EPA-600/3-76-060
August 1976
Ecological Research Series
                FATE AND  BIOLOGICAL EFFECTS OF
                         MERCURY  INTRODUCED  INTO
                                 ARTIFICIAL STREAMS
                                      Environmental Research Laboratory
                                     Office of Research and Development
                                    U.S. Environmental Protection Agency
                                            Athens, Georgia  30601

-------
                 RESEARCH REPORTING SERIES

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

      1.    Environmental Health Effects Research
      2.    Environmental Protection Technology
      3.    Ecological  Research
      4.    Environmental Monitoring
      5.    Socioeconomic Environmental Studies

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

-------
                                       EPA-600/3-76-060
                                       August 1976
            FATE AND BIOLOGICAL EFFECTS OF

      MERCURY INTRODUCED INTO ARTIFICIAL STREAMS
                          by

Henry J. Kania, Robert L. Knight, and Robert J. Beyers
           Savannah River Ecology Laboratory
                 Institute of Ecology
                 University of Georgia
                Athens, Georgia  30601
                  Grant No. R. 800510
                    Project Officer

                    Harvey W. Holm
             Environmental Systems Branch
           Environmental Research Laboratory
                Athens, Georgia  30601
           ENVIRONMENTAL RESEARCH LABORATORY
          OFFICE OF RESEARCH AND DEVELOPMENT
         U.S. ENVIRONMENTAL PROTECTION AGENCY
                ATHENS, GEORGIA  30601

-------
                         DISCLAIMER
     This report has been reviewed by the Environmental
Research Laboratory, U.S. Environmental Protection Agency,
and approved for publication.  Approval 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

-------
                         ABSTRACT
Mercuric ion was continuously input to artificial stream
channels to provide water concentrations of 0.01, 1.0 and
5.0 yg/1.  Channel components were periodically sampled for
total mercury analyses.  The effects of mercury on the algal
components of the periphytic communities and on the benthic
insects were determined.  The sampling program continued
one full year after mercury inputs were stopped.

Approximately 15% of the added mercury was removed from the
water.  The community components acquired very high
concentrations of mercury, although in most cases the
levels in these were not a linear function of the water
levels.  Concentrations in invertebrates decreased most
rapidly after mercury inputs were stopped while the sediment
levels decreased most slowly.

Periphytic algae showed several treatment responses including
total inhibition, inhibition of certain life stages, possible
stimulation of certain life stages and morphological
alterations.  One species was found only in the treated
channels and disappeared when mercury inputs were stopped.

The diversity and evenness of benthic insect communities
were affected by mercury treatment.

This report was submitted in fulfillment of Project Number
R800510 by the University of Georgia under the sponsorship
of the Environmental Protection Agency.  Work was completed
as of April 1, 1975.
                              111

-------
                          CONTENTS

Sections                                            Page
I     Introduction                                    1
II    Conclusions                                     4
III   Recommendations                                 6
IV    Experimental Design and Site Description        7
V     Methods                                        16
VI    Results and Discussion                         34
VII   References                                    115
VIII  Appendix                                      123

-------
                          FIGURES


 No.                                                   - Page

  1   Artificial  streams                                   8

  2   Limestone-filled treatment system               ,    10

  3   Average  change in channel water-hardness            11
     between  limestone changes

  4   Mercury  recovery system                             13

  5   Air sampling  system                                 18

  6   Continuous  flow chamber system                      32

  7   Total mercury levels in channel and effluent        35
     water

  8   Diurnal  radiomercury uptake curve, June, 1972       37

  9   Diurnal  radiomercury uptake curve, August, 1972     38

 10   Diurnal  mercury uptake in treated channels on       39
     January  22-23, 1974

 11   Proportion  of mercury associated with particulates  41
     in  water samples

 12   Bacterial populations in channel water              42

 13   Average  mercury releases at the air-water  interface 44
     of  the treated channels

 14   Diurnal  mercury releases on January 22-23, 1975,    46
     at  the air-water interface of channels  receiving
     an  input of 1.0 yg/1

 15  Diurnal  mercury releases on January 22-23, 1975,    47
    at  the air-water interface of channels  receiving
    an  input of 5.0 ug/1

16  Average  mercury concentrations  in the organic       49
    portion  of the sediments of the control channels
    and channels  receiving an input of 1.0  ug/1
                             VI

-------
                                                       Page

17  Average mercury concentrations in the organic       50
    portion of the sediments of channels receiving
    inputs of 1.0 and 5.0 yg/1

18  Average mercury concentrations in particulate       52
    matter leaving the treated channels

19  Average mercury concentration in periphyton         53
    removed from the PVC walls of the control
    channels and channels receiving an input of
    0.01 yg/1

20  Average mercury concentrations in periphyton        55
    removed from the PVC walls of channels receiving
    inputs of 1.0 and 5.0 yg/1

21  Average mercury concentrations in periphyton        56
    removed from plexiglass plates in channels
    receiving inputs of 1.0 and 5.0 yg/1

22  Average mercury concentrations in roots of Juncus   59
    diffusissimus removed from channels receiving
    inputs of 1.0 and 5.0 yg/1

23  Average mercury concentrations in leaves of Juncus  60
    diffusissimus removed from channels receiving
    inputs of 1.0 and 5.0 yg/1

24  Average mercury concentrations in damselfly nymphs  62
    collected from channels receiving inputs of 1.0
    and 5.0 yg/1

25  Mercury concentrations acquired by corixids         64
    confined to flowing water containing 5.0 yg/1 Hg++

26  Mercury concentrations in mosquitofish removed      66
    during the first year from channels receiving an
    input of 1.0 yg/1

27  Mercury concentrations in mosquitofish removed      68
    during the first year from the control channels
    and the channels receiving an input of 0.01 yg/1

28  Mercury concentrations in mosquitofish removed      70
    from treated channels during the last 21 months
    of the project
                            Vll

-------
                                                      Page

29  Change in mercury concentrations  in  mosquitofish    72
    moved from a contaminated area to clear  water

30  Biomass of periphyton scraped from PVC strips       78
    suspended in the channels

31  Biomass of periphyton scraped from channel  walls    79

32  Calculated volume of periphytic algae scraped       81
    from the channel walls

33  Diversity and evenness of periphytic algae          82
    collected from the channel walls

34  Biomass of periphyton scraped from vertical        83
    plexiglass plates

35  Calculated volume of periphytic algae scraped       85
    from vertical plexiglass plates

36  Diversity and evenness of periphytic algae          86
    collected from vertical plexiglass plates

37  Calculated volume of algae colonizing microscope    87
    slides during mercury inputs (started October  31,
    1973)

38  Calculated volume of algae colonizing microscope    88
    slides during mercury inputs (started December 7,
    1973)

39  Calculated volume of algae colonizing microscope    89
    slides after mercury inputs were stopped (started
    January 30, 1974)

40  Calculated volume of algae colonizing microscope   90
    slides after mercury inputs were stopped (started
    April 25, 1975)

41  Density of Oedogonium sp. on channel walls before  94
    and after mercury inputs were stopped

42  Density of Oedogonium sp.on plexiglass plates      95
    before and after mercury inputs were stopped

43  Density of Stigeoclonium elongatum on plexiglass   96
    plates before and after mercury inputs were
    stopped
                           VI11

-------
                                                      Page

44  Density of StigeocIonium elongatum on channel       98
    walls before and after mercury inputs were stopped

45  Geminella turfosa                                  100

46  Density of Cosmarium aspherosporum on plexiglass   102
    plates before and after mercury inputs were
    stopped

47  Density of Cosmarium aspherosporum on channel      103
    walls before and after mercury inputs were
    stopped

48  Density of Spondylosium planum on plexiglass       105
    plates before and after mercury inputs were
    stopped

49  Density of Spondylosium planum on channel walls    106
    before and after mercury inputs were stopped

50  Total macrophyte biomass in channels during the    109
    final year of the project

51  Densities of insects collected from benthic        111
    dishes before and after mercury inputs were
    stopped

52  Diversity and evenness of insects collected from   112
    benthic dishes before and after mercury inputs
    were stopped
                            IX

-------
                          TABLES


No.                                                   Page

1  Well water quality                                   12

2  Limestone composition                                14

3  Distribution of Juncus diffusissimus on April 2,     58
   1973, in the artificial streams

4  Organisms collected from the artificial streams,     76
   February 11-16, 1975

5  Incidence of abnormal Geminella turfosa cells on    101
   glass slides during and after mercury input

-------
                      ACKNOWLEDGMENTS
The support and assistance of the personnel of the EPA's
Environmental Research Laboratory and the
University of Georgia's Savannah River Ecology Laboratory
is gratefully acknowledged.  Special recognition is due
Dr. Harvey Holm of the EPA, who patiently provided
competent technical and administrative direction throughout
the project, and Mrs. Vicki Buley of the University of
Georgia, who has willingly participated in every aspect
of the program, from the tedious analytical work to the
typing and editing of manuscripts.
                             XI

-------
                         SECTION I

                       INTRODUCTION
The toxic effects of Hg and its compounds on both plants
and animals have been known for a long time.  It is only
within the last twenty-five years, however, that the extent
of man's uses of this element, and the actual and potential
problems resulting from his uses, have been recognized.
The history of Hg, its chemical and physical properties,
sources, uses, and toxicological effects as related to
current problems have been reviewed in a number of recent
books  (Jones, 1971; D'ltri, 1972; Friberg and Vostal,
1972; Hartung and Dinman, 1972), dissertations (Curtis,
1974; Ferens, 1974), and review papers (Nelson, 1971;
Wallace et al., 1971; Peakall and Lovett, 1972;
Putman, 1972; Saha, 1972; Lepple, 1973).

At present, high-level environmental releases of Hg from
point sources such as chlor-alkali plants and paper pulp
operations have been virtually eliminated.  Releases due
to agricultural usage of organomercurials have also been
greatly reduced.  It is unlikely that we will see another
Minimata Bay type incident, although the poisoning of man
by the misuse of grain treated with mercurials remains a
possibility.  The two major problems still facing man with
respect to mercury are:   (1) what to do about the areas
already contaminated by large releases in the past and,
(2) what are to be the consequences of the continuing large-
scale diffuse releases of Hg from the burning of fossil
fuels and the discarding and incineration of Hg-containing
products.

Since the recognition that Hg releases of any kind may
result in ecosystem alterations and even poisoning of man
himself, a great deal of research dealing with this metal
has been initiated.  The majority of the studies have been
surveys or experiments of short duration involving
relatively high Hg levels and/or very simple biological
systems, or performed in existing contaminated areas without
adequate controls and replication.  These studies have
demonstrated that:

     (1)  Mercury is universally distributed and occurs in
small amounts in all components of the biosphere.  The
inability to find Hg in a sample is probably a result of
the analytical techniques used.

-------
     (2)   There is  a natural biogeochemical cycle in which
 Hg circulates between  living  organisms and the environment.
 Inputs to the natural  cycle include releases  from the
 earth's  crust by the weathering of Hg-bearing rocks,
 volatilization of  Hg from covered deposits, leaching by
 ground water, and  releases associated with volcanic activity
 and thermal springs.

     (3)   Mercury can exist in three general forms  (organic
 compounds,  inorganic compounds and free metal) which differ
 greatly  in their physical and chemical properties and,
 therefore,  their behavior in  biological systems.  These
 forms have different accumulation and clearing rates, sites
 and physiological  effects on  living organisms.  All possible
 conversions between forms can and do take place in natural
 systems.

     (4)   Man adds  to the  Hg load of the environment in
 several  ways.  Point source releases have caused localized
 problems including human  deaths.  Agricultural use of
 mercurials has resulted in wildlife losses.   Although point
 source and agricultural releases have been almost entirely
 eliminated in the  past few years, contaminated aquatic areas
 still exist and fish from these areas remain  unfit for
 human consumption.   There continue to be extremely large
 releases of Hg wherever fossil fuels are utilized.

      (5)  From the standpoint of human health, methyl Hg
 compounds are the  most important because this is the form
 that tends to accumulate  in food organisms.   Methyl Hg is
 almost completely  absorbed in the gut of mammals, is very
 slowly eliminated,  has a  great effect on the  central
 nervous  system at  low  levels,  and easily passes through
 placenta! membranes.

      (6)   Relatively low  levels of all forms  of Hg may be
 toxic  to aquatic organisms.   The toxicity is  a function
 of  species  and water quality.

 In  spite of  the recent surge  of Hg-related research, several
 important questions  have  not  been fully answered.  The site
 of methylation of the  Hg  that occurs in fish  as methyl Hg
 has not  been  clearly demonstrated.  The effects of long-term
 exposures of  biological systems to slightly elevated Hg
 levels have not been adequately investigated.  The rates of
recovery  to be expected from  contaminated systems are not
clearly  defined.  And  the maximum levels that will not cause
environmental  alterations or  the contamination of human  food
resources have not been determined.  The research program

-------
reported here was directed to the last three points as
pertaining to aquatic systems.  More specifically, the
artificial stream research program had the followinq
general goals:   (a) to determine the fate of Hg introduced
on a continuous basis at low levels into biologically
complex aquatic communities over a long period of time,
(b) to determine the fate of Hg remaining in the systems
after the inputs were stopped, (c) to assess the impact of
the added Hg on the communities in the systems and, (d) to
compare results obtained from the artificial streams with
data from other sources in order to evaluate the artificial
streams as a potential tool for the future routine screening
of environmental pollutants.

-------
                        SECTION II

                        CONCLUSIONS

 1.  Water levels of only 0.01 yg/1 Hg++ above ambient
    resulted in Hg levels significantly greater than
    background in components of the communities that
    developed in the artificial streams.

 2.  At 1.0 and 5.0 yg/1 Hg++, very high Hg concentrations
    accumulated in the biota and sediments of the treated
    channels although, except for insects and macrophytes,
    the relationship between water levels and the levels
    measured in the community components was not linear.

 3.  In a simple food chain  (periphyton-midge larvae-
    mosquitofish) no biomagnification was observed.

 4.  At dosing levels of 1.0 and 5.0 yg/1, approximately
    15% of the Hg was removed from the water during the
    two-hour retention time of the channels.  Approximately
    33% of the Hg leaving the 1.0 yg/1 channels and 21%
    of that leaving the 5.0 yg/1 channels was attached
    to microscopic particulates.  No diurnal Hg uptake
    patterns were observed, indicating uptake was by
    passive adsorption in the channels.

 5.  The major loss of fixed Hg from the channels was with
    exported macroscopic matter.  These losses were erratic
    and very much affected by weather conditions.

 6.  A loss of elemental Hg from the channels at the air-
    water interface was observed and this loss rate was
    linearly related to water concentrations of Hg and
    not to sediment levels.  The release of Hg from the
    channels appeared to be related to the diurnal
    photosynthetic activities of channel communities.
    The maximum releases measured were always less than
    1% of the Hg present in the water.

7.   After Hg inputs were stopped, there were major
    differences  between the clearing rates of Hg from
    the various  channel components.  The sediments showed
    greatest Hg  retention (T% = 7.85 years, C.I.95 =
    5.6-13.0 years in 1.0 yg/1 channels, T% = 8.67 years,
    C.I.g5  = 6.11-15.0 years in 5.0 yg/1 channels), and
    an odonate nymph the least (T% ^ 15 days).  One year
    after  the Hg inputs were stopped, all system components
    analyzed still had elevated levels relative to the
    controls.

-------
 8.   During the project period the channel communities were
     continuously changing, and no steady-state systems
     developed.  There was continual downstream expansion
     of the macrophyte populations, a constant rearrangement
     of insect and algal populations, and a continuous
     influx of new plant and animal species into the channels,

 9.   Periphyton colonization was retarded in the Hg-treated
     channels with respect to controls although greater
     biomasses eventually developed in these channels along
     with decreased species diversities.

10.   Periphytic algae were affected by Hg treatment in a
     variety of ways including:

     (a) inhibition during all parts of the life history.

     (b) inhibition during the reproductive and stimulation
         during the vegetative part of the life history.

     (c) morphological alteration without a significant
         effect on growth, and

     (d) complete restriction of one species to treated
         channels with the greatest densities occurring
         at the highest Hg concentration.

     Short-term studies using continuous-flow chambers
     confirmed the growth retardation observed for some
     of the stream algae under more controlled conditions.

11.   The most abundant macrophyte, Juncus diffusissimus,
     showed a negative response to Hg treatment.

12.   Insect densities showed no consistent responses to
     the Hg concentrations tested; however, the diversity
     and evenness of the benthic insect communities were
     significantly lowered at the highest treatment level.

-------
                        SECTION III

                      RECOMMENDATIONS
1.  The present drinking water standards regarding Hg
    should be carefully reviewed since currently permitted
    levels can cause ecosystem changes, and also result in
    fish with concentrations too high for human consumption.

2.  Additional long-term experimental studies with
    replicable complex microcosms should be done with both
    Hg and other environmental contaminants.  Treatment
    levels should be realistically low, and a variety of
    environmental conditions (geographical areas) utilized.
    Interactions between contaminants should be considered.

3.  Modifications of the artificial stream facility used
    in this project should be made so as to decrease the
    variability between channels treated alike.

-------
                        SECTION IV

          EXPERIMENTAL DESIGN AND SITE DESCRIPTION
EXPERIMENTAL DESIGN

The basic experimental design was to input water of constant
quality at  a constant rate  into six artificial stream
systems.  After biological  communities became established
in the channels,  fish were  to be added, followed shortly
by the continuous introduction of Hg++ to establish water
levels of 0.01 and 1.0 yg Hg/1.  The Hg inputs were to be
continued for one full year, during which time samples of
water, lining material, biota and sediments were to be
taken and analyzed for total Hg content.  During this same
period, the structures of select portions of the channel
communities were  to be determined by detailed species
counts of periphytic algae  and benthic insects.  Estimates
of community productivity and respiration were to be made
from the continuous monitoring of upstream-downstream
changes in  dissolved oxygen and pH.  All sampling programs
were to be  continued for one additional year after the Hg
inputs were terminated.

The experimental  program deviated from the original plans
as work progressed.  Early  data indicated certain sampling
programs were not producing useable information and these
were discontinued.  Changes in sampling techniques were
made necessary by changes in the biological communities.
Analytical  procedures were  simplified to allow for more
efficient handling of samples.  And the original Hg dosing
schedule was changed when it became evident that the
0.01 yg/1 water concentrations were not providing
contamination levels sufficiently high to permit the
study of the clearing of Hg from the channels receiving
this input.

ARTIFICIAL  STREAM FACILITY

The artificial stream complex is an Environmental Protection
Agency (EPA) facility located on the Energy Research and
Development Administration's (ERDA) Savannah River Plant
(SRP), a 507-km2  reserve that includes portions of Aiken,
Barnwell and Allendale Counties in South Carolina.  The
six concrete Block channels (Figure 1) are 91.5 m long,
0.61 m wide and 0.31 m deep with concrete block pools
(1.5 m X 3.0 m X  0.92 m) at both ends of each stream.
The channels were  constructed on a level concrete slab

-------
FIGURE 1,   ARTIFICIAL STREAMS,  IN THIS PHOTOGRAPH TAKEN
FROM THE HEAD POOLS, CHANNEL 1 IS ON THE RIGHT AND CHANNEL 6
IS ON THE LEFT,   THE PLASTIC STRIPS AND FISH-HOLDING CAGES
CAN BE SEEN,

-------
which was once a building foundation.  For this study, the
pools and channels were lined with a 0.05-cm-thick black
polyvinyl chloride  (PVC) film.

CHANNEL WATER AND SUBSTRATE

Water for the channels is pumped from a well located near
the facility.  The chemical composition of this water/
based on analyses made over a 10-year period by personnel
of the E. I. duPont Corporation, is presented in Table 1.
During the period of this study, the average pH of the
well water was 4.6 rather than the .5.3 shown in Table 1.
The water was treated at the stream site by passage
through limestone-filled tubes  (Figure 2) in order to
decrease the free C02 content and increase the hardness
and pH.  The limestone, obtained at the beginning of the
project from a quarry in Rome, Georgia, was changed
approximately every 30 days.  An analysis of the
limestone as provided by the supplier is given in Table 2.
Figure 3 shows the average hardness of the water input to
the pools at the head of the channels as a function of time.
In addition to this time variation in water quality between
limestone changes, there were measurable differences
between channels during the latter part of the project.
This was due to leaks that developed in the head pools,
which made it necessary to input water at different rates
into these pools in order to maintain desired flows into
the channels.

Periodic measurements of dissolved organic matter, nitrate
nitrogen, orthophosphate, conductivity, pH, temperature
and dissolved oxygen were made on samples of the influent
water.  The dissolved organic carbon levels were always less
than 0.5 mg/1, the nitrate-N levels averaged 38 ug/1
 (Ferens, 1974), and the orthophosphate levels averaged 76
Vig/1  (Ferens, 1974) .  The conductivity decreased from
approximately 30 to 22 ymho/cm and the pH decreased from
6 to 5 during the period between limestone changes.  The
average temperature of the incoming water varied seasonally
but only slightly about the mean of 23°C.  Water entering
the channels was generally saturated with respect to
dissolved oxygen.

The rate of water entering each channel was controlled by
a manual valve located just before the limestone-filled
treatment tubes.  Flows were checked daily against
gradations on a V-notch weir separating the pool at the
head of each channel from the channel itself.

-------
FIGURE 2,    LIMESTONE-FILLED TREATMENT SYSTEM,

-------
            5        10       15       20       25
             TIME AFTER  LIMESTONE CHANGE, days
30
FIGURE 3, AVERAGE CHANGE IN CHANNEL WATER HARDNESS BETWEEN LIMESTONE CHANGES,

-------
            Table 1.   WELL WATER QUALITY 1963-1973
                    (After Ferens,  1974)
    Analysis
N
X
                              cv
Temp
PH
Cond
TDS
DO
Hardness (EDTA)
Mg++
Ca++
K+
Na+
Fe+++
Cu++
HC03-
ci-
504 =
SiO2
4
13
11
8
5
12
8
7
7
10
10
8
8
8
8
9
22.8°C
5.3
19.7 umho/cm
13.1 mg/1
7.1
6.56 (CaCo3)
1.87
0.63
0.47
2.48
0.04
0.09
0.39
2.59
1.90
7.82
15%
7
18
42
18
67
88
108
17
115
57
71
149
79
71
70
In September of  1971, washed builders sand was distributed
in the channels  to a uniform depth of 0.05 m and the water
input started.   Flows of 95 1/min into each channel, as
measured by V-notch weirs between the head pool and channels,
were maintained  until February, 1975, except for several
short intervals  of reduced flow due to water pump failures.
Stainless steel  end plates between the channels and tail
pools were adjusted to maintain a water depth of 20 cm over
the sand.  With  these settings, mean retention time of the
water in the channels was 2 hr and the average water
velocity was 0.013 m/sec.

