ of
               Siiqt S^fTaet A^to
                                       I     tcce
                                    ^\\-
     CASTOR CREEK
UplQX Dora o* function of
 mii e' Siiet
          0»> 24  6  •  10  12 14 it  l> 20 22  24 2*   025 I 2 •»  »  » K>  12  14 It  i« 2O 22  24 2*
                                        TIME Ihra)
Figure  A.4.1.   Accumulation  of    C  B(a)P  by  periphyton  communities  from
                 UTRC  and Castor  Creek  which  had been colonized for 3  or  5
                 wk.   Accumulation is  normalized to  a  surface  area (a  and b)
                 and dry  weight  basis  (c and d).   Each point  represents  the
                 mean  of  3   replications  with  confidence  interval   =  +  SD.

-------

                                                                 I M
Figure A.A.2.  Autoradiograplis  illustrating the  deposition of   C B(a)P  in
               A) DCS mi (Hum  coarctatum,  B) Spondyl osium pulrliriiin, C) Netrium
               di[;i tus,  D)  Ncidi uiii  i ri <\ i s , and  E)  Lunotia  sp.  All  auto-
               radiographs are  500X  magnification.
                                     A6

-------
Biotransformation

     Neither  UTRC  nor Castor  Creek periphyton  communities  exhibited sig-
nificant biotransformation  of    C-B(a)P within 24 h.  In both.living peri-
phyton and dead cell controls, approximately 85-90% of total   C after 24 h
was B(a)P  (Figs.  4.4.3 and 4.4.4).  Histograms representing the percent of
  C activity  as  B(a)P and B(a)P-transformation products at 0.25, 4, and 24
h exposures reveal no substantial differences between living and dead peri-
phyton  communities  for either  stream or  either colonization period (Fig.
4.4.5).  Significant  apparent  transformation of   C-B(a)P occurred over 24
h/in flasks containing experimental water only.  Approximately 90% of total
  C was B(a)P after 0.25 b, while  only 62% remained as  B(a)P  after 24 h.
In previous studies with fish, exposure water analyzed the same day without
freezing never contained  more than 12% B(a)P transformation products after
24 h.
Anthracene

      A set  of studies was  conducted to determine the uptake and depuration
rate  constants for anthracene in periphyton communities colonized in UTRC.
The accumulation of anthracene during 24  b  was linear, except for en ini-
tial   rapid  accumulation (Fig. 4.4.6).   Thie  initially very  rapid  uptake
results in  a positive y-intercept  for linear regressions.  The uptake flux
was estimated as  the slope of the overall  linear  regression.  The first-
order uptake rate  constant  was O.OAh

      After  24 h exposure to anthracene, plates and the associated periphy-
ton were transferred to clean water and  desorption followed  (Fig. 4.4.7).
Desorption  was rapid.  The rate  constant  for  depuration  was 0.17h   which
means that  the  half time  for  elimination was 4 h.   However, there was a
residual of  approximately  4 ng  anthracene*cm   which was  not eliminated
after 30 h  of depuration.
 DISCUSSION

      Results  of our  study suggest  that  community species composition can
 significantly affect  uptake  rates of PAH by periphyton.  This conclusion is
 consistent with  our observations that  !) Castor Creek periphyton had signi-
 ficantly greater B(a)P uptake rates,  2) Castor Creek periphyton, dominated
 by  structurally  ornate desmids with extensive gelatinous  sheaths, differed
 markedly from the diatom-dominited  UTRC  community,  and 3) autoradiographs
 demonstrated  the relative  importance  of B(a)P  sorption to cell surfaces,
 especially to sheath  material.  The larger biomass  on  Castor Creek slides
 did  not appear  to  cause the  greater  uptake  rates observed,  since results
 expressed  on  an  area  basis were not significantly different  between peri-
 phyton  from the  two streams.

     Our results suggest  that uptake rates vary  as  a  function of coloni-
 zation  period,  or  community  biomass.  One  would expect an  increase in
 biomass  to yield a  concomitant increase  in  the  concentration  of PAH per
                                    47

-------
                      Castor Creek
                      Benzo(a)pyrene  and Degradation Products:
                      Three Week Colonization      1/29/80
100

ro
o 60
"
*
S 40
cr
UJ
> 20
o
UJ
a: n



-

-

•



^ • ^
;;
;f
'•'//,
t
'•/,
', •
/

. 	

r


.
'/.
I- .




'.,'













	

'"'/•

'•'•/
lf
'.
/












_^

—

-;,












•


^_^

^
•''.

''.;:



'












'/
•'







-METABOLITES


-B(o)P





                          L  D      L  D     Live  Dead
                      Six  Week Colonization    2/19/80
o 100
t-
0 80
i-
— p
UJ 60
£T
a 40
20
0

-

-
-
PM^<
;,.

;:







—

-







-•
/

:•.






—











;
/






i —



-METABOLITES


-BMP

L D L D Live Dead
0.25 4 24
TIME (hrs)
Figure 4.4.3.
Percent  of  recovered   C  as B(a)P,  sluilcd,  and non-B(.i )P,
unshaded, extracurd from C.islor Creek pcriphyton.  The upper
figure represents peripliyton t rom .1 three week colonisation,
while the lower figure represents prripliyton from .1 six week
coloniza-ion  period.   S.nnples of  both live and dead peri-
phyton were t.iken after 0.25, A and 24 h.

-------
                    Upper Three Runs Creek
                    Ben;o(o)pyrene and Degradation Products:
100
80
60
o
t
o 40
LJ
CE
UJ
> 20
0
o
LJ
K 0


; I.;'-
. i-.-.
v
• v





'


^




i

._
-







• » ^.
J

v- •» ^* W

«




















1 V-< 1


1 1




/'
















»«» «






,

/



iv i » i












' *^»
•-,


•



.




' \

-< v^

t .









j





-METABOLITE!

-BlaJP







i
4

1
• & i* i
_J L D L D Live Dead
i—
g 100
u.
^^
H- 80
UJ
o
) P,
               unr.liaded, extracted  from  Upper  Three Kuns peri pliy ton.   The
               upper figure represents periphyton  from a  three week coloni-
               zation  while  the  lower   figure  represents  five  weeks  of
               colon i x.i I ion.  Samples of  both live and dead periphyton  were
               taken after 0.25, A and 24 h.

-------
                  Upper Three Runs Creek     Costor Creek
THREE
•£. 25
c
^ 2°
in
o 15
§ .0
cc
Q- 5
Z
O o
1/8/80

WEEK COLONIZATION
50
40
B_ TransformaHcn
Products 30
- B'0)P 3°
-

.- FI
P L 0 LO Liv<
<
20

10
-0- o
' 1/29/80
.
•
-
r
^ nn
i
^,
v
'!

„




Deod L D L 0 Live Deod

^ Upper Three Runs Creek Castor Creek
P FIVE
ft 25
Z
c 20
|_
r^
0. 15
o
<£ 10
CC
0 5
a.
o
CD 0
1/22/80
R
1
J
-

.
— nn
L 0 L D Live
WEEK COLONIZATION
50
40
D 30

20

10

" 2/19/80
•
_

-

_
.-i 1-1 M n
"•^
"
7


\

^"
""






— i — i 	 1 — i 	 1 — i 	
Dead L D L 0 Live Deod
0.25 4 24 0.25 4 24
                                   TIME (hrs)
Figure 4.4.5.  Histograms ^f  quantities  of  B(a)I»  shaded,  and  non-B(.i)P,
               unshaded,    C  in Upper  Three  Huns  and  Castor Creek pcri-
               phyton. Samples  of  live  and dead  peripliyton were taken after
               0.25, 4 and  24  h.
                                    50

-------
                                  I	1	1	L.  I   I   I   I  1   I   I _J
     0.25
                   10   12    14    16
                     TIME  (h)
Figure 4.4.6.
Accumulation of   C-anthracenc by periphyLon which colonized
gloss  slices  in  UTRC.   Mean  liiomass/s 1 i(\c  -  0.4  mg, dry
wciglit'cm *".  Anllir.iccne concent i~J t i on in w.itcr  = 22 (ig'Al
Concentrations  are  norni-lizoil  to  an JTCM  basis but can  bo
converted  to  a biomass  basis with  the  convo!sion factor
given.  Each  point  represents  ttic mean  of  3 slides.  Confi-
dence intervals are +  SD.
                                     51

-------
  27r
 E 23
 o
a:
 o>
      24
28
32
36
   4O
TIME
48
52  54
                                             (h)
Figure 4.4.7.  Desorption of,anthracene  from  UTKC periphyton which had been
               exposed to    C-anlhraccne  for 24  h.   Each point represents
               the mean  of  three slides.  Confidence  .intervals  are + SD.
               Mean pcriphyton biomass = .4 nig, dry weighfcm   .      ~
                                    52

-------
unit  substrate surface  area as more  of a  given substrate  is  covered by
periphyton growth.   The  maximal effect.of bionfass on uptake  rate should be
realized when the entire substrate becomes covered by a layer of periphyton
growth.  As  the  community becomes more  layered  with time, only the upper-
most  strata  should  actually be exposed to levels of PAH  within a .24 hour
period.  Any  further increase  in biomass  should not increase uptake rate,
but  rather,  should  increase the length  of  time  necessary for a particular
community to achieve steady  state, and  the absolute  capacity  of that commu-
nity  for  accumulation of compound.  Thus, uptake rates, on a concentration
basis, would be expected to  decrease after a critical minimum thickness, or
mass,  of  periphyton has been  reached.   Our  data show slower uptake rates,
on a  concentration  basis, after longer  colonization  times.  Similar results
have  been-.reported  for  'the accumulation  of selected radionuclides (Co   ,
Zn   ,  Ru    ,  Cs    ) by periphyton (Neal et  aj.., 1967).   Uptake was linear
over  a certain  range of biomass, but  as  mass increased,  accumulation was
not proportional.   Uptake rates are best expressed on a surface area rather
than  a concentration basis.

      The  statistically significant  differences  in  B(a)P  uptake rates be-
tween UTRC and Castor  Creek  periphyton  must be interpreted cautiously since
no  steady state equilibrium was achieved.   The  steady state bioconcentra-
tion factor  (BCF)  can be estimated from  the  ratio  of uptake rate coeffi-
cient (K )  to depuration  rate coefficient  (K.).   Preliminary depuration
results  "for  Castor Creek)  were  too  variable  to accurately estimate K,;
however, .the relatively slow  B(a)P depuration observed  (only a 25% reduc-
tion in   C activity  after  24 h)  indicates that B(a)P will  not be quickly
released  from periphyton once  complexed.  The range of expected values for
K ,  K.,  and BCF for periphyton from other streams,  such  as  northern hard-
waters, would be of interest.

      The  achievement of  steady state  is most probably' linked  to the rate at
which the compound  can diffuse through the multi-layered  periphyton commu-
nity. The plateau observed  in  the  low biomass UTRC three week old community
suggests  an approach  to saturation  of  this  thin periphyton layer.  In con-
trast, the 5 week  old UTRC  community,  and both  the  3 and  6 week old Castor
Creek communities,  are still in a  linear phase of uptake after 24 h.  These
findings  are in agreement with those  reported for Zn   adsorption by peri-
phyton communities  where accumulation was  largely a surface  phenomenon and
ultimate  biomass   saturation  was diffusion-dependent  (Rose  and  Gushing,
1970).   Complete saturation of  periphyton will  be  a function of community
thickness  and exposure  time.   Exposure times of several  weeks may be re-
quired to achieve  steady state for various PAH  in mature  periphyton commu-
nities.

      Biotransformation has  been  suggested  as  'a  significant degradative
pathway  for PAH in  aquatic  environments.   Pseudomonad bacteria in culture
have  been found  to trani.form  B(a)P  and fluoranthene (Barnsley, 1975), as
well  as anthracene  (Evans et al. ,  1965).   Bacteria.in petroleum-contain-'nat-
ed  stream sediments have been found to transfora   C-labelied naphthalene,
anthracene,   B(a)P   and  dibenzanthracene  to    CO *,  polar   trahsformation
products  and  bound material  (Herbes  and  Schwall,   197."; Herbes  e_t  al.,
1977).  In general,  both freshwater (Soto et  al. ,  1975)  and marine phyto-
                                     53

-------
plankton  (Lee  et  al.,  1977) have not  demonstrated an ability to biotrans-
fona PAH,  but  do  accumulate considerable  amounts.   Our data indicate-that
the periphyton communities  examined did not actively biotransform B(a)P in
24-h exposures. The, non-B(a)P   C material present  in  periphyton (as much
as  20% of  total   C)  was  attributed  to  non-specific  B(a)P  breakdown or
binding since  no difference  between living periphyton and that  fixed in
Lugol's solution  was  observed. No separate measure  of  viability or enzyme
activity  in  fixed controls  was made, but  apparent active bit-transformation
of  methyl parathion by  sediment bacteria autoclaved up to 20 minutes has
been  observed  (David  Lewis,  pers.   comm.).  The experiments reported here
weie  conducted under low intensity  gold fluorescent light  (X > 500 nm) to
reduce  photolysis of  B(a)P.   The  importance  of photosynthetically coupled
B(a)P  uptake and  transformation in  natural exposures  remains  to be deter-
mined.

     The  deposition of B(a)P in heterotrophically  dominated communities may
differ  from the  trend reported above.  Biotransformation  of  8(3)^ may be
significant  in environments such as  densely canopied stream areas where P/R
ratios  are redu:ed.   Mclntire  (1975)  observed  the tendency for periphyton
developing on rocks  in  a  stream  to exhibit characteristic  layering with
time.  He  noted an outer, photosynthetically active,  layer  and in time the
development  of a  heterotrophic layer between this outer layer and the sub-
stratum surface.  Thus, the  biotransformation capabilities of a given commu-
nity  may  not be uniform  throughout.  The concentration  of B(a)P transforma-
tion  products may  vary  along a vertical  transect down through the commu-
nity.   The growth of  Chlorella vulgaris #29,  a heterotrophic variety, has
reportedly been stimulated  by  exposure  to  crude  oil  (Graham and Hutchinson,
1975).    Chronic  low  level  exposure to  PAH,  such  as   B(a)P, may  work to
restructure  the periphyton  community and favor heterotrophic growth.

      Where periphyton cover a significant area  of stream bottom, this com-
munity may equal  or  exceed  the importance  of organic sediments in determin-
ing the fate of B(a)P  and other PAH.  The  observed property of composition-
dependent rates of uptake  for periphyton  communities may serve to signifi-
cantly alter  the  responses of distinct  lotic   systems  to  B(a)P perturba-
tions.   Rates  of accumulation and  capacity for  accumulation to B(a)P  input
will  vary between streams  supporting divergent  periphyton  flora and within
the same  stream as seasonal changes in periphyton composition occur.  On a
smaller scale, variations in  response  to  B(a)P input between reaches of a
particular stream,  e.g.  pool  vs.  riffle  area   or  canopied vs.  open  area,
will  occur due to associated  variation in algal species composition.  The
phenomenon of  patchiness  in the microdistribution  of algal  species has been
frequently reported (Bruno and  Lowe,   1980;  Douglas,  1957; Duthie,  1965).
This  property of periphyton distribution  ensures  that  local variations in
B(a)P  concentrations will occur.

     ieriphyton  may represent a significant pathway for food chain bioac-
cumulation.  While direct uptake from water  is  usually  considered the most
important pathway,  trophic  transfer of B(a)P from the  diatom Thalassiosira
pseudonana  to  larvae  of  the  marine  hard clam Hercenaria mercenaria was
reportedly  equal  to  that  of  direct uptake  from  water (Dobroski and Epi-

-------
fanio,  1980).   Grazing  benthic  invertebrates  may  accumulate  6(a)P  in  »
similar manner.

     When uptake and'depuration rate constants, which were derived in these
laboratory studies, were  used in the simulation model,  the  predicted con-
centrations of anthracene in periphyton were very similar to those observed
in periphyton in the channels.microcosm (see section 5.4).

    •In  summary,  periphyton  represent  a  significant and  previously unex-
amined  sink for  B(a)P  in  stream  systems.   Clearly,  additional work  is
needed  to  confirm  the observed low biotransformation rates,  as  well as to
determine  the  physical or physiological basis  of  community  differences in
B(a)P uptake rates.
                                    55

-------
                                 SECTION A.5

          UPTAKE DEPURATION AND BIOTRANSFORMATION OF BENZO(a)PYRENE

                      BY THE MIDGE, CHIRONOMUS RIPARIUS
INTRODUCTION

     Larval midges (Chironomidae) are abundant benthic organisms which bur-
row  *.a sediments  and  have  been  shown to  accumulate trace  contaminants
(Kawatski and Bittner,  1975; Derr and Zabik 1972).  Chironomids are a major
food source  for  larger macroinvertebrates  and smaller  fish, and  link both
aquatic and terrestrial food chains through emergent adults.
                                  *
     This study  of B(a)P kinetics in chironomids was  conducted to 1) con-
struct descriptive models of B(a)P uptake and elimination kinetics, 2) com-
pare uptake  and  elimination coefficients determined using one and two com-
partment  models,  3)  determine  the  ability  of   chironomids to metabolize
B(a)P  and the effect  of biotransformation on bioconcentration factor,  A)
determine  the contribution of exoskeleton to total  bioaccumulation  and  5)
determine the effect of bottom substrate on depuration rate.
MATERIALS AND METHODS

Uptake and Depuration

     Chironomus  riparius  larvae  were collected  from  a sewage  outfall  on
Badfish  Creek  near Madison, Wisconsin, and reared in the laboratory at the
Savannah  River Ecology Laboratory  for  several  generations.   Cultures were
maintained  in  gallon  glass jars  that contained  well  aerated substrate  of
previously  washed  and fermented  ground  paper towels.  Approximately one
gram  of a mixture  of dog biscuits, TetraMin   flake  fish  food shrimp pel-
lets,  and  Cerophyll  (powdered nettle leaves)  was  added  to the jars every
three  days during  the  three week larval stage.  After two days  as pupae,
adults emerged, mated and oviposited in the jars.  Fourth instar larvae had
G  mean dry weight  of  0.39  mg (SD = 0.2, n  = 50) and a dry to wet weight
ratio of 0.095.
                           1A
     Benzo(a)pyrene  (7,10-   C)  was used  as  purchased from Amersham/Searle
in two  lots,  CFAA72 Batch 26 (specific activity 60.7 mCi-1131101" ) and Batch
29 (specific  activity  21.7  mCi-mmol  ).   The specific activities were con-
firmed with a Varian Model 5000 high pressure liquid chromatography system.
Radiochemical purity determined  by  thin layer  chromatography was greater
than 99%.   All  preparative,  analytical,  and experimental  procedures* were
                                    56

-------
performed I'uder gold fluorescent light (\ > 500 run) to minimize photodegra-
dation of B(a)P.

     Well water (pH 7.1) was aerated and centrifuged to remove particulates
(> 0.15  M)-   Wacer was labeled in  bulk  (7-10 £) and dispensed into repli-
cate test chambers after a 1 to 2 h equilibration.  Actual water concentra-
tions were  determined  from calculations based on  dpm-ml    and known B(a)P
specific  activity.   Concentrations ranged between 0.6  -  1.5 |Jg B(a)P-£   .

     Uptake experiments  (in  triplicate)  were performed in 1) 0.35 £ stain-
ing  dishes  which  contained  0.2 £  of water  and  20 chironomids;  2)  6.0 £
aquaria  which contained 1-2  £ of water and 100 - 200 chironomids.  Samples
of  10 chironomids were  taken after  0.25,  0.5,  1.0,  1.5,  2,  4,  and 8.h.
Each  staining dish provided a single  sample of 10 chironomids for total   C
and  10  for  analysis of biotransfonnation products.  The  biomass to volume
ratios  ranged from 0.020 - 0.039  mg  dry wfml   water.  The range of bio-
mass  to volume ratios in both uptake and depuration experiments was due to
differences  in water volume.  The  chironomid larvae were approximately the
same  size in  all experiments.

      Depuration experiments were performed by transferring 100-200 chirono-
mids, labeled for  eight hours, into 6 £ glass aquaria, which contained 1 to
2 £ of clean water.   Ten chironomids as well as one ml of water were taken
from the aquaria after 0.5,  1.0,  1.4, 2.0, 4, 8,  15,__24 and 48 h.  Mass to
volume  ratios  ranged  from 0.035-0.070 mg dry  wfml    water.   Addition of
paper towel  with  associated microflora provided  a  food source,  to avoid
starvation  effects, and  a  substrate allowing  normal  behavioral patterns.
Substrate was not used in uptake  •experiments,  because  sorption and micro-
bial biotransfonnation would  have  confounded exposure calculations.

      Samples  of 10 chironomids were combusted in a Packard Model 306 sample
oxidizer aiid  collected in  a -scintillation  cocktail  consisting  of  17  ml
Penra-afluor   and 5  ml  Carbosorb  .   Internal and external standards  indicated
a   CO.  recovery  greater than 99% and no carry-over.   Water and solvent
samples were placed directly into  a  preraixed commercial cocktail  (Research
Products International 3»70B  ).  All  sample activities were measured with a
Beckman Model LS8100  liquid scintillation counter and corrected for quench
using internal  and external  standards and the sample channels ratio method.
 Biotransfonnation
                                               14
      Both  animals  and  water were  analyzed  for   C- fl(a)P and  transformation
 products.   Chironcraids were assayed  by  a combination of solvent extraction,
 thin  layer chroraf.tography (TLC),  and liquid scintillation counting  (LSC).
•Thirty  chironomids  (10  from  each  replicate)  were homogenized  in  a Tea
 Broeck  tissue homogenizer with five dropu  of concentrated HC1.   The acid
 homogenate  was extracted  sequentially  by homogenizing  with benzene  (5 ml
 nanograde),  diethylether  (2  x 10 ml,  anhydrous),  and ethylacetate  (5 ml
 nanograde).   The  organic  solvents  were  combined,  a   2  x  0.5  ml aliquot
 counted, and  the remaining volume determined.  The samples were dried with
 anhydrous  sodium sulfate and  the volume reduced  to approximately  100  pi by
                                    57

-------
rotary  flash evaporation  and  evaporation  under  a nitrogen  stream.   The
samples were brought to 500 pi vith diethylether and activity determined on
a 25  \tl  aliquot.   Recovery of   C-B(.a)P from spiked chironomids was 92.3 ±
3%.

     Samples  were spotted  onto thin  layer  plates and  chromatographed in
pentane:diethylether (9:1, V/V) in an unsaturated system.  Developed plates
were divided into five sections corresponding to B(a)P, hydroxylated metab-
olites,  the. origin, and  two others.   The  sections were  scraped from the
plate and   C activity measured.  Chironomid extracts were also analyzed by
high  pressure  liquid  chromatography  (see below).   All scmples  were  kept
frozen  (-20°C) prior to analysis.

     Water  from an eight  hour uptake  experiment  was  analyzed for metabo-
lites by 2-dimensioual TLC.  The pH was adjusted to 4.0 with glacial acetic
acid  and  water was passed through  100 ml  of wet XAD-4 resin precleaaed by
the method  of Garnas (1975).  No breakthrough of radioactivity was detected
by  counting aliquots  of water  which passed  through the resin column.  The
resin column was eluted sequentially  with 250  ml  each of diethylether and
acetone.   The solvents were combined  and   C activity of one ml subsamples
determined.   Solvents  were dried by the addition  of  50 ol petroleum ether
 followed  by passage over anhydrous sodium sulfate.   Volume  was reduced as
previously described  and  metabolites  were  isolated  by  2-dimensional TLC
with   pentane:ether  (9:1,   V/V)  and   toluene:methylene  chloride:methanol
 (25:10:1,  V/V/V) (Pitts et  al., 1978).  The  spots were identified with UV
.light and  quantified  by  LSC.  Standards of  B(a)P metabolites  for TLC co-
 chromatography  were obtained from the  National Cancer Institute's Standard
 Chemical  Reference  Repository.

      Water was  also  analyzed  for  B(a)P  by high  pressure  liquid chroma-
 tography  (HPLC).  After chironomids were removed,  200 ml water samples were
 taken from uptake aquaria and  uptake  controls  (no  chironomids) and extrac-
 ted  with  hexane  (3 x 50  ml).   The three hexane  extracts  from each water
 sample were  combined  and  dried with anhydrous  sodium  sulfate.  Extract
volumes were reduced to < 0.25 ml  as  previously described, diluted to 0.5
ml with methanol  and analyzed by a Varian Model 5000 HPLC system with a 254
nm  fixed  wavelength detector.   Recovery of   C - B(a)P from spiked water
was 77.6  ± 6%.   Separations  were made  with a Micro-Pak MCH-10 reverse-phase
column  (35 cm long) equipped with  a  Whatman guard column of Co-Pel C.-ODS
on  35  pm  particles Ufing gradient  programmed  elution  conditions  at 28 C
 (Johnson  et al., 1977).  The gradient was  from 75% acetonitrile:25% water
to  90% acetonitrile.   Acetonitrile (90%) was pumped through the column for
five  rain before  recreating the  initial  conditions.

      Chironomid  extracts  were  analyzed by HPLC  with  a Var.^an Fluorichrome
detector  with 7-54  and 7-60 ewcitat, ->c  filters  (bandpass  230-390 nm, peak
360 nm) and 4-76 and 3-72 emission  filters  (bandpass 430-6'iO nm, peak 525
nm).   Gradient  elution was  from 30%  acetonitrile to  90%  acetonitrile in
water at 2%'min
                                    bft

-------
 Data  Analysis

      Distribution of  B(a)P  and  metabolites  in a  static  uptake  test  is
 represented schematically in Figure  4.5.1.   The data supported the assump-
 tions of  1)  constant water  B(a)P concentration,  2)  low water  metabolite
 concentration (K.  ~ 0)-   Under these conditions, bioconcentration of total
   C was describedoy equation 4.5.1.
                    d C
                    _
                    — :
                     at

where
                           K  • C  - (K . + K , ) • C                  (4.5.1)
                         =  u    w     d    dm     a
                          14                               -1
                     C  =   C concentration in animal (ng-p  ,  wet wt)

                     C  =   C concentration in water (ng-ml  )

                     K  = B(a)P uptake rate constant (h  ;

                     K. = B(a)P depuration rate constant (h  )

                    K,  = Metabolite depuration rate constant (h  )


Low metabolite concentration in water and high metabolite concentration in
chironomids suggested that K.  « K , and a single overall depuration rate
constant designated K, was assumed.  Equation 4.5.2 is the integral form of
equation 4.5.1 using these assumptions.


                    c                          "(Kd't}
                     a     u  d     w

This is the familiar first order equilibrium model, and at steady
state, the bioconcentration factor (BCF) can be calculated as shown in
equation 4.5.3.

                    Cg/Cw = Ku/Kd = BCF                              (4.5.3)

A simple two compartment model (equation 4.5.4) which uj'd not assume con-
»tant water concentration but does assume constant mass within the
uysteoi was also used to estimate ,1  and K, (Giesy et al., 1980).
                                  u      d        — —

          "a = I(Ku ' 'total' ' (VKd» (1-e"(Ku * KdH)            <*'*.4)

where
                       14
          Q  = mass of   C in animals

                                 14
          Q    .  = total mass of   C in experimental system
           to ta L
                                     59

-------
       B (o)P
  METABOLITES
                       KU
                      KD
                       KD
                         m
     B(a)P
           Km
METABOLITES
      WATER
    ANIMAL
Figure A.5.1.  Distribution of B(a)P and nielnbolites in static uptake test.
                        60

-------
Large  concentrations  of metabolites in chironomids  indicated  a large bio-
transformation  rate.   There  were  insufficient  data to  merit independent
estimates of  K    using a model analagous  to  equation 4.5.1 for chironomid
metabolite  accumulation.   The  importance  of biotransformation  was.,shown
instead by comparing calculated BCF (K /K  ) with steady state total   C and
B(a)P.   Loss  of   C  in depuration experiments  was  biphasic.   Rate coef-
ficients  were estimated from  semi-log plots  by  linear least-square fits.

     Rate constants,  steady  state concentrations and asymptotic 95% confi-
dence  limits  were estimated  by the Marquardt iterative least  squares pro-
cedure  (Procedure NLIN, Statistical  Analysis System,  Barr et al., 1979).
RESULTS AND DISCUSSION

     Data  from triplicate  uptake  experiments were  fit to. one compartment
and  two  compartment models  (Table  4.5.1).   Kinetics  of    C accumulation
were  accurately described  by the one  compartment model.   Bioaccumulation
was  rapid  and steady  state activity  was  approached  after  eight  hours.

     The  uptake rate constant (K )  for the  one compartment model was 214 ±
20 li   when calculated  from water and animal  concentrations. A one compart-
ment  model is  only appropriate if  water  concentration is constant.  When
calculated from activity measurements,  water concentration was 1.38 ± 0.32
ng -ml    (X± SE, n  = 3)  initially  and did not  decrease^ significantly
during the  course  of  the  experiment  (1.11 ±  0.25  ng-ml    after  8 h).
Analysis   by  TLC  and  HPLC  indicated no   measurable  amounts  of non-B(a)P
materials in  the  control water  (no chironomids) or after two hours in water
containing chironomids.   After  eight hours, water  containing chironomids
had  86% of   C as B(a)P and 14% as unidentified metabolites.  It was con-
cluded that water activity through eight hours was an acceptable measure of
B(a)T  concentration.

      If  water concentration changes significantly,  and observed losses are
into  the  animals,  a two  compartment model   is required.   Two compartment
models have  a major  advantage  in  not  requiring a  constant  source term;
however,  a knowledge of total  mass  is  required.   They can  be used  in ex-
perimental systems  where  mass  additions   can be  controlled but  may be of
limited use in natural  systems with  unknown  mass additions.
                                                         .1
     The  depuration  rate  constant  (K   = 0.22 ± 0.04 h   )  was estimated
jsing  the  one-compartment model from  data   collected  during  the exposure
phase.   The time  required  for  loss  of  50% of   C was  between 3-4 h.  The
two compartment model estimate of the same constant  (K^. =  0.15 ± 0.001 h  )
was in good agreement.

     Depuration  rate  constants  were  also   estimated  by  placing exposed
animals in clean water  (Figure 4.5.2).  A  semi-log plot of depuration shows
a  slope change between  four and five hours,  which suggests biphasic depura-
tion.   Animals  in  chambers  containing cubstrate  depurated  about  60% of
initial.activity within  4 L.  The  initial  depuration  rate constant (K. =
0.22 h  )  agrees  well with  the overall depuration rate constant calculated
                                    61

-------
        Table  4.5.1.    Bioaccumulation parameter  estimates  for total l C in Chironomus riparius using one
                       and two - compartment  models.

                        ~~~~                                  ~^~  ~~  ~                       Depuration
        Model  and Assumptions                                Parameter Estimates           Half-life (hr)


        One Compartment                                       K               K^
                                   -CK.-t)                      "
          C  = (K /KJ •  C  • (1-e        )               214 ± 20        0.22 t 0.04             3.1
           a     u  d     w

          Assumes Constant Infusion
          B(ajP Concentration Units (ag-g  wet
            weight in animal; ng-ml   in water)
        Two Compartment                                        K              K.
S                                      -(K +K.)t                "    h
          Q =l(K -Q    ,)/(K -HK,)](l-e    U  d    )      0,039  ± 0.003°     0.15 ± 0.001            4.8
                       ,       ,
           a    u  total    u  a

          Assumes constant B(a)P Mass Balance
            within experimental system
         Assumes density of water is 1 g •  ml

         In concentration units 0.039 ± 0.003 h"  = 190 ± 15 h" ,  which compares  favorably with  one  com
              partment m<">del.

         All parameter estimates = Mean ± SE, for pooled data from triplicate  experiments.
                                     -1                                   -1              14
         K  = Uptake rate constant (h  ); K  = Depuration rate constant (h  ) ;  Qa = Mass    C as  B(a)P
              (ng) in experimental system;  t = hr.

-------
        1000

         800

         600
      E?
      I  400
      £
      C7>
         200
K-l=0.06 hr'1
o  No Substrote
•  Substrote
]  X ± S. D.
         100'  1''  ' '   '  .',
               0245  810   15   20  25   30   35  40  45   50   55
                                      HOURS
Figure 4.5.2.   Dopurytion of   C by C. £_ilKirj_ns wiLli and witliput  subst IM Lt;.
                ConconlrnLion  is  cxprc-usotl . ,>s  ng  B(o)P  •  fc   (wot  vcight)
                cliiroiioniid,   assuming all    C  is  B(.i)P.   Eacli point  is  the
                mo;in  of 3 determinations  +  SO.
                                     63

-------
from  the  accumulation data.   The initial  depuration  includes  a component
due to K-  from 4-24 h.  This  component  was comparatively small (K. = 0.04
h  )  and accounted  for 22% of initial activity released.  Between 24-55 h,
19% of the  maximum    C concentration attained  remained  as bound B(a)P and
metabolites.   Animals  exposed to  clean  water  in  the  absence of substrate
also  exhibited  biphasic  depuration but the rates were consistently slower.
Approximately 23% of the initial activity was rapidly released.   The slower
depuration  was  not  complete when the experiment was terminated after 55 h.
As noted,  the  initial depuration rate constant was  similar to  the overall
depuration  constant  derived from a monophasic model and that for B(a)P the
monophasic  model  was  adequate  to summarize  the  data.  This initial  rate
constant gives  the  best overall estimate of  net  flux  of B(a)P and metabo-
lites from  chironomids for  ecological fate studies.

      The mechanism  by which  substrate increased  depuration in  chironomids
is  unknown  but may involve increased biotransfonnation of B(a)P related to
gut  processes, an  active  role  by  bacterial food, behavioral  changes  by
chironomids ,on the  paper  towel substrate, or  increased  gut contents  into
which the   C  labeled compounds can  partition.   Simple  clearance  of gut
contents  was probably  not  important in  observed initial depuration since
animals were starved  for eight hr prior to exposure and for eight hr during
exposure.   More  rapid depuration by  feeding animals may be due  to  par-
titioning of B(a)P  from the animal into uncontaminated food.  Synthesis and
excretion  of the  peritrophic membrane has  been shown  to facilitate excre-
tion  of DDT by mosqaito  larvae  (Abedi and Brown,  1961)  and may be impor-
tant.  Alternatively,  Herbes and  Risi   (1978)  reported  that  feeding de-
creased  the rate  of depuration  of anthracene from Daphnia pulex.  Only 10%
of activity in chironomids was  associated with exoskeleton  (Figure 4.5.3).
Therefore,  surface  losses  were  probably  not  important in  the initial depu-
ration  rates we observed.  Exoskeleton may influence  initial uptake rates
in chironomids since  nearly  50% of  activity after 0.5 h was  in  the  exo-
skeleton.
 BID-TRANSFORMATION

      Information on  biotransfonnation is  important in  hazard assessment
 because  some metabolites of B(a)F  are  mutagenic  and carcinogenic (Lehr et
 al.  1978).   The  biotransfonnation of  I?(a)P by C.  riparius was raj.id.  After
 1  h 43 ± 2% (X  ±  SE) of chironomid    C  activity existed  as non-B(a)P me-
 tabolites,  as  determined by TLC, and  after 8 h 72 ±  2% was non-B(a)P (Table
 4.5.2  and Figure 4.5.4).  The  large  percentage  of metabolites  was not due
 to  B(a)P oxidation on TLC plates  as  evidenced by a  92.3  ±  3%  recovery of
  C  -  B(a)P from spiked chironomid controls.  Biotransfonnation of organic
 compounds  is  common  and  has been  reported  for  chironomids  (Kawatski and
 Bittner  1975,  Estenik and  Collins 1979).   The  rate of biotransformation,
 calculated  from  metabolite  activity and  known B(a)P  specific activity, was
 3.2  ± 0.5 nmol  • g  dry  weight   • h    (Table 4.5.2).   This rate compares
 favorably,  with  that for conversion  of aldrin to dieldrin (1.44 nmol  • g
 • h  ) reported  by Estenik and  Collins  (1979).  The  large biotransformation
 rate  at  0.5 h is likely  due  to sampling time errors,  as  suggested by the
                                    64

-------
 o>
                                                  Whole  Chironomids


                                                  Viscera

                                                  Exoskeleton
       08        10

      Uptake-* I Depuration
135
                                14
Figure  4.5.3.  Uptake nnd  loss of   C in whole C. riparius,  exoskclcton and

              viscera.  Each point represents the mean of 3 determinations

              + SD.
                                 65

-------
O>
              Table A.5.2.  biotransformation of   C - B(a)P by Chironomus  riparius
HOURS
0.5
1.0
2.0
4.0
8.0
% RECOVERY*
1A2 ± 33 C
103 ± 23
97 ± 13
80 ± A
77 ± 6
14C - B(a)P
TLC
767 ± 137
1166 ± 199
13A9 ± 217
1935 ± 115
2085 ± 291
(ng • g )
KPLC
N.D.
N.D.
2180 ± 286
2372 ± 115
2760 ± 286
% METABOLITES1*
57 ± 9
A3 ± 2
57 ± 2
60 ± 2
72 ± 2
BIOTRANSFORM|TION_£ATE
(nmol • g • h )
TLC HPLC
7. A ± 3.0
3.6 ± 0.7
3.6 ± 0.9
2.7 ± 0.3
2.7 ± 0.6
N.D.
N.D.
1.9 ± 0.7
2.2 ± 0.3
2. A ± 0.9
                DPM of sample extracted/DPM of sample combusted.

                PPM Total - PPM B(a)P  x 100.
                     DPM Total

                Mean ± SE, n = 3.

-------
   lOQOr
._  800
0)

c
    600
    400
    200
               •  TotalI4C
               *  I4C BaP
                I  X±S.D.
                                     4      5
                                   HOURS
                                                                  8
Figure  4.5.4.   Accumulation of   C R(a)P (A)  ns determined by  TLC,  and
              B(a)P plus metabolites (o) expressed as ng Ii(a)P * g   (wet
              weight)  chironomiu.   Each point  represents the mean of 3
              determinations ± SD.
                                 67

-------
large standard error.   While  the data are  not  sufficient for careful sta-
tistical analysis,  it was  concluded that there was  no apparent change in
biotransfonnation rate with time.

     The rapid  biotransfonnation observed  will  cause the bioconcentration
factor  (BCF)  reported  in terms of parent compound to be very low.  Indeed,
Lu £t il-  (1977)  reported a BCF  of  "0"  for B(a)P in  fish  due  to complete
biotransfacmation.  The  high  uptake  rate constant and ACT reported here in
terras of    C  suggest very large accumulation of B(a)P metabolites, some of
which may  be carcinogenic.   Accumulation of metabolities may  be the most
important  aspect  of  environmental  risk  as?  jsment  for  certain  PAH.   Tor
this  reason we have reported  both  an apparent BCF  (total    C)  and a true
BCF  (parent compound only).   The major   metabolites  depurated  into  water
during  the uptake phase were  polar  compounds.   Using co--chromatography of
authentic  standards,  the  major metabolite  was  determined  to  be  3-hy-
droxy-B(a)P and  represented 4.4% of water  activity  after 8 hr.  This com-
pound has  also been reported as the major metabolite in mammals (Nebert and
Gelboin 1968).   The 3-hydroxy B(a)P was fluorescent  blue  on the original
TLC  plate,  but became  a  fluorescent red air oxidation product on  co-chroma-
tography plates.

