Ecological Research Series
TOXIC EFFECT OF WATER SOLUBLE POLLUTANTS
                             ON FRESHWATER  FISH
                                     Environmental Research Laboratory
                                    Office of Research and Development
                                    U.S. Environmental Protection Agency
                                         Duluth, Minnesota 55804

-------
                RESEARCH REPORTING SERIES

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

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

This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials. Problems  are assessed for their long- and short-term influ-
ences. Investigations include formation, transport, and pathway studies to deter-
mine the fate of pollutants and their effects. This work provides the  technical basis
for setting standards to minimize undesirable changes in living organisms in the
aquatic, terrestrial, and atmospheric  environments.
 This document is available to the public through the National Technical Informa-
 tion Service, Springfield, Virginia 221.61.

-------
                                           EPA-600/3-77-057
                                           May 1977
TOXIC EFFECT OF WATER SOLUBLE POLLUTANTS

           ON FRESHWATER FISH
                   by
              Paul 0. Fromm
        Department of Physiology
        Michigan State University
      East Lansing, Michigan 48824
          Grant Number R-801034
             Project officer

             James M. McKim
Environmental Research Laboratory-Duluth
        Duluth, Minnesota 55804
ENVIRONMENTAL RESEARCH LABORATORY-DULUTH
   OFFICE OF RESEARCH AND DEVELOPMENT
  U.S. ENVIRONMENTAL PROTECTION AGENCY
         DULUTH,  MINNESOTA 55804

-------
                                DISCLAIMER


     This report has been reviewed by the Environmental Research Labor-
atory-Duluth,  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 consti
tute endorsement or recommendation for use.
                                   ii

-------
                                  FOREWORD
     Our nation's freshwaters are vital for all animals and plants, yet our
diverse uses of water — for recreation, food, energy, transportation,  and
industry — physically and chemically alter lakes, rivers,  and streams.  Such
alterations threaten terrestrial organisms, as well as those living in  water.
The Environmental Research Laboratory in Duluth, Minnesota  develops methods,
conducts laboratory and field studies, and extrapolates research findings

     —to determine how physical and chemical pollution affects aquatic
       life,

     —to assess the effects of ecosystems on pollutants,

     —to predict effects of pollutants on large lakes through use of
       models,  and

     —to measure bioaccumulation of pollutants in aquatic  organisms that
       are consumed by other animals, including man.

     While the  results of this project do not provide "numbers" for direct
agency use,  the study is no less important to EPA.  Only by understanding
modes of action and physiological pathways can aquatic toxicologists predict
effects of pollutants without long and expensive testing.   Otherwise our
data produced are applicable only to the conditions studied.

     This report moves us another step closer to better predictions of
toxicity.
                                      Donald  I.  Mount,  Ph.D.
                                      Director
                                      Environmental  Research  Laboratory
                                      Duluth, Minnesota
                                     ill

-------
                                  ABSTRACT
     Studies of the effect of inorganic and organic mercury on trout indicated
that uptake was primarily via the gills in non-feeding fish.  Organic, mercury
entered fish at a faster rate than inorganic mercury.  Exposure of trout to
10 yg Hg/1 (methyl forir.) had no effect on the gill oxygen consumption measured
in vitro or on the plasma electrolytes.  The hematocrit index increased
significantly.  Studies of the metabolism of iron hy normal and iron deficient
trout (made deficient by bleeding) indicated that the liver, spleen, and head
kidney are the major iron storage organs.   Liver iron was reduced by bleeding
whereas splenic iron was unaffected.   In iron deficient fish more radioiron
appeared in erythrocytes than in normal controls.  Studies of isolated-perfused
gills revealed the presence of both a and 3 adrenergic receptors and the data
obtained indicate the functional surface area of trout gills can be regulated
by changes in perfusion pathway through the gills.  Use of perfused gills
appears to be a very sensitive model  to detect deleterious action of pollutants
on fish.  Evaluation of heat exchange in perfused gills indicates that the
presence of epinephrine increased the transfer maximum of the gill but they
were unaffected by the administration of acetylcholine.  Analysis of a simple
model indicated that the gills may account for as much as 60% of the total
heat exchange by trout.
                                     iv

-------
                                  CONTENTS

Foreword	    iii
Abstract	„. •, 	     iv
Figures	,	,	,	     vi
Tables	„	„	   viii
Acknowledgments	     ix

   1.  Introduction	      1
   2.  Conclusions	      2
   3.  Recommenda t ions....,	»...      4
   4.  Effects of methyl  mercuric chloride and mercuric chloride on
         rainbow trout	      5
          A study of uptake pathways into whole fish	      5
          Uptake of mercury by trout erythrocytes in vitro	      7
          Mercury uptake  and ion distribution in gills of
             rainbow trout	„	     10
          Effect of methyl  mercury on gill metabolism and blood
             parameters of  rainbow trout	     12
          Effect of mercury exposure on rainbow trout gill
             ultrastructure	•.	     14
   5.  Iron metabolism	„	     17
          Metabolism of iron by normal and iron deficient
             rainbow trout	     17
   6.  Respiratory physiology	„	     24
          Control mechanisms for regulating the functional
             surface area of fish gills	„	     24
          The role of fish  gills in heat  exchange	     32
          The vascularity of trout gills	     44

References	„	„     48
Publications	     51
Index	     52

-------
                                   FIGURES

Number                                                                 Page

  1   Percent uptake of methyl mercury or divalent mercury by
        red blood cells as a function of time	     9
  2   Three component exponential curve of iron clearance from
        plasma of rainbow trout following i.p.  injection of ->9pe	    19

  3   Specific activity of iron in tissues of trout following
        intraperitoneal injection of ^'Fe	    21

  4   Specific activity of iron in tissues of iron deficient
        trout following intraper itoneal injection of -*'Fe	    21

  5   Diagram of gill perfusion apparatus	    26

  6   Results from a typical epinephrine experiment	    28

  7   The effect of epinephrine on ^C-urea influx	    29

  8   The effect of adrenergic blockade on epinephrine induced
          C-urea influx	    29

  9   The effect of acetylcholine on ^C-urea influx	    30

 10   Typical effect of acetylcholine on perfusion pressure	    31

 11   Typical effect of norepinephrine on perfusion pressure
        in ;'he absence and presence of adrenergic blockers	    32
 12   Diagram of apparatus used to perfuse isolated heads of
        rainbow trout	   35

 13   Experimental data for heat transfer plotted as a function of  _1
        ventilatory flow for one isolated head perfused at 20 ml min    36

 14   The effects of alterations in perfusion flow and heart rate
        on heat exchange from perfused head of rainbow trout	   37

 15   The effects of vasoactive agents on heat exchange from
        perfused head of rainbow trout	   38
                                    vi

-------
                                   FIGURES (continued)

Number                                                                 Page

 16   Diagram of model used to evaluate gill heat exchange in
        terms of whole body heat exchange	   41

 17   Graph of percent of total heat flux from the model gill
        with increasing perfusion flows	   42

 18   Model prediction of 7oQtot; as a function of ventilatory flow
        (VK) using experimental data at perfusion flows of 16 and
        20 ml rain'1	   43

 19   Microfil cast of a single pair of gill filaments from
        rainbow trout	,	   43
                                    vii

-------
                                  TABLES

Number                                                                page

  1   The 24-hr uptake of mercury by trout exposed to 275 ng/liter
        mercury as either C^HgCl or HgCl2	    6

  2   Values for some blood parameters of control and mercury
        exposed (10 ng Hg/liter) rainbow trout	   13

  3   In vitro oxygen consumption of gill tissue from control
        and mercury exposed rainbow trout in 100% and 107,
        phosphate  buffered saline (PT>G)	   14

  4   Calculated values for the iron content of tissues of a
        hatchery-reared lOOg rainbow trout	   18

  5   Hematocrit,  hemoglobin,  MCHC,  and percent reticulocytes of
        blood samples from control and experimental fish	   22

  6   Heat transfer (liA) of isolated-perfused second gill arches
        of rainbow trout	   36

  7   Heat transfer for isolated-perfused head of rainbow trout
        during control periods	   36
                                  v±ii

-------
                             ACKNOWLEDGMENTS
The research work reported herein was carried out in the Comparative
Physiology Laboratories, Department of Physiology, Michigan State Univer-
sity, East Lansing, Michigan 48824.

The research was supported in part by the Michigan Agricultural Experiment
Station, Project 122 and by the Institute of Water Research, Michigan State
University as OWRR Project No. A-064-MICH.  Some of the rainbow trout used
in the experiments were obtained from the Michigan Department of Natural
Resources, Grayling Research Station, Grayling, Michigan.  We are also in-
debted to Drs. W. Willford and R.E. Reinert, Great Lakes Fishery Laboratory,
Bureau of Sport Fisheries and Wildlife, Ann Arbor, Michigan for their sug-
gestions and generous supply of mercury exposed rainbow trout.

I am particularly indebted to Dr. J.R. Hoffert and Mrs. Esther Brenke not
only for their continued interest in this project but for their work in
preparing and photographing the many illustrations.  I also thank Dr. R.
Pax, Wm. Jackson, G. Eldred, T. Snow and J. Ubels for their help with
various technical aspects of the research and for many lively discussions
of the project.

Needless to say, this research work would not have been completed without
the aid of the support personnel listed below:

Principal Investigator;  Paul 0. Fromm, B.S. Zoology, M.S. Physiology,
Ph.D. Physiology all from Univ. of Illinois, Urbana, 111.  Professor,
Department of Physiology, Michigan State University, East Lansing, Mi.

Support personnel:  All persons other than R.L. Bergman who worked on or
were employed on this grant project have been graduate students in the
Department of Physiology at Michigan State University.  Appointments were
usually on a one-half time basis and were entitled 'Special Graduate
Research Assistant1.  A brief biographical sketch is given with each
person listed.

R.L. Walker, B.S. Biology, Alma College, M.S. Physiology, Ph.D. Physiology,
Michigan State University.  Currently an Instructor, Department of Biology,
Univ. of Calgary, Calgary, Alberta, Canada.

W.F. Jackson, B.S. Zoology, M.S. Physiology, Michigan State University.
Currently a graduate student in Dept. of Physiology, Michigan State
University.
                                     ix

-------
 D.V. O'Connor, B.S. Biology, Oakland Univ., M.S. Physiology, Michigan
 State University.  Currently employed as a biologist  for  the U.S. Co?-ps
 of Engineers, Baltimore, Md.

 P.R. Sorenson, B.S. Zoology, Michigan State Uni.vcisity.   Currently a
 graduate student, Dept. of Physiology, Michigan State University.

 K.R. Olson? B.S. Chemistry & Zoology, LaCrosse State University, M.S.
 Physiology, Ph.D. Physiology, Michigan State University.  Currently Asst.
 Prof. Indiana Univ. School of Medicine, U. of Notre Dame, Notre Dame, Ind,

 H..L. Bergman, B.A. Biology, M.S. Biology, Eastern Michigan University,
 Ph.D. Fisheries Biology, Michigan State University.  Currently Asst. Prof.
 Dept. of Zoology and Physiology, Univ. of Wyoming,  Laramie, Wyo.
* received no research assistant stipend, however, funds from Grant
  R-801034 were used to support research project.

-------
                                SECTION 1

                              INTRODUCTION
     The research work supported in part by Environmental Protection
Agency Grant No. R-801034 was conducted over a period of approximately
three calendar years.  There were essentially three general research areas,
all of which involved laboratory experiments with rainbow trout.  All
projects were carried out in the Comparative Physiology Laboratory, Depart-
ment of Physiology, Michigan State University, East Lansing, Michigan.  The
results of most of the research work have previously been published and the
reader should consult the respective publications for a more detailed dis-
cussion than that which appears below.  The information on the effects of
exposure to mercurials on the ultrastructure of gills is being prepared
for publication and the report of heat flux across the gills has been
accepted for publication in the Journal of Comparative Physiology,

     The experimental animals used in all experiments discussed herein
were hatchery-reared rainbow trout (Salmo gairdneri).  Fish were obtained
from the Michigan Department of Natural Resources hatchery at Grayling,
Michigan and a few wen  supplied by the Great Lakes Fishery Laboratory,
Bureau of Sport Fisheries and Wildlife at Ann Arbor, Michigan.  The most
recent experiments have been conducted with fish purchased from Midwest
Fish Farming Enterprises, Inc., Harrison, Michigan.  During the period of
acclimation to laboratory conditions fish were kept in large (300 liter)
fiberglass tanks in flowing dechlorinated tap water at about 12° C under
a controlled photoperiod of 16 hours light per day.  In our holding
facilities the fish were fed commercial trout food, either EWOS or Purina
Chow pellets.  Subsequent treatment and feeding of fish varied with each
experiment and are so indicated below.

-------
                                SECTION 2

                               CONCLUSIONS

      Twenty-four  hour  uptake  rate  of  either  203Hgd2  or  CH3203Hj',Cl  by
 rainbow trout  was  not  affected by  esophageal ligation.   Uptake  of  these
 two  mercurials in  non-feeding trout appears  to  be  by  way of  the  gills.
 Methyl  mercury enters  the  fish at  a faster rate than  the inorganic  form
 and  anomalous  tissue distribution  of  these two  mercurials  suggests  that
 inorganic mercury  does not  require methylation  prior  to  entry  into  the
 fish.   In vitro experiments using  radioactive mercurials demonstrated
 high affinity  of methyl mercury for red cells,  up  to  90% was bound  to red
 cells in 40 minutes.  Only  9% of inorganic mercury was taken up  by  red
 cells,  but, this percentage was  increased up to 65% if the cells were
 washed  and suspended in Ringer solution prior to incubation with mercury.

     Mercury was found in gills  of rainbow trout which had been  exposed
 to inorganic mercury but not  in  those  exposed to methyl  mercury.  No
 specific site  for mercury uptake was  identified and it is  suggested that
 inorganic mercury enters the  gill across the general  lamellar surface.
 High  concentrations were found associated with  the gill  cartilage.  Since
 little  ion diffusion occurs during tissue preparation, localization and/
 or identification of tissues  can be accomplished by scans  for various
 elements: sodium,  potassium chlorine and sulfur.  The technique  is not
 suitable for identification of  highly  volatile  compounds  such as methyl
 mercury due to  the necessity  of  subjecting tissues to high vacuum condi-
 tions, however, electron probe analyses may  prove  to be  useful in studies
 of active ion  transport systems  in gill tissue  and in investigations of
 the  effects of  heavy metal pollutants  on fishes.

     Exposure  of fish to methyl mercury for  12  weeks had no effect on the
 oxygen consumption of gill tissue or on plasma  electrolyte concentration
 but  the hematocrit index increased significantly.

     In studies of iron metabolism by  control and  iron deficient (bled)
 trout we found  that within 24  hours after injection most radioiron was
 absorbed from  the peritoneal  cavity and that clearance of  iron from the
 plasma  is a 3  component exponential process.   The erythropoietic response
 to bleeding is  slow.  The incorporation of radioiron into red cells
 probably represents (a) incorporation  of iron for hemoglobin formation
and  (b)  exchange of iron between cytopiasmic non-heme stores and the
plasma.   The liver, spleen and head kidney are  the main  iron storage
organs  in trout and bleeding  significantly reduced the liver iron content
but had little  effect on splenic iron.  Iron metabolism  in rainbow trout
appeared to be  a closed, recycling system as essentially no radioiron

-------
was excreted  in  feces and urine.

     The relative  functional  surface area  of  isolated-perfused gills was
evaluated by  measuring  the  influx of radiocarbon  labelled urea, a passively
diffusing molecule.  The catecholamines, norepinephrine and epinephrine,
increase functional gill surface area and  decrease overall branchial vas-
cular resistance.  Surface  area and resistance effects of adrenergic agon-
ists and blocking  agents demonstrated the  presence of both « and 3 adren-
ergic receptors  in rainbow  trout gills.  Stimulation of a adrenergic re-
ceptors increased  both  functional surface  area and branchial vascular
resistance, while  P> adrenergic receptor stimulation  increased functional
surface area  but decreased  branchial vascular resistance.  Acetylcholine
decreased functional gill surface area and increased overall branchial
vascular resistance.  The data presented strongly indicate that the func-
tional surface area of  rainbow trout gills can be regulated by changing
perfusion pathway  with  adjustments in the  relative vascular resistance
across the different pathways.