Water leaving  the treated channels was passed through beds
of shredded rubber,  (Figure 4)  a waste product of
commercial tire  recapping operations.  Results of water
analyses made  on a number of different occasions show
that at least  50% of the Hg leaving the channels was taken
up by the rubber particles.  Laboratory uptake studies
have shown that  a much higher  removal of Hg is possible  if
contact between  the water and  rubber is complete  (Kania  and
Beyers, 1974).   The use of shredded rubber for Hg removal
                             12

-------
H
•••
         FIGURE  L\,    MERCURY RECOVERY SYSTEM,   EFFLUENT WATER ENTERED POOLS FILLED WITH
         SHREDDED  RUBBER  OBTAINED FROM A TIRE-RECAPPING PLANT,

-------
               Table 2.   LIMESTONE COMPOSITION
               (Analysis provided by the quarry)

CaO
MgO
Si02
Fe2°3
A1203
Na2O
s
52.73%
1.70
3.42
0.25
0.33
0.10
Trace
was suggested by Edwin D. Russell of the E. I. duPont
Corporation.

CHANNEL BIOTA

The biological  communities  that became established in the
channels developed  from organisms carried  in by the wind
or by animals visiting the  stream facility.  A variety of
microorganisms  was  probably added with mosquitofish,
Gambusia affinis, that were introduced at  two different
times in the channels.

Mosquitofish were seined  from a nearby pond and were
acclimated for  two  weeks  in artificial ponds receiving
water similar in quality  to that input to  the channels.
In April, 1972, 400 fish  were introduced into each channel
and 40 into each of two 51-cm X 30.5-cm stainless-steel
screen cages suspended 10 m from both ends of each channel.
Fish were confined  to the channels by coarse stainless-
steel screens.  Caged fish  were fed three  times a week
with a commercial tropical  fish food  (Bi-0-Rell).  In
May, 1973, all  fish remaining in the channels and cages
were removed for analyses and new animals  introduced,
again 400 into  each channel,  40 into each  cage.

MERCURY INPUT

Mercury was pumped  continuously into the streams with a
four-channel, peristaltic  tubing pump  (Sage Instruments
Model 375).  Solutions for  each channel were made daily
with the concentrations adjusted to compensate for
                              14

-------
changes in pump characteristics.  The concentrated
solutions from which the daily solutions were prepared
were made by dissolving dried reagent grade HgCl2 in
2% (vol/vol) HNC>3 to provide Hg++ levels of 0.1 mg/1,
10 mg/1, and 20 mg/1.  Measured amounts of these stocks
were diluted with well water to 2 1 in a volumetric flask
and transferred to polyethylene bottles connected to the
input tubes of the pump.  Based on the volume removed,
and the elapsed time between changes, a new concentration
for the input mixture was calculated and new solutions
prepared.

Mercury inputs to four of the six channels were begun on
May 15, 1972, so as to establish water concentrations of
Hg++ at 0.01 yg/1 above ambient in two channels  (3 and 6
as counted  left to right looking upstream), and 1.0 yg/1
in another  two channels  (2  and 5).  Channels 1 and 4 were
used as controls.  Since the six channels are structured
as three pairs  (Figure 1),  the dosing arrangement was
chosen so that there was a  northern and a southern exposed
surface for each treatment.

On August 1, 1973, the input levels to channels 3 and 6
were increased to 5.0 yg/1.  This was done because it
was obvious that the 0.01 yg/1 dosing levels were not
providing high enough Hg concentrations in the channel
systems to  permit the study of the removal of this metal
after inputs were stopped.  All Hg inputs were stopped
on January  29, 1974, although routine sampling continued
until February 10, 1975, when the study was terminated.
                              15

-------
                         SECTION V

                          METHODS
 SAMPLING  PROCEDURES

 Water

 Prior  to  August  1, 1973, only occasional water samples
 were taken  for Hg analyses, and then only from the channels
 receiving an  input of 1.0 yg/1.  Mercury in the water
 from the  channels receiving inputs of 0.01 yg Hg/1 was
 below  detection  limits of the techniques used.  After
 August 1, 1973,  the date of the dosage change from 0.01
 yg/1 to 5.0 yg/1, samples were taken routinely from two
 locations in  each channel (including controls) and also
 from two  regions of the effluent channel leaving the
 stream facility.  Sampling stations were located
 approximately 6  m from the Hg input tubes at the heads
 of  the channels  and 1 m from the downstream ends of the
 channels.   Samples of the effluent water leaving the tail
 pools  were  taken just below where the individual outflows
 joined into a common stream and also 100 m farther
 downstream.

 The samples were taken by submersing a nitric-acid-rinsed
 250-ml Erlenmyer flask.  The contents of the flask were
 checked for visible particulates and a new sample taken
 if  particles  were present.  A clear sample was transferred
 to  a nitric-acid-rinsed polyethylene bottle containing
 2.5 ml of concentrated HNOj.

 Air

 In  the initial stages of the project, no attempt was made
 to  measure Hg losses from the channels at the air-water
 interface.  Based on reports by Fagerstrom and Jernelov
 (1972)   and Jernelov and Lann (1973), no losses were
 expected  from the acidic waters  (pH 5-6) of the channels
 because monomethyl Hg was supposed to be formed rather
 than the more volatile dimethylmercury.  A slight increase
 in Hg  levels  of  some components of the control channels,
however,  indicated a transfer might be occurring from the
 treated systems.  The possibility of leaks was remote since
each channel  was covered with a single continuous sheet of
PVC film  and  there was no pressure differential across
the wall  separating adjacent channels.  Therefore, the
possibility of air transfers was investigated.


                             16

-------
 Six  air  sampling systems, as shown in Figure  5, were
 constructed and the glass boxes,  which covered  0-187 m2,
 suspended at the midpoint of the  channels  by stainless-
 steel  straps.   Air was drawn from under the  boxes at the
 rate of  0.47 1/s,  dried with Mg(C104)2 and then passed
 through  gas-scrubbing bottles containing 100 ml of  a 10%
 (vol/vol)  H2S04, 1% (wt/vol)  KMn04 solution.  On three
 occasions,  two scrubbing bottles  were connected in  series.
 Analytical results showed that 98% of the  Hg was collected
 in the first scrubber.

 To define the form of the Hg at the air-water interface,
 traps  consisting of silver-plated copper turnings were
 placed between the drying tube and the gas scrubber
 containing the acid permanganate.   On three  occasions,
 adjacent boxes were placed in channel 6 (5.0 yg/1), one
 with and one without the Ag-Cu trap and the  air sampled
 for  a  two- or three-hour period.

 During routine sampling, air samples were  taken in  the
 early  afternoon on clear days.

 Sediment

 Several  different techniques were used to  obtain sediment
 samples.   Early in the project, a coring  device, which
 removed  frozen cores 14 mm in diameter,  was  used to
 determine the  vertical distribution of Hg  in  the sediments.
 The  frozen cores were sectioned into three zones and
 analyzed.   This method of sampling was discontinued when
 it became obvious  that the Hg in  the sediments  was  entirely
 in the organic component present  as a thin layer on top
 of the sand.

 As described in a  following section,  the processing of
 pyrex  dishes that  had been placed in the channels for
 invertebrate sampling resulted  in a separation  of the
 organic  portion of the sediments  from the  sand.  Both of
 these  components were analyzed  for Hg and  the organic
 content  of  the organic component  was measured.

 During the  last year of the study,  samples were taken with
 a plexiglass tube  5.5 cm in diameter and 40  cm  long.  The
 tube was  pushed vertically through the sediments until it
 touched  the channel  bottom.   The  material  was retained in
 the tube  by hand as  it was transferred to  a  container for
 transportation.  In  the laboratory,  the macrophytes were
 removed  and saved.   The core  material,  free  of  plants,
was blended for several seconds and the suspended sediments
                              17

-------
  AIR
INTAKE
    TO -^
VACUUM PUMP
GLASS  BOX
       TRAP
     FLOWMETER
                                     DRYING
                                      TUBE
   GAS
SCRUBBER
                                                Mg(CI04):
                                             H.2S04
                                               +
                                             KMn04
                 FIGURE 5.  AIR SAMPLING SYSTEM.

-------
and algae poured off from the heavier sand.  The sand
was rinsed once with tap water and was combined with the
first suspension removed.  The total volume of the resulting
suspension was measured, blended for one minute, and
subsampled for Hg analyses, organic matter determinations,
and algal counts.

Export

The original plans were such that the organic particulate
material leaving the channels would be collected
quantitatively by a series of screens and nets of
decreasing mesh and analyzed for both Hg and organic
content.  It became obvious very soon after sampling
started that this method would not work because of the
almost immediate plugging of the screens by very large
amounts of a slime associated with the material moving
out of the channels, and another method for collecting
export was developed.

In the early stages of the project, virtually all of the
material moving downstream was floating.  Therefore, two
stainless-steel screen skimmers were suspended in each
channel, one at the effluent end and one midway between
ends.  The mats that formed behind the skimmers were
harvested quantitatively with a slightly submerged plastic
dishpan and transferred to a 4-1 graduated cylinder for
measurement of total volume.  The contents of the graduated
cylinder were thoroughly mixed, a 200-ml volume removed
and blended for one minute, and two 10-ml subsamples
taken, one for organic content determination, one for Hg
analysis.

Periphyton

Channel Walls-
Shortly after the water input to channels was started in
September, 1971, 40 plastic strips (10 X 20 cm) of the same
PVC material lining the channels, were suspended at upstream
and downstream locations in each channel.  These strips
were removed periodically and the periphyton was scraped
off into a small amount of water.  Biomass determinations
and Hg analyses were made on samples of a blended suspension
of this material. Sampling of these plastic strips was
discontinued in October, 1973, because the large growth of
an emergent aquatic plant, Juncus diffusissimus, in the
channels interferred with the hanging strips to the extent
that they could no longer be considered representative of
the channel walls.  A method was then devised so that a
known area of channel wall could be scraped and the periphyton
                              19

-------
analyzed as above.  A plexiglass box with the top  and one
side removed was placed tightly against the wall with the
top above the water line.  The enclosed section of wall,
an area 47-mm wide and 125-mm deep, was scraped clean with
the flat end of a plastic stirring rod and the slurry was
transferred to a beaker by means of a hydrometer  suction
tube.  All samples were taken from the north-facing (shaded)
wall.  These samples were treated the same as the plastic
strip samples except that a 5-ml aliquot was saved for
algal enumeration.

Plexiglass Plates-
Plexiglass plates measuring 10 X 10 X 0.63 cm were placed
in the channels in December, 1971, to provide substrates
for periphytic growth.  Four trays carrying a total of 32
vertical plates  (the sides were parallel  to  the direction
of stream flow) were placed at the head and  tail of each of
the six channels.  The same number of horizontal plates
were placed on the sand  substrate  interspersed with the
racks holding the vertical slides.  During the first two
years, two plates of each orientation were removed monthly
for biomass determination and algal enumeration.   In the
following year, one vertical plate was removed monthly from
each position in  each stream for Hg analyses  in addition to
the algal counts  and biomass determinations.   The  horizontal
plates were abandoned after April,  1973,  because  of  the
high variability  between them  (Ferens, 1974).  Vacated
positions were refilled with new plates  that were  allowed
to colonize for at least six months before being  removed.
During the last year of  the study,  all of the plates  used
had been in the  streams  for at  least one  year.

Periphyt ic growth on the plates was carefully scraped with
a plastic stirring rod and washed  with a  small amount of
distilled water  into a beaker.  This sample  was homogenized
for one minute in a Waring blender,  the  total volume was
determined, and measured subsamples were  removed  after
additional mixing for Hg analysis,  biomass determination,
and algal counts.

Between July, 1972, and  April,  1973, two  clean plates were
placed each month at both positions  (head and tail)  in  all
channels and harvested after one week  of  exposure as a
study of short-term productivity.   These  plates were
sampled as described above except  that the  entire
suspension was filtered  for biomass  determinations.

In June of 1974,  five months after the mercury inputs had
been stopped, 10  clean plates were  placed in channel 6
                              20

-------
(formerly 5 ug/1).  Two plates were harvested biweekly over
a three month period and handled as described above.

Glass Slides-
Glass microscope slides (25 X 75 mm) were used to directly
study the initial colonization of periphytic algae in the
artificial streams.  Slides were hand-washed with a mild
cleanser, thoroughly rinsed with deionized water, and
placed at upstream and downstream locations in each
channel on five occasions, two before the Hg inputs were
stopped  (October 31, 1973 and December 7, 1973) and three
after the inputs were stopped (January 30, 1974, April 25,
1974 and October 23, 1974).  Eight to ten slides were
suspended 7.6 cm above the sand substrate at each location
and oriented with their flat sides facing the water flow.
One slide was removed from each location in each channel
at three-to-seven-day intervals for at least one month,
the downstream side was wiped clean of periphyton and the
upstream side was counted as described below.  On two
occasions (after 40 and 80 days of colonization), sets of
glass slides were scraped and rinsed with distilled water
and the biomass and Hg concentrations were determined.

Macrophytes

The initial project plans had made no provision for sampling
macrophytes.  By spring of 1973, however, it became evident
that a macrophyte, Juncus diffusissimus, was to assume
increasing importance in the channel communities.  Routine
sampling of this Juncus for Hg analyses was initiated in
June, 1973.  Individual plants were removed by hand from
the channels, the roots were washed free of sand, the
leaves were washed free of periphyton and these two portions
separated.  During the last year of the project,
quantitative estimates of the macrophyte standing crop
biomasses were made from plants removed with the large
core sediment samples previously described.  These plants
were also used for Hg analyses.

Aquatic Insects

In November, 1971, 16 rectangular pyrex baking dishes
(15.7 X 26 X 4 cm) were filled with sand and placed in
upstream and downstream locations in each channel.  One
dish from each position was removed monthly, the organisms
being retained by either a rectangular stainless-steel
screen box fitted inside the walls of the dish or by a
flat piece of screen held against the surface as the dish
was removed.  The contents of the dishes were transferred
                              21

-------
 to stainless-steel buckets,  the  dishes washed,  filled with
 sand and replaced in their former  positions.  Tap water
 was added to the bucket,  the contents stirred by hand and
 the suspended material poured into white  enamel pans from
 which the living organisms were  removed with forceps.
 The procedure was repeated until nothing  was found  in two
 consecutive rinsings.   Insects removed were preserved in
 70% or 80% ethanol until  they could be identified and
 counted.  All dishes sampled had been in  the channels
 approximately one year before removal.

 Aquatic insects were qualitatively sampled by sorting
 through the detritus collected on  the fish-retaining
 screens at the tails of the streams and also by collecting
 individuals emerging from the streams onto the  plants or
 into the air.  Adult odonates laying eggs in the streams
 were netted and preserved throughout the  warmer months to
 aid in the identification of nymphs.

 Fish

 From April, 1972, until January, 1973, the fish were
 collected from the channels with small nets.  This  method
 was inefficient because of the care required to prevent
 disturbance of the wall and bottom communities. A  small
 portable shocker was constructed and utilized for the
 remainder of the project.  Fish  were removed from the
 screen cages with a small dip net.

 ANALYTICAL PROCEDURES

 General

 All analyses were carried out on a Coleman MAS-50 Hg
 Analyzer System modified  by the  addition  of  a digital
 read-out and fused quartz windows  on the  absorption cell.
 Standards were prepared prior to each  set of  analyses  from
 a 1.0  yg/ml solution made weekly from a  commercially
 available 1000 yg/ml standard (Harleco Corp., Philadelphia,
 Pa.).   Standard curves derived by  the  "standard additions"
 method  with fish and periphyton  digestates did  not  differ
 significantly from those  derived by adding measured amounts
 of  Hg  to  100 mis of a  deionized  H90 solution containing
 5 ml concentrated 112804,  1 ml concentration  HN03, and  1  ml
 of  a 6%  (wt/vol)  KMn04 solution.  Therefore, the  simple
 acid-permanganate solution was used for the preparation
 of  standards.

Regardless  of the material sampled,  an acid-permanganate
solution  resulted from the digestion procedure.  This was
                              22

-------
reduced with a 1.5%  (wt/vol) hydroxylamine hydrochloride
solution, further reduced with  a  10%  (wt/vol) SnCl2 in 20%
(vol/vol) HCL solution,  and  the resulting free Hg sparged
from solution by a recirculating  current of air which
passed through a Mg(ClO4)2 drying tube  into the UV absorption
cell of the Hg detector  (Coleman,  1970).  The maximum
absorbance reached was recorded.   Analytical procedures
were checked by analyses made of  Hg standards in gelatin
(Eastman Kodak Co.,  Rochester,  N.Y.)  and by a series of
recovery experiments with samples spiked both with inorganic
(Harleco) and methylmercury  (Alfa Inorganics, Beverly, Mass.)
standards.  In all cases recoveries ranged from 95-103%.

The few methylmercury analyses  that are reported were
performed by personnel of the Environmental Protection
Agency's Southeast Environmenta.l  Research Laboratory under
the direction of Dr. Harvey  W.  Holm using the technique of
Longbottom et al.  (1972).

Water

Analyses were made of 100-ml aliquots of both filtered
(0.3 ym Gelman Type  A glass  fiber filters) and unfiltered
water.  These samples were treated with 5 ml of concentrated
H2S04, 1 ml of concentrated  HNO.,,  and 1 ml of a 6% (wt/vol)
KMnO4 solution and allowed to stand at  least 30 min before
being analyzed.  Results from this technique were compared
against those obtained from  a procedure which included the
addition of K2S2Og and a heat treatment.  These extra steps
were found to be unnecessary for  the channel water samples
although they may be required in  some cases (EPA, 1971).
Filters tested for Hg were found  to contain less than 0.002
yg  of this metal.   Filtration  of 100 ml of 0.1 yg Hg/1
solutions of HgCl2 in 1% (vol/vol) HN03 showed that the
filters removed less than 1% of the dissolved Hg from
solution.

Air

The gas scrubbers containing the  100 ml of H2304-KMn04
solution used for extracting Hg from the air were carried
to the laboratory and the excess  permanganate reduced by
the addition of crystaline hydroxylamine hydrochloride.
The contents of the  scrubbers were then transferred to BOD
bottles and handled  in the usual  manner, starting with the
addition of SnCl2.   In some  cases  10-ml aliquots were used
instead of the entire sample when high  levels were expected,
based on previous data.  These  were diluted to 100 ml with
deionized water.  Samples (100  ml) of the H2S04-KMn04
                              23

-------
solution used in the sample scrubbers were  analyzed so that
corrections could be made for the slight  amount of Hg in
the reagents.

Sediment

The small (14-mm) cores removed in the  early part of  the
project were cut, while frozen, into three parts.  The
well-defined organic layer, which ranged from 2-5 mm in
thickness, was removed and the remaining sand was divided
into two approximately equal portions.   The three fractions
from each core were dried at 60°C overnight, and cooled in
a dessicator.  Weighed subsamples were transferred to
100-ml volumetric flasks and treated with  5 ml of a 6%
(wt/vol) aqueous KMn04 solution.  The flasks were covered
with marbles and autoclaved for  15 min at  121°C  (EPA,
1974).  After the flasks had cooled, approximately 80 ml
of deionized water was added.  The flasks  were allowed to
stand for several hours, and then brought  to volume with
additional deionized water.  The contents  of the flasks
were transferred to BOD bottles  for analysis.  For the
quantification of organic  content, weighed subsamples of
the dried core fractions were  placed in  40-ml Vycor
evaporating dishes and ashed for approximately 18 hr at
400°C.

The sampling procedure for benthic insects effectively
separated the organic portion  of the sediment in the pyrex
dishes  from the sand.  Virtually all of the organic matter
was transferred  into the enamel  pans with  the wash water
used to remove the insects.  This was saved and  the
total volume measured when sampling was  complete.  The
water was then mixed to suspend  settled  matter and a
100-ml portion was removed -which was blended and subsampled
for Hg analysis and the determination of organic content.

Estimates of the organic matter  contained,  in the blended
samples were obtained in one of  two ways.   A 10-ml sample
of the homogenate was either filtered through a  prepared
0.3 nm Gelman Type A glass fiber filter, or was  placed in
a prepared 40-ml Vycor evaporating dish.   Filters were
washed with deionized water, dried at 110°C, fired at
500°C for at least four hours, and dessicated until  they
were removed for weighing prior  to use.  The same treatment
was applied to Vycor dishes except the dishes were scrubbed
with a mild abrasive (Bon-Ami) before being rinsed.

Samples on the filters or  in the Vycor evaporating dishes
were dried to constant weight  at 65°C, cooled and weighed.
                              24

-------
They were then ashed at 400°C overnight, cooled and
reweighed.  The loss between the  last two weighings was
considered to be an estimate of the ash-free dry weight
which was used to convert Hg readings to a relative  (yg
H
-------
results were consistently higher by about 5%.   This  indicates
that the blending process may release soluble  organic  cell
constituents which pass through the filter.

The relationship between wet weight and ash-free dry weight
was determined on three occasions.  Composite samples  of
export material were collected, blended, centrifuged and
weighed amounts of the resulting mass placed in Vycor dishes
for drying and ashing.  Data show that the weight loss on
ignition was  (5.9%) of the wet weight which agrees well with
the value of  6.4% calculated from data presented by Sladecek
and Sladeckova  (1963)  for periphyton colonizing vertical
glass slides.

Periphyton

All periphyton  sampling  procedures  resulted in aqueous
suspensions of  the material  scraped from the  various
substrates.   These suspensions  were blended for 1 min and
treated  in  the  same manner  as  described for the organic
portion  of  the  sediments except that an additional
subsample was taken  for  algal  counts.   The  volume taken
for digestion was  a  function of the density of the  suspended
matter in the sample.