      Small amounts  of three  other  metabolites  were  detected.   These had
relative  abundances  of I) 0.35,  II)  0.06 and III) 0.015 compared to 3-hy-
droxy B(a)P.  Metabolite I  had en Rt value  corresponding  to 7-hydroxy-B(a)P
in the original  2-dimensional TLC  plate,  but  apparently degraded and did
not  move from the origin after  transfer  and co-chromatography with stand-
ard.  Metabolites II and III  were tentatively  identified as the 9,10- and
7,8-dihydrodiols  of B(a)P  respectively.  In  general,  the metabolites were
quite  labile and  degraded  relatively rapidly,  even in  the  dark at -20°C
under  organic  solvent.   Some  of the  polar compounds  observed  may  have
 resulted   from  degradation  during   storage  of  both water  and  chironomid
extracts.   Rapid  analysis after  sampling  is recommended.
                               14
   ,   It was assumed that all    C was B(a)P and  an apparent BCF (ng B(a)P  •
g   wet  weight/ng B(a)P  -ml   )  of 650  was  calculated  from  chironomid
steady  state activity  at eight h.   An estimate of  970 was obtained using
the   ratio K /K..   The  presence  of substantial  quantities  of  non-B(a)P
materials  in chironomids  (Figure 4.5.4, Table  4.5.2) indicated fhat bio-
transformation  was significant,  and the actual BCF was about 200.  B(a)P
was  at apparent  steady  state after 4  - 8  h  (Figure 4.5.4).   Lu  et a±.
 (1977)  reported  a similarily  low   BCF  of  107-149  for  B(a)P  in mosquito
larvae, with only 46% of the  activity present  as  parent compound.  South-
worth et &l.  (1978) predicted a  BCF  of 13,000 for B(a)P  in D. magna calcu-
lated  from  the octanol-water partition  coefficient,  but assumed  no bio-
transformation.  The  difference in BCF between Daphnia and both chironomids
and  mosquito larvae  was probably related  to active biotransfonnation and
excretion  processes  in the  latter.

      These studies  demonstrate  that donor  controlled  linear compartment
models  may be  used  ^j>  predict  steady  state bioaccumulation of B(a)P and
metabolites   (total    C)  in  Chironomus  riparius.   However,  the  BCF for
parent  compound  is reduced by a  factor of  three relative to  that of   C by


                                    68

-------
active  biotransformation  of B(a)P  in this  species.   The frequently  used
correlations between  octanol-water  partition coefficient  and  bioaccumula-
tion may  not be  adequate  for predictions  of  B(a)P fate  in many  species.
                                    69

-------
                                 SECTION 4.6

            EFFECTS OF TEMPERATURE AND-ANTHRACENE CONCENTRATION

                ON UPTAKE,  DEPURATION AND BIOTRANSFORMATION

                    OF ANTHRACENE BY CHIRONOMUS RIPARIUS
INTRODUCTION

     Midges,  important  inhabitants  of  a  variety  of  aquatic  habitats
(Oliver, 1971), often  represent  the largest number of tax  and individuals
in  Lenthic  communities  (Jonasson,  1978).   Chironomus  riparius is  often
abundant  (10  individuals-m~ )  in polluted  streams  (Gower  and Buckland,
1978).   C.   riparius  is able  to metabolize  aldrin (Escenik  and  Collins,
1979)  and  the PAH  benzo(a)pyrene  (Leversee et al. ,  1981a).   Thus,  midges
living in periphyton and sediment may be important in the biotiansformation
of  PAH  and  other  lipophilic  compounds  which  concentrate  in  those  pub-
strates.   Midges  may  also  be  an  important  vector of  mobilization  of PAH
from  sediments  to  high  trophic  levels.  Emergence  of winged  adults  from
environments  polluted with  PAH may  increase the  export *rom  aquatic  to
terrestrial  environments.

     Anthracene was chosen  for study as a representative of the homologous
series  of  PAH because of its relatively low volatility (Southworth,  1979),
intermediate water  solubility  (Leo,  1975),  and presence  in contaminated
aquatic  systems (Sheldon and Kites, 1978).

     Regulatory agencies' are  faced with predicting the fates of large num-
bers  of compounds.   Mathematical simulation models may be  useful  for this
purpose.   However,  few rate constants  for uptake,  depuration or biotrans-
formation  are  available for  use   in  mechanistic, predictive,  simulation
models.  Also, the  effect of external forcing functions on conditional rate
constants  is unknown.  If conditional rate constants for up^ake, depuration
and  biotransformation  are  not independent of the  forcing function,  a sta-
tistical  relationship may  be needed  for  prediction of  these constants.

     Two  important  forcing  functions that affect the dynamics of compounds
within  aquatic  systems and organisms are  temperature  and concentration of
chemical.   One  aspect of  our study was  to  determine  the relationship be-
tween  PAH concentration and  the uptake  and  depuration  rate constants and
bioconcentration  factor  (BCF) in  midge larvae.    The  relationship between
concentration of  PAH  in water and in aquatic animals  has  beeu  tudied in
Daphnia  (Southworth et  aj..,   1978)  and  Callinectes  sapidus  (Lee  et al
1976).                                                              	
                                    70

-------
     A second aspect of this study was to determine the effects of tempera-
ture on uptake, depuration, biotransfonnation, and BCF.  Other authi  s'have
observed a negative correlation between temperature and uptake or retention
of trace  organic  compounds in coho salmon, copepods (Collier et al. t 1978)
and  clams  (Fucik and  Neff,  1977; Harris et  al.,  1977).   However,  Fossato
and  Canzonier  (1976)  reported that  BCF  was not a direct  function  of tem-
perature  in  mussels.  Also,  a  temperature optimum has been  found  for the
mixed  function oxidase system which is  responsible  for tne biotransfonna-
tion of  anthracene  and other xenobiotics in aquatic organisms (Lee et al. ,
1979; Estenik and Collins, 1979).
MATERIALS AND METHODS
     Chironomus  riparius  larvae  were collected  from  a sewage  outfall on
Badfish  Creek near Madison,  Wisconsin.   Cultures were maintained using a
method similar to that described by Leversee et a_l. (1981a) at the Savannah
River Ecology Laboratory for over 20 generations at 25°C.  Animals used for
experiments  at different temperatures were acclimated for 2 to 3 days prior
to exposure  to anthracene.  Overall mean weight  of  5  dry midges was 3.0 ±
0.1 mg (x ±  95% CI,  n  = 334) and ranged from 2.2 ± 0.2 mg (n=30) to 3.9 ±
0.2 (n=30).   The mean wet  to dry weight ratio was 8:1.
 Water and Anthracene

      Well water (pH^.l) was aerated for several days and filtered through
 0.45   pm   Milipoce     filters   to   remove   particulates.    Anthracene
 (3.3mCi'nmol  ,  9-  C) was  obtained  from  California  Bionuclear Corp. (Lot
 #770824) and  used  without  further  purification.   Radiochemical purity was
 determined by thin  layer  chromatogfaphy to be greater than 98%.  All prepa-
 rative  analytical  and experimental  procedures were performed under gold
 fluorescent light  (\  > SOOmn)  to  minimize  phototransformation of anthra-
 cene.
 Dosing and Sampling

      Accumulation flux and uptake and  depuration  rate const.ants were mea-
 sured at four concentrations  (1.7, 8.7. 22.3 and 30.5 jJg-£  ) at 25°C, and
 three temperatures  (16,  25,  and 30°C)  at 22  pg-JZ   .   Uptake experiments
 were conducted in flow-through systems with a  flushing  rate of 1-2 volumes
 per hour to maintain a constant  anthracene  concentration and minimize the
 accumulation of  both   C  labeled  biotransformation  products  and metabo-
 lites.   Anthracene  concentrations  were  monitored  throughout  the study.
 Three or four replicate samples  of 5 midges were taken  after approximately
 0.5,  1,  2,  4,  8,  12,  20, and 30 h (Figure  4.6.1).  While  sampling intervals
 varied  slightly  among experiments,  actual length of  exposure  was used for
 computations.  For  depuration  of  anthracene and biotransformation products,
 midges  exposed to  anthracene  for 9  h  were transferred to uncontaminated
                                    71

-------
2  ~ 40

go


cc  —
ui
o
    o •
    o

    UJ
    z

    o

    IT
    X
    h-
    z
 0
      20
      10
                                           Observed Volues  X±95%C.I.
                                            Predicted Values

                                            _l	1	•	1_
               2   4   6   8   10  12  14  16  18  20  22  24  26  28  30

                                      HOURS
Figure 4.6.1.
                            14
            Accumulation of   C anthracene  by  C.  rip.-irius at 25°C amJ  22

            MR*£  .   Tach  point  represents the mean + 2  SE  of 3 repli-

            cates.  The line represents the  least  squares  fit of equation

            2.
                                  72

-------
aerated water containing paper tovel substrate.  Three replicate samples of
5 midges  and  a  sample of water  taken  after 1, 2, 4,  8,  16,  24, and 30 h.
Water was changed if activity was twice background.

     Biotransfonnation rates  of  anthracene were measured at three tempera-
tures (16, 25, and 30°C) at 22 jJg-£~ .  Four replicate samples of 30 midget
were  taken  after 0.5,  1,  2,  and  4 h.   At 30 C, samples  were also taken
after 9 and 18 h.
Extraction and Analysis
                                                                         £*>
     Water samples were placed directly into the courting cocktail  (3a70P  ,
Research  Products International)  for   C  activity  determination.  Midges
were dessicated overnight at room temperature in glass dessication  chambers
and  weighed  on a model 4700 Cahn  Electrobalance.  Jotal   C in midges was
collected  in a 4 ml Carbosorb  and  15 ml RPI 3a70B  scintillation  cocktail
by combustion by a model 306 Packard sample oxidizer.   Internal and exter-
nal  standards indicated   C recovery greater than  99% and no carryover of
 *V

     Biotransfonnation products of anthracene were measured in undessicated
midges which had  been blotted dry.   Samples were  stored  in  a solution of
 ethylacetate:acetone   (4:1,  V/V)  at  -40  C  until  they  were  extracted.
 Samples were homogenized  in a Ten  Broeck  homogenizer and extracted  twice
 with 5 ml of ethylacetate:acetone (4:1,  V/V)  and once with 5 ml  of cyclo-
 hexane.  The extracts  were  combined  and  filtered through Whatman #40 paper.
 The extract  was evaporated  to 0.5 ml under a stream of  nitrogen.   Activity
 was determined from a  50-(Jl aliquant.  The filter paper was burned, count-
 ed, and reported as unextractable metabolite.  Mean (±  95% CI) recovery of
 spiked samples  was  88  ± 11% (n=5).

      Aliquants  of  200 ul  were spotted  and  chromatographed  on  precoated
 silica gel  thin  layer chromatography plates  (E.  Merck) in hexane:benzene
 (9:1,  V/V)  and pentane:ether  (9:1,  V/V).   Developed plates were  sectioned
 into  4  to  5 parts  corresponding  to the origin,  anthracene,  and  2  and  5
 o£her spots  visualized by UV  light.  Spots were scraped into RPI  3a70B for
   C activity determination.   The  origin,  non-anthracene  spots  and filter
 paper  are  reported  as total metabolite  (see Figure 4.6.6).  Biotransfor-
mation rates were  determined  by  dividing  total  metabolite by duration of
 exposure (see Figure 4.6.5).
      14
        C activity of all samples  was determined using a Beckman model 81Ci<
 liquid scintillation  counter.  The  samples were corrected for backsround,
quench and  counting  efficiency.   The quench  and counting efficiency were
corrected  using  a  quench  curve  and  the  samples  channels  ratio method.
Counting efficiency for  water  and combusted insects  were  85% and 8V?. re-
spectively.
                                     73

-------
Data  Analysis

      Rate constants, BCF  and asymptotic  95% confidence limits were  esti-
mated by  Harquardt  iterative least  squares proceedurcs  (proceedure  NLIN,
Statistical  Analysis System,  Barr et  al . ,  1979), with  a one  compartment
donor dependent  model   (equations  4.6.1  and 4.6.2)  (Giesy  et al,. ,  1980;
Hamelink 1977).

                      dC
                         a    (K -C )  -  (K.-C )                      (4.6.1)
                      — — =   u  w      da
                        dt

where :

     C  = concentration of   C in midge (anthracene equivalents)
      d

     C  = concentration of   C in water (anthracene equivalents)

     K  = uptake rate constant
      u    r

     K, — depuration rate constant

      t  = exposure time  (h) .

Since CW was constant and assuming 1 ml water = 1 g, the integrated analyti-
cal solution of equation 4.6.1 is given by equation 4.6.2.

                    Kn          '(V^
               C  = —   C   (1-e    d   )                            (4.6.2)
                aw
                    Kd

At steady rate state the BCF can be related to the rate constants for uptake
and depuration by equation  4.6.3.
                                                                    (4.6.3)
                F = FT = BCF
                w    d
       Since  K  and  K, were estimated  simultaneously  by numerical methods,
  the>  were highly correlated.   Bias introduced into one estimate by violat-
  ing assumptions  can affect  the other.  Therefore, we estimated K  and K. by
  several  other  techniques which made different assumptions.

       Independent estimates  of K  were also calculated  from  slopes  of tan-
  gents  to the uptake  curve  at  1  h.   This initial rates technique assumes a
  first  order  relationship with respect  to CW and no initial depuration (K  =
  slope  /  C ).  The  overall  depuration  rate constant (K  ) was also estimated
  from  data collected during the  clearance period, when  midges  were not ex-
  posed  to anthracene.   The  overall depuration  rate  constant was estimated
  from  the slope of  the  first  log-linear  component  of the depuration curve.
  The component  K, ,  contributing to the overall K,  were estimated concur-
  rently by equation  4.6.4.
                                     74

-------
                     n        -K..'t
              C   =  Z  C  -e    dl                                (4-6.4)
               at        ai
where:

     C   = concentration in midge at time t

       t = time from cessation of exposure

     C  . = initial concentration in i   component

       i = individual depuration component

     K.. = depuration rate constant for i   component

       n = number of independent depuration components


 RESULTS AND  DISCUSSION
      Uptake of   C-anthracene was  rapid  and  approached steady state in  25
 to 45  h (Figure 4.6.1).  The uptake  rate  constant  for   C increased by_a
 factor of approximately 3  over  the. concentration range 1.7 to 30.5  pg'£
 (Figure 4.6.2).  Few studies  have  computed rate constants for the accumu-
 lation of  PAH by  aquatic  i'ivertebrates.   However,  Herbes and Risi  (1978)
 report a  first-order  uptaKe  rate  constant of  1.01  h   for anthracene ac-
 cumulation by Daphnia  pulex.

                                              14
      As exposure time  increased, K   based on   C generally decreased  (Table
 4.6.1).  Since K   and  K,  were estimated  simultaneously from   C  data,  they
 may be biased  by  accumulation of biotransformation  products.  This will  be
 discussed more  fully  in the sections on depuration  and biotransformation.

      The decrease of K  may also, be due to a  decline in swimming  and  activ-
 ity.   The speed of the normal swimming motion of midges decreased as dura-
 tion of exposure increased.  Midges in uncontaminated water also  showed the
 same  activity patterns.   Slower swimming activity  could be expected  to
 increase the boundary  layer of less concentrated anthracene and thus  reduce
 exposure concentration.  Respiration  and metabolic  rates, as  manifested  in
 swimming rates  (Walshe, 1949)  could  also influence  uptake rate.  K   values
 measured at 25° and 30°C  were not  significantly different, but were  higher
 than that determined at 16° (Table  4.6.1).


 Depuration

      Depuration rate constants were greater at higher anthracene  concentra-
 tions (Table 4.6.1).  As exposure  time increased, a larger fraction  of the
   C was in the bound  than  extractable pools.  Therefore,  K decreased  with


                                     75

-------
           1000
             100
              10
                                    In KU= 3.99 + 0.045 CW
                  1.7
                (CW)
   6.7
|4C-ANTHRACENE
22.3       30.5
CONCENTRATION
                             IN  WATER  (ng  • mL~')
Figure 4.6.2.   Lo£  K  for C.  ri par ins at 4  different concent r.it ions of
              nnlhracenc  in water.  Futim:itc  + .isymptotic 95% CI.
                                 76

-------
 Table  4.6.1.   Uptake  (K  )  and depuration  (K ) rate constants in C. riparius at four concentrations
               and  three  temperatures  (estimate 1 95% CI).
Concentra-
tion
t% ™ -*
Tempera- K
ture "
Kd
                         Ih
                          ab
lOh
30h
lOh
30h
                                                                               3h
                                                             30h
1.7
8.7
22.3
30.5
22.2
22.3
22.7
25
25
25
25
16
25
30
82
79
193
2.»5
128
193
153
77±21
54±13
161112
222*32
116110
161112
157.114
68113
74110
155116
143120
io?.±n
155116
;52il6
0.07010.

0.10710.
0.27610.
0.09110.
0.10710.
0.08910.
073

021
070
023
021
026
0.03510
0.04110
0.08510
0.07810
0.05310
0.08510
0.08910
.019
.014
.014
.017
.011
.014
.015
0.07710.
0.21510.
0.15810.
0.09310.
0.14410.
0.15810.
0.26ciO.
023
078
176
014
027
018
139
0.31110.224
0.23610.098
0.24710.101
0.17710.120
0.25010.457
0.24710.102
0.36310.231
a.  Duration of test.  All values estimated from concentration of 143-anthracene in midges (ng-g
    wet weight).
b.  Estimated from tangent to slope at 1 h.
c.  Estimated from   C-anthracene uptake experiments using donor dependent uouel (equations 4.6.1
    and 4.6.2).
d.  Measured as slope during initial 3 /. depuration in uncontaminatcd water and paper towel sub-
    strate after 9 >; exposure.
e.  First component of overall depuration race constant calculat3vl-with equation 4.
                                                                                                -1

-------
time,  since  a  smaller proportion  of  the  * C  material was  available for
depuration.  Polar biotransformation products  have been shown to be elimi-
nated  more  slowly by  many  aquatic animals (Land runt and  Crosby,  1981a and
198lb).

     Depuration  rate  constant  was  not  affected by  temperature  (Table
4.6.1).   Because  most metabolic  functions are temperature  dependent over
the  normal temperature  range  encountered by animals one  would  expect that
both biotransformation and active elimination would be affected by tempera-
ture.  The fact that we see no effect of temperature ou the depuration rate
constant, based on    C activity, is probably due  to  a  complex interaction
of temperature effects on metabolism, excretion and anthracene biotransfor-
roation.

     When nidges,  which had  accumulated anthracene, were placed in water
with paper towel  substrate, no effects  of  temperature  or concentration on
K. were  observed.  K. values, calculated from the donor dependent model (Equa-
tion 4.6.2),  were generally lower  ia the  abs-nce  of  paper towel substrate
than when paper towel  substrate was present (Ttble 4.6.1).  Leversee et at .
(1981) observed  greater  rates  of depuration of  benzo(a)pyrene, (B(a)P),
froa* C.  riparius  in  the presence  of  paper towel  than in  the  absence of
paper  towel.

                                         14
     We observed  biphasic depuration of   Oanthracene  (Figure 4.6.3).  The
two  phases may represent depuration of anthracene  from two storage compart-
ments, or depuration  ol anthracene  and its biotransf onaation products which
are  more polar.  Herbes and Risi  (1978) observed  a multiphasic depuration
of  anthracene from Daphnia puKx.  Resolution of  these questions would be
difficult in the midge and our  studies  were  oot  designed to resolve them.

     However,  we  did resolve  the overall  depuration  rate  constant into
several  component parts.   In each  case  where we determined the overall and
more rapid depuration rate  constants the overall rate constant estimate was
slower than  the estimate of the  rapid phase (Table '«.6.1).  This indicates
that our estimate of  the overall  K, is  intermediate between the more rapid
and  slower  rates.   The  overall  depuration rate constant, *» estimated by
equations 1  and  2  should  be a weighted average of the multiple  rate con-
stants (equation  4.6.5).
                                                                  (4.6.5)
where:
          K.  - overall depuration rate constant
                                    78

-------
      '0
  CE ^^ C
  H- —
  Z  ?
  UJ .-
  u  «
  z ^

  o~
     O1
  Ul  .

  s  -
  u ,£.

  ce
  x
  o



                                       J_
                                                             J_
                                                                 J_
                               10   12   14   16   18

                                     TIME (h)
2O  22   24  26   28  30
Figure 4.6.3.
Depuration of   C anthrjrene by  C.  rip .iritis in

water and  pjpcr  towel  substr.ite  at 30°C.  X  *
           unconljminJLcd

            2 Sc, n  =  3.
                                    79

-------
          C   = initial  concentration  of  i    component

          K   = i   depuration rate  constant

     For predicting body burdens  in populations we feel that this level of
resolution is not necessary and advocate  use of the overall depuration rate
constant.   If  one needs to  resolve depuration rate  constants,  we suggest
that estimates be made by fitting the  data,  using equation 4.6.4, by inter-
ative  least  squares  techniques rather than the "back stripping" technique
of Wagner (1975).  It can be seen tuat unless the sample size is large, the
confidence intervals for the estimates are  very large (Table 4.6.1).

     After two depuration phases had been identified, a residual of approx-
imately 20%  of the initial   C still remained  bound in the animal at 30 C
(Figure 4.6.3).   This  residual  was .depurated  very  slowly or  not at all.
The  slope of  the log-transformed   C concentration  as a  function of time
was  not significantly different from zero.
 Biotransformation

      The  biotransformation  rate  was maximum  at  25°C  (Figures  4.6.4 and
 4.6.5).   Because the  biotransformation  rate was  high,  percent anthracene
 was minimum at  2S°C.   The mass of anthracene itt  the  midges,  however, was
 maximum  at 30°C because accumulation of   C was  as high, but biotransfor-
 mation was lower at  30°  than at 25°C.

      The  tetnperature effect indicates that the enzymes  responsible for the
 biotransformation  are responding  optimally  at 25°C.  Estenik  and Collins
 (1979) found in vitro aldrin expoxidation activity in C. riparius by mixed
 function  oxidase was greatest at 30°C.  Thus, the decline in biotransforma-
 tion froM 25° to 30°C was  probably due to a decline in  the overall physio-
 logical  state of the midges and cot  due  to  enzyme efficiency.  Additional
 evidence   for  reduced physiological  fitness  at  30°C  was seen  in culture
 growth.   Cultures  of midges maintained  at  30°C  did  not  remain viable,
 whereas  those at 25"C  propagated successfully.

      At  16°C,  the  overall biotransformation  rate was slower than at higher
 temperatures  (Figure 4.6.6).  Additionally,  the proportion of biotr^nsfor-
 mation anthracene  that was unextractable was smalle: -„ at 16°C.   This sug-
 gests that  biotransformation  involves  several   processes  which  respond
 differently  to   temperature.  Likewise,  Varinaui  et aj_. (1981)  found that
 the  relative  concentrations of   different  naphthalene biotransformation
 products  in starry flounder also changed with temperature.

      The  bfotransformatioo rate  decreased during  the  first  4  h (Figure
 4.6.5).   This stay  have been due to a decline  in respiration, as discussed
 earlier,  or to  a  build-up of  biotraoBforniation  products and  end product
 inhibition.  Leversee  et aj..  (1981a)  found a similar decrease in biotrans-
 formation rate in midges exposed to B(a)P.  However, at  30°C biotransforma-
 tion  rate tended to  increase after 4 h (Figure 4.6.5).
                                    80

-------
I<+U
Ul
< ~i2o
a •
.e

| O.'00
< £• 80
S T>
o: -"
0 7 60
(^ 0>
< - 40
S o
£ 6
0 c 20
CD ~~
Q
	 16'C
	 25"C
, 	 30°C
V 1
u
1 l
r
1
•\
V
n

f

-\ , 4
\ ^.-•-"""'"'^
r^ ^.-— •• — '
^-k x^ .. 	 -"
i -Os T 4— —
^ 	 1-
\ 	 ',
i ~~~\
*


• I I t 1 1 1 1 t l l 1 t t ! f 1 1 1
0 2 4 6 8 10 12 14 16 18
TIME(h)
Figure 4.6.4.   Biolransform.ition rate after  1  h  exposure as a  function of
               tempera lure,   X + 2  SE,  n = 4.
                                   81

-------
   ui
   5
   o
      ,4
      '2
   to

  «•« .0
   O
    «  6
    •e
80 - % Recovery
   - Unextroctoble metcbolito

   - Exlroctoble metobolite

    - Anthracene
                 78
                         73
            92
                              80
                                      66
                                            79
                                                         76
                                                    73
                                                                 76
                                                                         76
   Temp.CCne 25  30      162530      162530      162530     30      30
Length of
Exposure(h):   0.5            I             2            4          9       10
       4.6.5.  Biotrnnsfonn.TUon r.ttc .is A  fnnclio.i of  timo of cxjiosurc and
               tempc 1.1 Lure.  X  + 2 SE, n =  A.
                                      82

-------
     120
UJ ~-~


^  •

il
      60
 2 5 40
 CD
      20
                       10             20
                             TEMPERATURE  (°C)
                                                     30
Figure 4.6.6.  Biotr.inuforra.iliou  of    C  .inllir.'iccno at  16,  25  anJ  30°C.
              + ?. SK, n - it.
                                  83

-------
     At  30°C,  the percent of   C that was anthracene decreased with  length
of  exposure  from 48% at 0.5 h to 9% at  18 h (Figure 4.6.6).  The trend was
less  pronounced at the other temperatures^ Between 4 and  18 h, anthracene
concentration  decreased by 2/3 while the   C  activity  ro^ by a factor of
2.7.   However,  at all  temperatures percent unextractable    C increased with
time.
 Bioconcentration Factor
                                                           -1           14
      Bioconcentration factor (BCF), calculated using KU*K.   based on   C,
 was  not  affected  by  temperature  or  concentration  in the  water  (Table
 4.6.2).  Other  studies  with  aquatic  organisms,  Canton  et  al.  (1978),
 Blanchard ct al.
Table 4
                                         14
.6.2.   Bioconcentration  factors  (BCf) for   C-anthracene in the m^dge
       C.  ripai jus.   BCF for  total   C wad calculated from K^K^
       after 10 and  30 h of uptake.  BCF for anthracene was calculat-
       ed  from concentrations of anthracene at 4 hr.  x ± 95% CI.
Concentration
                     Temp
                                    BIOCONCENTRATION FACTORS
(ng-ml )
1.7
8.7
22.3
30.5
22.2
22.3
22.7
(C)
25
25
25
25
10
25
30

10 h
9151532

15031206
8041104
12771245
15031206
16971384
14C
30 b
19641722
18171388
18201136
18431180
1915*228
18201136
17021119
Anthracene
4 h




132139
47115
9515
 (1977)  and  Wilkes and Weiss  (1971)  have  likewise shown that BCF was  con-
 stant at different concentrations.   Similarly,  bioaccuamlation was  found to
 be  unaffected  by  temperature in  Mytilus edulia  (Fossoto and  Canzonier,
 1976).  Bioconcenlration factor predicted  from  the 30 h data  set  was  usual-
 ly higher than  BCF from  the 10 ^  data.   The difference is probably  due to
 accumulation of  non-anthracene   C.   The  BCF calculated for  anthracene in
 b£otransfonnation  studies  was more  than  an  order of magnitude  lower  than
   C-BCF, due  to rapid biotransformation  of  anthracene.  Steady state  had
 not been reached  by  4  h so a true  BCF can not  be  calculated.   Intcresting-
                                     84

-------
ly, at 30°C,  the  concentration factor lit 18 h was 68 ± 17, which is 2/3 of
the concentration factor  at  4 h.  Concentration  factor  was lowest at 25°C
and reflected the biotransformation rate.

     BCF  was more  strongly  affected  by differences  in biotransformation
rate due  to  temperature than by differences in uptake rate due to tempera-
ture.  For example,  BCF was highest at  16°C  because biotransformation was
lowest, even though uptake rate and concentration were also lowest at 16°C.
Consequently, caution  must  be exercised in comparing BCF for roetabolically
transformable  compounds.   As  we  found  with  anthracene  in  chironomids,
  C-BCF leads to overestimation of bioconcentration potential.
Benzo(a)pyrene and Anthracene

     Since  octanol:water  partitioning coefficient has been positively cor-
related with BCF  (Neely et aK,  1974), benzo(a)pyrene  (BaP) would be expec-
ted to have a larger BCF  than anthracene.   The   C-BCF of anthracene (915
at l.J ng-ml" )  was similar  to BCF for B(a)P  in C.  riparius (970 at  1.0
ng-ml" )  (Leversee, et. al., 1981a).   Though  both K  and K. were slower  for
anthracene, the   C-BCF  \alues were similar  for anthracene  and B(a)P  be-
cause  their relative proportions  remained  constant.   The  A h accumulation
 factor based on parent compound was  approximately 4 times greater for B(a)P
 (200)  than  for  anthracene  (47),  as might  *>e  expected from the  octonol:water
partitioning coefficient  predictions.  Bioaccumulation was calculated using
 parent compound  at  4 h,  since steady: .state  was not easily  recognizable.
 The difference between  BCF based on   C and  BCF based on parent  compound
 for anthracene compared respectively to  those for B(a)P indicates  a slower
 eliminaticn of  anthracene biotransformation  product.  Relative  rate.s of
 biotransformation on a molar  basis were  similar for B(a)P (2.7 mnol-g  -h
 at 4 nraol'Jfc  )  and anthracene (51  nmol-g   'h   at 125  nux>l'£~  ) considering
 the-difference  in concentration.
 CONCLUSIONS
                                                              dC
                  14                                          	a         1.
 Accumulation of    C-anthracene  is  described  by  the model  dt  =  (K  -C  )
     - (Kd-Ca)  (Equation  4.6.1).                                       u  w

 2.    Uptake  rate  constants  are  temperature and  concentration dependent.

 3.   Bioconcentration factor based on   C  is not affected by concentration
     or  temperature.

 4.   Bioconcentration factor based on anthracene is  minimum at  25°C and  iu
     inversely proportional to rate of biotranaforroation.
                                    85

-------
                                 SECTION 4.7

           UPTAKE,  DEPURATION AND BIOTRANSFORMATION OF ANTHRACENE

                      BY THE CLAM, ANODONTA IKBECILLIS
INTRODUCTION

     Uptake and release of polycyclic aromatic hydrocarbons (PAH) and other
petroleum hydrocarbon  compounds  by  marine  bivalve molluscs have  been the
focus  of much  research  in  recent  years.   Being sessile  filter  feeders,
clams  are readily  exposed to pollutants (Nunes and Bcnville, 1979), and it
has been suggested that these organisms could be used as worldwide monitors
of hydrocarbon  pollution  (Stegeman and Teal, 1973; Goldberg, 1975; Fossato
and Canzonier,  1975;  Dunn and Stitch, 1975; DeSal/o et al., 1975, Bravo et
al.,  1978).   The  general  conclusions of past studies of hydrocarbon inter-
actions  with  clams have been that  1)  uptake is rapid (Nunes and Benville,
1979), but  can be highly variable  (Jackira aud Wilson, 1977), 2) depuration
consists  of several  release  compartments  (i.e.  fast  and slow)  (Dunn and
Stitch,  1976)  and depuration half-life is directly dependent on the length
of  exposure  to a  pollutant (Bochra  and  Quinn,  1977; Jackim  and Wilson,
1977), and  3) biotransformation  of  hydrocarbons  by molluscs  is minimal or
nonexistent  (Carlson,  1971;  Lee  et  o_l., 1972; Dunn and Stitch, 1976; Nunes
and Benville,  1979).  This lack of  biotransformation is probably due to the
fact  that molluscs are  deficient in  mixed function oxidase enzymes (Neff et.
al_.,  1976;  Bend et aj_. ,  1977; Moore,  1979).  Molluscs have also been found
to  selectively concentrate  PAH  compounds  relative to  other  hydrocarbons
(Lake and Kershrner, 1977).

      PAH studies concerning  freshwater invertebrates have centered on Daph-
ni£ sp.  (Herbes and Risi, 1978; Southworth  et  aK,  1978) and chironomids,
which hive  been  shown  to  have  a high biotrensformation  potential (this
report).  The  primary objective  of  this study  was  to determine fluxes and
rate  constants for uptake,  depuration and  biotransforroation of anthracene
by the freshwater clam  Anodonta imbecillis.  The  symmetry of the anthracene
molecule  lowers the  number  of  possible metabolites  compared  to other PAH
compounds,  such as benzo(a)pyrene,  making  the  assessment  of biotransfor-
mation readily feasible.   An ideal  freshwater  monitor organism the paper-
shell  clam,  A.  imbed 11 is,  is a common member  of the benthic community of
lakes  and large rivers  of North America, Europe, and Asia (Pennak, :953).
                                    86

-------
MATERIALS AND METHODS

     Anodonta  imbectllis  were collected  August 4,  1980  by SCUHA from PAH
pond  (on  the  Savannah River Plant, Aiken,  S.C.)  at depths  from  1.5 to 3.0
meters.   Clams  were  returned to the lab and held for 2 days in a tank con-
taining  aerated, PAR pond  water  at  25°C.   Twenty-four  hours  before the
experiment  was initiated,  eight  clams of  similar  sii.-e  (X valve length =
7.07  ±.315 cm)  were  cleaned externally, and  placed in individual aquaria
containing  filtered   (0.45  M™).  aerated,  PAR  pond  water  to allow for gut
clearance.  The clams were blotted and weighed before  transfer to experi-
mental chcmbers.  Mean tissue weight was  17.41  ±.82  gm, wet weight.

      14-C anthracene  was  added in bulk to > 0.45 uffl  filtered PAR pond water
allowed  to equilibrate for  2 hours, and then 1.1 £ was added to each of the
5.0-£ (covered) glass experimental aquaria.  To allow for  adsorption to the
walls, the aquaria were left for an additional  hour.  The  fiaaljCquilibrat-
ed concentration of  14-C anthracene was  31 ug-£~   (0.68  uCi-£  ).  Experi-
nents were conducted  at 25  ± 1 C.

      Water samples were taken  after 0.0,  0.5,  1.0,  3.0, 5.0, 13.0, and 24.0
 h and the anthracene  concentrations were  determined  by liquid scintillation
 counting (LSC).  One ml  of water was  added to 12. ml of a  premixed counting
 cocktail  (Research Products  International  3a70B ).  All  activity measure-
 ments were made by  «  Beckman Model LS8100 liquid  scintillation counter, and
 were  corrected for  quenching  using  internal and  external standards and
 sample channels ratio.   All procedures  involving anthracene were performed
 under gold flourescent  light (A  >  500  run)  to minimize  photodegradation.

      After 24 h of exposure to anthracene,  three  of the remaining six  clams
 (2 died  during the experiment  and  were not used in the  analysis) were i e-
 moved and  frozen for extraction  and  analysis.  The other  three  clams w>,-re
 placed  in aquaria containing  2.0 £  of uncontaminited  filtered PAR pjnd
 water for  depuration studies.  Before  transfer to  :'reezer bags or depura-
 tion  chambers the  clams  were  dipped in clean pond water and  allowed  to
 drain to remove 14-C activity from loosely bound  water  on the surface  of
 the  shells.   Water samples were  taken after  0.0, 0.5, 1.0, 1.5, 3.0, 5.0,
 12.0, and  24.0 h, and  water activity was measured  by LSC to determine the
 rate of depuration  from the clams.

      For tissue analysis, individual  clams  were homogenized in a  Ten Broeck
 tissue homogenizer.   Emptied shells were  blotted  and wet  weight determined.
 14-C combustion analysis was used  to  determine the final  total  anthracene
 tissue concentration.  Tissue extracts were analyzed by  thin-layer chroma-
 tography (TLC)  and HPLC  to determine  concentrations of  possible biotrans-
 fomuition products  of anthracene.  Clam  horcogenate (0.25  ml) was combusted
 in a  Packard  Model   306  sample  oxidizet^ for  1 min.,  and  was collected  in
scintillation  vials  with  5 ml Carbosorb  and   13 ml Permaflour .  Previous
experiments showed,  by use of  internal and external standards, that recov-
ery of  14-O2  wao  90% but possible  contamination  of  the  oxidizer lowered
the  percent   recovery  to  81%  during  this  experiment.   Oxidized samples
accounted for  95 ± 27% of  the  14-C anthracene in  tissues as was predicted
from  water data and  the  mass balance  assumption.   Extractions  for TLC, and


                                    87

-------
extraction and  analysis  procedures for HPLC are, given in Section 4.6.  TIC
separations  were  made on  silica gel  plates,  one set  with hexane-lenzene
(9:1, V/V) and  another set with pentane-ether (9:1, V/V) as solvents.  The
position  of  anthracene was determined by  comparison  to standards parallel
to  the  samples.   Recovery by this technique was 70.7% ± 3.5 (95% C.I.) for
anthracene and  79.3% i 2.9 for anthraquinone, a possible biotransformation
product of anthracene.  After "isualization with UV light, areas 1 cm above
and below the  14-C  anthracene and metabolite spots  were  scraped from the
plates  and placed in LSC vials containing counting cocktail for subsequent
analysis.

      Since  changes in water anthracene concentration  were  used to measure
uptake  and  depuration by A. imbecillis ,  it  was  necessary to determine the
contribution of the shells alone.  Weighted and sealed shells were used in
an  uptake and depuration experiment which duplicated that with live clams.
Surface area was determined for  the shells  in all experiments.  Uptake and
depuration  data  obtained  from living clams  were  corrected for calculated
contribution from shells.

      A Marquardt iterative lear,t squares method was used to estimate first
order rate constants and asymptotic 95%  confidence limits  (PROC NLIN, Sta-
 tistical Analysis System,  Barr et al^. ,  1979).   Because  the dosing system
used in these studies was  closed,  a two compartment model was used to esti
mate the  first order rate  constants  for uptake and depuration (Equations
 4.7.1-4.7.4).

               dQa   K                                             (4.7.1)
               -^ =   uqw    dqa


                         AO  = Qw + Qa                              (4.7.2)

assuming equation 6.7.2 is  tiue  and integrating gives:
                         -OC+KJ-t
                           "
             and


* A0(l * K  - e'(Ku + Kd)
-------
    K  - first order uptake  rate  constant

    K  = first order depuration  rate  constant
     d
    AO = total mass of  anthracene in  system

      t = time

     t  = time zero
      o

RESULTS  AND DISCUSSION

Shells

     A two  compartment  rood--;! (Equation 4.7.4) was  us_e^J  to estimate KU and
K   for shells.   K  and K  were 0.092 h"  ?\ud 0.362 h"  respectively during
the first  5.0 h  oi the experiment.   Only  data  from the first 5.0 h of up-
take by  shells wers used  because of  loss  of anthracene  from  the water (>
20% aftei:  5.0 h).  Liss was primarily attributed to adsoiptioa of compound
to  dust particles which accumulated on the surface of the water during the
rarople period.   After  2k  h of  depuration,  steady state  was  reached, and
only 12% of  14-C  anthracene  hjd been  released by the shells (Figure 4.7.1).
This type  of  almor.t irreversible  adsorption raises  doubt about the use of a
two compartment model to predict  the  uptake of this compound by the shells,
or  at  least  in this case,  the  use of the term  'depuration rate constant'.
Instead  of a depuration  rate constant, K, reost  likely  represents a sab-
str.itc saturation rate  constant.