     Casts of the  gill  vasculature of rainbow trout perfused with Microfil
were examined visually.  The most conspicuous feature seen was enlargements,
called blebs, of the afferent filamental vessels that were located in the
outermost region of the interfilamental septa.  Several lamellae were
supplied by profusely branched afferent lamellar veseels, whereas most
lamellae were individually  connected to the efferent filamental vessel.
Vessels contained  in the septal area between filaments were also visualized
as well as vessels that ran parallel to the long axis of the filaments.

     Heat exchange was  evaluated in isolated-perfused second gill arches
and the intact branchial basket of rainbow trout.  The presence of ICT-'M
epinephrine in the perfusion solution increased the transfer maximum of the
gill suggesting a  change in perfusion pathway and/or vascular dimensions.
Changes in perfusion flow altered heat exchange by the gill which is again
due in part to changes  in vascular dimensions.  A model for the evaluation
of the gills  in relation to whole body heat exchange indicated that in the
range of perfusion flows from 4 to 20 ml min~l the gill may account for as
much as 30 to 607» of the total heat exchange in the animal.

-------
                                  SECTION 3

                               RECOMMENDATIONS
      1.  The electron microprobe  should  be  satisfactory  for  use  in investi-
gations of active  ion transport  systems  in  gill  tissues  and  in  research
on  the effects of  heavy metal  ions  (pollutants)  on  fishes.   The  technique
is  not suitable  for identification  of highly  volatile  compounds  such  as
methyl mercury due to the necessity of subjecting tissues  to high  vacuum
cond it ions.

      2.  Incorporation of radioiron into erythrocytes  is recommended  as a
very  sensitive parameter for study  of the effects of environmental  pollu-
tants on fish.  The study of any  one parameter,  however, would  probably
provide insufficient information  on which a meaningful i'ia«nosis could be
base.', thus it is  suggested that  detc-rmin.it >'o:is  nf  blood hemoglobin and
erythrocyte volume (hematocrit) as  well as  reticulocyte  counts  be made
concurrently with  experiments using radioiron.

      3.   It is highly recommended that the  isolated-perfused  gill  pre-
paration be further developed and standardized for  use as  an  assay  pro-
cedure to determine the effects of  water soluble pollutants  on  fish.
Using this preparation one could  quantitatively  describe the  effects  of
pollutants on gills in terms of changes in gill  permeability  to water,
gill  ion transport and possible biotransformation of pollutants by  pul-
monary (gill) metabolism.

-------
                                  SECTION 4

              EFFECTS OF METHYL MERCURIC CHLORIDE AND MERCURIC
                          CHLORIDE ON RAINBOW TROUT

A STUDY OF UPTAKE PATHWAYS INTO WHOLE FISH

     Uptake of mercury by fish can occur via three possible routes, the gas-
trointestinal (GI) tract, the skin, or the gills.  The importance of the GI
tract in mercury uptake has been demonstrated by Hannerz (1968), Backstrom
(1969), Miettinen et_ _al. (1969, 1970), Miettinen (1970), and others.  Al-
though orally administered mercury was often regurgitated in their experi-
ments, the general consensus was that mercury, especially methyl mercury, is
absorbed through the GI pathway.  In nonfeeding, freshwater teleosts, which
also accumulate high tissue levels of mercury, the amount of mercury absorbed
via the GI tract may be limited by a low drinking rate (around 13 yliters/hr
per 100 g for a freshwater adapted eel (Maetz and Skadhauge, 1968).  However,
if exposure to mercury stimulates drinking rates, the GI tract pathway could
play a significant role in mercury uptake.  To our knowledge there have been
no efforts to delineate the importance of each of these pathways in nonfeed-
ing teleosts.

Methods

     In order to determine the significance of the GI tract in mercury uptake
by nonfeeding trout, the GI pathway was eliminated in one group of fish by
esophageal ligation.  The fish were divided into three groups:  control, sham
ligated, and ligated.  The purpose of the sham ligation procedure was to de-
termine if any mercury was taken up through the body wall incision.  Eight
fish from each group were placed into 100 liters of aged tap water than con-
tained 275 ng/liter Hg as either 2°3HgCl2 or CH3203HgCl.  After 24-hr expo-
sure fish were randomly removed and blood samples taken from the caudal vein.
They were then killed and tissue samples from the gills, liver, kidney, heart,
skeletal muscle, stomach and intestine were removed and placed into preweighed
vials.  Tissue activity was determined with a 2-inch thallium activated sodi-
um iodide crystal well detector and Nuclear Chicago analyzer/sealer.  Tissue
dry weights were also determined.  Analysis of the data was done on a Control
Data Corp. 6500 computer.  One way analysis of variance was used for all sta-
tistical comparisons.

Results and Discussion

     Uptake of divalent mercuric ions, as well as methyl mercuric ions, occurs
primarily by way of the gills.  Ligation of the esophagus has no apparent ef-
fect on 24-hr accumulation of either form of mercury by any tissue with the

-------
 exception  of  gills  from  fish  exposed  to  inorganic mercury.   In  this  instance
 gills  from the  sham-operated  fish had  the  highest mercury concentration.
 There  was  a noticeable trend,  although not  statistically significant,  toward
 increased  tissue mercury  concentrations  for  sham-operated fish  in both experi-
 ments.  The explanation  for this is unknown  but  it possibly  could be related
 to variations in surgical trauma.  In  both methyl and inorganic mercury-ex-
 posed  fish, dry weight to wet  weight ratios  remained constant for all  tissues
 except liver  and kidney.  Even though  there  was  an increase  in  hydration of
 liver  and  kidney tissue  the change was no  greater than 1.5%  and probably does
 not reflect gross inward water flux which would  be needed to significantly
 affect tissue mercury concentrations.

     Methyl mercury was taken  up by all  tissues  at a much faster rate  than
 inorganic  mercury over the 24-hr period  and  the  large concentration differen-
 ces in gill tissue observed between the  two  compounds (28 vs. 405 ng Hg/g dry
 weight) can be explained only  by faster  penetration of the methylated  form.
 Mucus  which continually coats  the gill epithelium can act as an ion binding
 resin  to trap mercury and prevent its  access to  the tissue.  Since ambient
 concentrations of both mercurials were identical for the experiments and as-
 suming both ion species were trapped or  precipitated by the mucus, the con-
 centration of inorganic mercury at the gill  should be no less than one half
 that of the organic form.  This is based on  the  assumption that one sulfhy-
 dryl (SH)  binding site is occupied by  each monovalent methyl mercuric  ion and
 a maximum  of  two SH sites by each divalent inorganic mercury cation.

                                   TABLE 1

    THE 24-HR UPTAKE OF MERCURY BY TROUT EXPOSED TO 275 NG/LITER MERCURY
                     AS EITHER CH3HgCl or HgCl2  (n = 8)
Tissue
Gill
Liver
Stomach
Intestine
Heart
Kidney
Ku s c 1 e
Blood
	
	 	 ••— ' '• • 1 • •! « •• --•- ••-• — .U - -
Tissue activity
CH3HgCl
404.6
29.3
9.6
15.9
30.7
92.1
2.4
81.8
(ng Hg/g dry vt . )
HgCl2
28.1
2.6
2.3
12.3
2.6
4.6
0.2
15.4
Ratio
Me t hy 1 Hg : i nor na n i c Us;
14.4
11.3
3.5
1.3
11.8
19.8
15.4
5.3
     A ratio of gill methyl mercury concentration to gill inorganic mercury
concentration (Table 1) of 14.4 indicates that the increased uptake of the
methylated form must reflect either entry of mercury into the gill tissue
where additional binding sites are available or increased mucus secretion
which binds the methylated form on the exterior of the gill.  Observations of
fish from this and numerous other experiments have shown that methyl mercury
does not stimulate greater mucus secretion than inorganic mercury and, in
fact, the opposite is often the case.  The increased penetration of methyl

-------
mercury into gill tissue may simply reflect the greater lipid solubility of
this compound over that of inorganic mercury.  This explanation would hold if
mercury gains entry into gill epithelial cells across a lipid-containing cell
membrane and not via water-filled pores.  On the other hand the bivalency of
inorganic mercury may promote tighter binding to mucoproteins thereby re-
stricting its diffusion into gill tissue.

     Comparison of the distribution of the two mercury compounds within a tis-
sue demonstrates preferential accumulation of mercury based on its chemical
form.  Ratios of methyl mercury to inorganic mercury in various tissues after
a 24-hr exposure are given in Table 1.  The magnitude of these ratios is par-
tially attributable to increased uptake of methyl mercury by the gill, how-
ever, all the tissues should exhibit approximately the same ratios if there
is no selectivity by the tissues for specific forms of mercury.  The differ-
ences in these ratios show that tissues such as the kidney, muscle, gill,
heart and liver preferentially accumulate the methylated form and blood, sto-
mach and intestine do not.  Low ratios for the latter three tissues could be
due to selective accumulation of the inorganic form which is masked by higher
uptake rates of the organic compound.  Using radioautography Halbhuber et al.
(1970) have demonstrated selective accumulation of radiomercury by the Paneth
cells in the intestine of the rat and guinea pig.  It was postulated that one
function of these cells is to transfer divalent heavy metal ions from the
body fluids into the lumen of the gut.  Similarly Paneth cells in the trout
gut might sequester divalent mercury but not methylater forms.

     Backstrom (1969) administered methyl and inorganic mercury intravenously
to speckled trout.  After 24 hr the distribution patterns of methyl mercury
were similar to our data for rainbow trout with the exception of the low val-
ues for the gills.  He reported highest inorganic mercury concentrations in
the kidney; values for blood and gill were about 50% lower.  These small dis-
crepancies in distribution when compared to data presented here are probably
due to the mode of administration of the mercurial.  The tissue mercury dis-
tributions reported by Hannerz (1968) for cod exposed for 24 hr to waterborne
mercurials are similar to those found in the present study.  He postulated
gill uptake as the principal mode of entry of mercury into cod and pike.

     The wide variation in organic/inorganic tissue concentration ratios that
we noted supports the idea that inorganic mercury can cross the gill epithe-
lium as divalent mercury ions or as some complex without prior methylation.
The present experiments have eliminated the GI tract as a major pathway for
mercury uptake in nonfeeding trout and it is widely accepted that the skin is
relatively impermeable (Prosser, 1973, pp. 48).

UPTAKE OF MERCURY BY TROUT ERYTHROCYTES _IN VITRO

Methods

     Heparinized whole blood or washed cells were placed in 25-ml stoppered
Erlenmeyer flasks.  The mercury solution was added to the blood at time zero
and the flasks stirred at 16 rpm with a multipurpose rotator.  At timed inter-
vals 0.2 ml of blood was removed and placed in a 0.4 ml polyethylene

-------
m
  icrocentrifuge  tube.   Packed  red  cells were  then  resuspended  in Ringer  solu-
 tion,  centrifuged  and  the  subsequent  supernatant was  placed  in a third micro-
 centrifuge  tube.   Hematocrits  were determined  in duplicate  for all blood sam-
 ples  incubated.  The concentration of mercury  from the packed  red cells, plas-
 ma and wash was  determined by  scintillation counting.  Percent uptake of mer-
 cury  into the cellular  fraction was standardized by correction for equal vol-
 umes  of  red cells  and plasma.  All glassware used  was coated with silicone  to
 minimize absorption of  the mercurial onto  the  glass.

 Results  and Discussion

      The percent uptake of both mercurials into erythrocytes as a function  of
 time  is  plotted  in Figure 1.   With the exception of methyl mercury incubated
 with  whole  blood,  equilibrium  was  reached  in less  than 10 minutes.   Uptake  of
 methyl mercury did not reach equilibrium when  incubated with whole blood until
 after  30 to 40 minutes.  Analysis  of this  reaction revealed a  two-component
 uptake system, the first having a  half time of under 1 min, the second about
 11 min.  (Uptake values of less than 1 min are only estimates  due to the in-
 herent time lag incurred during blood centrifugation and separation.)  This
 can be explained by initial binding of a small percentage of mercury to plas-
ma proteins or thiols,  the remainder being taken up by the red cells as a
 function of the free or nonplasma  bound mercury.  After the red cells have  re-
moved most of the unbound mercury  from the plasma a new equilibrium (slow com-
 ponent)  is established that is dependent on the dissociation of bound mercury
 from  'plasma sites' prior to red cell uptake.  This postulate  is further sup-
 ported by the mode of uptake of methyl mercury into red blood  cells suspended
 in Ringer solution.  In this instance the uptake rate is so fast that only
one component can be determined.   The increased G^Rg"*" binding at equilibrium
by washed cells in Ringer solution compared to cells in plasma (95% vs.  90%)
is probably indicative of some irreversible mercury binding to plasma protein.
Failure of the washed cells to bind 100% of the mercury could  be due to slight
hemolysis coupled with some mercury binding by the incubating media or absorp-
 tion on glassware.

     Inorganic mercury incubated with whole blood binds almost exclusively to
the plasma fraction; less than 9% was taken up by red cells after 1440 min
 (not shown in Figure 1).  Washed cells suspended in Ringer solution take up
9-10 times more inorganic mercury  than cells do in plasma.  Removal of plasma
proteins undoubtedly is the critical factor.  There is also a  significant
difference  in percent uptake by red cells that is dependent on the suspending
media; cells suspended in Abbott Ringer take up less Kg"1""*" than cells suspen-
ded in phosphate buffered Ringer solution.  Since the only difference is the
presence of magnesium and the  buffering capacity of the PER either the magne-
sium or pH or perhaps both factors must enhance equilibrium binding by red
cells.

     In other experiments it was shown that uptake of either methyl or inor-
ganic mercury was,  in all instances, independent of the length of the incuba-
tion period which preceded addition of mercury to the cell suspension.   Thus,
secretion of a chelator by trout red cells is not a factor in  the failure of
these cells (especially washed cells) to bind all mercury present.   As

-------
 indicated in Figure 1 less than 100% of the available methyl or inorganic
 mercury was taken up by red cells in these experiments.
         LJ
            IOO-,
             90-
             80-
             70-
         o
         03  60-
         cr
             40-

             30-

             20-

             10-

                      10
30
60
                                                        -/A
180
                                  TIME (MIN)
Figure 1.  Percent uptake of methyl mercury or divalent mercury by red blood
cells as a function of time.  A = CI^HgCl incubated in whole blood (15);
HgCl2 in whole blood (6); A = CI^HgCl  incubated with washed cells suspended
in Abbott Ringer (4);    = HgCl2 incubated with washed cells suspended in
Abbott Ringer (4); ° = HgCl2 incubated with washed cells suspended in Phos-
phate Buffered Ringer (4).   Exposure concentrations approximately 0.0133 to-
tal ppm CH-jHgCl and 0.054 ppm total HgCl2-  Number of observations (N).  Stan-
dard errors are not shown on the graph as they were smaller than the diameter
of the data points.

-------
      It appears that saturation of  trout erythrocytes with mercurials must in-
volve either  (1) some group(s) such as SH, amono, or other which are respon-
sible for cell membrane integrity and are not saturated at the time of hemol-
ysis  or (2) a disruption of ionic balance and subsequent osmotic hemolysis
caused by mercury levels below those required for total cell saturation.  SH-
containing compounds such as albumin, cysteine and reduced glutathione failed
to remove either inorganic or organic mercury previously bound to red cells,
which suggests strong mercury-red cell binding.