Macrophytes

The washed  leaves  and roots  of the Juncus plants removed
from  the channels were allowed to dry at room temperature
for 24-48  hr.  The roots and stems were cut into small
pieces and weighed portions  placed in 100-ml  volumetric
flasks and  treated with  5 ml of concentrated H^SO^ and 1 ml of
concentrated  HNO3.  The  stoppered flasks were manipulated
so that  all the plant material  was  in contact with  the
acid, and then  allowed to stand for 18-20 hr.  Five ml of
6% KMnC>4 was  then added  and  the flasks stoppered after the
initial  frothing had  stopped.   The  samples  were allowed  to
digest at room  temperature  for  an additional
After this  period, the volumes  were adjusted  to 100 ml
and the  contents of the  flask  transferred to  BOD bottles
for analysis.   This procedure was sufficient  to dissolve
and oxidize the plant tissues  and no residues were  visible
when  the permanganate coloration was destroyed with
hydroxylamine hydrochloride  during  the analytical process.

Weighed  and dried samples of roots  and stems  were oven
dried at 60°C for 18-20  hr,  cooled in a dessicator, weighed,
ashed at 400°C,  cooled and weighed  again.  The weight lost
during the  ashing procedure was used to calculate the
                              26

-------
percent organic matter in the original air dried sample.
The figure was then used to calculate the weight of the
organic component of the plant samples analyzed for Hg.

Insects

Insects collected from the screens at the ends of the
channels were blotted dry, weighed, and placed in 100-ml
volumetric flasks.  Five ml of concentrated H2S04 and 1 ml
of concentrated HNOs were added to the flasks which were
allowed to stand, with occasional shaking, until a clear
solution resulted.  Five ml of a 6% (wt/vol) KMnC>4 solution
were then added to each flask.  The flasks were stoppered
after initial frothing had ceased and allowed to stand for at
least 16 hr at room temperature.  They were then filled to
100 ml with deionized water and transferred to BOD bottles
for analysis.

Fish

Mosquitofish were killed by submersion in hot water, blotted
dry, weighed and placed in 100-ml volumetric flasks.  They
were then processed in the same manner as described for
insects.  When the fish were large (> 0.5 g), twice the
usual amount of acids and KMn04 were added.  Reagent blanks
were prepared in the same manner.

ALGAL COUNTS

Representative subsamples  (approximately 5 ml)- of the
homogenates resulting from the processing of sediment cores,
channel walls and plexiglass plates were placed in 15 X 125-
mm tubes and refrigerated in the dark until counts could be
made.  Usually samples were counted the same day they were
collected, however, a few were stored for as long as two days,

For counting, the samples were brought to room temperature,
resuspended by shaking the test tube at high speed with a
vortex type mixer  (Cole-Palmer Super-Mixer) and a 50-ul
portion removed immediately with an Oxford micropipettor
fitted with a disposable tip.  The portion was placed as a
drop on a clean microscope slide and spread evenly under
a 22 X 22-mm, #1 glass coverslip.  The algal cells observed
in measured areas of the slide were identified and counted
at 400 X with a Nikon Model SKe phase-contrast microscope.
An attempt was made to identify and count a total of 200
cells from each slide although up to 1000 cells were
counted on very dense slides.  On a few occasions, because
of the low numbers of organisms in the samples, counting
                              27

-------
was terminated at 100 cells.  During the last part of  the
study, a replicate count from another measured drop was
made for each sample.

All cells in colonies and filaments were counted individually.
Vegetative cells were counted if chloroplasts were present,
and tiny, highly motile flagellates were not counted unless
they were observed in resting stages.

Qualitative observations of blended  and unblended periphytic
material indicated only one common algal species was altered
beyond recognition in the blending process  (Eremosphaeria
viridis) while  filamentous  forms were commonly broken up
into single cells or small  groups of cells  permitting more
accurate enumeration.

Algal counts  for  individual species  were converted  to
relative values (cells/mm2  of  substrate) for treatment
comparisons.  To  compare  total populations, the volume of
each species  was  estimated  from average measurements of  the
cells  (APHA,  1965) and  the  total cell volumes (in  cubic
standard units, 1 CSU = 8000 mm3)  per unit  area  (mm2)  of
substrate compared between  treatments. The total  organic
content of periphyton attached to each substrate  was
determined from dryed and ashed subsamples  of the same
suspension on which  the algal  counts were made.   Diversity
and evenness  values  were  calculated from  equations given
by Pielou  (1969)  for subsamples from large  populations,

                 S
            H=- E  (n./N) log, (nt/N) ,  and
                i=l   1        ^

            E=H/log2 S

where S is the  number of  species, N is the  total  number  of
cells counted,  n^ is the  number of  cells  in the  it'1 species,
H is the Shannon  Weaver diversity  and E is  the evenness.

Algae colonizing  the glass  microscope slides suspended in
the channels  were counted directly.   After  a slide was
removed, the  downstream surface was  wiped clean  and a
coverslip was placed on that end of  the upstream  surface
that had been deepest in  the water.   Again, a known area
was examined  and  the algal  species  identified and counted.

STATISTICAL TREATMENT

The amount of replication that  was  possible in sampling  the
channel components was  limited  by the physical size of the
                             28

-------
channels, the long duration of the study, and the amount of
personnel time available.  At some time during the project
period, however, at least one estimate was made of the
variability associated with each method and these
variability data have been presented in the appropriate
section.

In analyzing the large quantity of routine data collected,
the positional effects resulting from the difference between
upstream and downstream stations and the differences between
channels treated alike have been largely ignored.  Separation
of the effects of the most important factors influencing
the results, Hg level and time, were made by a two-way
analysis of variance  (ANOVA, fixed effects model) by means
of subroutines included in the Statistical Analysis System
(SAS)  (Service, 1972).  Correlations, regressions, and
basic calculations were done either with SAS subroutines
or with BASIC programs specifically written for the Nova
1200 computer associated with the stream facility.

RELATED STUDIES

General

During the project period, several short-term studies
relevant to the stream research were carried out.  Some
of these were initiated to confirm channel results, others
were done to provide information not available from the
channels.  The rationale  for these short-term studies and
the methods used will be presented in the following
paragraphs.  The results and conclusions will be discussed
in Section VII.
Bacterial Counts

Since the sensitivity of microorganisms to Hg salts is well
known, the initial project plans included daily sampling of
water for total bacterial counts.  From the beginning of
June, 1972, until the middle of October, 1972, samples of
water were taken between noon and 2 PM daily in sterile
dilution bottles from head and tail positions in each
channel.  The suspended bacteria were counted by serial
dilution pour plates made with tryptone glucose extract
agar.  Because no differences were found between treatments ,
sampling was reduced to a weekly schedule in November,
1972, and discontinued completely at the end of the year.
The Hg levels in the channels were below those found to have
an effect on heterotrophic bacteria by Albright et al.  (1972)
                              29

-------
Productivity Estimates

The original project also included plans for channel
productivity determinations based on upstream-downstream
measurements of dissolved oxygen (Odum, 1956)  and pH
(Beyers, 1970).  For this purpose, Schneider Instrument
Co. RM-15 water quality monitors were borrowed from the
EPA's Southeast Environmental Research Laboratory (SERL)
Athens, Georgia, and installed in small buildings located
at both ends of each channel and a system constructed which
switched the water input to the monitors from one channel
to another every 10 min.  Dissolved oxygen, conductivity,
pH, and temperature were recorded by both monitors and the
monitor near the head of the channels also recorded solar
input.  Plans to interface the monitors with a digital
data-processing system based on a Nova 1200 computer were
abandoned when it became apparent from the analog outputs
that there were no measurable differences, except with
respect to temperature, between the ends of the channels.
The water was generally saturated with oxygen at all times.
The maintenance and calibration of the monitors was
extremely time consuming and the fulltime monitoring of
water quality was discontinued.

During the period the monitors were in use, the standing
crop biomasses in the channels were relatively small.  To
see if the lack of. usable results was simply a result of
the lack of biota, upstream-downstream measurements of
oxygen and pH were made on several occasions shortly before
the termination of the project.  There were well-defined
differences between stations and it appears that
productivity measurements could be made in the channels
after sufficient growth has occurred.

Diurnal Uptake Patterns

There was, and still is, some question as to whether or
not the removal of Hg from water is mediated by the
physiological activities of organisms in aquatic systems.
In June and August of 1972, 25 mci of Hg-197 was input over
a  24-hr period into  the channels maintained at 1.0  yg Hg/1
(2 and 5).  The low-level channels set at 0.01 yg/1 could
not be used because the specific activity of commercially
available Hg-197 was too low.  During the  24-hr input
period, and for several hours afterwards, 5-ml water
samples  (unfiltered)  were taken hourly from three locations
in both channels, from the effluent stream below the
shredded rubber Hg removal systems, and from a small pond
connecting with the effluent channels.
                              30

-------
All samples were counted in a deep-well,  5-cm Nal  (11)
scintillation crystal coupled to a 400-channel pulse height
analyzer.  Counting times were adjusted to maintain a
relative counting error of approximately  10%.  The purpose
of taking the hourly water samples was to look for diurnal
patterns of Hg uptake by the channel communities as
determined by differences in activity between upstream
and downstream samples of the same water mass.

An intensive 26-hr  survey was also done in January, 1974,
when all six channels were sampled for water at upstream
and downstream stations, and for Hg losses at the air-water
interface at the mid-point of the channels every 2 hr.  Both
filtered and unfiltered water subsamples were analyzed for
Hg by the methods described previously.

Continuous-flow Chambers

To determine if the responses to Hg observed in the channels
could be verified under conditions of reduced interspecific
interactions, continuous-flow chambers (CFC) were designed
and built to test Hg uptake and growth responses of specific
channel organisms in a more-controlled fashion.  The system
is illustrated in Figure 6.  Water from the well feeding
the artificial streams was passed through a limestone
treatment similar to the system used for  the channels and
passed through the plexiglass chambers to give a velocity
equal to that used in the channels (0.013 m/sec.)  Mercuric
chloride was pumped into the input line of three of the
plexiglass chambers to provide a concentration of 5 yg
Hg/1   and the other three chambers served as controls.

The CFC system was first used to determine direct uptake
rates of Hg from water by mosquitofish and by a corixid
hemipteran (Hespercorixa sp.).  For these uptake studies,
30 animals were placed in each chamber and 5 were removed
at each sampling time.

Attempts were made to isolate the dominant algal species
of the artificial streams to use in the CFC system.
Suspensions of channel periphyton were sprayed with a
spray-diffuser (Pringsheim, 1946) onto agar plates of Hold's
medium (Bischoff and Bold, 1963), modified by the elimination
of buffers and the addition of vitamins,  or onto agar plates
of "Algal-Grow," a product of Carolina Biological Supply Co. ,
Burlington, N.  C.  Five species were isolated and maintained
on agar slants.  Transfers to aqueous media were made to
develop seeding stocks for the CFC system.
                              31

-------
                                               LIMESTONE
                     CHAMBERS
EFFLUENT
                            GLASS SLIDES

f
n r
n
T_rn r
ill
i rn r
d

                     SIDE VIEW
FIGURE 6.  CONTINUOUS FLOW CHAMBER  SYSTEM

-------
For the study of the effects of Hg on the algal species,
racks holding glass microscope slides perpendicular to the
direction of flow were installed in the chambers.  Mixed or
unialgal cultures were poured into the two mixing tanks.
The colonization of the glass slides was observed, by the
methods previously described for the channel glass slides,
for a period of up to 27 days.

Biological Half-life of Mercury in Mosquitofish

After the Hg inputs to the channels were stopped, 200
Hg-contaminated mosquitofish were collected from the
effluent system of the channels.  These fish were placed
in 200-1 tanks, 40 per tank, into which a constant  flow
of fresh water was piped.  The fish were fed with a
commercial fish food preparation (Bi-0-Rell) low in Hg
(0.20 ug/g dry wt.).  Samples of 10 fish were removed
periodically and analyzed by the methods previously
described.  The study was done to compare the rates of
Hg loss from contaminated fish transferred to clean systems
with those calculated from fish remaining in contaminated
systems (the stream channels).  It was also done to compare
estimates of the biological half-life of Hg in mosquitofish
exposed over a long period with estimates based on fish
exposed for a short period to relatively high levels
(Huckabee and Blaylock, 1973).
                              33

-------
                        SECTION VI

                  RESULTS AND DISCUSSION
FATE OP MERCURY

General

There are several common ways of expressing analytical
results and problems can arise when units are not explicitly
stated  (Olley, 1973).  In the discussions that follow, Hg
concentrations in fish and insects are expressed on a wet-
weight basis.  The concentrations in all other community
components, with the exception of the sand portion of the
sediment, are based on an ash-free dry weight.  The ash
free dry weight is the weight lost during the ashing
process previously described-, and is considered to be an
estimate of the total organic content of the sample.
Results of analysis made on the sand component of the
sediment are based on an oven-dry weight, the usual method
for expressing sediment results.

Analytical results of export, periphyton, and organic
sediment samples, can be converted to a wet-weight basis
by multiplying the ash-free dry weight value by 0.06.
This conversion is necessary for making comparisons between
trophic levels.

Water

During the first 12 months of the project, the average
Hg levels in the channels., as calculated from the daily
flow and Hg input records, were 1.04 and 1.05 yg/1 in
channels 2 and 5 respectively, and 0.01 yg/1 in channels
3 and 6.  Analyses of unfiltered water samples taken
approximately 10 m from the Hg input tubes in channels
2 and 5 support the calculated values for these channels.

A summary of Hg analyses of unfiltered water samples
collected between August 1, 1973 and January 29, 1974,  is
presented in Figure 7.  The data show that, during this
period, the input levels were slightly higher than desired
and also that the Hg concentrations in the water leaving
the channels were lower than those at the head end.  The
average uptake in the 1.0-yg/l channels  (2 and 5) was  17.5%
and in the 5.0-yg/l channels  (3 and 6) was 16.8%.  The
similarity in the relative uptake values at the two  different
doses indicates there is no saturation of adsorption sites

                              34

-------
    6r
    4
cr

I  3
o

I  2
    0
                            H3
    H6
?      }
T3
                                       T6
         H2
               El
                                                  I
                                                  E2
            I.O/m/l           5.0/^g/l
             EFFLUENT
               DITCH
                    SAMPLING LOCATION

  FIGURE 7,   TOTAL MERCURY  LEVELS (X ± 2SE) IN CHANNEL AND EFFLUENT
  WATER,  SAMPLES WERE TAKEN FROM UPSTREAM (H) AND DOWNSTREAM (T)
  POSITIONS IN EACH CHANNEL  AND FROM TWO POSITIONS (EL E2) IN THE
  EFFLUENT DITCH DRAINING THE CHANNEL SYSTEM,
                             35

-------
at the levels of Hg introduced.  The relative uptake values
^™?Ch,_10Wer, than. those  rePorted by Huckabee and Blaylock
(1973) who worked with a 100-m length of natural stream and
found that 75% of Hg added as Hg (1103)2 was removed.  The
natural stream Huckabee and Blaylock (1973) studied was
much more turbulent than the artificial streams, and
probably supported a higher total biomass per unit area.

That turbulence might affect the uptake of Hg is shown by
the data presented in Figure 7 for the two positions in
the effluent stream.  Although the distance between sampling
stations in this stream was about the same as between the
two channel stations, there was a greater relative uptake
of Hg  (54.6%) from the effluent water, even though the
retention time of the water in this section was short.  The
greater uptake might be explained by the high turbulence
in this region resulting in greater probability of contact
between a Hg atom and some portion of the stream wall.  The
biological communities and lining (poured concrete) in the
effluent region, of course, were quite different from those
in the experimental channels and the possibility of
differential uptakes resulting from differences in community
structure and channel lining cannot be disregarded.

The results of the two Hg-197 diurnal uptake studies are
summarized in Figures 8 and 9.  No uptake patterns are
evident, although the variability of the data is high,
probably due to the inclusion of particulates in the
sample.  In the June, 1972, experiment, an average of
12.5% of the radiomercury was removed from the water.
In August, 1972, this value increased to 17.9%, probably as
a result of the increased biomass in the channels.

The results of the 26-hr intensive water sampling done in
January, 1974, are presented in Figure 10.  The data show
that an average of 13.2% of the Hg added was retained in
the 1.0-yg/l channels and  15.6% in the 5.0-yg/l channels.
The bimodal uptake curve in both treatment levels is
difficult to relate to the biological activity in the
channels.  A diurnal curve of the type described by Odum
(1956) for dissolved oxygen levels or by Beyers  (1963,
1965) for carbon dioxide uptake and release in aquatic
systems, with a single maximum and minimum each  24-hr
period, would be expected  if Hg uptake were related to
the diurnal photosynthetic or respiratory patterns of
the channel communities.

All data pertaining to uptake  (Figures 7-10) show that  a
constant proportion of the added Hg was removed  regardless
                              36

-------
                         Q Stream  2
                         m Stream  5
-15
       i  i   i  i  i   i  i   i  i   i  i  i   i  i   i  i  i
i  i   i  i   i
10
am
2
pm
6
pm
10
pm
TIME
2
am
6
am
10
am
     FIGURE  8.  DIURNAL RADIOMERCURY UPTAKE CURVE/ JUNE. 1972,
     PLOTTED ARE THE DIFFERENCES BETWEEN RADIATION LEVELS IN
     UPSTREAM WATER SAMPLES TAKEN AT THE INDICATED TIME AND
     DOWNSTREAM SAMPLES TAKEN  TWO HOURS LATER,
                             37

-------
       20
K)
  =3
  o
  o
 UJ
 o
 z:
 UJ
 cr
 LU
 u_
 u_

 O
 o    -10
D  Stream 2

         am 5
,J — — —
10
am

2
pm

6
pm

10
pm
TIME
2
am

6
am

10
am

         FIGURE 9,   DIURNAL RADIOMERCURY  UPTAKE  CURVE, AUGUST, 1972,
         PLOTTED ARE THE DIFFERENCES  BETWEEN  RADIATION LEVELS  IN
         UPSTREAM WATER SAMPLES TAKEN AT  THE  INDICATED TIME AND
         DOWNSTREAM SAMPLES TAKEN TWO HOURS LATER,
                                   38

-------
            1.0 -
        u
            0.5
OJ
        CJ>
        X
            0.0
                                                        I.O^g/l
                                                        5.0/ig/l
                     2   4   6    8    10   I2M  2   4    6   8   10   I2NOON
                                PM                         AM
                                            TIME
        FIGURE 10,   DIURNAL MERCURY UPTAKE  IN  TREATED CHANNELS ON JANUARY 22-23,  1974,
        PLOTTED ARE THE DIFFERENCES BETWEEN  HG  LEVELS IN UNFILTERED UPSTREAM SAMPLES
        TAKEN AT THE INDICATED TIME AND DOWNSTREAM SAMPLES TAKEN TWO HOURS LATER,

-------
of dose, which indicates there was no saturation of  potential
binding sites at Hg levels up to 5.0 ug/1-   These results
differ from those reported for Chlorella by Matsui and
Gloyna  (1972) who found that the higher the Hg concentration,
the greater the percent uptake by this alga.  The difference
may be due to the complex nature of the channel communities
compared to unispecific Chlorella cultures.

Figures 8-10 show no clear pattern that would couple Hg
uptake with the metabolic activities of the channel
communities.  These results, which indicate passive uptake,
are consistent with the findings of Hannerz  (1968),
Glooschenko  (1969) and Fujita  and Hashizume  (1972) who attribute
inorganic Hg uptake in algae to surface adsorption although
Hannan and Patouillet  (1972) concluded that  in  the freshwater
algae they studied, Hg uptake  could not be  accounted for
by surface adsorption alone.   Burkett  (1975)  also found
indications of processes other than passive  adsorption in
the uptake of methylmeifcury 'by Cladophora,  a green alga.

Data showing the fraction of the Hg in the water  that was
attached to particulate material  in the January,  1974,
intensive sampling are presented  in Figure  11.   In all
treated channels, the proportion of the total Hg  removed
by filtration was greater at the downstream stations than
at the upstream stations.  The percent of  the total Hg
bound to particles was greater in  the  1.0-yg/l  channels
than in the 5.0-yg/l channels.  This  indicates  a saturation
phenomenon with respect to the suspended particulates which
may be related to the number of bacteria in the channel
water.  Figure 12 summarizes bacterial count data based on
results of daily samples taken between June 1,  1972, and
October 10, 1972.  Counts made after October 10,  1972 fall
in the limits presented in Figure  12, even  after the  dose
increase of August 1, 1973.  The constant number of bacteria
at a given point in all channels provide a  fixed number of
attachment sites for Hg.  If the number of  sites  were
limited, there would be a decreased relative uptake at
higher Hg levels.  This is what was observed (Figure  12)
and, although the results are by no means conclusive,  the
relationship between the increased number of bacteria at
the tail end of the channels and the greater proportion
of particulate-bound Hg at these stations strongly  implicates
the suspended bacteria as adsorbing agents.

After the Hg inputs to the channels were terminated  (9 AM,
January 29, 1974), water samples were taken hourly  for  6 hr
at the two sample stations in  each channel,  and on  several
occasions during the following month.  The  water levels
                               40

-------
    c
 .  o
en  V-
X  a
    JD
CD
O
I-
cr
<  b
°-  E
    CD
        50 i-
40
£  ~  30
        20
         0
          0
            a
                    T5
              H2
                                  I
                                 T3
                                                  T6
                H5
                                     H3
                                              H6
                 I.O/ig/l                5.0/ig/l

                      SAMPLING  LOCATION
          FIGURE 11.   PROPORTION OF HG  ASSOCIATED WITH PARTICIPATES
          IN WATER SAMPLES,  PLOTTED ARE MEANS (± 2SE) BASED ON THE
          13 SETS OF  SAMPLES TAKEN DURING THE JANUARY,, 1974,
          INTENSIVE SAMPLING PERIOD,

-------
Q)
JD
ce
LJ
H
O
<
CD
HOOr

1000

 900


 800

 700

 600

 500

 400


 300

 200


 100
         I
               4     36    25         14     36    25
                 UPSTREAM              DOWNSTREAM

                        CHANNEL  NUMBER
        FIGURE 12.  BACTERIAL POPULATIONS IN CHANNEL WATER,  RESULTS
        (X ± 2SE) ARE BASED ON 130 DAILY COUNTS MADE FROM JUNE TO
        OCTOBER OF 1974.
                            42

-------
decreased rapidly and were  approximately  0.2  yg/1   (filtered
samples) in all treated  channels  by January 31,  1974.  By
the end of February, 1974,  all  samples were below  the
0.05 ug/1 detection limit of  the  procedure used, even
though the Hg levels in  the channel periphyton and sediment
remained high.