     After 24 h  exposure  to 31  fJR anthracene'£   , clams  had accumulated
0.71  ± 0.098 pg anthracene/g wet  weight  (X t  SD) of soft tissue (Figure
4.7.2).   The predicted  steady  state  concentration (CSS)  of  0.75 had been
reached  by this  time.  The  first order rate  constants for uptake and depu-
ration are given in Table 4.7.1.   Uptake  appeared to be biphasic for this
experiment (Figure 4.7.2).   However,  due to  the absence of data points
between  13 and  24 h,  where the phase shift  is  suggested to occur, short
term multicompart.mentalization for  uptake  in  thest  clams  in merely specula-
tive.   The literature  on  compartment  theory  ib  incomplete for short term
uptake circumstances  since no previous experiment  has exhaustive'y studied
the first  24 h  of exposure of  clams  to  organic  compounds.   The overall
uptake  rate  constant was  estimated as a  single  value.   The  two component
nature of  the uptake may  have beer, due to saturation  of  a pool such as the
extrapallial  fluid,  lollowed by  partitioning into the  soft  tissue of the
clam.   In  otJer  to  estimate overall  rate constants  for the  24  h uptake
period,  the  13  h data was  omitted  for use  in the  two compartment model
since in  this experiment  raultiphasic uptake  was   asciuited  to  be additive.
The model  used  here  assumed that  all  of  the anthracene  accumulated by the
clams was  from exposure to anthracene dissolved  in  the wjter.  This assump-
tion is supported by  the work of Dobrcpki and Epifanio  (1980) which showed
that the hard clam (Mercinaria mercinaria) accumulated more B(a)P directly
from water than from B(a)P-containing  food.
                                    89

-------

J^~*
1
o
jC
(/)
•
O>
^.
*"•*
to
0
C..C.'
\ B-
1 • U

1.4-

1.0-


0.6
n ?
•
^______ 	 £
* s"1^

/*


s
^
J
£ N ANTHR'ACEN E.'^I N PU J .'^
                                                          • - Observed
                                                          & - Predicted
            tI   I   i  i   I   r  tT  T  t  I  ]i  (   I   i   iT   i  i   t   I   I  *
            2     6    10   14    18   22   26  30   34   38   42   45    50
                                      TIME (h)
Figure 4.7.1.  Accumulation  of  .mthr.iccnc  by  A.  imln-fj 11 i:;   :;lioll  .1::  a
               function of time.   The- fitted line i.s h.ir.c-1  on  tlio first 5 h
               of upt.ike only  because of loss of .-inthr.io'nc from the v.itor,
               which  ilid  net p,o to  the cl.ims.  Tlii:;  rrsulta  in a ('ccrc.ise
               in  the  .imounl  of  anthr.icone  .issoci.itcil   villi   tli'?  slu-lls
               (points Mtcr 24  h).
                                     90

-------
      1.0
u
ui   0.8
o _
< .c
^ .? 0.6

»- 5
< ^ 0.4


^   0.2


     0.0
             CA = CSS (
            •Observed
            - Predicted
                        ANTHRACENE    INPUT
              24   6    6   10   12   14   16   18   20  22  24
                                 TIME (h)
Figure U.I.2.  Arciimul.iliori of  antlirarcnc  by soft tissues of A.
             a» a  function of time.                       ~
                                91

-------
           Table 4.7.1.   First order rate constants for uptake
                         and depuration of anthracene by A.
                         imbecillis from 31 pg anthracene.
                         Estimated by Marquardt iterative
                         least squares procedure.

                                 SHELLS                    ~~

                                   Asymptotic
                       Estimate        SE        F and P of F
             K *        0.092        0.015       F-  , = 184
              U                                   2,2o
                                                 P<0.0001
             K.2        0.362        0.126
              d
                               SOFT TISSUE

              Kul        0.0052       0.0006      F     - 215
                                                 P<0.0001
                        0.216        0.033
              Kd         0.200        0.079       F2 ^ = 57

                                                 P<0.001

             Overall uptake rate constant with units £'(h'g)~ ,
             wet weight
            2
             Overall depuration rate constant estimated during
             uptake phase, with units h  .

             Single component depuration rate constant esti-
             mated during depuration phase, with units h  .


     Depuration  studies  showed  that compound release  is biphasic  during a
24  h  period with a shift in  the  depuration  rate occurring between 1.5 and
3.0 h  after removal  from exposure to 14-C anthracene.  A oultiphasic depu-
ration model (Equation  4.7.4)  gave estimates of 0.118 h   for  K   (0 - 1 5
hrs)  and 0.011 h   for K    (3.0  - 24.0 hrs) (Fig 4.7.3).   Theddepuration
rate constant  estimate  (O  from  the two compartment  model  was similar to
the overall  depuration  rate  constant K l estimated from the two component
depuration model.  The  dispersion in tne data was such that we also used a
single component depuration model to estimate K  phase (Table  471)   The
overall  depuration  rate constant  was  similar to that  estimated from data
collected during  the  anthracene  exposure.   However,   it can  be seen that
                                    92

-------
   0.2
 c
o
   O.lL-t
         • Actual
        — Predicted


                              10   12   14    16
                                 TIME (hr)
18
20   22   24
 4-7.3.   D,-pur-iLion of  anlhr.iccnc from  the  clim A
         H.U.,  |,.,vc  been  fit  to  .1  two  comnonnnT'
                               93

-------
this rate constant is too great and that there is a slow loss of anthracene
which is not  accounted  for by this model.   Therefore tne best estimate of
the  overall  depuration  rate  constant  is  O.llSh" .   However,  even this
constant alone  would predict  the  loss  of anthracene from  the  clams to be
more rapid than it really is.

     The  rates  of depuration  of PAH  from  molluscs  has  been  shown to be
non-selective,  and  slower than  from  fish or  shrimp (Neff  et  aj..,  1976).
Several  investigators  have  suggested  multiphasic deouration of  toxins by
aquatic  organisms  (Dunn and  Stitch 1976, Herbes  and  Rishi  1978,  Nunes and
Benville 1979, Boehio and Quinn 1977, Blumer et al., 1970, Stegemen and Teal
1977,  Fossato and Canzonier 1976) but  again  none of  these studies dealt
with  short term components as in  this  study.   The depuration  half-life of
14-C  anthracene in  A.  imbecillis had  not  been  reached by  the end of the
sample  period and,  assuming the occurrence of no asore  phase  shifts, the
estimated  half-life  is  approximately two days.  Hultiphasic depuration and
depuration times which  are long  relative to the length of the uptake period
are consistent  with  findings in  the literature.

      In_a study with Anodonta cataractae, McLeese et al. (1979)  found Ky =
0.125 h"  and  K  =  0.015 h    for the  uptake  and depuration of an  organo-
phosphate  pesticide.  While  the depuration  rate  constant is similar to the
K. obtained  in this experiment,  the  uptake  rate constant  in  the  HcLecse
 paper is  much  laiger.   A lowering of K  could possibly be  due to  the ab-
 sence of  food  in  this  experiment.   It can be  argued that  a clam which is
 actively  feeding will  accumulate more  toxic  burden, due  to digestion and
 assimilation of food material laden with the  toxic  substance,  than  a clam
 which is  siphoning  for  respiration only.   This  feeding effect is  unknown
 and deserves further study.  The different natures of the chemicals  used by
 HcLeesc   et  a_l. (1979)  and  this study may   also affect K  .  The predicted
 concentration  from  336 h  exposure   to 31   pg   anthracene'Z  , which was
 the  average  concentration  to which  clams  were  exposed  in the channels
 microcosms,  from  uptake and  depuration rate  constants  reported here for
 laboratory studies  is  1.2  pg  anthracene'g  ,  dry weight  of  soft  tissue.
 This value is more  than an order of magnitude  less than  tl «•  16.7  pg  anthra-
 cene-g , dry weight, actually  observed  in clams collected from  the channels
 microcosms after 336 h  exposure.  The  reasons  for this difference could be
many.  Foremost, is that the  laboratory studies were  conducted for short
periods  of time and may underestimate  the actual  rate of accumulation.  We
observed  clams  in  the  laboratory to remain closed much of the time.  Be-
cause the laboratory studies were done for  short periods of time the clams
may not  have been able  to acclimate  rapidly enough to  give rote constants
representative  of those observed in the field.  We conclude  that  laboratory
determined rate constants may not  adequately  describe  the actual kinetics
of accumulation of PAH  compounds in natural  environments.


Biotrans formation

     After 24 h of  exposure,  little  nf the  ^C-anthracene had been bio-
transformed.  Only 0.9  ±  0.6%  of the l\ had been biotransforaed  after 24 h
                                    94

-------
of uptake.  After  a  subsequent 24 h depuration period only 1 ± 0.8X of the
  C remaining in the soft tissues was biotransformed.

     These findings  are  similar to those of  Dunn  and Stitch's (1976) sug-
gestion that bivalve molluscs  do not possess the  capability to m-»abo)ize
petroleum  hydrocarbons,  and  that  indeed  "the  organism's  only  defense
against toxins is slow depuration" (Nunes and Benville 1979; Carlson 1972).
Stegeman and Teal  (1973) also found depuration  from clams to bs slow, and
accumulation to  be  dependent  on li, id content.   Lake  and Kerchner (..977)
reported that depuration of PHA from molluscs was non-selective.

     It was observed during the uptake portioa of this study that the clams
were sporadically  closing,  and were remaining closed for periods up to 0.5
h.   During depuration,  the  clams did  not exhibit  this  behavior and were
rarely observed  to be closed throughout the sample period.  It appears that
A.  imbecillis  are  sensitive  to either the  14-C anthracene or the acetone
which was  used as a  solvent for the anthracene,  la either case, the abil-
ity  of  these clams  to detect  small amounts of foreign substance  (30 pg'£
14-C anthracene  and  15 pl*£    acetone) is worth  noting.
 CONCLUSIONS

      Sealed and weighted shells of Acodonta imbecillis adsorb 14-C anthra-
 cene in what can be  termed  an  irreversible fashion.  Non-feeding clams take
 up significant amounts of  14-C  anthracene in  24 h, and this uptake may be
 biphasic, although further  study as needed to validate comparttncntalization
 of uptake.  A low uptake  rate  constant,  compared to that from an experiment
 with Anodonta cataractae, raay  be due  in  part to  the absence of food in this
 study.   The  first order  rate  constants  determined from short-term, simple
 laboratory jccumulation studies, result  in an overestimation of the concen-
 trations attained by A.  imbecillis when compared  to  clams in the channels
 microcosms.  This may  be due t<> many  factors,  such  as current, riltcriTg
 time, temperature, suspended particulates or other factors.  The important
 fact is  that the  laboratory derived  rate constants are poor predictors of
 field observations (see Section  6.5).
                                    95

-------
                               •  SECTION 4.6

           UPTAKE,  DEPURATION,  AND BIOTRANSFORMATION OF ANTHRACENE

                   AND BENZO(A)PYRENE IN BLUEGILL SUNFISH


     Direct uptake  from water to  fish  has been shown  for  several  PAH in-
cluding naphthalene, alfcylated naphthalenes, anthracene, and benzo(a)pyrene
Oe
-------
stream within  the U.  S.  Department of Energy's Savannah  River Plant near
Aiken, South  Carolina.  Because  of Its extremely  low  mineral content  (60
umhos'cm  ), the  water was  amended with salts to  produce a standard soft
test water.  Water quality  characteristics  were:   pH  7.4-7.6, hardness A3
mg-«    as  CaCO , alkalinity  30  Dig-*    as  CaC(>3,  -nd  conductivity  140
umhos-cm  .  Well water,  used for  several  uptake  experiments, had similar
quality but lacked the  colored  organic material  found in  the blackwater
stream water  ^nd  was  amended to give  the  same hardness characteilstics as
the amended stream water.  Humic acids (Aldrich Chemical Co., molecular wt.
>  300,000:  3.6 g dry wfml" ) were  added  to  well  water at a  concentration
of 1 mg-£ 1 for certain uptake experiments.
                                                                         14
     Uptake,  depuration,  and biotransformation were  studied  with   9-  C
anthracene   (California  Bionuclear   Corp.,   specific  activity   of   3.3
mCi-mmol   )  and  7-10-  C benzo(a)pyrene  (Amersham-Searle,  specific activity
of 21.7 mCi-maol ).  The  radiopurity  of  the  coippounds vas determined by
 two-dimensional,  thin layer chromatography (TLC)   The  purities of anthra-
 cene  and  benzo(a)pyrene were  97.4  + 2.6  and 96%, respectively.


 Uptake Phase

      Exposures were performed at  23-24.5"C in  covered  5-1 all-glass  aquar-
 ia. All test  water was  aerated,  passed through a  Sorvall SS-1  continuous-
 flow  centrifuge  to  remove participates,  and  spiked  directly with    C  an-
 thracene or  B(a)P.   After mixing,  the  test  water was distributed  to  the
 aquaria.   Fish were,  sorted  by size and  introduced  to the  tanks  at rates of
 100-500 rag  fish-£  .  No aeration or feeding was  provided during testing.
 At appropriate  time   intervals,   fish  were  netted,  blotted dry, wrapped in
 aluminum foil, and frozen at -20°C.


 Depuration Phase

      Fish used for depuration tests were transferred to  clean  flowing water
 after  4  h  exposure.   The  fish   were  held in  5-1 all-glass  flow-through
 chambers  with uncontaminated, unfiltered test  water fit a  flow  rate  of  1 -
 1.3  S.-h  g  Hydrocarbons  eliminated  by  the fish were  trapped on a   75 ml
 Amberlite  XAD-4 resin  column (Rohm  and  Haas) placed  downstream,  as  de-
 scribed by Crosby et  aj..  (1979).   No significant radioactivity was detected
 in the  column  eluent during any depuration  experiment.  The  resin  was
eluted with diethyl  ether,  acetone,  and  methanol according to  the method of
Carna* (1975).   Trapping efficiency  of  anthracene  was  97  -f 3%  and elution
efficiency was 89 +  10%.


Analysis

     Fish  to  be  analyzed were weighed before and after  being  freeze-dried
The  dry weight  to wet weight ratio  was 0.235 (n=45).   All  concentrations
are  expressed on a  wet weight  basis unless  noted otherwise.  Cried fish
vere  conbusted  for  one  minute   in «  Packard  Sample  Oxidizer Model  306.


                                    97

-------
                                                               14
Carry-over during  combustion was  negligible.   Recoveries  of    C activity
from injected dried fish were 46.9% for anthracene and 100% for B(a)P.  The
relatively low  recovery for anthracene suggested  some  loss by volatiliza-
tion during  the freeze drying process.   Therefore,  all  anthracene experi-
ments  were  repeated using  fish  dried  in a desiccator  (recovery 100%)  to
minimize volatilization loss.
                                                                 8
     Water  samples and  solvent  extracts  were counted  in  3a70B  cocktail
 (Research Products international Corp.).   Samples were counted to a pre-set
 time  of  10  min on  a  Beckman LS-8100 liquid  scintillation  counter.  The
 samples  were corrected  for  background,  quench,  and  counting efficiency.
 The quench and  counting efficiency correction  was made using a quench curve
 based  on  the sample-channels  ratio method.  The counting error was SD < 7%.

     Anthracene, B(a)P, and metabolites were extracted from dried fish in a
 40 ml TenBroeck homogecizer.   Each fish  was  ground with  successive 10  ml
 fractions of acidified  benzene, diethylether (twice), and ethylacetate. The
 fractions were decanted,  filtered,  combined,   dried over  anhydrous  sodium
 sulfate,  and the volumes  reduced.  The  remaining fish residue was combusted
 and counted to determine unextractable    C activity.   Recoveries from the
 extraction  of  injected  samples  were  82  + 9% for B(a)P and  93 + 11% for
 anthracene.

      Metabolites  were  determined  by TLC  using E.  Merck silica gel  plates
 without  fluorescent indicator.   Two solvent  systems,  hexane:benzene (9:1,
 V/V) and pentane:ether (9:1,  V/V)  were employed separately  to  resolve B(a)P
 from  its biotransformation  products.   Anthracene  samples  were developed
 either in  hexane .-benzene  (9:1,  V/V)  or  in a two-directional  system with
 pentane:ether  (95:5,  V/V)  as the  second solvent.  Hydrocarbon spots were
 visualized under UV light,  scraped,  and the   C activity determined.

      Extracts   were  also  analyzed by  high pressure liquid chromatography.
 Separations were made with  a tticro-Pak MCH-10 reverse-phase column  (35  cm
 long) equipped with a  Whatman guard  column of  Co-Pel C RODS on 35 pm  parti-
 cles  using  gradient  programmed  elution  conditions  at ;28°C.  The gradient
 was from 45% acetonitrile  in water  to 90% acetonitrile  iu water.  Acetoni-
 trile (90%) was  pumped through the  column for five min before  recreating
 the initial conditions.

      Samples were  detected  by both fluorescence and  254  nm  fixed wavelength
 detectors.   The percentage  of metabolite  was calculated  from the noc^nthra-
 cene activity  in  the  extract, plus  the  unextracted residue determined  by
 combustion.


 Rate Constants  and Bioconcentration Factors

     First-order rate  constants  for KC  activity were calculated, assuming
 a  single  animal compartment,  according  to the first-order model described
 by Branson et aj..  (1975) and  Neely (1979).

-------
              dc                                                       --
              -1 =  Ku  •  Cw '  Kd '  Ca
              dt

                  .                                               (A.8.2)
              C.  . ^ Cw (1 -  e-'V0)
                  Kd

where C  = concentration in fish (ng • g  )

       t =  time (h)

      C  = concentration in water (ng  • g  }
       w
      K  = uptake rate constantjdnl  •  g"1  • h"1) or  (h"  ) assuming a  tissue
       U   density of 1 g  • ml

      K, = depuration rate constant  (h )

 Underlying  assumptions  for the use of this  model are discussed by  Spacie
 and Hamelink  (1980).  The  initial uptake  rate was  estimated  by a  tangent  to
 the initial uptake  curve  (equation 4.8.3):

               C = C  + K   •  C   •  t                                (4.8.3)
                 a    o    u    w
 where C  is  the  initial  concentration in fish.  This was estimated  from a
 linear regression  and  initial rates  assumptions.   For  small values of  t,
 the value of C   is constant  and C   is negligible  (Xpnemann  and vanLecuven,
 1980).  The  apparent first-order  elimination  of    C  activity during  the
 depuration phase was calculated according to  equation  4.8.4.

                loge Cfl  = loge (Ct  -  Kd)        (4.8.4)

 where C  is  the  concentration in  fish at the beginning  of depuration.   The
 elimination half-life  is described  by  equation  4.8.5.
                                                                    (4.8.5)
                        ^
              *• L. —  ^^___^__
                * ~  ~K
                      Kd

 Ideally,  assuming  the model  described by equation  4.8.2, the  bioconcen-
 tration factor  (BCF)  at steady state  las  t  •* ») is described  by equation
 4.B.6.

                       C    K                                       (L ft A>
                        a     u                                       vt.o.o;
                 nrp  —  — -  	
                       C    K
                       Cw   Kd

 Since  the kinetic behavior of hydrocarbons  muy be more complex  than this
 model would  suggest, the ratio  (Ku/Kd) will  be termed  the  "estimated" BCF.
 The  BCF values for anthracene and  B(a)P were  also calculated  according to
 the  regression of Veith et  aj..  (1979b):

                                     99

-------
               loglc BCF = 0.85 logjo P - 0.70         (4.8.7)

which relates bioconcentration to the n-octanol-w«ter partition coefficient
(P)  of  a  chemical  (Tulp and  Hutzinger,  1978).  This  will be  termed the
"predicted  BCF."   The partition coefficients were  calculated according to
Leo et a_l.  (1971).

     The  above BCF  methods ordinarily  relate  to  the  bioconcentration of
parent material, not  to total   C activity which includes biotransformation
products.   To  find  a "corrected" BCF, the  BCF  calculated from the (K/Kg)
ratio  was  corrected  for  the % parent material present  in  the  fish at the
end  of  the 4 h exposure.   The equations  describing the models were fit to
the  data  by the Marquardt numerical methods (PROC NLIN, Barr et^ aK , 1979).
 RESULTS

 Uptake Rates

     Fish  ("mean weight 487  mgl.were exposed  to  20.5  Mg'£~  anthrecene in
 aquaria  for up to  16  h.   The   C activity  in the fish increasedlinearly
 (r   =  0.93) during the first 4 h with a flux  of  1375 + 105 ng  • g"   • h~  ,
 dry  weight. (Figure 4.8.1).  This represents an uptake rate constant (K ) of
 67 + 5 h  . The uptake rate  declined after 4 h  due  to the depletion of  a
 significant  proportion of  the  dissolved  anthracene:  8% after  1  h,  121
 after  4 h,  and 42% after  16 h (Figure 4.8.2).   By  comparison, a control
 tank without fish   lost  only  4.5%  of   initial   activity  in   16 h.  Ad-
 sorption  to  walls represented 0.2% of  the activity.   Accountability for
   C-anthracene  from  water,  walls,  and fish was 96  + 2% and 98.3%  for B(a)P
 after  4 h.  Subsequent exposures were ended after  4 h to avoid  problems of
 hydrocarbon  depletion or water quality deterioration.

     The effect of  exposure  concentration on K   was  investigated for an-
 thracene concentrations of  0.7 to 16.6 pg'£   (figure 4.8.3).   The results
 (Table  4.8.1)  show that K  is essentially independent of the concentration
 of anthracene  in  the water.  This is always assumed to be true but  seldom
 tested.  Another trialj using larger  fish  (5140 tng), at 6.7 Mg-JJ~  produced
 a  similar  K  of 31 h   .  A  slower  uptake rate constant  (K   =  14 h  ) was
 observed in  a  single group of 60 fish exposed to 11 pg-£   "anthracene in  a
 large  60-£ tank.   In general, the fish  in this  group showed less swimming
 activity than those  in the smaller 5-£ test  tanks.

     A-cumulation  of  anthracene  from well water  (Figure 4.8.4)  resulted in
 an uptake  rate  constant similar to  that  of  stream water.  The  addition of
 humic acids  does  not affect  the uptake  rate significantly (p = 0.1).  The
 uptake  rate  constants  for  equimolar solutions of B(a)P  and anthracene in
 stream  water were  similar  (Table  4.8.2).   However,   the  accumulation of
 B(a)P  from well water  was greater  t.han from the stream water containing
 refractory  organics.   The presence  oi humic  icids in  well water signifi-
 cantly  reduced the  rate  of  6(8)1'  accumulation  by  fish  (Figure 4.8.4).
                                    100

-------
          Anthracene
   1800
         0
     Benzo(a)pyrene

180-
                                  TIME (hr)
                               14
Figure 4.8.1.  Accumulation of   C-antlirarcnc and bcnzo(a)|>yrcne from w«-JL

              w.itor with  and  without  humic  acids.  liumics present (solid)

              circles), no htimirs (open circles).
                                   101

-------
      6r
             Anthracene Uptake
x
en
u.
 ui
 o >»
 X
 h-
 z:
 o
                               TIME (hr)
  o
   r
                                                            1.5
                                                     15
"P o
ia m

rm

   E
                                                                  -i
                                                                  m
                                                                  ;u
Figure it.il.2.  Accumulation  of   "*C-niU hrjcene  by  fish  ami  depletion  ol

                C-.intliraconc from water in .1 clor.od system.
                                  102

-------
              12,000
              10,000
           S  8000
           LL.
           z

           LJ
           Z
           Ul
           o
           <
           a:
           x
           i-
           •z.
           <
            i
           o
6000
4000
 2000
Y= 754.6X -15.44

T2=0.9985

n = 12
                    0' V ,,,,,,  i  ,,,.,-,,—,  , ,  -
                      0   2   4  6*8   10'  12  14  16

                     I4C-ANTHRACENE  IN WATER

                                  (/ig/L)
r.vgurc 4.8.3.   Conccnlr.-itioii  of . /  C-.»nt linicene  in  bluo^ills  aflor  four

              hours exposure to   C-.inthr.n cue JK  J  fui:clion  of .inLhr.iccnc

              CoiiccnLraLi on in w.ilcr.
                                 103

-------
      800
oo "o> 600
LL. QJ

     400
o en
CQ x
'  C*
o JE
     200
       0
Without Humics
K|= 27G ng/fj/hr

r2= 0.938
        0
                      •^   With Humics
                             K| = 92.2ng/g/hr
           ~r
            2
                     TIME (hr)
~T
 3
~r
 4
       Acc,,m,,lJtioil Of
       '-  the
                    or
                        10/1

-------
TABLE A.8.1.   EFFECT OF VARYING EXPOSURE CONCENTRATION ON THE RATE  OF UPTAKE
               r«p **r AUTWDAiTMT RV RT.IIFfiTl.T.S*
               OF   C ANTHRACENE BY BLUEGILLS.
Cone, in water
(Mg'*"1)
0.7
1.1
1.9
4.6
8.6
8.9
16.6
Cone, in fish at
, L / -l«b
4 hr (ng-g ;
94 + 4
170 + 9
328 + IS
786 + 26
1439 + 41
1591 + 48
2777 + 98
kl
(h"1)
36
39
43
43
42
40
42
 8 Mean weight 620 mg.

 b Mean + SE.


  Elimination Rates
                                    14
       Both  anthracene and  B(a)P   C  activity  in  fish appeared  to follow
  first-order kinetics  during  the depuration phase (Figure 4.8.5).  However,
  the  elimination of  anthracene residues was  approximately 4  times faster
  (Table 4.8.2).   Elimination  half-lives  for total   C  activity  were 17 and
  67 h  for  anthracene and B(a)F, respectively.   The  rate  constants measured
  here predict  that  bluegills  would reach 90%  of r.heir  steady state concen-
  trations in approximately 68 h for anthracene and 268 h for B(a)F.

       Anthracene  and  metabolites  excreted  by the fish were tr^ped on resin
  columns . placed  below  the depuration  chamber.  Approximately 47% of the
  total   C  recovered  from the colunns was anthracene, according to TLC; the
  remainder was  wore  polar material.   A similar trapping technique for B(a)P
  was not  completed  because of  the low  concentration of  excreted B(a)P in
  water and its' slow elimination from exposed fish.


  Tissue Distribution and Biotransformation

      _^.arger bluegills (4633  mg  mean  wt.,  n = 6) were exposed to either 5.8
  \tg-i    anthracene  or 0.70 pg-Jf!    B(a)P for  4 h to determine   C distri-
                                      105

-------
e> 100
z

E  80

<
2  60
LU
cr

o  40
       LU


       cr  20
       LJ
                                                r = -0.789
              L r=-
1 '
0
1
20
	 1 	
40
-x '
60
TIME

80
(hr)

100

120
                                 14
Figure A.8.5.  Depuration    of     C-.inthr.icene   (open    circles)    and

              benzo(a)iiyrcne  (solid  circles)  followiiiR an  initial A  h
              exposure x +; SE.
                                  106

-------
TABLE 4.8.2.   RATES OF UPTAKE AND ELIMINATION OF   C ACTIVITY  IN  BLUEGILLS
               EXPOSED TO ANTHRACENE OR BENZO(A)PYRENE  (B(A)P)  IN  STREAM
               WATER.
                                     Anthracene              B(a)P
Cone,  in water*                         0.7                    1.0
    (M8-D

Cone,  in fish at                     94 + 4b                207 +  15
   4 h
 kU (h"1)                             36+3°                 49+4

 kD (h"1)                          0.040 +  .006C           0.010 +  .002

 Elimination half-life  (h)            17                      67


   Approximately equitnolar  solutions.

 b Mean + SE.

   Slope + standard deviation of regression coefficient.


 buttons in  tissues  (Table 4.8.3).  Although  the carcaes  had  the largest
 quantity  of both hydrocarbons,  the  gall bladder  contained  the greatest
 concentration on a weight basis.   Liver,  viscera,  and brain all  contained
 similar concentrations of  anthracene.  However, B(a)P  residues  were pro-
 portionately greatest in liver  and  least  in  brain.   The very high rate of
 accumulation of   C-B(a)P in the gall bladder (14,000) suggests that B(a)P
 is rapidly metabolized and  conjugated in  the liver  before  transport to the
 bile.

     The  relative rate of biotransi. ,rmation, expressed as the percentage of
 rT8K?UC /oT/^  PCr hOUr'  WaS  «reater for  B(a>P  than  for  anthracene
 (Table 4.8.4).   The  proportion of non-B(a)P  biotransformation  products
 increased  from 36% after 1  h  to 89% after 4 h.  In contrast, only 7.9% of
 the anthracene activity  had been biotrans formed after 4 h.  The rapid rate
 of biotransfonnation  of  B(a)P indicates that its'  slower  rate  of elimina-
 tion is due to excretion of metabolites rather than simple partitioning of
 the parent material.                                                   e


 Predicted Rates and Bioconcentration Factors
 f onn      /o           COD8tants  Yield  estimated BCF values
of 900 and  4900  for anthracene and B(a)P, respectively (Table 4.8.5).  The
                                    107

-------
o
CO
                TABLE A.8.3.    DISTRIBUTION OF UC ACTIVITY IN BLUEGILLS EXPOSED TO 5.8  pg-fc"

                               0.70 pg-£   BENZO(A)PYRENE (BaP) FOR 4  HOURS.
ANTHRACENE (An) OR
Tissue
Call bladder
Liver
Viscera
Brain
Carcass

% Total
An
3-° h
(0.7)b
3.2
(0.8)
20.3
(1.9)
1.5
(0.4)
72.0
(2.0)
Activity
B(a)P
16.7
(2.7)
7.8
(1.7)
17.2
(1.3)
0.3
(0.03)
58.0
(3. A)
Tissue
ng'S
An
A3, 800
(34,200)
13,000
(1,100)
14,900
(1,200)
12,900
(2,310)
962
(194)
cone.
dry
B(a)P
39,000
(14,000)
4,600
(640)
2,200
(270)
250
(16)
370
(26)
Uptake
Coefficient
An
1890
561
6A1
555
42

B(a)P
14,000
1,600
770
90
130

                 (Tissue cone, at 4 h)/(conc. in water at 4 h).
                 Mean (SE) for n = 6.

-------
TABLE 4.8.4.   RATES OF BIOTRANSFORMATION IN BLUEGTLLS EXPOSED TO  8.9
                 '£   ANTHRACENE OR 0.98 M8'*   BENZO(A)PYRENE (B(a)P).
Anthracene

Cone, of
residue

total
(nmol'g )
I metabolite


1 h
6.2
(1.2)*
3.9
(1.0)
2 h
6.8
(C.3)
4.6
(0.4)
3 h
13.0
(0.4)
7.9
(0.2)
1 h
0.122
(-)
36
(5)
B(a)P
2 h
0.255
(-)
69
(2)

3 h
0.369
(-)
89
(4)
 Biotransformation        0.24     0.16       0.26    0.044     0.088     0.082
 rate (nmol-g  -hr   )     (0.06)    (0.01)     (0.03)   (0.007)   (0.004)   (0.009)

 % transformed per  hr     3.9      2.3        2.0         36        35        22
 * Mean (SE) for  n = 3.
 TABLE 4.8.5.   COMPARISON OF BIOCONCENTRATION FACTORS  DETERMINED BY THREE
                METHODS.a
Method
Predicted
(partition)
Estimated
Corrected
(parent only)
Anthracene
1209
900
675
B(a)P
28,250
4,900
490
   See Methods section for definitions.

   Using log P = 4.45  for  anthracene  and  6.06  for  B(a)P  (Leo,  1975).
                                     109

-------
estimated value for anthracene  is reasonably close  to  its  predicted value
of 1209, based  on  partition coefficient.  However, the estimated value for
B(a)P is  considerably lower  than that predicted  from  the  partition coef-
ficient.   Since the  regression of  BCF versus  log  P equation  is derived
primarily from inert or poorly metabolized chemicals (such as the polychlo-
rinated  biphenyls),   it  reflects  an accumulation  process  dominated  by a
reversible exchange of  the unaltered material.  B(a)P is a clear exception
to  that pattern  because  of  its rapid  conversion to other  products.   In
fact, a true steady state  condition between fish and water may never occur.

     The true  corrected  BCF values  for both  anthracene  and B(a)P are well
below  the  other estimates  (Table  4.8.5).  The  % biotransformation product
(Table  A.8.A) used to calculate  the  correct BCF may vary over time, depend-
ing  on the length of exposure,  enzyme  activity, or other  factors.  If so,
the  corrected BCF values based on parent material  would also change.

     Although  there  are many BCF values  in  the literature for comparison,
 relatively few  rate  constants  have been  reported for  fish.   Spacie and
Hamelink  (1980)  reported  correlations  for K  and Krf  versus  log P  for a
 series of chlorohydrocarbons and other organics.  The correlations predict
 a slight increase in K  and  a  relatively large decrease in K. with Jugher
 log P.   The predicted" K  values,  19  h   for  anthracene  and  33 h   for^
 B(a)P, give reasonably  good agreement with  the measured  values of 36 h
 and 49  h   .   The predicted  K.  values  based  o . partition coefficient are
 0.02 h   for  anthracene  and 0.004  h   for  B(a)P  .   They suggest slower
 rates  of  elimination and longer half-lives  than  were  actually  observed.
 DISCUSSION

      Anthracene and  B(a)P  were removed  from water at  similar rates  from
 equimolar  solutions.  The  rate  constants are proportional to  the  octanol-
 water  partition coefficient.  This  reflects  the tendency of more  nonpolar
 components to partition  out of water (Chiou et  al. ,  1977;  Leo et al.,  1971;
 Southworth et  aJL,  1978;  Veith  et  aj..,  1979b;  Kenaga  and GoTiog,  3980).

      Similarly, the elimination rite  constants  were inversely  proportional
 to the octanol-water partition coefficient.   However,  this  cannot be  total-
 ly attributed  to the partition  coefficients  since  significant  biotransfor-
 nation occurred, particularly for B(a)P.  The  foruation of macromolecule-
 bound and poLa^r biotransformation products will tend  to  reduce the depura-
 tion of  the    C label.   The  bound  material  is  not free to dissociate  and
 the polar components may,  in many  cases,  be more slowly eliminated  (Landrum
 and Crosby,  1981a).   Slow elimination  of polar components by fish  during
 these studies may have been  enhanced  if concentrated  metabolites  were  not
 being emptied from the gall  bladder due  to the  absence of  feeding  (Lech et
 al.,  1973).  The relative  tissue  distribution of anthracene and biotrans^
 formation products  observed  in  the  bluegill  sunfish  studied  here  were
 similar to that observed  for anthracene in  young  Cobo  salmon (Roubal  et
 al..,  1977).   In the  study with salmon,j|all bladder was the  tissue  contain^
 ing  the  greatest  concentration  of   C-anlhracene  and  biotransformation
 products.
                                     11J

-------
     The bioconcentration  factor  predicted from the  log  P relationship is
asuch greater  than that determined  from K /K , and much  greater than that
corrected for biotransformation.  This hasUalso been observed by Southvorth
et  al.  (1980) for ezaarenes  in the fathead minnow and by Leversee et aK
(T981b) for B(a)P  in Chironomus riparius.  Fish show this discrepancy du«
to  the  relatively  slow uptake  in proportion to the biot:.«nsfonnation rate.
Thus, predictions  of BCF  from  the  log  P  ran be expected  to be higher than
those  actually  found  for  compounds that  undergo  significant biotransfor-
nation.

     Differences  in  the partition coefficients of the anthracene and B(a)P
nay account  in  part  for  the differences  in  observed   biotransfonnation
 rates.  Since B(a)P  has a faster  uptake  rate  from  equimolar solutions,  a
 greater concentration  should be  available  for  biotransfonnation.  Addi-
 tionally,  B(a)P can  be expected to be bound to serum proteins.  This would
 result in  lower whole-body concentrations and greater relative distribution
 to the liver.   This  could  result in greater biotransfonnation, assuming the
 enzymes for biotransfonnation are not  saturated.  The biotransfonnation of
 B(a)P  by  fish  on a  dry  weight basis was less  than  that found for either
 Dapfania or  Chironomus riparius  (Leversee et  al.,   1981  a  and b).  This
 reflects the comparUnentalization  of  the biotransformation enzymes in the
 fish.

      The  rate  of biotransformation  appeared  to  increase in  the  case of
 B(a)P  between  1  and 2 h.   While  B(a)P is a  known cytochrome P-A50 inducer
 (Ahokas, 1979), the  short exposures  (A h) precluded the  synthesis  of more
 enzymes. However, adsorption of  B(a)P or  its metabolic products by endo-
 plasmic reticular membranes  may alter  the specific activity of  the  enzymes.
 Such  an  apparent induction  would  affect K. by  altering the pool  size of
 parent and  biotransformation  products.   Induction  would have  its major
 effect on K.  if  parent  compound  were  the  major excretion  product and  a
 minimal effect for   the excretion  of  biotransformation products (Lech and
 Bend,  1980;  Spacie  and Hammelink,  1981).  The  length of  exposure will tend
 to decrease the depuration rate constant, however, the uptake rate  constant
 will  not  be  affected by  this induction.  The  decreased depuration rate
 results from the accumulation of  a bound, unexchangeable,  biotransformation
 product pool.  This  irreversible  binding has  been described  for PAH as the
 mechanism  for  toxic  response.   Irreversible binding was observed in the
 kinetics  of anthracene   in  Chironomus  riparius  (Gerould et  al.,  1981).

      Thus, models used in describing  the kinetics of uptake  and depuration
 of PAH become biased when there is significant  biotransformation, induction
 of  PAH transforming  enzymes, binding  of  biotransformation  products,  or
 translocation.   In  this study, we  used ehort-tena exposures  and depuration
 periods to  estimate  rate constants.   This avoided many of the  problems
 mentioned  above.  Also,  we   chose  to  use simple one compartment,  constant
 infusion models instesd of two compartment mass balance models.   The deple-
 tion of anthracene from water after 4  h in the  static uptake  study  could be
 corrected  for  by a  two  compartment  model.   Howiver,  the other  problems
 mentioned  above would  have  resulted in  a bias  for  both K  and K  .  Thuu
 the creation of a multi-compartment model  would be raost  appropriate, how-
 ever,  these models require large  amounts of data for curve  fitting.'Since


                                     111

-------
the rate  constants  are highly  correlated,  many roots  of the equation  re-
gression solutions (toy not result in real values or mechanistically  realis-
tic rate constants.   Therefore,  the use of simple models which fit one  rate
constant at a time provide good approximations to the overall processes  for
the fate of xenobiotics.