MERCURY UPTAKE AND ION DISTRIBUTION IN GILLS OF RAINBOW TROUT

      Efforts to localize and/or characterize the pathway of uptake of mercury
in gills of trout using radioautographic techniques were largely unsuccess-
ful.  Autographs of gills of fish exposed to either methyl or inorganic mer-
cury  exhibited a diffuse grain pattern which precluded identification of the
uptake site.  Lack of grain density in areas containing many chloride cells
probably indicates that these cells are not involved in the uptake of either
mercurial.  The thinness of lamellae and the small interlamellar distances
prevented localization of preferential mercury binding sites.  Dense grain
patterns due to 203jjg-H- were -found associated with the filamental cartilage.
Our second attempt to localize the gill pathway for mercury uptake was done
with  the electron microprobe.  The  following is a report of these experi-
ments.

Methods

      Fish were exposed to the mercurials which were added to water in 100-
liter tanks.   The concentrations used were 0.25 and 0.05 ppm Hg for inorganic
and organic mercury, respectively.  After 23-hr exposure a second amount of
mercury equal to the first was added to the tanks and fish were removed 1 hr
later and killed.   The second pair of gill arches were removed from mercury-
exposed fish within 15 sec after the fish was killed.  The arches were quick
frozen in isopentane which has been cooled to near -160° C with liquid nitro-
gen.   Details of mounting, embedding, sectioning and preparation of sections
for use in the ARL Microprobe are given by Olson and Fromm, 1973.  Tissue
scans and image scan micrographs were recorded on Polaroid 200 black and
white film and the line scans were recorded on a Hewlett Packard x-y recorder.
Low peak to background rations during mercury analysis made the tissue scan
procedure difficult to interpret, however, line scan procedures proved ade-
quate for mercury identification.  Line and tissue scans were also used for
detection of other elements as both an aid in tissue orientation and a check
on diffusional ion movements which might have occurred during sample prepara-
tion.  The 0.5 y diameter of the primary electron beam permits fairly high
resolution in the localization of ions within a given area especially with
the line scan procedure.  All analytical data obtained were qualitative or
only  semiquantitative as the refinements and sophistication required for quan-
titative determinations were beyond the scope of this preliminary study.

Results and Discussion

      For the line scans as the beam entered the area known to contain large
numbers of chloride cells, concentrations of all ions except mercury

                                      10

-------
increased.  Interlamellar areas contained definite mercury peaks and the lo-
cation of lamellae was marked by increases in the concentrations of Na, K, Ca
and Cl.  Mercury content appeared to be inversely related to concentrations
of the other elements and it did not increase until the beam entered the
space normally filled by environmental water.  This confirms results of radio-
autographic studies and supports the concept that inorganic mercury is not
selectively taken up by chloride cells nor concentrated by them in 24 hr.  It
should be stated that we were unable to detect mercury in gills from tout ex-
posed to methyl mercury even though our previous study showed that fish ex-
posed to methyl mercuric chloride contained 10 times more mercury than those
fish exposed to mercuric chloride.  The inability to detect mercury after
methyl mercury exposure was probably due to the loss of this element during
tissue preparation.  Crystals of methyl mercury will sublime at room tempera-
ture, therefore, mercury could be lost during freeze drying, carbon evapora-
tion or in the vacuum volumn of the microprobe.  All of these procedures re-
quire a high vacuum.

    The mercury peaks observed were due in large part to mercury which was
added 1 hr prior to sampling and demonstrated the effectiveness of the inter-
lamellar mucus to trap ions.  Presumably mercury is contained in the mucus
from which it can either enter the gill or be eliminated with the sloughed off
mucus.  McKone e^ al. (1971) exposed goldfish to 0.25 ppm HgCl2 for 3 hr and
found that 79.3% of the mercury in excised gills was removed by washing with
80% ethanol.  They concluded that this fraction represented mercury trapped
by the mucus coating the gill, however, they did not examine gill tissues af-
ter the ethanol wash to verify that only mucus was removed.  Gills are very
fragile and it is quite possible that some epithelium was also removed in the
wash process.

     Tissue scans of other material showed that mercury concentrations were
highest in cartilage and were correlated with high calcium, sulfur and phos-
phorus levels.   Microprobe (and radioautographic) studies have clearly indi-
cated that there is no preferential accumulation of mercury in lamellae.  It
appears that mercury can enter gills through the general lamellar surface,
thus active uptake is probably of little consequence in mercury influx.  Once
in the filament, much inorganic mercury is accumulated by cartilagenous tis-
sue.   Other gill tissues exhibit no preferential uptake of mercury with re-
spect to a given cell type,  i.e. epithelial cells, pillar cells, or chloride
cells.

     Preparation of tissues for electron microprobe analysis avoids the use
of various aqueous solvents as are commonly employed in routine analytical
histochemistry.  Thus, localization of many ions (e.g.  small molecular weight
monovalents) with the probe probably reflects a fairly true tissue distribu-
tion of these ions.  We suggest that electron probe analyses should prove
beneficial in a study of ion distribution in chloride cells of fish exposed
to various salinities or to pharmacological agents.   Further information on
the ion transfer role of mucus, which covers the surface of chloride cells
from marine species, could be gained by studies utilizing the electron
microprobe.
                                      11

-------
 EFFECT  OF METHYL MERCURY  ON  GILL  METABOLISM AND  BLOOD  PARAMETERS  OF  TROUT

      The  gills  of  freshwater teleosts  function as  the  primary  site for  the
 active  absorption  of  ions from  the  external media  and  for  the  exchange  of
 respiratory  gases.  Large amounts of water  must  be passed  over the gill sur-
 faces to  meet the  oxygen  demands  of the  fish consequently  when fish  are in
 contaminated waters their gills can be exposed to  large amounts of water sol-
 uble  pollutants.   The  experiments discussed below  were performed  to  determine
 if methyl mercury  has  any effect  on the metabolism (oxygen consumption)  or
 physiological function (plasma electrolyte  regulation) of  the  gill.

 Methods

      The  experimental  design involved  exposing a group of  starved fish  to 10
 yg Hg/liter  administered  as  methyl  mercuric  chloride while another group of
 starved fish served as controls.  The  mercury was  administered  to the experi-
 mental fish  using  a gravity  feed  system in which the flow  rate of the concen-
 trated mercury  solution into  the  tank  was correlated to the flow of  water
 through the  tank to maintain a concentration of  10 yg Hg/liter.  At  the  end
 of 4, 8 and  12 weeks the  fish were  killed and determinations of the  follow-
 ing parameters were made:  hematocrit, oxygen consumption  of gill tissue,
 protein content of gill sample and  plasma electrolyte concentrations (Na+,
 K , Cl~, Mg++ and Ca~H").  Plasma  sodium and potassium were determined on a
 Beckman flame photometer, magnesium and calcium were determined on a Perkin
 Elmer atomic absorption spectrophotometer and chloride was measured with a
 Buchler chloridometer.  All plasma  electrolyte determinations were done in
 duplicate.   Samples of gill  filaments  from the second and  third branchial
 arches and were suspended in 3 ml of 10% and 100% phosphate buffered saline
 (PBS).  The  oxygen consumption of the  filaments was then measured polarograph-
 ically at  12° C with a YSI biological  oxygen monitor.  Calculation of oxygen
 consumption was based on  the solubility of oxygen in 10% and 100% PBS at 12°
 C and the  percent of initial oxygen consumed during a 10 minute period.   Fol-
 lowing oxygen consumption measurements the 3 ml samples of PBS containing the
gill filaments was sonified  to remove  as much tissue as possible from the
cartilaginous skeleton and the protein content of the mixture was determined.

Results and Discussion

     Results from  this investigation (Table 2) indicate that up to 12 weeks
 exposure  to  methyl mercuric  chloride (10 yg Hg/liter) does not significantly
 affect the in vitro metabolism of the  gill or the concentration of plasma
 electrolytes in rainbow trout.  There  was a significant increase in  the he-
matocrit of  experimental  fish after 12 weeks exposure.  Webb (1966) pointed
 out that many mercurials  are  known  to  cause hemolysis of RBC's which would
 tend  to cause a decrease  in  the hematocrit  index, not an increase as we found
 in our experiments.  Possibly methyl mercury directly stimulated erythropoi-
 esis causing an increase  in  hematocrit of trout or the fish may have over-
 compensated  erythropoietically in response  to RBC loss resulting from expo-
 sure to mercury.
                                      12

-------
                                   TABLE 2

           VALUES FOR SOME BLOOD PARAMETERS OF CONTROL AND MERCURY
                   EXPOSED (10 NG Hg/LITER) RAINBOW TROUT
Blood
Parameter
Plasma Na+
(meq/L)
Control
Hg-treated
Plasma K1"
(meq/L)
Control
Hg-treated
Plasma Cl"
(meq/L)
Control
Hg-treated
Plasma Ca~*~*"
(meq/L)
Control
Hg-treated
i i
Plasma Kg"1"1"
(meq/L)
Control
Hg-treated
Hematocrlt
(7. RBC)
Control
Hg-treated
Mean±S.E.(N)
Length of
4


140. 08±1. 35(12)
142.58±1.93(12)


2.74±0.17(12)
2.61±0.13(12)


123. 36±0. 95(12)
123. 27±2. 07(12)


4.36±0.28(12)
3.97±0.25(12)


1.82±0. 05(12)
1.66±0.09(12)


19. 44±1. 28(12)
18.00±0.88(12)

Exposure (Weeks)
8


142.68±3.48( 8)
149. 20±1. 83(12)


1.36±0.22( 8)
1.42±0.19(12)


122.02±5.13( 8)
128. 73±1. 71(12)


4.00±0.15( 8)
4.15±0.26(12)


1.96±0.04( 8)
2.02±0.06(12)


19.59±2.63( 8)
21. 52±1. 38(12)


12


139. 37±3. 13(8)
133.75±7.66(8)


1.70±0.27(8)
l.77±0.36(8)


121. 75±4. 00(8)
110.85±11.14(8)


3.76±0.15(8)
3.73±0.25(8)


2.15±0.09(8)
1.92±0.23(8)


21.31±1.46(8)
30.34±2.40(8)

     Although there were no significant differences in plasma electrolyte
levels for paired comparisons between control and mercury treated fish after
4, 8 and 12 weeks exposure, over the 12 week period several of the electro-
lytes did vary considerably.  Plasma sodium for controls remained fairly con-
stant at about 140 meq/liter while those for experimental fish varied between
133 and 149 meq/liter.  This greater variability in the experimental fish may
represent some effect of methyl mercury on the active transport of sodium by
gills but further work is needed to substantiate this hypothesis.  No plaus-
ible explanation can be offered for the increase in plasma magnesium levels
of control fish after 12 weeks exposure.

     Gill samples from control and mercury exposed fish measured in 100% PBS
showed a significantly higher rate of oxygen consumption than samples mea-
sured in 10% PBS (Table 3).  This could be the result of the sudden transi-
tion of gill tissue from fresh water (5-10 mOsm) to 100% PBS (285-290 mOsm)
which represents both an osmotic shock and a probable increased salt load for
the tissue.  Chloride cells present in the gill filaments could conceivably
have responded with an increased rate of active transport to maintain osmo-
tic equilibrium and thus account for at least some of the difference in oxy-
gen consumption.  It is interesting to note that the decrease in oxygen con-
sumption for control and mercury treated fish in 10% PBS occurred during the
same period of time that a decrease in plasma potassium concentration was
                                     13

-------
 seen.   It would appear that these changes represent some form of accommoda-
 tion to the effects of starvation since they were observed in both control
 and experimental fish.

                                    TABLE 3

          IN VITRO OXYGEN CONSUMPTION OF GILL TISSUE FROM CONTROL AND
          MERCURY EXPOSED (10 pG Hg/LITER)  RAINBOW TROUT IN 100%  AND
                      10% PHOSPHATE BUFFERED SALINE (PBS)
                Data expressed as pi 02 consumed/hr/mg protein.
Sample

100% PBS
Control
Hg-treated
10% PBS
Control
Hg-treated
MeaniS.E.
(N)


17.
18.
14.
13.


4
35±0
53±1
13±0
13±0

Length

.74(12)
.37(12)
.78(12)
.64(12)

of Exposure (Weeks)

15.
14.
8.
8.

8
56±0
48±1
74±0
64±0



.98( 8)
•15(12)
.53( 8)
.50(12)



16
17
9
9

12
.47±2
.27±1
.78±0
-57±0


.07(8)
.15(8)
.77(8)
.63(8)

      The data presented above seems  to  indicate  that  longterm exposure  to
methyl mercuric chloride does not alter the metabolism of  the gill or affect
its  role in plasma electrolyte regulation.  Such a generalization may at best
be premature.  McKim et al.  (1970) reported significant changes in seven blood
parameters of brook trout after exposure to three different concentrations of
copper for periods of 6 and  21 days, however, when the experiment was exten-
ded  to 337 days five of the  seven parameters were similar  to control values.
It is possible that methyl mercury exerted a similar  transient effect on the
fish in this study and remained undetected or possibly a longer period  of ex-
posure is necessary before significant changes in oxygen consumption or plas-
ma electrolyte concentrations occur.  The need for longer  term studies with
shorter time intervals between samplings is evident and would do much to fur-
ther our understanding of the physiological effects of mercurials on fish.

EFFECT  OF MERCURY EXPOSURE  ON RAINBOW TROUT GILL ULTRASTRUCTURE

     Several investigators using light microscopy and standard histological
techniques have described the gross histopathological or structural changes
in the gill epithelium after exposure of fish to mercurials.  Schweiger (1957)
found that fish exposed to 0.03 mg Hg/liter survived  longer than seven days
and considered this concentration harmless.  At higher concentrations the sur-
vival time was progressively shortened and massive damage  to the respiratory
epithelium was observed.  Below is a description of some morphological
changes that occurred in the ultrastructure of gills  of rainbow trout which
were exposed to methyl mercuric chloride and mercuric chloride.  Using light
microscopy, no significant histopathological changes were  noted in these
gills when compared to untreated controls.
                                      14

-------
Methods

     Rainbow trout which had been exposed to monomethyl mercury for 4 weeks
or longer were obtained from the Great Lakes Fisheries Laboratory, Bureau of
Sport Fisheries and Wildlife, Ann Arbor, Michigan.  These fish ranged from 35
to 60 g and were exposed to an average concentration of 200 ng Hg/liter (as
CH3HgCl) at 5, 10 and 15° C in a constant flow bioassay apparatus.  Gill sam-
ples from these fish were collected and fixed at Ann Arbor.  For transmission
electron microscopy (TEM) individual gill filaments were dissected from sup-
porting tissue of the second gill arch of a freshly killed fish, fixed in 4%
glutaraldehyde buffered with S«5rensen buffer and post fixed in 1% osmium tet-
roxide.  The tissues were then dehydrated in graded concentrations of ethanol
or acetone and embedded in Spurrs embedding media.  Thin sections were
stained in uranyl acetate and lead citrate and were examined with a Philips
EM 300 electron microscope.  For scanning electron microscopy (SEM) the
second gill arch (contralateral) side was removed, rinsed with tap water and
fixed in 50% glutaraldehyde.  After six rinses of 30 minutes each in S^rensen
buffer the tissues were either frozen in isopentane cooled with liquid nitro-
gen and subsequently lyophilized or they were dehydrated with ethanol and
amylacetate and dried by the critical point method.  Dried filaments were
separated from the supporting cartilage, attached to coverslips and coated
with carbon and gold palladium.  The tissues were then examined with an AMR
Model 900 scanning electron microscope.

Results and Discussion

     For detailed descriptions of normal rainbow trout gill ultrastructure
the reader should consult Olson and Fromm (1973) for SEM work and Morgan and
Tovell (1973) for TEM observations.  Investigations with the SEM indicated
that a 24 hr exposure to 0.05 mg/liter methyl mercury produced some slight
disarrangement of the microridges present on the outer surface of lamellar
epithelial cells although junctions between epithelial cells appeared essen-
tially normal.  This same exposure resulted in what is believed to be an in-
crease in the degenerative process in chloride cells which, in SEM micro-
graphs is typified by loss of surface microvilli.  There was also an increase
in the number of degenerating chloride cells in gills of fish exposed to 200
ng/liter methyl mercury for 4 to 8 weeks.  Exposure to HgCl2 produced some-
what similar changes in the surface features of chloride cells and also led
to the appearance of many smooth non-ridged lamellar epithelial cells.