Air

Quantitative data for Hg losses from  the  channels  at the
air-water interface were not  obtained until December, 1973.
No estimates exist, therefore,  for losses from the two
channels that were receiving  an Hg input  of 0.01 yg/1.

The air-loss data, which are  summarized in Figure  13, show
that the releases were highly variable, although related to
water concentration, and continued after  the  Hg inputs were
stopped.  No losses were measured from the control channels.
There appears to be no logical  explanation as to why the
losses from channel 5 were  consistently low,  or why the
losses decreased rapidly before the Hg inputs were stopped.
The data are insufficient to  show any seasonal patterns.

The form of Hg lost at the  air-water  interface is  mostly
elemental Hg.  The results  from two air sampling boxes
located next to each other  in the same channel, one with
a silver trap in the line,  showed that the filter  removed
approximately 97% of the Hg before it reached the  gas
scrubbing system.  According  to Braman and Johnson (1973) ,
silver does not adsorb dimethyl mercury,  and  will  only
partially retain monomethyl mercury compounds, but will
quantitatively retain elemental Hg.   The  specific  filters
used in this study were  checked by inserting  them  in the
line of the Coleman MAS  50  leading into the absorption
cell, and then processing a 1-yg  Hg standard. No  increase
in absorption was noted  indicating that none  of  the
elemental Hg released in the  analytical process passed
through the filter.  The loss of  elemental Hg from aquatic
systems has been found in other studies  (Fagerstrom and
Jernelov, 1972; Spangler et al., 1973; Parks,  1973; Holm
and Cox, 1974) and some  of  the  possible mechanisms of its
formation have been investigated  (Toribara et al., 1970;
Alberts et al., 1974; Wood, 1974; Bisogni and Lawrence, 1975).

During the air sampling  periods,  it appeared  that  the amount
of Hg collected in the traps  was  a function of the number
of bubbles rising from the  bottom.  Samples of the gas
rising from undisturbed  sediments were collected and
immediately injected into the closed  air  system  of the Hg
                              43

-------
    0.9
o>
     0.7
    0.6
LU
CO
UJ

*
en
X
    0.5
    0.4
    0.3
cr   0.2
    an-
    o.o
                                         a— I.O^g/l

                                         A—5.0/tg/l
        FEB MAR APR MAY JUN  JUL AUG SEP  OCT NOV DEC
                      SAMPLING  DATE 1974
 FIGURE 13,  AVERAGE MERCURY RELEASES AT THE AIR-WATER INTERFACE OF THE
 TREATED CHANNELS.

-------
analyzer.  This resulted in a significant  absorption of  the
UV beam, and, although no attempts were made  to quantify  the
data, it confirmed the presence of elemental Hg  in  the gas
bubbles.  Samples of gas from the control  channels  were
checked and the results were negative.  The  loss of volatile
Hg species with bubbles released  from sediments  has been
suggested as a Hg release mechanism by Bothner and  Carpenter
(1974).

There were no significant correlations between the  air
losses and either the organic contents or  the Hg content
of the sediments under the sampling boxes.   Perhaps a
correlation would have been observed if some quantification
of the living portion of the sediment, especially the
bacterial components, could have been made at the same
time the air samples were taken.  Bacteria are known to
be able to release elemental Hg from both  inorganic and
organic compounds of this metal  (Suzuki et al., 1968;
Matsumura et al., 1972; Furukawa and Tonomura, 1972),
although elemental Hg production can also  occur  in  the
absence of bacteria (Alberts et al., 1974; Bisogni and
Lawrence, 1975).

Results of the intensive  24-hr air  sampling  program of
January 22-23, 1974, are shown in Figures  14 and 15.  The
data again show that the releases were greater from the
high-level channels than from the low-level  channels.
The rates of the releases appear  to be related to the
water concentrations  (5:1) rather than to  the Hg levels  in
the sediment on that date  (1.6:1, Figure 17).  This
implies that the reaction-producing elemental Hg was
occurring at the water-sediment interface  rather than
deeper in the sediments.

The air release data shown in Figures 14 and 15  indicate
that there may be a diurnal release pattern  with a  maximum
occurring in the late afternoon.  This diurnal pattern
would not be reflected in the water concentrations  of Hg
because the maximum air losses observed, if  extrapolated
to the entire channel, were always less than 1%  of  the
total Hg in the water.

Sediments

When the water inputs to the channels were started  in
September, 1971, the bottom substrate consisted  of  a 10-
cm layer of washed builders sand.  As time passed,  a thin
mat community consisting of detritus, fungi, bacteria,
algae and small invertebrates formed on the  sand.   The
                              45

-------
   0.2 r-
LJ
UJ
-J
UJ
cr
cr>
   0.1
   0.0
                                        D-CHANNEL  2
                                        •-CHANNEL 5
           I
 7
PM
I!    I
  7
AM
                                   TIME
  FIGURE 14,   DIURNAL MERCURY RELEASES ON JANUARY 22-23, 1975, AT THE
  AIR-WATER INTERFACE OF CHANNELS RECEIVING AN  INPUT OF l,0yu.g/l.
  THE AREAS SAMPLED WERE 0,186 M2,

-------
    2.0
u
     .5 -
LLJ  ID

LJ
cr
CT>
cr  0.5
    0.0
                                           A—CHANNELS

                                           A—CHANNELS
                                         _L
I
                                                         I
             I
    9    II
              3579    II    I    357
                   PM                           AM
                                   TIME
FIGURE 15,   DIURNAL MERCURY RELEASES ON JANUARY 22-23, 1975, AT THE
AIR-WATER  INTERFACE OF CHANNELS RECEIVING AN  INPUT OF 5,0/lQ/l.
THE AREAS  SAMPLED WERE 0,186 M2,

-------
 thickness of this  mat  was  always greatest near the heads of
 the channels and became  thinner toward the tails.  Gradually
 this mat became less structured as emergent aquatic plants
 and larger' animals, especially amphibian larvae, became
 established in the channels.

 The sediment samples removed  in the  first part of the
 project consisted  of frozen cores which were sectioned to
 give a vertical profile  of Hg distribution.  After several
 sets of samples, it became evident that all of the Hg in
 the cores was associated with the mat community covering
 the sand.  Sand samples  from  the cores, and also from the
 benthic dishes used for  invertebrate sampling, contained
 Hg levels of less  than 0.2 yg/g except when they were
 contaminated with  organic  debris forced in by the coring
 procedure.  At no  time during the project were significant
 amounts of Hg found in the sand portion of the sediment.
 This was not a result  of the  drying  process used because
 portions of mat were also  dried and  analyzed without a
 significant loss of Hg.  Because of  the localization of Hg
 in the organic layer on  top of the sand, Hg values were
 calculated as yg/g ash-free dry weight of organic matter
 for comparisons between  channels.

 The results of analyses  of the organic component of the
 channel sediments  are  presented in Figure 16 for the
 control and 0.01 yg/1  treatments, and in Figure 17 for
 the 1.0 yg/1 and 5.0 yg/1  treatments.  The data show that
 even in the channels receiving Hg inputs of 0.01 yg/1,
 there were elevated Hg levels in the sediments relative
 to the controls.   When the 0.01 yg/1 input was changed to
 5.0 yg/1, there was a  rapid increase in sediment levels
 although the data  are  insufficient to show whether or not
 an equilibrium level was reached.  The sediment Hg levels
 in all treated channels  showed an increase four months
 after the Hg inputs were stopped.  There is no obvious
 explanation for this although it appears to be more than
 chance variability.  There was no marked decrease in any
 other components of the  channel communities at the time
 of  the increase in the sediments.

Figures  16  and 17  clearly  show the slow decrease in sediment
 levels  in the 13-month period following the cessation of
Hg  input.   The half-lives  calculated from data collected
after May,  1974 were 7.85  years  (€.1.95% = 5.62-13.0 years)
for the  1.0  yg/1 treatments and 8.67 years  (0.1.95% =
6.11-15.0 years) for the 5.0  yg/1 treatments, assuming
first order  kinetics.  These  half-lives are considerably
longer  than  the 1.3 years  calculated by Bothner and
                              48

-------
              20 r
vo
            1!
            2  15  -
            IT
            \~
            •z.
            LU
               10
            o
            o
            Ltl
            Q
            U
            cn
                0
                                     Hg INPUT
                                     INCREASED
Hg INPUT
   OFF
                                                               O
              CONTROL
              •0.01
                   j	i
                  SONDJFMAMJJASONDJ FMAMJJASONDJ
                   1972             1973                    1974             1975
                                         SAMPLING  DATE

            FIGURE 16.   AVERAGE MERCURY CONCENTRATIONS IN THE ORGANIC PORTION OF THE SEDIMENTS OF
            THE CONTROL CHANNELS AND CHANNELS RECEIVING AN INPUT OF 0,

-------
en
O
            2000
        CP
^   1500

<
        Ld
        o
        -z.
        o
        o
        Ui
        Q
        UJ
    1000
             500
               0
                     D-I.O/xg/l
                            Hg  INPUT    Hg  INPUT

                            INCREASED     OFF
                    0
              N
D
M
A
J
A
0
N
D
M
A
M
A
0
N
D
                   1972
                              1973                     1974
                                  SAMPLING  DATE
                                                     1975
          FIGURE  17.  AVERAGE MERCURY CONCENTRATIONS IN THE ORGANIC PORTION OF THE SEDIMENTS OF
          CHANNELS RECEIVING INPUTS OF 1.0 AND 5.0^ g / | .

-------
Carpenter  (1974) for losses from a contaminated marine
system but are similar to the values  they calculated from
the data of other workers  (Spangler et  al.,  1973).

Export

Figure 18 summarizes the Hg concentrations  measured in the
particulate matter leaving the channels.  Data for the
control channels are not presented because  of the generally
low values (<4 yg/g ash-free dry weight) found in these
samples.  Although Figure 18 shows the  typical high
variability associated with the Hg analyses, several trends
are evident.  The Hg levels in export from  channels 3 and 6
rose dramatically in the three-week period  following the
treatment changes in these channels from 0.01 to 5.0 yg/1.
Indeed, it appears that an equilibrium  level of approximately
1800 ug/g was reached in this short'period  of time.  This
concentration was a factor of only 2-3  times that measured
in the 1.0 ug/1 channels, even though  the water levels were
greater by a factor of 5.

Mercury concentrations were consistently higher in export
material than in the sediments (Figure  17)  until Hg inputs
were stopped in January, 1974.  Export  levels then decreased
more rapidly than the sediments and were much lower when the
project was terminated.  This indicates that the majority
of the exported particulate was not derived from the
sediments.

Periphyton

The most complete set of data for Hg  levels in the channel
periphyton is for material removed from the PVC lining.
The results for the control channels  and the channels
receiving an input of 0.01 yg/1 are presented in Figure 19.
Figure 19 shows that Hg levels in the periphyton exposed to
a water concentration of only 0.01 ug/1 were higher than in
samples from the control channels.  It also shows that there
was an increase in Hg concentrations  in the control
channels when the doses were changed  in August of 1973,
and a decrease in the control concentrations after the Hg
inputs were stopped.  Mercury was probably  entering the
control channels at the air-water interface since.a series
of wall periphyton samples taken every 20 m along the length
of channel 1 (control)  showed the Hg  levels were about the
same along the entire length.  If there were a leak between
adjacent channels, only areas downstream from the leak would
have been affected.
                             51

-------
Ul
         2500
        1
       ,~ 2000
      o
      h-
LU
O

O
O
          500
         1000
      n:  500
      h-
      tr
      o
      o_
      X
      UJ
       0
         Hg  INPUT     Hg INPUT
         INCREASED     OFF
                                                             D -  I.O/ig/l
                                                             o -  0.01/xg/l
                                                               -  5.0/xg/l
         S  0 N  D J  F
M  A M  J
J ASOND  JFMAMJ  JASONDJ F
                 1972
                              1973
                                    1974
                                            975
                                          SAMPLING  DATE
          FIGURE 18.  AVERAGE MERCURY CONCENTRATIONS IN PARTICULATE MATTER LEAVING THE TREATED CHANNELS.

-------
ui
U)
                                   Hg  INPUT    Hg INPUT
                                   INCREASED    OFF
                                                           O —
                              CONTROL
                               0.01/tg/l
                1972
1973

 SAMPLING DATE
1974
1975
        FIGURE 19,   AVERAGE MERCURY CONCENTRATION  IN PERIPHYTON REMOVED FROM THE PVC WALLS OF
        THE  CONTROL CHANNELS AND CHANNELS RECEIVING AN INPUT OF 0,01/ig/l.

-------
Results of the Hg analyses of periphyton removed from the
PVC in the 1.0 and 5.0 yg/1 channels are shown in Figure 20.
The levels in the 1.0-yg/l channels reached a peak of 3000
yg/g after about seven months.  They remained constant at
this level for about five months and then, for no obvious
reason, began to decrease until they reached 600 yg/g by
the time the Hg was turned off.  During the last six months
of the project, the levels remained relatively constant at
150-200 yg/g.  This was slightly lower than the concentration
measured in the export leaving the channels during this
same period  (Figure 18) which indicates at least some
contribution to the export from the sediments.

Figure 20 shows that the periphyton Hg levels in the 5.0
yg/1 channels were very erratic.  The high value observed
in October, 1973, which like all other points is an average
of four samples, appears to be a valid measurement since
all four samples were high.  After this peak, the decrease
in levels in the 5.0-yg/l channels parallels the observations
from the 1.0-yg/1 channels, although the values reached just
prior to the end of the project were slightly higher.

Data for periphyton removed from the plexiglass plates are
summarized in Figure 21.  Periphyton removed from the
plexiglass plates always had much higher Hg levels than
that scraped from the PVC walls.  This difference was
maintained until the end of the project when periphyton
removed from the plexiglass plates had concentrations of
approximately 1000 yg/g and PVC periphyton had concentrations
of 200 yg/g.  The highest Hg levels measured in plexiglass
plate samples  (>8000 yg/g in January, 1974) were still
lower than  those reported by Matida and Kumada (1969)  for
attached algae samples taken from a contaminated river in
Japan.  Concentration factors calculated from the maximum
periphyton levels measured in plexiglass plate samples
range from 1 X 104 to 3 X 104.  This range is considerably
higher than the values reported by Bannerz (1968) for algae
from experimental ponds but is very similar to the range of
concentration factors reported by Fujita and Hashizume  (1972)
for a freshwater diatom.

Although the water levels of Hg during the last year of the
project were below detection limits, periphyton colonizing
new plexiglass plates put into a channel formerly treated
with Hg levels of 5 yg/1 in June, 1974, acquired equilibrium
Hg concentrations of 45 yg/g after two months.
                              54

-------
Ul
Oi
        o>
        1:  3000
       o
       a:
        LU
        o
        o
        o
X
z
o
        X
        Q_

        ir
        UJ
        Q_
            2000
             000
        0
                                       Hg INPUT    Hg INPUT
                                       INCREASED     OFF
                                                         Q
                                                         A— 5.0 /tg/ 1
              0
NDJFMAMJJASONDJ
                 F M A
M J J A S 0
N
D J
             1972
1973

 SAMPLING DATE
 1974
    1975
           FIGURE 20,  AVERAGE MERCURY CONCENTRATIONS IN PERIPHYTON REMOVED FROM THE PVC WALLS
           OF CHANNELS RECEIVING INPUTS OF 1,0 AND 5.0/i.g/I.

-------
U1
         en
            9000 r
         o
         or
            8000  -
            7000  -
            6000
         LU
         O  5000
         "Z.
         o
            4000  -
         en

            3000  -
            2000 -
            1000 -
        o:
        LU
        Q_
                                        INPUT

                                      OFF
     MJJAS
                    1973
          1974

SAMPLING  DATE
1975
              FIGURE 21.  AVERAGE MERCURY CONCENTRATIONS IN PERIPHYTON REMOVED FROM PLEXIGLASS PLATES  IN
              CHANNELS RECEIVING  INPUTS OF 1.0 AND 5.0 fJ.g/\.

-------
Macrophytes

In the first 18 months of channel operation,  the bottom
communities were relatively simple with no rooted vegetation.
By the spring of 1973, however, a great number of small
rooted plants were noted, especially in the upstream portions
of the outer channels.  Although there appeared to be two
distinct species, these were later identified as variants
of a single species, Juncus diffusissimus.  The distribution
of this plant in the channels in April of 1973 is shown in
Table 3.  Quite obviously, channel one received a much
greater input of seeds than the other channels.  The upstream-
downstream gradient in the distribution pattern may be due
to the input of seeds to the pools above the  channels.

Starting in June, 1973, samples of Juncus were periodically
analyzed for Hg content.  Results of root analyses are
shown in Figure 22, results of leaf analyses  are shown in
Figure 23.  No data are presented for the control channels
or the channels receiving an input of 0.01 yg/1 -because the
values were so low.  In all cases, when Hg was being input
to the channels, the levels in the leaves were higher than
in the roots.  The root levels were significant, however,
indicating movement  of  Hg within the  plants
since the sand in which the roots were growing had an
extremely low Hg content.  The rapid increase in root
levels during the spring of 1974, several months after the
Hg inputs were stopped, also indicates internal shifting
of Hg.  The Hg in the leaves of the Juncus appears to have
been tightly bound since there was little Hg  loss from the
leaves for several months after the inputs were stopped.
The rapid decrease in leaf concentrations beginning in
May was probably due to dilution by growth and the inclusion
of newly grown plants, which had not been exposed to Hg, in
the samples.

During the latter part of the period when Hg  was being input,
the levels in the Juncus leaves were in approximately the
same ratio (1:5) as the water levels.  This is one of the
few components of the channel communities that showed this
relationship.  At the termination of the project, Hg levels
in the roots and stems were approximately equal in both
treatments, and were declining very slowly.

Concentration factors calculated from the maximum Hg levels
reached in the stems during the project are on the order of
2.5 X 103.  The conversion factor for estimating wet-weight
concentrations from the ash-free dry weight data was 0.13
and was derived experimentally from Juncus removed from the
                              57

-------
Table 3.  DISTRIBUTION OP JUNCUS DIFPUSISSIMUS ON APRIL 2,
1973, IN THE ARTIFICIAL STREAMS
(number of plants/interval)

Channel number
1
Distance control
from input
weir, meters
0-7.5
7.5-15
15-22.5
22.5-30
30-37.5
37.5-45
45-52.5
52.5-60
60-67.5
67.5-75
75-82.5
82.5-90
Total
453
261
144
98
97
124
222
70
38
26
49
236
1818
2
1.0
yg/i
112
131
30
23
13
18
31
13
6
0
17
18
412
3
0.01
yg/i
30
8
8
3
3
11
20
3
0
3
2
5
96
4
control
170
39
16
10
11
34
33
6
2
7
2
31
361
5
1.0
yg/i
139
10
9
5
8
18
25
1
4
1
12
5
237
6
0.01
ug/i
294
40
20
11
5
13
16
3
2
5
5
9
423
channels.  The concentration factors determined by Hannerz
(1968) for rooted aquatics seldom exceeded 100, although
Dietz (1972) calculated values > 430 for submerged
spermatophytes.
                             58

-------
VD
             400 -
 I
g
<
h- o» 300 -
       LU
       o
          o»
          o
       1
             200
              100
                0
Hg INPUT
INCREASED
Hg INPUT
   OFF
                                                                 1.0 jig/I
                                                                 5.0
                          973
                                  1974
                   SAMPLING DATE
                                      1975
        FIGURE 22,   AVERAGE MERCURY CONCENTRATIONS IN ROOTS OF JUNCUS DIFFUSISSIMUS REMOVED
        FROM CHANNELS RECEIVING INPUTS OF 1,0 AND 5,0 /xg/1.

-------
                                                           D —1.0/j.g/l
                                                             — 5.0 fig/1
             Hg INPUT
            INCREASED
                                                                                   975
                                     SAMPLING  DATE
FIGURE 23.   AVERAGE MERCURY CONCENTRATIONS IN LEAVES OF JJJNCUS DlEfiiS.LS_si/iu.s REMOVED FROM CHANNELS
RECEIVING INPUTS OF 1.0 AND 5.0 fJ.Q/\.

-------
Aquatic Insects

Early in the study, the larger aquatic  insects were
relatively rare in the channels and no  attempt was made
to collect these on a routine basis.  During  the entire
project period, the organisms removed from  the benthic
dishes were never available in sufficient numbers to
provide an adequate biomass for Hg analyses.  Also, the
taxonomic problems associated with the  groups most common
in the benthic samples made it impossible to  identify and
enumerate these organisms without first preserving them.

The number of species and population densities of aquatic
insects increased as the rooted macrophytes became an
important channel component.  During the last 14 months of
the project, the material collected in  the  fish-retaining
end screens was periodically searched for macroinvertebrates
for analysis.  The organism for which most  data were
collected was a damselfly of the genus  Ischnura.  The data
are summarized in Figure 24 in which each point represents
the average of two to nine analyses, each of  an individual
nymph.  The average coefficient of variation  (CV)
associated with the results from the 1.0-yg/l channels was
46% and from the 5.0-yg/l channels, 40%.  Results for
Ischnura collected from the control channels  are not shown
because the numbers were typically very low (<0.1 yg/g)
and also variable  (CV=116%).  Figure 24 shows that the Hg
levels in this damselfly were probably  a result of direct
water uptake.  The concentrations in animals  collected in
the two treatments were approximately in the  same ratio as
the water levels (5:1) and these concentrations converged
rapidly to low levels after the Hg inputs were stopped.

During the Hg input period, another damselfly, Argia sp.,
was collected occasionally from all channels  other than
those receiving an input of 5.0 yg/1-   Based  on a sample
of six animals collected from the 1.0-yg/l  channels during
January, 1974, the mean Hg level of 12.8 yg/g (SD=5.7)
for this damselfly was about the same as that found in
Ischnura sp. taken from the same channels at  about the same
time.  Argia collected from the same channels in February
and March of 1974 had levels that ranged from 3-5 yg/g,
values very similar to those observed for Ischnura.  In the
time interval between the Hg shut-off and the project
termination, Argia became established in channels 3 and 6
and at the end of the project, levels measured in this
species were 0.25, 1.7 and 2.2 yg/g respectively for
animals from the control, former 1.0-yg/1 and former-5.0
yg/1 channels.  Levels in Ischnura at the end of the  study
                             61

-------
NJ
                     Hg INPUT
                     OFF
                                                    D —1.0/z.g/l
                                                    A —5.0/xg/l
                  1974
                                          SAMPLING  DATE
1975
           FIGURE 24.  AVERAGE MERCURY CONCENTRATIONS IN DAMSELFLY NYMPHS (ISC.ENJJRA SP.)  COLLECTED
           FROM CHANNELS RECEIVING  INPUTS OF 1.0 AND 5.0 {J.Q/\.