     Veith et  al.  (1979a)  suggested a technique  for  determining the log  P
or partitioning coefficient for organic compounds.  This technique uses  the
retention  time  of   the compound  on  reverse-phase,   high-pressure   liquid
chroaatography  columns.   This   technique  makes  the  measurement  of  this
secondary property of compounds accurate and rapid.  Since the partitioning
coefficient  is  defined  by several properties of compounds, such as  molecu-
lar  weight aud polarity,  it has been suggested  as  a possible integrator,
which  would  be highly correlated with the bioconcentration factor (BCF) in
fish  (Leo et al.., 1971; Chiou  et  a_l.,  1977).   However,  the  results of  our
studies  indicate  that the BCF  predicted from the log P is greater than the
observed  BCF.   This  result is  in agreement with those of Southworth et  al.
 (1980) for azaarenes  in fish and is most probrbly due to biotransformation.
Therefore,  if  organisms  are  able  to  biotransform a  compound  of interest
 such as bluegills can biotransform  B(a)P and anthracene,  the  use of  the  log
 P to predict steady  state  concentrations may result in error.

      The biotransformation of  PAH  compounds,  such  as B(a)P and anthracene
 in fish,  has been widely  studied  (Pedersen e_t al., 1974;  Payne and Penrose,
 1975;  Gerhart  and  Carlson,  1978;  Schnell et  a_l.,  1980; James  and Bend,
 1980).  Because the  microsomal mixed  function oxidases can be induced,  the
 rate  of biotransformation can  change with  time  of exposure.  As discussed
 earlier, if the biotransformation  products are bound  tightly to macromole-
 cules, the  apparent BCF will  change  as a  function  of  time  dee to:   (1)
 induction of enzyme  activity;  (2)  changes in the  ratio of biotransformation
 products  to parent  compound;  and  (3)  redistribution of both  parent  and
 biotransformed compound.  Therefore,  the  short-term pharmacokinetic tech-
 nique  for determining  steady-state concentrations  of organic compounds by
 fish  which  was proposed  by  Branson et. al.  (1975) is appropriate only  when
 biotransformation is minimal  or the magnitude and  rate of biotransformation
 is well known and constant.

      The  uptake  rate  constant for B(a)P was  altered by the  presence of
 dissolved humics.  This reduced bioavailability was proportional to  the  log
 P, indicating  that  the more  nonpolar  materials  interact  strongly with  the
 dissolved humics.   Similar  results have been  observed for  the  uptake of
 hydrocarbons by sole (McCain et  a_l. ,  1978) and for selected  PAH by  Daphnia
 (this  report,  Section 4.3).                                          —	
                                    112

-------
                                  SECTION  5

                  SIMULATION MODEL FOR PREDICTING PAH FATES

                                 SECTION 5.1
INTRODUCTION

     It  is  desirable to be  able to  predict the concentrations  of  PAH to
vhicb aquatic organisms will be directly exposed and humans will be exposed
through  food and  drinking  water.  Thousands of different  species of PAH's
are  chemically  possible;  therefore,  application  of elaborate  screening
protocols  (Duthie,  1977) to  individual  PAH in  order to  quickly estimate
major processes of transport, accumulation, and degradation are impractical
for  purposes of  risk analysis.  Assessment of health risks associated with
introduction of  PAH  into the environment depends in  part  upon quantifica-
tion of  environmental   transport  and subsequent  concentrations  to  which
aquatic  organisms are exposed (Crawford and Leggett,  1980).   Estimates of
transport  and  dose are  independent of the specific nature of the threat to
human  health,  either directly through contamination  of  potable water, for
example,  or indirectly  through  damage to  ecological  life  support systems.

     Several  alternatives  exist  for  the development  of needed predictive
capacity.   Monitoring PAH  content of stream components of interest follow-
ing  accidential  pollution  by PAH should provide sufficient data  from which
to  generalize,  given enough accidents.   Statistical analysis of  these data
could  facilitate prediction by  providing  regression  models or correlation
between  measurable properties  of specific PAH compounds and their observed
fates  in  streams  and rivers.    Statistical  models,  however,  depend solely
upon available data.   In addition, coefficients in statistical models often
cannot be  translated into mechanistic processes subject  to control or mani-
pulation by environmental decisionmakers.

     Simulation  models  based  upon mechanistic processes  of PAH flux and
accumulation  represent  a  potentially more useful  alternative  to environ-
mental   monitoring  or  statistical  models.   These  m dels  are  based  upon
"state  of  the  art"  understanding  of  processes,  mathematical  analogs of
which  translate  directly into parameters or coefficients that can be esti-
mated  in  the  field  or  laboratory.   Data  collected  in monitoring programs
remain useful for model  validation.

     This  section describes a model (Fates of Aromatics Models, FOAM) which
simulates  the transport  of PAH co.j/ounds through lotic systems.   Similar to
models  developed by  Baughman  and Lassiter (1978) this  model  has been de-
signed to  predict fates of PAH  in 
-------
ever, model  structure permits  parameterization for  simulation of  PAII  dy-
namics it* specific streams.
                                    114

-------
                                 SECTION  5.2

                         MODEL  STRUCTURE  DESCRIPTION


     Simulation of PAH transport in lotic systems requires an understanding
of  basic  structure and  function of streams  and rivers,  as well  as,  the
physical,   chemical and  biological  processes  that  govern PAH  movement.
Basic ecological  information concerning  species  composition, standing crop
and, in some instances, energy flow has accumulated for a variety of stream
systems (Coffman  et al.,  1971;  Fisher  and  Likens, 1972;  Mclntire,  1973;
Mclntire  and  Phinney,  1965;   Minshall,  1978; Odum,  1957; Teal,  1957  and
Tilly,  1968).  Yet  streams  have not received the degree  of attention from
ecological  modelers as have other aquatic systems  (Patten, 1968).  Models
of  lotic  systems  have  classically been the purview of sanitation engineers.
State  variables in these models typically included dissolved  oxygen, bio-
logical  oxygen demand,  ana  total  solids.  Model  form  typically divided a
particular  stream or  river into  a series of reaches, which  individually may
be  quite  different  in hydraulic characteristics,  but  were assumed  to be
 internally   homogeneous.   Reach models  have become  sophisticated  in  the
 realm  of  sanitation engineering, but  little has been done  to enhance their
 application to increase basic ecological understanding of  stream structure
 and function.  However, ecologists have  recently  begun to simulate energy
 and material  flow  through lotic systems, in some cases borrowing the reach
 model  approach (Chen  and Wells, 1976; Knowles and Wakeford, 1978; Mclntire
 and Colby,  1978;  Sandoval  et. al.,  1976;  Zalucki, 1978).

      The  major objective  of  this  modeling effort  wjs  to predict the pat-
 terns  of flow and accumulation in aquatic  systems  based on easily deter-
 mined  parameters  for  the  compound of interest  and the receiving environ-
 ment.   In  this  study we  used  laboratory  studies  to ascertain  the most
 important vectors  and mechanisms  to  be  included  in the simulation model.
The effects of potentially  mitigating  factors  on  pathways and  rate con-
 stants  were also investigated.   In some cases  overall rate constants for
uptake,  depuration  and biotransformation by biotic and  abiotic matrices
were  determined  in the  laboratory for  use in  the anodel.   Finally,  the
simulation  model  was parameterized for  the conditions existing in the SREL
channels  microcosm  faciiity   (Ciesy  et  *±.,  1979;  Bowling et  a]..   1980).
Simulation  results were then compared to those  of  inputs of anthracene in
the channels microcosms.

     The  general  strategy was  to  adopt  the  reach  approach (Figure 5.2.1)
and to  include fundamental biological production processes common to lotic
systems.   Specific processes  known to  influence  PAH  flux somewhat inde-
pendent of ecological  processes  include photolytic degradation, volatiliza-
                                    115

-------
                         GENERAL  REACH MODEL
                   Segmented
                    Street
                                              REACHi
Figure 5.2.1.   Conceptualization  of  Keneralizc.il  stream  reach  transport
              model.
                                 116

-------
tion losses  and  sorption  to  sediments and  suspended  participates.  These
processes were, therefore, included in the model.

     The fate  of  aromatics model (FOAM) predicted concentration  of PAH in
phytoplankton, "periphyton,  macrophytes, zooplankton,  two benthic  inverte-
brate compartments, bacteria,  sediments,  suspended particulate matter, and
fish (Figure  5.2.2).   The overall simulation model  is  organized  as a main
program  (MAIN) which  calls  subroutines  which  calculate  solar  radiation
(SOLAR)  movement  of water (HYDRO), solubility of PAH, and resuspension and
calculates  and graphs the relative  and absolute mass of  PAH in  each com-
partment (Figure  5.2.3).   Other submodels calculate the photolysis (PHOTO)
volatilization (VOL)  and  sorption to  sediments  (SORP).   The biotic compo-
nents are coupled as in the representation given in F:.gure 5.2.4.

     FOAM   is  programmed  in   FORTRAN.   A modular  structure  was  employed
wherein  individual processes  were   separated  into  subroutines  (Figure
5.2.3).  For example,  subroutine HYDRO generates the hydraulic history used
 in subsequent calculations of state  variable dynamics.  Likewise, subrou-
 tine SOLAR  simulates  hourly values of  incident solar radiation.   The physi-
 cal-chemical processes importaut in PAH  dynamics:  photolysis, volatiliza-
 tion and sorption,  are modeled in routines PHOTO, VOL and  SORP, respective-
 ly.  Biological  equations are  contained in routine  EQUA.   Each routir"
 calculates  part of the overcll  derivative  for  change in biomass or PAH in
 each state variable  (eg. equation 5.2.1).   Modular -structure facilitates
 modification  of  the  submodels  as  directed by comparison  of  simulation
 results with observations to  be made  in the  experimental  streams.
 Model Output

      FOAM simulates the  dynamics  of  biological  production  and  PAH  dynamics
 through time and space (reach) (Table 5.2.1).  Each  state variable  value  is
 printed in matrix form where row elements  are  values at  a specific  time for
 each reach.   Column elements are temporal  values in  a particulsr  reach.   To
 assist  in  recognition  of  complex patterns  in  spatial-temporal dynamics,
 each  state  variable  can be  plotted sirau)taneojsly  against reach  number
 (distance downstream) and time  (hour)  to  provide a  three  dimensional fig-
 ure.  A  complete listing of  the model code  is beyond the  scope  of this
 report but  can be obtained from the authors.   It should  be  noted  that model
 development  is a dynamic process and  FOAM  is continually being  updated.   An
 overall  mass  balance approach was taken  to model changes of biomass and PAH
 through  time  and space.

      The first  term on  the  right hand side  of equation  5.2.1  represents
 time dep-ndent changes in PAH concentration that result  from  processes such
 as volatilization, photolysis,  biological  transformation and  translcration,
 and  sorption  to suspended particulate matter or  sediments.  The second term
 on  the right  hand side  of equation represents downscream convective  trans-
 port.

     Ths  general form of  equation 5.2  1  implies independence between time
 dependent change and ccnvective transport.  However,  as will be  seen, cur-


                                    117

-------
                     SIMULATION OF PAH FLUX THROUGH STREAMS
                               VOL*TI'.lI*TIOM

                       PHOTOLYSIS
                             SEDIMENT
    REACH .,
                                    REACH
REACH
Figure  5.2.2.  Srhcm.it.ic   rcprcsontaLion  of  F«iton  of  Aroni.itic  Molecules
                (FOA11) model.
                                       113

-------
                  MODULAR  PROGRAM  STRUCTURE
                                Inputs
                FATES  OF  AROMATICS  MODEL (FOAM)
Figure 5.2.3.   Schematic representation of modular subroutine  st
              FOAM.
ructure of
                                 119

-------
         t=0



     uj    t
     2
     I-

          t-t-l
             x=0
                       DISTANCE
x+l
Ax
Figure 5.2.4.  Movement of w.iter m.iss Lhroii);h sp.icc and Lime.
                               120

-------
Table 5.2.1.  Output from FOAM.  Each variable  is  simulated  hourly.
Irradiance                                           langleys  •  h      _2
Alloch;hanous organic matter                         8 •  drv w*2gbt '  m
Exteraal PAH loading                                 8 PAH-2 m
Periphyton biomass                                   g *  ">
Macrophyte biomass
Benthic  Invertebrate Biomass
Clan biomass
Bacteria biomass
Fish biomass
 Suspended  particulate  organics
 Dissolved  organics
 Bottom Sediments
 Suspended inorganic particulates
 Settled detritus                                      "   _j
 PAH in periphytoc                             M»°l PAH^ g  , dry weight
 PAH in macrophytes
 PAH in Benthic Insects                                "
 PAH in Clans                                          "
 PAH  in Bacteria                                       "
 PAH  in Fish
     Predacious                                         "
     Herbivorous                                        "
 Suspended organic particulates                        "
 PAH in  dissolved organic matter                       "
 PAH in  bottom sediments                               "
 PAH in  suspended inorganic particulates               "
 PAH in  settled detritus                               "
 PAH dissolved in water                            pmol PAH • t~
 PAH which has been transferred                        "
  rent velocity directly affects some of the rates of time dependent process-
  es, for example, rate of volatization of dissolved PAH or rates of settling
  and resuspension of particulate PAH.   Time and  space  are to an extent  the
  same thing coupled by current velocity.

      Similar  to  Sandoval et  al.  (1976), a numerical  approximatiou to  the
  solution  of  equation 5.2.1  was used wherein  the (x,t)  plane was divided
  into  an  array and  the  derivatives  approximated  by  differences  (Figure
  5.2.4).
                             P(x,t+l)
                                                                   5.2.1
                                     121

-------
                      8P    P(''t*1)  "  P(x>t)
                      3t~ =        At5.2.2

     An element  in  the  array  represents  the  PAH concentration  in  a given
compartment,  say periphyton, at a  particular  point  in time and space.  The
actual value is  the  sum of the time dependent  change (vertical arrow) and
convective change (horizontal  arrow).   In  another sense,  the vertical arrow
represents — in equation  5.2.2 and  the horizontal arrows,  actually their
sum, represents U.    „—   in  equation  5.2.2.   The time step was determined
                 (x,t>dxn
by a necessary  relationship At =  l!/Ax  = 0.01 H • S   between reach length
(M, 18.28 ro) and current velocity  (S  ) (Bella and Dobbins, 1968).
                                   122

-------
                                 SECTION 5.3

                                  SUBMODELS
Hydrological Sub-model

     Application  of  the model  was constrained  to conditions  of constant
reach morphometry under  a  constant discharge regime (Figure 5.3.1).  Reach
dimensions  and  flow rates  are  summarized in the  facility description.  A
simple  continuity equation  was  used to determine current velocity in rela-
tion  to constant sectional  area of each  reach  and discharge regime (equa-
tion  5.3.1).
                                   Q = V  • A         (5.3.1)
 whe re,
                                  3-1         3-1
       Q = discharge  into  reach,  (M  -t  ), 4.53 m -h

      V  = current  velocity  (m't  ), 18.3 m-h

                                          2         2
       A = cross  sectional area of reach (m ), 0.12 m

 This hydrologic submodel appeared  reasonable for  the artificial  streams,
 since the  channels  were  of constant  length,  width,  depth  and flow  rate.


 Temperature Dependent Processes

      In FOAM water  temperature  indirectly affected  flow of PAH's  through
 the food web by  directly modifying  rates of photosynthesis,  consumption  and
 respiration.   An empirically derived equation (O'Neill  et al., 1972; Shu-
 gart et  aj..,  1974) was used  to calculate the effect of water temperature in
 reach j  at time t on rate  estimates  of photosynthesis and  respiration  for
 biological  state variables  (equation 5.3.2).

                           f(T) = V* . ex(1-V)               5.3.2

 where

     V =  (Tm - T)/(Tm - To)
                                    123

-------
                U-2AI =AX
                              U = m/time
                              t = time




ltllllt®tllil
=^§lf§; fcs5j§§;5 ;
^S^^S?'




X- meters


-
         t+At
         t+2Al
                             X          X+AX
                               PLUG  FLOW
Figure 5.3.1.
Schrm.itic  r -prcsrntat ion  of  plug
w.itcr movement in reach model.
                                               flow  used  to represent
                                 124

-------
    X =  V2  (1  + (1  + 40/Y)^2/400

    W -  (ln V {T» ' To)

    Y =   Ce-'V          5.3.3

where

       I_ = radiation  at  surface  of reach (ly/min)

      !_._ = direct  beam solar  radiation, slope dependent
       XxO
       is = scattered  clear sky solar radiation

      I,.- = direct  beam radiation on  a horizontal surface

        C = percent cloud over

        F = slope correction factor for scattered clear sky radiation

        d = parameter  for seasonal cloud effects

Model  inputs included latitude of stream  (33  degrees  N),  slope direction,
slope  inclination  (0) and daily estimates of mean  cloud  over.   The model
was  calibrated with  measurements  recorded in the vicinity  of the streams
site.   Figure  5.3.2  represents  a simulation of the annual  cycle of solar
radiation.

     FOAM includes  the capacity  to  attenuate surface light with depth as a
function  of  suspended  sediments and  phytoplankton  biomass.   Hyperbolic


                                     125

-------
Table 5.3.1.  Parameter estimates for consumer components  of  fates  of  aromatic model  (FOAM)
              See Table 5.3.3 for explanation of abbreviations.
Parameter
Component

Zooplankton
Benthic Invert
Macro Invert
Bacteria
Carnivore
DMAX
(h'1)
0.0375
0.0150
0.0125
0.0042
0.0021
Q10C

2.0000
2.0000
2.0000
2.0000
2.3000
TOPIC TMAXC RMAX
(°0 (°
25.000 30.
27.000 32.
25.000 30.
27.000 32.
27.000 31.
'0 (h'1)
000 0.0042
000 0.0042
000 0.0042
000 0.0083
000 0.0010
Q10R

1.8500
2.8500
1.8500
1.8500
2.1000
TOPTR TMAXR
(°0 (°0
30.000 35.000
32.000 37.000
30.000 35.000
32.000 37.000
30.000 34.000
•*

Component

Zooplankton
Benthic Invert
Macro Invert
Bacteria
Carnivore
EXCR
(h'1)
0.0042
0.0042
0.0042
0.0083
0.0021
MORT
(h-1)
0.0001
0.0001
0.0001
0.0001
0.0001
EGEST
(h-1)
0.0042
0.0042
0.0042
0.0042
0.0042
LIPID
(%)
0.13
U.13
0.13
0.13
0.13
DMAX
(h-1)
2.4xlO~5
0.00
0.00
0.00
0.00
QCi KP
(h'1) (g-m'2)
4x10 1.00
0.0004 1.00
0.0004 1.00
0.0004 1.00
0.0004 1.00

-------
Table 5.3.2.  Parameter estimates for primary producer components of fates of aromatic model
              (FOAM).  See Table 5.3.3 for explanation of abbreviations.
Component
PSMAX
                           Q10P
                   TOPTP
TMAXP
RMAX
     Parameter

Q10R    TOPTR    TMAXR
                                                                        EXCR
Phytoplankton   0.0080   1.4900   20.000    35.000   0.0021   1.8000  30.000   37.000    0.0008

Periphyton      0.0060   1.4900   20.00     35.000   0.0021   1.8000  30.000   37.000    0.0008

Macrophytes     0.0050   1.4900   20.00     35.000   0.0021   1.8000  30.000   37.000    0.0008

Component MORT SINK RK
(h"1) (h'1) (FC)
Phytoplankton 0.0001 0.0021 5.
Periphy'.on 0.0001 0.0 5.
Macrophytes 0.0001 0.0 5.
LIPID DMAX QPi KP
(%) (h"1) (h"1) (g-m"2)
0.15 0.00 0.0042 l.OC
0.15 0.00 0.0042 1.00
0.15 0.00 0.0042 1.00

-------
        CC  61
        O
        I
        UJ  54
        O
        <
        S  47
        W  39
                 -T	1	r
                  37    74    110   147   IB3   219   256  292  329   365
                                 DAY  OF YEAR
Figure 5.3.2.  Simulation of solar radiation by sub-modrl r.oljr.
                                     128

-------
Table 5.3.3.  Key to parameter abbreviations used in FOAM  for  consumer  and
              producer components.
DMAX - Maximum depuration rate
Q10C - Temperature dependence for  consumption
TOPTC - Optimum  temperature  for  consumption
TMAXP - Temperature of maximum photosynthesis
RMAX  - Maximum  respiration  rate
Q10R  - Temperature dependence for respiration
TOPTR - Temperature of maximum respiration
TMAXR - Temperature of which respiration  goes to  zero
EXCR  - Rate of  execution
MORT  - Rate of  mortality
EGEST  - Rate of  egestion
LIPID  - Fraction lipid  content
QC    - Maximum  PAH uptake  rate  by consumers
 KP    -  Half saturation constant for uptake  (half of max solubility)
 PSMAX -  Maximum photosynthesis  rate
 Q10P  -  Temperature dependence  for producer
 SINK  - Sinking rate of phytoplankton
 RK    - Light saturation of photosynthesis
 TMAXR - Temperature upper limit for respiration
 QP    - Maximum PAH uptake rate by producers
 PUC   - Uptake  rate (flux) of PAH by consumers
 PUP   - Uptake  rate (flux) of PAH by producers
  functions provided partial extinction coefficients for each component which
  were  summed to  estimate  an  overall  extinction rate used in a  first order
  exponential decay function.  Depth in past simulations was 0.21  m.   Attenu-
  ated  light  drove tLe photosynthetic submodel for periphytou.   Surface light
  intensity   drove  the  phytoplankton  photosynthesis  submodel.    Photolytic
  degradation of PAH was calculated with light intensities at several depths,
  integrated  over  the water column.
 Photolytic degradation of PAH

      Light  dependent  degradation  of  aromatic  hydrocarbons  represents  a
 potentially  important  mechanism  for  physico-chemical  transformation  of
 these  compounds  in  aquatic  systems.   Direct  photochemical  breakdown  of
 dissolved PAH  and degradation  of  photosensitized PAH sorbed  to  suspended
 particulates may both be iraport.ant, however,  Zepp and Cline (1977) indicate
 that indirect photolysis  of  sorbed PAH is minor  ia  comparison  with direct
 photolytic degradation of dissolved  PAH.   In the current  version of FOAM,
 only direct  photolytic degradation  is simulated.
                                     129

-------
     Incident  light  intensity,  R ,  simulated by  equation 5.3.3 is attenu-
ated by  suspended particulates via a  diffuse  attenuation coefficient  cal-
culated  from  a relationship (equation 5.3.4; based on published data  (Zepp
and Cline, 1977).
                                               t)
                  K  = 0.20 +  0.00247 •  SS  (r   = 0.77)               5.3.4

where

      K  = diffuse attenuation coefficient, cm

     SS  = suspended  sediments, mg • £

     In  the subroutine that calculates phctol/tic degradation  of  PAH, dis-
 solved  PAH,  assumed -available for photolysis,  is  converted from units  of
 grams  per  square meter  to moles  per liter for use in  the equation  for
 photolytic  loss (equation 5.3.5) adapted  from Zepp and Cline (1977).

                     d PAH                                         5.3.5
                        /  =  * ^      = yield coefficient

      k.   =  light absorbed at  wavelength X and
       A
      PAH. =  dissolved PAH concentration
         d
      Reported  yield coefficients  range  between  0.001 and 0.01  for polycy-
 clic  aromatic  hydrocarbons (Zepp  and Schlotzhauer,  1981).   A linear  re-
 gression  based  upon data presented  by Zepp and  Schlotzhauer (1981)  was
 calculated to relate yield to molecular weight  of  PAH compounds with  equa-
 tion 5.3.6  (Figure  5.3.3) <)> =  0.0235 -  8.38x10   • molecular weight (r =
 -0.62).

      The value of k. reprerents light absorbance at  wavelength X by the PAH
 compound.  Values oT absorbance  (photons •  cm   • sec"  )  were  summed  at 10
 nra  increments  between  wavelengths of  300  and  500 run,  a  region of peak
 absorbance and  intense  photochemical  breakdown  for  PAH  compounds. Each k,
 was calculated  as  the  product ~>f  simulated light intensity and a measured
 molar extinction coefficient,  e ,  at lembda.   Values of  e. must be supplied
 as input to the model by the user.

      Hourly  light intensity simulated by subroutine solar is attenuated as
 a function of wavelength following some assumptions concerning the quality
 of light impinging  upon  the surface of each reach  in  the stream  of  inter-
 est.   First,   we  assume  that only  12% of the simulated  light intensity,  R
 is within  the 300  to  500 nm  range  (Zepp and Cline, 1977).   According  to
 data  presented  in Miller and  Zepp  (1979), we  determine  the relative  compo-
 sition   of  light at each  10  nm increment  between  300  and 500  nm  (Table
                                     130

-------
              -1
UJ
O
U.
U_
LJ
O
O

O
_l
UJ
              -2.
           CD
           O
           _J
               -3
                                        o
                "100              200
                       MOLECULAR  WEIGHT
                                         300
Figure 5.3.3.  Log  ns a function of 1'AII moJociiJnr weight.
                                131

-------
       Table 5.3.4.  Contribution of light of 10 nm increments to
                     total spectrum between 300 and 500 nm.
Wavelength
(nm)
300
' 310
320
330
340
350
360
370
380
390
400
410
420
430
440
450
460
470
480
490
500
Per Cent of Total
0.0050
0.1009
0.2749
1.7309
1.9876
2.1307
2.2715
2.5441
2.8167
3.3619
4.8157
6.3376
6.5193
6.2922
7.4279
8.3593
S.'iSOl
8.7227
8.9499
8.4501
8.6319
5.3.4).   Twelve per cent  of  R  is multiplied by the appropriate  value  in
Table  5.3.4  to estimate the incident light at each wavelength.   Then,  t'jis
quantity  is  attenuated  according to an equation fit to  the  curve of light
attenuation  versus  wavelength (equation 5.3.6) for  12  southeastern rivers
(Zepp  and Cline, 1979).


            ax=0.42e(-4-4366xl0~3-X), (r2 = 0.96)              5.3.6

where  a.  has  units  of cm  .  A depth specific rate of photolysis is calcu-
lated  at  1.0  cm  intervals until  either  the depth  of  the  reach  has  been
equalled or the depth specific rate is less than 10% the rate calculated at
the water surface.   The depth specific rates are  then  integrated over the
water  column  and  converted to g2PAH  •  m   • h  .   Initial light intensity
was  converted  to  photons   •  cm   •  sec    from  lyh~   by calculating the
joules per Einstein  for each  wavelength \  between  300  and 500 ma at 10 nm
increments.   While  as  a  constant  for a
                                    132

-------
given compound over all wavelengths.   Differences in photolytic degradation
of specific compounds, as simulated by FOAM,  occur as a  result of a  molecu-
lar weight dependent value of <}> and different absorbance characteristics of
each compound  in the 300 to 500  nm region of the  spectrum  (Table  5.3.5).
Photolytically degraded hydrocarbons  were  added  to the  pool  of transformed
PAH.
Table 5.3.5.  Comparison of observed quantum yield coefficients for
              reaction of PAH in air saturated water to predicted quantum
              yields based upon regression with molecular weight.
Compound
Naptbthalene
Phenanthrene
Anthracene
9-Methylanthracene
9 , 10-Dimethylanthracene
Chrysene
Naphthacene
Benz (a) anthracene
Benz(a)pyrene
Mol. Wt
128
178
178
192
206
228
228
228
252
regression based upon data in Zepp and
omitted from regression
r2 = 0.87
Volatilization of PAH
analysis, log <(>


a
Predicted
.0219
.0075
.0075
.0055
.0041 "~""~
.0026
.0026
.0026
.0015
Schlotzhauer (1981)
= -0.465 - 0.0093 • mol.


Observed
.0150
.0100
.0030
.0075
.0040
.0028
.0130
.0033
.009

wt.


      The  submodel that quantifies losses of PAH  through  volatilization  in
 reach j at  time  t was adapted from Southworth (1981).   The rate of volatil-
 ization was modeled as a function of current velocity, wind velocity,  reach
 depth and  molecular weight  of the PAH according to  a  hyperbolic equation
 involving Henry's Law (equation 5.3.7).
                                   V(H
5.3.7
                                    133

-------
where

     H   = molar  concentration  of PAH  in air  divided  by molar  concentra-
          tion of PAH in water

     K  = gas phase exchange constant, cm-h
      g                                       -j
     K  = liquid phase exchangj constant, cm-h

The  log of  H  can be  estimated from molecular weight  of the PAH  compound
(Zepp and Schlotzhauer, 1979, equation 5.3.8).   Similarly, Kg was estimated
from-equation 5.3.9.
                   log  H =  0.7854 -  0.0191  • molecular weight        5.3.8

                   K
K  =  1137.5  •  (V+W)  •  (18/molecular  weight)'     5.3.9
 where
      V = current velocity (m-sec  )

      W = wind velocity (in-sec  )

 Values of K.  result  from application of two regressions for different wind
 velocity regimes (equations 5.3.10 and 5.3.11).

             K£ = 23.51 (V°-S69/R°-673) • (32/molecular weight)*      5.3.10

                                        if W < 1.9 ra-sec"1

     K£  = 23.51  (V°-969/R°-673) - (32/molecular  weight)* -  (eO-526(W-1.9))

                                        if W > 1.9 m-sec"1            5.3.11

 where

      R = water depth  (m)

      FOAM  uses  calculated  current  velocities  plus  a  user  supplied  wind
 regime in the volatilization model.


 MicrobiajL Biotransformation of PAH

      Metabolic degradation of  assimilated  PAH is modeled as  a  function of
 respiration  rate and was  therefore  indirectly  affected  by water tempera-
 ture.  Respiration was multiplied by DMAX.  for compartment i.  The units of
 DMAX^ are grams  PAH  metabolized per gram dry weight respired per time.  At
 present, all DMAX values are set at  zero.
                                     134

-------
Sorption of PAH

     Polycyclic aromatic hydrocarbons  may' be quickly lost from solution as
a result of sorption to suspended particulate matter and settled sediments.
Sorption  to  particulate matter  provides  a  possible  entry to  the aquatic
food web via filter feeding or sediment feeding animals.

     The sorption sub-model in FOAM was based heavily upon data provided in
Karickhoff et  a_l.  (1979)  where the sorption of  PAH was modeled as a first
order  relationship  between  dissolved  PAH in the water and partitioning co-
efficient, Kp.  The partition coefficient increased linearly as a function
of  the organic content of the sorbent material.  The partition coefficient
also  correlated with the octanol/water partition  coefficient  (KQW)  of the
PAH.

      Depending  upon available data,  FOAM calculates  rates  of sorption in
one of three ways.  First, if K   and the organic fraction of the sorbent
were  known,  a partition coefficient is calculated  from equations 5.3.12 -
 5.3.15.
 where
                      K   = K    •  oc                                5.3.12
                       p     oc
      Log  (k   )  = Log  (K   )  - 0.21                                 5.3.13
            oc          ow

      oc = fractional organic  content of  the sorbent.

 Second,  if K   is not known, k   is  estimated  from the solubility of the
 particular PAff:                °C

      Log  (Koc)  = -0.54 •  Log(s) - 0.44                             5.3.14

 where

      s = solubility of the PAH mole fraction

 Solubility can  be estimated  from molecular  weight (Figure 5.3.4).  Third,
 if neither k   nor the organic fraction  of the sorbeni, were known, a parti-
 tion coefficient was estimated directly  from:
           Log(Kp) =  4.16  - 0.51  • Log(s)                           5.3.15

 where

      s = solubility of the PAH,  ppm

 Equation 5.3.15 resulted from  further analysis of data presented by Karick-
 hoff et a_l.  (1979).

      Values  of K   were converted to units of inverse  time by dividing by
 the  duration of experiments (24h) of the  experimeuts which generated data


                                    135

-------
           I  3
            m
            o
            CO
            O   0
            _j
                -I
• Aromotics
  y=6.50-.040X

A Cycloolfefms
  y = 4.97-.030X

a Cyr.loparoffins
  y = 4.M-.03IX
                                     100
                              MOLECULAR  WEIGHT
                                   200
Figure  5.3.4.  Log  solubility  of  PAH  in  water  as  a  function  of  PAH
              molecular weight.
                                  136

-------
for  the  regressions (equations 5.3.14  -  5.3-16).   Because the experiments
were performed with PAH  concentrations that  approached the maximum solu-
bility of the  compounds and because the sorbent was suspended by continuous
stirring, we assumed  that  the K   values provided estimates  of a maximum
rate of  sorption,  S   .   At eachpiteration of the model, the specific  rate
of sorption  is calculated from equation 5.3.16.
       S =  S    • SBATE  •  SBENT  •  oc /  [SBATE +  (SBEKT  *oc)l       5.3.16
             max
where
                                     -2  -1
         S =  rate  of  sorption, gPAH-m   *h

       S    =  maximum  sorption  rate, £*h
       max               .
      SBATE =  soluble  PAH,  g-n"2
                          -2
      SBENT =  sorbent, g*m
 Biological Production Sub-model
                         * .
 Primary producers

      Rates of  biomass change  were  determined by first order mass balance
 differential  equations.   The  overall  biomass equation for  producer i  in
 reach  j  at time  t (puytoplankton,  periphyton, and macrophytes) is  repre-
 sented by equation 5.3.17.


     dPi
     _ _ (FI^ - FO.) + P.  (PSi - Ri - Mi - IL -  S.)  -  G.          5.3.17

      dt
where
                                _
      P. = biomass, g dry wt • m

     FIj = inflow of P  from re-.ch j-1

     F0i = outflow of P. from reach j

     PS. = photosynthetic rate of P

      R. = respiration rate of P

      M. = non-predatory mortality rate

      U. = excretion rate

      S. = sinking  rate

      C. = grazing  loss  of  P

                                    137

-------
Photosynthesis

     In FOAM, gross photosynthesis (of phytoplankton; periphyton and macro
phytes) was modeled after Smith (1936) (equation 5.3.18).
                         Pmax
where
     P    = photosynthesis per unit producer i in reach j at time t

     P    = light  saturated photosynthetic rate for producer i
      max
         I =  light  intensity  supplied by solar submodel

        I =  light  saturation constant  for producer  i

      f(T) =  temperature  function,  (equation  5.3.2)

 Values  of  parameters for  three primary producers in FOAM  are listed  in
 Table 5.3.2.
 Consumption

      Food  consumption  provided  energy for  biomass  production and  another
 potential  PAH source  to the non-photosynthetic  biological components  of
 FOAM.   Consumption  ic  modeled as  a  function of predator and prey  biomass
 after DeAngelis et ai.  (1975) (equation 5.3.19).

                      C   .  • B.  •  W. .  • B.
          C.  . = f(T) -    - -  -                         5.3.19
                        B. + J (Wij . B.)

 where

       C. . = input of prey j biomass to consumer i
        *»J
      C   ,  = maximum feeding rate of consumer i (day  )
                                     _2
         B.  = consumer biomass (g • m  )
                                   _o
         B.  = prey j biomass (g • m  )

        W   = Bayesian preference of predator i for prey  j  (unitless)

       f(T)  =  temperature dependent modifier of consumption  rate  (equation
              5.3.2).

 The behavior  and  stability properties  of this function have been  studied
 (De Angelis  et al. ,  1975).  This  equation has  been  usefully applied  to
 model  feeding by  zooplankton,  benthic invertebrates and fish  (Kitchell  et

                                     138

-------
ajl., 1974; O'Neill, 1976; Smith et al. , 1975).  Figure 5.3.5 represents the
prey  items  allowed for  each consumer  by FOAM.  When  prey are plentiful,
consumption,  C..,  is proportional  to predator biomass.   Conversely, prey
abundance regulates  consumption  in  equation 5.3.19  when  prey are scarce.
The Bayesian  parameter,  W   (Table 5.3.6) can be estimated from data quan-
tifying the  relative  abundance of prey j in predator i stomach samples and
the relative  abundance of prey j  in the environment (O'Neill, 1971).
     Foodweb  transfer  of PAH was modeled as C. . multiplied by the PAH con-
centration  in prey j  summed over all j for consumer  i.  FOAM can therefore
be  used  to  examine the  relative  importance of direct uptake of PAH by con-
sumers in association with respiration and trophic transfer of PAH through
the  foodweb.   Appreciable  amounts  of polychlorinated  biphenyls  have been
shown  to accumulate from consumptive processes,  especially at higher tro-
phic  levels  (Weininger,  1978).  Behavior of PAH's in  this regard was essen-
tialxy  unknown.   Calculation  of  the PAH  derivative for  P.  results from
multiplying  each terra in equation 5.3.17 by  the  fractional TAH concentra-
tion  of  P.  and adding  two terms that  quantify PAH uptake and degradation as
functions of  photosynthesis and respiration, respectively (equation 5.3.20)

            H.  = (Q.  •  PU.  •  PAH.)  / (K. + PAH.)                   5.3.20
              ill      d       i      a   ,
                             -2     -1
 In equation 5.3.21 H. (g  • m    • t  ) was the rate,  of PAH uptake mediated
by photosynthesis, PAH. was dissolved PAH (g  •  m   ) available for uptake
 and PU.  was  a  linear  function  of photosynthetic rate  (equation 5.3.21)

                          Plh  = IP •  PSi                             5.3.21

 where

      PU. =  g PAH assimilated per  g producer i per time

      PS  =  photosynthetic rate of producer  i, g • g"  • t

        m =  slope,  of PAH  taken  up  per g photosynthetically fixed biomass

 Respiratory  mediated  PAH   biotransformation  is  represented  by  equation
 5.3.22.

                Di = Ri  ' di ' Pi '  hi                              5.3.22

 where

      D.  = rate of PAH degradation by  producer i

      d.  = fraction of respiration used  to degrade PAH

      hj  = fractional  PAH content  of P

     R  = temperature dependent respiration rate of P
      1

                                    139

-------
                                                                        .#
                                                   .   „
^ q£ ^£> ^0 <£«, ^C> ^C> Q\ ^ c*>
Zooplonkton
Benthic
1 nvertebrotes
Macro-
Invertebrates
Bacteria
Fish (Pred)
Fish (Herb.)




























































Figure 5.3.5.  M.-itrix  of  cnn^tunei-.s and  food  items which arc  considered by
               the production portion of FOAM.
                                     140

-------

-------
        Table 5.3.6.   Continued
10
Food Item j
Predator Predatory
Fish
Zooplankton
Benthic
Invertebrates
Clams
Bacteria
Fish (FRED)
Fish (Herb)
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Herbivorous
Fish
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Suspended
Organics
0.1
0.9
0.9
0.0
0.0
0.0
0.25
0.90
0.0
0.0
0.0
0.0
Dissolved
Organics
0.0
0.0
0.0
0.0
0.0
0.0
0.25
0.90
0.0
0.0
0.0
0.0
Sediments
0.0
0.0
0.0
0.05
0.0
0.0
0.0
0.0
0.0
0.0
0.5
0.10
Suspended
Inorganics
0.1
0.01
0.01
0.01
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Settled
Detritus
0.0
0.0
0.0
0.4
0.2
0.2
0.5
0.75
0.0
0.0
0.15
0.40

-------
Consumer organisms

     The overall  mass balance equation  for  consumer  compartment  biomass is
given by equation 5.3.23
       ; =  (FIt  - FCh)  +  I  C    -  N.  (R.  + F£ + IK  + M.)  - G.
                                                                    5.3.23
     dt                  j

where
                                                                _2
         N.  = biomass of consumer i in reach at 5,  (g dry wt • m  )

        C..  = rate of consumption of prey j by consumer i

         R.  = temperature dependent respiration rate of consumer i

         F.  = fraction of ingested food that is egested

         U  = rate of biomass lost through excretion

         M. = natural mortality rate of i

         G, = predatory losses of i

      FI-FO = net convective transport of consumer i


  .".espi ration

      For  producers, a fraction of photosynthesis or for consumers, a frac-
  tion of standing crop  is lost through respiration.  Respiration was modeled
  as a  temperature-dependent function by specifying an R    value, which oc-
  curs at T  ,  for  each biological  compartment.