     Observations with the TEM revealed that exposure to 0.05 mg/liter methyl
mercury had relatively little effect on the appearance of the epithelial mi-
croridges which appear more like villi in TEM micrographs.  Some vacuolation
of chloride cells was apparent but many of these cells appeared normal.  A 4
week exposure to 200 ng/liter methyl mercury produced red cell vacuolation
and slight loss of epithelial cell ridge integrity.  The vascular space was
often  filled with debris, possibly the contents of ruptured red cells.  De~
generating or abnormal chloride cells appeared to be present in greater num-
bers but normal chloride cells were also present.  Extensive vacuolation of
chloride cells was noted in fish exposed to 200 ng/liter methyl mercury for
8 weeks.  This vacuolation extended throughout the cell and appeared to be
                                      15

-------
associated with the smooth endoplasmic reticulum (SER) of these cells.  In
addition to their becoming vacuolated these cells were often detached from
subepithelial structures.  Many red cells in fish exposed to this concentra-
tion appeared to be normal but vacuolation was not uncommon.  Chloride cells
in gills of fish exposed for 5 days to 0.25 mg/liter inorganic mercury and
those exposed to lower concentrations for longer periods of time showed ex-
tensive vacuolation and loss of surface microvilli.  Small vacuoles with a
low electron density were found near the apical area of the chloride cells,
whereas, large (up to 0.5 um diameter) vacuoles were located in the basal
portion of the cell and were more electron dense.  This is in contrast to
methyl mercury exposed chloride cells in which the small vacuoles were pres-
ent throughout all but the extreme apical portion of the cell.  This differ-
ence may reflect different penetration rates of these two mercurials or dif-
ferent biochemical interactions.  Cross sections through the smooth areas on
the lamellar epithelial surface revealed that these areas were the outer or
apical surface of large cells which appear to push apart and replace regular
epithelial cells on the lamellar surface.  These smooth cells are somewhat
larger and less electron dense than normal epithelial cells; they contain a
relatively large nucleus, numerous mitochondria and cytoplasmic vacuoles.
Except for the absence of a slightly ridged epithelial surface, these cells
are similar to the 'new' type of epithelial cell seen in methyl mercury ex-
posed fish.
                                      16

-------
                                  SECTION 5

                               IRON METABOLISM

METABOLISM OF IRON BY NORMAL AND IRON DEFICIENT RAINBOW TROUT

     Iron is an essential element  involved in  the physiological functions of
oxygen  transport and cellular respiration.  An enormous volume of literature
has resulted from measurements made on human subjects and rats which deals
with iron absorption, the mode of  iron transport via the plasma, and the
utilization and conservation of iron.  There is, however, relatively little
information concerning iron metabolism in lower vertebrates.  The study
reported here was designed  to increase our knowledge of the metabolism of
iron in salmonidae.  The objectives were:  (1)  to study iron distribution and
utilization in normal control and  iron deficient trout over a 30 day period
following an intraperitoneal (i.p.) injection  of 59Fe,  (2) to investigate the
absorption of iron into red blood  cells  and (3) to evaluate  the relative im-
portance of various iron storage pools in providing iron to be used during
time of increased erythropoietic activity by head kidney tissue.

Methods

     One group of fish were made iron deficient by removal of 40% of the
blood volume in four bleedings of  10% of the blood volume over a seven day
period with 48 hours between bleedings.  On the day of the last bleeding
both experimental and control fish received an i.p. injection of 1 yCi-^FeCl-,/
100 g fish in 0.9% saline.   Six experimental and six control fish were sacri-
ficed 1, 2, 4,  8, 11, 16, 23, and  30 days following 59Fe injection.  Blood
samples were taken via the hemal arch and fish were killed with a sharp blow
to the head.  Tissue samples were  taken  and weighed  using a Roller-Smith
Balance.  Intestinal contents were tare  weighed  on a Mettler Model B5 Balance.
Tissues and intestinal contents were stored at 0°C in glass culture tubes
until analyzed for total iron and/or 59pe content.

     The percentage of pack red blood cells was determined by the micro-
hematocrit method and hemoglobin by the  cyanmethemoglobin method using Hycel
(Houston,  Texas) standards and reagents.  Reticulocyte counts were made on
blood smears prepared by supra-vital staining.  After staining a total of
1000 cells, both immature and mature red cells, were counted using an ocular
grid.  The number of reticulocytes was expressed as a percentage of this
population.  Plasma iron was determined  colorimetrically and expressed as
ug Fe/100 ml plasma (yg%).   Plasma from  any hemolyzed samples was discarded.
The plasma total iron binding capacity (TIBC) was calculated from the plasma
iron content and the unsaturated iron binding capacity  (UIBC).  Tissue iron
was measured colorimetrically after wet-ashing in sulfuric-nitric acid
(5:1).  Total blood volume and tissue blood volume was determined using 51(]r-
labelled RBCs.   The tissue blood volumes were used to correct tissue 59pe
activity and total iron concentration for the blood iron content.  To evaluate

                                     17

-------
 absorption of radioiron  by the intestine the method of De Benedetto and
 Farmanfarmaian (1975) was used to measure renal excretion of radioiron the
 urinary bladders of trout were cannulated and urine collected in a manner
 described by Froiran (1963).

 Results and Discussion

      With respect to normal control rainbow trout the data presented in Table
 4 indicate that the spleen has the highest iron concentration but the liver
 due to its mass,  represents the prime storage area for iron in trout.  Tissue
 iron concentration and/or total iron content corrected for blood iron content
 remained constant over the 30  day study period.   Plasma contained very small
 amounts of iron but plasma does contain transferrin,  the iron transport pro-
 tein which shuttles iron back  and forth between the various iron pools, thus
 it plays an essential role in  the metabolism of iron by trout.

                                    TABLE 4

             CALCULATED VALUES  FOR THE IRON CONTENT OF TISSUES OF A
                       HATCHERY-REARED 100 G RAINBOW TROUT

             Tissue	yg Fe/g (X + S.E.)       yg Fe Total
Liver
Spleen
Head kidney
Caeca
Intestine
Plasma
168
381
122
36
34
0.55
± 13
± 46
± 11
± 3
± 3
± 0.03
185.0
81.2
24.3
45.3
14.3
1.2
     The clearance of iron from the plasma  of  rainbow  trout  following  i.p.
injection of 59Fe was determined.  When plotted as  log concentration against
time, radioiron content of the plasma presents the  clearance curve  seen  in
Figure 2 upper left.  The curve is composed of three exponential  components
G!» C? anc* C3» similar to those characterized  for human plasma  iron clearance.
The equation for the line is given at the top  of Figure 2 where:  Ct = fraction
of the initial dose of 59Fe present in the  plasma at time t;  Co^, Co2, 003 = y-
intercepts for components 1, 2 and 3 expressed as decimal equivalents of
initial dose; k]_, k2, k3 = rate constants for  components 1,  2,  and  3; e =
base of natural logarithms and t = time in  days.

     The first component (Cj) with a half time (Tl/2)  of 5.98 hours represents
the flux of radioiron-labelled transferrin  from the plasma into the tissue
extravascular spaces.  The second component with a half-time of 1.16 days
represents the return of 59pe-transferrin to the plasma pool, a process which
is responsible for the change in the slope  of  the clearance  curve between
days 2 and 8 following i.p. injection of radioiron.  The attainment of
equilibrium between radioiron in the plasma pool and the erythropoietic and
storage labile iron pools occurs after day  8 and is represented by  the third
component of the plasma clearance curve (tl/2 = 39/38  days).  As indicated
above fish were made iron deficient by removal of 40%  of their blood volume.
No fish died as a result of the bleeding.  Recovery from hemorrhagic shock was
                                      18

-------
    100
                          <>3
                          k3-.OI76
                         11/2 -39.38 DAYS
                                               .25  .50  .75    1.0
                                                          DAY
  .01
                   345    67
                     DAY
Figure 2.  Three component exponential  curve  of  iron clearance from plasma
of rainbow trout following i.p.  injection of  -'%e.   Each component is
Ri-fiphed separately to show its y-intercept (Co), Tl/2 and slope (k) .  De-
finitions of the symbols used in the  equation are  Given in the text.

-------
rapid and  fish were actively  swimming within  minutes  after  bleeding.   Some
fish gupled air at the surface but  the majority  exhibited normal  activity.
There was  very little change  in  total iron  concentration in the various
tissues except the liver during  the 30 day  period  following the final  bleed-
ing and averages for the different  tissues  did not  differ from those for
controls.  The iron conent of the liver  dropped  significantly after day 11
and was significantly below that of controls.  The  plasma clearance curve
for iron deficient fish was similar to that shown  in  Figure 2.  The Tl/2
for the GI component appeared to be shorter than the  control but  statistically
there was  no difference between  the two  curves.

    The movement of 59pe to and  from the extravascular spaces of  tissues is
indicated  by plots of data in Figures 3  and 4 for  controls  and iron defici-
ent fish respectively.  It is evident that  59pe  activity of tissue is  in
equilibrium with the iron pool of the plasma  some  8 days after administration
of radioiron.  The clearance of ^Fe from spleen,  head kidney and cecal
tissue was rapid and muscle (not shown)  was similar in that respect.   The
specific activity of muscle iron was the lowest  of any tissue and only 3 to
5% of the  initial injected activity remained  in  the total muscle mass  16
days after injection (controls).  In contrast, an  average of 15% of the
initial injected radioactivity remained  in  the liver  of controls over  the
30 day experimental period and the  specific activity  of iron in the liver was
higher than that for any tissue from day 4  through day 30 (Figure 3).  This
indicates  that a significant amount  of 59pe was  deposited and/or stored in
the liver and remained there throughout  the 30 day study period.  In iron
deficient  fish the initial increase  in liver radioiron between days 1  and
4 is similar to that seen in controls but thereafter  the 59pe activity
decreased steadily in the liver of  bled  fish whereas  it remained unchanged
in controls.   There is some indication that the  head  kidney of bled fish is
more active in sequestering ^9pe than those from untreated  controls.   Mean
values for specific activity ( and  % of  initial  dose  given)  are higher in
the experimental fish head kidney than in the controls but  since these values
were not statistically different we  are  unable to state that bleeding  had
any verifiable effect on head kidney tissue.  Thus the experiments with
iron deficient trout tend to indicate that  the liver  contains the most
important labile iron storage pool which can be  utilized for the accelerated
erythropoietic process.   We found that the  spleen iron stores were not
significantly reduced following bleeding thus this storage  pool apparently
is non-responsive to acute induced  iron  deficiency.

    Between 70 and 80% of the injected dose of radioiron was found in  the
red blood cells of bled fish by day  16.   The difference in  red cell iron
uptake between the bled and control  groups was statistically significantly
different.   The peak erythrocyte radioiron  content coincided with the  in-
crease in reticulocyte population (Table 5).  Following day  16 there was
a decline in red blood cell radioactivity similar to  that seen in control
fish.   In cells from control fish initial uptake of 59pe was rapid and
reached a maximum of 55 to 60% on the initial injected activity by day 11.
                                      20

-------
       90n
 CONTROL
                                    ® LIVER
                                    A SPLEEN
                                    23 HEAD KIDNEY
                                    O CAECA
           I 2  4
                             DAY
".Figure 3.   Specific activity of iron in tissues  of trout following.
            intraperitoneal injection of -^Fe.
     150
EXPERIMENTAL
                                    © LIVER
                                    A SPLEEN
                                    0 HEAD KIDNEY
                                    O CAECA
                            DAY

 Figure 4.   Specific activity of iron in tissues  of  iron deficient
            trout  following intraperitoiieal injection  of ^Fe.
                                21

-------
                                    TABLE  5

        HEMATOCRIT, HEMOGLOBIN, MCHC,  AND PERCENT  RETICULOCYTES  OF
                BLOOD  SAMPLES FROM  CONTROL AND  EXPERIMENTAL  FISH.
                         INCLUDED ARE MEAN ±  S.E.  (N).
Days
after
inj_. j
1-30
1
2
4
8
11
16
23
30

Group

Control
Exp.
Exp.
Exp.
Exp.
Exp.
Exp.
Exp.
Exp.

Hematocrit

37.7 + 1.1(48)
16.3 ± 2.2 (6)*
14.2 ± 1.9 (6)*
19.8 ± 2.0 (6)*
26.3 ± 2.3 (6)*
24.7 ± 1.8 (6)*
31.7 ± 1.0 (6)*
33.5 ± 2.7 (6)
35.2 ± 3.8 (5)

Hemoglobin
(gram%)
7.9 ± 0.2(48)
3.5 ± 0.4 (6)*
2.9 ± 0.4 (6)*
3.2 ± 0.3 (6)*
4.7 ± 0.5 (6)*
3.9 ± 0.3 (6)*
5.1 ± 0.1 (6)*
5.9 ± 0.6 (6)*
5.8 ± 0.7 (5)*

MCHC

20.8
21.3
20.1
16.2
17.7
15.9
16.0
17.7
16.5

Reticuloytes
_ 	 _(%) 	
2.75 ± 0.3(28)
6.98 ± 0.8 (5)*

5.00 ± 1.5 (2)
11.07 ± 5.5 (5)
17.65 ± 6.7 (5)*
10.62 ± 2.1 (6)*
5.07 ± 2.1 (6)*
5.00 ± 0.9 (2)
  * Significantly differnet from controls at p = 0.05

    As shown in Table 5, the hematocrit index was lowest 2 days after the
last bleeding but thereafter the red cell volume gradually increased and was
back to the control level by day 23.  Although the red cell volume was
normal by day 23 the hemoglobin concentration remained below normal during
the 30 day study period.  The man corpuscular hemoblobin concentration
(MCHC) decreased from about 21% to 15-16% and remained below the normal
level throughout the remainder of the 30 day experimental period.  Appar-
ently the red cell volume was restored to normal by the release of young
cells, which lacked functional hemoglobin, into the circulating blood.  Even
though the accuracy of reticulocyte counting is poor there was some evidence
of an increase in the immature red cell population by day 11 to replace those
cells lost via bleeding.  It is apparent that the processes involved in
hemoglobin formation are significantly slower than those involved in the
formation and/or release of red blood cells from the erythropoietic head
kidney.

    The total iron binding capacity (TIBC) of the plasma, which is an in-
direct indicator of the transferrin concentration, was measured.  There was
a noticeable and significant increase in the TIBC following bleeding, however,
there was also an increase in the TIBC for controls during the same period.
This nullified any attempt to relate the increase to induced iron deficiency.
The TIBC values for trout plasma (400-700 yg%) are higher than values re-
ported for humans (250-350 pg%) and for the tench (250 yg) reported by
Hevesy et^ al. (1964).  The values for trout blood were reproducible with the
method used and human plasma tested in the exact same manner yielded values
within the accepted range.

    The results of the absorption studies indicated that there was a direct
correlation between the amount of iron injected into the gut segment and the
                                      22

-------
amount recovered in the tissues of the trout.  The dose  given was insuffi-
cient to exceed the absorptive capacity of the gut,  thus it was not possible
to establish a miniumum daily requirement for iron by  trout based upon  in-
testinal absorption.  Possibly some iron enters trout  by absorption across
the gill and uptake by this pathway has yet to be investigated.  However, it
seem unlikely that absorption of iron by the gill would  be of any great
importance since we have shown that iron is readily  absorbed via the gut.
Following i.p. injection, very little radioiron is lost  from trout via  the
feces.  The radioiron in fecal material taken from the intestine of fish
used in this study was initially as high as 4% of the  injected dose, but
by day 2 it was barely detectable and less than 0.08% was found infecal
material after day 8.

    To check iron loss in urine a group of trout which had been on the  same
feeding regimen as controls were cannulated (urinary bladder cannulation)
and urine was collected over a 4 day period following  i.p. injection of
No detectable quantity of radioiron was found in the urine.  Thus the data
from experiments with the gut and kidney indicate that the iron cycle in
fish is a closed recycling system with very little iron  input or excretion
and in this respect is similar to the iron cycle in  humans.