-------
were 0.09, 2.6 and 2.8 yg/g in these  same  channels, again
indicating the two species were quite similar with respect
to Hg uptake and elimination.

Analyses were made of other immature  insects, including
several other species of odonates, dipterans, aquatic
lepidopterans and several species of  adult coleopterans,
notonectids and corixids.  Levels for other odonate nymph
species were about the same as those  presented for Ischnura
and Argia.  Values for the adult forms (notonectids"]!
coleopterans and corixids) were extremely  variable,
probably  as a result  of movement of  these  groups in and
out of the channels.  During the course of the project,
corixids and other adult insects were seen flying from
one channel to another.  In January,  1974, corixids were
observed to be very common in the channels and 18 individuals
(Hesperocorixa sp.) were collected, three  from each channel.
Mean values (x + SD) for the Hg concentrations in the three
treatments were 0.11 + 0.13 ug/g, 0.34 + 0.20 yg/g and
3.91 + 5.5 yg/g for the controls, 1.0 yg/1 and 5.0 yg/1
treatments, respectively.

In August, 1974, a large number of uncontaminated
Hesperocorixa were collected from a nearby pond.  These
animals were placed in the continuous-flow chambers
previously described  to determine their ability to take
up Hg directly from the water  and to measure the
variability of their concentrating ability under controlled
conditions.  The results are presented in  Figure 25.  The
relatively rapid uptake by the unfed  animals indicates that
direct uptake by this species was of  major importance in
the channels.   The lower variability  in the CFC-treated
animals in contrast to the channel organisms was probably
due to the fact that they were confined to a given treatment
and there was no exchange between treatments as  took place
in the channels.

Midge larvae were an important link in the food  web of the
channel communities, and were numerically  the most abundant
invertebrates.   Because of the small  size  of the individual
organisms, composite samples of a number of individuals
were made for Hg analyses.  Average concentrations in midge
larvae at the time the Hg inputs were terminated ranged
from below detection limits in the control channels to
8 yg/g in the 1.0-yg/l channels and 20 yg/g in the 5.0-
yg/1 channels.

The concentrations of Hg found in the channel insects were
much lower than those found by Matida and  Kumada (1969) in
                              63

-------
en

 •»
^

O

H
cr
h-
LJ
O
O
O

CP
X

Q

X
cr
O
o
    4.0 r
    3.0
    2.0
           • —CONTROL

           A—5.0/ig/l
             I      23456789

                        EXPOSURE  TIME,  days

     FIGURE 25,   MERCURY  CONCENTRATIONS (X ± 1SD)  ACQUIRED BY CORIXIDS
     (HESPERQCORIXA SP.) CONFINED TO FLOWING WATER  CONTAINING 5,(
     HG++.  EACH  POINT is  BASED ON ANALYSES OF 10 ORGANISMS,

-------
a contaminated Japanese river although the concentration
factors that can be calculated from the channel results,
are much higher than those reported by Hannerz  (1968) for
his pond experiments with
Fish

The background level of Hg in the mosquitofish collected
for this project, as determined in April, 1972, was 0.04 +
0.01 yg/g (X + 2SE, n=20) .  This value is much lower than"
the 0.22 yg/g (wet weight) level in the experimental
mosquitofish used by Schoper  (1974).  The pond from which
the fish for this project were taken is located adjacent
to an installation on the SRP which consumes a large
quantity of coal.  The large stack releases from this
installation do not appear to have affected the pond
system, although other components have not been checked.

During the first year of the project, a number of fish
escaped from both the channels and cages when the screens
at the ends of the channels plugged and water levels became
excessive.  The reduced numbers of fish made sampling
difficult and also greatly limited the number of samples
that could be taken.  It was for this reason that all
remaining fish were removed at the end of the first year
and new ones introduced.

The results of analyses of fish removed from the 1.0-yg/l
channels  (2 and 5) between May, 1972, and May, 1973, are
presented in Figure 26.  Fish in channel 2 acquired Hg
faster and to a greater extent than those in channel 5.
In May, 1973, ,the average Hg concentration in_the 21 fish
removed from channel 2 was 5.30 + 0.96 yg/g (X + 2SE) , in
the 62 fish removed from channel 5, 3.41 + 0.30 (X + 2SE) .
The means are significantly different (p < 0-01)  based on
a t-test of the log transformed (to uncouple the variance
and mean) data.   There is no apparent reason for the
differences between channels.  The fish population in
channel 5 was greater than in channel 2 but the average
weight of the fish did not differ (0.52 g in channel 2,
0.57 g in channel 5) when the populations were removed in
May, 1973.

Figure 26 shows  that the Hg levels in the mosquitofish
exposed to 1.0 yg/1 rose slowly over a 10- month period to
equilibrium levels of 4-7 yg/g.  There may be a seasonal
cycle indicated,  with a peak occurring during the winter
months, especially if the values for the fish collected
during the population removal in May, 1973, are included.
                              65

-------
CPi
             ?  8
LU
O

O
O

C7>
X
            5
4

3

2
                 0
                         D — CHANNEL 2
                         • — CHANNEL 5
M
J
J
A
S
0
N
D
J
F
M
A
M
                     1972
                        SAMPLING DATE
                                          1973
           FIGURE 26.   MERCURY CONCENTRATIONS IN MOSQUITOFISH (GAMBUSIA AFFINIS)
           REMOVED DURING THE FIRST YEAR  FROM THE CHANNELS RECEIVING AN INPUT OF
           l.Oft.g/1.    EACH POINT is AN AVERAGE BASED ON TWO ANALYSES OF
           INDIVIDUAL FISH.

-------
A similar seasonal trend is indicated  in Figure  27 which
summarizes the results of analyses of  fish removed from the
control channels and those receiving an input of 0.01 yg/1
Hg.  Figure 27 shows that at levels of even  0.01 yg/1/
mosquitofish accumulate Hg above control levels.  The
increase in the control fish is probably due to  air transfers
of Hg between channels although this may be  a natural
seasonal phenomenon that has simply not been observed before.
Statistical tests of measurements made on the fish removed
from the control and 0.01-yg/1 channels show that there were
no significant differences in Hg levels and  weights of fish
between channels treated alike, but that there was a
significant (p < 0.01) difference between the Hg _
concentrations in the control fish (0.072 +_  .009, X + 2SE,
n=34) and those removed from the 0.01-yg/l~channels ~
(0.179 + 0.007, X + 2SE, n=99).

The Hg levels in fish removed from the control channels
after 1 year were significantly (p < 0.01) higher than
those measured in the fish at the beginning  of the project.
This may reflect a natural increase with age, or it may be
due to cross contamination between channels, as  mentioned
above.  The differences in population  sizes  between channels
can in no way be attributed to treatment since so many fish
escaped from the channels.

Linear correlation coefficients between the  body weights
and the Hg concentrations in fish were calculated on a
channel basis for all fish sampled between May,  1972,
and May, 1973.  Only in channel 6 was  there  a significant
(r=0.82, n=12, p < 0.01) correlation.  Correlation
coefficients for the same variables were calculated with
data collected from the fish populations removed  from the
channels in May, 1973.  There was a significant  negative
correlation (r=-.61, n=21, p < 0.01)  for animals  removed
from channel 2.  There does not appear to be a neat,
consistant relationship between weight and Hg levels in
mosquitofish exposed to low levels of  Hg over a  long time
period.   This is not consistent with the data reported by
Johnels et al. (1967), Johnels and Westermark (1970),
Fimreite et al. (1971), Knight and Herring (1972)  and
Kelso and Frank (1974) although Scott  and Armstrong (1972)
found all possible relationships (positive,  negative, and
zero correlation coefficients)  between Hg levels and fish
lengtl^ depending on what species they  were considering.

At the end of one year of exposure, there was no difference
between the Hg levels in fish living in the  channels and
those confined to cages and fed with a commercial tropical
                              67

-------
                  t  .5
CO
                 O
                 H
                 <
                 o:
                 UJ
                 o
                 z
                 o
                 o
                 I
                 1
                 5
A
      • —CONTROL
      0—0.01
    M
ASON   DJFMA
M
                 1972                      1973
                     SAMPLING DATE
                FIGURE  27.   MERCURY CONCENTRATIONS IN MOSQUITOFISH  (fiAMflii&iA AFFINIS)
                REMOVED DURING THE FIRST YEAR FROM THE CONTROL CHANNELS AND THE CHANNELS
                RECEIVING AN INPUT OF 0.01/J.g/l.     EACH POINT  IS AN AVERAGE BASED ON
                FOUR ANALYSES OF INDIVIDUAL FISH,

-------
fish food containing 0.20 + 0.02 yg/g Hg on a dry weight
basis.  This indicates that direct uptake of Hg  from  the
water makes a significant contribution to the body burdens
of mosquitofish.  Based on two sets of analyses  on composite
samples of several fish, the portion of the total Hg  present
as methyl mercury was 28% in mosquitofish exposed for one
year.  This value is low compared to published results
(Kamps et al., 1972; Bache et al., 1971) probably  because the
entire fish was digested for total Hg  (including gut
contents) rather than only the muscle tissue of  the fish.

Fish data collected the last 20 months of the project from
fish placed in the channels in June, 1973, are summarized
in Figure 28.  Mercury levels in fish from the control
channels ranged from 0.005 to 0.17 yg/g.  The data are
averaged for treatments since there were no significant
differences between channels treated alike during this
period.

Figure 28 shows that the fish in the 1.0-yg/l channels
acquired about the same Hg concentrations as the fish of the
previous year.  This equilibrium level was reached more
quickly however, (6 months as opposed to 10), which
indicates that fish placed in a contaminated area
concentrated Hg faster than those placed in a clean area
that was then contaminated.  This result implies that
direct uptake of Hg++ from the water is not the  only  method
by which mosquitofish acquire Hg.  Food chain uptake  may
be involved, although the conversion of Hg++ to elemental
Hg, a process already mentioned as occurring in  the
channels, may also be important.  Schoper (1974) found
that elemental Hg is more rapidly taken up by mosquitofish
than Hg++.

Figure 28 shows the rapid increase in Hg levels  in fish
living in the channels when the concentrations were changed
from 0.01 yg/1 to 5.0 yg/1.  The equilibrium level  of
11-14 yg/g in the fish was reached in only four  months.
Although the Hg concentrations in the water of the treated
channels were in the ratio of 5:1, the levels reached by
the fish were only 2:1.  Hence the concentrations attained
by fish were not a simple linear function of water
concentrations.

The levels in all treated fish decreased to approximately
3 yg/g within three months after the Hg inputs were turned
off.  These data are variable but indicate that  biological
half-lives may be a function of the initial concentration
levels of Hg.  From May, 1974, until the end of  the project.
                              69

-------
 en

  9+

O

h-

cr
h-

LJ
O

O
O

 en
                                                        D— 1.0 p.q/\
                                                          —.01 fj.q/\
         L Hg INPUT
           INCREASED
      0
                    973
                                                  1974
                                 SAMPLING  DATE
1975
FIGURE 28.   MERCURY CONCENTRATIONS IN MOSQUITOFISH (£AMM)SXA AEFJIUJS)  REMOVED FROM THE
TREATED CHANNELS DURING THE LAST 21 MONTHS OF THE PROJECT.  EACH POINT  IS AN AVERAGE BASED
ON FOUR ANALYSES OF  INDIVIDUAL FISH.

-------
there were no differences between fish exposed to the two
treatment levels.  By September, 1974, an equilibrium was
reached in the contaminated fish at a level significantly
above that measured in the controls.  This level was
maintained until the project was terminated in February,
1975.  At that time the average concentration in fish from
the treated channels was approximately 0.4 ug/g.  Only a
small fraction of the fish recovered at the end of the
project have been analyzed.

During the project period included in Figure 28, there was
successful mosquitofish reproduction in the channels.
The increased community complexity that developed provided
the necessary cover for young animals.  Fish collected
during this interval (June, 1973-March, 1975), might have
been born in the channels.  This was not true for fish
collected the first year.

A single set of methyl mercury analyses was made on
composite samples of fish removed from the channels in
August, 1973.  Although the total Hg concentrations ranged
as high as 5.6 yg/g, the portion of Hg present as methyl
mercury was on the order of 10%, considerably lower than
the 28% value calculated for fish removed from the channel
after a full year.  This suggests that the proportion of
total Hg present in fish as methyl mercury increases with
exposure time.  This result was also observed by Bache et al.
(1971) who found that the proportion of methyl mercury to
total Hg increased with age in lake trout.

Data obtained from the half-life study (Figure 29) show
that the Hg is removed from mosquitofish rapidly when the
animals are transferred to clean water.  The best estimate
of the biological half life, based on the data, is 13.5
days.  (C.I.95% = 9.9-23.1) based on a least squares fit
assuming first order kinetics.  This half-life, obtained
from fish that had been in a contaminated environment their
entire lives, is considerably lower than the 60-day half
life reported by Huckabee and Blaylock (1973) for
mosquitofish contaminated by a short-term direct exposure.
Schoper (1974) calculated a half-life of 45 days for Hg in
mosquitofish but her data are limited and she presents no
confidence interval for her estimate.

BIOLOGICAL RESPONSES

Introduction

During the 41-month project period, the biological
communities that developed in the artificial streams
                              71

-------
                 10     20    30   40    50   60    70

                   TIME  IN  Hg-FREE WATER, days
80
FIGURE 29,   CHANGE IN MERCURY CONCENTRATIONS IN MOSQUITOFISH (GAMBUSIA AFFINIS)
MOVED FROM A CONTAMINATED AREA TO CLEAR WATER,  EACH POINT IS BASED ON 10
ANALYSES, EACH OF AN  INDIVIDUAL FISH,

-------
were undergoing continuous changes and no  equilibrium
systems evolved.  There was a continual downstream
expansion of the macrophyte populations, a constant
rearrangement of insect and algal -populations,  and a
continuous influx of new plant and animal  species into
the channels.  It is within this constantly  changing
pattern that the effects of Hg must be considered.  In
the statistical analyses of the data, time-treatment
interactions often appeared as significant factors in
comparisons between treatments.

In the initial planning, it was assumed that the only
factor that would vary among the six channels would be
the concentration of Hg++ in the water.  There  were
several deviations from this ideal situation, however,
resulting from the physical structure and  positioning
of the channels.  These included:   (1) differential inputs
of seed material resulting from the orientation of the
channels with respect to the prevailing winds,  (2)
differential disturbances by vertebrates,  including
mammals and amphibians, which were largely confined to the
outermost channels, and (3) the input of dissimilar water
qualities which was a result of the different flow rates
through the limestone water treatment systems made necessary
by the development of major leaks in several of the head
pools.  These factors contributed greatly  to the variability
of the data and decreased the sensitivity  of the statistical
procedures to reject a null hypothesis of  "no effect  of Hg"
on the various biological parameters studied.   Nevertheless,
there were effects observed that can be related to the
input of Hg.

Ferens (1974) summarizes the data regarding  the effects of
Hg on the benthic and periphytic communities in the
channels prior to the dose increases in August, 1973.
Even at concentrations of 0-01 and 1.0 Ug/1/  levels not
uncommon in natural waters (Wershaw, 1970; Jenne, 1972),
she found small but significant reductions in algal
number, biomass, and diversity, although the number of
species affected was small, and there was  no indication
that these effects were transferred to consumer trophic
levels.

Taxonomy

Taxonomic efforts were concentrated on the algae, vascular
plants, and insects in the channels.  Early  in  the study,
for the identification of algae, a number  of samples  were
sent to Dr.  L.  A. Whitford at the North Carolina State
                             73

-------
University.  As new species were found, they were identified
by keys and descriptions presented in Huber-Pestalozzi
(1938, 1941, 1955), Prescott (1962, 1970), Ralfs  (1848),
Smith  (1950) , Tiffany and Britton  (1952) , and Whitford and
Schumacher  (1969) .  Vascular plants were identified by keys
in Radford et al.  (1968).  The identity of the dominant
plant, Juneus diffusissimus was confirmed by Dr. J. Coffee
of Queens College, Charlotte, N. C.  Most of the insects
were identified  as to genus by reference to keys  in Pennak
(1953) and Edmondson (1959) .  Certain dipterans were
further identified by Beck (1968) and Mason (1973) and
vertebrates were keyed with Blair et al. (1957).

A complete species list of identified channel biota is
presented in the Appendix.  In addition to these organisms,
a variety of bacteria, fungi, protozoa, nematodes, rotifers,
tardigrads, oligochaetes, cladocerans, ostracods, copepods,
and aquatic mites  was  also observed during the study.
This assemblage of organisms that colonized the streams,
including at least 50 species of algae, 5 species of vascular
plants, and 61 species of insects, approached a natural
ecosystem in complexity and possibilities of interaction.
In their review of artificial stream research, Warren and
Davis  (1971) stress the relationship between the complexity
of artificial stream systems and the conclusions that can
be extrapolated from them to natural systems.  Because of
the duration of the Hg study, the complexity of the
biological communities that evolved, and the similarity
of the water to natural streams in this region, data
obtained from the  artificial streams should be directly
applicable to soft-water streams of the southeastern
United States and  in a more general way to all aquatic
systems.

During the study,  with the exception of amphibians, no
attempt was made to limit recruitment of new species into
the channels.  Frog egg masses were removed because these
were found only in the  outermost pairs of channels.  When
it became obvious  that some eggs had been missed  (September,
1973) , an attempt  was made to place an equal number of eggs
(Rana pipiens) in  each channel.  Based on later observations,
it appears that few, if any, of the eggs placed in Hg-treated
channels hatched.  This was probably a direct result of  the
Hg treatment.  Birge et al.(1974)  found that levels of 10
Vtg/1 of inorganic  Hg were 100% effective in preventing the
development of Rana pipiens embryos when applied  during  the
cleavage of biastula stage.  During the period after the Hg
inputs were stopped, amphibian eggs successfully  completed
development in all channels and an unknown number of tadpoles
of at least three  species  (see Appendix) were observed.
                              74

-------
At the end of the project, an attempt was made to
quantitatively remove all vertebrates and larger
invertebrates from the channels.  The results of these
efforts are summarized in Table  4.  These results were
somewhat unexpected in that the  great differences that
existed between channels with respect to fish, snail and
tadpole densities were never obvious from the routine
sampling programs.  There is no  question but that the
density differences in these animals, which appear to be
totally unrelated to Hg treatment, affected other channel
components.  These effects contribute to the variability
of the data and make it even more difficult to determine
responses due to Hg treatment.
All of the algae identified in the artificial streams
during the course of this study could be described as
tychoplanktonic, attached or benthic algae characteristic
of soft, acidic waters of low-current velocity.  The algal
community was clearly dominated by members of Chlorophyta
(green algae) with desmids and filamentous forms the most
abundant.  Diatoms were very poorly represented with only
four species observed and only two of these occurring
regularly.  The diatoms were not  limited in the streams
by a lack of silicon in the water  (Table 1), as the level
observed was well above the silicon concentration in nearby
diatom-dominated systems, but rather from a combination of
adverse environmental parameters  such as the lack of shading,
the very slow current velocity, and the year-round water
temperatures above 20°C.  The Cyanophyta or blue-greens
were never dominant either, but one species, Merismopedia
punctata, was present in almost every sample taken.
Filtering of large volumes of water from the channels
indicated that truly planktonic algae were not common
although Dictyosphaerium pulchellum was occasionally seen.

Community Responses-
Of the various substrates in the  streams that were
regularly sampled, the walls contained the oldest and
least-disturbed communities of periphyton.  Figure 30
summarizes the biomass results obtained from the PVC strips
during the first part of the study.  Figure 31 shows the
biomass levels of the samples from the stream walls during
the latter part of study.  The apparent abrupt increase in
wall biomass between October, 1973, and January, 1974, with
respect to the 5-yg/l channels may have been a result of
the change in sampling technique  rather than a real increase
                              75

-------
 Table 4.  ORGANISMS COLLECTED  FROM  THE ARITIFICIAL STREAMS,
                        FEBRUARY  11-16, 1975

Organism
(* indicates
a new species
on this date)
Odonata :
Argia sp.
Ischnura sp.
Erythrodiplax
miniscula
Pantala hymenea
*Ladona deplanta
*Celithemis
fasciata
*Tetragoneuria
semiaquea
*Plathemis lydia
*Pacydiplax
longipennis
1
Channel number
4253
control
430
7
162
10
3
11
7
1
637
131
187
3
12
110
6
i.o yg/l 5.0 yg/i
545 591 465
101 3L8 314
858 451 297
4
14
12 89 84
1
2
53
296
47
8
  Ephemeroptera:

  Caenis  sp.
  Callibaetis  sp.
26
14
 4
 18
  4
68
 3
 15
  5
  Hemiptera:

 *Ranatra  sp.
  Pelocoris femoratus^
  Hesperocorixa spp.
21
44
 2
14
40
 31
 44
20
46
  9
 22
  2
154
 Megaloptera:

*Chauliodes sp.
            14
 Diptera:

 Tanypodinae
 Chironominae
 pupae
*Tipula sp.
 Helius sp.
 Dasyhelea sp.
10
53
 6
 6
 4
 1
  3
186
 12
 27
  2
  6
 4
 6

 3
 1
 1
140

  1
  2
  2
                               76

-------
Table 4 (continued)
        Channel Number
1	4     253
Control
  Organism
     1.0
                       b.U yig/r
 Lepidoptera:

 Parapoynx sp.
*Synclita sp.
 3
 1
13
             8
            32
19
23
34
 2
 3
 3
 Coleoptera:

 Berosus spp.
 Coptotomus sp.
* Trop i sternus II
*Tropisternus #2
 Hydocanthus sp.
 Hydroporus sp.
*Agabus fl
*Agabus #2
*Enochrus #1
*Enochrus #2
 Laccophilus sp.
 4
 2
 2
 1
44
 9
 6
 1

13
 1
            18
             6

             1
            62
             9

             1
             1
33
 2
 1

50
 1
24
 7
11
15
                         1
                         1
38
 6
 4
 9
 1
 3
 Others:
snails
(Phy^sa and Lymnea)
Rana spp .
Gambusia affinis

Total Insects
Insects/meter 2
£ insect species
H - insects
E - insects
Grand Total
# species
H
E


30
167

858
15.4
25
2.58
0.56
1055
27
2.91
0.61
405

30
246

1204
21.6
20
2.31
0.53
1886
23
2.84
0.63
4

284
217

2001
35.9
24
2.56
0.56
2506
27
2.98
0.63
262

489
172

1750
31.4
24
2.70
0.59
2673
27
3.20
0.67


129
148

1453
26.1
21
2.73
0.62
1730
23
3.08
0.68
1077

629
76

628
11.3
15
2.28
0.58
2410
18
2.28
0.55
                              77

-------
00
      CM
       E

       CP
       **
      CO
      CO
      o
      DQ  3

      -z.
      O
      cr
      LU
      Q_
          0
                                    •-CONTROL
                                    • -I.O/ig/l
                                    O-.OI/ig/l
                               Hg INPUT
                               INCREASED
                    A
0 I  N
D
M
A
M
0
           1972                      1973

                                SAMPLING  DATE

          FIGURE 30.  BIOMASS OF PERIPHYTON  SCRAPED FROM PVC  STRIPS
          SUSPENDED  IN THE  CHANNELS.