      For  consumers, a  respiration  dependent direct uptake  of PAH was in-
  cluded  in  FOAM.   A  limited  amount of evidence suggests that Daphnia concen-
  trate  PAH in its lipid tissue  (Southworth  et al. ,  1978).   These data fur-
  ther  suggested  an  active  uptake mechanism  which resulted in steady-state
  concentrations  after approximately  24  h.   In  FOAM,  a hyperbolic function
  which  relates respiratory  rate and lipid  concentration in  the biomass is
  used  to simulate direct PAH  uptake  by  zooplankton, benthic invertebrates,
  macroinvertebrates  and  carnivores.

      Biotransformation  of  PAH parent compound  is modeled  as a function of
  respiration  rate.   A variable fraction of the body  load is removed from all
  biological  compartments  as  a function of respiration rat»,  Biotransforma-
  tion products are shunted to a dissolved pool of biot Transformation products
  that acts as a final sink.
                                     143

-------
Excretion

     It  is  assumed that part of the biomass,of  the plant and  animal  compo-
nents  is lost  through  excretion.   Secretion  might be  a  more  appropriate
interpretation  of this process in  plant  metabolism.   Excretion is modeled
in a linear fashion.  It is  further assumed that PAH is  lost through  excre-
tion  in proportion  to  the  concentration  of  PAH in the bioinass component.

     Rates  of excretion (Tables 5.3.1  and 5.3.2) mul£ipl^ed by  the biomass
in  each component to calculate g,  dry  wt  excreted-m   *h   by  each biologi-
cal  component.   Multiplication of  these  rates  by the concentration  of,  PAH
in  each biological .component, results  in  estimates of PAH  excreted-m  -h
by  the  biota.   The model internally tracks  the PAH  concentration in  each
compartment during the  simulation.
 Mortality

      As for excretion,  it  is  assumed that a  constant fraction (Tables  5.3.1
 and 5.3.2) of biomass  is  lost from each compartment as  the  result  of non-
 predatory mortality.  Again,  multiplication  of mortality  losses  by the con-
 centration of PAH in  the biomass  produces PAH losses to mortality.


 Sinking losses

      Phytoplankton sinking  losses  are modeled separately from  settling  of
 suspended  particulate  matter.   Future versions  of FOAM  could  incorporate
 more  realistic  coupling  between hydraulic  aspects of  the  model and  net
 suspension  of both phytoplankton and suspended particulate  matter.   Peri-
 phyton  and  macrophyte  mortality  are added to the  sediment organic pool via
 the KSTL parameter.  The sinking rate used is 0.0021 h~ .


 Egestion

      A  fraction of the consumed  food is  egested  by the consumer components
 of the- model  (Tables  5.3.1 and 5.3.2). Egestion  losses enter the suspended
 particulate pools  for  bj^th  biomass  and PAH.   Egested PAH is  the product of
 egestion rate (0.0042 h  ) and PAH concentration  of the food  item.


 Direct uptake of dissolved PAH by producers  and consumers.

      Uptake of  PAH  by plants was  modeled  as a  second   order,  non-linear
 .Michaelis-Henten process (equation 5.3.24).
                  PUi = Pi  '  Qi  '  PAH/CKPi + PAH)                    5.3.24
                                     144

-------
where
                                                                           f
     PUC  or PUP. = uptake  of  PAH  by producer  (P)  or  consumer (C)  i
                                         _2
               P. = biomass (g,  dry wt'm  )  of producer i

       QP   or  QC. = maximum uptake rate of  PAH by producer (P) or consumei
                 1   (C)  i,  (g  PAH/g, dry r/t-b" )
                                                         _2
               PAH =  dissolved  PAH  concentration (g PAH-m  )
                                                                          2
       KP.  or  KG.  =  analagous  to a  half-saturation constant (g PAH-m )

      Direct uptake by the  producers was assumed,  in the model, to be^iude-
 pendent  of photosynthetic or respiration rates.  Q. was 20 g PAH • g  ,  dry
 weight producers-h  .   This constant  was derived from laboratory  accumula-
 tion  studies  and was specific  for anthracene in  this  case.   However,  the
 Monod model was used so  that structure/activity relationships could Le used
 in  future  versicus  of the model.  The Monod model  is  a relatively simple
 relationship to  describe the  saturation phenomenon observed in periphyton.

      Direct uptake by consumers was coupled to the rate of respiration with
 the assumption  that direct uptake takes place across respiratory  membranes
 and  is  an  active  process.  The  equation  is  similar to  (O'Neill  et al. ,
 1972),  except  respiration rate,   RSMAX.,  appears  in the  numerator as  a
 multiplier of Q. and Q.  has units of g PAH t;cumuldted'g   , dry wt-h  .   Q.
 was 42  • h    for all consumer components.


 External  loading

       Polycyclic aromatic  hydrocarbons  enter  the  artificial  streams  by
 external  loading of dissolved PAH from the head tanks.  This is reproduced
 in the  model structure, where dissolved PAH  enters  the uppermost reach as
 the product  of the  pre-defined  discharge regime (ra -day  ) and  the  PAH
 concentrations  of  the influent (g PAH-m  ).  Future versions of FOAM could
 accomodate  several point sources per reach.

       A  statistical  relationship between solubility and molecular weight of
 PAH compounds  was  derived from  published data  (Braunstein et  al.  1977)
 (equation 5.3.25 and Figure 5.3.4).

             log  (PAHd)  = 6.50 - 0-04  •  mol.  wt.,  r2 = 0.77         5.3.25

 This  regression was  used  to estimate  solubility  of  a  specific  PAH  and
 ensure  that  solubility  constraints were obeyed  throughout the simulation.
 PAH in excess of solubility limits was shunted to the sediments.
                                     145

-------
Initial conditions
                                                     —A
     Estimates  of initial conditions  (g dry wt  •  m  )  of the  biota  for
all 5  reaches  were derived from previous experimental  work at the channels
microcosm facility, located on the Savannah River Plant site (Giesy et aj..,
1979)  (Table 5.3.7).   Each biological  component of  the syrtem is discussed
in  greater  detail in  this section.   Simulations of other  aquatic systems
would  require parameterization for that system.   Alternatively, for simula-
tion to  compare fates  among compounds  a generalized system could be para-
meterized.
Table 5.3.7.  Initial biomass conditions for simulations reported here
              g dry weight • m
Compartment
Phytoplankton
Periphyton
Macrophytes
Zooplankton
Benthic Invertebrates
Clams
Bacteria
Predatory Fish
Herbivorous Fish
Suspended Organic Particulates
Dissolved Organic Matter
Sediments
Suspended Inorganic Matter
Settled Detritus
1
1
0.0
0.18
1.54
0.0
0.5
0.2
0.1
0.1
0.2
0.05
0.01
1.0
0.01
0.55
2
0.0
0.18
0.3
0.0
0.5
0.2
0.1
0.1
0.2
0.05
0.01
1.0
0.01
0.55
Reach
3
0.0
0.18
0.3
0.0
0.5
0.2
0.1
0.1
0.2
0.05
0.01
1.0
0.01
0.55
4
0.0
0.18
0.3
0.0
0.5
0.2
0.1
0.1
0.2
0.05
0.01
1.0
0.01
0.55
5
0.0
0.18
0.3
0.0
0.5
0.2
0.1
0.1
0.2
0.05
0.01
1.0
0.01
0.55
                                    146

-------
Primary procedures;  phytoplanktoa, periphyton. macrophytes

     Phytoplankton  contribute  to  the  production dynamics  of  large,  slow
moving  rivers;  however, previous experience with the SREL channels  micro-
cosms indicated that phytoplankton do not contribute  substantially  to over-
all  system p.-eduction.  While  parameters for  phytoplankton growth and  PAH
metabolism  have been  generated for this model component, initial biomass
was  defined  as  zero for the current simulations  (Table 5.3.7).

     The  algal  component  of the aufwuchs community was  measured to  be ap-
proximately  2 percent  of  the  live  volume  (Giesy  et aJL. , 1979), of which
members  of  the Chlorophyta  and Cyanophyta  comprised more  than 95%.   In
1977,  algal  volume collected from a reference  channel ranged between 6 and
12  cm -ra^.   Assuming  a  density of 1.0, this  volume translated to  0.12  -
0.24 g-m   .   Initial conditions for current simulations were  defined  as the
midpoint  of  this observed  range, 0.18 g-m

     Macrophyte flora  that  commonly  colonized  the artificial  channels  in
past research  included Juncus  diffusissimusi, Gratiola virginiana,  Calli-
 triche heterophylla and Bacopa caroliniana.   Colonization was assisted  by
 transplants  of  J.  diffusissitnus which subsequently dominated  the  biomass  of
 the macrophytic  vegetation in the  channels.   Estimates  of standing  crop
 from the  work  of Giesy et al. (1979)  suggested  spatial heterogeneity  in
 distribution of  J.  diffusissimus,  with the  greatest abundance upstream.
 Initial macrophyte biomass in  the  first reach was defined as 1.54  g  dry wt
 • m .  The regaining reaches were each set at 0.03 g • m


 Consumer  components:   zooplankton,  benthic   invertebrates,
 macroinvertebrates,  bacteria, fish

      As  in  the case  of phytoplankton,  pelagic zooplankton were  assumed  to
 be  insignificant  in the  overall  production  dynamics of the SREL  channels
 microcosms.  While  parameters  have  been  defined for the  growth and  PAH
 metabolism  by  zooplankton,  this  component  was  initialized as zero  for
 current  simulations.

      The benthic invertebrate component in  the model is derived  to  simulate
 the growth  and PAH dynamics of stream insects including the orders Ephe-
 meroptera,  Odonata,   Coleoptera  and   families   Notonectidae,   Corixidae,
 Centropogonidae and  Chironomidae.   An  estimate of  their  combined  standing
 crop  iu  previous  research in  the channels (Giesy et al. , 1979) was 0.5  g
 dry wt per  m .  This  value was used as an initial  condition for  all five
 reaches.

      The macroinvertebrate  compartment was defined  to  include  larger in-
 vertebrate organisms such  as crayfish and clams that have  been  incorporated
 in  previous ecological  studies involving  the  SREL channels  microcosms.
 This rompartment  represents  a  complex  fauna  in  natural  systems;  however,
 species  used  in  the   channels  included Corbicula  fluminea (Asian  clam)'
 Anodonta  irobecilus (papershell  clam),  and  Procamberus acutus acutus  (cray-
 fish).    In  experimental  tests   of  the  model,   individual  C. flmninea  were


                                     147

-------
added and.constrained  in each reach by wire mesh.   Clam  biomass  was 0.2 g
dry wt-m   in_each reach.  Bacterial biomass was arbitrarily defined as 0.1
g dry wt  • m  .   In experimental tests of  the  model,  10  juvenile bluegill
sunfish (Lepomis macrochirus) Mere  caged  in each reach.  Fish biomass was
estimated as 0.1 g dry wt • m  .

     The  biological  state  variable  for simulations  of the  dynamics  of
anthrcene in the channels microcosms facility are given in Tables  5.3.8 and
5.3.9.  The  physical parameters  for  simulating volatilization and  photo-
lysis of  anthracene  in the channels microcosms are  given in Tables  5.3.10
and 5.3.11.
                                    148

-------
Table 5.3.8.  Biological state variables for simulation of anthracene dynamics.  See Table 5.3.3  for
              explanation of abbreviations.
PRODUCERS       PSMAX   Q10P    TOPTP   TMAXP   RMAX    Q10R   TOPTR    TMAXR    EXCR    MORT    SINK
                (/HR)            (C)     (C)    (/HR)           (C)      (C)      (/HR)    (/HR)   (/HR)
Photoplankton  0.3000  1.4900  20.000  35.000  0.1000  2.3000  27.000  31.000   0.0300   0.0001   0.0021     5.4

Periphyton     0.0300  1.A900  20.000  35.000  0.1000  1.8000  30.000  37.000   0.0300   0.0001   0.0       5.4

Macrophytes    0.0200  1.4900  20.000  35.000  0.1000  1.8000  30.000  37.000   0.0300   0.0001   0.0       5.4


CONSUMERS       CMAX    Q10C   TOPIC   TMAXC    RMAX    Q10R   TOPTR   TMAXR    EXCR    MORT   EGEST
                (/HR)           (C)     (C)     (/HR)           (C)     (C)     (/HR)    (/HR)   (/HR)

Zooplankton    0.0375  2.0000  25.000  30.000  0.0080  1.8500  30.000  35.000  0.0042   0.0001   0.0042

Benth Invert   0.0150  2.0000  27.000  32.000  0.0042  1.8500  32.000  37.000  0.0042   0.0001   0.0042

Macro Invert   0.0125  2.0000  25.000  30.000  0.0042  1.8500  30.000  35.000  0.0042   0.0001   0.0042

Bacteria       0.0042  2.0000  27.000  32.000  0.0083  1.8500  32.000  37.000  0.0083   0.0001   0.0042
                                                        i
Fish (Pred)    0.0021  2.3000  27.000  31.000  0.0005  2.1000  30.000  34.000  0.0022  0.0001  0.0042

Fish (Herb)    0.0021  2.3000  27.000  31.000  0.0005  2.1000  30.000  34.000  0.0021  0.0001  0.0042

-------
Table 5.3.9.  Biological state variables for anthracene dynamics simulation.
                          Lipid
                         Content
Maximum
Uptake
(QMAX)
Uptake half
Saturation
Constant
Minimum
Depuration
  DMAX
Producers

Phytoplankton

Periphyton

Macrophytes


Consumers

Zooplankton

Benthic Invertebrates

Clams

Bacteria

Predatory Fish

Herbivorous Fish
                           15
   2x10
                                           -6
   1.0
                           13
 4.2x10

 2.4x10
                                    -5
 a.-l
  h

 bg PAH-m~2

 V1
                                    150

-------
          Table 5.3.10.   Parameters for the simula-

                         tion of volatilization of

                         anthracene.
                                           2
          Gas phase constant. K  = 3.6 x 10
                               g

          liquid phase constant, K- = 1.8 x 10


          overall rate constant, KL = 7.3 x 10
Table 5.3.11.  Parameters for photolytic degradationxof anthracene.

               Quantum yield coefficient = 7.5 x 10
Wave length
(nm)
300
310
320
330
3AO
350
360
370
380
390
400
Molar extinction coefficient
4.84 x 102
1.17 x 103
1.90 x 103
2.1 x 103
4.9 x 103
3.4 x 103
4.8 x 103
3.2 x 103
2.7 x 103
5.4 x 102
2.2 x 101
                                   151

-------
                                 SECTION 5.4

              SIMULATION OF ANTHRACENE IN CHANNELS MICROCOSMS

                           RESULTS AND DISCUSSION
Water

     FOAM was  used to  simulate a  constant infusion of  0.056 pmol-£   of
anthracene  into  the uppermost reach  of  the experimental channels.  Param-
eter estimates were based solely upon laboratory measurements or published
data; only the initial conditions were measured in the channels.

     During  daylight hours  the  model predicts  a  decrease  from  0.056 to
0.036 (jmoles  in  a distance of 81 meters, between reach one and reach five.
At night, this concentration gradient is absent and all reaches exhibited a
concentration  similar  to the  input  concentration.   Regardless  of light
conditions,  reach  one  exhibited  a  constant  anthracene  concentration of
0.056  [imol  anthracene-£~ .  The  model predictions  compared favorably with
the  average  day and  nighttime measurements  made  in the  channels vTable
5.4.1).   Simulated  values were somewhat lower, and  the  rate of loss down-
 stream  was  slower in the model.  Nevertheless, the general behavior of the
 dissolved "pool" was predicted  by the  model.  When the constant  infusion
 was  stopped,  the concentration of dissolved anthracene quickly decreased to
 0.006 nmol-£  within 24  h.
     Table 5.4.1.   Predicted  and  observed  concentrations of dissolved
                   anthracene in  channels  microcosms.
                                          pmol anthracene-"
          Time             Reach 1             Reach 3           Reach 5

     Dawn                                                      "
          observed          .066                  .065               .063
          predicted         .056                •  .056               .055
     Noon
          observed          .067                  .045               .028
          predicted         .056                  .045               Io36
                                    152

-------
Sediments

     Simulations of  anthracene  dynamics in the channels microcosms predic-
ted  that  the  bottom  sediments  would  accumulate  anthracene  linearly.
Steady-state concentrations  of  0.09, 0.004 and 0.0035 pmol-g   , dry weight
were predicted  for  reaches  1,  3 and 5,  respectively (Figure 5.4.1).  The
simulation predicted a rapid desorption  of  anthracene  from the sediments,
which was  not  consistent with the p "tern observed  in  the channels micro-
cosms  (Table 5.4.2).  FOAM  also underestimated the  rate  of  flux into the
sediments.   The  simulation model did,  however, predict, a downstream gradi-
ent of  anthracene  concentrations in  the  sediments  similar to the observed
gradient.

     The  routines  to  simulate  sorption  will  need  to be  improved to give
better  predictability in  future  simulation  models.   Sorption to sediments
is  complicated  by resuspension,  transport and corrective processes, which
are  difficult to represent  by  either deterministic  or  statistical models.
Long  term simulations will  be  affected by large rainfall events, which may
cause  great deviations from the  patterns of sediment transport throughout
most  of an annual cycle.  The  scaling  of such events will probably be such
that  good predictability will only be available on geologic time scales and
will  cause  large uncertainties  in  short-term  simulations.   Even beyond
this,  differences  in diffusion rates and^ micro-dynamics  have not yet been
adequately described by FOAM to -predict anthracene dynamics in bottom sedi-
ments.   Descriptious of thoroughly  mixed suspended sediments are much more
 adequately described by FOAM and other existing relationships.  We suggest
 that more research on sediment  sorptive and  desorptive processes as well  as
 interfacing to sediment transport models is  required before PAH concentra-
 tions ?n this important  PAH  accumulating  matrix can be made.


 Periphyton

      The  simulated  anthracene dynamics in periphyton showed a  rapid uptake
 to a quasi-equilibrium concentration of  about 0.17 pmol-g   dry weight,  by
 day  3   (Figure  5.4.2).   Anthracene  concentrations  in  the  periphyton de-
 creased downstream and are greater  at night  in all reaches.  These patterns
 probably  reflect both the greater  concentration of dissolved anthracene  at
 night  and the  initial assumption  that depuration rate is physiologically
 coupled  to  photosynthesis.   This assumption  is  probably violated because
 some depuration is  due  to passive  diffusion.from the surface  and interior
 of  the cells.   The day-night  pattern  results  in  the broad  envelope  of
 anthracene observed  in periphyton in the simulations represented in Figure
 3» H • £ •

      The  steady  state  concentration of anthracene in periphyton predicted
 by the simulation  was  similar to  the results  observed  in  the channels.
However,  the great  variability in  measured  anthracene concentrations made
 comparisons difficult (see Section 6.4).  The  2 SE confidence intervals for
the steady state anthracene concentrations  measured  in periphyton from the
channels microcosm ranged  from  0.07  to  O.lb  [imol anthracene-g   dry weight
which was in the  same range  as  the simulated  concentrations (Figure 5.4.2)'


                                     153

-------
     0.66
                                            Reach  5	
           .'//     ANTHRACENE
         0    67   134  201   269  336  403  470   537   609  672
                                   TIME  (h)
Figure 5.A.I.   Simulation of anthracene dynamics in bottom sediments of  the
              channel:; microcosms.
                                 154

-------
   LJ
   Z
     o
   o
   5
    a.
      o
        0.20
         0.16
         0.12
         0.08
         0.04
              0    67   134   201   268   336  403  470  537   604  672

                                SIMULATED  HOUR
Figure 5.4.2.  Simulated concentration of anthracene  in periphyton  through
              time  in  reaches  1  and  5.   Hatched  area contains  transient
              concentrations for  reaches 2-4.
                                  155

-------
Table 5.A.2.  Comparison of predicted and observed concentrations  of
              anthracene in stream sediments.
                       Anthracene Concentration
                                  , dry weight)
Time
(h)
24
67
134
201
269
336
403
470
537
672
Reach
Obs
.0024
.0044
.0070
.0096
.0149
.0137
.0108
.0113
.0092
.0072
1
Pre
.00026
.00059
.0014
.0027
.0073
.0079
.0084
.0092
.0096
.0092
Reach
Obs
.0023
.0040
.0060
.0086
.0134
.0118
.0104
.0116
.0076
.0065
3
Pre
.00020
.00045
.00096
.0018
.0040
.0041
.0043
.0044
.0044
.0042
Reach
Obs
.0021
.0034
.0053
.0073
.0123
.0106
.0104
.0107
.0065
.0053
5
Pre
.00017
.00037
.00080
.0015
.0033
.0034
.0035
.0035
.0035
.0034
      Parameter  values  for anthracene uptake were derived from experimental
 work with cultures of periphyton grown on  glass  slides in natural streams
 located on the  SRP.  The  uptake kinetics were determined for these actively
 growing cultures (see Section 4.4).  Periphyton in the artificial channels
 was comprised of less  luxuriant growth.  Extrapolation of laboratory deter-
 mined uptake kinetics, via  the model,  to  the  field  situation  seems  to be
 justified in  this case.   However, the  comoarison  between  half-lives of
 anthracene in simulated periphyton  and that  measured for periphyton shows
 the model results to be  in  close  agreement.   The depuration rate constant
 determined in the laboratory studies with  natural cultures appears to be i
 reasonable estimate of the  rate of depuration by  periphyton in the chan-
 nels.   During daylight hours,  photolysis  reduced the dissolved anthracene
 concentration, making  less compound  available for uptake.  In addition, the
 depuration process of  the periphyton should be operating to reduce the con-
 centration of anthracene  in periphyton.  In absolute  terms, the equilibrium
 concentration  docs not agree  within two orders  of magnitude with the an-
 thracene  concentrations  measured  in the  periphyton  in the channels.  The
 half-time  for elimination of anthracene in  periphyton following termination
                                    156

-------
of  the constant  input,  however,  agrees  within  2-3  h  with the measured
half-time  for elimination  of  anthracene  from  periphyton in the  channels.
Uptake might have been overestimated in the model, as parameters  for  uptake
rate were  derived  from actively growing  cultures  of periphyton that  colo-
nized  glass  slides placed  in streams located  on  the  Savannah River  Plant
site. -Vihereas, the periphyton in the  channels  microcosms  did not  appear  to
be  growing at the rate of  the cultured plants  and consisted  of a  different
flora.
Clams

     The  simulation  of anthracene  predicted that  the maximum  anthracene
concentration in clams  would not be  attained until  after the  anthracene
input  was terminated (Figure 5.4.3).   This  is because the  simulation model
assumes that the clams  are  accumulating anthracene both directly  from  the
water  and from  suspended  particulates, which still had anthracene associ-
ated with them  after the  anthracene  input was terminated.   The naximum  was
reached earlier  in  reach 1  than 3 or 5 because the amount of  anthracene
•associated with suspended, particulates  was  greatest downstream.   The maxi-
mum anthracene  concentration was attained in claas  in reach 3 because this
 reach had  the greatest exposure  concentration  between water and  particu-
 lates.

      The concentration  of anthracene  predicted to be in the soft  tissues of
 clams at  the time  of termination of  anthracene  input was  4.16 pmol anthra-
 cene- g   ,dry weight of soft tissue  (Figure 5.4.3).  This  is,equivalent to
 740 (Jg'g   ,  dry weight,  which  is greater than the 16.6 Mg'g,  , dry weight,
 actually observed in clams exposed to approximately 10 |Jg'£  anthracene in
 the channels.   This  value is probably an overestimate  of  actual accumula-
 tion  because  the  simulation allows  uptake  from particulates,   which have
 been  shown  to be  less  important for accumulation of PAH by   clams than
 accumulation directly  from  water (Dobroski and  Epifanio,  1980).   Also  the
 model  assumes  that filtering is a continuous process, which was not observ-
 ed in the channels microcosms.   The  clams living in the channels were also
 under  environmental  stress   (Giesy et  al.,  1979) because  of factors other
 than  anthracene, whioh nay have reduced the frequency and  rate of filtering
 as well as metabolism.

      FOAM considers  a component which we have called benthic invertebrates.
 This  term is  somewhat  misleading  and  needs  to  be defined here.   We have
 considered only macroinvertebrates such as  insects and oligochetes and  not
 protozoans in this  class  of organisms.  FOAM also  includes  another macro-
 invertebrate component which  for the  anthracene  simulation  is  the  clam
 component.   This component  can  also  be used to  include other macroinverte-
 brates such  as crayfish.   If better  precision of accumulation by clams is
 desired,  this  component  needs  to be  separated and  better  mechanistic pre-
 dictions made.

      The results of the simulation of anthracene dynamics  in benthic macro-
 invertebrates  cannot be  compared  to  analogous  results from tne  channels
microcosm because the benthic macroinvertebrate  community  had not developed


                                     157

-------
UJ
Z
UJ
o:
x
 <

 en
           Reach I

           Reach 3

           Reuc h 5
                               268  335  402   469  536  603  670
Figure 5.A.3.  Simulation  of  anthracene  dynamics  in  clams   in  channels

              microcosm.
                                 158

-------
to a  great  enough extent  to allow  sampling of enough biomass  for  represen-
tative, destructive sampling.
                                                                        *
     The  results  of the simulation  of  anthracene dynamics in the  macroin-
vertebrate  community  (Figure  5.4.4)  show that anthraceae concentration  is
greater in  the upstream reaches  because the anthracene  concentration in the
water and periphyton  is greater.  Also, it should be noted that  tho anthra-
cene  concentration in  the benthic  macroinverteurate component continues  to
increase  after tho anthracene input  was terminated.  This is  a  function  of
the ingestion of  anthracene-containing  food  (periphyton and  detritus) after
the anthracene  input  period.   Even  though we  cannot compare  this simulation
to  the macroinvertebrates in  the  channels microcosm study we  feel that the
prolonged increase predicted  by the simulation is unrealistic.   The main
reason is that the  relative  importance of  accumulation c   PAH,  anthracene
in  this  case, from   food and directly  from  the water  is  not  known.   We
suggest further research  be  conducted  to determine the relative importance
of  flow  materials along  these  pathways and determine better estimates  of
KP.,  KC., QP. and QC. (equation  5.3.24).


 fish

      As with the macroinvertebrate organisms, we were  unable to  compare re-
 sults from  the simulation to results  from the  channels  microcosms because
 of anthracene  induced mortality  (fee  section 6.6).   The simulation model
 includes components  for both nerbivorous and carnivorous fish.   The herbi-
 vorous  fish and  carnivorous fish attain approximately  the  same concentra-
 tions  of  anthracene  during  the  input  period  (Figures 5.4.5  and 5.4.6).
 However,  after the  input  of anthracene was stopped the  concentrations  of
 anthracene  in the herbivorous  fish stopped increasing  while that  of the
 carnivorous  fish continued  to increase.  This is a result of rapid depura-
 tion from the periphyton,  which was  the  major source  to  th?  herbivorous
  fish.   Alternatively,  loss  from  the   macrcinvertebrates,  which  were  the
  source  of ingestion for the carnivoious fish was much slower, which caused
  the  simulation  to predict a continued accumulation  of  anthracene.   Some
 preliminary laboratory studies  indicate that blueg?ll sunfish (see section
 4.8) accumulate  mo-^t of their body concentrations of PAH directly from the
 water.   Thus,  to improve the prediction capabilities  of  FOAM we  will need
 to  have better  estimates  of the maximum uptake  rates  and  relative impor-
 tance  of various  pathways under different conditions.


 Relative Importance Along Biotic and Abiotic Pathways
      We  havrt  calculated the relative importance of non-advective physical-
 chemical processes and  biological processes in determining the overall flux
 of anthracene through the simulated lotic system.  The periodic behavior of
 anthracene  corresponded to  the light-dark cycle and  emphasized  the impor-
 tance  of photolytic  degradation of  anthracene in the  overall  dynamics.
 During  the  c,.thracene  simulations,  the  continued  processes  of photolysis
 sorption and  volatilization accounted  for more than 50% of  the  total an-
 thracene flux along  pathway which allowed for either uptake and depuration


                                     159

-------
   57
          Benthic Invertebrates
           Reach  5 	
       ANTHRACENE
         INPUT
                 134  201   269
 336  403
TIME (h)
470  537   604  672
Figure 5.4.4.   Simulation of dynamics of antlicjcene in bcnthic macroinverte-
              brates other tli.'in clams.
                                 160

-------
-0.85
o
^0.68
 o> 0.51
 UJ
 §0.34
 (T
 I 0.1
ANTHRACENE
  INPUT
                                       Reach 5 	
134  201
          269  336   403
              TIME   (h)
                                               470   537   604  672
 Figure 5.4.5.  Simulation  of  dynamics o£ antlir.iccne  in  lirrhivn.-

-------
en
 o>  3

 UJ
 2
 UJ
 O
 <
 rr
 in
 H
 •z.
 <
         Carnivorous  Fish

         Reach I  	

         Reach 3 	

         Reach 5 	
      ANTHRACENE

       INPUT
     0
134  201
269   336   403  470  537

TIME  (h)
604  672
Figure 5.A.6.  SimulaLion of  anthracene  dynamics  in  carnivorous  fish
                                162

-------
or transformation.   As more  anthracene entered  biological  components the
relative  importance  of the  biological processes  increased.   This was es-
pecially  true  at night  when the  photosynthesis  rate was  zero.   When the
anthracene  input was  terminated,  the physical  chemical  processes became
unimportant  and  biological degradation   processes  assumed  control  over
the   fate of   anthracene.  Of all biological processes, consumer egestion
and  defecation of anthracene  laden prey items accounted  for  more than 40
percent  of  anthracene  flux.   Plant uptake  reprasented  about  6 percent of
the  total flux.  The  remaining  anthracene  moved fairly equally among the
Other biological pathways:  depuration  and  excretion by plants, mortality
losses  to plants  and  consumers,  flux along the  food web and  uptake and
depuration of  anthracene by  consumers.


Comparison of  Naphthalene  Ben2o(a)Pyrene and Anthracene  Simulations.

      One of  our initial premises  for  constructing a model from first prin-
ciples of  compounds  to  be  released  to  the  environment stated  that the
dynamics of  PAH compounds could  be  predicted  from a few readily available
pieces of information about a  compound.   These  include molecular weight,
 functional  groups  and  molecular stereochemistry.   There  was  given some
 agreement between trends  of  the simulations  and channels microcosms experi-
 ments for anthracene,  (mol.  wt.  = 178).  The experiment  was simulated again
 with equal  infusion rates of napthalene (mol. wt.  = 128)  and B(a)P (mol.
 wt.  = 252).   The initial  conditions and parameter estimates were identical
 to  the   anthracene  simulation.   Only  the molecular  weight  and absorbance
 spectra were changed to correspond to  naphthalene or  B(a)P.

      Without  data  for comparison with  the  FOAM simulation,  discussion of
 the  dynamics  of  individual state  variables  becomes  rather  meaningless.
 However,  it is  reasonable  to  examine the  relative  importance  of various
 pathways  simulated  in light of  available  information  concerning observed
 behavior of these compounds  in natural or  experimental  systems.  The simu-
 lated  fate  of naphthalene  was  similar to  anthracene  with  respect to the
 relative  importance  of  physical/chemical  versus   biological  processes.
 Physico-chemical processes  accounted  for more than  90% of the napthalene
 flux  through  the  channels  in  simulations  (Figure  5.4.7).   However, the
 absence  of periodic behavior corresponding  to  the light-dark cycle indicat-
 ed  that, unlike anthracene, photolysis was not  the  dominant  process but
 that volatilization was the  dominant physico-chemical  pathway of napthalene
 loss  from the channels,  according  to the simulations.   These results are
 similar to the conclusions of Lee et al.  (1978),  that volatilization was an
 important process in  the overall  loss of  naphthalene  from  large pelagic
 enclosures.   Similar to the  anthracene  simulation,  the  process of consumer
 egestion  and  defecation  assumed  importance in  the overall flux  of naph-
 thalene following the  trrminatioc of naphthalene  infusion in simulations of
 naphthalene  dynamics in the  channels.

      The transport  of benzo(a)pyrene   (B(a)P)  in the simulation was domi-
nated by biological processes  in contrast  to  the lighter weight anthracene
and  naphthalene  (Figure 5.4.7).   Following  an  initial 8 h period of  infu-
sion, where photolytic  degradation was important,  uptake by primary pro-


                                     163

-------
                        10
                        o.e
                        0.4
                        0.2
                                V1 ll  I
                                i H  i
                                   '
                                        i i
                                        i./1

                                                    (J)
AMI HH&CENt
                      5                    ',
                         c.e
                         0.4
                         o.j -

                         0.8 -
                         0.6 -
                         0.4
                            0  67  154 20* 263 Jlfi 401 -170 SJ/ 604 6 7J
                                     SIMULATED HOUB
Figure 514.7-   Rcl.itive   importance   of   non-advcctive   physical/chemical
                processos (solid  dots,  dashed line) arid  biological processes
                (open circles,  solid  liiie) in tot.il transport of (a) anLhra-
                ccnc,   (I))   naphthalene,   and  (c)-  benzo(a)pyrenc  through
                artificial streams.
                                     164

-------
ducers, movement  through the  foodweb, and  consumer processes of egestion
and  defecation played the determining  role  in the non-advective transport
and  fate  of B(a)P  in  the modeled stream.  Lu et  *I.  (1977) measured bio-
magnification of B(a)P by biota in a simple aquatic microcosm that contain-
ed  algae,  mosquito  larvae,  daphnia, snails, and  fish.   Lee et al.  (1978)
found  40  percent  of B(a)P added  to marine pelagic enclosures in samples of
sediments.   Of  the  physical/chemical  processes  simulated,   sorption  to
suspended  particulates  and settled sediments was  the most important  vector
simulated  for  B(a)P by FOAM.
Discussion           f

      To  be useful,  a model  of  the transport of  FAH should be capable  of
several  functions.   The model should  trace  the  entry of  PAH into ecosystem
components  potentially affected  by,  or  involved  in  the  transfer  of, PAH
throughout  the  system.  Second,  the model  should quantitatively  charac-
terize the  movement of PAH  among  system  components  and across system bound-
aries.  Third, the disposition of  PAH within components  should  result  from
model application.  Finally,  transport through the  system to ultimate sinks
 following removal of the  PAH input  should be simulated.

      Examination of initial  results of  FOAM simulations  suggests that  this
 model could become a  useful tool for  synthesizing  current understanding  of
 PAH dynamics and developing capacity  to  reasonably  prodict the transport  of
 PAH in lotic systems.

       Initial results of simulations showed that  FOAM predicted the  spatial-
 temporal dynamics  of dissolved anthracene with reasonable  accuracy (Table
 5.4.1).   This  success  might reflect  the observation that photolytic degra-
 dation  is  the major  pathway of degradation  for  this PAH.  The photolysis
 submodel has a firm foundation in theory and photolysis information for PAH
 in  general exists  in greater  amount  and  sophistication  than  for other
 processes.  Given  that the photolytic submodel  is  perhaps the most realis-
 tic  process  submodel within  FOAM  and photolysis  appears as  a key pathway
 for  degradation  of anthracene,  it  is  not surprising that FOAM predicts the
 diurnal  and downstream fluctuations measured for  soluble anthracene in the
 artificial stream channels.

       The absence of rate  coefficients  for biological transport processes  in
 relation to  molecular weight or  other  fundamental  chemical  properties  of
 PAH  reflected the limited amount of data available  for estimation of corre-
 lations.  Therefore,  the model  depended on extrapolation of rate constants
 measured in the  laboratory or published in the literature.   As the number
 of necessary  rate  parameters that  must  be directly  measured  in the system
 of  interest  increases,  specific  application  of  the model becomes  more
 difficult.   Lack of  close  agreement between simulated and measured concen-
 trations of anthracene in peripbyton,  clams,  and sediments likely reflected
 the  difficulties in extrapolating  laboratory determined  rate constants  to
 the streams.  For  example,  rate  of uptake by periphyton  was overestimated.
                                     165

-------
The  cultures of  algae  used in  the  laboratory  experiments  to determine
uptake  did  not  resemble  the  flora that  was growing  in  the experimental
channels.
                                   •
     While  no data  from the artificial  streams  are available to evaluate
the  simulated transport of naphthalene  and B(a)P,  the  agreement between
predicted  and measured anthra-ene transport  and  several general  aspects  of
the  simulated behavior of  naphthalene and B(a)P  generated some  confidence
in  the  model's  general  applicability over  a  wide  range  of possible PAH
compounds.   In the naphthalene simulation, volatilization proved to be the
single  most  important, process  that  determined  transport  of naphthalene.
This  agrees  favorably  with general  experience  with this  low molecular
weight  PAH.

     Photolysis  initially  accounted  for 40  percent of the  degradation  of
B(a)P  inputs  to  the  simulated  streams.   The modeled importance of photo-
lysis  agrees with the observations  of Zepp and Schlotzhauer  (1981) and are
consistent  with  expectations  based  upon  the   relatively  high molecular
seight  ot  this  compound.   However,  the low solubility  of B(a)P in  water
 (~lpg-ial  )  resulted  in most of  the external input being shunted to the
 sediments  where  it entered the  food web via the feeding activity of benthic
 invertebrates. Given the  observation  that  B(a)P  was input  in excess of its
 solubility  limits  during this  simulation,  strict comparison with the an-
 thracene and napthalene simulations  may  be unjustified.  In future  model
 application, equimolar  concentrations might provide results that are more
 valid for purposes of comparison.

      Initial  results  of  application of the model indicated  that molecular
 weight  carries  considerable information  concerning  transport of PAH com-
 pounds through, lotic systems.   Nevertheless,  much work is  still  required  to
 evaluate  the overall utility of FOAM.  Rigorous  sensitivity  analysis will
 identify  a  set  of parameters,  of which the values must be  measured with  as
 much accuracy as  possible, as  well as point out  those parameters that need
 only be estimated, say, within an order of magnitude.  Systematic variation
 of  individual  parameters  in  repeated  simulations  will  accomplish this
 aspect  of the sensitivity analysis.  From past  experience with functional
 forms  used  in FOAM (Bartell et al., 1981), likely candidates for sensitive
 parameters  in the  production  submodel  are the Bayesian food  prefeience
 values, W.., and assimilation values, a...