Summary

    Iron metabolism in normal and iron deficient rainbow trout was studied
after intraperitoneal injection of 59]7e<,  in both groups most of the 59pe
was absorbed from the peritoneal cavity within 24 hours  and equilibrium be-
tween the plasma 59pe pool and that of the tissues and attained 8 days  after
i.p. injection.  Liver iron, the main storage pool in  trout, was reduced from
the control level of 185 yg/g to below 100 Mg'g 16 days  after bleeding,
whereas,  splenic iron stores were unaffected.   In iron deficient fish the
RBC 59Fe content increased to 70-80% of the injected dose by day 16 com-
pared to 50% in controls.  This was attributed to the difference in reticu-
locyte count which was 10-20% for the bled fish and  2-3% in controls.   Some
iron accumulated by erythrocytes is temporarily stores as non-heme iron by
these cells.   An average of 15% of the i.p.  injected 59Fe was taken up by
hepatic tissue of control fish and remained there throughout the 30 day
study.  In iron deficient trout liver radioiron was  reduced from a high of
15% on day 2 to less than 1% of the initial dose by day 16 post bleeding.
There was essentially no detectable loss of 59Fe in  the urine or feces of
either normal control or iron deficient fish.
                                      23

-------
                                  SECTION  6

                          RESPIRATORY  PHYSIOLOGY

 CONTROL MECHANISMS  FOR REGULATING THE  FUNCTIONAL SURFACE AREA OF FISH GILLS

     Research on the effects  of  water borne  pollutants  is  gradually moving
 from the realm of determination of  LD-50  concentrations or TLm  values  into
 that of clinical determinations of  the cause  of death  of  test animals.
 Information  on the  toxicological and/or pharmacological action  of  pollutants
 obtained from experiments with  fish may prove to be  very  beneficial in  as-
 sessing the  effect(s) these  materials  may have on  mammals,  including man.
 During  the past several years we have been forced to  interpret data on  the
 toxic action of  materials on fish in light  of information  from  experiments
 with rats or other  'higher'  experimental  vertebrates.  With the advent  of
 many suitable micro-procedures  which can  be used to  obtain physiological  and
 biochemical  data and with more  sophisticated  chemical  and  physical ana-
 lytical systems  available we are in a  position to  extend  our collection of
 information  on the  toxic action of  pollutants on fish.

     Ventilation  of  the gills of  fish maintains at  the  gill surface the  con-
 centration of any pollutant  that is in the  aquatic environment.  It follows
 that  the study of gill structure and function can  contribute valuable in-
 sight into the action of pollutants.   Suspended solids and  sulfite liquor
 from paper-mill  wastes, ions of  heavy  metal salts  from mining industries,
 detergents,  phenols and a variety of other  discharges  from agricultural
 practices and industrial processes  as  well  as specific fish poisons,  e.g.,
 rotenone, are known to affect the gills of  fishes.

     Thickening of the gill epithelium  is  one  of the first changes noticeable
 in fish  that  are exposed to  heavy metal salts, detergents  and phenols.
 Epithelial thickening is usually followed by  fusion of adjacent secondary
 lamellae and  detachment of the  epithelium from the basement  membrane of
 the gill filaments and their secondary  lamellae.   Changes  in the gill
 epithelium during poisoning  by detergents are accompanied by vasodilation
 of the gill arterioles and blood  spaces of  the secondary lamellae,  producing
 extensive vascular stasis or hematomos.  Heavy metal ions are known to  coag-
 ulate mucus secreted on the  gills and  the gills of chronically exposed  fish
 have a thinner epithelium covering  the lamellae than those  acutely  exposed.
 These changes  occur to varying degrees in different parts of the gill so
 that diagnosis must be based on  a substantial  sample of filaments  from
different parts of gills.

    It is important to note  that  even though  histological changes  found
after exposure to various toxic materials have been described in detail,
 they do not necessarily indicate  the preceise cause of death.  Heavy metals
                                    24

-------
may react with enzymes in sill epithelial cells causing formation of lyso-
somes, vesicles and vacuoles in the gill.  Damage to the epithelium may
affect gas exchange, extrarenal excretion or gill ion exchange although os-
motic or ionic imbalance has not been found to be the precise cause of death
after exposure, to zinc or mercury.  It is probable that in many cases where
fish are exposed to toxic materials, death is due to hypoxia resulting from
the detachment of the gill epithelium from the vascular bed of the lamellae
causing a lengthening of the water-blood pathway.  Under these conditions a
significant decrease in the diffusing capacity of the gills would result and
even greatly increased gill ventilation would be unable to meet the oxygen
requirements of the fish.  Water borne toxic materials might also have
serious detrimental effects on blood flow through the gills so that even with
adequate external ventilation the internal perfusion might not be adequate
for transport of oxygen from the gills to the various body tissue where it
can be utilized.

    It is apparent from the above discussion that more detailed experiments
involving use of the isolated gill preparation are in order.  From experi-
ments of this type information should be obtained relating to critical or
detrimental effects that water soluble pollutants, or changes in certain
water quality criteria, might have on freshwater fish.  The objectives of
the experiments discussed below were to  (1) modify published isolated-
perfused gill techniques to permit longer experiments under conditions re-
sembling, as closely as possible, those  found in vivo, and (2) to add to
our existing knowledge of the physiology of the fish gill so that the
preparation could be utilized for studies of the effects of pollutants on
gills.  With respect to the latter we wished to confrim or deny regulation
of functional surface area of the teleost gill and determine the nature of
physiological control mechanisms which would be responsible for regulating
functional surface area.  For these experiments the influx of -^C-urea was
used as a relative measure of gill functional surface area.  Urea is not
metabolized by rainbow trout gill tissue and is not known to be actively
transported by teleost gills.  The method is based on the assumption that
diffusional influx of l^C-urea is limited principally by the extent of
secondary lamellae perfusion, i.e., the  functional surface area of the gill
available for diffusional influx of the  marker.

Methods

    Previously described methods for isolated gill perfusion studies were
modified by using a peristaltic pump to  deliver perfusion solutions and by
collecting perfusate samples in a fraction collector  (Figure 5).  Two in
dependent perfusion channels were used to simulataneously perfuse the second
pair of gill arches from each fish.  All experiments were conducted in a
cold room at 11 ± 1°C.  The peristaltic  pump aspirated perfusion solutions
from polyethylene bottles and delivered  them through PE 60 polyethylene
tubing and the afferent cannulae to the  gills.  Perfusate leaving the gills
passed through efferent cannulae and PE  60 tubing to photoelectric drop
counters above the fraction collector.   The channel B drop counter signal
drove the fraction collector, which was  set to turn after a pre-selected
number of drops had been collected.  Fraction collector turns along with
the individual drop signals from the channel A drop counter were recorded


                                      25

-------
     Figure 5.  Diagram of gill perfusion apparatus.   Solid  lines represent
     perfusion channels and connections; dashed  lines  represent  electronic
     connections.  A and B gill arches  in A and  B perfusion  channels; D
     photoelectric drop counters; FC  fraction collector; FCD  fraction col-
     lector drop count accumulator;  G Grass polygraph; PC pressure  cali-
     bration manometer; PP peristaltic  pump; PS  perfusion solutions; R
     polygraph recording; S 4-way stopcocks; T pressure transducers.

on a Grass 5jy polygraph.  To monitor perfusion pressures, 't' connectors
between the pump and gill arches (10 cm from the afferent cannula tips) were
connected to Statham P23^c pressure  transducers.  A calibration manometer was
also connected through valves and PE 60 tubing to the  transducers.  The can-
nula tips, manometer base, and transducers were  all positioned in the same
horizontal plane to facilitate accurate calibration and measurement of
pressures.  Output from each transducer was recorded continuously on the
Grass 5j) polygraph.

     The gill cannulation procedure has been described in detail (Bergman
et al.  1974).   After cannulation, the pump output to each arch was  increased
to 0.5 ml/min and held constant throughout the remainder of  each experiment.
Since only about 70% of the filaments were perfused in these gills  the
0.5 rnl/min flow rate was consistent with the expected single arch flow rate
(0.75 ml/min,  assuming equally divided  flow to all eight gill arches) cal-
culated from cardiac output estimates for resting 200-400 g  rainbow trout.
Drop size calibration for the photoelectric drop counters permitted calcula-
tion of flow rate through each gill arch.   In preparations where severe leak-
age developed  the measured flow rate fell below pump output and data from
these experiments were discarded.   Pressures usually ranged within  the limits
reported for ventral aortic blood pressure in rainbow trout.   To mimic

                                     26

-------
systemic resistance, efferent pressure was set at 15 mm Hg by elevating the
efferent tube outflow 20 cm above each gill arch.

     Perfusion solutions consisted of vasoactive drugs or hormones added to
Ringer solution.  At 11°C the pH of this solution was 7.5 which compares
closely to mixed venous-arterial blood pH for the same species and temperature
in this laboratory.  However, this value is lower than the arterial blood pH
of 7.9 reported at 10.5°C by Randall and Cameron (1973) for rainbow trout.
Before vasoactive agents were added, the perfusion solution was vacuum fil-
tered through a 0.22 ym millipore filter.  The solution was then vigorously
shaken to assure atmospheric equilibration.

     After cannulation, gill arches were perfused for an equilibration period
of about one hour  before experiments were begun.  During this time each gill
arch was perfused with the same solution that was to be used in that arch
during the initial period of the experiment.  After the initial period, drug
or hormone concentrations were increased in 10-fold steps in the experimental
arch while they were either not added or held constant in the control arches.
At the end of a few experiments under each protocol the vasoactive perfusion
pressure and l^C-urea flux to control values.  Starting with the initial
period, perfusate fractions of equal volume were collected directly into
vials for determination of -^C-urea activity in a liquid scintillation
counter.  The l^C-urea counts from the initial experimental period for the
arch.  The sample activities from the initial period were averaged for each
gill arch, and the average was defined at 100% of initial activity.  The
activities of samples collected during the remainder of the experiment were
then converted to a percentage of this mean initial activity.  To demonstrate
qualitative pressure responses to the drug and hormone treatments, typical
perfusion pressure records were selected to represent each type of experiment.
Where quantitative expressions of pressure responses were appropriate, the
data were treated in the same way as the l^C-urea data.  All statistics were
calculated with log transformed data, and the means of 95% confidence limits
were used for all comparisons and Tukey's w-procedure was used for multiple
comparisons among means.

Results and Discussion

     The influx of -^C-urea into isolated-perfused gills was altered
markedly by the vasoactive drugs and hormones used in this study.  The
results of a typical epinephrine (EPI) experiment are depicted in Figure 6.
In the figure the open circles = control arch fractions and x = experimental
arch fractions.  Vertical dashed lines delineate experimental periods with
EPI concentrations in the experimental arch shown above.  EPI concentration
in the control was 10"7M throughout.  The solid horizontal line represents
the control initial activity for both gill arches.  Typically, the percent
of initial activity increased (or decreased depending upon the agent used)
toward some new steady-state value in response to step-wise changes in the
vasoactive drug concentration.  For all experiments following one protocol,
steady-state percent of initial activity values are tabulated for the drug
concentrations used.
                                      27

-------
              TOO

              600
            b 500
            *20O

  100
                  lO'7 M
                    EPINEPHRINE CONCS.

                  IO'6 M     IF* M     IO'7 M
                                ,*>"*''
     •ffir—
    0 1.0
                         L5     2-0      2.5     3-0     3.5
                              POST CANNULATION TIME (MRS)
Figure 6.  Results from a typical epinephrine experiment.   See  text  for
details.

     The results from epinephrine (EPI) experiments  (Figures  7  and  8)  were
qualitatively the same as those found  for norepinephrine  (NEPI),  although
a narrower concentration of EPI was used.   EPI,  like  NEPI,  is a mixed  agonist
which can stimulate both a and 3 adrenergic receptors.  Because -^C-urea in-
flux data were less variable in the EPI experiments than  in NEPI experiments,
more clear-cut differences were evident when the EPI  effect was blocked  with
phenoxybenzamine (POB) and propronolol (PROP) as noted  in Figure 8  below.
The unblocked 10~5M EPI induced l^C-urea influx  is significantly greater
that  seen in the control or blocked gills. * Influx in  the single block
experiments (a or 3) was still significantly higher than  in the controls,
and double blockade (a and £) of the EPI effect  was necessary to reduce  in-
flux to the control value.

     Acetylcholine (ACH) significantly reduced ^C-urea influx  when perfused
alone (Figure 9) or when perfused along with 10~5M EPI.   The  reduction was
greater in experiments where ACH was perfused without the EPI.   Stimulation
of a and/or 8 adrenergic receptors with catecholamines  or pharmacological
drugs in this study produced marked and usually  significant Increases  in
14c-urea influx, whereas ACH significantly  reduced influx of  the marker.
The most tenable explanation for the observed changes in  l^C-urea influx and
branchial vascular resistance is that  alterations in  the  internal perfusion
pathway resulted in changes in functional surface area  of these gills.  It
could be aruged that the vasoactive agents  may have affected  the permeability
of the gills to l^C-urea but it is doubtful that the  changes  in permeability
could account ror the effects these agents  had on branchial vascular resistance
(Figures 10 and 11).  This is especially true in the  case where, with  opposite
effects on resistance, stimulation of  either a or 3  adrenergic  receptors
caused an increase in 14C-urea influx. We  conclude  that  these  vasoactive
                                      28

-------
                      TOO

                    t 600
                    o 900
                    3 400
                      300
                      £00
                      IOO
        1 CONTROL EPI


     [~~3 EXPERIMENTAL EPI
                                 £
              i?
            f
                                     icr6            to-3
                             EPINEPHHINE CONCS. (MOLES/LITER)
Figure 7.  The effect  of epinephrine on l^C-urea  influx.  During  the  initial
period (not  shown)  both control and experimental  gills were  perfused  with
10~^M epinephrine.   The mean,  957» confidence interval, and N value  are
shown for each control and treatment from successive experimental periods.
Significance  for  paired comparisons; *0.01  P  0.05; **0.001  P  0.01;
***  0.001.
7OO,
£ soo
5 300
J 400
z SOO
u.
0 COO
J*
100





--

•4m

•i?

Do
LJ E
1 1
^i 8 jr
M_i_i
                                                            OL EPI    ^ W BLOCKADE

                                                        EXPERIMENTAL EPIS /3 BLOCKADE

                                                                     ^ Ota/3 BLOCKADE
                            EPINEPHRINE  CONCS.  (MOLES/LITER)
Figure 8.  The  effect  of adrenergic blockade on epinephrine  induced l*C-
urea influx.  During the initial period (not shown) all arches  were per-
fused with 10~7M EPI.   In blocked arches, the blocking drug(s)  was  per-
fused throughout the entire experiment.  All values are from the  same ex-
perimental period, a block = 1Q-5M POB; 3 block = 10~5M PROP.   The  means,
957» confidence  interval and N values are shown.  The means  listed below
which are not underlined by the same line are significantly  different (0.05)
          Control     a + 8         a	B	Exp.
           EPI
           72
block
100
block
187
block
189
EPI
564
                                      29

-------
                                     **
140-1
120
^ "
§80
*g
z
u.40
o






















6
••



:;.-•
M


1

;i
1












6
*




««•
x. 1
';«-
•




1
y

•


^^



6





1]
                                                    1 CONTROL (RINGERS ONLY)


                                                     EXPERIMENTAL  ACH
                 10""        10*'        1C"
           ACETYLCKOLINE CONCS. (MOLES/LITER)


 Figure 9.   The effect of acetylcholine on l^C-urea influx.   During the
 initial period (not shown)  both control and experimental gills were perfused
 with  Ringer solution only.

 agents had little effect on permeability and that the observed changes in
 14C-urea uptake reflect  changes in the relative functional  surface area of
 the gills.   For any series  of experiments with a vasoactive agents, the
 pressure effect was qualitatively consistent,  but the sensitivity of different
 preparations to a drug or hormone varied.  This may have resulted from.
 differences in prevailing vascular tqne.