-------
vo
                       Hg INPUT OFF
                                                  •—CONTROL
                                                  D — I.O/ig/l
                                               A   S  0  N   D
                     1974
                                    SAMPLING DATE
1975
                 FIGURE 31.   BIOMASS OF PERIPHYTON SCRAPED FROM CHANNEL WALLS.

-------
 during this time interval.  There was a certain amount of
 loss from the PVC strips when they were removed from the
 channels and these losses were relatively greater when the
 growth was heavy, as it was in the 5.0-yg/l channels.

 The periphytic biomass in all channels fluctuated similarly
 during the summer of 1972 probably due to environmental
 changes, but in September, channels 3 and 6 (0.01 yg/1 at
 that time) did not follow the biomass reductions that
 occurred in the other streams, indicating a possible
 stimulation in growth at low-Hg levels.  After the Hg
 levels in channels 3 and 6 were increased to 5.0 yg/1 in
 August, 1973, the walls of these channels continued to
 support higher periphytic biomasses than in the other
 channels.  Figure 31 illustrates a direct correlation
 between the Hg concentration in the streams and the standing
 crops of periphyton inhabiting the walls, with a subsequent
 equalization in biomass levels after the input was shut off.
 A stimulation of algal growth at a Hg concentration as high
 as 50 pg/1 has been reported (Matsui and Gloyna, 1972).

 Figure 32 shows that the algal component of the wall
 periphytic communities, as summarized by total cell volumes,
 followed the same general pattern as shown in the biomass
 data.  In October, 1974, the total algal volume was greater
 in the control channels than in- the treated, although this
 pattern is not repeated in the biomass data of the same
 date.  This indicates that, at least at that sampling time,
 the non-algal component of the periphyton, either living
 or dead, made up a greater percentage of the organic matter
 on the walls of the Hg-treated channels.

 Figure 33 summarizes the calculated diversity and evenness
 values for these algal components of the periphytic
 communities in time.  The control and low-level streams were
 quite similar with respect to these parameters  (as they were
 in biomass and algal volume), and the high-level streams
 initially had decreased diversity and evenness values.  An
 analysis of variance (ANOVA) of the data collected while Hg
 was still being input to the channels showed there was a
 significant treatment effect (p < .05) on H, and highly
 significant effects (p  < .01) on E and cell volume.

Although biomass values of periphyton collected from
 vertical plexiglass plates were quite variable  (Figure  34),
 certain trends are evident.  During 1972, the biomass
 levels on this substrate had slightly lower values than
were found on the PVC strips.  Channels  3 and 6 again
 diverged from the other channels in the  fall of 1972 with
                              80

-------
                 Hg INPUT OFF
00

               I973
                                                                  I975
                                 I974
                           SAMPLING  DATE
FIGURE 32,  CALCULATED VOLUME OF PERIPHYTIC ALGAE SCRAPED FROM THE CHANNEL WALLS,

-------
00
NJ
          CO °'8
          00 07
          LU u-'
          ^ 0.6

          UJ 0.5

          UJ 0.4

             0.3
           02.5
              2.0
          CO
          tr  1.5
          LU
          Q
           I
          IX
1.0
                 I     1
                      Hg INPUT OFF
I    I
                                                        I    I
D   J    F !  M
                                     A
M    J   J
                                                  0
      N    D
                 1973
                              1974
                       SAMPLING  DATE
              1975
          FIGURE 33.  DIVERSITY AND EVENNESS OF PERIPHYTIC ALGAE COLLECTED FROM THE CHANNEL WALLS.

-------
00
Ul
e
\
en
       C/)
       CO
       o
       00
       X
       Q_

       LT
       LU
       Q.
                 »—CONTROL
                    .01/z.g/l
                    I.O/ig/l
                    5.0/tg/l
                          Hg INPUT    Hg INPUT
                          INCREASED     OFF
M
A
M
J
J
A
S
0
N
D
J
F
M
A
M

N
D
J
F
M
A
M
J
J
A
S
0
N
D
J
            0
                                   SAMPLING  DATE

           FIGURE 34,   BIOMASS OF PERIPHYTON SCRAPED FROM VERTICAL PLEXIGLASS PLATES,
                                                                         1975

-------
higher biomass values.  This pattern continued after the Hg
input levels were increased in August, 1973, until the
winter of 1974, 11 months after the Hg was shut off.  At
the end of the project, all channels were similar.  The
graph summarizing algal cell volumes for the plexiglass
plates (Figure 35) shows that the increase in biomass
observed in streams 3 and 6 was partially the result of
increased algal growth at that time.  Figure 36 shows
that algal diversity and evenness of periphyUc  samples
removed from plexiglass plates were lowered by Hg treatment
during the period Hg was input to the streams.  These
differences were significant  (ANOVA, p < 0.01).

The results obtained from the first two glass slide
experiments are summarized in Figures 37 and 38 in terms
of the calculated algal volumes.  These two experiments
were carried out while Hg was still being added to the
channels.  Both graphs demonstrate the significant delay
in the colonization of microscope slides by algae in the
5.0-yg/l channels and the similarity between the control
and 1.0-yg/l channels.

In both experiments, the diversity and evenness of the algal
component of the periphytj.c communities were significantly
(ANOVA, p < 0.01) affected by the treatment with the highest
diversities in the high-level streams.  On a small number
of slides from these experiments examined after 40 and 80
days of colonization, the trend towards lowered algal
biomass with higher Hg exposure had disappeared and even
reversed, with the highest algal densities occurring on
the slides from the 5.0-yg/l channels.  The observed
increases in algal biomass in the high-level streams were
not accompanied with any decreases in diversity or evenness.

The initial retardation of algal growth on microscope slides
placed in the 5.0-yg/l channels disappeared in the glass
slide experiments carried out after the inputs of Hg were
stopped.  Figure 39 illustrates algal volume changes on
slides placed in the channels one day after the Hg was
turned off.  There were no significant treatment effects
found for these data with respect to cell volume, diversity
or evenness (ANOVA, p > 0.05).  Figure 40 summarizes the
data from a glass slide colonization experiment started  •
three months after inputs were stopped.  While the  former
control and low-level channels were similar in cell volumes
and diversities throughout the study, the former high-level
channels developed larger cell volumes and  lower diversity
values.
                              84

-------
                      Hg INPUT OFF
00
Ul
       CM
        E
        CO
        O   _
t
CO

UJ
Q

UJ
S
ID

o
N
D
TIT-
M
A
M
J
J
A
S
0
N
D
J
                  1973
                                  1974

                           SAMPLING  DATE
1975
           FIGURE 35.   CALCULATED VOLUME OF PERIPHYTIC ALGAE SCRAPED FROM VERTICAL

           PLEXIGLASS PLATES,

-------
             co
             CO
             UJ
             UJ
             LU
     .8
     .7
     .6
     .5
     .4
     .3
                            Hg INPUT OFF
                               I    I    I	I
                                                           J	I
J	I    L   I
CO
CTl
             CD
             O
                2.5
>-- 2.0

co
LU  1.5
             Q
             IX  1.0
                      1973
                                                               •-CONTROL
                                                               Q-I.O/tg/l
                      N|D|J|F|M|A|M|j|j|A|StO|N|D|J
                                  1974
                           SAMPLING   DATE
        1975
            FIGURE 36,  DIVERSITY AND EVENNESS OF PERIPHYTIC ALGAE COLLECTED FROM VERTICAL
            PLEXIGLASS PLATES,

-------
           CJ
CD
             £ 20
                15
£-  10
CO


UJ
Q   K
ID


O

  r>
             Ld
                 0
                  0
>—CONTROL


k—5.0/ig/l
             5     10    15    20   25    30

                COLONIZATION  TIME,days
                                         35
               FIGURE 37,   CALCULATED VOLUME OF ALGAE COLONIZING MICROSCOPE
               SLIDES DURING MERCURY INPUTS (STARTED OCTOBER 31, 1973).

-------
            
    25
    20
                  15
                  10
                  0
• — CONTROL

Q—I.O/ig/l
A — 5.0/ig/l
                          5     10     15    20    25    30

                            COLONIZATION  TIME, days
                                                    35
               FIGURE 38,   CALCULATED VOLUME OF ALGAE COLONIZING MICROSCOPE
               SLIDES DURING MERCURY INPUTS (STARTED DECEMBER 7', 1973),

-------
00
10
            CVI
                 25
             E
             e
             en
             o
Z
LU
Q


UJ
             O
              .  15
                 10
                  0
—CONTROL

—I.O/^g/l
                          5     10     15    20    25    30

                            COLONIZATION  TIME, days
                                                    35
                FIGURE 39,   CALCULATED VOLUME OF ALGAE COLONIZING MICROSCOPE
                SLIDES AFTER MERCURY INPUTS WERE STOPPED (STARTED JANUARY 30,
                1974),

-------
CM
 E
 E
    60
50
    40
    30
 I  20
 o
 UJ
 o
 >
    10
     0
CONTROL
I.O/ig/l

5.0f«.g/l
             5     10     15     20    25     30
              COLONIZATION TIME, days
                                               35
FIGURE 40,   CALCULATED VOLUME OF ALGAE COLONIZING MICROSCOPE
SLIDES AFTER MERCURY INPUTS WERE STOPPED (STARTED APRIL 25, 1975),

-------
The final glass slide colonization experiment was begun
in October, 1974, nine months after Hg inputs were  stopped.
During this experiment, the control and former 5.0  pg/1
channels were very similar, and the algal volumes in the
former 1.0-yg/l channels were consistently lower.   There
were no significant differences between the treatments
with respect to diversity and evenness.

In August, 1974, experiments were initiated in the
continuous flow chambers       to check some of the results
observed in the channels.  In the first of three experiments
with algae, mixed cultures containing Oedogonium sp.,
Stigeoclonium elongatum, Chlamydomonas sp., Oscillatoria
geminata, and Microspora quadrata, species isolated from
the artificial streams, were used as seeding material.
Three boxes received a Hg input of 5.0 yg/1 and three were
used as controls.  Colonization on these glass slides was
observed for 10 days and was found to be dominated  by the
Oscillatoria geminata and Chlamydomonas sp. with very few
§.• elongatum or Oedogonium sp. and no M. quadrata observed
on the slides.  After 10 days, the mean number of cells/mm^
was 1256 + 357 (X + SE) for the controls and 10.0 +3.6
(X + SE) for the treated chambers.  The colonization of
both 0. geminata and Chlamydomonas sp. was inhibited in
the Hg-treated chambers.  The limited number of Chlamydomonas
sp. that were observed  were immobile in palmelloid groups
that did not stir while the slides were being counted.  When
the Hg was cut off in these boxes, these cells were observed
to develop flagella and start swimming as had been
continuously observed on slides from the control chambers.
Although the M. quadrata did not develop on the glass slides
in any of the~*chambers, lush growth of this filamentous
green algae did develop in the outflow holes of the control
chambers, while very little growth occurred in this position
in the Hg chambers.

In a single-species experiment with S_. elongatum, very little
growth was obtained over a 25-day period in the treated
chambers while this and other contaminating species grew
profusely in the control chambers.  After the Hg input was
discontinued, five clean slides were placed in each chamber.
The colonization of S. elongatum was observed and found not
to differ between the chambers.  These experiments  indicated
a very pronounced retardation of growth of certain  algae at
a Hg concentration of 5.0 yg/1, and confirm the trends seen
in the artificial streams.

The observations from the artificial substrates and the
system presented above can be summarized as follows.  While
                              91

-------
the Hg was input into the 5.0 yg/1 channels, there was an
initial inhibition of algal growth which disappeared after
about 40 days and was replaced by increased cell numbers
and periphytic biomasses compared to the controls.  After
the Hg inputs were shut off, the initial inhibition
previously found was nonexistent but the trend towards
higher biomasses in the former 5.0-yg/1 channels continued
for about nine months.  The algal population dynamics in
the 1.0-yg/l channels showed very little modification as
compared to the controls throughout the study.

Individual Responses-
Thirty-one of the algal species listed in the Appendix were
found at one time or another in all of the six channels.
Although the other species were not observed in all channels,
with one exception (Microthamnion strictissimum), their
absence was not related to Hg treatments.  M. strictissimum
was found only in the four treated channels.  The species
that appeared to be absent from one or more streams were
rare wherever they were found.  Two exceptions were
Navicula notha,a motile diatom that was very abundant in
channel 6 but never seen in channel 1, and Spirogyra sp.,
which became very common in the tail area of channel 1 but
never was found in channels 3, 5, and 6.  No.definite
reasons can be presented to explain the localized abundances
of these algae, but they probably illustrate differences
between the biota of the channels caused by a combination of
random events and possibly Hg treatments.

Detailed cell counts of algae from the various substrates
indicated that certain species were important in creating
biomass differences between the treatments.  A few species
of algae were found to be quite sensitive to the low Hg
levels used, and one species exhibited morphological
alterations apparently attributable to Hg concentrations.

Oedogonium sp. - This green alga was present throughout the
study and, although a few cells which may have been oogonia
were observed, it was never observed in a true fruiting
condition and identification as to species was not possible.
Dr. L. Whitford of N. C. State University tentatively called
it 0. reinschii because of the lack of fruiting filaments
and the presence of some diamond-shaped cells he observed.
However, continuous examination of this species during  the
last year of the study indicated that it was  smaller
(usually 3.5 ym in diameter) than the typical O.  reinschii
(5-11 ym) and fit very closely the description of  sterile
filaments of 0. pusilium.
                             92

-------
Initial colonization of this species on glass microscope
slides was rapid in the control and 1.0 yg/1 channels but
significantly slower in the 5.0-yg/1 channels during the
two experiments carried out when Hg was being input.  In
two of three colonization experiments performed after the
Hg input was shut off there was no significant difference
between the former treatments.  In the third experiment,
a significantly greater growth was observed in the former
1.0-yg/l streams.

The occurrence of this species on the long-term substrates
is summarized in Figures 41 and 42.  This species was much
more abundant in the high-level streams than in the low-level
streams or in the controls during the input of Hg, but
during the year of recovery this dominance disappeared on
both the walls and plexiglass plates.  No seasonal variation
was observed for the populations in the two control channels
where this species remained at a density of about 300
cells/mm^ throughout the year.

These data indicate that the reproductive and vegetative
portions of the life cycle of Oedogonium sp. were
differentially affected by low Hg levels.  In spite of the
large population densities in the Hg-treated channels, the
slower colonization of this species on glass slides indicates
that its reproduction was affected by the Hg treatment.
Since no fruiting filaments were observed in any of the
channels, reproduction was asexual by means of zoospores.
The zoospores, or zoospore production, therefore, must have
been sensitive to the Hg levels in the water although the
growth of vegetative filaments was stimulated.

StigeocIonium elongatum - This species was found to be very
sensitive to Hg.  During the first two years, a significant
growth stimulation on plexiglass plates was observed in the
0.01-yg/1 channels and a slight reduction in abundance was
found in the 1.0-yg/l channels as compared to the controls.
When the Hg concentrations in channels 3 and 6 were increased
to 5.0 yg/1, the numbers of cells of S. elongatum were
significantly reduced.  Averages (cells/mm?)based on 14
counts from each treatment while Hg was being input were
46 +12 (X + SE) for the controls, 25+7 for the 1.0-yg/l
and 11+5 for the 5.0-yg/1 levels.

Figure 43 summarizes plexiglass plate data from the last
year of the project.  After the Hg was shut off, all of the
streams went through similar seasonal fluctuations with
summer maxima lasting for about 4-5 months.  During this
latter period, S. elongatum was much more common in
                              93

-------
cvj
 E
 E
 CD
 o
 O)
 z:
 LJ
 Q
 LJ
 O
O
O
c
If)
I
         Hg INPUT OFF
                                     CONTROL
D
J|F
M
A
M
J
J
A
S
0
N
D
J
       1973
                         1974

                    SAMPLING  DATE
                                                   1975
 FIGURE 41,   DENSITY OF QEDOGONIUM SP, ON CHANNEL WALLS BEFORE AND AFTER
 MERCURY INPUTS WERE STOPPED,

-------
>D
> ^*

E
                   Hg INPUT OFF
       o
                                                     CONTROL
                                                     5-Oftg/l
N
D
J
F
M
A
M
J
J
A
S
0
N
L>
J
                1973
                                  1974

                       SAMPLING   DATE
1975
          FIGURE 42,  DENSITY OF QEDQGONIUM SP,  ON PLEXIGLASS PLATES BEFORE AND AFTER
          MERCURY INPUTS WERE STOPPED,

-------
                             Hg  INPUT OFF
vo
N
D
J
F
M
A
M ! J
J
A
S
0
N
D
J
                       1973
        1974

SAMPLING  DATE
1975
                   FIGURE 43.   DENSITY OF STIGEQCLONIUM ELQNGATUM ON PLEXIGLASS PLATES
                   BEFORE AND AFTER  MERCURY INPUTS WERE STOPPED.

-------
channels 3 and 6 than in the other channels.  At  the  end  of
the project, its density on plexiglass plates was similar
in all channels.

Counts for this species from wall samples  are summarized
in Figure 44.  The results are variable but indicate  that
S^. elongatum was most abundant in the 5.0-pg/l  streams.
During the year after the Hg input was shut off, there was
much less variation in the numbers of £3. elongatum cells
on this substrate than was seen on the plexiglass plates.
This may have been due to the shaded location of  the
southwest-facing wall used for sampling.

Data for this species from the glass slides showed
significant differences in cell densities  due to  treatment
in the first two colonization experiments  (ANOVA,  p < 0.01)
with reduced numbers in the Hg-treated channels.   After
the Hg was shut off, this reduction in initial  colonization
disappeared with a trend towards greater densities on the
slides in the former Hg streams.

SL elongatum was isolated and tested under more controlled
conditions with the CFC system.  In the experiment in  which
the Hg concentrations were held relatively constant at
5.0 yg/1, the number of cells of this species was
significantly reduced in the treated chambers (p  < 0.01)
compared to the controls.  When the Hg input was  shut off
and clean slides were placed in the chambers, S^.  elongatum
colonized slides in all chambers at similar rates.

Oscillatoria geminata - During the first summer of the study,
this species was found to comprise numerically more than  10%
of the algae inhabiting plexiglass plates  in the  streams,
but during the next two years it was greatly reduced  in
abundance in the plexiglass plate and channel wall samples.
It appears that 0. geminata is well adapted for initial
colonization and rapid growth on a new substrate  but  is
unable to compete at later stages of succession.

This species exhibited a highly significant negative  response
to Hg treatment during the two glass slide experiments
carried out just prior to the shut-off of  mercury
(ANOVA, p < 0.01).  In the glass slide experiments after  the
Hg input was stopped, this treatment effect disappeared and
0. geminata was found to be equally distributed in the
streams.   In two experiments with the CFC  system  this
species was found to grow profusely in the controls and
not to grow at all on the slides in the 5.0-yg/l  chambers.
                              97

-------
                   Hg INPUT OFF
00
         CM
          e
          .£
          o
           *•»
          cn
          •ZL
          LU
          Q
          LJ
          O
D
J
F
M
A
M
J
J
A
S
0
N
D
J
                1973
            1974

SAMPLING  DATE
1975
            FIGURE 44,   DENSITY OF STIGEQCLONIUM ELONGATUM ON CHANNEL WALLS
            BEFORE AND AFTER MERCURY  INPUTS WERE STOPPED,

-------
Chromulina pseudonebulosa - Although C. pseudonebulosa is
flagellated, it was most often found on glass slides in a
resting condition.  During the early part of the Hg project,
this species was observed to be most abundant on the
plexiglass plates in the 1.0-yg/l channels  (high level at
that time)  (Ferens, 1974); however, it did  not appear to
be common on this substrate in later samples.  This species
showed a significantly lowered colonization rate on glass
slides in streams receiving a Hg input.  In the glass
slide experiment started one day after the  Hg input was
turned off, greatest colonization was found on the slides
in the former Hg channels.  In the last two microscope
slide experiments this species was rare and showed no
differences in growth between channels.

Geminella turfosa - There is no record of this species
being observed in North America, but the relatively
abundant filaments of narrow cells with spiral chloroplasts
(Figure 45) resemble in every way the excellent description
of Gloeotila turfosa in Sweden (Skuja, 1956).  This species
was later transferred to the genus Geminella (Ramanathan,
1964) and will be referred to by that name.