      The  dynamics  of PAH  transport  implicit in  molecular  weight  are me-
 chanically   translated  by  several  regressions  into  variable   parameters
 associated  with the  processes  of photolysis, sorption, volatilization, and
 solubilization.   As  a  function  of available data, each regression has its'
 own band  of confidence  limits  about the regression equation. For a parti-
 cular  molecular  weight,   regression  derived  parameters   can be selected
 randomly about the regression within the 95 percent confidence limits.  Use
 of these  stochastic  parameters  in repeated simulations constitutes another
 aspect of the necessary  sensitivity analysis of  FOAM.  This  analysis will
 evaluate the robustness  of the  regressions and  indicate which relationships
 should  receive  additional attention  or  which  ones  require larger data
 bases.
                                     166

-------
     Even without  the results  of rigorous  sensitivity  testing, there  are
several aspects  of the  model that  require  modification.   The  version  of
FOAM  used in  present simulations  fails  to  link  hydraulic  power to  net
suspension.  Resuspension and settling rates of  particulates are now  inde-
pendent of current velocity, clearly an unrealistic assumption.  Given  the
importance of  PAH  sorption  to particulates,  dynamics of particulate sus-
pension are  potentially  important in terms  of removing  dissolved PAH from
the water column and introduction of PAH into organisms  that  either filter
feed  or work the sediments.  The submodel for net suspension  should incor-
porate  hydraulic power,   reach  dimensions,   and  size distributions of sus-
pended  and  settled particulates  to more realistically  simulate particle
processing in  lotic environments.

      The  hydrologic  routine is  only  capable  of  simulating constant flow
conditions.   All reaches are identical  in morphometry,  and discharge into
the  uppermost  reach equals  discharge  between all  reaches; the simulated
stream  is hydrologically in  a  dynamic steady-state.   While  this  limitation
does  not constrain application of  the  model  to  the  artificial channels,
model application  to natural  lotic systems  depends upon  the  degree   of
similarity between the stream segment of  interest and  the constant  geometry
of the  model reaches.  Clearly,  to make  the  model useful  to  the majority of
 streams and  rivers of interest,  many of which  are impounded, the  hydrologic
 submodel  needs to  be modified  or replaced by an algorithm  that  can handle
 variable  reach dimensions and discharge  regimes.

      Simulation of B(a)P  transport suggested,  counter-intuitively,  that
 sorption was  not  an  important  process  in determining  the fate of this PAH.
 Previous research  (Karickhoff et al.,  1979)  indicated  that this high molec-
 ular weight,  relatively insoluble PAH  should readily sorb to particulats
 matter.  The regression  was  derived  from analysis of data  that described
 sorption  kinetics  for   a  variety  of  PAH  compounds  and  river sediments
 (Karickhoff et  al.,  1979).  The  derived regression has an opposite  slope
 from a regression based upon sorption  data  from another  source  (SRI Report
 in EPA file).  Clearly,  the analyses that lead to  the  present  submodel must
 be  re-examined.   It may be  worthwhile to test  the  implication of the  re-
 gression based  upon  the SRI data base  in the  sorption submodel and overall
 transport of  compounds  that appear  readily  sorbed to  particulates, such as
 B(a)P.

      Another  potential   mechanical  modification  of  FOAM  resides  in  the
 current  integration scheme.  The mass balance equations,  while conceptual-
 ized as  a   set  of coupled partial  differential equations  (e.g.,  equation
 5.3.1), are mathematically handled  as a  set of different equations with a
 simple Euler   integration  routine.    Results   of  simulations  repeated   at
 different time steps should  be  compared  in  order  to  determine a time step
 that optimizes  precision and  computing time.   It  might be  necessary  to
 include a more  sophisticated numerical  integration  algorithm (e.g.,  vari-
 able step size Runge-Kutta method)  to maximize  precision at the  expense of
 increased computer costs per simulation.

      A  major  conceptual weakness of the model lies  in  the production  of
metabolic or  degradation products  of  the   particular PAH  simulated.    The


                                     167

-------
current revision of  FOAM directs a constant fraction of metabolized pareat
compound   into  the  metabolite   pool.   We  assume  for   the  moment  that
the  metabolite  pool  does not undergo  further degradation.   No molecular
weight was assigned  to the metabolite.  In future versions of the model it
may  be  possible to  simulate some characteristics of  the degradation pro-
ducts based  upon  the  structure  of the parent  compound.   For example, the
number of  rings and  the kind,  location,  and  number  of functional groups
that  comprise  the parent compound might  be  used to  predict the chemical
structure  of degradation compounds that result from photolytic and metabo-
lic processes.

     The physiological process structure of the equations for PAH transport
through the food web encourages linkage of lethal and suolethal toxicity to
the  overall transport model.   In the  past,   fate  and  effects  have  been
largely examined  independently.   As  information accumulates concerning the
processes  affected by toxic components of PAH, the component of the growth
derivative  can  be  modified appropriately.   For  example,  low molecular
weight  fractions  of  certain  oils have been shown to  decrease  the  rate of
photosynthesis  of  marine  algal  species.   This effect could be translated
into a concentration dependent term that multiplies the simulated photosyn-
thetic  rate  by  the appropriate factor to model the toxic effect.  Similar-
ly,  rates  of other processes (feeding, respiration,  excretion, etc.) might
be adjusted to  simulate  toxic effects.  Incorporation of these toxic effect
multipliers would  directly link  effects to transport through the food web.

      Feedback between the modeling effort, analysis of published data, and
evaluation of  laboratory and field  microcosm  experiments will  result  in
modification  of FOAM  to increase the model's  usefulness as  a  management
 tool for predicting  PAH  dynamics  in lotic systems.
                                    168

-------
                                  SECTION 6

                         CHANNELS MICROCOSM STUDIES

                                 SECTION 6.1

                            FACILITY DESCRIPTION


     The  channels  microcosm facility used in this study  is  located on the
Department of Energy's (DOE) Savannah River Plant (SRP),  a 507 km  reserve,
including portions of Aiken, Barnwell and Allendale Counties in South Caro-
lina, U.S.A.

     The  channels microcosms  facility  (Figure 6.1.1)  is a  pass-through
system consisting  of six separate cinder-block channels,  each 91.5 m long,
0.61 m wide,  and 0.31 m deep.  Located at the upper end of each channel is
a pool 3.1 m long, 1.5 m wide, and 0.9 m deep.  At the lower end of the six
channel facility is  a single large pool 10.2 m l^ig, 3.1 m wide, and 1.0 x
deep.  For these studies, the channels and headpools were lined with a 0.05
cm thick  black polyvinyl chloride (PVC) film and  covered with 0.05 meters
of washed quartz sand substratum.

     Water  for the  channels  was pumped from the Tuscaloosa  aquifer  via a
deep well located near  the facility, treated to remove  C0?,  and hydrated
lime added  to produce  inorganic water quality similar to  that  of surface
waters in the local  upper coastal plain  (Table 6.1.1).   Water  flows were
monitored by v-notch weirs on each head pool where  water enters the chan-
nel.  Flow  rates of  75.7 £*min   were maintained manually by input valves
located   above   each   headpool;  this   resulted j.n  a   current  velocity
of 1.0 x  10   m'sec    and  a  retention time  of  2.5 h.  The water depth in
each channel was maintained at 20 cm by use of control gates located at the
lower end of each channel.

     Effluent  water  from  the  channels was pumped from the tail  pool at a
rate of 108 gpm, into a large steel treatment tank where it was gravity fed
through a filter of activated charcoal and gravel prior to final release to
the local environment.   At the point  of  release  anthracene  and anthracene
degradation, compound  concentrations  in the process water  were  at or below
background levels for water supplying the facility.

     At the time water flow was commenced four channels were seeded with a
50-£ slurry of water, sediment and biological  material  (i.e.,  algae, zoo-
plankton, macroinvertebrates)  from  seven locations  around the  SRP  site:
Brinkley  Well,  Skinface Pond,  Ellenton Bay, Dicks Pond,  Asphalt Pond and
Upper Three Runs Creek.   Seeding was  conducted weekly from March 11, 1979


                                    169

-------
Figure 6.1.1.  View of channels microcosms facility.
                                    170

-------
Table 6.1.1.  Mean water quality of treated well water
Total alkalinity
Hardness
pH
Specific conductance
Ionic strength (I)
DO
co2
so4'2
Total P
Nitrogen (N02 + N03>
Ca
Cu
Co
Cd
Cr
Fe
K
Mg
Mn
Na
17
21
6.
31
2.
8.
3.
1.
10
19
3.
3.
2.
0.
0.
1.
1.
.5 mg/£ as CaCO-
«J
.5 mg/£ as CaCO,
6
jj mho/ cm
5 x 10"4
5 mg/£
25 mg/£
9 mg/£
.1 pg/*
.2 vg/S.
17 mg/£
4 pg/£
5 pg/£
023 \ig/H
3 pg/£
7 |Jg/£
1 Mg/J^
246 |Jg/£
7.
1.
0 pg/£
8 mg/£
                                     171

-------
through June  13,  1979 and daily  from June 13, 1979 to July 16, 1979 after
which these channels were allowed  10  weeks of  additional colonization prior
to  the  introduction  of  anthracene.  Also, during  the  early  stages of the
seeding processes the macrophytic  angiosperm  Juncus canadensis was trans-
planted to  the system.  Juncaceae  represent the dominant inacrophytic group
which has naturally  colonized  the channels during past  studies (Kania et
al., 1976, Giesy  et al., 1979).                        £•'


METHODS AND MATERIALS

     Anthracene   (Aldrich Chemical Co.)  was   made  up  to  649±86  mg-£~  in
pharmaceutical  grade  ethanol.  Stock was  introduced via a peristaltic pump
at  2.1210.2  ml-min   for a  nominal anthraceneconcentration of 18.2 |Jg*^
and a nominal  ethanol  concentration of 23  mg-Ji  .  Fresh stock solution was
prepared  daily and was introduced  through  stainless  steel tubing (excepting
15  cm of  silicon  rubber tubing within tb°  pump) directly into the turbulent
mixing  zone  of  the  channel.   Nominal anthracene  concentrations  were not
obtained. The  average  observed input  of anthracene  was  16.2 and 13.8 |Jg*£
for the abiotic and biotic experiments, respectively.

     Anthracene   was  introduced  into a   newly  lined  uncolonized  channel
microcosm (abiotic study) for five  days (August 18-22,  1980) and samples of
water,  sediment  (sand) and  plastic liner  material  were taken from reaches
one,  two  and five.  Samples  cf water, sand and plastic liner were taken on
days  two  and  four at dawn  and 1530.  Samples of  water,  sand and plastic
liner were  taken at noon on day one  and three and  at 1030 on day five.  No
samples were  taken after the five  day input period.

     Anthracene  was  subsequently  introduced  into  a biologically colonized
channel   (biotic  study)  for  14 d  of exposure (September  22 -  October 6,
1980).   Samples  of  water,   organic sediment,  periphyton,  'fish,  clams and
plastic liner were takeu from reaches one, three and five on days 1, 2, 3,
5,  8  and  is  during the period of anthracene input.   Samples were also taken
1,  2,  4,  8, and 14  days after  anthracene input  was  terminated (October
6-20).  For  sampling and modeling  purposes, the  channels  were divided con-
ceptually into five  8.3 m long reaches.   Samples ware  taken at 4.58, 9.15,
and 13.7  m  from the beginning  of  each  sampled   reach.   For  the abiotic
study,  water samples  only were taken on days  two and four at dawn and 1530
to  evaluate photolysis losses.   Water, sand,  and plastic were also sampled
at  noon on  days one and three, and  at 1030 on  day five.

      In the  biotic study, water samples were taken  at noon on days 1, 2, C,
5  and at  dawn on days 2, 5,  8, 15, 16, 17, 19, 23  and  29 from reaches one,
three  and five.   Sediments  and periphyton were  sampled  from reaches one,
three  and five  on days 2,  3,  5,  8, 15,  16,  17,   19, 23  and  29 at 0830.
Clams,  fish and plastic were sampled from reaches  one and five at 0830 on
the same  days  as  the  sediment and  periphyton.
                                     172

-------
Water

     One-liter  water samples were acidified with  1 ml of concentrated HC1.
Control  and  blank water samples were  taken  at the weir prior to the addi-
tion of  compound.  The control was  spiked  with 3135 ng (abiotic study) or
5371 ng  (biotic  study)  anthracene and 4578 ng  (abiotic study)  or 5076 ng
(biotic  study)  anthraquinone in 30 pi  acetonitrile.  The water samples were
passed   through  ArcberliteTCAD-4 resin (Rhom  and  Haas, Philadelphia,  PA)
macroreticular   adsorbant   for  anthracene  and  anthraquinone.   The  resin
columns  were 1.5 cm diameter x 25 cm  long equipped with a 250 ml reservoir
and a stopcock.   A plug of  glass wool  was placed in the column and 30 ml of
wet precleaned XAD-4  resin was added to the  column.   The  resin  was pre-
cleaaed  with and  stored under nanograde methanol.  Prior to use the columns
were  drained  and rinsed with  10 bed volumes of deionized  water.   Water
samples  were percolated  through   the  resin  at a  flow  rate  of  0.8-1.2  £ •
hr  .   The  columns  were  eluted sequentially  with two  bed  volumes each of
anhydrous  diethylether and  nancgrade  acetone.   The  solvents were  combined
and dried  over  anhydrous  sodium sulfate.  The  solvent volume was reduced to
~ 200 (Jl by  a combination of rotary  flash evaporation and evaporacion under
a  stream  of  nitrogen.   Acetonitrile (2.0 ml),  containing the  chrysene
internal standard,  was  added to  each sample.  This  was  mixed and  25 pi
analyzed by  high pressure  liquid  chromatography (HPLC).  The recovery from
fortified  control samples was 62.319.8%  (x ± SE) for anthracene (n = 7) and
84113%   for  anthraquinone (n =  7) in the  abiotic study.   For  the  biotic
study  the  recoveries were  89.212% for anthracene  (n = 24) and 10513.4% for
anthraquinone (n = 24).
Sediment

      Sand  sediment to be sampled during  the  abiotic  study was placed in the
channel  in 3.3 cm  deep  crucibles.   One  crucible was  removed from each sta-
tion  at each sampling time.   The interstitial water was removed by vacuum
filtration.   Samples  were  extracted in a soxhlet  apparatus  for 18 h with
benzene:acetonitrLie (65:35,  v/v).   The  samples were  dried over  sodium sul-
fate  and  prepared  for  HPLC  analysis  as previously described.   Blanks and
control  samples were  removed  from  the stream  prior  to  the addition of com-
pound,  filtered  and  stored  frozen  under acetoniLrile.   Prior to analysis,
controls  were  spiked with 1045  ng anthracene and  1526 ng anthraquinone.

      Organic  sediments for  the biotic  study  were collected from  a silt bank
at  Box Landing on  Upper Three Runs Creek.   The sediment was sieved through
a  5 mm stainless  steel  screen,  mixed  and placed in  144,  15  x  60 mm glass
petri dishs.    Twelve cf the  sediment  filled petri  dishes  were placed at
each  sampling  station in reaches  1,  3 and  5.   One randomly chosen petri
dish  from  each sampling  station was removed at each sampling time.  Three
control  sediments from  an adjacent channel,  which  did not receive anthra-
cene, were  removed  per sampling  time.   Figure 4.2.1 (Section  4.2  this
report)  shows  a  flow diagram  of the extractiou and analytical procedure
utilized  tor sediment camples  during tne  biotic study.  Each  sample was
filtered  through Whatman number  4"!  filter paper  usir.g ^20 ml of distil] ed
water to  rinse the  petri  dish.  The sediment was  weighed  and  MO g sub-
                                     173

-------
sampled for  extraction  and analysis.  The remainder was dried at •v50°C for
1 to  2  weeks and dry weight determined.  Each subsawple for extraction was
mixed with MO  g of granular anhydrous Na SO,, transferred to a Whatman 22
x 80  mo cellulose  extraction thimble  an/ stored under 35  ml  of acetoni-
trile,  at  -20°C.   The  three control  sediments  were  treated  identically
except  that  one sediment  was spiked  prior  to addition of  sodium sulfate
with  1045  or 1790 ng of anthracene  and 1526 ng or 1685 ng ^nthraquinone in
acetonitile.

     One sample  set corresponding to one sampling date  consisted of  9 ex-
periraentcls, 2 background  controls and 1 spiked control.  The sediments and
acetonitrile  for each  sample set were placed in soxhlet extractors with 65
nl of benzene and extracted  for  18 h.  The crude extracts were rotary flash
evaporated to ^2 ml; passed  through  a 7 ca column of Florisil ,  eluted with
three benzene  rinses (^2  ml  each) of  the  round bottom flask and  10  ml  of
901  Benzene:ieethylene   Cl,(9:l,  V/V).   The  sample  volume  *as  reduced  by
rotary  flash evaporation "and evaporation under a stream of nitrogen to 200
pi.   Two or  four milliliters of acetonitrile containing a  chrysene, inter-
nal standard was added to each  sample.  The  samples were  mixed and placed
in a  freezer overnight.  Precipitates that formed overnight were removed by
ceatrifugation  and  the superantant was  transferred  and  stored  at  -20°C
until HPLC analysis.  Recovery of anthracene from spiked sediment was 73.3
* 7%, n =  5.  Anthrnquinone could not be quantified in extracts from sedi-
ments because  of an interfering compound, which could not  be resolved from
anthraquinone by HPLC (see Section A.2).


Poriphyton
                              2
     A  constant  area of 75  cm  of periphyton was scraped from the walls of
the  channel.  This  was removed as  an  aqueous suspension.   After all th<*
samples  were taken, the water was  removed by filtration and a wet weight
determined.

     Subsaoiples  from each  reach were removed for subsequent species identi-
fication.  Samples  for  identification were stored in a mixture of 60% deio-
nizcd  water,  301  ethanol  (95%)  and  10% formalin,  prior  to  microscopic
analysis.  A second subsaraple was  stored  under ethylacetate:acetone  (4:1,
V/V)  at  -20°C.   Three  controls and  one blank were taken from a non-exposed
channel.   The controls were  spiked with  1790 ng  anthracene and  1685  ng
anthraqulnone in 10  pi  acetonitrile.

     The  samples were  extracted by homofienixation  in a  TenBroeck tissue
grinder with 2 x 12  tal  cthylacetate:acetone (4:1, V/V) and 2 x 12 ml cyclo-
hexane.  The  extracts were combined, dried over Na^SO,  and  the volume re-
duced  to •*•  300  pi.  The  internal  standard  was «3ded  initially  in  2  ml
acetonitrile.  However, problems of  a two phase separation occurred; subse-
quently, 1 ml  ethanol  was used  to  add  the internal standard.  The samples
were  analyzed  by HPLC.   The  recovery of anthracene  from fortified samples
was 70.1121.   Samples  were corrected for background, which averaged 107 ng
anthracene per sample.
                                     174

-------
Fish

     The conditions  for the subsequent experiments on  fish in the channels
were slightly  different.   Juvenile bluegills, Lepomis  cmcrochirus, of mean
length 5.0  + 0.1 c»  ( + 2SE,  N = 58), mean wet weight  1.2 + 0.1 grams (N =
58), and with  a dry  to wet  weight ratio of 0.222  (N  = 45),  were obtained
froa the  National   Fish  Hatcheries at  Millen,  Georgia and  Orangeburg,
South Carolina.  Only fish appearing  in excellent condition were used.  The
fish were acclimated for 48 b  in a control channel and then transferred to
glass  and   stainless steel  mesh cages  in the  experimental channel  12  h
before  the   initiation  of  anthracene dosing.   During  both the acclimation
and experimental periods  fish were fed with  Tctra  Hin  staple food flakes
once a day.

     Fish for  anthracene assays  were  collected by dip net, wet weights were
determined  and whole fish were  placed  into vials containing ethylacetate:
acetone  (4:1  v/v)  aud  stored at  -20°C  until  analyzed.  Prior to analysis
fish were  homogenized in 2  x  12 ml ethylacetate acetone  (4:1, V/V) and 2 x
23  ml  cyclohexane   in  a  TenBroeck  glass  tissue  grinder.   Extracts  were
combined,  dried  over anhydrous  N'a.SO,   and  volume reduced  to  ~  300 \il,
Control fish were spiked with  895  ng  anthracene and 843 ng anthraquinor.e in
5 pi of acetonitrile. A chrysene internal standard was  added to each sample
in 2 ml of  acetonitrile.  Analysis was by hPLC and anthracene recovery froa
fortified  controls   was 94.5 +  2.5%.  Reported  fish  whole boay anthracene
concentrations after  24  -  144 h of depuration are  freshly  dead animals
collected when observed or  taken at the end of the experiment.  Analyris of
dead  fish  became  necessary  due  to unexpected  acute  phototoxicity  (see
Section 6.6).
Claras

     Individual  clams were  sampled frota  each  sampling station  and  a wet
weight  determined.  The  tissue and all  fluid  from the closed animal were
blended  and a homogeneous subsomple taken for  analysis.  The weight of the
subsample  and shell were determined.  The sample was />ddcd to 40-60 ml of
anhydrous  Na^SO,  for desiccation  and  stored  under  either cyclohexane or
cycohexane:ethylacetate  (1:1,  V/V)  at -20°C.  Three controls and  two blanks
were  prepared in  the same manner  from  an unexposed  population.   Tne con-
trols were spiked with 895  ng anthracene and 843  ng anthraquinone respec-
tively.

     The  sodiuaj  fculf-Ue coaled with desiccated clats  was extracted by shak-
ing with 2 x  50  Hi  cyclohfx«ne:ethylacct,ite (1:1,  V/V)  and  2 x 50 sal cyclo-
hexane.   The  volume was reduced  to ^ 200 pi  «nd internal standard added in
acetonitrile.  The recovery of anthracene  from fortified samples was 8717%
(n = 6).
Plastic Liner

     Plastic  strips  were placed at  mid-reach  in reaches one, two and  five
(abiotic  study)  and  one,  three  and  five  (biotic  study).   These were  sampled

                                     175

-------
                                                   2
using a  cork  borer which yielded plugs  of  / .45 cm .   Six plugs were taken
«l  each sampling  time.   The plastic  was stored  in  the freezer  prior  to
Analysis. The  plastic was extracted by shaking 5 min in (2 x 20 ml) ethyl-
acetate racetone  (4:1, V/V).   Recovery for  fortified  samples  was  9117% for
anthracene.  The samples were prepared for HPLC analysis as previously des-
cribed  for water extracts.
Ueatber and Sunlight

     Air tecperature', water temperature and general weather conditions were
recorded at each  sampling  time.  Solar  radiation  was determined  using  a
Belfort  Instrument Company Pyranograph  Cat.  No. 5-3850A  located approxi-
mately 10 miles southwest of the channels microcosms.
Quantification

     The  samples  were  analyzed  for  anthracene and  anthraquinone  using  a
Varian  Model 5000 HPLC  system with a  254 run  fixed wavelength detector.
Separations  were made with a  35  cm Micro-Pak HCH-10 reverse-phase column,
which  was equipped with  a  Whatman guard column of  Co-Pel  C.,,ODS  on  35 Mm
particles.   Compounds were  eiutcd from  the  column  at 28°C by programing
the moble phase  from  a  /5«% acetonitrile:  25% water mixture to 90% acetoni-
trile:  10% water  (Johnson  et  ^1.,  1977).   Acetonitrile  (90%)  was pumped
through  the  column for five Bin  before  recreating the initial conditions.
The  sediments  required a  gradient  to  100% acetonitrile  to  remove  late
cluting  compounds.


Quality  Control/Qua 1ity Assurance

     The  recovery and precision  for  each of  the matrices was determined by
applying  the proposed method  to  fortified  samples (Table 6.1.2).   Many of
these  studies were  performed  with   C-anthracene.   The storage stability
was  checked  by  storing fortified  samples  the  same  length of time as the
samples.

     The  background determinations were  made  from samples analyzed at the
same  time as the  environmental  samples.  The limit of detection was taken
to  be 3  o * background (Table 6.1.3).   The  linearity of the analysis was
determined with  each  sample  set by creating a standard  curve.

     The  accuracy of  the analysis was  determined by analysing a  standard
PAH  mixture  Alltech  (PAH-I)   which  contained anthracene.   The   relative
amount of anthracene  determined over  several  months  was 96.5  +  8%  (n =  10).

     Double  peaks at or near  the retention time  for anthraquinone  for  some
sediment  samples were confirmed  by GC-mass spectrometry (see Section 4.2^.
                                     176

-------
Table 6.1.2.  Recovery and precision of anthracene and anthraquinone from
              various matrices.
   MATRIX                                 aX RECOVERIES
                                Anthracene           Anthraquinone
Sediments
A) Sand
B) Organic
Fish
Periphyton
Clams
Water
Plastic Liner

15.5 ± 2
71.2 ± 12
87.7 ± 6
78.5 ± 2
78.9 ± 2
81 ± 2
91.1 ± 7

b
b
92.3 ± 8
82.2 ± 3
93.2 ± 2
97.6 ± 4
b
ax ±  1SU

  not  determined  because  of  interferences
                                   177

-------
Table 6.1.3.  Background end  Units of detection  for analysis of Anthracene
              and Anthraquinone  from various matrices
      Matrix
          Background              Limit of Detection
Anthracene     Anthraquinone  Anthracene  Anthraquinone
Water
Input Phase
Depuration
Phase
Periphyton
Fish
Claa.s
Sediments
Plastic

56 + 44 ng-jT1
3.8 ± 8.5 ng-£-1
_2
0.68 ± .64 ng'cin
47.6 ng/fish
103 + 70 ng-R dry wt
53.3 ± 2(> ng-g dry wt"
KD

322 + 336
ND
ND
KD
m
1 ND
KD

188 ng-f"1 1330 ng'£~]
29.3 ng'i'1
_2
2.6 ng-cia
116.9 ng/fish
315.75 ng-g dry wt"1
131.6 ng-g dry wt

 ND Not  detected.   Instrument Detection limit 0.25 ,ig = S x 2N
                                      178

-------
Statistical Methods

     Rate  constants  for accumulation  and  elimination  from each geological
and biological matrix as well as steady state concentrations were estimated
by  least-square  fits of  the data  to first-order,  donor-dependent models.
Each data  set was fit  independently  by  the  methods of Ciesy et. al. (1980)
with the Marquardt iterative least  square procedures (NLIN of the^Jtatisti-
cal Analysis  System, Barr et aK , 19/9).

     Dynamics  of anthracene  in  each  component  were  calculated by fitting
data to equations with  the  general  form given in equation 6.1.1

               dC                                                  6.1.1
               —- = K  -C   - K.-C
                 dt    " W    d  a


where

     K  =  uptake rate constant
      u

     K, =  depuration rate constant

     C  =  concentration of  anthracene  in the water

     C  =  concentration of  anthracene  in the component

      t =  time

To  avoid biases  introduced  by making  inappropriate  assumptions, most param-
eters  were estimated by several  different equations,  which wade different
assumptions.
      Accumulation of  anthracene by  the  PVC  liner was  linear  during th««
 period  which anthracene  was input  into the channels (Fig. 6.2.3).  After
 anthracene inputs were terminated,  .nthracene  desorbed slowly from  the PVC.

      Because organic  compounds  such an  anthracene tend to partition into
 organic matrices  an  estimate of  the  possible loss  to  the  PVC liner  was made
 by exposing PVC  strips,  which were  not colonized by periphyton.   Sorptive
 losses  to PVC estimated  in  this manner are a conservative estimate.  That
 is,  because the  PVC  extracted  was  not covered  by periphyton,  this number
 represents a maximum loss to the liner  that would  be  expected.

      The uptake  flux  into any component can be  described by equation 6.1.2

                          J  =  C    •  K                                6.1.2
                               w     u
                                     179

-------
 By  making the initial  rates  assumptions, that the amount 01  anthracene  in
 the component of  interest is  small  during the initial uptake period, one
 can estimate  the  uptake  rate  constant by  taking  a tangent to the  uptake
 curve  (J) and dividing by the water concentration.

     Because the water  concentration  C  is relatively constant  for a given
 period of exposure  the  equation which describes the dynamics  of anthracene
 in  a given component can be simplified to

               ^n              -(K 'tl                             6.1.3
          C  = ~ • (C ) (1 -  e (Kd t})
               K
               Kd

the  concentration of anthracene in the component of interest  approaches
steady  state as t •» a.  Thus,  at t -» a

                     C    K
               BCF = — = —                                       6.1.4
                     Cw   Kd

Equation 6.1.4 gives the relationship between the bioconcentration  factor
(BCF) or steady state distribution between the water and component  of  inter-
est.  Also, from equation 6.1.4 it can be seen that the steady  state concen-
tration can be predicted from equation 6.1.5, when the water  concentration is
constant                                   '

               C  = C    (1 - e"(Kd*°)                    * '   '    6.1.5
                a    ss  v

where

     C   = steady state  concentration hi component.

      We  have given  equations  for  estimating  KU,  KD  and  CSS during  the
 period of exposure.  To test for biases in  fitting equations to estimate  KD
 during  only  the  uptake  period,  the  concentration  of anthracene in each
 component was followed  during the depuration period, whtn no anthracene  wss
 being  added  to  the channels.   Estimations of  multiple   depuration  rate
 constants were made by  fitting the observed data to equation 6.4.1

                           n      -(K.,'t)
                     C   = Z Ca -e   di                            6.1.6
 where
      C   = concentration in component at time t

      C  , = initial concentration in ith component

         i = individual depuration component
                                     18C

-------
K., = depuration rate constant for ith conponea'.




  n - number of independent depuration components
                                 181

-------
                           RESULTS AND DISCUSSION

                         CHANNELS MICROCOSM STUDIES

                                 SECTION 6.2

                                    WATER
     In the  abiotic  study the control samples for the water had recoveries
of 62.3  ± 9.8%  (x ±  SE)  for anthracene and 84 ±  13.1% for anthraquinone
(Table 6.2.1).   These  values were considerably lower than those determined


Table 6.2.1. Percent recovery of anthracene and anthraquinone from control
             water samples in abiotic channels experiment.''
Saaple Set
1
2
3
A
5,
6
7
Anthracene
ng-£
1745
2960
1917
2172
674
2265
3096
% Recovery
56
94
61
69
22
72
99
Anthraquinone
ng'£~l
5442
5178
3192
3493
1451
4298
5705
« r, b
i Recovery
119
113
70
76
32
94
125
 One liter control water samples were taken upstream of anthracene input
 (weir) at each sampling period and spiked as follows:  3135 ng anthracene
 and 4578 ng antbraquinone in 30 pi acetone.  Blank water (no additions)
 recorded anthracene (91.8 ± 78 ng-£  ) and anthraquinone (298 ± 213
      1
 Calculated .is:  ngv?   detected/ng*£   spiked.
                                     182

-------
for samples  of the same  water fortified in the  laboratory  --  81.1  ± 1.5%
for anthracene  and 97.6  ± 3.9 for  anthraquinone.   This  nay have occurred
due to  the  additional handling and  transport time (30 minutes) in the heat
from the channels microcosms to the  laboratory.

     Aftrr  correcting for  recovery, some loss  occurred  between  the input
aad the first sampling station which was not accounted  for  by the  sum of
the anthracene and anthraquinone  (Table  6.2.2).   The average accountability
was
         Table 6.2.2.  Calculated  input of anthracene to channel
                       and measured values in  reach one.
         Day     Calculated Anthracene    Measured Anthracene.
                      Input (jJg*£   )      Concentration
I
2

3
4
17.7
12.6

21.1
21.8
11. A ± 0.4a'b
11.4 ± 0.7C
K
7.3 ± 1.3°
18.4 ± 1.7C
          x ±  ISO,  all  values  corrected  for  average  control
          recovery

          Values  recorded  at noon

          Values  recorded  at dawn  (7:30  am)


 76  i 5.3% and  ranged  from  48-95%.   The average  actual  anthracene concentra-
 just below  theinput in both  the  biotic and abiotic  studies were approxi-
 mately  12  Mg'£~  •  Anthracene  concentrations  in the water showed signifi-
 cant  diurnal  variation.   Maximum  water  concentrations (steady state) were
 achieved and maintained during periods of  darkness.  However concentrations
 decreased  significantly with  distance  downstream during daylight periods.
 Photolysis  accounts for 100%  of the  loss  down the  channel, with the other
 matrices  falling  in the experimental error  of  our  ability  to measure water
 concentration.   In  the  absence of  photolysis  (dawn  samples) the anthracene
 concentration  in the water was at steady  state  within  24 h  in both the
 abiotic and biotic  studies (Table  6.2.3  and  6.2.4)  (Figs. 6.2.1 and 6.2.2).
 Samples taken  at  noon or  1530  h had lower  concentrations  of anthracene than
 dawn  samples  in  reaches   3 and 5   (Table  6.2,3 and 6.2.4,  Figs.  6.2.1 and
 6.2.2  ).   The  loss, rate  constant  for  all  noon and afternoon samples was
 0.016  ±  0.006 rain"  which is  a half  life  of 4T 3 ±  12 min.  Siuce the loss
 of  anthracene was  observed in both  the abiotic  and biotic study the most
 probable  cause  is  photolysis  rather than  photo-respiration  by microbial
 populations in  the  water and benthos.

                                     183

-------
                                       .-1
       Table 6.2.3.  Concentration (pg-J?  ) of anthracene and anthraquinone in water.
Day  Time       	Reach 1	
                Anthracene  Anthraquinone
                  III.O"'       lln.o"'
	Reach 2
Anthracene  Anthraqui
  ue-£         uc«£
                                                                         none
                                                                         	eac   	
                                                                         Anthracene  Anthraquinone
                                                                           pg-2~         pg-1
CD
1   12:00 paj   7.7 ± 0.3    2.4  ± 0.3

2    8:00 am   7.7 ± 0.4    0.09 ± 0.05

2    3:00 pm   5.3 ± 0.8    0.8  ± 0.4

3   12:00 po   4.9 ± 1.0    3.0  ± 0.4

4    8:00 am  15.8 ± 13.3   0.8  ± 0.5

4    3:00 pm  14.1 ± 12.2   3.6  ± 1.2

5   10:30 an.   5.2 ± 3.4    1.5  ± 0.6
                                             3.1 i 0.1     3.9 ±2         1.8 ± 0.1   2.A  ± 0.7

                                             8.2 ± 0.3    0.08 ± 0.03      7.9 ± 0.2   0.08 ± 0.05

                                             3.4 ± 0.3    2.1  ± 0.5       !.7 ± 0.3   1.7  ± 0.3

                                             3.4 ± 1.0    3.1  ± 0.4       1.3 ± 0.1   2.7  ± 0.3

                                            13.1 ± 0.3    0.6  ± 0.3      11.7 ± 1.3   0.3  ± 0.08

                                             4.7 ± 0.6    4.6  ± 1.4       0.9 ± 0.2   3.0  t 0.9

                                             9.9 ± 1.5    4.9  i 1.5       9.0 ± 0.2   4.8  i 1.2
       All values mean ± SD of three samr-ing stations in each teach.

-------
Table  6.2.4.   Percent accountability nM anthracene + nM anth;jqui-
               nonf/actual anthracene input (nM).
                          Reach 1           Reach 3          Reach 5
 Dawn                       102               106              103

 Noon                        96                77               48

 Dry 8 Heavy
 Cloud cover                110                95               80
     Noon
     Concentrations of  anthraquinone were more variable  than  those  of  an-
thracene.  Anthraquinone  concentrations  were near zero in raicples taken at
dawn (Table 6.2.5).  Concentrations of anthraquinone in water samples taken
during the day  ranged from 1-5 pg'£   •   This  suggests that the authraqui-
none was due to photolysis of anthracene.

Plastic Liner
     Accumulation  of  anthracene  by  -he PVC  liner  was  linear  durirg  the
period during  which anthracene  was input  into  the  channels  (Fig.  6.2.3).
Attei anthracene  inputs  were  terminated, anthracene  desoibed  slowly from
the PVC.
     Because organic  compounds such  as   anthracene  tend  to partition into
organic  matrices  an estimate  o*  the  possible loss to the  PVC linear  was
made by  exposing  PVC  strips,  which were not colonized by periphyton.  The
estimate of sorptive losses to PVC estimated in this manner are a conserva-
tive  estimate.   That  is,  because  the  PVC extracted  was not  covere'l by
periphyton, this number represents a maximum loss t  the liner that would be
expected.
                                     185

-------
       WATER  ANTHRACENE  CONCENTRATIONS
  9000



  8000



  7000



  6000
o»
c
Uj 5000
Z
UJ
o
< 4000
cc
I
\-
-  3000
   2000
   1000
                                     • Doyl   Noon

                                     • Ooy 2  Down
                          3

                       REACH
                               —r
                               4
                                                T"
                                                 5
Figure 6.2.1.
    Anlhr.icene ronrentr.it ionr. in rojch 1 tlurinr, ti'f input

    of tlic  abiol ic stii
-------
  o>
 LJ
 O
 <
 cc
 x
          BIOTIC   STUDY
       I5
        10
1*1*
      \\\v
              DAWN NOON
              REACH 1
                                = I5
dn
                                        * ± a S.E.

                DAWN NOON
                REACH 3

             Downstream
                                        T
                DAWN NOON
                REACH5
Figure 6.2.2.  lllstor.r.iin rcprcsmLittK .intlir.icc-nc  concent r.i lions
            from rr.ichcs 1,  3 ami  5, rluritij; ll>c l>iotic nlmly.
                                            .'a I c r
                              187

-------
    Table 6.2.5.   Input and mass balance of anthracene/anthraquioone  in  reach one.
no
CD
Day
1
2
2
3
4
4
5
Sample Calculated Input Measured Concentration
Time Anthracene (nM) Anthracene (nM)
12:00 pea
7:00 am
3:30 pta
12:00 pm
7:00 pea
3:30 pm
10:30 amb
99
99
71
118
113
122
0
63.9
63-6
44.7
40.4
108-7
58.4
43.5
Measured Concentration Percent
Anthraquinone (nM) Accountability
13.1
O.A
4.3
15.8
3.3
19.2
8.1
77.8
64.7
68.7
47.7
94.7
63.6
«-t
     Accountability  -  uM  anthracene +  nM  anthraquinone/calculated anthracene input (nM).
      Input  stopped  1  h  prior to  sampling.

-------
                     PVC   LINER - REACH I
                1200
                 1000
              o
              »x
              UJ
              z
              UJ
              o
  800
                 600
              «" 400
                 200
                             I
                             Augutt 1380
                         '  Abiotic
                         .Avguil 1980
                                *± ISO.
                                 N =6
                          •UPTAKE-
                      ,  Abiotic
                      * Jvpait I9BO

-------
                                 SECTION 6.3

                                  SEDIMENTS


     The rate  constants for uptake and depuration from  sediments  were  not
significantly  different  among  the  three  reaches  (Table  6.3.1  and  Fig.
6.3.1).  This  indicates  that  the rates of  accumulation  were  indeed  first-
order with  respect to water concentrations of anthracene because  the  mean
concentrations of  anthracene  in the water varied and fluxes into  the sedi-
ments  varied  in proportion  to  the water  concentrations.   Depuration  from
the sediments was  also  first order with respect to the concentration  in the
sediment  (Table  6.3.1  and  Fig. 6.3.1).  Because the  rate  of accumulation
could be  related to the rate of diffusion into the sediment,  we normalized
the  concentration  of  anthracene in the sediments on both  a  weight  (Ttble
6.3.1) and area basis  (Table 6.3.2).  The relative variabilities of the K ,
K. and  C   values  from  the three reaches were less when normalized to area
but not greatly so (Tables 6.3.1 and 6.3.2>.