      Perfusion pressure  was reduced by NEPI (Figure 10 ), although in a few
 experiments a transient  pressure increase was  seen first (Figure 10B).   At
 1Q-5M NEPI  the average percent of initial pressure (88%) was significantly
 lower than  in the 10~9M  NEPI  controls.   Although not statistically signifi-
 cant,  reductions   in perfusion pressure were observed in some preparations
 with  NEPI concentrations as low as 10~8M.  Figure IOC,  D and E show the typi-
 cal pressure responses when adrenergic  blockers were perfused along with NEPI
 just  as  in  the unblocked gills.   After  8 blockade,  however,  a pressure  in-
 crease was  usually unmasked and when both a and $ blockers were sued no pres-
 sure  change resulted from the increased NEPI concentration.   Qualitative
 pressure responses in gills perfused with EPI  alone or  with  adrenergic  block-
 ing agents  were identical to  those seen in the NEPI experiments,  hence  these
 data  are not  presented.   When the ACH content  of the perfusion solution was
 increased from 10~Jl to 10~7M  a slight but significant increase, in perfusion
pressure was  noted (Figure  11A).   Another ten-fold  increase  to 10~°M ACH
resulted in a very significant  pressure increase (Figure 11 B).   The effects
of four different  drugs  on  the  increased  perfusion  pressure  induced with
10-6M ACH were  also  noted.  The  muscarinic and nicotinic (ganglionic) block-
ers, atropine  and  hemamethonium respectively,  both  eliminated  most  of the
pressor effect  of  ACH.   The a adrenertic  block POP  completely  eliminated
the pressure  increase whil  phentolamine,  also  an a  adrenergic  blocker, had
little effect.
                                     30

-------
                     E
o
CO
£ 60r
o. 30«-
I
                       30L         i    ,	,
                                   1    I  MIN
 Figure  10.  Typical effect of norepinephrine on perfusion pressure  in  the
 absence and presence of adrenergic blockers.  Arrows indicate where NEPI con-
 centration was increased  from 10~6 to 10~5M.   (A) and (B) NEPI alone;  (C)  a
 blockard of NEPI with  10~5M FOB; (D) 3 blockade of NEPI with 1CT5M PROP; (E)
 a  and 3 blockade of NEPI with 10~5M FOB and 10~5M PROP.
 Summary

     The physiological viability of the perfused gills in this study was
probably due partly to the use of filtered perfusion fluid  (Rankin and Maetz,
1971) as well as pulsatile pump-perfusion, which appeared qualitatively
superior to perfusion with a hydrostatic head  (Richards and Fromm, 1969;
Rankin and Maetz, 1971; Randell e£ al. 1972; Wood, 1974) or a constant pres-
sure syringe pump (Shuttlcworth, 1972).  The appearance of the gill arches
usually remained good during and after perfusion except that secreted mucus
accumulated on the surface of a few gills.  Although no edema was evident,
colloid osmotic, pressure was low in our perfusion fluids and some edema could
have developed.  However, branchial vascular resistance remained constant and
vasoactive agent effects were reproducible for as long as 12 hours.  With
flow rates which approximated the expected single arch flow rate in intact,
resting rainbow trout (Stevens and Randall, 1967a), the perfusion pressures
in these experiments were generally within the range of ventral aortic blood
                                     31

-------
                  en

                 I60F
                    SOL
                 LU
                 CC.
                 tO
                 Ul
                 a:
                 o.
                                      ! MIN
 Figure HA and B.   Typical effect of acetylcholine on perfusion pressure.
 Arrows indicate where ACH concentration was increased.  (A)  ACH increase
 from 10~8 to 10~7M; (B)  ACH increase from 10~7 to 10~6M.

 pressure  reported  for this species (Stevens and Randall,  1967b).  Pressures
 in  excess of the normal  range were observed only in the gills perfused with
 10~% ACH.   Significant  effects on resistaace and 1/lC-urea uptake were seen
 at  NEPI and EPI concentrations which were comparable to the maximum plasma
 catecholamine levels reported for distrubed rainbow trout (Nakano and
 Tomlinson,  1967).   However, the adrenergic agonist concentrations which
 produced  threshold effects on resistance were sometimes 100-fold higher than
 previously reported for  the same drugs and hormones in isolated teleost gill
 preparations (Rankin and Maetz, 1971; Wood, 1974).  The reasons for this are
 probably  related to differences in pH, temperature, perfusion pressure and
 composition of the perfusion fluids.

      The  physiological control mechanisms responsible for adjusting the
 resistance to flow in teleost gills are probably both hormonal and neural in
 nature.   The catecholamine-induced decrease in branchial vascular resistance
 increase  in functional surface  area  seen in these experiments agree with pub-
 lished  observations on the  resistance  (Keys and Bateman, 1932; Ostlund and
 Fange,  1962;  Reite,  1969) and blood  pathway effects of catecholamines (Steen
 and  Kruysse,  1964;  Richards  and Fromm, 1969).  These findings, combined with
 the  observation  that  plasma  catecholamine concentrations are elevated during
 exercise  (Nakano and  Tomlison,  1967),  strongly support hormonal control of
 gill blood  flow  and functional  gill  surface area.  Wood (1974) has proven
 the  involvement  of  3^ adrenergic  receptors in this catecholamine effect.  Al-
 though  the  critical catecholamine potency-series experiments have not yet
 been reported  as proof, evidence  in  this and other studies suggest a ad-
 renergic  receptor  involvement in  the  catecholamine effect as well.
 THE  ROLE  OF  FISH GILLS IN HEAT EXCHANGE

     Another series of experiments dealing with heat exchange of fishes  were
conducted using isolated-perfused gills as well as isolated-perfused heads
of rainbow trout.  Investigations dealing with rates of thermal exchange in
                                     32

-------
fish have been few in number.  Stevens and Fry (1970, 1974) have reported
values for a coefficient of temperature exchange in a variety of species,
however, there appear to be no data directly in terms of heat exchange by
a teleost.  Fry (1967) speculated that the gills play a major role in overall
thermal exchange  but pertinent experiments to substantiate this hypothesis
are lacking.

     In attempting to quantitate heat exchange by the gills of rainbow trout
a generalized equation for a tubular heat exchanger was adopted (Kay, 1963):

                            Q - hAAln                        (1)

Where Q_= total heat flux in cal min"1; h = heat transfer factor in cal
min~-'-0C  cm" ; A = surface area in cm^ and AT^n = logarithmic mean temperature
gradient in °C.  The value of h will be determined by three factors:  (1)
convective transfer to the inner wall of the vasculature, (2) conduction
through the vessel wall and (3) convective transfer from the gill surface to
the environment.  The surface area will be determined by the mean length and
mean diameter of the vasculature.  Because we were unable to accurately de-
termine a value for surface area (A), Equation 1 was solved for the quantity
hA following determinations of Q and T^n using Equations 2 and 3 respectively
(Kay, 1963):

                          Q   - f Cpp(Ti-T0)                 (2)

                          Tln      Tj - TQ
                                .  Ti - To
                                   TO - Tb                    (3)

Where f = perfusion flow rate in ml min  ; Cp = specific heat of perfusion
fluid (= 0.962 cal gm-l'C"1 for 0.1% NaCl, CRC Handbook); p = density of
perfusion fluid flowing into the gills, out of the gills, and of the bath
respectively.  Equation 3 represents the mean temperature gradient for a
tube of uniform dimensions and may not be rigorous for the case of the
branching vasculature.  However, this equation provides the best means of
assessing the mean gradient driving heat across the gill.  Also, with uni-
directional ventilation, the increase in temperature of the ventilatory
stream as it passes the gills may be accounted for an equation similar to
(3), but makes an insignificant contribution to the results reported here.

     The experiments performed were designed to provide information on (1)
the magnitude of heat exchange occurring in isolated-perfused gills of
rainbow trout, (2) the effect of changes in vasculature gemoetry (i.e.,
changes in functional surface area) on heat exchange and (3) the effect of
ventilation on heat exchange.  Finally, a simple model is evaluated in an
attempt to relate gill heat exchange to the whole body thermal response.

Methods
     The second gill  arches from 200-400 gm rainbow trout  (Salmo gairdneri)
were prepared for cannulation as outlined by Bergman et^ a±.  (1974).  Inflow,
outflow and bath temperatures were measured (+ 0.05°C) with  thermistors
(32A7, Victory Engr.,) using a continuous balancing Wheatstone bridge re-

                                      33

-------
corder.  Inflow and outflow thermistors were  inserted  in  the  cannulas  to
within 1 cm of the supportive branchial bar of each gill  arch.  Heat ex-
change was determined in gills perfused at 0.5 ml min~l by  a  Ringer solution
alone or Ringer solution which contained 10~^M epinephrine  (EPI).  With
T± held constant at 10 + 1°C, data were obtained and Equation 1 solved for
hA at steady state values for Tfo of 2, 5, 8,  13, 16, and  19°C.  Statistical
treatment of the data was performed using the Student's _t_ test for sample
means.

     For whole head preparations fish were decapitated just posterior  to
the opercula and the intact branchial basket  was allowed  to clear of blood as
described by Bergman et al. (1974).  A cannula with a  thermistor probe
threaded through the lumen to the cannula tip was inserted  into the ventral
aorta and secured with a ligature around the  bulbus arteriosus.  A similar
cannula was placed in the dorsal aorta and secured by  a ligature around the
dorsal aorta and spinal column.  A third ligature was  placed  around the cut
end of the esophagus.  The head was transferred to a temperature regulated
water bath and a ventilatory flow (Vg) was diverted into  the  mouth via a
flowmeter and varied between 40 and 130 ml rain"1 (Figure  12).  Perfusion
solutions were delivered by a positive displacement type  pump with an  in-
dependently adjustable stroke volume and rate.  Perfusion pressure was
monitored continuously with mean perfusion pressure of all  control experi-
thermistor for a reliable measure of To precluded the  use of  an outflow
resistance resulting in larger than normal perfusion pulse  pressures.  Iso-
lated heads were perfused with four different solutions:  Ringer solution
alone and Ringer solution which contained 10~5M EPI, 10~8M  or 10~7M ACH.
Heat exchange was determined in the steady state with  an  inflow temperature
(Ti) of 12 + 1°C and bath temperature (Tt,) of 8 + 0.1°C.  The thermistor
probes were calibrated against a differential thermometer and temperatures
recorded + 0.01°C.  All experiments were preceded by a 30 min equilibration
period.  Changes in drug concentration were followed by a 30  min equilibration
period and changes in perfusion flow were followed by  a 15  min equilibration
period.
Results and Discussion

     In the isolated-perfused heads, the influence of  ventilatory flow
on heat transfer was evaluated and is shown for a representative experiment
in Figure 13.  Values for hA at Vg below 400 ml min~l were not determined
because of marked variations in hA due to (1) extraneous  heat picked up by
the fluid passing through the flowmeter which was in contact  with ambient
temperature (20°C) and (2) possible mixing of the low  temperature bath with
the warmer ventilatory fluid by back flow into the opercular  chamber at
low Vg.  Despite these experimental limitations, two constraints may be
placed on the values of hA with increasing Vg.  Since Vg  influences only the
convective heat loss from the gill surface, as Vg increases a point is reach-
ed where heat transfer is limited solely by the transfer  of heat to the inner
vessel wall and the conduction of heat through the vessel wall and hA  becomes
independent of Vg.  This constraint represents0an asymptotic  maximum of hA
(hAmax).   In a similar manner, for 0 ml min"1 Vg there will ideally be no
heat lost from the system in steady state and hA must be  zero.  We chose
to use a rectangular hyperbolic function to predict hA for a  given Vg:

                                      34

-------
   FLOWMETER
              TEMPERATURE


               REGULATED


                 WATER


                 BATH
Figure 12.  Diagram of apparatus used to perfuse isolated heads of rainbow

trout.  See text for details.
                              hA  = hA    V
                                     max  g
                                   V  - Vasym

                                    O
(4)
The  term -Vasym also  represents  the value  of V   necessary  to reach one-half

of hA    (Vn  c     ) and Equation 4 becomes:
     max   005 max      M
                              hA = hA    V
                                     max  g
                                   Va
(5)
                                         0.5 max
From equation 5,  the  transfer maximum  (hA   ) and ventilatory  flow at  one-

half the transfer maximum  (Vn _    ) were obtained  from  a  linear  regression
                            u•j max
of hA    -1 on V -1.  Changes in hA    and V_ _     were evaluated using  a  2

way analysis of variance and when appropriate"a §?udent-Newman-Keuls test

for significance group means  (Sokal and Rohlf, 1969).




Values  for heat transfer  (hA) In the isolated-perfused gills are shown in

Table 6.  Perfusion with  ICT^M EPI significantly increased hA.  There were

no uniform or systemic alterations In  hA  or perfusion pressure associated

with changes in bath  temperatures.
                                      35

-------
                  hA
               cal/sec/°C
                0.4 J
                0.3-
                0.2-
                0.1-
                                           max
                                               - 25.59 cal
                                        VQ.5 max " 205 ml nln
                                                        -1
                               400
        —,	1
         800        1200
              ml/min
Figure 13.  Experimental data for heat transfer plotted as a function of
ventilatory flow (Vg) for one isolated head perfused at 20 ml min"1.  hAmax
and VQ.Smax were determined by a double reciprocal linear regression and  the
solid line represents the theoretical curve.

                                  TABLE 6

       HEAT TRANSFER  (hA) OF ISOLATED-PERFUSED SECOND  GILL ARCHES
               OF RAINBOW TROUT.  PERFUSION RATE - 0.5 ML  MIN

     _P_e_rfusion Fluid
       Heat transfer (hA)cal min"
Ringer solution
10~->M EPI in Ringer solution
                  0.774 + 0.024*
                  0.858 + 0.024*+
*X + SE; + p less than 0.05

     Values of hAmax and V0.5jnax at perfusion  flow rates  of 16 and 20 ml
min"1 for the initial control periods  of  all experimental groups  are summarized
in Table 7.  The effects of altering perfusion flow are shown in  Figure 14.

                                   TABLE  7

          HEAT TRANSFER FOR ISOLATED-PERFUSED  HEAD OF  RAINBOW TROUT
                   DURING CONTROL PERIODS.  VALUES (X  + SE,  n)
                 ARE GIVEN FOR FLOW RATES OF 16 and 20 ML MIN"1.
                                           Perfusion  flow rate
                                    16 ml min"1	  20 ml  min'1
      (cal min"1^"1)
 0.5max
21.27 + 0.57 (21)
 144  + 17   (21)
24.79 + 0.77 (21)
 183  + 23   (21)
                                     36

-------
Increasing the perfusion flow increased
Also, at a given perfusion flow rate,
a 20% change in stroke volume.
                                              but had no effect on V
                                            and Vo.5max were unaffected by
    The effects of drug infusions are shown in Figure  14 (below) for a
perfusion flow of 20 ml min*"1.  Each experimental group was followed
through control, drug infusion and post-control periods.  During
each period hA was determined for perfusion flows of 16 and  20  ml min~   at
a stroke rate of 50 min"1.  VQ.Smax was unaltered for  any of the experiment-
al treatments.  With infusion of 1Q-5M EPI hA^x was significantly  increased
above the values obtained during the pre- and post-infusion  control periods,
whereas, perfusion of solutions containing 10~8M and 10" 'M acetylcholine
had no effect on hA^x-  These results are qualitatively the same at both
perfusion flow rates.  When perfusion flow was increased from 16 to 20  ml
min"1 during constant infusion of either EPI or ACH the change  in hA^x was
identical to that seen in the control experiments  (Figure  15).
                  'man
                col/»K/-C
                                             20 ml/mln
                                             50

0.5-
0.4-
0.3-
0.2-
0.1-
t
T



JL
5



5


5





S

1
fft
I
i
4




JL
5

I
5
.