No significant density differences were seen between
treatments for this species for plexiglass  plates, walls,
or glass microscope slides.  There were, however,
morphological changes in cell shape that were apparently
caused by Hg treatment levels.  Figure 45 illustrates
normal and various abnormal cells observed  on glass
microscope slides during the input of Hg.   Abnormal cells
were often only slightly swollen but some appeared as
spheres or branched cells.  The numbers of  abnormal cells
observed during routine counts of glass microscope slides,
both before and after the Hg was shut off,  were recorded.
Table 5 lists the relative abundance of abnormal cells
observed for the various treatments and indicates a highly
significant difference between the 5.0-yg/l treatment and
controls and also between the two treatments while Hg was
on.  After the Hg inputs were discontinued, the differences
disappeared and by the time of the last glass slide
experiment (October, 1974) no more abnormal cells were
observed.   This species was never isolated  into pure culture
so laboratory experiments were not conducted.  Working with
three marine algae in axenic cultures, Nuzzi (1972) also
found morphological abnormalities associated with Hg at
concentrations in the growth medium equal to those in the
artificial streams.
                              99

-------
FIGURE 45,   QEMINELLA IUREQSA,  A.  NORMAL FILAMENT SHOWING
LOCATION OF SHEATH, B,  VARIOUS ABNORMAL CELLS OBSERVED  IN
MERCURY CHANNELS,

                          100

-------
Table 5.   INCIDENCE OF ABNORMAL GEMINELLA TURFOSA CELLS ON
            GLASS SLIDES DURING AND AFTER MERCURY INPUT
                (% OF TOTAL GEMINELLA CELLS)
                   During Mercury Input

                         Treatment

X + 2SE
n
Control
0.19 + .24
21
1.0 yg/1
1.16 + 1.01
24
5.0 ug/1
45.99 + 10.
35

40


                    After Mercury Input

                     Previous Treatment
Control
X + 2SE
n
0.42 +
17
.84

1.0
5.43

yg/i
+ 5.42
21
5.
3.84

0
+
22
yg/i
4.12

Microthamnion strictissimum - This branched green alga was
unique because it was found to be common in the four
Hg-treated streams, reaching its highest abundances on
slides placed in the 5.0-yg/1 channels, and was never seen
in the control channels.  This species was observed only
during the winter of 1973 and showed a significant positive
response to the Hg input level.  These observations suggest
that M. strictissimum might be an indicator of the presence
of mercuric ions in low concentrations.  This species was
not isolated because it quickly disappeared after the Hg
inputs were stopped.

Other species - Several other species of algae were more
dense on the plexiglass plates and walls in the treated
channels, and contributed to the higher total biomasses
measured in these channels.  Cosmarium aspherosporum, a
small placoderm desmid that was common in all channels,•
was most dense in the controls  (Figure 46 and 47) .
                              101

-------
                 Hg  INPUT OFF
o
KJ
             1973
      1974

SAMPLING  DATE
                    ^^                     - PLEXIGLASS PLATES BEFORE

-------
o
to
           CM
            E
Hg INPUT OFF
            C/)
            o
           l-
            UJ

            Q
            _J

            UJ

            O
                                      CONTROL
               °T
               21
                    .0/xg/l
D
J
F
M
A
M
J
J
A
S
0
N
D
J
                  1973
                       1974

                  SAMPLING  DATE
1975
            FIGURE l\7,   DENSITY OF COSMARiUM ASPHERQSPQRUM ON CHANNEL WALLS BEFORE

            AND AFTER MERCURY  INPUTS WERE STOPPED,

-------
 The density  differences disappeared during the  last year of
 the project.   Spondylosium planum, a  filamentous desmid, was
 also most dense  in  the 5.0-yg/l treatments and  least dense
 in the controls  while the Hg  inputs were on  (Figure 48 and
 49).   This pattern  continued  after the Hg inputs were
 terminated.   In  the short-term glass  slide colonization
 experiments,  both of these desmids showed greater growth
 in the treated channels with  the highest numbers occurring
 in the 1.0-yg/l  treatments.

 Another species  that was more dense on the plexiglass
 plates and walls of the treated channels was Merismopedia
 punctata,  a  blue-green alga of common occurrence.  In the
 short-term colonization studies, however, this  species
 showed no significant response to Hg  levels.

 Past studies concerning toxicity of Hg compounds to algae
 have all been short term in nature.   These studies have
 shown various levels of toxicity by this metal  to chlorophyll
 production (Hannan  and Patouillet, 1972), primary production
 (Harriss et  al., 1970; Knauer and Martin, 1972), and cell
 numbers (Ukeles, 1962; Nuzzi, 1972; Ben-Bassat, et al.,
 1972 and Matsui  and Gloyna, 1972).  Working with three
 species of marine phytoplankton, Nuzzi  (1972) demonstrated
 reduced growth of all three species at a Hg concentration
 of 5.0 ng/1  as HgCl2-  Knauer and Martin  (1972) also found
 this same concentration to be inhibitory to marine
 phytoplankton.   Our short term studies using the CFC
 system also  showed  this reduction in  the initial growth
 of several fresh-water algae  at a low concentration of
 HgCl2;  however,  the study in  the artificial streams
 demonstrated that an initial  inhibition of growth may not
 always affect the long-term accumulation of a species.

 The  initial  slow colonization seen in the Hg-treated
 channels  seemed  to  be a result of a lowering of the
 reproductive  potential of the algae and was related to
 the  level  of  Hg  in  the water.  The lower colonization rates
 were  found in spite of denser periphytic communities on
 the older  substrates in the treated channels, which should
 have  provided larger seeding  stocks of colonizers relative
 to the  control channels.  Apparently  the cells  that actually
 colonized  a bare substrate such as a  microscope slide were
 less  common or less successful in the Hg channels.  For
 some  of the most abundant species such as Oedogonium or
 Stigeoclonium these would have been zoospores and for
 others  they would have been merely drifting cells or
 fragmented filaments (Cosmarium spp., Oscillatoria, and
Merismopedia).
                              104

-------
o
Ul
       CVJ
       E
       o
CO
z
LU
Q
        Ld
        O
                 Hg INPUT OFF
                                            CONTROL
N
D
J
F
M
A
M
J
J
A
S
0
N
D
J
               1973
                                 1974
                         SAMPLING   DATE
1975
             FIGURE 48,  DENSITY OF SPONDYLOSIUM PLANUM ON PLEXIGLASS PLATES BEFORE
             AND AFTER MERCURY INPUTS WERE STOPPED,

-------
CM
 E
 E
 o
 LU
 O
 LU
 O
          Hg  INPUT OFF
     CONTROL
       1973
       1974
SAMPLING  DATE
                                                         1975
   FIGURE 49,   DENSITY OF SPONDYLOSIUM PLANU^ ON CHANNEL WALLS BEFORE
   AND AFTER MERCURY INPUTS WERE STOPPED,

-------
The greater algal populations that were found in the
Hg-treated channels may have resulted from nutrient level
differences between channels, or they may have resulted
from differences in consumer densities  (including parasites
and pathogens).

With respect to nutrients, the input of nitrogen with the
Hg solutions never exceeded 0.5 ug/1, which represented an
increase of 1.3% over background.  This amount is probably
not sufficient to cause the observed differences between
control and treated channels.  Nutrient stripping by
macrophytes, especially in the dense stands that formed
just below the input weirs of the channels, might have been
an important factor.  These macrophytic stands were much
more dense in the control channels than in the treated,
especially just below the Hg-input tubes.  An enrichment
of the water caused by death and decay of organisms coming
into the channels from the head pools when these organisms
were subjected to Hg inputs may have been a factor.
Williams and Mount  (1965) concluded that the higher
periphytic biomasses they observed in Zn-treated channels
were due to this mechanism.  Unfortunately, personnel, time
and equipment were not available for routine water analyses.

Little can be said with respect to consumer densities in
the channels and their effects on periphytic growth.  It
has already been stated that tadpoles were established
in the control channels long before they were observed
in the treated.  Dickman (1968) has shown that tadpoles
enclosed in wooden cages severely reduced the standing
crop of algae in proportion to the number of individuals
present.  Tadpoles may have been partially responsible
for the reduced periphytic levels measured in the control
channels, although this was not the only factor since glass
slides suspended in the channels showed the same trends
as the walls and plexiglass plates, and these glass slides
were not grazed by tadpoles.

From the available information it is impossible to tell
if the ruling influence on the differential development
of algae in the streams was directly due to nutrient
differences, grazing differences, a combination of these
influences or some other factor dependent on Hg concentration.

Macrophytes

The vascular plants that became established in the channels
were  all typical of shallow ponds or very slow-moving water.
                              107

-------
By the end of the study, Juncus diffusissimus and Utricularia
biflora/ were common in all six channels, although their
densities were quite different between channels.  The other
three species of vascular plants found (see Appendix) were
represented by very small numbers of individuals and seemed
to be randomly distributed  among  the  streams.

Early colonization and long-term biomass accumulation in
the channels by vascular plants was dominated by Juncus
diffusissimus, a species of rush of common occurrence in
wet habitats of the southeastern coastal plain of North
America.  This plant was first observed in the streams
nearly a year after water input was begun and by 20
months was well established in all of the channels (Table 3).
It appears that the outside channels (especially channel 1)
received higher inputs of seeds, and also that this plant
was much more common at the heads of the streams than
downstream.  This increased abundance in the upper reaches
of the channels became increasingly evident during the next
20 months until it was necessary to remove some plants
to allow water to flow downstream.

Attempts to quantify the Juncus in the channels were not
made until  February, 1974.  At that time the control
channels were found to average 134 g/m^ ash-free dry weight
and the 1.0-and 5.0-ug/l channels averaged 60.3 and 62.5
g/m^ respectively.  The extremely poor growth of Juncus
immediately downstream from the Hg input tubes showed that
this plant was affected by Hg.  Data collected during the
last year of the project (with no Hg input) are summarized
in Figure 50.  There is no obvious explanation for the
decrease in standing crop biomass of Juncus in the control
channels except that this might be a normal successional
trend, not seen in the treated channels because of the
effects of the Hg.

The second most common plant by the end of the study was
Utricularia biflora which bloomed during the spring of
1974.  This species was not found to be abundant in any of
the streams until after the Hg was shut off.  Since no
differences were observed between the control and treated
channels, this appears to be another successional phenomenon.
The other two species of flowering plants present, Typha
latifolia and Callitriche heterophylla, were not observed
to flower or fruit during the study.

Insects

All of the species of insects collected from the artificial
streams were characteristic of sluggish streams or shallow,
                              108

-------
o
vo
ElSOh
      en
        ++
      en
      CO
      O
      QD
         100
       <  50
       _l
       Q_



            0
                                                       • — CONTROL
                                                       a— i.o/^g/l

FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
DEC
JAN
                                   1974
                                     SAMPLING DATE
                                                                 1975
         FIGURE 50,   TOTAL MACROPHYTE BIOMASS (INCLUDING STEMS AND ROOTS) IN CHANNELS
         DURING THE  FINAL YEAR OF THE PROJECT,

-------
 pond-type habitats of the southeastern United States.  New
 species were continuously discovered in benthic dish samples
 throughout the three years of study indicating that
 colonization continued throughout this period.  After 19
 months of study, only 15 different species had been collected
 (Ferens, 1974).  After 40 months of colonization, these
 original 15 species were still present but rare, and a total
 of  61 species had been encountered.  Fifteen of these were
 first collected when the streams were cleaned out in
 February    1975.

 The original species colonizing the streams were mostly
 midge larvae with one dragonfly, Pantala hymenea, present
 after the first year.  By the end of two years another
 dragonfly, Erythrodiplax minuscula, and a damselfly,
 Ischnura sp., were present in addition to several new
 midge species.  The first beetle larvae and adults were
 collected in January    1974, 29 months after input of
 water was started.  At this same time the first true bugs,
 Hesperocorixa sp., were also collected.  Four months later
 in  May, 1974, the first tricopterans were found and in
 August    1974 the first aquatic lepidopterans were observed.
 As  mentioned earlier, several species, including a helgrammite,
 a waterscorpion, several species of dragonflies and beetles
 and another lepidopteran, were not observed until the streams
 were cleaned out in February  1975.

 During the initial part of the study when input concentrations
 were 0.01 and 1.0 pg/1, no significant differences in mean
 numbers and relative abundances of dominant species of
 midges were found (Ferens, 1974).  Small but significant
 (p  < 0.05)  increases in diversity and evenness values were
 found for downstream samples from the 1.0-yg/l streams as
 compared to the controls; however, diversity values were
 very low during this time (about 1.0) because of the
 initially slow rate of colonization by insects.

Benthic dish data collected after the Hg inputs to channels
 3 and 6 were increased to 5.0 yg/1 in August, until the Hg
was shut off six months later, indicated that insect
densities were greatest in the 1.0-pg/l streams and very
similar between the controls and the 5.0-yg/l streams
 (Figure 51).  Fluctuations in these densities were observed
to be similar between treatments for the year following the
Hg  shut-off, but statistical analysis showed that the
density of insects was significantly lower  (ANOVA, p  < 0.01)
in  the previously treated streams.  Diversity and evenness
values for the benthic dish samples collected during  this
last year (Figure 52) were found to be significantly  higher
                              110

-------
        Hg INPUT OFF
CVJ
 c>
*^>^*^xv^,,^xlr*^5*'^'^
CONTROL!!
           '"'"
 H

 CO
 I-
 O
 LU

 CO
     o
D
J|F
M
A
M
J
J
A
S
0
N
n
J
         1973
                   1974


             SAMPLING  DATE
1975
   FIGURE 51,   DENSITIES OF INSECTS COLLECTED FROM BENTHIC DISHES  BEFORE
   AND AFTER MERCURY INPUTS WERE STOPPED,

-------
    1.0

co 0.9
UJ
z 0.8

g 0.7

   0.6
 Ld
    0.5
                Hg INPUT OFF
 CD
 O
Q
 i
IX
    2.5
t  1.5
co
o:
LU
    1.0
                                                        —CONTROL
          D
M
M
A
0
N
        1973
                                   1974

                            SAMPLING  DATE
D
                                           1975
 FIGURE
 BEFORE
        52,   DIVERSITY AND EVENNESS OF INSECTS COLLECTED FROM BENTHIC DISHES
        AND AFTER MERCURY INPUTS WERE STOPPED,

-------
in the control streams than in  the  5.0  yg/1  channels
(Students t, p. < 0.05).  This trend of  lower diversity
values for higher treatments was almost gone at the end
of the summer after the Hg had  been shut off.

Using ANOVA methods, three species  of insects were found
to be negatively affected by treatment  levels in the
streams.  These were a damselfly, Argia sp.,  that reached
considerable densities by the end of the study;  a tiny
mayfly, Callibaetis sp.; and one midge  species,
Paralauterborniella elachisa.   Another  mayfly,  Caenis sp.,
was not collected in channels 3 or  6 during  the  input of
Hg and this may have been related to the treatment.
Three species of chironomids  (Pseudochironomus  sp.,
Ablabesmyia ornata, and Psectrocladius  sp.)  and  one
ceratopogonid (Dasyhelea sp.) were  most abundant in the
1.0-yg/l streams.

When the water was turned off in February    1975, an
attempt was made to remove the  entire populations of the
larger insects, amphibians, molluscs, and mosquitofish in
the streams.  The water level in the channels was lowered
to about two inches above the sand.  Starting at the tails,
an 8.5-m stream section was enclosed with two screens, and
ten passes were made with a 3-mm mesh .dip net.   The lower
screen was then moved up 17 m thereby enclosing  the next
section.  This was repeated until the entire 100  m length
of each stream had been sampled.  The material collected
by this method was carefully examined in the laboratory
and all animals were removed, sorted and counted.
Representative organisms from each  group were frozen for
Hg analyses.  Table 4 summarizes the data from this final
sampling.  The insect densities shown in Table 4  are much
lower than those calculated from benthic dish data (Figure
51).  This is because the sampling  technique used did not
collect midge larvae, the most  common benthic insects.
Table 4 shows that the greatest densities of larger insects
occurred in channels 2 and 5, the 1.0-yg/l treatments, and
the lowest densities in channels 1  and  6  (the first a
control, the second a 5.0-yg/1  treatment). There  appears to
be no relationship between treatment levels and diversity,
density or evenness values.

Of the parameters measured for  insects,  diversity values
appeared to be most sensitive to Hg  levels with an initial
increase in diversity for the 1.0-yg/l  streams and an
eventual lowering of the diversities corresponding to
higher treatment levels.   No significant differences were
seen for insect densities while the  Hg was on.   In general
                              113

-------
the streams appeared to be variable with respect to insect
numbers in a manner unrelated to treatment levels.  This
variation    was probably the result of the finite number
of non-random visits by egg-laying adults.  None of the
common species of insects occurring in the streams
demonstrated a conclusive direct or indirect reaction to
Hg levels, although possible inhibitory effects were seen
for four less-abundant species.

Warnick and Bell (1969), studying the acute toxicity of
various metals to three insect species, found Hg in the
inorganic form to be lethal at 2.0 mg/1 for all of the
species tested.  Of the three species, the mayfly was
found to be most sensitive to the various metals.  In
our streams both mayfly species present showed a possible
inhibitory response to a Hg concentration as low as 5.0
pg/1; however, this highest concentration tested has a
very small effect on the total insect community.
                              114

-------
                        SECTION VII

                        REFERENCES

Alberts, J. J., J.  E.  Schindler,  and R.  W.  Miller.   1974.
     Elemental mercury evolution  mediated by humic  acid.
     Science  184:   895-896.

Albright, L.  J., J. W.  Wentworth,  and E.  M.  Wilson.   1972.
     Technique for  measuring  metallic salt  effects  upon
     the indigenous heterotrophic microflora of  a natural
     water.   Water  Res.  6:  1589-1596.

American Public Health Association.   1965.   Standard
     Methods  for the Examination  of  Water and Wastewater,
     12th ed. American Public Health Association, New
     York.  769 p.

Bache, C. A., W. H. Gutenmann,  and D.  J.  Lisk.   1971.
     Residues of total mercury and methylmercuric salts
     in lake  trout  as  a function  of  age.  Science 172:
     951-952.

Beck, W. M., Jr.  1968.  Chironomidae.   In:   Keys to
     Water Quality  Indicative Organisms  of  the South
     Eastern United States.   Fed.  Water  Pollut.  Con.
     Adm., Atlanta, p.  V 1-22.

Ben-Bassat, D. , G.  Shelef, N.  Gruner,  and H.  I.  Shuval.
     1972.  Growth  of  Chlamydomonas  in a  medium  containing
     mercury.  Nature  240:  43-44.

Beyers, R. J.  1963.  A  characteristic diurnal metabolic
     pattern in balanced microcosms.   Publ.  Inst. Mar.
     Sci. Univ. Tex. 9:  19-27.

Beyers, R. J.  1965.  The pattern  of photosynthesis and
     respiration in laboratory microecosystems.  Mem.
     1st. Ital. Idrobiol. 18  Suppl.:   61-74.

Beyers, R. J.  1970.  A  pH-carbon  dioxide method for
     measuring aquatic primary productivity.  Bull. Acad.
     Sci. 28:  55-68.

Birge,  W. J., J.  J. Just, A. Westerman, and  A. D. Rose.
     1974.   Sensitivity  of Vertebrate Embryos to Heavy
     Metals as a Criterion of Water Quality  Phase I.
     University of Kentucky Water  Resources  Research
     Institute'.   Lexington, Kentucky.  Report No. 71.


                             115

-------
Bischoff, H., and H. C. Bold.  1963.  Phycological studies
     IV.  Some soil algae from Enchanted Rock and related
     algal species.  Univ. Publ. No. 6318.  Austin.

Bisogni, J. J., and A. W. Lawrence.  1975.  Kinetics of
     mercury methylation in aerobic aquatic environments.
     J. Water Pollut. Control Fed. 47:  135-152.

Blair, W. F., A. P. Blair, P. Brodkorb, F. R. Cagle, and
     G. A. Moore.  1957.  Vertebrates of the United States.
     McGraw-Hill Book Co., New York.  819 p.

Bothner, M. H., and R. Carpenter.  1974,  The rate of
     mercury loss from contaminated estuarine sediments
     in Bellingham Bay, Washington.  Proceedings of 1st
     Annual NSF Trace Contaminants Conference.  Oak Ridge
     Nat. Lab. p. 198-210.

Braman, R. J., and D. L. Johnson.  1973.  Selective
     absorption tubes and an emission technique for the
     determination of ambient forms of mercury in air.
     Unpubl. paper presented at 166th Annual Meeting of
     the Amer. Chem. Soc. 31 p.

Burkett, R. D.  1975.  Uptake and release of methylmercury-
     203 by Cladophora glomerata.  J. Phycol. 11:  55-59.

Coleman.  1970.  Operating Directions for Coleman Mercury
     Analyzer MAS-50.  #50-900.  12 p.

Curtis, E. H.  1974.  The Influence of Temperature and
     Methyl Mercury Chloride Concentration on the Direct
     Uptake of Methyl Mercury Chloride from Water in
     Bluegill Sunfish  (Lepomis macrochirus).  Ph.D.
     dissertation.  Northwestern Univ., Evanston, 111.
     188 p.

Dickman, M.  1968.  The effect of grazing by  tadpoles
     on the structure of a periphyton community.
     Ecology 49:  1188-1190.

Dietz, F.  1972.  The enrichment of heavy metals  in
     submerged plants.  Proc. of the 6th  Internat. Water
     Pollut. Res. p. A/2/4/1-8.

D'ltri, F. M.  1972.  The Environmental Mercury Problem.
     CRC Press, Cleveland, Ohio.   124 p.

Edmondson, W. T., ed.  1959.  Fresh-Water Biology.   John
     Wiley & Sons, Inc., New York.   1248  p.
                             116

-------
Environmental Protection Agency  (EPA).  1974.  Methods for
     Chemical Analysis of Water  and Wastes.  Methods
     Development and Quality Assurance Research Lab.
     EPA-625/6-74-003.  298 p.

Fagerstrom, T. , and A. Jernelov.  1972.  Some aspects of
     the quantitative ecology of mercury.  Water Res.
     6:  1193-1202.

Ferens, M. C.  1974.  The Impact of Mercuric Ions on
     Benthos and Periphyton of   Artificial Streams.
     Ph.D. dissertation.  Univ.  Ga. , Athens, Ga.  100 p.

Fimreite, N. , W. N. Holsworth, J. A. Keith, P. A. Pearce,
     and I. M. Gruchy.  1971.  Mercury in fish-eating
     birds near sites of industrial contamination in
     Canada.  Can. Field-Nat. 85:  211-220.

Friberg, L. T. , and J. J. Vostal.  1972.  Mercury in the
     Environment.  CRC Press, Cleveland, Ohio.  215 p.

Fujita, M. , and K. Hashizume.  1972.  The accumulation
     of mercury by freshwater planktonic diatom.
     Chemosphere 5:  203-207.

Furukawa, K. , and K. Tonomura.   1972.  Metallic mercury
     releasing enzyme in mercury resistant Pseudomonas .
     Agr. Biol. Chem. 36:  217-226.
Glooschenko, W. A.  1969.  Accumulation of ^u°Hg by the
     marine diatom Chaetoceros costatum.  J. Phycol. 5:
     224-226.

Hannan, P. J. , and C. Pautouillet.  1972.  Effect of
     mercury on algal growth rates.  Biotechnol. Bioeng.
     14:  93-101.