     The  sorption and desorption of  PAH are some  of  the most  difficult
mechanisms  to  model  and predict.   Host  previous  models have  used  steady
state partitioning coefficients in completely mixed systems to predict the
proportion of PAH  associated with the sediment.  It is difficult to develop
a probabilistic  model to predict the PAH  sorbed  to  settled sediment or to
quiescent sediment.   Also,  the  rate of sorption and desorption of PAH from
quiescent sediment may  be determined by the rates of diffusion into and out
of the sediment as well  as settling and resuspension processes.

     The results of  the experiments presented in Section 4.2 indicate that
PAH diffuse into organics and clay lattice spaces as a function of time and
become  less  available.  One  of the most  difficult  hydraulic processes to
Model  is bed  movement of sediment.  This process may be potentially  impor-
tant  in determining  the dynamics  of PAH  associated  with sediments.   Fur-
thermore,  catastrophic  (i.e.,   storm)  events may  be  the most  important
process  in  determining  the  overall  movement  of  PAH  associated with sedi-
ment.   Bioturbatfon  has been found to be an important process in  determin-
ing  the transport of  PAH  into  and out of  sediments.  Even though many of
these  transport  and  .-csuspension mechanisms  are known and  quantifiable,
developing a mathematical  simulation  model will be difficult.  One  of the
main  reasons  is  the  temporal sc.iling of  catastrophic  effects,  relative to
the  time  steps  required to attain adequate  resolution  of other processes.
For  this  reason  we feel that simulation  of the  dynamics of PAH associated
with  sediments will  always  be one  of  the   least accurate  components of
simulation models.  This will be a severe  limitation  of simulation  models
becaus'   of the  great  affinity of  sediments  for  PAH  in  aquatic systems
(Giddings et al.,  1978). We suggest that for short-term  simulations the top
                                    190

-------
2 en be considered us a homogeneous completely mixed storage pool for which
the sorption and desorption fluxes are not diffusion limited.
                                      191

-------
TABLE 6.3.1.   First order  rate  constants for uptake and release of anthra-
               cene from  organic sediments during the channels microcosm
               experiment based  on dry weight of sedimeut.  Data was fit by
               the Marquardt  iterative least squares procedure.  X + SE,
               n = 3,  PF  <  0.001 for all regressions.


Ka
u
Kb
Kd
K C
Kd
c d
ss
a . -1
ml'g
Reach
1
1.3 + 0.093
~~
0.0048 + 0.0007
C.0022 + 0.0003
3265 + 266

•h"
Reach
3
1.4 + 0.13

0.0049 + 0.0009
0.0022 + 0.0004
2980 + 298


Reach
5
1.4 + 0.11

0.0040 + 0.0007
0.0025 + 0.0003
2917 + 306


   h   ,  assuming density of sediment is near 1,  estimated  from data collected
   during anthracene input.

 C  h"  ,  estimated from data collected during period  of  no  anthracene input.

   ng  aothracene-g   sediment, dry weight
                                     192

-------
                     SEDIMENTS: REAV.H i
                      41448
                     SEDIMENTS:REACH j
                                        TIME (hr»)
                     SEOIWEHTS: REACH 5
                                          JJ6 J444JJ
                                       TIME ( I-..*!
                                                              672
Figure  6.3.1.
                     ion  of jtithr.icrof  l>y or^.inic  ^nltmcnls  in
pc-tri   
-------
TABLE 6.3.2.
First order rate constaits for uptake and release of anthra-
cene from organic sediments, based on area of sediment.
Data was fit_by the Marquardt iterative least squares
procedure.  X + SE, n = 3, PF < 0.0001 for all regressions.
Reach
1
Kua 0.22 + 0.014
K.b 0.0034 + 0.00056
d —
K.° 0.002 + 0.0003
d —
C d 800 + 86
S3 —
Reach Reach
3 5
0.27 + 0.016 0.26 + 0.019
0.0039 + 0.0005 0.0032 + 0.0006
0.0023 + 0.00033 0.0027 +. 0.00028
685 + 57 680 + 87
~
a , -2 . -1
ml* on -b
   h"  ,  estimated from data collected during period of  anthracene exposure

   h  ,  estimated during period of no anthracene input
 j                  _^
   ng  anthracene'cm   sediscnt
                                     194

-------
                                 SECTION 6.4

                                 PERIPHYTON


     The  anthracene  concentrations reported are not  corrected  for  extrac-
tion efficiency  vhich  was 71% + 0.1  (X + SD).   Thus, the numbers reported
are approximately 30%  less than the total amount of anthracene contained  in
the periphyton.  Since the reported values would be corrected by a constant
this does  not  affect the trends reported.  Since there is an error  associ-
ated with determining the  extraction efficiency at  each  point  we  decided
not to confound  this error with that of the extractions and quantifications
of anthracene in periphyton.  The dry to wet ratio for periphyton was  0.034
+ 0.016 (SD, n = 11).

     Concentrations  of anthracene  in the periphyton community were  normal-
ized to both an  crea  and a wet weight basis.  Similar crends were observed
for data  reported  in both ways (Figs. 6.4.1, 6.4.2, and 6.4.3).  The  coef-
ficient of variation was generally  less  when  concentrations  were reported
on  an  area  basis  but this  was  not always true.  There was  a  general  in-
crease during  the  first  48 h of exposure, however, there was some decrease
between 24 and  48  h in reach 1.  Concentrations in periphyton from resch 1
were much greater than  those  in reach 3 which  were  in turn approximately
four times greater  than  those  in  reach  5.  The  trend  of greater  conc'«r>.-
trations  at  the  upstream sites was consistent at all sampling times during
anthracene exposure.

     Because of  the  variability of anthracene concentrations in periphyton,
no  rate  constants were  calculated.   Uptake of  anthracene was  rapid,  and
reached steady  state within 168 h.   Furthermore,  desorption of anthracene
was very   rapid  (Figs.  6.4.1,  6.4.2,  and 6.4.3).   The  anthracene  concen-
tration was  not significantly  different from b-ckground  after  48  h.   The
half-time  for elimination was less than 1 day.

     The  concentration of anthraquinone in periphyton w.js variable and not
very different from  background.  An  interferring peak could not be resolved
by HPLC so anthraquinone concentrations are not  reported here.

     Removal  of  anthracene  from  the water column  by periphyton was  not a
significant  vrctor.   After  day  4 and  14,  0.09  and 0.04%  of the anthracene
added  to  the channels had been removed by periphyton.  While this is not a
significant  reduction  of the mass of  anthracene  added  to the water, con-
sideration of the periphyton  component could be  much  more  important as a
potential  for food  web  accumulation.   While  it  is  our  opinion  that the
vector of food  is  not as important  as  that of  direct exposure  from water,
                                     195

-------
for  aquatic organisms,  more  data  are  needed 'to  determine  the  relative
importance of this vector.
                                      196

-------
CM
E
^ PER1PHYTON- REACH 1
ui 30
2
UJ
< 20
ce
I .0
2

Ul
2 700
UJ
o
£ 500
tr
X
2 300
c" IOO
-


:A j>^— H
NM I •*• Ka'nr
^ 0 \»
II*| 1 | I 1
b ^
ANTHRACENE INPUT
•
Ax 1(± SO
/ \' N - 3
T f' ^
' " / X

. M / \
i T /* ^
/ \ / \
* 1 \\ * \
i Y \
. * i i
. [ V-
\ BO
42448 96 I&3 360J044CO
TIME (hrs)
Figure  6. 'i. 1.
AnLlir.iccnt!   concent r.i t ions   in  ;icr i,>liyto.i  in  rr.ich  1  •1
funrtion  of Lime.   Upju'r  Hptire  rrprcsrnts .nithr.icrne -c
while   the   lower   fiK'»'<-'   rop resents   .inthfJtcne 'K  .
weight, pcriphyton.

                       197
                                                                             «ifV

-------
      e
      
      c
           I PERIPHYTON - REACH 3
               ANTHRACENE  INPUT
                                                       BG
              1 1   f     >
       1000



        8CO
    Id-
's*
tr  .
Xo
H-S
     o»
     c
        200
                            /
                         /
             4 24 48  96
                                   X±SD

                                    N- 3
                                                   x
                                                 L
                                                     -
                                                  BG
                         168

                         TIME (hrs)
                                               360 304 408
6./».2.  Anthr.icone  conrcntr.it i on:;  in  prriphytoii  in  rc.irli  3 -is_2a

        function of  time:.   Upper  f inure represents  .intlir.icene-cm   ,

        while.   the   lower   figure   represent.-.   .intliracenfj;  ,  dry

        weight,  pcri|.hyton.
                              198

-------
                    PERIPHYTON - REACH 5
9
8
7
N
S *
•^
LJ
Z 5
LJ
O
< 4
CC
X
£ 3
fc **
<
0- o
c ^

1
0
•^
UJ
2 - 300
LU •*"
o.?
< a>
Q: 5200
X :K
V- «-
z ^ ioo
*t
en
o
c 0
"i1
ANTHRACENE INPUT |
i '
,








/
ki •» /
N « 2 .
/
9/n ' 2
!
' I
ln*\

1
x±so |
1 1 ^
N - 3 . (I*
, I
• T
i
.
.
\ ^
\ ^^.-^N.I
\ ^^^^
V-^
N»l

S

3
s
^
^.
^
^
a
^



*•-! ; — i 	 1 	 r- • i '
_ H i 1

N »2 S

/n*2
_. fj a 1
f
A-
^ _v ^»
^^ N'l 	 . 	 — * \-
^^ , 	 	 H • 1 \ C,
1 \
42448 96 169 360 384 40«
TIME (hrs)
Fipuro 6.A.3.  Antl-r.icrnc  concoiilr^L ions   in  pcfipliy ton  in  rcarl'  5  -TS.T'"1
                Unction  of Lime.   Uppor  fiRiiro  represents ...it lir.icrnc-cm   ,
                while   tlie  lower   figure   n-i'>-csruts   nnthr.iccne-R   ,  
-------
                                 SECTION 6.5

                                    CLAMS
     After  336 h  of  exposure  to  anthracene,  papershell clams,  Anodonta
imbecillis, had attained 96% of the projected steady state concentration of
17,580 ng antarocene-g  , dry weight of soft tissue which could be attained
at approximately  500  h  of exposure (Fig. 6.5.1).  The  first  order  uptake
and  depuration rate  constants  estimated simultaneously  during the J36  h
exposure  period  were JL3.99  + 1.7  (asymptotic  standard  error,  ml-g  'h  )
and 0.0080  +  0.0015 h   respectively.  This depuration  rate  constant  cor-
responds to a half time for elimination of approximately 88 h.   See Section
4.7 for a discussion of PAH accumulation and depuration by molluscs and the
lack of oxidative biotransformation observed in most molluscs.
                                     200

-------
                  20,000
            	 o
            23
            < *»
            _J •-   I 5,000
            — uj
                   IO.COO
            o
                            CC =(KU/KD) CW(l-e"KD"T
                                                 F2.39 =302
                                                 P< 0.0001
5,000
4,000
3,000
2.0CO
I.OOC
n
> V
>• \f ANTHRACENE INPUT
l
1

           0 2448   96      168
                       TIME (hrs)
                                                             336
Figure 6.5.1.
Uptake  of  .ml lir.i erne  by soft  tissues of  cl.ims  during tlic
poriocl  of  anllirjccnc   input,  to  llio  cli.innels.   E.icli point
represents  the mean of  6 clams.   Confidence  interv.ils ol 1,'GE
arc presented.  Tlir le.isl squ.irrs predicted  repression curve
for the indicated movie 1  is Riven.
                                   201

-------
                                SECTION 6.6

                                   FISH


     During  the  channels  microcosm study, we observed mortality of bluegill
Runfish.   The first fish mortality was observed  as  early as 1530 hrs, 7 h
after  the  anthracene infusion was started.  By  1700 hrs, 9 h of exposure,
all  of the  fish in  reach  1 of  the  experimental channels  were  dead.  In
reach  5,  the reach  farthest  frora the  input,  no fish died during the first
day  of the  experiment.   All  of  the  fish in this  reach were alive  at 0800
hrs  of day 2 of  the experiment.  Fish in reach 5  started dying at 1000 hrs
and  were all dead within  an hour.

     Bluegills  were collected just  after death  in  reach 1  and alive from
reach  5 after 4  h  exposure.  Fish were also collected from  reach 5 at 0800
hrs  on day  2 (24  h after  anthracene  infusion was  initiated),  which was
prior  to  any  observed mortality.  The  tojLal  anthracene  concentrations in
fish in reach 1  after  4  h was 2389, ng'g" , wet  weight  (SD = 275, n = 3).
Anthracene concentrations in reach five were 457  (SD = 182, n = 3) and 7763
(SD  =  1325,  n = 3) after  4 and 24  h exposure,  respectively.   Subsequent
•studies  showed  that  the mortality was  not due to the ethanol carrier.

     From  the  anthracene concentrations measured at the times  when fish
died,  we  suspected a  photo-induced mortality  of anthracene contaminated
fish.   To  try and  elucidate  this possibility we  considered  several experi-
ments.   A  brief  description  of  the  design  as  well as the results and dis-
cussion of each  experiment  are given in this section.

      Initially,  we  repeated the anthracene  exposure to  determine  if the
mortality  observed   during  the channels microcosms study  could  be repeated
(Figure  6.6.1).   Again  all of the fish  in  both reaches  were killed.  The
onset  of  mortality was  more  rapid  in reach 1  than  in reach 5,  as in the
first  biotic channels microcosm  experiment.  However, in this second study
of  fish mortality,   which was conducted  on  8-10 January,  the mortality was
less rapid  than  in the  first  experiment.   Also,  the  solar radiation was
less during  the  second experiment, due  to pertly  cloudy skies.

     A second study was  conducted  to  determine if the  mortality observed
was  due directly to  light  on the fish or indirectly frora a photo-product.
In  this study fish  in  reaches  1  and 5 were  shaded so there  was no sunlight
and  one reach (reach  3)  which was between  then  was  unshaded.   In reaches  1
and  5, which  were   shaded,  only  1 mortality was observed  (Figure 6.6.2).
However,   in  reach  3,  which  was  not shaded, all of the  fish died between
0800 and  1800  hra of the second  day  of the  experiment.   These results
                                     202

-------
a:
o
       Reach
16
12
 8
 4
               P'D'ARK
FDARK~~TI
UJ
_J
ID
O
16
12
 8
       Reach 5
                                    E5
                                 r
       246    12        24        36        48       60        72
                  HOURS  OF  ANTHRACENE  EXPOSURE
 Figure 6.6.1.  Ct:mi't ;il ivtj  aiort.i 1 i ly  of blurpill sin\finh  in  the  ch.iiiiicls
               microcor;m.   Moan  il.iwn  anllir.iccne  conconLr.it ions  were  12
               I'R-S.  .  Two  cajjcs  .it  e.ich  n-.icli with 8 fish |>cr capo.   Time
               is Riven ,TS cumulative  i-xpopurc  to .inlhraccnc.   i'crioitr.  of
               lit;ht  .infl  d.irk  nro  gjvrn.  * donolor. time at which all  fish
               were
                                   203

-------
           Reach
        10
        8
        6
        4
        2
     a:
     o
     ui
       IV. DARK
16
12
 e
 4
           24 6   12

           Reach  5
                     Shaded
L.:. DARK
                   24       36

                      Shaded
                                             48
                                              60
                              72
{eocn o unsnoc
3
! X^

Tea

-------
support the  hypothesis that the mortality observed was an anthracene-light
mortality.

     In another  experiment fish were dosed for 72 h in the dark (reaches 1
and 5)  and in the  light  reach 3 (Figure 6.6.2).   After 72 h  of  dosing at
0800  hrs  the  anthracene  input was terminated and the  fish which  had been
shaded  (reaches  1 and  5)  were  unshaded at 1000 hrs.

     All of  the  fish which had been dosed in the unshaded reach 3 were dead
at  the  end of the  72  h dosing period.   One fish  had  died  in  both reach 1
and 5 after  the 72 h dosing period.   Upon  being exposed to  light  after
accumulating  anthracene,  all of the  fish in reaches 1 and 5 died within 24
h  (Figure  6.6.3).  These  results suggest  that the mode  of  action  is  by
anthracene  or an anthracene  transformation product  in the  organisms  since
we  allowed  a washout period to  remove  anthracene  from the  water  coluran
prior to removing the  shading.

     An additional  experiment  was  conducted to determine the tiate necessary
for fish  held in clean water  to depuratt  photoactive  contaminants.   Caged
fish  were  dosed continuously  for  48  h with anthracene  (14.6 Mg-£~ ) at the
shaded  9  meter station.   The  anthracene infusion  was stopped and fish were
transferred  from the  shaded 4ia  station,  after  24, 48, 72,  96 and 144 h of
depuration  in  shaded, unconf aminated water  (eight  fish) to  an  unshaded,
uacontaminated section.

     After  48 h of exposure  whole fish  had attained an average anthracene
concentration of 318 Mg'g' ,  dry weight  (n = 8, SD = 79).  No fish had died
during  dosing in the dark.

     The  first  group  was  removed  from  the dark  at 08CO  (following  24  h
depuration  in the  dark).   The mean  anthracene  concentration  in whole fish
was  47.3  Mg'g"  ,  dry  weight  (n = 6, SD =  18.9).   All of  these  fish were
dead  by 1500 h (7 h exposure to  light) (Figure  6.6.4).

     Fish  which were  allowed 48  h_pf  depuretion in  the dark had  a mean
anthracene  concentration  36.0 Mg'g"  (« = 5,  SD  =  28.3)  dry weight.  The
mortality  of fish   in  this group was  less than  those which had been allowed
only  24  h of  depuration  in the  dark  before  exposure to  light (Figure
6.6.4).   After  48  h   four  of  the eight  fish  were  still  alive.   The fish
which were  allowed  72  and 96 h of depuration  in the  dark exhibited rates of
mortality,  which were similar to  that observed tor  fish which were allowed
to  depurate for 48 h.  After  72. h of depuration  the  anthracene concentra-
tion  had  decreased  to 73  Mg'g   •  dry weight  (SD = 37).   After  144  h the
concentrations were  not   different  from those in control fish.   Some of
these fish  were  able to survive  in sunlight  and reduce  the concentration of
anthracene  in their  bodies until  they  were undetectable.   We observed no
mortality  in  fish which had been allowed to  depurate until their whole body
concentrations had  decreased to  background  (Figure 6.6.4).

      The  rapid  depuration of anthracene from  fish,  which we observed, is
consistent  with  those  reported for many  different  fish species  and PAH  (Lee
et  a_l.  , 1972;  RoubaJ  e^  a^. , 1978).  Elimination mechanisms  reported for
                                     205

-------
                16

                12


                 8


             >   4
             
             O  |6
                12
                   Reach I
                   Shaded during anthracene  dosing.
                   -i—i—r—
                    246
                        12
~T
 24
36
              Reoch 3
              Unshaded  during  anthracene  dosing.
              All  fish  were  dead at the  end  of
              the  anthracene dosing  period.
                   Reach 5
                    Shaded  during anthracene dosing.
                    246     iT~~        24          36

                HOURS AFTER  DOSING WAS  TERMINATED
Fic
6.6.3.   Cmuul.-iLivc  mortality  of  bliicnill  r.untish  after  72 he  of
        rxposiiro to .nithr.icrnc in sh.nli'd  or itnsh.idrd readies.  After
        72  hi-  of antlu-ici-nc  dosing, .mtlir-icrnc  inputs wero stoppoil
        and fish oxpni.cd  to lij;lit.
                                  206

-------
              0800 hr.
             O



' 1 1 i ' i ii_
24 h


;iDA_FU<_^ KOARK :., ]
L ' ' 1 1 J J-l 1 1 1 1 •• 1 1 1 1 1 1 . t 1 1 . 1 1 1 1 1 1 1 ! 1 1 1 1 I I.
•^ 24 36 46
r Depurotion
JLATIVE MORT/
e
4
0
8
4
0
: 48 hr.
/~^~
/-"
T r— i j , 	 	 1 1
•72 hr.
^ 	 /

                   96  hr.
                O^r—r
8
4
n
: 144 hr. "
- • . . — A/ n F" i r A /7 / yr fi ...... ..



I  4   8  12        24        36
         HOURS AFTER FISH
      WERE  PLACED  IN UNSHADED
         REACH  OF CHANNEL
                                                         48
Figure  6.6.4.  Cumul.it ive mortal i ty of bluer.! 11 sunfish as a  function of
              time.  Fish wore rxpoard to 1A.6 \>%  anlliraccno • £   for ^8 h
              in the .lark.  Sets of 8 fish wero allowed  to depurate- into
              cK-.in  w.itcr  for  2'«, AS, 72,  9C  and 144 li  then pl.icrd in
              clean w.itcr in llic  li^.tit.   E.ich not  of fish were placed in
              the lir.lil- at 0800 lirs and obr.orvcd for  48 h.
                                   207

-------
fish  include  partitioning of PAH  into  clean water as well as  active- bio-
transfonnation  and  excretion   (Neeley  et  al.,  1974; Melancon and  Lech,
1"79} .                                  —  —

     The  mechanism  of  anthracene  phototoxicity  is  unknown,  but  can  be
postulated  to involve photooxidation of anthracene to form  toxic  products
wxthin fish or, more  likely, to involve sensitized photooxidation of macro-
molecules binding  anthracene.   The oxidative photo-chemistry of anthracene
has been  well studied.  Solution phase formation of endoperoxides  by sing-
let oxygen  as well as further oxidation to anthraquinone has been  reported
(Foote, 1968).   Anthracene derived fluorescent dyes,  used as environmental
probes  of protein  structure,  cause photooxidation of histidine and  other
amino acid  residues  (Harrington et ajl., 1956).

      Photooxidation  in the presence of sensitizing dyes  has  been  reported
to  cause  cell death,  in vitro  and  in vivo damage to nucleic acids, enzyme
inactivation  and  degradation  of  proteins  and  carbohydrates   (Means  and
Feeney, 1971; Spikes,  1977).  Light  is also known to increase fish activity
lev-Is and  trigger many  complex physiological processes which could contri-
bute  indirectly  to mortality under  stress  conditions of anthracene  contami-
nation.

      Phototoxicity  effects may be widespread among PAH.  Death of  £.  Coli
due to  photodynamic action of  benzo(a)pyrene in the presence of oxygen has
been  reported (Harrison and Raabe,  1967).  Mutagenic and potentially toxic
photoproducts have  been  farmed  from fuel  oil  exposed to sunlight (Larson et
al.,  1977)  and on PAH contaminated atmospheric particulate matter (Fox and
Oliver,  1981).   While many PAH are  known  to  be  photolabile,  to our know-
ledge phototoxic effects  in aquatic  biota have not been  reported previous-
ly.   Tests  of  phototoxicity should  be included  in  protocols  for  risk as-
sessment  of  compounds  known  to  be photclabile.   The  common  practice of
studying  PAH fate  and effects  under laboratory conditions  which minimize
photo-degradation  should be re-examined  until additional studies determine
the  potential  ecological  significance  of  PAH  phototoxicity  in  aquatic
biota.
                                     208

-------
                                 APPENDIX I


     Open  literature  publications  containing information collected  during
this study are given below.

1.   Bartell, S. M., P. F. Landrum, J. P. Giesy,  and G.  J. Leversee.   1981.
     Simulated transport  of  polycyclic  aromatic  hydrocarbons  in artificial
     streams.   in:   Energy and  Ecological  Modeling Mitsch,  W.  J.,  R.  W.
     Bosserman and J. H. Klopatek (eds.), Elsevier,  N.  Y.  839 p.

2.   Leversee,  G.* J., J.  P.  Giesy,  P.  F. Landrum,  S.  Gerould, J. W.  Bow-
     ling, T. E. Fannin, J. D. Haddock, and S. H. Bartell.   1981.  Kinetics
     and  Biotransformation  of  Benzo(a)Pyrene  in  Chironomous  riparius.
     Arch. Environ. Contain. Toxicol.  11:25-31.

3.   Leversee, G.  J.,  J. P. Giesy, P. F. Landrum, S. Gerould, M.  Bruno,  A.
     Spacie,  S.  Bartell,  J.  Bowling, J. Haddock,  and T.  Fannin.   1981b.
     Disposition of benzo(a)pyrene  in aquatic systems components:  periphy-
     ton, chironomids,  fish.  pp. 357-366.  In:  Chemical Analysis and Bio-
     logical  Fate:   Polynuclear Aromatic Hydrocarbons.   M.  Cooke and A.  J.
     Dennis  (eds.), Battelle Press, Columbus, Ohio.

4.   Landrum,  P.   F.  and  J.  P.  Giesy.   1981.  Anomolous  Breakthrough  of
     Benzo(a)pyrene During Concentration with Amberlite XAD-4 from  Aqueous
     Solutions.  Chapter 22 in:  Advances  in the Identification and Analy-
     sis  of Organic  Pollutants in Water,  L.  H.  Keith (ed.), pp. 345-355,
     Ann Arbor Science,  Ann Arbor,  Michigan.

5.   Gerould, S.,  P. F.  Landrua and J.  P. Giesy.   1982.  Anthracene  biocon-
     centration and  biotransforniation in chironomids:  Effects of tempera-
     ture and concentration.  Environ.  Poll,  (in press).

6.   Bruno,  M. G., T. E.  Fannin and G.  J. Leversee.  1982.  The disposition
     of benzo(a)pyrene  in the  periphyton communities of two  South Carolina
     streams.  Can. J.  Bot.  (in press).

7.   J. W.  Bowling,  G. J.  Leversee,  P. F.  Landrura  and J.  P.  Giesy.   1982.
     Photo-induced  Toxicity  in  Anthracene  Contaminated   Fish  Exposed  to
     Sunlight.  Aquatic  Toxicol.  (in  press).

8.   Leversee,  G.  J.,  P.  F. Landrum,  J.  P.  Giesy, and T.  Fannin.   1982.
     Effect  of Humics on  Polycyclic Aromatic  Hydrocarbons:  Accumulation by
     Daphnia magna and  Coprecipitation  at  Estuarine  Salinities.   Can. J.
     Fish.  Aquat.  Scl.   (submitted).
                                     209

-------
9.  Spacie, A., P.  p.  Landrum and G. J. Leversee.  1982.  Uptake, Depura-
    tion  and  Biotransformation of Anthracene and  Benzo(a)pyrene  in Blue-
    gill Sunfish.  Ecotox. Environ. Safety.  (In Press).
                                     210

-------
                                 REFERENCES


Abedi   2.  H.  and  A.  W. A. Brown.   1961.  Peritrophic membrane  as  vehicle
     tor  UDT and  DDE excretion in  Acdes  aegypti  larvae.  Ann.  Ent.  Soc.
     Amer. 54:539-542.                     —  	           	   	   	

Ahokas,  J.  T  ,  H.  Saarni, D.  W.  Nebert and  0.  Pelkonen.   1979.  The  in
     vitro  metabolism and covalent binding of benzo(a)pyrene to  DNA  cata-
     lysed  by trout  liver microsomes.  Chea. Biol.  Interactions 25:  103-


Ahokas, J. T.  1979.   Cytochrome P-450 in fish liver microsomes  and  carcin-
     ogen activation.  In:  Pesticide and Xenobiotic Metabolism in  Aquatic
     Organisms.  M. A.  Q.  Kahn, J. J.  Lech, and J.  J. Menn (Eds.), American
     Chemical  Society,  ACS  Symposium  Series 99,   Washington,  D.  C.,  pp.
     279-296.

American  Petroleum   Institute.   1978.   Fate and  Effects  of  Polynuclear
     Aromatic  Hydrocarbons  in the  Aquatic   Environment.   Publication  No.
     4297, 23  p.

Anderson, J.  W. , J.  M.  Neif,  B.  A.  Cox, H.  E. Tatum and C-.  M.  Hightower.
     1974.   The  effects of  oil  on estuarine animals:   Toxicity,  uptake,
     depuration,  respiration.   pp. 285-310.    In:   Pollution and Physiology
     oT  Marine Organisms.  F.  J.  Vernberg  and W.  B. Vernberg (Eds.), Aca-
     demic Press,  New York.

Andelman, J.  B.  and  J.  E. Snodgrass.   1974.   Incidence  and significance of
     polynuclear  aromatic  hydrobarbons in   the  water  environment.   CRC
     Critical  Reviews in Environmental Control 4:69-83.

Ausmus,  B.  S., G. K.  Eddlemon, S. J. Draggan, J.  H.  Giddings, D.  R. Jack-
     son, R. J.  Luxmoore,  E.  G. O'Neill, R.  V. O'Neill,  Monte Ross-Todd and
     P.  Van  Voris.    1980.   Microcosms  as   Potential  Screening Tools for
     Evaluating  Transport and  Effects  of Toxic Substances.  Edited  by W. F.
     Harris.  ORNL/EPA  4,  EPA-600/3-80-042  U.  S.  Environmental  Protection
     Agency, Athens,  Ga.   p  379-

Archer,  S. R., T.  R.  Bljckwood and G.  E. Wilkins.   1979. Status  Assessment
     o£ Toxic Chemicals:   Polynuclear  Aromatic  Hydrocarbons I.   EPA-600/
     2-79-210L,  U.   S.  Environmental  Protection  Agency,  Cincinnati, OH.

Barnsley,  E.  A.   1975.   The  bacterial degradation of fluoranthrene and
     benzo(a)pyrene.   Can. J.  Microbiol. 21:1004-1008.
                                     211

-------
Barr.  A.  S.  J.  H. Goodnight,,  J.  P. Sail  asd  J. T.  Helwig.   1979.   SAS
     User s  Guide.   SAS  Institute,  Inc.,  Raleigh,  N.  C.   pp.  317-329.

Bartell,  S.  M.,  P.  y.  Landrum, J.  P.  Giesy and G.  J.  Leversee.   1981.
     bimulated  transport of polycyclic aromatic  hydrocarbons in artificial
     streams.   pp.   1-31.   jn:   Proceedings  of the International Symposium
     on Energy  and  Ecolgoical  Modeling.  Louisville, Kentucky, April 20-28,
     J 981.

Baughman,  G. L.  and R.  R.  Lassiter.  1976.  Prediction  of environmental
     pollutant  concentration.    pp.  35-54.   In:    Estimating  the Hazard of
     Cheeical  Substances  to Aquatij: Life.  J.  Cairns,  Jr.,  K.  L.  Dickson
     and  A. W.  Maki,  (eds.).   American Society  for Testing Materials,  ASTM
     STP  657.

Bella,  D.  A.  and  W.  E.  Dobbins.   1968.   Difference  modeling of streau
     pollution.  J. Sanitary Engineering ASCE.   94:995-1016.

Bend,  J   R. , L.  M.  Ball,  T. H. Elmamlouk,  M. 0.  James  and  R.  M. Philpot.
     1979.   Microsomal mixed-function oxidation  in untreated  and polycyclic
     aromatic  hydrocarbon-treated  marine  fish.   pp.  297-318.  In:   Pesti-
     cide and Xenobiotic  Metabolism  ijn Aquatic  Organisms    M.  A. Q. Kahn,
     J.  J.  Lech and  J.  J. Menn  (Eds.).    American Chemical Society,  ACS
     Symposium  Series 99, Washington, D. C.

Bend,  J.  R., M. 0.  James and P. M.  Dansetts.  1977.  In  vitro metabolism of
     xenobiotics in some  marine animals.   Ann.  N. Y.  Acad.   Sci. 298: 505-
     521.

Bjorseth,  A. and A.  J.  Dennis.  1979.  Polynuclear Aromatic Hydrocarbons:
     Chemistry  arid Biological  Effects.   Battelle  Press,  Columbus, Ohio.
      1097 p.

Blanchard,   F.  *. ,   I.  T.  Takahashi,  H.  C.  Alexander ^nd  E. A. Bartlett.
      1977.   Uptake, clearance  and bioconcentration of    C-sec-buty-14-chlo-
     rodiphenyl  oxide  in  rainbow trout.   pp. 162-177.  In:  Aquatic Toxi-
     cology and  Hazard  Evaluation,  F.  L.  Mayer and J.  L.  Hamelink (eds.),
     American Society for Testing and Materials.  ASTM STP 634.

Bliuner,  M., G.  Souza  and J. Sass.   1970.   Hydrocarbon  pollution of edible
     shellfish  by an oil spill.  Mar. Biol. 5:195-202.

Boehm, P.  D. and J. G. Quinn.    1974.  The solubility behavior of No. 2  fuel
     oil  in sea water.  Marine Pollut. Bull. 5:101-105.

Boehm, P.  D.  and  J.  G.  Quinn.   1976.   The  effect  of  dissolved organic
     matter in  sea water on the uptake of mixed individual hydrocarbons and
     number  2   fuel  oil  by  a marine  filter-feeding  bivalve  (Hercenaria
     mercenaria).  Estu. Coastal Mari. Sci. 4:93-105.
                                     212

-------
Bowling, J.  W.,  J. p. Giesyt H   j
     scale  microcosms  for  assessing fates  and  effect-, of  trace  contami-
     nants.   pp   224-247.   In:   Microcosms  in  Ecological Research  J.  P.
     OakSRide   TN °   S'  Department  of Energy Technical Information Center,


Branson, D.  R. ,  G. E.  Blau,  H.  C.  Alexander and W. B. Neely.  1975.   Bio-
     concentration of Z^'^^'-tetrachlorobiphenyl  in  rainbow  trout  as
     measured  by  an  accelerated  test.   Trans. Am.  Fish.  Soc. 104-785-792.

Braunstein,  H.   M.,  E. D. Copenhaver and H.  A.  Pfuder^r  (ed.).   1977.  En-
     vironmental,  Health  and  Control  Aspects  of Coal Conversion:  An Infor-
     mation  Overview  Vols.  1  and  2.   ORNL/EIS-94.

Bravo,  H.   S. ,   S.  L. Salazar,  A.  V. Botello and  E.   F.  Mandelli.   1978.
     Polyaromatic  hydrocarbons in  oysters  from coastal  lagoons  along  the
     eastern coa'it of the Gulf of  Mexico, Mexico.   Bull. Environ.  Contamn.
     Toxicol. 20:171-176.                           -  -  -

Brown,  R.   A.  and P.  K. Starnes.   1978.   Hydrocarbons  in  the water  and
     sediment of Wilderness  Lake  IT.  Marine  Poll.  Bull. 9:162-165.

Bruno,  H.  G. and  R.  L.  Lowe.  i960.  The distribution of some beg Diatoms:
     a  cluster analysis.   Am. Midi.  Nat.  104:70-79.

Canton, J.  H. , R.  C.  C.  Wegman,  T.  J. A.  Vulto,  C.  II.  Verhnef and  G. J.  van
     Esch.   1978.   Toxieity-, £ccumulation-,  and  elimination  studies of
     cr-Hexachlorocyclohexane (aHCH)  with  saltwater organisms of  different
     trophic levels.   Water  Res .  2:687-690.

Carlson,  G.  P.   1972.   Detoxification  of foreign organic compounds by the
     quahaug,  Kerccnaria mercenaria .  Ccjip .  Biochem.  Physio 1 .  438:295-302.

Chen,  C.  W. and J.  T. Wells, Jr.   1976.   Boise river ecological  modeling.
     pp.  171-203.   In:   Modeling Biochemical Processes in Aquatic Ecosys-
     tems .   R. P.  Canale, (ed.),  Ann Arbor Science, Ann Arbor,  Mich.  389 p.

Chiou,  C.  T..   V. H.  Freed, D. W.  Schmedding and R. K. Kohnert.  1977.
     Partition   coefficient  and  biojccuroulation of  selected  organic  chemi-
     cals.  Envii-.  Sex. Techno 1.  11:475-478.
 Christensen,  H.  E. ,  T.  T. Lugin Byhl and B. S. Carroll.  1975.   R£J?i£tr£ of
      toxic  effects  oj^ chemical substances.  U. S. Dept. of Health Education
      anor~Welfar2.   Rockville, MD.

 Coffman,  W.  P.,  K.  W. Curonings and J. C. Wuycheck.  1971.  Energy flow in a
      voodland stream ecosystem:   Tissue  support trophic  structure  of the
      autumnal community.  Arch. Hydrobiol. 68:232-276.

 Crawford, D. J. and R.  W. Leggett.   1980.   Assessing the risk of exposure
      to radioactivity.   Amer. Sci. 68:524-636.
                                     213

-------
Collier, T.  K. ,  L.  c.  Thomas and  D.  C. Malins.   1978.   Influence of en-
     vironmental temperature  on  disposition of dietary naphthalene in echo
     salmon  (On^oj^rnchus  kisutch):   isolation and identification of indi-
     vidual metabolites.  Comp. Biochejn. Physiojl. 6lC:23-28.

     y>  K-G '  *"'  *"" Landrum and C.  C.  Fischer.  1979.  Investigations of
     xenobiotic  metabolism in intact  aquatic animals.  pp.  217-231.  In:
     KCh    N   ~ Xenobiotic Metabolism  in Aquatir  Organisms.   M.  A.  Q.
     Arcc     J'  Lech  and J- J-  Menn  (Eds.),  American Chemical Society,
     ACS Symposium Series 99, Washington, D. C.

DeAngelis,  D.  L.,  Goldstein, R.  A. and  R.  V. O'Neill.  1975.  A model for
     trophic interaction.  Ecology 56:881-892.

Deickman,  W. B.  and H.  W.  Gerarde.   1969.   Toxicology of Drugs and Chemi-
     cals-   Academic Press,  New  York.  805 p.~~

Derr,  S.  K. and J.  J.  Zabik.  1972.   Biolobically active compounds in the
     aquatic environment:   the  uptake and distribution of  (1, 1-dichloro
     <.,2-bis (p-chlorophenyl) ethylene),  DDE by Chironomus  tenlans Fabri-
     cius   (Diptera: Chironomidae).   Trnns. Amer. Fish.  Soc'.   101:323-328.

DeSalvo,  L.  H.,  H.  E.  Guard  and  L.  Hunter.   1975.   Tissue hydrocarbon
     burden of  mussels as  potential  monitor of environmental hydrocarbon
     insult.   Environ.  Sci.  Technoj^.  9-.2A7-251.

Dobroski,  C. J. and C. E. Epifanio.   1980.   Accumulation of benzo(a)pyrene
     in a  larval  bivalve  via trophic transfer.   Can.  J. Fish. Aquat. Sci.
     37:2318-2322.                                    ~

Douglas,  B.   1958.   The ecology of  attached diatoms  and  other algae in a
     small  stony stream.  J. Ecol.  46:295-322.

Dunn,   B.  P.    1976.   Techniques  for determination  of  benzo(a)pyrene  iu
     marine organises  and sediments.   Environ.  Sc_i.  Technol.   10:1018-1021.