4





6

1
!!
i!:i
6
|

6




1P<0.05
mi/milt 1 1 Ring
               300-
               200-
                ICO-
                                                 ,-5.
T




I

9

11

9


5




|

9

^2 IOJMEpi
I
1
1
&
n


4





i

5

H] IO"8M Ach
{H IO"7M Ach
I

,s;
:-
I
4

flJlT
ji!»l
« 6 C
ijijii
 Figure 14_   The effects of vasoactive agents on hAmax and VQ.Smax
 (X + SE)  of the perfused head at a perfusion flow of 20 ml min~l.  + signif-
 icant change in hAmax during 10~5M EPI from pre- and post-infusion control
 periods.
                                      37

-------
  cal/sec/eC
                                                  t P<0.05
ml/min
0.5-

0.4-
0.3-
0.2-
0.1-
t
_ T





4


4






4



4





50 40 60 40 beat
16 20 ml

T



4
T


4



T T


4


4



-200

- 100

/
min 50 40 60 40
•nin 16 20
Figure 15.  The effects of alterations in perfusion flow and heart rate
on hAmax on the left and Vo.5max on the right (X + SE) for the isolated-
perfused head (n = 4) .   + significant change in hA^x at perfusion flow of
20 ml min"1 over 16 ml min"1.

    In interpreting changes in heat transfer  (hA) following experimental
manipulations, the basic determinants of h and A must be considered.  These
determinants are not independent.  A change in vascular dimensions ^will
alter not only the length and diameter of the vessels, but wall thickness
and perfusate velocity such that the value of both h  and A are affected.
The effects of ventilatory flow on external convection will be determined
by the velocity of the stream as well as its relation to the path of in-
ternal perfusion.
    The use of values for maximal heat transfer  (hA^x) and ventilatory
flow at one-half hAmax  (Vo.5max) allows some of  these factors to be evaluat-
ed independently.  An increase in Vo.5max without a change in hAmax would
result in a decrease in hA at any given $g.  Assuming the internal dimensions
and hence internal convection and conduction are stable during the determi-
nation of hA^x and V0.5max' the only alteration which would account for  the
increase in V0.5max would be a decrease in convective loss from the gill
surface.  A decrease in external convective loss can be brought about by
two means:  (1) a conformational change of the gill in the ventilatory
stream or (2) a change  in the mean ventilatory velocity over the perfused
lamellae.  If a gross conformational change of the gill has not occurred,
the patterns of ventilatory velocity over the gill should be unaltered.  The
increased VQ.Smax would then indicate a change in perfusion pathway away
                                     38

-------
from the better ventilated secondary lamellae with a resultant increase in
ventilation-perfusion inequality.  In contrast to Vo.5max» hAmax provides
a VK independent measurement of heat transfer and should be altered only by
changes in internal thermal convection from the perfusate to the vessel wall
and thermal conduction through the vessel wall.  By evaluating the two para-
meters of the hyperbolic function relating hA to Vg, independent changes in
either V0 5raax or hAmax wil1 provide information regarding alterations in
(1) the ventilation-per fusion relationship of the gill  (hAmax) .  Simultaneous
changes in hAmax and Vo.5max would indicate some indeterminate combination
of these factors.
     The  theoretical  determinants  of hA^x  could  be  correlated  to  experimental
 observations by noting  the  effects of  increasing perfusion flow.   The  in-
 creased  flow and  resultant  increase in perfusion pressure  may  cause  an in-
 crease in mean perfusate velocity and/or distension of  the vasculature.
 These changes alter  the internal  convection  and  conduction processes and
 result in a significant increase  in hAmax  in response to  the increased per-
 fusion flow  (Figure  14).  We were unable to  detect  any  significant alter-
 ation in VQ_5max  following  our  experimental  manipulations. This  may be due
 in  part  to a basic insensitivity  of V0.5max  to detect ventilation-perfusion
 changes.  However, the  importance of ventilation-perfusion matching  cannot
 be  overemphasized when  evaluating exchange processes in any respiratory
 organ.

     Holeton and Randall (1967)  reported that trout  exposed to  a hypoxic
 enviroment exhibit a marked bradycardia with no  alteration in  cardiac  out-
 put.  In this case,  the animal  should  attempt to maximize  oxygen  transport
 by  minimizing ventilation-perfusion inequalities.   The  evidence for  a  cap-
 activie  'bleb' in the afferent  filamental  vessels (Fromm,  1975 and see
 below) gave rise  to  speculation concerning the frequency  dependence  of the
 distensible gill  vasculature.   If the  gill vessels  could  exhibit  a frequency
 dependent nature  to  allow for passive  changes in perfusion pathways  in
 response to changes  in  heart rate, the hypoxic bradycardia could  give  rise
 to  a passive redistribution of  perfusion to  aid  in  oxygen  transport.  Thi
 is  apparently not the case  since, as shown in Figure 14,  a 30° change  in
 heart rate  (40 to 60 min"-1-) did not alter  the Vo.5max or hA^y of the  gills.
 However, the diffusion  distance and membrane permeability  characteristics
 than heat,  thus there are many  changes that  may  be  occurring in vascular
 flow and the lamellar membrane  that will alter oxygen uptake,  but have an
 insignificant effect on thermal exchange.  If a  passive vascular  response
 to  changes  in heart  rate occurs,  this  could  be coupled  with an active  mech-
 anism(s) to efficiently regulate  ventilation-perfusion  ratios.

     Along with the possibility  changes in  the gill  vasculature, there  have
 been several investigations dealing with active  regulation of  the gill
 vasculature  (Bergman e£ al. , 1974; Steen and Kruysse, 1964).   Bergman
 et  al. ,  (1974) concluded that EPI increased  functional  surface area  for
                                      39

-------
         uptake and that it was decreased by acetylcholine.  The primary
changes in these experiments were attributed to perfusion pathway changes
which should be reflected by alterations in the hA^x value for heat ex-
change.  As shown in Table 6 and Figure 15, 10-% EPI significantly in-
creased hA in the isolated perfused gill and hAmax in the isolated perfused
head.  This increase is not as dramatic as that seen by Bergman et_ ja. (1974),
and is indicative of the relative sensitivities of the two markers used.  In
the brachial basket the major change in vascular dimensions is presumed to
take place at the secondary lamellae which would greatly alter the l^C-urea
uptake and also change heat transfer.  However, the remainder of the large
vessels in the gill are readily exchanging heat but not ^C-urea indicating
there is a greater effective surface area for thermal exchange than for mo-
lecular exchange.  This could explain why there was a much smaller increase
in total heat flux (13%) than total 14C-urea flux (400%, Bergman*£t al.,
1974) for the same perfusion pathway change.  Acetylcholine at concentrations
of 10~8M and 10-?M did not significantly alter hA^x-  There are several pos-
sible reasons for this lack of supportive evidence of ACH modulation of the
exchange process:  (1) the relative sensitivities of a molecular versus ther-
mal diffusion phenomena to small changes in vascular geometry, (2) the ini-
tial control state of the gill vasculature and (3) the possibility of a dif-
ferential response by the gill arches to ACH that would mask overall changes
in the brachial basket but not in the isolated gill arch.  This final point
has received little attention and the redistribution of perfusion pathways
at the gill arch level as well as the lamellar level should be considered for
all vasoactive agents.

     Aside from the basic quantitation of heat exchange by the gill, a sim-
ple model  (see Figure 16) was constructed to determine the relative contri-
bution of the gill to whole body heat exchange.  The body of the fish con-
sists of a cylinder with thermal conductivity  (k) of 1.4 x 10~J cal°cnr*  (k
for water, Crc Handbook) and an internal core of 0.5 cm radius maintained at
some temperature (T^ above ambient  (Tb).  Blood is pumped through an ex-
changer and returned to the body at T0.  To simplify the model it was as-
sumed  that:   (1) there is no convective mixing in the body, (2) there is no
interference with thermal conduction from the body due to  the flow of blood
through the gills, and  (3) there is no heat conducted from the ends of  the
cylinder.  Thus, the total heat flux (Qtot) will be the sum of the convective
loss from  the gill  (Qcv) and the conductive loss from the body (Qcd)•

                      Qtot - Qcv + Qcd                         (5)

Qcd  can be determined using Equation 6  (Bennett and Meyers, 1962).

                  Qcd = k  2  lATi/(ln r2/ri)                   (6)

Where  1 =  length of cylinder;  TI - internal radius of cylinder; r2 -  external
radius of  cylinder  and  T±  = T1 - Tb.  Qcv  can  be determined by using  Equation
2.   The solid  lines in  Figure  17 represent  a plot of  the percent of total
heat flux  (%  Qtot)  exchanged via the model  gill as a  function of perfusion
flow when  the %  equilibration   (1  -  To/Ti)100   is held at 100% and 50%.  The
maintenance  of a constant  % equilibration  in the face of  increasing perfusion
flow will  probably  not  occur and the actual model response should  tend  to

                                      40

-------
Figure 16.  Diagram of model used to evaluate gill heat exchange in terms of
whole body heat exchange,  r^ = 0.5 cm, r<^ = 2.0 cm.  Other terms as given
in the text.
flatten as the heat delivered by the perfusion fluid exceeds the exchange
capability of the gills.
f C
     Experimental data were evaluated in a similar manner.  Since Q
   pP
                                                                   cv
          ~ T0) (Equation 2) and Qcv = hAATln (Equation 1) if we express To
as a fraction, X, of T-^ then:

               f Cpp (Tj - T0) = hA   T± -
                                  In  Ti - Tb
                                          - Tb
and:
In
                      - Tb
                  XT± - Tb
                                 hA (T± -
                                          - XT±)
Subtracting Tb from all temperatures, i.e., measuring all temperatures rela-
tive to Tb:
                                     41

-------
     In
                              (Ti - Tb)

                            X(T± - Tb)
 hA

f C
                               PP
Solving for X:
                              hA
                   X = e     f C
                                PP
                                          - hA
and:
% equilibration = (1-e
   ) 100
Using this value of X:
and:
                                            f  C
                                               PP
                       Qcv = f Cp          

                       _   f Cn        (1  - X)
                                                                (7)
                                       (1  -  X)  + k  2 ir 1  100   (8)

                                                 In
                                                             100
                                               __ _.	Equilibration
                        T	T
                        8       12       16
                         Cardiac  Output    ml/min
 Figure 17.   Graph of percent of total heat flux (% Qtot) from the model gill
 with  increasing perfusion flows.  Solid lines indicate the model response at
 a  constant  % equilibration, 50 and 100%.  Dotted line indicates model re-
 sponse using experimental data.
                                      42

-------
     For the specific model of Figure 16,  Equation 8  was  solved using  experi-
mental data.  In Figure 17 the data point  at 0.5 ml min~l is  the value ob-
tained for a single gill arch (Table 6)  and  the point at  4 ml min~l  is extrap-
olated assuming 8 independent gill arches  all exchanging  at the same rate as
the single arch perfused at 0.5 ml min~l.  The points at  perfusion flows of
16 and 20 ml min~l were calculated using the hA^jj values given in Table 7.
The dotted line connecting these data points lies between the 50 and 100%
equilibration curves for the idealized model and % Qtot is leveling  off with
higher flow rates.

     The model may also be used to determine the influence of gentilation on
% Qtot exchanged at the gills.  Values of  hA as a function of Vg were  deter-
mined using Equation 5 and substituting  for  hA^.,,. and V0.5 max at perfusion
flows of 16 and 20 ml min~l.   Values for % Qtot were  obtained by solving Equa-
tion 7 as a function of V  and using Equation 8 to determine  the % QtOf
marked effect on % Qtot as Vg is increased is shown in Figure 18.  Several
cautions should be placed on the interpretation of the model  responses.
   %Q
       tot
   lOO-i
                                                                20ml/min
                                                                16 ml/min
                    400
  1	T
800 .        1200
     V  ml/min
1600
                                                                         o
Figure 18.  Model prediction of % Qtot  as  a function of ventilatory flow (Vg)
using experimental data at perfusion flows of  16 and 20 ml min~l.

First, we are only able to estimate the effect of perfusion flow on the max-
imal rate of heat transfer by the gills, the In vivo state will lie at some
point below the dotted curve of Figure  17.  Second, we are only able to esti-
mate the effect of a continuous ventilatory stream at perfusion flows of 16
and 20 ml min~l.  Data for Intact animals  may  be at some point below the
                                     43

-------
curves of Figure 18 due to lower cardiac outputs and the complicating factor
of a phasic respiratory flow.  Finally, the model of the fish body as being
a purely conductive cylinder will be complicated by the convective mixing of
the blood.  This factor will tend to raise QC(j and lower all the values of
gill heat transfer when expressed as % Qtot.  The model is primarily illus-
trative of the maximum relative contribution the gills may make to whole body
heat exchange.  The actual contribution to heat exchange by the gills awaits
further investigation.

Summary

     The heat exchange has been evaluated in isolated-perfused sec-
ond gill arches and the intact branchial basket.  The presence of 10~5M epi-
nephrine increases the transfer maximum of the gill suggesting a change in
perfusion pathway and/or vascular dimensions.  Changes in perfusion flow al-
ter heat exchange by the gill which is again due in part to changes in vas-
cular dimensions.  A model for the evaluation of the gills in relation to
whole body heat exchange indicates that in the range of perfusion flows from
4 to 20 ml min~l the gill may account for as much as 30 to 60% of the total
heat exchange in the animal.

THE VASCULARITY OF TROUT GILLS

      The final  project which received financial support from Grant R-801034
 was work which  was an outgrowth of studies on circulation in trout gills in-
 jected with latex and examined after acid digestion (Richards and Fromm,
 1969).  More recently Microfil,  a vulcanizing liquid silicone compound that
 flows through vessels of capillary size much more readily than latex, became
 available.   The following is a summary of observations made on gill circula-
 tion in specimens infused with this compound.

 Methods

      Rainbow trout were anesthetized with MS-222 and placed ventral side up
 in a V-shaped trough, and an incision was made to expose the heart and bul-
 bus arteriosus.  Freshly mixed Microfil (2.5 ml MV 112 white, 2.0 ml MV di-
 luent, and 0.135 ml curing agent)(Canton Bio-Medical Products, Inc., Boulder,
 Colo.) was drawn into a 5 ml syringe and infused via a polyethylene cannula
 (PE 90) that had been inserted and secured into the ventral aorta.  Gentle
 pressure on the syringe plunger was applied manually and excessive disten-
 tion of the ventral aorta avoided by extending the infusion over a period of
 12-15 min.  The infusion pressure was not recorded.  About 5 ml Microfil was
 injected into fish weighing around 200 g, a much larger volume than that re-
 quired to fill the gill vasculature.  Significant amounts passed through the
 gills and appeared in vessels of the musculature and the splanchnic area.
 After injection, the specimens were stored in a refrigerator overnight to
 allow complete polymerization of the Microfil.  They were then cleared in
 methyl salicylate (Merck), and selected casts with adhering cleared tissue
 were submerged in methyl, salicylate and photographed (see Figure 19).

-------
Figure 19.  Microfil cast of a single pair of gill filaments from rainbow
trout.  AG, afferent gill vessel; AF, afferent filamental vessel; EG,  effer-
ent gill vessels; EF, efferent filamental vessel;  B,  afferent filamental
blebs; SL, secondary lamellae.  Calibration bar =  1 mm.
                                      45

-------
Results and Discussion

     The trout circulatory system from the ventricle to the secondary lamel-
lae  fit the classical description except that the efferent gill vessel  (EG)
was  found  to leave the arch a short distance below the upper limit of the
gill,  leaving a cul-de-sac extension to drain blood from vessels of the more
dorsally located filaments.  Many vessels that arise from the base of the
afferent filamental or from the afferent gill vessel were observed in the in-
terfilamental septal area.  Vessels on either side of the septum were seen to
loop toward the base of the filament and then leave the septal area and run
parallel to the afferent filamental vessel.   Although it is not as clearly
visible, a second somewhat larger vessel, distal to the bleb, also ran par-
allel  to the afferent filamental vessel toward the tip of the filament.  It
appeared as if this vessel might provide cross connections between afferent
lamellar vessels at the point where they empty into the space surrounding
the pillar cells.  Another septal vessel of smaller diameter than the effer-
ent filamental vessels was found in several but not all gill arches.  This
vessel was located in the basal portion of the septal area and ran parallel
to the afferent gill blood vessel.  Several additional short intraseptal vas-
cular  connections between successive pairs of filaments on a gill arch were
also noted.