Hannerz, L.  1968.  Experimental investigations on the
     accumulation of mercury in water organisms.  Inst.
     Freshwater Res. Drottingholm Rep. 48:  120-176.

Harriss, R. C. , D. B. White, and R. B. MacFarlane.  1970.
     Mercury compounds reduce photosynthesis by plankton.
     Science 170:  736-737.

Hartung, R. , and B.  D. Dinman, eds.  1972.  Environmental
     Mercury Contamination.  Ann Arbor Science Publishers
     Inc., Ann Arbor, Mich.  349 p.
                             117

-------
Holm, H. W. and M. F. Cox.  1974.  Mercury in Aquatic Systems;
     Methylation, Oxidation-Reduction, and Bioaccumulation.
     Natl. Environ. Res. Center/ Corvallis, Ore., EPA-660/3-
     74-021.  38 p.

Huber-Pestalozzi, G.  1938.  Das Phytoplankton des
     Susswassers:  Systematik und Biologie.  Allgemeiner
     Teil.  Blaualgen, Bukterien, Pilze.  Binnengewasser
     16, Teil 1.  342 p.

Huber -Pestalozzi, G.  1941.  Das Phytoplankton des
     Susswassers:  Systematik und Biologie.  Chrysophyceen,
     Farblose Flagellaten, Heterokonten.  Binnengewasser,
     16, Teil 2, Halfte.  365 p.

Huber-Pestalozzi, G.  1955.  Das Phytoplankton des
     Susswassers:  Systematik und Biologie.  Euglenophyceen.
     Binnengewasser, 16, Teil 4, ix.  606 p.

Huckabee, J. W., and B. G. Blaylock.  1973.  Transfer of
     mercury and cadmium from terrestrial to aquatic
     ecosystems.  In:  Metal Ions in Biological Systems,
     Dhar  (ed.), New York, N.Y., Plenum Publ. Corp.
     p. 125-160.

Jenne, E. A.  1972.  Mercury in Waters of the United States:
     1970-71.  U. S. Dept. of the Interior, Geological
     Survey, Water Resources Div., Menlo Park, Calif.  34 p.

Jernelov, A., and H. Lann.  1973.  Studies in Sweden on
     feasibility of some methods for restoration of
     mercury-contaminated bodies of water.  Environ. Sci.
     Tech. 7:  712-718.

Johnels, A. G., T. Westermark, W. Berg, P. I. Persson, and
     B. Sjostrand.  1967.  Pike  (Esox lucius L.) and some
     other aquatic organisms in Sweden as indicators of
     mercury contamination in the environment.  Oikos 18:
     323-333.

Johnels, A. G., and T. Westermark.  1970.  Mercury
     contamination of the environment in Sweden.  In:
     Chemical Fallout, Miller, M. W. and G. C. Berg  (eds.),
     Charles C. Thomas Publisher, Springfield, 111.
     p. 221-241.

Jones, H. R.  1971.  Mercury Pollution Control.  Noyes Dat
     Corp., Park Ridge, N.Y.  251 p.
                              118

-------
Kamps, L. R., R. Carr, and H. Miller.   1972.  Total
     mercury-monomethylmercury content  of  several species
     of fish.  Bull. Environ. Conta. Toxicol. 8:  273-279.

Kania, H. J., and R. J. Beyers.   1974.  NTA and Mercury in
     Artificial Stream Systems.   EPA-660/3-73-025.  25 p.

Kelso, J. R. M., and R. Frank.   1974.   Organochlorine
     residues, mercury, copper,  and cadmium in yellow
     perch, white bass and smallmouth bass, Long Point
     Bay, Lake Erie.  Trans. Am.  Fish.  Soc. 103:  577-581.

Knauer, G. A., and H. H. Martin.   1972.  Mercury in a
     marine pelagic food chain.   Limnol. Oceanog. 17:
     868-876.

Knight, L. A., Jr., and J. Herring.  1972.  Total mercury
     in largemouth bass  (Micropterus salmoides) in Ross
     Barnett Reservoir, Mississippi - 1970 and 1971.
     Pestic. Monit. J. 6:  103-106.

Lepple, F. K.  1973.  Mercury in the Environment.  A Global
     Review Including Recent Studies in the Delaware Bay
     Region.  Univ. of Delaware  Publ. DEL-SG-8-73.  Newark,
     Delaware.  75 p.

Longbottom, J. E. , R. C. Dressman, and J. J. Lichtenberg.
     1972.  The Determination of Methyl Mercury in Fish,
     Sediment, and Water.  Off.  Res. Monitor., Nat.
     Environ. Res. Center, EPA,  Cincinnati, Ohio.  33 p.

Mason, W. T.  1973.  An Introduction to the Identification
     of Chironomid Larvae.  EPA,  Cincinnati, Ohio.  90 p.

Matida, Y. , and H. Kumada.  1969.  Distribution of mercury
     in water, bottom mud, and aquatic organisms of
     Minamata Bay, the River Agano, and other water bodies
     in Japan.  Bull. Freshwater  Fish. Res. Lab. Tokyo
     19:  73-93.

Matsui, S., and E. F. Gloyna.  1972.  Radioactivity Transport
     in Water-Biological Accumulation of Mercuric and
     Methyl Mercury by Chlorella  pyrenoidosa.  Centr. Res.
     Water Resources., Austin, Tex., AEC Tech. Report #22.
     251 p.

Matsumura, F., Y. Gotoh, and G. M. Boush.  1972.  Factors
     influencing translocation and transformation of
     mercury in river sediment.   Bull. Environ. Contain.
     Toxicol. 8:   267-272.
                              119

-------
Nelson, N.  1971.  Hazards of mercury.  Environ. Res. 4:
     1-69.

Nuzzi/ R.  1972.  Toxicity of mercury to phytoplankton.
     Nature 237:  38-40.

Odum, H. T.  1956.  Primary production in flowing waters.
     Limnol. Oceanogr. 1:  102-116.

Olley, J.  1973.  Mercury in fish and the news media.  Food
     Technol. Aust. 25:  270.

Parks, G. A.  1973.  Trace Elements in Water:  Origin, Fate,
     and Control:  1.  Mercury.  Report of Progress, March 1,
     1972 to Feb. 1, 1973.  NSF, Stanford Univ., Stanford,
     Calif. 247 R.

Patrick, R., and C. W. Reimer.  1966.  The Diatoms of the
     United States:  Exclusive of Alaska and Hawaii.  Vol. I,
     Fragilariaceae, Eunotiaceae, Achnanthatceae, Naviculaceae.
     The Acad. Natr. Sci. Phila., Philadelphia, Penn.  688 p.
                                            s
Peakall, D. B., and R. J. Lovett.  1972.  Mercury:  its
     occurrence and effects in the ecosystem.  Bioscience
     22:  20-25.                                   :  '

Pennak, R. W.  1953.  Fresh-Water Invertebrates of the
     United States.  Ronald Press Co., New York.  769 p.

Pielou, E. C.  1969.  An Introduction to Mathematical
     Ecology.  Wiley-Interscience, New York.  286 p.

Prescott, G. W.  1962.  Algae of the Western Great Lakes
     Area.  Wm. C. Brown Co., Dubuque.  977 p.

Prescott, G. W.  1970.  How to Know the Fresh-Water Algae.
     Wm. C. Brown Co., Dubuque.  348 p.

Pringsheim, E. G.  1946.  Pure Cultures of Algae.  Cambridge
     Univ. Press, Cambridge.  119 p.

Putman, J. J.  1972.  Quicksilver and slow death.  Natl.
     Geogr. Mag. 142:  507-527.

Radford, A. E., H. E. Ahles, and C. R. Bell.  1968.   Manual
     of the Vascular Flora of the Carolinas.  Univ. N.  C.
     Press, Chapel Hill.  1183 p.

Ralfs, J.  1848.  The British Desmidieae.  Reeve, Benham,  and
     Reeve, London.  226 p.
                               120

-------
Ramanathan, K. R.  1964.  Ulotrichales.   Indian Council of
     Agr. Res., New Delhi.  188 p.

Saha, J. G.  1972.  Significance of mercury in the environment.
     Residue Rev. 42:  103-163.

Schoper, N. J.  1974.  The Uptake, Biotransformation, and
     Elimination of Elemental Mercury by  Fish.  M. S. thesis,
     Univ. Ga., Athens.   36 p.

Scott, D. P.,  and F. A. J. Armstrong.  1972.  Mercury
     concentration in relation to size in several species
     of freshwater fishes from Manitoba and Northwestern
     Ontario.  J. Fish. Res. Board Can. 29:  1685-1690.

Service, J.  1972.  A User's Guide to the Statistical
     Analysis  System.  N. C. State Univ., Raleigh.  260 p.

Skuja, H.  1956.  Taxonomische und biologische studien uber
     das phytoplankton Schwedischer binnengewasser.  Nova
     Acta Regiae Societatis Scientiarum Upsaliensis.  Ser.
     IV.  16:  1-404.

Sladecek, V.,  and A. Sladeckova.  1963.   Limnological study
     of the reservoir Sedlice near Zeliv. XXIII.  Periphyton
     production.  Technology of Water 7:  77-133.

Smith, G. M.   1950.  The Fresh-Water Algae of the United
     States, 2nd ed.  McGraw-Hill Book Co., New York, N. Y.
     719 p.

Spangler, W. J., J. L. Spigarelli, J. M.  Rose, and H. M.
     Miller.   1973.  Methylmercury:  bacteria degradation
     in lake sediments.  Science 180:  192-193.

Suzuki, T., K. Furukawa, and K. Tonomura.  1968.  Studies
     on the removal of inorganic mercurial.  J. Ferment.
     Technol.  46:  1048-1055.

Tiffany, L. H. , and M. E. Britton.  1952.  The Algae of
     Illinois.  Hafner Publ. Co., New York.  407 p.

Toribara, T. Y., C. P. Shields, and L. Koval.  1970.
     Behavior  of dilute solutions of mercury.  Talanta
     17:  1025-1028.

Ukeles, R.  1962.  Growth of pure cultures of marine
     phytoplankton in the presence of toxicants.  Appl.
     Microbiol. 10:  532-537.
                              121

-------
Wallace, R. A., W. Fulkerson, W. D. Shults, and W. S. Lyon.
     1971.  Mercury in the Environment:  The Human Element.
     Oak Ridge Nat. Lab. Pub.  ORNL NSF-EP-1.  Oak Ridge,
     Tennessee.  61 p.

Warnick, S. L., and H. L. Bell.  1969.  The acute toxicity
     of some heavy metals to different species of aquatic
     insects.  J. Water Pollut. Control Fed. 41:  280-284.

Warren, C. E., and G. E. Davis.  1971.  Laboratory stream
     research:  objectives, possibilities, and constraints.
     Ann. Rev. Ecol. Syst. 2:  111-144.

Wershaw, R. L.  1960.  Sources and behavior of mercury in
     surface waters.  In:  Mercury in the Environment,
     Geological Survey Professional Paper 713.  p. 29-31.

Whitford, L. A., and G. J. -Schumacher.  1969.  A Manual of
     the Fresh-Water Algae in North  Carolina.  N. C.
     Agr. Exp. Station, Raleigh, 313 p.

Williams, L. G., and D. I. Mount.  1965.  Influence of
     zinc on periphytic communities.  Am. J. Bot. 52:
     26-34.

Wood, J. M.  1974.  Biological cycles for toxic elements
     in the environment.  Science 183:  1049-1052.
                             122

-------
                        SECTION VIII

                         APPENDIX

BIOTA OBSERVED IN THE ARTIFICIAL STREAMS

Phylum Chlorophyta
  Sub-Phylum Chlorophyceae
    Order Volvocales
      Family Chlamydomonadaceae
        Chlamydomonas sp.
    Order Tetrasporales
      Family Gloeocystaceae
        Gloeocystis vesiculosa Naegeli
      Family Chaetochloridaceae
        Porochloris filamentorum Pascher
    Order Chlorococcales
      Family Chlorococcaceae
        Chlorococcum humicola (Naegeli)  Rabenhorst
        Desmatractum bipyramidatum (Chodat)  Pascher
      Family Oocystaceae
        Eremosphaera viridis DeBary
        Oocystis sp.
      Family Scenedesmaceae
        Scenedesmus acutiformis Schroeder
      Family Coccomyxaceae
        Dispora crucigenioides Printz
    Order Ulotrichales
      Family Ulotrichaceae
        Hormidium subtile (Koetzing)  Heering
        Geminella turfosa (Skuja)  Ramanathan
      Family Microsporaceae
        Microspora pachyderma (Wille)  Lagerheim
        M.  quadrata Hazen
    Order Chaetophorales
      Family Chaetophoraceae
        Microthamnion strictissimum Rabenhorst
        Stigeoclonium elongatum (Hassall)  Kuetzing
    Order Oedogoniales
      Family Oedogoniaceae
        Oedogonium sp. (probably 0.  pusillum Kirchner)
    Order Zygnematales
      Family Zygnemataceae
        Mougeotia sp.  (5 micrometer width)
        Mougeotfa sp.  (10 micrometer width)
        Spirogyra sp.
      Family Mesotaeniaceae
        Cylindrocystis brebissonii var.  minor West and West
        Mesotaenium macrococcum (Kuetzing) West  and West
                             123

-------
      Family Desmidiaceae
        Arthrodesmus incus (Brebisson)  Hassall
        Cosmarium asphaerosporum Nordstedt
        C. pseudoconnatum var. ornatum Allorge
        C. laeve var. septentrionale Wille
        C. undulatum var. minutum Wittrock
        C. viride var. minor West
        Penium spinospermum Joshua
        Shaerozosma excavata Ralfs
        Spondylosium planum West and West

Phylum Euglenophyta
  Order Euglenales
    Family Euglenaceae
      Euglena mutabilis Schmitz

Phylum Pyrrophyta
  Class Dinophyceae
    Order Dinokontae
      Family Glenodiniaceae
        Glenodinium sp.

Phylum Chrysophyta
  Sub-Phylum Xanthophyceae
    Order Rhizochloridales
      Family Rhizochloridaceae
        Stipitococcus vasiformis Tiffany
    Order Mischococcales
      Family Pleurochloridaceae
        Chlorocloster minimus Pascher
      Family Sciadaceae
        Ophiocytium deserturn var. minor Prescott
  Sub-Phylum Chrysophyceae
    Order Rhizochrysidales
      Family Rhizochrysidaceae
        Rhizochrysis sp.
      Family Stylococcaceae
        Lagynion macrotrachelum  (Stokes) Pascher
    Order Chromulinales
      Family Chromulinaceae
        Chromulina pseudonebulosa Pascher
    Order Ochromonadales
      Family Ochromonadaceae
        Ochromonas sp.
      Family Dinobryaceae
        Epipyxis sp.
  Sub-Phylum Bacillariophyceae
    Order Pennales
      Family Eunotiaceae
        Eunotia tenella  (Grun.) Cleve
                             124

-------
      Family Naviculaceae
        Navicula notha Wallace
        Pinnularia abaugensis  (Pant.) Ross
      Family Cymbellaceae
        Cymbella sp.

Phyltim Cyanophyta
  Order Chroococcales
    Family Chroococcaceae
      Merismopedia punctata Meyer
  Order Oscillatoriales
    Family Oscillatoriaceae
      Osciliatoria geminata Meneghini
      0. meslini Fremy
      P.- splendida Greville
  Order Nostocales
    Family Nostocaceae
      Anabaenopsis sp.
    Family Rivulariaceae
      Calothrix parietina  (Naegeli) Thuret

Phylum Spermophyta
  Class Monocotyledoneae
    Order Alismatales
      Family Typhaceae
        Typha latifolia Linnaeus
    Order Liliales
      Family Juncaceae
        Juneus diffusissimus Buckley
    Order Graminales
      Family Poaceae
        Panicum sp.
  Class Dicotyledoneae
    Order Geraniales
      Family Callitrichaceae
        Callitriche heterophylla Pursh
    Order Scrophulariales
      Family Lentibulariaceae
        Utricularia biflora Lamarck

Phylum Arthropoda
  Class Arachnoidea
    Order Araneae
      Family Pisauridae
        Dolimedes sexpunctatus Hentz
      Family Lycosidae
        Lycosa helluo Walckenaer
  Class Insecta
    Order Odonata
      Family Coenagrionidae
                             125

-------
    Argia sp.
    Ischnura sp.
  Family Libellulidae
    Celithemis fasciata
    Epicordulia sp.
    Erythrodxplax minus'cula
    Ladona deplanta
    Pacydiplax longipennis
    Pantala. hymenea
    Plathemis lydia
    Tetragoneuria semiaquea^
  Family Aeschnidae
    Anax junius
Order Coleoptera
  Family Dytiscidae
    Agabus spp.
    Coptotomus sp.
    Hydrocanthus sp.
    Kydroporus spp.
    Hydrovatus sp.
    Laccophilus sp.
    Suphisellus sp.
  Family Hydrophilidae
    Berosus spp.
    Enochrus spp.
    Tropisternus spp.
  Family Gyrinidae
    Gyrinus sp.
Order Ephemeroptera
  Family Baetidae
    Caenis sp.
    Callibaetis sp.
Order Trichoptera
  Family Leptoceridae
    Oecetis sp.
    Triaenodes sp.
Order Hemiptera
  Family Veliidae
    Microvelia sp.
    Velia sp.
  Family Notonectidae
    Notonecta indica
  Family Navcoridae
    Pelocoris femoratus
  Family Gerridae
    Gerris sp.
  Family Nepidae
    Ranatra sp.
  Family Corixidae
    Hesperocorixa spp.
                         126

-------
    Order Lepidoptera
      Family Pyralidae
        Parapoynx sp.
        Synclita sp.
    Order Megaloptera
      Family Corydalidae
        Chauliodes sp.
    Order Diptera
      Family Tipulidae
        Tipula sp.
        Helius sp.
      Family Chironomidae
        Subfamily Chironominae
          Chironomus spp.
          Cladotanytarsus sp.
          Cryptochironomus sp.
          Dicrotendipes sp.
          Microtendipes sp.
          Parachironomus sp.
          Paralauterborniella  elachista
          P. nigrohalterale
          Polypedilum sp.
          Pseudochironomus sp.
          Rheotanytarsus sp.
          Tanytarsus sp.
        Subfamily Tanypodinae
          Ablabesmyia ornata
          A. peleensis
          Larsia sp.
          Zavrelimyia sp.
        Subfamily Orthocladiinae
          Cardiocladius
          Cricotopus sp.
          Psectrocladius sp.
      Family Ceratopogonidae
        Subfamily Dasyheleinae
          Dasyhelea sp.
        Subfamily Heleinae
          Alluandomyia sp.
          Palpomyia tibialis

Phylum Mollusca
  Class Gastropoda
    Family Lymnaeidae
      Lymnea sp.
    Family Physidae
      Physa sp.
                             127

-------
Phylum Chordata
  Class Teleostorai
    Order Cyprinodontiformes
      Family Poeciliidae
        Gambusia affinis (Baird and Girard)
  Class Amphibia
    Order Anura
      Family Bufonidae
        Bufo terrestris (Bonnaterre)
      Family Ranidae
        Rana catesbeiana Shaw
        R. pipiens Schreber
  Class Aves
    Order Coraciiformes
      Family Alcedinidae
        Megaceryle aIcyon (Linnaeus)
  Class Mammalia
    Order Carnivora
      Family Procyonidae
        Procyon lotor (Linnaeus)
                              128

-------
                              TECHNICAL REPORT DATA
                       (Please read Instructions on the reverse before completing)
 . REPORT NO.
 EPA-600/3-76-060
           3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
 Fate  and Biological  Effects of Mercury
 Introduced into Artificial Streams
                                                  5. REPORT DATE
                                                   August 1976  (Issuing date
           6. PERFORMING ORGANIZATION CODE
 J. AUTHOR(S)
 Henry J.  Kania, Rbbert L.  Knight, and Robert
 J.  Beyers
           8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Savannah River Ecology  Laboratory
 Institute of Ecology
 University of Georgia
 Athens,  GA  30601
           10. PROGRAM ELEMENT NO.
              1BA023
           11. CONTRACT/GRANT NO.

              R800510
 12. SPONSORING AGENCY NAME AND ADDRESS
 Environmental Research Laboratory
 Office of Research  and Development
 U.S.  Environmental  Protection Agency
 Athens,  GA  30601
           13. TYPE OF REPORT AND PERIOD COVERED
             Final report	
           14. SPONSORING AGENCY CQDE
             EPA/ORD
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
     Mercuric ion was  continuously input to artificial stream channels to
 provide water concentrations of 0.01,  1.0, and 5.0 yg/1.  Channel com-
 ponents were periodically sampled for  total mercury analyses.   The
 effects of mercury on the algal components of the periphyton communities
 and on the benthic insects were determined.  The sampling program con-
 tinued one full year  after mercury inputs were stopped.
     Approximately 15% of the added mercury was removed from  the water.
 The community components acquired very high concentrations of mercury,
 although in most cases the levels in these were not a linear function of
 the water levels.  Concentrations in invertebrates decreased most
 rapidly after mercury inputs were stopped while the sediment levels
 decreased most slowly.
     Periphytic algae  showed several treatment responses including total
 inhibition,  inhibition of certain life stages, possible stimulation of
 certain life stages and morphological  alterations.  One species was
 found  only in the treated channels and disappeared when mercury inputs
 were stopped.
    The diversity and  evenness of benthic inject communities were
 affected  by  mercury treatment.	i	
                           KEY WORDS AND DOCUMENT ANALYSIS
               DESCRIPTORS
                                       b.lDENTIFIERS/OPEN ENDED TERMS
                         COS AT I Field/Group
Streams, Aquatic animals, Food chains
Purification,  Water pollution,
Aquatic microbiology-algae,  sediments
invertebrates, Fishes
 Heavy  metals,
 Mercury,  mercuric io
 Elemental mercury,
 Microcosms,  Fate,
 Transformations,
 Fish Food Organisms
 Periphyton
                                                                  6F
 8. DISTRIBUTION STATEMENT

"formal distribution
19. SECURITY CLASS (This Report/
 UNCLASSIFIED
                                                              21. .NO
141
20. SECURITY CLASS (Thispage)
 UNCLASSIFIED
                       22. PRICE
EPA Form 2220-1 (9-73)
                                      129
            *U.S. GOVERNMENT PRINTING OFFICE: 1976-657-695/5497

-------