Dunn,  B.  P.  and  H. F. Stitch.   1975.   The use  of mussels  in estimating
     benzo  (a)pyrene  contamination of  the marine environment.  Proc. Soc.
     Exp.  Biol.  Medicine 150:49-51.

Dunn,  B.  P.  jnd  H. F.  Stitch.   1976.  Release of  the  carcinogen  benzo(a)
     pyrene  fro"1  environmentally  contaminated  mussels.   Bull.   Environra.
     Contam.  Toxicol.  15:398-401.

Dut.iie,  H.  C.   1965.   A study of  the distribution  and p.eriodicity of some
     algae  in a  bog  pool.   J. Ecol. 53:343-349.

Duthie, J.  R.   1977.   The  importance  of  sequential  assessment in  test
     programs for  estimating  hazard  to  aquatic  life.   pp. 17-35.   In:
     Aquatic Toxicology and  H«rzaj-d Ev.ilnation, F. L. Mayer and J. L. Hame-
     link   (eds.).   American "society for  Testing and Materials,  ASTM  STP
     634.
                                     214

-------
Estenik, J.  p.  and W.  J. Collins.   1979.   In  vivo and in vitro studies of
     349-170  nCti°n  oxidase  in an aquatic  insect  Chironomus  riparius.  pp.
        -J/0.   In:   E£ltl£.>_des and Xenobiotic Metabolism  in  Aquatic  Orga-
     ££151.  M.  A. Q.  Khan,  J. J.  Lech and J.  J. Menn  (Ids.).   American
     Chemxcal Society, Washington, D. C.

 vans,  W.  C., H. N. Fercley and E. Griffiths.   1965.   Oxidation metabolism
                   "e 3nd anthracene by soil pseudomonads .   Biochem. J.  95:
Fisher,  S.  G.  and G. E.  Likens.   1972.   Stream ecosystem:   Organic energy
     budget.   Bi^scJ.. 22:33-35.
              9          Mechan*sras of  photosensitized oxidation.   Science


Fox,  M.  A.  and S.  Olive.   1981.  Photooxidation of  anthracene  on atmos-
     pheric  participate  matter.   Science 205:582-583.

Fossato,  V.  U.  and  W. J.  Canzonier.   1976.  Hydrocarbon uptake and loss by
     the mussel Mytilus  edulis.   Marine Biol. 36:243-250.

Fucik, K. W.  and  J.  M. Neff.   1977.  Effects of temperature and salinity of
     napthalenes  uptake  in  the  temperature clam  Rang i a  cuneata  and  the
     boreal  clam  Protothaca  staminea.  pp.  305-312.  In:  Fate and Effects
     p_f  Petroleum Hydrocarbons  i,n Marine Organisms  and Ecosystem's,  D. AT
     Wolfe  (ed.),  New York, Pergamon  Press.

Gardner,  W.  S.  and  D. W.  Menzel.  1974.   Phenolic aldehydes as indicators
     of  terrestrially derived  organic matter  in  the sea.   Geochirc. Cgsmo-
     chim. Acta .  38:813-822.

Garnas,  R.   L.   1975.   Comparative  metabolism  of pesticides by marine in-
     vertebrates.   Doctoral  dissertation, University of California, Davis.

Gearing,  J.  N. ,  P.  J.  Gearing,  T. F.  Lytle and  J.  S. Lytle.  1978.   Com-
     parison of  thin-layer  and  column  chroraatography  for  separation of
     sedimentary  hydrocarbons.  Ana 1 .  Chem.  50:1833-1835.

Gearing,  P.  J.  , J.  N.  Gearing, R. J.  Prueli,  T.  C. Wade and J. G. Quinn.
      1980.   Partitioning  of  No.  2  fuel  oil  in  controlled  estuarine  eco-
     systems,   Sediments  and  suspended particulate  matter.   Environ.  Sci .
     Technol.  14:  1129-1136.

Gehrs,  C. W. (coordinator).   1976.   Coal Conversion: description  of  tech-
     nologies  ajnd   necessary   bioroedical   arid   environmental   research.
     ORNL5192.  ~0ak Ridge  National  Laboratory,  Oak Ridge,  TN.   288  p.

Gerhart,  E.   H.  and  R.  M. Carlson.   1978.   Hepatic mixed-function oxidase
      activity  in  rainbow trout  exposed   to  several  polycyclic  aromatic
      compounds.  Environ.  Res. 17:284-295.
                                     215

-------
 erould, S.,  p.  Landrura  and  J.  P.  Giesy.   1981.  Anthracene  bioconcen-
     tration and biotransformation  in  chironomids:   Effects of temperature
     and concentration.   Environ.  Pollut.  (In Press).
Giddings,  j.  H^  B  T   Walton>  G  K   Eddlemon and  K.  G. Olson.   1979.
     iransport and  fate  of  anthracene  in aquatic microcosms,   pp. 312-320.
     FPA fiftlr^-^-^  Degradation  of  Pollutants  in  Marine  Environments,
     c,rA-t>U019-79-012.  U. S. Environmental Protection Agency,  Gulf Breeze,
         '   .'    '   1979.  Acute  toxicity  to  Selenastrum  capricornutum  of
     aromatic  compounds  from  coal  conversion.   Bull.  Environ.  Contain.
     Toxicol. 23:360-364.                          	   	   	

GleSy78noi P" i!(ed-)'   158°'   Microcosl"S  in  Ecological  Research.   CONF-
     R d>   TN    S  Department of Energy Technical Information Center, Oak


Giesy, J.  P. and L.  A. Briese.  1978.  Trace metal transport  by particu-
     lates  and  organic carbon in two South Carolina streams.   Vech. Inter-
     nal.. Vereln.  Limnol. 20:1401-1417.

Giesy, J.  P.,  },.  J. Kania, J. W. Bowling, R.  L.  Knight,  S. Hashburn and S.
     Clarkin.   1979.   Fate  and  Biological  Effects of Cadmium Introduced
     .into   Channel Microcosm^.   EPA-6002/3-79-039.   U.  S.  Environmental
     Protection  Agency, Athens,  Ga.   157 p.

Giesy,  J.   P.,  J.  W.   Bowling  and  H. J.  Kania.   1980.   Cadmium  and  zinc
     accumulation and  elimination  by freshwater  crayfish.  Arch.  Environ.
     Contain. Toxicol.  9:683-697.

Goldberg,  E.  D.   1975.  The  mussel  watch--A first step  in  global marine
     monitoring.   Har.  Poll.  Bull.  6:111.

Cower, A.  M.  and  P. J. Buckland.  1978.  Water quality and the occurrence
     of  Chironomus   riparius  Heigan (Diptera:Chironomidae)   in  a  stream
     receiving  sewage  effluent.   Freshwater Biol.  8:153-164.

Graham,  J.  E.  S.  and  T.  C.  Hutchinson.   1975.  The  role of temperature and
     heterotrophisra  in determining  crude oil  toxicity.  Proc.  10th Canadien
     Symp.  Water  Poll.  Re-.  Canada, p. 73-83.

Gude,  W. D.   1968.   Autorat!iographic Techniques:  Localization  of Radio-
     isotopes   i_n Biological  Material.   Prentice-Hall,   Inc.,  Englewood
     Cliffs, NJ.   113  p.

Hamelink,  J.  L.   1977.   Current bioconcentration test methods and theory.
     pp.  1&9-161.  In:   Aquatic Toxicology  and  Hazard Evaluation.  F. L.
     Mayer  and  J. L.  Hamelink (eds.), American Society for Testing and Ma-
     terials, ASTM STP 634.
                                     216

-------
Harrington,  W.  F. ,  p. Johnson  and R. H.  Ottewill.   1956.   Bovine  serum
     albumin  and  its  behavior  in acid solution.   Biochem.  J. 62:569-582.
Harri*' R-  P-»  v- Berdugo, E. D.  S.  Corner,  C. C. Kilvington and S. C. M.
       Hara.   1977.   Factors  affecting the retention of a petroleum hydro-
     carbon  by  marine  planktonic copepods.  pp.  286-304.   In:   Fate  and
     1* u°iS  2l Petroleum Hjdrocarboas in Marine Systems and Organisms, D.
     A. Wolfe  (ed.).  Pergamon Press, Kew'York.

Harrison,  A.  P.  and V. 'E.  Raabe.  1967.  Factors  influencing tne photo-
     93?6IUC  aCti°n °f benz°(aJPyrene on Excherjchia  coli.   J. Bacteriol.


Harrison,  R.  M., R. Perry  and R.  A. Veilings.  1975.  Polynuclear aromatic
     hydrocarbons in raw,  potable and waste  waters  (Review paper).  Water
     Res.  9:331-346.                                                 -

Hassett,  J.  P.  and M.  A. Anderson.   1979.   Association  of  hydrophobic
     organic  compounds with  dissolved organic matter  in aquatic systems.
     Environ.  Sci.  Tecbnol. 13:1526-1529.

Hedges,  J.  I.  aoi  P.  L.  Parker.  1976.  Land-derived  organic  matter in
     surface  sediments from the  Gulf  of  Mexico.   Geochim. Cosmochim. Acta.
     40:1019-1029.

Herbes, S.  E.  and G. F. Risi.   1978.   Metabolic alteration and  excretion of
     anthracene  by  Daphiiia  pulex.   Bull.  Environ.  Contain. Toxicol.  19:
     147-155.

Herbes,  S.  E.  and  L.  R.  Schwall.  1978.   Microbial  transformation of poly-
     cyclic aromatic  hydrocarbons  in prsitine and petroleum-contaminated
     sediments.  Appl.  Environ.  Microbiol . 35:306-316.

Herbes,  S.  E.,  L.  R.  Schwall and G.  A.  Williams.   1977.  Rate  of microbial
     transformation of polycyclic aromatic hydrocarbons:  a  chromatographic
     quantification  procedure.   Ap_p_l.   Environ.   Hicrobiol.  34:244-246.

Herbes,  S.  E. ,  G.  R.  Southworth  and C.  W.  Gehrs.   1976.  Organic  contami-
     nants  in  aqueous  coal conversion effluents:   Environmental  consequen-
     ces  and research  priorities.  pp.   295-303.  In:  Trace Substances i_n
     Environmental   Healtih,  D.  D. HemphiJl (ed.). University  of  Missouri,
     Columbia.

Hinga, K.  R. ,  Pilson, M.  E. Q. , R. F. Lee, J. W. Farrington, K. Tjessem and
     A. C.  Davis.  I960.   Biogeoche-aistry of ben/oanthracene in an enclosed
     marine ecosystem.   Environ.  Sci. Technol. 14:1136-1143.

Kites,  R.  A.,  R.  E.  Laflammc  and  J.  W. Farrington.   1977.   Sedimentary
     polycyclic  aromatic  hydrocarbons:    The  historical  record.   Science
     198:829-8J1.
                                     21?

-------
 ames, M. 0. and J. R. Bend.  1980.  Polycyclic aromatic hydrocarbon induc-
     tion of  cytochrome P-450-dependent mixed-function  oxidases  in marine
     tisn.  Toxicol. Appl.  Pnarra. 54:117-133.

Jackim,  E.  and  L. Wilson.   1977.   Benzo(a)pyrene  accumulation  and  depu-
     ration in the soft-shell clam (Hya arenia).  pp.  90.-94.   In:   Proceed-
     ings  Tenth   National  Shellfish  Sanitation  Workshop.   Hunt  Valley,
     Maryland.              —"		

Jonasson,  P.  M.   1978.   Edgardo  Baldi  Memorial  Lecture,  Zoobenthos  of
     lakes.  Verh.  Internal,.  Verein. Limnol.  20:13-37.

Johnson,  E.,  A.  Abu-Shumaya  and S.  R. Abbott.   1977.   Use of fluorescence
     detection  in high performance  liquid chromatography.   J.  Chromatogr.
     134:107-119.

Jones,  P. W.  and P.  Leber  (eds.).   1979.   Polynuclear Aromatic  Hydrocar-
     bons.  Third  Internationa^  Symposium  on Chemistry  and  Biology - Car-
     cinogenesis  and Mutagenesis.   Ann Arbor Science, Ann Arbor,  Mich. 892
     P-

Kania,  H. J., R.  L. Knight  and R.  J. Beyers.   1979.   Fate  and biologiccl
     effects  of  mercury   introduced  into  artificial  str-ams.   EPA-60013-
     76-060.  U.  S.  Environmental  Protection  Agency,  Athens,  Ga.   129  p.

Karickhoff,  S.  W.,  D.  S.  Brown  and  T. A. Scott.  1979.  Sorption of hydro-
     phobic pollutants on  natural  sediments.  Water Res. 13:241-248.

Kawatski,  J.  A.  and  M.  A.   Bittner.  1975.   Uptake,  elimination and bio-
     transformation  of the   lampricide 3-trifluormethyl 4nitrophenol (TFM)
     by  larvae  of the aqiatic  midge Chironomus tentans.   Toxicol. 4:183-
      194.

Kenaga,  E.  E.   and  C. A.  I.  Goring.  1980.   Relationship between water
     solubility,   soil sorption,  octanoi-water  partitioning,  and concen-
     tration  of  chemicals in biota.  pp.  62-77.  In:   Aquatic Toxicology.
     J.  G.  Eaton,  P.  R.  Porrish and A.  C.  Hendricks  (eds.), American Soci-
     ety for  Testing and  Materials, ASTM STP 707.

Kitchell,  J.  F.,  J.  F.  Koonce,  R.  V.  O'Neill,  H.  H. Shugart, Jr., J. J.
     Magnuson and R.  S.  Booth.   1974.  Model  of  fish biomass  dynamics.
     Trans. Amer.  Fish. Soc. 103:786-798.

Knowles,  G. and  A. C. Wakeford.   1978.  A mathematical deterministic  river-
     quality  modeJ..   Part   I:  Formulation and  description.   Water Res.
      12:1149-1153.

Konemann,  H.  and  K.  vanleeuwln.    1980.  Toxicokinetics  in fish:   Accumu-
      lation  and  elimination  of  six chlorobenzenes by guppies.   Chemosphere
     9:3-19.
                                     218

-------
If*  I
 00 CE t   M- M'' J' W>  A'  Lustenhouwer,  K. Olie and 0.  Hutzinger.   1S81.
       traction efficiencies of polychlorinated dibenzo-p-dioxins and poly-
     cnionnated  dibenzofurans  from  fly  ash.   Anal.  Chem.  53:461-463.

LaFlamme   R.  E.  and R. A. Hites.   1978.   The  global distribution of poly-
     cyciic   aromatic  hydrocarbons  in  recent sediments.   Geochimica  et
     losmocnimica Acta.  42:289-303.

LaFlamme,  R.  E.  and R. A. Hites.   1979.   TETRA-  and pentacyclic, natural-
     ly-occurring,  aromatic  hydrocarbons  in  recent  sediments.   Geochemica
     £t Losmochiinica Acta.  43:1687-1691.

LaVoie,  E. ,  L.  Tulley,  V. Bedenko  and  D. Hoffman.   1979.   Hutagenicity.
     tumor initiating  activity,  and  metabolism  of  tricyclic  polynuclear
     aromatic hydrocarbons.   pp.  1041-1058.   In:   Polynuclear  Aromatic
     Hydrocarbons:   Chemistry and  Biological  Effects.  A, Bjorseth  and A.
     J. Dennis (eds.).   Battelle  Press, Columbus, Ohio.   1097 p.

Lake,  J.  L. , C.  W.  Dimock  and  C.  B.  Norwood.   1980.   A  comparison of
     Methods  for  the Analysis  of  hydrocarbons  in Har:.ne Sediments.   Ad-
     vances   in   Chemistry;    Petroleum   in  the   Marine  Environment.   (in
     press.)

Lake,  J.  L.  and  C.  Kershner.   1977.   Petroleum sulfur-containing compounds
     and  aromatic  hydrocarbons  in the  marine mollusks  Modiolus cl;missus
     and  crassastrea virginicji.   Oil Spill Conf., pp. 627-632.

Landrum,  P.  F. and  D.  G. Ccosby.  1981a.  Comparison of the disposition ot
     several  nitrogen  containing  compounds   in   the  sea  urchin  and other
     marine  invertebrates.  Xenobiotica  11:351-361.

Landrum,  P.  F. and G.  D. Crosby.  198]b.  The Disposition  of p-Nitroanisole
     by  the  sea  urchin, Strongylocentrolus  purpuratus,  biotransformaticn
     and  bioconcentration.  Ecotoxicol.  Environ.  Safety   5:240-254.

Landrum,  P.  F. and  J.  P. Giesy.  1981.   Anomolous  breakthrough of benzo(a)
     pyrene  during  concentration with Arabcrlite® XAD-4 resin  from  aqueous
     solutions.    pp.  345-355.   In:  Advances  it»  the  Identification and
     Analysis 
-------
 Ce> Uotak' R' .Sauerheber a"d A.  A.  Benson.   1972.   Petroleum  hydrocarbons:
     344-346    Dlscharge by  the marine uussel Hytilus endulis. Sci.  177:
 Ce'
R' Sauerheber and G. H.
     Biol

                                     Dobbs.   1972b.   Uptake,  metabolism and
                              ar°raatic hydrocarbons  by  marine  fish.   Mar.
          ''  J'tC'.fS.in8er* K' R'  Ten°re,  W-  S.  Gardner and R. M.  Philpot.
     H«   A   . toxl?lcatl°n  system in polychaete worms:  importance  in  the
     aegraoation of  sediment  hydrocarbons,   pp.  23-37.   In:   Marine  Pollu-
     tion:   Functional  Res£onses,  W.  B.  Vernberg,  F.  P.  Thu7ber^7~A.
     Calabreso and F. J. Vernberg (eds.).   Academic Press,  New York.

Lee, R.  F.  and H. Takahashi.   1977.   The fate  and effect of  petroleum  in
     controlled  ecosystem enclosures.   Rapp. P.V-Reun. Cons.  int. Explor.
     Her. 171:150-156.                  ~ ^  -  - -   -  — * -

Lee, R.  F.,  W.  S.  Gardner,  J.  W.  Anderson,  J. W.  B.aylock and J. Bar-
     well-Clark.   1978.   Fate of polycyclic aromatic hydrocarbons in con-
     trolled  ecosystem  enclosures.   Environ.   Sci.  Technol.  12:832-838.

Lehr, R. E.. H. Yagi, D.  R. Thakker, W. Levin,  A. W. Wood,  A.  H.  Couney  and
     D.  H.  Jerina.   1978.   The  bay  region  theory of  polycyclic  aromatic
     hydrocarbon induced  carcinogenicity.  In:   Carcinogenesis , Vol.  3.   P.
     W.  Jones and R.  I.  Freudenthof  (eds.).    A,  Raven  Press, New  York.

Leo, A.,  C.  Hansch and  D. Elkins.   1971.  Partition coefficients and their
     uses.  Chemical  Reviews  71 (G) :525-6l6.

Leo, A.  J.   1975.   Calculation of  partition  coefficients useful  in  the
     evaluation  of the  relative hazards  of various  chemicals  in the  en-
     vironment.     In:    Symposium  on Structure-Activity Correlations  in
     Studies of Toxicity and  Bioconcentration with  Aquatic Organisms. G.  D.
     Veith and D.  E.  Konasewich (eds.), Internat. Joint Commission. Winsor,
     Ontario, Canada.

Leversee, G.  J. ,  J.   P.  Giesy,  P. F.  Landrum, S.  Gerould, J. W.  Bowling, T.
     Fannin,  J.  Haddock and  S.  Bartell.   1981a.   Kinetics and biotransfor-
     mation  of  Benzo(a)pyrene  by Chironooius  riparius.   Arch.  Environ.
     Contamn. Toxicol.  11:25-31.

Leversee,  G. J.  J.  P.  Giesy,   P.  F.  Landrura,  S.  Gerould,  M. Bruno,  A.
     Epacie,  S.  Bartell, J.  Bowling,  J. Haddock and  T.  Fannin.  198lb.
     Disposition of  benzo(a)pyrene in aquatic  systems components:  periphy-
     ton,  chironomids,  fish.   pp.  357-366.   In:   Chemical  Analysis   and
     Biological Fate:  Polynuclear Aromatic Hydrocarbons.  M.  Cooke and A.
     J. Dennis (eds.),  Battelle Press, Columbus, Ohio.
                                     220

-------
     fate of- '  "etcalf-  N- Plummer and D.  Handel.   1977.   The environmental
     ride evaul\^A Ci>rcino8ens:   Benzo(a)Pyrene,  Benzidine and Vinyl Chlo-
     Toxicol  63129 ln laboratory fficdel ecosystems.   Arch.  Environ.  Contain.


 "'  fiJp'  ^'- L"i "etcalf an'-i E-  M.  Carlson.   1978.   Environmental  fate of
     t«™ «rfT    lcd  coal  conversion by-products  evaluated in a  labora-
     tory model ecosystem.  Environ. Health Perspect. 24:201-208.

McCain, B  B., H  0  Hodgins, W. D. Gronlund,  J.  W.  Hawkes, D. W. Brown, M.
     ™,   •  Vi, Vandenneulen.  1978.  Bioavailability of  crudt oil  from
     !S  TK  ,   y  °iled sediments  to English  sole  (Parophrys vetulus),
     and  pathological  consequences.  J.  Fish. Res.  Boa~rd  Can.  357657^664.

Mclntire,  C,  D.   1973.  Periphyton  dynamics  in  laboratory streams:   A
     simulation  model  and  its  implications.   Ecol.  Monogr.  43:399-420.

     /nt'/o;  °'T 1975-    PeriPhyto" assemblages in laboratory streams,   pp.
     403-430.   In:    River  Ecology.  B. A.  Whittoa (ed.).   University of
     California  Press, Berkeley.

Mclntire,  C.  W.  and J.  A.  Colby.   1978.  A hierarchical model of lotic
     ecosystems.  Ecol.  Monogr.  48:167-190.

Mclntire, C. D.  and H. K. Phinney.   1965.  Laboratory studies of periphyton
     production  and  community  metabolism in  lotic  environments.   Ecol.
     Monogr. 35:237-258.

HcLeese,  D.  W. ,  V.  Zitko and D. B. Sergeant.  1979.  Uptake and excretion
     of  fenitrothione hy claras  and mussels.   Bull.  Environ.  Contarn. Toxi-
     col. 22:800-806.

Maher,  W.  A., J.  Bagg  and  D.  Smith.   1978.   Determination of polycyclic
     aromatic  hydrocarbons  in  marine sediments,  using solvent  extraction,
     thin loyer  chromatography and spectrofluorimetry.  Intern.  J. Environ.
     Anal. Chem. 4:1-11.

Means,  C.  E.  and R.  E.  Feeney.   1971.  Chemical Modifications of Proteins.
     Holden-Day, San  Francisco,  149 p.

Melancon, M. J., Jr. and J.  J. Lech.  1978.  Distribution  and  elimination
     of  naphthalene and 2-methylnaphthalene  in  rainbow trout during short
     and  long-term exposures.   Arch. Environ. Contain. Toxicol. 7:207-220.

Melancon, M. J. , Jr. and J.  J. Lech.  1979.  Uptake,  a  biotran^fonnation,
     disposition, and elimination of 2-metuylnaphtbalene  and naphthalene  in
     several  fish  species.   pp.  5-22.    In:   Aquatic Toxicology,  L.  L.
     Marking  and R.  A.  Kimerle  (Eds.),  pp. 5-22.   American  Society  for
     Testing and Materials,  ASTM STP 667.

Metcalf, R., C. K. Sangha and I. P. Kapoor.   1971.   Model  ecosystem  for the
     evaluation of pesticide biodegiadability and ecological magaification.
     Environ. Sci. Technol.  5:709-713.

                                     221

-------
Miller   G   r
     Photo'lvt' 8nd  R'  G' ZePP-   1979-   Effects of suspended  sediments  on
           *  S  rates  of  dissolved  pollutants.   Water  Res.  13:453-459.
Minshall,  G.  W   1970
     707-771   '          ftutotrophy in Stream Ecosystems.   Bio-science  28:


     «nH   ^K»  v.  i! . Ce*lular responses to polycyclic  aromatic hydrocarbons
     and   phenobarbltal  iu  Mytilus  edulis    Marine  Environ.  2:255-263.
Moore
         l,' aG'  Gri1mmer and H. Bohnke.  1977.  Sedimentary  record  of heavy
                                                         ^stance.   Natur-
                              (USA).   1972.   Particulate  Polycyclic  Organic


Neal, EB  Patten and C.  DePoe.   1967.   Periphyton  growth  on artificial
     suostrates  in a radioactively contaminated lake.   Ecology 4f. -918-924.

Nebert   D. ,  and H. Gelboin.   1968.   Substrate inducible microsomal  aryl
     k'^°"rb°n hydroxylases in «""»alian  cell cultures.  J.  Biol.  Chem.
     2'»3:62'>2-6249 .                                         ~  -  -

Neely, W.  B.   1979.  Estimating  rate  constants for the uptake  and clearance
     of chemicals  by fish.   Environ.  Sci. Technol. 13:1506-1510.

Neely, W.  B. , D.  R. Branson and G. E. Blau.   1974.  Partition coefficient
     to  measure bioconcentratioa  potential of organic  chemicals  in  fish.
     Env.  Sci. Technol.  8:1113-1115.

Neff, J.  M.,  B. A.  Cox,  D.  Dixit,  and J. W.  Anderson.   1976.   Accumulation
     and  release of petroleum-derived aromatic  hydrocarbons by  four species
     of marine a imals.   Har. Biol. 38:279-289.

Neff, J. M.   1982.   Polycyclic aromatic hydrocarbons in  the aquatic enviro-
     nment and  cancer  risk  to  aquatic organisms  and man.   In: Symposium:
     Carcinogenic  Polynuclear Aromatic Hydrocarbons ii\  the Marine Environ-
     ment.  N.  L.  Richards  and  B.  L. Jackson  (eds.), EPA-600/9-82-013, U.
     S. Environmental  Protection Agency,  Pensacola,  Florida.

Neff, J.   H.   1979.  Polycyclic Aromatic  Hydrocarbons  in  the Aquatic En-
     vironment.  Applied Science Publishers Ltd.,  London.   262 p.

Norden,  B.,  U.  F.dlund and  S. Wold.   1979.   Carcinogenicity of  polycyclic
     aromatic  hydrocarbons   studied  by SIMCA pattern  recognition.   pp.
     345-366.    In:    Polynuclear  Aromatic  Hydrocarbons:    Chemistry and
     Biological  Effects.  A, Bjorseth  and A. J.  Dennis  (cds.).   Battelle
     Press, Columbus,  Ohio.   1097 p.

Nunes, P.  and  P.  E. Benville.   1979.  Uptake and  depuration of petroleum
     hydrocarbons  in  the Manila  clam,  Tapes semidecussata  Reeve.   Bull.
     Environm. Contain.  Toxicol.   21:719-726.
                                     222

-------
       *,     V-    1971.   Systems approaches  to the  study  of forest  floor
     Ecol         PP'  44l'A77-   In:   Systems  Analysis  and  Simulation  in
     6crH^'  V°1-  l'   patte«\, B.  C.  (ed.).  Academic Pr*ss,  New  York.

O'Neill, R   V    ,0-,,    _
     Ecology 57-15/A    '-cosystein persistence and heterotrophir  regulation.

O'Neill,  R  v    ft  A   />  ,
     Eracuct.  'r         Goldstein  and H.  H.  Shugart.   1972.   Terrestrial
     Ecosystem Energy Model.   USIBP-EDFB Memo Rep  72-19.

Odum.  H.  T.  1957   T~~ u-
     FlnriH*   r   i  lr°Pnic  structure and productivity  of silver springs,
     xAoriaa.  fccol. Mpnogr.  27:55-112.
  1Ve-1971   Llfe history of the Chironoraidae.   Ann.  Rev.  Entomcl.
Overton   E  B., J   Bracken end J. L. Caseter.  1977.  Application of glass
      capillary columns  to monitor  petroleum-tvpe  hydrocarbons  in  marine
      sediments.   J.  Chroma tog. Sci.  15:169-173.'
Payer,  H.  D. and  C.  J.  Soeder.   1975.  Accumulation of polycyclic aromatic
      hydrocarbons in cultivated  micrcalgae.  Naturvissenschaf te 62:536-537.

Payne,  J.  F.  and W.  R.  Penrose.   1975.   Induction of  aryl hydrocarbon
      (benzo(a)pyrene)  hydroxylase  in  fish  by petrolem.    Bull.  Environ.
      Contaran. Toxicol. 14:112-116.                          -   -

Patten,  B.  C.   1968.   Mathematical models  of plankton  production.   Int.
      Rev.  Ges.  Hydrobiol. 53:357-408.

Pedersen,  M. G., W. K. Hershherger, and M. R. Juchau.   1974.   Metabolism of
      3,4-benzpyrene  in rainbow trout  (Salmo  gairdneri).   Bull. Environ.
      Contain. Toxicol. 12:481-486.

Pitts,  J.  N.,  K.  A.  Van  Cauwenbetghe,  D.  Grosjeac, J.  P. Schmid,  D. R.
      Fitz,  W.  L.   Belser,  G. B.  Knuds->n,  and P. M.  Hynds.   1978.  Atmos-
      pheric  reactions  of po'ycyclic aromatic hydrocarbons:   Facile forma-
      tion of mutagenic nitro derivatives.  Science 202:515-519.

Prabl,  F. G. and  R. Carpenter.   1979.   The role of  zooplankton fecal  pel-
      lets in the sedimentation  of polycyclic aromatic hydrocarbons in  Dabob
      Bay,  Washingtor.  Geochin.  Cosmochim. Acta.  43:1959-1972.

Rose, F.  L.  rnd  C.  E.  Cushing.   1970.   Periphyton:   autoradiography  of
      zinc-65 adsorption.  Science 168:576-577.

Koubal,  W.  T., T.  K.  Collier,  and D.  C. Malins.   1977.   Accumulation and
      metabolism of carbon-14  labeled benzene,  naphthalene,  and  anthracene
      by young  coho salmon  (Oncorhynchus kisutch).   Arch.  Environ, Cojitam.
      Tcocicol. 5'; 5 13 -529.
                                      223

-------
                                                                     •
Roubal, W.  T.,  S.  I. Stranahin, and D. C. Halins.  1978.  The accumulation
     of  low molecular weight  aromatic hydrocarbons  of  crude oil  by coho
     salmon (Oncorhynchus kisutch) and starry flounder (Platichthys stella-
     tus).  Arch. Euviron. Contain. Toxicoi. 7:237-244.

Saidoval, M., F. H. Verhoff and T. H. Cahill.  1976.  Mathematical modeling
     of nutrient  cycling in rivers,   pp. 205-232.  In:  Modeling Biochemi-
     cal  Processes  in Aquatic  Ecosystems.   R. P.  Canale  (ed.),  Ann Arbor
 »   Science, Ann Arbcr, Mich. 389 p.

Satterluad, D. R. and J. F.. Means.  1978.  Estimating solar radiation under
     variable cloud conditions.  Forest Sci. 24:363-373.

Schnell,  J. V.,  E.  H. Gruger, Jr. and D. C. Malins.  1980.  Mouo-oxygenase
     activities  of  coho  salmon   (Oncorhynchus  kisutch)  liver  microsomes
     using  three polycyclic aromatic  hydrocarbon substrates.  Xenobiotica
     10:  229-234.

Sheldon,  L.  S.,  Kites,  R.  A.  1978.   Organic  Compounds in  the Delaware
     River. Env. Sci. Techno!. 12:1188-1194.

Shindler,  D.  B. , B.  F.  Scott  and D. B.  Carlisle.   1975.   Effect of crude
     oil  on populations  of bacteria and  algae  in artificial ponds subject
     to winter  weather  and ice formation.  Verb. Internal. Verein. Limnol.
     19:2138-2144.

Shngart,  H. H., R.  A.  Goldstein  and R.  V. O'Neill.   197^.   TEEM:   A ter-
     restrial  ecosystem energy model for forests.   Oecol.  Plant. 25: 251-
     234.

Smith, E.  L.   1936.  Photosynthesis in  relation  to light and carbon diox-
     ide.   Proc. Nat. Acad. Sci. 22:504-511.

Smith, 0. L., H. H. Shugart, R. V. O'Neill, R. S. Booth and D. C. McNaught.
     1975.   Resource competitioa  and  an analytical uodel  of zooplankton
     feeding  on  phytoplankton.  Amer. Nat.  109:571-591.

Soto,  C.,  J.  A.  Hellebust and T. C. Hutchiuson.  1975.  Effect of naphtha-
     lene  and aqueous crude oil extracts on the green flagellate Chlamydo-
     tnonas  ajit;ulosd.   II.   Photosynthesis and  the uptake  and  release of
     naphthalene^Can.  J. Bot. 53:118-126.

Soto,  C.,  J.  A.  Hellebust, T. C.  Hutchinson  and R. G. Sheath.   1979.  The
     effect oi'  crude oil on the morphology of the  green flagellate Chlaniy-
     do£ona_s  angulo_s_a.   Can. J. Bot. 57:2717-2728.

Soto,  C.,  J.  A.  Hellebust, T. C. Hutchinson and R. G. Sheath.  1979b.  The
     effect of  the  hydrocarbon naphthalene on  the  morphology of the green
     flagellate  Chlamydomonas angulosa.   Can. J. Bat. 57:2729-2739.
                                    2'>4

-------
Southworth, G. R.  1979.  The role of volatilization in removing polycyclic
     aromatic  hydrocarbons  from  aquatic  environments.   Bull.  Environ.
     Contamn. Toxicol. 21:507-5)4.

SouLhworth, G.  R.,  J. J. Beauchamp and P. K. Schmieder.  1978.  Bioaccumu-
     lation potential of polycyclic aromatic hydrocarbons in Dephnia pulex.
     Water Res.  12:973-977.

Southworth, G. R., C. C. Keffer, and J. J. Beauchamp.  1980.  Potential and
     realized  bioconcentration.   A  comparison of  observed  and predicted
     bioconcentration  of azaarenes  in the  fathead  (Pimephales  promelas).
     Environ. Sci. Technol. 14:1529-1531.

Spacie,  A.  and  J.  Hamelink.   1981.   Recent advances  in  the  prediction of
     bioconcentration  in fish.   J.  Environm.  Toxicol. Chem.  (in  press).

Spikes, J. D.  1977.  Photosensitization.  pp.  87-112.  In:   The Science of
     Photobiology, K. C. Smith (ed.), Plenum Presr, New York,  87 p.

Statham, C. N.,  C. R. Elcombe, S. P. Szyjka, and J. J. Lech.  1978.   Effect
     of  polycyclic  aromatic  hydrocarbons on hepatic microsomal enzymes and
     disposition of methylnaphthalene in  rainbow  trout in  vivo.  Xenobio-
     tica 8:65-71.

Stegeman, J. J.  and S. M. Teal.  1973.  Accumulation, release  and retention
     of  petroleum hydrocarbons by the oyster Crassostrea virginica.   Marine
     Biol. 22:37-44.

Suess,  M.  J.   1976.   The environmental load and  cycle of  polycyclic aro-
     matic hydrocarbons.  Sci. Total Environ. 6:239-250.

Teal, J. M.  1957.  Community metabolism in a temperate cold spring.  Ecol.
     Honogr. 27:283-302.

Teal,  J. M.,  K.  Burns and  J.  Farrington.   1978.  Analyses  of aromatic
     hydrocarbons in  intertidal  sediments resulting from two spills of No.
     2  fuel  oil  in  Buzzards Bay, Massachusetts.  J.  Fish. Res  Board Can.
     15:510-520.

Tilly,  L. J.    1968.   The  structure and  dynamics  of  Cone Spring.   Ecol.
     Monogr. 38:169-197.

Tulp, M. T. M. and 0. Hutzinger.  1978.  Some thoughts on aqueous solubili-
     ties and  partition coefficients of PCB, and the mathematical correla-
     tion between bioaccumulation and physico-chemical properties.  Chemos-
     phere 10:849-860.

Varanasi,  1).,  D.  J.  Gmur,  and W. L. Weichert.   1981.   Effect of environ-
     mental temperature  on naphthalene metabolism by juvenil^ stcrry floun-
     der  (Platichthys  stellatus).   Arch.  Environm.  Contain.  Toxicol.   10:
     203-214.
                                    225

-------
Veith,  G.  D.,  D.  L.  DeFoe,  and B.  V.  Bergstedt.  1979a.   Measuring and
     estimating the bxoconcentration factor of chemicals in fish.  J. Fish.
     Res. Bd. Canada.  36:1040-1048.

Veith,  G.  D., K.  J.  Macek,  S.  R. Petrocelli, and J.  Carroll.   1979b.  An
     evaluation  of  using  partition  coefficients  and water  solubility to
     estimate bioconcentratiou  factors  for organic chemicals in fish. Fed.
     Regist.  15926  (March 15, 1979).

Wagner,  J.   G.    1975.   Fundamentals of Clinical Pharmacokinetics.   Drug
     Intelligence  Publications  Inc., Hamilton Press, Hamilton, IL.

Walsh,  G.  E., K.  E,  Ainsworth  and L. Faas.  1977.  Effects  and uptake of
     chlorinated  naphthalenes in marine unicellular algae.  Bull. Environ.
     Contamn. Toxicol.  18:297-301.

Walshe,  B.  M.   1949.   The function of  haemoglobin  in Chironomus plumosus
     under natural  conditions.  J. Zxp. Biol. 27:73-95.

Weininger, D. L.   1978.  Accumulation of PCB's by lake front in Lake Michi-
     gan.  Ph.D. Thesis, University of Wisconsin, Madison, 232 p.

Wershaw,  R.  L.,  P.  J. Burcar,  and M. C.  Goldberg.   1969.   Interaction of
     pesticides  and   natural  organic  material.   Environ.  Sci.  Technol.
     3(3):271-273.

Wilkfcs,  F.  G. ,  and C. M.  Weiss.   1071.  The  accumulation of  DDT  by the
     dragonfly  nymph Tetragoneuria.   Trans.  Amer.  Fish. Soc.  100:222-236.

Windsor, J. G. and  K.  A. Kites.    1979.  Polycyclic aromatic hydrocarbons in
     Gulf  of Maine  sediwents and Nova  Scotia  soils.   Geochitn.  Cosmochim.
     Acta. 43:27-33.

Zalucki, M.  P.   1978.  Modeling and  simulation  of the energy-flow through
     Root Spring,  Massachusetts.  Ecology 59:654-659.

Zepp,  R.  G.  and D. M. Cline.   1977.  Rates of direct photolysis in aquatic
     environment.   Environ. Sci. Technol.  11:359-366.

Zepp,  R.  G.  and  P. F.  Schlotzhauer.   1979.  Photoreactivity  of selected
     aromatic hydrocarbons  in  water.     pp.   141-158.   In:   Polynu-lear
     Aromatic Hydrocarbons.  Third International Symposium on Chemistry and
     Biology  ~  Carcinogensis and  Hutagenesis.   P.  W.  Jones and  P.  Leber
     (eds.).  Ann  Arbor Science, Ann Arbor, Mich.
                                    226

-------

-------