     Perhaps the most conspicuous feature of the gill casts was the presence
of blebs.  These aneurysm-like enlargements of the afferent filamental ves-
sels were located in the most distal part of the septal area, roughly one-
third of the distance from the base to the tips of filaments.  The casts ap-
peared as if those on opposing filaments would make contact with one another
when fully distended but no vascular connections between blebs were found.

     It appears presumptive that the blood vessels in the interfilamental
septal area and those that run parallel to the afferent filamental vessels
are not involved in gas exchange.  The function of the filamental blebs, on
the other hand, is even more speculative.

     In these experiments we were aware that perfusion pressure during cast-
ing was rather critical, especially with respect to flow into the central
filament space between afferent and efferent filamental vessels.  It was not
practical to record perfusion pressure but care was taken to avoid excessive
distension of the ventral aorta during infusion.  There was no evidence of
the application of excessive pressure,  e.g., rupture of gill vessels; thus,
we believe that the casts represent fairly accurately the size and shape of
vessels and blebs in trout gills.

     Using a dissecting microscope and intense direct illumination we have
observed a turbulent, pulsatile flow of blood into the bleb area of rainbow
trout gills in situ and the frequency of this flow into the blebs was found
to correlate perfectly with a simultaneous recording (EGG) of heart rate.
Also when a large nerve carrying fibers to an isolated-perfused gill arch
was stimulated, adduction of the filaments occurred, accompanied by an in-
crease of about 15% in the perfusion pressure.  When stimulation was exten-
ded for as long as 60 sec the increased pressure rapidly subsided, lasting
                                      46

-------
only about 15-20 sec.  It is tempting to speculate that contraction of the
adductor muscles may have increased tissue pressure surrounding the blebs
and facilitated their emptying.  If rhythmical isotonic or isometric con-
traction of these muscles occurred synchronously with heart beat, the pump-
ing effect would be analogous to that of gill hearts that aid in forcing
blood through the gill vasculature.  Unfortunately, this idea is the obser-
vation by Bijtel (1949) that adduction movement of the filaments (which might
not occur with isometric contraction of adductor muscles) appears sporadi-
cally and then only in connection with coughing movements.
                                     47

-------
                                 REFERENCES

Backstrom, J. 1969.  Distribution studies of mercuric pesticides in quail and
some fresh-water fishes.  Acta Pharmacol. Toxicol. 27:5-103.

Bennett, C.O. and J.E. Myers. 1962.  Momentum, heat and mass transfer.
McGraw-Hill Book Co., Inc., New York.

Bijtel, J.H. 1949.  The structure and mechanism of movement of the gill-fil-
aments in teleosts.  Arch. Neerl. Zool. 8:267-288.

Di Benedetto, F.E. and A. Farmanfarmaian. 1975.  Intestinal absorption of
naturally occurring sugars in the rainbow trout, Salmo gairdneri, under free-
swimming conditions.  Comp. Biochem. Physiol. 50:555-559.

Fromm, P.O. 1963.  Studies on renal and extra-renal excretion in a freshwater
teleost, Salmo gairdneri.  Comp. Biochem. Physiol. 10:121-128.

Fry, F.E.J. 1967.  Responses of vertebrate poikilotherms to temperature.
In;  Thermobiology, Edited by A.H. Rose.  Academic Press, New York.  pp. 375-
409.

Halbhuber, K.J. von, H.J. Stibenz, U. Halbhuber, and G. Geyer. 1970.  Auto-
radiographische Untersuchungen iiber die Verteilung einiger Metallisotope im
Darm von Laboratorimstieren.  Bin Beitrag zur Ausscheidungsfunken der Pan-
ethschen Kornerzellen.  Acta Histochem. 35:307-319.

Bannerz, L. 1968.  Experimental investigations on the accumulation of mer-
cury in water organisms.  Rep. Inst. Freshwater Res. Swed. 48:120-176.

Hevesy, G., D. Lockner, and K. Sletten. 1964.  Iron metabolism and erythro-
cyte formation in fish.  Acta Physiol. Scand. 60:256-266.

Hodgman, C.D. (Editor-in-chief). 1957.  Handbook of Chemistry and Physics,
39th Ed. Chemical Rubber Pub. Co., Cleveland,  p. E-4.

Holeton, G.F. and D.J. Randall. 1967.  The effect of hypoxia upon the par-
tial pressure of gases in the blood and water afferent and efferent to the
gills of rainbow trout.  J. Exp. Biol. 46:317-327.

Kay, J.M. 1963.  An Introduction to Fluid Mechanics and Heat Transfer, 2nd
Ed.  Cambridge Univ. Press, Cambridge.  1963.  425 p.

Keys, A. and J.B. Bateman. 1932.  Branchial responses to adrenaline and pi-
tressin in the eel.  Biol. Bull. 63:327-336.
                                     48

-------
Maetz, J. and E. Skadhauge. 1968.  Bringing rates and gill ionic turnover in
relation to external salinities in the eel.  Nature 217:371-373.

McKim, J.M., G.M. Christensen, and E.P. Hunt. 1970.  Changes in the blood of
brook trout after short term and long term exposure to copper.  J. Fish.
Res. Bd. Canada 27:1883-1889.

McKone, C.E., R.G. Young, C.A. Bache, and D.J. Lisk. 1971.  Rapid uptake of
mercuric ion by goldfish.  Environ. Sci. Technol. 5:1138-1139.

Miettinen, J.K., M. Heyraud, and S. Keckes. 1970.  Mercury as a hydrospheric
pollutant.  II.  Biological half-time of methyl mercury in four Mediterran-
ean species:  A fish, a crab, and two molluscs.  Presented at FAO Technical
Conference on Marine Pollution and Its Effects on Living Resources and Fish-
ing, Rome.  15 p.

	, M. Tillander, K. Rissanen, V. Miettinen, and T. Ohmomo.
1969.  Distribution and excretion rate of phenyl and methyl mercury nitrate
in fish, mussels, molluscs and crayfish.  Proceedings of the 9th Japan Con-
ference on Radioisotopes.  Atomic Energy Soc.  Jap., Tokyo.  10 p.

Miettinen, V., E. Blankenstein, K. Rissanen, M. Tallander, J.K. Miettinen,
and M. Valtonen. 1970.  Preliminary study of the distribution and effects
of two chemical forms of methyl mercury in pike and rainbow trout.  Presen-
ted at FAP Technical Conference on Marine Pollution and Its Effects on Liv-
ing Resources and Fishing, Rome.  9 p.

Morgan, M. and P.W.A. Tovell. 1973.  The structure of the gill of the trout,
Salmo gairdneri (Richardson).  Z. Zellforsch. 142:147-162.

Nakano, T. and N. Tomlinson. 1967.  Catecholamine and carbohydrate concen-
trations in rainbow trout (Salmo gairdneri) in relation to physical distur-
bance.  J. Fish. Res. Bd. Canada 24:1701-1715.

Ostlund, E. and R. Fange. 1962.  Vasodilation by adrenaline and noradrena-
line and the effects of some other substances on perfused fish gills.   Comp.
Biochem. Physiol. 5:307-309.

Prosser, C.L. (Editor). 1973.  Comparative Animal Physiology, 3rd edition,
W.B. Saunders Co., Philadelphia.  966 p.

Randall, D.J., B, Baumgarten, and M. Malyusz. 1972.   The relationship be-
tween gas and ion transfer across the gills of fish.  Comp. Biochem.  Physiol.
41A:629-637.

Randall, D.J. and J.N. Cameron. 1973.  Respiratory control of arterial pH
as temperature changes in rainbow trout Salmo gairdneri.   Am. J. Physiol.
225:997-1002.
                                      49

-------
Rankin, J.C. and J. Maetz. 1971.  A perfused teleostean gill preparation:
Vascular actions of neurohypophysial hormones and catecholamines.  J. Endocr.
51:621-635.

Reite, O.B. 1969.  The evolution of vascular smooth muscle responses to hista-
mine and 5-hydroxytryptamine.  I.  Occurrence of stimulatory actions in fish.
Acta Physiol. Scand. 75:221-239.

Richards, B.D. and P.O. Fromm. 1969.  Patterns of blood flow through fila-
ments and lamellae of isolated-perfused rainbow trout (Salmo gairdneri) gills.
Comp. Biochem. Physiol. 29:1063-1071.

Schweiger, G. 1957.  Die Toxikologische Einwirkung von Schwermetallizalzen
auf Fische und Fischnahrtiere.  Arch. f. Fishchereiwiss. 8:54-77.

Shuttleworth, T.J. 1972.  A new isolated perfused gill preparation for the
study of the mechanisms of ionic regulation in teleosts.  Comp. Biochem.
Physiol. 43A:59-64.

Sokal, R.R. and F.J. Rolf. 1969.  Biometry.  W.H. Freeman Co., San Francisco.
776 p.

Steen, J.B. and A. Kruysse. 1964.  The respiratory function of teleostean
gills.  Comp. Biochem. Physiol. 12:127-142.

Stevens, E.D. and F.E.J. Fry. 1970.  The rate of thermal exchange in a tele-
ost Tilapia mossambica.  Can. J. Zool. 48:221-226.

Stevens, E.D. and D.J. Randall. 1967a.  Changes of gas concentrations in
blood and water during moderate swimming activity in rainbow trout.  J. Exp.
Biol. 46:329-337.

	 and 	. 1967b.  Changes in blood pressure, heart rate
and breathing rate during moderate swimming activity in rainbow trout.  J.
Exp. Biol. 46:307-315.

Wood, C.M. 1974.  A critical examination of the physical and adrenergic fac-
tors affecting blood flow through the gills of the rainbow trout.  J. Exp.
Biol. 60:24-265.
                                      50

-------
                                PUBLICATIONS

Olson, K.R., H.L. Bergman, and P.O. Fromm. 1973.  Uptake of methyl mercuric
chloride and mercuric chloride by trout:  A study of uptake pathways into
the whole animal and uptake by erythrocytes in vitro.  J. Fish. Res. Bd.
Canada 30:1293-1299.

Olson, K.R. and P.O. Fromm. 1973.  Mercury uptake and ion distribution in
gills of rainbow trout (Salmo gairdneri):   Tissue scans with an electron
microprobe.  J. Fish. Res. Bd. Canada 30:1575-1578.

Olson, K.R. and P.O. Fromm. 1973.  A scanning electron microscopic study of
secondary lamellae and chloride cells of rainbow trout (Salmo gairdneri)  Z.
Zellforsch. 143:439-449.

O'Connor, D.V. and P.O. Fromm. 1974.  The effect of methyl mercury on gill
metabolism and blood parameters of rainbow trout.  Bull. Env. Cont. Tox. 13:
406-411.

Bergman, H.L., K.R. Olson, and P.O. Fromm. 1974.  The effects of vasoactive
agents on the functional surface area of isolated-perfused gills of rainbow
trout.  J. Comp. Physiol. 94:267-286.

Fromm, P.O. 1974.  Circulation in trout gills:  presence of "blebs"  in af-
ferent filamental vessels.  J. Fish. Res.  Bd. Canada 31:1793-1796.

Walker, R.L. and P.O. Fromm. 1976.  Metabolism of iron by normal and iron
deficient rainbow trout.  Comp. Biochem. Physiol. 55A:311-318.

Sorenson, P.R. and P.O. Fromm. 1977.  Heat transfer characteristics of iso-
lated-perfused gills of rainbow trout.  J. Comp. Physiol. 8112:345-357.
                                     51

-------
                                    INDEX
Acetylcholine, effect on
  permeability, 28
  perfusion pressure, 28
  heat exchange, 37
Acknowledgments, ix
blood parameters, 13, 22
14C-Urea, 24 et seq.
electron microscopy, 14
electron microprobe, 10
EPA 2220-1 form, 53
Epinephrine, effect on
  heat exchange, 35
  permeability 27, 28, 30
experimental animals, 1, 10, 12
gills
  circulation in, 44, 46
  functional surface area, 25, 31
  ion distribution in, 11
  oxygen consumption, 12
gill cannulation, 26
gill perfusion, 25, 27
heat exchange in fish, 33
hematological parameters, 22
holding facilities, 1
Introduction, 1
iron
  uptake by fish, 13 et seq.
  tissue content, 18
  deficiency, 17
  metabolism of, 23
mercury
  distribution in tissues, 6, 11
  uptake into fish, 6
  methyl form, 6
  inorganic, 7
  uptake by RBC's, 8
  uptake by gills, 10
mercury, effect on
  metabolism, 14
  blood parameters, 13
  gill ultrastructure, 15
Microfil, 44
nor-epinephrine, 28
Personnel,  ix
phenoxybenzamine, 28
plasma
  electrolytes, 13
  clearance of ^Fe, 13
propranolol, 28
Publications, 51
References, 48, 49, 50
statistics, 27
tissue, fixation of, 15
whole head cannulation, 34
                                 52

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 1. REPORT NO.
   EPA-6QQ/3-77-Q57
                                                           3. RECIPIENT'S ACCESSION-NO.
 4. TITLE AND SUBTITLE
  TOXIC  EFFECT OF WATER SOLUBLE POLLUTANTS ON
  FRESHWATER FISH
              5. REPORT DATE
                May 1977 issuing  date
              6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
  Paul  0.  Fromm
             8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Department of Physiology
  Michigan State University
  East  Lansing, Michigan  48824
                                                            10. PROGRAM ELEMENT NO.
              11. CONTRACT/GRANT NO.
                Grant No. R-801034
 12. SPONSORING AGENCY NAME AND ADDRESS
  Environmental Research Laboratory - Duluth, MN
  Office of Research and Development
  U.S.  Environmental Protection  Agency
  Duluth.  Minnesota  55804	
              13. TYPE OF REPORT AND PERIOD COVERED
               Final 3-1-72 through  6-3Q-JZ5.
              14. SPONSORING AGENCY CODE

                EPA/600/03
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
  Studies  of the effect of inorganic  and organic mercury on  trout indicated that uptake
  was  primarily via the gills  in  non-feeding fish.  Organic  mercury entered fish at a
  faster rate than inorganic mercury.   Exposure of trout to  10 yg Hg/1 (methyl form)
  haa  no effect on the gill oxygen  consumption measured in vitro or on the plasma
  electrolytes.  The hematocrit index increased significantly.   Studies of the
  metabolism of iron by normal and  iron deficient trout (made  deficient by bleeding)
  indicated that the liver, spleen, and head kidney are the  major iron storage organs.
  Liver iron was reduced by bleeding  whereas splenic iron was  unaffected.  In iron
  deficient fish more radioiron appeared in erythrocytes than  in normal controls.
  Studies  of isolated-perfused gills  revealed the presence of  both a and gadrenergic
  receptors and the data obtained indicate the functional surface area of trout gills
  can  be regulated by changes  in  perfusion pathway through the gills.   Use of perfused
  gills appears to be a very sensitive model to detect deleterious action of pollutants
  on fish.   Evaluation of heat exchange in perfused gills indicates that the presence
  of epinephrine increased the transfer maximum of the gill  but they were unaffected by
  the  administration of acetylcholine.   Analysis of a simple model indicated that the
  gills may account for as much as  60% of the total heat exchanged by trout.
 7.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS  C. COSATI Field/Group
  Trout
  Mercury
  Iron
  Metabolism
  Blood  Analysis
  Perfused  fish gills
  Iron metabolism
   06A
   060
   06T
 8. DISTRIBUTION STATEMENT

   RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
  UNCLASSIFIED
21. NO. OF PAGES
  63
                                              20. SECURITY CLASS (Thispagef
                                                UNCLASSIFIED
                           22. PRICE
EPA Form 2220-1 (9-73)
                                            53
              ft U.S. GOVERNMENT PRINTING OFFICE: 1977-767-066/6474

-------